6

Click here to load reader

Low temperature chemical vapor deposition of TaB2

Embed Size (px)

Citation preview

Page 1: Low temperature chemical vapor deposition of TaB2

Thin Solid Films, 72 (1980) 517-522 © Elsevier Sequoia S.A., Lausanne--Printed in the Netherlands 517

LOW T E M P E R A T U R E CHEMICAL VAPOR DEPOSITION OF TAB2*

E. RANDICH Sandia National Laboratories, Albuquerque, N.M. 87185 (U.S.A.) (Received April 7, 1980; accepted April 22, 1980)

Crystalline TaB 2 was deposited using the chemical vapor deposition reaction of TaC15 and B2H 6 in the temperature range 773-1200 K. Thermodynamic calculations were made which compared the use of both B2H6 and BC13 as boron source gases. The deposits obtained with B2H 6 had an extremely small crystal size and contained amorphous boron when the deposition temperature was below approximately 873 K but were substoichiometric in boron above this temperature. Carbon analysis indicated that carbon may be substituting for boron and thereby stabilizing the diboride structure at high deposition temperatures. The microhard- ness of the coatings decreased with increasing atomic ratio of boron to tantalum and decreasing crystal size.

1. INTRODUCTION

TAB2, like most diboridesof the group IV and V metals, is a refractory hard compound 1, 2. Unlike the other diborides, however, the chemical vapor deposition (CVD) of TaB 2 has been relatively difficult to accomplish. Early attempts using the hydrogen reduction of TaC15 and BC13 were not successful and resulted in the deposition of single-phase tantalum or unsatisfactory multiple-phase deposits 3. Thermal decomposition of the bromides was used to deposit single-phase TaB 2 at temperatures above 1773 K 4. Recent studies by Motojima and Sugiyama 5 have successfully used the H 2 reduction of TaC15 and BC13 to obtain single-phase TaB 2 deposits on SiO 2 substrates. These investigators used very low flow rates (not greater than 0.4 ml s-1) in the temperature range 1123-1473 K. These flow rates were one to two orders of magnitude lower than the rates other investigators have used and they resulted in the deposition of single-phase TaB 2. To produce TaB 2 on a commercial scale, these flow rates would have to be increased considerably. Other attempts including our own attempts to obtain TaB 2 at higher flow rates using H 2, TaCI 5 and BC1 a have not been successful 3' 6

A logical alternative to producing TaB 2 at low temperatures (less than 1200 K) is to change the source gases. TaC15 is a good tantalum source gas because it is relatively easy to make using direct chlorination of tantalum metal. Diborane

* PaPer presented at the International Conference on Metallurgical Coatings, San Diego, California, U.S.A., April 21-25, 1980.

Page 2: Low temperature chemical vapor deposition of TaB2

518 E. RANDICH

(B2H6) is a logical choice as a boron source to replace BCI 3 because it is thermally unstable in the temperature range of interest and it produces less corrosive byproducts than BCI 3 7. The basic CVD reaction is

mr TaC1 s + B2H 6 , TaB 2 + 5HCI + 1H 2 (1)

The purpose of this study is to determine the feasibility of reaction (1) in the temperature range 723-1200 K.

2. T a - B - C I - H - A r PHASE DIAGRAM

A section of the T a - B - C I - H - A r equilibrium phase diagram was calculated for the standard CVD conditions (for a hydrogen-to-chlorine atomic ratio of 0.76 and an argon-to-hydrogen atomic ratio of 4.17) using the method of Randich and Gerlach 8. This diagram for 800 K and our laboratory atmospheric pressure of 8.4 x 104 Pa is shown in Fig. 1. Bulk compositions (or CVD operating points) for two reactant gas choices are shown. The square represents a mixture of TaC15, B2H 6 and argon (the chosen standard condition) whereas the circle represents a mixture of TaC15, BC13, H 2 and argon. Application of the lever rule shows that the gas mixture containing B2H 6 would be more supersaturated in TaB2 than would the gas mixture containing BC13. This suggests that a larger equilibrium yield will result for the gas mixture containing B2H 6 (approximately 3.3~o compared with approximately 1.09/o for BCI3). Therefore, from thermodynamic considerations, B2H 6 is a better choice for a boron source than BC1 a.

