Kinetics of silica dissolution from rice husk char.

Benke, Dean J. ; Wainwright, Mark S. ; Nigam, K.D.P. 等

The effect of leaching time and temperature on the ash content of rice husk char and the resultant surface area development was investigated. Experiments on dissolution of rice husk char ash with 0.5 mol NaOH solution were conducted at 40[degrees]C, 60[degrees]C, 80[degrees]C and 95[degrees]C for different times up to 7200 s. The surface areas of the leached chars after drying were determined using Micromeritics 2200 surface area analyzer. The leaching data was analyzed using different models applicable for fluid-solid non-catalytic reactions. It was found that the data could be satisfactorily represented by diffusion-controlled mechanism as well as modified shrinking core models. The activation energies were in the range of 41.75 to 51.25 kJ/mole. The surface areas of the leached chars were found to depend linearly on the percentage of ash content in the chars.

On a etudie l'effet du temps de lessivage et de la temperature sur la teneur en cendres de cosses de riz carbonisees et le developpement de surface qui en resulte. Des experiences ont ete menees sur la dissolution des cendres de cosses de riz avec une solution de NaOH de 0,5 mole a 40 [degre]C, 60 [degre]C, 80 [degre]C et 95 [degre]C pour differents temps allant jusqu'a 7200 sec. Les aires de surfaces des cendres apres le sechage ont ete determinees a l'aide d'un analyseur de surface 2200 de Micromeritics. Les donnees de lessivage ont ete analysees a l'aide de differents modeles applicables a des reactions fluide-solide non catalytiques. On a trouve que les donnees pouvaient etre representees de maniere satisfaisante par le mecanisme controle par la diffusion et par les modeles de noyau retrecissant modifies. Les energies d'activation sont comprises entre 41,75 et 51,25 kJ/mole. On a trouve que les aires de surfaces des cendres lessivees dependaient lineairement du pourcentage de teneur en cendres dans les produits carbonises.

Keywords: silica dissolution, solid-liquid reactions, heterogeneous reaction, rice hull char, leaching

INTRODUCTION

Activated carbon is widely used for removal of hazardous pollutants like different phenols, chromium compounds, etc., from industrial waste water streams and also for adsorption of gaseous pollutants from gas streams. Due to high cost of activated carbon, cheaper resources of carbon are being developed for adsorption of pollutants. Agro based biomass materials are quite promising as they are available in large quantities and also cheap. Among these, rice husk is the most important material. Several methods are used to produce activated carbon and among them, the silica dissolution technique using sodium hydroxide is attractive due to co-production of sodium silicate. El Sayed et al. (1979), studied silica leaching from rice husks with sodium hydroxide solution of different concentrations and at different temperatures. They reported that maximum silica removal occurred at temperatures between 50-60[degrees]C. Moya Portugez et al. (1988), reported that sodium hydroxide not only remove silica but also some bio-organic compounds from rice hulls. The dissolution of silica from rice husk ash with lime was investigated by Rama Rao et al. (1989), Borthakur et al. (1980), and James and Rao (1986, 1987). They studied the lime reactivity of silica in terms of process variables including temperature, time of pyrolysis of ash preparation, and cationic impurities. Jain et al. (1995), reported process development methods for preparation of sodium silicate and potassium silicate by leaching of rice husk chars.

It has been found that silica dissolution process is affected by several factors. El Sayed et al. (1979) reported that at higher leaching temperatures the ash content in the treated rice husk increased which was attributed to reprecipitation of dissolved silica on the fibres. Rama Rao et al. (1989), and James and Rao (1988) reported that temporary formation of gelatinous silicate hydrate enveloped the silica grain as a semi-permeable membrane and inhibited the diffusion mass transfer rates of sodium hydroxide for reaction. Iler (1979) has reported that there is a possibility of re-adsorption of soluble silica onto the char surface by two different mechanisms. They also reported that an increase in the viscosity of sodium silicate formed within the solid could also inhibit the rate of diffusion of sodium hydroxide within the porous structure. There are few studies of the leaching of silica from rice husk char and no work has been reported on the kinetics of the silica dissolution. The present work attempts to determine the kinetic parameters for silica dissolution and also surface area development in the resultant leached chars. Cheng et al. (2000), studied the reaction of silica with ethylene glycol in the presence of high boiling organic amines as catalyst. It is reported that the silica dissolution is found to be pseudo zero order in silica. The activation energy reported is 58 kJ/mol.

