ARTICLE IN PRESS

0043-1354/$ - se

doi:10.1016/j.w

�Correspondfax: +1201 216

E-mail addr1New Jersey

P.O. Box 409, T

Water Research 39 (2005) 2327–2337

www.elsevier.com/locate/watres

Adsorption of As(V) and As(III) by nanocrystallinetitanium dioxide

Maria E. Pena, George P. Korfiatis, Manish Patel1, Lee Lippincott1,Xiaoguang Meng�

Center for Environmental Systems, Stevens Institute of Technology, Hoboken, NJ 07030, USA

Received 15 May 2004; received in revised form 11 March 2005

Available online 17 May 2005

Abstract

This study evaluated the effectiveness of nanocrystalline titanium dioxide (TiO2) in removing arsenate [As(V)] and

arsenite [As(III)] and in photocatalytic oxidation of As(III). Batch adsorption and oxidation experiments were

conducted with TiO2 suspensions prepared in a 0.04M NaCl solution and in a challenge water containing the

competing anions phosphate, silicate, and carbonate. The removal of As(V) and As(III) reached equilibrium within 4 h

and the adsorption kinetics were described by a pseudo-second-order equation. The TiO2 was effective for As(V)

removal at pHo8 and showed a maximum removal for As(III) at pH of about 7.5 in the challenge water. The

adsorption capacity of the TiO2 for As(V) and As(III) was much higher than fumed TiO2 (Degussa P25) and granular

ferric oxide. More than 0.5mmol/g of As(V) and As(III) was adsorbed by the TiO2 at an equilibrium arsenic

concentration of 0.6mM. The presence of the competing anions had a moderate effect on the adsorption capacities of

the TiO2 for As(III) and As(V) in a neutral pH range. In the presence of sunlight and dissolved oxygen, As(III) (26.7 mMor 2mg/L) was completely converted to As(V) in a 0.2 g/L TiO2 suspension through photocatalytic oxidation within

25min. The nanocrystalline TiO2 is an effective adsorbent for As(V) and As(III) and an efficient photocatalyst.

r 2005 Elsevier Ltd. All rights reserved.

Keywords: Adsorption; Challenge water; TiO2; Photocatalytic oxidation; Arsenic speciation

1. Introduction

Arsenic is one of the most toxic contaminants found

in the environment and has been recognized as a toxic

element for centuries. Arsenic contamination has

become a worldwide epidemic, especially in developing

countries where a significant percentage of the popula-

e front matter r 2005 Elsevier Ltd. All rights reserve

atres.2005.04.006

ing author. Tel.: +1201 216 8014;

8303.

ess: xmeng@stevens-tech.edu (X. Meng).

Department of Environmental Protection,

renton, NJ 08625, USA.

tion depends on groundwater for drinking. Elevated

concentrations of As in groundwater are found in many

countries such as India, Bangladesh, Vietnam and Chile

(Berg et al., 2001). Chronic exposure to arsenic-

contaminated drinking water can result in serious health

problems, such as skin lesions and cancers (Davis et al.,

1996). To minimize the health impact of arsenic, the

United States Environmental Protection Agency

adopted a new maximum contaminant level of 10 mg/Lin drinking water on January 22, 2001.

In natural waters, inorganic arsenic occurs primarily

in two oxidation states, As(V) and As(III). At circum-

neutral pH levels, the predominant As(V) species are the

d.

ARTICLE IN PRESSM.E. Pena et al. / Water Research 39 (2005) 2327–23372328

monovalent (H2AsO4�) and divalent (HAsO4

2�), while

the predominant arsenite species is neutral H3AsO3(Cullen and Reimer, 1989; Meng et al., 2000). As(V)

generally exhibits a low mobility in aquifer and sediment

systems due to its retention on mineral surfaces

controlled primarily by adsorption reactions with metal

hydroxide (Picheler et al., 1999). As(III) is more mobile

and more toxic (25–60 times) than As(V). This elevated

toxicity is due to its preferential reaction with sulfhydryl

groups in mammalian enzymes (Korte and Quintus,

1991).

Several treatment techniques are available for remov-

ing As in water, such as: coagulation/precipitation, ion

exchange, lime softening, reverse osmosis and electro-

dialysis. Coagulation/precipitation with ferric and alu-

minum salts is one of the conventional methods for

arsenic removal from aqueous systems (Gregor, 2001;

McNeill and Edwards, 1997; Gulledge and O’connor,

1973). However, this technique is usually encumbered by

problems associated with the treatment and disposal of

the resulting waste sludge. To overcome this problem,

adsorption processes using granular media have been

explored for their ability to remove arsenic from water.

Numerous laboratory and field filtration tests have been

conducted to assess the removal capacity of adsorbents

such as iron oxide-coated sand, granular ferric hydro-

xide, sulfur-modified iron and activated alumina (Hatch,

2002; Driehaus et al., 1998; Vaishya and Gupta, 2003).

Most technologies require the pre-oxidation of As(III)

to As(V) to enhance removal of As(III), since As(V)

adsorbs more strongly onto the solid phase than As(III).

Table 1

Photocatalytic reactions in TiO2 and As(III) system

Formation of radicals

1TiO2�!

hvhþ þ e�

2 hþ þ e� ! heat

3 e� þO2 ! Od�2

4 Od�2 þHþ ! HOd2

5 2HOd2 ! H2O2 þO26 HOd2 þOd�

2 ! HO�2 þO2

7 HO�2 þ hþ

! HOd28 hþþ2H2Oad ! OHdadþH

þ

9 hþþOH�ad ! OHdad

Possible As(III) oxidation reactions

10 AsðIIIÞ þ hþ ! AsðIVÞ

11 As(III)+OHd-As(IV)

12 As(III)+O2d�-As(IV)+H2O2

13 As(III)+H2O2-As(IV)+2OH�

14 As(IV)+O2-As(V)+O2d�

15 As(IV)+OHd-As(V)

16 As(IV)+O2d� -As(V)

17 As(IV)+O2+H+-As(V)+HO

As(III) oxidation is carried out by oxidants such as

chlorine compounds, ozone, H2O2, permanganate or

manganese oxide and Fenton’s reagent (Pettine et al.,

1999; Kim and Nriagu, 2000; Pettine and Millero, 2000;

Hug and Leupin, 2003). Photochemical oxidation of

As(III) by iron complexes such as Fe(III)OH2+,

Fe(III)Cl2+ and solid iron phases was also investigated

(Hug et al., 2001; Emett and Khoe, 2001; Kocar and

Inskeep, 2003). Photocatalytic oxidation with TiO2, as a

semiconductor, is effective in destroying a wide range of

contaminants in gaseous and aqueous phases. Experi-

mental results have demonstrated that As(III) can be

oxidized to As(V) in UV-illuminated TiO2 (Degussa

P25) suspension (Yang et al., 1999; Lee and Choi, 2002;

Bissen et al., 2001).

The photocatalytic oxidation involves absorption of

light with wavelength shorter than 387.5 nm by TiO2,

which results in the excitation of an electron from the

valence band (vb) to the conduction band (cb). This

excitation creates a positively charged hole, h+, in the

valence band (Table 1, reaction 1). An electron–hole

pair forms at the surface of TiO2, which may react with

adsorbed species such as oxygen to generate free radicals

such as Od�2 ;HOd2 ;OH

d (Table 1). Although the vastmajority of previous TiO2 photocatalytic oxidation

studies have focused on organic compounds, inorganic

species have also been investigated (Gruebel et al., 1995).

For the oxidation reactions to take place, it is necessary

that the valence band or conduction band is more

oxidizing than the oxidation potential of the species in

question.

Low et al. (1991)

Low et al. (1991)

Wong and Chu (2003)

Low et al. (1991)

Kocar and Inskeep (2003)

Low et al. (1991)

Low et al. (1991)

Chen and Ray (1998)

Chen and Ray (1998)

Lee and Choi (2002)

Hug and Leupin (2003)

Hug et al. (2001)

Lee and Choi (2002)

Klaning et al. (1989)

Lee and Choi (2002)

Lee and Choi (2002)

2d Emett and Khoe (2001)

ARTICLE IN PRESSM.E. Pena et al. / Water Research 39 (2005) 2327–2337 2329

As(III) may be oxidized to As(IV) by h+, hydroxyl

radical, superoxide ion, and hydrogen peroxide in a

TiO2/UV system (reactions 10–13). As(IV) is expected to

be oxidized rapidly to As(V) through reactions 14–17.

There is dispute over which reaction controls the As(III)

oxidation process. Lee and Choi (2002) and Ryu and

Choi (2004) showed that superoxide is the dominant

oxidant for As(III) oxidation, which was supported by

Ferguson et al. (2005). On other hand, Dutta et al.

(2005) demonstrated that hydroxyl radical was mainly

responsible for As(III) oxidation.

Based on the literature results, TiO2 is a promising

material for treatment of arsenic, especially As(III).

However, the TiO2 product (Degussa P25) used in most

of the research has very low adsorptive capacity for

arsenic. In the present study, the effectiveness of a

nanocrystalline TiO2 for treatment of arsenic was

examined. The main objectives were to: (a) determine

its adsorption capacity for arsenic; (b) establish the

kinetics of removal of As(V) and As(III); (c) investigate

the photocatalytic oxidation of As(III) to As(V); and (d)

evaluate the effects of common anions in natural water

on As(V) and As(III) adsorption.

2. Experimental methods

2.1. Materials and chemicals

All solutions were prepared using Fisher Scientific

ACS grade chemicals and deionized water (DI). Stock

solutions of 1000mg/L of arsenic were prepared using

sodium salt heptahydrate (Na2HAsO4 � 7H2O) and

sodium arsenite (Na3AsO3). Challenge water containing:

Mg2+ ¼ 12 ppm, SO42�

¼ 50 ppm, NO3�¼ 6 ppm, F� ¼

1ppm, SiO2 ¼ 20ppm, PO43�

¼ 40ppb, Ca2+ ¼ 40ppm,

CO32�

¼ 179ppm, and total chlorine ¼ 0.25–0.75ppm

(chlorine was added only to As(V)-contaminated water)

was prepared daily before the experiments were run

(USEPA/600/R-01/021).

The nanocrystalline TiO2 was produced by hydrolysis

of titanium sulfate solution (Meng et al., 2003). X-ray

powder diffraction analysis indicated that the oxide had

a primary crystallite size of about 6 nm. The material in

powder form costs less than commercial granular ferric

hydroxide (GFH) and granular ferric oxide (GFO). The

nanocrystalline TiO2 slurry was washed with DI water

to remove sulfate ions which could interfere with As

adsorption, and was dried at 105 1C. This metal oxide

had a specific surface area of 330m2/g and a total pore

volume of 0.42 cm3/g. The point of zero charge pH

(pHpzc) of the TiO2 was 5.8 which was determined in a

suspension containing 0.01 g of TiO2 in 0.04M NaCl

using ZetaSizer 3000 (Malvern Instrument). Scanning

electron microscopic images indicated that the nano-

crystalline TiO2 agglomerated to form particles with

sizes of 0.5–2.0 mm.

