Mass Transfer and Adsorption Kinetics of Phenolic Compounds onto Activated Carbon Prepared from Rice Husk

Transcription

1 157 Mass Transfer and Adsorption Kinetics of Phenolic Compounds onto Activated Carbon Prepared from Rice Husk Mamdouh M. Nassar 1*, Yehia H. Magdy 1, Abd El Hakim Daifullah 2 and H. Kelany 1 (1) Chemical Engineering Department, Faculty of Engineering, El-Minia University, El-Minia, Egypt. (2) Hot Laboratories and Waste Management Center, Atomic Energy Commission, Cairo, Egypt. (Received 2 February 28; revised form accepted 15 April 28) ABSTRACT: Activated carbon from an agriculture by-product (rice husk) was prepared by carbonization at 5 C (5 C/15 min) followed by activation at 85 C for 1 h. The adsorption of phenolic compounds onto this activated carbon was studied using batch adsorber methods. The experimental results showed that the prepared activated carbon removed phenolic compounds effectively from aqueous solution. Analysis of the contact time data gave an indication of the mechanism. The external mass-transfer constant,, involved in the adsorption process was determined using different initial adsorbate concentrations (C ) and adsorbent masses (M). The following relationships were obtained: =.23(M).55 and =.6(C ).17. A macropore rate parameter for intraparticle diffusion, k p, was defined and used to correlate the experimental data via the following expression: k p =.183(C ).48. A twin-resistance mass-transfer model based on external mass transfer and pore diffusion was applied successfully to the system under investigation. Concentration decay curves were predicted using a single modified mass-transfer coefficient,, and a single effective diffusion coefficient, D eff, for studies involving variable adsorbent masses and initial adsorbate concentrations. 1. INTRODUCTION Phenolic compounds are amongst priority water pollutants whose ecological risks and toxicity to aquatic flora and fauna are biomagnified through various food chains. Their sources are wastewater discharged by petroleum refiners, plastics and rubber manufacture, pharmaceuticals and woodpreservation processes. Such compounds, especially chlorophenols, are also used in pesticides. Although a number of processes have been developed for the removal of phenolic compounds, the most important from an economic standpoint is adsorption employing activated carbon. Mouraa et al. (26) have described the preparation of activated carbon by both physical and chemical activation methods. Activated carbon prepared from coconut shell pre-treated with KOH exhibits a high surface area and a good adsorption capacity towards the adsorption of phenol (Radhika and Palanivelu 26). Activated carbon from firewood has also been found to be a good adsorbent of phenol (Wu and Tseng 26). In the present work, rice husk has been employed as a source of activated carbon. In order to explore the potential of this prepared activated carbon towards the removal of phenolic * Author to whom all correspondence should be addressed. Mamdouh_Nassar@yahoo.com.

2 158 M.M. Nassar et al./adsorption Science & Technology Vol. 26 No compounds from aqueous media, the adsorption of four phenolic contaminants, viz. phenol, 2,4-dichlorophenol, 2,6-dichlorophenol and 2,4-dinitrophenol, was investigated. 2. EXPERIMENTAL Rice husk (RH), a by-product from rice milling, obtained from the rice mill of El-Mansoura City (Egypt) was used as the raw material for preparing activated carbon (ACRH). Activated carbon was prepared in two stages: carbonization followed by activation. Carbonization was carried out by heating the rice husk in a tubular furnace (a 4 cm i.d. stainless steel tube mounted inside calibrated tubular electric furnace). Thus, RH was heated gradually at a rate of 5 C/15 min in a limited supply of air. The activation step was carried out in an atmosphere of CO 2 with the temperature maintained at 85 C for 1 h. Aqueous solutions of all four phenolic compounds studied (phenol, 2,4-dichlorophenol, 2,6-dichlorophenol and 2,4-dinitrophenol) were prepared employing distilled water, the concentration of such compounds being measured by means of a Shimadzu model 161 spectrophotometer. All measurements were undertaken at the wavelength corresponding to the maximum absorbance, λ max, i.e. 27 nm for phenol, 36 nm for 2,4-dinitrophenol, nm for 2,4-dichlorophenol and 284 nm for 2,6-dichlorophenol Kinetic experiments A batch adsorber vessel with a standard tank configuration was used in all the experiments. This adsorber initially contained 1.7 dm 3 of the well-stirred aqueous solution of the phenolic compound under study at a known initial concentration. Then a prepared quantity of ACRH with a uniform particle size was introduced into the vessel and the resulting mixture agitated by means of a four-bladed turbine-type mixer. Samples were withdrawn from the system at selected time intervals and analyzed, the process being continued until equilibrium had been attained in the system. During this process, the temperature of the system was maintained at a constant value (2 ± 2 o C). The following variable parameters were studied: (i) Variation of the initial concentration of the phenolic compounds. Thus, the adsorption rate was measured using a range of initial concentrations, C, between 25 mg/dm 3 and 1 mg/dm 3. (ii) Variation of the adsorbent mass. For this purpose, different masses of activated carbon within the range 3 4 g were employed. Thus, the following masses and initial concentrations were employed in combination: (1) C = 25 mg/dm 3 M =.3,.61, 2.5 g (2) C = 5 mg/dm 3 M =.7, 1.3, 2.5 g (3) C = 1 mg/dm 3 M = 1., 2.5, 4. g All samples were shaken at a constant rate of 22 rpm. 3. RESULTS AND DISCUSSION 3.1. Adsorption rate equilibria Adsorption from aqueous solutions of phenolic compounds is controlled by mass transfer through the adsorbent phase. This diffusion process is, in turn, controlled by the transfer of adsorbate

3 Mass Transfer/Adsorption Kinetics of Phenolic Compounds onto Activated Carbon q t (mg/g) Time (h) Figure 1. Equilibrium time necessary for the adsorption of the phenolic compounds studied onto activated carbon. Data points relate to the following:, 2,4-dichlorophenol;, 2,6-dichlorophenol;, 2,4-dinitrophenol;, phenol. within the adsorbent particles, the process continuing until a characteristic equilibrium is attained. At such an equilibrium, a defined distribution of solute exists between the liquid and solid phases. The process equilibrium indicates the adsorption capacity of the solid phase, i.e. the amount of adsorbent necessary to remove a given amount of pollutant from aqueous solution. The development of equilibrium conditions for each of the four phenolic compounds studied is depicted in Figure 1 as a plot of q e (amount of solute adsorbed) versus time, t. In all cases, equilibrium was attained within ca. 2 h. ARCH exhibited the highest adsorption capacity towards phenol (7 mg/g) with the lowest adsorption capacity being exhibited towards nitrophenol (26 mg/g) Mechanism of sorption of phenolic compounds studied In terms of their maximum adsorption capacities (defined as the amount of phenolic compound adsorbed at the plateau of the adsorption isotherm), the phenolic compounds studied may be arranged in the following order (see Figure 1): phenol > 2,4-dichlorophenol > 2,6-dichlorophenol > 2,4-dinitrophenol. This finding is in agreement with that reported by Caturla et al. (1988) who found that the adsorption of phenol is greater than that of 2,4-dichlorophenol. This result is the reverse of that obtained by Juang et al. (1996) who reported that the adsorption of 2,4-dichlorophenol is higher than that of phenol. Such discrepancies probably arise from the differences in pore structure between the ACRH used in this study and the activated carbon fibre prepared by Juang et al. Activated carbon is essentially characterized by a broad pore-size distribution involving micro- (< 2 nm), meso- (2 5 nm) and macro-pores (> 25 nm). It is believed that the major adsorption sites on an activated carbon are located in the micropores rather than the mesopores, where only weaker adsorption is observed (Yang et al. 1993). Adsorption onto an activated carbon must proceed through a sequence of diffusion steps from the bulk phase into the mesopores and then into the micropores. Thus, the adsorption capacity would be lower for an adsorbate with a larger molecular volume (Caturla et al. 1988). Hence, the data listed in Table 1 provide an understanding of the higher adsorption of phenol relative to the other phenolic compounds observed in the present work. Compared to the other phenolic compounds studied, the phenol molecule has a lower molar volume, a lower cross-sectional area (A m ) and a lower effective molecular diameter (σ m ). The next in this sequence is 2,4-dichlorophenol, then 2,6-dichlorophenol

4 16 M.M. Nassar et al./adsorption Science & Technology Vol. 26 No TABLE 1. Some Physical Properties of the Phenolic Compounds Studied Compounds Mol. Solubility Mol. volume A m (nm 2 ) b σ m (nm) b λ m (nm) weight (g/1 g at 25 o C) a (mm 3 /mol) b Phenol ,4-Dichlorophenol ,6-Dichlorophenol ,4-Dinitrophenol a Morrison and Boyd (1983). b Caturla et al. (1988). and finally 2,4-dinitrophenol which has the largest molar volume, cross-sectional area and effective molecular diameter. It should be mentioned that several workers (Caturla et al. 1988; Dvorak et al. 1993; Streat et al. 1995) have reported that the adsorption capacity of activated carbon is higher when the adsorbate molecule has a high molar mass or a smaller solubility in water. According to the observations reported for the present study as well as the data listed in Table 1, the molar mass and molar volume of the adsorbate probably has a greater influence on the adsorption capacity than the solubility of the adsorbate in water Sorption dynamics The sorption mechanism could be controlled by reaction kinetics or by one or more diffusion processes such as film diffusion or pore diffusion. A high agitation speed (22 rpm) was selected for the kinetic and diffusion studies, thereby ensuring that the shear at the particle surface was high. As a result, the thickness of the boundary layer surrounding the particle should be minimal and boundary layer resistance or film diffusion should not be a major rate-controlling factor. Basically, the adsorption process may be described in terms of the following three steps: (a) mass transfer of solute from the bulk solution to the particle surface (external mass transfer); (b) adsorption of solute onto the active sites; and (c) internal diffusion of solute via pore diffusion. Step (b) is considered to be rapid in the establishment of the rate-limiting step and is thus not considered in any kinetic analysis employed in the present work. Consequently, the two rate-limiting steps considered are external mass transfer and intra-particle diffusion. The development of models based on the simultaneous occurrence of two such mass-transfer steps is quite complex and hence simplifying assumptions have been made in our attempts to describe the adsorption process Rate constant study The kinetics of the sorption of phenol and its derivatives onto activated carbon were studied on the basis of the Lagergren pseudo-first-order rate equation, viz. log(q e q t ) = log q e (k ads /2.33)t (1) where q t (mg/g) is the amount of solute sorbed at time t, q e (mg/g) is the amount sorbed at equilibrium and k ads is the equilibrium rate constant for sorption. The straight line plots of log(q e q t ) versus time depicted in Figure 2 for the various solutes studied demonstrate the validity of applying equation (1)

5 Mass Transfer/Adsorption Kinetics of Phenolic Compounds onto Activated Carbon log(q e q t ) Time (h) Figure 2. Lagergren plots for the adsorption of the phenolic compounds studied onto activated carbon. Data points relate to the following:, phenol;, 2,4-dichlorophenol;, 2,6-dinitrophenol;, 2,4-dinitrophenol. TABLE 2. Sorption Kinetic Parameters for the Various Phenolic Compounds Studied Compound k ads (h 1 ) k p [mg/(g h.5 )] Phenol ,4-Dichlorophenol ,6-Dinitrophenol ,4-Dinitrophenol in the present study and indicate that the process followed first-order rate kinetics. The magnitude of k ads for the solute was calculated from the slope of these plots and is listed in Table Single resistance mass-transfer models The kinetics of the adsorption of phenol onto activated carbon prepared from rice husk have been studied using single resistance models, the first being based on external mass transfer and the second on intra-particle diffusion. Such studies concentrated only on phenol due to its high removal rate External mass-transfer model As stated above, resistance to mass transfer occurs mainly in the boundary layer film surrounding the adsorbent particles. Thus, the use of well-agitated vessels to achieve completely mixed systems result in the initial resistance to mass transfer in the layer film being fairly low. As a consequence, the intrinsic adsorption rate is rapid until the external surface of the adsorbent particles is covered by a layer of the adsorbate (phenolic compounds). It is therefore possible to construct a model for determining the external mass-transfer coefficient ( ) from the boundary condition given in equation (2): = V/A p [d(c t /C )/dt] t = (2)

