w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 3 7 2 7 – 3 7 3 4

Avai lab le a t www.sc iencedi rec t .com

journa l homepage : www.e lsev ie r . com/ loca te /wat res

Synthesis of mesoporous magnetic g-Fe2O3 and itsapplication to Cr(VI) removal from contaminated water

Peng Wang, Irene M.C. Lo*

Department of Civil and Environmental Engineering, The Hong Kong University of Science and Technology, Hong Kong, China

a r t i c l e i n f o

Article history:

Received 13 February 2009

Received in revised form

24 May 2009

Accepted 27 May 2009

Published online 6 June 2009

Keywords:

Adsorption

Chromium

Magnetic separation

Mesoporous iron-oxide

* Corresponding author. Tel.: þ86 852 2358 7E-mail address: cemclo@ust.hk (I.M.C. Lo

0043-1354/$ – see front matter ª 2009 Elsevidoi:10.1016/j.watres.2009.05.041

a b s t r a c t

In this study, mesoporous magnetic iron-oxide (g-Fe2O3) was synthesized as an adsorbent

for Cr(VI) removal. For material synthesis, mesoporous silica (KIT-6) was used as a hard

template and to drive iron precursor into KIT-6, a ‘greener’, affinity based impregnation

method was employed, which involved using a nonpolar solvent (xylene) and led to

recycling of the solvent. The results of Cr(VI) removal experiments showed that the

synthesized mesoporous g-Fe2O3 has a Cr(VI) adsorption capacity comparable with 10 nm

nonporous g-Fe2O3 but simultaneously has a much faster separation than 10 nm nonpo-

rous g-Fe2O3 in the presence of an external magnetic field under the same experimental

conditions. Cr(VI) adsorption capacity onto the mesoporous g-Fe2O3 increased with

decreasing solution pH and could be readily regenerated. Therefore, mesoporous g-Fe2O3

presents a reusable adsorbent for a fast, convenient, and highly efficient removal of Cr(VI)

from contaminated water.

ª 2009 Elsevier Ltd. All rights reserved.

1. Introduction wastes, agricultural wastes and other polysaccharide mate-

Hexavalent chromium, Cr(VI), is a hard oxidant and a group A

carcinogen (Katz and Salem, 1993) and the industrial sources

of Cr(VI) mainly include: alloys and steel manufacturing,

metal finishing, electroplating, leather tanning, and pigments

synthesis and dyeing (Shevchenko et al., 2008).

During the past decade, considerable research attention has

been diverted to selective removal of Cr(VI) from contaminated

water via adsorption. In principle, adsorption not only can

remove contaminants but also can recover and recycle them

back to industrial processes (Singh and Tiwari, 1997). Various

types of adsorbents have been studied for their effectiveness in

this regard, including activated carbons (Aggarwal et al., 1999),

polymeric adsorbents (Dabrowski et al., 2004), metal oxides (Ai

et al., 2008), and even certain types of biosorbents (Volesky,

2007). It is worth mentioning that biosorption using bio-

sorbents, including bacteria, fungi, yeast, algae, industrial

157; fax: þ86 852 2358 153).er Ltd. All rights reserved

rials, etc. has been a research hotspot in the field of heavy metal

removal lately (Vijayaraghavan and Yun, 2008; Volesky, 2007).

However, given the fact that almost all of the biosorbents

require drying and chemical pretreatment, at least with acid or

alkali pretreatment, for their effective performance, and also

the concerns of safe disposal of spent biosorbents, large-scale

application is still not foreseeable in the near future.

Iron-oxide based materials, on the other hand, have

distinguished themselves due to their effectiveness and

selectiveness in Cr(VI) removal, quantitative recovery of

Cr(VI), and their innocuousness and chemical stability over

a wide pH range. Goethite (a-FeOOH) and hydrous ferric oxide

(HFO) have been among the most studied iron-oxide based

materials for Cr(VI) adsorption because of their natural

abundance (Fendorf et al., 1997).

In the past few years, another special property of some

iron-oxide based materials (Fe3O4, g-Fe2O3), magnetism, came

4.

.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 3 7 2 7 – 3 7 3 43728

to be realized and utilized in the context of environmental

remediation. Utilizing the magnetic properties of these

adsorbents, magnetic separation has been combined with

adsorption for heavy metal removal from contaminated water

at laboratory scales (Yavuz et al., 2006; Mayo et al., 2007; Hu

et al., 2004, 2005a, b). In industry, magnetic separation is

especially desirable because it overcomes many of the issues

present in filtration, centrifugation or gravitational separa-

tion, and requires much less energy to achieve a given level of

separation.

