Skip to main content

Full text of "DTIC ADA379966: Seventh Annual Report - AEC Fuels and Materials Development Program"

See other formats


73^ I •"'^7 

GEMP-1004 



SEVENTH ANNUAL REPORT - 

AEC FUELS AND MATERIALS DEVELOPMENT PROGRAM 








- LEGAL NOTICE- 

United States Atomic Energy Commission Contract No. AT (40-1J-2847 

This report was prepared as an account of Government sponsored work. Neither the 
United States, nor the Commission, nor any person acting on behalf of the Commission: 

A. Makes any warranty or representation, expressed or implied, with respect to the 
accuracy, completeness, or usefulness of the information contained in this report, 
or that the use of any information, apparatus, material, method, or process disclosed 
in this report may not infringe privately owned rights; or 

B. Assumes any liabilities with respect to the use of, or for damages resulting 
from the use of any information, apparatus, material, method, or process disclosed 
in this report. 

As used in the above, "person acting on behalf of the Commission" includes any 
employee or contractor of the Commission, or employee of such contractor, to the 
extent that such employee or contractor of the Commission, or employee of such 
contractor prepares, disseminates, or provides access to, any information pursuant 
to his employment or contract with the Commission or his employment with such 
contractor. 


Printed in the United States of America 
Available from 

Clearinghouse for Federal Scientific and Technical Information 
National Bureau of Standards, U.S. Department of Commerce 
Springfield, Virginia 22151 
Price: Printed Copy $3.00; Microfiche $0.65 


DISTRIBUTION 


US Atomic Ener g y Commission 
Washin g ton Headquarters 

G. W. Cunningham M. A. Rosen 

G. K. Dicker F. C. Schwenk 

D. E. Erb W. F. Sheely 

J. S. Griffo J. M. Simmons (2) 

H. G. Hembree R h. Steele (NR) 

J. R. Hunter s. A. Szawlewicz 

C. E, Johnson A Van Frhn 

E. E. Kintner j w Vaughn ^ 

J. A. Lieberman E s Wilson 

J. S. Powers Librarian, Germantown 

Chicago Operations Office 

M. E. Jackson, Senior Site Rep., ANL 
FFTF Pro j ect Office , PNL 
E. R. Astley 
Idaho O p erations Office 

D. A. Moss 

K. A. Trickett, Senior Site Rep., PNL 
LMFBR Pro g ram Office, ANL 

A. Amorosi 
J. R. Humphreys 

L. K. Kelman 

Oak Rid ge Op erations Office 
(Laboratory & Univ Div) (1) 

D. F. Cope, Senior Site Rep., ORNL 

C. L. Karl, CAO 

J. F. Weissenberg, GE-NMPO Site Rep. 
Richland O p erations Office 
P. G. Hoisted, Senior Site Rep., PNL 
J. H. Sako 

San Francisco Operations Office 

R. H. Ball, Senior Site Rep., GA, 

San Diego, California 
A. W. Larson, Senior Site Rep., GE, 
Sunnyvale, California 

R. L. Morgan, Senior Site, Rep., 

Canoga Park, California 
DUE (20) 

Air Force Materials Laborator y 

H. M. Burte 

G. Glenn 

Technical Director 
Air Force Wea p ons Laborator y 
Lt. D. Brooks 
Ames Laborator y 

O. N. Carlson 
W. L. Larsen 

F. H. Spedding 

Argonne National Laboratory 

L. Baker 
J. H. Kittel 

M. V. Nevitt 

P. Shewman 

R. C. Vogel 

Army Materials & Mechanics Research Center 

S. V. Arnold, Asst for Technology 
Atomics International 

S. C. Carniglia 

H. Pearlman 


Babcock and Wilcox Com pany_ 

C. Baroch 

C. Johnson 
J. Landis 
L. Weissert 

Battelle Memorial Institute, Columbus, Ohio 
W. Berry 

D. Keller 

S. Paprocki 
DMIC 

Brookhaven National Laborator y 

D. H. Gurinsky 

Bureau of Mines, Albany, Oregon 
H. Kato 

Combustion En g ineerin g 
W. P. Chernock 
S. S. Christopher 

E. I. duPont de Nemours and Co., SRL 
P. H. Permar 
S. Rideout 

General Atomic , Div of General Dynamics 

S. Jaye 

T. Watson 

General Electric Com pany 

APQ - Sunnyvale , California 

K. P. Cohen 
F. Comprelli 
R. Skavdahl 
C. Spalaris 

E. L. Zebroski 
KAPL 
A. E. Bibb 
Library 

Idaho Nuclear Co rp oration 
W. C. Francis 
Jet Pro p ulsion Laborator y 
J. Davis 

Lawrence Radiation Laborator y 
J. S. Kane 

L. W. Roberts 

A. J. Rothman 

B. Rubin 

Los Alamos Scientific Laborato ry_ 

R. D. Baker 
D. E. Hall 
J. A. Leary 
R. Perkins 
Mound Laborator y 

L. J. Wittenberg 
NASA Headquarters 

J. J. Lynch 

NASA Lewis Research Center 

M. Ault 

J. W. Creigh 

T. P. Moffitt 

T. A. Moss 

F. E. Rom 
L. Rosenblum 

N. D. Sanders 

N. Saunders 
H. Schwartz 


Naval Air Systems Command 

I. Machlin 

Naval Research Laborator y 
R. Hawthorne 

C. Serpan 

Nuclear Materials and E quip ment Cor p oration 

C. Caldwell 

K. Puechl 

Oak Rid g e National Laborator y 

G. E. Boyd 

J. E. Cunningham 
J. H. DeVan 
W. R. Grimes 
W. 0. Harmes 
W. R. Martin 
P. Patriarca 
J. Scott 
J. R. Weir 

Ohio State Universi ty 
M. Fontana 

Pacific Northwest Laborator y 

F. W. Albaugh 
A. L. Bement 
J. J. Cadwell 

D. R. deHalas 
R. L. Dillon 

E. Evans 

J. C. Tobin 
R. G. Wheeler 
Sandia Corporation 
J. Jacobs 
R. P. Stromberg 

TRW Space Technolo g y Laborator y 

H. Lurie 

United Nuclear Cor p oration 
A. Strasser 

University of California, Berkeley 

L. Brewer 

US Naval Air Systems Command (AIR-52031B) 
T. Kearns 

I. Machlin 

Westin g house Electric Cor p oration 
Advanced Reactor Division 

E. C. Bishop 

J. C. R. Kelly, Jr. 

W..E. Ray 

Astronuclear Laborator y 
R. T. Begley 

D. C. Goldberg 
Atomic Power Division 

R. Allio 
T. Stern 
Bettis 

W. J. Babyak 
R. H. Fillnow 

E. J. Kreigh 
B. Lustman 


D. H. Ahmann 


P. K. Conn 

J. Moteff 

J. E. VanHoomissen, NTPO 

N. K. Anderson (2) 


J. B. Conway (2) 

G. T. Muehlenkamp 

G. R. VanHouten 

W. G. Baxter 


E. W. Filer 

W. E. Niemuth 

H. E. Wagner 

J. T. Berling 


P. N. Flagella 

G. W. Pomeroy 

J. F. White 

J.J. Brady, FPD 


R. E. Fryxeli 

R. B. Richards, APED 

V. C. Wilson, R&DC 

H. C. Brassfield 


E. S. Funston 

F. C; Robertshaw 

R. E. Wood 

M. L. Bromberg 

R. D. Brooks 

V. P. Calkins 

L. Chockie, APED 

C. G. Collins (2) 

E. S. Collins 

J. F. Collins 


J. 0. Hibbits 

E. F. Juenke 

F. Kingsbury 

W. C. Kuhlman 

L. R. McCreight, MSD 

J. A. McGurty 

J. W. Morfitt 

C. S. Robertson 

E. J. Schmidt, ATS 

L. H. Sjodahl 

J. P. Smith 

H. R. Stephan 

C. 0. Tarr 

F. 0. Urban 

Library (20) 


3 



4 


* i ♦ 


i 


♦ 


3 > 


PREFACE 


This report, GEMP-1004, is the seventh annual report of the unclassified portion of the 
Fuels and Materials Development Programs being conducted by the General Electric 
Company's Nuclear Materials and Propulsion Operation under Contract AT(40-l)-2847, 
issued by the Fuels and Materials Branch, Division of Reactor Development and Tech¬ 
nology, of the Atomic Energy Commission. 

This report covers the period from January 31, 1967 to January 31, 1968, and thus also 
serves as the quarterly progress report for the final quarter of the year. During this 
period, two unclassified programs were terminated and the title and objectives of others 
were revised. The title change and status of each program is indicated below in the more 
detailed breakdown of this unclassfied annual report, GEMP-1004, and the classified 
annual report, GEMP-1003, which covers the period from January 1, to December 31, 1967. 

GEMP-1004,. Unclassified 

1. Physical and Mechanical Properties of Reactor Materials (same title). Task 1503, 
continuation. 

2. Radiation Effects on Fast Breeder Reactor Cladding and Structural Materials (same 
title), Task 1304, continuation. 

3. Advanced Fast Breeder Reactor Fuel Element Cladding Development (Refractory- 
Metal Alloy Research and Development),* Task 1115, continuation. 

4. Physical Metallurgy of Fast Breeder Reactor Cladding Materials and Refractory 
Metals (Physical Metallurgy of Refractory-Metal Alloys), Task 1177, continuation. 

5. Fast Breeder Reactor Fuel Element Cladding Research (Advanced Long-Life Reactor 
Fuel Cladding and Structural Materials Development), Task 1119, continuation. 

6. Evaluation of Plastic Fatigue Properties of Heat Resistant Alloys (same title), 

Task 1516, continuation. 

7. Advanced Pressure Vessel Materials (same title), Task 1521, continuation. 

8. Physico-Chemical Studies of Clad UO 2 in Potential Meltdown Environments (same 
title), Task 1175, continuation. 

9. Fast Breeder Reactor Thermocouple Development (High-Temperature Thermocouple 
and Electrical Materials Research), Task 1414, continuation. 

10. Physico-Chemical Studies of Fe-Cr-Al-Clad Fuel Systems (same title), Task 7076, 
terminated FY-67. 

11. High-Temperature Research on Carbides for Fuel and Structural Applications (same 
title). Task 7073, terminated FY-67. 

"The titles presented in parenthesis are the ones used in the previous quarterly reports (GEMP-1001 and -1002). 


4 



4 


* t 


% 


1 


i 


* / 


' > 


GEMP-1003, Classified 

1. High-Temperature Studies of Uranium Based Fuels (High-Temperature Studies of 
Substoichiometric Uranium and Uranium Solid Solutions), Task 1171, continuation. 

2. Refractory-Metal Fuel Element Materials Research (same title), Task 1105, con¬ 
tinuation. 

3. Fission Product Transport Processes in Refractory Metal Fuel Elements (same 
title), Task 1170, continuation. 

4. High-Temperature Materials Engineering Properties Evaluation (same title), Task 
7017, Terminated FY-1967. 

5. Fast Breeder Reactor Control, Shield, and Reflector Materials Development (Ad¬ 
vanced Long-Life Reactor Fuel Element Moderator, Control, and Shield Materials 
Development), Task 1220, continuation. 





3 


* > 


\ 

% 






CONTENTS 


Page 


INTRODUCTION AND SUMMARY. 11 

s/l. PHYSICAL AND MECHANICAL PROPERTIES OF REACTOR MATERIALS 

(1503) . 14 

1.1 CREEP-RUPTURE STUDIES. 14 

Tungsten . 14 

Tungsten-Base Alloys. 37 

Molybdenum. 41 

Rhenium . 58 

Niobium and Niobium Alloys. 63 

Constant-Stress Creep Testing . 65 

Single Crystals. 71 

Stress-Rupture Parameter Analysis. 71 

Comparison of Test Data. 79 

1. 2 THERMAL PROPERTY EVALUATIONS. 81 

Electrical Resistivity and Thermal Conductivity. 81 

Thermal Diffusivity. 82 

Enthalpy. 84 

1. 3 SUMMARY AND CONCLUSIONS. 85 

1. 4 PLANS AND RECOMMENDATIONS. 87 

2. RADIATION EFFECTS ON FAST REACTOR CLADDING AND 

STRUCTURAL MATERIALS (1304) . 89 

2.1 STATUS OF IRRADIATIONS . 89 

EBR-II Irradiation Program. 89 

ORR and ETR Irradiations. 90 

2. 2 HEAT-RESISTANT ALLOY PROGRAM . 90 

Creep-Rupture Testing. 90 

Hot Hardness . 100 

Resistivity Studies . 103 

Transmission Electron Microscopy. 107 

2. 3 REFRACTORY METALS AND ALLOYS PROGRAM. 113 

Creep-Rupture Testing. 113 

Hot Hardness . 131 

Hardness and Ultimate Strength Correlation. 134 

Tensile Testing . 137 

Resistivity Studies. 147 

Transmission Electron Microscopy . 154 

2. 4 REACTOR DOSIMETRY. 165 

Monte Carlo Spectrum Measurements. 165 

EBR-II Flux Density Measurements. 165 

2. 5 SUMMARY AND CONCLUSIONS. 171 

2. 6 PLANS AND RECOMMENDATIONS. 173 

2. 7 REFERENCES . 174 


6 











































Page 


^3. ADVANCED FAST BREEDER REACTOR FUEL ELEMENT CLADDING 

DEVELOPMENT (1115). 118 

3.1 REFRACTORY METAL ALLOY TUBING, SHEET, AND 

WIRE PRODUCTS . 118 

Seamless Tubing . 180 

Sheet. 181 

Shapes . 183 

Refractory Metal Alloy Wire Drawing. 183 

3. 2 DEVELOPMENT OF MOLYBDENUM AND ITS ALLOYS 

FOR FAST BREEDER REACTOR APPLICATIONS . 186 

Molybdenum Purification. 188 

Molybdenum Alloys. 193 

W-Re-Mo Alloys . 194 

3. 3 SUMMARY AND CONCLUSIONS . 194 

3.4 PLANS AND RECOMMENDATIONS. 196 

*/4. PHYSICAL METALLURGY OF FAST BREEDER REACTOR CLADDING 

MATERIALS AND REFRACTORY METALS (1111). 198 

4.1 TUNGSTEN-RHENIUM-MOLYBDENUM ALLOYS. 198 

Fabrication. 198 

Recrystallization .. 199 

Aging. 202 

Ductility. 208 

4. 2 MOLYBDENUM . 210 

Fabrication. 212 

Recrystallization. 212 

Ductility. 212 

Aging Studies . 214 

4. 3 FAST REACTOR FUEL CLADDING ALLOYS . 214 

4. 4 SUMMARY AND CONCLUSIONS. 214 

4. 5 PLANS AND RECOMMENDATIONS. 215 

/5. FAST BREEDER REACTOR FUEL ELEMENT CLADDING RESEARCH, (1119) . . 216 

5.1 CHROMIUM-BASE ALLOYS. 216 

Material Preparation . 211 

Material Evaluation. 219 

5. 2 Fe-Cr-Al-Y ALLOYS .. 230 

Material Preparation . 231 

Material Evaluation. 231 

5. 3 IRON-BASE (FERRITIC). 238 

5.4 IRON-BASE (AUSTENITIC) . 238 

5. 5 SUMMARY AND CONCLUSIONS. 238 

5. 6 PLANS AND RECOMMENDATIONS. 241 

yfi 5. EVALUATIONS OF PLASTIC FATIGUE PROPERTIES OF 

HEAT-RESISTANT ALLOYS (1516). 242 

6.1 MATERIALS SPECIFICATIONS. 242 

AISI 304 Stainless Steel Rod Stock. 242 

AISI 316 Stainless Steel Rod Stock. 242 

NMPO Processing — AISI 304 and 348 Stainless Steel . 242 

NMPO Processing — AISI 316 Stainless Steel. 242 

6.2 FATIGUE-TESTING EQUIPMENT. 243 


7 















































Page 


6. 3 TEST RESULTS. 

6. 4 FATIGUE DATA ANALYSIS . 

6. 5 METALLOGRAPHIC AND FRACTOGRAPHIC ANALYSIS OF 

LOW-CYCLE FATIGUE SPECIMENS . 

6. 6 SUMMARY AND CONCLUSIONS. 

. 6. 7 PLANS AND RECOMMENDATIONS. 

#7. ADVANCED PRESSURE VESSEL MATERIALS (1521). 

7.1 EXPERIMENTAL PROGRAM . 

12Ni — 5Cr — 3Mo Maraging Steel . 

PH13-8MO. 

Inconel Alloy 718. 

7. 2 CONSIDERATION OF PRESSURE VESSEL FAILURE MODES AND 

RELATED PROPERTIES. 

Burst Due to Over-Pressure . 

Fatigue Cracking. 

Fast Fracture by Either Brittle Fracture or Low-Energy Tearing . 

7. 3 GENERAL DISCUSSION . 

12Ni - 5Cr - 3Mo. 

PH13-8MO . 

Inconel Alloy 718.. 

7. 4 SUMMARY AND CONCLUSIONS. 

7. 5 PLANS AND RECOMMENDATIONS. 

>C PHYSICO-CHEMICAL STUDIES OF CLAD U0 2 IN POTENTIAL 

MELTDOWN ENVIRONMENTS (1175). 

8.1 DYNAMIC TESTING OF ZIRCONIUM-BASE ALLOYS. 

Internal Pressure Effects . 

Effect of Oxidation on Tube Deformation . 

Testing of 50-cm-Long Zircaloy-4-Clad UO 2 Fuel Elements . 

Dynamic Testing of Irradiated Zircaloy-2 at Constant Internal Pressure ... 

8. 2 REACTION MECHANISMS AND KINETICS . 

Oxidation of Zirconium Alloys by Steam or Air. 

Oxidation of Type 304 Stainless Steel in Steam or Air. 

Oxidation of UO 2 by Steam. 

8. 3 PHYSICAL AND MECHANICAL PROPERTIES. 

Tensile Strength of Type 304 L Stainless Steel. 

Thermal Conductivity of 304 L Stainless Steel. 

Spectral and Total Emittance Measurements of Oxidized Zircaloy-4. 

8. 4 SUMMARY AND CONCLUSIONS. 

8. 5 PLANS AND RECOMMENDATIONS. 

FAST BREEDER REACTOR THERMOCOUPLE DEVELOPMENT (1414). 

9.1 W VERSUS W - 25Re THERMOCOUPLE CHARACTERISTICS AT 

HIGH TEMPERATURE... 

Electrical Insulation Evaluation. 

Effect of Gaseous Environment . 

Sheathing Studies. 

Effect of Unmatched Thermocouple Wire Diameters. 

9. 2 ELECTRICAL INSULATION FOR HIGH-TEMPERATURE 

THERMOCOUPLES. 

Intrinsic Variables Affecting Thermocouple Voltage. 

8 


247 

252 

263 

265 

273 

275 

275 

275 

281 

289 

300 

301 
303 

309 

310 

311 
311 

313 

314 


316 

316 

316 

320 

324 

326 

326 

326 

333 

337 

342 

342 

343 
346 

349 

350 


351 

352 

352 

353 

353 

354 

355 













































Page 


9. 3 THERMOELECTRIC CHANGES IN W - 25Re DUE TO TRANSMUTATION... 358 
Reactor Stability of W Versus W — 25Re Thermocouple. 358 

Reactor Testing of Thermocouple Materials. 359 

9. 4 SUMMARY AND CONCLUSIONS. 359 

9. 5 PLANS AND RECOMMENDATIONS. 360 

PHYSICO-CHEMICAL STUDIES OF Fe-Cr-Al-CLAD FUEL SYSTEMS (7076)... 361 

10.1 DIFFUSION STUDIES ... 361 

Fe, Cr, and A1 Diffusion. 361 

Uranium Diffusion .. .. 362 

10.2 FUELED CAPSULE TESTS. 365 

10. 3 CONCLUSIONS. 367 

^ HIGH-TEMPERATURE RESEARCH ON CARBIDES FOR FUEL AND 

STRUCTURAL APPLICATIONS (7073). 369 

11.1 PREPARATION AND FABRICATION OF REFRACTORY CARBIDES. 369 

11.2 Ta CARBIDE. 371 

Thermal Stability. 371 

Effect of Strain on Lattice Parameters. 372 

11.3 NON-STOICHIOMETRY IN Ta AND U MONOCARBIDES. 373 

Experimental Procedure . 373 

Discussion . 376 

11.4 PLANS AND RECOMMENDATIONS. 378 


9 





















UNCLASSIFIED 


CONVERSION TABLE 


To Convert From 


To 


Multiply By 


Atmospheres .. 

Calories (mean). 

Calories/gram-°C .... 
Calories/sec-cm-°C .. 

Calories/ sec-cm 2 . 

Calories/sec-cm 2 -°C . 
Centimeters. 

Cubic centimeters .... 

Grams... 

Grams/cm 2 . 

Grams/cm 2 . 

Kilograms . 

Kilograms/cm 2 . 

Kilograms/mm 2 . 

Kilowatts . 

Liters. 

Meters .. 

Millimeters of mercury 
Square centimeters 

Torr. 

Watts/cm-°C. 

Watt-seconds. 

Watts/cm^ . 

Watts/cm 2 -°C. 

Centimeters/sec. 

Meters/sec .. 


Pounds/inch 2 . 

Btu (mean).... 

Btu/pound-°F.... 

Btu/hr-ft-°F... 

Btu/hr-ft 2 .. 

Btu/hr-ft 2 -°F. 

Feet. 

Inches . 

Cubic feet . 

Cubic inches. . . • „ 

Pounds . . 

Pounds/ft^. 

psi. 

Pounds . 

Atmospheres. 

Pounds/ft 2 . 

Pounds/inch 2 . 

Pounds/inch 2 . 

Btu/sec.*. 

Cubic feet . 

Inches .. 

Atmospheres. 

Square feet. 

Square inches. 

mm of Hg... 

Atmospheres. 

Btu/hr-ft-°F. 

Btu.... 

Btu/hr-ft 2 . 

Btu/hr-ft 2 -°F. 

Feet/sec... 

Feet/sec... 


14.7 

0.00397 

1.0 

241.8 

1.32 xlO 4 
7370 
0.03281 
0.3937 
3.531 x 10~ 5 
0.06103 

0.002205 

62.43 

0.01422 

2.205 

0.9678 

2048 

14.22 

1422.32 
0.948 
0.0353 
39.37 
0.001316 
0.001076 
0.155 
1.0 

0.001316 

57.8 

0.000948 

3170 

1760 

0.03281 

3.281 


10 

UNCLASSIFIED 





























































INTRODUCTION AND SUMMARY 


This report, GEMP-1004, is the seventh annual report on the unclassified portion of the 
GE-NMPO Fuels and Materials Development Program conducted during calendar year 1967 
under Contract No. AT(40-l)-2847. This report covers eleven unclassified jobs: (1) proper¬ 
ties of reactor materials from 1000° to 3000°C; (2) radiation effects on the time-, temper¬ 
ature-, and stress-dependent properties of fast breeder reactor (FBK) cladding and struc¬ 
tural materials; (3) fabrication of FBR advanced fuel element cladding; (4) physical metal¬ 
lurgy of FBR cladding materials and refractory metals; (5) development of advanced FBR 
fuel element cladding materials with improved performance capability; (6) parameters 
affecting low-cycle fatigue behavior of heat-resistant alloys; (7) applicability of high- 
strength steels to nuclear reactor pressure vessels; (8) behavior of Zircaloy-4-clad and 
Type 304 stainless steel-clad UC >2 in meltdown environments; (9) high-temperature 
thermocouple and electrical materials; (10) physico-chemical stability and reactions 
between Fe-Cr-Al alloys and UO 2 ; and (11) refractory carbides for fuel and structural 
applications. 

Significant results achieved in this program were as follows: 

From 1600° to 3000°C the creep-rupture data for arc-cast W was corrdatable in terms 
of diffusion compensated creep rate versus the modulus compensated stress. Above one- 
half the absolute melting temperature, creep is by dislocation climb or glide of jogged 
screw dislocations. Arc-cast W showed no tendency to form cavities in contrast to the 
grain boundary separation and cavitation observed in powder-metallurgy W. Creep-rupture 
data (1200° to 2400°C) for arc-cast Mo showed good correlation between time to rupture 
and linear creep rate except at high temperatures and low stress where a change in creep 
mechanism was indicated. Creep-rupture data from 1600° to 2600°C are presented for 
W — 30Re — 30Mo, Mo — 30W, W — 25Re, and Mo — 50Re. Single crystal material of SB* 
w at 2800°C showed no detectable creep 

with fracture being brittle in nature. A comprehensive study of stress-rupture parameters 
was completed. Thermal conductivity and electrical resistivity for W — 25Re to 2400°C, 
thermal diffusivity for W —25Re, 304L SS, and W from 300° to 1000°C, and enthalpies of 
W from 1200° to 3450°K are presented. 

Irradiation of Incoloy 800 to 3 x 10^0 n /cm^ (E n — 1 Mev) lowers rupture life at 540°C 
but increases it at 705°C with corresponding increase and decrease in minimum creep 
rate; ductility is decreased 1/3 to 1/2. Hastelloy X, irradiated in either ETR or EBR-II 
to comparable fast fluences have comparable creep-rupture properties. Hastelloy R-235 
with varying B-*-0 content at 50 ppm boron level showed increased radiation damage with 
increasing B 10 content when tested at 870°C. Hastelloy R-235 and A-286 show shells of 
localized damage and gas bubbles with radii approximating the recoil distance of Li and 
a-particles in Fe and Ni. Embrittlement of irradiated ASTM-A302B is probably caused by 
carbon-defect-complex formation which recovers at 300° to 500°C as evidenced by re¬ 
sistivity studies. 


11 


12 


Irradiation temperature has a pronounced effect on creep-rupture properties of Mo; for 
70°C irradiations the effect at 750°C is least; for 700° and 1000°C irradiations the time 
to rupture at 750°C was increased by factors of 12 and 18, respectively; annealing at 
1000°C of a 700°C irradiated specimen increased 750°C rupture life an additional 25 per¬ 
cent; delayed creep was observed at 600°, 700°, and 750°C for 70°C irradiations but not 
in control specimens, in 70°C irradiated specimens tested at 580°, 850°, or 900°C, or 
in 700°C and 1000°C irradiated specimens. Resistivity measurements indicated irradiation 
to saturate in Mo at about 1 x 10 2 ® n/cm 2 (E n s i Mev). Molybdenum, creep-rupture tested 
at 750°C, showed dislocation loops in matrix but not at grain boundaries in specimens ir¬ 
radiated at 70°C and 700°C; whereas, for the 1000° C irradiated Mo there were no matrix 
loops, with only occasional loops near grain boundaries. Irradiation of Nb and Nb — IZr 
at 70°C to 2.1 x 1020 n /cm 2 (E n s l Mev) increases the room-temperature tensile yield 
strength about 110 and 200 percent, respectively; the 600° to 650°C yield strength is in¬ 
creased by 93 and 110 percent, respectively; ductility is significantly reduced; and com¬ 
plete recovery occurs after 1000°C anneal. 

Facilities for producing refractory-metal sheet, bar, seamless tubing, and wire were 
put into operation and process procedures established. Purification procedures for molyb¬ 
denum resulted in sheet which was ductile and free of microporosity in weld-heat-affected 
areas. A Mo — 5W alloy, processed using purification procedures, had stress-rupture life 
and deformation resistance equal to Mo - 50Re at 1600°C and superior to that alloy at 2200°C. 

Lower-cost production procedures and improved room-temperature ductility properties 
were developed in a study of powder-metallurgy W - 30Re - 30Mo and W - 25Re - 30Mo 
alloys. Recrystallization, grain size, hardness, and ductility data for aging to 2000°C for 
up to 1000 hours are presented. Similar studies are being applied to Mo. 

Fe-Cr-Al-Y alloys and Cr- and Fe-base (ferritic and austenitic) alloys are compared to 
austenitic stainless steel mechanical properties. All program alloys were found to possess 
superior tensile yield strength up to 750°C. Of these, only Cr- and austenitic Fe-base alloys 
have indicated 650° C creep-rupture properties comparable or superior to stainless steel. 

Li low-cycle fatigue, Type 348 stainless steel exhibited better strain fatigue resistance at 
430°, 650°, and 816°C than Types 304 or 316 stainless steel. All three stainless steels 
showed cyclic strain hardening characteristics with the 316 stainless st.eel at 430°C ulti¬ 
mately showing cyclic strain softening. Results at 430°C for 348 and 316 stainless steel 
are in fair agreement with the Coffin-Manson and Manson-Halford relations. A hundred¬ 
fold decrease in strain rate caused, in general, a several-fold decrease in fatigue resis¬ 
tance of 304, 348, and 316 stainless steel at 650°C and 816°C. Fatigue curves based both 
on total axial and plastic strain are shifted to lower life with decrease in strain rate. Frac- 
tographic analyses indicate the crack, formed at 650°C, is approximately 1.7 mm in length 
at the N 5 point. The mode of crack initiation in 348 and 304 stainless steel at 650°C and 
816°C and strain rates of 4 x 10 - 3 to 4 x 10 - 5 sec”l was primarily intergranular except 
for the highest strain rate at 650°C. 

The potential of three high-strength materials for advanced pressure vessels is being 
evaluated. The 12Ni — 5Cr — 3Mo maraging steel possesses (1) aging characteristics which 
appear to be suited for aging of heavy sections, (2) promising tensile and toughness proper¬ 
ties, (3) adequate structural stability at least to 315°C, and (4) good weldability. The pre¬ 
cipitation-hardening stainless steel, PH13-8M0, at aging temperatures which produce good 
toughness (1) may not be suited to uniform aging of heavy sections, (2) has excellent tensile 
and good toughness properties, (3) is stable structurally to 425°C, and (4) has good welda¬ 
bility and weld strength. Inconel alloy 718 has (1) excellent strength, (2) relatively low 
toughness which was not significantly lowered when irradiated to 1 x 1019 n/cm 2 (E n =; 1 Mev), 
(3) good structural stability to 540°C, and (4) showed good weldability but with weld strength 


13 


and toughness below parent metal level. The types of failure which need to be prevented in 
pressure vessel applications and the special materials properties which are related (e. g., 
strain hardening coefficient, dynamic yield strength, and plane strain fracture toughness) 
are presented. 

Dynamic heating of Zircaloy-4 tubes with an internal pressure show that steam oxidation 
lowers deformation and deformation rates, and increases failure temperature; in the ab¬ 
sence of oxidation, tube failure has been expressed as a function of tensile strength and 
heating rate. Oxidation of Zircaloy-4 in air is greater than in steam and is attributed to 
reaction with nitrogen in the air. The spectral emittance (at 0.65 micron) of unoxidized 
Zircaloy-4 varied from 0. 56 at 885°C to 0.43 at 1550°C. Diffusion of oxygen in U02+ x lat¬ 
tice is the rate controlling mechanism for oxidation of U0 2 by steam. Up to 1370° C, the 
oxidation of 304L stainless steel in air is 10^ times less than in steam due to a protective 
01*203 film; above 1370°C, the rates are equal as molten FeO destroys the Cr 2 C >3 film. 

High-temperature thermocouple characteristics of W / W — 25Re are affected by the thermo 
electric emf produced along BeO, Hf02, or ThC >2 insulators and by the atmosphere for bare- 
wire (non-insulated) thermocouples. Irradiation of W versus W —25Re thermocouples for 
2 months in a 1 . 2 x 10 -^ neutron/cm^-sec thermal and 2.1 x 10*3 neutron/cm^-sec fast 
flux resulted in similar shifts in thermal emf, causing a small (5°C) positive error. 

Aluminum in Fe-Cr-Al-base claddings reduces UC >2 in cermet fuels above 1000°C. Free 
U formed dissolves in and diffuses through 0.38-mm cladding in 100 hours at 1000°C and 
in 6 hours at 1200° C. Liberated oxygen combines with A1 to form A^Og at the cladding — 
core interface where it inhibits subsequent diffusion of A1 and thus limits the quantity of 
U formed. 

The preparation and fabrication of well-characterized carbides of Ta, Zr, Hf; the thermal 
stability and effect of strain on lattice parameters of Ta carbide; and the non-stoichiometry 
in Ta- and U-monocarbides are summarized. 





<r> _,»• 

?s:i 





1A PHYSICAL AND lW yf.HANTflAI. PROPE RTIES 
OF REACTOR MATERIALS^ 


(1503) 

J. B. Conway,* P. N. Flagellat 


I The purpose of this program is to measure and evaluate high-temperature (to 3000°C) 
physical and mechanical properties of commercially available and newly developed ma¬ 
terials being considered for use in fueled and non-fueled high-temperature reactor 
applications. 

1.1 CREEP-RyPTURE STUDIES j 

A significant amount of creep-rupture data for wrought, unalloyed, tungsten and molyb¬ 
denum have been obtained over the last few years at temperatures from 1600° to 3000°C. 
Results have shown that wrought material, fabricated by the powder-metallurgy process, 
does not consistently exhibit the same strength (rupture life), creep resistance (linear 
creep rate), and ductility (elongation at rupture). Powder-metallurgy material was found 
to be characterized by considerable duplexing of the structure with a significant portion 
of deformation related to grain boundary separation or cavitation leading to fracture in 
the grain boundaries. 

Wrought, arc-cast material exhibited more consistent behavior. Extremely large grains 
formed in high-temperature tests and deformation occurred as crystallographic slip and 
grain boundary sliding. No void formation was observed in the grain boundaries. Arc-cast 
material was considerably more ductile with fracture being transgranular. 

\ TUNGSTEN j^ ^ (7 

Stress-Rupture and Creep 

Stress-rupture and creep data for wrought, unalloyed, arc-cast tungsten sheet, identi¬ 
fying code W(3), were obtained over a broad range of temperatures and stresses. The re¬ 
sults are summarized in Figures 1.1 and 1.2 and indicate a change in slope for both the 
stress-rupture and creep rate curves at 2200°C. The slope for the creep rate curve at 
low stresses is consistent with the isotherms for 2400°C and 2600°C. A similar change in 
slope is indicated for stress-rupture data at 2400°C but not at 2600°C. The mechanism 
associated with slope changes has not been identified. 

Creep Rate Correlation 

The creep data obtained for arc-cast tungsten, W(3), from 1600° to 3000°C were eval¬ 
uated based on the proposed method of Sherby 1 for pure polycrystalline metals above one- 
half the absolute melting temperature. The results are presented in Figure 1.3 in terms 
of the ratio of steady-state creep rate (e s ) to the diffusion coefficient (D) plotted as a 
function of creep stress (a) divided by the elastic modulus (E). Correlation of the data 

^Project leader. 

* Principal investigator. 

1 0. D. Sherby, “Factors Affecting the High Temperature Creep of Polycrystalline Solids," Acta Met., Vol. 10,1962, p. 135. 

14 


A 



¥ 


r 

15 




Time to rupture, hours 


Fig. 1.1 — Stress-rupture test results for wrought, arc-cast unalloyed tungsten, 
W{3), sheet tested in hydrogen 



Linear creep rate, min 1 

Fig. 1.2 - Creep rate test results for wrought, arc-cast unalloyed tungsten, W(3), 
sheet tested in hydrogen 


in this manner appears to be quite good with the curve shape the same as that determined 
by Sherby for some other pure polycrystalline metals. The slope, n, of the linear portion 
of the curve is 4.2 and is consistent with Sherby’s prediction that the value be approxi¬ 
mately 5. This portion of the curve is associated with a creep process controlled by dis¬ 
location climb involving an equilibrium vacancy concentration. The non-linear portion of 
the curve (higher stresses) involves dislocation climb under conditions where vacancy 
concentration is greater than the equilibrium value. 

^ Ductility 

Evaluations of rupture elongation, for both powder-metallurgy (PM) and arc-cast (AC) 
tungsten, were made to obtain the results shown in Figure 1. 4. (PM data are the result 
of testing three different vendor sources of material.) The PM sheet exhibited a decreas¬ 
ing trend in ductility with temperature from 1600° to 2800°C; the AC sheet showed a de¬ 
crease in ductility from 1600° to 1800°C followed by a significant increase, peaking at 
2200° to 2400°C, and then a decrease with further increasing temperature to 3000°C. Frac¬ 
ture in the PM sheet occurred only in the grain boundaries; no transgranular failures 


L 


J 






































































Elongation, percent 


16 







Fig. 1.3 - Ratio of steady-state creep rate (e s ) to diffusion coefficient (D) 
versus ratio of stress (a) to Young's modulus (E) for arc-cast W 



Fig. 1.4 - Elongation versus temperature for wrought powder-metallurgy and 
wrought arc-cast unalloyed W sheet tested in hydrogen 






17 


were observed. Failure in the AC sheet was principally transgranular with little or no 
grain.boundary fissuring observed. 

Although initial grain size appears to have an effect on the strength and ductility of 
these materials, the presence of many large grains after failure in the PM sheet indicates 
that some factor in addition to grain size must be exerting an effect. All fractures in the 
PM sheet occurred in grain boundaries indicating that the grain boundaries are weaker 
than the matrix at test conditions. The AC sheet grain boundaries remained strong and 
ductile leading to typical ductile transgranular deformation and ultimately resulting in 
"woody" tearing-type fractures with the grain boundaries appearing to play no direct 
part. 

Reasons for reduction in ductility of the AC sheet at about 1800°C are not clear. Indivi¬ 
dual creep curves as well as instantaneous strain rate values plotted as a function of strain 
show a distinct discontinuity associated with the drop in ductility. Some inhibiting mechan¬ 
ism is apparently operative in the region of the second stage of creep (usually associated 
with a constant rate of dislocation flow). 

(" Creepy Analysis ^ ^ go 

Studies of creep data for arc-cast tungsten, W(3), were performed using the method pro¬ 
posed by Woodford 2 for constant-load tests. Although most analyses and interpretations of 
creep mechanisms are based on constant-stress tests, most experimental data for materi¬ 
als are obtained using the constant-load technique. For constant-load tests, actual stress 
on the specimen increases as strain increases since the cross-sectional area is decreas¬ 
ing. In both types of testing, the assumption is made that deformation occurs uniformly 
over the gage length until local necking is initiated. 

A special computer program was written to analyze experimental strain — time data. 

This program was designed to yield calculated values of the instantaneous creep rate and 
instantaneous or true stress (the latter values are based on an assumption of constant vol¬ 
ume deformation). For the portion of the creep curve beyond that corresponding to mini¬ 
mum creep rate, a plot of instantaneous creep rate versus instantaneous stress was found 
to be linear on logarithmic coordinates (Figure 1.5). For each temperature, then, a rela¬ 
tionship of the form k = Aa 11 is indicated; where k is the instantaneous creep rate beyond 
the minimum, a is the instantaneous or true stress, and A and n are constants. Based on 
this approach, the following expression was obtained relating k to the initial stress, ct 0 , 
and the strain, e, in a constant-load Greep test: 

e Aa n = A[a 0 (i + e)] n . (1.1) 

Integration at constant o Q and solving for time, t, yields: 

t = B(1 + e) 1_n + C (1.2) 

where C is the constant of integration and n is the slope of the plot shown in Figure 1.5 at 
any given temperature. 

For a given value of Oq, this expression defines a linear relationship between t and 
(1 + e)*“ n for the region beyond the minimum creep rate. A typical plot of this type is 
shown in Figure 1.6 for the tungsten data at 2400°C for which n = 4. 291. Linearity is 
noted at all a 0 values. It has been found in this analysis that the value of C [i. e., the 


9 u 

D. A. Woodford, Constant-Load Creep Data Interpreted in Terms of the Stress Dependence of Dislocation Velocity, 
Trans. Met. Soc. of AIME, Vol. 239, May 1967, p. 655. 


Instantaneous stress, kg/mm 


18 



W at various temperatures 





19 


1.0 

0.9 

0.8 

c 

~ 0.7 
+ 

0.6 

0.5 

0.4 

0 10 20 30 40 50 60 70 80 

Time, hours 



— 

n = 4.291 

— 







O 

k 






1 Y 

\ 

o 

\% 


f 6T 

Qr^ 



ii 1 R 

111 l 


A. % 

\ 




v 

11 4 

_1 A* 

\ a 

\ ‘—i 

\ O 

: \ S 


\ 





m \v 








Fig. 1.6 — Tungsten W(3) data at 2400°C analyzed in terms of 
(1 + e) 1 —1n versus time 


intercept in Figure 1.6 at (1 + e)^“ n = 0] at any given temperature is related to a 0 ac¬ 
cording to: 

C = Da™ (1.3) 

leading to: 

e =---1 (1.4) 

(E - Ftf“t)P 


Expressions of this type for the tungsten data obtained in this study are: 


T = 2400°C 


€ = 


T = 2600°C 


T = 2800°C 


(0.95 - 6.1708 x 10 -15 a£* 225 t) 0> 30386 


(0.95 - 7.5289 x 10" 15 a*- 561 t) 0,2922 


-1 


€ = 


(0.94 - 1.6707 x 10" 16 CT 6 - 557 t) 0-2247 


where: 


(1.5) 

( 1 . 6 ) 

(1.7) 


€ is strain 
t is time, hr 

cr 0 is initial stress, kg/mm 2 x 1422.3 






The experimental results obtained at 2400°C and those determined from the above equa¬ 
tion are compared in Figure 1. 7. The agreement is excellent when considering the problems 
and uncertainties associated with testing at 2400°C. 



Fig. 1.7 — Strain versus time creep data for arc-cast tungsten tested at 2400°C 
in Ho atmosphere 

p" . 

Hot Hardness 1 

Wrought, arc-cast tungsten sheet, W(3), specimens were annealed in hydrogen at vari¬ 
ous temperatures for different times. Microhardness measurements were then made from 
room temperature to 1300°C. Results are given in Figure 1.8 for three specimens annealed 
at 1200°, 1400°, and 1600°C. As shown, a break in the curve occurs at approximately 300°C 
(0.16 T m , where T m represents the absolute melting temperature) consistent with data re¬ 
ported previously 3 for unalloyed tungsten. Also shown is a decrease in hardness with increas¬ 
ing annealing temperature over the total temperature range investigated. These differences 
are apparently related to the grain size differences resulting from the annealing treatments. 
Figure 1.9 compares the grain size and typical indents obtained in determining the micro¬ 
hardness of specimens annealed at 1200°C and 1600°C. The specimen receiving the highest- 
temperature anneal (1600°C) had such large grains that the indent and resulting symmetrical 
deformation pattern were contained within a single grain. The specimen annealed at 1200°C 
shows a relatively fine-grained structure with no deformation pattern apparent. Average 
grain sizes associated with the data given in Figure 1.8 are 20, 28, and 88 microns for an¬ 
nealing temperatures of 1200°, 1400°, and 1600°C, respectively. 

I Microstructural StudiesjR. C. Rau, S. F. Bartram, P. N. Flagella) -— ^-3' 

Microstructural studies, using optical and electron microscopy and X-ray diffraction 
techniques, were performed on high-purity polycrystalline tungsten after creep-rupture 
testing at stress levels ranging from 0.18 to 5.62 kg,/mm^ over the temperature range from 


"AEC Fuels and Materials Development Program Progress Report No. 67," GE-NMPO, GEMP-67,June 30,1967, p. 55. 



21 



Temperature, °C 

Fig. 1.8 — Hot microhardness of wrought, arc-cast tungsten sheet, W(3), 
after various annealing treatments 

1600° to 3000°C.The objectives of these studies were (1) to characterize the dislocation 
microstructure in arc-cast tungsten as a function of stress and temperature, and (2) to de¬ 
termine the failure mechanisms in arc-cast and powder-metallurgy processed tungsten. 

Dislocation Microstructures in Arc-Cast Tungsten - Development of dislocation sub¬ 
structures in metals during high-temperature deformation plays an important role in de¬ 
termining creep resistance. This substructure consists of dislocation networks or tangles, 
which form low-angle subgrain boundaries between slightly misoriented regions of materi¬ 
al, and free dislocations located within the subgrains. In the present investigation, sub¬ 
grain sizes, dislocation densities, and subgrain boundary tilt angles were determined in 
arc-cast, unalloyed tungsten and have been correlated with high-temperature creep condi¬ 
tions. Although the specimens used in this study had all been tested to failure, micro- 
structural examinations were carried out on regions somewhat removed from the fracture 
to avoid third-stage creep effects and thus concentrate on effects related more nearly to 
secondary creep. 

Material used for this study was 1.5-mm-thick wrought, arc-cast, polycrystalline tung¬ 
sten, W(3), sheet. Creep-rupture specimens with 6. 4-mm gage widths and 25.4-mm gage 



22 


1200°C anneal for 2 hours. 1600°C anneal for 200 hours. 

Indent at 1308°C. Indent at 1310°C. 

Fig. 1.9 — Photomicrographs of wrought, arc-cast unalloyed tungsten sheet showing 
typical indents and microstructure after hot hardness testing (200X) 

lengths were tested under constant load at temperatures ranging from 1600° to 3000°C and 
at stress levels ranging from 0.18 to 5.62 kg/mm^ in a hydrogen atmosphere. All speci¬ 
mens were annealed at test temperature for 2 hours before stress was applied. 

A number of experimental techniques were used in studying the substructure of the creep- 
tested specimens. Optical microscopy methods were used to determine subgrain sizes and 
free dislocation densities. These measurements were made using electrolytically polished 
and etched surfaces parallel to the stress axis. To gain increased resolution and magnifi¬ 
cation over that available by optical microscopy, replica electron microscopy techniques 
were used to study etch pits in some samples. Replication of the polished and etched sur¬ 
faces involved a two-stage plastic/carbon technique, using chromium shadowing. Direct 
observations of the dislocation microstructures were made by transmission electron micro¬ 
scopy of thin foils prepared by a double-jet electrolytic thinning technique. 4 These obser¬ 
vations were used primarily for determining the dislocation separations within subgrain 
boundaries and for direct measurements of tilt angles between subgrains. Finally, back- 
reflection Laue X-ray diffraction photographs were used for indirect determinations of 
subgrain sizes and subgrain boundary tilt angles. This method, described previously, 5 
has the advantage over direct microscopic methods of providing bulk measurements in¬ 
tegrated over a volume of the sample. A summary of the results for the tungsten samples 
examined is given in Table 1.1. 

^R. L. Ladd and R. C. Rau, "Immersed Double-Jet Technique for Electrothinning Tungsten for Transmission Electron 
Microscopy/' Rev. Sci. Instrum., Vol. 38, 1967, p. 1162. 

5 "AEC Fuels and Materials Development Program Progress Report No. 71," GE-NMPO, GEMP-1002, December 29,1967, 

pp. 18-28. 





23 


TABLE 1.1 

MICROSTRUCTURAL DATA ON ARC-CAST TUNGSTEN 


Free Dislocation Subgrain Dislocation Separation Dislocation 


Sample 

Number 

Creep Testli 
Temperature, 

°c 

ng Conditions 
Stress 

Subgrain Diameter, 

mm 

Density, 

10 4 per mm 2 

Boundary 
Tilt Angle, 
degrees 

in Sub-Boundaries, 

£ 

Density in 
Sub-Boundaries, 

10® per mm 2 

kg/mm 2 

psi 

Optical 

X-Ray 

Optical 

Electron 

X-Ray 

Electron 

W{3)-69 

1600 

5.62 

8000 

~0.01 

<0.01 


4.7 



525 


-71 

1600 

3.37 

4800 

-0.01 

<0.01 

3.5 






-50 

1800 

3.37 

4800 

0.013 

0.057 

7.4 

12.8 

0.294 


780 


-87 

1800 

2.46 

3500 

0.075 

0.084 

1.0 


0.207 

897 

1340 

2.73 

-52 

1800 

2.11 

3000 







1507 


-53 

2000 

2.11 

3000 

0.028 

0.090 

3.5 


0.364 

505 

1070 

4.47 

-58 

2000 

1.76 

2500 



0.3 






-39 

2000 

1.41 

2000 

0.083 

0.096 

0.30 

0.96 

0.265 

721 


3.05 

-46 

2000 

1.41 

2000 

0.075 

0.077 


1.5 

0.365 

512 


5.22 

-45 

2000 

1.05 

1500 

0.14 

0.10 

1.3 


0.257 

712 

500 

2.83 

-13 

2200 

1.41 

2000 

0.067 

0.10 

2.6 

11.1 

0.321 

589 


3.48 

-16 

2200 

0.70 

1000 

0.10 

0.12 

1.3 


0.522 

355 

1250 

4.79 

-38 

2400 

1.05 

1500 

0.05 

0.092 

4.2 

3.4 

0.333 

567 


4.00 

-21 

2400 

0.70 

1000 



0.71 






-85 

2400 

0.56 

800 

0.10 

0.21 

0.42 


0.160 

1152 

633 

8.42 

-24 

2400 

0.56 

800 



1.2 






-22 

2400 

0.46 

650 



0.53 






-19 

2600 

0.84 

1200 

0.057 

0.16 

0.9 

4.9 

0.366 

510 


2.55 

-27 

2600 

0.28 

400 

0.40 

0.55 

1.6 


0.342 

531 


6.85 

-28 

2800 

0.35 

500 

0.18 

0.33 

1.1 


0.220 

828 


7.36 

-36 

2800 

0.21 

300 

0.27 

0.30 

0.32 


0.344 

529 


12.64 

-49 

3000 

0.21 

300 

0.31 

0.42 

1.3 


0.272 

678 


7.14 


Optical microscope examination of the tested specimens showed that the structure al¬ 
most invariably consisted of very large grains, often covering the entire width and thick¬ 
ness of the pieces. This large grain size was verified by Laue back-reflection X-ray 
photographs, and permitted such photographs to be obtained from single grains. A few 
exceptions to this large-grained microstructure were specimens tested at the lower tem¬ 
peratures (1600° C and 1800° C) and higher stress levels (>3 kg/mm 2 ). In these specimens, 
the typical large-grained microstructure occurred at the surfaces, but a fine-grained 
polycrystalline texture was retained in the interior. 

Substructure developed within the large grains during testing consisted of subgrain 
boundaries and free dislocations, both of which are revealed by etch pits, as shown in 
the optical micrographs of Figure 1.10. Typical back-reflection Laue photographs are 
ehown in Figure 1.11. The pattern in Figure 1.11a shows typical clusters of small spots 
corresponding to the subgrain reflections from a single large tungsten grain, while that 
in Figure 1.11b shows spotty Debye-Scherrer rings from small randomly oriented grains 
in a low-temperature, high-stress specimen. Such a pattern, which must be produced by 
grains having a mean dimension of about 10“ 2 mm, cannot be used for subgrain measure¬ 
ments. 

Subgrain diameters measured by direct optical examination are listed in column five of 
Table 1.1; those determined indirectly from the Laue photographs are listed in column 
six. The agreement between the results is considered quite good, although the subgrain 
diameters indicated by X-ray tend to be slightly larger than those measured with the 
microscope. This is probably due to overlap and coincidences of spots in the Laue pat- 







25 


terns, leading to low estimates of spot densities within the clusters and therefore high 
calculated subgrain diameters. 

The subgrain sizes in tungsten are independent of test temperature and show an inverse 
relationship to creep stress, as indicated in Figure 1.12. This log-log plot shows a near¬ 
ly linear dependence, similar to that obtained for aluminum, 6 iron, 7 and steels 8 ? 9 tested at 
various stress levels. In a recent review, Sherby and Burke 10 correlated subgrain sizes, 

L, with stress, a, for a number of metals, and showed that the data seem to obey a power 
relation of the type 

a = KL'^, (1.8) 

where j3 is almost equal to unity. The tungsten data of Figure 1.12 appear to follow a simi¬ 
lar relation. 

Free dislocation densities were determined by counting the etch pits within subgrains in 
areas such as those seen in the optical micrographs of Figure 1.10. In addition, free etch 
pits in some specimens were counted from composite replica electron micrographs such 
as that shown in Figure 1.13. Results obtained from the two counts are listed in columns 
seven and eight, respectively, of Table 1.1. Unlike subgrain sizes, which are a function 
only of creep stress, the free dislocation density depends upon both temperature and 
stress. The relationship is illustrated in Figure 1.14. As shown by this plot, the number 
of etch pits, or the density of free dislocations, increases with increasing stress at con¬ 
stant temperature and increases with increasing temperature at constant stress. To main¬ 
tain a fixed concentration of free dislocations, the temperature must be decreased if the 
stress is increased. 

The relationship of the free dislocation density (p) to stress appears similar to that ob¬ 
served for the linear creep rate (e s ) as a function of stress (power law at constant tem¬ 
perature) for this same material. Good correlation of the creep rate data was obtained 
when e s was divided by the diffusion coefficient (Figure 1.3) as suggested by Sherby 11 for 
polycrystalline solids. When the same technique is applied to the dislocation density data, 
good correlation is also obtained, as shown in Figure 1.15. The same stress dependency 
(4.2) is obtained for the diffusion compensated p as for the diffusion compensated k s 
parameter. This linear portion of the curve is believed to result from creep by disloca¬ 
tion climb or by the motion of jogged screw dislocations involving an equilibrium concen¬ 
tration of vacancies. Barrett 12 indicates that slip models based on the glide of jogged screw 
dislocations can predict an approximate power law region with € g a 5 but require the p to 
be a strong function of stress (o^ or a^). The present data appear to support this model, at 
least for the high-temperature, low-stress region. The dislocation densities in the low- 
temperature, high-stress region, showing a deviation from the power law dependence, are 
apparently influenced by excess vacancies. 

®|. S. Servi and N. J. Grant, "Structure Observations of Aluminum Deformed in Creep at Elevated Temperatures/' 

Trans, of AIME, Vol. 191, 1951, p. 917. 

^A. Goldberg, "Influence of Prior Cold Work on the Creep Resistance and Microstructure of a 0.05% Carbon Steel," 

J. Iron and Steel Inst., Vol. 204, 1966, p. 268. 

O 9$ 

°F. Garofalo, 0. Richmond, W. F. Domis, and F. von Gemmingen, Strain-Time, Rate-Stress and Rate-Temperature 
Relations During Large Deformations in Creep," Joint International Conference on Creep, The Institution of Mechanical 
Engineers, London, 1963, p. 1-31. 

9 „ 

F. Garofalo, W. F. Domis, and F. von Gemmingen, Effect of Grain Size on the Creep Behavior of an Austenitic Iron-Base 
Alloy," Trans. Met. Soc. of AIME, Vol. 230, 1964, p. 1460. 

D. Sherby and P. M. Burke, "Mechanical Behavior of Crystalline Solids at Elevated Temperatures," Progress in Materials 
Science, Vol. 13, No. 7, 1967, p. 325. 

^O. D. Sherby, "Factors Affecting the High Temperature Creep of Polycrystalline Solids," Acta Met., Vol. 10,1962, p. 135. 

C. R. Barrett, On the Stress Dependence of High Temperature Creep, Trans, of AIME, Vol. 239, 1967, p. 1726. 



Creep stress, psi 

150 200 300 400 500 700 1,000 1,500 2,000 3,000 4,000 5,000 7,000 10,000 






I 


X Laue back-reflection X-ray 
# Optical 




I 


Creep stress, kg/mm* 


Fig. 1.12 — Effect of stress on the subgrain diameter of arc-cast tungsten as the 
result of creep-rupture testing from 1600° to 3000°C 

\ iWVW'W*** Fifin'. 

wm- - .(viMm*- W>*,’W>v;v 

■; x . 'iS (.■ :'5> >. -fli' Y-'f.'-v* V-. 

: •/* v" 


v.:vv' \* 


v. < r '• I • - , ttr "<$ . ' ' • V 




S -I - . .»• '.VfcxW ■ 1 

1^; A ■ ":£•,. Y Y. ;>!% ,* * '■•' 

'-V* ; 

'WiBm'M , r v - 

xY^YrV. t, ;l v \feYj/v vTOi 


#vi' ; ^ 


10 JJ 


> : >*• 


VTOlL 


Fig. 1.13 - Composite replica electron micrograph of overlapping areas used for etch 
pit counting. Tungsten tested at 1600°C and 5.62 kg/rnm^. 

















27 


Stress, psi 

200 300 400 500 700 1,000 1,500 2,000 3,000 4,000 5,000 7,000 10,000 



Stress, kg/mm 2 


Fig. 1.14 — Number of free etch pits (dislocations) as a function of stress and 
temperature for wrought, arc-cast tungsten after creep tests from 
1600° to 3000°C 

Average misorientation angles between subgrains, determined by the back-reflection Laue 
technique, are listed in column nine of Table 1.1. The values obtained range from 0.16 to 
0. 52 degree, with the average for all samples being 0. 31 degree. If the actual number of 
subgrains present has been underestimated, because of overlap of Laue spots, the calcu¬ 
lated tilt angles may be somewhat too large, but the error should not exceed about 0. 05 
degree. These values are in good agreement with subgrain misorientations reported for 
other creep tested metals. 13 ’ 14 To check the values of tilt angle directly, measurements 
of misorientations between subgrains in one sample were made from selected area elec¬ 
tron diffraction patterns obtained with the electron microscope. Values obtained by this 
technique ranged from 0.17 to 0.47 degree, with the average being 0.29 degree, in good 
agreement with the X-ray results. 

From the tilt angles, the average dislocation separation within the subgrain boundaries 
was calculated. This separation, S, is given by the expression 15 


13 F. Garofalo, L. Zweil, A. S. Keh, and S. Weissmannn, "Substructure Formation in Iron During Creep at 600°C," Acta Met., 
Vol. 9, 1961, p. 721. 

^R. W. Guard, "Discussion of Parker and Washburn Paper on the Role of the Boundary in Creep Phenomena," Creep and 
Recovery, American Society for Metals, 1957, p. 251. 

15 S. Amelinckx and W. Dekeyser, "The Structure and Properties of Grain Boundaries," Solid State Phys., Vol. 8,1959, p. 325. 






28 



0.1 


1.0 

Stress, kg/mm 2 


Fig. 1.15 — Correlation of diffusion-compensated dislocation density as a 
function of stress for wrought, arc-cast unalloyed tungsten, 
W(3), creep tested at temperatures from 1600° to 3000°C 


S = 


b 

2 sin 0/2 


(1.9) 


where b is the length of the Burgers vector of the dislocations and 0 is the tilt angle. At 
small angles this expression reduces to 


S=y (1.10) 

where 0 is expressed in radians. Using this expression and the tilt angles tabulated in col¬ 
umn nine of Table 1.1, the separation distances given in column ten were obtained. 

Direct observations of subgrain boundary dislocation configurations and measurements 
of separations between the dislocations were made by transmission electron microscopy. 
Two subgrain boundaries analyzed crystallographically are shown in Figures 1.16 and 
1.17, while typical measurements of dislocation separations are listed in column 11 of 
Table 1.1. Although these measured separations represent a small number of individual 
dislocation networks and cannot be considered good statistical values, they neverthe¬ 
less verify the accuracy of the calculated values given in column ten. 





£•’ - ■ - '*»■** 

***Z ’**»%$*& * 

$%%***& ** Mm p'* 


Transmission electron micrographs of nearly pure tilt boundary in 
tungsten tested at 1800°C and 2.11 kg/mm 2 , (a) Operating reflec¬ 
tion (110); intersections of the boundary with the bottom and top 
surfaces of the foil are indicated by B and T, respectively, (b) Operat¬ 
ing reflection (110). 


Transmission electron micrographs of twist boundary in tungsten 
at 1800°C and 3.37 kg/mm 2 , (a) Operating reflection (020). 

(b) Operating reflection (110). 



• § 

♦ w 









30 


Finally, from the subgrain diameter and tilt angle data, the dislocation density, p, in 
the subgrain boundaries was calculated from the expression 16 



where 6 is the tilt angle in radians, L is the subgrain diameter, and b is the Burgers vec¬ 
tor of the dislocations. This calculation thus gives the number of dislocations per unit area 
which are tied up in the subgrain boundary networks, rather than the number of free or mo¬ 
bile dislocations which can contribute to deformation under stress. The total number of sub¬ 
grain boundary dislocations must obviously increase with increasing tilt angles and decreas¬ 
ing subgrain sizes. 

Subgrain boundary dislocation densities were calculated from the subgrain diameters and 
tilt angles derived from the Laue photographs, columns six and nine, respectively, of Table 
1.1, and are given in column 12 of that table. The data reveal that the number of disloca¬ 
tions in the subgrain boundaries tends to increase with decreasing stress at constant tem¬ 
perature and gradually increases as the temperature is raised. These observations indi¬ 
cate that during creep, free mobile dislocations within the subgrains move to and accumu¬ 
late in the subgrain boundaries. Since dislocation movement during high-temperature creep 
is primarily diffusion controlled, longer times (i. e., lower stresses) and higher tempera¬ 
tures would be expected to increase the number of dislocations occupying subgrain bound¬ 
aries. 

The two subgrain boundaries shown in Figures 1.16 and 1.17 represent the two basic 
types of subgrain misorientations which can occur in body-centered cubic (bcc) metals. 
Figure 1.16 shows a nearly pure tilt boundary formed when subgrain misorientation oc¬ 
curs by a rotation about an axis in the boundary plane, while Figure 1.17 shows a twist 
boundary formed when such misorientation occurs about an axis perpendicular to the bound¬ 
ary plane. Most observed boundaries are formed by a combination of these two basic rota¬ 
tions. 

Tilting experiments in_the electron microscope reveal that the tilt boundary shown in 
Figure 1.16 lies on a (111) plane, inclined approximately 45 degrees from the (001) foil 
plane. Intersections of this boundary with the bottom and top surfaces of the foil are de¬ 
noted B and T, respectively, in Figure 1.16a. It can be seen that the parallel set of close¬ 
ly spaced straight dislocations comprising this boundary is imaged in strong contrast in 
Figure 1.16a, where the (110) reflection is operating (i. e., g = [110]), but is at extinction 
in Figure 1.16b where the (110) reflection is operating (i. e., g= [110], g-b = 0). This 
diffraction_contrast behavior is consistent with dislocations having a Burgers vector 
b = a/2[lll ], indicating that they are edge dislocations and that the boundary is composed 
primarily of extra half-planes inserted on the (111) boundary plane. The fact that a second 
set of widely spaced dislocations, visible in both micrographs, is superimposed on this 
network of parallel edge dislocations indicates that the boundary is not a pure tilt bound¬ 
ary, but has a small amount of twist character. 

Several direct measurements across the tilt boundary of Figure 1.16 indicate that the 
misorientation between the subgrains is about 0. 83 degree of arc, i. e., a higher than 
average subgrain misorientation. The measured separation of the edge dislocations in 
Figure 1.16a is approximately 105 X. Making a geometrical correction for the measured 
slope of the boundary plane, the true spacing of the dislocations is then 183 X. This is in 
very good agreement with the value of 188 X calculated from equation (1.10) for a pure 
tilt boundary composed of a/2 <111> dislocations and having a tilt angle of 0.83 degree. 

1 fi 

D. McLean, “Creep Processes in Coarse-Grained Aluminum," J. Inst. Metals, Vol. 80, 1951—52, p. 507. 



31 


The twist boundary of Figure 1.17 is a very extensive planar boundary, lying on the 
(001) plane parallel to the plane of the foil. The extinction conditions illustrated by the 
photomicrographs of Figure 1.17 indicate that the network is composed of three sets of 
straight dislocations, two having Burgers vectors of the type a/2<lll> and one having 
b = a <100>, as diagramed in Figure 1.18, plus a superimposed set of randomly oriented 
stranger dislocations which were not identified. One of the following two interactions has 
occurred to form the short a <100> dislocation segments: 

a/2 [111] + a/2 [111] = a [100] 

a/2 [111] + a/2 [111] = a [100]; 

however, the diffraction conditions observed were not sufficient to differentiate between 
these two equivalent possibilities. 

Similar networks have been identified in the bcc metals iron 17 * 18 and tantalum. 19 The ob¬ 
served mesh shape in Figure 1.17 agrees rather well with that calculated by Carrington, 
Hale, and McLean 17 for a network on {100}, from strictly theoretical considerations. It 
should be pointed out, however, that since this boundary does not lie on a {110} plane, it 
is not pure twist, but actually is a mixed boundary, i. e., it is not composed of pure screw 
dislocations. The <100> segments are in screw orientation, but the a/2 <111> segments 
have mixed edge and screw character. 15 

Using equation (1.10), the angle of misorientation between the two subgrains on either 
side of this boundary was calculated from the mesh size of the network. This gave a value 
of about 0. 47 degree, in good agreement with a number of measurements on other bound¬ 
aries and within the range of tilt angles determined by the back-reflection Laue method. 

Failure Mechanisms in Arc-Cast and Powder Metallurgy Tungsten - Using replica elec¬ 
tron microscopy, Stiegler et al. 20 studied the formation of voids or cavities at the grain 
boundaries of powder-metallurgy processed tungsten subjected to creep deformation at 
1650°C and 2200°C. They found that at both test temperatures specimens generally failed 
in a brittle manner, and they attributed this failure to the growth and linking up of grain 
boundary cavities. 

Recently, however, it has been shown that while powder-metallurgy tungsten creep- 
rupture tested at temperatures ranging from 1600° to 3000°C failed in a brittle, inter¬ 
granular manner, specimens prepared from arc-cast material and tested under similar 
conditions failed in a ductile manner. High-temperature stress-rupture tests of molyb¬ 
denum also showed similar differences in the failure behavior of powder-metallurgy and 
arc-cast material. 21 These results indicate that different mechanisms of failure are opera- 
ing in the two types of materials. To determine this basic difference, optical microscopy 
and electron fractography studies of creep-rupture-tested powder metallurgy and arc-cast 
tungsten were undertaken. 

The studies were performed using selected powder-metallurgy (W-4) and arc-cast (W-3) 
tungsten specimens. These specimens were formed from sheet material with initial thick¬ 
nesses of 0.5 mm for powder-metallurgy specimens and 1.5 mm for arc-cast specimens. 


17 W- Carrington, K. F. Hale, and D. McLean, "Arrangements of Dislocations in Iron," Proc. Roy. Soc., A Vol. 259,1960, p. 203. 

°S. M. Ohr and D. N. Beshers, "Crystallography of Dislocation Networks in Annealed Iron," Phil. Mag., Vol. 8,1963, p. 1343. 

19 

D. Hull, I. D. Mclvor, and W. S. Owen, "The Distribution of Dislocations in Annealed Tantalum," J. Less-Common Metals, 
Vol. 4, 1962, p.409. 

20 

J. O. Stiegler, K. Farrell, B. T. M. Loh, and H. E. McCoy, "Nature of Creep Cavities in Tungsten," Trans. ASM, Vol. 60, 

1967, p.494. 

21 a 

P. N. Flagella, High-Temperature Stress-Rupture Characteristics of Mo and Mo Alloys," AIAA Journal, Vol. 5,1967, p. 281. 


32 



Fig. 1.18 — Schematic drawing of the three sets of dislocations comprising 
the twist boundary network of Figure 1.17 


Specimens examined had been creep tested at 1600°, 2000°, 2200°, and 2600°C, with the 
results listed in Table 1. 2. 

For the fractographic study, fresh fracture surfaces in the highly stressed region near 
the original break were created by clamping the specimen in a vise with about 3 mm of the 
broken end protruding and sharply rapping the protruding end. These newly exposed sur¬ 
faces were then replicated by the standard two-stage plastic/carbon technique, using 
chromium shadowing, and examined in the electron microscope. 

After replication, these same pieces were mounted and mechanically polished for metal- 
lographic examination. The pieces were mounted edge-on, with the view perpendicular to 

TABLE 1.2 

TUNGSTEN SPECIMENS USED FOR STUDY OF FAILURE MECHANISM 


Sample 

No. 

Test 

Temperature, 

°C 

Stress 

kg/mm^ psi 

Time to 
Rupture, 
hr 

Elongation, 

% 

Maximum 
Void Content, 

% 

Measured 
Void Content, 
% 

Ratio of 
Measured/Max. 
Void Content, % 

Averai 

Grain S 
microi 

Powder Metallurgy 









W(4)-1-73 

1600 

4.22 

6000 

51.27 

32 

24.2 

5.58 

23.1 

32.5 

-79 

2000 

2.11 

3000 

23.17 

27 

21.3 

2.46 

11.5 

38.0 

-18 

2200 

1.41 

2000 

15.29 

15 

13.0 

2.02 

15.5 

48.5 

-10 

2200 

1.12 

1600 

60.65 

17 

14.5 

2.11 

14.6 

48.5 

-8 

2600 

0.84 

1200 

8.84 

8 

7.4 

1.11 

15.0 

215.5 

-7 

2600 

0.70 

1000 

25.08 

5 

4.8 

0.90 

18.8 

264.0 

Arc-Cast 










W(3)-48 

1600 

4.22 

6000 

30.90 

76 





-53 

2000 

2.11 

3000 

6.70 

70 





-14 

2200 

1.05 

1500 

40.70 

106 





-25 

2600 

0.56 

800 

4.77 

64 








33 


the stress axis, and were examined and photographed in the as-polished condition. In this 
way, any stress-induced porosity was observed in its original state, unaltered by chemical 
attack. Following examination for porosity, some of the specimens were etched for grain 
size evaluation. 

Optical microscopy of the as-polished surfaces showed an abundance of grain boundary 
cavities in the powder-metallurgy tungsten, and an almost total lack of such cavities in 
the arc-cast material. Representative micrographs are presented in Figure 1.19, where 
a and b are powder-metallurgy material, and c and d are arc-cast material. 

As illustrated by these micrographs, the powder-metallurgy material consisted of fairly 
equiaxed grains, which increased in average size with increasing test temperature. Grain 
boundary separation and cavitation were very pronounced, especially along boundaries 
which were transverse to the stress direction. On the other hand, arc-cast material was 
composed of very large grains extending across the total thickness of the specimens and 
showed practically no tendency to form cavities. A rare instance of a grain boundary cavity 
in the arc-cast specimen tested at 1600°C is indicated by the arrow in Figure 1.19c. 



Fig. 1.19 — Optical micrographs of creep-rupture-tested tungsten in as-polished 
condition, (a) Powder-metallurgy specimen tested at 1600°C and 
4.22 kg/mm^. (b) Powder-metallurgy specimen tested at 2000°C 
and 2.11 kg/mm^. (c) Arc-cast specimen tested at 1600°C and 
4.22 kg/mm^. (d) Arc-cast specimen tested at 2000°C and 2.11 
kg/mm^. 






34 


Replica electron microscopy of the powder-metallurgy samples showed that the fracture 
mode was almost totally intergranular, especially for the low-temperature, high-stress 
conditions, and that the exposed grain boundary surfaces contained a profusion of cavities. 
Typical examples are illustrated in Figure 1. 20. At low temperatures the voids were 
usually irregular in shape, tending to be relatively flattened along the boundary surface 
and often containing elongated "fingers" suggestive of void coalescence, Figure 1.20a. 
Such cavities are termed wedged-shaped, or w-type, 22 and are believed to originate 
primarily by grain boundary sliding. 23 ? 24 At higher temperatures, the shape of the cavities 
became more polyhedral, especially at low-stress levels, Figure 1.20b. These more 
rounded, or r-type 22 voids, are bounded by flat crystallographic facets which correspond 
to low-index planes of the cubic lattice. 25 Such polyhedral voids form under conditions 
where surface diffusion permits the cavities to reduce their surface tensions and assume 
equilibrium shapes. Although a few examples were seen of creep cavities aligned along 
grain boundary triple junctions, Figure 1.20c, there was nothing to indicate that such 
junctions were preferred sites for voids. In general, both w-type and r-type cavities tend¬ 
ed to be distributed uniformly over the grain boundary surfaces, in agreement with the ob¬ 
servations of Stiegler et al. 20 


In contrast to the powder-metallurgy material, replica electron micrographs of the arc- 
cast tungsten showed that the fracture at all temperatures was predominantly transgranu- 
lar, as illustrated in Figure 1.21a. When grain boundary surfaces were seen, they were 
always relatively clean and contained only a few small irregularities suggestive of cavi¬ 
ties, Figure 1.21b. In none of these samples was any evidence seen of cavity coalescence 
or polyhedral shaped voids. 


An attempt was made to quantitatively correlate the observed microstructural features 
with the mechanical properties of the creep-tested specimens. This involved measurements 
of the open porosity and grain size in the powder-metallurgy specimens, Table 1. 2. How¬ 
ever, similar measurements could not be made on the arc-cast material because of the lack 
of porosity and the extremely large grain size. 


For the powder-metallurgy tungsten, which failed with relatively little elongation in com¬ 
parison with the arc-cast material, it was of interest to determine whether all of the elonga¬ 
tion was caused by cavitation or whether other mechanisms also contributed. Accordingly, 
lineal measurements of open porosity were made from the as-polished surfaces at 250X 
magnification. These measurements were made parallel to the stress (elongation) direc¬ 
tion at approximately 3 to 4 mm from the original fracture of each specimen in regions 
similar to those shown in Figures 1.19a and b. The averages of ten measurements made 
on each sample were converted to percentages of the length sampled, and are listed in 
column eight of Table 1.2. 


Maximum possible void content in the stress direction, assuming that the total elongation 
was due to cavitation, was calculated from the measured elongations listed in column six 
of Table 1. 2 using the relation: 


Max void content (%) = 


elongation (%) x 100 
100 + elongation (%) 


Comparison of the values listed in column seven of the table with the measured values of 
column eight shows that only a portion of the total elongation can be accounted for by the 
cavities. This relative proportion, however, remains essentially constant as the test 
22 

F. Garofalo, Fundamentals of Creep and Creep-Rupture in Metals, Macmillan Series in Materials Science, New York, 1965, p. 213. 
23 

C. Zener, Elasticity and Anelasticity of Metals, University of Chicago Press, Chicago, 1948, p. 158. 

24 H. C. Chang and N. J. Grant, “Mechanism of Intercrystalline Fracture," Trans, of AIME, Vol. 206, 1956, p. 544. 

nc 

K. Farrell, B. T. M. Loh, and J. 0. Stiegler, "Morphologies of Bubbles and Voids in Tungsten," Trans. ASM, Vol. 60,1967, p. 485. 




35 



Fig. 1.20 - Electron fractographs of powder-metallurgy tungsten, (a) Irregular creep 
cavities in specimen tested at 1600°C and 4.22 kg/mm^. (b) Polyhedral 
shaped voids in specimen tested at 2600°C and 0.70 kg/mm^. (c) Cavities 
along a grain boundary triple junction in specimen tested at 2600°C and 
0.84 kg/mrn^. 

temperature increases and overall ductility decreases. As shown in column nine of the 
table, the portion of the elongation due to cavities is about 16 percent of that expected 
if cavitation were the only source of deformation. 

To determine if grain deformation had occurred during creep, lineal grain size meas¬ 
urements were made in the longitudinal (stress) and transverse directions both in the high- 
stress regions where porosity measurements had been made and in the unstressed holder 
ends of the specimens. These measurements showed the grains to be elongated approxi¬ 
mately 40 percent in the stress direction in all specimens, but this elongation was the 






36 



Fig. 1.21 - Electron fractographs of arc-cast tungsten, (a) Transgranular fracture 
in specimen tested at 2000°C and 2.11 kg/mm^. (b) Grain boundary 
surface.in specimen tested at 1600°C and 4.22 kg/mm^. 


same in unstressed regions as in highly stressed regions. This elongation is undoubtedly 
the result of prior fabrication history, since the specimens were formed with their 
lengths parallel to the rolling direction of the starting sheet material. The lack of meas- 
ureable grain deformation during creep indicates that part of the bulk deformation must, 
therefore, be due to grain boundary sliding. 

Although these observations illustrate the pronounced difference in failure mechanisms 
of powder-metallurgy and arc-cast tungsten, they do not indicate the underlying cause of 
this difference. Grain boundary failure, such as that observed in the powder-metallurgy 
material, can occur if the matrix is strengthened without a corresponding strengthening 
of the boundaries, or if the boundaries themselves are weakened. 26 Matrix hardening can 
be produced by a dispersion of impurities or precipitates while grain boundary weakening 
can occur with preferential precipitation in the boundaries, especially if a continuous film 
of precipitate forms. By the same reasoning, transgranular fracture, such as that ob¬ 
served in the arc-cast tungsten, will be favored if the grain boundaries are strengthened 
with respect to the matrix. Such strengthening can arise through a decrease in the amount 
of segregated impurities at grain boundaries, thus leading to increased grain boundary 
mobility and ductility. Alternatively, certain types of discontinuous precipitation at grain 
boundaries can strengthen those boundaries by blocking dislocation movement and pre¬ 
venting sliding. 

In the present work, no direct evidence was observed of precipitation either within the 
matrix or along grain boundaries. The techniques employed, however, would not be ex¬ 
pected to resolve very small precipitates. To detect such precipitates, transmission 
electron microscopy or replica electron microscopy of polished and etched Surfaces would 
be required. 


26 


’Zener, loc. cit. 




37 



Recent creep-rupture studies of arc-cast W - 25Re (wt %) sheet material have extended 
the range of data reported previously 27 for this material. A complete summary of all the 
creep-rupture data generated for arc-cast W - 25Re in this program is presented in Figures 
1. 22 and 1.23. As reported, 27 arc-cast W - 25Re is stronger than arc-cast tungsten up to 
2000°C and somewhat weaker than arc-cast tungsten at temperatures above 2000°C. It is im¬ 
portant to note that the isotherms in Figures 1. 2SLand l.SDJare linear even though non-linear 
isotherms were observed in tests of powder-metallurgy W - 25Re. Apparently, different 
deformation mechanisms are involved in these two types of material. Two isotherms are 
shown for 2400°C, in Figure 1 . 3 . 1 ; two are shown to indicate that at this temperature, the 
stress-rupture properties of W - 25Re depend upon elongation at rupture. 

vv 


30Re jj 3 0M 


*h 


-7 


P 


Sf 


Stress-rupture and creep data were obtained for a number of different lots of wrought, 
powder-metallurgy (PM) W - 30Re - 30Mo (at. %) sheet material and one lot of wrought 
arc-cast (AC) material in sheet form. The PM material was produced at GE-NMPO under 
Task 1115, "Advanced Fast Breeder Reactor Fuel Element Cladding Development," and 
the AC material purchased commercially. A comparison of the stress-rupture and linear 
creep rate test results are shown in Figure 1.24 and 1.25. The AC material exhibits a 
greater rupture life than the PM material at both 1600°C and 2200°C for the stress levels 
investigated. This is probably related to the fact that the AC material is considerably more 
ductile, based on the elongation at rupture. 


Photomicrographs of the PM sheet (lot 2) and the commercial AC sheet (lot 3) after creep- 
rupture testing are shown in Figure 1.26. Both materials tested at 1600°C and 3.37 kg/mm^ 
show uniformly dispersed Re-rich sigma phase. The grain size for both is essentially the 
same, but intergranular separation is more predominant in the PM material. This is appar¬ 
ently the reason for the lower ductility. After testing at 2200°C and 0. 56 kg/mm^, the two 
materials are single phase, but the grain size in the AC material is approximately four 
times that of the PM material. Deformation appears to be primarily intergranular for the 
AC material, whereas cavitation or grain-boundary separation appears to contribute sig¬ 
nificantly to the deformation of the PM material. These factors may also account for the 
large difference in ductility for the two materials. 


The linear creep rate data are essentially the same at 1600°C for the two types of ma¬ 
terial. At 2200°C, creep resistance of the AC material appears to be significantly greater 
at the lower stresses. This difference is apparently related to the difference in grain size 
at 2200°C which leads to different creep mechanisms. The PM material (lot 2) is fine 
grained (Figure 1. 26) and has a stress versus creep rate slope of approximately 1; this 
indicates that diffusional creep 28 is the primary mechanism. The AC material (lot 3) has 
large grains at 2200°C (Figure 1.26) and gives a stress versus creep rate slope of approx¬ 
imate 5; this indicates that dislocation climb is the primary creep mechanism. 29 


Special fabrication procedures were employed to produce high-purity W — 30Re — 30Mo 
sheet material. Creep-rupture tests were performed to compare this material with the 
previously produced (i.e., nonpurified) sheet material. 30 The in-process fabrication pro- 


^"Sixth Annual Report — High-Temperature Materials Program, Part A," GE-NMPO, GEMP-475A, March 31, 1967, p. 13. 
3 ®Sherby, loc. cit. 

29 C . Herring, "Diffusional Viscosity of a Polycrystalline Solid," J. Appl. Phys., Vol. 21, 1950, p. 437. 

30 GEMP-67, p. 25. 



38 



Time to rupture, hours 


Fig. 1.22 - Stress-rupture data for arc-cast W - 25Re (wt %) alloy sheet tested in hydrogen 



Linear creep rate, min' 


Fig. 1.23 — Linear creep rate data for arc-cast W — 25Re (wt %) alloy sheet tested in hydrogen 

cedure was modified to produce four sheets (0,06-cm thick x 7 cm x 10 cm) of material, 
as shown in' Table 1.3. Since the processing procedure differed slightly for each sheet, 
any effect on the mechanical properties could be determined. Specimens from each sheet 
were creep-rupture tested in hydrogen at 1600°C and 3. 37 kg/mm 2 and at 2200°C and 
0.56 kg/mm^. 

Typical creep curves, Figure 1.27, show no great difference in the rupture life, creep 
resistance, or ductility for the four sheets tested at the same conditions. Indication is 
that the variables associated with each sheet of high-purity material have a minor effect 
on the creep resistance and ductility, at least for the test conditions evaluated. Figure 
1.27 also contains the results for AC and nonpurified PM material of the same nominal 
composition. The PM material, obtained as regular laboratory-produced sheet, was the 
least creep resistant of the group at both temperatures. Data for the commercial, re¬ 
crystallized AC material show it to be somewhat more creep resistant than the regular 
nonpurified laboratory-produced sheet but considerably less creep resistant than the 
high-purity PM material. 
















































































Stress, kg/mm 2 Stres5/ kg/mi 






























Stress 

Direction 



306(3)-25 Powder metallurgy 


56.8 hours 
120% elongation 
15 microns 

(Neg, 9022, as-pol ished) 


1600°C, 3.37 kg/mm 2 


21.6 hours 
28% elongation 
20 microns 
(Neg. 8173, etched) 


306(8)-7 




306(3)-27 Powder metallurgy 

29.4 hours 

22.0 hours 

79% elongation 

265 microns 

2200°C, 0.56 kg/mm 2 32% elongation 

00 microns 

(Neg. 9026, etched) 

(Neg. 817], etched) 


306(8)-11 


Fig. 1.26 - Photomicrographs of wrought W - 30Re - 30Mo (at. %) sheet after 
creep-rupture testing in hydrogen; numbers below each photomicro¬ 
graph refer to time to rupture, elongation at fracture, and average 
grain size (250X) 



41 



TABLE 1.3 

FABRICATION PROCEDURE USED TO PRODUCE 

HIGH-PURITY W - 30Re - 30Mo SHEET 3 

Sheet No. 

Process 

306(4) 

Loose metal powders leached with 0.001 N HCI prior to blending. 
Not vapor treated. 

306(5) 

Same as 306(4) but powders leached with 0.001 N HNO 3 . 

306(6) 

Same as 306(4) but compacted powders treated 150 hours at 

1000°C in hydrogen-water vapor atmosphere prior to sintering. 

306(7) 

Same as 306(5) but treated as a compact similar to 306(6). 


a AII sheets were made from the identical metal powders and processed 
simultaneously. 




Photomicrographs of the high-purity samples after testing at 1600°C and2200°C are shown 
in Figures 1.28 and 1.29. No significant differences are noted between the samples tested 
at 2200°C. At 1600°C there appears to be some small variation in the second phase (sigma). 
Chemical analysis of the as-fabricated, high-purity sheet materials yielded the results 
given in Table 1.4. The major difference appears to be in the AC material where both the 
carbon and oxygen content are greater than any of the PM sheets. The special purification 
procedures apparently decrease the carbon content as the process is changed. These data 
along with the metallographic studies do not indicate the reason for the increased creep 
resistance of the high-purity material. 


MOLYBDENUM 


Stress-rupture data were previously reported 27 for wrought, arc-cast and wrought, 
powder-metallurgy, unalloyed molybdenum from 1200° to 2400°C in both hydrogen and 
argon atmospheres. Differences in the strength and structure for the two types of ma¬ 
terial were shown. It was also reported that essentially identical results were obtained 
for AC material purchased from a number of different vendor sources. 


Additional data were obtained for AC material to evaluate the creep properties of un¬ 
alloyed molybdenum. 

Constant-load tests of wrought, AC molybdenum sheet (0.05 cm thick) were performed 
in hydrogen at temperatures from 1200° to 2400°C. Chemical analyses of the as-received 
material are given in Table 1.5. As shown, sheet material from two different vendor 
sources was evaluated. Test results are given in Tables 1.6 and 1.7; Table 1.8 gives the 
stress-rupture data previously reported for wrought, AC molybdenum rod material (0.40- 
cm gage diameter) along with the creep data obtained in these same tests. Figure 1. 30 
shows the stress-rupture curves obtained using the three sources of material; good agree¬ 
ment is apparent at 1600°C and 2200°C where duplicate tests were performed. Linear creep 
rate data for the three materials are presented in Figure 1.31. These data also show good 
agreement for the three materials at 1600°C and 2200°C. At 2400°C the creep rate curve 
decreases in slope. In fact, the data points at 2200°C and 0.141 kg/mm^ (for M(9)4 material) 
deviate from the general trend, indicating a slope consistent with the data at 2400°C. The 
reason for this is not apparent but seems to be related to a unique behavior observed in an 
evaluation of the individual strain — time curves. 


These creep curves, at constant temperature, are presented in Figure 1. 32 as a function 
of the time required to give 10 percent strain. Plotting the curves in this way provides a 
comparison of the high- and low-stress tests at the same temperature. Additionally, tests 
at the same stress performed at different temperatures may also be compared conveniently. 
In this figure, creep curves obtained at 1200°C and 1600°C show a significant primary (first- 





i, percent 


(4,800 psi) 






8 9 10 11 12 13 14 15 16 17 18 19 20 

Time, hours 


0.56 kg/mm 2 

(800 psi) 



□ 

Arc-cast, 0.028 cm thick 


O 

Powder metallurgy 0.065 cm thick 

Sheet No 



5 

A 

| High-purity 

4 

a 

;> Powder metallurgy 

6 


l 0.052 cm thick 


2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 

Time, hours 


Fig. 1.27 - Creep curves for various lots of wrought W - 30Re - 30Mo (at. %) sheet 
tested in hydrogen 















43 



306(4)-12 

43.7 hr 

21.5% elongation 
165 microns 
{Neg. 9893) 


306(5)-17 

29.6 hr 

15% elongation 
135 microns 
(Neg. 9894) 



37.8 hr 

14.6% elongation 
165 microns 
(Neg. 9895) 


306(6)-18 


306(7)-23 

49.3 hr 

21 % elongation 
145 microns 
(Neg. 9896) 


Fig. 1.28 — Photomicrographs of various lots of wrought, powder-metallurgy W — 30Re - 30Mo 
(at. %) sheet after creep-rupture testing at 1600°C and 3.37 kg/mm 2 in hydrogen; 
numbers below each photomicrograph refer to time-to-rupture, elongation at fracture, 
and average grain diameter. The figure in parentheses following the 306, under each 
photograph, is the lot number. (Etched, 250X) 




44 



Fig. 1.29 - Photomicrographs of various lots of wrought, powder-metallurgy W - 30Re - 30Mo (at. %) sheet 
after creep-rupture testing at 2200°C and 0.56 kg/mm^ in hydrogen; numbers below each photo¬ 
micrograph refer to time-to-rupture, elongation at fracture, and average grain diameter. The figure 
in parentheses following the 306, under each photograph, is the lot number. (Etched, 250X) 




TABLE 1.4 


TABLE 1.5 


CHEMICAL ANALYSIS OF AS-FABRICATED W - 30Re - 30Mo 
SHEET SAMPLES FROM VARIOUS LOTS 


Lot 

No. 

“c 

Impurities, ppm 

0 H 

N 

Fabrication 

2 

19 

4.9 

0.5 

0.3 

PM regular laboratory process 

3 

<40 

80 

1.7 

11 

Per vendor r Commercial 


41 

33 

4 

4 

Per NMPO i AC 

4 

19 

12 

0.9 

25 i 

^ PM special purification 

5 

16 

9.8 

1 

0.2 


6 

13 

8.1 

0.9 

0.9 

procedures 

7 

7 

11 

0.9 

4 1 

1 


CHEMICAL ANALYSIS OF 
AS-RECEIVED WROUGHT, 
ARC-CAST MOLYBDENUM 


Material 

Impurity Content, 

ppm 

Source 

C 

H 

O 

N 

M(6)4 

140 

1 

17 

19 

M{9)4 

250 

1 

22 

5 


TABLE 1.6 


CREEP-RUPTURE RESULTS 3 FOR WROUGHT, ARC-CAST, UNALLOYED MOLYBDENUM SHEET, b 

M(6)4, TESTED IN HYDROGEN _ 


Specimen 

No. 

Test Conditions 

0 . 2 % 

Time to Indicated Strain, hr 
0.5% 1% 2% 3% 

5% 

10 % 

Rupture 

Linear 
Creep Rate, 
min -1 

Temperature, 

°C 

Stress, 

kg/mm 2 

Time, 

hr 

Elongation, 

% 

4 

1600 

2.11 


_ 

0.03 

0.13 

0.28 

0.60 

1.4 

2.5 

30 

9.8 x 10 —4 

5 

1600 

1.76 

_ 

0.05 

0.18 

0.53 

0.98 

1.9 

4.2 

6.9 

30 

3.6 x lO -4 

6 

1600 

1.41 

0.06 

0.20 

0.58 

1.5 

2.5 

4.7 

9.8 

25.1 

55 

1.6 x 10 —4 

7 

1600 

1.05 

0.26 

0.99 

2.6 

6.5 

10.4 

18.8 

38.7 

99.6 

64 

2.9 X 10 ~ 5 

16 

1600 

0.84 

9.7 

3.1 

7.4 

17.8 

29.6 

51.2 

100 . 

247. 

47 

1.4 x 10 “ 5 

3 

1600 

0.70 

3.8 

12.0 

24.2 

48.5 

73.0 

122 . 

234. 

554. 

45 

6.8 x 10 —6 

12 

1600 

0.56 

8.5 

26.5 

57.5 

124. 

198. 

345. 

- 

452. c 

6.4 C 

2.3 x 10“ 6 

15 

1600 

0.35 

90.5 

255. 

- 

- 

- 

- 

- 

348. c 

0 . 68 c 

2.9 x icr 7 

8 

2200 

0.56 


_ 

_ 


— 

0.02 

0.09 

0.22 

40 

- 

9 

2200 

0.35 

— 

— 

0.03 

0.12 

0.24 

0.53 

1.2 

2.88 

53 

1.1 x 10“ 3 

10 

2200 

0.21 

0.10 

0.40 

0.97 

2.0 

3.0 

5.1 

10.1 

23.8 

49 

1.6 x 10 -4 

11 

2200 

0.14 

1.8 

5.5 

11.5 

21.9 

29.8 

43.3 

67.0 

123. 

52 

1.3 x 10“ 5 

17T^ 

2200 

0.10 

10.0 

28.0 

55.0 

95.0 

126. 

169. 

232. 

244. c 

11 .OP 

2.7 x 10 “ 6 


a Annealed at test temperature for 2 hours in hydrogen before testing. 
b 0.05-cm-thick sheet with 0.64-cm by 2.54-cm gage section. 
c Test terminated prior to rupture at time and strain indicated. 
d T denotes specimen transverse to rolling direction. 

stage) creep as do the curves obtained at higher stresses at 2200°C. Curves for 0.141 kg/mm 
and lower at both 2200°C and 24^6 show no apparent primary or secondary stages of creep. 
The creep rates reported in Table 1. 5 are essentially instantaneous creep rates in the re¬ 
gion near time zero. This change in curve shape from the more conventional creep curves 
obtained at the higher stresses may be related to a change in creep mechanism at low 
stresses. 

Sherby 31 showed that when the secondary creep rate, divided by the diffusion coefficient, 
is presented as a function of the stress divided by the modulus a single curve is obtained 
related to the mechanism of creep. The result of correlating the AC molybdenum data in 
this manner is presented in Figure 1.33. The 1600°C and 2200°C data show good correla¬ 
tion with a slope of 4.85; Sherby reported that a slope of 4 to 5 is normally obtained. The 
1200°C data show a displacement that is not understood. The deviation of the data for 2400°C 
is apparently related to the effect of the creep-curve shape on the creep rate discussed 
above. 

31 


Sherby, loc. cit. 



46 


TABLE 1.7 

CREEP-RUPTURE RESULTS 9 FOR WROUGHT, ARC-CAST, UNALLOYED MOLYBDENUM SHEET, b 

M{9)4, TESTED IN HYDROGEN 


Test Conditions Rupture_ Linear 


Specimen Temperature, 
No. °C 

Stress, 

kg/mm ^ 

0.2% 

0.5% 

Time to Indicated Strain, hr 
1% 2% 3% 

5% 

10% 

Time, 

hr 

Elongation, 

% 

Creep Rate, 
min - 1 

40 

1200 

7.03 

_ 

_ 

_ 

0.01 

0.04 

0.12 

0.35 

1.06 

70 

3.6 x 10“ 3 

41 

1200 

5.62 

- 

- 

0.04 

0.20 

0.42 

0.96 

2.63 

9.70 

78 

4.8 x 10" 4 

42 

1200 

4.22 

- 

0.07 

0.34 

1.15 

2.25 

4.93 

12.4 

33.2 

72 

i.i x icr 4 

44 

1200 

3.37 

0.12 

0.42 

1.13 

3.05 

6.00 

13.6 

34.2 

98.2 

84 

4.0 x 10 -5 

36 

1200 

2.81 

0.76 

1.93 

4.62 

13.2 

24.8 

50.4 

112. 

214. 

44 

1.3 x 10“ 5 

43 

1200 

2.11 

5.20 

14.6 

37.5 

94.0 

159. 

278. 

611. 

1742. 

52 

2.4 x icr 6 

11 

1600 

1.41 

0.01 

0.08 

0.33 

1.03 

1.88 

3.71 

8.33 

17.5 

34 

1.8 x icr 4 

9 

1600 

1.05 

0.65 

1.35 

2.67 

5.30 

8.60 

14.8 

29.7 

59.3 

57 

5.o x icr 5 

29 

1600 

1.05 

0.26 

0.82 

2.09 

5.50 

9.00 

16.0 

33.0 

81.7 

51 

4.8 x 10~ 5 

30 

1600 

0.844 

1.20 

2.80 

6.00 

14.0 

23.4 

42.2 

87.2 

241. 

52 

1.8 x 10~ 5 

10 

1600 

0.703 

2.80 

7.15 

15.8 

37.2 

61.0 

110. 

228. 

488. 

35 

6.8 x 10 -6 

12 

2200 

0.352 

_ 

0.01 

0.04 

0.15 

0.27 

0.55 

1.25 

3.40 

78 

1.2 x 10-3 

13 

2200 

0.281 

- 

0.03 

0.15 

0.48 

0.87 

1.71 

3.72 

9.92 

52 

4.0 x 10 -4 

26 

2200 

0.211 

0.11 

0.58 

1.35 

2.90 

4.45 

7.45 

13.6 

32.1 

73 

1.1 x 10 -4 

27 

2200 

0.176 

0.54 

1.76 

3.82 

7.93 

11.9 

19.5 

35.0 

89.0 

82 

4.1 x 10“ 5 

28 

2200 

0.141 

5.45 

14.5 

27.0 

49.0 

68.5 

99.4 

145. 

247. 

50 

7.2 x 10“ 6 

14 

2200 

0.141 

5.00 

11.1 

20.3 

37.2 

53.3 

77.3 

108. 

200. d 

38 

8.5 x 10“ 6 

49T 

2400 

0.211 

_ 

0.02 

0.04 

0.12 

0.20 

0.36 

0.76 

1.88 

48 

2.0 x 10- 3 

60 

2400 

0.190 

0.02 

0.06 

0.14 

0.30 

0.56 

0.78 

1.52 

3.89 

52 

1.0 x 10- 3 

48T 

2400 

0.176 

— 

0.02 

0.07 

0.19 

0.32 

0.59 

1.13 

2.80 

54 

1.2 x 10-3 

23 

2400 

0.141 

0.12 

0.48 

1.10 

2.36 

3.44 

5.18 

8.37 

17.0 

63 

1.4 x 10~ 4 

53T 

2400 

0.120 

0.63 

1.90 

3.81 

6.86 

9.25 

13.0 

19.2 

28.3 

46 

4.4 x 10 —6 

47T 

2400 

0.105 

2.05 

4.90 

8.52 

14.3 

19.0 

25.9 

35.3 

48.5 

38 

— 

54T 

2400 

0.084 

15.1 

31.0 

47.2 

58.9 

66.1 

76.9 

88.7 

101. 

32 

2.6 x 10-® 


a Annealed at test temperature for 2 hours in hydrogen before testing. 
b 0.05-cm-thick sheet, 0.64-cm by 2.54-cm gage section. 