6 162 M.M. Nassar et al./adsorption Science & Technology Vol. 26 No where A p is the total surface area of ACRH exposed to the solution as given by equation (3): A p = 6m/[ρd p (1 ε p )] (3) The values of can be correlated by an equation of the general form: = A(variable) B (4) Including the adsorbent mass and the initial adsorbate concentration in the variables studied allows the constants A and B to be determined. Effect of adsorbent mass The influence of the mass of activated carbon employed in the experiments on the adsorption of phenol is shown by the data depicted in Figure 3. It will be seen that, in all cases, the rate of phenol (solute) uptake increased with increasing adsorbent mass. The values of the corresponding external mass-transfer coefficients,, were determined using equation (2). The external mass-transfer coefficient depends on the driving force per unit area and, since C was maintained at a constant value (5 mg/dm 3 ) in these experiments, increasing the adsorbent mass increased the surface area available for adsorption and hence the rate of adsorbate removal was increased. The resulting plot of log versus log M gave a linear relationship with the data fitting the expression: =.23[M].55 (5) In fact, the magnitude of decreased with increasing activated carbon mass. This effect is probably due to the fact that, with small adsorbent masses, only a small amount of external surface was exposed to phenol molecules and hence the driving force per unit surface area for phenol adsorption was larger. Effect of initial adsorbate concentration The initial concentration of phenol in wastewater is important since a given mass of activated carbon can only adsorb a certain amount of it. Hence, the more concentrated the adsorbate, the smaller the volume of adsorbate that can be purified by a fixed mass of adsorbent. For this reason, experiments to examine the influence of the initial.8 C t /C Time (h) Figure 3. Effect of the mass of activated carbon employed, m (g), on the concentration decay curves for the adsorption of phenol at an initial concentration of the latter (C ) of 5 mg/dm 3. Data points relate to the following:, m =.7 g;, m = 1.3 g;, m = 2.5 g.

7 Mass Transfer/Adsorption Kinetics of Phenolic Compounds onto Activated Carbon C t /C Time (h) Figure 4. Effect of the initial concentration of phenol, C, on the concentration decay curves. The data points relate to the following:, C = 25 mg/dm 3 ;, C = 5 mg/dm 3 ;, C = 1 mg/dm 3. concentration of phenol on the adsorption process were conducted. The experimental results obtained shown in Figure 4 indicate that at high concentrations the linear sections of the curves were close together with the fractional adsorption being low. However, at low initial adsorbate concentrations, the initial uptake of phenol was high. The magnitude of the external mass-transfer coefficients,, were determined as before. On plotting log versus log C, a linear relationship was obtained. The experimental data satisfied the equation: Intra-particle diffusion model =.6[C ].17 (6) In this model, intra-particle diffusion is assumed to be the only rate-controlling step for adsorption, since such diffusion represents 95% of the adsorption capacity. A classification of the pore-size distribution has been proposed by Dubinin (1967) in which the pores are sub-divided into three size ranges: micropores, transitional pores and macropores. The role of pore diffusion as the single rate-controlling step in the present work was tested through the use of the equation of Weber and Morris (1963), viz.: q = k p t.5 (7) This indicates that the effect of intraparticle diffusion on the adsorption process may be assessed via a plot of q t versus t.5 as shown in Figure 5 for different initial concentrations of phenol with a constant adsorbent mass of 2.5 g. It will be seen from the figure that all the plots exhibit the same general feature, i.e. a linear portion extending upwards to a curved plateau. According to Weber and Morris (1963) and Nassar (1997, 1999), such plots actually involve an initial curved portion (not shown) followed by a linear portion and, finally, a plateau. These features may be interpreted by assuming that the linear portions arise from intra-particle diffusion while the final plateau indicates complete saturation of the ACRH with phenol under equilibrium conditions. The sorption rates for intra-particle diffusion, k p, were calculated from the slopes of the linear portions of the respective plots with units of mg/(g h.5 ) (not the true reaction rate, but relative rates which are useful for comparative purposes) and are listed in Table 2. Plots of log k p versus log C were linear, indicating a gradual increase in k p with increasing C and supporting the view

