Rapid, photocatalytic, and deep debromination of polybrominated diphenyl ethers on Pd-TiO2: Intermediates and pathways

Article abstract of DOI:10.1002/chem.201402477

Titanium dioxide with surface-loaded palladium (Pd-TiO₂) was able to easily remove all ten bromine atoms from decabromodiphenyl ether (BDE209) within 1 h under the irradiation of sunlight or an artificial light source. By contrast, fewer than three bromine atoms were eliminated on the pristine TiO₂ even with prolonged irradiation (5 h). During the photocatalytic debromination, moreover, the formed BDE intermediates exhibited a significant difference between the Pd-TiO₂ and pristine TiO ₂ systems, and much less position selectivity for the debromination on Pd-TiO₂ was observed than that on the pristine TiO₂ surface. For another polybrominated diphenyl ether (BDE15), pristine TiO ₂ was incapable of its photocatalytic reduction, whereas the loading of Pd enabled its debromination to diphenyl ether within 20 min. In addition, an evident induction period appeared in the photocatalytic debromination of BDE15 on Pd-TiO₂. The experiments imply that the Pd-cocatalyzed effect changes significantly the photocatalytic reductive debromination pathways.

Full text of DOI:10.1002/chem.201402477

DOI: 10.1002/chem.201402477
Full Paper
&
Photocatalysis
Rapid, Photocatalytic, and Deep Debromination of
Polybrominated Diphenyl Ethers on Pd–TiO2: Intermediates and
Pathways
Lina Li, Wei Chang, Ying Wang, Hongwei Ji, Chuncheng Chen,* Wanhong Ma, and
Jincai Zhao[a]
Abstract: Titanium dioxide with surface-loaded palladium
(Pd–TiO2) was able to easily remove all ten bromine atoms
from decabromodiphenyl ether (BDE209) within 1 h under
the irradiation of sunlight or an artificial light source. By con-
trast, fewer than three bromine atoms were eliminated on
the pristine TiO2 even with prolonged irradiation (5 h).
During the photocatalytic debromination, moreover, the
formed BDE intermediates exhibited a significant difference
between the Pd–TiO2 and pristine TiO2 systems, and much
less position selectivity for the debromination on Pd–TiO2
was observed than that on the pristine TiO2 surface. For an-
other polybrominated diphenyl ether (BDE15), pristine TiO2
was incapable of its photocatalytic reduction, whereas the
loading of Pd enabled its debromination to diphenyl ether
within 20 min. In addition, an evident induction period ap-
peared in the photocatalytic debromination of BDE15 on
Pd–TiO2. The experiments imply that the Pd-cocatalyzed
effect changes significantly the photocatalytic reductive de-
bromination pathways.
Introduction
whereas the most-used deca-BDE has been detected with
much less frequency. Although these moderately brominated
congeners, as important components of historically used flame
retardants with penta-BDE formulations, can be directly re-
leased into the environment from the products, the most im-
portant reason for the accumulation is their resistance to deg-
radation. In nature, UV-induced hydrodebromination has been
identified as the major abiotic transformation pathway of
highly brominated PBDEs.[11–16] Previous reports showed that
the degradation of PBDEs by UV light was rather ineffective to
the low-brominated diphenyl ethers, which led to the accumu-
lation of hexa-, penta-, and tetra-BDEs.[11,12] The biotic reductive
debromination by microbes or fish, which represents another
transformation pathway of PBDEs in the environment, still fails
to completely transform BDE209 to nonbrominated com-
pounds, even at prolonged incubation periods (typically sever-
al months).[17–19] Accordingly, the accumulation of these moder-
ately brominated congeners in nature may result from the de-
bromination of higher-brominated congeners, but they them-
selves are more recalcitrant to further debromination. To make
matters worse, these moderately brominated PBDEs (e.g., con-
geners BDE47, 99, and 100) are known to be the most toxic
and bioaccumulative congeners.[20,21] Their existence in the en-
vironment would cause a more serious risk to human health.
The general resistance to degradation and the potential tox-
icity of PBDEs also triggered substantial investigations on the
development of efficient methods and techniques to properly
treat this type of compound. Because PBDEs are difficult to oxi-
datively degrade, most of their decontamination methods are
based on reductive debromination. Although these reduction
Polybrominated diphenyl ethers (PBDEs) are widely applied in
electronics, petroleum products, and textiles because of their
stability and efficient flame-retardant property. These PBDEs
can migrate from the products and release to the environment
continuously. As a result, congeners of PBDEs have been de-
tected globally in food, humans, wildlife, and environmental
media such as sediments, sewage sludge, and air,[1–6] and the
detected concentrations appear to increase rapidly over
time.[7,8] Recent studies on animals showed that some conge-
ners of PBDEs can disrupt the endocrine system, threaten re-
productive health, and impair the development of the immune
system in animal bodies.[9,10]
The penta-, octa-, and deca-BDE products are the common
commercially available PBDEs, among which deca-BDE
(BDE209), the fully brominated congener, is in greatest
demand. In environmental media and biological systems, the
most frequently detected PBDEs are the congeners with mod-
erate bromination, such as tri-, tetra-, penta-, and hexa-BDEs,
[a] Dr. L. Li, Dr. W. Chang, Dr. Y. Wang, Prof. H. Ji, Prof. C. Chen, Prof. W. Ma,
Prof. J. Zhao
Key Laboratory of Photochemistry
Beijing National Laboratory for Molecular Sciences
Institute of Chemistry, Chinese Academy of Sciences
Beijing 100190 (P. R. China)
Fax: (+86)82616495
E-mail: ccchen@iccas.ac.cn
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201402477.
