Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Kinetic and fractographic study of the stress corrosion cracking of Austenitic stainless steels Russell, Alan James 1977

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1977_A6_7 R78.pdf [ 21.96MB ]
Metadata
JSON: 831-1.0078739.json
JSON-LD: 831-1.0078739-ld.json
RDF/XML (Pretty): 831-1.0078739-rdf.xml
RDF/JSON: 831-1.0078739-rdf.json
Turtle: 831-1.0078739-turtle.txt
N-Triples: 831-1.0078739-rdf-ntriples.txt
Original Record: 831-1.0078739-source.json
Full Text
831-1.0078739-fulltext.txt
Citation
831-1.0078739.ris

Full Text

CP. <  A KINETIC AND FRACTOGRAPHIC STUDY OF THE STRESS CORROSION CRACKING OF AUSTENITIC STAINLESS STEELS by ALAN JAMES RUSSELL B.Sc., University of Glasgow, 1972  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE DEPARTMENT OF METALLURGY We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA January, 1977 fc)  Alan James Russell, 1977  In  presenting  this  thesis  an a d v a n c e d d e g r e e a t the L i b r a r y I  f u r t h e r agree  for  the U n i v e r s i t y  make i t  freely  that permission  this  written  representatives. thesis  for  for  financial  of  F e b r u a r y 21,  Columbia  1977  of  Columbia,  British for  extensive  the  requirements  reference copying of  I agree and this  that  not  copying or  for  that  study. thesis  by t h e Head o f my D e p a r t m e n t  gain shall  Metallurgy  The U n i v e r s i t y o f B r i t i s h V a n c o u v e r 8, Canada  of  is understood  permission.  Department  Date  It  fulfilment  available  s c h o l a r l y p u r p o s e s may be g r a n t e d  by h i s of  shall  in p a r t i a l  or  publication  be a l l o w e d w i t h o u t my  ABSTRACT A variation of the double cantilever beam specimen has been calibrated and used to study the propagation of stress corrosion cracks as a function of stress intensity ln Jl6 and 310 stainless steels, and a TRIP steel exposed to hot aqueous magnesium chloride solutions.  The effects of cold work, tem-  perature and applied potential on both the fractography and cracking rates have been examined. The effects of cold work and crack path on crack branching were also investigated. Bpth stress Intensity dependent (Region I) and stress intensity independent (Region II) cracking were observed. Region II having apparent activation energies from 15.1 to 18.1 kcal/g.mole.  The crack velocities of Z%  kcal/g.mole cold rolled  316 were found to be independent of applied potential pver a range of more than 5°  mV  in Region I and 75mV in Region II. In  the same material the crack path changed from solely transgranular at low stress intensities and noble potentials to more than 86% intergranular at high stress intensities and active potentials. The topography of the transgranular fracture was similar to that observed by others except i n the case of the TRIP steel where nodular features were observed. These observations have been discussed with respect to mechanisms involving the follpwingj olution, ( i i )  (i) electrochemical diss-  absorption of hydrogen and ( i i i )  a damaging species.  adsorption of  Of these, an adsorption assisted process i s  most compatible with the observations.  Qualitatively the adsorbed  species are envisioned as modifying the behaviour of the surface atoms at the crack t i p .  -11ACKNOWLEDGEMENT I would like to thank Dr. D. Tromans for his readily available help and encouragement.  I am also grateful to  the faculty, staff and fellow students of the Department of Metallurgy for a l l the useful advice given to me regarding this research.  Thanks are also due to Mrs. Anne  Russell f o r financial support, patience and understanding. The financial assistance provided by the National Research Council (grant number A-^534) i s also acknowledged.  -iiiTABLE OF CONTENTS PAGE 1.  2•  INTRODUCTION  ,,,  1  1.1  General Description  ,  1  1.2  Fracture Mechanics Approach to 3.C.C..  2  1.3  The Chloride Cracking o f  7  1.4  ahbice of Materials and Scope of the Present Work  Fe-Cr-Ni A l l o y s . . .  10  EXPERIMENTAL. ,•»•»,•••••,•*••••••••»•..«••.•»••••••.••••»•••*•»••.  12  2.1  Specimen Geometry,  2.2  Stress Intensity C a l i b r a t i o n .  2.3  Specimen Thickness  18  2.4  Advantages and Limitations of the Specimen Geometry.........  19  2.5  Materials and Specimen Preparation  20  2.6  Apparatus and Testing Procedure.............................  22  3.2  3.3  ,..., 12  . 28  .3. • RESULTS.., 3.1  12  Stress Corrosion Crack Morphology  28  3.1.1  316 Stainless S t e e l . . . . .  28  3.1.2  TRIP S t e e l . .  31  3.1.3  310 Stainless S t e e l  31  Considerations  of Errors.............  31  3.2.1  Errors i n Crack Length Measurements  31  3.2.2  Errors due to the Cracks Growing Downwards  33  K i n e t i c Data......  33  3.3.1  33  316 Stainless S t e e l  3.3.1.1 3.3.1.2  E f f e c t of Stress Intensity on Crack Velocity E f f e c t of P l a s t i c Deformation,  33  . 3 4  -ivPAGfi  3.3.2  3.3.1.3  Effect of Temperature.  34  3.3.1.4  Effect of Applied Potential  40  3.3.2.1 3.3.2.2 3.3/3  3.4  43  TRIP Steel Effect of Martensite on the Region II Velocity..  Effect of Temperature on the Region II Velocities.... .....43  310 Stainless Steel.....  43  .'  3.3.3.1  Effect of Stress Intensity on the Crack Velocity.....  43  3.3.3.2  Effect of Temperature on the Crack Velocity.....  50 50  Corrosion Potentials..,. 3.4.1  Effect of Stress Intensity on the Corrosion Potential.  3.4.2  Effect of Composition and Processing Treatment  3.4.3  50  .. 54  on the Corrosion Potential  3.5  43  Effect of Temperature on the Corrosion Potential.....  54  Frac tography  54  3.5.1  316 Stainless Steel  54  3.5.1.1  Effect of Plastic Deformation  54  3.5.1.2  Effects of Electrochemical  3.5.1.3 3.5.2  Potential  and Stress Intensity  57  Effect of Temperature  68  TRIP Steel 3 . 5 . 2 . 1 Comparison with the Fractography of the 316 Stainless Steel..., 3.5.2.2 Effects of Cold Working and the Presence of Martensite  74 74 74  - V -  PAQE  3.5.3  3.5.2.3 Effect of Stress Intensity  78  3.5.2.4 Effect of Temperature  78  310 Stainless Steel  78  3.5.3.1 General Description  78 83  4. DISCUSSION 4.1 Crack Branching.  83  4.2 Kinetics  85  4.2.1  85  Region I  4.2.2 Region II  86  4.2.3  Crack Mechanisms.  87  4.2.3.1 General Considerations  88  4.2.3.2 Dissolution  91  4.2.3.3 Adsorption  93  4.2.3.4 Absorption of Hydrogen...  96  4.3 Corrosion Potentials 4.4 Fractography 4.4.1  4.5 5.  97 »  General  •• 99 99  4.4.2 Effect of Applied Potential  100  4.4.3 Effect of Stress Intensity.....  101  4.4.4 Effect of Composition  103  4.4.5 Effect of Martensite  ...104  Assessment of S.C.C. Models  105  CONCLUSION........  108  5.1 Conclusions  108  5.2 Suggestions for Future Work  109  BIBLIOGRAPHY APPENDIX 1....  HO .....114  -viLIST OF TABLES TABLE I II  III IV  •V  VI  Al  PAGE Alloy Compositions  21  Mechanical Properties of the Various Alloys and Processing Treatments (Tensile data measured perpendicular to the rolling direction at 150 C.)  23  Variation with Temperature of Region II Velocities for 2% and 50% Cold Rolled 316 Stainless Steel............. 39 Variation with Temperature of Region II Velocities for 25% Cold Rolled Austenitic and Partially Martensitlc TRIP Steel  47  Variation with Temperature of I n i t i a l and Subsequent Constant Velocities for 50% Cold Rolled 310 Stainless Steel  51  Corrosion Potentials of the Various Alloys and Processing Treatments  Errors due to the Downward Growth of the Cracks  ... 56  11$  -viiLIST OF FIGURES FIGURE 1  PAGE Typical relationship between stress corrosion cracking velocity and stress intensity showing Region I, II and III and K^g  2  5  GC  Specimens used in the present work and by others to measure crack velocities as a function of stress intensity......  3  ,  •  Double cantilever beam specimen showing side cracking.  8x.  Fatigue crack at t i p of notch.  Other cracks are S.C.G 4  14  Compliance of the specimen used in the present work as a function of crack length  5  13  16  Stress intensity calibrations for the specimen used in the present work and for a three point bend specimen of similar size...  6  17  Experimental arrangement showing S.C.C. c e l l , reflux condenser and load stabilizing spring mounted in tensile testing machine  7  ..24  Crack morphology for cracking parallel to the rolling direction in Z%  cold rolled 316  stainless steel. (a.) Side view.  8  7x. 4,8x.  (b)  Fracture surface.  (c)  End view at 15 and 30MPa.m . ¥  23x  29  Crack morphologies i n JlS stainless steel. (a)  12x.  Cracking perpendicular to the rolling  direction. (b)  Cracking in the annealed alloy.  (The crack  has been opened to f a c i l i t a t e observation of cracking.)  30  -viiiFIGURE 9  PAGE Crack morphology in the 310 and TRIP steels. (a)  TRIP - 25% cold rolled at 200°C.  Cracking parallel to the rolling direction. (b)  310 - 50% cold rolled.  to the r o l l i n g direction. 10  7x,  Cracking parallel 17x.  ,  32  Variation of crack length and load with time for a specimen of 25% cold rolled Jl6 stainless steel having an i n i t i a l load of 500N.  11  MgClg solution boiling at 154°C  35  Variation of crack length and load with time for a specimen of 25% cold rolled 31° stainless steel at 154°C. having an i n i t i a l load of 1200N.  12  MgCl solution boiling at 154°C 2  36  Effect of stress intensity on crack velocity for 25% cold rolled JlS stainless steel exposed to boiling MgCl solution at 154°C  37  2  13  Effect of stress intensity on crack velocity for 50% cold rolled 316 stainless steel exposed to boiling MgCl solution at 154°C  38  2  14  Arrhenius plot of the Region II velocities of cold rolled Jl6 stainless steel cracking in a 41  h5% MgCl solution 2  15  Effect of applied potential and temperature on the Region I cracking of 25% cold rolled 316 stainless steel exposed to a ^5% MgCl solution 2  16  42  Effect of cathodic polarization on the Region II velocity of 25% cold rolled 316 stainless steel cracking in a MgCl solution boiling at 154°C. ?  44  -ixFIGURE 17  PAGE Effect of anodic polarization on the Region II velocity of 2% cold rolled Jlh stainless steel  154°G  cracking in a MgClg solution boiling at 18  45  Effect of stress Intensity on crack velocity for 25% cold rolled TRIP steel exposed to MgCl solution boiling at 154°C  46  2  19  Arrhenius plot of the Region II velocities of 25% cold rolled TRIP steel cracking in a l±5% MgCl solution  48  2  20  Variation of crack length, load and stress intensity with time for a specimen of 50% cold rolled 310 stainless steel exposed to boiling MgCl^ 40  solution at 154°C 21  Arrhenius plot of the crack velocities of 50% cold rolled 310 stainless steel cracking in 52  a k$% MgCl solution 2  22  Variation of corrosion potential with time as a function of applied load.  25% cold rolled Jl£ stainless steel  53  in boiling MgCl solution at 154°G 2  23  Effect of stress intensity on the  corrosion  potential of 25% cold rolled 316 stainless steel cracking in boiling MgCl solution at 2  154°G... 24  55  Effect of temperature on the corrosion  potential  of 25% cold rolled Jl6 stainless steel cracking in ^5% MgCig ° - 't s  1  u  lon a  "t  a  stress intensity of  25MPa.m.  55  2  25  Effect of cold work on the fractography of 316 o stainless steel at 154 G. 400x. A  (a)  Annealed, K  = lOMPa.m  (b)  25% cold rolled K  (c)  50% cold rolled K  2  1  = 20MPa.m  2  ±  n  = 20MPa.m  J  58  -X-  FIGURE 26  PAGE Matching fractographs from opposite sides of the crack in 2% cold, rolled 316 59  stainless steel. 400x. 27  Matching fractographs from opposite sides of the crack in 2$% cold rolled 316 stainless steel.  28  60  lOOOx  Effects of applied potential and stress intensity on the amount and distribution of intergranular cracking. 2% cold rolled 316 stainless steel exposed to boiling MgClg solution at 154°C  29  61  Examples of fractographs of 2% cold rolled^ 316 stainless steel at 154°G. E = -0.25V (S.C.E.).  30  (a)  Edge of crack  (b)  Centre of crack  = llMPa.m , 2  400x. 62  Examples of fractographs of 23% cold rolled^ 316 stainless steel at 154°C. K = 22MPa.m% x  E = -0.25V (S.C.E.).  31  (a)  Edge of crack  (b)  Centre of crack  400x. 63  Examples of fractographs of 25$ cold rolled^ 316 stainless steel at 154°C. E - -0.3V (S.C.E.).  32  (a)  Edge of crack  (b)  Centre of crack  = llMPa.m^,  400x. 64  Examples of fractographs of 25$ cold rolled^ 316 stainless steel at 154°C. K E = -0.3V (S.C.E.). (a)  Edge of crack  (b)  Centre of crack  1  = 22MPa.m%  400x. 65  -xiFIGURE 33  PAGE Examples of fractographs of 2% cold rolled 316 stainless steel at 154°C. Centre of crack.  34  K  n  = 33MPa.m. T  400x.  (a)  E = -O.325V (S.C.E.)  (b)  E - -0.35V (S.C.E.)  66  Example of fractograph of 25$ cold rolled 316 stainless steel cracking without an applied load.  35  800x  6?  Example of fractograph of 50% cold rolled 316 stainless steel showing small area of = 70MPa.m*. 400x  ductile failure. 36  69  Distribution of ductile fracture in a specimen of 50% cold rolled Jl6 stainless steel exposed to boiling MgClg solution at 154°C. (a)  As a function of position across the  specimen. (b) 37  70  As a function of stress intensity  Examples of precipitates formed on the fracture surface during cathodic polarization.  38  (a)  E = -0.325V (S.C.E.).  800x.  (b)  E = -0.35V (S.C.E.).  400x...  Effect of temperature on the fractography of 50% cold rolled 316 stainless steel.  39  71  (a)  T - 154°C, K  x  - 27MPa.m  (b)  T = 116°C, K  x  - 27MPa.m  lOOOx.  2  2  72  Effect of temperature on the fractography of 25$ cold rolled 316 stainless^steel. 400x. (a)  T - 154°C, K  = 27MPa.m  (b)  T - 116°C, K. = 27MPa.m  2  x  2  73  -xliFIGURE 40  PAGE E f f e c t of cold work on the fractography of the TRIP s t e e l .  400x. - lOMPa.m^, T - 154°C.  (a)  Annealed TRIP, K  (b)  25% cold r o l l e d (at 200°C.),  ±  - 9MPa.m^, 75  T - 154°C 41  E f f e c t of martensite on the fractography of 25% cold r o l l e d TRIP s t e e l .  42  lOOOx.  (a)  F u l l y a u s t e n i t i c , 1^ - 22MPa.m^, T - l l 6 ° C .  (b)  Containing martensite,  * 22MPa.m^, T - l l 6 ° C  76  Matching fractographs from opposite sides of the crack i n 25% cold r o l l e d (at 200 C.) TRIP G  steel. 43  77  800x  E f f e c t of stress i n t e n s i t y on the fractography of 25% cold r o l l e d (at 200°C.) TRIP s t e e l a t open c i r c u i t potentials,  44  lOOOx.  (a)  No applied load, T - 154°C.  (b)  K  (c)  K •- 40MPa.m , T - 154°C  x  - 8.5MPa.m^, T - 154°C. 79  2  x  E f f e c t of temperature on the fractography of the 25% c o l d r o l l e d (at 200°C.) TRIP s t e e l . (a) (b)  45  T ~ 154°C., K, - 8.2MPa.m^" _ 1 i T - 119 C , K - 6.6MPa.m  80  T  x  E f f e c t s of cold work and stress i n t e n s i t y on the fractography of 310 s t a i n l e s s s t e e l .  46  lOOOx.  400x.  - 20MPa.m^, T - 154°C.  (a)  Annealed,  (b)  50% cold r o l l e d , ^  - 25MPa.m^, T - 154°C.  (c)  50% cold r o l l e d ,  «= 50MPa.m \ T - 154°C 2  81  E f f e c t s of temperature and applied p o t e n t i a l on the fractography of 50% cold r o l l e d 310 s t a i n l e s s steel. (a)  400x. - 25MPa.m^, E » -0.27V(S.C.E.)  T - 120°C, x  (b)  T - 154°C., K  x  4  - 25MPa.m% E - -0.325V(S.C.E.)  82  -xiiiFIGURE 47  PAGE V-K  1  plot for 7039-T64 aluminum alloy in 5M  aqueous  solution showing two stress intensity  independent regions and the effect of temperature (after Hyatt and Speidel [5]) 48  90  Relationship between the heat of adsorption and the activation energy for chemisorption (after Glasstone, Laidler and Eyring [56])  49  Schematic Evens diagram illustrating the effect of stress intensity on the corrosion potential  50  98  Schematic representation of the effect of adsorption at the crack t i p . .  Al  95  107  Effect of the downward growth of the cracks on the measurements of V and K.  115  -1-  1. INTRODUCTION 1.1 General Description Stress corrosion cracking (S.C.C.) i s a time dependent, low energy fracture process which may occur in materials subject to the conjoint action of a tensile stress and a specific corrosive environment. Some common examples of susceptible systems are» (i) (ii) (iii) (iv)  Copper alloys In the presence of ammonia and water. Aluminum alloys in aqueous solutions containing chloride ions, Mild steel i n caustic solutions. Austenitic stainless steels in aqueous solutions containing  chloride ions. S.C.C. is a particularly annoying form of corrosion for several reasonst (i)  S.C.C. reduces the usefulness of materials having especially  desirable properties such as high strength to weight ratios, e.g., 7000 series aluminum alloys, Ti-Al-Mo alloys and 4340 type steels, or excellent corrosion resistance (i.e., to general dissolution), e.g.» (ii)  300 series stainless steels. Cracking can take place with l i t t l e or no applied stress  (the residual stress from machining, welding or quenching being sufficient), and with only a few parts per million of aggressive species present. (iii)  Cracks may propagate rapidly once they have been initiated,  -2making them d i f f i c u l t to detect prior to failure. S.C.C. has been responsible for a large number of service failures during the last 35 years and has been estimated to have cost the chemical industry in the U.S.A. alone over $30 million annually in maintenance and replacement of failed equipment [l]. The knowledge gleaned from failure analysis, as well as from experimental work, has enabled designers, manufacturers and corrosion engineers to reduce considerably the incidence of failure due to S.C.C. However, there is as yet  no theory that  can predict which materials w i l l be susceptible in a particular environment, or even how minor changes in an environment w i l l affect the resistance of a given material to S.C.C. 1.2  Fracture Mechanics Approach to S.C.C. Prior to 1965, most tests of susceptibility to S.C.C. were  performed on smooth specimens (e.g., U-bend and tensile specimens), the time to failure being the measure of susceptibility.  While  these tests have provided a great deal of useful, practical information, from the point of view of improving understanding of the mechanisims involved in S.C.C. they have several disadvantages, viz: (i)  The time to failure usually includes the crack initiation  time, which, especially in the case of thin specimens, may be a large fraction of the total time. (ii)  The stress i s not well defined (usually given as the net  section stress prior to cracking), and varies in an unknown manner as the specimen cracks.  -3(iii)  In the case of U-bend specimens and constant load tensile  specimens, overload failure w i l l occur, with less S.C.C. having taken place, i f either the i n i t i a l stress i s increased or the ultimate tensile strength i s reduced. The fracture mechanics approach to S.C.C. overcomes these disadvantages by studying the propagation of single planar cracks. The stress around the crack t i p i s measured by the stress intensity factor, designated 'K*. When the crack opens normal to the crack plane, as i s the case i n S.C.C, i t i s called opening mode I, and the stress intensity factor i s written as K^. For a straight through centre crack of length 2a, or an edge crack of length a, with a straight crack front  i n an infinite plate having a  tensile stress c normal to and distant from  the crack, the stress  intensity factor i s given by \_2~\  t  1.  = O ( -na )  2  Thus,  has the dimensions of STRESS x (LENGTH) . 2  In practice,  for small test specimens, i t i s given by  • % -1 .  Y(a)  where P i s the applied load B i s the specimen thickness and Y(a) i s a function of the crack length and the specimen geometry. The measure of susceptibility to cracking i s the crack propagation rate or crack velocity (V), and, since i t can vary over several orders of magnitude, data i s normally displayed graphically as log V against K,. The fracture mechanics  -4approach has been used to study the S.C.C. of the following materialsi  soda lime and other glasses  £3»*0>  various aluminum  alloys [ 5 » 6 ] , titanium alloys [ 7 ] , high strength steels [8], 70-30 brass [9] and an Fe-Mn-Cr austenitic steel [lO].  A l l these  materials have V-K^ plots similar to that shown in Fig. 1, although not a l l exhibited every feature. With reference to Fig. 1, in Region I the velocity increases with stress intensity, the variation often being approximated by a straight line on the log V vs.K^ graph. At low velocities the curve may turn to be parallel to the ordinate, in which case a threshold stress intensity, defined.  K-J^QQ  t  can be  If cracking continues to lower and lower stress  intensities an arbitrary  K^CJQQ  i  s  sometimes defined as the  stress Intensity resulting in some arbitrarily small velocity. In Region II the crack velocity i s constant and thus independent of stress intensity.  In Region III the crack velocity  increases rapidly with stress intensity and i s independent of the environment.  The transition from region to region may not  be as sharp as shown by the solid curve in Fig. 1, but may be quite gradual as shown by the broken line.  instead  The reason  for this may be that any local variation in stress intensity along the crack front w i l l smooth out any sudden changes in dV/dK^. Possible causes of variations could be changes i n stress state at the edges of the specimen, local changes in crack length and variations in the residual stress from grain to grain in cold worked material.  -5-  C9 O  ^ISCC  S T R E S S  FIGURE 1  INTENSITY  T y p i c a l r e l a t i o n s h i p between stress  corrosion  cracking v e l o c i t y and stress i n t e n s i t y showing Regions I , I I and I I I and K .  gcc  .  -6-  Most of this type of research has been carried out by Industry or other organisations concerned with the practical problems of S.C.C. Consequently much of the work has dealt with alloy development and measurement of K - ^ Q values, e.g., Boeing Airplane Co. on aluminum and titanium alloys, U.S. Naval Research Laboratory on high strength steels, and Brown Boveri Research Centre on Fe-Mn-Cr alloys. From a mechanistic point of view, however, the real merit of the fracture mechanics approach to S.C.C. i s i n i t s ability to distinguish stress dependent mechanisms (Region I) from stress independent phenomena (Region I i ) .  This i s best illustrated  by an example. Speidel obtained activation energies of 27 kcal/g.mole and 3.8 kcal/g.mole for Regions I and II respectively, for a 7079-T6 aluminum alloy in 3M aqueous potassium iodide solution £6].  Further, by determining the  effects of iodide concentration and solution viscosity on the Region II velocity, he was able to show that the stress independent cracking rate was controlled by mass transport i n the solution f i l l i n g the crack. Speidel did not come to any conclusion about the mechanism for Region I. On the other hand, Helfrich obtained activation energies from time to failure tests of a 7039-T6 alloy in a IN sodium chloride solution [ l l ] .  His values ranged from 20.4 kcal/g.mole  at 15 k s i to 16.9 kcal/g.mole at 55 ksi.  These values together  with a calculated activation volume of 28.5 c.c./g.mole led him to the conclusion that the rate controlling process for inter-  -7granular stress corrosion of the 7039 aluminum alloy might he based on stress and thermally activated dissolution of MgZn  2  at the grain boundaries.  It i s interesting to note that Helfrich's  activation energies l i e between those of Speidel for Regions I and II and that the decrease with increasing stress could be explained by the fact that at higher stresses more time would be spent in Region II (lower activation energy). 1.3  The Chloride Cracking of Fe-Cr-Nl Alloys The f i r s t reported industrial failure of austenitic  stainless steel, that was attributed to S.C.C, took place in 19^0 [12].  Since then i t has been well established that most  failures occur in the presence of chloride ions at temperatures above about 60°C £l3 l4]. r  However, other environments such as  sulphuric acid [15] 1 caustic solutions [ l 6 ] and low pH chloride solutions at room temperature [17] have also been shown to cause cracking. During the last 30 years much effort has gone into trying to understand the mechanisms involved in the chloride cracking of austenitic Fe-Cr-Ni alloys.  Many time to failure tests have been  carried out, as well as numerous phenomenological experiments, and, on the basis of these, several theories have been proposed. Rather than trying to summarise a l l of this work, only two of the more important aspects w i l l be considered at this stage. F u l l details of the previously published studies are presented in excellemt reviews of the subject by Latanision and Staehle  -8[18], Nielsen [ l ] and Staehle [19]. Nickel extends the time to failure [ 2 0 , 2 l ] but as yet i t s precise role has not been satisfactorily explained. Nickel i s more noble than either iron or chromium and according to Latanision and Staehle [ l 8 ] the crack sides might become enriched in nickel. Freshly exposed metal at the crack t i p i s then anodic to the crack sides and dissolves.  At higher nickel contents the driving  force for this dissolution i s reduced and so cracking i s slower. There i s , however, no good evidence for nickel enrichment on the crack sides and i t has now been established that the crack sides are covered with a passive oxide film [ 2 2 , 2 3 ] . Nickel has been shown to raise the stacking fault energy (S.F.E.) of austenitic Fe-Cr-Ni alloys [24],  Swann and Nutting  [ 2 5 ] demonstrated preferential dissolution along stacking faults in copper-zinc and copper-aluminum alloys.  It was later proposed  that the low S.F.E. led to co-planar dislocation arrangements which, in turn, result i n coarse s l i p [ 2 6 ] .  This coarse s l i p  has a greater chance of breaking the oxide film on the surface than i t would i f the slip was finer.  However, co-planar dis-  locations have been observed in the higher nickel alloys in spite of their higher S.F.E. [ 2 7 ] .  Thus, while nickel w i l l increase  the S.F.E., i t i s not clear exactly how the S.F.E. affects the susceptibility to S.C.C. It i s possible that nickel might affect the properties of the oxide film formed on the stress corrosion fracture surface.  Properties such as the conductivity and ductility  -9of the film are important ln film rupture models of S.C.C. [28]. The structure and composition of the films formed on the stress corrosion fracture surface of 304, 316 and 310 alloys in MgCl solution have been studied by Nikiforuk [29].  2  In a l l cases he  found the structure to be that of a chromium enriched spinel and that the compositions were similar.  Thus, i f film rupture  does occur, then the effect of nickel i s more likely to be on some other property such as the slip step height or the dissolution rate and not on i t s effect on the passive film. The role of hydrogen in the S.C.C. of austenitic stainless steels i s uncertain. The pH within the crack has been measured to be 1.0 or less in the case of 304 alloy cracking in boiling MgCl solution a,t 154°C. [ 3 0 ] .  This low pH results frpm hydrolysis  2  of the corrosion products [ 3 l ] and means that proton reduction, H  +  + e~5=? H  ,  ads is possible within the crack.  That this i s in fact the cathodic  reaction has been confirmed by the observation of hydrogen bubbles emanating from stress corrosion cracks [ l , 3 2 ] . Holzworth and Louthan [33] and Vaughn et a l . [34] showed that cathodic charging with hydrogen resulted in transformation of 304L alloy to € and a'phases. Birley and Tromans [35] detected a'martensite on the stress corrosion fracture surface of 304L stainless steel and showed that the topography of the fracture surface was consistent with the presence of a'martensite laths. Since a does not normally form under tensile stresses at the temperature of the test, they suggested that hydrogen absorbed  -10-  into the lattice might be responsible for the transformation. It i s not clear, however, whether the crack propagates through the a' phase by a hydrogen embrittlement mechanism or by dissolution. On the other hand, there i s no evidence for martensite involvement in the S.C.C, of the higher nickel alloys such as 310 and Inconel 600 [36]. Other mechanisms by which absorbed hydrogen might promote cracking have been proposed [l9»37]i but as yet there i s no consensus as to whether or not absorbed hydrogen i s a necessary feature of the S.C.C. of austenitic Fe-Cr-Ni alloys. 1.4  Choice of Materials and Scope of the Present Work Three alloys were selected for this investigation, namely  AISI 316, AISI 310 and a TRIP* s t e e l  v  The 316 and 310 alloys  are both known to crack in hot chloride solutions.  However,  the 310 with 9# more nickel takes approximately ten times as long to f a i l .  The TRIP steel was chosen because of i t s greater  tendency to form martensite on deformation.  The cracking of the  TRIP steel was examined in the f u l l y austenitic condition as well as partially transformed to martensite, but was not studied in the TRIP condition (i.e., both work and precipitation hardened by deforming 80% at400°C.)  As far as i s known there i s no  TRIP i s an acronym for Transformation Induced Plasticity and was proposed by Zackay [38]. TRIP steels make use of the strain induced martensite transformation to delay the onset of necking and thereby to increase the ductility.  -11-  published data on the susceptibility to S.C.C. of the TRIP steel, although there i s some information available on the hydrogen embrittlement of these steels [39]. Cracking of a l l the steels was examined in both the f u l l y annealed and cold worked conditions, however, quantitative results were only obtained from the cold worked specimens. The environment used for cracking was a concentrated solution of magnesium chloride in water having a boiling point of  154°C, Although such a medium i s not common industrially,  the cracking susceptibilities of alloys determined i n magnesium chloride compare well with those obtained from long term tests and from industrial experience [4o],  This environment also has the  advantage that reasonably fast cracking rates can be achieved without the need to go to higher than atmospheric pressure. The other experimental variables studied were the temperature, the electrochemical potential and the amount of prior deformation to the alloys.  The effects of these variables on  the topography of the stress corrosion fracture surface were also examined. It was hoped that by applying the fracture mechanics approach to S.C.C., to the chloride cracking of Fe-Cr-Ni alloys, that new insight could be gained into the cracking mechanism.  -122.  2.1  EXPERIMENTAL  Specimen Geometry The type of specimen most often used for measuring stress  corrosion crack velocities as a function of stress intensity i s the double cantilever beam specimen (D.C.B.) \_5~}.  See Fig. 2(a).  When wedge loaded (i.e.* constant deflection) i t has several important advantages: (i)  It i s compact while allowing long cracks to be studied,  (ii)  It gives a large range of stress intensity, decreasing  from the i n i t i a l value to that at which the crack stops. (iii)  ( igrjrj) • K  Initiation times are short. Thus, accurate data can be obtained reasonably quickly  without requiring elaborate apparatus or a great deal of material. For these reasons D.C.B. specimens were used for preliminary tests in the present study.  Cracking to one side, however, was always  favoured over cracking along the centre plane (see Fig. 3)»  in spite  of changing several different experimental parameters such as: the degree of cold work, the specimen thickness, the i n i t i a l crack length tp specimen width ratio, the surface finish and the manner in which the pre-crack was introduced.  In an attempt to  benefit from the observation of these side cracks, the specimen shown in Fig. 2.2  2(b) was tried and proved to be more useful.  Stress Intensity Calibration The stress intensity calibration for this specimen was  obtained by the method originally used by Irwin and Kies [ 4 l ] ,  -13-  L= 6-4 cm W = 3-2cm BENDING MOMENT  = Pl  2P BENDING MOMENT  P  P V//////////////////777  P  P  " 2P  Specimens used i n the present work and by others to measure crack v e l o c i t i e s as a function o f stress i n t e n s i t y . (a)  Double Cantilever Beam Specimen  (b)  Specimen used i n the present work  (c)  Two 3 point bend specimens face to face to  show the s i m i l a r i t y with (b)  -14-  PIGURE 3  Double cantilever beam specimen showing side cracking. Fatigue crack at t i p of notch.  Other cracks are S.C.G.  -15in which the specimen compliance i s measured as a function of the crack length. Use i s then made of the following relationships [42,43]; K  x  - (GM )  (1)  *  dC 2B ' da  and G where  2  (2)  «= stress intensity factor G  = energy release rate  M  = elastic modulus  P  «? load  G  = compliance  a  e  crack length  Equations (l) and (2) can be combined to give; K  3/2  x  . B  Bj~M dC 2 * da  (3)  The crack was made with a fine jewellers saw and the load and displacement measurements were made with an Instron tensile machine equipped with a D c e l l and a 0,2" crack opening displacement gauge.  The compliance curve i s shown in Fig. 4, A  polynomial of degree six was fitted to the data points and used to calculate the expression on the right hand side of Eqn. 3This i s plotted as a function of crack length to specimen width ratio (a/fy) i n Fig. 5. It i s interesting to compare this relationship, between  2 — Irwin [43] included a factor ( 1-1/ ) " in the right hand side of this equation, (y i s Poisson's ratio). 2  -16-  0  0-5 C R A C K  10  1-5  L E N G T H  20  (cm)  FIGURE 4 Compliance of the specimen used ln the present work (see Fig. 2fb))  as a function of crack length.  -17-  14  0  01  0-2  0-3  0-4  0-5  0 6  Q W FIGURE 5  Stress intensity calibrations f o r the specimen used in the present work (Fig. 2(b)) and for a three point bend specimen of similar size (Fig. 2(c)). 3 and 4.  See Eqns.  -18-  stress Intensity and crack length, with the case of three point bending.  The latter has been calibrated by means of boundary  collocation [44] which gives, for the case of S/W  = 4 (see Fig. 2(c)),  the relationship} h  • ^/l P  = 12B a Y(a/W ) 2  (4)  2  2  sir  r  where Y ( a/W ) = 1.93-3.07(a/V)+l4.53(a/M) -25.1l(a/V) +25.8(a/i0' 2  B  = thickness  W  - width  S  = span  3  The right hand side of Eqn. 4 i s also plotted agianst a/W Fig. 5» for the case of S * 2L.  IJ  in  There i s good agreement between  equations (3) and (4) which i s perhaps not too surprising considering the similarities in the geometry of the two specimens. The boundary collocation calibration for three point bending i s , therefore, applicable to the specimen used in the present study. 2.3  Specimen Thickness A criterion for the minimum thickness of specimens used  for fracture mechanics type stress corrosion tests has not yet been established, as i t has in the case of fracture mechanics. Since neither pop-in nor slant cracks are found in stress corrosion  2 i t does not seem reasonable to simply adopt the value of 2.5(K /o ) 1  y  used in fracture mechanics.  Some work has been done on the effect  of specimen thickness on K^g  GC  values in a titanium alloy, a  steel and an aluminum alloy, and the findings have been reviewed by Smith and Piper [45].  The results from these  -19investigations are conflicting.  No thickness effect was observed  with the titanium and aluminum alloys, however, in the case of the steel, a minimum K^g  GG  was found i f the thickness was greater  than or equal to 2.5(K^/o ) .  There i s no data available on the  effect of thickness on crack velocities. In view of this, a specimen thickness of 3.2mm. was selected as a compromise among four opposing factors: (i)  This thickness resulted in an almost straight crack front,  meaning that the crack lengths measured at the surface were representative of the crack throughout the thickness.  This is  not necessarily the case with thicker samples [46], (ii)  This thickness prevented the specimen from twisting  appreciably out of the plane of the specimen on loading (i.e., from introducing K J J J opening mode). (iii)  It made the f i n a l cold rolling and stress relieving  relatively straightforward. It might also be noted that the thickness of 3.2mm. does satisfy the condition, B > 2.5(K../c ) , for a l l the materials y used for quantitative measurements, up to a stress intensity of K j ^ 19MPa.m . 2  2.4  (Yield strength data i s given in Table II).  Advantages and Limitations of the Specimen Geometry There are several advantages to the specimen geometry  chosenj (i)  The specimen i s quite compact, easily loaded in the  corrosive environment and has a useful crack length of approximately l.?cm. (i.e., from 0.3cm. to 2,0cm.).  -20-  (ii)  The stress intensity changes by approximately a factor  of five, over the useful length of the crack, i f the load i s kept constant, see F i g . 5. (iii)  The specimen w i l l not be as prone to corrosion product *  wedging  as the self loading D.C.B, specimen since the stress  intensity increases rather than decreasing as the crack grows. There are however, two drawbacks to this type of specimen} (i)  The initiation time for cracking may be long since cracking  proceeds from low to high stress intensities, (ii)  ^2SCC  v a  ^-  u e s  w  i l l be time consuming to determine since  each test w i l l only give an upper or lower limit for K-^g^ 2.5  Materials and Specimen Preparation The compositions of the three alloys studied are shown in  Table 1.  The Jl6 and 310 stainless steels were commercial alloys  purchased as 0.64 cm, thick, hot rolled, annealed, and pickled plate. heat.  The TRIP steel was produced by vacuum melting a 2-^kg. The resulting ingot was then electro-slag refined and  hot rolled to a thickness of 0.64cm. Further reduction of a l l three alloys was by cold rolling, with or without an intermediate anneal, in order to arrive at the f i n a l thickness with the desired amount of cold work.  A l l annealing was carried out in sealed  Sen-Pak stainless steel containers at 1050°C. for one hour and was followed by quenching in water.  The TRIP steel was cold  Accumulation of corrosion product in the crack can produce sufficient hydrostatic force to invalidate the calculated stress intensities i n wedge loaded D.C.B. specimens, especially in the case of stainless steels, uranium and aluminum alloys [47],  -21-  TABLE I Alloy Compositions  TRIP wt %  316 wt %  310 wt %  Fe  66.35  53.67  73.20  Cr  16.85  21.8  11.0  Ni  12.08  21.8  8.8  ELEMENT  3.66  Mo  2.13  Mn  1.58  1.64  2.01  Si  0.89  1.10  1.10  C  0.0?  0.05  0.22  S  0.011  0.019  0.005  P  0.036  0.02  0.006  -22rolled at 200°C. when a f u l l y austenitic structure was required, and at room temperature when a partially martensitic structure was desired.  The crystal structure of a l l the alloys was checked  "by X-ray diffraction, with only the TRIP steel rolled at room temperature displaying an a peak (i.e., b.c.c. phase).  Also,  only the TRIP steel rolled at room temperature was found to be magnetic.  In order to reduce the residual stresses l e f t after  rolling, the rolled strip was strained 2% by pulling in tension. This "stress relieving" treatment was only carried out on material used to measure Region I (i.e., stress dependent) velocities. Table II l i s t s the mechanical properties of the different alloys and processing treatments used. Specimens were cut from the rolled strip, so that the cracks would, be either parallel or normal to the rolling direction, and were then machined to a f i n a l size of 11 x 6.5cm,  The central  slot and loading pin holes were then machined and the side notches cut, either by spark machining or with a jewellers saw, the normal length being about Jmm. Occasionally, fatigued pre-cracks were The specimens were then polished to a 600 grit finish and  used.  a 2cm. long millimetre scale  lightly scribed on each side.  The  entire specimen, except for the ends of the side notches, was then coated in transparent lacquer in order to prevent secondary cracks from initiating ahead of the primary ones. 2.6  Apparatus and Testing Procedure A photograph of the experimental arrangement i s shown in :  Fig. 6.  The tests were carried out in a 500ml. Pyrex glass c e l l  TABLE II Mechanical Properties of the Various Alloys and Processing Treatments (Tensile data measured perpendicular to the-rolling direction at 150°C.)  ALLOY  % REDUCTION  DEFORMATION TEMPERATURE  0.2% OFFSET YIELD STRENGTH  °C  MPa  *  ULTIMATE TENSILE STRENGTH  % ELONGATION  (G.L. - 2.5cm.)  HARDNESS Rc  MPa  29.0  316  25  R.T.  534  903  21.0  316  50  R.T.  903  1230  9.0  310  50  R.T.  696  883  10,5  26.5  TRIP  25  200°C.  580  952  18.0  36.5  TRIP  25  R.T.  655  1080  9.5  42.0  IMPa = 10 Newtons/metre  = 0.145 ksi  37  -24-  FIGURE 6  Experimental arrangement showing S.G.C. c e l l , r e f l u x condenser and load s t a b i l i z i n g spring mounted i n t e n s i l e t e s t i n g machine.  -25-  fitted with a Teflon cap through which two stainless steel grips were free to slide.  This c e l l was mounted in a Hounsfield tensile  testing machine i n series with an elastically 'soft* spring, which was used to reduce the drop in load which occurred as the cracks lengthened.  The c e l l was heated by a hot.plate and an immersion  heater, the latter being used to control the temperature to within +  -|C. by means of a temperature controller.(Model ?1A Yellow 0  Springs Industrial Go.).  The thermistor probe was immersed in  the solution close to the cracks.  The solution was stirred by a  magnetic s t i r r e r in order to reduce temperature and concentration gradients.  A reflux condenser was used to minimise the water loss  from the c e l l .  Electrochemical potentials were measured with  respect to a saturated calomel electrode at room temperature. A luggin capilliary was fixed in position about 1mm. from the specimen, midway between the cracks, and was connected to the reference electrode by means of a magnesium chloride salt bridge. counter electrode was used. Research model 173)  w  a  s  A platinum  A potentiostat (Princeton Applied  used to measure and control the potential.  Potentials quoted are with respect to the saturated calomel eleptrode and are not corrected for either liquid junction or thermal gradient potentials. To start a. test, a specimen was mounted in the c e l l and 350ml. of solution was added.  The solution was prepared by adding  d i s t i l l e d water to reagent grade magnesium chloride hexahydrate and heating until the boiling point reached 154°C. responds to a concentration of  MgCl_ [48],  This cor-  The c e l l was  -26heated and time allowed for temperature equilibrium to be attained. The mercury column in the tensile machine was zeroed and the desired load applied, or in some cases, to reduce initiation times, a greater load was applied and then reduced once the cracks had initiated.  If a temperature of less than 154°C. was required,  the solution was allowed to cool to the desired temperature and the temperature controller set. In spite of the reflux condenser some water vapour was lost from the c e l l .  For the boiling solution, sufficient d i s t i l l e d  water to restore the temperature to 154°G, was periodically added. At the lower temperatures the level of the solution was marked on the side of the c e l l and d i s t i l l e d water added periodically to maintain the level.  Typically, a few c c . of water were re-  quired each day. Measurements of the crack lengths were read, by eye, from the scales on both sides of the specimen, the cylindrical half c e l l acting as an adequate magnifying glass.  The cracks were  referred to as •north* and 'south* in order to distinguish one from the other. On completion of the experiment, the specimen was removed from the solution, pulled carefully apart i n the tensile machine, washed ultrasonically in hot d i s t i l l e d water, dried and stored in a dessicator.  The specimens were subsequently mounted for  ovservation of the fracture surfaces in a Scanning Electron Microscope (S.E.M,), (E T E C Corporation Auotscan). The scribed scale could be seen when the specimen was t i l t e d appropriately in  -27th e S.E.M. and so t h i s scale was used to locate p a r t i c u l a r portions of the f r a c t u r e surface.  