Dynamic strain ageing in Inconel® Alloy 783 under tension and low cycle fatigue

Materials Science and Engineering A 546 (2012) 34–39

Contents lists available at SciVerse ScienceDirect

Materials Science and Engineering A journal homepage: www.elsevier.com/locate/msea

Dynamic strain ageing in Inconel® Alloy 783 under tension and low cycle fatigue A. Nagesha a,∗ , Sunil Goyal a , M. Nandagopal a , P. Parameswaran b , R. Sandhya a , M.D. Mathew a , Sarwan K. Mannan c a b c

Mechanical Metallurgy Division, Indira Gandhi Centre for Atomic Research, Kalpakkam, India Physical Metallurgy Division, Indira Gandhi Centre for Atomic Research, Kalpakkam, India Special Metals Corporation, Huntington, WV 25705, USA

a r t i c l e

i n f o

Article history: Received 8 August 2011 Received in revised form 21 February 2012 Accepted 7 March 2012 Available online 21 March 2012 Keywords: Inconel 783 Low cycle fatigue Dynamic strain ageing Cyclic softening Friction stress

a b s t r a c t Low cycle fatigue (LCF) tests were performed on Inconel® 1 Alloy 783 at a strain rate of 3 × 10−3 s−1 and a strain amplitude of ±0.6%, employing various temperatures in the range 300–923 K. A continuous reduction in the LCF life was observed with increase in the test temperature. The material generally showed a stable stress response followed by a region of continuous softening up to failure. However, in the temperature range of 573–723 K, the alloy was seen to exhibit dynamic strain ageing (DSA) which was observed to reduce the extent of cyclic softening. With a view to identifying the operative mechanisms responsible for DSA, tensile tests were conducted at temperatures in the range, 473–798 K with strain rates varying from 3 × 10−5 s−1 to 3 × 10−3 s−1 . Interaction of dislocations with interstitial (C) and substitutional (Cr) atoms respectively, in the lower and higher temperature regimes was found to be responsible for DSA. Further, the friction stress, as determined using the stabilised stress–strain hysteresis loops, was seen to show a more prominent peak in the DSA range, compared to the maximum tensile stress. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Inconel® Alloy 783 is a new class of Ni–Co based superalloys with a low coefficient of thermal expansion (CTE) [1–4]. The alloy was developed with an objective of achieving a low CTE while maintaining a high strength and a good corrosion resistance so that it can be used in aircraft gas turbine components such as rings, casings and shrouds that can maintain tightened blade tip clearances at different operating temperatures, leading to an increased turbine efficiency [5,6]. The lowering of CTE is attributed to a reduced Cr content in the alloy which ensures that the Curie temperature stays high enough thereby avoiding the lattice thermal expansion associated with magnetic transformation [5]. Besides gas turbine engine parts, the low CTE superalloys have potential usage in the superconductive magnets for fusion reactors. The alloy contains a higher Al content of about 5% compared to the more conventional and better known variants of the above class of superalloys such as Incoloy 903, 907, 908 and 909 [5]. The increased Al addition facilitates the formation of incoherent NiAl-type ␤-phase in an austenitic matrix in addition to the coherent Ni3 Al-type ␥ precipitates. The ␤-phase that occurs discontinuously at both inter- and intra-granular locations, can be processed to resist stress accelerated grain boundary oxidation (SAGBO), a particular form of stress corrosion cracking to

which these class of superalloys are usually susceptible [1–4,6]. The alloy also possesses an improved resistance against salt spray corrosion compared to other low CTE alloys, which makes it a suitable candidate for offshore turbine applications as well [7]. While the austenitic matrix accounts for a good combination of mechanical properties and workability, the ␥ imparts precipitation strengthening, to the alloy. Components, such as turbine blades and vanes in gas turbine and aero engines are subjected to repeated cyclic thermal stresses resulting from startup and shutdown operations. The high rotational speeds and the associated centrifugal stresses experienced by the turbine discs call for good elevated temperature tensile strength and toughness. Furthermore, stresses that are associated with changes in the rotational speed of the turbine and the stresses that are of thermal origin are cyclic in nature, underlining the need for such alloys to possess a good resistance to low cycle fatigue (LCF) [8]. Though the fatigue crack growth behaviour of the alloy under different microstructural conditions has been studied in detail [2,6], literature on LCF is limited [9]. The present investigation is aimed at investigating the influence of DSA under LCF and tensile deformation at different temperature–strain rate combinations. 2. Experimental 2.1. Material and heat treatment

1

Inconel is the registered trademark of Special Metals Corporation, USA. ∗ Corresponding author. Fax: +91 44 27480075. E-mail address: [email protected] (A. Nagesha). 0921-5093/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2012.03.018

Detailed chemical composition of the alloy in wt.% is provided in Table 1. The alloy, received in hot worked condition, was machined

A. Nagesha et al. / Materials Science and Engineering A 546 (2012) 34–39

35

Table 1 Chemical composition (wt.%) of Inconel® Alloy 783. Ni

Co

Fe

Al

Cr

Nb

Ti

Mo

Si

Mn

Cu

C

B

P

S

28.11

34.75

25.05

5.37

3.21

3.00

0.20

0.14

0.08

0.04

0.03

0.01

0.005

0.004

<0.001

into bars, which were subsequently given a three-step heat treatment as detailed below: 1. Annealing at 1393 K for 1 h and air-cool. 2. Ageing at 1116 K for 4 h and air-cool (referred to as ␤-ageing). 3. Ageing at 991 K for 8 h and furnace cooling at 328 K per hour to 894 K, held for 8 h at this temperature followed by air cooling to room temperature (␥ -ageing). The ␤-phase in this alloy has been reported to contain approximately 31% Al, 30% Ni, 24% Co and 15% Fe (atom%) [10]. The ␥ precipitates formed in step-3 above have an average size of 30–40 nm [2]. 2.2. Tensile and low cycle fatigue tests Tensile tests were performed on cylindrical samples with a gauge length of 28.6 mm and a gauge diameter of 4 mm (Fig. 1a). Tests were conducted at different temperatures in the range, 473–798 K at strain rates varying from 3 × 10−5 s−1 to 3 × 10−3 s−1 . Smooth specimens of 25 mm gauge length and 10 mm gauge diameter were used for carrying out total axial strain controlled LCF tests in air. Fig. 1(b) presents the specimen geometry employed for the LCF tests. All tests were performed at a constant strain rate of 3 × 10−3 s−1 and a strain amplitude of ±0.6% in the temperature range, 300–923 K, using an Instron 1342 servohydraulic machine, equipped with a radiant heating furnace. The failure life was taken as the cycle number corresponding to a 20% drop in the peak tensile stress at half-life in LCF tests.

Fig. 1. (a) Specimen geometry used for tensile tests. (b) Specimen geometry used for LCF tests.

70 ␮m, followed by electropolishing using a solution containing 10% perchloric acid and 90% methanol at 243 K using a d.c. supply of 20 V. 3. Results and discussion 3.1. Prior microstructure

2.3. Metallography Optical metallography was carried out on samples swab etched for about 30 s using Kalling’s reagent (6 g CuCl2 , 100 ml HCl, 100 ml H2 O and 100 ml CH3 OH). Samples for transmission electron microscopy were initially mechanically polished down to about

The prior microstructure of the alloy subjected to the heat treatment detailed earlier consisted of three phases, an austenitic matrix, intra- and intergranularly distributed NiAl ␤-phase and a fine distribution of ␥ precipitates. Fig. 2 presents the typical microstructure of the untested material that shows the ␥

Fig. 2. Prior microstructure of the alloy in the heat-treated condition showing ␥ .

36

A. Nagesha et al. / Materials Science and Engineering A 546 (2012) 34–39

Fig. 3. Influence of temperature on the cyclic stress response [9].

precipitates distributed uniformly in an austenitic matrix. The ␤phase occurs as intra- or intergranular precipitates either in plate or globular form, with the precipitation temperature of the phase ranging from 973 to 1416 K [5].

Fig. 4. Gamma prime precipitation seen within the ␤-phase after LCF testing (±0.6%, 3 × 10−3 s−1 , 923 K).

3.2. Cyclic stress response The nature of cyclic stress response is closely related to microstructural changes. The cyclic stress response obtained in LCF tests conducted at different temperatures in the range, 300–923 K is presented in Fig. 3. As can be seen in the figure, after passing through a regime of stable stress response during the first 20–30 cycles, the alloy underwent continuous softening up to the final drop leading to failure. However, the softening was not significant in the temperature range of 573–723 K (Fig. 3). The cyclic softening in ␥ strengthened nickel base superalloys during LCF testing at room temperature has been attributed to a variety of factors such as, (a) scrambling or disordering of the ordered ␥ precipitates by the moving dislocations [11], (b) reduction in the size of the ␥ precipitates due to shearing during repeated cyclic deformation [12,13], and (c) their eventual dissolution [14]. On the other hand, the softening observed at the higher temperatures has been reported to be associated with the loss of coherency of ␥ following their coarsening [15]. However, the ␥ precipitates were generally known to be resistant against coarsening except at very high temperatures. Hwang et al. [16] have attributed the significant softening at a temperature of 1255 K to the coarsening of ␥ . Extensive cyclic softening at elevated temperatures in alloy 800H has been ascribed to recovery processes such as dislocation annihilation and thermally activated climb processes [17]. The loss of coherency of ␥ particles, aided by shearing and dissolution has been thought to lead to cyclic softening during LCF cycling of a ␥ strengthened superalloy at temperatures of 1033 and 1144 K [16]. In the present case, extensive cyclic softening has been observed even at room temperature (Fig. 3) and it would be reasonable to attribute this to the continuous shearing of ␥ by dislocations during successive cycles in the course of fatigue deformation. The ␤-phase present in the alloy is essentially devoid of ␥ in untested/unaged condition [5]. However, fatigue cycling at a temperature of 923 K was seen to promote precipitation of ␥ within the ␤-phase as shown in Fig. 4· Precipitation of ␥ within ␤-phase has also been reported upon elevated temperature exposure of the alloy [18]. Failure was generally seen to occur by the propagation of a single major crack to failure since secondary cracks were rarely seen. Fractography revealed that the fracture was largely transgranular at

temperatures up to 823 K, with some intergranularity seen in specimens tested at 923 K [9]. 3.3. Role of temperature and dynamic strain ageing Tensile tests showed that the alloy exhibits DSA over the temperature range, 573–748 K, as reflected by periodic serrations in the stress–strain curves. Fig. 5 presents segments of true stress–true strain plots obtained at different temperatures at a strain rate of 3 × 10−5 s−1 . The strain rate sensitivity (m) of flow stress  at different temperatures and strains was evaluated using the –ε plots using the expression,



m=

log(2 /1 )  log(ε˙ 2 /ε˙ 1 ) 

(1) ε,T

where  1 and  2 are the flow stresses at the strain rates, ε˙ 1 and ε˙ 2 respectively. Values of m were determined at different strain

Fig. 5. Segments of true stress–true strain plots at different temperatures, at a strain rate of 3 × 10−5 s−1 .

A. Nagesha et al. / Materials Science and Engineering A 546 (2012) 34–39

37

Fig. 6. Variation of strain rate sensitivity with temperature.

levels, using ε˙ 1 = 3 × 10−4 s−1 and ε˙ 2 = 3 × 10−3 s−1 . Fig. 6 presents the variation of the same (at ε = 5%) with temperature. As can be seen, m showed a negative value in the temperature range, 573–748 K, though it assumes a slightly positive value in the intermediate temperature of 673 K. The occurrence of DSA was accompanied by a negative strain rate sensitivity in the temperature range 573–748 K, as shown in Fig. 6. The occurrence of PLC effect has been extensively studied in several alloy systems. The influence of DSA was also seen to have a bearing on the nature of cyclic stress response; the cyclic softening that was prominently noticed in tests at room temperature and 923 K was almost non-existent in the intermediate temperature range, 573–723 K as shown in Fig. 3. In the above temperature range, the additional hardening mechanism was seen to counteract the cyclic softening seen in other temperatures. The interaction between the mobile dislocations and the solute atmosphere restricts the cross slip of dislocations, causing slip planarity in the temperature range where DSA is active. Consequently, the material gets hardened giving raise to an increase in the flow stress needed to impose the same total strain. As a result, a decrease in the magnitude of cyclic softening was observed. It may however be noted that the hysteresis loops did not exhibit serrations at any temperature. This could be attributed to the fact that the critical strain for the onset of serrated flow (>1%) remained higher compared to the strain amplitudes used for LCF tests (±0.6%).

Fig. 7. Variation of half-life, friction and back stresses with temperature.

right from the start of deformation and the associated load drops occurring below the general level of the stress–strain curves are a result of breakaway of the aged dislocations [22]. As shown in Fig. 5, the C-type serrations seen in the temperature range of 673–748 K required a high critical strain and were seen to continue almost till failure. The transition from A + B to C type serrations was seen to occur at the temperature of 673 K at all the strain rates investigated.

3.5. Analysis of friction and back stresses In order to investigate the role of DSA on the cyclic stress–strain behaviour, the friction (thermal) and back (athermal) stresses were evaluated from the stress–strain hysteresis loops according to the scheme proposed by Cottrell [23]. Following expressions were used to derive the values of the friction stress,  F and the back stress,  B : F =

Characteristics of the serrations differed in the temperature ranges, 573–623 K and 673–748 K. They were identified as mixtures of types A and B in the temperature range, 573–623 K whilst C-type serrations were dominant in the temperature range, 673–748 K (Fig. 5). Type A serrations, termed as ‘locking’ serrations, are associated with repetitive nucleation of deformation bands from one end and their continuous propagation to the other end of the gauge length, with each load drop corresponding to the nucleation of a new band [19]. Such serrations are believed to be a consequence of cyclic locking–unlocking of the leading dislocations due to segregation of solute atmospheres [20]. They occur in the low temperature or high strain rate part of DSA regime. Type B serrations, which are also considered locking serrations, are a result of discontinuous band propagation arising from moving dislocations within the band [21]. They are characterised by oscillations about the general level of the stress–strain curves. The C-type serrations, on the other hand, are known as ‘unlocking’ serrations and occur in the higher temperature end of the regime of DSA. The diffusion rates in this region are believed to be high enough for the dislocations to be aged

(2)

E − s 2

(3)

and B =

3.4. Type of serrations

E + s 2

where  E and  S are the peak tensile stress and yield stress, respectively. It was noticed that the friction stress displayed a welldefined peak in the temperature interval, 573–723 K. The peak was found to be significantly more pronounced in comparison to that observed in half-life stress in the above temperature regime, as is evident in Fig. 7. A close similarity between the variation in the friction stress and half-life stress (obtained from the half-life hysteresis loops under LCF tests) as a function of temperature can be observed (Fig. 7). Therefore it can be presumed that the stress amplitude attained during cyclic deformation is controlled by the friction stress generated as a result of an interaction between dislocations and solute atmospheres. It may be pointed out that the friction stress has a significant contribution from point defects [24] and that an increase in the former with increasing temperature in the domain where DSA is operative indicates an increase in the effective locking of mobile dislocations. Hong and Laird [25] have suggested that the elastic interaction of dislocations and segregated solute atoms (such as Al or Zn) is the dominant mechanism for increase in the value of frictional stress in the DSA regime in Cu-alloys.

38

A. Nagesha et al. / Materials Science and Engineering A 546 (2012) 34–39

Fig. 8. Influence of temperature on the ultimate tensile strength at different strain rates.

3.6. Activation energy for serrated flow It may be recalled that a change in the UTS, rather than the yield or the flow stress with temperature, is a more appropriate criterion for identifying DSA [26]. The ultimate tensile strength (UTS) vs. temperature plot in Fig. 8 presents a dual peak behaviour corresponding to different temperature regimes. The dual peak behaviour seen at the strain rates of 3 × 10−4 s−1 and 3 × 10−5 s−1 is suggestive of the involvement of more than one mechanism for the occurrence of DSA. Investigations pertaining to DSA in Ni and Ni-base alloys have generally attributed the phenomenon to interaction of interstitial solutes with the moving dislocations at lower temperatures [27] and to the diffusion and interaction of substitutional solutes with dislocations at higher temperatures [28]. The solute element responsible for dislocation locking in the DSA regime is usually identified by computing the activation energy for serrated flow evaluated from the slopes of εC vs. ε˙ and εC vs. 1/T plots using the relationship, εc m+ˇ = K ε˙ exp

Q  RT

(4)

where Q is the activation energy for the onset of serrated flow, R is the gas constant, T is the temperature and K is a constant. The critical plastic strain εC was evaluated as the minimum true plastic strain at which a perceptible load drop of 5 N occurs [29]. The value of Q so obtained is compared with the activation energy for migration of solutes in the matrix. The exponent m + ˇ is related to the vacancy concentration Cv and the mobile dislocation density m as, Cv ∝ εm and m ∝ εˇ . The value of Q, as determined from the slope of the plot of ln ε˙ vs. ln εC (Fig. 9a), was seen to lie in the range, 59–74 kJ mol−1 , which is comparable to the activation energy of 69 kJ mol−1 , obtained for pipe diffusion of C in Ni [30]. In the temperature range of 673–748 K, an inverse PLE was observed, coinciding with the occurrence of type-C serrations. Such a behaviour, characterised by a negative dependence of the critical strain for the onset of serrations with strain rate and a positive dependence of the same with temperature, has been reported in several Ni-based alloys [19,28,31–38]. The inverse PLE is also evident in the ln ε˙ vs. ln εC plot presented in Fig. 9(a) and the ln εC vs. 1000/T plots as shown in Fig. 9(b). It should be noted that Eq. (4) is valid only in the regime of normal PLC. Hence for the inverse PLE observed in the range, 673–748 K, the activation energy was determined using the stress drop method [35,39] wherein the heights

Fig. 9. (a) Strain rate vs. critical strain plots at different temperatures. (b) Critical strain vs. 1000/T plots, Inconel Alloy 783, strain rate: 3 × 10−5 s−1 .

of individual serrations,  are used to arrive at the value of Q, as follows: Q = −R

  ln ε˙  1/T

(5) ,ε

The average stress drop ( avg ) of the C-type serrations were determined and were plotted against strain rate, using the stress–strain curves obtained at 673, 723 and 748 K (Fig. 10a). The variation of  avg with strain rate was found to obey the following power functions: avg = 0.1422ε˙ −0.4631

(6)

avg = 0.2516ε˙ −0.5204

(7)

avg = 0.5187ε˙ −0.4945

(8)

The above equations (Eqs. (6)–(8)), which correspond to temperatures, 673, 723 and 748 K respectively, were used to establish the ε˙ vs. 1/T plots for different values of  avg . The value of Q was subsequently calculated using the slopes of ln ε˙ vs. 1000/T plots, as shown in Fig. 10(b). As seen, the above method yielded a Q value of 171–184 kJ mol−1 . This suggests a change in the mechanism responsible for the occurrence of serrated flow with an increase in the test temperature. It is well known that diffusion of substitutional solute atoms through the dislocation core can take place more rapidly with a lower activation energy of about 0.4–0.7 times that required for the bulk diffusion [40]. The value of Q for DSA obtained in the temperature range of 673–748 K is about 0.5 times that reported for the bulk diffusion of Cr in austenite (Q␥ Cr = 334.4 kJ mol−1 ) [41]. It is therefore reasonable to conclude that the serrated flow in the above temperature range was associated with the diffusion of Cr atoms and their interaction with dislocations. The average stress decrement, as can be observed

A. Nagesha et al. / Materials Science and Engineering A 546 (2012) 34–39

39

Acknowledgements The authors wish to record their sincere gratitude to Sri. P.K. Chaurasia for the help rendered during the investigation. Encouragement received from Dr. T. Jayakumar and Dr. A.K. Bhaduri is gratefully acknowledged. References

Fig. 10. (a)Variation of the average stress drop with strain rate at different temperatures. (b) Variation of ln(strain rate) vs. 1000/T at different values of .

from Fig. 10(a), was seen to increase with an increase in the test temperature which substantiates the conclusion that the serrations are a result of the dislocations tearing away from the solute atmospheres rather than their pileup at obstacles [24]. It is therefore concluded that the serrated flow observed in the temperature range, 673–748 K is associated with the diffusion of chromium. 4. Conclusions 1. Inconel® Alloy 783 was seen to exhibit DSA in the temperature range, 573–723 K under LCF loading. The extent of cyclic softening was found to be minimised in the above temperature regime. 2. Friction stress was found to be influenced by DSA to a greater extent than the half-life tensile stress. 3. Under tensile deformation, DSA was observed in the temperature regime, 573–748 K. 4. Occurrence of DSA in the alloy was a result of an interaction of interstitial (C) and substitutional (Cr) atoms in the lower and higher temperature regimes respectively.

[1] K.A. Heck, D.F. Smith, M.A. Holderby, J.S. Smith, in: S.D. Antolovich, R.W. Stusrud, R.A. MacKay, D.L. Anton, T. Khan, R.D. Kissinger, D.L. Klarstrom (Eds.), Superalloy, The Minerals, Metals and Materials Society, 1992, pp. 217–226. [2] L. Ma, K.M. Chang, S.K. Mannan, S.J. Patel, Scripta Mater. 48 (5) (2003) 551–557. [3] L. Ma, Micron 35 (4) (2004) 273–279. [4] L. Ma, K.M. Chang, S.K. Mannan, Scripta Mater. 48 (5) (2003) 583–588. [5] L. Ma, Effect of Ni–Al-β precipitates on crack growth behaviour of Inconel Alloy 783, Ph.D. Dissertation, West Virginia University, 2001. [6] X. Kang, C. Liu, K.M. Chang Cisloiu, Mater. Sci. Eng. A347 (2003) 205–213. [7] E.A. Ott, J.R. Groh, S.K. Mannan, in: K.A. Green, T.M. Pollock, H. Harada, T.E. Howson, R.C. Reed, J.J. Schirra (Eds.), Superalloy, The Minerals, Metals and Materials Society, 2004, pp. 643–652. [8] K. Gopinath, A.K. Gogia, S.V. Kamat, R. Balamuralikrishnan, U. Ramamurty, Acta Mater. 57 (2009) 3450–3459. [9] A. Nagesha, M. Sunil Goyal, K. Valsan, Bhanu Sankara Rao, S.K. Mannan, Trans. Indian Inst. Met. 63 (2-3) (2010) 575–579. [10] S.K. Mannan, J.J. deBarbadillo, International Gas Turbine and Aeroengine Congress and Exhibition, Stockholm, June 2–5, 1998. [11] R.E. Stoltz, A.G. Pineau, Mater. Sci. Eng. 34 (1978) 275–284. [12] B.A. Lerch, V. Gerold, Acta Metall. 33 (1985) 1709–1716. [13] M. Valsan, P. Parameswaran, K. Bhanu Sankara Rao, M. Vijayalakshmi, S.L. Mannan, D.H. Sastry, Metall. Trans. 23A (1992) 1751–1761. [14] M. Sundararaman, W. Chen, V. Singh, R.P. Wahi, Acta Metall. 38 (1990) 1813–1822. [15] S.D. Antolovich, S. Liu, R. Baur, Metall. Trans. 12A (1981) 473–481. [16] S.K. Hwang, H.N. Lee, B.H. Yoon, Metall. Trans. 20A (1989) 2793–2801. [17] K. Bhanu Sankara Rao, H. Schiffers, H. Schuster, G.R. Halford, Metall. Mater. Trans. 27A (1996) 255–267. [18] S.K. Mannan, G.D. Smith, S.J. Patel, in: K.A. Green, T.M. Pollock, H. Harada, T.E. Howson, R.C. Reed, J.J. Schirra (Eds.), Superalloy, The Minerals, Metals and Materials Society, 2004, pp. 627–635. [19] K. Gopinath, A.K. Gogia, S.V. Kamat, U. Ramamurty, Acta Mater. 57 (2009) 1243–1253. [20] I. Nikulin, R. Kaibyshev, Mater. Sci. Eng. A528 (2011) 1340–1347. [21] P. Rodriguez, Bull. Mater. Sci. 6 (1984) 653–663. [22] W.C. Charnock, Philos. Mag. 20 (1969) 427–432. [23] A.H. Cottrell, Dislocations and Plastic Flow in Crystals, Oxford University Press, London, 1953. [24] C. Laird, in: F.R.N. Nabarro (Ed.), Dislocations in Solids, North Holland Pub. Co., 1983, pp. 111–132. [25] S.I. Hong, C. Laird, Acta Mater. 38 (8) (1990) 1581–1594. [26] S. Lou, D.O. Northwood, Mater. Forum. 17 (2) (1993) 153–158. [27] R.S. Moshtaghin, S. Asgari, Mater. Sci. Eng. A486 (2008) 376–380. [28] C.L. Hale, W.S. Rollings, M.L. Weaver, Mater. Sci. Eng. A300 (2001) 153–164. [29] S. Venkadesan, C. Phaniraj, P.V. Sivaprasad, P. Rodriguez, Acta Mater. 40 (1992) 569–580. [30] W.F. Gale, T.C. Totemeier, C.J. Smithells (Eds.), Metals Reference Book, sixth ed., Plenum Press, New York, 1983. [31] W. Chen, M.C. Chaturvedi, Mater. Sci. Eng. A229 (1997) 163–168. [32] Y. Nakada, A.S. Keh, Acta Metall. 18 (1970) 437–443. [33] J.S. Blakemore, Metall. Trans. 1 (1970) 1281–1285. [34] R.A. Mulford, U.F. Kocks, Acta Metall. 27 (1979) 1125–1134. [35] R.H. Hayes, W.C. Hayes, Acta Metall. 30 (1982) 1295–1301. [36] R.W. Hayes, Acta Metall. 31 (1983) 365–371. [37] M. Vani Shankar, K. Valsan, Bhanu Sankara Rao, S.L. Mannan, Metall. Mater. Trans. 35A (2004) 3129–3139. [38] K. Bhanu Sankara Rao, V. Seetharaman, S.L. Mannan, P. Rodriguez, High Temp. Mater. Proc. 7 (1) (1986) 63–81. [39] E. Pink, A. Grinberg, Acta Metall. 30 (1982) 2153–2160. [40] R.W. Balluffi, Phys. Status Solidi 42 (1970) 11–34. [41] W. Fuxing, Metal Physics, Mechanical Industry Press, Beijing, 1981, p. 160.