Materials Transactions, Vol. 5, No. () pp. 19 to 19 # The Japan Institute of Metals Interreactions of Thin Film on Bulk -TiAl and on Bulk -Ti 3 Al Alloys at 7 1 C Min-Sheng Chu and Shyi-Kaan Wu* Department of Materials Science and Engineering, National Taiwan University, Taipei 1, Taiwan, R.O. China Interreactions of diffusion couples of film on bulk -TiAl and film on bulk -Ti 3 Al are investigated at high temperature. Experimental results show that layer and /-TiAl mixed layers are observed at the interfaces of film/bulk -TiAl and film/bulk -Ti 3 Al diffusion couples, respectively, at 71 C. In addition, the growth rates of and -TiAl product layers comply well with a parabolic law. The growth activation energy, Q k, of the phase in the film/bulk -TiAl system is calculated as 15.9 kj/ mol, and Q k values of -TiAl and phases in film/bulk -Ti 3 Al system are 13. kj/mol and 17.9 kj/mol, respectively. These Q k values are similar in magnitude with those of phase formation in Ti-Al thin film diffusion systems. In this study, formation is suggested to have been nucleated at the interface during the heating prior to reaching the set temperature. (Received December 11, 3; Accepted February 3, ) Keywords: sputtering, intermetallics, diffusion couple, kinetics, -TiAl and -Ti 3 Al 1. Introduction Titanium aluminides such as -TiAl and -Ti 3 Al alloys are high temperature structural intermetallics with rapidly growing technological importance. They have the characteristics of high temperature stability, good creep resistance and relatively high yield strength. 1,) Their superb properties have found applications in aerospace and automotive applications. 3,) The interdiffusion and interreaction phenomena of the Ti-Al diffusion couple system have previously been studied in detail. 5 1) As shown by these studies, phenomena occurring in thin film diffusion couples are very different from those observed in bulk couples. This comes from the fact that thin films contain more interfaces and surface than bulk specimens. Additionally, the thickness of thin films is usually limited to a few thousand nanometers. Hence, these inherent characteristics of thin films are often responsible for the particular behaviors occurring there. Among the various titanium aluminides in the Ti-Al binary alloys, it has been reported that is thermodynamically and kinetically more favorable to form first than the other titanium aluminides in the bulk or thin film interreactions of Ti and Al diffusion couples. 5 13) Diffusion tests in the Ti-Al bulk couple system have indicated that the growth of the layer obeys the parabolic time dependence. In addition, the growth of the phase in the Ti-Al thin film couple system obeys the diffusion controlled kinetics with activation energies varying from about 15 to 19 kj/mol according to different researchers. 9 1) To our best knowledge, most investigations on phase interreactions in the Ti-Al system have concentrated on the formation of the thin film diffusion couple or the bulk diffusion couple. In the case of the diffusion couple of thin film on -TiAl or on - Ti 3 Al bulk alloys, their interreactions to form and - TiAl phases at high temperature have not previously been reported. In this study, -TiAl and -Ti 3 Al intermetallics were *Corresponding author, E-mail: skw@ccms.ntu.edu.tw sputtered with a pure Al thin film. A subsequent interdiffusion treatment at C was conducted, in which an adhesive layer forms between the Al film and -TiAl or - Ti 3 Al. Therefore, simple diffusion couples of thin film on -TiAl bulk alloy or on -Ti 3 Al bulk alloy are obtained. We investigate the further interreactions of these simple diffusion couples at 71 C, including the phase formation sequence and kinetics. At the same time, based on the measured data of the thicknesses of formed phases in these diffusion couples, the activation energies of phases formation can also be calculated.. Experimental Procedures The Ti-5 at%al (-TiAl) and Ti-5 at%al ( -Ti 3 Al) alloys were prepared as ingots from the raw materials of titanium (99.7%) and aluminum (99.9%) by a vacuum arc remelter (VAR). The ingots were re-melted at least six times. After homogenization of the ingots at 1 C for 1 h, square specimens of 1 1 1 mm 3 were sectioned by a diamond saw and ground to a final polishing of.3 mm alumina, then ultrasonically cleaned with acetone, ethanol and deionized water and then blown dry before sputtering. The sputtering target of 99.9% pure aluminum was 5 mm in diameter and 5 mm in thickness. The Al films were sputter-deposited on all the surfaces of the -TiAl and -Ti 3 Al specimens using an r.f. magnetron sputtering apparatus with a turbo-pumped vacuum system at a base pressure of about :5 1 5 Pa. The working pressure was set at :5 1 1 Pa with ultra high purity argon. The r.f. power was 1 W and the substrate was unheated during the Al sputtering. Al films were deposited on specimens with their thickness measured by -step (Veeco Dektak 3 ST). 15,1) A subsequent interdiffusion treatment was carried out at C for h in a high vacuum furnace that reached about 3:99 1 5 Pa. After interdiffusion treatment, the specimens were then further heated in a muffle furnace at temperatures ranging from 71 C for 3 min to 3 h in ambient atmosphere.
