UDC 669, "/88
THEORY
OF
HYDROGEN
EMBRITTLEMENT
OF M E T A L S *
K. V. Popov and E. P. Nechai
Fiziko-Khimicheskaya Mekhanika Materialov, Vol. 3, No. 6, pp 631-650, 1967 The present state of the theory of hydrogen embrittlement of metals is analyzed and the shortcoming of several variants of this theory are discussed. A new hypothesis of hydrogen embrittlement is presented; the hypothesis is based on concepts of dynamic strain aging and makes it possible to overcome certain difficulties of the theory of this phenomenon. The influence of hydrogen on the strength and ductility of steel has been the subject of numerous investigations whose results differ widely not only in the quantitative estimates of the effect of hydrogen on mechanical properties of metals but also in respect to qualitative conclusions reached. These discrepancies are evidently attributable to differences in the experimental conditions. Among various kinds of brittleness associated with the presence of interstitial impurities and not caused by coarse structural changes (precipitation of new brittle phases, decarburization, etc.) there are two groups which differ in the character of the manifestation of the susceptibility of metals to brittle fracture [t]. One of these groups comprizes phenomena which have symptoms of cold brittleness. Embrittlement of this kind begins to be manifested only at a certain sufficiently low temperature, persists at all the temperatures below the critical, and becomes more pronounced with increasing strain rate. The other group includes embrittlement phenomena which are not intensely manifested except at sufficiently high temperatures and sufficiently slow strain rates; at fast strain rates brittleness of this kind is not manifested. The irreversible hydrogen embtittlement [2] may manifest itself as cold brittleness, i . e . , as brittleness of the first kind. True [3] or reversible [2] hydrogen embrittlement is typical of brittleness of the second kind. The increase produced by hydrogen in the susceptibility of steel to cold brittleness was observed by many workers [4-12, 14, 18]. It is significant that hydrogen-induced cold brittleness of steel begins to be manifested only at sufficiently high hydrogen contents [16] which, depending on the composition of steel, range from 2 to 8 ml/100 g. Small quantities of hydrogen, which do not lead to the appearance of a gaseous phase with a high hydrogen pressure, may even reduce the susceptibility of steel to cold brittleness [8, 13]. It was shown [17-28] that under dynamic Ioading conditions the influence of hydrogen on mechanical properties of metals is either not observed or very weak, and that hydrogen embrittlement of steel is substantially intensified at faster strain rates [18, 28-27, 80, 66]. The probability of brittle fracture of hydrogen-charged iron increases with decreasing strain rate and rising temperature [88]. Tensile tests at slow strain rates showed that hydrogen embrittlement of steel is manifested between 200 and - 1 6 0 " C [88] and also above - 6 0 ~ C [89]. One of the symptoms of reversible hydrogen embrittlement is retarded fracture or static fatigue [16, 28, 29, 84, 87]. Hydrogen embrittlement was observed also during fatigue tests in rotating bending [31]; a hydrogen-induced reduetion in the conditional fatigue limit of steel was observed in [32, 38]. Most of the studies of hydrogen embrittlement were carried out on steels with ferritic ~or ferrito-pearlitic structures. Austenitic steels and nickel were much less investigated in this respect. Several workers [27, 86-89] observed no evidence of the influence of hydrogen on the mechanical properties of these materials. A small reduction in the ductility of an anstenitic steel as a result of hydrogen charging was reported in [40, 41], whilst a marked hydrogen embrittlement of steel of this kind was observed in [42]. We established that both austenitic steels [43] and nickel [44] are markedly embrittled by hydrogen when present in sufficiently large quantities across the entire specimen cross section. A reduction in the ductility of nickel under the influence of absorbed hydrogen was observed also by other researchers [48-50]. Very contradictory is the data on the influence of hydrogen content in steel and nickel on their mechanical properties which reflect the resistance of these metals to plastic deformation at various stages of this process, including fracture. It was established [14, 60-64] that hydrogen reduces the UTS of metals especially when extended at slow strain rates [62]. As shown by certain investigations [65-69], hydrogen absorbed by steel during electrolysis or chemical pickling has no effect on its UTS. It was observed that metallurgical hydrogen at concentrations of up to 8.6 ml/100 g does not affect the UTS of steel either [28, 70]. When steel St. 3 is charged with electrolytic hydrogen during a tensil test, the stress/strain diagram undergoes no changes (up to the stress corresponding to the UTS [69]). Other studies [71] showed *This article is published as part of a discussion of the theory of hydrogen embrittlement of metals.
459
that although the UTS of steel is practically unaffected by hydrogen charging, its true tensile strength is greatly reduced. Finally, it was observed in [86] that hydrogen charging produces a 7-fold reduction in the bending strength of steel 5KhNM plate specimens. An increase in the UTS of steel as a result of hydrogen charging is observed ----'---veryseldom [72-74]. We studied hydrogen embrittlement of steel 1KhlSN9T which 9 had absorbed up to 35 rul H2/100g as a result of a high temperature treatment in hydrogen at high press~es, or up to 30 ml Hz/100g as a result of cathodic polarization (14-800 hr) in an electrolyte [43]. We observed no changes in the UTS in either case although the ductility of steel was markedly reduced in both cases. A substantial 80 9 b increase in the UTS of this steel was observed [54] on specimens with a hydrogen content of 50 m//100g (Fig. 1). 80
q0 r - 760
~
-80
0
*8~
7~
The results of our investigations [43, 54, 55] made it possible to postulate that the main cause of contradictions in published data on the influence of hydrogen on the UTS of austenitic steel and nickel is the fact that UTS of these metals is strongly dependent on their hydrogen content and that various data was obtained under different experimental conditions. It was shown for instance [55] that hydrogen, depending on its concentration, may either increase or reduce the UTS of nickel (Fig. 2).
Although UTS is an important characteristic of materials, it is not very useful in elucidating the effect of various factors on metal properties because it depends in a complex way on the strengthening and embrittleing influence of these factors. If any one external factor ( e . g . , dissolved hydrogen) strengthens or weakens a rnetal but has no effect on its abitity to deform plastically, and if specimens in the starting (u), strengthened (y) and weakened (py) condition form a neck when tested to rupture in tension, the ratio of ultimate stresses Oy, o V and opy (Fig. 3) gives a true picture of the influence of this factor on the resisfaneg of the'fnetal to deformation. If, however, a specimen in a strengthened condition fractures prematurely (as a result of the embrittling influence of the factor under consideration) at a strain s', the ultimate stress o' of such a material may be below the initial value of this stress (Ou). When the embrittling influence is very strong (i. e., when fracture takes place at ~'), the ultimate stress of material in the strengthened condition O"y will be lower than that of the material in the weakened but not embrittled condition Opy. Fig. 1. Temperature dependence of the UTS of steel 1Khl8NgT without hydrogen (curve a) and with a hydroyen content of 50 m//100g (curve b).
i
E
2O
I
-;20
I
-40
I
+40 r,*C
Fig. 2. Temperature dependence of the UTS of nickel: 1) without hydrogen; 2. 3, 4) with hydrogen contents of 35, 50 and 125 ml/100g, respecti rely.
Fig. 3. Schematic representation of the variation in the ultimate stress under the influence of strengthening and embrittling factors.
These views were supported by the results of an experimental study of the influence of hydrogen on mechanical properties of steel [85]. When the hydrogen content in the steel studied was increased to 5 ml/100g, the dependence of the true yield point on strain was unaffected and the true UTS was reduced only due to the fracture taking place at lower strains.