8 164 M.M. Nassar et al./adsorption Science & Technology Vol. 26 No q t (mg/g) Time.5 (h.5 ) Figure 5. Adsorption of phenol onto activated carbon. Plots of q t versus t.5 for various initial concentrations of phenol. Data points relate to the following:, C = 25 mg/dm 3 ;, C = 5 mg/dm 3 ;, C = 1 mg/dm 3. an increase in the adsorbate in the bulk liquid phase leads to an increase in the force driving phenol from the bulk solution onto the solid particles. The data obtained fitted the equation: 3.5. Twin-resistance mass-transfer model k p =.183(C ).478 (8) Pore diffusion model (external mass transfer + pore diffusion) In the present study, the pore diffusion model has been used to predict the theoretical concentration/time curves. This model is based on combined external mass transfer and pore diffusion, and has been developed to simulate the behaviour of a batch adsorber. The method employed for solving the corresponding rate relationship was the exponential curve technique of Provencher and Vogel (198). This is obviously more complex and necessitates a computer program to fit the experimental data and generate a theoretical concentration versus time curve. The experimental and theoretical data are then compared in the form of Sherwood numbers, Sh. The experimental Sherwood number, Sh *, is calculated using the experimental adsorption rate exp and may be evaluated from the relationship: Sh * exp = (Rρ p k pt q e )/(D eff C ) (9) while the theoretical Sherwood number, Sh *, may be obtained from the following equation: theor Sh * theor = (1 C h η)/[(1 η).67 (1 (1/B i ))(1 η)] (1) By comparing Sh *, and exp Sh*, it is possible to assign values for the important mass-transfer theor parameters, liquid-phase mass-transfer coefficient (K f ) and the effective diffusion coefficient (D eff ) for the phenol/activated carbon system. The extent to which the experimental and theoretical results

9 Mass Transfer/Adsorption Kinetics of Phenolic Compounds onto Activated Carbon 165 agree may be determined by calculating the residual of the experimental and theoretical Sherwood numbers: n ( ) = ( ) R 1 Sh Shtheor Sh 2 exp + Sh Sh i = 1 exp exp i 1 ( ) / theor i= / exp where (R ) is the residual, i corresponds to the experimental points and n is the number of experimental points. In the present model, a wide range of respective and D eff values derived from the experimental concentration/time data was supplied together with the initial concentration, C, the adsorbent mass, M, the volume of the solution, V, the adsorbent particle diameter, d p, the adsorbent density, ρ p, and the Langmuir isotherm constants, a L and K L. The program provided the five best-fit curves for the best combination of and D eff based on the minimum value of the residual (R ). The model has been applied to the adsorption of phenol onto activated carbon as studied in the present work. The magnitude of was originally evaluated using the initial system conditions in equation (11) and then a wide range around this original value was fed into the program to give it greater flexibility. Previous authors (McConvey and McKay 1985; Magdy 1992; Spahn and Schlunder 1975) have shown that a single value of is sufficient to describe a range of system conditions. In the present study, the value of was: (11) = (79.75 ± 1.7%) cm/s (12) Similarly, previous authors have demonstrated that a single effective pore diffusion coefficient (D eff ) is sufficient to describe the whole experimental range (McConvey and McKay 1985; Magdy 1992). In the present study, the value of D eff was: D eff = (13) A series of plots were constructed to determine the relationship between the Sherwood number and time for various initial concentrations of adsorbate in the adsorption of phenol onto activated carbon. A typical plot is shown in Figure 6. Generally, the application of this model yields good agreement between the experimental and theoretical results as shown by the data listed in Table 3. n Sh 3 Sherwood number, Sh exp Time (min) Figure 6. Plot of the Sherwood number, Sh * exp, versus time for the phenol/activated carbon system. Experimental conditions: C = 5 mg/dm 3 ; M = 2.5 g.

10 166 M.M. Nassar et al./adsorption Science & Technology Vol. 26 No TABLE 3. Kinetic Parameters, and D eff, for the Adsorption of Phenol onto Activated Carbon C (mg/dm 3 ) M (g) (cm/s) D eff R CONCLUSIONS The following conclusions arise from the work described herein. 1. The adsorption capacities of the phenolic compounds studied follow the sequence: phenol > 2,4-dichlorophenol > 2,6-dichlorophenol > 2,4-dinitrophenol. The molar mass and molar volume probably have a greater influence on the adsorption sequence of phenolic compounds onto ACRH than the corresponding solubilities of the phenolic compounds in water. 2. The external mass-transfer coefficient ( ) for phenol adsorption onto ACRH could be correlated with the mass of ACRH employed (M) and the initial concentration of phenol (C ) via the relationships: =.23[M].55 and =.6(C ) The intra-particle diffusion constant, k p, for phenol adsorption onto ARCH could be correlated with the initial concentration of phenol, C, via the relationship: k p =.183(C ) A twin-resistance mass-transfer model based on external mass transfer and pore diffusion was applied successfully to the system under investigation. The concentration decay curves could be predicted using the modified single mass-transfer coefficient,, and the single effective diffusion coefficient, D eff, for studies involving various adsorbent masses and initial adsorbate concentrations. 5. NOMENCLATURE A p external surface area of solute (cm 2 ) A m particle cross-sectional area (nm 2 ) B i Biot number (= R/D eff ) C initial liquid-phase concentration (mg/dm 3 ) C s saturation concentration (mg/dm 3 ) C constant

11 Mass Transfer/Adsorption Kinetics of Phenolic Compounds onto Activated Carbon 167 C h D eff capacity factor [= (mq *)/VC ] s effective pore diffusion coefficient (cm 2 /s) d p mean particle diameter (mm) k ads equilibrium rate constant (h 1 ) external mass-transfer coefficient (cm/s) k p intra-particle diffusion rate constant [mg/(g h.5 )] k pt internal mass-transfer coefficient m adsorbent mass (g) q e equilibrium solid-phase concentration at time t (mg/g) q* sh hypothetical equilibrium solid-phase concentration R residual of Sherwood number Sh * exp modified experimental Sherwood number Sh * modified theoretical Sherwood number theor t time (s/min/h) V volume of liquid in reactor (dm 3 ) V m volume of gas required to form a monolayer on the adsorbent surface η dimensionless solid-phase concentration ρ p particle density concentration (= q t /q eh ) λ max maximum wavelength (nm) effective molecular diameter (nm) σ m REFERENCES Caturla, F., Martin-Martinez, J.M., Molina-Sabio, M., Rodriguez-Reinoso, F. and Torregrosa, R. (1988) J. Colloid Interface Sci. 124, 528. Dubinin, M.M. (1967) J. Colloid Interface Sci. 23, 487. Dvorak, B.I., Lawler, D.F., Speitel, G.E., Jones, D.L. and Badway, D.A. (1993) Water Environ. Res. 65, 827. Finar, I. L. (1995) Organic Chemistry, 6th Edn, Longman-Pearson Education, Singapore. Juang, R.S., Feng, C.W. and Tseng, R.L. (1996) J. Chem. Eng. 41, 487. Magdy, Y.H. (1992) Ph.D.Thesis, University of El-Minia, Egypt. McConvey, I.F. and McKay, G. (1985) Chem. Eng. Process. 19, 267. Morrison, R.T. and Boyd, R.N. (1983) Organic Chemistry, 4th Edn, Allyn and Bacon, London, p Moura ~ o, P.A.M., Carrott, P.J.M. and Ribeiro Carrott, M.M.L (26) Carbon 44, Nassar, M.M. (1997) Adsorp. Sci. Technol. 15, 69. Nassar, M.M. (1999) Water Sci. Technol. 4, 133. Provencher, S.W. and Vogel, R.H. (198) Math. Biosci. 5, 251. Radhika, M. and Palanivelu, K. (26) J. Hazard. Mater. 138, 116. Spahn, H. and Schlunder, E.U. (1975) Chem. Eng. Sci. 3, 529. Streat, M., Patrick, J.W. and Camporro Perez, M.J. (1995) Water Res. 29, 467. Weber, Jr., W.J. and Morris, Jr., J.C. (1963) J. Sanit. Eng. Div., ASCE 9, 79. Wu, F.-C. and Tseng, R.-L. (26) J. Colloid Interface Sci. 294, 21. Yang, O.B., Kim, J.C., Lee, J.S. and Kim, Y.G. (1993) Ind. Eng. Chem., Res. 32, 1692.