Chemical Engineering Journal 228 (2013) 655–664

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier .com/locate /cej

Simultaneous co-adsorptive removal of phenol and cyanide from binarysolution using granular activated carbon

1385-8947/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.cej.2013.05.030

⇑ Corresponding author. Tel.: +91 1332 286651.E-mail addresses: bhumica.agarwal@gmail.com (B. Agarwal), chandfch@iitr.

ernet.in (C. Balomajumder), prabhat.thkr@gmail.com (P.K. Thakur).

Bhumica Agarwal ⇑, Chandrajit Balomajumder, Prabhat Kumar ThakurDepartment of Chemical Engineering, Indian Institute of Technology Roorkee, Roorkee, India

h i g h l i g h t s

� Optimization of co-adsorption of phenol and cyanide from synthetic coke waste water.� Multicomponent adsorption isotherms have been applied.� Kinetic modeling had been performed.� Thermodynamic parameters of the process have been envisaged.

a r t i c l e i n f o

Article history:Received 13 February 2013Received in revised form 5 May 2013Accepted 10 May 2013Available online 18 May 2013

Keywords:PhenolCyanideMulticomponentAdsorptionIsothermsKinetics

a b s t r a c t

The present study deals with the equilibrium, kinetic and thermodynamic modeling of simultaneous co-adsorption of phenol and cyanide from binary solution onto Granular Activated Carbon (GAC). The effectof process parameters like pH, temperature, adsorbent dose and contact time on the adsorptive efficiencyhas been evaluated. At an optimum pH 8, temperature 30 �C and adsorbent dose of 30 g/L, 79.9% of200 mg/L phenol and 93.6% of 20 mg/L cyanide were removed. Four multicomponent isotherms wereapplied to the experimental data conducted at an initial concentration range of 100–1000 mg/L. Singlecomponent isotherms viz. Langmuir and Freundlich were applied to determine the multicomponent iso-therm parameters. It was found that phenol adsorption followed extended Langmuir isotherm while cya-nide adsorption followed extended Freundlich isotherm in multicomponent system. The monolayeradsorption capacity of GAC was found to be 269.7 and 1.95 mg/g for phenol and cyanide, respectivelyas calculated by extended Langmuir isotherm. Adsorption of phenol and cyanide followed pseudo-secondorder kinetics indicating chemisorption to be the mechanism of adsorption. Thermodynamic parametersviz., DG0, DH0 and DS0 were �3.5174 KJ/mol, �10.326 KJ/mol, �0.0225 KJ/mol-K for phenol and�6.5575 KJ/mol, 14.044 KJ/mol and 0.0679 KJ/mol-K for cyanide adsorption, respectively. Thermody-namic studies established the process of phenol adsorption onto GAC as exothermic and spontaneousin nature while of cyanide as endothermic in nature.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

Remediation of wastewater from coke industries is of primeimportance in today’s scenario as it contains a number of pollu-tants in extremely high concentrations (Table 1). Among themphenol and cyanide have been found to be extremely toxic to bothhuman and aquatic life. The MCL (Maximum Contaminant Limit) ofphenol and cyanide in industrial discharge has been set as 0.5 mg/Land 0.2 mg/L, respectively by USEPA, WHO and CPCB, India [6].Exposure to even low concentrations of cyanide can cause coma,heart pains, breathing disorders, thyroid gland enlargement, head-

aches and even death. On the other hand, phenol exposure can leadto skin and eyes injuries, headache, vomiting, gastrointestinal dis-orders, central nervous system depression, lung, kidney, liver andheart damage ultimately leading to death [7,8].

In response to increasing health and environmental awarenessa number of physical, chemical, biological and combined methodshave been employed for removal of phenol and cyanide fromindustrial wastewater [6,9]. Adsorption of these toxicants has beenthe most widely accepted method for this purpose and has beenapplied by many researchers in single solute systems. However,co-adsorptive removal of phenol and cyanide has not been re-ported yet. Moreover equilibrium isotherms applicable to singlesolute systems viz. Langmuir, Freundlich, Redlich–Peterson, Toth,Temkin, etc. are not applicable to binary systems in most of thecases. The most probable reason could be the competitive nature

Nomenclature

BOD biological oxygen demand (mg/L)COD chemical oxygen demand (mg/L)q specific uptake capacity of adsorbent(mg/g of adsor-

bent)C0 initial pollutant concentration (mg/L)Ct pollutant concentration at time t (mg/L)V volume of the solution (L)M mass of the adsorbent used (g)qe specific uptake of adsorbent at equilibrium (mg/g of

adsorbent)KC equilibrium constantR gas constant (J/mol/K)DG0 change in Gibb’s free energy change (kJ/mol)T temperature (K)DS0 change in entropy (KJ/mol-K)DH0 change in enthalpy (kJ/mol)MPSD Marquardt’s percent standard deviationR2 coefficient of correlationqexp

e;i experimental specific uptake (mg/g)qcal

e;i calculated specific uptake (mg/g)N number of observations in the experimental isothermp number of parameters in the regression modelARE Average Relative Error

RL separation factorQmix adsorption capacity of one adsorbate in mixtureQ0 adsorption capacity of one adsorbate when present

aloneKf constant in Freundlich model (mg/g)/(mg/L)1/n

n constant in Freundlich modelQ0;i constant in modified Langmuir model for ith component

(mg/g)Ce;i concentration of ith component in the binary mixture at

equilibrium (mg/L)bi constant of ith component in Langmuir model (L/mg)qe;i amount of ith component adsorbed per gram of adsor-

bent at equilibrium (mg/g)KF;i constant in extended Freundlich constant for ith compo-

nent (mg/g)/(mg/L)1/n

k1 rate constant of pseudo-first order kinetic models (h�1)k2 rate constant of pseudo-second order kinetic model

(mg g�1 h�1)Ci intraparticle diffusion coefficient(mg g�1)kid intraparticle diffusion rate constant (mg g�1 h�0.5)xi, yi, zi constant in extended Freundlich model for ith compo-

nent

656 B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664

of both the anions and complexity of application of adsorption iso-therms owing to competition for the same binding sites as well assolute–surface interactions [10,11]. Thus modified and extendedisotherms have been proposed to compensate for inhibition ofadsorption of one component by another and vice versa. Multicom-ponent equilibrium modeling of various metals as well as phenolsand its derivatives have been studied extensively [11–13]. Butmulticomponent modeling of phenol and cyanide adsorption hasnot been reported till now to the best of our knowledge. For thepurpose of adsorption, GAC has been chosen as an adsorbent be-cause of its high adsorption capacity of phenol and cyanide as re-ported by many researchers [4,14].

