Physical characteristics of carbon materials derived from pyrolysed vascular plants

ARTICLE IN PRESS

Biomass and Bioenergy 30 (2006) 166–176 www.elsevier.com/locate/biombioe

Physical characteristics of carbon materials derived from pyrolysed vascular plants Marta Krzesin´skaa,, Barbara Pilawaa, S"awomira Pusza, Jonathan Ngb a

Institute of Coal Chemistry, Polish Academy of Sciences, Sowin´skiego 5, 44-121 Gliwice, Poland Department of Chemical Engineering, McMaster University, Hamilton, Ont., Canada L8S 4L7

b

Received 7 December 2004; accepted 17 November 2005

Abstract The purpose of this study was to develop new monolithic porous carbon materials from vascular plants using highly controlled pyrolysis. Perennial plants belonging to the grass family Poaceae such as bamboo (Bambusa vulgaris) and to the family Agavaceae such as yucca (Yucca flaccida) characterized by a homogeneous profile and homogenous vessel distribution were selected for the study. They were heat-treated at temperatures 550 and 950 1C in a nitrogen atmosphere to produce a crack-free monolithic porous carbon materials for which physical characteristics such as density, porosity, yield and dimensional changes were determined. The EPR spectroscopy, ultrasonic technique and optical microscopy were applied for further characterization. All samples studied demonstrated a reduction in apparent density and dimensions due to carbonisation. It was found that similarly as in the case of hardwoods, the higher the carbonisation temperature, the greater the dimensional shrinkage. The greatest changes were observed in ‘‘transverse’’ to plant fibres directions, i.e., for radial and tangential. It was found that the dimensional changes under heattreatment exhibited transverse isotropy. Carbonised plants were characterised by elastic moduli almost independent of apparent density in contrast to elasticity of precursors. Elastic moduli of samples carbonised to 950 1C were higher than those heat-treated to 550 1C. Results showed that materials carbonised at higher temperature were more stiff—more ordered in structure. Microscopic observations showed that during heat-treatment of yucca and bamboo, their tissue structure remained unaltered. There was the increase in order of aromatic layers in the walls of fibres expressed by the increase of optical reflectance values through the carbonisation process. It was found that heating plants to 950 1C quenched paramagnetic centres in carbonised samples. This effect resulted from an increase of multiring aromatic units in the samples. The observed lack of saturation of the EPR spectra evidenced that during slow pyrolysis defects were not created. Carbonised woody stems of perennials studied were found as very porous, but stiff materials, which can be excellent precursors (as skeleton) for new eco-materials, e.g., for wood-ceramics. r 2005 Elsevier Ltd. All rights reserved. Keywords: Vascular plants; Biotemplating; Carbonisation; Acoustical properties; EPR parameters; Optical reflectance

1. Introduction Biomass feedstocks including wood or agricultural residues and byproducts (e.g., wood chips, sawdust, tree prunings, corn stover, bagasse and rice husks) or dedicated energy crops (e.g., fast-growing trees, shrubs and grasses) usually applied as biomass fuels are of significant interest, because they form the world’s third largest primary energy resource after coal and oil [1,2]. However, some of these Corresponding author. Tel.: +48 32 238 07 94; fax +48 32 231 28 31.

E-mail address: [email protected] (M. Krzesin´ska). 0961-9534/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biombioe.2005.11.009

plants can also be precursors for the materials of various applications, especially for removal of colours, odours, organic and inorganic pollutants from industrial process to waste effluents. The microstructures of natural-grown plants are characterized by a unidirectional, open porous system on the micrometer level which provides a transportation path for water in the living plant and yields anisotropic structural and mechanical properties [3]. The microstructural features of naturally grown plants are an attractive template for the design of novel porous ceramics. By using various monolithic wood specimens (tulip poplar, red oak white