Interreactions of Thin Film on Bulk -TiAl and on Bulk -Ti 3 Al Alloys at 7 1 C 191 Epoxy Epoxy Fig. 1 SEM cross-sectional morphologies of Al-sputtered specimens after interdiffusion treatment at C for h in a vacuum. - TiAl with 5 mm Al film and -Ti 3 Al with 3 mm Al film. (d) Al O 3 1 3 5 (e) Al O 3 (c) (f) Al O 3 Fig. SEM cross-sectional images of -TiAl specimens with 5 mm Al films which have been interdiffusion treated, and then heated at C in air for 1 h 5 h (c) 1 h (d) h (e) h and (f) 3 h. The EDS analysis for the composition of points 15 shown in is listed in Table 1.
19 M.-S. Chu and S.-K. Wu An X-ray diffractometer (XRD, Philips PW179) with Cu k radiation at 3 kv, ma and /min scanning rate was used to identify the phases of sputtered-films formed by the interdiffusion treatment and the following interreaction tests. The cross-sectional microstructures of specimens were observed by means of a Leo 153 scanning electron microscope (SEM) coupled with energy dispersive spectrometry (EDS). The layer thickness of formed phase was determined by the measured function of the SEM, using the average of ten measurements. Some of the samples were examined by electron probe microanalyzer (EPMA, JEOL JXA- SX) to measure the chemical compositions of phases formed by sputtering, interdiffusion treatment and interreaction tests. Table 1 EDS analysis of Fig. for the composition of points 15. Composition Location Ti (at%) Al (at%) phase Point 1.5 75.35 Point 5.1 7. Point 3 33..5 Point.5 51.35 -TiAl Point 5 5.5 9.75 -TiAl 3. Results and Discussion 3.1 Interdiffusion treatment of as-sputtered Al films with bulk -TiAl and -Ti 3 Al alloys Al films with different thicknesses of 15 mm were sputtered on the -TiAl and -Ti 3 Al alloys. The appropriate thickness of Al film can be controlled accurately according to the sputtering rate of Al film as about.35 nm/s. 15,1) However, intrinsic stress in the film increases with increasing the film thickness, resulting in poor adhesion of coating to substrate. Once the thickness of Al film is over 5 mm, small parts of the film spall out from the -TiAl and -Ti 3 Al specimens after the interdiffusion treatment. Hence, in this study, the thicknesses of as-sputtered Al films are all less than 5 mm. The as-sputtered Al film reveals a crystalline structure in terms of XRD pattern. 15) The interdiffusion treatment at C for h can not only increase the adhesion of Al film with substrates but also promote the formation of phase, as shown in Fig. 1. 15,1) Figures 1 and show the cross-sectional SEM morphologies of Al-sputtered -TiAl and -Ti 3 Al, respectively, after interdiffusion treatment at C for h in a vacuum. From Fig. 1, a layer on the outer surface of the substrate can be seen, which has a uniform thickness with no cracks or voids at the interface. Thus, the interdiffusion treatment is beneficial to increase the bonding of the films and substrates, (c) Al O 3 (d) AlO3 Fig. 3 SEM cross-sectional morphologies of -TiAl specimens with 5 mm Al films heated at high temperatures in air for 9 C 3 min 9 C 1 h (C) 1 C 3 min and (d) 1 C 1 h. The specimens were interdiffusion treated in high vacuum before the interreaction.