460
Hardness measured by indentauon methods, especially under small loads producing relatively small strains, is one of the characteristics which reflect the resistance of materials to plastic deformation when the embrittling influence of various factors is suppressed. This is probably why the results of studies of the influence of hydrogen on hardness of steel are more positive than the corresponding data relating to the UTS. Even so, an increase in hardness of steel and other metals under the influence of absorbed hydrogen was reported in [4, 78-80, 82-84], whereas no effect of hydrogen on hardness of steel was observed by other workers [19, 75-77, 81]. A large number of investigations were concerned with the influence of hydrogen on the yield point, on a yield stress corresponding to a certain plastic strain, and on the shape of the stress/strain diagram in the i n i t i a l yield range. A noticeable reduction in the yield point of steel as a result of hydrogen absorption was observed during corrosion of steel in acid m e d i a [87] and during high-temperature hydrogen charging [14]. In isolated cases it was reported [60, 88] that under the influence of hydrogen the yield point is slightly reduced and the yield ledge disappears. The specific features of the influence of hydrogen in the initial deformation stages are regarded in [88] as a consequence of hydrogen driving carbon and nitrogen atoms away from dislocation centers, as a result of which the strength of blocking of d i s l o cations is reduced and so is the yield point of steel at any temperature. The disappearance of the sharp yield point and yield ledge under the influence of hydrogen [ 8 9 - 9 1 ] was attributed [90, 91] to the appearance of high internal stresses or local strains in regions of pores filled with m o l e c u l a r hydrogen at high pressures. Tensile tests carried out at - 150 ~ C on v a c u u m - m e l t e d iron after hydrogen charging and preliminary deformation at - 1 5 0 and - 7 0 ~ C (and also after aging at - 7 0 ~ C [98]) showed that the "hydrogen" yield ledge can be obtained when the "nitrogen-carbon" yield ledge is suppressed. On this basis it was postulated [93] that hydrogen in iron may interact with dislocations in the same way as other interstitial impurities. The appearance of the "hydrogen" yield ledge was attributed [92] to the formation of atomic hydrogen atmospheres on the segments of microslip and interaction of these atmospheres with moving microsIip segments. A hydrogen-induced increase in the yield point was observed in the case of molybdenum [94]. It should be pointed out, however, that many workers [65-69, 95] found no evidence of the influence of hydrogen on the yield point of c a t h o d i c a l l y polarized steel. Certain studies of the dependence of ductility and yield point of hydrogen-bearing steel and nickel on temperature and strain-rate are of a special theoretical interest. The ductility (reduction in area O) of hydrogen-charged low-carbon steel [51] vanes nonmonotonically between +60 and - 1 9 6 ~ C (Fig. 4). Hydrogen e m b d t t l e m e n t of steel at slow strain rates reaches its m a x i m u m at a temperature about 150 ~ C higher than the d u c t i l e - t o - b r i t t l e transformation temperature. The faster the strain rate, the less pronounced is the influence of hydrogen on the ductility of steel and the higher is the temperature corresponding to the m i n i m u m ductility. At the highest strain rates employed, no evidence of hydrogen e m b r i t t l e m e n t was observed.
~2~1/ ~
/ c
8
-200
-t~
o
+ tO0 - 2 0 0 - 700
0
7,~
+ 100-200 - 100
I o
T~
I
9 100 -200 -700
0
+tOO
T~
Fig. 4, Effect of temperature and strain rate on the ductility # of a steel c o n taining O. 2~ C and O. 7% Mn and tested at the following strain rates (s, min'l): a) 0,05; b) 100; c) 5 x lOS; d) 1,9 x 104. Curves I relate to hydrogen-bearing and curves II to hydrogen-free specimens. We obtained similar results [14] for armeo iron and iron-chromium alloys with a hydrogen content of 2 m l / 1 0 0 g . One of the graphs obtained, showing a ductility m i n i m u m at about - 8 0 ~ C is reproduced in Fig. 5. Other workers also observed [52] a nonmonotonie character of the temperature dependence of hydrogen e m b r i t t l e m e n t of steel with a h y drogen content of 2 the same applying to such metals as vanadium, tungsten and molybdenum [53].
ml/lOOg,
The effect of hydrogen on the dependence of the ductility of austenitic steels and nickel on temperature and strain rate has not been extensively studied, but one of our investigations E54] was concerned with this problem. The t e m p e r a ture dependence of the ductility of steel 1KhlSN9T is reproduced in Fig. 6. As in the case of f e r d t i e steels, the susceptibility of austenitic steel to hydrogen e m b r i t t l e m e n t (in a certain temperature interval) increases with decreasing strain rate. At the s a m e time, the hydrogen e m b r i t t l e m e n t of austenitic steels is less l o c a l i z e d in respect to temperature, and
461
varying the strain rate by two orders of magnitude has practically no effect on the temperature corresponding to the minimum ductility. A nonmotonic character of the temperature dependence of ductility was observed also in the case of nickel [45, 55] which, in this respect, behaves very much like austenitic steel (Fig. 7). It is interesting that athigh hydrogen concentrations (about 125 m//100g) nickel is brittle in the entire temperature interval studied, this being a clear evidence of irreversible hydrogen embrittlement. Analysis of the existing data on the influence of hydrogen on the ductility of iron, steel and nickel leads to the following conclusions: •0
iii
e
1) Hydrogen i n sufficiently high concentrations produces irreversible embrittlement of metals; this is manifested in bcc metals as increased susceptibility to brittle fracture at low temperatures and in fcc metals as brittleness which is temperature-independent.
~
40 1
9
2) The most characteristic manifestation of the influence of hydrogen on ductility is reversible hydrogen embrittlement; this is reflected in a loss of ductility to an extent which increases with decreasing strain rate, and in a nonmonotonic dependence of ductility on temperature which passes through a minimum in a certain low temperature range whose position again depends on the rate of strain.
l
20 O~
I
-t
-20Q -160 -Z20
I
I
I
-80
- 40
~7 7 ~
Fig. 5, Temperature dependence of the ductility ($) of a 5% Cr-Fe alloy tested at a strain rate of 0.6 ram/ /min. 1) Hydrogen-free specimens; 2) Hydrogen-bearing specimens.
At temperatures corresponding to the maximum reversible hydrogen embrittlement, all the metals we studied were characterized by these anomalies of the temperature dependence of the yield point (or flow stress).
A study of the temperature dependence of the yield point of armco iron at +20 to -160" C showed [96] that a hydrogen content of 2.5 m//100g increases the yield point of this metal at temperatures ranging from - 8 0 to - 1 2 0 ~ C by up to 10 kg/mm z, the corresponding increase at other temperatures being only 1-3 kg/mm z (Fig. 8).
8 0 12
0 I
_ o.~
I
i
~
IL
,
,
~ I
I
_
ll
l
_
,
.i
-200-720 -,~0 ~40 -200 -?~.0 -~0 +~O -,~00-1tO -OO +~O r, oc
r, "c
I
r, . c
Fig. 6. Temperature dependence of the ductility (~) of steel 1Khl8N9T tested at the following rates of strain (mm/min): a) 4.0; b) 0.4; c) 0.08. 1) Hydrogenfree specimens; 2) specimens with a hydrogen content of 25 ml/100 g. The point at which the curve os(T ) of hydrogen-charged specimens begins to diverge from the corresponding curve for hydrogen-free metal coincides with the minimum on the curve representing the temperature dependence of ductility of the hydrogen-charged material. One of our investigations [128] was concerned with the temperature dependence of the flow stress of nickel and an austenitic steel 1Khl8N9T at relatively large strains. The specimens were charged with hydrogen in autoclaves at 400500* C and a pressure of 600 atm. The resulting hydrogen content in the specimens was about 50 m//100g. It was shown that hydrogen produces a marked increase in the flow point of both metals in the entire range of strain (0.2-20% elongation) studied (Figs. 9 and 10). An unusually high resistance to plastic deformation was observed in the ease of hydrogen-eharged anstenitic steel at 2% strain. The temperature dependence of the flow stress of both nickel and austenitic steel has local maxima in the temperature interval corresponding to the maximum degree of hydrogen embrittlement; the height of these maxima increases with increasing strain. At present there are many theories of hydrogen embrittlement, the most numerous of which are based on the theory of internal pressure. An assumption common to all the theories in this group is the presence of voids in metals before they are exposed to the influence of hydrogen [97, 98]. According to these theories, the detrimental effect of absorbed hydrogen on the ductility of a metal is manifested only when the accumulation of molecular hydrogen in these voids takes place. When the concentration of dissolved hydrogen is sufficiently high, a hydrogen pressure in the voids increases leading to the appearance of tensile stresses in the adjacent metal volumes. These stresses not only facilitate the fracture of metals under an applied stress, but may lead to spontaneous cracking.
462
The internal pressure theory, even in its simple form, gives a satisfactory explanation of hydrogen embrittlement which is manifested by brittleness at subzero temperatures or temperature-independent brittleness of such f . c . c , metals as austenitic steel and nickel. Using this theory, one can also explain certain aspects of the reversible hydrogen embrittlement if one considers the process of formation of molecular hydrogen in voids in a metal undergoing plastic deformation. It is known that the degree of reversible embtittlement decreases with increasing strain rate and falling temperature. 9~ Both these factors inhibit the restoration of the dangerous molecular hydrogen pressure in voids growing in metals during plastic deformation, which explains 80 the reduced intensity of the embrittling action of hydrogen. Calculations showed [99] that at room temperature and not too slow strain rates, when the entire process of deformation and fracture of a specimen does not exceed qO 60 see, oriented diffusion of hydrogen toward regions of its concentration in microscopic cracks may maintain the hydrogen pressure in these cracks at a level of several thousands atmospheres. -200
-120
,-40
0 +~0
.Y*c
One of the shortcomings of the internal pressure theory is that it does not touch upon the subject of the role of hydrogen in the formation of voids Fig. 7. Temperature dependence of in which subsequently dangerously high hydrogen pressures are built up. The the ductility (r of nickel: 1) Hydromain shortcoming is in that it cannot explain the nonmonotonic character of gen-free specimens; 2, 3, 4) specimens the temperature dependence of ductility of hydrogen-charged metals. Cerwith hydrogen contents of 35, 50 and tain supplementary concepts [13, 100, 101] were incorporated in the internal 125 mi/100 g, respectively pressure theory with a view to elucidating the mechanism of transport of hydrogen to voids in metals. According to these concepts, hydrogen is delivered to sites of its recombination by moving dislocations which carry hydrogen atmospheres of the Cottrell atmosphere type; however, the introduction of these concepts did not eliminate the basic shortcomings of the internal pressure theory.
.
.
.
.
.
.
.
,
, , ,
a
s
b
9
30 w
2O
w
v
0
b
~'Pi
70
0
a
,~
I.
=:::
601~"~
^
"
~,l
.-;--
Q i
-J
-
~, H
08 ,.~
t
160 "120
I
-80
I
- 40
t
Toc
-200
0
-r
-40
":roc§
Fig. 9. Temperature dependence of the reduction in area (~) and flow stress at strains of 2 and 20% (02 and oz0, respectively) of hydrogen-charged (index "I-t") and hydrogen free (no index) nickel.
Fig. 8. Temperature dependence of the ~ e l d poing ~s) and reduction in area (~) of armco iron: a). Hydrogen charged; b) hydrogen free,
According to the Griffith-Orowan theory [102, 1031 a crack present in a metal can grow only if the stress acting on it reaches a certain critical level"
, / 2es ~
= F~C
(1)
where Ocr is the critical stress level for crack propagation, E is the Young modulus, S denotes the specific surface energy of the metal, and C is the crack half-length. If it is postulated [64, 104] that the surface energy of a crack in a hydrogencharged metal is reduced due to the adsorption of hydrogen liberated by the solid solution, a lower critical stress should - i n accordance with ( 1 ) - b e required to trigger off crack propagation. If it is borne in mind that hydrogen being
463
absorbed on a freshly formed crack surface finds its way from the solid solution to the region of tensile stresses in the crack tip by a diffusion process, it becomes clear why the degree of hydrogen embrittlement is increased at slow rates of strain. The problem of crack propagation in hydrogen=charged metals was also approached from the energy standpoint by other workers [105-107], who assumed that hydrogen promotes the propagation of Griffith's cracks by increasing the energy liberated during crack growth. Like the internal pressure theory, adsorption theories of hydrogen embrittlement cannot explain the nonmonotonie character of the temperature dependence of ductility of hydrogen=charged metals. In the framework of these theories the reduction in ductility may be explained as a result of premature crack propagation; however, they do not explain the observed increase in the flow stress. It was attempted recently [108, 109] to analyze the combined action of hydrogen adsorption and high hydrogen pressures in cracks during brittle fracture of hydrogen=charged metats. It is assumed [109] that V-shaped micro = cracks are nucleated in a metal acted upon by external stresses leading to the formation of dislocation pile-ups on obstacles. The fomlation of such a crack in a specimen (whose radius is larger than C) is shown schematically in Fig. 11. The critical crack length [20] is given [109] by
80 ~.4a
-750 -r2O
-4#
r.,c
+40
2C = 4 0 7 {~ (1.L. ~) r [ r + (~ + p ) sin O - - z cos 01} -1, Fig. 10. Temperature dependence of the reduction in area ($) and flow stress at strains of 2 and 20% (az and az0, respectively) of steel 1Kh18N9T (notation as in Fig. 9).
(2)
where G is the shear modulus, $ surface energy of hydrogen~charged metal, v Poisson's coefficient, o normal stress, r tangential stress, p hydrogen pressure in the crack, ~ the angle at which the crack is inclined to the sRp plane, and 9 the total force acting on the hydrogen-filled crack (ff = (o + p)Z +r").
A crack of a critical length determined by (2) wiU grow at a eatastrophic rate when the following condition is satisfied: ]/~(~ -F P ) ' q- z" q- [(0 + p ) sln 0 -- ~cos O] >/ 4~ nb'
(3)
where n is the number of dislocations in a pile-up or in a dislocation array in the slip plane, and b is the Burgers vector. And so. the catastrophic growth of a crack under the influence of external stresses depends on the ratio of the surface energy of the metal to the hydrogen pressure inside the crack [109].
floIJ
Fig. 11. A schematic representation of a V-shaped crack [109].
a
b
Fig. 12. Formation of a crack by merging of dislocation climbing in planes (i01) and (101).
It was calculated [109] that the quantity of hydrogen required to pmduoe severe embtittlement of steel as a result of the build-up of hydrogen pressure in micmcracks and the reduction in the metal surface energy due to hydrogen adsorption is very small: about 0.1% of the hydrogen dissolved in the metal. 464
Hydrogen e m b r i t t l e m e n t of metals may be considered [111] in the framework of the Cottrell theory [112] of brittle fracture. According to this theory, the i n i t i a l crack in the brittle fracture of fcc metals is formed as a result of the appearance in the plane (001) of stable sessile dislocations according to a reaction: . - t - ~a ( 1 1, 1 ) -~ a ( 0 0 1 ) "
(111)
The formation of such a dislocation is shown schematically in Fig. 12. This dislocation is a nucleus of a crack which will grow as a result of repeated merging of dislocations climbing in planes (101) and (10f). If the slip planes are at 45 ~ to the direction of applied force P, and if there are n dislocations moving along each of these planes to form a dislocation with the Burgers vector equal to na (this dislocation being the nucleus of a crack of a length C which grows under the influence of applied force P), the energy of a crack dislocation [111, 112] is given by
Ira-' a'-'
W -
In
4r~ (1 ..... ,)
4R
+ 2,~C - -
r: (1
-
C
-
~) P"C-'
PnaC
8~
,
(4)
2
where tl is the elasticity modulus, v Poisson's coefficient, and R the effective radius of a crack dislocation with the Burgers vector equal to na. The first term of Eq, (4) represents the energy of the stress field of a crack dislocation; the second and third terms represent, respectively, the surface energy of the crack and its elastic energy in the applied stress field, the fourth term representing the work done by the applied force i n increasing the crack volume. Taking C,- . . . . . . . .
~ C . ,"-
81~7
~
=
'
(5)
Cottrell simplified (4) and obtained
w =
c,
(4R
c=
2c ,
rc, '1, c 1
j.
Starting from the condition of the equilibrium length of the crack, we obtained an equation quadratic in C:
C~-
~ CC..-i [1 - - 2 (L)"I
CjC,,=O.