C T denotes specimen transverse to rolling direction, all others parallel to rolling direction. 
d Power failure at 88 hours caused increased strain rate. 


TABLE 1.8 

CREEP-RUPTURE RESULTS 9 FOR WROUGHT, ARC-CAST, UNALLOYED MOLYBDENUM ROD, b 
__M(8)1, TESTED IN HYDROGEN 


Test Conditions Rupture Linear 


Specimen Temperature, 
No. °C 

Stress, 
kg/mm ^ 

0.2% 

0.5% 

Time to Indicated Strain, hr 
1% 2% 3% 

5% 

10% 

Time, 

hr 

Elongation, 

% 

Creep Rate, 
min - 1 

9 

1600 

1.76 

— 

0.02 

0.12 

0.44 

0.82 

1.60 

3.30 

5.18 

64 

4.5 x 10 -4 

10 

1600 

1.41 

0.01 

0.14 

0.49 

1.40 

2.37 

4.35 

7.47 

11.7 

72 

1.7 x icr 4 

11 

1600 

1.05 

0.11 

0.74 

2.30 

5.60 

8.92 

13.4 

23.3 

58.2 

53 

5.0 x 10“ 5 

13 

1600 

0.844 

1.88 

3.48 

7.40 

16.7 

26.7 

48.7 

105. 

222. 

82 

1.4 x 10 -5 

12 

1600 

0.703 

5.20 

17.0 

26.5 

45.0 

68.5 

120. 

244. 

357. c 

15 

7.0 x 10 —6 

3 

2200 

0.316 

— 

0.02 

0.07 

0.24 

0.44 

0.82 

1.60 

2.66 

86 

8.9 x 10 -4 

5 

2200 

0.246 

0.02 

0.09 

0.17 

0.76 

1.33 

2.52 

5.20 

9.60 

78 

2.8 x 10 —4 

2 

2200 

0.197 

0.26 

0.79 

1.67 

3.43 

5.19 

8.70 

11.6 

17.0 

44 

9.3 x 10~ 5 

4 

2200 

0.162 

1.05 

3.32 

6.11 

11.56 

17.0 

26.2 

31.5 

50.0 

38 

3.1 x 10“ 5 

1 

2200 

0.141 

2.66 

6.00 

10.6 

19.2 

27.8 

48.5 

77.5 

145. 

64 

1.9 x 10“ 5 


a Annealed at test temperature for 2 hours in hydrogen before testing. 
b 0.41-cm-dia. by 2.54-cm-long gage section. 
c Test terminated, no rupture. 


000 01 


47 



jsd 'ssejjg 


Stress-rupture results for wrought, arc-cast unalloyed molybdenum 
tested in hydrogen 
















































10,000 


48 




red 'sssjjg 


Creep rate results for wrought, arc-cast unalloyed molybdenum teste< 
in hydrogen 































































juaojad 'wens 


Constant load creep curves for wrought, arc-cast molybdenum sheet 
at 1200°, 1600° 2200°, and 2400°C 












50 



Fig. 1.33 — Correlation of linear creep rate data for arc-cast molybdenum 
tested in hydrogen 

Monkman and Grant 32 reported that when the log of the time-to-rupture is plotted against 
the log of the linear creep rate, a straight-line relationship is obtained with a slope of 
minus m, in accordance with: 

log t R = log C - m log e s (1.12) 

where 

tp = rupture time 
e s = linear creep rate 
m, C = constants 

When m is equal to one, as it has been found to be for a number of materials, the equation 
becomes: 



(1.13) 


The data for molybdenum discussed above are shown in this form in Figure 1.34. The curves 
for both rod and sheet material are shown to possess a slope of minus 1 but are not coinci¬ 
dent. This displacement is due to the consistent difference in rupture life (Figure 1.30) since 


32 


F. C. Monkman and N. J. Grant, "An Empirical Relationship Between Rupture Life and Minimum Creep Rate in Creep Rupture 
Tests/' Proc. ASTM, Vol. 56, 1956, p. 593. 



Time to rupture (tp), hours 


51 



Linear creep rate (e $ ), min ^ 


Fig. 1.34 — Rupture life versus creep rate for arc-cast molybdenum 
showing correlation of data to tp e s = C 


good agreement for the linear creep rate is shown in Figure 1. 31. The data points in Figure 
1.34 that deviate significantly from the straight lines are those showing deviations from the 
linear creep rate plot of Figure 1.31. The effect of the change in the creep curve shape is 
shown in this Monkman-Grant type of plot. Only those data at the lower stresses and higher 
temperatures, where no primary creep was observed, deviate from the lines in Figure 1. 31. 


( ^Molybdenum-Base Alloys j 


Kmo 


5Remand Mo - 5W^ 


-^> i-S3 

Creep-rupture data were reported 33 for PM Mo — 5Re and Mo-5W 
(both at. %) alloy sheet fabricated at GE-NMPO. Sheet thicknesses were 0.05 cm and 0.08 cm 
for Mo — 5W and Mo — 5Re, respectively, with testing performed at 1600°C and 2200°C in hydro¬ 
gen. Data obtained in these tests are presented in Figures 1. 35 and 1. 36; the data for AC mo¬ 
lybdenum are shown for comparison. Each alloy is seen to be stronger than unalloyed, AC 


33 


GEMP-1002, p. 16. 








































53 


molybdenum. Two isotherms are presented for Mo — 5Re at 1600°C to reflect the different 
behavior observed in specimens from two different sheets of this material. Actually these 
two molybdenum alloys exhibit about the same creep-rupture behavior. 

rn m ^ p.gfa 

\ Mo — 50ReJ - Stress-rupture and creep data (constant load) for wrought, PM Mo — 50Re 
(wt %) alloy sheet were previously presented 34 for lot 1 material at 2200°C and for lot 2 ma¬ 
terial at 1600°, 2200°, and 2400°C. The lot 2 results were obtained using specimens from 
sheet 1. Additional tests were performed at 1600°, 2200°, and 2400°C using specimens ob¬ 
tained from sheet 2 of the lot 2 material. The objective was to compare results for two dif¬ 
ferent sheets of the same wrought, powder-metallurgy material obtained from the same lot. 
In addition, the lot 2, sheet 1 results indicated that at the higher temperatures and lower 
stresses, diffusional creep may be the controlling mechanism based on the increased slope 
of the stress versus linear creep rate plot. Further studies were therefore performed to 
verify this. 

Table 1.9 and Figures 1.37 and 1.38 give the stress-rupture and creep data obtained 
for the lot 2, sheet 2 material. The sheet 2 results appear to be consistent with the sheet 
1 material in terms of stress-rupture, linear creep rate, and ductility at 2200°C and 2400°C. 
At 1600°C, the rupture life is greater whereas the linear creep rate and ductility are lower 
for the sheet 1 material. 


Creep curves for lot 2, sheet 2 specimens tested at the lowest stress for each of the 
three temperatures involved are shown in Figure 1.39. At 1600°C, the curve shows third- 
stage-type creep from essentially time zero, while the 2200°C curve gives the more con¬ 
ventional type of creep curve. At 2400°C, the creep curve displays two regions of essen- 


TABLE 1.9 

CREEP-RUPTURE RESULTS 3 FOR WROUGHT, POWDER-METALLURGY Mo - 50Re SHEET b 


TESTED IN HYDROGEN (LOT 2, SHEET 2) 



Test Conditions 








Rupture 

Linear 

Specimen Temperature, 

Stress, 



Time to Indicated Strain, 

hr 


Time, 

Elongation, 

Creep Rate, 

No. 

°C 

kg/m 

0.2% 

0.5% 

1% 

2% 

3% 

5% 

10% 

hr 

% 

min”^ 

2-2 

1600 

3.37 

0.05 

0.10 

0.16 

0.25 

0.32 

0.46 

0.74 

2.02 

102 

_ 

-3 

1600 

2.81 

0.18 

0.40 

0.61 

0.95 

1.23 

1.67 

2.45 

5.25 

92 

8.0 x 10“ 5 

-4 

1600 

2.11 

0.72 

1.44 

2.42 

3.82 

4.90 

6.57 

9.30 

16.7 

85 

— 

-5 

1600 

1.76 

1.37 

2.90 

4.90 

8.10 

10.6 

14.7 

21.6 

37.9 

74 

2.3 x 10~ 5 

-6 

1600 

1.41 

2.85 

6.85 

12.4 

20.0 

25.5 

33.8 

47.2 

81.0 

60 

1.3 x 10“ 5 

-1 

1600 

1.05 

7.45 

18.2 

34.2 

59.6 

79.9 

111. 

162. 

262. 

58 

4.5 x 10“ 6 

2-13 

2200 

0,352 

0.29 

0.83 

1.87 

3.76 

5.51 

8.48 

12.6 

14.8 

21 

9.5 x 10 -5 

-15 

2200 

0.211 

0.72 

1.88 

4.06 

8.85 

14.0 

24.9 

47.6 

70.3 

24 

2.9 x 10~ 5 

-17 

2200 

0.141 

1.03 

2.77 

5.52 

14.4 

23.2 

42.0 

88.2 

145. 

24 

1.7 x 10“ 5 

-18 

2200 

0.105 

1.28 

3.50 

7.60 

16.5 

26.2 

46.7 

94.1 

174. 

45 

1.6 x 10“ 5 

2-8 

2400 

0.211 

0.14 

0.34 

0.68 

_ 

_ 

_ 

_ 

1.09 c 

1.8 

2.5 x 10-4 

-30 

2400 

0.211 

0.15 

0.37 

0.72 

1.33 

1.83 

2.51 

3.20 

3.48 

30 

2.4 x 10 -4 

-21 

2400 

0.141 

0.24 

0.57 

1.12 

2.19 

3.13 

4.63 

6.75 

8.65 

32 

1.5 x 10- 4 

-25 

2400 

0.105 

0.36 

0.89 

1.78 

3.56 

5.34 

8.30 

13.2 

20.9 

46 

9.4 x 10 ~ 5 

-29 

2400 

0.070 

0.61 

1.48 

2.94 

5.84 

8.75 

14.6 

27.5 

70.0 

44 

5.7 x 10—5“ 

-24 

2400 

0.035 

0.97 

2.66 

5.50 

11.3 

18.4 

33.6 

97.7 

224. 

36 

( 1.2 x 10 “ 5 
\2.9x 10— 5<J 

1-21 

2200 e 

0.141 

1.55 

3.93 

8.80 

18.6 

28.6 

48.5 

85.9 

112. 

20 

1.7 x 10~ 5 


3 Annealed at test temperature for 2 hours in hydrogen before testing. 
b 0.05-cm-thick sheet, 0.64-cm by 2.54-cm gage section. 
c Test terminated; load train support failed. 

^Diffusional creep rate. 

e Annealed at 2400°C for 2 hours in hydrogen before testing (lot 2, sheet 1). 


34 


GEMP-67, p. 15. 



Stress, psi 


54 



Fig. 1.37 - Stress-rupture results for wrought, powder-metallurgy Mo - 50Re 
(wt %) sheet (lot 2) tested in hydrogen 


tially linear creep rate followed by third-stage creep. If these curves are analyzed in terms 
of the creep-rate curve given in Figure 1.38, it appears that the initial portion of the creep 
curve at 2400°C and 0.035 kg/mm^ (Figure 1.39) is diffusion controlled since the creep- 
rate curve at 2400°C (Figure 1. 38) yields a slope of essentially 1. This is consistent with 
the Nabarro-Herring 35 model for diffusional creep which indicates stress to be directly 
proportional to creep rate. This is further confirmed by the analysis approach proposed 
by Sherby 36 for poly crystalline metals above one-half the absolute melting temperature. 
Figure 1.40 shows good correlation of the results using this approach. The curve at the 
low-stress levels also yields a slope of 1 indicating diffusional creep. At the higher stress 
levels the curve slope is 2.7. According to Sherby, when a slope approximating 3 is ob¬ 
tained, the creep mechanism is dislocation glide based on the microcreep theory of 
Weertman. 37 He showed that when dislocation motion under stress is controlled by the ve¬ 
locity of solute drag along the dislocation line the creep rate is given by 



(1.14) 


35 

erring, loc. cit. 

36 Sherby, loc. cit. 

37 J. Weertman, "Steady-State Creep of Crystals," J. Appl. Phys., Vol. 28, 1957, p. 1185. 




55 



Linear creep rate, mirf ^ 


Fig. 1.38 — Creep rate results for wrought, powder- 
metallurgy Mo — 50Re (wt %) sheet 


where k is a constant depending on the interaction force between the solute atmosphere and 
the dislocation, D s is the solute diffusion coefficient, andGis the shear modulus. This equa 
tion predicts that the creep rate of solid solutions should be proportional to the third power 
of stress in contrast to the five power law dependence observed for pure metals. 


Apparent activation energy calculations for creep were made based on the linear creep 
rate data obtained at the same stress level for two different temperatures. Data between 
1600° and 2200°C gave a value of 84 kcal/mole and the data between 2200° and 2400°C gave 
a value of 121 kcal/mole. 


A value of the diffusion coefficient for Mo - 50Re was calculated based on the Nabarro- 
Herring equation for creep rate: 

10 D ab 3 


=• 


(1.15) 


k T IT 


where e g is the steady-state creep rate in see"**, D is the diffusion coefficient in cm 2 /sec, 
a is the stress in dynes/cm 2 , k is Boltzmann 1 s constant (1.38 x 10~16 ergs per degree), 

T is the absolute temperature, L is the grain diameter in centimeters, and b is the lattice 
spacing in centimeters. Based on the test data obtained at 2400°C and 0.070 kg/mm 2 for 
lot 2, sheet 1 material, the value of the diffusion coefficient was calculated to be 1.143 x 
10“7 cm 2 /sec. The values of grain size and lattice spacing used were based on actual meas 


















56 


14 


12 


10 


8 

c 

a 


4 


2 


0 

0 20 40 60 80 100 120 140 160 180 200 

Time, hours 


Fig. 1.39 - Creep curves for wrought, powder-metallurgy Mo - 50Re (wt %) sheet (Lot 2) 
tested in hydrogen 

urements of 285 microns and 3.16 x 10"8 cm, respectively. The two activation energy 
values discussed above were applied to the conventional diffusion expression 

-_Q_ 

RT 

D = D 0 e (1.16) 

where D is the self-diffusion rate of atoms, D 0 is a constant, Q is the activation energy for 
self-diffusion, R is the gas constant, and T is the absolute temperature. Diffusion rate values 
could then be expressed by the relations shown in Figure 1.40. 

n w, 

Mo - 30W -)The stress-rupture characteristics for both wrought, unalloyed tungsten and 
molybdenuriTsheet were shown to vary above one-half the absolute melting temperature de¬ 
pending on whether the material was fabricated by an AC or PM process. 38 * 39 Results obtained 
for PM molybdenum varied depending on the material source (i. e., vendor); results for AC 
material were essentially identical independent of the source. Similar findings were observed 
for tungsten. When the stress-rupture test results for AC molybdenum and AC tungsten were 
analyzed in terms of homologous temperature, identical results were obtained; 40 this did not 
occur for the PM materials. 

The phase diagram for the tungsten-molybdenum system reveals that these metals form 
solid solutions over the complete range of compositions. The rupture-strength relationship 
for alloys of the two metals may be expected to be the same as that for the elements on a 
homologous temperature basis except for the possible strengthening effect of the lattice dis- 



3h tt 

P. N. Flagella, High-Temperature Stress-Rupture Characteristics of Mo and IVlo Alloys/' AIAA Journal, Vol. 5 1967 p. 281 

39 tt 

P. N. Flagella, High Temperature Creep-Rupture Behavior of Unalloyed Tungsten," GE-NMPO, GEMP-543, August 1967. 
40 GEMP-475A, p. 16. 



























































58 


tortion by the substitution of tungsten atoms for molybdenum atoms. Because of the simi¬ 
larity between tungsten and molybdenum, the stress-rupture characteristics of wrought, 
arc-cast Mo — 30W (wt %), purchased commercially (heat 30W 7694), were evaluated. 41 

The Mo - 30W alloy was of relatively high purity containing 69, 12, 1, and 41 ppm of C, 

O, H, and N, respectively. After rolling to sheet, with intermediate anneals, the material 
was stress-relieved for 1 hour at 1175°C in vacuum. Stress-rupture and creep tests were 
performed at 1600°C and 2200°C in hydrogen after annealing for 2 hours at the test tempera¬ 
ture. Creep curves for this material are presented in Figure 1.41. For all stress levels 
indicated, at constant temperature, the strain is presented as a function of the time to give 
10 percent strain. In this way the curve shapes over a wide range of stresses are easily 
shown and compared. All the creep curves for the Mo - 30W alloy display a third-stage- 
type creep from essentially time zero. As shown, the ratio of t^/tio% (Figure 1.41) de¬ 
creases with decreasing stress at the same temperature. This is apparently not related 
to the fact that at 1600°C the elongation for the material at rupture decreases with decreas¬ 
ing stress since this is not the case at 2200°C. 

Figure 1.42 gives the stress-rupture curves for the Mo-30Wmaterial atl600°C and2200°C 
along with curves for wrought, AC, unalloyed W, Mo, and W - 25 Re (wt %) for comparison. 

As expected, the addition of tungsten to molybdenum increases the rupture life. All the ma¬ 
terials exhibit a linear relationship between log stress and log rupture time. Isotherms for 
the Mo - 30W and W — 25Re alloys are nearly parallel and have slopes greater than those 
shown for the unalloyed tungsten and molybdenum. 

The 100-hour rupture data from Figure 1.42 are shown in Figure 1.43 in terms of the 
homologous temperature (percent of the absolute melting temperature). Agreement is shown 
between the tungsten and molybdenum data but the Mo - 30W data show a 17 percent increase 
in the rupture stress for the same percent of the absolute melting temperature. This may be 
due to the lattice distortion effect mentioned above. The W - 25Re data reflect a different 
behavior. 

Figure 1.44 shows typical specimens after testing at 1600° C and 2200° C and the resulting 
photomicrographs. The effect of temperature on the grain size is quite apparent being 145 
microns at 1600°C and 490 microns at 2200°C. Significant intergranular separation is also 
noted. For the specimen tested at 1600°C this appears to occur primarily transverse to the 
^stress direction. At 2200°C, the effect is not apparent due to the large grain size involved. 

RHENIUM J 

Studies of the creep-rupture behavior of wrought, PM (lot 2), unalloyed, rhenium sheet 
(0.05 cm thick) were reported 42 *. 43 for tests at 1600°, 2200°, and 2600°C in hydrogen. Rup¬ 
ture data obtained in these evaluations are presented in Figure 1.45 along with some pre¬ 
viously reported 44 data for lot 1 rhenium which had been obtained from the same vendor at 
an earlier date. Some slight difference between lots 1 and 2 is quite evident from Figure 
1.45 and was attributable 43 to different sintering procedures which led to different grain 
sizes in the two lots of material. Cavitation was observed in the grain boundaries of the 
material from both lots at fracture. Linear creep rate data for the lot 2 material are pre¬ 
sented in Figure 1.46. In terms of rupture strength, the lot 2 material has about the same 
strength as AC tungsten at 1600°C and is decidedly stronger than AC tungsten at all higher 
temperatures. 

41 GEMP-1002, p. 11. 

42 GEMP-67, p. 11. 

43 "AEC Fuels and Materials Development Program Progress Report No. 69," GE-NMPO, GEMP-69, September 29,1967, p. 11. 

44 "Fourth Annual Report — High-Temperature Materials and Reactor Component Development Programs, Volume I — 

Materials," GE-NMPO, GEMP-334A, February 26, 1965, p. 23. 




t/MO% 


Fig. 1.41 - Creep curves for wrought arc-cast Mo — 30W (wt %) sheet tested in hydrogen 








Stress, psi 


Time to rupture, hours 


Fig. 1.42 — Stress-rupture results for wrought arc-cast Mo — 30W (wt %) sheet 
compared to W, Mo, and W — 25Re (wt %) 



Fig. 1.43 — Stress required to cause rupture in W, Mo, Mo 
in 100 hours 


30W, and W 


25R- 

















Stress, kg/mm 2 Stress, kg/mm 





















63 


NIOBIUM AND NIOBIUM ALLOYSJ JjP h ~ ^ j - 

Creep-rupture testing of niobium was based on experience gained from tests 45 of tanta¬ 
lum at high temperatures, in either high-purity hydrogen or argon. These tests showed 
that the material becomes significantly contaminated with nitrogen and oxygen causing 
strengthening and embrittlement. The primary source of the contamination is not the test 
gas, since high-purity gas was used, but the result of the furnace components outgassing 
at high temperatures. Placing a protective tantalum foil around the specimen being tested 
decreased the contamination level but did not eliminate it. No such problem was encountered 
with the non-reactive refractory metals (W, Mo, Re) or their alloys. This is apparently due 
to the low level of solubility for interstitials. 

Niobium, tantalum, and vanadium (Group V elements) are highly reactive and have a high 
solubility for the interstitial elements and hence are more difficult to evaluate in the un¬ 
contaminated condition. To overcome some of the contaminating problems associated with 
evaluating the creep-rupture properties of niobium, a test system was modified by incor¬ 
porating an ion pump so that testing could be performed in a relatively high vacuum. In 
addition, modifications were incorporated to allow the system to outgas after reaching test 
temperature with the test specimen in a cold region of the furnace until outgassing had sub¬ 
sided and a good vacuum level was reached (~1 x 10“6 Torr). By operating the system at a 
higher temperature than required for testing, outgassing was accelerated and possible out¬ 
gassing during test was reduced. After outgassing was complete and thermal equilibrium 
obtained at the test temperature, the test specimen was positioned in the hot zone of the 
furnace, annealed for 2 hours, loaded to the desired stress level and the creep-rupture 
test performed. 

This procedure was used to test four wrought, AC, unalloyed niobium sheet samples and 
one alloy, Cb-753 (Nb-5V-1.25Zr, wt %), sample at 1600°C. Impurities in ppm for the as - 
received materials were as follows: for Nb; C = 22, O = 282, N = 16, H = 5; for Cb-753; 

C = 15, O = 208, N = 44, H = 2. The test results are given in Table 1.10 and the creep curves 
shown in Figure 1.47. The stress-rupture data are presented in Figure 1.48 including pre¬ 
viously published 46 data for the alloy Cb-753. The greater rupture life for the alloy over the 
unalloyed niobium at 1600°C and 0. 703 kg/mm 2 is obvious. The fact that the alloy showed 
considerable elongation at rupture (130%) after 342 hours at 1600°C is an indication of a 
high-purity non-contaminating environment. Additionally, the hardness of the alloy decreased 


TABLE 1.10 

CREEP-RUPTURE RESULTS 3 FOR WROUGHT, ARC-CAST, UNALLOYED NIOBIUM SHEET 6 
AND ALLOY Cb-753 SHEET 0 AT 1600°C TESTED IN VACUUM d 


Material 

Specimen 

No. 

Stress, 

kg/mm^ 

0.2% 

0.5% 

Time to Indicated Strain, hours 

1% 2% .3% 5% 

10% 

Rupture 

Time, Elongation, 
hr % 

Linear 
Creep Rate, 
min - 1 

Nb 

1-4 

0.211 

0.95 

2.60 

5.50 

12.2 

20.0 

35.3 

70.0 

182. 

41 

1.9 x 10-3 


-3 

0.352 

- 

0.03 

0.18 

0.86 

1.88 

4.28 

9.60 

19.3 

30 

5.1 x 10-3 


-5 

0.562 

- 

- 

- 

- 

0.01 

0.09 

0.51 

1.83 

41 

1.9 x 10-3 


-1 

0.703 

- 


- 

- 

- 

0.04 

0.18 

0.54 

30 

1.7 x 10~ 5 

Cb-753 

1 

0.703 

0.79 

2.52 

6.00 

14.4 

23.6 

42.8 

90.0 

342. 

130 

1.4 xIO- 4 


a Annealed at 1600°C for 2 hours in vacuum before testing. 
^0.053-cm-thick sheet, 0.64-cm x 2.54-cm gage section. 
c 0.15-cm-thick sheet, 0.64-cm x 2.54-cm gage section. 
d From 1 td 2 x 10“ 6 Torr. 


^ 6 "High-Temperature Materials Program Progress Report No. 25," GE-NMPO, GEMP-25A, July 31,1963, p. 12. 
46 Union Carbide Corp., Stellite Division, F-30, 269-1, September 1964. 





tested at 1600°C and in a vacuum of 1 x 10 6 Torr 

















































65 



10 

Time to rupture, hours 


Fig. 1.48 — Stress-rupture results for unalloyed niobium and alloy Cb-753 


r, 


from 183 DPH before test to 160 DPH after test. Chemical analysis and metallographic eval¬ 
uations are planned for all samples to determine the extent of any otherwise undetected 
contamination.. 


CO NSTANT-STRESS CREEP TESTING 


7/ 


All stress-rupture and creep data reported in this program to date have been obtained by 
constant-load testing. As the test specimen elongates, the cross-sectional area decreases 
and the stress increases. Most data reported in the literature are obtained in this manner, 
but to better understand the basic characteristics and mechanisms involved in creep, test 
data obtained at constant stress are desirable. 


One high-temperature creep-rupture test stand was modified to perform tests at con¬ 
stant stress. Figure 1.49 shows a schematic diagram of the system. The electro-optical 
strain measuring system 47 automatically records strain versus time for the test specimen. 
On the basis of the strain measurement, the automatic load-control unit decreases the load 
on the specimen to maintain a constant stress. The assumption involved with this approach 
is that the strain occurs uniformly over the gage length of the specimen and that the in¬ 
crease in strain is proportional to the decrease in cross section (i. e., constant volume 
deformation). This is a valid assumption until local necking occurs which, on the basis 
of constant stress, does not occur until third-stage creep is observed. 


Several constant-stress creep tests were performed to evaluate the system. Unalloyed 
molybdenum sheet specimens were tested at 1600°C and a constant stress of 2.11 kg/mm^ 
in hydrogen and compared to the results obtained at constant load for the same material. 
One such comparison is shown in Figure 1.50. As expected, due to the continually decreas¬ 
ing load with increasing strain, the constant-stress tests show a longer time to reach the 
same values of strain observed in the constant-load tests. 

Previous evaluations of equation forms have shown that a third-degree polynomial in t*/® de¬ 
scribes the primary and secondary stages of creep for a constant-load test very well. This 


L. McCullough and P. N. Flagella, “Experimental Techniques Employed in Stress-Rupture and Creep Measurements 
to 3000°C/' GE-NMPO, GE-TM 65-5-10, April 1965. 

























































67 


e = - 7.28 x 10“ 4 + 4.09x 10- 2 t 1/3 -2.20 x 10- 2 t 2/3 + 4.90x 10" 2 t 

t 1 ~ 2.60 x 10~ 4 

Time: 0.0042 1.6 hr 



€ = 3.48 x 10~ 4 + 3.99 x 10~ 2 t 1/3 -4.73 x 10" 3 t 2/3 + 3..17xl0- 2 t 
fi =2.90 x 10- 4 
Time: 0.005 -> 2.016 hr 



Fig. 1.50 — Comparison of constant-load and constant-stress creep tests of 
Mo sheet at 1600°C and 2.11 kg/mm^ in hydrogen 

same equation form is effective in describing the constant-stress test as shown in Figure 
1.50. 

Additional tests of wrought, AC molybdenum sheet were performed at both 1600°C and 
2200°C in hydrogen to compare the second-stage creep rates. Figures 1. 51 and 1.52 show 
the creep curves obtained with the time scale normalized to the time required to give 10 
percent strain. This permits a more direct comparison of the creep curves to evaluate 
the effect of stress on the strain. Although small differences are observed between the 
constant-load (increasing stress) and constant-stress creep curves, the differences are 
not significant until approximately 15 percent strain is obtained which is beyond the region 
of the linear creep rate for the constant-load test. The linear creep rates for the two types 
of tests are essentially the same. This being the case, constant-load creep test results 
may be analyzed in terms of the linear (secondary) creep rate with no significant error. 

The creep curves presented in Figures 1.51 and 1.52 for AC molybdenum, display the 
classical form at constant load and, therefore, permit analysis, correlation, and inter¬ 
pretation of linear creep rate behavior. Any conclusions reached in this type of analysis 
should be consistent with those based on an analysis of constant-stress data. Some ma¬ 
terials, notably solid-solution alloys such as W - 25Re, Mo - 30W, and some tantalum 
alloys, have displayed creep curves consisting of only third-stage creep when tested at 
constant load. Similar curves were obtained for AC molybdenum at 2400°C (0.93 T m ) as 
shown in Figure 1. 32. These data do not permit an analysis to be performed on the basis 
of linear creep rate. 

To evaluate the difference due to constant-stress testing, constant-stress tests of Mo-30W 
(wt %) were performed at 2200°C and 0. 352 kg/mm 2 . Creep curves for two constant-stress 
tests and one constant-load test are shown in Figure 1.53. The results show a marked dif- 




68 



Fig. 1.51 — Comparison of constant-load and constant-stress creep curves for arc-cast Mo sheet 
tested at 1600°C and 0.844 kg/mm 2 in hydrogen 



Fig. 1.52 — Comparison of constant-load and constant-stress creep curves for arc-cast Mo sheet 
tested at 2200°C and 0.35 kg/mm^ in hydrogen 


ference in the curve shapes. Whereas, the constant-load test displays a curve of increasing 
creep rate from time zero, one of the constant-stress tests gave essentially a linear creep 
curve to a strain of 65 percent. This indicates that uniform deformation over the length of the 
gage section occurred to at least this strain level. Strain measuring beyond this level was 
not possible with the test equipment and the test was terminated. The lower constant-stress 
creep curve in Figure 1. 53 shows a decreasing rate of deformation after approximately 25 
percent strain. This was caused by a greater load removal rate than actually needed to 
maintain constant stress. Deformation of the specimen from this test was quite uniform 
over the total length of the gage section (2. 54 cm) after 36 percent strain as shown in Fig¬ 
ure 1.54. The same observation was made for the sample tested to 65 percent strain. 






69 



Fig. 1.53 - Comparison of constant-load and constant-stress creep curves for arc-cast 
Mo - 30W (wt %) tested at 2200°C and 0.352 kg/mm 2 



Fig. 1.54 - Mo - 30W (wt %) creep sample after testing at 2200°C and 0.352 
kg/mm 2 showing 36% strain (2.54-cm gage length)!Neg. P68-M2B) 

Computer analysis procedures were developed to analyze constant-load creep tests in 
terms of instantaneous creep rate and instantaneous stress (see creep analysis section 
under Tungsten) to determine the dependence of strain rate on stress. This technique was 
employed to analyze the experimental results obtained at constant stress. Figure 1. 55 shows 
the results for both types of testing. The constant-load test of this one sample shows the 
creep rate to be proportional to stress to the 4.6 power over the range from 0. 35 to 0. 5 
kg/mm 2 . The data points for the two constant-stress tests show reasonably good agreement 
with the constant-load test in terms of creep rate. The difference in the two constant-stress 
tests is probably due to experimental variations associated with precise temperature meas¬ 
urement at 2200°C, the normal precision being ±10°C at 2200°C. 

Experimental creep data obtained at both constant-load and constant-stress indicate that 
constant-load creep data may be analyzed in terms of constant stress with no significant 
error due to the increasing stress as the result of strain until non-uniform deformation 
occurs. This appears to apply for materials exhibiting the conventional creep curve, con¬ 
sisting of the three stages of creep, as well as a material exhibiting only third-stage-type 
creep. 





Creep rate, mm 



500 550 600 650 700 750 800 

Instantaneous stress (a), psi 


Fig. 1.55 - Comparison of constant-load and constant-stress 
creep tests in terms of creep rate for Mo - 30W 
(wt %) at 2200°C and 0.352 kg/mm^ 






71 


n 


SINGLE ^RYSTALS 


Unalloyed, powder-metallurgy tungsten sheet straps, 0. 05 cm thick by 1. 9 cm wide by 
19 cm long, are used to support the sheet-type test specimens during creep testing at 
1600° to 3000°C in hydrogen. During testing, large grains are sometimes developed in the 
straps with the sizes apparently dependent on temperature, time, and stress. One such 
test performed at 2200°C for 315 hours produced a single crystal of tungsten over approx¬ 
imately one-half the length of a strap. Using the Laue back-reflection X-ray diffraction 
technique, it was confirmed that the strap was a single crystal and had a nearly perfect 
(114) orientation with the stress axis along the (110) direction. 


Creep-rupture tests of polycrystalline tungsten have been shown to exhibit correctable 
characteristics consistent with other pure materials. The creep characteristics of single 
crystals are not understood, in general, and no data at high temperatures seem to be 
available for single-crystal tungsten. A creep-rupture test of the single-crystal sheet 
material was performed in hydrogen on a specimen having a 0.64-cm-wide by 2.54-em- 
long gage section fabricated by the electrical discharge machining process (Elox). Table 
1.11 lists the test conditions and results. These data show the single-crystal material to 
be stronger than polycrystalline, wrought, AC tungsten. 


Figure 1.56 shows the single-crystal test specimen before and after testing. The sig¬ 
nificant observation is that essentially no deformation occurred prior to rupture other than 
that occurring locally at the 0.01-cm holes used as fiducial marks for strain measurements. 
The mode of fracture is not clear but near the center of the gage length there is an indication 
of slip along some plane of the crystal. Investigations are underway to study this region of 
the specimen using X-ray diffraction, metallography, and the electron microscope. These 
studies should provide some explanation of the mode of deformation in single-crystal, PM 
tungsten at extremely high temperatures. 


TABLE 1.11 


CREEP RESULTS OF SINGLE-CRYSTAL, UNALLOYED 
TUNGSTEN SHEET TESTED IN HYDROGEN 


Temperature, 

°C 

Stress, 

9 

kg/mm 

Time, 

hr 

Strain, 3 % 

2400 

0.46 

22 

0 (5.0) 


1.05 

22 

0 (rupture at 3.1 hr - 85%) 

11 : 

1.05 

22 

0 (rupture on loading) 

2800 

1.05 

53.6 

Rupture 


a Values In parentheses refer to data for polycrystalline 
arc-cast tungsten. 


STRESS-RUPTURE P. 


TER ANALYSIS 




"7 


%3 


A comprehensive study of stress-rupture parameters was completed. 48 In addition to a 
detailed treatment of the three most common parameters, Larson-Miller, Dorn, and 
Manson-Haferd, appropriate attention was given the Graham-Walles, Murry, Chitty- 
Duval, and Brozzo parametric approaches. In each case, special consideration was given 
to the mathematical procedures employed in identifying parameter constants and in de¬ 
veloping an analytical expression relating parameters to stress. Also, a detailed compari¬ 
son was made of the relative effectiveness of these parameters in the correlation and 
extrapolation of experimental data. As part of this comparison, mathematical optimi¬ 
zation procedures were employed to re-evaluate numerous sets of experimental data 
which have appeared in the literature. 


48 


J. B. Conway, “Stress-Rupture Parameters: Origin, Calculation and Use/ 7 GE-NMPO, GEMP-555, June 30, 1967. 










Fig. 1.56 — Single-crystal, unalloyed W sheet before and after 
testing for 53.6 hours at 2800°C and 1.05 kg/mm 
(Neg. P67-4-23A; Inset, Neg. P67-3-44A) 







73 


Several typical developments evolved during the above study are presented below. 
C^Timken 35-15 Stainless Steel__^ 


In a recent study by Clauss, 49 the stress-rupture data for Timken 35-15 stainless steel 
were correlated in terms of the Larson-Miller parameter using a constant C = 20. A plot 
of this correlation is shown in Figure 1. 57a indicating a high degree of scatter about an 
average behavior given by the solid curve. Dashed lines are used to demonstrate that the 
data at each temperature seem to describe a separate line and, hence, poor correlation 
is indicated. Obviously, when the solid curve is used for extrapolation to lower stresses 
at a given temperature, rupture times are predicted which are much higher than those 
actually observed. 


When these same data were analyzed using the optimization procedures described previ¬ 
ously, 50 the results shown in Figure 1. 57b were obtained. In this case, the Larson-Miller 
constant is found to be 12. 2 and results in a very excellent correlation. Extrapolation of 
these results to lower stresses would seem to be capable of yielding fairly acceptable 
results. 


In Figure 1. 57c, the correlation obtained when C = 5 is presented. Here, too, when the 
C value is different from the optimum an extremely poor correlation results. It is appar¬ 
ent that as the C value is decreased, the isotherms assume different positions with respect 
to the average parameter curve until, at the optimum C value, all isotherms appear to be 
coincident with an average curve which can be drawn through all the points. Actually, the 
solid curve in Figure 1. 57b represents a polynomial developed in the least-squares opti¬ 
mization procedure applied to the Larson-Miller parameter. 


\~ Arc-Cast W^.Mo, and W - 25Re^ J 

A linear relationship was identified 50 for the stress-rupture data of Hastelloy N when the 
natural logarithm of the Dorn parameter was plotted as a function of the natural logarithm 
of sinh (Ccr), where C is a constant and a is the stress. A similar approach was employed 
in analyzing the stress-rupture data for arc-cast W, Mo, and W — 25Re. As shown in Fig¬ 
ure 1. 58, linear relations are obtained. Such plots can be compared to the standard master- 
rupture plots which usually yield a curved relationship. 


When the hyperbolic sine approach is employed, it is possible to define the linear type of 
relationships shown. With this approach it would appear that extrapolation beyond the ex¬ 
perimental stress range can be effected with more confidence than when the master rupture 
plot is curved. 


In a similar analysis the Larson-Miller parameter was employed, and, as in Figure 1.58, 
a linear relationship was defined using the hyperbolic sine stress function. These results 
are presented in Figure 1. 59 where excellent linearity is seen to exist. 


\ Timken 25-20 Stainless Steel 
x - 




72 


In an evaluation of the rupture data for Timken 25-20 stainless steel, Clauss employed 
the Larson-Miller approach, with C = 20, and found that a very unsatisfactory correlation 
resulted. On the parameter plot, data for each temperature described individual lines and 
no general correlation was noted. In the present study, these same data were analyzed using 
optimization procedures with the results shown in Figure 1. 60. It can be seen that all three 
parameters appear equally effective in the representation of these data, at least if the stand- 


4 ^F. J. Clauss, "An Examination of High-Temperature Stress-Rupture Correlating Parameters," Proc. of ASTM, Vol. 60,1960, p. 905. 
High-Temperature Materials Program Progress Report No. 63," GE-NMPO, GEMP-63, December 30, 1966, p. 13. 



74 



Fig. 1.57 — Larson-Miller plots of Timken 35-15 data illustrating effect 
of various C values 




In sinh C 





























Stress, psi 


77 



20 22 24 26 28 30 32 34 36 38 

P =T (log t R +C) x 10- 3 



_19 -18 -17 -16 -15 -14 -13 -12 -11 -10 -9 

log t R _ b/RT 


<N 

E 

E 


t/i 

Ui 

O 

l/) 


10 5 



-150-140 -130 -120-110 -100 -90 
(T + 816.2)/log t R - 20.4) 


Fig. 1.60 — Rupture data for Timken 25-20 stainless steel analyzed 
in terms of various parameters 






78 


ard deviation, p (based on log time) is used as a guide. For the Larson-Miller approach, 
the optimized C value of 14.03 yields a much better correlation than that observed in the 
Clauss analysis. 

f lT-590 Alfojff / 

Some indication of the relative extrapolative effectiveness of the Larson-Miller, Dorn, and 
Manson-Haferd parameters was provided in an analysis of the rupture data for S-590. 51 Em¬ 
ploying optimization procedures, and assuming a second-degree polynomial relating the pa¬ 
rameter to log stress, master-rupture plots were prepared for the three parameters. In 
one case, the analysis was based on data points below 10, 000 hours while in a second eval¬ 
uation only data points below 2000 hours were included. 

Using the master rupture plots, the predicted rupture times corresponding to conditions 
beyond the range of data used in making the analysis were compared to actual experimental 
values. This comparison is presented in Table 1.12 where the standard deviations (based on 
log time) for each set of results indicate the Dorn parameter to be the most effective in the 
case of the 10,000-hour data with the Manson-Haferd slightly less effective. Rather poor 
results are seen for the Larson-Miller parameter. In the 2000-hour analysis, approximate¬ 
ly equal effectiveness is seen for the Dorn and Manson-Haferd parameters with the Larson- 
Miller parameter again somewhat less effective. While these results identify certain param¬ 
eters as being more effective than others, it cannot be concluded that this is a general re¬ 
sult. For other data, the Larson-Miller parameter may give the best extrapolations. 


TABLE 1.12 


COMPARISON OF PREDICTED AND EXPERIMENTAL RUPTURE TIMES 
FOR S-590 ALLOY FOR LARSON-MILLER, DORN, AND 
MANSON-HAFERD PARAMETERS 


Stress 

Temperature, 

°C 

Experimental 

Rupture 

Time, hr 

Predicted Rupture Times, hr 

kg/mm^ 

psi 

L-M 

Dorn 

M-H 



Based on S-590 data, tp < 10,000 hours 



17.6 

25,000 

1200 

43,978 

62,248 

42,958 

41,698 

21.1 

30,000 

1200 

11,937 

11,999 

9,977 

10,685 

12.3 

17,500 

1350 

16,964 

19,508 

13,493 

16,378 

7.0 

10,000 

1500 

15,335 

34,465 

17,436 

21,870 

7.0 

10,000 

1500 

11,257 

34,465 

17,436 

21 >870 





0.278 

0.105 

0.151 



Based on S-590 data, tp < 2,000 hours 



35.2 

50,000 

1100 

3,149 

1,661 

1,614 

1,467 

24.6 

35,000 

1200 

2,243 

2,582 

2,206 

2,304 

14.1 

20,000 

1350 

9,529 

5,921 

4,234 

5,143 

8.8 

12,500 

1500 

5,052 

6,428 

4,076 

5,084 

17.6 

25,000 

1200 

43,978 

50,747 

29,856 

29,860 

21.1 

30,000 

1200 

11,937 

10,069 

7,518 

7,715 

12.3 

17,500 

1350 

16,964 

17,761 

10,332 

13,067 

7.0 

10,000 

1500 

15,335 

35,918 

13,873 

19,445 

7.0 

10,000 

1500 

11,257 

35,918 

13,873 

19,445 





0.244 

0.195 

0.190 


51 R- m - Goldhoff, "Comparison of Parameter Methods for Extrapolating High Temperature Data " ASME Trans J of 
Vol. 81,1959, p. 629. 


Basic Engrg., 





79 



C OMPARISON OF^TEST DA TA 


~=y> <P 


‘El 


During the last few years, stress-rupture and creep data have been obtained for a number 
of refractory metals and alloys from 1600° to 2800°C. Stresses required to cause rupture in 
1, 10, 100, and 1000 hours at 1600°C and 2200°C are given in Table 1.13 for a number of ma¬ 
terials. Items of note are (1) at 1600°C arc-cast W — 25Re has the highest 100-hour rupture 
stress, whereas at 2200°C powder-metallurgy rhenium has the highest value followed by 
tungsten; (2) available data have indicated different behavior for arc-cast and powder- 
metallurgy forms of the same material. Figures 1. 61 and 1. 62 give isochronal stress- 
rupture data as a function of temperature for 1 and 100 hours. The change in rupture life 
with respect to temperature and time is readily seen. 


TABLE 1.13 

STRESS TO CAUSE RUPTURE AT TIMES INDICATED FOR VARIOUS 
REFRACTORY METALS AND ALLOYS 


Material 3 '* 3 


1600 

°C 



2200°C 


Stress,* 3 kg/mm^, at hours indicated 

1 10 100 1000 

Stress,* 3 

1 

kg/mm^, at hours indicated 
10 100 1000 

Tungsten 









AC 

7.7* 

5.2 

3.6 

2.5 

2.0* 

1.3 

0.88 

0.49 

PM 

0.84 

5.5 

3.5 


2.3 

1.6 

1.1 


Molybdenum 









AC 

2.5 

1.5 

0.98 

0.62 

0.46 

0.28 

0.17 

0.098* 

PM 



1.3 


0.70 

0.59 

0.38* 


Rhenjum 









PM No. 1 


5.6 

3.1 


3.5 

2.2 

1.2 


PM No. 2 


7.4 

4.8 

3.2* 

4.2* 

2.8 

1.7 


Niobium 









AC 

0.63 

0.39 







W—25Re (wt%) 









AC 

15.* 

8.8 

4.9 

2.8 

2.5 

1.3 

0.70 

0.38 

PM 

15.* 

7.9 

4.1 


1.9 

1.1 

0.44* 


Mo—30W ( wt %) 









AC 

5.2* 

3.0 

1.8 

1.0 

1.0* 

0.57 

0.32 

0.17 

Mo—50Re (wt %) 









PM No. 1 

4.3 

2.7 

1.7 

1.1* 

0.81 

0.35 

0.15 

0.063* 

PM No. 2 

4.3 

2.4 

1.3 

0.75* 

0.81 

0.35 

0.15 

0.063* 

W—30Re—30Mo (at. %) 









AC 


5.9* 

2.8* 



0.70* 

0.42 

0.25* 



4.2 

2.2 



0.74* 

0.32 


Mo — 5W (at. %) 









PM 

4.7* 

2.5 

1.3 


0.75* 

0.48 

0.31 


Mo —5Re (at. %) 









PM 

4.7* 

2.5 

1.3 


0.69* 

0.46 

0.31 


Re-IOOs (at. %) 









PM 


6.7 

4.5 



2.7 

1.1 


Ta — 10W (wt %) 









AC 


7.7 

4.9* 


2.1* 

1.2* 



Mo-TZM 









AC 





0.48 

0.28 



Tantalum 









AC 

1.4 




0.34* 




PM 









Cb-753 c 









AC 



0.84 







a Values with asterisk are based on extrapolations. 

^Wrought sheet material, AC = arc-cast, PM = powder metallurgy. 
c Nb —5V —1.25Zr (wt%). 





Stress, psi 


80 









PM — wrought powder-metallurgy 
AC — wrought arc-cast 


100 I-1- 

1200 1400 


1600 1800 2000 
Temperature, °C 


Fig. 1.61 — 1-hour isochronal stress-rupture data for various refractory 
metals and alloys 




81 



Fig. 1.62 — 100-hour isochronal stress-rupture data for various refractory 
metals and alloys 

1.2 THERMAL PROPERTY EVALUATIONS 

J ELECTRICAL RESISTIVITY AND THERMAL CONDUCTIVITY 

Measurements of the electrical resistivity of W - 26Re (wt %) and unalloyed tungsten were 
made from room temperature to 1600 °C in argon plus hydro gen. j These data 52 are presented 
in Figure 1.63 along with some representative literature values/ 3 " 57 A least squares analy¬ 
sis of these data yielded the following relationships: 

W — 26Re 

p = 1.86430 x 10 -5 + 3.37090 x 10' 8 T - 1. 29457 x 10" 12 T 2 (1.17) 

p= 2.43036 x 10-7 0 hm-cm 

Tungsten 

p = 4. 33471 x 10" 12 T 2 + 2.19691 x 10 -8 T - 1. 64011 x 10 -6 (1.18) 

p = 3. 03318 x 10" 7 ohm-cm 300°K ^ T < 1240°K 

52 GEMP-1002, pp. 26-31. 

L. McElroy, "High Temperature Materials Program Quarterly Report for Period Ending April 30, 1966," ORNL-TM 1520, 

Part 1, pp. 53-58. 

54 J. P. Moore, R. S. Graves, W. Fulkerson, and D. L. McElroy, 'The Physical Properties of Tungsten/' 1965 Conference on 
Thermal Conductivity, Denver, Colorado, October 1965, by permission. 

55 R. P. Tye, "Preliminary Measurements on the Thermal and Electrical Conductivities of Mo, Nb, Ta, and W," from Niobium, 
Tantalum, Molybdenum, and Tungsten, A. G. Quarell, Editor, Elsevier Publishing Co., Amsterdam, 1961. 

56 A. G. Worthing and E. M. Watson, "Resistance and Radiation of Tungsten as a Function of Temperature," J. Opt. Soc. A., 

Vol.24, 1934, p. 114. 

57 V. S. Gumenyuk and V. V. Lebedev, "Investigating Heat and Electro-Conduction of Tungsten and Graphite at High Tem¬ 
peratures," Fizika Metallov and Metalovedeniye (2 Jan. 1961), Nr. 1, pp. 29—33, Translation USAEC-NP-TR733. 












82 


and 4,tT)o°l3 

p =-4.06012 x 10- 12 T 2 + li - 0 3 03 -x 10 _8 T -1.97071 x 10" 5 (1.19) 

H = 1. 56328 x 10" 6 ohm-cm 1240°K < T < 2570°K 

where: 


p = electrical resistivity, ohm-cm 
T = temperature, °K 
fi = standard deviation 

Previously reported data 52 for the thermal conductivity of W - 26Re (density close to 98 
percent of theoretical) are summarized in Figure 1. 64. Using the resistivity data reported 
above, the thermal conductivity data were analyzed to yield: 

k = 2.443 x 10“ 8 —+---(1.20) 

P 0.7493 + 7. 3427 x 10 _3 T 

(standard deviation = 4.5818 x 10“2 watt/cm °C) 

where: 

k = thermal conductivity, watt/cm°C 
T = temperature, °K 
p = electrical resistivity, ohm-cm 

Also shown in Figure 1. 64 are data for W — 25Re reported by Jun and Hoch 58 and other 
data estimated by McElroy. 53 Fairly good agreement is noted between these data and those 
obtained in the NMPO studies. 

THERMAL ,DIFFUSIVITY y 

A pulse-type diffusivity technique 59 based on the use of a laser was employed to measure 
the thermal diffusivity of commercial, powder-metallurgy W - 25Re in the temperature 
range from 300° to 1000°C. Specimens were in the form of discs 0.63 cm in diameter by 
0.18 cm in thickness. A platinum-wound muffle furnace was used to provide the desired 
test temperatures and an oscilloscope was employed to determine the temperature - time 
transient of the specimen following the laser pulse. The time required for the back sur¬ 
face of the specimen to reach a temperature corresponding to one-half the maximum rise 
was then used in the usual manner to calculate diffusivity values. Data obtained for the 
W - 25Re alloy are presented in Figure 1. 65. Also shown in this figure are diffusivity 
values calculated from the thermal conductivity data for the W - 26Re composition shown 
in Figure 1. 64. As shown the agreement is quite excellent. 

Diffusivity data were also obtained for arc-cast and powder-metallurgy tungsten using 
specimens 0.152 and 0.203 cm in thickness and 0.63 cm in diameter. All diffusivity values 
were in good agreement indicating no effect due to specimen thickness or material form. 
Average diffusivity values obtained in these tests are presented in Figure 1. 66; all ex¬ 
perimental data were within ±2 percent of this average curve. Also shown in Figure 1. 66 
are diffusivity data for tungsten calculated from some recent thermal conductivity data 60 
reported for this material. Except at 300°C where a difference of slightly greater than 10 

C. K. Jun and M. Hoch, "Thermal Conductivity of Tantalum, Tungsten, Rhenium, Ta - 10W, T-222, W - 25Re in the 
Temperature Range 1500° - 2700°K," Third International Symposium on High-Temperature Technology, Stanford 
Research Institute, September 17-20, 1967. To be published. 

59 R. j * Freeman, "Thermal Diffusivity Measurements on Pre- and Post*Irradiated BeO," GE-NMPO, GEMP-452, 

November 15, 1966. 

60 

Moore, Graves, Fulkerson, and McElroy, loc. cit. 


Thermal conductivity, watt/cm-°C 


83 



Fig. 1.63 — Electrical resistivity versus temperature for W — 26Re alloy 
and unalloyed W (AS-745B) 




, i 

& C 

Lit 

T 



Hb 





1^64; 




3-— drj- 

- W-- 



— 





5 

u 

u 



r 







-- 

5 This 

-Jun c 

-Jun c 

itudy 
md Hoch 
and Hoc! 

i-powdei 
i-arc cas 

' metallu 
t 

rgy 







V McElroy new estimate 

A McElroy old estimate 

_l_ I _l_ 

1 _ 


400 600 800 1000 1200 1400 1600 1800 2000 2200 2400 

Temperature, °C 


Fig. 1.64 — Thermal conductivity versus temperature for W — 26Re alloy (AS-745D) 














84 



Fig. 1.65 — Thermal diffusivity versus temperature for W - 25Re (wt %) from 
300° to 1000°C in argon atmosphere 



300 400 . 500 600 700 800 900 1000 

Temperature, °C 

Fig. 1.66 — Average thermal diffusivity of tungsten from 
300° to 1000°C in argon atmosphere (speci¬ 
men thickness, 0.152 and 0.203 cm) 


percent is indicated, agreement between measured and calculated diffusivity values is 
within 5 percent. This is excellent agreement considering the thermal conductivity data 
were obtained using a radial-flow steady-state technique, whereas the measured diffusivi- 
ties were obtained by a transient procedure. 


n. 


Diffusivity data were also obtained for 304L stainless steel in the temperature range 
from 300° to 1100°C. Testing was performed in argon using specimens 0. 63 cm in diam¬ 
eter and 0.127 and 0.152 cm in thickness. The chemical composition of this material was: 
18.37% Cr, 9.89% Ni, 0.024%C, 1.31%Mn, 0.70% Si, 0.014%P, 0.012%S, 0.09% Mo, 
and 0.05% Cu. A plot of the diffusivity data for this material is presented in Figure 1. 67. 
It is important to note that for both specimen thicknesses the diffusivity increased slightly 
after the first heating to reveal an annealing effect. 


u 


ENTHALPY 
-- 

Modifications were made to the drop calorimeter to permit enthalpy measurements to 
be made to 3000 C. Enthalpies of commercially pure (99.94%), powder-metallurgy tung- 





85 


sten were measured 61 to 3250°K and are summarized in Figure 1. 68. A least squares 
analysis of these data yielded: 

h t - H 2 98°k = “ 1579. 3 + 5. 406T + 5. 047 x 10" 4 T 2 + 271485X (1.21) 

where: 

X = exp. (-35629. 5/RT) 

Hrp = enthalpy in cal/mole 
R = 1.9869 cal/mole - °K 
T = °K 

applicable temperature range 1200° to 3250°K. 

Differentiating the heat content equation yields the heat capacity equation: 

Cp = 5.406 + 10.094 x 10“ 4 T + 96728 x 10 5 X/RT 2 (1.22) 

where: 


C n = cal/mole - °K 

A comparison of the heat capacities obtained in this study with those reported in the most 
recent literature 62 " 65 is given in Figure 1. 69. Between 1200° and 2400°K agreement within 
4 percent is shown for most of the data; however, this study is about 8 percent higher than 
that reported by Kirillin, et al. 63 at 3100°K. A slightly different equation was applied to 
Kirillin’s data by West 65 to yield a slightly better representation; however, it only affected 
the Cp by about 2 percent (higher) at 3100°K. 

tTg SUMMARY AND CONCLUSIONS 

Stress-rupture and creep data for wrought, arc-cast tungsten are presented for tempera¬ 
tures from 1600° to 3000°C and rupture times approaching 4000 hours. Good correlation is 
shown when analyzed in terms of the diffusion-compensated creep rate versus the modulus - 
compensated stress. The rupture ductility for arc-cast tungsten is a maximum (~20%) in 
the 2200° to 2400°C temperature range, whereas powder-metallurgy material shows a de¬ 
crease in ductility (from 70%) with increasing temperature. Based on the counting of free 
dislocations for samples creep tested above one-half the absolute melting temperature, a 
power stress law is shown to apply indicating creep by dislocation climb or the glide of 
jogged screw dislocations. Optical microscopy and electron fractography studies of creep- 
rupture tested (1600° to 2800°C) powder-metallurgy and arc-cast tungsten indicate differ¬ 
ent mechanisms of failure operating in the two types of materials. The powder-metallurgy 
and arc-cast tungsten indicate different mechanisms of failure operating in the two types 
of materials. The powder-metallurgy material showed pronounced grain boundary separa¬ 
tion and cavitation, whereas the arc-cast material showed no tendency to form cavities. 


Stress-rupture and creep data are presented for wrought, arc-cast, unalloyed molyb¬ 
denum sheet for temperatures from 1200° to 2400°C. Good correlation between the time 


to rupture and linear creep rate is shown except for the extremely high temperatures 


61 GEMP-1002, p. 34. 

Hoch and H. L. Johnston, "A High Temperature Drop Calorimeter. The Heat Capacities of Tantalum and Tungsten 
Between 1000° and 3000°K/' J. Phys. Chem., Vol. 65, 1961, pp. 855-860. 

A. Kirillin, et al., '"Thermodynamic Properties of Tungsten at 0° — 3500°K," Zhurnal Fizi Khimii, Vol. 37, No. 10, 

1963, pp. 2249-2257. 

C. Lowenthal, "The Specific Heats of Metals Between 1200°K and 2400°K," Aust. J. Phys., Vol. 16, No. 1,1963, pp.47—67. 
65 E. D. West and S. Ishihara, National Bureau of Standards, Washington, D.C., Preliminary Data of the Enthalpy of Tungsten — 
private communication. 


86 



Fig. 1.67 — Thermal diffusivity of 304L stainless steel from 300° to 1100°C 
in argon atmosphere 



Fig. 1.68 — Enthalpy versus temperature for tungsten (99.94% purity) 

and low stresses. The deviation appears to be related to a change in the creep mechanism 
indicated by a charge in the form of the creep curve. 


Creep-rupture data are presented for Re, W - 30Re - 30Mo (at. %), Mo - 30W (wt %), 
and W — 25Re (wt %) sheet tested in hydrogen at temperatures from 1600° to 2600°C. 

Creep-rupture properties for powder-metallurgy Mo - 50Re (wt %) are presented for 
temperatures of 1600°, 2200°, and 2400°C. Diffusional creep is indicated at the higher 
temperatures and low stresses based on the creep rate versus stress relationship yield¬ 
ing a slope of unity. The diffusion coefficient for Mo - 50Re at 2400°C and 0.07 kg/mm 2 
was calculated to be 1.14 x 10 - ? cm 2 /sec. , 




87 




1000 1200 1400 1600 1800 2000 2200 2400 2600 2800 3000 3200 3400 

Temperature, °K 


Fig. 1.69 — Heat capacity versus temperature for tungsten 

Creep-rupture tests of Mo - 30W at 2200°C under constant load display a creep curve 
consisting only of third-stage-type creep with the elongation at rupture exceeding 100 
percent. Identical tests performed at constant stress display a linear (secondary) creep 
rate to a strain of 65 percent. 

Creep-rupture tests of powder-metallurgy tungsten at 2800°C and 1.05 kg/mm^, in 
single-crystal form, displayed no detectable creep with fracture being brittle in nature. 
This result appears to be consistent with creep deformation observations for most powder- 
metallurgy, polycrystalline materials tested at high temperatures in that cavitation and/or 
grain boundary separation are the predominant mechanisms. 

A comprehensive study of stress-rupture parameters was completed with special con¬ 
sideration given to the mathematical procedures employed in identifying parameter con¬ 
stants. Detailed comparisons were made of the relative effectiveness of these parameters 
in the correlation and extrapolation of experimental data. 

Based on experimental measurements of thermal conductivity and electrical resistivity 
for W - 26Re (wt %), an expression describing the thermal conductivity from 400° to 2400°C 
is presented. 

Enthalpies of unalloyed tungsten were measured from 1200° to 3450°K. Equations for A 
heat content and heat capacity were determined from the data using a least squares J 
analysis. j 

1. 4 PLANS AND RECOMMENDATIONS 


Stress-rupture and creep studies will be continued for commercially available and de¬ 
velopmental refractory metal alloys as well as materials having potential application to 





88 


the LMFBR program. Evaluations will be performed in inert, reducing, and vacuum 
environments. 

In an attempt to identify the mechanisms associated with creep resistance and deforma¬ 
tion, optical and transmission electron microscopy studies of tested samples will be per¬ 
formed. 

Creep tests at constant stress will be performed to afford a comparison with data ob¬ 
tained under constant load. The results will assist in determining the stress dependence 
of more complex materials. 

Some physical properties will be determined for materials of interest. Properties such 
as enthalpy, thermal expansion, and thermal conductivity will be determined to tempera¬ 
tures approaching the melting point for design purposes as well as hazards evaluation. 
Some thermal diffusivity measurements to 1000°C are also planned. 



I s*—- 

* o* k3 


& 


2.\ radiation effe his on fast reactor cladding 

AND STRUCTURAL MATERIALS_\ 

( 1304 ) 

J. Moteff,* F. D. Kingsbury.t J. P. Smith t 


The objective of this program is to determine the effect of radiation on the time-, tem¬ 
perature-, and stress-dependent properties of selected heat-resistant alloys and refrac¬ 
tory metals, to identify the causes of any changes observed in these properties, and to 
develop remedial measures. 

The experimental program to study effects of neutron irradiation on the creep-rupture, 
tensile, hardness, and resistivity properties of the above classes of metals and alloys is 
continuing. Direct observations of defect configurations by means of the transmission 
electron microscope, and analysis of the neutron-induced defects and their migration 
behavior, are being performed and correlated with the measured changes in the mechan¬ 
ical properties. Detailed spectra studies are being performed to determine an accurate 
profile of flux and energy distribution of reactor positions used in the program. 

Materials investigated in the heat-resistant alloys portion of the program include A-286, 
Hastelloy X, Hastelloy N (INOR-8), Hastelloy R-235, Fe-Cr-Al-Y alloys, AISI 304, 316, 
and 348 stainless steels, ASTM-A302B and A350-LF3 pressure vessel steels, and various 
Inconel and Incoloy alloys. 

The refractory metals portion of the program includes investigations of V, Mo, Nb, W, 
Ta, and their alloys. 

2.1 STATUS OF IRRADIATIONS (F. D. Kingsbury, J. P. Smith, W. S. Chenault) 

EBR-n IRRADIATION PROGRAM 

A planned program of fast neutron spectrum irradiations in the EBR-II reactor was 
initiated in conjunction with ORNL. This program is a reorientation of a previous pro¬ 
gram for a series of irradiations in the Fermi reactor. Two phases of the present pro¬ 
gram were implemented. 

Phase I consisted of irradiating Fe, Ti, Ni, and Co dosimeters and 16 refractory metal 
tensile/creep-rupture specimens. The irradiation of 148 dosimeters resulted in complete 
flux mapping of the EBR-II rows 2 and 7 positions with the new 91-element reactor core 
configuration. Irradiation was conducted at a maximum reactor power level of 30 Mw to 
a total integrated exposure of 688 Mw hours. The neutron flux density profile was deter¬ 
mined in order to interpret subsequent EBR-II material irradiations and to correlate ir¬ 
radiation test results from other reactors. The radiochemical analysis of the dosimeters 
are discussed in the dosimetry section. 


*Project leader. 
^Principal investigator. 


89 






90 


Phase HA of the program provides for the irradiation of resistivity, hot-hardness, and 
tensile/creep-rupture specimens at various elevated temperature and fluence values. 
Individual NMPO capsules, loaded into the tube bundles by ORNL personnel and awaiting 
insertion into the EBR-II, include 112 tensile/creep-rupture specimens, 64 hot-hardness 
specimens, and 264 resistivity (wire) specimens. Materials to be irradiated are the 
heat-resistant alloys Hastelloy X, Hastelloy R (three different B isotope dopings), 
Fe-Cr-Al-Y, and Mo, Cb, W, Ta, and their alloys. Additional materials including 
stainless steels and vanadium-base alloys are planned for Phase HB. 

ORR AND ETR IRRADIATIONS 

Four refractory metal specimen capsules were irradiated in the ORR and four heat- 
resistant alloy capsules were irradiated, three in the ETR, one in the ORR. 

2. 2 HEAT-RESISTANT ALLOY PROGRAM 

The objectives of this phase are to determine engineering and design data of heat- 
resistant alloys and to interpret these data (with supporting transmission electron micro¬ 
scopy) to determine the mechanisms of radiation damage. Materials tested include Incoloy 
800, Hastelloy X, A-286, Hastelloy R-235, and Fe - 15Cr -4A1 - 1Y (1541 alloy). 

CREEP-RUPTURE TESTING (R. A. Joseph, J. P. Smith, J. Moteff, J. A. Edwards) 

All creep-rupture testing except for the 1541 alloy was performed in standard lever- 
loaded (5:1 ratio) dead-weight equipment. Essentially all creep data to be discussed 
were measured from total load train movement. To determine the order of magnitude of 
error inherent in this technique, several specimens were instrumented with platinum 
strip-type extensometers attached directly to the gage length of the specimens. Com¬ 
parison of strain as determined by the load-train LVDT sensor against the optical showed 
that for strains above 2 percent the relative error was less than 10 percent. For strains 
less than 1 percent, the relative error was somewhat higher (up to 40%). These tests 
verified that the total load-train monitoring system utilized was satisfactory for record¬ 
ing strains above 1 to 2 percent, generally well below the range at which second-stage or 
minimum creep occurs with the possible exception of extremely brittle materials. 

Incoloy 800 

Specimens of Incoloy 800, irradiated at about 540°C in EBR-H to fast fluences of 0.27 
and 2.7 x 10^0 n/cm^ (E n ^ 1 Mev), were creep-rupture tested at 540°C and 705°C. Data 
are summarized in the graphs of Figures 2.1 and 2. 2. Radiation affects the creep prop¬ 
erties somewhat differently at the two temperatures. At 540°C the minimum creep rate 
(e) is increased and time to rupture (tp) is reduced compared to the unirradiated material; 
at 705°C, e is reduced and tp increased (except at high stresses) compared to the control 
specimens. At both test temperatures radiation reduced the rupture elongation by about a 
factor of 3. Another significant difference is that for 540°C tests the irradiated and con¬ 
trol lines tend to converge at low stresses but more tests are needed to verify this; on 
the other hand, the tendency for convergence of the unirradiated and irradiated control 
lines at 705°C is at high stresses. This convergence at high stress is unique among alloys 
investigated at NMPO; generally any convergence occurs at low-stress (longtime) tests, 
as appears to be the case at 540°C (Figure 2.1), indicating annealing of point defects. To 
determine whether the effect observed at 705°C was due to thermal aging or to combined 
stress and temperature, an irradiated and an unirradiated control specimen were aged 
at 705°C for 500 hours and then tested at 10.5 kg/mm^. The irradiated specimen rup¬ 
tured after 24.7 hours and had a minimum creep rate of 2. 4 x 10”® sec"*, and a fracture 
elongation of 37 percent; these properties are essentially identical to those of the as- 
irradiated tests shown in Figure 2.2. This indicates that the divergence of the control 




Applied stress, kg/mm 2 (log scale) Applied stress, kg/mm 2 (log scale) 



















93 


and irradiated 705°C properties (Figure 2.2) at low stress is not due solely to thermally 
activated migration or interaction of irradiation-induced defects; more tests are needed 
to determine the cause of the increased post-irradiation strength. 

The creep curves for the 540°C and 705°C tests (Figure 2.3) indicate reduction in post¬ 
irradiation ductility. The major effect of radiation on ductility is the reduced third-stage 
creep strain. 

The tests described above were of the creep-rupture variety performed at a fixed load 
to failure. Actual reactor operations do not result in this condition; instead, the stress 
varies over a rather wide range. To investigate the effects of varying stress and to deter¬ 
mine accurately how radiation affects the low strain part of the creep curve, tests were 
performed in which the stress on a given specimen was varied and the strain was meas¬ 
ured accurately with a cathetometer sighting directly on the gage length of the specimen. 
The low-strain range is of particular interest to designers since many cladding designs 
allow for 3 percent maximum strain. The procedure was to load a specimen to <Xj and 
establish the linear creep rate; the load was then increased to and again a linear creep 
rate was established; this was repeated for three stress levels. Finally the load was re- 




Fig. 2.3 — Creep curves for irradiated and unirradiated Incoloy 800 at 540°C and 705°C 





94 


duced to <Tj to compare with the original creep rate at aj and to determine whether the 
substructure had been altered. Creep data were obtained for irradiated and control speci¬ 
mens and are shown in Figures 2.4 and 2.5. For the unirradiated specimen (Figure 2.4) 
the minimum creep rate is apparently a function of prior stress or strain history, based 
on the fact that second-stage creep rate (e g ) for final is significantly higher than for 
aj initial. In the case of the irradiated specimen (Figure 2.5) the e g is essentially identi¬ 
cal for Oj initial and final. In the unirradiated results, for Oj initial, two separate linear 
portions e g appear to be present. The exact cause of this is not known; it has been ob¬ 
served on post-test metallography of other ruptured specimens that recrystallization is 
probably occurring during test since the grain size near the fracture is very fine and 
becomes progressively coarser in low-strain areas. This recrystallization could explain 
the two linear portions of the initial ctj and the decreased creep resistance of the final o^. 

The result on the irradiated specimen indicates that irradiation "freezes in the struc¬ 
ture" and that stresses or strains in the region investigated do not alter this structure. 



Fig. 2.4 — Instantaneous creep rate at 704°C of unirradiated Incoloy 800 as a 
function of creep time for various stress levels 


Metallography and transmission electron microscopy will be performed to help evaluate 
results. 

A similar series of tests was initiated at 540°C. Creep rates for the unirradiated speci¬ 
mens at the three lower stress levels of Figure 2.1a were all determined from one speci¬ 
men; they are in good agreement with tests performed at a constant load to failure. 

To more fully evaluate all available test data, correlations are being made between 
creep and tensile data. Tensile data discussed here were generated at GE-APO 1 on sam¬ 
ples.from the same heat and irradiation as for the specimens described above. Figure 2.6 
demonstrates a reasonably good correlation of the tensile and creep data both for strength 
and ductility. Of particular interest were (1) the effect of fluence on the ductility but not 
on the strength, and (2) the minimum in the ductility curve for the irradiated specimens. 






95 



o 100 200 300 


Time, hours 

Fig. 2.5 — Instantaneous creep rate at 704°C of irradiated (2.4 x 10^ n/cm^, 

E n > 1 Mev) Incoloy 800 as a function of creep time for various 
stress levels 

Stiegler and Weir 2 reported such a minimum for Hastelloy N. A unique approach to corre¬ 
lating the tensile and creep data is the application of the Monkman-Grant relationship 3 
(log € vs log t-^) as shown in Figure 2. 7. At both temperatures, data fall on lines close 
to the predicted slope of -1, and available tensile data fit the curve quite well. The effec¬ 
tive t^ for the tensile data was obtained by calculating e /£ for each test where e is the 
fracture strain. Refinements can probably be made, such as using uniform strain, al¬ 
though the above technique appears to work quite well to a first approximation. The slope 
is unity; hence e-tp = constant. Although the units of this constant are strain, its physical 
significance is not known. It may be used as a guide to predict the minimum fracture strain 
since what it represents is the strain at failure obtained if no third-stage creep occurred. 
Since most unirradiated materials do exhibit a third-stage creep, the fracture ductility is 
generally somewhat higher than predicted by the Monkman-Grant relationship. On the 
other hand, many irradiated materials exhibit little or no third-stage creep and this re¬ 
lationship may predict quite closely the fracture ductility. 

Hastelloy X* 

Concurrent to the work described above, creep-rupture testing of Hastelloy X irradi¬ 
ated in EBR-II was performed in the range of 540° to 705 °C. Most data were generated 
at 650 °C since the primary intent was to compare these data with previously determined 
data from ETR irradiations. The 540°C and 705 °C data are presented in Table 2.1 but 
will not be discussed because of the limited quantity. Stress-rupture life at 650 °C of the 

^Specimens fabricated and irradiated by GE-APO. 







96 



1 10 100 
True stress <ct), kg/mm 2 


Fig. 2.6 — Strain rate and elongation as a function of true stress for 
fncoloy 800 tested at 705°C. To minimize structure differences 
due to work-hardening the true flow stress at about 6% true 
strain was used to correlate the tensile data with the creep data. 

specimens from the EBR-II and ETR irradiations, as summarized in Figure 2.8, was 
essentially the same. Ductilities (not shown) from both irradiations were similar, but 
the EBR-II specimens showed slightly lower ductility due to the higher fluence. The 
point of convergence for control and irradiated lines is about 20 kg/mm^. Creep data 
were not obtained for the ETR irradiated specimens. The strain rate-stress relation¬ 
ship for the EBR-II specimens (Figure 2.9) shows the creep rate to be reduced by irra¬ 
diation; reduction is greater at lower stresses. The irradiated and control curves of 
Figure 2.9 converge at about 27 kg/mm^, a stress higher than for the tj^-a curves. Creep 
curves (not illustrated) for irradiated and control specimens at 650 °C show that the major 
reason for the loss in ductility is the reduction of third-stage creep; the same was true 
of Incoloy 800 (Figure 2.3). The Monkman-Grant 3 relationship was also applied to these 
data as shown in Figure 2.10. Although the lines are not so close to a slope of -1 as the 
Incoloy 800 data, the control and irradiated lines are parallel. Based on the criterion 
discussed above that e • tj^ = constant, the constants 0.06. Irradiated Hastelloy X ex¬ 
hibited virtually no third-stage creep and the fracture ductility was 8 to 9 percent which 
is not much above predicted minimum of 6 percent. 




97 



Fig. 2.7 — Monkman-Grant relationship for irradiated and unirradiated 
Incoloy 800 at 540°C and 705°C using tensile and creep data 


TABLE 2.1 

CREEP-RUPTURE DATA FOR IRRADIATED AND UNIRRADIATED HASTELLOY X 


EBR-H Irradiation Conditions 


Test Conditions 


Fluence, Temperature, 

n/cm 2 {E n > 1 Mev) °C 

Stress, 

kg/mm 2 

Temperature, 

°C 

Rupture Life, Minimum Creep 
hr Rate, sec - ^ 

Elongation, % 
in 2.54 cm 

Unirradiated 


39.4 

540 

On test 

1.1 x 10“ 8 

_ 

3.2 x 10 20 

540 

39.4 

540 

On test 



1.7 x 10 19 

540 

39.4 

540 

636 

1.7 x 10” 8 

21.7 

Unirradiated 


15.7 

705 

355 

3.0 x 10“ 7 

78 

Unirradiated 


17.1 

705 

229 

3.6 x 10~ 7 

58 

3.2 x 10 20 

540 

15.7 

705 

126 

7.7 x 10~ 8 

5 


Hastelloy R-235 

Creep-rupture measurements are continuing on the special split heat of Hastelloy R-235 
containing a total of 45 ppm boron with varying B* 0 concentration. Specimens were test¬ 
ed at 870°C following irradiation in the ORR at 70°C and in the ETR at 760°C. Results of 
the post-irradiation creep-rupture tests are shown in Figure 2„ 11. Only two specimens of 
each type were available from the ORR irradiation. Data in Figure 2.11 indicate that (1) 
the 760 °C irradiated specimens exhibit slightly poorer properties than the 70 °C specimens, 
and (2) the stress dependency is not significantly changed by irradiation. The amount of 
damage appears to be definitely a function of the concentration. One specimen of each 
heat from the ORR irradiation was re-solutioned and aged after irradiation and before 
testing; results are shown in Figure 2.11. As observed in other alloys, the rupture 
strength is further decreased by the re-heat treatment compared to the as-irradiated 




98 



H60 


30 


£ 

£ 20 




o 




a 

% 


ioi 


Pre-Test Treatment 
0 100 hours at 650°C 
□ Solution-treated 
O Solution-treated plus 
223 hours at 540°C 
^ ETR irradiated at 650°C 
4>p = 5 x 10^9 n/cm2 
4»j^ = 4 x 1019 n/cm2 

% EBR-II irradiated at 540°C 
$p = 3.2 x 1020 n/cm2 

_ _> Determined by least squares 

} fit through Q and ^ only. 

_I. .IN 


Specimen Type 
0.31-cm diameter 
0.25-cm x 0.076-cm sheet 
0.25-cm x 0.076-cm sheet 

0.31-cm diameter 


0.25-cm x 0.076-cm sheet 


Rupture life, hours 


Fig. 2.8 — Stress-rupture properties of irradiated and unirradiated Hastelloy X at 650°C 



Fig. 2.9 — Minimum creep rate as a function of applied stress for unirradiated 
and irradiated Hastelloy X at 650°C 

values. Analyses of these tests are continuing, to determine the relationship between 
creep rate, helium concentration, and defect density. 

A-286 

As discussed in a previous report, 4 Woodford 5 showed that the stress dependency, n, 
can be approximated from a single constant-load test for a fixed temperature. The ex¬ 
pression k ~ Aa n can be used in this analysis, with n, the stress dependency, equal to 
(d log e/d log a) and inversely proportional to the strain rate sensitivity m; i. e., m = 1/n = 


































(d loga/d loge ). Data presented previously 4 showed that three different tests of Hastelloy 
X at 650 °C and various initial stress levels yielded essentially the same value of n. This 
type of analysis was applied to irradiated and unirradiated test results of A-286 specimens. 

The results shown in Figure 2.12 indicate about a factor of 2 increase in the stress 
dependency. More significantly, the minimum creep rate decreases with increasing 
concentration of helium atoms as generated from the Bl0(n,cOLi7 reaction. In this 
figure the stress notation is omitted from the abscissa because the comparison is for creep 
tests performed at the same applied stress. Since there is essentially no primary creep, 
the intercept represents the minimum creep rate. 

The minimum creep rate is plotted as a function of the calculated helium atom concentra¬ 
tion in Figure 2.13; there appears to be a fairly good linear relationship. Minimum creep 
rate could be a function of the number of Bl0(n, a)Li^ reactions which occur and is probably 
a function of the helium atom concentration. Strain rate sensitivity (m) is independent of he¬ 
lium atom concentration in the range investigated as shown in Figure 2.12. The interaction 
of dislocations with helium bubbles may account for the increased creep resistance; this is 
consistent with transmission microscopy results reported previously. 4 The reduced ductility 
in irradiated A-286 could, however, be the result of decreased strain rate sensitivity (in¬ 
creased n or decreased m) which apparently is not a sensitive function of total helium atom 
concentration. The ductility change may also be a function of helium atom distribution; i. e., 
bubbles on those dislocations within grain boundaries may play a special role. Studies to 
correlate influence of changes in stress dependency and strain rate sensitivity on creep duc¬ 
tility and creep rates are continuing. 

Fe - 15Cr - 4A1 - 1Y 

Data for Fe - 15Cr - 4A1 - 1Y (1541 alloy) tensile and creep tests by ORNL, 6 Harwell, 7 
and NMPO were compiled. Results of unirradiated tests are presented in Figure 2.14 as 
the temperature-compensated creep rate (Zenner-Holloman parameter) versus stress. The 
ultimate strength was used for the tensile data, although the yield strength could have been 
used since they are virtually the same for this alloy. The applied stress was used for creep 
data. Considering the variation in test technique among the different laboratories and the 
inclusion of both tensile and creep data, a good least-squares fit to the experimental data 
was obtained with relatively little scatter. Figure 2.15 shows the similar parameter plots 
for the irradiated specimen data. This plot shows that irradiation reduces creep rate slight¬ 
ly but consistently for all test conditions considered, and that the stress dependency (n) for 
this irradiated material is not significantly changed. This is in contrast to the results dis¬ 
cussed above for Incoloy 800, Hastelloy X, and A-286. The Woodford technique 5 of calculat¬ 
ing stress dependency from a single constant-load test was applied to the NMPO creep data, 
and the analysis predicted that the stress dependency (n) was essentially unchanged by the 
irradiation. 

HOT HARDNESS (J. L. Kamphouse) 

Incoloy 800 

Hot hardness tests were performed on control and irradiated Incoloy 800 from the same 
material used to fabricate the control and irradiated spcimens used for the creep and tensile 
results discussed above. A summary of the hot-hardness data is shown in Figure 2.16. There 
is very little difference in hardness between the irradiated and unirradiated material at any 
given temperature, especially from 500 °C and below. At 600 °C and above, the irradiated 
specimen appears to be slightly harder than the unirradiated specimen, possibly because the 
irradiation temperature was approximately 540 °C; hence significant changes in strength would 
not be expected below this temperature. Of significance is the fact that tensile data 1 on these 















Stress, kg/mm2 










103 



Temperature, °C 

Fig. 2.16 — Hot microhardness of unirradiated and irradiated Incoloy 800 

same specimens in the 600° to 700°C range show only a small strength reduction (~ 5 to 10%) 
due to irradiation; hence these two types of tests are in qualitative agreement. 

The discontinuity at 200°C may be real because the tensile data 1 cited above show that the 
line through the data from the irradiated specimens crosses the line through the control data 
at about 200°C. That is, below 200° C the strength of the irradiated specimens is slightly 
higher than the control specimens, whereas above 200° C the strength of the irradiated 
specimens is generally lower than the control specimens. 

ASTM-A302B 

One unirradiated ASTM-A302B pressure vessel steel specimen and three irradiated in 
the IRL were hot-hardness tested from room temperature to 1000°C. During irradiation 
the three irradiated specimens were located at different distances from the core within a 
pressure vessel mock-up . 8 The spectrum changed somewhat from one side of the mockup 
to the other and the fluence decreased from the side nearest the core to the side farthest 
away. The three specimens showed essentially the same hot hardness regardless of posi¬ 
tion in the mockup and in fact showed the same hot hardness as the unirradiated specimen. 

RESISTIVITY STUDIES (L. K. Keys, J. Moteff) 

ASTM-A302B 

Preliminary studies presented on the A302B steels 8 demonstrated a good correlation 
between mechanical properties and resistivity, although these studies were performed over 
a limited fluence range (1 - 4 x 10*8 n/cm^). A second series of irradiation studies were 
carried out in the ORR at reactor ambient temperature in which a fluence range of approxi¬ 
mately 1 x 10^ to about 7:1 x 10 ^ n/cm^ was obtained, Table 2.2. In Figures 2.17 and 
2.18 the fractional resistivity recovery, Ap/p 0 , versus annealing temperature is given for 
the ORR and IRL irradiations. The general form of the recovery is the same. There are 





Fractional decrease in resistivity /-2. j at-196°C 


104 


TABLE 2.2 


IRRADIATION CONDITIONS AND PRINCIPLE RECOVERY PARAMETERS FOR A302B STEELS 


Sample 

No. 

Capsule 

Irradiation Conditions 
Fluence, nvt 

Fast (E R > 1 Mev) Thermal 

A p/p 

(Ap/p 0 ) min 

A(Ap/p 0 ) = 
l(Ap/p Q ) min—Ap/p 0 

2-3 

NRL-1-5 3 

4.8 x 10 18 

6.9 x 10 18 

-0.0223 

-0.0258 

0.0035 

2-5 

NRL-1-3 

2.7 x 10 18 

3.2 x 10 18 

-0.0142 

-0.0229 

0.0087 

2-7 

NRL-1-1 

1.1 x 10 18 

2.2 x 10 18 

-0.0004 

-0.0098 

0.0094 

2-8 

ORM-49-1 b 

1.0 x 10 18 

9.0 x 10 18 

-0.0093 

-0.0160 

0.0067 

2-9 

ORM-49-2 

1.7 x 10 18 

1.7 x 10 19 

-0.0118 

-0.0170 

0.0052 

2-10 

ORM-49-3 

6.0 x 10 18 

3.9 x 10 19 

-0.0098 

-0.0280 

0.0182 

2-11 

ORM-49-4 

1.1 x 10 19 

7.8 x 10 19 

-0.0064 

-0.0278 

0.0214 

2-12 

ORM-49-6 

7.1 x 10 19 

4.6 x 10 20 

-0.0427 

-0.0567 

0.0140 

2-13 

ORM-49-5 

4.5 x 10 19 

3.8 x 10 20 

-0.0365 

-0.0497 

0.0132 

-i_ 


industrial Reactor Laboratory (Plainsboro, New Jersey). 
b Oak Ridge Reactor (ORNL). 



Annealing temperature, °C 

Fig. 2.17 - Isochronal fractional resistivity recovery of irradiated A302B steel as a function 
of annealing temperature 



105 



Fig. 2.18 — Isochronal fractional resistivity recovery of IRL neutron irradiated A302B steel 
as a function of annealing temperature 

three important regions of resistivity change: (1) an initial negative resistivity change as a 
result of the reactor irradiation, (2) a further decrease occurring upon annealing to about 
250 °C, and (3) a region of increasing resistivity and recovery up to about 500°C above 
which recovery is complete and the resistivity behavior resembles that of the control speci¬ 
mens. The recovery minimum at about 250°C represents the maximum negative resistivity 
change due to the irradiation of these specimens. This minimum resistivity dependence 
(Ap/Pq) on the fast neutron fluence is shown in Figure 2.19. The value (Ap/po) minimum is 
defined as the largest negative Ap/p 0 value observed, i. e., Ap/p 0 at about 250°C. These 
values show a relatively smooth linear dependence on the fluence except for the lowest 
fluence IRL value. 

Based on the work of others, 9 ’ 10 the decrease in resistivity presumably results from the 
precipitation of carbon from solution by the formation of a carbon-defect complex. With in¬ 
creasing temperature, the carbon returns to solution and the vacancies recover and are an¬ 
nihilated at various sinks including carbide precipitate particles, causing the resistivity to 
return to its original value. Although carbon has been considered the only important inter¬ 
stitial impurity (because of its high concentration), other interstitials such as nitrogen may 
have similar effects. 

In Figure 2.20 the dependence of the radiation-induced resistivity incrementAp/p 0 on 
the fast neutron fluence is presented for the two different irradiation experiments. Both the 
IRL and ORR irradiations show a negative dependence of the resistivity increment fraction 
on the fast neutron fluence over the total fluence region rather than the normal positive de¬ 
pendence expected for radiation-induced defects (vacancies and interstitials). The IRL speci¬ 
mens, however, show a somewhat higher fluence dependence than the ORR specimens. Several 
factors can contribute to the difference: (1) differences in the metallurgical conditions be¬ 
tween the two sets of samples, (2) differences in the reactor spectrum between these irradia¬ 
tions, and (3) differences in the temperatures of the irradiation. 

Based on the known location of the test specimens within the original block of steel and on 
annealing studies of unirradiated specimens, the difference in fluence dependency is believed 
due primarily to variations of solutioned carbon from specimen to specimen. These annealing 
studies indicate that the IRL specimens were probably cut from a section which had more 
carbon in solution; hence more was available for radiation-induced precipitation. Carbon 




106 


analysis will be obtained to verify the implications of these resistivity studies. The effects 
of spectra and temperature were evaluated, but are probably minor compared to effects of 
pre-irradiation heat treatment. Other investigators 11 have also shown that the heat treat¬ 
ment of A302B and other mild steels has a major effect on the sensitivity to fast neutron 
irradiation. 

Figure 2. 21 shows the dependence of the nil ductile transition temperature (NDT) on the 
fast neutron fluence for A302B steel, 12 and illustrates a good correlation with the resis¬ 
tivity fluence dependence. 






— 






































































































J 

> 

"" 


















9 ^ 



























- 






_r 

C 





O 

Ap = P-P 0 

where 

p = resistivity after 1-hr anneals 
p Q = unirradiated resistivity 
{^) m j n = lowest observed value of (^) 

O-0RR 
□ - IRL 

1. . 1_1_1_1_1_1_l.I,- 




□ 

































_l—1 





































1018 10 19 10 20 

Fast neutron fluence, n/cm^ (E n > 1 Mev) 


Fig. 2.19 - Dependence of the fractional resistivity minimum, (Ap/p 0 ) m j n , 
of irradiated A302B steels on the fast neutron fluence 



Fig. 2.20 — Fractional decrease in resistivity of IRL and 
ORR irradiated A302B steel as a function of 
fast neutron fluence 



19 2 

Fast neutron fluence, 10 n/cm 


Fig. 2.21 — Increase in the nil ductility transition tem¬ 
perature 15 of irradiated A302B steel as a 
function of fast neutron fluence 




107 


TRANSMISSION ELECTRON MICROSCOPY (R, C. Rau, D. A. Woodford,* J. P. Smith, 

J. Moteff) 

Previous results 13 on irradiated unstressed A-286 containing 0. 001 and 0.010 weight 
percent natural boron indicated that all observable helium gas bubbles were associated 
with dislocations. There was no preference for bubbles to move to grain boundaries on 
post-irradiation thermal treatments up to 955 °C. To determine whether the combined 
effects of stress and temperature altered helium distributions, sections were cut from 
the stressed portion of the same specimens which had been creep-rupture tested at 660°C 
and previously evaluated in the unstressed condition. The specimens were sectioned 
within 0. 2 cm of the fracture and heat treated at 760°, 845 °, and 955 °C. Examination of 
the sections from the as-tested specimens revealed very few bubbles, too few and too 
small in diameter to evaluate. The post-test heat-treated specimens, however, had many 
bubbles large enough to be evaluated. The size and size distribution appeared to be un¬ 
affected by the combined effect of stress and temperature; i.e., the results were the 
same as found on the specimens which were only thermally treated and not stressed. The 
size distribution as a function of boron content and anneal temperature is shown in Figure 
2. 22. There is no systematic difference between the stressed and unstressed sections 
taken from the specimens. 

Figure 2.23 shows an area containing a grain boundary triple point and numerous ma¬ 
trix dislocation lines. Of significant interest are the dislocations leading into the grain 
boundaries; there appears to be no driving force for the helium on these dislocations to 
migrate to the boundary, although one would expect the dislocation to be a good diffusion 
pipe. Several small bubbles are present at the triple point. Figure 2.24 represents an 
area containing a rather high matrix bubble density compared to the few bubbles seen on 
the grain boundary. Significantly, all bubbles in the boundary also lie on dislocation lines 
within the boundary; these would probably impede dislocation movement within the boundary 
and thereby restrict grain boundary sliding. 

The presence of shells with a high bubble density around certain precipitate particles 
was reported previously for A-286. 13 Recent investigations on another alloy, Hastelloy 
R-235 containing 50 ppm boron enriched in isotope, revealed a similar post-irradiation 
microstructure. Figure 2. 25a illustrates the relative number of atom displacements as a 
function of recoil distance for both the alpha-particle and lithium. Essentially all damage 
is produced at the end of the recoiling particle path. The relative number of displacements 
caused by the lithium is somewhat higher than for the alpha-particle. Based on these fac¬ 
tors one would expect two relatively narrow damage bands containing numerous atomic dis¬ 
placements as shown schematically in Figure 2. 25b. Both bands (shells) were observed 
in A-286 and Hastelloy R-235 as shown in Figure 2.26. Figure 2„26a shows shells around 
three separate precipitate [Ti(N, B)] particles. For two particles an inner and outer shell 
are apparent. The bubbles in the outer shell are probably helium bubbles precipitated on 
dislocations; bubbles on the inner shell could be helium or vacancy clusters. The highly 
damaged matrix could limit the normal range of an alpha-particle by minimizing channel¬ 
ing, or the high vacancy concentration resulting from the high damage state around the 
lithium atom may form large voids visible on post-irradiation heat treatment. Figure 
2. 26b shows a similar condition in Hastelloy R-235 but probably in a different stage of 
development; all that is observed are concentric dislocation shells. Higher-temperature 
annealing will presumably reveal voids or bubbles. 


General Electric Company, Materials and Processes Laboratory, Schenectady, New York. 




108 



BUBBLE DIAMETER A 

0.010 weight percent natural boron 



BUBBLE DIAMETER A 

0.001 weight percent natural boron 


Fig. 2.22 — Size distribution of helium bubbles as a 
function of boron content and annealing 
temperature in stressed and unstressed 
A-286 alloy 









Fig. 2.23 — Transmission electron micrograph of irradiated A-286 showing helium bubbles on 
matrix dislocation lines and several small bubbles at the triple point 


Fig. 2.24 — Transmission electron micrograph of irradiated A-286 showing bubbles attached 
to dislocation lines in the matrix and in the grain boundary 






no 


To fully explain radiation-induced changes in metals it is first necessary to describe 
deformation mechanisms of unirradiated alloys. An important phase of the current work 
is the examination of unirradiated specimens which have the same history as the irradiated 
specimens except for neutron exposure. Transmission electron microscope observation 
of unirradiated Hastelloy R-235 specimens indicates that deformation at 870 °C was accom¬ 
panied by the pairwise motion of dislocations through both the matrix and the coherent y* 
particles. 14 A good example of a slip band containing paired dislocations in a low-stress 
region of a tested specimen is seen in the micrographs of Figure 2.27. As discussed pre¬ 
viously, 15 this pairing is due to the creation and annihilation of an antiphase boundary by 
successive dislocations as they pass through the ordered y’ particles. This process retards 
the motion of the first dislocation and accelerates the motion of the second, causing them 
to draw together and move as pairs. The mobility of dislocations is largely controlled by 
the energy requirements for creating and annihilating an antiphase boundary; hence strength¬ 
ening of coherent y*-hardened alloys depends upon the distribution and volume fraction of 
y* present. 16 ’ 17 


The effect of strain on the interaction of dislocations with y' particles was examined in 
detail, by cutting foils from three different parts of the deformed region; deformation is 
defined as the reduction in cross-section area. This technique involved examining foils from 
sections deformed at 870 °C to 1, 7, and 12 percent strain. 

Figure 2.28a shows the microstructure in a foil taken from a region of relatively low 
strain (~1% reduction in area). Mosty’ particles have lost coherency and are encircled by 
one or more dislocations. Figure 2. 28b shows the microstructure in a region of intermedi¬ 
ate strain, approximately 7 percent. At this strain level, all the y' particles are encased 
in dislocation networks. The dislocation density within these networks, however, is lower 
than in those surrounding y' in the high-strain ( ~ 12 %) region, as shown in Figure 2. 28c. 
Similar dislocation networks surrounding y' particles were reported recently 18 in strained 
Inconel X-750, but were not discussed. 

This streSs-induced formation of dislocation networks at the interface between y' particles 
and matrix is believed to arise from the pinching off of loops from moving dislocations whi ch 
bow around and are forced past the particles, leaving a loop for each passing dislocation. 
Increasing deformation deposits additional loops at the interface, adding to the dislocation 
density within the networks. As this dislocation density increases, mutual repulsive forces 
between the dislocations within the networks make it more difficult to add further loops, and 
thus lead to a hardening effect in the alloy. 

In regions of high-stress concentration, dislocations would be expected to bow around co¬ 
herent y* particles rather than to shear them if the retarding force on the leading dislocation, 
i. e., the antiphase boundary energy of the y', were high enough. This is apparently true of 
Hastelloy R-235; no areas were observed which showed massive shearing of y’ particles by 
glide dislocations, as have been observed in other alloys by Copley and Rear 17 and Gleiter 
and Hornbogen. 19 The mode of dislocation immobilization in y* precipitation hardening alloys 
is probably very sensitive to the alloy composition and heat treatment, since these factors 
influence the composition and degree of ordering of the y' phase which in turn determine the 
antiphase boundary energy. 


Ill 



Range, ft 


a. Relative displacement 


2.0fi 

i-:—i 



b. High-damage regions 

Fig. 2.25 — Relative atom displacement and expected high-damage regions 


In addition to the transmission electron microscopy discussed above, surface replication 
and electron microscopy is also being performed on these same Hastelloy R-235 specimens. 
Primary emphasis is being placed on the appearance of the structure immediately adjacent 
to the fracture, since past experience with similar alloys has shown that the microstructures 
in the unstressed condition are essentially identical before and after irradiation. Comparison 
of replicas from the stressed areas reveals significant differences between irradiated and 
unirradiated specimens, as illustrated in Figure 2.29. Of interest is the alignment of y f 




112 



a. Helium bubbles around precipitate particles in irradiated A-286 



b. Shell of dislocations around precipitate particle in irradiated Hastelloy R-235 

Fig. 2.26 — Helium bubbles and shell of dislocations around precipitate particles 
in irradiated A-286 and Hastelloy R-235 


within each grain of the unirradiated specimen (Figure 2.29a); in the irradiated specimen 
(Figure 2.29b) there is no tendency for alignment. This rearrangement ofy' is apparently 
associated with grain deformation since the unirradiated specimen exhibited approximately 
18 percent elongation at rupture, but the irradiated specimen elongated only 1 percent 
prior to fracture. No tendency for y' alignment was observed in the unstressed, unirradi¬ 
ated specimen (thermally treated but not stressed) as shown in Figure 2.30, which supports 
the above conclusion that the alignment is strain induced. 







113 



Fig. 2.27 - Slip band containing paired dislocations in low-stress region in Hastelloy R-235. 
(a) (111) reflection operating, (b) (200) reflection operating. 


The microstructures in Figures 2. 30a and b are from the same specimen; the difference 
is the etchant. The structure in Figure 2. 30a was obtained by using Marbles etchant; 2. 34b 
was obtained by electro etching with 10 percent H 3 P0 4 in H 2 0. The H 3 PO 4 etchant gives a 
much better definition of y* structure, particularly in grain boundaries, where it is shown 
in Figure 2. 30b that virtually all carbide particles (matrix as well as grain broundary) are 
coated with y’. This does not occur with the Marbles etchant. Further work will emphasize 
the H 3 PO 4 etchant because it more truly illustrates the structure. The lack of alignment of 
y ? in the irradiated specimen indicates that grain deformation is reduced or eliminated by 
irradiation. It cannot be directly concluded that this indicates matrix stiffening, since it 
could be the result of premature grain boundary failure. 

