Conversion of cellulose materials into nanostructured ceramics by biomineralization

Conversion of cellulose materials into nanostructured ceramics by biomineralization

(Parte 1 de 3)

Conversion of cellulose materials into nanostructured ceramics by biomineralization

Yongsoon Shin Æ Gregory J. Exarhos

Received: 12 September 2006/Accepted: 1 November 2006/Published online: 6 December 2006 Springer Science+Business Media B.V. 2006

Abstract Synthesis of hierarchically ordered silica materials having ordered wood cellular structures has been demonstrated through in-situ mineralization of wood by means of surfactantdirected mineralization in solutions of different pH. At low pH, silicic acid penetrates the buried interfaces of the wood cellular structure without clogging the pores to subsequently ‘‘molecularly paint’’ the interfaces thereby forming a positive replica following calcinations. At high pH, the hydrolyzed silica rapidly condenses to fill the open cells and pits within the structure resulting in a negative replica of the structure. Surfactanttemplated mineralization in acid solutions leads to the formation of micelles that hexagonally pack at the wood interfaces preserving structural integrity while integrating hexagonally ordered nanoporosity into the structure of the cell walls following thermal treatment in air. The carbothermal reduction of mineralized wood with silica at high temperature produces biomorphic silicon carbide (SiC) materials, which are typical aggregations of b-SiC nanoparticles. To understand the roles of each component (lignin, crystalline cellulose, amorphous cellulose) com- prising the natural biotemplates in the transformation to SiC rods, three different cellulose precursors including unbleached and bleached pulp, and cellulose nanocrystals have been utilized. Lignin in unbleached pulp blocked homogeneous penetration of silica into the pores between cellulose fibers resulting in non-uniform SiC fibers containing thick silica layers. Bleached pulp produced uniform SiC rods with camelback structures (80 nm in diameter; ~50 lm in length), indicating that more silica infiltrates into the amorphous constituent of cellulose to form chunky rather than straight rod structures. The cellulose nanocrystal (CNXL) material produced clean and uniform SiC nanowires (70 nm in diameter; >100 lm in length) without the camelback structure.

Keywords Biomimetic Cellulose Nanocrystal Silica Silicon carbide Silicon oxide


Synthesis of hierarchical porous materials from biological templates continues to generate significant interest because biological structures exhibit excellent strength at low density, high stiffness and elasticity, and high damage tolerance integrated into these structures through the evolutionary process (Sarikaya and Aksay 1993).

Y. Shin (&) G. J. Exarhos Chemical Science Division, Pacific Northwest National Laboratory, 902 Battelle Blvd, P.O. Box 9, MS K2-4, Richland, WA 99354, USA e-mail:

Cellulose (2007) 14:269–279 DOI 10.1007/s10570-006-9101-0

A particular research interest is the use of natural biological materials as templates to construct novel hierarchical inorganic materials such as oxides and carbides. This is an emerging research area due to the unique and sophisticated microstructures (Davis et al. 2001), which evolve and have been optimized by nature through ages. Compared to artificial templates, biological materials exhibit hierarchical structure and are abundant, complex, renewable, and environmentally benign. To date many different types of biological materials including diatoms (Anderson et al. 2000), bacteria (Davis et al. 1997; Zhang et al. 2000), pollen (Hall et al. 2003), cornstarch (Zhang et al. 2002), chitin (Ogasawara et al. 2000), living cells (Chia et al. 2000), and wood (Shin et al. 2001) have been templated in order to mimic their specific structures. The use of biotemplates generates multimodal pore systems, which contain adjustable macropores and tunable, interconnected mesopore types of different size (2–50 nm) in the macropore walls (Shin et al. 2001; Holland et al. 1999). This processing likely will overcome the main drawback of surfactant-mixing approaches, which often fail because the small surfactant molecule just acts as a co-template for the larger one, leading to a monomodal pore size distribution, rather than the desired bimodal pore distribution (Karlsson et al. 1999).

Wood is a natural cellular composite material that is highly anisotropic and requires minimal treatment for mineralization. Wood tissues are comprised of interconnected cells (tracheids) and open spaces (lumens). The cell walls usually contain a thin outer layer called the primary wall and a thicker inner layer called the secondary wall. These cells are glued together by an intercellular layer (or middle lamella) and are connected by openings of different shapes and sizes. These openings are called pits (bordered pits or simple pits) and are the communication channels between the cells. Pits mostly contain a pit chamber, a pit aperture, and a membrane (the central portion of the membrane is called torus). Two common wood types were used in this study. Poplar wood (Populus spp.), representative of hard wood species, contains close-packed tubular cells of different sizes, ranging from a few micrometers to about one hundred micrometers. Pine (Pinus) wood, representative of soft wood species, contains close-packed rectangular cells of roughly the same size.

In this paper, we describe the preparation of multimodal porous silica materials based upon wood cellular structures that are achieved by controlling precursor solution pH. To generate a positive replica of the structure, slow condensation of silica precursors is required. Subsequent thermal treatment of silica/wood composites at high temperature and in the absence of air produces biomorphic SiC ceramics (Shin et al. 2005). In addition, unbleached, bleached pulp, and CNXLs (Samir et al. 2005) have been used to investigate the role of cellulose and lignin in the synthesis of SiC rods. In the case of unbleached pulp, lignin blocks penetration of silicic acid into the cell walls and forms non-uniform SiC rods with thick silica layers. For bleached pulp, silicic acid deposits more in amorphous regions of the cellulose and generates uniform SiC rods with camelback structures. Homogeneous deposits of silicic acid on CNXL produce uniform SiC rods without any surface inhomogeneity.

Materials and methods Materials and chemicals

Poplar and pine wood were obtained from Lowe Hardware store (Kennewick, WA) and used without further treatment. Tetraethylorthosilicate (TEOS), sodium silicate (27.0 wt%), cetyltrimethyl bromide (CTAB), Hydrochloric acid (HCl), and ethyl alcohol (EtOH) were obtained from Aldrich Chemical Co., and cetyltrimethylammonium chloride (CTAC) was purchased from Kodak Co. CNXL was prepared by a previously published procedure (Edgar and Gray 2003).

Sample preparation

Biomorphic SiO2

For preparations involving acid-catalyzed mineralization, the molar ratio of the surfactant sol-gel solution, TEOS:CTAC:HCl:EtOH:H2O, was

1.0:0.24:0.05:4.0:2.0. A relatively high-volume ratio of EtOH to H2O was used to prevent silica from precipitating from the solution. In a typical proce- dure, the surfactant solution was prepared by adding 3.686 g CTAC to a mixed solution containing 8.830 g EtOH and 1.73 g of HCl solution (1.39 M).Then,10.0 gTEOS was added.Afterthe solution was cooled to room temperature, several pieces of wood on the order of 1.0 cubiccentimeter were soaked in the solution and kept at 40–60 C for 2 days in an unsealed polypropylene container. The wood samples were then removed from solution and similarly treated in a new solution for another 2 days at the same temperature. Finally, samples were removed from the sol mixture, air-dried, and calcined at 550 C for 6 h in air. Under neutral conditions, the molar ratio of

TEOS:CTAC:EtOH:H2O was 1.0:0.24:3.0:2.0. A surfactant solution was prepared by adding 0.05 g of CTAC to10 ml ofTris–HCl (50 mM, pH ~ 7.0) buffer and 15 ml of an EtOH mixture. To the clear solution 7.08 g of TEOS was added. Several wood samples were treated by the solution and kept at 40 C overnight. After 18 h, silicate started to precipitate. This process was repeated 3 times andthewoodwascalcinedat550 C.InthehighpH processing solution, 27.0% sodium silicate solution was used as the silicate precursor; Si:CTAC:-

NaOH:H2O = 1.0:0.5:2.0:120. All the processing steps were kept the same except samples were washed prior to calcination to remove adsorbed sodium ions.

Biomorphic SiC

Biomorphic SiC materials were prepared from dried wood/silica composites by thermal treatment in the absence of oxygen. Samples were placed in an alumina boat and then introduced into a horizontal alumina tube furnace. The temperature was increased to 1400 Ca ta4 C/min rate. This was followed by heating for 2 h in Ar. The samples were subsequently cooled to room temperature in Ar.

SiC nanorods from pulp

Pulp mats of unbleached and bleached pulps were soaked separately at room temperature for 18 h to acidic silicate solutions prepared from TEOS/

HCl/H2O (molar ratio = 1.0:1.58:46.24). Pulp mats were removed and air-dried overnight. The dried silica/pulp composite was placed in an alumina boat and positioned in the center of a horizontal alumina tube furnace. The temperature was increased to 1400 Ca t5 C min–1 rate and held for 2 h under an Ar flow. The furnace was cooled to room temperature at a rate of 5 C min–1.

The preparation of CNXL

In a typical preparation of 2.0% CNXL from Whatman No.1 filter paper, the filter paper was first ground in a house blender (Proctor Silex). The cellulose was hydrolyzed with 64% sulfuric acid at 45 C. About 10 g of cellulose was treated with 150 ml of acid. After hydrolyzing for 1 h, the suspension was diluted with 10-fold to stop the reaction. The suspension was then washed repeatedly by diluting with water and centrifuging until the pH of the supernatant was about 1. Next, the sample was dialyzed against water for 3–4 days and the colloidal solution was repeatedly sonication (Branson Sonifier, Model 2210) for 7 min interval for a total of 35 min before the suspension was allowed to stand over a mixed bed resin for 2 days. The suspension was then filtered through hardened ashless filter paper (Whatman 541) and a dilute suspension of CNXLs was finally filtered through a 0.45 lm syringe filter (2.0%, pH 2.7).

SiC nanorods from CNXL

To a 2% CNXL solution (20 ml), 10 ml of acidic silicate solution (same as above) was added. After hand-shaking, a white composite precipitates from the homogeneous solution. Centrifugation at 8000 rpm for 10 min gives a free-standing film after air-drying overnight. The preparation of SiC using CNXL/silica composites was identical to the procedure used in mineralizing the pulp.


XRD patterns were obtained from a Philips X’Pert MPD X-ray Powder Diffractometer with Cu anode (Model PW3040/0), running at 40 kV and 50 mA, and scanning from 1 through 75 at

0.04 /s. The N2 adsorption isotherms were measured with a Quantachrome Autosorb-6 and a Micromeritics ASAP 2010 adsorption analyzer independently. Before analysis, each sample was degassed at 100–200 C overnight under vacuum of about 10–3 Torr. Brunauer–Emmet–Teller (BET) surface areas, pore volumes, and pore size distributions were determined from N2 adsorption isotherms. Adsorption data over the relative pressure range of P/P0 = 0.05–0.15 were used to calculate the BET surface area. Pore size distri- butions were calculated from adsorption branches of N2 isotherms using the BJH method. Scanning electron microscope (SEM) measurements were carried out with a JEOL (JSEM 633F) instrument operating at an acceleration voltage of 5 kV. SEM samples were sputter-coated with about 5 nm of Au. TEM images were obtained from thin areas of the particles on a JEOL 1200 instrument at 120 kV. TEM images of CNXL were obtained by staining using 1.0% polytungstic acid. 13C and 29Si solid-state NMR spectra were obtained from a Chemagnetics spectrometer (300 MHz). The samples were loaded into 7 m zirconia PENCIL rotors and spun at 3–4 kHz.

Results and discussion

Biomorphic SiO2

Our approach to synthesize silica-based systems with multidimensional pores is to mineralize wood cellular structures using a surfactant templated sol-gel solution (Fig. 1). Heating at low temperature (40–60 C) enhances lignin leaching and penetration of silicate solution into the wood cellular networks. To avoid instant precipitation of silica/surfactant composites, low pH solutions were heated at 60 C and high pH systems were heated at 40 C with shaking. The surfactanttemplated sol-gel solution was replaced to insure complete reaction and avoid precipitation. The removal of residual organic by thermal treatment at 550 C in air produced white silica monoliths with wood cellular structures.

SEM images of the calcined samples mineralized in acid revealed that they retain their original cellular structures, which contain intact cells (Fig. 2a and d), cell walls and pit structures along the cell walls (Fig. 2b and f), ray pitting of poplar (Fig. 2c), rectangular-type cells with fine fibrous arrays (Fig. 2e), and bordered pits (Fig. 2g). In contrast, hierarchical silica samples prepared under neutral solution conditions show formation of negative wood replicas after calcination (Fig. 3). Wood cavities are filled with precipitated ceramic and the negative images of pits (Fig. 3a, b, and d) are seen; negative vessel pitting on the cell walls (Fig. 3c), and negative donut-types of bordered pits (Fig. 3e) were also observed. These negative structural replicas have been formed due to rapid condensation of silica precursors. Under acidic conditions, the hydrolyzed silica precursor (silicic acid) penetrates through cell walls since under these conditions condensation is slow. Occasionally, positive replicas have been isolated under neutral conditions when large portions of EtOH were used. Calcination of mineralized wood samples prepared under basic pH conditions leads to negative replicas as well (Fig. 4). Filled tube-types (Fig. 4a), negative vessel pitting (Fig. 4b), retangular type cells (Fig. 4c), and negative donut-types of bodered pits (Fig. 4d) were also observed.

Under all processing conditions, we observed significant lignin leaching, which enhanced precipitation of silica. EtOH, which is used as a cosolvent, not only keeps silica from precipitating, but also promotes leaching of non-crystalline lignin and hemicellulose. Under acidic conditions, at 60 C lignin was leached from the structure much quicker than under neutral conditions. More than 95% lignin was removed after 5 cycles. XRD patterns of as-synthesized and calcined samples for all processing conditions showed that the crystallinity of wood cellular structures was nearly preserved even after acid or base treatment (Fig. 5). The surfactant-templated liquid crystalline solution produced hexagonally ordered nanoporous networks. Figure 6 shows TEM images of calcined poplar samples prepared under acidic and basic conditions, indicating hexagonal ordering with 2 nm diameter pores. The BET measurement of surfactant-templated samples showed much higher surface area (up to 700 m2/g) than those of silica samples prepared

without the surfactant (£50 m2/g). 13C and 29Si MAS NMR spectra of silica/wood composite samples are presented in Fig. 7. The crystallinity of the wood saw dust significantly increased with increased leaching of lignin. The poplar tissue comprised of crystalline celluloses (Atalla et al.

1980) and lignins (Ono and Sudo 1989) showed dominant cellulose peaks at 67.5, 72.5, 84.0, 8.8, and 105.2 ppm. Lignin features were observed to be broad at 136.9, 154.0, 172.3, and 194.7 ppm, and overlapped at 56.6 and 84.5 ppm. After a 2 days treatment in acid/EtOH, the lignin content

TEOS or Sodium silicate

1. Dry 2. calcination

1. wash/dry 2. calcination


Fig. 1 Proposed mechanism for the formation of mineralized silica frameworks using wood cellular templates. (a) Mineralization of lignin-washed swollen spaces with mesostructured silica sol at low pH, (b) Infiltration and precipitation of silicate-surfactant mesophase in wood cellular structures by rapid condensation at high pH. A parallelgram on the left indicates wood cellular structures (20–50 lm), and a dot indicates silicate sol b c d e fg

100µm 50µm 20µm 100µm 10µm ab dce

was significantly lower as indicated by a decrease in peaks at 56.6 and 84.5 ppm. Crystalline cellulose peaks remained the same. Concurrently, peaks corresponding to CTAC were dominant in the low chemical shift region (14.6, 18.3, 23.3, 30.4, and 53.7 ppm). However, basic treatment of the silica/poplar composite for 2 days showed residual lignin. This caused the infiltration of a small amount of CTAC into the tissue network. Figure 7b shows the single pulse 29Si NMR spectrum of the silica poplar composite, indicating that the acidic condition promotes hydrolysis and allows polymerization to cyclic trimers and tetramers, but inhibits condensation (Brinker and Scherer 1990). However, under neutral conditions, a very high degree of condensation is evident with 100% polymerization seen at long times. The condensation reaction under acidic conditions is essentially irreversible because hydrolysis of the siloxane bond (the dissolution

20µm 100µm 10µm10µm ab cd

Fig. 4. SEM images of poplar (a, b) and pine (d, e) samples prepared under basic conditions after calcination at 550 C for 6h

2 theta2 theta bFig. 5 XRD traces of poplar/silica composites prepared under (a) acidic and (b) basic conditions as a function of sequential treatment: (1) poplar untreated, (2) 1 cycle, (3) 2 cycles, (4) 3 cycles, and (5) after calcinations.

5mn 20nm 20nm abFig. 6. TEM images of poplar samples prepared under (a) acidic and (b) basic conditions after calcination

(Parte 1 de 3)