An in situ UV-vis and FTIR spectroscopy study of the effect of H 2 and CO during the selective catalytic reduction of nitrogen oxides over a silver alumina catalyst

Article abstract of DOI:10.1016/j.jcat.2005.08.006

The effect of hydrogen and carbon monoxide on the selective catalytic reduction of NO_x with decane over a 2 wt% Ag/Al₂O ₃ catalyst was investigated. Hydrogen promoted NO_x conversion during the SCR reaction, in contrast to carbon monoxide, which had a slightly detrimental effect on catalyst activity. However, in situ UV-vis spectroscopy showed that both carbon monoxide and hydrogen promoted the formation of Ag clusters during the reaction. This, coupled with previous EXAFS experiments and UV-vis data on the rate of change of cluster formation, demonstrate that the enhanced activity from the addition of hydrogen to the NO_x-SCR mix cannot be attributed exclusively to the formation of Ag clusters. It is proposed that hydrogen itself participates directly in the reaction mechanism.

Full text of DOI:10.1016/j.jcat.2005.08.006

Journal of Catalysis 235 (2005) 195–200
www.elsevier.com/locate/jcat
An in situ UV–vis and FTIR spectroscopy study of the effect of H and CO
2
during the selective catalytic reduction of nitrogen oxides over
a silver alumina catalyst
a
a
b,∗
b
b
a
a
ˇ
B. Wichterlová , P. Sazama , J.P. Breen , R. Burch , C.J. Hill , L. Capek , Z. Sobalík
a
J. Heyrovský Institute of Physical Chemistry, Academy of Sciences of the Czech Republic, 182 23 Prague 8, Czech Republic
b
CenTACat, School of Chemistry, Queens University Belfast, Belfast BT9 5AG, N. Ireland
Received 4 April 2005; revised 3 August 2005; accepted 5 August 2005
Abstract
The effect of hydrogen and carbon monoxide on the selective catalytic reduction of NOx with decane over a 2 wt% Ag/Al O catalyst was
2
3
investigated. Hydrogen promoted NOx conversion during the SCR reaction, in contrast to carbon monoxide, which had a slightly detrimental
effect on catalyst activity. However, in situ UV–vis spectroscopy showed that both carbon monoxide and hydrogen promoted the formation of
Ag clusters during the reaction. This, coupled with previous EXAFS experiments and UV–vis data on the rate of change of cluster formation,
demonstrate that the enhanced activity from the addition of hydrogen to the NOx-SCR mix cannot be attributed exclusively to the formation of
Ag clusters. It is proposed that hydrogen itself participates directly in the reaction mechanism.

2005 Elsevier Inc. All rights reserved.
Keywords: SCR-NOx; Ag/alumina; Ag clusters; H effect in SCR-NOx
2
1
. Introduction
can be drawn from these investigations; first, it is evident that
hydrogen promotes NOx reduction over Ag/γ -Al2O3 catalysts
when using a range of lower alkanes and alkenes [6] and higher
alkanes [13,16] as reductants. The hydrogen effect is limited to
Ag-based catalysts but is very dependent on the support used.
The activity of Ag supported on α-Al2O3, TiO2, ZrO2, SiO2, or
Ga2O3 is not enhanced by the addition of hydrogen [6]; how-
ever, Ag/γ -Al O and Ag/MFI catalysts show markedly im-
proved performance when hydrogen is added to the feed [7,8].
Hydrogen has a significant promotional effect on the par-
tial oxidation of the hydrocarbon and the oxidation of NO to
NO2 [6,13]. Shibata et al. [9] claimed that the rate of NOx re-
duction in the SCR reaction is directly dependent on the rate
of partial oxidation of the hydrocarbon to surface acetate, and
that hydrogen has a remarkable effect in promoting this oxi-
dation reaction, thus having a positive effect on the proposed
rate-determining step of the NOx SCR reaction. There is no
disagreement as to the effect of hydrogen on the SCR reaction;
however, there is some debate as to the mechanism by which
the hydrocarbon and NO are activated.