EXPERIMENTAL

The rice husk char used was a single batch of material produced in an industrial pyrolyser/combustion unit by the company BIOCON that is a division of the Ricegrowers Cooperative. The char was produced from rice husk waste material as part of a heat recovery process. Rice husk char having 67.6% ash content was ground to powder form and the fraction passing through 250 microns sieve was collected and used in the experiments. Sodium hydroxide solution of 0.5M concentrations was used for leaching. 200 ml of the solution was placed in 250 ml conical flasks and immersed in a Paton's reciprocating shaking bath at a set temperature and a frequency of 1 rpm. The temperature in the shaker bath was monitored until the desired temperature of the solution was reached. The flasks were then taken out and 2 gm of the rice husk char were added to each flask and shaken thoroughly in order to ensure complete wetting of char in the solution and then placed back in the shaker bath and zero time recorded. The flasks were taken out from the shaker after predetermined leaching times and the char and the liquor were separated immediately by filtration. The treated char was then washed thoroughly with distilled water and dried in an air oven at 60[degrees]C. The ash contents were determined by placing the dried char samples in a muffle furnace at 600[degrees]C for 24 h. The surface areas of the treated chars were obtained using Micromeritics 2220 high-speed single point BET surface area analyzer. Rice hull char ash component was determined by using X-Ray Fluorescence. Approximately 96% of the ash content comprised of silica was obtained.

THEORETICAL

The leaching of silica with sodium hydroxide solution is a typical example of a liquid-solid non-catalytic reaction. Several models reported in the literature for these reactions have been used in the present study for interpreting the data and determining the rate parameters. These are given below for spherical solid reactant particles.

Shrinking Core Model

The time versus concentration relationship for different rate controlling steps, according to shrinking core model (Levenspiel, 1988) are given below:

(a) Fluid film diffusion rate controlling:

X = [K.sub.1]t (1)

[K.sup.-1.sub.1] = [rho]R/3[K.sub.g][C.sub.Ao] (2)

(b) Ash layer diffusion rate controlling:

1 - 3[(1-X).sup.2/3] + 2(1-X) = [K.sub.2]t (3)

[K.sup.-1.sub.2] = [rho][R.sup.2]/6[D.sub.eff][C.sub.Ao] (4)

(c) Surface chemical reaction rate controlling:

1 - [(1-X).sup.1/3] = [K.sub.3]t (5)

[K.sup.-1.sub.3] = [rho]R/[K.sub.r][C.sub.Ao] (6)

Modified Shrinking Core Model

Assuming that the limiting reaction step is surface chemical reaction on the unreacted cored and that the presence of cationic impurities and sodium silicate deposits on the unreacted core surface reduce the available surface area for reaction, which is proportional to the extent of conversion. The rate equation for this type of condition is given by Yoshioka et al. (1998):

[(1-X).sup.-2/3] -1 = [K.sub.4]t (7)

[K.sup.-1.sub.4] = R[rho]/2[K.sub.r][C.sub.Ao] (8)

Homogenous Model

According to this model, the diffusional resistance for the gaseous reactant in the solid is considered negligible and the reaction occurs throughout the solid homogenously. The rate equation is given by:

-dW/dt = [K.sub.5][W.sup.n][C.sup.m.sub.A0] (9)

For n = 1 and m = 1, integration of Equation (9) gives:

-Ln(l-X) = [K'.sub.5]t (10)

where X = 1 - (W/[W.sub.0]) (11)

[K'.sub.5] = [K.sub.5][C.sub.Ao] (12)

Modified Homogenous Model

Here, similar to the modified shrinking core model, the effect surface area per unit volume can be assumed to be reduced proportional to the extent of conversion. The rate equations are given by:

-dW/dt = [K.sub.5]W[C.sub.A0](l-X) (13)

dX/dt = [K'.sub.5][(l-X).sup.2] (14)

Integration of Equation (14) gives:

X/(1-X) = [K'.sub.5]t (15)

Equation (15) is similar to second-order homogenous rate equation.