2. 3 REFRACTORY METALS AND ALLOYS PROGRAM 

Data are presented on the creep-rupture properties of molybdenum specimens of two 
carbon levels. Tests were performed on specimens irradiated to neutron fluences greater 
than 1 x 10 20 n/cm 2 (E n ^ 1 Mev) at 70°, 700°, and 1000°C. In several cases specimens 
were post-irradiation annealed at various temperatures and times to determine the effect 
of annealing on the neutron-induced changes in the creep-rupture properties. 

Elevated-temperature and room-temperature tensile test results on niobium, Nb — IZr, 
and molybdenum are discussed. Post-irradiation anneals were given to a few specimens 
to study thermal hardening and recovery phenomena. Hot-hardness tests on several re¬ 
fractory metals and alloys were completed and are compared. The recovery of radiation- 
induced defects in molybdenum was investigated by isochronal resistivity studies. Studies 
continued on the effect of changes in defect structures upon the mechanical properties. An 
understanding of these effects will help to define the mechanisms of neutron-induced rad¬ 
iation changes in refractory materials. 

CREEP-RUPTURE TESTING (F. Kingsbury, R. Treinen, J. Moteff) 

Molybdenum Rod 

A study of the effect of irradiation temperature on the creep-rupture properties of 
molybdenum at 750°C was conducted. The material used in this study contained a 
relatively high carbon content, approximately 220 ppm. The effect of irradiation tem¬ 
perature was studied on recrystallized, poly crystalline specimens irradiated at 70° 
(reactor-ambient), 700°, and 1000° C, to fast fluences in the range 1. 3 x 10^ to 1. 5 x 10^ 




> * 


Relatively low-strain (—1 %) region 


14 



Medium-strain (~7%) region 


High-strain (~12%) region near fracture 


Fig. 2.28 — Electron micrographs showing formation of dislocation networks around 
Y particles as a function of strain in creep-tested Hastelloy R-235 



115 



a. Unirradiated, 21% strain, t R 200 hours b. Irradiated, 0.6% strain, t R 1 hour 

(p p = 2.5 x 10 19 n/cm 2 (E n > 1 Mev) 

<p Jh = 1.6 x 10 20 n/cm 2 at 70°C in ORR 

Fig. 2.29 — Longitudinal sections of creep-rupture-tested Hastelloy R-235 immediately 
adjacent to fracture. Creep tested at 870°C, 14.1 kg/mm 2 . (Marbles etchant) 




a. Marbles etchant 


b. 10 % H 3 PO 4 electro-etch 


Fig. 2.30 — Unirradiated Hastelloy R-235 exposed 200 hours at 870°C, unstressed 





116 


n/cm (E n ^ 1 Mev). All irradiated specimens were tested at a common initial stress 
level of 18.00 kg/mm^ (Table 2. 3). Two irradiated specimens were given 1-hour post¬ 
irradiation anneals at 1000° C. A test temperature of 750° C, based on the transmission 
electron microscope studies of Mastel and Brimhall, 20 was selected for this study. Their 
work showed that the observable spot and loop defect density in recrystallized molybdenum, 
irradiated at reactor-ambient temperature to a fast neutron fluence of about 1 x 10^ n/cm^ 
(E n l Mev), decreased rapidly following post-irradiation anneals in the temperature 
region above 750° C. Thus it was thought that the irradiation of molydenum above and below 
this temperature (~ 0.35 T m ) would result in significant differences in post-irradiation 
creep-rupture properties. Creep-rupture properties of molybdenum at 750°C for speci¬ 
mens irradiated at three different temperatures are compared with a control specimen 
in Figure 2. 31. Irradiation at the lower temperature (~ 70° C) resulted in the least effect 
on the time to rupture. There was, however, an initial period of "delayed" creep in which 
the creep rate remained very low for about 5 hours. The specimens irradiated at 700°C 
and 1000° C show a factor of 12 and 18 increase, respectively, in rupture life, with a slight 
decrease in ductility at the higher temperature. There is a relatively long second-stage 
creep behavior which increases with increasing irradiation temperature. 

Two irradiated specimens were post-irradiation annealed at 1000° C in vacuum for 1 hour 
and creep-rupture tested. The effects of these anneals are shown in Figures 2.32 and 2. 33, 
Annealing an ambient-temperature irradiated specimen altered the creep-rupture proper¬ 
ties; the period of delayed creep was shortened, creep rate was reduced, and time to 
rupture was increased compared to the as-irradiated specimen. Changes produced by a 
1000°C anneal in the creep properties of a 700° C irradiated specimen were not so pro- 
mounced. The minimum creep rate did not show a significant change, rupture life was 
increased by 25 percent. 



Fig. 2.31 - Creep-rupture properties of polycrystalline Mo at a test temperature of 750°C 
and an initial stress level of 18.00 kg/mm^ following neutron irradiations at 
temperatures of 70°, 700°, and 1000°C 



117 


00 

cm 


CQ 

< 


= < 


-Q " 
c 



CD 

O) 

nP 

q 

1 

to 

to 

to 

o 

3 

00 

d 


q 

q 

r> 

34.0 

- 


c 

o 


to 


to 

<3- 

to 


3, 


to 

to 


LU 













o 




1BI 



00 


■o^ 





D 

oT 


o 

CM 

rr 

CO 

CO 

«— 

to 


i— 

r- 

r— 

Q_ 

H- 

k- 


o 

o 

to 

00 

CO* 

r^‘ 

to 


d 

to 

'3 

DC 

1| 


a) 

CM 

CM 

to 


to 

00 

to 


o 

to 



to 







CM 

CM 

to 

to 


O a 

_ OJ 

“ VJ 

g> 2 

<a ra 
£ * 


H- Q. 
o 0 


o O 


■O I 


a +- 

X (O 

_ DC 


to (J 

n 

** fc 


O D 
CO £ o 

■o g.° 

S £ 

— 0) 


tt - 

CO A 

U- A 


oo 


to 


00 

to 

3 

co 

CM 

to 

1 to 


CM* 

CM 

1 <cfr 


co 

to 

<Ji 

03 

a> 

O) 

05 

03 

0) 

05 

03 


*P 


--.totococor^r-.coooco 
. I I I I I I I I I I 
ooooooooooo 

XXXXXXXXX XX 

[>.<ci’COtf>ir>CNCOC0C'0 *- cm 

tri Ifl ^ tf r-’ 00 oi N N N 


. L I M 

to LO 

o o l 


I I 1 


I I 


I-* r*. oo 
I I I 
o o o 

I X X X 
O’ 00 IT) 

HI 


I I I 


0) <D Q) <D CD CD OQ) <D 

ccccccTccT c 

o o o o o o I o o I o 

zzzzzz^zzpz 

o o 

s § 

o o 


I I I I I 


o o o o o o 
r-* r-» o o o 

1 * » 77 t 


D H) 0) 0 


c c c c c 
O O O ‘ 

z z z 


aj oh- ai oir 

o o *- o o o 

CM CM CM CM CM CM 
O O O O O O 

O T— T— <— T— T— 


X X X X X 


o o o o 


o o o o o 
2 Z Z Z Z 


oo^-ooo 

CM CM CM CM CM CM 
« ° O O O O O 

X X X X X X 

in to r> in in to 

^ 'O'* CM ^ 


toOrHO)^h*.«-CMtoCM 
rt-r-(3)0)0)0)CNtN i- 

cocDcototototocococD 


O 

°o 

o 


"8 


a 

o 

2 


c 

o 

'E 

o 

a 


a 

o 


§ 


c .C 
11 
.S3 

> in to 

J5 UJ ^ LU LU 

§.fIM 

• -5 ~ * ~ 

CD 


to 


*9 c 

£ 

V a 
00 
o 

to 

d = oo 

£ £ g 

P ^ o 


li a 

£ 


<0 <o 
Z •*- 4 ~ 
DC DC 

T t— H 

to LU LU 

3 c .£ c c 


a> J2 
8 c 

£ t 


O CM 


CO 

s s « 

™ CM CM 


c id o L p t; f 

25 ° 2 SSS£ 


£ ' “ O 

.£ S* 

8 e S 

Q. <D <1> 


TO 

E 
£ .2 
£ £ 

I'| 

to 

XJ o 


o § I 

,2 CO CD 

■o 3 *o 

J 2 2 

~ oT jE" 



Strain, percent 








119 


Specimen No. 1591 which was irradiated to a neutron fluence of 1. 5 x 10 21 n/cm 2 (E n ^ 

1 Mev) at ambient temperature permits a comparison of the effect of neutron fluence. 
Compared to specimen No. 1594 (1. 35 x 10 20 n/cm 2 ), Figure 2. 34, which was irradiated 
to a fluence an order of magnitude lower, specimen No. 1591 exhibited a factor of 4.6 
increase in time to tupture and almost an order of magnitude decrease in creep rate. A 
portion of the greater rupture life was due to the increased length of the delayed creep 
period in the higher fluence specimen. 

The differences in the creep characteristics between the various specimens can also 
be noted in the instantaneous creep rate versus plastic strain plots of Figures 2.35 through 
2.37. Specimens irradiated at ambient temperature (Figure 2. 35) all show a rapidly de¬ 
creasing creep rate in the initial 0. 2 percent strain. This is followed by a rapidly increas¬ 
ing creep rate which reaches a maximum value in the 2 to 6 percent strain region. After 
reaching this maximum value the creep rate decreases to the usually minimum second- 
stage creep rate. An expanded plot of the initial 1 percent strain (Figure 2. 36) shows the 
ordering of the inflection points in the creep rates. The initial rate of decrease is approx¬ 
imately constant and the minimum value at the inflection point shifts to higher strain values 
at decreasing creep rates. 

The creep-rate strain behavior of specimens irradiated at elevated temperatures 
(Figure 2.37) showed two differences in the characteristics compared to the ambient- 
temperature irradiated specimens. No maximum is observed in the creep rate following 
the initial inflection point. The three specimens exhibited an increase in creep rate with 
increasing strain following the initial decrease in creep rate. Irradiation at 1000°C did 
not produce the marked inflection at the minimum creep rate observed in the 700°C irra¬ 
diated specimen or the 700°C irradiated specimen with the 1000°C post-irradiation anneal. 
The 700° C irradiated specimens have higher creep rates over a greater range of strain 
(and hence time) than the 1000°C irradiated specimen. The 1000°C post-irradiated annealed 
specimens also showed creep rates lower than the 700°C as-irradiated specimen. These 



« 25 50 75 100 125 150 175 200 225 250 

Time, hours 


Fig. 2.34 — Creep characteristics of irradiated high-carbon Mo rod at two neutron fluence levels. 

Specimens irradiated at approximately 70°C and tested at 750°C and 18.00 kg/mm^. 







120 



Fig. 2.35 - Creep rate versus plastic strain for ambient 

temperature irradiated high-carbon (200 ppm) 

Mo specimens tested at 750°C and 18.00 
kg/mm 2 

results suggest that some of the more stable defects are possibly formed between 700°C 
and 1000° C. 

It is believed that the substructure, i. e., isolated dislocations, tangles, and subgrain 
boundaries formed during the initial straining establish the mode of deformation in creep. 
Structural studies are being conducted on these high-carbon molybdenum rod specimens, 
both optically and by transmission electron-microscope techniques. Initial electron micro¬ 
scope observations will be reported in a subsequent section. 

Molybdenum Sheet 

Creep-rupture testing of arc-cast molybdenum sheet was conducted to gain an under¬ 
standing of the effect of changes in the defect structure produced by post-irradiation 
annealing on the mechanical properties. Preliminary tests 21 exhibited accelerated creep 
at 580°C and delayed creep at 700°C. Additional tests were conducted on material at two 
carbon levels, 205 ppm and 26 ppm, to confirm these results. Most tests were conducted 
at fixed stress levels and test temperatures; the irradiated specimens received various 
annealing treatments in the temperature range of 700° to 1600°C (0.34 to 0.65 T m ). 





121 


A composite strain — time plot for control, as-irradiated, and 750 °C post-irradiated 
low-carbon molybdenum specimens tested at 580°C is shown in Figure 2.38. The accel¬ 
erated creep of the as-irradiated specimen is evident. The 750°C (0.36 T m ) anneal pro¬ 
duced a lower creep rate and longer rupture life than the as-irradiated specimen, but did 
not fully restore the creep characteristics to the control values; hence the defects causing 
the accelerated creep were only partly removed at 750°C. 

Tests at 700°C on low-carbon molybdenum included specimens which had two neutron 
fluence levels (Table 2.4). All irradiated specimens were tested at 17. 50 kg/mm 2 except 
for specimens which had been post-irradiation annealed above 1400 °C. These specimens 
were tested at lower stresses due to grain growth which had occurred in the annealing 
process. Originally all low-carbon specimens were annealed at 1200°C in vacuum for 1 
hour following fabrication. 



Plastic strain, percent 


Fig. 2.36 — Creep rate versus plastic strain for irradiated 
high-carbon (200 ppm) Mo specimens tested 
at 750°C and 18.00 kg/mm 2 




122 


N 

LU 

-I 

m 

< 


O 

§ 

pH 

S 

GO 


%- 

LU 

UJ 

X 

to 

2 

3 

Z 

LU 

Q 

CD 

> 

O 

5 

z 

o 

m 

cc 

5 

§ 


o 

LU 


3 G> 
4-* u-. 

Q. “ 


£ I 
O » 


8 > § 
&& 


4 _ a 

° S 

O o w 

\p — -C 
ro <u 

“ m 


li 

a> v 

Q 3 

— to 

- oc 


CO 


o 


.CM 

I| 

co °> 


1 o 

2 O 


tt 

o 

Q_ 


t5 

LU A\ 


lO 

q 

CO 

CM 

CO 

O 


q 


^ 1 - 

O 

00 



q 

in 

in 

q 

O 

CO 

pi 

cd 

t-T 

o 

CO 


r- 

o 

cd 

in 

cd 

o 

1 


o 

cd 


CM 


tp 

in 

in 

in 

in 





’Sp 


't 



TP 

TP 

in 

TP 


CO 

LO 

CO 

o 

pn 

CO 

r*. 

q 

CO 

00 

Ph 


CO 

CO 


CO 

CO 

00 

r -1 

cd 

pi 

cd 


in 

pi 

a> 

r»l 

cd 

d 

d 

in 

o 

o 

M-’ 

cd 

cd 

cd 

to 

a> 

CO 

CO 

CO 


tp 

CO 

CO 

4 d- 

CO 

05 

CO 

CO 

in 

P*. 

CO 

PH 

PH 

CM 











>r- 







*“ 

pH 

1 

^ CD CD 

i i i 

CO 

1 

pH 

CO 

1 

CO 

■ 


CO 

1 

CO 

r T 


CO 

1 

<9 

CO 

1 

CD 

1 

7 

PH 

1 

o 

o 

I 

o 

o 

o 

o 

o 

i 

o 

o 

o 

o 

o 

o 

o 

o 

o 

1 

o 

o 

1 

2 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

in 

Ta¬ 

Tt 

r> 


CO 

in 

00 

LD 

CO 

CO 

LO 

p* 

CO 

TP 

CM 

CM 

o 

«— 


rn 

CO 

00 

q 

o 

r- 

p*. 

O 

q 

in 



CO 

CM 

q 

o 

cd 

cd 

CM CO 

T“ 

*“ 

»- 

cd 


in 

in 


cm’ 

cd 

cd 

cd 

cd 




Y 











CO 


00 







0 ) 

CD 

CD 

CD 

CD 

ai 

a) 

CO 

o 

°? 


<D 

r— 

| 

? 

00 

t 

a> 

(D 

<D 

c 

C 

c 

o 

C 

c 

c 





C 


i 


t 

c 

C 

C 

o 

o 

o 

O 

o 

o 

o 

1 

1 

LO 

in 

O 

in 

in 

in 

tn 

o 

O 

o 

z 

z 

z 

z 

z 

z 

Z 

CM 

r— 

o 

o 

z 

o 

o 

d 

o 

z 

z 

z 








I s * 

I 

r> 


p« 


°? 

00 

i 

7 

7 











O 

o 

o 

o 


o 

o 

o 

o 




1 

1 

1 

1 

i 

1 

1 





1 





i 

1 

■ 

1 

1 

1 

1 

r 

1 

i 

X 

X 

X 

X 


X 

X 

X 

X 

i 

1 

i 








T— 


»— 



Ti¬ 


CM 

o 












q 

o 

*7 


pi 

cd 

cd 

CM 























o 

o 

o 

o 

o 

o 


o 

o 

o 

o 

O 

o 

o 

o 

o 

o 

o 

o 

o 

in 

in 

in 

q 

q 

o 

q 

q 

q 

q 

q 

q 

in 

q 

in 

q 

q 

o 

LO 

cd 

pi 

pi 

pi 

tp* 

cd 

p >1 

r>! 

p»l 

p*! 

pi 

pi 

pi 

pi 

pi 

pi 

TP 

cd 


r ” 

T~ 

T- 


r~ 

*— 




*— 


r** 















CM 













CM CM z 

CM 

X 

CM CM CM 

CM CM CM CM s 

M 




I 

i 

X 

T 

X 

1 


\_ 

X X 

1 1 

X 

1 


X 

X 

X 

X 

i 

f 

X 

1 




1 

lm 

k_ 

i 

1 

1 - 


-C 


i 

U- 

t- 



I 

J. 

l 

l. 

u. 

l 

k. 

CD 

CD 

CD 

JO 


iE 

-C 

(D 


£ 

JZ 


a) 

-C 


£ 


s: 

JO 

C 

o 

C 

o 

C 

o 

T- 

r~ 

T- 

r— 

C 

o 

r— 


t- 

T- 

c 

o 

T- 

T— 


«- 

«- 


z 

z 

z 

1 1 
o o 

1 

o 

1 

o 

z 

l 

a 

1 

o 

1 

O 

1 

o 

z 

1 

o 

1 

o 

1 

u 

1 

a 

1 

O 

1 

o 




o 

O 

o 

o 


o 

o 

o 

o 


o 

o 

o 

o 

o 

o 




o 

o 

o 

o 


o 

o 

o 

o 


o 

o 

O 

o 

o 

o 




o 

o 

in 

o 


o 

o 

o 

o 


in 

o 

in 

o 

in 

o 




CM 

CM 

tp 

CO 


p*. 

00 

05 

CM 


00 

05 

05 

CM 

TP 

CO 




r— 



«■“ 










T— 

















CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 








O 

O 

o 

O 

O 

O 

o 

o 

O 

O 

O 

O 








X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

CD 

(D 

a> 

a) 

CD 

a) 

a) 

T- 

<— 

»“ 

1— 

«— 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

c 

c 

c 

c 

C 

c 

c 

cd 

cd 

cd 

cd 

cd 

t— 



«— 

T-* 

r— 

»— 

O 

o 

o 

o 

o 

o 

o 













z 

z 

z 

z 

z 

z 

Z 




















o 

o 

o 

CD 

o 

o 

O 

o 

o 

O 

O 

o 









CM 

o 

CM 

O 

CM 

O 

% 

•M 

O 


% 

% 








*— 

*— 


T— 

T— 

.— 

T— 

r— 

T— 

.— 

*— 

t— 








X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 








CO 

CO 

CO 

CO 

CO 

05 

05 

05 

05 

05 

05 

05 








ci 

d 

d 

d 

d 

cd 

cd 

cd 

cd 

cd 

cd 

cd 


CM 

CO 

Tp 

00 


CD 

o 

CM 


in 

CO 

00 

CO 

CM 


o 

in 

CO 

o 

»— 


T— 

CM 

CO 


o 

o 

o 

o 

o 


CM 

CM 

CM 

CM 

CM 

CM 

r*. 

rn 

m 

m 

p* 

r*> 

P**- 

r^- 

P* 

p*. 

ph 

pH 

pH 

pH 

PH 

PH 

Ph 

PH 

pH 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 

CM 


E 

o 

l 

o 

°o 

o 

CM 


s 


c 

CD 

*a 

c 

I ■ 

tx" 

^ c 

CM ^ 

K s 

t* o 


c 

o 

« 


s 


O a) 

s r 

« < 

i i 

CD Q. 

E « 


°o 

c *-» 

1| 

i! 

E g 
“ 15 
£ E 
<5 ? 
15 o 

I « 

w CD 
£ 2 
o to 
S LU 
£ Z 

LL £“ 


7 Ji gE 


> £* 


QC UJ 

O - E 

E eg 

a . 5 2 

to £ * £ 

CM 5> 2 it 


E 

£ 

8 £ 5 CC L_ 

E #;oa 

o c o> a> o 

-.»*u 


£ £ 


U 


\ I 8 


E -P 
o „ 

4= o 

c CO •£ o o 

£ O) +3 

| r» C ffl ,2 

T O TJ ‘a 

_ i & g g 

CO Q s- — .== 

CD 5 O TJ 


8 


123 



Fig. 2.37 - Creep rate versus plastic strain for elevated-temperature irradiated 
high-carbon (200 ppm) Mo specimens tested at 750°C and 18.00 

kg/mm^. 


Specimens in the as-irradiated condition or post-irradiation annealed below 1000°C 
exhibited delayed creep (Table 2.4). The apparent length of this incubation period in which 
the creep rate is abnormally low tends to decrease upon annealing in the 700° to 950°C 
temperature range. The 700° to 800° C anneals produced only small changes in the creep 
rates or rupture times as shown (Figure 2.39) for the specimens irradiated to a fluence 
level of 3.1 x 10 20 n/cm 2 (E n 1 Mev). Annealing at 900°C or 950°C produces little fur¬ 
ther changes in the incubation period but does decrease the steady-state creep rates and 
increase the times to rupture. A complete change in creep characteristics occurs upon 
annealing at 1200° C. The period of incubation (i. e., delayed creep) is not exhibited and 
the strain - time plots show the normal three stages of creep behavior. This is shown in 
Figure 2. 40 for specimens irradiated to 1. 2 x 10 2 ® n/cm 2 (E n — 1 Mev). 

Tests at 700°C on specimens post-annealed above 1200°C are not directly comparable 
to the preceding results. Nevertheless, they may be compared to control specimens an¬ 
nealed at the same time and tested at the same stress level. Tests conducted on specimens 
annealed at 1450°C and 1600°C (0.60 and 0.65 T m , respectively) did not show delayed 





124 



Fig. 2.38 — Effect of 750°C post-irradiation anneal on the accelerated creep 
of irradiated low-carbon molybdenum. Tests conducted at 580°C 
and 19.40 kg/mm 2 on sheet specimens irradiated at reactor 
ambient temperature to a fluence level of 4.8 x 10^ n/cm 2 
(E n > 1 Mev). 



Fig. 2.39 — Effect of post-irradiation anneals at 700°, 800°, and 900°C on 

the creep properties of low-carbon molybdenum. Tests conducted 
at 700°C and 17.50 kg/mm 2 on sheet specimens irradiated at 
reactor ambient temperature to a fluence level of 3.1 x 10 20 n/cm 2 
(E n > 1 Mev) 




125 



Fig. 2.40 — Effect of post-irradiation anneals at 900°, 950°, and 1200°C on creep 

properties of low-carbon Mo. Tests conducted at 700°C and 17.50 kg/mm^ 
in a hydrogen atmosphere. Anneals performed at indicated temperatures 
for 1 hour in a hydrogen atmosphere. 

creep. Pronounced neutron-induced strengthening still remains under these test condi¬ 
tions, as can be seen in the 1600°C annealed specimens, Figure 2. 41. 

A limited number of creep tests were also conducted at 700°C on high-carbon (~200 
ppm) molybdenum specimens to observe the possible effect of particle — defect interac¬ 
tion on creep behavior. The differences in the two grades of material can be seen in the 
microstructures. The high-carbon molybdenum exhibited a dense, equiaxed, recrystal¬ 
lized grain structure with randomly distributed carbide particles. The particles were 
generally small but a few larger carbides were observed. Several larger carbides ap¬ 
peared to have voids associated with the ends of the particles (Figure 2.42). In contrast 
to the high-carbon molybdenum sheet, the low-carbon molybdenum microstructure (Fig¬ 
ure 2.43) consisted of single-phase grains with no carbides observable at a magnification 
of 600X. The grain size of the low-carbon material was slightly larger (30-micron versus 
22-micron grain sizes in the high-carbon Mo). The grains in the low-carbon material 
tended to be elongated in the direction of rolling. 

A composite strain — time plot for the results of high-carbon molybdenum tests given 
in Table 2.5 is shown in Figure 2.44. The 1200°C annealed specimen exhibited no incuba¬ 
tion period; creep rate was reduced significantly and rupture life increased 37 percent 
compared to the as-irradiated specimen. The 900°C anneal tended to remove the incubation 
period and produced a longer rupture life than the as-irradiated or 1200 °C post-irradiation 
annealed specimen. This effect is believed related to the thermal hardening generally ob¬ 
served in tensile tests of irradiated body-centered cubic (bcc) materials following elevated- 
temperature anneals. In the present case, the thermal hardening can be related to rupture 
times (rupture strength) rather than to tensile strength. 

The differences in the creep characteristics of the high-carbon molybdenum specimens 
can be readily observed in the instantaneous creep rate versus time plots of Figure 2. 45. 











Specimen 

No. 3 

Condition 

Post-Irradiation 

Anneal 

Stage-1 (Delayed) 
Creep Rate, sec - 1 

Stage-1 Creep 
Period, hr 

Stage- II Creep 
Rate, sec - 1 

Rupture 
Life, hr 

Elongation, 1 

% 

2103 

None 

None 

_ 

_ 

6.09 x 10-7 

118.9 

49.2 

2122 

Irradiated 0 

None 

~2.8 x 10-8 

1 to 10 

4.11 x 10-7 

180.5 

46.8 

2126 

Irradiated 0 

1200°C for 

1 hr in H 2 

_ 

— 

2.6 x 10 _ 7 

229.8 

41.3 

2127 

Irradiated 0 

900°C for 

1 hr in H 2 

~2.8 x 10-8 

1 to 5 

2.25 x 10-7 

271.4 

43.1 


aSpecimens from 0.05-cm-thick commercial vacuum arc-cast Mo sheet (heat C 6605) post-grind annealed at 1200°C for 1 hour in 
vacuum (~22 micron grain size — DPH 191). Carbon content 205 ppm; O 2 content 3 ppm. All tests at constant load in hydrogen. 
^Percent in 2.54-cm gage length. 

cIrradiation test ORM-38 in ORR facility A-2 at reactor ambient temperature (~70°C) to a fluence of 1.1 x 10 20 n/cm 2 (E n > 1 


Mev); thermal fluence of 5.2 x lO 2 ^ n/cm 2 . 


Normal creep-rate curves are exhibited by the unirradiated and 1200 °C annealed specimens; 
the 1200°C curve is displaced into the lower creep-rate regions. The 1-hour anneal at 
1200°C probably did not remove all the radiation-induced defects (possible stable vacancy 
clusters). This view is supported by the results of the 1600°C anneal in the low-carbon 
molybdenum test data. The lower creep rate following the 1200°C anneal reflects disloca¬ 
tion interaction with the remaining defects. 

The curves for as-irradiated and 900 °C annealed specimens show the nature of the 
changes in creep rate during the incubation and early stages of creep. The as-irradiated 
specimen shows a maximum in the creep-rate curve (at ~ 40 hours) following the initial 
rapid decrease in creep rate (delayed creep). This maximum is not observed in the 900°C 
post-irradiation annealed specimen, and the specimen tends to show a continuous increase 
in creep rate following the delayed creep stage. 







128 



Fig. 2.44 - Effect of post-irradiation anneals at 900°C and 1200°C on creep properties 
of high-carbon Mo. Tests conducted at 700°C and 20.00 kg/mm 2 in a hydro¬ 
gen atmosphere. Anneals performed at indicated temperature for 1 hour in 
hydrogen. 


In general, the results between the low-carbon and high-carbon molybdenum sheet series 
of tests show agreement when compared at a similar fast-neutron fluence. The major dif¬ 
ference between the two series exists in the effects of 900 °C and 1200 °C post-irradiation 
anneals. The 1200 °C anneal in the low-carbon molybdenum test produced a longer rupture 
life than the 900 °C anneal; the opposite was true in the high-carbon molybdenum. The 
temperature of maximum thermal hardening (i. e., longer time to rupture) is probably re¬ 
lated to the carbon content since the fluence, grain size, hardness, and basic compositions 
of the two grades of molybdenum were very similar. 

Fracture Examinations 

Fracture areas of selected 700 °C creep-rupture tested specimens were examined metal- 
lographically to determine possible differences between the microstructure of the unirradi¬ 
ated, irradiated, and post-irradiation annealed specimens. In addition to the basic param¬ 
eter of post-irradiation annealing temperature, the examination included a consideration 
of the two carbon content levels, 26 ppm and 205 ppm. 

In general, the examinations did not reveal basic differences in fracture behavior between 
the various conditions. The basic difference between the two carbon levels at the immediate 
fracture tip was the presence of elongated voids associated with the carbide particles in 
high-carbon molybdenum. These voids were previously observed in the original material 
(Figure 2.42) but they were extremely small and localized. Nevertheless, the presence of 
the elongated voids did not impair the fracture ductility nor change the mode of fracture. 

The excellent ductility of the matrix, even in the as-irradiated specimens, is apparently 
controlling flow and fracture behavior. Fractures were all transgranular (Figure 2.46) and 
were accompanied by a relatively large amount of local plastic deformation (necking down 
to almost a knife edge). 

Microstructures of regions away from the immediate fracture area varied considerably 
between the low-carbon and high-carbon molybdenum materials. Three differences were 




Unirradiated and 1200°C post-irradiation annealed specimens (AS-746K) b. As-irradiated and 900°C post-irradiation annealed specimens (AS-746L) 










:ested at 70CrC and 17.50 kg/mm 








131 


noted: (1) A much finer plastically deformed grain size (Figure 2.47) in the high-carbon 
molybdenum due to differences in the grain sizes of the two original recrystallized 
molybdenum materials, (2) a considerably heavier density of low-angle grain boundary 
substructure in the high-carbon molybdenum, and (3) a concentration of relatively massive 
angular carbides in the high-carbon molybdenum (Figure 2. 48a). Only a trace of carbides 
was noted in the low-carbon material (Figure 2. 48b) and none of this carbide compared in 
size or shape to the chunky highly friable type present in the high-carbon molybdenum. 

The grain growth which occurred in the 1600 °C annealing of the low-carbon molybdenum 
is quite evident in the irradiated and unirradiated microstructures (Figure 2. 49). When 
viewed at a higher magnification (Figure 2. 50) the microstructure shows a relatively 
carbide-free matrix with agglomerated larger carbides distributed both intergranularly 
and intragranularly. The isolated particles of intragranular carbides are probably the re¬ 
sult of grain boundary migration away from the particles in the grain growth process. 

HOT HARDNESS (J. L. Kamphouse, J. Moteff) 

Warm-Worked Tungsten 

The objective of the hot-hardness study of warm-worked and irradiated tungsten was to 
determine the effect of substructure on the generation of radiation-induced defects as 
measured by changes in the hot hardness. Four specimens were prepared, each from 
sections of 73 percent and 98 percent warm-worked 0.51-mm-thick tungsten sheet. Two 
specimens of each worked condition were recrystallized at 1760 °C for 1 hour in hydrogen. 
One specimen of each of the four conditions (73% worked, 73% worked plus recrystalliza¬ 
tion, 98% worked, and 98% worked plus recrystallization) was irradiated at reactor- 
ambient temperature in the ORR Rabbit 14 facility to a fluence level of 4.1 x 10*9 n / cm 2 
(E n ^ 1 Mev). One specimen of each of the original unirradiated four conditions was used 
for comparison. 

As shown by the microhardness increment fraction curves of Figure 2.51, 73 percent 
worked and recrystallized material and the 98 percent worked and recrystallized material 
showed about the same radiation-induced hardening, primarily in the athermal component. 
The 73 percent warm-worked material showed much less radiation-induced hardening and 
the 98 percent worked material showed virtually no additional hardening. 

22 

Studies on the influence of dislocations on the damage structure in neutron-irradiated 
(~7x 10 18 n/cm 2 , E n ^ 1 Mev) molybdenum show that the spot density is about a factor of 
two greater in the annealed material than in the material deformed 50 percent prior to 
irradiation. Assuming that this type of spot defect contributes to the hardness of the irra¬ 
diated tungsten, it is believed that the increase in spot density in the recrystallized tung¬ 
sten and the relative ineffectiveness of the spot defects to further harden the worked 
tungsten may account for the greater radiation-induced hardening in the recrystallized 
material compared to that observed in the worked tungsten. 

A series of isothermal anneals was performed on 98 percent warm-worked tungsten. 
Results presented in Figure 2. 52 show some scatter in the data because some data were 
obtained from single indents rather than from an average of three indents as is customary 
when obtaining isochronal data. Because of this unavoidable data scatter at annealing 
times less than 10 3 seconds, a very accurate activation energy for recrystallization 
could not be obtained. An attempt was made to calculate the activation energy using the 
crosscut method, 23 and a value of 4.8 ev was obtained using data from specimens annealed 
at 1300°C and 1373 °C as shown in Figure 2.52. 




132 



Fig. 2.47 — Photomicrographs of fracture area of high-carbon Mo sheet specimens creep-rupture 
tested at 700°C and 20.00 kg/mm 2 (100X) 




133 



a. High-carbon Mo-heat C-6605 (Neg. R-2267) b. Low-carbon Mo-heat C-7722 (Neg. R-2268) 

Fig. 2.48 — Photomicrographs showing differences in carbide size and distribution in irradiated 

Mo sheet specimens post-irradiation annealed at 1200°C for 1 hour and creep-rupture 
tested at 700°C (500X) 

Niobium and Nb - IZr 

Hot-microhardness data were obtained on irradiated and unirradiated niobium and Nb — 
IZr. The Nb — IZr is harder over the entire temperature range due to solid-solution 
hardening and/or ZrC precipitation hardening. Some thermal hardening apparently started 
at a temperature of about 500 °C and persisted at temperatures to 1000 °C for the unirra¬ 
diated material. These data are given in Figures 2.53 and 2.54, respectively, and the 
microhardness increment fraction curves obtained are shown in Figure 2.55. These 
curves show recovery at the same temperatures observed in ultimate tensile strength tests. 

There are apparently peaks in the increment fraction curves for tungsten, molybdenum, 
and niobium at 0.16 T m and approximately 0. 35 T m ; the 0.16 T m peak is the more prom¬ 
inent. Figure 2.56 gives a comparison on a homologous temperature basis of the increment 
fraction curves for W, Mo, and Nb 24 , where it appears that the radiation-induced hardening 
of these unalloyed metals is recovered at temperatures above 0. 35 T m . 

Molybdenum Single Crystals 

Hot-hardness data have been obtained on two single crystals of molybdenum of known 
orientations, {110} and {001}. There is a distinct difference in the data dependent on the 
crystallographic orientation (Figure 2.57). The specimen of orientation {110} showed 
little difference from polycrystalline molybdenum previously tested, although the specimen 
of orientation {001} was considereably softer at all temperatures. Photomicrographs 
(Figure 2.58) of the indents taken at 1200®C show markedly different flow patterns for 
the two orientations. 



134 



Unirradiated — 1600°C anneal for Irradiated {1.2 x 10^® n/cm^) plus 

1 hour in H 2 (R-2263) 1600°C anneal for 1 hour (R-2255) 

Fig. 2.49 — Photomicrographs of creep-rupture tested low-carbon Mo sheet specimens tested at 
700°C and 13.0 kg/mm^ (100X) 

Other Refractory Alloys 

26 

Hot-microhardness data were determined on five materials in the unirradiated re¬ 
crystallized condition: Ta, Ta - 10W, W - 30Re - 30Mo (at.%; 306 alloy), W - 25Re - 
30Mo (at. %; 256 alloy), and Mo — 0.5Ti. These investigations were conducted to study 
the effects of alloying and to provide comparison data for similar specimens being irra¬ 
diated in the EBR-IL 

A previous discussion 27 of the tantalum and Ta — 10W data noted an anomalous peak at 
approximately 300°C. Subsequent analyses show that the tantalum material contained 33 
ppm oxygen prior to testing and 65 ppm after testing. Hence the peak may be due to in¬ 
terstitial impurities such as oxygen. 28 

HARDNESS AND ULTIMATE STRENGTH CORRELATION (J. L. Kamphouse, J. Moteff) 

29-32 

A relationship appears to exist between hot hardness and other properties such as 
ultimate tensile strength and yield stress. Figures 2.59 through 2.61 show approximately 
linear relationships for several materials between the ultimate tensile strength, cr u , in 
kg/mm^ and the VPH number, H v , in kg/mm^. These linear relationships are not perfectly 
accurate, but they do give an estimate of the ultimate tensile strength of the material if 
the relatively convenient hot-hardness tests are performed first. 





135 



Fig. 2.50 — Photomicrograph of unirradiated low-carbon Mo sheet specimen 
(1600°C - 1 hour anneal) showing agglomerated carbides. Speci¬ 
men creep-rupture tested at 700°C and 13.00 kg/mm^. 

(Neg. R-2269, 500X) 



Temperature, °C 


Fig. 2.51 — Hot-microhardness increment fraction for irradiated W in various conditions as 

a function of temperature. Specimens irradiated at ambient temperature (~70°C) 
in a water-moderated reactor to a fluence of 4.1 x 10^ n/cm^ (E p > 1 Mev). 
Recrystallized material annealed at 1760°C for 1 hour in H 2 prior to irradiation. 





Log VPN 


136 



Vickers Pyramid Hardness, kg/mm^ 














137 



Temperature, °C 


Fig. 2.54 — Hot microhardness of unirradiated and irradiated Nb — IZr 
as a function of temperature 


TENSILE TESTING (W. J. Stapp, A. R. Begany) 

Niobium and Nb - 1 Zr 

A study of the effect of annealing on the room-temperature and elevated-temperature 
tensile properties of irradiated (2.1 x 10 ^ n/cm^, E n ^ 1 Mev at ~ 70 °C) recrystallized 
niobium and Nb-lZr was conducted. Irradiated and control sheet specimens were post¬ 
irradiation annealed at temperatures determined by hot-hardness studies. The tempera¬ 
tures were selected to show most advantageously the room-temperature and elevated- 
temperature radiation-damage recovery behavior. 

The effect of post-irradiation anneals on the room-temperature engineering tensile 
properties of niobium appears in Figure 2.62 and Table 2.6. Recovery was complete 
following the 1000°C anneal. In the hot-hardness studies recovery was complete (Figure 
2.53) following the 900° C anneal. Comparison of the data shows that neutron irradiation 
increased ultimate strength less than yield strength. Irradiation drastically reduced the 
uniform elongation. The tensile behavior observed in the post-irradiation anneals is 
similar to that observed in irradiated molybdenum, 33 > 34 tungsten, 35 and niobium 36 in other 
studies, except that no evidence of lower yield points was observed. 

The elevated-temperature (300 °C and 650°C) tests were conducted after holding at test 
temperature for 1 hour, giving specimens, in effect, a 1-hour anneal at test temperature. 
The results (Figure 2.63 and Table 2.6) do not show any indication of thermal hardening, 
as might be suggested by an increase in the ratio of irradiated to unirradiated yield 
strength values at the elevated test temperatures. The dotted lines on the ultimate strength 
plots represent values determined from linear cross plots of hot hardness (DPH) versus 
ultimate tensile strength. Pronounced thermal hardening had been observed by Makin and 
Minter in room-temperature tensile tests following 200° to 300°C anneals for 1 hour. 





Microhardness increment fraction, AH 



Fig. 2.55 — Hot-mlcrohardness increment fraction for irradiated (2.1 x 10‘ 
n/cm^; E n > 1 Mev) Nb as a function of temperature 



0.20 


0.25 


Homologous temperature, 


0.30 

T°K 


Fig. 2.56 - Microhardness increment fraction, AHf, for W, Mo, 
and Nb compared on a homologous temperature 
basis 







139 



0 200 400 600 800 1000 1200 1400 

Temperature, °C 

Fig. 2.57 — Hot microhardness of unirradiated single-crystal molybdenum 
of orientations {l 10} and {00l} 

The present results do not agree with the 600 °C temperature shown to produce complete 
recovery in the earlier work. 36 The NMPQ material contained 130 ppm oxygen compared 
to 1600 ppm oxygen in the earlier tests. This difference and a higher neutron fluence 
level probably account for the higher recovery temperature. The higher strength and 
lower ductility levels in the specimens containing the higher oxygen level may have masked 
out the influence of radiation-induced defects on the mechanical properties. 

The effect of post-irradiation anneals on the conventional room-temperature tensile 
properties of Nb — IZr shown in Figure 2.64 and Table 2. 7 exhibit a type of thermal 
hardening in the 300° to 600°C range which agrees with hot-hardness results. Complete 
recovery of the radiation-induced defect structure occurs after the 1000°C anneal. Com¬ 
plete recovery did not occur until 1200°C in the hot-hardness studies. 

Nb — IZr exhibited greater strength in the irradiated condition up to a test temperature 
of 600°C (Figure 2.65). Reduced ductilities were observed in both irradiated and unirra¬ 
diated specimens at the 300°C and 600°C test temperatures. 

Molybdenum 

A study of the effect of post-irradiation annealing on the room-temperature tensile proper¬ 
ties of irradiated, recrystallized molybdenum was conducted and reported. 37 Data were ex¬ 
amined 38 in terms of the true stress versus true strain relationship to evaluate strain-hard¬ 
ening characteristics. The analysis employed the empirical power law expression: 

a = k (e) n (2.1) 

where: 

a = true stress 

7 = true strain (or natural strain) 
k = constant (strength coefficient at «T= 1) 
n = constant (strain-hardening exponent) 




{lOO } orientation 



{llO} orientation 

Fig. 2.58 — Photomicrographs of hot-microhardness (1200°C) indents on Mo 
single-crystals of two orientations (200X) 

It was concluded that the increase in strength of the irradiated molybdenum was due to both 
source and lattice defect mechanisms. 

The data were re-examined in terms of the exponential function 42 relating true stress (a) 
and true strain (e): 

= CT 00 " (a « " a 0 > eX P (-^C> 

where: 

<? Q = stress at which plastic deformation begins 
°oo = asymptotic stress attained after deformation 
c c = characteristic strain 




Ultimate tensile strength, kg/mm 2 Ultimate tensile strength, %, kg/mm : 


141 










✓ 

s 

_^_ 

\ 

\ 

\ 

\ 

\ 

\ 








s 

s 

Jf - 

\ 

\ 

\ 















z 

A 


Y' 

0 Alloy 

a 

256 (W - 30Mo 
emperature ran; 
u data obtained 
GEMP-497, Ai 

- 25Re) = 
ge ~ 25° to 13( 
from: P. N. Fla 
aril 14, 1967, p 

0.352 H v 
)0°C 
gel la, 

4. 





m 


□ 304 SS o = 0.395 H v 

Temperature range 20° to 1000°C 
o u data obtained from memo by 

P. N. Flagella, February 2, 1967. 

O Ta — 10W a u “ 0.301 H v 

Temperature range 20° to 1200°C 
o u data obtained from: "Fourth Annual 

Report - High-Temperature Materials 
and Reactor Component Development 
Programs, Vol. 1 — Materials," GE-NMPO, 
GEMP-334A, February 26,1965, p. 38. 

O Alloy 1541 (Fe-15Cr-4AI-1Y) 

% = 0.344 

o u data obtained from UKAEA (Harwell), 
ORNL-3970, and GE-NMPO; material from 
GE-NMPO Heat Ms-51. 

1 1 1 



m 




y 





j 

_i 


40 80 120 160 200 240 280 320 360 400 

Vickers Pyramid Hardness, kg/mm 2 


Fig. 2.59 — Ultimate tensile strength, o u , as a function of Vickers Pyramid Hardness, H v , 
for various alloys 



Fig. 2.60 - Ultimate tensile strength, a u , as a function of Vickers Pyramid Hardness, H y , for Nb 








142 



Fig. 2.61 — Ultimate tensile strength, <j u , as a function of Vickers Pyramid Hardness, H , for 
Nb-IZr 


Results (Table 2.8 and Figure 2.66) of the exponential function show that: 

1. a Q is increased by radiation-induced defects to a point of elastic fracture and decreased 
by post-irradiation annealing to the a Q value of the nonirradiated material. a Q may be 
considered the lattice friction stress at zero strain or slip propagation stress. 

2. Molybdenum will elongate and work-harden by the generation of dislocations. Size and/or 
number of radiation-induced defects are varied by annealing heat treatments, thereby 
affecting the characteristic strain, e c . 

3. The asymptotic stress is not significantly sensitive to the number or size of the radia¬ 
tion-induced defects. 

Hence it is again concluded that the radiation-induced increase in strength of recrystal¬ 
lized molybdenum is due to both the source-hardening and lattice-hardening defect 
mechanisms. 

Fracture areas of selected room-temperature, tensile-tested molybdenum specimens 
were examined to determine possible differences in microstructural flow and fracture 
characteristics between the unirradiated, irradiated, and post-irradiation annealed con¬ 
ditions. Since mixed modes of fracture were expected in some cases, it was believed that 
optical fractography would complement the metallographic examinations and provide de¬ 
tailed insight into the differences in modes of fracture between the various conditions. 
Specifically, examination of the actual fracture faces for evidence of dimple rupture fea¬ 
tures characteristic of the ductile mode of fracture was deemed expedient. The fracto- 
graphic studies were conducted at 500 diameters using dark-field illumination. Specimens 
in this study possessed irregular surfaces with varying amounts of localized necking on 
gage sections whose original dimensions were 0.5 by 6.3 mm. In general, irradiated speci¬ 
mens receiving a post-irradiation anneal lower than about 950°C displayed considerable 
embrittlement with little or no localized necking (Figure 2. 67). 



143 




_ 

' Annealed 1 hour at 2 x 10—7 Torr “ 

O 0 Irradiated to 2.1 x 10 20 n/cm 2 

(En>1 Mev) inGEFP-2-239 i 



— 

r 

racture 

Q 0 Unirrat 

, '• -I 

iated co 

ntrol 




h 

elongation 

















i Uniform 

elongation 





77: 



— 







_ ^ 

7 


p- 

— 

L 



/ 





•ractun 




—-- 

-o 


Uniforn 

=< 

-— 

— 

-- 






0 100 200 300 400 500 600 700 800 900 1000 

Annealing temperature, °C 


Fig. 2.62 — Room-temperature tensile properties of Nb sheet specimens 
versus post-irradiation annealing temperature 


TABLE 2.6 


NIOBIUM TENSILE TEST RESULTS 


Specimen 3 

No. 

Condition 

Post Irradiation 
Anneal Temperature, 
°C 

Test 

Temperature, 

°C 

Yield Strength 
0.2% Offset, 
kg/mm 2 

Tensile 

Strength, 

kg/mm 2 

Uniform Elongation 
in 2.48 cm, 

% 

Fracture Elongation 
in 2.48 cm, 

% 

1860 

Unirradiated 

None 

25 

13.2 

22.2 

24.4 

35.5 

1862 

Unirradiated 

300—1 hr—vac b 

300 

13.2 

20.2 

9.2 

13.4 

1861 

Unirradiated 

650-1 hr-vac b 

650 

6.9 

11.8 

9.0 

14.7 

1850 

Irradiated 0 

None 

25 

27.6 

28.2 

0.5 

18.6 

1848 

Irradiated 

300-1 hr—vac b 

300 

23.2 

23.2 

0.4 

5.9 

1849 

Irradiated 

650-1 hr-vac b 

650 

12.3 

12.3 

1.0 

12.6 

1859 

Irradiated 

300—1 hr—vac e 

25 

29.5 

31.1 

2.5 

15.9 

1852 

Irradiated 

650—1 hr—vac e 

25 

23.3 

29.0 

5.4 

13.5 

1854 

Irradiated 

1000-1 hr—vac e 

25 

13.8 

23.4 

28.3 

43.0 

1855 

Unirradiated 

1000-1 hr-vac 

25 

14.1 

21.6 

25.3 

35.3 


Specimens from 0.5-mm-thick sheet No. 80B792, vacuum annealed at 1150°C for 1 hour after fabrication. All tests conducted in 
vacuum ~8 x 10~® Torr. 

^Heating rate to test temperature 20°C/min f annealed for 1 hour prior to load application. Power turned off upon specimen fracture, 
irradiation test GEFP2-239 in ETR facility at reactor-ambient temperature (~70°C) to a fiuence of 2.1 x 10 2 ^ n/cm 2 (E n > 1 Mev) 
and 4.7 x 10 20 n/cm 2 (thermal). 

^Elastic fracture. 

e Annealed for 1 hour in a vacuum of 2 x Torr. 








144 




£ ' 
TO CM 

« E , 


a 30 






















o — 

— 

— « 

1 



-O 







_ Nb s 

heet spt 

LI 

"-O 





Irradiated to 2.1 x 10 20 n/cm 2 
(E n > 1 Mev) in GEFP-2-239 





“ 2.48-cm gage length 

Strain rate 8.3 x 10~ 5 sec“^ 

A_.--1 i u_ao-5-r 






□ N on-irradiated 

O Irradiated 







Tensile strength from 
hot hardness relationship 























0= 



d 



HJ 





0 tOO 200 300 400 500 600 700 

Test temperature, °C 


800 900 1000 






“ Strain rate 8.3 x 10 ° sec ' 
Annealed 1 hour at 2 x 10“®Torr 

n :_i:_. 






O Ir 

radiated 

























Tn 










z_ 











O— 



>TTT 


f 

A 

7^ 




300 400 500 600 700 

Annealing temperature, °C 


800 900 1000 


Fig. 2.63 — Elevated-temperature tensile properties 
of unirradiated and irradiated 
niobium sheet specimens 


Fig. 2.64 — Room-temperature tensile properties 
of unirradiated and irradiated 
Nb — 1 Zr sheet specimens 


TABLE 2.7 


Nb - IZr TENSILE TEST RESULTS 


Specimen 

No. a 

Condition 

Post Irradiation 
Anneal Temperature, 
°C 

Test 

Temperature, 

°C 

Yield Strength 
0.2% Offset, 
kg/mm 2 

Tensile 

Strength, 

kg/mm 2 

Uniform Elongation 
in 2.48 cm, 

% 

Fracture Elongation 
in 2.48 cm, 

% 

1726 

Unirradiated 

None 

25 

12.7 

24.9 

25.1 

33.5 

1729 

Unirradiated 

300—1 hr—vac b 

300 

8.3 

20.0 

10.9 

15.5 

1730 

Unirradiated 

600—1 hr—vac b 

600 

13.2 

21.8 

14.1 

21.0 

1721 

Irradiated 

None 

25 

38.7 

38.7 

0.2 

0.4 

1723 

irradiated 

300-1 hr—vac b 

300 

27.4 

31.3 

1.6 

1.6 

1727 

irradiated 

600—1 hr—vac b 

600 

27.8 

31.5 

1.5 

2.6 

1725 

Irradiated 

300—1 hr—vac e 

25 

38.7 

38.6 

0.7 

4.9 

1722 

Irradiated 

600-1 hr—vac e 

25 

40.0 

40.5 

0.9 

8.0 

1719 

Irradiated 

1000-1 hr—vac e 

25 

14.0 

26.4 

22.4 

33.4 

1724 

Unirradiated 

1000-1 hr—vac e 

25 

13.9 

26.6 

26.2 

34.6 


Specimens from 0.5-mm-thick sheet (heat No. 27605), vacuum annealed at 1200°C for 1 hour after fabrication. All tests conducted 
in vacuum ~8 x 10“^ Torr. 

b Heating rate to test temperature 20°C/min, annealed for 1 hour prior to load application. Power turned off upon specimen fracture, 
irradiation test GEFP2-239 in ETR facility at reactor-ambient temperature (~70°C) to a fluence of 2.1 x 10 20 n/cm 2 (E p > 1 Mev) 
and 4.7 x 10 20 n/cm 2 (thermal), 
dElastic fracture. 

e Annealed for 1 hour in a vacuum of 2 x 10~ 7 Torr. 








145 




(E n > 1 Mev) in GEFP-2-239 



Fig. 2.65 — Elevated-temperature tensile properties of unirradiated and irradiated 
Nb — IZr sheet specimens 


Metallographic examination of the unirradiated specimens showed failure to be entirely 
transgranular (Figure 2.68) without any trace of intergranular fracture. A moderate 
amount of plastic flow (necking) in the fracture area had occured. Fractographic study of 
these fractures indicated a condition of quasi-cleavage; i. e., the fracture displayed charac¬ 
teristics of feathery cleavage (little or no plastic deformation) while possessing some plas¬ 
tic deformation features such as tear ridges and stretching. 

The as-irradiated specimen exhibited mixed mode of fracture with intergranularity 
slightly in preponderence. No evidence of any appreciable wedging or ductility was observed. 

The fractographic study of the post-irradiated annealed specimens showed that fracture 
occurred predominately by cleavage. Specimens annealed at 1025 °C and higher tended to 
exhibit quasi-cleavage. No post-irradiated specimens exhibited evidence of ductile dimple 
rupture. Specimens annealed at 950°C and lower displayed mixed modes of fracture. In 
general, lowering the annealing temperature resulted in increased evidence of intergranu¬ 
lar fracturing. 

Although these observations were substantiated during metallographic examination, a pos¬ 
sible exception to this was the following microstructural anomaly: the secondary cracking 
condition possessed by the 830 °C post-irradiation annealed specimen was predominantly 
intergranular, whereas the secondary cracking within the 700 °C annealed specimen was pre¬ 
dominantly transgranular (see Figure 2.69). 




146 


00 

CN 

LU 

-I 

00 

2 


a -6 

u <u £ 

| .E | 
f® <o E 


D 

L. 

H 


s n 

CO 

o 

00 

to 

i— 


»-r £ 

o 

CN 

q 

CO 

«— 


E $ 'p- 


cvi 

»- 

o 


toi 

V O O) 
$ * * 

CO 

p> 

r-. 

r> 

CD 

co 

< CO 







~0 







to 







m ^CM 







:.E£ 

CO 


LO 

r- 

o 

00 

% £ £ 

to 

CN 

*“ 

lo 

o 




ai 

csi 


cd 

T— 

4-> CO CD 

w o * 

CO 

CO 

LO 


CO 

CO 


N 

l! § 

s |S* 

i o« 

LU .E 


§■2 5 
£ S.S SS 

c § ^ 

^ rn c 


4= C 

g | 

si 

03 ^ 


-o s> 3= t 

« g ° E 

>- £ 3? o> 
CO CM •* 
O 


fi te o 

H Q.° 

E 


^ i 2 
2 c 2 O 

M &° 

S .1 


E ro . 
*5 o 


O r- O CN 00 O 

:gg§°g 

o' o' o o o o“ 


O CN 

co 


O) 

00 

o ^ 

O)’ 

C\I 

03 




00 


rt 


00 h» (D O) fO 

od o oo co cd to 


CN CO »— q CO 03 
O «— 00 CD CN 00 

to n to in in Tt 


CO <D Xfr CO ^ P- 
^ cd ^ 

CO CD rf CO CO 


O O O O O 

.2 ■§ "g "g 1g "g 


C ll 

D ± 






S = 

03 

CN 

03 

CD 


E 2 

00 


O 

O 

CN 

,E x: 
8 g 
&o 

<0 43 

5 

CN 

CN 

CN 

CN 

X 

o° 


8 


- 2 


g a 


8 


ai 

8 *5 2 

|*l 

51 i 

C »*- 4-* 


§.sf 

i 

.—. CO ,* 

g^o 

CD C O 

<9 ® g 

O E »- 

5 ■? 5 

® O *- 

"S A C 

^ 2 LU 

S> C C 
o OO 
- 0>CN 

o 


(O ° 

f ?£ 

Hi 

in .E -p 

o w ® 

E re 2 

a) ^ 

2 i2 

TO 



147 



Fig. 2.66 — Room-temperature stress versus strain curves for irradiated (ORM-38) 
molybdenum post-irradiation annealed at indicated temperature 

RESISTIVITY STUDIES (L. K. Keys, J, Moteff) 

Molybdenum 

The isochronal resistivity recovery of fast neutron (E ^ 1 Mev) radiation-induced defects 
in recrystallized molybdenum specimens has continued. 39 Six specimens were irradiated at 
reactor-ambient temperature (~70°C) to fast neutron fluences ranging from 1.1 x lO 1 ^ to 
lo5 x 10 2 * n/cm 2 (E n ^ 1 Mev). The various samples used and the pertinent irradiation con¬ 
ditions are shown in Table 2.9. The dependence of the radiation-induced resistivity incre¬ 
ment, Ap ? on the fast neutron fluence (for NMPO studies as well as those of Peacock and 
Johnson 40 ) is presented in Figure 2.70. The resistivity increment, Ap, is defined as Ap = 
p x - p 0 , wherePj is the as-irradiated resistivity andp Q is the pre-irradiation resistivity. The 
resistivity increment reaches a maximum between 5 x 10*9 n/cm^ and 2 x 10 2 ^ n/cm 2 , 
then decreases by about a factor of 2 at 1.5 x 10 2 ^ n/cm 2 . This result is somewhat surpris- 
ing in view of the asymptotic approach to saturation considered in other investigations. » 

The slope of the linear portion of the curve (a slope of about 0. 58) is in good agreement 
with that reported by Peacock and Johnson (0.54), although they observed what appeared 
to be saturation effects at about 3 to 4 x 10^® n/cm 2 . Results of NMPO studies and those 
of Kissinger, Brimhall, and Mastel yield saturation fluences much higher than those of 
Peacock and Johnson. Kissinger et al. observed saturation below NMPO saturation 
fluences. These differences result from different techniques of observation (X-ray para¬ 
meter and length changes), as will be shown. 

The recovery of the irradiation-induced defects in molybdenum is presented in Figure 
2.71, in which the radiation-induced resistivity, Ap, is plotted versus annealing tempera- 




148 



Fig. 2.67 — Photomicrograph of fracture area of room- 
temperature tensile-tested Mo sheet specimen after 
irradiating to 1.1 x 10^0 n/cm^ (E n > 1 Mev). 
Specimen given an 830°C anneal for 1 hour prior 
to testing. (Neg. R-2227, 100X) 



Fig. 2.68 — Photomicrograph of fracture area of unirradiated 
room-temperature tensile-tested Mo sheet specimen 
(Neg. R-2236, 500X) 


149 



Fig. 2.69 — Photomicrograph of fracture area of room- 
temperature tensile-tested Mo specimen showing 
predominantly transgranular secondary cracking. 
Irradiated (1.1 x 10 20 n/cm 2 , E n > 1 Mev) speci¬ 
men given a 700°C anneal for 1 hour prior to 
testing. (Neg. R-2257, 100X) 


TABLE 2.9 

PRINCIPLE DEFECT RECOVERY STAGES IN NEUTRON-IRRADIATED MOLYBDENUM 
AS DETERMINED BY ISOCHRONAL RESISTIVITY STUDIES _ 


Peak Temperature of Recovery Spectra, and 
Irradiation Condition 3 _Area Under Respective Recovery Peak b 


Sample 

No. 

Rod 

Capsule 

Fluence, nvt 

Fast 

(E n > 1 Mev) Thermal 

0.15 T m . 
°C 

(Apm) 

Area, 

micro-ohm-cm 

0.31 T m , 
°C 

(Apiy) 

Area, 

micro-ohm-cm 

(Ap III/ Apjy) 
Area, 

micro-ohm-cm 

520 

C 

ORM-14 

1.1 x 10 19 

9.3 x 10 19 

153 

0.396 

624 

0.163 

2.43 

527 

C 

ORM-18 

2.3 x 10 19 

1.1 x 10 20 

163 

0.476 

620 

0.227 

2.10 

2277 

M 

ORM-54 

5.8 x 10 19 

4.1 x 10 20 

150 

0.693 

616 

0.406 

1.71 

909 

L 

MT-138 

9.8 x 10 19 

3.7 x 10 20 

160 

0.608 

619 

0.536 

1.13 

923 

L 

ORM-38 

1.1 x 10 20 

5.2 x 10 20 

157 

0.701 

620 

0.567 

1.24 

914 

L 

MT 2-234 

1.5 x 10 21 

2.6 x 10 21 

- 

0 

606 

0.596 

0 


Specimens irradiated at reactor-ambient temperatures in a water-moderated reactor {ORR or ETR). 
bpeak temperature of recovery refers to temperature of maxima in the derivative of the recovery curves. This recovery 
maximun is related to the melting temperature, T m , as shown. 



150 



Fig. 2.70 — Total radiation-induced resistivity increment for Mo as a function of 
fast neutron fluence 

ture. The recovery of radiation-induced defects after reactor-ambient (~70°C) irradiations 
resembles that for tungsten. 43 ’ 44 Two principal recovery regions are evident in all but the 
highest fluence specimen. These recovery regions are more distinct in Figure 2.72, in 
which the derivative d(Ap/Ap Q )/dT of the normalized resistivity increment is plotted versus 
annealing temperature. The two principle recovery regions centered at about 160°C and 
620 °C, respectively, occur at almost exactly 0.15 T m (stage III) and 0.31 T m (stage IV) 
and agree quite well with the recovery observed by Ibragimov et al. 45 after a 2 x 10^° 
n/cmr irradiation with fast neutrons. As for tungsten, 43 the lower temperature recovery 
region (0.15 T m ) is believed to represent the migration and recovery of self-interstitials, 
and the higher-temperature recovery region (0. 31 T m ) represents the recovery of vacan¬ 
cies. (The term interstitial will be considered, for the remainder of this discussion, to 
mean self-interstitial unless specified otherwise.) Other smaller recovery regions are 
evident in some of the specimens, similar to tungsten. 46 The reproducibility of these 
smaller peaks is uncertain in NMPO studies and in the studies of others; 40 ? 45 ? 47 their 
importance should not be overlooked, however, since the defects responsible for these 
peaks apparently contribute to the thermal hardening in this recovery region. 48 In Figure 
2.73 the resistivity recovered in stage in (Apjjj) and stage IV (A pjy) is presented. The 
stage ID resistivity recovery completely saturates over this range, dropping to zero at the 
highest fluence. The stage IV recovery, however, shows a linear fluence dependence up to 
about 1020 n/cm 2 , where a normal asymptotic approach to saturation is evident. These 
results are in relatively good agreement with those of Kissinger, Brimhall, and Mastel 42 
who have investigated, by lattice parameter and length change studies, the fast-neutron, 
radiation-induced fluence dependence in molybdenum over a range of 1C)19 to 10^0 n/cm . 

The results of the length change study indicate that vacancies saturate at about lO 2 ^n/cm^ 
or above, and that interstitials saturate at approximately 2 to 5 x 10*9 n/cm^. The ratio of 
the stage HI resistivity (Ap m ) recovery to the stage IV recovery (Apjy), presented in Fig¬ 
ure 2.74, shows a linear dependence on the fast neutron fluence, on a semilog plot, from 
which intermediate values of this ratio can be predicted. Results from a resistivity recovery 
study of molybdenum by Ibragimov, et al. 45 after fast neutron irradiation to 2 x lO 2 ^ n/cm 2 , 
agree quite well with the prediction from Figure 2.74 in that a Apjjj/Apjy ratio of almost 
one is obtained for approximately 2 x lO 2 ^ n/cm 2 . 

These results are all consistent with the formation of increasing number of interstitial 
clusters (i. e., increase in the ratio of the number of interstitials in clusters to the number 
of free interstitials) until a fast fluence of 10 21 n/cm 2 is reached. There are practically 
no free interstitials remaining above this neutron fluence level. 






































151 



3 0 96L— i B uuo-Luqo-ojojUJ 'c/y ^ustuajoui AjjAijsisay 


Fig. 2.71 — Isochronal resistivity recovery of neutron-irradiated, recrystallized molybdenum 
as a function of annealing temperature 

















153 



Fig. 2.73 - Resistivity recovery increments of stage III (Apj j ( ) to stage IV 
(Apjy) as a function of fast neutron fluence 



Fast neutron fluence, n/cm^ (E n > 1 Mev) 


Fig. 2.74 — Ratio of resistivity recovered in stage III to stage IV as a function of fast neutron fluence 

The major differences in the recovery behavior observed between tungsten and molybdenum 
undoubtedly arise from the differences in the homologous temperature of irradiation. 49 For 
tungsten, the irradiation temperature of approximately 70°C corresponds to about 0.093 T m , 
where T m is the absolute melting temperature. For molybdenum, however, the irradiation 
temperature of 70 °C represents about 0.12 T m . This represents an irradiation temperature 
corresponding to the onset of the stage III region in molybdenum and thus considerable re¬ 
covery occurs during the irradiation. Recovery actually appears to begin at around O°C. 50j51 
The 0. 09 T m irradiation temperature for tungsten is well below the recovery region for 






154 


stage HI; 52 hence from the recovery point of view numerous point defects would be expected. 
Apparently irradiation within a recovery stage generates a larger ratio of cluster defects/ 
point defects for the defects recovering in that recovery stage. This explanation is support¬ 
ed by the facts that the radiation-induced resistivity values observed for molybdenum are 
lower than for tungsten at the same radiation fluence, and that the slope observed for the 
neutron fluence dependence of molybdenum 53 is smaller than for tungsten. 


The stage m recovery mechanism, considered to be the migration of point defects (inter¬ 
stitials), can also be compared to other metals with respect to the number of jumps such a 
migrating defect can make. It has been shown that a defect migrating with an activation 
energy Q, makes n jumps in a lifetime t before annihilation, with a vibrational frequency 
(v) of approximately 10 14 sec -1 , at a temperature T according to: 54 


, n Q/kT 
t = — e 
v 


( 2 . 2 ) 


Rearranging this formula, for an isochronal anneal of a time t at a temperature T, a 
point defect can make n jumps according to: 


n = t v e 


-Q/kT 


(2.3) 


Thus, for a 1-hour isochronal anneal, considering Q ~ 1.3 ev 55 at 160°C, about 300 jumps 
can be made. This parallels the number of jumps which have been found in stage III for 
several metals. 54 


TRANSMISSION ELECTRON MICROSCOPY (R. C. Rau) 

Microstructural studies using transmission electron microscopy were performed on 
neutron-irradiated tungsten and molybdenum. The objective of this work was to charac¬ 
terize radiation-induced microstructural features and, if possible, to relate those features 
to mechanical and physical properties. These studies concerned (1) characterization of 
dislocation loops in irradiated and annealed tungsten, (2) annealing of defect clusters in 
irradiated tungsten, and (3) effect of high irradiation temperatures on molybdenum. 

Tungsten 

Characterization of Dislocation Loops — Previous work 56 ? 57 has sh6wn that post-irra¬ 
diation annealing at temperatures near 1100 °C produces resolvable dislocation loops in 
tungsten irradiated to fast neutron fluences of about 4 x 10 19 n/cm 2 (E n > 1 Mev) or 
greater. It had been postulated that the loops in tungsten were probably interstitial, based 
mainly on results in irradiated and annealed molybdenum, 58 ~ 60 but no diffraction contrast 
analyses had been carried out. To crystallographically characterize the loops in tungsten, 
tilting experiments in the electron microscope were performed on a specimen which had 
been irradiated to 1.5 x 10 2 * n/cm 2 at reactor-ambient temperature (~70°C), and sub¬ 
sequently creep-rupture tested (i. e., annealed) for 315 hours at 1100°C. 


Thin foil specimens were prepared electrolytically 61 from an unstressed button head of 
the tested specimen. Preliminary scanning showed the microstructure to consist of well 
formed dislocation networks, making up subgrain boundaries, and a population of large 
dislocation loops averaging about 1000 A in diameter. 57 Because of the polycrystalline 
nature of the specimens and the apparent lack of preferred orientation, grains sampled 
in the thin foils had random orientations. Hence it was necessary to scan many areas to 
find grains in reasonably simple orientations from which meaningful tilting experiments 
could be performed. Figure 2.75 shows six micrographs of an extensive series obtained 
from an especially suitable area. The orientation of this grain was near (001), and tilting 
was carried out about axes near [100], [110], and [110]. The micrographs shown in Figure 




155 








156 


2.75 illustrate the complete characterization of the loops, i.e., their Burgers vectors, 
habit planes, and nature. 

Burgers vectors of the loops were determined by noting diffraction conditions for which 
the loop images were at extinction; i. e., g. b = 0. 62 For example, loops labeled A and B 
are out of contrast in Figure 2.75a with g = [lTO]; loops labeled C and D are out of contrast 
in Figure 2.75b with g = [110]. In Figure 2.75c and d, with g = ± [231], only B loops are 
out of contrast; in Figure 2.75f, with g = [121], C loops are at extinction. These conditions 
unambiguously identify the Burgers vector of B loops as ± 1/2 [111] and that of C loops as 
± 1/2 [111]. Similar analyses identified the Burgers vector of A loops as ± 1/2 [111] and 
that of D loops as ± 1/2 [Ill]. The four sets of loops and their diffraction conditions in 
Figure 2.75 are summarized in Table 2.10. 

Habit planes of the loops were determined by observing the manner in which their image 
shapes changed during tilting about known crystallographic directions, assuming their 
true shape to be circular. These observations, compared with observations of shape 
changes of circles drawn on a transparent tetrahedron, indicated that the loops lay on or 

TABLE 2.10 


MAGNITUDES OF g • b FOR OPERATING REFLECTIONS 
IN FIGURE 2.75 


Loop 

\ 9 

b \ 

[110] 

[110] 

[231] 

[231] 

[200] 

[121] 

A 

±% tun 

0 

± 1 

± 1 

+ 1 

±1 

±2 

B 

±'A [111] 

0 

± 1 

0 

0 

± 1 

± 1 

C 

±’/2 [iTl] 

± 1 

0 

+ 2 

±2 

± 1 

0 

D 

±'A [Til] 

+1 

0 

±3 

+ 3 

+ 1 

± 1 


near {111} planes, with each loop lying on the particular {111} plane normal to its Burgers 
vector. Thus the loops are in edge orientation. The relative orientations of the four sets 
of {111} loops, both with respect to each other and with respect to the micrographs of 
Figure 2. 75, are shown in the stereoscopic drawing of Figure 2.76. 

The nature of the loops was determined by the formal method of analysis of the image 
contrast as the sign of the quantity (g.b)s was changed. 63 Using standard conventions to 
define the geometry of dislocations and their Burgers vectors, 64 the nature of the A, C, 
and D loops can be deduced from Figures 2.75c and d. For example, from the identifica¬ 
tion of the habit planes, it is known that C loops slope downward to the right in the micro¬ 
graphs. In Figure 2. 75c these loops have outside contrast, indicating that (g.b)s is positive; 
hence the positive edge component of the Burgers vector has the same directional sense 
as the g vector. This situation obeys a right-hand rule, which indicates that the C loops 
are vacancy in nature. 63 Similar analysis of the contrast of the other loops indicates that 
they are all vacancy type. These results are summarized in Table 2.11. 

Annealing of Defect Clusters in Irradiated Tungsten — To study the progressive effects 
of post-irradiation annealing on the defect structure of neutron-irradiated tungsten, a 
series of specimens were examined which had been irradiated to 4.2 x 10 19 n/cm 2 
(® n — 1 Mev) at reactor-ambient temperature (~ 70°C). Following irradiation, the speci¬ 
mens received 1-hour anneals in argon at various temperatures and were tensile tested 
at 240°C. 35 The annealing temperatures and test results for both irradiated and unirra¬ 
diated control samples are listed in Table 2.12. 

Thin foils for transmission electron microscopy were prepared from both the unstressed 
button heads and the reduced areas near the fracture. Examination of foils from the button 
heads of unirradiated control specimens revealed that the starting microstructure contained 



Fig. 2.76 — Stereoscopic drawing showing the appearance of loops on the four 
sets of {l 11} planes. The tetrahedron is oriented in approximate 
agreement with the foil orientation of Figure 2.75. 


TABLE 2.11 

ANALYSIS OF THE NATURE OF LOOPS IN FIGURES 2.75c AND 2.75d 
Possible 9“ [231], s + g=[23f],s + 


Loop 

b 

9 * b 

Image 

g-b 

Image 

Loop Nature 

A 

±% [mi 

± 1 

inside 

g"b<0 

+ 1 

outside 
g ■ b > 0 

vacancy 

TT= y 2 HTT] 

B 

±% [1 IT] 

0 

- 

0 

- 

vacancy 3 

b = y 2 HTl] 

C 

±% [1T1] 

+ 2 

outside 
g- b > 0 

±2 

inside 
g * b < 0 

vacancy 

b = % [TlTl 

D 

±% [Til] 

+ 3 

inside 
g ‘ b < 0 

+ 3 

outside 

g-b>0 

vacancy 

b = % [ITT] 


a Nature of B loops determined from other micrographs not shown in Figure 2.75 


TABLE 2.12 


MECHANICAL PROPERTIES AND DEFECT CLUSTER DATA FOR 
IRRADIATED AND ANNEALED TUNGSTEN 


Specimen 

No. 

Fast Neutron 
Fluence , a 

n/cm 2 <E n > 1 Mev) 

Post-Irradiation 

Annealing 

Temperature, 

°c t/t m 

Ultimate 

Strength, 

kg/mnn^ 

Elongation 
in 3.18 cm, 

% 

Reduction 
in Area, 

% 

Cluster 

Size, 

a 

Cluster 

Density, 

No./cm 9 

10W 

Unirradiated 

240 b 

0.14 b 

49.9 

37.7 

55.2 

— 

— 

7W 

4.2 x 10 19 

240 b 

0.14 b 

66.4 

0.3 

8.6 

35 

1.0 x 10 17 

16W 

Unirradiated 

435 

0.19 

49.3 

34.6 

64.7 

— 

- 

11W 

4.2 x 10 19 

435 

0.19 

61.7 

1.1 

8.2 

35 

2.2 x 10 16 

26W 

Unirradiated 

743 

0.28 

50.7 

21.9 

33.6 

- 

- 

23W 

4.2x 10 19 

743 

0.28 

77.6 

0.5 

1.8 

50 

8.7 x 10 16 

24W 

4.2 x 10 19 

897 

0.32 

60.0 

0.6 

2/1 

50 

1.1 x 10 16 

25W 

4.2 x 10 19 

1043 

0.36 

52.3 

12.0 

12.3 

100 

3.5 x 10 14 


a Irradiated at reactor-ambient temperature (~70°C). 
^Tensile test temperature. 



158 


numerous dislocation tangles surrounding relatively dislocation-free cells, typical of worked 
and incompletely recrystallized metals. It was noted during subsequent examination of the 
irradiated specimen annealed at the highest temperature, 1043 °C, that these random tangles 
were partly replaced by more ordered networks of dislocations such as occur in well annealed 
and recrystallized metals. 65 

Examination of the irradiated specimens showed the presence of radiation-induced black- 
dot defect clusters and the effect of post-irradiation annealing on those clusters. Figure 
2.77 shows photomicrographs illustrating typical microstructures in button head foils. In 
the unannealed condition (Figure 2.77a) a dispersion of fine dot clusters was present 
throughout the matrix. Annealing at 435 °C (Figure 2. 77b) produced an apparent slight de¬ 
crease in cluster concentration, but annealing at 735 °C (Figure 2.77c) resulted in a notice¬ 
able densification of the clusters. Finally, annealing at 897 °C (Figure 2.77d) and 1043 °C 
(Figure 2.77e) caused a definite decrease in cluster concentration and led to the formation 
of resolvable dislocation loops. 

To obtain a quantitative comparison between defect cluster densities and tensile strengths 
of the specimens, cluster counts were made from the electron micrographs of Figure 2.77. 66 
The resulting cluster densities listed in Table 2.12 are plotted as a function of annealing 
temperature in Figure 2.78a. This plot bears a striking resemblance to a plot of the tensile 
strengths of these samples shown in Figure 2.78b. These curves and the data in Table 2.12 
show that annealing produces a minimum in the strength curve at 435 °C which corresponds 
to a depletion of defect clusters; a maximum occurs at 743 °C which corresponds to a pro¬ 
nounced coarsening and densification of the clusters. Higher-temperature annealing results 
in a loss of strength and a reduction in the population of defect clusters, in good agreement 
with previous findings. 56 > 57 

This complex annealing behavior is believed to be caused by two overlapping defect 
recovery peaks, one occurring below 435 °C and one above. The observed cluster density 
curve and hence the strength curve represent the sum of the two separate recovery curves. 
These separate recovery curves are due to the diffusion and agglomeration of two different 
defect species which require different thermal energies for migration. It is significant that 
the 435 °C minimum occurs somewhat beyond the 0.15 T m (stage m) recovery peak observed 
in electrical resistivity studies; the 743 °C maximum occurs just below the 0. 31 T m (stage 
IV) recovery peak. 67 ? 68 Both the 897°C and 1043°C annealing temperatures, which produce 
marked recovery in the observable microstructures, are above the 0.31 T m recovery peak. 

The assignment of specific migrating defects to the stage HI and stage IV recovery re¬ 
gions has been the subject of much controversy in the literature, primarily because of the 
indirect measurement techniques involved. Recent direct observations on annealing in 
irradiated tungsten by field ion microscopy have provided conclusive evidence that stage HI 
corresponds to the migration of free interstitials and stage IV corresponds to the migration 
of vacancies. 69 ’ 71 On this basis the defect clusters seen in the unannealed specimen in the 
present study are probably interstitial clusters, and those seen in the specimens annealed 
at the higher temperatures are probably vacancies. The clusters in the 435 °C annealed 
specimen might be remnants of the early interstitial cluster population, the beginnings of 
the buildup in vacancy cluster population, or a combination of both. 

To check these conclusions, it would be desirable to carry out electron diffraction con¬ 
trast experiments on the different specimens to identify the clusters present after each 
annealing temperature. Such experiments cannot be performed reliably, however, when 
the cluster density is high or when the clusters are not resolvable as dislocation loops. 

Hence in the present case only, the specimen annealed at the highest temperature, 1043 °C, 
was suitable for analysis. Large-angle tilting experiments were carried out on a foil from 
this specimen following the method of Edmondson and Williamson. 72 Figure 2.79 shows the 



159 



Fig. 2.77 - Transmission electron micrographs used for defect cluster 
counting in button heads of irradiated and annealed W. 

(a) Unannealed, (b) Annealed at 435°C. (c) Annealed at 
743°C. (d) Annealed at 897°C. (e) Annealed at 1043°C. 



(a) 


O Irradiated 
□ Unirradiated 


Annealing temperature, °C Annealing temperature, °C 

Fig. 2.78 - Effect of post-irradiation annealing temperature on defect cluster density, and ultimate 
tensile strength at 240°C of polycrystalline W irradiated at ~70°C to 4.2 x 10^ n/cm^ 
(E n > 1 Mev) 








W 






¥. 




p 








Fig. 2.79 — Series of electron micrographs used for identifying loops in tungsten irradiated to 
4.2 x lO^n/cm^ at 70°C and anneajed a^1043°C. All micrographs taken with 
s > o. (a) g = 112JLtilt +18° 40'; (b) g = [200], tilt +2° 40'; (c) g = [200],tilt 
+0° 50'; (d) g = [110],tilt -6° 20'; (e) g = [110],tilt -8° 00'; (f) g = [30T],tilt 













161 


resulting micrographs. The figure shows that loops labeled A increase in apparent size 
due to purely geometrical factors as the foil is continuously tilted through a relatively 
large angle from +18° 40 T to -17° 50 T , but the loop labeled B decreases in apparent size. 
But when the diffraction conditions are changed by reversing the sign of the quantity (g.b)s 
(i.e., by reversing g from [200] to [200] or from [110] to [110]), A loop images shrink 
and the B loop image expands. Thus the apparent size changes due to diffraction conditions 
are in the opposite sense to those due to geometrical factors for both A and B loops, and 
both are vacancy in nature . 72 This analysis agrees with the previous, more rigorous 
analysis of large loops in tungsten annealed at 1100°C (Figure 2.75), and supports the 
contention that vacancies migrate in stage IV. 

In addition to specimens from the button heads, foils prepared from the gage sections 
near the fracture were also examined in the electron microscope. Detailed cluster counts 
were not made in the gage section foils, but it was apparent that the microstructures were 
generally similar to those in the button heads shown in Figure 2 . 77 . 66 The sample annealed 
at 435 °C was an exception; the observed cluster density in the gage section was noticeably 
lower than in the button head. This implies that the small clusters present after the 435°C 
anneal interact with and are possibly swept out by the dislocations moving along slip planes 
during tensile testing . 73 

Molybdenum 

Defect Structures - Transmission electron microscopy studies were begun on a series 
of polycrystalline molybdenum specimens irradiated at three different temperatures and 
creep-rupture tested at 750°C. 74 This test temperature was selected on the basis of micro- 
structural observations by Mastel and Brimhall , 75 who found that post-irradiation anneal¬ 
ing at temperatures above 750°C led to a rapid decrease in defect cluster densities in 
molybdenum which had been irradiated at reactor-ambient temperatures. The specimens 
used in the present investigation, and their irradiation conditions and test results, are 
listed in Table 2.13. 


TABLE 2.13 


CREEP-RUPTURE DATA 3 FOR MOLYBDENUM TESTED 
AT 750°C AND 18.00 kg/mm 2 


Specimen 

No. 

Fast Neutron 
Fluence, 

n/cm 2 (E n > 1 Mev) 

Irradiation 

Temperature, 

°c 

Rupture 

Life, 

hr 

Elongation 
in 4.45 cm, 

% 

Reduction 
in Area, 

% 

1610 

Unirradiated 

_ 

20.38 

39.5 

95.7 

1594 

1.4 X 10 20 

70 

56.06 

40.4 

92.8 

1622 

1.8 x 10 20 

700 

244.06 

37.6 

94.4 

1612 

1.8 x 10 20 

1000 

355.09 

31.1 

95.2 


a See Table 2.3 for complete history of specimens. 


Thin foils were prepared electrolytically, using a mixture of 12. 5 percent H 2 SO 4 in 
methanol, from the unstressed button heads and the stressed regions near the fracture 
of the tested specimens. Examination of foils from the button head of the unirradiated 
control specimen showed a relatively defect-free microstructure containing only a few 
random dislocations. Stress at 750°C and the resulting deformation led to a microstruc¬ 
ture of tangled networks of dislocations outlining subgrain boundaries, and free disloca¬ 
tions within the subgrains; these were revealed by foils from the gage section of this 
specimen. 

Examination of button head foils from the specimen irradiated at pile-ambient temper¬ 
ature (~70°C) and tested at 750°C showed an abundance of resolvable dislocation loops, 
many quite large (up to 4000A), shown in Figure 2.80a. These loops were uniformly distributed 





163 


throughout the grains, and showed no tendency of denuding near grain boundaries. Pre¬ 
liminary diffraction contrast experiments indicated that the majority of the loops probably 
had a/2 <111 > Burgers vectors, similar to the large loops in tungsten and in agreement 
with previous work on irradiated molybdenum. 58-60 Large-angle tilting experiments, 72 also 
preliminary, indicated that the loops were interstitial in nature. This result is consistent 

7 SR— fiO 

with other work on irradiated molybdenum. 


The microstructure of button head foils from the specimen irradiated at 700°C and 
tested at 750°C also contained a dense population of large dislocation loops within the 
grains (Figure 2. 80b), but in this case a zone denuded of loops was present along the 
grain boundaries. The width of this denuded zone was about 1 micron, in agreement with 
recent observations by Brimhall et al. 76 on molybdenum irradiated to 3. 5 x 10^ n/cm^ 
(E n — 1 Mev) at 600°C. Preliminary tilting experiments indicated that the loops were 
again interstitial in nature, and probably were in edge orientation, lying on {ill}habit 
planes and having a/2 <111> Burgers vectors. 


Examination of foils from the button head of the specimen irradiated at 1000°C and 
tested at 750°C showed a completely different microstructure than in the other two con¬ 
ditions. This microstructure (Figure 2.80c) was essentially the reverse of that seen in 
the 700°C irradiated specimen, and consisted of an almost structureless matrix contain¬ 
ing a scattering of small black dots together with a few scattered colonies of loops located 
in regions within 1 micron of the grain boundaries. These loops, which averaged about 
500 to 1000 A in diameter, were tentatively identified as vacancy in nature. This result 
is consistent with the recent findings of Brimhall et al. who reported that 1000 C anneal¬ 
ing of molybdenum previously irradiated at 600 °C led to the nucleation and growth of 
vacancy clusters within the 1-micron-wide denuded zone along grain boundaries. 


Examination of the foils from the stressed regions in the gage sections of the molybdenum 
specimens irradiated at 70 °C and 700 °C showed dislocation tangles and networks outlining 
subgrain boundaries, typical of deformed metals. Gage section foils from the 1000°C 
irradiated specimen showed similar dislocation structures and evidence of pinning of 
dislocations by the small black dots in the matrix. Several such pinned dislocations can 
be seen bowing out from black dots in the micrograph of Figure 2.81. The identity of these 
black dots is not certain; they are not believed to be clusters originating from displace¬ 
ment events, but rather small carbide particles which precipitated during the 1000°C 
irradiation. The carbon content of this series of specimens was approximately 220 ppm, 
well in excess of the equilibrium solid solubility limit in molybdenum. The 1000°C 
irradiation temperature may cause precipitation of carbides from solution. Dislocation 
pinning by these particles probably accounts for the enhanced strength indicated for this 
specimen in Table 2.13. 

Large dislocation loops were scattered throughout the matrix in gage section foils 
from the specimen irradiated at 1000°C and tested at 750°C. A fairly complete crystallo¬ 
graphic characterization of these loops was made from a series of micrographs, two of 
which are shown in Figure 2.82. This diffraction contrast analysis indicated that most of 
the loops were edge loops, lying on{lll)planes and having a/2 <111> Burgers vectors, 
but that they were interstitial in nature. This is in contrast to the tentative identification 
of vacancy loops in button head foils from this same specimen. 

Finally, one rather unusual dislocation loop was found in the gage section of the 1000 °C 
irradiated specimen. That loop is marked A in Figure 2.82a, taken with the (211) reflec¬ 
tion operating. This loop is at extinction in Figure 2.82b, taken with the (002) reflection 
operating; i.e., g. b = 0. This behavior indicates that loop A cannot have a Burgers 
vector of the type a/2 < 111> commonly found in bcc metals, but suggests that it must 



Fig. 2.82 — Dislocation loops in stressed region of molybdenum irradiate^at 1000°C and 
tested at 750^C. Loop marked A is non-a/2 <111> type, (a) g = 1211|, tilt 
+9° 40'; (b>"g = |002|, tilt -5° 40'. 





r f f 


165 


have a higher-energy Burgers vector such as a < 100 > or a <110 ^ Although the available 
diffraction conditions were not sufficient to differentiate between these two possibilities, 
dislocation energy considerations strongly favor the a <100> choice. 77 Loops having 
Burgers vectors of this type have been found in iron bombarded with 150-kev Fe + ions, 78 
but apparently have never been observed in neutron-irradiated bcc metals* 

2.4 REACTOR DOSIMETRY 

MONTE CARLO SPECTRUM CALCULATIONS (L. S. Burns, D. G. Besco, J. L. Kamphouse, 
J. Moteff) 

The first phase of the Monte Carlo calculations of the neutron energy spectrum for 
various regions of the EBR-II was completed and some preliminary data are presented 
in this section. The neutron energy spectrum was determined for each region shown in 
the computer nuclear mockup in Figure 2.83. Gaps above and below the core were made 
equal in the mockup to utilize symmetry in the calculations. The spectrum was composed 
of 40 energy groups of equal lethargy between 0. 01 and 14 Mev. The quadrilateral areas 
shown in the mockup are rotated about the reactor centerline to give a quasi three- 
dimensional geometry, described in the r-z plane and assumed in the 4> rotational direc¬ 
tion. Each region is composed of a homogeneous mixture of materials present in those 
regions. Neutron histories are started in the reactor and followed until they escape the 
configuration, are absorbed, or fall below the energy cutoff. The total number of neu¬ 
trons below 0.01-Mev energy are also recorded for each region. The calculation con¬ 
sidered elastic scattering, inelastic scattering, radiative capture, (n,a) reactions, 

(n-2n) reactions in beryllium, and neutron absorption without secondary emission. The 
core was composed of 91 elements with twelve control rods and two safety rods; it was 
surrounded by two rows of stainless steel and the U^8 blanket. Appropriate axial and 
radial power distributions with a Watts fission spectrum were used to obtain source 
neutrons. 

For presentation the integrals of the differential neutron flux densities, 0’ (E), were 
normalized so that the number of neutrons above 1 Mev are equal to unity. These curves 
are given for four regions of the core in Figure 2. 84. Regions 2A and 7A represent the 
third of rows 2 and 7, respectively, at the core midplane and above. Regions 2B and 7B 
represent the third of rows 2 and 7, respectively, above the midplane nearest the top of 
the core. Row 7 is actually within the stainless steel reflector surrounding the core. 

These four regions are clearly delineated in Figure 2. 83. 

The integrated neutron flux densities giving the fraction of neutrons above a given energy 

r 14 

E, 0(E) = j 0'(E)dE, for each one of the four regions, along with the Watt fission spec- 

E 

trum for comparative purposes, are shown in Figure 2.85. The energy E has the limits 
0. 01 < E< 14 Mev. In this case0 T (E) was normalized so that the area under the complete 
curve from 0.01 to 14 Mev is equal to unity. The fraction of neutrons, 0(E), above 0.01 
Mev as obtained from these curves are simply those considered in the detailed spectrum 
calculation. Since there are also neutrons present at energies below 0.01 Mev, the frac¬ 
tion of the total neutrons greater than E in the respective regions will be smaller than that 
obtained from Figure 2.85 by the factor k which will also be different for each region. 
These factors are listed in Table 2.14. 

EBR-H FLUX DENSITY MEA SUREMENTS (R. L. Stuart, J. L. Kamphouse, J. Moteff) 

Wire monitors which were used for the flux mapping experiment in the EBR-II rows 
2 and 7 positions were counted and preliminary results for the fast neutron flux densities 
were determined by using the threshold reactions Fe^ 4 (n,p)Mn^ 4 , Ni^^(n,p)Co^^, and 


























Lethargy, n (lethargy units) 



0.01 0.1 1 10 100 

Neutron energy (E), Mev 


Fig. 2.85 — Monte Carlo neutron flux densities above energy E for four core regions of 
EBR-II. Flux densities normalized to unity above 0.01 Mev. 







169 


Ti46(n,p)Sc46. The Co^^(n,y)Co^^ and Fe^(n,y)Fe59 reactions were used to obtain in¬ 
formation of the neutron environment at neutron energies below those obtained with the 
threshold reactions. A total of 148 dosimeters were used in this experiment. 

The saturated activity of the Ni^ fast neutron dosimeters and the Co 5 ^ and Fe^ slow 
neutron dosimeters was normalized to unity at the core midplane and is given as a func¬ 
tion of position in Figure 2. 86. The actual values of the saturated activity for rows 2 and 
7 at the core midplane are given in Table 2.14, for regions 2A and 7A. The Ni-Co dosi¬ 
meter saturated activities for regions 2B and 7B were obtained from dosimeters located 
15. 2 cm above the core midplane, and the iron-saturated activity for regions 2B and 7B 
was obtained from dosimeters located 12. 7 cm above the core midplane. 

Using the Monte Carlo calculated spectra shown in Figure 2.84, the Ni®®(n,p)Co^® and 
Fe 54 (n,p)Mn 54 spectrum-averaged cross sections were determined for regipns 2A, 2B, 

7A, and 7B. These cross sections are given in Table 2.15 along with those obtained if 
one were to use a Watt fission spectrum. The neutron flux density (E n ^ 1 Mev) was 

TABLE 2.14 


SATURATED ACTIVITY VALUES FOR FOILS USED IN EBR-II DOSIMETRY EVALUATION 




Saturated Activity, A^, dis. sec ^ mg ^ 


Region 

Ni 58 (n,p)Co 58 

Fe 84 (n,p)Mn 84 

Co 59 (n, 7 )Co 88 

Fe 58 (n/y)Fe 59 

Row 2 

A 

2.61 x 10 8 

1.66 x 10 7 

2.17 x 10 8 

2.41 x 10 5 

•B 

1.56 x 10 8 

1.17 x 10 7 

4.70 x 10 8 

2.35 x 10 5 

Row 7 

A 

9.37 x 10 7 

5.62 x 10 6 

3.61 x 10 8 

2.07 x 10 5 

B 

5.77 x 10 7 

4.23 x 10 6 

6.92 x 10 8 

2.07 x 10 5 


TABLE 2.15 


SPECTRUM-AVERAGED CROSS SECTIONS AND FLUX DENSITIES FOR 
SEVERAL CORE REGIONS OF EBR-II 


Region 

Cross Section, 

0 ^ (E n > 1 Mev), millibarns 3 

Neutron Flux Density,* 3 

0(E n > 1 Mev), n/cm^-sec 

Ni 88 (n,p)Co 88 

Fe 84 (n,p)Mn 84 

Ni 58 (n,p)Co 58 

Fe 84 (n,p)Mn 84 

Row 2 





A 

117 

103 

3.19 x 10 14 (2.31 x 10 14 ) 

i 2.45 x 10 14 (1.78 x 10 14 ) 

B 

84 

74 

2.65 x 10 14 (1.38 x 10 14 ) 

i 2.43 x 10 14 (1.27 x 10 14 ) 

Row 7 





A 

89 

80 

1.49 x 10 14 (8.23 x 10 13 ) 

i 1.05 x 10 14 (6.02 x 10 13 ) 

B 

64 

59 

1.28 x 10 14 (5.07 x 10 13 ) 

1.10 x 10 14 (4.59 x 10 13 ) 

Watt fission 





spectrum 

162 

141 




a Cross sections calculated using AFWL TR-65-34 (WL1) values; they are based on the EBR-II Monte Carlo 
spectra normalized to unity above 1 Mev {see Figure 2.84). 

bValues in parentheses were obtained using the Watt fission spectrum-averaged cross sections. Of (E n ^ 1 Mev), 
as listed in this table. The average cross sections, a f , in a Watt fission spectrum are 112 and 98 mb, respec¬ 
tively, for the Ni 88 and the Fe 84 (n,p) reactions; i.e.. Of (E n ^ 1 Mev) - Of/0.693. 



Fig. 2.86 — Saturated activity for Ni^®, Co^, and Fe®® as a function of position in 
EBR-I I, normalized to unity at the core midplane 













































171 


calculated from various dosimeters located in the four regions using the spectrum- 
corrected cross sections. The Ni^ dosimeters in regions and were located 15.2 
cm above the core midplane. The Fe 5 ^ dosimeters in regions 2A and 7A were located 
2. 54 cm above the core midplane, and the Fe 5 ^ dosimeters in regions 2B and 7B were 
located 12.7 cm above the core midplane. 

The neutron flux density (E n ^ 1 Mev) as measured by the Fe^ dosimeters listed in 
Table 2.15 is essentially constant in going from the midplane to 12.7 cm above the mid¬ 
plane for either row 2 or row 7. Although the saturated activity is reduced about 40 per¬ 
cent, the spectrum-averaged cross sections are increased by almost the same amount 
and consequently the neutron flux density (E n ^ 1 Mev) remains virtually constant. The 
Ni 5 ^ dosimeters show a similar trend. Serious errors can result if a Watt fission spec¬ 
trum is assumed for obtaining the cross sections to be used in determining neutron flux 
densities from foil activities. The uncertainties can range from a factor of 1.4 in region 
2A up to a factor of 2.4 in region 7B. 


CONCLUSIONS 


5 SUMMARY 


Post-irradiation creep- rupture tests of Incoloy 800 specimens irradiated in EBR-II to 
fast neutron fluences (E n ^ 1 Mev) up to approximately 3 x 10^0 n /cm^ showed that the 
rupture life is reduced by irradiation when tested at 540° C but increased for a test tem¬ 
perature of 705°C. Correspondingly the minimum creep rate is increased at 540°C but 
decreased at 705°C. Post-irradiation rupture elongation values and in the region of 15 to 
20 percent or about one-third the unirradiated value at 540°C, and in the region of 20 to 
50 percent at 705°C, which is about one-half the unirradiated^values. 

Post-irradiation creep-rupture properties of Hastelloy i£ specimens irradiated in either 
ETR or EBR-II comparable fast fluences are essentially the same for both types of irra¬ 
diation. The ductility determined by elongation measurements is about 5 percent for speci¬ 
mens irradiated to a fast neutron fluence of 3. 2 x 10^0 n/cm^ (E n ^ 1 Mev) compared to 
about 80 percent foi^he control at a test temperature of 704°C. 

Hastelloy R-235 containing the same total boron (50 ppm) concentration but varying 
B*0 isotope content and post-irradiation creep-rupture tested at 870°C showed increas¬ 
ing damage with increasing B^ content. This clearly supports the proposed mechanisms 
that the B^ isotope, in some manner, contributes to reduced ductility of irradiated alloys. 

