Removal of phenol from aqueous solutions by adsorption

Removal of phenol from aqueous solutions by adsorption

Journal of Environmental Management 70 (2004) 157–164 www.elsevier.com/locate/jenvman Removal of phenol from aqueous solutions by adsorption Nadia Ro...

175KB Sizes 7 Downloads 142 Views

Journal of Environmental Management 70 (2004) 157–164 www.elsevier.com/locate/jenvman

Removal of phenol from aqueous solutions by adsorption Nadia Roostaei, F. Handan Tezel* Department of Chemical Engineering, University of Ottawa, 161 Louis-Pasteur, Ottawa, Ont., Canada K1N 6N5 Received 18 June 2003; revised 24 October 2003; accepted 2 November 2003

Abstract Experiments have been conducted to examine the liquid-phase adsorption of phenol from water by silica gel, HiSiv 3000, activated alumina, activated carbon, Filtrasorb-400, and HiSiv 1000. Experiments were carried out for the analysis of adsorption equilibrium capacities and kinetics. The adsorption isotherm model of the Langmuir – Freundlich type was the best to describe adsorption equilibrium data for phenol for the adsorbents studied. Results of kinetic experiments indicated that HiSiv 1000 had the highest rate of adsorption among the adsorbents studied and therefore more detailed studies were carried out with this adsorbent. The influence of particle size, temperature, and thermal regeneration on adsorption of phenol by HiSiv 1000 was evaluated. From particle size experiments it appeared that adsorption capacity of HiSiv 1000 did not change by changing the particle size, but the rate of adsorption decreased considerably by increasing the particle size. The effect of temperature on adsorption was studied by determining equilibrium isotherms for HiSiv 1000 at 25, 40, and 55 8C. The results showed that adsorption capacity decreased with increasing temperature. Thermal regeneration of HiSiv 1000 was performed at 360 8C. It was observed that adsorption capacity of HiSiv 1000 did not change after 14 regeneration cycles. Equilibrium experiments showed that the adsorption capacities of activated carbon and Filtrasorb-400 were several times higher than that of HiSiv 1000. q 2003 Elsevier Ltd. All rights reserved. Keywords: Phenol adsorption; Removal of phenol; Liquid adsorption; Activated carbon; ZSM-5 zeolite; Zeolite Y

1. Introduction Phenols and related compounds are toxic to humans and aquatic life; create an oxygen demand in receiving waters. By increasing the industrial wastewaters, the demand for removal of organic compounds including phenol has been increased. Adsorption of phenol by different adsorbents has been investigated to find the relation between adsorption capacity and adsorbent characteristics such as surface area, and pore size distribution for separation applications in the drinking water concentration range. An excellent review of carbon materials as adsorbents in aqueous solutions is given by Radovic et al. (2000). Peel and Benedec (1980) examined the adsorption of phenol from aqueous solution by Filtrasorb-400. Although adsorption took up to 3 weeks to reach equilibrium, up to 80% of adsorption equilibrium capacity was reached in the first few hours, and the remaining capacity was utilized very slowly. Hsieh and Teng (2000) studied the liquid-phase adsorption of phenol onto activated carbons prepared with different * Corresponding author. Tel.: þ 1-613-562-5800; fax: þ1-613-562-5172. E-mail address: [email protected] (F.H. Tezel). 0301-4797/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.jenvman.2003.11.004

activation levels. Viraraghavan and Alfaro (1992) investigated the adsorption of phenol by peat, fly ash, and bentonite. The capacities of these adsorbents were much lower than that of activated carbon. Batch kinetic studies showed that an equilibrium time of 16, 15, and 16 h was needed for the adsorption of phenol by peat, fly ash, and bentonite, respectively. Isotherm experiments showed that phenol removal by peat, fly ash, and bentonite were 46.1, 41.6, and 42.5%, respectively, from initial concentration of approximately 1000 mg/l. It has been observed that increasing the temperature enhanced the adsorption capacity of phenol by some adsorbents while the opposite phenomenon occurred for other adsorbents. Zogorski and Faust (1978) reported that adsorptive capacity of phenol on granular activated carbon was increased with decreasing temperature. Vidic and Suidan (1991) observed the same phenomenon. They investigated the effect of dissolved oxygen on the capacity of phenol adsorption by granular activated carbon at 21 and 35 8C. However, Singh and Rawat (1994) found that sorption of phenol by fly ash was increased as temperature increased in the range of 30 – 50 8C. The effect of temperature on rate of adsorption always showed that kinetic rate increased with increasing