Since the pioneering work of Iwamoto et al. [1] and Held et
al. [2], considerable effort has been expended in searching for
active catalysts for the selective catalytic reduction of NOx and
also in studying the mechanism of reaction over these catalysts.
It is well established that Ag/Al2O3 catalysts are among the
most active and selective for the selective catalytic reduction
of NOx to N2 [3]. However, until recently, one of the draw-
backs preventing commercialization of these catalysts was very
2
3
◦
poor activity at temperatures below ca. 350 C. This problem
has been overcome by the recent discovery that adding hydro-
gen to the feed stream can result in a remarkable improvement
in the level of NOx reduction and the range of temperature
over which NOx can be reduced [4,5]. Several authors have
investigated this “hydrogen effect” using different catalyst for-
mulations and different gas mixes [6–17]. Several conclusions
*
Corresponding author. Fax: +44-2890382117.
E-mail address: j.breen@qub.ac.uk (J.P. Breen).
0021-9517/$ – see front matter  2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2005.08.006

196
B. Wichterlová et al. / Journal of Catalysis 235 (2005) 195–200
δ+
monitoring of NOx conversion and the UV–vis bands of Agn
Satokawa and co-workers [5–9] support an Ag cluster
model. They used UV–vis spectroscopy to show that during
the SCR reaction (in the absence of H2), the silver in a low-
loading Ag catalyst is predominantly in the form of isolated
clusters revealed that the sharp decrease in NOx conversion af-
ter the removal of hydrogen from the feed was not accompanied
by a similar decrease in the number of Ag clusters, this num-
ber declined only gradually. Thus the high SCR-NOx activity of
Ag/alumina in the presence of hydrogen cannot be exclusively
ascribed to the presence of Ag clusters. Hydrogen itself dramat-
ically enhances the SCR-NOx reaction by initiating most likely
radical-type reactions and increasing the rate of all of the reac-
tion steps.
+
Ag ions. However, they noticed that adding small concen-
trations of hydrogen (<1%) promoted the formation of small
δ+
clusters of Agn (with estimated 2 ꢀ n ꢁ 4). They suggested
that at higher hydrogen concentrations (>1%), metallic Agm
(
3 ꢀ m ꢁ 5) clusters were formed, promoting unselective ox-
δ+
idation of the hydrocarbon. They argued that it is the Ag
n
clusters that are responsible for the high activity of the Ag-
based catalysts (Ag/MFI and Ag/γ -Al2O3), although they did
not give any details of the mechanism by which these clusters
promote the reaction, other than to comment that they promote
the formation of surface acetates and oxidation of NO to NO2.
The aim of the present study was to use in situ UV–vis and
FTIR techniques to probe the effects of hydrogen and carbon
monoxide on the state of silver in the catalyst and on the surface
reaction intermediates formed during the reaction. We found
that although both of these co-reductants significantly increased
the number of small metallic silver clusters, only hydrogen in-
creased NOx reduction to nitrogen by increasing the rates of all
reaction steps, manifested in the detected surface intermediate
species.
0
Richter et al. [11] invoked the importance of Ag2O/Ag nan-
oclusters in their interpretation of the reaction mechanism dur-
ing the hydrogen-promoted SCR reaction. They proposed that
hydrogen rapidly reduces the nano-sized clusters of Ag2O to
0
Ag , and that these smaller silver clusters promote dissociation
2
. Experimental
of NO and O2. Although the authors found no direct evidence
for the formation of adsorbed O ions on the surface of the Ag,
they suggested that such a species is likely to form and to be ac-
tive in the conversion of NO to nitrite species and of propane to
O-functionalized intermediates or to surface acetate.
The catalyst used in this work was prepared by impregnat-
ing Al2O3 with a silver nitrate solution, followed by drying and
calcination as described previously [18]. Catalyst testing was
performed in a 5-mm-i.d. quartz tubular downflow reactor. The
sample was held in place between plugs of quartz wool, and a
thermocouple was placed in the center of the catalyst bed. The
reactant gases (NO, NO2, and O2) and the carrier gas (He) were
fed from independent mass flow controllers, whereas n-decane
and H2O were fed using a Razel syringe pump. All reactor
lines were heated to prevent condensation. The NO and total
NOx were determined by a Signal 4000 series chemilumines-
cence detector. In these studies NOx conversion is defined as
the reduction of NO and NO2 to N2 and N2O. The experimen-
tal conditions for the SCR reaction were as follows: total gas
Brosius et al. [12] recently investigated the effect of adding
hydrogen to a NOx-containing stream (in the absence of hy-
drocarbons) over Ag/Al2O3. They provided evidence that the
promotional effect of hydrogen is related to its role in reducing
0
Ag(I) to Ag . Extrapolating this to the SCR of NOx reaction,
they suggested that the hydrogen reacts with stable Ag nitrate
0
to form Ag , which can activate the hydrocarbon. This is con-
sistent with the model proposed earlier by Burch [14].
In earlier publications [13,16], we proposed that the enhanc-
ing effect of hydrogen was not related to Ag cluster formation,
but that the hydrogen reacts directly to either promote the for-
mation and storage of a reactive species that can then read-
ily reduce NOx or, alternatively, to remove a species that is
strongly bound at low temperatures. We found that hydrogen
3
−1 −1
flow: GHSV, 62,256 cm g
h ; NO, 720 ppm; O2, 4.3%;
n-C10H22, 4350 ppm (as C1); H2O, 7.2%; CO2, 7.2%; CO or
H2: 7200 ppm; balance He.
+
3+
promoted the conversion of Ag CN to Al NCO at temper-
atures >520 K. We suggested that at temperatures <520 K,
hydrogen could remove strongly bound nitrates, which can be
a potential poison to the reaction in the absence of hydro-
gen. This direct mechanism does not require the formation of
Ag clusters; however, it does not rule out the possibility that
clusters may form as a consequence of the introduction of hy-
drogen. More recent in situ EXAFS experiments demonstrated
that small clusters of Ag, containing an average of 3 atoms,
were formed under SCR reaction conditions, but that adding
hydrogen or carbon monoxide did not lead to any further clus-
ter formation [15]. In another study [16], we provided evidence
that hydrogen directly helps enhance the SCR-NOx reaction
and all of the related reaction steps. By recording in situ UV–vis
spectra during the decane-SCR-NOx reaction, we showed that
2
.1. In situ UV–vis spectroscopy measurements
In situ UV–Vis diffuse reflectance spectra of the silver alu-
mina catalysts were collected under the reaction mixture at the
reaction temperature. The spectra were recorded using a Perkin
Elmer Lambda 19 UV–vis-NIR spectrophotometer with a dif-
fuse reflectance attachment equipped with a BaSO4-coated in-
tegration sphere and a heated “homemade” flow-through quartz
optical cell. The spectra were monitored in the range 4000–
−1
5
0,000 cm . The absorption intensity was calculated using the
Schuster–Kubelka–Munk equation,
2
F(R ∝) = (1 − R ) /2R .
∞
∞
2.2. In situ FTIR spectroscopy measurements
δ+
small metallic Agn clusters (n ꢀ 8) are formed even in the
absence of hydrogen. Experiments involving adding/removing
hydrogen from the reactant feed and simultaneous time-scale
FTIR spectra were collected under reaction conditions us-
ing a FTIR spectrometer (Thermo Nicolet) equipped with a

B. Wichterlová et al. / Journal of Catalysis 235 (2005) 195–200
197
(a)
Fig. 1. NOx conversion as a function of temperature for the SCR of NOx with
decane (F); with 0.72% CO added (Q) and with 0.72% H added (2) using
2
conditions given in the text.
stainless steel flow-through IR cell reactor (in situ Research
Instruments) with CaF2 windows. A DTGS KBr detector was
used, and FTIR spectra were measured with resolution of
−
1
2
(
cm . The catalyst, in the form of a thin self-supporting wafer
−2
ca. 10 mg cm ), was calcined at 723 K in an oxygen+helium
stream for 1 h before all experiments and then cooled to the re-
action temperature. The spectrum of the calcined sample (mon-
itored in a helium stream with 4.3% water vapor) before the
introduction of reactants was subtracted from the spectra col-
lected under in situ conditions during reaction.
(b)
3
. Results
Fig. 2. (a) In situ diffuse reflectance UV–vis-NIR spectra of Ag/Al O during
2
3
hydrogen or CO oxidation over Ag/Al O at 473 K. Feed: 7200 ppm H or
2
3
2
7
200 ppm CO, 4.3% O , rest He. (b) Difference spectra to that of Ag/Al O in
The effects of adding CO and H2 to the feed of the SCR-NOx
2
2
3
flow of O .
2
with n-decane are depicted in Fig. 1. The addition of hydrogen
results in markedly improved catalyst activity over the tem-
perature range of the tests. These results are similar to those
obtained for the SCR reaction using n-octane as the primary re-
ductant, although the effect of hydrogen was more dramatic in
that case [13]. In contrast, the addition of CO has no (or a small
negative) effect on catalyst activity.
Agnδ+
clusters for the CO + O2 reaction were substantially
lower that those obtained during the H2 + O2 reaction. The
difference spectra for both reactions clearly indicate that the
+
intensity of the bands of the Ag ions did not change signif-
icantly when Agn
δ+
clusters were formed. This finding is in
agreement with the very high absorbance of Ag clusters com-
pared with Ag ions and the fact that not all Ag ion species
are detected by UV–vis spectra [16]. Apparently, only a very
small amount of the Ag ions (<3%) were reduced to Ag clus-
ters. The apparent increased intensity of the bands of Ag ions
Fig. 2a shows in situ UV–vis spectra of the silver alu-
mina catalyst under flows of O2, H2 + O2, and CO + O2 at
+
+
4
73 K at similar conditions to that of the SCR-NOx reaction.
+
With Ag/alumina exposed to a stream of oxygen, bands at
+
−
1
4
1,600 and 46,600 cm , corresponding to the electronic tran-
+
+
sitions of isolated Ag ions, were observed (the spectra of Ag
ion and small metallic Agn clusters were interpreted previ-
ously [16]), along with very low-intensity absorption around
in mixtures of H2 +O2 and CO+O2 compared with those in O2
δ+
may be caused by an overlap of the absorbance of Ag clusters
−1
to the region above 40,000 cm
.
−1
2
8,000–30,000 cm , attributed to a small number of silver
Figs. 3a and 3b show the in situ UV–vis spectra of Ag/
alumina collected during the decane-SCR-NOx reactions under
similar conditions as in the catalytic tests. The spectra revealed
that even at 473 K, where low conversion of NO to N2 (<10%)
occurs, traces of Agn clusters are formed (low absorption
intensity around 28,000 and 38,000 cm ). The presence of ad-
clusters. The spectrum obtained during catalyst exposure to O2
was subtracted from those obtained during the H2 + O2 and
CO + O2 reactions (Fig. 2b). For the H2 + O2 reaction, the ab-
sorption maxima observed at ca. 24,000, 29,000, 31,000, and
δ+
−
1
−1
3
4,000 cm (obtained from the second derivative mode, not
δ+
shown) were attributed to Ag
tail below 24,000 cm indicates the presence of some larger
clusters, for n ꢀ 8 [16]. The
sorbed nitrates is reflected in increased absorbance intensity,
with a maximum at 47,000 cm . Adding CO to the reaction
n
−1
−1
metallic Ag particles. For the CO + O2 reaction, the maxima of
mixture produced increased absorption intensities, with max-
ima at ca. 25,000, 34,000, and 38,000 cm (obtained from the
second derivative mode). These bands can be attributed to the
δ+
−1
absorbance due to Ag
clusters were located at ca. 29,000,
n
−1
3
1,000, and 34,000 cm . The intensities of the bands due to

198
B. Wichterlová et al. / Journal of Catalysis 235 (2005) 195–200
Fig. 4. In situ FTIR spectra of adsorbed species during C
H -SCR NO (- - -),
10 22
(a)
with CO added (···) and with H added (—) at 473 K. Feed: 720 ppm NO;
2
4
35 ppm n-C H ; 4.3% O ; 4.3% H O; (7200 ppm CO). Subtracted spec-
10 22 2 2
trum was taken from sample treated in helium/water mixture.
Table 1
Bands observed on Ag/Al O catalyst during in situ FTIR experiments and the
2
3
corresponding surface species and vibrations to which they are assigned
Wavenumber
Surface species
Vibration
References
−
1
(cm
)
−
a
1
580
Carboxylate COO
ν
[19,20]
OCO
s
1457
ν
OCO
−
a
1595
1390
1373
Formate HCOO
ν
[19–21]
OCO
δ
ν
CH
s
OCO
1
550
245
Monodentate
ν
ν
[22–24]
[22–24]
N=O
a
(
b)
−
1
nitrate NO
3
ONO
Fig. 3. (a) In situ diffuse reflectance UV–vis-NIR spectra of Ag/Al O during
exposure to different reactant gas mixes at 473 K. SCR feed: 720 ppm NO;
35 ppm n-C ; 4.3% O ; 4.3% H O; (7200 ppm CO or H ). (b) Differ-
2
3
1590
295
Bidentate
ν
ν
N=O
a
ONO
−
−
1
nitrate NO
3
4
H
22
1
0
2
2
2
a
ence spectra to that of Ag/Al O in flow of O .
1220
560
1300
Nitrite NO2
ν
[19]
[25]
2
3
2
ONO
1
Ad-NOx
δ+
−
formation of small Agn clusters. Adding H2 to the feed in-
stead of CO substantially increased the intensity of the broad
band ranging from 20,000 to 40,000. This band showed sim-
ilar features (intensity maxima at 24,000, 28,000, 34,000, and
Bulk-like NO
2
1645
2230
2260
Acrylate
ν
[9]
C=C
Isocyanate—(AlIVNCO)
Isocyanate—(AlVINCO)
Cyanide—AgCN
[26]
[26]
[21]
−
1
3
8,000 cm ) to those obtained under a CO-containing feed, in-
2146
dicating that similar types of silver clusters are formed in each
−1
case. The increase in spectrum intensity around 47,000 cm
tant to identify the surface species formed during the addition
after hydrogen or CO addition can be accounted for by the pres-
ence of some adsorbed species on Ag or Al ions (e.g., nitrates).
Fig. 4 compares the FTIR spectra measured at 473 K under
the decane-SCR-NO reaction mix and under the same mix with
the addition of CO or H2. This figure clearly shows that hydro-
gen addition, unlike CO addition, produces a large increase in
the quantity of surface species on the catalyst. The quantity of
surface species slowly increases during the first 60 min of SCR-
NOx reaction; however, admission of H2 into the feed stream
produces a sudden, dramatic increase in the integrated peak ar-
eas. This is in marked contrast to the negligible effect on peak
areas from adding CO. Clearly, H2 promotes the formation of
various surface species, whereas CO does not.
of H and CO. Figs. 5a–5c zoom in on specific regions of the
IR spectrum and show the changes resulting from adding CO
2
and H . Table 1 summarizes the attribution of peaks to various
2
surface species.
Fig. 5a shows the evolution of several bands over time in
the SCR mix (the first four spectra) and then after introduc-
tion of CO to the feed (the last four spectra). The peaks cen-
−
1
tered at 1457 and 1587 cm can be attributed to a carboxylate
−
1
species, and that at 1558 cm can be attributed to an ad-NO
species. The peaks at 1373 and 1390 cm are most likely due
x
−1
−1
to formate species. The broad band at 1230 cm reflects ni-
trite species. All bands increase with time on stream except
the broad peak at 1230 cm , which decreases slowly. This
demonstrates that nitrite species, which are formed very quickly
after introduction of the feed, are gradually oxidized to nitrate
species with increasing time on stream.
−
1
Once we found that adding both H2 and CO to the SCR-NOx
feed promotes the formation of Ag clusters but only the pres-
ence of hydrogen enhances NOx–N2 conversion, it was impor-

B. Wichterlová et al. / Journal of Catalysis 235 (2005) 195–200
199
160 cm 1) was detected, but no –NCO (2230–2250 cm
−
−1
)
2
species were observed, probably due to the low NOx conver-
sion (<10%). The spectrum intensity of the –CN species was
the same with and without CO; this is not shown in Fig. 5a
but is similar to that shown in Fig. 5c without hydrogen. The
addition of hydrogen resulted in a dramatic increase in the com-
position and number of adsorbed species. Fig. 5b shows a sub-
stantial increase in the number of oxidized carbon-containing
−
1
species, such as carboxylates (1457 and 1587 cm ) and for-
−1
mates (1373 and 1390 cm ). The nitrites are immediately
transformed to nitrates (1295, 1245, 1550, and 1590 cm ), of
which bidentate nitrates (1295 and 1590 cm ) increased more
than monodentate species (1245 and 1550 cm ), with the lat-
ter consumed faster during propagation of the SCR process. The
peak at 1558 cm due to an ad-NOx species also increased.
In the absence of hydrogen at 473 K, only –CN groups are
detected (Fig. 5c; note that NOx conversion was <10%). The
addition of hydrogen results in the appearance of isocyanate
species formed on different Al sites, AlIVNCO at 2230 cm
and AlVINCO at 2260 cm , and the simultaneous slow in-
crease of –CN species (2150 cm ). Such changes were not
observed under CO addition, and in this case the spectra re-
semble those of the SCR mix without any added reductant
When higher conversion of NOx was achieved by increasing
the contact time (∼20%), similar development of the spec-
tra was observed with formation of the –NCO bands at 2230
and 2260 cm . At higher temperatures (>523 K), a similar
trend was observed for the SCR of NOx with octane and hy-
drogen over the same catalyst. But in this case the appearance
of the –NCO species was accompanied by a disappearance of
the –CN species [13]. A new species appeared as a shoulder
peak at 1644 cm on addition of H2, this peak occurred at a
position similar to that observed during addition of H2 to an oc-
tane SCR-NOx feed [13]. This was tentatively assigned to an
organic species containing a C=N double bond, but could also
be attributed to the C=C stretch of an acrylate species [9].
−1
−
1
−
1
(a)
−1
−1
−1
−1
−1
(b)
−1
4
. Discussion
Hydrogen has a significant promotional effect on the con-
version of NOx for the decane SCR reaction over Ag/Al2O3,
although the extent of the promotional effect is not as great as
that observed with other alkanes and alkenes [6,13]. CO was
added to the reactant mix to investigate whether another re-
ductant could have a similar effect to hydrogen. CO is readily
(
c)
Fig. 5. (a) In situ FTIR spectra of adsorbed species during C
H -SCR NO
0 22
1
(
—) and with CO added (···) as a function of time (10 min intervals) at 473 K.
Feed: 720 ppm NO; 435 ppm n-C
H ; 4.3% O ; 4.3% H O; (7200 ppm
22 2 2
1
0
+
oxidized to CO2 over supported Ag catalysts; in fact, Ag ions
CO). (b) In situ FTIR spectra of adsorbed species (—) and with H added (···)
2
as a function of time (10 min intervals) at 473 K. SCR feed: 720 ppm NO;
will preferentially oxidize CO in the presence of excess H at
2
4
35 ppm n-C
H
; 4.3% O ; 4.3% H O; (7200 ppm H ). (c) In situ FTIR
◦
10
22
2
2
2
low temperatures (<100 C) [27]. This may be an indication
spectra of adsorbed species during C
H
-SCR NO at the CN– and NCO–
1
0
22
that in the absence of a hydrocarbon, CO is a stronger reduc-
tant of oxidized Ag than H2. Clearly, CO does not promote
the SCR reaction, and in fact it has a slightly detrimental ef-
spectral regions before and after H addition measured at 10 min intervals at
73 K. Feed: 720 ppm NO; 435 ppm n-C H ; 4.3% O ; (7200 ppm H ).
After addition of CO to the reaction mix the spectrum was identical to that
without CO.
2
4
10 22 2 2
fect on the reduction of NO . Although CO has no direct effect
x
on the rate of NOx conversion, in situ diffuse reflectance UV–
The addition of CO (Fig. 5a) does not result in the formation
of additional bands; neither does it appear to increase the inten-
sity of the bands above what would be expected had the switch
to CO not occurred. A low concentration of –CN species (2130–
vis spectroscopy results show that, similar to hydrogen, it does
promote the formation of small Ag clusters during the CO ox-
idation and CO/decane-SCR-NO reactions. The extent of Ag
x
cluster formation from the addition of CO is less than that ob-

200
B. Wichterlová et al. / Journal of Catalysis 235 (2005) 195–200
served from hydrogen addition but is nevertheless significant.
Moreover, similar types of Ag clusters are formed.
effect on NOx conversion. These findings support our hypothe-
sis that hydrogen has a direct chemical function in the reaction
mechanism, as evidenced by the dramatic, steeply increased
concentration of adsorbed species and NOx conversion after the
addition of hydrogen to the SCR reaction. It is this direct action
of hydrogen, rather than Ag cluster formation, which is respon-
sible for the enhanced activity.
The results of the in situ FTIR measurements are revealing.
They show (in agreement with Eränen et al. [17] and Shibata
et al. [9]) that the addition of H2 to the reactant feed resulted
in a huge increase in the number and variety of surface species
formed during the reaction. This is in marked contrast to the re-
sults obtained from adding CO to the feed, which showed little
or no change in the number of adsorbed species. This demon-
strated that although CO can cause the formation of Ag clusters
on the catalyst surface, it has no impact on the rate of reaction
or on the nature and quantity of surface species formed during
the reaction.
Acknowledgments
The authors from the Heyrovský Institute wish to ac-
knowledge financial support from the Grant Agency of the
Academy of Sciences of the Czech Republic through project
T400400413. The CenTACat authors wish to acknowledge fi-
nancial support from the Support Programme for University
Research (SPUR-2).
The effect of hydrogen on promoting surface species forma-
tion is complex and remains to be elucidated. Development of
FTIR spectra after hydrogen addition clearly indicates enhance-
ment of all reaction steps. At low temperatures (470–570 K)
and under the reaction conditions of this study, hydrogen en-
hances the transformation of nitrites to nitrates and their further
transformation, and also promotes the conversion of cyanides to
isocyanates [13,16]. Isocyanate (–NCO) species are formed on
Al sites, and the concentrations of cyanide species also increase
slowly after the addition of hydrogen. Bion et al. [28] proposed
that the slow step in the SCR reaction was the transformation
References
[1] M. Iwamoto, H. Yahiro, Y. Yu-u, S. Shundo, N. Mizuno, Shokubai (Cata-
lyst) 32 (1990) 430.
[
[
2] W. Held, A. König, T. Richter, L. Pupper, SAE Paper 900496 (1990).
3] R. Burch, J.P. Breen, F.C. Meunier, Appl. Catal. B 39 (2002) 283, and
references therein.
[4] B. Krutzsch, Ch. Goerigk, S. Kurze, G. Wenninger, W. Boegner, F. Wirbe-
leit, U.S. Patent 5,921,076 (1999).
+
3+
of Ag CN species to Al NCO species.
tetra
[5] S. Satokawa, Chem. Lett. (2000) 294.
Alternatively, we speculated in an earlier paper on the SCR
of NOx with octane [13] that at low temperatures (<520 K)
the Ag/Al2O3 catalyst might be poisoned by strongly bound
nitrates, and that hydrogen could promote the reaction by re-
moving these species. Recently, Brosius et al. [12] showed that
in the absence of a hydrocarbon, hydrogen can promote the
removal of nitrates from Ag and promote the formation of ni-
trates on the Al2O3 support. These authors speculated that the
reduced Ag may then promote partial oxidation of the hydro-
[
6] S. Satokawa, J. Shibata, K. Shimizu, A. Satsuma, T. Hattori, Appl. Catal.
B 42 (2003) 179.
[7] J. Shibata, Y. Takada, A. Shichi, S. Satokawa, A. Satsuma, T. Hattori,
J. Catal. 222 (2004) 368.
[8] J. Shibata, K. Shimizu, Y. Takada, A. Shichi, H. Yoshida, S. Satokawa,
A. Satsuma, T. Hattori, J. Catal. 227 (2004) 367.
[9] J. Shibata, K. Shimizu, S. Satokawa, A. Satsuma, T. Hattori, Phys. Chem.
Chem. Phys. 5 (2003) 2154.
[10] M. Richter, R. Fricke, R. Eckelt, Catal. Lett. 94 (2004) 115.
11] M. Richter, U. Bentrup, R. Eckelt, M. Scneider, M.-M. Pohl, R. Fricke,
[
[
[
−1
Appl. Catal. B 51 (2004) 261.
carbon. The large number of species in the 1250–1700 cm
12] R. Brosius, K. Arve, M.H. Groothaert, J.A. Martens, J. Catal. 231 (2005)
region of our spectra makes it impossible to determine whether
reduction of Ag nitrate species and formation of Al nitrate
species occur after the introduction of hydrogen. Therefore, this
proposed mechanism of promotion cannot be ruled in or out.
The key point of this paper is to show that the highly posi-
tive “hydrogen effect” cannot be attributed to formation of Ag
clusters due to hydrogen as a co-reductant, but is associated
with a chemical function of hydrogen itself. We propose that the
role of hydrogen is to accelerate the critical SCR-NOx reaction
steps, particularly the slow steps in the reaction. The critical
rate-determining step may itself be a function of temperature
and reaction conditions. However, the formation of silver clus-
ters is not a sufficient condition for a catalyst that is active at
low temperatures.
344.
13] R. Burch, J.P. Breen, C.J. Hill, B. Krutzsch, B. Konrad, E. Jobson, L. Ci-
der, K. Eränen, F. Klingstedt, L.-E. Lindfors, Top. Catal. 30 (2004) 19.
14] R. Burch, Catal. Rev. Sci.-Eng. 46 (2004) 271.
[
[
15] J.P. Breen, R. Burch, C. Hardacre, C.J. Hill, J. Phys. Chem. B 109 (2005)
4805.
ˇ
[
16] P. Sazama, L. Capek, H. Drobná, Z. Sobalík, J. D eˇ de cˇ ek, K. Arve, B. Wi-
chterlová, J. Catal. 232 (2005) 302.
[17] K. Eränen, F. Klingstedt, K. Arve, L.E. Lindfors, D.Y. Murzin, J. Ca-
tal. 227 (2004) 328.
[
18] K. Eränen, L.-E. Lindfors, A. Niemi, P. Elfving, L. Cider, SAE Paper,
000-01-2813 (2000).
19] G. Busca, J. Lamotte, J.C. Lavalley, V. Lorenzelli, J. Am. Chem. Soc. 109
1987) 5197.
[20] V.A. Mathyshak, O.V. Krylov, Catal. Today 25 (1995) 1.
21] K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination
Compounds, fourth ed., Wiley–Interscience, New York, 1986.
22] M. Schraml-Marth, A. Wokaun, A. Baiker, J. Catal. 138 (1992) 306.
23] W.S. Kijlstra, D.S. Brands, E.K. Poels, A. Bliek, J. Catal. 171 (1997) 208.
24] K.I. Hadjiivanov, Catal. Rev. Sci.-Eng. 42 (2000) 71.
25] A.D. Cross, Introduction to Practical IR Spectroscopy, Butterworths, Lon-
don, 1964.
2
[
(
[
[
[
[
[
5
. Conclusions
The formation of small metallic Ag clusters does not explain
the dramatic increase in HC-SCR-NOx activity in the presence
of hydrogen. Ag clusters are also formed during the decane-
SCR-NOx reaction without hydrogen, and adding CO into the
feed increases the number of Ag clusters but has no significant
[
26] N. Bion, J. Saussey, C. Hedouin, T. Seguelong, M. Daturi, Phys. Chem.
Chem. Phys. 3 (2001) 4811.
[27] Z.P. Qu, S.T. Zhou, W.C. Wu, C. Li, X.H. Bao, Catal. Lett. 101 (2005) 21.
[28] N. Bion, J. Saussey, M. Haneda, M. Daturi, J. Catal. 217 (2003) 47.