Parabolic Law

In this case (Kubaschewski and Hopkins, 1953) it is assumed that rate of thickening of the product layer is inversely proportional to the thickness of the product layer at any time. The rate equation is given by:

dy/dt = [K.sub.6]/y (16)

Integration of Equation (16) gives:

1-2[(1 - X).sup.1/3] + [(1 - X).sup.2/3] = [K'.sub.6]t (17)

where, y = (R-[r.sub.c]) (18)

X = 1 - [([r.sub.c]/R).sup.3] (19)

[K'.sub.6] = [K.sub.6]/[R.sup.2] (20)

RESULTS AND DISCUSSION

The fractional conversion versus time data for different temperatures is shown in Figure 1. It is seen that the extent of removal of silica is not complete but increases with temperature. Complete removal of silica would not be achieved due to the inability of caustic soda penetration to all the silica site as same are covered by resistant carbon structure. It could also be due to some reprecipitation of silica. Also, high viscosity of caustic soda at lower temperatures causes resistance for diffusion of caustic soda to penetrate into the pore structure resulting in lower values of silica removal at lower temperatures.

[FIGURE 1 OMITTED]

Reaction of silica with caustic soda is a typical non-catalytic fluid-solid reaction which involves different rate processes like mass transfer through external fluid film surrounding the solid, diffusional mass transfer within the porous solid and chemical reaction on the surface of the solid reactant. Although the chemical reaction step involves adsorption, desorption and surface chemical reaction with the adsorbed species, quantitative formulation of these processes is not possible due to continuous changes in the solid reactant surface and transient behaviour. Several models have been reported in literature for non-catalytic fluid-solid reactions. In the preset case, due to the small size of the reactant, solid particles models, which are appropriate, are considered for analysis of the data. These are discussed above.

From the experimental data shown in Figure 1, it is clear that the time versus conversion data is not linear and hence Equation (1) corresponding to shrinking core model with fluid film mass transfer as rate controlling is not satisfactory. Therefore, Equations (3), (5), (7), (10), (15) and (17) corresponding to other situations have been used for analyzing the data and to determine the rate constants. This is done using Microsoft Excel computer programming. Figure 2 shows typical results for the data obtained at temperature of 60[degrees]C according to the above model equations. It is observed that there is same scattering of the data points from linear relationship with some models. From the complete analysis of the data for all temperatures, it is found that the homogeneous model with first order kinetics, Equation (10), and shrinking core model with chemical reaction step as rate controlling, Equation (5), does not show satisfactory linear trends and the regression coefficients are below 0.7. Hence, these two models are also considered not satisfactory. These two models represent chemical reaction as the rate controlling mechanisms with negligible diffusional mass transfer resistance within the porous solid. The diffusional mass transfer may be considerable in view of the pore structure of the solid and high viscosity of the caustic soda. This fact is further supported by the activation energy values obtained with other models.

[FIGURE 2 OMITTED]

