Abstract
Silicon nitride (Si3 N4) nanowires have been prepared by carbothermal reduction followed by the nitridation (CTRN) of silica gel containing ultrafine excess carbon obtained by the decomposition of dextrose over the temperature range of 1200–1350 °C. This innovative process involves repeated evacuation followed by purging of nitrogen gas so that the interconnected nanopores of the gel are filled with nitrogen gas prior to heat treatment. During heat treatment at higher temperatures, the presence of nitrogen gas in the nanopores of the gel starts the CTRN reaction simultaneously throughout the bulk of the gel, leading to the formation of Si3 N4 nanowires. The in situ generated ultrafine carbon obtained by the decomposition of dextrose decreases the partial pressure of oxygen in the system to stabilize the nanowires. The nanowires synthesized by this process are of ∼500 nm diameter and ∼0.2 mm length. The product was characterized by scanning electron microscope (SEM), energy dispersive x-ray analysis (EDX), x-ray diffraction (XRD) and infrared (IR) spectra.
Keywords: sol-gel, nanowires, α-Si3N4
Introduction
Si3N4 has attracted considerable interest because of its excellent high-temperature mechanical properties, such as strength, hardness, wear resistance, thermal shock resistance and chemical inertness. This material is utilized for structural applications at high temperatures and its domains of application extend from vessels for chemical reactions to heat-exchanger bearings, engines and gas turbine components [1–3]. Si3 N4 is also a wide-band-gap semiconductor used in the optical and electrical device industries [4]. However, Si3 N4 nanotubes and nanowires have recently attracted much attention because of their enhanced hardness associated with their unique optical and other mechanical properties [5–7] as compared with coarse-grained Si3 N4 ceramic materials. As a result, different physical and chemical routes have been proposed and developed for the synthesis of Si3 N4 nanowires. Chemical vapor deposition [8, 9], the carbothermal reduction of silica and carbon in nitrogen-containing ambient [10, 11], the nitridation of silicon powder [12, 13] and the nitridation of a Fe–Si catalyst [14] are frequently used for nanowire synthesis. Wang and Wada synthesized Si3 N4 nanowires by the carbothermal reduction of silica followed by 10 h nitridation at 1350 °C in ultrapure nitrogen atmosphere [15]. The same route has been used by Ramesh and Rao [16] to synthesize α-Si3 N4. Si3 N4 forms only after the carbothermal reduction of preoxidized SiO1.7 for 10 h followed by nitridation in very high purity nitrogen (partial pressure of oxygen: 10−20 atmosphere) at 1350 °C but long fibers of α-Si3 N4 are only formed after 24 h of the above treatment [16]. The carbothermal reduction of silica xerogels followed by 30 h of nitridation at 1300 °C in ultrahigh-purity nitrogen atmosphere results in the generation of long and pure α-Si3 N4 nanowires [17]. All these synthesis methods of Si3 N4 nanowires are kinetically slow and require very lengthy heat treatment. In all the above processes, the nitrogen gas has to first diffuse to the reaction sites and only then does the CTRN reaction occur for the synthesis of Si3 N4 nanowires. It is therefore expected that if nitrogen gas was present beforehand at the reaction sites, the carbothermal reduction followed by the nitridation (CTRN) reaction kinetics would be much faster during heat treatment. In this paper, we report the effect of the presence of nitrogen gas in the nanopores of the gel, obtained by repeated evacuation followed by purging of nitrogen gas to enhance the kinetics of the CTRN reaction for the synthesis of Si3 N4 nanowires.
Thermodynamics of Si–N–O system
The thermodynamics of the carbon-saturated Si–N–O system can be used to determine the permissible oxygen content in nitrogen that ensures the stability of the Si3 N4 phase at a given temperature. This is important because the nitrogen used is not completely free from oxygen. Figure 1 shows a stability diagram of the Si–N–O system at 1350 °C. The equilibrium isobar of pSiO= 10−4 superimposed on the stability diagram will help us to determine the mechanism of the synthesis of Si3 N4 nanowires, as explained later.
Figure 1.

Stability diagram of Si–N–O system at 1623K.
