Biomorphic ceramics from wood-derived precursors

Biomorphic Ceramics from Wood Derived Precursors J. Ramírez-Ricoa * , J. Martínez-Fernandeza, M. Singhb a Dpto. Física de la Materia Condensada. Instituto de Ciencia de Materiales de Sevilla (ICMS). Universidad de Sevilla. Consejo Superior de Investigaciones Científicas (CSIC). Avda. Reina Mercedes S/N, 41012 Sevilla, Spain 2 Ohio Aerospace Institute, 22800 Cedar Point Road, Cleveland, OH 44142, USA Joaquín Ramírez Rico obtained his PhD from the University of Seville in 2008, where he is currently Assistant Professor in the Solid State Physics Department at the University of Seville. He has authored more than 55 papers in peer-reviewed journals and his current interests are in natural and bioinspired material, novel structural ceramics and ceramicmatrix composites, and in-situ materials characterization using x-ray techniques. Julián Martinez Fernandez is Professor of Solid State Physics and Materials Science, at the University of Seville, Spain, where he also serves as the Director of Research Facilities. He has published over 160 papers in peer-reviewed journals and has been the PI of over 20 nationally and internationally funded research grants. His research interests include the study of biomorphic silicon carbide, single crystal and polycrystalline zirconia and alumina, alumina-based composites, silicon nitride, electronic ceramics, ceramic joining, and single crystals fibers. Mrityunjay Singh is Chief Scientist, Ohio Aerospace Institute, NASA Glenn Research Center, Cleveland, Ohio. He has edited/co-edited thirty eight books and five journal volumes, authored/co-authored ten book chapters/invited reviews and more than two hundred fifty papers in journals and edited volumes. He is involved with various activities in processing, manufacturing, joining and attachment technologies, and characterization of advanced ceramics and composites, lightweight cellular ceramics, environment conscious ceramics, and porous ceramic foams, high conductivity composites and graphite foams for thermal management systems, ceramic matrix composites for turbomachinery and propulsion systems, and a wide variety of materials for ultra high temperature and extreme environment applications. * Corresponding author – Email: [email protected]s, Tel: +34 954 550 936 Biomorphic Ceramics from Wood Derived Precursors Materials development is driven by microstructural complexity, in many cases inspired by biological systems like bones, shells, and wood. In one approach, one selects the main microstructural features responsible for improved properties a designs processes to obtain materials with such microstructures (continuous-fiber-reinforced ceramics, porous ceramics, fibrous ceramic monoliths, etc.). In a different approach, it is possible to use natural materials directly as microstructural templates. Biomorphic ceramics are produced from natural and renewable resources (wood or wood-derived products). A wide variety of SiC based ceramics can be fabricated by infiltration of silicon or silicon alloys into cellulose-derived carbonaceous templates, providing a low-cost route to advanced ceramic materials with near-net shape potential and amenable to rapid prototyping. These materials have tailorable microstructure and properties, and behave like ceramic materials manufactured by advanced ceramic processing approaches. This review aims to be a comprehensive description of the development of bioSiC ceramics: from wood templates and their microstructure to potential applications of bioSiC materials. Keywords: biomorphic; ceramics; wood; silicon carbide; mechanical properties; high temperature; porous materials 1 Introduction Silicon carbide-based ceramics are interesting materials for key engineering applications, especially those requiring good thermomechanical performance and specific properties. However, in many cases SiC materials do not find extensive usage due to their high cost, as they are obtained at very high sintering temperatures and require expensive manufacturing (extrusion, molding, and machining, etc) techniques to produce complexshaped components. For this reason, SiC ceramics are often limited to niche applications. Macroporous ceramics [1] and specifically macroporous silicon carbide materials [2] are also interesting candidates for a wide range of applications, especially in high temperature environments involving corrosive gases and high mechanical loading. Envisioned applications could be catalyst supports, thermal insulation and management, or filtration of molten metals, liquids or gases (such as in diesel particulate filters or in gasification processes). Thus, there is technological potential in any process that allows one to obtain SiC monoliths with reduced processing and machining costs, but also that enables the production of porous SiC ceramics, ideally with the same low cost attributes as outlined above. A way to overcome these problems is to synthesize SiC ceramics by melt infiltration of a porous carbonaceous preform, typically obtained from the pyrolysis of a cast polymer, to produce Reaction Formed SiC (RFSC) [3-10], a process developed in the early 90s that is near-net shape and allows to obtain monolithic SiC materials with small amounts of unreacted carbon and silicon. The microstructure of these