Adsorption of naphthalene on zeolite from aqueous solution

(1)

www.elsevier.com/locate/jcis

Adsorption of naphthalene on zeolite from aqueous solution

Chiung-Fen Chang

a

_{, Ching-Yuan Chang}

a,∗

_{, Ken-Hung Chen}

a

_{, Wen-Tien Tsai}

b

_{, Je-Lueng Shie}

a

_,

Yi-Hung Chen

a

a_{Graduate Institute of Environmental Engineering, National Taiwan University, Taipei 106, Taiwan}

b_{Department of Environmental Engineering and Health, Chia Nan University of Pharmacy and Science, Tainan 717, Taiwan} Received 18 January 2004; accepted 14 April 2004

Available online 18 May 2004

Abstract

Polynuclear aromatic hydrocarbons (PAHs), which are environmental hormones and carcinogens, are viewed as the priority pollutants to deal with by many countries. Most PAHs are hydrophobic with high boiling and melting points and high electrochemical stability, but with low water solubility. Compared with other PAH species, naphthalene has less toxicity and is easily found in the environment. Thus, naphthalene is usually adopted as a model compound to examine the environmental and health aspects of PAHs. This study attempted to use an adsorption process to remove naphthalene from a water environment. The adsorption equilibrium of naphthalene on zeolite from water–butanol solution, which is a surfactant-enriched scrubbing liquid, was successfully evaluated by Langmuir, Freundlich, and linear isotherms. Among the tested kinetics models in this study (e.g., pseudo-first-order, pseudo-second-order, and Elovich rate equations), the pseudo-second-order equation successfully predicted the adsorption.

Keywords: Naphthalene; Zeolite; Adsorption isotherms; Adsorption kinetics; Pseudo-first-order process; Pseudo-second-order process; Elovich rate equation

1. Introduction

Polynuclear aromatic hydrocarbons (PAHs), which many countries view as the priority pollutants to deal with, are environmental hormones and carcinogens. PAHs are made up of only carbon and hydrogen, and can be divided into two groups: kata-annellated and peri-condensed[1]. The sat-urated vapor pressures of PAHs at 298 K are lower than 0.1 mm Hg and belong to semivolatile organic compounds (SVOCs). Most PAHs are hydrophobic with high boiling and melting points and possess low water solubility and electro-chemical stability. Therefore, they can exist and be accumu-lated in the environment for long times. The sources of PAHs can be divided into two categories, artificial and natural sources, in which the amount of the former is far greater than that of the later and surpasses the self-purification capacity.

Due to the refractoriness and difficulty of biological degradation of PAHs, it is found that PAHs have accumu-lated in the air, water bodies, soil, and food. Therefore,

* _{Corresponding author. Fax: +886-2-2363-8994.}

E-mail address: [email protected] (C.-Y. Chang).

immediate attention to the effective treatment of PAHs is needed. The methods for treating PAHs mainly include biodegradation[2,3], scrubber absorption[4–6], high-energy electron beam irradiation[7], ozonation, and catalytic com-bustion. In order to avoid a secondary public nuisance, it is suitable to treat PAHs with medium-high concentrations by catalytic combustion and ozonation. The adsorption process is widely applied to organic compound removal in wa-ter/wastewater treatment with high removal efficiency. Al-though activated carbon is the most commonly used sorbent, it still has some limitations in application, such as flamma-bility, difficulty of regenerating adsorbed high-boiling-point organics, etc. In contrast to the activated carbon, zeolite (aluminosilicates renowned for their microporosity, catalytic properties, and extensive applications) can be used in more specific situations with the advantages of the framework’s open structure, rich ion-exchange chemistry, etc. [8], but without the disadvantages of activated carbon. Therefore, ze-olites were used as adsorbents to concentrate organics from streams[9–14].

