Synthesis of SiC powder by RF plasma technique

This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright

Author's personal copy Powder Technology 214 (2011) 300–305

Contents lists available at SciVerse ScienceDirect

Powder Technology journal homepage: www.elsevier.com/locate/powtec

Synthesis of SiC powder by RF plasma technique Z. Károly a,⁎, I. Mohai a, Sz. Klébert a, A. Keszler a, I.E. Sajó b, J. Szépvölgyi a, c a b c

Institute of Materials and Environmental Chemistry, Chemical Research Center, HAS, 59-67 Pusztaszeri út, Budapest, 1025, Hungary Chemical Research Center, Hungarian Academy of Sciences, 59-67 Pusztaszeri út, Budapest, 1025, Hungary Research Institute of Chemical and Process Engineering, University of Pannonia, 10 Egyetem u. Veszprém, 8200, Hungary

a r t i c l e

i n f o

Article history: Received 10 May 2011 Received in revised form 21 July 2011 Accepted 27 August 2011 Available online 2 September 2011 Keywords: Powders–gas phase reaction Carbon SiC Synthesis

a b s t r a c t Continuous synthesis of SiC nanoparticles by RF thermal plasma method has been studied. Precursor mixtures comprised commercial silica powder and various types of carbon source including graphite, char, carbon black as well as the carbonaceous residue of tire pyrolyses. The obtained SiC consisted of nanosized particles that were crystallized mainly in β phase with traces of α. The conversion rate of the silica precursor to SiC varied between 60% and 73% depending on the type of carbonaceous material and on the carbon excess. The main obstacle to achieve higher conversion is the rapid cooling of reactive species that can also be attributed to formation of nanosized particles. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Silicon carbide (SiC) is the most widely used non-oxide ceramics for many industrial applications because of its attractive mechanical and thermal properties such as high mechanical strength and hardness, high thermal conductivity, excellent corrosion and thermal shock resistance, semiconductivity, etc. [1–6]. SiC finds application as free-standing parts, as thin layer depositions as well as in composites as filler material. Since nanoparticulate materials exhibit unique properties and nanocomposites are reported to have superior mechanical properties too over their conventional counterparts [7–11], over the last decade attempts have been directed to the synthesis of nanosized SiC powders. The main method of SiC production is the Acheson process, which is the carbothermal reduction of SiO2 by coke at 2200–2500 °C [12]. Due to the high reaction temperatures and long reaction time of the process, the product has large grain size and invariably contaminated with more or less oxygen. For the synthesis of nanosized SiC powder numerous synthesis routes have been developed, although most of them remained only in laboratory scale. Among these processes the most studied include mechanical milling [13,14], rapid carbothermal synthesis [15–17], SHS processes [18], microwave synthesis [19], polymer pyrolisis [20], sol–gel processes [21], CVD [22] and laser synthesis [23]. All the mentioned processes have their merits and limitations over the others such as the cheaper precursors they use or the lower reaction temperature, the higher purity of the obtained product, and so on. It's beyond the scope of this article to make detailed

⁎ Corresponding author. E-mail address: [email protected] (Z. Károly). 0032-5910/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.powtec.2011.08.027

comparison of these processes and judge them in terms of feasibility. Instead, in the present article the authors report on a plasmathermal process also for the synthesis of nanometer-sized SiC powder which combine the benefits of low cost precursors with continuous processing. Thermal plasmas have been mainly utilized in such processes where the extremely high temperature that can reach several thousand degrees provides advantage to establish a more economical processing route [24–29]. In addition, it makes possible a continuous process. The main question is whether or not the residence times of solid precursors in the plasma were enough for an in-flight reaction. In respect of residence time RF plasma has an advantage over arc plasma due to the more extended plasma flames that result in longer mean residence time of reactive species in the hot plasma region [30]. In the present paper attempt has been made to synthesize nanosized SiC powder from low cost raw materials using RF thermal plasma. We investigated the feasibility of an in-flight reaction when reactants are fed in solid form and the effect of the carbon excess on the SiC yield. Various types of carbon sources were compared in terms of efficiency of the reduction and subsequent carburization. Beyond economic point of view, environmental aspects were also considered when carbon powder that remained from the pyrolytic destruction of waste car tires was also involved in the tests. 2. Experimental Starting powders were of commercial fine silica (C600, Sifraco) and different carbon powders as well as char residue from tire pyrolysis. Powder mixtures were prepared and processed by milling and thermal plasma processing. Not only the type and source of carbon powders