158

N. Roostaei, F.H. Tezel / Journal of Environmental Management 70 (2004) 157–164

temperature. As a diffusion-limited process, the rate of adsorption of the organics from solution increases as the temperature of the system is increased. Zogorski et al. (1976) showed that raising the temperature from 10 to 30 8C increases the removal rate of phenol by 21%. Effects of temperature, as well the effect of particle size were studied by Leyva-Ramos et al. (1999). Studies of thermal regeneration have focused on the optimum condition to restore the maximum adsorbent capacity and to retain, as much as possible, the original pore structure of the adsorbent (Moreno-Castilla et al., 1995). Cooke and Lables (1994) reported that during thermal regeneration of granular activated carbon, physically adsorbed phenol reacts with oxygenated groups of the carbon, and is converted into chemically adsorbed phenol. Between 600 and 900 K (327 and 627 8C), the chemically adsorbed phenol decomposes and is released in the form of light gases. Above 900 K (627 8C), the chemically adsorbed phenol is degraded into condensation products and graphite. Waer et al. (1992) have investigated fluidized-bed regeneration of coal-based fiber activated carbon loaded with naturally occurring organic material and methylene blue. The capacity of adsorption was returned to that of virgin carbon at a temperature of 850 8C in 15 min. Shorter time and lower temperature for regeneration resulted in a loss of adsorptive capacity. van Vliet and Venter (1984) found that the optimum regeneration condition for spent granular activated carbon (came form GAC contactor treating secondary effluent) was 8-min oxidation at 800 8C in an infrared furnace. They were able to restore only 70% of the virgin micropore volume (on a unit bed volume basis) after the fourth regeneration cycle. The main disadvantages of the thermal regeneration of the activated carbon were the loss of carbon (5–10% per cycle) due to oxidation and attrition, and the energy cost of heating (Cen, 1994). Chiang et al. (1997) have studied the comparison of chemical and thermal regeneration of aromatic compounds on exhausted activated carbon. The objective of this study was to screen various adsorbents for potential application for the removal of phenol for drinking water concentration range. The adsorbents used in this study are silica gel, activated alumina, HiSiv 3000, activated carbon, Filtrasorb-400, and HiSiv 1000. Kinetic experiments were

conducted at 25 8C, to determine the rate of adsorption for each adsorbent. Then the adsorbents with higher rate of adsorption were chosen and used in equilibrium experiments to find the capacity of adsorption for each of them. Effects of temperature, particle size and regeneration on adsorption of phenol by HiSiv 1000 were also investigated. 2. Materials and methods Adsorbents that are used in this study are listed in Table 1. Activated carbon and powder adsorbents such as activated alumina, silica gel, and HiSiv 1000, and HiSiv 3000 were used as is. Filtrasorb-400, which had a large range of particle size, was sieved through the mesh (14 £ 20). To investigate the effect of particle size on HiSiv 1000, pellet adsorbent was crushed and sieved through the meshes (20 £ 50) and (50 £ 70) to obtain smaller particle sizes. Pre-treatment of each adsorbent was performed at 360 8C in air atmosphere for 16 h to remove any organic contamination, which might exist in the adsorbent pores. All the adsorbents were stored in airtight desiccators until they were used. Reagent grade phenol (99.99% from Aldrich chemical company) was used to prepare phenol solutions in water. Deionized distilled water filtered through Millipore purification system (Zenopure corporation, Quatra 90 LC), was used for the preparation of the sorbate solutions. 2.1. Kinetic experiments Two hundred milligrams per liter of phenol solution was prepared and transferred to a 2-l glass vessel. The vessel was submerged in a water bath controlled at a predetermined constant temperature by a thermostat (precision ¼ ^0.5 8C). After the temperature of the adsorbate solution was stabilized at a predetermined level, a known weight of adsorbent was immersed in the vessel and agitation was immediately initiated (speed of stirring ¼ 220 rpm). This was considered as time zero for the kinetic experiment. Small-volume liquid samples (6 ml) were withdrawn from the vessel at predetermined time intervals. Then samples were filtered by using glass-fiber filter papers (Pore size ¼ 0.5–1 mm, Whatman company). The concentration of the adsorbate was determined by UV

