Journal of
Catalysis 225 (2004) 128–137
Catalytic oxidation of pyridine on the
supported copper catalysts in the
presence of excess oxygen
J. Zhou, Q.-H.
Xia,∗ S.-C.
Shen, S. Kawi, and K. Hidajat
Department of Chemical and
Environmental Engineering, National University of Singapore, Singapore,
Republic of Singapore
Received
19 September 2003; revised 24 March 2004; accepted 24 March 2004
Available
online 12 May 2004
Abstract
The catalytic oxidation
of pyridine pollutant
on a series of supported
metals/zeolites catalysts in the presence
of excess oxygen was studied. All the catalysts were further characterized by means of BET, XRD, H2-TPR, XPS, and FTIR techniques. The catalyst that could be reduced
at
lower temperatures had a better oxidation activity. The better NOx control
ability could be attributed to the importance of the Lewis acidity of the
samples, rather than to that of Brønsted acidity. On supported copper
catalysts, Cu(H2O)62+ ions had higher activity for
the
NOx
control but poorer activity for the pyridine
oxidation. The cointeraction of some factors, such as the surface acidity of
catalysts, the structure of supports, and the amount of Cu(H2O)62+ ions on the
supports, played a deciding role in the NOx
control ability of the catalysts.
Comparatively,
Cu/beta was the most active for the pyridine oxidation and NOx control,
possibly being a potential catalyst for the catalytic removal of pyridine
pollutant.
Ó 2004
Elsevier Inc. All rights reserved.
Keywords: Supported copper catalysts; Cu/β; Catalytic oxidation of pyridine; NOx yield
1. Introduction
Nowadays vehicle exhaust as a major source of air pollu- tion is attracting increasing
attention from governments and society. Nitrated polycyclic aromatic
hydrocarbons (NPAC) have been
identified as potent mutagens and possible car- cinogens. Since 1977, when the EPA (Environment Protec- tion Association
of America) first identified diesel exhaust particulates as mutagens, NPAC has been thought to be re- sponsible
for up to 90% of the total mutagenicity of diesel particulates. Although
much research has been performed
to determine the concentration of NPAC
in the diesel exhaust, the reports on the removal of NPAC from the diesel exhaust are quite
limited. Since the automobile emission
legisla- tion is becoming stringent, the
automotive industry is inter- ested in the development of a catalyst applied
for removing NPAC from the exhaust
of diesel engines.
For understanding the catalytic decomposition of nitro-PACs, nitro-compounds may be more suitable than
pyridine as the model com- pounds. Indeed, Ismagilov et al. [1] reported higher NOx
|
|
|
* Corresponding
author.
yield from nitrobenzene than from pyridine
when both were catalytically oxidized. However, as one important pollutant, the
catalytic oxidation of pyridine has been extensively in- vestigated, which thus
led to our studies on the oxidation decomposition of pyridine over a series
of zeolite-supported
catalysts.
To date, there have been two main routes dealing
with the catalytic removal
of pyridine. One is the catalytic oxida- tion (or the catalytic combustion),
the other is the catalytic supercritical water oxidation (CSCWO). The catalytic
ox- idation of pyridine was reported in
1983 by Ismagilov et al. [2,3], who
observed the formation of nitrogen oxides in the catalytic oxidation of
pyridine over the catalysts, such as 0.64% Pt/γ -Al2O3, 26% CuO/γ
-Al2O3, 5% CuCr2O4/ γ -Al2O3, 11% CuO/γ -Al2O3, and 30% CuCr2O4/γ -Al2O3. All the catalysts showed similar specific activity
toward the oxidation of pyridine, and
the oxides catalyst exhib- ited much better NOx control ability (the ability to con-
trol the yield of NOx ) than the platinum one. It was found that the
NOx control ability tended to decrease in the or- der of 30% CuCr2O4/γ -Al2O3 > 11% CuO/γ -Al2O3 > 5%
CuCr2O4/γ -Al2O3. They also reported that at temperatures between 240 and 550 ◦C, the main products
were CO2, H2O,
0021-9517/$ – see front matter Ó 2004 Elsevier Inc.
All rights reserved. doi:10.1016/j.jcat.2004.03.042
N2,
and NOx , in which the yield of NOx increased with the increase of temperature.
Luo et al. [4] studied the pyridine oxidation over γ - Al2O3-supported
metal oxide catalysts,
such as Ag/ γ -Al2O3, Cu/γ -Al2O3, Mn/γ -Al2O3, Cr/γ -Al2O3, Fe/
γ -Al2O3, Co/γ -Al2O3, Ni/γ -Al2O3, V/γ -Al2O3, and Ce/ γ -Al2O3. They found that the content of NOx generating from the pyridine combustion went through a
maximum value as the temperature increased, and
that the oxida- tion
activity of individual catalyst was proportional to its NOx control ability. The oxidation activity and NOx control ability of the catalysts decreased in the
sequence Ag/γ - Al2O3 > Cu/γ -Al2O3 > Mn/γ -Al2O3 > Cr/γ -Al2O3 >
Fe/γ -Al2O3 ≈ Co/γ -Al2O3 > Ni/γ -Al2O3, V/γ -Al2O3 >
Ce/γ -Al2O3.
The results further showed that the oxidation activity and NOx control ability of individual Ag/γ -Al2O3 catalyst could be promoted by certain amount
of Mn or Co.
Their other researches showed that increasing the metal
loading remarkably improved the oxidation activity and NOx control ability of
the catalyst until reaching an equilib- rium
value [5]. They further suggested that the ability
of the catalyst to control
the formation of NOx species be closely relevant to the acidity
on the catalyst surface.
Crain
et al. [6] reported the catalytic
supercritical wa- ter oxidation of
pyridine in the temperature range of 425– 527 ◦C, in which the conversion of pyridine increased from
0.03 (at 426 ◦C) to 0.68 (a.u.)
(at 527 ◦C) at a residence time of 10 s. During the CSCWO of
pyridine, ammonia, dimeth- ylamine, and several carboxylic acids such as formic
acid, acetic acid, glutaconic acid, and oxalic acid could be de- tected. Wightman
et al. [7] reported the CSCWO of pyridine
in a flowing system with a
conversion of around 16% (at 400 ◦C and a pressure
of 40.82 MPa). Katritzky and Barcock
[8] observed that pyridine was nonreactive in the absence
of oxygen, even in supercritical water. Aki and Abraham [9]
reported the absence of external mass-transfer
limit and the presence of internal mass-transfer limit during the CSCWO of
pyridine over Pt/γ -Al2O3.
As noted above, research
on the catalytic removal of pyri-
dine has been carried out over metal
oxide-supported cat- alysts such as Pt/γ -Al2O3 or
CuO/γ -Al2O3. However, no study has been, to date, focused on the
zeolite-supported catalysts, which are
being widely used in the SCR-deNOx reaction and petrochemical industry. The aim
of this work is to test the catalytic
oxidation of pyridine over zeolite- supported catalysts and to compare the
difference in their activities for the oxidation of pyridine.
2.
Experimental
2.1.
Preparation
of catalysts
Supported-copper catalysts were prepared by wet impreg- nation of powder support solids
with an aqueous copper(II) nitrate solution. Excess water was evaporated off at
90 ◦C under stirring. The samples were, then, dried
at 100 ◦C for
12 h, followed by fine grinding,
and calcination at 700 ◦C for 10 h in the dry air. Metal/beta catalysts were also pre- pared by wet impregnation of zeolite
β with the aqueous solutions containing different metal salts,
such as cobalt
ac- etate, nickel nitrate,
and iron sulfate. The powdered support materials used were beta, ZSM-5, MCM-41,
and γ -Al2O3 (Strem
Chemicals, USA). Zeolite beta with a Si/Al
ratio of 20 was synthesized according to the procedure
reported elsewhere [10].
Zeolite ZSM-5 with
a Si/Al ratio of 15 was
synthesized hydrothermally at
170 ◦C using tetrapropylam-
monium hydroxide (TPAOH) as the structure-directing agent
[11,12]. The
mesoporous MCM-41 with a Si/Al ratio
of 20 was hydrothermally synthesized at 100 ◦C using C16TMABr as a template
[13,14]. The removal
of organic template mole- cules was carried out by calcinations at 550 ◦C for 10 h in a
flowing air. Note that the unit of metal
loadings on the sup- port was wt%.
2.2. Characterization of catalysts
Autosorb-1 was used to determine
the N2
adsorption– desorption isotherms
of the samples. The BET specific sur- face area was calculated in terms of the
Brunauer–Emmett– Teller (BET) equation. The BJH method was used to calcu- late
the pore volume and the pore-size distributions of the
samples. Prior to the
measurements, the samples were de- gassed at 300 ◦C
under a vacuum of 10−2 mbar for 10 h.
X-ray diffraction (XRD) analysis
was carried out using a Shimadzu
XRD-6000 diffractometer with Ni-filtered Cu- Kα
radiation operating at 40 kV
and 30 mA. The scanning conditions for ZSM-5, beta, and Al2O3 were as follows: di-
vergence slit, 1.0 (◦);
scattering slit, 1.0 (◦);
and receiving
slit, 0.3 (mm). The scanning
range was from 5 to 80◦ 2θ with a scanning speed of 2◦ /min. The measuring conditions for MCM-41 were
changed to divergence slit, 0.05 (◦); scatter- ing slit, 1.0 (◦); and receiving slit, 0.15 (mm), with a scan- ning range
from 1.5 to 10◦ 2θ .
Temperature-programmed
reduction (TPR) was carried out in a He flow of 50 ml/min (5 vol% H2), from room temperature to 800 ◦C at a heat- ing rate of 5 ◦C/min.
A TCD was used to monitor the H2
consumption. Before starting
the TPR procedure, the sam-
ple was pretreated at 500 ◦C for 1 h in a helium flow of 100 ml/min
and then cooled down to room temperature.
X-ray photoelectron spectroscopy
(XPS) spectra were recorded on a Shimadzu Kratos AXIS spectrometer, using a monochromatic Al-Kα radiation (1486.6 eV,
225 W), op- erating at a constant transmission energy pass (80 eV). The
C 1s photoelectron peak (binding
energy at 284.6 eV with an
accuracy of ±0.2 eV) was
used as an energy reference. All the spectra were fitted using a Gaussian
method in order to determine the number of components in the peaks. Shirley
background was subtracted in the least-square
fitting. The peak shape was fixed to a mixture of 20% Lorentzian and 80%
Gaussian with an asymmetric parameter of 0.
IR spectra of adsorbed pyridine
were recorded to deter- mine the presence of Brønsted and Lewis acid sites
over
the catalysts, which was performed using a Shimadzu
FTIR- 8700 spectrophotometer with
a resolution of 2 cm−1 and
connected to a PFEIFFER vacuum system. Fifteen
mil- ligrams of the sample was ground and then pressed into a self-supporting
wafer at the pressure of 6000 kg/cm2 for 15 min, which was then mounted into a quartz
IR cell with
CaF2 windows.
Before scanning, the self-supporting wafer of the catalyst was heated at 400 ◦C for 3 h under a resid- ual pressure of 10−6 mbar
before adsorbing an excess of
pyridine
at room temperature, followed by evacuation at 200 ◦C
for 30 min. Note that the FTIR background was
in
situ recorded at room temperature at a vacuum
of 10−6 mbar without
adsorbing pyridine after the self-supporting
wafer
had been
pretreated at 400 ◦C for 3 h under a residual pres- sure of 10−6 mbar. Brønsted and Lewis were quantified from the integrated areas of the bands at
1540 and at 1445 cm−1,
respectively. The integrated areas of these
two bands pro- vide information only about the relative amount
of pyridine, which interacts
with Brønsted and Lewis sites; therefore the
B and L values must be regarded
as relative indications.
2.3.
Catalytic tests
The
reaction was carried out in a fixed-bed microreac- tor, loaded with 0.3 g of
the catalyst, operating at normal pressure in the temperature range 250–700 ◦C
with a feed
gas (120 ml/min)
consisting of 600 ppm pyridine and 5% O2 in helium. Each reaction reached steady state in 30 min;
the products were then analyzed on stream by a Shimadzu GC-17A gas
chromatograph in order to detect the composi- tion under steady state. The
conversion of pyridine and the
yield of NOx were, respectively, calculated based on the inlet
and outlet compositions of the reactor, namely,
Xpyr = (inlet pyridine − outlet pyridine)/(inlet pyridine) × 100% and the yield of NOx = (the detected concentration of NOx )/(the theoretical concentration of NOx) × 100%. An on-line
Shi- madzu NOA-7000 NOx analyzer monitored the concentra-
tion of NOx
in the effluent gas. A
parameter T100, charac- terizing the pyridine oxidation ability of the
catalysts, was defined as the lowest
temperature at which 100% of pyri- dine conversion was realized. Generally, the
lower the T100 value, the higher the pyridine oxidation ability of the
cata- lyst. TNOx,5% stood for the temperature at which the yield
of
NOx is 5%. Note that a new parameter 6T was defined as (TNOx,5% − T100),
where the bigger the 6T value, the
better the NOx control ability the catalyst retained.
