Environ. Sci. Technol.
2007, 41, 4613-4619
Removal Mechanism of As(III) by a
Novel Fe-Mn Binary Oxide
Adsorbent: Oxidation and Sorption
G A O - S H E N G Z H A N G , †,‡ J I U - H U I Q U , * ,†
HUI-JUAN LIU,† RUI-PING LIU,† AND
G U O - T I N G L I †,‡
State Key Laboratory of Environmental Aquatic Chemistry,
Research Center for Eco-Environmental Sciences, Chinese
Academy of Sciences, Beijing, China 100085, and Graduate
School, Chinese Academy of Sciences, Beijing, China 100039
A novel Fe-Mn binary oxide adsorbent was developed
for effective As(III) removal, which is more difficult to remove
from drinking water and much more toxic to humans
than As(V). The synthetic adsorbent showed a significantly
higher As(III) uptake than As(V). The mechanism study is
therefore necessary for interpreting such result and
understanding the As(III) removal process. A control
experiment was conducted to investigate the effect of Na2SO3-
treatment on arsenic removal, which can provide useful
information on As(III) removal mechanism. The adsorbent
was first treated by Na2SO3, which can lower its oxidizing
capacity by reductive dissolution of the Mn oxide and then
reacted with As(V) or As(III). The results showed that
the As(V) uptake was enhanced while the As(III) removal
was inhibited after the pretreatment, indicating the
important role of manganese dioxide during the As(III)
removal. FTIR along with XPS was used to analyze the
surface change of the original Fe-Mn adsorbent and the
pretreated adsorbent before and after reaction with
As(V) or As(III). Change in characteristic surface hydroxyl
groups (Fe-OH, 1130, 1048, and 973 cm-1) was observed
by the FTIR. The determination of arsenic oxidation state on
the solid surface after reaction with As(III) revealed that
the manganese dioxide instead of the iron oxide oxidized
As(III) to As(V). The iron oxide was dominant for adsorbing
the formed As(V). An oxidation and sorption mechanism
for As(III) removal was developed. The relatively higher As(III)
uptake may be attributed to the formation of fresh
adsorption sites at the solid surface during As(III) oxidation.
Introduction
Arsenic, a relatively scarce but ubiquitous element, is of
serious concern due to its toxicity and carcinogenicity even
at low concentrations (1). In natural water, arsenic is primarily
present in inorganic forms and exists in two predominant
species: arsenate [As(V)] and arsenite [As(III)]. As(III) is much
more toxic (2), soluble, and mobile (3) than As(V). Therefore,
developing an economical, effective, and reliable water
treatment technique that is capable of removing both As(V)
and As(III) from contaminated drinking water and under-
standing arsenic behavior in the treatment processes are
gaining considerable attentions in recent years.
* Corresponding author phone: +86 10 62849151; fax: +86 10
6292355; e-mail: jhqu@rcees.ac.cn.
† State Key Laboratory of Environmental Aquatic Chemistry.
‡ Chinese Academy of Sciences.
10.1021/es063010u CCC: $37.00  2007 American Chemical Society
Published on Web 05/26/2007
Adsorption is considered to be one of the most promising
technologies because it can be simple in operation and cost-
effective (4). Iron hydroxides and oxides are the most used
in the adsorption due to the higher affinity of iron oxide
toward inorganic arsenic species. Numerous studies (5-7)
have documented the adsorption of As(V) and As(III) on iron-
(III) (hydr)oxides. At the concentrations normally found in
natural water, the As(III) adsorption is normally less effective
than the As(V) adsorption by these adsorbents. To achieve
higher arsenic removal, a pretreatment for As(III) oxidation
is usually adopted. The adsorption mechanism has also been
investigated by a number of researchers using EXAFS and IR
spectroscopic techniques and they show that arsenic forms
bi-nuclear bidentate as well as monodentae complexes on
the iron oxide surface (8-10).
Manganese oxides have been extensively investigated as
oxidizing agents for arsenite (11-14), and some researchers
observed the overall oxidation reaction (shown in reaction
1). Recent works by Moore et al. (15) and Nesbitt et al. (13)
demonstrated that the oxidation of As(III) by the synthetic
birnessite surface proceeds by a two step pathway (given in
reactions 2 and 3), involving the reduction of Mn(IV) to
Mn(III) and then Mn(III) to Mn(II).