Transmission electron microscopy studies on irradiated A-286 and Hastelloy R-235 show 
shells of localized high-damage regions (dislocations) and gas bubbles with radii approxi¬ 
mating the recoil distance of lithium and alpha particles in iron or nickel. The range of 
the lithium particle is calculated to be approximately 1 micron and that of the alpha to be 
about 2 microns. 

Resistivity studies of irradiated AJSTM-A302BjLndicate that the embrittlement is proba¬ 
bly caused by the formation of a carbon-defect complex. This complex recovers in the 
temperature range of 300° to 500°C with the subsequent resolutioning of carbon. Resis¬ 
tivity studies on this material correlate, over the wide range investigated, quite well with 
radiation-induced changes and the subsequent recovery of mechanical properties. 

/ 4 & 

The effect of irradiation temperature on creep -ruptur e properties of molybdenum at 
750°C was determined. Irradiation at 70°C resulted in the least effect on the time to rup¬ 
ture; irradiation at 700°C and 1000°C showed increases in the time to rupture by factors 
of 12 and 18, respectively. There was also a relatively long second-stage creep period 
which increases with an increase in the irradiation temperature. Post-irradiation anneal¬ 
ing at 1000°C of a specimen irradiated at 700°C did not produce a significant change in 
the creep rate but did increase the rupture life by 25 percent over that observed in the 
as-irradiated (700°C) condition. —-3 



172 


(Accelerated creep of irradiated molybdenum was observed at 580°C. Delayed creep, 
a period of time during the initial portion of the creep curve in which essentially no creep 
is observed to occur, was observed at 600°, 700°, and 750°C in specimens irradiated at 
reactor-ambient temperature. The accelerated creep was observed only for those speci¬ 
mens irradiated at the lower fluence (4.8 x 10 18 n/cm 2 , E n > 1 Mev). Delayed creep to 
date has not been observed in control specimens, in tests of ambient temperature ir¬ 
radiated specimens at 580°, 850°, or 900°C, nor in700°C and 1000°C irradiated specimens. 

The effect of post-irradiation annealing on the 700°C creep-rupture properties of irra¬ 
diated (1-3 x 1020 n / cm 2) molybdenum at two carbon levels, 26 ppm and 205 ppm, was 
studied. An annealing temperature of 900° C is necessary to remove the incubation period 
for delayed creep. Although an anneal at 1600°C eliminated the incubation period, this 
heat treatment does not restore the properties to control values. The data suggest the 
formation of stable defects at temperatures between 900° and 1200°C which tend to 
strengthen the irradiated material for those test conditions. 

Initial results of the effects of irradiation at reactor-ambient (~70°C) temperatures to 
fast neutron fluences of 2.1 x 10^0 n / C m2 (E n ^ 1 Mev) on niobium and Nb — lZrumsile 
specimens show that the room-temperatureyie^ftfrengths are increasedabouTllO and 
200 percent respectively for the two materialsTrKe yield strength increases at 650°C 
are about 93 and 110 percent, respectively, for niobium and Nb - IZr. The ductility of 
360°C and 650°C of both metals is significantly reduced as a result of irradiation. Com¬ 
plete recovery occurs in the room-temperature strength and ductility of both metals 
following a 1000°C anneal for 1 hour. 

A detailed study of the various recovery stages observed in resistivity measurements 
of isochronally annealed irradiated molybdenum was conducted. Molybdenum appears to 
saturate at about 1 x 10^0 n/cm^ (E^ 1 Mev) with the saturation effect occurring as a 

resistivity maximum rather than the normal asymptotic approach. No significant stage HI 
(~0.15 T m ) temperature shift nor residual resistivity was observed at high fluences or 
anneal temperatures; this indicates negligible transmutation reactions. The unusual satur¬ 
ation behavior apparently arises from the relatively high irradiation temperature (0.12 T m ) 
leading to a removal of free self-interstitials, with only excess vacancies left behind. 


Hot-hardness studies of irradiated 73 percent warm-worked tungsten and irradiated re¬ 
crystallized tungsten of the same sheet showed a factor of 3 greater radiation-induced 
hardening in the recrystallized material. Hardening occurred primarily in the athermal 
portion of the curve from 280°C (0.15 T m ) to 1050°C (0.36 T m ). Tests on recrystallized 
tungsten following either 73 or 98 percent warm working show the same values for either 
original material in the unirradiated and irradiated conditions. The hardness increase 
was the same for both cases, indicating that the substructure was the same for both ma¬ 
terials. The 73 percent warm-worked material showed a radiation-induced hardness in¬ 
crease at 316°C (0.16 T 
increase. 


i); the 98 percent warm-worked material showed no additional 






, -l * * j i ‘ 

A /I iV JaUA 

(*/',' ' ' 

Hot-microhardness data were also /letermined for Nb, Nb 

25tfalloy (W — 30Mo — 25Re, at. 


&r- . ]fj 

IZr , Mo, Mo - 0.5T i, I&S- 
Ta, Ta^ lOWj^ 


alloy (W — 30Re — 30Mo, at. 7 o), \w — ouiviu — at. 7 c 

from 25° to 1200°C. Most tests showed that the ultimate yield strength (a u ) increases 
linearly with an increase in the hot-hardness (Hy) values when compared at correspond¬ 
ing temperatures. The relationship u u = A H v , where A is a dimensionless number with 
values ranging from 0. 3 to 0. 4. 


Transmission electron microscopy was used to study the annealing behavior of small 
defects in irradiated tungsten and to determine the crystallography of dislocation loops 
formed during this annealing. These studies revealed a complex annealing process attri- 



173 


buted to the presence of two overlapping temperature regions within which different defect 
species become mobile. At temperatures below 435°C annealing results from migration 
of free interstitials; annealing above this temperature is related to vacancy migration. 
Dislocation loops formed during high-temperature annealing (1100°C) are vacancy in na¬ 
ture and pure edge in character, lying on {ill} planes and having a/2<lll> Burgers 
vectors. 

Micro structures of molybdenum irradiated at three different temperatures and creep- 
rupture tested at 750°C were studied by transmission electron microscopy. These ob¬ 
servations revealed many interstitial dislocation loops throughout the matrix in speci¬ 
mens irradiated at 70° C and 700° C. In the specimen irradiated at 700° C, a zone about 
1 micron wide, denuded of loops, parallels the grain boundaries. The specimen irra¬ 
diated at 1000°C contained no loops within the matrix, but showed occasional loops 
thought to be vacancy in nature near grain boundaries. This specimen also showed evi¬ 
dence of small precipitates, believed to be carbides, which pin mobile dislocations. 

Preliminary evaluation of the Monte Carlo spectrum calculations and the influence on 
fast neutron cross sections for the Ni^(n,p) and Fe54( n? p) reactions of wires irradiated 
in EBR-II was completed. Results show that although the experimentally determined 
saturated activity measured with the wire detectors is reduced about 40 percent from 
the core center to the core edge, the spectrum average cross sections (E n > 1 Mev) are 
increased almost by the same amount. Consequently, the fast neutron fluence (E n ^ 1 Mev) 
remains virtually constant throughout the core region. Serious errors can result if aWatt 
fission spectrum is assumed in the determination of the average reaction cross section 
of these detector wires. The resulting errors can be low by a factor of 1.4 in the center 
of the core up to a factor of 2.4 in the core region near the edge, j ^ 

2. 6 PLANS AND RECOMMENDATIONS ' 

Investigations will continue to study the basic mechanisms of the effects of neutron 
irradiation on the stress-, strain-, time-, and temperature-dependent properties of 
selected heat-resistant alloys and refractory metals. Contributing mechanisms will be 
identified by considering current dislocation theories coupled with analyses of experi¬ 
mental data generated by creep, tensile, resistivity, and hot-hardness measurements 
and by direct observations of radiation-induced defects with the transmission electron 
microscope. 

Particular emphasis will be given to the problem of radiation-induced changes in the 
ductility of some of these metals and alloys and to the importance of neutron spectra 
and fluence on both the substructure and the resulting mechanical properties. 

In view of the need for information concerning the effect of neutron-induced changes 
in welds, a weld evaluation program will be initiated on selected candidate materials for 
fast breeder reactors. Selected materials will initially be limited to the heat-resistant 
alloys. Early work will concentrate on the general behavior of welds in unirradiated 
alloys. 

Current investigation of the mechanisms of neutron-radiation-induced defects in vari¬ 
ous metals and alloys by isochronal and isothermal resistivity studies will be continued. 
Isothermal studies will be expanded to furnish migration energies and jump frequencies 
for various defect stages. Resistivity studies will be expanded to include iron- and 
nickel-base alloys, vanadium, and vanadium alloys. These resistivity studies will also 
be used to investigate the kinetics of bubble (or cavity) diameter changes and their in¬ 
fluence on the cladding void volume. 

Isothermal hot-hardness studies will be continued to investigate activation energies 
associated with various defect recovery stages in irradiated materials. Use of the hot- 



174 


hardness apparatus to study creep phenomena by varying indenter holding times will be 
conducted to employ hot-hardness methods for defining important temperature regions 
to be studied in tensile tests and to provide data for correlating and interpolating ulti¬ 
mate tensile strength of various irradiated alloys. 

Detailed transmission electron microscopy studies to further define the kinetics of defect 
cluster and loop formation will be conducted. Emphasis will be placed on electron micro¬ 
scopy, resistivity, and density studies to define the number and size of defect loops, cavi¬ 
ties, and/or gas bubbles for correlation with corresponding strength and hardness proper¬ 
ties and changes in ductility. 

Creep-rupture and tensile testing of irradiated specimens will be continued at temper¬ 
atures above and below the prominent recovery regions to evaluate the magnitude and mecha¬ 
nisms of the neutron-irradiation-induced changes in properties. Particular emphasis will 
be given to those materials irradiated in EBR-II to fluences up to about 5 x 10 22 n/cm 2 . 
These studies will include detailed investigations on the radiation-induced changes in first - 
and second-stage creep behavior by the technique of applied stress changes on single 
specimens. The study of molybdenum and niobium alloys will be continued on potential 
LMFBR candidate materials including the vanadium-base alloys. 

2.7 REFERENCES 

1. Busboom, H. J. and Mathay, P. W., "Fast Neutron Damage Studies in High Nickel 
Alloys," GE-APED, GEAP 4985, August 1966. 

2. Stiegler, J. O. and Weir, J. R., "Effects of Irradiation on Ductility, " ORNL, 
ORNL-TM-2019, January 1968. 

3. Monkman and Grant, Trans. ASTM, Vol. 56, 1956, p. 593. 

4. "AEC Fuels and Materials Development Program Progress Report No. 69," GE-NMPO, 
GEMP-69, September 29, 1967, pp. 50-58. 

5. Woodford, D. A., "Constant Load Creep Data Interpreted in Terms of the Stress De¬ 
pendence of Dislocation Velocity, " Trans. Met. Soc. AIME, Vol. 239, May 1967, 
pp. 655-659. 

6 . "Annual Progress Report for Period Ending June 30, 1966," ORNL, Metals and 
Ceramics Division, ORNL-3970, Table 20. 3, p. 116. 

7. Harwell, UK-AEA, private communications. 

8 . "Sixth Annual Report — High-Temperature Materials Program, Part A, " GE-NMPO, 
GEMP-475A, March 31, 1967, pp. 205-208. 

9. Wagenblast, H., Fujita, F. E., and Damask, A. C., "Kintics of Carbon Precipitation 
in Irradiated Iron-IV Electron Microscope Observations, " Acta Met., Vol. 12, 1964, 
pp. 347-353. 

10. Fujita, F. E. and Damask, A. C., "Kinetics of Carbon Precipitation in Irradiated 
Iron-II Electrical Resistivity Measurements, " Acta Met., Vol. 12, 1964, pp. 331-339. 

11. Carpenter, G. F., Knopf, N. R., and Byron, E. S., "Anomalous Embrittling Effects 
Observed During Irradiation Studies on Pressure Vessel Steels," Nucl. Sci. and Engr., 
Vol. 19, 1964, pp. 18-38. 

12. Steele, L. E. and Hawthorne, J. R., "New Information on Neutron Embrittlement 
Relief of Reactor Pressure Vessel Steels," ASTM, STP-380, 1965, p. 289. 

13. GEMP-475A, pp. 201-204. 

14. "AEC Fuels and Materials Development Program Progress Report No. 71," GE-NMPO, 
GEMP-1002, December 29, 1967, pp. 48-50. 

15. GEMP-69, pp. 58-61. 

16. Gleiter, H. and Hornbogen, E., "Theory of the Interaction of Dislocations with 
Coherent Ordered Zones, (1)," (In German), Phys. Status Solidi, Vol. 12, 1965, 
pp. 235-250. 


175 


17. Copley, S. M. and Kear, B. H., "A Dynamic Theory of Coherent Precipitation 
Hardening with Application to Nickel-Base Superalloys, " Trans. AIME, Vol. 239, 

1967, pp. 984-992. 

18. Raymond, E. L., "Effect of Grain Boundary Denudation of Gamma Prime on Notch- 
Rupture Ductility of Inconel Nickel-Chromium Alloys X-750 and 718," Trans. AIME, 
Vol. 239, 1967, pp. 1415-1422. 

19. Gleiter and Hornbogen, op. cit., pp. 251-264. 

20. Mastel, B. and Brimhall, J. L., "The Combined Effect of Carbon and Neutron Radia¬ 
tion on Molybdenum, " Acta Met., Vol. 13, 1965, pp. 1109-1116. 

21. GEMP-475A, pp. 69, 70. 

22. Brimhall, J. L. and Mastel, B., "Effect of Prestrain and Subsequent Neutron Irra¬ 
diation on Molybdenum, " Acta. Met. Vol. 14, No. 4, April 1966, pp. 539-541. 

23. Damask, A. C. and Dienes, C. J., "Point Defects in Metals, ” Gordon and Breach, 
New York, 1963, pp. 146-147. 

24. GEMP-400A, pp. 86-90. 

25. GEMP-475A, pp. 80-82. 

26. GEMP-1002, pp. 61-63. 

27. GEMP-1002, p. 60. 

28. Tietz, T. E. and Wilson, J., "Behavior and Properties of Refractory Metals," 
Stanford University Press, 1965, p. 20. 

29. Geary, A. L., "Fundamental and Applied Research in Metallurgy, " Nuclear Metals, 
Inc., NMI-1258, September 17, 1963, p. 11. 

30. Petty, E. R. and O' Neill, H., "Hot Hardness Values in Relation to the Physical 
Properties of Metals," Metallurgia, Vol. 63, 1961, pp. 25-30. 

31. Westbrook, J. H., "Temperature Dependence of the Hardness of Pure Metals," 

Trans. ASM, Vol. 45, 1953, pp. 221-248. 

32. O’ Neill, H., "Hardness Measurement of Metals and Alloys,," 2nd Ed., Chapman and 
Hall, Ltd., London, England, pp. 153-207. 

33. Wronski, A. S. and Johnson, A. A."A Hardening Effect Associated with Stage ni 
Recovery in Neutron-Irradiated Molybdenum," Phil., Vol. 8, No. 90, June 1963, 
pp. 1067-1070. 

34. GEMP-475A, pp. 73-77. 

35. GEMP-400A, pp. 76-77. 

36. Makin, M. J. and Minter, F. J., "The Mechanical Properties of Irradiated Niobium," 
Acta Met., Vol. 7, June 1959, pp. 361-366. 

37. GEMP-475A, pp. 73-74. 

38. "AEC Fuels and Materials Development Program Progress Report No. 67, GE-NMPO, 
GEMP-67, June 30, 1967, p. 55. 

39. GEMP-475A, pp. 93-96. 

40. Peacock, D. R. and Johnson, A. A., "Stage III Recovery in Neutron-Irradiated 
Molybdenum and Niobium," Phil. Mag. Vol. 8, 1963, pp. 275-284. 

41. Holmes, D. K., "The Interaction of Radiation with Solids," North Holland Publishing 
Company, 1964, p. 147. 

42. Kissinger, H. E., Brimhall, J. L., and Mastel, B., "Physical Characterization of 
Molybdenum Single Crystals for Irradiation Experiments, ” International Conference 
on Characterization of Materials," Pennsylvania State University, November 16-18, 
1966. 

43. GEMP-475A, p. 83. 

44. GEMP-69, pp. 37-44. 

45. Ibragimov, S. S., Lyashenko, V. S., and Zavyalov, A. I., "An Investigation of the 
Structure and Properties of Several Steels and Other Metals After Irradiation with 
Fast Neutrons," J. Nucl. Energy, Vol. 16, 1962, pp. 45-49. 




176 


46. Keys, L. K., Smith, J. P., and Moteff, J., "High Temperature Recovery of Neutron- 
Irradiated Tungsten, " Bulletin of the American Physical Society, Vol. 13, 1968, 

p. 463. 

47. Kinchin, G. H. and Thompson, M. W., "Irradiation Damage and Recovery in Moly¬ 
bdenum and Tungsten, " J. Nucl. Energy, Vol. 8, 1958, pp. 275-284. 

48. GEMP-69, p. 46. 

49. GEMP-69, p. 42. 

50. Kinchin and Thompson, op. cit. p. 279. 

51. Nihoul, J., "The Recovery of Radiation Damage in Molybdenum, " Phys. Stat. Solidi, 
Vol. 2, 1962, p. 310. 

52. Kinchin and Thompson, op. cit., p. 280. 

53. GEMP-475A, Figure 2.23, p. 87. 

54. Lomer, W. H. and Cottrell, A. H., "Analysis of Point Defects in Metals and Alloys, " 
Phil. Mag., Vol. 46, 1955, pp. 701-719. 

55. DeJong, M. and Afman, H. B., "Resistometric Measurements on Molybdenum Irra¬ 
diated with 2.5 Mev Electrons," Acta Met., Vol. 15, 1967, pp. 1-12. 

56. Lacefield, K., Moteff, J., and Smith, J. P., "Neutron Radiation Damage in Tungsten 
Single Crystals," Phil. Mag., Vol. 13, 1966, p. 1079. 

57. Rau, R. C., Moteff, J., and Ladd, R. L., "Comparison of Microstructure with 
Mechanical Properties of Irradiated Tungsten," J. Nucl. Mat., Vol. 24, 1967, p. 164. 

58. Higgins, P. R. B. and Roberts, A. C., "The Nature and Annealing Behavior of 
Fission Fragment Damage in Molybdenum, " J. Less-Common Met., Vol. 6, 1964, 
p. 472. 

59. Downey, M. E. and Eyre, B. L., "Neutron Irradiation Damage in Molybdenum, " 

Phil. Mag., Vol. 11, 1965, p. 53. 

60. Meakin, J. D. and Greenfield, I. G., "Interstitial Loops in Neutron-Irradiated 
Molybdenum," Phil. Mag., Vol. 11, 1965, p. 277. 

61. Ladd, R. L. and Rau, R. C., "Immersed Double-Jet Technique for Electro thinning 
Tungsten for Transmission Electron Microscopy, " Rev. Sci. Instr., Vol. 38, 1967, 

p. 1162. 

62. Hirsch, P. B., Howie, A., and Whelan, M. J., "A Kinematical Theory of Diffraction 
Contrast of Electron Transmission Microscope Images of Dislocations and Other 
Defects," Phil. Trans. Roy. Soc. A, Vol. 252, 1960, p. 499. 

63. Mazey, D. J., Barnes, R. S., and Howie, A., "On Interstitial Dislocation Loops in 
Aluminium Bombarded with Alpha-particles," Phil. Mag., Vol. 7, 1962, p. 1861. 

64. Bilby, B. A., Bullough, R., and Smith, E., "Continuous Distributions of Dislocations: 
A New Application of the Methods of Non-Riemannian Geometry, " Proc. Roy. Soc. A, 
Vol. 231, 1955, p. 263. 

65. GEMP-67, pp. 60-61. 

66. GEMP-69, pp. 44-48. 

67. Moteff, J. and Smith, J. P., "Recovery of Defects in Neutron-Irradiated Tungsten, " 
ASTM STP-380, 1965, p. 171. 

68. GEMP-475A, pp. 82-89. 

69. Attardo, M. and Galligan, J. M., "A Field Ion Microscope Study of Neutron-Irradiated 
Tungsten," Phys. Stat. Sol., Vol. 16, 1966, p. 449. 

70. Attardo, J. J., Galligan, J. M., and Chow, J. G. Y., "Interstitial Removal in Stage- 
El Recovery of Neutron-Irradiated W, " Phys. Rev. Letters, Vol. 19, 1967, p. 73. 

71. Jeannotte, D. and Galligan, J. M., "Energy of Motion of Vacancies in Tungsten, " 
Phys. Rev. Letters, Vol. 19, 1967, p. 232. 

72. Edmondson, B. and Williamson, G. K., "On the Determination of the Nature of Dis¬ 
location Loops," Phil. Mag., Vol. 9, 1964, p. 277. 

73. Mastel, B., Kissinger, H. E., Laidler, J. J., and Bierlein, T. K., "Dislocation 


177 


Channeling in Neutron-Irradiated Molybdenum," J. Appl. Phys., Vol. 34, 1963, 
p. 3637. 

74. GEMP-69, pp. 29-31. 

75. Mastel and Brimhall, op. cit., p. 1109. 

76. Brimhall, J. L., Mastel, B., and Bierlein, T. K., "Thermal Stability of Radiation 
Produced Defects in Molybdenum," Acta Met. 

77. Dingley, D. J. and Hale, K. F., "Burgers Vectors of Dislocations in Deformed Iron 
and Iron Alloys," Proc. Roy. Soc. A, Vol. 295, 1966, p. 55. 

78. Masters, B. C., "Dislocation Loops in Irradiated Iron," Phil. Mag., Vol. 11, 1965, 

p. 881. 




73083 


3 ^ADVANCED FAST BREEDER REACTOR 
FUEL ELEMENT f.I.AItniNC DEVELOPMENT) 

(iiiij 

E. S. Funston,* C. 0. Tarrt 


[ThTol 


| The objective of this program is to further the technology of refractory metals and, in 
particular, the new refractory metals developed under this task, to permit their early 
utilization in the high-temperature reactor programs sponsored by the AEC. Research 
will continue to develop and obtain new refractory metals which will make possible marked 
improvements in the temperature and life capabilities of nuclear powerplants. j tj 


This task was redirected during calendar year 1967 to emphasize the advanced fast breeder 
fuel element cladding development activities. Included in the reoriented task is the require¬ 
ment to produce refractory metal thermoelement wire (W-Os) and small-diameter tungsten 
alloy tubing that cannot be obtained commercially. 


3.1 REFRACTORY METAL ALLOY TUBING, SHEET, AND WIRE PRODUCTS 


The major effort during CY-67 was directed toward developing processing procedures 
for the production of quality seamless W-Re-Mo tubing, sheet, and bar and W-Os wire. 
Special emphasis was placed on the development of fabrication procedures for seamless 
tubing and trial production of specific tubing sizes. In order to carry out these tasks, spe¬ 
cial processing equipment was designed and installed. 1 


The key items of equipment include: 

1. Facilities for the purification of metal powders, sintered compacts, and sintered mill 
products (acid leaching facilities and various annealing furnaces with controlled hydro¬ 
gen atmosphere; -50°C to +5°C). 

2. An experimental 20,000-pound capacity variable-speed draw bench for mandrel draw¬ 
ing of seamless tubes up to 5.0-cm OD by 366-cm finished length, Figure 3.1. A spe¬ 
cial feature of the draw bench includes an electronic control console for manual or 
automatic control of the drawing operation by selection of wave amplitude (force) and 
frequency (variable from 30 to 30,000 kc/sec). Data obtained from this equipment, 
including load cell read-out, yield information directly applicable to the working 
characteristics of each alloy. 

3. Six hydraulically controlled draw benches for repetitive production runs of seamless 
tubing to 2.0-cm OD by 160-cm finished length, Figure 3.2. 

4. Accessory equipment required to meet the quality standards for refractory alloy seam¬ 
less tubing (quenching fixture to assure meeting straightness tolerances of hardened 
tool steel mandrels; honing fixtures to eliminate gouging or scoring of the ED surface 
of the tubes during the honing operation). 


‘Project leader. 

+ 

Principal investigator. 

^"AEC Fuels and Materials Development Program Progress Report No. 71," GE-NMPO, GEMP-1002, December 29, 1967, 
pp. 68-76. 


178 




179 



Fig. 3.1 - Draw bench for developing procedures in drawing tubing up to 5.0-cm 
diameter by 366-cm long, finished size. Actual draw stroke of the ma¬ 
chine is 426 cm. Photo shows a work piece on a hardened mandrel 
passing through a die. The dark cylinder with an attached electric cable 
at the rear of the draw chuck is a tensile load cell attached to a recorder. 

Valves and burners for warming the workpiece and the hardened 
mandrel can be observed near the die holder. Commercial stroke align¬ 
ment rails appear as 3.8-cm-diameter bars on each side of the draw 
length. (Neg. P67-10-94) 

5. A beryllia-insulated, molybdenum-lined annealing furnace capable of operating con¬ 
tinuously at 1650°C in hydrogen atmosphere, Figure 3. 3. A water-cooled chamber 
is attached to the front of the molybdenum muffle to permit purging of the furnace 
with inert gas and hydrogen before the work load passes into the hot zone. 

6. A wire drawing facility capable of warm drawing W-Os and W-Re alloy wire to 0.0025 
cm diameter. Diamond-coated dies are employed and heating of the wire is accom¬ 
plished by induction or gas burners. 

7. Facilities for reclamation of rhenium from W-Re, W-Re-Mo, and Mo-Re alloys. 






Fig. 3.2 - Repetitive work tube-drawing benches. One operator controls four benches of this type. Three 

benches, as shown, are controlled by hydraulic valving with motion established by a foot-operated 
valve mounted on the horizontal pedestal bracket below the die station of each draw bench. Speed 
control is provided by the micrometer flow-control valve mounted above the foot-control position. 
(N6g. P67-10-9B) 

The processing procedures developed or selected for the fabrication of the various 
W-Re-Mo alloy mill products required in the program are shown schematically in Fig¬ 
ure 3.4 and apply to both powder-metallurgy and arc-cast materials. Although the flow 
diagram was developed specifically for W-Re-Mo alloys, the general processing sequence 
can be adapted to the fabrication of tungsten and molybdenum and other solid-solution al¬ 
loys of tungsten and molybdenum. The material processed from arc-cast ingots was pro¬ 
duced from metal powders which were purified, isostatically compacted, and hydrogen 
sintered at NMPO. The sintered electrodes were double arc melted through the coopera¬ 
tion of the U. S. Department of the Interior, Bureau of Mines, Albany, Oregon. 

Primary breakdown of the arc-cast or sintered billets is accomplished by extrusion at 
NMPO using a 1250-ton Loewy horizontal extrusion press. In FY 1968, a total of 70 billets 
were extruded at NMPO; a summary of the alloys which were extruded is presented in 
Table 3.1 and pertinent processing data are presented in Table 3.2. 

SEAMLESS TUBING 

For the production of seamless tubing, machined hollows (2.50 — 5.14-cm OD) of both 
arc-cast ingots and sintered billets were encased in molybdenum and coextruded to 1.5- 
or 2. 0-cm-diameter product out of a 13.18-cm ID container. After extrusion, the molyb¬ 
denum is removed by leaching in a concentrated nitric acid solution leaving the W-Re-Mo 
alloy tube hollow. The resulting tube hollow is then cut to appropriate lengths and sized 
and honed for warm drawing to 0.39- to 1.27-cm OD seamless tubing. Drawing is carried 
out at temperatures of 200°C using a WS2 lubricant. Frequent intermediate stress relief 
anneals are required to eliminate laminations or in severe cases fracture of the tube. 

In the case of purified molybdenum, Mo — 35Re and Mo — 5W alloys, it was necessary 
to wrap the hollow extrusion billet with 0.005-cm thick tungsten foil so that the nitric acid 
leaching treatment designed to remove the molybdenum jacket would not attack the purified 




181 



Fig. 3.3 — Beryllia-insulated, hydrogen-atmosphere furnace operated continuously at 1650°C 
in stress-relieving W-base alloys during processing to tubing or sheet. The fore chamber 
is water-cooled, and permits purging the work load with inert gas and hydrogen before 
entering the furnace hot zone. (Neg. P67-5-12C) 

molybdenum or molybdenum alloy extrusion. The tungsten alloy foil also has been used for 
W —30Re — 30Mo alloy extrusions to suppress surface irregularities due to selective acid 
attack. 

Seamless tube processing of refractory alloys must be adaptable to a wide range of sizes. 
Extrusions of W — 25Re — 30Mo alloy with an ID/OD ratio of 0.471, as-extruded, represent 
the heaviest-walled tubular extrusions made at NMPO; the finished tube had an ID/OD ratio 
of 0.27. Normally, tubes are processed with a relatively thin wall having an ID/OD ratio 
of 0.96. The heaviest arc-cast billet of W - 25Re — 30Mo alloy that was extruded weighed 
38 kg; and the heaviest powder-metallurgy billet processed at NMPO weighed 29 kg. Larg¬ 
er sizes exceed the isostatic pressing and sintering furnace capacity for metal powders 
at NMPO. However, through the cooperation of the Oak Ridge National Laboratory Y-12 
plant, a large 14.5-cm diameter W — 25Re — 30Mo alloy powder compact was prepared for 
purification at NMPO, Figure 3. 5. After sintering, the compact shrank to 11.1-cm diam¬ 
eter. No further processing is planned with this billet at this time. 

SHEET 

Hot working procedures for powder-metallurgy W — 25Re alloy specifies an 80 percent 
gage reduction to eliminate sintering voids. A study was conducted to determine the per¬ 
cent hot reduction required to eliminate voids after sintering purified W-Re-Mo alloy com¬ 
pacts. As shown in Figure 3. 6, remnants of voids remain after as much as 69 percent re¬ 
duction of a sintered W — 25Re — 30Mo compact. After 75 percent reduction by hot rolling, 
however, sintering voids appear to be eliminated. Similar results were obtained for 
W — 30Re - 30Mo alloy. In this series hot rolling was carried out by heating the material 
to 1400°C in hydrogen; light reductions were interspaced with thorough reheating between 
rolling passes. These results confirm current GE-NMPO practice of specifying hot work 
reductions of over 80 percent to prepare W-Re-Mo alloys for subsequent cold working. 






Tubing as small as 0.394 cm OD x 0.343 cm ID x 50.8 cm lengths has been drawn in quantity. Smaller sizes 
would be considered a development item. 

^Larger sizes could be fabricated using commercial extrusion and sintering facilities. 



















































183 


TABLE 3.1 


FY-68 EXTRUSION SUMMARY 


Alloy, at. % 

Bar or Rod 

Tube 

Remarks 

W - 25Re 

- 

4 a 


W - 25Re - 30Mo 

18 b 

22° 

Five bars and one tube from arc-cast 
ingots of W - 25Re — 30Mo 

W — 30Re — 30Mo 

4 

8 

Two bars from arc-cast ingots 

Mo — 35Re 

2 

3 

Commercial Mo — 50Re (wt %) alloy 

W - 0.5 Os 

5 d 

- 

W — 0.5 Os for wire 

Purified Mo 

4 

- 

Two bars from arc-cast ingots 

Total 

33 

37 e 



a Ultrasonic evaluation disclosed defects in one machined tube billet which 
was withheld from extrusion. 

bIncludes one set of isostatically pressed slabs not intended for extrusion and 
one large billet prepared but not extruded through July 1, 1968. 
includes one W - 26Re - 30Mo alloy and one W - 30Re - 30 Mo alloy 
prepared for arc-melting by Bureau of Mines and mandrel extruded as tube 
by ORNL. Both items are now awaiting additional processing under this task. 
•^Extruded from 9.65-cm-diameter billet, all others were processed from a 
13.17-cm-diameter extrusion chamber, 
includes eight items started for powder-metallurgy process product but 
diverted to arc-melted material. 


SHAPES 

In an auxiliary phase of this program, machining techniques on W-Re-Mo alloys were 
developed to permit ready-to-use test components to be supplied to advanced programs. 

A complex junction or welding connection of W — 25Re — 30Mo alloy produced from arc- 
cast ingots, to be used for a dynamic liquid metal test loop, is shown in the as-machined 
condition in Figure 3.7. Machining operations utilized in making this single part included 
surface grinding, electro-discharge machining (Elox), single-point internal thread cutting 
and turning, and thread grinding. 

REFRACTORY METAL ALLOY WIRE DRAWING 

Process development to obtain special refractory metal wire for thermocouple applica¬ 
tions was undertaken in CY-67 and, in the last quarter, 0.0076-cm-diameter W — 0.5 Os 
(wt %) alloy wire was produced. The osmium addition to tungsten and W — 25Re alloy is 
required to compensate for the change in output emf of tungsten versus W — 25Re alloy 
thermocouples that occur in reactor service as a result of tungsten transmutation to rhe¬ 
nium and rhenium transmutation to osmium. Commercial vendors have rejected requisi¬ 
tions for osmium-containing alloy wires because of the toxicity of osmium. Injurious ef¬ 
fects to NMPO personnel were avoided by use of complete coverage respiratory face 
masks and by the use of coatings over the material during high-temperature processing 
operations. 

Procedures used in this work generally were conventional. Initial drawing temperatures 
were on the order of 650°C, light reductions were employed, and the material was given 
intermediate process anneals of 1400° C between each draw pass. The preferred method 
of cleaning the wire prior to the in-process anneals is swabbing in caustic detergent fol¬ 
lowed by a wet hydrogen treatment at 1000° C. 



184 


TABLE 3.2 


PROCESSING DATA FOR REFRACTORY ALLOY EXTRUSIONS, FY 1968 


Extrusion 

Number 3 

Alloy 6 

Product 

Objective 

Extrusion Blank Prior to Extrusion 
Length, OD, ID, Weight, 

cm cm cm g 

Completed Extrusion 
Length, OD at Center, 

cm cm 

Extrusion 
Reduction 
Ratio By 
Area 

Ratio 

ID to OD 

As 

Extruded 

Drawn 

Tube 

ID to OD 

Ratio 

D 76 

306 

Solid 

5.7886 

4.8158 

_ 

1,730 

48.26 

1.473 

8.73 

_ 

_ 

D 77 

50Mo-50Re 

Solid 

5.9055 

4.8158 

- 

1,450 

33.02 

1.473 

8.73 

_ 

_ 

D 78 

Arc-cast 306 

Solid 

16.0325 

7.8740 

- 

12,780 

— 

— 

8.20 

— 

_ 

D 79 

50Mo—50Re 

Tube 

9.2075 

4.9530 

3.5941 

1,171 

— 

— 

6.77 

0.725 

_ 

D 80 

Mo 

Solid 

Scrapped in Purification Treatment 

— 

— 

— 

— 

_ 

D 81 

256 

Arc-cast Solid 

12.0294 

6.9215 

- 

7,620 

- 

- 

7.43 

— 

— 

D 82 

50Mo—50Re 

Tube 

11.8364 

4.6990 

3.4925 

1,145 

— 

1.930 

6.45 

0.743 

— 

D 83 

306 

Solid 

8.9408 

4.3307 

- 

2,133 

— 

1.520 

10.61 

_ 

_ 

D 84 

50Mo—50Re 

Solid 

5.6890 

4.8869 

- 

1,405 

— 

1.520 

10.61 

_ 

_ 

D 85 

256 

Solid 

10.0022 

4.3891 

- 

2,451 

_ 

1.520 

8.56 

_ 

_ 

D 86 

306 

Tube 

11.6586 

4.7371 

3.7338 

1,259 

— 

1.930 

6.51 

0.788 

_ 

D 87 

50Mo—50Re 

Tube 

11.8948 

4.6.177 

3.6474 

995 

- 

— 

5.49 

0.789 

_ 

D 88 

256 

Tube 

11.2191 

4.3535 

3.3020 

- 

- 

1.930 

5.50 

0.759 

_ 

D 89 

256 

Tube 

11.760 

4.6355 

3.3147 

1,551 

- 

1.390 

10.48 

0.715 

_ 

D 90 

Mo 

Arc-cast Solid 


Extruded as D117 in two pieces of about equal length 


— 

— 

D 91 

256 

Arc-cast Solid 


Processed for Bureau of Mines melting and ORNL extruding 


— 

— 

D 92 

256 

Tube 

11.1125 

4.4729 

3.3604 

1,232 

- 

1.390 

6.11 

0.751 

0.87 

D 93 

306 

Tube 

11.8618 

4.5948 

3.6195 

1,266 

— 

1.930 

5.75 

0.786 

0.84 

D 94 

306 

Tube 



Scrapped in Sinter Operation 

- 

- 

- 

— 

D 95 

256 

Solid 


Processed for Bureau of Mines melting and ORNL extruding 


— 

— 

D 96 

306 

Tube 

11.7627 

4.9047 

3.5560 

1,457 


1.390 

11.45 

0.724 

Various 

D 97 





Not processed to completion 





D 98 





Not processed to completion 





D 99 





Not processed to completion 





D100 

256 

Tube 

13.5864 

5.9334 

4.3180 

2,852 

- 

1.981 

8.96 

0.726 

0.84 

D101 

256 

Tube 

13.3350 

5.9486 

4.3078 

2,905 

- 

1.981 

9.28 

0.724 

0.84 

D102 

256 

Tube 

13.4416 

5.9131 

4.3688 

2,786 

- 

1.981 

8.88 

0.737 

0.84 

D103 

256 

Tube 

5.8420 

6.0325 

4.4450 

1,240 

— 

1.981 

8.29 

0.736 

0.84 

D104 

256 

Tube 

Used as melt stock in D118 


— 

— 

— 

_ 

_ 

D105 

256 

Tube 

Used as melt stock in D118 



— 


_ 

__ 

D106 

Mo 

Arc-cast Solid 

18.4912 

7.5057 

— 

8,126 

— 

1.905 

13.69 

_ 


D107 

256 

Tube 

Scrapped — Distorted in extrusion 

1270 

1.270 

11.92 

— 

— 

D108 

256 

Tube 

11.8643 

4.5694 

3.4010 

1,384 

1295 

1.270 

11.92 

0.758 

0.87 

D109 

256 

Solid 

10.1244 

4.3865 

- 

4,070 

68 

1.574 

7.75 

_ 

_ 

D110 

256 

Solid 

11.2522 

5.3035 

- 

4,029 

45 

2.540 

4.33 


_ 

Dill 

256 

Melt Stock 

Used for arc-cast D118 

1,341 

— 

— 

__ 

_ 

_ 

D112 

256 

Tube 

11.5621 

4.5034 

3.3807 

1,355 

1271 

1.270 

11.92 

0.751 

0.87 

D113 

306 

Tube 

11.7043 

4.5720 

3.5255 

1,334 

66 

1.981 

5.81 

0.770 

_ 

D114 

306 

Tube 

11.7881 

4.5517 

3.5255 

1,295 

- 

1.270 

11.92 

0.773 

0.87 

D115 

306 

Tube 

Scrapped — Distorted in extrusion 

1270 

1.270 

11.92 

_ 

0.87 

D116 

306 

Tube 

11.7957 

4.5135 

3.3223 

1,445 

1271 

1.270 

11.92 

0.736 

0.87 

D117 

Mo 

Solid 

20.1827 

3.9700 

- 

2,482 

132 

5.227 

6.30 

_ 

_ 

D118 

256 

Arc-cast Solid 

15.1739 

11.8287 

- 

25,015 

39 

5.715 

5.30 

— 

_ 

D119 

256 

Tube 

10.4394 

5.4838 

2.5882 

3,130 

- 

1.981 

8.28 

0.471 

0.27 

D120 

256 

Tube 

12.7406 

5.7327 

2.6212 

4,236 

- ■ 

1.981 

8.28 

0.457 

0.27 


185 


TABLE 3.2 (Cent.) 


PROCESSING DATA FOR REFRACTORY ALLOY EXTRUSIONS, FY 1968 









Extrusion Ratio 

Drawn 




Extrusion Blank Prior to Extrusion 

Completed Extrusion 

Reduction IDtoOD 

Tube 

Extrusion 


Product 

Length, 

OD, ID, Weight, 

Length, OD at Center, 

Ratio By As 

IDtoOD 

Number 3 

Alloy b 

Objective 

cm 

cm cm g 

cm 

cm 

Area Extruded 

Ratio 

D121 

256 

Tube 

12.7000 

4.5542 2.7432 1,861 

_ 

1.270 

11.92 0.602 

0.82 

D122 

256 

Tube 

12.8422 

4.7625 3.5991 1,595 

- 

1.270 

11.92 0.754 

0.82 

D123 

256 

Solid 

12.0446 

2.987 - 1,386 

- 

0.889 

21.39 

- 

D124 




Number not used 





D125 

256 

Tube 

12.5730 

4.8361 3.7733 1,506 

- 

1.905 

6.38 0.781 

0.96 

D126 

256 

Tube 

12.6060 

4.7244 3.5204 1,557 

- 

1.270 

11.92 0.745 

0.87 

D127 

256 

Tube 

12.4561 

4.7625 3.5560 1,620 

- 

1.270 

11.92 0.745 

0.87 

D128 

256 

Tube 

12.7685 

4.7548 3.6830 1,492 

- 

1.270 

12.94 0.774 

0.82 

D129 

256 

Tube 


Used for D158 


- 

- 

0.87 

D130 

256 

Tube 


Used for D158 


1.270 

- - 

0.87 

D131 

256 

Solid 


Used for D132 


- 

_ - 

- 

D132 

256 

Solid 

11.3030 

6.4338 1.6433 15,000 

107 

- 

12.08 0.255 

- 

D133 

256 

Solid 


Used for D132 




- 

D134 

256 

Tube 

Used as melt stock in D158. Melt by Bureau of Mines. 

Made into sheet for NASA. 

- 

D135 

256 

Tube 

Used as melt stock in D158. Melt by Bureau of Mines. 

Made into sheet for NASA. 

- 

D136 

256 

Tube 

Used as melt stock in D158. Melt by Bureau of Mines. 

Made into sheet for NASA. 

- 

D137 

256 

Tube 

Used as melt stock in D158. Melt by Bureau of Mines. 

Made into sheet for NASA. 

- 

D138 

256 

Tube 

Used as melt stock in D158. Melt by Bureau of Mines. 

Made into sheet for NASA. 

- 

D139 

250 

Tube 

13.0556 

5.0114 3.5306 

- 

1.390 

11.92 0.704 

- 

D140 

250 

Tube 

Scrapped in preparation for extrusion 


- 

- - 

- 

D141 

250 

Tube 




1.270 

- - 

- 

D142 

250 

Tube 

12.8905 

4.7498 3.3578 2,189 

- 

1.905 

6.44 0.706 

- 

D143 

256 

Solid 

11.3030 

5.1739 - 38,590 

- 

2.540 

4.12 

- 

D144 

256 

Solid 

10.2946 

7.5082 - 7,439 

- 

3.556 

4.29 

- 

D145 

256 

Solid 

Not extruded in FY 1968 27,830 

- 

- 

- - 

_ 

D146 

256 

Solid 


Four pieces used as heavy bars, and 1 piece used as melt 

stock in D158 


D147 

256 

Solid 

11.0490 

4.3738 - 2,724 

- 

1.346 

10.35 

- 

D148 

256 

Tube 

12.7635 

6.0960 4.4450 2,951 

- 

1.981 

9.37 0.729 

- 

D149 

256 

Tube 


Used for arc-cast D118 


- 

- 

- 

D150 

306 

Arc-cast Solid 12.3545 

7.7012 - 9,661 

- 

3.492 

4.78 

- 

D151 

W—0.5 Os 

Solid 

11.4782 

1.5240 - 342 

- 

0.485 

10.00 

- 

D152 

256 

Tube 

12.0650 

4.7345 3.6195 1,473 

- 

1.778 

6.70 0.763 

- 

D153 

W—0.5 Os 

Solid 

For thermocouple wire. Sizes as D151. 


0.485 

10.00 

- 

D154 

W—0.5 Os 

Solid 

For thermocouple wire. Sizes as D151. 


0.435 

10.00 

- 

D155 

W—0.5 Os 

Solid 

For thermocouple wire. Sizes as D151. 


0.485 

10.00 

- 

D156 

W—0.5 Os 

Solid 

For thermocouple wire. Sizes as D151. 


0.485 

10.00 

- 

D157 

256 

Tube 

13.3350 

4.9530 3.9116 1,629 

- 

1.270 

11.92 0.789 

0.87 

D158 

256 

Arc-cast Solid 


Melted by Bureau of Mines; extruded and processed to sheet for NASA evaluation. 



a Four numbers not used in this series. 

^256 is GE-NSP designation for W—25Re—30Mo alloy. 

306 is GE-NSP designation for W—30Re—30Mo alloy. 

250 is GE-NSP designation for W—25Re alloy. 

Mo-50Re is commercial designation for Mo—35% Re alloy. 
Mo is GE-NSP purified unalloyed Mo. 




186 


3. 2 DEVELOPMENT OF MOLYBDENUM AND ITS ALLOYS FOR FAST BREEDER 
REACTOR APPLICATIONS 

Molybdenum is a promising candidate for containing the liquid-metal heat-transfer 
media in fast breeder reactors. The element resists liquid alkali metal corrosion and 
hydrogen embrittlement. Fuel element cladding temperatures can be increased to higher 
levels, i. e., greater than 650°C, because molybdenum has a significantly higher melting 
point than stainless steels (2610°C versus 1400° — 1450°C). In addition, the resistance to 
radiation damage of molybdenum at service temperatures of 750°C is expected to be su¬ 
perior to ferritic or austenitic steels. This is particularly true with respect to the forma¬ 
tion of helium bubbles. The low nucleation rate of helium bubbles in molybdenum and the 
low diffusion rates of helium to grain boundaries in molybdenum at 650°C are significant 
advantages over stainless steel. 



Fig. 3.5 — Compacted W — 25Re - 30Mo alloy with rubber mold used in isostatically compacting the metal 
powders. The specimen is laying on the mold cover. This pressing, successfully completed by 
Oak Ridge National Laboratory, Y-12 plant in support of this program, was conducted following 
purification treatment by GE-NMPO. (Neg. P67-8-3B) 








187 



a. Sintering voids unclosed in W — 25Re — 30Mo 
powder metal compact hot rolled from 0.82 cm to 
0.53 cm thickness. Material treated 275 hr at 1000°C 
in hydrogen - water vapor atmosphere (5°C dew¬ 
point) prior to sintering 3 hours at 2640°C in hydro¬ 
gen. (Neg. 9421) 


b. Material as in a with hot rolling continued to 0.25 
cm thickness. Treated 10 hr at 1400°C in hydrogen 
after hot rolling. This thickness shows partial void 
retention with 69% hot-work reduction. (Neg. 9422) 



c. Material as in a and b shown after hot rolling to usual 
GE-NMPO practice of 0.1 cm thickness. Sintering 
porosity is believed eliminated — reduction average 
75% in hot rolling. (Neg. 9423) 


d. Results for W — 30Re — 30Mo alloy hot rolled with 
elimination of sintering pores with reductions equiva¬ 
lent to those shown for W — 25Re — 30Mo alloy in c. 
(Neg. 9425) 


Fig. 3.6 — Photomicrograph showing the microstructures developed at major 
stages of the hot-rolling process on sintered W-Re-Mo alloys 
(Murakamis etchant, 100X) 

A major limitation preventing molybdenum from immediate consideration for LMFBR 
applications is the problem associated with ductility, particularly in weldments. Commer¬ 
cial molybdenum rod, tubing, and sheet stock produced by powder metallurgy* techniques 
have been shown to develop a highly porous structure in the heat-affected area of TIG or 
EB weldments. Porosity in fuel element cladding stock is unacceptable. Although sound 
weldments can be produced in commercial arc-cast molybdenum products, the welds are 
brittle. 









188 



D 

&.- .. 


Fig. 3.7 — Welding junction and liquid-metal flow direction control, component machined from 
W — 25Re — 30Mo alloy. Welding of connecting tubes entering from the opposite 
side as shown in the photograph is accomplished through the holes shown in the fore¬ 
ground. The plugs shown close the welding parts. {Neg. P68-2-5C) 


Research under this Task was directed toward obtaining unalloyed molybdenum products 
which will yield ductile non-porous EB welds. The technical approach was based upon im¬ 
proving the purity of the starting metal powders and maintaining the high purity during all 
stages of fabrication. 

MOLYBDENUM PURIFICATION 

The initial work directed toward improving unalloyed molybdenum for nuclear application 
was discussed in the preceding annual report. 2 At that time, treatment of metal powder 
compacts in hydrogen - water vapor atmospheres definitely improved the finished sheet, 
producing good room-temperature bend ductility in weld areas. Chemical analyses of vari¬ 
ous lots of molybdenum produced during CY-67 repeatedly showed the effectiveness of the 
purification treatments in lowering the carbon and oxygen levels of the molybdenum with 
the resultant beneficial effect on weld ductility. Reduction in the level of carbon impurities 
is accomplished by 1000°C, 6-week (1000 hours) treatment of the as-pressed compact in a 
hydrogen — water vapor atmosphere (+5°C dewpoint); carbon generally is reduced from an 
average of 65 ppm to 6 to 12 ppm. Oxygen impurities can be reduced from levels of 2000 
to 3000 ppm in the bulk powders to approximately 40 ppm by a treatment of more than 
64 hours at 1000°C in hydrogen with a low dewpoint (-40°C). Further reduction of oxygen 
is accomplished by similar thermal treatments after hot working operations which expose 
the molybdenum to air. Oxygen analyses of finished sheet were less than 20 ppm. 

Although good room-temperature ductility was achieved in weldments produced in the 
NMPO high-purity molybdenum sheet, microporosity was not eliminated completely from 
the weld-affected areas. During CY-67, particular emphasis was placed on eliminating 

2 „ 

Sixth Annual Report - High-Temperature Materials Program, Part A," GE-NMPO, GEMP-475A, March 31, 1967, p. 113. 






189 


all microporosity and on confirming the adequacy of the procedures developed. In the be¬ 
lief that microporosity in the heat-affected zone of weldments may be caused by volatiliza¬ 
tion of residual trace impurities of relatively high vapor pressure, i. e., alkali metals, 
etc., in the molybdenum sheet during the welding operation, attempts to eliminate or mini¬ 
mize the suspect trace impurities were made by leaching of the bulk powders before con¬ 
solidation with various solutions including HC1, HN0 3 , HN0 3 + HF + HC1, propanol, xylene, 
acetone, and distilled water. Credence to this hypothesis was the fact that porosity was 
not observed in the fused weld zone where the highly volatile metal vapors could escape 
when the molybdenum metal was in the liquid state (2610° C). Significantly it was found 
that weldments made in molybdenum sheet produced from powders which had been leached 
in dilute solutions of HC1 (0.012M) and HN0 3 (0.014M) were free of microporosity; weld¬ 
ments in sheet produced from powders leached in all other solutions listed above all con¬ 
tained various degrees of microporosity. 

To confirm the encouraging initial findings, four different batches of 0.050-cm-thick 
molybdenum sheet were produced using the following purification treatment: 

1. Leaching bulk metal powders in dilute HC1 and HN0 3 acid solutions before compacting. 

2. Exposing as-compacted metal powders for prolonged periods in hydrogen — water vapor 
atmospheres (+5°C dewpoint) at 1000°C to lower the carbon content. 

Electron beam weldments were made on sheet from each batch and all weldments were 
free of microporosity. The micro structure of the heat-affected zone of a weldment free of 
microporosity is shown in Figure 3.8. A room-temperature 2T bend across the weld of a 
purified unalloyed molybdenum sheet is shown in Figure 3. 9. This bend represents the 
most stringent bend test combination considered; both the weld seam and the bend crease 
are parallel to the rolling direction of the sheet. 



Fig. 3.8 - Microstructure of heat-affected zone in electron beam welded, purified 
unalloyed Mo showing absence of microporosity 
(Neg. MA6221, Murakamis etchant, 250X) 




Fig. 3.9 - Microstructure of weld zone in purified molybdenum 0.050-cm-thick 
sheet after 90° bend (2T) at room temperature. Weld seam and bend 
crease are parallel to the rolling direction. 

(Neg. 10102, as-polished, 50X) 

Although sheet produced from metal powders leached by dilute acid treatments have re¬ 
peatedly shown a minimum of microporosity in the EB weld - heat effective zone, presum¬ 
ably the result of decreasing or completely removing trace element impurities, the specific 
trace elements involved have not been positively identified. This is because the overall level 
of impurities is in the low ppm range and changes in these impurities due to dilute acid leach¬ 
ing treatments cannot be detected and isolated by the arc-source mass spectrograph. Fur¬ 
ther studies on the purification process are in process to identify the elements contributing 
to microporosity in weldments. 

Because of the detrimental influence of carbon and oxygen impurities on the properties 
of weldments, it is important to minimize contamination by these elements during the vari¬ 
ous hot and cold working operations in the processing of the sintered compact to sheet. For 
this reason the individual and accumulative change in carbon and gaseous impurities were 
determined after each major operation during processing of 24 hydrogen sintered compacts 
of unalloyed molybdenum to 0.038-cm-thick sheet (12 compacts were produced from powders 
leached in 0.012M HC1; 12 compacts were produced from powders leached in 0.014M HNO 3 ). 
Some of the pertinent results are shown in Table 3. 3. Several important conclusions were 
reached from these studies: 

1. Hot rolling purified molybdenum sheet in air from a hydrogen atmosphere results in 
surface contamination by oxygen, nitrogen, and carbon from the air and mill lubricants. 

2. Cold rolling appears to impregnate the metal surface with rolling lubricants. 

3. A hydrogen treatment (-40°C dewpoint) at 1000° to 1400°C reduces the carbon and gas¬ 
eous impurities from both hot- and cold-rolled sheet. 

To evaluate the influence of the dilute HC1 and HNO 3 acid leaching treatments of the powder 
on the properties of the finished sheet, three groups of companion specimens were produced 
from purified molybdenum powders for bend ductility, hardness, and grain size determina¬ 
tions as a function of stress relief temperature (1100° — 1400°C). Processing procedures 
that were followed for the production of the high-purity sheet used in the evaluation were: 



191 


TABLE 3.3 


CARBON AND GASEOUS IMPURITIES DETERMINED AT VARIOUS PROCESSING STAGES 
IN PRODUCING UNALLOYED Mo a _ 


Sheet 

Identity 

Processing Steps Completed 

Carbon 

Reported Analytical Results, ppm 

Nitrogen Oxygen Hydrogen 

N-1 

Hot-roll {reduce from 0.6 cm to 0.1 cm) 

19.0 

6.0 

77.0 

1.0 

C-13 

Hot-roll (reduce from 0.6 cm to 0.1 cm) 

15.0 

5.0 

41.0 

2.0 

C-17 

Hot-roll (reduce from 0.6 cm to 0.1 cm) 

27.0 

- 

45,9,31 

- 

N-8 

Hot-roll (reduce from 0.6 cm to 0.1 cm) 

9.0 

— 

21,51 

— 

N-1 

Hot-roll + 208 hr at 1050°C in H 2 

13.0 

1.0 

20.0 

1.0 

C-13 

Hot-roll + 208 hr at 1050°C in H 2 

9.0 

4.0 

18.0 

1.0 

C-17 

Hot-roll + 60 hr at 1400°C in H 2 

4.0 

- 

13,37 

- 

N-8' 

Hot-roll + 60 hr at 1400°C in H 2 

9.0 

- 

9,11 

— 

N-3 

Hot-roll + 130 hr at 1000°C in H 2 

5.0 

1.0 

18.0 

0.8 

C-16 

Hot-roll + 130 hr at 1000°Cin H 2 

7.0 

0.6 

16.0 

0.6 

N-3-A 

Hot-roll + 130 hr at 1000°C in H 2 , cold-rolled 
to 0.05 cm 

27.0 

2.0 

22.0 

1.0 

N-3-A 

Plus stress-relieve 20 min at 1000°C, in H 2 

7.0 

1.0 

16.0 

1.0 

C-16 A 

Hot-roll + 130 hr at 1000°C, in H 2 , cold-rolled 
to 0.05 cm 

22.0 

5.0 

35.0 

0.9 

C-16A 

Plusstress-relieve 20 min at 1000°C in H 2 

8.0 

0.6 

10.0 

0.6 


o 

a Mo received as one lot minus 325 mesh-size powder. Vendor reported 1.53 g/cm apparent density by ASTM 
B-325-58 specification, and 1.8 micron average Fisher sub-sieve diameter. Powder contained 23 ppm-C and 
2920 ppm-0 2 impurities according to vendor. 


1. Hydrogen-sintered compacts were hot rolled to 0.075 cm thickness in air out of a 
hydrogen furnace set at 1315°C. The total hot reduction was greater than 75 percent. 

2. The hot-rolled sheet was treated for 175 hours in hydrogen (-40°C dewpoint) at 1000°C 
to remove oxygen absorbed during hot rolling. 

3. The de-oxygenated sheet was cold rolled from 0.075 cm thickness to 0.050 cm thick¬ 
ness for a total of 33-1/3 percent reduction and stress relieved for 20 minutes at tem¬ 
peratures ranging from 1100° to 1400°C in hydrogen. 

Table 3.4 presents the results of the tests conducted on the three companion lots of puri¬ 
fied molybdenum sheet. Simulated EB butt welds were made on sheets from all three lots 
and, as demonstrated repeatedly, no evidence of microporosity could be found in the heat- 
affected zone of weldments made in sheet which was produced from powders leached either 
in dilute HC1 or HNO 3 acids. Bend specimens oriented with the bend parallel and trans¬ 
verse to the rolling direction of both types of purified sheet and stress relieved at 1100°C 
and 1400°C, with one exception, withstood 90-degree bends over a 2T radius. A ram speed 
of 0.005 cm/min was employed. No explanation is apparent for the one specimen (oriented 
parallel to the rolling direction) which failed at an angle of 10 degrees other than possibly 
an edge effect. At a ram speed of 0.127 cm/min all specimens oriented transverse to the 
rolling direction and stress relieved at 1100°, 1200°, and 1300°C also withstood a 90-de- 
gree/2T bend at room temperature. However, considerably less ductility was observed in 
specimens oriented parallel to the rolling direction at the higher strain rate. There appears 
to be no difference in bend properties of the sheet as affected by the acid leaching treatment. 

Grain size determinations indicate that sheet produced from the HNO 3 -treated powders 
is slightly more resistant to recrystallization and grain growth than the sheet produced 
from the HCl-treated powders. This is shown in Figure 3.10 for material stress relieved 
at 1050°C for 20 minutes in hydrogen following cold rolling. At 1050°C, the HNC> 3 -treated 
molybdenum sheet showed incipient recrystallization and the HCl-treated molybdenum sheet 




192 


CO 


00 

£ 


On 
NO I 


ro 

6 


QJ 


< CO 


0^3 
CD O " 


C ^ cc 

111 

ZZ 4-> 

< CO 


CD 4- 
C ™ 

■S „ 

ro QC 

■*- </> 

c/> 


i 2 
£ £ 


i 2 C 


a to 
£ C 


£ ® - 


o 







o 

o 

CD 

03 

o 

CO 

q 

co 

o 

CO 

1 ^ 1 

r-.* 

r-’ | 

r>.‘ 

r^* 


CO 

1 CO 1 

CO 

00 1 

CO 

co 

o 







o 

Q 

CM 

CD 

CM 

o 

o 

o 

o 

d 


co’ 

03 | 

03 

03 

CO 

03 

CO 1 

CO 

00 1 

co 

00 

o 







o 

O 

o 

CM 

o 

CO 

o 

00 

o 

o 

1 05 ! 

03* 

1 °* 1 

03* 

1 00* 

CN 

03 

1 CO 1 

CO 

1 03 1 

00 

1 00 

o 







o 

o 

CO 

CM 

CO 

q 

ID 

ID 

o 

o 


o’ 

1 Q 1 

d 

1 d 


03 

1 03 1 

03 

1 03 1 

03 

1 03 


8 V 

t E : 

CO .0> 


£ 

a: 


h- 

-O 


> E .2 

CoS 
cc •- 
Q 


CO 


1 

2 ‘55 03 ro 
.« o ■§: ® 
5 o c < 

Q_- 


I - 


CD 

I •£ 

8 | 
a ^ 
co 


•S 2 
£ x ro 

=; cd a) 
.E ■= = c 
■o 2. <o o 

m O r — 

^ UJ t QJ 

S 3 - CO 3 - 


l 'O i. 13 

O) OJ CD CD 
w CD cn CD 
□ CD =3 CD 

o c o c 


I O O 
1 03 t- 


O 

2 


CO CO 

_O O 

0 0 2 2 
XXXI 


E 

3 

TJ I 


E 

3 

I T) 


& 

O 

lO -□ 

I 03 

1 X 

— CD 


o 

o 


E a 

- UJ Sill 


CO 


o o o o 

*- 03 T- 03 


o o o o 
03 03 03 03 


O O O O 
CO 03 CM 03 


O O O O 
LO 03 03 03 


O O O 


r~ r- 


CO CO 

_O O 

0 0 2 2 
X X X X 


o o o o 
2 2 2 2 


CO CO 

_O O 

0 0 2 2 
X X X X 


0 0 2 2 


0 0 2 2 


O O 


£ 

o 

o 

o 

o 

■'cr 


I 

o' 

03 

00 

II 

CM 

o 


o 

J. c 

< g) 

CD 2 
CM TO 

I C 


Q. ~ 
CD C 

O t 


£ 2 4= -E 


§ c cd 

2 ^ c 


£ -a 


■a .E 
§ E 
> o 

X) w 
TO C 


« ■“ M 

t' CD W 


i_ ^3 CD 


E .2 
9- a; 


*o 

CD 

E 

C/5 
T3 
C 

^ CD 

m w -a 


•55 ® 
> 


< < 

CD X3 


< f 

O X 


* 




193 



Fig. 3.10 — Photomicrograph of purified Mo before and after stress-relief at 1050°C. Metal 
powders used to fabricate the Mo sheet were treated with the acids indicated 
(Etched in methanol-sulfuric acid solution, (~200X) 


showed some grain growth. As-cold-rolled, the HNO 3 -treated material was fibrous; the 
HC1-treated material contained a few moderate sized grains remaining from the hot-rolling 
operation (1350°C). 

No trends were observed from the hardness data other than the normal decrease as a 
function of stress relief temperature. 

MOLYBDENUM ALLOYS 

The purification technology developed for unalloyed molybdenum for LMFBR applications 
also is being extended to its alloys. 

Three alloys (Mo - 5W, Mo - 3Re - 5W, and Mo - 5Re, all at. %) were processed using 
the NMPO HC1- and wet-hydrogen purification technique which was developed for unalloyed 
molybdenum, as discussed above. 















194 


Sheet stock was prepared from the three alloys to verify the original test data. The HC1- 
treated metal powders were compacted into flat bars, treated 360 hours at 1000° C in a 
hydrogen - water vapor atmosphere (5°C dewpoint), and hot rolled from 0.56- to 0.1-cm- 
thick sheet at 1400° C. The hot-rolled sheet then was treated for 100 hours in hydrogen 
(-40°C dewpoint) at 1000°C to decrease the oxygen absorbed during hot rolling. The de- 
oxygenated sheet was cold rolled to 0.05-cm thickness. Intermediate and final anneals 
were carried out at 1400° C in hydrogen. 

Finished sheet was evaluated for resistance to microporosity in weldments and stress 
rupture properties at 1600°C and 2200°C. Simulated EB butt welds were made in the longi¬ 
tudinal and transverse directions with respect to the rolling direction of the sheet. Radio- 
graphic inspection revealed that only the weldments in the Mo - 5W alloy and unalloyed 
molybdenum control sheet were free of microporosity. Further studies will be required 
to understand the role of small rhenium additions on microporosity. 

Room-temperature bend tests were conducted on weld specimens over a 2T radius at a 
ram speed of 0.058 cm/min to 0.127 cm/min. Specimens from each material successfully 
withstood 90-degree bends when the bend crease was transverse to the weld seam and the 
weld seam was either transverse or parallel to the rolling direction of the sheet. When 
the bend crease was imposed parallel to and directly on the weld seam, bend angles of 
3 5 to 40 degrees were recorded for the Mo-5W alloy, 40 to 50 degrees for the W-3Re—5W 
alloy and 90 degrees for the unalloyed molybdenum control samples with the higher bend 
angle corresponding to the slower bend rate. 


Stress-rupture data at 1600°C and 2200°C (Figure 3.11a) indicate that the NMPO-purified 
Mo - 5W and Mo - 5Re alloys are equal to the rupture strength of commercial Mo - 50Re. 
Creep data at 1600°C and 2200°C to 2 percent and 5 percent strain (Figures 3.11b and 3.11c) 
indicate that the Mo — 5Re and Mo — 5W alloys are more creep resistant than the Mo —50Re 
alloy tested under identical stress - temperature. One-hour isochronal stress-rupture data 
for a number of refractory metals and alloys are given in Figure 3.12. These data show the 
1-hour rupture stress of Mo - 5W and Mo - 5Re is greater at all temperatures (1200° - 
2400°C) than niobium, tantalum, unalloyed molybdenum, and Mo-TZM alloy. 

W-Re-Mo ALLOYS 

The purification treatments developed for unalloyed molybdenum were successfully ap¬ 
plied to W-Re-Mo alloys in a continuing effort to produce refractory alloy mill products 
that have usable room-temperature ductility in the weld area and that are resistant to 
microporosity in heat-affected zones of welds or in the base metal after longtime thermal^ 
exposure. Chemical analyses of W - 30Re - 30Mo seamless tubes that were produced indi¬ 
cate the carbon content to be 1.0 ppm. Similarly a Mo — 35Re tube analyzed 5.0 and 2.0 
ppm carbon on samples taken 150 cm apart. These products had received a 324-hour treat¬ 
ment in hydrogen (5°C dewpoint) at 1000°C. Finished sheet of W-25Re -30Mo and 
W — 30Re — 30Mo similarly treated analyzed 2.0 and 6.0 ppm carbon, respectively. These 
data confirm the reliability of the hydrogen - water vapor treatment in reducing the carbon 
levels in W-Re-Mo alloys. 


3. 3 SUMMARY AND CONCLUSIONS 

/"Equipment was installed for producin g sheet , bar, seamless tubing, and wirl of refrac¬ 
tory metals including an electronically controllec^^lraw bench to control the drawing speeds 
of seamless refractory metal tubing. Fine W — 0.5 Os wire (0.0076-cm diameter) was suc¬ 
cessfully drawn for thermocouple applications 


ions 


~Lr f- / 


n 






























196 



Fig. 3.12 — One-hour isochronal stress-rupture data for various refractory 
metals and alloys 


, Purification procedures developed for molybdenum have resulted in the production of 
unalloyed molybdenum sheet which is ductile and free of microporosity in weld-heat- 
affected areas. A Mo — 5W alloy processed by recently developed purification procedures 
has stress-rupture life and deformation resistance equal to Mo — 50Re alloy at 1600° C, 
and is superior to that alloy at 2200°C. j , 

Several areas of significant accomplishment under this Task during CY-67 will be re¬ 
ported following resolution of patent actions. 

3. 4 PLANS AND RECOMMENDATIONS 

Development and selection of processing procedures for obtaining purified molybdenum 
as nuclear-quality tubing and sheet will be emphasized. Tube-drawing procedures involv¬ 
ing plug drawing will be evaluated for ductile molybdenum. This technique was prohibited 
by the high working stresses required for the stronger W-Re-Mo alloys. 

Inspection and testing specifications for ductile and weldable purified molybdenum mill 
products will be developed. 












197 


Processing of small-diameter tubing of W-Re-Mo alloys for advanced thermocouple 
applications of interest to the AEC will continue. Tube-drawing procedures developed for 
W-Re-Mo alloys will be applied to W - 25Re and Re alloys which are of interest for LMFBR 
thermocouple applications. Processing of fine W — 0. 5 Os thermocouple wire will continue 
for LMFBR applications. 



iG 


EHYSICAL METALLURG Y OF FAST BREEDER REACTOR 
^LAU DING MATERIALS AND REFRACTORY METALS) 

(1177) 

C. 6. Collins,* K. M. Bohlander 


The objective of this program is to define heat treatment, composition, and micro¬ 
structures that result in optimum creep ductility of selected fuel cladding alloys. This 
objective represents a redirection of this task as of January 1968 toward emphasis on 
fast breeder reactor fuel cladding materials. During 1967, the task was concerned with 
the physical metallurgy of refractory metals, specifically W-Re-Mo alloys and molyb¬ 
denum, in support of the alloy research and development efforts under task 1115. Work 
on refractory metals, particularly molybdenum, will continue through the remainder of 
the current fiscal year, with the major effort henceforth devoted to stainless steels and 
other selected fuel cladding alloys noted subsequently. 

4.1 TUNGSTEN-RHENIUM-MOLYBDENUM ALLOYS 

Studies of the W - 25Re - 30Mo and W - 30Re - 30Mo (at. %) alloys were concerned with 
recrystallization, ductility, hardness, and aging characteristics of original process and 
specially purified powder metallurgy materials. Specific objectives were to improve pro¬ 
duction procedures and properties of the sheet product. 

FABRICATION 

Fabrication studies centered on means of incorporating purification procedures without 
impairing yields or properties. As indicated in Figure 4.1, the original fabrication proc¬ 
ess for sheet material of the two alloys involved hot rolling at 1400°C to about 1-mm 
thickness with subsequent 5 percent reduction steps obtained in multiple passes at 200°C 
to final thickness. Intermediate anneals of about 15 minutes at 1400°C were made after 
each reduction step. Subsequently, special purification steps outlined in Figure 4.1 were 
incorporated in the procedure to improve weldability. Unfortunately, the temperatures 
initially utilized in the final purification treatment after hot rolling caused poor yields 
in working and limited room-temperature ductility in both alloys. 

W - 25Re - 30Mo 

Simpler production procedures were established for W - 25Re - 30Mo using the modi¬ 
fied procedures indicated in Figure 4.1. Since this alloy exhibits the brittleness typical 
of refractory metals in the recrystallized condition, the wrought structure necessary for 
cold working could be retained by the following modifications: (1) carrying out the final 
purification treatment at temperatures below that at which recrystallization occurred, 

(2) omission of this final treatment, or (3) carrying out the purification treatment prior 
to the final hot rolling step. With a sufficiently wrought structure, it was possible to 
achieve reductions of 10 percent per pass at room temperature and total reductions of 
up to 80 percent without annealing or stress-relief treatments. With modifications No. 1 


‘Project leader and principal investigator. 



199 


Original 

Process 


Special 

Purification 

Process 


Modified 

Production 

Procedures 



Finished sheet Finished sheet 


Finished sheet 


Finished sheet 


Fig. 4.1 — Production procedures for W-Re-Mo alloys 

and 2 (Figure 4.1), improved yields were obtained in the processing of sheet and the final 
stress-relieved material possessed excellent room-temperature ductility. Present analy¬ 
tical data do not indicate whether the concentration of interstitial impurities, particularly 
oxygen, differs appreciably in sheet produced by modifications No. 1 and 2, i. e., when 
the final purification treatment is omitted. Weld porosity was not adversely affected by 
the latter procedure, however. Pending clarification of this point, studies of modification 
No. 3 have not been pursued. 

W - 30Re - 30Mo 

Similar modifications in sheet fabrication were successfully applied to the W-30Re- 30Mo 
alloy with improvements in yield and ductility. This alloy is ductile in the recrystallized 
condition, although relatively poor ductility properties can arise from precipitation of a 
sigma phase which forms at temperatures below approximately 1700°C. Material recry¬ 
stallized above the sigma solution temperatures was definitely more ductile than that re¬ 
crystallized at temperatures at which sigma precipitated, even though interstitial impuri¬ 
ties may also be involved in the ductility. Modifications in the production process therefore 
involved reduction and stress-relieving treatments identical to those described for the 
W - 25Re - 30Mo alloy (see Figure 4.1), and a final recrystallization anneal at 1700°C or 
higher to obtain a ductile-to-brittle bend transition temperature below -70°C. The final 
anneal also served as a further deoxidation treatment. 

RECRYSTALLIZATION 

Recrystallization studies to aid fabrication development were conducted on both W-Re-Mo 
alloys after cold reduction of sheet from the as-processed condition (produced by the orig¬ 
inal procedure outlined in Figure 4.1) and from the annealed (2000°C, 2 hr) condition. Also 





200 


included were W - 30Re — 30Mo specimens processed using special purification procedures. 
All materials were cold rolled at room temperature to reductions ranging from about 10 to 75 
percent without intermediate anneals. A summary of the specimens prepared for this re¬ 
crystallization study appears in Table 4.1. 

In general, the two alloys exhibited typical recrystallization behavior; a pronounced de¬ 
crease of grain size on recrystallization occurred at reductions greater than 15 to 30 per¬ 
cent. Data established that the final purification treatment at 1400°C for 70 hours after hot 
rolling amounted to a recrystallization anneal, and that the purification treatment should be 
1200°C or lower if recrystallization is to be avoided. 

W - 25Re - 30Mo 


In the as-reduced condition, pre-annealed specimens retained the large equiaxed grain 
structure up to about 13.5 percent cold reduction, at which point a wrought structure be¬ 
came apparent. Grain sizes decreased with increasing amounts of reduction as shown in 
Figure 4. 2, the largest decrease occurring at reductions in the range of 15 to 30 percent. 
Grain size in this instance refers to grain height measured in the thickness direction of 
the material. 

TABLE 4.1 


RECRYSTALLIZATION AND AGING SPECIMENS BEFORE TESTING 


Specimen 

No. 

Pre-Rolling 

Treatment 3 

Reduction, 

% 

Final Sheet 
Thickness, cm 

Hardness, 

DPH b 

Grain Size, 
Width 

microns 0 

Height 

W — 25Re 

1 

— 30Mo 

A 

13.5 

0.046 

508 

62.3 

62 

2 

A 

24.6 

0.039 

525 

49.4 d 

30 

3 

A 

39.8 

0.030 

560 

50.4 d 

24.8 

4 

A 

45.5 

0.027 

560 

51 d 

28.5 

5 

A 

54.0 

0.023 

588 

49.4 d 

19 

6 

A 

59.1 

0.019 

595 

55.7 d < e 

17.2 

7 

B 

33.5 

0.036 

575 

32.2 d ' e 

10.5 

8 

B 

43.3 

0.031 

610 

31.2 d > e 

9.5 

9 

B 

51.2 

0.027 

610 

30.8 d - e 

9.0 

10 

B 

65.1 

0.019 

615 

34.3 d,e 

8.3 

11 

B 

56.9 

0.047 

583 

35.1 d ' e 

7.3 

12 

B 

76.2 

0.026 

625 

32 d ' e 

6.8 

W — 30Re 

1 

— 30Mo 

C 

19.4 

0.042 

525 

94.8 

95 

2 

C 

32.7 

0.034 

553 

33.8 

16.5 

3 

c 

43.1 

0.029 

573 

32.5 d 

13.9 

4 

c 

55.8 

0.022 

578 

31.8 d ' e 

14.2 

5 

c 

66.8 

0.015 

583 

39.5 d ' e 

9.7 

6 

D 

18 

0.043 

554 

54.5 

55 

7 

D 

35.1 

0.033 

563 

42.7 d 

15.1 

8 

D 

45.7 

0.027 

573 

57.4 d 

23.4 

9 

D 

54.8 

0.022 

580 

33.8 d 

11.3 

10 

B 

31.2 

0.036 

600 

115.6 d 

19.3 

11 

B 

41.5 

0.030 

630 

105.3 d ' e 

17.9 

12 

B 

49.8 

0.026 

635 

105.3 d>e 

13.9 

13 

B 

68.3 

0.019 

638 

131.6 d>e 

14 


a A — Annealed 2 hours at 2000°C. 

B — Stress relieved at 1400°C. 

C - Purified, solution annealed for 2 hours at 2000°C. 

D - Solution annealed 2 hours at 2000°C. 

b Average of 3 to 5 indentations with 2.5-kg load using microhardness tester: 
usual spread of values ± 10 DPH. 

c Grain sizes were determined by the linear intercept method on transverse sections. 
^Individual grains elongated. 
e Microstructure heavily worked, wrought. 



201 



Reduction, percent 


Fig. 4.2 - Influence of pre-treatment (2 hours at 2000°C) and amount of 
reduction in thickness by cold rolling on W —25Re-30Mo 
grain size 


Isochronal anneals of 0.5 hour at 1200°, 1300°, 1400°, 1600°, 1800°, and 2000°C for both 
the as-processed and pre-annealed materials produced complete recrystallization at 1800°C 
and 2000°C regardless of the amount of reduction; at 1600°C for reductions of 40 percent or 
more, and at 1400°C for a reduction of 90 percent. Grain sizes after recrystallization de¬ 
creased markedly at reductions of 20 to 30 percent with relatively small additional decreases 
for the larger amounts of reduction. Data for recrystallization at 1800 P C shown in Figure 
4. 3 are representative of other temperatures. 

Isothermal anneals for various times of material reduced 60 percent (a region of parti¬ 
cular interest in fabrication) indicated essentially complete recrystallization at 1400°C in 
8 to 9 hours, at 1300°C in approximately 100 hours, and at 1200°C in about 1000 hours. 
Measurements of the amount of recrystallization at various temperatures indicated that 
the activation energy for recrystallization was approximately 100 kcal. 

W - 30Re - 30Mo 

The recrystallization behavior of the W — 30Re — 30Mo alloy in the different purities and 
conditions examined was similar to that of the W — 25Re —30Mo alloy. In grain sizes obtained 
in 0.5-hour anneals at 1800°C (Figure 4.4) the decrease in the recrystallized grain size 
after 10 to 30 percent reduction is similar to that of the W — 25Re — 30Mo alloy. Although 
the different starting conditions led to differences at low reductions, the grain size after 
annealing was roughly equal at reductions greater than 30 percent; hence the recrystalli¬ 
zation behavior was not markedly influenced by purification or solution treatments or by 
the presence of some sigma phase. For reductions greater than 30 percent, recrystalli¬ 
zation was complete after 0. 5 hour anneals at 1600°C or above and after 2 hours at 1400°C. 




202 



Fig. 4.3 - Grain size of W - 25Re - 30Mo as a function of reduction in 
thickness after 0.5 hour at 1800°C recrystallization anneal 


Hardness 

The hardness of the two alloys increased rapidly with increasing amounts of reduction, as 
indicated in Figures 4. 5 and 4.6, but reductions greater than 40 percent produced little ad¬ 
ditional change. Relatively little effect of the 2-hour pre-anneal at 2000°C remained after 
about 20 percent cold work. In the W — 30Re — 30Mo alloy, the as-processed material at¬ 
tained a higher hardness with reduction than the annealed materials; this behavior is attri¬ 
buted to a rhenium-rich sigma phase present in this alloy in the as-processed condition. 
Hardness changes due to annealing and aging are noted subsequently. 

AGING 

Aging studies of the W — 25Re — 30Mo and W — 30Re — 30Mo alloys were conducted to assess 
grain growth and hardness changes of the original process materials and specially purified 
W - 30Re - 30Mo. Specimens were in the same pre-annealed, as-processed, and cold- 
worked conditions utilized in the recrystallization studies (Table 4.1). Measurements were 
completed on the two alloys after aging to 1000 hours at 1400°, 1600°, and 1800°C, and to 
500 hours at 2000°C in helium. 

Grain Growth 

In general, the grain growth behavior of the two alloys was comparable. The major dif¬ 
ference noted for the different purities and structures was that the materials pre-annealed 
two hours at 2000°C prior to reduction were less susceptible to secondary recrystallization. 

In the W - 25Re - 30Mo alloy, the pre-annealed material exhibited good grain size stabi¬ 
lity at all aging temperatures and times, except for one specimen reduced only 5 percent 
which exhibited secondary recrystallization after 500 hours at 2000°C. Fairly typical re¬ 
sults for these specimens are shown in Figure 4. 7 for aging at 1800°C. The maximum 
grain size attained in aged specimens that had been reduced 40 percent or more was in 




203 



0 10 20 30 40 50 60 70 80 90 

Reduction, percent 


Fig. 4.4 - Grain size of W -30Re-30Mo as a function of reduction in 
fhickness after 0.5-hour recrystallization anneal at 1800°C 

the range of 40 to 50 microns. The lightly reduced specimens attained maximum grain 
sizes ranging up to 100 microns. 

In contrast, the W - 25Re - 30Mo specimens in the as-processed condition, i. e., those 
with a wrought structure obtained by 5 percent reductions with anneals at 1400 C, exhibited 
secondary recrystallization in amounts roughly proportional to the amount of reduction. At 
1600°C, only the specimens reduced 51 and 65 percent were affected. At 1800°C all speci¬ 
mens were affected, although the amount of secondary recrystallization in the two smallest 
reductions was not large. The 1800°C results are shown in Figure 4.8. 

The overall behavior of the W - 30Re - 30Mo alloy including the specially purified ma¬ 
terial was essentially the same as that described for the W — 25Re — 30Mo alloy. As indi¬ 
cated in Figures 4.9 and 4.10, one specimen of the standard process material exhibited 
secondary recrystallization, whereas none was observed in the specially purified material. 
Secondary recrystallization occurred at 1800°C in the as-processed W - 30Re - 30Mo in 
all specimens reduced more than approximately 30 percent. 

Secondary Recrystallization 

Secondary recrystallization is a fairly common problem in both powder-metallurgy and 
arc-melted refractory metals. Its occurrence can cause loss of ductility and changes in 
other mechanical and physical properties. As with ordinary grain growth, the driving force 
for the process is the surface energy of grain boundaries. Although the causes are not well 



204 



Fig. 4.5 — Hardness changes of W — 25Re — 30Mo alloy as a 
function of reduction in thickness by cold rolling 



Fig. 4.6 — Hardness changes of W - 30Re - 30Mo alloy as a 
function of reduction in thickness by cold rolling 



2 


<2 70 



IMIIII 


IK =mm§ 




iSisgiiiissS 


O 14% reduction 
0 25% reduction 
O 40% reduction 
A 46% reduction 
□ 54% reduction 
^7 59% reduction 


Time, hours 

Fig. 4.7 - Grain size changes as a function of time at 1800°C for W-25Re-30Mo 
alloy pre-treated for 2 hours at 2000°C in H 2 before cold rolling and 
recrystallization anneals 


. 700 microns 


Q 34% reduction - 

O 43% reduction 

A 51% reduction - 

□ 65% .reduction 

V 76% reduction -- 

i I I I I I 

S-secondary recrystallization 




IBS 




Fig. 4.8 — Grain size changes as a function of time at 1800°C for W—25Re—30Mo 
alloy. Samples were original process material cold-rolled from the as- 
fabricated condition. 



























Grain size, m/crons Grain size, microns 


206 



Fig. 4.9 — Grain size changes as a function of time at 1800°C for standard process 
W — 30Re — 30Mo solution annealed for 2 hours at 2000°C before cold 
rolling and recrystallization anneals 



Fig. 4.10 - Grain size changes as a function of time at 1800°C for purified W-30Re-30Mo 
alloy solution annealed for 2 hours at 2000°C before cold rolling and recrystal¬ 
lization anneals 



























207 


defined, purity, preferred orientation, and deformation are known to influence the phenom¬ 
enon. 

As noted in the preceding paragraphs, several specimens of both W-Re-Mo alloys ex¬ 
hibited areas of secondary recrystallization after aging at 1600°C or higher. Further inves¬ 
tigation indicated that this occurred in about half the production sheets and confirmed the 
aging studies which indicated that it was related to amount of reduction and to purity. No 
definitive cause or complete solution to the problem was found. 

In examinations of recent production lots of sheet materials of the two alloys, second¬ 
ary recrystallization occurred in five of ten sheets produced by the original rolling pro¬ 
cedure (small cold reductions with intermediate stress-relief anneals at 1400°C). Second¬ 
ary recrystallization areas were found in four specimens after 17 hours at 2000°C in hy¬ 
drogen and in a fifth specimen after an induction period of about 60 hours at 2000°C. In 
those materials susceptible to the phenomenon, there appeared to be a direct relationship 
between the amount of cold working and the occurrence of secondary recrystallization, as 
in the grain growth measurements described previously. This relationship was particular¬ 
ly apparent upon aging specimens that were cold rolled from 5 to 75 percent with no inter¬ 
mediate anneals, and is probably caused by strain-induced grain growth since secondary 
recrystallization is enhanced by both light and heavy plastic deformations. 1 ’ 2 It is not 
solely a function of the amount of deformation, however. Some W-Re-Mo specimens have 
been cold reduced 50 to 75 percent without secondary recrystallization occurring on sub¬ 
sequent aging treatments, whether with or without anneals during reduction. 

The possibility that the impurity concentration or inclusions - or rather the lack of 
these - is a factor in secondary recrystallization was investigated with some success. 
Indirect evidence is the fact that materials produced some time ago, with a reportedly 
higher impurity concentration, 3 showed no secondary recrystallization in aging studies. 4 
Furthermore, although there is no statistical sampling, • se condary recrystallization has 
only recently been observed in work under various other tasks using these materials. 5 
Direct evidence was obtained when specimens from a large sheet of W - 30Re - 30Mo, 
known to undergo extensive secondary recrystallization, were heated in a nitrogen at¬ 
mosphere at temperatures ranging from 1000° to 1200°C and times from 3 minutes to 
3 hours, annealed in argon at 2000°C for 2 hours, then cold rolled to 60 percent reduc¬ 
tion in thickness. These specimens did not form secondary growth when aged for 100 
hours at 2000°C even though some were strained by bending. In contrast, the control 
specimen without the nitrogen treatment formed extensive secondary growth within 17 
hours at 2000°C in the same test. Chemical analyses failed to establish clearly that the 
nitrogen content was changed; both the nitrided and control samples indicated 10 to 20 
ppm nitrogen. Considering that not all production materials exhibited secondary recrys¬ 
tallization, these data suggest the existence of a critical concentration of impurities which 
may prevent the phenomenon. For example, Fiedler 6 found that nitrogen concentrations 
in excess of 0.0184 percent inhibited secondary recrystallization in silicon-iron. 


1 C. D. Calhoon, "Exaggerated Grain Growth in Reactor-Grade Hafnium After Small Deformations," KAPL-3193, 

September 2, 1966. 

2 K. T. Aust, "Crystal Growth from the Solid State," Genera! Electric Research and Development Center, Report No. 66-C-294, 
September 1966, p. 11. 

3 "AEC Fuels and Materials Development Program Progress Report No. 67," GE-NMPO, GEMP-67, June 30,1967, p. 65. 

^"Sixth Annual Report — High-Temperature Materials Program, Part A," GE-NMPO, GEMP-475A, March 31,1967, pp.133—135. 

^"710 Reactor Program Progress Report No. 24," GE-NMPO, GEMP-529 (Conf.), July 31, 1967, pp. 37—42. 

6 H. C. Fiedler, "The Behavior of Nitrogen in 3.1% Silicon-Iron," General Electric Research and Development Center, Report 
No. 67-C-225, June 1967. 



208 


Efforts to relate the secondary recrystallization to local inhomogeneities or to preferred 
orientation were not successful. Local inhomogeneities of the order of 1 percent concen¬ 
tration variations can be detected in the materials by X-ray fluorescence techniques. These 
areas probably extend only to thicknesses of a few grains in diameter since both etching 
and polishing resulted in variations of the composition of the same spot of 3- to 6-mm di¬ 
ameter. Although the measurements may not be conclusive, no correlation between these 
inhomogeneities and the secondary recrystallization was obtained comparing sheets known 
to be susceptible to it and free of it. Similar negative results were obtained in X-ray dif¬ 
fraction analyses of preferred orientation. At present there is no way of detecting sus¬ 
ceptibility to the phenomenon other than actual aging tests. 

Hardness Changes 

The general trend of the hardness changes of the purified W - 25Re - 30Mo and 
W - 30Re - 30Mo alloys was almost identical to that in material not specially purified, al¬ 
though the purified alloys were about 20 VPH units higher (400 versus 380) than the normal 
material for the softest condition. Data up to 1000 hours are summarized in Figure 4.11 
for specimens aged at 1200 to 2000°C. As noted previously, the increase in hardness of 
W - 30Re - 30Mo alloy at 1400°C and 1600°C at the longer aging times is attributed to the 
sigma phase that develops in amounts up to about 7.5 percent at these temperatures. 




Fig. 4.11 — Change in room-temperature hardness of cold-rolled W-Re-Mo 
alloys after aging at elevated temperatures 


DUCTILITY 

Bend ductility measurements under a variety of heat treatment and microstructural con¬ 
ditions established that good room-temperature ductility could be achieved in W-25Re-30Mo 
with stress relieving at 1150°C following about 80 percent cold reduction or following a com¬ 
bination of hot and cold reduction of recrystallized material. This behavior was a factor in 
the selection of modified production procedures outlined in Figure 4.1. Similar treatment 




209 


provided only moderate ductility (40- to 50-degree bends at room temperature) in the 
W - 30Re - 30Mo alloy. This material is ductile in the recrystallized condition, however, 
provided recrystallization is carried out at temperatures above that at which the sigma 
phase is in solution, approximately 1700°C. 

Bend ductilities were determined by means of the 4T bend test. Because of the sample 
quantities required, the earlier rather than the later definition of this bend test recom¬ 
mended by the Materials Advisory Board was followed. 7 Bends were made both parallel 
and normal to the rolling direction on 0. 05-cm-thick sheet specimens. 

The bend ductility of the two alloys was measured as a function of amount of cold reduc¬ 
tion. Prior to reduction, the materials were annealed 2 hours at 2000°C, a recrystalliza¬ 
tion anneal for both alloys. They were then cold rolled to reductions in thickness ranging 
from 10 to 60 percent. Bend tests (4T) were made on the as-reduced sheet parallel to the 
rolling direction. Results showed decreases in ductility of the W — 30Re - 30Mo with in¬ 
creasing amounts of reduction up to about 35 percent. The ductile-to-brittle transition 
temperatures (DBTT) were still below room temperature after 10 percent reduction, but 
with 35 percent reduction, room-temperature bends (parallel) of only 10 to 20 degrees 
were obtained. No further reduction of ductility occurred with additional cold reductions 
in thickness up to 60 percent. The ductile-to-brittle behavior of W - 25Re - 30Mo alloy 
differed from that of the W - 30Re - 30Mo in that it did not change after the 2000°C re¬ 
crystallization anneal; the DBTT remained at 250° to 300°C. There was no apparent re¬ 
duction in ductility in this alloy as a function of increased cold working. 

Ductility changes as a function of the annealing temperature following various amounts 
of cold reduction of the recrystallized structure indicated that approximately 80 percent 
reduction was necessary to attain good room-temperature ductility (with 90-degree bends) 
with non-recrystallization anneals of 0.5 hour at 1150°C. Recrystallized specimens 0.5 mm 
thick were subjected to room-temperature bend tests parallel to the rolling direction after 
50 percent reduction. Results following 0.5-hour anneals at temperatures from 800° to 
2000°C in 200°C steps (Figure 4.12) show an increase in bend ductility of the W-30Re-30Mo 
alloy with increasing annealing temperature; 90-degree bends or greater were obtained at 
or above 1800°C. For the 0. 5-hour time period involved, 1800°C was the lowest tempera¬ 
ture of those examined that recrystallized the material and did not result in precipitation 
of sigma phase. The W - 25Re - 30Mo alloy increased in ductility to a maximum 30-degree 
bend after 0.5 hour at 1600°C and did not increase beyond this value at higher temperatures. 

Ductility changes on aging the alloys are essentially predictable on the basis of recrys¬ 
tallization and sigma formation. The W - 30Re - 30Mo alloy is ductile in the recrystallized 
condition; hence aging at 1600°C or higher temperatures at which sigma does not precipi¬ 
tate does not induce brittleness, although secondary recrystallization in some cases reduced 
the bend angles at room temperature to 50 to 70 degrees. At 1400°C and lower, sigma for¬ 
mation in this alloy apparently contributed to embrittlement even when recrystallization 
occurred and parallel room-temperature bends of only 10 to 20 degrees were obtained after 
aging times greater than about 70 hours. The W — 25Re — 30Mo alloy is brittle in the re¬ 
crystallized condition; hence aging for sufficient time at temperatures greater than about 
1200°C resulted in parallel bends of only 10 to 20 degrees. Somewhat surprisingly, aging 
periods of 7 to 25 hours at 1800°C resulted in 70-degree parallel bends in this material; 
this relatively good ductility region was not observed at 1600°C or 2000°C. 

The DBTT for different production procedures are summarized in Table 4.2. Other than 
material produced by the improved production procedure, which had a DBTT of less than 

7The earlier bend definition was given in MAB-176-M, "Evaluation Test Methods for Refractory Metal Sheet Materials," 

November 1961. The later definition is given in MAB-192-M (same title) April 1963. Both reports by National Academy 
of Sciences, National Research Council, Washington, D.C. 




210 



Fig. 4.12 - Influence of isochronal (0.5 hour) anneals in H 2 on the 
bend ductility of two W-Re-Mo alloys cold-rolled 50 per 
cent to 0.51-mm thickness before annealing 


-70 C, the W — 25Re — 30Mo yielded 10- to 20-degree bends at room temperature under all 
conditions investigated. The normal sharp transition from ductile-to-brittle bending was 
evident usually from 50 to 90 degrees rather than from 0 to 90 degrees. In all cases other 
than the improved process material, bends parallel to the rolling direction had a higher 
transition temperature than those normal to the rolling direction. Parallel bends of ap¬ 
proximately 90 degrees were obtained in the original process material at approximately 
300°C for the as-processed, for the 2000°C anneal, and for the aged conditions, although 
some change in the shape of the curve occurred with these different treatments. Ninety- 
degree bends normal to the rolling direction occurred at approximately 150° to 200°C and 
were approximately independent of the condition of the material. 

As with the W- 25Re - 30Mo alloy, the improved production procedure yielded W - 30Re - 30Mo 
material with a DBTT of less than -70 C. The ductile-to-brittle behavior of normally proc¬ 
essed W - 30Re - 30Mo alloy, however, differed from that of the W - 25Re - 30Mo alloy, 
particularly with respect to the effects of 2000°C anneals and aging treatments at 1400°Cor 
anneals at°2000°C. As indicated in Table 4. 2, the original process material had a DBTT of 
100° to 150°C (normal to the rolling direction) in the as-produced condition, somewhat lower 
than that of the W - 25Re - 30Mo alloy. In the specially purified material, the DBTT for 
bends normal to the rolling direction was about 400° to 500°C; that for bends parallel to the 
rolling direction was greater than 500 C. This difference in behavior of the purified ma¬ 
terial is attributed, at least partly, to the sigma developed as a result of the final purifi¬ 
cation treatment. This conjecture is supported by the results for materials annealed 2 
hours at 2000°C to place the sigma in solution; in this condition both the original process 
and purified materials had a DBTT of less than -70°C, appreciably lower than that of the 
original process material. Aging of the original process and purified materials at 1400°C 
raised the DBTT higher than 500 C in both the as-produced and solution-treated conditions; 
again, presumably because of the growth of the sgima phase at this temperature. 

4.2 MOLYBDENUM 

Studies of NMPO-processed molybdenum under this task are directed toward improving 
fabrication and properties of powder metallurgy material. Since molybdenum is brittle in 
the recrystallized condition, its fabrication in high-purity form presents the same problem 




TABLE 4.2 


BEND TEST DUCT!LE-TO-BRITTLE TRANSITION TEMPERATURES 
OF W-Re-Mo ALLOYS 


Alloy 

Production 

Process 3 

Treatment 

Bend 

Direction* 3 DBTT, °C c '^ 

W - 25Re - 30Mo 

Original Process 

15 min final 
anneal at 1400°C 

P 250 - 300 

N 150-175 

W — 25Re — 30Mo 

Original Process 

520 hr at 1400°C 

P 200 - 250 

N 100-125 

W — 25Re — 30Mo 

Original Process 

2 hr at 2000°C 

(30° at 150 

P (90° at 300 

N 150-190 

W - 25Re - 30Mo 

Original Process 

2 hr at 2000°C, 

520 hr at 1400°C 

120° at 200 

190° at 300 

120° at 100 

N '90° at 190 

W - 25Re - 30Mo 

Improved Process 

0.5 hr at 1150°C, 

P <-70 


Modification No. 2 

stress relief 

N <-70 

W - 30Re - 30Mo 

Original Process 

As-produced, 15 min 
final anneal at 

1400°C 

150° at 25 

P (90° at 150 

N 100 -r 150 

W - 30Re - 30Mo 

Original Process 

2 hr at 2000°C 

150° at 25 

P (90° at 100 

N -50 

W - 30Re - 30Mo 

Original Process 

As-produced, 520 hr 
at 1400°C 

P -410 

(10° at 200 

150° at 500 

W - 30Re - 30Mo 

Original Process 

2 hr at 2000°C then 

520 hr at 1400°C 

110° at 200 

P (30° at 350-500 

(10° at 200 
(40° at 500 

W— 30Re — 30Mo 

Special Purified 

As-produced, 15 min 
final anneal at 1400°C 

110° at 200 

P 125° at 500 

4 10° at 100 

N j 45° at 400 

• 90° at 500 

W - 30Re - 30Mo 

Special Purified 

2 hr at 2000°C 

o o 

V V 

Q. Z 

W - 30Re - 30Mo 

Special Purified 

1 hr at 1700°C 

P <-70 


Modification No. 2 

or higher 

N <-70 

W - 30Re - 30Mo 

Special Purified 

As-produced, 520 hr 
at 1400°C 

/10° at 200 

P 130°—40° at 500 

/15° at 200 

N (30°-40° at 500 

W - 30Re - 30Mo 

Special Purified 

2 hr at 2000°C, then 
520 hr at 1400°C 

(10° at 200 

P l40°-50° at 500 

(25° at 200 

N 145° at 500 


a Refer to Figure 4.1 for outline of the different production processes. 

b P = bend made parallel with the rolling direction; N = bend made normal to rolling direction. 
c The DBTT was based on a 90-degree bend in 4T bend tests on 0.05-cm-thick sheet. The 4T bend 


test was as defined in MAB-176-M. 

