of the Less-Common
CHEMICAL VAPOUR DEPOSITION OF RHENIUM ON GRAPHITE Y. ISOBE, M. TANAKA, Department Yamadaoka
of Nuclear Engineering, Faculty 2-1, Suita, Osaka 565 (Japan)
(Received September 13,1988;
and M. MIYAKE of Engineering,
in revised form October 30, 1988)
Summary Adhesive and dense rhenium coatings were fabricated on graphite substrates by thermal decomposition of Re,(CO)i,,. The rhenium coatings possessed columnar grains at deposition temperatures above 600 “C and fine fibrous grains below 500 “C. The degree of crystallinity increased as the deposition temperature increased. The (110) preferred orientation was favoured above 600 “C. Hardness and grain size of the coating strongly depended upon the deposition temperature. Annealing above 1000 “C caused recrystallization of the coatings. All the annealed coatings showed no thermal damage, such as ablation or microcracking, at 800 - 1100 “C. Lattice parameters and microhardness increased after the annealing due to carbon diffusion from the graphite substrate. However, no carbide layer growth was observed at the interface between the rhenium coating and graphite substrate.
1. Introduction In order to realize wider applications of graphite as an engineering material, it is necessary to improve such surface properties such as its corrosion resistance and its gas permeation characteristics. Protective coatings may offer a means of modifying the surface to meet the requirements for a variety of operating environments. From this point of view, we have studied the deposition of some refractory metals and compounds on graphite [l - 31. Of the refractory metals, rhenium has desirable corrosion and thermal resistance and may be used as a barrier for preventing diffusion of carbon because rhenium carbide can be produced above 800 “C at an extremely high pressure, above 60 kbar [ 41. In the present study, chemical vapour deposition (CVD) of rhenium was therefore developed for graphite substrates and the coatings prepared were characterized. Although most of the present-day investigations on rhenium coatings are related to halide decomposition processes [5,6], the method chosen for this study was a carbonyl decomposition process which enabled us to reduce the deposition temperature and to 0022-5088139i$3.50
@ Elsevier Sequoia/~inted
in The Netherlands
produce the coating without damage to the substrates because of the noncorrosive nature of Re2(CO)i0 and its decomposition products, rhenium and CO. This paper also describes the annealing behaviour of the coating at elevated temperatures.
2. Experimental details Isotropic graphite sheets (10 mm X 20 mm X 1 mm, ISOGRAPH-88, Toyo Tanso Ltd) and Re2(CO)i0 powder of purity 99 wt.% were used as substrates and as a precursor respectively. Prior to the CVD of rhenium, the graphite substrates were mechanically polished and then degassed at 1100 “C for 2 h in a vacuum below 10m4Pa. The deposition of rhenium was carried out in a conventional CVD system at temperatures ranging from 500 to 700 “C and with the precursor at a temperature of 70 “C. Coating layers of 6 pm thickness were employed for all the experimental runs. For the coatings prepared at the deposition temperature of 700 “C, isothermal annealing was performed at temperatures from 800 to 1100 “C in a vacuum below lop4 Pa. The annealing period was between 0.5 and 9 h. Microstructures of as-deposited and annealed rhenium coatings were characterized by means of scanning electron microscopy, surface profilometry and microhardness measurements. In order to elucidate the crystallographic properties such as orientation and lattice parameters, X-ray diffraction techniques were also employed.
3. Results and discussion 3.1. Characterization of rhenium coatings The fundamental characteristics for the coatings prepared are summarized in Table 1 as a function of the deposition temperature. The deposition rates of all these coating experiments were about 1.0 pm h-’ in spite of their different deposition temperatures. This suggests that the Re,(CO)ia vapour above 500 “C decomposed completely into rhenium and CO under our deposition conditions. Results obtained by X-ray diffraction analysis have indicated that all the coatings were composed of metallic rhenium and that no carbide was present in the coatings. The coating prepared at the deposition temperature of 500 “C exhibited a broadened diffraction pattern without resolved doublets. At higher deposition temperatures, the diffraction patterns became sharpened. As shown in Table 1, the half-width estimated from the (110) diffraction peak tended to decrease as the deposition temperature increased. It is evident from the diffraction patterns and the half-widths that the degree of crystallinity increased with increasing deposition temperature. It can be seen from Table 1 that the crystal orientation of the rhenium deposited depended on the deposition temperature. At a deposition tempera-
TABLE 1 Characteristics of rhenium coatings prepared at various deposition temperatures Deposition
Deposition rate (pm hh’) Phase Orientation Half-width (deg)b Lattice parameters a (A) c (A) Grain size (pm) Roughness (pm) Hardness (kg mm-L)C
1.0 Re _a
0.6 0.3 (approx.) 0.39 1478
0.3 2.764 4.457 0.81 0.63 322
0.3 2.764 4.457 1.12 1.18 254
aThe coating showed a random texture. bEstimated from the (110) diffraction peak. CVickers microhardness number obtained at a load of 25 g.
ture of 500 “C, the coating revealed a random texture, the diffraction pattern of which was almost the same as that of rhenium powder. For the coatings prepared at 600 and 700 “C, some growth texture occurred, which tended to be only in the (110) preferential orientation. These results indicate that at higher deposition temperatures the rhenium coatings had a (110) preferred orientation. Certain types of orientation in deposited materials have been thought to originate from the initial stage where a densely populated atomic plane forms parallel to the substrate surface [7,8]. Although the (001) preferred orientation is expected to be formed in the case of hexagonal close-packed metal, this prediction is inconsistent with the present results. The lattice parameters of the coatings prepared at 600 and 700 “C were found to be a = 2.764 A and c = 4.457 A, which are almost the same for the published data for pure rhenium powder. It can therefore be concluded that at higher deposition temperatures pure rhenium with (110) orientation is deposited on the graphite substrate. Figures l(A), l(B) and l(C) show surface microstructures of the rhenium coatings prepared at different deposition temperatures. At a deposition temperature of 500 “C, the surface had a characteristic domed surface composed of fine grains as indicated in Fig. l(A). The typical faceted surface could be detected for the coatings prepared at higher deposition temperatures (Figs. l(B) and l(C)). The grain size calculated from the number of grains observed in a known area of the surface increased from about 0.3 to 1.12 E.trnwith increasing deposition temperature, as listed in Table 1. These surface features are consistent with the results for the surface roughness measurement of the coatings. With an increase in deposition temperature, the surface roughness changed from 0.39 to 1.18 ,um. Cross-sectional micrographs (Figs. l(a) - l(c)) reveal that the coatings adhering to the graphite substrate were obtained without voids or cracks. It can be seen from Fig. l(a) that the coating obtained at 500 “C consisted of
Fig. 1. Surface ((A), (B), (C)) an d cross-sectional ((a), (b), (c)) scanning electron micrographs of rhenium coatings prepared at various deposition temperatures: (A), (a) 500 “C; (B), (b) 600 “C; (C), (c) 700 “C.