Table 1 List of adsorbents used in this study Adsorbent

Particle size

Surface area (m2/g)

Name of company

Activated carbon Filtrasorb (F-400) Activated alumina (Acidic) Activated alumina (Basic) Silica gel HiSiv 3000 (Zeolite ZSM-5 structure) HiSiv 1000 (Zeolite-Y structure)

0.8 mm 0.55– 0.75 mm 63–150 mm 63–150 mm 63–200 mm Powder ,100 mm Powder ,100 mm, 50 £ 70 mesh (212–297 mm), 20 £ 50 mesh (297–850 mm), Extrude (1.5 mm)

858 900

Sigma-Aldrich, Canada Calgon, Mississauga, Ont., Canada Fisher Scientific, Canada Alcan Chemical, Brockville, Ont., Canada Fisher Scientific, Canada UOP Molecular Sieve, USA UOP Molecular Sieve, USA

403 355

N. Roostaei, F.H. Tezel / Journal of Environmental Management 70 (2004) 157–164

spectrophotometer (Backman, Model: DU640), however, the concentration of the samples containing fine particles or colorful pigments were determined by TOC analyzer (Rosemount Analytical, Inc., Model DC 190). Then the relation between the adsorbate concentration in the liquid phase and the contact time was determined to generate a concentration decay curve. 2.2. Equilibrium experiments An adsorption isotherm describes the relationship between the amount of adsorbate that is adsorbed on the adsorbent and the concentration of dissolved adsorbate in the liquid at equilibrium. Accurately weighed portions of adsorbent were placed into a series of 125-ml bottles. Different amount of adsorbent was used for different bottles to determine the isotherm. After the addition of 75 ml of phenol solution, bottles were sealed with glass stoppers. Initial concentration of phenol in liquid phase was 200 mg/l for all the bottles. The bottles were placed in an air-bath shaker (Eberbach Corporation, Ann Arbor, Michigan) until they reached equilibrium. The speed of shaker was adjusted to 180 rpm, to mix the solutions well. Included in each set of bottles, were two control bottles. One of the control bottles contained just the phenol solution with no adsorbent, and the other one contained just water and adsorbent with no phenol in the solution. The first control bottle was used to check for phenol volatilization and/or adsorption onto the walls of the container during the equilibration period. The second control bottle was used to check the existence of organic contamination in adsorbent or in deionized distilled water. Equilibrium times for different adsorbents were determined from the kinetic studies as follows: 1.5, 19 and 30 days for HiSiv 1000, activated carbon and F-400, respectively. The bottles were then taken off the shaker and the suspensions were left standing for a while to allow the adsorbent particles to settle. Ten milliliter of the supernatant was removed from the bottles with syringe and filtered through glass-fiber filter paper to remove any remaining adsorbent particles. Two milliliter of the sample was filtered and discarded before filtrate samples were taken for analysis, to minimize the adsorption of the substrate on filters Equilibrium experiments were conducted at room temperature (25 ^ 1 8C) for all of the adsorbents.

