An unexpected fluctuating reactivity for odd and even carbon numbers in the TiO2-based photocatalytic decarboxylation of C2-C6 dicarboxylic acids

Article abstract of DOI:10.1002/chem.201303236

The degradation behaviours of five straight-chain dicarboxylic acids (from ethanedioic acid to hexanedioic acid) were compared in aqueous TiO ₂-based photocatalysis. When all other conditions were identical, the degradation rates were found to fluctuate regularly with the parity of the number of carbon atoms. Dicarboxylic acids with an even number of carbon atoms (e-DAs) always degraded more slowly than those acids with an odd number of carbon atoms (o-DAs). This unusual fluctuation in the reactivity for the degradation of dicarboxylic acids by TiO₂-based photocatalysis is very closely related to the different pre-coordination modes of the acids with the photocatalyst. Attenuated total reflection FTIR (ATR-FTIR) of e-DAs labelled with ¹³C showed that both carboxyl groups of the acid coordinate to TiO₂ through bidentate chelating forms. In contrast, only one carboxyl group of the o-DAs coordinated to TiO₂ in a bidentate chelating manner, whereas the other formed a monodentate binding linkage. The bidentate chelating form with bilateral symmetric coordination did not favour degradation. Isotope-labelling experiments were performed with ¹⁸O₂ to observe the different ways in which incorporated oxygen entered the initial decarboxylated products of e- and o-DAs. For the degradation of butanedioic acid, (45.9±0.5) % of the oxygen in the formed propanedioic acid came from H₂O, whereas for pentanedioic acid, (97.4±0.2) % of the oxygen in the formed butanedioic acid came from H₂O. Our results demonstrate that in TiO₂-based photocatalysis, the reactivity of active species, such as ^.OH/h _vb⁺, is far from non-selective and that the attacks of these active species on organic substrates are significantly affected by the coordination patterns of the substrates on the TiO₂ surface.

Full text of DOI:10.1002/chem.201303236

DOI: 10.1002/chem.201303236
Full Paper
&
Photocatalysis
An Unexpected Fluctuating Reactivity for Odd and Even Carbon
Numbers in the TiO -Based Photocatalytic Decarboxylation of
2
C2-C6 Dicarboxylic Acids
[
a]
Yiran Sun, Wei Chang, Hongwei Ji, Chuncheng Chen, Wanhong Ma,* and Jincai Zhao*
Abstract: The degradation behaviours of five straight-chain
dicarboxylic acids (from ethanedioic acid to hexanedioic
dentate chelating manner, whereas the other formed a mon-
odentate binding linkage. The bidentate chelating form with
bilateral symmetric coordination did not favour degradation.
acid) were compared in aqueous TiO -based photocatalysis.
2
1
8
When all other conditions were identical, the degradation
rates were found to fluctuate regularly with the parity of the
number of carbon atoms. Dicarboxylic acids with an even
number of carbon atoms (e-DAs) always degraded more
slowly than those acids with an odd number of carbon
atoms (o-DAs). This unusual fluctuation in the reactivity for
Isotope-labelling experiments were performed with O to
2
observe the different ways in which incorporated oxygen en-
tered the initial decarboxylated products of e- and o-DAs.
For the degradation of butanedioic acid, (45.9Æ0.5)% of the
oxygen in the formed propanedioic acid came from H O,
2
whereas for pentanedioic acid, (97.4Æ0.2)% of the oxygen
the degradation of dicarboxylic acids by TiO -based photoca-
in the formed butanedioic acid came from H O. Our results
2
2
talysis is very closely related to the different pre-coordina-
tion modes of the acids with the photocatalyst. Attenuated
demonstrate that in TiO -based photocatalysis, the reactivity
of active species, such as COH/hvb , is far from non-selective
and that the attacks of these active species on organic sub-
strates are significantly affected by the coordination patterns
2
+
13
total reflection FTIR (ATR-FTIR) of e-DAs labelled with
C
showed that both carboxyl groups of the acid coordinate to
TiO2 through bidentate chelating forms. In contrast, only
of the substrates on the TiO surface.
2
one carboxyl group of the o-DAs coordinated to TiO in a bi-
2
À
+
Introduction
by ecb /hvb of a TiO photocatalyst is completely changed by
2
+
the introduction of dioxygen to the process (Scheme 1a); h
vb
À
Photo-assisted heterogeneous reactions on TiO particles have
begins oxidative decarboxylation to form CO and e partici-
2
2
cb
attracted much attention because of their promising applica-
tions in solar energy conversion, water splitting and treatment
pates in nÀ1 RH alkylation in accordance with the cooperative
formation mechanism. In the presence of dioxygen, the major
decarboxylated product is the nÀ1 acid rather than the nÀ1
hydrocarbon. Acid formation is commonly attributed to the
capture of the active alkyl radical by dissolved dioxygen and
[
1–3]
of environmental contamination.
Under UV irradiation, most
surface redox processes are performed by photo-induced elec-
À
+
tron–hole pairs (e /hvb ) on the surfaces of the semiconduc-
cb
[
4–6]
[15]
tors.
There is a consensus about the initial occurrence of
the subsequent reaction between these species. From this
À
+
charge separation and recombination of ecb /hvb on the sur-
face of the semiconductor, but to date insight into, and deep
comprehension of, the subsequent induced chemical reactions,
especially those involving the cleavage and formation of vari-
traditional perspective, the oxygen of the carboxylic group
within any formed acid intermediates should have two sour-
ces: dioxygen in air and water in the solvent (Scheme 1a, with
O2).
[
7,8]
ous chemical bonds, is lacking. For example, in the degrada-
tion decarboxylation of organic carboxylic acids in water, the
The participation of the latter source of oxygen is possible
but is variable because the oxidative hydrolysis of the sub-
strate or the COH radical reaction are involved to different ex-
final step by which the organic compounds release CO still re-
2
mains unclear. Generally, the photo-Kolbe reaction mediated
tents, whereas the former source ought to be present in the
formed acid product; otherwise, the alkane with the nÀ1 RH
product should also be formed in the Kolbe reaction without
dioxygen participation. However, our recent work using
[
a] Y. Sun, Dr. W. Chang, Dr. H. Ji, Dr. C. Chen, Prof. W. Ma, Prof. J. Zhao
Beijing National Laboratory for Molecular Sciences
Key Laboratory of Photochemistry
[9]
oxygen-18 labelling strangely showed that in the formed car-
boxyl group of the acid there was no oxygen from dioxygen
during the initial period. All of the oxygen in the formed acid
intermediate came from water in the solvent. The proportion
coming from dioxygen gradually increased, however, from 0 to
finally ~40% until the reaction had undergone approximately
Institute of Chemistry, Chinese Academy of Sciences
Beijing 100190 (China)
E-mail: whma@iccas.ac.cn
jczhao@iccas.ac.cn
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201303236.
