The Effect of Environment, Chemistry, and Microstructure on the Corrosion Fatig ie Behavior of
ARCHIVES
Austenitic Stainless Steels in High Temperature Water
MASSACHUSETTS INSTITUTE
OF TECHNOLOLGY
by
Lindsay Beth O'Brien
MAY 0 6 2015
B.S. Nuclear and Mechanical Engineering, 2011
Rensselaer Polytechnic Institute
LIBRARIES
SUBMITTED TO THE DEPARTMENT OF NUCLEAR SCIENCE
AND ENGINEERING
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE IN NUCLEAR SCIENCE AND ENGINEERING
AT THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
September 2014
2014 Massachusetts Institute of Technology. All rights reserved.
Signature of Author:
Signature redacted
6'
Certified by:
- r-
Lindsay Beth O'Brien
Department of Nuclear Science and Engineering
August 29, 2014
ignature reda cted
Ronald G. Ballinger
rofessor of Nuc1ejr Science and Engineering, Materials Science and Engineering
Thesis Supervisor
Certified by:
ignature redacted
Michael P. Short
Assistant Professor of Nuclear Science and Engineering
Thesis Reader
Signature redacted
Accepted by:
Mujid S. Kazimi
CO Professor of Nuclear Engineering
Chair, Department Committee on Graduate Students
The Effect of Environment, Chemistry, and Microstructure on the Corrosion Fatigue Behavior of
Austenitic Stainless Steels in High Temperature Water
by
Lindsay Beth O'Brien
B.S. Nuclear and Mechanical Engineering, 2011
Rensselaer Polytechnic Institute
SUBMITTED TO THE DEPARTMENT OF NUCLEAR SCIENCE
AND ENGINEERING
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE IN NUCLEAR SCIENCE AND ENGINEERING
AT THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
September 2014
2014 Massachusetts Institute of Technology. All rights reserved.
Signature of Author:
Signature redacted
Lindsay Beth O'Brien
Department of Nuclear Science and Engineering
August 21, 2014
Certified by:
Ronald G. Ballinger
Professor of Nuclear Science and Engineering, Materials Science and Engineering
Thesis Supervisor
Accepted by:
Mujid S. Kazimi
TEPCO Professor of Nuclear Engineering
Chair, Department Committee on Graduate Students
The Effect of Environment, Chemistry, and Microstructure on the Corrosion Fatigue
Behavior of Austenitic Stainless Steels in High Temperature Water
By
Lindsay Beth O'Brien
Submitted to the Department of Nuclear Science and Engineering
On August 29, 2014, in partial fulfillment of the
requirements for the degree of
Masters of Science in Nuclear Science and Engineering
Abstract
The effect of sulfur on the corrosion fatigue crack growth of austenitic stainless steel was evaluated
under Light Water Reactor (LWR) conditions of 288'C deaerated (less than 5ppb 02) water, to
shed light on the accelerating effect of the LWR environment and to explore the effect of high
sulfur content on the retardation of fatigue crack growth rates. Fatigue tests were performed using
a trapezoidal loading pattern with rise times of 5.1, 51, 510, and 5100 seconds (fall time of 0.9, 9,
90, and 900 seconds), with Kmzx of 28.6 or 31.9 MPa'm and stress ratios (R, Pmin/Pmax) of 0.4 or
0.7. Two test materials were used to evaluate the effect of sulfur: (1) a low sulfur (<0.0025 wt%)
stainless steel and, (2) a high sulfur (0.032 wt% stainless steel. The low sulfur stainless steel
exhibited increasing crack growth rates from 9.4 x10-5 mm/cycle to 1.2x 1 04 mm/cycle as rise times
were increased from 5.1 to 5100 seconds with a stress ratio of 0.7. The high sulfur stainless steel
exhibited decreasing crack growth rates from 1.4 x10-4 mm/cycle to 7.9 x10 5 mm/cycle as rise
times were increased for a stress ratio of 0.4, and crack growth rates from 6.4 x10 5 mm/cycle to
3.6 x10 5 mm/cycle with increasing rise time at a stress ratio of 0.7. Evaluation of the crack growth
rates showed environmental enhancement of the crack growth rates for the low sulfur stainless
steel, while the high sulfur stainless steel showed retardation of environmental crack growth rates,
likely due to the increased corrosion at the crack tip associated with the high sulfur content.
The crack surfaces were characterized using Scanning Electron Microscopy (SEM). The low sulfur
material showed a light layer of corrosion product that decreased in thickness as the testing
progressed, and faceting on the surface was highly crystallographic. Faceting ran both
perpendicular and parallel to the crack for the short rise time steps of the test, but fewer
perpendicular facets were evident at the longer rise times. The high sulfur material was heavily
corroded throughout the fracture surface, and crystallographic faceting was seen for stages of the
test with R=0.4 For R=0.7, the heavy oxidation on the surface made the facets hard to resolve.
Striations were apparent during the 5100 second rise time for the low sulfur material (where
corrosion was almost nonexistent) and throughout the entirety of the crack surface for the high
sulfur material. Materials were also characterized by optical microscopy. The low sulfur material
showed pitting along the grain boundaries, due to the boron concentration in this material, which
resulted in boron precipitates, while the high sulfur material showed pitting throughout the surface,
due to the MnS inclusions.
2
Electrochemical tests were also performed at room temperature on both materials in pH 4 (using
H 2 SO 4 ), 7, and 10 (using NH 40H). Peaks in the passive region of the high sulfur material were
seen at potentials of 160, 630, and 1400 mVsHE, due to dissolution of the MnS inclusions.
The results suggest that the high sulfur material provides an increase in corrosion when exposed
to the environment, which leads to the retardation of crack growth rates at the longer rise times
due to prolonged exposure of the crack tip to the environment. At low stress ratios, the proposed
mechanism for retardation of crack growth rates is crack tip closure, due to a buildup of corrosion
product at the fracture surface, which lowers the effective load that the crack tip experiences. At
high stress ratios, the proposed mechanism for retardation is an increased in injected vacancies and
enhanced creep, which disrupt the slip bands ahead of the crack tip, reducing the crack tip stresses.
Fractography of the fracture surface and crack growth rate comparisons of the low and high sulfur
material provide supportive evidence for the proposed mechanisms, and further work is proposed
to examine the effect of increased corrosion ahead of the crack tip.
Thesis Supervisor: Ronald G. Ballinger
Title: Professor of Nuclear Science and Engineering, Material Science and Engineering
3
Acknowledgements
First, I would like to express my sincerest appreciation to my advisor, Professor Ronald Ballinger.
Without his guidance, wisdom, and ideas, this project would not have been possible. I am
extremely grateful for his expertise, through which I have learned both inside and outside of the
classroom. He has brought new ideas and topics to the table throughout this project, progressing
the subject above and beyond its original form.
I would like to thank Dr. Denise Paraventi, my fellowship mentor at Bettis Atomic Power
Laboratory. The amount she has taught me through this process is immeasurable, and her guidance,
both professional and personal, has been above and beyond a mentorship role. From challenging
engineering concepts that she has illuminated, to difficult times that she has helped me overcome,
I cannot thank her enough.This project is a result of years of previous research and study performed
in part by Bill Mills. In the short amount of time I spent learning from him, I was able to develop
skills in both fractography and corrosion fatigue, and his wealth of knowledge, along with his
friendship and patience, will be sorely missed.
I would also like to thank everyone at Bettis Atomic Power Laboratory that has guided and helped
me through this process, including Bill Moshier, Tom Webb, Earl Johns, Bryan Miller, and Kevin
Fisher. I would also like to thank the countless other coworkers who have helped me prepare
specimens, navigate the lab, and learn various concepts. I look forward to working with all of them
in the future.
I would like to thank the other researchers I have had the pleasure of working with at MIT. I would
like to express my sincerest thanks to Dr. Yusaku Maruno, who not only helped significantly with
the electrochemistry portion of this project, but also helped greatly in all portions of the work that
was performed. His expertise and friendship has made this work successful. Special thanks to Pete
Stahle, without whom I would not have been able to complete this project. From start to finish, his
help in the lab has been constant, and the autoclave system relied heavily on his problem solving
in order to be successful. Additionally, I would like to thank Lun Yu, who I have had the pleasure
of working side by side with the past year, and who will be taking over this project as it evolves in
the coming future. Last but not least, my sincerest thanks to Dr. Michael Short for reading my
thesis and providing insight and help in the process.
4
I would like to thank Rachel Clark for her help as part of the Undergraduate Research
Opportunities Program at MIT. Her participation, both small and large, was an important part of
this work.
I would like to thank Van Christie at Altran Solutions, Inc., for his help in the fractography of the
high sulfur material that was studied.
Last but not least, I would like to thank my family, including my parents, David and Phylis
O'Brien, and my brother, Matthew O'Brien. They have been a solid foundation throughout my
education, and their encouragements have helped me overcome many difficult times.
This research project was performed as a collaboration between Bechtel Marine Propulsion
Corporation: Bettis Atomic Power Laboratory and MIT. Targeted, specific experiments were
performed at MIT to expand upon a large database of experimental data from Bettis. Microscopy,
experimental work, and knowledge were shared between both sites in order to complete this
research.
This research was performed under appointment to the Rickover Fellowship Program in Nuclear
Engineering sponsored by Naval Reactors Division of the U.S. Department of Energy.
5
Table of Contents
A b stract ............................................................................................................................................
2
Table of Figures ...............................................................................................................................
8
Chapter
-
Introduction.................................................................................................................16
1.1 Environm ental Degradation of N uclear M aterials ..........................................................
16
1.2 Enhancement and Retardation of Fatigue Crack Growth in Austenitic Stainless Steels ..... 17
1.3 Research Objectives ........................................................................................................
Chapter 2
-
22
Background.................................................................................................................24
2.1 Corrosion Fatigue.................................................................................................................24
2.2 A SM E A ir Curves ................................................................................................................
25
2.3 PW R Environm ent ...............................................................................................................
27
2.4 Environm ental Enhancem ent ..........................................................................................
29
2.4.1 Proposed Mechanisms for Environmental Acceleration.......................................29
2.4.2 W ater Flow Rate .......................................................................................................
32
2.4.3 Water Temperature ...............................................................................................
36
2.4.4 Rise Tim e..................................................................................................................41
2.4.5 Stress Ratio and AK .............................................................................................
2.5 Environm ental Crack Retardation....................................................................................
44
46
2.5.1 Proposed Mechanisms for Retardation in Corrosion Fatigue ...............................
47
2.5.2 Sulfur Content......................................................................................................
50
Chapter 3 - Experim ental...............................................................................................................53
3.1 A utoclave System .................................................................................................................
53
3.1.1 Autoclave Conditioning and pH Control ..................................................................
54
3.1.2 A utoclave Load Control System ...........................................................................
57
3.1.3 D ata Acquisition and Test Data .............................................................................
57
3.2 M aterial and Specim en Preparation .................................................................................
58
3.3 Electrochem istry...................................................................................................................59
3.4 M icroscopy...........................................................................................................................61
Chapter 4 - Fatigue Crack Growth Rate Test Results ...............................................................
4.1 Low Sulfur M aterial (H eat E5174-Specim en 5174-LR- 11) ................................................
4.1.1 Crack Growth Rate Testing .................................................................................
62
62
62
6
4.1.2 Fractography ........................................................................................................
66
4.1.3 Optical M icroscopy ...............................................................................................
76
4.2 High Sulfur Material (Heat A16830-Specimen A16-LR-10) ..........................................
78
4.2.1 Crack Growth Rate Testing .................................................................................
78
4.2.2 Fractography ........................................................................................................
82
4.2.3 Optical M icroscopy ...............................................................................................
91
Chapter 5 - Electrochem istry Results........................................................................................
93
5.1 Cyclic Polarization ...............................................................................................................
93
5.2 Optical M icroscopy (after Polarization Tests) .................................................................
96
5.3 Potentiostatic Tests ..............................................................................................................
96
Chapter 6 - D iscussion...................................................................................................................98
6.1 Crack G row th Rate Com parisons....................................................................................
98
6.2 Effect of Sulfur (w ith Respect to Proposed M echanism s).................................................104
Chapter 7 - Conclusions and Future W ork..................................................................................108
7.1 Conclusions for Proposed M echanism s .............................................................................
108
7.2 Future W ork Recom mendations.........................................................................................109
References....................................................................................................................................110
Appendix A -Load Cell Calibration............................................................................................112
Appendix B - Crack G rowth Rate D ata ......................................................................................
114
Appendix C -Electrochem istry D ata ...........................................................................................
121
7
Table of Figures
Figure 1: Time domain plot of crack growth rates in a PWR environment, shown for a range of
tem p eratu res. [2 ] ...........................................................................................................................
18
Figure 2: Crack growth rates as a function of rise time, shown for a range of stress ratio (R) and
A K . [2 ]...........................................................................................................................................
18
Figure 3: Crack growth rate given as a time domain plot with equivalent ASME air rates. Both
low sulfur heats, shown in red and yellow, have sulfur contents of 0.0005 wt%, and show fully
enhanced rates close to those predicted by the dotted line. The high sulfur heat has a sulfur
content of 0.034 wt%, and results are close to crack growth rates predicted by the ASME air
rates. The test was conducted under short rise times, with an R of 0.7, at 2500 C. [8]............. 20
Figure 4: Crack growth under SCC of stainless steel in high purity water. [3]........................
21
Figure 5: Fatigue Striations. Striations, when the distance between each striation is measured,
can often provide an estimate of the macroscopic crack growth rate. [3] ................................. 25
Figure 6: ASME air curves, predicting crack growth rates for room temperature and 2880 C at a
26
ran ge o f R v alu es. [4 ]....................................................................................................................
Figure 7: Evaluation of proposed crack growth rates from PWR curve with respect to
experimental data. Test parameters are given in the upper left corner. [10] ............................
28
Figure 8: Blitzkrieg process. Panel (a) shows the crack advance process. When a crack resistant
area is found, such as in (b), the crack proceeds around in the area, though more susceptible
areas. Once the crack overtakes the area, the crack can then grow along the edges of the crack
331
resistant area, due to environmental attack, as seen in (c). [3] ...................................................
Figure 9: Crack growth rate as a function of AK under various rise times and flow conditions.
Arrows denote crack growth rates that decreased or increased over time. Open symbols denote
low flow, while closed symbols are used for high flow conditions. Red squares are used for
tR=0.85 minutes, blue triangles for tR=8.5 minutes. The data designated by green diamonds was
produced after the data in red and blue are produced. The high flow pump was then shut off,
where the rates shown by red circles were produced. The high flow pump was then reinitiated,
producing the data shown as purple diamonds. [2]...................................................33
Figure 10: Time domain plot of low-alloy, high sulfur steel. Under low flow (denoted by
squares), crack growth rates are well predicted by previous methodology assuming full
environmental enhancement at the test temperature, at both the surface and deepest penetration
of the crack. Under high flow, denoted by triangles, these rates drop below both the prediction
for no environmnetal enhancement, as well as the ASME air rates. [11]................................... 355
8
Figure 11: Crack growth rate as a function of rise time for two temperatures. Heat B is a 304
stainless steel with sulfur content <0.01 wt%. For both temperatures, crack growth rate decreases
at a 500 second rise tim e. [7] ......................................................................................................
377
Figure 12: Fatigue crack growth surface in room temperature air, with irregular and rough
faceting, (a). Fatigue crack growth in 288"C air, with faceting more crystallographic in nature,
(b). Fatigue crack growth in 288"C, highly crystallographic, with obvious river patterning, (c),
L WR type w ater. [14] .................................................................................................................
399
Figure 13: Time domain plot of crack growth rates in a PWR environment, shown for a range of
temperatures. The black and red arrows denote locations of retardation in environmental crack
grow th rates. [2 ].......................................................................................................................
4 04 0
Figure 14: Crack growth rates as a function of AK and rise time
(tR).
2 ...............................
4141
Figure 15: Crack growth rates as a function of rise time for Heats A and B. Crack growth rates
are similar at short rise times, but deviate at longer rise times. [5] .......................................... 4242
Figure 16: Surface features for 50 s rise time (top of figure) and 1500 s rise time (bottom of
figure). The fan shaped lines in the bottom part of the figure are possible areas of crack tip
b lun tin g . [5 ].................................................................................................................................
43 3
Figure 17: Crack growth rate as a function of AK and stress ratio. [5]......................................
444
Figure 18: Crack growth rate effect of stress ratio and AK as a function of rise time. The overall
trend for the data is increasing crack growth with increasing rise time. [2].............................. 455
Figure 19: Time domain plot of crack growth rate as a function of stress ratio. [2] .................. 466
Figure 20: Low R crack growth rate as a function of AK. Certain crack growth rates for Heat
61115 (intermediate sulfur) are retarded below the predicted values for full environmental
enh an cem ent. [3]...........................................................................................................................
48
Figure 21: Oxides from mating surfaces can be "smashed" together. [3] ................................
49
Figure 22: IVEC mechanism. On the left, a fully enhanced crack tip is shown, having planar
slip bands. On the right, a retarded crack tip is shown, as suggested by the IVEC mechanism,
where the planar slip bands are disrupted, and the crack tip stresses are reduced. [3].............. 50
Figure 23: Crack growth rate as a function of AK, for R of 0.1, 0.3. Green data points are high
sulfur material (Heat A16830), red data points are low sulfur material (Heat 42322). [3] ..... 51
Figure 24: Potential-pH diagram for MnS in LWR environment. Estimated crack tip pH and
potential values are show n in red. [17].....................................................................................
52
Figure 25: Schematic of waterboard used for autoclave testing...............................................
55
9
Figure 26: Images of waterboard and pump on left. On the right, the fully closed fatigue machine
56
is sh ow n . .......................................................................................................................................
Figure 27: On the left, the specimen setup is shown within the fatigue machine. On the right is an
56
example of the computer software used to run the fatigue test..................................................