159

3. Results and discussion 3.1. Kinetic experiments Kinetic experiments were performed for silica gel, activated alumina (basic and acidic base), HiSiv 3000, activated carbon, Filtrasorb-400, and HiSiv 1000. The results indicated that silica gel and activated alumina did not show any affinity for phenol adsorption. Silica gel is a partially dehydrated form of polymeric colloidal silicic acid. The chemical composition can be expressed as SiO2, n H2O. Its surface area is in the range of 100 –850 m2/g. The pore size distribution is generally unimodal with the pore ˚ . Activated alumina is a porous size between 1 and 40 A high-area form of aluminum oxide (Al2O3, n H2O), with a strongly polar surface compared to silica gel and has both acidic and basic character. Surface area is in the range of 250 – 350 m2/g. Silica gel and activated alumina are used as desiccants, so their affinity for water molecules is higher than their affinity for phenol molecules. The results of kinetic experiments for HiSiv 3000, HiSiv 1000 (50 £ 70 mesh), activated carbon and Filtrasorb-400 at 25 8C, are shown in Figs. 1 and 2. Concentrations were normalized with respect to the initial concentration and given as ‘Reduced concentration’ in these figures. Fig. 1 shows the data for HiSiv 3000 and HiSiv 1000 (50 £ 70 mesh), where time axis is given in hours. Fig. 2 shows the data for activated carbon and F-400, where time axis goes up to 30 days. It was found that removal of phenol by HiSiv 1000 was obtained in less than 3 h, whereas equilibrium was not reached for F-400 and activated carbon even after 29 days. This is due to the fact that F-400 and activated carbon have very wide pore size distribution, whereas HiSiv 3000 and HiSiv 1000 have only one size for pores, determined by their crystalline structure. In the case of F-400 and activated

2.3. Regeneration experiments The effect of regeneration on adsorption capacity of phenol was studied using HiSiv 1000. The adsorbent was regenerated in air atmosphere at 360 8C for 16 h in a Fisher Isotemp Junior Model oven. This temperature was chosen since it was suggested in the literature for phenol adsorption (Meytal et al., 1997). The structure of HiSiv 1000 will not be damaged at 360 8C, since this adsorbent is thermally stable up to 500 8C (product information sheet, UOP, USA).

Fig. 1. Kinetic data for adsorption of phenol by HiSiv 3000 and HiSiv 1000 (50 £ 70) mesh at 25 8C.

160

N. Roostaei, F.H. Tezel / Journal of Environmental Management 70 (2004) 157–164

cations cause the surface of the zeolite to be more heterogeneous, attracting more polar molecules. As a result of this higher affinity of HiSiv 1000 for polar molecules, the adsorption capacity of phenol on this adsorbent will be higher than that of phenol on HiSiv 3000.

Fig. 2. Comparison of adsorption rates of phenol by F-400 and activated carbon at 25 8C.

carbon, phenol initially fills larger pores very fast, in the fast adsorption range, and then slowly keeps filling the smaller pores. For practical applications, if the times of rapid adsorption stage are compared for these adsorbents, they were observed to be 30 min for HiSiv 1000, 5 h for F-400, and 23 h for activated carbon. For the determination of equilibrium capacities for isotherms, following times were allowed before analysis: 1.5 days for HiSiv 1000, 19 days for activated carbon and 30 days for F-400. HiSiv 1000 adsorbent has a crystalline, inorganic silica– alumina structure which is developed from high silica molecular sieve, also referred to as ‘high silica zeolites’ (UOP, 1998b). The secondary building unit of HiSiv 1000 is similar to those of Zeolite Y, with Si/Al ratio less than 20 (UOP, 1998b). HiSiv 3000 and HiSiv 1000 are zeolites with different micropore sizes as a result of different intracrystalline channel structures. ZSM-5 zeolites ˚ channel including HiSiv 3000 are characterized by 5.7 A size formed by 10-member oxygen rings. Zeolite Y adsorbents including HiSiv 1000 are characterized by 12-member oxygen ˚ . Therefore phenol rings which have free diameters of 7–7.4 A ˚ can be more molecules with molecular size of around 6 A easily adsorbed by HiSiv 1000 than by HiSiv 3000. This will explain the faster equilibrium time observed with HiSiv 1000, as opposed to HiSiv 3000. On the other hand the Si/Al ratio plays a very important role in determining the adsorptive properties of zeolites. The Si/Al ratio in a zeolite is never less than 1.0, but there is no upper limit and pure silica analogs of some of the zeolite structures have been prepared. The adsorptive properties show a systematic transition from the aluminum-rich sieves, which have very high affinities for water and other polar molecules, to the microporous silica such as silicalite, which are essentially hydrophobic and adsorb n-paraffins in preference to water (Ruthven, 1984). The Si/Al ratio of HiSiv 3000 is in the range of thousands, whereas this ratio for HiSiv 1000 is less than 20. The lower this ratio in zeolites, the more cations will be present in the structure, to balance the charge difference between Siþ4 and Alþ3. These