Interreactions of Thin Film on Bulk -TiAl and on Bulk -Ti 3 Al Alloys at 7 1 C 193 T hickness, t h 1 1 1 1 C 5 1 15 5 3 35 5 1 15 5 3 1 9 C 1 1 1 1 1 1 1 (c) 1 C 5 1 15 5 3 35 5 1 15 5 3 1 Fig. Dependence of the thicknesses of different formed phases in film/bulk -TiAl. vs. interreaction time. C 9 C and (c) 1 C. and also beneficial to investigate the subsequent interreaction tests between the thin film and its substrates. Therefore, film/bulk -TiAl and film/bulk -Ti 3 Al diffusion couples are formed via interdiffusion treatment. According to an effective heat of formation model using the solid-state diffusion theory, the predicted first phase formed in Al film/-tial film and in Al film/ -Ti 3 Al film diffusion couples is. 17) This is because the formation of compound has a more negative Gibbs energy level than the other titanium aluminides. 17) In Fig. 1, the formation of phase continues until all the sputtered Al film is exhausted during the C h interdiffusion treatment. This indicates that the thickness of layer formed by interdiffusion is dependent on the thickness of sputtered Al film. From Fig. 1, with a 5 mm sputtered-al film before the interdiffusion treatment, the thickness of layer is about mm. In contrast, from Fig. 1, with a 3 mm sputtered-al film before the interdiffusion treatment, the thickness of layer is only about mm. Clearly, the growth rate of the layer on -TiAl is higher than that on -Ti 3 Al, which shows that the Al-atom diffusion in -TiAl is faster than that in -Ti 3 Al during the C interdiffusion. This feature agrees with the reported results of Ref. 1), in which the interdiffusion coefficient of Al-atoms was determined by calculating the Al tracer in -TiAl and -Ti 3 Al single-phase bulk diffusion couples. Moreover, from many reported diffusion couples tests in the Ti-Al system, the layer is observed to grow rapidly and complies with the parabolic time dependence.,9 1) In our cases, we suggest that the growth in Fig. 1 is also following the parabolic law, although its growth kinetics is not investigated in this study. 3. Interreaction in between film and bulk -TiAl at 1 C It has been reported that, in the and -TiAl diffusion couple, the phase can be formed first by the interreaction at high temperature. 7,15) In order to understand the subsequent phase formation developed between bulk -TiAl and thin film, specimens of Fig. 1 were heated at C, 9 C and 1 C for different time intervals in the ambient atmosphere. Figure shows the SEM crosssectional microstructure of the specimen heated at C for 1 h. From Fig., the outer layer (points 1 and of Fig. ) and the inner layer (point 3 of Fig. ) coexist, as shown by the accompanying EDS analysis, which is listed in Table 1. The appearance of a layer indicates that the film has interreacted with bulk -TiAl at C, indicated by the Ti-Al phase diagram. 1) With increasing interreaction time at C, the thickness of the layer decreases whereas that of the layer increases, as shown in Figs. and (c). This is because of the solid-state diffusion between thin film and bulk - TiAl at high temperature. From Fig., comparing the thickness of the layer to that of the layer, the layer will be exhausted completely within approximately h at C. Meanwhile, the thickness of layer reaches its maximum of about 15 mm at h, as shown in Fig. (d). In addition, Figs. (d), (e) and (f) reveal that, not only has the formed, but also a stable Al O 3 layer adheres on the outer surface of the specimen at C. In Fig., the Al O 3 layer is not completed during the initial period, but becomes a continuous layer with further heating time. This stable Al O 3 layer becomes a good barrier against the inward oxygen attack, resulting in the significantly improved oxidation resistance of -TiAl at high temperature. 