Chem. Eur. J. 2014, 20, 1861 – 1870
1861
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Scheme 1. a) Traditional photo-Kolbe reaction with and without dioxygen. b) Our previous observation of the step-by-step cleavage of the C
monocarboxylic acids. c) A possible pre-coordination process allowing the incorporation of dioxygen into the product during the decarboxylation of a-keto
acids. Right inset: average rates of degradation as a function of number of carbon atoms for both full conversion (&) and total organic carbon (TOC) removal
1
ÀC
2
bonds of
(
2
~) of the four monocarboxylic acids by TiO -based photocatalysis. The initial concentrations of the four acids were 2 mm in a volume of 30 mL, and the sur-
2
[3,9,10–16].
face area of the P25 film was 6 cm in all experiments.
5
0% conversion (Scheme 1b, acetic acid case). Coincidentally,
greater difficulty of converting the methyl group of acetic acid
into the required a-keto acid intermediate than the methylene
group of propanoic acid. However, for aliphatic straight-chain
monocarboxylic acids with four or more carbon atoms, there
were no significant differences in their decarboxylation rates
because of the similarities in both the reactivity of the methyl-
ene group and the coordination of the a-keto acid to TiO2
(right inset of Scheme 1).
the hydroxylation of benzene and benzoic acids by photocatal-
ysis with TiO also revealed that the oxygen composition in the
2
18
formed phenol from H O decreased as the conversion in-
2
[
9,17]
creased.
For a single product, this variation of the oxygen
source with reaction progress strongly indicated that the reac-
tion of the substrate with the active species, such as COH radi-
cals and O C/COOH radicals, was not unalterable as presumed
À
2
by the traditional view of diffusive control. A reasonable ex-
planation for such an alternation could be competitive reac-
tions controlled by changes in the interactions between the
In previous work, there are almost no exceptions to the de-
pendence of the decarboxylation process on the carbon chain
length for the degradation of straight-chain monocarboxylic
[9]
TiO surface sites and the substrates (including solvents such
acids (as described by the inset in Scheme 1). Such a unidirec-
tional dependence of the transformation on the number of
carbon atoms suggests that the formation of the a-keto acids
2
as H O), intermediates and products. Different coordination po-
2
tentials for the substrates with the TiO surface sites might
2
À
+
play a key role in revealing the initial reactions with ecb /hvb
.
or the coordination of these acids with the TiO surface plays
2
+
Indeed, our IR experiment with H O and O labelled with the
a key role to dictate the COH radical/hvb breaking of the C ÀC
2
2
1
2
oxygen-18 isotope identified an a-keto acid intermediate,
a common precursor shared by subacids, an intermediate that
emerged during the degradation of aliphatic straight-chain
bond of the initial acids. Additionally, one end of the carboxylic
group exclusively chelates with TiO and prevents the remote
2
CÀC or CÀH bonds at the other end of the acid from accessing
[
9]
monocarboxylic acids. Tentatively, an underlying bidentate
the active species of TiO . In particular, such coordination by
2
[
9]
coordination of the nascent a-keto acid intermediate with Ti
sites was proposed to be the most convenient precursor to aid
in the incorporation of dioxygen into the split of C ÀC bond,
a monocarboxylic donor allows only the a-CÀH bond of the
acid to be preferentially oxidised; therefore, the cleavage of
every C ÀC bond of the acid takes place step-by-step (see
1
2
1
2
[
18]
likely through a Crigee rearrangement (Scheme 1c). The car-
boxylic acids with shorter carbon chain lengths, such as acetic
acid, should show slower degradation than those with longer
chains, such as propanoic acid. This effect was attributed to
Scheme 1b). Consequently, every inferior acid is always a mon-
ocarboxylic acid rather than a dicarboxylic acid, aldehyde acid
or hydroxyl acid. In this way, is the degradation always linearly
dependent on the number of carbon atoms if the substrates
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1862
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
have multiple coordinative groups within a straight chain? Is
the reactivity of TiO -based photocatalysis of a straight-chain
2
dicarboxylic acid, for example, the simplest case of multiple co-
ordination sites, similar to that of straight-chain monocarboxyl-
ic acids? Few theoretical predictions and little experimental
evidence have been reported on this issue.
Analogous to the studies of the degradation of monocar-
[
9]
boxylic acids by TiO -based photocatalysis, in this work,
2
a range of dicarboxylic acids will be employed (from ethane-
dioic to hexanedioic acid) to investigate their degradation be-
haviours. Some excellent studies have been performed to com-
pare the adsorption of dicarboxylic acids, including ethanedioic
(
oxalic), propanedioic and butanedioic acid, on the surface of
[
19,20]
TiO under dark conditions.
The role of adsorption modes
2
Figure 1. Average rates for both full conversion (&) and TOC removal (~)
by TiO -based photocatalysis for the five dicarboxylic acids as a function of
2
their carbon number. The initial concentration of every acid was 0.5 mm in
in photocatalytic events was also reported in relation to buta-
[20b]
nedioic acid and its mono- and di-hydroxyl derivatives.
In
essence, decarboxylation required direct carboxylic group–TiO
2
a volume of 50 mL, and 25 mg of P25 TiO was added. P25 TiO was pre-
2
2
[
19b]
contact.
Most characterisation results from attenuated total
treated according to the method described in ref. [23]. Similar to our prior
[
9]
work on carboxylic acids, the pH value of each acid solution was not ad-
justed during the reaction period.
reflection (ATR-FTIR) spectra indicated that oxalic acid typically
exhibited bidentate coordination by both of the carboxyl
groups on the TiO surface, whereas propanedioic acid was sig-
2
nificantly different from oxalic acid, in that one carboxyl had
bidentate coordination and the other had monodentate coor-
In sharp contrast to the linear carbon-number-dependent re-
activity of monocarboxylic acids (inset of Scheme 1), the degra-
dation rates of dicarboxylic acids showed an abnormal fluctua-
tion as the carbon number increased. Those species with an
odd number of carbon atoms (o-DAs) degraded faster than
those with an even number of carbon atoms (e-DAs). The deg-
radation of oxalic acid was somewhat faster than expected be-
cause of the auto-decomposition of oxalic acid under UV irradi-
[
20]
dination with the TiO2. Inspired by these fine characterisation
studies, in this work we systematically correlate the photocata-
lytic transformation and mineralization of dicarboxylic acids
with the different pre-coordination patterns on the TiO sur-
2
face. An unusual oscillating reactivity is revealed for dicarboxyl-
ic acids as a function of the number of carbon atoms. Through
13
[21]
ATR-FTIR spectra of dicarboxylic acids with the C isotope as
well as the performance of degradation experiments with iso-
ation . Although every dicarboxylic acid had different inter-
mediates (Table 2) and these intermediates probably affected
the transformation rates of the initial substrates, every single
dicarboxylic acid maintained this fluctuated reactivity until
total transformation. This peculiar aspect of the degradation
18
topically labelled O , we examine whether the same biden-
2
tate binding with the TiO surface by two carboxyl groups de-
2
termines whether dioxygen or H O participates in the splitting,
2
À
+
induced by ecb /hvb , of C ÀC bond(s).Our results demonstrate
reminded us to simultaneously observe the CO formation rate
1
2
2
the importance of understanding the pre-coordination of sub-
as a function of the carbon number of the substrates
strates with TiO and the effect of this pre-coordination on the
roles of dioxygen and COH radical/hvb in photocatalytic reac-
(Figure 1). Surprisingly, the fluctuation pattern described by
2
+
[21]
CO formation was perfect even for oxalic acid. The order of
2
tions.
mineralization for dicarboxylic acids is clearly completely differ-
ent from that of monocarboxylic acids in TiO -based photoca-
2
talysis. In principle, the decarboxylation of organic acids fol-
+
lowed the direct single-step reaction of hvb oxidation de-
[
9,12–16]
Results and Discussion
scribed by the photo-Kolbe reaction mechanism.
If the
+
other conditions were the same, consequently, the hvb yield
was constant; thus, the time required for full decarboxylation
per carboxylic group should be the same. The longer the
carbon chain of the acid was, the longer the time required was
for full mineralization. This case was similar to most homoge-
nous oxidations that have been examined (see the next sec-
Typical fluctuating reactivity as a function of number of
carbon atoms for straight-chain dicarboxylic acids
To determine the nature of the decarboxylation process by
TiO -based photocatalysis as generally as possible, we selected
2
five straight-chain dicarboxylic acids (from oxalic acid to hexa-
nedioic acid) to systematically compare their reactivities in
identical aqueous solutions. The changes in the concentration
of every substrate and generated intermediate during the
course of the reaction were measured by ion chromatography
2À
tion on S O8 photooxidation). However, this fundamental in-
2
ference had never been clearly displayed by the degradation
of straight-chain monocarboxylic acids in TiO -based photoca-
2
talysis. Quite the opposite was true; the sequence of reactivity
was acetic acid<propionic acid<butanoic acid<pentanoic
acid (inset of Scheme 1). This result at least illustrated that radi-
(IC) at optimum conditions (see the Experimental Section and
Figure 4 for details). Figure 1 gives the average rate for full
conversion of every dicarboxylic acid as a function of the
number of carbon atoms under identical reaction conditions.
+
cal oxidation by COH/hvb was neither the control step of de-
carboxylation nor the chief contributor to the decarboxylation
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1863
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
of monocarboxylic acids in the heterogeneous system. This op-
posite trend in reactivity relative to homogenous oxidation in-
dicated that the surface sites of the catalyst play a crucial role.
The formation of the a-keto acid followed by bidentate chelat-
ing coordination on the TiO surface was thought most likely
2
to determine the rate of the degradation of monocarboxylic
[
9]
acids. However, the degradation of dicarboxylic acids seem-
ingly did not follow the mechanism of either homogenous or
heterogeneous TiO -based photocatalytic oxidation of mono-
2
carboxylic acids. To our knowledge, the consistent fluctuation
patterns for the rates of substrate transformation and CO for-
2
mation as a function of the alternation between odd and even
carbon numbers has never been reported or attracted atten-
tion. Apparently, this unique reactivity, compared with the deg-
radation of monocarboxylic acids, arose exclusively from the
Figure 2. Average rates for both full conversion (&) and TOC removal (~)
2À
of five dicarboxylic acids by S
2
O
8
photooxidation as a function of carbon
[
20b,c]
number. The initial reactive solution (50 mL) for the degradation of each di-
carboxylic acid sample contained 0.5 mm dicarboxylic acid and
addition of another carboxylic group.
There are two possi-
ble ways that a dicarboxylic acid can cause such an oscillation
in reactivity as a function of the carbon number: i) through the
variety of inherent redox nature of dicarboxylic acids with the
carbon chain fluctuations; ii) through significantly different co-
ordination modes of substrates with odd and even carbon
1
2 2 8
5 mmK S O .
were not influenced by any properties of the substrate except
for the carbon number. Furthermore, this suggested that the
numbers with the TiO surface. The former was easily excluded
coordination of the substrates to the TiO surface altered the
2
2
by the following control experiments. Thus, it is likely that the
oxidation sequence despite the fact that the oxidising species
+
two carboxylic units in dicarboxylic acids bind to the TiO sur-
such as the COH radical/h
were more powerful than the
2
vb
À
face and dramatically alter the conventional route, namely that
SO C radical and could react with the substrate very fast. Thus,
4
+
active species, such as COH radical/hvb , dominate TiO -based
we tentatively attributed this fluctuating reactivity to the dif-
ferences in the pre-coordination states of the two carboxylic
2
photocatalysis.
groups on the TiO surface between e-DAs and o-DAs.
2
2
À
S2O8 photooxidation
Nature of pre-coordination of dicarboxylic acids on the TiO2
2
À
S2O8 photooxidation is a classic one-electron-oxidation pro-
surface
2À
cess that takes place through homogenous cleavage of S O
into the highly active SO C radical under UV light and is usual-
ly used to degrade organic pollutants in water. The oxidative
2
8
À
Intrinsically, the coordination model was correlated with the
distribution of the electronic charge density of the whole mol-
ecule and this distribution had great impact on the oxidative
degradation of the molecule. The carboxylic groups at both
ends of dicarboxylic acid molecules made the location of the
4
[
22]
À
activity of the SO C radical is so high that once formed, the
4
À
SO C radical reacts quickly with the substrate regardless of the
4
redox potential of the substrate. Therefore, for this homolo-
gous oxidation, the time for the entire transformation process
for an equivalent of any substrate, that is, the average conver-
sion rate, definitely depended on the length of the carbon
chain of the substrates. The concentration changes during the
course of the reactions of five dicarboxylic acids were mea-
sured by IC. Figure 2 illustrates the average rates for both the
complete conversion and the mineralization of the five dicar-
framework of the molecule on the surface of the TiO conspic-
2
[19,20]
uously different from that of the monocarboxylic acids.
However, this difference could not account for the regular fluc-
tuations in reactivity with the odd and even carbon number of
the homologues. It has been reported that the coordination
models for the two terminal carboxylic groups with TiO could
2
be identical or different, based on the carbon number of the
2À
[19]
boxylic acids as a function of carbon number during S O
dicarboxylic acid.
For example, ATR-FTIR results indicated
2
8
photooxidation. The average degradation rates of the five di-
carboxylic acids decreased almost linearly as the carbon
number increased (except for oxalic acid), whereas the average
mineralization rates slowed significantly as the carbon number
increased. The correlation of both the degradation and the
mineralization with the carbon number of the substrates was
that both of the carboxylic groups of oxalic acid formed a bi-
dentate chelating mode with TiO , whereas only one carboxylic
2
group of propanedioic acid formed a bidentate chelating
mode and the other carboxylic group formed a monodentate
bridge to a TiO site (Scheme 2a and b). To obtain more de-
2
tailed results for the adsorption states on the TiO surface for
2
different from that seen in the TiO -based photocatalysis. Addi-
all five dicarboxylic acids, we carefully studied the pre-coordi-
2
tionally, the features did not agree with the degradation of
nation states of butanedioic and pentanedioic acid on the TiO
2
[19,20]
monocarboxylic acids by TiO -based photocatalysis either. This
surface by using ATR-FTIR.
Figure 3a shows the ATR-FTIR
2
dependence between the reactivity and the carbon chain
length of the substrate demonstrated that the formation of
spectrum of pentanedioic acid and the IR spectrum of penta-
nedioic acid in solution, and Figure 3b shows the ATR-FTIR
À
13
SO C and the subsequent reactions in homogenous solutions
spectrum of butanedioic acid and butanedioic acid with C la-
4
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1864
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Regardless of which pair was true, the values indicated that
D >D and D <D ; D where stands for the difference of
as
sy
as
sy
sy
nas and n in dissolved solution on the TiO2 film, while Das
s
stands for the difference of nas and n in dilute solution on the
s
ZnSe substrate. Thus, the two carboxylic groups in pentanedio-
ic acid were adsorbed on the TiO surface with one group bi-
2
dentate (D >D ) and one group monodentate (D <D ;
as
sy
as
s
Scheme 2d). Similarly, Dolamic et al. reported that the adsorp-
tion mode for the two carboxylate groups in propanedioic acid
Scheme 2. Two modes of adsorption characterized by bidentate chelation
and mono-/bidentate chelation with the TiO
respectively.
2
surface for e-DAs and o-DAs,
on the TiO surface is with one group bidentate and the other
2
[19,20,24]
group monodentate (Scheme 2b ).
As shown in Fig-
ure 3b, butanedioic acid with a different initial concentration
to that previously applied by others was first used to carry out
ATR-FTIR experiments. Distinctly opposite to the case of penta-
À1
nedioic acid, there were only two peaks, 1548 and 1421 cm ,
that appeared as chemical adsorption of the carboxylic groups.
This indicated that the two carboxylic units of butanedioic acid
adopted a single chelating mode when bound to TiO . To con-
2
firm that the two peaks belong to the carboxylic group ad-
13
sorption, C isotopically labelled butanedioic acid (in which
the two carbon atoms of the carboxyl groups were labelled)
was employed in ATR-FTIR experiments. On the substitution of
13
the C atoms, nCOO changed in accordance with Hooke’s law.
Accordingly, we identified that the peaks at 1548 and
À1
1
421 cm of the normal butanedioic acid and the peaks at
À1
13
1
512 and 1390 cm of the C-labelled butanedioic acid be-
longed to the asymmetric and symmetric stretching vibrations
À
of COO , respectively. The difference between the two carbox-
ylate stretching modes, that is, n
À
À
and ns(COO ), was
)
as(COO
À1
1
27 cm represented by D . According to well-established ex-
sy
[
20,24]
perimental and theoretical calculations,
the difference for
À1
butanedioic acid free in solution is 157 cm , represented by
À1
D , and it was smaller than the former difference (127 cm ) .
as
This difference signified that both of the carboxylate groups
coordinated to TiO2 through the same bidentate chelating
[
20,24]
mode (Scheme 2c).
Similar to the pentanedioic acid case,
À1
the peak at 1716 cm was assigned to the physical adsorption
of the carboxylic group. To confirm these adsorption character-
istics, we carried out ATR-FTIR measurements of pentanedioic
acid and butanedioic acid for high initial concentrations (in the
range 0.05 mm to 100 mm; Figure S1 in the Supporting Infor-
mation). All of the features were in agreement with the results
Figure 3. ATR-FTIR spectra of a) pentanedioic acid in solution (bottom) and
adsorbed on the surface of TiO film (top) and b) normal butanedioic acid
2
1
3
(
(
bottom) and butanedioic acid labelled with C on both carboxyl groups
top). The concentration of the substrates was 5 mm, except for the penta-
nedioic acid in solution (a, bottom) the concentration of which was
.02 mm.
0
À1
in Figure 3, except for the peak at 1712 cm for pentanedioic
À1
acid (Figure S1b; 1690 cm for butanedioic acid, Figure S1c)
belling. In Figure 3a (bottom), the nas(COO
À
and ns(COO )
À
for
that appeared as a non-dissociated ÀCOOH vibration absorb-
)
a sample of dissolved pentanedioic acid (on the ZnSe sub-
ance peak at high concentration. In particular, the ATR-FTIR
À1
13
strate) are 1549 and 1402 cm , respectively, in the normal IR
spectra of C-labelled butanedioic acid with a different initial
À1
spectra; thus, D =147 cm , where n and n stand for the
concentration (Figure S1d) showed only two peaks, 1512 and
as
as
s
À1
wavenumbers of the of the asymmetric and symmetric vibra-
tion of dicarboxylic acid, respectively. However, in the ATR-FTIR
spectrum of pentanedioic acid (absorbed on the coated TiO2
1390 cm , very similar to the low-concentration spectra in Fig-
ure 3b, indicating that the symmetrically bidentate chelating
mode remained largely unchanged with over-loading of the
À1
film), four peaks appeared, at 1641, 1543, 1462 and 1404 cm ,
butanedioic acid on the surface of the TiO film. Thus, all the
2
indicating different chemical adsorption of two carboxylic
e-DAs clearly had their two carboxylic groups adopt the same
bidentate coordination pattern, whereas those of o-DAs dif-
fered, with one group adopting monodentate coordination
and the other group adopting bidentate coordination with the
groups and TiO surface. In these peaks, two pairs of carboxyl-
2
ate vibrations undoubtedly belonged to different binding pat-
terns. Therefore, two different combinations were possible for
À1
À1
the Dsy values: either 237 and 81 cm or 179 and 139 cm .