Quantitative measurements  of the r e l a t i v e amounts of intergranular S.C.C, transgranular S.C.C. and d u c t i l e overload f r a c t u r e were made by measuring l i n e a r intercepts of each d i r e c t l y from the viewing screen.  Crack velocities were obtained from the crack length versus time data by f i t t i n g curves to the data points using the method of least squares and differentiating these with respect to time.  For Region I a polynomial of degree less  than or equal to six was used while f o r Region II a straight line f i t was obtained.  -283. 3.1  RESULTS  Stress Corrosion Crack Morphology 3.1.1  316 Stainless Steel  The nature of the cracks obtained in 316 stainless steel depended on the degree of cold work and on whether the crack was parallel or perpendicular to the rolling direction.  In the case  of the 2$% and 50% cold rolled material with the cracks propagating in the rolling direction, the cracks were usually quite straight and unbranched, although there was some microbranching, especially at high stress intensities (K^ > ^OMPa.m  2  ),  An  example i s shown in Fig. ?(a). The crack front was usually quite straight and perpendicular to the propagating direction, Fig. 7(b). The roughness of the fracture surface increased with stress, as shown In Fig. 7(c). There was a tendency, particularly at low 1.  stress intensities (K^ < 15MPa.m ), for the cracks to grow downi;  wards at an angle to the normal to the edge of the sample. angle was usually less than 10°.  This  However, when the rolling direction  was normal to the intended cracking direction, the cracks grew downwards at an angle of about ^5° and tended to branch.  An  example of this i s shown in Fig. 8(a). Consequently, quantitative data could only be obtained from.specimens in which the crack was parallel to the rolling direction. Cracks obtained in annealed 316 tended to branch, lead on one side of the specimen and to twist and "bend out of the crack *  i i IMPa.m - 0.909 k s l ( i n . ) 2  2  -29-  (a)  NOTCH  S.C.C.  OVERLOAD FRACTURE  (c)  K  = 15MPa.m  K  2  x  - 30MPa.m  2  ±  Crack morphology for cracking parallel to the r o l l i n g direction i n 2% cold rolled 316 stainless steel. 7x.  (a)  Side view.  (b)  Fracture surface. 4 . 8 x .  (c)  End view at 15 and 30MPa.m . 23x .  4 2  -30-  FIGURE 8  Crack morphologies i n Jib s t a i n l e s s s t e e l .  12x.  (a)  Cracking perpendicular to the r o l l i n g d i r e c t i o n .  (b)  Cracking i n the annealed a l l o y .  (The crack has  been opened to f a c i l i t a t e observation of cracking.)  -31plane.  An example of the branching i s shown i n F i g . 8(b).  quantitative measurements were made of cracking i n annealed  Tic 316  since the branching would have affected both the crack v e l o c i t y and stress i n t e n s i t y . 3-1.2  TRIP S t e e l  Cracking i n the TRIP s t e e l was very s i m i l a r to that found i n the 31& s t a i n l e s s s t e e l .  An example of a crack propagating  p a r a l l e l to the r o l l i n g d i r e c t i o n i n TRIP deformed Z$% at 200°C. i s shown i n F i g , 9(a).  Similar straight and unbranched cracks  were found i n the TRIP s t e e l r o l l e d a t 20°C.  Cracks i n the  annealed TRIP were branched and unsuitable f o r quantitative measurements. 3.1.3  310 Stainless Steel  An example of cracking i n 50% cold r o l l e d 3 1 0 i s shown i n F i g . 9(b). branching.  The crack i s not s t r a i g h t and displays multiple  In spite of t h i s , some measurements were made of the  crack v e l o c i t y at high Region I I .  i n the hope that the cracking was i n  Fatigued pre~cracks were used since these were help-  f u l i n i n i t i a t i n g single cracks. also  branched.  3.2  Considerations of Errors 3.2.1  Cracking i n annealed 310 was  Errors i n Crack Length Measurements  Usually the crack t i p could be c l e a r l y seen and the p o s i t i o n determined to within ±0,02cm.  At very high stress  This error was estimated by comparing readings taken just before and a f t e r unloading and removing the specimens from the solution.  -32-  FIGURE 9  Crack morphology in the 310 and TRIP steels. (a)  TRIP - Z% cold rolled at 200°C.  parallel to the rolling direction. (b)  310 - 50% cold rolled.  rolling direction.  17x.  Cracking  7x.  Cracking parallel to the  -33intensities, especially in the softer materials, the size of the plastic zone at the crack t i p made determination of the crack length more d i f f i c u l t resulting in an error of approximately ±0.C4cm. 3.2,2 Errors due to the Cracks Growing Downwards If the crack grows downwards, then i t crosses the scale at an angle which w i l l result i n an underestimation of the velocity.  I t can be shown that, i f the crack makes an angle of  10° with the horizontal, then the measured velocity w i l l be i n error by about 1.5% and for an angle of 20° the error w i l l be about 5%>  See Appendix 1.  Also, as the crack grows downwards, the length of the moment arm (L, see Fig. 2(b)) Increases causing the actual stress intensity to be greater than that calculated using the i n i t i a l value of L. It can be shown that for a 2cm. long crack at an angle of 10°,  i s underestimated by about 5%-  The downward growth of  the crack also results in a! measured ligament length which i s less than the actual length. This w i l l result i n a calculated stress intensity greater than the true value.  The magnitude of this  error can be estimated, and for a 2cm. long crack at 10° to the horizontal, i t i s found to be 1.5$. 3.3  See Appendix 1.  Kinetic Data 3.3.1  316 Stainless Steel  3.3.1.1 Effect of Stress Intensity on Crack Velocity It was observed that both stress dependent (Regions I and III) and stress independent (Region II) stress corrosion  cracking were present l n cold worked Jl6 exposed to concentrated magnesium chloride solution.  F i g s . 10 and 11 show the v a r i a t i o n  with time of crack length and load f o r two specimens of 25% cold r o l l e d 316 cracking a t 15^0. with i n i t i a l loads of 500N and 1200N respectively. against  i n F i g . 12.  This same data i s plotted as l o g ^ V At low stress i n t e n s i t i e s the v e l o c i t y  increases r a p i d l y f o r small increases i n K^, Values of K-jV V i s independent of K^  while a t higher stress  t  i n t e n s i t i e s the v e l o c i t y increases again.  Thus, Region I and  I I and the s t a r t of Region I I I are present. not studied i n the present work).  At intermediate  (Region I I I was  I f the specimen was unloaded  a f t e r cracking had i n i t i a t e d , i t was found that the crack continued to grow slowly. a K  1 S G C  Thus, i t was not possible to determine  value. 3.3.1.2  E f f e c t of P l a s t i c  Deformation  F i g . 13 shows the v a r i a t i o n of crack v e l o c i t y with stress i n t e n s i t y f o r two specimens of 50% cold r o l l e d 316 a t 15^°G. The v a r i a t i o n i s quite s i m i l a r to that found i n the 25% reduced material, except that Region I I extends to higher stress intensities.  A comparision of the Region I I v e l o c i t i e s of the 25%  and 50% reduced a l l o y was made at three temperatures by cracking several specimens.  The mean values obtained are shown i n Table I I I .  At a l l temperatures  the 50% reduced material cracked about 30%  f a s t e r than the 25% reduced material. 3.3.1.3  E f f e c t of Temperature  The data i n Table I I I also show  the e f f e c t of temperature  T I M E (hours) FIGURE  10  Variation of crack length and load with time f o r a specimen of 25% cold r o l l e d 3 i 6 stainless steel - having an i n i t i a l load "of 500N, -  at 154°c.  HgCl  solution b a i l i n g - '  1200  800  Q < O  4400  JL  0  •  LOAD  •  CRACK  6  LENGTH  1 8  T I M E (hours) FIGURE 11  V a r i a t i o n o f c r a c k l e n g t h and load with t i m e f o r a specimen o f ?.%• c o l d 316 s t a i n l e s s s t e e l a t 15k°C. having an i n i t i a l b o l l i n s a t 1C4°C.  lo*.d o f 1200N,  0  /0  rolled  Mg^l.^ s o l u t i o n  'TGURE 12  Effect of stress intensity on c r a c k velocity f.or-7.% cold r o l l e d 316 stainless steel'tfxponed t o boiling KgOL, solution  at 1S^°0.  -39-  TABLE III Variation with Temperature of Region II Velocities for 25% and 50% Cold Rolled 316 Stainless Steel  136°C  154°C.  116°C.  50%  V  m/s  >.7pxl0~  7  2.20xl0~  cold  S.D.  m/s  0.4lxl0~  7  0.20xl0~  rolled  rolled  7  3  •9 154°C.  25% cold  7  m/s  3.80xlC~  S.D.  m/s  0.26xl.(f  7  7  5  V i s the average Region II velocity S.D. i s the standard deviation of V n i s the number of velocity measurements  1.56xl0~  0.034xl0"  7  116°C. 7  0.1?xlO" 3  -7  3 135°C  V  0.8Q5xlO  ?  0.59xl0"  7  0.048xl0"  7  -toon the Region II velocities for "both the cold rolled 316 alloys. The data was plotted as log V against l/T in Fig. Ik to see i f 1Q  the crack velocity followed an Arrhenius rate law of the form: V -V e ^ o  where V i s the velocity  T  V i s a constant o Q i s the activation energy R i s the gas constant and  T i s the temperature (°K.) As can he seen, both sets of points l i e close to straight  lines which yield apparent activation energies of l 6 . 0 i 0 . 5 kcal/g.mole for the 25% reduced 316 and 15.1±0.5 kcal/g.mole for the 50% reduced Jl6. Fig. 15 shows the effect of temperature on Region I. At o 116 G. and  A  < 15MPa.m, the cracking rate was extremely slow. 2  Because of this and the steepness of the log V -  slope i t was  not f e l t that the effect of temperature on the Region I crack velocity could be determined with sufficient accuracy within a reasonable length of time. 3.3.1A  Effect of Applied Potential  The effect of applied potential (E) on Region I velocities of 25% cold, rolled 316 alloy at 15^°C. i s also shown in Fig. 15. There does not appear to be a dependence of V on E between -0.25V(S.G.E.) and -0.3V(S.C.E.).  The scatter i n the data,  however, would mask a l l but a large effect. 1  there was no cracking at or below 12MPa.m . ?  For E = -0.325V(S.C.E.)  -41-  FIGURE 14  Arrhenius plot of the Region II velocities of cold rolled 316 stainless steel cracking in a 45^ MgCl, solution.  -42-  50  20  O  x >t  0  5  o o _J  LU >  o 0-2 <t  •  •  T(°C)  E (SCE)  154  -0-25V -0 26V  it  it  A  O  01  5  10 S T R E S S  FIGURE 15  •  II  O  116  15 N T E N S I T Y  -0-275V -0-30V 0/C  20 (MPa/m)  Effect of applied potential and temperature on the Region I cracking of 25$ cold rolled Jl6 stainless steel exposed to a h% MgCl solution. 2  ( 0/C • open circuit potential).  -43The effect of applied potential on Region II velocities of 25/6 cold rolled 316 at 15^°C. i s shown in Figs. 16 and 17. Between -0.25V(S.G.E.) and -0.325V(S.C.E.) there i s no detectable On reducing the potential to -0.35V(S.G.E.),  dependence on E.  the crack slowed down and appeared to stop after a distance of approximately 1mm. 3.3.2  TRIP Steel  3.3.2.1  Effect of Martensite on the Region II Velocity for 2%  Fig. 18 shows the relationship between V and  cold rolled TRIP steel at 15^°C. i n both the f u l l y austenitic and partially martensltlc conditions.  The steel containing no  martensite has a Region II which extends to higher Also, the martensitic TRIP cracks faster In Region II.  values. This  was found to be the case at lower temperatures as well, as shown in Table IV. 3.3.2.2  Effect of Temperature on the Region II Velocities  The data in Table IV also show the increase i n velocity with temperature of the TRIP. shown in Fig. 19.  An Arrhenius plot of the data i s  Both sets of data l i e close to straight lines  having gradients which yield apparent activation energies of 15.5±0.6 kcal/g.mole for the TRIP steel deformed at 200°C. and 15.6±0.6 kcal/g.mole for the TRIP steel deformed at 20°C. 3.3.3  310 Stainless Steel  3.3.3.1  Effect of Stress Intensity on the Crack Velocity  Fig. 20 shows the variation of crack length with time, at 154°G., for a specimen of 5°$ cold rolled 310 stainless steel.  -44-  V=3 77x|0" m/s 7  l l  V = 3-76x|0" rTVS 7  •  NORTH  CRACK  •  SOUTH  CRACK  30<K<:35 MPa/m 0-350V-  0  . 3  4  T I M E  ( h o u r s )  5  Effect of cathodic polarization on the Region II velocity of 2% cold rolled 316 stainless steel cracking in a MgCL solution boiling at 154°C. saturated calomel electrode. juk ~ 10"^ amps).  Potentials are w.r.t. a (o/C «= open c i r c u i t ,  7  -45-  FIGURE 17  Effect of anodic polarization on the Region II velocity of 25% cold rolled 316 stainless steel cracking in a MgClg solution boiling at 154°C, Potentials are w.r.t. a saturated calomel electrode. yUA • 10"  (o/C - open circuit,  amps, mA - 10~ amps). J  -46-  x  3  Austenite  100%  Partially transformed to Martensite  0  10  20  STRESS  FIGURE 18  30  40  INTENSITY  50 (MPa/m)  Effect of stress intensity on crack velocity for Z% cold rolled TRIP steel exposed to MgCl solution 2  boiling at 154°C.  60  -47-  TABLE IV Variation with Temperature of Region II Velocities for 2% Gold Rolled Austenitic and Partially Martensitic TRIP Steel  154°C Austenitic  V  TRIP  n  Partially Martensitic TRIP  V n  m/s  3.28xl0"  7  2 m/s  3.96xl0"  116°G.  135°C 1.46xl0"  7  2 7  2  V i s the average Region II velocity n i s the number of velocity measurements  1.57xl0" 2  o.550xio"  7  2 7  0.656xl0" 2  7  FIGURE 19  Arrhenius p l o t of the Region II v e l o c i t i e s of 2% r o l l e d TRIP s t e e l cracking i n a h%  cold  MgCl_ solution.  -49-  60  NORTH SOUTH  2  3  CRACK "  4  TIME ( D A Y S )  FIGURE 20  Variation of Crack Length, load and stress intensity with time for a specimen of 50% cold rolled 310 stainless steel exposed to boiling MgCl solution 2  at 154°C.  -50Gracking i n the 310 was slower by more than a factor of ten than i t was in either the 316 or TRIP steels at the same stress intensity. After cracking for approximately 1.5mm., the cracks slowed down and continued to crack at a new constant velocity, between one half and one third as fast as the i n i t i a l rate.  This change in  velocity was quite reproducible and was not accompanied by a change l n corrosion potential.  Also, i t was not related to the  fatigue pre-cracks since i t occurred even when these were not used. 3.3.3.2  Effect of Temperature on the Crack Velocity  The effect of temperature on both the i n i t i a l and subsequent velocities i s shown i n Table V. The mean values are plotted as log^ V against l/T in Fig. 21. Q  The data points for  the slower velocity l i e close to a straight line whose gradient gives a value of Q » 18.1±0,8 kcal/g.mole.  The two points for  the i n i t i a l velocity have a similar temperature dependence. 3.4  Corrosion Potentials 3.4.1  Effect of Stress Intensity on the Corrosion Potential  The corrosion potential was found to become more negative (active) on increasing the stress intensity. Fig. 22 shows the change i n potential with time of a specimen of 25$ cold rolled 316 at 154°C. as the load was increased in increments of 220N. The i n i t i a l potential was obtained by initiating cracking, unloading the specimen and leaving i t overnight in the solution, during which time cracking continued slowly.  