fine fibrous grains. At higher deposition temperatures of 600 “C and 700 “C, clearly defined columnar grains with faceted ends were observed. It was confirmed by the X-ray diffraction analysis that the coatings with the columnar structure were highly oriented. Movchan and Demchishin  found that the microstructures of metal coatings prepared by evaporation with electron beam heating could be represented as a function of T/T, by three zones, where T and T, indicate the substrate temperature and the melting point of the coating material, respectively. Zone 1 (T/T,< 0.3)consists of tapered crystallites with a characteristic domed surface; zone 2 (0.3 < T/T,,., < 0.45-0.5)consists of columnar grains with a smooth matt surface; and zone 3 (T/T, > 0.45-0.5) consists of equiaxial grains and a bright surface. In our study of CVD of rhenium, the columnar grains appear above 600 “C, which corresponds to zone 1. This inconsistency with the three-zone model is probably due to the
fact that the deposition rate of rhenium by CVD is lower (about 300 times) and the fabrication techniques of the coatings are different. The results of the microhardness measurement on the rhenium coatings are given in Table 1. The microhardness of the coatings depended strongly on the deposition temperature. For the coating prepared at 500 “C, the microhardness was 1478 kg mm -‘, but it decreased with increasing deposition temperature. The hardness of 254 kg mme2 for the coating prepared at 700 “C is in good agreement with the reported value of 279 kg mmh2 for bulk rhenium [lo]. Syrkin et al. [ll] have reported that for rhenium coatings fabricated by a carbonyl method the microhardness decreased with increasing deposition temperature. It is known that in a carbonyl process low deposition temperatures favour a high carbon content in the deposit. Consequently, the observed decrease in hardness is mainly due to a decrease in the carbon content in the coating. The extremely high value of microhardness at the deposition temperature of 500 “C suggests that the carbon content in the coating exceeds the maximum amount (11.7 at.%) of soluble carbon in rhenium proposed by Hughes [ 121. 3.2. Post-deposition annealing of rhenium coatings The coatings prepared at the deposition temperature of 700 “C were subjected to isothermal annealing, and changes in their microstructure are illustrated in Fig. 2. Annealing below 900 “C for 9 h brought about no change in surface and cross-sectional structure of the coating as shown in Figs. 2(A) and 2(a). At an annealing temperature of 1000 “C, recrystallization appeared to begin after 9 h (Figs. 2(B) and 2(b)), while no morphological change could be observed after annealing for 1 h. As illustrated in Figs. 2(C) and 2(c), for annealing at 1100 “C, the faceted surface for the asdeposited specimen disappeared and the columnar grains were annihilated during grain growth. However, intermediate layer growth did not take place at the interface between the rhenium coating and the graphite substrate and microstructural observation provided no evidence of ablation or microcracking. Annealing above 1000 “C yielded an appreciable reduction in the surface roughness of the coatings. The roughness values of the coating surfaces annealed at 1000 “C and 1100 “C for 9 h were 1.05 and 0.89 pm respectively. It was verified from these results that recrystallization commenced above 1000 “C for rhenium coatings prepared by a carbonyl decomposition process. Belomytsev and Yaroshevich  found that vacuum-condensed rhenium recrystallized at 1200 “C for 1 h. Sims et al. [lo] have given information on the recrystallization of bulk rhenium and reported that annealing at 1100 “C for 1 h caused an abrupt drop in hardness and that recrystallization and grain growth occurred completely at temperatures ranging from 1300 to 1500 “C. There is no marked difference in the recrystallization behaviour of rhenium as determined by us and by others. This means that the structural state (e.g. lattice defects and residual stress) in the rhenium coating is similar to that of bulk rhenium.
Fig. 2. Surface ((A), (B), (C)) and cross-sectional ((a), (b), (c)) scanning electron micrographs of rhenium coatings (depositiori temperature 700 “C) after annealing at various temperatures for 9 h: (A), (a) 900 “C;(B), (b) 1000 “C; (C), (c) 1100 “C.
The X-ray diffraction patterns for all the coatings annealed at 800 1100 “C indicated only the presence of rhenium metal; no rhenium carbide phase was detectable after annealing. The degree of crystallinity of the coatings increased as the annealing temperature or period increased. The (110) orientation in the coating was thermally stable, at least at 1100 “C for 9 h, even though considerable re~~s~l~~tion occurred. Figure 3 shows the dependence of the lattice parameters a and c and the c:a ratio for the rhenium coating on the annealing period at various temperatures. As shown in this figure, the values of a and c increase with the annealing period, whereas the value of c:a decreases with the annealing period. It should be noted that the lattice parameters gradually changed at annealing periods longer than 1 h. The extent of increase in the lattice parameters due to annealing was larger at higher temperatures. Microhardness measurements were performed for the coatings annealed at 1000 “C. The value of microhardness varied from 254 to 429 kg mm-*
Fig. 3. Change in the lattice parameters a and c and c:u ratio of rhenium coatings with annealing period at various temperatures: A, 1100 “C; 0, 1000 “C; A, 900 T; 0, 800 “C; *, pure rhenium powder.
