COMBUSTION SYNTHESIS OF ZnS IN

0 downloads 0 Views 969KB Size Report
The self-heating nature of SHS (Self-propagating High-temperature ... the zinc with molten sulfur, which allows the synthesis of a high-density and high-purity product. .... with the zinc powder: (i) initial powder mixture; (2) compacted powder ...
Missing:
Twenty-Fifth Symposimn(International)on Combustion/TheCombustion Institute, 1994/pp. 1651-1657

C O M B U S T I O N SYNTHESIS O F ZnS IN MICROGRAVITY S. GOROSHIN, J. H. S. LEE AND D. L. FROST Department of Mechanical Engineering McGill University 817 Sherbrook Street West, Montreal, Quebec, H3A 2K6, Canada The self-heating nature of SHS (Self-propagating High-temperature Synthesis) makes it particularly suitable for microgravity processing of materials where weight and power requirements are severely restricted. The absence of convection, hydrostatic pressure, and phase separation permits the combustion front dynamics and solidification processes of SHS to be studied under controlled conditions. This paper describes recent ground-based and microgravity (NASA KC-135 parabolic flight) ex~periments on SHS processing of ZnS. A novel technique was used for preparing the precursor mixture of Zn + S by mixing the zinc with molten sulfur, which allows the synthesis of a high-density and high-purity product. The flame speed, quenching diameter, and temperature profile in the flame front and crystal structure of the synthesized samples have been determined. Thermocouple measurement of the temperature profile in the flame front indicates that the thermal thickness of the flame is less than 0.3 mm. The average flame speed is of the order of 7 mm/s, and slightly lower values (~4 mm/s) are observed near the quenching limit. It was found that the flame speed is not stable along the samples with diameters more than 12 mm. The quenching diameter is found to be of the order of 5 mm (in microgravity less than 4 ram). X-ray diffraction data show a wurtzite structure both in ground-based and in-flight synthesized samples, and the lattice's parameters are most similar to the ideal ZnS wlrtzite structure in the outer part of samples synthesized in mierogravity. The ability to provide containerless SHS processing of molten ZnS in mierogravity also has been demonstrated.

Introduction Access to the microgravity environment provided by the use of drop towers, parabolic trajectories of aircraft and rockets, and shuttle and space station orbits has permitted material scientists to investigate a number of problems associated with gravitational effects on material processing. In the absence of gravity, buoyant convection, hydrostatic pressure, and the separation of phases of different densities are eliminated, providing the ability to synthesize new or improved materials. Conventional methods of material processing require a heat source (i.e., a furnace); thus, the maximum temperature and sample size are severely limited in a microgravity environment as a result of the size and operational limits of the furnace. Combustion synthesis does not require an external energy source; hence, it is particularly advantageous in a mierogravity environment where weight, volume, and electrical power requirements must be strictly limited. It is anticipated that containerless material processing and solidification in a microgravity environment will very likely result in the manufacture of materials with a high purity and the possible synthesis of new, unique materials. The use of combustion synthesis eliminates the furnace, allowing for the first time a true containerless environment.

Preliminary studies of Self-propagating, Hightemperature Synthesis (SHS) in mierogravity have already been initiated by Shteinberg and co-workers [1], who demonstrated that TiC produced in microgravity can have a highly porous (95% porosity) structure. Odawara et al. [2] have reported work on thermite reactions in short-duration microgravity experiments, and Moore [3] has investigated the influence of mierogravity on the porosity and structure of combustion-synthesized metal-matrix composites. The present paper describes the SHS studies recently initiated at McGill University on the SHS of ZnS in mierogravity.

General Considerations: The material of interest in our study is ZnS. Zinc sulfide has a variety of luminescent properties (e.g., x-ray, photo-, electro-, as well as triboluminescence) depending on the dopant used, as well as semiconductor and piezoelectric properties. The conventional method of producing fine crystalline or amorphous ZnS powders for luminescent applications is via precipitation from an aqueous solution of zinc sulfate with hydrogen sulfide [4]. Larger single crystals can be obtained from melt or by condensation from the gaseous phase [5,6]. Uniform doping with

1651

1652

MATERIALSSYNTHESIS

small concentrations of impurities and the growth of large crystals are difficult using conventional methods of synthesis [7]; therefore, a suitable industrial technology for fabricating large ZnS single crystals does not currently exist. Zinc sulfide can also be produced by the SHS process. This was first achieved by the group at Odessa in 1986 [8] where a compacted mixture of zinc and sulfur powders reacted to form ZnS. The SHS-synthesized ZnS was found to have a polycrystalline structure with a density of less than 91% of the theoretical value. The structure of the products depends on the combustion and solidification front dynamics, and the purity of the product was not high enough for electronic applications. Therefore, it is necessary to achieve a better understanding of these processes before the desired structure and purity of ZnS synthesized by SHS can be achieved. Both the combustion and the solidification processes are strongly dependent on temperature, and hence on the mode of heat transport, as well as boundary conditions. In mierogravity, convection- and gravity-induced flow of the melt is absent, permitting better observation of the combustion and solidification processes. In addition, in a long-duration mierogravity environment, there is the possibility of growing single crystals from SHS-generated ZnS melt.

Experimental Details Charge Preparation: High-purity zinc ( - 325 mesh) and sulfur ( - 100 mesh) powders (supplied by Aldrich Chemical Company, Milwaukee, WI) were used for preparation of the charges. The powders were mixed in a proportion close to stoichiometric in accordance with the reaction Zn + S --+ ZnS. Since part of the Zn in the zinc powder is in the form of an oxide (ZnO), a small empirically determined excess of sulfur (0.01-0.05%) is added in order to compensate for the oxygen. In order to increase the luminescent efficiency of the synthesized samples, some quantity of regular activators such as Cu (0.01-0.05%) or Mn (0.1-1%) was added to the Zn-S powder mixture. The traditional method of sample preparation in SHS technology is to compact the powder mixture before firing in a steel die by a cylindrical press [9]. This requires a pressing force of up to 50 tons/cm2, but the final density of the sample for Zn + S is typically less than 85% of the theoretical density [10]. Our innovation in SHS sample preparation is to melt the component with the lowest melting point. For the Zn + S, this corresponds to mixing of molten sulfur with zinc powder. This novelty not only greatly simplifies the preparation of samples of different shape and geometry but eliminates trapped gases, improving the quality of the synthesized product.

The difference of the sample structure prepared by cold-pressing and mixing molten sulfur with zinc powder is illustrated in Fig. 1. Eliminating the gases trapped in the pores of the pressed reactant mixture allows us to produce a final product with a density that is very close to the theoretical value. The synthesized ZnS samples have a density not less than 98.5% of theoretical, compared with the best results of 91% obtained with a coldpressing method [8]. It can also be seen from Fig. 1 that the zinc particles in a cold-pressed mixture are concentrated along the deformed borders of the initial sulfur particles. This nonuniformity in zinc concentration after cold-pressing decreases the completeness of the chemical reaction and is the cause of the dependence of the purity of the final product on sulfur particle size [10]. Unlike the cold-pressed powders, a mixture prepared from molten sulfur is highly uniform, and completeness of the chemical reaction is higher.

Reactor and Microgravity Experimental Package: The boiling points of sulfur (717 K) and zinc (1179 K) are less than the temperature in the combustion front (2200 K [8]), and at low pressures, the sublimation temperature of ZnS is lower than the melting point [11]. Therefore, one needs to carry out the process of ZnS-SHS synthesis within a pressurized inert gas environment. The first experiments of ZnS-SH8 synthesis [81 showed that the final product could be synthesized in molten form only if the total pressure in the reactor was more than 50 atm. In general, the completeness of the chemical reaction and the quality of the product are also higher for higher pressures

[81. The SHS-ZnS synthesis equipment consists primarily of a high-pressure reactor vessel (Clover Leaf Reactor, Model CL-9, High Pressure Equipment Company, Erie, PA) capable of withstanding a static pressure >1000 atm. The reactor is rated for a working temperature up to 120 ~ and includes a quickopening cover. A schematic of the experimental flight package and the pneumatic control circuit is given in Fig. 2. The amount of gas in the high-pressure bottle (400 atm) is enough to refill the reactor up to 35 times to the 100 _+ 10-arm working pressure. The high-pressure in-line relief valve prevents the high-pressure manifold from overpressurization, and the low-pressure in-line relief valve prevents overpressurization of the plane's overboard vent manifold. The overboard pressure under regular flight conditions is considerably lower than the pressure inside the cabin. Therefore, a special manifold with a low-pressure gauge is used to equalize the pressure in the reactor and cabin before it is opened. An electric ignition circuit and flame-speed measurement circuit are linked to the reactor cover.

COMBUSTION SYNTHESIS OF ZnS IN MICROGRAVITY

1653

I

Ii~

I

9

x

FIG. 1. Schematic illustrations of the structure of the samples prepared by cold-pressing and mixing of the liquid sulfur with the zinc powder: (i) initial powder mixture; (2) compacted powder mixture; (3) melted sulfur mixed with zinc powder,

Tung + 45V

C1

-q~---

Le

G2 qF-C3

I

C4

q~---

Le,

G5

qp---7-

K-

"1.5 m

-),4

C1-C5 = R 1 - R 5 -" R 6 = 1 .~

FIG. 2. The schematic of the flight experiment package (A) and the pneumatic circuit (B). FIG. 3. The arrangement of the sample for tlame-speed, temperature, and quenching-distance measurements. The microgravity flights are carried out on the NASA KC-135 aircraft. Usually, a total of 40-50 parabolas are flown per flight with a 5-minute break in between each set of 10 parabolas. Retrieving the sample and reloading the reactor with a new charge are carried out during the 5-minute break periods. Henee, about 4 or 5 samples are synthesized per flight.

Samplesfor the Flame-Speedand QuenchingDistance Meca~urements: The final temperature of the synthesized sample is dose to the adiabatic temperature, i.e., 2200 K. To minimize the heating of the gas inside the reactor, each sample was placed inside a heavy-walled aluminum container (12-era long, 8.5-cm diameter), as shown in Fig. 3. The container absorbs almost all of the enthalpy of reaction, and its final temperature does not exceed 150 ~ Figure 3 also shows the five

"ionization" probes used to measure flame speed and the fine thermocouple embedded into the central axis of the charge used to measure the temperature profile as the flame propagates past the thermoeoupie. Charges of the mixture of Zn + S of different diameters are placed in an inner aluminum cylinder, which in turn is inserted into the larger heavy-walled aluminum container. At one end of the Zn + S charge is a pressed pellet ofTi + C, which serves as the ignition source. The Ti + C pellet is ignited by a tungsten wire. Tile initial Zn + S mixture, because of the very high resistance of sullhr, is an insulator. Tile final product, ZnS, is a semiconductor at room temperature and has a high resistance of the order of l0 s10 l~ s cm [6J. However, the molten ZnS within the combustion wave apparently has a conductivity that

MATERIALS SYNTHESIS

1654

approaches that of metals.* We used this phenomenon to design the circuit that is similar to the conventional one used for ionization probe flame-speed measurements of detonation and deflagration modes of propagation in a combustible gas. Unlike the ionization probes used in gases, as a result of the high conductivity of the ZnS melt and the low flame speed, the voltage is comparatively low, i.e., 45 V. The 0.5-ram graphite rods are used as probes and allow easy replacement of the sample. Because of the large charging time of the RC circuit, all of the probes are connected to one channel of the digital oscilloscope. The sequential triggering of the ionization probes signaling the arrival of the flame at each of the probe locations is used to determine the flame speed. For the unconfined charge experiments, the inner aluminum cylinder that surrounds the sample was removed, leaving an air space between the charge and the outer aluminum container wall. In order to synthesize pure samples suitable for x-ray and luminescent analysis, some samples were enclosed in quartz ampoules. The space between the walls of the quartz ampoule and the wall of the aluminum container was filled with 2-3-mm alumina beads. Results

Flame Speed and Quenching Distance in Confined Charges: For confined charges in aluminum cylinders where the heat losses are very high, quenching diameters have been measured. In normal gravity, a minimum diameter of about 4.5-5 mm is required for flame propagation. However, in microgravity, self-sustained propagation is observed for charges less than 4 mm.** The absence of convection in the ZnS melt in microgravity may play a role in reduction in the quenching diameter. In particular, the elimination of convection in microgravity will reduce the heat transfer rate and hence reduce the quenching diameter. The variation of the flame velocity with distance is shown in Fig. 4 for a few charge diameters. For the small-diameter charges, the velocity is found to be very stable, and the flame propagates at 4-5 mm/s. For the larger charge diameters, the velocity fluctuates, indicating a nonsteady flame propagation. In this case, the flame may exhibit a pulsating or spinning instability as observed by other investigators in different SHS systems [9,12]. The phenomena of unstable combustion may be the major source of ina*The conductivity of the molten ZnS has not yet been measured. We did this rough estimation using our own preliminary measurements. **We did not manage to obtain quenching of the flame in microgravity as yet.