^Entries with degree symbol indicate bend angle at temperature cited. 



212 


described previously for W — 25Re — 30Mo; specifically, avoiding recrystallization condi¬ 
tions in final purification treatments. The principal options are the three modified produc¬ 
tion procedures outlined in Figure 4.1, as applied to molybdenum. 8 Pending better definition 
of the important impurity concentrations, efforts have been concerned mainly with modifi¬ 
cation No. 2, i. e., omission of the final purification treatment. For this procedure the re¬ 
duction, recrystallization, grain growth, and ductility measurements described in the fol¬ 
lowing paragraphs indicated that stress-relieving treatments of 0.5 hour at 1000°C after 
40 percent reduction at room temperature yielded sheet material with a DBTT below -70°C 
and good grain size stability on aging. 

FABRICATION 

The powder metallurgy molybdenum is readily cold rolled in the wrought condition in¬ 
duced by hot rolling sintered compacts from 1.1-cm to 0.1-cm thickness; reductions of 
up to 25 percent per pass are achieved at room temperature. As illustrated in Figure 
4.13, material in the as-hot-rolled condition 8 could be satisfactorily cold reduced up to 
60 percent with and without stress relief treatments of 0. 5 hour at 1000°C including cross 
rolling, but fracturing occurred if the material was recrystallized. Cold reductions up to 
80 percent without stress relieving were achieved but the sheet contained laminations. 

Some lamination occurred at 60 percent reduction, but 40 percent reductions appeared 
sound. The sheet was successfully cross rolled provided a stress-relief treatment pre¬ 
ceded changes in rolling direction. 

RECRYSTALLIZATION 

A few recrystallization anneals of cold-reduced sheet material were made to verify that 
the powder-metallurgy molybdenum conformed to existing data on recrystallization. 9 Speci¬ 
mens of "N"- and "C"-processed molybdenum sheets were annealed at 1000° to 1200°C in 
the as-hot-rolled condition and after room-temperature reductions of 20, 40, 60, and 80 
percent without intermediate stress-relief treatments. ("N" and "C" materials refer to 
nitric or hydrochloric acid leaching of powders prior to compaction. ) 10 In 1-hour anneals 
of the as-hot-rolled condition, recrystallization of the "N" material was approximately 
50 percent complete at 1100°C and complete at 1200°C; the "C" material was about 50 
percent recrystallized at 1050°C and complete at 1100°C. These recrystallization tem¬ 
peratures were lowered only about 50°C by the additional room-temperature reductions, 
evidently adding little to the wrought structure developed in the hot rolling. 

DUCTILITY 

Bend ductility measurements indicated reasonable ductility in the as-reduced condition 
and very good ductility after stress-relieving treatments. All bends were made parallel to 
the rolling direction on strip 6.4 mm wide. Tests were made at room temperature on a die 
with 6.4-mm span and a 2T radius punch operated at 1.27 mm per minute until the first in¬ 
dication of failure or until a 90-degree bend was attained. In the as-reduced condition, the 
hot-rolled material bent 55 degrees and the materials cold rolled 40 and 60 percent bent 
approximately 40 and 20 degrees, respectively. After stress relieving 0.5 hour at 1000°C, 
the hot-rolled material bent 80 to 90 degrees and the cold-rolled materials bent 90 degrees 
or more. Recrystallized material yielded bends of about 5 degrees regardless of reduc¬ 
tion history. 

8 GEMP-475A, p. 114. 

9 " „ 

Refractory Metals and Alloys, M. Semchyshen and J. J. Harwood, editors, Interscience Publishers, New York, N.Y., 

1960, pp. 290-293. 

10 "AEC Fuels and Materials Development Program Progress Report No. 71," GE-NMPO, GEMP-1002, December 29,1967, 
pp. 65-66. 



*•. r* 1 • - 


from 1200°c H 2 furnace 
1.02 mm thick sheet 


Unalloyed powder-metallurgy molybdenum reduced at room temperature. 
Preferred procedure involves 0.5 hour at 1000°C stress-relief treatment. 











214 


Additional measurements established that the DBTT of the hot-rolled material given a 
subsequent 40 to 60 percent reduction at room temperature and stress relieved 0.5 hour 
at 1000°C was below -70°C. For material with this same reduction, a stress relief of 0.5 
hour at 1100°C raised the DBTT to 0°C. 

AGING STUDIES 

The "C" and ”N" materials were aged in hydrogen for 1000 hours at temperatures of 
1200 to 2000 C to assess the stability of the structures developed in processing. As in the 
other studies, the specimens were obtained from as-reduced hot-rolled material that was 
subsequently cold reduced up to 80 percent. 

All materials exhibited very good microstructural stability at temperatures to 1400°C, 
and those reduced 20 percent prior to aging were stable at 1600°C. In this stable tempera¬ 
ture region grain growth appeared normal; e. g., the specimen reduced 60 percent recrys¬ 
tallized at 1200°C to a grain size of 27 microns which, on aging 1000 hours at this tempera¬ 
ture, increased to 39 microns. Secondary recrystallization occurred at 1600°C in materials 
reduced 40 percent or more and in all materials aged at higher temperatures. 

4. 3 FAST REACTOR FUEL CLADDING ALLOYS 

The program on fast reactor fuel cladding alloys is concerned with varying heat treat¬ 
ment of the alloys to obtain better ductility in creep-rupture tests and, ultimately, under 
irradiation. Principal areas to be pursued initially are fine grain size materials, various 
carbon distributions obtained in work - heat-treatment combinations and in single and 
double aging treatments, and some work on cavitation fracture. Evaluation of the heat 
treatments will be based primarily on creep and rupture tests of pressurized tube speci¬ 
mens and of sheet specimens. All tests will be conducted in argon. Tensile tests as a 
function of strain rates extending to very low strain rates will be utilized as a supplemen¬ 
tary means of evaluating heat treatments. Tests will be conducted mainly at 650°C; se¬ 
lected materials will also be investigated at 538°C and 760°C. 

Materials included in the program are the following: 

316 stainless steel 
316 + Nb stainless steel 
Incoloy 800 
12R72HV * 

19-9DL 

Vanadium alloys 11 
HSV 207 
HSV 208 
HSV 209 
V — 15Cr — 5Ti 
Haynes 56 

Initial efforts are concerned with 316 stainless steel and 19-9DL. 
f4. 4 SUMMARY AND CONCLUSIONS 

’ - ?—V - u - 

Studies on W - 30Re - 30Mo and W - 25Re - 30Mo (at. %) alloys led to simpler, lower- 
cost production procedures and greatly improved room-temperature ductility properties 
for powder metallurgy materials. Data were obtained on recrystallization, grain size,^ 

y “ 

"Product of Sandvik Steel, Inc., Fairlawn, N.J. Nominal composition is Fe-15Cr-15Ni-1.2Mo-0.45Ti-1.8Mn-0.5Si-0.1C. 

^"Vanadium Cladding Alloy Development, Quarterly Progress Report December 31, 1967," Westinghouse Electric Corp., 

Adv. Reactors Division, Madison, Pennsylvania, WARD-3791-13. 



215 


\+y i kJ 

hardness, and ductilit y changes on aging of these alloys at temperatures to 2000 C for 
periods up to 1000 hours. 

Similar studies have resulted in procedures for producing molybdenum sheet; parts of 
this procedure are still under s tudy 

4.5 PLANS AND RECOMMENDATIONS 


Work on fast reactor fuel cladding alloys will include investigation of heat treatment and 
work combinations on 316 stainless steel, and 19-9DL in terms of ductility changes as a 
function of strain rates; creep-rupture tests will be initiated on selected materials. Other 
materials will be included in the studies as they become available. 

Effort on the refractory metals will concentrate in two areas: (1) definition of impurity 
concentrations obtained in molybdenum processed by modifications No. 1 and 2 through in¬ 
ternal friction and improved chromatographic analysis, and (2) completion of weld integrity 
and porosity studies currently in progress to define material purity requirements and weld¬ 
ing conditions. 



5.[fast breeder reactor > 

FUEL ELEMENT CLADDING RESEARCH 
0119 ) " 

F. C. Robertshaw,* J. L. Bartos 


The objective of this program is to develop improved fuel element cladding materials 
for fast breeder reactors and to define the properties of these materials which are of im¬ 
portance for fast breeder reactor applications. 

More than half the material prepared for this report period under this task has been 
withheld pending patent clarification, at the request of the AEC Oak Ridge patent office. 

The alloy systems selected for study under this program are shown in Table 5.1, to¬ 
gether with an indication of their potential utility in various coolants. The specific alloys 
studied include a number which originated during the course of this program and others 
which were developed previously for other purposes. 

TABLE 5.1 


ALLOY SYSTEMS SELECTED FOR STUDY AS 
FAST BREEDER REACTOR CLADDING MATERIALS 


Alloy System 

Applicable Coolants 

Cr-base 

Na, inert gases 

Fe-Cr-AI-Y alloys 

Na, steam, C0 2 , inert gases, air 

Fe-base (ferritic) 

Na, inert gases 

Fe-base (austenitic) 

Na, inert gases 


Tin the alloy studies conducted during the past year, primary emphasis was on developing 
alloys which have mechanical properties comparable or superior to austenitic stainless 
steel. Exploratory work was conducted to examine other properties and characteristics 
which are important in fast breeder cladding alloys; notably, environmental compatibility, 
radiation effects, and processability^ 

f».l C HRQMIUM - BAS E ALLOYS j _^ p . &dO 

Although specific data were previously unavailable, interest in chromium stemmed 
from the prospects for high strength and sodium corrosion resistance at temperatures of 
interest for fast breeder reactor fuel cladding. It was thought that, unlike most other appli¬ 
cations for which chromium-base alloys had been studied, operation in sodium or relative¬ 
ly inert atmospheres would minimize corrosion problems as well as the absorption of in¬ 
terstitial elements which cause embrittlement. Previous NMPO experience in the prepara¬ 
tion of chromium-base alloys indicated that fabricability problems would not be prohibitive. 
It was concluded that chromium might play a useful role in fast breeder reactor technology; 
efforts were undertaken to characterize it at relevant temperatures and environments. 


*Project leader and principal investigator. 



217 


Three alloys, Cr-Y, Cr-Hf-Th-Y, and Cr-TZC-Y, were selected for evaluation; previ¬ 
ous experience had proven them relatively ductile and fairly strong. It was soon discovered 
that some degree of warm working was necessary to achieve strength levels better than 
those of stainless steel. Moreover, prepared in a practical way, these alloys had transi¬ 
tion temperatures in bending and impact higher than room temperature. This has led to 
two other studies, one of which is to establish whether improved ductility can be achieved 
with high-purity starting materials and processing modifications. The other study is to 
develop new, or to evaluate existing, high-strength chromium alloys to determine how 
severe a ductility penalty is associated with the higher strength. 

Warm-worked primary alloys and high-strength chromium-base alloys are being devel¬ 
oped, and special procedures are being employed to improve ductility. 

MATERIAL PREPARATION 

Material used to evaluate the three primary chromium-base alloys was provided by 5.4-kg 
vacuum-induction-melted heats (M-313, M-314, M-315). Chemical analyses are listed in 
Table 5.2; fabrication procedures were presented previously. 1 


TABLE 5.2 

CHEMICAL ANALYSIS OF CHROMIUM-BASE ALLOYS 
_ Composition_ 


wt % _ppm 


Alloy 

Heat No. 

Y 

Th 

Hf 

Zr 

Ti 

C 

°2 

n 2 

h 2 

S 

Si 

Ca 

P 

Cr-Y 

M-313 

0.097 

— 

— 

0.001 

- 

413 

18 

77 

5 

<10 

200 

10 

39 

Cr-Hf-Th-Y 

M-314 

0.13 

0.074 

0.24 

0.007 

- 

160 

15 

90 

6 

<10 

70 

10 

8 

Cr-TZC+Y 

M-315 

0.12 

- 

- 

0.37 

0.12 

349 

12 

26 

7 

<10 

10 

10 

3 


Studies relating to higher-strength chromium-base alloys included preparation of binary 
chromium-base alloys by button arc melting to study the individual effects of alloying addi¬ 
tions on the strength of chromium. A plot of as-cast room-temperature hardness versus 
alloy content (Figure 5.1) reveals the significant increase in hardness afforded by vanadium. 

Based largely on these results, two 3.2-kg vacuum-induction-melted ingots were pre¬ 
pared and extruded. Table 5.3 indicates composition and extrusion data. After extrusion 
to 2.54-cm diameter, segments of alloy M-300 were encapsulated in mild steel and suc¬ 
cessfully press-forged and hot rolled at 980°C to 0.254-cm thickness. The Cr —20V—0.36Y 
(at. %) alloy was cold rolled 10 percent per pass with no intermediate anneals to 0.1'96 cm DiQ"7& o>nr\ 
thickness.^One extrusion segment of the Cr - 15V — 0.7Y (at. %) alloy was clad in drilled 
molybdenum rod, sealed by inert arc welding, soaked at 1370°C in argon for 15 minutes, 
and extruded to approximately 0.7-cm diameter. Subsequent zyglo inspection revealed the 
re-extruded rod to be crack-free. Cylindrical button-head tensile specimens have been 
machined from segments of the re-extruded Cr — 15V — 0.7Y alloy. 

In other efforts to exploit high-strength chromium-base alloys, three alloys were con¬ 
sidered which had been developed under a NASA contract in which GE-NMPO partici¬ 
pated.* Fabrication difficulties have hindered evaluation of these alloys. The compositions 
of these previously melted 3.2-kg vacuum induction-melted heats appear in Table 5.4. The 
ingots were originally extruded at approximately 1510°C from 5.4-cm diameter to 1.91-cm 

*NAS3-7260 with GE-AEG. 

^"Sixth Annual Report — High-Temperature Materials Program, Part A," GE-NMPO, GEMP-475A, March 31, 1967, p. 158. 



218 



Fig. 5.1 — Room-temperature hardness versus alloy content for Cr-V alloys 

TABLE 5.3 

COMPOSITION AND EXTRUSION DATA FOR EXPERIMENTAL Cr-V ALLOYS 


Alloy® 

Nominal 

Composition, at. % 

Extrusion 

Original Diameter, 

Final Diameter, 

V 

Y 

Cr 

Temperature, °C 

cm 

cm 

M-364 

15 

0.7 

Bal 

1205 

5.4 

2.54 

M-300 

20 

0.36 

Bal 

1205 

5.4 

2.54 


a All alloys are 3.2-kg heats. 


TABLE 5.4 

NOMINAL COMPOSITIONS AND IMPURITY ANALYSES 


FOR HIGH-STRENGTH Cr-BASE ALLOYS 


Alloy 

Designation 

NMPO 

Heat No. 

Approximate 
Analysis, at. % 

Gas Content, 
ppm 

W 

Mo 

Y 

O 

N 

H 

CI-5 

M-245 


6.0 

0.1 

35 

72 

6 

CI-7 

M-246 

4.0 


0.1 

28 

89 

13 

Cl -8 

M-248 

6.0 


0.1 

29 

89 

7 


diameter, and swaged to 0.635-cm diameter. Cylindrical button-head tensile specimens 
were prepared and tested in the range 1204° to 1315°C. Attempts were made to re-extrude 
remaining segments of the three alloys from 1.91-cm diameter to 0.7-cm diameter after 
encapsulating in 5. 08-cm-diameter molybdenum rod, sealing by inert arc welding, and 
soaking at 1480° to 1510°C in argon for 15 minutes. The extruded rods appeared to be 
crack-free, but subsequent zyglo inspection revealed severe transverse cracking which 
seemed to be caused by the differential thermal expansion of chromium and molybdenum. A 
second 3. 2-kg vacuum-induction-melted ingot of the Cr - 6W - 0.1Y (at. %) alloy has been 
cast and will be brought to size using a modified extrusion procedure. 

In efforts to assess the effect on ductility of highest-purity starting materials, experi¬ 
mental 80-gram button heats of several Cr-V alloys were prepared using high-purity vana¬ 
dium obtained from the US Bureau of Mines, Boulder City, Nevada and iodide Cr. Nominal 



219 


analyses of the compositions melted appear in Table 5.5. Vanadium was obtained in crystal 
form from the US Bureau of Mines; impurity content was 20 ppm carbon, 75 ppm oxygen, 10 
ppm nitrogen. Chromium was obtained in iodide crystals from Battelle Memorial Institute. 

It had an impurity content of 40 ppm carbon, 8 ppm oxygen, 2 ppm nitrogen. 

The buttons were arc-melted under a partial pressure of argon (50.8 cm) in the presence 
of a titanium button to getter interstitials. Drop casting into a cylindrical copper mold was 
conducted in a smaller arc-melting furnace (P^ = 50.8 cm) without a titanium getter. The 
interstitial element content of the Cr - 20V alloy was determined to include 35 ppm oxygen 
and less than 1 ppm nitrogen, compared to calculated levels (based on raw material compo¬ 
sition) of 21 ppm oxygen and approximately 4 ppm nitrogen. The alloys were encapsulated in 
mild steel and rolled in hydrogen at 980°C to 0.59 cm thickness where major cracking oc¬ 
curred. A second series of Cr-V alloys (see Table 5.5) is being prepared employing the same 
raw materials and melting procedures. The arc-cast buttons will be encapsulated and ex¬ 
truded to approximately 0.95 cm diameter from which cylindrical button-head tensile 
specimens will be machined to determine mechanical properties. 

MATERIAL EVALUATION 

Environmental Compatibility 

Exploratory static sodium environment tests conducted at 650°C for 14 days on the three 
primary alloys and the Cr - 20V - 1Y (at. %, heat M-300) alloy were completed at ANL. 2 
Details of these tests were reported previously. 3 Weight change data are reported in Table 
5.6, along with data reported by ANL for a V - 20Ti (wt %) sample. All four chromium- 
base alloys are clearly superior to the V - 20Ti. ANL concludes that t! these exploratory 
tests indicate chromium-base alloys may be useful materials for service in low-oxygen 
sodium at 650°C." 2 

TABLE 5.5 

NOMINAL COMPOSITIONS 
OF H!GH-PURlTY a Cr-V 
ARC-MELTED 
BUTTON HEATS 


Composition, 
at. % 


Button No. 

Cr 

V 

First Series 

1 

95 

5 

2 

90 

10 


3 

85 

15 


TABLE 5.6 


4 

80 

20 




5 

70 

30 


STATIC SODIUM TEST WEIGHT CHANGE DATA FOR 


6 

60 

40 


DEVELOPMENTAL Cr-BASE ALLOYS AT 650°C 


7 

50 

50 




Second Series 



Alloy Heat No. 

Composition, wt % Weight Loss, 9 mg/cm 2 






la 

95 

5 

M- 313 

Cr - 0.097Y 

0.017 

. 2a 

85 

15 

M—314 

Cr - 0.24Hf - 0.07Th - 0.13Y - 0.007Zr 

0.008 

3a 

75 

15 

M—315 

Cr - 0.37Zr - 0.12Ti - 0.12Y - 349 ppm C 

0.005 

4a 

65 

35 

M-300 

Cr — 20V — 1Y 

0.014 

5a 

50 

50 

- 

V - 20Ti 

0.500 


a Raw materials consisted of ^Weight losses are accurate to ±0.001 mg/cm2. 

iodide Cr and V from U.S. 

Bureau of Mines, 

Boulder City, Nevada. 

2 Private communication, Sherman Greenberg, September 11, 1967. 

3 "AEC Fuels and Materials Development Program Progress Report No. 71," GE-NMPO, GEMP-1002, December 29, 1967, p. 83. 



220 


A previously described sodium capsule 4 heated to 750°C for 1000 hours showed very small 
weight changes for all chromium-base alloys. 

Strength and Ductility 

p Tensile Properties - T ensi le data at 550°, 650°, and 750°Chave been obtained on the three 
primary alloys and various experimental compositio ns.\ Table 5.7 and Figure 5.2 present 
these data with information concerning working conditions and pre-test hardness values. 
Differences in finish rolling temperature caused some variability in the hardness and short- 
time strength of the materials. Data indicate that warm rolling can have a profound effect 
on the tensile properties of these chromium-base alloys. Moreover, it is obvious that warm 
rolling is necessary in these alloys to achieve short-time strength levels superior to auste¬ 
nitic stainless steels. 

Included in Table 5.7 are preliminary data on cylindrical button-head Cr — 15V - 0.7Y 
(at. %) alloy specimens in the as-extruded condition. Its tensile strength in the hot-extruded 
condition exceeds those of the warm-finished primary alloys at 750°C and 550°C. The low 
tensile ductility of the Cr — 15V — 0.7Y alloy at 550°C is a preliminary indication of a rela¬ 
tively high transition temperature. 



Tensile properties of the high-strength chromium-base alloys developed under a NASA 
contract also appear in Table 5. 7. Additional tests will be performed pending the success¬ 
ful fabrication of specimens. ^ ^ 

r Creep-Ruptiire Strength - Initial c reep-rupture tests at 750°C were conducted on as- 
warm-rolled Specimens of the three primary alloys]in the type of argon-filled test cap¬ 
sule pictured in Figure 5.3. Data are presented in Table 5.8. 

In these tests some attention focuses upon the possible influence of atmosphere con¬ 
taminants on strength and ductility. The initial Cr-Hf-Th-Y specimens tested were dis¬ 
colored after rupture. Analyses (Table 5.8) confirmed the contamination of the specimens 
by nitrogen, oxygen, and carbon. Subsequently, specimens were wrapped in tantalum foil 
in an effort to reduce interstitial pickup. Analysis of several longtime tests wrapped in 
tantalum, however, revealed a significant increase in interstitial concentration and hard¬ 
ness (Figure 5.4). Fabrication and test procedures were checked and/or modified in fur¬ 
ther efforts to minimize contamination. 5 The use of titanium shrouding material and a new 
encapsulation procedure 6 are being evaluated. Data in Table 5.7 compare interstitial anal¬ 
yses of a Cr-TZC-Y specimen (CC-7) which was enveloped in a tantalum shroud and was 
on test for 1307 hours, with analyses of two other specimens (219 and AA-14) tested with 
titanium foil as a getter, using the improved encapsulation procedure. These limited data 
indicate that the contamination problem has been markedly reduced. Post-test analyses 
for interstitial element content will continue to confirm initial results. 

All creep-rupture data involving the three primary alloys (warm-rolled) in which the 
specimens were wrapped in titanium foil and encapsulated under the improved procedure 
are plotted in Figure 5.5 along with data for Types 304 and 316 stainless steel. All three 
alloys exhibit a definite advantage over Types 304 and 316 stainless steel in creep-rupture 
at 750°C. Testing at 750°C and 650°C will continue in order to complete creep-rupture 
characterization of the primary alloys. Time extension data at 650°C and 750°C for speci¬ 
mens of all primary alloys compared to the stainless steels are presented in Figures 5. 6 

4 GEMP-475A, p. 158. 

5 GEMP-1002, pp. 85-88. 

6 GEMP-1002, p. 86. 



221 


TABLE 5.7 


TENSILE PROPERTIES OF Cr-BASE ALLOYS IN ARGON 3 


Alloy 


Heat No. 

Condition* 3 

Test 

Temperature, 

°C 

Yield Strength? 
ksi kg/cm 2 

Tensile Strength, 
ksi kg/cm 2 

Elongation, 

% 

Pre-Test 

Hardness, 

DPH 

Cr-Y 


M-313 

Warm rolled 

750 

36.9 

2580 

38.6 

2700 

— 

- 

Cr-Y 


M-313 

Warm rolled 

750 

27.3 

1910 

28.8 

2020 

- 

— 

Cr-Y 


M-313 

Warm rolled 
+ annealed 

750 

28.5 

1930 

37.4 

2620 

3.4 

227 

Cr-Y 


M-313 

Warm rolled 
+ annealed 

750 

31.4 

2200 

38.0 

2660 

2.0 

227 

Cr-Y 


M-216 

Hot rolled 

750 

12.6 

880 

20.1 

1410 

28.4 

119 

Cr-Y 


M-313 

Hot rolled 
+ annealed 

750 

10.2 

710 

22.5 

1570 

41.4 

153 

Cr-Y 


M-313 

Warm rolled 

650 

- 

- 

50.0 

3500 

14.7 

- 

Cr-Y 


M-313 

Warm rolled 

650 

42.4 

2960 

43.7 

3060 

2.1 

226 

Cr-Y 


M-313 

Warm rolled 

550 

49.2 

3940 

49.9 

3490 

3.2 

- 

Cr-Y 


M-313 

Warm rolled 

550 

58.8 

4110 

62.6 

4390 

2.2 

266 

Cr-Hf-Th-Y 


M-314 

Warm rolled 

750 

45.5 

3180 

54.7 

3830 

3.3 

240 

Cr-Hf-Th-Y 


M-314 

Warm rolled 

750 

44.9 

3140 

55.5 

3880 

5.3 

- 

Cr-Hf-Th-Y 


M-314 

Warm rolled 

750 

41.8 

2920 

53.2 

3720 

3.7 


Cr-Hf-Th-Y 


M-314 

Warm rolled 
+ annealed 

750 

52.3 

3660 

56.7 

3970 

2.9 

258 

Cr-Hf-Th-Y 


M-314 

Warm rolled 
+ annealed 

750 

50.3 

3520 

57.0 

3990 

3.1 

258 

Cr-Hf-Th-Y 


M-191 

Hot rolled 

750 

21.2 

1480 

27.7 

1940 

6.5 

170 

Cr-Hf-Th-Y 


M-314 

Hot rolled 
+ annealed 

750 

40.2 

2820 

41.2 

2880 

6.1 

203 

Cr-Hf-Th-Y 


M-314 

Warm rolled 

650 

51.3 

3600 

55.7 

3900 

- 

- 

Cr-Hf-Th-Y 


M-314 

Warm rolled 

650 

51.1 

3590 

59.4 

4150 

4.5 

278 

Cr-Hf-Th-Y 


M-314 

Warm rolled 

550 

48.3 

3380 

57.8 

4050 

2.9 

240 

Cr-Hf-Th-Y 


M-314 

Warm rolled 

550 

56.8 

3980 

65.0 

4550 

4.5 

- 

Cr-Hf-Th-Y 


M-314 

Warm rolled 

550 

59.6 

4180 

66.9 

4680 

3.3 

278 

Cr-Hf-Th-Y 


M-191 

Hot rolled 

550 

19.1 

1340 

31.5 

2200 

16.3 

140 

Cr-TZC-Y 


M-315 

Warm rolled 

750 

43.8 

3060 

61.9 

4330 

1.9 

266 

Cr-TZC-Y 


M-315 

Warm rolled 

750 

45.3 

3170 

48.4 

3420 

4.3 

- 

Cr-TZC-Y 


M-315 

Warm rolled 

750 

45.6 

3200 

57.3 

4010 

5.1 

262 

Cr-TZC-Y 


M-315 

Warm rolled 

+ annealed 

750 

53.3 

3730 

62.0 

4340 

4.1 

266 

Cr-TZC-Y 


M-298 

Hot rolled 

750 

8.3 

580 

26.9 

1880 

35.4 

131 

Cr-TZC-Y 


M-298 

Hot rolled 
+ annealed 

750 

13.2 

925 

20.1 

1410 

8.1 

139 

Cr-TZC-Y 


M-315 

Warm rolled 

650 

59.1 

4140 

63.6 

4460 

5.3 

- 

Cr-TZC-Y 


M-315 

Warm rolled 

650 

61.2 

4280 

65.4 

4570 

3.4 

266 

Cr-TZC-Y 


M-315 

Warm rolled 

550 

62.6 

4390 

66.4 

4650 

2.9 

- 

Cr-TZC-Y 


M-315 

Warm rolled 

550 

49.2 

3440 

49.9 

3490 

3.2 

- 

Cr-TZC-Y 


M-315 

Warm rolled 

550 

63.3 

4430 

69.7 

4880 

4.4 

270 

Cr-TZC-Y 


M-298 

Hot rolled 

550 

10.7 

750 

30.9 

2160 

15.2 

134 

Cr- 15V- 

0.7Y 

M-364 d 

As-extruded 

750 

_ 

— 

68.5 

4800 

29.0 

- 

O- 15V- 

0.7Y 

M-364^ As-extruded 

550 

- 

- 

80.8 

5650 

0.5 

- 

CI-5 


_ 

As-hot-swaged 

1310 

22.8 

1600 

24.0 

1680 

14.2 

- 

CI-7 


— 

As-hot-swaged 

1040 

58.9 

4120 

63.6 

4450 

20.7 

- 

CI-7 


- 

As-hot-swaged 

1310 

22.7 

1590 

24.3 

1700 

19.5 

- 

CI-8 


— 

As-hot-swaged 

1040 

73.2 

5120 

81.0 

5670 

19.3 

- 

Cl-9 


- 

As-hot-swaged 

1310 

28.5 

1990 

30.1 

2110 

48.6 

— 


specimens were 0.076 cm thick with a 0.63-cm-wide reduced section and a 2.54-cm gage length. Major axis of each 
specimen parallels rolling direction. 

^Warm rolled from 850°C; annealed in vacuum at 750°C for 2 hours. 

Hot rolled from 1100°C; annealed in vacuum at 750°C for 2 hours. 
c Yield strength value is approximate, based on deflectometer measurement of total load train elongation. 
^Cylindrical button-head specimens. 




222 



Fig. 5.2 — Ultimate tensile strengths of Type 304 and 316 stainless steel and 
warm rolled (850°C) Cr-base primary alloys 

and 5.7. In all capsule creep tests conducted to date in this program, creep is based on 
total load train movement. Thus actual specimen creep is no greater, and in some cases 
is less than, indicated creep, which is assumed to occur within the gage section in the 
central 1 inch of the overall specimen length. All stainless steel data are based on a 
1 -inch gage length and optical readings of strain from a platinum strip extensometer. 

Two new systems are being procured which will permit optical strain measurements on 
specimens tested in argon. 

f pBTT in Bending - Sheet bend tests were performed on the three primary alloys and a 
Cr — 20V — 1 Y (M-300) alloy after various annealing treatments. Testing was performed 
on a tensile machine at a crosshead speed of 0. 076 cm/min over a 2T radius. Bending 
was continued until fracture or a 90-degree bend was achie ved./ Results in Table 5.9 and 
Figure 5.8 may be summarized as follow: 


1 • The DBTT for Cr-Y is about 100°C and does not vary with the annealing treatments 
employed. 


2. The DBTT for Cr-Hf-Th-Y in the as-warm-rolled condition has not been determined, 
but after a 750°C anneal it is about 200°C and decreases for increasing annealing tem¬ 
peratures and/or times, although accompanied by a significant loss in hardness 
(strength). 

3. The DBTT for Cr-TZC-Y has not been determined in the as-warm-rolled condition; 
with annealing at 750°C and 870°C after warm rolling it is apparently above 100°C. 
Higher annealing temperatures tend to reduce it to room temperature, again accom¬ 
panied by a significant loss in hardness (strength). 


4. The limited data on Cr - 20V - Y in the as-warm-rolled condition, annealed at 750°C, 
indicate that its DBTT exceeds 200°C. 

f DBTT injgyxic^— Subsize cylindrical notched specimens were prepared as previously 


described 7 from both hot (1100°C) and warm (850°C) swaged 0.64-cm-diameter rod of the 
three primary alloys. Test results appear in Figure 5.9 and suggest that warm swaging 
and annealing at 750°C lower the DBTT of Cr-Y from 400°C for hot-swaged to approxi- >->?> 


7 GEMP-475A, p. 161. 




































STRESS RUPTURE RESULTS 3 ON THE THREE PRIMARY Cr-BASE ALLOYS IN ARGON 


224 


CD h* CN LO LO o 
00 O CM 00 1- O 
CM CM CM CM N 


O 

£1111 


ID O 03 

ID CM | | | 00 I 

CO LO 1 I 1 LO I 


1 I I I 


't LD CM 

00 10 I I I ® | 


® I I ! I J « 


Co I I I 1 LO 


. i-OOLDlDOCDI^OOOO , 
IcoiDttcor^o^-t-cMCo ! 
CMCMCMCMCMCMCOCMCMCM 


,co , io n o oi 

I CD I ** t- co <3- 
CM CM CM ID CM 


.oococMtoco^rcoor^h* . 
I COlDCOCDlOCMCOr-CMCO ! 
CMCMCMCMCMCMCMCMCMCM 


, 03 . CO O r- 03 

I ^ I *3* ^ CO ^ 

CM CM CM CM CM 


CM CM CM CM CM 


O g-sS 


r-LOCOCDt-ID^tOOLOCq 

CJ) fN |N Lfi CM* LC (V) 


rs o o o n to 0 co ct> »- 0 
trJ in' (d 6 cm o co 03 rs co 


j CM r- 


CO 0 CO CD fs CO 

CO 1- CD O CO 

<3- co cm 


cor^oi'-M- 
CD fs (M M* 
ID ^ O CM 
** CM 


^tCMCMOOOOOO^tO 
| CDt-t— (OCOCOt-NCMCO 
1 Tf»-'c-r>»r^pv'^-p-0O'3- 

CM CNJ CM t— »— »— CO CM CM 


O <3- o o o 

I r-» CO CO CO CD N 

CO CM CM 1- t- CO 


O O <3" O CM 

I CD T- CO 03 T- 

1 r->. ^ *3 - cm •— 

r* p- CM CM CM 


t LDOOLDLDtDOlDOLDO 
COCOCOCNCMCMCM'^-'^CO 


I ID ID ID 00 ID ID 
M - O CO CM CM M" 


I ID O ID CM O 
CM CM CO CO 00 


•—TOT- o CO ~o 

03 5 T T 7 7 9^ t 

a cJoOQDODOCOCOi^aOCOLDlDT- 

tD <OOOOCOQOOOCMOOOQCMCMCM 


§ <3- oo-o-o T- 

« o o “ lo 6 6 
< O O ™ cm O O 


c -a “o 

TO iD rf 

>*0 -O 7“ T3 «7 

« <3* CD < LO < 

< CM CM < CM < 


r^.r^.r^r^r^r^r^cDcDcDr^ 


i—. r— i— co 


o o o o o 

ID ID LO LO LD 

r^. O'. CO CD CD 


Ta shroud. 

Test terminated — no rupture. 
*Ti shroud. 


Elongation, in. Stress, psi Interstitial contamination, ppm 


225 



Fig. 5.4 — Effect of interstitial contamination on rupture life of Cr—Hf—Th+Y 



10,000 


Time, hours 


Fig. 5.5 — Stress-rupture properties of the three primary Cr-base alloys in argon. 



Fig. 5.6 - 750°C time extension curves for Type 316 stainless steel and the Cr-base primary alloys 
at various stress levels 
























































































226 



Fig. 5.7 — 650°C time — extension curves for Type 316 SS and the Cr-base primary alloys at 
various stress levels 

mately 250°C. Annealing at 980°C after warm swaging raises the transition temperature 
of Cr-Y by approximately 125°C. Warm swaging and annealing, compared to hot swaging, 
apparently have little effect on the DBTT of Cr-Hf-Th-Y. Compared with previous results 
on hot-swaged material, 8 warm swaging and annealing at 750°C of Cr-TZC-Y reduces the 
DBTT from approximately 500°C to 400°C. 

Hot-Hardness - Hot- hardness^ ests were performed on the three primary alloys and sev¬ 
eral experimental compositions utilizing a previously described test apparatus. 9 Figure 
5.10 compares the hot-hardness of the three primary alloys in the warm-rolled condition 
with two high-strength chromium-base alloys and two experimental chromium-base alloys. 
The primary alloys are generally softer than the high-strength chromium-base alloys. 

Thermal Stability J _ _ ^^330 

Vacuum annealing stability studies were performed on the three primary alloys wrapped 
in tantalum foil at temperatures from 750° to 1100°C for 2 and 10 hours. Figure 5.11 incor¬ 
porates previous 10 2-hour vacuum annealing data on warm-rolled sheet with 10-hour data. 
The change in hardness which appears when the annealing time is increased from 2 to 10 
hours has initiated a more complete investigation of longtime (1000 hr) annealing and its 
effect on mechanical properties, taking into account the two competing processes of soft¬ 
ening during annealing and hardening due to interstitial contamination from the test en¬ 
vironment. Photomicrographs 11 of warm- and hot-rolled Cr-Y after 2-hour vacuum an¬ 
neals at 750° to 1090°C show partial recrystallization in the warm-rolled specimens after 
the 870°C anneal. Significant grain growth occurs after the 1090°C anneal. No change in 
the structure of the hot-rolled material can be noted, even after the 1090° C anneal. 

8 GEMP-475A, pp. 161-162. 

9 " 

AEC Fuels and Materials Development Program Progress Report No. 69," GE-NMPO, GEMP-69, September 29, 1967, 

pp. 106-108. 

10 GEMP-475A, pp. 162, 164. 

11 "AEC Fuels and Materials Development Program Progress Report No. 67," GE-NMPO, GEMP-67, June 30, 1967, p. 97. 
























































TABLE 5.9 


BEND DATA FOR Cr-BASE ALLOY AS A FUNCTION OF VACUUM ANNEALING CONDITIONS 


Alloy 

Designation 

Heat No. 

Vacuum 

Annealing 

Conditions 

Hardness, 

DPH 

Temperature, 9 

°C 

Bend Angle, 
degrees 

Comment 

Cr-Y, W/R b 

M-313 

As-rolled 

195 

RT 

9 

Fractured 

Cr-Y, W/R 


750°C - 2 hr 

195 

RT 

6 

Fractured 

Cr-Y, W/R 


980°C - 2 hr 

127 

RT 

9 

Fractured 

Cr-Y, W/R 


As-rolled 

195 

106 

90 

No fracture 

Cr-Y, W/R 


750°C - 2 hr 

195 

104 

90 

No fracture 

Cr-Y, W/R 


870°C - 2 hr 

178 

104 

90 

No fracture 

Cr-Y, W/R 


980°C - 2 hr 

127 

100 

90 

No fracture 

Cr-Y, W/R 


750°C - 10 hr 

200 

110 

90 

No fracture 

Cr-Y, W/R 


870°C — 10 hr 

136 

106 

90 

No fracture 

Cr-Y, H/R 


750°C - 2 hr 

140 

109 

5 

Fractured 

Cr-Hf-Th-Y, W/R 

M-314 

980°C - 2 hr 

175 

RT 

90 

No fracture 

Cr-Hf-Th-Y, W/R 


870°C - 10 hr 

197 

RT 

90 

No,fracture 

Cr-Hf-Th-Y, H/R 


980°C - 3.5 hr 

145 

RT 

90 

No fracture 

Cr-Hf-Th-Y, W/R 


750°C - 2 hr 

232 

100 

9 

Fractured 

Cr-Hf-Th-Y, W/R 


750°C — 2 hr 

232 

150 

10 

Fractured 

Cr-Hf-Th-Y, W/R 


750°C - 2 hr 

232 

204 

90 

Fractured 

Cr-Hf-Th-Y, W/R 


750°C — 10 hr 

222 

110 

72 

Fractured 

Cr-Hf-Th-Y, W/R 


870°C - 2 hr 

216 

110 

18 

Fractured 

Cr-Hf-Th-Y, W/R 


870°C - 10 hr 

197 

110 

90 

No fracture 

Cr-Hf-Th-Y, W/R 


870°C - 2 hr 

216 

150 

90 

No fracture 

Cr-Hf-Th-Y, W/R 


870°C - 2 hr 

216 

209 

90 

No fracture 

Cr-Hf-Th-Y, W/R 


980°C - 2 hr 

175 

108 

90 

No fracture 

Cr-Hf-Th-Y, W/R 


1090°C — 2 hr 

140 

110 

90 

No fracture 

Cr-Hf-Th-Y, H/R 


750°C - 2 hr 

196 

105 

7 

Fractured 

Cr-Hf-Th-Y, H/R 


980°C - 3.5 hr 

145 

109 

90 

No fracture 

Cr-TZC-Y, W/R 

M-315 

980°C - 2 hr 

154 

RT 

7 

Fractured 

Cr-TZC-Y, W/R 


1090°C — 2 hr 

129 

RT 

90 

No fracture 

Cr-TZC-Y, H/R 


980°C - 3.5 hr 

142 

RT 

5 

Fractured 

Cr-TZC-Y, W/R 


750°C - 2 hr 

259 

RT 

12 

Fractured 

Cr-TZC-Y, W/R 


750°C - 2 hr 

259 

150 

7 

Fractured 

Cr-TZC-Y, W/R 


750°C - 2 hr 

259 

204 

9 

Fractured 

Cr-TZC-Y, W/R 


750°C — 10 hr 

252 

110 

15 

Fractured 

Cr-TZC-Y, W/R 


870°C - 2 hr 

227 

110 

5 

Fractured 

Cr-TZC-Y, W/R 


870°C - 2 hr 

227 

150 

18 

Fractured 

Cr-TZC-Y, W/R 


870°C - 2 hr 

227 

204 

9 

Fractured 

Cr-TZC-Y, W/R 


870°C - 10 hr 

211 

106 

9 

Fractured 

Cr-TZC-Y, W/R 


980°C - 2 hr 

154 

108 

90 

No fracture 

Cr-TZC-Y, H/R 


750°C - 2 hr 

181 

114 

7 

Fractured 

Cr-TZC-Y, H/R 


980°C - 3.5 hr 

142 

108 

90 

No fracture 

Cr-V-Y, W/R 

M-300 

750°C - 2 hr 

366 

110 

9 

Fractured 

Cr-V-Y, W/R 


750°C - 2 hr 

366 

152 

10 

Fractured 

Cr-V-Y, W/R 


750°C - 2 hr 

366 

204 

9 

Fractured 


a For tests above room temperature, the specimens were heated in test position in a resistance furnace. 
^W/R - warm-rolled from 850°C. 

H/R - hot-rolled from 1100°C. 



228 



u 

O 

o' 

3 

o 

o 

Q. 

E 


01 x 6>f-uj 'pscjaosqD XBjoug 




sGdjBsp 'uoip0[jaQ 


Fig. 5.8 - Bend data for Cr-base alloys as a function of Fig. 5.9 - Impact data for Cr-base alloys as a function of 

vacuum annealing conditions vacuum annealing conditions 






229 



Fig. 5.10 - Hot hardness of Cr-base alloys 



Annealing temperature, °C 

Fig. 5.11 — Room-temperature hardness versus annealing temperature for warm-rolled Cr-base alloys 

Hot-hardness at 750°C as a function of time was determined for the three primary alloys 
and the Cr - 20V - Y (M-300) alloy in the warm-rolled condition. These tests were con¬ 
ducted in argon, in the previously described hot-hardness tester. The impressions were 
made periodically and read after the 137-hour period when the specimens cooled to room 
temperature. The results (Figure 5.12) show that all alloys retained essentially uniform 
hardness except for the Cr-Y specimen which softened after about 75 hours of exposure. 

Processability 

Previous exploratory efforts 12 to determine the feasibility of fabricating a chromium- 
base alloy in tubular form were moderately successful. Preliminary welding experiments 
were performed on bend specimens of warm-rolled Cr-Y sheet. 13 A TIG weld pass parallel 

12 GEMP-475A, pp. 162-166. 

13 GEMP-67, pp. 102-103. 







230 



Time, hours 


Fig. 5.12 — Hot-hardness of Cr-base alloys at 750°C 

to the major axis produced a weld zone 3.2 mm wide with about 95 percent penetration in 
all cases. Both the weld zone and heat-affected zone were free of cracks and voids and 
had a common hardness of 160 DPH, compared to the 197 DPH hardness of the warm- 
rolled parent metal before welding. Bend tests conducted on the welded specimens indi¬ 
cated a DBTT similar to that obtained for the parent metal under similar annealing condi¬ 
tions. A typical TIG welded specimen before and after bend testing is shown in Figure 5.13. 
Exploratory electron-beam welds were attempted at 10~ 5 Torr, but prohibitive volatiliza¬ 
tion of chromium occurred. 


Radiation Stability 


Hot-rolled creep-rupture and bend specimens of the three primary alloys were included 
in irradiation experiment capsule ORM-53, which received an estimated dosage of 2.3xl0 20 
n/cm 2 (E n ^ 1 Mev). The capsule was irradiated in the A-2 facility of the ORR for four cy¬ 
cles. The specimens appeared in good condition following capsule disassembly. Testing has 


not yet begun. ^ ^ 

5.2 Fe^Cr-Al-YALLOYS 


<r 


* 


The Fe-Cr-Al-Y alloys were investigated with the knowledge that they were low in strength 
compared to austenitic stainless steel, but that their oxidation resistance and tensile ductility 




231 


Bent 90 degrees at 100°C after 
vacuum annealing at 980°C 


As-welded 



Fig. 5.13 - Welded Cr-Y alloy (M-313) bend specimens (Neg. P67-5-21) 

were excellent. There were indications that their resistance to irradiation might be outstand¬ 
ing. 14 It was also considered possible that retention of ductility in an alloy during service 
might be more important ultimately than an initial high-strength level compromised by loss 
of ductility during servi ce, j ~=7f • 5 

During the past year a number of previously developed Fe-Cr-Al-Y alloys have been 
characterized, particularly with respect to strength in the temperature range of interest. 
Although a few additional studies remain to be completed, no new work is contemplated on 
alloys within this system. 

MATERIAL PREPARATION 

The sheet material used in the characterization of the Fe-Cr-Al-Y alloys was provided 
by six 45.5-kg heats. Melting and fabrication details were previously described. 16 Analyses 
of the six heats appear in Table 5.10. 

MATERIAL EVALUATION 
Environmental Compatibility 

Two Fe-Cr-Al-Y specimens were included in the capsule tested at ANL. The results in 
Table 5.11 indicate weight losses comparable to the reference V — 20 Ti specimen and in 
excess of the chromium-base alloys. Previous data 16 on a number of Fe-Cr-Al-Y specimens 
exposed to static metallic sodium (containing 22 ppm O 2 ) at 750°C for 1000 hours indicated 
that most alloys were more susceptible than chromium-base alloys similarly tested. To de¬ 
termine effect of longer exposure times, two additional capsules have completed over 2000 
hours on test. One capsule contains vapor-blasted bend specimens (0. 076 by 0.947 by 3.8 cm) 
of 1541* (heatMS-51), 0561 (heat MS-59), 1061 (heat MS-61), 2541 (heatMS-62), 0561 + 3Mo 
(heat MS-64), and 304 and 316 stainless steel. The other capsule contains specimens 
pre-oxidized in air at 1090°C for 1 hour after vapor blasting. These await inspection. 


‘Designation for Fe-Cr-Al-Y alloy. The first two digits denote Cr content, the third Al, and the fourth Y; i.e., 1541 is 
Fe - 15Cr - 4AI - 1Y, 0561 is Fe - 5Cr - 6AI - 1Y, etc. 

14 GEMP-67, pp. 25,126. 

15 GEMP-475A, p. 144. 

16 


’GEMP-475A, pp. 145-147. 






232 


TABLE 5.10 


COMPOSITION OF Fe-Cr-AI-Y 45.5-kg INDUCTION MELTED HEATS 


Alloy 

Designation 

Heat No. 


Nominal Composition, wt % 



1 nterstitial 
Concentration, 

DDm 

Cr 

At 

Y 

C 

Si 

Mo 

Fe 

°2 

n 2 

0561 

MS-59 

4.75 

5.96 

1.02 

0.05 

_ 

_ 

Bal 

9 

1 

1041 

MS-60 

9.82 

3.86 

0.82 

0.05 

0.03 

— 

Bal 

36 

4 

1541 

MS-51 

15.11 

4.01 

0.77 

- 

- 

- 

Bal 

15.6 

1.9 

1061 

MS-61 

11.87 

5.89 

1.07 

0.03 

0.04 

_ 

Bal 

239 

39 

2541 

MS-62 

24.55 

3.83 

1.23 

0.04 

0.15 

— 

Bal 

11 

6 

0561 + 3Mo 

MS-64 

5.34 

5.87 

1.46 

0.04 

0.09 

3.42 

Bal 

65 

12 


TABLE 5.11 

WEIGHT CHANGE DATA FOR Fe-Cr-AI-Y ALLOYS 


Alloy Heat No. 

Nominal Composition, wt % 

Weight Loss, 3 mg/cm2 

MS-51 

Fe - 15Cr-4Al - 1Y 

0.800 

MS-60 

Fe - 10Cr-4AI - 1Y 

0.130 


V - 20Ti 

0.500 


a Weight losses are accurate to ±0.001 mg/cm2. 

| Strength and Ductility 

" f-c'3 

Tensijg properties - Tensile tests were performed at room temperature, 550°, and 
750°C on annealed sheet specimens of the six Fe-Cr-AI-Y alloys listed in Table 5.10. 

Results appear in Table 5.12 and Figure 5.14, including NMPO data for austenitic stain¬ 
less steel.t Strength apparently increases somewhat with increasing chromium and alum¬ 
inum until 750°C, when the strengths are nearly the same. The strength increase afforded 
by the 3 percent molybdenum addition to 0561 alloy is noteworthy. None of these alloys 
have tensile strengths up to 750°C as high as the austenitic stainless steels, but most of 
them possess equivalent or superior yield strengths at all test temperatures. 

F&13 

Stress-Rupture Strength - Stress-rupture tests on the Fe-Cr-AI-Y alloys were conducted 
in air at 550°C and 750°C using 0.076-cm-thick sheet specimens. The results shown in Table 
5.13 and Figure 5.15 are conclusive in demonstrating the low stress-rupture strength of the 
Fe-Cr alloys compared to austenitic stainless steel. For the Fe-Cr-AI-Y compositions, how¬ 
ever, the higher strength at 750°C of the 2541 alloy and the beneficial effect of molybdenum 
should be noted. At 550 C chromium level does not significantly affect stress-rupture strength, 
but the molybdenum addition remains advantageous. The reservoir of ductility suggested by 
the generally high elongation values of the Fe-Cr-AI-Y alloys is probably related to the favor¬ 
able radiation behavior which has been reported 17 for the 1541 alloy. 

^ J 

DBTT in Bsadjpg - Annealed (1000°C, 10 min) bend specimens (0.076 by 0.95 by 3.81 cm) 
of the six 45.5-kg heats listed in Ta.ble 5.10 were tested before and after aging at 450°C for 
1000 hours. The specimens were bent over a 2T radius in a tensile machine at a crosshead 
speed of 0.69 cm/min until fracture or a 90-degree bend was achieved. The results (Table 
5.14) indicate a rise in the DBTT of 1541 and 2541 as a result of aging. Transition tempera¬ 
tures of the remaining alloys were unchanged by aging. The DBTT of all annealed specimens 
was 0°C or lower.j _ 

^Chemical analysis, weight percent: 

304 - 19.08 Cr, 9.21 Ni, 0.63 Si, 1.88 Mn, 460 ppm N 2 , 417 ppm C. 

316 - 17.61 Cr, 12.5 Ni, 2.62 Mo, 0.63 Si, 1.9 Mn, 580 ppm N 2 , 393 ppm C. 

17 


GEMP-67, pp. 125-126. 




TABLE 5.12 


TENSILE DATA 3 FOR Fe-Cr-AI-Y ALLOYS 


{100-lb Vacuum-Induction-Melted Heats Converted to 0.076-cm Sheet) 



Test 

0.2% 





Temperature, 

Yield Strength, 

Ultimate Strength, 

Elongation in 

Alloy 

Heat No. 

°C 

psi 

kg/cm'^ 

psi 

kg/cm^ 

2.54 cm, % 

0561 

MS-59 

24 

45,100 

3170 

63,400 

4460 

19.4 



24 

45,100 

3170 

64,100 

4510 

22.9 



24 

45,700 

3210 

65,600 

4610 

25J 



550 

26,500 

1860 

33,300 

2340 

46.0 



550 

25,800 

1810 

31,600 

2220 

33.1 



750 

8,900 

630 

9,600 

670 

24.7 



750 

— 

- 

10,600 

750 

19.2 



750 

8,400 

590 

9,400 

660 

22.1 

1041 

MS-60 

24 

43,300 

3040 

66,600 

4680 

28.0 



24 

41,100 

2890 

67,000 

4710 

29.1 



24 

40,600 

2850 

67,500 

4750 

28.4 



550 

22,200 

1560 

30,900 

2170 

26.6 



550 

24,600 

1730 

32,000 

2250 

20.8 



750 

8,100 

570 

9,200 

650 

39.5 



750 

8,000 

560 

9,100 

640 

28.4 

1061 

MS-61 

24 

54,500 

3830 

74,800 

5260 

22.7 



24 

54,800 

3850 

74,000 

5200 

21.4 



24 

55,700 

3920 

75,000 

5270 

22.3 



550 

31,900 

2240 

37,600 

2640 

25.0 



550 

30,900 

2170 

35,500 

2500 

38.7 



750 

7,700 

540 

8,300 

580 

13.6 



750 

- 

- 

9,500 

670 

22.4 

1541 

MS-51 

24 

49,600 

3490 

71,800 

5050 

25.7 



24 

48,900 

3440 

71,400 

5020 

25.1 



24 

49,800 

3500 

71,300 

5010 

24.2 



550 

27,300 

1920 

35,000 

2460 

26.7 



550 

27,300 

1920 

34,500 

2430 

26.2 



550 

27,100 

1900 

33,000 

2320 

35.9 



750 

8,000 

560 

9,200 

650 

16.9 



750 

9,500 

670 

10,000 

700 

20.8 

2541 

Ms^r~ 

24 

56,000 

3940 

79,300 

5580 

19.1 


7r 

,7- 24 

54,300 

3820 

78,800 

5540 

21.2 


24 

55,300 

3890 

79,300 

5580 

18.5 



^ 550 

35,300 

2480 

41,700 

2930 

35.5 


(a ^ 

750 

- 

- 

15,800 

1110 

19.7 

l 


11,600 

12,200 

820 

860 

11,900 

13,300 

840 

940 

20.6 

0561 + 3Mo 

MS-64 

24 

58,500 

4110 

79,300 

5580 

20.4 



24 

58,400 

4100 

79,100 

5560 

21.8 



24 

57,900 

4070 

78,500 

5520 

21.2 



550 

37,400 

2630 

50,000 

3520 

29.3 



750 

15,900 

- 

15,900 

1120 

18.7 



750 

14,500 

1120 

14,500 

1020 

11.6 



750 

16,100 

1130 

16,200 

1140 

11.6 

Type 304 SS 

_ 

24 

28,300 

1980 

85,500 

5990 

74.6 



550 

10,100 

700 

53,300 

3740 

35.5 



750 

8,600 

600 

26,000 

1820 

19.1 

Type 316 SS 

_ 

24 

35,700 

2500 

83,200 

5830 

72.5 



550 

16,000 

1120 

64,000 

4470 

42.5 



750 

13,000 

910 

34,000 

2380 

52.1 


Specimens were 0.076 cm thick with a 1.27-cm-wide reduced section and a 2.54-cm gage length. 
The major axis of all specimens parallels the rolling direction. 



234 




Fig. 5.14 - Tensile and yield strength of Fe-Cr-AI-Y alloys versus temperature 


r- P*£> 

I DBTT in Impact - The DBTT of 0561 (heat MS-59) and 1041 (heat MS-60) were determined 
by testing Charpy V-notch specimens at various temperatures. Specimens were prepared 
from a 2.22-cm diameter hot-rolled (982°C) stock originating from 3.15-cm diameter extru¬ 
sions. Results (Figure 5.16) suggest a DBTT between 100° and 200°C for both alloys, which 
is substantiated by the appearance of the fracture surfaces of 1041 (heat MS-60) shown in 
Figure 5.17. 

A study was made to determine the correlation between DBTT as measured by subsize 
cylindrical notched impact specimens and Charpy V-notch specimens. Subsize specimens 
were machined from both halves of fractured Charpy V-notch specimens of an Fe-Cr-Al-Y 
alloy (MS-60—1041). The results (Figure 5.18) reveal that the subsize specimen indicates a 
lower DBTT than the standard Charpy V-notch. Although the subsize specimen will continue 
to be used for screening purposes, the ASTM standard Charpy V-notch or drop-weight test 
will be used in circumstances demanding comparative engineering data. 








235 


TABLE 5.13 


CREEP-RUPTURE DATA 3 FOR Fe-Cr-AI-Y ALLOYS IN AIR 


Alloy 

Designation 

Heat No. 

Test 

Temperature, 

°C 

Stress 

psi kg/cm 2 

Time to Indicated 
Strain, hr 

0.5% 1% 5% 

Time to 
Rupture, hr 

Linear Creep 
Rate, hr~^ 

Elongation, 

% 

0561 

MS-59 

550 

20,000 

1400 

0.2 

0.5 

2.5 

9.5 

2.0 x 10 -2 

37.9 




15,000 

1050 

1.1 

2.9 

18.8 

87.2 

2.3 x 10 ** 

60.2 




11,000 

770 

10.3 

23.5 

126.0 

619 

3.7 x 10 -4 

71.7 



750 

3,000 

210 

0.7 

1.4 

6.6 

35.9 

7.4 x 10 -3 

44.7 




2,000 

140 

1.8 

6.6 

51 

331.2 

6.4 x 10~ 4 

61.0 




1,000 

77 

158.0 

880.0 

- 

2015.3 

1.1 x 10"° 

1.5 

0561 +3Mo 

MS-64 

550 

40,000 

2800 

— 

— 

1.6 

4.4 

1.7 xIO -2 

27.1 




30,000 

2100 

2.8 

6.4 

25.8 

70.2 

1.8x10 3 

39.3 




20,000 

1400 

36.5 

78.5 

318.0 

948.0 

1.2 x 10 -4 

70.2 




15,000 

1050 

190.0 

334.0 

- 

867.6 

4.1 x 10" & 

3.2 b 



750 

5,000 

350 

0.4 

0.8 

3.3 

14.6 


76.9 




4,000 

280 

1.7 

3.4 

12.9 

52.5 

3.0 x 10~ J 

89.8 




3,000 

210 

5.8 

13.0 

76.3 

441.3 

5.8 x 10~ 4 

42.3 

1041 

MS-60 

550 

20,000 

1400 

_ 

0.4 

3.9 

23.9 

1.1 x 10 —2 

62.7 




15,000 

1050 

2.3 

6.4 

38.8 

168.8 

1.23 x 10 3 

48.1 




13,000 

910 

6.5 

14.5 

74.5 

385.9 

6.05 x 10“ 4 

73.4 




7,000 

490 

181.0 

435.1 

- 

492.0 b 

2.0 x 10 -5 

1.10 b 



750 

4,000 

280 

0.3 

0.7 

3.8 

14.8 

1.2 x 10~ 2 

57.6 




3,000 

210 

1.0 

2.3 

12.1 

58.9 

4.0 x 10“ 3 

51.9 




2,000 

140 

10.0 

32.0 

310.0 

- 

1.5 x 10 -4 

- 




1,400 

98 

163.0 

512.0 

- 

600.1 b 

1.7 x 10 -5 

1.1 b 

1061 

MS-61 

550 

20,000 

1400 

0.2 

0.5 

2.5 

11.4 

2.0 x 10 —2 

52.2 




15,000 

1050 

0.5 

1.7 

13.5 

84.7 

3.3 x 10" 3 

52.4 




11,000 

770 

12.0 

24.0 

112.0 

489.2 

4.2 x 10 -4 

105.3 



750 

4,000 

280 

0.2 

0.3 

1.73 

8.2 

2.8 x 10 -2 

65.6 




3,000 

210 

0.6 

1.3 

7.0 

38.5 

7.0 x 10~ 3 

25.9 




2,500 

175 

2.9 

5.8 

28.6 

142.9 

1.8 x 10 -3 

- 




2,000 

140 

6.1 

14.3 

92.3 

395.0 

4.6 x 10’*' 4 

34.5 

1541 

MS-51 

550 

20,000 

1400 

0.7 

1.7 

— 

49.4 

2.4 x 10 -3 

44.8 




15,000 

1050 

16.5 

56.6 

193.0 

491.4 

1.2 x 10“ 4 

44.3 




11,000 

770 

330.0 

850.0 

1808.0 

3567.8 

— 

75.0 



750 

3,000 

210 

1.0 

2.2 

11.5 

42.9 

4.3 x 10 -3 

34.7 




2,000 

140 

12.3 

26.7 

145.0 

624.2 

3.5 x 10“^ 

59.2 




1,100 

77 

510.0 

- 

- 

2278.0 b 

1.5 x 10 -6 

0.59 b 

2541 

MS-62 

550 

25,000 

1750 

0.2 

0.3 

2.7 

10.8 

1.5 x 10~ 2 

Q 

43.2 




20,000 

1400 

0.8 

2.3 

13.1 

34.8 

3.1 x 10"“ 3 

47.6 




15,000 

1050 

2.8 

6.7 

46.7 

213.5 

9.0 x 10 -4 

49.9 




5,000 

350 

680.0 

- 

- 

2188.4 b 

2.5 x 10~ 6 

0.59 b 



750 

5,000 

350 

0.3 

0.6 

2.5 

16.1 

- 

83.4 




3,500 

245 

1.1 

2.8 

16.9 

73.7 

2.9 x 10 -3 

79.5 




3,000 

210 

3.4 

7.2 

36.3 

243.2 

1.3 x lO -4 

60.5 




1,100 

77 

272.0 

1800.0 

- 

1946.8 b 

2.6 x 10~ 6 

1.1 b 


specimens were 0.076 cm thick with 1.27-cm-wide reduced section with 2.54-cm gage length. The major axis of all 
specimens parallels the rolling direction. Annealed at 1000°C for 10 minutes and water-quenched. 
b Test terminated — no rupture. 



ass 

s 



■■■III 

■1 

s 

5 

5551 SI 

■■mi ii 
■■■■ ii 
■■■1 ii 

■1 

m 






■ 

u 


Hill ■ 




1 

■ 

■1 

IIIIH 

■ 

nil 

nr 

■II 

1 

III 

III!!!! 

!§g 



Ml 

■Him 

■IIIIH 



■mil 

■■■I”) 




■iiiim 

M 

■mi ii 

11 

■ hiiiii 

■ 

mu ii 


wmm 

■mu 


■■■■■■■ 

■■■■III 


■■■■nil 
■■■■■III 

■hiiiii 


= 

H H ■ B B ■ 

■ ■■■■I 

■ ■■■■1 


■ ■■■■1 

in 

■■llll 


Mill 

■llllll 


■■■■■■IB 

■■■■■III 


■Vllin 




Time to rupture, hours 

Fig. 5.15 - Creep-rupture data for Fe-Cr-AI-Y alloys 


• MS-59-0561 Annealed 
■ MS-60-1041 Annealed 
O □ Denotes coinciding points 


■ 3 : 

lllli.. 

HIIIII 

■III 


Temperature, °C 

Fig. 5.16 — Charpy V-notch impact data for Fe-Cr-AI-Y alloys 


Stress, kg/cm 4 

























R.T. 100 200 300 400 500 

Test temperature, °C 

Fig. 5.17 — Fracture surfaces of 1041 alloy (Heat MS-60) (Neg. P67-8-1) 



4.4 

3.9 


E 

2-8 -g 


2.2 *8 
> 

I 

t.7 LU 

1.4 

0.55 


Temperature, °C 



0.192 

0.165 
0.137 „ 

e 

0.110 1! 
■e 

j 

0.082 ■ 

I 

0.055 w 

0.027 


Fig. 5.18 — Impact specimen correlation study results 


Alloys containing 10 weight percent chromium or less showed no significant long-term 
aging; the 1541 and 2541 alloys, particularly the latter, aged extensively. Aging was con¬ 
siderable even after 100 hours for the 2541 alloy; most of the aging in the 1541 alloy oc¬ 
curred beyond 500 hours. The reason for the apparent aging of the 0561 + 3 Mo alloy after 
100 and 250 hours, but not at longer times, is unknown. 

Results of bend tests conducted on these specimens following aging are presented under 
the section, DBTT in bending. 




238 


TABLE 5.14 


SHEET BEND TEST OF ANNEALED AND AGED 


Fe-Cr-AI-Y ALLOYS 3 


Alloy 

Specimen 
Heat No. Condition 

Bend Angle at Indicated 
Temperature, degrees 
FIT 0°C -78°C 

0561 

MS-59 

Annealed 

90 

90 

90 

0561 

MS-59 

Aged 



90 

1041 

MS-60 

Annealed 

90 

90 

90 

1041 

MS-60 

Aged 



90 

1061 

MS-61 

Annealed 

90 

90 

45 

1061 

MS-61 

Aged 


90 

30 

1541 

MS-51 

Annealed 

90 

90 

90 

1541 

MS-51 

Aged 

90 

30 


2541 

MS-62 

Annealed 

90 

90 

90 b 

2541 

MS-62 

Aged 

0 



0561 +3Mo 

MS-64 

Annealed 

90 

90 

35 

0561 + 3Mo 

MS-64 

Aged 

90 

90 

30 


a Annealed: 1000°C for 10 minutes; water quenched. 

Aged: 450°C for 1000 hours. 

b Broke as bend reached 90°; other 90° bends did not break. 


TABLE 5.15 


AGING OF Fe-Cr-AI-Y a ALLOYS AT 450°C 


Alloy 

Heat No. 

Hardness versus Time at 450°C, DPH 

Annealed 

100 hr 

250 hr 

500 hr 

1000 hr 

0561 

MS-59 

188 

207 

193 

178 

187 

0561 +3Mo MS-64 

216 

256 

255 

206 

224 

1041 

MS-60 

179 

183 

179 

165 

177 

1061 

MS-61 

209 

214 

217 

194 

220 

1541 

MS-51 

196 

200 

238 

220 

267 

2541 

MS-62 

228 

274 

341 

325 

391 


a 45.5-kg vacuum-induction heats converted to 0.076-cm sheet. 


Radiation Stability 

Capsule ORM-53, which contained five bend and five creep-rupture specimens of 0561 
(heat MS-59), 0561 + 3Mo (heat MS-64), 1541 (heat MS-51), 1061 (heat MS-61), and 1041 
(heat MS-60), had completed irradiation under conditions described earlier in this report. 
Post -testing remains to be performed. 

5. 3 IRON-BASE (FERRITIC) 

Details of this work withheld pending patent clarification. 

5. 4 IRON-BASE (AUSTENITIC) 

Details of this work withheld pending patent clarification. 

5.5 SUMMARY AND CONCLUSIONS 

In the alloy studies conducted during the past year, primary emphasis has been on devel¬ 
oping alloys which have mechanical properties comparable or superior to austenitic stain¬ 
less steel. Compositions have been investigated from four alloy systems: chromium-base, 
Fe-Cr-Al-Y alloys, iron-base (ferritic), and iron-base (austenitic). Comparative tensile, 
yield, stress-rupture and creep strength, and other pertinent data are presented for the 
various alloys under study. Considerably more work will be required before thorough com¬ 
parisons can be made, but sufficient data have been obtained for preliminary comparisons. 

A comparison of the tensile and yield strengths of representative alloys within the vari¬ 
ous systems is contained in Figures 5.19 and 5.20. Data included in these and subsequent 
comparisons, except for the austenitic stainless steels, are based on tests of specimens 
from 3.6- to 5.5-kg vacuum-induction-melted laboratory heats which were extruded and 
rolled to approximately 0.076-cm sheet. Tests on all program alloys, other than the air- 
tested Fe-Cr-Al-Y system, were conducted in an argon atmosphere. The stainless steel 
data (based on air tests) are for commercial heats converted to 0.076-cm sheet and an¬ 
nealed at 1120°C for 0.5 hour and water quenched. Figure 5.19 illustrates tensile strength 
as a function of temperature for the various alloys. Beyond about 650° C, iron-base (aus¬ 
tenitic) and the chromium-base alloys are clearly superior to the austenitic stainless 
steels. This advantage appears to increase with reducing temperatures for the iron-base 







240 


(austenitic) but not for the chromium alloys. The outstanding comparative strength levels 
attained by the iron-base (ferritic) alloys up to about 550°C is noteworthy. 

Figure 5. 20 indicates that nearly all the alloys possess an advantage in yield strength 
over the stainless steels at all temperatures up to 750°C. This can be said even though, 
as indicated previously, the hot yield strength values for the program alloys are approxi¬ 
mate because of the measurement technique employed. 

A comparison of the 650°C stress-rupture properties of the various alloys is illustrated 
in Figure 5. 21. Except for Fe-Cr-Al-Y alloys, only relatively limited data are available 
for the program alloys. Nonetheless, it seems apparent that only the chromium-and iron- 
base (austenitic) alloys can be expected to have stress-rupture strength comparable or 
superior to the stainless steels. 

Creep (time — extension) curves taken from several of the stress-rupture tests are pre¬ 
sented in Figure 5. 22. These further indicate, in a preliminary way, that the chromium- 
and iron-base (austenitic) alloys are in a favorable strength position with respect to stain¬ 
less steels. Because of the measurement technique employed in these argon tests, data for 
the program alloys tend to show somewhat greater creep as a function of time, short of rup¬ 
ture, than actually occurs. 



Fig. 5.21 — 650°C stress-rupture strengths of potential fast breeder reactor 
cladding materials 


In the light of the comparative data presented, two alloy systems appear to be emerging 
as alternatives to stainless steel: iron-base (austenitic) and chromium-base. The iron-base 
(austenitic) alloys are especially attractive for their prospect of improved strength over 
the stainless steels. They should also possess improved and more predictable radiation and 
structural stability even at equivalent strength, since they are relatively pure. 

Three factors other than strength can be considered with respect to the chromium-base 
alloys: ductility, sodium corrosion resistance, and thermal stability. None of the chromium- 
base alloys tested in this program have ductile-to-brittle transition temperatures in bending 
or impact below 100°C. Current emphasis is placed on improving this situation. The sodium 
corrosion resistance of chromium-base alloys appears most promising based on preliminary 
tests; more extensive dynamic testing is necessary to characterize the behavior of these al¬ 
loys in sodium. Finally, thermal stability of chromium-base alloys, which depend upon warm- 
finishing for much of their strength, must be more clearly established, based upon annealing 
studies to date. 
































241 


OComm 304 (1400 kg/cm 2 ) □ M-313 Cr - 0.1 Y (2100 kg/cm 2 ) 

AFe-base (austenitic) — heat M-389 (1440 kg/cm 2 ) 
OComm 316 (1400 kg/cnn2) Q Fe-base (ferritic) - heat M-384 (1440 kg/cm 2 ) 



Time, hours 


Fig. 5.22 - 650°C time-extension curves at various stresses for potential fast breeder 
reactor cladding materials 


5.6 PLANS AND RECOMMENDATIONS 

Program emphasis in the immediate future will be given to iron-base (austenitic) and 
chromium-base alloys. Characterization of iron-base (ferritic) and Fe-Cr-Al-Y alloys 
will be completed since other nuclear or non-nuclear applications may benefit, but no 
new alloy development within these systems will be undertaken. 

Specific plans include the following: 

1. Initiate strength characterization of vacuum-induction-melted developmental iron-base 
(austenitic) alloys. 

2. Complete additional annealing and creep studies on warm-finished primary chromium- 
base alloys. 

3. Evaluate tensile strength and ductility of high-purity and high-strength chromium-base 
alloys. 

4. Extend strength characterization of iron-base (ferritic) alloys. 

5. Continue preliminary sodium compatibility experiments on alloys from all systems. 

6. Initiate post-irradiation studies on chromium-base and Fe-Cr-Al-Y alloys. 

























































‘•f T * *1 ‘i- O 



EVALUATIONS OF PLASTIC FATTfiiiR properties 
OF HEAT RESISTANT ALLOYS ) 


( 1516 ) 


J. B. Conway* 


rihe objective of this program is to determine the parameters affecting low-cycle fatigue 
life of metals and alloys at elevated temperatures and to generate low-cycle fatigue data for 
use in the design of structural components of high-performance nuclear reactor systems. 

Materials currently being studied in this program are the AISI 304, 348, and 316 stain¬ 
less steels. Test parameters under investigation are temperature, strain amplitude, strain 
rate, and length of hold times at peak strain in each cycle. 

6.1 MATERIAL SPECIFICATIONS 

AISI 304 STAINLESS STEEL ROD STOCK 

AISI grade—heat 55697 from PNL—billets 49 cm2 cross-sectional area, rolled at 1180°C 
to rods 16 mm in diameter-rod coiled, annealed 60 minutes at 1066°C, and water quenched - 
sections cut from coil, straightened, stress-relieved 30 minutes at 1010°C, and water 
quenched. 

AISI 348 STAINLESS STEEL ROD STOCK 



AISI grade - heat 55700 from PNL 
304 stainless steel. 


rolling and heat treatment sequence same as for Type 


AISI 316 STAINLESS STEEL ROD STOCK 


AISI grade — heat 65808 from PN.L — ladle chemistry by weight percent: 


C Mn Si Cr Ni Mo Co Sn Cu S 

0.086 1.73 0.52 18.16 13.60 2.47 0.74 0.040 0.078 0.006 

NMPO PROCESSING - AISI 304 AND 348 STAINLESS STEEL 


P _J?2_ 

0.010 0.050 


Ground specimens to hourglass configuration - diametral gage section 0.635 cm - sur¬ 
face of gage section longitudinally polished - annealed 30 minutes at 1092°C in argon, cool¬ 
ing rate approximately 100°C per minute - average VHN 139 for 304 SS and 155 for 348 
SS - grain size ASTM 3-5. 


NMPO PROCESSING - AISI 316 STAINLESS STEEL 

Sample blanks annealed 30 minutes at 1070°C in air and water quenched — ground samples 
to hourglass configuration - diametral gage section 0. 635 cm - surface of gage section 
longitudinally polished - stress relieved 60 minutes at 760°C in argon, cooling rate approxi¬ 
mately 100°C per minute - average VHN 171 - grain size ASTM 3-5. 


^Project leader and principal investigator. 


242 


? 




243 


6.2 FATIGUE-TESTING EQUIPMENT 

During the past year equipment and test technique modifications were made to improve the 
precision and accuracy of low-cycle fatigue data being generated. A more precise method 
was devised for calibrating the extensometer employed to measure diametral strain; Figure 
6.1 shows a photograph of the new calibration fixture. This fixture has permitted the exten¬ 
someter to be calibrated in a manner that closely simulates the mode of operation during 
an actual test. 

Another improvement consisted of fabricating Plexiglas enclosures for two uniaxial fatigue 
systems. One enclosure shown in Figure 6.2 permits operation in air while isolating the test¬ 
ing or sample zone from the immediate laboratory environment. In this way the effects of the 
unavoidable air currents present in the laboratory are eliminated and the test specimen tem¬ 
perature is more accurately controlled. This modification also permits tests to operate in 
inert environments if such testing becomes desirable. 

Another modification which promises to improve testing accuracy involves the reposition¬ 
ing of the load-measuring device in the load train. This change was needed to eliminate the 
extraneous load resulting from frictional forces in the die-set assembly; this is presently 
summed along with the force on the test specimen. A sketch of this new modification is pre¬ 
sented in Figure 6. 3. The new die-set fixture is being assembled and will soon be incorpo¬ 
rated into the current test program. 




Fig. 6.1 — Calibration fixture for diametral strain 
extensometer (Neg. P68-2-2D) 


Fig. 6.2 — Typical uniaxial fatigue-testing system with 
Plexiglas enclosure (Neg. P68-2-2C) 
























2 ‘ 


Load 










245 


In another important equipment modification, a strain computer shown in Figure 6.4 
was developed and installed on each fatigue-testing machine. This small analog computer 
continually determines relative amounts of elastic and plastic strain in the specimen and 
produces a signal which represents instantaneous axial strain. Once this signal has been 
generated it can be used for control purposes; the advantages of the diametral strain 
measurements in this evaluation can be combined with the ability to program and control 
the axial strain. Figure 6.5 compares the different strain and stress wave forms ob- 


ELASTIC PLASTIC TOTAL 

DIAMETRAL DIAMETRAL AXIAL 

LOAD STRAIN STRAIN STRAIN 



LOW CYCLE FATIGUE STRAIN COMPUTER 

Mode by 

NUCLEAR MATERIALS ond PROPULSION OPERATION Cincinnoti, Ohio 



COMPLIANCE 



S TA 


CONTROL MODE 

POISSON’S 

RATIO PD A ID 



PD ED F STABILITY 


Fig. 6.4 — Front panel of low-cycle fatigue strain computer (Neg. P67-7-32) 


- Axial strain controlled in triangular cyclic mode 

-Diametral strain controlled in triangular cyclic mode 



(b) Diametral strain versus time 



(c) Stress versus time 


Fig. 6.5 — Comparison of stress and strain wave forms obtained 
with programming of axial and diametral strain 
(Type 348 stainless steel at 800°C) 







246 


tained 1 in diametral and axial strain control. A triangular wave form (constant strain rate) 
was programmed in both cases. The strain computer was used to generate the axial strain 
signal; the diametral strain signal was obtained directly from the diametral extensometer. 
Such a diagram clearly shows the intra-cycle strain rate differences, but not the more 
significant, long-term strain reorientation which will change the relationship between axial 
and diametral strain. For example, the ratio of axial strain range to diametral strain 
range will increase with cyclic strain hardening and decrease with strain softening. The 
strain range being controlled, however, will remain at a constant value. 

The analog of axial strain is produced by operating upon the signals from the diametral 
strain extensometer and the load cell. The following relationships form the basis for the 
axial strain simulation: 


€ = e e + E p 

(6.1) 

“ e de 


e e = - 

^e 

(6.2) 

, _ " e dp 

P "p 

(6.3) 

-Fu e 


6de ~ AE 

(6.4) 

e dp = e d~ e de 

(6.5) 


where: 

e - total axial strain 
e e = elastic axial strain 
6p = plastic axial strain 
= total diametral strain 
e^e = elastic diametral strain 
€ dp = Plastic diametral strain 
F = force on the specimen 

A = cross sectional area of the specimen at the minimum diameter 
v e = Poisson's ratio in the elastic region (assumed to be constant) 

= Poisson's ratio in the plastic region (assumed to be 0.5) 

E = Young's modulus 

Values of A, E, v e , and L- p are readily inserted into the computer as constants, but 
F and e d are measured. The other variables are electronically determined using d-c 
amplifier summing circuits. Sufficient signals are produced to provide the option of 
programming and controlling diametral strain, axial strain, plastic strain, or force. 
To calibrate the computer it is necessary only to adjust a potentiometer in the force 
circuit for zero plastic strain while cycling the specimen in the elastic region. This 
procedure establishes the ratio v e /AE and thus the value of v e if E and A are known. 
The value of v e is then set on a front-panel potentiometer dial. When plastic strain 
control is used it is not necessary to know vq or E, since the procedure of adjusting 
for zero plastic strain in the elastic region defines all necessary relationship. 


^T. Slot and R. H. Stentz, Experimental Methods for Low-Cycle Fatigue Research at High Temperature " GE-NMPO 
GEMP-527, June 1967. 


GEMP-1004 

Errata 


f/ 


Section 2 

Figure 2.59, page 141, key to figure, fourth line from bottom 
/ Should read a u = 0.344 H v 

instead of 0,344 E v 




Figure 2.70, page 150, key to figure 
Should read A - taken from reference 40 


instead of ... reference 41 
Figure 2,84, page 167 

Identification left off of the four different parts of the figure. 
l/i°v left should be identified as Region 2A; top right. Region 7A; 
bottom left. Region 2B; and bottom right. Region 7B. 

Figure 2.86, page 170, key to figure 

' ' “ " CQ 

Row 2 for Ni.58 should be large diamond shape (^) and row 7 for Fe 3 
should be small diamond shape (Q). 

Page 171, second line from top of page 
/ Should read "The Ni58 dosimeters in regions 2B and 7B were . • 
instead of ". . • in regions 2A and 7A . . ." 

Section 5 

Page 217, second paragraph from bottom, third sentence 
Should read ", , , no intermediate anneals to 0,076 cm thickness," 
^ instead of ", . . to 0.125 cm thickness." 

Also remove the asterisk (*) after the work "thickness." 

Section 6 

The A It appearing on pages 247, 253, and 256; Tables 6.1, 6.2, 6.3, 
and 6.4; and Figures 6.6, 6.8, 6.10, 6.12, 6.17, 6.25, 6.28, 6.29, 
6.30, and 6.31 should be A e t 

Page 247, first paragraph, last sentence 

Change to read "... presented for axial strain ranges ..." 
instead of ". . . strain amplitudes . . ." 

Figure 6.18, page 260 

Left hand ordinate should read: 

Stress range, psi 


GEMP-1004 

Errata 


Section 7 

Replace Figures 7.16 (p. 305), 7.17 (p. 306), 7.18 (p. 306), and 
7.19 (p. 307) with the following: 



Fig. 7.16 - Comparison of calculated design fatigue curves for PH13-8Mo and 
12Ni—5Cr—3Mo (using the Linear Rule to correct for the maximum 
effect of mean stress) with design fatigue curve for carbon and alloy 
steels from Section III of ASME Boiler and Pressure Vessel Code for 
Nuclear Vessels 



Fig. 7.17 — Comparison of calculated design fatigue curve for Inconel 718 (using 
the Linear Rule to correct for the maximum effect of mean stress) 
with design fatigue curve for 18-8 stainless steels and nickel-chrome- 
iron alloy from Section III of ASME Boiler and Pressure Vessel Code 
for Nuclear Vessels 


Alternating stress, kg/cm 4 
















































































Alternating stress, psi Alternating stress, psi 



Fig. 7.18 — Comparison of calculated design fatigue curves for PH 13—8Mo and 
12Ni—5Cr—3Mo (using the Peterson Cubic Rule to correct for the 
maximum effect of mean stress) with design fatigue curve for carbon 
and alloy steels from Section III of ASME Boiler and Pressure Vessel 
Code for Nuclear Vessels 



Fig. 7.19 — Comparison of calculated design fatigue curve for Inconel 718 (using the 
Peterson Cubic Rule to correct for the maximum effect of mean stress) 
with design fatigue curve for 18-8 stainless steels and nickel-chrome-iron 
alloy from Section III of ASME Boiler and Pressure Vessel Code for 
Nuclear Vessels 

























6.3 TEST RESULTS 


A I, ^ £iZ 

247 


During the past year fully reversed strain cycling fatigue data were obtained in an air 
environment for fully annealed, AISI 348, 304, and 316 stainless steel to determine ef¬ 
fects of strain rate, strain range, and temperature. A summary of these data is pre¬ 
sented in Tables 6 .1, 6.2, and 6.3. As indicated, three strain rates, 4xl0" 3 , 4xl0 -4 , 
and 4x10"^ sec’*, were evaluated at test temperatures of 430°, 650°, and 816°C. Data 
are also presented for axial strain amplitudes (Ae t ) ranging from approximately 0.5 to 
approximately 2.5 percent. 

The general effects of strain rate, strain range, and temperature on the cyclic strain 
fatigue characteristics of 348, 304, and 316 SS are shown graphically in Figures 6.6 
through 6 .11. Both total axial strain range (Ae t ) and plastic strain range (Ae p ) are pre¬ 
sented in terms of N 5 , the cycles corresponding to a 5-percent reduction in load. Special 
tests were performed to identify Poisson* s ratio, employed in conversion 2 from diametral 
to axial strain, for the 316 SS used in this program. Values for Poisson 1 s ratio for the 
304 and 348 SS have not been obtained yet; hence literature values were employed in the 
present analysis. This property is being measured for these two materials and any effect 
which these new data may have on the data presented here will be discussed in a subse¬ 
quent report. A summary of the short-term tensile data generated to date is presented in 
Table 6.4. 

As indicated in Figures 6 . 6 and 6 .7, the general effect of a decrease in strain rate at 
650°C and 816°C results in a severalfold decrease in the fatigue resistance of 348 SS for 
a given strain (Aej. or Aep) range. Figure 6.6 also shows that data obtained at 650°C and 
an axial strain rate of 4x10"^ sec - * seem to coincide with data obtained at816°C and 
4x10“** sec"*. The degree of coincidence is reduced, however, if plastic strain range 
data are used as the basis for the comparison, as shown in Figure 6.7. 

Relationships similar to those indicated by the 348 SS data were obtained for the 304 SS 
as shown in Figure 6.8 and 6.9. The only apparent differences seem to be an increase in 
slope of the logarithmic plots of plastic strain range versus fatigue life, and slightly lower 
fatigue resistance for the 304 SS at the same strain levels compared to the 348 SS data. 
More data are needed for a meaningful interpretation of these slope differences. 

Fatigue data for 316 SS are shown in Figures 6.10 and 6.11. The effect of strain rate 
on fatigue resistance of 316 SS is about the same as that noted for 348 and 304 SS but the 
temperature effect is quite different. The effect of temperature on the fatigue resistance 
of 316 SS is much less pronounced than that observed in tests of the other two materials; 
at the lowest strain rate little, if any, temperature effect on fatigue resistance is noted. 
This suggests that metallurgical reactions are occurring at 816°C which tend to enhance 
the fatigue resistance of 316 SS compared to the 650°C results. Another surprising re¬ 
sult concerns the lower-than-expected fatigue resistance of the 316 SS at 650°C and the 
highest strain rate. Metallurgical evaluations are being conducted to identify the mecha¬ 
nisms involved. 

The effects of temperature and strain range (Ae t and Ae p ) on the fatigue resistance of 
316 SS at a constant strain rate of 4xl0 ~ 3 sec“* are shown in Figures 6 .12 and 6.13. 
Based on these limited data, some consistency with the Coffin-Manson relationship, 

Aep = ANf“ m , can be identified by the linear isotherms. Although N 5 was used in Figure 
6.13, the same general observations follow if Nf is employed. At 430°C the value of m 
is essentially 0 .5 but increases slightly with increasing temperature. 


2"AEC Fuels and Materials Development Program Progress Report No. 67/' GE-NMPO, GEMP-67, June 30, 1967, p. 135. 



LOW-CYCLE FATIGUE DATA 3 FOR ANNEALED AISI 348 STAINLESS STEEL TESTED IN AIR AT 650°C AND 816°C 


248 


O -5“ LOOOCDCNCMCOtDOOLfit— Tt0'vf*-000’<fr 

• — OOinOMflCONr OOCOr-p^OOO'^-COOOOO<D 

cqoqo5 0)cqo)oqps^ooo)fficq^o)0)0)0)03 

DC 2 O O O O O O O O o' O O O O d O O O O O 


in^-coqcncNr^'^'^^cooLncoooo*- 
^CNt-LncoinininNcocoi-coifi^^coin 
CO iq IN IT) oq S (D IV 00 0 (0 pH 00 00 h* 00 N O) 

o o’ o’ o’ o o o o o o’ o’ o o o o o o’ o’ 


—1 ■ “ WMi50r-lfl00ffiWCMN(£)^Wffl0(D(M 


« 05hP)NhOO©COCOP)i- 

u T-r-OQCO^-lOr^COOr-^ 


T-ontD^0<-isin 
’^CDCN^CDCMOCDtD 
(NCMdlflOMflCDO 
«- t— t- CM 


COCNOOCDCDLDLOOLDCMCOCM'^-CO^PCOOCD 

*-CMuni^(Drj-ooooo^-ooir)^ri^r^ix) 

*-r-T-T-cMCMCO'spLO-^-r^cDmcocof^roco 

T- T- T“ CM CO ^ 


3 *-». NCOOOIOtDt-OQOlOLni-r-CDSCMtOCM 

X -2r *-Of''.c\iLnor^-LnLnocooocD<— ooor-co 

gJZ CMCMCOCOLOLOCOOCMCOCDCOCDCOLnOO»— O'sf 

ii — ’-T-cNCMCor^t-r^iDpN.T- 

LL *- i — i— CM 


£2£2S Lf5r ^ CD S5£ir 0 > 00)CNJ o jLn O'-<o 

*-CNCNC 0 C 0 C 0 ini 0 C 0 f^<“O 0000 ^r-CM 00 

r- C- T- OvJ 00 


LninM-'tnnLOLn^rtcoro 0 Lf)^^nnro 

llllllllllillllllll 

ooooooooooooooooooo 

xxxxxxxxxxxxxxxxxxx 
co co ^ ^ co in ir) iflinifl(ooiffioooo)o 
^ ^ ^ ^ -t ^ lo in d ^ in 


ww't^cocoiflio^^cocotnifl^^coco 
I I I I I I I I I I I I I I I | | I 

oooooooooooooooooo 

xxxxxxxxxxxxxxxxxx 

»^'— > CNCNCNCNJ 00 C 0 ^P^t^tCO^fLrJCDCDCDCD 
^ ^ ^ "Cf ^ rfr ^ ^ ^ ^ 


it tt cm cm 

SS^cr o o cn cm oo*-^- 

oqqqt-T-qqqocNJCNOOOO^t^^- 

dddddddddddddddddo’d 


<r cm cm < 3 - 

O O t- O O CM CM O O ^ 'f 

qqqqT-^-qqoqqcMOOoO'd-^ 

oooddddddddddddddo 


o) E 
£ d £ 

DC 5-^, 

$ ? ^ 


CDCOCD'CfOOJLOCO^— i— p-. W t- CM CM CM N CM 0 
UO r-CO^CMCMCMCOr-^fMintDCOOQCOtDnCM 

0 " 10 d d cm cn in d iri cm »- is ri id’ ifj pi ri M 1 ’ 


r-cogmMQ^^-M-M-coisrNoiconhCM 

q^rqp^qqr^cDcocDococMCDCDOocM 

CMCMC0C0C0^'CM*“CMCMC0C0CMr— CMCMC0C0 


C CD 

£ 0= ? O, CM«-CMCMCMCMOOOOO*-03030QOCOCOO> 
wroc <U t— »— t— r-T— t— 0»— i—t—OO'— OOOO 


IllI C/3 DC ^ 


ooooooooooooooooooo 


525 ^ 50 99 Dcr o° go ^ lo lo lo io 

qqqqqqqqqooooooooo 
•dddddddddod dodo odd 


.0 C CD 

:K’cdc"q_ S°SJS2 SS22!: 0 2 05IsrNin(:0|s i s ® 

TO £ to "D COOOOCOCOOO^'CP'tf-^^CQr-t— r-r-T— r-T- 

ol oo a: ^ ddddddddddddddddddd 


NtOM-^CMCOIflinPJCOCMCM'-r-OOOO 

0)0)05fflfflG)'tM;^M -, tM , C'ICMCMCMCMCM 

dodddddddddddddddd 


*« c j:rc:^S^ooooooininifl(Diflii)ifl 

£ m jp qqqqqqiniflLnioininc\jCMCM(MCMCMCM 

c/30C<j ^^odddddddddddd 


»— Or— OOOOOOOOOinLOLOLOLC3LjO 

qqqqqqiflu)inininif)(\|CMPj!NNcvi 

T-T-^r-odddddodddod 


ft *ro C-O, 0£OOOr-COCM»-COOOP-.COCOCMir>r'>NO 

to £ ro ^ qq^^^^cocorococococMCNOocMCMCMCo 

ujcoDC<j oodddddddddddddddod 


oir3coop'.por>p'>.cooop*Nir)CN<frr-oor^r-. 

dddddddddddddodddo 


tS ’*5 c ojocooococoot-cmoooococoon^^oj 

TO £ TO yp- NOO^NNISOOCOCOCOCOpsCOMCOOCOCOCO 

to 0C <j t-*ddddddodddddd 


M'r-COCOM'fflr-OCOtO^lflCOCMOOfflO) 

qqqqqqqqqqooq^^^qcoq 

r-oddddddddddd 


_ c <1> S'- 

‘In : r^>P^C0COCDO3COCOCMC0COLD»— r- CM CM CM '— CM 

o t r , ir (DCD(0 ® (DtD ® 

I— to oc *<j CMCMCMCMCMpjr-r^r-'i-’ r-i-ddddddd 


CMCMCMCMCMCMr-i-i- 


*-000 000 


CM O CM CM t- 

rs0oooDi0M , »-'-0r-o)Ln^r-^oon<-M- 

cocorocococo444444coLi3cod)ind5d5 


t— o 

qr^q^fr-T^(MLnrv.(j)to^— 03^— incococM 
000000 oid)d)C)d)d)(6(D0(D(6(D 


« <a O 
£ 


a Control mode: diametral strain. 

Strain wave form: triangular with zero mean strain 





250 


CO 

CD 

LU 

_I 

CO 

< 

h- 







4- 

CM 

LO 

co 

CO 

<3- 

LO 

CD 

LO 

o 

05 

LO 

00 

CO 

o 

LO 

Mf 

Mf 

o 

Mf 

CO 

r-x 

CD 

o 






z 

CO 

00 

oi 

CO 

r-. 

00 

CO 

CD 

LO 

CM 



CD 

CM 

Mf 

CM 

LO 

CM 



co 

00 

03 







CD 

CD 

CD 

CD 

CD 

00 

CD 

00 

Cl 

O) 

03 

01 

CO 

03 

03 

03 

03 

03 

03 

CO 

05 

03 

05 





cr 

z 

o' 

O 

O 

o 

o 

o 

o 

O 

O 

o 

o 

o 

O 

O 

o 

O 

o 

O 

O 

o 

o 

o 

d 






To 

























a> 

4- 



O T3 

z 

























_l 

03 



+ J CO 
t/5 O 

c 



























® —1 

O 

T— 

00 

T— 

Mf 

CD 

o 

LO 

CO 

CO 

Mf 

CO 

o 

1^. 

CM 

Mf 


r> 


CO 

o 

on 

r- 



O) 


O SO 
>» O'- 

O « 


CO 

CO 

o 

CD 

CM 

CM 

CD 

05 

05 


8 

00 

Mf 

CO 

CO 

CO 

CM 

CO 

00 

o 

o 

CO 

m 







r- 


O 

CO 

O 



Mf 

LO 

CD 

00 

LO 

r-> 

05 

CM 

03 


CO 

CM 

CD 

co 





1 

CM 


Mf 

CO 













CM 

CO 

Mf 

CO 

Mf 


co 


LL 








CM 

r-. 























DC 
























o 





























o 

CD 




O aj 

■M u. 


CM 


00 

LO 

00 


00 

CO 

CM 

CO 

CM 

CM 

LO 


00 

03 

CO 

Mf 

CM 

o 

o 


r>. 





Q) 

H- 

O 

CO 


CM 

01 

CO 

O) 

CM 

CM 

LO 

CO 

O 

Mf 


o 

00 

CO 

o 

CO 

o 

05 

CD 

CM 





71 y 

Z 

CM 

r-* 


CO 

00 

CM 


CM 

LO 

LO 

o 

o 

ai 

CO 

00 

o 

CO 

CO 

Mf 

s 

Mf 

r>. 