3. Results and
discussion
3.1.
Effect of copper content
on the activity of catalysts
Copper, which is a cheap metal compared with cobalt and nickel, as an active metal component
has been widely ap- plied in the SCR deNOx
reaction, to achieve a rather
high deNOx activity. It is well
known that the loading amount
Fig. 1. Effect of copper loading
on the pyridine conversion and NOx control. Table 1
Characteristic temperatures and BET surface
areas of Cu/beta
catalysts with different
copper loadings
Catalysts
|
T100
(◦ C)
|
TNOx,5%
(◦C)
|
6T
(◦C)
|
BET specific
surface area (m2/g)
|
Beta
|
530
|
550
|
20
|
542
|
1% Cu/beta
|
500
|
600
|
100
|
516
|
2% Cu/beta
|
410
|
600
|
190
|
500
|
4% Cu/beta
|
340
|
500
|
160
|
464
|
6% Cu/beta
|
310
|
450
|
140
|
398
|
10% Cu/beta
|
305
|
420
|
115
|
381
|
15% Cu/beta
|
310
|
395
|
85
|
368
|
T100: the lowest
temperature at which the pyridine conversion is 100%.
TNOx,5%: the temperature at which the yield
of NOx is
5%.
6T = TNOx,5% − T100 : an indication of the NOx control
ability.
of active component has a great impact on the
performance of the catalyst [2,3], and highly dispersed active particles have higher
activity than big crystal particles. Many stud- ies have shown that the
activity of catalysts cannot be in- finitely improved through simply increasing
the content of active component when it has reached a certain value, in which
the relationship between the activity and the loading shows a bell-shaped curve. In this research, copper/beta zeo- lite sample was used as the catalyst.
To investigate the effect
of copper loading on the pyridine oxidation activity and the NOx control ability of Cu/beta catalysts, the copper loading
was varied from 0 to 15% (wt%).
Fig.
1 and Table 1 show the pyridine conversion and NOx yield as a function of temperature over the catalysts,
where the curves on the left correspond to the pyridine conversion, while ones on the right correspond to the NOx yield. The ac- tivity of 1% Cu/beta for the pyridine
oxidation was slightly better than that of pure beta. A parameter T100, character- izing the ability of the catalysts to
oxidize pyridine, was defined as the lowest temperature at which 100% of pyri-
dine conversion was realized. Generally, the lower the T100 value, the higher the
ability of the catalyst to oxidize pyri-
Fig. 2. XRD patterns
of Cu/beta with different copper
loadings and different metals/beta samples.
dine. The T100 values of 1% Cu/beta
and pure beta were 500 and 530 ◦C, respectively. When the copper content was in- creased to
2 wt%, the oxidation activity of the catalyst was
much improved, with an obvious
decrease of T100 value
to 410 ◦C. With a continuous increase of the copper content to
4 wt%, the oxidation activity of
the resulting catalyst was further improved, accompanied with an abrupt
reduction of T100 value to 340 ◦C. However, when the copper content was
increased to 6 wt%, the
improvement of oxidation activity of the resultant catalyst was badly limited, with a quite
small drop of T100 value
to 310 ◦C. Whereas, the copper loadings
of 10 and 15 wt% showed a
similar activity for the pyri- dine oxidation to that of 6 wt%, where T100 values
of three catalysts were 310 (6 wt% Cu), 305 (10 wt% Cu), 310 ◦C
(15 wt% Cu), respectively, with an almost same
pyridine conversion behavior, in agreement with the phenomenon of pyridine
oxidation over Ag/γ -Al2O3 catalysts
with differ- ent silver loadings reported by Luo et al. [4]. Based on the
variation of T100 value, the oxidation activity of all the
copper/beta catalysts increased in the following sequence:
beta (T100 = 530
◦C) < 1%
Cu/beta (500 ◦C) < 2% Cu/beta
(410 ◦ C) < 4% Cu/beta (340 ◦C) < 6% Cu/beta (310 ◦ C) ≈
10% Cu/beta (305 ◦ C)
≈ 15% Cu/beta (310 ◦ C).
The XRD patterns (in Fig. 2) of the catalysts show that
below 6 wt% of the copper content on zeolite
beta, only ze- olite beta diffraction peaks are observed. Neither CuO nor metal
Cu diffraction peaks were detected under our experi- mental conditions,
indicating that all the copper on beta has been highly dispersed with smaller
than 5 nm of the parti- cle size. However, it
is not the case for 10% Cu/beta and 15% Cu/beta catalysts, which present rather
clear diffraction peaks of
crystalline CuO oxides in their XRD patterns, in- dicative of the formation of big CuO crystals on zeolite beta. Moreover, increasing the copper loading from 10 to 15 wt% further promoted the growth of big CuO crystal
particles.
Both 1% Cu/beta and 2% Cu/beta catalysts
exhibited a similar NOx yield curve (in Fig. 1). For other four catalysts
4% Cu/beta, 6% Cu/beta, 10% Cu/beta, and 15% Cu/beta, despite their similar
pyridine oxidation activities, their NOx yield behaviors were noticeably different from
one another.
For example, at 500 ◦C
the NOx yield of 4% Cu/beta was
4.8%, and that of 15% Cu/beta was 20.1%. When the tem-
perature was increased to 700 ◦C, the NOx yield of 4% Cu/beta increased to 27.2%,
while that of 15% Cu/beta
even
to 42.5%. Table 1 presents
the characteristic temperatures of the catalysts, which are used to evaluate the activity of
the catalysts. From Table 1, we can observe that the
NOx control ability (6T = TNOx,5% − T100)
of the catalysts first shows an increase trend with increasing the copper
content
from 0 to 2
wt%, then decreases with a
further increase of copper
content to above 2 wt%, typical of a clear bell- shaped curve. Amongst the
Cu/beta catalysts, 2% Cu/beta retained the best NOx
control ability. BET data in Table 1 clearly show that
increasing the copper content in the range of 1–6 wt% largely retarded the
specific surface area of the catalysts, which is in agreement with those for
the catalysts used in the carbon monoxide and methane oxidation [15], but
strikingly improved the pyridine oxidation activity of the catalyst. However, when the copper loading was above
6 wt%, the pyridine conversion (T100 value) could not be correlated with the
specific surface area at all.
3.2.
