Location and nature of Cu species in Cu/SAPO-34 for selective catalytic reduction of NO with NH3

Article abstract of DOI:10.1016/j.jcat.2012.01.012

Cu/SAPO-34 catalysts have been reported to show much improved activities and durability for selective catalytic reduction (SCR) of NO_x. In this work, we examined the nature of Cu species in SAPO-34 catalysts in order to shed light on the active

Full text of DOI:10.1016/j.jcat.2012.01.012

Journal of Catalysis 289 (2012) 21–29
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Location and nature of Cu species in Cu/SAPO-34 for selective catalytic reduction
of NO with NH3
a
b,
⇑
b
a,c,
⇑
Lei Wang , Wei Li , Gongshin Qi , Duan Weng
a
Key Laboratory for Advanced Materials of Ministry of Education, Department of Materials Science and Engineering, Tsinghua University, Beijing 100084, China
Chemical Sciences & Materials Systems Laboratory, General Motors R&D Center, Warren, MI, USA
State Key Laboratory of New Ceramics and Fine Processing, Department of Materials Science and Engineering, Tsinghua University, Beijing 100084, China
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Cu/SAPO-34 catalysts have been reported to show much improved activities and durability for selective
Received 31 October 2011
Revised 5 January 2012
Accepted 19 January 2012
Available online 2 March 2012
x
catalytic reduction (SCR) of NO . In this work, we examined the nature of Cu species in SAPO-34 catalysts
in order to shed light on the active site requirements. Cu/SAPO-34 catalysts, prepared by ion-exchange
and precipitation methods, were characterized in detail. The results consistently indicate that Cu species
exist predominantly as isolated ions at the exchange sites in the ion-exchanged sample, whereas in the
precipitated sample CuO on the external surface is the dominant species. The ion-exchanged Cu/SAPO-34
Keywords:
NH -SCR
3
Copper
SAPO-34
Active sites
Cu species
was found to show superior NH
3
-SCR activity than the precipitated sample, suggesting that isolated Cu
ions at the exchange sites are likely the active sites.
Ó 2012 Elsevier Inc. All rights reserved.
1
. Introduction
SAPO-34, with a small-pore chabazite structure, exhibits high ther-
mal stability even in the presence of water [12]. Cu ion-exchanged
SAPO-n has been investigated for the selective reduction of NO by
x
Selective catalytic reduction of NO with hydrocarbons (HC-
SCR) or with ammonia (NH
extensively studied for lean NO
first reported on copper ion-exchanged zeolites by Iwamoto et al.
1,2]. NH -SCR technology has been commonly used for NO con-
trol in stationary applications [3] and more recently for diesel vehi-
cle emissions [4–10].
3
-SCR) as the reducing agent has been
C
x
H
y
[12], N
2 3
O decomposition [19], and recently for NH SCR [26].
x
control. HC-SCR technology was
The SCR catalytic performance of Cu-modified zeolites has been
found to depend on the microstructure and the content of copper
[2]. The SCR activity of exchanged Cu/ZSM-5 increased with
increasing Cu exchange level and reached a maximum at the ex-
change level of 100% [27,28]. Several types of copper species, such
[
3
x
2
+
+
Copper-based zeolites have been investigated as the catalysts
2
as isolated Cu and Cu ions [29,30], CuO or Cu O clusters [31,32]
2
+
for direct decomposition of NO [11], HC-SCR [12–14], NH
3
-SCR
and dimeric copper species ([Cu–O–Cu] ) [33,34], have been pro-
posed as the active sites over Cu/ZSM-5 catalysts. The most active
species in Cu-Beta zeolites were found to be copper ions rapidly
[
15–18], decomposition of N O [19,20], and N O reduction [21].
2
2
Cu ion-exchanged ZSM-5 zeolites exhibit high NO decomposition
rates and SCR activities, but ZSM-5 zeolite possesses poor hydro-
thermal stability due to dealumination. Consequently, other types
of zeolites have been explored to improve the activity and durabil-
ity of the SCR catalysts. Cu-exchanged beta zeolite [22,23] has been
shown to have excellent activity with better hydrothermal stability
than ZSM-5 catalysts. Recently, it was reported that Cu ion-ex-
2
+
+
transitioning between Cu and Cu species under SCR reaction
conditions [35]. However, the active site requirements for Cu ion-
exchanged small-pore zeolites (e.g., Cu/SAPO-34), especially the
location and state of the active Cu species, are not well understood.
The state of copper species could be strongly affected by the
preparation method. With the coexistence of multiple copper spe-
cies, it is difficult to distinguish the contribution of various copper
species to the SCR activity separately. In addition, there has been
no systematic study of the SCR activities of copper species at differ-
ent locations over Cu/SAPO-34. In this work, Cu-modified SAPO-34
catalysts were prepared via liquid ion-exchange and precipitation
methods, and the goal was to prepare samples with predominately
x
changed SSZ-n [24,25] showed higher NO conversions over a
broad temperature range with higher N selectivities and hydro-
2
thermal stability compared with Cu/ZSM-5 and Cu/beta. Silico-alu-
mino-phosphates (SAPO-n) possess ion-exchange properties as a
result of the isomorphous substitution of P in AlPO by Si, and
4
⇑
Corresponding authors.
E-mail addresses: wei.1.li@gm.com (W. Li), duanweng@tsinghua.edu.cn (D.
Weng).
single type of Cu species. Various techniques (XRD, H
and DRIFTS) were used to probe the location and state of the Cu
2
-TPR, STEM,
0
021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2012.01.012

2
2
L. Wang et al. / Journal of Catalysis 289 (2012) 21–29
species in these samples. The impact of different Cu species on NH
SCR activity was evaluated to shed insight on the active sites in
these catalysts.
3
½NH
3
ꢁ
inlet ꢀ ½NH
3
outlet
ꢁ
NH
3
Conversion ¼
ꢂ 100%
ð2Þ
½NH ꢁ
3
inlet
½
NH
3
ꢁinlet ꢀ ½NOꢁ
ꢀ ½NO
2
ꢁ
ꢀ 2½N
2
Oꢁoutlet ꢀ ½NH ꢁoutlet
3
outlet
outlet
N
2
Selectivity ¼
½
NH
3
ꢁ
ꢀ ½NH ꢁ
3
outlet
2
. Experimental
inlet
ꢂ 100%
2.1. Catalysts preparation
ð3Þ
The ion-exchanged Cu/SAPO-34 catalyst was prepared by a
2
.3. Characterizations of Cu-modified SAPO-34 catalysts
two-step liquid ion-exchange method. A commercial H/SAPO-34
powder (Novel, Al/Si/P = 1:0.1:0.9, measured by inductively cou-
pled plasma and atomic emission spectrometry) was ion ex-
The copper contents in the samples were determined by induc-
tively coupled plasma and atomic emission spectrometry (ICP-
AES), as shown in Table 1. The Cu loadings in the two samples were
chosen to be similar to probe the effect of Cu species.
The phase compositions and structures of the catalysts were
determined by X-ray diffraction (D8 advance diffractometer) using
changed using a NH
h to obtain the NH4 form. Then, the solid was filtered and washed
with distilled water. The NH /SAPO-34 was dried at 100 °C for 16 h
4
NO
3
(Alfa Aesar, >95%) solution at 80 °C for
þ
1
þ
4
before repeating the ammonium exchange process for a total of
two exchanges. Cu ion-exchange was performed by mixing the
þ
CuKa radiation (k = 1.5418 Å). The X-ray powder diffractograms
NH /SAPO-34 with a Cu(CH
3
COO)
2
(Riedel de Haën, >99%) solution
4
were recorded at 0.02° intervals in the 2h range from 10° to 40°.
The transmission electron microscopy (TEM) images were mea-
sured on a Cs corrected JEOL 2100F TEM/STEM operated at 200 kV.
(
0.05 mol/L) at ambient temperature for 6 h. A low-concentration
copper solution was used for liquid ion exchange to avoid copper
aggregating on the surface or in the pores. After the powder was
filtered and thoroughly washed with distilled water, it was dried
at 100 °C for 16 h and calcined at 550 °C for 4 h. This sample will
be hereafter denoted as ‘‘Cu/SAPO-34-E,’’ where ‘‘E’’ stands for
the exchange method.
Temperature-programmed reduction by H
2
2
(H -TPR) and tem-
perature-programmed desorption of NH (NH
3
3
-TPD) were carried
out on a commercial instrument (Micromeritics, Autochem 2920)
with a TCD detector. Before the TPR/TPD measurements, the sam-
ples were pre-oxidized at 500 °C followed by helium flow to purge
The CuO/SAPO-34 catalyst was prepared by precipitation of
the residual gaseous and weakly adsorbed oxygen. The H
files were measured in a flow of 10 vol.% H /Ar (30 ml min ) from
room temperature to 800 °C at a rate of 10 °C min . For NH
measurements, the catalysts were exposed to NH at 30 °C for
0 min, followed with a temperature ramp at a rate of 5 °C min
Diffuse reflectance infrared Fourier transform spectra (DRIFTS)
2
-TPR pro-
CuC
Haën, >99%) as the precursor and H
cipitator. A Cu(CH COO) solution was added dropwise into a slur-
ry of H aqueous solution and H/SAPO-34 powder to minimize
2
O
4
3 2
on H/SAPO-34 powder with Cu(CH COO) (Riedel de
ꢀ1
2
2 2 4
C O
(Fisher, >99%) as the pre-
ꢀ1
3
-TPD
3
2
3
2 2 4
C O
ꢀ1
6
.
liquid ion exchange of Cu. The precipitate was filtered, washed and
then dried at 100 °C overnight and calcined at 550 °C for 4 h. This
sample will be hereafter denoted as ‘‘Cu/SAPO-34-P,’’ where ‘‘P’’
stands for the precipitation method.
were measured on a FT-IR spectrometer (Thermo Nicolet Mag-
na670) using a heatable environmental reaction cell with ZnSe
windows, which was connected to a gas-dosing system. The oxida-
tion pretreatments at 500 °C for 1 h were executed to remove the
adsorbed water and clean the surface before each measurement.
The DRIFTS spectra were recorded in the range of 4000–
2
3 3
.2. NH -SCR catalytic activity and NH oxidation ability
The catalytic activity evaluation was carried out using a flow-
through powder reactor system equipped with a Fourier Transform
Infrared (FT-IR) spectrometer (Nicolet 670) with a gas-sampling
ꢀ1
ꢀ1
6
50 cm with a resolution of 4 cm . Nicolet OMNIC™ software
(
Ver. 7.3) was used to convert the raw reflectance data (absorbance
units) to Kubelka–Munk units. The ex situ spectra of the catalysts
were recorded at 30°C and the spectrum of KBr under the same
condition was used as the background. The in situ DRIFTS of chemi-
sorption were performed to characterize the surface adsorbed spe-
cies, and the background spectra were collected before the samples
were exposed to the absorbates. CO chemisorption was performed
in 1 vol.% CO/He (100 ml min ) at room temperature. The spectra
were recorded over the CO-saturated samples. A helium flow was
introduced to purge the extra gas-phase CO molecules before the
sampling. NH
in 200 ppm NH
every 5 min during the NH
with NH , the flow was switched to helium, and the samples were
heated up to 500 °C at a rate of 10 °C min in the He flow. The
spectra of the temperature-programmed desorption were collected
every 50 °C.
cell. The gas mixture consisted of 200 ppm NO, 200 ppm NH
balanced with N for the NH -SCR reaction and 200 ppm NH
% O balanced with N for the NH oxidation reaction. 10% H
and 10% CO were present all the time. The gas hourly space veloc-
ity (GHSV) was 300,000 h for the standard SCR and NH
3
, 8%
O
8
2
2
3
3
,
2
2
3
2
O
2
ꢀ1
3
oxida-
ꢀ1
tion measurements and 1,200,000 h for the rate measurements
and activation energy evaluation. Prior to the activity measure-
ments, the catalysts were pretreated at 550 °C for 30 min in 8%
ꢀ1
O
2
/N
varying from 180 to 550 °C. The NO and NH
culated based on the inlet and outlet gas concentrations (Eqs. (1)
2
flow. The catalytic activities were measured at temperatures
3
chemisorption was operated at room temperature
3
conversions were cal-
ꢀ1
3
/He (100 ml min ). The spectra were recorded
adsorption process. After saturated
3
and (2)) at steady state. The selectivity to N
2
in the NH
3
oxidation
3
was estimated based on nitrogen balance (Eq. (3)).
ꢀ1
½
NOꢁ
inlet
ꢀ ½NOꢁoutlet
NO Conversion ¼
ꢂ 100%
ð1Þ
½
NOꢁinlet
3
. Results
Table 1
Chemical composition of the catalysts by ICP analysis.
3.1. NH -SCR catalytic activity and NH oxidation activity
3
3
Sample
Concentration (wt.%)
NO and NH
catalysts at reaction temperatures between 180 and 550 °C
Fig. 1). Due to the high activities of these catalysts, high space
velocity conditions were used in order to differentiate the activities
of the catalyst samples. The H/SAPO-34 itself showed very low
3
conversions were measured on the Cu/SAPO-34
Al
P
Si
Cu
H/SAPO-34
Cu/SAPO-34-E
Cu/SAPO-34-P
18
19
17
19
20
19
2.1
2.3
2.1
0
0.8
0.8
(

L. Wang et al. / Journal of Catalysis 289 (2012) 21–29
23
Fig. 1. NO (left) and NH
3
(right) conversions as a function of reaction temperature; reaction conditions: 200 ppm NO, 200 ppm NH
3
, 8% O
2
, 10% H
2
O, 10% CO
2
balanced with
ꢀ
1
ꢀ1
N
2
; flow rate: 1 L min ; GHSV: 300,000 h
.
Fig. 2. NH
3
oxidation conversions and N
2
selectivity as a function of reaction
, 8% O , 10% H O, 10% CO balanced
temperature; reaction conditions: 200 ppm NH
with N ; flow rate: 1 L min ; GHSV: 300,000 h .
2
3
ꢀ1
2
2
2
ꢀ
1
Fig. 3. XRD patterns of H/SAPO-34 (a), Cu/SAPO-34-E (b), Cu/SAPO-34-P (c).
activities in this temperature range; hence, the activities of the Cu-
containing samples could be attributed to contribution of the cop-
per species. The exchanged sample Cu/SAPO-34-E was much more
active than the precipitated catalyst Cu/SAPO-34-P, even though
3.2. Structural properties
The XRD patterns of the Cu-containing samples show a chabaz-
ite phase (Fig. 3, labeled by solid circle symbols), which is consis-
tent with the pattern of H/SAPO-34, suggesting that neither the
ion-exchange nor the precipitation method significantly affects
the SAPO-34 structure. No diffraction peak of copper species (Cu,
the Cu loadings in these two samples were similar. NO
detected in H/SAPO-34 at low temperature. For the Cu-containing
catalysts, there was no NO formed during the NH -SCR reaction
in the whole temperature range, suggesting that either the cata-
lysts have a low NO oxidation ability or the formed NO was
quickly reacting with NH to form N
Fig. 2 shows the NH oxidation performance of the catalysts as a
function of temperature during the steady state reaction in the ab-
sence of NO. H/SAPO-34 not only exhibits low NH oxidation abil-
ity, but also poor N selectivity. However, after copper ion
exchange, the NH conversion could reach 40% at 550 °C with a
selectivity of 90% (Fig. 2, Cu/SAPO-34-E). Consistent with the
performance in SCR reaction, the precipitated sample Cu/SAPO-
4-P shows a stronger NH oxidation ability, but a lower N selec-
2
was only
2
3
2
Cu O, CuO) was observed for the ion-exchanged sample Cu/
2
SAPO-34-E (Fig. 3b), suggesting that copper is either well dispersed
or exists as isolated ions at the exchange sites. For Cu/SAPO-34-P
(Fig. 3c), two diffraction peaks located at 35.29° and 38.49° (la-
beled by open diamond symbols) were observed, and these peaks
can be related to a CuO phase. The CuO crystallite, estimated based
on the Sherrer’s equation, was found to have an average size of
5.0 nm, which is much too large to be inside the pores of SAPO-
34 (pore diameter = 0.38 nm) [27]. Therefore, it can be inferred
that CuO particles locate on the external surface of SAPO-34 in
the precipitated sample.
3
2
.
3
3
2
3
N
2
3
3
2
tivity compared with Cu/SAPO-34-E. Considering that different Cu
species may be present in these two samples due to the different
preparation methods, it seems that the location and the status of
Cu species play a critical role in determining the catalyst activity
and selectivity.
Although the introduction of Cu ions does not significantly
influence the crystal structure of the SAPO-34, different Cu species
may affect the positions of the diffraction peaks. In comparison
with H/SAPO-34, the peaks for the Cu ion-exchanged sample
shifted to lower diffraction angles, suggesting a lattice expansion

2
4
L. Wang et al. / Journal of Catalysis 289 (2012) 21–29
Fig. 4. STEM images of Cu/SAPO-34-P.
of SAPO-34. This could be attributed to the incorporation of Cu ions
inside the SAPO-34 pores. On the other hand, the diffraction peaks
in the precipitated sample (Cu/SAPO-34-P) did not shift positions,
consistent with the Cu species existing as CuO on the external sur-
face of SAPO-34.
2 2
Fig. 5. H -TPR profiles of all the catalysts; conditions: 10 vol.% H /Ar; ramp rate:
ꢀ1
10 °C min .
Table 2
H
2
2
consumption and composition of copper ions in H -TPR.
a
Catalyst
Cu loading (wt.%)
H
2
consumption (ml/g)
2
H /Cu
3
.3. Morphology of Cu/SAPO-34 catalysts
Cu/SAPO-34-E
Cu/SAPO-34-P
0.8
0.8
0.67
2.79
0.24
0.98
STEM images of Cu/SAPO-34 catalysts were taken (Fig. 4) to
a
examine the morphology of Cu species and the support (SAPO-
Obtained by ICP.
3
3
4). SAPO-34 showed the typical cubic crystals in the size of 1–
m. The isolated copper ions in the exchanged sample (Cu/
l
SAPO-34-E, not shown here) cannot be detected by the STEM tech-
nique. However, CuO particles (bright area) were indeed found to
aggregate on the external surface of SAPO-34 over the precipitated
catalyst (Cu/SAPO-34-P, Fig. 4), and the crystallite size of CuO ob-
served in the image is in the range of 10–30 nm, which is higher
than the averaged crystallite size value calculated by XRD. This
suggests that either the observed CuO particles are agglomerates
of smaller crystallites or there are probably smaller CuO particles
that are invisible in the STEM images.
3.4. Reducibility of copper species
Temperature-programmed reduction (TPR) measurements
were carried out using H
reducibility of Cu species. No reduction peak was observed for H/
SAPO-34 (Fig. 5), indicating that the H consumption peaks for
2
for the catalyst samples to probe the
2
the Cu-containing samples are attributed to the reduction of cop-
per species. The precipitated sample (Cu/SAPO-34-P) exhibited a
sharp peak centered at 200 °C, attributed to the reduction of CuO
x
particles [36]. Contrary to that, a broad reduction peak in the tem-
perature range of 150–350 °C was obtained in the ion-exchanged
samples (Cu/SAPO-34-E). The reducibility of copper species at the
Fig. 6. NH
temperature ramp rate: 5 °C min
3
-TPD profiles of all the catalysts; ammonia adsorbed at 30 °C for 60 min;
1
ꢀ
.
ion-exchange sites is still under debate. According to the literature
+
[
2
36], the isolated Cu2 ions could be reduced in the temperature of
3
3.5. NH -TPD
+
00–300 °C, but the Cu ions at the ion-exchange sites could be re-
0
duced to Cu at distinctly higher temperatures or too stabile to be
reduced in the H -TPR temperature range.
The amount of H consumption during TPR provides quantita-
tive information on the number of reducible Cu ions in the samples
Table 2). The H /Cu ratio in the precipitated sample is close to 1,
Temperature-programmed desorption of ammonia was carried
out to determine the strength and amount of different acid sites.
The normalized NH
2
2
3
-TPD profiles are displayed in Fig. 6. All the
desorption curves with three peaks,
catalysts show similar NH
3
(
2
corresponding to the acid sites with different acid strengths. The
low-temperature (LT) desorption peak is centered at 80 °C with a
shoulder peak at around 155 °C, which can be attributed to weakly
2+
suggesting that the Cu ions mainly exist as Cu , which again con-
firms that CuO is the dominant species. As for the ion-exchanged
sample, the amount of H
2
consumption is much less than the the-
3
adsorbed NH or ammonium species adsorbed at weak Lewis acid
oretical value, even if we assume that all the copper ions exist as
sites [37]. The high-temperature (HT) desorption peak in the tem-
perature range of 250–450 °C could be assigned to ammonia ad-
sorbed at strong Brønsted acid sites [37,38]. H/SAPO-34 shows
+
Cu . This indicates that some copper species (e.g., Cu ions at the
0
ion-exchange sites) are too stable to be reduced to Cu at temper-
atures below 800 °C.
3
the most intense NH desorption peaks at high temperature as well

L. Wang et al. / Journal of Catalysis 289 (2012) 21–29
25
is related to Cu2+ (oxygen bridged) ions and the band at the higher
+
wavenumber is related to Cu (isolated) ions [39,40]. The presence
of these two bands in the exchanged samples provides direct evi-
dence for the Cu species occupying the exchange sites inside the
SAPO-34 pores. These two bands were not detected for Cu/SAPO-
3
4-P, which suggests that there were few copper ions at the ex-
change sites in the precipitated sample. The peak assignments
and references for the DRIFT spectra are summarized in Table 3.
3.6.2. CO chemisorption
Diffuse reflectance IR spectra of CO adsorption were measured
to probe the oxidation state and coordination of Cu ions in the cat-
+
+
alysts. CO can absorb strongly on Cu sites to form stable Cu –CO
2
+
species [41], while it does not interact with Cu cations at ambient
temperature. The H/SAPO-34 zeolite does not show any CO adsorp-
ꢀ1
tion in the range of 2200–2100 cm (Fig. 8). A band at around
ꢀ1
+
2
154 cm , characteristic of Cu –CO carbonyls species on isolated
Fig. 7. Ex situ DRIFT spectra of H/SAPO-34, Cu/SAPO-34-P and Cu/SAPO-34-E; units
in absorbance.
+
Cu sites, was observed on the exchanged catalysts (Cu/SAPO-34-
E), consistent with isolated Cu cations occupying the exchange
sites. For the precipitated sample (Cu/SAPO-34-P), in addition to
the main band at 2154 cm , a shoulder at around 2134 cm
was observed. This shoulder band has been assigned to CO ad-
ꢀ1
ꢀ1
as the precipitated catalyst (Cu/SAPO-34-P), suggesting the largest
number of strong Brønsted acid sites. However, the ion-exchanged
catalyst (Cu/SAPO-34-E) exhibits a stronger LT desorption peak and
a slightly weaker HT desorption peak compared with H/SAPO-34.
The difference in LT desorption peak could be attributed to the
Cu species in the ion-exchanged sample, which will be further con-
firmed in the following section.
+
+
+
sorbed on dimeric Cu sites (Cu –CO–Cu species) [41]. The pres-
ence of this lower frequency band may suggest the existence of
x
CuO aggregates, consistent with the XRD and STEM results.
3
.6.3. NH
Ammonia has been extensively used as a probe molecule to
identify the different types of surface acidic sites. Fig. 9 illustrates
the DRIFT spectra of all the samples after NH adsorption at room
temperature for 5 min and 30 min. Adsorbed species can be
3
chemisorption
3
3
.6. Diffuse reflectance infrared spectra
3
ꢀ1
.6.1. H/SAPO-34 and Cu/SAPO-34
roughly divided into three regions: 3800–3500 cm (O–H stretch-
ing mode region), 3500–1800 cm (N–H stretching mode region)
and 1800–1200 cm (N–H deformation region) [37,38].
Generally, the similarity of NH adsorption behaviors of H/
SAPO-34 and Cu/SAPO-34-P implies that the CuO species on the
external surface of SAPO-34 has little impact on the acidic prop-
ꢀ1
The ex situ DRIFT spectra of the Cu/SAPO-34 catalysts were mea-
ꢀ1
sured to probe the effect of Cu on the SAPO-34 framework vibra-
tions (Fig. 7). The spectra in absorbance units were acquired after
pretreating the samples at 500 °C to remove the adsorbed water
3
x
ꢀ1
and clean the surface. Peaks in the range of 2000–650 cm are
associated with the internal and external stretching vibrations of
erty. At the beginning of NH
3
adsorption (Fig. 9, gray lines), they
ꢀ1
ꢀ1
the SAPO-34 framework. The peaks in the 3500–3800 cm region
are related to the OH group stretching vibration modes ( OH). The
both show strong band at around 1465 cm assigned to asymmet-
þ
þ
t
ric vibration of NH4 adsorbed on the Brønsted acid sites (
r
as(NH )),
4
ꢀ1
ꢀ1
þ
weaker bands at 3745 and 3678 cm are assigned to P–OH and Si–
as well as the bands at 3377, 3335 and 3294 cm related to NH4
OH species located on the external surface of the sample particles,
[37]. Meanwhile, two distinct negative bands at 3600 and
ꢀ1
ꢀ1
respectively. Two stronger bands at 3627 and 3600 cm are as-
signed as the stretching mode of bridged OH groups Al–(OH)–Si,
which are related to the Brønsted OH groups [37,38].
3627 cm , assigned to OH vibrations at Brønsted acid sites, are
þ
progressively intensified and it is thus indicative of NH4 adsorbed
at these Brønsted acid sites over these two catalysts. With the
increasing extent of ammonia chemisorption, ammonia is eventu-
ꢀ1
Two new bands at 891 and 844 cm only appeared in the ex-
changed samples, which have been associated with an internal
asymmetric framework vibration perturbed by copper cations
ꢀ
1
ally adsorbed molecularly. Band at 3338 cm assigned as NH
3
ꢀ1
molecule adsorption increases with band at 3293 cm intensified,
þ
[
39,40]. It was proposed that the band at the lower wavenumber
then superimposed on bands of NH4 groups.
Table 3
Peak positions and assignments in the DRIFT spectra.
Spectra
Wavenumber (cmꢀ1)
Species
Reference
ex situ
3745
Si–OH on the external surface
P–OH on the external surface
Bridged Al–(OH)–Si Brønsted groups
Internal asymmetric framework vibration perturbed by copper cations
[37,38]
[37,38]
[37,38]
[39,40]
3
3
8
678
627, 3600
91, 844
+
CO chemisorption
2154
134
Cu –CO carbonyls
[41]
[41]
+
+
2
Cu –CO–Cu species
þ
NH
3
chemisorption
3377, 3335, 3294
NH species
[37,38]
[37,38]
[42]
[37,38,42]
[37,38,42]
4
3
3
1
1
338, 3293 (3344, 3286)
179
617
465–1496
NH
NH
3
3
molecules
+
–Cu species
r
as(NH
3
þ
) on Lewis acid sites
ras(NH ) on Brønsted acid sites
4

2
6
L. Wang et al. / Journal of Catalysis 289 (2012) 21–29
strengths. Consistent with the NH
evolution during the NH desorption process for Cu/SAPO-34-P
Fig. 10c) is much similar with that for H/SAPO-34 (Fig. 10a). The
3
adsorption results, the spectra
3
(
ꢀ
1
ꢀ1
doublet bands at 3338 and 3293 cm (3344 and 3286 cm for
Cu/SAPO-34-E) firstly decrease at lower temperatures (below
2
ꢀ1
00 °C), and meanwhile, the negative band at 3678 cm assigned
to the external hydroxyl groups (i.e., P–OH, Si–OH, and Al–OH) re-
stores. This process is corresponding to the release of molecularly
adsorbed NH
of the NH -TPD profiles (Fig. 6). For Cu/SAPO-34-E, an extra quan-
tity of NH comes from the contribution of copper species. As
3
, which is the main product in the LT desorption peak
3
3
ꢀ1
shown in Fig. 10b, the bands at 1617 and 3179 cm , assigned to
ammonia adsorbed on copper related Lewis sites, also decrease
in this temperature range, mainly due to the release of NH
3
ad-
sorbed on copper species. At the same time, the negative bands
ꢀ1
at 891 and 844 cm gradually restore, which reveals the available
copper ions sites resumed. With further increasing the tempera-
tures, the intensities of the spectra continue to decrease. The pro-
Fig. 8. DRIFT spectra of CO adsorption over H/SAPO-34, Cu/SAPO-34-P and Cu/
SAPO-34-E; units in Kubelka–Munk; conditions: 1% CO/He, adsorption of CO for
ꢀ1
1
5 min and purging with He at 30 °C.
gressive disappearance of the complex bands in 1469–1496 cm ,
ꢀ1
and 3377–3294 cm , indicates that adsorbed on the Brønsted acid
sites are decomposed. Simultaneously, the bridged OH groups gen-
erally come out with the temperature increasing, as testified by the
ꢀ
1
growth of the bands at 3627 and 3600 cm . This again confirms
that the HT desorption peak in NH -TPD process (Fig. 6) is assigned
to ammonia adsorbed at strong Brønsted acid sites.
3
4
. Discussion
4.1. The structure of copper species
The Cu species in Cu/SAPO-34 prepared by precipitation and li-
quid ion-exchange methods appear to exist in different forms. In
Cu/SAPO-34-P, CuO on the external surface is the dominant spe-
x
cies. The XRD patterns (Fig. 3) prove that Cu ions in the precipi-
tated sample (Cu/SAPO-34-P) exist as copper oxide phase and
STEM images (Fig. 4) illustrate that these CuO particles locate on
the external surface of SAPO-34 as expected. From the H
file, the sharp reduction peak at low temperatures is due to the
reduction of CuO as reported in the literature [36].
2
-TPR pro-
x
In the ion-exchanged sample Cu/SAPO-34-E, Cu species exist
predominantly as isolated Cu ions at the exchange sites. No feature
peaks of copper species were detected by XRD, and no copper
aggregation was observed over the entire zeolite in the STEM
micrographs, which means the copper ions uniformly distribute
over the SAPO-34 zeolite. However, the absence of Cu species
phase in the XRD pattern of Cu/SAPO-34-E sample could not confi-
dently prove the copper ions locate at the cation exchange sites.
3
Fig. 9. DRIFT spectra of NH adsorption at room temperature for 5 min (gray lines)
and 30 min (black line); units in Kubelka–Munk; a:H/SAPO-34; b:Cu/SAPO-34-E; c:
Cu/SAPO-34-P.
3
NH chemisorption spectra of Cu/SAPO-34-E are different from
those of H/SAPO-34 and Cu/SAPO-34-P, which exhibit more intense
ꢀ1
bands at 1617 cm , assigned to asymmetric vibration of NH
3
ad-
sorbed on Lewis acid sites ( as(NH )) at the first stage of adsorption,
r
3
2
But the broad reduction peak in the H -TPR profile suggests that
ꢀ
1
but less intense bands in the range 1465–1496 cm assigned to
the copper species in Cu/SAPO-34-E are much more difficult to re-
duce than the copper oxides. In addition, the appearance of two
þ
NH adsorbed on the Brønsted acid sites. A new band at
4
ꢀ1
+
ꢀ1
3
179 cm corresponding to NH
3
–Cu species on the exchanged
bands at 891 and 844 cm in the ex situ IR spectra of the ex-
sites was observed [42]. Concurrently, two obvious negative bands
changed samples (Fig. 7), which have been associated with an
internal asymmetric framework vibration perturbed by copper cat-
ions, directly proves the occupancy of the ion-exchange sites by
copper ions inside the framework. On the contrary, the absence
of these two bands in the precipitated catalyst (Cu/SAPO-34-P,
Fig. 7), the spectrum of which resembles that of the parent H/
SAPO-34, suggests that the Cu species in this sample do not affect
the SAPO-34 framework, which again confirms that CuO species
are mostly on the external surface of SAPO-34.
ꢀ1
at 891 and 844 cm appeared which are associated with internal
asymmetric framework vibrations perturbed by copper cations. This
further confirms that Cu ions species in the framework were occu-
pied by NH
exchange enhances the NH
Lewis acid sites and reduces the Brønsted acidic strength of zeolite.
3
3
with the formation of Cu–NH species. As a result, Cu
3
adsorption ability by introducing more
3
.6.4. NH
The decomposition behaviors of ammonium species over all the
catalysts were also studied by DRIFT spectra as shown in Fig. 10.
Contrary to what was observed during NH adsorption process,
3
desorption
4.2. Relation between acidity and the structure of copper species
3
the adsorbed species are released at different temperatures while
the intrinsic species of catalysts are restored, due to different acidic
The acidity of the catalysts is considered as an essential factor
for SCR reaction. Two stronger bands at 3627 and 3600 cm in
ꢀ1

L. Wang et al. / Journal of Catalysis 289 (2012) 21–29
27
Fig. 10. DRIFT spectra of NH
f:500 °C;
3
-TPD over H/SAPO-34 (a), Cu/SAPO-34-E (b), and Cu/SAPO-34-P (c); units in Kubelka–Munk; a:50 °C; b:100 °C; c:200 °C; d:300 °C; e:400 °C;
of Brønsted acid sites. It could also be concluded from the ammo-
nia desorption results (Fig. 6) that the copper ion exchange en-
Table 4
Amount of Brønsted acid sites in Cu/SAPO-34 measured by IR.
hances the Lewis acidity but reduces the Brønsted acidity of
+
x+
Catalyst
H/SAPO-
Cu/SAPO-34- Cu/SAPO-34-
SAPO-34, due to the replacement of H by Cu at the bridge hydro-
xyl sites Si–(OH)–Al, whereas the copper oxide loaded on the
external surface has little impact on the acidity of zeolite, which
could maintain a strong Brønsted acidity of the catalyst.
3
4
P
E
Integrated area (3560–
105.60
—
95.54
72.87
ꢀ
1
3
660 cm )
[
B
Catal.]/[BH/SAPO-34
]
0.90
0.69
4.3. Relation between SCR activity and nature of copper species
the ex situ IR spectra (Fig. 7) are related to the Brønsted acid sites in
The present work demonstrates that copper species could locate
the catalysts [37,38]. The integrated area in the range of 3560–
at different positions in SAPO-34 via different preparation meth-
ods, resulting in significant differences in both NH -SCR activities
ꢀ
1
ꢀ1
3
660 cm , calibrated by the intensity of the band at 1362 cm
3
which is assigned to the framework vibration as an internal stan-
dard, represents the amount of the Brønsted acid sites in the sam-
ples. Normalized to that in H/SAPO-34, the relative amounts of
Brønsted acid sites in Cu/SAPO-34 are listed in Table 4, suggesting
that the ion-exchange consumed Brønsted acid sites to form iso-
lated copper ions while precipitation maintained a large quantity
and structural/chemical properties. The SCR activity results in
Fig. 1 reveal that the Cu-modified SAPO-34 catalysts with different
structures of copper species exhibit significant difference in NH
SCR activities. Generally, the Cu ion-exchanged SAPO-34 is the
most active catalyst for NH -SCR reaction in the whole tempera-
ture range, while the precipitated catalyst Cu/SAPO-34-P shows
3
-
3

2
8
L. Wang et al. / Journal of Catalysis 289 (2012) 21–29
Table 5
Reaction rates at 220 °C and apparent activation energies.
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
ꢀ1
Catalysts
Reaction rate (mol g
s
)
Cu content (mol g
)
TOF (s
)
a
Apparent activation energy E (kJ mol )
ꢀ
7
ꢀ4
Cu/SAPO-34-E
Cu/SAPO-34-P
9.63 ꢂ 10
1.25ꢂ10
0.0077
0.0017
9.8
16.8
ꢀ
7
ꢀ4
2.06ꢂ10
1.25ꢂ10
the estimated apparent activation energy may be affected by
factors such as NO or NH adsorption on the catalyst surface.
3
Although the CuO species on the external surface are capable of
undergoing through reduction/oxidation cycles, their weak inter-
action with the zeolite acid sites may lead to lower NH
tion rates and the stronger Brønsted acidity and external CuO
species could accelerate the NH oxidation at high temperatures.
3
-SCR reac-
3
On the other hand, isolated Cu ions at the exchange sites are the
predominant species in the exchanged sample, which show much
faster rates of NH SCR reactions. The Cu ions residing in the prox-
3
imity of the acid sites in the zeolite with a stronger Lewis acidity
could provide high catalytic reactivity even at low temperatures.
5
. Conclusions
Ion-exchange and precipitation methods were used to prepare
Cu-loaded SAPO-34 catalysts. Cu species existed as isolated cations
inside the SAPO-34 pores in the ion-exchanged samples, whereas
the predominant species in the precipitated sample were CuO
clusters dispersed on the external surface of the zeolite. The ion-
exchanged sample was found to have superior catalytic activities
x
Fig. 11. Arrhenius plots of the turnover frequency (TOF) with respect to the total Cu
amount on Cu catalysts; conditions: 200 ppm NO, 200 ppm NH
3
, 8% O
2
, 10% H
2
O,
ꢀ
1
ꢀ1
1
0% CO
2
balanced with N
2
; flow rate: 1 L min ; GHSV: 1,200,000 h
.
for ammonia SCR. For Cu/SAPO-34, the active sites for NH
3
-SCR
reactions are likely the isolated Cu ions inside the pores at the ex-
change sites, and copper oxide species could promote NH
tion reaction prior to SCR reaction at high temperatures.
3
oxida-
3
strong NH oxidation ability rather than SCR activity at the high
temperatures (>350 °C), suggesting that the isolated copper ions
at the exchange sites are active for the SCR reaction and that cop-
Acknowledgments
per oxide species could promote NH
SCR reaction.
3
oxidation reaction prior to
The authors would like to acknowledge the financial support
from GM Global R&D on this project, and Ms. Lei Wang is grateful
for the China Scholarship Council Postgraduate Scholarship Pro-
gram provided by the Ministry of Education, China. The authors
would also like to thank Misle Tessema, Michael P. Balogh, and
Nicholas Irish for the XRD, STEM, and ICP measurements.
In order to quantitatively estimate the contribution of different
copper species to the SCR reaction, NO reaction rates were mea-
sured for the two Cu/SAPO-34 samples. A higher space velocity
ꢀ1
(
1,200,000 h ) was used to achieve the kinetically controlled con-
ditions and keep the NO conversions low (less than 40%). As listed
in Table 5, the reaction rate at 220 °C on the Cu ion-exchanged
ꢀ
7
ꢀ1 ꢀ1
sample (Cu/SAPO-34-E) is 9.63 ꢂ 10 mol g
s , which is more
than four times of that on the Cu-precipitated catalyst
s ). In addition, the turnover frequency
TOF), which is defined as the number of NO molecule converted
Appendix A. Supplementary data
ꢀ7
ꢀ1 ꢀ1
(
2.06 ꢂ 10 mol g
(
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2012.01.012.
per Cu per second, could be calculated with respect to the total
copper contents and also shows the similar trend. These rates are
comparable to those reported on Fe/ZSM-5 [43], but at much lower
temperatures (220 °C versus 300 °C). Fig. 11 shows the Arrhenius
plots of the turnover frequency (TOF) for the SCR reaction with
respect to the total copper contents in the temperature range of
References
[
[
1] M. Iwamoto, H. Furukawa, Y. Mine, F. Uemura, S. Mikuriya, S. Kagawa, J. Chem.
Soc. Chem. Commun. (1986) 1272.
2] M. Iwamoto, H. Yahiro, Y. Yu-u, S. Shundo, N. Mizuno, Shokubai (Catalyst) 32
(1990) 430.
2
20–350 °C. Over this temperature range, the NH
are consistent with the NO conversions for all the catalysts with
no N O and NO detected, suggesting that the NH oxidation, NO
oxidation, and N O formation are negligible. The apparent activa-
tion energy of the SCR reaction for Cu/SAPO-34-E is 9.8 kJ mol
which is lower than values reported for Cu/ZSM-5 catalysts
3
conversions
[
[
[
3] G. Busca, L. Lietti, G. Ramis, F. Berti, Appl. Catal. B 18 (1998) 1–36.
4] G. Qi, R.T. Yang, R. Chang, Appl. Catal. B 51 (2004) 93–106.
5] Z.C. Si, D. Weng, X.D. Wu, J. Li, G. Li, J. Catal. 271 (2010) 43–51.
2
2
3
2
[6] L.J. Lobree, I.C. Hwang, J.A. Reimer, A.T. Bell, J. Catal. 186 (1999) 242–253.
ꢀ
1
[
[
[
7] C.C. Liu, H. Teng, Appl. Catal. B 58 (2005) 69–77.
8] H. Sjövall, R.J. Blint, L. Olsson, Appl. Catal. B 92 (2009) 138–153.
9] H. Praliaud, S. Mikhailenko, Z. Chajar, M. Primet, Appl. Catal. B 16 (1998) 359–
374.
,
ꢀ1
ꢀ1
(
85 kJ mol [44] and 49 kJ mol [45]). The reason for the differ-
ence is not clear and might suggest a different reaction mechanism
on these catalysts. We examined the effect of external mass trans-
port and found that the reaction rates were not affected by external
mass transport (results are included in the Supporting Informa-
tion). Small size (less than 200 mesh size) catalyst particles were
used in this study and internal diffusion limitation was unlikely
but could not be completely ruled out. It is important to note that
[10] F. Seyedeyn-Azad, D.K. Zhang, Catal. Today 68 (2001) 161–171.
[
[
[
11] M. Iwamoto, H. Hamada, Catal. Today 10 (1991) 57.
12] T. Ishihara, M. Kagawa, F. Hadama, Y. Takita, J. Catal. 169 (1997) 93–102.
13] C. Torre-Abreu, C. Henriques, F.R. Ribeiro, G. Delahay, M.F. Ribeiro, Catal. Today
54 (1999) 407–418.
[
[
14] J.M. Fedeyko, B. Chen, H.Y. Chen, Catal. Today 151 (2010) 231–236.
15] T. Komatsu, M. Nunokawa, I.S. Moon, T. Takahara, S. Namba, T. Yashima, J.
Catal. 148 (1994) 427.
[16] S. Kieger, G. Delahay, B. Coq, Appl. Catal. B 25 (2000) 1.

L. Wang et al. / Journal of Catalysis 289 (2012) 21–29
29
[
[
17] A.V. Salket, W. Weisweiler, Appl. Catal. A 203 (2000) 221.
[31] J.O. Petunchi, G. Sill, W.K. Hall, Appl. Catal. B 2 (1993) 303.
[32] J. Valyon, W.K. Hall, Catal. Lett. 19 (1993) 109.
[33] M. Iwamoto, H. Yahiro, N. Mizuno, W.Z. Zhang, Y. Mine, H. Furukawa, J. Phys.
Chem. B 96 (1992) 9360.
18] R. Moreno-Tost, J. Santamaría-González, E. Rodríguez-Castellón, A. Jiménez-
López, M.A. Autié, E. González, M. Carreras Glacial, C. De las Pozas, Appl. Catal.
B 50 (2004) 279–288.
[
19] A. Frache, B.I. Palella, M. Cadoni, R. Pirone, H.O. Pastore, L. Marchese, Top. Catal.
[34] J. Valyon, W.K. Hall, J. Phys. Chem. B 97 (1993) 1204.
[35] A. Corma, A. Palomares, F. Márquez, J. Catal. 170 (1997) 132–139.
[36] M. Richter, M.J.G. Fait, R. Eckelt, M. Schneider, J. Radnik, D. Heidemann, R.
Fricke, J. Catal. 245 (2007) 11–24.
[37] G.V.A. Martins, G. Berlier, C. Bisio, S. Coluccia, H.O. Pastore, L. Marchese, J. Phys.
Chem. C 112 (2008) 7193–7200.
[38] B. Onida, Z. Gabelica, J. Lourencüo, E. Garrone, J. Phys. Chem. 100 (1996)
11072–11079.
[39] P.E. Fanning, M.A. Vannice, J. Catal. 207 (2002) 166–182.
[40] G.D. Lei, B.J. Adelman, J. Sárkány, W.M.H. Sachtler, Appl. Catal. B 5 (1995) 245–
256.
2
2 (2003) 53–57.
[
[
[
20] G.I. Panov, V.I. Sobolev, A.S. Kharitonov, J. Mol. Catal. 61 (1990) 85.
21] K. Ebitani, M. Morokuma, A. Morikawa, Stud. Surf. Sci. Catal. 84 (1994) 1501.
22] R. Kefirov, A. Penkova, K. Hadjiivanov, S. Dzwigaj, M. Che, Micro. Meso. Mater.
1
16 (2008) 180–187.
23] S. Brandenberger, O. Kröcher, A. Tissler, R. Althoff, Catal. Rev. Sci. Eng. 50
2008) 492.
[
[
[
[
(
24] I. Bull, W.-M. Xue, P. Burk, R.S. Boorse, W.M. Jaglowski, G.S. Koermer, A. Moini,
J.A. Patchett, J.C. Dettling, M.T. Caudle, US Patent 7610,662, 2009.
25] J. Hun Kwak, R.G. Tonkyn, D. Heui Kim, J. Szanyi, C.H.F. Peden, J. Catal. 275
(
2010) 187–190.
26] D.W. Fickel, E. D’Addio, J.A. Lauterbach, R.F. Lobo, Appl. Catal. B 102 (2011)
41–448.
[41] T. Thuy Hanh Dang, H.L. Zubowa, U. Bentrup, M. Richter, A. Martin, Micro.
Meso. Mater. 123 (2009) 209–220.
4
[42] H. Sjövall, E. Fridell, R.J. Blint, L. Olsson, Topics Catal. 42–43 (2007) 113–117.
[43] R.Q. Long, R.T. Yang, J. Catal. 188 (1999) 332–339.
[44] L. Olsson, H. Sjövall, R.J. Blint, Appl. Catal. B 81 (2008) 203–217.
[45] D. Chatterjee, T. Burkhardt, M. Weibel, I. Nova, A. Grossale, E. Tronconi, SAE
Paper 2007-01-1136, 2007.
[
[
[
[
27] M. Hartmann, L. Kaven, Chem. Rev. 99 (1999) 635–663.
28] K.C.C. Kharas, Appl. Catal. B 2 (1993) 207.
29] H. Praliaud, S. Mikhailenko, Z. Chajar, M. Primet, Appl. Catal. B 16 (1998) 359.
30] B.K. Cho, J. Catal. 155 (1994) 84.