On loading tfie specimen  the potential decreased (i.e., became more active) quite quickly attaining a new steady state value i n less than five minutes.  -51-  TABLE V Variation with Temperature of I n i t i a l and Subsequent Constant Velocities for 50% Cold Rolled 310 Stainless Steel  154°C. Initial  V  m/s  2.63xl0"  velocity-  S.D.  m/s  0.37xic~  n  8  i.i5xio"  8  8  2  k  Subsequent  V  m/s  l.OlxlO"  8  velocity  S.D.  m/s  O.llxlO"  8  n  120°C.  135°C  6  V Is the average Region II velocity S.D. i s the standard deviation of V n i s the number of velocity measurements  0.395xl0" 2  8  0.156x10" 2  8  FIGURE 21  Arrhenius plot of the crack velocities of 50% cold rolled 310 stainless steel cracking in a k% MgCl, solution.  -0-250  ^  -0-300  ZERO LOAD  220 N  440  i 200  1  LU  N  660 N  i 400  1  600  d (/>  TIME  C/)  (S)  I-I o  > -0-275  _J <  660  N  880  N  I I 00 N  hz  LU -0-325 r-  65a  800  1000  o  12 00  Q_  T I M E (S)  Z o  t o o  o r  -0-275  o  1100  N  660  N  u  -0-325 1300 FIGURE 22  2000  4000  6000  Variation of Corrosion P o t e n t i a l with time as a function of applied  load.  25% cold r o l l e d 316 s t a i n l e s s s t e e l l n b o i l i n g MgCl solution a t 154°C. 2  -54On increasing the load, the potential continued to decrease in a similar fashion.  On decreasing the load to an Intermediate  value, the potential increased immediately, hut required nearly an hour to return to the steady state value.  The steady potentials  are plotted against stress intensity in Fig. 23. 3.4.2  Effect of Composition and Processing Treatment on the  Corrosion Potential Table VI shows the average corrosion potentials of the different alloys and processing treatments at a temperature of o •!• 154 C. and a stress intensity of 30MPa.m . There was considerable variation in these values and none of them are thought to be more accurate than by±5mV,  It can be seen, however, that the TRIP  steel i s the most active and the 310 the most noble. 3.4.3  Effect of Temperature on the Corrosion Potential  Fig. 24 shows the variation of the corrosion potential (E^Qpp) as a function of the temperature for 2$% cold rolled 316 cracking at a stress intensity of 3°MPa..m . 2  3.5  Fractography 3.5.1  316 Stainless Steel  3.5.1.1  Effect of Plastic Deformation  The stress corrosion crack path was solely transgranular (T.G.) in the annealed material, whereas both T.G. and Intergranular (i.G.) cracking were observed in the cold worked material.  A l l S.E.M. fractographs are oriented so that the crack propagation direction i s from l e f t to right.  -55-027,  -0-28 LU O CO tr a: o o LU  -029  -0-30  -0-31 0 FIGURE 23  10  20  30  STRESS INTENSITY  40  50  (MPaVm)  Effect of stress Intensity on the corrosion potential of Z$% cold rolled 316 stainless steel cracking i n boiling MgCl solution at 154°C. 2  -0-28  120  130  140  TEMPERATURE (°C)  FIGURE 24  150  160  Effect of temperature on the corrosion potential of 2% cold rolled 316 stainless steel cracking i n h% MgClg solution at a stress intensity of 25MPa.nJ  -56-  TABLE VI Corrosion Potentials of the Various Alloys and Processing Treatments  ALLOY AND PROCESSING TREATMENT  CORR V(S.C.E.) ,  E  25% cold rolled 316  -0.295  25% cold rolled (200°C.) TRIP  -0.331  25% cold rolled (20°C.) TRIP  -0.333  50% cold rolled 310  -0.270  Values measured at K, = 30MPa.m and T = 154°C. 2  -57Fig. 25 compares cracking under similar conditions in annealed and 2$% and 50$ cold rolled Jl6 stainless steel.  The morpho-  logical features of the T.G. cracking are similar in a l l three conditions, although the fan shaped features described by Neilson  are not so clearly defined in the cold worked samples.  Both the transgranular features and the striations on the intergranular facets were found to match on opposite sides of the cracks as shown in Figs. 26 and 27.  This indicates that these  features were present when the surfaces were formed. 3.5.1.2  Effects of Electrochemical Potential and  Stress Intensity Both the electrochemical potential and the stress intensity influenced the mode of cracking.  Experiments were  carried out at several different applied potentials to determine the effect of each variable separately, since changing the stress intensity also altered the corrosion potential.  Also, i t was  observed that the amount of intergranular cracking was greater in the centre of the crack than at the edges. Consequently, the fraction of I.G. cracking was measured as a function of the distance from the edges of the crack when varying E and K^. data obtained i s displayed on histograms in Fig. 28.  The  It can  be seen that the fraction of I.G. cracking increases both with increasing stress intensity independent of applied potential and with decreasing potential independent of stress intensity. Figs. 29 to 33 show some examples of fractography from the specimens used to obtain the data in Fig. 28. Fig. 34 shows an example of the fractography of a crack  -58-  FIGURE 25  E f f e c t of cold work on the fractography of 316 s t a i n l e s s s t e e l at 15k°C.  kOOx. 1  (a) (b)  Annealed, K. = lOMPa.m 25% cold r o l l e d K-. = 20MPa.m  (c)  50% cold r o l l e d ^  2  2  = 20MPa.m  2  -59-  FIGURE 26  Matching fractographs from opposite sides of the crack in 25$ cold rolled Jl6 stainless steel.  400x.  -6o-  FIGURE 2?  Matching fractographs from opposite sides of the crack i n 2$% cold r o l l e d 316 stainless s t e e l .  lOOOx.  DI-  STRESS INTENSITY (MPa^m) 220  60-  | 33-0  c3  40-  o  •sp 20 •  CM  6i  0 80-  u  CJ CO  6040-  in Is-  20-  CM  0CJ CO  80-  co  cj 6 0 CJ CO  O >  40-  O  o l<  200  ro z  6  L  I  UJ h-  O a  UJ _J  a. FIGURE 28  Effects of applied potential and stress intensity on the amount and distribution of Intergranular cracking. 2.5$ cold rolled 316 stainless steel exposed to boiling MgClg solution at 154°C. Each histogram shows the variation in cracking mode across the width of the specimen.  Q_ <  -62-  FIGURE 2 9  Examples of fractographs of 2|# cold rolled 3 1 6 stainless steel at 154°G.  ^  (a)  Edge of crack  (b)  Centre of crack  - llHPa.m  T t  E - -0.25V (S.C.E.)  400x.  -63-  PIGURE 30  Examples of fractographs of 25$ cold rolled 316 stainless steel at 154°C. K  ±  - 22MPa.m, E = -0.25V (S.C.E.) 400x.  (a)  Edge of crack  (b)  Centre of crack  2  -64-  FIGURE 31  Examples of fractographs of 2% cold rolled 316 stainless steel at 154°C. K  ±  (a)  Edge of crack  (b)  Centre of crack  = llMPa.m"", E = -0.3V (S.C.E.) 2  400x.  -65-  FIGURE 32  Examples of fractographs of 25% cold r o l l e d 316 s t a i n l e s s s t e e l a t 154°G.  KL - 22MPa.m , E = -0.3V  (a)  Edge of crack  (b)  Centre of crack  2  (S.C.E.)  400x.  -66-  FIGURE 33  Examples of fractographs of 2$% cold rolled 316 stainless steel at 154°C. K  - 33MPa.m. Centre of crack. 2  ±  (a)  E - -0.325V (S.C.E.)  (b)  E - -0.35V (S.C.E.)  400x.  -67  FIGURE 34  Example of fractograph of 2% cold rolled 316 stainless steel cracking without an applied load.  800x.  -68propagating in the absence of an applied load at an open-circuit potential of -0.275V(S.C.E.). There i s no evidence of I.G. fracture, which i s in agreement with the prediction from the data in Fig. 28. i  At very high stress intensities (>55MPa.nr ), patches of z  ductile fraoture were observed amongst the stress corroded fracture, Fig. 35. The amount of ductile fracture increased i  with Increasing stress intensity, above K^= 55MPa.m, and was 2  found mostly at the edges of the specimen, Fig. 36. At active potentials, E < -0.325V(S.C.E.) precipitates were observed on the fracture surface, Fig. 37t and X-ray energy analysis of these, conducted in the S.E.M., showed them to contain magnesium. Wilde [493 identified precipitates formed under similar conditions as Mg(OH)g. 3.5.1.3 Effect of Temperature Varying the temperature had l i t t l e effect on the appearance of the fracture surface.  Examples of fractographs  from specimens of Z$% and 50$ cold rolled 316 stainless steel, cracked at similar stress intensities (K^ «• 25MPa.m*) at both 2  154°C. and ll6°G., at open circuit potentials are shown in Figs. 38 and 39. with  This implies that the variation of fractography  Is similar at different temperatures, but the variation  with E must be different at different temperatures, since changing the temperature has been shown to alter the corrosion potential, see Fig. 24.  -69-  FIGURE 35  Example of fractograph of 50$ cold r o l l e d 316 s t a i n l e s s s t e e l showing small area of d u c t i l e f a i l u r e . K  = 70MPa.m . 2  x  400x .  -70-  LJ 12  CC  h- 10 O <  (A)  8  L_  UJ _J  rO 3 O 2  $5  0125 cm  FIGURE 36  Distribution of ductile fracture in a specimen of 50$ cold rolled Jl6 stainless steel exposed to boiling MgCl solution at 154°C. (a) As a function of position across the specimen. 2  (b)  As a function of stress intensity.  -71-  (a)  00  FIGURE 37  Examples of precipitates formed on the fracture surface during cathodic polarization. (a)  E = -0.325V (S.G.E.) 800x.  (b)  E - -0.35V (S.G.E.) 400x.  -72-  FIGURE 38  Effect of temperature on the fractography of 50$ cold rolled 316 stainless steel.  lOOOx .  (a)  T = 154°C. K, - 27MPa.m^  (b)  T - 116°C., K.. - 27MPa.m  f  2  -73-  (a)  OO  FIGURE 39  Effect of temperature on the fractography of 25$ cold, rolled 316 stainless steel. 400x. (a)  T - 154°G. K - 27MPa.m^  (b)  T - ll6°G. K, - 27MPa.m^  f  f  x  -74-  3.5.2  TRIP S t e e l  3.5.2.1  Comparison with the Fractography of the  316  The fractography of the TRIP s t e e l was s i m i l a r to that of the 316, the major difference being the greater amount of I.G. cracking i n the TRIP c f . F i g s . 39 and 43.  3.5.2.2  E f f e c t s of Cold Working and the Presence of Martensite  The c o l d worked austenite showed more I.G. cracking than d i d the annealed material a t the same stress i n t e n s i t y .  This i s  shown i n F i g . 40. The TRIP containing the martensite had a dark coloured f r a c t u r e surface covered with corrosion product i n contrast t o the l i g h t grey clean looking surface of the TRIP containing only austenite.  This made i t rather d i f f i c u l t to  obtain sharp fractographs from the p a r t i a l l y martensitic s t e e l . However, the e f f e c t of the martensite i s shown i n F i g . 41. I t can be seen that there i s more T.G. cracking i n the martensitic TRIP.  Also, although the nature of the transgranular surface i s  the same f o r both TRIP materials, i t i s d i f f e r e n t from that of the 316 i n that i t has a nodular appearance rather than a s t r i a t e d or ridged look.  To check that t h i s was not due t o excessive  corrosion of the TRIP s t e e l i n s o l u t i o n a f t e r cracking had taken place, an attempt was made to match the features present on opposite sides of the crack.  The r e s u l t i s shown i n F i g . 42.  I t can be seen that the nodules on one side correspond to p i t s on the other, thus revealing that the nodular features r e s u l t from the cracking and not from subsequent corrosion.  -75-  FIGURE kO  Effect of cold work on the fractography of the TRIP steel.  400x.  (a)  Annealed TRIP, K j - lOMPa.m , T - 154°G.  (b)  25$ cold rolled (at 200°G.), K^- 9MPa.m, T - 154°C.  2  2  FIGURE kl  Effect of martensite on the fractography of 25$ cold rolled TRIP steel. lOOOx. (a) Fully austenitic, K - 22MPa.m% T - ll6°C. (D)  Containing martensite, K - 22MPa.m% T = ll6°G.  FIGURE 42  Matching fractographs from opposite sides of the crack in 25$ cold rolled (at 200°C.) TRIP steel.  800x.  -78-  3.5.2.3  Effect of Stress Intensity  The amount of I.G. cracking i n the f u l l y austenitic TRIP increased with increasing stress intensity, as was found for the 316.  This i s shown in Fig. 43. Also, at lower stress  intensities, the T.G. surface shows more ridges and striations and resembles that found in 316, cf. Fig. 34.  3.5.2.4  Effect of Temperature  Fig. 44 shows examples of fractographs of the austenitic TRIP cracked at  154°C. and 119°C. at similar stress intensities.  The specimen cracked at  119°G. shows more of the nodular type of 1  0  fracture than does the one at 154 G. 3.5.3  310 Stainless Steel  3.5.3.I General Description Cracking i n 310 was solely transgranular under a l l the conditions investigated.  See Figs. 45 and 46. The surface has  a more angular appearance than that of the 316,  especially i n  the case of the annealed material, Fig. 45(a).  The stress  intensity (approximate because of branching) does not appear to alter the fractography, cf. Fig. 45(b) and 45(c).  Also, neither  lowering the temperature nor changing the electrochemical potential from the open circuit value of - 0.27V(S.C.E.) to -0.325V(S.C.E.) had any observable effect. and 46(b) respectively.  See Figs. 46(a)  -79-  FIGURE 43  Effect of stress intensity on the fractography of 25$ cold rolled (at 200°G.) TRIP steel at open circuit potentials.  lOOOx.  (a)  No applied load, T = 154°G.  (b)  ^  (c)  K  - 8.5MPa.m , T = 154°G. 2  - 40MPa.m, T - 154°C. 2  x  -80-  FIGURE kk  Effect of temperature on the fractography of the 25$ cold rolled (at 200°C.) TRIP steel.  lOOOx.  (a)  T = 154°C, K  (D)  T - 119°0., K, - 6.6MPa.m  - 8.2MPa.m  2  x  2  -81-  FIGURE 45  Effects of cold work and stress intensity on the fractography of 310 stainless steel. 400x. (a) Annealed, K - 20MPa.m, T - 154°C. 2  x  (h) 50$ cold rolled, ^ - 25MPa.m2, T - 154°C. (c) 50$ cold rolled, K., - 50MPa.m% T - 154°C.  -82-  FIGURE k6  E f f e c t s o f temperature and a p p l i e d p o t e n t i a l on t h e f r a c t o g r a p h y o f 50$ c o l d r o l l e d 310 s t a i n l e s s s t e e l .  (a) (b)  -0.27V  T -  120°G. K-- 25MPa.m, E -  T -  154°G. K - 25MPa.m , E = -0.325V (S.G.E.)  2  t  2  f  1  (S.C.E.)  400x.  -6> 4. 4.1  DISCUSSION  Crack Branching The observations made on crack branching can be summarized  as followst (i) (ii)  There was more branching i n the low y i e l d strength a l l o y s , Transgranular cracks branched more than intergranular  cracks. (iii)  Intergranular cracks branched only when propagating normal  to the r o l l i n g d i r e c t i o n . The lower y i e l d strength (O  ) of the annealed a l l o y s w i l l y r e s u l t i n a larger p l a s t i c zone ( r ) a t the crack t i p since [42], y  Consequently,  I f crack propagation i s dependent on p l a s t i c s t r a i n i n g  at the crack t i p (e.g., to rupture a passive film) then, i n the low y i e l d strength a l l o y s , t h i s w i l l occur over a larger area and i n crease the l i k e l i h o o d of forming more than one crack.  