Chem. Eur. J. 2014, 20, 11163 – 11170
11163
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
technologies are effective for the decomposition of highly bro-
minated PBDEs such as BDE209, the debromination reaction
becomes rather difficult for the low-brominated intermediates.
For example, the reductive debromination of PBDEs by widely
studied zerovalent iron (ZVI) particles was ineffective in reduc-
ing the low-brominated congeners of PBDEs, which led to the
accumulation of a lot of PBDEs with three to six bromides in
the systems at the end of reaction (weeks to months).[22,23] For
these reasons, it is urgent to develop effective remediation
technologies to achieve the deep debromination of the PBDEs.
In our previous study, we reported that TiO2-mediated pho-
tocatalysis can be used to achieve rapid reductive debromina-
tion of BDE209, in which nearly all the original BDE209 could
be degraded within 8 min of UV irradiation. However, this
system still suffered from the problem of ineffectiveness for
the low-brominated PBDEs. After prolonged irradiation (24 h),
the BDE209 was observed to be reduced only to hepta- and
hexa-BDEs, with a small amount of penta- and tetra-BDEs.[14]
Herein, we report that BDE209 (10 mm) can be debrominated
rapidly and completely by surface-palladized TiO2 (Pd–TiO2)
photocatalyst (0.2 gLÀ1) within 1 h under UV irradiation. In ad-
dition, the low-debrominated PBDE (BDE15), which is resistant
to photocatalytic debromination on the pristine TiO2, also un-
dergoes a rapid reaction in the Pd–TiO2 system. By a detailed
comparison of the photocatalytic degradation kinetics of
PBDEs and the reaction intermediates with those in the pris-
tine TiO2 system, the mechanism of photocatalytic debromina-
tion of PBDEs on Pd–TiO2 is further discussed. The present
study provides an efficient method for the rapid full debromi-
nation of PBDEs and an insightful understanding of the TiO2 or
Pd–TiO2 photocatalyzed debromination reaction mechanism,
which may extend to other photocatalytic dehalogenation re-
actions of halogenated pollutants.
Figure 1. TEM images of a) Pd–TiO2 (2 wt% Pd) and b) TiO2 particles;
c) HRTEM image of Pd–TiO2 particles; d) X-ray photoelectron spectrum of
the prepared Pd–TiO2 nanoparticles at the Pd3d region. CPS=counts per
second.
from Pd3d5/2 and Pd3d3/2, respectively, which are consistent
with the binding energy of metallic Pd0.[24,25] As shown in Fig-
ure S3 in the Supporting Information, besides the diffraction
peaks of anatase and rutile phases, the Pd(111) and Pd(200) dif-
fraction peaks with their respective 2q=40.2 and 46.88 were
clearly observed (JCPDS 87-0638).[26–28] An inductively coupled
plasma–atomic emission spectroscopy measurement indicates
that the actual loading of palladium on TiO2 is 1.37 wt%,
which is less than the nominal content (2 wt%), probably be-
cause of its loss during the loading process.
Results and Discussion
Catalyst characterization
Reductive debromination of BDE209 on Pd–TiO2
After loading of palladium, the color of TiO2 turned from milky
white to gray, and Pd–TiO2 exhibited a strong absorption in
the visible-light region (Figure S1 in the Supporting Informa-
tion) accordingly, which suggested that Pd should be in its
metallic state. Figure 1a and b show the TEM images of Pd–
TiO2 (2 wt% Pd) and the pristine TiO2, respectively. It was
found that, in comparison with the clean features of the TEM
image of pristine TiO2 (Figure 1b), the TEM image of the as-
prepared Pd–TiO2 catalyst was clearly covered by dispersed
spots (Figure 1a). As indicated by high-resolution TEM (HRTEM,
Figure 1c), these spots had a lattice spacing of 0.19 nm, corre-
sponding to the (200) facet of metallic Pd and confirming that
these spots were well-defined palladium particles. The HRTEM
image also showed that the palladium particles have a diame-
ter of about 7 nm.
A mercury lamp was chosen as the light source. To eliminate
the direct photolysis of BDE209, the irradiation below the
wavelength of 360 nm was filtered off. Figure 2 shows the
temporal concentration changes (c/c0) of BDE209 under differ-
ent conditions. Although, as reported in our previous study,[14]
BDE209 undergoes efficient degradation in anoxic methanol
suspensions of TiO2 under UV irradiation (Figure 2, curve c), its
degradation was much more rapid on the Pd–TiO2 surface
(Figure 2, curve d). BDE209 disappeared completely after only
4 min of irradiation on the Pd–TiO2, whereas only about 40%
of BDE209 was degraded in the pristine TiO2 system at the
same irradiation time, thus indicating a much superior reactivi-
ty of the Pd–TiO2 toward the debromination of PBDE. It is
known that the supported Pd particles can catalytically activate
donors of hydrogen/electron (such as H2) to reduce many halo-
genated organic compounds.[29,30] To exclude the possibility
that the enhancing effect of Pd–TiO2 is caused by such thermal
catalytic reactions, the concentration change of BDE209 was
Furthermore, the existence of palladium particles was also
confirmed by X-ray photoelectron spectroscopy (XPS; Fig-
ure 1d and Figure S2 in the Supporting Information). The
Pd3d spectrum in XPS shows two peaks at 334.4 and 349.5 eV
Chem. Eur. J. 2014, 20, 11163 – 11170
www.chemeurj.org
11164
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
was also negligible (Figure 2, curve b). This result implies that,
compared to O2, BDE209 is less competitive for the photogen-
erated conduction band electron (ecbÀ), despite the catalytic
properties of Pd for the reduction of BDE209. It should be
pointed out that the ineffectiveness of debromination under
air-saturated conditions may limit the practical application of
Pd–TiO2 in many environmental systems. Nevertheless, besides
its fundamental interest, photocatalytic debromination on Pd–
TiO2 can be used to treat wastewater with a high concentra-
tion of PBDEs before its discharge.