(7)
From Eq. (7) it follows that: 1) At P = 0 a crack is formed only due to the action of dislocation forces since C 2 = ~o and C = C1, i . e . , C 1 is the length of the crack formed by dislocations; 2) at n = 0 ( i . e . , when there are no dislocations) and p > 0 we have C1 = 0 and C = C2, i . e . , C 2 is the length which should be reached by a Griffith crack under the action of stress P; 3) when n dislocations are present in a m e t a l subjected to a stress P, the crack becomes unstable and infinitely grows; this occurs at (Gl/C2)~/z = 0.25, i . e . , when Pna = 2 7. The Cottrell mechanism may be applied to explain hydrogen e m b r i t t l e m e n t [111]. This explanation is based on the fact that the free energy of a m e t a l becomes reduced if its hydrogen content is below the solubility limit, and is increased if the quantity of absorbed hydrogen is larger than the solubility l i m i t . At the same time it is postulated that the increase in the free energy of the m e t a l due to intense hydrogen charging leads to an increase in the energy of the dislocation pile-up which forms a crack. Depending on the crack length the increase in the dislocation energy is described by an arbitrary function
A
f
Cm
(8)
where A is a constant depending on the hydrogen content and n is a certain positive coefficient, The energy of a crack dislocation in a hydrogen-charged m e t a l is described [111] by an equation which was o b tained from (6) by adding a term f to its right side:
WH -----2#
C , In
-~ C
c,
2C._,
,. 2 kC~]
] -]-
(9)
Analyzing this equation for cases when the hydrogen content is above and below the solubility limit, the authors of [111] concluded that noticeable hydrogen e m b r i t t l e m e n t is produced only when a given metal is supersaturated with hydrogen; it is also necessary to apply to the m e t a l a sufficiently high stress to initiate the m o v e m e n t of dislocations. A crack nucleus formed as a result of merging of c l i m b i n g dislocation is produced only at a certain definite stress.
465
Hydrogen atoms concentrated in the stretched regions of glissile dislocations form clusters in the tip region of the crack nucleus when these dislocations merge. If the metal is supersaturated with hydrogen, the energy of a crack dislocation is considerably increased with a corresponding reduction in the stress level at which the crack nucleus is formed. Although it is satisfactory in many respects including the fact that it considers the mechanism of hydrogen embrittlement of metals from the standpoint of the dislocation theory of plastic deformation, the theory formulated in [111] is not free from shortcomings of all the previous theories: It leaves unexplained the increase produced by hydrogen in the flow stress and the complex temperature dependence of the ductility of hydrogen-charged metals. These aspects of hydrogen embrittlement cannot be explained either in the framework of the theory of maximum stresses [110]. And so, none of the existing theories of hydrogen embrittlement can offer a satisfactory explanation of two established facts: The increased flow stress of hydrogen-charged metals and the complex temperature dependence of their ductility with a minimum in a certain low-temperature interval. The latter fact could be explained in the framework of the above reviewed theories if one could verify even one of the following hypotheses:
a) At temperatures above that corresponding to the minimum ductility, hydrogen is liberated by specimens undergoing mechanical tests; b) the ductility of hydrogen-charged metals passes through a minimum because the hydrogen pressure in voids reaches its equilibrium level; when the temperature is raised the hydrogen pressure falls which leads to an increase in ductility; c) at temperatures higher than that corresponding to the minimum ductility, there is a substantial reduction in the adsorption of hydrogen on the walls of internal mieroeraeks formed during the deformation; this weakens the influence of hydrogen in its role of a surface-active substance and leads to an increase in ductility. The first of thesd hypotheses is refuted by the results of many of our experiments which showed that the initial hydrogen content in the metals studied remains practically the same in the course of testing. To check the second hypothesis, let us compare the temperature dependence of the equilibrium hydrogen pressure at various concentrations of hydrogen in solid solution (Fig. 13) constructed by a method suggested in [113] with c a l culated data we obtained [99] for the temperature dependence of hydrogen pres2 sure in micropores which may be reached during tests lasting 3 hr. When one compares the two groups of curves in Fig. 13, it is easy to see that if the hypothesis under consideration were true, the ductility of a metal with a hydrogen 42 content of several m//100g would reach its minimum above room temperature, i . e . , at a temperature at least 100" C higher than that observed in practice. Moreover, the temperature of this minimum should be strongly dependent on the r t hydrogen content, which is not shown by experiment. / ~"V 1 0 _ ~_~ ~--_~S_--_'.~~- --1 The third hypothesis must be rejected if one considers the temperature dependence [114] of the adsorption of hydrogen on iron at normal pressure. -100 -80 -20 7,,~20 According to this data, raising the temperature from - 1 0 0 to 0* C produces a linear variation in hydrogen adsorption which in this temperature interval is Fig. 12. Temperature dependreduced by about 25%. ence of the equilibrium hydrogen pressure (continuous curves) and the pressure which may be reached in micro pores in the course of deformation (broken curves) plotted for a metal with the following hydrogen contents (ml/100 g): 1) 5.0 g; 2) 2.0; 3) 0.5.
Since increasing the temperature produces an increase in the hydrogen pressure in micro-pores, the adsorption of hydrogen on the walls of these pores will, in fact, deorease even less. Consequentiy, the low-temperature maximum of hydrogen embrittlement cannot be related to the temperature dependence of adsorption either.
The above discussed singularities of hydrogen embritttement of steel, iron and nickel can be explained if in considering the mechanism of this phenomenon one takes into account the interaction of impurity atoms with dislocations during the plastic deformation process, as was done in studies of certain aspects of the temperature- and strain rate-dependence of mechanical properties of steel associated with the interaction between dislocations and nitrogen and carbon atoms dissolved in ferrite [115-117]. A postulate was then formulated and indirectly proved, according to which there is a certain combination of temperature and strain rate at which impurity atoms may actively interact with new dislocations formed during the deformation process. This interaction consists in the formation
466
of Cottrell atmospheres around the newly generated dislocations, as a result of which the active dislocations, rapidly pinned by the Cottrell atmospheres and unable to be discharged as a result of slip, form additional stationary barriers to the motion of other, not yet pinned, dislocations. The number of piled up dislocations rapidly increases, plastic deformation becomes inhibited, and the metal becomes brittle. It was postulated that the movement of dislocations is inhibited by dislocations pinned by atom atmospheres when the impurity atoms can migrate to the dislocations and form there sufficiently effective (dense) atmosphere before the yield point is reached. By comparing the time required to reach the yield point at a given strain rate with the time necessary for a certain number of impurity atoms (sufficient to ensure effective blocking) to reach dislocations [118], an equation was derived which makes it possible to estimate the minimum temperature at which the secondary pinning of dislocations by impurity atom atmospheres (formed during the deformation 80 process) is possible. Calculations carried out for the case o~carbon in vacuumannealed low-carbon steel extended at a loading rate of 10 ' ~ kg/mm z see [116] showed that this temperature is 12" C, i . e . , near room temperature. The temperature dependence of the yield point and elongation of the steel studied is reproduced in Fig. 14, where the temperature dependence of the 40 g yield point calculated from data obtained by other workers [119-120] without taking into account the secondary pinning of dislocations during the deforma2o tion process is shown for comparison. The temperature at which the temperature dependence of the yield point begins to diverge from the theoretical [119, 120] is in good agreement with our calculated value.
'~O
\\
40
Anomalies in the temperature dependence of elongation and yield point of steel (Fig. 14) are characteristic for the temperature interval of blue brit20 tleness. This phenomenon [121-127] is a result of the onset (during deformation at certain temperatures and strain rates) of an elastic interaction of ~0 J i dislocations with carbon or nitrogen atoms in solid solution which, in the -200 0 200 400 T, ~ temperature interval, can diffuse at sufficiently fast rates. As shown by the results of many investigations [121, 122, 128], blue brittleness is characFig. 14. Temperature dependence terized not only by anomalies on the curves ~(T) and Os(T), observed usually of elongation (curve 1) and yield between 50 and 300* C, but also by a discontinuous character of plastic depoint (experimental curve 2 and formation which is reflected in serrated strain/stress diagrams plotted from theoretical [119, 120] curve 3) of results of tensile tests on "hard" testing machines. The results of several a low-carbon steel. studies [121, 123, 128, 129] lead to a conclusion that btue brittleness is a special case of a general phenomenon, i.e., dynamic strain aging which takes place in systems metal-impurity (interstitial or substitutional). A theory of strain aging of sohd solutions was formulated by Cottrell [124-127]. According to this theory, the rate of displacement of dislocations carrying atmospheres of dissolved impurity atoms is controlled by the rate of diffusion of these atoms. The critical velocity Per of an edge dislocation [126]
4D v
cr=
- -
[
,
(10)
where I is the ~characteristic" radius of the atmosphere, and D is the diffusion coefficient of dissolved atoms. It follows from (i0) that if a dislocation is moving at a rate faster than that of an impurity atom at a small distance r -- 1/2, it will leave its atmosphere behind. Consequently, dislocations formed during the deformation process will be freed from their atmospheres. As a result of periodic disruptions in the loading force due to dislocations being freed from the force fields of their atmospheres, the deformation of a metal will proceed in jumps, and serrations, i.e., a series of successive sharp yield points, will appear in the stress/straindiagram.
According to Cottrell, the critical strain rate ecr = VcrPb,
(11)
where b is Burgers vector and p the dislocation density. Assuming 4 b / l ~ 1 and substituting this value in (11), we obtain a simple expression for the critical strain rate:
ecr---- Dp.
(12)
467
From (12) we can determine the minimum value of the diffusion coefficiSnt at which impurity atoms will dynamically interact with dislocations during deformation. At p = 109 em "~- and strain rates of the order of 10 ~ see "~ the minimum value of D is 10"tSem'sec'l. According to our calculations, such values of D of interstitial impurities (C, N, H) in various metals correspond to various temperature intervals (Table !). Table 1 Temperature Intervals in Which Strain Aging may Occur in Metals Containing Various Impurities
Temperature interval, "C
.Metals
Mo Ni
Fe(a) Fe(~) V
C
N
H
700--800 250-300 100--150: 300--400 300--500
..-'- -100--200
from -1501 to --80 ~ from--1201to ,--30 [ from--100 to --80 I Ifrom-150 to--80
Literature source from w h i c h the value D was obtained
[ 131 ] [132,. 1331 [134, 135] [135] [136]
The calculated data is in satisfactory agreement with experiment. And so, dynamic strain aging associated with the presence of carbon and nitrogen was observed in the following temperature intervals: 600-800* C in molybdenum [137, 138]; 30-300* C in nickel [139]; 50-800* C in armeo-iron and low-carbon steel [121, 128, 129]; 200-500* C in a chromium-nickel austenitie steel [128]; 300-350* C in vanadium [136]. In the case of the above metals discontinuous plastic flow and anomalous variation in their ductility and flow stress in the temperature intervals indicated were observed. At certain temperatures the ductility was sharply reduced to its minimum level, this being accompanied by a marked ine/ease in the flow stress. As shown in Table 1, similar phenomena in metals containing hydrogen should take place at subzero temperatures. As predicted by calculations, these effects were in fact observed in hydrogen-charged nickel at temperatures ranging from - 1 2 0 to -45* C [45] and from - 1 4 0 to -60* C [133]. "Hydrogen" dynamic strain aging at temperatures predicted by calculations was observed also in armco iron [96, 140], austenitie steel [143] and vanadium [53]. In this ease, like in the ease of "carbon-nitrogen" dynamic strain aging, discontinuous flow of the metal was also observed, this being aecompartied by the appearance of a minimum on the curve @(T) and a maximum on the curveo(T) (see Figs. 9, 10, 15). It follows from the foregoing that both the blue brittleness and reversible hydrogen embrittlement of bcc and fcc metals studied are special cases of dynamic strain aging of metals containing impurities. Our [96, 115-117, 133] and other studies [136, 139, 141, 142, 144] showed that the Cottrell theory of elastic interaction of impurity atoms with dislocations is equally well applicable to hydrogen atoms as to other elements (interstitial and substitutional). And so, the begining of the temperature interval of the onset of discontinuous deformation of hydrogen-charged nickel (-140" C) which we calculated [133] with the aid of Cottrell's formulas for carbon and nitrogen impurities [125, 126] is in good agreement with experimental data. Similarly, the experimental point at -120" C (see Fig. 8) which marks the beginning of the temperature interval of inhibited plastic deformation of hydrogen-charged armeo iron (due to secondary blocking of dislocations by hydrogen atom atomspheres) is in good agreement ~vith the value we calculated [96] from an equation derived in [116] for the blue brittleness. The fact that the various impurities, including hydrogen, have essentially the same effect on the behavior of metals during plastic deformation may serve as a proof that the reversible hydrogen embrittiement is due to the action of the atomic or protonized hydrogen which, being in solid solution, distorts the metal crystal lattice as a result of which the Cottrell type interaction of this impurity with dislocations becomes possible. As shown above, the general effect of various impurities which dynamically interact with freshly formed dislocations on the behavior of metals undergoing deformation is the same. However, depending on the nature of impurity, certain specific effects may be observed in any given ease. And so, in the case of blue brittleness the temperature range of inhibited deformation (anomalous yield point) is correlated with the temperature of reduced ductility. Hydrogen e m brittlement differs from the blue brittleness in that the anomalous variation in the yield point is in the former ease localized in the initial stage of the reduced ductility range at subzero temperatures. If hydrogen embrittlement were associated only with the inhibition of plastic deformation due to the secondary pinning of dislocations by hydrogen atmospheres, the anomalous variation in the yield point should extend over the entire temperature range of reduced ductility, as happens in the case of the blue brittleness. This discrepancy between experiment and the above outlined
468
hypothesis can be explained i f it is assumed that in the reduced ductility range, starting from the temperature at which the anomalous variation in the yield point ceases, another process comes into p l a y leading to a reduction in both d u c t i l i t y and flow stress. This process consists probably in the form of cracks facilitated by the diffusion of hydrogen to the regions of c r a c k nucleation and growth. At a given strain rate, this process ~0, should be manifested at higher temperatures than the process of secondary pinning of dislocations because it requires the diffusion o f larger quantities of !~ i i hydrogen through larger distances. ~. ...~. 7 ~- .~, t ~ .-x~\\
~ ~ %"
?O~_3_f~_~"~e' 6 "Q~~.2~[.~.~?.#
-80
0
V
60
~
- ~0
.2P J
-200 0
I
I
200 4,00 r,~
0
Fig. 15. Temperature dependence of m e c h a n i c a l properties of steel (0.21% C) a n d a r m c o iron. 1 ) o s of steel; 2) ditto, theoretical curve; 3) o s of hydrogen-charged armco iron; 4) ditto, for hydrogenfree m a t e r i a l ; 5) 8 of steel; 6) @ of hydrogen-charged armco iron; 7) ditto, for hydrogen-free m a t e rial.
