- Email: [email protected]
Effect of starting Si contents on the properties and structure of biomorphic SiC ceramics
Journal of Materials Processing Technology 182 (2007) 34–38
Effect of starting Si contents on the properties and structure of biomorphic SiC ceramics Guangya Hou ∗ , Zhihao Jin, Junmin Qian State Key Laboratory for Mechanical Behavior of Materials, Xi’an Jiaotong University, Xi’an 710049, PR China Received 22 March 2006; received in revised form 12 June 2006; accepted 8 July 2006
Abstract Biomorphic SiC ceramics were prepared by infiltrating liquid Si into biocarbon templates derived from beech at 1550 ◦ C for 1.5 h with different starting Si/C ratios. Microstructure observation and phase identification of the biocarbon templates and the resulting products were examined by scanning electron microscopy (SEM) and X-ray diffraction (XRD), respectively. Open porosity and density of the resulting products were measured by the Archimedes method. The bending strength and fracture toughness were tested by the three-point bending method and the single edge notched beam (SENB) method, respectively. The conversion degrees from biocarbon templates to biomorphic SiC ceramics were also discussed. Experimental results showed that, with increasing starting Si contents, the resulting products were biomorphic C/SiC, SiC and SiC/Si, the porosity decreased, and bending strength and fracture toughness increased. The key factors affecting properties of the resulting products were conversion degree and the amount of the residual silicon. The bending strength and toughness fracture of the axial samples were much higher than those of the radial samples because of loading direction and anisotropic pore orientation derived from microstructures of original wood. © 2006 Elsevier B.V. All rights reserved. Keywords: Biomorphic ceramics; Properties; Structure; SiC
1. Introduction In recent years, there has been an increasing interest in using biomimetic-based processing approaches to fabricate a variety of oxide and non-oxide based structural and functional materials [1]. Biomorphic ceramics or composites such as SiC, Al2 O3 , TiC, TiO2 and Si–Mo–C have been fabricated from wood, wood wastes or papers via using the microstructural features of naturally grown materials [2–9]. Biomorphic SiC ceramics are firstly and extensively studied because of easy preparation processing and good properties. It is interesting of technical applications such as heat insulation structures, substrates, filter and catalyst carriers at high temperatures, thermally and mechanically loaded lightweight structures, as well as for medical implant structures (e.g., porous materials for bone-substitution) [10]. Previous methods converting wood into various SiC ceramic materials are the infiltration of the pyrolysed biocarbon template with gaseous or liquid sili-
∗
Corresponding author. Tel.: +86 29 2667942; fax: +86 29 2665443. E-mail address: [email protected] (G. Hou).
0924-0136/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2006.07.003
con bearing precursors such as silicon melt, silicon, and silicon monoxide vapors, and organosilicon compounds at high temperatures, sol–gel and carbothermal reduction processing and chemical vapor infiltration and reaction (CVI-R) processing [7,11–14]. However, most of works on infiltrating liquid Si into biocarbon templates to manufacture biomorphic SiC or SiC/Si materials were focused on fabrication and properties of the final products. As far as we know, the relationship of starting Si contents and properties of biomorphic SiC ceramics has not been reported. In the present work, it focus on investigating the effects of starting Si contents on the properties and structures of biomorphic SiC ceramics. Biomorphic SiC ceramics were prepared by infiltrated different contents liquid Si into the biocarbon templates derived from beech. 2. Experimental procedure 2.1. Material preparation The wood pieces with 10 mm × 10 mm × 50 mm were cut from beech and dried at 105 ◦ C for more than 24 h, which were divided to axial samples and radial samples according to the direction of length perpendicular and parallel to the direction of wood growth, as shown in Fig. 1a and b. The samples were
G. Hou et al. / Journal of Materials Processing Technology 182 (2007) 34–38
35
Fig. 1. Sketch of bending strength test (a and b) and fracture toughness test (c and d): axial samples (a and c) and radial samples (b and d).
impregnated with phenolic resin in order to reinforce the cell wall and to avoid distortion or cracking of samples during pyrolyzation process. The effects of phenolic resin on the basic properties of woodceramics and biomorphic SiC ceramics were discussed in another our paper. Industrial silicon powder with a diameter range of 1–3 mm was used as a silicon source. All wood samples were pyrolyzed at 800 ◦ C for 4 h with flowing N2 protection to prepare the biocarbon templates, and then, packed in different contents silicon powder and heated at 1550 ◦ C for 1.5 h to form biomorphic SiC in N2 atmosphere. Finally, after the over of holding time, the samples were continually heated up to 1700 ◦ C for 40 min in vacuum to get rid of remnant silicon. The silicon at the temperature was transformed into gaseous Si and expelled from the samples. The crystal structures of the biocarbon, silicon and -SiC at room temperature are amorphous, cubic and face-centered cubic type, respectively, and the corresponding densities are 1.78, 2.32 and 3.21 g/cm−3 , respectively. The weight ratio of the starting silicon powder to the biocarbon templates was defined as ϕ (ϕ = WSi /Wbio-C ).
2.2. Characterization The XRD patterns of the biocarbon templates and the resulting products were recorded using X-ray diffractometer (D/MAX-RA, Rigaku, Japan) with nickel filtered Cu K␣ radiation produced at 35 kV and 20 mA. The tested samples were broken and ground into powders in a carnelian mortar. The microstructure morphology was observed with scanning electron microscope (SEM) (S-2700, Hitachi, Japan) operated at 20 kV and 20 mA. Open porosity and density of resulting products were determined by the Archimedes method with distilled water as the liquid medium. Strength was tested by the three-point bending method with a 20 mm span. Fracture toughness was tested with the single edge notched beam (SENB) method (Fig. 1). A notch with 2 mm depth and 0.28 mm width was cut in the middle of each SENB specimen. All the mechanical properties testing were carried out by a SANS-5104A machine (SANS Ltd., Tianshui, China) and crosshead speed was 0.5 mm/min. The degree of conversion from biocarbon to biomorphic SiC ceramics according to the following relationship: conversion degree (%) =
Wfinal − WC × 100 WC × SiAM /CAM
Fig. 2. XRD pattern of biocarbon (a) and biomorphic SiC ceramics prepared at different starting Si/C ratios (ϕ): 1 (b), 2 (c), 3 (d), 4 (e) and 5 (f).
of 15–50◦ . It can be seen that two broad peaks in Figs. 2a and 3a suggest the biocarbon is amorphous. With increasing starting silicon powder content or ϕ values, the intensity of amorphous carbon decreased or even disappeared. The resulting product may be biomorphic C/SiC ceramics when ϕ values were less than 3 (Figs. 2c and 3c). Pure SiC phase may obtained when ϕ value was 3. However, the diffraction intensity of SiC is significant higher than those of amorphous carbon, the carbon was not completely disappeared under the condition. When ϕ values were 4 and 5, there exist the diffraction peaks corresponding to -SiC and the peaks corresponding to Si in the XRD pattern of the resulting products, which indicated that the resultant were a diphase composites consisting of major phase SiC and secondary phase Si. The intensity of peaks of Si phase increased significantly when ϕ value was 5. It indicates that the residual silicon was increased with the contents of starting silicon powder increasing. Although all samples were gone through removing residual silicon process, there were not pure SiC phase when ϕ values was higher. According to the reaction equation of C and Si, pure SiC can be obtained when the weight ratio of starting silicon powder to biocarbon (ϕ) is about 2.34. From the XRD
(1)
where Wfinal is the weight of the resulting products after the reaction and WC the weight of the biocarbon template, SiAM and CAM are the atomic mass of silicon and carbon, respectively.
3. Results and discussion 3.1. XRD analysis Fig. 2 shows the effect of starting Si contents on the XRD patterns of biocarbon templates and biomorphic SiC ceramics. Fig. 3 is enlarged patterns of Fig. 2 at diffraction angle (2θ) range
Fig. 3. Enlarged pattern of Fig. 2 at diffraction angle (2θ) range of 15–50.
36
G. Hou et al. / Journal of Materials Processing Technology 182 (2007) 34–38
Fig. 4. The degree of conversion from biocabron to biomorphic SiC at different starting Si/C ratios (ϕ).
Fig. 5. porosity and density of axial and radial samples prepared with different starting Si/C ratios. ϕ = 0 means biocarbon templates.
pattern (Fig. 2d), the resulting product may consist of pure SiC phase when ϕ value was in the range of 3–4.
3.3. Open porosity and density of resulting products
3.2. Conversion degree from biocarbon to resulting products The degree of conversion from biocarbon to biomorphic SiC is shown in Fig. 4. It can be seen that the conversion degree of both axial samples and radial samples increased with increasing starting Si/C ratio. When ϕ was 3, the conversion degree of the axial samples and the radial samples were 68.42% and 72.72%, respectively. It showed the reaction did not finish and there were residual carbon in the resulting products. The Xray diffraction pattern shows only -SiC peaks, but considering the low degree of conversion, amorphous carbon is probably present in this sample [15]. The conversion degree exceeded 100% when ϕ value was 5. From XRD patterns of the resulting products (Fig. 2), there were the present of residual silicon in the resulting products, which was called as biomorphic SiC/Si material. Therefore, in this case the conversion degree exceeded 100% was reasonable. In fact, the conversion degree values may result overestimated because the presence of residual silicon in the resulting products was neglected in Eq. (1). There was the fastest conversion stage while starting Si/C ratio (ϕ) increased to 2. This was starting stage of reaction. Spontaneous wetting and infiltration of the porous carbon performs occurred when the tracheidal pore system was brought into contact with the silicon melt [16] and then reacted to form SiC. After amorphous carbon of the surface of tracheidal pore converted to -SiC, liquid silicon must permeate through the primary solid SiC layer formed, latter reaction would occur. Although starting Si contents increased, the conversion speed would decline. The conversion degree of the radial samples was higher than that of the axial sample. The orientation of wood channels in respect of the direction of infiltration affects the ceramization process [15]. The channels of the radial and axial samples were parallel and perpendicular to the direction of infiltration, respectively. Therefore the former was easier than the later during infiltrating silicon processing.
Fig. 5 shows the porosity and density of the axial and radial samples prepared with different starting Si/C ratios. Porosity of all samples declined and density ascended while starting Si contents increased. The volume expansion was caused by conversion from biocarbon templates to SiC and the presence of free silicon were mainly causes of a decrease of porosity. It should illuminate that the key factor was conversion or reaction degree from biocarbon templates to biomorphic C/SiC when starting Si/C ratio was less than 3. The residual silicon and conversion together decided the porosity and density of resulting product when ϕ value exceeded 3. If starting Si content kept on increase and the reaction finished, the residual silicon become factor only alone. 3.4. Microstructures of biocarbon template and resulting products Microstructure of the biocarbon template pyrolyzed at 800 ◦ C is shown in Fig. 6a. Beech wood is of hardwood, which is less homogeneous than softwood (such as pine wood) and consists of basic tissues with libriform fibres and vessel elements [17]. The pores of biocarbon templates have smooth surfaces. During the infiltrating liquid silicon process, liquid silicon firstly filled into the smaller vessel pores and reacted with carbon to form SiC. However, not all of smaller pores were filled as shown in Fig. 6a. There was an amount of amorphous carbon in the samples. When starting Si/C ratio was 2, almost all vessel pores were brought into contact with the silicon melt and then reacted to form SiC (Fig. 6c), but the reaction had not completed, there may be amorphous carbon in strut material. With increasing starting Si content, most of pores were filled by liquid silicon and biocarbon partly or totally returned to SiC, which resulted in those smaller pores disappeared and the dimensions of larger pores decreased. It resulted in decrease of porosity and enhancing of mechanical property of the samples. It is showed that the thinner walls between the larger pores were destroyed or disappeared, and partly larger pores incorporated, as shown in Fig. 6d and e. When the process had finished, the resulting products had
G. Hou et al. / Journal of Materials Processing Technology 182 (2007) 34–38
37
Fig. 6. SEM pattern of biocarbon (a) and biomorphic SiC ceramics prepared with different starting Si/C ratios (ϕ): 1 (b), 2 (c), 3 (d), 4 (e) and 5 (f). Insets of (b) and (c) show magnified images of selected area.
coarse surface of pores because of recrystallization of fabricated SiC.
The results of bending strength and fracture toughness of the resulting products are shown in Figs. 7 and 8, respectively. With increasing of starting Si contents, both bending strength and fracture toughness generally enhanced. While ϕ values were 1 and 5, the average bending strength of resulting products were 27.45
and 174.22 MPa for the axial samples, and 15.82 and 59.43 MPa for the radial samples, respectively. With an increase starting Si/C ratio from 1 to 5, the average fracture toughness enhanced from 0.72 to 2.54 MPa m0.5 for the axial samples, and from 0.63 to 1.58 MPa m0.5 for the radial samples, respectively. The strength and toughness of the axial samples had significantly higher than those of the radial samples, which is caused by the force direction and the anisotropic pore orientation derived from microstructures of original wood. In bending strength test, the loading direction is perpendicular to the direction of wood growth for the axial samples, and parallel to the direction of
Fig. 7. Bending strength of resulting products prepared with different starting Si/C ratio (ϕ).
Fig. 8. Fracture toughness of resulting products prepared with different starting Si/C ratio (ϕ).
3.5. Bending strength and fracture toughness of resulting products
38
G. Hou et al. / Journal of Materials Processing Technology 182 (2007) 34–38
wood growth for the radial samples. In fracture toughness test, although the loading direction is perpendicular to the direction of wood growth for all samples, the notches are perpendicular and parallel to the direction of wood growth for the axial samples and radial samples, respectively. Comparing in the axial samples, cracks are easy to expand in the radial samples. Both the transformation from biocarbon to SiC and increase of the residual silicon bring on enhancing bending strength of the resulting samples. Mechanical properties of highly porous brittle structures such as ceramic honeycombes and foams were considered to depend on the cellular structure size and the variation of strength distribution in the solid materials [18]. The content and relative density of SiC in strut material determined mechanical properties of resulting products. In addition, the residual silicon completed the strut and enhanced the density of resulting products. 4. Conclusions Biomorphic SiC ceramics were prepared by infiltrating different contents liquid Si into biocarbon template from beech. When the starting Si/C ratios (ϕ) increased, the resulting products were biomorphic C/SiC, SiC and SiC/Si. The resulting products had coarse surface of pores because of recrystallization of fabricated SiC. With increasing starting Si contents, the porosity decreased and the bending strength and fracture toughness enhanced. The key factors affecting properties of resulting products were conversion degree from biocarbon to SiC and the residual silicon. The strength and toughness of the axial samples had significantly higher than those of the radial samples because of the force direction and anisotropic pore orientation derived from microstructures of original wood. References [1] M. Singh, B.M. Yee, Reactive processing of environmentally conscious, biomorphic ceramics from natural wood precursors, J. Eur. Ceram. Soc. 24 (2004) 209–217. [2] C.E. Byrne, D.E. Nagle, Cellulose derived composites—a new method for materials processing, Mater. Res. Innov. 1 (1997) 137–144.
[3] E. Vogli, J. Mukerji, C. Hoffman, R. Kladny, H. Sieber, P. Greil, Conversion of oak to cellular silicon carbide ceramic by gas phase reaction with silicon monoxide, J. Am. Ceram. Soc. 84 (2001) 1236– 1240. [4] T. Ota, M. Imaeda, H. Takase, M. Kobayashi, N. Kinoshia, T. Hirashita, H. Miyazaki, Y. Hikichi, Porous titania ceramic prepared by mimicking silicified wood, J. Am. Ceram. Soc. 83 (2000) 1521–1523. [5] C. Zollfrank, R. Kladny, G. Motz, P. Greil, Manufacturing of anisotropic ceramics from preceramic polymers and wood, Ceram. Trans. 114 (2001) 43–46. [6] D.A. Streitwieser, N. Popovska, H. Gerhard, G. Emig, Application of the chemical vapor infiltration and reaction (M-R) technique for the preparation of highly porous biomorphic SiC ceramics derived from paper, J. Eur. Ceram. Soc. 25 (2005) 817–828. [7] O. Chakrabarti, L. Weisensel, H. Sieber, Reactive melt infiltration processing of biomorphic Si–Mo–C ceramics from wood, J. Am. Ceram. Soc. 88 (2005) 1792–1798. [8] C.R. Rambo, H. Sieber, Novel synthetic route to biomorphic Al2 O3 ceramics, Adv. Mater. 17 (2005) 1088–1093. [9] J. Cao, O. Rusina, H. Sieber, Processing of porous TiO2 -ceramics from biological performs, Ceram. Int. 30 (2004) 1971–1974. [10] H. Sieber, C. Hoffmann, A. Kaindl, P. Greil, Biomorphic cellular ceramics, Adv. Eng. Mater. 2 (2000) 105–109. [11] G.J. Qiao, R. Ma, N. Cai, C.G. Zhang, Z.H. Jin, Mechanical properties and micro structure of Si/SiC materials derived from native wood, Mater. Sci. Eng. A 323 (2002) 301–305. [12] J.M. Qian, J.P. Wang, Z.H. Jin, G.J. Qiao, Preparation of macroporous SiC from Si and wood powder using infiltration reaction process, Mater. Sci. Eng. A 358 (2003) 304–309. [13] F.M. Varela-Feria, J. Mart´ınez-Fern´andeza, A.R. de Arellano-L´opez, M. Singh, Low density biomorphic silicon carbide: microstructure and mechanical properties, J. Eur. Ceram. Soc. 22 (2002) 2719– 2725. [14] J.M. Qian, J.P. Wang, G.J. Qiao, Z.H. Jin, Preparation of porous SiC ceramic with a woodlike microstructure by sol–gel and carbothermal reduction processing, J. Eur. Ceram. Soc. 24 (2004) 3251–3259. [15] L. Esposito, D. Sciti, A. Piancastelli, A. Bellosi, Microstructure and properties of porous -SiC templated from softwoods, J. Eur. Ceram. Soc. 24 (2004) 533–540. [16] P. Greil, T. Lifka, A. Kaindl, Biomorphic cellular silicon carbide ceramics from wood. I. Processing and microstructure, J. Eur. Ceram. Soc. 18 (1998) 1961–1974. [17] C. Zollfrank, H. Sieber, Microstructure and phase morphology of wood derived biomorphous SiSiC-ceramics, J. Eur. Ceram. Soc. 24 (2004) 495–506. [18] J.S. Huang, L.J. Gibson, Fracture toughness of brittle honeycombe, Acta Metal. Mater. 39 (1991) 1617–1625.
Effect of starting Si contents on the properties and structure of biomorphic SiC ceramics Guangya Hou ∗ , Zhihao Jin, Junmin Qian State Key Laboratory for Mechanical Behavior of Materials, Xi’an Jiaotong University, Xi’an 710049, PR China Received 22 March 2006; received in revised form 12 June 2006; accepted 8 July 2006
Abstract Biomorphic SiC ceramics were prepared by infiltrating liquid Si into biocarbon templates derived from beech at 1550 ◦ C for 1.5 h with different starting Si/C ratios. Microstructure observation and phase identification of the biocarbon templates and the resulting products were examined by scanning electron microscopy (SEM) and X-ray diffraction (XRD), respectively. Open porosity and density of the resulting products were measured by the Archimedes method. The bending strength and fracture toughness were tested by the three-point bending method and the single edge notched beam (SENB) method, respectively. The conversion degrees from biocarbon templates to biomorphic SiC ceramics were also discussed. Experimental results showed that, with increasing starting Si contents, the resulting products were biomorphic C/SiC, SiC and SiC/Si, the porosity decreased, and bending strength and fracture toughness increased. The key factors affecting properties of the resulting products were conversion degree and the amount of the residual silicon. The bending strength and toughness fracture of the axial samples were much higher than those of the radial samples because of loading direction and anisotropic pore orientation derived from microstructures of original wood. © 2006 Elsevier B.V. All rights reserved. Keywords: Biomorphic ceramics; Properties; Structure; SiC
1. Introduction In recent years, there has been an increasing interest in using biomimetic-based processing approaches to fabricate a variety of oxide and non-oxide based structural and functional materials [1]. Biomorphic ceramics or composites such as SiC, Al2 O3 , TiC, TiO2 and Si–Mo–C have been fabricated from wood, wood wastes or papers via using the microstructural features of naturally grown materials [2–9]. Biomorphic SiC ceramics are firstly and extensively studied because of easy preparation processing and good properties. It is interesting of technical applications such as heat insulation structures, substrates, filter and catalyst carriers at high temperatures, thermally and mechanically loaded lightweight structures, as well as for medical implant structures (e.g., porous materials for bone-substitution) [10]. Previous methods converting wood into various SiC ceramic materials are the infiltration of the pyrolysed biocarbon template with gaseous or liquid sili-
∗
Corresponding author. Tel.: +86 29 2667942; fax: +86 29 2665443. E-mail address: [email protected] (G. Hou).
0924-0136/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2006.07.003
con bearing precursors such as silicon melt, silicon, and silicon monoxide vapors, and organosilicon compounds at high temperatures, sol–gel and carbothermal reduction processing and chemical vapor infiltration and reaction (CVI-R) processing [7,11–14]. However, most of works on infiltrating liquid Si into biocarbon templates to manufacture biomorphic SiC or SiC/Si materials were focused on fabrication and properties of the final products. As far as we know, the relationship of starting Si contents and properties of biomorphic SiC ceramics has not been reported. In the present work, it focus on investigating the effects of starting Si contents on the properties and structures of biomorphic SiC ceramics. Biomorphic SiC ceramics were prepared by infiltrated different contents liquid Si into the biocarbon templates derived from beech. 2. Experimental procedure 2.1. Material preparation The wood pieces with 10 mm × 10 mm × 50 mm were cut from beech and dried at 105 ◦ C for more than 24 h, which were divided to axial samples and radial samples according to the direction of length perpendicular and parallel to the direction of wood growth, as shown in Fig. 1a and b. The samples were
G. Hou et al. / Journal of Materials Processing Technology 182 (2007) 34–38
35
Fig. 1. Sketch of bending strength test (a and b) and fracture toughness test (c and d): axial samples (a and c) and radial samples (b and d).
impregnated with phenolic resin in order to reinforce the cell wall and to avoid distortion or cracking of samples during pyrolyzation process. The effects of phenolic resin on the basic properties of woodceramics and biomorphic SiC ceramics were discussed in another our paper. Industrial silicon powder with a diameter range of 1–3 mm was used as a silicon source. All wood samples were pyrolyzed at 800 ◦ C for 4 h with flowing N2 protection to prepare the biocarbon templates, and then, packed in different contents silicon powder and heated at 1550 ◦ C for 1.5 h to form biomorphic SiC in N2 atmosphere. Finally, after the over of holding time, the samples were continually heated up to 1700 ◦ C for 40 min in vacuum to get rid of remnant silicon. The silicon at the temperature was transformed into gaseous Si and expelled from the samples. The crystal structures of the biocarbon, silicon and -SiC at room temperature are amorphous, cubic and face-centered cubic type, respectively, and the corresponding densities are 1.78, 2.32 and 3.21 g/cm−3 , respectively. The weight ratio of the starting silicon powder to the biocarbon templates was defined as ϕ (ϕ = WSi /Wbio-C ).
2.2. Characterization The XRD patterns of the biocarbon templates and the resulting products were recorded using X-ray diffractometer (D/MAX-RA, Rigaku, Japan) with nickel filtered Cu K␣ radiation produced at 35 kV and 20 mA. The tested samples were broken and ground into powders in a carnelian mortar. The microstructure morphology was observed with scanning electron microscope (SEM) (S-2700, Hitachi, Japan) operated at 20 kV and 20 mA. Open porosity and density of resulting products were determined by the Archimedes method with distilled water as the liquid medium. Strength was tested by the three-point bending method with a 20 mm span. Fracture toughness was tested with the single edge notched beam (SENB) method (Fig. 1). A notch with 2 mm depth and 0.28 mm width was cut in the middle of each SENB specimen. All the mechanical properties testing were carried out by a SANS-5104A machine (SANS Ltd., Tianshui, China) and crosshead speed was 0.5 mm/min. The degree of conversion from biocarbon to biomorphic SiC ceramics according to the following relationship: conversion degree (%) =
Wfinal − WC × 100 WC × SiAM /CAM
Fig. 2. XRD pattern of biocarbon (a) and biomorphic SiC ceramics prepared at different starting Si/C ratios (ϕ): 1 (b), 2 (c), 3 (d), 4 (e) and 5 (f).
of 15–50◦ . It can be seen that two broad peaks in Figs. 2a and 3a suggest the biocarbon is amorphous. With increasing starting silicon powder content or ϕ values, the intensity of amorphous carbon decreased or even disappeared. The resulting product may be biomorphic C/SiC ceramics when ϕ values were less than 3 (Figs. 2c and 3c). Pure SiC phase may obtained when ϕ value was 3. However, the diffraction intensity of SiC is significant higher than those of amorphous carbon, the carbon was not completely disappeared under the condition. When ϕ values were 4 and 5, there exist the diffraction peaks corresponding to -SiC and the peaks corresponding to Si in the XRD pattern of the resulting products, which indicated that the resultant were a diphase composites consisting of major phase SiC and secondary phase Si. The intensity of peaks of Si phase increased significantly when ϕ value was 5. It indicates that the residual silicon was increased with the contents of starting silicon powder increasing. Although all samples were gone through removing residual silicon process, there were not pure SiC phase when ϕ values was higher. According to the reaction equation of C and Si, pure SiC can be obtained when the weight ratio of starting silicon powder to biocarbon (ϕ) is about 2.34. From the XRD
(1)
where Wfinal is the weight of the resulting products after the reaction and WC the weight of the biocarbon template, SiAM and CAM are the atomic mass of silicon and carbon, respectively.
3. Results and discussion 3.1. XRD analysis Fig. 2 shows the effect of starting Si contents on the XRD patterns of biocarbon templates and biomorphic SiC ceramics. Fig. 3 is enlarged patterns of Fig. 2 at diffraction angle (2θ) range
Fig. 3. Enlarged pattern of Fig. 2 at diffraction angle (2θ) range of 15–50.
36
G. Hou et al. / Journal of Materials Processing Technology 182 (2007) 34–38
Fig. 4. The degree of conversion from biocabron to biomorphic SiC at different starting Si/C ratios (ϕ).
Fig. 5. porosity and density of axial and radial samples prepared with different starting Si/C ratios. ϕ = 0 means biocarbon templates.
pattern (Fig. 2d), the resulting product may consist of pure SiC phase when ϕ value was in the range of 3–4.
3.3. Open porosity and density of resulting products
3.2. Conversion degree from biocarbon to resulting products The degree of conversion from biocarbon to biomorphic SiC is shown in Fig. 4. It can be seen that the conversion degree of both axial samples and radial samples increased with increasing starting Si/C ratio. When ϕ was 3, the conversion degree of the axial samples and the radial samples were 68.42% and 72.72%, respectively. It showed the reaction did not finish and there were residual carbon in the resulting products. The Xray diffraction pattern shows only -SiC peaks, but considering the low degree of conversion, amorphous carbon is probably present in this sample [15]. The conversion degree exceeded 100% when ϕ value was 5. From XRD patterns of the resulting products (Fig. 2), there were the present of residual silicon in the resulting products, which was called as biomorphic SiC/Si material. Therefore, in this case the conversion degree exceeded 100% was reasonable. In fact, the conversion degree values may result overestimated because the presence of residual silicon in the resulting products was neglected in Eq. (1). There was the fastest conversion stage while starting Si/C ratio (ϕ) increased to 2. This was starting stage of reaction. Spontaneous wetting and infiltration of the porous carbon performs occurred when the tracheidal pore system was brought into contact with the silicon melt [16] and then reacted to form SiC. After amorphous carbon of the surface of tracheidal pore converted to -SiC, liquid silicon must permeate through the primary solid SiC layer formed, latter reaction would occur. Although starting Si contents increased, the conversion speed would decline. The conversion degree of the radial samples was higher than that of the axial sample. The orientation of wood channels in respect of the direction of infiltration affects the ceramization process [15]. The channels of the radial and axial samples were parallel and perpendicular to the direction of infiltration, respectively. Therefore the former was easier than the later during infiltrating silicon processing.
Fig. 5 shows the porosity and density of the axial and radial samples prepared with different starting Si/C ratios. Porosity of all samples declined and density ascended while starting Si contents increased. The volume expansion was caused by conversion from biocarbon templates to SiC and the presence of free silicon were mainly causes of a decrease of porosity. It should illuminate that the key factor was conversion or reaction degree from biocarbon templates to biomorphic C/SiC when starting Si/C ratio was less than 3. The residual silicon and conversion together decided the porosity and density of resulting product when ϕ value exceeded 3. If starting Si content kept on increase and the reaction finished, the residual silicon become factor only alone. 3.4. Microstructures of biocarbon template and resulting products Microstructure of the biocarbon template pyrolyzed at 800 ◦ C is shown in Fig. 6a. Beech wood is of hardwood, which is less homogeneous than softwood (such as pine wood) and consists of basic tissues with libriform fibres and vessel elements [17]. The pores of biocarbon templates have smooth surfaces. During the infiltrating liquid silicon process, liquid silicon firstly filled into the smaller vessel pores and reacted with carbon to form SiC. However, not all of smaller pores were filled as shown in Fig. 6a. There was an amount of amorphous carbon in the samples. When starting Si/C ratio was 2, almost all vessel pores were brought into contact with the silicon melt and then reacted to form SiC (Fig. 6c), but the reaction had not completed, there may be amorphous carbon in strut material. With increasing starting Si content, most of pores were filled by liquid silicon and biocarbon partly or totally returned to SiC, which resulted in those smaller pores disappeared and the dimensions of larger pores decreased. It resulted in decrease of porosity and enhancing of mechanical property of the samples. It is showed that the thinner walls between the larger pores were destroyed or disappeared, and partly larger pores incorporated, as shown in Fig. 6d and e. When the process had finished, the resulting products had
G. Hou et al. / Journal of Materials Processing Technology 182 (2007) 34–38
37
Fig. 6. SEM pattern of biocarbon (a) and biomorphic SiC ceramics prepared with different starting Si/C ratios (ϕ): 1 (b), 2 (c), 3 (d), 4 (e) and 5 (f). Insets of (b) and (c) show magnified images of selected area.
coarse surface of pores because of recrystallization of fabricated SiC.
The results of bending strength and fracture toughness of the resulting products are shown in Figs. 7 and 8, respectively. With increasing of starting Si contents, both bending strength and fracture toughness generally enhanced. While ϕ values were 1 and 5, the average bending strength of resulting products were 27.45
and 174.22 MPa for the axial samples, and 15.82 and 59.43 MPa for the radial samples, respectively. With an increase starting Si/C ratio from 1 to 5, the average fracture toughness enhanced from 0.72 to 2.54 MPa m0.5 for the axial samples, and from 0.63 to 1.58 MPa m0.5 for the radial samples, respectively. The strength and toughness of the axial samples had significantly higher than those of the radial samples, which is caused by the force direction and the anisotropic pore orientation derived from microstructures of original wood. In bending strength test, the loading direction is perpendicular to the direction of wood growth for the axial samples, and parallel to the direction of
Fig. 7. Bending strength of resulting products prepared with different starting Si/C ratio (ϕ).
Fig. 8. Fracture toughness of resulting products prepared with different starting Si/C ratio (ϕ).
3.5. Bending strength and fracture toughness of resulting products
38
G. Hou et al. / Journal of Materials Processing Technology 182 (2007) 34–38
wood growth for the radial samples. In fracture toughness test, although the loading direction is perpendicular to the direction of wood growth for all samples, the notches are perpendicular and parallel to the direction of wood growth for the axial samples and radial samples, respectively. Comparing in the axial samples, cracks are easy to expand in the radial samples. Both the transformation from biocarbon to SiC and increase of the residual silicon bring on enhancing bending strength of the resulting samples. Mechanical properties of highly porous brittle structures such as ceramic honeycombes and foams were considered to depend on the cellular structure size and the variation of strength distribution in the solid materials [18]. The content and relative density of SiC in strut material determined mechanical properties of resulting products. In addition, the residual silicon completed the strut and enhanced the density of resulting products. 4. Conclusions Biomorphic SiC ceramics were prepared by infiltrating different contents liquid Si into biocarbon template from beech. When the starting Si/C ratios (ϕ) increased, the resulting products were biomorphic C/SiC, SiC and SiC/Si. The resulting products had coarse surface of pores because of recrystallization of fabricated SiC. With increasing starting Si contents, the porosity decreased and the bending strength and fracture toughness enhanced. The key factors affecting properties of resulting products were conversion degree from biocarbon to SiC and the residual silicon. The strength and toughness of the axial samples had significantly higher than those of the radial samples because of the force direction and anisotropic pore orientation derived from microstructures of original wood. References [1] M. Singh, B.M. Yee, Reactive processing of environmentally conscious, biomorphic ceramics from natural wood precursors, J. Eur. Ceram. Soc. 24 (2004) 209–217. [2] C.E. Byrne, D.E. Nagle, Cellulose derived composites—a new method for materials processing, Mater. Res. Innov. 1 (1997) 137–144.
[3] E. Vogli, J. Mukerji, C. Hoffman, R. Kladny, H. Sieber, P. Greil, Conversion of oak to cellular silicon carbide ceramic by gas phase reaction with silicon monoxide, J. Am. Ceram. Soc. 84 (2001) 1236– 1240. [4] T. Ota, M. Imaeda, H. Takase, M. Kobayashi, N. Kinoshia, T. Hirashita, H. Miyazaki, Y. Hikichi, Porous titania ceramic prepared by mimicking silicified wood, J. Am. Ceram. Soc. 83 (2000) 1521–1523. [5] C. Zollfrank, R. Kladny, G. Motz, P. Greil, Manufacturing of anisotropic ceramics from preceramic polymers and wood, Ceram. Trans. 114 (2001) 43–46. [6] D.A. Streitwieser, N. Popovska, H. Gerhard, G. Emig, Application of the chemical vapor infiltration and reaction (M-R) technique for the preparation of highly porous biomorphic SiC ceramics derived from paper, J. Eur. Ceram. Soc. 25 (2005) 817–828. [7] O. Chakrabarti, L. Weisensel, H. Sieber, Reactive melt infiltration processing of biomorphic Si–Mo–C ceramics from wood, J. Am. Ceram. Soc. 88 (2005) 1792–1798. [8] C.R. Rambo, H. Sieber, Novel synthetic route to biomorphic Al2 O3 ceramics, Adv. Mater. 17 (2005) 1088–1093. [9] J. Cao, O. Rusina, H. Sieber, Processing of porous TiO2 -ceramics from biological performs, Ceram. Int. 30 (2004) 1971–1974. [10] H. Sieber, C. Hoffmann, A. Kaindl, P. Greil, Biomorphic cellular ceramics, Adv. Eng. Mater. 2 (2000) 105–109. [11] G.J. Qiao, R. Ma, N. Cai, C.G. Zhang, Z.H. Jin, Mechanical properties and micro structure of Si/SiC materials derived from native wood, Mater. Sci. Eng. A 323 (2002) 301–305. [12] J.M. Qian, J.P. Wang, Z.H. Jin, G.J. Qiao, Preparation of macroporous SiC from Si and wood powder using infiltration reaction process, Mater. Sci. Eng. A 358 (2003) 304–309. [13] F.M. Varela-Feria, J. Mart´ınez-Fern´andeza, A.R. de Arellano-L´opez, M. Singh, Low density biomorphic silicon carbide: microstructure and mechanical properties, J. Eur. Ceram. Soc. 22 (2002) 2719– 2725. [14] J.M. Qian, J.P. Wang, G.J. Qiao, Z.H. Jin, Preparation of porous SiC ceramic with a woodlike microstructure by sol–gel and carbothermal reduction processing, J. Eur. Ceram. Soc. 24 (2004) 3251–3259. [15] L. Esposito, D. Sciti, A. Piancastelli, A. Bellosi, Microstructure and properties of porous -SiC templated from softwoods, J. Eur. Ceram. Soc. 24 (2004) 533–540. [16] P. Greil, T. Lifka, A. Kaindl, Biomorphic cellular silicon carbide ceramics from wood. I. Processing and microstructure, J. Eur. Ceram. Soc. 18 (1998) 1961–1974. [17] C. Zollfrank, H. Sieber, Microstructure and phase morphology of wood derived biomorphous SiSiC-ceramics, J. Eur. Ceram. Soc. 24 (2004) 495–506. [18] J.S. Huang, L.J. Gibson, Fracture toughness of brittle honeycombe, Acta Metal. Mater. 39 (1991) 1617–1625.