Cross-sectional transmission electron microscopy characterisation of plasma immersion ion implanted austenitic stainless steel

Cross-sectional transmission electron microscopy characterisation of plasma immersion ion implanted austenitic stainless steel

ELSEVIER Surfaceand CoatingsTechnology85 (1996)28-36 Cross-sectional transmission [email protected]&rn~microscopy characterisation of plasma immersion ion implant...

1MB Sizes 0 Downloads 14 Views

Recommend Documents

No documents

Surfaceand CoatingsTechnology85 (1996)28-36

Cross-sectional transmission [email protected]&rn~microscopy characterisation of plasma immersion ion implanted austenitic stainless steel X. Li a, M. Samandi a, D. Dune a, G. Collins b, J. Tendys b, K. Short bYR. Hutchings b a Department of Materials Engineering,Universityof Wollongong,LockedBag 8844,SouthCoastMail Centre, Wollongong, NSW2521,Australia bAustralianNuclearScienceand TechnologyOrganisation,Private Mail Bag I, Mend, NSW2234,Amtrdia

Abstract Cross-sectional transmission electron microscopy (XTEM), selected area diffraction (SAD) and nano-beam diffraction (NBD) techniques were used to investigate the surface microstructure of 316 stainless steel, implanted with high doses of nitrogen ions at 150, 250, 350, 450 and 520 “C using plasma immersion ion implantation. It has been found that the treatment temperature has a strong influence on the evolution of the microstructure. An amorphous layer of about 1 ,um thick with a heavily stressedsubstrate underneath wasobserved on the 150 “C implanted sample. Both the 250 and 350 “C implanted samples showed a thin nanocrystalline sublayer at the outermost surface and an amorphous sublayer between the nanocrystalline sublayer and the substrate. A thick amorphous layer up to 3 pm thick was formed on the 450 “C implanted sample whereasat 520 “C, cellular precipitation of CrN and a-ferrite dominated the system. It is suggested that a solid state chemical reaction and the poor mobility of the reactant atoms are the key factors for the solid state amorphisation by nitrogen ion implantation into austenite. Keywords:

Ion implantation; Stainless steel; Amorphous; P13; XTEM

1. Introduction The corrosion performance as well as the wear resistance of austenitic stainless steels can be significantly improved by nitrogen ion implantation at appropriate temperatures [l-6]. However, despite many investigations, the exact nature of the microstructural changes responsible for these improvements is not very well understood. In fact, the literature on this subject appears to be contradictory and fragmented. The main problem appears to be the limited knowledge from the traditional surface microstructure characterisation techniques, such as XRD (X-ray diffraction), GAXD (glancing angle X-ray diffraction), TEM + SAD (transmission electron microscopy + selected area diffraction) etc., which are not able to characterise such a heterogeneous thin layer in which different nanocrystalline precipitates with similar crystal structures and d-spacings are formed. In the present work, cross-sectional transmission electron microscopy (XTEM), SAD and nano-beam

diffraction (NBD) techniques were used to investigate the surface microstructure

of 316 stainless


implanted with high doses of nitrogen ions at a range of temperatures using plasma immersion ion implantation (P13). Preliminary results of these investigations 0257-8972/96/$15,00 0 1996ElsevierScienceS.A. All rights reserved PII


have appeared elsewhere [5-71 where we noted the formation of a thick (2-3 1u-n)amorphous zone which may contain nano-crystalline precipitates of CrN and CIferrite. Only partial reconciliation of the TEM results with those obtained by other techniquessuch as GAXD has been possible [6]. In this article, we present and discussin detail the microstructures obtained from PI3 treated samples at 150, 250, 350, 450 and 520 “C. In particular, we show how the treatment temperature has a strong influence on the evolution of the microstructure.

2. Experimental


A 25 mm diameter bar of AISI 316 stainless steel supplied by ASSAB Steel was used to produce test pieces.Chemical analysisby spark emissionspectroscopy indicated that the steelwas within the AISI specification for 316 grade (Table 1). Several discs approximately 3 mm in thickness were cut from the bar. The samples were polished to a 1 pm diamond finish, prior to treatment. PI3 treatment was conducted at ANSTO using the developmental equipment. Sampleswere treated at 150,

X Li et al./Surface and Coatings Technology 85 (1996) 28-36 Table 1 Chemical composition of untreated AISI 316 C




0.043 16.40 10.56 2.34







0.012 0.24





Table 2 PI3 treatment details Temperature (“Cl

Nitrogen dose ( x lOi atoms/cm’)

