Hydrogenation of pyrene using Pd catalysts supported on tungstated metal oxides

Article abstract of DOI:10.1016/j.apcata.2010.08.014

A series of Pd catalyst supported on tungstated metal oxides (Pd/W-MOx) were prepared and applied for pyrene hydrogenation, and the role of acid sites of supports was investigated. Among Pd/W-MOx catalysts, Pd/W-TiO₂ showed the highest activity which was comparable to those of Pd/BEA, Pd/Y and Pd/SiO₂-Al₂O₃. As for the Pd catalysts supported on metal oxides without tungstate (Pd/MOx), the hydrogenation activity became higher with the increase in the acid amount of supports measured by calorimetric measurement of ammonia adsorption. The important role of acid sites on hydrogenation activity was demonstrated. On the other hand, the hydrogenation activity of Pd/W-MOx catalysts was not correlated to the acid amount of supports measured by ammonia adsorption. On-site generation of protonic acid on tungstated metal oxide supports was estimated from kinetic analysis of reduced W species in the presence of hydrogen by in situ UV-visible measurement. From a good correlation between the kinetic parameters of on-site protonic acid formation and the hydrogenation activity, the important role of protonic acid formation on tungstated metal oxide supports was clarified.

Full text of DOI:10.1016/j.apcata.2010.08.014

Applied Catalysis A: General 387 (2010) 166–172
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Hydrogenation of pyrene using Pd catalysts supported on
tungstated metal oxides
Qiang Lin, Ken-ichi Shimizu, Atsushi Satsuma∗
Graduate School of Engineering, Nagoya University, Chikusa-ku, Nagoya 464-8603, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 May 2010
Received in revised form 6 August 2010
Accepted 11 August 2010
Available online 17 August 2010
A series of Pd catalyst supported on tungstated metal oxides (Pd/W–MOx) were prepared and applied
for pyrene hydrogenation, and the role of acid sites of supports was investigated. Among Pd/W–MOx
catalysts, Pd/W–TiO2 showed the highest activity which was comparable to those of Pd/BEA, Pd/Y and
Pd/SiO2–Al2O3. As for the Pd catalysts supported on metal oxides without tungstate (Pd/MOx), the
hydrogenation activity became higher with the increase in the acid amount of supports measured by
calorimetric measurement of ammonia adsorption. The important role of acid sites on hydrogenation
activity was demonstrated. On the other hand, the hydrogenation activity of Pd/W–MOx catalysts was
not correlated to the acid amount of supports measured by ammonia adsorption. On-site generation
of protonic acid on tungstated metal oxide supports was estimated from kinetic analysis of reduced W
species in the presence of hydrogen by in situ UV–visible measurement. From a good correlation between
the kinetic parameters of on-site protonic acid formation and the hydrogenation activity, the important
role of protonic acid formation on tungstated metal oxide supports was clarified.
Keywords:
Solid acid
Pd catalysts
Hydrogenation
Pyrene
Tungstated metal oxides
UV–visible spectroscopy
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
The main reasons for the use of silicates and alumina as sup-
ports for hydrogenation catalysts are their high surface area and
High aromatic content in diesel fuel lowers the fuel quality
and contributes significantly to the formation of undesired emis-
sions in exhaust gases [1–7]. As a simple solution to remove the
aromatics, deep hydrogenation of aromatic hydrocarbons over sup-
ported noble metal catalysts in mild conditions has drawn much
attention [1,2]. Metal oxides having strong acidity is known to be
effective as supports of noble metals [1,3,6]. Due to their strong
acidity, zeolites are the attractive materials for supports [3,6,8–12].
However, the pore sizes are usually too small to allow access of
polycyclic aromatic hydrocarbons. Actually, the minimum cross-
sectional diameters of the aromatic molecules in diesel fuels are
much larger than the pore-mouth opening of the supercage of Y
zeolite [13]. One of the solutions to overcome the large barrier
of aromatics is the introduction of meso- or macropores to zeo-
lite crystals [13–16]. For example, Tang et al. [15] showed that
the greater acidity. Actually, the important role of the acid sites
of supports has been well investigated [6,9,11,12,18,19]. Yasuda
et al. [11] investigated the effect of SiO2/Al2O3 ratio of USY zeo-
lites as supports for Pd–Pt bimetallic catalysts, and reported that
the activity for tetralin hydrogenation was strongly dependent on
the SiO2/Al2O3 ratio. Zheng et al. [12] also examined the effect
of SiO2/Al2O3 ratio of Pd/USY on the hydrogenation of naphtha-
lene. They found the decrease of the hydrogenation activity with
the increase of SiO2/Al2O3 ratio was in harmony with the acid
amount of USY determined by n-buthylamine-temperature pro-
grammed desorption. Pawelec et al. compared the hydrogenation
activity of Pd/BEA, Pd/SiO2–Al2O3, and Pd/Al2O3 using a model
feed containing toluene, naphthalene and dibenzothiophene [18],
respectively. They concluded that the highest hydrogenation activ-
ity of Pd/SiO2–Al2O3 can be attributed to the greater Brønsted
acidity of Pd/SiO2–Al2O3 and the degree of reduction of Pd species.