15) The formation of an Al O 3 layer needs a supply of Al atoms from the layer, or from the layer after the layer has been exhausted. At the same time, Al
19 M.-S. Chu and S.-K. Wu (d) 1 3 5 (e) Al O 3 (c) (f) Al O 3 Ti-Al-O compound Fig. 5 SEM cross-sectional images of -Ti 3 Al specimens with mm Al films which have been interdiffusion treated, and then heated at C in air for 1 h 5 h (c) 1 h (d) h (e) h and (f) 3 h. The EDS analysis for the composition of points 15 shown in is listed in Table. Table EDS analysis of Fig. 5 for the composition of points 15. Composition Location Ti (at%) Al (at%) Phase Point 1 5.5 7.35 Point 33.5.75 Point 3 5.3 9.77 -TiAl Point 73.15.5 -Ti 3 Al Point 5 7.37 5.3 -Ti 3 Al atoms of the layer also diffuse into the -TiAl substrate, as shown in points and 5 of Fig. with EDS analysis listed in Table 1. From Fig., the layer stops its growth at C h, and starts to become thinner with further heating time. After C 3 h, as shown in Fig. (f), the layer is reduced to about mm in thickness, and will be exhausted eventually. This is because Al atoms in the layer must diffuse outward to form the Al O 3 layer, as well as diffuse inward to increase the Al content in the -TiAl substrate. Figures 3 and show SEM cross-sectional morphologies of specimens of Fig. 1 heated at 9 C in air for 3 min and 1 h, respectively. Clearly, the layer shown in Fig. 3 is thicker than that in Fig.. The and layers coexist in Fig. (c), but only the layer is observed in Fig. 3. This feature accounts for the fact that growth rate of the layer increases significantly with increasing temperature, while the is completely exhausted within h at 9 C but within h at C. While the specimen was heated at 1 C in air for 3 min
Interreactions of Thin Film on Bulk -TiAl and on Bulk -Ti 3 Al Alloys at 7 1 C 195 (c) (d) Fig. SEM cross-sectional morphologies of -Ti 3 Al specimens with mm Al films heated at high temperatures in air for 7 C 3 min 7 C 1 h (C) 9 C 3 min and (d) 9 C 1 h. The specimens were interdiffusion treated in high vacuum before the interreaction. and 1 h, as shown in Figs. 3(c) and (d), respectively, the layer is estimated to have been exhausted within 1 h. In addition, according to the thickness of layer shown in Fig. 3(d), the exhausting time of layer at 1 Cis estimated to be shorter than 15 h. From Figs. and 3, the dependence of the thicknesses of and layers vs. interreaction time for film on bulk -TiAl heated at C, 9 C and 1 C can be plotted, as shown in Figs., and (c), respectively. From Fig., the growth of the layer can be described by the diffusion-controlled model within the initial interreaction (< h at C, <1 h at 9 C and <5 h at 1 C), as discussed further in Section 3.. 3.3 Interreaction between film and bulk -Ti 3 Al at 79 C In order to realize the phase formation sequence between bulk -Ti 3 Al and thin film, specimens of Fig. 1 were heated at C for 3 min3 h, and their crosssectional SEM images with accompanying EDS analyses are shown in Fig. 5 and Table, respectively. From Fig. 5 of C 1 h, the and -TiAl phases were formed between the and -Ti 3 Al layers where a mixture layer of / /-TiAl/ -Ti 3 Al coexists. The EDS results of points and 3 confirm the formation of and -TiAl phases. It has been reported that and -TiAl phases can be formed simultaneously at the interface of and -Ti 3 Al in a Ti-Al bulk diffusion couple. 7) After 5 h heating in air, the is exhausted and the thicknesses of and -TiAl layers increase, as shown in Fig. 5. With further heating time, the thickness of decreases gradually, as compared with Figs. 5 and (c). The layer will be exhausted completely at h, as shown in Fig. 5(d). At this time, a mixture layer of Al O 3 /-TiAl/ -Ti 3 Al is observed. An adhered Al O 3 layer is formed on the outer surface of -TiAl layer, as shown in Fig. 5(e), and it provides excellent improvement in the oxidation resistance of the - Ti 3 Al alloy at high temperature. 1) From the EDS analysis of Fig. 5, in the -Ti 3 Al substrate, the Al content is higher at the position nearer the -TiAl layer, as indicated in points and 5 of Fig. 5. While the heating time reaches 3 h, there is a Ti-Al-O compound existing within the -TiAl layer, as shown in Fig. 5(f). This feature implies that the -TiAl beneath the protective Al O 3 layer will be destroyed by this Ti-Al-O compound with longer heating time. Figure shows the SEM cross-sectional morphologies of Fig. 1 specimens heated at 7 C (Figs. and ) and 9 C (Figs. (c) and (d)) in air for 3 min and 1 h, respectively. From the inserted photo of Fig., very thin layers of and -TiAl phases are formed at the interface of film and -Ti 3 Al substrate in which the and -TiAl phases are formed at the same time. Figure shows a sandwich layer of / /-TiAl/ -Ti 3 Al at 7 C for 1 h heating. This microstructure coincides with Fig. 5,
19 M.-S. Chu and S.-K. Wu 5 3 1 1 1 7 C 1 3 5 7 9 1 C 5 1 15 5 3 35 5 1 15 5 3 (c) 9 C Thickness of, t h Thickness of, t h 1 1 1 1 1 5..5. 3.5 3..5. 1.5 1..5. C 9 C 1 C 1 (Interreaction Time) 1/, t / h 1/.5 1. 1.5..5 3. 3.5..5 5. (Interreaction Time) 1/, t / h 1/ 7 C C Fig. Growth kinetics of the phase in between film and bulk -TiAl at C, 9 C and 1 C in between film and bulk -Ti 3 Al at 7 C and C. 1 Fig. 7 Dependence of the thicknesses of different formed phases in film/bulk -Ti 3 Al vs. interreaction time. 7 C C and (c) 9 C. which was heated at C for 1 h. Figure (c) shows that /-TiAl/ -Ti 3 Al layers are coexisting without a layer after 9 C 3 min interreaction. With further heating time, only the -TiAl layer remains at 9 C 1 h, as shown in Fig. (d). Figures 7, and (c) show the dependence of the thicknesses of, and -TiAl phases vs. interreaction time for film on bulk -Ti 3 Al heated at 7 C, C and 9 C, respectively. From Fig. 7, the Thickness of γ -TiAl, t h 11 1 9 7 5 3 1 7 C C 9 C 1 (Interreaction Time) 1/, t / h 1/ Fig. 9 Growth kinetics of the -TiAl phase between film and bulk -Ti 3 Al at 7 C, C and 9 C. growth of and -TiAl layers is also in agreement with diffusion-controlled behavior within 1 h at 7 C, 5 h at C and 3 min at 9 C for formation; and within h at 7 C, h at C and 1 h at 9 C for -TiAl formation, as discussed further in Section 3..
Interreactions of Thin Film on Bulk -TiAl and on Bulk -Ti 3 Al Alloys at 7 1 C 197 3. Interreaction kinetics of phase formation between film/bulk -TiAl and film/bulk -Ti 3 Al systems Figures and 7 show the thicknesses of the, and -TiAl layers vs. interreaction time at high temperature in film/bulk -TiAl and film/bulk -Ti 3 Al, respectively. From Figs. and 7, the curves of and - TiAl thicknesses vs. heating time exhibit a complete interreaction and are suitable for investigating the kinetics of phase formation in this study. The and -TiAl curves shown in Figs. and 7 begin from the origin, increase rapidly in the initial heating time and reach the curve maximum. After reaching their maxima, the and -TiAl curves decrease gradually because the Al-atoms diffuse outward to form the Al O 3 layer and diffuse inward to increase the Al concentration in the substrate. Obviously, only the portion of and -TiAl curves from the origin to their maximum is related to the phase formation and can be used to analyze the growth kinetics, as shown in Fig. for formation and in Fig. 9 for -TiAl formation. In Figs. and, the dependence of the layer thickness is plotted against the square root of the interreaction time for film on -TiAl bulk alloy and for film on -Ti 3 Al one, respectively. The heating temperatures are C, 9 C and 1 C for Fig., and 7 C and C for Fig.. In all heating temperatures of Fig., each set of data points can be fitted well to straight lines. This characteristic behaves a diffusioncontrolled process for phase growth and complies fairly well with the parabolic law. The same situation also occurs in Fig. 9 for -TiAl formation in film on bulk -Ti 3 Al system. Therefore, the growth constants from the slopes of these straight lines can be determined, as discussed in detail below. Notice also from Fig., that all fitted straight lines do not intercept the origin of the plot. This may be because a few phases have nucleated at the interface before the specimens were heated up to the set temperature. However, all straight lines shown in Fig. 9 pass near the origin of the plot. This feature indicates that -TiAl formation may be not nucleated during the heating and it only needs a little incubation time at the set temperature, which is quite different from the behavior of formation at the same temperature. From Figs. and 9, the growth kinetics of and - TiAl phases in film/bulk -TiAl and in film/ bulk -Ti 3 Al all belong to diffusion-controlled process. There are Arrhenius p plots of the growth constant, k, which is defined by W ¼ k ffiffi t, where W is the thickness of product layer and t the interreaction time. Hence, the growth constants can be determined from the slopes of the straights lines shown in Figs. and 9. The temperature dependence of k can be expressed by the Arrhenius equation: k ¼ k o expðq k =RTÞ ð1þ where Q k is the activation energy of phase growth, k o is a pre-exponential factor and R is the gas constant. The logarithm of the growth constant k is plotted as a function of reciprocal temperature (1=T), as shown in Fig. 1. Figure 1 shows that the straight lines can be fitted well to the data points, thus the growth of and -TiAl is thermally activated, but with different Q k and k o. From Fig. 1, the parameters Q k and k o were calculated from eq. (1) and are listed in Table 3. From Table 3, for the film/ bulk -TiAl diffusion couple, the Q k of phase formation is about 15.9 kj/mol, and the k o is approximately 3:9 1 m /s at 1 C. In the case of film/bulk -Ti 3 Al diffusion couple, the Q k and k o of - TiAl and phases formation are about 13. kj/mol and Growth Constant, k / m. s -1 1E-9 1E-1 1E-11 1E-1 / (k TiAl ) /α 3 -Ti 3 Al (k TiAl ) /α 3 -Ti 3 Al (k γ -TiAl ).75..5.9.95 1. 1.5 1.1 T -1 / 1-3 K -1 Fig. 1 Arrehenius plots of the growth constants versus 1=T (K 1 ) for and -TiAl phases. The growth constant data are calculated from Fig. and Fig. 9. Table 3 Data of various product phases in Ti-Al diffusion couples. Formed Q k k o or D o Temperature range Diffusion couples Phase (kj/mol) (cm /s) ( C) Reference No. bulk Al-bulk Ti 179:5 5 1 3 5 bulk -bulk 3:5 1 7 95 7 bulk Ti 5 Al 5 -bulk Ti Al 5 95 1 : 1 3 131 bulk Ti 7:5 Al 7:5 -bulk Ti Al 3 31 1 1 3 9 1131 Al film-ti film 1 1.7 515 9 Al film-ti film 15 1 3: 1 3 375 5 1 Al film-ti film 173 1 3 5 11 Al film-ti film 15 1 :5 1 35 5 1 film-bulk -TiAl 15.9 3:9 1 1 This work film-bulk -Ti 3 Al 17.9 1: 1 7 9 This work -TiAl film-bulk -Ti 3 Al 13. : 1 7 9 This work