LO 

o 




> 2 


CM 


Mf 

CO 


O 










CM 

CM 

Mf 

Mf 

Mf 


CO 

z 




O ll 






CM 

00 

















r_ 

< 





























o' 




Q) 


CD 

CO 

CO 

CO 

co 

CO 

LO 

Mf- 

CO 

CO 

CO 

CO 

co 

LO 

Mf 

CO 

CO 

CO 

CO 

LO 

Mf 

co 

7 

o 




+-< 

| 

I 

1 

1 

I 

I 

1 

1 

I 

| 

| 

1 

1 

I 

| 

I 


| 

t 

1 

t 

| 

1 

ID 

CD 




CO 

u 

O 

o 

o 

o 

O 

o 

o 

o 

O 

O 

o 

o 

o 

o 

o 

O 

o 

o 

o 

o 

O 

o 

o 




CO u - 

CD 

C/5 





r— 

1“ 






T— 





r— 

r— 

t— 

V— 

T— 

r* 

t— 

o' 

0> 



X C 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

X 

O 

CO 



< ro 

+-■ 

r-> 

CO 

T— 

T— 

rr 

co 

CM 

Mr 

o 

o 

CO 

Mf 

CO 

Mf 

00 

CD 

LO 

Mf 

CO 

CO 

CM 

o 

on 

CO 



CD 

«<u 

; 

o 

CM 

T_ 

co 

co 

CM 

CO 

o 

o 


CM 

T~ 

LO 

Mf 

CO 

Mf 

Mf 

LO 

CO 

CO 

00 

co 

H 

DC 




Mf 


Mf 

Mf 


Ml- 

Ml"* 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

< 





























DC 

O 

> 



> CD 

g £ 

o 



00 





















< 



CD XJ 

"«* 



CM 

CO 

Mt 





00 

CO 

00 

CM 


CM 

CM 

CO 

CO 

Mf 


Mf 

Mf 





3 5 

s 

t— 

t— 

Mf 

CM 

CD 

Mj- 

O 


*— 

T— 

Mf 

Mf 

Mf 

O 

CM 

CM 

CM 

03 

03 

o 

Mf 

Mf 

Mf 

Z 




o 

*5 

t— 


*— 

CM 

CM 

Ml- 

o 

o 






o 

O 

CM 

CM 

CM 

CM 

o 

O 

Mf 

Mf 

o 

LU 




> 

o 

o 

o 

o 

O 

O 

o 

o 

o 

o 

o 

o 

o 

o 

o 

O 

O 

o 

o 

o 

o 

d 

o 

o 

j_ 





























CD 

LU 

1- 

% 

03 



CD 

CD 

c 

CM 

E 
























_J 




co CM 

E 

r^. 

CM 

00 


o 

CM 

CM 

CD 

CO 

CO 

T- 

r»» 

CO 

CM 

Mf 

Mf 

CD 

CO 

LO 

<0 

Mf 

C0 

05 

LU 




a_ H_ 

CD 

o 

<J> 

CM 

CD 

CM 

O) 

Mf 

LO 

05 

o 

00 

CD 

Mf 

CO 

r» 

03 

O 

r*«. 

p-. 

CO 

CO 

Mf 

03 

LU 

<0 



ffl 2 

u: 

co 

co 

CO 

Mf 

O 

05 

o 

O 

Mf 

CM 

on 

r*- 

CD 

O 

00 

CM 

1— 

CM 

t— 

CM 

05 

CD 

CO 

|— 

X 



CD 


r- 

r-» 

CO 

LO 

LO 

Ml- 

CD 

r-* 

r-- 

r^. 

CO 

CD 

CO 

LO 

LO 

CO 

CO 

LO 

LO 

Mf 

Mf 

Mf 

Mf 

CD 

< 



CD 

o 

<1 
























(/) 





























LU 





























_J 





























z 





s? 
























< 


.y 

c s, 


00 

LO 

LO 

03 

CD 

CO 


CO 

Mf 

Mf 

Mf 

Mf 

Mf 

O 


CO 

CO 


CM 

r-. 

O) 

03 

05 

1 — 
CD 
CD 


CO 

LU 

2 c 
■3 co 
CD QC 

TJ 

CD 

o 

o 

o 

C_) 

d 

O 

o 

o 

o 

o 

o 

o 

d 

d 

o 

o 

o 

o 

o 

d 

o 

o 

o 

o 

O 

d 

O 

o 

O 

o 

CO 

c 




























CD 

< 

2 

CD 

O 

on 

CO 

c & 
*5 cP 

ss 

‘"'q. 

00 

LO 

LO 



CM 

CD 

r- 

co 

CO 

CO 

00 

LO 

o 

03 

r*. 


o 

05 

co 

CO 



Q 



TJ 

co 

CO 

Mf 

CO 

CM 

’— 

CO 

00 

co 

CD 

Mf 

Mf 

Mf 

Mf 

CO 

CM 

CM 

CM 




*— 

T — 

LU 

03 

e 

O- 

CD DC 

CD 

< 

o 

o 

o 

d 

d 

o 

o 

o 

o 

d 

d 

d 

o 

o 

o 

o 

o 

d 

o 

o 

o 

o 

o 

< 

CO 




























LU 

Q 




























Z 

z 




C o 

s? 
























< 


2 

CO C* 


*— 

o 

o 

o 

o 

o 

CD 

o 

a 

o 

o 

CM 

o 

o 

CD 

o 

T _ 

o 

o 

LO 

LO 

o 

o 

DC 

o 


c 

0 

2 CO 

c? 

< 

CO 

00 

CO 

■'tf 

CO 

CM 

05 

o 

00 

00 

co 

CD 

CD 

LO 


Mf 

Mf 

CO 

CO 

CM 

CM 

CM 

CM 


M- 

CD DC 

o 

d 

d 

o 

o 

d 

o 


d 

o 

o 

o 

o 

o 

o 

o 

o 

o 

d 

d 

o 

o 

d 

LL 





























< 

h- 

< 


CJ 

.£ & 
2 £= 

85 

o> 


p> 

00 

CO 

LO 

LO 

CM 

LO 

LO 

LO 

LO 

LO 

CM 

LO 



LO 

00 

CM 

co 

03 

03 

Q 



_ 

CD tr 

-05 

LO 

'Sf 

Mf 

CM 

CM 

CM 

CO 

Mf 

Mf 

Mf 

Mf 

Mf 

Mf 

CO 

CO 

Mf 


CO 

co 

CM 

CM 

CM 

CM 

LU 


r 

< 

d 

o 

o 

o 

o 

o 

d 

o 

o 

o 

o 

o 

d 

o 

d 

o 

o 

o 

o 

o 

o 

o 

o 

=> 






























C 




























f- 

'co 

■M 

_o 

C CD 

CT'' 
























< 

CD 


CT 
CD c 

—T 

co 

o 

<— 

co 

CM 

MT 

CD 

LO 

CM 

CM 

CM 

LO 

r~ 

05 

rv 

Mf 

LO 

05 

00 

CO 

CM 

CM 

co 

LL 

15 

<0 


CL 

CD 

<\ 

CM 

co 

O) 

CO 


CM 

r^ 

r>. 

CO 

CO 

05 

CD 

01 

r>. 


LO 

LO 

CO 

co 

CO 

CO 

CM 

CM 

LU 

x 

CL 

cd tr 


1— 

d 

o 

o 

d 

T™ 


T“■ 

T—’ 

d 

O 

o 

o 

o 

d 

o 

d 

o 

o 

d 

o 

o 

_J 

o 

< 




























> 





























o 




c ^ 
























£ 

o 

-1 


CO 

'co c 

~ 

LO 


00 

r— 

o 

CD 

T— 

CO 

T--. 

r-» 

1^. 

o 

co 

T— 

CM 

LO 

co 

Mf 

co 

CO 

o 

,— 

CM 


O 

1- 

H CO 

CD QC 

•CD - 

< 

00 

r>- 

CO 

03 

d 

d 

Ml- 

o 

CM 

cm’ 

r> 


CO 

Mf 

co 


- 

03 

d 

03 

d 

r-*. 

o 

r*> 

o 

LO 

d 

co 

o 

LO 

o 

LO 

o 





c 

CD 

E 


CO 


LO 



00 

CO 





CM 



o 


CM 

LO 

co 



co 






6 

Mf 

r>. 

co 

LO 


CM 

T- 

Mf 


CM 


CM 

CO 



r>. 

r— 

CO 

Mf 





8 

z 

Mf 

4 

Mf 

O’ 



T— 

CO 

CO 

Mf 

Mf 

CO 


CO 

A 

r^ 

CO 

r^. 


rL 

CD 

CO 

fi. 





Q. 


r— 

T— 

’— 

»— 

r- 

t— 

CM 


CO 

T— 

t— 

CO 

CM 

r— 

CM 

CM 

CO 

CM 

CM 

CM 


CO 

CM 





CD 

0) 





























Test 

eratu 

<_> 

o 

430 






650 





















CL 





























E 





























i2 


























T3 

O 

E 

o 


o 

o 


Strain wave form: triangular with zero mean strain. 



251 


o 

O 


CO 

a 

z 

< 

o ‘ 
o 
in 
to 

o' 

o 

CO 

h- 

< 

oc 

< 

z 

Q 

LU 


o ■= 


CO 

to 


CO 

< 



4- 

CO 


to 

in 

CO 

CM 

o 

CO 

LO 

CO 

CO 

p*. 

LO 

CO 

05 

LO 

r-> 

o 

05 

o 

z 

CO 

05 

CD 

l n 


Ml* 

CM 


CO 

CM 

to 

o 

LO 

o 

05 

CO 

05 


CM 



to 


r-. 

r» 

CO 

oo 

CO 

o> 

CO 

CO 

r>. 

CO 

CO 

CO 

CO 

00 

r-« 

00 

05 

CU 

oc 

z 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

d 

o 

o 

O 

o 

o 


LO 

o ,5 

*■ 5 c 

jg 3 .1 
> S? I 


O Q3 

W 3 

® s 

o £ s 

O u. 


X 
< 




c a 

.11 


Q> CM 

? E 
ir ^ 

8 z" ® 


co 


to 

< 


•S ~ s> * 

to CD C flj 
CD W. CD -S 
LU GO OC <jf 


s£ 

o c Q> . 

t{ 5 c Q. 

<2 +- «j -o 

s: CO DC 

<1 


5? 

15 .E §> -: 
ti « c *o 

O h (D (u 

r w dc <i 


.52 .E 8) 3 s 


CO DC ® 

< 


o r ® 

*3 .S 05 - 

V) CD C —_ 

.2 £ « uj a 

OL CO DC 


se 

■ .£ g> 

J ™ C H 
o ^ flj .lu 
h w DC <1 


.i d 
8 z 
«& 


v- 3 

CO +± O 

® £ o 
r- 03 

a 

E 


r»T-ioo»-ioooooo^-r^t35t-cocMOOco^'LO 

ocor-f-o)coincM^cM^O)oon^coo«-cofN 

i-T-CMCMCMCO<DCO^-LOLOT-r~0)r*.intOr^O)M 


a)LOOCOOOCOCM^rNr^OJLD'-(DOOC\tD’“COr- 

(O(ONNinO5OCOO)M*-COOOOnCOON^0 

p-r.rMMOOPOOOtD^tDNfn^O)COO«-^0) 

t-r-CMr-r.OJNtD't 


inio^t’^cocorocoiOLO'^tcocorocoLniO'd-coro 

i i i i i i i i i i i i t i i L L k L L 

oooooooooooooooooooo 

xxxxxxxxxxxxxxxxxxxx 
t0inOC0r-r-05inCMpNr-00inC0C0CM05in05P-. 
i->-fOMCOcococ»^^«-ooqr-rs(scqcNifl 
^ ’t n to co co ^ ^ 't n-’ ^ ^ ^ 


t~ r- r- t- CO CO CM CM ^(NtDtOJJ ^-^ 

OOi-r-t-'-^^OOCNIMtNffimOO'JTt^ 

0000 *-«“*-*- 000 CMCMCMCMqqq«tf^ 

ddddddddddddddo' ddddo 


foSSo^^cddjcdfficqifl^bifl 
tricdcococo'r^cip^cor^irir^cdirir^iriiricM'dcM 
CMCMcoco'd-^frcococMCMCocococococMCMcococo 


oooooooqi 

ddddddddo'ddddooooooo 


©’d’rr'WNNN^^COMfOCOnOlOO^^ 

O)O)O)OlNMnifi^^f0C0COCM(Nr-WCNr-r. 

ddddddddddddddoddddd 


1-0000000000000 


o o o o o 


p'.r-co(ocococop'.coiop'.cocoina 5 *-*-CMina 5 

’•“CMCOCOCOCOCOCOCMCMCMCOCOCMCMCMCMCMCMCM 

ddddddddddddddoddddd 


CMNNi"'f’f^^O)N(OlD0MOO)O5fflC)® 

0)CD00C0^^oOC0C0N<D(0't'tC0C0D1fMCN 

P j^: r j T -: f j T -: 1 j T -:dddddddddddd 



CO 

LO 


r>. 

Pn 

p>. 



CM 

CO 

CO 

05 

CM 

in 

05 

o 



CO 


o 

o 



p*. 

p-. 

co 


*— 

*— 

o 

05 

05 

p* 

p> 

in 

CO 

CO 

in 

in 


CM 

CM 

CM 

CM 


t “ 

r ~ 





o 

o 

d 

d 

o 

d 

d 

d 

o 

C 





















« 





















CD 





to 

p^ 


CO 




05 

't 

CO 

o 

o 



in 

CM 

t_ 

CM 

*$ 

05 

CO 

CO 

CO 

r— 


«— 

05 

in 

CO 


*— 

<3- 

.— 

in 

CO 

CO 

«“ 

Q) 

,1 


to 


CO 

CO 

f^. 

p* 

CO 


to 

CO 

p> 

p- 

co 

CO 

■ r— 

to 

CO 

P^ 

E 

CM 

CM 


CM 

CO 

CO 

CM 

CM 




CO 

CM 

CM 

CO 

r- 

CM 


CO 

CM 

.2 


to 

CO 


E -C 


f I 

1 § 
_ ra 

2 5 

i I 
s* & 




252 



Fig. 6.6 - Total axial strain range versus fatigue life (N 5 ) for AISI 348 stainless steel tested in air 

A comparison of plastic strain fatigue characteristics of 348, 304, and 316 SS obtained 
at 430°, 650°, and 816°C and a strain rate of 4x 10' 3 sec -1 is shown in Figures 6.14, 6.15, 
and 6.16. At 43.0°C 3 the 316 SS is slightly less fatigue resistant than the 348 SS. No data 
have been obtained for 304 SS at this temperature. At 650°C the 316 SS is considerably 
less fatigue resistant than the 348 SS and slightly less than the 304 SS. Based on the data 
at 816°C, the 348 and 316 SS exhibit similar fatigue characteristics for a given strain 
level, whereas 304 SS is slightly less fatigue resistant than either of these two materials 
over the strain range evaluated. 

Plots of stress range versus total axial strain range for 316 SS at 650°C and 816°C and 
for a constant strain rate of 4x 10" 3 sec -1 are shown in Figure 6.17. The stress values 
represent steady-state stress range data and were usually obtained at Nf/2. In general, 

316 SS data for 650°C and 816°C exhibited cyclic hardening followed by the steady-state 
stress range indicated in Tables 6.1, 6.2, and 6.3. At 430°C, however, it exhibited 
cyclic hardening followed by cyclic softening; this phenomenon was observed over the en¬ 
tire strain range evaluated, as shown in Figure 6.18. In this evaluation, axial strain 
range data were calculated from diametral strain measurements obtained at the steady- 
state stress range which occurred after most cyclic hardening or cyclic softening had 
been completed. 

6.4 FATIGUE DATA ANALYSIS 

An analysis of fatigue data generated to date on 304, 348, and 316 SS at 650°C and 816°C 
for an approximate axial strain rate of 4xl0' 3 sec' 1 is shown in Figures 6.19 through 

3 AEC Fuels and Materials Development Program Progress Report No. 71," GE-NMPO, GEMP-1002, December 29, 1967, p. 96. 






253 


TABLE 6.4 


SUMMARY OF SHORT-TERM TENSILE DATA FOR 
AISI TYPES 348, 304, AND 316 STAINLESS STEEL 


Fully Annealed 
Stainless Steel 

Test 

Temperature, 

°C 

Young's 
Modulus (E), 
10 3 kg/mm 2 

Axial Strain 
Rate (e t ), 
sec”^ 

Reduction 
in Area, 

% 

Poisson's 
Ratio (elastic), 

v e 

AISI 316 

20 

21.16 

4 x 10“ 3 


a 

a 


430 

16.87 

4 x icr 3 

60.0 

0.32 


430 

16.87 

4 x10-5 

57.1 

0.32 


650 

15.43 

4 x 10“ 3 

60.4 

0.31 


650 

15.43 

4 x icr 5 

33.4 

0.31 


816 

12.94 

4 x10“ 3 

56.5 

0.26 


816 

12.94 

4 x icr 5 

49.8 

0.26 

AISI 304 

20 

20.18 

4x10 —3 


i 

0.3 b 


20 

20.18 

4 x icr 5 





430 

16.45 

4x10-3 



0.3 b 


430 

16.45 

4 x icr 5 





650 

15.19 

4 x10-3 

i 


0.3 b 


650 

15.19 

4x nr 5 

1 

f 



816 

a 

a 

I 


0.3 b 

AISI 348 

20 

19.82 

4 x10-3 

! 

) a 

0.3 b 


20 

19.82 

4 x10“ 5 

[ 




430 

16.73 

4 x10-3 

V 


0.3 b 


430 

16.73 

4x 10— 5 

| 




650 

15.33 

4x10-3 



0.3 b 


650 

15.33 

4 x10“ 5 





816 

13.39 

4 x10~3 

i 


0.3 b 


816 

13.39 

4 x10 -5 

1 




a Values are being determined, 
kAssumed values. 

Poisson's Ratio (plastic) = 0.5. 


6 . 24. In this analysis, the Manson-Halford 4 approach was employed to predict low-cycle 
fatigue behavior. These calculations followed from the expression: 


3. 5 0.. 


Aet =■ 


' N f 


0 - 12 + d 0.6 m .- 0*6 


N. 


f 


( 6 . 6 ) 


where: 

o u is the ultimate tensile strength 
E is the modulus of elasticity 


RA is reduction in area expressed as a fraction 
Nf is the number of cycles to failure 

Using the short-term tensile data presented in Table 6.5, the Nf values corresponding 
to assumed Ae^ values were calculated. Based on the Manson-Halford suggestions, the 
lower-, average-, and upper-bound curves were positioned. 

The short-term tensile property data and Poisson's Ratio needed for the conversion of 
diametral strain to axial strain are handbook values; hence the correlation obtained should 
be considered preliminary. Short-term tensile data and values for Poisson's Ratio are be¬ 
ing generated for the alloys under investigation to secure a better correlation. As shown 
in Figures 6.19 through 6.24, data for the three alloys evaluated generally lie near the 

^S. S. Manson and G. Halford, "A Method of Estimating High-Temperature Low-Cycle Fatigue Behavior of Metals," NASA 
TM X-52270, 1967. 



Total axial strain range (Ae\), % Axial plastic strain range (Ae p ), % 


254 



Fig. 6.7 — Axial plastic strain range versus fatigue life (N 5 ) for AISI 348 stainless steel tested in air 



Fig. 6.8 — Total axial strain range versus fatigue life (N 5 ) for AISI 304 stainless steel tested in air 










Fig. 6.10 - Total axial strain range versus fatigue life (N 5 ) for AISI 316 stainless steel tested in air 































256 



1 <£ 103 104 

Fatigue life, N 5 {cycles to 5% reduction in load) 


Fig. 6.11 - Axial plastic strain range versus fatigue life (N 5 ) for AISl 316 stainless steel tested in air 

upper-bound of the predicted curve, except for the 348 SS data at 816°C. Data in the high- 
life region (Ae^O. 5%) for both 650°C and 816°C become more conservative than those pre¬ 
dicted for the upper-bound limit using the expression mentioned. Work will continue using 
the short-term property data characteristic of the material being evaluated. Additional 
correlations that include the lower strain rate data will also be evaluated. 

In all evaluations mentioned above, two definitions of fatigue life were encountered. In 
addition to the number of cycles to failure, Nf, the number of cycles to a 5-percent reduc¬ 
tion in load, N 5 , was employed. A comparison of N 5 to Nf is presented in Tables 6.1, 6.2, 
and 6.3. Since N5 corresponds to a fairly significant drop in the load, this point is expect¬ 
ed to occur near the failure point. For this reason the ratio N$/Nf should be a fraction 
closer to unity than to zero. If a value of 0. 75 is selected, data in Tables 6.1, 6.2, and 
6 . 3 indicate only 16 of 99 data points with Ns/Nf values below 0. 75. Other interesting 
data obtained in this evaluation are: 


1 . 


2 . 


3. 

4. 


Only 1 of 42 data points for 316 SS has a ratio of Ns/Nf below 0.75; this compares to 
5 of 20 for 304 SS and 10 of 37 for 348 SS. Obviously for the 316 SS fracture is closer 
at hand when the N 5 point is reached. 

Ranges of the Ng/Nf values are: 

0. 524 to 0. 988 for 348 SS 
0.615 to 0.970 for 304 SS 
0.633 to 0.996 for 316 SS. 

Twelve of the 16 N§/Nf values below 0.75 correspond to tests at 816°C. 

Only two of the 16 Ns/Nf values below 0.75 were obtained in tests at a strain rate of 
4x10“^ sec~*. 


5. Only three of the 16 Ns/Nf values below 0.75 correspond to strain range tests below 
1 percent. 





Axial plastic strain range (Ae ), % 


257 



Fig. 6.12 - Total axial strain range versus fatigue life (N 5 ) 
axial strain rate = 4 x 10“3 sec -1 


for AISI 316 stainless steel tested in air; 



Fig. 6.13 - Axial plastic strain range versus fatigue life (N 5 ) for AISI 316 stainless steel tested in air; 
axial strain rate ^ = 4 x 1 0 — 3 sec~1 


































lo.o- 



range for AISI 316 stainless steel at 
a strain rate of 4 x 10""^ sec“"1 
























260 



Fig. 6.18 - Stress range versus cycles for AISI 316 stainless steel at 430 U C and an axial 
strain rate of 4 x 10“3 sec“^ 



Fig. 6.19 — Correlation of high-temperature fatigue data obtained for 
AISI 304 stainless steel at 650°C and a constant strain 
rate of ~4 x 10 — 3 sec“1 



100 101 102 103 104 105 


Number of cycles to failure, Nf 


Fig. 6.20 - Correlation of high-temperature fatigue data obtained for 
AISI 348 stainless steel at 650°C and a constant strain 
rate of ~4 x 10 — 3 sec - ^ 


Stress range, kg/mm' 

















































































Total axial attain range IA„], % To,al “ lal strain ran 3 a % Total axial attain range (A« t ), % 


261 



Number of cycles to failure, Nf 


Fig. 6.21 — Correlation of high-temperature fatigue data obtained for 
AISI 316 stainless steel at 650°C and a constant strain 
rate of ~4 x 10“3 sec -1 



Fig. 6.22 — Correlation of high-temperature fatigue data obtained for 
AISI 304 stainless steel at 816°C and a constant strain 
rate of ~4 x 10 -3 sec” 1 



S 





■■ 

mm 

SB 






■I 




B 

■ 

■ 

■ 

■■i 

SIS 

S 

5 



■ 

5 

Hi 


s 



5 

■ 

■ ■a 

■ ■a 

■ ■1 




21 






■1 





Bib 

S 



S 

5 



B 














■ 








■ 

i 






« 

i 

a 

Bl 


■ 







■ 





HH 


■ 







■ 

i 






■ 

M 

2 

9 

11 


■ 





BB 







H 



i 




| 

■ 


■ 





s 




i 



■ 

1 




HP 


■ 





■ 



1 




B 

i 

I 

1 

1 



1 

s 

j 



i 

hi! 


s 


I 



1 

• 

B 

i 

1 

| 




H 

B 

s 

M 

■ 


■ 



S 




m 






■Si 



MU 


■ 

■ 

■ 

Si 

IBB 




■ 






■ 

■ 




HBf 


■ 



in 

i b 



■ 


■HI 


■■ 


■ 


■ 

■1 


^B 

IB 

■ 

■ 

■ 

in 



■ 

■ 

■ 








i 

in 


BB 


■ 


Bll 




k 

s 

1 

II 

i„^B 

iftv 


■ 

■ 


in 




1 

1 



jjjg} 

■ 

■ 

■ 

1 

1 

in 


H 

■ 

■ 


III 

■ 


ft 

S 

5 

1 

1 



ft 

£ 

i 

1 

ii* 

n 

H 

■ 







■ 

■ 


■ 


hi 



■ 



w 


■ 

1 

1 

■ 

1 

H 

■ 

1 

m 

1 

1 


1 




1 




i 

in 

■ 

■ 




III 








■ 

■ 



1 

l 

m 


0 T ■ 1 1 L-J-JJLiJJ -1-1—I 1 1 I ill-1-1—i i i H U-'-i— ' i I i I 11 . 

" /- |00 101 102 103 104 105 

Number of cycles to failure, Nf 


Fig. 6.23 - Correlation of high-temperature fatigue data obtained tor 
AISI 348 stainless steel at 816°C and a constant strain 
rate of ~4 x 10”^ sec” 1 



































































































262 



Fig. 6.24 - Correlation of high-temperature fatigue data obtained for 
AISI 316 stainless steel at 816°C and a constant strain 
rate of ~4 x 10~3 sec”1 


TABLE 6.5 


SHORT-TERM TENSILE DATA 3 FOR CORRELATION OF 
HIGH-TEMPERATURE FATIGUE RESULTS OBTAINED FOR 
AISI 304, 348, AND 316 STAINLESS STEEL 


Material 

Temperature, 

°C 

Ultimate Tensile 
Strength, kg/cm^ 

Young's Modulus, 

kg/cm^ x 10"® 

D b 

304 

650 

2882 

1.546 

0.867 

348 

65Q 

3233 

1.398 

1.295 

316 

650 

3093 

1.476 

0.756 

304 

816 

1406 

1.335 

0.545 

348 

816 

1575 

1.272 

1.87 

316 

816 

2039 

1.335 

0.60 


a Literature data. 

b D is In —1- 

1-RA 


6 . Only three of the 16 Ng/Nf values below 0.75 correspond to frequencies above 0.01 


sec 


-1 


Although no detailed study has been made of this relationship, attention has been focused 
recently on evaluating the condition of the specimen at the N 5 point. Visual observations 
during test have shown that the specimen is definitely cracked at this point but no quantita¬ 
tive data were available to define the extent of cracking. Post-test metallographic evalua¬ 
tions have been initiated in an attempt to acquire more specific information relating to 
crack initiation and propagation. A fractographic analysis based on both light and electron 
microscopy has been made to yield measurements of the fatigue striations. These data 
were used to plot crack length as a function of the number of fatigue cycles. In the two 
plots obtained to date (Figure 6.25), the measured N 5 point was found to correspond to a 
crack length of about 1.7 mm. As this study continues, some indication should be ob¬ 
tained of the crack initiation point or of the crack length, where the demand strain remains 
fairly uniform over the controlled cross section. 








































263 



Fig. 6.25 - Relationship of fatigue cycles to crack length for AISI 304 stainless 
steel tested in air at 650°C; Ae t = 2.1% 

6.5 METALLOGRAPHIC AND FRACTOGRAPHIC ANALYSIS OF LOW-CYCLE FATIGUE 
SPECIMENS 

Post-test metallographic and fractographic analyses of low-cycle fatigue specimens 
were initiated. Information obtained in this study is expected to provide some understand¬ 
ing of behavior.observed in the experimental fatigue testing program, and of basic mech¬ 
anisms involved in crack initiation, crack propagation, and fracture phases of metal 
fatigue. 

Standard metallographic techniques described previously, 5 were used to characterize 
the fracture mode of AISI 304 and 348 stainless steel as a function of strain rate, strain 
range, and temperature. The specimens were sectioned longitudinally through the area 
where both the initial and final fracture occurred, and this cross section was prepared 
for metallographic evaluation. Photomicrographs of the structure of these two materials 
in the pre-test condition are shown in Figures 6. 26 and 6. 27. AISI 304 SS had a uniform 
equiaxed grain structure with twins in some grains and an ASTM grain size of 2 to 3. 

AISI 348 SS had a duplex grain structure with a fine-grained center surrounded by a coarse¬ 
grained outer layer. The fine-grained portion varied from one fatigue specimen to another. 
Because of this duplex structure, two ASTM grain sizes are reported: fine-grained ma¬ 
terial 10, and coarse-grained material 4 to 5. 

Metallographic analyses were made of the cross sections of the fractures of 304 SS 
tested at 650°C, at two strain ranges (2 and 0.6%), and at three strain rates (4xl0 -5 , 
xlO" 4 , and xlO -3 sec -1 ). The crack initiated and propagated transgranularly in the spec¬ 
imens tested at the highest strain rate at both strain ranges. In specimens tested at the 
lower strain rates, the cracks initiated intergranularly and propagated both trans- and 
intergranularly. At the final fracture in all specimens, the grains were elongated and 
cracks appeared on the outer surfaces. Metallographic examination revealed a grain 
boundary phase in all fatigue specimens. Etch pits were seen in the highly stressed areas, 
indicating the dislocations which formed in the stressed metal. 

AISI 304 SS specimens tested at 816°C, at two different strain ranges, and at three dif¬ 
ferent strain rates were also analyzed. In these specimens the crack initiated and propa- 


5 GEMP-1002, pp. 102-105. 





Fig. 6.26 — Photomicrograph showing pre-test structure of AISI 304 stainless steel 
(longitudinal cross section) used in low-cycle fatigue testing. 

(Neg. 10282; 22 H 2 SO 4 , 12 H 2 O 2 [30%] ,661^0, electrolytic 
etchant; 100X) 



Fig. 6.27 — Photomicrograph showing pre-test structure of AISI 348 stainless steel 
(longitudinal cross section) used in low-cycle fatigue testing. 

(Neg. 10281; 5 OHNO 3 , 5 OH 2 O, electrolytic etchant; 100X) 








265 


gated intergranularly; at the final fracture the grains were elongated. In the highly strained 
regions of the 2-percent strain range specimens, subgrains were observed at the lowest 
strain rate, subgrains and etch pits at the intermediate strain rate, and etch pits at the 
highest strain rate. In the highly strained regions of the 0.6-percent strain range speci¬ 
mens, no etch pits or subgrains were observed at the lowest strain rate; at the interme¬ 
diate strain rate etch pits were observed, and at the highest strain rate subgrains appeared. 
At the lowest strain rate the grain boundary phase was intermittent; at the highest strain 
rates it was continuous (see Figure 6.28). 

Post-test metallographic analyses were performed on AISI 348 SS specimens tested at 
two temperatures, three strain rates, and two strain ranges. As stated previously, this 
material had a duplex grain structure, and the amount of fine-grained material varied 
from specimen to specimen. There was no change in the grain boundary phase in this ma¬ 
terial, as was seen in the 304 SS. In each specimen the crack initiated intergranularly in 
the large-grained areas and propagated transgranularly. Numerous secondary cracks oc¬ 
curred on the outer periphery of the specimens tested at the 2-percent strain range at 
both temperatures, but at the 0.6-percent strain range, only a few small secondary cracks 
appeared. The fractures varied considerably, depending upon the strain range. At the low¬ 
est strain range fractures were almost straight across except where the final separation 
took place (compare Figures 6.29 and 6. 30). Another difference appeared in the fatigue 
specimens tested at 816°C at the lowest strain range; a series of cracks developed at an 
angle approximately 60 degrees to the fractured surface in the direction the crack was 
propagating (see Figure 6.31). Indications of this condition appeared in other tests but 
they were not nearly so pronounced. 

Results of metallographic analyses of 304 and 348 SS specimens are summarized in 
Table 6.6. 

To characterize the AISI 316 SS for the low-cycle fatigue program, longitudinal and 
transverse cross sections were prepared for metallographic analyses. The analysis and 
pre-test processing were reported previously. 6 No differences were noted in the two 
cross sections, and the material had been fully recrystallized with a uniform equiaxed 
grain structure (Figure 6.32); ASTM grain size was 4 to 5. Twins were noted in some of 
the grains. 

To study the fractured surfaces of low-cycle fatigue specimens at higher magnification, 
techniques were developed for replicating the area of interest for study in the electron mi¬ 
croscope. 7 These procedures are used extensively. Each striation on the fractured surface 
(Figure 6.33) is the result of one cycle in the fatigue test. A study is in progress to corre¬ 
late this striation density with strain rate, strain range, and temperature of materials be¬ 
ing tested. Striations could not be resolved at 1500X at the crack initiation point for the 
304 SS tested at the lowest strain rate (~4xl0 - ^ sec"■*■), and a strain range of 2 percent. 
This is apparently due to the low strain rate imposed during testing. In the 304 SS tested 
at 650°C, 2-percent strain range, and at the two highest strain rates, the cracks had propa¬ 
gated approximately 30 percent of the diameter (~2 mm) of the test specimen at the N 5 
point. 

6.6 SUMMARY AND CONCLUSIONS 

Several modifications incorporated in the four low-cycle fatigue systems this year im¬ 
proved strain programming versatility, load measurement sensitivity, and strain cali¬ 
bration accuracy. 

6 GEMP-1002, p. 94. 

7 GEMP-1002, p. 105. 




Test 7—1, Ae t = 2.08%, e t = 4.2 x 10 _o sec _l (Neg. 10138) 

Fig. 6:28 — Photomicrographs taken near the crack initiation point showing the 
grain boundary phase in AISI 304 stainless steel tested at 816°C. 
( 22 H 2 S 04 , 12 H 2 O 2 [30%], 66 H 2 O, electrolytic etchant; 1000X) 






267 




Neg. 10122 Test 3-4, Ae t = 2.19%, e t = 4.4 x 10 3 sec 1 Neg. 10121 


Fig. 6.29 — Composite photomicrograph of longitudinal cross section of 
AISI 348 stainless steel tested at 650°C.(22H2SO4,12H2O2 
[30%], 66 H 2 O, electrolytic etchant; ~25X) 






268 



Fig. 6.30 - Composite photomicrograph of longitudinal cross section of 
AISI 348 stainless steel tested at 650°C. ( 22 H 2 SO 4 , 12 H 2 O 2 
[30%], 661^0, electrolytic etchant; ~25X) 











Neg. 10078 Test 6-2, Ae t = 0.57%, e t = 4.5 x 10 3 sec 1 Neg. 10077 

Fig. 6.31 — Composite photomicrograph of longitudinal cross section of 
AISI 348 stainless steel tested at 816°C.(22H2SC>4, 12 H 2 O 2 
[30%], 66 H 2 O, electrolytic etchant; ~25X) 






270 


CO 

UJ 


OQ 

< 


o o s ^ 

si®! 

■§ 

S £ 


-2 1± +- c 

3 g\E o 
c ci> o +5 

2 J Q. 10 C 

g* * £ •e E 

S id g - 


o> _* .u 
*B o +S 
O <o .2 

s “? 


2 .2 T3 

03 ^ O O ID 

o w □ J Z 

u ® .E ~ 

° oc 


75 1 * g 
X 2 ra « 
< 00 c ~ 

•to 


n c “ ^ 

X £ £ "£• 

<5 <23 


tt O 


fc £P 

a o 
E 


E 

'§ 

CL 

</) 


.b c 

■g 1 
£ > 


CL « 

■8 *5- 
£ -5 

■M fi 

c fo 

a> .* 

*n c/a 

1 4 

3 co 

t/> *D 

3 i 
§.8 g 
1 .= .2 
(9 2 S’ 

U Ol u 


XJ ro 
® h 

t ts 

« > 
-Q J= 
O t O) 

o) !E 

w _ 
<0 C 

D. £3 


■ 4 = £ 
c o> 

® . * 

? 4 

3 ra 
w "D 
3 C 
O 3 
3 O 
.£ -° c 

E .E .2 

g 2 g> 

U D1 u 


O) o 

c 'a 
0) 

XI »- 

Si 


C -C 


II 

i 5 


2 jj s 

w o w 

> O "O 

— w c 

f a 43 


•— CO 

c .E 

® CO 

s fc 

C -Q 

2 ? 


E-S 
fe § 
53° 


3 CO 

CO 


.8 f 


3 

o 

2 5 "3 
.E X3 ® 
£ c .E 
c 3 co 
o o b 


O n 


CO 


.2 -g 

-o .E 

0) CO 

g s 

JS >• 

O -C 
03 .E 5 
CO -«= 

x: c 
a — 
-a P 
.2 a 


3 4= 

to ■ - 
§1 
1 -3 

o ^ 
(J 5 


co 


co 


3 3 3 3 3 3 3 

C > C > C C > C > C >C 

+ C0 .*3 CO ” tO _l <0 “ CO “ CO “m 

w J- 1_ + t- U 1- I L z £ 

I O) (o D) (o O) 1 O) m O) m O) m 03 

CO C/3 f— CO L (A ^ V) VI 2 c 

®cEcEc®cEc£c E a> 
c £ := 2 *= 2 c £ ? 2 ■= 2 !? £ 

— 4 -< D_ 4 -> Q_ 4 -* — +-- Q- +J Q_ +-* CL .b 


i2 J5 

. 3 ^3 

> c 2? c 

£2 ■= 2 

co o> <0 o> 

E fc ,£ fe 

£ .E a! '.E 


co *- 

o o 


SP 

c 


CM 

o 


CD -2 


£ £ TO 


o 


I- 1- I t c 


r*. r- 

05 00 

«- CM 


CM 

UD 


CO 

CO 

ID 


CO 

CM 

CM 


ID 

| 

1 

CO 

ID 

1 


7 

in 

1 


7 

O 

O 

O 

O 

0 

0 

0 

O 

0 

*— 

*“ 


T— 

’— 



T— 


X 

X 

X 

X 

X 

X 

X 

X 

X 

CM 

CO 

CM 

<0 

00 

q 

T— 

T— 

CM 


<fr 








CM 

to 


00 

0 

CM 



CO 

*— 

T— 

«— 

in 

CO 

CO 

O 

0 

0 

CM 

CM 

CM 

d 

d 

d 

CM 

CM* 

CM* 


O 

ID 

CO 


G) 

d 


o 

ID 

CO 


CO 

05 

CM 


05 

CM 


LD 

CD 


S S 

CO CO 


CO 

CO 


CO 

5 


CO 

00 


FRACTURE MODE CHARACTERIZATION OF AISI 304 AND 348 STAINLESS STEEL LOW-CYCLE FATIGUE SPECIMENS 


271 


4 * 132 * 

> £ c ~ 

O flC — 


now 

O) CM 00 

«- co id 


S CO O CM 00 CD 

-<t r- CM ID CO 

*3* 03 O «“ «- CM 

ID O ID' 


7 ? 7 

ooo 

x x x 
00 ^ ^ 


ID Tl- 00 ID *3- 00 

III III 

ooo ooo 


03 o O 1- CM CM 
5T iO CD Tf Tf 


T5 .£ S> * 

5 S 

< co cc ^ 




CM t- 

*7 «7 oo fj* id *7 

LD 00 ID ID ID ID 


1*1 
i ^ = 

CT 3 Q. 

T3 ® 0 ) 

c 0) « 

- ! r 
» » 3 
£ 

ra £ 3 
*5 O tj 

g\<2 
2- x ^ 

> » £ 

Jr c <0 

8 g E e 

g .E v .2 
ro Q a 

5 x o & 

_ o. w o> 

.1 s S | 

+5 i= O) P 

s> fe c5 °* 

| 8 I ° 

a "2 S’ o 

| gS s 

S =5 


o o <5 

D C 

o a w o 

in- 


is <0 (0 to 

I 3 3 3 

= § £ £ £ 

to O) V 3 5 V? U ? 5 

C Jr C C C 

| 4 £ £ £ 

£ £ H H H 


3 3 3 3 3 3 

C C C C C C 

<0 CD CO (0 CO <0 

k. L. k. w k. k. 

D) D) CT 0)0)0) 

00 Co t/) W W M 

C C C C C C 

£ £ £ £ £ £ 

i— I— i— Hhh 


® ti v c 

3 g\E o 
2 J i ,2 E 

|| If E 


CM »- t- 

o o o 


cm «- 
o-dd 


<D -X 

•not; 
5 g .2 
2 6 £ 


3 3 3 3 3 3 

c c c c c c 

(0 CO CO co to CO 


0 ) 0 ) 0 ) 
W> k« 

o a> a> 

■M +■* 4-* 

c c c 


8 

E 


OOO OOO CO CO CD 

ID ID LD ID ID ID 
COCOCO COCOCO CD 00 00 



Fig. 6.33 - Electron fractograph showing striations on the fractured 
surface of AISI 304 stainless steel tested at 0.51% strain 
range,4.5x 10 — 3 sec — 1 strain rate, and 430°C. Note rub- 
marks in the top center portion of the fractograph (two- 
stage acetate-carbon replica). Test No. 347U (Neg. 1197B) 







273 


Low-cycle fatigue data were generated for AISI 348, 304, and 316 stainless steel to de¬ 
termine effects of strain range (0. 3 to 2. 2 percent), temperature, (430 , 650 , and 816 C), 
and strain rate, ('4x10-3, 4xl0^ 4 , and4xl0 -5 sec" 1 ). 

In general, the 348 SS exhibited slightly better strain fatigue resistance for a given total 
axial or plastic strain level than 304 or 316 SS; this effect was observed for all three tem¬ 
peratures evaluated. The fatigue resistance of 316 SS for 650°C, however, is less than 
either the 304 or 348 SS. The 348, 304, and 316 SS exhibited cyclic strain-hardening char¬ 
acteristics for the various test conditions evaluated, except for the 316 SS tested at 430°C. 
The 316 SS exhibited cyclic strain hardening followed by cyclic strain softening over the en¬ 
tire strain range evaluated. 

Results for 348 and 316 SS tested at 430°C are in fair agreement with the Coffin-Manson 
relationship that correlates low-cycle fatigue life at temperatures below the creep range. 

Fair agreement was obtained using the Manson-Halford approach for predicting the fatigue 
life of materials tested in the creep range. In general data tended to be more conservative 
than the average prediction had indicated. 

A hundredfold decrease in strain rate caused a severalfold decrease in the fatigue resist¬ 
ance of 304 and 348 SS tested at 650°C and 816°C. Similar results were obtained for 316 SS 
tested at 650°C. At a test temperature of 816°C, however, only slight differences in the fa¬ 
tigue resistance of 316 SS were observed at strain ranges less than 1.0 percent for a change 
in strain rate from 4x 10 “ 4 sec -1 to 4x 10 " 5 sec -1 . This improvement in the fatigue char¬ 
acteristics of 316 SS at the lowest strain rate is presumably due to a metallurgical reaction. 

The shape of the curves shown in plots of total axial or plastic strain versus fatigue life, 

N 5 , for the three materials evaluated does not seem to change as the strain rate is decreased 
several hundredfold but the curve is shifted to positions of lower life for given strain values. 

A comparison of the ratio of Ns/Nf for the overall data showed that this ratio is generally 
decreased from a range of approximately 0.99 to 0.90 representing the 650°C data, to a 
range of approximately 0.7 to 0.85 for the 816°C data. 

An attempt was made to characterize the N 5 failure criterion from the standpoint of crack 
length. Preliminary results of a fractographic analysis for samples tested at 650°C indicate 
that the crack is approximately 1.7 mm long at the N 5 point. 

Results of a metallographic and fractographic analysis of 348 and 304 SS specimens test¬ 
ed at 650°C and 816°C and strain rates of 4x10-3, 4xl0- 4 , and 4xl0 “ 5 sec -1 , showed that 
the mode of crack initiation was primarily intergranular for all conditions tested, except 
for the highest strain rate at 650°C. Under these conditions the 304 SS exhibited a trans- 
granular type of crack initiation at both the highest and lowest strain range evaluated. 

6.7 PLANS AND RECOMMENDATIONS 

The effect of strain rate on fatigue characteristics of AISI 348, 304, and 316 stainless 
steels will be further evaluated. This continuing work will duplicate several test condi¬ 
tions for which only a few data points are presently available. Overall data will be sta¬ 
tistically analyzed when these tests are completed. Several tests will also be conducted 
to determine the strain-rate sensitivity at 430°C of the austenitic stainless steels pres¬ 
ently being evaluated. 

Testing will be initiated to determine the effect of hold times at peak strain for AISI 
348, 304, and 316 stainless steels. These data will be obtained at temperatures of 430°, 

650°, and 816°C. 



274 


The various aspects of biaxial fatigue testing are being investigated. Several techniques 
are being reviewed for conducting biaxial fatigue tests at NMPO. 

Metallographic and fractographic analyses of fatigue specimens will continue. Special 
emphasis is being placed on fracture analysis, although future work will include an in¬ 
vestigation of morphology changes in the various materials being evaluated. 



KM* 


7.^ADVANCED PRESSURE VE_Sj 


( 1521 ) 



F. C. Robertshaw,* H. R. Stephan,t J. E. McConnelee 


f The^bjective of this program is to determine the applicability of high-strength steels 
and alloys to nuclear reactor pressureVessels ^nd similar critical applications. Ma¬ 
terials under investigation include a 12Ni — 5Cr — 3Mo managing steel^pd orecipitation- 
hardening stainless steel, PHKHEMq. The nickel-base alloy, Inconel alloy718, is 
also being investigated together with a quenched and tempered alloy steel, BP“t>Ni — 

4Co - 0. 2C. 

All alloys studied in this program can be processed to far greater tensile and yield 
strength levels than the current alloys which are ASME Boiler and Pressure Vessel Code- 
approved and in current use. If it can be established that these alloys meet strength cri¬ 
teria and other significant requirements for pressure vessel applications, their eventual 
use may achieve one or more of the following: 

1. Greater pressure vessel reliability and safety. 

2. Reduced pressure vessel weight and section size. 

3. Improved reactor performance resulting from higher permissible pressures and/or 
temperatures. | 

The type of materials being studied will necessitate greater unit costs for materials and 
processing, but it is believed that overall reactor economics may disclose an acceptable 
cost balance, particularly in view of such intangibles as greater safety or equal safety with 
less weight. 

The experimental approach being followed encompasses two phases. The first phase con¬ 
sists of selecting and procuring candidate materials in the form of approximately 2. 54-cm- 
thick plate, and evaluating pertinent properties, response to heat treatment, weldability, 
and radiation resistance. The second phase involves procurement of heavier plate (2 10 cm 
thick) to determine size effects on properties, response to heat treatment, and weldability. 
Ultimately one or more materials are expected to emerge as candidates for high-strength, 
pressure vessels. For the most part all alloys currently under study are well into the first 
phase. The 12-5-3 maraging steel is entering the second phase. 

7.1 EXPERIMENTAL PROGRAM 

12Ni - 5Cr - 3Mo MARAGING STEEL 


The 12-5-3 maraging steel is a heat-treatable alloy capable of high strength with excel¬ 
lent toughness. Heat treatment consists of solution annealing, air cooling or quenching to 
form martensite, and aging to effect precipitation strengthening. Because of the low carbon 
level (0. 03C max, Table 7.1) the alloy is relatively soft (approximately Rc 30) and is work- 

‘Project leader. 

* Principal investigator. 


275 




DESCRIPTION OF 12NI 


276 




X 



1 

1 

1 

1 

in 

l 


1 

1 


E 

Z 



1 

1 

o 

cd 

o 


O 

1 


a 






00 


00 


in 



a 

O 



1 

1 

in 

r> 

o 


o 

1 







T— 


CM 






u! 

(0 


15 

"co 

"co 

15 








CO 


CO 

CO 

CO 

00 







o 

1 

If) 

If) 

o 

00 00 

CO 

rf 


O 


CD 

CO 



S 

r». 

CN 

o o 

CO 

CM 

co 

o 


O 

o 




cm 

CO 

cd cd 

cm’ 

cd 

CM 

cd 


cd 

cd 




1 

JD 

u 








V 5 



o 

O 

o 

O) 

r-* 

in 




00 

‘35 


< 

CM 

in 

co ^ 

o 

CM 

o 

CM 


CM 

co 

>~ 

CO 



o 

d 

o o 

o 

o 

d 

o 


o 

d 

c 













< 



1 

JD 

o 








— 



o 

in 

CO 'Ct 

CO 

co 

co 

in 


CD 

< 3 - 

.1 


r— 

f" 

CO 

*— *— 

CO 

CM 


CM 


CM 

CM 

E 



o 

o 

o d 

d 

o 

o 

d 


o 

o 

05 













o 



1 


o 

C"* 



in 


CM 

o 

OJ 


•— 

in 

in 

O) 

CD 

*— 

O) 

r- 


O 


*w. 


z 


CM 

f— | 

r-’ 

CM 

CM 

CM 


CM 

cm’ 

o 












T“ 

TJ 













C 













05 

> 


CJ 

1 

in 

in 

o 

in 

O 

in 

00 


O 



% 

r- 

CM 

9 1 


o 

o 

o 


O 

q 




"T 

in 

in 

CM 

in 

in 

in 


in 







o 











CM 

x 

<D CM 

CM 

o 

CM 

CM 



00 



<75 

*— 

(0 

o o 

O 


o 

o 


o 

o 




o 

2 

o o 

o 

d 

d 

o 


o 

o 




o 


CO 

00 

r** 

'd- 




<3- 



CO 


X 

o o 

o 

o 

o 

o 


O 

0 



o 

CO 

o o 

o 

o 

o 

o 


o 

0 




o 

2 

d o 

o 

d 

o 

o 


o 

d 






o 











o 


<d in 

m 

CM 

CM 

in 


LD 

r^. 



CL 


X 

o o 

o 

O 

o 

o 


O 

0 




o 

CO 

o o 

o 

O 

o 

o 


O 

0 




o 


o d 

o 

o 

o 

o 


o 

d 




o 

X 

u 

CD *- 


CD 


t 


r 

in 



c 

’— 

CO 

o o 

o 

o 

T— 

o 


o 

0 



:> 

d 


o d 

o 

d 

o 

o 


o 

d 






V 













o 











CO 


in 


co 

< 3 - 

CD 


LD 




Cj 

X 

CM <- 

CM 


o 

o 


O 

0 




o 

CO 

o o 

O 

o 

o 

o 


O 





o 

2 

o d 

o 

o 

o 

o 


o 






1 


1 


8 












\ 05 












t- c 





c 



-C 


-C 


-C ^ 





o 



»— 


T— 


T— .E 








L. 



E 

»- In 





•5 

c 



o 

H— 

1 

o 

s 

(0 

o +* 
**- E 

o 

o 

1 

I 


o 



O O 


u o 

■D 

O o 

CJ 




CJ 



o g 
in “ 


o g 
in “ 

TJ 

O 

°-s 

.E 

'co 







00 'co 


00 'to 

o 

00 CM* 

1 








_ 

r-- t- 

_ 

CM 


CM 







o 

in co 

o 

CM 


CM 

05 


Size, cm 



5 

X 

5 

x 

o _ 
a — 

C/5 CO 

X 05 

co E 

S * 

X X 

CD CM 

in in 

CO t- 

X X 

o 

Cl¬ 

io 

X 

CO 

*a 05 

in •* 

X 

q __ 

O to 
CM 

CJ 

X 05 


X 

q 

° 1 

X 05 

05 

i 

to 

05 ^ 

jz in 




1"- 

CM — 
CO 

"d- * 3 - 
in lo 

in o 

cm ’a 


cm ’a 

0 £ 





CM 

d t 

CM CM 

o l 

2 £i 


O r- 

^ d 


CO 












05 

c 

05 

3 



+ m 

+ 2 








*05 

"□ 

05 

O 

O 



2 o 
< > 

_ LU 

2 o 
< > 

AM 

IAIAI 

IVM 


IVM 

IVM 


ol 












o 

V) 




CO 

0) 

o 






z 

\n 

> 


00 

r*v 

"St* 

CO 

CO 

CM 

O) 

CO 


o 

■^r 


co 

to 


co 


o 


LD 


LD 

LD 


X 

C 

CO 


o 

o 

a. 

CD 

> 

O 

CD 


O 

CD 

§ 



15 












c 











o 

£ 











z 

o 











■M 

Z 


*- 

CM 

co 


in 


in 

CD 


_J 












O 

a 

E 

o 


a 

H- 

o 


O • “ 
+-* to 


> -Q 

05 

+J -C 
05 T* 

E 03 
« 5 

TJ ° 
O O 

| i 
0 ) £ 
a; o 
n §" 

E o 

w 

O F- 

c * 

E lo 

3 CM 


£ 

V 

o 


05 

i 

o 


o 

Q. 

05 

-Q 

i 

c/i O 

> . £ 

CO '5; — 

O O .E 

CL _ 

;> "o .2 

2 . 05 05 

Ljj O (D 
CJ_ 


S CO 

< o w _ « 

CO JD o TJ 05 


277 


able as solution annealed; after aging the alloy has a hardness of Rq 40 to 44 and is not 
easily worked. A standard heat treatment consists of 815°C for 1 hour, air cool, plus 480°C 
for 3 hours, air cool. 

The aging treatment for optimum properties may vary between different heats of materi¬ 
al. The alloy can be prepared by air melting techniques, but vacuum melting techniques are 
preferable. Ibis reportedly weldable by either TIG or MIG methods and requires no pre- 
or post-heating. For best properties, heat treatment by aging or solution treatment and 
aging are required after welding. 

Table 7.1 describes the 12-5-3 material procured to date. Lots 1 and 2 were obtained 
first and most data subsequently reported involve these materials. Lots 3 and 4 were ob¬ 
tained more recently, principally to evaluate the 12-5-3 wire composition compared to the 
12-3-3 composition. Lots 5 and 6 will be used to study size effects on properties, response 
to heat treatment, and weldability. Initial inquiries have been made to procure 20-cm-thick 
material. 

Base Metal Studies 

Aging Studies - Table 7.2 and Figure 7.1 show the effect of aging at temperatures of 
425°, 450°, 480°, and 510°C for times up to 25 hours on the hardness of the four heats of 
the 12-5-3 alloy. Specimens were approximately 1.27-cm to 2-cm cubes. Heat No. 01378 
developed the lowest hardness for the three higher aging temperatures; heat No. 1P0193 
developed the highest hardness under every aging condition. This is probably due to a more 
favorable Ti/Al ratio and a higher molybdenum level. Heat No. 60539 and 60540, fabrica¬ 
ted into 10.2-cm by 20.3-cm forgings, show intermediate to high hardness values except 
at the lowest aging temperature. 

These studies indicate that the generally recommended aging treatment of 480°C for 3 
hours results in about the same hardness for all four heats, but longer aging times can 
be used to increase the hardness from 1.5 to 4 Re- Increased aging times may be useful 
when greater strength with somewhat lower toughness is desired. 


TABLE 7.2 

12Ni - 5Cr - 3Mo AGING STUDIES 



Aging 






Aging Time, hr 







Heat 

Temperature, 

1 2 

3 

4 

5 

6 

8 9 10 11 

12 

14 15 

20 

21 

24 

25 

No. 

°C 






Rockwell C Hardness 






01378 

425 

38 

39.5 


40 


42 





42 



450 

39 

40 



40 






42 



480 

38.5 

40 



41 

41.5 






41.5 


510 

39 

39 



39.5 

39.5 






39.5 

IP0193 

425 

38 

39.5 



41.5 

42 

42 

43 


43 




450 

39 

41 



42 

42.5 

43 


43.5 


44 



480 

39 

41 



42 

42.5 

43 


44 





510 

39.5 

40 



41 

42.5 

43 

44 

43.5 


44 


60539 

425 

36 


36.5 



38 






41.5 


450 

37 

40.5 



41.5 

42 






43.5 


480 

37.5 

40 


40.5 


41 


41.5 


42.5 


42.5 


510 

39 

39.5 


40 


39.5 


40 


41 


41 

60540 

425 

36 



37 


38.5 






41.5 


450 

37.5 

41 



41 

42.5 






44.5 


480 

37 

41 


41 


41.5 


42 


43.5 


42.5 


510 

39 

39 


40 


40 


40 


41 


41 



278 




0 5 10 15 20 25 

Aging time, hours 




0 5 10 15 20 25 

Aging time, hours 


Fig. 7.1 — 12Ni — 5Cr — 3Mo aging studies 

The relatively slow rate of aging to peak hardness values is an excellent feature of this 
alloy since it indicates no serious aging problem associated with thick sections or com¬ 
plicated structures. 

Tensile Properties - Tensile data obtained on specimens from the 12-5-3 plate from 
heat No. 01378 at 25°, 315°, and 425°C in orientations transverse and parallel to the roll¬ 
ing direction are summarized in Table 7.3 and Figure 7. 2. Although earlier tests 1 had 
indicated a difference in values due to orientation, additional aging has produced essen¬ 
tially equal and higher strength levels in both directions. 

Toughness - Charpy V-notch impact tests at 25°, -40°, and -78°C on specimens orient¬ 
ed both transverse and parallel to the rolling direction from heat No. 01378 are shown in 
Table 7.4 and Figure 7.2. Additional aging has reduced the transverse values somewhat, 
which has reduced but not eliminated the difference between values for the two rolling di¬ 
rections. 

As a further measure of fracture toughness, specimens are being prepared for drop- 
weight tear testing (DWTT) at the Naval Research Laboratory. 

Fatigue Properties — Hourglass specimens are being prepared for preliminary low- 
cycle fatigue testing. Material from heat 1P0193 is being employed. 

Radiation Effects - Tensile blanks and Charpy V-notch specimens of the 12-5-3 alloy 
and similar specimens of PH13-8Mo and Inconel alloy 718, were irradiated in the MTR 
at pile-ambient temperature to an approximate fluence of 4 to 5xl0 19 n/cm 2 (E n ^ 1 Mev). 
Testing will be conducted at the Naval Research Laboratory. 

1 "Fuels and Materials Development Program Progress Report No. 71," GE-NMPO, GEMP-1002, December 29,1967, p. 109. 





279 


TABLE 7.3 


TENSILE PROPERTIES OF 12Ni - 5Cr - 3Mo PLATE 3 


Specimen 

Orientation 

Test 

Specimen Specimen Temperature, 
No. Type b °C 

0.2% Yield Strength 

Tensile Strength 

Elongation 0 

Reduction 

kg/cm^ 

psi 

kg/cm^ 

psi 

in 4D, % 

of Area, % 

Transverse 

K1 1 

25 

11,330 

161,200 

11,860 

168,700 

22 

72 

to rolling 

K2 1 

25 

11,450 

162,900 

11,950 

169,900 

22 

70 

direction 

K3 1 

315 

9,010 

128,100 

10,030 

142,700 

16 

69 


K4 1 

315 

9,060 

128,900 

10,060 

143,100 

14 

69 


K5 1 

425 

8,100 

115,200 

9,170 

130,500 

9 

73 


K6 1 

425 

8,090 

114,800 

9,140 

129,700 

18 

74 


2E II 

315 

8,700 

124,200 

9,580 

139,200 

14 

66 


3E II 

425 

8,100 

115,300 

9,150 

130,000 

17 

70 


L1 d 1 

25 

12,250 

175,700 

12,600 

180,100 

19 

70 


L2 d 1 

25 

12,200 

175,400 

12,600 

180,000 

19 

70 


L3 e 1 

25 

12,650 

180,900 

13,000 

185,400 

18 

70 


L4 e 1 

25 

12,700 

181,400 

13,100 

186,200 

18 

69 

Parallel 

IB 1 

25 

12,330 

175,400 

12,570 

178,800 

22 

68 

to rolling 

2B 1 

25 

12,200 

173,900 

12,520 

178,000 

22 

67 

direction 

3B 1 

315 

9,630 

136,900 

10,410 

148,100 

17 

41 


4B 1 

315 

9,560 

136,000 

10,440 

148,500 

16 

66 


5B 1 

425 

8,340 

119,500 

9,400 

133,900 

18 

70 


6B 1 

425 

8,020 

114,300 

9,030 

127,900 

14 

72 


1A II 

25 

12,340 

175,200 

12,600 

179,200 

17 

65 


2A II 

315 

9,400 

133,700 

10,400 

148,200 

15 

65 


3A II 

425 

7,950 

113,600 

9,200 

131,100 

17 

69 


1 E d 1 

25 

12,300 

176,320 

12,600 

180,700 

18 

68 


2E d 1 

25 

12,400 

176,570 

12,600 

180,500 

19 

67 


a Plate heat treated at 480°C for 3 hours and air cooled before machining to specimen configuration. 
^All specimens were standard ASTM type: I — 0.636-cm-diameter reduced section. 

II — 1.27‘Cm-diameter reduced section. 

c Elongation in four x diameter (4D). 

^Specimen aged an additional 3 hours at 480°C and air cooled. 
e Specimen aged an additional 6 hours at 480°C and air cooled. 


Thermal Stability - Specimens of the 12-5-3 alloy aged at 480°C for 3 hours were heated 
for 1000 hours at 315°, 370°, 425°, and 480°C. Exposures up to 100 hours resulted in no 
change in hardness except at 480°C where a slight increase was noted. After 1000-hour 
exposures at 315°C and 370°C, hardness increases of 3 to 4 R c were observed. The speci¬ 
mens heated for 1000 hours at 425°C and 480°C tended to drop in hardness. Properly aged, 
the alloy can apparently be made stable to at least 315°C. 

Weldment Studies 

Four weldments of the 12-5-3 alloy were prepared using the 2. 54-cm-thick plate from 
heat 01378. Joint designs were similar to those shown earlier for Inconel alloy 718. 2 Of 
these, two were made using the U-joint and two with the double U-joint preparation. The 
former is utilized in developing weld metal properties and the latter in determining weld 
joint properties. The manual TIG welding method was used. The filler metal was 0. 32-cm- 
diameter spooled bare wire of 12Ni — 3Cr - 3Mo (heat 01434). Additional details of weld¬ 
ment preparation were given in an earlier report. 3 

Weld Zone Hardness versus Heat Treatment — The effect of post-weld heat treatment 
on weld zone hardness was studied on several sections cut from the two U-joint weldments 
of 12-5-3 plate made using 12-3-3 weld wire. 4 The first heat treatment consisted of post- 

^"Sixth Annual Report — High-Temperature Materials Program, Part A," GE-NMPO, GEMP-475A, March 31,1967, Fig. 9.1, p.232. 

3"AEC Fuels and Materials Development Program Progress Report No. 67," GE-NMPO, GEMP-67, June 30, 1967, pp. 141—145. 

4 GEMP-67, Fig. 10.1, p. 144. 




280 


Re-age, hours 

3 6 Id- 

□ B 
O O O 


Parallel to R.D. 
Transverse to R.D. 
Weld joint (W+ A) 



Temperature, °C 
Tensile Strength 



Temperature, °C 
Yield Strength 



0 100 200 300 400 500 

Temperature, °C 

Elongation 



Temperature, °C 


Charpy V-notch Impact Strength 


Fig. 7.2 — Tensile and Charpy V-notch impact properties of 12Ni — 5Cr — 3Mo 
maraging steel plate and weldments. Data points taken from Tables 7.3 
through 7.6. 

weld aging at 480°C for 3 hours. Figure 7.3 illustrates the results of a Rockwell C hard¬ 
ness transverse of weldment sections which received this treatment. Weld center hard¬ 
nesses varied from Rc 30 to 35. Two sections, B and C, were then annealed for 1 hour 
at 815°C (air cooled) and aged at 450°C for 24 hours. Weld center hardnesses after this 
treatment rose to R^ 37 to 38. Figure 7.4 illustrates sections from a weldment annealed 
at 815 C for 1 hour (air cooled) and aged at 480°C for 3 hours. Weld center hardnesses 
ranged from R^ 32 to 35. 

Both base metal and weld metal hardnesses are low after aging at 480°C for 3 hours 
(Figure 7.4). Re-aging such segments at 450°C for 24 hours or at 480°C for an additional 
3 to 6 hours achieved higher centerline weld hardnesses in the range of Rq 36 to 38; base 
metal hardnesses increased to Rq 40 to 42. This is a further indication that aging prac¬ 
tice must be adjusted to effect optimum properties. The disparity in weld metal and parent 
metal hardness is particularly important because it indicates that this 12-3-3.wire compo¬ 
sition will not achieve optimum weld properties. 

Tensile Properties - Joint tensile strength data at 25°C and 315°C of weldments heat 
treated in various ways are summarized in Table 7. 5 and Figure 7.2. Aging of weldments 
at 480°C for 3 hours resulted in strengths lower than anticipated, based on base metal 



281 


TABLE 7.4 


CHARPY V-NOTCH IMPACT STRENGTH OF 12Ni-5Cr-3Mo PLATE 3 


Orientation of Major 
Specimen Axis* 3 


Energy Absorbed 



25°C 

- 

40°C 

- 

78°C 

ft-lb 

m-kg 

ft-lb 

m-kg 

ft-lb 

m-kg 

Transverse to 

104 

14.4 

71 

9.8 

59 

8.2 

rolling direction 

103 

14.3 

66 

9.2 

54 

7.5 


114 

15.8 

68 

9.4 

56 

7.8 


83 c 

11.5 






95C 

14.2 






90 c 

12.4 






84 d 

11.6 






78 d 

11.0 






92 d 

12.6 





Parallel to 

70 

9.7 

49 

6.8 

42 

5.8 

rolling direction 

68 

9.4 

49 

6.8 

41 

5.7 


72 

10.0 

50 

6.9 

39 

5.4 


76 c 

10.5 






64 c 

8.9 






72 c 

10.0 






62 d 

8.6 






67 d 

9.3 






83 d 

11.5 






a Plate heat treated at 480°C for 3 hours and air cooled before 
machining to specimen configuration. 
d AII notches perpendicular to surface plate. 

Specimens aged an additional 3 hours at 480°C. 

^Specimens aged an additional 6 hours at 480°C. 


properties. Re-aging at 450°Cfor 24 hours resulted in marked improvements to levels 
about equivalent to base metal transverse properties reported earlier for the 3-hour age 
at 480°C. Post-weld annealing at 815°C for 1 hour plus aging at 480°C for 3 hours resulted 
in about the same properties noted above for the direct aged joint. Re-aging this weldment 
at 450°C for 24 hours again resulted in about the same improvement noted above for the 
direct aged and re-aged weldment. Both treatments with re-aging resulted in joint proper¬ 
ties about equivalent to the transverse base metal values. Elongation values for most weld 
joint specimens were 12 to 15 percent except for one which was 9 percent. These values 
are lower than those of the base metal. 

Toughness - Charpy V-notch specimens of the above weldments of the 12-5-3 plate made 
using the 12-3-3 weld wire showed the welded, annealed, and aged specimens to have an 
average absorbed energy value of 11.0 m-kg (80 ft-lb) at 25°C, compared to 8. 6 m-kg (62 
ft-lb) for welded and aged specimens. The values obtained at -40°C were 8.4 m-kg (61 ft- 
lb) and 4.7 m-kg (34 ft-lb), respectively. These values were lower than the base metal 
for the direct aged weld but were midway between the transverse and parallel rolling di¬ 
rection values for base metal when the weld was annealed and aged. Table 7.6 and Figure 
7. 2 summarize the Charpy V-notch data. 

PH13-8MO 

The PH13-8MO alloy is a martensitic precipitation-hardening stainless steel which has 
excellent toughness at high-strength levels. It is heat treated by solution annealing, air 
cooling or quenching, and aging. In the solution-treated condition the alloy has a hardness 
of approximately Rq 35 maximum; as-aged, depending on aging temperature, its hard¬ 
ness ranges from about R^ 36 to 48. Normally, the material is supplied from the vendor 
in the solution-treated condition (925°C for 0.5 hour and air cooled or quenched to below 
15°C) and is only aged by the user at 510° to 590°C, depending on the properties required. 




Fig. 7.3 - Sections of weldments of the 12Ni - 5Cr - 3Mo maraging steel made using 12Ni - 3Cr - 3Mo filler metal. 

Sections were maraged after welding at 480°C for 3 hours which resulted in the Rockwell C hardness values 
shown beside the indentations. After maraging, sections B and C were annealed 1 hour at 815°C, air cooled, 
and aged at 450°C for 24 hours. The hardnesses achieved by this treatment are shown by the numbers be¬ 
side the X-marks. (Neg. P67-8-33B) 

















284 


TABLE 7.5 

TENSILE PROPERTIES OF 12NI - 5Cr - 3Mo WELDMENTS 


Test 


Specimen 

Type 9 

Specimen 

No. 

Specimen 

Treatment* 3 

Temperature, 

°C 

0.2% Yield Strength 
kg/cm 2 psi 

Tensile Strength 0 
kg/cm 2 psi 

Elongation 
in 4D, % 

Reduction 
of Area, % 

1 

22E 

W+A 

25 

10,430 

148,800 

11,100 

157,100 

15 

64 

1 

22K 

W+A 

25 

10,400 

148,600 

10,800 

154,800 

14 

67 

1 

22F 

W+A 

315 

8,300 

118,900 

9,130 

132,300 

13 

63 

1 

22L 

W+A 

315 

8,400 

119,700 

9,080 

129,300 

13 

53 

II 

22A 

W+A 

25 

10,900 

155,300 

11,600 

161,300 

11 

61 

II 

22B 

W+A 

315 

8,450 

120,300 

9,000 

127,200 

11 

62 

1 

22G 

W+A+R 

25 

11,280 

159,800 

11,845 

167,900 

14 

59 

1 

22H 

W+A+R 

25 

11,200 

159,200 

11,840 

167,700 

14 

60 

1 

221 

W+A+R 

315 

9,060 

130,100 

10,000 

142,400 

12 

60 

1 

22J 

W+A+R 

315 

9,300 

132,100 

10,000 

142,700 

12 

59 

1 

24A 

W+An+A 

25 

10,500 

149,500 

11,100 

156,500 

15 

65 

1 

24 B 

W+An+A 

25 

10,800 

153,600 

11,250 

158,600 

14 

64 

1 

24C 

W+An+A 

315 

9,030 

127,700 

9,550 

135,100 

13 

63 

1 

24D 

W+An+A 

315 

9,100 

130,000 

9,550 

136,800 

13 

63 

1 

24 E 

W+An+A+R 

25 

11,700 

166,300 

11,900 

169,300 

9 

33 

1 

24F 

W+An+A+R 

315 

9,800 

139,100 

10,250 

145,800 

12 

59 


a AH specimens were standard ASTM type: I - 0.636-cm-diameter reduced section. 

II — 1.27-cm-diameter reduced section. 
Reduced sections contained weld metal, heat-affected zone, and base metal. 

^Heat Treatments: W — Weld. 

A - Aged 3 hours at 480°C and air cooled prior to machining. 

R — Re-aged an additional 24 hours at 450°C. 

An — Annealed 1 hour at 815°C and air cooled to room temperature. 
C AII specimens fractured in the weld. 


TABLE 7.6 

CHARPY V-NOTCH a IMPACT STRENGTH 
OF 12Ni - 5Cr - 30Mo WELDMENTS 


Energy Absorbed 

Condition 25°C -40°C 

of Specimen* 3 ft-lb m-kg ft-lb m-kg 


W + A 

68 

9.4 

35 

4.8 


58 

8.0 

30 

4.1 


60 

8.3 

37 

5.1 

W+An+A 

80 

11.0 

56 

7.8 


81 

11.1 

66 

9.1 


78 

10.8 

61 

8.4 


a Notch oriented perpendicular to surface of 
plate and through weld metal. 

^Heat treatment: W — Welded. 

A — Aged 3 hours at 480°C. 
An - Annealed 1 hour at 
815°C and air cooled 
to room temperature. 


A standard aging temperature for a good combination of strength and toughness is 565°C 
for 5 hours, air cool. The alloy is best prepared by a combination of vacuum induction 
melting plus vacuum consumable arc remelting. It is reportedly weldable by techniques 
used for other precipitation-hardening stainless steels with no pre- or post-heating. For 
best properties, heat treatment by aging or solution treating and aging are required after 
welding. 

Table 7. 7 describes the PH13-8Mo material procured to date. Lots 1 and 3 were ob¬ 
tained first, and the majority of the data reported are for these materials. Lot 2 was 



285 



286 


obtained to provide additional material for testing and to further study aging response 
following welding on somewhat heavier plate. The study of plate 10-cm thick or greater 
has been considered. It is not certain, however, that a post-weld heat treatment can 
be developed which is practical for large pressure vessels, and the cost involved for ex¬ 
perimental quantities in these thicknesses is high. Accordingly a more certain qualifica¬ 
tion will be made in the 2. 54-cm thickness before proceeding to heavier plate. 

Base Metal Studies 

Aging Studies - Specimens from the 1. 27-cm-thick plate (925°C solution-treated con¬ 
dition) were aged in air for various times up to 25 hours at 510°, 540°, 565°, and 590°C. 
Hardnesses ranging from R c 46 to 48 were attained within 1 to 2 hours at 510°C and 540°C; 
after 24 hours the hardnesses dropped slightly to about Rq 45. Hardnesses of the order 
of Rq 42 to 43 were attained in 1 hour at 565°C and 590°C; these dropped to Rq 40 to 42 
in about 4 hours, and to about Rq 36 to 39 after 24 hours. Data are given in Table 7.8 
and aging curves are shown in Figure 7.5. 

The alloy apparently hardens extremely rapidly to a peak hardness level and then over - 
ages to a lower hardness at longer times. This is particularly apparent at aging tempera¬ 
tures of 565°C and 590°C which, incidentally, produce the highest toughness levels. Aging 
temperatures above about 540°C can apparently present significant problems in large com¬ 
plicated structures; high- and low-strength zones might develop during aging, depending 
on section thickness and/or design configuration. 

Tensile Properties - Tensile data obtained on 1.27-cm PH13-8Mo plate were determined 
at 25°, 315°, and 425°C, and are presented in Table 7.9 and Figure 7.6. The figure shows 
that PH13-8Mo possesses a tensile strength greater than 11, 500 kg/cm 2 (156, 000 psi) up 
to about 300°C, combined with a yield strength of nearly 11, 000 kg/cm 2 . These values are 
the lower for the two orientations tested,i. e.,parallel to the rolling direction. The tensile 
and yield strengths of the specimens whose principal axis was transverse to the direction 
of rolling were higher by as much as 15 percent. Tensile elongation was approximately the 
same in both directions. 

Toughness - Charpy V-notch impact tests were made on 1.27-cm-thick PH13-8Mo plate 
at 25°, -40°, and -78°C. Energy absorption values at 25°C were above 9 m-kg for both ori¬ 
entations; the transverse direction showed values above 10 m-kg. At lower temperatures, 
however, Charpy V-notch values for the transverse orientation fell off quickly to about 2.7 
m-kg at -78°C, and those for the parallel orientation dropped to 4. 3 m-kg. Table 7.10 and 
Figure 7.6 summarize impact data obtained on PH13-8Mo 1.27-cm-thick plate. 

To further evaluate the fracture toughness of the PH13-8MO alloy, specimens for drop- 
weight tear testing are being prepared of the 2. 54-cm plate. Testing will take place at the 
Naval Research Laboratory. 

Fatigue Properties - Hourglass specimens of material from the 2. 54-cm-thick plate are 
being prepared for low-cycle fatigue testing. 

Radiation Effects (NRL) — Tensile blanks and Charpy V-notch specimens have been irra¬ 
diated in the MTR at pile-ambient temperature. The fluence level to which they were ex- 
posed is approximately 4 to 5x 10 19 n/cm 2 (E n > 1 Mev). Post-irradiation testing will be 
conducted at the Naval Research Laboratory. 

Thermal Stability - The effects on hardness and microstructure of the aged alloy after 
longtime heating at various temperatures are being determined. No change in hardness or 
microstructure was noted after exposure at 315°, 370°, 425°, 480', and 540°C for 100 hours. 
Moreover, no significant hardness changes were noted after heating for 1000 hours at 315°, 
370°, and 425°C. Heating for 1000 hours at 480°C and 540°C, however, resulted in softening 



287 


TABLE 7.8 

PH13 - 8Mo AGING STUDIES 3 

Aging _ Aging Time, hr 


Specimen Temperature, 0 0.25 0.5 1 _ 2 3 4 5 6 7 8 10 11 12 14 16 19 20 24 25 

No. °C _ Rockwell C Hardness _ 


H-1 

510 

35 


46.5 

46 

46 


46 

44.5 


44.5 

H-2 

540 

35.5 

47 

47.5 


47 

46 

46 


45.5 

45.5 

H-4 

565 


43 

42.5 

42 


40 

39 

37.5 

37 

36.5 

H-6 

565 

35 45 

45.5 46 

44 


44 


40 



39 

H-5 

590 


42 

42.5 

40 


39 

37.5 37 


36 

35.5 

H-3 

590 




40 








a Heat No. VC5281, 1.27-cm-thick plate. 



0 5 10 15 20 25 

Aging time, hours 


Fig. 7.5 — Aging of PH 13 — 8Mo precipitation hardening stainless steel, 
heat No. VC 5281 


TABLE 7.9 

TENSILE PROPERTIES OF PH13 - 8Mo PLATE 3 


Test 


Specimen 

Specimen Temperature, 

0.2% Yield Strength 

Tensile 

Strength 

Elongation 

Reduction 

Orientation 

No. 

°c 

kg/cm 2 

psi 

kg/cm 2 

psi 

in 2.54 cm,% 

of Area, % 

Parallel to 

N1 

25 

13,200 

188,900 

13,710 

195,000 

19.4 

69.8 

rolling direction 

01 

25 

13,210 

187,900 

13,660 

194,200 

21.2 

71.7 


PI 

315 

11,130 

158,300 

11,590 

164,800 

15.8 

69.7 


01 

315 

11,070 

157,500 

11,470 

163,100 

14.3 

69.1 


R1 

315 

11,065 

157,400 

11,870 

163,800 

16.7 

69.1 


SI 

425 

9,740 

138,500 

10,220 

145,300 

22.4 

74.9 

T ransverse to 

Ba 

25 

14,360 

204,200 

15,790 

224,600 

16.8 

64.9 

rolling direction 

BB 

25 

14,510 

206,300 

15,440 

219,600 

18.7 

65.8 


BC 

315 

12,100 

172,100 

13,100 

186,300 

15.0 

64.9 


BD 

315 

10,890 

154,900 

12,950 

184,200 

14.2 

65.0 


BE 

425 

10,320 

146,800 

11,250 

160,000 

21.1 

72.8 


BF 

425 

10,350 

147,200 

11,130 

158,300 

22.0 

73.1 


a Portions of the PH 13 - 8Mo plate were heat treated at 565°C for 4 hours and air cooled prior to machining 
to specimen configuration. 

^All specimens were standard ASTM type with 0.636-cm-diameter reduced section. 





Elongation, percent 


288 


□ Parallel to R.D. 

O Transverse to R. D. 



Temperature, °C 

Tensile strength 



Elongation 



Yield strength 


12 

10 e 

8 1 

l 

6 I 

A C 
H LU 

2 

-100 0 100 
Temperature, °C 

Charpy V-notch impact strength 



Fig. 7.6 — Tensile and Charpy V-notch impact properties of PH13-8Mo 
precipitation hardening stainless steel plate and weldments 


TABLE 7.10 


CHARPY V-NOTCH IMPACT STRENGTH OF PH13 - 8Mo PLATE 3 


Major Specimen Axis b 



Energy Absorbed 



25°C 

ft-lb m-kg 

—40°C 
ft-lb m-kg 

—78°C 
ft-lb m-kg 

Transverse to rolling direction 

75.5 

10.4 

21.5 

3.0 

10.0 

1.4 


78.5 

10.8 

23.5 

3.2 

24.0 

3.3 


80.0 

11.0 

24.0 

3.3 

19.5 

2.7 

Parallel to rolling direction 

69.0 

9.5 

28.5 

3.9 

31.0 

4.3 


69.0 

9.5 

36.4 

5.0 

31.0 

4.3 


74.0 

10.2 

36.0 

5.0 

36.0 

4.3 


a Portions of the PH 13 — 8Mo plate were heat treated at 565°C for 4 hours and 
air cooled prior to machining to specimen configuration. 

^All notches perpendicular to the surface of the plate. 


~0.2% yield strength, kg/cm2 





289 


from an original hardness of R c 43 to R c 35 to 36. For stable operation of the heat-treated 
alloy, the maximum use temperature is apparently less than 480°C. 

Weldment Studies 

Four weldments were prepared from the 1.27-cm-thick plate of PH13-8MO, using the 
0.155-cm-diameter spooled wire of normally matching composition. The manual TIG weld¬ 
ing process was employed in conjunction with both single and double bevel weld preparations. 
Welding procedures were reported previously. 5 

Weld Zone Hardness Versus Heat Treatment - Figure 7.7 illustrates weld zone hardness 
values of two weldment sections, one after aging only (specimen W11A) and the other after 
annealing plus aging (specimen W11AA). The weldments shown are for the 1.27-cm-thick 
plate with a double V-joint weld. The improved weld hardness pattern achieved by anneal¬ 
ing and aging after welding is apparent. More pertinent weld data are anticipated from weld¬ 
ments in the 2. 54-cm-thick plate. 

Tensile Properties - Weld-joint tensile properties of the PHI 3-8Mo weldments aged after 
welding were determined at 25°, 315°, and 425°C, and are presented in Table 7.11 and Fig¬ 
ure 7.6. Tensile strength is intermediate between the transverse and parallel orientation 



Fig. 7.7 — Macrqetched cross sections from double-V PH13-8Mo weldments. W11A aged after welding and W11AA 
annealed and aged after welding. Numbers beside the indentations are Rockwell C hardness values 
(P67-8-33N) 

values of the base metal; yield strength varies from about 93 percent at room temperature 
to 96 percent at 415°C of the base metal, parallel orientation values. Ductility of the welds 
is significantly lower and more variable than observed in base metal tests. Effects of other 
post-weld heat treatments are being studied. 

All-weld specimen data in two post-weld heat-treated conditions are also presented in 
Table 7.11. The properties are generally quite good, although post-weld annealing and 
aging apparently produce higher yield strength values than aging alone. 

Toughness - Charpy V-notch impact tests were completed on specimens machined from 
weldments aged only after welding. Table 7.12 and Figure 7.6 summarize the absorbed 
energy values obtained at 25°, -40°, and -78°C. The values decrease approximately lin¬ 
early from 25°C to -78°C, and are about the same as those reported earlier for base metal 
transverse specimens at -40°C, but much lower at 25°C and -78 C. Effects of other post¬ 
weld heat treatments are being studied. 

INCONEL ALLOY 718 

Inconel alloy 718 is a precipitation-hardening nickel-base alloy which has a nominal com¬ 
position of 53Ni, 19Cr, 19Fe, 5Nb, 3Mo, ITi, 0.5A1, 0.05C, and 0. 004B (wt%). It is 
strengthened by a coherent Nb-rich gamma-prime precipitate believed to be Ni3 (Nb, Ti 

5 GEMP-67, p. 146. 





290 


TABLE 7.11 


TENSILE PROPERTIES OF PH13-8Mo WELDMENTS 


Specimen Type 3 

Specimen 

No. 

Specimen 

Treatment^ 

Test 

Temperature, 

°C 

0.2% Yield 
Strength 
kg/cm 2 psi 

Tensile Strength 
kg/cm 2 psi 

Elongation 
in 2.54 cm, 

% 

Reduction 

of Area 

% 

Location 

of 

Fracture 

Joint tensile 

C 

W + A 

25 

12,160 

172,800 

14,120 

201,300 

14.0 

68 

Base metal 

Joint tensile 

E 

W + A 

25 

12,390 

176,000 

14,280 

203,000 

8.2 

22 

Weld 

Joint tensile 

D 

W + A 

315 

10,410 

148,300 

12,130 

172,100 

12.6 

52 

Weld 

Joint tensile 

F 

W + A 

315 

10,540 

150,000 

12,110 

172,000 

9.2 

38 

Weld 

Joint tensile 

G 

W + A 

425 

9,330 

133,300 

11,120 

158,100 

10.8 

35 

Weld 

Joint tpnsile 

H 

W + A 

425 

9,400 

133,700 

11,030 

157,400 

9.3 

34 

Weld 

All weld tensile 

D 

W + A 

25 

11,620 

165,000 

14,300 

203,200 

12.9 

27 

_ 

All weld tensile 

C 

W + A 

25 

11,600 

164,600 

14,300 

203,200 

20.8 

58 

_ 

All weld tensile 

4 

W+An+A 

25 

13,030 

186,400 

14,520 

206,500 

7.8 

20 

_ 

All weld tensile 

3 

W+An+A 

25 

12,500 

176,800 

14,400 

207,400 

11.0 

21 

- 


a AII specimens were standard ASTM type with.0.636-cm diameter reduced section. 
t>Heat treatments: W — Welded. 


A — Aged 4 hours at 565°C, air cooled to room temperature. 
An — Annealed at 925°C for 0.5 hour, air cooled to 15°C. 


TABLE 7.12 

CHARPY V-NOTCHa IMPACT STRENGTH 
OF PH13-8Mo WELDMENTS 


Condition 
of Specimen 13 



Enery Absorbed 



25°C 

—40°C 

—78°C 

ft—lb 

m-kg 

ft—lb 

m-kg 

ft—lb 

m-kg 

W + A 

44 

6.1 

14 

1.94 

8 

1.08 

W + A 

46 

6.4 

26 

3.6 

7 

0.97 

W + A 

41 

5.7 

22 

3.0 

8 

1.08 


a Notch oriented perpendicular to surface of plate and through weld metal. 
^Heat treatment: W — Welded. 

A — Aged for 4 hours at 565°C, air cooled to room 
temperature. 


Al). The alloy has better strength at higher temperatures than any other compositions be¬ 
ing studied, retaining approximately 90 percent of its room-temperature strength above 
480°C. This high-temperature strength was the principal reason for selection, since it 
offered an opportunity for temperature growth in advanced reactor systems. Several heat 
treatments are employed, depending on the properties required. A typical heat treatment 
consists of solution treating at 950°C for 1 hour, air cool followed by aging at 715°C for 8 
hours, furnace cool to 620°C, hold 12 hours, and air cool. The alloy can be prepared by 
air melting techniques, but vacuum melting techniques are preferred. For some applica¬ 
tions, double vacuum melting is employed. It is reportedly quite weldable by inert arc 
techniques, its sluggish response to age hardening favors good weldability, and it requires 
no pre- or post-heating. For best properties, heat treatment by aging or solution treating 
and aging are required after welding. 

Table 7.13 describes the Inconel alloy 718 material procured for test. Preliminary in¬ 
quiries were made for 10-cm-thick material, but further procurement action has been de¬ 
layed pending additional qualification of the alloy, particularly with respect to fracture 
toughness. 

Base Metal Studies 

Tensile Properties - The tensile properties of the Inconel alloy 718 plate were deter¬ 
mined at 25°, 425°, 540°, and 650°C using three different heat treatments. These data are 
shown in Tables 7.14, 15, and 16 and in Figure 7.8. Heat treatment I resulted in the low- 


291 




| 

to 


o 

z 

2 



1 

Bal 

18.30 

18.95 



o 




1 

1 


m 

S 

■o 

c 

§s § 8 



2 

6 2 do 



1 



3 

z 



o 

*D 

8 2 3 8 



1 

d 5 d d 



o 




Q_ 

1 


5 

z 

1 

O S 8 8 



> 

r-‘ 2 oo 



8 

5 

z 

5.20 


-Q 

Z 

o 

*o 

IS .£ g 2 ° IS 



1 

■c 2 iri 2 id ir 



o 




a. 

2 

8 


Mo 

Z 

CO 



■§ 

CO .E CO 2 O £ 




cn 2 to 2 cd co 



1 

8 


_ 

z 

o 



0 

T> 

c 

<D 

0.80 

Max 

0.54 

0.44 

% 


> 


.2 

> 

< 


2 

2 

Z 

680 

1 

I 


O 

I 

0.65 

Min 

1.40 

Max 

1.01 

0.84 

O 


o 




1 

l 


Z 

o 

■o 

CO to 

O c O X CN O) 



> 

Q = Ifl | CO CN 

55 2 in 2 in in 



| 

1 



z 



6 

o 

O o R W 



I 

£ J si 2 “ * 



o 

o 



I 

z 

d 



o 

1 

$ & 2 « 



5 

d 2 do 



2 

2 

Z 

8 

d 



o 

"O 

5 2 8 8 



a> 

> 

o' 2 o o 



2 

8 



| 



0. 

1 

5 2 o 5 




o 2 o o 



2 



c 

2 

z 

1 


2 

o 

■D 

2 * ° 2 

^ to I- CO 



> 

o 2 do 



2 




2 

8 


O 

Z 

o’ 



O 

■O 

c 

5 

0.08 

Max 

0.02 

0.04 

I 

c 

o 

J | 

QJ 

8 

1 

o t 3 

Z ■O 

DC 

i 

3 

11 


s 

! «! 


1 

*g .2 E 

® Jr 

X cm * 



3 St 

I 

Procedure 3 

1 1 

> 


o 

Z 

i g . 

p 9 ® 


X 

•2 5 

Z o <5 


L 




292 


TABLE 7.14 

TENSILE PROPERTIES OF INCONEL ALLOY 718 PLATE AFTER HE AT TREATMENT l a 
Test 


Specimen Specimen Temperature, 0.2% Yield Strength Tensile Strength Elongation Reduction 
Orientation No. b _°C_kg/cm 2 psi kg/cm 2 psi in 4D, % of Area, % 


Parallel to 

Y1 c 

25 

10,000 

140,200 

13,960 

197,000 

15.9 

23 

rolling direction 

Y2 C 

425 

9,150 

129,900 

12,900 

183,200 

19.9 

33 


Y3 C 

540 

8,800 

125,200 

12,700 

180,200 

17.4 

35 


D 

25 

9,650 

137,000 

13,700 

193,800 

17.8 

22 


H 

25 

9,650 

137,000 

13,600 

192,200 

13.3 

18 


B 

425 

8,300 

118,000 

12,570 

178,700 

21.6 

32 


1 

425 

8,700 

124,000 

12,600 

179,800 

12.9 

15 


C 

540 

8,350 

119,000 

12,400 

176,500 

16.3 

33 


J 

540 

8,900 

119,900 

12,450 

177,100 

12.9 

21 


E 

650 

8,150 

116,000 

10,400 

148,000 

21.0 

39 


K 

650 

8,100 

115,800 

10,500 

149,000 

18.9 

50 

Transverse to 

Cl 

25 

9,900 

140,800 

13,300 

189,100 

13.7 

15.3 

rolling direction 

C2 

25 

9,940 

141,300 

13,270 

188,700 

11.9 

14.8 


C3 

425 

8,920 

126,900 

12,370 

175,900 

17.6 

19.1 


C4 

425 

8,940 

127,100 

12,360 

175,800 

19.1 

20.3 


B1 

540 

8,630 

122,700 

12,300 

174,900 

20.8 

31.6 


B3 

540 

8,590 

122,200 

12,310 

175,000 

20.5 

31.6 


C5 

650 

8,290 

117,900 

10,570 

150,300 

16.5 

21.6 


C6 

650 

8,300 

118,000 

10,630 

151,200 

23.9 

30.7 


a Heat treatment: 950°C for 1 hr, furnace cool 100 min to 715°C, hold 8 hr, furnace cool to 620°C, 
hold for 12 hr {including furnace cooling time), air cool to room temperature. 
Standard ASTM type with 0.636-cm-diameter reduced section. 
c 1.27-cfn-diameter reduced section. 


est strength levels. Heat treatment n, which is similar to I except for more rapid cooling 
from the solutioning temperature, results in the highest strength levels. Heat treatment 
III, involving a much higher solutioning temperature and a modified aging cycle, resulted 
in intermediate strength values. In general, heat treatment in, which was selected for im¬ 
proved toughness, resulted in better elongation and reduction of area values than the other 
two heat treatments. It is clear from heat treatments I and II that the cooling rate from 

TABLE 7.15 

TENSILE PROPERTIES OF INCONEL ALLOY 718 PLATE AFTER HEAT TREATMENT ll a 
Test 


Specimen Specimen Temperature, 0.2% Yield Strength Tensile Strength Elongation Reduction 
Orientation No. b _°C_kg/cm 2 psi kg/cm 2 psi in4D,% of Area, % 


Parallel to 

PI 

25 

11,600 

165,000 

14,700 

209,500 

21 

24 

rolling direction 

AA 

25 

12,150 

172,400 

14,800 

211,100 

21 

33 


Q1 

425 

10,400 

147,900 

13,120 

186,160 

21 

35 


R1 

425 

10,300 

145,580 

13,150 

186,500 

15 

37 


SI 

540 

10,090 

143,500 

12,810 

182,270 

24 

26 


T1 

540 

10,330 

146,680 

12,750 

181,330 

17 

34 


U1 

650 

9,950 

141,610 

11,830 

168,220 

14 

32 


VI 

650 

- 

c 

11,970 

170,120 

21 

40 

Transverse to 

FA 

25 

12,970 

170,000 

14,400 

204,700 

19 

23 

rolling direction 

FD 

25 

12,970 

170,000 

14,350 

204,200 

18 

23 


FB 

425 

10,760 

152,830 

12,810 

182,210 

17 

29 


FE 

425 

10,790 

153,300 

12,690 

180,490 

18 

29 


FC 

540 

10,290 

146,340 

12,490 

177,670 

14 

24 


FF 

540 

10,630 

150,980 

12,410 

176,440 

12 

23 


FH 

650 

c 

c 

11,290 

160,600 

14 

26 


FG 

650 

9,310 

132,570 

11,030 

156,980 

6 

32 


a Heat treatment: 950°C for 1 hr, air cool to room temperature, heat to 715°C, hold 8 hr, furnace cool 
to 620°C, hold for 12 hr (including furnace cool time), air cool to room temperature. 
b AII specimens were standard ASTM type with 0.636-cm-diameter reduced section. 
c Extensometer slipped. 


rength, 10 3 


293 


TABLE 7.16 


TENSILE PROPERTIES OF INCONEL ALLOY 718 PLATE AFTER HEAT TREATMENT III 3 


Specimen 

Orientation 

Test 

Specimen Temperature, 
No. b °C 

0.2% Yield Strength 
kg/cm^ psi 

Tensile Strength 
kg/cm ^ psi 

Elongation Reduction 
in 4D, % of Area, % 

Parallel to 

M 

25 

10,020 

142,490 

13,590 

193,290 

24 

31 

rolling direction 

N 

425 

8,830 

125,620 

12,425 

176,950 

24 

37 


0 

540 

8,580 

122,300 

12,180 

173,140 

23 

37 

T ransverse to 

A 

25 

11,300 

160,740 

13,870 

197,300 

18 

23 

rolling direction 

D 

25 

11,350 

161,600 

13,930 

198,300 

20 

26 


B 

425 

- 

c 

12,510 

177,930 

23 

33 


E 

425 

10,140 

143,990 

12,510 

177,900 

22 

33 


C 

540 

9,780 

139,060 

12,310 

175,000 

20 

33 


F 

540 

9,550 

135,800 

12,270 

174,340 

22 

33 


a Heat treatment: 1070°C for 1 hr, air cool to room temperature, heat to 760°C, hold for 10 hr, furnace 
cool to 650°C, hold for 10 hr (including furnace cool time), air cool to room temperature. 

^All specimens were standard ASTM type with 0.636-cm-diameter reduced section. 
c Extensometer slipped. 


Tensile Strength 



Elongation 



Temperature, °C 


Yield Strength 




Temperature, °C 


Fig. 7.8 — Tensile and Charpy V-notch impact properties of Inconel alloy 718 plate and weldments 


~0.2% yield strength, kg/cm 2 




294 


the annealing temperature has a major effect on tensile properties. This suggests that the 
quenching procedures employed in the solution annealing of the type of heavier plate po¬ 
tentially useful in nuclear pressure vessel construction must be carefully selected to 
assure optimum properties. 

Toughness - Charpy V-notch impact tests were completed for the parallel and transverse 
orientations at 25°, -40°, and -78°C for heat treatments I and II. Table 7.17 and Figure 7.8 
show these data. No evidence of a transition from ductile-to-brittle fracture was noted and 
the cooling rate apparently had little effect on the absorbed energy levels, but a distinct 
toughness anisotropy exists. Charpy V-notch specimens which received heat treatment HI 
will be evaluated. Drop-weight tear test specimens for each of the three heat treatment are 
being prepared and will be tested at NRL to further define the toughness of this fee alloy. 

Fatigue Properties - Specimens are being prepared for preliminary low-cycle fatigue test¬ 
ing. These specimens are being made from material which has received heat treatment m. 

Radiation Effects - Charpy V-notch data have been obtained at NRL on Inconel alloy 718 
specimens irradiated to a fluence of 1.15 x 10 19 n/cm 2 (E n > 1 Mev). Specimens were ir¬ 
radiated in the Union Carbide Research Reactor in a helium environment at less than 120°C. 
The specimens received heat treatment I prior to machining for irradiation. Table 7.18 and 
Figure 7.9 summarize data obtained. The irradiation exposure apparently did not reduce 
impact resistance, at least when similar test temperatures were employed. Additional con¬ 
trol tests will be made at higher temperatures. 

A second capsule containing tensile blanks and Charpy V-notch specimens of Inconel alloy 
718 had been irradiated in the MTR at pile-ambient temperature to a fluence level of about 
4 to 5 x 10 19 n/cm 2 (E n > 1 Mev). Post-testing will be conducted at NRL. 