2.2. Adsorption kinetics

The rate of arsenic uptake is an important factor for

arsenic removal, and is influenced by the surface and

pore properties of the adsorbent. Batch experiments

were conducted to determine the reaction time required

to reach adsorption equilibrium. All suspensions were

prepared in a 0.04M NaCl solution in 1-L glass beakers.

Aliquots of As(V) and As(III) stock solutions were

added to make 2.0mg/L (i.e. 26.7 mmol/L) of As(V) orAs(III) concentration. After the solution pH was

adjusted to 7.0 by adding hydrochloric acid and sodium

hydroxide at room temperature (21–25 1C), TiO2 was

added to attain a 0.2 g/L suspension. The suspension

was mixed with a magnetic stirrer, and the pH was

maintained at 7.070.1 throughout the experiment byaddition of the acid and base solutions. Approximately

25ml aliquots were taken from the suspension at the

following intervals: 0.1, 0.15, 0.25, 0.3, 0.6, 0.9, 2, 4, 6,

10, 20, 40 and 72 h of reaction. The samples were filtered

through a 0.45mm membrane filter. Total As in the

filtered solution was determined using a Furnace Atomic

Absorption Spectrometer (FAAS) (Varian Spectra AA-

400). Speciation of As(III) and As(V) in the solution

samples was performed by passing a solution sample

through an arsenic speciation cartridge (Meng et al.,

2001). The cartridge removed all of the As(V), but did

not remove As(III). The soluble As(V) concentration

was calculated from the difference between the total

soluble arsenic and the soluble As(III) concentrations.

For quality control purposes, the batch experiments

were performed at least in duplicate.

In order to evaluate the effect of dissolved oxygen and

light irradiation on the photocatalytic oxidation of

As(III) in TiO2 suspensions, As(III) removal tests were

conducted in five experimental systems: (1) air and light

system (air–light), where the suspension was open to the

air and under room light (natural and florescent light);

(2) air–dark system, where the reactor was wrapped with

aluminum foil and oxygen in the air was dissolved in the

suspension; (3) N2–dark system, which was obtained by

purging the suspension with high purity nitrogen gas in

the dark; (4) air–sunlight system, where the suspension

in a 1-L glass beaker was mixed directly under the

sunlight outside of the laboratory (this experiment was

performed between 9 a.m. and 5 p.m. on a sunny

summer day (July 8, 2003)), and (5) N2–light system.

2.3. Effect of pH on As adsorption

A suspension containing 0.04M NaCl, 1.0mg/L of

As(V) or As(III) and 0.2 g/L of TiO2 was prepared in a

1-L beaker. Then, 50ml of the uniform suspension was

ARTICLE IN PRESSM.E. Pena et al. / Water Research 39 (2005) 2327–23372330

transferred into 10 50-ml polyethylene terephthalate

(PETE) tubes. The pH values of the suspension in each

tube were adjusted to between 3 and 13 using dilute

NaOH and HCl solutions. The tubes were placed in a

tumbler and mixed for 22 h. After mixing, the final pH

of the suspension was measured as the equilibrium pH.

The solids were then separated from the solution by

centrifugation for 35min at 13,000 rpm in a Model IEC/

Micromax centrifuge. The solutions were acidified using

HNO3 and the total soluble arsenic concentrations were

determined by FAAS. Similar batch tests were also

carried out using challenge water to determine the effects

of common anions on arsenic adsorption.

2.4. Adsorption isotherms

Adsorption isotherms were obtained by adding

different amounts of stock solutions of As(V) and

As(III) in a suspension containing 1.0 g/L TiO2.

Separate batch tests were performed using 0.04M NaCl

and challenge water. The pH value was adjusted to

7.070.1 by adding hydrochloric acid and sodium

hydroxide at room temperature. After 22 h of mixing,

suspension samples were withdrawn and centrifuged for

35min. Arsenic concentrations in the supernatant

solution were analyzed by FAAS.

3. Results and discussion

3.1. Adsorption kinetics

Arsenic removal by TiO2 occurred rapidly in all

systems except the air–light system (Fig. 1). The uptake

of As(V) and As(III) in the air–dark and N2–dark

systems reached equilibrium in approximately 4 h.

Adsorption kinetics of As(III) in the sunlight system

was similar to that of As(V). The amount of As(III)

adsorbed in the air–light system increased at a slow pace

and reached the same level as that observed in the As(V)

system after 14 h of mixing. The adsorption of As(III) by

commercial TiO2 (Degussa P25) reached equilibrium in

about 1 h, which was similar to the reported rate of

As(V) removal by P25 (Dutta et al., 2004). P25 was

made of nonporous TiO2 particles and the nanocrystal-

line TiO2 was composed of aggregated particles with

high porosity. Therefore, it was expected that longer

time was needed to reach adsorption equilibrium in

porous adsorbent.

The total arsenic to TiO2 ratio in all the suspensions

was 133mmol-As/g–TiO2. At equilibrium approximately120 mmol/g of arsenic or 90% of the total arsenic was

adsorbed in the As(V), As(III)-air–light and As(III)-

air–sunlight systems. The removal of As(III) in the

N2–dark system was lower than that of As(V), reaching

an adsorption density of 95 mmol/g. It should be noted

that only about 50 mmol/g of As(III) in the air–lightsystem was adsorbed by P25 TiO2. The results indicated

that the nanocrystalline TiO2 used in this study had a

much higher adsorption capacity than P25. This could

be attributed to much higher specific surface area of the

nanocrystalline TiO2 than that of P25 (Table 2). The

lower pHpzc value of the TiO2 than that of P25 indicated

that the titanyl group, Ti-OH, on the TiO2 was more

acidic than the functional groups on P25.

