- Email: [email protected]
N composite powder and nano SiC powder at high frequencies
Physica E 9 (2001) 679–685
www.elsevier.nl/locate/physe
Dielectric properties of nano Si= C= N composite powder and nano SiC powder at high frequencies Donglin Zhao ∗ , Hongsheng Zhao, Wancheng Zhou State Key Laboratory of Solidiÿcation Processing, Northwestern Polytechnical University, Xi’an, Shaanxi, 710072, People’s Republic of China Received 27 July 1999; accepted 27 June 2000
Abstract The dielectric properties of nano Si=C=N composite powder and nano SiC powder at high frequencies have been studied. The nano Si=C=N composite powder and nano SiC powder were synthesized from hexamethyldisilazane ((Me3 Si)2 NH) (Me:CH3 ) and SiH4 –C2 H2 , respectively, by a laser-induced gas-phase reaction. The complex permittivities of the nano Si=C=N composite powder and nano SiC powder were measured at a frequency range of 8.2–12.4 GHz. The real part (0 ) and imaginary part (00 ) of the complex permittivity, and dissipation factor (tg = 00 =0 ) of nano Si=C=N composite powder are much higher than those of nano SiC powder and bulk SiC, Si3 N4 ; SiO2 , and Si, especially the tg . The promising features of nano Si=C=N composite powder would be due to more complicated Si, C, and N atomic chemical environment than in a mixture of pure SiC and Si3 N4 phase. The charged defects and quasi-free electrons moved in response to the electric eld, di usion or polarization current resulted from the eld propagation. Because there exists graphite in the nano Si=C=N composite powder, some charge carries are related to the sp3 dangling bonds (of silicon and carbon) and unsaturated sp2 carbons. The high 00 and tg of nano Si=C=N composite powder were due to the dielectric relaxation. The nano Si=C=N composite powder would be a good candidate for electromagnetic interface shielding material. ? 2001 Elsevier Science B.V. All rights reserved. Keywords: Nano Si=C=N composite powder; Nano SiC powder; Dielectric properties; Microstructure
1. Introduction Nano materials are increasingly receiving recognition as practical structural and functional materials ∗
Corresponding author. Current address: Institute of Carbon Fiber and Composites, Beijing University of Chemical Technology, Box 101, Beijing 100029, People’s Republic of China. Tel.: +86-10-64434914; fax: +86-10-64434653. E-mail address: [email protected] (D. Zhao).
with good prospects, and have been developed extensively in recent years. The nano Si=C=N composite powder and nano SiC powder have received an increasing interest since this kind of powder could be synthesized by laser-induced gas-phase reaction. Recently, nano Si=C=N composite materials have stimulated wide interest because they exhibit some new features compared with those of pure Si3 N4 and SiC phase mixtures. Several papers concerning the
1386-9477/01/$ - see front matter ? 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 9 4 7 7 ( 0 0 ) 0 0 1 9 6 - X
680
D. Zhao et al. / Physica E 9 (2001) 679–685
microstructure of ultra ne Si=C=N particles or deposits, suggest that promising features of nano Si=C=N composite materials would be due to more complicated Si, C, and N atomic chemical environment than in a mixture of pure SiC and Si3 N4 phase. Because of the decrement in particle size, some physical and mechanical properties of the nano materials will be changed greatly. The machinability and superplasticity properties were found for the hot-pressed Si3 N4 /SiC crystal nano-composite. The use of nano Si=C=N composite powder and nano SiC powder, involving a uniform dispersion of the phase, would greatly improve materials properties. The characterizations of the nano Si=C=N composite powder and nano SiC powder, such as chemical and phase composition, particle size, grain size, grain morphology, defect, speci c surface area, lattice constant, and local structure of microcrystallines, have been studied eciently. Much work has been reported on the synthesis, sinterability and heat-treatment of these nano powders [1–12]. However, to our knowledge, no work has been reported on the dielectric properties of the nano Si=C=N composite powder and nano SiC powder, especially at high frequencies. Our research work found that the nano Si=C=N composite powder exhibits some new feature of dielectric property at high frequencies compared with those of nano SiC powder and bulk SiC, Si3 N4 ; SiO2 , and Si. Dielectric speci c materials are required for microwave components such as isolators, circulators, phase shifters, lenses, and dielectric waveguides. Commercially available materials often limit the design of these devices including frequency selective surfaces such as bandpass and lowpass lters as well as radar absorbing materials (RAM) [13,14]. In our recent work, we found that the microwave permittivity and the dissipation factor of the nano Si=C=N composite powder are very high. In an e ort to nd a moldable material with a continuously variable dielectric constant for microwave applications, mixtures of paran wax and nano Si=C=N composite power were prepared in the laboratory and tested in the frequency range of 8.2–12.4 GHz. In comparison and investigating the interaction mechanism between the nano Si=C=N composite powder and microwave, the dielectric property of nano SiC powder and bulk SiC, Si3 N4 and SiO2 has been studied too.