Figure 28: Specim en orientation from ASTM E3.........................................................................
59
Figure 29: Electrochemistry specimen (front and side views)..................................................
60
Figure 30: Gamry electrochemical cell and electrode arrangements used for electrochemical
m easu rem en ts................................................................................................................................
60
Figure 31: Crack length versus time for low sulfur test specimen (Heat E5174).....................
63
Figure 32: Sectioning of specimens for further analysis. Cuts were made using wire EDM
(copper w ire in a w ater bath). ....................................................................................................
64
Figure 33: Side measurement of the crack from one of the outside pieces of specimen.......... 64
Figure 34: Measurements of the crack length, as measured from the notch, along the specimen.65
Figure 35: Corrected crack growth rate values as a function of cycles for the low sulfur test
specim en (H eat E5174)...................................................................................................
66
Figure 36: Crack surface from precrack (A), to Step 1 (5.1s tR, B), to Step 2 (51s tR, C). Image is
at 300x magnification. Blue marks show approximate changes in steps.................................. 67
Figure 37: Crack surface from Step 2 (51s tR, C), to stop Step 3 (51 Os tR, D), to Step 4 (5100s tR,
67
E), to the fatigue apart (F).............................................................................................................
Figure 38: Step 1, 5.1 s rise time. Arrow A shows river patterns. Arrows B and C show areas
where the cracking occurred perpendicular to the crack growth direction, which is possible
68
evidence for a B litzkrieg attack. ................................................................................................
Figure 39: High magnification (1200x) of Step 1 (5.1 s
tR)........................................
Figure 40: Step 2 (51 s tR). Yellow arrow marks possible DSA event. .....................................
69
70
Figure 41: Transition between Step 2 (51 s tR) and Step 3 (510 S tR), with transistion shown by
the blue line. The large particles on the right are contaminants from the EDM process...... 71
Figure 42: Steps 2 and 3 (51s and 51 Os tR). Arrow 1 shows an example of slip offsets, while
arrow s 2 and 3 show striations....................................................................................................
Figure 43: Step 3 (51 Os
tR).
72
River patterns run parallel to the crack growth direction.............. 73
10
Figure 44: Step 3 (5 1Os tR) and Step 4 (5 100s tR). Step 4 starts at arrow 1, and the fatigue apart
74
starts at arrow 2 .............................................................................................................................
Figure 45: Step 4 (5 100s tR). Arrow 1 shows an example of a grain boundary. Arrow 2 shows
fatigue striations. The spacings were within a factor of 2 of the macroscopic crack growth rate
75
fo r S tep 4 .......................................................................................................................................
Figure 46: Microstructure of the low sulfur test material (E5174)...........................................
76
Figure 47: Microstructure of the low sulfur material (2739-LR-2)..........................................
77
Figure 48: Crack growth rate versus time for high sulfur material (Heat A 16830).................80
Figure 49: Crack growth versus cycle for high sulfur material (Heat A 16830).......................
81
Figure 50: DE correction for high sulfur material Heat A 16830...................................82
Figure 51: Entire fracture surface, with precrack (A), fatigue crack (B), Step 9 (51 00s tR, C) and
83
fatigue apart (D ). ..........................................................................................
Figure 52: Example of fracture surface, which was heavily oxidized......................................
83
Figure 53: Fatigue surface covered in heavy layer of oxide......................................................
84
Figure 54: Example of fatigue striations on the fracture surface...............................................
85
Figure 55: Example of fatigue striations on the fracture surface. Striations were about
3 .2 p m ...................................................................................................
. . 86
Figure 56: Fracture surface for the low stress ratio (R=04) portion of the test...................87
Figure 57: Fracture surface for the high stress ratio (R=0.7) portion of the test. .....................
88
Figure 58: Three steps (changes in rise times) of the fatigue test. Arrows show the transition in
89
rise tim es on the fracture surface. ..............................................................................................
Figure 59: Fracture surface, with areas of crack arrest (A), heavy oxidation, and locations of
Mn S disso lution (B , C ). ...............................................................................................................................
90
Figure 60: Fracture surface, with heavy oxidation and locations of MnS dissolution. ............ 91
Figure 61: Optical microscopy for the high sulfur material (Heat A16830). ...............................
92
Figure 62: Cyclic polarization of the low and high sulfur materials. Peak in the passive region of
94
the low sulfur material are most obvious at this pH of 10. ..............................................................
11
Figure 63: Cyclic polarization of the low and high sulfur materials. Peaks in the passive region of
95
the high sulfur material are more obvious. Environment at pH 7. ..................................................
Figure 64: Cyclic polarization of the three materials. Peaks in the passive region in the high
sulfur material are not seen in the low sulfur materials. Environment at pH 4. ..........................
95
Figure 65: Surface appearance of high and low sulfur materials after potentiostatic tests. ....... 96
Figure 66: Surface features after potentiostatic testing of high sulfur heat (Heat A 16830). Small
. 97
pits are seen at 500x m agnification . ....................................................................................................
Figure 67: Crack growth rate equations for air (ASME) and LWR-type environments. ........... 99
Figure 68: Crack growth rates as a function of AK with JSME, ASME, and experimental data for
100
the low sulfur m aterial. Stress ratio is R=0.4..............................................................................
Figure 69: Crack growth rates as a function of AK with JSME, ASME, and experimental data for
10 1
the high sulfur material. Stress ratio is R=0.4. ......................................................................................
Figure 70: Crack growth rates as a function of AK with JSME, ASME, and experimental data for
102
the high sulfur m aterial. Stress ratio is R=0.7. .....................................................................................
Figure 71 : Crack growth rates as a function of rise time for low and high sulfur materials......103
Figure 72: Comparison between high (A) sulfur and low (B) sulfur materials at R=0.4, tR >5I Os.
The arrow in A shows an example of an area where the corrosion product appeared damaged,
due to crack closure. The change in appearance of the oxides, in both shape and amount, is
104
ob v iou s betw een the tw o im ages. ...........................................................................................................
Figure 73: Comparison between high (A) sulfur and low (B) sulfur materials at R=0. 1, 0.3.
Features are sim ilar to those in Figure 72. [3]..........................................................105
Figure 74: Step 9 (5100 s tR) of the high sulfur specimen. There is less cleavage faceting, and the
fracture features are nondescript, as compared to fractrography from shorter rise times in Chapter
1 06
4 . ....................................................................................................................................................................
Figure 75: Intermediate sulfur (Heat 61115) material. The arrow denotes a 600 s rise time, which
produced retarded crack growth rates, and appears nondescript. [3] ............................................... 107
Figure 76: Applied Load Readings from Model #661.21A-0.3; Serial #3245...............................112
Figure 77: Comparison of load vs. voltage between Instron load cell, left, and WMC-5000
c alibra te d lo ad c e ll, rig h t. ........................................................................................................................
1 12
12
Figure 78: Applied Load Readings from Model #1020AJ-25K-B; Serial #186674A.......113
Figure 79: Comparison of load vs. voltage between Instron load cell, left, and WMC-5000
c alib rate d lo a d c e ll, rig h t. .........................................................................................................................
1 13
Figure 80: Crack length vs. time for Step #1 (5.1 second rise time) for the low sulfur heat. ...... 114
Figure 81: Crack length vs. time for Step #2 (51 second rise time) for the low sulfur heat. ....... 114
Figure 82: Crack length vs. time for Step #3 (510 second rise time) for the low sulfur heat. ..... 115
Figure 83: Crack length vs. time for Step #4 (5100 second rise time) for the low sulfur heat.
1 15
......................................................................................................................................................................
Figure 84: Crack length vs. time for Step #1 (5.1 second rise time) for the high sulfur heat.
.............................................................
...
. ............................................................................................
1 16
Figure 85:. Crack length vs. time for Step #2 (51 second rise time) for the high sulfur heat.
.........................................................................................................................................................................
1 16
Figure 86: Crack length vs. time for Step #3-1st Section (510 second rise time) for the high
sulfur heat. ......................................................................................
117
Figure 87: Crack length vs. time for Step #3 -2nd Section (510 second rise time) for the high
117
sulfur heat. ......................................................................................
Figure 88: Crack length vs. time for Step #4 (5.1 second rise time) for the high sulfur heat.......118
Figure 89: Crack length vs. time for Step #5 (5100 second rise time) for the high sulfur heat... 118
Figure 90: Crack length vs. time for Step #6 (5.1 second rise time) for the high sulfur heat.
119
Figure 91: Crack length vs. time for Step #7 (51 second rise time) for the high sulfur heat. ...... 119
Figure 92: Crack length vs. time for Step #8 (5100 second rise time) for the high sulfur heat....120
Figure 93: Crack length vs. time for Step #9 (5100 second rise time) for the high sulfur heat....120
Figure 94: Example of Pre-Treatment for Potentiostatic Specimens. ...............................................
121
Figure 95: Results of Potentiostatic Test at 220mVsHE for High S Specimens in 0.1N Na2 S04
S o lutio n (p H 7 ) . ...........................................................................................................................................
121
13
Figure 96: Results of Potentiostatic Test at 650mVSHE for High S Specimens in 0.1N Na2 S04
S o lutio n (p H 7 ). ..........................................................................................................................................
1 22
14
List of Abbreviations and Symbols
APW
Aerated Pressurized Water
BWR
Boiling Water Reactor
CT
Compact Tension
da/dN
Crack growth/Cycle
DPW
Deaerated Pressurized Water
DSA
Dynamic Strain Aging
EDM
Electrical Discharge Machining
EDX
Energy-Dispersive X-ray Spectroscopy
EAC
Environmentally Assisted Cracking
FCGR
Fatigue Crack Growth Rate
HIC
Hydrogen Induced Cracking
LWR
Light Water Reactor
MnS
Manganese Sulfide
Kmax
Maximum Stress Intensity Factor
PWR
Pressurized Water Reactor
AK
Range in Stress Intensity Factor, Kmax-Kmin
tR
Rise Time
SEM
Scanning Electron Microscopy
SCC
Stress Corrosion Cracking
R
Stress Ratio, Kmin/Kmax
15
Chapter 1 - Introduction
1.1 Environmental Degradation of Nuclear Materials
Material performance is often a limiting factor for nuclear power plant design and
operations; parameters such as peak fuel temperature, heat transfer, performance during transients,
and neutron flux, to name a few, set limits on material performance. Material selection is a critical
part of the design process to ensure safe and economic operation. In a corrosive environment, such
as that found in a water cooled nuclear reactor, component lifetimes are often estimated using a
combination of experimental data and conservative estimates. However, component behavior
estimates are often still uncertain, resulting in parts that either fail before their estimated lifetime
(often from a misunderstanding of how the material behaves in the chosen environment) or that
require replacement due to manifestations of degradation phenomenon unknown at the design
stage, or estimates that are too conservative, resulting in economical poor operations (unnecessary
inspections, etc.). Both of these outcomes are costly, but in many cases could have been mitigated
by a better understanding of the mechanisms underlying material response in the environment.
Environmental degradation in the form of stress corrosion cracking, hydrogen
embrittlement, corrosion fatigue, or a combination of these, is often the limiting factors for material
performance in a Light Water Reactor (LWR). Specifically, austenitic stainless steels, which are
used extensively in Pressurized Water Reactor (PWR) and Boiling Water Reactor (BWR) system
components, are susceptible to corrosion fatigue crack growth when exposed to the LWR
environment. The extent of cracking is dependent on the material chemistry, water chemistry
(including temperature), and mechanical loading conditions. For instance, laboratory testing of
stainless steels in LWR environments both in fatigue life (S-N) testing and fatigue crack growth
testing have shown reduced cycles to failure and increases in measured crack growth rates by
factors of up to 100 over those in air. [1,2,3] As a result, the industry has been working to revise
existing S-N and fatigue crack growth rate (FCGR) models to account for the environmental
enhancement measured in the laboratory. In order to compensate for the decrease in component
lifetime that occurs due to corrosion fatigue, conservative factors are applied to the air fatigue
crack growth curves from Section XI of the ASME Boiler and Pressure Vessel Code (referred to
as the "ASME air rates"), [4] based on experimental data in elevated temperature water. These
16
conservative factors lead to reduced calculated lifetimes and potentially increased inspection
intervals, increasing cost. While emphasis in recent years has been on fatigue life, there has also
been work on factoring environmental effects on the crack growth rate curves. The focus of this
work is on the crack growth rate behavior in high temperature water as a function of material
chemistry.
1.2 Enhancement and Retardation of Fatigue Crack Growth in Austenitic
Stainless Steels
Experimental data and lifetime performance in LWRs of stainless steels show that fatigue
crack growth is greatly accelerated in a PWR environment compared to air. [5] On the primary
side, a PWR operates at 15.2 MPa and at about 2870 C to 324'C, with deaerated, high purity water
(< 5ppb 02). In high temperature, deaerated water, the environmental enhancement of the crack
growth rates increases with increasing temperature, well over the rates predicted by the ASME air
curves. Figure 1 shows an example of environmental enhancement of FCGRs for a low sulfur
(0.001 wt%) austenitic stainless steel in PWR water. [2] Shown in Figure 1 are the rates predicted
by the ASME curve in air, compared to the measured rates in high temperature water, with the
different temperatures denoted by the symbols. The dashed line represents the one-to-one line
where the air rates equal the high temperature water rates. Deviations from the one-to-one line
(above the line) indicate environmental enhancement. As the temperature is increased, crack
growth rates also increase, further deviating from the ASME air rates. Experimental crack growth
rates vary in the extent of enhancement, due to the effects of test conditions including the stress
ratio R (ratio of minimum to maximum load, P), rise time (the length of time the load is increasing
in the total cycle), and the range in stress intensity factor, AK. [2] Figure 2 shows an example of
the effect of rise time on crack growth rates during trapezoidal loading as a function of rise time
and stress ratio for austenitic stainless steel. It is observed that the crack growth rates increase with
increasing rise time in most cases, indicating increases in environmental enhancement at longer
the rise times.
17
i.OE-07
PWRwaO
R-0, 7 -0 8
300C
'I
A
A2W0L
250~~
*02=U
0
....
...
. ..
. ...
A'
-
1.0E-12
1.OE-14
1.OE-13
1.OE-12
1.0E-09
1.OE-10
1.0E-11
1.OE-08
dWt* (sk) U S
Figure 1: Time domain plot of crack growth rates in a PWR environment, shown for a range of
temperatures. [2]
1.OE05
T
AK vakue in MPaan
I .OE06
a ",4 1. Deft K 41*
a R-03. Do". K labS
OR-01 DeaK-11
.O
-
-
I.OE-0
ER-C.7 De.a X 12-15
R-as Do"K 1-10
0
*R-00
*R.0.
Do"ta9b4
DWI& K 4-
1.OE-09
1
10
1000
100
Me dw I sws
10000
10
Figure 2: Crack growth rates as a function of rise time, shown for a range of stress ratio (R) and
AK. [2]
While there is a large body of laboratory data showing significant environmental
enhancement in the literature [6], recent research has shown that corrosion FCGRs in austenitic
stainless steels in PWR type environments do not always exhibit environmentally accelerated rates.
This phenomenon is referred to as crack growth retardation, meaning the crack growth rates are
rates
reduced relative to the "normally" environmentally enhanced rates. In some cases, retarded
at
have been observed equal to those measured in air. [3] Retardation effects have been observed
18
high temperatures (around 300'C, shown in Figure 1 at 300'C) [2], certain material chemistries
(high sulfur) [7], specific loading conditions (high R) [2,7], and at longer rise times under
trapezoidal loading. [8]
The focus of this thesis is on the effect of material chemistry on crack growth rates,
specifically the sulfur content, which has been shown to slow environmentally enhanced crack
growth rates in austenitic stainless steels with high sulfur content (but still within the ASTM
specifications). In this thesis, compositional ranges of sulfur chemistry are defined in Table 1. The
effect of material sulfur content on environmental cracking has been observed in both corrosion
fatigue (under long cycle times), as shown in Figure 3, and stress corrosion cracking (SCC), as
shown in Figure 4. The results for the high sulfur heat shown in Figure 3 show retardation of the
crack growth rates compared to those predicted by the ASME air curve. This is shown by the blue
data points, designating a high sulfur material, which fall along the ASME air curve prediction,
whereas the red and yellow data points, designated for low sulfur materials, are at higher rates,
close to those predicted for environmentally enhanced crack growth rates. The test in Figure 3 was
a rise time test, with rise times from 5.1 s to 600s. In SCC tests, the high sulfur heat crack growth
rates, shown in Figure 4 as the dark blue data, appear to fall to rates that are below measureable at
around 2000 hours. Tests were performed with sensitized and cold worked with specimens with
low (Heat ECT-57) and high (Heat GCT-40) sulfur in aerated and deaerated pressurized water
(APW and DPW). The specimens were initially in the environment at 249"C with Ippm 02 (APW)
for 1700 hours. The environment was then degassed and hydrogen added to obtain (30cc/kg) DPW
conditions. The first 1891 hours was performed with a load cycle of 9000s, and then held at
constant load after until 3019 hours, when unload and load cycles of 12 (at 3370 hours) and 24 (at
3845 hours) hours were introduced. Finally, constant load was applied at 4152 hours. The
corrosion potential for 304 stainless steel is shown in purple at the top of the figure. SCC tests are
relevant to corrosion fatigue at long rise times because the low frequency of the loading pattern in
fatigue approaches static conditions. The environmental effect is increased in SCC and longer rise
time fatigue tests, compared to shorter rise times.
19
Table 1: Compositional ranges for sulfur in austenitic stainless steel.
Designation
Low sulfur chemistry
Intermediate sulfur chemistry
High sulfur chemistry
Compositional Range
0.001-0.002 wt%
0.005-0.0012 wt%
0.030-0.015 wt%,
AK - 5MPa m, R=0.7
(Blue-High Sulfur) fully enhanced, (Red, Yellow-Low Sulfur) nonenhanced rates throughout
1
~
4.
