Sei sulla pagina 1di 9

Chemical Engineering and Processing 46 (2007) 477485

Application of lm-pore diffusion model for the adsorption of metal ions on coir in a xed-bed column
S.Y. Quek a, , B. Al-Duri b
a b

Chemistry Department, University of Auckland, Private Bag 92019, Auckland, New Zealand Chemical Engineering, University of Birmingham, Edgbaston, Birmingham, United Kingdom

Received 10 November 2005; received in revised form 30 April 2006; accepted 5 June 2006 Available online 22 September 2006

Abstract The most important criterion in the design of xed-bed adsorption systems is the prediction of column breakthrough or the shape of the adsorption wave front, which determines the operating life-span of the bed and the regeneration time. In this study, the investigation of column breakthrough was carried out using coir, a by-product from coconut processing industry, as adsorbent to remove Pb(II) and Cu(II) from aqueous solutions. A two resistance lm-pore diffusion (FPD) model was applied to predict the concentration proles for various system variables in this study, including metal solution concentration, ow rate and bed height. It was found that the FPD model could be applied to predict the breakthrough kinetics of the column. The systems under investigation (Pb/coir and Cu/coir) could be described by a single effective diffusivity (Deff ) over the operating range of ow rates and initial metal concentrations. The Deff values were (1.31 0.36) 105 cm2 /s and (8.24 1.03) 106 cm2 /s for Pb/coir and Cu/coir systems, respectively. On the other hand, the values of external mass transfer coefcient (c ) were found to increase with decreasing ow rates and remained constant for the variation in initial adsorbate concentration. Biot number was used as an indicator for the intraparticle diffusion. The Biot number was found to increase with increasing ow rate and initial concentration, indicating an increase in intraparticle diffusion resistance. 2006 Elsevier B.V. All rights reserved.
Keywords: Adsorption; Film-pore diffusion model; Fixed-bed column; Coir; Metal ions; Diffusion; Mass transfer; Biot number

1. Introduction Many waste by-products from food industry have been evaluated for their ability to remove metal ions from water in search of cost-effective adsorbents [119]. Coir or coconut husk bre, is a waste by-products from the coconut processing industries. It is abundantly available in the tropical countries and can be obtained free or at a minimal cost. Coir was previously proven to have ability to remove heavy metals from aqueous solution in batch system [20]. However, the information obtained from adsorption isotherms and contact time study in a batch system is useful in determining the effectiveness of the metal-adsorbent system. It is not sufcient to give all the accurate scale-up data required when designing efuent treatment systems which employ adsorption columns. For example, xed-bed columns do not necessarily operate under equilibrium conditions because the

Corresponding author. Tel.: +64 9 373 7599x85852; fax: +64 9 373 7422. E-mail address: sy.quek@auckland.ac.nz (S.Y. Quek).

contact time is not sufciently long for the attainment of equilibrium. Besides, other operational problems such as uneven ow pattern (chanelling) in the column, recycling and regeneration cannot be studied in batch experiments. Therefore, it is necessary to perform ow tests using columns. Mathematical models have been incorporated into adsorption design in order to predict the concentration proles within adsorption columns at a wide range of system conditions. A number of two-resistance models have been presented which described adsorption systems with varying accuracy [21]. These models tend to differ in their description of the adsorption mechanism occurring within the adsorbent particle, assuming either pore diffusion within the liquid-phase, or solid diffusion within the solid phase. In this study, a lm-pore diffusion (FPD) model is applied to investigate the mass transfer process in the xed-bed column. This model has been successfully employed to describe the adsorption of phenol and dyes onto carbon, silica, lignite and pith in both column and batch experiments [2226] as well as the adsorption of cadmium and copper ions onto bone char [15,16]. However, it has not been applied to coir previously.

0255-2701/$ see front matter 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.cep.2006.06.019

478

S.Y. Quek, B. Al-Duri / Chemical Engineering and Processing 46 (2007) 477485

(4) The average concentration in the solid is given by


h q = qe

3 rf R

(4)

By introducing the following dimensionless parameters: = q , h qe kf R , Deff = Ct , Co


h qe m Co V

t Co Deff , h qe R2

Bi =

Ch =

Fig. 1. A conceptual diagram of unreacted core theory.

The adsorption rate for a single particle can be expressed as a function of adsorbate concentration, ; in the adsorbent phase, ; the capacity factor, Ch and of the Biot number, Bi, and is represented by 3(1 Ch )(1 )0.33 d = d 1 1 (1/Bi)(1 )0.33 (5)

2. Theory The lm-pore diffusion (FPD) model of interest in this work was proposed by Spahn and Schl nder [27] and Brauch and u Schl nder [28], based on the unreacted core theory [29,30]. u This model describes the occurance of adsorption by external lm mass transfer followed by intraparticle pore diffusion to the sorption sites where solute molecules (adsorbate) are taken up. The adsorbent particle is regarded as a porous solid. The adsorbate in the solution is adsorbed in a well dened concentration front which starts at the outer surface of adsorbent particles, moving radially inward with a certain velocity, leaving an unreacted zone at the center. A conceptual diagram of the unreacted core theory is shown in Fig. 1. A few assumptions are made for this theory: (i) the transfer of solute molecules within the pores of an adsorbent particle occurred only by molecular diffusion; (ii) adsorption equilibrium occurs between the pore and solute solution and adsorbent surface throughout the particle; (iii) the adsorption is irreversible; (iv) the concentration of solute molecules in the pore water is negligible as compared with that on the adsorbent per unit volume. The main mathematical steps of the FPD model are as outlined below: (1) External mass transfer from the external liquid phase: N(t) = kf A(Ct Ce ) = kf 4R2 (Ct Ce ) (1)

The adsorption rate equation (Eq. (5)) can be incorporated into the xed bed kinetic Eq. (6): (6) = f ()Bi This Eq. (6) is combined with the differential mass balance in the column: + =0 (7) and yields a differential equation for the prediction of the local and time dependent concentration prole in the solid phase: 2 f () f () = 0 f () (8)

A general solution of Eq. (8) was developed by Van Meel [31] and Brauch and Schl nder [28] for the following boundary u conditions: (, 0) (, 0) = 0, =0 Hence, Eq. (8) becomes: 1 +=0 f (n) Integration of Eq. (9) yields the general solution: = f ()

(9)

(2) Diffusion in the liquid-lled pore occurs according to Ficks rst law: N(t) = 4Deff Ce (t) (1/rf ) (1/R) (2)

(10)

(3) The velocity of the concentration front is obtained from the mass balance on a spherical element:
2 h N(t) = 4rf qe

This equation was solved by Brauch and Schl nder [28] by u dividing the integration region into two sections separated by a time, 1 . This time 1 is reached when the rst adsorbent layer ( = 0) at the adsorber entrance is saturated. The mathematical equations leading to the solution were described by McKay [32]. As a summary, the limit for the rst integration region are:

drf dt

(3)
1

(0,t)

(11)

S.Y. Quek, B. Al-Duri / Chemical Engineering and Processing 46 (2007) 477485

479

Hence: (, ) = (, ) (0, ) (12)

3. Materials and method 3.1. Materials Coir was dried in an oven (Gallenkamp, Model OV-160, England) at 105 C in a large size tray for 24 h. It was then screened through a mesh sieve to obtain adsorbents with a particle size range of 5001000 m. Stock solutions (1000 mg/l) of lead and copper were prepared from lead nitrate (Pb(NO3 )2 ) (SigmaAldrich Dorset, England) and hydrated copper sulphate (CuSO4 5H2 O) (Fisons, Loughborough, England) by dissolving the appropriate amounts of metal salts in 1 l distilled water. Solutions of various concentrations were obtained by diluting the stock solution with distilled water to the desired concentration. 3.2. Column experiments The apparatus for column studies are shown schematically in Fig. 2. The adsorption unit is a glass column with inner diameter 2.5 cm and length 25.0 cm. The column was designed in such a way that samples could be taken at various points (i.e. 2.5 cm, 7.5 cm, 12.5 cm, 17.5 cm and 22.5 cm) in the column for the evaluation of breakthrough curves at different bed heights. The column was packed with 14.5 g coir to give a total column height

The limit for the second integration region are:


1

(13)

yielding: (, ) = (, ) (14)

Eq. (12) predicts the liquid phase sorbate concentrations up to time 1 , and Eq. (14) predicts the sorbate concentration after the constant pattern breakthrough curve is fully developed. These mathematical steps have been developed in a Fortran programme which related the solute on adsorbent, , to dimensionless time, 1 , and solute in the efuent, , to dimensionless time, [32]. Since monitoring the solute in efuent is the particular aim of this study, the Fortran programme is adapted to print out the dimensionless solute concentration in efuent Ct /C0 () against dimensionless time, and also the volume treated. The theoretical breakthrough curves can then be compared with experimental curves.

Fig. 2. Schematic diagram of the experimental apparatus.

480

S.Y. Quek, B. Al-Duri / Chemical Engineering and Processing 46 (2007) 477485

of 24.5 cm approximately. The adsorbent was boiled for 5 min prior to the packing process to get rid of air bubbles. A ne mesh and a layer of gravel (about 0.5 cm) were placed at the bottom of the adsorbent bed to support the bed while another layer of gravel was placed on the top of the adsorbent bed to prevent carry over of sorbent particles. The column was then sealed with grease and a glass stopper (contains ne mesh at the end). During the experiments, the column was fed continuously with a metal solution (either Pb or Cu) which was kept in a constant head inuent tank (60 l capacity) using a pump (Stuart Turner Ltd., Oxon, UK). The pH of metal solutions was adjusted to 4.5 and pH 5.0 for lead and copper respectively, prior to the experiments. The ow rates were regulated by a rotameter (Model OMM 1037 NGT, CT Platon Ltd., Basingstoke, UK) and the overow from the rotameter was collected back to the inuent tank. Up-ow conditions were chosen to facilitate the accurate control of adsorbate ow rate. The column was operated at four different ow rates ranging from 25 ml/min to 75 ml/min for copper solutions and 50200 ml/min for lead solutions. The effect of adsorbate concentration was studied using initial adsorbate concentrations ranging from 10 mg/l to 50 mg/l for copper solutions and from 25 mg/l to 100 mg/l for lead solutions. Samples (3 ml) were withdrawn using syringes with needles in certain intervals of time at different bed heights and then ltered instantly through 0.45 m cellulose acetate membrane lter (Whatman). The metal concentrations of the samples were analysed using atomic absorption spectrophotometer (ATI UNICAM Model 939, Cambridge, England) with an acetyleneair ame. 4. Results and discussion 4.1. Effect of bed height The adsorption column used in this study was designed to contain sampling points at 5 cm interval, to facilitate the study of the effect of bed height on the life-span of the column. Typical breakthrough curves for different bed heights are illustrated in Fig. 3 for Pb/coir system as plots of dimensionless concen-

tration versus the volume of inuent treated. The resulting plot showed the typical S shape of a packed-bed adsorption system with an initial period of minimal solute, followed by gradual breakthrough that would slowly reach the feed metal concentration. This ow pattern was repeated over the entire duration of the experiment for different bed heights. The S shape prole for column adsorption is generally associated with adsorbate of smaller molecular weight and simple structure and is not observed in larger molecular such as dyes [3336]. This ow prole has been observed previously by other researchers for the adsorption of Cu(II), Ni(II) and Cd(II) onto banana pith and moss [10,37]. Results indicate that an increase in bed height increases the breakthrough volume and hence the breakthrough time, resulting in longer service time. This is because of the increase in the amount of adsorbent packed and therefore in the binding sites with increasing bed height. From Fig. 3, an increase of about 70% in breakthrough volume was obtained at 50% breakthrough upon increasing the bed height from 7.5 cm to 12.5 cm at a ow rate of 75 ml/min for Pb/coir system. Similarly, the service time was extended from 53.33 min to 90.67 min. 4.2. Effect of ow rate Fig. 4 shows the plot of breakthrough curves at four different ow rates for Pb/coir system, at a xed bed height of 12.5 cm and initial metal concentration of 50 mg/l. The shape of the breakthrough curve would indicate the internal resistance within the column and the relative effects of mass transfer parameters throughout the operating conditions. It is observed that at higher ow rates, the breakthrough point occurred earlier, indicating a shorter column life. Also, the breakthrough curves were sharper for higher ow rates, implying higher intraparticle diffusion effect and a smaller mass transfer zone [28]. The breakthrough curves were atter for lower ow rates indicating a more prominent effect of lm transfer resistance, larger mass transfer zone and longer service time for the column. This is reasonable because at higher ow rates the boundary layer around the particles is thinner, which reduces external mass transfer

Fig. 3. Typical breakthrough curves for Pb/coir system (ow rate = 75 ml/min, initial metal concentration = 50 mg/l).

Fig. 4. Breakthrough curves for Pb/coir at four different ow rates (initial Pb(II) concentration = 50 mg/l, bed height = 12.5 cm).

S.Y. Quek, B. Al-Duri / Chemical Engineering and Processing 46 (2007) 477485 Table 1 Effect of ow rates on the volume of wastewater treated when Ct /C0 = 0.2 (20% breakthrough) at a bed height of 12.5 cm Pb/coir Flow rate (ml/min) 50 75 100 200 Volume treated (l) when Ct /C0 = 0.2 6.3 5.7 5.4 5.2 Time (min) 125.0 75.0 54.2 24.0 Cu/coir Flow rate (ml/min) 20 35 50 75 Volume treated (l) when Ct /C0 = 0.2 2.9 2.7 2.5 2.5

481

Time (min) 146.6 77.0 52.5 35.0

(lm) resistance. The opposite is true for systems with lower ow rates. The shape of breakthrough curves for Cu/coir system show similar trend as those of Pb/coir system. The effect of lm and intraparticle resistance and both external mass transfer coefcient (c ) and intraparticle diffusion (Deff ) will be evaluated in Section 4.4. Results also indicate that for a xed bed height, an increase in ow rate decreases the volume treated until breakthrough (Fig. 10). However, this decrease was not very signicant (Table 1). For example, a breakthrough volume (for 20% breakthrough) increases from 5.2 l to 6.3 l was observed on changing ow rate from 200 ml/min to 50 ml/min, for Pb/coir system, at a xed bed height of 12.5 cm. However, the time for the breakthrough volume was increased from 24 min to 125 min. The same results were also observed for the Cu/coir system. This suggests that the adsorption of these metals on the column is rapid and the use of slower ow rates do not signicantly improve the efciency of the column performance. 4.3. Effect of initial metal concentration Fig. 5 shows the breakthrough curves for various initial solute concentrations of Pb(II) at a ow rate of 50 ml/min and bed height of 12.5 cm. The difference in shape observed in the breakthrough curves would be attributed to the variation in column resistance and adsorption driving forces because these systems have similar ow rates, i.e., same hydraulic loading. At lower concentrations, the breakthrough curves were atter, indicating a relatively wider mass transfer zone and lm controlled process.

On the contrary, the breakthrough curves were sharper at higher concentrations, implying a relatively smaller mass transfer zone and a more intraparticle diffusion controlled process. Clearly, the volumes of the metal solution treated decreased (Fig. 5) and hence the service times were shorter for the systems of higher sorbate concentrations because of the high adsorbate concentrations saturating the adsorbent more quickly. The same trend was also observed for the Cu/coir system. 4.4. Film-pore diffusion (FPD) model tting The Fortran computer program that solves the FPD model required system parameters as input. These include the inuent ow rate, initial metal concentration, adsorbent particle size and density, liquid viscosity, column dimensions, equilibrium capacity of the adsorbent, adsorbate molecular diffusivity and effective pore diffusivity. The molecular diffusivities (Dmol ) were initially determined using the WilkeChang correlation [38] as follows: Dmol = 7.4 108 (xa M)0.5 T 0.6 VM (15)

where xa is the association parameter, M the molecular weight of solute (g/mol), T the temperature (K), the viscosity of solution (cP) and VM is the molal volume of solute at normal boiling point (ml/g mol). WilkeChang correlation has an average error of 10%, as estimated by Skelland (1974) [32]. In this work, Dmol values of 1.146 105 cm2 /s and 9.00 106 cm2 /s has been estimated for lead(II) solution and copper(II) solution, respectively. The equilibrium capacities are obtained from the Langmuir isotherm because the FPD model is based on irreversible adsorption. Mathematically, the Langmuir isotherm is the closest isotherm to rectangular shape of isotherm which implied irreversible adsorption. Results have shown that the Langmuir isotherm could be applied satisfactory to the two systems under investigation, namely, Pb(II)/coir and Cu(II)/coir, where the monolayer capacities for Pb(II)/coir system and Cu(II)/coir system were 48.84 mg/g and 19.30 mg/g, respectively [20]. The value of the mass transfer coefcient, c , in the liquid phase in the column is computed by the program using a correlation proposed by Carberry [39] and Kataoka et al. [40]: c = 1.15 Dmol dp dp 0
1/2

Fig. 5. Breakthrough curves for Pb/coir system for four initial metal concentrations (ow rate = 50 ml/min, bed height = 12.5 cm).

Dmol

1/3

(16)

482

S.Y. Quek, B. Al-Duri / Chemical Engineering and Processing 46 (2007) 477485

Fig. 6. Application of the FPDM to various bed heights for Pb/coir system at a ow rate of 50 ml/min and initial metal concentration of 25 mg/l.

Fig. 8. Application of the FPDM to various ow rates at a bed height of 12.5 cm for Pb/coir system.

where Dmol is the molecular diffusivity, dp the diameter of adsorbent particle, the density of adsorbent, 0 the initial solute solution velocity and is the kinematic viscosity. The rate of intraparticle diffusion is described by the effective pore diffusivity (Deff ), which is obtained by tting the theoretical plots to experimental data, by trial and error, using Deff values from the literature as guideline. Figs. 6 and 7 show the application of the FPD model to Pb(II)/coir and Cu(II)/coir systems at various bed heights. The FPD model tting of Pb/coir system at different ow rates and initial metal concentrations is shown in Figs. 810. All gures are illustrated as plots of dimensionless concentration versus service time except Fig. 10, which is plotted to show the variation of dimensionless concentration with the volume of metal solution treated. Generally, good agreement was obtained between the experiment and theory. This implied that the FPD model could be used to predict the breakthrough kinetics of these adsorption systems. However, in some cases, especially for Cu/coir system, the theoretical simulation could not t the experimental curves well at high Ct /C0 value. This deviation is possibly due to the incomplete attainment of equilibrium in the column and also some reversibility of the adsorption system as previously reported by

Fig. 9. Application of the FPDM to various initial Pb(II) concentrations at a bed height of 12.5 cm.

McKay and Bino [22]. It was found that the solid phase equilibrium adsorption capacity value tting the breakthrough curves was 15.5 mg/g for Cu/coir system even though this value was 19.30 mg/g when calculated from the Langmuir isotherm.

Fig. 7. Application of the FPD model to various bed heights for Cu/coir system at a ow rate of 50 ml/min and initial concentration of 10 mg/l.

Fig. 10. Application of the FPDM to various ow rates at a bed height of 12.5 cm for Pb/coir system: plot of Ct/Co vs. volume treated (l).

S.Y. Quek, B. Al-Duri / Chemical Engineering and Processing 46 (2007) 477485 Table 2 The value of c , Deff and Bi for the systems under investigation System Flow rate (ml/min) c (103 cm/s) Deff (105 cm2 /s) Bi

483

(a) Effect of ow ratea Pb coir 50 75 100 200 Cu coir 20 35 50 75 C0 (mg/l)

6.71 8.22 9.50 13.4 3.60 4.77 5.70 6.89 c (103 cm/s) concentrationb 6.71 6.71 6.71 6.71 5.70 5.70 5.70 5.70

1.00 1.25 1.45 1.65 0.70 0.84 0.90 1.00 Deff (105 cm2 /s) 2.00 1.00 0.95 0.90 1.00 0.90 0.75 0.68

9.73 9.83 9.86 12.2 7.72 8.52 9.50 10.5 Bi 5.03 9.73 10.6 11.2 8.55 9.50 11.4 12.6 Fig. 11. Correlation between ow rate and Biot number.

System

(b) Effect of initial metal Pb/coir 25 50 75 100 Cu/coir 10 25 35 50

a b

C0 for Pb/coir and Cu/coir systems were 50 mg/l and 25 mg/l, respectively. Flow rates for both Pb/coir and Cu/coir systems were 50 ml/min.

Table 2 gives the values of the mass transfer parameters, namely the external mass transfer coefcient, c (cm/s), and intraparticle diffusivity, Deff (cm2 /s), which describe the systems under investigation. Deff was found to increase with increasing ow rate and decreasing initial metal concentration. However, for the two present systems, a single Deff could be adequately employed to describe each system, over the operating range of ow rates and adsorbate concentrations. These value were (1.31 0.36) 105 cm2 /s and (8.24 1.03) 106 cm2 /s for Pb/coir and Cu/coir system, respectively. The same observations have been reported by other researchers. Using the FPD model for column studies, Murray and Allen [23] had reported the value of 1.6 107 cm2 /s to 3.0 107 cm2 /s for the adsorption of basic red 22 onto lignite; while McKay and Bino [22] found the value of 5.0 107 cm2 /s for the adsorption of astrazone blue on silica and 2.5 105 cm2 /s for the adsorption of phenol on carbon. On the other hand, Buck and Al-Duri [24] reported the Deff values of (2.5 0.25) 107 cm2 /s and (4.5 0.50) 107 cm2 /s for the adsorption of basic blue 41 and basic red 46 onto carbon. As a comparison, the Deff values obtained from the current study are close to the value for phenol/carbon system but much larger than the adsorption of dyes. This is probably attributed to the relatively smaller diameters of both phenol and metal ions compared to dyes molecules, the factor that gives them much higher diffusivity values. However, when compared with the results of Ko et al. [16] who studied sorption of copper and cadmium ions onto bone char, the Deff values obtained from this study were higher. This means that Deff values are specic to each different adsorbate/adsorbent system. The external mass transfer coefcient, c , was found to vary with different ow rates but to remain constant for the adsor-

bate concentration ranges under investigation. This is in good agreement with the results of Ko et al. [16]. The values of c were 6.71 103 cm/s and 5.70 103 cm/s for Pb/coir and Cu/coir system, respectively. c increased with increasing the ow rates because the turbulence increased in the column causing reduction in the lm resistance as the boundary layer around the particles became thinner. For xed-bed adsorption, lm mass transfer would have a more pronounced effect than in batch adsorption. This is because of the presence of the boundary layer that would be prominent due to the nature of xed-bed adsorption, and thus would enhance the effect of lm mass transfer. The external mass transfer coefcient, c , remained constant for different concentration ranges because the hydraulic loading was constant (i.e., 50 ml/min for both Pb/coir and Cu/coir systems) for all concentration ranges. The Fortran program has also been adapted to calculate the Biot number (Bi). Bi measures the ratio of internal to external mass transfer resistances within the column. An increased Bi value would indicate an increase of intraparticle diffusion resistance. Therefore, Bi would approach for an intraparticle diffusion controlled process, and approach 0 for a lm diffusion controlled process [28]. The Bi values for the systems under investigation are shown in Table 2. It is observed that Bi, slowly increase with the increased of ow rates (Fig. 11). This showed that as ow rate increased, lm resistance decreased and the process would become more intraparticle diffusion controlled, hence breakthrough curves became sharper as observed earlier (Figs. 4 and 8). However, the residence time would decrease with increasing ow rates and the bed service time became shorter. At the same time, the volume of metal solution treated until breakthrough, i.e., the liquid removal efciency, would decrease as ow rate increased (Fig. 10). In contrast, a decrease in ow rate lead to a longer column life and an increase in liquid removal efciency, provided that the bed length and adsorbent particle size were constant. Nevertheless, this does not mean that the bed performance will improve much because lm transfer resistance

484

S.Y. Quek, B. Al-Duri / Chemical Engineering and Processing 46 (2007) 477485

Ce Ch Ct C0 dp Deff Dmol kf M qe h qe q(t) r r2 rf R t T VM x

Fig. 12. Correlation between initial concentration and Biot number.

will increase and the liquid removal efciency will not improve very signicantly, as discussed in Section 4.2. Therefore, for treating large volume of efuent, it is more efcient to operate a few columns in parallel with moderate ow rate (i.e., 75 ml/min or 50 ml/min), rather than using one column with very high ow rate. In addition, applying very high ow rate will increase the pumping cost. Bi was also found to increase with the increase of initial metal concentration (Table 2, Fig. 12). This indicated a more intraparticle controlled process and a shorter bed service time as initial metal concentration increase. Therefore, if the initial metal concentration of the waste water is high, column with bigger diameter and longer length may be used. Alternatively, a few columns operating in parallel can be applied. Overall, the system under current investigation is more suitable for treating low metal concentration wastewater in small volume using moderate ow rates. 5. Conclusions The theoretical mass transfer model based on external mass transfer and pore diffusion was successfully applied to the Pb/coir and Cu/coir systems. Breakthrough curves were predicted using a single effective pore diffusivity (Deff ) for each system for the studies involving various bed heights, ow rates and initial metal concentrations. The external mass transfer coefcients (c ) increased with increasing ow rate but remained constant at different initial concentrations. The Biot number increased with increasing ow rate and initial concentration, indicating an increase in intraparticle diffusion resistance. Appendix A. Nomenclature

equilibrium liquid-phase concentration of adsorbate (mg/l) h capacity factor in FPDM (qe m/C0 V ) liquid-phase concentration of adsorbate at time t (mg/l) initial liquid-phase concentration of adsorbate (mg/l) mean particle size (cm, m) effective diffusivity (cm2 /s) molecular diffusivity (cm2 /s) external mass transfer coefcient (cm/s) molecular weight of solute (g/mol) equilibrium solid-phase concentration (mg/g) hypothetical concentration of solute on adsorbent at equilibrium (mg/g) mean concentration of solute on adsorbent (mg/g) radial co-ordinate (cm) correlation coefcient radius of the unreacted zone at the particle centre (cm) radius of particle (cm) time (min, s) temperature ( C, K) molal volume of solute at normal boiling point (ml/g mol) association parameter in WilkeChang correlation of Dmol

Greek symbols dimensionless column bed height c external mass transfer coefcient in FPD model (cm/s) particle porosity q/qe , dimensionless solid phase concentration viscosity of solution (cP) 0 solute solution velocity (m/s) kinematic viscosity (Stoke or cm2 /s) particle density (solid phase) (g/cm3 ) dimensionless time Ct /C0 , dimensionless liquid phase concentration References
[1] G. Macchi, D. Marani, G. Tirivanti, Uptake of mercury by exhausted coffee grounds, Environ. Technol. Lett. 7 (1986) 431444. [2] R. Suemitsu, R. Uenishi, I. Akashi, M. Nakano, The use of dyestuff-treated rice hulls for removal of heavy metals from wastewater, J. Appl. Polym. Sci. 31 (1986) 7583. [3] W.T. Tan, A.R.M. Khan, Removal of lead, cadmium and zinc by waste tea leaves, Environ. Technol. Lett. 9 (1988) 1223 1232. [4] E. Maranon, H. Sastre, Heavy metal removal in packed beds using apple waste, Bioresour. Technol. 38 (1991) 3943. [5] F.E. Okieimen, E.U. Okundia, D.E. Ogbeifun, Sorption of cadmium and lead ions on modied groundnut (arachis hypogea) husks, J. Chem. Technol. Biotechnol. 51 (1991) 97103. [6] K.S. Low, C.K. Lee, K.P. Lee, Sorption of copper by dye-treated oil-palm bers, Bioresour. Technol. 44 (1993) 109112. [7] W.E. Marshall, E.T. Campagne, W.J. Evans, The use of rice milling by products (hulls and bran) to remove metal ions from solution, J. Environ. Sci. Health 28 (1993) 19771992. [8] C. Namasivayam, K. Periasamy, Bicarbonate-treated peanut hull carbon for mercury (ii) removal from aqueous solution, Water Res. 27 (1993) 16631668.

A Bi

total surface area of an adsorbent (m2 /g) c R/Deff , Biot number

S.Y. Quek, B. Al-Duri / Chemical Engineering and Processing 46 (2007) 477485 [9] W.T. Tan, S.T. Ooi, C.K. Lee, Removal of chromium (vi) from solution by coconut husk and palm pressed bres, Environ. Technol. 14 (1993) 277282. [10] K.S. Low, C.K. Lee, A.C. Leo, Removal of metals from electroplating wastes using banana pith, Bioresour. Technol. 51 (1995) 227231. [11] V. Chui, K. Mok, C. Ng, R. Luong, K. Ma, Removal and recovery of copper(II), chromium(III), and nickel(II) from solutions using crude shrimp chitin packed in small columns, Environ. Int. 22 (1996) 463. [12] K. Periasamy, C. Namasivayam, Removal of copper(II) by adsorption onto peanut hull carbon from water and copper plating industry wastewater, Chemosphere 32 (1996) 769. [13] W.T. Tan, C.K. Lee, K.L. Ng, Column studies of copper(II) and nickel(II) ions sorption on palm pressed bres, Environ. Technol. 17 (1996) 621. [14] Y. Sa , Y. Aktay, Mass transfer and equilibrium studies for the sorption of g chromium ions onto chitin, Process Biochem. 36 (2000) 157173. [15] C.W. Cheung, C.K. Chan, J.F. Porter, G. McKay, Film-pore diffusion control for the batch sorption of cadmium ions frm efuent onto bone char, J. Colloid Interf. Sci. 234 (2001) 328336. [16] D.C.K. Ko, J.F. Porter, G. McKay, Film-pore diffusion model for the xedbed sorption of copper and cadmium ions onto bone char, Water Res. 35 (2001) 38763886. [17] M. Ahmedna, W.E. Marshall, A.A. Hussieny, R.M. Rao, I. Goktepe, The use of nutshell carbons in drinking water lters for removal of trace metals, Water Res. 38 (2004) 10621068. [18] B.S. Inbaraj, N. Sulochana, Carbonised jackfruit peel as an adsorbent for the removal of Cd(II) from aqueous solution, Bioresour. Technol. 94 (2004) 4952. [19] T. Tokimoto, N. Kawasaki, T. Nakamura, J. Akutagawa, S. Tanada, Removal of lead ions in drinking water by coffee grounds as vegetable biomass, J. Colloid Interf. Sci. 281 (2005) 5661. [20] S.Y. Quek, B. Al-Duri, D.A.J. Wase, C.F. Forster, Coir as a biosorbent of copper and lead, Trans. IChemE 76 (part B) (1998) 5054. [21] B. Al-Duri, Adsorption modelling and mass transfer, in: G. McKay (Ed.), Use of Adsorbents for the Removal of Pollutants from Wastewaters, CRC Press, Inc., Boca Raton, 1996. [22] G. McKay, M.J. Bino, Application of two resistance mass transfer model to adsorption systems, Chem. Eng. Res. Des. 63 (1985) 168174. [23] M. Murray, S.J. Allen, Decolourisation of efuent by adsorption onto lgnite in xed beds, in: Proceedings of the 1995 IChemE Research Event/First European Conference, vol. 1, 1995, pp. 225227.

485

[24] G.L. Buck, B. Al-Duri, Modelling multicomponent adsorption of basic dyes onto granular activated carbon in a xed-bed column, in: Proceedings of the 1997 Jubilee IChemE Research Event, vol. 1, 1997, pp. 161164. [25] C.W. Hui, B. Chen, G. McKay, Contact time optimisation of two-stage batch adsorber systems using the modied lm-pore diffusion model, Chem. Eng. Sci. 57 (2002) 28632873. [26] B. Chen, C.W. Hui, G. McKay, Film-pore diffusion modelling and contact time optimization for the adsorption of dyestuffs on pith, Chem. Eng. J. 84 (2001) 7794. [27] H. Spahn, E.U. Schl nder, The scale-up of activated carbon columns for u water purication, based on results from batch tests. I, Chem. Eng. Sci. 30 (1975) 529537. [28] V. Brauch, E.U. Schl nder, The scale-up of activated carbon columns for u water purication, based on results from batch tests. II, Chem. Eng. Sci. 30 (1975) 539548. [29] O. Levenspiel, Chemical Reaction Engineering, Wiley, New York, 1972. [30] S. Yagi, D. Kunii, Effective thermal conductivities in packed beds, Chem. Eng. Sci. 16 (1961) 372. [31] D.A. van Meel, Adiabatic convection batch drying with recirculation of air, Chem. Eng. Sci. 9 (1958) 3644. [32] G. McKay, Analytical solution using a pore diffusion model for a pseudoirreversible isotherm for the adsorption of basic dye on silica, AIChE J. 30 (1984) 692697. [33] G. McKay, S.J. Allen, A mathematical model for the xed bed adsorption of telon blue dye on peat, J. Sep. Proc. Technol. 4 (1983) 815. [34] G. McKay, S.J. Allen, I.F. McConvey, H.R. Walters, External mass transfer and homogeneous solid-phase diffusion effects during the adsorption of dyestuffs, Ind. Eng. Chem. Process Des. Dev. 23 (1984) 221226. [35] G.M. Walker, B. Al-Duri, Acid dye adsorption using xed gac columns, in: Proceedings of the 1994 IChemE Research Event, vol. 2, London, 1994, pp. 109110. [36] G.M. Walker, L.R. Weatherley, Adsorption of acid dyes onto granular activated carbon on xed beds, Water Res. 31 (1997) 20932101. [37] K.S. Low, C.K. Lee, Cadmium uptake by the moss, Calymperes delessertii Besch, Bioresour. Technol. 38 (1991) 16. [38] C.R. Wilke, P. Chang, Correlation of diffusion coefcients in dilute solutions, AIChE J. 1 (1955) 264. [39] J.J. Carberry, Equilibrium and kinetics in the N-oxidesH2 O systems, AIChE J. 6 (1960) 460. [40] T. Kataoka, H. Yoshida, K. Ueyama, J. Chem. Eng. Jpn. 5 (1972) 132.

Potrebbero piacerti anche