Photocatalytic hydrodenitrogenation of aromatic cyanides on TiO2 loaded with Pd nanoparticles

Article abstract of DOI:10.1039/c3cy20748j

Photocatalytic hydrodenitrogenation of aromatic cyanides has been carried out on TiO₂ loaded with Pd nanoparticles (Pd-TiO₂) in the presence of ethanol as a hydrogen source. Photoirradiation of Pd-TiO₂ at λ > 300 nm in ethanol containing aromatic cyanides successfully produces the corresponding toluene derivatives and triethylamine with almost quantitative yields. Photoexcited Pd-TiO₂ catalysts promote oxidation of ethanol and reduction of protons, producing acetaldehyde and hydrogen atoms on the surface of Pd particles (H-Pd species). Aromatic cyanides undergo hydrodenitrogenation via consecutive reactions involving hydrogenation by the H-Pd species and condensation with aldehydes. The catalytic activity strongly depends on the amount of Pd loaded. The catalyst containing 2 wt% Pd, with a relatively low Schottky barrier height at the Pd-TiO₂ heterojunction and a large number of surface Pd atoms, exhibits the highest denitrogenation activity.

Full text of DOI:10.1039/c3cy20748j

Catalysis
Science & Technology
View Article Online
View Journal | View Issue
PAPER
Photocatalytic hydrodenitrogenation of aromatic
cyanides on TiO2 loaded with Pd nanoparticles
Yoshitsune Sugano,a Keisuke Fujiwara,a Yasuhiro Shiraishi,*a Satoshi Ichikawab and
Takayuki Hiraia
Cite this: Catal. Sci. Technol., 2013,
3, 1718
Photocatalytic hydrodenitrogenation of aromatic cyanides has been carried out on TiO2 loaded with Pd
nanoparticles (Pd–TiO2) in the presence of ethanol as a hydrogen source. Photoirradiation of Pd–TiO2 at
l > 300 nm in ethanol containing aromatic cyanides successfully produces the corresponding toluene
derivatives and triethylamine with almost quantitative yields. Photoexcited Pd–TiO2 catalysts promote
oxidation of ethanol and reduction of protons, producing acetaldehyde and hydrogen atoms on the
surface of Pd particles (H–Pd species). Aromatic cyanides undergo hydrodenitrogenation via consecutive
reactions involving hydrogenation by the H–Pd species and condensation with aldehydes. The catalytic
activity strongly depends on the amount of Pd loaded. The catalyst containing 2 wt% Pd, with a
relatively low Schottky barrier height at the Pd–TiO2 heterojunction and a large number of surface Pd
atoms, exhibits the highest denitrogenation activity.
Received 5th November 2012,
Accepted 28th January 2013
DOI: 10.1039/c3cy20748j
www.rsc.org/catalysis
Supported Pd nanoparticles are often employed for catalytic
hydrogenation of various compounds such as olefins6–9 and nitro
1. Introduction
compounds10–12 with molecular hydrogen (H2) as a hydrogen
source. Hydrogenation of aromatic cyanides on Pd catalysts has
also been studied at elevated temperatures.13 These compounds
are successfully converted into toluene derivatives and NH3 via
hydrogenation and subsequent cleavage of the C–N moiety. This
hydrodenitrogenation is promoted by the reaction of aromatic
cyanides with activated hydrogen species (H–Pd), formed via a
dissociative adsorption of H2 on the surface of Pd particles. The
reaction proceeds selectively without decomposition of aromatic
nuclei and enables the reuse of products for upstream processes.
Photoexcitation of semiconductor materials loaded with
noble metal particles creates a positive hole (h+) and conduction
band electrons (eꢁ), and allows subsequent eꢁ transfer to metal
particles.14 Photoexcitation of semiconductors loaded with Pd
particles, when performed in protic solvents such as alcohols,
produces aldehydes via oxidation of alcohols by h+ and H–Pd
species via reduction of protons (H+) by eꢁ on the Pd particles.15
This photocatalytic reaction, if performed with aromatic cyanides,
may successfully promote hydrodenitrogenation by the photo-
formed H–Pd species at room temperature.
Aromatic cyanides are some of the most important chemicals
used as solvents, extractants, and intermediates for the synthesis
of pharmaceuticals, plastics, rubbers, herbicides, and pesti-
cides.1 These compounds are therefore often contained in
industrial effluents. Since they are highly toxic to living
organisms,2 their decomposition and detoxification are very
important tasks. Bioremediation processes have been studied
for the treatment of aromatic cyanides.3 These processes
decompose the compounds into harmless ones such as CO2
and H2O; however, they require restricted operation conditions
in pH and temperature. Photocatalytic decomposition of aromatic
cyanides has also been studied with titanium dioxide (TiO2).4
Photoexcitation of TiO2 in water under O2 produces active oxygen
speꢀcies such as hydroxyl radicals (OHꢀ) or superoxide anions
(O2 ꢁ), and decomposes aromatic cyanides into CO2, NO3ꢁ, and
H2O.5 These oxidative processes, however, completely decompose
the aromatic nuclei of the compounds. Selective decomposition of
the –CN moiety and the reuse of resulting aromatic nuclei for
upstream processes are desirable for the development of green
and sustainable processes.
In the present work, TiO2 loaded with Pd particles (Pd–TiO2)
was used for the photocatalytic reaction of aromatic cyanides
with ethanol under UV irradiation (l > 300 nm). This catalytic
system successfully promotes hydrodenitrogenation of aromatic
cyanides and produces toluene derivatives with high selectivity
at room temperature. The denitrogenation mechanism was
a Research Center for Solar Energy Chemistry, and Division of Chemical
Engineering, Graduate School of Engineering Science, Osaka University, Toyonaka,
560-8531, Japan. E-mail: shiraish@cheng.es.osaka-u.ac.jp; Fax: +81-6-6850-6273;
Tel: +81-6-6850-6271
b Institute for NanoScience Design, Osaka University, Toyonaka, 560-8531, Japan
c
1718 Catal. Sci. Technol., 2013, 3, 1718--1724
This journal is The Royal Society of Chemistry 2013

View Article Online
Paper
Catalysis Science & Technology
clarified, and the effect of the Pd amount on the catalytic 3. Results and discussion
activity was studied.
3.1 Synthesis and characterization of catalysts
The Pdx–TiO2 catalysts with different Pd loadings [x (wt%) =
Pd/(Pd + TiO2) ꢂ 100; x = 0.5, 1, 2, 4] were prepared by
impregnation of Pd(NO3)2 on to TiO2 followed by reduction with
H2. Fig. 1 shows the typical TEM images of respective Pdx–TiO2
catalysts. All catalysts contain spherical Pd particles. The size
of the Pd particles increases with the Pd loadings: the average
diameters for x = 0.5, 1, 2, and 4 catalysts were 2.6, 4.1, 5.0, and
2. Experimental
2.1 Materials
All of the reagents used were purchased from Wako, Tokyo
Kasei, and Sigma-Aldrich, and used without further purifica-
tion. Water was purified by the Milli Q system. JRC-TIO-4 TiO2
(equivalent to Deggusa P25; particle diameter, 24 nm; BET
surface area, 54 m2 gꢁ1; anatase/rutile = ca. 83/17) was kindly
supplied by the Catalyst Society of Japan.
The Pdx–TiO2 catalysts [x (wt%) = Pd/(Pd + TiO2) ꢂ 100;
x = 0.5, 1, 2, 4] were prepared according to the literature
procedure16 as follows: TiO2 (1.0 g) and Pd(NO3)2 (10.9, 21.9,
44.2, or 90.2 mg) were added to water (40 mL), and solvents were
evaporated with vigorous stirring at 353 K for 12 h. The obtained
powders were calcined at 673 K under an air flow (0.5 L minꢁ1
)
and then reduced at 673 K under an H2 flow (0.2 L minꢁ1). The
heating rate and the holding time at 673 K for these treatments
were 2 K minꢁ1 and 2 h, respectively. Ag2–TiO2 and Pt2–TiO2
were prepared in a similar manner to Pdx–TiO2, using AgNO3
(16 mg) or H2PtCl6ꢃ6H2O (54 mg) as a precursor.
Au2–TiO2 was prepared by a deposition–precipitation method
as follows:17 HAuCl4ꢃ4H2O (45.8 mg) was added to water (50 mL).
The pH of the solution was adjusted to 7 by an addition of 1 M
NaOH. TiO2 (1.0 g) was added to the solution and stirred
vigorously at 353 K for 3 h. The particles were recovered by
centrifugation and washed with water. They were calcined at
673 K under an air flow (0.5 L minꢁ1).
2.2 Photoreaction procedure
Each of the respective catalysts (10 mg) was added to ethanol
(5 mL) containing aromatic cyanides within a Pyrex glass tube
(20 cm3; j 16.5 mm). The tube was sealed using a rubber septum
cap and purged with argon gas. Each tube was photoirradiated
with magnetic stirring by an Xe lamp (2 kW; Ushio Inc.). The
temperature of the solution during photoirradiation was 298 K,
and the light intensity was 18.2 W mꢁ2 (at 300–400 nm; through
a water filter). After photoirradiation, the gas phase product was
analyzed by GC-TCD (Shimadzu; GC-8A). The resulting solution
was recovered by centrifugation and analyzed by GC-FID
(Shimadzu; GC-2010), where the products were identified by
GC-MS (EI) (Shimadzu; 40 GCMS-QP5050A).
2.3 Analysis
Total amounts of metals on the catalysts were determined by an
inductively-coupled argon plasma atomic emission spectrometer
(ICAP-AES; SII Nanotechnology, SPS 7800), after dissolution of
catalysts in aqua regia.16 Diffuse reflectance UV-vis spectra were
measured on an UV-vis spectrophotometer (Jasco Corp.; V-550
with an Integrated Sphere Apparatus ISV-469) using BaSO4 as a
reference. Transmission electron microscopy (TEM) observations
were carried out using an FEI Tecnai G2 20ST analytical electron
microscope operated at 200 kV.
Fig. 1 TEM images and size distribution of Pd particles for (a) Pd0.5–TiO2,
(b) Pd1–TiO2, (c) Pd2–TiO2, and (d) Pd4–TiO2 catalysts. The black bars show the
data for fresh catalysts, and the white bars show the data for catalysts recovered
after the 3rd reuse for reactions (Table 2, run 4).
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1718--1724 1719

View Article Online
Catalysis Science & Technology
Paper
Table 1 Hydrodenitrogenation of benzonitrile with the Pd2–TiO2 catalyst
Yield/%
Run
System
Temp./K
Conversion of 1/%
2
3
4
1a
2
H2
H2
H2
Photo
298
313
333
298
84
98
>99
99
15
61
99
99
30
14
0
38
23
0
Fig. 2 High-resolution TEM images of the Pd2–TiO2 catalyst. (a) The incident
beam direction is [0, ꢁ1, 1]. (b) This particle is a twinned particle. The incident
beam directions are [0, ꢁ1, 1] and [0, 1, ꢁ1].
3
4b
0
0
a
Dibenzylamine (1%) was also detected as a product. This is produced
by the nucleophilic attack of a lone pair of nitrogen atoms of benzyl-
amine to an electrophilic carbon atom of the semi-hydrogenated
6.5 nm, respectively. As shown in Fig. 2, high-resolution TEM
images of catalysts revealed that Pd particles can be indexed as
fcc structures, as same as bulk metallic Pd (JCPDS 46-1043).
Fig. 3 shows diffuse-reflectance UV-vis spectra of catalysts.
The higher Pd loading catalysts exhibit increased absorbance
at l > 300 nm due to the light scattering by the Pd particles.15 As
shown in the inset, TiO2 loaded with 2 wt% Pt (Pt2–TiO2)
exhibits spectra similar to that of Pdx–TiO2,18 and TiO2 loaded
with 2 wt% Au (Au2–TiO2) and Ag (Ag2–TiO2) exhibits distinctive
absorption bands at 400–500 nm assigned to localized surface
plasmon resonance.17,19
b
intermediate (ref. 13). After photoreaction, acetaldehyde (197 mmol)
and trace amounts of N,N-diethylbenzylamine (5) were formed.
toluene yields, and the reaction at 333 K facilitates almost
quantitative conversion to toluene (run 3). The photocatalytic
reaction of benzonitrile with alcohol as a hydrogen donor was
carried out. The reaction was performed by photoirradiation
(l > 300 nm) of the solution under argon (1 atm) for 6 h. As
shown in run 4, the reaction facilitates complete conversion of
benzonitrile to toluene at 298 K. This suggests that the photo-
catalytic reaction with alcohol as a hydrogen source successfully
promotes hydrodenitrogenation even at room temperature.
The catalytic activity of Pd–TiO2 for hydrodenitrogenation
was studied with benzonitrile (1) as a model compound.13
Table 1 (runs 1–3) summarizes the results for the hydro-
denitrogenation of benzonitrile with H2 (1 atm) as a hydrogen
3.2 Photocatalytic activity of Pd–TiO2 catalysts
source in the dark. Benzonitrile (2 mM) and Pd2–TiO2 (10 mg) The photocatalytic activity of Pdx–TiO2 with different Pd loadings
were added to an n-hexane solution (5 mL) containing 5% (x) was studied in ethanol. Table 2 summarizes the results
ethanol, and the solution was stirred under H2 (1 atm) for 6 h for photocatalytic reaction of benzonitrile by a 6 h reaction. As
at different temperatures. As shown in run 1, the benzonitrile shown in run 1, pure TiO2 promotes almost no reaction of
conversion at 298 K is 84% but the yield of toluene (2) is benzonitrile. In contrast, increased Pd loadings (runs 2–5)
only 15%, where large amounts of benzylamine (3, 30%) and enhance the reaction and produce toluene with high selectivity
N-ethylbenzylamine (4, 38%) remain. This suggests that the (>95%). Among the catalysts, Pd2–TiO2 exhibits the highest
C–N cleavage is difficult to achieve at this temperature. As denitrogenation activity, and further Pd loading (Pd4–TiO2)
shown in runs 2 and 3, the rise in temperature increases the decreases the activity. Loading of other metal particles is
inefficient for denitrogenation. As shown in runs 6 and 7,
Au2–TiO2 or Ag2–TiO2 promotes almost no reaction of benzo-
nitrile. Pt2–TiO2 (run 8) shows a relatively high conversion of
benzonitrile (45%); however, the toluene yield is only 8%, where
N,N-diethylbenzylamine (5) is produced mainly (36%). These
data indicate that the Pd–TiO2 catalyst, especially loaded with
2 wt% Pd, promotes efficient denitrogenation. It must be noted
that the catalyst is reusable for further reaction. As shown in
run 4, the Pd2–TiO2 catalyst, when reused for further reaction,
shows activity and selectivity similar to those of the virgin
catalyst. In addition, as shown in Fig. 1c (white bars), TEM
analysis of the catalyst recovered after the reaction revealed that
the size of Pd particles scarcely changes during the reactions.
This indicates that the catalyst is reusable without loss of
activity and selectivity.
As shown in runs 2–5 (Table 2), photoreactions with Pd–TiO2
catalysts produced triethylamine (6) with the amount similar to
Fig. 3 Diffuse-reflectance UV-vis spectra of Pdx–TiO2, Pt2–TiO2, Au2–TiO2, and
Ag2–TiO2 catalysts.
that of toluene formed. Fig. 4 shows the time-dependent
c
1720 Catal. Sci. Technol., 2013, 3, 1718--1724
This journal is The Royal Society of Chemistry 2013

View Article Online
Paper
Catalysis Science & Technology
Table 2 Photocatalytic hydrodenitrogenation of benzonitrile (1) in ethanol with various catalystsa
Yield/%
Conversion
Acetaldehyde
H2 formed/
Run
Catalyst
d/nmb
of 1/%
Toluene (2)
N,N-Diethylbenzylamine (5)
Triethylamine (6)
formed/mmol
mmol
1
2
3
4
TiO2
0
42
58
>99
>99
>99
>99
86
0
40
57
98
99
97
98
85
0
0
0
38
54
96
96
95
97
83
0
6
293
256
206
o0.1
289
242
Pd0.5–TiO2
Pd1–TiO2
Pd2–TiO2
1st Reusec
2nd Reusec
3rd Reusec
Pd4–TiO2
Au2–TiO2
Ag2–TiO2
Pt2–TiO2
2.6
4.1
5.0
Trace
Trace
Trace
Trace
Trace
Trace
Trace
0
176
5.1
6.5
3.7
4.6
3.1
5.0
5
6
7
8
9
143
207
55
153
263
119
202
51
144
248
0
0
45
0
8
0
36
0
5
d
Pd2–TiO2
a
Reaction conditions: catalyst (10 mg), benzonitrile (10 mmol), ethanol (5 mL), Xe lamp (l > 300 nm), Ar (1 atm), temperature (298 K), time (6 h).
Average diameter for metal nanoparticles determined by TEM observations. Catalysts were reused after simple washing with ethanol.
Photoreaction was performed without benzonitrile.
b
d
c
H+ + eꢁ + Pd - H–Pd
(3)
The parts of the hydrogen atoms on the Pd particles are
removed by the coalescence as a H2 gas.
H–Pd + H–Pd - H2m + 2Pd
(4)
The stoichiometrical conversion of benzonitrile to toluene
and triethylamine (Fig. 4) indicates that three aldehyde
molecules formed by the photocatalytic oxidation of ethanol
(eqn (2)) are involved in the hydrodenitrogenation of one
benzonitrile molecule. The reaction mechanism can therefore
be summarized in Scheme 1, which involves twelve H–Pd species.
The substrate benzonitrile (1) undergoes hydrogenation by the
H–Pd species and produces benzylamine (3) intermediately. This
is transformed to an imine intermediate (7) by condensation
with acetaldehyde, and then converted to N-ethylbenzylamine (4)
via hydrogenation by the H–Pd species. Condensation of 4 with
acetaldehyde and subsequent hydrogenation by H–Pd species
afford N,N-diethylbenzylamine (5).21 Hydrogenolysis of 5 by
the H–Pd species produces toluene (2) and diethylamine (8).
Condensation of 8 with acetaldehyde and subsequent hydro-
genation produce triethylamine (6). During the photoreaction
(Fig. 4), trace amounts of 4 and 5 (o0.1 mmol) were detected by
GC analysis. In addition, as summarized in Table 3, photocatalytic
Fig. 4 Time-dependent change in the amounts of benzonitrile (1) and the
products obtained during photoreaction of benzonitrile with the Pd2–TiO2
catalyst. Reaction conditions are identical to those in Table 2.
change in the amounts of benzonitrile and products during
photoreaction with the Pd2–TiO2 catalyst. Photoirradiation
leads to a decrease in the amount of benzonitrile, along with
the formation of toluene and triethylamine. The profile for the
toluene formation is very similar to that for the triethylamine
formation. These data clearly suggest that the nitrogen atom of
benzonitrile is removed as triethylamine.
3.3 Mechanism for photocatalytic hydrodenitrogenation
Photocatalytic hydrodenitrogenation of benzonitrile is initiated
by photoexcitation of TiO2. This produces the positive hole (h+)
and electron (eꢁ) pairs, as follows.
TiO2 + hn - h+ + eꢁ
(1)
The h+ oxidizes ethanol and produces acetaldehyde and protons
(H+) on the TiO2 surface.18,20
CH3CH2OH + 2h+ - CH3CHO + 2H+
(2)
The photoformed conduction band eꢁ is transferred to the Pd
particles. This reduces H+ and produces a hydrogen atom on
the particles (H–Pd species).
Scheme 1 Proposed mechanism for photocatalytic hydrodenitrogenation of
benzonitrile with Pd–TiO2 catalysts in the presence of acetaldehyde.
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1718--1724 1721

View Article Online
Catalysis Science & Technology
Paper
Table
3
Results of photocatalytic reaction of various substrates with the
Table 4 Results of photocatalytic hydrodenitrogenation of various aromatic
cyanides with the Pd2–TiO2 catalysta
Pd2–TiO2 catalysta
Yield/%
Substrate
Time/h conversion/% Product
Run Substrate
Yield/%
98
Substrate
Conversion/%
Toluene (2)
Triethylamine (6)
3
4
5
>99
>99
>99
99
99
>99
97
98
97
1
2
12
48
>99
>99
a
98
Reaction conditions: Pd2–TiO2 (10 mg), substrate (10 mmol), ethanol
(5 mL), Xe lamp (l > 300 nm), Ar (1 atm), temperature (298 K),
photoirradiation time (6 h).
3
24
>99
93
reactions of compounds 3, 4, or 5 as the starting material with
the Pd2–TiO2 catalyst also produce toluene and triethylamine
with almost quantitative yields. These findings clearly support
the proposed denitrogenation mechanism involving the
condensation with aldehyde and the hydrogenation by the
H–Pd species (Scheme 1). As shown in Fig. 4, the mass balance
of benzonitrile (1) and toluene (2) is almost 100% during the
reaction. This suggests that the hydrogenation of 1 to 3 is
the rate-determining step for this reaction sequence. Table 2
(run 9) shows the result of photocatalytic reaction performed in
the absence of benzonitrile. The photoreaction produces
248 mmol H2, which is larger than that obtained with benzonitrile
(176 mmol, run 4). This indicates that hydrodenitrogenation and
H2 evolution (eqn (4)) take place competitively.
4
5
6
7
12
12
48
>99
>99
>99
93
>99
99
12
12
>99
>99
93
99
8
a
Reaction conditions: Pd2–TiO2 (10 mg), substrate (10 mmol), ethanol
(5 mL), Xe lamp (l > 300 nm), Ar (1 atm), temperature (298 K).
3.4 Hydrodenitrogenation of substituted benzonitrile
The Pd–TiO2 catalyst was employed for hydrodenitrogenation
of aromatic cyanides with several substituents. The results are
summarized in Table 4. As shown in run 2, terephthalonitrile is
successfully transformed to p-xylene with 98% yield via the
hydrodenitrogenation of two –CN groups. Reaction of
1-naphthonitrile (run 3) produces 1-methylnaphthalene with
93% yield. Reactions of benzonitrile with methyl, methoxy, or
hydroxyl substituents (runs 4–6) produce the corresponding
toluene derivatives while maintaining these substituents. In
contrast, halogens or carbonyl substituents also undergo reac-
tion. As shown in run 7, the reaction of p-chlorobenzonitrile
produces toluene because the halogen groups are removed via
the reaction with H–Pd species.15 In addition, the carbonyl
substituent is converted to the hydroxyl group (run 8) due to the
Fig. 5 Hammett plot for the photocatalytic reaction of p-substituted benzo-
nitriles with Pd2–TiO2. The k denotes the first-order rate constant for the decrease
in substrate concentration (mM hꢁ1) determined by 1 h photoreaction. The
reduction by the H–Pd species. These results suggest that the reaction conditions are identical to those in Table 4.
Pd–TiO2 catalyst promotes hydrodenitrogenation of the –CN
group even in the presence of substituents, although some
substituents are also transformed during the reaction.
in other words, the reaction proceeds via a nucleophilic attack
by a nucleophile.23–25 As proposed in Scheme 1, the reaction of
benzonitrile (1) would occur via the hydrogenation of the –CN
group (formation of benzylamine 3), via the nucleophilic attack
by the H–Pd species. The Hammett plot results clearly support
this mechanism.
The Hammett plot analysis was carried out to clarify the
effect of substituents on the reaction kinetics. Photocatalytic
reactions of the above p-substituted benzonitriles were carried
out with Pd2–TiO2 for 1 h, and the first-order rate constants for
the decrease in substrate concentration, k (mM hꢁ1), were
determined. Fig. 5 shows the relationship between log k and
the substituent constant, s.22 A linear correlation is observed,
3.5 Effect of Pd amount on the catalytic activity
and the slope of the plot (r) is determined to be +0.795. The As shown in Table 2, the denitrogenation activity depends on
positive r value indicates that stabilization of the negative the amount of Pd loaded, and the Pd2–TiO2 catalyst shows the
charge in the transition state efficiently promotes the reaction; highest activity. The H–Pd species are formed on the surface Pd
c
1722 Catal. Sci. Technol., 2013, 3, 1718--1724
This journal is The Royal Society of Chemistry 2013

View Article Online
Paper
Catalysis Science & Technology
atoms and, hence, the number of surface Pd atoms would
strongly affect the denitrogenation activity. As shown in
Fig. 2, the high-resolution TEM images of catalysts revealed
that the shape of Pd particles is a part of a cuboctahedron
surrounded by the (111) and (100) facets. The Pd particles on
TiO2 can therefore simply be modeled as an fcc cubocta-
hedron.26 This structure is generally used as a model for cubic
fcc metal nanoparticles. The numbers of surface metal atoms
calculated based on this model are often employed for the
determination of the active site in several catalytic systems such
as hydrogenation of allyl alcohols with H2 on Pd particles,27
steam reforming of methane on Pt particles,28 and oxidation of
cinnamyl alcohols on Au particles.29 The fcc cuboctahedron
model therefore allows rough determination of the number of
surface Pd atoms on the nanoparticles. Considering the full
shell close packing cuboctahedron for Pd particles where one
Pd atom is surrounded by twelve others, the number of total Pd
atoms per Pd particle (Ntꢄotal) is expressed by eqn (5) using the
number of shells (m). Ntꢄotal is rewritten with the average
diameter of Pd particles (dPd/nm) and the atomic diameter of
Pd (0.274 nm).30 The number of surface Pd atoms per Pd
particle (Nsꢄurface) is expressed by eqn (6).
Fig. 6 (Open symbol) The number of surface Pd atoms per gram of Pdx–TiO2
catalysts (Nsurface), and (closed symbol) the turnover number for toluene
formation per surface Pd atom on respective catalysts during photocatalytic
reaction of benzonitrile (TONsurface). Reaction conditions for benzonitrile were
identical to those in Table 2.
species with similar efficiency. The higher activity of Pd2–TiO2
than Pd0.5–TiO2 and Pd1–TiO2 (Table 2) is therefore because a
larger number of surface H–Pd species are produced due to the
larger number of surface Pd atoms (Nsurface).
