Fabrication of ZrC particles and its formation mechanism by self-propagating high-temperature synthesis from Fe–Zr–C elemental powders
M.X. Zhang,Q.D. Hu,B. Huang,J.G. Li
DOI: https://doi.org/10.1016/j.jallcom.2011.05.078
IF: 6.2
2011-01-01
Journal of Alloys and Compounds
Abstract:ZrC particles were prepared via self-propagating high-temperature synthesis (SHS) reaction from 0 to 30 wt.% Fe–Zr–C elemental powder mixtures. The ZrC particles of the SHS products greatly decreased from about 10 μm with an irregular shape in free Fe addition to nano-meter order with a nearly spherical shape in 30 wt.% Fe addition. The reaction mechanism of ZrC during the SHS processing was explored through the microstructural observation and phase constituents analysis on the combustion-wave quenched sample in combination with differential thermal analysis (DTA). For the low Fe contents, the solid-state reaction between C and Zr powders should be responsible for the formation of ZrC. While for the high Fe contents, the formation mechanism of ZrC could be ascribed to the dissolution of C into the Fe–Zr melt and the precipitation of ZrC. The addition of Fe to Zr–C reactants not only inhibits the growth of ZrC particles but also promotes the occurrence of ZrC-forming reaction. Keywords SHS ZrC Reaction mechanism Fe 1 Introduction Zirconium carbide, ZrC has attracted more and more attention due to its outstanding physical and chemical properties, such as high hardness and melting point, good electrical conductivity, high resistance to corrosion and wear, as well as excellent chemical stability [1–3] . In addition, ZrC particle is a good obstacle for dislocation movement and can increase recrystallization temperature and hinder grain growth, which lead to strength increment at elevated temperature [4] . As one of the potential and promising materials, ZrC has been applied for cutting tools, high temperature ceramic crucibles in the mechanical and nuclear industry, electrical devices and the reinforcing particles in the composites [5–8] . Traditionally, ZrC is produced by the carbothermal reduction of ZrO 2 with C at elevated temperature, mechanical alloying, the solid-state metathesis route, chemical vapor deposition [9–12] . However, in most cases, these methods would result in impurity of the products, complication of the process and equipment. In recent years, as an effective and promising method to synthesize inorganic solids, combustion synthesis (CS) or self-propagating high-temperature synthesis (SHS) has received intensive interest in virtue of its relative simplicity of the equipment and process, higher purity of the products, and efficiency in energy and time. So far, a wide range of its application covers intermetallic compounds, advanced high-temperature materials, ceramic particles, reinforced-particle of composites and functional gradient materials [1] . At present, most studies concentrate on the synthesis of carbides and borides ceramic particles of Ti and Si with different metal additions, while the fabrication of ZrC particles through SHS route and the effect of additive metals such as Fe to the Zr–C mixtures on the reaction mechanism still remain relatively limited. However, to effectively control the SHS process and eventually to obtain the desired reaction products, knowledge of the reaction mechanism is of vital importance. Therefore, in this work, ZrC particles were prepared by SHS reaction from Fe–Zr–C powder mixtures, and the main purpose is to report the results of studies on the reaction mechanism of ZrC. It is expected that the insights into this work would be helpful to understand and control the SHS reaction in the Fe–Zr–C system. 2 Experimental procedure In the present study, the starting materials are commercial powders of Zr (98% purity, <38 μm), carbon black (>98% purity, in nano-scale) and Fe (99% purity, ∼10 μm). The samples were prepared with zirconium and carbon at a mole ratio corresponding to that of stoichiometric ZrC plus 0, 10, 20 or 30 wt.% Fe in the Fe–Zr–C powder mixtures, respectively. The reactant powders were mixed sufficiently in a conventional ball-milling pot for 5 h under argon atmosphere. Subsequently, the mixed powders were pressed into cylindrical preforms in dimensions of Ø22 mm × 10 mm with 65% theoretical density. The SHS experiments were carried out in a vacuum vessel full of Ar protective atmosphere. The reaction was ignited by a tungsten arc heating installed under the compact and the temperatures were measured by a thermocouple inserted into the preforms. The phase constituents of SHS and DTA products were identified by X-ray diffraction (D/Max 2500PC Rigaku, Japan) using Cu-K α radiation source. And the phase constituents of the quenched sample were determined by X-ray micro-diffraction (D8 Discover with GADDS, Bruker AXS, Germany). The observation of microstructure of SHS products and quenched sample was performed by field emission scanning electron microscopy (FE-SEM) (JSM-6700F, Japan). 