3. EXPERIMENTAL

The CVD apparatus is similar to the apparatus that was used previously 6 to deposit (Ti,Zr)B 2 and (Ti,Ta)B 2. It consists of a fused silica reactor 6.5 cm in diameter, a separate tantalum chlorinator and a mixing chamber for the B2H 6 and argon carrier gas. TaC15 is produced by direct chlorination of tantalum chips at 575 K in a fused silica chlorinator 3 cm in diameter and 25 cm in length. All TaCls passages to the reactor ate maintained at 575 K to prevent condensation of the chloride. BzH 6 decomposes rapidly and polymerizes to higher boranes at temperatures slightly above ambient and cannot be heated before entering the reaction zone. It is therefore brought in at room temperature via a separate chamber where it is mixed with argon as a diluent carrier gas. The substrates and the susceptor were Poco-AXF 5Q graphite and were heated with an r.f. generator. The temperature was controlled to - 5 K with a P t - (P t -10~Rh) thermocouple inserted into the susceptor. It should be noted that B2H 6 requires special handling and careful CVD procedures because it can be explosive and it is toxic. All venting of gas mixtures containing B2H 6 was through an alcohol bubbler to react any remaining B2H 6 to B203 and H20. The system was thoroughly purged before and after the CVD reaction using argon. The standard conditions were a boron-to-tantalum atomic ratio of 1.33, a hydrogen-to-chlorine atomic ratio of 0.76 and an argon-to- hydrogen atomic ratio of 4.17; the total gas throughput was 20 ml s- 1. Preliminary measurements showed that the chlorination of tantalum was not complete and some

Page 3: Low temperature chemical vapor deposition of TaB2

LOW TEMPERATURE CVD OF T a B 2 519

unreacted C12 w a s passed into the system. The chlorinator output had a chlorine-to- tantalum atomic ratio of 5.64 + 0.15, indicating an 88.6% conversion efficiency of CI 2 to TaCI v

Analysis techniques included optical metallography, scanning electron microscopy and X-ray diffraction. The Vickers microhardness was measured at a load of 50 gf using a Shimadzu M tester. Carbon analysis was performed with an ARL-EMX electron beam microanalyzer equipped with a diffraction grating having a line spacing of 600 mm-t . The boron-to-tantalum atomic ratios were determined by atomic emission spectroscopy using inductively coupled plasma (ICP) excitation. The ICP samples were prepared by dissolving the deposits in a solution of 839/0 H N O 3 and 17% HF acid. Commercially available standards of tantalum and boron were made in a similar fashion.

4. RESULTS AND DISCUSSION

TaB 2 coatings were obtained in the temperature range 773-1200 K accord- ing to reaction (1). The coatings were fully dense and adherent on graphite,

S ° : ^ = o f e o o oo

(a)

, ~ Ta-B-CI-H-Ar ~ J [ soo K (b)

f m 0

Fig. 1. Section of the T a - B - C I - H - A r phase diagram at 800 K for standard CVD conditions: hydrogen- to-chlorine atomic ratio, 0.76; argon-to-chlorine atomic ratio, 4.17; 0.84 bar; I , bulk composition for B2H 6 source gas; O, bulk composition for BCI 3 source gas.

Fig. 2. X-ray diffraction profiles for deposition temperatures of (a) 773 K, (b) 873 K and (c) 973 K (Cu Ku radiation).

Page 4: Low temperature chemical vapor deposition of TaB2

520 E. RANDICH

tantalum, Kovar and A120 a but not on low carbon steel. Graphite was chosen as the standard substrate material to facilitate interpretation of X-ray diffraction data. Deposition rates varied from 6.0 x 10 -3 mg cm -2 s -1 at 773 K to 9.8 x 10 -3 mg cm- 2 s- 1 at 1023 K.

At low deposition temperatures the TaB2 coatings had an extremely fine crystallite size but, as the temperature increased, so did the crystaUinity (Fig,. 2). The X-ray diffraction peak broadening is substantial and can be due to either a fine grain size or a non-uniform strain in the coating 9. The temperature dependence of the broadening shown in Fig. 2 most probably indicates very small crystallites. Use of the Scherrer formula, which is strictly applicable to strain-free materials only, should give a reasonable estimate of the crystallite size. At 773 K this estimate is approximately 50 A whereas at 973 K it is approximately 105 A. These extremely fine grain sizes correlate well with the observation that attempts to etch the coatings metallographically revealed no grain boundary structures. The peak broadening, e.g. the crystallinity of the coating deposited at 773 K, is equivalent to that observed by Motojima and Sugiyama 5 at 1223 K using BCI 3 as a boron source gas. At 1073 K these investigators observed only tantalum deposition, so there is no direct comparison with results obtained at lower temperatures. The present study indicates a high degree of crystallinity at temperatures of 973 K and above.