~)0

'

''1

\

. . . . . . . . . . . .

I . . . .

~ . . . .

9' :

~ i

i--.-n --~

i ~ ',

10

\

~sm;-

I

-2Sin,', I -12.5mm |

.............. ].............~ i ..................... ~..................... i....... | - 7.5mm I .....

15

g

~ . . . .

~ Oha~.o,am.,.,

............... ~}

!

!1 ' ' - ~ ' ~

i'll

I

i

!.....................:.~-.i

..................... "5,,.. 1 1 '~..................

>

.-~---"-i . . . . .

0

..... 10

~ ............ 20

30

i 40

i--

o...

....

50

60

70

80

Distance (ram)

Fro. 4. Changing of the flame speed along the charge. purities in the combustion synthesis of ZnS. If the SHS product is solid, it consists after unstable combustion of layers of different chemical composition and structure that can be used to study the parameters of unstable combustion [9]. But this conventional method of studying unstable SHS regimes does not work, of course, if the combustion product is in the liquid phase, as in the case of ZnS-SHS processing. The violent convection that occurs in the nonuniformly cooled high-temperature melt at normal gravity annihilates any footprint of the unstable burning. The fast solidification of the ZnS-SHS melt in microgravity is the only way to have this "footprint" available for study. The variation of the average flame speed with the charge diameter is shown in Fig. 5. In general, a slight increase in the flame speed with charge diameter is observed, indicating the effect of heat losses. The downward-propagating flame in a wide channel (d > 12 ram) usually has a speed lower than that of an upward-propagating flame, and the value of the flame speed in microgravity is close to the speed of a downward propagating flame. However, as a result of the limited data available in microgravity as well as the instability of the flame speed itself, this dependence is not clear.

Synthesis of Unconfined Charges: Since the product of the ZnS combustion synthesis is in the liquid state, in normal gravity it is not possible to carry out experiments with an unconfined charge because of tile gravity-induced flow of the melt. In our microgravity experiments, it was found that complete combustion of an unconfined charge was possible. However, as a result of the poor heat

COMBUSTION SYNTHESIS OF ZnS IN MICROGRAVITY

1655

12 9 n 9

10

Upwardspropagation Downwardspropagation iicrogravity results

9

..................................................

~i

....................

............................................................ ~ ............................ ~ ............................ i.......................

g

t3

it3

E .......................

r

i .................

.......................................................................................................

B r a~

'i ,-~'~" ,i 'i 0

5

Quenching diameter for g = 1

10

15

20

Charge Diameter (mm)

transfer between the melt and the gaseous environment, the ZnS was still in the molten state upon return to normal gravity. G-jitter and the large acceleration associated with the sudden return of normal gravity at the end of the 20-s mierogravity period can introduce significant inertial forces on the melt and cause it to fragment into droplets. Even during the microgravity period, strong Marangoni convection and turbulence effects may also disrupt the melt interface. For the unconfined charge experiments in microgravity, we recovered solid spherical ZnS particles with a wide size spectrum ranging from fractions of a millimeter to 5 mm after the synthesis. If an integral solid sample of ZnS is desired, it is necessary to arrange for rapid quenching within the 20-s parabolic flight microgravity duration. ,

Temperature Profile in the Flame Front: The first objective of the thermocouple measurements was to estimate the thickness and temperature profile of the combustion wave. In order to accomplish this, a fine thermocouple was embedded into the central axis of the charge to measure the temperature profile as the flame propagates past the thermocouple. Three types of unsheathed fine-gauge thermocouples were used: type K (chromel-alumel, maximum temperature 152,3 K), diameter 25/~m; type S (Pt/Pt-10% Rh, maximum temperature 1723 K), diameter 50/lm; and type C (W/5% Re-W/26% Re, maximum temperature 2593 K), diameter 75/tm. The thermocouple was shielded in a two-hole alumina eeramic tube with outside diameter 1.5 mm and a length of 60 mm. The junction and 5 mm of the thermoeouple wire were exposed. The output from the thermoeouple was amplified with a 50 • DC amplifier and recorded by a digital oscilloscope.

25

30

FIG. 5. The variation of the flame speed with the charge diameter.

The theoretically estimated adiabatic flame temperature in the Zn-S combustion front is about 2200 K. Only high-temperature thermocouples based on W-Re alloy can cover this temperature range. At the same time, the thiekness of the combustion front in our ease is very small, so that thermocouples with a very small response time are required to resolve the reaction front. The thermal thickness of the flame is of the order o f L - 2JCsp~V, where 2~, Cs, and p~, are the conductivity, specific heat, and density of the unburned mixture, respectively, and V is the flame speed. Assuming that the thermophysical properties of the mixture are close to those of liquid sulfur, one can estimate for a measured flame speed of 0.5 cm/s that the flame thickness is of the order of 50-i00 /tmJ Correspondingly, the flame passes through the thermocouple in about 10-20 ms. The response time estimated for the 75-urn W-Re thermocouple is about 50 ms. Therefore, the measured value of the flame thickness should be considered as an estimate. The results of the flame structure measurements with the different thermoeouples are given in Fig. 6. To plot this graph, the time from the direct temperature measurements was multiplied by the average flame speed in the channel of that diameter (0.7 era/ s), and the point at which the temperature first begins to rise was set to a value of 0.2 mm on the horizontal axis. It is clear that the measured flame thickness shrinks as the thermocouple size decreases, and therefore, the flame thickness is clearly not more than the lowest measured value of about 300/L m. It was found that the high-temperature Zn-S melt ~There are no data available fbr the sulfur properties at such high temperatures in the literature. We did the estimation by extrapolation of the available data into the hightemperature region.

1656

MATERIALS SYNTHESIS 2500

'

'

'

I

.

.

.

.

.

.

.

Thermocouple Type i

1-.-o I

2000

-s

........ ]

I ........ i ....................... i........................ i ...................

|

.........

'

1500

& E I--

1000

500

~

........J.............................................

,

0

J

,

I

0.2

~

L

~

I

r

,

0.4

,

I

~

,

L

0.6 Distance

l

0.8

(ram)

,

,

t

I

1

,

t

.2

FIG. 6. Temperature profile in the flame front.

is an extremely chemically aggressive environment for all types of thermocouples. It is quite possible, as a result of the chemical reaction with sulfur, that thermocouples usually break at temperatures much lower than the maximum thermoeouple temperature limit (see above). The W-Re thermocouples break at temperatures close to 2000 K. Even the Pt-Rh thermocouple, which is usually resistant to the oxidizing environment, breaks at a lower temperature than its theoretical temperature limit.

Crystal Structure and Composition of Synthesized Samples: Thin polished sections (-50-pro thick) of samples synthesized in the quartz ampoules (i.d. 20 mm) were prepared in order to determine their crystal structure. The sections were studied with an optical microscope under polarized light. The typical structure of the cylindrical sample consists of three zones along the sample's radius, as shown in Fig. 7. The first zone, on the outer boundary of the sample, has a thickness of about 1-2 mm. This zone consists of small ( - 5 0 100/tm) crystals of approximately rectangular shape. The borders of the crystals are very close to each other. The bulk of the sample consists of a region containing very long (5-7 ram) radially oriented columnar crystals. The thickness of the crystals increases and reaches about 0.1-0.2 mm near the center of the sample. Columnar crystals are separated by "dark spaces." (the presence of the dark spaces under polarized light usually means the existence of the amorphous phase). As the thickness of the columnar crystals increases, the dark spaces between crystals also become wider. The central third zone has a diameter on the order of 5-7 mm and consists of equiaxial crystals. The size of the crystals is about 0.5-1 ram, and they are separated by wide dark zones.

FIG. 7. The typical structure of the synthesized sample: (1) rectangular crystal zone; (2) columnar crystal zone; (3) equiaxial crystal zone. The samples that were synthesized in microgravity also exhibit the tree zones observed in ground-based samples, although there are a number of differences present. For example, for the microgravity samples, the outer zone 1 is about 30-50% wider and consists of rectangular crystals almost twice as large as the ground-based samples. It is also interesting to note