Effect of different metals
on the activity of catalysts
3.2.1. Property and oxidation activity
of catalysts
In this study, zeolite beta-supported copper (Cu), cobalt (Co), iron
(Fe), and nickel (Ni) catalysts were tested. Since 2% Cu/beta showed the best
NOx control ability (which is the most concerned,
more than the pyridine oxidation abil- ity), the metal loading in the unspecified
studies is main- tained at about 2 wt%. The BET surface area and total pore
volume of the resulting catalysts are shown in Table 2. The surface
areas of all the catalysts were in the range of 471– 500 m2/g, without
significant decrease of the BET surface area
for all the samples. On the other hand, the total pore vol-
umes of the catalysts kept more or less constant
ranging from
0.30 to 0.32 cm3/g. The results
suggest that partial
blockage of some pores has occurred due to the metal-species disper-
sion either in the channels
or outer surface
of beta zeolite. As shown in Fig. 2, only
diffraction peaks of beta are detected for four catalysts without any diffraction
peaks of metal ox- ides, indicative of the high dispersion of these metal oxide
constituents on the surface of beta.
For the catalytic oxidation of
pyridine, the ideally ex- pected reaction is to combust
pyridine completely into CO2, H2O, N2, and NOx
. The lesser amount of NOx the catalyst produces, the better its performance. Fig. 3 depicts the re- lation of pyridine conversion and NOx yield to temperature over beta-supported metal catalysts. Pyridine conversion and NOx yield increase with increasing the reaction temperature, consistent with those
observed by Ismagilov et al. [2,3] but
Catalysts
|
BET surface
area (m2/g)
|
Total pore volume (cm3/g)
|
Catalysts
|
T100
(◦C)
|
TNOx,5%
(◦ C)
|
6T
(◦C)
|
Beta
|
542
|
0.33
|
2% Cu/beta
|
410
|
600
|
190
|
2% Cu/beta
|
500
|
0.32
|
2% Fe/beta
|
450
|
560
|
110
|
2% Co/beta
|
472
|
0.30
|
2% Co/beta
|
470
|
500
|
30
|
2% Ni/beta
|
491
|
0.31
|
2% Ni/beta
|
530
|
560
|
30
|
2% Fe/beta
|
473
|
0.30
|
|
|
|
|
|
|
Specific surface area and total pore volume
of beta supported metal catalysts
Fig. 3. Pyridine conversion and NOx yield over different beta-supported metal catalysts.
Table 3
Characteristic temperatures of
pyridine oxidation on various catalysts
different from the
results of Luo et al. [4], who found that the NOx yield went through a maximum value as the tem- perature increased. The difference may result from different
reactant compositions, in which the reactant mixture
used by Luo was 500 ppm
pyridine in air, while ours was 600 ppm in helium balanced
with 5% O2. Table 3 lists the characteris- tic temperatures of pyridine
oxidation over various
catalysts, at which temperature 100% conversion of pyridine was re-
Fig. 4. H2-TPR profiles
of different beta-supported metal catalysts.
in the following
mechanism: CuO + H2 → Cu0 + H2O,
(1)
alized. Comparatively, 2% Cu/beta still exhibited
the best
pyridine oxidation activity and NOx control ability, and
its
Cu2+
+ 0.5H2 → Cu+
+ H+,
(2)
T100 and 6T values were 410 and 190 ◦C,
respectively. The
(3)
oxidation activity of these catalysts
decreased in the order of 2% Cu/beta >
2% Fe/beta > 2% Co/beta > 2% Ni/beta; while their NOx control ability showed a new sequence 2%
Cu/beta > 2% Fe/beta > 2%
Co/beta ≈ 2% Ni/beta. The
re-
sults show that the catalyst
with a better pyridine oxidation
activity also retains a
stronger NOx control ability, in agree- ment with that of Luo et al. [4].
3.2.2. Temperature-programmed
reduction tests
TPR tests were conducted to determine the chemical state of metals on zeolite beta. In the
TPR profiles of four beta- supported metal catalysts, two main peaks at 222 and 332 ◦C
with a shoulder at 262 ◦C were observed for the sample 2% Cu/beta (in Fig. 4). According to the reports
[16–18], the re-
actions involved in the copper-containing catalysts proceed
where reactions (1) and (2) occur at a lower
temperature than reaction (3) [19]. Thus,
the first peak at 222 ◦C for the sam-
ple 2% Cu/beta
is ascribable to the reduction of CuO to metal
Cu0, and the shoulder at 262 ◦C is assigned to that of Cu2+ to Cu+, while the peak at 332 ◦C is the signal of reduction of Cu+ species to metal Cu0. The TPR
profile of the sample 2% Co/beta
shows one small peak centering at 322 ◦C, at-
tributable to the reduction of cobalt oxides
deposited outside the zeolite [20]. The
absence of reduction peaks from Co2+ to
Co+ or to metal Co0 is
not surprising, as Co2+ species is not readily reduced to Co+ or
to metal Co0 [21].
TPR profile of 2%
Fe/beta shows that the reduction of
iron species occhree steps with
a maximum signal at 317, 382, and 487 ◦ C,
respectively. According to the literature [22–24], the reduction
of unsupported or supported Fe2O3
to metal
Fe0 occurs in three steps below
630 ◦C, with Fe3O4 and FeO as the intermediates. The reduction of Fe3+ species
to metal Fe0 in the charge compensation sites of zeolite re-
Table 4
Specific
surface area and total pore volume of different supports and cata- lysts as
well as surface acidity of different supports
quires the temperatures above 730 ◦C,
but that of Fe3+ to
Fe2+ is
observed to occur at lower
temperatures [25–27].
Therefore, the former two peaks could be assigned to the
Sample BET surface area
(m2/g)
Total pore volume (cm3/g)
B acid
(1540 cm−1)
L acid
(1445 cm−1)
reduction of Fe2O3 “clusters” to metal Fe0,
with Fe3O4 as
intermediate species [28,29], and the third one
correspond-
Beta 542 0.34 1.66 4.23
ZSM-5 256 0.15 3.80 1.05
MCM-41 1330 1.04 0.59 0.62
ing to the reduction of Fe3+ to
Fe2+ [30,31], where Fe3+
γ -Al2O3
128 0.23 1.41 0.75
species come from the oxidation of ion-exchanged Fe2+ ions into zeolite by calcinations [32].
A main peak locating at 557 ◦C,
together with a very small shoulder at 422 ◦C,
is observed in the TPR profile of
2% Ni/beta. According to the literature [33,34], the
main peak at 557 ◦C is ascribable to the reduction of Ni2+ locating
in 12-ring channels [35], while the shoulder
at 422 ◦C to the
reduction of NiO oxide to metal Ni0. As shown in Fig. 4, the
reduction temperature of the samples displays a decreases in the order of 2% Ni/beta > 2% Co/beta > 2% Fe/beta > 2% Cu/beta,
inversely proportional to that of pyridine oxida- tion activity of these catalysts:
2% Cu/beta > 2% Fe/beta > 2% Co/beta > 2% Ni/beta (in Fig. 3), suggesting that the metal reduced readily at a lower
temperature has higher ox- idation activity.
3.3. Effect of different supports
on the activity of catalysts
2% Cu/beta 500 0.32
2% Cu/ZSM-5 235 0.15
2% Cu/MCM-41 695 0.47
2% Cu/γ -Al2 O3 94 0.16
Brønsted and Lewis acidities are quantified into integrated
areas of the ab- sorbances at 1540 cm−1 and at 1445 cm−1, respectively.
3.3.1. Structures of different supports
The copper-loaded catalysts on
different supports were prepared by wet impregnation. It was reported that the
sup- port took an important role in the catalyst activity. In this work, the effect of supports was investigated on the catalysts of Cu/γ -Al2O3, Cu/ZSM-5, Cu/beta, and Cu/MCM-41 with a loading of 2 wt% Cu, which is an
optimal value for show- ing the best NOx
control ability of 2% Cu/beta
catalyst. The X-ray diffraction patterns
of ZSM-5 and beta show that both are highly crystallized. The XRD
pattern of MCM-41 dis- plays a sharp peak of d100 at 39 Å, without
distinct peaks of (110), (200), and (210), suggesting its short-range ordered
mesostructure after the aluminum incorporation. The dif- fraction intensity of γ -Al2O3 is
not so strong as those of zeolites ZSM-5 and
beta.
Table 4
reports the BET surface area and total pore
vol- ume of supports and catalysts. The results show that the impregnation of
copper on the supports led to a decrease of surface area and total pore volume of the catalysts, indicative of the partial blockage
of some pores of these supports by the
loaded copper. Compared with pure MCM-41, the surface area and total pore
volume of 2% Cu/MCM-41 are reduced a lot.
3.3.2. Contribution of supports to the oxidation activity
To reveal the effect of support materials on the
activ- ity of catalysts, the pyridine oxidation was directly con- ducted
over pure supports, such as beta, ZSM-5, MCM-41, and γ -Al2O3. Fig. 5 has shown that the NOx control abil-
Fig. 5.
Pyridine conversion and NOx yield over
different supports.
ity of different supports decreases in the
order of beta >
ZSM-5 >
γ -Al2O3 > MCM-41, in agreement with the or- der of 2%
Cu/beta > 2% Cu/ZSM-5 > 2% Cu/γ -Al2O3 ≈ 2%
Cu/MCM-41 depicted in Fig. 6, where the introduction
of copper species
noticeably improved both the oxidation ac- tivity and the NOx control ability of the catalysts but did not change the trend of NOx control ability of the supports.
This suggests that the NOx control ability of supported copper catalysts is, at least partially, determined by the nature of the supports.
The surface acidity of the
supports was determined by the FTIR spectrum
of pyridine adsorption, and Brønsted and Lewis
acidities were quantified from the
integrated areas
of the bands at 1540 and 1445 cm−1,
respectively. Fig. 7
shows that beta and ZSM-5 are more acidic than γ -Al2O3 and
MCM-41, in which the total acidity of γ -Al2O3 is
stronger than that of MCM-41, corresponding to the ten- dency of NOx control ability of the supports. The data in Table 4 show that the number
of Brønsted acid sites of four acidic supports decreases
in the following sequence, ZSM-5 (3.80) > beta (1.66) > γ -Al2O3 (1.41) >
MCM-41 (0.59);
while that of their Lewis acid
sites decrease in a new order
Fig. 6. Pyridine conversion and NOx yield over different supported copper catalysts.
Fig. 7. Pyridine-adsorption FTIR spectra of different
supports.
of beta (4.23) > ZSM-5
(1.05) >γ -Al2O3 (0.75) > MCM-
41 (0.62), consistent with their NOx control ability
order of beta > ZSM-5 >γ -Al2O3 > MCM-41.
It seems that the bet- ter NOx control ability can be attributed to the
importance of Lewis acidity of samples, rather than to that of Brøn- sted
acidity. Especially, beta showed a better NOx control ability than ZSM-5, but the Brønsted
acidity of beta (1.66) was much weaker than that of ZSM-5 (3.80). This phenom-
enon could be partially assigned with the structure of beta and ZSM-5. It can
be assumed that pyridine molecules or
NOx molecules resulting from the pyridine
oxidation could readily diffuse into or out of large pores (∼ 7
Å) of beta in-
Table 5
Characteristic temperatures of
different copper-loaded catalysts
Catalysts
|
T100
(◦ C)
|
TNOx,5%
(◦C)
|
6T
(◦C)
|
2%
Cu/beta
|
410
|
600
|
190
|
2% Cu/ZSM-5
|
390
|
520
|
130
|
2% Cu/MCM-41
|
430
|
450
|
20
|
2%
Cu/γ -Al2O3
|
400
|
450
|
50
|
stead of small pores (∼ 5.5 Å) of ZSM-5, thus, leading to the
better NOx control ability of zeolite beta [36].
3.3.3. Activity of copper-loaded catalysts
Fig. 6 illustrates the
pyridine conversion and NOx yield over different copper-loaded catalysts.
2% Cu/beta, 2% Cu/ ZSM-5, and 2% Cu/γ -Al2O3 exhibited
a similar pyridine oxidation activity tendency and a close T100 value about
390 ◦C.
2% Cu/MCM-41 was less active for the
pyridine
oxidation as its T100 value was high up to 430 ◦C. These catalysts have shown quite different
NOx control abilities, which are clearly shown on
the right-hand side in Fig. 6. At a reaction temperature as high as 700 ◦ C, the NOx
yield of 2% Cu/beta, 2%
Cu/ZSM-5, 2% Cu/MCM-41, and 2% Cu/
γ -Al2O3 was 20, 30, 37, and
33%, respectively. The oxi- dation activity of pyridine on these catalysts
shows a de- scending tendency of 2% Cu/ZSM-5 > 2% Cu/γ -Al2O3 > 2%
Cu/beta > 2% Cu/MCM-41.
The characteristic temperatures of
the catalysts are listed in Table 5. The lowest T100 = 390 ◦C value showed the best pyridine oxidation activity of 2% Cu/ZSM-5 catalyst; in con-
trast, 2% Cu/MCM-41 catalyst had the worst pyridine oxi-
dation activity because of its highest T100 = 430
◦C value.
Also, 2% Cu/MCM-41 showed the poorest ability
to control
the yield of NOx , attributable to its narrow
temperature win- dow 6T value
of only 20 ◦C. However, 2%
Cu/ZSM-5 did
not show the best NOx control ability, although
its oxidation activity was the
best. 2% Cu/beta presented the best NOx control ability, and its 6T value was as wide as 190 ◦ C,
in- dicating that this catalyst could work very well in a relatively large temperature range.
3.3.4.
Characterization
of surface copper species
It is evident that the supports
have a marked effect on the pyridine oxidation
reaction, especially on the NOx yield. To explain
this phenomenon, the nature and chemical state of copper species on different
supports are characterized by means of XRD, XPS, and TPR techniques.
3.3.4.1. XRD XRD was carried out to examine the crys- talline structure of
the catalysts. Fig. 8 clearly shows the XRD patterns of all
the copper-loaded catalysts. Neither CuO oxide nor metal Cu diffraction
peaks were detected by our method, indicative of high
dispersion of all the Cu species on the supports.
Fig. 8. XRD patterns of supported copper catalysts and
pure CuO.
Fig.
9. XPS spectra of Cu 2p of the catalysts 2% Cu/beta (fresh, dehy- drated, and
treated hydrothermally), 2% Cu/MCM-41, 2% Cu/ZSM-5, and 2% Cu/γ -Al2O3.
3.3.4.2.
XPS
XPS was conducted to provide the informa- tion on the elemental
compositions and the oxidation state. Fig. 9 illustrates
the XPS spectrum of 2% Cu/beta, where peak 1 at 933.29 eV is assigned to 2p3/2 of Cu2+ in
the form
of CuO according to the
literature [37,38]. As we know, the binding energy of Cu2+ species (3d9)
is about 10 eV higher
than that of Cu 2p3/2 transition;
this characteristic is usu-
ally used to determine the
presence of Cu2+ [39]. Fig. 9 clearly shows such a peak 4 at 943.09 eV, which is 9.8 eV
higher than peak 1, in agreement with the result reported in
the literature [37], showing the
existence of CuO species in the 2% Cu/beta sample. Cu2+ 2p1/2 in
CuO and its satel-
Fig.
10. H2-TPR
profiles of different copper-containing catalysts.
lite peak were detected to locate at 953.50 eV (peak
5) and
961.90 eV (peak 7), respectively, consistent
with those re- ported in the literature [38], further
confirming the existence of CuO species in the 2% Cu/beta
sample. Moreover, there were still another
two peaks, namely peak 2 at 935.60
eV and peak 6 at 956.00 eV (in
Fig. 9). To identify peaks 2 and 6, 2% Cu/beta
sample was first dehydrated under vacuum
at
300 ◦C
for 12 h to result in a dull gray dehydrated
sample,
and then the XPS spectrum of
the dehydrated sample was measured (in Fig. 9). Very clearly, the peak at 935.60 eV was
totally removed, and only a very small peak appeared at 956.00 eV. Narayana et al. [40] have attributed the afore- said two peaks to Cu(H2O)62+ complex locating
in the large cages of zeolite,
namely the isolated
Cu2+ ions. Actually, the existence of hydrated Cu2+ ions has been indicated by the light green color of as-prepared 2% Cu/beta catalyst.
The
XPS spectra of Cu species obtained from 2% Cu/ ZSM-5 (933.4,
935.6, 940.5, 943.1,
953.5, 956.0, and
961.8 eV), 2%
Cu/MCM-41 (933.2, 935.6, 943.0, 953.4,
956.0, and 961.8
eV) and 2% Cu/γ -Al2O3 (933.3, 935.6,
943.1, 953.5, 956.0, and 961.9) (in Fig. 9),
showing that both CuO and Cu2+ ions also coexisted in these catalysts. By comparing
the area of XPS peaks of CuO and Cu2+ ions,
we can come to a conclusion that the copper
species in these catalysts are mainly present in the
form of CuO, which is further confirmed by TPR
tests.
3.3.4.3.
TPR
TPR profiles of different copper-loaded cat- alysts, and pure
CuO oxide which is used as a reference,
are presented in Fig. 10. It was found
that for pure CuO ox- ide only one reduction peak emerged at 257 ◦ C, while for all the supported
copper catalysts there were two reduction peaks. 2% Cu/beta displayed two main
TPR peaks at 222 and 332 ◦C, with a shoulder peak at 262 ◦C. The peak at 222 ◦C was ascribable to the reduction of CuO to metal Cu0, the peak at 262 ◦C to the reduction of Cu2+ to
Cu+ ions,
and the peak at 332 ◦C to the reduction of Cu+ to metal Cu0, respectively. It has been reported
that the supports
could en- hance the reduction
of copper oxide species [41], and the reduction temperature of CuO in the CuO/SiO2 decreased
from 297 to 227 ◦C,
along with increasing the dispersion of
CuO on silica [42,43], in
agreement with other results re- ported elsewhere [44,45]. The ratio of the area of the first peak to the second
peak (in Fig. 10)
indicated that highly dispersed CuO was a main form of copper species on beta.
The reduction peak temperature of the catalysts (in Fig. 10)
increased in the order of 2%
Cu/ZSM-5 < 2% Cu/beta ≈
2% Cu/γ -Al2O3 < 2% Cu/MCM-41, inversely proportional to the decreasing
sequence of their pyridine oxidation
ac-
tivity, namely
2% Cu/MCM-41 < 2% Cu/beta < 2% Cu/ γ -Al2O3 < 2% Cu/ZSM-5, suggesting that the copper species reduced
easily at lower
temperatures had higher
ox- idation activity.
The
preceding results have shown that there were two kinds of copper species on
supports, namely, CuO and Cu(H2O)62+.
The percentage of Cu(H2O)62+ ions in to-
tal copper amount was determined
to be around
26.5% for 2% Cu/beta,
21.9% for 2%
Cu/MCM-41, 18.1% for
2% Cu/ZSM-5, and 15.3% for 2% Cu/γ -Al2O3, respec- tively.
This sequence was almost the same as that of NOx control ability of these
catalysts with an
exception of 2% Cu/MCM-41, which showed the poorest NOx control ability
in the pyridine
oxidation, further suggesting
that
Cu(H2O)62+ together with some other factors like the na-
ture of the support took critical roles in retarding the NOx
yield of pyridine oxidation. Many researchers have reported
that Cu2+ ions were much more active than CuO in the SCR
deNOx reaction [46], consistent
with our results that
Cu(H2O)62+ ions really possessed a better NOx control abil-
ity than CuO. This has been further
proven in our hydrother-
mal treatment experiments in which
2% Cu/beta catalyst
was hydrothermally treated at 600 ◦C for 12 h in a helium flow
containing 10% water vapor, prior to catalytic
tests. It was found that the hydrothermally treated 2% Cu/beta showed a slightly
improved oxidation activity but a noticeably worse NOx control ability in comparison with the untreated one for the pyridine oxidation.
3.3.5. Effect of hydrothermal treatment on the Cu/beta catalyst
As described above, Cu/beta
was one of the most promis-
ing catalysts for the pyridine oxidation. Recently, some re- searchers [45,47] reported the low stability of Cu/ZSM-5 catalyst in the
presence of water vapor, which is a
typi- cal condition of diesel engine exhaust gas. As a result, it is worthwhile to investigate the
resistance of Cu/beta cata- lyst to water vapor. To fulfill this target, 2% Cu/beta catalyst
was treated hydrothermally at
600 ◦C for 12 h in a helium
flow containing 10% water vapor, prior to
catalytic tests. The hydrothermal treatment slightly improved the oxidation
activity of 2% Cu/beta catalyst for the pyridine oxidation, in which the
hydrothermally treated 2% Cu/beta showed a
stronger pyridine oxidation
ability than the untreated one at reaction temperatures between
300 and 350 ◦C. The temper-
ature (T100) required for the complete oxidation of
pyridine
was 400 ◦C for the treated
2% Cu/beta and 410 ◦C for the un- treated 2% Cu/beta, respectively.
Additionally, an obvious
increase of NOx yield was observed to be from 20% over the untreated catalyst to 26% over the treated one at 700 ◦C, and 6T value
dropped from 190 ◦C of untreated catalyst to 140 ◦C of treated one.
The hydrothermal treatment has
resulted in the appear- ance of a new XPS peak at the XPS spectrum of
treated 2% Cu/beta sample. In Fig. 9, peak
3 at about 940.6 eV can be observed
for both 2% Cu/beta treated
hydrothermally (933.2, 935.7, 940.6,
943.0, 953.4, 956.0,
and 961.8 eV) and
2% Cu/ZSM-5
(933.4, 935.6, 940.5, 943.1, 953.5,
956.0,
and 961.8 eV).
Thus, it is not likely
that peak 3 at about
940.6 eV belongs to ZSM-5. In addition, this
peak has also been observed from XPS spectrum (not shown in Fig. 9) of 6%
Cu/MCM-41 (933.2, 935.4, 940.6, 942.8, 953.4, 956.0, and 961.9 eV), further
indicating that peak 3 at 940.6 eV is not from ZSM-5. However, peak 3 at about 940.6 eV can- not be observed at all
for 2% Cu/γ -Al2O3, 2% Cu/beta and 2% Cu/MCM-41. Based on these experiments, we can come to a conclusion that peak 3 could
result from Cu species, especially from highly crystallized CuO instead of
highly dispersed CuO.
By comparing the XPS spectra of
untreated 2% Cu/beta and treated one (in Fig. 9), it was found
that the hy- drothermal treatment
led to a
reduction of the amount of
Cu(H2O)62+ in the
catalyst, with a percentage drop from
26.5 to 20.7%.
As reported by other authors
[45,46], the de- crease of activity of ion-exchanged
Cu/ZSM-5 catalyst in the deNOx reaction is that most of the isolated Cu2+ ions
were converted to CuO by the hydrothermal
treatment. The aforesaid results have shown that CuO had a better oxi- dation
activity but a poorer ability to reduce the yield of
NOx than Cu2+ ions,
and that the NOx control ability var- ied with different supports.
Thus, the loss of Cu(H2O)62+
ions converted to CuO oxides on beta during
the hydrother- mal treatment could be a partial reason for the fact that the 2%
Cu/beta catalyst treated hydrothermally suffered from a loss in the NOx control ability. Generally speaking, a de- ciding role in
the NOx control ability of the catalysts could involve the cointeraction of some factors,
such as the surface acidity of catalysts, the structure of supports, and the amount
of Cu(H2O)62+ ions on the supports.
4.
Conclusions
The activity of supported copper catalysts for the pyridine oxidation increased with the increase of copper content
from 0 to 6 wt%. Further
increasing the copper content showed
no contribution in improving the pyridine oxidation activity of the
resulting catalyst. The NOx control ability (6T ) of Cu- loaded catalysts went
through a maximum value with the
increase of
copper content from 0 to 15 wt%. 2% Cu/beta exhibited
the best NOx control ability with
a 6T value of 190 ◦ C,
while the 6T value of 6% Cu/beta
dropped to
140 ◦C.
When the catalyst worked in a rather low tempera-
ture range, 6% Cu/beta showed a
better activity, because it could completely convert
pyridine at only 310 ◦C. However, in a high temperature range from 400 to 700 ◦ C, 2% Cu/beta was supposed to be a better choice, as it showed
a better NOx control ability.
Pyridine oxidation ability and NOx control
ability could not be correlated with the
surface area of the catalysts at all.
Cu/beta was the most active one among Cu/beta,
Fe/beta, Co/beta, and Ni/beta
catalysts, and the catalyst that could be reduced at lower temperature had a
better oxidation activ- ity. The number of Brønsted acid sites of four acidic sup- ports
decreases in the following sequence ZSM-5 (3.80) > beta (1.66) > γ -Al2O3 (1.41) > MCM-41 (0.59); while
that of their Lewis acid sites
decreases in a new order of
beta (4.23) > ZSM-5
(1.05) >γ -Al2O3 (0.75) > MCM-41
(0.62). It seems that the better NOx control ability can be
at- tributed
to the importance of Lewis
acidity of samples,
rather than to that of Brønsted acidity.
2% Cu/beta, 2% Cu/ZSM-5 and 2% Cu/γ -Al2O3 exhib-
ited a similar pyridine oxidation activity tendency and a close T100 value about 390 ◦C. 2% Cu/MCM-41 was less ac- tive for the pyridine oxidation as
its T100 value
was high up
to 430 ◦C.
At a reaction temperature as high as 700 ◦ C,
the
NOx yield of 2%
Cu/beta, 2% Cu/ZSM-5, 2% Cu/MCM-41,
and 2% Cu/γ -Al2O3 were 20, 30, 37, and 33%, respectively. It
was found that Cu(H2O)62+ ions
showed a better NOx
control
ability but poorer activity for the pyridine oxidation than CuO oxides, and
that the NOx control ability varied with different
supports. The hydrothermal treatment retarded the NOx control ability of 2% Cu/beta catalyst,
partially due to the loss of Cu(H2O)62+ ions converted to CuO oxides on beta during the hydrothermal treatment. The cointeraction
of some factors, such as the surface acidity of
catalysts, the structure of
supports, and the amount of Cu(H2O)62+ ions
on the supports, played a deciding role in the NOx
control ability of the
catalysts.
References
[1] Z.R. Ismagilov, M.A. Kerzhentsev, Catal.
Rev.-Sci. Eng. 32 (1990)
51. [2] Z.R. Ismagilov, M.A. Kerzhentsev,
V.I. Besedin, T.L. Susharina, Re-
act. Kinet.
Catal. Lett. 23 (1983) 43.
[3] Z.R. Ismagilov, M.A. Kerzhentsev, V.I.
Besedin, T.L. Susharina, Re- act. Kinet. Catal. Lett. 23 (1983) 49.
[4] M.F. Luo, Y.J. Zhong, M. Chen, B. Zhu, X.X.
Yuan, Environ. Sci. 17 (1996) 52.
[5] M.F. Luo, M. Yu, X.X. Yuan, Chinese J.
Appl. Chem. 12 (1995) 87– 89.
[6] N. Crain,
S. Tebbal, L. Li, E.F. Golyna, Ind. Eng. Chem. Res. 32 (1993)
2259.
[7] T.J. Wightman, Studies in Supercritical Wet
Air Oxidation, M.S. the- sis, University of California at Berkeley, Berkeley,
CA, 1981.
[8] A.R. Katritzky, R.A.
Barcock, Energy Fuels 8 (1994) 990.
[9] S.N.V.K. Aki, M.A. Abraham,
Chem. Eng. Sci. 54 (1999) 3533.
[10] Q.H. Xia, S.C. Shen, J. Song, S. Kawi, K.
Hidajat, J. Catal. 219 (2003) 74.
[11] K.J. Chao, T.C. Tasi, M.S. Chen, I. Wang,
J. Chem. Soc., Faraday Trans. 1 76 (1981) 77.
[12] R.A.
Kensington, G.R. Landolt, US patent 3,702,886, 1972.
[13] S.C. Shen,
S. Kawi, Chem. Lett. (1999) 1293.
[14] S.C. Shen,
S. Kawi, Stud. Surf. Sci. Catal. 129 (2000) 227.
[15] P.W. Park,
J.S. Ledford, Appl. Catal. B 15 (1998) 221.
[16] J. Sárkány, J.L. d’Itri, W.M.H. Sachtler, Catal. Lett. 16
(1992) 241. [17] P.A. Jacobs, M.
Tielen, J. Linart,
J.B. Uytterhoeven, H.
Beyer,
J. Chem. Soc., Faraday Trans. 1 72 (1976) 2793.
[18] S.J. Gentry, N.W. Hurst, A. Jones, J.
Chem. Soc., Faraday Trans. 1 75 (1979) 1688.
[19] M.C.
Campa, V. Indovina, G. Minelli, G.
Moretti, I. Pettiti, P. Porta,
A. Riccio,
Catal. Lett. 23 (1993) 141.
[20] R.S. Cruz, A.J.S. Mascarenhas, H.M.C. Andrade,
Appl. Catal. B 18 (1998) 223.
[21] Y.J. Li,
T.L. Slager, J.N. Armor, J. Catal. 150 (1994) 388.
[22] E.E.
Unmuth, L.H. Schwartz, J.B. Butt, J. Catal. 63 (1980) 404.
[23] R. Brown,
M.E. Cooper, D.A. Whan, Appl. Catal. 3 (1982) 177.
[24] O.J. Wimmers, P. Arnoldy, J.A. Moulijn, J.
Phys. Chem. 90 (1986) 1331.
[25] F. Mahoney, R. Rudham, J.V. Summers, J. Chem. Soc., Faraday Trans. 1 75 (1979) 314.
[26] K. Inamura, R. Iwamoto, A. Iino, T. Takyu, J. Catal. 142
(1993) 274. [27] H.Y. Chen, W.M.H. Sachtler, Catal. Today 42 (1998) 73.
[28]
B. Coq, M. Mauvezin, G. Delahay, S. Kieger, J. Catal. 195 (2000) 298. [29] T.V. Voskoboinikov, H.Y.
Chen, W.M.H. Sachtler, Appl.
Catal. B 19
(1998) 279.
[30] J.F. Jia, J.Y. Shen, L.W. Lin, Z.S. Xu, T.
Zhang, D.B. Liang, J. Mol.
Catal. A: Chem.
138 (1999) 177.
[31] M.V. Cagnoli, N.G. Gallegos, A.M. Alvarez,
J.F. Bengoa, A.A. Ye- ramián, M. Schmal, S.G. Marchetti, Appl. Catal. A 230
(2002) 169.
[32] R.Q. Long,
R.T. Yang, J. Catal. 194 (2000) 80.
[33] B.W.
Hoffer, A.D. Langeveld, J.P. Janssens, R.L.C. Bonné, C.M. Lok,
J.A. Moulijn,
J. Catal. 192 (2000) 432.
[34] B. Vos, E.
Poels, A. Bliek, J. Catal. 198 (2001) 77.
[35] M. Afzal, G. Yasmeen, M. Saleem, J. Afzal,
J. Therm. Anal. Calor. 62 (2000) 277.
[36] J.Y. Yan, G.D. Lei, W.M.H. Sachtler, H.H.
Kung, J. Catal. 161 (1996) 43.
[37] J.F. Xu,
W. Ji, Z.X. Shen, S.H. Tang, X.R. Ye, D.Z. Jia, X.Q. Xin,
J. Solid State
Chem. 147 (1997) 516.
[38] R.P.
Vasquez, Surf. Sci. Spect. 5 (1998) 262.
[39] M.A. Kohler, H.E. Curry-Hyde, A.E. Hughes,
B.A. Sexton, N.W. Cant, J. Catal. 108 (1987) 323.
[40] M.
Narayana, S. Contatini, L. Kevan, J. Catal. 94 (1985) 370.
[41] Y.H. Hu, L. Dong, M.M. Shen, D. Liu, J.
Wang, W.P. Ding, Y. Chen, Appl. Catal. B 31 (2001) 61.
[42] S.D. Robertson, B.D. McNicol, J.H. de
Baas, S.C. Cloet, J.W. Jenkins, J. Catal. 37 (1975) 424.
[43] F.S. Delk,
A. Vavera, J. Catal. 85 (1984) 380.
[44] G.
Delahay, B. Coq, L. Broussous, Appl. Catal. B 12 (1997) 49.
[45] C.
Torre-Abreu, M.F. Ribeiro, C. Henriques, G. Delahay, Appl. Catal.
B 12 (1997)
249.
[46] K.C.C. Kharas, H.J. Robota, D.J. Liu, Appl. Catal. B 2
(1993) 225. [47] C.E. Quincoces, A. Kikot, E.I. Basaldella, M.G. Gonzalez,
Ind. Eng.
Chem. Res. 38 (1999) 4236.