A
*
01
I WA
SE WOA )W
Al wh~aA JL
a JJ1IFA
I WF-u
1 OF-14
A
ME )a crack g#me tiso mas
Figure 3: Crack growth rate given as a time domain plot with equivalent ASME air rates. Both
low sulfur heats, shown in red and yellow, have sulfur contents of 0.0005 wt%, and show fully
enhanced rates close to those predicted by the dotted line. The high sulfur heat has a sulfur content
of 0.034 wt%, and results are close to crack growth rates predicted by the ASME air rates. The
test was conducted under short rise times (5.1s to 210s), with an R of 0.7, at 250*C. [8]
20
1
I 249'C Nonfaulted APW/DPW
(
1(5x1O )0
5'x10
(6x10')
200
-200 IE
-400
-600
-800 W
L
21
20
"
(1
1x104)
xeo)
29.9
MPa m
E
0x0'
z
27.7
MPavm
20
(datdt=5x006 minis)
S
19
.
(4x1
I
4
(6x10 )
18
(da/dt=2x10' mmnis)
0
1000
25.7
MPa8m
__
(6x10 )f(9x10)
C
o
2000
ECT-47 (R-L, 10%CW, 0.001%S)
GCT-31 (R-L, 10%CW, 0.026%S)
3000
4000
24.0
MPa m
5000
TIME, hours
Figure 4: Crack growth under SCC of stainless steel in high purity water. [3]
The retardation effect of high sulfur content is not observed in all types of steels, and the
increase in the corrosive environment as the MnS particles (which are prevalent in high sulfur
steels) dissolve often has a negative effect on component lifetime. This is the case with ferritic
(low alloy) steels, where a high sulfur content will lead to an enhancement in environmentally
assisted cracking. [9] Nonetheless, it could be beneficial to lifetime analyses if reduced FCGRs
can be justified for austenitic stainless steels in specific environments. However, in order to
account for the retardation effects that are seen in austenitic stainless steels on the prediction of
crack growth rates, a better understanding of the underlying mechanism is necessary.
21
1.3 Research Objectives
The goal of this thesis is to
*
Examine the effect of sulfur on the corrosion fatigue behavior of austenitic stainless steels
in high pH LWR-like environments, and
" Provide additional insight into the likely mechanisms controlling retardation of fatigue
crack growth in higher sulfur material.
To accomplish the above goals, the areas of focus will include FCGR testing and
electrochemical characterization of low and high sulfur containing 304/304L stainless steel heats
to fully quantify the effect of high sulfur on fatigue crack growth. Fractographic analysis including
Scanning Electron Microscopy (SEM) and metallographic analysis of the fracture surfaces were
also performed. The discussion below further expands upon the tasks that will be used to
accomplish the goals.
1. Materials & Characterization: Two low carbon (304L) heats of stainless steels were used
in the program. Their chemistries are provided in Table 2. Material included a low sulfur
material (Heat E5174, used to develop techniques and analysis) and a high sulfur material
(Heat A16830). The high sulfur material was at the maximum allowed by the ASTM
specification
(A666,
0.030 wt%).
Materials
were characterized
using standard
metallographic as well as electrochemical techniques.
Special Note: The low sulfur heat had a higher than normal, but still within specification, boron
content of 18ppm. This resulted in the precipitation of (Cr,Fe)2B-type borides which may have
affected the corrosion behavior. However, there was no observed effect on the FCGR when
compared with other lower sulfur material from the literature.
2. Corrosion FCGR testing at LWR conditions (288'C, pH 10, 15.2 MPa). These conditions
are similar to an LWR environment, but with a higher pH of 10.7 (tests were run at pH
values of 10 and 10.7). Fatigue testing was conducted using a trapezoidal wave form with
variable "rise times" and "fall times". The effect of rise times of 5.1, 51, 510, and 5100
seconds (with fall times of 0.9, 9, 90, and 900 seconds, respectively) was explored.
22
Heat
Table 2: Material Chemistries (wt%)
Ni
Cr
C
Mn S
N
Fe
E5174 (Low Sulfur)
0.02
1.6
<0.0025
19.8
10.1
0.085
Bal
A16830 (High Sulfur)
0.021
1.73
0.032
18.5
8.2
0.089
Bal
3. Post-Test Characterization: Post-test analysis of fracture surfaces using SEM and other
metallographic techniques to provide insights into the nature of the processes involved in
controlling behavior.
23
Chapter 2 - Background
2.1 Corrosion Fatigue
Corrosion fatigue is the result of the combined effects of cyclic stresses, a susceptible
material, and a corrosive environment. [9] Cyclic stresses during corrosion fatigue are defined by
parameters such as frequency of the loading cycle, stress ratio (the ratio of the minimum load to
the maximum load, R), cyclic stress intensity (AK), and rise time of the cycle for trapezoidal
loading (tR). Failure of materials due to corrosion fatigue processes occurs at much lower stresses
than failure due to static loads, or failure due to cyclic loads in air. Furthermore, crack initiation
and growth can occur even faster when the material contains stress risers, such as notches, which
increase local stresses and susceptibility to corrosion fatigue. Both alloys and pure metals are
susceptible to corrosion fatigue. [9]
Failure due to corrosion fatigue, rather than other environmental cracking phenomena such
as SCC and hydrogen induced cracking (HIC), can be confirmed, at least in part, through study of
the fracture surface. Several surface features offer clues as to the nature of the crack growth
mechanism, such as crack growth rate, crack arrest, and crack propagation direction. Crack growth
in fatigue is often transgranular in nature (both SCC and HIC can be either transgranular or
intergranular), and the crack tip will appear blunted. Striations, a result of the crack opening in
increments as the load is applied cyclically, are unique to corrosion fatigue, and can be used to
determine the macroscopic da/dN (crack growth per cycle, or change in crack length with cycle).
Figure 5 shows an example of clearly defined striations, which appear as periodic ripples on the
crack surface. Striations appear perpendicular to the cracking direction. Crack arrest, or a change
in crack growth in which the crack propagation rate is changed from one rate to another, can result
24
in a macroscopic "beach mark" on the surface due to the buildup of corrosion products or a change
in the character of the propagation process. [3,8]
Figure 5: Fatigue Striations. Striations, when the distance between each striation is measured,
can often provide an estimate of the macroscopic crack growth rate. [3]
2.2 ASME Air Curves
To account for fatigue crack growth in design and performance of stainless steels at a given
stress ratio and temperature in air, the ASME air curve can be used. [4] The ASME air rates as a
function of applied AK for austenitic stainless steels are shown in Figure 6, for multiple stress
ratios for room temperature (21"C) and a higher temperature (288"C), such as those temperatures
seen in an LWR environment. Crack growth rate curves are plotted in inches per cycle as a function
of the cyclic stress intensity factor, AK. To predict corrosion fatigue crack growth in an
environment, experimental data is used to determine conservative factors that can be applied to
ASME reference curves in air, which can then be used to predict component lifetime. Although
based on experimental results, these predictions still contain considerable uncertainty due to the
lack of mechanistic understanding of the environmental processes that can enhance or retard the
crack growth rates; these include varying load conditions during operation, material chemistry, or
environmental factors, and can greatly affect the crack growth rate. Enhancement of crack growth
rates above the air curve can be as high as 100 times the predicted value [3], while retardation
under some loading and material conditions can produce environmental crack growth rates that
approach those values observed in air.
25
It,
j
~-t--T
A
r.
1J.
Li
r
-~
1-A'
Son
II
104
WS'
1I
*
111271A:
r
U
I
rA.'Ii
1111I11
Ill,
F~U4 1sN&W
I
C-28
U IIvv
I-
104
1
2
VIA
9
_
OSUMM 9"
-t
I*
go
E-
Figure 6: ASME air curves, predicting crack growth rates for room temperature and 288"C at a
range of R values. [4]
26
2.3 PWR environment
In the United States, the most common LWR type is the PWR, which is comprised of a
primary coolant system which operates at a pressure of 15.2 MPa to maintain a single phase in the
water that is used for cooling and neutron moderation, and a secondary heat transfer side (via a
steam generator) that operates at a maximum pressure of approximately 6.9 MPa. PWRs are
markedly different from another, less common LWR, the BWR, which operates using a direct
cycle. The two systems differ in such areas as water chemistry and operating conditions. PWR
primary side environments contain several important features that can play a large role in materials
performance, especially environmentally enhanced fatigue crack growth and stress corrosion
cracking (SCC). PWR coolant and moderator temperature ranges from about 2870 C to 324'C on
the primary side. Plants operate with a hydrogen overpressure, and the 02 concentration is thus
maintained at a minimum (less than 5 ppb), so PWR environments are much less oxidizing than
BWR environments, which contain small amounts of oxygen due to air in-leakage as well as water
radiolysis. This difference results in different performance of stainless steels in the two
environments. Nomura et al. discussed the inappropriateness of the use of BWR fatigue crack
growth curves that were prescribed by the Japan Society of Mechanical Engineers (JSME) in 2003
for austenitic stainless steels. [10] Due to the marked differences in water chemistry, as well as the
lack of consideration for temperature effects in the BWR reference curves, these curves were
judged to not be representative of PWR conditions. Instead, PWR curves were recommended based
on calculations considering only data generated in PWR conditions, considering variations in
temperature, stress ratio, and loading frequency. Material chemistry was not generally varied, or
accounted for, during experiments. Environmental temperatures tested were between 1 000 C and
3500C. The stress ratio, R, was varied from 0.1 up to 0.95. The data, along with the proposed PWR
curve, is shown in Figure 7. The application of a factor of 2.7 multiplied by the reference curve
gives a conservative estimate of FCGRs for almost all of the experimental data. The general forms
of both the BWR and PWR corrosion fatigue crack growth curve equations are given in Equations
1 and 2, respectively, where C", n, m' and m are constants that depend on the environment, such
as temperature (as given, there is no temperature dependence in the BWR reference curves). [10]
27
a = C-
T r-
AK"'/(1 - R)"
dN
1
da = 1.61X10~10 -T 0.63 -Tr O.33 - AK 3.0/(1 _ R) 1.56
dN
2
1.1-04
d.MHw4 35N1O'4AK'
231
3
n
.1**
a/
I
t
II
I
I
1.-06
......
dam". 1.611
01A Aa
.5-07
to
t0
&K (MPaf a)
Figure 7: Evaluation of proposed crack growth rates from PWR curve with respect to experimental
data. Test parameters are given in the upper left corner. [10]
28
2.4 Environmental Enhancement
It has been suggested that environmentally assisted crack growth is promoted by a critical
concentration of dissolved sulfur at the crack tip, as studied by James et al. on low alloy steels.
[11] This sulfur is a product of the dissolution in high temperature water of MnS inclusions that
exist in low-alloy as well as stainless steels, which are detrimental to the corrosion fatigue
properties of low-alloy steels; this is not necessarily true for stainless steels, as is discussed earlier.
Since conditions at the crack tip can promote dissolution of the MnS inclusions one would expect
that aqueous conditions (temperature, electrochemical potential, transport of species by diffusion,
ion migration, convection) will have a strong effect on behavior. Furthermore, the combination of
environmental conditions and mechanical loading conditions leads to a large degree of variation
in crack growth rate behavior. It is important to note that the mechanisms for crack growth are
different in low-alloy and stainless steels. Low-alloy steels experience hydrogen embrittlement
ahead of the crack tip, whereas stainless steel crack growth is largely dependent on the state of the
slip bands ahead of the crack. The potential effect of all variables is discussed below.
2.4.1 Proposed Mechanisms for Environmental Acceleration
The fracture surface for low AK regimes and high AK regimes have different appearances;
at AK values below approximately 13 MPalm, crystallographic faceting dominates, whereas a
combination of crystallographic faceting and fatigue striations is more dominant at higher values.
The change in fracture morphology is hypothesized to be a result of differences in slip behavior
between a large plastic zone that develops at the maximum load and a smaller plastic zone that
develops during reversed cyclic loading. The existence of both crystallographic faceting and
fatigue striations at higher AK suggests that there are two slip modes occurring in the different
regions of the plastic zone. The largest zone, the cyclic plastic zone, which expands ahead of the
crack tip, experiences heterogeneous slip: planar slip bands are weakened by heterogeneous slip
mechanisms as hydrogen is diffused ahead of the crack, creating a path for faceting to occur. In
the smaller, fully reversed plastic zone, homogenous slip is responsible for the creation of the
striation formation. Crack growth at AK above 13 MPalm therefore results in striations that occur
in addition to or within the crystallographic facets. Under low AK testing, both plastic zones
29
undergo the same heterogeneous slip. Mills also noted faceting that was oriented normal to the
direction of crack propagation, the occurrence of which is described by him as "blitzkrieg" attack,
because susceptible regions will grow quickly around crack resistant regions; the resistant region
eventually will be "attacked" by cracking orthogonal to the primary growth direction as a result of
high stresses developing in the uncracked region. This attack occurs at both low and high AK
regimes, and allows for the crack growth to continue without arresting. Mills suggested that this
process is due the presence of crack resistant areas, which the advancing crack will grow adjacent
to, along less resistant regions. The edges of the crack resistant region will then undergo
environmental attack, and the crack will grow along the more susceptible direction. The local crack
growth direction will then appear perpendicular to the macroscopic crack growth direction. Figure
8 shows a schematic of the blitzkrieg attack, which is thought to contribute to the environmentally
enhanced cracks growth rates, due to the growth of the crack not being restricted by these crack
resistant regions. [3]
30
Figure 88: Blitzkrieg process. Panel (a) shows the crack advance process. When a crack resistant
area is found, such as in (b), the crack proceeds around in the area, though more susceptible areas.
Once the crack overtakes the area, the crack can then grow along the edges of the crack resistant
area, due to environmental attack, as seen in (c). [3]
During corrosion fatigue, an additional process of dynamic strain aging (DSA) occurs in
materials exhibiting both environmentally enhanced and retarded FCGRs. DSA is a phenomenon
observed during high temperature tensile deformation, and manifests itself as "serrated yielding"
in tensile tests. Interactions between dislocations and carbon and other interstitial atoms in the
material result in temporary "pinning" of dislocations. Continued straining breaks the dislocations
free from the carbon atmospheres; with sufficient diffusion the carbon moves to the dislocations
again, resulting again in dislocation pinning. The pinning and unpinning of dislocations by carbon
results in small load increases and decreases during deformation (the serrated yielding observed in
the plastic regime of a tensile test). [12] During fatigue crack growth, DSA can occur and cause
31
the crack growth to experience bursts of plastic deformation, and they can appear as "super"
striations the crack surface, although they do not correspond to the macroscopic crack growth rates.
Even though they may blunt the crack tip, as the crack is briefly arrested, they do not lead to a
change in macroscopic crack growth rate; the crack arrest is temporary, and is reinitiated easily
along the same path. Furthermore, DSA events appear in both low and high sulfur materials, and
therefore is not expected to be the mechanism leading to retardation of the environmentally
enhanced crack growth rate. [3]
2.4.2 Water Flow Rate
Sulfur can be removed via various types of mass transport (diffusion, convection, ion
migration), and therefore transport into and out of the crack tip can influence the crack growth
behavior. [11] In an LWR environment, the mass flow rate of the coolant can be as high as 20000
kg/s (Reynolds number of 500000), which can have a strong effect on chemistry within a crack tip
and the movement or removal of species and corrosion products within the crack face. [13] Tice
et al. modified fatigue crack growth rate testing for austenitic stainless steels in order to create a
turbulent, rather than quasi-static (which is generally assumed for corrosion fatigue testing), flow
rate environment around the compact tension (CT) specimen in order to study the effect of flow
rate on FCGR. His results are shown in Figure 9. Under turbulent/high flow (with a flow rate of
20 L/min, around a Reynolds number of 200000), the crack growth rates are about 2 times lower
than those in quasi-stagnant conditions. His test was performed at 2500 C, with an R of 0.7. Rise
times are reported on the figure in the upper left hand comer. For austenitic stainless steels with
low sulfur content, as was used in Figure 9, crack growth rates would be expected to increase with
increase in rise time. In terms of flow, the crack growth rates decrease in turbulent flow over the
quasi-static flow crack growth rates. [2]
32
1.OE-06
00.05min Low Flow
a .5min Low Flow
0 86mWn High Flow
&8.6min High Flow
* 51min High Flow
* 51min Low Flow
o 8.5 min Low Flow
* S1min Low Flow
* 61min High Flow
ASME XIA R-0
6
1
.E-9
Arrows show time
dependent changes
to growth rate
1.OE-08
1.0
10.0
100,0
K / MPUIm
Figure 9: Crack growth rate as a function of AK under various rise times and flow conditions.
Arrows denote crack growth rates that decreased or increased over time. Open symbols denote low
85
flow, while closed symbols are used for high flow conditions. Red squares are used for tR=0.
minutes, blue triangles for tR= 8 . 5 minutes. The data designated by green diamonds was produced
after the data in red and blue are produced. The high flow pump was then shut off, where the rates
shown by red circles were produced. The high flow pump was then reinitiated, producing the data
shown as purple diamonds. [2]
James et al. also investigated the mitigating effect of water flow rate, this time on SCC in
low alloy steel, which again may apply to FCGR testing at long rise times. [11, 14] He conducted
experiments using flow rates of 78.9 L/min, 27.0 L/min, and 8.69 L/min with corresponding
Reynolds numbers of 643000, 221000, and 72200, respectively. Materials used had high sulfur
33
contents, ranging from 0.024-0.027 sulfur (wt.%). [11] Specimens were machined to include a
semi-elliptical surface crack in order to provide easier mass transport. Test conditions were similar
to an LWR environment, with a temperature of 288"C and a pressure of 15.2 MPa. The results of
his experiments are shown in Figure 10. The highest flow rate shows a complete mitigation of
environmentally enhanced cracking, while the intermediate flow rate also shows some success in
mitigating environmental effects to well below the environmentally assisted cracking (EAC) line.
The lowest flow rate results demonstrate a lack of EAC mitigation, suggesting flushing of the crack
tip to remove the deleterious sulfur species led to migration of the cracking, which contributes to
environmental enhancement in low alloy steels. Water flow rate, therefore, may be one of the
largest contributors to the presence (or lack thereof) of EAC under the conditions of his
experiments, at least for the removal of harmful species, and when no other competing retardation
mechanism is at play (as discussed in Section 2.5). Since his experimental conditions were similar
to those in a PWR, this behavior could at least partially explain the general lack of environmentally
assisted fatigue cracking over a reactor's lifetime under actual plant conditions, especially if
conservative correction factors to account for EAC are applied.
34
1E-4
t5.4
4444
Ip
ommn
c.1
64640
"Msmebee
141
WIP
d
1546
TkmeSmd Ak Rim,
4
1646
144
mvoboend
Figure 10: Time domain plot of low-alloy, high sulfur steel. Under low flow (denoted by squares),
crack growth rates are well predicted by previous methodology assuming full environmental
enhancement at the test temperature, at both the surface and deepest penetration of the crack. Under
high flow, denoted by triangles, these rates drop below both the prediction for no environmnetal
enhancement, as well as the ASME air rates. [11]
James et al., in a second report, [14] also studied the effect of water flow rate on low-alloy
steel (high sulfur) that was overlaid with corrosion resistant cladding of Alloy EN82H (low sulfur),
in order to study the flow pattern changes, as well as galvanic effects that could impact transport
of dissolved sulfur from the crack tip. Flow velocities of 1.69 and 4.74 m/s were used,
corresponding to the intermediate and high velocities used in Reference 11. Unlike previous results
from James et al. [11], the high flow rate did not mitigate EAC with as much success as the unclad
specimens, and effects of the high flow were not seen until the very end of the test (around 2200
hours, only for a flow rate of 4.74 m/s). Computational Fluid Dynamics (CFD) modeling using the
FIDAP code (Fluid Dynamics International) demonstrated hindrance of flow to the crack tip region
due to the cladding, which further implies that mass transport at the crack tip is successful at
mitigating EAC. The CFD code was used to evaluate the mass transport that occurred within the
35
crack, accounting for the crack, the surface surrounding the crack, and the free stream flow. The
cladding provided a greater surface roughness on the surface of the crack, impeding the free flow
of species from the crack tip; sulfur species were not as easily removed as they had been in the
previously non-clad specimens. [14]
In austenitic stainless steel, Tice et al. [2] reported that the effect of flow rate is similar to
that of low-alloy steels, but disagreed that the environmental enhancement was due to sulfur
species. Due to the lower crack growth rates of stainless steel in oxygenated high temperature
water, in which sulfides remain at the crack tip due to the electrochemical potential gradient, Tice
suggested that sulfides were not the species responsible. He suggested that some other soluble
species was responsible.
2.4.3 Water Temperature
Wire et al. performed fatigue tests at 288'C and 243'C in order to measure the temperature
dependence of the FCGR of austenitic stainless steels. [7] His data was obtained at a temperature
of 288"C under a fully reversed stress ratio (R=-1) and at high stress ratios (R=0.7, 0.83), and had
produced rates that were as high as 20 times those in air. Experimental conditions were LWR-type,
with a pH from 10.1 to 10.3, and 02<20 ppb. Figure 11 shows the crack growth rate as a function
of rise time and temperature. Measurements taken at the lower temperature showed lower crack
growth rates at rise times of 5 and 50 seconds than the higher temperature, though the lower
temperature rates were still accelerated above the air data by 7-8 times. At the longest rise time of
5000 seconds, both temperatures produced similar crack growth rates. The crack growth rates at
288"C were lower at the 500 second rise time than at the 50 second rise time for this temperature,
suggesting a retardation mechanism. The possible explanation for this retardation is the increased
sulfur content in the material, discussed in Section 2.5.
36
10-1
o 2
A 24ft
:R4.7,1S1-13MPaan
8
U
0
Ak RoO.7. 8K1I I WAPdm
10
100
1000
Me Time, s
Figure 91: Crack growth rate as a function of rise time for two temperatures. Heat B is a 304
stainless steel with sulfur content <0.01 wt%. For both temperatures, crack growth rate decreases
at a 500 second rise time. [7]
In an attempt to qualify the effects of environmental enhancement and retardation, Wire
and Mills presented fractographic evidence to describe and characterize the different regimes of
cracking. [15] Figure 12 shows three fractographs of 304 stainless steel (low sulfur), for material
tested in room temperature air (24 0C), 288 0C air, and 288"C water (LWR-type). The images in
Figure 12 show features that exemplify the testing condition. Fatigue tests were performed under
different stress ratios and rise times, but the fractography is discussed only for temperature and
environment changes. At room temperature, facets were irregular, compared to the more
crystallographic nature of the facets in air at 288"C. Since the temperature for martensitic
transformation is 100"C, the test in air at 2880 C is above the temperature for martensite to form
upon cold working, which explains the lack of quasi-cleavage faceting in the higher temperature,
and there are clear river patterns on the specimen face following the crack growth direction. [15]
Under high temperature water conditions, facets are crystallographic, as opposed to rough and
irregular in air. The river patterns are even more visible and crisp in high temperature water (as
37
opposed to 288'C air), and are observed for all environmentally enhanced crack growth rates.
Furthermore, for the high temperature water conditions, the crack paths (direction) in specific areas
can deviate from the overall crack path, which is a result of specific regions of susceptibility and
hydrogen embrittlement that occurs in water, and the authors suggested that these regions
contributed to the enhanced rates observed in this environment. Susceptibility is relative; resistant
ligaments in the material will hinder crack growth in that direction, forcing the crack to locally
grow in less resistant directions. Lastly, the appearance of cleavage-like sharp facets at high
temperature (in water) suggests a hydrogen embrittlement mechanism, due to absorption of
hydrogen from the environment into the crack tip. [15]
38
Figure 12: Fatigue crack growth surface in room temperature air, with irregular and rough faceting,
(a). Fatigue crack growth in 288C air, with faceting more crystallographic in nature, (b). Fatigue
crack growth in 2880 C, highly crystallographic, with obvious river patterning, (c), LWR type
water. [15]
39
In addition to the previously discussed research of Wire and Mills [15] regarding
temperature effects, Tice et al. also investigated the influence of four different temperatures on
fatigue crack growth in a PWR environment. [2] Their results are shown in Figure 13. The results
show an increase in crack growth rate with an increase in temperature, where crack growth rates
are accelerated over the AMSE curve, shown by the dotted line. Tice et al. also reported crack
growth retardation at 300'C at a rise times greater than 60 minutes (denoted by the black arrow),
which agrees with data from Wire et al, where the environmental enhancement effects of water
temperature decrease at long rise times. The same is also seen for 250'C. [2]
1.OE-07
I
PWR water
R-0.7 - 0.8
'.oE-08
300C
1.0(-09
4..
t.OE-il
A1 flE-12
1OEE-14
1.OE-13
1.OE-12
1.E-10
1.OE-11
daA (A&)
1.OE-09
t.OE-08
/S
Figure 13: Time domain plot of crack growth rates in a PWR environment, shown for a range of
temperatures. The black and red arrows denote locations of retardation in environmental crack
growth rates. [2]
Tice et. al. [2] further explained that while environmentally assisted fatigue cracking is
retarded for long rise times at high temperatures (3000C), lower stress ratios, on the order of 0.3,
require longer rise times in order to exhibit the retardation effect. Rise times effects are further
discussed in the next section.
40
2.4.4 Rise Time
At a constant temperature, FCGRs in PWR water increase as the rise time increases for low
sulfur material, as measured by Tice et al. [2], shown in Figure 14. Here, the maximum
environmental effect occurred at the longest rise time of 567 minutes, at a low AK, where crack
growth rates were enhanced around 100 times over those in air. Experimental conditions are the
same as mentioned in previous sections for Reference 1, which are LWR-type conditions.
1.OE-06
II
I
R4..7, 2500i 1
1.OE-07
" mink-O
0~
I
minm
-
____
- or
0 min
o-bi um
AL
A k-309 Pull
* r-Iu mln
t.E-0
I
10
100
Cyclic Straa ntenalty Factor Range (MPaIrm)
Figure 14: Crack growth rates as a function of AK and rise time
(tR).
[2]
Wire et al. (in continuation of work from Evans and Wire [16]) measured the effect of rise
time using very similar heats, but with different sulfur contents, with chemistries shown in
Table 3. For short rise times of 8.5 and 50 seconds, the crack growth rates increased with increasing
rise time, as shown in Figure 15, with very slight differences between the heats. [5]
41
Table 3: Heat chemistries, with a range of sulfur content.
-r200_0--
-T(
0OW
a:
T6U
2S6C, Re7 11-13 MPmMn
100
eC
O0
Q
0
I
H"a A
H"a B. CT-3P
H"I 8, CT-0P
Alt RvO.7_. Kw1 IMPWM
10
100
1000
Ri)m TBm, a
Figure 105: Crack growth rates as a function of rise time for Heats A and B. Crack growth rates
are similar at short rise times, but deviate at longer rise times. [5]
At the longest rise time of 500s, the two heats varied greatly, with Heat B (<0.01 wt.% sulfur, but
higher than Heat A) showing a strong retardation in crack growth rate, while Heat A (0.001 wt.%
sulfur) continued the trend established during the short rise times. Additionally, study of the
fracture surfaces at the 50 second rise time for Heat B showed transgranular cracking with sharp
crystallographic faceting, suggesting heterogeneous deformation along planar slip bands,
suggestive of a dominant hydrogen enhanced cracking mechanism. [3,4] However, these features
42
were not obvious at the longest rise time, suggesting a homogeneous deformation ahead of the
crack tip. Figure 16 shows the region of crack surface where the rise time changes from 1500s to
50s, where the bottom (1500s rise time, (a)) shows the less descript features of the longest rise
time, which changes to sharp crystallographic faceting at the 50 second rise time (top, (b)), above.
Mills proposed a mechanism to describe the features seen along the crack surface for Heat B. He
proposed that the homogeneous deformation temporarily blunts the crack tip, decreasing the crack
propagation along the planar slip bands, as the slip bands become disrupted. [3] This deformation
pattern was hypothesized to occur as a result of injected vacancies from the corrosion process,
which is increased due to the higher sulfur content of Heat B.
Figure 116: Surface features for 50 s rise time (top of figure) and 1500 s rise time (bottom of
figure). The fan shaped lines in the bottom part of the figure are possible areas of crack tip blunting.
[5]
43
2.4.5 Stress Ratio and AK
Figure 17 shows the crack growth rate dependency on both AK and stress ratio, as reported
by Wire et al., where stress ratios of 0.3 and 0.5 were tested, and a stress ratio of 0.7 was from
previous experiments. [5]
0
+
2~**cusms.Thfl.
".0'
ROS
0~
104
I
y
o
/
:4 .,W
/
*
I Or
/
I
9/
U
~.uo ~
-~--
e~.o
7
/
-
V
100
8
10
20
30
40
AK, MPMin
Figure 127: Crack growth rate as a function of AK and stress ratio. [5]
In general, the crack growth rate increases as the AK is increased, though environmental
effects are more prevalent at the lower AK, in that the crack growth is increased by a larger margin
over the ASME air curves at lower AK. Crack growth rates increase with the same dependence as
those in air at AK from 10-15 MPadm. [5] This result was also reported by Tice et al., shown in
Figure 18, where the amount of environmental enhancement was the least at the highest AK values,
compared to the ASME air curves. [2] The environmental effect on crack growth rate decreases
with shorter rise time, because at shorter rise times, the crack tip is driven too rapidly for the
environment to have as large an effect. Additionally, the effect of stress ratio between values of
0.3 and 0.5 at these conditions is minimal; differences are no greater than those that occur from
normal data scatter. When compared to crack growth rates for a stress ratio of 0.7, the higher stress
ratio results in rates about twice that of those at the lower values, likely due to a greater
44
environmental effect that occurs at loads close to maximum loads. [5] Tice et al. reported the
existence of a possible threshold effect with high R. He observed that crack growth rate values
decrease with stress ratios greater than 0.85, as illustrated in Figures 18 and 19. In Figure 19, the
arrows indicate rise times where the crack growth rate decreased over the testing period. [2]
1.OE-05
AK values in MPa1m
1.OE-06
-t
t
I
1
ma
A
*a
1.OE-07
z
M
1.OE-08
"
I
I
I
SR=0.7, Delta K 4.5
* R=0.3, Delta K 10.5
A
A
* R=0.7 Delta K 9-11
*R-0.7 Delta K 12 -15
AR=0.8 Delta K 7-10
0
A*
* R-0.85 Delta K 5-6
9
9
* R0.9 Delta K 4-5
1.OE-09
1
10
1000
100
Rise time / secs
10000
100000
Figure 18: Crack growth rate effect of stress ratio and AK as a function of rise time. The overall
trend for the data is increasing crack growth with increasing rise time. [2]
45
1.E-08
11-09
SA
V1.E-10
*
0
A*
R-0.7, DK 9-15
A
o R=o.7, DK<5
0.
A R=).8, DK 7-9
0 R=0.85, DK 56
0 R=4.9, DK 4-6
1.E-11
1.E-12
1.E-11
I.E-10
1.E-09
1.E-08
daldt (air) m/s
Figure 19: Time domain plot of crack growth rate as a function of stress ratio for low sulfur
material. [2]
46
2.5 Environmental Crack Retardation
As previously discussed, under certain testing conditions and material chemistries,
environmental crack growth rates may be retarded, rather than enhanced, to values close to or less
than those found in air. In the previous section, several conditions, such as high coolant flow rates
and long rise times, did not produce the expected environmental enhancement, resulting in FCGRs
that were much lower than the environmentally enhanced rates. Additionally, high temperature
studies and certain material chemistries have also shown some degree of retardation. This section
will discuss these additional environmental retardation effects, with a focus on sulfur content of
the stainless steel.
2.5.1 Proposed Mechanisms for Retardation in Corrosion Fatigue
Retardation in crack growth rates has been previously explained by two different
mechanisms, separated between low and high stress ratios. [3] Both of these mechanisms depend
on increased corrosion as their driving factor, due to the material's interaction with the
environment. For low stress ratios (R less than 0.4), it is proposed that a buildup of corrosion
product on the fracture surface will cause crack closure, reducing the effective load the crack tip
experiences during fatigue. [20] Figure 20 shows low R (behavior of crack growth rates for two
different 304 stainless steel heats in 288'C water. [3]
At AK values below 15 MPaNm, crack growth rates were retarded below the
environmentally enhanced rate as seen for Heat 61115 (intermediate sulfur, 0.006 wt%), which
was shown to have a thick corrosion product on the fracture surface. In contrast, Heat 42322 (low
sulfur, 0.001 wt%) had a much thinner corrosion product layer, and maintained environmentally
enhanced rates at much lower values of AK. At higher AK levels, the crack growth rates converge.
Crack growth rate results for these heats at a stress ratio of 0.1 are shown in Figure 20, and the
effect of corrosion product thickness is shown schematically in Figure 21. When the mating
surfaces are in contact, the thick corrosion product may be appear "smashed". In some instances,
corrosion products on mating surfaces may attach to the opposite surface and break off. The
difference in crack growth rates for thin and thick oxides can be explained through the effective
47
AK (AKeftective), given as the maximum K minus the K when the crack is just above the closure
value. [3] For a thin oxide layer, the AK effective is close to AK that would result from the
maximum and minimum applied loads, so the crack growth rates align with the expected
environmentally enhanced rates. For a thick film, the effective AK is much smaller, because the
minimum load on the crack occurs when the oxidized crack flanks come into contact, which can
be well above the applied load. Hence, the effective AK will be smaller than the expected AK from
the applied load, and the mechanical driving force for crack growth is greatly diminished. As
AKeffective decreases, the R ratio increases to the point where there can be little driving force for
fatigue crack growth, and the crack arrests. [3]
304 SS Tested in 288*C DPW
R = 0.1, tR = 50 s
da/dN =4.55 x 10-
a 131 0
(AK) 2 25
00
10-3
/
V
/ 288*C AIR, R =-.0.1
0
/ ASME
E
2f
1 0-4
/
0
/
0
10-5
I
6
I.t., 't T
10
0
Heat 61115, CT-7P
Heat61115. CT-11P
Heat 42322, CT-12
60
AK, MPa'lm
Figure 13: Low R crack growth rate as a function of AK. Certain crack growth rates for Heat
61115 (intermediate sulfur) are retarded below the predicted values for full environmental
enhancement. [3]
48
Figure 14: Oxides from mating surfaces can be "smashed" together. [3]
For high stress ratio conditions, retarded FCGRs can also slow down to rates well below
the environmentally enhanced rates. It has been postulated that the mechanism leading to the
retardation of environmentally enhanced rates is due to the injection of vacancies, leading to
enhanced creep (IVEC), proposed by Mills. [3] Under this process, the increase in corrosion
product production, which is greatly accelerated by the presence of sulfur, leads to vacancies being
injected at the crack tip. These vacancies disrupt the planar slip bands by allowing edge
dislocations to climb, leading to homogenized slip and reducing the stresses at the crack tip, which
in turn reduces the driving force of crack growth. The IVEC process is shown schematically in
Figure 22 on the right, and the fully enhanced crack growth rate process is shown on the left. When
the crack growth is fully enhanced, the hydrogen will diffuse ahead of the crack tip, causing planar
slip, which localized cyclic damage and enhanced crack growth allows the crack to proceed
quickly. The hydrogen that results from the corrosion process absorbs into the metal crack tip
region. During the IVEC process, at long rise times and higher temperatures, corrosion processes
have a greater opportunity to increase vacancy concentration in the material that can diffusion to
dislocations and enhance climb and creep. The dissolution of MnS places sulfides into an anoxic
environment at the crack tip, which depassivates the material and greatly enhances corrosion,
resulting in retardation.
49
14,
T
Fuly Enhaned Crac Growth
Severly Retwded Crack Growth i
unwdw Long HoWd TUes or in"I
Su*r 85
s
Figure 15: IVEC mechanism. On the left, a fully enhanced crack tip is shown, having planar
slip bands. On the right, a retarded crack tip is shown, as suggested by the IVEC mechanism,
where the planar slip bands are disrupted, and the crack tip stresses are reduced. [3]
2.5.2 Sulfur Content
As previously discussed in Section 2.5.1, high sulfur stainless steels have a much greater
propensity to undergo retardation. In addition to low sulfur (Heat 42322) and intermediate sulfur
(Heat 61115) materials, Mills also tested a third, high sulfur material (Heat A16830), which
contained a sulfur content of 0.030 wt% (whereas Heat 42322 had a sulfur content of 0.001 wt%,
and Heat 61115 had a sulfur content of 0.006 wt%). The test results for the low and high sulfur
heats are compared in Figure 23, for low R, where the green symbols represent the high sulfur
material. At low AK values, below 15 MPa\m, the high sulfur material (Heat A16830) shows a
significant drop off in crack growth rates, where some values fall below the ASME air curve; this
behavior is indicative of retardation due to crack closure. Retardation for high R is seen in Figure
3. The proposed cause of the retardation for both low and high stress ratios in high sulfur stainless
steels is the enhanced crack tip corrosion that occurs as a result of the elevated sulfur content in
the alloy.
50
304 / 304L SS
2500C DPW
tR 51S
10-3
da/dN = 3.935 x 10 (AK) 2,25
AA
-250*C Air, R = I.
ASME
1 0-
E
AL
E
z 10O-5
V
'A
10-6
V
vR = 0.1
AR
= 0.3
Ht 42322
VR = 0.1
A R= 0.3
4 AA
10-7
5
50
10
AK, MPam
Figure 16: Crack growth rate as a function of AK, for R of 0.1, 0.3. Green data points are high
sulfur material (Heat A16830), red data points are low sulfur material (Heat 42322). [3]
The sulfur in type 304 stainless steel is in the form of MnS inclusions. The potential-pH
diagram for MnS-H 20-Cl~ is shown in Figure 24. [17] Chlorides were present in the environment
specified by Eklund at 0.1 M/liter, which were not present in the test environment used in this
thesis. [17] The potential diagram-pH is shown for room temperature, however, MnS is known to
dissolve in this test environment at high temperatures. Estimated conditions at room temperature
are shown in red (at higher temperature, the potentials will be in the -700 to -800 mVsHE range).
When the MnS inclusions dissolve, hydrogen sulfide is produced inside the pit that is created,
-
lowering the pH of the pit. MnS inclusions could be dissolved in a pH 7 environment from
1 O0mVsHE to 400mVsHE. Electrochemical reactions for MnS are discussed in Chapter 3.
51
E nW
HS Z
600
400
09
200
2
s
7
-200
M n( H)2
Mn 2
'
-400
-600
000
2
4
6
8
10
12
pH
Figure 17: Potential-pH diagram for MnS in LWR environment. Estimated crack tip pH and
potential values are shown in red. [17]
52
Chapter 3 - Experimental
The experiments that were used to support the thesis goals involved corrosion FCGR
testing as a function of rise time (all other loading conditions were generally kept constant),
electrochemical characterization of the materials, and optical and electron microscopy of the
materials after both electrochemical and fatigue tests. Tests were performed to provide further
insight into the mechanisms behind environmental enhancement and retardation for certain loading
and material conditions, in order to examine the proposed mechanisms that were previously
mentioned.
3.1 Autoclave System
Two autoclaves systems were used during this study. Each autoclave consists of a
autoclave system combined with either an MTS servo-hydraulic or an Instron electromechanical
fatigue machine. The servo-hydraulic system is composed of stainless steel components (autoclave
and piping). The hot (T>1000 C) sections of the electromechanical fatigue system are made of
titanium with the other parts fabricated from and stainless steel components. The remaining
components of the fatigue machines are a water circulation system and a heating and pressure
control system. Although each autoclave has its own independent circulation system, the water for
both systems is supplied from a main column located in the center of the "waterboard", which
contains all of the components except for the high pressure pump and autoclave. Figure 25 shows
a schematic of the waterboard, and Figures 26 and 27 show images of the autoclave system,
including waterboard, fatigue machine, and control instrumentation. The autoclave operates at
288'C and 10.2 MPa.
The pH for these tests was controlled to either 10 or 10.7 at room
temperature; the details of pH control will be discussed in more detail in Section 3.1.1. All tests
were conducted under deaerated conditions (with oxygen less than 5 ppb).
The flow was
recirculating at a rate of 3.8L/hr.
53
3.1.1 Autoclave Conditioning and pH Control
Before each test, the autoclave was conditioned to produce a uniform film of corrosion
product on internal components, and the conditioning was performed until the outlet conductivity
was steady at approximately 0.11 pS/cm. Conditioning was performed by closing the system,
sparging with argon gas until the desired conductivity was met, and then sparging with 100%
oxygen gas to reach a concentration of approximately 3000 ppb, to control conductivity. Lastly,
100% hydrogen was bubbled in to establish deaerated conditions and the system was allowed to
fall back to the desired conductivity value. After conditioning, ammonium hydroxide in a 28%
solution was added to the makeup tank in order to reach a pH at room temperature between 9 to
10.5. Ammonium hydroxide was used since it is volatile and will not be incorporated into the oxide
at the crack tip like solid chemistries such as lithium hydroxide would. The pH was originally to
be measured using an in-line pH probe. However, during the first test it was discovered that the
meter was reading inaccurately, resulting in part of the first tests being run at a pH of 10.7 vs. 10.0.
This was ultimately identified to be due to the position of the pH sensor with respect to the water
flow direction. As a result, the pH was subsequently manually measured by sampling the water in
the makeup tank and measuring it using a table top pH monitor (which was checked against a
standard solution) in order to ensure proper values.
54
&
I
I
IW
I
I
i~.
9
~Tn~
4U1
t
64
* v'
,Ii~
t-
I
Figure 18: Schematic of waterboard used for autoclave testing.
55
Figure 19: Images of waterboard and pump on left. On the right, the fully closed fatigue machine
is shown.
Figure 20: On the left, the specimen setup is shown within the fatigue machine. On the right is an
example of the computer software used to run the fatigue test.
56
3.1.2
Autoclave Load Control System
Fatigue tests were performed using a constant cyclic stress intensity (AK) during crack
growth. The control software, which allows the user to specify maximum stress intensity
(Kmax),
stress ratio (R), crack growth increment per test segment, waveform, amplitude, and frequency of
the applied load, interacts with an Instron controller operating in either of the fatigue machines.
Since the autoclaves are at high pressure, internal pressure applies 2.7 kN of load to the specimen.
Because of this, the load control also relies on feedback from the autoclave pressure gage in order
to adjust the externally applied load to meet the specified
Kmax
(i.e. the applied load is autoclave
pressure compensated).
Load cell calibrations were performed in order to ensure the accuracy of the fatigue
machines. The calibrations were performed in room temperature air by applying 1OV to a
calibrated load cell (Model WMC-5000; Serial #452804), and then applying a range of loads from
0-22.2 kN while recording applied load readings from the machine and output readings from the
digital voltmeter (output readings were then interpreted to obtain the applied load from the
calibrated load cell). Results from this calibration can be found in Appendix A.
3.1.3 Data Acquisition and Test Data
Crack length measurements were made using a direct current potential drop (DCPD) system.
Using a known initial crack length, the system measures crack growth based on the measured
voltage across the crack face and returns the results in terms of crack length over specimen width.
This is done by passing a current through the specimen, and then measuring changes in potential
drop as the crack grows, which can be converted to crack length via the known initial values and
a standard correlation. [18] The software also reports real-time values of autoclave temperature,
pressure, pH, water conductivity, oxygen content, and potential (measured using internal Fe-Fe
Oxide reference electrodes, supplied by GE-Global Research & Development, and Pt electrodes).
Tests were conducted at stress ratios (R) of either 0.4 or 0.7, with Kmax of 28.6 or 31.9 MPa1m.
57
3.2 Materials and Specimen Preparation
Materials were provided by Bechtel Marine Propulsion Corporation (BMPC), Bettis
Laboratory, and all materials came from forged bars. Material chemistry was also provided by
Bettis Laboratory. Heat E5174 (5174-LR-1 1) and Heat A16830 (A16-LR-10) were tested in this
study. A third heat, Heat D2739 (2739-LR-2) will be tested in a future test program. The three
material chemistries are given in Table 4. The low sulfur material (Heat E5174), contained a boron
content of 18 ppm, while the other two materials contained boron contents around 11 ppm. At a
boron content of 18 ppm, precipitation of intergranular (Cr,Fe)2B-type boride precipitates lead to
higher sensitization than a boron content of 11 ppm. [19] Each 1.0 CT specimen was precracked
before fatigue testing within ASTM E647 guidelines. Each specimen was precracked to lengths
shown in Table 5. Additionally, specimens 5174-LR-1 1, A16-LR-10, and 2739-LR-2 were tested
in air to measure air fatigue crack growth data, in order to compare to ASME air data. Specimen
2739-LR-2 is part of an additional low sulfur material (Heat D2739) that is to be used in further
research; the specimen was precracked and used for air fatigue crack growth data. Figure 28 shows
the orientations and origination of the metallographic sections.
Heat
Table 4: Material chemistries, wt%
Ni
Cr
S
Mn
C
N
Fe
E5174 (Low Sulfur)
0.020
1.60
<0.0025
19.8
10.1
0.085
Bal
A16830 (High Sulfur)
0.021
1.73
0.032
18.5
8.2
0.089
Bal
D2739 (Low Sulfur)
0.019
1.60
<0.0025
18.3
9.4
0.051
Bal
Table 5: Specimen Precrack Lengths, as measured from notch.
Additional Length in Air
n
g
Aa
Precrack Length (a/W)
Specimen
(a/WV)
5174-LR-11
0.505
0.554
A16-LR-10
0.388
0.434
2739-LR-2
0.401
0.448
58
S
n-
Syrrbol in
Diagrvtm
A
B
C
D
E
F
G
H
Suggested Designation
Roled surface
Dtection of roing
Roled edge
Planar section
Lorgikdinal section perpendicular to rolled surtace
Transverse section
Radal longudinal section
Tangental lontludinal section
Figure 21: Specimen orientation from ASTM E3
3.3 Electrochemistry
The electrochemical testing was performed with help from Dr. Yusaku Maruno from
Hitachi-General Electric (Hitachi-GE), Tokyo, with additional help from students Lun Yu and
Rachel Clark from MIT. In addition to the fatigue crack growth tests, cyclic polarization
experiments were performed in order to understand the role of MnS inclusion and sulfur content
in the corrosion behavior of the material. Tests were performed according to ASTM standard G6 1.
The setup for the experiments is shown in Figures 29 and 30. Figure 29 shows an example of
specimen preparation, which is then placed in an electrochemical cell, as shown in Figure 30. Each
specimen was carefully measured to determine surface area once mounted in resin. Nickel wire
was attached to the back of the specimens and then covered with a glass tube (attached with resin).
Lastly, lacquer was applied to the edges of the specimen surface and the resin to prevent any
exposure to the environment of crevices and spaces that would occur during mounting. The
electrochemical cell was composed of an Ag/AgCl reference electrode with saturated KCl Haber59
Luggin capillary salt bridge, and 1000 ml volume of 0. IN Na2SO4 solution. In order to maintain a
constant temperature, a water bath system was assembled to maintain 25 'C. The pH was varied
to achieve values of 10 (using NH 4 0H, to mimic autoclave conditions), 7, and 4 (using H 2 SO4 ).
Ni Wire
G lass Tube
Resin
Specimen
Figure 22: Electrochemistry specimen (front and side views).
RE. AgAgCL
Sat, KO
WE
Flow of
SArgas
Counter
Figure 23: Gamry electrochemical cell and electrode arrangements used for electrochemical
measurements.
60
In order to create deaerated conditions, Ar gas was bubbled through the system. Cyclic
polarization was performed with a scanning voltage range of -0.55VsHEto 1 .5VsHE, with a scanning
rate of 0.167 mV/s. Additionally, potentiostatic tests were performed, in order to narrow in on
certain potential peaks that were observed during cyclic polarization. Using the same test setup
(with a pH of 7) and specimens, a potential of -250mVsHE was applied for 1200 seconds, followed
by a lOOmVsHE potential, in order to create a passive film. Potentials of 220 and 650 mVSHE, which
had produced peaks during the cyclic polarization tests, were then applied for 7200 seconds, and
current density was measured. After the tests, optical microscopy was performed to examine the
specimen surfaces.
3.4 Microscopy
Scanning Electron Microscopy (SEM) on the low sulfur material (Heat E5174) was performed at
Carnegie Mellon University using an FEI Quanta 600 SEM and FEI Novalab 600 Focused Ion
Beam (FIB). SEM on the high sulfur material (Heat A16830) was performed at in part at both MIT
and Altran Solutions, Inc, [451 D Street] Boston, MA.
61
Chapter 4: Fatigue Crack Growth Rate Test Results
4.1 Low Sulfur Material (Heat E51 74 -Specimen 51 74-LR-11)
In this section the corrosion fatigue results for the low sulfur material (Heat E5174,
specimen 5174-LR- 11) are presented. Crack growth rate testing, electrochemical measurements,
and fractography from this heat are discussed in this section. The boron content was noticeable
during electrochemical testing due to pitting observed at grain boundaries, but was not observed
to be a factor in the fatigue crack growth behavior. Due to errors in the control software, the load
ratio was incorrectly set to 0.4 for the entire tests. The lower stress ratio data was still acceptable,
and were used to compare to similar data produced in the high sulfur material.
4.1.1 Crack Growth Rate Testing
The raw crack growth rates ("DCPD indicated") for the low sulfur material (Heat E5174)
are given in Table 6, and the crack length is plotted versus time in Figure 31. The "jump" in crack
growth rate during Step 2 (51 second rise time) is due to a machine failure that occurred during
the step, which involved shutting the test down, and moving the specimen from one autoclave to
the other. The spike in crack length is due to the renormalization of the DCPD measurement, and
is an artifact. Crack growth rates were calculated with the second data set during this step, and is
designated by the arrow in Figure 31. As expected, the corrosion FCGRs increased with increasing
rise time, as discussed in Chapter 2. Data for each individual step are found in Appendix B.
Table 6: Crack growth results for low sulfur material (Heat E5174)
Step Rise
R
AK
Time
DE Corrected
DE Corrected
da/dt (mm/s)
da/dt (mm/s)
da/dN
(mm/cycles)
DCPD
DCPD
indicated
indicated
da/dN
(mm/cycle)
1
5.1
0.4
17.1
l.00x10- 4
1.67x10 5
1.17x10- 4
1.98x10-5
2
51
0.4
17.1
1.81x10-4
3.02x10- 6
2.51x10-4
4.19x10-6
3
510
0.4
17.1
3.56x10-4
5.92x10- 7
4.17x10-4
6.93x10- 7
4
5100
0.4
17.1
8.13x10-4
1.36x10- 7
9.47x10-4
1.58x10-7
62
Cracklength (a) vs. Time
30.5
30
Step 4
E 29.5
E
Step 3
-Step
1
-Step
2
Step 3
Step 4
29
Step 2
28.5
Step 1
280.
0.00
50.
100.0
50.00
100.00
150.00
200.00
250.00
300.00
350.00
400.00
450.00
500.00
Time (hrs)
Figure 24: Crack length versus time for low sulfur test specimen (Heat E5174).
Specimen Sectioning and DestructiveEvaluation (DE) Correction
Upon completion of testing, the specimen was sectioned into three parts, in order to have
enough sections for the different analyses. The method of specimen sectioning was taken from
previous work done by Gibbs [18], and the cuts are shown in Figure 32. The two outside, 6.35 mm
pieces were saved for further microscopy, and the middle, 12.7 mm piece was used for correction
of the DCPD data for the physical crack length and SEM fractography.
63
num
0
Wire EDM specimen into X",
W", and 4" sections, with B
section being " 1thick
number -C1
ID
ID number -81
ID number -Al
IDnumber
0
-C2
_7%-
11numb-, -82
ID number -A2
Figure 25: Sectioning of specimens for further analysis. Cuts were made using wire EDM
(copper wire in a water bath).
In order to correct for any inaccuracies in the crack length measurement using DCPD, the
in air DCPD data was corrected using actual measurements of the crack length by cyclically
loading the specimen at high frequency until failure and destructively examining (DE) the fracture
surface. Since the individual steps were not easily discernible, measurements were taken along the
crack front in increments of 0.668 mm, at both the end of the fatigue precrack and the end of the
final corrosion fatigue crack. Additionally, measurements were taken at the outsides and insides
of the two other specimen sections, in order to fully understand the shape of the crack front; an
example of the side measurements is shown in Figure 33.
LCF CrackL
I 2.01 5mnf'
Fatigue Pre-Crack
12.442mm
Figure 26: Side measurement of the crack from one of the outside pieces of specimen.
64
The crack length measurements were made relative to the machined specimen notch, which
had a length of 15.24 mm from the load line, are shown in Figure 34. Hence, all measurements
could be referenced to the load line, allowing K to be determined. The DCPD measurements
correlate well to the DE correction, with a less than 4% difference when compared to the center
measurements, and from 4-8.6% difference on the sides. The DE correction, per ASTM 647,
requires measurements at two locations along the crack front (for example, at the precrack and the
end of the fatigue crack). The DE correction also requires a minimum spacing along the contour
of the crack at a spacing of 0.25W, or for a 1.0 CT specimen, four points. The DE correction for
this specimen was performed with 28 points, much more than the required four.
DE Measurements with Side Measurements
19.00
18.00
-
17.00
5.00
-j
4.00
* Precrack DE
13.00
* Precrack, sides
12.00
-
-
---
11.00
0
-
-
5
15
20
10
Position along Width, mm
25
Fatigue Crack DE
AFatigue Crack, sides
30
Figure 27: Measurements of the crack length, as measured from the notch, along the specimen.
The corrected DCPD crack length as a function of cycles is shown in Figure 35, with the
corrected crack growth rates reported in Table 6. Crack growth rates increased with increasing
rise time in throughout the test.
65
32.5
5174-LR-11, DE Corrected Only
2,288E+01
y = 9.466E-04x+
2
R = 3.330E-01
32
y= 4.160E-04x+ 2.796E+01
2
R = 9.914E-01
Step 4, 5100s
31.5
E
VA
y = 2.514E-04x + 2.920E+01
2
R = 9.872E-01
31
Step 3, 510s
5.1s
--
51S
.510
---
s
5100s
30.5
Step 2, 51s
y =1. 166E-04x + 2.970E+t01
R' = 9.990E-01
30
Step
1,
5.1s
29.5
0
2000
4000
6000
8000
10000
Total Cycles (N)
Figure 28: Corrected crack growth rate values as a function of cycles for the low sulfur test
specimen (Heat E5174).
4.1.2 Fractography
SEM was used to examine the fracture surface after testing, without removal of corrosion
product from the surface. Each step was examined to determine differences in fracture morphology
between the step increments (rise times). The fractography that was observed is indicative of full
environmental enhancement, as compared to fractography in References 3 and 15, and features
discussed in Section 2.4. Figures 36 and 37 show the crack features step by step, starting from the
precrack and ending in the post test air fatigue region. Figure 36 shows the precrack (a), Step 1
(5.1 s rise time, (b)), and Step 2 (51 s rise time, (c)). Figure 37 shows Step 2 (51 s rise time, (c)),
Step 3 (510 s rise time, (d)), Step 4 (5100 s rise time, (e)) and the fatigue apart (in air, (f)). Crack
growth direction is denoted by the arrow in the bottom left corner.
66
Figure 29: Crack surface from precrack (A), to Step 1 (5.1s tR, B), to Step 2 (5 Is tR, C). Image is
at 300x magnification. Blue marks show approximate changes in steps.
Figure 30: Crack surface from Step 2 (51s
E), to the fatigue apart (F).
tR,
C), to stop Step 3 (510s
tR,
D), to Step 4 (5100s
tR,
Figure 38 shows the fracture surface from the first step, with a 5.1 second rise time. Arrows
show examples of river patterns, heavy oxidation, and "blitzkrieg" attacks reported in the
literature, where the cracking occurs adjacent to areas that are crack resistant. [2] Facets can be
seen running north to south, instead of east to west (the crack growth direction). The change in
crack direction may be due to either the acceleration of the crack, in areas of less fatigue resistance,
or different crystallographic planes on which the crack is growing. The heavy oxidation is most
likely a result of this part of the test being in the autoclave for the longest amount of time, but the
67
features and locations of the oxides are still relevant and consistent between all four rise times
tested.
Figure 31: Step 1, 5.1 s rise time. Arrow A shows river patterns. Arrows B and C show areas
where the cracking occurred perpendicular to the crack growth direction, which is possible
evidence for a Blitzkrieg attack.
Figure 39 shows a higher magnification view of the center of Figure 38. The fracture
surface for Step 1 (5.1 s
tR)
was heavily oxidized, with large and small oxides. Figure 39 shows an
example of both, where the large oxides are seen on the ridges of the facet, and the smaller oxides
in the valleys. The smaller oxides appear crisper than the larger oxides (which may be small areas
of surface contamination from the post-test sectioning). At the center of the image, the crack path
68
appears to "dive" under the fracture surface, where the faceting has proceeded along parallel slip
planes below the surface, causing secondary cracks.
Figure 32: High magnification (1200x) of Step 1 (5.1 s
Figure 40 shows only Step 2 (51 s
tR),
tR).
with clear river patterns showing the crack path.
Oxides in Step 2 were smaller than in Step 1 (5.1 s tR), most likely due to the longer exposure time
of the step surface in the autoclave. There are fewer sideways facets in this step as in the previous
step, possibly due to the increased corrosion that occurs during longer rise times, and the higher
crack growth rate, leading to fewer areas of resistances. Also shown in Figure 40 is an example of
a possible DSA event, as described in the Chapter 2. As discussed, DSA events have been thought
to potentially contribute to environmentally enhanced FCGRs. The arrow in Figure 40 indicates
such a DSA event. Note that the faceting before and after the DSA event is continuous, as there is
no disruption in the crack growth process. As discussed in Reference 3, Mills states that facet
69
orientation and morphology is not affected by DSA, and the faceted growth mechanism is
reinitiated immediately.
Figure 33: Step 2 (51 StR). Yellow arrow marks possible DSA event.
Between Step 2 (51 S tR) and Step 3 (510 s
shape and size again.
tR),
shown in Figure 41, the oxides appear to change
However, Energy-Dispersive X-ray Spectroscopy (EDX) that was
performed on a specimen from similar material, sectioned in the same way, indicates that these
may be contaminants from the Electrical Discharge Machining (EDM) that was used to section the
specimen, since the particle chemistry contained copper. These large, spherical particles were thus
ignored in analysis of the crack growth process.
70
Figure 34: Transition between Step 2 (51 s tR) and Step 3 (510 s tR), with transistion shown by
the blue line. The large particles on the right are contaminants from the EDM process.
Figure 42 shows a higher magnification view of the area given by the arrow. The bands
shown by arrow 1 are most likely slip offsets, rather than striations, due to the large spacing, their
straightness, and the fact that they are not in the correct orientation. Arrows 2 and 3, on the other
hand, show an example of striations; although hard to see, the bands are parallel with the crack
front. In order to ensure these were striations, the spacing was measured, and compared to the
macroscopic crack growth. The existence of both striations and slip bands in the region is evidence
of both heterogeneous and homogeneous cracking.
71
Figure 35: Steps 2 and 3 (51s and 51 Os tR). Arrow 1 shows an example of slip offsets, while
arrows 2 and 3 show striations.
The fracture surface from Step 3 (51 Os
tR),
due to the higher crack growth rates, shows
much less growth of cracks with local orientations perpendicular to the macroscopic crack front.
Figure 43 shows the faceting and river patterns following the direction of crack growth, which
occurs from left to right.
72
Figure 36: Step 3 (5 1Os
tR).
River patterns run parallel to the crack growth direction.
The fracture surface characteristics of Step 4 (51 00s
tR)
is shown in Figure 44, and is the
small, dark band that occurs in the center of the image, starting at arrow 1. Although the step is
small, the striations (which are difficult to see at this magnification) are about three times larger
than the preceeding step, which is around the same order of difference between the rates for those
steps. Also shown in this figure is the post test fatigue apart region, which occurs on the right side,
denoted by arrow 2, and was performed in air at room temperature. Faceting from Step 4 continues
on to the early part of the fatigue apart, due to the plastic zone that is formed during testing in the
environment. Ahead of the crack tip, the embrittlement process occurs before cracking. Step 4
(5100s
tR)
is shown in Figure 45. The arrows, while all indicating "bands" that are occuring
perpindicular to the crack path, give examples of three different features. Arrow 1 points most
likely to a grain boundary, due to its straight nature, rather than curved, as most other features
73
would be. Arrow 2 shows striations, which correlate within a factor of two to the macroscopic
crack growth rates, an appropriate level of accuracy.
Figure 37: Step 3 (510s tR) and Step 4 (5100s
starts at arrow 2.
tR).
Step 4 starts at arrow 1, and the fatigue apart
74
Figure 38: Step 4 (51 00s tR). Arrow 1 shows an example of a grain boundary. Arrow 2 shows
fatigue striations. The spacings were within a factor of 2 of the macroscopic crack growth rate
for Step 4.
75
4.1.3 Optical Microscopy
In order to more completely characterize each material, optical microscopy was performed on each
heat of material. Each material was mounted and polished, and then etched in a 10% oxalic acid
solution for 90 seconds. Figure 46 shows the microstructure for the low sulfur test material (Heat
E5174), with specimens taken from the longitudinal and transverse directions of a CT specimen.
Magn~tfion
Sa$0,09n
P04o
X200
z500
Figure 39: Microstructure of the low sulfur test material (E5174).
Figure 47 shows the microstructure for another the low sulfur material (Heat D2739) which did
not have the higher boron content present in Heat E5174. As expected, there is much less pitting
in this low sulfur material than in the higher boron low sulfur material (Heat E5174).
76
Sampling Position
owysufuateMagnification
x200
L
(
x5W0
Figure 40: Microstructure of the low sulfur material (2739-LR-2).
77
4.2 High Sulfur Material (Heat A16830 - Specimen A16-LR-10)
In this section the corrosion fatigue results for the high sulfur material (Heat A16580,
specimen A16-LR-10) with a sulfur content of 0.03 wt% are presented. The test was originally
planned for five steps with various rise times, but was expanded due to various complications that
arose during testing. Due to errors in the control software, the load ratio was incorrectly set to 0.4
for the initial portion of the test. The software issue was corrected and the test continued at a load
ratio of 0.7. While not a specific objective originally, the lower load ratio data were acceptable
and provided an opportunity to evaluate both low and high stress ratio. Additionally, due to then
unknown flow effects on the pH probe, the pH was actually higher than the values that were
recorded by the probe (actual 10.7 while measured at 10.0). After investigation of the discrepancy
between the measured conductivity and the actual pH, the pH was then adjusted to 10, and
measured periodically using standard solutions and a bench-top pH meter. Lastly, the pressure
compensation was not present during the last step of the test due to another software error which
resulted in a higher AK.
4.2.1 Crack Growth Rate Testing
Crack growth rates for each step are given in Table 7, with both raw and the corrected data
shown. Crack length is shown as a function of time in Figure 48 for raw values, with the DE
corrected values in Figure 49.
78
Table 7: Crack growth rates for high sulfur material (A16830)
A16830
(High
S)
Step
Step
Step
Step
1
2
3
4
Step 5
Step 5-
pH
R
Tr
(s)
AK,
MPaNlii
da/dN
(mm/cycle)
DCPD
Indicated
da/dt
(mm/sec)
DCPD
Indicated
da/dN
(mm/cycle)
Corrected
10.7
10.7
10.7
10.7
0.4
0.4
0.4
0.4
5.1
51
510
5.1
17.1
17.1
17.1
17.1
1.34x10~ 4
1.81x10-4
9.37x10-5
1.27x10-4
2.24x10- 5
3.03x10-5
1.56x10- 7
2.1 1x10-5
1.44x10- 4
1.95x10- 4
6.63x10- 5
1.36x10- 4
2.40
3.24
1.10
2.26
1.32x10-8
1.39x10- 4
2.32 x10-8
x10 5
x10-6
x10-7
x10-5
10.7
10
0.4
0.4
5100
17.1
7.92x10- 5
5100
17.1
3.43x10-4
5.71x10-8
3.40x10-4
5.67x10-8
10
10
10
10
0.7
0.7
0.7
0.7
5.1
51
5100
5100
8.6
8.6
8.6
9.6
6.25x10- 5
3.66x10- 5
3.51x10-5
8.15x10- 5
1.04x10-4
6.10x10-6
6.71x10-5
3.91x10-5
7.62x10- 5
8.64x10- 5
1.12 x10-5
6.52 x10-7
1.27 x10-8
1.44 x10-8
21
Step 6 2
Step 7
Step 8
Step 9 3
da/dN
(mm/cycle)
Corrected
5.84x10-8
1.36x10 7
1. Due to location of pH meter in the autoclave, pH values were not accurate
2. Test conditions were for R of 0.7, but due to system inputs, were initially run at 0.4
3. Due to a computer failure, the system pressure compensation was removed, increasing the max load.
Using the same procedures that were used for the heat E5174 specimen, the specimen was
sectioned into three pieces, the fracture surface of the middle of the specimen was measured after
being fatigued apart, and the DCPD-indicated crack lengths were corrected for the physical crack
length. The DCPD-indicated crack length is shown as a function of time in Figure 48, with the
corrected crack length as a function of cycles shown in Figure 49. The corrected rates for each
step are also summarized in given in Table 7. The corrected data is again included in Figure 50.
This time, only 4 measurements were made, in accordance with ASTM 647.
79
A16 Raw Crack Length
27.4
269
Step 9
Step 8
26.4
Step 7
25,9
E
Step 6
25.
Step 5
Step 4
249
24.4
Step 3
Step 2
23.9
Step 1
.4
22.9
0
1000
2000
3000
4000
Time (hrs)
5000
(000
000
8000
Figure 48: Crack growth rate versus time for high sulfur material (Heat A16830)
80
A16 DE Corrected
29.9
y = 8.62E-05x + 2.37E+01
R= 5.90E-02
28.9
Step 9
ep 8
y =3.92E-05x+ 2.62E+01
R' = 9.82E-01
y =7E62E-0x+ 2.42E+01
R2 2.47E-01
27.9
y =6 71E-05x + 2.52E+01
R' 9.77E-0
y
E 26.9
*
$tep
1.36E-04x + 2.35E+01
R
Step 5
y
Se
$e
6
1.50E-04x + 2.33E+01
R= 4.56E 01
Step 4
y
25.9
1.95E-04x+2.28E+01
R2 = 9.64E-01
Step
6.62E-05x + 2.50E+01
R2= 7.47 E-01
$tep 2
24.9
y
Step 1
23.9
0
10())
1.44E 04x + 2.34E+01
9.91E-01
20000
4WX)
30000
51*
60000
70000
Total Cycles (N)
Figure 49: Crack growth versus cycle for high sulfur material (Heat A16830)
81
DE Measurements along Sides
16
14
---------------- U
12
* Precrack, Sides
E
*E lo
t Final Fatigue
Crack, Sides
6
4
2
-
0
--- --
0
[---
10
Position Along Width (mni 0
Figure 50: DE correction for high sulfur material Heat A 16830.
In the high sulfur material, the crack growth rates decrease with increasing rise time. The
crack growth rates for the second portion of the test, where the stress ratio was increased from
0.4 to 0.7, were lower than the crack growth rates at the lower stress ratio as would be expected
given the lower AK at higher R for the same Kmax.
4.2.2 Fractography
Fractography on the high sulfur material (Heat A16830) was performed at both MIT and
at Altran Solutions, Inc, (451 D Street, Boston, MA). Individual steps were even less discernible
than they were in the low sulfur material (E5174), due to the heavy layer of corrosion product on
the surface for this material. The entire fracture surface is shown in Figure 51; area (a) denotes the
precrack, (b) denotes the fracture surface, (c) is the last step, Step 9 (51 00s tR), and (d) is the fatigue
apart in air. Although not obvious from the figure, the entire surface is covered in a layer of
corrosion product, as shown in the next images. Figures 52 and 53 show the characteristics of the
heavy corrosion product that was present over the entire fracture surface. Crystallographic faceting
is still present, beneath the oxide.
82
Figure 51: Entire fracture surface, with precrack (A), fatigue crack (B), Step 9 (51 00s tR, C) and
fatigue apart (D).
Mag a 200 X
EHT a 20.00 k
WD a 9.0 mm
BLTRdlo
UOLurAJ1Om
Figure 52: Example of fracture surface, which was heavily oxidized.
83
Meg a
200 X
EHT
20.00 kV
WD = 9.0 mm
OLI
NC I I
sNmXAJTCXS
Figure 53: Fatigue surface covered in heavy layer of oxide.
Figures 54 and 55 show what are possibly fatigue striations, though heavily covered in
corrosion product. Measurements taken on the striations in Figure 55 correspond well to the
macroscopic crack growth rate. The striation spacing was about 3.2 pm, which falls between 1.34
pm and 7.66 pm, the range of crack growth per cycle for the entire test.
84
Mag a 1.00 K X
EHT a 20.00 kV
WD a 9.0 mm
aLTRafl
SMAI11MN
Figure 54: Example of fatigue striations on the fracture surface.
85
Figure 55: Example of fatigue striations on the fracture surface. Striations were about 3.2 im.
Figure 56 shows the fracture surface for the low stress ratio (R=0.4) portion of the test,
which is highly faceted, and generally uniform throughout the entire test. Figure 57, on the other
hand, shows the fracture surface for end of the test, in the high stress ratio regime (R=0.7) where
FCGRs were retarded. As the rise time increases to where retarded rates are seen, the faceting
disappears, leaving a smoother, heavily corroded crack surface. The band, designated by the red
arrow, is the Step 9 in the test, and is shown in higher magnification in Figure 58. The surface,
again, shows little crystallographic faceting.
86
Mag
19 X
EHT = 20.00 kV
WD
9.0 m
TR
Figure 56: Fracture surface for the low stress ratio (R=04) portion of the test.
87
Figure 417: Fracture surface for the high stress ratio (R=0.7) portion of the test.
88
Figure 58: Three steps (changes in rise times) of the fatigue test. Arrows show the transition in
rise times on the fracture surface.
Figure 59 shows possible area where the crack was briefly arrested, due to changes in steps
(rise times). Additionally, the band down the middle of the image is most likely a change in step,
where the rise time changed; there is little difference between the morphology on either side of the
step change, but the oxide size and shape appears to change.
89
Figure 59: Fracture surface, with areas of crack arrest (A), heavy oxidation, and locations of
MnS dissolution (B, C).
Figure 60 shows locations where the MnS inclusions were dissolved. During optical
microscopy, images were taken at 200 and 500 magnification, and areas of dissolution appear to
be the same size between the two methods of imaging.
90
Figure 60: Fracture surface, with heavy oxidation and locations of MnS dissolution.
4.2.3 Optical Microscopy
Optical microscopy was performed on the high sulfur material (Heat A16830) in the same
manner as the low sulfur materials. The material shows pits on the surface after polishing and
etching, due to the dissolution of the MnS inclusions. Pits appear small and circular, or elongated,
depending on the orientation of the inclusions, as seen in Figure 61; the elongated pits occur along
the longitudinal section of the bar. The inclusions appear to be dispersed within the material
preferentially in the rolling direction, with no segregation towards the grain boundaries.
91
Moaoketo
Samp4ne Posftio
4w0
I
I
Figure 61: Optical microscopy for the high sulfur material (Heat A16830).
92
Chapter 5: Electrochemistry Results
Cyclic polarization and potentiostatic tests were performed at room temperature to
understand the basic electrochemical behavior of the material, as well as analyzing the effect of
specific potentials on the high sulfur material.'
5.1 Cyclic Polarization
Room temperature cyclic polarization results are shown in Figures 62 through 64; Heat
A16830 is designated as High Sulfur, Heat D2739 is designated as Low Sulfur, and Heat E5174,
also low sulfur (test specimen), is designated as "Dummy". At a pH of 10, the corrosion potential
for all specimens is between -260 mVsHE to -170 mVSHE, as shown in Figure 62. The low sulfur
material, Heat D2739, shows a stable passive region from -100mVsHE to lOOOmVsHE, with no
discernible differences due to specimen orientation. The other low sulfur material, Heat E5174,
which also contains boron concentration of 18 ppm, exhibits a much higher current density than
the other low sulfur material, with a markedly different behavior in the passive region. Heat
A 16830 exhibits a passive region similar to that for Heat D2739, though a peak appears at around
160mVsHE. The low sulfur specimens were cut from the CT specimen, where C refers to the face
of the CT specimen, and crack growth occurred in the A direction. Orientations from there are the
same as those in the optical microscopy results.
1.
Electrochemical tests were performed with help from Dr. Yusaku Maruno from GE, Tokyo. Additional help
proved by Mr. Lun Yu and Ms. Rachel Clark from MIT.
93
16
--
14 HOS 8
A
-
C:
DUMMY
12
10
-I 08
w
i
04
'U
02
02
00
00
-04
-02
-04
I E-09
7
tE-08
1E-7
I E-06
Cimyut 0.nsky 4km2)
E-0
1 E-04
-06
I E-09
IE-gE
1 E-07
E-06
1 E-
1 E-04
QOxert Duniy #Na2}
Figure 62: Cyclic polarization of the low and high sulfur materials. Peak in the passive region of
the low sulfur material are most obvious at this pH of 10.
In pH 7 conditions, this peak becomes much more apparent, and a second peak can also
been seen around 630mVsHE, resulting in a much different passive region than the low sulfur
material, as seen in Figure 63. In pH 4 conditions, peaks become even more noticeable, with a
third peak at 1400mVSHE, shown in Figure 64. These peaks in the passive region are the most
obvious difference between the low and high sulfur materials, causing a jump in current density in
a region where the material should be passive. This high exchange current density is due to the
dissolution of MnS inclusions, dissolving by either Equation 3 for potentials around -lOOmVsHE
to 400mVsHE (for pH 7 conditions), or Equations 4 and 5 for the potential around 650mVsHE.
MnS + 4H2 0= Mn2+ + S0 4 - + 8H+ + 8e-
(3)
2Mn2+ + 3H20 = Mn203 + 6H+ + 2e-
(4)
Mn203 + H20= 2MnO2 + 2H+ + 2e-
(5)
94
16
16
A
Low 9
14
14
C21
12
r.
C
-
RaN Evca
12
10
I"
-4
10
A-
I
08
--
0.6
06
ij
w
04
04
~
-~
-~
-
-- 02
02
-~
-~
~
-~
-iii
~ -Tv
0.0
-0.2
-04
-04
-06
-1,~
-46
I E-t
tAnE-04
I E-B
I E-07
CLInet
I E-OB
E-05
Oenrtyf (AnWO
I E-04
i
I &OR
E-49
I E-Q
W
Gurren
1
1,E-8
Densidy (Aean2
Figure 63: Cyclic polarization of the low and high sulfur materials. Peaks in the passive region of
the high sulfur material are more obvious. Environment at pH 7.
18
14
LOW S
A-
A-
B
HOgb S
-~
12
t2
10
to
B
C
08
i
06
w
04
06
Ro~cve~sct~
w.
02
--
_
-41--------i.'
-
-
7T~
-
_
TT,
-~
-~
1
~
-~
44
-04
IE-08
IE-
I E-07
I E-06
CnenI Dinlay NrW2)
1 E-06
IE-04
CWrr
Don*%y (At-}
Figure 64: Cyclic polarization of the three materials. Peaks in the passive region in the high
sulfur material are not seen in the low sulfur materials. Environment at pH 4.
95
5.2 Optical Microscopy (after Polarization Tests)
After cyclic polarization tests, the specimens were examined for any surface pitting or
corrosion that might have occurred during testing. Figure 65 shows the appearance of both low
sulfur (Heat D2739) and high sulfur (Heat A16830) specimens after polarization. The low sulfur
material shows no pitting or tarnish, except for around the edges of the specimen, where lacquer
was applied. The high sulfur specimens, in contrast, are heavily pitted, most likely in locations of
MnS inclusions. These surface features were present at all pH conditions.
(1)IwS (-
(3)Hi* S(=iti
I A)
&s~A)
(2) IwS (wic
A)
(4)H1* S (a ~biim A)
Figure 65: Surface appearance of high and low sulfur materials after potentiostatic tests.
96
5.3 Potentiostatic Tests
In order to further examine the peaks in the passive region that were found during cyclic
polarization in the high sulfur heat (Heat A16830), potentiostatic tests were performed at
220mVsHE and at 650mVsHE. After fixing the potential at 220mVsHE, the current densities after
7200 seconds were 1.13 x 10-7 (A/cm 2) for orientation A, 1.30 x 10-7 (A/cm 2 ) for orientation B,
and 3.41
x
10-7 (A/cm 2) for orientation C. Optical microscopy showed shallow pits on the surface
of the specimens, shown in Figure 66. For 650mVsHE polarization, the current densities were
)
2.55x10-7 (A/cm 2) for orientation A, 2.05 x 10- 7 (A/cm 2) for orientation B, and 1.03 x 10-6 (A/cm 2
for orientation C. Again, pits were observed on the surface, but a higher density of pits were
observed than in cyclic polarization, as shown in Figure 66. Potentiostatic results are given in
Appendix C.
SSamlig
P:osMgiiiij
A
B
Ro
ng direcson
C
Figure 66: Surface features after potentiostatic testing of high sulfur heat (Heat A 16830). Small
pits are seen at 500x magnification.
97
Chapter 6: Discussion
6.1 Crack Growth Rate Comparisons
The crack growth rates measured in this thesis work showed the following trends:
1) In low sulfur materials, environmentally enhanced crack growth rates increased continually
with increasing rise time for low sulfur materials.
2) In high sulfur material, environmentally enhanced crack growth rates were observed at
short rise times, but as rise time increased the rates either were unchanged or approached
air rates, indicative of crack growth retardation.
The measured rates in this thesis were compared to established stainless steel FCGR curves for
air [4] and for environmentally enhanced fatigue crack growth in water. [10] The JSME equations
are used to predict crack growth rates for austenitic stainless steels in PWR environments, shown
in Figure 67, and plotted in the following figures. ASME air rates are also shown. Crack growth
rates for the low sulfur material (Heat E5174) compared to the predicted JSME curves [10] are
shown in Figure 68, for each rise time tested. Arrows designate the direction of increasing rise
time.
98
ASME Section XI
FatigueCrack Growth Rate o
d
ss Steels in Air
= CN(,K lf
Co
=
c = 1d[
S =
CS
(16)
- 8.714 + 1.34 x 10~ 3T - 3.34
x
daldV: m/( cle,
106 T2 + 5.95 x 10
9
T3
T: Temperaturedegree C
1.0whenR!5 0
=
1.0 + 1.8R when 0 < R S 0.79
=
- 43.35 + 57.97R when 0.79 < R
< 1.0
JSME S NA1-2010 (The Japan Society of Mechanical Engineers)
Code for Nuclear Power Generation Facilities
[Rules on Fitness-for-Service for Nuclear Power Plants]
CorrosionFatigueCrack Growth Rate in the PWR Primary Water Chemistry
da/dN = 4.35 x
10-3 x Tc- 63 x Tr
33 x
AK3-0 x (1-R)1 -5 6
da/d : nm/cycle, AK: MPcarm, Tc: Temperature degree C, Tr: Rising Thme (Second)
Figure 67: Crack growth rate equations for air (ASME) and LWR-type environments.
99
1.E-04
1 1Z
E5174 (Low S), R=0.4
zff
" Tr=5.1s
A
Y1
" Tr=51s
1.E-05
" Tr=510s
" Tr=5100s
1.E-06
z
M
- - --
1.E-07
1.E-08
1
JSME PW
Curves
10
100
AK (MPa4m)
Figure 68: Crack growth rates as a function of AK with JSME, ASME, and experimental data for
the low sulfur material. Stress ratio is R=0.4.
The experimental results in Figure 68 are slightly lower than the predicted JSME water
curves, but for all rise times, the crack growth rates are above the ASME air curve. As expected,
the crack growth rates increase as the rise time increases, and are consistent in rise time
dependence when compared to the JSME curves. The behavior of the experimental results
corresponds well to the reference data in Chapter 2.
Figure 69 shows the experimental results for the high sulfur material (Heat A 16830),
again plotted with the JSME water curves and the ASME air curve. The data are only shown here
for the low stress ratio, R=0.4. Unlike the low sulfur material, the high sulfur material does not
follow the pattern of increasing crack growth rate with increasing rise time. Rather, at the longest
rise time, 5100 s, the lowest crack growth rate is seen. In Chapter 2, the retardation in high sulfur
100
austenitic stainless steels occurred at longer rise times, around 500 seconds, which corresponds
well to the result that is seen in Figure 67. The 5Is rise time crack growth rate increases from the
5.1 s rise time, but 51 Os rise time crack growth rate is lower than the both of the shorter rise
times. The conclusion is that retardation in crack growth rates occurs for high sulfur materials at
longer rise times. The retardation occurs, in theory, due to crack closure (for low stress ratio).
[20]
1.E-04
iA16 (High S),
R=0.4
-
__
________
0 Tr=5.1s
* Tr=51s
1.E-05
* Tr=51Os
* Tr=5100s
1.E-06
z
ff
ff
---
Iff zAV
f
1.E-07
1.E-08
1
10
100
AK (MPa4m)
Figure 69: Crack growth rates as a function of AK with JSME, ASME, and experimental data for
the high sulfur material. Stress ratio is R=0.4.
Figure 70 shows the crack growth rate data for the high sulfur material at a high stress
ratio of R=0.7, corresponding to steps 6-9 of the test (Step 9 [5100s tR] has a AK of 9.6 MPalm,
rather than 8.6 MPa'm. The behavior at this stress ratio differs from the R=0.4 data, as the rise
time of 51s has a lower crack growth rate compared to the 5.1s rise time. The crack growth rates
then remain almost constant as the test transitions to a longer rise time of 51 00s. Even though the
101
crack growth rates at 51 and 51 00s rise times nearly identical, the predicted growth rates from
the JSME curve increase with increasing rise time. The crack growth rate at the longest rise time
is retarded compared to the same rise time in the low sulfur material. The retardation for this
high stress ratio, in theory, is due to the IVEC mechanism. [3]
1.E-04
A16 (High S), R=0.7
ff1/ 11 1 A
-4
" Tr=-5.1 s
" Tr=-51sr
1.E-05
" Tr=-5100s
*Tr=-5100s (K=9.6)
U 1.E-06
z
1.E-07
1.E-08
1
10
100
AK (MPavm)
Figure 70: Crack growth rates as a function of AK with JSME, ASME, and experimental data for
the high sulfur material. Stress ratio is R=0.7.
To compare the low and high sulfur heat, the crack growth rates are shown as a function
of rise time in Figure 71. As expected, the crack growth rates for the high and low sulfur deviate
the most at the highest rise times, due to the retardation from the sulfur. The high sulfur material,
shown in blue (R=0.4) and red (R=0.7), has crack growth rates that remain the same or decrease
with rise time, while the low sulfur material, shown in green (R=0.4) has crack growth rates that
increase with rise time. This results in almost an order of magnitude reduction in measured rates
at the longest rise time.
102
da/dN vs. Cycle
1.00E-03
1 .00E-04
ze
E
--
1.00E-05
Low Sulfur (R-0.4), AK -17.1 MPa
Mills, Low Sulfor(R-0.1) [3], AK - 10-17 MPa
- Tice, Low Sulfur (R-0.7) [2], AK - 12-15 Ma
-S-High Sulfur (R-0.4), AK - 17.1 MPa
0
-4-High Sulfur (R-0.7), AK - 17.1 MPa
Mills, High Sulfiur(R-0.1) [3], AK - 10-17 MPa
--- Wire, High Sulfur (R-0.7) [6], AK
11-13 MPa
I.OOE-06
10
100
1000
10000
Rise Time (s)
Figure 71: Crack growth rates as a function of rise time for low and high sulfur materials, including
literature data in similar environments, for reference.
A comparison of crack growth rates at low stress ratios for high and low sulfur materials
can be compared to values reported by Mills. [3] At a rise time of 51s, the measured crack growth
rates are 1.95x 10-4 mm/cycle and 2.5x 10-4 mm/cycle for low and high sulfur, respectively. From
Figure 23, Mills reported values on the order of 10-4 for both low and high sulfur materials, with
the crack growth rate of the high sulfur material being slightly lower than that of the low sulfur
material at the given stress intensity range (17.14 MPalm). Mills data at low (0.001 wt.%) and
high (0.006%) sulfur contents is shown in Figure 71 [3], along with data from Tice for low sulfur
content (0.001 wt%) [2], and data from Wire et al. for high sulfur content (<0.01 wt%). [7]
Literature data is also shown for low and high stress ratio, to compare. The values produced in this
thesis correspond well with the values from the literature, and follow the same rise time trends.
103
6.2 Effect of Sulfur (with Respect to Proposed Mechanisms)
At the low stress ratio of R=0.4, high sulfur materials possibly experience retardation due
to crack closure, as described in Chapter 2. As discussed, the appearance of the fatigue surface will
show areas of deformed corrosion product due to the mating surfaces being smashed together.
Figure 72 shows a comparison between the surfaces for the low and high sulfur materials at the
longer rise times (510 s and above) for R=0.4. On the left, the oxides that are on the surface of the
high sulfur material have a slightly blunted nature on the top, while the low sulfur material on the
right shows crisp oxides. The corrosion product is also much heavier and thicker on the high sulfur
specimen, as expected by the higher material sulfur content.
Figure 72: Comparison between high (A) sulfur and low (B) sulfur materials at R=0.4, tR >5 1Os.
The arrow in A shows an example of an area where the corrosion product appeared damaged,
due to crack closure. The change in appearance of the oxides, in both shape and amount, is
obvious between the two images.
Additionally, the fatigue surfaces show similar features to those given by Mills. [3] Figure
73 shows high and low sulfur materials for low R (R=0.1, 0.3) as shown in Reference 3. On the
surface of the high sulfur material, left, the mating surface has actually detached from its original
surface, and is found on the mating surface shown below. The low sulfur material is shown on
right.
104
Figure 73: Comparison between high (A) sulfur and low (B) sulfur materials at R=O.1, 0.3.
Features are similar to those in Figure 72. [3]
At the high stress ratio of R=0.7, features of importance on the crack surface at long rise
times (5100s) are the prevalence of nondescript features and less cleavage faceting, shown in
Figure 74. These features are predicted by Mills [3] as part of the IVEC mechanism. The heavy
corrosion product suggests that an active strain component of corrosion is effective; as the material
is subjected to the maximum load for longer periods of time, which occurs at the longest rise times,
the environmental corrosion processes are increased. As suggested in Chapter 2, this increased
corrosion, in turn, generates the increased vacancies that are injected ahead of the crack tip. The
surface in Figure 74 closely resembles the surface that is shown in Figure 75 of intermediate sulfur
material from Reference 2, shown for a rise time of 600 seconds by the arrow, with a stress ratio
of 0.65. Although the images are different in appearance, the nondescript features are apparent in
both figures.
105
Figure 74: Step 9 (5100 s tR) of the high sulfur specimen. There is less cleavage faceting, and the
fracture features are nondescript, as compared to fractrography from shorter rise times in Chapter
4.
The crack growth rate data that was produced for the high sulfur material at high stress
ratio shows retarded cracking that is predicted by the IVEC mechanism. The degree of retardation
at high stress ratio, for long rise times, is similar to that shown in Figure 3. [8] At long rise times,
crack growth rates for high sulfur stainless steel were reduced around an order of magnitude in
both the literature and the results reported here over the low sulfur stainless steel. Since the fatigue
crack surfaces match those reported by Mills, the retardation is assumed to be related to the features
mentioned above, and again, related to the high sulfur content in the stainless steel. The IVEC
mechanism, as opposed to the crack closure that occurs at low R, acts ahead of the crack tip, and
in order to support IVEC, this region must be examined in further detail. The fatigue surface, while
106
it supports the mechanism, does not provide direct evidence of the micro processes underlying the
IVEC mechanism.
Figure 75: Intermediate sulfur (Heat 61115) material. The arrow denotes a 600 s rise time, which
produced retarded crack growth rates, and appears nondescript. [3]
107
Chapter 7: Conclusions and Future Work
7.1 Conclusions for Proposed Mechanisms
In this thesis, FCGR tests were conducted on low and high sulfur materials in order to
examine the effect of sulfur on corrosion fatigue behavior of austenitic stainless steels. Results
showed environmental enhancement in the low sulfur material; FCGRs increased with increasing
rise times. In the high sulfur material, results showed retardation of environmentally enhanced
crack growth rates, and FCGRs mostly decreased with increasing rise time. Fractographic analysis
was conducted on the fatigued materials in order to examine the fracture surfaces of both materials
and to compare with equivalent results in the literature, as well as to provide insight into the
proposed mechanisms for environmental enhancement and retardation. Electrochemical testing
was conducted on both materials in order to examine the corrosion effects of sulfur in austenitic
stainless steels at specific pH values, to gain an understanding of the processes that are occurring
at the crack tip.
With the combination of crack growth rate testing, metallography, fractography, and
electrochemistry, it is apparent that the higher material sulfur content has a retarding effect on
FCGRs in Type 304/304L stainless steel under LWR-type environmental conditions. Retardation
is a result of the increased corrosion that occurs due to the presence of sulfur in the form of MnS
inclusions which dissolved in the high temperature LWR environment. However, the processes
leading to retardation are most likely different at high and low stress ratios. As discussed in Chapter
2, mechanisms for retardation include crack closure due to oxide buildup and IVEC. Crack closure
is seen on the fracture surface of the low stress ratio test, supporting oxide induced crack closure
at low stress ratio, as discussed in Section 2.5.1. However, at high stress ratios, the surface features
do not provide enough evidence to fully support the IVEC mechanism.
108
7.2 Future Work Recommendations
In order to elucidate the IVEC processes, two targeted methods using the high sulfur
material have been proposed by the researchers, in order to examine features of IVEC ahead and
at the crack tip. First, dislocation structures ahead of the crack tip can be evaluated using
Transmission Electron Microscopy (TEM). During TEM, a thin foil of the relevant material is
created, and electrons are passed through the material in order to show features that could not be
seen using SEM, such as dislocations. After testing, regions of the crack tip, and the regions ahead
of the crack tip, could be analyzed to understand the dislocation structure. Since the IVEC
mechanism proposes that planar slip bands are disrupted ahead of the crack tip, the dislocation
structure could provide evidence of this process.
Second, advanced micro-characterization techniques, such as Atom Probe Tomography
(APT), could be used to study the chemistry at the crack tip. This technique has been previously
used by Gibbs [19] to study oxide compositions of tunnels that are created ahead of the crack tip
during SCC tests of nickel-based Alloy X-750. To support IVEC, APT could be used to examine
the oxide compositions that form in crack tips under fully enhanced and retarded FCGR regimes
to gain insight into the corrosion processes that may be controlling possible vacancy injection into
the crack tip region.
109
References
1. Chopra, O.K. and Shack, W.J. (2007). Effect of LWR Coolant Environments on the
Fatigue Life of Reactor Materials - Final Report. NUREG/CR-6909, February 2007.
2. Tice, D., Platts, N., Rigby, K., Stairmand, J., & Swan, D. (2005). Influence of PWR
Primary Coolant Environment on Corrosion Fatigue Crack Growth of Austenitic
Stainless Steels. Pressure Vessel and PipingDivision Conference, 2005 (pp. 193-205).
Denver, CO. ASME.
3. Mills, W.J. (2013). Accelerated and Retarded Corrosion Fatigue Crack Growth Rates for
304 Stainless Steel in an Elevated Temperature Aqueous Environment. 16th International
Conference on EnvironmentalDegradationof Materials in Nuclear Power SystemsWater Reactors. Asheville, NC.
4. American Society of Mechanical Engineers (2010). Rules for Inservice Inspection of
Nuclear Power Plant Components. Boiler andPressure Vessel Codes, Section XI.
5. Wire, G., Evans, W.M., & Mills, W.J. (2004). Fatigue Crack Propagation Tests on 304
Stainless Steel in High Temperature Water - Accelerated Cracking Rates and Transition
to Lower Rates. Pressure Vessel and Piping Conference, 2004 (pp. 71-81). San Diego,
CA. ASME.
6. Seifert, H., Ritter, S., & Leber, H. (2012). Corrosion Fatigue Crack Growth Behavior of
Austenitic Stainless Steels Under Light Water Reactor Conditions. CorrosionScience
(pp. 61-75).
7. Wire, G., Evans, W.M., & Mills, W.J. (2005). Fatigue Crack Propagation of 304 Stainless
Steel in High Temperature Water - Additional Tests and Data Correlation. Pressure
Vessel and Piping Codes and Standards. PVP2005-7 1680
8. Platts, N., Tice, D., Mottershead, K., McIntyre, L., & Scenini, F. (2012). Effects of
Material Composition on Corrosion Fatigue Crack Growth of Austenitic Stainless Steels
in High Temperature Water. 15th InternationalConference on Environmental
Degradationof Materials in Nuclear Power Systems- Water Reactors (pp.561-579).
Colorado Springs, CO.
9.
Jones, D.A. (1995). Principlesand Prevention of Corrosion, 2 nd Edition. Prentice Hall,
NJ.
10. Nomura, Y., Tsutsumi, K., Kanasaki, H., & Chigusa, N. (2004). Fatigue Crack Growth
Curve for Austenitic Stainless Steels in PWR Environment. Pressure Vessel and Piping
Codes and Standards (pp 63-70)
11. James, L., Lee, H., & Wire, G. (1997). The Effect of Water Flow Rate Upon the
Environmentally Assisted Cracking Response of a Low-Alloy Steel: Experimental
Results Plus Modeling. PressureVessel Technology, ASME (pp. 83-90)
12. Hill, R.E. and Abbaschian, R. PhysicalMetallurgy Principles, Third Edition. PWS-
KENT Publishing, 1992, pp 294-298.
13. Buongiorno, J. (2010). PWR Description. 22.06 EngineeringofNuclear Systems
14. James, L., Lee, H., Wire, G., Novak, S., & Cullen, W. (1997). Corrosion Fatigue Crack
Growth in Clad Low-Alloy Steels-Part II: Water Flow Rate Effects in High Sulfur Plate
Steel. Pressure Vessel Technology, ASME (pp. 255-263)
110
15. Wire, G. L., & Mills, W. J. (2004). Fatigue Crack Propagation Rates for Notched 304
Stainless Steel Specimens in Elevated Temperature Water. JournalofPressure Vessels,
ASME (pp. 318-326)
16. Evans, W. M. and Wire, G. L. (2002). Results of High Stress Ratio and Low Stress
Intensity on Fatigue Crack Growth Rates for 304 Stainless Steel in 288 0 C Water.
Pressure Vessel and Piping Codes and Standards, ASME. PWP2002-1226
17. Eklund, G.S. (1974). Initiation of Pitting at Sulfide Inclusions in Stainless Steel, J.
Electrochem. Soc., Vol. 121, No.4.
18. Gibbs, J. (2011). Stress CorrosionCracking and Crack Tip CharacterizationofAlloy X750 in Light Water Reactor Environments. PhD thesis, MIT
19. Johns, E.C. and Miller, B.D. (2013). Effects of Boron and Mechanical Processing on
Intergranular Attack in Dual Certified Type 304 Stainless Steel Determined Using DLEPR. 1 6th InternationalConference on Environmental Degradationof Materials in
Nuclear Power Systems-Water Reactors. Asheville, NC.
20. Lias, P.K. (1988). Overview of Crack Closure at Near-Threshold Fatigue Crack Growth
Levels. Mechanics of Fatigue Crack Closure, ASTMSTP 982 (pp. 62-92). ASTM,
Philadelpia, PA.
111
Appendix A - Load Cell Calibration
The calibration of the Instron load cell (Model #661.21A-0.3; Serial #3245) is shown Figures 76
and 77, and the calibration of the Instron load cell (Model #1020AJ-25K-B; Serial #186674A is
shown in Figures 78 and 79 . The applied load was interpolated using the actual readings from the
DVM.
Actual Readings
Applied Load
from DVM (V)
from Calibrated
Load Difference
(Fwatm-FrAh)
load cell (Ibs)
(Ibs)
0
0.00000000
-
1001.9
20013
300L9
40019
50019
40019
3001-9
20013
1001.9
0
0.00385800
0.00771900
0.01158900
0.01544800
0.01931200
0.01545800
0.01158900
0.00772400
0.00386300
1001.3
2003.4
3007.9
4009.5
5012.4
4011.8
3007.2
2003.7
1002
0.6
-1.5
-6
-7.6
-10.5
-9.9
-5.3
-1.8
-0.1
0.00000200
-
-
Error (%Full Scale)
-
Applied Load
from Instron (Ibs)
-
0.002668932
0.006672331
0.026689323
0.033806476
0.046706315
0.044037383
0.023575569
0.008006797
0.000444822
Figure 76: Applied Load Readings from Model #661.21A-0.3; Serial #3245
Instron Load Cell
WMC-5000 Calibration Shet Serial
(Model #661,21A~-03: Serial #3245)
0.025
#452804
025
--- Act Data
-uCal Sht
0.02
0 02
0 015
C 015
00
0.005
0
0
0
1000
2000 3000 4000
Load (lbs)
'6000 6000
0
2000
'000
Load (Ibs)
6000
Figure 77: Comparison of load vs. voltage between Instron load cell, left, and WMC-5000
calibrated load cell, right.
112
Load
Applied
Load from Difference
Calibrated (Fm.load cell
Fjjjcd)
(Ibs)
(bs)
Error
(%Full
Scale)
Applied Load from Instron
(lbs)
Actual Readings from DVM (V)
RUN
1.0
RUN2
0.7
Average
0.0
RUNI
0.000000
RUN2
0.000000
Average
0.000000
0.00
0.00
0.0000
1000.0
1000.0
999.2
0.003819
0-003834
0.003827
993.15
6.00
0.0240
2000.0
3000.0
4000.0
5000.0
4000.0
3000.0
2000.0
1000.0
1.0
2000.0
3000-0
4000.0
5000.0
4000.0
3000.0
2000.0
1000.0
0.7
1999.2
2999-2
3999.2
4999.2
3999.2
2999.2
1999.2
999.2
0.0
0.007677
0.011522
0.015376
0.019203
0.015384
0.011530
0.007671
0.003821
0.000000
0.007683
0.011533
0.015393
0.019230
0.015395
0.011532
0.007682
0.003836
0.000003
0.007680
0-011528
0.015385
0.019217
0.015390
0.011531
0.007677
0.003829
0.000002
1993.30
2991-90
3993.00
4987.60
3994.00
2992.20
1991.40
993.05
0.13
5.85
7.25
6.15
11.55
5.15
6.95
7.75
6.10
-0.13
0.0234
0-0290
0.0246
0.0462
0.0206
0.0278
0.0310
0.0244
0.0005
Figure 78: Applied Load Readings from Model #1 020AJ-25K-B; Serial #1 86674A
Instron Load Cell
(Model #1020AJ-25K-B; Serial #186674A)
WMC-5000 Calibration Sheet Serial
#452804
0 025
0.025
-.-
Act Data
2
---
Cal Sht
0.02
0.01
0 015
0
0.0(
0 005
C,
0
1WO
2( O
3(X)O 4(OX
Load (Ibs)
SWO
6X)(
0
2000
4000
Load (Ibs)
6000
Figure 79: Comparison of load vs. voltage between Instron load cell, left, and WMC-5000
calibrated load cell, right.
113
Appendix B - Crack Growth Rate Data
Heat E51 74 - Low Sulfur Material CrackLength vs. Time
Step 1 Crack length vs. Time
1.14
1.12
1.1
1.08
1.06
1.04
1.02
----- ----- --------
1
0
10
20
30
40
60
50
70
80
90
Elapsod Time (hrs)
Figure 80: Crack length vs. time for Step #1 (5.1 second rise time) for the low sulfur heat.
Step 2 Crack Length vs. Time
0.59
0.58
0.57
0.56
0.55
0.54
0.63
0.52
______
0.51
______il _____
_______
_______
_______
-
0.5
0.0 0
_______
50.00
100.00
150.00
200.00
250.00
300.00
350.00
ElapsedTime (hr.)
Figure 81: Crack length vs. time for Step #2 (51 second rise time) for the low sulfur heat.
114
Step 3 Crack length vs. Time
1.2
1.19
.-.
..
.
.
1.18
1.17
1.16
S1.15
1.14
1.13
1.12
1.11
1.1
350
300
250
500
450
400
550
600
650
700
Elapsed Time (hrs)
Figure 82: Crack length vs. time for Step #3 (510 second rise time) for the low sulfur heat.
Step 4 Crack length vs. Time
1.195
1.19
1.185
--.
1.18
WWAA
Md
1.175
1.17
1,165
1.16
1.155
1.15
640
660
680
700
720
740
760
Elapsed Time (hrs)
Figure 83: Crack length vs. time for Step #4 (5100 second rise time) for the low sulfur heat.
115
Heat A16830 - High Sulfur MaterialCrack Length vs. Time
Step I Crack Length vs. Time
0 94
003
092
0 91
-s
09
ae~e
LAW a
00
080
087
obe
5
0
15
10
20
TC Is.t
Figure 84: Crack length vs. time for Step # 1 (5. 1 second rise time) for the high sulfur heat.
Step 2 Crack Length vs. Time
097
0 905
0 96
I
0 6
-Raw
a
-Lows
0045
0094
0.935
093
0 925
0
10
20
30
40
50
00
70
s0
90
100
TNMe fars)
Figure 85:. Crack length vs. time for Step #2 (51 second rise time) for the high sulfur heat.
116
Stop 3 Crack Length vs. Time (Start to 387.6 Elapsed Tim.)
0978
0 976
0.974
IIIII
0.972
AOl
hL-i7di
t
I0966
-Rw a
S098
0964
0962
096
090"
50
100
150
2WX
250
3W
350
400
Uie (Mrs)
Figure 86: Crack length vs. time for Step #3-1"t Section (510 second rise time) for the high
sulfur heat.
Step 3 Crack Length vs. Time (387.6 Elapsed to End Stop 3)
-----
0.96S
350
450
550
650
750
650
950
Tbl hnr,)
Figure 87: Crack length vs. time for Step #3 -2nd Section (510 second rise time) for the high
sulfur heat.
117
-
0.99
...
----------
0099
98
o.988
4--
o-84
o.982
-Raw
-----
a
098
0.978
0.976
0.974
91 7
917 5
918
9185
9195
919
920.5
920
Figure 88: Crack length vs. time for Step #4 (5.1 second rise time) for the high sulfur heat
Stop 5 Crack Length vs. Time
-
1.02
0.98
-Raw
a
Linear a
..-S0.94
092
0.9
900
1100
1300
1500
1700
1900
2100
2300
Thme (hr.)
Figure 89: Crack length vs. time for Step #5 (5100 second rise time) for the high sulfur heat
118
Step 6 Crack Length vs. Time
1.025
1.02
---
1.015
1.01
i
1.005
U
VT
-Raw aa
-Rw~'
-- Lnear a
'
C
1
0
0.995
0.99
-
-
___________
---
______________
-______________
--
-___________-
___- -+-______- - - --______
fl GAr
2340
2350
2345
2300
2355
2385
Time (hrs)
Figure 90: Crack length vs. time for Step #6 (5.1 second rise time) for the high sulfur heat.
1.06
-
Step 7 Crack ength vs. Time
---------
1.055
1.05
1.045
1.04
-a
a
-1mner a
1.035
-
1.03
1.025
1.02
1.015
2300
2350
2400
2450
2550
2500
Time (hrs)
2600
2650
2700
2750
Figure 91: Crack length vs. time for Step #7 (51 second rise time) for the high sulfur heat.
119
Step 8 Crack Length vs. Time
1.07
___---_
1.06
1,02
1.05
4-
C
qwqw-V-".i
1.04
-I
-Raw
-Linear
1.03
a
a
1.01
1
2700
2900
3100
3300
3700
3500
3900
Time (hrs)
Figure 92: Crack length vs. time for Step #8 (5100 second rise time) for the high sulfur heat.
Step 9 Crack Length vs. Time
1.065
1.064
1.063
-
1.062
1.061
,
lb
1,06
-Rw
a
-Linear a
1.059
1.058
-
-
1.057
1.055
-
1.058
3800
4000
4400
4200
4600
4800
TWme (hrs)
Figure 93: Crack length vs. time for Step #9 (5100 second rise time) for the high sulfur heat.
120
Appendix C - Electrochemistry Data
PotentiostaticTests
The specimen preparation, as outlined in Chapter 3, is shown in Figure 94. The results are shown
in Figures 95 and 96.
130
1.00
030
0.00
-oin
I.WE-04
1.DE-05
1.0D-6 P
1.0DE-08
'
I
1.DE-07
0
500
1000
150
200
Tmh (sec)
Figure 94: Example of Pre-Treatment for Potentiostatic Specimens.
1.OE-O5
_HighS A
-FEghS
B
Bipb S C
22OmVsm
T17 71
%
1.OE-06
I .0E-07
0
1000
2000
3000
4000
5000
6000
'000
8000
Tmie (sec)
Figure 95: Results of Potentiostatic Test at 220mVsHE for High S Specimens in 0.1N Na2 S04
Solution (pH 7).
121
1.OE-04
-.
SA
-High
H
---
U
1.OE-05
HghS B
S C-
1.OE-06
- - -
-
- ---
U
1.OE-07
0
1000
2000
3000
4000
5000
Tne (sec)
6000
7000
8000
Figure 96: Results of Potentiostatic Test at 650mVSHE for High S Specimens in O.IN Na2 S04
Solution (pH 7).
122