Time (minute)

150 250 350 450 520

2 8 8 8 8

430 250 130 70 40

250, 350, 450 and 520 “C respectively. The details of the treatment parameters are given in Table 2. A detailed description of the XTEM sample preparation has been published elsewhere [7]. First, the PI3 treated discs were carefully thinned from the untreated side to a thickness of lmm, with the modified surface carefully protected. Then the thinned discs were sliced to give 2 mm wide slabs. These slabs were subsequently sandwiched using G-l epoxy from Gatan, with the treated surfaces facing each other. The specimen sandwich was then cast into a brass tube with a 3 mm outer diameter, using E-SOLDER 3021 silver epoxy supplied by ACME Division of Allied Products Corporation, Discs of 400-500 pm thickness were sectioned from the tube assembly using a LECO VC-50 thin section slitting wheel. The direction of the cutting was perpendicular to the epoxy-bonded joint to minimise the loss of the epoxy. A TENUPOL-2 jet polisher and a multi-function EDWARDS IBT Auto Model 306 Coater equipped with an ion gun were used for pre-thinning and final ionbeam thinning of the XTEM samples. TEM investigations were conducted on a Jeol JEM-2000FX transmission electron microscope at an accelerating voltage of 200 kV. The size of the electron beam for NBD is less than 5 nm.

Fig. 1. XTEM image of the modified amorphous layer with correlated NBD pattern, PI3 treated at 150 “C.

3.2. PI3 at 250 “C

The total thickness of the modified layer at 250 “C is about 4.5 pm and can be divided into 3 sublayers. The top nanocrystalline sublayer I is about 0.9 pm thick, composed mainly of CrN and a-ferrite, with traces of s-nitride (Fig. 2). The NBD patterns of the nanocrystalline first sublayer show that both Bain and N-W relationships exist between CrN and u. (Fig. 3). The second sublayer is semi-amorphous and about 2 pm thick, showing strip-like modulation with a strip space between 4

3. Results 3.1. PP at 150 “C

The modified layer is amorphous, about 0.9-l pm thick (Fig. 1). The modified layer has a very sharp interface with the substrate. The substrate under the modified layer is stressed to form a heavily stressed zone which is up to 4 pm thick. a-ferrite and small amounts of s-martensite are detected in the stressed zone.

40 nm



Fig. 2. The first nano-crystalline sublayer and correlated SAD pattern, PI3 treated at 250 “C.

X Li et al.lSurface and Coatings Technology 85 (1996) 28-36

0 CrN snots,Bain 0 0-N SD&?,N-W Fig. 3. NBD pattern shows that the first nano-crystalline sublayer is compc bsed of very fine CrN and c( precipitates. Both Bain and N-W relationships are obsexved between these two phases. PI3 treated at 250 “C.

20nm 1 I Fig. 4. The second semi-amorphous sublayer and correlated NBD pattern, PI3 treated at 250 “C.

Fig. 5. The third amorphous sublayer and correlated NBD pattern, PI3 treated at 250 “C

and 6 nm (Fig. 4). The third sublayer is amorphous and is about 1.5 pm thick (Fig. 5). The amorphous structure showed similar clustered structure to the 150“C implanted sample but with a sharper contrast of the clusters.The size of the clustersis about 4-7 nm. Close to the interface, the contrast of the cluster gradually fades towards semi-amorphous. A sharp interface was observed between the third sublayer and the substrate (Fig. 6). The substrate under the modified layer appears to be the normal untreated 316 steel.

a three sublayer structure is also found in the 350 “C PI3 treated sample. The first sublayer is nanocystalline and about 0.9-l pm thick, composed mainly of CrN and CI,with tracesof s-nitride. The width of the tiny, strip-like crystal colonies is less than 20 nm. Once again, both Bain and N-W relationships are observed between the CrN and a phases. Fig. 7 shows the transition from the first nanocrystalline sublayer to the second amorphous sublayer. The second sublayer is about 0.9 ltrn thick. The amorphous second sublayer also shows a clustered structure though the contrast of the clusters is not as sharp as it is in the amorphous structure of the 250 “C treated sample.

3.3. PI3 at 350 “C

The total thickness of the 350 “C treated sample is about 2.3 urn. Similar to the 250 “C PI3 treated sample,

X Li et al.jSurface and Coatings Technology 85 (1996) 28-36

50 nm

I Fig. 6. Interface between the third amorphous substrate, PI3 treated at 250 “C.

sublayer and the

Fig. 8. Interface between the third semi-amorphous sublayer and the substrate, PI3 treated at 350 “C.

modified layer and the substrate. The substrate below the interface shows a normal untreated microstructure. 3.4. PI3 at 450 “C The 450 “C implanted sample shows a 3 pm thick modified layer which has a sharp interface with the substrate (Fig. 9). Detail of the modified layer shows it is amorphous with a cluster size of 4-7 nm (Fig. 10). Towards the interface, the cluster structure gradually

40 nm


Fig. 7. Transition from first nanocrystalline sublayer to second amorphous sublayer, PI3 treated at 350 “C. The NBD pattern of the nanocrystalline structure shows Bain relationship between the CrN and o! precipitates.

The third sublayer is semi-amorphous and about 0.5-0.6 urn thick. Fig. 8 shows the details of the third sublayer and the interface between the third sublayer and the substrate. Towards the substrate, the contrast of the clusters in the third semi-amorphous sublayer gradually fades. Some dark spots of less than 10 nm are observed in the third sublayer. The presence of prohibited diffraction spots of the NBD pattern reveals they are tiny (Cr,Fe)N crystals, as well as a and y’-Fe4N crystals. A sharp interface is observed between the

,lw, Fig. 9. XTEM

image of the modified layer, PI3 treated at 450 “C.


X Li et al,/Swfaee

and Coatings


85 (1996)


20 nm Fig 10. Detail of the amorphous structure, PI3 treated at 450 “C.


I Fig. 11. XTEM


image of the modified layer, PI3 treated at 520 “C.

fades. The NBD pattern taken near to the interface shows semi-amorphous features. The substrate below the modified layer shows a very high dislocation density. The thickness of this high dislocation zone is about 1.5-2 urn. The NBD pattern of the high dislocation zone shows that it has the same orientation as the substrate but with about 3% lattice expansion. If we include this high dislocation zone as a second sublayer, the total thickness of the modified layer would be about 4.5-5 urn. 3.5. PI3 nt 520 “C

The sample treated at 520 “C comprises two sublayers, a 0.4-0.5 urn thick sublayer I and a 7.5-8.5 urn thick sublayer II. The total thickness of the modified layer is about 8-9 urn (Fig. 11). The first sublayer is nanocrystalline, characterised by very fine randomly dispersed precipitates ranging in size from a few nm to around 50 nm. The NBD pattern shows that the first sublayer is composed of mainly CrN and c( phases, together with a small amount of h.c.p. a-nitride, hexagonal Cr,N and h.c.p. (Cr,Mo)N, [ 81. Sublayer II is made up of colonies displaying a lamellar structure (Fig. 12). The colony and precipitates appear to have been formed by a cellular precipitation process and show a pearlite-type appearance. With the aid of NBD, the phases present were identified to be CrN and CI. Both ‘Bain and N-W relationships are observed between the CrN and a phases. A small amount of c-nitride phase is detected between the colonies at the near surface region in sublayer II. Like all the other samples, the 520 “C treated sample shows a very sharp interface with the substrate (Fig. 13).

50 nm Fig. 12. Detail of the lamellar CrN+cc structure, PI3 treated at 520 “C. s-nitride locates between the CrN f 5 colonies (dark area with fine fringes). Correlated NBD pattern (top left of the photo) shows [ll?O]s The NBD pattern at the bottom right shows Bain relation between CrN and cc-ferrite.

A few tiny Cr,N and CrMoN, crystals at a size of about lo-20 nm were found to exist between the CrN and CI colonies near the interface area.

4. Discussion Fig. 14 shows the schematic description of the surface microstructure of the 150, 250, 350, 450 and 520 “C PI3 treated 316 stainless steel samples.

X. Li et al./Swfaee and Coatings Technology85 (1996) 28-36




Fig. 13. Interface between the second sublayer and the substrate, PI3 treated at 520 “C. Some Cr,N or CrMoN, islands can be seen between the CrN + CIcolonies (top left area of the photo).

From the results shown above, it is obvious that the temperature of the implantation treatment, as well as the treatment time, play a decisive role in the evolution of the surface microstructure. A thick amorphous layer or a thick amorphous sublayer, which is interestingly located below the nano-crystalline sublayer, can be formed by nitrogen ion implantation at temperatures below 500 “C. In the crystalline layers or sublayers formed at all temperatures, CrN was found to be the dominant nitride which was cellularly decomposed with CIto form a pearlite-type structure. 4.1. Amorphous structure in nitrogen implanted austenitic stainlesssteel

The detailed structure of amorphous phases is a topic which has been debated for some time. However, short range order and middle range ordering in amorphous metals have been widely reported in recent years. Ke proposed that the size of the ordered cluster determines the degree of amorphisation [9], i.e., smaller clusters correspond to a higher degree of amorphisation. Since amorphousness is a relative concept, in this article we define amorphous by NBD patterns. The microstructure corresponding to NBD ring patterns is defined as amorphous. The obvious NBD ring patterns with clear diffraction spots correspond to the semi-amorphous structure. The NBD patterns with or without dim rings only correspond to the crystalline structure. The formation of a thick amorphous modified layer obtained in this study, especially the thick amorphous sublayer under the nanocrystalline sublayer, argues


strongly against the idea of radiation induced amorphisation in ion implanted steel. In fact, quite a few articles in the past few years have reported that implantation induced amorphisation only occurs if there is a chemical reaction between the implanted ions and the target metal. No amorphous structure has ever been found in self-implanted or inert gas implanted metals or alloys [ 10-121. Furthermore, it has been reported that Fe, Cr and Ni implanted into 304L steel did not activate y to a transformation at very high doses [ 131. From some previous work of implanting nitrogen into pure gold, it was found that nitrogen implantation did not change the dislocation density even at doses up to 2 x 10” ions cmw2 but increased the density of nitrogen bubbles formed in the gold matrix [14]. All the results mentioned above strongly support the idea that radiation damage has almost nothing to do with the amorphisation of ion implanted metals. In fact, the stable disordering due to simple displacement-cascade-damage overlap has so far been proven only in semiconductors but has not been observed in pure metal even at very low temperatures [ 151. Implantation induced amorphisation in metals has only been found in cases where the disorder was produced by significant quantities of chemically active solute atoms. Several factors may influence the formation of the amorphous phase by solid state reaction. Firstly, it has been reported that the composition of most amorphous alloys is close to eutectic concentration [ 161. The widely different composition of the precipitating phases requires large rearrangement of the atoms, making nucleation and growth of new crystals more difficult. In Fe-Cr-Ni austenitic stainless steel, it has been found that implanted nitrogen always binds with the Cr and the two diffuse together [ 171. Cr-N bonding makes the mobility of Cr even poorer. It is therefore conceivable that, by analogy, the cellular precipitation of CrN + a from austenite will result in amorphisation if the temperature is not high enough to start the precipitation process. Secondly, amorphisation can be greatly enhanced when a restricted solubility range exists for the compound which could precipitate [18]. A slight deviation from the compound’s stoichiometric concentration can result in a large rise in the free energy of the system. When the free energy of the crystalline state becomes greater than that of the amorphous state, a spontaneous amorphous transformation is favoured. CrN is one such line compound, without composition range. Summing up all the above discussion, it is fair to say that a solid state chemical reaction and poor mobility of the reactant atoms are the key factors for solid state amorphisation by nitrogen ion implantation of austenitic stainless steel. The amorphisation obtained at the 450 “C treated

X Li et al.lSurface and Coatings Technology 85 (1996) 28-36 Nano-crjstdUne sublayer CrNta (@Ypm)

Amorphous layer

Sembamorphws s”grgl Stress



\ x





(1.5 pm) y substrate

iiiiililili!ililiil! .:::::::::::::::::::: .. ... .... ... ... ...*. :::::::::::::::::::: :::::::::::::::::::: .I**III.III, . ... ... ....I.. . . ,, .*..*.*..., *........*, . . ..I..... **.......* .:.:.:.:.:.:.:.:.:.: ’ E


About 100 ’ nm thick slightly expanded y with a below the interface


a. PI3 treated at 150°C 2x1017atoms/cm2,430 min.




1cK)run [email protected] transition iDIK: (Cr,Fe)Ntaty’t Cr2N(tncc)


b. PI3 treated at 250°C 8x1017atoms/cm2,2.50min,

Nano.c~3tallinc sublayer I CrNta with

Nano-uystaliine sublayer I tiNta+E(trace)


(a 91 pm),

layer Amwphous~ sublayer II (0.9 pm)

[3 pm)

Ll2SStll~lCO~ semi-amorphous --) transition zcne with (Cr,Fe)Ntaty’

Semi-amorphous sublayer III with very fine (Cr,Fe)N+ay’

High density Y dislocation i, (1.5 am)


y substrate#

y substrate


c.PI3 treatedat 350°C 8~10’~ atoms/cm2,130 min.

d. PI3 treatedat 450°C 8x1017atoms/cm2,70 min.

e. PI? treated at 520°C 8x1017atoms/cm2,40 min.

Fig. 14. Schematic description of the PI3 treatment modified surface layer on 316 stainless steel,

sample shows that the amorphous phase can be stable up to quite high temperatures in austenitic stainlesssteel if conditions such as the nitrogen concentration are suitable. Recent reports on Ti-(30-40%)Cr alloy show that not only can the amorphous phase be stable at high temperatures, but the transformation between the partially amorphous and crystalline phasescan be reversed by the application of alternating annealing steps at 600-800 “C [ 191. Basedon thesereports, one can imagine that at a proper concentration and temperature, the amorphous structure may be in a stable phase at much higher temperatures than is usually expected. Unfortunately, most of the discussionson the micro-

structure of nitrogen ion implanted stainlesssteelshave been confined to the Fe-N or Cr-N binary phase diagrams, which is inadequate for multi-phase systems. To get a clearer picture of the amorphous structure of the high nitrogen concentration Fe-Cr-Ni system, a phase diagram of the Fe-Cr-Ni-N system would be very helpful. 4.2. Why ON is the do~nirtnnt nitride

It is generally accepted that thermodynamic factors still prevail even in the non-equilibrium process of ion implantation [20]. Therefore, the more negative the

X Li et al./Surface

and Coatings

heat of formation for a nitride, the greater is the driving force of its nucleation and growth and, as a result, this nitride is easily detected in the modified layer. However, to determine whether a nitride could be formed by nitrogen implantation, not only thermodynamics but also kinetics should be considered. By implanting nitrogen into pure Cr, Zhao et al. [21] found that most of the products of the implantation were CrN. Cr,N could only be found at the node area. The explanation of this result could be based on the fact that f.c.c. CrN can grow in 6 directions which can transfer to each other, and can also grow very large. By contrast, CrzN (hexagonal) grows in exact crystal directions, is difficult to grow and can easily be stopped. Table 3 shows the heat of formation for some of the metal nitrides [22]. From the dHf values shown in the table, it can be seen that Cr and MO nitride are thermodynamically favoured to form in nitrogen implanted 316 stainless steel. Considering the kinetic factor, only f.c.c. nitrides like CrN and Mo,N are favoured to form. However, the ,4Hr value of Mo,N is a little higher than that of CrN, and the content of MO in 316 steel is also very low (about 2.34%). So even if Mo,N forms, the amount of Mo,N will be very small. More importantly, not only do both CrN and MozN have the same f.c.c. lattice, the lattice parameter of Mo,N (4.16 A) and CrN (4.14 A) are also very close. Because it is almost impossible to identify a small amount of Mo,N from CrN by the techniques used in this study, CrN is obviously the most detectable nitride. From a wide study of nitriding, Jack also pointed out that below 575 “C, CrN is more stable than Cr,N [23]. During nitriding of Cr-alloy steel, CrN always precipitates except at a low nitrogen potential where CrzN may precipitate.




tration obtained could be very different, resulting in different microstructures. For better control of implantation, we suggest that the dose rate should be considered.

5. Conclusions

(1) A

thick amorphous layer or sublayer up to 3 pm thick can be obtained in nitrogen implanted 316 austenitic stainless steel at temperatures up to 450 “C. The amorphous structure has nothing to do with radiation damage. The amorphisation is controlled by a solid state chemical reaction and the low mobility of implanted and substitutional atoms. (2) Cellular CrN + a precipitates obtained at a high temperature or at high nitrogen concentrations are the dominant phases in the crystalline structure of the nitrogen implanted 316 austenitic stainless steel. Both N-W and Bain relationships are observed between CrN and a. (3) Temperature, time and dose rate play an important role in PI3 treatment of 316 austenitic stainless steel.

Acknowledgments This work is supported by the Australian Institute of Nuclear Science and Engineering. Xinyang Li acknowledges the support of an ANSTO postgraduate scholarship. Xinyang Li would also like to thank Dr. David Wexler of the University of Wollongong, Mr. Mark Blackford and Dr. David Mitchell of ANSTO and Mr. Phillip Renwick of BHP for their help on TEM work.


4.3. The important factors for implantation From these discussions, it is obvious that temperature and time of implantation are both very important factors which decide the results of the implantation in austenitic stainless steel. However, since both thermodynamic and kinetic criteria still play a role, the concentration of the implanted atoms is another key factor controlling the degree of lattice distortion and concentration gradient and hence the diffusivity of the nitrogen atoms. Therefore, if only the total implantation dose is controlled, neglecting the time of implantation, the concen-

[l] [2] [3] [4] [S]

[6] Table 3 Heat of formation (AH,) of some metal nitride


85 (1996)


CrN (fee.)

Cr,N (hex.)

Fe4N (Etc.)

Fe,N (orth.)

MoN (hex.)

Mo,N (Etc.)

Ni,N (hex.)








[8] [9] [lOl

M. Samandi, B.A. Shedden, D.I. Smith, G.A. Collins, R. Hutchings and J. Tendys, Surf. Coat. Technol., 59 (1993) 261-266. D.L. Williamson, 0. Ozturk and S. Glick, Nncl. Instrum. Meth. B, 59/60 (1991) 737. S. Shrivastava, R.D. Tarey, MC. Bhatnagar, A. Jain and K.L. Cohpra, Surf. Coat. Technol., 50 (1991) 41. S. Fayeulle and D. Treheux, Appl. Surf. Sci., 25 (1986) 288. X. Li, M. Samandi, D. Dunne and R. Hutchings, in C. Subramanian and K.N. Strafford (eds.), Proc. 2nd Aust. Int. Conf. on Surface Enginering: Coatings and Surface Treatments in Munufncturing, Adelaide, March 1994, Vol. 2, Surface Engineering Research Group, University of South Australia, 1994, p. c201. G.A. Collins, R. Hutchings, K.T. Short, J. Tendys, X. Li and M. Samandi, Nitriding of austenitic stainless steel by plasma immersion ion implantation, submitted to Surf. Coat. Technol. M. Samandi, X. Li and B.A.Shedden, Conf. Proc., Microscopy: Materials and Techniques, Mascot, Australia, Institute of Metals and Materials Australia, 21-22 September, 1993, p. 31. B. Snyder and V. Snyder, Trans. Am. Sot. Metals, 45 (1953) 397. L. Ke, Acta Metall. B, 5 (1992) 245-254. W.A. Grant, A. Ah, L.T. Chadderton, P.J. Grundy and E.



[12] [13] [ 141 [15]

.K Li et al./Surface and Coatings Technology 85 (1996) 28-36 Johnson, in B. Cantor (ed.), Rapidly Quenched Metals III: Proc. of the Third Int. Conf. on Rapidly Quenched Metals, University of Sussex, Brighton, 3-7 July 1978. N.L. Lee, G.B. Fisher and Y.T. Cheng, in G.S. Was and KS. Grabowski (eds.), Enuironmental Degradation of Ion and Lasej Beam Treated Surfaces, TMS, 1989, pp. 107-126. J.M. Hampikian and D.I. Potter, Oxid. Metals, 38 (1992) 125-137. EC. Cooney and D.I. Potter, Surf. Coat. Technol., 51 (1992) 420-428. X. Li, The ejject of nitrogen implnntation on gold, Univ. Adelaide internal memo., November 1991. C. Cohen, A. Benyagoub, H. Bernas, J. Chaumont, L. Thorn&, M. Berti and A.V. Drigo, Phys. Reu. B, 31 (1985) 5.


S. Steeb and P. Lamparter, J. Not+Cryst. Solids, 150 (1993) 24-33. r171 D.L. Williamson, 0. Ozturk, R. Wei and P.J. Wilbur, Surf. Coat. Technol., 65 (1994) 15-23. 1181 J.L. Brimhall and E.P. Simonen, Nwl. Iustr. hfeth. B, 16 (1986) 187. [191 R. Bormann, Mater. Sci. Eng. A, 179,480 (1989) 31-35. I331 B.X. Liu, S. Zhou and H.D. Li, Phys. Stat. Sol. A, I13 (1989) 11-22. WI Y.F. Zhao, S.Z. Sun, L.L. He, H.Q. Ye and S.J. Pang, I+rcunr~t, 43 (1992) 1065-1067. [221 P.C.P. Bouten and A.R. Miedema, J, Less-Corrrm. Met., 65 (1979) 217. ~231 K.H. Jack, Heat Treatment 73, Metals Society, London, 1975, pp. 39-50.