This hypothesis is shown s c h e m a t i c a l l y in Fig. 16. Depending on the relative position of curves Os(T ) and o c ( r ) , the resulting curve of e x p e r i m e n t a l l y determined stress producing the onset of residual strain m a y have o n l y o n e region of "anomalous" increase in thisstress (at T1), as is observed in the case of armco iron (see Fig. 8), or two such regions (at TI and T2), as happens in the case of austenitic steel and nickel (Figs. 9 and 10). In a c c o r d a n c e with our theory of hydrogen e m b r i t t l e m e n t , it should be postulated that the recovery of plasticity starting from the temperature corresponding to the m i n i m u m ductility is associated with a reduction in the number of additional barriers to the m o v e m e n t of dislocations formed during the d e formation. It is further postulated that this reduction is due to a reduction in the number of dislocations pinned by hydrogen atmospheres since the latter are dispersed at e l e v a t e d temperatures. It is known [145] that the "settled l i f e t i m e " of an atom in a position of the m i n i m u m potential energy is given by = % exp
(f/kr),
(13)
where E is the energy barrier between two a d j a c e n t equilibrium states, k is the Boltzmann constant, T absolute temperature and r0 the natural atom oscillation frequency. The "settled l i f e t i m e " of an atom in a dislocation field is = r exp [(E +
U)/kTI,
(a4)
where U denotes the bonding energy between a dissolved atom and a d i s l o c a tion. Let us take for hydrogen r0 = 10 "~3 see and E = 3 . 6 x 10 "13 e r g / a t o m [86. 146]. The value U is found [126] from the equation U = (A s!n
O)/r,,
(15)
where A is a coefficient depending on the m e t a l properties and the degree of its l a t t i c e distortion due to dissolved i r a purities; r and I) are polar coordinates of the atom in relation to the dislocation. If we take r = 2 x 10 .8 cm and sin I~ = 1, the value A is determined [118] by A --
G), 1 -k- ~ 3~l--v
Av,
(16)
where v is the Poisson ratio equal to 0.28, G is the shear modulus which for iron is equal to 7.2 • 10 l l d y n e s / c m z, and Ay is the change in the e l e m e n t a r y c e l l volume due to dissolved hydrogen. Dissolved impurity atoms tend to congregate in the dislocation r, r, field, where, as a result, the impurity concentration will be higher than average. Taking [118] n'/n = 0.5, where n' is the number of hydrogen Fig. 16. Schematic representation of the atoms dissolved in a crystal with n iron atoms, and using data [12] on temperature dependence of the yield the expansion o f iron l a t t i c e due to dissolved hydrogen, we find that at point (os), stress required to produce n'/n = 0.5 ; 33.3 at % = 7 x 10 s m l / 1 0 0 g , we have Ay = 0.3 • 10 .23 cracking (Oe) and e x p e r i m e n t a l l y d e cm s. Then, in accordance with (16), A = 10 .20 dynes cmz. Substituting termined m i m m u m stress required to this value of A in (15), we c a l c u l a t e the energy of interaction between produce residual elongation (Oe) of a hydrogen atom and a dislocation, and from (14) we find the mean hydrogen-charged armco iron. *settled l i f e t i m e " of a hydrogen atom in an atmosphere formed around a dislocation; the m e a n l i f e t i m e of an atmosphere m a y be taken (to simplify calculations) to be of the same order of magnitude.
469
The time of the formation of a CottreU atmosphere is given by
V-2 kr
IIV/,Vs
,
(17)
where k is the Boltzmann constant, T absolute temperature, N/N S the degree of occupation of the atmosphere, no the concentration of the solution, )~ the minimum interplanar spacing in the crystal lattice, and D the diffusion coefficient of the impurity atoms. Let us take the coefficient of diffusion of hydrogen in an iron alloy (containing 0.5% Cr) to'be 2.3 • 10 -s cml/sec[99, Table 2 147]. Having determined the diffusioncoefficients at temperaTemperature Dependence of the Time of Formatures other than room temperature and taking N/N~ = 0.1 [118], 18 tion (t) and the Lifetime (r) of a Hydrogen Atno = 2 m l / 1 0 0 g 9 cm .JI , and k = 2. 4 8 x 1 0~" S cm. we mosphere Around a Dislocation find from (17) the time of the formation of a hydrogen atmosphere around a dislocation. Values of t and r calculated in this t, sec, st way are given in Table 2. T, o C
N/Ns=O.I
N/Ns=0.3
"~, SeC
The dispersion of hydrogen atmospheres at room temperature is unlikely to take place because the time of the formation 10-7 5.10-T 10--4 20 of an atmosphere at this temperature is by three orders of m a g 1,5,10 -11 7,5:10-11 500 3.10 -~~ nitude shorter than its mean lifetime; hydrogen atom atmos5.5.10 "-11 700 2,5.10 -1~ 10-u pheres become unstable at much higher temperatures. Consequently, the increase in the ductility of a hydrogen-charged iron-chromium alloy at temperatures ranging from -100 to 20" C (see Fig. 5) cannot be attributedm the dispersionof hydrogen atom atmospheres normally surrounding dislocations. It m a y be postulated, as an alternative explanation, that the density of dislocationspinned during the deformation at a given strain rate begins to decrease due to the fact that dislocationsbecome able to move with the surrounding atmospheres. The critical dislocation velocity Vcr, at which an atmosphere can move with a dislocation, can be determined [126] from Eq. (10) and the corresponding critical strain rate ecr from Eq. (11). Taking in (10) l ~ A / k T [126] and solving simultaneously equations (10) and (11). we obtain
%r= 4A ' D,, kpbT e x p ( - Q/RT).
(18)
It follows from this equation that increasing the strainram produces an increase in the minimum temperature at which an impurity atom atomosphere can move with its dislocation. In a study of the mmperamre- and strainram-dependence of hydrogen embrittlement of iron-chromium aUoys [130] such a shiftof the temperature TnCOrresponding to the m i n i m u m ductilitytoward higher mmperatures was. in fact, observed. Consequently, it m a y be concluded that the recovery of ductilityat temperatures above T n is associated with a reduction in the number of blocked dislocationdue to the fact that hydrogen atmospheres in these circumstances become unable to move with their dislocations. The hypothesis which explains the sharp reduction in the ducfihty of hydrogen-charged specimens in terms of pinning of glissiledislocations, and the theory according to which the recovery of ductilityis associated with the ability of hydrogen atmospheres to move with their dislocationsat temperatures above Tn are applicable not only to iron and its alloys. Both these concepts m a y be used to explain the nonmonotonic characmr of the temperature dependence of ductility of nickel and ausmnitic smels saturated with hydrogen (the case of reversiblehydrogen embrittlement) [55]. In spite of some differences in the temperature dependence of ductilityand flow stressof bcc and fcc metals, the basic laws governing the manifestation of reversiblehydrogen embrltflement of these metals are analogous (see Figs. 8, 9, 10). Analysis of resultsobtained in studies of the influence of hydrogen (in quantities not exceeding 0.005%) on the behavior of these metals during deformation at slow strainrates shows that the mechanism of thisinfluence is the same in both cases, involving dynamic interactionof dissolved hydrogen with freshly formed edge dislocations(in the case of fcc metals). Further developments in the theory of hydrogen embrittlement and the formulation of a more rational explanation of experimemally observed singularitiesof this phenomenon are possible if it is taken into account that hydrogen dismlved in a metal interactsduring the deformation with crystal latticedefects.
470
However, before a qualitative and more detailed theory of hydrogen embrittlement can be formulated, it is necessary to carry out theoretical and experimental studies directed toward elucidating the following problems: 1) the state of hydrogen in metal crystal lattices and the nature of the interaction between hydrogen and lattice defects, impurities and alloying elements; 2) the relation between hydrogen embrittlement (due m dynamic strain aging) and weakening (due to facilitated crack formation) of hydrogen-charged metals; 3) the distribution of absorbed hydrogen in respect of its states and the dependence of this distribution on temperature, on time elapsed from the compIetion of hydrogen-charging treatment and on the rate of strain. REFERENCES 1. 6, 1960. 2. 3~ 4. 5. 6. 7. 8. 9. i0.
K. V. Popov and V. G. Savitskii, in: Investigations of Heat-Resistant Alloys [in Russian], Izd. AN SSSR, vol.
L. S. Moroz and T. E. Mingin, MiTOM, no. 4, 1962. H. C. Rogers, Mat. Protection, 1, no. 4, 26, 1962. N. A. Galaktionova, Hydrogen in Metals [in Russian], Metallurgizdat, Moscow, 1959. M. Moneyr0n and D. Sgfdrian, sondage et techn, connexes, 11, no. 9'10, 227, 1957. Mima and Midauta, RZh. Met., no. 11, 22303, 1951. F. grdman-Jesmitzer, Neue Hatte, 4, no. 11, 669, 1959. P~ V. Sklyuev, L. Ts. Kvater, and V. E. Sharipo, Stal, no. 10, 1956. V. Ya. Dubovoi and V. A. Romanov, Stal, no. 8, 1947. P. V. Sklyuev, L. I. Kvater, and B. D. Pert.v, in: Problems of Metallography and Heat Treatment [in Russian], Moscow, Sverdlovsk, Mashgiz, 1956. 11. B. P. Sinyutin, MiTOM, no. 1, 1956. 12. V. A. Yagunova and K. V. Pop.v, FMM, 12, no. 2, 1961. 13. P. Bastien and P. Azou, Rev. mgta!lurgie, 54, no. 2, 93, 1957. 14. L. A. Glikman, V. P. Teodorovich, N. N. Kolgatin, an d V. I. Deryab~na, in: Chemistry of Organic Compounds of Sulfur and Nitrogen Present in Oil and Products [in Russian1 vol. 3, Ufa, 1960. 15. W. Zabik, Przegl. mechan~ ! 6, no., 2, 58, 1957. 16. A. R. Trojan., Metal. pmgr., 77, no. 2, 112, 1960. 17. J. Wilson, V. A t Bqnner, H. D. Burnham, and I. T. Skei, The IVth International Oil Congress [Russian transIation], 8, Gostoptekhizdat, Moscow, 1956. 18. J. Colbus, Z. Schweisstechnik 46, no. 12, 1956. 19. C. H. van Ness, Petrol. Refiner., 35, no. 3, 160, 1956. 20. N. Christensen, Sveisetekr~kk, 11, no. 2, 14, 1956. 21. W. A. Bell, Prod. Engng., 26, no.~ 3, 189, 1955. 22. I. D. Hobson and C . Sykes, J. Iron and Steel Inst., 169, 209, 1951. 23. I. Ohnishi and Y. Kikuta, Proc. II Japan CongT. Test. Mater., Kyoto, Japan S.c. Test. Mater., 113, 1959. 24. J. Seabrook, N. J. Grant, and D. Carney, Trans. AIME, 188, 1317, 1950. 25. Onisi, Kikuta, Kavan, and Kakuno, J. Japan Weld S.c., 25, no. 11, 27, 1956. 26. A. R. Troiano, Spec. Rep. Iron and Steel Inst., no. 7% 1, 1962, 27. W. Hofmann and W. Rauls, Abhandl. Braunsehw. Wiss. Oes., 18, 158, 1961. 28. E. R. Slaughter, Open Hearth Proc., 39, 114, 1956. 29. H. J. Read, Hydrogen Embrittlement Mater. Finish., N. Y.-I~ondon, 1961. 30. Ya. M. Potak and O. P. Breslavtseva, in Some Problems of the Strength of Solids [in Russian], Izd. AN SSSR, 1959. 31. 32. 33. 34~ 35. 36. 37. 38. 39. 40. 41. 42. 43~
G. Schevek, G. Sachs, and K. Tong, proc. Amer. S.c. Test. Mater., 57, 682, 1,957 (1958). G. V. Karpenko, Yu. I. Babel, and R. I. Kripyakevich, DAN SSSR, 145, no. 1, 1962. J. A. Gurklis, D. McGraw, and C. L. Faust, Plating, 47, no. 10, 1146, 1960. M. Smialowski, g. Gasior, A. Waclawski, and Z. Mlynamzyk, Prate Inst. Hutn., 14, no. 2, 95, 1962. J. D. Fast and D. J. van O0ijen, Philips techn, tijdschr., 24, no. 8, 25, 1962. D. D. Perlmutter and B. F' Dodge, Industr, and Engng. Chem., 48, no. 5, 885, 1956. H. C. van Ness and B. F. Dodge, Chem. Engng. Progr., 51, no. 6, 266, 1955. R. L. Mills and F. I. Edeskuty, Chem. Engng. Progr., 52, no. 11, 477, 1956. R. Santt, Rev. Nickel, 26, no. 1, 16, 1960. J. D. Hobson and I. Hewitt, J. Iron and Steel Inst., 178, no. 2, 131, 1953. F. Eisenkolb and G. Ehrlieh, Arch. Eisenhattenw.,125, no. S/4, !87, 1954. L. I. Tushinokii, Trudy Vost.-Sib. ill. AN SSSIL Vyp., 6, 86, 1957. E. P. Neehai and K. V. Pop.v, RMM, 11, no.2, 1961.
471
44. E. P, Neehai and K. V. Popov, in: Investigations of Steels and Alloys [in Russian], Izd. "Nauka, " 1964. 45. T. Boniszewski and G. C. Smith, Aeta met., 11, no. 8, 165+ 1968. 46. M. Satoru, J. lapan ~nst. Metals, 25, no. 2~ 101, 1961. 47. E. P. Nechai, in: Effect of Hydrogen on the Properties of Steel, 1968. 48. B. A. Wilcox and G. C. Smith, Acta met., 12, no. 4, 871~ 1964. and V. I. Deryabina, F M M , 16, no. 5, 1968. 49. L. S. Moroz, N. N. Kolgatin, V. P. Tr 50. O. S. Popova and A. T. Sanzharovskii, D A N SSSR, 186, no. 3, !961. 51. J. T. Brown and W. M. Baldwin, J. Metals, 6~ ser. 2, no. 2, 298, 1954. 52. P. Bastien and P. Amiot, C. r. Aead. Se./,, 288, no. 28, 2288, 1954. 53. W. M. Baldwin, The Metal Molybdenum, Cleveland, Ohio, 279+ 1988, 54. E. P. Neehai and K. V. Popov, F M M , 14, no. 2, 1962. 55. K. V. Popov and E. P. Nechai, Izv. A N SSSR, Metally, i, 172, 1966. 56. g. P. Klier, B. B. Muvdi, and G. Sachs, J. Metals, 9, no. I, 106, i~57. 57.H. H. Johnson, R. D. Johnson, R+ P. Fromberg, and A. R. Troiano, Case Inst. Technol., Cleveland, Ohio,
1955. 58. Iosii, RZh, Met., no. 11, 22405, 1957. 59. K. Winterton, Brit. Weld. 1., 2, rio. 9, 885, 1955. 60. H. G. Vaughan and M. E. Morton, Brit. Weld. J., 4, no. 1, 40, 1957, 61. A. Dvo)'ak, Hutnik, 6, n0+ ii, 337, 1956. 62. K. Sachs and S. H. Melbourn, Spee. Rept. Iron and Steel Inst., 1961. 63. M. I. Zamotorin and T. S. Kara~0va, Trtldy LPI, Metallovedellie, no. 218, 1962. 64. N. J. Perch, Phil. Mag., i, set. 8, rio. 4, 881, 1956. 65. R. P. Fromberg, W. J. Barnett, arld A. R. Troiano, Trans. ASM, 4"/, 892, 1955. 66. G. V. Karpenko and R. I. Krip'yall~vich, D A N URSR, no t i, !958. 67. G. V. Karpenko and R. I. Krip'yakevieh, D A N SSSR, 120, no. 4, !958. 68. G. V. Karpenko and R. I. Krip'yakevieh, FMM, 8, no. i, 1969. 69. G. V. Karpenko and R. I. Krip'yakevieh, prikladna mekhanika, 4, no~ 4, 1958. 70. Yu. A. Nekhendzi and V. P. Kalellov, Izv. vuzov, Chernaya metallyrg~ya, no. 7, 1959. 71. G. V. Karpenko and R. I. Krip'yakevlch, Effect of Hydrogen on the Properties of Steel [in Russian], MetaUttrgizdat, Moscow, 1962. 72. K. V. Popov, V. A. Y~tgunova, and N. A. Khvorostukhina, in: Investigatlonsof Heat-Resistant Alloys [in Russian], Izd.AN SSSR, 5, 119, 1959. 78. K. V. Popov and N. A. Khvorostukhina, Izv, SO A N SSSR, no~ 8, 89, 1988. 74. V. Ya. Dubovoi, Floccules in Steel [in Russian], Metallurgizdat, Moscow, 1950. 75. I. H. Andrew, H. Lee, H. K. Lloyd, and N. Stephenson, J, Iron and S~eel Inst., 156, no. 2, 208, 1947. 76. N. J. Grant and J. L. Lunsford, Iron Age, !75, no. 22, 192, 1955. 77. E. R. Slaughter, J. Metals, 8, no. 4, 430, 1956. 78. M. I. Zamotorin and T. S, Kosovtseva, in Metallurgy in USSR [in Russian], Izd. A N SSSR, 1957, 79. H. G. Vaughan and M. E. Morton, Nature, 177, no. 4522, 1225, 1956. 80. P. I. Strel'nikov, A. I. Fedorenko arld •. P. Klyucharev, in: Influence of Nuclear Radiation on Materials [in Russian], Izd. A N SSSR, 1962. 81. M. M. Shved and F. P. Yanchishln, in; Effeqt of Working Media on the Properties of Steel [in Russianl Izd. A N USSR, Kiev, vol. i, 1961. 82. V. V. Kuznetsov, N. I. Konstantinova, and V. A. Frolov, FMM, 12, p.O. 2, 1961. 83. G. V. Karpenko and A. K. Litvin, in: Effect of Working Media on the Properties of Steel [in Russian]/ Izd. A N USSR, Kiev, vol. i, 1961. 84. S. M. Beloglazov, FMM, 15, no. 6, 1963. 85, 1. D. Hobson and C. Sykes, J. Iron and Steel Inst., 169, no. 3, 209, 1951. 86. L. S. Moroz and T. E. Mingin, in: Metallorgraphy [in Russian], Sudpromgi;% vol. 2, 1958. 87. E. Herzog, Teehn. et applic, petrole0 18, no. 147, 1958. 88. A. Siede and W. Rostoker, Trans. Met. Soc. AIME, 212, no. 6, 852, 1958 (1959). 89. H. C. Rogers, Acta met., 4, no. 3, 114, 1956. 90. H. C. Rogers, Trans. Met. Soc. A!ME. 915, no. 4, 666, 1959 (!960), 91. A. Cracknell and N. I. Petch, Acta met., 8, no. 2, 200, 1955. 92. H. Schumann, Metallurgie und Gisser~itechnik, 4, no. 9, 385, 1954. 9S. A. M. Adair and R. E. Hook, Acta met., 10, no. 8, 741, 1962. 94. A. Lawley, W. Liebmann, and R. Mad~lin, Acta met., 9, no. 9, 841, 1.961, 95. Murayama and Hamahasi, J. Iron and steel Inst., Japan, 41, no. 3, 292, 1955.
472
96. 97. 98. 99. i00. I01. 102. 103. 104. 105. 106. 107. 108. 109. ii0. iii. 112. 113. 114. 115. 116. 117.
K. V. Popov and V. A. Yagunova, FMM, 19, no. 2, 1965. C. A. Zapffe and C. Sims, Trans. AIME, 145, 225, 1941. C. A. Zapffe, trans. ASM, 39, 190, 241, 1947. K. V. Popov and V. A. Yagunova, FMM, 8, no. 2, 1959. P. Bastien and P. Azou, Proc. I World Metal1 Congress, Amer. Soc. Met., 535, 1951. P, Bastien, Bergakademie, 11, no. 2, 70, 1959. E. Orovan, Fatigue and Fracture of Metals, N. Y. 135, 1952. A. N. Stroh, Advances Phis., 6, no. 24, 418, 1957. N. J. Perch and P. Stables, Nature, 169, 842, 1982. F. de Kazinczy, J. Iron and Steel Inst., 177, 1, 85, 1954. F. de Kazinczy, Jernkontorets ann., 139, no. 11, 885, 1955. Ya. M. Potak, Brittle Fracture of Steel and Steel Parts [in Russian], Oborongiz, Moscow, 1955. F. Carafoto, G. T. Chou, and V. Ambegaokar, Acta met., 8, no. 8, 504, 1960. B. A. Bilby and J. Hewitt, Acta met., 10, no. 6, 587, 1962. I. J. Morlet, H. H. Johnson, and A. R. Troiano, J. Iron and Steel Inst., 189, p. 1, 37, 1958. p. Blanchard and A. R. Troiano, Mere. Sci. Rev. Metallurg., 57, no. 6, 409, 1960. A. H. Cottrell, Trans. Met. Soe. AIME, 212, no. 4, 192, 1958. G. Phragmen, Jernkontorets ann, 128, 537, 1944. P. Emmet and R. Harckness, J. Amer. Chem. Soc., 57, 1631, 1935. V. G. Savitskii and K. V. Popov, Izv. SO AN SSSR, no. 2, 15, 1960. V. G. Savitskii and K. V. Popov, Izv. SO AN SSSR, no. 8, 138, 1960. V. G. Savitskii and K. V. Popov, in: Investigations of Heat-Resistant Alloys [in Russian], Izv. AN SSSR, vol. 9,
1962.
A. H. Cottrell and B. A. Bilby, Proc. Phys. Soc., 62, 1, 349A, 49, 1949. T. Yokobori, Phys. Rev., 88, Ser. 2, 1423, 1952. I. D. Campbell, Acta met., 1, 706, 1953. J. D. Baird, J. Iron and Steel Inst., 86, 10, 450, 1963. 122. C. Crussard, J. Plateau, R, Tamhankar, G. Henry, and D. Lajeunesse, in: Atomic Mechanism of Fracture [Russian translation], Metallurgizdat, Moscow, 1963. 123. D. McLean, Mechanical Properties of Metals [Russian translation], Metallurgizdat, Moscow, 1965. 124. A. H. Cottrell, Report Conf. Strength of Solids, London, Physical Society, 30, 1948. 125. A. I I. ~ottrell, Phil. Mag., 44, 829, 1953. 126. Ao H. Cottrell, Dislocations and Plastic Flow in Crystals [Russian translation], Moscow, 1958. 127. Ao H. Cottrell, in: Structure and Properties of Metals [Russian translation], Moscow, 1957. 128. R. Tamhankar, Trans. of Indian Inst. of Metals, 11, 37, 1958. 129. E. A. Steigerwald and G. L. Hanna, Trans. Amer. Soc. of Metals, 56, 656, 1963. 130. V. A. Yagunova and K. V. Popov, in: Investigationsof Heat-Resistant Alloys [in Russian], Izd. AN SSSR, Moscow, vol. 8, 1962. 131. G. V. Samsonov and A. P. Epfk, in: Investigations of Steels and Alloys [in Russian], Izd. "Nauka," Moscow, 1964. 132. P. L. Gruzin, Yu. A. Polikarpov, and G. B. Fedor0v, FMM, 4, no. 1, 1957. 133. E. P. Nechai and K. V. Popov, FMM, 19, no. 4, 1965. 134. V. ZeAt, Diffusion in Metals [Russian translation], IL, Moscow, 1958. 135. M. A. Krishtal, Diffusion Phenomena in Iron Alloys [in Russianl Metallurgizdat, Moscow, 1963. 136. J. W. Edington, T. C. Lindley~ and R. E. Smillman, Acta met., 12, 9, 1025, 1964. 13% R. P. Carrekerand R. W. Guard, Trans. AIME, 206, 178, 1956. 138. J. H. Bechtold, Trans. AIME, 197, 1469, 1953. 139. V. F. Sukhovarov and R. P. Kharlova, FMM, 10, no. 6, 1960. 140. G. M. Grigor'eva, O. G. Mamneva, Eo P. Nechai, K. P. Popov, and E. A. Chipcheeva, FKhMM [Soviet Materials Science], no. 3, 1965. 141. V. F. Sukhovarov, FMM, 13, no. 2, 1962. 142. V. F. Sukhovarov, N. A. Aleksandrov, and L. A. Kudryavtseva, FMM, 14, no. 6, 1962. 143. K. V. Popov and E. P. Nechai, FMM, 21, no. 1, 1965. 144. B. A. Wilcox, A. W. Brisbane, and R. F. Klinger, Trans. Amer. Soc. Metals, 55, 1, 179, 1962. 145. Ya. I. Frenkel', Introduction to the Theory of Metals [in Russian], GTII, Moscow-Leningrad, 1950. 146. R. M. Berrer, Tram, Farad. Sot., 36, no. 12, 12350 1940. 147. V. A. Yagunova, K. V. Popov, and K. P. Zhdanova, in: Investigations of Heat-Resistant Alloys [in Russian], Izd. AN SSSR, 1960. 25 January 1966 Institute of Hydrodynamics at the Siberian Filial of AS USSR, Novosibirsk 118, 119. 120. 121.
473