The intermediates during the degradation of BDE209 in TiO2
and Pd–TiO2 systems were analyzed by gas chromatography
with micro electron capture detection (GC-mECD). As shown in
Figure 3, in the pristine TiO2 system the nona-BDE congeners
were observed as the dominant degradation intermediates of
BDE209 (with minor octa-BDEs) after irradiation for 10 min. Fur-
ther irradiation led to the gradual decrease of the nona-BDEs,
and concomitantly the accumulation of octa-BDEs. After 2 h,
octa-BDEs became the only dominant products. However, fur-
ther transformation of octa-BDEs was rather slow. After up to
5 h of irradiation, very few hepta-BDEs were observed (Fig-
ure 3a and c), which is in line with the general trend that the
less brominated BDEs are much more resistant to further re-
ductive debromination than BDE209. In contrast, in the
BDE209/Pd–TiO2 suspension, nona-, octa-, and hepta-BDEs
Figure 2. Temporal curves of the photocatalytic degradation of BDE209 in
25 mL methanol solution under different conditions. TiO2: in a suspension of
pristine TiO2; Pd–TiO2: in a suspension of surface-palladized TiO2; UV: under
irradiation at wavelength >360 nm; Ar: purged with argon; air: air-saturat-
ed; BDE209: 10 mmolLÀ1; photocatalysts: 0.2 gLÀ1
.
examined by keeping the BDE209/Pd–TiO2 suspensions in the
dark. No degradation of BDE209 in this control experiment
(Figure 2, curve a) indicates that the enhancing effect of Pd is
not attributable to the Pd-catalyzed thermal reaction between
BDE209 and methanol. In another control experiment, the pho-
tocatalytic reaction was performed in an air-saturated Pd–TiO2
suspension under UV irradiation. The degradation of BDE209
Figure 3. GC-mECD chromatograms and relative changes in distribution of photodegradation products of BDE209 in a,c) 0.2 gLÀ1 TiO2 and b,d) 0.2 gLÀ1 Pd–
TiO2 suspensions at different irradiation times. For comparison, the chromatogram of the standard sample of PBDEs (EO5113) under identical conditions is
also given in panels (a) and (b) (bottom curves).
Chem. Eur. J. 2014, 20, 11163 – 11170
www.chemeurj.org
11165
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
were all detected in considerable amounts within only 2 min
of irradiation. From 2Br to 9Br, BDEs with the dominance of
5Br to 8Br congeners were observed after 4 min, and further ir-
radiation shifted the main intermediates to the 2Br–4Br BDEs.
After irradiation for 1 h, no brominated intermediates were de-
tected in the gas chromatogram (Figure 3b and d), which indi-
cated the full debromination of PBDEs. Meanwhile, a measure-
ment of BrÀ concentration by ion chromatography (Figure S4
in the Supporting Information) showed that the bromine
atoms in BDE209 (100 mm) were released nearly stoichiometri-
cally in the form of inorganic BrÀ at the end of the reaction in
the Pd–TiO2 system after irradiation for 60 min.
mation). This difference in light intensity can explain the differ-
ent reduction rate of BDE209 under artificial mercury lamp and
sunlight irradiation.
The bioaccumulation and toxicity of PBDEs were known to
depend greatly on both the number of bromine atoms and
their substituted positions. It is therefore necessary to identify
the intermediates and the products of BDE209 in a more de-
tailed manner during its photocatalytic degradation. During
the debromination of BDE209, the loss of the first bromine
atom may lead to three nona-BDE congeners (BDE206,
BDE207, and BDE208, corresponding to removing one bromine
atom from the ortho, meta, and para positions of BDE209, re-
spectively). In the present study, all three nona-BDE congeners
were detected in considerable amounts at the beginning of ir-
radiation in both the TiO2 and Pd–TiO2 systems. The relative
distributions (ratios of GC peak areas) of these three nona-
BDEs remained quite stable along with the reaction time in
each system (Figure 4, Figure S8 in the Supporting Informa-
To detect intermediates other than PBDEs, the supernatants
at a given irradiation time were concentrated to 1/20 of the
original volume by freeze-drying, and analyzed by HPLC. More
than a dozen non-PBDE intermediates were observed during
the photocatalytic reactions in the Pd–TiO2 system (Figure S5a
in the Supporting Information). The fully debrominated prod-
uct diphenyl ether (DE) was detected as the main product at
the late stage of irradiation (Figure S5a in the Supporting Infor-
mation). After irradiation for 60 min, only 15.7% of the DE
yield relative to BDE209 was monitored. By comparing with
the standard samples, another two non-PBDE intermediates
were also identified to be 2,4-dibromophenol and pentabro-
mophenol (PBP). These phenolic intermediates originating
from BDE-209 are derived from the cleavage of the CÀOÀC
ester bond, probably due to the Pd-catalyzed hydrogenation
of the oxygen. In addition, PBP was observed as a main inter-
mediate at an early stage (10 min) but disappeared with fur-
ther irradiation. These results indicate that the cleavage of the
ester bond begins to occur at an early irradiation time and the
bromophenols can be further debrominated on the Pd–TiO2
surface. On the pristine TiO2 surface, however, few intermedi-
ates were detected at the same HPLC region during the whole
irradiation time up to 300 min (Figure S5b in the Supporting
Information), thus indicating that the cleavage of the ester
bond of PBDEs does not occur on the pristine surface.
