Pathway of oxygen incorporation from O2 in TiO2 photocatalytic hydroxylation of aromatics: Oxygen isotope labeling studies

Article abstract of DOI:10.1002/chem.201103446

The hydroxylation process is the primary, and even the rate-determining step of the photocatalytic degradation of aromatic compounds. To make clear the hydroxylation pathway of aromatics, the TiO₂ photocatalytic hydroxylation of several model substrates, such as benzoic acid, benzene, nitrobenzene, and benzonitrile, has been studied by an oxygen-isotope-labeling method, which can definitively assign the origin of the O atoms (from oxidant O₂ or solvent H₂O) in the hydroxyl groups of the hydroxylated products. It is found that the oxygen source of the hydroxylated products depends markedly on the reaction conditions. The percentage of the products with O₂-derived hydroxyl O atoms increases with the irradiation time, while it decreases with the increase of substrate concentration. More intriguingly, when photogenerated valence-band holes (h _vb⁺) are removed, nearly all the O atoms (>97%) in the hydroxyl groups of the hydroxylated products of benzoic acid come from O ₂, whereas the scavenging of conduction-band electrons (e _cb^-) makes almost all the hydroxyl O atoms (>95%) originate from solvent H₂O. In the photocatalytic oxidation system with benzoic acid and benzene coexisting in the same dispersion, the percentage of O₂-derived hydroxyl O atoms in the hydroxylated products of strongly adsorbed benzoic acid (ca. 30%) is much less than in that of weakly adsorbed benzene (phenol) (>60%). Such dependences provide unique clues to uncover the photocatalytic hydroxylation pathway. Our experiments show that the main O₂-incorporation pathway involves the reduction of O₂ by e_cb^- and the subsequent formation of free ^.OH via H₂O₂, which was usually overlooked in the past photocatalytic studies. Moreover, in the hydroxylation initiated by h_vb⁺, unlike the conventional mechanism, the O atom in O₂ cannot incorporate into the product through the direct coupling between molecular O₂ and the substrate-based radicals.

Full text of DOI:10.1002/chem.201103446

DOI: 10.1002/chem.201103446
Pathway of Oxygen Incorporation from O2 in TiO2 Photocatalytic
Hydroxylation of Aromatics: Oxygen Isotope Labeling Studies
Yue Li, Bo Wen, Cailan Yu, Chuncheng Chen,* Hongwei Ji, Wanhong Ma, and
Jincai Zhao[a]
Abstract: The hydroxylation process is
the primary, and even the rate-deter-
mining step of the photocatalytic deg-
radation of aromatic compounds. To
make clear the hydroxylation pathway
of aromatics, the TiO2 photocatalytic
hydroxylation of several model sub-
strates, such as benzoic acid, benzene,
nitrobenzene, and benzonitrile, has
been studied by an oxygen-isotope-la-
beling method, which can definitively
assign the origin of the O atoms (from
oxidant O2 or solvent H2O) in the hy-
droxyl groups of the hydroxylated
products. It is found that the oxygen
source of the hydroxylated products
depends markedly on the reaction con-
ditions. The percentage of the products
with O2-derived hydroxyl O atoms in-
creases with the irradiation time, while
it decreases with the increase of sub-
strate concentration. More intriguingly,
when photogenerated valence-band
holes (hvb+) are removed, nearly all the
products of strongly adsorbed benzoic
acid (ca. 30%) is much less than in
that of weakly adsorbed benzene
(phenol) (>60%). Such dependences
provide unique clues to uncover the
photocatalytic hydroxylation pathway.
Our experiments show that the main
O2-incorporation pathway involves the
O
atoms (>97%) in the hydroxyl
groups of the hydroxylated products of
benzoic acid come from O2, whereas
the scavenging of conduction-band
electrons (ecbꢀ) makes almost all the
hydroxyl O atoms (>95%) originate
from solvent H2O. In the photocatalytic
oxidation system with benzoic acid and
benzene coexisting in the same disper-
sion, the percentage of O2-derived hy-
droxyl O atoms in the hydroxylated
ꢀ
reduction of O2 by ecb and the subse-
C
quent formation of free OH via H2O2,
which was usually overlooked in the
past photocatalytic studies. Moreover,
+
in the hydroxylation initiated by hvb
,
unlike the conventional mechanism,
the O atom in O2 cannot incorporate
into the product through the direct
coupling between molecular O2 and
the substrate-based radicals.
Keywords: aromatic compounds ·
hydroxylation · isotope labeling ·
photocatalysis · titanium dioxide
Introduction
pounds have been the most frequently used model sub-
strates to investigate photocatalytic mechanisms[5] and to
test the activity of the photocatalysts.[6] During the photoca-
talytic oxidation of aromatic compounds, the hydroxylated
products are always among the main intermediates and the
hydroxylation of the aromatics is usually regarded as the
primary step,[7] and even the rate-determining step,[8] of the
TiO2 photocatalysis has attracted persistent attention due to
its potential application in water purification,[1] water split-
ting,[2] and synthesis of chemicals.[3] During semiconductor
photocatalysis, the absorption of irradiation with energy
larger than the band gap leads to the formation of valence-
band holes (hvb+) and conduction-band electrons (ecbꢀ),
which could initiate the oxidation and reduction reactions,
respectively. For example, nearly all kinds of organic com-
pounds could be degraded, even completely mineralized, in
the air-saturated aqueous TiO2 suspensions with UV irradia-
tion.[4]
whole photocatalytic mineralization process. The hydroxyl-
[9]
C
ation is also utilized to detect the generation of OH, be-
cause the hydroxylated products of aromatics, such as cou-
marin and terephthalic acid, are easily detected by fluores-
cence emission or chromatography. Besides its great signifi-
cance in photocatalysis, the hydroxylation of aromatic com-
pounds also attracts extensive attention from scientists in a
variety of disciplines and fields, such as biochemistry,[10] en-
vironmental chemistry,[11] and synthetic organic chemistry.[12]
Despite extensive attention, the detailed reaction pathway
of photocatalytic hydroxylation of aromatics still eludes the
understanding of the researchers. Most of the past studies
mainly focused on the detection of hydroxylated products
and their formation kinetics,[5b,13] while little work has been
dedicated to the specific hydroxylation pathway under pho-
tocatalytic conditions so far. During the photocatalytic pro-
cess, O2 is the oxidant and the final electron acceptor.
As a large group of organic chemicals, and due to their
great effect on environmental pollution, aromatic com-
[a] Dr. Y. Li, Dr. B. Wen, C. Yu, Dr. C. Chen, Dr. H. Ji, Dr. W. Ma,
Dr. J. Zhao
Beijing National Laboratory for Molecular Sciences
Key Laboratory of Photochemistry, Institute of Chemistry
Chinese Academy of Sciences, Beijing 100190 (P.R. China)
Fax : (+86)108-261-6495
E-mail: ccchen@iccas.ac.cn
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201103446.
2030
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2030 – 2039

FULL PAPER
+
[5b,13a]
Therefore, how it takes part in the reaction is extremely im-
portant and can provide essential information on the whole
hydroxylation process. Generally, the primary role of molec-
ular O2 is to depress the charge recombination and regener-
ate the photocatalyst by removing ecbꢀ. O2 was also consid-
ered to be able to directly participate in the oxidation pro-
cess and enter the oxidized products. Recently, by using iso-
tope-labeling methods, Matsumura, et al.[8] experimentally
showed that O2 contributes to the formation of the hydroxy-
lated product of benzene. In the anatase and rutile systems,
10–30 and 60–80%, respectively, of the hydroxyl O atom in
the formed phenol were found to be from O2 during the
photocatalytic hydroxylation of benzene. However, the de-
tailed pathway of the O2 incorporation is not fully under-
stood.
mainly from the H2O/OHꢀ oxidation by hvb
.
Thus the
O atom in radical II, formed through both pathways, would
be from H2O/OHꢀ. The most probable fate of radical II is
to simply undergo one-electron oxidation and deprotonation
to form the hydroxylated product (path c in Scheme 1B). By
this mechanism, the hydroxyl group of the hydroxylated
product would retain the O atom of radical II, that is, its hy-
droxyl O atom comes from H2O. In principle, besides the
hvb oxidation of H2O/OHꢀ, OH could also be formed
+
C
ꢀ
through the reduction of dissolved O2 by ecb [by means of
Eqs. (1)–(5)], by which the O atom of O2 is incorporated
into radical II and subsequently into the product through
path c. However, the importance of this pathway in the pho-
tocatalytic oxidation is usually overlooked in the literature.
C
Most of the earlier experiments indicated that the OH radi-
It is well accepted that, as shown in Scheme 1A, the hy-
droxylation of aromatics is initiated by direct oxidation by
cal from the reduction of in situ formed H2O2 [Eq. (5)] is
generally not important for the photocatalytic oxidation of
organic compounds,[15] although some reports showed that
the addition of H2O2 could enhance the rate of the photoca-
talytic oxidation.[16]
+
hvb (form the cationic radical I) followed by hydrolysis
(path a)[14] or the attack of OH (path b) in the photocatalyt-
C
ic systems. Both pathways lead to the formation of HO
adduct radical II, which bears the O atom from H2O/OHꢀ
Cꢀ
ꢀ
O2 þ ecb ! O2
ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
C
C
or OH. Conventionally, OH is believed to be derived
O2 þ Hþ ! OOH
ðpKa ¼ 4:8Þ
Cꢀ
C
C
2 OOH ! H2O2 þ O2
OOH þ ecbꢀ þ Hþ ! H2O2
C
ꢀ
ꢀ
C
H2O2 þ þecb ! OH þ OH
To explain the experimental observation of the incorpora-
tion of O atom from O2 into the products, several mecha-
nisms involving the formation of radical O2 adducts have
been proposed. Generally, these pathways are adapted from
[17]
C
C
the so-called Russell mechanism (R +O2!ROO ), which
is normally applicable to the oxidation of aliphatic sub-
strates.[4,18] The radical HO adduct II was proposed to react
with molecular O2 to generate the radical O2 adduct III
(path d in Scheme 1B) in both photocatalytic and other oxi-
dation systems.[19] The further reactions of radical III could
introduce the O atom of either O2 or H2O into the hydroxy-
lated product (paths d1 and d2). Analogously, molecular O2
Cꢀ
(or O2 ) has been considered to directly couple with formed
cationic radical I to generate radical O2 adducts IV and V
(paths e and f in Scheme 1C).[8] The products formed by the
further reactions of these adducts would bear the O atoms
from O2.
Herein, as part of our mechanistic studies on the photoca-
talytic oxidation of organic compounds by an 18O-labeling
method,[18c,20] we report our experimental evidence on the
origin of the hydroxyl O atom and reaction pathway in the
photocatalytic hydroxylation of aromatics. We first show
that, in a 16O2/H218O system, the 16O (from 16O2) percentage
in the hydroxylated product depends markedly on reaction
conditions and changes with irradiation time. Further de-
tailed examination on the oxygen source of the hydroxyl-
ation shows that a pathway involving the reduction of O2 by
Scheme 1. Traditional mechanisms of the hydroxylation of aromati-
cs.[5b,8,13a,14,17]
Chem. Eur. J. 2012, 18, 2030 – 2039
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2031

C. Chen et al.
ꢀ
C
ecb and the subsequent formation of free OH via H2O2,
which was usually overlooked in the past photocatalytic
studies, is the main O2-incorporation mechanism. Such a
mechanism can interpret perfectly all our experimental ob-
servations. In our study, no experimental evidence supports
that the radical O2 coupling mechanisms (paths d–f in
Scheme 1) can play a significant role in the incorporation of
O2. These results help to deepen our understanding on the
photocatalytic hydroxylation of aromatic compounds.
with the irradiation time. For example, the 16O abundance
increased from 33.0 to 40.1%, when the irradiation time in-
creased from 1 to 2 h at a BA initial concentration of 3 mm.
A similar trend was observed when higher substrate concen-
trations were used (Figure 1). It was also observed that the
proportion of BA-16OH decreased as the initial concentra-
tion of the substrate increased. For example, when a saturat-
ed solution of BA (ca. 25 mm) was applied, only 13.4% (vs.
33.0% at 3 mm BA) of the hydroxyl O atom was derived
from aerial O2 after 1 h of irradiation. Such a dependence of
the proportion of O2-derived hydroxyl O atoms on the reac-
tion time and initial substrate concentration were also con-
firmed by using H216O as solvent and 18O2 as oxidant (Fig-
ure S2 in the Supporting Information). The O2 incorporation
and the enhancement of its proportion with the irradiation
time were also observed in the photocatalytic hydroxylation
of nitrobenzene and benzonitrile (Table S1 in the Support-
ing Information).
Results and Discussion
Oxygen source of the hydroxylated products formed in the
photocatalytic oxidation of aromatic compounds: The pho-
tocatalytic reaction of benzoic acid (BA) was first carried
out in 18O-enriched water (H218O) using 16O2 in the air as ox-
idant. The measured isotope abundance[21] was corrected to
determine the proportion of the origination of the hydroxyl
O atom. In blank experiments, a TiO2–H218O suspension
containing BA or hydroxylbenzoic acid (BA-OH) was
stirred for 10 h, with and without UV irradiation, and no
change in the isotope abundance in the remaining BA or
BA-OH was detected. These experiments indicate that the
exchange of the O atoms in the carboxyl groups of BA and
BA-OH and the hydroxyl group of BA-OH with H218O is
rather slow and can be ignored under the present experi-
mental conditions. The percentages of 16O in the hydroxyl
group of BA-OH are shown in Figure 1.[22] At an initial sub-
Another interesting observation was that the oxygen iso-
tope abundance of the hydroxylated product was also affect-
ed by the adsorption ability of the substrate. As shown in
Table 1, when BA and benzene, which have different ad-
Table 1. Percentages of 16O in the hydroxyl groups of BA-OH and phe-
nol.[a]
0
Entry
cBA
BA
BA-OH [%]
BA + benzene
BA-OH [%] phenol [%]
[mm]
1
2
3
20
7
0
26.6
38.5
—
22.3
32.9
—
67.9
62.9
47.1
[a] The H218O solution of BA (1 mL) containing TiO2 (2 gLꢀ1), with or
without the addition of benzene (50 mL), under aerated (16O2) conditions,
irradiation time: 2 h.
sorption abilities on TiO2,[23] coexisted in the same disper-
sion, the 16O abundance (from 16O2) in the formed BA-OH
was much lower than that in phenol (the hydroxylated prod-
uct of benzene). For example, in the dispersion with 20 mm
BA and 23 mm (saturated) benzene, BA-16OH accounted for
only 22.3% of the hydroxylated product of BA, whereas up
to 67.9% of the O atoms in phenol was from O2 (Table 1,
entry 1). Further, the addition of benzene slightly lowered
the proportion of 16O in BA-OH (for example, from 38.5 to
32.9% at initial BA concentration of 7 mm), while the pres-
ence of BA notably increased the 16O abundance in phenol
from 47.1% (Table 1, entry 3) to 60–70%.
Figure 1. Percentages of O2-derived hydroxyl O atom in BA-OH, using
aerial 16O2 as oxidant and H218O as solvent. TiO2 (P25; 2 gLꢀ1), H218O
(1 mL), irradiation time: 1 and 2 h.
The time-dependence of the 16O abundance gives us a
hint that the O2 incorporation might be determined by some
intermediate formed during the photoreaction, while the dif-
ference in isotope distribution of the hydroxylated product
between BA and benzene suggests that the adsorption of
the substrate might have an influence on the O2 incorpora-
tion. In the following study, by detailed isotope-labeled ex-
periments, we will identify the intermediate that controls
the O2-incorporation process, and reveal the oxygen sources
of the hydroxylated product and the corresponding hydrox-
ylation pathways.
strate concentration of 3 mm, the proportion of 16O hydrox-
ylbenzoic acid (BA-16OH) was 33.0% after 1 h of irradia-
tion, which means that about one-third of the O atoms in
the hydroxyl group originates from oxidant O2, while the
other two-thirds comes from solvent H2O (18O). More inter-
estingly, we found that the isotope abundance of the hydrox-
yl O atom of the product was markedly dependent on the
reaction conditions, such as the irradiation time and sub-
strate concentration. The proportion of BA-16OH increased
2032
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2030 – 2039

Photocatalytic Hydroxylation
FULL PAPER
Contribution of hvb and ecbꢀ: Since the photocatalytic reac-
droxyl O atoms in BA-OH was lowered significantly, for ex-
ample, from 18.2 to 10.0%, in the presence of 1 mm BQ
(cBQ0:cBA0 =1:20) after 1 h of irradiation (Table 2, entry 6).
Nearly all the hydroxyl O atom was found to originate from
solvent H2O and the oxygen incorporation from O2 into the
hydroxyl group was completely inhibited when a large
amount of BQ was added (Table 2, entries 7 and 8). By
using 18O2 as the oxidant and H216O as solvent, BA-16OH
was nearly the exclusive hydroxylated product in the pres-
+
tion is initiated by photogenerated hvb and ecbꢀ, we first ex-
amined their effects on the isotope distribution of the hy-
droxylated product. To this end, we designed some special
experiments, in which hvb or ecb was selectively removed,
and then the oxygen isotope abundance in BA-OH was ana-
lyzed. The results are summarized in Table 2. The role of
+
+
ꢀ
Table 2. Abundance of 16O in the hydroxyl group of BA-OH formed in
the photocatalytic oxidation with the addition of the scavenger of hvb or
ꢀ
+
ence of 40 mm BQ (Table 2, entry 9). After ecb is removed
ꢀ [a,b]
ꢀ
ecb
.
and the activation of O2 by ecb is inhibited, the hydroxyl-
Abundance of 16O [%]
1 h
ation process should be initiated only by hvb+. The complete
inhibition of the O2 incorporation under such conditions
Entry
Scavenger
(c [mm])
2 h
26.6
33.2
97.5
97.7
0.6
19.2
0.3
ca. 0
94.8
1[c]
2[c]
3[c]
4[c]
5[d]
6[c]
7[c]
8[c]
9[d]
none
18.2
38.3
97.7
97.8
1.9
10.0
0.1
ca. 0
95.8
+
suggests that, in the hydroxylation pathway initiated by hvb
ꢀ
FA (2)
, molecular O2 without activation by ecb cannot enter the
FA (40)
FA (700)
FA (700)
BQ (1)
BQ (10)
BQ (40)
BQ (40)
product. Therefore, the mechanisms involving the direct re-
actions between substrate-based radicals and O2 (paths d
and e in Scheme 1)[8,19] are not likely to be the main path-
ways for the O2 incorporation during the hydroxylation of
aromatic compounds.
C
It is also noted that FA is much less competitive for OH
[a] TiO2 (P25; 2 gLꢀ1), cBA0 =20 mm, irradiation time: 1 and 2 h. [b] The
additives inhibit the degradation of BA. For example, the addition of
700 mm FA reduced the consumption rate of BA from 1.86 mmminꢀ1 to
0.23 mmminꢀ1. [c] In H218O (1 mL), under aerated (16O2) conditions. [d] In
H216O (5 mL), 1 atm 18O2.
than BA, since the rate constant of the reaction between FA
and OH (kFA =1.3ꢁ108 mꢀ1 sꢀ1)[26] is much less than that of
C
BA (kBA =4.3ꢁ109 mꢀ1 sꢀ1).[27] For example, in the presence
of 2 mm of FA and 20 mm of BA (Table 2, entry 2), only
C
0.3% of the OH is estimated to be captured by FA. There-
fore, the hydroxylation reaction of BA could still occur
+
C
hvb was first examined by adding formic acid (FA) to
through the reaction between BA and OH, if generated,
remove it selectively.[24] In the presence of only 2 mm of FA,
which was 10% of the BA concentration, the abundance of
16O (from 16O2) in the hydroxylated product increased re-
markably (Table 2, entry 2, for example, from 18.2 to 38.3%
for 1 h of irradiation), which means that the incorporation
of the O atom from solvent H2O was greatly suppressed.
Further increase of the FA concentration could make nearly
all the O atoms in the hydroxyl group of the product come
from O2 (Table 2, entries 3 and 4). When 18O2 was used, and
the solvent water was in its natural isotope abundance
(H216O), almost all the hydroxyl O atoms in BA-OH were
18O if 700 mm of FA was added (Table 2, entry 5). The sup-
pression of the oxygen incorporation from H2O by FA indi-
even in the presence of a high concentration of FA. On the
other hand, in the BQ system, the reaction between BQ and
9
OH has a rate constant of 1.2ꢁ10 mꢀ1 sꢀ1.[26] Even at the
C
highest addition concentration (40 mm), BQ can only cap-
C
ture about 36% of the free OH in the competition with BA
(20 mm). Because of the strong adsorption of BA on the sur-
+
face of TiO2, the competitiveness of BQ for hvb and surface
+
C
OH should be much less than that of BA. Therefore, hvb
C
and OH can still initiate the hydroxylation of BA in the
presence of BQ.
C
In another experiment, OH was produced by the photoly-
sis of H2O2 [Eq. (6)][6a,25a] to model the reaction initiated by
C
OH. When the photochemical reaction was carried out in
H216O/H216O2 solution but 18O2 atmosphere, no incorporation
of the 18O atom of 18O2 was observed (Table 3, entry 1). In
H218O/H216O2 solution and 16O2 atmosphere, almost all the O
atoms in the hydroxyl group was 16O (Table 3, entry 2). It
clearly indicates that the hydroxyl O atom is predominantly
+
cates that hvb is indispensable for the H2O incorporation,
that is, H2O cannot participate directly in the hydroxylation
of aromatic compounds initiated by ecbꢀ. In addition, most
of the earlier studies seemed to indicate that the hydroxyl-
+
ation of aromatic compounds is directly related to hvb or
[8,13a,14]
+
C
C
OH generated from it.
In our study, however, hvb
derived from H2O2, that is, OH is generated from H2O2 in
was expected to be completely scavenged, when the concen-
tration of FA was high enough (for example, 700 mm as
shown in Table 2, entries 4 and 5). The formation of BA-OH
under these conditions suggests the presence of another hy-
droxylation pathway, which incorporates the O atom of oxi-
Table 3. Percentages of the hydroxyl O atom of BA-OH originating from
18O enriched reagent in the H2O2/UV and S2O82ꢀ/UV systems.[a]
Entry
System
Abundance of 18O [%]
1 h
2 h
dant O2 into the hydroxylated product (see below), but does
1
2
3
18O2/H216O2/H216O
ca. 0
3.3
98.4
ca. 0
2.9
98.4
+
16O
16O
2A
2A
(air)/S216O82ꢀ/H218O
not involve hvb
.
In contrast, after benzoquinone (BQ) was employed to
CTHUNGTRENNUNG
ꢀ
[25]
Cꢀ
scavenge ecb and/or oxidize O2 back to O2 to hinder the
[a] The solution (1 mL) containing 10 mm H2O2 or K2S2O8 and 20 mm
BA, irradiation time: 1 and 2 h.
activation of O2 by ecbꢀ, the percentage of O2-derived hy-
Chem. Eur. J. 2012, 18, 2030 – 2039
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2033

C. Chen et al.
Table 4. Abundances of 16O in the hydroxyl group of BA-OH formed in
photocatalysis with different additives.[a]
this system. This experiment also suggests that the hydroxyl-
C
ation process initiated by OH cannot insert the O atom of
Entry
Additive
Abundance of 16O [%]
molecular O2 into the product, which further implies that
path d in Scheme 1B may play a minor role in the O2 incor-
poration. Another implication of this experiment is that the
1 h
2 h
1
none
SOD
18.2
25.9
4.9
48.8
34.8
98.0
26.6
32.7
4.9
41.9
23.3
97.7
2[b]
3[c]
4[d]
5[e]
6[f]
C
C
isotope exchange between OH or the OH adduct II and
H2O is rather slow under the present experimental condi-
tions, and hence the 18O isotope labeling is a reliable tech-
nique to study the hydroxylation process.
POD
H216O2
18O2/H216O2
Ar/FA/H216O2
[a] TiO2 (P25; 2 gLꢀ1), H218O (1 mL), cBA0 =20 mm, under aerated (16O2)
C
conditions, irradiation time: 1 and 2 h. [b] SOD (500 U initially and 40 U
H2O2 ! 2 OH
ð6Þ
per 5 min during the reaction) was added. [c] POD (34 mg initially and
0
2
3.4 mg per 5 min after the beginning of irradiation) was added. [d] cH O
=
+
2
The direct oxidation process of hvb was modeled by the
oxidation of BA by the one-electron oxidant SO4
5.2 mm. [e] cH O 0 =5 mm, 1 atm of 18O2. [f] cFA0 =0.7m, cH O 0 =20 mm, in
2
2
2
2
Cꢀ [28]
It was
.
Ar atmosphere.
Cꢀ
reported that SO4 reacts with aromatic compounds through
one-electron transfer [Eq. (7)], rather than adding to it.[29]
·ꢀ
Also, SO4 is stable in acidic and neutral aqueous solutions
the hydroxylated product, which excludes path f in
Scheme 1C as the main O2-incorporation pathway.[8] More-
over, although both the addition of SOD and BQ could
remove the formed superoxide species, the change in 16O
abundance of the hydroxylated product exhibited complete-
ly opposite trends, that is, the O2 incorporation was slightly
facilitated by SOD, while it was markedly suppressed in the
presence of BQ (Table 1, entries 6–9). In the BQ system,
[30]
C
and cannot oxidize solvent H2O to OH.
In our study,
SO4 was generated by the photolysis of S2O8 [Eq. (8)].[31]
Cꢀ
2ꢀ
The hydroxylated product formed from the one-electron ox-
idation of BA by SO4 had the hydroxyl O atom dominantly
Cꢀ
from H218O and the incorporation of 16O2 was in the range
of measurement error (Table 3, entry 3). This result is simi-
lar to the case in which BQ was used to block the activation
of O2 in the TiO2 photocatalytic system (Table 2, entries 6–
9). Together with the above FA- and BQ addition and H2O2/
UV experiments, we can exclude safely the earlier proposed
possibility that the O atom of O2 incorporates into the hy-
droxylated product by the direct addition of molecular O2 to
the formed radical species (paths d and e in Scheme 1),[8,19]
because according to these pathways, the O2 incorporation
should always be observed no matter how the reaction is ini-
tiated. However, it should be pointed out that we cannot
rule out the formation of these radical O2 adducts, because
the further reactions of theses high-energy peroxyl radicals
lead to the cleavage of aromatic ring,[32] which is not consid-
ered in the present study. ,
Cꢀ C
O2 / OOH is oxidized back to O2, while the superoxide spe-
cies disproportionates catalytically to H2O2 and O2 in the
SOD system.[25] The contrastive effects between SOD and
BQ on the oxygen isotope abundance suggest that H2O2
might play an important role in the O2 incorporation.
H2O2 is the two-electron reduction product of O2. Its role
was further examined by the addition of horseradish perox-
idase (POD) to decompose H2O2 selectively to H2O and O2
[Eq. (9)].[34] It was found that POD quenched H2O2 com-
pletely (curve c in Figure S4 in the Supporting Information),
and hindered remarkably the incorporation of the O atom
of O2 into BA-OH (Table 4, entry 3). For example, the 16O
percentage decreased from 18.2 to 4.9% after the addition
of POD for the 1 h irradiated system. On the other hand,
adding H216O2 to the TiO2/H218O system increased signifi-
cantly the 16O abundance in the hydroxyl group of the prod-
uct (Table 4, entry 4). For instance, after 1 h of irradiation,
the addition of H216O2 increased the 16O percentage of the
product from 18.2 to 48.8%. Similarly, in the SOD system,
which also generated the product with higher 16O abun-
dance, the accumulation of H2O2 was observed to be en-
hanced to some extent (curve b in Figure S4 in the Support-
ing Information, probably because the rapid conversion of
2ꢀ
Cꢀ
Cþ
ð7Þ
ð8Þ
BA þ SO4 ! BA þ SO4
Cꢀ
2ꢀ
S2O8 ! 2SO4
Roles of reactive oxygen species: After clarifying the impor-
tance of O2 activation by ecb and excluding the pathways
through organic radical O2 adducts, we decided to follow the
pathway of the activation of O2 by ecb to identify the reac-
tive oxygen species that is responsible for the O2 incorpora-
tion. Superoxide species (O2 / OOH) are the first reactive
oxygen species formed by the one-electron reduction of O2
by ecb [Eqs. (1) and (2)]. The addition of superoxide dismu-
ꢀ
ꢀ
Cꢀ C
Cꢀ C
O2 / OOH into H2O2 and O2 suppresses the charge recom-
bination[35]). The correlation between the 16O2 incorporation
and the concentration of H2O2 in all above systems suggests
again that H2O2, which is generated primarily from the re-
ꢀ
tase (SOD), which catalyzes selectively the disproportiona-
[33]
ꢀ
Cꢀ C
tion of O2 / OOH to form H2O2 and O2 [Eq. (3)], caused
duction of O2 by ecb [Eqs. (1)–(4)],[36] plays a key role in
a slight enhancement of the 16O percentage in the hydroxyl
group of BA-OH (Table 4, entry 2, for example, from 18.2
the incorporation of O2 into the hydroxylated product. To
verify this, H216O2 was added into the TiO2/H218O/18O2
system. It was found that a significant proportion of the O
atom in the hydroxyl group of the product was 16O from
H216O2 (for example, 34.8% after 1 h of irradiation, Table 4,
to 25.9% after 1 h of irradiation). The facilitation of the
O2 incorporation indicates that O2 / OOH is not a necessa-
ry species that directly introduces the O atom from O2 into
16
Cꢀ C
2034
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2030 – 2039

Photocatalytic Hydroxylation
FULL PAPER
entry 5). It definitely proves that the O atom in H2O2 could
be incorporated into the hydroxylated product in the photo-
catalytic systems.
droxylated product, and the 16O proportion of the product
formed (60–90%) was higher than that using aerial 16O2 as
oxidant (10–40%, Figure 1). Further, the addition of FA
lead to nearly all the hydroxyl O atoms in the product (ca.
98%) arising from H2O2 (Table 4, entry 6). These results
suggest that the oxygen incorporation from H2O2 into the
hydroxylated product is initiated by ecbꢀ, while the H2O in-
corporation is triggered by hvb+, which is similar to the case
of O2 as oxidant. However, because three electrons are
2 H2O2 ! 2 H2O þ O2
ð9Þ
It is well-known that both the further reduction and pho-
C
tolysis of H2O2 could lead to the formation of OH [Eqs. (5)
and (6)]. Spin-trapping electron spin resonance (ESR) was
C
needed for the reduction of O2 (via H2O2) to OH, while the
C
used to detect the formation of OH in the photocatalytic
C
reduction from H2O2 to OH needs only one electron
systems (Figure 2). The addition of SOD enhanced the in-
[Eqs. (1)–(5)], more 16OH should be generated from the
C
ꢀ
direct capture of ecb by H216O2 than by 16O2. Therefore, the
system using H2O2 as oxidant formed the hydroxylated
product with much higher 16O abundance.
Next, we will show that our experimental observations of
the O2-incorporation dependence on the reaction conditions
can be well interpreted by the mechanism in which the
oxygen incorporation from O2 into the hydroxylated product
ꢀ
is attributed to the reduction of O2 by ecb and the subse-
C
quent formation of OH (via H2O2). Unlike other reactive
C
C
oxygen species such as OOH and OH, H2O2 is so stable
that its accumulation in the solution can be quantitatively
measured by the steady-state technique. As shown in
Figure 3, at all the tested substrate concentrations, the con-
Figure 2. The ESR signals of the BA/TiO2/UV systems with different ad-
ditives. TiO2 (2 gLꢀ1), BA (20 mm) and DMPO (50 mm), H2O (1 mL),
with the addition of a) none, b) SOD (500 U), c) POD (40 mg) and d)
iPrOH (0.6 mmol), respectively, under aerated conditions, irradiation
time: 5 min. (Fourfold peaks with intensity of 1:2:2:1 labeled by “*” are
attributed to DMPO-OH, sixfold peaks labeled by “o” are due to
DMPO-C(OH)Me2.)
C
tensity of the signal of trapped OH slightly, while POD dra-
matically suppressed its signal, which is in good agreement
with their effects on the isotope distribution of the hydroxy-
lated product (Table 4, entries 2 and 3) and the concentra-
tion of formed H2O2 (Figure S4 in the Supporting Informa-
Figure 3. The change of H2O2 concentration along with the irradiation
time at different initial substrate concentrations. TiO2 (P25; 2 gLꢀ1), H2O
(1 mL), under aerated conditions. The inset tabulates the corresponding
apparent formation and consumption rate constants.
C
tion). These results suggest that the OH generated from
H2O2 may be an important intermediate in the oxygen incor-
poration from O2 into the hydroxylated product. In the pres-
ent photocatalytic system, because the UV absorption of
TiO2 is much stronger than that of H2O2, the photolysis of
centration of H2O2 increased with the reaction time, which
is consistent with the increase of the 16O abundance of the
hydroxylated product with irradiation time (Figure 1; see
also Figure S2 and Tables S1 and S2 in the Supporting Infor-
mation). By the addition of H216O2 to the H218O system at
the beginning of the reaction, however, the 16O abundance
of the product was observed to decrease with prolonged ir-
radiation time (Table 4, entry 4), and the concentration of
H2O2 declined from 5.2 mm to 2.4 mm during 2 h of irradia-
tion. All these facts indicate that the accumulation of H2O2
during the reaction is responsible for the increase of the 16O
abundance with the irradiation time.
ꢀ
H2O2 should be much slower than the ecb reduction pro-
cess.[36,37] Therefore, the formation of OH from H2O2 should
C
ꢀ
be mainly ascribed to the surface reduction of H2O2 by ecb
,
but the direct photolysis of H2O2 cannot play a significant
C
role in the OH formation under the photocatalytic condi-
tions.
C
To further study the role of H2O2 and the OH from H2O2,
ꢀ
we used H216O2 to replace 16O2 as oxidant to capture ecb (in
Ar atmosphere) in the TiO2/H218O/UV systems. The results
are shown in Figure S3 in the Supporting Information. It is
evident that the O atom of H2O2 incorporated into the hy-
Chem. Eur. J. 2012, 18, 2030 – 2039
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2035

C. Chen et al.
In addition, the formed H2O2 accumulation was observed
to be dependent on the concentration of the substrate. At
the beginning stage of the photocatalytic reaction (0–
30 min), the concentration of H2O2 was not seriously affect-
ed by the substrate concentration. However, after 30 min of
irradiation, the concentration of H2O2 increased with the ini-
tial BA concentration at the same irradiation time. The con-
centrations of H2O2 were 0.29, 0.45, and 0.59 mm for the sys-
tems with initial BA concentration of 5, 20, and 25 mm, re-
spectively, at irradiation of 120 min. The accumulation of
H2O2 is determined by its formation and consumption
during the photocatalytic reaction. The formation and con-
sumption of H2O2 are fitted by zero-order and first-order re-
actions, respectively.[36,38] The apparent formation rate con-
stants of H2O2 were 0.016, 0.013, and 0.012 mmminꢀ1 for
these systems with increasing BA concentration (inset in
Figure 3), indicating that the formation rate of H2O2 does
not change significantly among these systems. By contrast,
the consumption rate constants were found to be greatly de-
pendent on the concentration of the substrate (0.054, 0.027,
and 0.017 minꢀ1, respectively), that is, the higher concentra-
tion of BA leaded to the slower consumption of H2O2. As
stated above, in the photocatalytic system, H2O2 decompos-
Scheme 2. Surface reduction of H2O2.
ꢀ
one-electron reduction of H2O2 by ecb leads to the breaking
of O O bond and formation of OH and OH.[40] Because of
ꢀ
ꢀ
C
the stronger coordination ability of OHꢀ with surface Ti
sites, surface adsorbed OHꢀ and free OH, which may dif-
C
fuse to some extent away from the surface, should be
formed. This argument was corroborated by earlier stud-
ies.[41] For example, Choi found that the OH generated from
C
the reduction of H2O2 on the TiO2 surface can be released
into the gas phase and initiate remote photocatalytic oxida-
tion.[41a] Salvador concluded that the free OH may only be
C
generated by the electroreduction of dissolved O2 via
H2O2.[41b] On the other hand, the hydroxylation could also
+
C
be initiated by hvb , or the OH from it, and generates the
product with H2O-derived hydroxyl O atoms, as shown in
Scheme 1. In the first process, hvb can only oxidize the sub-
+
strate adsorbed on the surface of TiO2. In the second pro-
ꢀ
+
cess, the OH formed by hvb oxidation of H2O/OHꢀ should
C
es mainly by means of the ecb reduction process on the sur-
face of TiO2 [Eq. (5)]. Therefore, the inhibitory effect of
substrate on the consumption of H2O2 could be rationalized
by the competitive adsorption on TiO2 between H2O2 and
BA.[5c,39]
be adsorbed on the TiO2 surface, and its desorption proved
to be difficult.[42] Thus, in both processes, the hvb+-initiated
hydroxylation takes place only on the surface of TiO2. Ac-
cordingly, the reaction on the surface tends to incorporate
the O atom from H2O into the product, while the hydroxyl-
ation of the unadsorbed substrate leads to the formation of
the product containing O2-derived O atoms. Because the ad-
sorption of BA is much stronger than that of benzene,[23] the
surface process would take a higher proportion in the hy-
droxylation of BA than in that of benzene, while the reac-
tion of unadsorbed substrate is more important for benzene.
Therefore, BA-OH has lower 16O (from aerial 16O2) abun-
dance than phenol. Furthermore, in the presence of ben-
To validate that, the effect of BA on the adsorption of
H2O2 on TiO2 was examined (Figure S6 in the Supporting
Information). The inhibitory effect of BA on the adsorption
of H2O2 was significant. For instance, with a total concentra-
tion of 0.15 mm H2O2, the adsorbed amount of H2O2 de-
creased from 5 mmolgꢀ1 in the presence of 5 mm BA to
3 mmolgꢀ1 for 25 mm BA solution (saturated). As a result,
the presence of BA markedly hindered the degradation of
H2O2 (Figure S7 in the Supporting Information). These ex-
periments also corroborate that the direct photolysis of
H2O2, compared with its surface reduction, is not important
in the photocatalytic systems. Since the reductive degrada-
zene, the competition of benzene for free 16OH with BA
C
could lower slightly the proportion of 16O in BA-OH
(Table 1, entries 1 and 2), since BA and benzene have simi-
ꢀ
C
tion of H2O2 by ecb is the major pathway by which the O
lar reaction rate constants with OH (kBA =(4.3ꢁ0.8)ꢁ
atom from O2 incorporates into the hydroxylated product,
the inhibition of the adsorption and degradation of H2O2 at
high BA concentration will hinder the O2 incorporation. On
the other hand, at high substrate concentration, the ad-
sorbed amount of BA will increase. The surface oxidation
process initiated by hvb+, which incorporates the O atom of
H2O into the hydroxylated product, will play a more impor-
tant role. All these facts favor the oxygen incorporation
from water (H218O), and block the pathway from 16O2, so
the hydroxylated product with lower 16O proportion is
formed.
Finally, we will explain the difference in oxygen source
between the hydroxylated products of BA and benzene ac-
cording to our H2O2 reduction mechanism. It was reported
that H2O2 binds in a monodentate fashhion to the surface Ti
atom through its O atom.[39a,b] As shown in Scheme 2, the
109 mꢀ1 sꢀ1 and kbenzene =(4.3ꢁ0.9)ꢁ109 mꢀ1 sꢀ1).[27,43] Similarly,
the presence of BA could enhance strongly the abundance
of 16O in phenol, because of the competition of BA with
+
benzene for the surface hvb
18OH.
C
/
Further, if the hydroxylation processes of BA and ben-
zene occurred only on the surface of TiO2, the adsorption
ability of substrates, which affects their distribution among
surface and bulk solution, could not have an effect on the
contribution fractions of diverse pathways in the whole hy-
droxylation process. The only possible explanation to the
oxygen isotope abundance difference between BA-OH and
phenol would be the different relative importance of the
direct hvb oxidation pathway (only from H218O, path a in
+
Scheme 1) and the OH addition pathway (from both H218O
C
and 16O2, path b in Scheme 1). Because BA and benzene
have nearly identical rate constants for their reactions with
2036
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2030 – 2039

Photocatalytic Hydroxylation
FULL PAPER
C
OH (see above), and benzene is more easily oxidized than
cies that introduces the O2-derived O atoms into the hy-
droxylated product. The findings in the present study also
reveal that, during the photocatalytic oxidation, hvb is not
BA (one-electron oxidation potential: EBA0 =2.6 V and
Ebenzene0 =2.1 V vs. NHE),[44] the direct oxidation pathway
should be more important for benzene than for BA. It
means that the 18O abundance of phenol (from H218O)
would be higher than that of BA-OH, which is completely
contradictory to our experimental observations shown in
Table 1. On the other hand, if the hydroxylation processes
of BA and benzene only took place away from the surface
of TiO2 (in the bulk solution), no difference in isotope abun-
dance between BA-OH and phenol would be measured
+
the only active species that initiates the oxidation. Even in
the absence of hvb+, the substrate can still be oxidized by
the OH formed from the reduction of O2 by ecbꢀ. The con-
C
tribution of this pathway cannot be neglected for the photo-
catalytic oxidation, and its proportion is comparable to that
of hvb+, at least for the hydroxylation of aromatic com-
pounds. Our experiments also show that the photocatalytic
oxidation products of the adsorbed substrates bear less
oxygen from oxidant O2 than those of the unadsorbed ones.
(i.e., 16O%
= O%
= O%·OH). It is not consistent
phenol
16
16
BAꢀOH
with our experimental results either. Therefore, to explain
the observed isotope abundance difference between BA-OH
and phenol, both the surface reaction and bulk solution re-
action should be considered at the same time.
Other supports for the different oxygen sources of the
photocatalytic hydroxylation between the surface reaction
Experimental Section
Materials: TiO2 (P25, ca. 80% anatase, 20% rutile; surface area, ca.
50 m2 gꢀ1) was kindly supplied by the Degussa Company (Clausthal-Zel-
lerfeld, Germany). H218O was purchased from Jiangsu Changshu Chemi-
cal Limited (Changshu, Jiangsu, P.R. China), the isotope abundance of
which was 98%. 18O2 (18O: 97%) was purchased from Cambridge Isotope
Laboratories, Inc (Andover, USA). The mass spectrometry analysis of
the post-reaction solvent/atmosphere showed that the oxygen isotope
abundance of the 18O-enriched reagent did not change significantly (<
2%) in all the reactions. Benzoic acid (BA), benzene, nitrobenzene, ben-
zonitrile, and their hydroxylated products were all of analytical grade,
and obtained from the Beijing Chemical Company (Beijing, P.R. China).
5,5-Dimethyl-1-pyrroline-N-oxide (DMPO), a spin trapping reagent for
ESR measurements, was supplied by Sigma–Aldrich Co. (Shanghai, P.R.
China). Horseradish peroxidase (POD), which was used in the measure-
ment of H2O2, was purchased from the Huamei Biologic Engineering Co.
(Luoyang, Henan, China), while N,N-dialkyl-p-phenylenediamine (DPD)
was from Merk (p.a.) (Whitehouse Station, NJ, USA). Chlorotrimethylsi-
lane (TMSCl) and 1,1,1,3,3,3,-hexamethyldisilazane (HMDS) were pur-
chased from Acros Organics (Beijing, P.R. China). All reagents were
used as received without further purification.
C
and bulk solution reaction came from the free OH quench-
ing experiments (Table S2 in the Supporting Information).
Both tert-butyl alcohol (tBuOH) and isopropanol (iPrOH)
[5a,45]
C
were reported to capture preferably free OH,
because
their low adsorption on TiO2 makes them unfavorable to
+
C
compete with BA for hvb or surface-bound OH (details of
the estimation and discussion about the proportions of the
C
OH captured by tBuOH or iPrOH on the TiO2 surface and
in the bulk solution are shown in Tables S3 and S4 in the
Supporting Information). In the spin-trapping ESR experi-
ment, the addition of iPrOH markedly weakened the
C
trapped OH signal with the appearance of the signal of
DMPO-C(OH)Me2 (Figure 2; DMPO=5,5-dimethyl-1-pyr-
roline-N-oxide, a spin-trapping reagent used in the ESR
measurements), indicating the quenching effect of iPrOH on
Photocatalytic reactions: A 100 W Hg Lamp (ToshibaIn SHL-100UVQ)
was used as the light source for the photocatalytic reaction. In a typical
H218O-isotope-labeled reaction, TiO2 (P25, 2 mg) was dispersed in H218O
(1 mL) with a given concentration of BA under aerated conditions (16O2).
Prior to irradiation, the suspension was magnetically stirred in the dark
for about 30 min to ensure the establishment of an adsorption/desorption
equilibrium. After irradiation for a definite time, the oxygen isotope
abundance of the hydroxylated product was analyzed by HPLC-ESI
method (Agilent LC 1200/Ion Trap 6310) with C-18 column (250 mmꢁ
2.1 mm). Each measurement was repeated at least three times to assure
the accuracy. The standard deviation of MS analysis is estimated to be ca.
0.6%. To minimize the disturbance caused by the change of substrate
concentration and the formation of intermediates, only the initial stage
(0–2 h, <7% of substrate conversion) of photocatalytic oxidation was
studied. Figure S1 in the Supporting Information gives the typical HPLC
chromatograms (acquired by UV and MS detectors) and ESI-MS spectra.
The measured isotope abundance of the product was corrected with the
oxygen isotope abundance of solvent H2O and the natural isotope abun-
dance of the product [Eqs. (10) and (11)] in which Cp, Cn, and Cw are the
18O percentages of the measured isotope abundance of the product, natu-
ral isotope abundance of the product, and measured isotope abundance
of solvent H2O, respectively.
C
OH. Accordingly, the isotope labeled experiments showed
that the presence of tBuOH and iPrOH decreased the pro-
portion of 16O in the hydroxyl group of BA-OH from 18.2%
to 13.9% and 11.8%, respectively, for the 1 h irradiated sys-
tems (Table S2, entries 2 and 3), indicating again the impor-
tant role of free OH, which mainly initiates the hydroxyl-
ation of unadsorbed substrates, in the oxygen incorporation
from O2.
C
Conclusion
We studied the photocatalytic hydroxylation of aromatics by
18O-labeling method. The experiments show that 18O-isotope
labeling is a powerful method to identify the oxygen incor-
poration pathway in the photocatalytic hydroxylation. Our
experimental results do not support the mechanism that the
addition of molecular O2 onto the cationic radical or a radi-
cal HO adduct of the substrate could incorporate the O
atom of O2 into the hydroxylated product. In contrast, we
show that the reduction of O2 by ecb is indispensable for
the O2 incorporation. The formed intermediate H2O2 plays
an essential role in the O2 incorporation process. The OH
generated from the reduction of H2O2 is the final active spe-
Cp ꢀ Cn
ð10Þ
ð11Þ
H2O% ¼
ꢂ 100
Cw ꢀ Cn
ꢀ
Cw ꢀ Cp
Cw ꢀ Cn
O2% ¼
ꢂ 100
C
In the experiments using 18O2 as oxidant, the photocatalytic oxidation of
Chem. Eur. J. 2012, 18, 2030 – 2039
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2037

C. Chen et al.
BA was performed in Pyrex vessels (10 mL) with H216O suspension
(5 mL) containing BA (20 mm) and TiO2 (P25; 2 gLꢀ1 ). Before the 18O2
reaction, the suspension in the vessel underwent twenty cycles of pump-
ing to vacuum and then purging with argon to remove the air, followed
by saturation with 1 atm of 18O2. After reaction, the oxygen isotope abun-
dance of the hydroxylated product was measured with the same method
as that of H218O labeled experiments, and the correction was also applied
[Eqs. (12) and (13)] in which CO is the 18O abundance of 18O-enriched O2.
[10] H. Kaur, B. Halliwell, Oxygen Radicals in Biological Systems Pt C
1994, 233, 67–82.
[11] D. L. Sedlak, A. W. Andren, Environ. Sci. Technol. 1991, 25, 777–
782.
[12] a) G. Palmisano, M. Addamo, V. Augugliaro, T. Caronna, A. Di Pao-
la, E. G. Lopez, V. Loddo, G. Marci, L. Palmisano, M. Schiavello,
Catal. Today 2007, 122, 118–127; b) H. Yoshida, H. Yuzawa, M.
Aoki, K. Otake, H. Itoh, T. Hattori, Chem. Commun. 2008, 4634–
4636.
CO ꢀ Cp
[13] a) R. W. Matthews, J. Chem. Soc. Faraday Trans. 1 1984, 80, 457–
471; b) G. Palmisano, M. Addamo, V. Augugliaro, T. Caronna, E.
Garcia-Lopez, V. Loddo, L. Palmisano, Chem. Commun. 2006,
1012–1014.
[14] S. Goldstein, G. Czapski, J. Rabani, J. Phys. Chem. 1994, 98, 6586–
6591.
ð12Þ
ð13Þ
H2O% ¼
ꢂ 100
CO ꢀ Cn
Cp ꢀ Cn
CO ꢀ Cn
O2% ¼
ꢂ 100
In the measurement of the isotope abundance of the hydroxyl O atoms
of the hydroxylated products formed in the photocatalytic oxidation of
benzene, nitrobenzene, and benzonitrile, the reacted suspension was ex-
tracted with ethyl ether (5 mL). The organic phase was dried with anhy-
drous Na2SO4, and evaporated to about 0.1 mL by Ar purging. The con-
centrated sample was treated with HMDS (100 mL) and TMSCl (50 mL),
and then centrifuged. The TMS derivations of phenols were analyzed by
[15] a) M. A. Barakat, J. M. Tseng, C. P. Huang, Appl. Catal. B 2005, 59,
99–104; b) V. Sukharev, R. Kershaw, J. Photochem. Photobiol. A
1996, 98, 165–169; c) L. Rideh, A. Wehrer, D. Ronze, A. Zoulalian,
Ind. Eng. Chem. Res. 1997, 36, 4712–4718.
[16] a) D. F. Ollis, E. Pelizzetti, N. Serpone, Environ. Sci. Technol. 1991,
25, 1522–1529; b) D. D. Dionysiou, M. T. Suidan, I. Baudin, J.-M.
Laꢃnꢄ, Appl. Catal. B 2004, 50, 259–269.
[17] G. A. Russell, J. Am. Chem. Soc. 1957, 79, 3871–3877.
[18] a) J. Schwitzgebel, J. G. Ekerdt, H. Gerischer, A. Heller, J. Phys.
Chem. 1995, 99, 5633–5638; b) E. R. Carraway, A. J. Hoffman,
M. R. Hoffmann, Environ. Sci. Technol. 1994, 28, 786–793; c) B.
Wen, Y. Li, C. C. Chen, W. H. Ma, J. C. Zhao, Chem. Eur. J. 2010,
16, 11859–11866.
[19] N. Narita, T. Tezuka, J. Am. Chem. Soc. 1982, 104, 7316–7318.
[20] M. Zhang, Q. Wang, C. C. Chen, L. Zang, W. H. Ma, J. C. Zhao,
Angew. Chem. 2009, 121, 6197–6200; Angew. Chem. Int. Ed. 2009,
48, 6081–6084.
GC-MS (Thermo-Finingan; Trace 2000/Trace DSQ) with
a DB-5MS
column (30 mꢁ0.25 mm).[7a] The MS peaks corresponding to [MꢀCH3]+
were used to calculate the oxygen isotope abundance.
The concentration of H2O2 formed during the reaction was measured by
the spectrophotometric DPD method.[46] Electron spin resonance (ESR)
signals of the radicals trapped by DMPO were detected at ambient tem-
perature with a Brucker ESP 300 E spectrometer, under the reaction
conditions that were the same as those in the isotope labeled reactions
except the addition of DMPO. A given amount of suspension (30 mL)
was collected at a given time and filled into a quartz capillary for ESR
measurement. The settings for the ESR spectrometer were as follows:
center field, 3443 G; sweep width, 100 G; microwave frequency,
9.64 GHz; modulation frequency, 100 kHz; power, 10.05 mW.
[21] The photocatalytic reactions of aromatics involve a series of rather
complicated processes, such as cleavage of the aromatic rings, decar-
boxylation, and hydroxylation. In the present study, we only focus
on the hydroxylation process, but other processes are not consid-
ered.
[22] It should be pointed out that three (ortho-, meta-, and para-) posi-
tions on BA are available for the hydroxyl to add on. Depending on
the adding position, the isotope abundance of the hydroxylated
products differs slightly from each other. For example, 31.7, 30.8,
and 36.8% of 16O-hydroxyl were measured after 1 h of irradiation in
the para-, meta-, and ortho-hydroxylbenzoic acids, respectively,
when the initial concentration of BA was 3 mm. In this paper, only
the general incorporation amount, which reflects the origin of hy-
droxyl oxygen in the products, is considered. This slight difference
cannot affect the discussion. Accordingly, the incorporation is repre-
sented by the average isotope ratios among the three isomers.
[23] a) Y. Suda, Langmuir 1988, 4, 147–152; b) Y. H. Hsien, C. F. Chang,
Y. H. Chen, S. F. Cheng, Appl. Catal. B 2001, 31, 241–249; c) D.
Gumy, S. A. Giraldo, J. Rengifo, C. Pulgarin, Appl. Catal. B 2008,
78, 19–29.
[24] a) T. Tan, D. Beydoun, R. Amal, J. Photochem. Photobiol. A 2003,
159, 273–280; b) H. Liu, A. Imanishi, Y. Nakato, J. Phys. Chem. C
2007, 111, 8603–8610; c) K. Imamura, S.-i. Iwasaki, T. Maeda, K.
Hashimoto, B. Ohtani, H. Kominami, Phys. Chem. Chem. Phys.
2011, 13, 5114–5119; d) S.-H. Yoon, S.-E. Oh, J. E. Yang, J. H. Lee,
M. Lee, S. Yu, D. Pak, Environ. Sci. Technol. 2009, 43, 864–869;
e) S. Rengaraj, X. Z. Li, Chemosphere 2007, 66, 930–938.
[25] a) X. Li, W. S. Jenks, J. Am. Chem. Soc. 2000, 122, 11864–11870;
b) L. E. Manring, M. K. Kramer, C. S. Foote, Tetrahedron Lett. 1984,
25, 2523–2526.
Acknowledgements
Financial support from 973 project (No. 2010CB933503), NSFC (No.
20920102034, 21137004, and 20907056), and CAS is grateful acknowl-
edged.
[1] U. I. Gaya, A. H. Abdullah, J. Photochem. Photobiol. C 2008, 9, 1–
12.
[2] A. Kudo, Y. Miseki, Chem. Soc. Rev. 2009, 38, 253–278.
[3] G. Palmisano, V. Augugliaro, M. Pagliaro, L. Palmisano, Chem.
Commun. 2007, 3425–3437.
[4] M. R. Hoffmann, S. T. Martin, W. Y. Choi, D. W. Bahnemann,
Chem. Rev. 1995, 95, 69–96.
[5] a) C. Minero, G. Mariella, V. Maurino, D. Vione, E. Pelizzetti, Lang-
muir 2000, 16, 8964–8972; b) V. Brezovꢂ, M. Ceppan, E. Brandste-
terova, M. Breza, L. Lapcik, J. Photochem. Photobiol. A 1991, 59,
385–391; c) M. Mrowetz, E. Selli, J. Photochem. Photobiol. A 2006,
180, 15–22.
[6] a) M. Pera-Titus, V. Garcia-Molina, M. A. Banos, J. Gimenez, S. Es-
plugas, Appl. Catal. B 2004, 47, 219–256; b) D. Vione, C. Minero, V.
Maurino, A. E. Carlotti, T. Picatonotto, E. Pelizzetti, Appl. Catal. B
2005, 58, 79–88.
[7] a) X. Li, J. W. Cubbage, T. A. Tetzlaff, W. S. Jenks, J. Org. Chem.
1999, 64, 8509–8524; b) M. Canle, J. A. Santaballa, E. Vulliet, J.
Photochem. Photobiol. A 2005, 175, 192–200.
[8] T. D. Bui, A. Kimura, S. Ikeda, M. Matsumura, J. Am. Chem. Soc.
2010, 132, 8453–8458.
[26] G. V. Buxton, C. L. Greenstock, W. P. Helman, A. B. Ross, J. Phys.
Chem. Ref. Data 1988, 17, 513–886.
[27] R. Wander, P. Neta, L. M. Dorfman, J. Phys. Chem. 1968, 72, 2946–
2949.
[28] P. Neta, R. E. Huie, A. B. Ross, J. Phys. Chem. Ref. Data 1988, 17,
1027–1284.
[9] a) K. Ishibashi, A. Fujishima, T. Watanabe, K. Hashimoto, Electro-
chem. Commun. 2000, 2, 207–210; b) H. Czili, A. Horvath, Appl.
Catal. B 2008, 81, 295–302.
2038
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2030 – 2039

Photocatalytic Hydroxylation
FULL PAPER
[29] a) P. Neta, V. Madhavan, H. Zemel, R. W. Fessenden, J. Am. Chem.
Soc. 1977, 99, 163–164; b) V. Madhavan, H. Levanon, P. Neta,
Radiat. Res. 1978, 76, 15–22.
[40] a) X. Z. Li, C. C. Chen, J. C. Zhao, Langmuir 2001, 17, 4118–4122;
b) A. Fujishima, X. T. Zhang, D. A. Tryk, Surf. Sci. Rep. 2008, 63,
515–582.
[30] L. Dogliotti, E. Hayon, J. Phys. Chem. 1967, 71, 2511-2516.
[31] a) D. Salari, A. Niaei, S. Aber, M. H. Rasoulifard, J. Hazard. Mater.
2009, 166, 61–66; b) T. K. Lau, W. Chu, N. J. D. Graham, Environ.
Sci. Technol. 2007, 41, 613–619.
[41] a) J. S. Park, W. Choi, Langmuir 2004, 20, 11523–11527; b) P. Salva-
dor, J. Phys. Chem. C 2007, 111, 17038–17043.
[42] D. Lawless, N. Serpone, D. Meisel, J. Phys. Chem. 1991, 95, 5166–
5170.
[32] L. Cermenati, P. Pichat, C. Guillard, A. Albini, J. Phys. Chem. B
1997, 101, 2650–2658.
[43] L. M. Dorfman, I. A. Taub, R. E. Buhler, J. Chem. Phys. 1962, 36,
3051–3061.
[33] D. P. Riley, Chem. Rev. 1999, 99, 2573–2588.
[44] a) M. Jonsson, J. Lind, T. Reitberger, T. E. Eriksen, G. Merenyi, J.
Phys. Chem. 1993, 97, 11278–11282; b) C. Hansch, A. Leo, R. W.
Taft, Chem. Rev. 1991, 91, 165–195; c) R. G. Pearson, J. Am. Chem.
Soc. 1986, 108, 6109–6114.
[45] a) Y. F. Sun, J. J. Pignatello, Environ. Sci. Technol. 1995, 29, 2065–
2072; b) S. Kim, W. Choi, Environ. Sci. Technol. 2002, 36, 2019–
2025; c) M. A. Ferguson, M. R. Hoffmann, J. G. Hering, Environ.
Sci. Technol. 2005, 39, 1880–1886; d) P. Kormali, A. Troupis, T. Tri-
antis, A. Hiskia, E. Papaconstantinou, Catal. Today 2007, 124, 149–
155.
[34] A. D. Bokare, W. Choi, Environ. Sci. Technol. 2009, 43, 7130–7135.
[35] a) T. Hirakawa, Y. Nosaka, Langmuir 2002, 18, 3247–3254; b) T.
Daimon, Y. Nosaka, J. Phys. Chem. C 2007, 111, 4420–4424.
[36] A. J. Hoffman, E. R. Carraway, M. R. Hoffmann, Environ. Sci. Tech-
nol. 1994, 28, 776–785.
[37] a) T. Hirakawa, C. Koga, N. Negishi, K. Takeuchi, S. Matsuzawa,
Appl. Catal. B 2009, 87, 46–55; b) B. Jenny, P. Pichat, Langmuir
1991, 7, 947–954.
[38] C. Kormann, D. W. Bahnemann, M. R. Hoffmann, Environ. Sci.
Technol. 1988, 22, 798–806.
[39] a) W. F. Huang, P. Raghunath, M. C. Lin, J. Comput. Chem. 2011,
32, 1065–1081; b) J. Kiwi, M. Graetzel, J. Mol. Catal. 1987, 39, 63–
70; c) Q. Guo, I. Cocks, E. M. Williams, Surf. Sci. 1997, 393, 1–11;
d) S. Tunesi, M. A. Anderson, Langmuir 1992, 8, 487–495.
[46] H. Bader, V. Sturzenegger, J. Hoigne, Water Res. 1988, 22, 1109–
1115.
Received: November 2, 2011
Published online: January 20, 2012
Chem. Eur. J. 2012, 18, 2030 – 2039
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
2039