The positive role of acid sites of supports were attributed to both
the hydrogen spill-over mechanism and the electron-deficient state
of supported noble metals on acidic supports [6,12,18]. However,
such investigations on the role of acidity have been limited in the
silicates and alumina.
mesoporous-loaded Pd/BEA (mesoporous volume of 0.17 cm3 g−1
)
represented higher activity for pyrene hydrogenation than normal
Pd/BEA (mesoporous volume of 0.06 cm3 g−1), though the surface
area and the acid amount of the supports were comparable. Due to
the merit of diffusion of polycyclic aromatics, mesoporous silicas
or amorphous silica–alumina are also preferable as a support for
noble metals [16–19].
As described above, the promotional role of silicate supports on
the hydrogenation is essentially based on their greater amount of
strong acid sites. Other materials, such as transition metal oxides,
having strong acid sites would be other candidates for the supports
of precious metals for the hydrogenation of polycyclic aromatics.
∗
Corresponding author. Tel.: +81 52 789 4608; fax: +81 52 789 3193.
E-mail address: satsuma@apchem.nagoya-u.ac.jp (A. Satsuma).
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.08.014

Q. Lin et al. / Applied Catalysis A: General 387 (2010) 166–172
167
Table 1
List of tungstated metal oxides prepared in the present study.
Abbreviation
Tungstated metal oxides
W concentration (wt.% of WO3)
Surface area (m2 g−1
)
Surface W density (nm−2
)
10WZ
25WZ
45WZ
WMg
WTi
WO3/ZrO2
WO3/ZrO2
WO3/ZrO2
WO3/MgO
WO3/TiO2
WO3/Al2O3
WO3/CeO2
10
25
45
18
9
105
108
101
98
46
178
53
2.5
6.0
11.6
4.8
5.0
4.3
WAl
WCe
30
10
4.8
In the present study, we apply tungstated metal oxides as supports
of Pd catalysts for the hydrogenation of pyrene. Tungstated metal
oxides, such as WO3/ZrO2, are well known as solid acid materials
which can be applied for various applications such as hydrodesul-
furization, hydrocarbon cracking and light alkane isomerization
[20–23]. It has been reported that the types of oxide support and the
supported W density on the surface affect the amount and strength
of acid sites on tungstated metal oxides [20–26]. Due to their redox
property, supported tungstate is reduced in the presence of hydro-
gen to form partially reduced W species involving simultaneous
formation of protonic acids, i.e., Brønsted acids [21–29]. This pro-
cess can be significantly enhanced by an introduction of noble
metals [25,30]. Therefore, promotion effect of tungstated metal
oxides can be expected in the hydrogenation reactions.
The aims of this study are to clarify (1) the potential of
tungstated metal oxides as supports of Pd catalysts for the hydro-
genation of polycyclic aromatic hydrocarbons and (2) the role of
solid acid on tungstated metal oxides for the hydrogenation reac-
tion. Pyrene is used as a typical model compound of polycyclic
aromatic hydrocarbons. The following two techniques are applied
to estimate the acid sites of supports; (1) a microcalorimetric mea-
surement of NH3 adsorption and (2) a kinetic analysis of surface
tungstate using in situ UV–visible spectroscopy for estimation of
on-site formation of protonic acids in the presence of hydrogen
[30].
of Pd(NH3)4Cl2 with these supports, followed by drying at 373 K
overnight, and calcination in air at 723 K for 2 h.
2.3. Characterization
BET surface areas were measured using a flow apparatus for N2
adsorption at 77 K according to the one-point method. Microcalori-
metric measurement on NH3 adsorption was performed at 423 K
using a microcalorimeter (Tokyo Rikou Co., HAC-450G). Before the
adsorption experiments, the sample (0.1 g) was pretreated under
vacuum for 6 h at 423 K.
Diffuse reflectance UV–visible measurements were made using
a UV–visible spectrometer (JASCO V-550) equipped with an in situ
flow cell having a quartz window, a heating system inside, and con-
nected with a gas flow system as reported in our previous paper
[30]. A thermocouple located under the bed of the sample cell was
used to measure the local temperature. The light source is led to the
center of an integrating sphere by an optical fiber. Reflectance was
converted to pseudo-absorbance using the Kubelka-Munk func-
tion. BaSO4 was used to collect a background spectrum. Model gas
mixtures were fed at a flow rate of 100 cm3 min−1 to supported
tungsten oxide samples (50 mg) which were pretreated in a flow
of O2 (10%)/He at 823 K for 20 min. Then the time resolved in situ
UV–visible spectra were recorded at 473 K.
3. Hydrogenation of pyrene
2. Experimental
The hydrogenation of pyrene was carried out in a batch stainless
autoclave having a diameter of 2 cm and a volume of 40 mL. In a typ-
ical run of reaction, the reaction conditions were as follows: a 0.1 g
of catalyst, 1 mmol of pyrene in 4 mL of n-decane, reaction tempera-
ture of 523 K, initial total pressure of 6.2 MPa (hydrogen pressure of
about 6.0 MPa), reaction time of 80 min, and stirring rate of 800 rpm.
Liquid products were analyzed by GC-FID (Shimadzu GC-14A) and
GC–MS (Shimadzu GC-17A) with Rtx-65 or DB-1 capillary column
(Shimadzu). The yields of products were determined by quantita-
tive GC analysis using n-dodecane as an internal standard, and the
conversion was determined by the sum of the yields. Average mate-
rial balance, estimated from comparison of the total yield and the
amount of recovered pyrene, was 90.4%.
The reaction network of pyrene hydrogenation has been pro-
posed as Fig. 1 [15,31]. Pyrene hydrogenation is a consecutive
reaction that creates products in the order of dihydropyrene (B),
tetrahydropyrene (E), hexahydropyrene (C and D), decahydropy-
rene (G and F) and perhydropyrene (H). It was also reported that
the rate of the first ring hydrogenation is fast, the hydrogenation
rate gradually decreases with the hydrogenation getting deeper,
and the hydrogenation of the last ring is very difficult [1,15,16,31].
2.1. Preparation of tungstated metal oxides
ZrO2 was prepared by hydrolysis of zirconium oxynitrate 2-
hydrate in distilled water by gradually adding an aqueous NH4OH
solution (1.0 mol dm−3), filtration of precipitate, washing with
distilled water three times, and dryness at 373 K. ␥-Al2O3 was
prepared by calcination of ␥-AlOOH (Catapal B Alumina pur-
chased from Sasol) at 873 K for 3 h. Tungsten oxide (WO3, 99.9%,
16 m2 g−1) is supplied by Kishida Chemical Co, Ltd. Amorphous
silica (JRC-SIO-8, 303 m2 g−1), TiO2 (JRC-TiO-4, 50 m2 g−1), MgO
(JRC-MGO-1, 180 m2 g−1) and CeO2 (JRC-CeO-1, 53 m2 g−1) were
supplied from the Catalysis Society of Japan. Tungstated metal
oxides were prepared by impregnation of an aqueous solution of
ammonium paratungstate with above-mentioned supports. Sam-
ples were dried at 373 K overnight and then calcined in ambient
air at 873 K for 3 h. The properties and abbreviations of tungstated
metal oxides are listed in Table 1.
2.2. Preparation of Pd-supported catalysts
In addition to above-mentioned metal oxides and
tungstated metal oxides, proton-type Y zeolite (Y, JRC-Z-HY4.8,
SiO2/Al2O3 = 4.8), proton-type BEA zeolite (BEA, JRC-Z-HB25,
SiO2/Al2O3 = 25) and SiO2–Al2O3 (JRC-SAL-2, SiO2/Al2O3 = 5.6,
560 m2 g−1) supplied from the Catalysis Society of Japan were
used as supports. Supported Pd catalysts with Pd loading of
0.02 wt.% were prepared by impregnation of an aqueous solution
4. Result and discussion
4.1. Acidic properties of support
Fig. 2 shows the profiles of the differential heat of NH3 adsorp-
tion on various supports. BEA and Y zeolites show the initial

168
Q. Lin et al. / Applied Catalysis A: General 387 (2010) 166–172
Fig. 1. Proposed network of pyrene hydrogenation: (A) pyrene, (B) 4,5-dihydropyrene, (C) 1,2,3,6,7,8-hexahydropyrene, (D) 1,2,3,3a,4,5-hexahydropyrene, (E) 4,5,9,10-
tetrahydropyrene, (F) 1,2,3,3a,4,5,5a,6,7,8-decahydropyrene, (G) 1,2,3,3a,4,5,9,10,10a,10b-decahydropyrene and (H) perhydropyrene.
heat of adsorption in the range of 130–150 kJ mol−1, which are
4.2. Catalytic hydrogenation
in agreement with previous reports [32–34]. The differential heat
of adsorption gradually decreases with the loading amounts of
NH3, and finally reaches to around 50 kJ mol−1 above the surface
coverage of 1.0 mmol g−1. Since this minimum heat of adsorption
can be assigned to hydrogen-bonded NH3, we have defined the
total acid amount from adsorbed NH3 having the differential heat
above 50 kJ mol−1. Tungstated metal oxides show the initial heat
of adsorption in the range of 60–90 kJ mol−1. Other metal oxides
(not shown), i.e. Al2O3, TiO2 and ZrO2, showed the initial heat of
adsorption in the range of 60–90 kJ mol−1. On the other hand, the
initial heat of adsorption on SiO2 is only 37 kJ mol−1, which is due to
weak adsorption of NH3 via hydrogen-bonding [35]. The amounts
of acid site on all the supports thus determined are summarized in
Table 2.
Table 2 shows the results of catalytic tests for pyrene hydro-
genation over various supported Pd catalysts. When WZ samples
are used as supports, the catalytic activity is strongly affected by
WO3 content (Runs 1–3). Pd/25WZ shows the highest pyrene con-
version of 56.6% among these three catalysts. Taking the surface
area of ZrO2 into account, WO3 content of 25 wt.% is nearly the the-
oretical monolayer W density (7 W nm−2). As we have reported,
supported WO3 is present as surface polytungstates and tungsten
oxo-clusters up to the theoretical monolayer W density, and the
further increase in the WO3 content results in the formation of crys-
talline WO3 [30]. The decrease in the pyrene conversion at Pd/45WZ
suggests that the presence of WO3 crystalline is not favorable.
The effect of supports on tungstated metal oxides having the
theoretical monolayer W density is compared in the Runs 2 and
4–7. This series of catalysts is abbreviated as Pd/W–MOx, hereafter.
Among the series of catalysts, Pd/WTi shows the highest conversion
(76.9%). It should be noted that the conversion of pyrene hydro-
genation over Pd/WTi is comparable to those of the state-of-the-art
catalysts, i.e., Pd/SiO2–Al2O3, Pd/Y and Pd/BEA (Runs 8–10).
The pyrene conversion on Pd/25WZ, Pd/WAl and Pd/WTi are
higher than those on Pd/ZrO2, Pd/Al2O3 and Pd/TiO2, respectively.
The addition of WO3 generally enhances the hydrogenation activ-
ity. This is contrary to the results reported by Albertazzi et al.
[36–38]. They reported that addition of WO3 onto TiO2 decreased
the catalytic activity for tetralin hydrogenation on 1 wt.% Pd/Pt sup-
ported catalysts. In this study, Pd loading is 0.02 wt.% and far lower
than that of their study. The difference in the effect of WO3 addi-
tion may be due to the content of noble metals. The lower content
of Pd in this study is preferable for the positive effects of tungstated
metal oxides.
The amount of H2 consumption for various catalysts in Table 2
was calculated from the sum of produced hydropyrenes. Obviously,
higher pyrene conversion leads to broader distribution of prod-
ucts and higher selectivity of deeply hydrogenated products (C, D,
F and G). It should be noted that zeolite-supported Pd catalysts,
Fig. 2. Differential heat of adsorption of NH3 on (a) WMg, (b) SiO2, (c) WTi, (d)
25WZr, (e) WAl, (f) SiO2-Al2O3, (g) BEA zeolite and (h) Y zeolite.

Q. Lin et al. / Applied Catalysis A: General 387 (2010) 166–172
169
Table 2
Conversion and selectivity for hydrogenation of pyrene over various supported Pd catalystsa.
Run
Catalysts
Pyrene
Conv. (%)
Product selectivity (%)
H2 consumptionc
(mmol)
Acid amountd
(mmol g−1
)
B
C
D
E
F
G
H
Othersb
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Pd/10WZr
Pd/25WZr
Pd/45WZr
Pd/WMg
Pd/WCe
Pd/WAl
Pd/WTi
Pd/SiO2–Al2O3
Pd/Y
Pd/BEA
Pd/Al2O3
Pd/TiO2
Pd/WO3
Pd/SiO2
Pd/ZrO2
48.0
56.6
35.7
15.8
35.1
52.6
76.9
78.6
68.4
84.2
19.8
13.4
12.4
9.4
35.1
40.8
27.3
13.7
26.9
40.2
42.1
38.7
37.3
36.8
14.6
12.0
8.8
2.1
2.7
1.0
0.3
1.3
1.4
8.4
8.9
7.3
3.5
0.8
0.3
0.5
0.2
–
1.2
1.5
0.6
0.2
0.7
0.7
3.0
4.8
3.2
3.5
0.5
–
9.4
11.3
6.7
1.6
6.2
0.2
0.2
0.1
–
–
0.1
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
0.73
0.78
0.51
0.20
0.51
0.66
1.27
1.58
1.09
1.11
0.27
0.15
0.19
0.14
0.07
0.13
0.20
0.14
0.02
0.07
0.40
0.10
0.33
1.45
1.00
0.05
0.02
0.02
0.00
0.01
–
9.3
1.0
0.7
2.1
0.8
1.7
–
–
–
–
–
–
22.2
23.1
16.4
19.3
3.9
1.1
2.8
2.8
0.9
0.5
1.0
0.4
1.0
–
–
–
–
–
3.0
18.4
–
–
–
–
–
0.3
–
–
6.4
4.4
5.3
a
Reaction conditions: temperature, 523 K; initial total pressure, 6.2 MPa; reaction time, 80 min.
Major products are ring opening derivatives from pyrene.
Calculated from sum of H2 consumed on pyrene hydrogenation products from B to H in Fig. 1.
Acid amount of support determined from heat of NH3 adsorption.
b
c
d
Pd/Y and Pd/BEA, produced some other products of higher molec-
reaction could take place on metal sites as well as on acid sites
over noble metal/acid zeolite catalysts [27]. Furthermore, Simon
et al. proposed that benzene hydrogenation over 1.0 or 1.1 wt.%
Pt/mordenites includes two routes: (i) the monofunctional hydro-
genation of benzene on the metal itself and (ii) the hydrogenation
of Brønsted acid bound benzene using hydrogen dissociated on
the metal surface nearby [42]. They also showed a model that the
hydrogenation mainly proceeds via acidic path when the activity of
the metal sites decreased during surface Pt metal poisoned by thio-
phene. Accordingly, in the case of our catalysts with only small Pd
loading, the acidic pathway mainly contributes to hydrogenation of
aromatic molecules. As shown in the correlation of Pd/MOx in Fig. 3,
the promotion effect of solid acids on the hydrogenation activity
is confirmed in the case of Pd catalysts supported on the transi-
tion metal oxides. This result suggests that the possible reaction
pathway should be similar to the proposals of Vannice and co-
workers [39–41] and Simon et al. [42]. The hydrogenation mainly
proceeds as adsorption and then activation of pyrene on Brønsted
acid sites to form carbocation intermediate followed by hydrogena-
tion using dissociated hydrogen atoms migrated from metal surface
ular mass, which were assigned mainly to hydro-benzopyrenes or
alkyl-pyrenes by using GC–MS. Because they are produced through
alkylation of pyrene and the solvent of decane on acid sites of
zeolites, the influence of these by-products on hydrogenation on
Pd surface should be negligible. Therefore, the amount of H2 con-
sumption estimated from hydropyrene yield can be regarded as a
measure of both the hydrogenation activity and selectivity.
In the case of Pd catalysts supported on the simple metal
oxides (Runs 11–15), the H2 consumptions are correlated with the
amount of acid sites, i.e., the highest H2 consumption is observed
on Pd/Al2O3, while Pd/SiO2 and Pd/ZrO2 show very low activity.
The H2 consumption is more than 6 times higher on more acidic
Pd/SiO2–Al2O3 catalyst than that of Pd/Al2O3. However, this corre-
lation cannot be observed in other series of catalysts. For example,
Pd/WTi is the most active among Pd/W–MOx catalysts but its
acid amount estimated by calorimeter is very low (0.10 mmol g−1).
The zeolite-supported Pd catalysts also show lower hydrogenation
activities than that of Pd/SiO2–Al2O3, although they have larger
amount of acid sites. The dependence of pyrene hydrogenation on
the acidity of support will be discussed in the next section.
4.3. Effect of acid sites on hydrogenation
The H2 consumption over the supported Pd catalysts are plotted
in Fig. 3 as a function of the amount of acid sites on the supports
estimated by the differential heat of NH3 adsorption. In the cases
of Pd-supported on SiO2–Al2O3, Al2O3, TiO2, WO3, SiO2 and ZrO2
(abbreviated as Pd/MOx, hereafter), the H2 consumption is fairly
correlated with the amount of acid sites as shown in the straight
line in the figure. This result suggests the hydrogenation activity is
strongly influenced by the acidity of supports for metal-supported
catalysts with a low content of Pd (0.02 wt.%). Although it is well
known that benzene hydrogenation proceeds via a purely metal
catalyzed reaction [1], the important role of acid sites on supports
also been pointed out for hydrogenation of aromatics over silicates
and alumina [6,11,16,24,27]. Quantitative correlation between the
acid amount and the hydrogenation activity was reported by Paw-
elec et al. [6,18]. As proposed by Vannice and co-workers [39–41]
and other researchers [27,42,43], acid sites can produce protons
from spill-over hydrogen migrated from metal surface, which can
activate aromatic molecules adsorbed on acidic sites by the forma-
tion of carbocation. Wang et al. suggested that the hydrogenation
Fig. 3. H2 consumptions for pyrene hydrogenation over Pd catalysts supported on
(ꢀ) zeolites, (ꢀ) tungstated metal oxides and (ꢁ) other metal oxides.

170
Q. Lin et al. / Applied Catalysis A: General 387 (2010) 166–172
or hydride. The formation of hydride on acidic supports is proposed
by Hattori and Shishido [44] as follows. The hydrogen molecule dis-
sociates on metal surface and hydrogen atoms spill-over onto the
acidic support and migrate to Lewis acid sites, where the hydrogen
atom releases an electron to become a proton. The electron trapped
on the Lewis acid site may react with a second hydrogen atom to
form a hydride which is stabilized on the Lewis acid site.
Although the amounts of acid site are very high on Pd/Y and
Pd/BEA, these catalysts show lower H2 consumptions than that of
Pd/WTi and Pd/SiO2–Al2O3. This result suggests a negative effect
of diffusion limitation of pyrene and hydropyrenes in the zeo-
lite channel. This is because the molecular diameter of pyrene
(0.72–0.92 nm in its planer directions) is larger than the pore diam-
eter of Y (0.74 nm) and BEA (0.66–0.77 nm). Actually, Meng et al.
[13] and Zhang et al. [16] reported that the introduction of meso-
pores into zeolite Y results in the higher hydrogenation activity for
naphthalene and pyrene over Pd–Pt/Y, respectively. Tang et al. [15]
also demonstrated that the pyrene hydrogenation activity was pro-
moted by the increase in the mesoporous volume of Pd/BEA. These
results indicate the diffusion limitation of aromatic compounds to
access the acid sites or metal clusters in zeolite micropores. In the
present study, the lower hydrogenation activity of Pd/Y and Pd/BEA
than that expected from the acid amount of supports can be ratio-
nalized by the diffusion limitation of pyrene and hydropyrenes in
the zeolite channels.
As shown in open circles in Fig. 3, the correlation between
the H2 consumption and the amount of acid site is not clear on
Pd/W–MOx, indicating that other controlling factors work for the
pyrene hydrogenation activity. The influence of diffusion limita-
tion can be excluded because the surface areas of Pd/W–MOx are
comparable or lower than those of Pd/MOx. It can be expected that
some chemical properties of surface W species affected the activity
for pyrene hydrogenation. As a characteristic function of WO3/ZrO2,
it is well known that protonic acid sites, i.e., Brønsted acid sites, are
formed from gaseous hydrogen molecules involving reduction of
surface W6+ to W(6-␦)+ [21–29]. This reduction–oxidation process
including the formation of protonic acids can be simply expressed
as the following Eq. (1) [30]:
Fig. 4. UV–vis spectra of Pd/25WZr in a flow of pure He and 0.5%H2/He at 473 K.
troscopy for the estimation of the on-site formation of protonic
acids on Pd/W–MOx catalysts.
Fig. 4 shows UV–visible spectra of Pd/25WZr in a flow of
0.5%H2/He and pure He at 473 K after pretreatment in a flow of
O2/He at 823 K and cooling to 473 K in flowing He. In pure He,
Pd/25WZ shows only a major broad band below 410 nm assigned
to the LMCT band of tungstate [23,30]. When the catalyst reduced
in a flow of 0.5% H2/He, another broad band appears in the range
of 350–800 nm together with the LMCT band below 410 nm. After
the switching the flow gas from H2/He to pure He, the broad band
gradually decreases and finally returns to the original level (fig-
ure not shown). Since the UV–visible spectra of pure ZrO2 gave no
absorption band in the range of 350–800 nm under both oxidative
and reductive atmospheres, oxidation state of ZrO2 do not affect
the spectra of Pd/25WZ. For pure CeO2, the change in the band
intensity in the same range was less than 0.001 KM unit between
the oxidative and reductive atmospheres. The negligible contribu-
tion of supported 0.02 wt.% of Pd to the spectra of Pd/W–MOx was
also confirmed by a set of separate experiments. For example, the
sample of 0.02 wt.%Pd/ZrO2 without WO3 only showed very small
absorption band below 0.02 KM unit in the range of 300–800 nm
and below 0.0035 KM unit at 600 nm, and the change in the band
intensity was only below 0.004 KM unit with the reduction in 0.5%
H2/He. Therefore, the change in the band intensity in the range of
350–800 nm represents the reduction–oxidation process between
surface W6+ and W(6-␦)+ species. Although the pyrene hydrogena-
tion was carried out at 523 K, the in situ UV–visible measurements
were carried out at 473 K because a part of surface tungstate was
irreversibly reduced at 523 K but all the surface tungstate reversibly
reduced and oxidized at 473 K.
K
R
W
6+ + 0.5H2ꢂW(6-␦)+ + H␦+
(1)
K
O
These temporary protonic acids act as active Brønsted acids to
catalyze o-xylene or n-alkane isomerization in the presence of H2
[24]. Furthermore, the addition of precious metals on solid acids
generally promotes the formation of protonic acid sites [20,25,44].
Actually, a loading of Pt onto WO3/ZrO2 significantly improves iso-
merization of pentane at lower temperatures [24]. Therefore, the
on-site formation of protonic acid sites by H2 dissociation should
affect the catalytic activity of pyrene hydrogenation. In the next
section, we will examine the contribution of the on-site formation
of protonic acid sites in the pyrene hydrogenation.
Fig. 5 shows the transient response of the band at 600 nm of
Pd/W–MOx catalysts in a flow of pure He (0–50 s) and 0.5% H2/He
(after 50 s) at 473 K. As shown in the figure, the band intensity at
600 nm initially increases after the feed of 0.5% H2/He due to the
reduction of surface W6+ to W(6-␦)+ species, and then reaches to
the equilibrium [30]. It is clear that the response magnitude of
the band is the greatest on Pd/WTi, and Pd/25WZr also show a
large change in the band intensity. On the other hand, the reduc-
tion of surface W species is negligible on Pd/WMg. The maximum
response magnitude between the oxidative state and the equilib-
rium in 0.5% H2/He, indicated as ꢀKM hereafter, is in the order of
Pd/WTi > Pd/WZr > Pd/WAl > Pd/WCe > Pd/WMg. When the flowing
gas was switched to pure He, the intensity of the band decreased by
re-oxidation of W(6-␦)+ to W6+ species, as reported in the previous
paper (figure not shown) [30].
4.4. Contribution of on-site formation of protonic acid on pyrene
hydrogenation
According to the Eq. (1), the on-site formation of protonic acids
over W–MOx can be estimated from the formation of reduced
W(6-␦)+ species. Diffuse reflectance UV–visible spectroscopy is a
powerful tool for characterizing the electronic states of surface
tungstate species [21–24]. Iglesia and co-workers examined the
color centers in WOx domains during reduction by hydrogen
[22,24]. In the previous paper, we reported an kinetic analysis of
the reduction–oxidation process between surface W6+ and W(6-␦)+
species using time resolved in situ UV–visible spectroscopy [30].
In this study, we apply the time resolved in situ UV–visible spec-

Q. Lin et al. / Applied Catalysis A: General 387 (2010) 166–172
171
Fig. 5. Transient response of the band at 600 nm of Pd/W–MOx catalysts at 473 K.
Fig. 6. H2 consumption for pyrene hydrogenation as
a
function of kR/kO on
The flowing gas was switched from pure He to 0.5%H2/He at 50 s.
Pd/W–MOx.
As reported in the previous paper [30], the rate constants of
reduction (kR) and that of re-oxidation (kO) in Eq. (1) were esti-
mated from the responses of the band in the initial stage of the
reduction or the re-oxidation using the following the first-order
reaction equation:
kt = ln [W(6-␦)+] − ln [W(6-␦)+]0
(2)
Table 3 summaries the rate constants of reduction (kR) and
re-oxidation (kO), their ratio (kR/kO), and ꢀKM. These param-
eters are strongly affected by the type of oxide supports.
The rate constants of kR and kO increased in the order of
Pd/WMg < Pd/WCe < Pd/WAl < Pd/WZr < Pd/WTi. Since the varia-
tion of the rate constant of kR is more significant, the ratio of kR/kO,
which is the equilibrium constant of the Eq. (1), also increases in
the same order. The order of the maximum response magnitude
(ꢀKM) is also in the same order.
In Fig. 6, the H2 consumptions for pyrene hydrogenation of
Pd/W–MOx catalysts are plotted as a function of kR/kO. As the
increase in the equilibrium constant of kR/kO, the H2 consump-
tion of Pd/W–MOx catalysts increases. Such a good correlation is
also observed between the H2 consumptions and ꢀKM as shown
in Fig. 7, indicating ꢀKM can be used as more convenient param-
eter of the equilibrium, alternatively. These correlations clearly
demonstrate that the on-site formation of Brønsted acid sites on
Pd/W–MOx is the controlling factor for the catalytic activity in the
pyrene hydrogenation. This result is in agreement with the dis-
cussion for Pd/MOx catalysts in Section 4.3 and also supports the
proposed possible mechanism. The on-site formation of Brønsted
acid should proceed via dissociative adsorption of a hydrogen
molecule on noble metal followed by spill-over of a hydrogen atom
and migration on the support surface to form a proton, as proposed
by Hattori and co-workers [25,44]. In their proposal, a hydride
Fig. 7. H2 consumption for pyrene hydrogenation as a function of degree of W
reduction on Pd/W–MOx.
should produce via reaction of a hydrogen atom with a trapped
electron which was generated with the formation of Brønsted acid
on solid supports. In the case of Pd/W–MOx, generation of a hydride
could be available via reaction of a spilt hydrogen atom with W(6-␦)+
species because W(6-␦)+ species on metal-supported catalysts is
easily re-oxidized to W6+ in the presence of noble metal [30]. Con-
sequently, as a characteristic feature of tungstated metal oxide
supports, the promoted hydrogenation activities of Pd/W–MOx can
be attributed to the on-site formation of protonic acids involving
the reduction of W(6-␦)+ species.
5. Conclusion
Table 3
Rate constant of reduction under 0.2%H2/He (kR) and re-oxidation under pure He
(kO), their ratio (kR/kO), and change in the band intensity at 600 nm under 0.2%H2/He
(ꢀKM).
The effect of acid sites of supports on the pyrene hydro-
genation over Pd catalysts was investigated using BEA and Y
zeolites, SiO2–Al2O3, Al2O3 and other transition metal oxides,
and tungstated metal oxides as supports. Pd/WO3–TiO2 catalyst
showed the comparable hydrogenation activity to Pd/BEA, Pd/Y and
Pd/SiO2–Al2O3. The catalytic activities of Pd catalysts supported
on non-tungstated metal oxides were fairly dependent on the acid
amount of supports measured by the heat of NH3 adsorption. On
Catalyst
kR (s−1
)
kO (s−1
)
kR/kO
ꢀKM
Pd/WMg
Pd/WCe
Pd/WAl
Pd/25WZr
Pd/WTi
1.1 × 10−6
4.6 × 10−5
2.9 × 10−4
1.5 × 10−3
2.7 × 10−3
1.9 × 10−5
2.8 × 10−4
5.8 × 10−4
1.2 × 10−3
1.7 × 10−3
0.06
0.16
0.50
1.2
0.0004
0.02
0.035
0.09
1.6
0.15

172
Q. Lin et al. / Applied Catalysis A: General 387 (2010) 166–172
the other hand, this correlation was not observed when tungstated
metal oxides were used for supports. The hydrogenation activity
of Pd/W–MOx was well correlated to the equilibrium constant of
the reversible reduction–oxidation of the surface W species, which
represents the on-site formation of protonic acid from gaseous H2.
It was concluded that the promoted hydrogenation activities of
Pd/W–MOx can be attributed to the on-site formation of protonic
acids involving the reduction of surface tungstate.
[19] A. Corma, A. Martínez, V. Martínez-Soriay, J. Catal. 169 (1997) 480–489.
[20] I.E. Wachs, T. Kim, E.I. Ross, Catal. Today 116 (2006) 162–168.
[21] E. Iglesia, D.G. Barton, S.L. Soled, S. Miseo, J.E. Baumgartner, W.E. Gates, G.A.
Fuentes, G.D. Meitzner, Stud. Surf. Sci. Catal. 101 (1996) 533–542.
[22] D.G. Barton, S.L. Soled, E. Iglesia, Top. Catal. 6 (1998) 87–99.
[23] D.G. Barton, M. Shtein, R.D. Wilson, S.L. Soled, E. Iglesia, J. Phys. Chem. B 103
(1999) 630–640.
[24] J. Macht, E. Iglesia, Phys. Chem. Chem. Phys. 10 (2008) 5331–5343.
[25] S. Triwahyono, T. Yamada, H. Hattori, Appl. Catal. A 242 (2003) 101–109.
[26] S. Triwahyono, T. Yamada, H. Hattori, Appl. Catal. A: Gen. 250 (2003) 65–73.
[27] J. Wang, Q. Li, J. Yao, Appl. Catal. A 184 (1999) 181–188.
[28] S. Kuba, P. Concepción Heydorn, R.K. Grasselli, B.C. Gates, M. Che, H. Knözinger,
Phys. Chem. Chem. Phys. 3 (2001) 146–154.
[29] S. Kuba, P. Lukinskas, R.K. Grasselli, B.C. Gates, H. Knözinger, J. Catal. 216 (2003)
353–361.
[30] Q. Lin, K. Shimizu, A. Satsuma, Appl. Catal. A 365 (2009) 55–61.
[31] S.C. Korre, M.T. Klein, Ind. Eng. Chem. Res. 34 (1995) 101–117.
[32] N. Cardona-Martinez, J.A. Dumesic, Adv. Catal. 38 (1992) 149–224.
[33] A. Auroux, Top. Catal. 4 (1997) 71–89.
[34] V. Solinas, I. Ferino, Catal. Today 41 (1998) 179–189.
[35] K. Shimizu, E. Hayashi, T. Hatamachi, T. Kodama, T. Higuchi, A. Satsuma, Y.
Kitayama, J. Catal. 231 (2005) 131–138.
[36] S. Albertazzi, E. Rodrí guez-Castellón, M. Livi, A. Jimépez-López, A. Vaccari, J.
Catal. 228 (2004) 218–224.
[37] S. Albertazzi, C. Barbieri, A. Infantes-Molina, A. Jimépez-López, R. Moreno-Tost,
E. Rodrí guez-Castellón, A. Vaccari, Catal. Lett. 104 (2005) 29–36.
[38] S. Albonetti, G. Baldi, A. Barzanti, E.R. Castellon, A.J. Lopez, D.E. Quesada, A.
Vaccari, Catal. Lett. 108 (2006) 197–207.
References
[1] A. Stanislaus, B.H. Cooper, Catal. Rev. Sci. Eng. 36 (1994) 75–123.
[2] B.H. Cooper, B.B.L. Donnis, Appl. Catal. A 137 (1996) 203–223.
[3] C. Song, X.L. Ma, Appl. Catal. B 41 (2003) 207–238.
[4] D. Minicucci, X.Y. Zou, J.M. Shaw, Fluid Phase Equilib. 353 (2002) 194–197.
[5] T. Yuan, W.D. Marshall, J. Hazard. Mater. B 126 (2005) 149–157.
[6] B. Pawelec, R. Mariscal, R.M. Navarro, S. van Bokhorst, S. Rojas, J.L.G. Fierro,
Appl. Catal. A 225 (2002) 223–237.
[7] K.M. Lin, C. Song, Catal. Today 31 (1996) 93–104.
[8] A.D. Schmitz, G. Bowers, C. Song, Catal. Today 31 (1996) 45–56.
[9] C. Song, A.D. Schmitz, Energy Fuels 11 (1997) 656–661.
[10] H. Yasuda, Y. Yoshimura, Catal. Lett. 46 (1997) 43–48.
[11] H. Yasuda, T. Sato, Y. Yoshimura, Catal. Today 50 (1999) 63–71.
[12] J. Zheng, M. Guo, C. Song, Fuel Process. Technol. 89 (2008) 467–474.
[13] X.C. Meng, Y.X. Wu, Y.D. Li, J. Porous Mater. 13 (2006) 365–371.
[14] K.M. Reddy, C. Song, Catal. Today 31 (1996) 137–144.
[15] T. Tang, C. Yin, L. Wang, Y. Ji, F.S. Xiao, J. Catal. 249 (2007) 111–115.
[16] H. Zhang, X. Meng, Y. Li, Y.S. Lin, Ind. Eng. Chem. Res. 46 (2007) 4186–4192.
[17] L.Y. Chen, Z. Ping, G.K. Chuah, S. Jaenicke, G. Simon, Micropor. Mesopor. Mater.
27 (1999) 231–242.
[39] P. Chou, A.A. Vannice, J. Catal. 107 (1987) 129–139.
[40] S.D. Lin, M.A. Vannice, J. Catal. 143 (1993), pp. 539–543, 554–562, 563–572.
[41] D. Poondi, A.A. Vannice, J. Catal. 161 (1996) 742–751.
[42] L.J. Simon, J.G. van Ommen, A. Jentys, J.A. Lercher, Catal. Today 73 (2002)
105–112.
[43] J. Wang, L. Huang, Q. Li, Appl. Catal. A. Gen. 175 (1998) 191–199.
[44] H. Hattori, T. Shishido, Catal. Surv. Jpn. 1 (1997) 205–213.
[18] B. Pawelec, J.M. Campos-Martin, E. Cano-Serrano, R.M. Navarro, S. Thomas, J.L.G.
Fierro, Environ. Sci. Technol. 39 (2005) 3374–3381.