Practically, natural solar light is much more desirable for irra-
diation than an artificial light source. Hence, the photocatalytic
debromination of BDE209 was also examined under the irradia-
tion of outdoor sunlight. It was observed that, under sunlight,
the surface palladization of TiO2 exhibited similar effects on
the degradation rates and intermediate distribution to those
under the artificial light source (Figure S6 in the Supporting In-
formation). In the Pd-loaded system, BDE209 was almost de-
graded after sunlight irradiation for 30 min, and 5Br- to 8Br-
BDEs were all detected as primary intermediates. Sunlight ex-
posure for 1 h transformed the BDE209 to its congeners with
fewer than 3Br. However, the octa-BDEs were still the domi-
nant intermediates in the pristine TiO2 systems even with sun-
light exposure for 5 h. We also noted that the degradation rate
and emergence of intermediates under sunlight irradiation
were somewhat slower than that in the artificial mercury lamp
cases. At 365 nm wavelength, the mercury lamp has a constant
light intensity of about 130 mWcmÀ2, whereas the intensity of
sunlight changes against time, with the maximum about
92 mWcmÀ2 at noon (12:00) (Figure S7 in the Supporting Infor-
Figure 4. Comparison of the abundance of the nona- (a) and octa-BDE (b) in-
termediates during the photocatalytic debromination of BDE209 on the pris-
tine and the palladized TiO2.
tion), but were quite different between the two systems. In the
pristine TiO2 system, the relative peak areas were in the order
of BDE206>BDE207>BDE208, indicative of the preference for
ortho debromination. However, the peak area of BDE207 was
larger than those of BDE206 and BDE208 in the Pd–TiO2 disper-
sions, which indicated that the meta-Br is more sensitive to Pd-
cocatalyzed debromination. The sensitivity of the meta-Br was
also observed in the PBDE reduction by nZVI.[31]
Chem. Eur. J. 2014, 20, 11163 – 11170
www.chemeurj.org
11166
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
In both the pristine and palladized TiO2 catalytic systems, all
three nona-BDE intermediates were found to further undergo
reductive debromination, which was concomitant with the
dominant accumulation of the octa-BDEs (Figure 3). There are
12 potential congeners for the octa-BDEs. Because not all of
these octa-BDE standards are available, some intermediates
had to be assigned according to the relative retention times
from the literature.[32] As shown in Figure S9 in the Supporting
Information, in the pristine TiO2 systems, seven peaks of octa-
BDEs were observed after 10 min of irradiation. With well-iso-
lated GC peaks, the congeners of BDE195 (2,3,4,4’,5,5’,6,6’-
BDE), BDE194 (3,3’,4,4’,5,5’,6,6’-BDE), BDE196 (2,3’,4,4’,5,5’,6,6’-
BDE), BDE197 (2,2’,4,4’,5,5’,6,6’-BDE), BDE204 (2,2’,3,4,4’,5,6,6’-
BDE), and BDE201 (2,2’,3,4’,5,5’,6,6’-BDE) were confidently iden-
tified. Two congeners of BDE198 (2,3,3’,4,5,5’,6,6’-BDE) and
BDE199 (2,3,3’,4’,5,5’,6,6’-BDE) with retention time at 24.44 min,
which were generated from the debromination of ortho- and
meta-Br from one ring and from two rings, respectively, were
actually inseparable under GC conditions.[32] Hence, at most
eight congeners of octa-BDEs might be detected throughout
the photoreaction on pristine TiO2. Further irradiation for
60 min led to the disappearance of BDE195, which indicated
that it can be further debrominated, whereas another six
peaks remained and their relative intensities did not change
much during all the testing time (300 min). A common charac-
teristic for the detected 8Br products, except BDE204, in the
pristine TiO2 system is that the two bromine atoms are re-
moved from the different benzene rings of BDE209. To satisfy
this rule, it is plausible that the BDE199, formed by the loss of
two Br atoms of BDE209 from different rings, should be domi-
nant in the peak at 24.44 min. On the Pd–TiO2, however, many
more peaks in the octa-BDE region were detected. The most
remarkable difference is the appearance of a strong peak with
retention time of 24.51 min, which is absent in the pristine
TiO2 systems. This peak was assigned to BDE200
(2,3,4,4’,5,5’,6,6’-BDE) and/or BDE203 (2,3,3’,4,4’,5,6,6’-BDE),
which are inseparable under the GC conditions. It is notable
that these two congeners are formed by debromination of two
Br atoms on the same aromatic ring of BDE209. Two minor in-
termediates with retention times of 23.62 and 25.04 min were
also observed on Pd–TiO2, and assigned to BDE202
(2,2’,3,3’,5,5’,6,6’-BDE) and BDE205 (2,3,3’,4,4’,5,5’,6-BDE). Thus,
all the 12 congeners of octa-BDEs during the BDE209 reduction
on Pd–TiO2 can be detected, which indicates that the bromides
at all the positions of PBDEs could be removed. However, the
formation of only partial 8Br intermediates (seven of 12) on
the pristine TiO2 means that the debromination reaction is dif-
ficult for the Br atom in some positions of PBDEs. In this sense,
the photocatalytic debromination reaction is less selective on
Pd–TiO2 than on the pristine TiO2 surface. Evidently, the lack of
position selectivity on the Pd–TiO2 is attributable to the cocata-
lytic effect of Pd, which makes the cleavage of the CÀBr bond
at every position of the PBDEs occur at comparable rates, in
agreement with the rapid debromination of low PBDEs on Pd–
TiO2 (see below). Moreover, on the Pd–TiO2, the octa-BDEs
formed by loss of one or both the meta-Br atoms (seven in
total, i.e., BDE195, 196, 197, 200, 201, 203, and 204) generally
appeared as main intermediates, whereas those in which all
the meta-Br are preserved (five in total, i.e., BDE194, 198, 199,
202, and 205,) were always detected in minor amounts (Fig-
ure 4b). This is consistent with the observation in Figure 4a
that the meta-Br is preferential to leave on Pd–TiO2.