TiO surface (Scheme 2). Our investigations into the coordina-
2
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1865
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
tion of dicarboxylic acids with the TiO surface agreed with
less than the parent dicarboxylic acid. In the formed nÀ1 dicar-
boxylic acid molecule, one carboxylic group was considered to
be new, whereas the other group was classed as the original
group. The e- and o-DA showed a striking contrast in the way
2
previous work concerning the adsorption in the dark of dicar-
[
19,20]
[25]
[24]
boxylic acids on TiO2,
ZnO, and other oxides. For the
substrates, this alternating coordination as a function of odd
or even carbon number perfectly agreed with the regular fluc-
tuations in the degradation rates. All of the e-DAs degraded
more slowly than the o-DAs. That is, both oxidative degrada-
tion and mineralization of dicarboxylic acids with symmetric bi-
dentate chelating coordination were more difficult than these
processes for acids with asymmetric coordination. Next, we will
look further at the details of the differences in the cleavage of
the C ÀC bond of the dicarboxylic acids with even and odd
H O/dioxygen were combined into their corresponding nÀ1 di-
2
carboxylic acids; nearly all ([97.4Æ0.2]%) oxygen came from
H O for the degradation of pentanedioic acid, whereas (45.9Æ
2
0.5)% came from H O in that of butanedioic acid. In spite of
2
some disputes about the way in which dioxygen/H O partici-
2
[9,17,23,27–29]
pates in the degradation,
the results for different
oxygen sources in the formed carboxylic groups unambiguous-
+
ly indicated the competition between activated H O/h
and
1
2
2
vb
À
carbon numbers through oxygen-18 isotope-labelling experi-
O /e . The asymmetrically chelating donor of the o-DA fav-
2
cb
18
16
2
+
ments ( O /H O). By identifying differences in the incorporat-
oured activated H O/h
vb
attack, whereas the symmetric coor-
2
2
ed oxygen-18 in the initial decarboxylation products for dicar-
boxylic acids with even and odd carbon numbers, we expected
dination model of e-DAs, such as butanedioic acid, significantly
favoured attack by dioxygen. The asymmetrically chelating o-
DA favoured COH radical/hvb decarboxylation; therefore, the
+
À
+
to determine which attack, namely H O/h
or O /e , was
2
vb
2
cb
significantly affected by the two different coordination envi-
ronments.
degradation of these e-DAs was faster than that of symmetri-
cally chelating o-DAs. This result entirely agreed with the fluc-
tuating degradation rates for e-and o-DAs.
Comparison of photocatalytic degradation products be-
tween pentanedioic and butanedioic acid by oxygen-18 iso-
tope-labelling experiments
Identification of a possible degradation process
In our previous work on the degradation of straight-chain ali-
[
9a]
To describe the relationship between the reactivity and coordi-
phatic monocarboxylic acids , the step-by-step C ÀC bond
1
2
nation environment during the TiO -based photocatalytic reac-
cleavage dominated throughout the photocatalytic process.
Every secondary monocarboxylic acid (nÀ1, nÀ2 and nÀ3) was
formed, and no dicarboxylic acid products were observed.
Therefore, the degradation of the monocarboxylic acid began
neither from the end of the alkyl group nor with the cleavage
of any other CÀC bond. This result strongly indicated that the
2
tion in detail, butanedioic and pentanedioic acids were chosen
18
for isotope-labelling experiments with O (95%) to observe
2
18
the O distribution in the product after elimination of one
carbon atom (Table 1).
Blank experiments were performed to examine the oxygen
exchange between the carboxyl groups of the dicarboxylic
+
attack of either COH radical (activated H O by h ) or dioxygen
2
[27]
vb
18
À
acids and O and these showed that such exchange was neg-
(activated by ecb or organic radical RC) exclusively began at
the end with the coordinating carboxylic group. In the present
work, we also tried to identify the degradation products as
much as possible for every dicarboxylic acid by using IC
(Figure 4). The detailed results are summarized in Table 2. We
expected to find more evidence to support the established re-
lationship between the coordination pattern (especially based
on the even or odd carbon number of the dicarboxylic acids)
and the breakage of any C ÀC bond in the presence of multi-
2
ligible (see the Supporting Information, Figure S2). In addition,
similar to the exchange between monocarboxylic acids and
1
8
[9]
18
H2 O, the product acids formed with O did not exchange
16
with H2 O (see the Supporting Information, Figure S3). This
was in very good agreement with earlier studies in both
[
9,26]
homo- and heterogeneous systems
in which little oxygen
exchange took place between dioxygen and any carboxyl, al-
dehyde or alcohol groups. Therefore, we considered the results
1
2
1
8
16
from O /H2 O where the nÀ1 degradation products of dicar-
ple chelating coordination modes. Firstly, oxalic acid was gen-
erated during the degradations of all the other dicarboxylic
acid substrates but did not become the major product, partly
because it was too unstable to accumulate. Secondly, in the
2
boxylic acid were analysed by GC-MS (Table 1).
Table 1 shows the respective oxygen isotopic proportion in
the produced dicarboxylic acid, which had one carbon atom
1
8
16
18
Table 1. Distribution of
2 2 2 2
O isotope in the products formed from pentanedioic and butanedioic acid degradation by TiO in H O solution under an O
atmosphere.
1
6
18
16
18
O
O
O
O
[
a]
O isotopomers [%]
9
7.4Æ0.2
2.6Æ0.3
45.9Æ0.5
54.1Æ0.6
[a] The calculation assumed that one of the carboxyl groups of the produced butanedioic acid or propanedioic acid was inherited from the parent penta-
nedioic acid or butanedioic acid without any change. The difference in the distribution of oxygen isotope appeared only in the other carboxyl group.
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1866
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
carbon chain. For example, the maximum yield of
acetic acid was approximately 76% in the butanedio-
ic acid degradation, and the maximum yield for the
sum of propanoic acid and acetic acid was 90% in
the transformation of hexanedioic acid (Table 2). De-
spite this, these degradation products for o-DAs had
yields of no more than 50%. Apparently, during the
nonocarboxylic acid degradation, the latter case was
more coincident with that from the step-by-step
cleavage of the C ÀC bond to form acids without
2
Table 2. Distribution of products and their maximum concentrations during TiO -
based photocatalytic degradation of five dicarboxylic acids.
e-DAs
Major
products
o-DAs
Major
products
Initial conc.
.5 mM
Max.
conc.
Initial conc.
0.5 mm
Max.
conc.
[mm]
0
[mm]
formic acid
0.060
propanedioic
acid
acetic acid
0.241
oxalic acid
acetic acid
propanedioic
acid
0.065
0.041
propanoic acid 0.077
1
2
the selectivity. Given that the formation of monocar-
boxylic acids exclusively originated from the photo-
Kolbe reaction, it could be speculated that when one
component became a major product, the cleavage of
another CÀC bond, for example, C ÀC for butanedio-
acetic acid
0.380
pentanedioic
acid
acetic acid
0.185
propionic acid 0.132
propanoic acid 0.243
butanedioic
acid
oxalic acid
0.026
propanedioic
acid
0.014
0.077
2
3
ic acid or C ÀC for hexanedioic acid, was likely to
propanedioic
acid
0.090
butanedioic
acid
3
4
become the predominant process. In addition, the
C ÀC bonds of the symmetrically chelating e-DAs
1
2
hexanedioic
acid
propanoic acid 0.218
broke with more difficultly than those of the asym-
metrically chelating o-DAs (Scheme 3). We believe
that the weaker monodentate coordination of the
carboxyl group favoured approach and attack by the
degradation of elementary dicarboxylic acids, such as oxalic
and propanedioic acids, there were small amounts of either di-
carboxylic acid products with larger carbon numbers or some
monocarboxylic acid products with the same carbon number
as the initial substrate. Namely, propanedioic and acetic acid
were produced in the oxalic acid degradation, and propanoic
acid was produced in the propanedioic acid degradation. This
result was in agreement with the extension of the carbon
underlying oxidant. Even dioxygen, as a weakly coordinating
and attacking ligand on the TiO surface, had a fairly high effi-
2
ciency, a fact that was shown by the oxygen isotope labelling
in the products.
Such a carbon-number-dependent reactivity is fundamental
to understanding the most basic oxidative reactions of sub-
+
strates by COH radical/hvb and by dioxygen (Scheme 3). Tradi-
tionally, for homologues, the rates of degradation were
thought to correlate with the redox potential in the homoge-
chain through CO radical or other carbon-centred radical addi-
2
[
15]
[30]
tion reactions. However, generally speaking, these products
were in very low yields. During the degradation process of all
of the five dicarboxylic acids, none of the mono- or dicarboxyl-
ic acid intermediates formed had undergone substitution at
any site of the carbon chain. Similar to the serial degradation
nous solution based on the Marcus equation. Otherwise, if
substrates had similar redox potentials or if the oxidant species
À
were active enough, such as the SO C radical, the rates de-
4
pended on the carbon number. Almost no reports concerned
with coordination on the TiO surface noted the significant in-
2
[
9a]
of monocarboxylic acids, dicarboxylic acids were not seen to
duced regular fluctuations in reactivity related to the odd and
even carbon numbers. Of course, there are still many uncertain
oxidize any other CÀH or CÀC bond except for the C ÀC
1
2
bond. In addition, mono- or di-
carboxylic acid products that
had lost only one carbon atom
were observed in the degrada-
tion of all five initial dicarboxylic
acids. The generation of the nÀ1
monocarboxylic acid was signifi-
cant, and it should arise from
the reduction of the alkyl radical
À
by proton-coupled ecb transfer
or by hydrogen abstraction; this
result clearly demonstrates that
À
+
ecb and hvb simultaneously in-
duced different chemical reac-
tions. More importantly, there
was a notable tendency in e-DAs
to form or accumulate monocar-
boxylic acids from cleavage at
Scheme 3. Schematic diagrams of photocatalytic degradation. a) Pentanedioic acid represents the dicarboxylic
acids with an odd number of carbon atoms and b) butanedioic acid represents the dicarboxylic acids with an
the middle position of the even number of carbon atoms.
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1867
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
straight-chain
monocarboxylic
acids. Our results demonstrate
that the traditional insight that
+
COH radicals/hvb oxidise sub-
strates fast is only associated
with redox property or total
carbon number and cannot be
compared with TiO -based pho-
2
tocatalysis at all. Through ATR-
FTIR experiments and isotope-la-
belling methods, we showed
that the differences in coordina-
tion patterns determined wheth-
er selective cleavage of the C À
1
C bond of the dicarboxylic acids
2
was predominant. The two car-
boxyl groups for e-DAs adopted
symmetrical bidentate chelation
to the TiO surface. This strong
2
coordination meaningfully re-
strained any approach or attack
by an active species, partly be-
cause this coordination model
led to C ÀC bond electron defi-
1
2
ciency, which inhibited oxidative
cleavage. On the other hand, o-
DAs exhibited asymmetrical che-
lation, which permitted all oxi-
dant species, even dioxygen, to
approach and attack the weak-
ened C ÀC bond associated
1
2
with the monodentate chelation
and thereby greatly enhanced
degradation and mineralisation.
All of our results clearly illustrate
that the coordination character-
istics of the substrates’ function-
Figure 4. IC spectra of a) oxalic acid, b) propanedioic acid, c) butanedioic acid, d) pentanedioic acid and e) hexane-
dioic acid during photocatalytic degradation. Degradation times are marked on the right-hand side of each graph
and the retention time of every sample are marked on the bottom of the graph.
al groups with the TiO surface
2
are very important and are
sometimes the main factor that
+
interactions among the chelating groups of the substrates or
the intermediates, moieties remote from the anchored active
sites, and subsequent active species, and these interactions re-
quire investigation with more direct methods to observe
determines the active species, such as hvb /COH radical and O2,
targeted attack on the CÀC or CÀH bond of the substrates.
[31]
single-molecular behaviour in situ on the TiO crystal facet.
Experimental Section
2
Catalyst and chemicals
Conclusion
Both C1-C6 monocarboxylic acids and C2-C6 dicarboxylic acids
were obtained from Alfa Aesar. A silylating reagent consisting of
bistrimethylsilurea (BSU) and hexamethyldisilazane (HMN) was
We revealed a peculiarly fluctuating reactivity for straight-chain
dicarboxylic acids with even and odd carbon numbers (e-DAs
bought from Acros. All the chemicals were of analytical-grade
and o-DAs) in aqueous solution under TiO -based photocataly-
2
18
purity and were used as received. O was purchased from Cam-
2
sis. Among the five dicarboxylic acids investigated (from oxalic
acid to hexanedioic acid), all of the o-DAs degraded faster than
the e-DAs when the photocatalytic conditions were controlled
to be identical. This odd/even dependence strongly differed
bridge Isotope Laboratories, Inc. The isotopic enrichment was
18
9
5%. H2 O was purchased from Jiangsu Changshu Chemical Limit-
ed. The isotopic enrichment was 96%, as determined by mass
spectrometry.