with an annealing period of 0 - 9 h. The hardness increased rapidly with the annealing period until no more changes in hardness took place after 1 h. The trend of change in hardness with annealing period resembles the observation of the increase in lattice parameters with annealing period. These changes in lattice parameters and microhardness after annealing are attributable to dissolution of carbon into the rhenium coating, which is associated with carbon diffusion from the graphite substrate and carbon solubility in the rhenium coating, because the presence of carbon dissolved in rhenium leads to an expansion of the lattice and an increase in hardness 1121. The values of lattice parameters and microhardness obtained suggest that the rhenium coating might contain a maximum of 4 at.% carbon after annealing at 1100 “C, based on the data given in the literature . From the results shown in Fig. 3, the equilibrium concentration of carbon in the rhenium coating appears almost to be attained in the initial period of annealing. Since, in the temperature range employed in the present study, the bulk diffusion coefficient of carbon in rhenium is too small to attain the equilibrium concentration within a short period, grain boundary diffusion of carbon may contribute considerably to carbon migration in the rhenium coating.
4. Conclusion The following conclusions were drawn from the present study. (i) Adhesive and dense rhenium coatings could be fabricated on graphite substrates by thermal decomposition of rhenium carbonyl. The
coating possessed fine fibrous grains with a domed surface at a deposition temperature of 500 “G and columnax grains with a faceted surface at deposition temperat~es of 600 ‘% and 700 “G. The degree of c~sta~~inity increased as the deposition temperature increased. The (110) preferred orientation was favoured by higher deposition temperatures. The hardness increased with decreasing deposition temperature, and the grain size and surface roughness increased with increasing deposition temperature. (ii) The mi~ros~~ct~e of the rhenium coating prepared at the deposition temperature of 700 “G was unaltered by annealing at temperatures of 800 “G and 900 “G. Annealing above 1000 “G resulted in a structural change in the coating due to recrystallization. However, all the annealed coatings showed no damage from exfoliation or microcracking. After annealing, there was no formation of carbide at the interface between the coating and the substrate, while the lattice parameters and microhardness increased, This was because of diffusion of carbon from the graphite substrate into the rhenium coating.
Acknowledgment This research was supported in part by a grant-in-aid for fusion research from the ninety of Education (Japan).
References 1 2 3 4 5 6 7 8 9 10 11 12 13
M. Miyake, Y. I-Iirooka, R. Imoto and T. Sane, Thin Solid Films, 63 (1979) 303. M. Miyake, Y. Hirooka, T. Imoto and T. Sano, Thin Solid Films, 79 (1981) 75. S. Yamanaka, I-I. Ohara, P. Son and M. Miyake, J. Nucl. Mater., 123 (1984) 1304. S. V. Popava and L. E. Boiko, High Temp. High Pressures, 3 (1971) 237. F, A. Glaski, Tkermionic Conversion Spec. Gonf., 1970, p. 128. L. Yang, R. G. Hudson and J. J. Ward, in F. A. Glaski (ed.), Proc. 3rd Int. Conf, cm Chemical Vapor Deposition, Hinsdale, IL, 1972, American Nuclear Society, p. 253. K. R. Dixit, Phil. Mug., 16 (1933) 1049. D. M. Evans and N. W~lrn~, Acta Crysta~~ogr., 5 (1952) 731. B. A. Movchan and A. V. ~mchishin, Fiz. Metal. ~etalioved., 28 (1969) 653. C. T. Sims, C. M. Craighead and R. I. Jaffee, d. Met., 7 (1955) 168. V. G. Syrkin, A. A. Uel’skii, R. I. Akmeeva and L. N. Romanova, Zh. prikl. Skim. ~~e~~~d~, 45 (1972) 2261. J. E. Hughes, J. Less-Common Met., I (1959) 377. Y. S. Belomytsev and P. Y. Yaroshevich, Fiz. Metal. Metalloved., 33 (1972) 758.