Ligand-to-metal charge transfer of a pyridine surface complex on TiO2for selective dehydrogenative cross-coupling with benzene

Article abstract of DOI:10.1039/d1cp00496d

Dehydrogenative cross-coupling (DCC) between pyridine and benzene proceeded selectively using a TiO2 photocatalyst under visible light irradiation at optimized concentrations of the substrates. Visible light induces ligand-to-metal charge transfer (LMCT) between pyridine and a TiO2 surface to give a pyridine radical cation, which produces a pyridyl radical by its deprotonation or oxidation of another pyridine molecule. The pyridyl radical attacks a benzene ring to form an sp2C-sp2C bond and a hydrogen atom is subsequently removed to complete DCC. Selective excitation of the pyridine LMCT complex in the presence of an excess amount of benzene would be the key for higher selectivity. This journal is

Full text of DOI:10.1039/d1cp00496d

PCCP
View Article Online
View Journal
PAPER
Ligand-to-metal charge transfer of a pyridine
surface complex on TiO for selective
2
Cite this: DOI: 10.1039/d1cp00496d
dehydrogenative cross-coupling with benzene†
a
a
ab
Shimpei Naniwa,
Hisao Yoshida
Shinichiro Hishitani, Akira Yamamoto
and
ab
*
Dehydrogenative cross-coupling (DCC) between pyridine and benzene proceeded selectively using a
TiO photocatalyst under visible light irradiation at optimized concentrations of the substrates. Visible
light induces ligand-to-metal charge transfer (LMCT) between pyridine and a TiO surface to give a
pyridine radical cation, which produces a pyridyl radical by its deprotonation or oxidation of another
2
Received 2nd February 2021,
Accepted 13th April 2021
2
2
2
pyridine molecule. The pyridyl radical attacks a benzene ring to form an sp C–sp C bond and a hydrogen
atom is subsequently removed to complete DCC. Selective excitation of the pyridine LMCT complex in
the presence of an excess amount of benzene would be the key for higher selectivity.
DOI: 10.1039/d1cp00496d
rsc.li/pccp
improve the selectivity of DCC by TiO
oxidation of the substrate should be avoided. One of the ways to
2
photocatalysis, unselective
1
. Introduction
Photocatalytic organic transformation is a promising reaction suppress unselective oxidation in TiO photocatalysis is to utilize
2
system which can achieve unique and highly difficult reactions ligand-to-metal charge transfer (LMCT) between the TiO surface
2
20,21
such as dehydrogenative cross-coupling (DCC) under ambient and adsorbed molecules.
In this system, a photoformed hole
conditions. DCC is one of the most environmental-friendly and is generated and present on the adsorbed molecule. Since the
1
,2
thus important reactions in organic synthesis. Many DCC oxidative potential of this hole is less positive than that of
systems have been reported both in homogeneous and hetero- an excited TiO photocatalyst, undesired oxidation tends to be
geneous catalyses, usually utilizing heat or light as a driving limited, allowing selective oxidation of the substrate. Although
2
3
–7
force. The reaction often proceeds in an oxidative manner by few studies were available on the utilization of the LMCT system
using an external oxidant such as molecular oxygen, in which to DCC, we have recently reported selective DCC between benzene
8
the byproduct is water. In semiconductor photocatalysis, or pyridine and cyclohexane through LMCT between TiO and the
2
22,23
however, DCC can be achieved in a dehydrogenative manner aromatics.
In our systems, an aromatic radical cation
where the byproduct generated by the LMCT of the aromatic-TiO surface complex
is hydrogen gas, which is a promising energy carrier. Therefore, oxidizes cyclohexane to a cyclohexyl radical. The radical attacks an
9–12
under ambient pressure and temperature,
2
13
2
3
photocatalytic DCC by semiconductor material has great potential aromatic molecule to complete DCC between sp C and sp C
and should be explored. through radical addition and elimination mechanism.
Photocatalytic DCC typically proceeds through a reaction Formation of such surface complexes on the TiO surface has
between radical species generated by hole oxidation of organic been reported with various molecules such as alcohols,
molecules. Especially over TiO , one of the most practical amines,
a
2
24,25
26,27
28
22,23,29,30
sulfides, and aromatics.
Therefore, in order
2
photocatalysts with enough oxidation power to oxidize various to extend the application of the LMCT-based system, we need to
1
4–18
molecules,
radical coupling mechanism,
undesired reactions such as homo-coupling. In order to surface complexation. Moreover, the scope of previous systems
the reaction often proceeds through a radical– investigate how the system performs and whether the selectivity
1
1,12,19
which also allows can be improved even in the presence of multiple adsorbates for
2
3
was limited to cross-coupling between sp C and sp C.
a
2
2
Department of Interdisciplinary Environment, Graduate School of Human and
Environmental Studies, Kyoto University, Yoshida Nihonmatsu-cho, Sakyo-ku,
Kyoto 606-8501, Japan. E-mail: yoshida.hisao.2a@kyoto-u.ac.jp
Cross-coupling between sp C and sp C such as the formation of
an arene–arene linkage is also important in organic synthesis
because of a large number of important compounds containing
the linkage in pharmaceutical, agrochemical, and natural product
b
Elements Strategy Initiative for Catalysts & Batteries (ESICB), Kyoto University,
1
-30 Goryo-Ohara, Nishikyo-ku, Kyoto 615-8245, Japan
31
chemistry. In the present study, we report a new LMCT-based
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
2
2
d1cp00496d
DCC between sp C and sp C in the presence of two adsorbates for
This journal is © the Owner Societies 2021
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
surface complexation: pyridine and benzene. We found that 2.2. Reaction test
selective DCC between these molecules can be achieved through
All of the chemicals were of analytical grade and were used
without further purification. Before a photocatalytic reaction
test, the M/TiO sample (0.1 g) in a Pyrex glass tube (20 mL)
2
LMCT excitation of the pyridine surface complex at optimized
concentrations of the substrates in which an excess amount of
benzene exists.
was subjected to pre-treatment for 30 min under the xenon
lamp light so as to clean its surface. Then, the reactants
were added into the test tube, followed by 10 minutes of Ar
bubbling, and stirring under light for the desired reaction time
2. Experimental section
ꢀ2
2
.1. Catalyst preparation
Various TiO powder samples were employed, which were
donated by the Catalysis Society of Japan and denoted as
at room temperature. The light intensity was 160 mW cm
measured at 415 ꢂ 55 nm in wavelength. After irradiation,
.5 mL of the gaseous phase was collected by an air-tight
2
0
2
ꢀ1
syringe and was analyzed by a GC-TCD (Shimadzu, GC-8A).
The reaction mixture was sampled by a syringe with a
JRC-TIO-2 (anatase phase, 18 m g ), JRC-TIO-6 (rutile phase,
2
ꢀ1
2
ꢀ1
1
(
2
00 m g ), JRC-TIO-7 (anatase phase, 279 m g ), JRC-TIO-9
2
ꢀ1
PTFE filter to separate the M/TiO
2
sample and then
anatase phase, 290–310 m g ), JRC-TIO-12 (anatase phase,
2
ꢀ1
2
ꢀ1
analyzed by a GC–MS (Shimadzu, GCMS-QP5050A). The
amounts of products were determined from the GC–MS cali-
bration curve of the authentic samples procured from the
industries.
90 m g ), JRC-TIO-13 (anatase phase, 59 m g ), and
2
ꢀ1
JRC-TIO-14 (anatase phase, 338 m g ). All metal-loaded TiO2
catalysts were prepared by a photodeposition method using TiO2
powder and an aqueous methanolic solution of an appropriate
metal precursor such as PdCl
2
(Kishida, 99%) or H
powder (4.0 g) was dispersed in
2
PtCl
6
ꢁ6H
2
O
(Wako, 99.9%) as follows. TiO
2
2
.3. Diffuse reflectance UV-vis spectroscopy
The UV-vis spectra of the powder samples were recorded in a
diffuse-reflectance (DR) mode. A Pd/TiO sample (JRC-TIO-14)
was diluted with BaSO by a factor of 100. For samples with an
ion-exchanged water (300 mL) and irradiated using a ceramic
xenon lamp (PE300BUV) for 30 min. Then, methanol (100 mL)
and the desired amount of the metal precursor solution were
added to the suspension, and the contents were magnetically
stirred for 15 min without irradiation, followed by stirring under
the light for 1 h. The amount of the metal precursor solution was
determined so as to fix the metal loading amount to 0.1 wt%.
The product was then filtered by suction, washed with ion-
exchanged water, and dried at 323 K for 12 h to obtain metal-
2
4
adsorbate, 50 mL of pyridine or benzene was added to 0.15 g of
the diluted sample, followed by stirring for 30 min. Then, the
desired amount of the sample (typically 0.13 g) was placed in a
cell to fix the amount of the Pd/TiO
spectrum was recorded using a UV-vis spectrophotometer
Jasco V-570) equipped with an integrating sphere, where BaSO
was used as a reference.
2
in the sample, and the
(
4
2
loaded TiO photocatalysts. The catalysts were referred to as
M/TiO , where M represents Pd or Pt.
2
a
Table 1 Results of reaction tests performed under photoirradiation with the light of various limited wavelength ranges
d
Selectivity to PhPs
(%)
b
c
Products /mmol
Yield of PhPs (%)
Entry
l/nm
Time/h
Metal
PhPs
BPs
BPh
Y
Py
S
Py
S
Be
1
2
3
4
5
6
4350
4400
4422
4400
4400
4400
0.2
0.5
6
2
2
Pd
Pd
Pd
Pd
Pt
1.7
1.6
1.7
5.2
6.1
1.6
0.99
0.17
Trace
0.49
1.1
1.2
0.14
0.13
0.14
0.43
0.51
0.13
44
84
499
84
73
42
73
90
73
73
87
0.32
0.09
0.96
1.1
2
—
0.14
0.12
85
a
2
Reaction conditions: pyridine (0.10 mL, 1.2 mmol) and benzene (1.9 mL, 21 mmol) were used with a M/TiO photocatalyst (JRC-TIO-14, 0.1 g), and
ꢀ2
b
the light intensity was 160 mW cm measured at a wavelength of 415 ꢂ 55 nm. PhPs: total amount of 2-PhP, 3-PhP, and 4-PhP. BPs: total amount
0
0
0
c
d
of 2,2 -BP, 2,3 -BP, and 2,4 -BP. Yield of PhPs based on the introduced pyridine. Selectivity of DCC. The selectivity based on pyridine was
calculated as SPy = [100 ꢃ PhPs (mmol)]/[(PhPs + 2 ꢃ BPs) (mmol)]; the selectivity based on benzene was calculated as SBe = [100 ꢃ PhPs (mmol)]/
[
(PhPs + 2 ꢃ BPh) (mmol)].