The boron-to-tantalum atomic ratios measured using ICP analysis varied with deposition temperature (Fig. 3). At temperatures below 873 K the coatings contained codeposited boron which from X-ray analysis appears to be amorphous. Above 873 K the coatings were substoichiometric in boron. The Ta-B binary phase diagram indicates stoichiometry limits ~° for TaB2 from TaB~.s2 to TaB2.o2. The present study shows the existence of the diboride structure down to TaBL42 (Fig. 3)--clearly outside the expected limits. However, X-ray results for the substoichiometric coatings show only TaB2 peaks. Many of the boride-carbide- nitride systems of the transition metals are solid solution seriesXl; therefore, impurity elements such as carbon, oxygen or nitrogen may be substituting for boron on the boride lattice and may stabilize the diboride structure. Of these, carbon is the most probable candidate because of its abundance in the CVD syste m as the substrate and susceptor. Carbon contents of the coatings are shown in Fig. 3. The results show a definite trend: :the carbon content increases as the boron content decreases. Quantitatively the carbon content does not compensate entirely for the boron deficiency. For example, at 923 K the measured boron-to-tantalum atomic ratio is 1.59. To have;sufficient substitutional carbon to meet the lower stoichiometry limit of Ta(B,C)Ls2 (e.g, TaBL59Co.23) requires 1.3 wt.% C (8.2 at.% C). The measured amount is approximately 0.65 wt.% C (4.1 at.% C) or only half of the necessary~ amount. Similarly at 1023 K the measured amount is approximately 1.0 wt.% C or less than half of the necessary 2.4 wt.% C needed to obtain Ta(B,C)Ls2. The lack of a sufficient amount of a substitutional element is attributed to the semiquantitative nature of the carbon analysis as well as to the possible presence of other substitutional species.

The hardness of the coatings is a strong function of deposition temperature (Fig. 4). The reported hardness ~2 of TaB2 is 2500 HV. Deposits produced above 873 K show good agreement with the results of other investigators 5. Below 873 K the hardness decreases rapidly. This is coincidental with boron-to-tantalum atomic

Page 5: Low temperature chemical vapor deposition of TaB2

Low TEMPERATURE CVD OF TaB, 521

ratios in excess of 2.0, e.g. with the appearance of amorphous boron in the TaB, and with the decrease in crystallinity of the deposits (Fig. 2).

I 1 , 4 4

800 9w Irm 1100 1mJ

DEPOSITION TEMPERATURE IKI

7w 800 9&l IKIO 1100 1700

DEPOSITION TEMPERATIJRE IKI

Fig. 3. Effect of deposition temperature on the boron-to-tantalum atomic ratio (0) and the carbon impurity content (0).

Fig. 4. Effect of deposition temperature on hardness.

5. CONCLUSIONS

Thermodynamic considerations show that the use of either BCl, or B,H, as the boron source gas should result in the deposition of TaB, in the temperature range of interest. Since the use of BCl, is successful only at very low flow rates, the reaction kinetics in a flowing CVD reactor must play an important role in determining which condensed phases form in the Ta-B-Cl-H system. Experimental results show that using B,H, results in rapid deposition of hard crystalline TaB, at temperatures where iron-based alloys can be used as substrates. Under certain conditions, impurities such as carbon which are substitutional for boron apparently stabilize the diboride lattice well beyond the expected limits of B-Ta stoichiometry. Such impurities do not affect the hardness although lack of crystallinity and codeposited amorphous boron substantially decrease the hardness. The results show that B,H, is an excellent replacement for BCl, as a boron source for the CVD of TaB, and could possibly be used for the CVD of other refractory metal borides.

ACKNOWLEDGMENT

This work is supported by the CMice of Fusion Energy, U.S. Department of Energy, under Contract DE-ACO4-76-DPOO 789.

REFERENCES

1 C. F. Powell. in I. E. Campbell (ed.), Higk Temperature Technology, Wiley, New York, 1956, p. 131. 2 L. Kaufman and E. V. Clougherty, Proc. 5th Plansee Semin., Reutte, Tyrol, 1965, Springer, Vienna,

1966, p. 722. 3 C. F. Powell, J. H. Oxley and J. M. Blocher, Jr. (eds), Vapor Deposifion, New York, 1966. 4 B. Armas, C. Combescure and F. Trombe, f. Electrochpm. Sot., 123 (2) (1976) 308.

Page 6: Low temperature chemical vapor deposition of TaB2

522 E. RANDICrl

5 S. Motojima and K. Sugiyama, J. Mater. Sci., 14 (1979) 2859. 6 E. Randich, Thin Solid Films, 63 (1979) 309. 7 P. Casadesus, C. Frantz and M. Gantois, Metall. Trans. A, 10 (1979) 1739. 8 E. Randich and T. M. Gerlach, Rep. SANDSO-0308, 1980, Sandia Laboratories, Albuquerque, New

Mexico. 9 B.D. CuUity, Elements of X-ray Diffraction, Addison-Wesley, Reading, Massachusetts, 1956.

10 F.A. Shunk, Constitution of Binary Alloys, McGraw-Hill, New York, 1969, Suppl. 2, p. 99. 11 P. Schwarzkopf and R. Kieffer, Refractory Hard Metals, Macmillan, New York, 1953. 12 B.W. Mott, Micro-indentation Hardness Testing, Butterworths, London, 1956.