2À
from the characteristics for S O8 homogenous oxidation and
Degussa P25 TiO , containing 80% anatase and 20% rutile phases
2
2
2
À1
for TiO -based photocatalytic degradation of a series of
with a surface area of 50 m g and an average primary particle
2
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1868 ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
size of 25 nm, was used in the photo-assisted reactions. Before
every experiment, P25 was pre-treated according to the method
described in ref. [23].
Photocatalytic tests for isotopic studies
All the isotope-labelling experiments, unless otherwise specified,
were performed in Pyrex vessels (10 mL) with a solution volume of
8
mL. The initial concentration for all the substrate dicarboxylic
À1
Instrumentation
acids was 5 mm, and 0.5 gL of P25 was added. Before use in the
18
O reaction, the solution in the vessel was placed under vacuum
2
The photocatalytic process was analysed by ion chromatography
with a pump, purged with argon six times to remove air, and then
(
IC) and total organic carbon analysis (TOC). Samples of 2.5 mL
18
was saturated with O . After the photocatalytic isotope-labelling
2
were removed and filtered at regular time intervals during the irra-
diation, and 0.3 mL of the samples were analysed by IC. The con-
centrations of carboxylic acids and dicarboxylic acids were mea-
sured by a DX-900 Ion Chromatograph with an IonPac AS23
column (Dionex). The eluent for IC used 13 mm KOH solution. The
reaction was stopped, the solution was extracted with ethyl ether
three times. Then, the ethyl ether was silanized with silylating re-
agents (BSU and HMN). After centrifugation, the supernatant was
removed and was analysed by GC-MS to give the oxygen isotopic
profile in the target degraded products.
À1
flow rate was 0.7 mLmin . The identification of the intermediates
by IC was performed by comparing the retention times with those
of standards. The remaining 2.2 mL solution was used for TOC.
TOC was measured by using a Tekmar Dohrmann Apollo 9000 TOC
Analyser.
ATR-FTIR measurements
According to ref. [26], a layer of water was added dripwise onto
GC-MS analyses were performed with a Finnigan Trace GC ultra
gas chromatograph with a 25 m DB-5 column, coupled with a Finni-
gan Trace DSQ mass spectrometer. The injector port was set for
split operation at 2508C. The temperature program of the column
was as follows: hold at 508C for 4 min, then the temperature was
the surface of the ZnSe crystal coated with a TiO film, the crystal
was scanned, and the background spectrum was obtained. After-
wards, the dicarboxylic acid solution was added dripwise onto the
2
surface and the adsorption equilibrium with the TiO film was at-
2
tained (Figure 5).
À1
increased at a rate of 158Cmin until it reached 2508C and was
maintained at this temperature for 3 min. The isotopic enrichment
of each chemical was examined by EI-MS. The response factors for
the isotopomers of these chemicals were within a small experimen-
tal error; thus, it was assumed that all the isotopomers would have
identical response factors.
The attenuated total reflection Fourier transform infrared spectros-
copy (ATR-FTIR) setup followed ref. [26]. The ATR-FTIR setup con-
sisted of a Harrick Horizon multiple internal reflection accessory
coupled to a 4 mL flow-through cell containing a ZnSe crystal on
the bottom plate and a quartz window on the top plate. The 458
3
internal reflection element of 50ꢁ10ꢁ2 mm allowed approximate-
Figure 5. a) Schematic of the ATR-FTIR cell, including the quartz window on
the top plate (1) and the internal reflection element on the bottom plate (2)
ly 11 infrared bounces. The Fourier transform infrared measure-
ments were performed with a Nicolet 8700 FT-IR equipped with
a DTGS detector.
; b) Representation of the IR path and penetration into the TiO coating. In
2
practice, approximately 11 bounces were allowed by the crystal dimensions
rather than the 4 bounces shown.
Photocatalytic experiments for kinetic studies
TiO film preparation
2
The photocatalytic degradation reactions of the dicarboxylic acids
for the kinetic studies were all performed in Pyrex vessels (50 mL)
with a solution volume of 50 mL. The initial concentration for all
the substrate dicarboxylic acids was 0.5 mm. P25 powder was
À1
P25 was suspended in distilled water at a concentration of 5 gL
.
The suspension was treated for 1 h in a 40 kHz ultrasonic bath and
then ground in an agate mortar for 30 min; 0.5 mL of this suspen-
sion was spread on the ZnSe crystal and was dried for 1 h at
À1
added at 0.5 gL . Irradiation was performed with a 100 W mercury
lamp (Toshiba SHL–100UVQ–2), and the wavelength range was ap-
proximately 310 to 400 nm.
1008C. The conditions used formed uniform thick TiO coatings
that could be used in ATR-FTIR experiments.
2
To reveal the redox nature of all the dicarboxylic acids used, we
performed controlled experiments by using homogenous oxidation
Acknowledgements
2À
of persulfate (S O ), during which UV irradiation was employed
2
8
to homogenously break the peroxide bonds and produce active
4À
2À
We gratefully acknowledge the National Basic Research Pro-
gram of China (973 program Nos. 2010CB933503,
2013CB632405), the National Science Foundation of China
(Grant No. 21137004, 21077110, 21277147), and the Chinese
Academy of Sciences for financial support.
SO C free radicals. This activation of S O by UV irradiation could
2
8
be compared with the TiO /UV photocatalysis in which the possi-
2
ble effects of direct photolysis by UV irradiation on the photocata-
lytic degradation of dicarboxylic acid might be offset. The kinetic
studies concerning the photooxidation of dicarboxylic acids by
2À [22]
)
8
persulfate ion (S O
were also all performed in Pyrex vessels
2
(
50 mL) with a solution volume of 50 mL. The initial concentration
Keywords: chelating coordination
dicarboxylic acids · isotope labelling experiments · TiO
photocatalysis
·
decarboxylation
·
for all the substrate dicarboxylic acids was 0.5 mm, and the initial
concentration of persulfate was 15 mm. The irradiation was also
performed with a 100 W mercury lamp. The eluent was 15 mm
-based
2
À1
KOH solution, the flow rate was 1.0 mLmin in the IC analysis.
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1869
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
[
[
1] A. Fujishima, X. Zhang, D. A. Tryk, Surf. Sci. Rep. 2008, 63, 515–582.
2] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 2001, 293,
[19] a) I. Dolamic, T. Bꢃrgi, J. Phys. Chem. B 2006, 110, 14898–14904; b) R.
Enriquez, A. G. Agrios, P. Pichat, Catal. Today 2007, 120, 196–202.
[20] a) I. Dolamic, T. Bꢃrgi, J. Catal. 2007, 248, 268–276; b) W. Irawaty, D.
Friedmann, J. Scott, P. Pichat, R. Amal, Catal. Today 2011, 178, 51–57;
c) Q. Chen, J. M. Song, F. Pan, F. L. Xia, J. Y. Yuan, Environ. Technol. 2009,
30, 1103–1109.
2
69–271.
3] a) P. Pichat, Photocatalysis and water purification, Wiley-VCH, Weinheim,
013; b) M. R. Hoffmann, S. T. Martin, W. Y. Choi, D. W. Bahnemann,
[
2
Chem. Rev. 1995, 95, 69–96.
[
[
4] a) M. A. Henderson, Surf. Sci. Rep. 2011, 66, 185–297; b) A. L. Linsebigler,
G. Lu, J. T. Yates, Chem. Rev. 1995, 95, 735–758.
5] S. H. Szczepankiewicz, A. J. Colussi, M. R. Hoffmann, J. Phys. Chem. B
[21] The degradation of oxalic acid was seemingly an exception. We believe
that the singular degradation of oxalic acid partly originated from
direct UV-light photolysis because of the typical instability of oxalic acid
and that this photolysis may conceal a slower photocatalytic rate than
that of propanedioic acid (Figure 1a). Thus, we conducted control ex-
periments for direct photocatalysis, and indeed found that oxalic acid
partly degraded under the UV-light irradiation used. After correlation
with the control value, the genuine photocatalytic degradation rate of
2000, 104, 9842–9850.
[6] R. B. Draper, M. A. Fox, Langmuir 1990, 6, 1396–1402.
[7] T. Sumita, T. Yamaki, S. Yamamoto, A. Miyashita, Appl. Surf. Sci. 2002,
2
00, 21–26.
8] P. Fernꢂndez-IbꢂÇez, J. Blanco, S. Malato, F. J. de Las Nieves, Water Res.
003, 37, 3180–3188.
9] a) B. Wen, Y. Li, C. Chen, W. Ma, J. Zhao, Chem. Eur. J. 2010, 16, 11859–
1866; b) Y. Li, B. Wen, C. Yu, C. Chen, H. Ji, W. Ma, J. Zhao, Chem. Eur. J.
012, 18, 2030–2039.
10] D. F. Ollis, E. Pelizzetti, N. Serpone, Environ. Sci. Technol. 1991, 25, 1522–
529.
[
[
2
2
oxalic acid was not faster than that of propanedioic acid. For CO for-
mation, the mineralisation rate of oxalic acid was not correlated with
the control mineralisation experiment.
1
2
[22] B. B. Wang, M. H. Cao, Z. J. Tan, L. L. Wang, S. H. Yuan, J. Chen, J. Hazard.
Mater. 2010, 181, 187–192.
[23] J. Yang, C. Chen, H. Ji, W. Ma, J. Zhao, J. Phys. Chem. B 2005, 109,
21900–21907.
[
1
[
[
11] G. A. Russell, J. Am. Chem. Soc. 1957, 79, 3871–3877.
12] a) S. Yurdakal, G. Palmisano, V. Loddo, V. Augugliaro, L. Palmisano, J.
Am. Chem. Soc. 2008, 130, 1568–1569; b) M. Zhang, C. Chen, W. Ma, J.
Zhao, Angew. Chem. 2008, 120, 9876–9879; Angew. Chem. Int. Ed. 2008,
[24] S. Kang, B. Xing, Langmuir 2007, 23, 7024–7031.
[25] N. J. Nicholas, G. V. Franks, W. A. Ducker, Langmuir 2012, 28, 7189–
7196.
4
1
7, 9730–9733; c) A. M. Khenkin, R. Neumann, J. Am. Chem. Soc. 2008,
30, 14474–14476; d) M. Zhang, Q. Wang, C. Chen, L. Zang, W. Ma, J.
[26] A. R. Almeida, J. A. Moulijn, J. Phys. Chem. C 2008, 112, 1552–1561.
[27] H. Gerischer, A. Heller, J. Phys. Chem. 1991, 95, 5261–5267.
[28] X. Li, J. W. Cubbage, T. A. Tetzlaff, W. S. Jenks, J. Org. Chem. 1999, 64,
8509–8524.
[29] L. F. Liao, C.-F. Lien, D.-L. Shieh, M.-T. Chen, J.-L. Lin, J. Phys. Chem. B
2002, 106, 11240–11245.
Zhao, Angew. Chem. 2009, 121, 6197–6200; Angew. Chem. Int. Ed. 2009,
8, 6081–6084.
13] T. Ohno, S. Izumi, K. Fujihara, Y. Masaki, M. Matsumura, J. Phys. Chem. B
000, 104, 6801–6803.
14] N. Serpone, J. Martin, S. Horikoshi, H. Hidaka, J. Photochem. Photobiol. A
005, 169, 235–251.
4
[
[
[
[
[
[
2
[30] R. A. Marcus, J. Chem. Phys. 1956, 24, 966; R. A. Marcus, Annu. Rev. Phys.
Chem. 1964, 15, 155.
2
15] a) B. Kraeutler, A. J. Bard, J. Am. Chem. Soc. 1978, 100, 5985–5992;
b) M. A. Fox, M. T. Dulay, Chem. Rev. 1993, 93, 341–357.
16] J. Schwitzgebel, J. G. Ekerdt, H. Gerischer, A. Heller, J. Phys. Chem. 1995,
[31] a) C. Xu, W. Yang, Q. Guo, D. Dai, M. Chen, X. Yang, J. Am. Chem. Soc.
2013, 135, 10206–10209; b) M. Shen, D. P Acharya, Z. Dohnꢂlek, M. A.
Henderson, J. Phys. Chem. C 2012, 116, 25465–25469.
9
9, 5633–5638.
17] T. D. Bui, A. Kimura, S. Ikeda, M. Matsumura, J. Am. Chem. Soc. 2010,
32, 8453–8458.
18] R. R. Sever, T. W. Root, J. Phys. Chem. B 2003, 107, 10521–10530.
1
Received: August 16, 2013
Published online on January 8, 2014
Chem. Eur. J. 2014, 20, 1861 – 1870
www.chemeurj.org
1870
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim