Removal of Phenol

Colloids and Surfaces A: Physicochem. Eng. Aspects 272 (2006) 89–104

Adsorptive removal of phenol by bagasse fly ash and activated carbon:Equilibrium, kinetics and thermodynamics

Vimal C. Srivastava, Mahadeva M. Swamy1, Indra D. Mall∗, Basheswar Prasad, Indra M. Mishra

Department of Chemical Engineering, Indian Institute of Technology Roorkee, Roorkee, Uttaranchal 247667, India

Received 3 April 2005; received in revised form 9 July 2005; accepted 14 July 2005Available online 26 September 2005

Abstract

Present study deals with the adsorption of phenol on carbon rich bagasse fly ash (BFA) and activated carbon-commercial grade (ACC)and laboratory grade (ACL). BFA is a solid waste obtained from the particulate collection equipment attached to the flue gas line of thebagasse-fired boilers of cane sugar mills. Batch studies were performed to evaluate the influences of various experimental parameters likeinitial pH (pH0), contact time, adsorbent dose and initial concentration (C0) on the removal of phenol.C0 varied from 75 to 300 mg/l for thea pHa nitials nol is oft ngmuir,T on isothermwf Gibbs freee d with thes©

K

1

rpmsli

f

a

gwn

eststedantsans

ndistur-dept-hasl inters,ntra-

0d

dsorption isotherm studies and the effect of temperature on adsorption. Optimum conditions for phenol removal were found to be0 ≈ 6.5,dsorbent dose≈10 g/l of solution and equilibrium time≈5 h. Adsorption of phenol followed pseudo-second order kinetics with the iorption rate for adsorption on ACL being the highest followed by those on BFA and ACC. The effective diffusion coefficient of phehe order of 10−10 m2/s. Equilibrium isotherms for the adsorption of phenol on BFA, ACC and ACL were analysed by Freundlich, Laemkin, Redlich–Peterson, Radke–Prausnitz and Toth isotherm models using non-linear regression technique. Redlich–Petersas found to best represent the data for phenol adsorption on all the adsorbents. The change in entropy (�S◦) and heat of adsorption (�H◦)

or phenol adsorption on BFA were estimated as 1.8 MJ/kg K and 0.5 MJ/kg, respectively. The high negative value of change innergy (�G◦) indicates the feasible and spontaneous adsorption of phenol on BFA. The values of isosteric heat of adsorption varieurface loading of phenol.2005 Elsevier B.V. All rights reserved.

eywords: Adsorption; Phenol removal; Bagasse fly ash (BFA); Activated carbon; Adsorption thermodynamics; Temperature; Kinetics; Isotherms

. Introduction

Phenol, a derivative of benzene, is an important raw mate-ial and/or product of chemical and allied industries (e.g.etrochemicals, oil refineries, plastics, leather, paint, phar-aceutical, steel industries and pesticides)[1,2]. It is highly

oluble in water and is very toxic in nature. It is a colour-ess, hygroscopic and crystalline substance, which turns pinkn air owing to its oxidation. Its solubility in water is 98 g/l

∗ Corresponding author. Tel.: +91 1332 285319 (O)/285106 (R);ax: + 91 1332 276535/273560.

E-mail address: id [email protected] (I.D. Mall).1 Present address: Department of Environmental Engineering, Sri Jay-chamarajendra College of Engineering, Mysore, Karnataka, India.

and its melting point is 181◦C. It is a weak acid dissociatinslightly in aqueous solution: for this reason it is also knoas carbolic acid. The Ministry of Environment and For(MOEF), Government of India and EPA, USA, have lisphenol and phenolic compounds on the priority-pollutlist. Chronic toxic effects due to phenols reported in huminclude vomiting, difficulty in swallowing, anorexia, liver akidney damage, headache, fainting and other mental dbances[3]. That phenol is highly toxic and difficult to degrabiologically have led to setting up of rigid limits on the acceable level of phenol in the environment. While the MOEFset a maximum concentration level of 1.0 mg/l of phenothe industrial effluents for safe discharge into surface wathe WHO recommends the permissible phenolic concetion of 0.001 mg/l in potable waters[4].

927-7757/$ – see front matter © 2005 Elsevier B.V. All rights reserved.oi:10.1016/j.colsurfa.2005.07.016

90 V.C. Srivastava et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 272 (2006) 89–104

Nomenclature

1/n heterogeneity factor, dimensionlessaR constant of Redlich–Peterson isotherm (l/mg)A adsorbateACC activated carbon-commercial gradeACL activated carbon-laboratory gradeA.S activated complexBFA bagasse fly ashC constant that gives idea about the thickness of

boundary layer (mg/g)C0 initial concentration of adsorbate in solution

(mg/l)Ce equilibrium liquid phase concentration (mg/l)CS adsorbent concentration in the solutionDe effective diffusion coefficient of adsorbate in

the adsorbent phase (m2/s)f exchange of number of moles of water per

mole of adsorbate at the adsorption siteF(t) fractional uptake of adsorbate on adsorbent,

0 <F(t) < 1h initial sorption rate (mg/g min)HYBRID hybrid fractional error functionk0 constant in Bangham equationkf rate constant of pseudo-first-order adsorption

model (min−1)kid intra-particle diffusion rate constant

(mg/g min1/2)kA adsorption rate constant for the adsorption

equilibriumkD desorption rate constant for the adsorption

equilibriumkRP constant in Radke–Prausnitz isotherm

((mg/g)/(mg/l)1/P)kS rate constant of pseudo-second-order adsorp-

tion model (g/mg min)KA equilibrium adsorption constantKF constant of Freundlich isotherm

((mg/g)/(l/mg)1/n)KL constant of Langmuir isotherm (l/mg)KR constant of Redlich–Peterson isotherm (l/g)KRP constant in Radke–Prausnitz isotherm (l/g)KTh constant in Toth isotherm ((mg/l)Th)m mass of adsorbent per liter of solution (g/l)n number of data pointsN number of data pointsp number of parametersP constant in Radke–Prausnitz isotherm, dimen-

sionlessMPSD Marquardt’s percent standard deviationqe equilibrium solid phase concentration (mg/g)qe,cal calculated value of solid phase concentration

of adsorbate at equilibrium (mg/g)

qe,exp experimental value of solid phase concentra-tion of adsorbate at equilibrium (mg/g)

q∞e monolayer adsorption capacity parameter in

Toth isotherm (mg/g)qm maximum adsorption capacity of adsorbent

(mg/g)qt amount of adsorbate adsorbed by adsorbent at

time t (mg/g)R universal gas constant (8.314 J/K mol)Ra radius of the adsorbent particle assumed to be

spherical (m)RL separation factor, dimensionlessS active site on the adsorbentt time (min)T absolute temperature (K)Th constant in Toth isotherm, dimensionlessV volume of the solution (l)w mass of the adsorbent (g)XAe fraction of the adsorbate adsorbed on the adsor-

bent under equilibriumz an integer�G◦ Gibbs free energy of adsorption (KJ/mol)�H◦ enthalpy of adsorption (KJ/mol)�Hw heat of adsorption of water (normally assumed

to be zero)�Hsol heat of solution�Hst,0 isosteric heat of adsorption with zero coverage�Hst,a apparent isosteric heat of adsorption�Hst,net net isosteric heat of adsorption�S◦ entropy of adsorption (J/K mol)

Greek symbolsα Bangham constant (<1)β constant of Redlich–Peterson isotherm

(0 <β < 1)λvap heat of vapourisation of phenol

Several methods for the treatment of phenolic waste waterhave been proposed in the literature. These include physico-chemical treatment processes, chemical oxidation and bio-logical degradation. The physico-chemical processes includeadsorption and ion exchange. Various oxidizing agents (oxy-gen, hydrogen peroxide, ozone, etc.) have been used forwet oxidation of phenolic waste waters. For high strengthand low volume of phenolic waste waters, phenol removalby adsorption using granular/powdered activated carbon hasbeen widely used[5–12]. However, high costs of activatedcarbon and 10–15% loss during regeneration and the dif-ficulties faced in the recovery/disposal of phenol make theutilization of activated carbon prohibitive in the developingcountries[8]. This has led to search for cheaper carbonaceoussubstitutes to activated carbon.

V.C. Srivastava et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 272 (2006) 89–104 91

Most of the studies on phenol removal dealt with theadsorptive equilibrium and kinetics, both in batch and contin-uous mode. The equilibrium relationships were representedby various isotherm models, viz. two parameter modelsof Freundlich[8–21] and Langmuir[9,11–13,17–22]; threeparameter models of Redlich–Peterson[18]. The details ofthese models and their characteristics have been extensivelydealt with by Swamy et al.[8]. Most of the authors have usedVermeulen’s[23] approximations to the analytical solution ofa differential equation to represent the fractional approach toequilibrium. Mechanistic studies to understand sorption char-acteristics have been reported by many investigators[24,25].Temperature is known to have pronounced effect on adsorp-tion phenomenon and intraparticle diffusional process.

In the present investigation, sugarcane bagasse fly ash(BFA) collected from the particulate collection deviceattached to the flue gas line of a bagasse-fired boiler stackof a nearby canesugar mill, and activated carbon-commercialgrade (ACC) and laboratory grade (ACL) have been usedas adsorbents for the removal of phenol from aqueous solu-tion. The aim of the present work is to explore the possibilityof BFA, which is a carbonaceous waste, being utilized asan adsorbent for the removal of phenol from wastewater.BFA has fairly good amount of carbon and silica in it. ACLhas been used as the standard adsorbent for the comparisonof adsorptive capacities of BFA and ACC. Effects of suchpc fp ptiono iffer-e beens men-t minet Thee FAh ptionp nerga alsob

2

2

tionL adea rew-e aoli,D an-u liedI Thea

entsw anal-y

Bulk densities were determined by using MAC bulk den-sity meter whereas particle size analysis was done usingstandard sieves. The specific surface area and the porediameter of the samples were measured by N2 adsorptionisotherm using an ASAP 2010 Micromeritics instrumentand by Brunauer–Emmett–Teller (BET) method, using thesoftware of Micromeritics. Nitrogen was used as cold bath(77.15 K).

X-ray diffraction analyses of the adsorbents were carriedout using Phillips diffraction unit (Model PW 1140/90), usingcopper as the target with nickel as filter media, and K radiationmaintained at 1.542A. Goniometer speed was maintainedat 1◦/min. Scanning electron microscopic (SEM) analysesof the adsorbents were carried out by using a scanningelectron microscope (Model SEM-501, Phillips, Holland)[27].

2.2. Adsorbate

Phenol (C6H5OH) of analytical reagent (AR) grade sup-plied by Ranbaxi Laboratories Ltd., India, was used for thepreparation of the synthetic adsorbate solutions of variousC0in the range of 75–300 mg/l. The required quantity of phenolwas accurately weighed and dissolved in a small amount ofdistilled water and subsequently made-up to 1 l in a measur-ing flask. Fresh stock solution as required was prepared everyd f 5 lc db

2

ingo ngthu hi-m rsusc p to4 gherc is-t ationl phe-n thec enolv yingf

2

ofk eret wasa at ac wasw lterp olc con-

arameters as initial pH (pH0), adsorbent dose (m), initialoncentration (C0), and contact time (t) on the sorption ohenol have been investigated. The kinetics of adsorf phenol on the adsorbents has been studied using dnt models. The sorption capacity of the adsorbents hastudied using the adsorption isotherm technique. Experial data were fitted to various isotherm equations to deterhe best isotherm to correlate the experimental data.ffect of temperature for the adsorption of phenol onto Bas also been investigated. Thermodynamics of adsorrocess have been studied and the change in Gibbs free end the enthalpy, and isosteric heat of adsorption haveeen determined.

. Material and methods

.1. Adsorbents and their characterization

BFA was obtained from U.P. State Sugar Corporatd., Doiwala Unit, Dehradun, India. The commercial grctivated carbon (ACC) manufactured by Rajasthan Bries Ltd. was procured from the open market (Kharibelhi) and the laboratory grade activated carbon (ACL) mfactured by GSE Chemical Testing Laboratory and Al

ndustry, New Delhi, was procured from local suppliers.dsorbents were used as procured.

The physico-chemical characteristics of the adsorbere determined using standard procedures. Proximatesis was carried out using the standard procedure[26].

y

ay and was stored in a brown colour glass reservoir oapacity to prevent photo-oxidation. TheC0 was ascertaineefore the start of each experimental run.

.3. Analytical measurements

The concentration of phenol was determined by findut the absorbance of the solution at 270 nm wavelesing UV/vis spectrophotometer (model UV 210 A; Scadzu, Japan). The calibration plot of absorbance ve

oncentration for phenol showed a linear variation u0 mg/l concentration. Therefore, the samples with hioncentration of phenol (>40 mg/l) were diluted with dilled water, whenever necessary, to make the concentress than 40 mg/l, for the accurate determination of theol concentration with the help of the linear portion ofalibration curve. The pH of the aqueous solution of pharied from 7.04 to 7.59 for the phenol concentration varrom 10 to 200 mg/l.

.4. Batch experimental programme

For each experiment, 50 ml of the phenol solutionnownC0, pH0 and a known amount of the adsorbents waken in a 100 ml stoppered conical flask. This mixturegitated in a temperature-controlled shaking water bathonstant speed of 145 rpm. Small amount of the sampleithdrawn after 5 h and was filtered through Whatman fiaper no. 42 (pore size ca. 2.5�m) and analysed for phenoncentration. The sample was again transferred to the


ical flask. The pH0 of the adsorbate solutions was adjustedusing 1 M aqueous solution of either HCI or NaOH.

The percentage removal of phenol and equilibrium adsorp-tion uptake,qe (mg/g), were calculated using the followingrelationships:

% removal= 100(C0 − Ce)

C0, (1)

amount adsorbedqe = (C0 − Ce)V

w

(mg of adsorbate/g of adsorbent), (2)

whereC0 is the initial sorbate concentration (mg/l),Ce theequilibrium sorbate concentration (mg/l),V the volume of thesolution (l) andw is the mass of the adsorbent (g).

2.5. Effect of initial pH (pH0)

The sorption of phenol by the adsorbents was studied overa pH0 range of 3–10 at 30◦C and the studies were carried outfor 5 h.C0 was 100 mg/l and the adsorbent dose was kept at10 g/l for ACL and BFA and 12 g/l for ACC.

2.6. Effect of temperature and estimation ofthermodynamic parameters

sticsw s at2 ofa minedu7 eda

2

thea en ina dsor-b pt ina que-o . Att n,t phe-n aperd allt nola than0 thea rptionk fter1 batef

ther ilib-

rium adsorption data. For adsorption isotherms, experimentswere carried out at pH0 of 6.5 by contacting fixed amount ofadsorbent with 50 ml of phenol solution at theC0 varying overthe range of 75–300 mg/l. Adsorbents were separated fromthe solution after 5 and 24 h, and the phenol concentration inthe solution and in the adsorbent was estimated.

In order to investigate the kinetics of adsorption of phe-nol on the adsorbents, various kinetic models, like pseudo-first-order, pseudo-second-order, Bangham and intraparticlediffusion models were used. Two parameters model, viz.,Langmuir, Freundlich and Temkin, and three-parameter mod-els, viz., Redlich–Peterson, Radke–Prausnitz and Toth havebeen used to describe the equilibrium nature of adsorption ofphenol in the present study.

Two different error functions of non-linear regressionbasin were employed in this study to find out the mostsuitable kinetic and isotherm models to represent the experi-mental data respectively. The hybrid fractional error function(HYBRID) [28] and the Marquardt’s percent standard devi-ation (MPSD) error function[29] have been used previouslyby a number of researchers in the field[30,31]. These errorfunctions are given as:

HYBRID = 100

n − p

n∑i=1

[qe,meas− qe,calc

qe,meas

]i

(3)

M

H ofe D iss ribu-t domo

3

3

bentsa eni n-s etc.F isl easo giveni icles ss ofB

m-i BFA,Al andh BFA

The effect of temperature on the sorption characterias investigated by determining the adsorption isotherm5, 30, 35, 40 and 45◦C. Apparent and net isosteric heatsdsorption at various surface coverages have been detersing classical thermodynamic equations.C0 was varied from5 to 300 mg/l but the pH0 of the solutions was maintaint 6.5.

.7. Batch kinetic and isotherm study and error analysis

To determine thet necessary for adsorption, 50 ml ofqueous solution containing 100 mg/l of phenol was takseries of conical flasks. Preweighed amounts of the aents were added to different flasks. The flasks were ketemperature-controlled shaking water bath and the a

us solution–adsorbent mixtures were stirred at 145 rpmhe end of the predeterminedt, the flasks were withdrawheir contents were filtered, and the filtrate analyzed forol. The amount of phenol adsorbed onto the filter puring the filtration was found to be less than 1% for

he C0 between 20 and 60 mg/l. Above 75 mg/l, the phedsorption onto the filter paper was found to be less.7%. Hence, the effect of filter paper on adsorption bydsorbents has been neglected during the studies. Adsoinetics was followed for 24 h and it was observed that ah, there was gradual but very slow removal of adsor

rom the solution.To optimize the design of an adsorption system for

emoval of adsorbate, it is important to obtain the equ

PSD= 100

√√√√ 1

n − p

n∑i=1

(qe,meas− qe,calc

qe,meas

)2

i

(4)

YBRID was developed to improve the fit of the squarerrors function at low concentration values. The MPSimilar in some respects to a geometric mean error distion modified according to the number of degrees of freef the system.

. Results and discussion

.1. Characterisation of adsorbents

Physico-chemical characteristics of the three adsorre presented inTable 1and particle size analysis is giv

n Table 2. The ash obtained by incineration of BFA coists mainly of carbon, silica, alumina, calcium oxide,romTable 1, it is observed that the bulk density of BFA

ower than that of ACC or ACL. The specific surface arf the adsorbents used in the present investigations are

n Table 1. From geometrical considerations of the parthape, it is expected that the surface area per unit maFA will be much lower than that of either ACC or ACL.The morphologies of BFA, ACC and ACL were exa

ned under scanning electron microscope. The SEMs ofCC and ACL as obtained are shown inFig. 1. The BFA has

inear type of fibres with holes in it and at other placesas skeletal structure. Further, the number of pores in


Table 1Characteristics of bagasse fly ash and activated carbons

Characteristics ACLa ACC BFA

Proximate analysisMoisture (%) 0.31 9.00 2.51Ash (%) 2.65 8.78 30.98Volatile matter (%) 7.74 24.08 23.48Fixed carbon (%) 89.30 58.14 43.03Bulk density (kg/m3) 750.82 610.00 270.50

Chemical analysis of ashSiO2 (%) 8.00 54.93 51.05Al2O3 (%) ND 6.84 10.75CaO (%) ND 11.15 6.04Fe2O3 (%) ND 1.26 3.45MgO (%) ND 2.36 1.10Rest others – – –

Surface area of pores (m2/g)(i) BET 492.00 336.60 168.39(ii) BJH

(a) adsorption cumulative – 64.43 70.90(b) desorption cumulative – 34.03 45.30

BJH cumulative pore volume (cm3/g)(i) Single point total – 0.1622b 0.1067c

(ii) BJH adsorption – 0.0425 0.0844(iii) BJH desorption – 0.0224 0.0622

Average pore diameter (A)(i) BET – 19.72 25.54(ii) BJH adsorption – 26.62 49.85(iii) BJH desorption – 26.34 58.44

a Detailed analyses of pore surface/volume not performed.b Pores less than 2239A.c Pores less than 2194A.

are found to be less than that of ACL and these are also rela-tively larger in size, but the numbers of pores are more thanthat present in ACC. Activated carbon is generally describedas an amorphous form of graphite with a random structure ofgraphite plates; having highly porous structure with a rangeof cracks and crevices reaching molecular dimensions. Thesummary of porous structure of BFA and ACC are shown in

Table 1. Fig. 2 shows the pore size distribution of BFA andACC. While ACC has a narrower pore size distribution, BFAhas a much wider size distribution. The average pore size dis-tribution of BFA is also much larger. Although, the detailedpore size analysis of the ACL is not available, it is expectedthat the pore size distribution shall be in between that of BFAand ACC. It has both macro- and micro-pores[32].

X-ray diffraction patterns for BFA, ACC and ACL areshown inFig. 3. Major components identified in BFA arecrystalline quartz, alumina, cristobalite, and calcium orthosil-icate, whereas, tridymite and silicate carbon are the majorcomponents of ACC and ACL[33]. Diffraction peaks cor-responding to crystalline carbon were not observed in BFAand activated carbons. Nowackiite was also observed in ACL.The broad peaks in all the three samples indicate the pres-ence of amorphous form of silica. From the micrographs andX-ray diffractograms, it is seen that the activated carbons andthe BFA have heterogeneous surface. Proximate analysis ofadsorbents showed 43.03, 89.03 and 58.14% fixed carbon inBFA, ACL and ACC, respectively. Thus, BFA has about one-half of the carbon present in ACL and about three-fourth ofthe carbon present in ACC.

3.2. Effect of adsorbent dosage (m)

Car se ina osagel se int ted tog ptions atedw n isl esd ounto val

Table 2Particle size analysis of different adsorbents

ACL ACC

Size (mm) Weight (%) Size

>4.00 1.42 >1.19 mm−4.00 + 3.33 3.23 −1.19 + 1.18 mm−3.33 + 2.80 11.70 −1.18 + 1.14 mm−2.80 + 2.36 45.30 −1.14 + 1.0 mm−2.36 + 2.00 22.85 −1.0 mm + 850�m−2.00 + 1.70 9.51 −850 + 710�m−1.70 + 1.40 4.89 −710 + 425�m<1.40 1.10 <425�m

The effect ofm on the uptake of phenol on BFA, ACnd ACL was studied and is shown inFig. 4. This figureeveals that the removal of phenol increases with increadsorbent dosage. The removal of phenol at adsorbent d

arger than 10 g/l remains almost unchanged. An increahe adsorption with the adsorbent dosage can be attribureater surface area and the availability of more adsorites. Atm < 5 g/l, the adsorbent surface become saturith phenol and the residual concentration in the solutio

arge. With increase in them, the phenol removal increasue to increased phenol uptake by the increases amf adsorbent. Atm > 5 g/l, the incremental phenol remo

BFA

Weight (%) Size (�m) Weight (%)

6.72 >1000 0.7317.21 −1000 + 850 0.8014.92 −850 + 710 3.5821.80 −710 + 600 1.42

32.07 −600 + 500 0.845.00 −500 + 355 3.842.23 −355 + 300 2.780.35 −300 + 150 19.20

−150 + 125 13.84−125 + 106 22.83−106 + 90 0.42−90 + 75 14.81−75 + 63 9.45−63 + 45 1.78<45 3.68


Fig. 1. Scanning electron micrograph of adsorbents at 150× magnification(1 cm = 100�m).

Fig. 2. Pore size distribution of adsorbents (a) BFA and (b) ACC.

becomes very low, as the surface phenol concentration and thesolution phenol concentration come to equilibrium with eachother. At aboutm = 10 g/l, the removal efficiency becomesalmost constant. There is very little difference between thephenol uptake by ACL and BFA.

3.3. Effect of initial pH (pH0)

The pH of the solution affects the surface charge of theadsorbents as well as the degree of ionization and specia-tion of different pollutants[34]. Change in pH affects theadsorptive process through dissociation of functional groupson the adsorbent surface active sites. This subsequently leadsto a shift in reaction kinetics and equilibrium characteris-tics of adsorption process. Adsorption of various anionic andcationic species on such adsorbents has been explained on thebasis of the competitive adsorption of H+ and OH− ions withthe adsorbates[35]. It is a common observation that the sur-face adsorbs anions favourably at lower pH due to presenceof H+ ions, whereas, the surface is active for the adsorptionof cations at higher pH due to the deposition of OH− ions[36].

The influence of the pH0 of phenolic solution on the extentof adsorption of phenol is shown inFig. 5. Adsorption of phe-nol decreases with increase in pH0. Up to pH0 6.5 the decreasei allyf iuma s to

n adsorption is gradual, which, however, drops drasticor pH0 > 6.5. The presence of oxides of aluminum, calcnd silicon on the adsorbents (BFA, ACC and ACL) lead


Fig. 3. X-ray diffraction patterns for BFA, ACC and ACL.

Fig. 4. Effect of adsorbent dosage on the removal of phenol. pH0: 6.5; T:30◦C; t: 5 h; C0: 100 mg/l.

Fig. 5. Effect of pH0 on the removal of phenol.T: 30◦C; t: 5 h;C0: 100 mg/l;ACL dosage: 10 g/l; BFA dosage: 10 g/l; ACC dosage: 12 g/l.

the development of charge when adsorbent is in contact withwater according to the pH of the solution as follows:

M OH + H+ → M OH2+

M OH + OH− → M O− + H2O

where M stands for Al, Ca and Si. Except silica, all otheroxides will possess positive charge for a pH range of inter-est because the zero point charge of SiO2, Al2O3 and CaOare 2.2, 8.3 and 11.0, respectively[37]. For pH0 below 6.5, asignificantly high electrostatic attraction exists between thepositively charged surface of the adsorbent and the phenolateion, C6H5O− ion. Phenol being a weak acid is adsorbed toa lesser extent at higher pH values as the negatively chargedsurfaces of the adsorbent did not favour the adsorption ofC6H5O− ion due to electrostatic repulsion. Such observa-tions were also reported by other workers[38]. FromFig. 5,it is observed that the adsorption of phenol by ACL is slightlyhigher than that by BFA for phenol while the ACC adsorptioncapacity is much lower. The phenol adsorption is limited bythe micropore volume and the acid–basic characteristics ofthe adsorbents[39]. Since the diameter of phenol moleculeis about 6A [40], the mesoporous adsorbents with sufficientpore surface area may show better adsorption characteristics.This is clearly discerned from the pore size characteristicss ofm pos-s forh ncet ngef pac-i entsw

3

s af to asedw ng

hown inTable 1andFig. 2. Due to higher surface areaesoporous structure of BFA than that of ACC, and

ibly favourable surface functionalities, BFA accountsigher phenol uptake than ACC. Similar reasons influe

he adsorption of phenol onto ACL as well. Since the charom acidic pH to alkaline pH reduces the adsorption caty of the adsorbents drastically, the rest of the experimere performed at slightly acidic pH (∼6.5).