COMBUSTION SYNTHESIS OF ZnS IN MICROGRAVITY that the samples synthesized in the flight have a shorter columnar crystal zone. Because the complete solidification time of the cylindrical samples in a quartz ampoule is longer than the microgravity duration, only the outer part of the sample can solidify in mierogravity. Therefore, the larger rectangular crystals obtained during microgravity flights may be due to the more uniform cooling of the liquid in the absence of convection. However, the effect of the shrinking of the zone of columnar crystals is probably the consequence of the intense mixing of the melt by the violent convection that starts in the nonuniformly cooled liquid during the "pull down" 2-g acceleration. The x-ray diffraction analysis of the synthesized samples was provided with a Juniaer-H~igg focusing camera and CuKa radiation. A small amount of MgA1204 (spinel) was added as an internal standard (a = 8.0833 A at room temperature). In contrast to the results published in Refs. 8 and 10, where it was found that SHS-synthesized ZnS samples have both wurtzite and sphalerite crystal modification, only wurtzite-type crystals were found in our samples. The lattice's parameters for the outer part of the samples (a = 3.8214 A, c = 6.2657 A) that were synthesized and presumably solidified in a mierogravity environment are closer to the ideal wurzite structure than, for example, the lattiee's parameters of standard ZnS industrial powder (Canada Colors and Chemicals Ltd.). Conclusions The results reported in this paper are based on ground-based experiments and one KC-135, NASA microgravity parabola flight campaign and have successfully demonstrated the possibility of SHS-ZnS processing in microgravity. The highlights of the important results are as follows. 1. The novel technique developed for preparing the precursor mixture of Zn-S by mixing the zinc powder in the molten sulfur allowed an increase in the quality of the product. The density of synthesized samples reaches a value of 3.92 g/era 3 (98.5% of the theoretical density), and the completeness of the chemical reaction was 99%. 2. Average flame speeds are typically of the order of 7 mm/s, and slightly lower values ( - 4 mm/s) are observed near the quenching limit. The flame speed is not stable along the samples with diameters more than 12 mm. 3. The quenching diameter is found to be of the order of 5 mm in ground-based experiments (and in microgravity less than 4 mm), permitting quite small samples to be studied.

1657

4. Thermocouple measurements of the temperature profile in the flame front indicate that the thermal thickness of the flame is less than 0.3 mm. 5. Containerless processing of ZnS is achieved in microgravity, although the abrupt return to gravity and G-jitter cause the melt to break up into a spectrnm of drop sizes prior to solidification. 6. X-ray diffraction data show a wurtzite (hexagonal) structure both in ground-based and flight-synthesized samples. The lattiee's parameters are the most similar to the ideal ZnS wurtzite structure in the outer part of the sample synthesized in microgravity. Acknowledgments This work is supported under CSA Contract No. 9F0073-6019/01SW, and the authors are grateful for the assistance of Dr. O. Peraldi in designing the microgravityhardware.

REFERENCES 1. Shteinberg, A. S., Sherbakov, V. A., Marttynov, V. V., Muldaoyan, M. Z., and Merzhanov, A. G. Proceedings of AIAA/1KI Microgravity Science Symposium, AIAA, Washington, DC, 1991, pp. 268-272. 2. Odawara, O., Mori, K., Tanji, A., and Yoda, S., J. Mater. Synth. Process. 1:203-207 (1993). 3. Moore, J. J., 31st Aerospace Meeting and Exhibit, AIAA Paper 93-0830, January 1993. 4. Kane, P., and Larrabee, G., Characterization of Semiconductor Materials, McGraw-Hill Book Co., New York, 1970. 5. Shionoya, S., Kobayashi,Y., and Koda, T.,J. Phys. Soc. Jpn. 20:385-389 (1965). 6. Abricosov, N. Kh., Banldna, V. F., and Poteskaya, L. V., Semiconducting II-IV, IV-VI and V-VI Compounds, Plenum Press, New York, 1969, pp. 1~10. 7. Indradev, L. J., Van Ruyven, B., and Williams, F., Proceedings" of the International Conference on Luminescence, Akademiai Kiado, Budapest, 1968, pp. 10441050. 8. Kozitskiy,S. V., Pisarskiy,V. P., and Polischuk, D. D., Izv. Acad. Nauk SSSR 10:2472-2475 (1990) (in Russian). 9. Merzhanov, A. G., in Combustion and Plasma Synthesis of High-Temperature Materials (L. A. Munir and J. B. Holt, Eds.), VCH, New York, 1990, pp. 1-53. 10. Molodetskaya I. E., and Pisarskiy, V. P., Combust. Explos. Shock Waves 28:385-389 (1992). 11. Rubensteine, M.,J. Cryst. Growth 41:311-316 (1977). 12. Margolis, S., and Matkowski,B. J., in Combustion and Plasma Synthesis of High-Temperature Materials (L. A. Munir and J. B. Holt, Eds.), VCH, New York, 1990, pp. 73-83.