Figure 3 shows typical data for the parabolic law model for the four temperatures indicating satisfactory representation of the model. Table 1 gives the rate constants in Arrehenius form and the corresponding regression coefficient values for the remaining four models. These four models represent satisfactorily the experimental results. The activation energy values obtained are in the range of 41-51 kJ/mol. The activation energy values obtained in the present case are close to the value of 58 kJ/mol reported by Cheng et al. (2000) for silica dissolution with ethylene glycol and potassium hydroxide mixture. It is known that apparent activation energies for chemical reactions are typically in the range of 80-250 kJ/mol while those for diffusional processes it is much lower. The activation energy values obtained in the present case support the view that diffusional resistance strongly influence the leaching process and the overall reaction is not truly chemical reaction control. Also, it is known that influence of diffusional process on chemical reaction decreases the apparent activation energy with increasing temperatures. This aspect is clearly seen with respect to parabolic law and shrinking core model with ash layer diffusion control condition. These two models are typical diffusion controlled processes. The activation energies are 62 kJ/mol, 54 kJ/mol and 51 kJ/mol for temperature intervals of 40-60[degrees]C, 40-80[degrees]C and 40-95[degrees]C, respectively, for parabolic law model. Similarly for shrinking core model with ash layer diffusion control, these are 58 kJ/mol, 50 kJ/mol and 46 kJ/mol for temperature intervals of 40-60[degrees]C, 40-80[degrees]C and 40-95[degrees]C, respectively. The decrease in activation energies is clearly observed from the rate constant values in the Arrhenius plot shown in Figure 4. This suggests that the overall mechanism of reaction is predominantly pore diffusion controlled.

[FIGURES 3-4 OMITTED]

In the case of modified shrinking core and homogeneous models the activation energies increases from 41-42 and 46-47 as the temperature interval increases. This is contrary to the above observation.

These two models assume chemical reaction step as the main mechanism of reaction with reduction in surface areas for reaction. The presence of cationic impurities and reprecipitation of silica justifies the above models. It can be concluded that the above four models represent satisfactory the leaching process of silica with caustic soda. A decisive conclusion of specific representative model is difficult and requires exact analysis of the structure of the rice husk char and reactant distribution within the solid.

Figure 5 shows the surface area against fractional conversion of silica in the treated char for the data of 80[degrees]C. It can be seen that the surface area in the treated char is well represented by a linear relation with the percentage conversion of silica or the ash content in the treated chars. It was found that the data of the experiments conducted at 40[degrees] C, 60[degrees]C and 80[degrees]C followed approximately the same relationship between the surface area and the silica conversion. The data of 95[degrees]C showed a slightly different relationship with higher values of surface area corresponding to fractional conversion. This could be due to removal of impurities both from silica as well as carbon content resulting in development of higher surface areas. The relationship between surface area against fractional conversion for data of 40[degrees]C-80[degrees]C and for 40[degrees]C-95[degrees]C data points are given by Equations (21) and (22), respectively.

A = 3.17X + 51.26 (21) [R.sup.2.sub.0] = 0.92

A = 2.65X + 64.3 (22) [R.sup.2.sub.0] = 0.92

[FIGURE 5 OMITTED]

The intercept value of 51.3[m.sup.2]/g is slightly greater than the surface area of the untreated char measured by single point nitrogen adsorption (35 [m.sup.2]/g). However, the value is almost equal to the surface area of the char (52[m.sup.2]/g) measured by the Omnisorp 100CX.

NOMENCLATURE

A surface area of the char ([m.sup.2]/g)
[C.sub.A0] concentration of sodium hydroxide solution
 (mol/[m.sup.3])
[D.sub.eff] effective diffusion coefficient ([m.sup.2]/s)
[K.sub.1] ... [K.sub.4] apparent rate constants ([min.sup.-1])
[K'.sub.5], [K'.sub.6] apparent rate constants ([min.sup.-1])
[K.sub.g] film mass transfer coefficient (m/s)
[K.sub.r] intrinsic reaction rate constant (m/s)
[r.sub.c] radius of the unreacted core of the particle
 (m)
R radius of the particle (m)
[R.sup.2.sub.0] regression coefficient (-)
[R.sub.g] gas constant (J/mol.K)
t leaching time (s)
W weight of the solid reactant (kg)
[W.sub.0] initial weight of the solid reactant (kg)
X fractional conversion
Y thickness of the solid product layer (m)

Greek Symbols

[rho] molar density of the solid reactant
 (mol/[m.sup.3])

REFERENCES

Borthakur, P. C., P. C . Saika and S. N. Dutta, "Physicochemical Characteristics of Silica from Paddy Husk, Its Reactivity and Probable Field of Application," Indian Ceramics 23(2), 25-29 (1980).