Author's personal copy Z. Károly et al. / Powder Technology 214 (2011) 300–305

were varied in the particular tests but the carbon to silica ratio as well. The characteristics of the used carbon powders and the performed tests are listed in Table 1. The main object of milling was not attrition but much rather thorough homogenization to achieve good contact between the reactant particles. For this reason milling was carried out in a ball mill for 4 h in ethanol suspension. After milling the ethanol was evaporated then milled further for another 30 min to disintegrate the agglomerated particles. Thermal plasma processing of the powder mixtures was carried out in an RF system. The experimental setup consists of an induction plasma torch (Tekna Plasma systems Inc, PL-35) mounted on the top of a reaction chamber connected to a filter system and a vacuum pump as well as a powder feeding unit. The five turn plasma torch is connected to a radio frequency (RF) generator (Lepel, USA) delivering a maximum power of 30 kW at 3–5 MHz. Detailed description of the system and schematic presentation is given elsewhere [31]. Plasma (central) gas was a mixture of Ar (slpm) and He (slpm) and sheath gas was also Ar (slpm) mixed with He (slpm). The role of He was to improve heat conductivity and the enthalpy of the plasma. Mixed powders were fed with a carrier gas He (slpm) and injected into the Ar–He plasma via a water cooled stainless steel probe. The probe was positioned axially into the torch down to the middle length of the coil. The plate power was tuned to 25 kW and the precursor feed rate was set to 50 g·h −1 in each experiment. The products obtained in the form of loosely agglomerated small particles contained carbon, silica and silicon as impurities. The excess carbon was burnt off by heating the powder at 610 °C for 1 h. Silica was removed by treating the product with HF. The obtained powders were characterized with respect to phase composition, elemental analysis and morphology. X-ray diffraction (XRD) measurements were performed on a Siemens D5000 diffractometer using CuKα1 radiation. Diffractograms were obtained in a step scanning mode 20° b 2θ b 80°. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) investigations were performed on a Hitachi HR-SEM S4800 and a Philips CM-20 (200 kV) microscope, respectively. Plasma synthesized powders were characterized by 7 point BET analysis by an Autosorb 1C apparatus. The mean particle size was calculated from the specific surface area according to the following equation: d¼

6 ρ·SSA

where ρ is the powder effective density calculated from chemical analysis results.

resolved parameters need special instrumentation and they are very time consuming even for one set of plasma parameters. Short residence time of the reagents in the hot regions might not allow achieving the chemical composition corresponding to thermal equilibrium conditions. Yet, thermodynamic calculations based on the minimization of the system total free energy can provide suitable data for potential products and their concentrations. In a plasma reactor complete evaporation and atomization or even ionization of the starting materials is supposed to take place. The most important deviation of a thermal plasma process from the equilibrium reaction is that during cooling of the reagents the first solid phase that may form according to the Gibbs phase diagram will be the dominant product. In plasma conditions the first condensed phase cannot react further due to the high cooling rates. Hence, thermodynamically unstable phases can be frozen even at lower temperatures. Thermodynamic and kinetic aspects of carbothermic reduction of SiO2 were extensively analyzed in several papers [28,34]. It has been generally accepted that the strongly endothermic carbonization process in equilibrium conditions takes place as a result of two subsequent reaction steps in accordance to: SiO2 þ C→SiOðgÞ þ CO SiOðgÞ þ C→SiC þ CO: It can be assumed, however, that formation of SiC in RF thermal plasma takes place in a different way. In RF plasma the injected particles below a certain particle size can evaporate and the subsequent reactions occur in gas or vapor phase. In a higher distance from the plasma core the reaction products will condense from the gaseous phase. To estimate the reaction products in the SiO2 + C reaction in plasma conditions thermodynamic computations were performed by code FACTSAGE, which determines the thermodynamically most stable phases and their relative amounts at a given temperature and pressure by Gibbs energy minimization. In the calculations silica to carbon ratios were varied in the range of 1:2–1:6. Considering the complex reaction equation, the above settings represent a carbon deficient, a stochiometric, and a carbon excess case, as well. Fig.1. shows the equilibrium phases against temperature for the case of 1:2 SiO2 to C ratio, which corresponds to a carbon deficient state. The major reactions that occur above 3000 °C are as follows: SiO2 →SiOðgÞ þ 1=2 O2

Table 1 Major characteristics of the used carbon powders and the examined tests types. Test

Type of used carbon

C to SiO2 molar ratio

Size of the C-powder, μm

SSA of C powder m2·g−1

PS1 PS2 PS3 PCS PSC PCh PGr

Pyrolytic soot Pyrolytic soot Pyrolytic soot Columbian soot Cancarb soot Char Graphite

2:1 5:1 9:1 6:1 6:1 6:1 6:1

0.32 0.32 0.32 0.23 0.45 10–20 7–10

350 350 350 130 10 61 13

ð1Þ

SiOðgÞ þ C→SiðgÞ þ CO

ð2Þ

CðsÞ→CðgÞ:

ð3Þ

3. Thermodynamic calculations Estimation of the possible products of a reaction under thermal plasma conditions is difficult because of the different thermal history of the particles in the plasma. Data for temperature distribution and reagent trajectories (particles, molecules, atoms, radicals, etc.) in the reactor are generally based on mathematical simulations using more or less simplifications [32,33]. This is because measurements of these spatially

301

In this temperature range major part of the SiO(g) is reduced, thus only Si and CO are present considerably, while SiO and C(g) only in minor amount. With decreasing temperature, under 4000 °C Si(g) starts at first slightly then under 3500 °C sharply decreasing with simultaneous increase in the amount of SiO due to its reoxidation by CO. Between 2700 and 4000 °C Si2C and SiC2 phases appear too as intermediate phases. Although, the formation of these phases could be described by several reaction routes, regarding the prevailing phases at the given temperatures, the relevant reaction routes can be restricted to the following ones: 2SiðgÞ þ C→Si2 CðsÞ; 3SiðgÞ þ CO→Si2 C þ SiOðgÞ SiðgÞ þ 2CðgÞ→SiC2 ðsÞ; 3SiðgÞ þ 2CO→SiC2 ðsÞ þ 2SiOðgÞ: The above reactions and thermodynamic calculations, however, do not consider kinetic aspects i.e. how realistic a three body collision

Author's personal copy 302

Z. Károly et al. / Powder Technology 214 (2011) 300–305

Fig. 1. Chemical equilibrium diagram for the SiO2–C system assuming a 2:1 C to SiO2 molar ratio.

involved reaction to occur. For this reason it is not necessary to take into consideration these routes in a given case. Below 2700 °C silicon appears as liquid phase. Once a vapor phase is condensed within plasma conditions it has little chance to be involved in additional reactions on account of rapid cooling. Formation of SiC(s) becomes also possible under 2700 °C according to the equation as 2Si(g,l) +CO →SiC (s)+ SiO(g). Its amount gradually increases along with that of SiO(g) with decreasing temperature to finally reach a local maximum. The as-formed phases then can be theoretically oxidized by CO under 1800 °C. Oxidation of SiO is, however, preferred over SiC in terms of thermodynamics and starts at a bit higher temperature. A gas phase reaction has also got a greater probability to take place, while the condensed SiC remains unreacted. The equilibrium phases in Fig. 2. represent the case of stochiometric carbon to silica precursor ratio. The situation over 4000 °C is also determined by the reactions (1–3) described before. Due to the higher carbon input, the amount of C(g) is well increased as compared to the former case, while SiO(g) almost disappeared. Si(g) remains the dominant phase down to 3000 °C, although, its amount gradually decreases along with that of C(g) with decreasing temperature. Simultaneously the amount of SiC2(s) increases instead of SiO(g). Si2C also appears under 4000 °C as an intermediate phase but in a much less extent. Since all the reactions for the formation of SiC2(s) involves the collision of three bodies or so and thereby are quite unrealistic in terms of kinetics, it is assumed that it does not occur in fact during the extremely short residence time in the plasma flame. Therefore we suggest that neither the amount of Si(g) nor that of C(g) decreases in this region in actual conditions. Hence the governing reaction of SiC formation under 2700 °C can be described as Si(g) + C(s,g) → SiC(s), with minor participation of C(s). Obviously there are a number of other reaction routes listed below that result in SiC but they probably can be ruled out due to kinetic reasons. SiC2→SiC þ C; Si2C→SiC þ Si; Si2C þ SiC2→3SiC; Si2C þ C→2SiC; SiC2 þ Si→2SiC The amount of formed SiO(g) is negligible basically because of the excess C(s,g), while the final amount of SiC(s) is doubled as compared to the former case.

Fig. 2. Chemical equilibrium diagram for the SiO2–C system assuming a 3:1 C to SiO2 molar ratio.

Further increasing the initial carbon content in the precursor mixture the prevailing phases are not changed only their relative amounts at any temperature. Even C2(g) and C3(g) appear in considerable amounts above 4000 °C. For this reason the formation of SiC2(s), which is the dominant phase at lower temperatures, can be described according to the reaction: Si(g) + C2(g) → SiC2(s). As a result, the reaction of SiC2(s) + Si(g) → 2SiC(s) under 2700 °C must play also a greater role. In effect the subsequent reactions end up with formation of SiC with negligible amount of SiO(g). Thermodynamic computations certainly provide information about equilibrium states but it cannot take into consideration kinetic limitations as it was referred in above paragraphs. 4. Results and discussion According to SEM images (Fig. 3) the plasma synthesized pure SiC powders were found in the form of loosely agglomerated particles having a size around 100 nm in all tests. To have a more precise assessment for the mean size it was calculated from the value of the surface area assuming spherical shape. This way the mean particles size varied between 20 and 30 nm in the particular runs irrespective of the type of the used carbon precursor. In higher magnification TEM examinations revealed (Fig.4) that the particular particles are not spherical in fact and their size varies between 20 and 120 nm. The crystalline phases of the reaction products obtained in particular tests can be identified on the XRD patterns (Fig.5). The presence of SiC is indicated by intensive peaks in all the performed tests except PS1. Beyond SiC, carbon, metallic silicon as well as a considerable amount of amorphous phase presumably amorphous silica and carbon were also detected. Comparing the XRD peaks of tests with different carbon excess (Tests PS1–PS3) one can discover significant dissimilarities. Although SiC was detected invariably in all tests, its quantity seems by far the lowest in Test PS1. On the other hand, at the same tests the highest amount of Si was obtained as compared to any other tests. Still it does not contradict to thermodynamic calculations. In contrast, it is only the result of rapid cooling conditions and low residence times (ca. 1–20 ms) of the species in the plasma flame. Thus, the first phase to be condensated was hindered from additional reactions. In addition, in Test PS1 the largest amount of product powder, as it is indicated by an increased background on the XRD plot, was composed of some amorphous phase most probably SiO2. The amount of silica was by far the highest compared with the other tests. This silica,

Author's personal copy Z. Károly et al. / Powder Technology 214 (2011) 300–305

303

Fig. 3. SEM image of pure SiC powder.

however, cannot be considered as unprocessed part of the precursor, as the precursor silica was crystalline. In contrast, it probably formed by rapid oxidation of SiO at lower temperatures. On the effect of a slight increase in the C to SiO2 ratio a significant decrease in the amount of glassy phase resulted with a simultaneous increase in the amount of SiC. Further increasing the carbon excess, the amorphous phase seemingly increased again, probably due to the higher content of the unreacted part of the amorphous carbon soot precursor. A less intensive peak at 2θ = 47° refers to the smaller amount of metallic zinc that remained in the pyrolytic soot as impurity. The zinc was condensed also to the reactor wall along with SiC. It, however, could be easily removed in the subsequent leaching procedure. Using different types of carbon powders the most significant change is the less intensive peaks of metallic Si. In test PGr it could not be detected at all that suggests a better conversion. Neither any amorphous phase was present. The most plausible explanation for this is that in this test the precursor carbon was graphite and the unreacted part of the fed carbon preserved its crystalline form. The absence of the diffraction peaks of zinc is not surprising as other carbon precursors were not contaminated with zinc at all. The obtained SiC was crystallized both in α and β phases as well. The latter was represented by a 20–40% higher amount irrespective of the type of the used carbon powder. Other production routes generally results in only one crystalline phase. Silicon carbide produced by the Acheson process is crystallized in the α phase [35], while gas phase reaction usually results in β phase [36–38] that is thermodynamically stable at lower temperatures (Tb 1500 °C). Yet, it is not entirely surprising that the RF plasma process yielded mixed phases in the product. It was observed earlier that the processed material composed of even several crystalline phases of the same

Fig. 4. TEM image of pure SiC powder.

Fig. 5. XRD pattern of selected tests.

compound [39]. The reason of the presence of both phases in this case could be the fact that the precipitated particles had different trajectories and thus were subjected to quite different heat effects (thermal history). This is confirmed by the fact that quartz phase was also found in the product in smaller or higher quantity but in all cases less than 5%. This phase represents the smaller part of the silica in the feedstock material, which passed through the plasma flame without any chemical or physical transformation. As the knowledge of the conversion rate of the process is of paramount importance, the amount of the different phases had to be determined. Considering that the obtained phases were in part amorphous the quantitative XRD phase analysis was not reliable. Instead, we used carbon and oxygen analysis as well as ignition tests. Unreacted amount of carbon was determined by heating the product powders above 600 °C in air, since at this temperature oxidation of SiC is negligible [34]. As a result the powder mixture composed of SiO2 and SiC only. Oxygen and carbon analyses were also performed both on the reactor products and the powders after heating. SiC was finally extracted from the heated powders by leaching the oxides with HF acid solution. After the elemental analysis composition of phases can be calculated. Composition of the powders of particular runs, the LOI (loss of ignition) value, C and SiO2 contents are listed in Table 2. According to LOI values the C content (2 wt.%) was utilized in the highest proportion in the test with the lowest carbon content in the starting mixture (PS1). The high carbon consumption, however, did not accompany with high conversion rate. On the contrary, the conversion was the lowest of all tests. This fact, however, is in accordance with thermodynamics, because the main cause of the low conversion is that during cooling the SiC and SiO formed nearly in the same amount. Being an unstable compound the SiO was subsequently oxidized to SiO2. With increased carbon ratio in the precursor mixture the amount of the unreacted carbon also increased. At the same time it came with significant improvement in the conversion. The obtained value of 73%, however, seems to reach its maximum, since further C excess did not lead to additional improvement. This fact cannot be reasoned by thermodynamics considering that the amount of formed SiO below 1800 °C is negligible as compared to SiC. Kinetics has a lot more to do with experienced lower conversion rates. Presence of metallic silicon, the amount of which was around 7 wt.% in the processed powder, suggests that rapid cooling conditions prevented its transformation into carbide in further reactions. Another loss can be attributed to incomplete evaporation of silica precursor, which passed through the

Author's personal copy 304

Z. Károly et al. / Powder Technology 214 (2011) 300–305

Acknowledgments

Table 2 Main characteristics and results of the performed tests. Test

Type of the carbon source

C to SiO2 molar ratio

LOI%

SiO2%

PS1 PS2 PS3 PCS PSC PCh PGr

Pyrolytic soot Pyrolytic soot Pyrolytic soot Columbian soot Cancarb soot Char Graphite

2:1 5:1 9:1 6:1 6:1 6:1 6:1

2.0 24.0 36.2 43.1 33.5 20.6 24.3

11.7 18.0 22.5 21.2 19.1 12.9

Conversion rate, %

O in SiC%

0.22 0.73 0.65 0.60 0.66 0.69 0.71

18 3.2 4.3 4.9 4.2 3.1 4.2

plasma without being reacted. These findings drive us to decide that additional tests using different commercial carbon sources (Tests of PSC, PCS, PCh, PGr) were performed using a double carbon excess. Comparing the conversion rates of the particular tests with respect to type of used carbon the maximum conversion of 73% was obtained at test PS2. In other tests, however, the conversion rate was only slightly lower and varied between 60% and 71%. It suggests that the initial size and specific surface area have negligible effect on the conversion. This also confirms the hypothesis that the reaction takes place in the gas phase after vaporization of the injected starting materials. It is worth noting that lower conversion rates are accompanied to higher LOI values. It follows that in tests having lower conversion rates the reduction by carbon was not complete. This was probably due to the insufficient evaporation of the injected carbon. For this reason there is no sense to increase the carbon excess over a certain value in the hope of better conversion without increasing the plasma power to provide enough energy for vaporization of all the fed carbon powder. These results show that in addition to the carbon excess in the precursor powders, the type of the carbon source has also some influence on the conversion rate.

5. Conclusions Experimental investigation has been performed for continuous synthesis of SiC powder in RF plasma reactor. Mixture of fine particulates of commercial silica and different types of carbon powder were used as precursor. The particular runs differed in type of the carbonaceous material and the silica to carbon ratio in the precursor. The studied carbon sources included soots, char coal, graphite powder and soot from waste tire pyrolysis. In spite of the extremely short residence time of the reactants in the high temperature plasma flame the precursors were reacted to eventually form fine SiC particles. The SiC synthesis could only take place by evaporation of the fed material in subsequent sequence of reactions. Considering the different carbon sources the conversion rate of silica to carbide varied from 60% to 73%. Surprisingly, the highest value was achieved using the pyrolytic soot. Significant difference in conversion ratio occurred, however, varying the carbon excess. A slight carbon excess is mandatory to approach or reach the maximum value but further increasing the carbon excess does not accompany any higher conversion rate. The reason of the moderated conversion is related in part to low residence times of the flying particles through the plasma, which prevent the evaporation of all the reactant species and in part the rapid cooling conditions characteristic to plasma flame that could hinder the completion of carburization. Even if higher power may help to offset the limitations of low residence time, rapid cooling is more difficult to avoid. On the other hand rapid cooling is essential to achieve homogenous nucleation that ultimately results in nanoparticles. The final product of SiC composed of mainly β crystallites with minor amounts of α phase. The mean particle size was around 20–30 nm. The obtained results demonstrate that RF plasmathermal method is a useful tool to produce nanosized SiC particulates by a continuous method using low cost precursor materials.

The research leading to these results has received funding from the European Union's Seventh Framework Programme [FP7/2007–2013] under grant agreement No. 226549. References [1] A. Okada, Automotive and industrial applications of structural ceramics in Japan, Journal of the European Ceramic Society 28 (5) (2008) 1097–1104. [2] G. Gabriel, I. Erill, J. Caro, R. Gomez, D. Riera, R. Villa, et al., Manufacturing and full characterization of silicon carbide-based multi-sensor micro-probes for biomedical applications, Microelectronics Journal 38 (3) (2007) 406–415. [3] R.J. Ciora, B. Fayyaz, P.K.T. Liu, V. Suwanmethanond, R. Mallada, M. Sahimi, et al., Preparation and reactive applications of nanoporous silicon carbide membranes, Chemical Engineering Science 59 (22–23) (2004) 4957–4965. [4] S. Dimitrijev, Silicon carbide as a material for mainstream electronics, Microelectronic Engineering 83 (1) (2006) 123–125. [5] C. Brylinski, Silicon carbide for microwave power applications, Diamond and Related Materials 6 (10) (1997) 1405–1413. [6] B. Elyassi, M. Sahimi, T.T. Tsotsis, Silicon carbide membranes for gas separation applications, Journal of Membrane Science 288 (1–2) (2007) 290–297. [7] H. Awaji, S.M. Choi, E. Yagi, Mechanisms of toughening and strengthening in ceramic-based nanocomposites, Mechanics of Materials 34 (7) (2002) 411–422. [8] B. Derby, Ceramic nanocomposites: mechanical properties, Current Opinion in Solid State and Materials Science 3 (5) (1998) 490–495. [9] P. Descamps, D. O'Sullivan, M. Poorteman, J.C. Descamps, A. Leriche, F. Cambier, Creep behaviour of Al2O3SiC nanocomposites, Journal of the European Ceramic Society 19 (13–14) (1999) 2475–2485. [10] H.R. Rezaie, W.M. Rainforth, W.E. Lee, Fabrication and mechanical properties of SiC platelet reinforced mullite matrix composites, Journal of the European Ceramic Society 19 (9) (1999) 1777–1787. [11] G.J. Zhang, Z.Y. Deng, N. Kondo, J.F. Yang, T. Ohji, Reactive hot pressing of ZrB2–SiC composites, Journal of the American Ceramic Society 83 (9) (2000) 2330–2332. [12] Carbide, Nitride and Boride Materials Synthesis and Processing, in: A.W. Weimer (Ed.), Chapman and Hall, New York, 1997. [13] D. Chaira, B.K. Mishra, S. Sangal, Synthesis and characterization of silicon carbide by reaction milling in a dual-drive planetary mill, Materials Science and Engineering A, Structural Materials Properties Microstructure Processing 460 (2007) 111–120. [14] R.M. Ren, Z.G. Yang, L.L. Shaw, Synthesis of nanostructured silicon carbide through an integrated mechanical and thermal activation process, Journal of the American Ceramic Society 85 (4) (2002) 819–827. [15] C. Real, D. Alcala, J.M. Criado, Synthesis of silicon carbide whiskers from carbothermal reduction of silica gel by means of the constant rate thermal analysis (CRTA) method, Solid State Ionics 95 (1–2) (1997) 29–32. [16] S. Pan, J. Zhang, Y. Yang, G. Song, Effect of process parameters on the production of nanocrystalline silicon carbide from water glass, Ceramics International 34 (2) (2008) 391–395. [17] F.J. Narciso-Romero, F. Rodriguez-Reinoso, M.A. Diez, Influence of the carbon material on the synthesis of silicon carbide, Carbon 37 (11) (1999) 1771–1778. [18] Y. Yang, Z.M. Lin, J.T. Li, Synthesis of SiC by silicon and carbon combustion in air, Journal of the European Ceramic Society 29 (1) (2009) 175–180. [19] T. Ebadzadeh, E. Marzban-Rad, Microwave hybrid synthesis of silicon carbide nanopowders, Materials Characterization 60 (1) (2009) 69–72. [20] M. Narisawa, M. Shimoda, K. Okamura, M. Sugimoto, T. Seguchi, Reaction-mechanism of the pyrolisis of polycarbosilane and polycarbosilazane as ceramic precursors, Bulletin of the Chemical Society of Japan 68 (4) (1995) 1098–1104. [21] L. Cerovic, S.K. Milonjic, S.P. Zec, A comparison of sol–gel derived silicon-carbide powders from saccharose and activated carbon, Ceramics International 21 (4) (1995) 271–276. [22] L.Z. Cao, H. Jiang, H. Song, Z.M. Li, G.Q. Miao, Thermal CVD synthesis and photoluminescence of SiC/SiO2 core–shell structure nanoparticles, Journal of Alloys and Compounds 489 (2) (2010) 562–565. [23] P. Schaaf, M. Kahle, E. Carpene, Reactive laser synthesis of carbides and nitrides, Applied Surface Science 247 (1–4) (2005) 607–615. [24] L.R. Tong, R.G. Reddy, Thermal plasma synthesis of SiC nano-powders/nano-fibers, Materials Research Bulletin 41 (12) (2006) 2303–2310. [25] P. Fauchais, A. Vardelle, A. Denoirjean, Reactive thermal plasmas: ultrafine particle synthesis and coating deposition, Surface and Coating Technology 97 (1–3) (1997) 66–78. [26] N. Rao, B. Micheel, D. Hansen, C. Fandrey, M. Bench, S. Girshick, et al., Synthesis of nanophase silicon, carbon, and silicon-carbide powders using a plasma expansion process, Journal of Materials Research 10 (8) (1995) 2073–2084. [27] S.M. Oh, M. Cappelli, D.W. Park, Preparation of nano-sized silicon carbide powder using thermal plasma, Korean Journal of Chemical Engineering 19 (5) (2002) 903–907. [28] P. Kong, T.T. Huang, E. Pfender, Synthesis of ultrafine silicon-carbide powders in thermal arc plasmas, IEEE Transactions on Plasma Science 14 (4) (1986) 357–369. [29] Y. Leconte, M. Leparoux, X. Portier, N. Herlin-Boime, Controlled synthesis of beta-SiC nanopowders with variable stoichiometry using inductively coupled plasma, Plasma Chemistry and Plasma Processing 28 (2) (2008) 233–248. [30] B. Todorovic-Markovic, Z. Markovic, I. Mohai, Z. Karoly, L. Gal, K. Foglein, P.T. Szabo, J. Szepvolgyi, Efficient synthesis of fullerenes in RF-thermal plasma reactor, Chemical Physics Letters 378 (3–4) (2003) 434–439.

Author's personal copy Z. Károly et al. / Powder Technology 214 (2011) 300–305 [31] I. Mohai, J. Szepvolgyi, Treatment of particulate metallurgical wastes in thermal plasmas, Chemical Engineering and Processing 44 (2) (2005) 225–229. [32] D. Bernardi, V. Colomboa, E. Ghedini, A. Mentrelli, Three-dimensional effects in the modelling of ICPTs, European Physical Journal D: Atomic, Molecular, Optical and Plasma Physics 25 (2003) 271–277. [33] D. Bernardi, V. Colomboa, E. Ghedini, A. Mentrelli, Comparison of different techniques for the FLUENTc_-based treatment of the electromagnetic field in inductively coupled plasma torches, European Physical Journal D: Atomic, Molecular, Optical and Plasma Physics 27 (2003) 55–72. [34] L. Stachowicz, S.K. Singh, E. Pfender, S.L. Girshick, Synthesis of ultrafine SiC from rice hulls (husks) — a plasma process, Plasma Chemistry and Plasma Processing 13 (3) (1993) 447–461. [35] W.E. Lee, WMR, Ceramic Microstructures: Property Control by Processing, Chapman and Hall, London, 1994.

305

[36] W.Z. Zhu, M. Yan, Synthesis of nanometer-sized SiC powders through chemical vapour deposition — I. Effect of reaction temperature, Materials Chemistry and Physics 53 (3) (1998) 262–266. [37] W.Z. Zhu, M. Yan, Synthesis of nanometer-sized SiC powders through chemical vapour deposition II. Effect of C2H4/SiH4 mole ratio, Materials Chemistry and Physics 55 (1) (1998) 68–73. [38] W.Z. Zhu, M. Yan, Synthesis of nanometer-sized SiC powders through chemical vapour deposition III. Effect of overall cell pressure, Materials Chemistry and Physics 55 (1) (1998) 74–78. [39] P.K. Mishra, B.B. Nayak, B. Mohanty, Synthesis of silicon-carbide from rice husk in a packed-bed arc reactor, Journal of the American Ceramic Society 78 (9) (1995) 2381–2384.