The ten-  dency t o branch, however, must be dependent on more factors than just the size of the p l a s t i c zone, since no branching was observed i n the 25$ cold r o l l e d 316 a t a stress i n t e n s i t y of 25MPa.in \ 2  (At  I  25MPa.m the anticipated p l a s t i c zone s i z e w i l l be the same as that 7  at which branching occurred i n the annealed condition i . e . , a t " lOMPa.m ). 2  Other factors could be the radius of the crack  t i p , the d i s l o c a t i o n density and the nature of the a c t i v e s l i p systems a t the crack t i p and how these might a f f e c t the s l i p step height.  The s l i p step height i s important since i t must be  -84a t least as large as the passive f i l m thickness i n order to expose the underlying metal. The downward cracking, shown i n F i g . 8(a), indicates a strong preference f o r the I.G. cracks to follow the r o l l i n g d i r e c t i o n , i n agreement with the observation of Hyatt and Speidel [5]  that I.G. cracking followed the grain flow i n c o l d  extruded aluminum a l l o y s .  The elongation of the grains i n the  c o l d r o l l e d a l l o y s w i l l r e s u l t i n an almost s t r a i g h t i n t e r granular path i n the r o l l i n g d i r e c t i o n .  In the perpendicular  d i r e c t i o n , an I.G. crack would have to wind back and f o r t h between the grains, making cracking more d i f f i c u l t and providing many opportunities f o r the crack to branch. Crack branching i n high strength aluminum a l l o y s has been examined by Speidel [50]•  His observations led him to  the conclusion that branching only occurred i n a l l o y s having a r e l a t i v e l y i s o t r o p i c mlcrostructure and only i f the stress i n t e n s i t y was greater than 1.4 Kp, where Kp was the at which Region I I began.  value  The present r e s u l t s are i n agreement  with Speidel's conditions f o r branching, except f o r the annealed a l l o y s , which branched at stress i n t e n s i t i e s as low as 12MPa.m  2  and were c e r t a i n l y cracking i n Region I. In explanation of h i s observations, Speidel proposed that two cracks could form and continue to grow, only i f there was s u f f i c i e n t energy a v a i l a b l e to maintain each crack with a Region II v e l o c i t y .  I f less energy was a v a i l a b l e , one crack would have  a slower v e l o c i t y than the other and would be q u i c k l y s t i f l e d  -85as t h e r e g i o n o f maximum s t r e s s moved away from i t .  Thus, t h e  c r i t i c a l energy r e l e a s e r a t e f o r c r a c k b r a n c h i n g would be 2Gp,  where Gp i s t h e G v a l u e c o r r e s p o n d i n g t o Kp.  i n a c r i t i c a l s t r e s s i n t e n s i t y of 2.Kp.  This results  (See Eqn. l ) .  The above argument i s p r o b a b l y v a l i d f o r a wedge loaded D.C.B. specimen, (which i s what S p e i d e l used) s i n c e K^  decreases  as t h e c r a c k grows, but i t would not a p p l y t o t h e p r e s e n t ment because K^ i n c r e a s e s .  arrange-  I n g e n e r a l , i f a c r a c k branches  s t r e s s i n t e n s i t y o f l e s s than 1.4  Kp,  at a  then, although the longer  c r a c k w i l l grow more r a p i d l y , t h e s h o r t e r one w i l l c o n t i n u e t o grow  p r o v i d e d i t s s t r e s s I n t e n s i t y remains above K ^ ^ .  t h e p r e s e n t s t u d y , w h i l e the  Regarding  value f o r the s h o r t e r crack  will  c o n t i n u e t o d e c r e a s e as t h e l o n g e r c r a c k grows f u r t h e r ahead o f it,  t h e i n c r e a s e i n K^ w i t h c r a c k l e n g t h and the c o m p l i a n t s p r i n g  i n s e r i e s w i t h t h e specimen w i l l both h e l p t o d e l a y the i n the s t r e s s i n t e n s i t y of the s h o r t e r crack. r e a s o n t h a t c r a c k b r a n c h i n g was  observed  work, but not i n t h a t o f S p e i d e l , may  decrease  I n essence,  i n Region  the  I i n the present  be due t o the n a t u r e o f the  specimen geometry and the l o a d i n g c o n d i t i o n s r a t h e r t h a n t o a d i f f e r e n c e i n t h e S.C.C. p r o p e r t i e s of t h e a l l o y s .  4.2  Kinetics 4.2.1  Region  I  The c r a c k i n g r a t e i n Region  I , f o r t h e 2%  cold rolled  316  s t a i n l e s s s t e e l , i n c r e a s e d r a p i d l y w i t h s t r e s s I n t e n s i t y , the g r a d i e n t d log^v/dK^^ b e i n g e q u a l t o a p p r o x i m a t e l y (See F i g . 15).  0.5(MPa.m )~ . 2  T h i s i s c l o s e t o the v a l u e s o b t a i n e d f o r h i g h  -86-  strength aluminum alloysj 0.5 to 1.5(MPa.m )~ [5], 2  titanium alloys; < 1.0(MPa.m )~ [7]. 2  and for  There have been com-  paratively few attempts to explain this  dependence. H i l l i g and  Charles [51] proposed a model for Region I S.C.C. based on stress activated dissolution.  Their treatment was modified by Wiederhorn  and Bolz [52] who obtained the relationship V oc exp [(-2V K-jA-n/?) )/^] 2  where V  i s an activation volume  and /O i s the crack t i p radius While this has the observed dependence of velocity on stress, the model does not explain the significance of V or the mechanism by which the stress intensity increases the dissolution rate. The cracking  which was observed i n the unloaded specimens  was probably caused by the corrosion product in the crack preventing the specimen from completely relaxing when the load was removed [ l ] .  Residual stresses from the cold rolling may also  have been responsible. 4.2.2 Region II The Region II velocities of the Jl6 and TRIP steels were similar, as were the apparent activation energies (see Figs. 14 and 19).  The crack velocity of the 310» on the other hand, was  considerably slower although the apparent activation energy was similar, (see Fig. 21). The extent of Region II was dependent on the alloy and the processing treatment.  The start of Region II varied con-  siderably from specimen to specimen, (see Fig. 15), the average A stress intensity being approximately 20MPa.m for both the 316 2  -87and TRIP steels.  The stress intensity corresponding to the end  of Region II was related to the onset of ductile fracture. A comparison of Figs. 13 and 36 shows that for 50% cold rolled 316 ductile fracture was observed on the fracture surface at 65MPa.m% but that the transition  stress intensities above  1.  from Region II to Region III started at a  value of 50MPa.m . 2  1.  It i s possible that some tearing occurred below 65MPa.m, but 2  that i t was too localized to be observed.  While the higher  yield strength of the 50% cold rolled 316 extended Region II to higher values of K^, the lower ductility of the TRIP containing martensite caused i t to f a i l completely at a stress intensity of 1.  l i t t l e more than 50MPa.m, (see Figs. 12, 13 and 18). 2  The reason for the decrease in velocity which occurred in the 310 after the crack had grown by approximately 1-| mm. i s not known. The observed crack branching may have been responsible, either by lowering  into Region I at the instant of branching,  or else, by causing the crack to follow a crooked path which would reduce the average velocity.  However, i t was not possible  to say with certainty whether either of these effects were taking place.  The stress intensity did Increase significantly during  crack growth i n the 310» as was evident from the appearance of the plastic zone at the crack tip.  Even though  branching may  have had some influence on the stress Intensity and crack velocity, i t i s thought  that the constant velocities measured were the  result of a stress intensity independent process. 4.2.3  Cracking Mechanisms  4.2.3.1  General Considerations  Any mechanism of cracking capable of accounting for  -88the Region I and Region II results must satisfy the following conditions s (i)  The crack velocity should increase rapidly with stress  intensity up to approximately  = 20MPa.m and then be i n 2  dependent of stress intensity u n t i l the onset of ductile fracture. (ii)  The crack velocity should be independent of applied  potential between -0.25V (S.C.E.) and -0.30V (S.C.E.) in Region I and between -0.25V (S.C.E.) and -O.325V (S.C.E.) i n Region II. At potentials more active than these cracking should stop, (iii)  The apparent activation energy for Region II cracking  should l i e between 15 and 18 kcal/g.mole. Furthermore, i f the mechanism i s the same for a l l the materials studied i t must be compatible with the followlngj 50$ cold rolled 316 cracked faster ln Region II than  (iv)  2%  cold rolled 316. (v)  The TRIP steel containing martensite cracked faster in  Region II than the f u l l y austenitic TRIP. (vi) (vii)  The crack velocity was independent of the crack path i n 316. Region II cracking was slower by more than an order of  magnitude i n the 310 than in either the 316 or TRIP steels. These results differ l n several ways from those found in other systems.  Most of the published work on the temperature  dependence of Region II, gives apparent activation energies of between 3 and 5 kcal/g.mole, e.g., glasses [ 4 ] , [7] and high strength aluminum alloys [ 5 ] . [5],  titanium alloys  Hyatt and Speidel  however, obtained an activation energy of 20 kcal/g.mole  -89-  f o r Region I I cracking i n a 7039-T61 a l l o y and, i n a s i m i l a r a l l o y under c e r t a i n conditions, they observed two stress i n t e n s i t y independent regions, F i g . 47.  The values of the apparent a c t i -  vation energies, as measured from F i g . 47, are 17.3 and 2.6  kcal/g.mole  kcal/g.mole f o r the slower and f a s t e r plateaus respectively. Furthermore, the independence of crack v e l o c i t y on applied  p o t e n t i a l was not found f o r e i t h e r aluminum or titanium a l l o y s , both of which cracked f a s t e r at more noble potentials [5, 7] although i n these a l l o y s the applied p o t e n t i a l was varied by more than 1 volt.  The range of p o t e n t i a l over which cracking can occur,  l n the present system, i s l i m i t e d to less than lOOmV [53!•  At  potentials noble with respect to the corrosion p o t e n t i a l the steels p i t r a p i d l y while a t a c t i v e potentials ( i . e . , w.r.t. CORR^  E  -ck±ng  CTa  stops altogether.  Time to f a i l u r e t e s t s f o r 304  s t a i n l e s s s t e e l i n MgClg solution show a change of less than a f a c t o r of two i n f a i l u r e times over the a v a i l a b l e range of p o t e n t i a l [54]].  This i s consistent with the present r e s u l t s  f o r Regions I and I I i f i t i s assumed that the i n i t i a t i o n time i s p o t e n t i a l dependent. Before discussing stress independent mechanisms which might explain the Region I I r e s u l t s , the p o s s i b i l i t y that the actual stress at the crack t i p does not increase i n Region I I should be considered.  I n g l i s [55]] showed that the stress a t the  4 crack t i p ( o^j_p) was equal to 2o( a / / 3  where o i s the applied  stress and /O the radius of the crack t i p .  4  •= o(fTa)r (see page 3)t  Also, since  the crack t i p stress can be expressed  -90STRESS INTENSITY (kg - m m ~  10"  10  20  30  I  I  ,  40 ,  50 !  60 ,  3 / 2  j  70 _  80  90  100 10"  r  " A L L O Y 7039-T64 C R A C K O R I E N T A T I O N : T L (SHORT T R A N S V E R S E ) 5 M AQUEOUS Kl P O T E N T I A L : - 7 0 0 m V V S SCE ©  10  70° C  10  o  -2  10"  3  >-  H O O  o o _J ai >  ~  1  tr  10  10"  _l  LU >  o  o  < cc  < cc  o  o o 2:  CO  10 -5  -2  O 10  CO  O  CC  cc  cc  o o  CO CO Ui  CO CO LU CC  o u  cc co  10 -6  10 -3  .—7 IU  10  10  15  20  25  30  STRESS INTENSITY (MPa.ra ) 2  FIGURE 47  V - ^ plot for 7039-T64 aluminum alloy i n 5M aqueous K  1  solution showing two stress intensity independent  regions and the effect of temperature (after Hyatt and Speidel [ 5 ] ) .  -91as  °ti  P  Thus, If  • V<  v>>*  is increased, the stress at the crack tip  may not change i f the radius of the crack tip also increases i.e., i f the crack is blunted. If Region II is due to crack blunting, then i t might be expected that i f the yield strength was Increased then the stress intensity at which Region II begins would be greater, resulting in a greater velocity which is in agreement with condition (iv) above. 4.2.3.2  Dissolution  Electrochemical dissolution alone does not explain the cracking kinetics very well.  If crack growth occurs by dissolution  then i t might be expected that anodic polarization would increase the crack velocity. Over the range of potentials in which cracking occurs in MgClg solution, the anodic current density, as measured on a static surface, changes by more than two orders of magnitude [19]»  The dissolution rate in Region II could be independent  of potential i f the anodic reaction was under diffusion control, but then the activation energy would have to be closer to that for mass transport in the electrolyte, i.e., 3 - 5  kcal/g.mole [56].  Nevertheless, an activation energy of between 15 and 18 kcal/g.mole is in the observed range for dissolution e.g., Staehle reported values between 12 and 18 kcal/g.mole for the dissolution of nickel in lNHgSO^ on static and straining electrodes [19]. The faster velocities in Region II, caused by the cold work in the 316 and the martensite in the TRIP could be due to  -92an increase i n the exchange current density resulting from an increase i n the number of active sites for dissolution such as at dislocations [ 5 7 3 . Nickel, being more noble than iron or chromium, could exert i t s effect on the cracking rate by raising the reversible electrode potential E . Q  This could i n turn raise  the corrosion potential, (as was observed) and lead to a reduction in the corrosion current, as described by Nikiforuk [293.  While  this explanation w i l l result i n a slower crack velocity for the 310, i t w i l l also require that the velocity be dependent on the applied potential, which was not observed, at least i n the case of the 316. If dissolution alternates with a period of film rupture then an electrochemical process can explain the presence of a c r i t i c a l potential for S.G.G. At a given stress intensity sufficient dissolution must take place at the crack t i p to allow build up of enough strain to break the passive film [583. As the potential i s made more active w.r.t. QQRR» the dissolution e  rate w i l l decrease and the repassivation rate w i l l increase, thereby reducing the total amount of dissolution possible and preventing film rupture. strain would result  At higher stress intensities more  from a given amount of dissolution, thus  requiring a more negative potential to prevent cracking, as was observed. The slowly decreasing velocity observed on changing the potential to -0.35V (S.C.E.), (see Fig. 16) could be related to a change of pH i n the crack.  