MnO2 + H3AsO3 +2H+ ) Mn2+ + H3AsO4 + H2O (1)
2MnO2 + H3AsO3 ) 2MnOOH* + H3AsO4 (2)
2MnOOH* + H3AsO3 + 4H+) 2Mn2+ + H3AsO4 + 3H2O
(3)
where MnOOH* represents Mn(III) intermediate reaction
product. An electron-transfer mechanism and a substitution
mechanism were established and could well interpret the
redox reaction process (13).
In our study, a novel Fe-Mn binary oxide adsorbent
developed by a chemical coprecipitation method was found
to have a much higher adsorption capacity toward As(III)
than that of As(V). The similar phenomenon was observed
by Deschamps et al. (16) who used a natural Fe and Mn
enriched material to remove arsenic. They ascribed the higher
As(III) uptake to the oxidation ability of manganese oxides
content. However, the detailed mechanism of As(III) removal
by the Fe-Mn material was yet to be elucidated. Infrared
vibrational spectroscopy (IR) and X-ray photoelectron spec-
troscopy (XPS) are powerful techniques for the study of As(III)
removal mechanisms by iron oxides and manganese oxides
and were used by many researchers (9, 10, 13). Their results
provide very useful information on As(III) removal mech-
anism by single iron oxides or manganese oxides. Neverthe-
less, these results cannot be easily applied to predict/explain
the sorption behavior of As(III) on binary oxides.
Thus, a control experiment was designed to investigate
the effect of the MnO2 content in Fe-Mn binary oxide on
arsenic removal, in which Na2SO3 was used as reducing agent
to lower its oxidizing capacity by reductive dissolution of the
Mn oxide. A detailed study on the oxidation and adsorption
of As(III) by Fe-Mn binary oxide has been described using
FTIR and XPS techniques in the present paper. The objectives
of this study were (i) to investigate the respective roles of
iron oxide and manganese oxide content on As(III) removal;
(ii) to compare the FTIR and XPS spectra of the adsorbent
before and after reaction with As(V) or As(III); and finally (iii)
to explain the removal mechanism of As(III) by Fe-Mn binary
oxide.
VOL. 41, NO. 13, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4613
Materials and Methods
Materials. All chemicals were analytical grade and were
purchased from Beijing Chemical Co. (Beijing, China). The
As(III) and As(V) stock solutions were prepared with deionized
water using NaAsO2 and NaHAsO4‚7H2O, respectively. Arsenic
working solutions were freshly prepared by diluting arsenic
solutions with deionized water.
Adsorbent Preparation. The Fe-Mn binary oxide ad-
sorbent was prepared according to the method described in
detail in our previous publication (17). Fe oxide and MnO2
were synthesized by similar methods. XRD pattern of Fe-
Mn binary adsorbent shows that both the Fe oxide and Mn
oxide exist mainly in amorphous form.
Batch Adsorption Tests. Experiments to determine
isotherms were performed by adding 10 mg of the dry
adsorbent to a 150 mL glass vessel containing 50 mL of arsenic
solution. Initial arsenic concentrations were varied from 6.67
× 10-3 to 6.67 × 10-1 mM. As(III) and As(V) adsorption
isotherms were determined at pH of 6.9 and 4.8, respectively.
pH 6.9 was controlled by adding 0.1 M NaOH or HCl while
pH 4.8 was maintained using a buffer of 5 mM sodium acetate.
Ionic strength was adjusted to 0.01 M with 1 M NaNO3
solution. The supernatant was filtered through a 0.45 µm
membrane after the solutions were mixed for 24 h.
To determine the respective role of manganese oxide and
iron oxide content on arsenic removal and elucidate the
mechanism of As(III) removal, a control experiment was
carried out with an adsorbent content of 200 mg L-1 and pH
4.8 (5 mM sodium acetate buffer). Our primary study
indicated that Mn oxide was reduced by sodium sulfite while
Fe oxide could not under the employed conditions. Sodium
sulfite was therefore used to react with Fe-Mn binary oxide
to lower its Mn oxide content and oxidizing capacity. The
Fe-Mn binary oxide was first reduced by reacting with
different amounts (from 0 to 10 mM) of sodium sulfite
(Na2SO3) for 24 h, and then As(V) or As(III) (initial concen-
tration 0.20 mmol L-1) was introduced into this system and
mixed for another 24 h. The supernatant was filtered through
a 0.45 µm membrane and total arsenic in the filtered solution
was determined using an ICP-OES. All batch experiments
were carried out at 25 ( 1 °C and all the suspensions were
shaken on an orbit shaker at 140 rpm. The quantity of
adsorbed arsenic was calculated by the difference of the initial
and residual amounts of arsenic in solution divided by the
weight of the adsorbent. The control experiments were
repeated three times.
Analytical Methods. Total arsenic (As(III) + As(V))
concentrations were determined using an inductively coupled
plasma atomic emission spectroscopy (ICP-OES) (SCIEX
Perkin-Elmer Elan mode 5000). Prior to the analysis, the
aqueous samples were acidified with concentrated HCl, and
stored in acid-washed glass vessels. Selective As(III) analysis
was performed using hydride generation-atomic fluorescence
spectroscopy (HG-AFS) with an AF-610A instrument (Beijing
Ruili Analytical Instrument Co., Ltd. China) (18). A 5 mL
portion of filtered supernatant was transferred into a 10 mL
test tube and 2 mL of 0.5 mol L-1 citric acid solution was
added. The solution was diluted to 10 mL with deionized
water (solution pH 3.1) and then was analyzed. Citric acid
(0.1 mol L-1) was used as carrying fluid. All samples were
analyzed within 24 h of collection.
Characterization. To determine the arsenic species
adsorbed on the surface of the adsorbent after reaction with
As(V) or As(III), some selected samples were freeze-dried for
further analysis using FTIR and XPS. The change of oxidation
state of surface elements was simultaneously investigated.
To exclude the interference of sodium acetate in the IR and
XPS spectra, 0.1 M NaOH and HCl were used to control the
suspensions pH value of 4.8, instead of 5 mM sodium acetate
4614 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 13, 2007
buffer in the control experiment. In addition, the initial
arsenic concentration changed from 0.20 to 0.133 mmol L-1.
FTIR spectra were collected on a Nicolet 5700 spectrom-
eter using transmission mode. Samples for FTIR determi-
nation were ground with spectral grade KBr in an agate
mortar. To analyze Fe-OH quantitatively, a fixed amount of
sample (1% w:w) in KBr was used to prepare the pellet. All
IR measurements were carried out at room temperature.
X-ray photoelectron spectroscopy (XPS) data were col-
lected on an ESCA-Lab-220i-XL spectrometer with mono-
chromatic Al KR radiation (1486.6 eV). C1s peaks were used
as an inner standard calibration peak at 284.7 eV. XPS data
process and peak fitting was performed using a nonlinear
least-squares fitting program (XPSpeak software 4.1, Raymund
W. M. Kwork).
Results and Discussion
The determination of residual arsenic species in solution is
essential to investigate the oxidation of As(III) by the
adsorbent. This is also helpful for understanding the arsenic
species adsorbed on the adsorbent, which can be further
confirmed by FTIR and XPS analysis. Supernatant analysis
showed that arsenic species remaining in solution was As(V)
whether the initial arsenic was As(V) or As(III). When the
adsorbent was pretreated with 10 mM Na2SO3 solution and
then reacted with As(V) and As(III), however, residual arsenic
in the solution was As(V) and As(III), respectively. These
results showed that the Fe-Mn binary oxide could convert
As(III) to As(V) and Na2SO3-treatment significantly lowered
its oxidizing capacity.
Arsenic Adsorption Isotherms. The adsorption capacities
of the adsorbent for As(V) and As(III) were assessed using
the isotherms (Supporting Information (SI) Figure 2). The
adsorbent had high adsorption capacities toward both As(V)
and As(III) at two selected pH values. The maximum
adsorptive amount of As(V) at pH 4.8 was 0.96 mmol g-1.
This amount decreased to 0.72 mmol g-1 with pH increasing
to 6.9. This typical sorption behavior of As(V) is well
documented in the literature. The same trend was observed
with the initially added As(III) species. The maximum
adsorptive amount was 1.77 mmol g-1 at pH 4.8 and reduced
to 1.34 mmol g-1 at pH 6.9 and much higher than the results
obtained with As(V) at two selected pH values. However, for
pure iron systems, As(III) typically shows greater adsorption
with increased pH up to a maximum value at about pH 7.0.
So the conversion of initial As(III) to As(V) must occur during
the adsorption process. Previous study (18) also shows that
the adsorbent can effectively oxidize As(III) to As(V). The
higher maximal As(III) adsorption capacity than that of As-
(V) indicated that something else must take place on the
removal of As(III), besides only oxidation reaction. Production
of fresh adsorption sites at the solid surface during As(III)
oxidation may be responsible for this.
Effect of Na2SO3-Treatment on Arsenic Removal and
Mn2+ Release. No significant soluble iron was observed in
solution after the adsorbent reacted with As(III). The same
result was obtained when pure iron oxide reacted with As-
(III). The conversion of As(III) to As(V) by the adsorbent in
the solution can therefore be attributed to the presence of
manganese dioxide. If the oxidizing capacity of adsorbent
was lowered with pretreatment by Na2SO3 which can
reductively dissolve the Mn oxide, the arsenic uptake and
Mn2+ release would change.
The redox reaction between the adsorbent and Na2SO3 is
similar to that of As(III). The overall reaction can be
represented by eq 4.
MnO2* + SO32- + 2H+ ) Mn2+ + SO42- + H2O (4)
where MnO2* represents the manganese dioxide in the
adsorbent.
FIGURE 1. FTIR spectra of (A) Fe oxide, Fe oxide + As(V), and Fe oxide + As(III); (B) Mn oxide, Mn oxide + As(V), and Mn oxide + As(III);
(C) Fe-Mn oxide, Fe-Mn oxide + As(V), and Fe-Mn oxide + As(III); (D) pretreated Fe-Mn oxide, pretreated Fe-Mn oxide + As(V),
and pretreated Fe-Mn oxide + As(III). Initial arsenic concentration was 0.133 mM and a solid concentration of 200 mg L-1 was used
for all samples. Solution pH value was 4.8 (maintained by intermittent addition of dilute HCl or NaOH).
The effects of Na2SO3-treatment on arsenic removal and ment. The results of both arsenic removal and Mn2+ release
Mn2+ release are shown in SI Figure 3. Removal of As(V) reveal that MnO2 in the adsorbent plays a key role in
increased with an increase in reduction extent of the enhancing As(III) removal.
adsorbent, whereas the removal of As(III) decreased as the Analysis of FTIR Spectra. FTIR spectra of original
Na2SO3-treatment increased. The increase in As(V) uptake adsorbent before and after reaction with arsenic are shown
would be expected given that more active adsorption sites in Figure 1C. Figure 1-D presents the FTIR spectra for the
were produced, due to the reductive dissolution of MnO2 in pretreated adsorbent and the pretreated then reacted with
the adsorbent. The decrease in As(III) removal could be also arsenic samples. The FTIR spectra of pure Fe(III) oxide and
explained. The Na2SO3-treatment lowered the oxidizing pure Mn oxide before and after reaction with arsenic were
capacity of the adsorbent and consequently significantly also collected and are demonstrated in Figure 1A and B,
inhibited the As(III) oxidization. As(III) removal is not as respectively. The band at 1625 cm-1 was assigned to the
effective as As(V) removal for iron oxides. As a result, the deformation of water molecules and indicated the presence
As(III) removal decreased, despite that more fresh reaction of physisorbed water on the oxides. The peak at 1384 cm-1
sites were created during the reductive dissolution of MnO2. was ascribed to the vibration of NO3- because sodium nitrate
Without the Na2SO3-pretreatment, the concentration of was used to adjust the solution ion strength.
Mn2+ in the solution was as low as about 1.0 mg L-1 (SI For the Fe oxide spectra, three peaks at 1125, 1050, and
Figure 3b) after the adsorbent reacted with As(V). The release 976 cm-1 correspond to the bending vibration of the hydroxyl
of soluble manganese might be ascribed to the reductive group (Fe-OH) (19). After reaction with As(V), the peaks at
dissolution with the possibility of ligand-assisted dissolution 1125 and 976 cm-1 disappeared and the peak at 1050 cm-1
of MnO2 since 5.0 mM sodium acetate was used as buffer. weakened greatly, while a new band, corresponding to As-O
However, the concentration of soluble manganese in the stretching vibration, appeared at 820 cm-1. This indicates
solution was as high as 7.2 mg L-1 after the adsorbent reacted that the As(V) is bound as a surface complex and not as a
with As(III) alone, mainly due to the reductive dissolution of precipitated solid phase. After reaction with As(III), the peaks
MnO2. Without readsorption, the released Mn2+ concentra- at 1125 and 976 cm-1 also disappeared and the peak at 1050
tion in the solution should be 11 mg L-1. This indicated that cm-1 slightly weakened. A new peak that appeared at 585
some Mn2+ was re-adsorbed onto the solid surface. With the cm-1 instead of 820 cm-1 may be attributed to the As-O
increase in Na2SO3-treatment, the Mn2+ concentration in vibration in As(III) species.
the solution increased. For the As(V) system, the increase in For the Mn oxide spectra, the weak absorbance at 1048
As(V) removal corresponded to that of Mn2+ concentration cm-1 may be assigned to vibration of the hydroxyl group. No
in the solution, indicating the formation of new adsorptive obvious change was observed for this peak after the adsorp-
sites with the Mn2+ release. However, for the As(III) system, tion with As(V), while a new weak peak appeared at around
the As(III) removal declined as the Mn2+ concentration 820 cm-1. The weak peak shows that As(V) binds weakly to
increased, corresponding to the increase in Na2SO3-treat- Mn oxide. The spectrum of Mn oxide after reaction with

VOL. 41, NO. 13, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4615
TABLE 1. Composition of Unreacted and Reacted Fe-Mn Binary Oxide Surface
Fe(t) Mn(t) O(t) As(t)
sample at.% at.% at.% at.%
Fe-Mn oxide 18.2 6.4 75.4 0.0
Fe-Mn oxide after reaction with As(V) 17.2 6.3 74.1 2.4
Fe-Mn oxide after reaction with As(III) 17.4 5.4 73.7 3.5
Fe-Mn oxide after pretreatment with Na2SO3 19.8 3.4 76.8 0.0
pretreated Fe-Mn oxide after reaction with As(V) 18.6 3.3 74.6 3.5
pretreated Fe-Mn oxide after reaction with As(III) 19.1 3.1 75.8 2.0

FIGURE 2. Fe2p core level photoelectron spectra of (a) Fe-Mn binary oxide, (b) Fe-Mn binary oxide after reaction with As(V) (Ci ) 0.133
mM); (c) Fe-Mn binary oxide after reaction with As(III) (Ci ) 0.133 mM); (d) Fe-Mn binary oxide after pretreatment with Na2SO3 (Ci )
10 mM); (e) Fe-Mn binary oxide after pretreatment as (d) and then reaction with As(V) (Ci ) 0.133 mM); (f) Fe-Mn binary oxide after
pretreatment as (d) and then reaction with As(III) (Ci ) 0.133 mM).

As(III) was quite similar to that of Mn oxide after reaction
with As(V), indicating that initial As(III) was oxidized to As-
(V).
For original adsorbent spectra, three peaks at 1130, 1048,
and 973 cm-1 are due primarily to Fe-OH vibration. The
occurrence of a peak at 1538 cm-1 may be attributed to the
interaction between Mn oxide and Fe oxide in the adsorbent
because this peak appeared in neither the spectra of Fe oxide
or the spectra of Mn oxide. After reaction with arsenate, the
peak at 1538 cm-1 did not change, while those three peaks
disappeared completely. Simultaneously, a new band, cor-
responding to As-O stretching vibration, appeared at 820
cm-1. After reaction with arsenite, the peak at 1538 cm-1
disappeared. This may be due to a surface alteration for the
occurrence of redox reaction between As(III) and Fe-Mn
binary oxide. The other changes were the same as that of
As(V), which suggested indirectly that the arsenic species
adsorbed onto the surface of the original adsorbent was As-
(V).
Compared to the original absorbent, FTIR spectra for
pretreated adsorbent did not significantly change expect that
the peak at 1538 cm-1 almost completely disappeared.
However, the spectra of pretreated adsorbent then reacted
with arsenic were greatly different from those of the original
after reaction with arsenic. The peaks at 1130, 1048, and 973
cm-1 did not disappear completely for samples reacted with
both As(V) and As(III). And the intensity of hydroxyl group
vibration for the sample reacted with As(III) was higher than
that of sample reacted with As(V). This suggested that more
hydroxyl groups were consumed after reaction with As(V)
than that of As(III). Furthermore, the As-O stretching
vibration peak shifted from 820 cm-1 to 811 cm-1 for the
sample reacted with As(III). This would indicate that the
arsenic adsorbed on the adsorbent contained As(III) species
since the sorptive band corresponding to As(III)-O vibration
located at 794 cm-1 (10). Furthermore, the intensity of As-O
stretching vibration at 820 cm-1 for sample reacted with As-
(V) increased compared with that of the original one and
contrary result was obtained with the sample reacted with
As(III). These results also indicated that the conversion of
As(III) to As(V) was responsible for the high As(III) removal.
Surface Analysis of the Fe-Mn Binary Oxides. The
surface compositions of the selected samples were deter-
mined by XPS and results were presented in Table 1. Mn
atom content did not change after reaction with As(V),
whereas it decreased a little after reaction with As(III),
resulting from the reductive dissolution of MnO2. After Na2-
SO3-treatment, the Mn atom content decreased obviously.
No significant change of the Mn atom content was found
after reaction with As(V) or As(III), indicating available MnO2
for As(III) oxidation had been depleted after pretreatment.
The XPS spectra of Fe2p are illustrated in Figure 2. The
binding energy of 711.0 eV and the peak shape indicate that
the oxidation state of Fe in the adsorbent was +III. It did not
change after reaction with arsenic or the Na2SO3-treatment.
The content of Fe atom on the surface did not change greatly
as demonstrated in Table 1. However, a decrease of Fe2p
spectra intensity was observed after reaction with As(V) or
As(III), indicating the occurrence of strong interactions
between As(V) and Fe atoms (19). A similar phenomenon

4616 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 13, 2007
FIGURE 3. Mn2p core level photoelectron spectra of (a) Fe-Mn binary oxide, (b) Fe-Mn binary oxide after reaction with As(V) (Ci )
0.133 mM); (c) Fe-Mn binary oxide after reaction with As(III) (Ci ) 0.133 mM); (d) Fe-Mn binary oxide after pretreatment with Na2SO3(Ci
) 10 mM); (e) Fe-Mn binary oxide after pretreatment as (d) and then reaction with As(V) (Ci ) 0.133 mM); (f) Fe-Mn binary oxide after
pretreatment as (d) and then reaction with As(III) (Ci ) 0.133 mM).

FIGURE 4. As3d core level of the Fe-Mn adsorbent after reaction with As(V) or As(III). (a) After reaction with As(V) (Ci ) 0.133 mM);
(b) after reaction with As(III) (Ci ) 0.133 mM); (c) same as (a) but with pretreatment using Na2SO3; (d) same as (b) but with pretreatment
using Na2SO3.
was also observed after the Na2SO3-treatment. No obvious respectively. The fit data and peak shape demonstrated that
change was observed for the Fe2p spectra intensity of no great change occurred in Mn2p spectrum after reaction
pretreated adsorbent after reaction with arsenic. with As(V), unlike that of Fe2p. This indicated that no strong
The Mn2p spectra of the adsorbent before and after interactions occurred between As(V) and Mn atoms. Nev-
reaction with As(V) or As(III) are shown in Figure 3a-c, ertheless, a little decrease in intensity (by 18%) and binding

VOL. 41, NO. 13, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4617
energy (from 642.6 to 642.5 eV) of Mn2p spectrum after
reaction with As(III) was observed. This could be explained
by increasing the proportion of reduced Mn species relative
to that of Mn(IV), which resulted only from reaction with
As(III). Nesbitt et al. analyzed thoroughly the Mn2p spectra
of synthetic birnessite before and after reaction with As(III).
They found that binding energies of Mn(IV), Mn(III), and
Mn(II) species were very close to each other and the binding
energy of Mn with low oxidization state was located on the
low-energy side (13). After pretreatment, the intensity of
Mn2p spectrum (Figure 3d) was sharply decreased (by 55%)
and the binding energy shifted to the low region (from 642.6
to 642.3 eV). This indicated that part of MnO2 was reductively
dissolved by Na2SO3 and the near surface was enriched in
reduced forms of Mn relative to untreated Fe-Mn oxide.
The Mn2p spectra of the samples pretreated then reacted
with As(V) or As(III) are shown in Figure 3e and f, respectively.
The relative intensity and shape of Mn2p spectra did not
change significantly after reaction with As(V), similar to that
of the untreated one after reaction with As(V). The binding
energy shifted from 642.6 eV down to 642.1 eV, compared
with that of the untreated one after reaction with As(III).