materials depend on the characteristics of the carbon template and are, in a way, inherited from it, so by tailoring the template and increasing its microstructural complexity, a wide variety of SiC materials could be obtained. Biomorphic silicon carbide (bioSiC) are bio-templated SiC materials that start with a porous carbon scaffold derived from natural precursors such as wood, which is later converted into a SiC ceramic by a variety of methods, most commonly by melt infiltration. In this case, bioSiC is simply a type of reaction formed silicon carbide in which the carbon template is obtained by pyrolysis of wood, with the advantage of a complex, hierarchical microstructure that is the result of millions of years of evolution and is optimized for properties such as fluid transport or high strength to density ratios [11]. Several routes have been developed to obtain SiC ceramics starting from wood-derived carbon preforms, in a way that retains most of the microstructural features of the template. Figure 1 summarizes the three most common ones: melt-infiltration of liquid Si, vapor infiltration of Si either directly or through decomposition of SiO, or infiltration of SiO2 sols or gel precursors that react with carbon via carbothermal reduction to form SiC. All three processes start with a carbon template obtained from a wood precursor that is chosen according to the desired properties of the final material, although in some works wood is impregnated with a SiO2 precursor so carbonization and carbothermal reduction is carried out in a single step. All three routes also have in common that they are near-net- shape processes because after conversion the size and shape of the final material is equal to that of the preform, and any volume change associated with the reaction is buffered by the precursor’s porosity. This makes the conversion of carbonaceous precursors into SiC a highly cost effective method to produce SiC ceramics, porous or otherwise, as any desired machining can be done to the carbon preform and only finishing will need to be performed in the final component, greatly decreasing processing costs when compared to typical sintering and hot-pressing processes that might require extensive machining with expensive diamond tools. As an example of the complexity of shapes that can be produced, Figure 2 shows some components manufactured by melt infiltration of MDF derived bioSiC/Si, where all machining was performed in the carbon stage. From a historical point of view, the first mentions in the literature of using wood as a template for advanced materials processing is from Ota et al. [12] who synthesized “SiC wood” by tetraethyl-orthosilicate infiltration and later carbothermal reduction of wood chars. Byrne and Nagle made important contributions to the field by laying the groundworks of wood pyrolysis with materials applications in view, such as optimization of heating rates, densities and shrinkage ratios, as well as their effect on mechanical properties [13-15]. Later reports on biomorphic SiC are from Greil and coworkers [16-19], as well as Singh, who coined the term “Ecoceramics” to refer to this new class of woodderived materials [20, 21]. Several general papers exist that deal with the basics of biomorphic SiC materials [19, 20, 22-24]. 2 Carbon Templates from Natural Wood Precursors 2.1 Wood Structure and Composition Wood can be considered a natural material with a complex microstructure optimized by evolution for a good strength-to-density ratio and fluid transport [25]. It is a porous material with densities in the range 0.3-0.8 g cm-3 depending on species. The wide variety of commercially available wood from different species, and the possibility of using industrial wood products (paper, cardboard, cellulose paste, particle and fiberboards) allows the processing of bioSiC materials with a wide range of properties depending on the targeted application. The structure of wood has been studied in detail [26-29] and is composed of a stacking of elongated, tubular cells aligned in the tree’s growth direction. Wood can be broadly classified into two groups attending on its microstructure: in softwood from gymnosperm trees such as conifers, wood is composed of only one type of cell, called tracheids, while in angiosperm wood these are surrounded by smaller sclerenchymatic cells with a distribution that is characteristic of each species. Among them, small sized, thick walled cells are responsible for the mechanical strength of wood and receive the name of libriform “fibers”. In angiosperms, tracheids can have diameters over one hundred microns and are responsible for fluid transport (sap channels). Due to this structure, pore distribution in angiosperm wood is bimodal, while in wood from gymnosperm species the pore size distribution is monomodal with channels in the range of 30-50 μm. Figure 3 shows a schematic of wood from both angiosperm and gymnosperm precursors. The structure and composition of cell walls determines the properties of wood and woodderived products [30] and is similar for soft and hardwoods. This can be described as a lamellar or layered microstructure where each layer is composed of cellulose, hemicellulose and lignin (lignocellulose). Celluloses are long polymeric chains (up to 1 µm in length) aggregated into microfilaments normally aligned with the cell’s longest dimension. Hemicelluloses are shorter chains interconnecting cellulosic microfilaments while lignin occupies the spaces left by the cellulose and hemicellulose networks. By its very nature, wood has a highly anisotropic structure and its properties are described along three planes and their respective perpendicular directions: axial, radial and tangential (Figure 3). In addition to the presence of pores and cells, wood shows two types of macrostructures: rays and growth rings. Rays are horizontally aligned cells used for nutrient storage and radiate from the axis of the three, while growth rings are concentric rings that result from changing pore sizes due to seasonal variations in tree growth. Due to the absence of sap channels, softwoods from gymnosperms are usually referred to as non-porous, while hardwoods from angiosperms are classified according to their growth ring morphology into ring porous or diffuse porous: in ring porous hardwoods sap channels are segregated into bands while in diffuse-porous wood there is no such segregation (Figure 4) [28, 29]. 2.2 Pyrolysis of Wood Pyrolysis of wood for the production of carbon templates to be further converted into SiC ceramics is performed in inert atmospheres (typically Ar or N2) by slowly heating the wood precursors up to temperatures in excess of 800°C. During pyrolysis the different polymeric components of wood decompose in a step-wise manner with hemicellulose breaking-down first at 200-260 °C, followed by cellulose at 240-350ºC and lignin at 280- 500°C and up. Figure 5 shows typical thermogravimetric and loss rate per unit temperature curves for thermal decomposition of Quercus rubra (red oak) wood, exhibiting the step-wise process described earlier [14]. Labeled regions correspond to (1) loss of adsorbed water, (2) polymer decomposition of hemicellulose and (3) cellulose as well as lignin, (4) total decomposition of cellulose and decomposition of remaining lignin above 400 ºC (5). These stages more or less correspond to those described by Tang and Bacon and others [31], involving (a) the desorption of adsorbed water up to 150 ºC; (b) the splitting off of cellulose structure water between 150 and 240 °C; (c) depolymerization and breaking of C-O and C-C bonds within ring units evolving water, CO and CO2; (d) aromatization forming graphitic layers above 400 °C. In lignocellulosic materials such as wood the carbon-to-volatiles mass ratio increases with pyrolysis temperature, yielding ~90% carbon at 600 °C, while the rest is mostly nitrogen and oxygen, with small amount of other elements. Ash content varies with species and is between 0.5 to 3 wt. % of the carbon monoliths. Ashes from wood derived carbon contain mostly sodium, magnesium, phosphor, potassium, calcium, manganese and iron. Since crack-free monoliths are required for use as templates in SiC ceramic synthesis, pyrolysis must be carefully controlled and performed at very low heating rates, in the range of 1ºC/min up to 400-500 °C, where most of the thermal decomposition of polymers takes place. Such a slow rate allows for gaseous reaction products to evacuate the material through the pores without building up pressure that could lead to cracking. After that slow initial stage, pyrolysis can be accelerated to rates of the order of 10-20 °C/min without affecting the integrity of the final template. This slow process has the advantage of producing high carbon yields, in the 25-30 wt. % range when compared to the initial wood mass. It has been shown that the density of the carbon scaffold obtained by pyrolysis is proportional to that of the wood precursor, as was first observed by Byrne and Nagle as well as others [14]. Some examples of this are shown in Figure 6, where a linear fit of the form 𝜌𝑐=𝛼𝜌𝑤𝑜𝑜𝑑 is included. Values of the parameter 𝛼 vary with pyrolysis conditions, being mostly affected by the heating rate, but is roughly ~0.8 for slow pyrolysis conditions such as those typically used to obtain preforms for Si infiltration. Pyrolysis is accompanied by a shrinkage of the wood-carbon monolith, the extent of which is determined by the wood employed. It has been shown that shrinkage in the radial and tangential directions vary greatly from species to species, in the range of 25-40 % length reduction, while axial shrinkage is similar for all types of wood and results in ca. 20 % reduction, although these values themselves depend on the pyrolysis temperature, remaining constant over 1000 ºC (Table 1). 3 Biomorphic ceramics by melt infiltration of carbon preforms Processing of bioSiC by melt infiltration is well documented in the literature, and protected by patents [18, 22, 24, 32-40]. Basically, it involves the infiltration of liquid silicon into the carbon preform, by heating silicon powder in contact with it at temperatures in excess of the melting point of silicon (typically > 1450°C) in vacuum. Silicon quantities well above the stoichiometric ratio to carbon are typically used to ensure almost-complete conversion of the carbon template to SiC. During reaction, the carbon walls in the template are converted into β-SiC and the smaller pores in the template, up to a pore diameter of approximately 50 µm, are filled with residual silicon. The Si, SiC and residual carbon content of the final bioSiC/Si composite, as well as the remaining pores’ size and distribution, if any, depend on the processing conditions, amount of excess Si, and the anatomy of the wooden preform. Infiltration of SiC is a near-net shape process, as the volume expansion associated with the conversion of C into SiC is absorbed by the porous structure leaving the outer dimensions unchanged. Varela-Feria [41] performed a series of infiltrations using carbon templates shaped as parallelepipeds with sides of different lengths and measured the dimensions of the resulting bioSiC/Si materials (Figure 8), with the result that external sample dimensions did not change upon melt infiltration regardless of initial dimensions and aspect ratio. In melt infiltration, conversion of carbon into SiC takes place by two mechanisms which are active depending on the carbon wall thickness and topology. As will be elaborated further below, the main mechanism is dissolution of carbon into the melt followed by reprecipitation of micron-sized (ca. 10 µm) β-SiC grains at the carbon-silicon interface: silicon penetrates through the carbon preform by capillary effects and reacts with the solid C spontaneously and exothermically, with an enthalpy of ΔH0=-117.77 kJ/mol [42]. In denser areas of the carbon template the carbon density can exceed the critical limit of about 0.97 g cm-3 which can result in clogging of the pores preventing further SiC formation by precipitation [5-7]. In these cases solid state diffusion and reaction control the formation of SiC now at the SiC-carbon interface, resulting in the formation of a nanograined SiC phase (< 100 nm) [24, 39, 40, 43, 44]. This mechanism is corroborated first by microstructural observations. Figure 9 shows the main microstructural features in bioSiC. Three different phases can be distinguished: polycrystalline SiC shows intermediate contrast under backscattered electrons in the SEM (A, B and C), while Si appears light grey and residual carbon appears black. Micron-sized SiC grains can be found inside channels where Si is majority (large channels with diameter over 5 µm). In small channels where Si is usually depleted in the reaction, a layer of nanosized SiC grains can be observed at the interface between unreacted carbon and micronsized SiC grains, which are typically majority (panel D). Nanosized SiC is found at SiC- carbon interfaces, but not at SiC-Si interfaces. Where thin carbon walls surrounding the large channels existed, large grained SiC appears forming layers. Occasionally these layers are discontinuous and even some isolated SiC grains are observed (figure 9B, marked with arrows). At the SiC-carbon interfaces nanosized grains can be found forming rosettes which is typical of diffusion-controlled growth [45], while micron-sized SiC grains are faceted, typical of a precipitation mechanism. Figure 10 shows transmission electron microscopy images of the interfaces in the region of narrow and large channels, along with selected area diffraction patterns, corroborating the previous observations. Microstructural parameters of the resulting materials are inherited from the wood template and therefore show great variability, which means that they can be tailored by adequate selection of the precursor. Table 1 shows several parameters such as density, silicon volume fraction or porosity for melt infiltrated bioSiC from different precursors and in comparison with the carbon template. The β-SiC formation has been shown to follow first order kinetics, so the concentration of each phase can be modelled using: [𝐶𝑖]=[𝐶𝑖𝑓]+([𝐶𝑖0]−[𝐶𝑖𝑓])exp𝑘𝑡 Where k is the reaction constant and Ci is the concentration of phase i. [𝐶𝑖𝑓] final concentration for each phase, which is not necessarily zero since there is excess Si as well as unreacted carbon in the final material. [𝐶𝑖0] is the initial concentration of phase i. Figure 11 shows the evolution of each phase’s concentration with time, along with fits using the previous equation, from a work by Varela-Feria et al. [39]. In their work, the reaction constant was determined as k=1.8 · 10-2 s-1, similar to that measured by Pampuch et al., who studied the SiC formation process by reaction of bundles of C fibres 4-6 µm in diameter with liquid Si at 1422 °C and 1439 °C [46, 47], and determined the reaction constant k using DTA analysis. The possibility of a diffusion-controlled reaction rate is not compatible with the measured k values, as using data from Hon et al. [45, 48] and assuming the reaction was controlled by diffusion through the SiC layer, a value in the range of 𝑘 =1.0−3.0·10−10𝑠−1 would be expected. 4 Macroporous SiC Ceramics 4.1 Residual Si Removal In the processing of wood-derived SiC ceramics obtained by melt infiltration Si is usually added in excess with respect to amount needed for stoichiometric reaction, which results in the presence of residual Si filling some or all of original channels of the wood precursor. Removal of this residual Si results in a macroporous SiC ceramic with elongated, anisotropic and hierarchical porosity that is interesting for a wide range of applications, some of which are discussed in section 6. Several techniques are available for removal of this secondary phase, the one most often used being chemical etching using a mixture of HNO3 and HF in a molar ratio of 1.67 according to the following reactions: 3Si + 4HNO3 → 3SiO2 + 4NO + 4H2O SiO2 + 4HF → SiF4 + 2H2O The reaction has been shown to be diffusion limited and the etching rate 𝑅 follows that of a boundary layer problem: 𝑅~(𝐷𝑡)−1/2 The evolution of the reaction front as a function of etching time is shown for transverse sections in Figure 12 while Figure 13 [49, 50] shows the reaction front advance as well as the reaction rate in a log scale. The differences in etching rates for the different directions with respect to the growth axis of the wood precursor were attributed to an effective diffusion coefficient 𝐷𝑒𝑓𝑓 which depends on the apparent porosity in each direction due to the anisotropic permeability of the porous bioSiC. The porosity of bioSiC after removal of remaining silicon resembles that of the original wood precursor, however due to the volume expansion associated with the 𝐶+𝑆𝑖(𝑙)→ 𝑆𝑖𝐶 reaction, pores with sizes smaller than ~1 𝜇𝑚 close during the reactive infiltration step and are therefore not present in the porous bioSiC scaffold. Pappacena et al. used mercury intrusion porosimetry to study the pore diameter distribution in both carbon scaffolds obtained from different wood precursors and resulting porous SiC materials and found a total porosity reduction of ~10%, depending on species, mostly derived from the closure of pores < 1 𝜇𝑚 during the infiltration step [51]. The large macroporosity, with diameters 10 𝜇𝑚 and larger, remained basically unchanged (Figure 14). Connectivity properties and excellent thermomechanical performance in high temperature environments. By using wood as the initial template, highly complex microstructures can be obtained, and porous SiC ceramics with hierarchical porosity can be obtained. Since the process is near net-shape, biomorphic SiC can be processed into relatively complex shapes cheaply. All these advantages make them interesting candidates for a wide range of technological applications including catalyst supports, filtration elements in combustion environments, o biomedical devices, to name a few. Acknowledgements Joaquin Ramirez-Rico and Julian Martinez-Fernandez gratefully acknowledge support from the Ministerio de Economía, Industria y Competitividad (MINECO), under grants MAT2016-76526-R and MAT2013-41233-R. We also thank F. M. Varela-Feria and J. Quispe for fruitful discussions and for their help in preparing some of the figures. Figure Captions Figure 1. Different routes to biomorphic SiC materials and composites by infiltration of a porous carbon precursor. Figure 2. Some components made of melt-infiltrated bioSiCp, in this case using MDF as a precursor. Courtesy of biomorphic EBT. Figure 3. (top) Schematic of wood structure and definition of the three main planes used to describe for both gymnosperm and angiosperm specimens. (bottom) Actual microstructure determined from optical microscopy in sections along the three planes. From [28]. Figure 4. Axial sections of wood from three different species, a non-porous softwood (white pine), a ring porous hardwood (red oak) and a diffuse porous hardwood (tulip tree), highlighting the three different types of growth rings. Rays are signaled by red arrows, while growth rings are marked by white arrows. Adapted from [28]. Figure 5. Weight loss (TGA) analysis of the pyrolysis of Quercus rubra (red oak) wood along with the rate of loss with temperature. Labeled regions correspond to (1) loss of adsorbed water, (2) polymer decomposition of hemicellulose and (3) cellulose as well as lignin, (4) total decomposition of cellulose and decomposition of remaining lignin above 400 ºC (5). Figure 6. Final carbon density as a function of wood density before pyrolysis. A linear fit to the data is included. Solid points correspond to data by Byrne and Nagle while hollow points correspond to our own unpublished data. Figure 7. SEM micrographs of axial sections of carbon obtained through pyrolysis of different wood precursors, highlighting the resemblance of the carbon scaffold microstructure to that of the original wood precursor. Figure 8. Final dimensions of melt-infiltrated beech-derived carbon as a function of initial dimension, for parallelepipeds of different sizes and aspect ratios. The straight line has a slope of one, indicating that the dimensions of the original templates were conserved. Figure 9. SEM micrographs of bioSiC obtained from beech wood: A) Low magnification B) High magnification micrograph of SiC-Si interfaces. C) High magnification micrograph of SiC-carbon interfaces. D) TEM micrograph of a SiC-carbon interface showing a rosette of nano-sized grains [39]. Figure 10. Transmission electron microscopy images and associated diffraction patterns of melt infiltrated bioSiC in the vicinity of narrow channels (left) and large channels (right), highlighting the differences stated in the text. In narrow channels a SiC/carbon interface is formed with a layer of nanosized grains, whereas in large channels SiC/silicon interfaces are formed between large (micron-sized) grains. Figure 11. Time dependent phase fractions of carbon, silicon and SiC for silicon melt infiltration of wood-derived carbon preforms [39]. Figure 12. SEM micrographs of transverse sections of porous bioSiC for different etching times, from [49]. From left to right and top to bottom, etching times are 0.5, 1, 2, 6, 18 and 64 h. Figure 13. Etching rate obtained from SEM micrographs of sections of porous bioSiC for different etching times, for sipo wood [49] and MDF derived bioSiCp [50]. The etching rate follows a t -1/2 law that confirms the diffusion limited nature of the reaction. Figure 14. Pore size distributions and porosities of (a) carbon samples pyrolyzed to 1000 ◦C and (b) silicon carbide samples derived from carbon pyrolyzed to 1000 ◦C from five wood precursors as determined by mercury porosimetry. Curves are offset by (a) 20 volume% and (b) 30 volume% for ease of viewing. From [51]. Figure 15. Transverse cross section from a 3D reconstruction obtained by x-ray microcomputed tomography of porous bioSiC. Each colored region represents an interconnected structure, from [52]. Figure 16. Microstructure of porous bioSiC obtained from Si vapor infiltrated-pine char. From [53]. Figure 17. Young’s Modulus and bending strength of carbon preforms and bioSiC/Si composites obtained from Balsa, Pine, Maple, Oak, Beech and Ebony precursors, as a function of density, when the load was applied in the axial or radial orientations. From [16] Figure 18. High temperature stress vs. strain plots for Refel RBSiC (89% SiC), Cerastar RB (80% SiC), Cerastar RX (74% SiC), LS-RFSiC (91% SiC) and bioSiC (63% SiC, fabricated from eucalyptus wood) in the axial direction. A) 1000 °C and B) 1330 °C. From [22]. Figure 19. Compressive strength of different siliconized SiC materials compared to the strength of sintered SiC as a function of SiC fraction in the material. BioSiC obtained from wood with a range of densities from 0.48-0.97 g/cm3 was studied in both the axial and radial directions in the range of 1000-1450 ºC. From [104, 129]. Figure 20. Relative strength of several siliconized SiC ceramics including bioSiC, as a function of SiC volume fraction, at temperatures 1000 – 1300 ºC. The predictions from the minimum solid area models for several ideal microstructures are plotted as solid lines [102, 103]. The experimental values for several siliconized SiC materials including bioSiC are from [99]. Figure 21. Creep strain rates of bioSiC at 1175 ºC and 1600 ºC, at 250 MPa of applied stress. From [99]. Figure 22. Porous BioSiC candle prototypes used for hot-gas filtering experiments at 800ºC, from [113]. Tables Table 1. Properties of carbon preforms and Si-infiltrated SiC materials from different types of wood [16, 18, 95] Balsa Pine Oak Maple Beech Ebony Paulownia MDF Pyrolysis weight loss (wt.%) 73.5 73.8 70.4 74.9 74.2 64.6 70.17 71.6 Pyrolysis shrinkage (%) axial 21 23 17 20 22 14 20.06 24.75 radial 22 28 28 30 32 25 20.81 23.95 tangential 22 31 33 40 38 30 36.55 45.77 Density (g/cm3) Pyrolyzed 0.06 0.31 0.50 0.51 0.55 0.87 0.17 0.72 Siinfiltrated 2.02 2.22 2.16 2.58 2.57 - 2.26 2.94 Porosity (%) (open/closed) Pyrolyzed 22/70 21/57 30/40 43/22 42/21 23/20 75.81 42.65 Siinfiltrated 11/14 11/14 8/5 3/5 3/2 3/- 14 3 Mean pore diameter (µm)* 40 20 170 35 30 - Si-content (wt%) 67 50 27 23 37 - 57 10 * Open pore channels free of Si. Table 2. Microstructural features of pine-derived biomorphic SiC obtained by vapor infiltration of different precursors, from [60] Wood Carbon template Silicon carbide SiO Si CH3SiCl3 Density (g cm-3) Geometrical 0.47 0.35 0.6 1.0 1.2 Struts 1.4 1.4 2.5 3.1 3.1 Porosity (%) 67 76 80 70 60 Surface area (m2 g-1) - 46 16.1 3.3 0.5 Strut thickness (µm) 2 1.5 1 2 4 Biaxial strength (MPa) 4 13 21 Table 3. Room temperature mechanical properties of biomorphic SiC. 𝜎𝑐 and 𝜎𝑓 represent compressive and flexural strength respectively. Wood 𝝆 (g·cm-3) P (%) Direction 𝝈𝒇 (MPa) 𝝈𝒄 (MPa) E (MPa) KIC (MPa·m1/2) Ref. Bamboo 2.38-2.51 51 Axial 120-180 1.75 [130] Beech 2.50-2.61 Axial 180-200 250-290 [16] 2.50-2.61 Radial 90-130 220-280 [16] 2.10 Axial 228 438 2 [92] 2.50 Axial 216 1080 3 [92] Porous 53-55 Axial 330-480 115-130 1.1-1.4 [97] Porous 52-55 Radial 24-24 10-27 0.5-0.7 [97] Birch 2.71-2.77 9 Axial 210-320 2.4-4.3 [130] Bubinga Axial 186-240 2.4-2.8 [22] Eucalyptus 2.60 Axial 226 1410 4.2 [92] 2.63 Axial 290 1400 160-180 2.6-2.8 [91] Mahogany Porous 45-53 Axial 160-190 35-110 0.7-1.22 [97] Porous 46-50 Radial 57-75 15-21 0.46-0.6 [97] Porous 48-50 Tangential 31-35 12-20 - [97] Axial 144 150-195 2 [21] Maple 2.36 Axial 286-402 240-208 2.4-2.8 [21, 22] MDF 2.68-2.72 <1 - 125-245 2.4-3.2 [130] Pine 2.57-2.67 2 Axial 175-225 1.8-4.1 [130] 2.25-2.30 Axial 125 180 [16] 2.25-2.30 Radial 50-65 150-200 [16] Poplar Porous 59-64 Axial 135-250 30-76 0.8-1.1 [97] Porous 64-65 Radial 12 1 0-3 [97] Porous 58-60 Tangential 16-24 6-8 - [97] Red Oak Porous 47-50 Axial 220-270 28-60 0.9-1.8 [97] Porous 44-46 Radial 8-18 5 [97] Porous 44-47 Tangential 5-28 8 [97] 2.05-2.10 Axial 100-120 160-210 [16] 2.05-2.10 Radial 30-60 150-200 [16] Sapelly Porous 54-59 Axial 50-127 15-22 [97] Porous 55-59 Radial 12-15 6-14 [97] Porous 54-57 Tangential 13-18 7-10 [97] References [1] A.R. Studart, U.T. Gonzenbach, E. Tervoort, L.J. Gauckler, Processing routes to macroporous ceramics: A review, J Am Ceram Soc 89(6) (2006) 1771-1789. [2] J.-H. Eom, Y.-W. Kim, S. Raju, Processing and properties of macroporous silicon carbide ceramics: a review, Journal of Asian Ceramic Societies 1(3) (2013) 220-242. [3] Y.M. Chiang, R.P. Messner, C. Terwilliger, D.R. Behrendt, Reaction-Formed Silicon- Carbide, Abstr Pap Am Chem S 200 (1990) 80-Ched. [4] Y.M. Chiang, R.P. Messner, C.D. Terwilliger, D.R. Behrendt, Reaction-Formed Silicon- Carbide, Mat Sci Eng a-Struct 144 (1991) 63-74. [5] M. Singh, D.R. Behrendt, Microstructure and Mechanical-Properties of Reaction- Formed Silicon-Carbide (Rfsc) Ceramics, Mat Sci Eng a-Struct 187(2) (1994) 183-187. [6] M. Singh, D.R. Behrendt, Reactive Melt Infiltration of Silicon-Niobium Alloys in Microporous Carbons, J Mater Res 9(7) (1994) 1701-1708. [7] M. Singh, D.R. Behrendt, Reactive Melt Infiltration of Silicon-Molybdenum Alloys into Microporous Carbon Preforms, Mat Sci Eng a-Struct 194(2) (1995) 193-200. [8] M. Singh, T.A. Leonhardt, Microstructural characterization of reaction-formed silicon carbide ceramics, Mater Charact 35(4) (1995) 221-228. [9] A. Wolfenden, P.J. Rynn, M. Singh, Measurement of Elastic and Anelastic Properties of Reaction-Formed Silicon Carbide-Based Materials, J Mater Sci 30(21) (1995) 5502-5507. [10] M. Singh, R.M. Dickerson, Characterization of SiC fiber (SCS-6) reinforced-reaction- formed silicon carbide matrix composites, J Mater Res 11(3) (1996) 746-751. [11] T.X. Fan, S.K. Chow, Z. Di, Biomorphic mineralization: From biology to materials, Prog Mater Sci 54(5) (2009) 542-659. [12] T. Ota, M. Takahashi, T. Hibi, M. Ozawa, S. Suzuki, Y. Hikichi, H. Suzuki, Biomimetic process for producing SiC ''wood'', J Am Ceram Soc 78(12) (1995) 3409-3411. [13] C.E. Byrne, D.C. Nagle, Carbonized wood monoliths - Characterization, Carbon 35(2) (1997) 267-273. [14] C.E. Byrne, D.C. Nagle, Carbonization of wood for advanced materials applications, Carbon 35(2) (1997) 259-266. [15] C.E. Byrne, D.C. Nagle, Cellulose derived composites - A new method for materials processing, Mater Res Innov 1(3) (1997) 137-144. [16] P. Greil, T. Lifka, A. Kaindl, Biomorphic cellular silicon carbide ceramics from wood: II. Mechanical properties, J Eur Ceram Soc 18(14) (1998) 1975-1983. [17] P. Greil, T. Lifka, A. Kaindl, Biomorphic cellular silicon carbide ceramics from wood: I. Processing and microstructure, J Eur Ceram Soc 18(14) (1998) 1961-1973. [18] H. Sieber, C. Hoffmann, A. Kaindl, P. Greil, Biomorphic Cellular Ceramics, Adv Eng Mater 2(3) (2000) 105-109. [19] P. Greil, Biomorphous ceramics from lignocellulosics, J Eur Ceram Soc 21(2) (2001) 105-118. [20] M. Singh, Ecoceramics: Ceramics from wood, Adv Mater Process 160(3) (2002) 39-41. [21] M. Singh, J.A. Salem, Mechanical properties and microstructure of biomorphic silicon carbide ceramics fabricated from wood precursors, J Eur Ceram Soc 22(14-15) (2002) 2709-2717. [22] M. Singh, J. Martinez-Fernandez, A.R. de Arellano-Lopez, Environmentally conscious ceramics (ecoceramics) from natural wood precursors, Curr Opin Solid St M 7(3) (2003) 247-254. [23] H. Sieber, M. Singh, Microcellular ceramics from wood, Cellular Ceramics: Structure, Manufacturing, Properties and Applications (2005) 122-136. [24] A.R. de Arellano-Lopez, J. Martinez-Fernandez, P. Gonzalez, C. Dominguez, V. Fernandez-Quero, M. Singh, Biomorphic SiC: A new engineering ceramic material, Int J Appl Ceram Tec 1(1) (2004) 56-67. [25] A.P. Scniewind, Concise Encyclopedia of Wood and Wood-based Materials, Pergamon Press, New York, 1989. [26] A.J. Panshin, C. de Zeeuw, Textbook of Wood Technology, Mc Graw-Hill, New York, 1980. [27] L.J. Gibson, M.F. Ashby, Cellular Solids: Structure and Properties, Pergamon Press1988. [28] R.F. Evert, K. Esau, K. Esau, Esau's Plant anatomy : meristems, cells, and tissues of the plant body : their structure, function, and development, 3rd ed., Wiley-Interscience, Hoboken, N.J., 2006. [29] J.M. Dinwoodie, Institute of Metals., Wood : nature's cellular, polymeric, fibrecomposite, Institute of Metals, London ; Brookfield, VT, USA, 1989. [30] R.H. Atalla, J.M. Hackney, Structural Polysaccharides in the Molecular Architecture of Plant Cell Walls, from Algae to Hardwoods, Materials Research Society Symposium 255 (1992) 399-405. [31] H.P. Yang, R. Yan, H.P. Chen, D.H. Lee, C.G. Zheng, Characteristics of hemicellulose, cellulose and lignin pyrolysis, Fuel 86(12-13) (2007) 1781-1788. [32] A.R. de Arellano-Lopez, J. Martinez Fernandez, F.M. Varela Feria, R.E. Sepulveda, M.J. Lopez Robledo, J. Llorca, J.Y. Pastor, M. Presas, K.T. Faber, V.S. Kaul, K.E. Pappacena, T.E. Wilkes, Processing, microstructure and mechanical behavior of SiC-based ceramics VIA naturally derived scaffolds, Ceramics Engineering and Science Proceedings 27 (2006) 635- 650. [33] P. Gonzalez, J.P. Borrajo, J. Serra, S. Chiussi, B. Leon, J. Martinez Fernandez, F.M. Varela Feria, A.R. de Arellano-Lopez, A. De Carlos, A. Munoz, M. Lopez, M. Singh, A new generation of bio-derived ceramic materials for medical applications, J. Biom. Mater. Res. A (2008). [34] P. Greil, T. Lifka, A. Kaindl, Biomorphic Cellular Silicon Carbide Ceramics from Wood: I. Processing and Microstructure, J. Eur. Ceram. Soc. 18 (1998) 1961-1973. [35] J. Martinez-Fernandez, A.R. de Arellano-Lopez, F.M. Varela-Feria, M. Singh, Procedimiento para la Fabricación de Carburo de Silicio a partir de Precursores Vegetales, Spain, 2001. [36] J. Martinez-Fernandez, F.M. Varela-Feria, S. Lopez-Pombero, A.R. De Arellano-Lopez, M. Singh, Creep resistant biomorphic silicon-carbide based ceramics, Ceramics Engineering and Science Proceedings 22 (2001) 135-143. [37] A. Muñoz, J. Martinez Fernandez, M. Singh, High temperature compressive mechanical behavior of joined biomorphic silicon carbide ceramics, J Eur Ceram Soc 22(14-15) (2002) 2727-2733. [38] T. Ota, M. Takahashi, T. Hibi, M. Ozawa, S. Suzuki, Y. Hikichi, H. Suzuki, Biomimetic process for producing SiC 'wood', J. Am. Ceram. Soc. 78(12) (1995) 3409-3411. [39] F.M. Varela-Feria, J. Ramirez-Rico, A.R. de Arellano-Lopez, J. Martinez-Fernandez, M. Singh, Reaction-formation mechanisms and microstructure evolution of biomorphic SiC, J Mater Sci 43(3) (2008) 933-941. [40] F.M. Varela-Feria, J. Ramírez-Rico, J. Martínez-Fernández, A.R. De Arellano-López, M. Singh, Infiltration and reaction-formation mechanism and microstructural evolution of biomorphic SiC fabricated by Si-melt infiltration, Ceramic Transactions 177 (2006) 93- 101. [41] F.M. Varela Feria, Processing, Characterization and Mechanical Properties of Biomorphic Silicon Carbide, Condensed Matter Physics, Universidad de Sevilla, Sevilla, 2004. [42] R.C.E. Weast, Handbook of Chemistry and Physics, 55th ed., CRC Press, Cleveland, OH, 1974. [43] C. Zollfrank, H. Sieber, Microstructure evolution and reaction mechanism of biomorphous SiSiC ceramics, J Am Ceram Soc 88(1) (2005) 51-58. [44] C. Zollfrank, H. Sieber, Microstructure and phase morphology of wood derived biomorphous SiSiC-ceramics, J Eur Ceram Soc 24(2) (2004) 495-506. [45] M.H. Hon, R.F. Davis, Self-Diffusion of C-14 in Polycrystalline Beta-SiC, J Mater Sci 14(10) (1979) 2411-2421. [46] R. Pampuch, J. Bialoskorski, E. Walasek, Mechanism of Reactions in the Sii + Cf System and the Self-Propagating High-Temperature Synthesis of Silicon-Carbide, Ceram Int 13(1) (1987) 63-68. [47] R. Pampuch, E. Walasek, J. Bialoskorski, Reaction-Mechanism in Carbon Liquid Silicon Systems at Elevated-Temperatures, Ceram Int 12(2) (1986) 99-106. [48] M.H. Hon, R.F. Davis, D.E. Newbury, Self-Diffusion of Si-30 in Polycrystalline Beta-SiC, J Mater Sci 15(8) (1980) 2073-2080. [49] C. Torres-Raya, D. Hernandez-Maldonado, J. Ramirez-Rico, C. Garcia-Ganan, A.R. de Arellano-Lopez, J. Martinez-Fernandez, Fabrication, chemical etching, and compressive strength of porous biomimetic SiC for medical implants, J Mater Res 23(12) (2008) 3247- 3254. [50] J. Ramirez-Rico, C. Torres-Raya, D. Hernandez-Maldonado, C. Garcia-Gañan, J. Martinez-Fernandez, A.R. De Arellano-López, Porous biomorphic SiC for medical implants processed from natural and artificial precursors, Ceramic Engineering and Science Proceedings, 2010, pp. 203-214. [51] K.E. Pappacena, S.P. Gentry, T.E. Wilkes, M.T. Johnson, S. Xie, A. Davis, K.T. Faber, Effect of pyrolyzation temperature on wood-derived carbon and silicon carbide, J Eur Ceram Soc 29(14) (2009) 3069-3077. [52] T.E. Wilkes, S.R. Stock, F. De Carlo, X. Xiao, K.T. Faber, X-ray micro-computed tomography of beech wood and biomorphic C, SiC and Al/SiC composites, Philos Mag 89(17) (2009) 1373-1389. [53] E. Vogli, H. Sieber, P. Greil, Biomorphic SiC-ceramic prepared by Si-vapor phase infiltration of wood, J Eur Ceram Soc 22(14-15) (2002) 2663-2668. [54] 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(6) (2001) 1236-1240. [55] J.M. Qian, J.P. Wang, G.Y. Hou, G.J. Qiao, Z.H. Jin, Preparation and characterization of biomorphic SiC hollow fibers from wood by chemical vapor infiltration, Scripta Mater 53(12) (2005) 1363-1368. [56] J.W. Kim, S.W. Myoung, H.C. Kim, J.H. Lee, Y.G. Jung, C.Y. Jo, Synthesis of SiC microtubes with radial morphology using biomorphic carbon template, Mat Sci Eng a-Struct 434(1-2) (2006) 171-177. [57] D.A. Streitwieser, N. Popovska, H. Gerhard, Optimization of the ceramization process for the production of three-dimensional biomorphic porous SiC ceramics by chemical vapor infiltration (CVI), J Eur Ceram Soc 26(12) (2006) 2381-2387. [58] 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(6) (2005) 817-828. [59] H. Ghanem, E. Alkhateeb, H. Gerhard, N. Popovska, Oxidation behavior of silicon carbide based biomorphic ceramics prepared by chemical vapor infiltration and reaction technique, Ceram Int 35(7) (2009) 2767-2774. [60] P. Greil, E. Vogli, T. Fey, A. Bezold, N. Popovska, H. Gerhard, H. Sieber, Effect of microstructure on the fracture behavior of biomorphous silicon carbide ceramics, J Eur Ceram Soc 22(14-15) (2002) 2697-2707. [61] H. Sieber, C. Zollfrank, N. Popovska, D. Almeida, H. Gerhard, Gas phase processing of porous, biomorphous TiC-ceramics, Euro Ceramics Viii, Pts 1-3 264-268 (2004) 2227- 2230. [62] N. Popovska, D. Almeida-Streitwieser, C. Xu, H. Gerhard, H. Sieber, Kinetic analysis of the processing of porous biomorphic titanium carbide ceramics by chemical vapor infiltration, Chem Vapor Depos 11(3) (2005) 153-158. FIGURE 2 FIGURE 3 FIGURE 4 FIGURE 5 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Cork Sapelly Beech White Eucalyptus Oak Bamboo Red Eucalyptus Balsa Redwood White Pine Basswood Red oak Hard Maple Lignum vitae Wood density (g cm-3) FIGURE 6 FIGURE 7 FIGURE 8 FIGURE 9 01x1042x1043x1044x1045x1046x104 0 20 40 60 80 Si C SiC Time [s] FIGURE 10 FIGURE 11 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 10-3 10-2 10-1 100 bioSiC axial bioSiC r/t RF SiC RB SiC /SiC FIGURE 18 0.0 0.2 0.4 0.6 0.8 1.0 10-2 10-1 100 RBSiC bioSiC/Si (radial) RFSiC bioSiC/Si (axial) FIGURE 19 0.0 0.1 0.2 0.3 0.4 0.5 10-8 10-7 10-6 1175 ºC 1600 ºC Strain (%) FIGURE 20 FIGURE 21