The further debromination of the octa-BDE intermediates on
the pristine TiO2 surface was rather slow. Only a small amount
of 7Br intermediates appeared after irradiation for 5 h. In the
Pd–TiO2 system, in contrast, plenty of intermediates with fewer
than 8Br were detected. However, only a few of them were
identified by EO5113 standard solution (with 39 known PBDE
congeners), including BDE183, 154,138, 116, and 33/28
(Figure 5). Most of the main intermediates were not among
the 39 congeners of EO5113. Significantly, BDE47, which is the
Figure 5. Formation of the photodegradation intermediates of BDE209 in
Pd–TiO2 dispersions at different irradiation times. The chromatogram of the
standard sample of PBDEs (EO5113) is displayed in the top panel for com-
parison. The corresponding peaks of the solid lines are assigned to BDE33,
BDE28, BDE47, BDE116, BDE154, and BDE183, respectively, but the peaks on
the dotted lines are not identified.
most commonly detected PBDE in the environment and in
samples collected from organisms,[33,34] was hardly detected
throughout the whole debromination process.
The accumulation of octa-BDEs on pristine TiO2 during the
photocatalytic degradation of BDE209 demonstrates its ineffec-
tiveness on the low-brominated PBDEs. The remarkable charac-
teristic of the photocatalytic reductive debromination on the
Pd–TiO2 system is the emergence of the deep-debrominated
congeners even at the very early stage of the reaction, which
implies that the Pd–TiO2 is able to efficiently debrominate the
low-bromo DEs. This assertion is in agreement with the obser-
vation in the ZVI–Pd system in which the less brominated
BDEs exhibited a more significant enhancement in the reduc-
tion rates by palladization.[35] To further support this perspec-
tive, the photocatalytic reduction of a lower-bromo DE, 4,4’-di-
bromodiphenyl ether (BDE15), was examined. As shown in
Figure 6, the photocatalytic reduction reaction for BDE15 was
not detected under UV irradiation on the pristine TiO2. Howev-
er, in the Pd–TiO2 photocatalytic system, BDE15 was degraded
quickly within 20 min of irradiation, and concomitantly the 4-
bromodiphenyl ether intermediate was formed. Further irradia-
tion resulted in the disappearance of 4-bromodiphenyl ether
and the accumulation of diphenyl ether. The final yield of di-
Chem. Eur. J. 2014, 20, 11163 – 11170
www.chemeurj.org
11167
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
on the TiO2 surface, the photogenerated conduction-band
electron flows from the semiconductor to the Pd particle,
which would lead to the formation of a Schottky junction at
the metal/semiconductor interface. The electron transfer will
continue until the Fermi levels of metal and semiconductor are
aligned. The enhancement of BDE209 debromination by sur-
À
face palladization suggests that the ecb can be trapped on Pd
at the initial irradiation, and the reduction reaction takes place
on the Pd particles. However, the reduction of BDE15 does not
occur at this stage, evidently due to the low Fermi level of the
photocatalyst. With accumulation of ecbÀ, the reduction reac-
tion of BDE15 gradually becomes possible because of the in-
creased Fermi level. Accordingly, the presence of an induction
period for the photocatalytic reduction of BDE15 implies that
the electron transfer from TiO2 to palladium may play an es-
sential role in the Pd-cocatalyzed debromination on the TiO2
surface. On the other hand, the debromination of BDE15 did
not occur on the pristine TiO2 systems even at the prolonged
irradiation time, at which an even higher Fermi level of TiO2 is
expected, thus suggesting that the catalytic function of Pd is
also indispensable to the debromination of BDE15. It has been
proposed that activated H species on the Pd surface are re-
sponsible for the accelerated rates of PBDE debromination
through nucleophilic attack at a halogen substituent.[22,38] If
this is the case in our study, the reduction of both BDE15 and
BDE209 would be enhanced at the initial irradiation, since the
activated H species, which accelerate the debromination of
BDE209, should be equally effective for the reduction of
BDE15. The different reaction behaviors of the BDE209 and
BDE15 on Pd–TiO2 imply that this reaction is not in the rate-de-
termining step. CÀX bond homolytic cleavage on the Pd cata-
lyst is also reported to be the rate-limiting step in the Pd-cata-
lyzed dehalogenation,[39] which means that the cleavage of the
CÀX bond takes place prior to the reduction step. The absence
of the dark reaction even for unstable BDE209 in the Pd–TiO2
systems does not seem to support this possibility. More rea-
sonably, both electron transfer and CÀBr cleavage are in the
rate-determining step for the debromination on Pd–TiO2, in
which the electron transfer and CÀBr cleavage proceed simul-
taneously, whereas the electron-transfer process should be
rate-determining on the pristine TiO2, as our previous reports
mentioned.[14] According to this mechanism, the cocatalytic
Figure 6. Temporal curves of the photodegradation of BDE15 and the forma-
tion of 4-dibromodiphenyl ethers (BDE3) and diphenyl ethers in 25 mL solu-
tion under different conditions. BDE15: 10 mmolLÀ1; UV: under irradiation of
wavelength >360 nm; Ar: purged with argon; air: air-saturated; Pd–TiO2,
TiO2: 0.2 gLÀ1
.