2. Experimental procedure 2.1. Nano Si/C/N composite powder and SiC powder preparation The nano Si=C=N composite powder was synthesized from hexamethyldisilazane ((Me3 Si)2 NH) (Me:CH3 ) by a laser-induced gas-phase reaction. The nano SiC powder was synthesized from SiH4 and C2 H2 by the same method. The chemical composition of the Si=C=N composite powder was determined with a nitrogen analyzer (LECO-TN114) for the nitrogen content, with an oxygen analyzer (LECO-RO316) for the oxygen content, and with a carbon analyzer (LECO-CS344) for the carbon content. The silicon content was calculated as the remaining. The lattice constants of ÿ-SiC were measured through X-ray di raction (XRD) pattern for the ÿ-SiC (3 1 1) line and were corrected by highly pure Si (99.99%) as an external standard. The mean crystalline size of the particles was determined by the line width of di raction line using Scherrer’s method. The crystalline phase of the powder was analyzed by XRD (Cu K ). Microstructure characterization was performed by transmission electron microscopy (TEM), high-resolution electron microscopy (HRTEM) and energy-dispersive X-ray spectroscopy (EDXS). 2.2. Microwave permittivity measurement The permittivity of nano Si=C=N composite powder and nano SiC powder was measured by the method, which is based on measurements of the re ection and transmission moduli between 8.2 and 12.4 GHz, in the fundamental waveguide mode TE10 , using rectangular samples (10:16 × 22:86 mm2 ) set in a brass holder which lls the rectangular waveguide. After calibration with an intermediate of a short circuit and blank holder, re ection and transmission coecients were obtained with the help of an automated measuring system (HP8510C network analyzer). Both the real and imaginary parts of the permittivity and permeability were calculated. For dielectric materials (0 = 1; 00 = 0), the relative error varies between 1% (pure dielectric) and 10% (highly conductive materials). The nano Si=C=N composite powder and nano SiC powder were dispersed in melting paran wax, and then the mixtures were cast into molds
D. Zhao et al. / Physica E 9 (2001) 679–685
(10:16 × 22:86 × 2 mm3 ). Commercially available paran wax was used. Because the melting point ◦ of paran wax is low (∼ 60 C), the task of mixing liquid paran wax with the nano Si=C=N composite powder is simply a matter of blending the desired volumes of the two constituents in a suitable container until a uniform consistency is obtained. To prepare samples for testing, rectangular waveguide sections were lled carefully to prevent void formation. The samples consisted of 10 wt% nano Si=C=N composite powder or nano SiC powder and 90 wt% paran wax. 3. Results and discussion Fig. 1 shows the TEM photomicrographs of nano Si=C=N composite powder (Fig. 1(a)) and nano SiC powder (Fig. 1(b)). The TEM photomicrographs reveal that the powders are spherical with size in the range 20 –30 nm and loosely agglomerate. Only ÿ-SiC and -SiC are shown in the XRD pattern of nano Si=C=N composite powder. No Si3 N4 peaks are found. The mean crystalline size of ÿ-SiC in the nano Si=C=N composite powder calculated from the full-width at half-maximum of (1 1 1) peak using Scherrer’s method, is approximately 5.9 nm. The lattice constant of ÿ-SiC measured from the ÿ-SiC (3 . The chemical composition of 1 1) line, is 4.324 A the nano Si=C=N composites is silicon, 54.5 wt%, carbon, 26.4 wt%, nitrogen, 10.1 wt%, oxygen, 9.0 wt%. The peaks in the XRD pattern of nano SiC powder are indexed as ÿ-SiC and -SiC too. The mean crystalline size of ÿ-SiC in nano SiC is approximately 4.8 nm. The lattice constant of ÿ-SiC measured from the . ÿ-SiC (3 1 1) line, is 4.356 A Electrical properties can be determined at various frequencies [15 –20]. The interaction between electromagnetic waves and condensed matter can be described by using complex permittivity, ∗ (∗ = 0 + i00 , where 0 is the real part, 00 the imaginary part), and conductivity, ∗ . The relation between the real part of the (polarization) conductivity 0 (!) and the imaginary part of the permittivity 00 (!) is 0 (!) = ! 00 (!), where ! is the angular frequency. Fig. 2(a) shows 0 and 00 of the pure nano Si=C=N powder pressed at 100 MPa versus frequency (0 = 46:46–27.69, 00 = 57:89–38.556). Fig. 2(b)
681
shows tg of the pure nano Si=C=N composite powder (tg = 1:25–1.71). The density of this pure sample is 2:45 g=cm3 . In an e ort to nd a moldable material with a continuously variable dielectric constant for microwave applications, mixtures of paran wax and nano Si=C=N composite powder were prepared, because the melting point of paran wax is low and one can easily gain a uniform consistency. The 0 and 00 of paran wax are 2.26 and 0 at the frequency range of 8.2–12:4 GHz, respectively. Fig. 3(a) shows 0 and 00 of the nano Si=C=N particles imbedded in the paran wax matrix versus the frequency (0 = 5:79– 6.33, 00 = 3:51– 4.31). Fig. 3(b) shows tg of the nano Si=C=N composite powder imbedded in the paran wax matrix versus frequency (tg = 0:61– 0.68). The permittivities and dissipation factors of the nano Si=C=N composite powder imbedded in the paran wax matrix are much less than those of pure nano Si=C=N composite powder. So the permittivity of the mixture of nano Si=C=N composite powder and paran wax (or other dielectric materials) can be tailored by the content of this composite powder. Fig. 4(a) shows 0 and 00 of the nano SiC particles imbedded in the paran wax matrix versus the frequency (0 = 1:97–2.06, 00 = 0:09–0.19). Fig. 4(b) shows tg of the nano SiC powder imbedded in the paran wax matrix versus frequency (tg = 0:045– 0.094). So the 0 , 00 and tg of nano Si=C=N composite powder are much higher than those of nano SiC powder. The chemical composition of the nano Si=C=N composite powder shows that the nitrogen content is 10:1 wt%, while the XRD pattern shows only ÿ- and -SiC. No Si3 N4 peaks are found. So it is thought that the SiC microcrystalline in the nano Si=C=N composite powder dissolved a great deal of nitrogen. The EDXS analysis indicated that the chemical composition of SiC microcrystalline in nano Si=C=N composite powder is silicon, 54:86 wt%, carbon, 33:58 wt%, nitrogen, 9:53 wt%, oxygen, 2:03 wt%. The amount of dissolved nitrogen in a SiC–N solid solution has not been studied precisely; Komath has reported that the nitrogen content of the solid solution is at most about 0:3 wt% [21]. But considering the total nitrogen content, the particle homogeneity and EDXS analysis, it is likely that the amount of dissolved nitrogen is larger than that reported. Suzuki’s study revealed the same result [22].
682
D. Zhao et al. / Physica E 9 (2001) 679–685
Fig. 1. TEM photomicrographs of nano Si=C=N composite powder (a) and nano SiC powder (b).
Fig. 2. The 0 ; 00 (a) and tg (b) of pure nano Si=C=N composite powder versus frequency.
Fig. 3. The 0 ; 00 (a) and tg (b) of nano Si=C=N composite powder imbedded in the paran wax matrix versus frequency.
It is well known that nitrogen can exist in SiC as a solid solution, and nitrogen atoms are considered to occupy carbon sites [23]. Seo et al., reported that the lattice constant of SiC decreased by solubilization of nitrogen into SiC [24], and Suzuki found that as the nitrogen content increased, the lattice constant of the ÿ-SiC microcrystalline decreased rapidly and linearly
in nano Si=C=N composite powder [22]. However, the relationship between the amount of dissolved nitrogen and the lattice constant has been unclear. The atomic radii of nitrogen, carbon, and silicon are 0.070, 0.077, and 0:117 nm, respectively. The representative bond length is 0.188 nm for the Si–C bond and 0:173 nm for Si–N bond [25], so the lattice constant of SiC–N
D. Zhao et al. / Physica E 9 (2001) 679–685
683
Fig. 4. The 0 ; 00 (a) and tg (b) of nano SiC powder imbedded in the paran wax matrix versus frequency.
Fig. 5. Con guration of nitrogen atoms in SiC existing as a solid solution [27].
solid solution decreases. So the lattice constant of ÿ-SiC in nano Si=C=N composite powder is less than that of ÿ-SiC in nano SiC powder. The local structure of ÿ-SiC microcrystalline in nano Si=C=N composite powder has been studied by Suzuki [26]. In ÿ-SiC microcrystallines containing dissolved N, N atoms were considered to be tetrahedrally surrounded by four Si atoms in the lattice (SiC3=4 N1=4 ) (Fig. 5(a)). In this case, N atoms in the ÿ-SiC lattice should be paramagnetic centers by unpaired electrons because of the valence di erence between N and C. There were two possible nonparamagnetic N forms. One was a trivalent N bonded to three Si atoms (SiC3=4 N1=3 ) (Fig. 5(b)). If N atoms substitute for the C atoms, N could exist in a nonparamagnetic trivalent form (N(Si)3 ). The other possible nonparamagnetic form was a positively charged N atom bonded to four Si atoms (⊕N(Si)4 ) making pairs of negatively charged defects such as Si(C)3 or C(Si)3 [26]. Fig. 5 shows the structure of nitrogen existing in SiC as a solid solution.
In the high-frequency region, i.e. the infrared region, there are various kinds of dielectric resonance arising from ionic, molecular and (at higher energy) electronic motion. At lower frequency, i.e. in the microwave and radio frequency range, dipolar and space charge relaxation are observed. The low-frequency domain corresponds to free charges which can move in response to the electric eld (a di usion current results from the eld propagation), while the region at high frequencies is dominated by bound charges (dipoles) corresponding to the oscillating character of the electromagnetic wave (polarization current) [19,20,27]. Overlap with relaxation is often observed [27]. Much work has been done on the dielectric constant of bulk silicon carbide. The dielectric constant of pure single silicon carbide crystal is 10.2 at 100 kHz. The dissipation factors (tg ) are 3 × 10−3 (at 20 K) and 8 × 10−3 (at 70 K) [28]. Patric and Choyke found the static dielectric constant of cubic SiC to be s = 9:72, and s (⊥) = 9:66, and s (||) = 10:03 for 6HsiC [29]. Spitzer et al., reported that the optical dielectric constant of cubic silicon carbide ∞ is 6.7 [30]. From the literature the -values of diamond and silicon are known to be 5.7 and 11.7, respectively [28]. Chauvert et al., have studied the dielectric constant of SiC ber with SiC crystallites and graphitic clusters 2 nm in size embedded in an amorphous matrix at 0 –10 GHz. Their results show that the dielectric constant of the SiC ber is very high at 100 kHz ( ∼ 104 ), remains high even at 1 GHz ( ∼ 400), and drops to the usual value ( ∼ 20) only at GHz range. Charge carries in the SiC ber are related to sp3 dangling bonds (of silicon and carbon) and unsaturated sp2 carbons [31].
