Selective catalytic reduction of NO over Cu-Al-MCM-41

Article abstract of DOI:10.1016/j.jcat.2004.07.016

Al-MCM-41 and Cu-Al-MCM-41 were prepared by a modified hydrothermal method. The characterization using XRD, NMR, and N₂ adsorption isotherms revealed that the structure of Al-MCM-41 remained unchanged after the complete exchange of protons for copper ions. Cu-Al-MCM-41 catalysts were studied for the selective catalytic reduction (SCR) of NO by C₃H₆ in the presence of excess oxygen. Particular attention was paid to the influence of both copper content and Si/Al ratio on the nature of the active copper sites and, therefore, on the performance of the catalysts. It was observed that the most active catalyst have ～100% of copper exchange. For the underexchanged catalyst of the same copper content, it was found that the NO reduction activity decreased with decreasing Si/Al ratio. H₂-TPR and NO-TPD experiments showed that copper is mainly in the form of isolated Cu²⁺ ions in Cu-Al-MCM-41 catalysts with copper exchange ≤100%, whereas at higher copper loadings CuO species are also present. Moreover, when the catalyst Si/Al ratio increases copper is more easily reduced. These results are in agreement with the view accepted in the field of Cu-zeolite catalysts that the isolated Cu ²⁺ ions are involved in the SCR of NO by propylene.

Full text of DOI:10.1016/j.jcat.2004.07.016

Journal of Catalysis 227 (2004) 242–252
www.elsevier.com/locate/jcat
Selective catalytic reduction of NO over Cu-Al-MCM-41
a
b,∗
c
b
d
Ying Wan , Jianxin Ma , Zheng Wang , Wei Zhou , Serge Kaliaguine
a
Department of Chemistry, Shanghai Normal University, Shanghai 200234, China
b
Clean Energy Automotive Engineering Center, Tongji University, Shanghai 200092, China
c
Institute of Industrial Catalysis, East China University of Science & Technology, Shanghai 200237, China
d
Department of Chemical Engineering, Laval University, Québec, Canada
Received 3 March 2004; revised 6 July 2004; accepted 10 July 2004
Abstract
Al-MCM-41 and Cu-Al-MCM-41 were prepared by a modified hydrothermal method. The characterization using XRD, NMR, and N
2
adsorption isotherms revealed that the structure of Al-MCM-41 remained unchanged after the complete exchange of protons for copper
ions. Cu-Al-MCM-41 catalysts were studied for the selective catalytic reduction (SCR) of NO by C H in the presence of excess oxygen.
3
6
Particular attention was paid to the influence of both copper content and Si/Al ratio on the nature of the active copper sites and, therefore,
on the performance of the catalysts. It was observed that the most active catalyst have ∼100% of copper exchange. For the underexchanged
catalyst of the same copper content, it was found that the NO reduction activity decreased with decreasing Si/Al ratio. H -TPR and NO-TPD
2
2
+
experiments showed that copper is mainly in the form of isolated Cu ions in Cu-Al-MCM-41 catalysts with copper exchange ꢀ 100%,
whereas at higher copper loadings CuO species are also present. Moreover, when the catalyst Si/Al ratio increases copper is more easily
2
+
reduced. These results are in agreement with the view accepted in the field of Cu-zeolite catalysts that the isolated Cu ions are involved in
the SCR of NO by propylene.

2004 Elsevier Inc. All rights reserved.
Keywords: Cu-Al-MCM-41 catalyst; Nitric oxide; Selective catalytic reduction; Copper content; Si/Al ratio
1
. Introduction
exchanged ZSM-5, mordenite and ferrite were found active,
whereas zeolites X and Y were shown to be poor supports
[
2,3]. This indicated that the zeolite framework composi-
Lean-burn gasoline and diesel engines are of high fuel ef-
tion played an important role for the copper ion-exchanged
zeolites in the SCR reaction. Considering the similarities be-
tween MCM-41 and zeolites [8,9], it is worth exploring its
properties as a support for Cu-based deNOx catalysts. Inten-
sive studies have however not yet been reported [4].
The catalytic activity of Cu-exchanged zeolites for NO
conversion to N2 is known to depend on the amount of cop-
per as well as on the aluminum content of the host zeolite
ficiency and the application of this technology can therefore
reduce the amount of CO2 emission. However, due to the en-
gine operation at a high air to fuel (A/F) ratio, excess oxygen
exists in the exhaust gas and thus, the removal of NOx be-
comes difficult [1]. The selective catalytic reduction of NOx
by hydrocarbons (HC-SCR) in the presence of excess oxy-
gen has been recognized as a most challenging technology.
Ever since Iwamoto et al. [2] reported that transition-metal
ion-exchanged zeolites, in particular Cu-ZSM-5, showed
high HC-SCR activity, the copper-containing zeolite cata-
lysts have been comprehensivelyinvestigated [3–7]. Copper-
[
5–7]. For the different zeolite framework, the copper con-
tent and Si/Al ratio have different effects on the catalytic ac-
tivity. On Cu-MOR catalysts, the catalytic activity increases
with the increment of the exchange level, while it is almost
independent of copper content in Cu-FAU [6,7]. Addition-
ally, the nature and role of copper have been widely dis-
cussed [10,11]. Under reaction conditions it has been shown
*
Corresponding author. Fax: +86-21-65981022.
E-mail address: jxma@fcv-sh.com (J. Ma).
0021-9517/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2004.07.016

Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
243
that copper exhibits dynamic interconversion between +1
Table 1
Textural characteristics of Al-MCM-41 and Cu-Al-MCM-41 catalysts
and +2 oxidation states [12,13] and that copper is present as
isolated ions or small oxide clusters [7,13]. Therefore, the ef-
fects of Si/Al ratio and copper loading on the performance of
Cu-Al-MCM-41 catalysts deserve attention. Moreover, the
nature of copper in Cu-Al-MCM-41 requires further specifi-
cation.
Catalyst
Si/Al
S
BET
Pore size Mesopore volume
2
(m /g) (nm)
(ml/g)
Al-MCM-41-100
Al-MCM-41-30
Al-MCM-41-10
Al-MCM-41-3
Cu-Al-MCM-41-10-31
Cu-Al-MCM-41-10-61
Cu-Al-MCM-41-10-93
Cu-Al-MCM-41-10-120
100
30
10
3
10
10
10
10
1080
1050
984
873
962
846
842
728
3.7
3.6
3.4
3.2
3.3
3.1
3.1
2.9
0.92
0.92
0.81
0.76
0.79
0.69
0.67
0.61
The present contribution reports a study of NO SCR with
propylene under lean-burn conditions over Cu-exchanged
Al-MCM-41 catalysts and of the nature of the copper species
present in these catalysts. Particular attention is paid to the
influence of both copper-exchange level and Si/Al ratio on
the nature of the active copper sites and, therefore, on the
performance of the catalysts.
(
II) acetate, of appropriate concentration, at room temper-
ature under continuous stirring. The above process was re-
peated several times, each time being carried out for 24 h.
After the exchange, the catalysts were filtered, washed thor-
2
. Experimental
◦
oughly with deionized water, dried at 120 C, and heated at
◦
5
40 C for 4 h. The catalysts were analyzed using atomic
2
.1. Catalyst preparation and characterization
The Al-MCM-41 samples were synthesized according to
absorption spectroscopy, the whole rock fusion, and induc-
tively coupled plasma (ICP) methods to determine the cop-
per content and Si/Al ratio. The extent of copper-exchange
level was calculated assuming that one divalent copper ion
can be exchanged for two protons or sodium ions. The spec-
ifications of the catalysts calculated using the results of the
analysis are given in Table 1. For example, Cu-Al-MCM-41-
the procedure given in the literature [14]. An aqueous so-
lution of tetramethylammonium hydroxide (TMAOH, 25%,
Shanghai Chemical Reagent Co. (SHCR)) was well mixed
with tetraethyl orthosilicate (TEOS, 98%, SHCR) to form
mixture A. Sodium aluminate (NaAlO2, 41 wt% Na2O,
1
0-93 stands for a Cu ion-exchanged Al-MCM-41 catalyst
5
4 wt% Al2O3, SHCR) was added to a clear aqueous solu-
tion composed of n-tetradecyltrimethylammonium bromide
CTMABr, 99%, SHCR), n-tetradecyltrimethylammonium
with an initial Si/Al ratio of 10 and Cu ion-exchanged level
of 93%.
The catalysts used in this study were characterized using
X-ray powder diffraction, N2 adsorption isotherm measure-
(
hydroxide (CTMAOH), and NaOH (98%, SHCR) under stir-
ring to form mixture B. CTMAOH was obtained by the
ion exchange of CTMABr in a column packed with anion-
exchange resin. Mixture B was added to mixture A under
stirring. After 30 min, a certain amount of deionized water
was added and stirring was continued for 10 min further un-
til the pH value of the reaction mixture remained constant.
A gel of the following molar composition was obtained,
27
ments, and Al MAS NMR techniques. The X-ray pow-
der diffraction pattern was carried out on a Rigaku D/Max-
RB diffractometer with CuKα radiation. The samples were
loaded on a sample holder with a depth of 1 mm. N2 ad-
◦
sorption isotherm was measured at −196 C using a ST-03A
◦
analyzer. The catalysts were outgassed at 300 C overnight
before measurement. The catalyst pore size distribution was
calculated using the Dollimore-Heal formula. The specific
surface area (SBET) of samples was determined from the lin-
+
1
SiO2 : xAl2O3 : 0.23CTMA : 0.15Na2O
:
0.089TMAOH : 125H2O,
2
7
ear part of BET plot (p/p0 = 0.05–0.25). Al MAS NMR
where x could be regulated by controlling the ratio of TEOS
to NaAlO2. The gel was thereafter transferred into an auto-
clave with a polytetrafluoroethylene (PTFE) lining and sub-
jected to hydrothermal treatment under autogenous pressure
spectra were recorded at 104.26 MHz using a Brucker AMX-
4
00 spectrometer. The catalysts were equilibrated under the
saturated water vapor of an NH4Cl solution before packing
into NMR-MAS rotors. Chemical shifts were referred to ex-
ternal Al(H2O)6 in AlCl3 aqueous solution.
◦
3+
at 110 C, for a period of 72 h. After the solution was fil-
◦
tered, the solid was washed, dried, and calcined at 540 C
for 10 h in a flow of air. The obtained samples exhibited an
ordered structure of pore arrays. All Al atoms were incor-
porated tetrahedrally in the MCM-41 structure after calci-
nation. Moreover, the samples showed a higher thermal and
hydrothermal stability. A high order of mesoporous struc-
2.2. Temperature-programmed reduction and desorption
The catalyst reducibility was characterized by tempera-
ture-programmed reduction (TPR) analysis. A U-shaped
quartz cell of 6 mm o.d., 4 mm i.d. was used as the sam-
ple chamber. The temperature-programmedreduction exper-
◦
ture was kept after aging at 700 C in dry air whereas a fair
mesoporous structure was maintained after treatment in wet
air (10% H2O) at the same temperature.
The copper-containing catalysts were prepared by ion ex-
change of Al-MCM-41 with an aqueous solution of copper
iments were performed with H as the reducing agent. Prior
2
to TPR experiment, the catalyst was heated in situ under
◦
flowing He (50 ml/min) at 500 C for 2 h. Reduction of the
◦
catalyst was carried out from room temperature to 700 C in

244
Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
Fig. 1. X-ray diffraction pattern of Al-MCM-41 and Cu-Al-MCM-41
catalysts: (a) Al-MCM-41-10, (b) Cu-Al-MCM-41-10-31, and (c) Cu-
Al-MCM-41-10-93.
◦
a flow of 5% H2/N2 (30 ml/min) at 10 C/min. The con-
Fig. 2. NMR spectra of Al-MCM-41 and Cu-Al-MCM-41 catalysts: (a)
Al-MCM-41-10, (b) Cu-Al-MCM-41-10-31, (c) Cu-Al-MCM-41-10-63,
and (d) Cu-Al-MCM-41-10-93.
sumption of H2 was monitored continuously using a thermal
conductivity detector. The water produced during the reduc-
tion was trapped in a 5 A molecular sieve column.
3
. Results
Prior to the TPD measurements, the catalysts (about
5
00 mg) were pretreated in a way similar to that de-
3
.1. Catalyst characterization
scribed for TPR experiments. The adsorption of NO was
performed at room temperature by flowing 50 ml/min of
NO (0.8 vol%)/He for 2 h. The catalyst was then flushed
with helium (50 ml/min) at room temperature to remove
all the physically adsorbed species. The TPD experiments
were then carried out from room temperature to 700 C, at
1
catalysts were analyzed using a chemiluminescence gas an-
alyzer (ML9841AS, Monitor Labs, Inc.).
The XRD patterns of Al-MCM-41-10 and copper-ex-
changed Al-MCM-41 samples (Cu-Al-MCM-41-10-31 and
Cu-Al-MCM-41-10-93)are depicted in Fig. 1. They are sim-
ilar to those reported in the literature for MCM-41 [15]. Ta-
ble 1 lists the BET surface area, pore size, and mesopore vol-
ume for Al-MCM-41 and Cu-Al-MCM-41-10 catalysts. The
specific surface area and pore diameter of the catalysts were
◦
◦
0 C/min, under He flow. The gases desorbing from the
2
higher are 700 m /g and about 3.0 nm, respectively. Solid-
state 27Al MAS-NMR investigation (Fig. 2) shows further
that the Al atoms were in tetrahedral coordination [16]. The
results indicated that the uniformly arranged mesoporous
framework of MCM-41 could still remain even after it has
gone through the process of ion exchange, drying, and high-
temperature calcination.
2
.3. Evaluation of catalytic activity
The reduction reaction was carried out in a fixed-bed flow
reactor (a 10-mm diameter by 270-mm-long quartz tube).
The reactor was heated externally in an electrical furnace
and the temperature of the catalyst bed was controlled to
◦
within 1 C. For each measurement 0.1 g of catalyst with
6
0
0–80 mesh was used. The feed gas containing 0.1 vol% NO,
.1 vol% C3H6, 2 vol% O2, and balanced with highly pure
3.2. Catalytic performance
helium was fed into the reactor at a total flow of 500 ml/min.
3.2.1. Effect of copper content
−1
The GHSV was 100,000 h . The concentrations of NO and
NOx in the feed and effluent were monitored continuously
using the chemiluminescence gas analyzer. Two parallel-
arranged gas chromatographs (one with a FID detector and
a GDX-403 column and one with a TCD detector and a 5 A
molecular sieve column) were used to analyze the concen-
trations of C3H6, CO, CO2, O2, and N2 in the reaction gas
stream.
The reduction of NO by C3H6 in the presence of oxy-
gen was first conducted over the nonexchanged Al-MCM-41
sample with a Si/Al ratio of 10. Despite a C3H6 oxidation
conversion of about 80%, the concentration of NO does not
◦
change over the temperature range up to 550 C, indicating
that Al-MCM-41 itself is inactive for SCR of NO. However,
with copper-exchanged Al-MCM-41, a significant NO con-
version is reached.

Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
245
Fig. 4. The ratio of converted C H /converted NO as a function of
3
6
temperature: (2) Cu-Al-MCM-41-10-31, (1) Cu-Al-MCM-41-10-61, (×)
Cu-Al-MCM-41-10-93, and (a) Cu-Al-MCM-41-10-120 (reaction condi-
tions: GHSV = 100,000 h 1, 0.1 vol% NO, 0.1 vol% C H , 2 vol% O ).
−
3
6
2
N2 is the only product of NO reduction, but as noted above
some NO2 was observed at higher temperatures. This be-
havior indicates that there are three reaction stages during
the C3H6-SCR. At lower temperatures, the 1/1 stoichiomet-
ric ratio indicates that NO reduction and C3H6 oxidation
are successive reactions of the same intermediate adsorbed
species over Cu-exchanged catalyst. Some parallel C3H6 ox-
idation occurs under these conditions when overexchanged
oxidic copper is also present. As temperature is increased
competing propylene oxidation occurs, parallel to NO reduc-
tion. At still higher temperature propylene deep oxidation
predominates and NO reduction is inhibited due to the cor-
responding depletion of the reducing agent. This reduced
NO conversion rate leads to the observed maximum in NO
conversion (Fig. 3a) and seems to be associated with the ap-
pearance of NO2.
Fig. 3. Catalytic performance of Cu-MCM-41 catalysts with Si/Al =
0 and different copper contents for C H -SCR reaction: (2) Cu-
1
3
6
Al-MCM-41-10-31, (1) Cu-Al-MCM-41-10-61, (×) Cu-Al-MCM-41-
1
0-93, and (a) Cu-Al-MCM-41-10-120 (reaction conditions: GHSV =
1
−
100,000 h , 0.1 vol% NO, 0.1 vol% C H , 2 vol% O ).
3
6
2
In order to investigate whether the effect of the copper
content on NO SCR is dependent on the Al-MCM-41 Si/Al
ratio, catalytic tests were carried out on another series of cat-
alysts with Si/Al = 30 and different copper contents (0.22
Fig. 3a shows the variation of NO conversion for four
Cu-Al-MCM-41 catalysts with a Si/Al ratio of 10 but differ-
ent copper loadings (0.85 ꢀ Cu ꢀ 3.2 wt%). As the copper
loading in the catalyst increased from 0.85 to 2.45 wt%, the
maximum conversion in NO increased remarkably from 36
ꢀ
Cu ꢀ 1.1 wt%). Fig. 5 shows the catalytic results as a
function of temperature. The same dependenceon the copper
content existed in the conversion of C3H6 as for the catalysts
with Si/Al = 10. The increase in copper content resulted
in a steady increase in C H conversion at constant reac-
to 79%, and the corresponding temperature decreased from
◦
3
95 to 360 C. Nevertheless, further introduction of copper
decreased the NO reduction. N2 is the only product at tem-
peratures below the maximum conversion. Some NO2 (yield
no more than 2%) was also observed at higher temperatures.
A different behavior is observed in the conversion of C3H6.
Fig. 3b indeed shows that the reaction temperature required
for a given conversion of C3H6 steadily decreases with an
increase in copper loading.
Fig. 4 shows the curves of converted C3H6/converted NO
as a function of temperature. For the catalysts with copper
exchange lower than 100%, the ratio of the conversions for
the two reactants is close to 1 at lower temperatures. This ra-
tio increases at higher temperatures. At lower temperatures,
3
6
tion temperature. Moreover, the catalyst with copper content
close to the exchange capacity was the most active for NO
reduction. There still exist three steps in NO conversion. Ta-
ble 2 lists the maximal NO conversion and corresponding
temperature over Cu-Al-MCM-41. The maximal NO con-
version appeared at different copper loadings, i.e., 2.45 wt%
vs 0.85 wt%, corresponding to 100% exchanged level. This
was in contrast to the results shown for Cu-ZSM-5 by Kha-
ras et al. [17], who reported that the most active catalyst for
reduction of NO using propene had ion-exchange levels be-
tween 150 and 200. However, this result was obtained on the

246
Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
Fig. 5. Catalytic performance of Cu-MCM-41 with different copper contents for C H -SCR reaction with Si/Al = 30: (2) Cu-Al-MCM-41-30-25, (1)
3
6
−1
Cu-Al-MCM-41-30-60, (×) Cu-Al-MCM-41-30-97, and (a) Cu-Al-MCM-41-30-12 (reaction conditions: GHSV = 100,000 h , 0.1 vol% NO, 0.1 vol%
C H , 2 vol% O ).
3
6
2
Table 2
Effect of exchange of copper on C H -SCR reaction over Cu-Al-MCM-41
3
6
b
Catalyst
Si/Al
Cu content
wt%)
Cu exchanged
level (%)
NO conversion
a
(
◦
Xmax (%)
Tmax ( C)
Cu-Al-MCM-41-10-31
Cu-Al-MCM-41-10-61
Cu-Al-MCM-41-10-93
Cu-Al-MCM-41-10-120
Cu-Al-MCM-41-30-25
Cu-Al-MCM-41-30-60
Cu-Al-MCM-41-30-97
Cu-Al-MCM-41-30-127
Cu-Al-MCM-41-100-322
Cu-Al-MCM-41-3-9
10
10
10
10
30
30
30
30
100
3
0.85
1.6
2.45
3.2
0.22
0.52
0.85
1.1
0.85
0.85
31
61
93
120
25
60
36
61
79
65
24
49
60
53
28
26
395
380
360
350
455
385
370
360
356
480
97
127
322
9
a
The % of exchange is calculated as 2(n(Cu)/n(Al)) × 100 %.
−1
b
Reaction conditions: GHSV = 100,000 h , 0.1 vol% NO, 0.1 vol% C H , 2 vol% O .
3
6
2
basis of only five catalysts with a comparatively wide range
of 60 to 500%. In addition, it was reported that the most
active Cu-MFI catalysts for selective reduction of NOx were
with exchange levels between 80 and 100%, which coincides
with the present results [6,18].
However, when the Si/Al ratio increased to 100, the catalyst
exhibited a higher NO conversion at low temperatures but a
lower conversion at high temperatures than the catalyst with
Si/Al = 30. Cu-Al-MCM-41-100-322 was clearly overex-
changed and the complete oxidation of propylene over this
◦
Both Cu-Al-MCM-41-10-120 and Cu-Al-MCM-41-30-
catalyst above 360 C inhibited the reduction of nitric oxide.
1
27 catalysts showed a lower NO conversion but a higher
C3H6 conversion at low temperatures than Cu-Al-MCM-41-
0-93 and Cu-Al-MCM-41-30-97, respectively, suggesting
These results show a trend similar to that reported by Torre-
Abreu et al. over Cu-ZSM-5 [6]. These authors reported that
a catalyst with a Si/Al ratio of 27 was the most active among
catalysts having the same copper loading and Si/Al ratios of
11, 27, 50, and 100.
1
again the participation of non-ionic copper species in propy-
lene oxidation.
3
.2.2. Effect of Si/Al ratio
To analyze the effect of Si/Al ratio in Al-MCM-41 on
3.2.3. Preliminary study of the effect of water vapor
Since water was shown to be a poison for Cu-ZSM-5, it
is especially important to examine the effect of water vapor
on the activity of Cu-Al-MCM-41 in NO SCR. A full study
is currently undertaken in our laboratory and only some pre-
liminary results are reported here as Figs. 7 and 8.
The results reported in Fig. 7 were obtained for NO and
C3H6 conversions over Cu-Al-MCM-41-10-61 both with a
dry feed and in the presence of 10% of H2O in the feed
NO SCR, four catalysts with approximately the same copper
content of 0.85 wt% and different Si/Al ratios (3 ꢀ Si/Al ꢀ
1
00) were prepared. The catalytic behavior of these samples
as a function of temperature is illustrated in Fig. 6. Com-
paring the underexchanged catalysts with Si/Al ratio from 3
to 30 shows the Si/Al ratio of 30 yielded the highest NO
◦
conversion, this effect being more pronounced at 370 C.

Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
247
Fig. 6. Catalytic performance of Cu-MCM-41 catalysts with different Si/Al ratios for C H -SCR reaction: (2) Cu-Al-MCM-41-100-322, (1) Cu-Al-
3
6
−1
MCM-41-30-97, (×) Cu-Al-MCM-41-10-31, and (a) Cu-Al-MCM-41-3-9 (reaction conditions: GHSV = 100,000 h , 0.1 vol% NO, 0.1 vol% C H ,
3
6
2
vol% O ).
2
Fig. 7. Effect of 10% H O on the selective reduction of NO over Cu-Al-MCM-41-10-61 (2 1, NO and C H conversion under dry condition; Q P, NO and
2
3 6
−1
C H conversion under wet condition; 2 Q, NO conversion; 1 P, C H conversion) (reaction conditions: GHSV = 200,000 h , 0.1 vol% NO, 0.1 vol%
3
6
3 6
C H , 2 vol% O ).
3
6
2
gas. Only a minor decrease in catalytic activity is observed
upon introduction of water vapor. The maximum NO conver-
sion reached 37% at 370 C with the dry feed. Introducing
ꢀ Cu ꢀ 3.2 wt%) are illustrated in Fig. 9. For the underex-
changed catalysts, the profiles were similar, whereas that of
the overexchanged sample was different. The former showed
◦
◦
1
0% H2O in the feed gas decreased the maximum conver-
two reduction peaks: the first one at 170–215 C, and the
◦
sion to 34% and the corresponding peak temperature was
second one in the range 480–570 C, which suggests a two-
◦
2+
about 380 C.
step reduction process of isolated Cu species [19]. The
Fig. 8 shows the dynamic effect of introducing 10% H2O
in the feed, on catalytic activity at 400 C. The decrease in
NO conversion was sensible but still not very high and by
sequential suppression of the water gas feed the effect was
found to be reversible.
low-temperature peak was assigned to the reduction of iso-
◦
2+ +
lated Cu to Cu ions. The peak at a higher temperature
+
0
was attributed to the reduction of Cu ion to Cu , as was
done for Cu-ZSM-5 [19]. Since the CuO aggregates would
0
◦
be reduced to Cu by H in one step at about 350 C [19],
2
◦
the peak at 348 C in Cu-Al-MCM-41-10-120 indicated the
3
.3. Temperature-programmed reduction
presence of CuO in this sample. It is suggested that copper in
underexchanged catalysts was mainly in the form of isolated
copper ions, while in overexchanged catalyst, there was a
proportion of CuO species. Delahay et al. [20] also reported
The TPR profiles of Cu-Al-MCM-41 catalysts with the
same Si/Al ratio of 10, but different copper contents (0.85

248
Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
◦
Fig. 8. Effect of water vapor on NO conversion over Cu-Al-MCM-41 at 400 C.
Fig. 9. H -TPR profiles of Cu-Al-MCM-41 samples with Si/Al = 10: (a) Cu-Al-MCM-41-10-31, (b) Cu-Al-MCM-41-10-61, (c) Cu-Al-MCM-41-10-93, and
2
(d) Cu-Al-MCM-41-10-120.
that copper was mainly present as isolated Cu2 species in
underexchanged Cu-Beta and Cu-MFI zeolites.
+
tios of the two reduction peaks, we observed that in catalyst
Cu-Al-MCM-41-3-9, copper was mainly present as Cu
2
+
A similar dependence was found in catalysts with differ-
ent copper contents (0.22 ꢀ Cu ꢀ 1.1 wt%) of Si/Al = 30
ion, since the areas of the two peaks were very close. In-
creasing the Si/Al ratio led to a decrease of the H2/Cu ratio
for the first peak and an increase for the second. This phe-
nomenon could be accounted for as the result of partial re-
(
see Fig. 10). Isolated copper ions were the main species
in catalysts of a copper-exchanged level lower than 100%,
while CuO occurred in the catalyst with 127% copper ex-
change.
Fig. 11 shows the TPR profiles of Cu-Al-MCM-41 cata-
lysts with the same copper contents of 0.85 wt% but different
Si/Al ratio from 3 to 100; the respective temperatures and
H2/Cu ratios are given in Table 3. It was also observed that
there existed CuO species in the overexchanged catalyst Cu-
Al-MCM-41-100-322, while only copper ions were detected
in the underexchanged catalysts. Comparing the H2/Cu ra-
duction of Cu2 to Cu during the course of catalyst prepa-
+
+
◦
◦
ration (calcinations at 540 C) or pretreatment at 500 C in
a flow of He [4]. It is indicated that lowering the Si/Al ratio
enhanced the reduction of isolated Cu2 ions.
+
The peaks maxima in Figs. 9–11 appeared at a lower tem-
perature as the catalyst copper-exchanged level and Si/Al
ratio increased. It is suggested that the Cu2 and Cu ions
were more easily reduced, which is likely due to a decrease
of the electrostatic interaction between copper ions and zeo-
+
+

Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
249
Fig. 10. H -TPR profiles of the catalyst samples with Si/Al = 30: (a) Cu-Al-MCM-41-30-25, (b) Cu-Al-MCM-41-30-60, (c) Cu-Al-MCM-41-30-97, and (d)
2
Cu-Al-MCM-41-30-127.
Fig. 11. H -TPR profiles of Cu-Al-MCM-41 with 0.85 wt% Cu and various Si/Al ratios: (a) Cu-Al-MCM-41-100-322, (b) Cu-Al-MCM-41-30-97, (c)
2
Cu-Al-MCM-41-10-31, and (d) Cu-Al-MCM-41-3-9.
2
+
Table 3
alysts [21]. The decrease in reduction temperature of Cu
Peak temperature and area in the TPR profiles underexchanged Cu-Al-
MCM-41 with 0.85 wt% Cu
+
and Cu upon increasing the Cu-exchange loading is likely
associated with a distribution of the strength of electrosta-
tic interactions, the stronger sites attracting first the Cu ions
at lower Cu loading. A broader distribution is certainly to
be expected in Al-MCM-41 than in zeolites. The similar ef-
fect of an increased Al content on reduction temperatures is
in line with the well-documented increase in lattice basicity
for zeolites with lower Si/Al ratios. It was also verified that
the CuO concentration depends on the zeolite Si/Al ratio and
copper content; however, its reduction temperature is almost
independent of these parameters. This may be due to CuO
species not strongly interacting with the zeolite framework.
Sample
Si/Al Temperature of peak H /Cu(TPR)
2
◦
maximum ( C)
(mol/mol)
First
Second
First
Second
Cu-Al-MCM-41-3-9
Cu-Al-MCM-41-10-31 10
Cu-Al-MCM-41-30-97 30
3
260
220
195
630
570
517
0.47
0.42
0.38
0.51
0.53
0.56
lite framework as the content of copper ions increased and Al
content decreased. A similar behavior with however less sig-
nificant temperature changes was observed for Cu-MFI cat-

250
Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
Fig. 12. TPD profiles of NO adsorbed on Al-MCM-41 and Cu-Al-MCM-41: (a) Al-MCM-41-10, (b) Cu-Al-MCM-41-10-31, (c) Cu-Al-MCM-41-10-61, (d)
Cu-Al-MCM-41-10-93, and (e) Cu-Al-MCM-41-10-120.
Table 4
Characteristics of NO-TPD of Cu-Al-MCM-41 with Si/Al ratio of 10
a
a
AL/AL + AHb
Sample
NO desorption
NO desorption
(µmol µmol Cu
−
1
−1
(µmol g
)
)
Cu-Al-MCM-41-10-31
Cu-Al-MCM-41-10-61
Cu-Al-MCM-41-10-93
Cu-Al-MCM-41-10-120
14.2
34.8
40.6
41.8
0.106
0.138
0.108
0.083
1
0.95
0.9
0.9
a
Based on the peak area at low temperature.
AL: peak area at low temperature; AH: peak area at high temperature.
b
3
.4. NO temperature-programmed desorption (TPD)
the TPR results, this feature indicates that copper mainly ex-
ists as isolated Cu2 in these catalysts.
+
The total amount of NO desorbed increased with the
copper-exchanged level up to 100% and then tended to a
plateau. This fact may be related to the concentration of cop-
per sites for NO adsorption no more increasing after the
exchange saturation. Nevertheless, Coq et al. [7] reported
a sharp decrease of the accessibility of copper sites to NO
upon formation of copper oxide in Cu-MOR catalysts due
to the blocking of the mordenite channels by CuO species.
Since the pore channel of Al-MCM-41 is much larger than in
mordenite, we believe that the small amount of CuO aggre-
gates would not clog the mesopore and, therefore, had little
effect on NO adsorption.
The TPD profiles of Cu-Al-MCM-41 catalysts with Si/Al
=
10 but different copper contents (0 ꢀ Cu ꢀ 3.2 wt%)
are depicted in Fig. 12. The corresponding desorption data
are shown in Table 4. The copper-free sample, Al-MCM-41-
1
0, desorbed trace amounts of NO in comparison with the
copper-containing samples. This suggests that copper was
the adsorption center for nitric oxide. With the exception of
the sample having 31% exchange, the NO desorption curves
had the same shape. Two NO adsorption peaks were detected
◦
at about 150 and 410 C. In agreement with the literature [7],
the low-temperature peak was ascribed to the desorption of
NO from Cu2 ions, while the second peak was associated
+
Fig. 13 shows the NO TPD profiles of Cu-Al-MCM-41
catalysts with 0.85 wt% of copper content but different Si/Al
ratios (3 ꢀ Si/Al ꢀ 100); the respective desorption data are
presented in Table 5. Only one NO desorption peak is ob-
−
−
with the decomposition of nitrate (NO ), nitrite (NO ), or
NO adsorbed species. At low exchange levels, the amount
3
2
+
2
of desorbed species at high temperature may be too small to
detect.
◦
served at around 190 C for samples with low Si/Al ratios.
Assuming that the sensitivity factor is the same for the
peaks at low and high temperatures, their areas could be used
to estimate a normalized desorption of NO and the effect of
copper content and Si/Al ratio on NO desorption. As listed in
Table 4, the ratio of the amount of desorbed NO at low tem-
perature to the total amount was close to 1. Combined with
On the contrary, the other samples with a higher Si/Al ra-
tio present two peaks: a principal peak at low temperatures
◦
(about 140 C) and a second NO desorption peak at higher
◦
temperatures (440 C), which are related to the desorption of
2
+
−
−
+
Cu –NO, and NO , NO or NO species, respectively. It
3
2
2
is clear that increasing the Si/Al ratio in the underexchanged

Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
251
Fig. 13. TPD profiles of NO adsorbed on Cu-Al-MCM-41 with similar copper contents but different Si/Al ratios: (a) Cu-Al-MCM-41-100-322, (b)
Cu-Al-MCM-41-30-97, (c) Cu-Al-MCM-41-10-31, and (d) Cu-Al-MCM-41-3-9.
Table 5
centers for both the SCR of NO and the oxidation of propy-
lene.
In the present study, we investigated the effect of copper
content and Si/Al ratio on the catalytic activity of Cu-Al-
MCM-41 catalysts for NO SCR by propylene. The results
demonstrated that not all copper sites present in Al-MCM-
Amount of NO desorbed from Cu-Al-MCM-41 with different Si/Al ratio
Sample
Cu
NO
AL/AL + AHb
a
)
exchanged desorption
−1
(%)
(µmol g
Cu-Al-MCM-41-100-322
Cu-Al-MCM-41-30-97
Cu-Al-MCM-41-10-31
Cu-Al-MCM-41-3-9
322
97
31
9
14.6
33.2
14.2
8.4
0.98
0.9
1
4
1 catalysts were equal in activity for NO SCR, since the
catalysts with the same Si/Al ratio but different copper con-
tents or the same copper content but different Si/Al ratio
exhibited different catalytic activities for NO and propylene
conversions. The catalyst with the highest amount of copper
was not the most active. Furthermore, our results indicated
that the activity for NO conversion into N2 was related to
the ability of the catalyst for propylene oxidation to carbon
dioxide. Kintaichi et al. [22] also observed that the NO selec-
tive reduction is closely related to the extent of hydrocarbon
oxidation.
TPR and NO TPD experiments showed that the nature
of copper and the relative quantity of the different cop-
per species exchanged into Al-MCM-41 depended on the
copper-exchanged level and Si/Al ratio of the catalysts.
The results revealed that copper was mainly in the form
of isolated copper ions in Cu-Al-MCM-41 catalysts at a
copper-exchanged level ꢀ 100%, while at higher copper
contents CuO species were also present. It can be concluded
that the isolated copper ions are more effective for adsorp-
tion and reduction of NO than CuO species, since indeed
the maximal NO conversion and adsorption amounts were
obtained at the copper-exchange level of about 100%. In un-
derexchanged catalysts, the activity of NO reduction and the
amount of NO adsorption increased as the copper contents
(i.e., isolated copper ions) increased. On the other hand,
the CuO species in overexchanged catalysts promoted the
oxidation of propylene at low temperatures and therefore,
inhibited the reduction of NO. Moreover, in addition to the
1
a
Based on the peak area at low temperatures.
AL: peak area at low temperature; AH: peak area at high temperature.
b
catalysts enhanced the NO adsorption, which indicated fur-
ther that the interaction of copper ions and Al atoms in the
framework was affected by the Si/Al ratio. The oxidic phase
may also cover part of the cationic sites in the overexchanged
catalysts.
Since the SCR of NO is going on at temperatures ex-
ceeding 250 C it is likely that the low-temperature adsorbed
◦
NO will not be stable under reaction conditions. This would
mean that one of the NOx species that contribute to the high-
temperature desorption peak is involved in the SCR surface
process. Thus a very small fraction of the Cu ions would
constitute the SCR active sites most of the other Cu atoms
being spectator species.
4
. Discussion
The characterization using XRD, NMR, and N2 isother-
mal adsorption revealed that the structure of Al-MCM-41
remained unchanged after the exchange of copper ion. Since
Al-MCM-41 was proved to be inactive for the reduction of
NO and low active for the oxidation of propylene, the cop-
per ions exchanged to Al-MCM-41 should provide the active

252
Y. Wan et al. / Journal of Catalysis 227 (2004) 242–252
variation in average distance between surface Al ions, the
Si/Al ratio had an effect on the electrostatic interaction be-
tween copper ions and Al-MCM-41 framework and, thus, on
the activity for NO reduction.
4. When the catalyst Si/Al ratio increases copper is more
easily reduced.
5. The isolated copper ions play an important role in the
SCR of NO by propylene in the presence of an excess
of oxygen, although only a minor fraction of these ions
constitutes the active SCR sites.
6. Preliminary results suggest that Cu-Al-MCM-41 cata-
lysts of NO-SCR are only slightly deactivated by water
vapor and this deactivation is reversible.
All these results suggest that the copper ions in Cu-Al-
MCM-41 behave as NO-SCR catalysts in a manner quite
similar to isolated copper ions in Cu-ZSM-5. This is some-
what surprising as the Brønsted acid sites in H-ZSM-5 are
much stronger acids than the ones in Al-MCM-41. The lat-
ter are indeed low acid strength amorphous materials. The
Cu ions are therefore expected to be lower strength Lewis
acids. Thus the Lewis acidity of the copper ions does not
seem to be the dominant factor that determines the SCR ac-
tive sites.
It seems also that the oxidic copper (CuO) forms only
once the lattice acid sites are fully copper exchanged. This
is indeed associated with the large specific surface area and
large pore diameter of MCM-41 which allow an easy access
of the copper ions during ion exchange.
Acknowledgment
This work was funded by the China National Education
Committee (No. 93039).
References
The low and reversible deactivation by water vapor dur-
ing NO-SCR is also significant for potential commercial ap-
plication of these new catalysts. The results shown in Figs. 7
and 8 are indeed preliminary and more work should be per-
formed before a valuable discussion of these effects could be
made. It is however already clear that the surface chemistry
of the support, its hydrophilicity, will require investigations
before any definitive conclusion can be made.
[
1] B.K. Cho, J. Catal. 142 (1995) 418.
[
2] M. Iwamoto, in: Proceedings of the Meeting on Catalytic Technology
for Removal of Nitrogen Monoxides, Tokyo, 1990, p. 17.
3] M. Iwamoto, N. Mizuno, J. Automob. Eng. 207 (1993) 23.
4] R.Q. Long, R.T. Yang, Ind. Eng. Chem. Res. 38 (1999) 878.
[
[
[5] A.P. Walher, Catal. Today 26 (1995) 107.
[6] C. Torre-Abreu, C. Henriques, F.R. Ribero, G. Delahay, M.F. Ribero,
Catal. Today 54 (1999) 407.
[7] B. Coq, D. Tachnon, F. Figueras, G. Mabilon, M. Prigent, Appl. Catal.
B 6 (1995) 271.
[
8] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D.
Schmitt, C.T.-W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen,
J.B. Higgins, J.L. Schlenker, J. Am. Chem. Soc. 114 (1992) 10834.
5
. Conclusions
The main conclusions obtained from XRD, N2 adsorption
isotherm, NMR, TPR, NO TPD, and the analysis of reactiv-
ity results can be summarized as follows:
[9] D. Trong On, D. Desplantier-Giscard, C. Danumah, S. Kaliaguine,
Appl. Catal. A 253 (2003) 545.
10] D.J. Parrillo, D. Dolenec, R.J. Gorte, R.W. McCabe, J. Catal. 142
[
(1993) 708.
[
11] D. Liu, H.J. Robota, Appl. Catal. B 4 (1994) 155.
1
2
. The uniformly arranged mesoporous framework of Al-
MCM-41 could still be well maintained after copper
exchange.
[12] Z. Schay, V. Samuel Jaems, G. Pal-Borbely, A. Beck, A.V. Ra-
maswamy, L. Guczi, J. Mol. Catal. A 162 (2000) 191.
[13] E.S. Shpiro, W. Grunert, R.W. Joyner, G.N. Baeva, Catal. Lett. 24
1994) 159.
14] Y. Wan, Z. Wang, J.X. Ma, X.M. Jin, W. Zhou, Acta Chim. Sin. 60,
2002) 71.
(
. Cu-Al-MCM-41 are active catalysts for NO selec-
tive catalytic reduction by propylene in an oxidiz-
ing atmosphere. Their activity depends on the copper-
exchanged level and Si/Al ratio. At constant Si/Al ratio,
this activity, as indicated by NO conversion to N2, in-
creases with increasing Cu ion-exchange level up to
about 100% and then decreases. At constant copper
content, the maximal NO conversion decreases with de-
creasing Si/Al ratio in the underexchanged catalysts.
[
[
(
15] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.C. Beck, Na-
ture 359 (1992) 710.
[16] G. Engelhardt, Stud. Surf. Sci. Catal. 58 (1991) 285.
[17] K.C.C. Kharas, Appl. Catal. B 2 (1993) 207.
[18] S. Sato, Y. Yu-u, H. Yahiro, N. Mizuno, M. Iwamoto, Appl. Catal. 70
(1997) L1.
[19] C. Lee, B. Ha, Stud. Surf. Sci. Catal. 84 (1994) 1563.
[20] G. Delahay, B. Coq, L. Broussous, Appl. Catal. B 12 (1997) 49.
[21] C. Torre-Abreu, M.F. Ribero, C. Henriques, F.R. Ribero, Appl. Catal.
B 11 (1997) 383.
. Copper is mainly in the form of isolated Cu2 ions in
+
3
Cu-Al-MCM-41 catalysts with copper-exchangelevel ꢀ
1
00%, whereas at higher copper-exchange CuO species
[22] Y. Kintaichi, H. Hamada, M. Tabata, M. Sasaki, T. Ito, Catal. Lett. 6
(1990) 239.
are also present.