ꢀ
ꢁ
3
10m3 ꢁ 15m2 þ 11m ꢁ 3
dPd
NtꢄotalðꢁÞ ¼
¼
(5)
In contrast, the TONsurface value of Pd4–TiO2 is much lower
than that of lower Pd loading catalysts (Fig. 6), suggesting that
the efficiency for H–Pd formation on the catalyst is lower. This
is because the photocatalytic oxidation and reduction efficiency
is decreased by the amount of Pd loaded. As shown in runs 2–5
(Table 2), the amounts of acetaldehyde and H2 produced on
Pd4–TiO2 are much lower than those on lower Pd loading
catalysts. It is well known that, as shown in Fig. 7, the metal–
semiconductor heterojunction creates a Schottky barrier (fB),
and the fB height increases with an increase in the amount of
metal loaded.31 The increased fB height by the increased Pd
loadings therefore suppresses smooth transfer of the conduc-
tion band eꢁ to the Pd particles. This may result in inefficient
charge separation between h+ and eꢁ and exhibit decreased
photocatalytic activity. As a consequence, the H–Pd formation
is suppressed, thus resulting in decreased denitrogenation
3
1:105 ꢂ 0:274
NsꢄurfaceðꢁÞ ¼ 10m2 ꢁ 20m þ 12
(6)
The number of Pd particles per gram of the catalyst (nparticle) is
expressed by eqn (7), using the percent amount of Pd loaded on
the catalyst [x (wt%) = Pd/(Pd + TiO2) ꢂ 100], molecular weight
of Pd [MW (= 106.42 g molꢁ1)], and Ntꢄotal. The number of surface
Pd atoms per gram of the catalyst (Nsurface) is therefore
expressed by eqn (8).
x
nparticleðmol gꢁ1Þ ¼
(7)
100 ꢂ MW ꢂ Ntꢄotal
Nsurfaceðmol gꢁ1Þ ¼ Nsꢄurface ꢂ nparticle
(8)
The Nsurface values for respective Pdx–TiO2 catalysts can there-
fore be calculated using their dPd values determined by the TEM
observations (Fig. 1). As shown by the open symbols in Fig. 6,
the Nsurface values increase with an increase in the amount
of Pd loaded; the values are 19.8 mmol gꢁ1 (Pd0.5–TiO2), 27.8
(Pd1–TiO2), 45.5 (Pd2–TiO2), and 73.1 (Pd4–TiO2), respectively.
To clarify the hydrodenitrogenation activity per surface Pd
atom, the turnover number for the reaction per surface Pd
atom on respective catalyst (TONsurface) was determined based
on eqn (9), using the amount of toluene formed (mol) during
the photocatalytic reaction of benzonitrile for 6 h in the
presence of 10 mg catalyst (Table 2).
½toluene formedꢅ
TONsurfaceðꢁÞ ¼
(9)
Nsurface ꢂ 10 ꢂ 10ꢁ3
As shown by the black symbols in Fig. 6, the TONsurface values
for Pd0.5–TiO2, Pd1–TiO2, and Pd2–TiO2 are almost the same.
This suggests that these catalysts produce the surface H–Pd
Fig. 7 Schematic representation of the Schottky barrier created at the Pd–TiO2
heterojunction.
c
This journal is The Royal Society of Chemistry 2013
Catal. Sci. Technol., 2013, 3, 1718--1724 1723

View Article Online
Catalysis Science & Technology
Paper
9 W. Ludwig, A. Savara, K.-H. Dostert and S. Schauermann,
J. Catal., 2011, 284, 148–156.
10 D. K. Yi, S. S. Lee and J. Y. Ying, Chem. Mater., 2006, 18,
2459–2461.
activity. These findings indicate that the Pd2–TiO2 catalyst, with
relatively low fB height and a large number of surface
Pd atoms, exhibits the highest activity for photocatalytic hydro-
denitrogenation of aromatic cyanides.
11 X. Huang, Y. Wang, X. Liao and B. Shi, Chem. Commun.,
2009, 4687–4689.
12 H. Wu, L. Zhuo, Q. He, X. Liao and B. Shi, Appl. Catal., A,
2009, 366, 44–56.
4. Conclusion
TiO2 loaded with Pd particles (Pd–TiO2) was used as a catalyst
for photocatalytic hydrodenitrogenation of aromatic cyanides
in ethanol as a hydrogen source. These catalysts, when irradiated
by UV light at room temperature, promote denitrogenation and
produce the corresponding toluene derivatives and triethylamine
with very high selectivity. Photoexcited Pd–TiO2 produces
acetaldehyde and the active H–Pd species. Consecutive reac-
tions involving hydrogenation by H–Pd species and condensa-
tion with aldehyde facilitate efficient hydrodenitrogenation.
The amount of Pd loaded and the size of Pd particles strongly
affect the denitrogenation activity. The Pd–TiO2 catalyst with a
relatively low Schottky barrier height at the Pd–TiO2 hetero-
junction and a large number of surface Pd atoms is necessary
for efficient denitrogenation.
13 J. J. W. Bakker, A. G. van der Neut, M. T. Kreutzer,
J. A. Moulijn and F. Kapteijn, J. Catal., 2010, 274, 176–191.
14 T. Sreethawong and S. Yoshikawa, Catal. Commun., 2005, 6,
661–668.
15 Y. Shiraishi, Y. Takeda, Y. Sugano, S. Ichikawa, S. Tanaka
and T. Hirai, Chem. Commun., 2011, 47, 7863–7865.
16 Y. Shiraishi, D. Tsukamoto, Y. Sugano, A. Shiro, S. Ichikawa,
S. Tanaka and T. Hirai, ACS Catal., 2012, 2, 1984–1992.
17 D. Tsukamoto, Y. Shiraishi, Y. Sugano, S. Ichikawa,
S. Tanaka and T. Hirai, J. Am. Chem. Soc., 2012, 134,
6309–6315.
18 Y. Shiraishi, Y. Sugano, S. Tanaka and T. Hirai, Angew.
Chem., Int. Ed., 2010, 49, 1656–1660.
19 D. Tsukamoto, A. Shiro, Y. Sugano, Y. Shiraishi, S. Ichikawa,
S. Tanaka and T. Hirai, ACS Catal., 2012, 2, 599–603.
20 Y. Shiraishi and T. Hirai, J. Photochem. Photobiol., C, 2008, 9,
157–170.
Acknowledgements
This work was supported by the Grant-in-Aid for Scientific
Research (no. 23360349) from the Ministry of Education,
Culture, Sports, Science and Technology, Japan (MEXT).
21 Y. Matsushita, N. Ohba, T. Suzuki and T. Ichimura, Catal.
Today, 2008, 132, 153–158.
22 K. A. Connors, Chemcal Kinetics, The Study of Reaction Rates
in Solution, VCH, New York, 1990.
23 X. Yang, L. Zhao, T. Fox, Z.-X. Wang and H. Berke, Angew.
Chem., Int. Ed., 2010, 49, 2058–2062.
24 R. L. Donkers, Y. Song and R. W. Murray, Langmuir, 2004,
20, 4703–4707.
Notes and references
1 T. Li, J. Liu, R. Bai, D.-G. Ohandja and F.-S. Wong, Water
Res., 2007, 41, 3465–3473.
`
´
´
2 G. Marcı, A. D. Paola, E. Garcıa-Lopez and L. Palmisano,
Catal. Today, 2007, 129, 16–21.
25 P. Lu and N. Toshima, Bull. Chem. Soc. Jpn., 2000, 73,
751–758.
´
´
´
´
´
ˇ
3 L. Martınkova, B. Uhnakova, M. Patek, J. Nesvera and
ˇ
V. Kren, Environ. Int., 2009, 35, 162–177.
26 R. E. J. Benfield, J. Chem. Soc., Faraday Trans., 1992, 88,
1107–1110.
27 O. M. Wilson, M. R. Knecht, J. C. Gareia-Martinez and
R. M. Crooks, J. Am. Chem. Soc., 2006, 128, 4510–4511.
28 Y. Qu, A. M. Sutherland, J. Lien, G. D. Suarez and T. Guo,
J. Phys. Chem. Lett., 2010, 1, 254–259.
4 P. Calza, E. Pelizzetti and C. Minero, J. Appl. Electrochem.,
2005, 35, 665–673.
5 K. Okamoto, Y. Yamamoto, H. Tanaka, M. Tanaka and
A. Itaya, Bull. Chem. Soc. Jpn., 1985, 58, 2015–2022.
6 H. Ohde, C. M. Wai, H. Kim, J. Kim and M. Ohde, J. Am.
Chem. Soc., 2002, 124, 4540–4541.
´
29 A. Abad, A. Corma and H. Garcıa, Chem.–Eur. J., 2008, 14,
7 J. Silvestre-Albero, G. Rupprechter and H.-J. Freund, Chem.
Commun., 2006, 80–82.
212–222.
30 L. Pauling, J. Am. Chem. Soc., 1947, 69, 542–553.
8 W. Ludwig, A. Savara, S. Schauermann and H.-J. Freund, 31 T. Uchihara, M. Matsumura, A. Yamamoto and H. Tsubomura,
ChemPhysChem, 2010, 11, 2319–2322.
J. Phys. Chem., 1989, 93, 5870–6874.
c
1724 Catal. Sci. Technol., 2013, 3, 1718--1724
This journal is The Royal Society of Chemistry 2013