phenyl ether was about 60% of the parent BDE15. Intriguingly,
unlike the reaction of BDE209 on Pd–TiO2, in which the debro-
mination of the substrate decreased rapidly at the early stages
of the irradiation (Figure 2), the photocatalytic reduction of
BDE15 exhibited an evident induction period of 10 min. In this
induction period, the degradation of BDE15 was quite slow,
and after that the reaction was markedly accelerated
(Figure 6).
The presence of an induction period may have much mech-
anistic implication for the Pd-cocatalyzed debromination of
BDE15 to a large extent. In the earlier studies on the noncata-
lytic debromination of PBDEs, a good correlation between the
reduction rates and the heat of formation (Hf, which reflects
the thermodynamic chemical stability) and the energy level for
the lowest unoccupied molecular orbital (ELUMO, which is direct-
ly related to the electron affinity) of the PBDEs was usually ob-
served.[22,35,36] In the present study, the rapid reduction of
BDE209 but the ineffective reduction of low-brominated BDEs
on the pristine TiO2 is evidently attributable to their different
chemical stabilities and electron affinities, since BDE209 has
a much lower stability (Hf =414.2 kJmolÀ1) and much higher
electron affinity (ELUMO =À1.343 eV) than those of BDE15 (Hf =
139.8 kJmolÀ1 and ELUMO =À0.403 eV).[37] In addition, the obser-
vation that most of the 8Br products were formed by debromi-
nation from different benzene rings of BDE209 (Figure 4b) on
pristine TiO2 suggests that, once one Br atom is removed from
one aromatic ring, the second debromination reaction be-
comes difficult on this ring. Rather, the second Br atom prefers
to leave from the untouched rings with five Br atoms. This can
rationalize the dominant accumulation of 8Br congeners, and
also implies that the photocatalytic debromination reaction on
pristine TiO2 might proceed by an electron transfer followed
by CÀBr cleavage, as proposed in our earlier study.[14]
À
effect of Pd is attributable to two factors: Pd can trap the ecb
to enrich electrons on its surface and enhance the photogener-
ated charge separation, and further transfer the electron to
PBDEs by assisting the cleavage of the CÀBr bond, probably
owing to the high affinity of Pd metal for bromine atoms. Such
a concerted electron transfer and CÀBr cleavage process
avoids the formation of high-energy anionic radical intermedi-
ates, and instead a neutral radical intermediate is formed.
Under the photocatalytic conditions in the presence of an
electron donor (methanol in this study), electrons can accumu-
late in TiO2 particles if they cannot be efficiently removed by
the electron acceptor (PBDEs). The accumulation of electrons
in the conduction band would increase the Fermi level, and
consequently enhance the reduction ability of TiO2. After load-
ing of the noble metal with a high work function such as Pd
Conclusion
By using Pd–TiO2, a fast debromination of decabromodiphenyl
ether (BDE209) has been achieved under the irradiation of sun-
light or an artificial light source. Distinguished with the pristine
TiO2, the low-BDEs, usually the final products in TiO2 photoca-
Chem. Eur. J. 2014, 20, 11163 – 11170
www.chemeurj.org
11168
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
dation. The oven temperature was held initially at 1008C for 2 min,
increased at a rate of 158CminÀ1 to 2308C, 58CminÀ1 to 2708C,
and 108CminÀ1 to 3208C and kept at this temperature for 12 min.
The injection temperature was 3008C, and the detector tempera-
ture was 3258C. The carrier gas was helium at a constant flow rate
of 1 mLminÀ1. The concentration of BrÀ was measured by a DX-900
ion chromatograph with an IonPac AS23 column (Dionex). The
eluent for ion chromatography was 8 mm KOH solution with flow
talysis, can be further reduced to diphenyl oxide in Pd–TiO2
systems. Also, the formed BDE intermediates during the photo-
catalysis exhibit a significant difference between the Pd–TiO2
and pristine TiO2 systems: much less position selectivity for the
debromination on Pd–TiO2 was observed than that on the pris-
tine TiO2 surface. For another PBDE (BDE15), it was found that
only with the Pd-loaded system can the photoreduction of
BDE15 be efficiently conducted, and during this photoreduc-
tion process an evident induction period is observed. Such
a discrepancy between the pristine and Pd-loaded systems can
be attributed to the Pd-cocatalyzed effect, in which the elec-
trons can be first stored on the Pd surface and then further
transferred to PBDEs for the cleavage of the CÀBr bonds.
rate of 1 mLminÀ1
.