3 Results and discussion 3.1 Phases and microstructures of SHS products Fig. 1 shows the typical XRD patterns of SHS products with (a) free, (b) 10 wt.%, (c) 20 wt.% and (d) 30 wt.% Fe contents, respectively. The XRD results reveal that, when Fe content in the reactants increases from 0 to 20 wt.%, only Fe and ZrC phases can be formed in the final products, without any zirconium iron compounds. It is suggested that the synthesis of ZrC ceramic particles in the Fe–Zr–C system is completely feasible by SHS reaction. Nevertheless, it should be noted that slight peaks of the Fe 2 Zr intermetallic compound were also identified in the SHS product with 30 wt.% Fe addition, indicating that the full conversion of ZrC has not been achieved once the addition of Fe exceeds 30 wt.%. It is reported [14] that when synthesizing the transition-metal carbides, the full conversion of Zr and C to ZrC is generally difficult, and the formation of the binary phase even ternary phase is hardly avoided. However, in our work, there is no apparent presence for those phases except the designed ZrC and Fe unless the Fe addition in the mixtures exceeds 30 wt.%. It can be explained that on the one hand, those phases may be indeed residual in the SHS products and are too slight to be obviously identified by XRD. On the other hand, compared to “bulky” and “crystalline” graphite, carbon black has a high reactivity for its non-crystalline and fine particle sizes, leading to the complete conversion of ZrC during SHS processing. Fig. 2 reveals the typical microstructures of SHS products with (a) free, (b) 10 wt.%, (c) 20 wt.% and (d) 30 wt.% Fe contents, respectively. As seen, ZrC particles are about 10 μm in size with an irregular block-like morphology in the case of free Fe addition; while in the case of 30 wt.% Fe addition, ZrC particles greatly decreased to nano-meter order with a spherical morphology. The increase of Fe in the mixtures would decrease the reaction temperature and the sizes of ZrC particles because the growth of grains is an exponential function of the combustion temperature [15] . On the other hand, as Fe content increases, the coefficient of thermal conductivity and cooling rate would also increase. Therefore, the dwell time of the reactant at high temperature would be shortened, which prevents the formed ZrC from growing. Song [13] synthesized ZrC particles via SHS reaction from 0 to 40 wt.% Al–Zr–C elemental powder mixtures and considered that, for low Al content (≤10 wt.%), the solid-state reaction between C and Zr particles is the important formation mechanism of ZrC phase during SHS processing. It is believed that formation mechanism of ZrC at low Fe (i.e. 0 and 10 wt.%) content should have the similar one and thus relatively coarse ZrC particles were produced. While in the case of 20 wt.% and 30 wt.% Fe addition, the ZrC particle size greatly decreased to nano-meter order with a spherical morphology. It is suspected that these finer ZrC particles were formed by solution–precipitation mechanism rather than through solid-state reaction. 3.2 Reaction mechanisms The typical temperature–time profiles for the SHS process of 10 and 30 wt.% Fe–Zr–C compacts are shown in Fig. 3 . In the case of 10 wt.% Fe–Zr–C system, the combustion temperature rapidly rose up to 2521 °C after a relative long time of preheating. The heat released from the tungsten arc heating could improve the reactivity of mixtures and induce the Zr–C solid-state reaction. While in the case of 30 wt.% Fe–Zr–C system, after preheating the compact for about 5 s at about 374 °C, some solid-state reactions characterized by temperature-fluctuating may take place. Subsequently, the temperature rapidly increased to about 1095 °C and then fell down. It can be deduced that during this period, some liquid phases appeared and undoubtedly consumed the heat released from solid-state reactions, resulting in the decreasing of temperature. However, this process is merely temporary. Because the liquids provide an easier route for the diffusion of powders, more Fe, Zr and C particles can spontaneously diffuse and dissolve into the liquids. As the carbon atoms in the liquid phase become sufficiently supersaturated, precipitation of ZrC grains occurred, in which a plenty of heat was liberated and the temperature abruptly increased (up to 1962 °C). To understand the reaction mechanism in the Fe–Zr–C system, DTA (DTA/TG-Rigaku, Japan) experiments for loose powder mixtures of Fe–Zr, Zr–C, 10 wt.% and 30 wt.% Fe–Zr–C were performed in these individual systems, respectively. A small amount of mixture weighing about 30 mg was held in an alumina crucible and heated to 1250 °C at a rate of 40 °C/min. Then the DTA samples were crushed into powders and analyzed for their phase constituents by XRD. The DTA curves and XRD patterns of those samples are presented in Figs. 4 and 5 , respectively. From the DAT curve of Fe–Zr system ( Fig. 4 (a)), it can be observed that only one endothermic peak at 763 °C and one exothermic peak at 1035 °C appeared. Obviously, the endothermic peak at 763 °C corresponds to the allotropic change Fe α → Fe γ . According to the XRD pattern in Fig. 5 (a), the DTA product consists of ZrO 2 and Fe 2 Zr and residual Fe, Zr. It suggests that solid-state diffusion reaction of Fe 2 Zr had occurred between Fe and Zr powders at 1035 °C. Since the Gibbs free energy Fe 2 Zr is about −76,000 J/mol [16] , the formation of Fe 2 Zr is thermodynamically feasible. At the same time, Fe 2 Zr is rather stable compared to other intermetallic compounds presented in the Fe–Zr system. The reason for the presence of a small amount of ZrO 2 is due to invasion of oxygen during the experiments. In the DTA curve of the Zr–C powder mixtures ( Fig. 4 (b)), there is only an exothermic peak at 1220 °C, and the XRD result ( Fig. 5 (b)) shows that the products consist of ZrO 2 and ZrC in the final product, which indicates that the evolved heat may result mainly from the solid-state formation reaction of ZrC. The ZrC-forming reaction was not fierce due to the limited contact surface areas between powders. Fig. 4 (c) shows the DTA curve of the 10 wt.% Fe–Zr–C powder mixtures. As seen, an endothermic appears at 748 °C, corresponding to the allotropic change of Fe. Continuous heating bring an exothermic at about 1208 °C and only ZrC, Zr and Fe phases were detected by XRD. It suggests that the ZrC was formed by the solid-state reaction between Zr and C and the Fe 2 Zr-forming reaction between of Fe and Zr had not took place, possibly due to the low Fe content. Comparing Fig. 4 (b) with (c), it can be found that the reaction in the 10 wt.% Fe–Zr–C system is similar to that in the Zr–C system, which further offer evidence that for low Fe contents in the reactants, the solid-state reaction between C and Zr particles is the main formation mechanism of ZrC. The DTA curve of the 30 wt.% Fe–Zr–C powder mixtures is presented in Fig. 4 (d). There are an endothermic peak at 752 °C and two exothermic peaks at about 1048 °C and 1172 °C, respectively. To identify the exothermic reaction at 1048 °C, the same DTA experiment were performed and timely interrupted just after being heated to 1048 °C. The XRD analysis ( Fig. 5 (d)) reveals that the DTA products consist of Fe, Zr, ZrO 2 and Fe 2 Zr phases, which implies that the exothermic event may mainly originate from the formation of the Fe 2 Zr phase. When the temperature increased to 1172 °C, a large number of ZrC particles were detected by XRD ( Fig. 5 (e)). Hence, the exothermal peak at 1172 °C results from the formation of the ZrC phase. However, there is no apparent endothermic event related to Fe–C liquid formation from the reaction between Fe and C at 1152 °C in view of the Fe–C phase diagram [17] , possibly due to the fast heating rate of 40 °C/min. Considering the ternary phase Fe–Zr–C, one would expect that during the heating up stage, the second reaction occurring in the Fe–Zr–C powder mixtures would be the formation of a liquid phase Fe–C at 1152 °C due to the spatial proximity of Fe and C particles and the relatively high diffusion coefficient of carbon. This reaction is endothermic, but the endothermic reaction probably will be superseded by the reaction of the already formed carbon-rich Fe-liquid with the Zr particles to form ZrC particles, which would be highly exothermic. Comparing Fig. 4 (a) with (d), it can be observed that the formation route of ZrC in the 30 wt.% Fe–Zr–C powder mixtures differed from that in the Zr–C powder mixtures. Yang [18] considered that the increase of metal content may change the reaction path and DTA curves. Moreover, the ZrC-forming temperature in the 30 wt.% Fe–Zr–C powder mixtures is lower than that in the Zr–C powder mixtures. Therefore, it is accepted that the addition of Fe to Zr–C reactants could promote the reaction occurrence by the prior formation of Fe–C liquid, which increases the contact surface area between powders and provides an easier route for powders to move. It should be noted that, a small quantity of Fe 2 Zr remains in the DTA products, which suggests that the reaction in the DTA apparatus tends to be incomplete with the increase of Fe content. The reason may be that the heat released by the little volume blends in the DTA apparatus dissipates into the surrounding too rapidly to achieve the full conversion of equilibrium phase. The information obtained from the DTA experiment is valuable to help understanding the reaction mechanism in the SHS processing. However, it may not be fully reliable since the experimental conditions, such as heating rate and green density of the reactants, differ considerably in the DTA and SHS processes, which could have significant influence on the reaction kinetics and mechanism [19] . Therefore, to further investigate the reaction mechanism of ZrC during the SHS process, it is necessary to quench the combustion wave during its passage through the sample. In a vacuum vessel full of Ar, the copper-mold-aided combustion front quenching experiment was conducted with 30 wt.% Fe in the compact. It was observed that the combustion wave propagated some distance and then extinguished by increasing heat loss. The quenched sample could be roughly divided into four regions, i.e. the unreacted region, preheated region, reacting region and reacted region in turn along the propagation direction of combustion wave front. According to their microstructural examinations in Fig. 6 and X-ray microdiffraction patterns in Fig. 7 , the formation mechanism of ZrC was analyzed. The microstructure of the initial reactants in the unreacted region is presented in Fig. 6 (a). As indicated, the bright particles were Fe, the grey ones represented Zr and the carbon black particles were too fine to be found. The XRD result in Fig. 7 (a) shows the presence of only original Zr and Fe phases in this region, which means that the powder mixtures were not affected by the heat generated at the combustion front. Fig. 6 (b) reveals the typical microstructures in the preheated region. Spectrum 1 totally consists of Fe and Zr elements, suggesting that Fe x Zr y compounds should be formed in this region. The XRD analysis in this region shows that Fe 2 Zr intermetallic compound was formed, as shown in Fig. 7 (b). The presence of Fe 2 Zr phase should be the Fe–Zr solid-state reaction triggered by heat released from the reacting region. Furthermore, the heat generated by the formation of the Fe 2 Zr phase may contribute to activating the subsequent primary reaction between Fe and C to form Fe–C liquid phase, even though the endothermic event is not observed in the DTA curve. The liquid phase provides an easier route for the movement of Zr particles. Subsequently, Zr atoms dissolved into the liquid and reacted with C to form ZrC particles, during which a vast of heat was released. Li [20] synthesized ZrC powder with zirconium and carbon powder mixtures and investigated the effects of NaCl on the combustion temperature during SHS process. The results showed that NaCl (boiling point of NaCl: 1413 °C) would be evaporated by the heat released from ZrC-forming reaction during the SHS process. It is deduced that the heat generated by the ZrC-forming reaction significantly promotes the binary eutectic in the Fe–Zr system between Fe and Fe 2 Zr at 1337 °C and the pseudobinary eutectic reaction Fe + ZrC = liquid at 1400 °C. Hence, more C particles diffused and dissolved into the liquids because of the capillary penetration and the increasing diffusion path of liquids. Once the thermodynamic and kinetic formation conditions of ZrC in the Fe–Zr–C ternary liquid phase are favorably met, the more thermodynamically stable ZrC particles would spontaneously precipitate from the melt, leading to an abrupt increased in temperature. Obviously, it is this reaction that initiates the SHS reaction and stimulates the combustion wave propagation. Fig. 6 (c) reveals the microstructure in the reacting region. There exist Fe and Zr elements in spectrum 2, corresponding to the Fe–Zr liquid phase. Comparing to spectrum 2, spectrum 3 consists of Fe, Zr and C elements, which means that the particles had developed at the periphery of Fe–Zr liquids should contain ZrC phase. The XRD analysis in this region shows the presence of Fe 2 Zr and ZrC phases, as indicated in Fig. 7 (c). With the ZrC-forming reaction proceeding, the precipitated ZrC and free Fe become more and more, while Fe 2 Zr intermetallic compound is correspondingly fewer. However, the nonequilibrium structures are expected to exist in the SHS powder because of high cooling rates and high defect concentration during the SHS process [21] . Finally, there existed Fe, Fe 2 Zr and ZrC phases in the reacted region. Fig. 6 (d) shows the microstructure in the reacted region. In this region, ZrC grains have greatly formed. The morphology of these ceramic grains is generally close to that in the reacting region. The presence of the remnant of Fe–Zr implies that the reaction was unable to get to completion even after the passage of the combustion wave. On the basis of the microstructural observation and phase constituents analysis on the combustion-wave quenched sample in combination with differential thermal analysis, the reaction mechanism for the synthesis of ZrC in the Fe–Zr–C system during the SHS process is proposed. As the temperature increased, the Fe 2 Zr intermetallic compounds were formed initially via the solid-state diffusion reaction between Fe and Zr powders. Utilizing the heat released by the Fe 2 Zr-forming reaction, Fe and C eutectic reaction took place and the Fe–C liquid phase was formed. Zr particles spontaneously dissolved into the Fe–C liquid and reacted with C to form ZrC and liberated a mass of heat. At the same time, the binary eutectic in the Fe–Zr system between Fe and Fe 2 Zr at 1337 °C and the pseudobinary eutectic reaction Fe + ZrC = liquid at 1400 °C occurred, resulting in the formation of more liquids. As more and more C particles continuously dissolved into the Fe–Zr–C liquid and ZrC particulates gradually precipitated from the saturated liquid. 4 Conclusion ZrC particles were successfully prepared through SHS route by adding 0–30 wt.% Fe to the Zr–C powder mixtures. The ZrC particles in the SHS products greatly decreased from about 10 μm with an irregular shape in free Fe addition to nano-meter order with a nearly spherical shape in 30 wt.% Fe addition. In the case of 0 wt.% and 10 wt.% Fe contents, the formation of ZrC was controlled by the solid-state reaction between C and Zr powders. While in the case of 20 wt.% and 30 wt.% Fe contents, the ZrC particles were governed by the solution–precipitation mechanism. The addition of Fe to Zr–C reactants promotes the occurrence of ZrC reaction. Acknowledgements This work was supported by the National Natural Science Foundation of China (Grant No. 50531030 and No. 50804030 ) and the special grade China Postdoctoral Science Foundation (Grant No. 200902230 ). The authors gratefully acknowledge the help of the Department of Materials Science and Engineering of Jilin University and the Instrumental Analysis Centre of SJTU. References [1] M.S. Song M.W. Ran Y.Y. Kong Int. J. Refract. Met. Hard Mater. 26 2008 392 [2] A.A. Mahday M.S. El-Eskandarany H.A. Ahmed A.A. Amer J. Alloy Compd. 299 2000 244 [3] J.Q. Han F. Wan Z.Y. Zhu Y. Liao T. Ji Appl. Phys. Lett. 87 2005 172107 [4] M.Q. Zhang J.J. He W.J. Liu M.L. Zhong Surf. Coat. Technol. 162 2003 140 [5] W.A. Mackie P.R. Davis IEEE Trans. Electron. Dev. 36 1989 220 [6] S. Shinada M. Nishisako J. Am. Ceram. Soc. 78 1995 41 [7] D. Karabi T.K. Bandyopadhyay Mater. Sci. Eng. A. 379 2004 83 [8] C.R. Rambo J. Cao Carbon 43 2005 1174 [9] A.M. Nartowski I.P. Parkin A.J. Craven M. Mackenzie Adv. Mater. 10 1998 805 [10] G.Z. Shen D. Chen J. Cryst. Growth 262 2004 277 [11] Y. Kurokawa S. Kobayashi J. Mater. Res. 13 1998 1275 [12] T. Tsuchisa M. Kawaguchi Solid State Ionics 101 1997 149 [13] M.S. Song M.X. Zhang B. Huang J.G. Li Powder Technol. 191 2009 19134 [14] U. Leela-adisorn S.M. Choi Ceram. Int. 32 2006 431 [15] Y. Choi Y.S.W. Rhee J. Mater. Sci. 28 1993 6669 [16] S. Chatain B. Larousse C. Maillault C. Guéneau C. Chatillon Alloys Compd. 220 1995 19 [17] P. Villars A. Prince H. Okamoto Handbook of Ternary Alloy Phase Diagrams 1995 ASM [18] Y.F. Yang H.Y. Wang R.Y. Zhao Y.H. Liang Q.C. Jiang Int. J. Refract. Met. Hard Mater. 26 2008 77 [19] P. Shen B.L. Zou S.B. Jin Q.C. Jiang Mater. Sci. Eng. A 454 2007 300 [20] J. Li Z.Y. Fu W.M. Wang H. Wang L. Soowohn N. Kochi J. Chin. Ceram. Soc. 38 2010 979 [21] A. Makino C.K. Low J. Am. Ceram. Soc. 77 1994 778