The aim of present work is: (i) to determine the optimum pro-cess parameters viz., pH, temperature, adsorbent dose and contacttime for efficient co-adsorption of phenol and cyanide from binarysolution onto Granular Activated Carbon (GAC); (ii) to determinethe extent of competition and applicability of multicomponentequilibrium models; (iii) to study the kinetics of the process and(iv) to describe the thermodynamic nature of the process.

Table 1Composition of typical coke wastewater.

Reference/source pH Thiocyanates(mg/L)

Thiosulphates(mg/L)

Prasad and Singh 1989 [1]/IndianStandards 8073

8.5–9.5

50–100 110–220

Ghose et al. 2006 [2]/Jharia coalfieldDhanbad

8.2 – –

Mishra and Bhattacharya [3] 2006 – – –Vazquez et al. 2007 [4]Australia – 184 –Germany 200–500

Spain 215Machon et al. 2007 [5]/Arcelor group

steelworks Spain8.1 363 –

2. Materials and methods

2.1. Chemicals and adsorbent

All the chemicals used in this study were of analytical grade andobtained from Himedia Laboratories Pvt. Ltd. Mumbai India. Stocksolution containing 100 mg/L cyanide was prepared by dissolving0.25 g of KCN in 1 L of millipore water (Q-H2O, Millipore Corp. withresistivity of 18.2 MX-cm) whose pH was pre-adjusted to 10 using1N NaOH. Stock solution containing 1000 mg/L of phenol was pre-pared by dissolving 1 g of pure phenol crystal in 1 L of milliporewater. GAC was washed with millipore water and soaked in0.5 M H2SO4 for 24 h in 2:1 ratio of liquid to solid to increase thesurface area and pore volume of GAC. The adsorbent was thenwashed several times with millipore water and dried in hot airoven at 110 �C for 2 h to completely remove moisture, cooled toroom temperature and stored in polybags until further use. Thesurface area (BET) and total pore volume of the adsorbent was cal-culated by physisorption surface analysis on surface area analyzer

Total nitrogen(mg/L)

Cyanide(mg/L)

Phenol(mg/L)

BOD(mg/L)

COD(mg/L)

Chloride(mg/L)

800–1400 10–50 500–1000 – – 4000–4200

510 10.3 92.82 80.6 692.11 –

340 80 400 300 700 500

602 93 333 610 2200350–650 4–15 400–1200 1600–

26004000–6500

2255 50 485 1150 30301848 31.8 207 579 1102 1290

B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664 657

(ASAP 2010 Micrometrics, USA). Fourier Transform Infrared Spec-troscopy (FTIR, Nicolet 6700, USA) was employed to determinethe type of functional groups present on adsorbent’s surface beforeand after adsorption since the extent of adsorption depends greatlyon surface characteristics of adsorbent. The pellet to be analyzedwas obtained by compressing a mixture of sample and KBR in a ra-tio of 1:10 in hydraulic press at 15 ton pressure for 30 s. Unloadedand loaded samples were also visualized through SEM analysis totake into account the changes in surface morphology due to theadsorption of phenol and cyanide.

2.2. Batch experiments

Batch experiments for optimization of process parameters werecarried out in 250 mL round bottom flasks with working volume of100 mL at 125 rpm in an incubator cum orbital shaker (Metrex,MO-250, India). Multi-component system was obtained by takingconcentration of phenol and cyanide in the ratio of 10:1 as phenoland cyanide are discharged from coke wastewaters generally inthis ratio (Table 1). To avoid photo-oxidation of phenol the incuba-tor was covered with black cardboard properly throughout theexperiment. All the experiments were carried out in triplicatesand average results were used. Initial adsorbate concentrationswere selected as 200 mg/L for phenol and 20 mg/L for cyanide.The optimum pH and temperature were selected from a range ofpH 4–12 and temperature 20–40 �C. GAC dose of 5–40 g/L wereused for adsorption of phenol and cyanide and optimum dosewas decided on the basis of maximum percentage removal of phe-nol and cyanide. All the experiments were carried out for 48 hallowing sufficient time for achieving equilibrium. After every 2 hpH of the mixture was tested and readjusted to predefined valuewith 1N NaOH or HCl in case of any change during operation. Forstudy of adsorption isotherms initial concentrations of phenoland cyanide were varied from 100 to 1000 mg/L and 10–100 mg/L respectively. For study of adsorption kinetics an appropriate vol-ume of sample was withdrawn at an interval of every 2 h till theequilibrium conditions was achieved, filtered with standard What-man filter paper Cat No. 1001 125 and the filtrate was analyzed forcyanide and phenol by colorimetric picric acid and 4-aminoantipy-rene methods, respectively [15]. The amount of cyanide and phenoladsorbed per unit mass of the adsorbent was evaluated by the fol-lowing mass balance equation:

q ¼ ðC0 � CtÞV=M ð1Þ

where C0 is the initial pollutant concentration (mg/L) and Ct is thepollutant concentration (mg/L) at any time t, V is the volume ofthe solution (L) and M is the mass of the adsorbent used (g).

2.3. Equilibrium isotherms

Every adsorbate has a typical adsorption pattern for a particularadsorbent which could be specified with the help of simple equilib-rium models such as Langmuir model, Freundlich model, Tothmodel, Redlich–Peterson model, Radke–Prausnitz model, Temkinand Fritz–Schlundermodels. Among single component modelsLangmuir and Freundlich isotherms were used in the present studyand the equations used are described below [16]:

Langmuir: qe ¼ ðQ 0bCeÞ=ð1þ bCeÞ ð2Þ

Freundlich: qe ¼ KFC1=ne ð3Þ

In simple terms adsorption equilibrium models determine thetype of interaction between adsorbate and adsorbent at equilib-rium. However the real wastewater seldom contains a single pollu-tant. The presence of other pollutants tends to change the

adsorption pattern of the component of interest either by interac-tion between adsorbates or by competition for same binding sites.Additionally if the aim of the work is to optimize the simultaneousremoval of more than one component, it is desirable to keep an ac-count of the effect of such interactions as well as extent of adsorp-tion of one component in the presence of other component(s). Ingeneral amount of component adsorbed may increase, decreaseor remain unaltered in the presence of other component(s). Forsuch complex systems equilibrium data needs to be modeled usingvariants of single isotherm models discussed above. Eqs. (4)–(8)describes the multicomponent models used to model the equilib-rium data of phenol and cyanide adsorption [12,17–19].

Non-modified competitive Langmuir : qe;i

¼ ðQ0;ibiCe;iÞ=ð1þXN

j¼1

bjðCe;j=njÞÞ ð4Þ

Modified competitive Langmuir : qe;i

¼ ðQ0;ibiCe;i=niÞ=ð1þXN

j¼1

bjðCe;j=njÞ ð5Þ

Extended Langmuir: qe;i ¼ ðQ0;ibiCe;iÞ=ð1þXN

j¼1

bjCe;jÞ ð6Þ

Extended Freundlich: qe;1 ¼ ðKF;1Ce;11=n1þx1Þ=ðCx1e;1 þ y1Cz1

e;2Þ ð7Þ

Qe;2 ¼ ðKF;2C1=n2þx2e;2 Þ=ðCx2

e;2 þ y2Cz2e;1Þ ð8Þ

2.4. Kinetic studies

Kinetic modeling of an adsorption process is important frompractical point of view as it generates the time profile of utilizationof adsorption capacity of the adsorbent. An intensive study of liter-ature reveals that adsorption onto GAC follows either physisorp-tion (pseudo-first order kinetics) or chemisorptions (pseudo-second order kinetics). Adsorption on highly porous structures likethat of GAC is also affected by diffusional and mass transfer limita-tions [20]. Different models like pseudo-first order and pseudo-second order for the determination of the nature of adsorption:physical or chemical and additional model: Weber and Morris (orintraparticle) model for the study of mass transfer effects havebeen applied in the present study. Eqs. (9)–(11) describes variouskinetic models used in the present study [21–23].

Pseudo-first order : qt ¼ qeð1� expð�k1tÞÞ ð9Þ

Pseudo-second order : qt ¼ k2q2e t=ð1þ qek2tÞ ð10Þ

Intraparticle : qt ¼ kid � t0:5 ð11Þ

2.5. Thermodynamic studies

Linearized Van’t Hoff equation in the following form (Eqs. (12)and (13)), was used to calculate the thermodynamic parameterssuch as Gibbs’s free energy change (DG0), enthalpy change (DH0)and entropy change (DS0) during the process [24].

log Kc ¼ DS0=2:303R� DH0=2:303RT ð12Þ

DG0 ¼ �RT log Kc ð13Þ

where R is the gas constant = 8.314 J/mol/K, DG0 is the kJ/mol, T isthe temperature in K, DS0 is the KJ/mol-K, DH0 is the kJ/mol and

658 B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664

Kc is the equilibrium constant (amount of adsorbate on adsorbent/amount of adsorbate in solution).

2.6. Model validation

The Marquardt’s percent standard deviation (MPSD) was usedto validate the equilibrium adsorption data according to the fol-lowing equation:

MPSD ¼ 100�

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1

N � P

X qexpe;i � qcal

e;i

qexpe;i

!2vuut ð14Þ

where qexpe;i = experimental specific uptake (mg/g), qcal

e;i = calculatedspecific uptake (mg/g) for corresponding qexp

e;i , N is the number ofobservations in the experimental isotherm and P is the number ofparameters in the regression model. The smaller MPSD values indi-cate more accurate estimation of qe value [25].

To evaluate the goodness of fit of kinetic experimental data, thestatistical indice, Average Relative Error (ARE) between the exper-imental and calculated values was used [26]. The equation for eval-uating ARE is given by the following equation:

AREð%Þ ¼ 100=N �

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiX qexpe;i � qcal

e;i

qexpe;i

!22

vuut ð15Þ

The smaller ARE values indicate more accurate estimation of qt

values [25].

3. Results and discussion

3.1. Characterization of GAC

Table 2 demonstrates the surface area and pore volume of theraw, unloaded acid treated and loaded GAC. Acid treated GAC offersvery high surface area (23.8782 m2/g) making it a good adsorbentfor wastewater treatment (Table 2). As could be inferred from theTable 2, the surface area, monolayer volume as well as total porevolume of GAC decreases after adsorption indicating adsorptionof phenol and cyanide onto the GAC.

Fig. 1 depicts the morphology of unloaded and loaded GACas seen through scanning electron microscopy. The porousconfiguration and smooth morphology of GAC (Fig. 1a) makesit suitable adsorbent as it enhances the adsorption capacity. Achange in surface morphology from being smooth to roughand occupation of pores indicates adsorption of phenol andcyanide onto the surface and pores of GAC giving it a roughtexture (Fig. 1b).

As could be seen from Fig. 2 different FTIR spectra are obtainedfor acid treated GAC before and after adsorption. A number of addi-tional peaks, in the spectra of GAC after adsorption, viz. 2721, 1595and 1358 cm�1 could be observed which indicated adsorption ofphenol and cyanide onto GAC. Peak at 1595 cm�1 corresponds to–CH stretching due to the presence of conjugated hydrocarbongroups, aromatic hydrocarbons, carboxylic groups and carboxyland carbonate structures indicating adsorption of phenol whereaspeak at 1358 cm�1 corresponds to inorganic nitrates marking a po-tential adsorption site for cyanides. Shift in peak from 3438.28 to3431.03 cm�1corresponds to vibrations of –OH and –NH functionalgroups while shift in peak from 2922.92 to 2928.11 cm�1 corre-sponds to vibrations of –CH groups. Additional peak near720 cm�1 corresponds to rocking absorption due to the presenceof methylene groups.