This also suggested more reduced Mn species adsorbed on
the surface of the adsorbent. It can be concluded that
manganese dioxide mainly responses for oxidizing As(III)
during As(III) removal.
Figure 4 exhibits the As3d core level of the untreated
adsorbent after reaction with As(V) and As(III) and the
pretreated adsorbent after reaction with As(V) and As(III).
There was no obvious difference between the As3d spectrum
of the adsorbent after reaction with As(V) (Figure 4a) and
that after reaction with As(III) (Figure 4b). This indicated
that the arsenic species adsorbed after reaction with As(III)
was As(V) and further confirmed that the adsorbent was
effective for As(III) oxidation. After Na2SO3-treatment, the
As3d spectrum of the adsorbent after reaction with As(V)
(Figure 4c) was almost the same as that without pretreatment
(Figure 4a) except for higher intensity, indicating more As(V)
was adsorbed on the pretreated adsorbent. However, the
As3d spectrum of the Fe-Mn oxide after reaction with As-
(III) (Figure 4d) was quite different from that of the untreated
one. Fit results (SI Table 1) show that A3d line can be fitted
with two components having binding energies at 44.34 and
45.55 eV, respectively. This indicates that both As(III) and
As(V) species are on the surface of the pretreated then reacted
with As(III) sample. Commonly, binding energy of As3d core
level for As(III) and As(V) in arsenic oxides are 44.3-44.5 and
45.2-45.6 eV, respectively (13, 20, 21). They can be slowly
shifted up to 44.6 ( 0.13 eV for As(III) and 46.0 ( 0.17 eV for
As(V) when arsenic anions or molecules are adsorbed onto
iron oxide (22).
From the above discussion, it may be concluded that As-
(III) removal mechanism by Fe-Mn binary oxide is an
oxidation coupled with sorption approach. The whole process
can be briefly represented by reaction eqs 5-7.
As(III)(aq) + (-SFe-Mn) f As(III)-SFe-Mn (5)
As(III)-SFe-Mn MnO2 + 2H + f
As(V)(aq) + Mn2+ + H2O (6)
As(V)(aq) + As(III)-SFe-Mn f As(V)-SFe-Mn + As(III)
(aq) (7)
where (-SFe-Mn) represents an adsorption site on the Fe-
Mn adsorbent surface. As(III)-SFe-Mn represents the As(III)
surface species and As(V)-SFe-Mn represents the As(V) surface
species. This oxidation and sorption mechanism could well
4618 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 13, 2007
explain the higher As(III) uptake and other experimental
results.
Acknowledgments
This work was supported by the Funds for Creative Research
Groups of China (Grant 50621804) and by National Natural
Science Foundation of China (Grant 20577063, 20577063).
Supporting Information Available
Additional figures are provided. This material is available
free of charge via the Internet at http://pubs.acs.org.
Literature Cited
(1) Hughes, M. F. Arsenic toxicity and potential mechanisms of
action. Toxicol. Lett. 2002, 133, 1-16.
(2) Ferguson, J. F.; Gavis, J. A review of the arsenic cycle in nature
waters. Water Res. 1972, 6, 1259-1274.
(3) Duel, L. E.; Swoboda, A. R. Arsenic toxicity to cotton and
soybeans. J. Environ. Qual. 1972, 1, 317-320.
(4) Jang, M.; Min, S. H.; Kim, T. H.; Park, J. K. Removal of arsenite
and arsenate using hydrous ferric oxide incorporated into
naturally occurring porous diatomite. Environ. Sci. Technol.
2006, 40, 1636-1643.
(5) Pierce, M. L.; Moore, C. B. Adsorption of arsenite and arsenate
on amorphous iron hydroxide. Water Res. 1982, 16, 1247-1253.
(6) Raven, K. P.; Jain, A.; Loeppert, R. H. Arsenite and arsenate
adsorption on ferrihydrite: kinetics, equilibrium, and adsorption
envelopes. Environ. Sci. Technol. 1998, 32, 344-349.
(7) Dixit, S.; Hering, J. G. Comparison of arsenic(V) and arsenic(III)
sorption onto iron oxide minerals: implications for arsenic
mobility. Environ. Sci. Technol. 2003, 37, 4182-4189.
(8) Waychunas, G. A.; Rea, B. A.; Fuller, C. C.; Davis, J. A. Surface
chemistry of ferrihydrite. 1. EXAFS studies of the geometry of
coprecipitated and adsorbed arsenate. Geochim. Cosmochim.
Acta 1993, 57, 2251-2269.
(9) Sun, X. H.; Doner, H. E. An investigation of arsenite and arsenite
bonding structures on goethite by FTIR. Soil. Sci. 1996, 161
(12), 865-872.
(10) Goldberg, S.; Johnston, C. T. Mechanisms of Arsenic Adsorption
on Amorphous Oxides Evaluated Using Macroscopic Measure-
ments, Vibrational Spectroscopy, and Surface Complexation
Modeling. J. Colloid Interface Sci. 2001, 234 (1), 204-216.
(11) Oscarson, D. W.; Huang, P. M.; Defosse, C.; Herbillon, A.
Oxidative power of Mn(IV) and Fe(III) oxides with respect to
As(III) in terrestrial and aquatic environments. Nature 1981,
291, 50-51.
(12) Scott, M. J.; Morgan, J. J. Reactions at oxide surfaces. 1. Oxidation
of As(III) by synthetic birnessite. Environ. Sci. Technol. 1995,
29, 1898-1905.
(13) Nesbitt, H. W.; Canning, G. W.; Bancroft, G. M. XPS study of
reductive dissolution of 7Å-birnessite by H3AsO3 with constraints
on reaction mechanism. Geochim. Cosmochim. Acta. 1998, 62
(12), 2097-2110.
(14) Manning, B. A.; Fendorf, S. E.; Bostick, B.; Suarez, D. L. Arsenic-
(III) oxidation and arsenic(V) adsorption reactions on synthetic
birnessite. Environ. Sci. Technol. 2002, 36, 976-981.
(15) Moore, J. N.; Walker, J. R.; Hayes, T. H. Reaction scheme for the
oxidation of As(III) to As(V) by birnessite. Clays Clay Miner.
1990, 38 (5), 549-555.
(16) Deschamps, E.; Ciminelli, V. S. T.; Höll, W. H. Removal of As(III)
and As(V) from water using a natural Fe and Mn enriched
sample. Water Res. 2005, 39 (20), 5212-5220.
(17) Zhang, G. S; Qu, J. H.; Liu, H. J.; Liu, R. P.; Wu, R. C. Preparation
and evaluation of a novel Fe-Mn binary oxide adsorbent for
effective arsenite removal. Water Res. 2007, 41 (9), 1921-1928.
(18) Shi, J. B.; Tang, Z. Y.; Jin, Z. X.; Chi, Q.; He, B.; Jiang, G. B.
Determination of As(III) and As(V) in soils using sequential
extraction combined with flow injection hydride generation
atomic fluorescence detection. Anal. Chim. Acta. 2003, 477,
139-147.
(19) Zhang, Y.; Yang, M.; Dou, X.-M.; He, H.; Wang, D.-S. Arsenate
Adsorption on an Fe-Ce Bimetal Oxide Adsorbent: Role of
Surface Properties. Environ. Sci. Technol. 2005, 39 (18), 7246-
7253.
(20) Fullston, D.; Fornasiero, D.; Ralston, J. Oxidation of synthetic
and natural samples of enargite and tennantite. 2. X-ray
photoelectron spectroscopic study. Langmuir 1999, 15, 4530-
4536.
(21) Ouvrard, S.; Donato, P.; Simonnot, M.O.; Begin, S.; Ghanbaja,
J.; Alnot, M.; Duval, Y. B.; Lhote, F.; Barres, O.; Sardin, M. Natural
manganese oxide: Combined analytical approach for solid
characterization and arsenic retention, Geochim. Cosmochim.
Acta. 2005, 69, 2715-2724.
(22) Ding, M.; De Jong, B. H. W. S.; Roosendall, S. J.; Vredenberg, A.
XPS studies on the electronic structure of bonding between
solid and solutes. Adsorption of arsenate, chromate, phosphate,
Pb2+ and Zn2+ ions on amorphous black ferric oxyhydroxide.
Geochim. Cosmochim. Acta. 2000, 64, 1209-1219.
Received for review December 19, 2006. Revised manuscript
received April 13, 2007. Accepted April 19, 2007.
ES063010U
VOL. 41, NO. 13, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4619