Catalytic reduction of N2O and NO2 with methane over sol-gel palladium-based catalysts

Article abstract of DOI:10.1016/j.molcata.2006.06.024

Pd/TiO₂ and Gd-doped Pd/TiO₂ prepared through a 'one-pot' sol-gel method are shown to be active for the reduction of both N₂O and NO₂ using CH₄ as a reducing agent under excess O₂ conditions. In previous studies these catalysts were shown to effectively reduce NO with CH₄, however activity was strongly dependant on the oxidation state of Pd. Through the electropositive nature of Gd, Pd was maintained in metallic state under higher oxygen concentrations compared to the Gd-free catalysts. Characterization through XPS and in situ DRIFTS studies indicate that in the reduction of NO₂, even under excess oxygen, a significant portion of the surface Pd is maintained in the active metallic state.

Full text of DOI:10.1016/j.molcata.2006.06.024

Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
Catalytic reduction of N2O and NO2 with methane
over sol–gel palladium-based catalysts
Nuray Oktar1, Junko Mitome, Erik M. Holmgreen, Umit S. Ozkan∗
Department of Chemical and Biomolecular Engineering, The Ohio State University, 140 W. 19th Avenue, Columbus, OH 43210, USA
Received 3 May 2006; received in revised form 8 June 2006; accepted 9 June 2006
Available online 24 July 2006
Abstract
Pd/TiO2 and Gd-doped Pd/TiO2 prepared through a ‘one-pot’ sol–gel method are shown to be active for the reduction of both N2O and NO2
using CH4 as a reducing agent under excess O2 conditions. In previous studies these catalysts were shown to effectively reduce NO with CH4,
however activity was strongly dependant on the oxidation state of Pd. Through the electropositive nature of Gd, Pd was maintained in metallic
state under higher oxygen concentrations compared to the Gd-free catalysts. Characterization through XPS and in situ DRIFTS studies indicate
that in the reduction of NO2, even under excess oxygen, a significant portion of the surface Pd is maintained in the active metallic state.
© 2006 Elsevier B.V. All rights reserved.
Keywords: N2O; NO2 reduction; Palladium; Gadolinium; Methane
1. Introduction
Pd–Ce–H–ZSM-5 [4–6]. Vannice and co-workers [7,8] have
studied a group of non-zeolitic catalysts based on La2O3, CeO2,
Nd2O3, Sn2O3, Tm2O3, and Lu2O3. Nobel metals such as Pd,
Pt, and Rh supported on silica and/or alumina were also inves-
tigated for NO–CH4 reaction with and without the presence of
oxygen [9–10]. A few researchers have investigated the use of
titania supported palladium catalysts for the reduction of NO.
Ueda et al. [11] studied the reduction of NO using hydrogen in
the presence of oxygen and water vapor.
We have previously reported the effective use of Pd-based
catalysts supported on titania for the NO reduction with CH4 in
the presence of O2 [12–22]. The extent of the reaction was shown
to be highly dependent on the palladium oxidation state, with the
metallic phase of palladium necessary for the reduction of NO to
N2. We extensively studied the mechanistic aspects of NO/CH4
reactions using isotopic labeling techniques under both steady-
state and transient conditions. From a series of unsteady-state
and steady-state isotopic labeling studies using labeled species
such as 15N16O, 15N18O, 13CH4, 18O2, it was concluded that N2
is formed through direct participation of CH4, possibly through
a methyl–nitrosyl type intermediate, whereas N2O formation
was mainly a result of the NO decomposition reaction [13]. We
also used steady-state oscillations as a probe to gain insight into
the reaction network and found that the oxidation state of the
active metal was responsible for three competing reactions; NO
reduction with CH4, direct CH4 oxidation, and NO decomposi-
As a significant source of pollution from combustion sources,
nitrogen oxides (NOx) have been the focus of much regulatory
and research attention. Various NOx species are known to con-
tribute to the greenhouse effect, acid rain formation, and to play
a key role in initiating the formation of ground level smog and
ozone. The best current NOx control technology for stationary
sources is selective catalytic reduction (SCR) with NH3. Prob-
lems associated with the current SCR technology, particularly
the storage and safety issues associated with NH3 use, have
raised interest in the development of catalysts capable of using
alternative reducing agents. Methane is an especially attractive
candidate because of its relative high abundance and low cost.
The principle challenges to methane use are difficulty in acti-
vating it, and selectively using it as a NOx reduction agent in
oxygen rich streams.
Since the early 1990s many different catalyst formulations
have been investigated for the catalytic reduction of NO with
methane in the presence of oxygen, including metal exchanged
zeolitic catalysts such as Co/ZSM-5 [1–3], Pd–H–ZSM-5 and
∗
Corresponding author. Tel.: +1 614 292 6623; fax: +1 614 292 3769.
E-mail address: ozkan.1@osu.edu (U.S. Ozkan).
Present address: Department of Chemical Engineering, Hacettepe Univer-
1
sity, Ankara, Turkey.
1381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.06.024

172
N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
tion [14]. In situ DRIFTS was used to identify adsorbed species
under NO, NO + CH4, and NO + CH4 + O2 flow, and the forma-
tion of various nitrogen-oxo adspecies such as bridged/bidentate
nitrate, monodentate nitrate, nitro, and linear NO was confirmed.
It was found that linearly adsorbed NO species on Pd, with a
IR absorption band at 1780 cm−1, was dominant on reduced
Gd–Pd/TiO2 at high temperatures under NO flow. However, the
intensity of nitrate species was increased on an oxidized surface
relative to the linear NO. Based on these results, the key surface
species under reaction conditions were determined to be CHx,
NHx, monodentate nitrate, nitro, and linear NO species [20–22].
Other groups have also shown in situ FTIR to be a powerful
technique in the elucidation of reaction mechanisms [23–26].
Over Co-FER, Li and Armor [24] used DRIFTS techniques
to propose a possible mechanism for the NO–CH4–O2 reac-
tion. According to their proposed mechanism, formation of
bound NO2 on Co2+ sites is the first step, which is fo•llowed
by CH4 activation by the bound NO2, forming CH3 radi-
used instead of the Gd solution. After hydrolysis the resulting
gel was dried overnight and then calcined at 500 ◦C for 4 h in
air.
2.2. Catalyst characterization
Surface characterization techniques used in this study
included X-ray photoelectron spectroscopy (XPS), temperature
programmed desorption (TPD), and diffuse reflectance infrared
Fourier transform spectroscopy (DRIFTS).
The controlled-atmosphere X-ray photoelectron spectra of
post reaction samples were obtained by an ESCALAB MKII
ESCA/Auger spectrometer, with a Mg K␣ (hν = 1253.6 eV)
source operated at 15 kV and 20 mA. The charge shift was
corrected based upon the standard binding energy of C 1s
(284.5 eV).
Temperature programmed desorption (TPD) experiments
wereconductedusingalaboratory-madeTPR/TPDsystem. Des-
orbing species were detected by an HP5890GC-MS. 75 mg of
sample was loaded into a U-shaped quartz reactor and calcined
in-situ with 10% oxygen in balance helium at 500 ◦C for 30 min,
followed by flushing in helium for 30 min. Samples were then
reduced under 33% hydrogen in balance helium at 200 ◦C for
30 min, followed again by flushing for 30 min in helium. The
sample was then cooled to room temperature under helium flow.
Adsorption was performed at room temperature for 1 h followed
by helium flushing for 1 h to remove physically adsorbed gas.
Diffuse reflectance infrared Fourier transform spectroscopy
(DRIFTS) experiments were performed using a Bruker IFS66
equipped with a DTGS detector and a KBr beamsplitter. Cata-
lysts were placed in a heated sample cup inside a Spectratech
diffuse reflectance cell equipped with KBr windows. An inter-
nal thermocouple mount allowed for the direct measurement of
the catalyst surface temperature. Each catalyst was pre-treated
in situ either by calcination under 10% oxygen at 400 ◦C or by
reduction under 30% hydrogen in balance helium at 200 ◦C for
30 min, followed by cooling under He. For sequential adsorp-
tion studies performed at 300 ◦C, background spectra were taken
at 300 ◦C under helium before gas species were introduced.
For adsorption spectra obtained at several different tempera-
ture levels, background spectra were taken under helium at each
investigated temperature prior to the introduction of gas species.
Each spectrum was averaged over 1000 scans in the mid-IR
range (400–4000 cm−1) to a nominal 2 cm−1 resolution.
•
cals. The CH3 radical could react with bound NO2 to form
nitromethane (CH3NO2), which further reacts with gas phase
NO to N2, H2O, and CO2 (23). Lobree et al. observed Co2+–CN
and Al3+–NCO species during the NO reduction with CH4.
The reaction rate of CN species was determined to be faster
with NO2 than with O2 and NO. Possible mechanistic steps
include the formation of surfacenitrosomethane(CH3NO). Over
Mn–ZSM-5 catalyst, Mn2+ is slowly oxidized to Mn3+ by NO
adsorption and simultaneously, Mn2+(NO) and Mn3+(O−)(NO)
species are formed. Again, CN species were formed and shown
to react rapidly with NO2 (25). Other researchers propose that
methyl species combine directly with adsorbed NO2 to form
nitromethane (CH3NO2). A mechanistic study of NO–CH4–O2
reaction over Co–ZSM-5 was also conducted by Sun et al. [26]
using DRIFTS. The presence of O2 enhanced NO oxidation to
form adsorbed NO2. The authors suggested that the adsorbed
NO2 could react with methane to generate nitromethane inter-
mediates.
Here we focus on the activity of Pd/TiO2 and Gd–Pd/TiO2
catalysts toward the reduction of N2O and NO2 with CH4 in the
presence of O2. Additionally, characterization of the adsorption
properties of these catalysts studied through TPD and in situ
DRIFTS is presented.
2. Experimental
2.1. Catalyst preparation
2.3. Reaction studies
The catalysts examined in this study were 2%Pd/TiO2 and
1%Gd/2%Pd/TiO2, prepared by a modified ‘one-pot’ sol–gel
method. The precursors utilized were palladium acetate, tita-
nium(IV) isopropoxide, and gadolinium nitrate, all obtained
from Aldrich.
In both catalyst preparations, palladium acetate was dissolved
in isopropyl alcohol and titanium(IV) isopropoxide was added
to the palladium solution under continuous stirring at room tem-
perature. For the bimetallic catalyst a solution of gadolinium
nitrate in water was added drop-wise with a syringe pump to
this suspension. For the Pd-only catalyst, distilled water was
A 1/4 in OD stainless steel reactor was used to perform steady
state reaction experiments. All gas lines were heated to 130 ◦C to
prevent water condensation in the system. In all experiments, a
catalyst sample of 69 mg was placed inside the reactor and held
in place by quartz wool packing. Reaction temperatures were
varied from 300 to 500 ◦C, under control of a PID temperature
controller (Omega).
Prior to each run the catalyst bed was reduced under 33%
H2 in He at 200 ◦C for 30 min. Reaction feed concentrations
were 1.065% methane, 500 ppm NO2 or N2O, and O2 to reach

N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
173
Table 1
Steady state reaction results for the reduction of N2O with CH4 over 1%Gd/2%Pd/TiO2 and 2%Pd/TiO2
Catalyst
O2/CH4 = 1.88
O2/CH4 = 2.35
N2O conversion (%)
N2 selectivity (%)
CH4 conversion (%)
N2O conversion (%)
N2 selectivity (%)
CH4 conversion (%)
1%Gd/2%Pd/TiO2
2%Pd/TiO2
100
100
100
98
94
84
55
7
100
97
95
93
The reaction was performed under O2 deficient and excess conditions at 500 ◦C.
Table 2
an O2/CH4 ratio of 1.88 or 2.35%. Helium (minimum purity
99.995%) was used as the balance gas. Steady-state was gen-
erally reached within 1 h, however typical reaction times were
4 h. Both reactants and reaction products were quantified by
a HP 5890A gas chromatograph equipped with a molecu-
lar sieve and a pora-pak column operating at 80 ◦C (nitrous
oxide, nitrogen, oxygen, methane, carbon monoxide, and carbon
dioxide), by an IR ammonia analyzer (Siemens Ultramat 5F),
and by a chemiluminescence NO–NO2–NOx analyzer (Thermo
Environmental Instruments, Model 42H). Nitrogen and car-
bon balances were always close to 100%. For this work NO2
and N2O conversions were defined as (NOx consumed)/(NOx
Fed). N2 selectivity was calculated as (N2 produced)/(NOx con-
sumed) for N2O reactions, and twice this number for NO2
reactions.
Effect of temperature on the N2O reduction with CH4 over 1%Gd/2%Pd/TiO2
and 2%Pd/TiO2 catalysts
350 ◦C (%) 400 ◦C (%) 450 ◦C (%) 500 ◦C (%)
1%Gd/2%Pd/TiO2
N2O conversion 19
N2 selectivity
CH4 conversion
25
97
51
68
99
92
100
99
94
99
27
2%Pd/TiO2
N2O conversion
N2 selectivity
CH4 conversion
4
97
18
10
98
38
84
98
75
100
98
84
tivity was nearly 100% for both catalysts across the temperature
range. CH4 conversion was observed to increase with temper-
ature over both catalysts. At a given reaction temperature the
conversion was higher over the bimetallic catalyst, which cor-
responded to higher N2 yields.
3. Results and discussion
3.1. Reduction of N2O with CH4
Reduction of N2O over Pd/TiO2 and Gd-Pd/TiO2 catalysts
was examined at O2/CH4 concentration ratios of 1.88 and 2.35.
The N2O conversions and N2 selectivities obtained at 500 ◦C
are presented in Table 1. At the oxygen-deficient conditions,
complete N2O conversion is achieved over both catalysts with
N2 selectivities approaching 100%. The only other N-containing
product besides N2 was NH3 and was observed only on Pd/TiO2
catalyst. When the O2 concentration was increased, there was
a decrease in N2O conversion over both catalysts, but the drop
was much more significant over the Pd-only catalyst. CH4 con-
versions in these experiments was high and was observed to
increase with increasing O2 concentration.
The N2O reduction with CH4 was also investigated at differ-
ent temperatures and an O2/CH4 ratio of 1.88 (Table 2). In these
experiments, the only nitrogen containing products observed
were nitrogen and ammonia. N2O conversion increased with
temperature in the range of 350–500 ◦C, and was slightly higher
over the bimetallic catalyst at low temperature. By 500 ◦C, com-
plete N2O conversion was obtained over both catalysts. Selec-
3.2. Reduction of NO2 with CH4
Table 3 shows the reduction of NO2 with CH4 at 500 ◦C over
Pd/TiO2 and Gd–Pd/TiO2. NO2 conversion was above 95% for
both oxygen deficient and excess conditions. At an O2/CH4 ratio
of 1.88 N2 selectivity on the Pd-only catalyst was nearly 100%,
but decreased in excess oxygen reaching only 55%. Correspond-
ing to the lowered selectivity, the other nitrogen-containing
product was NO. In comparison, the Gd–Pd/TiO2 catalyst did
not suffer from a significant N2 selectivity drop under excess
oxygen, where N2 selectivity remained above 90%. CH4 con-
version over the Gd/Pd catalyst was above 95% for both O2
feeds. Over the Pd-only catalyst CH4 conversion increased with
the addition of more O2.
The effect of temperature on the reduction of NO2 at an
O2/CH4 ratio of 1.88 over Gd–Pd and Pd catalysts are shown
in Table 4. NO2 conversion was over 95% at all temperatures
between 350 and 500 ◦C over Gd–Pd, but near 100% N2 selec-
tivity was achieved only at 500 ◦C. At the lower temperatures N2
Table 3
Steady state reaction results for the reduction of NO2 with CH4 over 1%Gd/2%Pd/TiO2 and 2%Pd/TiO2
Catalyst
O2/CH4 = 1.88
O2/CH4 = 2.35
NO2 conversion (%)
N2 selectivity (%)
CH4 conversion (%)
NO2 conversion (%)
N2 selectivity (%)
CH4 conversion (%)
1%Gd/2%Pd/TiO2
2%Pd/TiO2
100
100
94
98
97
76
100
97
91
54
97
95
The reaction was performed under O2 deficient and excess conditions at 500 ◦C.

174
N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
Table 4
gests that, in this case, decomposition of NO2 may be more
prominent than its direct reduction.
Effect of temperature on the NO2 reduction with CH4 over 1%Gd/2%Pd/TiO2
and 2%Pd/TiO2 catalysts
350 ◦C (%) 400 ◦C (%) 450 ◦C (%) 500 ◦C (%)
3.3. Post-reaction XPS
1%Gd/2%Pd/TiO2
Analysis of the Pd oxidation state over post-reaction catalysts
was performed by XPS. After being kept on-stream at 500 ◦C
for 20 h in the NO2 + CH4 + O2 reaction with excess oxygen,
the catalyst was transferred from the reactor to the XPS analysis
chamber without exposure to the atmosphere. In the XPS results
two Pd oxidation states are visible. The zero oxidation state
was observed by Pd5/2 and Pd3/2 peaks at 334.7 and 340.0 eV,
respectively. The +2 oxidation state was assigned to peaks at
336.9 and 342.0 eV. Deconvolution of the peaks showed that
40% was present as Pd0 and 60% as Pd2+. It is important to
observe that some Pd was maintained in the reduced state even
under excess oxygen reaction conditions, corresponding to the
observed reaction activity. For comparison XPS results for the
reduced Gd–Pd/TiO2 showed that 84% of the Pd is present in
the metallic phase.
NO2 conversion 95
96
51
30
99
67
76
100
97
97
N2 selectivity
CH4 conversion
53
15
2%Pd/TiO2
NO2 conversion 90
90
57
43
92
61
55
97
54
76
N2 selectivity
CH4 conversion
50
20
selectivity was between 50 and 60%, with the other N-containing
product being NO. NO2 conversions over the Pd only cata-
lysts were somewhat lower, but still remained above 90% at all
temperatures tested. N2 selectivities however, were low in the
entire temperature range, with the highest N2 selectivity being
around 60%. Over both catalysts CH4 conversion increased with
reaction temperature, with higher conversions reached over the
bimetallic catalyst.
3.4. TPD experiments
In our previous work, we reported the strong dependence of
NO–CH4 reaction on the oxidation state of Pd, where metallic Pd
was shown to be necessary for the selective NO reduction with
CH4 whiles Pd2+ led to combustion of the hydrocarbon [17].
Using the oscillatory behavior of the CH4 + O2 system over Pd
catalysts, we have also shown that Pd can exist in either the
metallic or the oxide phase depending on the temperature and
oxygen partial pressure (14). At any given oxygen concentra-
tion, decreasing the temperature results in the thermodynamic
equilibrium shifting toward the oxide phase. In the NO2 + CH4
experiments, it is possible that, a larger fraction of Pd exists
in the Pd+2 state, especially at lower temperatures. This would
explainthelowerN2 formation, whichisdependentonthemetal-
lic Pd sites. The high NO2 conversions, on the other hand, can
be explained by the conversion of NO2 to NO and O2 through a
decomposition reaction.
Bamwenda and co-workers [45] studied NO2 decomposition
to NO and 1/2O2 during NOx reduction with C3H6 over Rh, Pt,
and Au catalysts supported on alumina. They found that both Rh
and Pt were active in converting NO2 to NO in the presence of
oxygen over a temperature range of 200–500 ◦C, approaching
the thermodynamic equilibrium above 400 ◦C. Since Pd is also a
precious metal in the same group as Rh and Pt, it is possible that
formation of NO through decomposition of NO2 took place dur-
ing NO2–CH4–O2 reaction. Then, NO could be further reduced
to N2 by methane, which is activated on Pd as reported in our
previous studies [17]. It is possible that overall NO formation
was lower on the Gd-containing catalyst because the rate of reac-
tion for NO reduction with activated methane is faster than the
decompositionrate. ThefactthatonlytheGd-containingcatalyst
is able to reduce NO is consistent with our previous results that
this catalyst can maintain Pd in metallic state at higher oxygen
concentrations compared to Gd-free catalyst. Direct reduction
of NO2 with CH4 is also a possibility, however, lower N2 yields
obtained from NO2 compared to those obtained from NO sug-
3.4.1. N2O TPD
Temperature-programmed desorption of N2O was performed
on both reduced and oxidized catalyst samples. Mass to charge
ratios of m/e = 46 (NO2), 44 (N2O), and 30 (NO) are shown.
Since m/e = 44 corresponds to CO2 as well as N2O, m/e = 12
corresponding to atomic carbon was also monitored. In all
experiments this signal was negligible, indicating that m/e = 44
was due only to N2O desorption. Fig. 1a shows the TPD pro-
files obtained after adsorption of N2O over the reduced Gd–Pd
bimetallic catalyst. Over this sample, there were N2O (m/e = 44)
peaks at 108, 200, and a broad desorption feature with max-
ima at 414, 552, and 590 ◦C. We also observed a small NO
(m/e = 30) desorption peak at around 302 ◦C, indicating that the
catalyst converted N2O to NO. N2O TPD from oxidized Gd–Pd
sample resulted in desorption features at 100, 180, 358, 420,
and 665 ◦C, as shown in Fig. 1b. The intensity of the low tem-
perature feature is much larger over the oxidized sample. NO
desorption was again observed, though the peak temperature was
lowered to 210 ◦C. NO2 desorption was not observed over either
sample.
N2O desorption profiles from Pd/TiO2 showed similar behav-
ior to the bimetallic catalysts. Over the reduced Pd/TiO2 shown
in Fig. 2a, desorption features were observed in a broad range
with maxima at 384, 458, and 485 ◦C. In comparison, the oxi-
dized catalyst (Fig. 2b) showed more prominent desorption
features at low temperatures, with peaks at 100, 311, 360, and
705 ◦C. Significant differences in the desorption of NO were
observed in comparison to the Gd–Pd samples. Over the reduced
sample a large NO desorption peak was observed at 197 ◦C.
This desorption occurs at lower temperature and is of larger
intensity than over the reduced bimetallic catalyst. Over the oxi-
dized Pd/TiO2 shown in Fig. 2b NO desorption was observed at
240 ◦C, slightly higher temperature than on the bimetallic sam-
ple. NO2 was not observed in either experiment.

N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
175
Fig. 1. N2O TPD profiles of 1%Gd/2%Pd/TiO2 (a) reduced (b) oxidized. m/e
ratios of 30(ꢀ), 44 (ꢁ), 46(᭹) correspond to NO, N2O, NO2, respectively.
Fig. 2. N2O TPD profiles of 2%Pd/TiO2 (a) reduced (b) oxidized. m/e ratios of
30(ꢀ), 44 (ꢁ), 46(᭹) correspond to NO, N2O, NO2.
3.4.2. NO2 TPD
NO2 TPD experiments were performed on Gd–Pd and Pd in
both the reduced and oxidized states. In examining the results of
NO2 TPD it is important to note that m/e = 30 (NO) is the princi-
ple ionization fragment of NO2. Over the reduced Gd-Pd/TiO2
catalyst shown in Fig. 3a, NO2 desorption peaks were observed
at 120, 249, and 324 ◦C. The profile for m/e = 30 matches that for
m/e = 46 closely, indicating that it is due solely to the fragmen-
tation of NO2. A small amount of m/e = 44 (N2O) is observed in
a broad desorption feature between 200–350 ◦C. Fig. 3b shows
the desorption profiles of the oxidized Gd–Pd sample. Desorp-
tion peaks for NO were seen at 117 and 301 ◦C. The peak at
117 ◦C was not accompanied by a corresponding m/e = 46 peak,
suggesting that it is NO desorption. The feature at 301 ◦C is due
to fragmentation of NO2 as the profiles match closely. The des-
orption of N2O was also observed and coincided with the NO2
feature at 301 ◦C. These results indicate that NO2 decomposi-
tiontookplaceovertheoxidizedcatalyst. NO2 desorptionresults
over reduced and oxidized Pd catalysts in Fig. 4a and b, respec-
tively, were nearly identical to their Gd–Pd counterparts. Peaks
were observed in similar locations as well as the same trends
in NOx species production. On the reduced Pd catalyst the two
main desorption features were slightly closer in temperature,
149 and 251 ◦C. Observed NO again seems to be due to NO2
fragmentation, and only a small amount of N2O is observed. On
the calcined Pd sample a shoulder at 367 ◦C is observed on the
main desorption peak, but otherwise peak shape and location are
very similar to the bimetallic sample. This oxidized sample also

176
N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
Fig. 3. NO2 TPD profiles of 1%Gd/2%Pd/TiO2 (a) reduced (b) oxidized. m/e
ratios of 30(ꢀ), 44 (ꢁ), 46(᭹) correspond to NO, N2O, NO2.
Fig. 4. NO2 TPD profiles of 2%Pd/TiO2 (a) reduced (b) oxidized. m/e ratios of
30 (ꢀ), 44(ꢁ), 46(᭹) correspond to NO, N2O, NO2.
appears to be active for NO2 decomposition as evidenced by NO
desorption at 117 ◦C and N2O seen near 315 ◦C. The NO2 TPD
results support our assertion that NO2 decomposition is more
likely to take place over the oxidized catalyst and help explain
the trends seen in the reaction experiments.
concentrations were N2O = 1780 ppm, CH4 = 2.13%, O2 = 0.3
and 4.5%, respectively. Spectra were taken under the gas flows,
after the catalysts had been exposed to each gas mixture for
30 min at 300 C surface temperature.
Fig. 5 shows spectra taken over the reduced Gd–Pd cata-
lyst. Under N2O flow only, we observed bands at 2200–2240,
2333, 1499, and 1347 cm−1. The doublet at 2200–2240 cm−1
could correspond to gas phase N2O, but is also indicative
of adsorbed N2O. Wood and co-workers [27] examined N2O
adsorption on zeolitic systems and reported bands in the
region 2226–2282 cm−1 associated with N–N stretching, and
in 1344–1308 cm−1 due to N–O stretching. They have reported
3.5. DRIFTS experiments
3.5.1. N2O → N2O + CH4 → N2O + CH4 + O2 → He flush
DRIFT spectra were taken over both reduced and oxidized
Gd–Pd/TiO2 catalysts following exposure to feeds of N2O, N2O
+ CH4, N2O + CH4 + O2 (low O2 concentration), and N2O +
CH4 + O2 (high O2 concentration), performed in sequence. Gas

N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
177
The N2O–CH4–O2 reaction data on reduced Gd–Pd catalyst at
300 ◦C showed 100% selectivity to N2, which is in agreement
with our observation in the IR spectrum. These trends continue
with the change to excess oxygen by increasing the oxygen con-
centration from 0.3 to 4.5%. With the introduction of excess
oxygen conditions, the intensity of the N2O peaks at 2199 and
2235 cm−1 further increased, again due to a loss of available
CH4 to the reduction reaction.
The final gas change of this experiment had the reaction feed
turned off and the catalyst flushed under He. After flushing,
the gas phase N2O and CH4 are no longer visible. Adsorbed
+
dinitrogen species at 2333 cm−1 and the surface NH4 groups
observed at 1499 cm−1 remain stable on the surface at 300 ◦C.
The nitrate species observed at 1347 cm−1 also remains on the
surface after flushing.
The same DRIFTS experiment was performed over the oxi-
dized Gd–Pd catalyst, which was inactive for NOx reduction.
Fig. 6 shows that under a flow of 1780 ppm N2O the only
observed bands were the doublet of gas phase N2O at 2199
and 2235 cm−1. With the addition of CH4 to the feed, gas phase
CH4 was seen through bands at 3016 and 1306 cm−1. The forma-
tion of new surface species with absorption bands at 1430, 1520,
1610, 1875, and a doublet around 2350 cm−1 was also observed.
The formation of the bands at 1430 and 1520 cm−1 was observed
only on the oxidized surface. The latter species was due to the
formation monodentate nitrate [31–34], while the former species
could arise either from CHx or nitrogen-oxo species. The forma-
Fig. 5. DRIFT spectra taken during the sequential introduction of reac-
tants at 300 ◦C to reduced 1%Gd/2%Pd/TiO2. (a) N2O, (b) N2O + CH4, (c)
N2O + CH4 + O2 (low), (d) N2O + CH4 + O2 (excess) and (e) He flush.
that the surface acidity plays a strong role in N2O adsorption.
The band at 2333 cm−1 is in the range reported to be adsorbed
dinitrogen [28], which suggests N2O decomposition to N2 on
the surface. The band at 1499 cm−1 was assigned to NH4
+
as
[29], likely the result of interaction with Brønsted acid sites on
the surface. It is also possible that chemisorbed hydrogen from
the reduction step remained on the surface after treatment and
interacted with N2O.
With the introduction of CH4 to the N2O flow, gas phase
CH4 is confirmed by the presence of the bands at 3016 and
1306 cm−1. A band at 1875 cm−1 was observed due to the for-
mationofNO, eitheradsorbedonmetal(Pd)cations[30]orinthe
gas phase. Evidence of the N2O reduction with CH4 is observed
through several changes in surface species. The intensity of the
N2O doublet at 2199 and 2235 cm−1 decreased upon the addi-
tion of CH4. Simultaneously the band at 2333 cm−1 assigned to
adsorbed dinitrogen increases. The formation of H2O, a product
of the reduction reaction, was also observed through a small peak
at 1610 cm−1. The doublet in the region of 2350 cm−1 charac-
teristic of CO2 may not be visible due to the large dinitrogen
band.
When0.3%O2 wasaddedtotheN2O + CH4 flowtheintensity
of N2O species increased in comparison to the O2 free experi-
ment. This would suggest that there is less CH4 available to react
with N2O species due to the competitive combustion reaction.
The NO band at 1875 cm−1 disappeared upon the addition of
Fig. 6. DRIFT spectra taken during the sequential introduction of reac-
tants at 300 ◦C to oxidized 1%Gd/2%Pd/TiO2. (a) N2O, (b) N2O + CH4, (c)
N2O + CH4 + O2 (low), (d) N2O + CH4 + O2 (excess) and (e) He flush.
+
O2, and an increase in the NH4 band at 1499 cm−1 was seen.

178
N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
tion of nitrate in the presence of CH4 was surprising considering
that we did not observe nitrate formation under N2O-only flow.
However we could speculate that NO, observed at 1875 cm−1
,
may be converted to nitrate species over the oxidized surface.
TPD profiles following N2O adsorption on the oxidized Gd-Pd
catalyst showed a NO desorption peak at 210 ◦C, which is lower
than that observed over the reduced catalyst by 80 ◦C. The lower
desorption temperature may make conversion of NO to surface
nitrate species possible. The doublet at 2350 cm−1 and the band
at 1610 cm−1 indicate the presence of CO2 and H2O on the
surface, respectively. The surface dinitrogen species observed
on the reduced Gd–Pd sample is not present, which is consis-
tent with the lack of reduction activity on the oxidized catalyst.
With the addition of O2 to the gas feed the CO2 and H2O band
increase in intensity, due to CH4 combustion. The other bands
remain fairly constant. After the final flushing step all of the gas
phase species disappear, as well as the bands of CO2 and H2O.
Only the bands at 1430 and 1520 cm−1 remain.
3.5.2. NO2 adsorption
The spectra in Figs. 7 and 8 were taken under a flow
of 1780 ppm NO2 at 300 ◦C over oxidized and reduced
Gd–Pd/TiO2 catalysts, respectively. Changes in surface species
were monitored with time up to 35 min. Over the oxidized sam-
ple, bands at 1612, 1577, and 1231 cm−1 were formed immedi-
ately upon introduction of NO2. The IR band at 1612 cm−1 is due
tothepresenceofbridgednitratespecies, andthepairat1578and
1231 cm−1 was assigned to bidentate nitrate species [31,33,35].
Fig. 8. DRIFT spectra of reduced 1%Gd/2%Pd, taken as a function of time
under a flow of 1780 ppm NO2, at 300 ◦C.
Over the first 15 min of exposure, these bands increased in inten-
sity and thereafter remained constant. The species observed at
1351 cm−1 lags behind in forming on the surface, but increases
in intensity with exposure time. Vibrational frequencies in this
region are typical of nitro (NO2−) species [33,36].
On the reduced Gd–Pd catalyst, the spectrum taken after
1 min of exposure showed bands at 1605 and 1191 cm−1. A band
in the region of 1605 cm−1 is due to bridged nitrate species as
on the oxidized sample, though in comparison, the intensity was
much smaller over the reduced sample. The band at 1191 cm−1
corresponds to molecularly adsorbed NH3 on Lewis acid sites.
The ammonia formation could be due to a reaction between
NO2 and chemisorbed hydrogen that was left on the surface
after the in situ reduction step. This band decreases with time
as the adsorbed NH3 is removed from the surface. Between 25
and 35 min, the development of bridged nitrate (1605 cm−1),
bidentate nitrate (1577 cm−1), and nitro species (1352 cm−1) are
observed. The growth of these bands corresponds to the appear-
ance and removal of linear adsorbed NO on Pd+ characterized
by a band at 1793 cm−1. Linear NO adsorption on Pd has been
studied on single crystals as well as supported catalysts [37–43].
Previously, we concluded that the 1793 cm−1 band indicated
linear NO adsorption on partially reduced Pd sites (Pd+) [22].
This evolution of surface species results in spectra similar to
those observed over the oxidized catalyst, which indicated that
the catalyst surface was oxidized as the duration of NO2 expo-
sure increased, perhaps making more lattice oxygen available for
bridged and bidentate nitrate formation. Reduction in the inten-
Fig. 7. DRIFT spectra of oxidized 1%Gd/2%Pd, taken as a function of time
under a flow of 1780 ppm NO2, at 300 ◦C.

N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
179
sity of Pd–NO suggests that reduced Pd sites are less available
due to complete oxidation of Pd in the presence of NO2.
3.5.3. NO2 → NO2 + CH4 → NO2 + CH4 + O2 adsorption
DRIFT spectra were taken over oxidized and reduced Gd-
Pd/TiO2 catalysts following exposure to NO2, NO2 + CH4,
NO2 + CH4 + O2 (low O2 concentration), and NO2 + CH4 + O2
(high O2 concentration) flows in sequence. Feed concentrations
were NO2 = 1780 ppm, CH4 = 2.13%, O2 = 0.3 and 4.5%. Spec-
tra were taken after the catalysts were exposed to each gas
mixture for 30 min at a surface temperature of 300 ◦C.
The high and low wavenumber regions for the oxidized sam-
ple are shown in Figs. 9 and 10. Upon the introduction of NO2,
bands at 1605, 1577, and 1222 cm−1 were observed, correspond-
ing to bridged and bidentate nitrates as in the previous exper-
iment. A surface nitro species at 1349 cm−1 is also observed.
With the addition of CH4 to the gas feed the bridged and biden-
tate species decrease in intensity while an increase in the nitro
band occurs. Concurrently, bands at 3016 and 1304 cm−1 are
observed due to gas phase CH4. Multiplets in the 3200 to
3400 cm−1 region and a band at 1191 cm−1 also appeared. The
multiplets corresponded to N–H stretching of coordinatively
adsorbed NH3, and the 1191 cm−1 band is a characteristic fre-
quency of NH3sy coordinated to Lewis acid sites. The formation
of NHx species suggested that activation of CH4 was achieved
on the oxidized catalyst, possibly through hydrogen abstraction
at bridged/bidentate nitrate sites. Previously, for the NO-CH4-
Fig. 10. DRIFT spectra taken during the sequential introduction of reac-
tants at 300 ◦C to oxidized 1%Gd/2%Pd/TiO2. (a) NO2, (b) NO2 + CH4, (c)
NO2 + CH4 + O2 (low) and (d) NO2 + CH4 + O2 (excess). High wavenumber
region.
O2 reaction over the oxidized Gd-Pd/TiO2 catalyst we did not
observe the formation of NHx species, indicating that activation
of CH4 was not possible on the oxidized catalyst after exposure
to NO [22]. The formation of bridged/bidentate nitrate was sig-
nificantly less after NO exposure than from NO2 exposure on the
catalyst. ThisseemstosupportoursuppositionthatCH4 couldbe
activated on NO2 covered Gd–Pd/TiO2 to form CHx and dissoci-
ated hydrogen, which further react with nitrogen-oxo species on
the surface to form NHx. CH4 activation through adsorbed NO2
has been reported in the literature [23–26,44]. Over Co-FER,
Li and Armor proposed a mechanism that involves the forma-
tion of adsorbed NO2 on Co2+ sites as the first•step, followed
by CH4 activation by the adsorbed NO2 to CH3 radicals. The
CH3• radical could react with bound NO2 to form nitromethane
(CH3NO2), which further can react with gas phase NO to N2,
H2O, and C•O2 (23). Hall and co-workers investigated the role
of free CH3 radicals over Co–ZSM-5 and H–ZSM-5 catalysts
•
and proposed that after the oxidation of NO to NO2, CH3 free
radicals, formed by interaction with NO2, react with NOx and
O2 [20].
When low concentrations of O2 were introduced (3000 ppm)
to the NO2 + CH4 flow, disappearance of NHx species was
observed, suggesting that the species is easily oxidized in the
presence of oxygen. This corresponded to a significant increase
in the intensity of nitro species at 1349 cm−1. Adsorbed CO2
Fig. 9. DRIFT spectra taken during the sequential introduction of reac-
tants at 300 ◦C to oxidized 1%Gd/2%Pd/TiO2. (a) NO2, (b) NO2 + CH4, (c)
NO2 + CH4 + O2 (low), (d) NO2 + CH4 + O2 (excess). Low wavenumber region.
was also observed through a doublet centered near 2350 cm−1
.

180
N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
Fig. 11. DRIFT spectra taken during the sequential introduction of reac-
tants at 300 ◦C to reduced 1%Gd/2%Pd/TiO2. (a) NO2, (b) NO2 + CH4, (c)
NO2 + CH4 + O2 (low) and (d) NO2 + CH4 + O2 (excess). Low wavenumber
region.
Fig. 12. DRIFT spectra taken during the sequential introduction of reactants at
300 ◦Ctoreduced1%Gd/2%Pd/TiO2. (a)NO2, (b)NO2 + CH4, (c)NO2 + CH4 +
O2 (low) and (d) NO2 + CH4 + O2 (excess). High wavenumber region.
Figs. 11 and 12 present the results from the sequential intro-
duction performed on the reduced catalyst. Under NO2 flow
only, the intensity of bridged nitrate and nitro species were sim-
ilar to the oxidized sample. Additionally only a small band
at 1789 cm−1 associated with Pd–NO was seen, suggesting
that most of the Pd sites were oxidized. When methane was
introduced we again observed the formation of NHx species
at 3200–3400 cm−1 and 1185 cm−1, the appearance of CH4 at
3016 and 1304 cm−1, and an intensity decrease of bridged nitrate
species. ThisconsistentevidencestrengthensourproposedCH4-
nitrate interaction. Under the low O2 feed the ammonia species
disappear, which coincides with an increase in nitro groups at
1347 cm−1. A strong CO2 band is observed, and these trends
continue with increased O2 concentration.
3.6. In situ DRIFTS during reaction of NO2 + CH4 + O2 at
different temperatures
DRIFTS spectra were taken over reduced Gd–Pd/TiO2 dur-
ing reaction of 1780 ppm NO2, 2.13% CH4, and 3000 ppm
O2 in balance He, at temperatures of 25, 75, 150, 225, and
300 ◦C (Figs. 13 and 14). At low temperatures bridged nitrate
at 1610 cm−1, bidentate nitrate at 1581/1240 cm−1, and mon-
odentate nitrito at 1486 cm−1 were observed. With increasing
reaction temperature a decrease in bidentate nitrate and mon-
odentate nitrito species occurred, along with the formation of
Fig. 13. In situ DRIFT spectra of reduced 1%Gd/2%Pd/TiO2 during reaction.
Low wavenumber region.

N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
181
of NO2 seems to proceed through the interaction of CH4 with
surface bidentate nitrates. However, under reaction conditions
nitro species, not linearly adsorbed NO, are the dominant surface
species.
Acknowledgments
This work was funded in part by the National Science Foun-
dation and the Ohio Coal Development Office.
References
[1] Y. Li, J.N. Armor, Appl. Catal. B: Environ. 1 (1992) L31.
[2] Y. Li, P.J. Battavio, J.N. Armor, J. Catal. 142 (1993) 561.
[3] Y. Li, J.N. Armor, Appl. Catal. B: Environ. 2 (1993) 239.
[4] Y. Nishizaka, M. Misono, Chem. Lett. (1993) 1295.
[5] Y. Nishizaka, M. Misono, Chem. Lett. (1994) 2237.
[6] M. Misono, Y. Hirao, C. Yokoyama, Catal. Today 38 (1997) 157.
[7] X. Zhang, A.B. Walter, M.A. Vannice, Appl. Catal. B: Environ. 4 (1994)
237.
[8] X. Zhang, A.B. Walter, M.A. Vannice, J. Catal. 155 (1995) 290.
[9] R. Burch, A. Ramli, Appl. Catal. B 15 (1998) 49.
[10] R. Burch, A. Ramli, Appl. Catal. B 15 (1998) 63.
[11] A. Ueda, T. Nakao, M. Azuma, T. Kobayashi, Catal. Today 45 (1998)
135.
[12] M.W. Kumthekar, U.S. Ozkan, J. Catal. 171 (1997) 45.
[13] M.W. Kumthekar, U.S. Ozkan, J. Catal. 171 (1997) 54.
[14] M.W. Kumthekar, U.S. Ozkan, J. Catal. 171 (1997) 67.
[15] M.W. Kumthekar, U.S. Ozkan, Catal. Today 35 (1997) 107.
[16] J. Mitome, G. Karakas, K.A. Bryan, U.S. Ozkan, Catal. Today 42 (1998)
3.
Fig. 14. In situ DRIFT spectra of reduced 1%Gd/2%Pd/TiO2 during reaction.
High wavenumber region.
[17] J. Mitome, E. Aceves, U.S. Ozkan, Catal. Today 53 (1999) 597.
[18] A. Rath, E. Aceves, J. Mitome, J. Liu, U.S. Ozkan, S.G. Shore, J. Mol.
Catal. A 165 (2001) 103.
[19] G. Karakas, J. Mitome-Watson, U.S. Ozkan, Catal. Commun. 3 (2002)
199.
[20] J.M. Watson, U.S. Ozkan, J. Catal. 210 (2002) 295.
[21] J.M. Watson, U.S. Ozkan, J. Mol. Catal. A 192 (2003) 79.
[22] J.M. Watson, U.S. Ozkan, J. Catal. 217 (2003) 1.
[23] Y. Li, J.N. Armor, J. Catal. 150 (1994) 388.
monodentate nitrate species with a band at 1524 cm−1 and nitro
species at 1347 cm−1 (28–29). At 300 ◦C all the nitrate species in
the 1450–1625 cm−1 range disappeared and CO2 was observed
around 2350 cm−1. These changes indicate light off of the reduc-
tion reaction, and support the conclusion of CH4 activation
through interaction with surface nitrate species.
[24] L.J. Lobree, A.W. Aylor, J.A. Reimer, A.T. Bell, J. Catal. 169 (1997)
188.
[25] A.W. Aylor, L.J. Lobree, J.A. Reimer, A.T. Bell, J. Catal. 170 (1997)
390.
4. Conclusions
The versatility of Gd–Pd/TiO2 catalyst for reduction of
not only NO but also other nitrogen oxides (N2O and NO2)
with methane has been demonstrated. Consistent with previous
results, the bimetallic Gd–Pd/TiO2 catalyst was shown to main-
tain the Pd sites in metallic state in higher O2 concentrations
compared to the Gd-free catalysts, retaining activity for NOx
reduction. As suggested previously (17, 20, 21, 22) the highly
electropositive nature of Gd may be primary reason for this
phenomenon. Mechanistic studies of the N2O reduction sug-
gest a role for an intermediate NH4+ species resulting in the
appearance of adsorbed dinitrogen. Although DRIFT spectra
show that the catalyst is partially oxidized under a flow of NO2,
under reaction conditions this may not be the case. Post reaction
XPS shows that 40% of Pd remains in the active metallic state
even after exposure to NO2 and in an O2 rich feed. Furthermore
comparisons of in situ DRIFTS results yield very similar spec-
tra when the reaction is performed over both the oxidized and
reduced catalyst. As with the reduction of NO, the reduction
[26] T. Sun, M.D. Fokema, J.Y. Ying, Catal. Today 33 (1997) 251.
[27] B.R. Wood, A.J. Reimer, A. Bell, J. Catal. 209 (2002) 151.
[28] J. Lin, H.Y. Chen, L. Chen, L. Tan, H.C. Zeng, Appl. Surf. Sci. 103 (1996)
307.
[29] K. Shimizu, F. Okada, Y. Nakamura, A. Satsuma, T. Hattori, J. Catal. 195
(2000) 151.
[30] L.J. Lobree, I.C. Hwang, J.A. Reimer, A.T. Bell, J. Catal. 186 (1999) 242.
[31] K. Hadjiivanov, V. Bushev, M. Kantcheva, D. Klissurski, Langmuir 10
(1994) 464.
[32] S.-J. Huang, A.B. Walters, M.A. Vannice, Appl. Catal. B: Environ. 26
(2000) 101.
[33] A.A. Davydov, Infrared Spectroscopy of Adsorbed Species on the Surface
of Transition Metal Oxides, Wiley, Chichester, 1990.
[34] Y. Chi, S.C. Chaung, J. Phys. Chem. 104 (2000) 4673.
[35] D.K. Captain, M.D. Amiridis, J. Catal. 184 (1999) 377.
[36] K. Hadijivanov, H. Knozinger, Phys. Chem. Chem. Phys. 2 (2000)
2803.
[37] T.E. Hoost, K. Otto, K.A. Laframboise, J. Catal. 155 (1995) 303.
[38] X. Xu, P. Chen, D.W. Goodman, J. Phys. Chem. 98 (1994) 9242.
[39] C.M. Grill, R.D. Gonzalez, J. Phys. Chem. 84 (1980) 878.
[40] D.T. Wickham, B.A. Banse, B.E. Koel, Surf. Sci. 243 (1991) 83.

182
N. Oktar et al. / Journal of Molecular Catalysis A: Chemical 259 (2006) 171–182
[41] S. Moriki, Y. Inoue, E. Miyazaki, I. Yasumori, J. Chem. Soc. Faraday Trans.
78 (1982) 171.
[44] D.B. Lukyanov, E.A. Lambardo, G.A. Sill, J.L. d’Itri, W.K. Hall, J. Catal.
163 (1996) 447.
[42] C. Nyberg, P. Uvdal, Surf. Sci. 250 (1991) 42.
[43] K. Almusaiteer, S.S.C. Chuang, J. Catal. 184 (1999) 189.
[45] G. Bamwenda, A. Obuchi, A. Ogata, J. Oi, S. Kushiyama, K. Mizuno, Reac.
Kinet. Catal. Lett. 63 (1998) 53.