It is believed that the contaminant adsorption capacity of

an adsorbent is largely determined by the surface area avail-

able for adsorption, and an increase in the surface area is

generally obtained by decreasing the particle size of the

adsorbent. Not surprisingly, with the introduction of nano-

scaled iron-oxide based materials, the contaminant removal

efficiency can be increased dramatically. For example, Yavuz

et al. (2006), Mayo et al. (2007), and Hu et al. (2005a) reported

a significant increase in heavy metal removal capacities using

10 nm nonporous iron-oxide nanoparticles (Fe3O4, g-Fe2O3) as

compared with big sized iron-oxide particles or bulk mate-

rials. However, as the size of a magnetic adsorbent decreases,

it has also been noticed that its response to an external

magnetic field decreases undesirably, making it increasingly

difficult to retract the adsorbent after treatment is completed

(Yavuz et al., 2006). Although an even higher magnetic field

can still be applied to achieve a complete separation of the

adsorbent, the cost of applying such a field might be so high

that magnetic separation loses ground against other conven-

tional separation methods. Thus, the ultimate objective of this

study is to remove Cr(VI) from contaminated water using

a newly synthesized magnetic iron-oxide based adsorbent

that simultaneously possesses a high surface area and a high

response to external magnetic fields.

Recently, mesoporous materials have attracted a lot of

attention in both scientific and industrial communities since

the introduction of well-ordered mesoporous silicas in the

1990s because of their large surface areas and uniform and

tunable pore sizes (2–50 nm) (Kresge et al., 1992; Zhao et al.,

1998). The characteristics of mesoporous materials are also

attractive to researchers seeking adsorbents for environ-

mental remediation, not only because of their high surface

area but also of their fast contaminant adsorption kinetics

(Wang et al., 2009). Although the synthesis of mesoporous

silicas and related materials has been well documented, well-

ordered mesoporous magnetic Fe2O3 (i.e., g-Fe2O3) was not

synthesized until recently (Jiao et al., 2006a, b). In addition to

a high surface area, synthesized mesoporous g-Fe2O3 usually

has a much bigger particle size (>200 nm) compared with

10 nm nonporous g-Fe2O3 and presumably its response to an

external magnetic field would be stronger than that of 10 nm

nonporous g-Fe2O3, resulting in a faster separation of meso-

porous g-Fe2O3. Therefore, mesoporous g-Fe2O3 has the

potential of being a more efficient and cost-effective adsor-

bent for removal of Cr(VI) from wastewater because a short-

ened separation time of an adsorbent reduces the operation

cost.

The specific objectives of the study are (1) to synthesize

mesoporous g-Fe2O3 using a modified, ‘greener’ method,

which involves recycling and reusing the solvent; and (2) to

evaluate the effectiveness of the synthesized mesoporous

g-Fe2O3 as a potential adsorbent for Cr(VI) removal from

wastewater as compared with 10 nm nonporous g-Fe2O3 in

terms of Cr(VI) adsorption capacities and magnetic

separation.

2. Materials and methods

2.1. Synthesis and characterization of mesoporousg-Fe2O3

The synthesis of mesoporous g-Fe2O3 involved three steps:

(1) synthesis of mesoporous a-Fe2O3 using a hard-templating

method; (2) reduction of a-Fe2O3 to Fe3O4 by H2 treatment;

(3) oxidation of Fe3O4 to g-Fe2O3. In a typical synthesis of

mesoporous a-Fe2O3, 3.0 g of Fe(NO3)3$9H2O was mixed with

80 mL of xylene at 60 �C, followed by the addition of 2.0 g of

mesoporous silica template, KIT-6. The preparation of KIT-6

was conducted following the procedure described by Kleitz

et al. (2003). After stirring at 60 �C for 4 h, the mixture was

quickly filtered. The filtrate (xylene) was collected and stored

for reuse again, whereas the filtered powder was dried over-

night at 50 �C. The dry powder was then slowly heated to

600 �C and calcined at that temperature for 6 h. The resulting

samples were treated with hot 2.0 M NaOH for three times to

remove the silica template, followed by drying overnight at

50 �C. This procedure leads to mesoporous a-Fe2O3, which is

nonmagnetic. Reduction of a-Fe2O3 to Fe3O4 was achieved by

heating the previously prepared mesoporous a-Fe2O3 at 300 �C

for 3 h under a 5% H2 atmosphere, and a mild oxidation of

Fe3O4 to g-Fe2O3 was obtained by heating the as-prepared

mesoporous Fe3O4 at 150 �C in air for 2 h.

A transmission electron microscope (TEM) (JEOL JEM2010,

Japan), equipped with an energy dispersive X-ray analyzer

(EDX), was used to characterize the structure properties of

the synthesized materials. Scanning electron microscopy

(SEM) studies were performed using an FEI XL40 Sirion FEG

microscope. The composition of the materials was identified

by powder X-ray diffraction (XRD) (Philips PW-1830,

Netherlands). Magnetization measurement was performed

using a vibrating sample magnetometer (VSM) (LakeShore

EM7037/9509-P, USA) at room temperature. Surface area

measurements were taken with a Brunauer, Emmett, Teller

(BET) (Coulter SA-3100, USA) analyzer at liquid nitrogen

temperature using conventional gas adsorption apparatus.

Fourier transform infrared spectroscopy (FTIR) measurement

was also conducted, and infrared spectra of diluted samples in

KBr were recorded between 4000 and 400 cm�1 in a Bruker

IFS-66 V FTIR.

2.2. Synthesis of 10 nm nonporous g-Fe2O3

In the literature, 10 nm nonporous magnetic iron-oxide

nanoparticles (Fe3O4, g-Fe2O3), due to their high surface areas,

have been reported as effective magnetic adsorbents for

heavy metal removal. In this study, to compare the effec-

tiveness of mesoporous g-Fe2O3 in terms of Cr(VI) removal and

magnetic separation, 10 nm g-Fe2O3 nanoparticles were

synthesized and used for Cr(VI) removal as well. Single crystal

250

550

20 30 40 50 60 70

2θ (degrees)

Inte

nsit

y

(b)

(a)

Fig. 1 – Wide-angle XRD for mesoporous (a) a-Fe2O3;

(b) g-Fe2O3.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 3 7 2 7 – 3 7 3 4 3729

10 nm Fe3O4 nanoparticles were first synthesized using

a method modified from Kang et al. (1996). Briefly, 6.1 g of

FeCl3$6H2O and 4.2 g FeSO4.7H2O were dissolved in 100 mL

ultrapure water. A total of 25 mL 6.5 M NaOH was then slowly

added and mixed with the above solution. The system was

mixed for a further hour after the addition of NaOH was

completed. The formed black precipitates were washed with

ultrapure water several times with an assistance of an

external magnetic field. This procedure leads to Fe3O4 nano-

particles with a size of around 10 nm. The 10 nm Fe3O4

nanoparticles were then oxidized in air at 150 �C for 2 h to

obtain 10 nm nonporous g-Fe2O3.

2.3. Batch experiments

A chromium stock solution was prepared by dissolving

a known quantity of potassium chromate (K2CrO4) in ultra-

pure water. Batch Cr(VI) adsorption studies were performed by

mixing 0.05 g of mesoporous g-Fe2O3 with 40 mL of solution of

varying Cr(VI) concentrations (5–100 mg/L) in a 40 mL glass

vial end-over-end to reach Cr(VI) adsorption equilibrium.

After adsorption reached its equilibrium, the adsorbent was

separated by using a hand-held permanent magnet and the

supernatant was collected for Cr(VI) concentration measure-

ments. The Cr(VI) adsorption on the mesoporous g-Fe2O3 was

first studied at three different pH values (2.5, 5.0, 7.0) to

investigate the dependence of Cr(VI) adsorption on solution

pH. Standard acid of 0.1 M HCl and a base of 0.1 M NaOH

solution were used for pH adjustment. Solution pH was stable

over the course of the experiments. All the adsorption

experiments were carried out at a room temperature of

22� 2 �C and were performed in duplicate. Based on the

results of pH dependence of Cr(VI) adsorption, further Cr(VI)

adsorption studies were conducted only at a pH of 2.5 unless

otherwise specified. The total aqueous concentrations of

chromium were measured using a flame atomic absorption

spectrometer (AAS, Varian 220FS). Sample dilution was con-

ducted before AAS measurement where necessary.

In some cases, adsorption experiments were immediately

followed by desorption experiments. Briefly, after adsorption

was completed and the supernatant was decanted, the chro-

mium-loaded mesoporous g-Fe2O3 (0.05 g) was then mixed

with 40 mL of 0.01 M NaOH to reach Cr(VI) desorption equi-

librium. To test whether any chemical redox reaction occur-

ring during the adsorption process, the concentration of Cr(VI)

in the desorption solution was measured by 1,5-diphe-

nylcarbazide colorimetric method, using a UV/visible spec-

trophotometer (Ultrospec 4300 Pro) at wavelengths of 540 nm,

to check both the concentration of Cr(VI) and the speciation of

chromium during the adsorption/desorption process. The

kinetics of both Cr(VI) adsorption and desorption was inves-

tigated as well. The Cr(VI) adsorption capacity onto the 10 nm

nonporous g-Fe2O3 nanoparticles was also measured at

pH¼ 2.5 and compared with that onto the mesoporous

g-Fe2O3 under the same conditions.

2.4. Regeneration and reuse of mesoporous g-Fe2O3

To study the regeneration and reusability of the mesoporous

g-Fe2O3 as an adsorbent for Cr(VI) removal, experiments

pertaining to mesoporous g-Fe2O3 regeneration and Cr(VI)

re-adsorption were carried out in 5 consecutive adsorption/

desorption cycles. For each cycle, 40 mL of 50 mg/L Cr(VI)

solution was adsorbed first by 0.05 g of mesoporous g-Fe2O3

for 120 min to reach adsorption equilibrium. The supernatant

was then decanted with an assistance of the permanent

magnet and the adsorbed Cr(VI) on mesoporous g-Fe2O3 was

then desorbed with 40 mL of 0.01 M NaOH for 120 min. After

each cycle of adsorption/desorption, the mesoporous g-Fe2O3

was washed thoroughly with ultrapure water to neutrality and

then reconditioned for adsorption in the succeeding cycle.

3. Results and discussions

3.1. Synthesis and characterization of mesoporousFe2O3

Unlike the synthesis method reported by Jiao et al. (2006a, b),

xylene was used in place of ethanol as the solvent. Ethanol is

a polar solvent, in which Fe(NO3)3$9H2O is soluble. Therefore,

in Jiao’s method, evaporation induced impregnation was

employed to get the iron precursor into the mesoporous silica

template, which inevitably led to the loss of the solvent. In the

adapted synthesis, a nonpolar solvent (xylene) was used,

within which the iron precursor entered the mesopores of the

templating silica via an entropy driven mechanism. Since the

internal surfaces of the silica mesopores are largely hydro-

philic (�SiOH), Fe3þ, in the form of hydrated Fe3þ, much

prefers to enter these mesopores rather than staying in the

nonpolar bulk solution. Thus, xylene serves as a medium

which drives the iron precursor into the mesopores and it itself

is not consumed at all. Based on this, the advantage of the use

of a nonpolar solvent in the synthesis is that the solvent can be

recycled and reused, leading to a ‘greener’ synthesis.

Fig. 1 presents wide-angle XRD patterns of the synthesized

a-Fe2O3 and g-Fe2O3. As can be seen, well-defined peaks cor-

responding to the crystal structure of a-Fe2O3 and g-Fe2O3 are

clearly evident, in agreement with the previous studies (Jiao

et al., 2006a, b), suggesting that the walls of both the meso-

porous a-Fe2O3 and g-Fe2O3 are crystalline. It should be noted

that because Fe3O4 gradually converts to g-Fe2O3 phase on

Fig. 2 – TEM images of mesoporous (a) a-Fe2O3; (b) g-Fe2O3; SEM image of mesoporous g-Fe2O3 (c) and TEM image of 10 nm

g-Fe2O3 (d). Note: the scale bar for (b) is 20 nm; that for (c) is 500 nm.

0

300

600

900

0.5 1 1.5

2θ (degrees)2 2.5 3

Inte

nsit

y

(a)

(b)

Fig. 3 – Low-angle XRD for (a) a-Fe2O3; (b) g-Fe2O3.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 3 7 2 7 – 3 7 3 43730

exposure to air, no attempts were made to characterize mes-

oporous Fe3O4 in this study.

Fig. 2(a) and (b) present TEM images of the mesoporous

a-Fe2O3 and g-Fe2O3. As seen, the hard-templating method

produces a well-ordered mesoporous a-Fe2O3 structure (a) and

converting it to Fe3O4 by reduction and then to g-Fe2O3 (b) by

oxidation retains largely the same ordered mesostructure

throughout. The ability to carry out solid/solid trans-

formations in mesoporous solids, with retention of the mes-

ostructure, has been reported previously for metal-oxide

based materials (Brezesinski et al., 2006; Shi et al., 2008).

Fig. 2(c) presents an SEM image of the synthesized meso-

porous g-Fe2O3. As shown, although the mesoporous g-Fe2O3

particles are not monodispersed in size, they have a size

greater than 200 nm. The results of low-angle XRD for mes-

oporous a-Fe2O3 and g-Fe2O3 are shown in Fig. 3. Both

materials exhibit a 2q peak around 0.9�, reflecting an ordered

pore structure. N2 sorption analysis showed that the surface

areas of the mesoporous a-Fe2O3 and g-Fe2O3 were 108 and

88 m2/g respectively and their mean pore sizes were both

about 4 nm. A TEM image of the synthesized 10 nm nonpo-

rous g-Fe2O3 is presented in Fig. 2(d). The surface area of the

10 nm nonporous g-Fe2O3 was measured to be 95 m2/g, which

is close to the value reported in the literature (Tuutijarvi

et al., 2009) and to that of the mesoporous g-Fe2O3 synthe-

sized in this study.

Fig. 4(a) presents the VSM measurement of 10 nm nonpo-

rous g-Fe2O3 and mesoporous g-Fe2O3. Although Fig. 4(a) is not

intended to show the difference between the two types of

materials in terms of their saturation magnetization, the

magnitude of the saturation magnetization indicates that

both materials are highly magnetic (Lu et al., 2007). Also, as

indicated by a zero magnetic moment when the external

magnetic field is absent, 10 nm g-Fe2O3 is superparamagnetic,

while mesoporous g-Fe2O3 is not. More importantly, the

ease with which the two types of materials can be separated

Mesoporousγ-Fe2O3

10nmγ-Fe2O3

Mesoporousγ-Fe2O3

10nmγ-Fe2O3

-50

-25

0

25

50

-6000 -4000 -2000 0 2000 4000 6000

Field (Oe)

Mom

ent

(em

u/g)

mesoporous γ-Fe2O3

10nm γ-Fe2O3

b

a

c

Fig. 4 – (a) VSM measurements for 10 nm g-Fe2O3 and mesoporous g-Fe2O3; demonstration of magnetic separation at

(b) 5 min; and (c) 6 h.

0

2

4

6

8

10

12

14

16

18

0 20 40 60 80 100

Equilibrium Cr(VI) concentration (mg/L)

Cr(

VI)

ads

orbe

d co

ncen

trat

ion

(mg/

g)

pH=2.5 pH=5.0 pH=7.0

Fig. 5 – Cr(VI) adsorption onto mesoporous g-Fe2O3 at

various pH values.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 3 7 2 7 – 3 7 3 4 3731

by a magnetic field significantly differentiates each type.

A qualitative ranking of the magnetic susceptibility of the two

materials was demonstrated in a simple laboratory setup with

a hand-held magnet as presented in Fig. 4(b) and (c). The

mesoporous g-Fe2O3 could be completely separated from the

aqueous solution within 5 min, while it took hours for 10 nm

g-Fe2O3 nanoparticles.

Presumably, the reasons for the differentiated magnetic

separation behaviors are twofold: first, since mesoporous

g-Fe2O3 is not superparamagnetic while 10 nm nonporous

g-Fe2O3 is, mesoporous g-Fe2O3 responds to the same external

magnetic field more strongly than 10 nm nonporous g-Fe2O3;

secondly, because 10 nm nonporous g-Fe2O3 nanoparticles are

virtually individual single iron-oxide crystals with a size

around 10 nm (Fig. 2(d)), while mesoporous g-Fe2O3 is actually

an cluster of single iron-oxide crystals with crystal size

comparable with 10 nm g-Fe2O3. However, the overall particle

size of mesoporous g-Fe2O3 is greater than 200 nm (Fig. 2(c)).

With a lot of single crystals being assembled together, the

average resistance to each single crystal of mesoporous

g-Fe2O3, exerted by the medium, against its directed move-

ment under an applied magnetic field is much reduced,

leading to a much easier separation for mesoporous g-Fe2O3.

This result is similar to that of settling separation in an

aqueous phase: although the gravitational acceleration

constant (g) is the same all the time, larger particles settle

faster than smaller ones due to less resistance exerted by the

aqueous medium on each unit weight of the larger particles

than of the smaller ones.

3.2. Cr(VI) adsorption onto mesoporous g-Fe2O3

The experimental data on Cr(VI) adsorption onto mesoporous

g-Fe2O3 at various pH values are presented in Fig. 5. The Cr(VI)

0%

20%

40%

60%

80%

100%

0 20 40 60 80 100 120

Time (minutes)

Rel

ativ

e %

Cr(

VI)

rem

oved

0%

20%

40%

60%

80%

100%

0 20 40 60 80 100

Time (minutes)

Rel

ativ

e %

Cr(

VI)

rec

over

ed

a b

Fig. 6 – Cr(VI) (a) adsorption and (b) desorption kinetics from mesoporous g-Fe2O3.

0400600800100012001400

Wave number (cm-1)

Abs

orba

nce

mesoporous γ -Fe2O3 with adsorbed Cr(VI)

mesoporous γ -Fe2O3 only

CrO42-

Fig. 7 – FTIR spectra of chromium adsorbed mesoporous

g-Fe2O3.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 3 7 2 7 – 3 7 3 43732

adsorption onto mesoporous g-Fe2O3 showed a saturation

adsorption behavior at high equilibrium aqueous Cr(VI)

concentrations. Clearly, the Cr(VI) saturation adsorption

capacities increased sharply from around 4.5–15.6 mg/g

(calculated as the average of the top three points) when the pH

was decreased from 7 to 2.5. It has been proposed that, at low

pH values, iron-oxide surfaces are protonated so that the net

surface charge is positive, which enhances the adsorption of

the negatively charged oxyanionic Cr(VI) (CrO42�) species. As

the pH increases, the iron-oxide surfaces are increasingly

deprotonated so that the net surface positive charges are

decreasing, leading to a reduction in Cr(VI) adsorption (Hu

et al., 2005a). Thus, the pH-dependent behavior of Cr(VI)

adsorption onto the mesoporous g-Fe2O3 suggests that Cr(VI)

adsorption is via physical adsorption, i.e., electrostatic

attraction, at low pH. The surface charge is neutral at the point

of zero charge (PZC), which is 6.5 for the mesoporous g-Fe2O3.

At pH higher than PZC, the Cr(VI) adsorption can be explained

by anion exchange on the metal-oxide surfaces (Hu et al.,

2005b).

Although a direct comparison of the mesoporous g-Fe2O3

with other adsorbents is difficult, due to the different experi-

mental conditions, it was found, in general, that the adsorp-

tion capacity of the mesoporous g-Fe2O3 for Cr(VI) at pH of 2.5

(15.6 mg/g) is higher than or comparable with those of diato-

mite (11.55 mg/g), anatase (14.56 mg/g), commercial activated

carbon (15.47 mg/g), and beech sawdust (16.13 mg/g) (Sandhya

and Tonnin, 2004; Hu et al., 2005a). The adsorption kinetics of

Cr(VI) onto mesoporous g-Fe2O3 at a pH of 2.5 is shown in

Fig. 6. The rate of Cr(VI) adsorption (Fig. 6(a)) was quite fast.

About 90% of the total Cr(VI) adsorption occurred during the

first 5 min of the reaction, while only a very small part of the

additional adsorption appeared during the following 15 min of

contact. The rapid adsorption of Cr(VI) by mesoporous g-Fe2O3

further suggests the adsorption mechanism is mainly due to

electrostatic attraction and it also implies that the mesopores

are not a limiting factor for Cr(VI) diffusion into the interior of

the mesoporous g-Fe2O3, which is consistent with other

studies using mesoporous materials (Wang et al., 2009). In

comparison, the equilibrium time for adsorption of Cr(VI) by

some other adsorbents is much longer. For instance, adsorp-

tion of Cr(VI) onto activated carbon is around 10–50 h (Aggar-

wal et al., 1999).

The effect of common ions in chrome-plating wastewater

(Naþ, Ca2þ, Mg2þ, Cu2þ, Ni2þ, NO3� and Cl�) on Cr(VI) adsorp-

tion onto g-Fe2O3 has been thoroughly studied in our previous

work and the competition of these ions on Cr(VI) adsorption

was found negligible (Hu et al., 2005a). These results indicate

a high selectivity of g-Fe2O3 for Cr(VI), implying a potential of

recycling Cr(VI) back to industries by using mesoporous

g-Fe2O3.

In this study, for the purpose of comparison, the same

adsorption experiments were conducted with 10 nm g-Fe2O3.

The saturation adsorption capacity of Cr(VI) onto 10 nm

g-Fe2O3 was measured as 14.6 mg/g at pH 2.5, which is close to

that of mesoporous g-Fe2O3 (15.6 mg/g). Given the fact that

mesoporous g-Fe2O3 can be separated much more easily than

10 nm g-Fe2O3, the operation cost for Cr(VI) removal can be

much reduced using mesoporous g-Fe2O3.

FTIR technique provides information on the state of

adsorbed molecules (particularly anions), thereby shedding

light on adsorption mechanisms. The FTIR spectra of the

sample collected after Cr(VI) adsorption onto the mesoporous

g-Fe2O3 at pH 2.5 is shown in Fig. 7. A new peak at 948 cm�1 in

the FTIR spectrum occurred, which belongs to CrO42� (Hu et al.,

2007). This serves as another evidence of physical adsorption

because in a physical adsorption at the mineral–water inter-

face, an oxyanion will retain its hydration shell and will not

form a direct chemical bond with the oxide surface (Petit et al.,

1995).

Fig. 8 – (a) TEM image of Cr(VI)-loaded mesoporous g-Fe2O3 and (b) the relevant EDX spectra of the circled area (A) in (a).

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 3 7 2 7 – 3 7 3 4 3733

TEM image of the mesoporous g-Fe2O3 sample after Cr(VI)

adsorption at pH of 2.5 with 40 mL of 100 mg/L Cr(VI) and the

relevant EDX spectra of the marked area of the sample is

shown in Fig. 8. EDX analysis is an analytical technique used

for the elemental analysis or chemical characterization of

a sample. The percentage of chromium within the circled area

(A as in Fig. 8(a)) is around 1.8%, which corresponds well with

the measured amount of Cr(VI) adsorbed onto the sample

(15.6 mg/g), implying a largely uniform Cr(VI) adsorption

within the mesoporous structure.

3.3. Regeneration and reuse of mesoporous g-Fe2O3

As shown earlier, since the adsorption of Cr(VI) onto the mes-

oporous g-Fe2O3 was highly dependent on the solution pH, the

desorption of Cr(VI) can be achieved by increasing the solution

pH. In this study, the Cr(VI)-loaded mesoporous g-Fe2O3 was

desorbed by using 0.01 M NaOH. Fig. 6(b) presents the desorp-

tion kinetics of the adsorbed Cr(VI). It should be noted that the

amount of Cr(VI) desorbed was measured with 1,5-diphe-

nylcarbazide colorimetric method using a UV/visible spectro-

photometer to check the speciation of chromium as this

method only responds to Cr(VI). An almost full recovery of

Cr(VI) supports the hypothesis that there is no redox reaction,

confirming that there is no chemical adsorption taking place.

This is one advantage of using iron-oxide based materials, with

80%

90%

100%

1 2 3 4 5

Cycles

% R

emov

ed

60%

70%

80%

90%

100%

% R

ecovered

Removed

Recovered

Fig. 9 – Percentage Cr(VI) removed and recovered during

five adsorption/desorption cycles.

iron oxidation state beingþ3, because the highest iron valency

(þ3) ensures that there is no redox taking place to reduce Cr(VI)

to Cr(III), and hence permits the regeneration of the adsorbent.

In the case with Fe3O4, where part of the iron is in an oxidation

state of þ2, chemical adsorption of Cr(VI) has been reported

and the formation of precipitated Cr(III) on the adsorbent

surfaces led to irreversible adsorption (Hu et al., 2004). In

addition, the rapid desorption kinetics of the adsorbed Cr(VI)

also serves as a proof of electrostatic interaction involved in the

Cr(VI) adsorption onto the mesoporous g-Fe2O3.

In a wastewater treatment process that uses adsorption,

regeneration of the adsorbent is crucially important. Nowa-

days, in many applications, reuse of the adsorbent through

regeneration of its adsorption properties is an economic

necessity. As mentioned previously, a complete desorption of

Cr(VI) can be achieved by using 0.01 M NaOH solution as an

extraction medium. To test the regeneration and reusability of

the mesoporous g-Fe2O3, the mesoporous g-Fe2O3 was used in

five consecutive adsorption/desorption cycles, with 40 mL of

50 mg/L Cr(VI) solution being used in the adsorption step

while 40 mL of 0.01 M NaOH in the desorption cycle as

desorbing agent. As shown in Fig. 9, at the end of the fifth

cycle, the mesoporous g-Fe2O3 retained more than 90% of its

original Cr(VI) adsorption capacity, and >90% of the total

adsorbed Cr(VI) can be recovered during the desorption steps.

4. Conclusions

In this study, mesoporous magnetic Fe2O3 (g-Fe2O3) was

synthesized using a ‘greener’ synthesis method. The synthe-

sized mesoporous g-Fe2O3 simultaneously has a surface area

comparable with and a higher susceptibility to magnetic

separation than 10 nm nonporous g-Fe2O3. Cr(VI) adsorption

onto mesoporous g-Fe2O3 exhibited a highly pH-dependent

behavior, with the Cr(VI) adsorption capacity increasing with

decreasing pH. Cr(VI) adsorption capacity of mesoporous

g-Fe2O3 is comparable with that of 10 nm nonporous g-Fe2O3,

but a much faster magnetic separation of mesoporous g-Fe2O3

makes it a better adsorbent for Cr(VI) than 10 nm nonporous

g-Fe2O3. The results showed that mesoporous g-Fe2O3 can be

regenerated, maintaining almost the same Cr(VI) adsorption

capacity. Therefore, mesoporous g-Fe2O3 is a promising

adsorbent for Cr(VI) removal from contaminated water.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 3 7 2 7 – 3 7 3 43734

Acknowledgements

This research was supported by the Hong Kong University of

Science and Technology Research Project Competition

Program under grant RPC07/08.EG03.

r e f e r e n c e s

Aggarwal, D., Goyal, M., Bansal, R.C., 1999. Adsorption ofchromium by activated carbon from aqueous solution. Carbon37 (12), 1989–1997.

Ai, Z.H., Cheng, Y., Zhang, L.Z., Qiu, J.R., 2008. Efficient removal ofCr(VI) from aqueous solution with Fe@Fe2O3 core-shellnanowires. Environ. Sci. Technol. 42 (18), 6955–6960.

Brezesinski, T., Groenewolt, M., Antonietti, M., Smarsly, B., 2006.Crystal-to-crystal phase transition in self-assembledmesoporous iron oxide films. Angew. Chem. Int. Ed. 45 (5),781–784.

Dabrowski, A., Hubicki, Z., Podkoscielny, P., Robens, E., 2004.Selective removal of the heavy metal ions from waters andindustrial wastewaters by ion-exchange method.Chemosphere 56 (2), 91–106.

Fendorf, S., Eick, M.J., Grossl, P., Sparks, D.L., 1997. Arsenate andchromate retention mechanisms on goethite. 1. Surfacestructure. Environ. Sci. Technol. 31 (2), 315–320.

Hu, J., Chen, G.H., Lo, I.M.C., 2005a. Removal and recovery of Cr(VI) from wastewater by maghemite nanoparticles. Water.Res. 39 (18), 4528–4536.

Hu, J., Lo, I.M.C., Chen, G.H., 2005b. Fast removal and recovery ofCr(VI) using surface-modified jacobsite (MnFe2O4)nanoparticles. Langmuir 21 (24), 11173–11179.

Hu, J., Lo, I.M.C., Chen, G.H., 2007. Performance and mechanism ofchromate (VI) adsorption by delta-FeOOH-coated maghemite(gamma-Fe2O3) nanoparticles. Sep. Purif. Technol. 58 (1), 76–82.

Hu, J., Lo, I.M.C., Chen, G., 2004. Removal of Cr(VI) by magnetitenanoparticle. Water Sci. Technol. 50 (12), 139–146.

Jiao, F., Harrison, A., Jumas, J., Chadwick, A.V., Kockelmann, W.,Bruce, P.G., 2006a. Ordered mesoporous Fe2O3 with crystallinewalls. J. Am. Chem. Soc. 128 (16), 5468–5474.

Jiao, F., Jumas, J., Womes, M., Chadwick, A.V., Harrison, A.,Bruce, P.G., 2006b. Synthesis of ordered mesoporous Fe3O4 andg-Fe2O3 with crystalline walls using post-template reduction/oxidation. J. Am. Chem. Soc. 128 (39), 12905–12909.

Kang, Y.S., Rishud, S., Rabolt, J.F., Stroeve, P., 1996. Synthesis andcharacterization of nanoparticle-size Fe3O4 and g-Fe2O3

particles. Chem. Mater. 8 (9), 2209–2211.Katz, S.A., Salem, H., 1993. The toxicology of chromium with

respect to its chemical speciation: a review. J. Appl. Toxicol. 13(3), 217–224.

Kleitz, F., Choi, S.H., Ryoo, R., 2003. Cubic Ia3d large mesoporoussilica: synthesis and replication to platinum nanowires,carbon nanorodes and carbon nanotubes. Chem. Comm. 17,2136–2137.

Kresge, C.T., Leonowica, M.E., Roth, W.J., Vartuli, J.C., Beck, J.S.,1992. Ordered mesoporous molecular sieves synthesizedby a liquid-crystal template mechanism. Nature 359,710–712.

Lu, H.M., Zheng, W.T., Jiang, Q., 2007. Saturation magnetization offerromagnetic and ferrimagnetic nanocrystals at roomtemperature. J. Phys. D: Appl. Phys. 40 (2), 320–325.

Mayo, J.T., Yavuz, C., Yean, S., Cong, L., Shipley, H., Yu, W.,Falkner, J., Kan, A., Tomson, M., Colvin, V.L., 2007. The effect ofnanocrystalline magnetite size on arsenic removal. Sci.Technol. Adv. Mater. 8, 71–75.

Petit, F., Debontride, H., Lenglet, F., Juhel, G., Verchere, D., 1995.Contribution of spectrometric methods to the study of theconstituents of chromating layers. Appl. Spectrosc. 49 (2),207–210.

Sandhya, B., Tonnin, A.K., 2004. Cr(VI) removal from syntheticwastewater using coconut shell charcoal and commercialactivated carbon modified with oxidizing agents and/orchitosan. Chemosphere 54 (7), 951–967.

Shevchenko, N., Zaitsev, V., Walcarius, A., 2008. Bifunctionalizedmesoporous silicas for Cr(VI) reduction and concomitantCr(III) immobilization. Environ. Sci. Technol. 42 (18),6922–6928.

Shi, Y.F., Wan, Y., Zhang, R.Y., Zhao, D.Y., 2008. Synthesis of self-supported ordered mesoporous cobalt and chromium nitrides.Adv. Funct. Mater. 18 (16), 2436–2443.

Singh, V.K., Tiwari, P.N., 1997. Removal and recovery ofchromium(VI) from industrial wastewater. J. Chem. Technol.Biotechnol. 69 (3), 376–382.

Tuutijarvi, T., Lu, J., Sillanpaa, M., Chen, G., 2009. As(V) adsorptionon maghemite nanoparticles. J. Hazard. Mater. 166 (2–3), 1415–1420.

Vijayaraghavan, K., Yun, Y.S., 2008. Bacterial biosorbents andbiosorption. Biotechnol. Adv. 26 (3), 266–291.

Volesky, B., 2007. Biosorption and me. Water. Res. 41 (18),4017–4029.

Wang, P., Shi, Q.H., Shi, Y.F., Clark, C.K., Stucky, G.D., Keller, A.A.,2009. Magnetic permanently confined micelle arrays fortreating hydrophobic organic compound contamination.J. Am. Chem. Soc. 131 (1), 182–188.

Yavuz, C.T., Mayo, J.T., Yu, W.W., Prakash, A., Falkner, J.C.,Yean, S., Cong, L.L., Shipley, H.J., Kan, A., Tomson, M.,Natelson, D., Colvin, V.L., 2006. Low-field magnetic separationof monodisperse Fe3O4 nanocrystals. Science 314 (5801),964–967.

Zhao, D.Y., Feng, J.L., Huo, Q.S., Melosh, N., Fredrickson, G.H.,Chmelka, B.F., Stucky, G.D., 1998. Triblock copolymersyntheses of mesoporous silica with periodic 50 to 300angstrom pores. Science 279 (5350), 548–552.