ARTICLE IN PRESS M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

oak, lignum vitae, sugar maple, basswood, white pine, redwood and balsa) and highly controlled slow pyrolysis, Byrne and Nagle [4,5] have demonstrated for the first time a carbonisation process to convert wood materials into crack-free, porous, monolithic charcoals, so-called ‘‘carbonised woods’’ which were readily shaped by conventional methods. In spite of pure solid woods, wood-based materials were also intensively studied during the last years. Basic properties of woods or wood-based materials carbonised in neutral atmosphere are still being investigated [6–14]. Slow-carbonised cellulosic materials (wood blocks or wood-based fiberboards) were found as materials of electrical resistivity varied by several orders of magnitude for the range of heat-treatment temperature (HTT) between 600 and 1400 1C [6–8]. Volumetric shrinkage at HTT over 600 1C suggested that turbostratic crystallites were drawn closer together as the low-density disordered carbon was converted into high-density graphene sheets [9,14]. At approximately 900 1C, the large graphene sheets and the large turbostratic crystallites significantly impinged on each other. This increasing impingement of conductive phases with increasing carbonisation temperature may cause nonmetal-metal transition [9]. Klose and Schinkel [11] developed a mathematical model based on the population balance and mass balance to investigate the change in pore size distribution during the pyrolysis of woods. Another method of carbonisation was proposed by Kurosaki et al. [12]. They prepared carbonised wood by flash heating at 800 1C for 1 h that, in contrast to conventional slow heating (4 1C/min to 800 1C for 1 h), exhibited pores that were surrounded by aggregates of carbon structures. The carbon structures were built up of clearly visible graphene layers that were often curved and overlap each other in a disordered manner. Preparation of monolithic or granular microporous activated carbon adsorbents from precursors such as wood or agricultural waste materials (e.g., coconut shells) needs physical or chemical activation [15–18]. Cited above papers have focused on the effect of the preparation parameters (kind of the activation agent, temperature, time of agent action, etc.) on the physical properties of the activated carbons (surface area, pore size distribution, homogeneity within the monolith) as well as, on adsorption yield and adsorption kinetics. In the case of monolithic wood precursors chemical activation appeared to be a better method than physical activation to prepare activated carbon, but the choice of an appropriate wood is of paramount importance to obtain homogenous porosity and high surface area. In the first papers [4,5,19], Byrne and Nagle have already suggested that net-shape polymer, ceramic and carbon composites can be produced using wood as a precursor. The controlled thermal decomposition of wood allows the formation of a monolithic carbon template without the presence of macro-cracks. Ozao et al. [20], as well as Hirose et al. [21] developed a new eco-material, wood-ceramic—

167

a new porous carbon material obtained by carbonisation of wood or woody material impregnated with phenolic resin or liquefied wood. Another group of porous composites with carbonised monolithic softwood or wood-based templates produced using an infiltration technique of reactive Si-vapor (SiC-ceramics) or zirconium-oxychloride (YSZ ceramics) were proposed by Fey et al. [22] and Rambo et al. [3]. Recently, a new type of porous ceramic with wood-like microstructures was prepared by mimicking silicified wood by Mizutani et al. [23]. Titania, alumina and zirconia ceramic woods were produced by the sol–gel method using various natural hardwoods and softwoods as templates. Their microstructure was found the same structure as that of the raw wood with the pore sizes corresponding to those of the original wood. All these highly porous ceramics, especially the YSZ ceramics, are expected to be suitable for specific applications including sensors, filters, catalysts carrier, thermal isolation in hightemperature processes or electrolyte material for the solid oxide fuel cell. All studies mentioned above involved monolithic specimens, generally cut from hardwoods, rarely from softwoods. It would be interesting to prepare carbon materials from plants of a different structure than that of hardwoods or softwoods. We decided to prepare and study carbonised monoliths from woody stems of plants such as bamboo and yucca. These plants belong to the category of perennial plants. They exhibit no branches or seasonal rings and are characterized by a homogenous profile and vessel distribution. The homogeneity of these plants can be important because of possible applications. Yucca and bamboo differ in terms of the stem cross-section. Yucca stem is characterized by quasi-cylinder shape, while bamboo stem is a collar with outside diameter varying within a wide range. High-speed growth of selected plants (especially of bamboo) is very advantageous in comparison with the growth of trees usually used for hardwood or softwood. We chose two limit temperatures of heat-treatment, i.e., 550 and 950 1C. A choice of these temperatures was not accidental. As it was reported in early works of Byrne and Nagle [4,5] the physical parameters were sharply changed in the temperature range of about 500
1000 1C. For the HTT over this range the magnitudes of parameters studied were almost constant. The aim of the investigation presented in this paper was to produce under slow, highly controlled pyrolysis monolithic carbon materials from stems of perennials such as yucca and bamboo, and characterise their physical structure. 2. Experimental section 2.1. Precursors Perennial plants with woody stems such as bamboo (Bambusa vulgaris) and yucca (Yucca flaccida) were selected for the study.

ARTICLE IN PRESS 168

M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

Bamboos are a group of woody perennial evergreen plants in the grass family Poaceae, subfamily Bambusoideae [24]. Some of its members are giants, forming by far the largest members of the grass family. The stems, or ‘‘culms’’, can range in height from a few centimetres to 40 m, with stem diameters ranging from 1 mm to 30 cm. In very general terms bamboo consists of 50–70% hemicellulose, 30% pentosans, and 20–25% lignin. Bamboo culms have a number of important chemical and anatomical differences from hardwoods and softwoods. The ray cells in hardwoods and softwoods are linked to form a radial transport system. These structures are absent in bamboo where there are no cells to facilitate an easy movement of liquids in the radial direction. The penetration of liquids into the culm takes place through the vessels in the axial direction, from end to end. Bamboo forms a very hard wood therefore it will be ideally placed to become a principal engineering and construction material for the twenty first century and beyond. The Yuccas are a genus of 40–50 species of perennials, shrubs and trees, notable for their rosettes of tough, swordshaped leaves and large clusters of white or whitish flowers [24]. Yucca is a flowering plant in the family Agavaceae (division Magnoliophyta). It belongs to the vascular plants that have specialized cells for conducting water and sap within their tissues. Yucca is not used as a building material in contrast to bamboo. Y. flaccida is the evergreen shrub.

cuboids. Description of shapes and dimensions of samples used are shown in Table 1 and Fig. 1. Slow carbonisation of block samples was performed in a furnace fitted with a quartz pipe under a gas flow of nitrogen (about 0.1 l/min). The carbonisation was followed until the maximum carbonisation temperature was achieved: 550 and 950 1C. The rate of heating was constant for these final temperatures, equal to 3 1C/min. Samples were held for 1 h at desired temperature. Reduction in dimension (RD) along axial, radial and tangential directions in a sample, caused by heat-treatment, was calculated with an equation: RDð%Þ ¼ ðDraw
Dcarb Þ=Draw  100,

(1)

where Draw and Dcarb denote dimensions of raw and carbonised samples, respectively. Carbonisation weight yield Yd was calculated from weight of a sample before and after carbonisation using a relation: Ydð%Þ ¼ ðwcarb =wraw Þ  100.

(2)

Monolithic (cuboid or cylinder) samples were used for the ultrasonic, densitometric and optical studies, while for the EPR studies samples were ground to a powder.

2.2. Preparation of samples Monolithic pieces of woody stems from selected precursors were used in this experiment. Two series of bamboo samples were prepared from a larger and smaller diameter shoot. Quasi-cuboids with two mutually perpendicular planes were cut from the larger diameter bamboo approximately 3 cm  1 cm, while rings of approximately 2 cm in diameter and 3 cm in length were cut from the smaller diameter bamboo shoot. Similarly, yucca stems were cut into cylinders of approximately 2 cm in diameter and 3 cm in length. Some samples were cut into quasi-

(a)

(b)

(c)

Fig. 1. Schematic drawing of the shapes of samples studied: (a) quasicuboid from a thick bamboo (series B1); (b) collar from a thin bamboo (series B2) and (c) cut of cylinder from yucca (series Y1-2). The arrow denotes of fibres orientation.

Table 1 Description of samples studied Initial material

Shape of samples

Dimensions of samples (mm)

Direction of ultrasonic studies

Thick bamboo of diameter 43.5 mm (series B1)

Quasi-cuboids

xaxial ¼ {30.8–40.6} xtang ¼ {11.6–16.7} xradial ¼ 5.6–5.7

Axial Tangential Radial

Thin bamboo of diameter 18.8 mm (series B2)

Collars

xaxial ¼ {26.8–30.5} Outside diameter ¼ 18.3–18.8 Wall thickness 3.1–3.3

Axial

Yucca of diameter 20.7 mm (series Y1-2)

Cut off cylinders with two pairs of flat surfaces parallel or perpendicular to axial direction

xaxial ¼ {28.8–32.0} xradial ¼ {11.3–14.3}

Axial Radial

ARTICLE IN PRESS M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

2.3. Porosity

dynamic elastic modulus and r the density of the material. The dynamic elastic moduli were determined using ultrasonic velocity measurements. The velocity of longitudinal ultrasonic waves at a frequency of 100 kHz was measured along parallel (axial direction) and perpendicular (radial or tangential directions) to fibre direction, using an ultrasonic tester (Tester CT1, UNIPAN-ULTRASONIC, Poland). Details of an ultrasonic method are described in our previously published papers [26–28]. Directions in a sample, along which the ultrasonic measurements were made, are shown in Fig. 2. Dynamic elastic moduli were determined from apparent density and ultrasonic velocity using Eq. (4). Elastic anisotropy was calculated from a relation Eaxial /Eradial. An isotropic structure is described by an elastic anisotropy parameter equal to 1. Materials with Eaxial /Eradial41 are mechanically anisotropic. The higher the value of Eaxial /Eradial, the larger the mechanical anisotropy. Ultrasonic velocities in both parallel and perpendicular to plant fibers directions were measured before and after carbonisation.

Both apparent and true densities were measured to determine the bulk porosity of samples. The true density was measured using a helium gas displacement pycnometer type 1305 Micromeriticss. The porosity of carbonised samples was calculated using an expression: Pð%Þ ¼ ððrtrue
rapp Þ=ðrtrue ÞÞ  100,

(3)

where P is a porosity, r true and r app are true and apparent densities of a sample. 2.4. Ultrasonic velocity and dynamic elastic modulus Scattering of ultrasonic waves at frequencies up to about 20 MHz on pores as defects in studied materials, does not occur for above frequencies. Thus, samples of yucca and bamboo may be treated as homogeneous materials in this frequency range and the following equation can be applied for the determination of dynamic elastic moduli [25]: E ¼ rv2 ,

169

(4)

where v is the velocity, with which a stress wave is propagated through a homogeneous material, E the

2.5. EPR studies The EPR spectra of the samples were measured using an X-band (9.3 GHz) electron paramagnetic resonance spectrometer (RADIOPAN, Poland) with a magnetic modulation frequency of 100 kHz, at low microwave power (0.7 mW) to avoid signal saturation. Microwave frequency was measured with MCM 101 recorder. Ultramarine was used as the reference for concentration of paramagnetic centres. EPR spectra of the studied samples were asymmetric lines. The parameters of the EPR, namely linewidth DBpp, g-factor and integral intensity were calculated [29,30]. Total concentration of paramagnetic centres in the studied samples were also determined.

Axial

Tangential

Radial

2.6. Optical parameters Optical texture and reflectance values of yucca and bamboo were determined with a reflected light microscope

Fig. 2. Principle directions used for describing physical and mechanical properties of plants, studied in the shape of cuboid.

Table 2 Averaged values of ultrasonic velocity, porosity and the carbonisation weight yield Sample

Direction of measurement

Raw

Carbonised to 550 1C

Carbonised to 950 1C

Velocity (m/s)

P (%)

Velocity (m/s)

Yd (%)

P (%)

Velocity (m/s)

Yd (%)

P (%)

Axial Radial Tangential

4176 1499 1850

27.8

2343 726 1813

30.0

85.3

3077 655 2191

26.8

87.5

Thin bamboo (series B2)

Axial

4436

17.7

2274

31.8

74.4

3386

28.1

65.3

Yucca (series Y1-2)

Axial Radial

3074 1064

47.1

2431 929

31.3

86.6

3581 1236

26

96

Thick bamboo (series B1)

ARTICLE IN PRESS M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

170

Axioskop MPM 200 (Opton-Zeiss, Germany), using monochromatic linearly polarized light of l ¼ 546 nm, with objective 8  , 16  in air and 50  in immersion oil. Optical reflectance R was measured in the direction perpendicular to the length of a stem, in several (15–25) points, each in various locations on the sample. Reflectance is a function of refractive and absorption indices of the substance which is considered and a medium in which the measurement is made: R ¼ ½ðn
n0 Þ2 þ n2 k2 =½ðn þ n0 Þ2 þ n2 k2 ,

where R is the reflectance of substance, n the refractive index of substance, n0 the refractive index of a medium in which the measurement is made (air, oil) and k the absorption index of substance. Details of optical reflectance measurement are described in our previously published paper [31].

(5)

Elastic modulus (GPa)

20

15

10

5

0

Fig. 3. The plot of the reduction in dimension of carbonised to 550 and 950 1C bamboo and yucca for axial, radial and tangential directions.

0.85

(a)

0.90 Apparent density (g/cc)

0.95

4

3 0.6

Elastic modulus (GPa)

App. density of carbonized samples (g/cc)

0.8

Bamboo

0.4

2

1

0.2 Yucca

0.0

0 0.4 0.8 App. density of raw samples (g/cc)

1.2

Fig. 4. Apparent density change from carbonisation of plants studied. Values are for carbonisation in a nitrogen atmosphere at 3 1C/min to 550 1C ( ) and 950 1C (K). Straight line denotes the fitting of the experimental data for 17 points: Y ¼ 0:6779X ðr ¼ 0:996Þ.

0.24

(b)

0.28 0.32 Apparent density (g/cc)

Fig. 5. The plot of the dynamic elastic modulus of initial bamboo (a) and initial yucca (b) measured along axial ( ), radial (’) and tangential ( ) directions vs. apparent density. The data for collar bamboo (axial direction) are denoted by symbol (m).

ARTICLE IN PRESS M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

between precursor plant and resulting carbonised plant which can be expressed as [4]

3. Results and discussion 3.1. Weight, dimensions and density after carbonisation

rcarb ¼ crraw ,

(6)

where rcarb and rraw are the apparent densities of the carbonised plant and the precursor plant, respectively, and c is a constant for the particular carbonisation condition. Byrne and Nagle [4] obtained for hardwoods c ¼ 0:8176 for highly controlled pyrolysis at temperatures within the range of 400
2500 1C. We obtained for the perennials woody stems studied the lower value of c ¼ 0:6779 for the pyrolysis temperatures of 550 and 950 1C (for 1 h) with the rate of heating 3 1C/min. The value of c was determined by fitting of the experimental data using the Eq. (6) (the correlation coefficient r ¼ 0:996). It can be seen from Table 2 that porosity (calculated with Eq. (3)) of carbonised species increased to the distinct magnitudes to about 90%. Thus carbonised woody stems of perennials studied can be very good precursors (as skeleton) for various wood-ceramics.

Heat-treatment in neutral atmosphere of plants caused a loss in weight, dependent on temperature used. Results of carbonisation weight yield are put in Table 2. It can be seen that yield varied for T ¼ 550 1C between 30 and 32 wt%, and for T ¼ 950 1C between 26 and 28 wt% for all samples studied. The carbonisation yield changes with sample and the kind of plant in an inconsiderable degree, i.e., within of 2%. The lower carbonisation yield for the higher temperatures was already observed for hardwoods and reported in Refs. [5,9,13,21]. Fig. 3 shows the dimensional changes due to carbonisation in principal directions for yucca and cuboid bamboo carbonised to 550 and 950 1C. It was found that as in the case of hardwoods [5,9,21] the higher the carbonisation temperature, the greater the dimensional shrinkage. The greatest changes are observed in ‘‘transverse’’ to plant fibres directions, i.e., for radial and tangential. For the case of cuboid bamboo it was obtained the similar values of RDrad and RDtan, observed for both carbonisation temperatures. Byrne and Nagle [4] reported that the dimensional changes in principal directions for all hardwoods carbonised were anisotropic, and the reduction of dimension from heat-treatment in each studied species was greatest in the tangential direction (except for balsa). In the case of perennials investigated in this work, we can rather say about transverse isotropy of the dimensional changes under heat-treatment. All samples studied demonstrated a reduction in apparent density due to carbonisation. A remarkable relationship between values for raw plants apparent density and those for carbonised plants was observed. This is shown in Fig. 4. For the particular heat-treatment conditions when a monolithic piece of plant stem was carbonised, the resulting char had 67.8% of the apparent density of the precursor plant. The slope of the linear fit to the density change data demonstrates a simple relationship

3.2. Ultrasonic velocity, dynamic elastic modulus and elastic anisotropy In comparison with raw material, carbonisation resulted in decreased ultrasonic velocity in all principal directions of samples studied. It can be seen from Table 2 that the velocity in the axial direction was equal to about 4200–4400 m/s for raw bamboo and about 3100 m/s for raw yucca, and it was reduced by 44–49% for carbonised bamboo and 21% for carbonised yucca. For bamboo the reduction in ultrasonic velocity for radial direction was about 52%, while the velocity for tangential direction was almost the same. It was found that vtangential4vradial for bamboo, raw and carbonised to both temperatures. This result is different as that reported by Byrne and Nagle [5] for hardwoods that exhibited the relation vradial4vtangential. The data from Table 2 demonstrate that the carbonised samples were characterized by ultrasonic

3.5

15

10

5

Elastic modulus (GPa)

Elastic modulus (GPa)

20 Elastic modulus (GPa)

171

2

1

3.0

2.5

2.0 0

0 0.6

(a)

0.8

0.6

1.0

Apparent density (g/cc)

(b)

0.8

0.6

1.0

Apparent density (g/cc)

(c)

0.8 1.0 Apparent density (g/cc)

Fig. 6. The plot of the dynamic elastic modulus of bamboo measured along axial (a), radial (b) and tangential (c) directions vs. apparent density. Notation: raw samples; carbonised to 550 1C; K carbonised to 950 1C.

ARTICLE IN PRESS M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

172

velocity increasing with a rise of the carbonisation temperature. For bamboo an increase of ultrasonic velocity in axial direction was distinctly greater than that in tangential, while along radial direction the velocity was slightly lower. The same trend of increase in velocity measured along axial direction in comparison with perpendicular to stem fibres direction was observed in the case of yucca. The trends in acoustic velocity with HTT serve to support the mechanism of atomic ordering within

the carbonised plant. The ultrasonic velocity changes with HTT can all be explained by atomic reordering in a solid carbon with crystallites of preferred orientation. It is known [5] that atomic ordering will increase material stiffness if a preferred orientation of layers is present. Increased stiffness is indicated by the increased ultrasonic velocities. In our studies, the changes of the velocity measured along stem fibres (an axial direction) due to carbonisation suggest the existence of a preferred orientation of graphitic layer planes along stem fibres.

4

20 Elastic anisotropy

Elastic modulus (GPa)

3

2

10

1

0 0.1

(a)

0.2

0.3

0.4

Apparent density (g/cc)

0.84

0.88 0.92 0.96 App. density of raw samples (g/cc)

(a)

1.00

12

0.4

0.3 Elastic anisotropy

Elastic modulus (GPa)

10

0.2

8

6

0.1

4

0.1

(b)

0.2 0.3 Apparent density (g/cc)

Fig. 7. The plot of the dynamic elastic modulus of yucca measured along axial (a) and radial (b) directions vs. apparent density. Notation: raw samples; carbonised to 550 1C; K carbonised to 950 1C.

0.24

(b)

0.28

0.32

App. density of raw samples (g/cc)

Fig. 8. The plot of the dynamic elastic anisotropy of initial and carbonised samples of bamboo (a) and yucca (b). Notation: raw samples; carbonised to 550 1C; K carbonised to 950 1C.

ARTICLE IN PRESS M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

Fig. 5 shows the plot of the dynamic elastic modulus of initial samples of bamboo and yucca vs. apparent density. As opposed to hardwoods studied by Byrne and Nagle [5] bamboo exhibits transverse isotropy of elastic properties, i.e., modulus for directions perpendicular to plant fibres 326

330

334

B [mT]

338

B2

Y1

B1

(a) 326

330

334

338

B [mT]

Y2

B2 B1

(b) Fig. 9. The plot of the EPR lines for series of samples studied: B1, B2 and Y1-2, carbonised to the temperatures 550 1C (a) and 950 1C (b).

173

(radial and tangential) is almost the same. Both plants studied have elastic moduli almost independent of apparent density, especially in the case of radial or tangential directions. Fig. 6 shows the plot of the dynamic elastic modulus of carbonised bamboo vs. apparent density related to that of the raw material. The data for carbonised samples is considerably less scattered than those for raw samples (except for Etang (T ¼ 950 1C)). This means that the heat-treated bamboo pieces at both temperatures have a more homogenous structure than those of initial material. What more, carbonised material is characterized by an almost constant value of E. The Eaxial values for samples pyrolysed to 950 1C are higher than those for material pyrolysed at the lower temperature, while the Eradial values are almost independent of the HTT. Modulus of carbonised material is lower, mainly because of the distinctly higher porosity (Table 2). Thus, bamboo carbonised to 950 1C is more porous, but stiffer. It is evident that the matrix of bamboo, carbonised to the higher temperature is a highly oriented structure that is more condensed. Fig. 7 shows the plot of the dynamic elastic modulus of raw and carbonised yucca vs. apparent density. The data of E for raw material depend distinctly on apparent density, which changes in a wide range. Similarly as in the case of bamboo, carbonised samples have closer values of E which is observed in both directions. The E value for carbonisation to 950 1C is higher than that for samples carbonised to the lower temperature. The difference between Eraw and E950 1C is lower than that in the case of bamboo. As it was earlier reported, raw bamboo was found to be a transverse isotropic material in contrast to hardwoods containing seasonal rings. However after heat-treatment in neutral atmosphere, the structure of bamboo became more anisotropic (Table 2). Fig. 8 shows the plot for the elastic anisotropy of raw and carbonised samples of bamboo and yucca. In the case of bamboo, elastic anisotropy clearly increases with the HTT but the character of the dependence on apparent density is completely different than the dependence for raw samples. Elastic anisotropy of raw bamboo increases with increasing apparent density, but it decreases with apparent density for carbonised species. The dependence of elastic anisotropy of yucca (distinctly less dense material than bamboo) is less clear but the increase of elastic anisotropy with the HTT is also seen.

Table 3 Concentration of paramagnetic centres N, g factor and linewidths DBpp of EPR spectra as well as the optical reflectance R of carbonised samples Sample

Carbonised to 550 1C N (spin/g)

Thick bamboo (series B1) Thin bamboo (series B2) Yucca (series Y1) Yucca (series Y2)

20

8.1  10 7.7  1020 2.5  1020 4.5  1020

g 2.0029 2.0029 2.0030 2.0030

Carbonised to 950 1C DBpp (mT) 0.53 0.57 0.63 0.56

R (%) 2.83 3.18 2.81 2.60

N (spin/g) 17

1.3  10 9.6  1017 3.2  1017 8.7  1017

g

DBpp (mT)

R (%)

2.0027 2.0028 2.0030 2.0028

1.27 0.46 0.39 0.56

4.58 6.38 3.75 4.24

ARTICLE IN PRESS M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

174 800 B1 B2 Y2

Amplitude [a.u.]

600

400

200

0 0.0

0.2

(a)

0.4 0.6 (M/Mo)1/2

0.8

1.0

6 B2 B1 Y2

Amplitude [a.u.]

4

2

0 0

0.2

(b)

0.4 0.6 (M/Mo)1/2

0.8

1

Fig. 10. The plot of the amplitude of the EPR line vs. microwave power determined for three series of samples studied: B1, B2 and Y2, carbonised to the temperatures 550 1C (a) and 950 1C (b).

3.3. EPR parameters Paramagnetic centres properties of the studied carbonised plants were examined with the use of electron paramagnetic resonance (EPR) spectroscopy. Fig. 9 shows the plot of the EPR spectra of yucca and bamboo heated to 550 and 950 1C. The EPR spectra are single asymmetric

lines. It can be seen from Fig. 9 that strong EPR signals are characteristic for the plants thermally decomposed at 550 1C, while plants heated to 950 1C are described by weak EPR lines with high level of noise. Concentrations of paramagnetic centres (N) and parameters of EPR lines: g-factors and linewidths (DBpp), are contained in Table 3. High concentrations (1020 spin/g) of paramagnetic centres resulted from breaking of chemical bonds during thermal decomposition of plants exist in the studied yucca and bamboo heated to 550 1C. The magnitudes of the free radical concentrations of chars observed in this study of range 1020 spin/g are qualitatively consistent with values that have been reported for cellulose chars by Feng et al. [32]. Strong dipolar interactions are responsible for the observed EPR line broadening (DBpp: 0.53
0.63 mT). g-Values in the range of 2.0029
2.0030 indicate that unpaired electrons are connected with carbon atoms, because of low spin-orbit coupling constant. Heating to 950 1C quenches paramagnetic centres in the samples of yucca and bamboo. Concentrations of paramagnetic centres decreased from 1020 to 1017 spin/g. This follows from the increase of multi-ring aromatic units in the samples studied. Dipolar interactions of unpaired electrons increases in thick bamboo (series B1) heated to 950 1C and its EPR lines broaden (DBpp ¼ 1.27 mT). Decrease of dipole-dipole interactions and linewidths was stated for thin bamboo (series B2) and yucca of series Y1. Magnetic interactions in yucca of series Y2 heated to 950 1C remain unchanged. The same linewidths (DBpp ¼ 0.56 mT) were obtained for the EPR spectra of yucca of series Y2 heated to both 550 and 950 1C. Fig. 10 shows the plot of the amplitude of the EPR line vs. microwave power determined for three series of samples studied: B1, B2 and Y2, carbonised to the temperatures 550 and 950 1C. Influence of microwave power on amplitudes of EPR spectra indicates existence of multi-ring aromatic units in plant samples thermally treated at both 550 and 950 1C. Microwave saturation of EPR lines was not observed up to 70 mW. Amplitudes of the EPR lines increase with increasing microwave power. It can be concluded that fast spin-lattice relaxation processes occur in the studied samples. Changes of amplitudes and linewidths of the analysed EPR lines with microwave power are characteristic for homogeneous distribution of paramagnetic centres in the samples. The lack of saturation of the EPR spectra of samples studied means that during slow highly controlled pyrolysis defects were not created. 3.4. Optical reflectance and texture Table 3 contains the optical reflectance R values of carbonised samples. The R values of samples carbonised to the higher temperature are distinctly greater than those of samples heat-treated at the lower temperature. Reflectance value is connected with the order of aromatic layer/clusters in carbonaceous materials. The maximum reflectance value occurs in the direction perpendicular to

ARTICLE IN PRESS M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

175

Fig. 11. Optical microscopy images (magnification 80  ) of carbonised to 950 1C samples of bamboo (series B1): (a) cross-section; (b) along fibres (long edge, centre of plant) and yucca (series Y2): (c) cross-section; (d) along fibres (large flat side).

aromatic layer. Heat-treatment leads to changes in the arrangement and dimensions of aromatic layers, as expressed in the changes of reflectance values. In the case of all samples studied, the higher temperature causes an increase of their reflectance values which indicates the increase of dimensions of aromatic layers and their ordering along preferential plane parallel to the length of stem. Paris et al. [14] deduced from SAXS/WAXS measurements that after the complete decomposition of the raw softwood structure, with increasing temperature the charring processes commenced with the formation and successive ordering of aromatic structures. The developing graphene sheets showed a slight preferred orientation with respect to the oriented cellular structure of the material. Our results are similar to those obtained by Paris et al. [14] for softwoods using the different technique. The value of reflectance changes depends on the sample individually, but is higher for bamboo samples than for yucca. Fig. 11 shows the typical structure of vascular plants studied, i.e., of bamboo and yucca according to the crosssection and along the length of stem, observed after carbonisation to 950 1C. Microscopic observations showed that during thermal treatment of yucca and bamboo their tissue structure remain unaltered. There is only the increase in order of aromatic layers in the walls of fibres expressed

by the increase of reflectance values during the carbonisation process. 4. Conclusions In the present work the properties of porous carbon materials derived from woody stems of different perennials such as bamboo and yucca were investigated. The results are summarized as follows: 1. Mass and dimensions were found to decrease with rising carbonisation temperatures. 2. Changes in the ultrasonic velocity due to carbonisation suggested the existence of a preferred orientation of graphitic layer planes along axial direction, i.e., along stem fibres. It was found that materials carbonised at the higher temperature were stiffer—more ordered in structure. 3. Anisotropy of elasticity of raw perennial stems was different than that of hardwoods containing seasonal rings. A transverse isotropy was observed. Character of elastic anisotropy was changed through carbonisation and its value was dependent on HTT. 4. Strong EPR signals were detected for the plants thermally decomposed at 550 1C, while plants heated

ARTICLE IN PRESS 176

M. Krzesin´ska et al. / Biomass and Bioenergy 30 (2006) 166–176

to 950 1C were described by weak EPR lines. Decrease in concentration of paramagnetic centres with increasing HTT as well as the results of microwave power saturation evidenced the increase of multi-ring aromatic units in carbonised plants. 5. Optical reflectance of samples carbonised to the higher temperature was found to be distinctly greater than that of samples heat-treated at the lower temperature. An increase of the R values measured perpendicularly to fibres indicated the increase of dimensions of aromatic layers and their ordering along the plane parallel to the length of stem. The difference between structure and values of the physical parameters of hardwoods (described in the cited literature) and of woody perennials stems studied in this work was noticed. It was shown that as in the case of hardwoods, the original structure of plant stems was retained in the porous carbon material during carbonisation. The increase in order of aromatic layers in the walls of fibres was confirmed by the physical parameter values changing through the carbonisation process. The difference in stem structure of carbonised grasses (bamboo) and flowering shrubs (yucca) was found: structure of bamboo was more ordered than that of yucca. Carbonised woody stems of perennials studied, especially of bamboo were found as very porous, but stiff materials, which can be excellent precursors (as skeleton) for new eco-materials, e.g., for wood-ceramics. Acknowledgement The authors gratefully acknowledge Mrs. Marianna Jasin´ska for fruitful laboratory cooperation. References [1] Serio MA, Kroo E, Wo´jtowicz MA. Biomass pyrolysis for distributed energy generation. Preprints of Papers—American Chemical Society, Division of Fuel Chemistry 2003;48(2):584–9. [2] Overend RP. Biomass resources to support a low carbon future. Preprints of Papers—American Chemical Society, Division of Fuel Chemistry 2003;48(2):582–3. [3] Rambo CR, Cao J, Sieber H. Preparation and properties of highly porous, biomorphic YSZ ceramics. Materials Chemistry and Physics 2004;87:345–52. [4] Byrne CE, Nagle DC. Carbonization of wood for advanced materials applications. Carbon 1997;35:259–66. [5] Byrne CE, Nagle DC. Carbonized wood monoliths—characterization. Carbon 1997;35:267–73. [6] Kercher AK, Nagle DC. Evaluation of carbonized medium-density fiberboard for electrical applications. Carbon 2002;40:1321–30. [7] Sugimoto H, Norimoto M. Dielectric relaxation due to interfacial polarization for heat-treated wood. Carbon 2004;42:211–8. [8] Kercher AK, Nagle DC. AC electrical measurements support microstructure model for carbonization: a comment on dielectric relaxation due to interfacial polarization for heat-treated wood. Carbon 2004;42:219–21. [9] Kercher AK, Nagle DC. Microstructural evolution during charcoal carbonization by X-ray diffraction analysis. Carbon 2003;41:15–27.

[10] Kercher AK, Nagle DC. Monolithic activated carbon sheets from carbonized medium-density fiberboard. Carbon 2003;41:3–13. [11] Klose W, Schinkel A. Measurement and modelling of the development of pore size distribution of wood during pyrolysis. Fuel Processing Technology 2002;77–78:459–66. [12] Kurosaki F, Ishimaru K, Hata T, Bronsveld P, Kobayashi E, Imamura Y. Microstructure of wood charcoal prepared by flash heating. Carbon 2003;41:3057–62. [13] Treusch O, Hofenauer A, Troger F, Fromm J, Wegener G. Basic properties of specific wood-based materials carbonised in a nitrogen atmosphere. Wood Science and Technology 2004;38(5):323–33. [14] Paris O, Zollfrank C, Zickler GA. Decomposition and carbonisation of wood biopolymers—a microstructural study of softwood pyrolysis. Carbon 2005;43:53–66. [15] Lopez M, Labady M, Laine J. Preparation of activated carbon from wood monolith. Carbon 1996;34:825–7. [16] Tseng R-L, Wu F-C, Juang R-S. Liquid-phase adsorption of dyes and phenols using pinewood-based activated carbons. Carbon 2003;41: 487–95. [17] Mohan D, Singh KP, Sinha S, Gosh D. Removal of pyridine from aqueous solution using low cost activated carbons derived from agricultural waste materials. Carbon 2004;42:2409–21. [18] Su W, Zhou L, Zhou Y. Preparation of microporous activated carbon from coconut shells without activating agents. Carbon 2003;41:861–3. [19] Byrne CE, Nagle DC. Cellulose derived composites—a new method for materials processing. Materials Research Innovations 1997;1(3): 137–44. [20] Ozao R, Pan W-P, Whitely N, Okabe T. Coal-like thermal behavior of a carbon-based environmentally benign new material: woodceramics. Energy & Fuels 2004;18:638–43. [21] Hirose T, Fujino T, Fan T, Endo H, Okabe T, Yoshimura M. Effect of carbonisation temperature on the structural changes of woodceramics impregnated with liquefied wood. Carbon 2002;40: 761–5. [22] Fey T, Sieber H, Greil P. Stress distribution in biomorphous SiCceramics under radial tensile loading. J. Eur. Ceram. Soc. 2005;25: 1015–24. [23] Mizutani M, Takase H, Adachi N, Ota T, Dajmon K, Hikichi Y. Porous ceramics prepared by mimicking silicified wood. Science and Technology of Advanced Materials 2005;6:76–83. [24] From the Wikipedia encyclopedia. [25] Stein RS, Wilkes GL. In: Ward IM, editor. Structure and Properties of Oriented Polymers. London: Applied Science; 1979. p. 136. [26] Krzesin´ska M, Celzard A, Mareche J-F, Puricelli S. Elastic properties of anisotropic monolithic samples of compressed expanded graphite studied with ultrasounds. Journal of Materials Research 2001;16: 606–14. [27] Celzard A, Krzesin´ska M, Begin D, Mareche J-F, Puricelli S, Furdin G. Preparation, electrical and elastic properties of new anisotropic expanded graphite-based composites. Carbon 2002;40:557–66. [28] Krzesin´ska M. Structure and properties of natural and processed carbon materials studied with ultrasounds: a review. In: Richard R, editor. Current Topics in Acoustical Research. Research Trends Trivandrum India, vol. 3; 2003. p. 43–61. [29] Pilawa B, Wi˛eckowski AB, Pietrzak R, Wachowska H. Oxidation of demineralized coal and coal free of pyrite examined by EPR spectroscopy. Fuel 2002;81:1925–31. [30] Pilawa B, Wi˛eckowski AB, Wachowska H, Koz"owski M. Effect of reduction and butylation on coal. Application of microwave saturation of two-component EPR spectra. Applied Magnetic Resonance 2003;24:73–83. [31] Pusz S, Kwiecin´ska BK, Duber S. Textural transformation of thermally treated anthracites. International Journal of Coal Geology 2003;54:115–23. [32] Feng J- W, Zheng S, Maciel GE. EPR investigations of charring and char/air interaction of cellulose, pectin, and tobacco. Energy & Fuels 2004;18:560–8.