3.2. Effect of process parameters

3.2.1. Effect of pHPhenol and cyanide are both easily dissociable species in aque-

ous solutions at controlled pH. As the extent of ionization and spe-ciation of organic and/or inorganic pollutants as well as surfaceproperties of the porous amphoteric adsorbent such as GAC varywith changing pH therefore pH plays an important role in optimiz-ing adsorptive removal of these compounds [14,27]. Fig. 3 repre-sents the effect of pH on adsorption of phenol and cyanide ontoGAC. Adsorption of phenol remains almost constant in acidic pHrange where the phenol remains in undissociated forms. Howeverat pH greater than its pKa (9.96) a sharp decrease in percentage re-moval is observed indicating adsorption of phenol mainly in itsundissociated form. A similar trend was observed by Kilic et al.[28] who found pH 7 to be most appropriate for phenol adsorption.

Since cyanide has a pKa of 9.39 maximum removal of cyanide isobtained in the region of pH 8–10. In lower pH range (4–8) there isa sharp decrease in percentage removal of cyanide which could beattributed to hydrolysis of weak acid dissociable cyanides to HCN.Since HCN is highly hydrophilic its tendency to be adsorbed at lowpH is markedly decreased. However at higher pH most of the cya-nide exists in undissociated form. Also hydration of GAC takesplace giving it the properties of an ion-exchanger thus readilyadsorbing cyanide ions from the solution [29,30]. At pH above 10the percentage removal of cyanide remains almost constant. Theresults are in compliance with several researches carried out re-cently on adsorption of cyanide by various adsorbents where opti-mum pH for maximum cyanide removal lies between 8 and 11[9,31,32]. From the above results (Fig. 3) pH 8 was selected for fur-ther optimization studies.

3.2.2. Effect Of temperatureExperiments were conducted to study the effect of temperature

on adsorption of phenol and cyanide. It was observed that adsorp-tion of cyanide increased when the temperature was increasedfrom 20 �C to 40 �C (Fig. 4). However, the increase in percentage re-moval was higher when the temperature changed from 20 �C to30 �C and became more gradual on further increasing the temper-ature. The initial increase in adsorption with increase in tempera-ture is probably due to increase in active adsorption sites resultingfrom the breaking of some of the internal bonds near the edge ofthe active surface sites of the adsorbent [33]. Also increase in tem-perature leads to decrease in solution viscosity and thus increase indiffusion rate of adsorbate within the pores [34]. On the other handadsorption of phenol onto GAC decreased with increasing temper-ature (Fig. 4) which could be due to the increased tendency ofdesorption of phenol at increased temperature resulting due toweakening of adsorptive forces between active sites of adsorbentand adsorbate as well as between the adjacent molecules of adsor-bate [35]. The results are in compliance with several other re-searches on adsorption of phenol and cyanide [28,36]. FromFig. 4 it could be inferred that percentage removal of cyanide(91.62–94.04%) and phenol (82.14–77.81%) does not vary consider-ably with variation in temperature (20–40 �C). Since at industriallevel narrow temperature ranges are preferred for pollution abate-ment temperature of 30 �C was selected for further studies [37].

3.2.3. Effect of adsorbent doseAs the adsorbent dose was increased percentage removal of

adsorbate increased owing to increase in surface area as well asnumber of possible active sites. However above a certain value per-centage removal either reaches an asymptotic value or decreases. Apossible explanation for this phenomenon could be partial overlap-ping of GAC resulting in decrease in effective surface area availablefor adsorption [28]. As is evident from Fig. 5 percentage removal of

Table 2Surface properties of unloaded and loaded GAC.

BET surfacearea (m2/g)

Monolayervolume (cm3/g)

Total porevolume (m3/g)

Raw GAC 6.5 1.495 0.0033Unloaded acid

treated GAC23.8782 5.485 0.0120

Loaded GAC 3.9099 0.898 0.0020

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

2721.30

717.57

3438.28

3431.03

2922.92

2928.11

1595.98

1358.71

(b)

(a)

Abs

orba

nce

B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664 659

phenol and cyanide increases on increasing GAC dose from 5 g/L to30 g/L but becomes almost constant above this dose. Therefore30 g/L GAC was selected as adsorbent dose for equilibrium and ki-netic studies.

Fig. 1. SEM photographs of (a) unloaded and (b) loaded GAG at 1000�magnification.

4000 3500 3000 2500 2000 1500 1000 500-0.10

Wavenumber (cm-1)

Fig. 2. FTIR spectrum of (a) acid treated GAC and (b) GAC after adsorption.

3.2.4. Effect of contact timeTo evaluate the time of equilibrium, when the concentration of

adsorbate in solution and on adsorbent does not change with time,contact time studies were carried out. From Fig. 6 it could be ob-served that after 18 h no or very little increase in percentage re-moval of phenol took place. Therefore it was assumed thatadsorption of phenol onto GAC in binary mixture achieved equilib-rium within 18 h. This result is in accordance with the studies ofDaifullah and Girgis [38] who reported 20 h as equilibrium timeof phenol onto GAC.

Similar study of cyanide adsorption revealed that adsorption ofcyanide onto GAC in the presence of phenol achieved asymptoticstate within 24 h. As could be observed from Fig. 6 during initialstage of the process, percentage removal of both phenol and cya-nide was extremely fast indicating physical adsorption involvingweak Van der Waal forces only. Dash et al. [9] attributed the rapiduptake of adsorbent to high concentration gradient between thebulk liquid and adsorbent’s surface. Thereafter adsorption of bothphenol and cyanide proceeded at almost a constant rate (4–16 h).However percentage removal rate was found to decrease withtime. This phase is indicative of the strong bond formation be-tween the adsorbent and adsorbate which requires increased timeas well as energy to overcome repulsive forces [33].

Fig. 3. Effect of varying pH onto adsorption of cyanide and phenol in a multicom-ponent system.

Fig. 4. Effect of increasing temperature on percentage removal of cyanide andphenol in a multicomponent system.

Fig. 6. Effect of contact time on percentage removal of cyanide and phenol in amulticomponent system.

660 B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664

3.2.5. Effect of initial concentrationFig. 7a and b represents the effect of increasing cyanide and

phenol initial concentration on their percentage removal respec-tively. It could be observed that percentage removal decreasedon increasing initial concentration however specific uptake, qe

(mg of adsorbate/g of adsorbent) increased. The percentage re-moval decreases from 85.59% to 33.76% and from 96.65% to71.58% in case of phenol and cyanide respectively on increasinginitial concentration of phenol from 100 to 1000 mg/L and initialconcentration of cyanide from 10 to 100 mg/L while specific uptakefor phenol and cyanide increases from 2.85 to 11.25 mg/g of adsor-bent and from 0.32 to 2.39 mg/g of adsorbent respectively. This in-crease in specific uptake could be attributed to greater massdriving force at higher concentration which helps in overcomingmass transfer limitations [39].

3.3. Equilibrium modeling

3.3.1. Estimation of parameters of single component modelsThe adsorptive equilibrium data for single components viz.,

phenol and cyanide onto GAC were analyzed using Microsoft Excel2010. Single component isotherm modeling was carried out to findthe best fit model among above mentioned isotherms and estima-tion of parameters for multicomponent adsorption studies. Table 3summarizes the values of parameters, coefficient of regression andMPSD obtained by the application of single component models

Fig. 5. Effect of adsorbent dose on percentage removal of cyanide and phenol in amulticomponent system.

along with data from other similar studies for comparison. Lang-muir isotherm predicts the specific uptake of phenol better thanFreundlich isotherm on the account of lower MPSD suggestingmonolayer adsorption. In case of cyanide experimental data wasclosely predicted by Freundlich model. The results are in compli-ance with the studies carried out by various researchers who foundLangmuir and Freundlich to respectively represent adsorption ofphenol and cyanide on GAC [26,40–43].

The Langmuir constant ‘b’ measures the affinity of the adsor-bent for the solute. It is clear from Table 3 that the values of bare more for cyanide than that of phenol, thereby indicating morestable cyanide:carbon interactions in comparison to phenol:carboninteractions [10]. However Qo values for phenol (13.005 mg/g)were found to be greater than that of cyanide (1.79 mg/g) whichcould be explained on the basis of higher concentration of phenolin the binary solution. Further the value of constant n gives an indi-cation about the favourability of adsorption. It has been mathemat-ically derived that values of n in the range 2–10 represent good, 1–2 moderately difficult, and less than 1 poor adsorption characteris-tics [44]. From Table 3 it is evident that the process of phenol andcyanide lies in the range of favorable process.

The parameters derived from equilibrium modeling are used todetermine the separation factor, RL which describes the feasibilityof adsorption process. RL is calculated according to the followingequation:

RL ¼1

1þ bC0ð16Þ

According to Hall et al. [45] a process is unfavorable whenRL > 1, linear adsorption when RL = 1, favorable adsorption for0 < RL < 1 and irreversible adsorption when RL = 0.

The values of RL for all the concentrations of phenol and cyanidelies between 0 and 1 (data not shown) indicating adsorption ofphenol and cyanide on GAC is very much favorable. However asthe initial concentration increased from 100 to 1000 mg/L and10–100 mg/L for phenol and cyanide respectively, the value of RL

decreased indicating favorable adsorption at lower concentration[46]. The value of n for phenol and cyanide was found to lie be-tween 1 and 10 indicating favorable adsorption of phenol and cya-nide onto GAC [47].

3.3.2. Estimation of parameters of multicomponent modelsThe multicomponent equilibrium data of phenol and cyanide

onto GAC have been fitted to different multicomponent isothermmodels i.e. non-modified Langmuir, modified Langmuir, extendedLangmuir and extended Freundlich (Table 4). It could be clearly

(a)

(b)Fig. 7. Effect of increasing initial concentration of (a) cyanide and (b) phenol onpercentage removal in a multicomponent system.

Table 4Parameters of phenol and cyanide adsorption onto GAC as estimated by multicom-ponent modeling.

Isotherm model Parameter Phenol Cyanide

Non-modified Langmuir MPSD 56.4295 38.4918Modified Langmuir nj 0.77963 1.01994

MPSD 48.691 43.691Qo,i 269.7 1.94784

Extended Langmuir bi 0.00087 0.41316MPSD 14.9837 26.3608xi �3.60401 0.27716

Extended Freundlich yi 5 � 10�7 3.96512zi �3.0925 �0.55523MPSD 15.43329 13.66

B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664 661

seen from the Table 4 that non-modified Langmuir which uses thecoefficients from single component Langmuir isotherm could notbe used for prediction of qe in case of multicomponent adsorptionof phenol and cyanide because of high values of MPSD. The use ofcorrection factor i.e. 0.77963 and 1.0994 for phenol and cyanide,respectively lowered MPSD values in case of modified Langmuirmodel indicating better fit. However best fit was obtained by appli-cation of extended Langmuir and extended Freundlich for adsorp-tion of phenol and cyanide, respectively.

In a multicomponent system, there are following three possibil-ities: synergism (the effect of the mixture is greater than that ofeach of the individual effects of the constituents in the mixture),antagonism (the effect of the mixture is less than that of each ofthe individual effects of the constituents in the mixture) andnon-interaction (the effect of the mixture is neither more nor lessthan that of each of the individual effects of the constituents in the

Table 3Parameters from single component modeling of adsorption of phenol and cyanide.

Reference Adsorbate/adsorbent Langmuir

Q0 b

This study Phenol/GAC 13.005 0.025This study Cyanide/GAC 1.79 0.87Kumar et al. [40] Phenol/GAC 216.1 0.032Vasu [41] Phenol/AC 0.79 9.78Maroof et al. [42] Phenol/NAC D10 166.67 0.5Behnamfard and Salarirad [26] Cyanide/AC 47.62 0.007Giraldo and Pirajan [43] Cyanide/ AC 226.76 0.029

a Comparison made between Langmuir and Freundlich models only.

mixture) [48]. Effect of interaction of two adsorbates could be ac-counted by the ratio of adsorption capacity of one adsorbate (Qmix

or KFmix) in the mixture to that of the same adsorbate when presentalone (Q0 or KF) according to following conditions [49]:

Qmix=Q 0 > 1; synergism

Qmix=Q 0 ¼ 0; non-interaction and

Qmix=Q 0 < 1; antagonism

In the present study the ratio of Qmix/Q0 for phenol and cyanidewas calculated as 20.73and 1.09 hence establishing the fact thatphenol and cyanide show synergism in the process of adsorptionfrom wastewater. Though the synergistic adsorption mechanismis not very clear, synergism could be attributed to hydrogen bond-ing between lone pair of electron of nitrogen atom of cyanide andhydrogen atom of hydroxyl group of phenol. In a similar kind ofstudy with phenol and aniline, Zhang et al. [50] attributed hydro-gen bonding to be the counteracting force against the electrostaticrepulsion of aniline and phenol molecules.

3.4. Kinetic modeling

To determine whether the adsorption of phenol and cyanide oc-curs through physisorption or chemisorptions, pseudo-first orderand pseudo-second order kinetic models were applied and the re-sults are summarized in Table 5 and Fig. 8. In case of phenol kineticdata is well predicted by pseudo-second order data which is alsoconfirmed by lower ARE values. However in case of cyanide bothpseudo-first and second order kinetics describes the experimentaldata well. Since ARE obtained from pseudo-second order is lowerthan that of pseudo-first order, it could be safely said that cyanideadsorption follows pseudo-second-order kinetics. Thus both phe-nol and cyanide are adsorbed through chemisorption on activatedcarbon [32].

To determine the nature of diffusion a plot of qt vs. t0.5 wasdrawn (Fig. 9). From the figure it could be inferred that in case of

Freundlich Best fit model

MPSD KF n MPSD

5.62 1.407 2.782 20.24 Langmuir61.03 0.89 4.35 24.75 Freundlich– 25.41 2.618 – Langmuira

– 14.24 2.69 – Langmuir– 52.589 1.337 – Langmuir and Freundlich12.12 3.14 2.49 3.69 Freundlicha

– 11.98 1.63 – Langmuir and Freundlich

Table 5Parameters of kinetic and intraparticle modeling of adsorption of phenol and cyanideon GAC.

Component Phenol Cyanide

Pseudo-first order qe,cal 5.06 0.66K1 0.191 0.13ARE 3.05 0.90

Pseudo-second order qe,cal 6.88 0.89K2 0.024 0.12ARE 1.78 0.79

Intraparticle Kid1 1.266 0.138

R21

0.992 0.962

Kid2 0.007 0.002

R22

0.958 0.997

(a)

662 B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664

both phenol and cyanide two clearly demarcated regions existswhich are indicative of both surface and intraparticle diffusion.However better fit of data in first region (R2

1 > R22 from Table 5) indi-

cates the dominance of surface diffusion in case of phenol adsorp-tion while rate limiting step in cyanide adsorption is intraparticlediffusion (R2

1 < R22) [51]. Importance of intraparticle diffusion as a

rate controlling agent is also reinforced from the fact that the plotdoes not pass through the origin [52].

3.5. Thermodynamic modeling

A graph between log kc vs. 1/T was plotted and DH0 and DS0

were calculated from the slope and intercept respectively and aresummarized in Table 6.

(a)

(b)Fig. 8. Comparative plot of experimental and calculated values of qt by pseudo-firstand second order kinetic model for (a) phenol adsorption and (b) cyanideadsorption.

(b) Fig. 9. Weber and Morris plot of (a) cyanide and (b) phenol adsorption on GAC.

Table 6Parameters from thermodynamic modeling.

Adsorbate Temperature(�C)

DG0 (KJ/mol)

DH0 (KJ/mol)

DS0 (KJ/mol-K)

Phenol 20 �3.7173 �10.326 �0.022530 �3.517440 �3.2655

Cyanide 20 �5.8267 14.044 0.067930 �6.557540 �7.182

Thermodynamic study of phenol adsorption reveals negativevalues of DG0, DH0 and DS0 which are indicative of feasible, exo-thermic and spontaneous nature of the process [53,54]. Negativevalue of DH0 also indicates free diffusion of adsorbate moleculesthrough bulk phase [55]. On the contrary positive values of DH0andDS0 for cyanide adsorption indicate endothermic process with in-creased randomness at the surface of solute–solution interfaceduring adsorption. Further the value of DS0 (which is less than 1)indicates that process is highly reversible for both phenol and cya-nide [24].

4. Conclusion

The process of simultaneous adsorption of phenol and cyanidefrom binary solution was studied at various pH, temperature andadsorbent doses. At an optimum pH 8, temperature 30 �C and

B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664 663

adsorbent dose 30 g/L, 79.9% of phenol and 93.6% of cyanide wasremoved from binary solution containing 200 mg/L of phenol and20 mg/L of cyanide. Equilibrium isotherm, kinetic and thermody-namic studies were also carried out at optimized parameters. Bothsingle and multicomponent equilibrium isotherms were applied inorder to explore the effect of presence of one adsorbate on theadsorption of other. Phenol adsorption in multicomponent systemwas found to be expressed by extended Langmuir isotherm whilecyanide adsorption followed extended Freundlich isotherm. The ef-fect of presence of phenol and cyanide on each other was also eval-uated and it was found to be synergistic. Kinetic studies pertainingto adsorption of phenol and cyanide revealed chemisorption to bethe mechanism of adsorption and the process was governed byboth surface and intraparticle diffusion. Adsorption of phenol ontoGAC was found to be exothermic and spontaneous while that ofcyanide was endothermic with increased randomness as depictedby thermodynamic studies.

Acknowledgements

The authors are thankful to Ministry of Human Resource Devel-opment, Government of India and Institute’s Instrumentation Cen-ter, IIT Roorkee for extending their financial and technical supportfor present research work.

References

[1] B. Prasad, G. Singh, Characterization of coke-plant effluents with subsequenttreatment – an appraisal, Indian J. Environ. Prot. 9 (7) (1989) 525–529.

[2] M.K. Ghose, S.K. Bhasa, A. Jena, Sustainable technologies for coke plant effluenttreatment: an Indian case study, Indian Chem. Eng. 48 (4) (2006) 278–287.

[3] S. Mishra, J. Bhattacharya, Potential of leaf litter for phenol adsorption – akinetic solution, Indian J. Chem. Eng. 13 (3) (2006) 298–301.

[4] Vazquez, J.R. Iglesias, E. Maranon, L. Castrillon, M.A. Ivarez, Removal ofresidual phenols from coke wastewater by adsorption, J. Hazard. Mater. 147(1–2) (2007) 395–400.

[5] Machon, H. Lopez, J.R. Iglesias, E. Maranon, I. Vazquez, Simulation of a cokewastewater nitrification process using a feed-forward neuronal net, Environ.Modell. Softw. 22 (9) (2007) 1382–1387.

[6] G. Busca, S. Berardinelli, C. Resini, L. Arrighi, Technologies for removal ofphenol from fluid streams: a short review on recent developments, J. Hazard.Mater. 160 (2–3) (2008) 265–288.

[7] ATSDR, Agency for Toxic Substances and Disease Registry, Public HealthStatement, Cyanide, 2006 <http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=72&tid=19>.

[8] ATSDR, Agency for Toxic Substances and Disease Registry, Toxicological profilefor Phenol, US, Department of Health and Human Services, Public HealthService, 2008, <http://www.atsdr.cdc.gov/toxprofiles/tp.asp?id=148&tid=27>.

[9] R.R. Dash, C. Balomajumder, A. Kumar, Treatment of cyanide bearing water/wastewater by plain and biological activated carbon, Ind. Eng. Chem. Res. 48(7) (2009) 3619–3627.

[10] D. Mohan, S. Chander, Single component and multi-component adsorption ofmetal ions by activated carbons, Colloids Surf. A 177 (2–3) (2001) 183–196.

[11] C.H. Wu, C.Y. Kuo, C.F. Lin, S.L. Lo, Modeling competitive adsorption ofmolybdate, sulfate, selenate, and selenite using a Freundlich-type multi-component isotherm, Chemosphere 47 (3) (2002) 283–292.

[12] W. Fritz, E.U. Schluender, Simultaneous adsorption equilibria of organicsolutes in dilute aqueous solutions on activated carbon, Chem. Eng. Sci. 29(5) (1974) 1279–1282.

[13] D. Mohan, C.U. Pittman, P.H. Steele, Single, binary and multi-componentadsorption of copper and cadmium from aqueous solutions on Kraft lignin—abiosorbent, J. Colloid Interface Sci. 297 (2) (2006) 489–504.

[14] R.R. Dash, C. Balomajumder, A. Kumar, Removal of cyanide from water andwastewater using granular activated carbon, Chem. Eng. J. 146 (1) (2009) 408–413.

[15] APHA, Standard Methods for the Examination of Water and Wastewater, 20thed, American Public Health Association, Washington, DC, 2001.

[16] B. Ozkaya, Adsorption and desorption of phenol on activated carbon and acomparison of isotherm models, J. Hazard. Mater. 129 (1–3) (2006) 158–163.

[17] R.T. Yang, Gas Separation by Adsorption Processes, Butterworths, Boston, MA,1987.

[18] J.C. Bellot, J.S. Condoret, Modeling of liquid chromatography equilibrium,Process Biochem. 28 (6) (1993) 365–376.

[19] A. Leitao, R. Serrao, Adsorption of phenolic compounds from water onactivated carbon: prediction of multicomponent equilibrium isotherms usingsingle component data, Adsorption 11 (2) (2005) 167–179.

[20] I. Tsibranska, E. Hristova, Comparison of different kinetic models foradsorption of heavy metals onto activated carbon from apricot stones, Bulg.Chem. Commun. 43 (3) (2011) 370–377.

[21] S. Lagergren, About the theory of so called adsorption of soluble substances,Kungl. VetenskapsakadHandl. 24 (4) (1898) 1–39.

[22] W.J. Weber, S.J.C. Morris, Proceedings of International Conference on WaterPollution Symposium, vol. 2, Pergamon, Oxford, 1962, pp. 231–266.

[23] Y.S. Ho, G. McKay, Pseudo-second order model for sorption processes, ProcessBiochem. 34 (5) (1999) 451–465.

[24] V. Srihari, A. Das, The kinetic and thermodynamic studies of phenol-sorptiononto three agro-based carbons, Desalination 225 (1–3) (2008) 220–234.

[25] K.V. Kumar, K. Porkodi, F. Rocha, Isotherms and thermodynamics by linear andnon-linear regression analysis for the sorption of methylene blue ontoactivated carbon: comparison of various error functions, J. Hazard. Mater.151 (2–3) (2008) 794–804.

[26] A. Behnamfard, M.M. Salarirad, Equilibrium and kinetic studies on free cyanideadsorption from aqueous solution by activated carbon, J. Hazard. Mater. 170(1) (2009) 127–133.

[27] M. Ahmaruzzaman, D.K. Sharma, Adsorption of phenols from wastewater, J.Colloid Interface Sci. 287 (1) (2005) 14–24.

[28] M. Kilic, E.A. Varol, A.E. Putun, Adsorptive removal of phenol from aqueoussolutions on activated carbon prepared from tobacco residues: equilibrium,kinetics and thermodynamics, J. Hazard. Mater. 189 (1–2) (2011) 397–403.

[29] R. Guo, C.L. Chakrabarti, K.S. Subramanian, X. Ma, Y. Lu, J. Cheng, W.F.Pickering, Sorption of low levels of cyanide by granular activated carbon,Water Environ. Res. 65 (5) (1993) 640–644.

[30] T. Depci, Comparison of activated carbon and iron impregnated activatedcarbon derived from Golbasi lignite to remove cyanide from water, Chem. Eng.J. 181–182 (2012) 467–478.

[31] S. Saxena, M. Prasad, S.S. Amritphale, N. Chandra, Adsorption of cyanide fromaqueous solutions at pyrophyllite surface, Sep. Purif. Technol. 24 (1–2) (2001)263–270.

[32] G. Moussavi, R. Khosravi, Removal of cyanide from wastewater by adsorptiononto pistachio hull wastes: parametric experiments, kinetics and equilibriumanalysis, J. Hazard. Mater. 183 (1–3) (2010) 724–730.

[33] M.M. Swamy, Studies on the treatment of phenolic wastewaters usingadsorption and immobilized whole cells. Ph.D. thesis, University of Roorkee,India, 1998.

[34] A.H. Sulaymon, A.A. Mohammed, T.J. Al-Musawi, Competitive biosorption oflead, cadmium, copper, and arsenic ions using algae, Environ. Sci. Pollut. Res.,doi:10.1007/s11356-012-1208-2.

[35] K. Saltali, A. Sari, M. Aydin, Removal of ammonium ion from aqueous solutionby natural Turkish (Yıldızeli) zeolite for environmental quality, J. Hazard.Mater. 141 (3) (2007) 258–263.

[36] P. Canizares, M. Carmona, O. Baraza, A. Delgado, M.A. Rodrigo, Adsorptionequilibrium of phenol onto chemically modified activated carbon F400, J.Hazard. Mater. 131 (1–3) (2006) 243–248.

[37] A. Seidel, E. Tzscheutschler, K.H. Radeke, D. Gelbin, Adsorption equilibria ofaqueous phenol and indol solutions on activated carbons, Chem. Eng. Sci. 40(2) (1985) 215–222.

[38] A.A.M. Daifullah, B.S. Girgis, Removal of some substituted phenols by activatedcarbons obtained from agricultural waste, Water Res. 32 (4) (1998) 1169–1177.

[39] A.T.M. Din, B.H. Hameed, A.L. Ahmad, Batch adsorption of phenol ontophysiochemical-activated coconut shell, J. Hazard. Mater. 161 (2–3) (2009)1522–1529.

[40] S. Kumar, M. Zafar, J.K. Prajapati, S. Kumar, S. Kannepalli, Modeling studies onsimultaneous adsorption of phenol and resorcinol onto granular activatedcarbon from simulated aqueous solution, J. Hazard. Mater. 185 (1) (2011) 287–294.

[41] A.E. Vasu, Removal of phenol and o-cresol by adsorption onto activated carbon,E-J. Chemistry 5 (2) (2008) 224–232. <http://www.e-journals.net>.

[42] H.I. Maarof, B.H. Hameed, A.L. Ahmad, Adsorption isotherms for phenol ontoactivated carbon, AJChE 4 (1) (2004) 70–76.

[43] L. Giraldo, J.C.M. Pirajan, Adsorption studies of cyanide onto activated carbonand c-alumina impregnated with cooper ions, Nat. Sci. 2 (10) (2010) 1066–1072.

[44] R.E. Treybal, Mass-Transfer Operations, third ed., McGraw-Hill, 1981.[45] K.R. Hall, L.C. Eagleton, A. Acrivos, T. Vermeulen, Pore and solid diffusion

kinetics in fixed bed adsorption under constant pattern conditions, Ind. Eng.Chem. Fund. 5 (2) (1966) 212–223.

[46] B.H. Hameed, I.A.W. Tan, A.L. Ahmad, Adsorption isotherm, kinetic modelingand mechanism of 2,4,6-trichlorophenol on coconut husk-based activatedcarbon, Chem. Eng. J. 144 (2) (2008) 235–244.

[47] M.F. Carvalo, A.F. Duque, I.C. Goncalves, P.M.L. Castro, Adsorption offluorobenzene onto granular activated carbon: isotherm and bioavailabilitystudies, Bioresour. Technol. 98 (18) (2007) 3424–3430.

[48] Z. Aksu, H. Gulen, Binary biosorption of iron (III) and iron (III)-cyanide complexions on Rhizopusarrhizus: modeling of synergistic interaction, Process Biochem.38 (2) (2002) 161–173.

[49] D. Mohan, K.P. Singh, Single- and multi-component adsorption of cadmiumand zinc using activated carbon derived from bagasse – an agricultural waste,Water Res. 36 (9) (2002) 2304–2318.

[50] W. Zhang, J. Chen, B. Pan, Q. Chen, M. He, Q. Zhang, F. Wang, B. Zang,Synergistic effect on phenol/ aniline mixture adsorption on nonpolar resinadsorbents from aqueous solution, React. Funct. Polym. 66 (4) (2006) 395–401.

664 B. Agarwal et al. / Chemical Engineering Journal 228 (2013) 655–664

[51] S.J. Allen, G. Mckay, K.Y.H. Khader, Intraparticle diffusion of basic dye duringadsorption on to Sphagnum peat, Environ. Pollut. 56 (1) (1989) 39–50.

[52] K. Mohanty, M. Jha, B.C. Meikap, M.N. Biswas, Removal of chromium (VI)from dilute aqueous solutions by activated carbon developed fromTerminalia arjuna nuts activated with zinc chloride, Chem. Eng. Sci. 60(11) (2005) 3049–3059.

[53] W.B. McCloskey, D.E. Bayer, Thermodynamics of fluridone adsorption anddesorption on three California soils, Soil Sci. Soc. Am. J. 51 (3) (1987) 605–612.

[54] J.P. Silva, S. Sousa, J. Rodrigues, H. Antunes, J.J. Porter, I. Gonçalves, S.F. Dias,Adsorption of acid orange 7 dye in aqueous solutions by spent brewery grains,Sep. Purif. Technol. 40 (3) (2004) 309–315.

[55] C. Raymond, Chemistry: Thermodynamics, McGraw-Hill, Boston, 1998.