.4. Effect of initial phenol concentration (C0)

The effect ofC0 on the extent of adsorption on BFA aunction of time is shown inFig. 6. At any time the amounf phenol adsorbed per unit weight of adsorbent increith increasingC0. The C0 provides the necessary drivi


Fig. 6. Effect of initial phenol concentrations on removal of phenol by BFA.pH0: 6.5;T: 30◦C; t: 5 h; BFA dosage: 10 g/l.

force to overcome the resistances to the mass transfer of phe-nol between the aqueous and the solid phases. The increasein C0 also enhances the interaction between phenol and theBFA. Therefore, an increase inC0 of phenol enhances theadsorption uptake of phenol.

3.5. Effect of contact time

Available adsorption results reveal that the uptake ofadsorbate species is fast at the initial stages of the contactperiod, and thereafter, it becomes slower near the equilib-rium. In between these two stages of the uptake, the rate ofadsorption is found to be nearly constant. This is obviousfrom the fact that a large number of vacant surface sites areavailable for adsorption during the initial stage, and after alapse of time, the remaining vacant surface sites are difficultto be occupied due to repulsive forces between the solutemolecules on the solid and bulk phases.

Aqueous phenol solutions with differentC0 were kept incontact with the adsorbents for 24 h. The values of the resid-ual concentrations at 5 h contact time were found to be higherby a maximum of∼2% than those obtained after 24 h contacttime. Therefore, after 5 h contact time, a steady state approx-imation was assumed and a quasi-equilibrium situation wasaccepted. Accordingly all the batch experiments were con-d kingc tion

F1

of phenol in solution versus contact time for the adsorbents atanC0 of 100 mg/l. The rate of phenol removal is found to bevery rapid during the initial 30 min and thereafter the rate ofphenol removal decreases. No significant change in phenolremoval is observed after about 120 min. It is also found thatthe removal of phenol by BFA is only slightly less than thatby ACL at any contact time and that the removal of phenolby ACC is around 15% less than that by BFA. Similar trendswere observed for differentC0, viz. 50, 150 and 200 mg/l. Itwas found that the adsorptive removal of the phenol ceasesafter 300 min of contacting with the three adsorbents.

3.6. Adsorption kinetic study

Four kinetic models, viz. pseudo-first-order, pseudo-second-order, Bangham and intraparticle diffusion models,were used to investigate the adsorption process of phenol onBFA, ACC and ACL.

3.6.1. Pseudo-first-order modelThe sorption of organic molecules from a liquid phase to

a solid phase can be considered as a reversible process withequilibrium being established between the solution and thesolid phase. Assuming a non-dissociating molecular adsorp-tion of phenols on BFA particles, the sorption phenomenoncan be described as the diffusion controlled process[41].

A

w dsor-ba Usingfi bateif n bee

w theait

l

w

k

q

(t ta igin.H nted

ucted with a contact time of 5 h under vigorous shaonditions.Fig. 7presents the plot of residual concentra

ig. 7. Effect of contact time on removal of phenol. pH0: 6.5;T: 30◦C; C0:00 mg/l; ACL and BFA dosage: 10 g/l; ACC dosage: 12 g/l.

+ SkA�kD

AS (5)

here A is adsorbate and S is the active site on the aent and AS is the activated complex.kA and kD are thedsorption and desorption rate constants, respectively.rst-order kinetics it can be shown that with no adsornitially present on the adsorbent (i.e.CAS0 = 0 at t = 0) theractional uptake of the adsorbate by the adsorbent caxpressed as:

XA

XAe= 1 − exp

[kACS + kA

KS

]t (6)

hereXAe is the fraction of the adsorbate adsorbed ondsorbent under equilibrium condition,KS = kA/kD and CS

s the adsorbent concentration in the solution Eq.(6) can beransformed as:

og10

(qe − q

qe

)= −kf

2.303t (7)

here

f =(

kACS + kA

kS

)(8)

= XA andqe = XAeThis equation is the so-called Lagergren equation[42]. Eq.

7)when plotted (not shown here) as log10((qe− q)/qe) versusfor the adsorption of phenolCA0 = 100 mg/l at 30◦C and apH0 6.5, showed a straight line passing through the orowever, after 15 min of contact time, the data is represe


by the two distinct straight lines. The adsorption data do notfit this equation, as theR2 values for BFA, ACC and ACL arefound to be 0.8931, 0.4996 and 0.9509, respectively.

3.7. Pseudo-second-order model

The pseudo-second-order model can be represented in thefollowing form [43]:

qt = tKSq2e

1 + tKSqe(9)

the initial sorption rate,h (mg/g min), ast → 0 can be definedas:

h = kSq2e (10)

The equilibrium adsorption capacity (qe) and the initialsorption rate (h) alongwith the pseudo-second-order con-stantkS were determined from the non-linear fitting of thedata. Kinetic parameters alongwith correlation coefficientfor pseudo-second order kinetic model are listed inTable 3.The calculated correlation coefficient is closer to unity forpseudo-second-order kinetic model than the pseudo-first-order kinetic model. Therefore, the sorption reaction can beapproximated more favourably by the pseudo-second-orderkinetic model for all the three adsorbents. Error functionsa do-s oft

3

thes s. Theo oneo ion,

surface diffusion and adsorption on the pore surface, or acombination of more than one step. In a rapidly stirred batchadsorption, the diffusive mass transfer can be related by anapparent diffusion coefficient, which will fit the experimentalsorption-rate data. Generally, a process is diffusion controlledif its rate is dependent upon the rate at which componentsdiffuse towards one another. The possibility of intra-particlediffusion was explored by using the intra-particle diffusionmodel[25,44–46].

qt = kidt1/2 + C (11)

where, kid is the intra-particle diffusion rate constant(mg/g min1/2) and C (mg/g) is a constant that gives ideaabout the thickness of the boundary layer, i.e., larger thevalue ofC the greater is the boundary layer effect[47]. Ifthe Weber–Morris[25] plot of qt versust0.5 gives a straightline, then the sorption process is controlled by intra-particlediffusion only. However, if the data exhibit multi-linear plots,then two or more steps influence the sorption process. Themathematical dependence of fractional uptake of adsorbateon t1/2 is obtained if the sorption process is considered to beinfluenced by diffusion in the cylindrical (or spherical) andconvective diffusion in the adsorbate solution. It is assumedthat the external resistance to mass transfer surrounding theparticles is significant only in the early stages of adsorption.T inearp ticled

peru efi thefi theso or-t ch

TK

A S (g/mg

P0.0396 50.0292 30.0370 7

A MPSD

W28.12 632.49 333.00 7

A

B

1

s shown inTable 3are also considerably less for pseuecond-order kinetic model reinforcing the applicabilityhe pseudo-second-order kinetic model.

.8. Intra-particle diffusion study

The adsorbate transport from the solution phase tourface of the adsorbent particles occurs in several stepverall adsorption process may be controlled either byr more steps, e.g. film or external diffusion, pore diffus

able 3inetic parameters for the removal of phenol by different adsorbents

dba h (mg/g min) qe (mg/g) k

seudo-second-order constantsACL 3.6784 9.6436ACC 1.4591 7.0689BFA 3.3379 9.4989

dba kid,1b R2 HYBRID

eber–Morris constantsACL 0.48742 0.9527 −19.8824ACC 0.32786 0.9820 −22.9889BFA 0.45137 0.9902 −23.3487

dba k0 (l/(g/l)) α

angham constantsACL 5.0262 0.3566ACC 3.6370 0.3091BFA 4.8033 0.3468

a Adsorbent.b Unit (mg/g min1/2).

his is represented by first sharper portion. The second lortion is the gradual adsorption stage with intra-pariffusion dominating.