From the stability diagram, we can observe that Si3 N4 exists as a stable phase at a maximum permissible partial pressure of oxygen of 10−17.7 atmosphere at 1623 K corresponding to a pressure of 1 atmosphere of nitrogen. Similar analysis also for 1300 and 1200 °C show that the maximum permissible partial pressure of oxygen ensuring the stability of Si3 N4 gradually decreases to nearly 10−19 and 10−20 atmospheres, respectively. Excess carbon, over the theoretical amount required for the synthesis of Si3 N4, is added to the system and is responsible for the generation of the extremely low partial pressure of oxygen required for the stability of Si3 N4.
Experimental
Tetraethyl orthosilicate (TEOS) and dextrose (AR grade) are used as precursors of Si and C, respectively. A clear solution of TEOS and alcohol (99.9%) is obtained by constant stirring. A clear solution of dextrose, double-distilled water and ethyl alcohol was added dropwise to the previous solution under constant stirring. The resultant solution was left in a beaker at room temperature to form a gel. After about seven days a transparent gel was obtained. The molar composition of the gel thus prepared is TEOS : water : ethyl alcohol =1 : 1 : 2, containing 200% excess of stoichiometric dextrose.
Dried gels were initially preheated at 300 °C for 30 min in an inert atmosphere to completely remove moisture and to generate extremely fine C in the nanopores of the gel by the decomposition of the dextrose. Each preheated sample was introduced into a vacuum furnace evacuated using a double-stage rotary pump, then purged with nitrogen gas at ambient temperature. Repeated evacuation and purging was carried out to ensure the maximum replacement of air with nitrogen. The temperature of the furnace was raised to 1200–1350 °C and was held isothermally for a period of 2 or 3 h under a constant flow of nitrogen. For comparison, another sample was heat-treated at 11350 °C for 3 h in a furnace under a constant flow of nitrogen gas but without prior evacuation followed by purging with nitrogen gas.
The heat-treated samples are ground in an agate mortar and passed through a 300-mesh sieve. X-ray diffraction (XRD) patterns of the powder samples were recorded using an x-ray diffractometer (Rigaku Miniflex) with CuKα radiation (λ= 0.15418 nm) in a continuous scan mode from 20° to 80° at a scan rate of 0.1 s.
A powder sample was placed on a conducting carbon tape, and a scanning electron microscope (SEM) (JEOL, JSM 6063) was used to observe the morphological details of phases present. The elemental analysis of the phases was carried out by energy dispersive x-ray analysis (EDX) (Oxford Instruments, INCA).
Infrared (IR) spectra were obtained using a PerkinElmer 833-IR spectrometer. A small amount of ground sample was mixed with KBr and a pellet was formed. The spectrum was obtained over 4000 –450 cm−1 at room temperature. The spectrum of KBr powder was used as the background.
Results and discussions
Figures 2(a)–(c) show SEM images and an EDX plot, respectively, of the samples heat treated in dynamic nitrogen atmosphere at 1350 °C for 3 h after repeated evacuation followed by purging of nitrogen gas. Figure 2(a) shows that Si3N4 nanowires were formed with a diameter of ∼500 nm (later confirmed by XRD). Figure 2(b) shows an SEM image of the same sample at lower magnification and shows that the wires are sufficiently long. No accurate measurement of the length of the wires could be carried out, but the wires formed after 3 h of carbothermal reduction followed by nitridation were always more than ∼0.2 mm long. The EDX plot of the same sample taken from the white wires is shown in figure 2(c) and it shows the presence of Si, N and O. The adjacent table contains the atomic percentages of the elements, and the atomic ratio of Si : N is ∼ 3 : 4. This suggests that the nanowires are Si3N4.
Figure 2.

(a) SEM image of sample with 200% excess dextrose gel heat treated at 1350 °C for 3 h under a constant flow of nitrogen with prior evacuation followed by purging with nitrogen gas. (b) SEM image of sample shown in (a) at lower magnification. (c) EDX plot of sample shown in (a).
White nanowires of α-Si3N4 were obtained only after 3 h of heat treatment using ordinary nitrogen gas, in contrast with the 24 or 30 h of heat treatment using ultrahigh-purity nitrogen gas required to form similar nanowires by other methods [16, 17]. This experiment clearly indicates the beneficial role of the presence of nitrogen in the nanopores of the gel.
Table 1 contains EDX data for the sample heat-treated at 1350 °C for 2 h under constant flow of nitrogen with prior evacuation of the furnace followed by purging with nitrogen gas and the EDX data of the sample heat-treated at 1350 °C for 3 h under a constant flow of nitrogen gas but without prior evacuation followed by purging with nitrogen gas. Compared with figure 2(c) the table shows that less nitrogen is present in the sample reduced at 1350 °C for 2 h under constant flow of nitrogen with prior evacuation and purging with nitrogen, and this indicates that a lower percentage of the sample was converted to Si3N4. On the other hand, when reduction is carried out at 1350 °C in nitrogen atmosphere for 3 h without prior evacuation of the furnace and followed by purging with nitrogen is completely absent and thus it is concluded that hardly any Si3N4 is formed.
Table 1.
EDX data for the samples heat-treated as described.
| Sample | 200% excess dextrose gel heat-treated at 1350 °C for | 200% excess dextrose gel heat-treated at 1350 °C for |
|---|---|---|
| 2 h under constant flow of nitrogen with prior | 3 h under constant flow of nitrogen but without | |
| evacuation of the furnace followed by purging with | prior evacuation of the furnace followed by | |
| nitrogen gas | purging with nitrogen gas | |
| Si (wt. %) | 50.49 | 37.84 |
| O (wt. %) | 34.22 | 62.16 |
| N (wt. %) | 15.29 | nil |
| Total | 100 | 100 |
Figures 3(a)–(c) show XRD patterns of the samples obtained after heat treatment at 1200, 1300 and 1350 °C, respectively, in nitrogen atmosphere for 3 h with prior repeated evacuation followed by purging with nitrogen gas. Figures 3(a) and (b) confirm the presence of a broad amorphous silica peak between 15° and 30° [18] and diffraction peaks corresponding to α- and β-Si3N4 phases. It is further observed that with increasing heat treatment temperature, the amount of α-Si3N4 formed increases and that of β-Si3N4 decreases. Figure 3(c) shows only diffraction peaks corresponding to the α-Si3N4 phase with a conspicuous absence of the amorphous silica peak. It is clear from the above discussion that the conversion from SiO2 to α-Si3N4 increases with temperature. An almost complete conversion to α-Si3N4 is obtained at a temperature of 1350 °C within 3 h. Figure 3(d) depicts the XRD pattern of the sample heat treated at 1350 °C for 2 h under constant flow of nitrogen with prior evacuation of the furnace followed by purging with nitrogen gas. It consists of diffraction peaks corresponding to α- and β-Si3N4 and a small amorphous silica peak. This means that the conversion to α-Si3N4 is not completed within 2 h.
Figure 3.

(a) XRD pattern of sample with 200% excess dextrose gel heat treated at 1200 °C for 3 h under a constant flow of nitrogen with prior evacuation of the furnace followed by purging with nitrogen gas. (b) XRD pattern of sample with 200% excess dextrose gel heat-treated at 1300 °C for 3 h under a constant flow of nitrogen with prior evacuation of the furnace followed by purging with nitrogen gas. (c) XRD pattern of sample with 200% excess dextrose gel heat treated at 1350 °C for 3 h under a constant flow of nitrogen with prior evacuation of the furnace followed by purging with nitrogen gas. (d) XRD pattern of sample with 200% excess dextrose gel heat treated at 1350 °C for 2 h under a constant flow of nitrogen with prior evacuation of the furnace followed by purging with nitrogen gas. (e) XRD pattern of sample with 200% excess dextrose gel heat treated at 1350 °C for 3 h under a constant flow of nitrogen but without prior evacuation of the furnace followed by purging with nitrogen gas.
Figure 3(e) shows the XRD pattern of the sample heat-treated at 1350 °C for 3 h in nitrogen atmosphere without prior evacuation and purging with nitrogen gas. There is a major amorphous silica peak and hardly any diffraction peaks for Si3N4. This is consistent with the EDX data (see table 1), which reflects the absence of nitrogen, resulting in the formation of hardly any Si3N4. This clearly establishes the importance of the presence of nitrogen in the nanopores of the gel prior to the heat treatment.
Observations based on SEM-EDX and XRD analyses are further confirmed by FTIR spectroscopic analysis. Figure 4(a) shows the IR spectrum of the sample with 200% excess dextrose gel heat-treated at 1350 °C for 3 h under a constant flow of nitrogen with prior evacuation of the furnace followed by purging with nitrogen gas. The IR absorption bands at 847, 682, 597 and 492 cm−1 correspond to the characteristics absorption bands of α-Si3N4 [19, 20], and this confirms the presence of α-Si3N4. Figure 4(b) shows the IR spectrum of the sample heat-treated at 1350 °C for 3 h under a constant flow of nitrogen gas without prior evacuation of the furnace followed by purging with nitrogen gas. This shows the presence of the characteristic absorption band at 1100 cm−1 corresponding to the Si–O bond in silicon oxide [21]. However, the absorption bands of α-Si3N4 are absent indicating the absence of Si3N4.
Figure 4.

(a) IR spectrum of sample with 200% excess dextrose gel heat-treated at 1350 °C for 3 h under a constant flow of nitrogen with prior evacuation of the furnace followed by purging with nitrogen gas. (b) IR spectrum of the sample with 200% excess dextrose gel heat treated at 1350 °C for 3 h under a constant flow of nitrogen but without evacuation of the furnace followed by purging with nitrogen gas.
Role of evacuation and proposed mechanism of synthesis
The results of different characterization techniques reveal that repeated evacuation followed by purging with nitrogen plays an important role in the synthesis of Si3N4 nanowires. Hardly any nanowires were formed without prior evacuation and purging with nitrogen gas. This leads us to propose the following mechanism for the synthesis of silicon nitride nanowires.
Dry silica gel contains interconnected nanopores filled with air and dextrose. Carbon is generated in situ by the decomposition of dextrose in the nanopores of the gel at 300 °C. When the gel is heat treated directly at a high temperature in a dynamic nitrogen atmosphere, the nitrogen gas has to first reach the nanopores by a slow diffusion process before nitridation can start. Thus, the process is kinetically slow and is the most plausible reason for the absence of Si3N4 (see figure 3(e)). On the other hand, if nitrogen gas was present a priori throughout the nanopores of the bulk gel, the reaction is much faster. This is exactly what was achieved by repeated evacuation followed by purging with nitrogen gas. In each evacuation–purging cycle, the nanopores of the gel are evacuated and subsequently filled with nitrogen gas. Repeated cycles ensure better filling of the nanopores with nitrogen. Then the temperature of the furnace is raised to the heat treatment temperature in a dynamic nitrogen atmosphere, and unlike the first case, the diffusion of nitrogen to the nanopores is no longer necessary; thus, the carbothermal reduction followed by nitridation starts simultaneously throughout the entire porous network of the gel, leading to the much faster synthesis of Si3N4. The following is the probable reaction mechanism. First the in situ generated carbon uniformly distributed in the nanopores reacts with silica gel on the pore surface to form SiO (g) and CO (g) as given by the reaction (1) below [16]. SiO maintains an equilibrium partial pressure of 10−4 atmosphere in 1 atmosphere of nitrogen at 1623 K (see figure 1). This gaseous SiO produced in the nanopores of the gel mixes homogeneously with the nitrogen and CO. This gas mixture undergoes a reaction to yield Si3N4 by reaction (2) as stated below. The other possibility is that Si3N4 is yielded by reaction (3) [16]. The solid Si3 N4 is deposited on the carbon particles, which act as nucleating sites. As Si3N4 is formed and deposited, the partial pressure of SiO decreases from its equilibrium value, which is restored by the reduction of SiO2 and drives equilibrium in the forward direction as in reaction (1). Since the reaction occurs within the porous network of the gel, the Si3N4 thus formed is expected to be in the form of nanowires. Thus, the most likely reactions, under the experimental conditions, for the synthesis of silicon nitride nanowires are given below:
The other possible gas-phase reaction is
Conclusions
Si3 N4 nanowires of diameter ∼500 nm and length ∼0.2 mm have been synthesized by the carbothermal reduction of silica followed by nitridation through the sol-gel route.
Repeated evacuation followed by purging with N2 fills the nanopores of the silica gel with N2 prior to heat treatment, and this greatly enhances the kinetics of the conversion from SiO2 to α-Si3N4.
The thermodynamic study of Si–N–O system shows that extremely low partial pressure of oxygen is required for the stability of Si3N4 phase, which is achieved by the addition of excess dextrose.
Acknowledgments
We gratefully acknowledge the Council of Scientific and Industrial Research, New Delhi, for sponsoring this research. The work was also partially supported by UGC Potential for Excellence, Nanoscience and Technology, Jadavpur University. We also acknowledge the instrumental assistance of Dr R Bhar and Mr P Das Gupta of the Department of Instrumentation Science, Jadavpur University.
References
- Backhaus-Ricoult M, Guerin V, Huntz A-M. and Urbanovich V S. J. Am. Ceram. Soc. 2002;85:385. [Google Scholar]
- Riley F L. J. Am. Ceram. Soc. 2000;83:245. [Google Scholar]
- Buljan S T. and Baldoni J G. Mater. Sci. Forum. 1989;47:249. [Google Scholar]
- Oba F, Tatsumi K, Tanaka I. and Adachi H. J. Am. Ceram. Soc. 2002;85:97. [Google Scholar]
- Carduner K R, Blackwell C S, Hammond W B, Reidinger F. and Hatfield G R. J. Am. Chem. Soc. 1990;112:4676. doi: 10.1021/ja00168a008. [DOI] [Google Scholar]
- Xu Y J, Cao C B, Chen Z, Li J, Wang F C and Cai H N. 2006. J. Phys. Chem. B 110 3088 10.1021/jp0560258 [DOI] [PubMed] [Google Scholar]
- Tang C C, Bando Y, Sato T. and Kurashima K. J. Mater. Chem. 2002;12:1910. doi: 10.1039/b200861k. [DOI] [Google Scholar]
- Kijima K, Setaka N. and Tanaka H. J. Cryst. Growth. 1974;24–25:183. doi: 10.1016/0022-0248(74)90301-7. [DOI] [Google Scholar]
- Motojima S, Yamana T, Araki T. and Iwanage H. J. Electrochem. Soc. 1995;142:3141. doi: 10.1149/1.2048702. [DOI] [Google Scholar]
- Arik H. J. Eur. Ceram. Soc. 2003;23:2005. doi: 10.1016/S0955-2219(03)00038-4. [DOI] [Google Scholar]
- Wu X C, Song W H, Zhao B, Huang W D, Pu M H, Sun Y P. and Du J. J. Solid State Commun. 2000;115:683. doi: 10.1016/S0038-1098(00)00255-6. [DOI] [Google Scholar]
- Gopalakrishnan P S. and Laksminarsimhan P S. J. Mater. Sci. Lett. 1993;12:1422. doi: 10.1007/BF00591595. [DOI] [Google Scholar]
- Inomata Y. and Yamane T. J. Cryst. Growth. 1974;21:317. doi: 10.1016/0022-0248(74)90024-4. [DOI] [Google Scholar]
- Huo K F, Ma Y W, Hu Y M, Fu J J, Lu B, Lu Y N, Hu Z. and Chen Y. Nanotechnology. 2005;16:2282. doi: 10.1088/0957-4484/16/10/050. [DOI] [PubMed] [Google Scholar]
- Wang M-J. and Wada H. J. Mater. Sci. 1990;25:1690. doi: 10.1007/BF01045372. [DOI] [Google Scholar]
- Ramesh P D. and Rao K J. J. Mater. Res. 1994;9:2330. doi: 10.1557/JMR.1994.2330. [DOI] [Google Scholar]
- Wang F, Jin G and Guo X-Y. 2006. J. Phys. Chem. B 110 14546 10.1021/jp0619282 [DOI] [PubMed] [Google Scholar]
- Liu Y Z, Jhou Y Q, Shi W Q, Jhao L L, Sun B Y. and Ye T C. Mater. Lett. 2004;58:2397. doi: 10.1016/j.matlet.2004.02.015. [DOI] [Google Scholar]
- Theodossiu E, Bauman H, Matz W and Mucklich A. 2002. Phys. Status Solidi a 194 47 [DOI] [Google Scholar]
- Mazdiyasni K S. and Cooke C M. J. Am. Ceram. Soc. 1973;56:628. doi: 10.1111/j.1151-2916.1973.tb12440.x. [DOI] [Google Scholar]
- Luo Y M, Zheng Z M, Xie Z M. and Zhang Z J. J. Mater. Lett. 2004;58:2114. doi: 10.1016/j.matlet.2003.12.032. [DOI] [Google Scholar]