Cheng, H., R. Tamaki, R. M. Laine, F. Babonneau, Y. Chujo and D. R. Treadwell, "Neutral Alkoxysilanes from Silica," J. Am. Chem. Soc. 122, 10063-10072 (2000).

El Sayed, H., A. E. El-Ashmawy and M. A. Hamad, "Waste Treatment and Utilization, Theory and Practices of Waste Management," Procs. of the International Symposium, University of Waterloo, ON 1978, M. Moo Young and G. J. Pergamon, Eds., Oxford, U.S. 363 (1979).

Iler, P. K., "The Chemistry of Silica Solubility Polymerization, Colloid and Surface Properties and Biochemistry," John Wiley and Sons, U.S. (1979).

Jain, A., T. R. Rao, S. S. Sambi and P. D. Grover, "Energy and Chemicals from Rice Husk," Biomass Bioenergy 2, 285-289 (1995).

James, J. and M. S. Rao, "Reactivity of Rice Husk Ash," Cement Concrete Res. 16, 296-302 (1986).

James, J. and M. S. Rao, "A Reply to a Discussion by D. J. Cook of the paper, 'Reaction Product of Lime and Silica from Rice Husk Ash'," Cement Concrete Res. 17, 687-690 (1987).

James, J. and M. S. Rao, "Optimization of Ash/Lime Ratio for Rice Husk Ash-Lime Cements," Cemento 85(2) 101-10 (1988).

Kubaschewski, O. and B. E. Hopkins, "Oxidation of Metals and Alloys," Academic Press Inc., N.Y. (1953).

Levenspiel, O., "Chemical Reaction Engineering," 2nd ed., Wiley Eastern Limited, New Delhi (1988).

Moya Portuguez, M. E., C. M. Duran and B. R. Sibaja, "Use of Agroindustrial Wastes. Plastification of Pineapple (Ananas Comusus) Peel, Rice (Oryza Sativa) Husks, and Terminalia Amazonia Sawdust," Ingenieria y Ciencia Quimica 12(1-2), 24-27 (1988).

Rama Rao, G., A. R. K. Sastry and P. K. Rohatgi, "Nature and Reactivity of Silica Available in Rice Husk and Its Ashes," Bull. Mater. Sci. 12, 469-480 (1989).

Yoshioka, T., N. Okayama and A. Okuwaki, "Kinetics of Hydrolsis of PET Powder in Nitric Acid by Modified Shrinking Core Model," Ind Eng. Chem. RES. 37, 336-340 (1998).

Dean J. Benke (1), Mark S. Wainwright (2), K. D. P. Nigam (3) and T. R. Rao (3)*

(1.) Glaxo-SmithKline, Consumer Healthcare, Sydney, NSW, Australia

(2.) School of Chemical Engineering and Industrial Chemistry, The University of New South Wales, UNSW Sydney, NSW 2052, Australia

(3.) Department of Chemical Engineering, Indian Institute of Technology, Hauz Khas, New Delhi, Delhi, India

* Author to whom correspondence may be addressed. E-mail address: trrao@chemical.iitd.ernet.in

Manuscript received September 2, 2005; revised manuscript received July 26, 2006; accepted for publication July 28, 2006.

Table 1. Rate constants for different models

Model Rate constant

Modified shrinking core Ln[K.sub.4] = -41755/
model (Equation (7)) [R.sub.g]T+10.2

Shrinking core model with ash layer Ln[K.sub.2] = -46379/
diffusion control (Equation (3)) [R.sub.g]T+10.1

Modified homogeneous model Ln[K'.sub.5] = -47420/
(Equation (15)) [R.sub.g]T + 12.8

Parabolic law model Ln[K'.sub.6] = -51232/
(Equation (17)) [R.sub.g]T+ 10.9

Model [R.sup.2.sub.0]

Modified shrinking core 0.99
model (Equation (7))

Shrinking core model with ash layer 0.99
diffusion control (Equation (3))

Modified homogeneous model 0.99
(Equation (15))

Parabolic law model 0.99
(Equation (17))