Phys. Chem. Chem. Phys.
This journal is © the Owner Societies 2021

View Article Online
PCCP
Paper
3
. Results and discussion
We have previously reported that a Pd cocatalyst on TiO
2
can
act as a catalyst which promotes the radical addition to benzene
3.1. Photocatalytic reaction tests
9
,10,12,22
in some cross-coupling reactions.
Thus, the higher
A Pd/TiO sample was examined for the reaction tests under
2
selectivity of the Pd-loaded sample in the present system would
also be contributed from the presence of Pd metal catalysis.
After optimization of the TiO photocatalyst (Table S3, ESI†),
2
JRC-TIO-14, which has a large specific surface area, was chosen
for the following experiments.
Next, we investigated the effect of the concentration of the
reactants on the reaction under visible light irradiation
photoirradiation with the light of various limited wavelength
ranges (Table 1, entries 1–3). Several products were obtained
such as 2-phenylpyridine (2-PhP), 3-phenylpyridine (3-PhP),
and 4-phenylpyridine (4-PhP) as the DCC products (PhPs) as
well as homo-coupled products formed from pyridine (bipyridyls
0 0 0
(BPs): 2,2 -BP, 2,3 -BP, and 2,4 -BP) and benzene (biphenyl: BPh).
The regioselectivity for the three PhPs was almost identical in
most cases regardless of the irradiation wavelength, i.e., about
(Table 2). We carried out reaction tests with a fixed volume of
the reaction solution (2.0 mL) but with a different ratio of
pyridine and benzene. Compared to the ratio used above
90% of 2-PhP, 4% of 3-PhP, and 6% of 4-PhP were formed. This
is consistent with the C–H bond dissociation energy (BDE) of
pyridine (Table S1, 439.3 kJ mol for site 2 and 468.6 kJ mol
(Table 2, entry 1), an increase of the ratio of pyridine to benzene
ꢀ
1
ꢀ1
decreased SPy but increased SBe (Table 2, entries 2–5), and a
decrease of the ratio increased SPy but decreased SBe (Table 2,
entries 6 and 7). These results were plotted in Fig. 1. A low
concentration of pyridine is necessary to achieve the selective
cross-coupling with high SPy under visible light. The ratio of
for sites 3 and 4, ESI†). Trimers from two pyridine molecules and
one benzene molecule or one pyridine and two benzene
molecules were not observed. The formation of BPh indicates
that benzene was also activated during the reaction. Thus, a
certain number of polymers could be formed although they
cannot be detected by GC. The reaction did not proceed in the
dark or in the absence of the photocatalyst (Table S2 entries 1
and 2, ESI,†), which confirmed that the reaction proceeds
photocatalytically. Concomitant hydrogen production also con-
firmed that the reaction took place dehydrogenatively (Table S2,
entry 3, ESI†).
0
.05 with 1.2 mmol of pyridine and 21 mmol of benzene
provided moderately high values in both selectivities, SPy and
Be, which was thus employed as standard conditions in the
S
present study. This means that selective DCC can be achieved
even in the presence of two adsorbates for surface complexation
by optimizing their concentration. When a long-time reaction
test was carried out under the optimized conditions, higher
selectivities were maintained after 12 h and the yield reached
Under photoirradiation of both UV and visible light (l 4
3
50 nm, Table 1, entry 1), PhPs, BPs and BPh were obtained
unselectively, where the selectivities to PhPs were calculated as
4% and 42% based on the products formed from pyridine (SPy
1
.0% (Fig. 2). Any successive reaction products such as trimers
were still not observed. We also calculated the apparent
quantum efficiency (AQE) based on the following equation:
4
)
and benzene (SBe), respectively. The unselective product formation AQE (%) = 100 ꢃ (amount of PhPs)/(number of incident photons
ꢄ
ꢄ
6
suggests that both pyridyl ( C H N) and phenyl ( C H ) radicals irradiated to the reactor). The AQE at 400 nm was 0.22%.
5
4
5
were formed by the photocatalytic oxidation of the reactants by
2
The bare TiO promoted the DCC reaction (Table 1, entry 6),
holes generated on the UV-light-activated TiO
followed by successive radical-radical coupling.
2
photocatalyst, and TiO2 photocatalyst and these molecules, pyridine and
benzene, do not absorb visible light when they have no inter-
Under visible light (l 4 400 nm, Table 1, entry 2), in actions. These facts propose that other visible-light responsive
contrast, PhPs as cross-coupling products were successfully species must be generated in this system. According to our
23
22
obtained as main products, which led to a higher selectivity previous reports, both pyridine and benzene can form a
Py = 84% and SBe = 73%) than that under irradiation including surface complex with TiO and absorb visible light (l 4 400 nm)
(S
2
UV light, although the production rates were lower due to the
limited photon number for the limitation of wavelength range.
The selectivity was further improved under more-limited-
wavelength light (l 4 422 nm, Table 1, entry 3) such as
Table 2 Results of the reaction tests with various concentrations of
pyridine and benzene under visible light
a
b
499% for SPy and 90% for SBe. This result indicates a different
Selectivity
Ratio of
pyridine to
b
Substrate/mmol
Products /mmol
(%)
reaction mechanism under visible light from that in the
presence of UV light.
Entry Pyridine Benzene benzene
PhPs BPs BPh
S
Py
S
Be
The effect of metal co-catalysts was also investigated
Table 1, entries 4 and 5). Generally, it is considered that metal
c
1
2
3
4
5
6
7
1.2
6.0
12
18
24
0.48
0.24
21
18
12
6.0
0.50
24
0.050
0.25
0.50
0.75
0.98
0.020
0.010
5.2
4.5
3.1
1.5
0.10 9.7
3.4
1.9
0.49 0.96 84
73
87
93
499
(
2.8
5.7
8.9
0.35 45
0.11 22
Trace 10
co-catalyst on photocatalysts can suppress the recombination
of photogenerated electrons and holes to enhance the photo-
Trace 0.50 499
3
2–34
catalytic activity.
Both Pd and Pt loaded samples showed
sample (Table 1, entry 6),
0.23 0.98 88 63
Trace 0.99 499 49
24
higher activity than a pristine TiO
2
a
which suggests that these metals actually work as a charge
separator. At the same time, metal deposition decreased the
Reaction conditions: pyridine and benzene were used with a Pd/TiO
2
photocatalyst (JRC-TIO-14, 0.1 g), the reaction time was 2 h, the
irradiation wavelength was l 4 400 nm, and the light intensity was
selectivity. The Pd/TiO
2
sample showed higher Spy (Table 1,
ꢀ2
b
1
60 mW cmc measured at a wavelength of 415 ꢂ 55 nm. See the caption
entry 4) than the Pt metal loaded sample (Table 1, entry 5). of Table 1. These data were the same as those in Table 1 entry 4.
This journal is © the Owner Societies 2021
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
We first measured DR UV-vis spectra of pyridine or benzene
adsorbed on a Pd/TiO
2
(JRC-TIO-14) sample (Fig. 3). The Pd/
TiO sample was diluted with BaSO for accurate measurement.
2
4
The Pd/TiO2 sample without adsorption of the reactants
showed no photoabsorption beyond 400 nm in wavelength
(
Fig. 3A(a) and B(a)). The sample of pyridine adsorbed on the
Pd/TiO exhibited a large absorption band in the UV region
Fig. 3A(b)), and the additional absorption band of low intensity
2
(
appeared beyond 400 nm (Fig. 3B(b)). The band which centers
around 260 nm is attributed to pyridine molecules adsorbed
physically or in the liquid phase and that centers around
Fig. 1 Selectivities based on pyridine (Spy) or benzene (SBe) in the reaction
tests with various ratios of pyridine to benzene in visible light. Data were
from Table 2.
300 nm and extends beyond 400 nm is LMCT excitation of
the surface complex of adsorbed pyridine and the TiO
2
2
3
surface. On the other hand, the sample adsorbing benzene
Fig. 3A(c)) showed a slightly larger absorption band compared
with the Pd/TiO (Fig. 3A(a)) but much lower than the sample
(
2
adsorbing pyridine (Fig. 3A(b)).
Next, we carried out adsorption tests to compare the abilities
of pyridine and benzene to form a surface complex with TiO2
(
Table 3). Hexane was used as a solvent. When 24 mmol of
pyridine was introduced, more than half of the introduced
pyridine adsorbed on TiO (Table 3, entry 1). When 27 mmol
of benzene was introduced, only 2.0 mmol of benzene adsorbed
Table 3, entry 2). These results confirmed that pyridine is
adsorbed on TiO more preferentially than benzene, which
2
(
2
could be due to the stronger interaction of pyridine with the
surface. We also carried out reaction tests with neat substrates
under photoirradiation (Table 4). In the light including UV
light, both reaction tests gave homo-coupling products, and the
Fig. 2 Time course of the cross-coupling reaction between pyridine and
benzene. Pyridine (0.1 mL, 1.2 mmol) and cyclohexane (1.9 mL, 21 mmol)
2
with the Pd/TiO photocatalyst (0.1 g) were used. The irradiation wave-
length was l 4 400 nm. See the caption of Table 1 for the definitions of ratio of BPs to BPh was 2.2 (Table 4, entries 1 and 2), meaning that
the selectivity of DCC, SPy and SBe
.
photoexcited TiO photocatalysts promote the homocoupling
2
reaction of pyridine with two times higher rate. In visible light,
the amount of the homo-coupling products was much larger
to be excited; the former forms through an acid–base interaction with pyridine (Table 4, entry 1) than with benzene (Table 4,
and the latter does through a p interaction with the surface. entry 2) and the ratio of BPs to BPh was 11 (Table 4, entries 3
Thus, many reaction routes are possibly speculated. Since both and 4). When we consider that the photoinduced DCC proceeds
homo-coupling products (BPs and BPh) were produced, both through adsorption, photoexcitation, and the reaction of the
pyridine and benzene should be activated in the present radical species with other radicals or molecules, the large
conditions. Furthermore, in order to confirm which is the major difference in the ratio of BPs to BPh in the homo-coupling
active surface complex to produce PhPs under visible light reaction tests under the two photoirradiation conditions would
conditions, we carried out the following experiments.
be determined by the difference of the photoexcitation process
Fig. 3 (A) DR UV-vis spectra of (a) the Pd/TiO
2
sample diluted with BaSO
4 2
, (b) pyridine adsorbed on the diluted Pd/TiO sample, and (c) benzene
adsorbed on the diluted Pd/TiO
2
sample. (B) Enlarged view of (A).