684
D. Zhao et al. / Physica E 9 (2001) 679–685
In an e ort to nd materials with high dissipation factor for microwave applications, we have investigated the dielectric property of the bulk materials such as SiO2 , hot-pressed SiC and Si3 N4 at 10 GHz. The dielectric constants of bulk SiC, Si3 N4 and SiO2 are 8.32, 9.47 and 3.80, respectively, while the dissipation factors tg are 0.12, 0.003 and 0.002. From the above discussions, the dielectric property of the nano Si=C=N composite powder is very di erent from those of bulk SiC, Si3 N4 , Si, SiO2 and nano SiC. The 0 , 00 and tg of nano Si=C=N composite powder are much higher than those of nano SiC powder and bulk SiC, Si3 N4 , SiO2 , and Si, especially the tg . The 0 , 00 and tg of the nano Si=C=N composite powder imbedded in the paran wax matrix are much less than those of pure nano Si=C=N composite powder. So the permittivity of the mixture of nano Si=C=N composite powder and paran wax (or other dielectric materials) can be tailored by the content of this composite powder. The nano Si=C=N composite powder would be a good candidate for EMI shielding material. In the nano Si=C=N composite powder, SiC microcrystallines dissolved a great deal of nitrogen. Nitrogen atoms exist in SiC as a solid solution, and nitrogen atoms occupy carbon sites. There will exist charged defects and unpaired electrons. Because there exists graphite in the nano Si=C=N composite powder, some charge carries are related to the sp3 dangling bonds (of silicon and carbon) and unsaturated sp2 carbons too. The charged defects and quasi-free electrons would move in response to the electric eld, and di usion or polarization current results from the eld propagation. The dielectric relaxation increased the 00 and tg . The ÿ-SiC and -SiC in nano SiC powder dissolved no nitrogen; therefore the charged defects and unpaired electrons in the nano SiC powder are much less than those in nano Si=C=N composite powder. This is the main reason why 00 and loss factor tg of nano Si=C=N composite powder are much higher than those of nano SiC powder. 4. Conclusion (1) The dielectric properties of the nano Si=C=N composite powder and nano SiC powder prepared by a laser-induced gas-phase reaction were measured at
the frequency range of 8.2–12:4 GHz. The permittivities and dissipation factors of the nano Si=C=N composite powder imbedded in the paran wax matrix are much less than those of pure nano Si=C=N composite powder. So the permittivity of the mixture of nano Si=C=N composite powder and paran wax (or other dielectric materials) can be tailored by the content of this composite powder. The nano Si=C=N composite powder would be a good candidate for EMI shielding material. (2) The dielectric properties of the Si=C=N composite powder are very di erent from those of bulk SiC, Si3 N4 , Si, SiO2 and nano SiC. The 0 , 00 and tg of nano Si=C=N composite powder are much higher than those of nano SiC powder and bulk SiC, Si3 N4 , SiO2 , and Si, especially the tg . The promising features of nano Si=C=N composite powder would be due to more complicated Si, C, and N atomic chemical environment than in a mixture of pure SiC and Si3 N4 phase. (3) In the nano Si=C=N composite powder, the Si, C, and N atoms were intimately mixed. The SiC microcrystallines in the nano composite powder dissolved a great deal of nitrogen. So the charged defects and quasi-free electrons moved in response to the electric eld, and a di usion or polarization current resulted from the eld propagation. Because there exists graphite in the nano Si=C=N composite powder, some charge carries are related to the sp3 dangling bonds (of silicon and carbon) and unsaturated sp2 carbons. The high 00 and loss factor tg were due to the dielectric relaxation. (4) The ÿ- and -SiC in nano SiC powder dissolved no nitrogen; therefore, the charged defects and unpaired electrons in the nano SiC powder are much less than those in nano Si=C=N composite powder. This is the main reason why 00 and loss factor tg of nano Si=C=N composite powder are much higher than those of nano SiC powder. References [1] K. Niihara, K. Izaki, T. Kawakami, J. Mater. Sci. Lett. 10 (1990) 112. [2] K. Niihara, K. Nakahira, Ann. Chem. (Paris) 16 (1991) 479. [3] M. Cauchetier, N. Ocroix, N. Herlin, M. Luce, J. Am. Ceram. Soc. 77 (1994) 993. [4] K.E. Gonsalves, P.R. Strutt, T.K. Xiao, P.G. Clemeans, J. Mater. Sci. 27 (1992) 3231.
D. Zhao et al. / Physica E 9 (2001) 679–685 [5] K. Niihara, J. Ceram. Soc. Japan 99 (1991) 974. [6] K. Niihara, K. Izaki, A. Nakahira, J. Jpn. Soc. Powder Powder Metall. 37 (1990) 352. [7] X.G. Li, A. Chiba, Y. Nakata, H. Nagai, M. Suzuki, Mater. Sci. Eng. A 219 (1996) 95. [8] S.M. Dong, D.L. Jiang, S.H. Tan, J.G. Guo, J. Mater. Sci. Lett. 16 (1997) 1080. [9] G.W. Rice, J. Am. Ceram. Soc. 69 (1986) C-183. [10] D. Bahloul, M. Pereira, P. Goursat, J. Am. Ceram. Soc. 76 (1993) 1156. [11] A. Bendeddouche, R. Berjoan, E. Beche et al., J. Appl. Phys. 81 (1997) 6147. [12] R. Riedel, G. Passing, H. Schonfelder, R.J. Brook, Nature 355 (1992) 714. [13] K. Bober, R.H. Giles, J. Waldman, Int. J. Infrared Millimeter Waves 18 (1997) 101. [14] G.F. Dionne, J.F. Fitzgerald, R.C. Aucoin, J. Appl. Phys. 47 (1976) 1708. [15] N.F. Mott, E.A. Davis, Electronic Processes in NonCrystalline Materials, Clarendon, Oxford, 1979, p. 149. [16] S.R. Elliot, A.P. Owens, Philos. Mag. B 80 (1989) 777. [17] H. Kahnt, Ber. Bunsenges, Phys. Chem. 95 (1991) 1021.
685
[18] A.K. Jonscher, Dielectric Relaxation in Solids, Chelsea Dielectric Press, London, 1983. [19] P. Debye, Polar Molecules, Dover, New York, 1945. [20] E. Mouchon, Ph. Colomban, J. Mater. Sci. 31 (1996) 323. [21] G.S. Kamath, Mater. Res. Bull. 4 (1969) S57. [22] M. Suzuki, Y. Maniette, Y. Nakata, T. Okutani, J. Am. Ceram. Soc. 76 (1993) 1195. [23] G.E.G. Hardeman, J. Phys. Chem. Solid. 24 (1963) 1223. [24] W.S. Seo, C.H. Pai, K. Koumoto, H. Yanagida, Seramikkusu Ronbun-shi 99 (1991) 443. [25] L. Pauling, The Nature of the Chemical Bond, 3rd Edition, Cornell University Press, Ithaca, New York, 1960, p. 137. [26] M. Suzuki, Y. Hasegawa, M. Aizawa, Y. Nakara, T. Okutani, J. Am. Ceram. Soc. 78 (1995) 83. [27] Ph. Colomban, J.C. Badot, Solid State Ionics 61 (1993) 55. [28] D. Hofman, J.A. Lely, J. Volger, Physica 23 (1967) 236. [29] L. Patrick, W.J. Choyke, Phys. Rev. B 2 (1970) 2255. [30] W.G. Spitzer, D.A. Kleinman, C.J. Frosch, Phys. Rev. 113 (1959) 133. [31] O. Chauvert, T. Stoto, L. Zuppiroli, Phys. Rev. B 46 (1992) 8139.
www.elsevier.nl/locate/physe
Dielectric properties of nano Si= C= N composite powder and nano SiC powder at high frequencies Donglin Zhao ∗ , Hongsheng Zhao, Wancheng Zhou State Key Laboratory of Solidiÿcation Processing, Northwestern Polytechnical University, Xi’an, Shaanxi, 710072, People’s Republic of China Received 27 July 1999; accepted 27 June 2000
Abstract The dielectric properties of nano Si=C=N composite powder and nano SiC powder at high frequencies have been studied. The nano Si=C=N composite powder and nano SiC powder were synthesized from hexamethyldisilazane ((Me3 Si)2 NH) (Me:CH3 ) and SiH4 –C2 H2 , respectively, by a laser-induced gas-phase reaction. The complex permittivities of the nano Si=C=N composite powder and nano SiC powder were measured at a frequency range of 8.2–12.4 GHz. The real part (0 ) and imaginary part (00 ) of the complex permittivity, and dissipation factor (tg = 00 =0 ) of nano Si=C=N composite powder are much higher than those of nano SiC powder and bulk SiC, Si3 N4 ; SiO2 , and Si, especially the tg . The promising features of nano Si=C=N composite powder would be due to more complicated Si, C, and N atomic chemical environment than in a mixture of pure SiC and Si3 N4 phase. The charged defects and quasi-free electrons moved in response to the electric eld, di usion or polarization current resulted from the eld propagation. Because there exists graphite in the nano Si=C=N composite powder, some charge carries are related to the sp3 dangling bonds (of silicon and carbon) and unsaturated sp2 carbons. The high 00 and tg of nano Si=C=N composite powder were due to the dielectric relaxation. The nano Si=C=N composite powder would be a good candidate for electromagnetic interface shielding material. ? 2001 Elsevier Science B.V. All rights reserved. Keywords: Nano Si=C=N composite powder; Nano SiC powder; Dielectric properties; Microstructure
1. Introduction Nano materials are increasingly receiving recognition as practical structural and functional materials ∗
Corresponding author. Current address: Institute of Carbon Fiber and Composites, Beijing University of Chemical Technology, Box 101, Beijing 100029, People’s Republic of China. Tel.: +86-10-64434914; fax: +86-10-64434653. E-mail address: [email protected] (D. Zhao).
with good prospects, and have been developed extensively in recent years. The nano Si=C=N composite powder and nano SiC powder have received an increasing interest since this kind of powder could be synthesized by laser-induced gas-phase reaction. Recently, nano Si=C=N composite materials have stimulated wide interest because they exhibit some new features compared with those of pure Si3 N4 and SiC phase mixtures. Several papers concerning the
1386-9477/01/$ - see front matter ? 2001 Elsevier Science B.V. All rights reserved. PII: S 1 3 8 6 - 9 4 7 7 ( 0 0 ) 0 0 1 9 6 - X
680
D. Zhao et al. / Physica E 9 (2001) 679–685
microstructure of ultra ne Si=C=N particles or deposits, suggest that promising features of nano Si=C=N composite materials would be due to more complicated Si, C, and N atomic chemical environment than in a mixture of pure SiC and Si3 N4 phase. Because of the decrement in particle size, some physical and mechanical properties of the nano materials will be changed greatly. The machinability and superplasticity properties were found for the hot-pressed Si3 N4 /SiC crystal nano-composite. The use of nano Si=C=N composite powder and nano SiC powder, involving a uniform dispersion of the phase, would greatly improve materials properties. The characterizations of the nano Si=C=N composite powder and nano SiC powder, such as chemical and phase composition, particle size, grain size, grain morphology, defect, speci c surface area, lattice constant, and local structure of microcrystallines, have been studied eciently. Much work has been reported on the synthesis, sinterability and heat-treatment of these nano powders [1–12]. However, to our knowledge, no work has been reported on the dielectric properties of the nano Si=C=N composite powder and nano SiC powder, especially at high frequencies. Our research work found that the nano Si=C=N composite powder exhibits some new feature of dielectric property at high frequencies compared with those of nano SiC powder and bulk SiC, Si3 N4 ; SiO2 , and Si. Dielectric speci c materials are required for microwave components such as isolators, circulators, phase shifters, lenses, and dielectric waveguides. Commercially available materials often limit the design of these devices including frequency selective surfaces such as bandpass and lowpass lters as well as radar absorbing materials (RAM) [13,14]. In our recent work, we found that the microwave permittivity and the dissipation factor of the nano Si=C=N composite powder are very high. In an e ort to nd a moldable material with a continuously variable dielectric constant for microwave applications, mixtures of paran wax and nano Si=C=N composite power were prepared in the laboratory and tested in the frequency range of 8.2–12.4 GHz. In comparison and investigating the interaction mechanism between the nano Si=C=N composite powder and microwave, the dielectric property of nano SiC powder and bulk SiC, Si3 N4 and SiO2 has been studied too.
2. Experimental procedure 2.1. Nano Si/C/N composite powder and SiC powder preparation The nano Si=C=N composite powder was synthesized from hexamethyldisilazane ((Me3 Si)2 NH) (Me:CH3 ) by a laser-induced gas-phase reaction. The nano SiC powder was synthesized from SiH4 and C2 H2 by the same method. The chemical composition of the Si=C=N composite powder was determined with a nitrogen analyzer (LECO-TN114) for the nitrogen content, with an oxygen analyzer (LECO-RO316) for the oxygen content, and with a carbon analyzer (LECO-CS344) for the carbon content. The silicon content was calculated as the remaining. The lattice constants of ÿ-SiC were measured through X-ray di raction (XRD) pattern for the ÿ-SiC (3 1 1) line and were corrected by highly pure Si (99.99%) as an external standard. The mean crystalline size of the particles was determined by the line width of di raction line using Scherrer’s method. The crystalline phase of the powder was analyzed by XRD (Cu K ). Microstructure characterization was performed by transmission electron microscopy (TEM), high-resolution electron microscopy (HRTEM) and energy-dispersive X-ray spectroscopy (EDXS). 2.2. Microwave permittivity measurement The permittivity of nano Si=C=N composite powder and nano SiC powder was measured by the method, which is based on measurements of the re ection and transmission moduli between 8.2 and 12.4 GHz, in the fundamental waveguide mode TE10 , using rectangular samples (10:16 × 22:86 mm2 ) set in a brass holder which lls the rectangular waveguide. After calibration with an intermediate of a short circuit and blank holder, re ection and transmission coecients were obtained with the help of an automated measuring system (HP8510C network analyzer). Both the real and imaginary parts of the permittivity and permeability were calculated. For dielectric materials (0 = 1; 00 = 0), the relative error varies between 1% (pure dielectric) and 10% (highly conductive materials). The nano Si=C=N composite powder and nano SiC powder were dispersed in melting paran wax, and then the mixtures were cast into molds
D. Zhao et al. / Physica E 9 (2001) 679–685
(10:16 × 22:86 × 2 mm3 ). Commercially available paran wax was used. Because the melting point ◦ of paran wax is low (∼ 60 C), the task of mixing liquid paran wax with the nano Si=C=N composite powder is simply a matter of blending the desired volumes of the two constituents in a suitable container until a uniform consistency is obtained. To prepare samples for testing, rectangular waveguide sections were lled carefully to prevent void formation. The samples consisted of 10 wt% nano Si=C=N composite powder or nano SiC powder and 90 wt% paran wax. 3. Results and discussion Fig. 1 shows the TEM photomicrographs of nano Si=C=N composite powder (Fig. 1(a)) and nano SiC powder (Fig. 1(b)). The TEM photomicrographs reveal that the powders are spherical with size in the range 20 –30 nm and loosely agglomerate. Only ÿ-SiC and -SiC are shown in the XRD pattern of nano Si=C=N composite powder. No Si3 N4 peaks are found. The mean crystalline size of ÿ-SiC in the nano Si=C=N composite powder calculated from the full-width at half-maximum of (1 1 1) peak using Scherrer’s method, is approximately 5.9 nm. The lattice constant of ÿ-SiC measured from the ÿ-SiC (3 . The chemical composition of 1 1) line, is 4.324 A the nano Si=C=N composites is silicon, 54.5 wt%, carbon, 26.4 wt%, nitrogen, 10.1 wt%, oxygen, 9.0 wt%. The peaks in the XRD pattern of nano SiC powder are indexed as ÿ-SiC and -SiC too. The mean crystalline size of ÿ-SiC in nano SiC is approximately 4.8 nm. The lattice constant of ÿ-SiC measured from the . ÿ-SiC (3 1 1) line, is 4.356 A Electrical properties can be determined at various frequencies [15 –20]. The interaction between electromagnetic waves and condensed matter can be described by using complex permittivity, ∗ (∗ = 0 + i00 , where 0 is the real part, 00 the imaginary part), and conductivity, ∗ . The relation between the real part of the (polarization) conductivity 0 (!) and the imaginary part of the permittivity 00 (!) is 0 (!) = ! 00 (!), where ! is the angular frequency. Fig. 2(a) shows 0 and 00 of the pure nano Si=C=N powder pressed at 100 MPa versus frequency (0 = 46:46–27.69, 00 = 57:89–38.556). Fig. 2(b)
681
shows tg of the pure nano Si=C=N composite powder (tg = 1:25–1.71). The density of this pure sample is 2:45 g=cm3 . In an e ort to nd a moldable material with a continuously variable dielectric constant for microwave applications, mixtures of paran wax and nano Si=C=N composite powder were prepared, because the melting point of paran wax is low and one can easily gain a uniform consistency. The 0 and 00 of paran wax are 2.26 and 0 at the frequency range of 8.2–12:4 GHz, respectively. Fig. 3(a) shows 0 and 00 of the nano Si=C=N particles imbedded in the paran wax matrix versus the frequency (0 = 5:79– 6.33, 00 = 3:51– 4.31). Fig. 3(b) shows tg of the nano Si=C=N composite powder imbedded in the paran wax matrix versus frequency (tg = 0:61– 0.68). The permittivities and dissipation factors of the nano Si=C=N composite powder imbedded in the paran wax matrix are much less than those of pure nano Si=C=N composite powder. So the permittivity of the mixture of nano Si=C=N composite powder and paran wax (or other dielectric materials) can be tailored by the content of this composite powder. Fig. 4(a) shows 0 and 00 of the nano SiC particles imbedded in the paran wax matrix versus the frequency (0 = 1:97–2.06, 00 = 0:09–0.19). Fig. 4(b) shows tg of the nano SiC powder imbedded in the paran wax matrix versus frequency (tg = 0:045– 0.094). So the 0 , 00 and tg of nano Si=C=N composite powder are much higher than those of nano SiC powder. The chemical composition of the nano Si=C=N composite powder shows that the nitrogen content is 10:1 wt%, while the XRD pattern shows only ÿ- and -SiC. No Si3 N4 peaks are found. So it is thought that the SiC microcrystalline in the nano Si=C=N composite powder dissolved a great deal of nitrogen. The EDXS analysis indicated that the chemical composition of SiC microcrystalline in nano Si=C=N composite powder is silicon, 54:86 wt%, carbon, 33:58 wt%, nitrogen, 9:53 wt%, oxygen, 2:03 wt%. The amount of dissolved nitrogen in a SiC–N solid solution has not been studied precisely; Komath has reported that the nitrogen content of the solid solution is at most about 0:3 wt% [21]. But considering the total nitrogen content, the particle homogeneity and EDXS analysis, it is likely that the amount of dissolved nitrogen is larger than that reported. Suzuki’s study revealed the same result [22].
682
D. Zhao et al. / Physica E 9 (2001) 679–685
Fig. 1. TEM photomicrographs of nano Si=C=N composite powder (a) and nano SiC powder (b).
Fig. 2. The 0 ; 00 (a) and tg (b) of pure nano Si=C=N composite powder versus frequency.
Fig. 3. The 0 ; 00 (a) and tg (b) of nano Si=C=N composite powder imbedded in the paran wax matrix versus frequency.
It is well known that nitrogen can exist in SiC as a solid solution, and nitrogen atoms are considered to occupy carbon sites [23]. Seo et al., reported that the lattice constant of SiC decreased by solubilization of nitrogen into SiC [24], and Suzuki found that as the nitrogen content increased, the lattice constant of the ÿ-SiC microcrystalline decreased rapidly and linearly
in nano Si=C=N composite powder [22]. However, the relationship between the amount of dissolved nitrogen and the lattice constant has been unclear. The atomic radii of nitrogen, carbon, and silicon are 0.070, 0.077, and 0:117 nm, respectively. The representative bond length is 0.188 nm for the Si–C bond and 0:173 nm for Si–N bond [25], so the lattice constant of SiC–N
D. Zhao et al. / Physica E 9 (2001) 679–685
683
Fig. 4. The 0 ; 00 (a) and tg (b) of nano SiC powder imbedded in the paran wax matrix versus frequency.
Fig. 5. Con guration of nitrogen atoms in SiC existing as a solid solution [27].
solid solution decreases. So the lattice constant of ÿ-SiC in nano Si=C=N composite powder is less than that of ÿ-SiC in nano SiC powder. The local structure of ÿ-SiC microcrystalline in nano Si=C=N composite powder has been studied by Suzuki [26]. In ÿ-SiC microcrystallines containing dissolved N, N atoms were considered to be tetrahedrally surrounded by four Si atoms in the lattice (SiC3=4 N1=4 ) (Fig. 5(a)). In this case, N atoms in the ÿ-SiC lattice should be paramagnetic centers by unpaired electrons because of the valence di erence between N and C. There were two possible nonparamagnetic N forms. One was a trivalent N bonded to three Si atoms (SiC3=4 N1=3 ) (Fig. 5(b)). If N atoms substitute for the C atoms, N could exist in a nonparamagnetic trivalent form (N(Si)3 ). The other possible nonparamagnetic form was a positively charged N atom bonded to four Si atoms (⊕N(Si)4 ) making pairs of negatively charged defects such as Si(C)3 or C(Si)3 [26]. Fig. 5 shows the structure of nitrogen existing in SiC as a solid solution.
In the high-frequency region, i.e. the infrared region, there are various kinds of dielectric resonance arising from ionic, molecular and (at higher energy) electronic motion. At lower frequency, i.e. in the microwave and radio frequency range, dipolar and space charge relaxation are observed. The low-frequency domain corresponds to free charges which can move in response to the electric eld (a di usion current results from the eld propagation), while the region at high frequencies is dominated by bound charges (dipoles) corresponding to the oscillating character of the electromagnetic wave (polarization current) [19,20,27]. Overlap with relaxation is often observed [27]. Much work has been done on the dielectric constant of bulk silicon carbide. The dielectric constant of pure single silicon carbide crystal is 10.2 at 100 kHz. The dissipation factors (tg ) are 3 × 10−3 (at 20 K) and 8 × 10−3 (at 70 K) [28]. Patric and Choyke found the static dielectric constant of cubic SiC to be s = 9:72, and s (⊥) = 9:66, and s (||) = 10:03 for 6HsiC [29]. Spitzer et al., reported that the optical dielectric constant of cubic silicon carbide ∞ is 6.7 [30]. From the literature the -values of diamond and silicon are known to be 5.7 and 11.7, respectively [28]. Chauvert et al., have studied the dielectric constant of SiC ber with SiC crystallites and graphitic clusters 2 nm in size embedded in an amorphous matrix at 0 –10 GHz. Their results show that the dielectric constant of the SiC ber is very high at 100 kHz ( ∼ 104 ), remains high even at 1 GHz ( ∼ 400), and drops to the usual value ( ∼ 20) only at GHz range. Charge carries in the SiC ber are related to sp3 dangling bonds (of silicon and carbon) and unsaturated sp2 carbons [31].
684
D. Zhao et al. / Physica E 9 (2001) 679–685
In an e ort to nd materials with high dissipation factor for microwave applications, we have investigated the dielectric property of the bulk materials such as SiO2 , hot-pressed SiC and Si3 N4 at 10 GHz. The dielectric constants of bulk SiC, Si3 N4 and SiO2 are 8.32, 9.47 and 3.80, respectively, while the dissipation factors tg are 0.12, 0.003 and 0.002. From the above discussions, the dielectric property of the nano Si=C=N composite powder is very di erent from those of bulk SiC, Si3 N4 , Si, SiO2 and nano SiC. The 0 , 00 and tg of nano Si=C=N composite powder are much higher than those of nano SiC powder and bulk SiC, Si3 N4 , SiO2 , and Si, especially the tg . The 0 , 00 and tg of the nano Si=C=N composite powder imbedded in the paran wax matrix are much less than those of pure nano Si=C=N composite powder. So the permittivity of the mixture of nano Si=C=N composite powder and paran wax (or other dielectric materials) can be tailored by the content of this composite powder. The nano Si=C=N composite powder would be a good candidate for EMI shielding material. In the nano Si=C=N composite powder, SiC microcrystallines dissolved a great deal of nitrogen. Nitrogen atoms exist in SiC as a solid solution, and nitrogen atoms occupy carbon sites. There will exist charged defects and unpaired electrons. Because there exists graphite in the nano Si=C=N composite powder, some charge carries are related to the sp3 dangling bonds (of silicon and carbon) and unsaturated sp2 carbons too. The charged defects and quasi-free electrons would move in response to the electric eld, and di usion or polarization current results from the eld propagation. The dielectric relaxation increased the 00 and tg . The ÿ-SiC and -SiC in nano SiC powder dissolved no nitrogen; therefore the charged defects and unpaired electrons in the nano SiC powder are much less than those in nano Si=C=N composite powder. This is the main reason why 00 and loss factor tg of nano Si=C=N composite powder are much higher than those of nano SiC powder. 4. Conclusion (1) The dielectric properties of the nano Si=C=N composite powder and nano SiC powder prepared by a laser-induced gas-phase reaction were measured at
the frequency range of 8.2–12:4 GHz. The permittivities and dissipation factors of the nano Si=C=N composite powder imbedded in the paran wax matrix are much less than those of pure nano Si=C=N composite powder. So the permittivity of the mixture of nano Si=C=N composite powder and paran wax (or other dielectric materials) can be tailored by the content of this composite powder. The nano Si=C=N composite powder would be a good candidate for EMI shielding material. (2) The dielectric properties of the Si=C=N composite powder are very di erent from those of bulk SiC, Si3 N4 , Si, SiO2 and nano SiC. The 0 , 00 and tg of nano Si=C=N composite powder are much higher than those of nano SiC powder and bulk SiC, Si3 N4 , SiO2 , and Si, especially the tg . The promising features of nano Si=C=N composite powder would be due to more complicated Si, C, and N atomic chemical environment than in a mixture of pure SiC and Si3 N4 phase. (3) In the nano Si=C=N composite powder, the Si, C, and N atoms were intimately mixed. The SiC microcrystallines in the nano composite powder dissolved a great deal of nitrogen. So the charged defects and quasi-free electrons moved in response to the electric eld, and a di usion or polarization current resulted from the eld propagation. Because there exists graphite in the nano Si=C=N composite powder, some charge carries are related to the sp3 dangling bonds (of silicon and carbon) and unsaturated sp2 carbons. The high 00 and loss factor tg were due to the dielectric relaxation. (4) The ÿ- and -SiC in nano SiC powder dissolved no nitrogen; therefore, the charged defects and unpaired electrons in the nano SiC powder are much less than those in nano Si=C=N composite powder. This is the main reason why 00 and loss factor tg of nano Si=C=N composite powder are much higher than those of nano SiC powder. References [1] K. Niihara, K. Izaki, T. Kawakami, J. Mater. Sci. Lett. 10 (1990) 112. [2] K. Niihara, K. Nakahira, Ann. Chem. (Paris) 16 (1991) 479. [3] M. Cauchetier, N. Ocroix, N. Herlin, M. Luce, J. Am. Ceram. Soc. 77 (1994) 993. [4] K.E. Gonsalves, P.R. Strutt, T.K. Xiao, P.G. Clemeans, J. Mater. Sci. 27 (1992) 3231.
D. Zhao et al. / Physica E 9 (2001) 679–685 [5] K. Niihara, J. Ceram. Soc. Japan 99 (1991) 974. [6] K. Niihara, K. Izaki, A. Nakahira, J. Jpn. Soc. Powder Powder Metall. 37 (1990) 352. [7] X.G. Li, A. Chiba, Y. Nakata, H. Nagai, M. Suzuki, Mater. Sci. Eng. A 219 (1996) 95. [8] S.M. Dong, D.L. Jiang, S.H. Tan, J.G. Guo, J. Mater. Sci. Lett. 16 (1997) 1080. [9] G.W. Rice, J. Am. Ceram. Soc. 69 (1986) C-183. [10] D. Bahloul, M. Pereira, P. Goursat, J. Am. Ceram. Soc. 76 (1993) 1156. [11] A. Bendeddouche, R. Berjoan, E. Beche et al., J. Appl. Phys. 81 (1997) 6147. [12] R. Riedel, G. Passing, H. Schonfelder, R.J. Brook, Nature 355 (1992) 714. [13] K. Bober, R.H. Giles, J. Waldman, Int. J. Infrared Millimeter Waves 18 (1997) 101. [14] G.F. Dionne, J.F. Fitzgerald, R.C. Aucoin, J. Appl. Phys. 47 (1976) 1708. [15] N.F. Mott, E.A. Davis, Electronic Processes in NonCrystalline Materials, Clarendon, Oxford, 1979, p. 149. [16] S.R. Elliot, A.P. Owens, Philos. Mag. B 80 (1989) 777. [17] H. Kahnt, Ber. Bunsenges, Phys. Chem. 95 (1991) 1021.
685
[18] A.K. Jonscher, Dielectric Relaxation in Solids, Chelsea Dielectric Press, London, 1983. [19] P. Debye, Polar Molecules, Dover, New York, 1945. [20] E. Mouchon, Ph. Colomban, J. Mater. Sci. 31 (1996) 323. [21] G.S. Kamath, Mater. Res. Bull. 4 (1969) S57. [22] M. Suzuki, Y. Maniette, Y. Nakata, T. Okutani, J. Am. Ceram. Soc. 76 (1993) 1195. [23] G.E.G. Hardeman, J. Phys. Chem. Solid. 24 (1963) 1223. [24] W.S. Seo, C.H. Pai, K. Koumoto, H. Yanagida, Seramikkusu Ronbun-shi 99 (1991) 443. [25] L. Pauling, The Nature of the Chemical Bond, 3rd Edition, Cornell University Press, Ithaca, New York, 1960, p. 137. [26] M. Suzuki, Y. Hasegawa, M. Aizawa, Y. Nakara, T. Okutani, J. Am. Ceram. Soc. 78 (1995) 83. [27] Ph. Colomban, J.C. Badot, Solid State Ionics 61 (1993) 55. [28] D. Hofman, J.A. Lely, J. Volger, Physica 23 (1967) 236. [29] L. Patrick, W.J. Choyke, Phys. Rev. B 2 (1970) 2255. [30] W.G. Spitzer, D.A. Kleinman, C.J. Frosch, Phys. Rev. 113 (1959) 133. [31] O. Chauvert, T. Stoto, L. Zuppiroli, Phys. Rev. B 46 (1992) 8139.