3.1.1. Effect of particle size Kinetic experiments were performed for HiSiv 1000 using four different particle sizes. The results, given in Fig. 3 show that, the smaller the particle size, the faster it can reach equilibrium. Since the initial concentrations were different for these experiments for different particle sizes, the equilibrium concentrations were different in this figure. It can be seen from Fig. 3 that powdered HiSiv 1000 has the highest rate of adsorption. For HiSiv 1000, which has the structure of Zeolite Y, most of the adsorption capacity is in the microporous channels. When adsorbent particles are small, these micropores are more exposed to the adsorbate molecule and therefore the diffusion will be faster. Although powdered HiSiv 1000 gave the best kinetics result, working with powdered particles was not easy, since they were stuck to the wall of the containers after each experiment. 3.1.2. Kinetics modeling An adsorption model similar to first order reversible kinetic reaction model (Bhattacharya and Venkobachar, 1984) was used to determine the rate of adsorption of phenol for adsorbents studied. In this model, the relation between fractional uptake ½UðtÞ; and overall rate constant ½k0  can be expressed as: Ln½1 2 UðtÞ ¼ k0 t 0

ð1Þ

k ¼ k1 þ k2

ð2Þ

UðtÞ ¼ XA =XAe

ð3Þ

where: k1 ; k2 ¼ First order rate constants for adsorption and desorption, respectively (1/h)

Fig. 3. Effect of particle size on rate of adsorption by HiSiv 1000 at 25 8C.

N. Roostaei, F.H. Tezel / Journal of Environmental Management 70 (2004) 157–164

161

3.2. Equilibrium experiments

Table 2 The values of rate constants for different adsorbents Adsorbents

k0

t1=2

HiSiv 1000 (powder) HiSiv 1000 (50 £ 70 mesh) HiSiv 1000 (20 £ 50 mesh) HiSiv 1000 (extrude) Activated carbon F-400

25.51 5.194 2.19 0.141 0.625 0.926

0.0271 0.133 0.316 4.914 1.108 0.748

XA ¼ Fractional adsorption of phenol XAe ¼ Fractional adsorption of phenol at equilibrium. It was assumed that adsorbents were initially free of phenol. By plotting the value of ln½1 2 UðtÞ versus time, the overall rate constant ðk0 Þ can be determined from the initial slope of this curve. After finding k0 ; the value of t1=2 (time required to reach half of the equilibrium capacity) can be calculated from Eq. (1) by substituting UðtÞ ¼ 1=2: Table 2 shows the value of rate constants ðk0 Þ for activated carbon, F-400 and different particle size of HiSiv 1000. Comparing the values of k0 for HiSiv 1000 it is observed that this value is decreased when the particle size of the adsorbent is increased. It takes longer for adsorbate molecules to diffuse into the larger particles. Comparing k0 values for HiSiv 1000, activated carbon, and F-400 it is concluded that HiSiv 1000 has the highest overall rate constant.

Results of phenol adsorption equilibrium experiments at 25 8C for HiSiv 1000 (50 £ 70 mesh), activated carbon, and F-400 are given and compared in Fig. 4, up to 220 mg/l concentration in the liquid phase. Since the interest in this work was in the drinking water concentration range, higher concentrations were not considered. The points represent experimental data, and the lines represent different isotherm models used. Langmuir, Freundlich and three-parameter Langmuir –Freundlich isotherms were used to model the experimental isotherms obtained according to the following equations: Langmuir:



Qbc 1 þ bc

ð4Þ

where q : adsorption capacity (mg/g) b : Langmuir constant (l/mg) Q : maximum saturation capacity at the isotherm temperature (mg/g) c : concentration in the liquid phase (mg/l) Freundlich: q ¼ kc1=n

Fig. 4. Equilibrium adsorption isotherm comparison for activated carbon, F-400 and HiSiv 1000 at 25 8C.

ð5Þ

162

N. Roostaei, F.H. Tezel / Journal of Environmental Management 70 (2004) 157–164

where q : adsorption capacity (mg/g) k : Freundlich constant (mg/g)(l/mg)1/n c : concentration in the liquid phase (mg/l) 1=n : Freundlich constant (dimensionless) Three-parameter Langmuir –Freundlich (Sips) isotherm:

q¼Q

ðbcÞ1=n 1 þ ðbcÞ1=n

ð6Þ

where q : adsorption capacity (mg/g) Q : constant (mg/g) c : concentration in the liquid phase (mg/l) b : constant (l/mg) 1=n : constant (dimensionless) The parameters for each model were determined for the isotherms experimentally determined. Table 3 shows the values of these parameters for activated carbon, F400, and HiSiv 1000 (including different temperatures and different particle sizes for HiSiv 1000). As can be seen from Fig. 4, all of the isotherm models fit the HiSiv 1000 data, since they were in the linear region. Freundlich isotherm did not represent activated carbon data, at all. Although the experimental data points started to level off, the Freundlich isotherm predicted that the equilibrium capacity would keep increasing steadily with increasing equilibrium concentration in the liquid phase. Three-parameter isotherm model predicted the data better at low concentrations for activated carbon, whereas Langmuir gave a better fit at higher concentrations. For F-400, all three-isotherm model predictions were close, although Freundlich model should not be used for the extrapolation of this data to higher concentrations. Like the prediction for the activated carbon, it had a trend to increase capacity as concentration increased, whereas the experimental data showed some leveling off.

It was found that activated carbon had the highest capacity for phenol, followed by F-400 and HiSiv 1000 for the concentration range studied in this work. The surface area determined during this study for HiSiv 1000, F-400 and activated carbon are 355, 900 and 858 m2/g, respectively. The difference in this value for these adsorbents explains the higher adsorption capacities for F-400 and activated carbon, compared to HiSiv 1000. Activated carbon almost reached its maximum limiting saturation capacity, as can be seen in Fig. 4 with capacity leveling off as concentration increases. F-400 and HiSiv 1000 showed steady increase in adsorption capacity with increasing concentration under these conditions. 3.2.1. Effect of particle size Effect of particle size on adsorption capacity of phenol by HiSiv 1000 was also studied. The results are given in Fig. 5. It was observed that changing the particle size did not affect the capacity of adsorption as can be seen from this figure. Although adsorption kinetics got faster with decreasing particle size (as previously discussed under ‘kinetic experiments’), equilibrium capacity was not affected by changing particle size. Most of the adsorption capacity for HiSiv 1000 is in the micropores. Therefore once the adsorbate molecule is given enough time to diffuse into the micropores, the capacity is controlled by the structure of the HiSiv 1000, no matter what the particle size is. 3.2.2. Effect of temperature To investigate the effect of temperature on adsorption capacity of HiSiv 1000, isotherm experiments were conducted at 25, 40, and 55 8C for 50 £ 70-mesh particle size. This particle size was chosen to work with, since it had fast kinetics. It was difficult to work with powdered HiSiv 1000. The results indicated that adsorption capacity was decreased as temperature was increased as shown in Fig. 6. This comes from the fact that phenol adsorption by HiSiv 1000 is an exothermic process. That is also observed with the Q parameters of the three-parameter fits for different temperatures for this adsorbent. Although the Q parameters for the Langmuir fits

Table 3 Equilibrium model parameters Adsorbent

Activated carbon F-400 HiSiv 1000 (Powder) HiSiv 1000 (50 £ 70 mesh) HiSiv 1000 (20 £ 50 mesh) HiSiv 1000 (extrude) HiSiv 1000 (50 £ 70 mesh), T ¼ 408 HiSiv 1000 (50 £ 70 mesh), T ¼ 558

Three-parameter

Freundlich

Langmuir 1/n

Q (mg/g)

b (l/mg)

1=n (–)

k (mg/g)(l/mg)

1=n (–)

Q (mg/g)

b (l/mg)

264.4 265.7 199.3 198.9 198.4 198.3 170.0 88.1

0.0821 0.0192 0.00155 0.00044 0.00026 0.00020 0.00060 0.00015

1.317 0.628 1.337 0.799 0.742 0.674 0.900 0.522

37.0 36.3 0.047 0.442 0.644 0.680 0.299 0.500

0.420 0.319 1.252 0.767 0.647 0.639 0.813 0.638

309.7 205.1 319.0 319.0 319.0 319.0 319.0 318.9

0.0533 0.0420 0.00055 0.00051 0.00045 0.00056 0.00041 0.00026

Temperature is 25 8C unless otherwise stated.

N. Roostaei, F.H. Tezel / Journal of Environmental Management 70 (2004) 157–164

163

Fig. 5. Effect of particle size on adsorption capacity of HiSiv 1000 for phenol at 25 8C.

were the same for different temperatures, the b parameter decreased as temperature increased. These parameters for different fits should be used with caution above the concentration range studied in this work, as extrapolation of these isotherms may not be reliable. 3.2.3. Effect of regeneration The effect of regeneration at 360 8C for 16 h on adsorption capacity of phenol was studied using HiSiv 1000. After each regeneration, adsorption isotherm was determined to see whether the capacity has changed after regeneration. This process was repeated up to 14 times to investigate the effect of regeneration on capacity of adsorption and also to determine the loss of adsorbent

Fig. 7. Effect of regeneration (cycles 11 –14) on adsorption isotherms of phenol on HiSiv 1000 (50 £ 70 mesh) at 25 8C.

particles during regeneration experiments. Fig. 7 compares the isotherm of the virgin (without regeneration) HiSiv 1000 to the isotherms of the same adsorbent after 11 – 14 regeneration –adsorption cycles. It was observed that there was no apparent change in adsorption capacity after regeneration up to 14 cycles. The average loss of particles during regeneration was about 2.5 – 5% per cycle for HiSiv 1000 while the same amount for activated carbon was 5– 10% per cycle due to oxidation and attrition (Cen, 1994). Comparing HiSiv 1000 particles with activated carbon particles it was observed that HiSiv 1000 particles were more rigid than activated carbon particles. As a result activated carbon lost more particles during adsorption – regeneration cycles.

4. Conclusions Following conclusions were drawn from the present study: – – – Fig. 6. Adsorption isotherms of phenol with HiSiv 1000 (50 £ 70 mesh), at different temperatures. Points represent experimental data and curves represent three-parameter isotherm fits for 25 and 40 8C and Freundlich fit for 55 8C.

There was no significant phenol adsorption by silica gel and activated alumina. HiSiv 1000 had the fastest adsorption kinetics among the adsorbents studied. Experiments carried out with different particle sizes of HiSiv 1000 showed that the smaller the particle size, the faster was the diffusion of phenol into adsorption sites within the adsorbent. They also showed that equilibrium capacity of HiSiv 1000 did not change with changing particle size.

164

– –





N. Roostaei, F.H. Tezel / Journal of Environmental Management 70 (2004) 157–164

Adsorption capacity of HiSiv 1000 decreased with increasing temperature. Thermal regeneration experiments carried out with HiSiv 1000 showed that its adsorption capacity for phenol did not change after 14 regeneration cycles. It was found that activated carbon had the highest capacity for adsorption of phenol, followed by F-400 and HiSiv 1000 for the concentration range studied in this work. Activated carbon almost reached its limiting saturation capacity, whereas F-400 and HiSiv 1000 showed steady increase in adsorption capacity under these conditions. Langmuir –Freundlich isotherm model was the best to describe adsorption equilibrium data for phenol.

Acknowledgements Financial support received from Natural Sciences and Engineering Research Council (NSERC) of Canada is gratefully acknowledged. Authors also acknowledge the contribution of Ms. Isil Toreci who determined the surface areas of the adsorbents. References Bhattacharya, A.K., Venkobachar, C., 1984. Removal of cadmium (II) by low cost adsorbents. Environ. Engng 110 (1). Cen, J., 1994. Electrochemical Regeneration of Granular Activated Carbon. MASc Thesis, Department of Civil Engineering, University of Ottawa. Chiang, P.C., Chang, E.E., Wu, J.S., 1997. Comparison of chemical and thermal regeneration of aromatic compounds on exhausted activated carbon. Water Sci. Technol. 35 (7), 279 –286. Cooke, S., Lables, M.M., 1994. Destruction of the environmentally hazardous monochlorinated phenols via pyrolysis in an inert atmosphere. Carbon 32, 1055.

Hsieh, C.-T., Teng, H., 2000. Liquid-phase adsorption of phenol onto activated carbons prepared with different activation levels. J. Colloid Interface Sci. 230 (1), 171 –175. Leyva-Ramos, R., Soto Zuniga, J.H., Mendoza Barron, J., Guerrero Coronado, R.M., 1999. Adsorption of phenol from aqueous solution on to activated carbon. Effect of solvent temperature and particle size. Adsorp. Sci. Technol. 17 (7), 533–544. Meytal, Yu.L., Sheintuch, M., Shter, G.E., Grader, G.S., 1997. Optimal temperature for catalytic regeneration of activated carbon. Carbon 35 (10), 1527–1531. Moreno-Castilla, C., Rivera-Utrilla, J., Joly, J.P., Lopez-Ramon, M.V., Ferro-Garcia, M.A., Carrasco-Marin, F., 1995. Thermal regeneration of an activated carbon exhausted with different substituted phenols. Carbon 33 (10), 1417– 1424. Peel, R.G., Benedec, A., 1980. Attainment of equilibrium in activated carbon isotherm studies. Environ. Sci. Technol. 14 (1), 66–71. Radovic, L.R., Moreno-Castilla, C., Rivera-Utrilla, J., 2000. Carbon materials as adsorbents in aqueous solutions. Chem. Phys. Carbon 27, 227. Ruthven, D.M., 1984. Principles of Adsorption and Adsorption Processes, Wiley, New York. Singh, B.Kr., Rawat, N.S., 1994. Comparative sorption equilibrium studies of toxic phenols on fly ash and impregnated fly ash. J. Chem. Tech. Biotech. 61, 307– 313. UOP Company, 1998. Product Information Sheet of HiSiv 1000, Molecular Sieve Department, Mt Laurel, NJ, USA. Vidic, R.D., Suidan, M.T., 1991. Role of dissolved oxygen on the adsorptive capacity of activated carbon for synthetic and natural organic matter. Environ. Sci. Technol. 25 (9), 1612–1618. Viraraghavan, T., Alfaro, J., 1992. Removal of phenol from wastewater by adsorption on peat, fly ash and bentonite. J. Environ. Engng 5, 311–317. van Vliet, B.M., Venter, L., 1984. Infrared thermal regeneration of spend carbon from water reclamation. Water Sci. Technol. 17, 1029–1042. Waer, M.A., Snoeyink, V.L., Mallon, K.L., 1992. Carbon regeneration dependence in time and temperature. J. Am. Water Works Assoc. 84 (3), 82–91. Zogorski, J.S., Faust, S.D., 1978. Equilibria of adsorption of phenols by granular activated carbon. In: Rubin, A.J., (Ed.), Chemistry of Wastewater Technology, Ann Arbor Science Publishers, Ann Arbor, Michigan, Chapter 9. Zogorski, J.S.M., Faust, S.D., Haas, J.H. Jr., 1976. The kinetics of adsorption of phenols by granular activated carbon. J. Colloid Interface Sci. 55 (2), 329 –341.