The increase In the rate of proton  reduction caused by cathodic polarization would raise the pH i n  -93the crack.  This higher pH, together with the more negative  potential, would reduce the dissolution rate, increase the repassivation rate and hence reduce the crack velocity. A reduction in overall dissolution would decrease the supply of hydrogen ions from hydrolysis of the metal ions, causing a further increase i n pH and hence a further reduction i n crack velocity. 4.2.3.3  Adsorption  The cracking kinetics can he explained qualitatively by a process in which an adsorbed species weakens the already stressed metal bonds at the crack t i p [59].  In Region II the  crack velocity could be controlled by the rate of adsorption of the damaging species. Activation energies for chemisorption can cover a wide range of values, but are frequently of the order of 20 kcal/g.mole [j>6]].  In Region I the cracking rate  could be controlled by a deformation process.  Deformation at  the crack t i p could increase the number of adsorption sites, either by breaking a passive film, or else by increasing the number of surface imperfections. An adsorption process i s also consistent with the observed independence of crack velocity on applied potential.  In Region  I a process controlled by deformation would not be affected by potential, provided there were sufficient species available to adsorb.  While chemisorption i s often restricted to a limited  range of potential, It is possible that over the limited range studied there was always more than sufficient sites available having suitable potentials.  -94The critical potential for S.C.C. in 18-8 stainless steel exposed to MgClg solution has been explained by Uhlig and Cook [60] as the potential below which Insufficient chloride ions can adsorb on surface imperfections at the crack tip. Raising could lower the critical potential either by increasing the number of adsorption sites or else, by making these sites more active. The faster cracking rates in Region II caused by the cold work in the Jl6 and the martensite in the TRIP could be due to an Increase in the number of adsorption sites resulting from the increased dislocation density In the Jl6 or the lattice strains produced by the martensite transformation in the TRIP. The slower cracking rate observed in the J10 could be caused by a greater activation energy for chemisorption due to its higher nickel content. The activation energy for chemisorption depends on the heat of adsorption as shown In Fig. 48. in the magnitude of the heat of adsorption AH  A decrease  will result in ct  curve II shifting towards higher potential energies, making Q greater. Heats of adsorption for chloride ions on stainless steels are not available, but a qualitative estimate of the effect of nickel can be obtained from the heats of formation of NiClg, FeClg and CrCl which are - 6 5 . 1 , - 7 2 . 2 and -85.2 kcal respectively [ 6 l ] ,  2  The  lower affinity of nickel for chloride ions will likely result in a lower heat of adsorption and consequently a greater activation energy. (See Fig. 48). The observed activation energy for the 310 of 18.1 kcal/g.mole is greater than that for the other alloys (15.1 to 16.0 kcal/g.mole). The ratio of the velocities predicted by an Arrhenius rate law,  -95-  DISTANCE OF ADSORBED UNIT FROM THE S U R F A C E  FIGURE 48  Relationship between the heat of adsorption and a c t i v a t i o n energy f o r chemlsorption E a i d l e r and Eyring [56]  the  ( a f t e r Glasstone,  ).  I i s the potential-energy distance curve f o r a c h l o r i d e ion  in solution.  I I are p o t e n t i a l energy distances  curves f o r a chemisorbed chloride ion.  A smaller  heat of adsorption A Ha (b) w i l l l i k e l y r e s u l t i n a greater a c t i v a t i o n energy f o r chemisorption Q (b).  -96-  V = V  q  exp (-Q/RT), for values of Q of 15.5 and 18.1 kcal/g.mole  is given by:  The above calculation assumes that the pre-exponential term V  q  i s the same in both cases.  Comparing Tables III, IV and V  shows that at 154°C. the 310 cracked approximately 35 times slower than the other alloys, which considering the errors i n the apparent activation energies Is i n reasonable agreement with the above explanation. 4.2.3.4  Absorption of Hydrogen  Diffusion of hydrogen i n austenite could be a rate controlling process for Region II. Austin and Elleman [ 6 2 ] obtained a value of 0 . 6 l eV/atom (i.e., 14.2 kcal/g.mole) for the bulk diffusion of tritium i n 304 and 316 stainless steels over the temperature range of 25°C. to 222°C.  (The result was  obtained by injecting tritium into the sample via the reaction ^ L i ( n,0t) ^H, diffusion annealing and then measuring the tritium concentration profiles, thus avoiding the surface trapping effects found in methods involving permeation rates through membranes [ 6 3 ] ) . The effect of martensite in the TRIP steel could be to shorten the length of the slow diffusion path through the austenite, diffusion in the b.c.c. a'martensite being faster than i n the f.c.c. austenite [ 6 3 ] . The Region II results dp not support a process i n which the rate of transformation to martensite at the crack t i p i s important, because i f this were the case the 25$ cold rolled  -97austenltic TRIP might have been expected to crack faster than the 2%  cold rolled 316.  In fact, the 316 cracked slightly faster  although the difference may not have been significant.  (See  Tables III and IV). 4.3  Corrosion Potentials The effect of stress on the corrosion potential E  been observed by Barnartt and Van Rooyen [64] for 18-8 steel in MgClg solution.  has  C Q R R  stainless  They found that QQpp became less E  noble after the crack had initiated (on stressed wires) and that the potential continued to decrease until failure occurred.  The  total potential change was -40mV. They explained their observations by a shift In the anodic polarization curve towards greater currents. Such an explanation i s valid for the present observations also. Fig. 49 shows a schematic Evan's diagram Illustrating the effect of stress on the corrosion potential.  Curve 1 i s  the anodic polarization curve representing a film covered surface remote from the crack t i p . Curves 2 (a), (b) and (c) represent the anodic dissolution at the film free surface at the crack tip at different stress intensities. Curves 3 (a), (b) and  (c)  are the composite anodic polarization curves for different values of K^.  If i t i s assumed that the cathodic reaction (curve 4) i s  not significantly affected by the stress intensity then the corrosion potential w i l l become less noble on increasing the stress intensity. The increase in the amount of dissolution at the crack tip with K., can be explained in several ways:  -98-  LOG(CURRENT)  FIGURE 49  Schematic Evans diagram illustrating the effect of stress intensity on the corrosion potential. (See text for explanation).  -99-  (i)  The exchange current density may Increase due to the  existence of more active sites at higher stresses. (ii)  As the crack velocity increases the rate of formation of  unfilmed surface w i l l Increase and raise the overall dissolution rate. (iii)  I f the crack grows by a film rupture/dissolution process  then at higher stresses the larger plastic zone w i l l result i n a larger area of exposed metal and hence more dissolution, i.e., a wider crack. (iv)  At higher stress intensities the effective surface area for  dissolution i s increased due to the greater roughness of the surface.  (See Fig. 7(°))»  Of these four possible effects, the f i r s t three would result In a rapid change of E  nrtt3  _, with K.., whereas the fourth  would require time for the topography of the fracture surface to change. Thus, the f i r s t three could explain the almost instantaneous change ofQQRR e  o  n  increasing  but only (iv) could  explain the slow return to equilibrium on reducing K^. 4.4 Fractography 4.4.1 General The fractography of the Jl6 indicated that the mode of cracking was subject to the following conditions« (i)  I.G. cracking was more probable in the cold worked condition  than i n the annealed alloy. (ii)  The amount of T.G. cracking increased as the applied  potential was made more noble.  -100(ili)  The amount of I.G. cracking increased as the stress  intensity was raised. (iv)  There was more I.G. cracking in the centre of the crack  than at the edges. Nielsen []l"j showed a photograph of a U-bend specimen of 304 stainless steel cracked in 42$ MgClg solution in which the fraction of I.G. cracking was greatest in the last part of the specimen to fail.  Nielsen did not relate the change in crack path to stress,  but instead to the fact that the part which had cracked last was i n i t i a l l y in compression and this, together with diffusion of hydrogen into the lattice, caused decohesion of the grain boundaries.  Okada and co-workers [65]  observed a change from  T.G. to I.G. cracking between the f i r s t and last parts to crack of a tensile specimen of 304 stainless steel under constant load in MgCl.2 solution and related this to an increase in stress. These workers also observed a change from I.G. to T.G. cracking on anodic polarization of JlG stainless steel. It i s interesting to note that in the case of both aluminum and titanium alloys the cracking mode changes from I.G. to T.G. as the stress intensity i s increased [_5~], 4.4.2  Effect of Applied Potential  The effect of applied potential on the cracking mode i s compatible with either a dissolution/film rupture process or with an adsorption mechanism. Staehle [vf\  has reported that  S.C.C. of Fe-Cr-Ni alloys in concentrated MgClg solutions occurs at potentials close to the transpassive knee of the anodic polarization curve, i.e., close to the region where localized film  -101-  breakdown occurs.  Also Payer and Staehle [66] reported that under  suitable conditions corrosion pits initiate preferentially at grain boundaries, indicating the less passive nature of the film there. Thus, i n terms of a film rupture process, at active applied potentials the grain boundaries may be the only sites where sufficient dissolution can take place between film rupture events to sustain cracking. At more noble applied potentials no sites on the oxide film w i l l be particularly passive and so the effect of the grain boundaries w i l l be less important, the crack path being determined by a film rupture process such as s l i p . In terms of adsorption the crack path would be dependent on the potential of the adsorption sites.  Due to the lattice  disorder at the grain boundaries,adsorption sites there would be at more active potentials than sites distant from the grain boundaries.  As the potential i s made more noble than the c r i t i c a l  potential for cracking, the f i r s t sites for adsorption w i l l l i e on the grain boundaries, resulting i n I.G. cracking. A further increase i n applied potential w i l l allow adsorption to take place on less active sites away from the grain boundaries, resulting i n T.G. cracking as well. 4.4.3  Effect of Stress Intensity  The effect of stress intensity on the crack path may have been due to the potential at the crack t i p becoming more active as the stress intensity was Increased, even applied potential was constant.  though the  This could occur as a result  of an IH drop along the crack, the value of which would increase  -102as the crack v e l o c i t y Increased.  (See explanation below).  If  the crack path i s determined only by the p o t e n t i a l , then from F i g . 28 I t can be seen that at -0.25V ( S . C . E . ) a p o t e n t i a l change of a t least 50mV. would be required to account f o r the change i n cracking mode caused by an increase i n  4  from lOMPa.nr to  4 30MPa.m .  Beck estimated that the IR drop along the stress  corrosion cracks l n titanium a l l o y s was greater than 150mV,  [[6?].  Assuming a p a r a l l e l sided crack, the IR drop would be given as a f i r s t approximation byj • i.l/c  AE  (5)  where A E Is the IR drop along the crack 1 i s the average distance between anodic  and  cathodic s i t e s . I i s the current density at the crack t i p c i s the e l e o t r i c a l conductivity of the s o l u t i o n . Also i f i t i s assumed that the crack propagates s o l e l y by d i s s o l u t i o n then the anodic current density ( i n view of conservation of mass considerations) i s r e l a t e d to the crack v e l o c i t y by the following expressions 1 where p  V/O  zF/A  (6)  •» density of the a l l o y  z • average valence of the metal ions A and  m  average atomic weight  F « Faraday  This does not r u l e out adsorption since weakening of the metal bonds by adsorption could r e s u l t i n e i t h e r bond rupture or dissolution.  -103For V - 4 x 10~ m/s,/3 - 8.0 g/cc, z - 2 and A • 56, i «= 1.2 amps/cm . 7  2  In order to maintain electroneutrallty at a l l points in the electrolyte there w i l l be a similar current density of anions towards the crack t i p .  For a total current density of 2.4 amps/cm ,  AE - 50mV and a conductivity of 2.0 ohnf cnf 1  1  [68], Eqn. (5)  gives a value for 1 of 0.4mm. This i s not an unreasonable distance i n view of the large differences between the anodic and cathodic current densities i.e., the cathode must be considerably larger than the anode.  An IR drop effect can also explain the change  in crack path across the width of the crack since the potential drop w i l l be less at the edges of the specimen. The above explanation i s also valid i f the crack grows by a process i n which adsorption results in bond rupture, but in that case the current density might be less since only sufficient current to passivate the sides of the crack would be required. It i s also possible that the change in crack path across the width of the specimen could be due to a change in stress state.  It i s not clear, however, why the plane strain conditions  at the centre of the specimen should favour intergranular cracking. Patches of dimple fracture on stress corrosion fracture surfaces, formed at high stresses, have been observed before [69]. That these occur predominantly at the edges of the specimen, i n the present case (see Fig. 36), i s probably due to the greater shear stress in the plane stress regions at the edges of the specimen. 4.4.4 Effect of Composition The different cracking modes of the three alloys could be due to their different alloy compositions or else to their  -104different corrosion potentials. Specifically, Mo additions have been shown to increase the amount of I.G. cracking in Fe-l6Cr-15Ni alloys [65].  Also, Mo improves the pitting resistance of austenitic  stainless steels [70].  A recent paper by Suglmoto and Sawada £?l]  Indicates that addition of molybdate ions to the corrosive environment can also inhibit pitting.  