Several kinetic models including pseudo-first-order of

Lagergren, pseudo-second-order, Elovich, parabolic

diffusion and power function were tested for simulation

of the experimental data in Fig. 1 (Ho and McKay,

1998; Sparks, 1989). The experimental results were best

described by pseudo-second-order and Elovich kinetic

equations.

A pseudo-second-order rate expression is defined by

in Eq. (1) (Sag and Aktay, 2002) and the integrated form

is given in Eq. (2).

dqt=dt ¼ Kadðqe � qtÞ2, (1)

qe � qt ¼ 1=ðð1=qeÞ þ ðKad � tÞÞ, (2)

where qe and qt are the amount (mmol g�1) of arsenic

adsorbed at equilibrium and at time t, respectively. Kadis the rate constant of sorption (mmol g�1 h�1).The Elovich kinetic equation is defined by Eq (3) and

the integrated form is expressed in Eq. (4) (Sparks,

1989).

dqt=dt ¼ a expð�bqtÞ, (3)

qt ¼ ð1=bÞ ln ðt þ toÞ � ð1=bÞ ln to, (4)

where a is an experimental fitting constant and 1=b ¼

kad is the reaction rate constant (Liang and Tsai, 1995).

The Elovich equation was previously used to study the

sorption kinetics of arsenic (Elkhabit et al., 1984) and

phosphate (Chien and Clayton, 1980) on soils, and

cadmium adsorption by bone char (Cheung et al., 2001).

The lines in Fig. 1 were calculated using the kinetic

equations. The pseudo-second-order kinetic equation

described the experimental data well for all systems

except the air–light condition. The adsorption kinetic

data in the air–light system was simulated well by the

Elovich equation. The modeling results suggested that

the mechanism controlling the rate of arsenic adsorption

in the air–light system was different from that in the

other systems. The best fit constants were summarized in

Table 3. The goodness of the model was quantified by

the mean weighted square error (MWSE). The small

MWSE values (o0.005) indicated that the model fittedthe data accurately.

The rate constants (Kad) of the pseudo-second-order

equation were 0.118–0.199 and 0.073 mmol g�1 h�1 forAs(III) and As(V) systems, respectively (Table 3). The

rate values indicated that the removal of As(III) was

ARTICLE IN PRESS

Table 2

Physicochemical properties of TiO2

Characteristics P25 TiO2 Nanocrystalline

TiO2

Crystal structure 80% anatase,

20% rutilea100% anatase

pHpzc 6.8a 5.8

Specific surface area 55m2/ga 330m2/g

Average primary

particle

30 nma 6 nm

aDutta et al. (2004).

0

20

40

60

80

100

120

140

0 10 12 14 16 18 20 22Time (hr)

As

adso

rbed

(µm

ol o

f A

s/g-

TiO

2)

As VAs III, air-darkAs III, air-lightAs III, N2-darkAs III, P25, air-lightAs III, air-sunlight

8642

Fig. 1. Adsorption kinetics of As(V) and As(III) on TiO2 in 0.04M NaCl solution. The lines were model calculations. TiO2 ¼ 0.2 g/L,

initial As concentration ¼ 26.7mM, pH ¼ 7.070.1.

M.E. Pena et al. / Water Research 39 (2005) 2327–2337 2331

faster than As(V). Similar results were observed for

adsorption of As(V) and As(III) onto ferrihydrite

surface (Raven et al., 1998). Dutta et al. (2004) reported

that the adsorption rate of As(V) onto TiO2 decreased

with increasing pH and attributed it to increased

electrostatic repulsion between the negatively charged

surface and anionic As(V) species. The surface potential

of TiO2 decreases with increasing pH and becomes

negative when the pH is higher than the pHpzc. Arsenic

acid H3AsO4 deprotonates progressively to form more

negatively charged anions (i.e., H2AsO4–, HAsO4

2�, and

AsO43�) as pH increases. The kinetics experiments

presented in Fig. 1 were carried out at pH 7.0. At this

pH the surface potential was negative and As(III) was

present in neutral H3AsO3 species. The higher adsorp-

tion rate of As(III) than As(V) could be attributed to

lack of electrostatic repulsion between the surface and

the As(III) species.

3.2. Photocatalytic oxidation of As(III)

The results in Fig. 2 showed the formation of soluble

As(V) species in the kinetic experiments. Soluble As(III)

was completely oxidized to As(V) in the air–sunlight

system within 25min. The sun emits UV light that

reaches the earth’s surface as UVA (400–320 nm) and

UVB (320–290 nm). In the air–light system, soluble

As(III) was slowly oxidized to As(V) due to the presence

of very low intensity of UV light in the room light.

Soluble As(III) was completely converted to As(V) in

about 24 h. Lee and Choi (2002) reported complete

oxidation of As(III) in a suspension containing 500mMAs(III) and 1.5 g/L of P25 and under UV light

irradiation within 2 h. Since the As(III) to TiO2 ratios

and the light sources used in their tests were different

from those employed in the present study, it was not

possible to compare the oxidation rates in different

systems. Dutta et al. (2005) observed decreased As(III)

oxidation rate with increasing As(III) to TiO2 ratio.

As(III) was completely oxidized in a 0.1 g P25 TiO2suspension containing 40 mM As(III) within 13min.

When As(III) concentration was increased to 200mMin the same system, only about 50% of the As(III) was

oxidized in 30min.

Approximately 10% of the total soluble arsenic was

As(V) in the N2–light and N2–dark systems. The

formation of small amounts of As(V) in the anoxic

suspensions could be attributed to the presence of minor

amounts of oxygen adsorbed on TiO2, which were not

removed by purging with nitrogen gas. In addition, the

suspensions could have been exposed to light for a short

ARTICLE IN PRESS

Table 3

Freudlich isotherm and sorption kinetics parameters for As(V) and As(III)

Parameters for isotherms Kinetics parameters

Kf 1/n R2 MWSE Kad R2 MWSE

Solution (0.04 M NaCl)

As(V) air–light system 0.500 0.185 0.988 0.00189 0.073a 0.988 0.00018

As(III) air–light system 0.615 0.267 0.989 0.00268 12.11b 0.984 0.00120

As(III) air–dark system 0.767 0.411 0.989 0.00426 0.199a 0.987 0.00598

As(III) N2–dark system 0.629 0.476 0.998 0.00076 0.118a 0.969 0.00097

As(III) air–sunligh system 0.173a 0.986 0.00362

Solution (challenge water)

As(V) air system, visible light 0.615 0.219 0.975 0.00489

As(III) air system, visible light 0.712 0.319 0.975 0.00749

As(III) air system, dark 0.665 0.404 0.989 0.00286

As(III) close system, dark 0.753 0.531 0.996 0.00130

MWSE ¼ meanweighted square error ¼ ð1=MPM

i¼1ðCexpi � Ccalci Þ=C

expi Þ

2.aPseudo-second-order equation.bFor Elovich equation.

0

20

40

60

80

100

0 3 6 9 12 15 18 21 24 27

Time (hr)

% A

s(V

) in

sol

utio

n

As III, air-lightAs III, N2-lightAs III, air-darkAs III, air-sunlightAs III, N2-darkAs III, air-sunlight model

Fig. 2. Oxidation of As(III) to As(V) in a suspension containing 0.2 g of TiO2/L and 0.04M NaCl. Initial As ¼ 26.7mM,pH ¼ 7.070.1.

M.E. Pena et al. / Water Research 39 (2005) 2327–23372332

period of time when samples were taken. The amount of

As(V) produced in the air–dark system was slightly

higher than in the N2 systems since dissolved oxygen was

allowed in the suspensions. Oxidation of small amount

of As(III) was also observed in a P25 TiO2-air–dark

system (Lee and Choi, 2002).

The kinetics of arsenic removal in the As(III) systems

(Fig. 1) could be controlled by adsorption or oxidation

processes. The rate of As(V) and As(III) removal in

N2–dark and air–dark systems should be controlled by

the adsorption process since no redox reactions occurred

for As(V) and only a small fraction of As(III) was

oxidized to As(V). In the air–sunlight system, the rate of

As(III) oxidation was much faster than the adsorption

rate of As(V). Therefore, the removal of arsenic in the

system should be controlled by the As(V) adsorption

process. This assumption was supported by the observa-

tion that arsenic removal rates were similar in the As(V)

and the air–sunlight systems. On the other hand, the

oxidation rate of As(III) to As(V) in the air–light system

was much slower than the kinetics of As(V) and As(III)

adsorption. Therefore, arsenic removal was controlled

by the oxidation process. As soon as As(V) was formed,

it reached new adsorption equilibrium.

ARTICLE IN PRESSM.E. Pena et al. / Water Research 39 (2005) 2327–2337 2333

Arsenic species on the TiO2 surface could be deduced

from soluble arsenic species and partition of arsenic

species between the solid and liquid phases. The results

in Fig. 1 showed that at equilibrium 95 mmol/g-TiO2 or19mM of arsenic was adsorbed and the soluble arsenic

concentration was 7.7 mM (i.e. 577 mg/L) in the N2–darksystem. Therefore, when soluble arsenic concentration

was below the method detection limit of FAAS (i.e.

0.7mg/L), the amount of As(III) in the solid phaseshould be very low. Moreover, the oxidation of As(III)

should take place at the solid–liquid interface region

since the oxidizing radicals were produced at TiO2surface. In the air–sunlight and air–light systems no

solid phase As(III) should exist when soluble As(III) was

not detected.

A color change of TiO2 from white to gray was

observed in the N2–light suspension. The gray color

became more obvious with time due to excess amount of

electrons trapped in the TiO2 particles (Lee and Choi,

2002). This process was reversed with the addition of

dissolved oxygen which consumed the electrons on the

surface of TiO2.

Control experiments were conducted to evaluate

the oxidation of As(III) in air–light and air–sunlight

systems in the absence of TiO2. About 20% of

the As(III) was converted to As(V) in the air–sunlight

system after 8 h of reaction. The oxidation of

As(III) was negligible in the air–light system during

20 h of the experiment. Therefore, the rapid oxidation of

As(III) in the systems was caused by TiO2 photocatalysis

(Fig. 2).

0

20

40

60

80

100

4 6

% A

s re

mov

al

As V

As III, air-lightAs III, N2-dark

As III, air-dark

875

Fig. 3. Removal of arsenic as a function of equilibrium pH in a su

As ¼ 13.4mM, equilibrium time ¼ 22 h.

3.3. Effect of pH on as adsorption

Fig. 3 illustrated the effect of pH on As(V) and As(III)

removal under different conditions at an arsenic to TiO2ratio of 67mmol-As/g–TiO2 and an equilibrium time of22 h. In a pH range between 4 and 13 arsenic adsorption

in As(V) and As(III)-air–light systems was similar due to

nearly complete oxidation of As(III) to As(V) in the

presence of dissolved oxygen and light. When pH was

below 10, more than 95% of the arsenic was removed in

both systems. A steep decrease in sorption occurred

when the pH increased from 10 to 13. Similar As(V)

adsorption behavior by P25 (Lee and Choi, 2002; Dutta

et al., 2004) and by ferric hydroxides (Meng et al., 2000)

was observed. Adsorption of strong acid anions by

metal oxides and hydroxides typically decreases with

increasing pH (Stumm, 1992). The adsorption of arsenic

acid is through the formation of surface complexa-

tion reaction: Ti-OH+H3AsO4 ¼ Ti-HnAsO4(3�n�1)�+

H2O+(2�n)H+, where n ¼ 0, 1, 2. The reaction

indicates that As(V) adsorption increases with decreas-

ing proton concentration or increasing pH. In addition,

the surface potential of metal oxides becomes more

negative when pH is greater than the pHpzc, which does

not favor the adsorption of anions.