Fig. 8 presents the plots of mass of phenol adsorbednit mass of adsorbent versust1/2 for all the adsorbents. In thgure the data points are related by two straight lines—rst straight portion depicting macropore diffusion andecond representing micro-pore diffusion[46]. These shownly the pore diffusion data. Extrapolation of the linear p

ions of the plots back to they-axis gives the intercepts, whi

min) R2 HYBRID MPSD

0.9999 −5.4414 11.6980.9998 −3.8630 6.1930.9999 −5.5468 12.080

kid,2b R2 HYBRID MPSD

77 0.0126 0.9591 −0.00021 0.14488 0.0286 0.9490 −0.0059 0.76560 0.0137 0.8967 −0.00033 0.179

R2 HYBRID MPSD

0.9326 −26.2841 41.60860.9515 −95.9866 103.5810.9305 −30.3758 46.6792


Fig. 8. Weber and Morris intra-particle diffusion plots for removal of phenol.pH0: 6.5;T: 30◦C;C0: 100 mg/l; ACL and BFA dosage: 10 g/l; ACC dosage:12 g/l.

provide the measure of the boundary layer thickness. Thedeviation of straight lines from the origin (Fig. 8) may be dueto difference in rate of mass transfer in the initial and finalstages of adsorption. Further, such deviation of straight linefrom the origin indicates that the pore diffusion is not thesole rate-controlling step. The adsorption data forq versust1/2 for the initial period show curvature, usually attributedto boundary layer diffusion effects or external mass transfereffects[45,48,49]. The slope of the Weber and Morris plots –q versust1/2 – are defined as a rate parameter, characteristicof the rate of adsorption in the region where intra-particle dif-fusion is rate controlling. The values of rate parameters (kid,1andkid,2) as given inTable 3show that ACL has highest valueof thekid,1 followed by BFA and ACC in that order but thetrend is just opposite forkid,2. The values of error functionsare given inTable 3. They show that the Weber–Morris modelshows better representation of the data than pseudo-first orderkinetic model.

3.8.1. Determination of diffusivityKinetic data could be treated by models given by Boyd et

al. [50], which are valid under the experimental conditionsused. With diffusion rate controlling in the adsorption onparticles of spherical shape, the solution of the simultaneousset of differential and algebric equations leads to:

F

[ ]

w ma tesi ntp eB ylin-d lacer tobw

F

wherebn’s are roots ofJ0(bnpR) = 0.Vermeulen’s approximation[23] of the Eq.(12) fits the

whole range 0 <F(t) < 1, for adsorption on spherical particles.

F (t) =[1 − exp

(−π2Det

R2a

)]1/2

(14)

This equation could further be simplified to cover most of thedata points for calculating effective particle diffusivity.

ln

[1

1 − F2(t)

]= π2Det

R2a

(15)

Thus the slope of the plot of ln[1/(1− F2(t))] versust wouldgive De. Table 4also presents the value of effective diffu-sion coefficient (De) as calculated from Eq.(15). Value ofDe is 1.800, 1.052 and 0.388× 10−10 m2/s, respectively, forphenol adsorption on ACL, BFA and ACC. This shows thatACL has highest overall pore diffusion rate.Table 4showsthe values of the effective pore diffusivities and the homoge-neous surface diffusion model (HSDM) surface diffusivitiesfor the adsorption of phenol on activated carbon and otheradsorbents. Wide variation is seen in the data with minimumvalue of surface diffusivity being 3.5× 10−13 m2/s and themaximum value being that ofDe for wood derived activatedcarbon,De = 3.80× 10−8 m2/s at 21◦C. Our values are for atemperature of 30◦C, but fall in between those forDs andD

3ore-

d orp-t

l

w ataa ea ver,E ata,i oresoi cesso amp luesa .60e

3

fort t toe riumc d tod um-b eringh tions

(t) = 1 − 6

π2

∞∑z=1

1

z2 exp−z2π2Det

R2a

(12)

hereF(t) = qt/qe is the fractional attainment of equilibriut timet, De the effective diffusion coefficient of adsorba

n the adsorbent phase (m2/s),Ra the radius of the adsorbearticle assumed to be spherical (m), andz is an integer. ThFA particles are not spherical; they may be taken as crical particles. If one assumes that the diffusion takes padially with diffusion in the angular and axial directione negligible, one gets the solution given by Skelland[51],hich after rearrangement is:

(t) = 1 − 4

π2

∞∑z=1

1

b2n

exp[−Deb2nt] (13)

e.

.8.2. Bangham’s equationKinetic data can further be used to check whether p

iffusion is the only rate-controlling step or not in the adsion system using Bangham’s equation[52].

og log

(C0

C0 − qtm

)= log

(k0m

2.303V

)+ α log(t) (16)

hereα (<1) andk0 are constants. If the experimental dre represented by Eq.(16), then it is an indication that thdsorption kinetics is limited by the pore diffusion. Howeq. (16) does not give a good fit of the experimental d

ndicating thereby that the diffusion of adsorbate into pf the sorbent is not the only rate-controlling step[53]. With

ncrease in the contact time, the effect of diffusion pron overall sorption could be ignored. Values of Bangharameters, correlation coefficient and error function vare given inTable 3. MPSD values are greater than 41xposing poorer fit of the model.

.9. Adsorption equilibrium study

To optimize the design of an adsorption systemhe removal of adsorption of adsorbates, it is importanstablish the most appropriate correlation for the equiliburves. Various isotherm equations have been useescribe the equilibrium nature of adsorption. Large ners of researchers in the field of environmental engineave used Freundlich and Langmuir isotherm equa


Table 4Comparison of effective pore diffusivity and HSDM surface diffusivities for adsorption of phenol on activated carbon

Adsorbent Conditions De × 1010 (m2/s) Reference

AC (coal; cecarbon GAC 40) T = 21◦C 141.0 [10]AC (coconut shell; 208C) T = 21◦C 276.0 [10]AC (wood; Pica 103) T = 21◦C 380.0 [10]AC (Coal; SP 207A) T = 21◦C 368.0 [10]AC (straw type 5; 800/200/12) T = 21◦C 221.0 [10]AC (tyre type 6; 900/100/18) T = 21◦C 263.0 [10]Granular AC pH = 3,T = 21◦C 0.012a [65]Granular AC pH = 7,T = 21◦C 0.0076a [65]Granular AC pH = 11,T = 21◦C 0.0035a [65]AC – 0.0138a [66]BFA pH = 6.5,T = 30◦C 1.052 Present workACC pH = 6.5,T = 30◦C 0.388 Present workACL pH = 6.5,T = 30◦C 1.800 Present work

AC: activated carbon.a HSDM surface diffusivity.

to represent equilibrium adsorption data using activatedcarbon–organic contaminants systems. This, despite the factthat these equations have serious limitations on their usage,the most popular Freundlich isotherm is suitable for highlyheterogeneous surfaces, however, it is valid for adsorptiondata over a restricted range of concentrations. For highlyheterogeneous surfaces and extremely low concentrations,Henry’s law is valid. However, Freundlich equation[54] doesnot approach Henry’s law at vanishing concentrations. TheLangmuir equation[55], although follows Henery’s law atvanishing concentrations, is valid for homogeneous surfaces.Thus, both these isotherm equations may not be suitablefor phenol adsorption on activated carbons and BFA for thewhole range of concentrations used in the study. Temkinisotherm contains a factor that explicitly takes into accountthe interactions between adsorbing species and the adsorbate.This isotherm assumes that (i) the heat of adsorption of all themolecules in the layer decreases linearly with coverage due toadsorbate–adsorbate interactions, and (ii) adsorption is char-acterized by a uniform distribution of binding energies, up tosome maximum binding energy[56,57]. Radke and Prausnitz[58] presented a simple equation, based on thermodynamicideal solution concept, for dilute solutions. The Redlichand Peterson equation[59] and the Toth equation[60] arethree parameter-equations, often used to represent soluteadsorption data on heterogeneous surfaces. These equationsr

aina ns.N tionsg son,R forp thei singt arame reef rms)w rsion

5.0 for Windows to minimize the deviation between calcu-lated and experimental values.