19 M.-S. Chu and S.-K. Wu : 1 m /s, and 17.9 kj/mol and 1: 1 m /s at 79 C, respectively. Data obtained from this study and more obtained from published reports are listed in Table 3 for comparison. From Table 3, the Q k values of and -TiAl phases formation in this study are very similar in magnitude with those of the phase in the Ti-Al thin film diffusion systems, 9 1) but are lower than those of Ti-Al product formation in Ti-Al bulk systems. ) A possible explanation in Q k for this discrepancy could be that the intrinsic defects, voids and the residual stress inherent in the thin film are more than those in the bulk form.. Conclusions In this study, film/bulk -TiAl and film/bulk -Ti 3 Al diffusion couples are obtained by sputtered-al film on -TiAl or -Ti 3 Al substrates and then annealing at C for h. Subsequent interreaction at 71 C for 3 min3 h in the ambient atmosphere shows that layer and /-TiAl mixed layers are observed at the interfaces of film/bulk -TiAl and film/bulk - Ti 3 Al diffusion couples, respectively. All growth kinetics of and -TiAl phase formation obey a parabolic law. Hence, the growth activation energy, Q k, of product phases, as well as their pre-exponential factor, k o, can be determined by an Arrhenius equation. Experimental results show that Q k of phase foramtion is about 15.9 kj/mol with k o of 3:9 1 m /s in film/bulk -TiAl system at 1 C, and those of -TiAl and phases formation are 13. kj/mol and 17.9 kj/mol with k o of : 1 m /s and 1: 1 m /s in the film/bulk -Ti 3 Al system at 79 C, respectively. In this study, the Q k values of and -TiAl phase formation are similar to the reported Q k values of phase formation in Ti-Al thin film diffusion couples. At the same time, none of the fitted straight lines of thickness vs. the square root of the interreaction time intercepts the origin of the plot. This implies that only a few phases had nucleated at the interface prior to the specimen being heated up to the set temperature. However, all fitted straight lines of -TiAl thickness vs. the square root of the interreaction time for the film on -Ti 3 Al alloy pass near the origin of the plot. This indicates that -TiAl formation may be not nucleated during the heating and it needs only a brief incubation time at the set temperature. Acknowledgements The authors gratefully acknowledge the financial support for this research provided by the National Science Council (NSC), Taiwan, Republic of China, under Grant No. NSC91- CS-7--. REFERENCES 1) F. Appel, P. A. Beaven and R. Wagner: Acta Metall. Mater. 1 (1993) 171 173. ) Y. W. Kim: J. Metals (199) 3 39. 3) C. T. Liu in: I. Baker, R. Darolia, J. D. Whittenberger and M. Yoo, eds.: High-Temperature Order Intermetallic Alloys V, (MRS, 1993) pp. 3 19. ) T. Tetsui and S. Ono: Intermetallics 7 (1999) 9 97. 5) S. Wöhlet and R. Bormann: J. Appl. Phys. 5 (1999) 5 3. ) F. J. J. van Loo and G. D. Rieck: Acta Metall. Mater. 1 (1973) 1 71. 7) F. J. J. van Loo and G. D. Rieck: Acta Metall. Mater. 1 (1973) 73 5. ) W. Sprengel, H. Nakajima and H. Oikawa: Mat. Sci. Eng. A 13 (199) 5 5. 9) X. A. Zhao, F. C. T. So and M. A. Nicolet: J. Appl. Phys. 3 (19) 7. 1) M. Wittmer, F. Le Goues and H. C. W. Huang: J. Electrochem. Soc. 13 (195) 15 155. 11) I. Krafcsik, J. Gyulai, C. J. Palmström and J. W. Myer: Appl. Phys. Lett. 3 (193) 115 117. 1) J. Tardy and K. N. Tu: Phys. Rev. B 3 (195) 7 1. 13) R. Bormann: Mater. Res. Soc. Symp. Proc. 19 (199) 33 3. 1) Y. Mishin and Chr. Herzig: Acta Mater. () 59 3. 15) M. S. Chu and S. K. Wu: Acta Mater. 51 (3) 319 31. 1) M. S. Chu and S. K. Wu: Surf. Coat. Technol. 179 () 57. 17) R. Pretorius, T. K. Marais and C. C. Theron: Mater. Sci. Eng. R 1 (1993) 1 3. 1) ASM Handbook, Vol. 3, Alloy Phase Diagrams, (ASM International, Metal Park, OH, 199) 5.