As(III) in the N2–dark and air–dark systems had a

maximum sorption at a pH of approximately 9, which is

characteristic of As(III) adsorption (Meng et al., 2000).

Adsorption of weak acids by metal oxides usually

reaches a maximum at pH values similar to pka1 of the

acid (Stumm, 1992). The pka1 of arsenious acid is 9.2.

9 10 11 12 13pH

spension containing 0.2 g of TiO2/L and 0.04M NaCl. Initial

ARTICLE IN PRESS

0

20

40

60

80

100

4 5 6 7 8 9 10 11 12 13pH

% A

s re

mov

al

As VAs III, air-light

As III, air-darkAs III, N2-dark

Fig. 4. Removal of arsenic as a function of equilibrium pH in a suspension containing 0.2 g of TiO2/L in challenge water. Initial

As ¼ 13.4mM, equilibrium time ¼ 22 h.

M.E. Pena et al. / Water Research 39 (2005) 2327–23372334

The effect of common anions in natural water on the

removal of arsenic was assessed with the results shown

in Fig. 4. In challenge water, TiO2 effectively adsorbed

As(V) at pH values ranging from 4 to 8. However, a

considerable decrease in As(V) removal was observed at

pH48.5, compared to As(V) removal in 0.04M NaCl

solution (Fig. 3). The maximum removal of As(III) in

N2–dark and air–dark systems was reduced from about

95% in the NaCl solution to about 65% in the challenge

water. The maximum adsorption pH shifted from about

9 in the NaCl solution to 7.5 in the challenge water,

indicating that the adverse effect of the competing

anions on As(III) adsorption was more dramatic in a

high pH range. It has been reported that the presence of

phosphate, silicate, and bicarbonate anions dramatically

decreased the removal of As(III) by ferric hydroxide

(Meng et al., 2000, 2002). Relatively high concentrations

of anions such as silicate and phosphate could hinder the

adsorption of As(V).

3.4. Adsorption isotherms

The adsorption capacities of TiO2 in the NaCl

solution and in challenge water at pH 7.070.1 arepresented in Figs. 5 and 6, respectively. When the

equilibrium arsenic concentration was lower than about

100 mM, the amount of As(V) adsorbed was higher thanAs(III) removal and arsenic removal in the N2–dark

system was the lowest (Fig. 5). These observations were

consistent with the results in Figs. 1 and 3. However,

arsenic adsorption had a steeper increase with increasing

equilibrium concentration in the As(III) systems than in

the As(V) system. When the equilibrium concentrations

were higher than 400mM, the amount of arsenicadsorbed in the As(V) system became the lowest. It

appears that As(V) adsorption reached saturation at an

adsorption content of about 0.5mmol/g while about

0.8mmol/g of arsenic was adsorbed in the As(III)

air–dark system at an equilibrium arsenic concentration

of 1100 mM. Since much higher As(III) concentrationwas used in the isotherm experiments than in the kinetic

tests (Fig. 2), the oxidation of As(III) was slower in the

high concentration systems. Speciation analysis indi-

cated that only 15% of the soluble As(III) was oxidized

in the air–dark system at an initial As(III) concentration

of 800 mM As(III) after 22 h of mixing.

The adsorption isotherms in the challenge water

showed that the amount of As(V) adsorbed was higher

than As(III) uptake in the air–dark and N2–dark

systems (Fig. 6). The difference in the amount of

As(V) and As(III) adsorbed became smaller with

increasing equilibrium arsenic concentration. Approxi-

mately 0.5mmol/g of arsenic was adsorbed at an

equilibrium arsenic concentration of 500 mM in all

systems at neutral pH. The adsorption capacity of the

nanocrystalline TiO2 was higher than other TiO2products reported in the literature. Dutta et al. (2004)

observed adsorption of about 0.35 and 0.15mmol/g

As(III) by Hombikat UV100 (an anatase product with

surface area of 334m2) and P25, respectively, at pH 9.

The Freundlich isotherm (Eq. (5)) was used to

describe the adsorption data in Figs. 5 and 6.

G ¼ K f ½C1=n, (5)

where G is the quantity of arsenic adsorbed in mmol g�1,C is the equilibrium arsenic concentration (mmol/l), and

ARTICLE IN PRESS

Equilibrium As (µmol/L)0 100 200 300 400 500 600 700

As

adso

rbed

(m

mol

/g)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8As V

As III, air-light

As III, air-dark

As III, N2-dark

As V, GFO

As III, air-dark, GFO

Fig. 6. Arsenic adsorption isotherms in a 1.0 TiO2 and challenge water at equilibrium pH 7.070.1, equilibrium time ¼ 22 h. The lineswere model calculations.

Equilibrium As (µmol/L)0 200 400 600 800 1000 1200

As

adso

rbed

(m

mol

/g)

0.0

0.2

0.4

0.6

0.8

1.0As VAs III, air-lightAs III, air-dark

As III, N2-dark

0 20 40 60 80 1000.0

0.1

0.2

0.3

0.4

Fig. 5. Arsenic adsorption isotherms in a 1.0 TiO2 and 0.04M NaCl suspension at equilibrium pH 7.070.1, equilibrium time ¼ 22 h.The lines were model calculations.

M.E. Pena et al. / Water Research 39 (2005) 2327–2337 2335

Kf and n are constants. The optimal constants in Table 3

were determined by fitting model-calculated values to

the experimental data. As(III) and As(V) adsorption on

TiO2 were found to obey the Freundlich isotherm in

NaCl solution and challenge water (Figs. 5 and 6).

Arsenic removal by a common iron-based adsorbent,

GFO (Apyron Technologies Inc, Georgia) was also

tested to compare with the adsorption capacity of the

TiO2 in Fig. 6. The GFO product was in granular form

with particle size between 0.5 and 10mm. In order to

ARTICLE IN PRESSM.E. Pena et al. / Water Research 39 (2005) 2327–23372336

compare with the TiO2 powder, the adsorbent was

pulverized to pass 60 mesh standard sieve (0.25mm

opening) and used in the adsorption tests. The BET

surface area of the GFO powder was 138m2/g. The

isotherms in Fig. 6 indicated that the TiO2 had a higher

adsorption capacity for As(III) and As(V) than the GFO

powder. At an equilibrium arsenic concentration of

600mmol/L, approximately 0.3mmol/g of arsenic was

adsorbed by the GFO and more than 0.5mmol/g of

arsenic was adsorbed by TiO2.

4. Conclusions

Adsorption of As(V) and As(III) by nanocrystalline

TiO2 reached equilibrium within 4 h and the adsorption

followed pseudo-second-order kinetics. Gradual oxida-

tion of As(III) to As(V) in the air–light system resulted

in a slow removal rate which could be described by the

Elovich equation. The TiO2 caused rapid photocatalytic

oxidation of As(III) to As(V) in the presence of oxygen

in sunlight. The adsorbent effectively removed As(V) in

a pH range of less than 8 and had a maximum

adsorption for As(III) at pH of approximately 7.5 in

the challenge water. The presence of silicate, carbonate,

and phosphate had a moderate effect on the adsorption

of As(V) and As(III) at pH 7. The nanocrystalline TiO2was an effective adsorbent for As(V) and As(III).

References

Berg, M., Tran, H.C., Nguyen, T.C., Pham, H.V., Schertenleib,

R., Giger, W., 2001. Arsenic contamination of groundwater

and drinking water in Vietnam: A human health threat.

Environ. Sci. Technol. 35 (13), 2621–2626.

Bissen, M., Vieillard-Baron, M.M., Schindelin, A.J., Frimmel,

F.H., 2001. TiO2-catalyzed photooxidation of arsenite(III)

to arsenate(V) in aqueous samples. Chemosphere 44,

751–757.

Chen, D.W., Ray, A.K., 1998. Photodegradation kinetics of 4-

nitrophenol in TiO2 suspension. Water Res. 32 (11),

3223–3234.

Cheung, C.W., Porter, J.F., Mckay, G., 2001. Sorption kinetics

analysis for the removal of cadmium ions from effluents

using bone char. Water Res. 35, 605–612.

Chien, S.H., Clayton, W.R., 1980. Application of Elovich

equation to the kinetics of phosphate release and sorption in

soils. Soil Sci. Soc. Am. J. 44, 265–268.

Cullen, W.R., Reimer, K.J., 1989. Arsenic speciation in the

environment. Chem. Rev. 89, 713–764.

Davis, A., Ruby, M.V., Bloom, M., Schoof, R., Freeman, G.,

Bergstrom, P.D., 1996. Mineralogic constraints on the

biovailabililty of arsenic in smelter-impacted soils. Environ.

Sci. Technol. 30, 392–399.

Driehaus, W., Jekel, M., Hildebrandt, U., 1998. Granular ferric

hydroxide—a new adsorbent for the removal of arsenic

from natural water. J. Water SRT–Aqua 47, 30–35.

Dutta, P.K., Ray, A.K., Sharma, V.K., Millero, F.J., 2004.

Adsorption of arsenate and arsenite on titanium dioxide

suspensions. J. Colloid Interface Sci. 278, 270–275.

Dutta, P.K., Pehkonen, S.O., Sharma, V.K., Ray, A.K., 2005.

Photocatalytic oxidation of Arsenic (III): evidence of

hydroxyl radicals. Environ. Sci. Technol. 39 (6), 1827–1834.

Emett, M.T., Khoe, G.H., 2001. Photochemical oxidation of

arsenic by oxygen and iron in acidic solutions. Water Res.

35 (3), 649–656.

Elkhabit, E.A., Bennett, O.L., Wright, R.J., 1984. Kinetics of

arsenate sorption in soils. Soil Sci. Soc. Am. J. 48, 758.

Ferguson, M.A., Hoffmann, M.R., Hering, J.G., 2005. TiO2-

photocatalyzed As(III) oxidation in aqueous suspensions:

reaction kinetics and effects of adsorption. Environ. Sci.

Technol. ASAP Article.

Gregor, J., 2001. Arsenic removal during conventional alumi-

num-base drinking water treatment. Water Res. 35 (7),

1659–1664.

Gruebel, K.A., Davis, J.A., Leckie, J.O., 1995. kinetics of

oxidation of selenite to selenate in the presence of oxygen,

titania, and light. Environ. Sci. Technol. 29, 586–594.

Gulledge, J.H., O’connor, J.T., 1973. Removal of arsenic(V)

from water by adsorption on aluminum and ferric hydro-

xides. J. Am. Water Works Assoc. 65 (8), 548–552.

Hatch, G.L., 2002. Meeting the new arsenic standard with a

new iron-based adsorbent media: POU applications. Water

Conditioning and Purification, pp. 52–56.

Ho, Y.S., McKay, G.A., 1998. Comparison of chemisorption

kinetics models applied to pollutant removal on various

sorbents. Inst. Chem. Eng. 76b, 332–340.

Hug, S.J., Leupin, O., 2003. Iron-catalyzed oxidation of arsenic

(III) by oxygen and by hydrogen peroxide: pH-dependent

formation of oxidants in the Fenton reaction. Environ. Sci.

Technol. 37 (12), 2734–2742.

Hug, S.J., Canonica, L., Wegelin, M., Gechter, D., Von

Gunten, U.V., 2001. Solar oxidation and removal of arsenic

at circumneutral pH in iron containing waters. Environ. Sci.

Technol. 35 (10), 2114–2121.

Kim, M.J., Nriagu, J., 2000. Oxidation of arsenite in ground-

water using ozone and oxygen. Sci. Total Environ. 247,

71–79.

Klaning, U.K., Bielski, B.H.J., Sehested, K., 1989. Arsenic

(IV). A pulse-radiolysis study. Inorg. Chem. 28 (14),

2717–2724.

Kocar, B.D., Inskeep, W.P., 2003. Photochemical oxidation of

As(III) in ferrioxalate solutions. Environ. Sci. Technol. 37,

1581–1588.

Korte, N.E., Quintus, F., 1991. A review of arsenic (III) in

groundwater. Crit. Rev. Environ. Control 21 (1), 1–39.

Lee, H., Choi, W., 2002. Photocatalytic oxidation of arsenite in

TiO2 suspension: kinetics and mechanisms. Environ. Sci.

Technol. 36, 3872–3878.

Liang, T., Tsai, J.Y., 1995. Sorption kinetics of cesium on

natural mordenite. Appl. Radiat. Isot. 46, 7–12.

Low, G.K.C., McEvoy, S.R., Matthews, R., 1991. Formation

of nitrate and ammonium ions in titanium dioxide mediated

photocatalytic degradation of organic compounds contain-

ing nitrogen atoms. Environ. Sci. Technol. 25, 460–467.

McNeill, L.S., Edwards, M., 1997. Predicting As removal

during metal hydroxide precipitation. J. Am. Water Works

Assoc. 89 (1), 75–86.

ARTICLE IN PRESSM.E. Pena et al. / Water Research 39 (2005) 2327–2337 2337

Meng, X.G., Bang, S.B., Korfiatis, G.P., 2000. Effect of silicate,

sulfate and carbonate on arsenic removal by ferric hydro-

xide. Water Res. 34, 1255–1261.

Meng, X.G., Korfiatis, G.P., Chuanyong, J., Christodoulatos,

C., 2001. Redox transformations of arsenic and iron in

water treatment sludge during aging and TCLP extraction.

Environ. Sci. Technol. 35, 3476–3481.

Meng, X.G., Korfiatis, G.P., Bang, S.B., Bang, K.W., 2002.

Combined effects of anions on arsenic removal by iron

hydroxides. Toxicol. Lett. 133, 103–111.

Meng, X.G., Dadachov, M., Korfiatis, G.P., Christodoulatos, C.,

2003. Method of preparing a surface-activated titanium oxide

product and of using the same in water treatment processes.

Patent pending, application number 20030155302.

Pettine, M., Millero, F.J., 2000. Effect of metals on the

oxidation of As(III) with H2O2. Mar. Chem. 70, 223–234.

Pettine, M., Campanella, L., Millero, F.J., 1999. Arsenite

oxidation by H2O2 in aqueous solutions. Geochim. Cosmo-

chim. 63 (18), 2727–2733.

Picheler, T., Veizer, J., Hall, G.E.M., 1999. Natural impute of

arsenic into a coral-reef ecosystem by hydrothermal fluids

and its removal by Fe(III) oxyhydroxides. Environ. Sci.

Technol. 33, 1373–1378.

Raven, K.P., Jain, A., Loeppert, R.H., 1998. Arsenite and

arsenate adsorption on ferrihydrite: kinetics, equilibrium

and adsorption envelopes. Environ. Sci. Technol. 32,

344–349.

Ryu, J., Choi, W., 2004. Effects of TiO2 surface modi-

fications on photocatalytic oxidation of arsenite: the

role of superoxides. Environ. Sci. Technol. 38 (10),

2928–2933.

Sag, Y., Aktay, Y., 2002. Kinetic studies on sorption of Cr(VI)

and Cu(II) ions by chitin, chitosan and rhizopus arrhizus.

Biochem. Eng. J. 12, 143–153.

Sparks, D.L., 1989. Kinetics of Soil Chemical Processes.

Academic Press, New York.

Stumm, W., 1992. Chemistry of the Solid–Water Interface.

Wiley-Interscience, New York, pp. 23–26.

USEPA, 2001. Laboratory study on the oxidation of arsenic III

to arsenic V, EPA/600/R-01/021, March 2001.

Vaishya, R.C., Gupta, S.K., 2003. Arsenic removal from

groundwater by iron impregnated sand. J. Environ. Eng.

129 (1), 89–92.

Wong, C.C., Chu, W., 2003. The direct photolysis and

photocatalytic degradation of alachlor at different TiO2and UV sources. Chemosphere 50 (8), 981–987.

Yang, H., Lin, W.Y., Rajeshwar, K., 1999. Homogeneous and

heterogeneous photocatalytic reactions involving As(III)

and As(V) species in aqueous media. J. Photochem.

Photobiol. 123, 137–143.