3.9.1. Choosing best isotherm modelSince each of the error functions produce a different

set of isotherm parameters, an overall optimum parame-ter set is difficult to identify directly. Thus, a normalisa-tion of each parameter is employed in order to have a bet-ter comparison between the parameter sets for the singleisotherm model[31]. In the normalisation processes firsteach error function was selected in turn and the results foreach parameter set were determined. Secondly, the errorsdetermined for a given error function were divided by themaximum to obtain the normalised errors for each parameterset. Lastly, the normalised errors for each parameter set weresummed up. Parameters as calculated for different isothermsfor all phenol–adsorbent systems are tabulated inTable 5.By comparing the results of the values for the error func-tion and correlation coefficients (Table 5), similar ‘best-fit’results for phenol adsorption on ACL, ACC and BFA areobtained. According to them, Redlich–Peterson isotherm bestfitted the equilibrium data for all phenol–adsorbent systems.Fig. 9 presents how well the six equations fit the data for

F

educe to Henry’s equation at very low concentrations.Due to experimental limitations, we could not obt

dsorption equilibrium data at very low concentratioone the less, we tried to use the six isotherm equaiven by, Freundlich, Langmuir, Temkin, Redlich–Peteradke–Prausnitz and Toth to fit the experimental datahenol on ACL, BFA and ACC. For each system and

sotherm equation all the experimental data were used. Uhe least-squares data reduction method, the isotherm pters (two for Freundlich, Langmuir and Temkin; and th

or Redlich–Peterson, Radke–Prausnitz and Toth isotheere obtained using a statistical software Statistica ve

-

ig. 9. Equilibrium isotherms for the removal of phenol by BFA at 30◦C.


Table 5Isotherm parameters for different phenol–adsorbent systems at 30◦C

Adsorbent KF [(mg/g)/(mg/l)1/n]

1/n R2 HYBRID MPSD

Freundlicha

BFA 4.5782 0.3678 0.9945 1.6465 6.5314ACC 2.4143 0.5031 0.9774 −0.1389 1.6692ACL 7.3771 0.3146 0.99451 −0.4508 6.4218

Adsorbent qm (mg/g) KL (l/mg) R2 HYBRID MPSD

Langmuirb

BFA 23.832 0.0884 0.9897 −1.3836 6.7864ACC 30.2187 0.0291 0.9916 2.74078 9.1324ACL 24.6458 0.2391 0.9543 3.7889 16.0369

Adsorbent KT (l/mg) B1 R2 HYBRID MPSD

Temkinc

BFA 5.0303 0.9317 0.9965 0.1702 3.2803ACC 0.3023 6.5015 0.9864 0.6562 10.2645ACL 3.4117 4.6443 0.9844 0.2722 8.2107

Adsorbent KR (l/g) aR (l/mg) β R2 HYBRID MPSD

Redlich–Petersond

BFA 4.4167 0.5233 0.7643 0.9976−0.3101 3.1300ACC 7.3820 2.5652 0.4384 0.9994 0.0242 1.5865ACL 57.3165 7.1016 0.7078 0.9945−0.3903 6.3051

Adsorbent kRP [(mg/g)/(mg/l)(1/P)]

KRP (l/g) P R2 HYBRID MPSD

Radke–Prausnitze

BFA 4.1629 13013.63 2.414 0.9955−6.6353 8.9522ACC 1.8300 9882.36 1.654 0.6806−8.7421 12.2868ACL 6.9933 9953.78 2.823 0.9913−4.8782 9.4707

Adsorbent q∞e (mg/g) KTh

[(mg/l)Th]Th R2 HYBRID MPSD

Tothf

BFA 53.0560 1.6612 0.3501 0.9971−0.3729 3.5018ACC 47.7308 7.9121 0.5878 0.9968 1.4123 5.6561ACL 80.6903 0.7791 0.2309 0.9914 0.4199 6.5198

a qe = KFC1/ne .

b qe = qmKLCe1+KLCe

.c qe = B1 ln KT + B1 ln Ce.d qe = KRCe

1+aRCβe

.

e 1qe

= 1KRPCe

+ 1

kRPC1/Pe

.

f qe = q∞e Ce

[KTh+CThe ]

1/Th .

phenol–BFA system. Following equilibrium relationships arerecommended for the three adsorbate–adsorbent systems.

qe = 57.3165Ce

1 + 7.1016C0.7078e

(phenol–ACL system) (17)

qe = 7.3820Ce

1 + 2.5652C0.4348e

, (phenol–ACC system) (18)

qe = 4.41670Ce

1 + 0.5233C0.7643e

, (phenol–BFA system) (19)

Several authors have reported Freundlich and Langmuirconstants for adsorption of phenol on various adsorbents. TheFreundlich and Langmuir constants values obtained in someof these works, although under different environmental con-ditions, are compared with the values obtained in the presentwork in Table 6. It may be seen that the isotherm parametersdiffer widely in their values for activated carbons of differ-ent origins. Hence one should be cautious while using thesevalues in design of adsorption systems.

3.10. Effect of temperature

Temperature has a pronounced effect on the adsorptioncapacity of the adsorbents.Fig. 10shows the plots of adsorp-tion isotherms,qe versusCe for phenol–BFA system at dif-ferent temperatures ranging from 25 to 45◦C. It shows thatwith the increase in temperature the adsorptivity of phenolincreases. This figure also shows that at lower adsorbateconcentrations,qe rises sharply and thereafter the increaseis gradual with solute concentration in the solution. Sincesorption is an exothermic process, it would be expected thatan increase in temperature of the adsorbate–adsorbent sys-tem would result in decreased sorption capacity. However, ifthe adsorption process is controlled by the diffusion process(intraparticle transport-pore diffusion), the sorption capacitywill show an increase with an increase in temperatures. Thisi s ane re,t rdingf reas-i howne nt isn essc , thei aturem enols h thei inves-t bona ion.

per-a

F forp

s basically due to the fact that the diffusion process indothermic process[24]. With an increase in temperatu

he mobility of the phenolate ions increases and the retaorces acting on the diffusing ions decrease, thereby incng the sorptive capacity of adsorbent. As has been sarlier, the diffusion of adsorbate into pores of the sorbeot the only rate-controlling step, and the diffusion procould be ignored with adequate contact time. Thereforencrease in sorption capacity with an increase in temper

ay be attributed to chemisorption. The increase in phorption capacity of the carbonaceous adsorbents witncrease in temperature has also been reported by otherigators[38,61]. These investigators used activated carnd activated date pits, respectively, for phenol adsorpt

Isotherm parameters for six isotherms at different temtures for phenol–BFA system are given inTable 7along with

ig. 10. Equilibrium adsorption isotherms at different temperaturehenol–BFA system. pH0: 6.5; BFA dosage: 10 g/l.


Table 6Freundlich and Langmuir constants for adsorption of phenol on various adsorbents

Adsorbent KF [(mg/g)/(mg/l)1/n] 1/n qm (mg/g) KL (l/mg) Reference

AC – – 216.43 0.0690 [9]Na–Y zeolites – – 75.28 0.0096 [9]Ni/Na–Y zeolites – – 84.69 0.00106 [9]AC (coal; cecarbon GAC 40) 36.902 0.293 – – [10]AC (coconut shell; 208C) 32.083 0.278 – – [10]AC (wood; pica 103) 9.9906 0.436 – – [10]AC (coal; SP 207A) 29.300 0.245 – – [10]AC (straw type 5; 800/200/12) 60.871 0.103 – - [10]AC (tyre type 6; 900/100/18) 26.436 0.272 – – [10]AC 39.97 0.188 1.1989 3.4443 [11]AC (pinewood-based) 23.33 0.256 240.6 0.034 [12]Dried activated sludge 15.10 0.45 236.8 0.0146 [13]Amnerlite XAD-4 resin 0.259 1.639 – – [14]NJ-8 0.808 2.356 – – [14]Modified-bentonite 3.64 0.464 – – [15]A-pillared bentonite 1.81 0.605 – – [15]CTAB-bentonite 1.48 0.642 – – [15]Thermal bentonite 1.30 0.652 – – [15]TiO2 surface 0.676 0.919 – – [16]Bentonite 0.2013 4.840 – – [17]Peat 0.0362 1.570 – – [17]Fly ash 19.8 0.26 [17]Dried aerobic activated sludge 1.28 0.820 194.2 0.004 [18]Bentonite 0.100 0.473 1.712 0.0141 [19]Amnerlite XAD-16 resin 0.0748 0.8720 1.5029 0.0511 [20]AC 37.0 0.17 309.7 0.0533 [21]Filtrasorb (F-400) 36.3 0.60 205.1 0.0420 [21]HiSiv 1000 (zeolite-Y structure) 0.047 0.70 319.0 0.00055 [21]Activated carbon cloth (ACC) – – 402.74 0.0253 [22]Pd (1%)/ACC – – 400.86 0.0176 [22]Pd (9.2%)/ACC – – 376.40 0.0156 [22]BFA 4.5782 0.3678 23.832 0.0884 Present workACC 2.4143 0.5031 30.2187 0.0291 Present workACL 7.3771 0.3146 24.6458 0.2391 Present work

AC: activated carbon; A-pillared: aluminum-hydroxypolycation as a pillaring agent; CTAB: cetyltrimethyl ammonium bromide.

the values of the correlation coefficient and the error func-tions values. On increasing the temperature from 25 to 45◦C,the value of the maximum adsorption capacity,qm increasesfrom 23.33 to 26.09 mg/g confirming endothermic natureof overall-sorption process.Table 7clearly reinforces ear-lier observation at 30◦C that the Redlich–Peterson isothermshows slightly better fit than all other isotherms. However,these values have been used to define and determine theisosteric heat of adsorption�Hst,0 that corresponds to zerosurface coverage (qe = 0).

3.11. Estimation of thermodynamic parameters

�S◦ can be determined by the slope of the linear Van’tHoff plot i.e. as lnK versus (1/T), using equation:

�H◦ =[R

d lnK

d(1/T )

](20)

�H◦ obtained here corresponds to isosteric heat of adsorp-tion (�Hst,0) with zero surface coverage (i.e.qe = 0) [62].Fig. 11 shows the Van’t Hoff’s plot for Redlich–Peterson

isotherm, from which �Hst,0= 503.88 kJ/kg and�S◦ = 1798.14 kJ/kg K have been obtained. For signif-icant adsorption to occur, the free energy changes ofadsorption,�G◦, must be negative. The thermodynamicsrelation between�G◦, �H◦ and �S◦ suggests that either(i) �H◦ is positive and�S◦ is positive and that the value

Fig. 11. Van’t Hoff plot of adsorption equilibrium constant,K usingRedlich–Peterson isotherm.


Table 7Isotherm parameters for phenol–BFA system at different temperatures

Temperature (◦C) KF [(mg/g)/(mg/l)1/n] 1/n R2 HYBRID MPSD

Freundlich25 3.6268 0.3993 0.9976 1.8853 8.217030 4.5782 0.3678 0.9945 1.6465 6.531435 5.2582 0.3528 0.9955 2.3981 7.980540 5.9182 0.3457 0.9951 2.9751 9.287145 7.1621 0.3146 0.9958 3.8357 12.3222

Temperature (◦C) qm (mg/g) KL (l/mg) R2 HYBRID MPSD

Langmuir25 23.331 0.0586 0.9846 −0.2849 3.627030 23.832 0.0884 0.9897 −1.3836 6.786435 24.645 0.1052 0.9901 −1.5441 6.762240 25.578 0.1282 0.9873 −1.7707 7.207045 26.088 0.1706 0.9811 −1.5823 6.9270

Tempeature (◦C) KT (l/mg) B1 R2 HYBRID MPSD

Temkin25 5.1695 0.5964 0.9912 0.3360 5.525830 5.0303 0.9317 0.9965 0.1702 3.280335 5.0602 1.1969 0.9979 0.3570 3.411940 5.1588 1.5278 0.9973 0.4322 4.006445 4.9700 2.3986 0.9963 0.5558 4.9311

Tempeature (◦C) KR (l/g) aR (l/mg) β R2 HYBRID MPSD

Redlich–Peterson25 3.8696 0.6336 0.7083 0.9991 −0.0712 2.784230 4.4167 0.5233 0.7643 0.9976 −0.3101 3.130035 5.5747 0.6003 0.7734 0.9995 −0.1202 1.425240 7.5220 0.7681 0.7697 0.9992 −0.1371 1.619345 13.438 1.3349 0.7661 0.9993 −0.0815 1.2575

Temperature (◦C) kRP [(mg/g)/(mg/l)(1/P)] KRP (l/g) P R2 HYBRID MPSD

Radke–Prausnitz25 3.0492 9938.99 2.1168 0.9945 −8.9098 11.463030 4.1629 13013.63 2.4143 0.9955 −6.6353 8.952235 4.6378 9210.27 2.3956 0.9948 −8.3398 10.748240 5.4111 9460.33 2.5049 0.9948 −5.4963 7.694745 6.6894 12097.61 2.7991 0.9913 −5.8189 8.2302

Temperature (◦C) q∞e (mg/g) KTh [(mg/l)Th] Th R2 HYBRID MPSD

Toth25 63.677 1.916 0.3288 0.9961 0.1991 3.162230 53.056 1.6612 0.3501 0.9971 −0.3729 3.501835 50.353 1.5528 0.3626 0.9992 −0.1845 1.862140 53.566 1.3541 0.3470 0.9989 −0.3067 2.057145 69.532 0.9161 0.2660 0.9990 −0.2052 1.8865

of T�S is much larger than�H◦, or (ii) �H◦ is negativeand�S◦ is positive or that the value of�H◦ is more thanT�S. Phenol adsorption is endothermic in nature, giving apositive value of�H◦. Hence,�S◦ has to be positive andthat the positive value ofT�S has to be larger than�H◦.The positive�Hst,0 value confirms the endothermic natureof the overall-sorption process. The adsorption process inthe solid–liquid system is a combination of two processes:(a) the desorption of the molecules of solvent (water)previously adsorbed, and (b) the adsorption of adsorbatespecies. The phenolate ions have to displace more than one

water molecule for their adsorption and this results in theendothermicity of the adsorption process. Therefore, the�Hst,0 will be positive. The positive value of�S◦ suggestsincreased randomness at the solid/solution interface withsome structural changes in the adsorbate and adsorbent andan affinity of the BFA towards phenol. Also, positive�S◦value corresponds to an increase in the degree of freedomof the adsorbed species[63]. �G◦ values were negativeindicating that the sorption process led to a decrease inGibbs free energy. Negative�G◦ indicates the feasibilityand spontaneity of the adsorption process.


Fig. 12. Adsorption isosters for determining isosteric heat of adsorption.

3.11.1. Isosteric heat of adsorptionApparent isosteric heat of adsorption (�Hst,a) at constant

surface coverage (qe = 5, 10, 15, 20, 25 mg/g) is calculatedusing Clausius–Clapeyron equation[64].

d lnCe

dT= −�Hst,a

RT 2 (21)

�Hst,a = Rd lnCe

d(1/T )

∣∣∣∣qe

(22)

For this purpose, the equilibrium concentration (Ce) at con-stant amount of adsorbed phenol is obtained from the adsorp-tion isotherm data at different temperatures.�Hst,a is cal-culated from the slope of the lnCe versus (1/T) for differ-ent amount of phenol adsorption on BFA (Fig. 12). The�Hst,a varies with the surface loading (Fig. 13), indicatingthat the BFA has an energetically heterogeneous surface.The variation in�Hst,a with surface loading can be alsobe attributed to the possibility of having lateral interactionsbetween adsorbed phenolate ions.

The apparent isosteric heat of adsorption in aqueousadsorption�Hst,a, is a resultant of net isosteric heat of adsorp-tion �Hst,net, heat of solution�Hsol, and heat of adsorptionof water,�Hw, i.e.:

�Hst,a = �Hst,net − �Hsol − f�Hw (23)

w oleo d

to be zero and�Hsol of the adsorbate in the solvent may becalculated from heats of formation data. Since�Hsol of phe-nol is 116.49 kJ/kg and�Hw should be negligible,�Hst,netascalculated from Eq.(23) is equal to 890.41, 713.59, 610.15,536.77 and 479.84 kJ/kg forqe = 5, 10, 15, 20 and 25 mg/g,respectively. Since heat of vapourisation of phenol is givenasλvap= 478.48 kJ/kg, therefore,�Hst,net lies between 1.00and 1.86 timesλvap.

4. Conclusions

The present study shows that the bagasse fly ash (BFA)is an effective adsorbent for the removal of phenol fromaqueous solution. Higher percentage of phenol removal byBFA, ACC and ACL was possible provided that theC0 in thesolution was low. Optimum conditions for phenol removalwere found to be pH0 ≈ 6.5, adsorbent dose≈10 g/l of solu-tion for phenol concentration up to 50 mg/l. The equilibriumbetween the adsorbate in the solution and on the adsorbentsurface was practically achieved in 5 h. Adsorption kineticswas found to follow second-order rate expression withinitial sorption rate being highest for adsorption on BFA.Equilibrium adsorption data for phenol on all the adsorbents,viz., ACL, ACC and BFA were best represented by theRedlich–Peterson isotherm. Adsorption of phenol on BFA isf re oft ss isc tivev l onB withs eouss tivec aters t duet anda

R

paa,

4)

nts,

ing

nolsy ofrbor,

16

00)

[ 67.

here,f is the exchange of number of moles of water per mf adsorbate at the adsorption site.�Hw is normally assume

Fig. 13. Variation of�Hst,a with respect to surface loading.

avourably influenced by an increase in the temperatuhe operation. This shows that the overall sorption proceontrolled by intraparticle diffusion of phenol. The negaalue of�G◦

adsindicate spontaneous adsorption of phenoFA. The values of isosteric heat of adsorption variedurface coverage implying an energetically heterogenurface of BFA. Overall, BFA showed excellent adsorpharacteristics for the removal of phenol from wastewamples and could be used as a very good adsorbeno its high carbon content and the presence of silicalumina.

eferences

[1] W. Kujawski, A. Warszawski, W. Ratajczak, T. Porebski, W. CaI. Ostrowska, Sep. Purif. Technol. 40 (2004) 123.

[2] H. Li, M. Xu, Z. Shi, B. He, J. Colloid Interf. Sci. 271 (20047.

[3] J.K. Fawell, S. Hunt, Environmental Toxicology: Organic PollutaHalsted Press, John Wiley & Sons, NY, 1988, 398.

[4] World Health Organization. International Standards for DrinkWater, Geneva, 1963, p. 40.

[5] J.S. Zogorski, S.D. Faust, Equilibria of adsorption of pheby granular activated carbon, in: A.J. Rubin (Ed.), ChemistrWastewater Technology, Ann Arbor Science Publishers, Ann AMichigan, 1978 (Chapter 9).

[6] C.T. Hsieh, H. Teng, J. Colloid Interf. Sci. 230 (2000) 171.[7] J.T. Paprowicz, Environ. Technol. 11 (1990) 71.[8] M.M. Swamy, I.D. Mall, B. Prasad, I.M. Mishra, Pollut. Series

(1997) 170.[9] B. Okolo, C. Park, M.A. Keane, J. Colloid Interf. Sci. 226 (20

308.10] M. Streat, J.W. Patrick, M.J.C. Perez, Water Res. 29 (1995) 4


[11] A.R. Khan, T.A. Al-Bahri, A. Al-Haddad, Water Res. 31 (1997)2102.

[12] R.L. Tseng, F.C. Wu, R.S. Juang, Carbon 41 (2003) 487.[13] Z. Aksu, J. Yener, Process Biochem. 33 (1998) 649.[14] A. Li, Q. Zhang, G. Zhang, J. Chen, Z. Fei, F. Liu, Chemosphere

47 (2002) 981.[15] S. Al-Asheh, F. Banat, L. Abu-Aitah, Sep. Purif. Technol. 33 (2003)

1.[16] D. Robert, S. Parra, C. Pulgarin, A. Krzton, J.V. Weber, App. Surface

Sci. 167 (2000) 51.[17] T. Viraraghavan, F.D.M. Alfaro, J. Hazard Mater. 57 (1998) 59.[18] Z. Aksu, D. Akpkinar, Sep. Purif. Technol. 21 (2000) 87.[19] F.A. Banat, B. Al-Bashir, S. Al-Asheh, O. Hayajneh, Environ. Pollut.

107 (2000) 391.[20] K. Abburi, J. Hazard Mater. B105 (2003) 143.[21] N. Roostaei, F.H. Tezel, J. Environ. Manage. 70 (2004) 157.[22] G.E. Shter, Y. Shindler, Y. Matatov-Meytal, G.S. Grader, M. Shein-

tuch, Carbon 40 (2002) 2547.[23] T. Vermeulen, Ind. Eng. Chem. 45 (1953) 1664.[24] W.J. Weber Jr., Physicochemical Processes for Water Quality Con-

trol, Wiley Interscience, New York, 1972, p. 206.[25] W.J. Weber Jr., J.C. Morris, J. Sanitary Eng. Div. ASCE 89 (SA2)

(1963) 31.[26] IS 1350 (Part I), Methods of Test for Coal and Coke Proximate

Analysis, Bureau of Indian Standards, Manak Bhawan, New Delhi,India, 1984.

[27] V.C. Srivastava, M.M. Swamy, I.D. Mall, B. Prasad, I.M. Mishra, J.Chem. Technol. Biotechnol. (2005), Communicated.

[28] J.F. Porter, G. McKay, K.H. Choy, Chem. Eng. Sci. 54 (1999) 5863.[29] D.W. Marquardt, J. Soc. Ind. Appl. Math. 11 (1963) 431.[30] I.D. Mall, V.C. Srivastava, N.K. Agarwal, I.M. Mishra, Colloids Surf.

[ em.

[[ for

[[ 224.[[ (7)

[38] F. Banat, A.A. Sameer, A.M. Leema, Chem. Eng. Technol. 27 (2004)80.

[39] A.P. Terzyk, J. Colloid Interf. Sci. 268 (2003) 301.[40] D.A. Mooney, F. Muller-Plathe, K. Kreneg, Chem. Phys. Lett. 294

(1998) 135.[41] H.S. Fogler, Elements of Chemical Reaction Engineering, 3rd ed.,

Prentice-Hall PTR, 1998.[42] S. Lagergren, Ksver. Veterskapsakad. Handl. 24 (1898) 1.[43] Y.S. Ho, G. McKay, Process Biochem. 34 (1999) 451.[44] V.J.P. Poots, G. McKay, J.J. Healy, J. Water Pollut. Control Fed. 50

(1978) 926.[45] G. McKay, M.S. Otterburn, A.G. Sweeney, Water Res. 14 (1980)

15.[46] S.J. Allen, G. Mckay, K.Y.H. Khader, Environ. Pollut. 56 (1989) 39.[47] K. Kannan, M.M. Sundaram, Dyes Pigments 51 (2001) 25.[48] J. Crank, The Mathematics of Diffusion, vol. 84, 1st ed., Oxford

Clarendon Press, London, 1965.[49] H.M. Asfour, M.M. Nassar, O.A. Fadali, M.S. El-Geundi, J. Chem.

Technol. Biotechnol. A 35 (1985) 28.[50] G.E. Boyd, A.W. Adamson, L.S. Meyers, J. Am. Chem. Soc. 69

(1947) 2836.[51] A.H.P. Skelland, Diffusional Mass Transfer, Wiley, NY, 1974.[52] C. Aharoni, S. Sideman, E. Hoffer, J. Chem. Technol. Biotechnol.

29 (1979) 404.[53] E. Tutem, R. Apak, C.F. Unal, Water Res. 32 (1998) 2315.[54] H.M.F. Freundlich, J. Phys. Chem. A 57 (1906) 385.[55] I. Langmuir, J. Am. Chem. Soc. 40 (1918) 1361.[56] M.I. Temkin, V. Pyzhev, Acta Physiochim. URSS 12 (1940) 327.[57] Y. Kim, C. Kim, I. Choi, S. Rengraj, J. Yi, Environ. Sci. Technol.

38 (2004) 924.[58] C.J. Radke, J.M. Prausnitz, AIChE J. 18 (1972) 761.[[[ hnol.

[[ ton,

[ ter-

[[

A: Physicochem. Eng. Aspects 264 (2005) 17.31] Y.C. Wong, Y.S. Szeto, W.H. Cheung, G. McKay, Process Bioch

39 (2004) 693.32] H.M. Stenzel, Chem. Eng. Prog. 89 (1993) 36.33] Powder Diffraction File (PDF), JCPDS International Centre

Diffraction Data, Swarthmore, PA, USA, 1979.34] H.A. Elliott, C.P. Huang, Water Res. 15 (1981) 849.35] P. Khanna, S.K. Malhotra, Indian J. Environ. Health 19 (1977)36] C. Huang, C.T. Stumm, J. Colloid Interf. Sci. 43 (1973) 409.37] K.K. Panday, G. Prasad, V.N. Singh, Environ. Technol. Lett. 50

(1986) 547.

59] O. Redlich, D.L. Peterson, J. Phys. Chem. 63 (1959) 1024.60] J. Toth, Acta Chem. Acad. Hung. 69 (1971) 311.61] P.R. Vijayalakshmi, V.J. Raksh, J. Rodriguez, J. Chem. Tec

Biotechnol. 71 (1998) 173.62] M. Suzuki, T. Fujii, AIChE J. 28 (1982) 380.63] C. Raymon, Chemistry: Thermodynamic, McGraw-Hill, Bos

1998, p. 737.64] D.M. Young, A.D. Crowell, Physical Adsorption of Gases, But

worths, London, 1962, p. 426.65] N.S. Abuzaid, G.F. Nakhla, J. Hazard Mater. 49 (1996) 217.66] V.K.C. Lee, G. McKay, Chem. Eng. J. 98 (2004) 255.

Documents

Removal of Phenol