Thermal Stability - Specimens of Inconel alloy 718 representing all three heat treatments 
have been exposed in air at 540°C for 1000 hours. No change in hardness was noted. Another 
plate which had received heat treatment II was exposed for 1000 hours at 650°C and a s ma ll 


TABLE 7.17 

CHARPY V-NOTCH IMPACT STRENGTH OF INCONEL ALLOY 718 PLATE 


Orientation of 
Major Specimen Axis 

Heat 

Treatment 3 



Energy Absorbed 



25°C 

—40°C 

—78°C 

ft—lb 

m-kg 

ft—lb 

m-kg 

ft—lb 

m-kg 

Parallel to 

1 

20 

2.76 

20 

2.76 

20 

2.76 

rolling direction 


20 

2.76 

20 

2.76 

18 

2.50 



20 

2.76 

19 

2.64 

19 

2.64 

Parallel to 

II 

20 

2.76 

18 

2.50 

17 

2.35 

rolling direction 


20 

2.76 

17 

2.35 

17 

2.35 



18 

2.50 

18 

2.50 

17 

2.35 

Transverse to 

1 

11.0 

1.52 

12 

1.66 

11.5 

1.59 

rolling direction 


12.0 

1.66 

12 

1.66 

9.0 

1.24 



12.0 

1.66 

12 

1.66 

11.0 

1.52 

Transverse to 

II 

10 

1.38 

9 

1.24 

_ 


rolling direction 


9 

1.24 

8 

1.08 

— 

— 



- 

- 

11 

1.52 

- 

- 


a Heat treatment: I - 950°C for 1 hr, furnace cool 100 min. to 715°C, hold 8 hr, furnace 
cool to 620°C, hold 12 hr (including furnace cooling time), then air 
cool tb room temperature. 

II — 950°C for 1 hr, air cool to room temperature, heat to 715°C, hold 
8 hr, furnace cool to 620°C, hold 12 hr (including furnace cool time), 
then air cool to room temperature. 



Orientation to R.D. 

Unirradiated, 

NMPO 

Irradiated, 

NRL 

Parallel 

□ 

s 

Transverse 

o 



Temperature, °C 



-200 -100 0 100 200 300 400 500 600 700 

Temperature, °F 


Fig. 7.9 — Effect of nuclear irradiation (<122°C 1.15 x 10^ n/cm^; 
E n > 1 Mev) on Incone! alloy 718 


TABLE 7.18 


INCONEL 718 CHARPY V-NOTCH DATA FOR IRRADIATED SPECIMENS 3 


Specimen 

No. 

Charpy V Properties 0 
Temperature Energy 

Orientation^ °F °C ft-lb mg-kg 

Rockwell 

Hardness, 

Rc 

T9 

T 

+80 

27 

13 

1.8 


E3 

P 

+300 

150 

23 

3.2 


T1 

T 

+500 

260 

16 

2.2 

40 d 

T15 

T 

+400 

205 

15 

2.1 


T7 

T 

+200 

93 

20 

2.8 


G3 

P 

+200 

93 

23 

3.2 


D1 

P 

+100 

38 

23 

3.2 


J3 

P 

+300 

150 

22 

3.0 


C3 

P 

±0 

-18 

21 

2.9 


F3 

P 

+200 

93 

22 

3.0 


K3 

P 

+100 

38 

21 

2.9 


13 

P 

-160 

-107 

18 

2.5 


K1 

P 

+400 

205 

22 

3.0 

40 d 

T5 

T 

+80 

27 

12 

1.7 


D3 

P 

+0 

-18 

21 

2.9 


T13 

T 

+300 

150 

14 

1.9 


El 

P 

-160 

-107 

18 

* 2.5 


HI 

P 

-160 

-107 

19 

2.6 


G1 

P 

+500 

260 

23- 

3.2 


LI 

P 

+400 

205 

23 

3.2 


H3 

P 

+500 

260 

26 

3.6 



a 1.15 x 10^ n/cm^, E n > 1 Mev (a 68 mb Mn^ fission). 
b T = Major specimens axis transverse to rolling direction. 

P = Major specimen axis parallel to rolling direction. 
C Material given Heat Treatment I prior to radiation exposure. 
^ Average of 3 readings. 





296 


increase in hardness was noted. Tensile and Charpy V-notch impact specimens are being 
prepared from each plate. 

Weldment Studies 

Four weldments of the Inconel 718 alloy were prepared in 2.54-cm-thick plate, using 
joint designs described previously. 6 Of these, two were made using the U-joint for all-weld 
metal specimens, and two were made with the double U-joint preparation for determining 
joint properties. The manual TIG welding method was employed according to procedures 
described earlier. 6 

Weld Zone Hardness Versus Heat Treatment - Sections of double U weldments with two 
different post-weld heat treatments were ground, etched, and surveyed for Rockwell C 
hardness. Figure 7.10 illustrates two sections from a weldment which, after welding, were 
aged at 715°C for 8 hours, furnace cooled to 620°C and held for 12 hours including furnace 
cooling time, and air cooled. Prior to welding the plate was heated at 950°C for 1 hour and 
air cooled. Weld center hardnesses ranged from Rq 34 to 44; base metal hardnesses ranged 
from Rc 38 to 46. The reason for the disparity in hardness between the two weld areas is 
not known but must tentatively be assigned to weld metal heterogeneity. Differences in base 
metal hardness are probably associated with inadvertent variations in pre-weld heat treat¬ 
ment. 

Figure 7.11 presents the results of hardness surveys on two sections of a weldment sub¬ 
jected, following welding, to a solution anneal at 950°C for 1 hour followed by furnace cool¬ 
ing to 715°C, holding 8 hours, furnace cooling to 620°C, holding for 12 hours, and air cooling. 
Weld center hardnesses range from Rq 35 to 41, with base metal hardnesses of Rq 41 to43. 

Tensile Properties — Data obtained on weld-joint tensile specimens machined from dou¬ 
ble U weldments are presented in Table 7.19 and Figure 7.8. These specimens possess 
significantly lower strength and ductility than the base metal similarly treated. The dele¬ 
terious effect on properties of slow cooling from the solution annealing temperature is 
again apparent. The W2 series specimens were cut from plate air cooled from 950°C 
prior to welding; hence they should be compared with base metal specimens given heat 
treatment n listed in Table 7.15. The effect of post-weld solution annealing and air 
cooling, and aging on weldment properties will be determined. All-weld specimens in 
various conditions of heat treatment are being prepared. 

Toughness - Table 7. 20 and Figure 7.8 contain results of Charpy V-notch tests con¬ 
ducted on specimens machined from weldments. The specimens aged only after welding 
(solution annealed at 950°C for 1 hour and air cooled prior to welding) possessed approx¬ 
imately the same impact strength as the base metal at 25°C but only half as much at -40°C. 
Even lower values at -40°C were obtained on specimens from weldments given a solution 
anneal (furnace cooled) and aged after welding. 

HP 9Ni — 4Co — 0.20C 

HP 9Ni — 4Co —0.20C, a heat-treatable steel, has been added to the program and will 
be evaluated. This alloy can achieve high tensile strengths while retaining excellent tough¬ 
ness. It is recommended for use in the quenched and double-tempered condition. The ma¬ 
terial is reported to be quite weldable in the quenched and tempered condition, and no post¬ 
weld heat treatment is required to achieve virtually 100 percent weld efficiencies. 

A plate 2.54 cm thick by 91. 4 by 122 cm has been received from heat No. 3932354* and 
is being machined into test pieces for tensile, impact, low-cycle fatigue, DWTT, and weld 

‘Composition in wt %of heat 3932354: C -0.017, Mn -0.29, P - 0.004, S - 0.007, Si -0.01, Ni -9.11, Cr - 0.79, 

V — 0.09, Co — 4.40, Fe — balance. 


'GEMP-475A, Fig. 9.1, p. 232. 



the indentations are Rockwell C values. (Neg. P67-8-33L) 












299 


TABLE 7.19 


TENSILE PROPERTIES OF INCONEL ALLOY 718 WELDMENTS 


Specimen 

Specimen 

No. 

Specimen 

Treatment 8 

Test 

Temperature, 

°C 

0.2% Yield Strength 
kg/cm 2 psi 

Tensile Strength 
kg/cm 2 psi 

Elongation 
in 4D, % 

Reduction 
of Area, % 

Location 

of 

Fracture 

Joint 

W3A 

W+An+A 

25 

9,700 

138,300 

10,900 

155,200 

12.7 

11 

W 

tensile 

W3A 

W+An+A 

25 

9,300 

132,900 

10,300 

147,000 

12.3 

13 

W 


W3C 

W+An+A 

425 

8,100 

115,400 

10,300 

146,900 

7.3 

22 

w 


W3D 

W+An+A 

425 

8,900 

119,800 

10,000 

143,200 

4.9 

13 

HAZb 


W3E 

W+An+A 

540 

8,300 

118,200 

10,500 

148,700 

8.9 

34 

w 


W3F 

W+An+A 

540 

8,100 

115,290 

9,200 

131,100 

4.1 

15 

HAZb 


W3G 

W+An+A 

540 

8,300 

118,000 

9,700 

137,600 

14.0 

14 

W 


W2E 

W + A 

25 

10,500 

149,600 

12,500 

178,200 

6.9 

30 

w 


W2K 

W + A 

25 

10,600 

150,700 

12,700 

179,800 

7.6 

28 

w 


W2G 

W + A 

425 

9,500 

134,700 

10,900 

154,800 

8.5 

33 

HAZb 


W2L 

W + A 

425 

9,200 

130,400 

10,100 

143,800 

4.1 

22 

W 


W2I 

W + A 

540 

9,200 

131,140 

10,400 

148,300 

7.3 

31 

W 


W2H 

W + A 

540 

8,700 

123,700 

9,500 

135,800 

4.5 

22 

W 


a Heat treatment: W — Welded. 

An - Annealed at 950°C for 1 hour, furnace cooled to aging temperature. 

A — Aged at 715°C for 8 hours, furnace cooled to 620°C and held 12 hours, including furance cooling time. 
b HAZ - Heat-affected zone. 


TABLE 7.20 


CHARPY V-NOTCH IMPACT STRENGTH 
OF INCONEL ALLOY 718 WELDMENTS 





Energy Absorbed 

Notch 

Specimen 

25°C 


—40°C 

Location 3 

Condition b 

ft—lb 

m-kg 

ft—lb 

m-kg 

HAZ C 

W + A 

21 

2.90 

10 

1.38 

Weld 

W + A 

16 

2.11 

11 

1.52 

Weld 

W + A 

20 

2.76 

11 

1.52 

BM d 

A 

20 

2.76 

17 

2.35 

Parallel to RD 

BM d 

A 

10 

1.38 

9 

1.24 

Transverse to RD 

HAC c 

W+An+A 

- 

- 

10 

1.38 

HAZ C 

W+An+A 

— 

- 

8 

1.08 

HAZ C 

W+An+A 

- 

- 

8 

1.08 

Weld 

W+An+A 

- 

- 

6 

0.83 

Weld 

W+An+A 

- 

- 

6 

0.83 

Weld 

W+An+A 

- 

- 

6 

0.83 

BM d 

An+A 

20 

2.76 

20 

2.76 

Parallel to RD 

BM d 

An+A 

10 

1.38 

9 

1.24 

Transverse to RD 


a AI! notches perpendicular to surface of plate. 

^Heat treatment: W — Welded. 

A - Aged at 715°C for 8 hours, furnace cooled to 
620°C, held 12 hours, including furnace cooling 
time. 

An+A — Annealed at 950°C for 1 hour, furnace cooled 

to 715°C, held 8 hours, furnace cooled to 620°C, 
held 12 hours including furnace cooling time. 
C HAZ - Heat-affected zone. 
d BM - Base metal. 



300 


specimens. This material was air melted and vacuum consumable electrode remelted and 
arrived in the heat-treated condition. Nominal heat-treating procedures are as follow: 

Normalize: 900°-930°C - 1 hr/2.54 cm thickness, air cool 

Austenitize: 845°±10°C - 1 hr/2.54 cm thickness, water or oil quench 

Temper: Double temper - each temper consists of 2 hours minimum at 540°C (for 
13, 360 kg/cm 2 minimum tensile strength). 

The vendor reports the following room-temperature properties for this material: 

Tensile strength, kg/cm 2 13,910 (198, 000 psi) 

Yield strength, kg/cm 2 13, 290 (188, 600 psi) 

Elongation in 5.08 cm, % 17 

Reduction of area, % 68.5 

Charpy V-notch impact energy (avg), m-kg 9.1 (66 ft-lb) 

Welding wire designed to match the base metal properties in the as-welded condition is 
being procured. 

7. 2 CONSIDERATION OF PRESSURE VESSEL FAILURE MODES AND RELATED 
PROPERTIES 

Studies have been made to identify special material properties beyond those normally 
obtained which may eventually be required to properly qualify a potential pressure ves¬ 
sel material. This gives reasonable assurance that no major basic limitation of the con¬ 
cept of using high-yield-strength materials or of using a given advanced pressure vessel 
material is overlooked. 

The first step in these studies is to consider potential failure modes and to identify 
special material properties or tests required to evaluate a factor of safety for each mode 
of failure. Six principal modes of failure have been considered: 

1. Burst due to over-pressure. 

2. Fatigue cracking. 

3. Fast fracture including both brittle fracture and low-energy tearing. 

4. Buckling collapse due to external pressure. 

5. Corrosion, including stress corrosion and corrosion fatigue. 

6. Progressive distortion (ratchetting). 

Although the foregoing six modes of failure are potentially all significant failure modes, 
only the first three are considered immediately pertinent to the development of advanced 
pressure vessel materials. The other three failure modes are being eliminated from 
further studies of advanced pressure vessel materials for the following reasons. 

Buckling collapse is eliminated from further consideration simply because the vast ma¬ 
jority of pressure vessels pertinent to this program are vessels subject only to internal 
pressure. There is a possibility of developing an instability at the knuckle of a poorly de¬ 
signed dished head, but this is considered primarily a design problem unrelated to devel¬ 
opment of advanced pressure vessel materials. 

Corrosion, on the other hand, is definitely a materials problem, but it is assumed that, 
should corrosion be a problem with these materials, a corrosion-resistant cladding could 
be applied in a manner similar to that currently used for the conventional low-carbon steels. 

Progressive distortion or ratchetting is eliminated from further consideration because 
it is primarily a design problem. A well designed structure (including adequate analysis) 
should not be subjected to ratchetting. 



301 


The remaining three failure modes require careful consideration in the development of 
advanced pressure vessel materials, and are treated in some detail in the following para¬ 
graphs. 

BURST DUE TO OVER-PRESSURE 

Investigators have shown 7 that burst pressure can be expressed in terms of ultimate ten¬ 
sile strength, strain-hardening exponent, and geometry of the vessel. For a thin-walled 
cylinder, Langer 7 gives the following equation for the burst pressure: 

p b = a u F cyl (d[) ( X - d[) (7 ‘ 1} 

where 


P^ = burst pressure 

Ou = nominal stress at ultimate load, i. e., nominal engineering ultimate tensile 
strength 

t = vessel wall thickness 
di = vessel inside diameter 
[ 0.25 1 /e\ €u 

Fc y! l_e u + 0.227J 

e = base of the natural logarithms 
c u = true strain at ultimate load 

Langer states that equation (7.1) should be used for 

W< 1.4 


di+2t 2t 

where W = do/di = —— = 1 + ~r 
a i a i 

and d Q = vessel outside diameter 

when the strain-hardening exponent, n, is defined as the true strain at ultimate load n = e u 
and equation (7.1) then defines the burst pressure in terms of the nominal tensile strength, 
strain-hardening exponent, and vessel geometric parameters. 


For a cylindrical vessel designed to Section III of the ASME Boiler and Pressure Vessel 
Code with the minimum permissible thickness, the design pressure, P^, is given by the 
following: 


p <i =2s «» (djt+t/dj 

a ^ U)! 1 ■ i] 


where 


S m = membrane stress allowable. 

If one defines the factor of safety, FS, on burst as the ratio of the burst pressure to the 
design pressure, and if it is further assumed that for the high-strength materials in this 
program the value of S m will always be based on the ultimate tensile strength, and using 
the Section III value of 


S m = 1/3 CXy , 

^B. F. Langer, "PVRC Interpretive Report of Pressure Vessel Research, Section I - Design Considerations," Welding Research 
Council Bulletin, No. 95, April 1964. 



302 


the factor of safety on burst due to over-pressure for a thin-walled cylinder becomes 
simply: 


FS = 3F cy i. (7.3) 

The definition of F cy | indicates that it is a function only of the strain-hardening expon¬ 
ent (n = e u ), Langer' s tabulated values of F cy i as a function of n together with equation 
(7. 3) were used in preparing Figure 7.12 which shows the relationship between the factor 
of safety and the strain-hardening exponent. 

Gross 8 gives plots which show the variation of yield strength and strain-hardening ex¬ 
ponent with tensile strength for typical pressure vessel steels. Figure 7.13 combines 



0 0.2 0.4 0.6 0.8 1.0 

Strain-hardening exponent, n 


Fig. 7.12 — Factor of safety of burst of over pressure for a thin-walled 
cylindrical pressure Vessel 



0.5 0.6 0.7 0.8 0.9 1.0 

Yield/tensile strength ratio 

Fig. 7.13 — Variation of the strain-hardening exponent with the yield-to- 
tensile strength ratio for typical pressure vessel steels — based 
on the data of Gross 

8 •• 

J. H. Gross, "PVRC Interpretive Report of Pressure Vessel Research, Section II - Materials Considerations," Welding Research 
Countil Bulletin, No. 101, November 1969. 






303 



0.5 0.6 0.7 0.8 0.9 1.0 


Yield/tensile strength ratio 


Fig. 7.14 — Variation of factor of safety with the yield-to-tensile strength ratio — 
obtained by cross plotting information contained in Figures 7.12 
and 7.13 


these plots of Gross to obtain a plot of the variation of the strain-hardening exponent with 
the yield-to-tensile strength ratio for the typical pressure vessel steels. The information 
contained in Figures 7.12 and 7.13 was combined to obtain the plot of factor of safety 
versus the yield-to-tensile strength ratio shown in Figure 7.14. This plot indicates that 
the factor of safety on burst due to over-pressure tends to increase with increasing yield- 
to-tensile strength ratio. 

The foregoing discussion demonstrates the need for somewhat more detailed tensile data 
and evaluation than are normally obtained from routine tensile tests in order to evaluate the 
factor of safety on burst due to over-pressure. In addition to the normal data, the true strain 
at the ultimate (maximum) load is required. The true strain at the ultimate load is needed 
to define the strain-hardening exponent which, in turn, can be used to obtain a calculated 
factor of safety on burst. Strain-hardening exponents obtained for candidate pressure ves¬ 
sel materials can be compared with the data of Gross presented in Figure 7.13 to assess 
the behavior of these materials compared to current typical pressure vessel materials. 


FATIGUE CRACKING 


Coffin, 9 Coffin and Tavernelli, 10 and Manson 11 have shown that low-cycle (or strain) fa¬ 
tigue data can be related to the tensile properties of a material. Manson has presented 
several different techniques for doing this, but his method of Universal slopes which has 
been correlated with fatigue data on numerous materials was selected for these studies due 
to its simplicity of application. Manson gives the following equation for relating the fatigue 
strain range with the tensile properties: 


a u 

Ac = 3.5 — 
E 



(7.4) 


^L. F. Coffin, Jr., "A Study of the Effects of Cyclic Thermal Stressses on a Ductile Metal," Trans. ASME, Vol. 76,1954, 
pp. 931-950. 

^L. F. Coffin, Jr., and J. F. Tavernelli, "Experimental Support for a Generalized Equation Predicting Low-Cycle Fatigue," 
Trans. ASME, Vol. 84, Series D, 1962, pp. 533-591. 

1 ^S. S. Manson, "Fatigue: A Complex Subject — Some Simple Approximations," Exp. Mech., Vol. 5, No. 7,1965, pp. 2-35. 







304 


where 

Ae = total strain range, elastic plus plastic 
E = modulus of elasticity 
ct u = nominal stress at the ultimate (max) load 
Nf = number of cycles to failure 
D = ductility = [1/(1-RA)] 

RA = reduction in area expressed as a fraction 

Manson does not propose that this equation be used in lieu of actual fatigue data. Since 
it has correlated well with fatigue data for a wide range of materials, however, it is con¬ 
sidered to give a good indication of the potential fatigue performance of materials. Accord¬ 
ingly* Manson's equation (7. 4) has been used to give a preliminary indication of the poten¬ 
tial fatigue performance of the candidate advanced pressure vessel materials. 

Since these advanced pressure vessel materials are specifically intended for nuclear 
service, the Manson equation was used to compare potential fatigue properties of these 
materials with design fatigue strength curves for the currently approved materials pre¬ 
sented in Section III of the ASME Boiler and Pressure Vessel Code for Nuclear Vessels. 

To make this comparison, the Code philosophy as delineated in the supplementary criteria 12 
issued with the first edition of Section III was used to apply corrections for the effect of 
mean stress and factors of safety on stress and cycles. 

Code criteria in essence assume that the materials behave in an elastic - perfectly plas¬ 
tic manner and that, due to residual stresses of unknown magnitude in'the as-fabricated 
vessel, one must assume that the maximum possible mean stress exists at any point in the 
vessel. The Code criteria also assume a linear rule for correcting for the effect of mean 
stress. 

Figure 7.15 helps to illustrate the manner in which the correction for mean stress has 
been made. After shakedown,* all possible stress states lie either within the triangular 
region A-O-B along line A-B, or on the alternating stress axis above point A. Line C-D 
from the point on the alternating stress axis represents the stress, S N , required to pro¬ 
duce failure in N cycles at zero mean stress to the ultimate tensile strength on the mean 
stress axis. This line represents the limiting combinations of mean and alternating stress 
required to produce failure in N cycles according to the linear rule. 

However, since line A-B represents the limits of the possible stress states, point E 
corresponds to the highest possible mean stress point corresponding to N cycles. The 
corresponding alternating stress value, is the alternating stress amplitude corrected 
for the maximum possible effect of mean stress. Applying this procedure for a number of 
different values of N gives the fatigue curve corrected for the maximum effect of mean 
stress. The next step in obtaining the calculated fatigue design curve is to apply the Code 
factors of safety of 2 on stress and 20 on cycles to the fatigue curve corrected for the 
maximum effect of mean stress. The calculated design fatigue strength curves obtained 
in this manner are compared with the applicable Section III design fatigue strength curves 
in Figures 7.16 and 7.17. 

For many materials the Peterson Cubic Rule 13 is a better representation of the fatigue 
behavior under combined mean and alternating stress; hence the foregoing comparison 
was repeated using this rule, and results are shown in Figures 7.18 and 7.19. 


‘Shakedown refers to an elastic cyclic condition after a few cycles which is assumed to be assured by the Code's 3S limit 
on the primary plus secondary stresses. m 

12 " 

Criteria of Section III of the ASME Boiler and Pressure Vessel Code for Nuclear Vessels," ASME, New York, N.Y., 1964. 

13 i# 

R. E. Peterson, Brittle Fracture and Fatigue in Machinery," Fatigue and Fracture of Metals, Wiley New York NY 1952 
pp. 79-102. ' ' ‘ *' 




305 



Fig. 7.15 — Modified Goodman diagram showing the application of the linear rule to obtain the 
correction for the maximum effect of mean stress 



Fig. 7.16 — Comparison of calculated design fatigue curves for PH 13—8Mo and 
12Ni—5Cr—3Mo (using the Linear Rule to correct for the maximum 
effect of mean stress) with design fatigue curve for carbon and alloy 
steels from Section III of ASME Boiler and Pressure Vessel Code for 
Nuclear Vessels 

Another item which should be considered in evaluating the maximum effect of mean stress 
correction is the fact that, in general, high yield strength materials tend to exhibit a high 
degree of cyclic strain softening, a decrease in the stress range with cycles during strain 
cycling tests. This amounts to a. decrease in the effective yield strength with cycles, even¬ 
tually stabilizing at some asymptotic value after many cycles. This asymptotic value is 
commonly referred to as the dynamic yield strength. It has been suggested 14 that it may be 
more realistic to base the mean stress correction on the dynamic yield strength rather than 

^Dr. W. E. Cooper, private communication. 





























































Alternating stress, psi Alternating stress, psi 


306 



70,000 


7,000 


10 2 


10 3 10 4 
Number of cycles 


105 


70 


loB 


Fig. 7.17 - Comparison of calculated design fatigue curve for Inconel 718 (using the 
Linear Rule to correct for the maximum effect of mean stress) with 
design fatigue curve for 18-8 stainless steels and nickel-chrome-iron 
alloy from Section III of ASME Boiler and Pressure Vessel Code for 
Nuclear Vessels 



Fig. 7.18 - Comparison of calculated design fatigue curves for PH13 - 8Mo and 
12Ni — 5Cr — 3Mo (using the Peterson Cubic Rule to correct for 
the maximum effect of mean stress) with design fatigue curve for 
carbon and alloy steels from Section III of ASME Boiler and Pressure 
Vessel Code for Nuclear Vessels 


Alternating stress, kg/cm 2 Alternating stress, kg/cm : 





































































307 



Fig. 7.19 - Comparison of calculated design fatigue curve for Inconel 718 (using the 
Peterson Cubic Rule to correct for the maximum effect of mean stress) 
with design fatigue curve for 18-8 stainless steels and nickel-chrome-iron 
alloy from Section III of ASME Boiler and Pressure Vessel Code for 
Nuclear Vessels 

the static yield strength for these high yield strength materials. Manson 15 shows data for a 
4340 steel in which the asymptotic value of the stress range is only about two-thirds its ini¬ 
tial value. To examine the potential significance of this factor, assumed values of dynamic 
yield strength, y, of 8790 and 10, 540 kg/cm^ were used to obtain calculated design fatigue 
strength curves for the 12Ni - 5Cr - 3Mo material, and these are compared with the calcu¬ 
lated curve based on the static yield strength, a sy , and with the corresponding Section III 
design fatigue strength curve in Figure 7. 20 using the linear rule and in Figure 7. 21 using 
the Peterson Cubic Rule. These figures demonstrate that the use of dynamic yield strength 
and the Peterson Cubic Rule can greatly influence assessment of the fatigue behavior of 
these high-strength materials. 

The preceding discussion of fatigue properties is only an indicator, at best, of the poten¬ 
tial fatigue properties of the candidate pressure vessel materials, but it does serve to iden¬ 
tify significant facets of the fatigue aspects which will be considered in evaluating the can¬ 
didate materials. 

In summary, the foregoing analysis indicates: 

1. Up to about 400 cycles, the 12Ni - 5Cr - 3Mo and PH13-8Mo materials appear to be 
potentially somewhat better than the conventional materials regardless of the type of 
mean stress correction used, but this remains to be demonstrated through actual fatigue 
test data. (Many reactor pressure vessels are currently designed on the basis of fewer 
than 500 specified operating cycles.) 

2. The Inconel alloy 718 appears to have a somewhat reduced fatigue strength over the 
full cyclic range compared to the Section III curve for the 18-8 stainless steels and 

Ni-Cr-Fe alloy. This remains to be demonstrated through actual fatigue tests. 

\ 

15 


'Manson, op. cit. 




Alternating stress, psi Alternating stress, psi 


308 



Number of cycles 


Fig. 7.20 - Comparison of calculated design fatigue curves for 12Ni — 5Cr - 3Mo 
corrected (Linear Rule) for mean stress using assumed values for the 
dynamic yield strength, y , with the uncorrected curve and with the 
curve corrected using the static yield strength, a sy 



Fig. 7.21 — Comparison of calculated design fatigue curves for 12Ni — 5Cr — 3Mo 
corrected (Peterson Cubic Rule) for mean stress using assumed values 
for the dynamic yield strength, y , with the uncorrected curve and 
with the curve corrected using the static yield strength, a sy 


J 


Alternating stress, kg/cm 2 Alternating stress, kg/cm2 




















































309 


3. The use of the Peterson Cubic Rule indicates significant potential gains in design fa¬ 
tigue life. The validity of the cubic rule should be investigated experimentally if such 
gains need to be exploited. 

4. Potentially even more significant than the use of the cubic rule are the gains indicated 
through the use of the dynamic yield strength in lieu of the static yield strength in mak¬ 
ing mean stress corrections. These results identify the need for experimental investi¬ 
gations exploring use of the dynamic yield strength in lieu of the static yield strength 
in making the mean stress corrections. They demonstrate specifically the need to es¬ 
tablish the actual dynamic yield strength of these materials in the presently planned 
fatigue tests to determine whether it would be advantageous to use dynamic data in 
making the mean stress correction. 

FAST FRACTURE BY EITHER BRITTLE FRACTURE OR LOW-ENERGY TEARING 

As used herein the term fast fracture refers to either brittle fracture or low-energy tear¬ 
ing. These types of failure are characterized by high-velocity crack propagation with little 
prior plastic deformation. These types of failure have been studied extensively in recent 
years, but no single approach has achieved wide acceptance for the full spectrum of signifi¬ 
cant strength levels. The two approaches most widely accepted are the NRL Fracture Analy¬ 
sis Diagram Procedure 16 and the fracture mechanics approach which is the outgrowth of 
Griffith's failure theory for brittle materials and its subsequent modifications for metallic 
materials by Orowan 17 and Irwin. 18 

Brittle Fracture 

NRL Fractural Analysis Diagram Procedures - These procedures are generally accepted 
as providing a usable procedure for the lower-strength steels with yield strengths below 
about 3500 kg/cm 2 . Pellini and Puzak have used the fracture analysis diagram procedures 
to correlate numerous service failures, 16 including some examples for higher-strength ma¬ 
terials. This procedure applies only to materials having a "well defined transition tempera¬ 
ture, " according to Pellini and Puzak; which is apparently not true of Inconel 718 material. 19 
Since this procedure has not been specifically validated for materials of the strength levels in¬ 
volved in this program, considerable experimental data will be needed if this approach is to 
be employed in fracture-safe evaluation of these advanced pressure vessel materials. To 
establish a fracture analysis diagram for these materials one needs (1) crack initiation data 
to define the crack initiation curves, (2) crack arrest data to define the crack arrest (CAT) 
curve, (3) Charpy V-notch or drop-weight test data to define the nil-ductility transition tem¬ 
perature (NDT), (4) explosion tear test 16 data to define the fracture transition plastic (FTP) 
point, and (5) correlation with service failures of model pressure vessels to establish vali¬ 
dity of the procedure in this strength range. 

Fracture Mechanics Approach - This approach is currently limited to the linear fracture 
mechanics range in which stress levels are in the elastic range everywhere except very 
locally at the crack tip. This limits current usefulness to very high strength materials 
(yield strengths 14, 000 kg/cm 2 ) and to materials operating below their nil-ductility transi¬ 
tion temperature (NDT). Since the materials of interest in this program approach the 
14, 000 kg/cm 2 yield strength level, it appears appropriate to consider determination of 

7 ®W. S. Pellini and P. P. Puzak, "Fracture Analysis Diagram Procedures for the Fracture-Safe Engineering of Steel Structures," 

NRL Report 5920, March 1963. 

^E. Orowen, "Fundamentals of Brittle Behavior in Metals," Fatigue and Fracture of Metals, J. Wiley and Sons, New York, N.Y., 

1952. 

R. Irwin, "Fracture Mechanics in Structural Mechanics," Proceedings of the First Symposium on Naval Structural 
Mechanics, Permagon Press, New York, N.Y., 1960. 

19 


GEMP-1002, p. 115. 




310 


plane-strain fracture toughness (K IC ) values for these materials. Of the many different 
test specimens currently in use for the determination of K IC values, it has been suggest¬ 
ed 20 that the wedge-opening loading (WOL) specimen developed by Westinghouse is most 
applicable for use with the heavy sections appropriate for pressure vessels. 

Low-Energy Tearing - The term low-energy tearing has been used by Pellini and Puzak 
to describe "fracture propagation not involving a cleavage mode but resulting from a very 
low energy absorption in tearing." Neither the fracture mechanics approach nor the NRL 
fracture analysis diagram procedures have been proved applicable for analysis of the low- 
energy tearing problem. Pellini and Puzak 21 indicate modifications to the fracture analysis 
diagram which they apparently feel would make the same type of procedure applicable to 
the low-energy tear case. This remains to be demonstrated and much more work is required 
before this procedure or the specific tests required to define the procedure can be expli¬ 
citly identified. Pellini and Puzak point out that the higher-strength materials may be par¬ 
ticularly susceptible to low-energy tearing of the welds or the heat-affected zone; they 
recommend using their explosion bulge test of welded sections to obtain at least a relative 
picture of the weld and heat-affected zone low-energy tearing problem. 

7.3 GENERAL DISCUSSION 

All alloys for which data have been obtained have relatively high yield-to-tensile strength 
ratios which would cause their Design Stress Intensity Values (S m ) to be based on the one- 
third tensile strength criterion of the Code. Table 7.21 is a compilation of tentative values 
for three alloys studied in this program, based upon the tests made to date. Modified heat 

TABLE 7.21 


ESTIMATED DESIGN STRESS INTENSITY VALUES, Sm, 
FOR ADVANCED PRESSURE VESSEL MATERIALS 


Grade 

Maximum Allowable Stress? 
for Maximum Metal Temperature, psi 


38°C 

315°C 

371°C 

427°C 

538°C 

12IMI -5Cr-3Mo 

59,000 

47,800 

_ 

41,000 


PH13-8MO 

65,000 

54,000 

— 

48,000 

_ 

Inconel alloy 718 

63,300 

— 

_ 

59,000 

56,600 

SA212B b 

23,300 

18,700 

18,300 

__ 

_ 

SA302B b 

26,700 

26,700 

26,700 

_ 

_ 

SA533B, b Class 1 

26,700 

26,700 

26,700 

- 

- 


a The maximum allowable stress values for the materials based on the lowest 
of the following: 

1/3 of the specified minimum tensile strength at room temperature. 

1/3 of the tensile strength at temperature. 

2/3 of the specified minimum yield strength at room temperature. 
b Actual values listed in Section III of ASME Boiler and Pressure Vessel Code. 


treatments and multiheat data would alter these values somewhat. Included in the table are 
corresponding Code-approved values for three materials of current use in pressure vessels 
Of significance is the potential strength margin possessed by the advanced materials, al¬ 
though this is only one of many aspects which must be considered before any materials can 
be proved suitable for advanced pressure vessels. In the following paragraphs, an attempt 
has been made to assess the relative potential of these alloys, based on results to date. 


20 

Dr. W. E. Cooper, private communication. 

21 u 

W. S. Pellini and P. P. Puzak, "Practical Considerations in Applying Laboratory Fracture Test Criteria to the Fracture-Safe 
Design of Pressure Vessels/' NRL Report 6030, November 1963. 


311 


12Ni - 5Cr - 3Mo 

Properties determined thus far for base metal and welded joints have involved one plate 
composition (heat No. 01378) and one filler metal composition, 12Ni - 3Cr - 3Mo (heat No. 
01434). Both base metal and weldments, when given the recommended aging treatment of 
480°C for 3 hours, showed room-temperature properties which were lower than expected. 
Re-aging of the base metal and weldments for longer times, however, resulted in improved 
hardness and strength. A major advantage of the alloy is its favorable aging characteristics 
(slow rise to peak hardness) in the temperature range which produces a favorable balance 
in strength and toughness. 

Weld efficiencies have exceeded 90 percent and the possibility exists that improvements 
can be made with modified aging procedures. The 12Ni — 3Cr - 3Mo filler metal on hand 
is apparently not optimum from a weld efficiency standpoint, however, and current efforts 
center on a matching 12Ni - 5Cr — 3Mo filler metal. The new 2. 54-cm plate (heat No. 
1P0193) and weld filler wire (heat No. V91230) are expected to develop higher properties 
under the same heat treatments with better joint efficiencies. Welding characteristics of 
heavier sections remain to be determined. 

It was expected from the outset that this alloy could achieve satisfactory weld properties 
by aging only after welding. The attractiveness of 12Ni - 5Cr - 3Mo, as well as the other 
two age-hardening alloys, PH13-8M0 and Inconel alloy 718, would be much reduced if a 
complete heat-treating cycle after welding is required. It appears that a suitable aging 
treatment can be devised which will make it unnecessary to solution-treat prior to aging 
after welding the 12Ni - 5Cr - 3Mo alloy, although this is not clearly established. 

Table 7.22 and Figure 7. 22 compare some properties of this alloy with those of PHI 3- 
8Mo and Inconel alloy 718. The 12-5-3 (heat 01378) shows the lowest tensile strength, is 
intermediate in yield strength, shows little or no orientation effects in tensile tests but 
some in impact tests, and has the best overall Charpy V-notch energy absorption values. 

If the alloy responds favorably relative to the foregoing considerations, achieves the low 
cycle fatigue life estimated for it, and continues to show the radiation resistance found in 
early tests at NRL, 22 it must be considered a primary candidate for advanced pressure 
vessels. 

In any case, continued consideration of this alloy is planned with emphasis on determin¬ 
ing DWTT, low-cycle fatigue behavior, radiation effects, weldability, and weld properties 
in heavy sections. 

PH13-8MO 

Studies of PH13-8Mo have involved one heat (heat No. VC5281) of 1. 27-cm-thick plate 
and one heat of wire filler metal (heat No. 5178). The alloy showed the highest room- 
temperature tensile and yield strengths, but also the highest degree of anisotropy. Its 
Charpy V-notch impact energy absorption level fell somewhat more rapidly with tempera¬ 
ture than the 12-5-3 alloy. 

PH13-8Mo developed the highest weld strength and had the best joint efficiency (98%), 
based on an average plate strength of the two orientations tested. It is not yet clear, how¬ 
ever, whether the alloy will develop satisfactory properties after welding by aging alone. 
Weldments planned for the 2. 54-cm-thick plate will be used to further evaluate this ques¬ 
tion. Aging studies on parent metal indicate that for aging temperatures which produce 
adequate toughness (e.g., 565°C), the alloy ages quite rapidly to a very high level and then 

22L. E. Steele, J. R. Hawthorne, R. A. Gray, "Neutron Irradiation Embrittlement of Several Higher Strength Steels/' 

NRL-6419, September 7, 1966. 





COMPARISON OF ROOM-TEMPERATURE PROPERTIES OF THREE ALLOYS IN PLATE AND WELD JOINT 3 


312 


> 

+-> o 
C C 

o •£ * 


cn 

Z e 
> >: 
o o> 


! * 
a d 

c ^ 

o C 


> >: 

O o) 


> to 

.1 I c 

5 — ~ 

S*8 


I I 

LO LO 
CO CO 


I I 


I I 


c/) </> 

0 cu 

r- c- 


co 00 
cm cm' 


^2 ii ^ 


i i 


i i 


i i 


o o 
o o 

o' CD 


O 

00 


O O 
in in 

CN CN 

cm" cm' 


o o 
o o 

CM CM 

^t «t 


r-* r*^ 

^ O* 


o 
cm o 


o o 
o o 
in lo 
o' o' 


o o 
o o 

CD CD 


CO CO 
V - 1 Csl 


o o 
m in 
in in 
o' o' 


o o 
o o 

CD CD 

cm' cm' 


1 1 


1 I 


a) o | 


o 

o 

o 

o 

o 

o 

o 

o 

o 

O 

o 

o 

o 

in 

o 

o . 

o 

o 

o 

in 

r>» 

o 

o 

o 

"st 

CM 

CM 

CM 1 


CM 

05 

CD 

o 

o 

CO 

o 

T— 

cm' 

cm' 

cm' 

't 

co' 

o' 

o' 

cm' 


▼— 

o' 

*” 

*“ 

t- 

r~ 

T_ 

T “ 



*“ 

r- 



o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

o 

in 

o 

o i 

o 

o 

o 

o 

in 

o 

o 

o 

05 

LO 

CD 

CD 1 

LO 

1 ". 

CO 

in 

CO 


o 

in 

T—' 

cm' 

cm' 

cm' 

in' 

co' 

co' 

co' 

'St 

't 

co' 

co' 

1 


' 

' 

*— 

4 - 

«■“ 

r— 






HO- H Q_ 


H o_ 


H a. Ho. H a. 


5 

4— 

o 

_0 

a 

£ 


‘ca +3 

■H* CD 

•a 
o 

CD 


~o 
o 
in ^ 


o 

o _j 

o o 


i o 


CD CD 

I « 


1 


7 * 

1 CO 

O | 
o 1 
o (J 
CO o 
't 


•= 't 
t I 
CM I 

I ° 
1 o 
O in 
o CD 


X £ 
ro w 

CD CD 

E £ 

1 = 

H- CD ~ 
° 0> £ 
£ -Q g 

cn O F 

c +J £ 

£ S S 

a t s 

jl) -U 
‘tfj CD 

c x: _ 

0 _ . 

•£ 1 s 

111 


0 


CD <d 

E E 


<s> — 


O Crt — 

CD a> 

fas 



CO 

in 


H 

H 

H 


4 -’ 

0 

't 

't- 

0 

+j 

■M 


03 

</> 

0 

O) 

+ 

+ 

0 

O) 

0 

0 

0 

0 

0 

0 

"0 

03 

< 

1 

< 

1 

< 

1 

< 

X 

1 

X 

1 

X 

1 

0 

CL 

|| 

to 

C 

0 

< 

CO 

o 

o 

LU 

UL 

o 

CL 

II 


x: ; 

■M - 
° 1 
3 i 

-O H 


0 


0 c 

Q_ 0 

2 g 

Q- ro 

CD 0 

cn +3 

0 4 _> 

s s 

< X 


0 0 Mr co 

| iH± 

« rr c n 
£ h= 0 o 
2 0 E "o 
1“ c 0 19 

» o £ 5 

1 — “3 H 5 

u ~o 0 


J 




313 


Orientation 12Ni-5Cr-3Mo PH13-8Mo Inconel 718* 

o □ O 

• ■ ♦ 


Perpendicular 
to R.D. 
Parallel to R.D. 





Fig. 7.22 — Comparison of tensile and impact properties of 12Ni — 5Cr — 3Mo, PHI3 — 8Mo, and 
Inconel 718 alloys at various temperatures 

overages to a significant degree within about a 20-hour period. This type of aging behavior 
is not favorable for heavy, complex shapes. 

The estimated low-cycle fatigue behavior and indicated radiation resistance of PH13-8M0 23 
appear to favor its consideration as an advanced pressure vessel material. 

Continued evaluation of this alloy is planned for the immediate future, with emphasis on 
further delineating aging behavior, weldability, weld properties, DWTT, preliminary low- 
cycle fatigue characteristics, and completing radiation effects studies currently in process. 

INCONEL ALLOY 718 

Investigation of Inconel alloy 718 has involved one composition of plate and one lot of fill¬ 
er metal. Figure 7.22 and Table 7. 22 show a comparison of certain properties of Inconel 
alloy 718 with the PH13-8M0 and 12Ni - 5Cr - 3Mo alloys. Tensile properties of this alloy 

23 


'Unpublished data, NRL. 





314 


are only slightly lower than those of PH13-8Mo at room temperature but, as expected, they 
eventually exceed that alloy as test temperature increases. The low toughness values for 
both parent metal and welds place the potential of this alloy in some question, however. An 
alternate heat treatment (III) expected to improve toughness is being investigated. Toughness 
values do not change significantly with temperature nor are they apparently worsened by 
radiation to approximately 10 19 n/cm 2 (E n > 1 Mev). Drop-weight tear tests will be conducted 
at NRL for additional information on the toughness of this fee alloy. 

Weld joint properties of Inconel alloy 718 show welding efficiencies of 80 percent with 
heat treatment I and 87 percent with heat treatment II. No testing has been done on weld¬ 
ments given heat treatment III. Compared to the parent metal, weld joint properties of 
Inconel alloy 718 show lower tensile ductility and reduced toughness. 


No new work is contemplated on this alloy in the immediate future, although efforts will 
be made to complete studies already initiated. 

7. ^SUMMARY AND CONCLUSIONS ^ 

The following data have been obtained on approximately 2. 54-cm-thick plate of 12Ni - 
5Cr - 3Mo, PH13-8Mo, and Inconel alloy 718, relative to the potential of these candidate 
high-strength materials for advanced pressure vessels. 


12Ni - 5Cr - 3Mo 

The 12Ni - 5Cr - 3Mo alloy possesses aging characteristics which appear to be suited 
to the uniform agtog of heavy sections. Its tensile properties and toughness are most prom¬ 
ising. Low-cv cleiatigue data and effects of irradiation on tough n ess remain to be deter¬ 
mined. The alloy give^vjdence of adequate structural stability at temperatures up to at 
least 315 C. The weldability of the 12-5-3 composition with 12-3-3 filler metal is good but 
the resulting weld efficiency should be even better if a filler metal composition (12-5-3) 
matching the base metal is employed. Further data are necessary to indicate whether 
post-weld aging alone will produce adequate properties. 

The 12-5-3 alloy continues to be regarded as a promising candidate for advanced pres¬ 
sure vessels, and will be evaluated in thicknesses &10 cm. 


PH13-8MO 

£> 

The .a ging characteristics of PH13-8Mo, at least for aging temperatures which produce 
higj* Roughness levels, may not bgjjrited to the uniform aging of heavy sections.^Excellent 
tensile properties and good toughness can be developed in the alloy. 
data and effects of irradiation on toughness remain to be determined. PH13-8Mo appears 
acceptably stitolj, structurally upon longtime heating at temperatures to 425°C. The alloy 
has good weldability and weld strength with filler metal of matching composition; weld 
ductility and toughness are only fair. It is not yet clear whether aging alone will be an 
acceptable post-weld heat treatment. 


Further qualification of this alloy will be made on a 2. 54-cm-thick plate, before heavy 
section studies are considered. 


INCONEL ALLOY 718 

Excellent strength properties characterize Inconel alloy 718, particularly at higher tem¬ 
peratures. It is much superior to the 12-5-3 alloy or PH13-8Mo in this respect. Its tough¬ 
ness is quit^lgw, based on Charpy V-notch tests.jojjder two conditions of heat treatment; 
a third heat treatment is being evaluatedJtor tj>ugtoiess. Initial radiation effects indicate 
no reduction in toughness levels. Low - c ycl e fatigue data remain to be determined. The 
alloy apparently possesses good structural stability”at temperatures to at least 540°C. In- 





315 


^ com 


conel alJLoy^718 developed sound welds with filler metal of matching composition, although 
weld Strength and toughness were significantly lower than those of the parent metal. 


No new work will be initiated on this alloy; studies presently underway will be completed. 

Fj & 

A heat-treatable steel, HP 9Ni — 4Co — 0. 20C, has been added to the program. 


Consideration has been given to the types of failure which need to be prevented in pres¬ 
sure vessel applications, and special material properties which are related, e.g., strain¬ 
hardening coefficient, dynamic yield strength, and plane strain fracture toughness. It is 
estimated that the 12-5-3 and PH13-8MO alloys, at least, possess satisfactory low-cycle 
fatigue behavior, particularly if dynamic yield strength and the Peterson Cubic relation¬ 


ship can be employed in applying mean stress corrections. 
7. 5 PLANS AND RECOMMENDATIONS 




Prepare additional weldments of 2. 54-cm-thick plate of the 12-5-3 alloy using filler 
metal of matching composition. 

Initiate preparation for welding of 10-cm-thick plate of 12-5-3 alloy. 

Complete initial property determinations on the 10-cm-thick plate of 12-5-3 alloy. 

Continue heat-treating studies on 12-5-3 and PH13-8Mo alloys. 

Prepare weldments of 2. 54-cm-thick plate of PH13-8Mo. 

Continue study of heat treatment influence on toughness of Inconel alloy 718. 

Complete initial determination of irradiation effects on toughness of 12-5-3 and PHl3-8Mo. 

Complete preparation of low-cycle fatigue specimens of 12-5-3, PH13-8Mo, and Inconel 
alloy 718. 

Complete preparation of DWTT specimens of 12-5-3, PH13-8Mo, and Inconel alloy 718. 




8. PHYSICO CHEMICAL STUDIES OF CLAD U0 2 
IN POTENTIAL MELTDOWN ENVIRONMENTS 


(1175) 

J. F. White* 


The objective of this program is to study the behavior of zirconium-alloy-clad UO 2 fuel 
elements in steam and steam plus air from 1000°C to the melting point of U0 2 . This in¬ 
cludes measurement, when necessary, of physical and mechanical properties of the clad¬ 
ding, the oxidized cladding, and U0 2 to their respective melting points. 

During the period covered by this report, CY-67, the properties of stainless steel clad¬ 
ding material were also investigated. Since recent reactor designs have employed Zircaloy-4 
exclusively for all claddings, spacers, and coolant channels, the effort on stainless steel 
property measurements is being phased out of the program. 

The behavior of zirconium-alloy-clad and stainless-steel-clad U0 2 fuel elements during 
temperature excursions to the melted condition in the presence of steam or steam plus air 
has not been studied sufficiently to permit suitable hazard analyses when loss of coolant 
occurs. This program is being conducted to investigate the reactions that occur during melt¬ 
down in these atmospheres at heating rates which approximate those likely to be encountered 
in nuclear afterheat temperature excursions when the coolant has been lost. These investiga¬ 
tions include the following studies: 

1. Dynamic testing to determine the behavior of clad U0 2 when oxidized by steam, re¬ 
actions between the various reactor core components, internal pressure effects, hy¬ 
drogen release by the metal - steam reactions, and flow aspects of the molten con¬ 
stituents. 

2. Measurements of the rates of metal - steam reactions, oxidation of U0 2 by steam, 
and reactions between core constituents. 

3. Property measurements, including thermal properties of U0 2 , Zircaloy-4, and Type 
304 stainless steel, and the mechanical properties of Zircaloy-4 and Type 304 stain¬ 
less steel. 

8.1 DYNAMIC TESTING OF ZIRCONIUM-BASE ALLOYS (K. M. Emmerich and E. F. Juenke) 
INTERNAL PRESSURE EFFECTS 

The effects of internal gas pressure and temperature on the deformation behavior of Zirca¬ 
loy-4 cladding material during a thermal surge were investigated using the pressure test fa¬ 
cility shown schematically in Figure 8.1. Figure 8. 2c shows a typical tube-sample prepared 
for testing. The composition and characterization of the Zircaloy-4 material used in these 
experiments meet LOFT specifications reported previously. 1 Details of the experimental pro¬ 
cedure have been reported. 2 

^Project leader and principal investigator. 

1 "Fifth Annual Report - High-Temperature Materials Program, Part A," GE-NMPO, GEMP-400A, February 28, 1966, p. 205. 

2 „ 

AEC Fuels and Materials Development Program Progress Report No. 71/' GE-NMPO, GEMP-1002, December 29, 1967, p. 123. 




317 


Vacuum pump 



Fig. 8.1 - Schematic diagram of pressure test circuit 



Fig. 8.2 — Typical pressure-test samples of Zircaloy-4 before and 

after testing at 1.11 kg/mm 2 . Sample length is 5.7 cm. 

(Neg. P67-11-1) 

a. Heated at 0.3°C per second to 668°C. 

b. Heated at 20°C per second to 753°C. 

c. Unheated. 

In this series of tests, heating rates of 20°C and 0. 3°C per second and internal helium 
pressures from 0. 024 to 1. 21 kg/mm 2 were used. The tubes were heated at a constant rate 
under fixed pressure in an argon atmosphere until 34 percent diametral expansion occurred, 
approximately the expansion which would result in major blockage of coolant channels in a 
typical reactor. Results for this series of tests are given in Table 8.1 which lists final inter¬ 
nal pressure, surface (maximum) temperature when the sample had expanded 34 percent dia- 
metrally, tensile strength of the material at the maximum temperature, and calculated ef¬ 
fective stress. Two typical samples tested at the same internal pressure (~1.11 kg/mm2) are 
compared with an untested sample in Figure 8.2. The 34 percent diametral expansion was 









318 


reached at 668°C for sample A, heated at 0.3°C per second; and at 753°C for sample B, heated 
at 20°C-per second. 

Tensile tests at various constant-load levels were conducted on Zircaloy-4 for compari¬ 
son with the internal pressure tests. In these experiments, sheet specimens cut parallel to 
the rolling direction were heated from room temperature to rupture temperature at rates of 
approximately 0. 3°, 3°, and 20°C per second. Data are presented in Table 8. 2. 

In both the pressurized-tube tests and the constant-load tensile tests, the temperature at¬ 
tained at failure (34% expansion in the case of the pressurized tubes or rupture of the ten¬ 
sile specimens) showed heating rate dependence. At fixed pressures or load levels, higher 
maximum temperatures were obtained for the higher heating rates in all instances. 

TABLE 8.1 


EFFECTS OF INTERNAL PRESSURE ON ZiRCALOY-4 TUBING HEATED IN ARGON 


Heating 

Rate, 

°C/sec 

Internal Pressure 
(Gage), 
kg/mm 2 

Maximum 

Temperature, 

°C 

Calculated 
Effective Stress, 3 
kg/mm 2 

Tensile Strength, at 
Maximum Temperature, 

kg/mm^ 

0.3 

1.125 

666 

8.114 

6.9 

0.3 

1.118 

668 

8.065 

6.9 

0.3 

0.721 

723 

5.205 

4.6 

0.3 

0.541 

758 

3.901 

3.7 

0.3 

0.214 

860 

1.550 

1.8 

0.3 

0.144 

915 

1.040 

1.4 

0.3 

0.065 

924 

0.469 

1.2 

0.3 

0.065 

933 

0.469 

1.15 

0.3 

0.074 

953 

0.534 

1.0 

0.3 

0.037 

1020 

0.267 

0.68 

0.3 

0.036 

1027 

0.260 

0.68 

0.3 

0.025 

1187 

0.181 

0.38 

23.1 

1.209 

737 

8.727 

4.0 

22.8 

1.055 

753 

7.615 

3.8 

20.9 

0.703 

820 

5.074 

2.5 

22.0 

0.352 

881 

2.541 

1.6 

21.9 

0.352 

886 

2.541 

1.6 

20.4 

0.141 

955 

1.031 

0.99 

21.5 

0.141 

1003 

1.013 

0.74 

22.3 

0.141 

1024 

1.013 

0.68 

22.6 

0.037 

1165 

0.263 

0.39 

20.1 

0.037 

1282 

0.266 

0.29 

aBased on pre-test tube dimensions. 


TABLE 8.2 

CONSTANT LOAD CONTROLLED HEATING-RATE TEST 

RESULTS FOR ZIRCALOY-4 a IN ARGON 






DPH Hardness 




Rupture 


(2-1/2 kg) 


Heating Rate, 

Initial Stress, 

Temperature, 


Before 

After 

Sample No. 

°C/sec 

kg/mm2 

°C 

Elongation, % 

Test 

Test 

22 

0.3 

1.195 

945 

68 

209 

245 

18 

3.0 

6.327 

765 

48 

220 

215 

19 

3.0 

12.654 

665 

28 

225 

206 

17 

3.0 

3.164 

900 

54 

224 

218 

15 

3.0 

1.195 

1020 

81 

198 

223.5 

16 

3.0 

0.504 

1120 

102 

221 

206 

21 

20.0 

1.195 

1120 

69 

227 

218 

23 

20.0 

0.504 

1235 

102 

215 

200 


specimens were 0.076 cm thick, 0.63 cm wide, with a gage length of 2.54 cm. 



319 


Data obtained in the two series of tests were analyzed to determine the inter-relation¬ 
ship of applied stress, rate of uniform heating, total strain, and maximum temperature. 
This was accomplished by examining the theoretical solutions for a uniaxial test under 
constant load and a linearly increasing temperature, and a test of a thin-walled tube 
with constant internal pressure and linearly increasing temperature under the following 
assumptions: 

1. Deformations are primarily those due to second-stage creep. 

2. Second-stage creep behavior can be represented by an equation of the form 

— = Ae "Q/R T o n (8.1) 

dt 

where 

e - strain 
t = time 

T= absolute temperature 
R= gas constant 
Q = activation energy for creep 
a = effective stress * 

A and n = material parameters. 


3. Deformations are small. 

4. For the case of the tube, the radial stress is negligible, i.e., a r = 0. 

Under these assumptions, and letting T = j3t, the governing equations are: for the uniaxial 
case, 


and for the tube case, 


where 


de x = Ae-Q/ R ^a n dt 

dee^Ae-Q/R^andt 

Li e 



p = pressure 
p = inside radius 
h = wall thickness 
eg = hoop (tangential) strain 
e x = axial strain 


( 8 . 2 ) 


(8.3) 


Letting Q/RT = z and integrating these equations [(8. 2) and (8. 3)], gives the general 
result 


c = 


a S n Q 
/3R 


-z 


{-Ei(-z)} 


z = 00 


J z = Z i 


(8.4) 


where, for the uniaxial case and the tube case, a = A and /372 A, S = cr and a Q , and 
e = e x and €q , respectively. 


*Based on the von Mises flow criterion, the effective stress for the multi-axial (tube) case is 

0 e = [(<>! - 0 2 ) 2 + (o 2 -®3l 2 + < CT 3 - cr,) 2 ! 0 - 5 . 

where o- j, c^, <*3 are the principal stresses which, for the case of the tubes, are the radial (a r ), hoop {oq), and axial (a x ) stresses. 



320 


The exponential integral function may be approximated for zj > 10 by 


-z. 


-Ei(-z) = e * 


Z 1 


1 2 ! 
2 + 3 
Z 1 Z 1 


3J, 

"4 

Z 1 


(8.5) 


Since Q is the order of 40 T m cal/mole, z^ will always be much greater than 10. There¬ 
fore, neglecting all but the first two terms of the expansion, equation (8.4) becomes: 


a S n RT 2 -Q/RT 
i3Q e 

Taking the logarithms of both sides and rearranging terms gives: 


In S = — 
n 


aT? Q 

In e - log 7 ^-+In j3 - 2 In T + ^ 
Q RT 


( 8 . 6 ) 


(8.7) 


The exponent n for stress dependence and the activation energy Q are essentially con¬ 
stant for temperatures above about one half the absolute melting point (788°C for Zircaloy-4). 


Logarithmic (true) strain, In (1 + e), is more valid in this case than engineering strain 
because the expansion cannot be considered small. Accordingly, the combined data in Tables 
8.1 and 8.2 for temperatures above 788°C were fitted to an equation of the form of (8.6), 
giving the final equation with a standard deviation of ±40 percent: 


S = 0.0113 k exp 


/ 78,470 ± 8900 \ / re i\ 1 ^ 3 ' 96 
V 3.96 RT / \ T 2 / 


( 8 . 8 ) 


where k = 1 for tensile tests and (V’ 372 ) 1 / 2 ’®® for pressurized tube tests, and 

S = stress, kg/mm 2 
R = gas constant, 1.987 cal/mole-°K 
T = temperature, °K 
r = heating rate, °K/sec 
= logarithmic strain = ln(l + e). 

The logarithm of stress versus 1/T for this equation is shown in Figures 8.3 and 8.4 for 
the several heating rates used in the tests described above. Experimental data for all tem¬ 
peratures are also plotted on these figures. A stress dependence (exponent n) of 3.96 and 
an activation energy (Q) of 78.5 kcal/mole for creep in Zircaloy-4 are indicated by this 
analysis. The value of n agrees well with the expected value of 4.0, 3 and the value of the 
activation energy agrees with the estimated value of 81 kcal/mole for zirconium self¬ 
diffusion, using the approximation Q = 38T m , where T m is the melting point in °K. 4 

This analysis does not consider the effect of the a-fi zirconium phase transformation 
which may occur in the temperature range of 900° to 1000°C in Zircaloy-4. Tensile data 
appear to show a difference in behavior above and below this range, but separate analyses 
of the two temperature regions are of dubious validity, due to the limited number of data 
points. The pressurized tube experiments showed no difference in behavior in the two 
temperature regions. 


EFFECT OF OXIDATION ON TUBE DEFORMATION 

The effects of oxidation on the deformation behavior of Zircaloy-4 tubing were studied 
under dynamic heating conditions. Initial tests were made with 0.3°C/sec heating rates 
and at various internal pressures of 0.040 to 0.141 kg/mm^ to study the effect of steam 

o 

J- M. Dorn, editor, The Mechanical Behavior of Materials at Elevated Temperatures, McGraw-Hill, 1961, p. 95. 

4 lbid., p. 88. 



Temperature, °C Temperature, 


321 




















322 


atmosphere. Results of these measurements are shown in Table 8.3 with the results of 
comparison tests in argon. These data show that in a steam atmosphere failure of par¬ 
tially oxidized Zircaloy-4 tubes due to internal pressure will occur at temperatures 
higher than predicted from the base metal strength data. Typical fracture failure for a 
tube heated in steam is compared in Figure 8.5 to a sample heated in argon. 


TABLE 8.3 


EFFECTS OF INTERNAL PRESSURE ON ZIRCALOY-4 
TUBING HEATED AT 0.3°C PER SECOND 



Internal 

Maximum 

Diametral 


Gage Pressure, 

Temperature, 

Expansion, 

Atmosphere 

kg/mm 2 

°C 

% 

Argon 

0.040 

1180 

33 

Steam 

0.040 

1475 

12.5 a 

Argon 

0.057 

1055 

34 

Steam 

0.058 

1460 

11.4 a 

Argon 

0.078 

970 

34 

Steam 

0.077 

1035 

18.6 a 

Argon 

0.141 

905 

34 

Steam 

0.141 

905 

33 


a Failed by fracture. 



Fig. 8.5 - Effect of atmosphere on the failure of Zircaloy-4 tubes heated 

at 0.3°C/sec with 0.040 kg/mm 2 internal pressure (Neg. 68-1-42) 



323 


The effects of atmosphere on the expansion rate of Zircaloy-4 tubes at constant tem¬ 
perature and constant internal pressure were studied in argon and in steam by measur¬ 
ing the rate of diametral expansion with a cathetometer. Test temperatures selected for 
a given pressure were about 50° C below the maximum temperature achieved in the argon 
atmosphere dynamic heating tests at 0.3°C per second. Two comparisons have been made 
between the behavior of Zircaloy-4 in argon and in steam. At 0.131 kg/mm 2 (195 psig), 
a Zircaloy-4 tube heated at 860°C required 260 minutes for 34 percent expansion in steam, 
but only 74 minutes in argon. At 995°C and 0.051 kg/mm 2 (72 psig), 34 percent expansion 
required 106 minutes in steam, compared to 20 minutes in argon. Figure 8.6 shows the 
post-test appearance of these latter two samples. 


These tests show that (1) under the dynamic conditions of a thermal surge, oxidation 
permits the Zircaloy cladding to achieve higher temperatures before failure, and (2) at 
constant temperature and pressure, the rate of expansion is retarded by oxygen absorp¬ 
tion in the Zircaloy. Figure 8.6 shows that the presence of steam results in a more uni¬ 
form expansion of the tube. Steam oxidation, like strain hardening, results in more uni¬ 
form expansion. 



a. Heated in argon required 20 minutes for 34% expansion (Neg. P68-1-34A) 



b. Heated in steam required 106 minutes for 34% expansion (Neg. P68-1-34B) 


Fig. 8.6 — Zircaloy-4 tubes heated at 995°C with an internal 
pressure of 0.051 kg/mm^ 



324 


TESTING OF 50-cm-LONG ZIRCALOY-4-CLAD U0 2 FUEL ELEMENTS 

A furnace capable of heating a 50-cm-long Zircaloy-4-clad U0 2 sample in steam was 
constructed for simulating the axial thermal gradient imposed upon a fuel element in a 
reactor loss-of-coolant accident. This equipment is being used to determine whether 
there are corrosion rate effects in Zircaloy-4 cladding due to a hydrogen boundary layer 
buildup, and to establish effects of internal pressure, conditions under which cladding 
rupture occurs with loss of internal pressure, and behavior under meltdown. Figure 8.7 
is a schematic drawing of the furnace; construction and operation details were reported 
previously. 5 Figure 8.8 shows atypical 50-cm-long Zircaloy-4-clad UO 2 sample assem¬ 
bled for test. 



Fig. 8.7 — Schematic diagram of furnace for simulating reactor thermal and 
atmospheric environment during a loss-of-coolant accident 


5 * 


A EC Fuels and Materials Development Program Progress Report No. 69," GE-NMPO, GEMP-69, September 30, 1967, p. 139. 



325 




326 


Oxidation tests were run to study thickness of the oxide layers formed on 50-cm-long 
Zircaloy-4-clad U0 2 samples heated at rates of 0.3°C and 1.0°C per second in steam 
flowing at the rate of 3.4 standard liters (6 x 10' 3 lb) per minute. The oxide thickness 
at the midpoint of the sample and at 7.6 cm from each end were determined metallo- 
graphically and are summarized in Table 8.4. Circumferentially the thickness varied 
less than 50 percent. Longitudinally the oxide thickness at the midpoint, the hottest 
portion of the specimen, was from 1.5 to 2.5 times the thickness at the ends of the tube. 
For these tests, no difference was observed between the oxide thicknesses at the two 
ends. Consequently, assuming laminar flow under the test conditions, there was no evi¬ 
dence of a hydrogen-rich boundary layer at the metal - gas interface. 


TABLE 8.4 

VARIATION OF OXIDE THICKNESS ON 50-cm-LONG 
ZIRCALOY-4-CLAD U0 2 ELEMENTS HEATED IN STEAM 


Heating 


Rate, 

°C/sec 

Maximum Temperature, °C 
Top Middle Bottom 

Oxide Thickness, microns 3 
Top b Middle Bottom 

0.3 

786 

900 

776 

<13 

13 

<13 

0.3 

934 

1098 

922 

19 

32 

16 

0.3 

c 

1313 

1222 

25 

86 

60 

0.3 

c 

1500 

1430 

117 

200 

102 

1.0 

c 

1100 

923 

13 

19 

13 

1.0 

c 

1300 

1170 

32 

70 

28 

1.0 

c 

1510 

1475 

66 

140 

91 

1.0 

c 

1510 

1300 

73 

133 

60 


a Average of four circumferential measurements. 
b Steam entered furnace from the top. 

c Top thermocouple failed before maximum temperature was attained. 


DYNAMIC TESTING OF IRRADIATED ZIRCALOY-2 AT CONSTANT INTERNAL PRESSURE 

A test fixture for pressure testing irradiated Zircaloy-2 tubing at various heating rates 
is being constructed. Test pieces will be Zircaloy-2 and fuel rod cladding tubes from the 
spent fuel of an industrial reactor from which UO 2 has been removed. Tubes will be cleaned 
and cut to 8-inch lengths. The material is highly radioactive and must be handled remotely; 
equipment is being designed with the simplified assembly and disassembly features shown ’ 
in Figure 8.9. The Zircaloy tube will be pressurized with helium from a line attached to 
the bottom of the fixture. The top and bottom of the tube will be made pressure tight with 
seals. The thermocouple will be attached in the form of a loop which is drawn 
against the sample by a constant weight. This permits about 1 cm of the thermocouple 
wire from the bead to contact the tubing, thus minimizing errors in temperature meas¬ 
urement arising from thermal losses due to heat transmission along the wires. 

8. 2 REACTION MECHANISMS AND KINETICS (J. T. Bittel and L. H. Sjodahl) 

OXIDATION OF ZIRCONIUM ALLOYS BY STEAM OR AIR 

Effects of the variation of steam flow rate and oxygen partial pressure on the oxidation 
kinetics of zirconium alloys were determined to 1585°C. Studies of Zircaloy-4 in steam 
were also extended to 1835°C. 

Steam Oxidation 

A thermogravimetric technique 6 was used to determine the oxidation kinetics of Zircaloy-4 
in steam at various flow rates. Parabolic rate constants determined from these tests are 

0 

R. E. Latta, J. T. Bittel, and G. B. Hadesty, "Continuous Recording Strain Gauge Thermobalance," Rev. of Sci. Instr. Vol. 38 
No. 11, November 1967, p. 1667. 



327 



Fig. 8.9 — Furnace assembly for internal pressure 
tests on irradiated Zircaloy-2 tubing 

shown in Table 8.5 and plotted as log rate versus reciprocal temperature in Figure 8.10, 
along with the curve determined previously for steam oxidation of zirconium alloys. 7 Re¬ 
sults of the steam tests at various flow rates agree with previously reported corrosion 
data using lower flow rates; this indicates that the steam oxidation rate of Zircaloy-4 is 
independent of the steam flow rate in the range examined. 

Although zirconium would not be expected to hydride appreciably at the temperatures of 
the steam corrosion tests, Lemmon 8 reported that part of the hydrogen which formed as 
a result of the Zr-I^O reaction was found in tested Zircaloy-2 samples; 12.5 percent of 
the hydrogen formed in a 1600°C test and 6.8 percent in a 1000°C test. Hence hydrogen 
content of Zircaloy-4 samples tested in steam at 1200°C and at 1600°C was determined 
by hot extraction analysis. Two samples tested at 1200° C showed 40 and 332 ppm; the 
1600°C test sample showed 98 ppm. The highest concentration of hydrogen found, 332 
ppm or 1.3 percent of the hydrogen produced, is equivalent to less than 0.5-mg weight 
g ain in this sample. This would have an insignificant effect (0.16%) on the total weight 
gain. The greater amounts of hydrogen reported by Lemmon may result from his testing 
in steam at a total pressure of 50 psi (0.035 kg/mm 2 ) or from the fact that Zircaloy-2 
absorbs hydrogen more readily than Zircaloy-4. 

Tests on the steam oxidation of Zircaloy-4 were extended to 1835°C using solid cylindri¬ 
cal samples in an inductively heated furnace. Experimental details have been reported 
previously. 9 A spectrographic analysis of the Zircaloy used in the present tests is given 
in Table 8.6. 


7 "Sixth Annual Report - High-Temperature Materials Program, Part A," GE-NMPO, GEMP-475A, March 31,1967, pp.240-245. 

®A. W. Lemmon, "Studies Relating to the Reaction Between Zirconium and Water at High Temperatures," BMI-1154, 1957. 

9 GEMP-1002, p. 134. 







Temperature, °C 

1900 1800 1700 1600 1500 1400 1300 1200 1100 



Fig. 8.10 — Oxidation of zirconium alloys by steam and air 


TABLE 8.5 

RESULTS OF STEAM AND AIR CORROSION 


OF ZIRCONIUM ALLOYS 


Sample 

Material 

Gas 

Composition 

Gas Flow 
Rate, std. 
ml/min. 

Temperature, 

°C 

Parabolic Rate 
Constant, 

(mg C^^/crrAsec 

Zr-4 

Steam 

300 

1200 

0.63 

Zr-4 

Steam 

800 

1500 

5.96 

Zr-4 

Steam 

2400 

1500 

8.3 

Zr-4 

Air 3 

300 

1200 

2.99 

Zr-2 

Air 3 

300 

1200 

3.17 

Zr-2 

Air 3 

300 

1400 

9.87 

Zr-4 

Air 3 

300 

1500 

12.2 

Zr-4 

Air 3 

300 

1585 

29.5 

Zr b 

°2 

— 

1100 

0.158 

Zr b 

0 2 

- 

1200 

0.415 

Zr b 

0 2 

- 

1290 

0.800 


a Dewpoint - 66°C. 
b Pemsler, gas flow rate unknown. 




329 


Test results from the steam oxidation of Zircaloy-4 in this high-temperature range 
are shown in Table 8.7 and Figure 8.11. Although these results fall nearly within the 
95 percent confidence limits calculated from the lower-temperature data, 10 they are 
about 36 to 70 percent higher than the extrapolated low-temperature Arrhenius curve. 
This indicates a possible change in controlling mechanism to one with higher activation 
energy for these higher temperatures. Fitting the In W^/1 for the four data points above 
1600° C, the previously measured data at 1600° C and 1615° C, 10 and the datum point of 
Baker 11 for 1852°C to a linear equation by least squares, yields: 

2/ - ...h /-60,600 ± 5,400 \ , a Q . 

W z /t = 1.5 x 10 8 exp I- ’ RT - 1 (8. 9) 


where 

W = weight gain per unit area, mg/cm^ 
t = time, sec 

R = gas constant, cal/mole-°K 
T = temperature, °K 

This can be compared to the expression derived for the least-squares fit to the data for 
temperatures between 1000° and 1600° C 10 : 

o, r „ _5 /-39,700 ± 2300\ , 0 

VT/t = 5.56 x 10° exp I- 2 ——-) (8.10) 

TABLE 8.6 


COMPOSITION OF ZIRCALOY-4 a 
USED IN OXIDATION STUDIES 


Element 

Analysis 

percent 

ppm 

Sn 

1.45 

- 

Fe 

0.186 

- 

Cr 

0.108 

- 

Fe+Cr+Ni 

0.294 

- 

Al 

- 

67.0 

B 

- 

0.3 

C 

- 

186.0 

Cd 

- 

< 0.25 

Co 

- 

< 5.0 


Cu 

Hf 

Mg 

Mn 


< 

< 

< 

20.0 

60.0 

10.0 

20.0 

TABLE 8.7 

RESULTS OF STEAM CORROSION OF ZIRCALOY-4 
BETWEEN 1500° and 1800°C 

Mo 


< 

20.0 






Na 


< 

10.0 





Parabolic Rate Constant 

Ni 


< 

20.0 

Temperature, 

Time, 

Weight 

Area, 

(Kp), 

Pb 


< 

20.0 

oc 

min 

Gain, g 

cm2 

(mg 0) 2 /cm4-sec 

Si 



31.0 






Ti 



20.0 

1540 

12 

0.9827 

11.09 

10.9 

V 


< 

20.0 

1700 

10 

1.3687 

10.83 

26.6 

W 


< 

20.0 

1733 

10 

1.8820 

11.19 

47.2 

u 

- 

< 

0.2 

1742 

10 

1.9041 

10.92 

50.0 

Zr 



Bal 

1800 

10 

1.9437 

10.77 

54.3 


a lngot analysis provided by vendor. 


10 GEMP-475A, pp. 240-242. 

11 L. Baker and L. C. Just, "Studies of Metal-Water Reactions at High Temperatures, III. Experimental and Theoretical Studies 
of the Zirconium-Water Reaction/' ANL-6548, p. 1962. 



330 


Temperature, °C 

1800 1700 1600 1500 1400 1300 1200 1100 



Fig. 8.11 — Parabolic rate constant versus reciprocal temperature for 
oxidation of Zircaloy-4 by steam. 

The oxidation of zirconium and Zircaloy proceeds by a combination of oxygen solution 
in the metal and growth of an oxide coating. Debuigne, et al. , 12 from investigation of zir¬ 
conium in oxygen, concluded that diffusion of oxygen in the alpha metal was the rate- 
controlling step in the temperature range from 668° to 1200°C and that the process is 
parabolic but has exponents slightly greater than 2. Oxidation rates calculated from his 
data using an exponent of 2 agree with those of Pemsler 13 for zirconium heated in oxygen. 
Although the crystal structure of the oxide films on zirconium are monoclinic at room 
temperature, the work of Ruh and Garrett on the system Zr-O 14 predicts that the crystal 

12 J. Debuigne, J. P. Guerlet, and P. Lehr, "Aspects Fundementaux des Phenomenes D'oxydation du Zirconium sous Oxygene Pur," 
Etude Sur La Corrosion et la Protection Du Zirconium et de Ses Alliages, Presses Universitaires de France, Paris, 1966. 

^J. P. Pemsler, "Studies on the Oxygen Gradients in Oxidizing Metals, V. The Oxidation of Oxygen Saturated Zirconium," 

J. Electrochem. Soc., 1966, pp. 1241—1244. 

^R. Ruh and H. J. Garrett, "Nonstoichiometry of Zr02 and Its Relation to Tetragonal-Cubic Inversion in Zr02,” J. Am. Ceram. 
Soc., Vol. 50, No. 5, May 1967, p. 257. 



331 


structure is tetragonal above about 1100°C and that it is cubic at temperatures as low as 
about 1500°C. Klepfer 15 has suggested that since oxygen may diffuse at different rates 
through the different oxide crystallographic forms, the activation energies for the oxida¬ 
tion process may be different also. If so, the higher rates measured in this present work 
may be related to the change in crystallographic structures of the oxide film. 

Air Oxidation 

Corrosion tests in dry (-66°C dewpoint) air flowing at 300 std ml/min were conducted 
with Zircaloy-4 at 1200°, 1500°, and 1585°C and, for comparison, with Zircaloy-2 at 
1200°C and 1400°C. Results of these tests are given in Tables 8.5 and 8.10. Also given, 
for comparison, are data on steam corrosion at 1200° C and 1500°C, and Pemsler's data 
on oxygen oxidation of unalloyed zirconium at 1100°, 1200°, and 1290°C. These data show 
that Zircaloy corrosion in air is more erratic and two to five times as rapid as in steam. 

Figure 8.12 shows the microstructure of Zircaloy-4 corroded in air at 1200°C. A sample 
corroded in steam at the same temperature is shown in Figure 8.13. The coating on the 
air-tested sample varied greatly in thickness and density around the periphery of a cross 
section. Figure 8.12a and b show, respectively, the relatively thin, dense reacted layer 
and the thicker, more porous reacted layer which contains up to 10 percent crescent- 
shaped separations or large voids. Below the dense areas of coating on the air-tested 
samples is an oxygen-rich alpha zirconium zone similar to that seen in steam corrosion. 
Below the more porous oxide is a porous yellow phase identified by X-ray diffraction as 
zirconium nitride. The oxygen-stabilized alpha-phase metal layer beneath the yellow 
phase is very thin compared to that observed in the microstructure of steam-oxidized 
metal. Table 8.8 records measurements of the total sample diameter and of thicknesses 
of the coating and of the oxygen-stabilized, alpha-phase metal formed at 1200°C in both 
atmospheres. 


TABLE 8.8 

ZIRCALOY-4 OXIDIZED FOR 1 HOUR AT 1200°C 
IN AIR AND IN STEAM 


Atmosphere 

Coating Thickness, 
microns 

Alpha-Phase 
Metal Thickness, 
microns 

Original 

Diameter, 

mm 

Post-Test 

Diameter, 

mm 

Air 

356 to 1016 

0 to 200 

12.6 

13.0 

Steam 

200 

180 

12.6 

12.8 


The higher rates for air corrosion relative to steam oxidation apparently result from 
the formation of zirconium nitride and its effect on the coating structure. Nelson 16 reports 
that zirconium reacts more rapidly with air than with oxygen, producing a mixture of the 
nitride, oxide, and oxynitride of zirconium. Since Pemsler's results 17 for zirconium in 
oxygen show even lower rates than those reported here for steam, the higher corrosion 
rates in air do not result from higher oxygen partial pressure in air, but rather from 
the presence of nitrogen. 


1 R 

H. H. Klepfer, private communication. 

^L. s. Nelson, “Combustion of Zirconium Droplets Ignited by Flash Heating/' Western States Section, Combustion Institute, 
Paper WSS/CI 64-23, 1964. 

17 


Pemsler, loc. cit. 






Reaction layer 
(Zr0 2 ) 


a zone 


Zr-4 

metal 


a. Dense area of reaction layer (Neg. 9029) 


Zr nitride 
and ZrC >2 


Zr-4 

metal 


b. Porous reaction layer (Neg. 8889) 


Fig. 8.12 — Photomicrographs of Zircaloy-4 tested at 1200°C 
for 1 hour in air (As-polished, 100X) 







333 


Fig. 8.13 — Photomicrograph of Zircaloy-4 tested at 1200°C for 
1 hour in steam (Neg. 8890, as-polished, 100X) 

OXIDATION OF TYPE 304 STAINLESS STEEL IN STEAM OR AIR 
Steam Oxidation 

In the previous annual report, 18 results of studies of the oxidation kinetics of Type 304 
stainless steel in steam at temperatures between 1000° and 1350°C were presented. 

Scales which formed at 1200°C and above remained in place upon cooling although they 
cracked; those formed at lower temperatures tended to spall when cooled. The micro¬ 
structure of the oxidized surface formed at the higher temperatures showed a multi¬ 
layer coating. The scales, formed on the surface of 304L stainless steel samples oxi¬ 
dized in steam at 1200°C and 1300°C, have been studied by microprobe analysis. 

These microprobe analyses previously discussed 19 in detail are summarized in Table 8.9. 
Both samples contain high-chromium oxide inner layers of a spinel structure, as revealed 
by X-ray diffraction analysis. Similar high concentrations of chromium as Fe-Cr spinel 
in the inner oxide layer have been observed by Fujii and Meussner 20 in studies of the oxi¬ 
dation of Fe-Cr alloys in argon plus steam at temperatures between 700° and 1100°C. 

Previous studies of steam corrosion kinetics of 304 stainless steel at temperatures be¬ 
tween 1000° and 1375°C showed that an initial linear oxidation rate (for 6 to 28 minutes) pre¬ 
ceded the parabolic behavior. 21 The parabolic rate was represented by the equation 

W 2 /t = 2.4 x 10 12 exp ( - 84 > 3 °^ 2400 \ (8.11) 

18 GEMP-475A, pp. 246-251. 

19 GEMP-69,pp. 144-147. 

20 n 

C. T. Fujii and R. A. Meussner, High Temperature Oxidation of Iron-Chromium Binary Alloys in Water Vapor, Part 1/' 

U.S. Naval Research Laboratory, NRL Report 5506, September 21, 1960. 

21 



GEMP-475A, p. 246. 




334 


TABLE 8.9 

DESCRIPTION OF OXIDE LAVERS FORMED ON STAINLESS STEEL 
DURING OXIDATION IN STEAM AT 1200°CAND 1300°C 


Temperature, 

°C 

Oxide 

Layer 

Electron Probe 
Analyses, 8 
wt % 

Fe Cr Ni 

Metallography of Oxide 

1200 

Outer Oxide 

75 

- 

2 

Two phases with large grains 
and large voids. 


Inner: Oxide 

45 

27 

4 

Small grains and voids 


Metal 

47 

10 

43 

plus metal. 

1300 

Outer: Oxide 

77 

- 

_b 

Two phases, large oxide grains. 


Metal 

30 

- 

77 c 

large voids, metallic inclusions. 


Inner: Oxide 

47 

22 

7 

Two phases, small grains and 


Metal 

40 

5 

60° 

voids, metallic particle. 


a Average values for weight percent metal at center of oxide layers. 
Approximately 2% Mn concentrated near surface of oxide layer. 
c Although corrected for absorption and fluorescence, the calculated compositions 
totaled more than 100% in the metal phase. This does not affect the conclusions 
drawn in the text. 


Recent analysis of the linear portion of the data yields a least-squares fit given by the 
equation 

, 1 in5 /-44,350 ± 230()\ ,„ 

W/t = 1.1 x 10° exp (- RT - I. (8.12) 

The linear rates for steam oxidation are plotted in Figure 8.14. The results of Higgins 22 
for steam oxidation of Type 321 stainless steel at 1260° C are included for comparison. 

Air Oxidation 

To compare the behavior of 304L stainless steel in air with its oxidation in steam, 
tests were run in flowing dry air (-66°C dewpoint) between 1100° and 1375°C using the 
thermogravimetric technique. Although weight gains in air at 1360°C and below were 
too small to accurately determine rate behavior, parabolic behavior is assumed because 
the low rates imply protective film formation. Parabolic rate constants, calculated for 
comparison with the parabolic and linear portions of the steam oxidation of 304L stainless 
steel, are included in Table 8.10. 

The logarithms of parabolic rate constants in air are plotted in Figure 8.15 together 
with those for steam oxidation. The rate in air is less than that in steam by a factor of 
about 10 3 . Oxidation behavior of 304 stainless steel in air including the point at 1100°C may 
be represented by Hagel's 23 curve for chromium oxidation in air, which is included in 
Figure 8.15. Data of Caplan and Cohen 24 for 302 stainless steel in dry air, which show 
good agreement with the data of Hagel 23 for chromium, are also included in Figure 8.15. 
Oxidation rates determined by Hagel 23 and by Gulbransen 25 for chromium in oxygen are 
shown in this figure; they differ only slightly from the rates in air. 

Tests of 304L stainless steel at 1375° C in air show oxidation rates comparable to those 
for steam at this temperature, much higher than the rates in air between 1100° and 1360°C. 

22 h 

H. M. Higgins, A Study of the Reaction of Metals and Water," AECD-3664, April 15, 1955. 

23 „ 

W. G. Hagel, Factors Controlling the High Temperature Oxidation of Chromium," ASTM Trans., Vol. 56, 1960, p. 583. 

24 D. Caplan and M. Cohen, "High Temperature Oxidation of Chromium-Nickel Steels," Corrosion, Vol. 15 1959 p. I4lt 

25 ' ' 

E. A. Gulbransen and K. F. Andrew, "Kinetics of the Oxidation of Chromium," J. Electrochem. Soc., Vol. 104,1957, p. 334. 


335 


Temperature, °C 



Fig. 8.14 — Linear rate constant versus temperature for 
steam oxidation of Type 304 stainless steel 


TABLE 8.10 

RESULTS OF OXIDATION OF 304L SS FROM 
1000° TO 1375°C IN STEAM AND IN AIR 3 
Rate Constants* 3 

Temperature, Linear, Parabolic, 

Atmosphere °C mg/cm 2 -sec mg 2 /cm 4 -sec 


Steam 


Air 


1000 

0.0030 

0.0095 

1050 

0.0039 

0.030 

1100 

0.0087 

0.095 

1150 

0.016 

0.18 

1150 

0.017 

- 

1200 

0.033 

0.75 

1200 

0.034 

0.80 

1250 

0.038 

2.1 

1300 

- 

4.4 

1350 

0.11 

13 

1375 

Not observed 

17 

1100 

— 

0.00025 

1200 

- 

0.00005 

1225 

- 

0.00012 

1350 

- 

0.0074 

1360 

- 

0.0079 

1375 

- 

21 

1375 

- 

25 


a Dry air (-66°C dewpoint). 

^For tests in steam, there was a change from linear to para¬ 
bolic behavior after times ranging from 6 to 28 minutes. 


The increased rate reflects the presence of a liquid phase, consistent with a temperature 
above the Fe-FeO solidus. 26 A sample after testing at this temperature is shown in Fig¬ 
ure 8.16, where it is compared with an unoxidized sample and a sample oxidized at 1360° C 
in air. The scale formed at 1360°C is thin and uniform compared to that at 1375°C which 
is much thicker, very irregular, and shows evidence of melting. X-ray diffraction re¬ 
vealed the composition of the scale surface formed in air at 1360° C and below to be princi¬ 
pally Fe^O^, with a minor amount of Fe20g. The surface of the scale formed at 1375°C 
was Fe 3 & 4 . When the poorly adherent outer layer was removed from a sample tested at 
1332°C,* however, a green oxide layer was exposed which was identified by X-ray diffrac¬ 
tion as rhombohedral CrgOg with less than 20 percent FegOg. 

Electron microprobe analysis of the sample oxidized in air at 1375°C showed a distribu¬ 
tion of the iron, nickel, and chromium in both oxide and metallic phases which was very 
similar to that observed in the sample oxidized in steam at 1300°C. A sample oxidized 
at 1310°C in air, however, showed a very high chromium concentration (48%) and only 
8 percent iron in the oxide layer attached to the metal substrate. This is consistent with 
the presence of a protective layer of Cr 2 0 3 and agrees with the data of Yearian. 27 

Scales formed in air at 1360° C and below generally spalled when the samples were 
cooled and removed from the furnace; this effect was not so noticeable on samples run 
in steam above 1200° C. 

*These samples were not used in weight gain analysis because much of the oxide layer was lost by spallation. 

26 L. S. Darkin and R. W. Gurry, "The System Iron - Oxygen, II: Equilibrium and Thermodynamics of Liquid Oxide and 
Other Phases," J. Am. Chem. Soc., Vol. 68, 1946, pp. 799-816. 

27 H. J. Yearian, H. E. Boren, Jr., and R. E. Warr, "Structure of Oxide Scales on Nickel-Chromium Steels," Corrosion, Vol. 12, 

1956, p. 5611. 




Temperature, °C 



Reciprocal absolute temperature, 10 4 /°K 


Fig. 8.15 — Parabolic rate constant versus temperature for steam and 
air oxidation of Type 304L stainless steel 



Unoxidized sample Oxidized at 1360°C for 100 minutes Oxidized at 1375°C for 4.4 minutes 


Fig. 8.16 — 304L stainless steel samples oxidized in air 








337 


A sample of mild steel was heated in air for 1 hour at 1280° C for comparison with the 
304L stainless steel behavior. The parabolic rate constant was 1.26 mg'Vcm 4 -sec com¬ 
pared to about 0.001 mg^/cm^-sec for 304L stainless steel under the same conditions. 

The unusually low oxidation rate of stainless steel in air compared to that in steam at 
1360°C and below is due to formation of a protective Cr 2 0 3 layer which forms in air but 
not in steam. Seybolt 28 has shown that at 1300°C the oxide of an 0.8Fe — 0.2Cr alloy is 
rhombohedral (Cr, Fe^Og only at oxygen partial pressures above 0.05 atmosphere, 
whereas spinel forms below 0.01 atmosphere. Since oxygen pressures in steam vary 
from about 2 x 10 -4 to 8.8 x 10 -6 atmospheres in the temperature range employed, the 
Fe-Cr spinel layer observed in the steam oxidation of 304L stainless steel is consistent 
with the results of Seybolt. 28 An increase of chromium in the oxide of alloys which have 
up to 0.85Cr, 0.15Fe, approximately the composition observed in the oxide formed on 
304L stainless steel by air below 1360°C, stabilizes the rhombohedral phase down to 
10 -4 atmospheres oxygen, accounting for the C^Og layer in the air-oxidized samples. 
Possibly because of reactions of Ci^Og with liquid Fe-FeO, no was observed at 

1375°C in air, and protection was no greater than that in steam. 


OXIDATION OF U0 2 BY STEAM 
Oxidation Kinetics 

The steam oxidation of U0 2 was investigated in the temperature range from 885° to 
1835°C. Data for the range from 1100° to 1500°C were presented previously. 29 Recent 
oxidation studies at 885°C and 1000°C were conducted using the thermobalance apparatus 
described earlier. 30 Investigations from 1600° to 1835°C were carried out in an induction- 
heated furnace with a dense Al 2 Og muffle to prevent oxidation of the molybdenum susceptor. 
Test specimens of U0 2 qo 2 or uo 2.004 were right circular cylinders sintered to about 95 
percent of theoretical density at 1700° C. The spectrographic analysis of a typical sample 
is given in Table 8.11. Sample dimensions were varied with oxidation temperature to com¬ 
pensate for the variation in oxidation rates. Details of the experimental procedures were 
given in earlier reports. 31 * 32 


Exact dimensions of each sample, experimental data, calculated rate constants, and 
calculated oxygen diffusion coefficients for all U0 2 steam oxidation tests are listed in 
Table 8.12. The parabolic rate constants are plotted in Figure 8.17. The least-squares 
equation for these rates is: 


,„2/. ,„4 /-48,000 ± 1900 ' 

W /t = 8.4 x 10* expl ——-. 


(8.13) 


Experiments using U0 2 of 98 percent theoretical density were run at 1390° C and 1520°C 
for comparison with the results of the tests using 95-percent-dense material. Data and 
results of these tests are included in Table 8.12 and Figure 8.17. The good agreement be¬ 
tween these data and the data from 95-percent-dense samples is predictable using the data 
of Belle, 33 which show no open porosity in U0 2 for densities greater than 91 percent. 


(J, Seybolt, "Observations on the Fe-Cr-0 System," J. Electrochem. Soc., Vol. 107, 1960, p. 147. 

29 GEMP-475A, pp. 251-256. 

39 Latta, Bittel, and Hadesty, loc. cit. 

•^"High-Temperature Materials Program Progress Report No. 63," GE-NMPO, GEMP-63, December 30,1966, p. 120. 
32 GEMP-1002, pp. 130-131. 

33 J. Belle, "Uranium Dioxide: Properties and Nuclear Applications," Naval Reactors, Div. of Reactor Development, USAEC, 
1961, p.333. 



TABLE 8.11 RESULTS OF U0 2 OXIDATION BY STEAM 


338 


r- C C O 

§ .2 .£ ® 

CD </> U 

££ E" 

O lt= ® E 

° Q o « 


ro 

’^tir)Lnir)ir)LOLr)CDir)<DcD<DcococDr^p^pvoooocn 

ooooooooooooooooooooo 

xxxxxxxxxxxxxxxxxxxxx 

(DOOONO«-oocotonra«-i-(-corawo^co 

^r-rOCNCOOCNltDCVOOCO^flNOOOCM^tOt-OlO 
*-* 00 LO CO CM* CO ^ 00 00 00 00 ID CM* CM Lf) st CO r-‘ CM 


ell 


in co co cm 

to IN 0 lO CM r- 

cmt— or^cocNCO^r cocj)OCDr^cMc>cor-ooo 
cnop^CNicNOcoocMr-.Lno^j-cN'^oooooo 

OOCD^COCMCOCO’—OO'—^OOOOOOOOO' 

ddddddddddddddodddddd 


Q) J3 CC 

O E 3 
w 3 < 
£ ,?0 


oocMiocoOr-T-cor^r^r^cococo-^oooor«.r^^)?j 

Locococor^r^r>.r^r^r>.c>-coooooooo)a)0o^<o 

T r ,_ . *" r ! ,,, ; r ‘'"'" r,T " N C N|CN1 CM 

CM CM CM cvi CM* CM* CM CM CM* CM* CM CM* CM CM* CM* CM* CM CM CM* CM* CM* 


O’^LOLnco^O'-^-CDCM’^-T-ocor^.^r^coOCT) 
or*.a)LnCT)r^or^>cj)Ocor^co»-LOuocococMLfico 
OOOCJOOt-^OOr-r-OT-i-OOOOOOO 
CM* CM CM CM CM CM* CM* CM* CM* CM* CM CM CM CM CM CM CM* CM CM* CM CM* 


.2» ~ OJ 
m CO 


*-ooo)cor-uor>«r^.coo)coi^.coo^05m*-CN o 

'tCMNOOOOOMNOCMWONlO^nCOOOOi- 
COONOCOLOt— OOOCMCOCMOOCQr-'CO»— t— CM 
0(D't'JM;0cof\jq'--r-ioT-i-r\jooqo^o 

ddddddddddddddodddddd 

t— 'd-oot-0)t—cMoor^.r>.r^.cocMT-o^-oocor>.ooo 
COCOCMCOCOOCOO>COOOr-COCOCT)0)inLOr««t~CDCM 
fMfMO)^OrsCJ0M-M^0O)M't^|scn^00 
lOO)COO)COCqiOO)^CM*-;a)qOOO)COlDa)inOOq 

cdodM- od'tcocdcoiocDidodcdcDtsddcDcdirid 

r-CDCMCM«—T— «— 


OT--5tcM^ooLOOococ-OLor^ooT-r^cor^LOCM 

OlOC)0)C)0)OJOONCM(NCnCMCMOCOC\ICM(MCMr* 

»—»— o»— »— «— t— "^T-idd 


0) Q) 

| I I 

”5 


t-OCDCOOCQ^r-OininCQ'tf'S-LOCOCOCOOOCOO 
0)000000000000101011)03101000)000(0100 
qOOO^OOr-.CMCMCMOCMCM(OCO^tCMCMCM*!t 
H CO CO ri ri P) CO N p* t- r- W r* r C\i t- r* r-’ t-’ r- 


OOOOOOOOOOOOOOOOOOOOO 

OOOOOOOO-^OCOOOOOCMOOCOOOO 

CDCDCOCOCMCMCOCMt-^LOCOr^O^OOOlOCO'^'^- 

^ r . r _ T _ r _ CN ^ tcolotDCN j r ^ OOOTd . r ^ 


CM O p 1 
O -D E 
D 1 « 


lOLOOlOLOOOOOOOOOOOOOOOOLO 
CO O CM r- CM»-OCMOOO*-OO0>OOOOO00 

oor^r^r^cocDcoioioioio^-^'d-cocMCMr-t-ooo 


£ U) LO CM CM in 

q. odioor^coiooddooor>oooi<OT-oooo 

a CD CD ^ V CM V V V '“'“C'lr-r-r-i-r-r-r-r-T-r- 

V v vv VVV V vvvv 


O) — ■OOL-D<D a5C 0.— _n (~ m 

O2oU-Q-<<00OOOOLJ_^22ZQ:C0CO|-a 


J 



339 


Temperature, oc 



5.0 6.0 7.0 8.0 9.0 


Reciprocal absolute temperature, 


Fig. 8.17 — Parabolic rate constant versus reciprocal temperature 
for oxidation of UO 2 by steam 


Oxygen Diffusion in UO ? 

Experimental — The controlling mechanism of steam oxidation of UO 2 is believed to be 
primarily interstitial bulk diffusion of oxygen. If oxidation of UO 2 is controlled by diffu¬ 
sion of oxygen through the crystal lattice, it follows that a composition gradient should 
exist in tested samples whose average composition is below that at equilibrium with the 
steam atmosphere. To test this hypothesis the outer surface, a layer approximately 80 
microns below the surface, and the center of a UO 2 sample oxidized in steam for 1 hour 
at 1400°C were studied by X-ray diffraction. Lattice parameters of the major phase 
(U02+ x ) from the center outward were 5.4705, 5.4695, and 5.4689, corresponding to 
O/U ratios of 2.003, 2.023, and 2.030. There was evidence in all areas of a second phase. 
Both the amount of the second phase and its oxygen content (estimated from lattice param¬ 
eter measurements) increased from the center to the surface. In the vicinity of the sur¬ 
face the second phase, with a composition of approximately t^Og, represented 30 percent 
of the material. From these data the overall surface region O/U ratio was estimated to 
be 2.03 x 0.7 + 2.25 x 0.3 = 2.10. An O/U ratio of 2.180 is calculated to be the O/U ratio 



340 


in equilibrium with steam at 1400° C. The entire sample of this test had a final average 
O/U ratio of 2.074 calculated from the weight gain. 

Portions of this sample taken from the center and from the surface regions were ana¬ 
lyzed chemically and O/U ratios of 2.072 in the interior and 2.084 toward the exterior 
were obtained, consistent with X-ray and weight gain analyses. 

The O/U ratio for various portions of a sample heated 10 minutes at 1715°C showed 
2.007 in the interior and 2.075 on the surface by X-ray analyses, and an overall average 
of 2.055 from the weight gain data; the O/U at equilibrium for this temperature is 2.166. 
Some redistribution of oxygen in the surface layer of the test sample occurs during the 
cooling cycle, but the discrepancy between the surface O/U and the equilibrium O/U value 
appears somewhat larger than can be explained on this basis. 

Overall results indicate that the oxidation is primarily controlled by diffusion of oxygen, 
but the difference between the equilibrium O/U ratios and that measured for the exterior 
indicates that the surface reaction rate may be low enough to affect the results. K so, 
the true diffusion coefficients may be slightly higher than the values reported below. 

Oxygen diffusion coefficients were calculated from the weight gain of the UO 2 tested 
in steam assuming a bulk diffusion mechanism and using the following equation derived 
by Jain 34 for right circular cylinders: 


v = 1 - 




(8.14) 


where 

v = fractional completion of the oxidation 
t x = Dt/x^ 

D = diffusion coefficient, cm/sec^ 
x = radius (p) or height (H), cm 
t = time, seconds 

Values of r were obtained using v calculated from the weight gain and the U02+ x composi¬ 
tion in equilibrium with steam at test temperature. The latter was calculated from basic 
thermodynamic data and appropriate relationships between UC> 2 +X composition and equili¬ 
brium oxygen partial pressure. Data of Anthony, et al. , 35 were used for temperatures of 
1100° C and above, and those of Aronson and Belle 36 were used for the lower temperatures. 


Figure 8.18 shows the results of these calculations plotted as log D versus 10 4 /T. The 
least-squares equation for the curve fitted to those data and previously reported data 37 
for the temperature range 1100° to 1600° C yields the following for the diffusion coefficient 
in cm2/sec. 


D chem = 151 ex P 


-58,000 ± 2000 
RT 


(8.15) 


34 S. C. Jain, "Simple Solutions of the Partial Differential Equation for Diffusion," Proc. Roy, Soc., A243,1957, pp. 359-374. 

35 ,, 

A. M. Anthony, R. Kiyoura, and T. Sata, Les Equilibres des Composes Oxygenes de {'Uranium entre 1500 et 2000°K," 

J. Nuc. Mat., Vol. 10, 1963, pp. 8-14. 

36 

S. Aronson and J. Belle, "Nonstoichiometry in Uranium Dioxide," J. Chem. Phys., Vol. 29, 1958, pp. 151-158. 

37 " 

AEC Fuels and Materials Development Program Progress Report No. 67," GE-NMPO, GEMP-67, June 30,1967, p. 160. 


341 


Theoretical - Thorn and Winslow 38 have derived the following theoretical relationship 
between oxygen self-diffusion coefficients in UOg+x an< ^ the composition of U 02 + X J 

, 1/2 


D = \ < l 2 > 

4 


v l x + 


2 8q i 

X“ + -i 


exp 


-(E v - Ei) 
RT 


/-u l \ 

exp et) (8 ' ,6> 


where 

D - self-diffusion coefficient for oxygen in UC>2+ X 
Z = number of nearest neighbors 
<£ 2 > = mean square of diffusional jump 

v - mean vibrational frequency of lattice 

U 1 = energy composed of energies associated with correlation of the diffusing species 
vibrating out of phase with its shell of neighbors 
E y = energy required to produce vacancies 
Ei = energy required to remove atoms from interstitial sites 
q^ = vibrational partition functions associated with interstitial sites 
q y = vibrational partition functions associated with vacancy sites 
x = fractional deviation from stoichiometric (x in UC> 2 + X ) 

They selected the following values for the parameters of the above equation: 


f <jf 2 > = 22.4 x 10 " 6 cm 2 

(U 

v = 18 x 10 12 sec " 1 
E y = 131.1 kcal/mole 
= 93.8 kcal/mole 
U 4 = 35.1 kcal/mole 
8 qi/q v = 500 

All the above values were derived from basic information except for the shell energy 
value which was derived from the experimental diffusion data of Auskern and Belle 39 for 
UO 2 002 * a * ater publication Thorn and Winslow 40 introduce new values for E v and U 4 
whereby the quantity (E v - Ej) becomes 40.9 kcal/mole and is 29.7 kcal/mole (the 
activation energy of Auskern and Belle for oxygen diffusion in UC> 2 + X having an O/U of 
2.004 and 2.067 in the temperature range 300° to 600°C). Using either set of values for 
(E y - E t ) and in equation (8.16) gives correlation with the results of Auskern and 
Belle; the later set of values better explains their results for nearly stoichiometric 
material (UC> 2 . 002 )- 

Chemical diffusion coefficients for oxygen in UG>2+ X (diffusion resulting from a chemical 
potential gradient) can be related to the coefficients for oxygen self diffusion (determined 
by tracer experiments) by the following equation: 


^chem ^self 



(8.17) 


where: P n 
°2 


= oxygen partial pressure 


Using Thorn and Winslow’s equation (8.16) to calculate D se if values for UO 2.05 an d 
converting these to D chem by equation (8.17) gives good correlation between theory and 


38 R. J. Thorn and G. H. Winslow, “Correlation of Thermodynamic Properties and Atomic Transport in the Uranium Dioxide 
Phase/' Thermodynamics: Proceedings of the Symposium on Thermodynamics, Vol. 2, IAEA, 1966, p. 234. 

39 A. B. Auskern and J. Belle, "Oxygen Ion Self Diffusion in Uranium Dioxide," J. Nucl. Mat., No. 3,1961, pp. 267-276. 

J. Thorn and G. H. Winslow, "Oxygen Self-Diffusion in Uranium Dioxide," J. Chem. Phys., Vol. 44, April 1966, p. 2822. 




342 


the experimental results of the present study, as shown in Figure 8.18. The dashed line 
is the theoretical curve and the solid line is the least-squares fit to the experimental 
data. The theoretical line in Figure 8.18 represents the composition UO2.05> for which 
there is general agreement in the literature concerning the value of the thermodynamic 
/d in PqA 

function I-£] . Furthermore, the composition UO 205 selected here is a reasonable 

\ d In x J 

first approximation to the composition of the test samples since the average O/U values 
changed from 2.003 to between 2.023 and 2.132 during the present experiments. 


Temperature, °C 



Fig. 8.18 — Oxygen diffusion coefficient versus reciprocal temperature for 
oxidation of UO 2 by steam compared to theoretical curve for 
oxygen diffusion coefficient in UO 2.05 

Although good correlation is established between theoretical and experimental values, 
effort is continuing to determine the best theoretical values with which to compare experi¬ 
mental results. This includes selection of the best set of constants for the Thorn and 
Winslow equation (8.16) and of the most valid equation for determining the thermodynamic 
function to be used in the conversion from self- to chemical-diffusion coefficients in equa¬ 
tion (8.17). It also remains to be determined which O/U value best represents experimental 
data obtained under oxidizing conditions in which the average O/U is constantly changing 
with time. 

8 . 3 PHYSICAL AND MECHANICAL PROPERTIES (A. D. Feith, E. F. Juenke, and 
W. L. McCullough) 

TENSILE STRENGTH OF TYPE 304 L STAINLESS STEEL 

The tensile strength of annealed 304 L stainless steel sheet was measured from room 
temperature to 1375°C. The test specimens were fabricated from 0.076-cm-thick sheet 
and had a 2.54-cm gage length in the major rolling direction of the sheet. Table 8.13 
shows the chemical analysis for this material. 






343 


TABLE 8.13 

CHEMICAL ANALYSIS OF TYPE 304L 


STAINLESS STEEL TENSILE SPECIMENS 


Element 

Percent 

C 

0.024 

Co 

0.05 

Cr 

18.00 

Cu 

0.28 

Mn 

1.69 

Mo 

0.16 

Ni 

10.95 

P 

0.024 

Si 

0.70 

S 

0.011 


Measurements from room temperature to 800°C were performed in air, utilizing an 
extensometer to determine the 0.2-percent yield point and to control the strain rate at 
0.005 cm per minute. The measurements from 800° to 1375°C were performed in argon 
at a crosshead speed of 0.15 cm per minute without using an extensometer. In all cases, 
the measurements were performed 0.6 hour after start of heat-up to temperature, allow¬ 
ing 0.3 hour to reach test temperature and 0.3 hour at temperature to attain constant tem¬ 
perature throughout the specimen. Ductility was measured as elongation at rupture, and 
the hardness (DPH) of each tested specimen was measured after test. 

Strength, ductility, and hardness results are given in Table 8.14. The tensile strength 
data versus temperature are plotted in Figure 8.19. Strength data for 800°C measured in 
argon are somewhat higher than in air, possibly because of the effect of oxidation or be¬ 
cause of a higher strain rate. Ductility and hardness data are plotted in Figure 8.20. 

THERMAL CONDUCTIVITY OF 304 L STAINLESS STEEL 

The thermal conductivity of Type 304 L stainless steel was measured in the 520° to 
1340° C temperature range by the radial heat flow method described previously. 41 ’ 42 Single 
Pt / Pt — lORh thermocouples were used to measure the temperatures at two radial loca¬ 
tions of the test specimen over the entire temperature range studied. Although the thermal 
gradient was small (3° to 8°C), it was determined to a reasonable degree of precision (ap¬ 
proximately ± 5%) by dividing the difference between the thermocouple outputs at the two 
radial locations by the sensitivity of the couple at the mean temperature value. This tech¬ 
nique is widely employed when measuring small gradients, rather than using differential 
thermocouples. 

The specimen was fabricated from a commercial grade 304L stainless steel of the com¬ 
position listed in Table 8.15. Shown also in this table are the compositions of other 300 
series stainless steels whose conductivities are compared with that of the tested material 
in Figure 8.21. 

The thermal conductivity of this material increased with temperature; this appears to 
be typical of the stainless steels. The results of this study have been fit to the following 
linear equation with a standard deviation of 1.770 x 10 “ 2 watt/cm-°C: 

k = 8.924 x IQ’ 2 + 1.970 x 10" 4 T (8.18) 


A. D. Feith, "A Radial Heat Flow Apparatus for High Temperature Thermal Conductivity Measurements/' GE-NMPO, 
GEMP-296, August 1963. 

4^A. D. Feith, "Measurements of the Thermal Conductivity and Electrical Resistivity of Molybdenum/' GE-NMPO, 
GE-TM 65-10-1, October 1965. 



TABLE 8.14 


TENSILE TEST RESULTS FOR TYPE 304 STAINLESS STEEL SHEET 


Specimen 

No. a 

Atmosphere 

Temperature, 

°C* 

0.2% Yield 
Strength, 
kg/mm 2 

Tensile 

Strength, 

kg/mm 2 

Elongation at rupture 
(in 2.54-cm gage 
length), percent 

Hardness, 

DPH b 

1 

Air 

25 

27.3 

65.4 C 

65 

174 

10 


25 

28.5 

66.5 

64 

174 

2 


200 

22.4 

46.4 

36 

203 

11 


200 

22.9 

47.2 

38 

227 

3 


400 

20.0 

44.9 

32 

266 

12 


400 

20.3 

44.8 

31 

254 

4 


600 

16.1 

33.2 

23 d 

218 

13 


600 

15.9 

33.9 

34 

218 

5 


800 

8.1 

10.4 

44 e 

173 

14 


800 

8.0 

10.0 

46 

164 

6 

Ar 

800 

- 

13.0 f 

53 

164 

15 


800 

- 

13.0 

52 

170 

7 


1000 

- 

4.9 

51 

143 

16 


1000 

- 

5.0 

44 

145 

8 


1200 

- 

1.9 

65 

157 

17 


1200 

- 

2.1 

53 

127 

9 


1300 

- 

1.4 

46 

105 

18 


1300 

- 

1.3 

48 

110 

19 


1350 

- 

0.91 

41 

104 

20 


1375 

- 

0.86 

43 

107 


a 0.076-cm-thick by 2.54-cm gage length, annealed condition. 

bAfter test; kg load. 

c Strain rate in air, 0.005/min. 

d Fracture at extensometer knife edge. 

e Fracture outside gage length. 

"^Loading rate in argon, 0.15 cm/min. (cross-head speed). 



Fig. 8.19 — Ultimate tensile strength of Type 304 stainless steel sheet 
















345 



Temperature, °C 



Fig. 8.20 — Ductility and hardness versus temperature as a result 
of tensile testing Type 304 stainless steel sheet 

where 

k = thermal conductivity, watt/cm-°C 
T = temperature, °C 

Values below 700°C are somewhat lower than other reported values, although the high- 
temperature data appear to agree well with values recommended by the TPRC data book. 43 
The recommended values for Type 304 stainless steel were extrapolated above 900°K 
(627°C) and apparently maintain a uniform displacement above the values for Type 347 
stainless steel. Values for Type 347 stainless steel above 1500°K (1227°C) were also 
extrapolated. 

Thermal diffusivity data for this same 304L stainless steel have been obtained in the 
300° to 1100°C temperature range and are reported under Task 1503. Enthalpy meas¬ 
urements are being made over the same temperature range and electrical resistivity 
measurements are planned. When these studies are complete, an adequate correlation 


^Thermal Physical Properties Research Center Data Book, Table 1089R, Vol. 1, June 1966. 






346 


TABLE 8.15 


COMPOSITION OF STAINLESS STEELS USED IN 
VARIOUS THERMAL CONDUCTIVITY TESTS 


Element 

This Study 
Type 304L 

TPRC a 
Type 304 

TPRC a 
Type 347 

Silverman^ 
Type 302 

Lucks c 
Type 301 d 

Powell e 
Type 18/8 

C 

0.024 

0.053 

0.06 

0.116 

0.15 

0.08 

Cu 



0.09 




Mn 

1.131 

0.67 

1.64 


2.0 

1.23 

Si 

0.70 


0.58 

0.13 

1.0 Max. 

0.62 

P 

0.014 

0.025 

0.013 

0.021 



S 

0.12 


0.017 

0.013 



Cr 

18.37 

18.51 

17.65 

18.4 

16-18 

18.68 

Ni 

9.89 

9.09 

10.94 

9.6 

6-8 

8.85 

Mo 

0.09 


0.02 




Nb 






0.99 

Ti 






0.14 


a Thermal Physical Properties Research Center Data Book, Vol. 1, June 1966, Table 1089R. 

^L. Silverman, "Thermal Conductivity Data Presented for Several Metals and Alloys Up to 
900°C," J. Metals, Vol. 5 (1953) pp. 631-2. 

c C. F. Lucks and H. W. Deem, "Thermal Properties of Thirteen Metals," ASTM Special 
Tech. Pub. No. 227, 1958. 

d Analysis not reported - values presented are typical for this material based on stainless 
steel catalogs. 

e R. W. Powell and R. P. Tye, "New Measurements on Thermal Conductivity Reference 
Materials," Proceedings of the Sixth Conference on Thermal Conductivity, Dayton, Ohio, 
1966. 



0 200 400 600 800 1000 1200 1400 160 ( 

Temperature, °C 


Fig. 8.21 - Thermal conductivity as a function of temperature for stainless steels 

between the thermal conductivity and thermal diffusivity of this material can be made. 
This correlation will be used to ascertain the correct positioning of the low-temperature 
portion of the curve representing conductivity data presented above. 

SPECTRAL AND TOTAL EMITTANCE MEASUREMENTS OF OXIDIZED ZIRCALOY-4 

Spectral and total hemispherical emittance measurements were carried out on both 
unoxidized and oxygen-saturated Zircaloy-4. The hole-in-tube method of measurement 
used in these studies is similar to that employed by Lemmon 44 for similar measurements 
on zirconium alloys and by Jain and Krishnan 45 in their studies on graphite. Figure 8.22 

44 A. W. Lemmon, op. cit., p. A-1. 

45 S. C. Jain and K. S. Krishnan, "The Distribution of Temperature Along a Thin Rod Electrically Heated in Vacuo. Ill: 
Experimental," Proc. Roy. Soc., Series A, Vol. 225, September 1954, p. 7. 







347 



To vacuum tube 
voltmeter and 
oscilloscopp 


Fig. 8.22 - Schematic diagram of furnace used 
for emittance measurements 


is a schematic diagram of the sample and important features of the apparatus. The sample 
was a tube of Zircaloy-4 about 15.2 cm long and 0.71 cm in diameter with a wall thickness 
of 0.038 cm. A hole, 0.079 cm in diameter, was drilled through one wall of the tube near 
the center. Two 0.025-cm-diameter tungsten wire voltage probes were spot-welded to 
the tube approximately 1.27 cm above and below the hole. Electrical power was supplied 
to the sample through water-cooled copper electrodes. The sample was surrounded by a 
Vycor chamber; the system was capable of operating with atmospheres of argon, steam, 
or vacuum. The black-body temperature and the surface temperature adjacent to the hole 
were measured optically through a quartz window. A vacuum tube voltmeter was used to 
determine both the voltage between the voltage probes and the sample current, calculated 
from the voltage drop across a calibrated standard resistance in series with the sample. 
An oscilloscope across the standard shunt was used to ensure a pure sinusoidal wave form 
for the sample current. 

The tube used in these experiments was long enough that the temperature at the middle 
of the tube in the region of the black-body hole was sensibly constant over a considerable 
length of the tube, approximately 5 cm. Under these conditions, the loss of heat from the 
region of the temperature measurement was due almost entirely to radiation from its sur¬ 
face, and little loss resulted from either thermal conduction toward the ends of the tube 
or from radiative transfer in the tube cavity. 


The spectral emittance was determined using the relationship: 


" C 2 

In E 2 = -g 


(T-S) 


TS 


(8.19) 







348 


where 

= spectral emittance 
X = observed wavelength 
C 2 = radiation constant, = 1.438 cm/°K 
S = surface temperature, °K 
T = black-body temperature, °K 

The wavelength observed in these measurements was 0.65 microns. 

Total hemispherical emittance was determined using the following equation: 

IV 

E T = :——4 ( 8 . 20) 

T 27rr4aT* v ' 


where 

E t = total hemispherical emittance 
I = sample current, amperes 
V = potential drop across measured section, volts 
r = radius of sample, cm 
i = length of measured section, cm 
a = Boltzman's constant = 5.673 x 10"12 watt/cm^-°K^ 

T = black-body temperature, °K 

Preliminary tests were run to compare spectral and total emittance values between 
900° and 1300°C for unoxidized Zircaloy-4, oxygen-saturated Zircaloy-4, and Zircaloy-4 
oxidized sufficiently to have a substantial ZrC >2 coating on the surface. The Zircaloy-4 
tubes used in the experiments meet LOFT specifications and ASTM specifications 
B-353-645 Grade RA-2, and have a surface finish of 63 RMS or better inside and out. 

Table 8.16 gives the vendor's ingot analysis. 

The measurements for unoxidized Zircaloy-4 were conducted on two samples heated in 
a vacuum (<10"3 Torr). On initial heat-up of these samples the spectral emittance was 
in the range of 0.6 to 0.7 up to a temperature between 1100° and 1200°C, at which point 
it dropped to about 0. 4. It is assumed that a change in surface condition, possibly oxida¬ 
tion or evaporation, accounts for this change in emittance behavior. The high emittance 
values are believed to result from oxidation of the metal by residual oxygen in the furnace 
during the initial heat-up. At high temperatures the oxygen dissolves rapidly in the metal 
and the emittance approaches that of the unoxidized metal. These initial experiments 
show that spectral emittance, E x , varies between 0.56 at 885° C and 0.43 at 1550°C. The 
value for 1100° to 1500°C is relatively constant, between 0.40 and 0.45. The total emit¬ 
tance, E t , for the same temperature range lies between 0.20 and 0.28. These values are 
in fair agreement with the values obtained by Lemmon 44 for unoxidized Zircaloy-2 and 
Zircaloy-B in the same temperature ranges; E^ = 0.43 to 0.44 and E T = 0.22 to 0.26. 

Application of an oxide layer by steam oxidation appreciably increased the spectral and 
total emittance. Several samples were oxidized for 30 minutes at temperatures between 
950° and 1050°C. In each case both the spectral and total emittance increased to between 
0.80 and 0.90 and remained thus through several thermal cycles. When steam oxidation 
was carried out for shorter periods of time, 5 to 10 minutes, the emittance reached high 
values as above. As heating continued, however, the oxygen of the scale dissolved in the 
metal and both spectral and total emittance values dropped, approaching the original values 
for the unoxidized metal. 

These observations prove that in order to measure the effect of oxide layer thickness 
on the emittance of Zircaloy-4, it is necessary to saturate the metal with oxygen before 


349 


applying the oxide layers. Table 8.17 gives the calculated conditions for saturation of 
Zircaloy-4 tubes of the dimensions used in these experiments by steam oxidation at vari¬ 
ous temperatures. The time required to deposit enough oxygen to saturate the metal and 
the time required to homogenize the sample by diffusing oxygen into the metal have been 
calculated using equation (8.10): 

o. „„ r /-39,700 ± 2300 \ 

W 2 /t = 5.56 x 10 5 exp I-—- 1 

developed in the reaction rate study described above and Lemmon's 44 equation for oxygen 
diffusion in alpha-zirconium, respectively. After saturation, oxide layers will be formed 
by additional oxidation in steam. Emittance values will then be determined and the oxide 
layer thickness will be measured metallographically. 

8 . 4 SUMMARY AND CONCLUSIONS 

Dynamic heating tests of Zircaloy-4 tubes with an internal pressure show that steam 
oxidation of the metal will result in lower deformation rates, less deformation, and higher 
failure temperatures. A model has been developed for expressing the tube failure by de¬ 
formation when oxidation is negligible, as a function of tensile strength and rate of heating. 


TABLE 8.16 

CHEMICAL ANALYSIS OF ZIRCALOY-4 
TUBING USED IN EMITTANCE MEASUREMENTS 3 


Ingot Analysis, 
wt % 


Element 

Top 

Bottom 

Cr 

0.09 

0.09 

Fe 

0.19 

0.20 

Sn 

1.45 

1.40 

Ni 

<0.004 

<0.004 

Zr 

Bal 

Bal 


Impurities 

Ingot Impurities, 
ppm 

Top Bottom 

C 

<80 

100 

H 

11 

6 

N 

25 

35 

O 

1400 

1000 

Al 

<20 

<20 

B 

<0.2 

<0.2 

Cd 

<0.2 

<0.2 

Cl 

<10 

<10 

Co 

<10 

<10 

Cu 

<20 

40 

Hf 

<100 

<100 

Mg 

<10 

<10 

Mn 

<20 

<20 

Mo 

<20 

<20 

Pb 

<20 

<20 

Si 

<30 

30 

Ti 

<20 

<20 

V 

<20 

<20 

W 

<50 

<50 


TABLE 8.17 


CALCULATED TIMES FOR SATURATION 
OF ZIRCALOY-4 TUBES 3 BY STEAM OXIDATION 


Temperature, 

°C 

Oxidation 

Time,* 5 

min 

Solution Time, 0 
min 

1000 

17 

337 

1050 

10 

183 

1100 

5 

103 

1150 

3 

61 

1200 

2 

37 

1250 

1 

23 

1300 

0.8 

15 

1350 

0.6 

10 

1400 

0.4 

7 

1450 

0.3 

5 

1500 

0.2 

3 


Calculations based on a sample dimension of 15.2-cm 
length, 0.71-cm diameter, 0.038-cm wall thickness. 
^Calculated from equation (8.10}: 

o c /—39,700 ± 2300\ 

W z /t = 5.56 x 10 5 exp (-—-I 

r /—41,000 ± 1500\ 

''Calculated from D a = 0.196 exp (-1 


a Analysis by vendor. 



350 


The oxidation of Zircaloy-4 and zirconium by steam shows a higher rate above about 
1600°C which can be expressed as 

60,600 ± 540o\ 

RT )' 

Oxidation of Zircaloy-4 in air is somewhat erratic and occurs at a greater rate than in 
steam; the difference is attributed to the reaction with nitrogen in the air. 

The rate-controlling mechanism for oxidation of UO 2 by steam is considered to be dif¬ 
fusion of oxygen through the UO£+ x lattice. This proposed mechanism was verified by 
measuring an oxygen gradient in samples which had not reached equilibrium conditions. 

The oxidation rate of 304 L stainless steel in air is about 10 3 times less than in steam 
up to 1370°C. This behavior is believed to be due to the presence of Cr 2 0 3 as a protec¬ 
tive film. During steam oxidation, however, formation of a film having the spinel struc¬ 
ture provides less resistance to oxidation. The air oxidation rate becomes equal to or 
greater than that for steam at 1370°C and above, since molten Fe-FeO destroys the pro¬ 
tective Cr 2 03 film. 

The tensile strength of annealed 304L stainless steel decreased from about 66 kg/mm 2 
at room temperature to 0.86 kg/mm 2 at 1375°C. 

The thermal conductivity of Type 304L stainless steel increased with temperature be¬ 
tween 520° and 1340° C. The variation of thermal conductivity with temperature for this 
temperature range may be expressed as follows: 

k = 8.924 x 10" 2 + 1.970 x 10‘ 4 T. 

The spectral emittance of unoxidized Zircaloy-4, at a wave length of 0.65 microns, 
varied between 0.56 at 885°C and 0.43 at 1550°C. Total hemispherical emittance varied 
between 0.20 and 0.28 over the same range The presence of an oxide layer increased 
the values for both spectral and total emittance to between 0.8 and 0.9. For thin oxide 
films, however, solution of the oxygen in the metal during prolonged heating caused both 
spectral and total emittance to approach original values for the unoxidized metal. 

8 . 5 PLANS AND RECOMMENDATIONS 

Spectral and total hemispherical emittance measurements will be continued on unoxi¬ 
dized, oxygen-saturated, and oxidized Zircaloy-4 to determine the effect of oxide layer 
thickness on the emittance of oxidized Zircaloy-4. 

A thermal-arrest technique will be employed to study the solidus temperatures for the 
system Zr-UC >2 from the zirconium-rich end to compositions in the range likely to be 
encountered in a loss-of-coolant meltdown accident. Phases present in quenched samples 
will be examined to augment the lower-temperature, pseudo-binary information presently 
available. 

Measurement of the effect of internal pressure on Zircaloy-4 tubing will continue with 
emphasis on the effect of oxide film thickness on the deformation and failure character¬ 
istics of this material. 

Irradiated Zircaloy-2 from the spent fuel discharge of a power-producing reactor will 
be studied to determine the effect of radiation damage on the behavior of pressurized 
tubes under dynamic heating conditions. 




9. FAST BREEDER REACTOR THERMOCOUPLE DEVELOPMENT 


(1414) 

E. S. Funston,* W. C. Kuhlmant 


The objective of this program is to identify and establish the properties of reliable high- 
temperature thermocouple, electrical, and electronic materials to provide instrumentation 
and electrical components for use in fast breeder reactors. 

9.1 W VERSUS W - 25Re THERMOCOUPLE CHARACTERISTICS AT HIGH TEMPERATURE 

Determining the centerline fuel temperature of LMFBR fuel elements requires temper¬ 
ature measurements to approximately 2500°C. Thermocouple systems used to measure 
this temperature must be constructed of materials that are metallurgically and chemically 
compatible with each other and with the anticipated environment. Although these materials 
may be compatible, the electrical properties (resistivity and thermoelectric power) of the 
sheath, insulators, and environmental gas may still influence the thermoelectric signal, 
making validity of indicated temperatures uncertain. Loss of effective electrical insulating 
properties of ceramic oxides when used in conjunction with a high resistance of thermo¬ 
elements has been observed at high temperatures by several investigators, 1 ’ 2 and this effect 
was measured as a function of temperature. 

Investigations during the past year have been primarily concerned with the many factors 
influencing the thermoelectric output generated in a thermocouple system. The various 
parameters that have received attention are: effects of gaseous environments, size of 
thermocouple wires, electrical insulation, and sheathing materials. By varying materials 
combinations and gaseous environments, and measuring the thermoelectric output as a 
function of temperature against calibrated standards, abnormal thermoelectric responses 
could be easily detected. Thermocouples using wire of 0.125-mm and 0.25-mm diameter 
were made and tested to determine whether optimization of wire diameters would be neces¬ 
sary. Electrical insulating characteristics of thoria and hafnia were compared at temper¬ 
atures ranging from 1600° to 2600°C. Research on different sheathing materials has been 
limited to the Mo — 50Re alloy tubing and pure molybdenum tubing. The effects of high- 
temperature gaseous environments on the thermoelectric response from W versus 
W — 25Ret thermocouples were determined in hydrogen and helium. An in-pile nuclear 
irradiation program was performed to determine the reactor stability of W versus W—25Re 
thermocouple systems. Data were obtained showing thermoelectric deviations as a function 
of temperature and neutron dosage. Each parameter investigated and significant results 
and conclusions of the individual experiments are discussed and illustrated. 

*Project leader. 

^Principal Investigator. 

* All compositions are in weight percent unless otherwise noted. 

1 E. A. Brown, B. G. Goodier, J. E. Perry, Jr., R. L. Petty, W. R. Prince, and C. R. Tallman, "Thermocouple Development for 
Project Rover," High Temperature Thermometry Seminar, Washington, D.C., February 24-26, 1965, WASH-1067. 

^G. F. Popper and T. Z. Zeren, "Refractory Oxide Insulated Thermocouple Analysis and Design," High Temperature 
Thermometry Seminar, Washington, D.C., February 24-26, 1965, 


351 





352 


ELECTRICAL INSULATION EVALUATION 

Comparative tests were made at temperatures ranging from 1600° to 2600°C to deter¬ 
mine the electrical insulating properties of the hafnia and thoria insulators. The construc¬ 
tion of each thermocouple used was essentially identical for these experiments, except for 
insulating materials and the web thickness between the two thermocouple holes separating 
the thermocouple legs. The web thickness of the HfC >2 insulator measured 0.3 mm; the 
ThC >2 insulator had a web thickness of 0.75 mm. No sheathing was used, and tests were 
performed in an electrical heat-resistant furnace (Figure 9.1). Both helium and hydrogen 
atmospheres were used as test environments. The differences in thermocouple response 
during the experimental testing for electrical insulating characteristics are shown in Fig¬ 
ure 9.2. The thermoelectric signals measured are shown by two curves which reveal that 
the thermocouple insulated with HfC>2 had the strongest thermoelectric response, even 
though the web thickness separating the W and W — 25Re thermoelements was less than 
one-half the thickness of the ThC >2 insulator. 

EFFECT OF GASEOUS ENVIRONMENT 

The next series of tests were primarily to establish effects on the thermoelectric re¬ 
sponse from thermocouples insulated with HfC>2 and ThC^ when used in hydrogen or helium 
environments. The thermocouples were made of 0.25-mm stranded W and W — 25Re alloy 
wire. The stranded wires were separated with HfC >2 and ThC >2 electrical insulators. The 



Fig. 9.1 — Test equipment to measure thermoelectric and thermionic 
anomalies in thermocouple wires 




353 



1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 

Tamptratur*, °C 


Fig. 9.2 — Comparative curves showing measured thermoelectric 
signals as a function of temperature in helium atmos¬ 
phere from W versus W — 25Re thermocouples insu¬ 
lated with thoria and hafnia 

Hf0 2 insulator measured 2.1-mm OD with two 0.6-mm holes. The Th0 2 insulator measured 
3-mm OD with two 0.3-mm diameter holes. The thermoelectric output as a function of tem¬ 
perature was determined using a direct-current electrical heat-resistant furnace (Figure 9.1). 
Initially the thermoelectric response of Hf02-insulated thermoelements was measured in a 
hydrogen environment. After data had been compiled, the system was purged with helium 
and electrical measurements were taken in helium. The experiments were repeated using 
thermocouples insulated with thoria. The changes measured are graphically illustrated in 
Figure 9. 3. The two sets of curves show that the response was lowered considerably for 
the Th0 2 -insulated thermocouple in hydrogen. The helium atmosphere shifted the thermo¬ 
electric output downward, but its overall effect was far less than hydrogen on the thoria 
insulator. The data show that hafnia as an electrical insulator is much less affected by 
hydrogen or helium atmosphere. All measurements in this series of tests were refer¬ 
enced against black-body measurements using an optical pyrometer calibrated against a 
NBS calibrated instrument. 

SHEATHING STUDIES 

Concurrent with the tests on thermocouple evaluation, test thermocouples were sheathed 
in molybdenum and Mo — 50Re alloy seamless tubing. Measuring the voltage output as a func¬ 
tion of temperature for thermoelements insulated with Hf0 2 or Th0 2 in either hydrogen or 
heliur gas showed only relatively small changes from calibrated values. The direction of 
the s all shift in voltage output was similar to those previously determined for unsheathed 
thermocouples. 

EFFECT OF UNMATCHED THERMOCOUPLE WIRE DIAMETERS 

To determine the effect of unmatched wire sizes in the construction of a thermocouple 
probe, a series of tests was initiated studying the effects of unmatched pairs on the emf signal. 
Wire sizes chosen were 0.25-mm- and 0.125-mm-diameter W and W — 25Re. Unmatched 
systems were made using 0.25-mm W versus 0.125-mm W — 25Re wire; the matched 
thermocouples had only 0.25-mm-diameter wire. The wires comprising the two thermo¬ 
elements of the couple were separated physically by HfC>2 ceramic insulation. The thermo¬ 
elements and electric ceramic insulator were sheathed in Mo — 50Re and unalloyed molyb¬ 
denum tubing. 





354 



1600 1700 1800 1900 2000 2100 2200 2300 2400 2500 2600 

Temperature, °C 


Fig. 9.3 - Comparative curves showing measured thermoelectric 
signals as a function of temperature in hydrogen and 
helium from W versus W — 25Re thermocouples insu¬ 
lated with thoria and hafnia 

The insulation and sheathing materials were selected for their maximum reliability in 
high-temperature thermocouple measurements, as shown by previous studies. The major vari¬ 
able in the thermocouple probe was in wire diameters. Testing was again performed in a 
direct-current electrical heat-resistant furnace (Figure 9.1). The thermoelectric signals 
were measured against a standard calibration for the W versus W — 25Re wire. Temper¬ 
atures were verified against optical pyrometer temperature readings on a black-body hole. 
The family of four curves shown in Figure 9.4 represent the deviations found, and appear 
to be well within the sensitivity of measuring techniques. The curves overlap each other 
up to 2200°C. In the higher temperature ranges the spread is insignificant. 

These tests strongly indicate that thermoelement wire sizes do not need to be matched 
in thermocouple systems. 

9. 2 ELECTRICAL INSULATION FOR HIGH-TEMPERATURE THERMOCOUPLES 

A vital part of the thermocouple system is the electrical insulator. Measuring temper¬ 
atures up to 1700°C is usually fairly easy, because there is a wide choice of materials. 
Alumina is generally preferred for its economy, excellent electrical insulator properties, 
and fabricability in all sizes. 

For temperature measurements above 1700°C the choice of materials narrows. In an 
attempt to expand present choices and discover better insulator materials, work was 
undertaken on compound oxide systems. Previous research studies 3 ’ 4 showed that the 
CaZr0 3 system had a potential for high-temperature insulation. Although its melting 
point was not high enough (-2375°C), similar compounds such as SrHf0 3 , SrZr0 3 , and 
CaHf0 3 have higher melting points: 2680°, 2550°, and 2500°C, respectively. These ma¬ 
terials have been considered possible insulators of thermocouples for center-core fuel 
element temperature measurements above 2500°C. Initial studies were made below 

3 " 

AEC Fuels and Materials Development Program Progress Report No. 69," GE-NMPO, GEMP-69, September 29, 1967, 
pp. 67—69. 

4 "Sixth Annual Report - High-Temperature Materials Program, Part A," GE-NMPO, GEMP-475A March 31 1967 
pp. 270-271. 




355 



1600 1700 1800 1900 2000 2100 2200 2300 2400 2b00 2600 

Temperature, °C 


Fig. 9.4 — Comparative curves showing measured thermoelectric 
signals generated between matched and unmatched 
wire sizes as a function of temperature 

this temperature so that the insulators could be compared with BeO (MP-2450 C), the 
highest resistivity insulator currently available as a thermocouple insulator. 

In three separate tests, each of the above insulators was evaluated by comparing the 
thermoelectric output of W versus W — 25Re thermocouples insulated with BeO and one 
of the above three materials. The thermocouples to be compared were placed in 
Mo -30Re sheaths which were closed only at the hot end. The sheathed thermocouples 
were placed together in an apparatus similar to that shown in Figure 9.1. Temperatures 
measured were referenced against black-body hole temperatures determined by a NBS 
calibrated optical pyrometer. The calibrations of the three different BeO-insulated 
thermocouples in all three tests were identical within ±0.5 percent of the same temper¬ 
ature for the same millivoltage output. The thermoelectric outputs of the SrZrOg-, SrHfC> 3 -, 
and CaHfC> 3 -insulated thermocouples were within ±0.5 percent of the BeO-insulated 
thermocouples up to 2000°C. Above this temperature, deviations occurred which are 
shown in Figure 9.5. The curve for SrHf 03 shows the least deviation from the BeO- 
insulated thermocouple of the three materials and is therefore a potential high-temper¬ 
ature thermocouple insulator. 

INTRINSIC VARIABLES AFFECTING THERMOCOUPLE VOLTAGE 

During the course of thermocouple testing, several synergistic effects were noted which 
resulted in abnormalities of the thermocouple signal, especially above 2000°C. In laboratory 
tests exploring these effects, experiments were performed on a non-insulated loop of Wwire 
and on a Hf0 2 -insulated loop of W wire inserted into a high-temperature furnace; the thermo¬ 
electric output between the two ends of the W wires was measured. Appreciable electric out¬ 
puts were obtained in both cases, although theoretically no output should be obtained. 

In the test furnace (Figure 9.1) an electrically insulated W T-bar made of 3-mm-diameter 
material was suspended inside a W muffle. Two 0.05-mm-diameter W wires, one with and 
one without an electrical insulator, were connected to the W T-bar as shown. This assem¬ 
bly was suspended in a resistively heated W tube furnace. Thermoelectric potential meas¬ 
urements were made between the W T-bar and both the bare and insulated 0.05-mm-diame- 






356 



Fig. 9.5 — Thermoelectric output of W-versus W — 25Re-stranded wire (ten 
0.075-cm thick wires) thermocouples using BeO.SrZrC^, SrHfC> 3 , 
and CaHf 03 for electrical insulation in hydrogen or helium 


ter wires as a function of temperature. Temperatures were measured using an optical pyro¬ 
meter by focusing on the bottom surface of the W muffle. The relatively large-diameter 
(3 mm) W T-bar was used to provide a thermoelectrically stable reference material which 
could then be used to measure thermoelectric anomalies in the small-diameter W wires by 
forming an electric circuit with the 0.05-mm-diameter wires. A thermoelectric study was 
made on bare 0.05-mm wires and on 0.05-mm wires surrounded by BeO or HfC>2. Small- 
diameter (0.05 mm) wires were used because it was believed they would give an exaggerated 
effect to any causes of abnormalities. 

The curves shown in Figures 9.6 through 9.8 are plots of measured Voltage changes as a 
function of temperature between the W T-bar and either the 0.05-mm-diameter W wire or 
the BeO-insulated 0.05-mm-diameter W wire in hydrogen, helium, or argon. Each test was 
run with a new piece of 0.05-mm-diameter W wire. In these three curves, the thermoelec¬ 
tric output developed between the reference W T-bar and the BeO-insulated 0.05-mm- 
diameter W wire was more consistent than between the W T-bar and the bare 0.05-mm- 
diameter W wire, as the environmental gas was changed from hydrogen to helium to argon. 
The reference W T-bar versus 0.05-mm W bare wire output in hydrogen increased rapidly 
to approximately 0.6 millivolts and then remained fairly steady to 2000°C. In argon the 
output increased to approximately 1800°C, then began to decrease. In helium, the output 
increased rapidly to approximately 1.7 millivolts at approximately 1872°C, three times 
the values in hydrogen or argon. 

This wide variation in thermoelectric signals generated between the tungsten T-bar 
versus bare tungsten wire in the three gases, compared to the relatively stable signals 
found in the insulated wire in the same gases, suggests that insulation prevents pickup of 
extraneous electrical signals within the muffle. 

The electrical measurements made in the specialized setup (Figure 9.1) indicate that 
the calibration of bare wire thermocouples must be done with care since there are a 
number of possible sources leading to errors in calibration measurements: 




Millivolts Millivolts 


357 



1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 

Temperature, °C 


Fig. 9.6 - Measured thermoelectric voltage in a hydrogen atmo¬ 
sphere test apparatus (Figure 9.1) designed to detect 
causes of anomalies noted during thermocouple 
calibration experiments 



Fig. 9.7 — Measured thermoelectric voltage in an argon atmo¬ 
sphere test apparatus (Figure 9.1) designed to detect 
causes of anomalies noted during thermocouple 
calibration experiments 




358 



1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 

Temperature, °C 


Fig. 9.8 — Measured thermoelectric voltage in a helium atmo¬ 
sphere test apparatus (Figure 9.1) designed to detect 
causes of anomalies noted during thermocouple 
calibration experiments 


1. Electrical leakage through the insulation. 

2. Electrical leakage through ionized gases. 

3. Induced emf if container is charged. 

4. Thermionics. 

5. Thermoelectric effect between "insulator" and thermocouple wire. 

9. 3 THERMOELECTRIC CHANGES IN W - 25Re DUE TO TRANSMUTATION 


The thermoelectric changes of an alloy representing the composition of W —25Re after 
1 year in 1014 neutron/cm 2 -sec thermal flux was previously evaluated and described in 
detail. 5 6 Theoretical calculations on transmuted effects indicate that the W - 25Re alloy 
composition will change to W - 21.5Re - 5.4 Os (at. %) after 1 year. An alloy rod of this 
composition was fabricated for thermoelectric tests after heat treating at 2770°C for 3 
hours; it was a single-phase alloy. When this rod was placed in a calibrating furnace 
which imposed a temperature gradient from room temperature to 2300°C along its length, 
two other phases precipitated in the rod between 2000° to 2300°C causing inhomogeneities 
along the rod length. This temperature-induced inhomogeneity caused differences (as 
much as 0.9 mv) in the emf-versus-temperature relationship, depending upon whether 
calibration was made while increasing or decreasing the temperature. The combination 
of neutron-induced thermoelectric changes (transmutation) and temperature-induced 
thermoelectric changes (phase precipitation) would make questionable the use of W-25Re 


as a thermoelement for temperature measurements when 1-year exposure in a 10*4 ne u- 
tron/cm 3 -sec thermal flux is required. 


REACTOR STABILITY OF W VERSUS W - 25Re THERMOCOUPLE 


A second reactor test of W versus W — 25Re thermocouples was performed under condi¬ 
tions very similar to one previously reported. 6 This test was designed to measure the 
neutron-induced thermoelectric changes in the W and W — 25Re thermoelements and to 


5 GEMP-69, pp. 67-69. 

6 GEMP-475A, pp. 268-270. 



359 


minimize the uncertainty previously experienced 7 in interpreting data relating to the 
W — 25Re leg. The test had considerable difficulty with W wires breaking; the small 
amount of W data obtained were inconclusive. Figures 9.9 and 9.10 show the results of 
thermoelectric deviations of W and W — 25Re as a function of temperature and time in 
the ORR with a flux of 1.2 x 10^ neutron/cm^-sec thermal and 2.1 x 10*3 neutron/cm^ 
sec fast (E n ^ 1 Mev). Data of Figure 9.9 indicate that the W thermoelement at low tem¬ 
perature (600°C) becomes more thermoelectrically positive with dosage, and at approxi¬ 
mately 1400°C changes very little with dosage. At approximately 1800°C no data were 
obtained, but the shape of the curve indicates that W becomes more thermoelectrically 
negative in this region. The W data appear to confirm previous findings 6 although at 600° C 
the latest test shows a greater positive increase in thermoelectric potential. The data of 
Figure 9.10 for W - 25Re indicate that the shapes of the curves are very similar to those 
obtained for W; this differs from conclusions of the first test 7 whose data were difficult to 
interpret. The similarity of the W and W — 25Re curves indicates that changes in the W 
thermoelement are compensated by changes in the W — 25Re thermoelement. There is a 
degree of difference between the W and the W — 25Re thermoelectric changes which results 
in a small net increase in thermoelectric output of a W versus W — 25Re thermocouple. 

For the time period represented by the two curves, the maximum increase in indicated 
temperature would be approximately 5°C. 



0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 


Reference W versus W - 25Re thermocouple emf, millivolts 


Fig. 9.9 — Relative stability of W thermoelements as a function of temperature and time 
in a 1.2 x 10 14 n/cm 2 -sec thermal flux 

REACTOR TESTING OF THERMOCOUPLE MATERIALS 

Test capsules containing 0.25-mm-diameter wires of W, W — 25Re, W — 3Re, Mo, Fe, 
Chromel, Alumel, Constantan, BeO, and AI 2 O 3 were shipped to the EBR-II. They were 
originally scheduled for insertion in December 1967, but have been delayed until March 
1968. These capsules contain 15-cm lengths of the above thermocouple materials. After 
exposure to approximately 10 22 to 10 22 neutron/cm 2 fast neutrons, all materials will be 
removed from the reactor and each will be compared thermoelectrically in the laboratory 
with unirradiated specimens of the identical materials. 

9. 4 SUMMARY AND CONCLUSIONS 

Studies of very high-temperature thermocouple characteristics revealed that the thermo¬ 
electric emf produced along BeO, Hf 02 , or TI 1 O 2 insulators can affect the output of the 

^"High-Temperature Materials Program Progress Report No. 63," GE-NMPO, GEMP-63, December 30, 1966, pp. 129—133. 





360 



Fig. 9.10 — Relative stability of W — 25Re thermoelements as a function of temperature 
and time in a 1.2 x lO^ 4 n/cm^-sec thermal flux 


thermoelements. Bare-wire thermocouples were found to be influenced by the environment. 

The nuclear radiation effect on W versus W — 25Re thermocouples in the ORR after ap¬ 
proximately 2 months in a 1. 2 x 10 14 neutron/cm 2 -sec thermal and 2.1 x 10 13 neutron/cm 2 - 
sec fast flux was studied. Results indicate that the W and the W-25Re thermoelements shift 
similarly in thermal emf, resulting in a small (5°C) positive error. 

SrHfC> 3 , SrZr0 3 , and CaHf0 3 were found to be inferior to BeO and Hf0 2 as thermocouple 
insulator materials. 

9. 5 PLANS AND RECOMMENDATIONS 

Evaluations will continue of neutron-flux-induced thermoelectric changes with emphasis 
on the effects of fast neutron dosage. 

Methods will be investigated to improve thermocouple performance at very high temper¬ 
atures by improving techniques and optimizing material selection. This effort is aimed at 
improving reliability and accuracy of center-core temperature measurement of LMFBR 
fuel pins. 

Studies will be made to improve the reliability of thermocouples for measuring the liquid 
metal coolant temperature of fast breeder reactors. 






10. PHYSICO CHEMICAL STUDIES OF Fe Cr Al CLAD FUEL SYSTEMS 


(7076) 

H. S. Edwards,* K. M. Bohlandert 


The objective of this program, which was terminated at the end of FY-66, was to define 
and to understand the reactions which can occur between oxidation-resistant Fe-Cr-Al-base 
alloys and UO 2 in the temperature region from 500° to 1200°C. 

Early attempts to use Fe-Cr-Al-base alloys as cladding for UO 2 or metal-UC^ fuels re¬ 
sulted in the appearance of uranium as surface contamination on specimens tested in air 
at elevated temperatures for extended periods of time. The tentative explanation of this 
result was that UO 2 was reduced by aluminum in the cladding, thus producing uranium 
which diffused to the surface. Final work on this program during CY-67 included diffusion 
studies and fueled capsule tests. A summary of the conclusions derived from the overall 
work program is presented. 

10.1 DIFFUSION STUDIES 

Fe, Cr, AND A1 DIFFUSION 

Diffusion couples of Fe-Cr-Al versus Fe and Fe-Cr-Al versus Cr were studied to deter¬ 
mine the rate of influx of cladding alloy constituents, such as Al, into the core area, which 
normally contains a cermet fuel of Fe-UC >2 or Cr-UC^. The couples were fabricated from 
2.54-cm-long by 1.27-cm-diameter rods of the Fe-Cr-Al alloys listed in Table 10.1. A 
0.32-cm-diameter axial hole was drilled into the rods and the core portion of the diffusion 
couple (usually Fe or Cr) press-fitted into the opening in the cladding. The diffusion cap¬ 
sules were completed by fitting a closure rod of the same alloy as the cladding into the top 
of the capsule and sealing by eleetron-beam welding in vacuum. After checking for leaks, 
the capsules were autoclaved for 2 hours at 1000°C and 700 kg/cm^ gas pressure to insure 
core — cladding contact. Sufficient couples were placed on test at 1000°C that a complete 
set could be removed for destructive analysis after 100, 300, and 1000 hours. 

The initial results of transverse electron microprobe analyses presented previously 1 
were not quantitatively accurate; trends in the data, however, indicated that cladding ma¬ 
terials containing greater quantities of Al will increase the distance Al diffuses into Fe 
cores, and that Cr most effectively curtails the diffusion of cladding constituents. The 
diffusion couples were re-examined using the electron microprobe in the step-scan scal¬ 
ing mode with fixed-time counting. Corrections were applied to the data and plots made 
of the normalized concentrations as a function of radial distance. 2 A solution of the diffu¬ 
sion equation, appropriate for the geometry of the test specimens, was determined and 

* Project leader. 

* Principal investigator. 

1# 'Sixth Annual Report — High-Temperature Materials Program, Part A/' GE-NMPO, GEMP-475A, March 31, 1967, pp. 178—179. 

2 ' j AEC Fuels and Materials Development Program Progress Report No. 67," GE-NMPO, GEMP-67, June 30, 1967, pp. 79-86. 


361 



362 


numerical values for diffusion coefficients of Fe, Cr, and A1 in Fe-base alloys were cal¬ 
culated (Table 10.2). No diffusion coefficients were calculated for samples with Cr cores. 
The markedly lower diffusion rates in Cr were apparent from a qualitative examination of 
the plots. 

URANIUM DIFFUSION 

The objective of the uranium diffusion studies was to determine the manner in which 
the U-bearing reaction products formed in the core were transported through the cladding 
to the surface. An attempt was made to determine the diffusion coefficient of U inFe-Cr-Al 
alloys and to differentiate between volume diffusion and grain boundary diffusion of U in 
these alloys. 

Determination of the diffusion coefficients of U in Fe-Cr-Al alloys was hampered by the 
lack of an accurate method for measuring very small concentrations of U in these alloys. 
Initially, the electron microprobe and spark-source mass spectrograph were used in at¬ 
tempts to measure U concentration gradients and solubilities. The extremely low U solu- 

TABLE 10.1 

COMPOSITIONS 3 OF ALLOYS FABRICATED FOR PROGRAM USE 

Fe - 5Cr - 4AI Fe - 5Cr - 7AI Fe - 5Cr - 10AI Fe - 4AI 

Fe — 15Cr — 4AI Fe-15Cr-7AI Fe-15Cr-10AI Fe-IOAI 

Fe — 25Cr - 4AI Fe-25Cr-7AI Fe-25Cr-10AI Fe - 5Cr 

__ Fe - 25Cr 

a Nominal compositions are in weight percent. 


TABLE 10.2 


DIFFUSION COEFFICIENTS AT 1000°C 


Nomina! Composition, wt % 


Diffusion Coefficient, 

cm^/sec 

Core 

Cladding 

Time, hr 

D Fe 

D C r 

Dai 

Fe 

Fe - 4AI 

100 

_a 

- 

6.5 x 10~ 10 

Fe 

Fe - 4AI 

300 

- 

- 

5.0 x 10“ 10 

Fe 

Fe-IOAI 

100 

7.8 x 10- 10 

- 

6.7 x 10- 10 

Fe 

Fe-IOAI 

300 

- 

- 

8.0 x 10 -10 

Fe 

Fe - 25Cr 

100 

2.9 x 10- 10 

3.1 x 10 -10 

- 

Fe 

Fe - 25Cr 

300 

- 

9.0 x 10" 10 

- 

Fe 

Fe - 5Cr — 4AI 

100 

- 

2.9 x 10 -10 

5.8 x 10 -10 

Fe 

Fe - 5Cr - 10AI 

100 

6.3 x 10~ 10 

2.8 x 10 —10 

5.3 x 10“ 10 






to 

21.0 x 10~ 10 

Fe 

Fe - 25Cr - 4A! 

100 

2.9 x 10 —10 

2.4 x 10- 10 

9.5 x 10- 10 

Fe 

Fe — 25Cr — 10AI 

300 

8.1 x 10- 11 

3.5 x 10“ 10 

1.3 x 10“ 9 

Fe - 10AI 

Fe — 5Cr — 10AI 

1000 

- 

5.5 x 10“ 10 

— 

Fe - 10AI 

Fe — 25Cr — 10AI 

100 

4.7x10“ 10 

3.4 x 10~ 10 

- 

Fe - 25Cr 

Fe - 25Cr - 4AI 

100 

- 

- 

3.5 x 10“ 10 

Fe - 25Cr 

Fe - 25Cr - 4AI 

300 

- 

— 

4.0 x 10- 10 

Fe - 25Cr 

■ . _ 

Fe — 25Cr — 10AI 

100 

- 

- 

6.8 x 10 —10 



363 


bilities in Fe-Cr-Al alloys (~0.1 wt %) were at or below the detection limit of the electron 
microprobe, and the relatively low spatial resolution of the spark-source mass spectro¬ 
graph rendered it unsuitable for accurate analysis of a solid diffusion couple. 

The high sensitivity and accuracy of the X-ray fluorescence spectrograph in determin¬ 
ing microgram quantities of U on the surface of Fe-Cr-Al-clad fueled specimens (dis¬ 
cussed later), prompted an investigation of this instrument as a potential tool in deter¬ 
mining U diffusion coefficients. As originally fabricated, the diffusion couple to be ana¬ 
lyzed consisted of two solid right cylinders of Fe — 5Cr — 4A1 — 1U and Fe - 5Cr - 4A1 
(both wt %) butted together and sealed in Fe-Cr-Al tubing. The couple was autoclaved at 
1000°C to insure contact and then heat treated for 3450 hours at 1000°C. Parallel sec¬ 
tions were removed by machining and analyses were made perpendicular to the specimen 
length with the periphery masked to eliminate surface effects. A graph of the U concen¬ 
tration as a function of distance (shown in Figure 10.1) agreed with pertinent chemical 
analyses, microprobe, and metallographic data; however, part of the experimental data 
in the region of the interface appeared suspect and would need to be verified. Also, the 



Fig. 10.1 - Uranium concentration profile in Fe — 5Cr — 4AI - 1U versus Fe - 5Cr - 4A! (wt %) 
diffusion couple after 3450 hours at 1000°C 


unsymmetrical shape of the data points in the vicinity of the interface does not fit available 
solutions of the diffusion equation. However, the extended length of diffusion time and the 
length of the diffusion zone may partly compensate for experimental errors. The maxi¬ 
mum diffusion distance measured by this technique correlates well with data obtained from 
the fission track etching method (discussed later). This correlation adds credence to the 
present data and prompted calculating an approximate value of the diffusion coefficient for 
U in the Fe — 5Cr — 4A1 alloy. Using a semi-infinite diffusion model with an extended source, 
the diffusion coefficient was determined to be in the range 10“® to 10“10 em^/sec, or ap¬ 
proximately the same as the coefficients determined for Fe, Cr, and Al. Although sup¬ 
porting data are necessary before this coefficient can be considered accurate, the results 




364 


do show the potentialities of the X-ray fluorescence spectrograph to perform difficult and 
sensitive analyses. It is believed that improved sectioning and standardizing techniques 
could significantly increase the accuracy of this technique for future diffusion studies of 
this type. 

An attempt was made to differentiate between volume and grain boundary diffusion of U 
in Fe-Cr-Al alloys using the fission track etching technique. 3 Diffusion couples of the type 
Fe-Cr-Al-U versus Fe-Cr-Al were heat treated for 500 and 1000 hours at 1000°C and 
100 hours at 1050°C. These specimens were longitudinally sectioned, metallographically 
polished, and shipped to a reactor site where they were coated with plastic and irradiated. 
After irradiation, the location of U in the Fe-Cr-Al side of the couple was determined in 
relation to existing grain boundaries. This method showed that U diffusion through Fe-Cr-Al 
alloys at 1000° to 1050°C is by volume diffusion, unaffected by grain boundaries. This me¬ 
thod also confirmed the existence of U in the U-depleted zone of the original Fe-Cr-Al-U 
portion and an abrupt decrease of U at the depleted zone - Fe-Cr-Al-U boundary. Because 
of the relatively small number of diffusion couples irradiated and the failure of several to 
develop reliable fission track patterns, no diffusion constant was obtained; however, maxi¬ 
mum U diffusion distances were measured and are shown in Figure 10.2 together with the 
maximum distance of U penetration as measured by X-ray fluorescence on Fe-5Cr-4Al-lU 
versus Fe — 5Cr — 4A1 after 3450 hours at 1000°C. These data verify that the diffusion of 
U is unaffected at 1000° C by grain boundaries and is independent of Fe-Cr-Al composition 
(within the limits studied). 



Fig. 10.2 - Maximum penetration of uranium in Fe-Cr-AI-base alloys at 1000°C as measured from 
the interface by fission track etching or sectioning methods 


3 

H. S. Rosenbaum and J. S. Armijo, “Fission Track Etching as a Metallographic Tool/' Journal of Nuclear Materials Vol 22 
April 1967, pp. 115-116. 



365 


10. 2 FUELED CAPSULE TESTS 

Capsule tests were performed to quantitatively determine the rate of U accumulation on 
the surface of Fe-Cr-Al-clad UO 2 -containing cermet fuel specimens as a function of the 
cladding and core composition. The specimens in these tests were 2.5-cm-diameter discs 
approximately 0.25 cm thick, with 0.038-cm-thick cladding on one flat side and 0.076-cm- 
thick cladding on the other. The cladding compositions were those listed in Table 10.1 and 
the fueled cores were predominantly UO 2 qo> UOj 939, and Fe - 4 OUO 200 ( vo1 %) sintered 
to densities exceeding 95 percent of theoretical. A few Cr — 4 OUO 2 00 ( vo1 ^ cores clad 
with 0.051-cm-thick Fe -25Cr -4A1 (wt %) were fabricated. The capsules were evacuated, 
sealed by electron-beam welding, and autoclaved at 600°C and 140 kg/cm^ gas pressure 
to achieve uniform core - cladding contact during testing. Testing was done under 2 atmos¬ 
pheres of argon gas containing 3 volume percent oxygen to maintain core — cladding con¬ 
tact, oxidize any U diffusing to the cladding surface, and prevent catastrophic oxidation 
of the low-alloy claddings. 

Partial results from testing capsules at 1200° C and 800° C were reported previously, 4 
together with information for a series of capsules containing Fe — 4 OUO 2 . 0 O cores anc ^ 

UOi ggg cores which completed 1000 hours at 600°C. Analysis for surface U was by X-ray 
fluorescence spectrograph with a 7 to 9 microgram U/cm^ lower limit of detection. 

Two series of clad specimens completed 1000 hours at 1200° C; one series contained 
UO 2.00 cores 311(1 the other UO 1989 cores. Figure 10.3, which shows the quantity of U 
on the surface of specimens containing UO^ 939 cores, demonstrates that free U (present 
in the core) diffuses completely through the cladding within 6 hours at 1200°C. The quantity 
of A1 present in-the Fe-Cr-Al cladding appeared to have little, if any, influence on the rate 
of accumulation of surface U; however, complete absence of A1 from the cladding (Fe-5Cr 
and Fe - 25Cr) decreased the U diffusion rate. The presence of U in the claddings had a 
detrimental effect on the cladding or cladding - core interface and resulted in many clad¬ 
ding failures due to radial cracking. 

Results for representative capsules containing UO 200 cores during 1000 hours at 1200°C 
are shown in Figure 10.4. Only minor differences arising from different cladding thickness 
are indicated, the first detection of surface U on the 0.76-mm-thick cladding occurred 20 
to 40 hours after the first indication on the 0.38-mm-thick side (13 hours). No definite re¬ 
lationship was established between the quantity of A1 in the Fe-Cr-Al cladding and the 
amount of surface U contamination. The data show that binary Fe-Al claddings will pro¬ 
duce greater amounts of U on the surface than Fe-Cr-Al alloy claddings and claddings 
containing no A1 will produce no detectable U during 1000 hours testing at 1200°C. Thus, 
it appears that A1 in the cladding does produce subsequent U contamination on the surface, 
probably due to reducing UO 2 to U. The comparatively high U concentration on the cap¬ 
sules without Cr in the cladding may indicate that Cr has an inhibiting effect on the diffu¬ 
sion of Al. A slight but similar trend was noted among the Fe-Cr-Al-clad capsules; those 
with 25 percent Cr had smaller surface concentrations than those with 5 percent Cr, even 
though the Al concentrations of the claddings were the same. In this group of specimens, 
there was no evidence of oxidation or cracking in any of the claddings as in the group of 
substoichiometric specimens. 

The preceding observations were in agreement with the diffusion results on Cr and Al 
and suggested that the diffusion of Al into the core (containing 1102 , 00 ) coul<1 be curtailed 
or eliminated with a Cr barrier layer; thus, no U would be produced to diffuse to the sur¬ 
face of the cladding. Accordingly, several capsules were fabricated from 0.51-mm-thick 
Fe — 25Cr — 4A1 (wt %) cladding and Cr — 4 OUO 2.00 ( vo1 %) luel was fabricated as the usual 

4 GEMP-475A, pp. 181-182. 



Uranium on cladding surface, /jg/crn 2 Uranium on cladding surface, fig/cm 2 


_ Fe - 4AI 


F«_ 1W_ inAI 














- 

25Cr 





100 200 300 400 500 600 700 800 

Time, hours 

900 

1000 

Fig. 10.3 — Concentration of U on surface of various claddings after 1000 hours 
at 1200°C (clad specimens contained UO-j .989) 





♦ Fe-25Cr-4AI 
□ Fe-25Cr-7AI 
O Fe - 25Cr - 10AI 
■ Fe_4AI 

A Fe - 10AI 
▼ Fe — 5Cr 

• Fe — 25Cr 


7 to 9 fig U/cm 2 detection limit 


Time, hours 


Fig. 10.4 — Concentration of U on surface of various claddings after 1000 hours 
at 1200°C {clad specimens contained UOo no) 



















367 


disc-shaped core. A 0. 08-mm-thick layer of Cr seperated the core and cladding. Capsules 
were tested in the usual manner at 1200°C with control capsules and a capsule containing 
a substoichiometric UOj 939 core separated from the Fe — 25Cr-4 A1 cladding by a Cr 
barrier. Results from thie control capsules duplicated previous results (Figures 10. 3 and 
10. 4). The capsule containing the UOj. 989 core showed no improvement over previously 
tested capsules with no Cr layer; hence, a Cr barrier does not impede the diffusion rate 
of U through the cladding after U is formed. Results from capsules containing Cr-UC> 2 . 00 
fuel and Cr barriers showed no surface U in 900 hours at 1200°C plus 125 hours at 1300°C. 
This appears to be a significant improvement over non-Cr-barrier capsules, but the ex¬ 
tent was difficult to assess because of the lower limit of detection of 7 to 9 micrograms 
U/cm 2 and the usual range of U concentration formed on Fe-Cr-Al-clad capsules (10 to 20 
micrograms U/cm 2 ) after 1000 hours at 1200°C. 

When examined, a metallurgical bond between the Cr and the core and cladding made 
the exact interface difficult to determine. There were a few small voids and oxide parti¬ 
cles in the interface area, fewer than in capsules containing Fe-U0 2 . 00 or Cr_uo 2 . 00 with 
no Cr barrier. From this limited amount of testing, it appears that a Cr barrier impedes 
the inward diffusion of A1 to the UO 2 core and thus prevents the formation of U; it does 
not, however, prevent the outward diffusion of U if it is already present in the fuel. 

Two series of capsules containing Fe - 40U0 2> 00 ( vo1 %) cores and UOj 939 cores were 
initially tested at 600° C for 1000 hours in argon containing 3 percent oxygen. Intermediate 
analyses by X-ray fluorescence during and after the test showed no surface U contamina¬ 
tion on any of the capsules. The temperature was increased to 800° C and the same speci¬ 
mens were tested an additional 1000 hours. Intermediate and final analyses indicated again 
that no surface U was on the claddings (within the 7 to 9 micrograms U/cm 2 detection 
limit of the instrument). 

The capsules were returned to test at 1000°C. After 100 hours, analyses showed U on 
all Fe-Cr-Al- and Fe-Al-clad UOj ggg core capsules. The amount varied from 10 to 60 
micrograms U/cm 2 and increased regularly with time until after 1000 hours it ranged 
from 30 to 380 micrograms U/cm 2 . The results were similar to the previous test at 
1200°C in that no definite relationship was established between Cr or A1 content of the 
cladding and the amount of U measured. No U was detected on the cladding alloys which 
did not contain Al; thus corroborating 1200° C test data showing enhanced U diffusion rates 
through cladding alloys containing Al. There were no cladding failures due to cracks dur¬ 
ing this 1000°C test as there were at 1200° C; however, metallographic examination dis¬ 
closed that U diffusing through the claddings had a detrimental effect. Most of the clad¬ 
dings bore evidence of intergranular oxidation; oxide penetration up to 0.35 mm was found 
in some specimens. 

Analyses of the series of clad capsules containing Fe - 4 OUO 2 . 0 O’ k° th during and after 
the 1000-hour test at 1000°C, detected no U on the surface of any capsule. Post-test metal¬ 
lographic examination showed no deterioration of the cladding (as occurred with substoichio¬ 
metric fuels) and good core-to-cladding contact with a thin discontinuous film, approxi¬ 
mately 3 microns thick, of AI 2 O 3 at the interface. 

10. 3 CONCLUSIONS 

Aluminum in Fe-Cr-Al-base cladding alloys reduces U0 2 in cermet fuels, beginning in 
the temperature range between 1000° and 1200° C. Free U formed in this reaction dissolves 
in and diffuses rapidly through the cladding to the surface where it is converted to an oxide. 
At 1000°C and below, no reaction of Al with UO 2 takes place within the lower limits of de¬ 
tection of the surface analytical method used (7 to 9 micrograms U/cm 2 ). Oxygen liberated 
during the reduction of U0 2 combined with Al from the cladding to form AI 2 O 3 at the clad- 



368 


ding - core interface. This stable compound inhibits subsequent diffusion of A1 into the 
UO 2 -containing fuel and consequently limits the quantity of U formed. This process of 
AI 2 O 3 barrier formation indicates that increasing the A1 content of the cladding above 
4 weight percent, to increase oxidation resistance, would not significantly influence the 
amount of surface U contamination. 

The diffusion coefficient of A1 in Cr was found to be significantly lower than other ele¬ 
ments in the Fe-Cr-Al-U system and led to the development of a Cr barrier between the 
fuel and cladding to reduce AI-UO 2 interaction and potentially eliminate surface U con¬ 
tamination. 

If free U is initially present in the fuel, for example, as substoichiometric UC> 2 _ X , or 
if free U is formed in the fuel during operation, then the U would diffuse through 0.38-mm- 
thick cladding within 100 hours at 1000° C or within 6 hours at 1200° C. Uranium diffusion 
through binary Fe-Cr alloys is curtailed in comparison with ternary Fe-Cr-Al alloys, in 
which the diffusion rate is virtually unaffected by the composition of the cladding (within 
the composition limits of this study 5 to 25% Cr and 4 to 10% Al). No U was detected on 
the surface of any specimen after 1000 hours at 600° C or after the test temperature was 
increased to 800°C for an additional 1000 hours. At 1000°C, U solubility in Fe-Cr-Al 
alloys is 0.1 to 0.2 percent, and its diffusion coefficient is in the range 10~9 to lO^cm^/sec. 



11. HIGH TEMPERATURE RESEARCH ON CARBIDES 
FOR FUEL AND STRUCTURAL APPLICATIONS 

(7073) 

J. F. White,* E. F. Juenket 


The objective of this program was to determine the high-temperature physico-chemical 
properties of refractory carbides for fuel and structural applications. 

This program terminated at the end of Fiscal Year 1967; significant results are pre¬ 
sented below. 

11.1 PREPARATION AND FABRICATION OF REFRACTORY CARBIDES 

The results of reaction studies for the preparation of well-characterized carbides of 
Ta, Zr, and Hf from their respective oxides are shown in Table 11.1. Spectrographic- 
grade graphite - oxide reactions require higher temperatures, in the range of 2200° to 
2300°C, to form well-crystallized single-phase material. While zirconium oxide reacts 
with C to form ZrC in about 1 to 2 hours at temperatures between 1850° and 2000 °C, com- 

TABLE 11.1 


X-RAY DIFFRACTION AND CHEMICAL ANALYSES OF CARBIDES PRODUCED FROM THE METAL OXIDES IN VACUUM 


Material 

Reaction 

Temperature, 

°C 

Time at 
Temperature, 
hr 

Lattice 

Parameter, 

a 

Combined 
Carbon, 
wt % (± 0.05) a 

Uncombined 
Carbon, 
wt % 

Oxygen, 

ppm 

Nitrogen, 

ppm 

Combined Carbon- 
to-Metal Ratio 

TaC c 

1800 

1 

Two phases 
TaC + Ta 2 C 

4.91 

1.48 ±0.05 

700 ±100 

23 ±20 

0.79 ± 0.01 

TaC° 

1800 

5 

TaC + Ta 2 C 

- 

- 

- 

- 

— 

TaC c 

2000 

1 

4.455 

5.74 

0.37 ± 0.05 

700±100 

<5 ±5 

0.93 ±0.01 

TaC c 

2000 

5 

4.456 

5.86 

0.28 ± 0.02 

600 ±100 

<5± 5 

0.94 ± 0.01 

TaC c 

2200 

1 

4.455 

6.07 

0.12 ±0.02 

350 ± 30 

<5± 5 

0.98 ± 0.01 

TaC c 

2200 

5 

4.456 

6.23 

0.10 ±0.02 

1640 ±100 

< 5 ± 5 

1.00 ± 0.01 

ZrC b 

1850 

2 

4.698 

11.47 

0.529 ± 0.05 

_ 

330 ± 30 

0.99 ± 0.01 

ZrC b 

2000 

1% 

4.649 

12.39 

0.18 ±0.02 

450 ± 30 

77 ±30 

1.08 ± 0.01 

ZrC b 

2150 

% 

4.699 

11.74 

0.18 ±0.02 

1140 ±100 

398 ± 30 

1.01 + 0.01 

ZrC c 

2000 

5 

4.698 

10.07 

1.31 ±0.05 

430 ± 30 

580±100 

0.86 ± 0.01 

ZrC c 

2200 

1 

4.699 

11.31 

0.62 ± 0.05 

1340 ±100 

252 ±30 

0.97 ± 0.01 

ZrC c 

2200 

5 

4.698 

11.59 

0.56 ± 0.05 

1300 ±100 

5 ± 5 

1.00 ± 0.01 

HfC c 

2200 

2 

4.638 

5.84 

0.36 ± 0.05 

920 ±100 

350 ±30 

0.92 ±0.02 

HfC c 

2300 

2 

4.638 

6.35 

0.36 ± 0.05 

720 ±100 

280 ± 30 

1.01 ± 0.01 

HfC c 

2300 

5 

4.638 

5.94 

0.62 ± 0.05 

470 ± 30 

350 ± 30 

0.95 ±0.01 


a Percent carbon for stoichiometric carbides {1:1 C/M) are as follows: 


TaC 6.22 
ZrC 11.63 
HfC 6.30 

^Prepared from oxide and carbon. 

c Prepared from oxide and spectrographic graphite. 


* Project leader. 

* Principal investigator. 


369 





370 


plete reaction between this oxide and spectrographic-grade graphite requires 5 hours at 
2200°C to produce single-phase ZrC. Hafnium oxide - graphite reactions to form HfC re¬ 
quire temperatures of at least 2300 °C to obtain a well-crystallized, homogeneous, single¬ 
phase product. 

Typical analyses of scaled-up lots (2 kg) of TaC, ZrC, and HfC (preparation based on 
the conditions determined in the above experiments) are shown in Table 11. 2. 

Analyses of Ta and Zr carbides prepared from the hydrides appear in Table 11.3. These 
data indicate that a 2200°C treatment was required to form single-phase TaC, whereas 
single-phase ZrC was formed at 1800 °C. Reacting the Zr hydride with C at 2000 °C and 
above, however, produced a higher carbon-to-metal ratio. In general, the oxygen content 
of the product is lower than in the graphite - oxide reactions; the nitrogen content is about 
the same. 

Vacuum hot pressing was used to form dense cylindrical specimens of various carbides. 
Table 11.4 summarizes the results. TiC and ZrC were fabricated at nearly theoretical den¬ 
sity by hot pressing in graphite dies at 2125 °C for 5 to 30 minutes. TaC and HfC required 
temperatures of 2200° to 2300 °C for 15 to 30 minutes to achieve densities from 94 to 97 
percent of theoretical. 


TABLE 11.2 


CHEMICAL AND X-RAY ANALYSES OF TYPICAL BATCHES OF TaC, ZrC, 
AND HfC PREPARED BY OXIDE-GRAPHITE REACTION IN VACUUM 


Material 

Reaction 

Conditions 

Free 

Carbon, % 

Oxygen, 

ppm 

Nitrogen, 

ppm 

X-Ray Analysis 

TaC 

5 hr at 2200°C 

0.16 

950 

8 

Single-phase fee 

ZrC 

5 hr at 2200°C 

0.56 

1300 

<5 

Single-phase fee 

HfC 

2 hr at 2400°C 

0.28 

91 

52 

Single-phase fee 


TABLE 11.3 


X-RAY DIFFRACTION AND CHEMICAL ANALYSES OF ZrC AND TaC PRODUCED FROM 
_METAL HYDRIDE - GRAPHITE REACTIONS IN VACUUM 


Material 

Reaction 

Temperature, 

°C 

Phases 

Combined 
Carbon, 
wt % (±0.05) 

Uncombined 
Carbon, 
wt % 

Oxygen, 

ppm 

Nitrogen, 

ppm 

Combined 
Carbon-to-Metal 
Ratio (± 0.01) 

TaC a 

1800 

Inhomogeneous 
TaC plus graphite 

5.67 

0.39 ± 0.05 

320 ± 30 

13 ± 5 

- 

TaC b 

1800 

Inhomogeneous 
TaC plus graphite 

5.50 

0.56 ± 0.05 

280 ± 30 

34 ±10 

— 

TaC a 

2000 

Inhomogeneous 

TaC 

5.73 

0.29 ± 0.05 

330 ± 30 

<5± 5 

— 

TaC b 

2000 

Inhomogeneous 

TaC 

5.89 

0.27 ± 0.05 

2.90 ± 30 

18 ± 5 

— 

TaC a 

2200 

TaC, a Q = 4.453 y 

6.07 

0.050 ± 0.007 

260 ± 30 

29 ± 10 

0.97 

TaC b 

2200 

TaC, a 0 = 4.454 1 

6.07 

0.032 ± 0.005 

180 ±30 

<5 ± 5 

0.97 

ZrC a 

1800 

ZrC, 3 q = 4.698g 

10.83 

0.40 ± 0.05 

650 ± 50 

345 ± 30 

0.93 

ZrC b 

1800 

ZrC, a Q - 4.698 1 

10.94 

0.38 ± 0.05 

590 ± 50 

369 ± 30 

0.94 

ZrC a 

2000 

11.55 

0.18 ± 0.02 

310 ±30 

299 ± 30 

0.99 

ZrC b 

2000 

- 

11.16 

0.33 ± 0.05 

350 ± 30 

351 ± 30 

0.96 

ZrC a 

2200 

ZrC, a 0 = 4.697 g 

11.33 

0.075 ± 0.005 

1570 ±100 

96+20 

0.97 

ZrC b 

2200 

ZrC, a Q = 4.697 ? 

11.36 

0.19 ± 0.02 

2910 ±100 

133 ± 20 

0.98 


^Material prepared by reaction of stoichiometric 1:1 graphite-to-metal ratio. 
Material prepared with 0.5 percent excess graphite. 





371 


TABLE 11.4 

DENSITY OF VARIOUS CARBIDE COMPACTS PREPARED 


BY VACUUM HOT-PRESSING 


Material 3 

Vacuum Hot-Pressing Conditions 

Density, 
percent of 
theoretical 

Temperature, 

°C 

Time, 

min 

Pressure, 

kg/cm^ 

Vacuum, 10 2 
Torr 

TiC 

1750 

5 

105.4 

5.0 

97.4 

TiC 

2030 

5 

105.4 

2.0 

100.0 

TiC 

2135 

15 

140.6 

1.5 

100.0 

ZrC 

2000 

45 

126.5 

2.5 

99.9 

ZrC 

2125 

30 

140.6 

1.5 

99.76 

ZrC 

2125 

30 

140.6 

1.5 

99.2 

HfC 

2175 

30 

105.4 

3.5 

90.1 

HfC 

2225 

30 

140.6 

2.5 

94.5 

TaC b 

2150 

30 

105.4 

8.0 

89.6 

TaC c 

2225 

30 

140.6 

2.5 

-85 

TaC d 

2250 

20 

140.6 

0.8 

95.3 

TaC d 

2235 

15 

140.6 

2.5 

96.8 


Cylindrical samples about 2.54 cm diameter x 3.18 cm high. 

^Pressed from commercial powder. 

c Pressed from powder prepared by vacuum reaction of Ta and carbon. 
^Pressed from powder prepared by vacuum reaction of Ta20g and carbon. 


This work was discussed in detail in a previous report. 1 
11.2 Ta CARBIDE 
THERMAL STABILITY 

The kinetics of the evaporation of TaC were studied in Ar + 3H£ (vol %) and in hydrogen 
atmospheres, at temperatures between 2000° and 3000°C. Weight loss from TaC in its 
homogeneity range followed a parabolic rate law indicating a diffusion-controlled mech¬ 
anism. Rates of C loss at the surface, determined by following surface compositional 
changes by X-ray analysis, were atmosphere dependent. In each atmosphere the surface 
lattice parameter (and, therefore, the composition) decreased more rapidly at higher 
temperatures. 

Color changes were observed on the surface of the sample during vaporization tests. 
Surface depletion of C results in a change from the characteristic golden color of TaC to 
a silver-gray color. The interior of the samples, however, retained the golden color, as, 
for example, did samples heated for 100 hours in hydrogen at 2000 °C and 55 hours at 2930 °C 
in Ar + 3^. X-ray analysis showed that the center was nearly stoichiometric TaC in both 
cases. These observations indicate a severe compositional gradient near the surface of the 
test samples and are consistent with a diffusion-controlled mechanism for loss of C from 
TaC in this temperature range. 

Diffusion coefficients were calculated from the weight changes for the loss of C from 
TaC through a C deficient (TaC-^) layer. The results are shown in Table 11.5. 

The loss of carbon from TaC is a reversible process. Carbon evaporates preferentially 
from TaC in an atmosphere of C activity lower than that of the solid throughout its homo¬ 
geneity range. Substoichiometric TaC will absorb C from atmospheres of higher C activi¬ 
ties. Thus, by controlling the C activity of the atmosphere, preferential evaporation of 


^'Fourth Annual Report - High-Temperature Materials and Reactor Component Development Programs, Volume II — 
Materials," GE-NMPO, GEMP-334B, February 26, 1965, pp. 107-112. 




372 


TABLE 11.5 


DIFFUSION COEFFICIENTS FOR CARBON IN TaC,^ 


Specimen 

Atmosphere 

Temperature, 

°C 

Diffusion Constant, 
cm^/sec 

22A 

Ar + 3H 2 

2000 

1.45 x 10 -10 

22B 

Ar + 3H 2 

2300 

( 3.22 x 10- 10 




(9.42 x 10~ 10 

16 

Ar + 3H 2 

2930 

1.37 x 10~ 8 

18B 

h 2 

2300 

(9.32 x 10~ 10 




(1.99 x 10“ 9 

20 

h 2 

2650 

2.74 x 10~ 9 


C from a TaCi_ x sample of any composition in the single-phase region can be suppressed. 
When this is accomplished, the C activity of the solid in equilibrium with an atmosphere 
of known activity is also known. 

The stability studies have been reported in detail previously. 2 ’ 3 
EFFECT OF STRAIN ON LATTICE PARAMETERS 

As noted above, severe compositional gradients occur due to preferential C evaporation 
from heat-treated TaC. X-ray lattice parameter measurements indicate that carbide sur¬ 
faces are often strained due to this C depletion. These strains, in turn, cause the calcu¬ 
lated lattice parameters to be biased, and significant errors are introduced into composi¬ 
tions estimated from lattice parameter. X-ray data from the present studies on TaC, as 
well as from similar data on NbC reported in the literature, 4 were analyzed to develop a 
means for determining when strain exists and to develop an analytical technique for taking 
strain into account when estimating composition from lattice parameter measurements. 

The relative magnitude and direction of the strain may be calculated from lattice para¬ 
meters measurements on carbides with the cubic NaCl structure. The parameters of TaCj_ x 
in this study were calculated from both the Cu Kdj and Cu Kff 2 peaks of the (331), (420), 
(422), and (333) (511) planes. These parameters were analyzed statistically by multiple 
regression against the variables cos# cote for the instrumental error, and a strain para¬ 
meter T for the error due to strain. 

These studies have shown that strain effects are present in lattice parameter measure¬ 
ments whenever Ta or Nb carbides are heated under non-equilibrium conditions. The degree 
and sense of the strain is dependent upon the thermal history of the specimen. 

On the basis of these analyses, an analytic expression was derived which permits deter¬ 
mination of the lattice parameter (a 0 ) from the measured parameter (a) corrected for in¬ 
strumental error (cos# cot0 term) and the strain parameter (r term) by the following equation: 

a = a Q + b cos 9 cot# + c (r - P 0 ). 

A positive strain coefficient (c) indicates that the strain is due to C depletion from the 
surface, whereas a negative coefficient indicates a higher C concentration at the surface. 

*r is a parameter which varies roughly in proportion to Poisson's ratio, and which for planes in cubic lattices 

equals (h 2 k 2 + k 2 U 2 + & 2 h 2 )/(h 2 + k 2 + i 2 ) 2 . 

2 GEMP-334B, pp. 114-120. 

3,, 

Fifth Annual Report - High-Temperature Materials Program, Part B," GE-NMPO, GEMP-400B, February 28 1966 

pp. 140-143. 

4 E. K. Storms, N. H. Kirkorian, and C. D. Kempter, "Crystallographic Data: Niobium Carbide," Analytical Chemistry 

Vol. 32, No. 12, 1960, p. 1722. 



373 


The coefficient is zero for a constant composition within the layer penetrated by the X-ray 
beam, as, for example, when the carbide has been equilibrated with an atmosphere of 
fixed C activity. 

The analysis of compositional gradients in carbides and strain analyses based on these 
measurements were discussed more fully in a previous report . 5 

11.3 NON-STOICHIOMETRY IN Ta AND U MONOCARBIDES 

Refractory carbides, such as TaC x and UC X , are of interest for high-temperature ap¬ 
plications principally because of their high melting points. These materials, however, 
characteristically undergo compositional variations and deviations from stoichiometry at 
high temperatures; hence, the reactivity of C must be examined from the standpoint of C 
activity as a function of composition. The object of this work was, therefore, to establish 
the C activity of Ta and U monocarbides as a function of composition and temperature. Re¬ 
lated thermodynamic properties were deduced from an analysis of the data making use of 
the model proposed by Hoch 6 for nonsubstitutional solid solutions. 

EXPERIMENTAL PROCEDURE 

The experimental procedure consisted of equilibrating the monocarbide powder with an 
atmosphere of controlled C activity. The desired C activities were obtained by introducing 
hydrogen-methane gas mixtures at a pressure of 1.2 atmosphere into a Re muffle, as shown 
in Figure 11.1. The exterior portions of the furnace were maintained in hydrogen at 1.1 at¬ 
mosphere. Carbon content of the controlled atmosphere was checked by chemical analysis 
of the exit gas, and oxygen content, as water, by a Beckman Electrolytic Hygrometer. Unit 
C activity was achieved in some runs by heating the powders in hydrogen in a closed graphite 
crucible in the presence of graphite powder. The principal gas species in the atmosphere 
in each run were determined by use of the thermodynamic functions of McBride, et al. T 
The measured amount of CH 4 at room temperature was taken to be the total C in all species 
at the temperature of the experiment, except where unit C activity was assumed to calculate 
the concentration of the gaseous species in hydrogen in equilibrium with graphite. The 
methane content of the gas was varied between 1 and 100 ppm, and the oxygen content was 
found to vary between 1 and 15 ppm. 

Tantalum monocarbide samples consisted of powder of average particle size < 44 microns 
whose X-ray and chemical analyses are given in Table 11. 6 . These powders, in Re crucibles, 
were equilibrated with the gas mixtures at temperatures between 1800° and 2350°C. X-ray 
diffraction analysis of equilibrated powders showed no evidence of strain due to composi¬ 
tional gradients. 

Samples were heated in increments of 1 to 12 hours depending upon temperature. After 
each interval, the samples were weighed and lattice parameters were measured to establish 
when equilibrium was being approached; i. e., powders were heated until the strain effect 
due to compositional gradients was insignificant and lattice parameters varied by less than 
0.0010 A between successive readings. Compositions were calculated from the lattice 
parameter a Q , corrected for strain effects and instrumental error, using the equation de¬ 
termined by Bowman 8 for the variation of lattice parameter with C-to-Ta ratio. 

5 GEMP-400B, pp. 143-146. 

Hoch, "Statistical Model for Nonsubstitutional Solutions," Trans. Met. Soc. of AIME, Volume 230, 1964, p. 138. 

^B. J. McBride, S. Heimel, J. C. Ehler, and S. Gordon, "Thermodynamic Properties to 6000°K for 210 Substances 
Involving the First 18 Elements/' NASA SP-3001, 1963. 

8a. L. Bowman, "The Variation of Lattice Parameter with Carbon Content of Tantalum Carbide," J. Phys. Chem., 

Vol. 65, 1961, p.1596. 


Ho + C Ha mixture - 4 - 



Fig. 11.1— Schematic diagram of experimental set-up for equilibrium studies 


TABLE 11.6 


CHEMICAL AND X-RAY ANALYSES OF TANTALUM CARBIDE 
COMPOSITIONS USED IN EQUILIBRATION EXPERIMENTS 


Nominal Composition 

TaC 0.8 

TaC 0.90 

TaC 0.99 

Carbon, % 

4.94 

5.76 

6.32 

Uncombined carbon, ppm 

134 

<100 

2700 

Oxygen, ppm 

167 

177 

163 

Nitrogen, ppm 

105 

40 

21 

X-ray lattice parameter (a Q ), X 

4.42577 

4.44303 

4.45585 

C/Ta (chemical analysis) 

0.784 

0.921 

0.973 

C/Ta {X-ray analysis) 3 

0.800 

0.910 

0.992 


a Based on: A. L. Bowman, "The Variation of Lattice Parameter 
with Carbon Content of Tantalum Carbide/' Journal of Physical 
Chemistry, Vol. 65, 1961, pp. 1596-1598. 











375 


To ensure equilibrium, TaC 0o 80 and TaC 0.99 P owders were heated side-by-side in a 
divided crucible in some tests. The change in composition of each powder was determined 
periodically by lattice parameter measurements. An exponential curve of the form 
A = A.oo + Be"Ct fitted to the data yielded the asymptotic lattice parameter from which 
equilibri um compositions were calculated by Bowman's 8 equation. Results of a typical run 
are shown in Figure 11.2. The calculated compositions for samples approaching equilibrium 
from both sides in this manner agree to within 1.0 percent. 



Fig. 11.2 - Extrapolated lattice parameter versus time at 1800°C for 
TbCq qqq and TaCg gg4 in H 2 +14 ppm CH 4 


Uranium monocarbide samples consisted of powders (< 44 microns particle size) en¬ 
capsulated in 0.05-cm-thick Re capsules formed from Re sheet and sealed in vacuum by 
electron-beam welding. Since Re does not form a stable carbide but dissolves up to 11.7 
atomic percent C, 9 it serves as a membrane permeable to C through which the carbide may 
be equilibrated with the C activity of the outside gas atmosphere. In this way, evaporation 
of U is suppressed. 

The composition of the commercially processed starting material is given in Table 11.7. 

The samples were heated at 1500 °C and 1700 °C in the hydrogen - methane mixtures. 

They were weighed at regular intervals and assumed to have achieved equilibrium when 
a weight change less than 0. 01 percent of the sample weight was observed in a 6-hour period. 
The total weight loss in a 2.5-g sample was about 0.030 g. Samples for X-ray diffraction 
analysis of the starting powder and of equilibrated powders after each series of runs were 
sealed in glass capillary sample holders under argon to avoid exposure of the sample to air. 


9 J. E. Hughes, "A Survey of the Rhenium-Carbon System," J. Less Common Metals, Vol. 1, 1959, p. 377. 






1 


376 


Results 

The C activity versus composition data measured for TaC x and UC X at various tempera¬ 
tures are summarized in Tables 11.8 and 11.9, respectively, along with a calculated activity 
coefficient for each data point, using a method described below. 

In the TaC studies, chemical analyses showed the oxygen content to be less than 600 ppm 
in all the samples analyzed. This concentration is low enough to be ignored in subsequent 
discussions. In the case of UC X , however, the oxygen content is high enough that the system 
under study must be considered as a portion of the U-C-O ternary system, as discussed 
below. 

DISCUSSION 

The variation of the C activity with composition and temperature was analyzed in terms 
of the statistical model of Hoch 6 to(l) study the implications of deviations from stoichiometry 
in TaC and UC, (2) establish the defect interaction energies, and (3) to determine the Frenkel 


TABLE 11.7 


CHEMICAL AND X-RAY ANALYSES 
OF URANIUM CARBIDE USED IN 
EQUILIBRIUM EXPERIMENTS 


Carbon, % 

4.77 

Uncombined carbon, ppm 

2300 

Oxygen, ppm 

6300 

Nitrogen, ppm 

140 

X-ray lattice parameter (a Q ) 

, A 4.960 a 

C/U (chemical analysis) 

0.951 

a Very diffuse lines. 


TABLE 11.8 


CARBON ACTIVITY AS A FUNCTION OF TEMPERATURE AND COMPOSITION IN TaC x 


Temperature, 

°C 

x in TaC x 

X-Ray 

Analysis Chemistry 

Oxygen, ppm 

Carbon 

Activity 

< a c> 

X 

1800 

0.805 



0.0014 

-32.8 

1800 

0.926 



0.0247 

-25.7 

1800 

0.931 



0.0283 

-25.4 

1800 

0.925 



0.0147 

-25.6 

1800 

0.936 



0.0283 

-25.7 

1830 

0.927 

0.864 

390 

0.0418 

-23.9 

2035 

0.928 

0.959 

561 

0.0287 

-28.0 

2035 

0.933 

0.955 

369 

0.0450 

-26.3 

2150 

0.845 



0.0085 

-31.0 

2150 

0.823 



0.0073 

-31.0 

2220 

0.914 

0.869 

547 

0.0342 

-28.4 

2220 

0.924 



0.0449 

-27.5 

2250 

0.995 



1.0000 

-26.5 

2260 

0.897 



0.0401 

-27.1 

2300 

0.872 



0.0185 

-30.2 

2300 

0.875 



0.0225 

-29.3 

2300 

0.869 



0.0225 

-29.1 

2300 

0.866 



0.0185 

-30.0 

2300 

0.994 



1.0000 

-26.3 

2345 

0.900 



0.0292 

-29.8 

2345 

0.994 



1.0000 

-26.6 






avg. -27.9 ±1.9 




377 


TABLE 11.9 


CARBON ACTIVITY AS A FUNCTION OF TEMPERATURE 
AND COMPOSITION IN UC x O Q 099 _ 


Temperature, 

°c 

Composition, 
x + y 

Carbon 

Activity 

a c 

/ 1—x—y\ 
RT In la 1 

\ C X ) 

kcal/mole 

1500 

0.992 

0.010 

-32.8 

1500 

0.994 

0.044 

-29.4 

1500 

0.991 

0.039 

-28.4 

1700 

0.990 

0.061 

-29.5 

1700 

0.971 

0.010 

-32.5 

1700 

0.993 

0.033 

-32.4 

1700 

0.952 

0.012 

-29.9 




avg. —30.7 ± 1.6 


energy associated with a transfer of a C atom from a singly occupied site to a site already 
occupied by another C atom in the U monocarbide phase region. 

Activity coefficients for TaC derived in this analysis are included in Table 11.8. The 
activity coefficient, to a first approximation, is independent of composition for 0.805 < x 
< 0.995 and may be only slightly dependent on temperature for 1800° < T < 2350°C. The 

average value for the activity coefficient, RT In (s-c — ^ , is -27.9 ±1.9 kcal/mole. 

This agrees with the value of approximately -27 kcal/mole obtained by Hoch 10 in an analysis 
of the evaporation rate data of Fries 11 for NbC, and of Coffman, et al., 12 for TaC. 

The experimental results for U carbide (UC x O 0 .099) are given in Table 11. 9. The data 
show that the activity coefficient is fairly constant as a function of temperature and com¬ 
position, with an average value of -30.7 ±1.6 kcal/mole. If the carbon-oxygen interaction, 
energy is assumed to be equal to the value deduced by Hoch® for the carbon-oxygen inter¬ 
action in the iron-carbon-oxygen system (-26.4 kcal/mole), the activity coefficient for C 
in UC X (x < 1) to a first approximation is 

RT In ^a c 1 ~ X -28.1 ± 1.6 kcal/mole. 


This value was deduced from data on compositions close to stoichiometry (average x = 
0.983), ignoring double occupancy of some C lattice sites. When double occupancy is taken 
into account, the activity coefficient becomes: 


RT 


In ^a c —j = -26.9 ±1.6 kcal/mole. 


Calculations of the C activity coefficient using the data of Storms 13 for UCq. ggg gives 

RT in (a —-) = -25.9 kcal/mole (T = 2100°K) 

' X ' = -25.3 kcal/mole (T = 2300°K) 


This agrees with the above value. 


10 M. Hoch, private communication. 

11R. J. Fries, "Vaporization Behavior of Niobium Carbide/ Journal of Chemical Physics, Vol. 37, No. 2, July 1962, 
pp. 320-327. 

12 J. A. Coffman, G. M. Kibler, T. R. Riethoff, and A. A. Watts, "Carbonization of Plastics and Refractory Materials 
Research, Part l/ # WADD-TR-60-646, Part I, February 1961. 

13 e . k # storms, "A Mass Spectrographic Study of the Vapor Pressure of U(g) and UC 2 (g) Over Various Compositions 
in the Uranium-Carbon System," Thermodynamics, Vol. 1, IAEA, Vienna, 1966. 



378 


The activity coefficient for the second G atom on a C site in U monocarbide is 6.8 ± 0.5 
kcal/mole. Therefore, the Frenkel energy for UC X , i. e., the energy required to move a 
C atom from a C site to a second site already occupied by a C atom, is 33.7 ±1.7 kcal/mole. 

Based on this analysis, the disorder in stoichiometric uranium carbide (UC^ q) was 
estimated. The results indicated that at 2000 °K the fraction of available C sites occupied 
is 0.986; at 2500 °K, it is 0.967. In other words, in UCjq at 2000 °K, 1.4 percent of the 
C sites are doubly occupied; at 2500°K, 3.3 percent are doubly occupied. 

The experimental procedure and results for the equilibration studies have been discussed 
extensively in previous reports. 14 ’ 15 A detailed discussion of the data analysis in terms 
of the statistical model has been published. 16 

11.4 PLANS AND RECOMMENDATIONS 

This task terminated June 30, 1967. A topical report will be issued covering the work 
accomplished in this program. 


14 GEMP-400B, p. 146-155. 

15 "Sixth Annual Report - High-Temperature Materials Program, Part A," GE-NMPO, GEMP-475A, March 31, 1967, 
16 

M. Hoch, E. F. Juenke, and L. H. Sjodahl, "Non-Stoichiometry of Tantalum Carbide and Uranium Monocarbide," 
GE-NMPO, GEMP-540, prepared for presentation at IAEA Symposium on Thermodynamics of Nuclear Materials, 
Vienna, Austria, September 4—8, 1967.