Acknowledgements
We gratefully acknowledge the National Basic Research Pro-
gram of China (973 program Nos. 2010CB933503 and
2013CB632405), the National Science Foundation of China
(Grant Nos. 21137004, 21273245, and 21277147), and the “Stra-
tegic Priority Research Program” of the Chinese Academy of
Sciences (No. XDA 09030200) for financial support.
Experimental Section
Catalyst preparation
The Pd–TiO2 catalyst with a loading of 2.0 wt% Pd was prepared
by a photodeposition method. First, TiO2–water/isopropanol dis-
persions were prepared by adding TiO2 particulates (5 gLÀ1) to
a mixture (100 mL) of water/isopropanol (1:1) in a Pyrex vessel. The
vessel was purged with argon for 1 h to remove O2 and protected
under an argon atmosphere during the irradiation. A 300 W xenon
lamp (PLS-SXE-300/300UV, Trustech, China) was used as the light
source. During the irradiation, a solution of PdCl2 dissolved in
water was introduced dropwise into the TiO2 suspension under
stirring. After irradiation for 2 h, the as-prepared samples were cen-
trifuged and thoroughly washed by water. Finally, they were
freeze-dried under vacuum overnight before use.
Keywords: debromination · diphenyl ethers · palladium ·
photocatalysis · titanium dioxide
[1] A. Bradman, R. Castorina, A. Sjçdin, L. Fenster, R. S. Jones, K. G. Harley, J.
Chevrier, N. T. Holland, B. Eskenazi, Environ. Sci. Technol. 2012, 46, 7373–
7381.
[2] C. M. Butt, M. L. Diamond, J. Truong, M. G. Ikonomou, A. F. H. terSchure,
Environ. Sci. Technol. 2004, 38, 724–731.
[3] L. S. Morf, J. Tremp, R. Gloor, Y. Huber, M. Stengele, M. Zennegg, Envi-
ron. Sci. Technol. 2005, 39, 8691–8699.
[4] N. Wu, T. Herrmann, O. Paepke, J. Tickner, R. Hale, E. Harvey, M. La Guar-
dia, M. D. Mc Clean, T. F. Webster, Environ. Sci. Technol. 2007, 41, 1584–
1589.
Debromination experiments and intermediates analysis
[5] A. Salamova, R. A. Hites, Environ. Sci. Technol. 2013, 47, 349–354.
[6] M. J. La Guardia, R. C. Hale, E. Harvey, Environ. Sci. Technol. 2006, 40,
6247–6254.
[7] R. A. Hites, Environ. Sci. Technol. 2004, 38, 945–956.
[8] A. O. W. Leung, W. J. Luksemburg, A. S. Wong, M. H. Wong, Environ. Sci.
Technol. 2007, 41, 2730–2737.
[9] I. Branchi, E. Alleva, L. G. Costa, Neurotoxicology 2002, 23, 375–384.
[10] I. Branchi, F. Capone, E. Alleva, L. G. Costa, Neurotoxicology 2003, 24,
449–462.
[11] M.-Y. Ahn, T. R. Filley, C. T. Jafvert, L. Nies, I. Hua, J. Bezares-Cruz, Environ.
Sci. Technol. 2006, 40, 215–220.
[12] G. Sçderstrçm, U. Sellstrçm, C. A. De Wit, M. Tysklind, Environ. Sci. Tech-
nol. 2004, 38, 127–132.
[13] Y. W. Suh, G. R. Buettner, S. Venkataraman, S. E. Treimer, L. W. Robertson,
G. Ludewig, Environ. Sci. Technol. 2009, 43, 2581–2588.
[14] C. Sun, D. Zhao, C. Chen, W. Ma, J. Zhao, Environ. Sci. Technol. 2009, 43,
157–162.
[15] C. Sun, W. Chang, W. Ma, C. Chen, J. Zhao, Environ. Sci. Technol. 2013,
47, 2370–2377.
[16] J. Eriksson, N. Green, G. Marsh, ꢂ. Bergman, Environ. Sci. Technol. 2004,
38, 3119–3125.
[17] A. C. Gerecke, P. C. Hartmann, N. V. Heeb, H.-P. E. Kohler, W. Giger, P.
Schmid, M. Zennegg, M. Kohler, Environ. Sci. Technol. 2005, 39, 1078–
1083.
[18] J. A. Tokarz, M.-Y. Ahn, J. Leng, T. R. Filley, L. Nies, Environ. Sci. Technol.
2008, 42, 1157–1164.
[19] K. R. Robrock, P. Korytꢃr, L. Alvarez-Cohen, Environ. Sci. Technol. 2008,
42, 2845–2852.
[20] P. Imm, L. Knobeloch, C. Buelow, H. A. Anderson, Environ. Health Per-
spect. 2009, 117, 1890–1895.
BDE209 stock solution (1 mmolLÀ1) in tetrahydrofuran was diluted
with methanol to 10 mmolLÀ1. The BDE209/Pd–TiO2 or BDE209/TiO2
suspensions were prepared by adding Pd–TiO2 (0.2 gLÀ1) or TiO2
(0.2 gLÀ1) particulates to a solution (25 mL) of BDE209 in a Pyrex
vessel. The vessel was sealed with an aluminum cap/rubber plug,
and mixed ultrasonically in a water bath for 2 min to disperse the
particles. For experiments under anoxic conditions, the vessel was
purged with argon for 30 min to remove O2, and the reaction
system was protected under an argon atmosphere during the irra-
diation. The suspensions were stirred in the dark for 1 h before irra-
diation to reach adsorption/desorption equilibrium. A mercury
lamp (100 W, Toyozumi Dengenkiki Co. Ltd.) was used as the light
source. To avoid the direct photolysis of BDE209, a cutoff filter
(UVCUT 360, Beijing Trusttech Co. Ltd.) was used to cut off the irra-
diation below 360 nm. During the irradiation, the suspension
(1 mL) was collected at given time intervals and centrifuged imme-
diately at 10000 rpm for 10 min to remove the Pd–TiO2 or TiO2 par-
ticles. The supernatants were used for HPLC and GC-mECD analysis.
The debromination of BDE15 was carried out under the same ex-
perimental conditions. To test the reproducibility and uncertainty
of the degradation experiments, the photochemical reactions were
repeated three times. Substrate in samples was quantified by HPLC
(Agilent 1260) monitoring at 240 nm, using a DikmaPlatisil ODS C-
18 column (250 mmꢁ4.6 mmꢁ5 mm). The mobile phase was 100%
acetonitrile at a flow rate of 1 mLminÀ1 for BDE209, and 85%
CH3CN/15% H2O for BDE15.
A gas chromatograph (Agilent 7890A) equipped with an electron
capture detector and a DB-5 capillary column (30 mꢁ250 mmꢁ
0.1 mm) was used for analyzing the intermediates of BDE209 degra-
[21] A. Sjçdin, L. Y. Wong, R. S. Jones, A. Park, Y. Zhang, C. Hodge, E. Dipietro,
C. McClure, W. Turner, L. L. Needham, D. G. Patterson, Environ. Sci. Tech-
nol. 2008, 42, 1377–1384.
Chem. Eur. J. 2014, 20, 11163 – 11170
www.chemeurj.org
11169
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[22] Y.-S. Keum, Q. X. Li, Environ. Sci. Technol. 2005, 39, 2280–2286.
[23] A. Li, C. Tai, Z. Zhao, Y. Wang, Q. Zhang, G. Jiang, J. Hu, Environ. Sci.
Technol. 2007, 41, 6841–6846.
[34] J. P. Boon, W. E. Lewis, M. R. Tjoen-A-Choy, C. R. Allchin, R. J. Law, J. B.
De, C. C. T. Hallers-Tjabbes, B. N. Zegers, Environ. Sci. Technol. 2002, 36,
4025–4032.
[24] Z. Zhang, Y. Yu, P. Wang, ACS Appl. Mater. Interfaces 2012, 4, 990–996.
[25] J. J. Zou, Y.-P. Zhang, C.-J. Liu, Langmuir 2006, 22, 11388–11394.
[26] B. Dong, B. L. He, J. Huang, G. Y. Gao, Z. Yang, H. L. Li, J. Power Sources
2008, 175, 266–271.
[35] Y. Zhuang, S. Ahn, R. G. Luthy, Environ. Sci. Technol. 2010, 44, 8236–
8242.
[36] J. W. Hu, Y. Zhuang, J. Luo, X. H. Wei, X. F. Huang, Int. J. Mol. Sci. 2012,
13, 9332–9342.
[27] M. Wang, D. J. Guo, H. L. Li, J. Solid State Chem. 2005, 178, 1996–2000.
[28] S. K. Mohapatra, N. Kondamudi, S. Banerjee, M. Misra, Langmuir 2008,
24, 11276–11281.
[29] M. A. Keane, Appl. Catal. A 2004, 271, 109–118.
[30] C. Schuðth, M. Reinhard, Appl. Catal. B 1998, 18, 215–221.
[31] Y. Zhuang, S. Ahn, A. L. Seyfferth, Y. Masue-Slowey, S. Fendorf, Environ.
Sci. Technol. 2011, 45, 4896–4903.
[37] J. W. Hu, L. Eriksson, A. Bergman, E. Jakobssonc, E. Kolehmainen, J.
Knuutinen, R. Suontamo, X. H. Wei, Chemosphere 2005, 59, 1043–1057.
[38] B. P. Chaplin, M. Reinhard, W. F. Schneider, C. Schꢄth, J. R. Shapley, T. J.
Strathmann, C. J. Werth, Environ. Sci. Technol. 2012, 46, 3655–3670.
[39] Y. Shiraishi, Y. Takeda, Y. Sugano, S. Ichikawa, S. Tanaka, T. Hirai, Chem.
Commun. 2011, 47, 7863–7865.
[32] H. Wei, R. Q. Yang, A. Li, E. R. Christensen, K. J. Rockne, J. Chromatogr. A
2010, 1217, 2964–2972.
Received: March 5, 2014
[33] A. F. H. ter Schure, P. Larsson, J. Merila, K. I. Jonsson, Environ. Sci. Technol.
2002, 36, 5057–5061.
Published online on July 25, 2014
Chem. Eur. J. 2014, 20, 11163 – 11170
www.chemeurj.org
11170
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim