Ceramics International 20 (1994) 195-199
Preparation of Si3N 4 by Carbotherrnal Reduction of Digested Rice Husk I. A. R a h m a n School of Chemical Sciences, University of Sains Malaysia, 11800 Penang, Malaysia (Received 3 June 1993; accepted 12 September 1993) Abstract: Rice husk was digested by using various concentrations of HNO3 at 60°C and it was found that rice husk containing about 29% carbon and about 79% silica was produced by digestion with 12M HNO 3 for 3 h. The carbon and silica obtained are in stoichiometric ratio for the overall carbothermal reduction reaction. An equiaxed, submicrometre and high ~-phase silicon nitride powder was obtained by heating pyrolysed digested husk at about 1430°C under flowing nitrogen. The homogeneity of the carbon and silica is the main factor that allows completion of the reaction.
A high-purity, predominantly e-phase silicon nitride powder can be obtained by carbothermal reduction and nitridation of a silica-carbon m i x t u r e ? - 9 To increase the reaction rate, homogeneous mixing of the starting powder is essential. This can be achieved by using fine silica and carbon, which are normally prepared by the sol-gel technique. 4'9 In all cases, excess carbon is required to complete the reaction in a reasonable time. Rice husk is an agricultural waste containing 80-85% organic materials and 15-20% silica. Its composition varies depending on the variety, climate and geographical location. Pyrolysis (burning in an inert atmosphere) of the husk produces a material containing a uniform mixture of silica and carbon which is useful for the preparation of silicon nitride powder. 1°'~'~5 The husk contains excess carbon which can be removed after nitridation by heating in air at about 700°C. In the present work, the possibility of producing carbon and silica through digestion of rice husk using nitric acid was investigated. In addition, fine and pure husk powder can be prepared for use as a starting powder for silicon nitride production. Nitric acid was used because of its strong oxidising property, which allows it to disintegrate organic material in the husk.
About 30 g of prewashed rice husk was digested in 300ml of HNO3 at 60°C. The concentration of acid was varied from 10M to 14M. During digestion (for up to 7 h with continuous stirring), the temperature was carefully controlled, as the reaction between organic constituents of the husk and HNO3 is exothermic. The husk was then filtered and repeatedly washed with distilled water until neutral, as shown by a pH value of approximately 7 for the wash water, and dried. The digested husk was ground in a ball mill. Chemical analysis of the combusted (700°C, I h in air) digested husk was carried out by atomic absorption spectrophotometry (M250, Inst. Labs., USA). Organic constituents of raw and digested husk were analysed using a method adopted from TAPPI (Test Analysis of Pulp and Paper Industry). The digested husk was pyrolysed at 800°C to a constant weight under a flow of argon gas. The carbon content of pyrolysed digested husk (PDH) was determined by heating at 700°C in air for 1 h. The weight loss after heating was considered to be the amount of carbon present in the pyrolysed digested husk, and the residue was considered to be pure silica. Approximately 2.0 g of P D H powder in a graphite boat was placed in a molybdenum disilicide 195
Ceramics International 0272-8842/94/$7.00 © 1994 Elsevier Science Limited, England and Techna S.r.l. Printed in Great Britain
L A. Rahman
(Kanthal Super-33) controlled-atmosphere heating furnace. Before heating, the furnace was purged with nitrogen at a high flow rate for about 1 min to remove trapped air. The heating rate was about 5°Cmin -~ up to 1200°C, and increased to 10°C min- ~ to the reaction temperature of 1430°C. The flow rate of nitrogen gas was about 50 ml minfrom room temperature to 1200°C, and was increased to 7 5 0 m l m i n - : at the nitriding temperature. After nitridation, the products were burned in a muffle furnace at about 700°C for 30 min to remove any excess carbon. Nitrogen analyses of the nitrided materials were carried out by alkali fusion and acid-base titration. Samples (0.1g) were fused with 10g of sodium hydroxide (99.8%, AnalaR Grade) in a Pyrex tube. The NH 3 gas evolved was absorbed in boric acid solution (2%) and then titrated with standardised HC1 solution (0"IM). The nitrogen content was calculated on the basis of the following reactions: SiaN4(s ) + 6NaOH(s) + 3H20(g) --* 3NaESiO3(1) +4NHa(g ) (1) HaBO3(eq) + 3NHa(g ) --~ (NH4)aBOa(eq) (NH4)aBOa(eq) + 3HCl(eq) --~ 3NH4CI(eq)+ HaBOa(eq)
The amount of carbon was determined by using a CHN-Analyser (M240A, Control Equipment Corp., USA). The phase composition of the product was determined by X-ray diffraction (Philip PW 1820/1729). Scanning electron micrographs were obtained using a Stereoscan $200 microscope (Cambridge, UK). RESULTS A N D D I S C U S S I O N
Digestion of rice husk Table 1 shows the composition of rice husk before and after digestion with 12M HNO 3 for 3 h at 60°C. Hollocellulose is known to be constituted of hemicellulose and cellulose. ~4 The results show that 50% and 98% of hollocellulose and lignin, respecT a b l e 1. C o m p o s i t i o n of r a w rice h u s k b e f o r e and a f t e r d i g e s t i o n w i t h 1 2 M H N O a f o r 3 h at 60°C Constituent
Silica Holocellulose Lignin Extractives Loss on digestion
15.05 58.15 25.06 2.95 --
15.00 19-59 0.50 -53.83
tively, were digested. This leads to the lower carbon content of the digested husk. Silica was found to be unaffected by HNO3. A scanning electron micrograph of the outer epidermis of a typical rice husk is shown in Fig. I(A). The morphology is similar to that reported elsewhere, 12'1a and consists of an array of protuberances with 'spines' emerging between them. After digestion, grooves appear between the protuberances, and some of the husk has disintegrated into fibrous structure (Figs I(B) and 1(C)), as a result of the action of HNO 3 on the organic constituents. Sharma and William 12 and Krishnarao and Godkhindi ~3 suggested that organic constituents are mainly present between the protuberances. The results shown in Figs I(B), 1(C) and I(D) are consistent with their findings. Nitric acid is known as a strong oxidising agent which is able to degrade the holocellulose and lignin and turn the husk into a disintegrated form. The extent of degradation depends on acid concentration and conditions. 14 Weight loss in the range of 50-60% was observed after digestion. The colour of the husk changed from golden yellow to light yellow, and the size of the husk decreased as the concentration of the acid increased. After pyrolysis, the digested husk turned black and was very fragile. Pyrolysis of digested rice husk in an inert atmosphere results in decomposition of holocellulose (cellulose and hemicellulose) into organic volatiles, carbon monoxide, carbon dioxide, water and char. The black char consists of intimately mixed carbonaceous material (later turning to carbon at temperatures above 700°C) and silica. Figure 2 illustrates the effect of temperature of pyrolysis of digested husk under argon atmosphere for 2h. The weight loss in the region 400-800°C corresponds to the decomposition of holocellulose. Lignin does not contribute to weight loss, as it is present at only about 0.5%. Decomposition of holocellulose appears to be completed at temperatures above 800°C, as indicated by the constant weight loss of about 60%. Table 2 shows C/SiO 2 ratios after pyrolysis at 800°C for 2h after digestion with various concentrations of HNO 3. It is interesting to note that after more than 3 h digestion in 12M HNO 3 the product obtained is in the stoichiometric ratio of 2C/SiO 2. Higher concentration could not produce the appropriate C/SiO 2 ratio, as the reaction is vigorous and exothermic, making it difficult to maintain the temperature. Lower concentrations required a longer time to produce a suitable carbon-silica mixture. The chemical composition of rice husk ash after digestion with 12M HNO a at 60°C for 3 h and combustion of carbon is shown in Table 3.
Si3N4 prepared from digested rice husk
Fig. 1. Scanning electron micrographs of rice husk. (A) Epidermis layer; (B), (C) and (D) digested husk.
T a b l e 2. E f f e c t o f C / S i O 2 r a t i o o n d i g e s t i o n o f r i c e husk w i t h v a r i o u s c o n c e n t r a t i o n s o f H N O 3 a t 60°C Time
'~ 501 o a~
Temperature ( 0£ )
Fig. 2. Effect of temperature on pyrolysis of digested rice husk.
1 2 3 4 5 6
C/SiO 2 ratio of pyrolysed digested husk at 800°C under argon atmosphere IOM
3"07 2'46 2"43 2.32 2"23 --
2"24 2" 14 2'00 2"00 2'00
1 '89 1 "68 1 "98 1 "73 1 '76
I . A . Rahman
Table 3. Chemical composition of digested rice husk ash Constituent
SiO= AI=O3 CaO Fe203 MgO Na20 K=O
99-75(9) a 0.01 (5) 0.01 (9) 0.02(1 ) 0.01 (8) 0-07(8) 0.09(1 )
• Theoreticaltoss= ~.~*/~
a Estimated from the total of metallic oxides.
Fig. 3. Weight loss after heating PDH at 1430°C under nitrogen atmosphere.
Nitridation of digested husk Nitridation of rice husk is, in principle, identical to the carbothermal reduction and nitridation of a carbon-silica mixture. The chemical reaction can be written as 3SiO2(s ) + 6C(s) + 2N2(g) ~ Si3N4(s ) + 6CO(g)
Although the overall reaction is presented in solidstate form, the actual reaction mechanism is generally agreed to involve SiO gas. 1- ~1 In this case, a high surface area and homogeneous mixing of reactants are normally required. The theoretical weight loss based on reaction (4) is 44.4%. The excess carbon after nitriding at 1430°C, and the composition of the powder are shown in Table 4. The changes in weight loss after nitriding P D H at 1430°C for various durations up to 6 h are shown in Fig. 3. The result shows that the reaction is completed after 6 h. Weight loss is higher than the theoretical value, indicating that some SiO was released unreacted. The theoretical yield of Si3N 4 calculated based on reaction (4) is 55-6% (assuming that C and SiO2 are in stoichiometric ratio). In the present work, the quantity of Si3N 4 produced after nitriding for 6 h is 40.5%; therefore, the Si3N 4 yield was found to be about 73%. Preparation of Si3N 4 powder by carbothermal Table 4. Excess carbon a f t e r nitridation of pyrolysed digested husk ( P D H ) at 1430°C, and chemical c o m p o s i t i o n of the Si3N 4 p o w d e r Nitridation time (h)
Carbon excess after nitridation
Composition in powder (%) N
22.0 29-47 34.59 37.14
1 '50 1.11 0.23 0.09
1 2 4 6
3.45 2.50 1 '60 0.00
a Estimated from composition of the ash (Table 3), assuming that Na20 and K20 evaporated at 1430°C.
reduction and nitridation of pure carbon and silica mixtures was previously reported by several researchers. 1-3'6-9 Komeya and Inoue 1 prepared ~-Si3N 4 powder from silica at C/SiO2 = 2-20, at 1400-1450°C. They concluded that the reaction rate significantly increased with increase in carbon content and surface area of the raw powders. Mori et al. 3 reported that a complete conversion of very fine silica to ~-Si3N 4 was achieved at C/SiO 2 = 5 and 15 after 10 h at 1500°C. Zhang and Cannon 4 found that the reaction rate increased with increase in surface area of silica and carbon. They concluded that homogeneous mixing and the addition of excess carbon contributed to the completeness of the carbothermal reduction and nitridation of silica at 1400oc. Others 6-9 confirmed that the addition of excess carbon and a high surface area of the reactant are important in attaining uniform contact between silica and carbon particles. In earlier works, l°'x~ ~-Si3N 4 was produced from rice husk of C/SiO2 = 7 at 1425°C. In this work, Si3N 4 powder ( N = 3 7 % , C = 0-09%) was prepared at 1430°C from rice husk containing carbon and silica in stoichiometric ratio. It is shown that the most important factor contributing to the completeness of reaction is the homogeneity of mixing. Thus, the use of rice husk is an advantage, as silica and carbon are naturally mixed. Silicon carbide was not detected in the present work. This result is consistent with that of a previous report, 11 which showed that the maximum temperature for Si3N 4 formation from rice husk is about 1425°C. Scanning electron micrographs of Si3N 4 powder obtained after nitriding P D H at 1430°C are shown in Fig. 4. The powders consist of equiaxed and short whiskers (Fig. 4(A)). The fact that a high packing densil~y ~-Si3N 4 powder with equiaxed grains would be preferable as a starting material for the production of sintered nitride prompted further investigation. P D H was ball milled for 6 h, after which the product consisted mainly of equiaxed Si3N 4 grains
Si3N 4 prepared f r o m digested rice husk
CONCLUSION Submicrometre Si3N 4 powder was prepared by the carbothermal reduction of digested rice husk. The carbon and silica content of the husk was controlled by digestion with hot HNO 3 in appropriate concentration and time to match the stoichiometric ratio of the reaction. The completeness of the reaction depends on the homogeneous mixing of carbon and silica, which occurs naturally in rice husk.
ACKN OWLEDG EM ENTS I thank the Universiti Sains Malaysia for financial support in terms of an R & D grant (123/3205/2501), and the XRD and SEM section (USM, Perak Branch), and M.S.A. Ghani, for their help.
Fig. 4. Scanning electron micrographs of Si3N4 powder obtained by nitridation of PDH. (A) Unmilled; (B) milled for 6 h.
(Fig. 4(B)). This result is contrary to those reported by Kom.eya and Inoue I and Rahman and Riley, 1° who showed that a preformed Si3N 4 must be added to the system SiO2-C-Si3N 4 to promote the formation of equiaxed Si3N 4 grains. Figure 5 is a typical X-ray diffraction pattern of the Si3N 4 powder synthesised from PDH. The pattern indicates that the powder contains mainly ~-Si3N 4. xlO 3 3-0
Fig. 5. XRD spectrum ofSi3N 4 powder obtained by nitridation
of PDH at 1430°C for 6 h under nitrogen atmosphere.
1. KOMEYA, K. & INOUE, H., Synthesis of the a-form silicon nitride from silica H. J. Mater. Sci. Lett., 10 (1975) 1243-6. 2. INOUE, H., KOMEYA, K. & TSUGE, A., Synthesis of silicon nitride powder from silica reduction. Commun. Am. Ceram. Sot., C (1982) C205. 3. MORI, M., INOUE, H. & OCHIAI, T., Preparation of silicon nitride powder from silica. In Progress in Nitrogen Ceramic. ed. F. L. Riley, Martinus Nijhoff, The Hague, 1983, p. 149-55. 4. ZHANG, S. C. & CANNON, W. R., Preparation of silicon nitride from silica. J. Am. Ceram. Soc, 67(10) (1985) 691-5. 5. SIDDIQI, S. A. & HENDRY, A., Influence of iron on the preparation of silicon nitride from silica. J. Mater. Sei., 20 (1985) 3230-8. 6. PERERA, D. S., Conversion of precipitated silica from geothermal water to silicon nitride. J. Mater. Sci., 22 (1987) 2411-15. 7. BANDYOPADHYAY, S. & MUKERJI, J., Reaction sequences in the synthesis of silicon nitride from quartz. Ceram. Int., 17 (1991) 171-9. 8. E K E L U N D , M. & FORSLUND, B., Carbothermal preparation of silicon nitride: influence of starting material and synthesis parameters. J. Am. Ceram. Soc., 75(3) (1992) 532-9. 9. POPPER, P. (ed.), In Special Ceramics, Vol. 7. Br. Ceram. Res. Assoc., 1981, pp. 107-18. 10. RAHMAN, I. A. & RILEY, F. L., The control of morphology in silicon nitride powder prepared from rice husk. J. Eur. Ceram. Soc., 5 (1989) 11-22. 11. HANNA, S. B., MANSOUR, N. A. L., TAHA, A. S. & ABD ALLAH, H. M. S., Silicon carbide and nitride from rice husks III. Formation of silicon nitride. Br. Ceram. Trans. J., 84(1) (1985) 18-21. 12. SHARMA, N. K. & WILLIAMS, W. S., Formation and structure of silicon carbide whiskers from rice husk. J. Am. Ceram. Soc., 67(11) (1984) 715-20. 13. KRISHNARAO, R. V. & G O D K H I N D I , N. M., Distribution of silica in rice husks and its effect on the formation of silicon carbide. Ceram. Int., 18 (1992) 243-9. 14. ROWELL, R. (ed.), The Chemistry of Solid Wood. Am. Chem. Soc., Washington, DC, 1984, pp. 69, 575-86. 15. K U M A R DAS, V. G., WENG, N. S. & GIELEN, M. (ed.), Chemistry and Technology of Silicon and Tin. Oxford University Press, New York, 1992, pp. 377-83.