Phys. Chem. Chem. Phys.
This journal is © the Owner Societies 2021

View Article Online
PCCP
Paper
a
a
Table 3 Results of adsorption tests
Table 5 Results of isotope experiments
b
Amount in the solution
Initial amount in at the adsorption
the solution/mmol equilibrium/mmol
H D
Product/mmol, (k /k )
Deuterated
compound
Adsorbed
amount /mmol
b
Entry
l/nm
PhPs
BPs
BPh
1
2
3
4
5
6
4350
—
17
4.3
1.5 (2.9)
6.3 (0.70)
0.45
0.27 (1.7)
0.47 (0.96)
3.0
Entry Pyridine Benzene Pyridine Benzene
Pyridine Benzene
C
C
6
D
D
5
N
N
8.8 (1.9)
15 (1.1)
5.2
2.7 (1.9)
4.3 (1.2)
4.1 (0.74)
0.76 (4.0)
0.80
1.1 (0.71)
0.12 (7.0)
1
2
24
0
0
27
11
0
0
25
13
—
—
2.0
6
6
4400
—
C
C
6
D
D
5
a
Conditions: hexane was used as a solvent. The solution (2.0 mL, 12 mM
photocatalyst (JRC-
TIO-14, 50 mg), and then stirred for 1 h without photoirradiation.
6
6
pyridine or 14 mM benzene) was mixed with a TiO
2
a
Reaction conditions: pyridine (0.05 mL, 0.62 mmol) and benzene
photocatalyst (0.1 g) were used,
b
Adsorbed amount was calculated as: (Initial amount in the solution (0.95 mL, 11 mmol) with the Pd/TiO
2
b
(
(
mm mm oo ll )) )) . ꢀ (Amount in the solution at the adsorption equilibrium the reaction time was 1 h. The values in parentheses: k
H D
/k = (amount
of the products in entry 1)/(that in entry 2 or 3) or (that in entry 4)/(that
in entry 5 or 6).
ꢄ
the formation of PhPs, which gives a pyridyl radical ( C
5
H
4
N).
Table 4 Results of the photocatalytic reaction tests with neat pyridine or
benzene under different wavelength light
a
A step of radical formation of benzene should not be included
in the major reaction of the pyridine radical since the C–H
bond cleavage is more difficult in benzene than in pyridine in
b
Products /mmol
Entry l/nm
Substrate BPs
BPh
Ratio of BPs to BPh terms of their C–H BDEs (Table S1, ESI†). Thus, the pyridyl
ꢄ
radical ( C
5
H
4
N) attacks a molecular benzene ring to form an
1
2
3
4
4350 Pyridine
57
n.d.
19
n.d.
26
n.d.
1.7
2.2
3
Benzene
4400 Pyridine
Benzene
sp -like transition state (eqn (1)), which would release a
hydrogen atom to give PhPs (eqn (2)). The small KIE observed
11
n.d.
6 H D
in the reaction with benzene-d (Table 5, entry 3, k /k value
a
Reaction conditions: pyridine or benzene (24 mmol) was used with a
photocatalyst (0.1 g), the reaction time was 2 h, and the light
was 1.1) indicates the presence of another route to form PhPs
which involves the C–H bond cleavage in benzene. KIEs for BPs
in the reaction with pyridine–d5 and for BPh in that with
Pd/TiO
2
ꢀ2
intensity was 160 mW cm measured at a wavelength of 415 ꢂ 55 nm.
b
See the caption in Table 1.
benzene-d (Table 5, entries 2 and 3) also evidenced that RDSs
6
in the formation of BPs and BPh are the C–H bond cleavage in
pyridine and benzene, respectively.
under these light conditions since adsorption and the reaction
would be in common. These results suggest that pyridine can
form the surface complex more easily and produce the homo-
coupling product under visible-light irradiation in the presence
H D
On the other hand, the k /k value was less than unity for
the homo-coupling of non-deuterated compounds (Table 5,
BPh in entry 2 and BPs in entry 3), that is, inverse KIEs were
observed. This can be interpreted by the competitive oxidation
of pyridine and benzene. During the reaction, the two molecules
are supposed to competitively consume the limited number of
of TiO than benzene. Therefore, it is suggested that the major
2
visible-light-excited species in the mixture of pyridine and
benzene would be the pyridine–TiO surface complex. As discussed
2
23
2
previously, pyridine adsorbs on the TiO surface through an acid–
2
photogenerated holes in TiO . Since the C–D bond is stronger
base interaction between the N atom in the pyridine ring and Ti
cation on the surface. At the LMCT, an electron of the lone pair on
the N atom would be excited to a localized 3d orbital on the Ti
cation, and the excited state should be a pyridine radical cation
than the C–H bond, the use of a deuterated compound instead
of the normal compound would slow down the oxidation of the
corresponding compound. This would increase the availability
of the hole for the counter compound and accelerate its
oxidation, resulting in a higher yield of the homo-coupling
products of the non-deuterated compound. Thus, together with
the discussion for the result of PhPs, it is suggested that at least
two processes take place for the cross-coupling in the presence
of UV light; radical addition and elimination between the
pyridyl radical and benzene molecule (eqn (1) and (2)), and
minor radical–radical coupling between a pyridyl radical and
phenyl radical (eqn (3)). Another possible process is radical
addition and elimination between pyridine and a phenyl radi-
cal (eqn (4) and (5)). The lower selectivity in the presence of UV
light (Table 1, entry 1) than that under visible light (Table 1,
entries 2 and 3) suggests that large numbers of pyridyl and
phenyl radical species give these products through many reac-
tion routes.
ꢄ
C H N ) adsorbed on the surface. The visible-light-induced
5 5
+
(
reaction will start with the generation of this surface radical cation
species.
3.2. Isotope experiments
To get an insight into the reaction mechanism under each
wavelength light, we carried out isotope experiments (Table 5).
In the presence of UV light, the yields of PhPs in the reaction
with pyridine–d
the reaction with normal reagents (Table 5, entry 1). The k
5
(Table 5, entry 2) were smaller than those in
/k
H
D
value was greater than unity such as 1.9, which is recognized as
the kinetic isotope effect (KIE). This means that the rate-
determining step (RDS) in the formation of PhPs is the C–H
bond cleavage in pyridine under irradiation with the light
including UV light. This indicates that the C–H bond cleavage
by the photogenerated hole takes place mainly in pyridine for
ꢄ
ꢄ
C
5
H
4
N + C
6
H
6
- [C
5
H
4
NꢀC
6
H
6
]
(1)
This journal is © the Owner Societies 2021
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
ꢄ
ꢄ
ꢄ
[
[
C H NꢀC H ] - C H N (PhPs) + H
(2)
(3)
(4)
(5)
between the radical and benzene. Due to the low concentration
of pyridine, homo-coupling of pyridyl radicals to form BPs
would also be suppressed for the same reason. The low rate
for BPh formation suggests the lower reactivity of the surface
pyridine cation to benzene for phenyl radical formation.
5
4
6
6
11
9
ꢄ
ꢄ
C
5
H
4
N + C
6
H
5
- C
5
H
5
NꢀC
NꢀC
N (PhPs) + H
6 5
H
ꢄ
ꢄ
6 5
H ]
C
5
H
5
N + C
6
H
5
- [C
5
H
5
ꢄ
C
5
H
5
NꢀC
6
H
5
] - C11
H
9
5
The inverse KIE for BPh in the reaction with pyridine–d could
be due to the faster formation of the phenyl radical through the
oxidation of benzene by the pyridine radical cation.
Under visible light, the k
reaction with pyridine–d
Table 5, entry 5), and the k
benzene-d was 7.0 (Table 5, entry 6). These KIEs confirmed
H
/k
were 1.9 and 1.7, respectively
/k value for BPh in that with
D
values for PhPs and BPs in the
5
(
H
D
3.3. Proposed reaction mechanism
6
that RDSs in their formations under visible light are the same Based on all of the results, a tentative reaction mechanism for
as discussed above in the UV light case: RDS in the formation of the photocatalytic DCC between pyridine and benzene is
PhPs and BPs is C–H bond cleavage in pyridine, while RDS in proposed as follows, where the Pd/TiO photocatalyst is shown
2
the formation of BPh is that in benzene. Simultaneously, it was as a representative. Since the different light conditions gave
suggested that the C–H bond dissociation by the hole takes different results, it is obvious that the UV light and the visible
place only in pyridine during the formation of PhPs by light give different reaction mechanisms.
considering their C–H BDEs. On the other hand, the yield of
BPs was not changed in the reaction with benzene-d (Table 5, tribution by the UV light excitation is considered to be very
entry 6, the k /k value was 0.96) while inverse KIE was still predominant (Fig. 4A). The UV light excites the TiO photo-
observed for BPh in the reaction with pyridine–d (Table 5, catalyst to give photogenerated electrons and holes in the
entry 5, the k /k value was 0.71). This different result from the conduction band and valence band, respectively. The electrons
In the light including both UV and visible light, the con-
6
H
D
2
5
H
D
UV light case suggests that the competitive oxidation of are transferred to the Pd nanoparticles, while the holes are
pyridine and benzene is not the case in this condition. Reaction trapped on the TiO surface (step (i)). Pyridine and benzene are
2
under visible light would be initiated by the LMCT excitation of competitively oxidized by the holes to the corresponding radical
the pyridine–TiO surface complex as suggested from the larger species (steps (ii) and (iii)). Then, the radical–radical coupling
2
adsorption amount and higher visible-light activity of pyridine of these radicals affords the DCC products (PhPs, step (iv)) as
than benzene (Tables 3 and 4). The surface radical cation well as the homo-coupling products (BPs and BPh, steps (v) and
produces a pyridyl radical by its deprotonation or oxidation of (vi)). Protons are reduced by electrons on the Pd cocatalyst to
another pyridine molecule, and the radical attacks the benzene give molecular hydrogen (step (vii)). This is a two-photon
ring. The cross-coupling proceeds through a radical addition process (Fig. 4A). Another possible route to form the cross-
and elimination mechanism. The use of an excess amount of coupling product is radical addition–elimination: the addition
3
benzene to pyridine would allow selective addition of the of the pyridyl radical to a benzene molecule to form an sp -
pyridyl radical to benzene due to the high collision frequency centered transition state, and the elimination of a hydrogen
Fig. 4 Schematic image of tentative reaction mechanism for photocatalytic DCC between pyridine and benzene under UV and visible light (A), and
0
visible light (B). 2-PhP and 2,2 -BP were chosen as representative products among PhPs and BPs, respectively.
Phys. Chem. Chem. Phys.
This journal is © the Owner Societies 2021

View Article Online
PCCP
Paper
ꢄ
+
ꢀ
radical from this transition state (steps (viii) and (ix)). In this
case, the proton is reduced by the electron on the Pd nano-
particles, followed by a reaction with the hydrogen radical to
give molecular hydrogen (step (x)). The step (ix) itself would be
thermodynamically favorable since the molecule can regain its
aromaticity by releasing a H radical. Therefore, acceleration of a
H radical elimination by other species or proton elimination by
further photoexcitation of the transition state would not be
necessary. This reaction between the pyridyl radical and benzene
C H NꢀTi
(surface complex) - C H N + TiTiO2 + e
5
5
TiO2
5
5
Pd
(xii)
ꢄ
+
ꢄ
+
C H N - C H N + H
(xiii)
(xiv)
(xv)
5
5
5
4
ꢄ
+
ꢄ
+
C
5
H
5
N
+ C
5
H
5
N - C
5
H
5
N + C
5
H
4
N + H
ꢄ
ꢄ
C H N + C H - [C H NꢀC H ]
5
4
6
6
5
5
6
6
ꢄ
ꢄ
[
C
5
H
5
NꢀC
6
H
6
] - C11
H
9
N (PhPs) + H
(xvi)
(xvii)
+
ꢄ
ꢀ
Pd
9
,10,12,22
H + H + e
- H2
molecule can be catalyzed by the Pd nanoparticles.
Under visible-light irradiation (Fig. 4B), the surface complex
consisting of adsorbed pyridine and a Ti cation, in which an
acid–base N–Ti coordinate bond is formed using the lone pair
on the nitrogen atom (step (xi)), is photoexcited to generate an
excited electron from the lone pair. The electron is injected into
4
. Conclusions
LMCT excitation of the pyridine–TiO2 surface complex by
visible light was utilized to promote DCC between pyridine
and benzene. The presence of an excess amount of benzene is
important to obtain high selectivity. The selective formation of
radical species due to the unique mechanism was successfully
demonstrated. Although the efficiency of the reaction and
product yield are not satisfactory, this study expanded
the possible use of aromatic-semiconductor surface complex
systems for DCC reactions, in which the selectivity can be
improved merely by changing the irradiation wavelength
without any additives or modifications of the photocatalyst.
2
the conduction band of TiO via the surface Ti cation, followed
by transfer to the deposited Pd nanoparticles, while a hole
remains as a pyridine radical cation (step (xii)). The radical
cation is deprotonated (step (xiii)) or oxidizes another pyridine
molecule (step (xiv)) to produce a pyridyl radical and a proton,
and the DCC reaction proceeds through the radical addition–
elimination process between the pyridyl radical and a benzene
molecule (step (xv) and (xvi)). The adsorbed pyridine molecules
would not be static on the surface but rather dynamic in an
adsorption–desorption equilibrium, that is, they act not only as
a visible-light-harvesting group but also as a substrate. Apart
from the LMCT sites, the proton is reduced by the electron on
the Pd nanoparticles, followed by a reaction with the hydrogen Conflicts of interest
radical to give molecular hydrogen (step (xvii)). The reaction
under visible light should be a one-photon process (Fig. 4B).
There are no conflicts to declare.
In this mechanism, pyridine radical cation forms selectively
instead of the photogenerated holes, suppressing the formation
of phenyl radical. Therefore, the homo-coupling of benzene is
Acknowledgements
This work was supported by ISHIZUE 2020 of Kyoto University
suppressed. In addition, the attack of pyridine radical to the
benzene molecule is dominant due to an excess amount of
benzene, suppressing the homo-coupling of pyridine.
Under UV light
Research Development Program, the joint research program of
the Artificial Photosynthesis, Osaka City University, Grant-in-Aid
for Challenging Research (Exploratory, 20K21108) from the Japan
Society for the Promotion of Science (JSPS), and the Program for
Element Strategy Initiative for Catalysts & Batteries (ESICB,
JPMXP0112101003) commissioned by the MEXT of Japan.
2
+ hn - e
ꢀPd
+ h TiO2
+
(i)
(ii)
Pd/TiO
+
ꢄ
+
C
5
H
5
N + h TiO2 - C
5
H
H
4
5
N + H
+
ꢄ
+
C
6
H
6
+ h TiO2 - C
6
+ H
(iii)
(iv)
(v)
References
ꢄ
ꢄ
C
5
H
4
N + C
6
H
5
- C11
H
9
N (PhPs)
(BPs)
12 (BPh)
1
C. S. Yeung and V. M. Dong, Chem. Rev., 2011, 111, 1215.
ꢄ
ꢄ
C
5
H
4
N + C
5
H
4
N - C10
H
8
N
2
2 C. Li, Acc. Chem. Res., 2009, 42, 335.
3
4
5
S. A. Girard, T. Knauber and C. J. Li, Angew. Chem., Int. Ed.,
014, 53, 74.
J. Xie, H. Jin, P. Xu and C. Zhu, Tetrahedron Lett., 2014,
5, 36.
ꢄ
ꢄ
C
6
H
6
+ C
6
H
6
- C12
H
(vi)
(vii)
(viii)
(ix)
(x)
2
+
ꢀPd
2H + 2e
- H
2
5
ꢄ
ꢄ
6 6
H ]
C
5
H
4
N + C
6
H
6
- [C
5
H
5
NꢀC
M. Rueping, J. Zoller, D. C. Fabry, K. Poscharny,
ꢄ
ꢄ
R. M. Koenigs, T. E. Weirich and J. Mayer, Chem. – Eur. J.,
[
C
5
H
5
NꢀC
6
H
6
] - C11
H
9
N (PhPs) + H
2012, 18, 3478.
+
ꢄ
ꢀPd
H + H + e - H
2
6 H. Yi, L. Niu, C. Song, Y. Li, B. Dou, A. K. Singh and A. Lei,
Angew. Chem., Int. Ed., 2017, 56, 1120.
Q. Liu, Y. N. Li, H. H. Zhang, B. Chen, C. H. Tung and
L. Z. Wu, Chem. – Eur. J., 2012, 18, 620.
Under visible light
N + TiTiO2 - C
7
C
5
H
5
5
H
5
NꢀTiTiO2 (surface complex) (xi)
This journal is © the Owner Societies 2021
Phys. Chem. Chem. Phys.

View Article Online
Paper
PCCP
8
C. Liu, J. Yuan, M. Gao, S. Tang, W. Li, R. Shi and A. Lei, 22 A. Yamamoto, T. Ohara and H. Yoshida, Catal. Sci. Technol.,
Chem. Rev., 2015, 115, 12138. 2018, 8, 2046.
H. Yoshida, Y. Fujimura, H. Yuzawa, J. Kumagai and 23 S. Naniwa, A. Tyagi, A. Yamamoto and H. Yoshida, Phys.
T. Yoshida, Chem. Commun., 2013, 49, 3793. Chem. Chem. Phys., 2018, 20, 28375.
0 A. Tyagi, T. Matsumoto, T. Kato and H. Yoshida, Catal. Sci. 24 S. Higashimoto, N. Kitao, N. Yoshida, T. Sakura, M. Azuma,
Technol., 2016, 6, 4577. H. Ohue and Y. Sakata, J. Catal., 2009, 266, 279.
1 A. Tyagi, A. Yamamoto, T. Kato and H. Yoshida, Catal. Sci. 25 S. Zhu, S. Liang, Y. Wang, X. Zhang, F. Li, H. Lin, Z. Zhang
Technol., 2017, 7, 2616. and X. Wang, Appl. Catal., B, 2016, 187, 11.
2 E. Wada, T. Takeuchi, Y. Fujimura, A. Tyagi, T. Kato and 26 S. Higashimoto, Y. Hatada, R. Ishikawa, M. Azuma,
9
1
1
1
1
H. Yoshida, Catal. Sci. Technol., 2017, 7, 2457.
3 K. Mazloomi and C. Gomes, Renewable Sustainable Energy
Rev., 2012, 16, 3024.
Y. Sakata and H. Kobayashi, Curr. Org. Chem., 2013,
17, 2374.
27 X. Lang, W. Ma, Y. Zhao, C. Chen, H. Ji and J. Zhao, Chem. –
Eur. J., 2012, 18, 2624.
14 N. Hoffmann, Aust. J. Chem., 2015, 68, 1621.
1
5 L. E. Oi, M.-Y. Choo, H. V. Lee, H. C. Ong, S. B. A. Hamid 28 X. Lang, W. Hao, W. R. Leow, S. Li, J. Zhao and X. Chen,
and J. C. Juan, RSC Adv., 2016, 6, 108741. Chem. Sci., 2015, 6, 5000.
6 A. Liu, C. Lu, C. Chen, S. Li and D. Ma, Catal. Sci. Technol., 29 T. Kamegawa, H. Seto, S. Matsuura and H. Yamashita, ACS
018, 8, 2030. Appl. Mater. Interfaces, 2012, 4, 6635.
7 Y. Wang, A. Liu, D. Ma, S. Li, C. Lu, T. Li and C. Chen, 30 T. Kamegawa, S. Matsuura, H. Seto and H. Yamashita,
Catalysts, 2018, 8, 355. Angew. Chem., Int. Ed., 2013, 52, 916.
8 S. Park, J. Jeong, K. I. Fujita, A. Yamamoto and H. Yoshida, 31 J. Hassan, M. S ´e vignon, C. Gozzi, E. Schulz and M. Lemaire,
J. Am. Chem. Soc., 2020, 142, 12708. Chem. Rev., 2002, 102, 1359.
9 A. Tyagi, A. Yamamoto, M. Yamamoto, T. Yoshida and 32 C. M. Wang, A. Heller and H. Gerischer, J. Am. Chem. Soc.,
H. Yoshida, Catal. Sci. Technol., 2018, 8, 2546. 1992, 114, 5230.
0 G. Zhang, G. Kim and W. Choi, Energy Environ. Sci., 2014, 33 V. Subramanian, E. E. Wolf and P. V. Kamat, J. Am. Chem.
, 954. Soc., 2004, 126, 4943.
1 D. Franchi and Z. Amara, ACS Sustainable Chem. Eng., 2020, 34 J. Yang, D. Wang, H. Han and C. Li, Acc. Chem. Res., 2013,
, 15405. 46, 1900.
1
1
1
1
2
2
2
7
8
Phys. Chem. Chem. Phys.
This journal is © the Owner Societies 2021