Naphthalene is a natural constituent of coal tar and com-monly used as a wood preservative, moth repellent, and raw material to produce methylanthranilate, phthalate

(2)

ters, chloronaphthalene, synthetic resins, etc. The sources of naphthalene in the air are mainly from the burning of coal and oil, the use of mothballs, the production of coal tar, etc. Being compared with other PAHs species, naph-thalene, the simplest PAH, has less toxicity and is easily found in the environment. Therefore, the adoption of naph-thalene as the target compound in this study can not only provide useful information for the treatment of naphthalene but also be viewed as the primary method of inquiry for dealing with complicated PAHs. The aim of this study is to investigate the feasibility of using hydrophobic zeolites to remove naphthalene from a water–butanol solution which serves as a surfactant-enriched scrubbing liquid. The adsorp-tion of naphthalene on zeolites from water–butanol soluadsorp-tions is further evaluated by the common isotherms (e.g., Lang-muir, Freundlich, and linear isotherms) and kinetic models (e.g., pseudo-first-order, pseudo-second-order, and Elovich rate equations).

2. Materials and methods

2.1. Adsorbent

Hydrophobic zeolite (type: DAY-Zeolith, Degussa, Ger-many) with a particle size range between 20 and 30 meshes (with sieve opening of 0.841 and 0.595 mm) was used as the adsorbent. The mean particle size dpof DAY zeolite is

0.72 mm. The physical characteristics of DAY zeolite are shown inTable 1. The pretreatment of the adsorbent com-prised several steps. First, the adsorbent was washed with distilled water to remove the crushed fines. Then, it was dried overnight at 383 K in a vacuum oven and then stored in a desiccator. Finally, it was wetted in the water–butanol solu-tion under vacuum, prior to the adsorpsolu-tion experiments. The weights of the adsorbents used in this study ranged between 0.5 to 3 g.

Table 1

Physical characteristics of hydrophobic DAY zeolite

Property Value

Mesh size 20–30

Average particle diameter,adp(mm) 0.72 Specific external surface area,bas(m2/kg) 5.48 Average true particle density,aρs(kg/m3) 2380 Apparent particle density,aρp(kg/m3) 1520

Particle porosity,cεp(–) 0.36

BET specific surface areaa(m2/g) 601

Total pore volumea(cm3/g) 0.237

Average pore diametera(Å) 15.8

Average pore hydraulic radiusd(Å) 3.01 a _{Data source: Ref.}_[14]_.

b _{Assumed as a spherical particle and calculated using as}_{= 6/(ρ}_pdp). c _{Calculated using εp}_{= 1 − (ρ}_p/ρs).

d _{Data from the analysis by the micropore method (MP method) with the} pore size between 2.4 and 17.2 Å.

2.2. Adsorbate

Naphthalene (C10H8), with a molecular weight of 128.16,

was scintillation grade, provided by Merck, and was used as the representative compound of PAHs in this study. The initial concentrations of naphthalene for adsorption experi-ments were 4.5 to 33 g m−3. A certain amount of naphtha-lene was dissolved in 1-butanol (reagent grade supplied by Nacalai Tesque, Inc., Japan) of 37 ml, and subsequently di-luted by distilled water of 3668 ml to prepare solutions at various concentrations, the total volume of which is fixed at 3705 ml.

2.3. Aqueous systems

Due to the low water solubility of naphthalene, the water– butanol solution was used as the aqueous solution. It is noted that butanol is a good solvent used to enhance the adsorp-tion of hydrophobic naphthalene from the hydrophobic envi-ronment, forming the naphthalene-containing solutions. The water–butanol solution was composed of 37 ml butyl alcohol and 3668 ml distilled water.

2.4. Analytical measurements

Prior to the analysis, all the samples were filtrated through a 0.45-µm membrane. Two analyzers, i.e., a total organic carbon (TOC) analyzer (O.I.C. M-700/carbon ana-lyzer Dohrmann DC-80), and UV spectrophotometer (GBC UV–visible Cintra 20 spectrometer) were used to determine the concentrations of butanol and naphthalene, respectively, in this study. The wavelength used in the UV spectropho-tometer for naphthalene was 276.3 nm.