On the basis of surface  analysis, these authors conclude that i t i s the adsorption of 2MoO^  ions from solution which prevents pitting.  With regard to  adsorption controlled S.C.C., competitive adsorption of MoO^  ions  could restrict adsorption of the damaging species to the more active adsorption sites at the grain boundaries. It i s also possible that the more active corrosion potentials of the TRIP and yi6 alloys resulted in the greater amounts of I.G. cracking in these alloys since lowering the applied potential favoured I.G. cracking in Jl6. 4.4.5  Effect of Martensite  Nodular features similar to those on the fracture surface of the TRIP steel, especially when i t contained martensite, have not been reported before.  In view of the fact that they were  more prominent at lower temperatures and at higher stress Intensities (see Figs. 43 and 44), I.e., conditions which favour the formation of stress induced martensite, i t Is likely that they are associated with martensite.  This implies that the  martensite i s formed in the f u l l y austenitic TRIP, before or at the same time as the new fracture surface, since the nodular features were found to match on opposite sides of the crack. (See Fig. 42).  -105The fractography associated with a martensite in the case of 304L stainless steel exposed to MgCl^ solution [35] is different from that observed for the TRIP steel. This difference could be related to the higher carbon content of the TRIP steel (0.22 wt. % G as opposed to < 0.03 wt. % G for the 304L). It is possible that precipitation of carbides from the a phase in 7  the TRIP steel at 154°C. could result in the nodular features because the carbon content exceeds the equilibrium solubility in the b.c.c. a phase. 4.5  Assessment of S.C.C. Models The results have been discussed in terms of three processes  thought to be important ln the S.C.C. of austenitic stainless steels, namely, electrochemical dissolution, absorption of hydrogen and adsorption of a damaging species. A film rupture model for S.C.C. can account for most of the present observations with the exception of the effects of applied potential and temperature on the Region II cracking. While a model involving hydrogen absorption can explain some of the kinetic results i t is difficult to see how i t can account for the effect of applied potential on the fractography. A model of cracking involving adsorption is consistent with a l l the present observations although there is not enough known about the details of chemisorption to thoroughly test the model, e.g.» activation energies for adsorption of specific species on the specific alloys. Foley [72] has pointed out that the concentrated solutions encountered in localized corrosion cells will result in the anions  -106and cations being very close together making adsorption quite probable.  In 45$ wt. $ magnesium chloride solution of pH2 there  are approximately 11 moles/litre of chloride ions, 10~  moles/  —12 l i t r e of hydrogen Ions and 10 ~  moles/litre of hydroxyl ions.  Thus, the adsorbing species Is very likely to be a chloride ion or perhaps a complex metal chloride ion, although the possibility that i t could be a hydrogen ion cannot be excluded. A model for cracking involving adsorption does not rule out the possibility that other processes such as dissolution, deformation or film rupture may be important.  As mentioned  previously plastic deformation or film rupture might be necessary to produce sufficient adsorption sites.  Also i t i s possible that  adsorption could result i n either bond rupture or dissolution. With reference to Fig. 50 adsorption on atom A could weaken bonds AB, AC and AD by causing a redistribution of the electrons surrounding atom A.  This could then result in either the rupture  of bonds AB or AC or else i n the dissolution of any of atoms A, B or C.  Thus, the crack could propagate by adsorption induced  bond rupture or adsorption assisted dissolution.  -107-  O  METAL  X  ADSORBED  FIGURE 50  ATOM SPECIES  Schematic representation of the effect of adsorption at the crack t i p .  (See text for explanation).  -1085. CONCLUSION 5.1  Conclusions The present studies on stress corrosion cracking of  316 and 310 stainless steels, and a TRIP steel in hot aqueous magnesium chloride solutions have led to the following conclusions! (i) V-K^plots of these ductile alloys are best obtained in the cold worked condition with a variation of the double cantilever beam specimen whose behaviour Is similar to a 3 point bend specimen. (li)  The alloys exhibit both stress intensity dependent cracking  (Region I) and stress intensity independent cracking (Region II). The cracking rates in Region II at 154°C. of the 316 and TRIP steels (approximately 3.5 x 10 of the  310  (iii)  Crack branching occurs in both Regions I and II, particularly  (1.01  x 10"  m/s) are considerably faster than that  8  m/s).  in the annealed condition and when the cracking mode is transgranular. (iv) Cracking in 25$ cold rolled 316 at 154°C. does not occur below a critical potential whose value is  dependent. Above  this -value cracking is independent of potential over a range of 50mV in Region I and ?5mV in Region II.  (v)  In Region II, the crack velocities of a l l the alloys  obey an Arrhenius rate law, the apparent activation energies varying from 15.1 kcal/g.mole to 18.1 kcal/g.mole. (vi) The corrosion potential of the 25$ cold rolled 316 alloy cracking at 154°C. becomes more active as the stress intensity is increased consistent with a shift of the anodic polarization curve to higher currents.  -109(vil)  The amount of Intergranular cracking decreases with  increasing alloy content and i n 25$ cold rolled 316 i t i n creases with stress intensity and cathodic polarization and i s greater in the centre of the crack front.  The effects of  and position along the crack front may he attributed to a change in potential at the crack t i p . (viii)  Temperature does not affect the fractography of the  316 and 310 alloys, but at lower temperatures there are nodular features on the fracture surface of the TRIP steel which may be related to the presence of martensite. (ix)  While this work does not exclude a film  rupture/dissolution  mechanism the results are most readily explained by a process in which an adsorbed species modifies the behaviour of surface atoms at the crack t i p . 5'2  Suggestions for Future Work  (i)  Determine quantitatively the effects of composition, cold  work and temperature on the Region I kinetics.  This may provide  some insight Into the rate controlling processes i n Region I. ( i i ) Determine the effect of varying the solution concentration on the cracking rates in Regions I and II and on the activation energy for Region II. This could help to distinguish between dissolution and adsorption controlled cracking.  -110BIBLIOGRAPHY 1.  Nielsen, N.A.  Corrosion, Vol. 27, p. 173 (1971).  2.  Brown, W.F. and Srawley, J.E. Plane Strain Crack Toughness Testing of High Strength Materials, p. 6, A.S.T.M. S.T.P. 410 (1966).  3.  Wiederhorn, S.M.  J. Amer. Ceramic S o c , Vol. 50, p. 407 (1967).  4.  Wiederhorn, S.M.  Int. J. of Fract. Mech., Vol. 4, p. 171 (1968).  5.  Speidel, M.O. and Hyatt, M.V. Advances in Corrosion Science and Technology, Vol. 2, pp. 115-335» Plenum Press, New York (1972).  6.  Speidel, M.O.  7.  Blackburn, M.J., Feeney, J.A., and Beck, T.R. Advances i n Corrosion Science and Technology, Vol. 3, Plenum Press, New York (1973).  8.  Sandoz, G.  Met. Trans, Vol. 6A, p. 631 (1975).  S.C.C. in High Strength Steels and i n Titanium and  Aluminum Alloys, pp. 79-145, U.S. Gov. Printing Office (1972). 9.  McEvily, A. and Bond, A. J . Electrochem S o c , Vol. 112, p. 131 (1965)  10.  Speidel, M.O.  11.  Helfrich, W.J.  Corrosion, Vol. 32, p. 187 (1976). Stress Corrosion Testing, pp. 21-30, A.S.T.M.  S.T.P. 425 (1966). 12.  Hodge, J.G., and Miller, J.L. Trans. A.S.M. Vol. 28, p. 25 (1940).  13.  Osterholm, R.  14.  Van Droffelaar, H., and Hudson, G.  Paper Trade Journal, p. 26, Oct. 1975. Canadian Chemical Processing,  Vol. 50, p. 78, FEB. and p. 63 MAR. (1966). 15.  Grafen, H.  16. 17. 18.  Snowden, P.P. J.I.S.I., Vol. 197, P. 136 (1961). Harston, J.D., and Scully, J.G. Corrosion, Vol. 25, p. 493 (1969). Latanision, R.M., and Staehle, R.W. Proc of Conf. on Fund. Aspects of Stress Corrosion Cracking, pp. 214-296, N.A.C.E. (1969). Staehle, R.W. The Theory, of Stress Corrosion Cracking in Alloys, p. 223, J.C. Scully - Ed., N.A.T.0. (1971).  19.  Werkstoffe und Korrosion, Vol. 16, p. 344 (1968).  -11120.  Copson, H.R. Physical Metallurgy of Stress Corrosion Fracture, p. 247, T. Rhodin - Ed., Interscience, New York (1959).  21.  Edeleanu, C.  J. of the Iron and Steel Industry, Vol. 175, P. 390  (1953). 22.  Blrley, S. Phd. Thesis 1972, University of British Columbia.  23.  Baker, H.R., Bloom, M., Bolster, R., and Singleterry, C. Corrosion, Vol. 26, p. 420 (1970). Corrosion, Vol. 19, p. 102t (1963).  24.  Swann, P.R.  25.  Swann, P.R., and Nutting, J. J. of Inst, of Metals, Vol. 88,  26.  Swann, P.R., and Embury, J.D. High Strength Materials, Zackay - Ed., Wiley, New York (1965).  27.  Douglas, D.L., Thomas, G., and Rosser, W.  p. 477 (I960).  Corrosion, Vol. 20,  p. 15t (1964). 28.  Vermilyea, D.A. Physics Today, p. 23, SEPT 1976.  29. 30.  Nikiforuk, T.P. Masters thesis 1976, University of British Columbia. Morek, M., and Hochman, R. Corrosion, Vol. 26, p. 5 (1970).  31.  Rhodes, P.R.  32.  Barnartt, S., and Van Rooyen, D.  Corrosion, Vol. 25, p. 462 (1969). J. of the Electrochem S o c ,  Vol. 108, p. 222 (1961). 33.  Holzworth, M.L., and Louthan, M.R.  Corrosion, Vol. 24, p. 110 (1968).  34.  Vaughn, D.A., Phalen, D., Peterson, C , and Boyd, W.  Corrosion,  Vol. 21, p. 315 (1965). 35.  Birley, S.S., and Tromans, D.  Corrosion, Vol. 27, p. 63 (1971)•  36.  Holzworth, M.L.  37.  Mehta, M.L., and Burke, J. Corrosion, Vol. 31, p. 108 (1975).  38.  Zackay, V., Parker, E.., Fahr, D., and Bursch, R. Quart., Vol. 60, p. 252 (1967).  39.  McCoy, R.A., and Gerberch, W.W.  40.  Sedriks, A.J. Corrosion, Vol. 31f p. 339 (1975).  Corrosion, Vol. 25, p. 107 (1969).  A.S.M. Trans.  Met. Trans, Vol. 4, p. 539 (1973).  -11241.  Irwin, G.R., and Kies, J.A. Welding Journal Research Supplement,  Vol. 33, P. 193 (1954). 42.  Knott, J.P.  Fundamentals of Fracture Mechanics, pp. 108-125,  Butterworths, London  (1973).  43.  Irwin, G.R.  44.  Gross, B., and Srawley, J.E. Stress Intensity Factors f o r Three Point Bend Specimens by Boundary C o l l o c a t i o n , Technical note D-3092, N.A.S.A. DEC. 1965.  45.  Smith, H.R., and Piper, D.E. p. 65 of Ref. # 7.  46.  Blau, P.J.  47.  Magnani, N.J.  48.  Streicher, M.A., and Ga3ale, I.B. p. 305 of Ref. # 18.  49.  Wilde, B.E.  50.  Speidel, M.O.  51.  H i l l l g , W.B., and Charles, R.J.  J . Appl. Mech., V o l . 24, p.  Met. Trans. V o l . 7A, p. Corrosion, V o l .  463 (1976).  31# p. 337 (1975).  J . of Electrochem. Soc., V o l . 118, p.  1717 (1971).  p. 35 of Ref. # 19. High Strength Materials, p. 682,  V. Zackay - Ed., Wilely, New York 52.  361 (1957).  (1965).  Wiederhorn, S.M., and Bolz, L.H. J . of Amer. Ceramic Soc.  Vol. 53, P. 543 (1970). 53.  Brennert, S.  Recent Advances i n Stress Corrosion, Royal  Swedish Acadamy of Eng. Sciences, Stockholm  54.  Wilde, B.E., and Kim, G.D.  55.  I n g l i s , C.E.  56.  Glasstone, S., L a i d l e r , K., and Eyring, H.  Corrosion, V o l . 28, p.  Trans. Inst, of Naval A r c h i t , , V o l .  Processes, McGraw H i l l , New York 57.  West, J.M.  (1961).  55, p. 2 (1913).  Theory of Rate  (1941).  Electrodeposition and Corrosion Processes, p. 39,  D. Van Nostrand, London  (I965).  58.  Vermilyea, D.A., and Diegle, R.B. Corrosion, V o l .  59.  Uhllg, H.H.  60.  U h l l g , H.H., and Cook, E.W.  61.  350 (1972).  32,  p.  pp. 86-91 of Ref. # 18. J . Electrochem. S o c , V o l . 117,  p. 18 (1970). Latimer, W.  26 (1976).  Oxidation P o t e n t i a l s , Prentice H a l l , New York  (1952).  -11362.  Austin, J.H., and Elleman, T.S. J. of Nuclear Mat. Vol. 43,  p. 119 (1972). 63.  Jost, W. Diffusion in Solids, Liquids and Gases, pp. 305-307, Academic Press, New York (1952),  64. Barnartt, S., and Van Rooyen, D. J. of Electrochem. S o c ,  Vol. 108, p. 222 (1961). 65.  Okada, H., Hosoi, Y., and Abe, S. Corrosion, Vol. 27, p. 424 (1971).  66. Payer, J.H., and Staehle, R.W. Corrosion Fatigue - Chemistry, Mechanics and Microstructure, p. 211, N.A.C.E. (1971). 67. Beck, T.R. p. 64 of Ref. # 19. 68. Chemical Engineers Handbook, J.H. Perry - Ed., McGraw H i l l , New York (1941). 69.  Scully, J.C.  p. 127 of Ref. # 19.  70. Streicher, M.A. 71.  Corrosion, Vol. 30, p. 77 (1974).  Sugimoto, K., and Sawada, Y. Corrosion, Vol. 32, p. 347 (1976).  72. Foley, R.T. J* Electrochem. S o c , Vol. 122, p. 1493 (1975).  -114APPENDIX 1 Calculation of Errors due to the Downward Growth of the Cracks From Fig. Alt sin 3 ~ ?!  CB - a sin 3  At) CA cos 3 -£g AC •» a cos 3 OD - OB - (OC + C B ) ^ ' - [ ( R + a cos 3 ) + (a sin 3 ) 2 ] ^ 2  2  - (R + 2aR cos 3 +  2  2  a  )^  Error i n a « A B - A D « a - (OD - OA) i.e.,  2 2-4a « a + R - (R + 2aR cos 3 + a )*  For 3 constant, the fractional error in da/dt, i.e., i n V, i s the same as the fractional error i n a. The % error in V for different values of a and 3 i s shown i n Table A l . The % error i n  due to the increase i n length of L i s equal to  the % increase i n L, since A L - CB - a sin 3  Oc L, see Eqn. (4). From Fig. A l . AL/L «• a sin 3/L  Values of AL/L f o r different a and 3 values are shown i n Table A l . The error i n  due to the actual ligament length BE, being greater  than the measured value DF, has been calculated as the difference ln  due to horizontal cracks of lengths AC and AD, i.e., of  lengths a cos 3 and [R + 2aR cos 3 + a ) * - R] respectively. The % error i n K^, obtained in this way i s shown i n Table Al for different values of a and 3. The net error i n errors i s also shown i n Table A l .  due to both of the above  -115-  F-IGURE A l  E f f e c t of the downward growth o f the cracks on the measurements of V and K^. crack length of a  AB 0 0 DB  at  angle between the crack and the h o r i z o n t a l  - centre of scribed scale - arc subtended by 0  AD m distance measured as crack length BE m a c t u a l ligament length DP m measured ligament length CB  s  the e r r o r i n moment arm L  OA » distance R  TABLE A l Errors due to the Downward Growth of the Cracks  % Error i n YL^  Crack  10°  20°  N.B.  Length  % Error ln  (cm.)  Velocity  % Error i n due to L  due to  Net % Error  Ligament Error  in K  x  0.5  -1.3  -1.4  +0.1  -1.3  1.0  -1.1  -2.7  +0.47  -2.2  2.0  -0.9  -5.5  +1.5  -4.0  0.5  -5.2  -2.7  +0.52  -2.2  1.0  -4.5  -5.4  +1.81  -3.6  2.0  -3.6  -10.8  +5.7  -5.1  negative errors indicate that the measured value i s less than the actual value  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0078739/manifest

Comment

Related Items