2.5. Adsorption behavior

Samples with various weight ratios of naphthalene to ze-olite were prepared to get various values of final equilibrium concentrations. The experiments were performed at 298 K in the completely stirred tank reactor (CSTR) until the con-centrations of filtrate did not change within a range of±3%. Since it is preferable to use weight concentration units for in-vestigating the removal efficiency in wastewater treatments, the units of g m−3and g kg−1for the liquid and solid phases were used for the organics in this study. The amount of naph-thalene adsorbed at time t and equilibrium were calculated using the equations

(1) qt= (C0− Ct)× V /W,

(2) qe= (C0− Ce)× V /W,

where qeand qtare the adsorbate concentrations in the solid

at equilibrium and time t , respectively; C0, Ct, and Ce are

the initial concentration, the liquid-phase concentration at time t , and the equilibrium concentration of naphthalene, respectively; V is the volume of the aqueous solution and equal to 3.705 L; and W is the mass of zeolite used in the experiments.

(3)

3. Results and discussion

3.1. Adsorption equilibria

In order to clarify the existence of competitive adsorption between naphthalene and butanol on zeolite, the adsorption of butanol on DAY zeolite was first conducted, as illustrated inFig. 1. The result shows that butanol was not adsorbed onto the zeolite, since it did not possess competitive adsorp-tion on zeolite from the water–butanol soluadsorp-tion containing naphthalene. For investigation of the adsorption of naphtha-lene on DAY zeolite in this study, the empirical Freundlich and Langmuir isotherms, which correspond to the hetero-geneous and homohetero-geneous adsorbent surfaces, respectively, were used to correlate the experimental data as follows:

(3) qe= kFCe1/nF, (4) qe= QLKLCe 1+ KLCe .

kF and nF are the Freundlich isotherm constants. QL and KL are the Langmuir isotherm constants, representing the

monolayer adsorption capacity and equilibrium constant, re-spectively. The constants in the models can be obtained by linearizing the above equations as follows:

Fig. 1. Dependence of butanol concentration (Ct) on time using DAY

zeo-lite as the adsorbent in a stirred tank at 800 rpm. The volume of solution is 3.705 L. The mass of zeolite is 3 g. The initial concentration of butanol (C0) is 11 g/m3. (5) ln qe= ln kF+ nFln Ce, (6) 1 qe= 1 QLKL 1 Ce + 1 QL .

The results are shown inTable 2andFig. 2. Both adsorp-tion isotherms can well predict the adsorpadsorp-tion of naphthalene on zeolite with high correlation coefficients (r2). It is noted that the value of nF is very close to 1, suggesting that the

adsorption of naphthalene on zeolite approximately follows a linear isotherm. Furthermore, according to the lower limit of the Langmuir isotherm (i.e., KLCe 1), the Langmuir

isotherm can be reduced to a linear isotherm as follows: (7) qe= KdCe,

where Kdis the equilibrium distribution coefficient

describ-ing the distribution of liquid in the adsorbent.Equation (7)is similar to Henry’s law (qe= KHCe), of which the constant KHdescribes the gas solubility in the liquid. Compared with

the Freundlich and Langmuir isotherms, the linear isotherm was also tested to predict the experimental data, as seen in Table 2andFig. 2. The results showed that good applica-bility was also obtained by the linear isotherm. As a result, Freundlich, Langmuir, and linear isotherms were all suitable to be applied in predicting the adsorption of naphthalene on DAY zeolite from water–butanol solution.

Fig. 2. Isothermal adsorption of naphthalene on DAY zeolite by dynamic experiments in the stirred tank at 800 rpm. (!), (—), (- - -) and (— -): expe-rimental data, simulated by Langmuir, Freundlich, and linear isotherms. Table 2

Isotherm parameters for adsorption of naphthalene onto DAY zeolite in the water–butanol solution

Freundlich isotherm Langmuir isotherm Linear isotherm

kF(g/kg)/((g/m3)1/nF) nF(–) r2 KL(m3/kg) QL(g/kg) r2 Kd(m3/kg) r2

(4)

3.2. Adsorption kinetics

Numerous publications in literature on adsorption kinet-ics are available[14–18]. The mass transfers occur within the boundary layer around the adsorbent and proceed in the liquid-filled pores or along the walls of the pores of adsor-bent, which are called external and internal mass transfers, respectively. The typical kinetic models normally consider both the external and internal mass transfers. Examples of these models are film-pore diffusion, film-surface diffusion, pore diffusion, surface diffusion, and combined pore and sur-face diffusion models. The models involve a complicated mathematical computation to obtain the related diffusion co-efficients of the models. Furthermore, the mass transfers of adsorption often involve many controlling mechanisms, of which the individual contribution may not be clearly recog-nized, at the same time during the course to approach adsorp-tion equilibria. Therefore, for the simplicity and practical use of engineering applications, the global kinetic expressions such as Lagergren pseudo-first-order, pseudo-second-order, and Elovich rate equations, were adopted to describe the ad-sorption kinetics in the study[18]by means of the lumped analysis of kinetics data.

3.2.1. Pseudo-first-order process

For the pseudo-first-order process, the Lagergren equa-tion was expressed as

(8) dqt

dt = ke1(qe− qt).

IntegratingEq. (8)with the conditions (qt= 0 at t = 0; qt= qt at t= t) gives

(9) ln(qe− qt)= ln(qe)− ke1t,

where qt and qehave the same meaning as described earlier

and ke1is the equilibrium rate constant of pseudo-first-order

sorption.

3.2.2. Pseudo-second-order process

The pseudo-second-order process can be presented as fol-lows:

(10) dqt

dt = ke2(qe− qt) 2_.

IntegratingEq. (10) with the conditions (qt = 0 at t = 0; qt= qt at t= t) yields (11) t qt = 1 ke2qe2 + t qe ,

where ke2 is the equilibrium rate constant of the

pseudo-second-order sorption. 3.2.3. Elovich rate equation

The Elovich equation is as follows:

(12) dqt

dt = a exp(−bqt).

IntegratingEq. (12) with the conditions (qt = 0 at t = 0; qt= qt at t= t) and subsequently linearizing the integrated

equation result in (13) qt= 1 bln(ab)+ 1 bln(t+ t0),

where a and b are the parameters of the Elovich rate equa-tion; t0is equal to 1/(ab).

If abt 1,Eq. (13)can further be simplified as

(14) qt= 1 bln(ab)+ 1 bln(t). Table 3

Parameters and determination coefficients (R2) of various kinetic models

Initial concentration C0 Pseudo-first-order equation Pseudo-second-order equation Elovich rate equation

(g/m3) r2a ke1 R2 r2b ke2 R2 r2c a b R2 5 0.9961 0.0244 0.8752 0.9950 0.00595 0.9874 0.9684 0.3048 0.9061 0.9684 15 0.9924 0.0179 0.6030 0.9984 0.00227 0.9896 0.9540 1.1548 0.3402 0.9595 23 0.9675 0.0109 0.9461 0.9990 0.00048 0.9959 0.9868 0.6838 0.1869 0.9895 28 0.9975 0.0103 0.7467 0.9996 0.00045 0.9743 0.9639 1.3394 0.1723 0.9656 33 0.9849 0.0089 0.9263 0.9950 0.00040 0.9792 0.9875 1.2813 0.1215 0.9875

a _{Correlation coefficient for linear regression of}_{Eq. (9)}_. b _{Correlation coefficient for linear regression of}_{Eq. (11)}_. c _{Correlation coefficient for linear regression of}_{Eq. (14)}_. Table 4

Comparison of equilibrium adsorption capacities (qe) Initial concentration C0 (g/m3) Experimental data qe (g/kg) Langmuir isotherm qe (g/kg) Pseudo-first-order equation qe (g/kg) Pseudo-second-order equation qe (g/kg) 5 4.22 4.21 3.87 4.96 15 13.23 17.47 10.93 14.58 23 22.00 20.23 21.78 25.64 28 27.30 21.00 24.86 30.77 33 31.23 29.59 29.18 37.88

(5)

Fig. 3. Pseudo-first-order adsorption kinetics of naphthalene on zeolite at various values of C0. (E), (P), (!), (1), and (e): C0= 5, 15, 23, 28, and 33 g/m3, respectively. (—): simulated by pseudo-first-order kinetic equa-tion.

The results are shown inTables 3–4, andFigs. 3–5. In ad-dition, the coefficient of determination (R2) used to compare the validities of the fitting of the three models is defined as

(15) R2= 1 − (ye− yc)2 (ye− ym)2 ,

where ye, yc, and ymare the experimental and predicted data,

and the average of the experimental values, respectively. Comparing the values of R2 of pseudo-first-order and second-order equations, the latter is better than the former and can be used to predict the adsorption kinetics of naphtha-lene on the zeolite. The values of qepredicted by the

pseudo-second-order equation are also in good agreement with the experimental data, as shown inTable 3. Among all the cal-culated qethe data based on the pseudo-first-order equation

are the closest to the experimental data. The Elovich rate equation is commonly used to describe the sorption behav-ior with a rapid equilibrium rate in the early period, while it slows down the equilibrium at later periods of the sorp-tion process. The constants a and b, in the Elovich rate equations represent the rate of sorption and surface cover-age, respectively[16]. Along with increasing initial concen-trations, the value of b decreases due to the less available surface for naphthalene. Furthermore, the value of a should increase with initial concentrations because of the higher driving force. However, this result was not obtained in this study, although the values of R2of prediction fitted by the Elovich rate equation are rather high. The physical meaning should overwhelm the determination coefficient. Therefore, the Elovich rate equation may not be suitable to describe the kinetics of naphthalene adsorbed onto the zeolite for this rea-son.

Fig. 4. Pseudo-second-order adsorption kinetics of naphthalene on zeolite at various values of C0. (E), (P), (!), (1), and (e): C0= 5, 15, 23, 28, and 33 g/m3, respectively. (—): simulated by pseudo-second-order kinetic equation.

Fig. 5. Elovich adsorption kinetics of naphthalene on zeolite at various val-ues of C0. (E), (P), (!), (1), and (e): C0= 5, 15, 23, 28, and 33 g/m3, respectively. (—): simulated by Elovich kinetic equation.

4. Conclusions

The adsorption behavior of naphthalene on the zeolite from water–butanol solution has been investigated in this study. Langmuir, Freundlich, and linear isotherms can be used to describe the adsorption equilibria of naphthalene on the zeolite. The simple kinetic model of the

(6)

pseudo-second-order equation has been successfully applied to pre-dict the adsorption of naphthalene on the zeolite. Although the pseudo-first-order equation can well predict the equi-librium capacity of naphthalene, the low coefficient of de-termination suggested that this model cannot be adapted to describe the kinetics. Regarding the Elovich rate equation, its value of parameter a, which represents the rate of adsorp-tion, is inconsistent with the initial concentration. Therefore, it is not recommended to describe the adsorption kinetics in this study.

Acknowledgment

The authors thank the Powder Technology Laboratory of the Chemical Engineering Department of National Taiwan University for assistance in powder characterization.

Appendix A. Nomenclature

a Elovich rate equation constant representing rate of sorption, as specified inEq. (12)(g/kg/min) as Specific external surface area (m2/kg)

b Elovich rate equation constant representing surface coverage, as specified inEq. (12)(kg/g)

BET The specific surface area obtained following the Brunauer, Emmett, and Teller theory

Ce Adsorbate concentration in the liquid phase at

equi-librium with qe(g m−3)

C0 Initial adsorbate concentration in the liquid phase

(g m−3)

Ct Adsorbate concentration in the liquid phase at

time t (g m−3)

dp Mean particle size (mm)

kF Freundlich isotherm constant as specified inEq. (3),

(g/kg)/((g/m3)1/nF₎

Kd Equilibrium distribution coefficient as specified in

Eq. (7)(m3/kg)

KH Equilibrium constant of Henry’s law (m3/mol) KL Langmuir isotherm constant as specified inEq. (4)

(m3/kg)

ke1 Equilibrium rate constant of pseudo-first-order

equation as specified inEq. (8)(1/min)

ke2 Equilibrium rate constant of pseudo-second-order

equation as specified inEq. (10)(kg/g/min)

MP Micropore

nF Freundlich isotherm constant as specified inEq. (3)

(–)

PAHs Polynuclear aromatic hydrocarbons

qe Adsorbate concentration in solid phase at

equilib-rium with Ce(g/kg)

qt Adsorbate concentration in solid phase at time t

(g/kg)

QL The monolayer adsorption capacity as specified in

Eq. (4)(g/kg)

R2 Determination coefficient defined byEq. (15)

r2 Correlation coefficient TOC Total organic carbon

t Adsorption time or elapsed time (min)

t0 1/(ab) (1/min)

V Volume of aqueous solution, equal to 3.705 dm3 W Mass of zeolite used in experiments (g)

yc Experimental data as specified inEq. (15) ye Predicted values as specified inEq. (15)

ym Average of the experimental values as specified in

Eq. (15)

εp Adsorbent porosity (–)

ρp Apparent particle density (kg/m3) ρs Average true particle density (kg/m3)

References

[1] M. Zander, in: A. Bjrseth (Ed.), Handbook of Polycyclic Aromatic Hydrocarbons, Dekker, New York, 1983, pp. 1–26.

[2] T. Ebihara, P.L. Bishop, J. Environ. Eng. 128 (6) (2002) 505–513. [3] H.N. Holma, K. Nieman, D.L. Sorensen, C.D. Miller, M.C. Martin,

T. Borch, W.R. McKinny, R.C. Sims, Environ. Sci. Technol. 36 (6) (2002) 1276–1280.

[4] K. Friedman, Removal of Naphthalene from Coke Oven Gas, available at:http://www.lehigh.edu/~Kaf3/cchemdata/naphth.html, 1994. [5] M. Platts, Thyssenkrupp Encoke USA, Qureshi, 2002. [6] H. Huang, W. Lee, J. Environ. Eng. 128 (1) (2002) 60–67.

[7] W. Cooper, M.G. Nickelsen, R.V. Green, S.P. Mezyk, Radiation Phys. Chem. 65 (4–5) (2002) 571–577.

[8] S.L. Suib, Science 302 (5649) (2003) 1335–1336. [9] D. Drohan, Metal Finishing 91 (11) (1993) 60–62. [10] S.B. Riffat, Int. J. Energy Res. 18 (7) (1994) 657–664.

[11] N.A. Booker, E.L. Cooney, A.J. Priestly, Water Sci. Technol. 34 (9) (1996) 17–24.

[12] W.T. Tsai, C.Y. Chang, H.C. Lee, J. Environ. Sci. Health A 31 (1996) 995.

[13] I.V. Babich, A.D. Langeveld, W. Zhu, W.J.W. Bakker, J.A. Moulijn, Ind. Eng. Chem. Res. 40 (1) (2001) 257–363.

[14] C.Y. Chang, W.T. Tsai, C.H. Ing, C.F. Chang, J. Colloid Interface Sci. 260 (2003) 273–279.

[15] Y.S. Ho, G. McKay, Chem. Eng. J. 70 (1998) 115–124. [16] H. Teng, C. Hsieh, Ind. Eng. Chem. Res. 38 (1999) 292–297. [17] C.F. Chang, C.Y. Chang, W. Höll, Ind. Eng. Chem. Res. 42 (2003)

6904–6910.

[18] C.W. Cheung, J.F. Porter, G. McKay, Water Res. 35 (3) (2001) 605– 612.