Kinetic studies of the stability of Pt for NO oxidation: Effect of sulfur and long-term aging

Article abstract of DOI:10.1016/j.jcat.2011.05.007

The stability of Pt catalysts for NO oxidation was analyzed by observing the effect of pre-adsorbed sulfur on the reaction kinetics using a series of Pt/SBA-15 catalysts with varying Pt particle sizes (ca 2-9 nm). Our results indicate that sulfur addition did not influence catalyst deactivation of any of the Pt catalysts, resulting in unchanged turnover rates (TOR) and reaction kinetics. The presence of sulfur on Pt was confirmed by X-ray absorption fine structure spectroscopy (EXAFS) under reducing environments. However, exposure of the catalyst to NO oxidation conditions displaced sulfur from the first coordination shell of Pt, yielding Pt-O bonds instead. Re-reduction fully recovered the Pt-S backscattering, implying that sulfur remained near the Pt under oxidizing conditions. X-ray photoelectron spectroscopy (XPS) and chemisorption measurements confirmed the presence of sulfur near platinum. The invariance of the NO oxidation reaction to sulfur poisoning is explained by sulfur displacement to interfacial sites and/or sulfur binding on kinetically irrelevant sites. Formation of Pt oxides remains as the main source of catalyst deactivation as observed by kinetic and X-ray absorption spectroscopy (XAS) measurements.

Full text of DOI:10.1016/j.jcat.2011.05.007

Journal of Catalysis 282 (2011) 13–24
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Kinetic studies of the stability of Pt for NO oxidation: Effect of sulfur and
long-term aging
Jorge H. Pazmiño a, Jeffrey T. Miller b, Shadab S. Mulla a, W. Nicholas Delgass a, Fabio H. Ribeiro a,
⇑
a School of Chemical Engineering, Purdue University, 480 Stadium Mall Drive, West Lafayette, IN 47907-2100, USA
b Chemical Technology Division, Argonne National Laboratory, 9700 S. Cass Avenue, Argonne, IL 60439, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 February 2011
Revised 2 May 2011
Accepted 4 May 2011
Available online 12 June 2011
The stability of Pt catalysts for NO oxidation was analyzed by observing the effect of pre-adsorbed sulfur
on the reaction kinetics using a series of Pt/SBA-15 catalysts with varying Pt particle sizes (ca 2–9 nm).
Our results indicate that sulfur addition did not influence catalyst deactivation of any of the Pt catalysts,
resulting in unchanged turnover rates (TOR) and reaction kinetics. The presence of sulfur on Pt was con-
firmed by X-ray absorption fine structure spectroscopy (EXAFS) under reducing environments. However,
exposure of the catalyst to NO oxidation conditions displaced sulfur from the first coordination shell of Pt,
yielding Pt–O bonds instead. Re-reduction fully recovered the Pt–S backscattering, implying that sulfur
remained near the Pt under oxidizing conditions. X-ray photoelectron spectroscopy (XPS) and chemisorp-
tion measurements confirmed the presence of sulfur near platinum. The invariance of the NO oxidation
reaction to sulfur poisoning is explained by sulfur displacement to interfacial sites and/or sulfur binding
on kinetically irrelevant sites. Formation of Pt oxides remains as the main source of catalyst deactivation
as observed by kinetic and X-ray absorption spectroscopy (XAS) measurements.
Keywords:
NO oxidation
Sulfur
EXAFS
Pt oxidation
Deactivation
Stability
Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction
on a Pt/Rh/BaO/Al2O3 possibly caused by site blocking. Under sim-
ilar experimental conditions, Amberntsson et al. [6,12] observed an
NO oxidation over supported Pt catalysts plays a key role in
developing NOx abatement technologies for lean-burn engines,
such as the NOx storage reduction (NSR) catalysts [1]. Given the
increasing market prices for Pt and other noble metals, broad com-
mercialization of NSR catalysts still requires further improvements,
especially in terms of durability and resistance to poisons. The
presence of sulfur-containing compounds in fuel and lubricants
has been reported to favor the formation of thermally stable sulfate
species on NOx storage sites, detrimentally affecting catalyst per-
formance [2–4]. Such an outcome has been reported to occur to
similar extents when SO2, H2S, or COS were introduced, with rich
conditions causing greater catalyst deactivation as compared with
lean periods [5–8]. Regeneration of sulfur-poisoned catalysts usu-
ally requires high-temperature processes that lead to thermal
aging of the catalyst by Pt sintering and/or reaction of the storage
component with the support [9–11]. Sulfur has also been reported
to affect the metal functionality for NO oxidation [12–17]. While
SO2 has been shown to severely decrease the activity on rho-
dium-based catalysts [18,19], discrepancies still remain regarding
the effect of sulfur on Pt catalysts. For example, Engström et al.
[3] reported an inhibition effect of SO2 on the NO oxidation activity
increase in oxidation activity with the addition of SO2 on Pt/BaO/
Al2O3 and Pt/Rh/BaO/Al2O3 and concluded that the enhancement
in NO2 formation was caused by an increased resistivity to Pt oxi-
dation. However, in later studies [13,16] done by the same group,
the inclusion of SO2 in the NOx mixture showed a detrimental ef-
fect in the short time range and a slow recovery of activity in the
long term due to S-promoted platinum sintering. These contradic-
tory observations may be related to changes in platinum–sulfur
interactions with varying conditions and/or to contributions from
the storage component in the vicinity of Pt. In their kinetic study,
Ji et al. [14] reported a total suppression of NO oxidation activity
upon the addition of SO2 in the feed. When SO2 was introduced
prior to reaction, the activity of Pt/SiO2 was mostly restored, while
only a partial regeneration was observed on Pt/CeO2. Their results
pointed out the preferential oxidation of SO2 over NO and the effect
of support in catalyst regeneration. However, their kinetic analysis
was biased by the exclusion of the product NO2 in the feed stream
to account for product inhibition. Evidently, the effect of sulfur on
the catalytic role of Pt for the NO oxidation reaction still requires a
systematic kinetic evaluation without the complexities derived
from supports with high affinity for sulfur, such as Al2O3 [20–23]
or BaO/Al2O3 [7,9,24].
Sulfur adsorption on platinum has been widely studied on sup-
ported and single-crystal catalysts [25]. Wang et al. [26] showed
⇑
Corresponding author.
E-mail address: fabio@purdue.edu (F.H. Ribeiro).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.05.007

14
J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
that bulk sulfides are readily formed on SiO2-supported Pt, Pd, and
Rh catalysts upon heating in 10%H2S/N2, reaching complete surface
sulfidation at 300 °C. Also, they found that sulfides are partially
transformed to PtSO4 and PtO after calcinations in air at 300 °C
and that complete removal of sulfur was achieved at 500 °C. Other
studies on Pt (1 0 0) and (1 1 1) crystals [27,28] have also observed
sulfur adsorption with saturation coverage ranging from 0.4 to 1.
Their results suggest that more sulfur was adsorbed at lower tem-
peratures, reaching sulfur coverage (hs) of 0.92, but heating above
300 °C resulted in the desorption of weakly bonded S resulting in
a hs of 0.38–0.5. Given the high affinity of Pt to form stable com-
pounds with sulfur, saturation coverage can be achieved with
PH2S/PH2 in the order of 10ꢀ8 between 200 and 300 °C [25]. Catalyst
supports can form stable sulfates either by direct reaction at high
temperatures (>700 °C) or as a result of oxidation of sulfur on the
Pt surface and spillover SOx species. Therefore, depending on the
capacity of the support to store sulfur, platinum will be more or
less susceptible to be poisoned.
The chemical state of adsorbed sulfur has been reported to be
dynamic, changing with reaction environments. For example, EX-
AFS measurements taken by Gracia et al. [22] on a highly dis-
persed Pt/SiO2 catalyst for CO oxidation suggested that
adsorbed sulfur existed in two configurations depending on the
reaction atmosphere: (i) directly bonded to Pt under reducing
conditions and (ii) through an oxygen atom as Pt–OSOx when
oxidized. Sulfur can also modify the structure and electronic
properties of platinum and interact with the support as well.
Many studies [29–32] have reported on S-induced surface recon-
struction and particle agglomeration due to lowering of the sur-
face free energy for the sulfided Pt. Similarly, sulfur modifies the
co-adsorption of other molecules such as, H2, O2, and CO. These
molecules are of particular interest for many reactions as they
participate as reactants and/or they are used to quantify the sur-
face metal atoms by chemisorption techniques. In that regard,
Apesteguia et al. [33] reported that in the case of sulfided Pt/
Al2O3 catalysts, the Pt surface area calculated by H2 chemisorp-
tion was significantly lower than the values from O2 and CO
chemisorption.
To provide a better insight into the effect of sulfur on NO oxida-
tion and the nature of catalyst instability, we make use of kinetic
analysis and spectroscopic techniques (XPS, XAS) to investigate
the chemical state of Pt under NO oxidation conditions. Our ap-
proach consists of (i) using Pt/SBA-15 catalysts to study the effect
of sulfur and the regenerability of the Pt under operating condi-
tions and (ii) assessing the nature of deactivation and its effects
on kinetic parameters over long times by using a non-sulfided Pt/
Al2O3 monolith catalyst.
2. Experimental methods
2.1. Catalysts
The 1%Pt/SBA-15 catalysts with average particle size of 3.8 and
9.0 nm were provided by Professor R.M. Rioux. The samples were
prepared by a solution-based alcohol reduction method, which al-
lows sufficient control over the metal particle size to tune it to the
mesoporous SBA-15 silica support pore diameter [45]. Synthesis
and incorporation of Pt nanoparticles in the mesoporous silicate
support were done using low-power sonication. The Pt/SBA-15
with an average Pt particle size of 2.3 nm was prepared by incipi-
ent wetness impregnation as follows: after mixing 1 g of SBA-15
with 50 mL of H2O, 1 mL of NH4OH (pH > 10) was added and mixed
for 30 min. Once a homogeneous SBA-15 slurry was formed, a solu-
tion containing 0.02 g of Pt(NH3)4(NO3)2 dissolved in 25 mL of H2O
and 0.5 mL of NH4OH was added, and the mixture was stirred for
30 min at room temperature. The resulting solid was filtered and
rinsed twice with 50 mL of H2O and then dried at 125 °C overnight.
The sample was later calcined in flowing air at 200 °C for 2 h and
reduced at 250 °C for 1 h. Before exposing the catalyst to the room
conditions, He was flown at 250 °C to desorb H2, and the sample
was cooled to room temperature in He. To simplify catalyst nota-
tion, the particle size of the S-free catalysts was used to differenti-
ate each set of Pt/SBA-15 and its corresponding sulfided samples.
For example, Pt/SBA-2.3-S refers to 2.4%Pt/SBA-15 with mean Pt
particle size of 2.3 nm, which was subjected to the H2S/H2 treat-
ment described in Section 2.5.
Considering that NSR catalysts are designed to perform over
long periods of time, it is also relevant to gain a better understand-
ing of sulfur regenerability and catalyst stability. Partial sulfur re-
moval has been reported on Pt/Al2O3 catalysts upon reduction
treatments with H2 at temperatures above 350 °C, while oxidation
prior to low-temperature reduction was seen to completely re-
move the sulfur from the metal [34,35]. The mechanism of regen-
eration was proposed to occur via the formation of sulfates on the
Al2O3 support [34], a material that is known to interact with sulfur.
Apart from any sulfur effects on catalyst performance, platinum-
based formulations have been shown to deactivate over time under
NO oxidation conditions. Previous studies on non-sulfided cata-
lysts [36–38] have attributed the decrease in catalytic activity for
NO oxidation to the oxidation of Pt particles (as seen by XPS) pro-
moted by the presence of NO2. In fact, Segner et al. [39] and Parker
and Koel [40] have shown that NO2 is a very effective source of sur-
face oxygen because of its high sticking coefficient. As a result, NO2
is known to kinetically inhibit the reaction as it competes for
adsorption sites for O2 [41]. The inhibition effect, however, cannot
explain the persistent deactivation with time observed in studies
where the product NO2 was included in the feed [37]. Such insta-
bility has also been reported to occur over long time periods, com-
plicating the collection of reproducible kinetic measurements
[16,42]. Other researchers [43] have suggested that platinum re-
mains metallic regardless of the cluster size over a large range of
oxygen virtual pressures and suggest that deactivation could be re-
lated to impurities in the NO2-containing gas mixture, possibly in
the form of bromine or chlorine species [44].
The Pt/Al2O3 catalyst in the monolithic form was supplied by
EmeraChem. The monolith was dipped into an aqueous slurry,
which contained the appropriate amount of
c-Al2O3 based on a
known liquid uptake of the bare monolith. The monolith was then
drained, dried, and calcined in flowing air at 500 °C for 1 h. Depo-
sition of Pt was done by dipping the treated monolith into a Pt-
containing aqueous solution in such a way that the final Pt loading
was approximately 1.8 ꢁ 10ꢀ3 g cmꢀ3 of monolith. An amine-based
solution was used as the Pt precursor. The resulting monolith had a
Pt loading of ca. 0.3 wt.% and a cell density of 31 channels cmꢀ2
.
The fraction of Pt surface metal to the total Pt atoms was measured
using the CO titration technique described in the next section.
2.2. Metal dispersion measurements by CO titration, H2 chemisorption,
and H2 titration
CO titration experiments were conducted in the NO oxidation
reactor, which is described in Section 2.3. The catalysts were sub-
jected to an oxygen exposure at room temperature with 30% O2/N2
(1 L minꢀ1) for 1 h. According to Benson and Boudart [46], this
room-temperature oxygen exposure leads to a monolayer of ad-
sorbed oxygen on Pt with an O/PtS ratio of unity. After this pre-
treatment with oxygen, the flow was switched to Ar
(750 mL minꢀ1) in order to purge out any residual gases from the
pre-treatment, and the reactor temperature was set to 150 °C. At
this titration temperature, the flow was changed to 3% CO/Ar
(750 mL minꢀ1), and the exit gases were analyzed in an FT-IR gas

J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
15
analyzer (Thermo Electron Corporation, Nicolet Antaris IGS) for
CO2. The CO2 trace was followed until it decreased to below the
detection limit (typically about 10 min). The resulting CO2 trace
was integrated and, assuming a reaction stoichiometry CO + O–
PtS ? CO2 + PtS, the total amount of oxygen on the Pt was calcu-
lated. At 150 °C, CO can titrate the oxygen monolayer completely
from Pt, and hence, the CO2 amount formed (or the atomic oxygen
removed) will be the same as the number of exposed surface Pt
atoms [37]. It is thus possible to calculate the number of surface
Pt atoms from the CO titration at 150 °C and consequently the Pt
dispersion (ratio of surface Pt to total Pt) while the catalysts are
still in the reactor. The Pt average particle size was then calculated
as (d in nm ꢂ 1.1/dispersion).
catalyst bed and right below the quartz frit. Mass flow controllers
were used to control all the gas flows, and their concentrations
were varied by adjusting their respective flow rates. A pressure
gauge with an analog output was used to monitor the feed pres-
sure to the reactor.
2.4. Reaction kinetic measurements
The monolith catalyst was pre-treated at 250 °C with 10% O2 in
N2 for 1 h to oxidize any carbonaceous impurities, followed by
reduction with 0.5–1% H2 in N2 for 1.5 h with a constant total flow
rate of 6.6 L minꢀ1. For the powder samples, the pre-treatment in
10% O2/N2 and the reduction in 1%H2/N2 were done at 340–
350 °C, at a total flow rate of 3.5 L minꢀ1. The reactor was operated
in a differential regime by keeping the NO conversions below 10%
and by including NO2 in the feed to minimize the contribution of
the NO2 formed to the total NO2 concentration. The NO and NO2
concentrations in the outlet gas were detected with a FT-IR gas
analyzer (Thermo Electron Corporation, Nicolet Antaris IGS). Fac-
tory-supplied calibrations were used for the analysis of the spectra
collected at a resolution of 0.5 cmꢀ1. The kinetic measurements at
different gas composition or temperature conditions were taken
after the catalyst was on stream for at least 1 h, once the sample
has been stabilized. Stabilization treatments consisted of exposing
the catalysts to reaction conditions for at least 16–18 h. The
repeatability of the measurements was checked by periodically
returning to a base case condition (300 ppm NO, 170 ppm NO2,
10%O2, balance N2, 3.4 L minꢀ1, 1 atm) at a chosen temperature
(generally 300 °C). Acceptance of data as stable required repeat-
ability of the base case condition over the entire time period of
the experiment. We were not able to use the Madon–Boudart
[47] test to verify experimentally that no heat and mass transport
limitations were present, because of the sensitivity of rates with
dispersion and deactivation. Instead, the criteria proposed by Dek-
ker et al. [48] were used for both the monolith and powder sam-
ples. Detailed calculations reported in the Supplemental
information confirmed negligible transport effects. All parameters
were obtained from direct measurements, correlations [49], or data
from handbooks [50].
The estimate of the apparent activation energy at low and high
reaction temperatures was determined from data collected be-
tween 150–200 °C and 250–320 °C, respectively. The temperature
was varied in a random fashion while maintaining the standard
feed composition at 300 ppm NO, 170 ppm NO2, 10% O2, and bal-
ance N2 with a total flow rate of 6.6 L minꢀ1 for the monolith and
3.4 L minꢀ1 for the powder samples. The reaction orders were ob-
tained by performing one-component-at-a-time variation experi-
ments over the ranges 150–420 ppm for NO, 5–30% for O2, and
120–420 ppm for NO2, at a constant baseline temperature of
160 °C for Pt/Al2O3 monolith or 300 °C for Pt-SBA-15 catalysts.
Least-squares analysis of the Arrhenius plot and the linear form
of the power law expression were used to estimate the apparent
activation energies, orders of reaction, and their confidence limits.
Following the standard methods for H2 chemisorption and H2–
O2 titration, the Pt surface area for clean and sulfided catalysts was
determined [46]. The measurements were taken on a Micromeri-
tics ASAP 2020 instrument. The reduction temperature was
250 °C for all catalysts, and the maximum temperature for evacu-
ation was set at 200 °C as a precaution in case sulfur were to des-
orb from the samples and contaminate the analysis manifold.
Hydrogen chemisorption was done at 35 °C. Upon completion of
the H2 chemisorption experiment, the catalyst was purged in He,
heated to 100 °C, and oxidized in UHP O2 for 30 min at 100 °C.
Titration with hydrogen was subsequently performed at 100 °C.
2.3. Reactor setup
The reactor system was designed to operate with two types of
reactors depending on the form of the catalyst. For the monolith
catalyst, a bench-top, plug-flow stainless steel reactor (2.6 cm in
OD) was used for measuring the conversion of NO. The monolith
was wrapped with high-temperature Zetex insulation to minimize
the gas flow bypassing the catalyst. A 6–8-cm bed containing clean
glass beads was placed upstream in the reactor to ensure unifor-
mity of the mixture inlet gas. The reactor was placed inside a tem-
perature-controlled furnace, and a pre-heater was used before the
reactor to avoid temperature gradients along the monolith. The
temperature was monitored upstream and downstream of the
monolith with two K-type thermocouples, each at a distance of
6 mm with respect to the edge of the sample. A reactor bypass line
located after the pre-heater was used to measure the inlet concen-
trations of NO and NO2 after each catalyst evaluation. The gas-
phase bypass concentrations were used as the inlet concentrations
for all the rate calculations in order to account for the reaction over
the catalyst.
The reactor design for powder samples consisted of a bench-
top, plug-flow 1.3-cm-diameter vertical Pyrex reactor with a
190–233-lm glass frit fitted at the midpoint of the reactor to sup-
port the powdered catalysts. Since the particle size of some of the
powder catalysts was smaller than the openings of the glass frit, a
thin (about 2 mm) plug of clean quartz wool was placed over the
frit in order to support the catalyst powder. The two ends of the
1.3-cm-diameter glass reactor were joined to a 7.6-cm-long, 2.6-
cm-OD-diameter, thick-walled glass tubing. The reactor was
cleaned prior to loading the catalyst with a phosphate-free clean-
ing solution, followed by pouring in fresh aqua regia followed by
rinsing with DI water and drying. A 4–5-mm-deep bed of glass
beads was supported on a metallic mesh located at 1 cm above
the catalyst. Another layer of quartz wool was added on top of
the catalyst and slightly pressed against it to ensure uniformity
of the bed. In order to maintain isothermal conditions across the
catalyst bed, a 19-cm-long aluminum sheath was placed around
the 1.3-cm-diameter portion of the glass reactor. In addition, the
glass was wrapped with a 5-cm-long aluminum foil centered in
the portion of the reactor where the catalyst bed was located.
Two K-type thermocouples were placed at ꢃ2 mm above the
2.5. Ex situ sulfur pre-treatment
The Pt/SBA-15 powder catalysts were treated in an independent
bench-top, plug-flow, 1.3-cm-diameter vertical Pyrex reactor as-
signed only for sulfur treatments. The choice of catalyst for S poi-
soning was based on earlier findings that concluded that SiO2
supports are less susceptible to the formation of surface sulfates
compared to Al2O3 or TiO2 [20,22]. Therefore, sulfur will interact
mostly with Pt instead of being stored in the support. With the
aid of a separate temperature-controlled furnace, the fresh cata-
lysts were pre-treated with H2S under either reducing (H2S/H2
labeled as ‘‘-S’’) or oxidizing (Air/H2S labeled as ‘‘-SO’’) conditions.

16
J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
The concentration of H2S was selected so as to exceed the 50–
200 ppm sulfur concentration values typical of engine exhaust.
For the first pre-treatment, ꢃ150 mg of Pt/SBA-15 was heated to
250 °C in Ar and reduced in 100 mL minꢀ1 of H2 at 250 °C for 2 h.
Then, a mixture containing 100 mL minꢀ1 of 538 ppm H2S/N2 and
25 mL minꢀ1 H2 was flowed through the catalyst bed for 2 h, and
the presence of sulfur in the effluent was verified by the formation
of the precipitate silver sulfide while the effluent was bubbled
through a solution of 1000 ppm AgNO3 (Fluka Scientific, certified
ACS). Weakly adsorbed sulfur was removed by reducing the cata-
lyst in H2 for 2 h at 250 °C before the samples were cooled in Ar
and exposed to air. The second pre-treatment consisted of room-
temperature exposure to flowing air (ꢃ100 mL minꢀ1) for about
20 min followed by a temperature ramp to 150 °C to calcine the
sample for 30 min at 150 °C. The catalysts (ꢃ150 mg) were purged
in Ar at 150 °C and treated in 538 ppm H2S/N2 for 2 h, all at a con-
stant flow rate of 100 mL minꢀ1. The same AgNO3 solution was
used to verify the presence of sulfur flowing through the catalyst
bed. After the (second) sulfidation treatment, the reactor was
flushed with Ar for 15 min before increasing the temperature to
250 °C for reduction in 100 mL minꢀ1 flow of H2 for 1 h. Based on
previous observations [51], the sample was heated in
100 mL minꢀ1 Ar to 350 °C and treated in air for 5 h to decompose
any bulk PtS into SO2 and thus prevent undesirable contaminations
in the NO oxidation kinetic reactor and instruments. During either
sulfur pre-treatment, the 150 mg of Pt/SBA-15 catalyst containing
about 10ꢀ6 moles of surface Pt was exposed to a total of 3 ꢁ 10ꢀ4
moles of H2S.
The two sulfur treatments were intended to poison the catalysts
to different degrees so as to explore possible correlations with the
NO oxidation rate. Under reducing conditions, the S-free metallic
catalysts samples were exposed to a mixture of H2S/H2 at 250 °C.
The introduction of H2 in the feed will shift the equilibrium toward
H2S, preventing the formation of large quantities of bulk platinum
sulfide (which could lead to particle sintering) and also maintain-
ing a gradientless poisoning process throughout the catalyst bed.
At the poisoning conditions (PH2S/PH2 = 2 ꢁ 10ꢀ3, 250 °C, 2 h), en-
ough sulfur is available to reach saturation coverage, typically
around 0.4–1.0 for Pt depending on temperature and sulfur con-
centration. The second sulfur pre-treatment aimed to poison Pt
to a greater extent by oxidizing the surface and then reacting the
platinum oxides with H2S to form stable (at T > 350 °C) bulk PtS.
In addition, hydrogen was not included in the feed, and the tem-
perature during the sulfur treatment was lowered to 150 °C to fa-
vor the equilibrium toward adsorbed sulfur.
yield about 2.0 total absorbance (
lx) at the Pt LIII edge and a Pt
edge step ( x) of about 0.5. The measurements were taken on
D
l
the clean and sulfided Pt/SBA-15 (2.3 nm, 3.8 nm, and 9.0 nm) cat-
alysts. The XAS were collected after the following ex situ pre-treat-
ments done on both clean and sulfided samples: (i) after reduction
in 3.4% H2/He at 300 °C for 1 h followed by purging with He for 10–
20 min while cooling to room temperature and (ii) after exposure
to the standard gas composition for NO oxidation (300 ppm NO,
170 ppm NO2 and balance air) for 1 h at 300 °C and 350 mL minꢀ1
followed by purging with He while cooling to room temperature
(Short NOx). For clean samples, a second pre-treatment was done
consisting of exposing the samples to a highly oxidized NOx gas
mixture of 200 ppm NO, 500 ppm NO2, and balance air for 18 h
at 300 °C and 350 mL minꢀ1 followed by purging with He while
cooling (Long NOx). The catalysts were analyzed in a continuous-
flow EXAFS reactor cell (46 cm long, 1.9 cm diameter) fitted at both
ends with polyimide windows and valves to isolate the reactor
from the atmosphere. All the spectra were obtained at room
temperature.
In situ XANES experiments were done on all the S-free Pt/SBA-
15 catalysts using the same procedures described above. The anal-
ysis was performed on these catalysts for each of the following pre-
treatments: (i) reduction in 2–4% H2/He at 250 °C for 30 min or un-
til the EXAFS showed no Pt–O backscatter followed by (ii) purging
in He at 250 °C for about 5 min and (iii) exposure to NO oxidation
reaction conditions, similar to those used for the kinetic measure-
ments (300 ppm NO, 170 ppm NO2, 10% O2, balance He, 250 °C).
The XAS spectra were collected at 90-s intervals starting at 210 s
after introducing the gases. All the data were obtained at 250 °C
until no significant changes were seen.
Reference compounds were used to correct for phase shifts and
backscattering amplitudes and for fitting the XANES data. For Pt0
(XANES) and the Pt–Pt bond distance (EXAFS), a Pt foil was used.
Similarly, the reference for Pt2+ was taken from an oxidized
(100 °C) 1 nm Pt/SiO2, which showed 4 Pt-O bonds and no metallic
Pt in the EXAFS. Phase and amplitude reference for Pt–S was ob-
tained from H2PtCl6 with a Pt–Cl distance of 2.32 Å. Note that scat-
tering is dependent on the number of electrons of the neighbor and
is identical for scattering pairs that differ by only a few electrons.
Thus, Pt–Cl is a good reference for the Pt–S coordination in these
samples. Lastly, a XANES spectrum for Pt4+ was collected from
transmission data on Na2Pt(OH)6 at 25 °C. The EXAFS data were
processed and analyzed using WINXAS97 software [52]. A least-
squares fit in R-space of the coordination k2-weighted Fourier
transform data allowed the calculation of the nearest neighbor
first-shell Pt–O, Pt–S, and Pt–Pt parameters. The data with k1 and
k3 weightings also resulted in satisfactory fits with negligible var-
iation in the parameters.
2.6. X-ray absorption spectroscopy (XAS) experiments
X-ray absorption measurements were taken at the Advanced
Photon Source in Argonne National Laboratory using the inser-
tion-device beam line of the Materials Research Collaborative Ac-
cess Team (MRCAT). To minimize the background from
harmonics, a cryogenically cooled double-crystal Si(1 1 1) mono-
chromator and an uncoated glass mirror were used. During mea-
surements, the monochromator was scanned constantly, and the
data were integrated over 0.5 eV for 0.07 s per data point. All
experiments were done in transmission mode. To optimize the cur-
rent (109–1010 photons detected per second) with a linear re-
sponse, a mixture of 10% Ar in N2 (10% adsorption) was flown
into the initial ionization chamber, and a mixture of 80% Ar in N2
(70% adsorption) was used for purging the transmission ionization
chamber. Energy calibration was done by simultaneously acquiring
the spectrum of a platinum foil.
2.7. X-ray photoelectron spectroscopy (XPS) measurements
A Kratos Ultra DLD SPS spectrometer was used for the analysis
of sulfur species on the Pt/SBA-15 sample with 2.2 nm average Pt
particle size. A pellet of about 1.42 cm in diameter and 1.3 mm
thickness was prepared by pressing about 100 mg of the powdered
catalyst in a bench-top hydraulic press under 5000 psi. Once
loaded, the sample was treated with different gas mixtures and
transferred to the UHV chamber without exposing it to the atmo-
sphere. The analysis was performed on the sulfided Pt/SBA-15
(Pt/SBA-2.3-S) after: (i) short exposure to air at room temperature
and (ii) reduction in 5%H2/Ar at 300 °C for 1 h and then cooling in
the same reducing mixture to 40 °C. Monochromatic Al K
a radia-
tion was used to record the XPS spectra. All binding energy (BE)
values were calibrated to the Si 2p energy of 104.3 eV with respect
to the Fermi level [53]. The spectra were fitted using CasaXPS soft-
ware (Casa Software Limited) assuming a line shape of a Hybrid
A cylindrical holder containing up to six catalyst pellets was
used for the transmission XAS experiments. The thickness of the
Pt/SBA-15 pellets (typically 40–70 mg of catalyst) was chosen to

J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
17
Table 2
Doniach Sunjic/Lorentzian function [54] for Pt 4f and a pure Gauss-
ian function for S 2s. The separation of the Pt 4f doublet was fixed
to 3.3 eV, and the 4f5/2:4f7/2 area ratio was set to 0.75. A Shirley-
type background was fit and subtracted in the Pt 4f region [55].
Comparison with literature values for turnover rates of NO oxidation on Pt catalysts.
a
Catalyst
Pt loading
(wt.%)
Pt dispersion
(%)
TOR
Ref.
(10ꢀ3 sꢀ1
)
Pt/Al2O3
Pt/Al2O3
Pt/SiO2
1.2
0.3
2.5
47
42
14.2
6.2
3.5
42
Weiss and Iglesia [43]
Mulla et al. [77]
Despres et al. [78]
Smeltz et al. [44]
3. Results
Pt(1 1 1)
113b
3.1. Effect of pre-adsorbed sulfur on NO oxidation kinetics
a
Turnover rate (TOR) at 300 °C corrected to 300 ppm NO, 170 ppm NO2, 10%O2,
balance N2. The active sites are the Pt atoms on the surface after reduction.
Kinetic measurements taken on the set of Pt/SBA-15 powder
catalysts after ex situ treatments with H2S along with the ones on
the unsulfided samples are summarized in Table 1. The forward
rate is reported as a turnover rate (TOR), which is defined as the
number of moles of NO reacted per second per mole of surface
Pt. The observed rates, robs, are corrected by a factor b, which is
the approach to equilibrium using the expression:
b
Oxygen coverage of 0.76 0.06 ML.
factor of two. Such variations were not greatly influenced by the
normalization to the Pt surface area since H2S/H2 treatments did
not affect the Pt dispersion (as measured by CO titration) and only
a slight sintering was observed after Air/H2S treatments.
In the absence of sulfur, the kinetic parameters are in good
agreement with our proposed mechanism [41] with a near-first-or-
der dependence of the TOR on NO and O2 and a near-negative-first-
order dependence on NO2 and an apparent activation energy (Ea)
varying from 82 2 to 102 4 kJ molꢀ1. As will be discussed in la-
ter sections, the values of Ea reported in Table 1 were obtained
after the catalysts were stabilized for over 40 h on stream and they
differ (by about 20–40 kJ molꢀ1) from the Ea observed with shorter
time on stream (TOS). Within experimental error, the addition of
sulfur by H2S/H2 showed no effect on the apparent activation en-
ergy or the dependence on gas composition for any of the Pt/
SBA-15 catalysts as the Pt particle size varied. A similar outcome
was observed when the catalysts were oxidized prior to the second
sulfur treatment at 150 °C. Given that the second sulfur treatment
caused no differences in TOR, it is unlikely that the orders of reac-
tion, which proved to be the less-sensitive parameters, were af-
fected. To explore the possibility of sulfur desorbing/reacting
during reaction, we also compared the transient measurement of
the rate as shown in Fig. 1 for Pt/SBA-2.3 and correspondingly in
Fig. S1 for Pt/SBA-3.8 and Fig. S2 for Pt/SBA-9.0 included in the
Supplemental information section. The similarities in the shape
of the transient-rate-versus-time curves suggest the existence of
a source of deactivation that does not depend on the presence of
sulfur and confirm that the NO oxidation reaction is insensitive
to the presence of pre-adsorbed sulfur on Pt/SBA-15 catalysts.
rf ¼ robsð1 ꢀ bÞꢀ1
ð1Þ
ð2Þ
2
½NO2ꢄ
b ¼
2
K½NOꢄ ½O2ꢄ
where K is the equilibrium constant of the overall reaction,
2NO(g) + O2(g) ? 2NO2(g), as written. A typical range for b during
our experiments was 0.01–0.15, verifying that the reaction was
far from equilibrium. All values reported in Table 1 correspond to
the ones measured after the catalysts were stabilized for 20–40 h,
and no further deactivation was observed. Comparison of measured
turnover rates with literature values is reported in Table 2.
A clear trend emerges from the data in Table 1 when one com-
pares the TOR as a function of platinum particle size. There is an
increase in the rate by about 60 times upon an increase in the aver-
age Pt particle size increased from 2 to 9 nm and by about 35 times
for an increase in size from 2 to 4 nm. The results are in line with
kinetic observations and DFT calculations for the effect of Pt parti-
cle size on the NO oxidation rate [37,38,56–60], which have sug-
gested that smaller Pt particles exhibit lower rates because of
their propensity to form inactive oxides. Alternatively, this depen-
dency has also been explained in mechanistic terms from the dif-
ference in binding energy of chemisorbed oxygen on large Pt
particles, which bind Oꢅ more weakly than low coordinated Pt
atoms typically present on smaller particles [43]. While the effect
of pre-sulfiding the catalysts on the TOR was within a factor of 2,
we observed that, as the Pt particle size increased, the variations
in TOR with respect to the those on the S-free catalysts became
more evident, yielding a higher rate for the samples treated in
H2S/H2 and a lower rate for those treated in Air/H2S. For the cata-
lyst with the smallest Pt particle size (Pt/SBA-2.3), the TOR of the
sulfided samples was nearly identical to the TOR of the clean cat-
alyst. Similarly, the TOR for the catalysts with medium (Pt/SBA-
3.8) and large (Pt/SBA-9.0) Pt particle size deviated by, at most, a
3.2. EXAFS and XPS characterization of the S-treated catalysts
The main goal of the EXAFS measurements was to study the lo-
cal environment of the Pt atoms in search for evidence of the pres-
ence of sulfur. Fig. 2 shows the k2-weighted Fourier transform (FT)
collected on the set of fresh and sulfided Pt/SBA-2.3 and Pt/SBA-3.8
catalysts. The FT of the spectra resulted in backscattering signal
Table 1
Summary of NO oxidation kinetics on Pt/SBA-15 samples as a function of particle size and different sulfur pre-treatments.
a
b
Catalyst
D (%) (d, nm)
TOR (10ꢀ3 sꢀ1
)
Ea (kJ molꢀ1
)
NO order
O2 order
NO2 order
Pt/SBA-2.3
Pt/SBA-3.8
Pt/SBA-9.0
Pt/SBA-2.3-S
Pt/SBA-3.8-S
Pt/SBA-9.0-S
Pt/SBA-2.3-SO
Pt/SBA-3.8-SO
Pt/SBA-9.0-SO
47 (2.3)
29 (3.8)
12 (9.0)
49 (2.2)
32 (3.4)
11 (10.1)
41 (2.7)
25 (4.4)
10 (11.0)
1.1 0.1
82
89
102
85
82
94
69
111
99
2
7
4
5
4
3
7
8
8
1.06
1.04
0.90
0.93
0.94
1.17
0.86
0.86
0.79
1.07
0.87
0.86
ꢀ0.95
ꢀ1.02
ꢀ1.00
ꢀ0.84
ꢀ1.22
ꢀ1.16
38
67
2
6
1.2 0.1
57
136
2
3
1.4 0.1
26
49
2
5
a
Average Pt particle size from CO titration, calculated using d (nm) ꢂ 1.1/(Pt dispersion, D).
b
Turnover rate (TOR) at 300 °C measured at 300 ppm NO, 170 ppm NO2, 10%O2, balance N2 after stabilization for >40 h, and normalized by the surface Pt atoms measured
by CO titration.

18
J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
decreases, making it more difficult to resolve it from the Pt–Pt
backscattering contribution. As a consequence, it was not possible
to detect Pt–S bonds on Pt/SBA-9.0-S (Fig. S3, Table S1).
0.10
0.08
0.06
0.04
0.02
0.00
Reduced
H2S/H2
2 nm Pt
To understand the state of sulfur under reaction conditions, EX-
AFS were collected after treating the samples ex situ in a typical NO
oxidation mixture for 1 h (Short NOx) and also after a subsequent
reduction to study the reversibility of the Pt–S bonds. Fig. 3 shows
the corresponding FT EXAFS for the case of the highly dispersed
(Pt/SBA-2.3) catalyst. The results indicate that all the Pt–S back-
scattering vanished after NO oxidation, leaving only Pt–O contribu-
tions at a distance of 2.04 Å (Table S1). This is consistent with the
XANES fittings that show that Pt is oxidized after exposure to NOx.
Interestingly, when the sample was reduced after NO oxidation,
the intensity (in Fig. 3) and the coordination number (in
Table S1) of the Pt–S bonds were completely recovered. This inter-
action was further confirmed by the presence of Pt–S backscatter-
ing on a reduced catalyst that was previously aged by exposure to
reaction conditions for a long time on stream (>40 h). Similar Pt–Pt
and Pt–S coordination numbers and bond distances were obtained
on this aged catalyst as seen on the fresh one (Table S1). A qualita-
tive comparison of the FT EXAFS on fresh and aged Pt/SBA-2.3-S
catalysts is presented in Fig. S4.
Air/H2S
0
2000
4000
Time /s
6000
Fig. 1. Transient NO oxidation TOR measurements for the first 2 h on stream for
2.4% Pt/SBA-15 (Pt/SBA-2.3), Pt/SBA-2.3-S and Pt/SBA-2.3-SO at 300 °C and 300 ppm
NO, 170 ppm NO2, 10%O2, balance N2.
The sulfided Pt/SBA-2.3-S sample was also analyzed by XPS in
order to identify the chemical states of sulfur and estimate surface
coverage. Fig. 4 shows the S 2s spectra obtained for Pt/SBA-2.3-S
(A)
Pt-Pt
Pt-S
Sulfided Air/H2S
0.01
Sulfided H2S/H2
S-free
Pt-O
Pt-S
0.01
Pt-Pt
0.00
0.02
NOx + Reduced
Reduced
1 h NOx
0
2
4
6
R[Å]
Pt-Pt
0.00
(B)
Pt-S
0
2
4
6
Sulfided Air/H2S
R[Å]
Fig. 3. Magnitude of the Fourier transform EXAFS (k2:
3.1 Å) for the sulfided 2.4%Pt/SBA-15 (Pt/SBA-2.3-S) catalyst after a series of ex situ
treatments. The NOx treatment consisted of flowing a NOx gas mixture of 300 ppm
NO, 170 ppm NO2 in air for 1 h at 300 °C, and 350 mL minꢀ1 total flow rate.
D , DR = 1.5–
k = 2.8–11 Åꢀ1
Sulfided H2S/H2
S-free
0.01
0.00
(B)
(A)
(SO4)2-
0
2
4
6
R[Å]
Fig. 2. Comparison of the magnitude of the Fourier transform EXAFS (k2:
11 Åꢀ1
D
k = 2.8–
,
D
R = 1.5–3.1 Å) for the set of clean and sulfided (A) 2.4% Pt/SBA-15 (Pt/SBA-
2.3) and (B) 1%Pt/SBA-15 (Pt/SBA-3.8) catalysts. The data were collected after
reduction at 300 °C in 3.4%H2/He for 1 h and purging in He to room temperature.
(SO3)2-
S-Pt
corresponding to both Pt–Pt and Pt–S bonds, confirming the pres-
ence of sulfur on the reduced catalysts. Fitting of the EXAFS in
Table S1 resulted in bond distances of 2.74 0.02 Å for Pt–Pt and
2.28 Å for Pt–S. Table S1 also shows that Pt–S bonds were present
on the reduced Pt/SBA-2.3-SO and Pt/SBA-3.8-SO catalysts. As the
Pt particle size increases, the signal is dominated by the bulk struc-
ture and therefore, the relative intensity of the Pt–S signal
235
230
B.E. / eV
225
Fig. 4. XPS spectra in the S 2s region for the sulfided 2.4%Pt/SBA-15 (Pt/SBA-2.3-S)
catalyst after (A) sulfur treatment and exposure to ambient conditions and (B) after
reduction in 5%H2/Ar for 1 h at 300 °C followed by cooling in the same reducing
mixture to 40 °C.

J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
19
Table 4
catalyst after the two different conditions described in Section 2.7.
Estimation of sulfur coverage on Pt/SBA-15 samples by chemisorption.
The S 2s region was chosen for analysis since the main sulfur peaks
in the 2p region were not observed due to overlapping by inelastic
losses from the nearby Si 2s peak at 154.94 eV. Three types of sul-
fur species can be suggested from the XPS spectra, based on previ-
ous findings: (i) chemisorbed sulfur on platinum, S–Pt, located at
226–227 eV [61], (ii) adsorbed SO23ꢀ species at 229–230 eV [61],
and (iii) adsorbed SO24ꢀ at 234–235 eV [62,63]. Quantification of
relative atomic ratios, FWHM, and peak positions for sulfur species
are summarized in Table 3 along with the corresponding values for
Pt 4f7/2 and Pt2+ 4f7/2. The accuracy of the measurements is 0.2 eV
for the binding energy and 0.4 eV for the FWHM. By comparing
the atomic ratios of S-containing species after exposure to air
and reduction at 300 °C, it is observed that the reduction treatment
was able to increase the sulfide intensity by 21%. Such increase is
accompanied by a decrease of 12% and 9% in S4+ and S6+ intensities,
respectively. Here, the total sulfur intensity for both analysis con-
ditions differed only by 3%.
Under the assumption that all sulfur species are adsorbed on Pt,
a surface coverage (hs) of 0.7–0.9 ML was calculated by dividing the
atomic ratios of sulfur and surface platinum (see Table S3). Since
the particles on Pt/SBA-2.3-S are small, we assumed that the inten-
sity of surface Pt is proportional to its total intensity multiplied by
the dispersion. If only sulfide and sulfite species are assumed to
interact with Pt, then hs = 0.4–0.5 ML, and if sulfides are the only
adsorbed species, then hs was only about 0.2–0.3 ML. If one as-
sumes that, after reduction at 300 °C, no Pt2+ comes from surface
oxides but from PtꢀS, a sulfur coverage around 0.65–0.7 ML can
be estimated by dividing the atomic ratio of Pt2+ by that of total
surface platinum (Pt0 and Pt2+). Also, the results from Table 3 con-
firmed that short exposures to ambient conditions after H2S/H2
only increased the surface oxidation by 10% compared with the re-
duced catalyst. The Pt 4f7/2 peak for the reduced sample was ob-
served at 71.5–71.6 eV, which is in line with reported values for
reduced platinum on Pt/SBA-15 [64].
Catalyst
Pt dispersion (%)
CO titration
hs
H2 chem.
H2 titration
Pt/SBA-2.3
Pt/SBA-3.8
Pt/SBA-9.0
Pt/SBA-2.3-S
Pt/SBA-3.8-S
Pt/SBA-9.0-S
Pt/SBA-2.3-SO
Pt/SBA-3.8-SO
Pt/SBA-9.0-SO
50
22
10
46
31
11
52 11
25
7
5
3
1
4
2
1
49
22
8
18
9
9
10
8
5
3
1
5
1
1
1
1
1
55
20
8
47
21
9
47
25
8
2
1
1
4
3
1
1
1
1
0.6 0.12
0.5 0.24
0.0 0.30
0.8 0.02
0.7 0.16
0.7 0.42
1
1
2
The results for hs in Table 4 range from 0.5 to 0.8 and reveal that
sulfidation under reducing conditions was not sufficient to poison
the catalyst with the largest mean particle size (Pt/SBA-9.0-S). Inter-
estingly, the second sulfur treatment, i.e., 538 ppm H2S/N2 over the
pre-oxidized catalyst, effectively poisoned around 70% of the sur-
face of reduced platinum on all samples regardless of particle size.
3.4. XAS evidence for deactivation during NO oxidation
Following previous studies on NO oxidation and NO2 dissocia-
tion [36,38], which showed that deactivation is likely to be caused
by platinum oxidation, we performed a series of in situ XANES
experiments on the set of Pt/SBA-15 catalysts to investigate the
chemical state of Pt under reaction conditions. After the catalysts
were reduced and purged in He, they were exposed to the NO oxi-
dation mixture, and the XANES spectra were collected every 60–
120 s. Our results indicate that Pt oxidizes with increasing time
on stream (TOS) in all the samples, but the fraction of Pt0 strongly
depends on the particle size, as shown in Fig. 5. The Pt0 fraction
seemed to reach a steady value after 10 min on stream in all cases.
While Pt/SBA-2.3 shows complete oxidation of Pt, only 40% and
20% of the Pt gets oxidized on Pt/SBA-3.8 and Pt/SBA-9.0, respec-
tively. That the greatest decay in metallic character occurred dur-
ing the first 400 s does not explain the persistent decrease in
turnover rate observed in Fig. 1 under similar conditions. Since
the in situ experiments were done within 25 min and our steady
state kinetic measurements required at least 16 h for stabilization,
a second set of experiments was conducted to verify the state of Pt
after longer time on stream (TOS).
The S-free Pt/SBA-15 catalysts were pre-treated ex situ, as de-
scribed in Section 2.6, and the results are summarized in
Table S2. A treatment in 4%H2/He at 300 °C was sufficient to com-
pletely reduce platinum, according to the fitted XANES spectra. The
Pt particle sizes were estimated from the coordination numbers
using previous calibrations. These results are in good agreement
with the corresponding values for dispersion obtained by the inde-
pendent CO titration method reported in Table 1. The Pt–Pt bond
distance on reduced catalysts was around 2.76 0.02 Å, except
for Pt/SBA-2.3, where a shorter distance of 2.66 0.01 Å was
observed. This contraction of the Pt bond distance occurs in the
absence of chemisorbed species when the particles are small
3.3. Quantification of sulfur coverage by H2 chemisorption–titration
methods
Measurements of Pt surface area by chemisorption methods are
summarized in Table 4. From the data, two trends emerged. First,
the dispersion calculated from hydrogen chemisorption was signif-
icantly lower than those from CO or H2 titration on all samples,
with the exception of the one with the largest Pt particle size
(Pt/SBA-9.0-S). However, the dispersion from titration of surface
oxygen with hydrogen yielded values similar to those calculated
from the CO titration method. Based on the EXAFS results, if one as-
sumes that oxidation allows oxygen to occupy all the Pt sites
regardless of the presence of sulfur and that the same (Pt:O) stoi-
chiometry is preserved, then a surface coverage can be estimated
by taking the ratio:
2
3 ðH2jtitrationÞ ꢀ 2ðH2jchemÞ
hs ¼
ð3Þ
2
3 ðH2jtitrationÞ
where the numbers in parenthesis are the specific gas uptake
(mL gꢀ1) as measured by chemisorption and titration of hydrogen.
Table 3
XPS S 2s and Pt 4f7/2 peak positions, FWHM and relative atomic ratios (R.A.) for the sulfided 2.4%Pt/SBA-15 catalyst (Pt/SBA-2.3-S).
Treatment
S–Pt, 2s
BE (eV)
Pt, 4f7/2
BE (eV)
Pt2+, 4f7/2
BE (eV)
SO23ꢀ; 2s
SO24ꢀ; 2s
R.A. (%)
FWHM (eV)
BE (eV)
R.A. (%)
BE (eV)
R.A (%)
R.A. (%)
FWHM (eV)
R.A. (%)
a
As received
5%H2/He
227.1
227.2
0.089
0.128
2.9
2.7
229.9
230.0
0.085
0.037
234.9
235.0
0.158
0.103
71.5
71.4
0.58
0.57
2.2
2.3
72.9
72.8
0.22
0.24
b
a
After H2S/H2 treatment and exposure to ambient conditions.
Reduction in 5%H2/Ar at 300 °C for 1 h, then cool in reducing mixture to 40 °C.
b

20
J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
-5.0
t>16 h, 60 2 kJ mol-1
D=55%
1.0
0.8
0.6
0.4
0.2
0.0
Steady state
-5.5
-6.0
-6.5
-7.0
-7.5
Pt/SBA-9.0
Pt/SBA-3.8
Pt/SBA-2.3
D=47%
t>40 h, 82 2 kJ mol-1
0
400
800
1200
1.6
1.7
1.8
1.9
Time /s
T-1 /10-3 K-1
Fig. 5. Fraction of Pt0 vs. time on stream measured by in situ XANES from 11.54 to
11.96 keV for all unsulfided Pt/SBA-15 catalysts. Time zero corresponds to the
measurement in He at 250 °C immediately following the 250 °C reduction in 3.4%
H2/He and preceding the NO oxidation reaction (300 ppm NO, 170 ppm NO2, 10%
O2, He at 250 °C).
Fig. 6. Arrhenius plot for NO oxidation on 2.4%Pt/SBA-15 (Pt/SBA-2.3) showing
changes in turnover rate (TOR) and Ea with time. TOR measured at 260–350 °C,
1 atm, and base case conditions. Platinum dispersion (D) was measured by CO
titration.
(1–3 nm) [65]. This phenomenon is not unique to Pt as it has also
been observed on catalysts with small Au particles [66]. The degree
of Pt oxidation was measured after a short (1 h) exposure to NO
oxidation conditions and resulted in values identical to those
observed by in situ XANES experiments. The fitted XANES spectra
suggest the formation of PtO on all three catalysts and PtO2 on
Pt/SBA-2.3 only. The XANES and EXAFS data are consistent with
the oxidation of Pt as evidenced by the reduction in Pt–Pt coordi-
nation number (CN) and the associated increase in CN for Pt–O. La-
ter, the samples were subjected to a long pre-treatment (Long NOx)
under oxidizing conditions. No changes in oxidation state were
detected on the samples with medium and large average particle
sizes, namely Pt/SBA-3.8 and Pt/SBA-9.0. However, a 20% increase
in PtO2 was seen on Pt/SBA-2.3. The particle size obtained from
fitting the EXAFS also showed that the catalyst did not suffer from
sintering during the time period of the experiments.
Table 5
Summary of observed changes in turnover rate (TOR) and Ea for Pt/SBA-15 catalysts as
a function of time on stream (TOS).
a
b
Catalyst
D (%) (d, nm)
TOR (10ꢀ3 sꢀ1) (Ea, kJ molꢀ1
)
TOS > 16 h
TOS > 40 h
Pt/SBA-2.3
Pt/SBA-2.3-S
Pt/SBA-3.8
Pt/SBA-9.0
47 (2.3)
49 (2.2)
29 (3.8)
12 (9.0)
3.8 0.5 (60 2)
3.7 0.3 (59 3)
126 6 (58 3)
151 1 (66 2)
1.1 0.1 (82 2)
1.2 0.1 (85 5)
38 2 (89 7)
67 6 (102 4)
a
Average Pt particle size by CO titration, calculated using d (nm) ꢂ 1.1/(Pt dis-
persion, D).
b
Turnover rate (TOR) measured at 300 °C, 1 atm and 300 ppm NO, 170 ppm NO2,
10%O2, balance N2.
4. Discussion
3.5. Instability of kinetic parameters with time on stream (TOS)
4.1. Effect of sulfur on NO oxidation
One of the reasons why we emphasized catalyst deactivation
when measuring the NO oxidation kinetics was the change in
parameters with TOS. An apparent steady state of the rate was
maintained after the fresh catalysts were stabilized for 2–4 h, but
as the exposure time increased, the catalysts showed further deac-
tivation. As presented in Fig. 6 for Pt/SBA-2.3, the rate stabilized
again after 40 h of exposure at lower levels with an apparent acti-
vation energy about 20 kJ molꢀ1 higher than previously measured.
This aging phenomenon is not dependent on the type or geometry
of the support since we also observed changes in Ea on a Pt/Al2O3
monolith catalyst (Fig. S5). In this case, our experiments indicated
that after exposure to reaction conditions for 40 h or more, the
(ꢃ30 kJ molꢀ1) increase in Ea was maintained but the turnover rate
continued to decrease with time on stream up to seven times
(t > 150 h) compared with the rate measured after 40 h of reaction.
As noted in Fig. S5, during the course of these experiments, we ob-
served less than 10% decrease in metal dispersion, which indicated
that no significant sintering occurred.
After treatment of the samples in two different S-containing
environments, the TORs after stabilization on Pt/SBA-15 catalysts
were identical within a factor of two, as shown in Table 1. All TORs
on Pt/SBA-15 compared within a factor of 5 (Table 2) to literature
values on other Pt catalysts with similar dispersions. Regarding
sulfur effects on the NO oxidation rate, previous studies
[6,12,16,17] on Pt/Al2O3 have reported a beneficial result when
adding SO2 in the feed stream. The increase in NO2 production
was explained as being a consequence of S-promoted sintering
and the resulting decrease in Pt oxide formation. This is not our
case, as the CO titration and EXAFS coordination numbers
(Table S1) suggest that no significant sintering had occurred on
Pt after the ex situ sulfur treatments. Note that CO titration was
not able to capture the expected decrease in surface area on any
of the sulfur-poisoned catalysts. Nonetheless, the dispersion values
estimated by this technique were used to normalize the rates be-
cause the catalyst operates under oxidation conditions. Other
researchers have proposed that adsorbed sulfur in the form of sul-
fite or sulfate species can withdraw charge from neighboring Pt
and thus influence to keep its reduced character [67]. For systems
that suffer from deactivation due to Pt oxidation, the addition of
sulfur may increase the stability of the catalyst at the cost of block-
ing a fraction of active sites, but we have seen no loss of surface Pt
under oxidizing reaction conditions. Additionally, we measured no
drastic changes in Ea and reaction orders with either of the two
A summary of the observed changes in Ea on the Pt/SBA-15 cat-
alysts is presented in Table 5. After a stabilization period of 16 h,
the Ea on all samples was consistently around 60–68 kJ molꢀ1
.
After deactivation for more than 40 h, Ea increased to about
90 10 molꢀ1 on all the Pt/SBA-15 catalysts. The data in Table 5
are for catalysts that showed no further deactivation during the
measurements of the activation energy.

J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
21
sulfur treatments. The changes that were observed were similar to
those for unsulfided samples, and we attribute them to catalyst
instability rather than the presence of sulfur. These results indicate
that sulfur is not playing an active role in the NO oxidation mech-
anism. Overall, the kinetic measurements suggest that under our
experimental conditions, the reaction is insensitive to the presence
of pre-adsorbed sulfur on all Pt/SBA-15 catalysts.
As pointed out in Figs. S1 and S2, the catalysts deactivated dur-
ing reaction to different degrees depending on the average Pt par-
ticle size. For example, the turnover rate after stabilization (>40 h)
was found to be about 50 times lower than the initial rate on the
set of Pt/SBA-2.3 samples with the smallest particle size (2 nm).
Correspondingly, the initial rate on the set of Pt/SBA-3.8 and Pt/
SBA-9.0 catalysts decayed about 20 times after stabilization. When
comparing the transient rates, we observed that catalyst deactiva-
tion is independent of the presence of pre-adsorbed sulfur. This
suggests that sulfur adsorbed preferentially on kinetically irrele-
vant sites, namely sites that become inactive even in the absence
of sulfur. The number of these sites is a function of the particle size
distribution of each sample. Alternatively, sulfur may also adsorb
on active sites but it is displaced by oxygen shortly after introduc-
ing the reactants such that the observed rates are effectively equal.
results in the sense that CO and H2 titration methods, which titrate
the adsorbed oxygen, were unable to differentiate the fraction of
unpoisoned Pt, while H2 chemisorption did. Both of these hypoth-
eses are in line with the idea that sulfur remained near the Pt and
was not removed from the catalysts. In fact, the EXAFS data for the
reduced Pt/SBA-2.3-S verified the presence of sulfur after the cata-
lyst was subjected to long exposures to NOx mixtures during ki-
netic measurements (Fig. S4, Table S1). This finding is relevant as
it reveals that sulfur was not fully removed from the vicinity of
Pt under oxidation conditions, yet no detrimental effect on the rate
was found.
Other researchers have observed a similar reversible behavior of
sulfur on Ni-supported catalysts when attempting to regenerate
them with oxygen. For example, Colby [69] reported that atmo-
spheric pressure O2 at 300 °C resulted in the rapid formation of a
layer of nickel oxide. A reduction treatment with H2 at 400 °C for
2 h reduced the metal and restored the sulfur layer on the surface.
The use of surface-sensitive Auger spectroscopy allowed the
authors to conclude that in the presence of high pressures (10–
760 Torr) of oxygen, the formation of oxides (in 27–527 °C range)
occurs faster than the rate of sulfur oxidation, and therefore, the
poison is trapped underneath a layer of oxide. This picture, how-
ever, is not consistent with our EXAFS results because this tech-
nique provides a volume average of the type and number of
neighboring atoms. If sulfur were to be buried below the surface,
the Pt–S contribution should still be present along with Pt–O
bonds, which is clearly not the case. Instead, the formation of Pt–
O–SOx species and/or the displacement of sulfur by oxygen to
perimeter or support sites are more reasonable explanations for
our results.
The XPS experiments on Pt/SBA-2.3-S gave further information
on the chemical state of sulfur as the catalyst was exposed to dif-
ferent environments. The Pt/SBA-2.3-S catalyst was chosen for
analysis since it had a higher Pt loading and dispersion, thus
increasing the chances of observing surface sulfur. Despite the
low concentration of sulfur, estimated from EXAFS to be about
0.06 wt.% on Pt/SBA-2.3-S, the S 2s XPS spectra shown in Fig. 4 al-
lowed the identification of three chemical states of sulfur. Based on
reported literature values for ZnSO4 [62] and PbSO4 [70], we sug-
gest that the peak located at 235 eV corresponds to sulfate species.
Similarly, we used the BE for PbSO3 at 229.2 eV [71] and the corre-
sponding one for S–Pt at 226.5 eV [61] to justify the presence of
sulfite and sulfide species. When the sample was reduced in
5%H2/Ar, the intensities of sulfite and sulfate species decreased,
resulting in an equivalent rise of intensity for reduced sulfur, as
shown in Table 3. These changes in sulfur oxidation state support
the idea of sulfur being in the region of influence of Pt; however,
the presence of S6+ after reduction opens the possibility of sulfur
adsorption on the SBA-15 support. Since the precision of the data
was insufficient to determine the fraction of sulfur species ad-
sorbed on Pt, we estimated bounds on the surface coverages
according to the measured XPS atomic ratios and which species
were assumed to adsorb on the metal.
The results presented in Table S3 indicate that sulfur species
could be covering the platinum surface in a range of 20–90%. Note
than even if all sulfur species are assumed to adsorb on Pt, hs did
not exceed 1, suggesting that under oxidation environments like
those in the NO oxidation reaction, sulfur may be bonded to Pt
as sulfites/sulfates. Although we do not have XPS evidence to dis-
tinguish the chemical state of sulfur during NO oxidation, previous
studies on a model Pt(3 5 5) surface showed that sulfur transforms
into the SO4 intermediate during oxidation (6 ꢁ 10ꢀ7 mbar O2
pressure) at 77 °C [72]. Therefore, is it feasible to propose that
the formation of sulfates under NO oxidation conditions (10%O2,
300 °C) must be favored. This picture of sulfates adsorbed on Pt
is also consistent with the EXAFS data that suggest Pt–O–SOx
4.2. EXAFS/XPS/Chemisorption analysis for sulfur
The kinetic measurements showed no apparent effect of sulfur,
but the EXAFS experiments clearly verified the presence of Pt–S
bonds on the reduced catalysts. The FT of the EXAFS raw data pre-
sented in Fig. 2 shows that both sulfur treatments were able to par-
tially sulfide Pt on Pt/SBA-2.3 and Pt/SBA-3.8. Even though Pt–S
bonds could not be resolved on the sample with the highest aver-
age Pt particle size (Pt/SBA-9.0-S), one cannot rule out the possibil-
ity of sulfur adsorption without further verification. This ambiguity
is caused by the increased magnitude of the characteristic Pt–Pt
shoulder overlapping the EXAFS from the Pt–S bonds (Fig. S3). As
presented in Table S1, the EXAFS fit confirmed the presence of
Pt–S backscattering at a distance of 2.25–2.28 Å with a first-shell
CN around 1 on the samples with 2 and 4 nm average Pt cluster
size. The Pt–S distance we observed is shorter than the one for bulk
PtS (2.31 Å) [22,29,68] or PtS2 (2.40 Å), reflecting the fact that sul-
fur is chemisorbed on metallic platinum.
The results from EXAFS on the reduced Pt/SBA-2.3-S are in line
with the XPS (Table S3) and chemisorption (Table 4) results that
suggest sulfur coverages varying from 0.6 to 1 ML. Take as an
example the absorption of sulfur on 3-fold hollow Pt sites on a
closed packed surface (Scheme S1). Within the boundaries of a unit
cell, there are four Pt atoms and four S atoms, giving a surface cov-
erage equal or close to unity. In this lattice, each Pt has three sulfur
neighbors, but only half of the Pt atoms in Pt/SBA-2.3-S are located
on the surface (50% dispersion), which brings the averaged coordi-
nation number to 1.5, a value well within the experimental error of
our XAS measurements.
Exposure to a mixture of NOx at conditions similar to the ones
used for our kinetic measurements revealed a reversible interac-
tion between sulfur and platinum. After confirming the presence
of Pt–S bonds on the catalysts after reduction, we observed that
oxidation of NOx resulted in first-shell backscattering from the
Pt–O neighbors only. Interestingly, further reduction of the sample
brought back the Pt–S bond contribution and the metallic charac-
ter of the Pt, as can be seen in Fig. 3 and Table S1. Consistent with
our kinetic measurements, two feasible interpretations of these re-
sults are (i) the formation of Pt–O–S–Ox species under oxidation
conditions, as suggested by Gracia et al. [22], and (ii) sulfur dis-
placement from surface Pt by oxygen and subsequent migration
to the metal–support interface or even to regions on the support
near Pt. This second hypothesis is supported by the chemisorption

22
J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
formation on Pt/SBA-2.3-S and Pt/SBA-3.8-S after reaction. Based
on the XPS evidence, however, we cannot exclude the possibility
that sulfur species co-exist on support sites in light that formation
of sulfates on other supports such as Al2O3 [21–23] or ZrO2 [73] has
been observed. For instance, it is possible that oxidation of sulfur to
SOx takes place on Pt followed by a partial migration of these spe-
cies to interfacial or support sites.
fur. Although the data do not provide sufficient information on
how sulfur behaves during NO oxidation, they revealed that treat-
ing Pt with sulfur on a poor S-adsorbing support does not change
the rate or the deactivation behavior caused by oxygen. These re-
sults on the effect of sulfur pointed out the importance of platinum
oxidation on the short- and long-term stability of Pt for NO oxida-
tion, as discussed in the next section.
Independent chemisorption measurements presented in Table 4
also support the hypothesis of sulfur displacement. The first indi-
cation comes from the similarities in available surface area be-
tween clean and sulfided catalysts when adsorbed oxygen was
titrated by flowing CO. Second, the fact that similar dispersion val-
ues were obtained by using H2 titration corroborates the finding
that one oxygen atom per Pt site is available for reaction, regard-
less of the presence of sulfur. For this to happen, most of the sulfur
would have to be displaced from Pt resulting in a layer of Pt–O. An-
other possibility can also be envisioned in which sulfur forms reac-
tive Pt–O–SOx species that interact with NO and NO2 and
effectively assume the same role as Pt alone in providing oxygen.
If sulfur species were taking an active role in the reaction, however,
one would expect to see changes in reaction orders and/or activa-
tion energies, which in our case were minimal. Furthermore, sur-
face crowding by SOx species might also be expected to affect the
kinetics.
After reduction, the surface area measured by H2 chemisorption
decreased on all the samples but Pt/SBA-9.0-S, thus confirming
that the titration results are not just a consequence of Pt being
resistant to sulfidation under our experimental conditions. Based
on these results, sulfur coverages were obtained by assuming that
titration of surface oxygen measures the total available Pt area,
while H2 chemisorption only counts the unpoisoned Pt sites.
According to the values of hs in Table 4, the samples with 2- and
4-nm Pt particles were poisoned to ꢃ½ ML sulfur coverage under
reducing conditions. Not surprisingly, the sample with large Pt
clusters (9 nm) showed more resistance to sulfur as the noble char-
acter of Pt is enhanced with particle size. In fact, studies of single
crystals [74] have shown that the heat of sulfur adsorption on Pt
increases with surface roughness, which is an indicative of a selec-
tive poisoning process. Interestingly, the second sulfur treatment
seemed more effective in poisoning most of the Pt surface
(ꢃ0.7 ML) at all particle sizes. If our above assumptions are correct,
the chemisorption results allowed not only an independent verifi-
cation of the presence of sulfur but also a quantitative analysis on
all of the Pt/SBA-15 samples.
In light of the strong evidence that sulfur interacts with Pt, rec-
onciling the kinetic and spectroscopic data requires that the insen-
sitivity of the NO oxidation to sulfur is caused by the greater
affinity of Pt to form oxides versus sulfides. According to our XPS
and chemisorption results, sulfur was able to at least partly cover
the surface of Pt, and yet, when that sample was exposed to NO
oxidation conditions, the transient and stabilized TORs did not cor-
relate with the expected loss in active sites. In addition, the EXAFS
data on S-treated samples after Short NOx (Table S1) not only
showed complete removal of Pt–S bonds but also indicated that
Pt oxidized to the same extent as for the unsulfided catalysts
(Table S2). If all the sulfur present after reduction formed Pt–O–
SOx species under NOx mixtures, then at least the initial rate should
have changed correspondingly. Therefore, migration of sulfur must
occur to some extent, particularly on Pt sites that are likely to
maintain their metallic character (and their activity) under NO oxi-
dation conditions. On the other hand, sulfur may remain adsorbed
on platinum sites in the form of sulfates, with increasing coverage
as the Pt particles become smaller. Small Pt clusters are character-
ized by defect sites (step/kinks/terraces) that tend to form oxides
and become inactive (or much less active), and their contribution
to the observed rate is negligible, regardless of the presence of sul-
4.3. Deactivation and instability of kinetic parameters
To address the question of deactivation caused by platinum oxi-
dation, we performed a series of in situ XANES experiments as de-
scribed in Section 2.6. The results shown in Fig. 5 clearly indicate
that Pt oxidation occurred within the first 6 min on stream. How-
ever, the subsequent stabilization of the average oxidation state
of Pt did not correlate with the continuous decay in activity ob-
served in our kinetic measurements within a similar time period
of 1500 s (Fig. 1). Differences in the results from the two experi-
ments may arise from the fact that the XAS experiments were con-
ducted on a cylindrical holder with the samples pressed into
pellets. The space between the holder and the glass tube was not
blocked, allowing most of the incoming gas flow to bypass the cat-
alyst. As the gases diffuse inside the catalyst, the NO2 concentra-
tion will increase to close to equilibrium levels, resulting in a
high driving force for Pt oxidation. On the contrary, during our ki-
netic measurements, the gas is forced to pass through the catalyst
bed, and the NOx concentrations do not change significantly as we
operate under differential conditions. In spite of the differences in
the time scale for deactivation, the data in Fig. 5 show that under
typical NO oxidation environments platinum oxides are readily
formed especially as the particle size decreases.
Given that our kinetic data required long stabilization periods
to measure reproducible values, we investigated the effect of pro-
longed time on stream (TOS) on the state of Pt by XANES. The re-
sults presented in Table S2 show that the degree of Pt oxidation
after a 1-h ex situ NOx treatment is identical to that achieved at
steady state during the in situ XANES experiments (Table S4,
Fig. 5). That the Pt oxidation state is invariant after 18-h exposure
to a NO2-rich gas mixture strongly suggests that no bulk oxidation
takes place after long TOS. Only the catalyst with average 2-nm Pt
particle showed an increase in oxidation with TOS, which is indic-
ative of its greater instability. As the surface sensitivity of XANES
decreases with increasing particle size, it is possible that the de-
gree of surface oxidation on the 4- and 9-nm Pt clusters also in-
creases with TOS, but XANES was unable to measure such
changes. What is evident is that deactivation is not related to the
formation of bulk oxides, but rather to a surface phenomenon, pos-
sibly caused by the slow oxidation of the surfaces of the relatively
larger particles since we believe that platinum oxides are inactive
compared to metal sites. In fact, recent DFT calculations [75] pre-
dict that NO oxidation on a PtO2(1 1 0) surface will exhibit a TOR
(T = 600 K, PO2 = 0.1, PNO = 3 ꢁ 10ꢀ4
,
PNO2 = 1.7 ꢁ 10ꢀ4
)
of
5.1 ꢁ 10ꢀ13 sꢀ1, which is negligible relative to their calculated
TOR of 1.9 ꢁ 10ꢀ4 sꢀ1 on a Pt(1 1 1) with low oxygen coverage
(3 ꢁ 10ꢀ6 ML). The discrepancy of two orders of magnitude in their
calculated TOR with experimental work in our group [44] arises
from the fact that only platinum sites with enough O coverage
(0.4–0.8 ML) are responsible for the observed turnover rate. How-
ever, at these O coverages, Pt is metastable to the formation of
inactive surface oxides, which results in catalyst deactivation.
The special emphasis we make on catalyst deactivation is justi-
fied by the observed changes in kinetic parameters with time on
stream. During the course of the kinetic experiments, we have
observed changes in TOR and Ea on both Pt/SBA-15 and Pt/Al2O3
catalysts (Fig. 6 and Fig. S5). As presented in Table 5, a 20- to 40-
kJ molꢀ1 increase in Ea was measured for the Pt/SBA-15 catalysts

J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
23
after aging for about 40 h on stream. We note that Ea values around
60 kJ molꢀ1 are representative of all catalysts after 16 h on stream
and that after 40 h, greater increments in activation energies were
measured with increasing Pt particle size. Since our XANES data
indicated that no oxidation of bulk Pt occurs at extended times
on stream, we believe that catalyst aging is driven by surface rear-
rangement and oxidation that only takes place over long exposure
times. Further oxidation of the surface by a fraction of a monolayer
would not be detectable by XANES. Such restructuring may change
the resistivity of particles to oxidation by creating coordinately
unsaturated sites that nucleate oxidation. As the number of such
sites increases, the effect of aging with time on stream becomes
more evident. We envision that small particles are likely to remain
oxidized as reaction conditions changes, while a fraction of larger
particles may oxidize on the surface, depending on the tempera-
ture and gas compositions. This behavior is supported by thermo-
dynamic calculations, which suggest that under our typical
operation conditions, Pt is in a metastable state between Pt with
reactive chemisorbed oxygen and PtOx [76]. Consequently, the
changes we observe with time may just be a consequence of the
continuous changes in platinum as it adjusts to varying conditions
and relaxes toward the thermodynamically favored oxide state.
supported by the US Department of Energy, Office of Science, Office
of Basic Energy Sciences, under Contract No. DE-AC02-06CH11357.
We are greatly thankful to Dr. Dmitry Zemlyanov for collecting the
XPS data and providing valuable comments on the data analysis
and Professor R.M. Rioux for providing the Pt/SBA-15 samples.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.jcat.2011.05.007.
References
[1] S. Roy, A. Baiker, Chem. Rev. 109 (2009) 4054–4091.
[2] A. Amberntsson, B. Westerberg, P. Engström, E. Fridell, M. Skoglundh, B.D.a.G.F.
Froment, Stud. Surf. Sci. Catal. 126 (1999) 317–324.
[3] P. Engström, A. Amberntsson, M. Skoglundh, E. Fridell, G. Smedler, Appl. Catal.
B 22 (1999) L241–L248.
[4] S.i. Matsumoto, Y. Ikeda, H. Suzuki, M. Ogai, N. Miyoshi, Appl. Catal. B 25 (2000)
115–124.
[5] A. Amberntsson, M. Skoglundh, M. Jonsson, E. Fridell, Catal. Today 73 (2002)
279–286.
[6] A. Amberntsson, M. Skoglundh, S. Ljungström, E. Fridell, J. Catal. 217 (2003)
253–263.
[7] J. Dawody, M. Skoglundh, L. Olsson, E. Fridell, J. Catal. 234 (2005) 206–218.
[8] E. Fridell, H. Persson, L. Olsson, B. Westerberg, A. Amberntsson, M. Skoglundh,
Top. Catal. 16 (2001) 133–137.
5. Conclusions
[9] S. Elbouazzaoui, E.C. Corbos, X. Courtois, P. Marecot, D. Duprez, Appl. Catal. B
61 (2005) 236–243.
[10] D.H. Kim, Y.H. Chin, G.G. Muntean, A. Yezeretz, N.W. Currier, W.S. Epling, H.Y.
Chen, H. Hess, C.H.F. Peden, Ind. Eng. Chem. Res. 45 (2006) 8815–8821.
[11] S. Poulston, R.R. Rajaram, Catal. Today 81 (2003) 603–610.
[12] A. Amberntsson, E. Fridell, M. Skoglundh, Appl. Catal. B 46 (2003) 429–439.
[13] J. Dawody, M. Skoglundh, E. Fridell, J. Mol. Catal. A: Chem. 209 (2004) 215–225.
[14] Y.Y. Ji, T.J. Toops, U.M. Graham, G. Jacobs, M. Crocker, Catal. Lett. 110 (2006)
29–37.
[15] O. Krocher, M. Widmer, M. Elsener, D. Rothe, Ind. Eng. Chem. Res. 48 (2009)
9847–9857.
[16] L. Olsson, H. Karlsson, Catal. Today 147 (2009) S290–S294.
[17] A. Toubeli, E.A. Efthimiadis, I.A. Vasalos, Catal. Lett. 69 (2000) 157–164.
[18] E.A. Efthimiadis, S.C. Christoforou, A.A. Nikolopoulos, I.A. Vasalos, Appl. Catal. B
22 (1999) 91–106.
[19] A.A. Nikolopoulos, E.S. Stergioula, E.A. Efthimiadis, I.A. Vasalos, Catal. Today 54
(1999) 439–450.
[20] C.C. Chang, J. Catal. 53 (1978) 374–385.
[21] P.D. Clark, N.I. Dowling, M. Huang, O. Okemona, G.D. Butlin, R. Hou, W.S.
Kijlstra, Appl. Catal. A 235 (2002) 61–69.
[22] F.J. Gracia, S. Guerrero, E.E. Wolf, J.T. Miller, A.J. Kropf, J. Catal. 233 (2005) 372–
387.
[23] J.C. Summers, Environ. Sci. Technol. 13 (1979) 321–325.
[24] H. Abdulhamid, E. Fridell, J. Dawody, M. Skoglundh, J. Catal. 241 (2006) 200–
210.
[25] C.H. Bartholomew, G.D. Weatherbee, G.A. Jarvi, J. Catal. 60 (1979) 257–269.
[26] T. Wang, A. Vazquez, A. Kato, L.D. Schmidt, J. Catal. 78 (1982) 306–318.
[27] Y. Berthier, M. Perderea, J. Oudar, Surf. Sci. 36 (1973) 225–241.
[28] W. Heegemann, K.H. Meister, E. Bechtold, K. Hayek, Surf. Sci. 49 (1975) 161–
180.
[29] J.R. Chang, S.L. Chang, T.B. Lin, J. Catal. 169 (1997) 338–346.
[30] P.J.F. Harris, Nature 323 (1986) 792–794.
[31] L.D. Schmidt, D. Luss, J. Catal. 22 (1971) 269–279.
[32] G.A. Somorjai, J. Catal. 27 (1972) 453–456.
[33] C.R. Apesteguia, C.E. Brema, T.F. Garetto, A. Borgna, J.M. Parera, J. Catal. 89
(1984) 52–59.
[34] M.V. Mathieu, M. Primet, Appl. Catal. 9 (1984) 361–370.
[35] A. Melchor, E. Garbowski, M.V. Mathieu, M. Primet, React. Kinet. Catal. Lett. 29
(1985) 371–377.
The kinetic and spectroscopic evidence presented here shows
that the presence of pre-adsorbed sulfur on Pt/SBA-15 catalysts
has no significant effect on the rate or the kinetics of the NO oxida-
tion reaction. The insensitivity of the steady turnover rates to the
presence of sulfur can be explained as a combination of two ef-
fects: (i) displacement of sulfur from Pt to interfacial or support
sites driven by Pt oxidation and (ii) selective poisoning by S on
coordinately unsaturated surfaces that are effectively inactive for
NO oxidation, even in the absence of sulfur. This argument is based
on EXAFS data that showed a complete reversal of Pt–S (observed
after reduction) to Pt–O bonds when the catalysts are exposed to
typical NO oxidation conditions. Both XPS and chemisorption mea-
surements confirmed a substantial presence of sulfur on reduced
Pt, reaching up to 0.8 ML coverage. Nevertheless, sulfur treatments
failed to affect the catalyst performance under reaction conditions,
leaving Pt oxidation as the main source of catalyst deactivation.
This knowledge can be directly applied for designing better NSR
formulations focusing on improving the selectivity of the support
toward NOx uptake (vs. SOx) and the resistivity of Pt to form oxides.
In that regard, our in situ XAS data on S-free Pt/SBA-15 and Pt/Al2O3
confirmed that the propensity to form oxides decreases with
increasing particle size and revealed that on relatively large Pt
clusters no bulk oxidation occurs with prolonged exposures to
NOx. Although the results showed an increase in Pt4+ on small par-
ticles (2 nm) after long times on stream, we conclude that those
particles become inactive quickly by forming PtO. These findings
are consistent with our theory that the active sites for NO oxida-
tion are Pt surfaces that resist oxidation, while defect sites (omni-
present on the surface of small particles) quickly deactivate and do
not turn over. Finally, we postulate that the long-term rate changes
are caused by the rearrangement of Pt surfaces on medium (4–
6 nm) and large particles (9–12 nm) possibly by additional oxida-
tion by a fraction of a monolayer that leads to the formation of
metastable surface oxides not detectable by XANES.
[36] J. Després, M. Elsener, M. Koebel, O. Kröcher, B. Schnyder, A. Wokaun, Appl.
Catal. B 50 (2004) 73–82.
[37] S.S. Mulla, N. Chen, L. Cumaranatunge, G.E. Blau, D.Y. Zemlyanov, W.N. Delgass,
W.S. Epling, F.H. Ribeiro, J. Catal. 241 (2006) 389–399.
[38] L. Olsson, E. Fridell, J. Catal. 210 (2002) 340–353.
[39] J. Segner, W. Vielhaber, G. Ertl, Isr. J. Chem. 22 (1982) 375–379.
[40] D.H. Parker, B.E. Koel, J. Vac. Sci. Technol. A 8 (1990) 2585–2590.
[41] S.S. Mulla, N. Chen, W.N. Delgass, W.S. Epling, F.H. Ribeiro, Catal. Lett. 100
(2005) 267–270.
[42] D. Bhatia, R.W. McCabe, M.P. Harold, V. Balakotaiah, J. Catal. 266 (2009) 106–
119.
Acknowledgments
[43] B.M. Weiss, E. Iglesia, J. Phys. Chem. C 113 (2009) 13331–13340.
[44] A.D. Smeltz, R.B. Getman, W.F. Schneider, F.H. Ribeiro, Catal. Today 136 (2008)
84–92.
[45] R.M. Rioux, H. Song, J.D. Hoefelmeyer, P. Yang, G.A. Somorjai, J. Phys. Chem. B
109 (2005) 2192–2202.
Support from the Department of Energy, Office of Basic Energy
Sciences, Chemical Sciences, under Grant DE-FG02-03ER15408, is
gratefully acknowledged. Use of the Advanced Photon Source was

24
J.H. Pazmiño et al. / Journal of Catalysis 282 (2011) 13–24
[46] J.E. Benson, M. Boudart, J. Catal. 4 (1965) 704–710.
[64] A.B. Chen, W.P. Zhang, X.Y. Li, D.L. Tan, X.W. Han, X.H. Bao, Catal. Lett. 119
(2007) 159–164.
[65] N. Guo, B.R. Fingland, W.D. Williams, V.F. Kispersky, J. Jelic, W.N. Delgass, F.H.
Ribeiro, R.J. Meyer, J.T. Miller, Phys. Chem. Chem. Phys. 12 (2010) 5678–5693.
[66] J.T. Miller, A.J. Kropf, Y. Zha, J.R. Regalbuto, L. Delannoy, C. Louis, E. Bus, J.A. van
Bokhoven, J. Catal. 240 (2006) 222–234.
[67] A.F. Lee, K. Wilson, R.M. Lambert, C.P. Hubbard, R.G. Hurley, R.W. McCabe, H.S.
Gandhi, J. Catal. 184 (1999) 491–498.
[68] J.R. Chang, S.L. Chang, J. Catal. 176 (1998) 42–51.
[69] S.A. Colby, Auger study of oxidative removal of sulfur from nickel, in:
University of Delaware, Newark, 1977.
[70] D.S. Zingg, D.M. Hercules, J. Phys. Chem. 82 (1978) 1992–1995.
[71] J.M. Thomas, R.H. Williams, M. Barber, I. Adams, J. Chem. Soc. Faraday Trans. 2
(68) (1972) 755.
[47] R.J. Madon, M. Boudart, Ind. Eng. Chem., Fundam. 21 (1982) 438–447.
[48] F.H.M. Dekker, A. Bliek, F. Kapteijn, J.A. Moulijn, Chem. Eng. Sci. 50 (1995)
3573–3580.
[49] R.E. Hayes, S.T. Kolaczkowski, Catal. Today 47 (1999) 295–303.
[50] D.R. Lide (Ed.), CRC Handbook of Chemistry and Physics, 91st ed., CRC Press,
Boca Raton, FL, 2011.
[51] B.L.K. Yang, H. Harold, Ind. Eng. Chem. Res. 33 (1994) 1090–1097.
[52] T. Ressler, J. Synchrotron Radiat. 5 (1998) 118–122.
[53] C.D. Wagner, W.M. Riggs, L.E. Davis, J.F. Moulder, G.E. Muilenberg, Handbook of
X-ray Photoelectron Spectroscopy, 1st ed.., Perkin-Elmer Corporation, Eden
Prairie, MN, 1979.
[54] S. Doniach, M. Sunjic, J. Phys. C: Solid State Phys. 3 (1970) 285–291.
[55] D.A. Shirley, Phys. Rev. B 5 (1972) 4709.
[56] S. Benard, L. Retailleau, F. Gaillard, P. Vernoux, A. Giroir-Fendler, Appl. Catal. B
55 (2005) 11–21.
[72] R. Streber, C. Papp, M.P.A. Lorenz, A. Bayer, R. Denecke, H.P. Steinruck, Angew.
Chem., Int. Ed. 48 (2009) 9743–9746.
[57] J.H. Lee, H.H. Kung, Catal. Lett. 51 (1998) 1–4.
[73] M.-D. Appay, J.-M. Manoli, C. Potvin, M. Muhler, U. Wild, O. Pozdnyakova, Z.
Paál, J. Catal. 222 (2004) 419–428.
[74] J. Benard, J. Oudar, N. Barbouth, E. Margot, Y. Berthier, Surf. Sci. 88 (1979) L35–
L41.
[75] H.-F. Wang, Y.-L. Guo, G. Lu, P. Hu, J. Phys. Chem. C 113 (2009) 18746–18752.
[76] R.B. Getman, Y. Xu, W.F. Schneider, J. Phys. Chem. C 112 (2008) 9559–9572.
[77] S.S. Mulla, N. Chen, L. Cumaranatunge, W.N. Delgass, W.S. Epling, F.H. Ribeiro,
Catal. Today 114 (2006) 57–63.
[78] J. Despres, M. Elsener, M. Koebel, O. Krocher, B. Schnyder, A. Wokaun, Appl.
Catal. B 50 (2004) 73–82.
[58] P.J. Schmitz, R.J. Kudla, A.R. Drews, A.E. Chen, C.K. Lowe-Ma, R.W. McCabe, W.F.
Schneider, C.T. Goralski, Appl. Catal. B 67 (2006) 246–256.
[59] Y. Xu, R.B. Getman, W.A. Shelton, W.F. Schneider, Phys. Chem. Chem. Phys. 10
(2008) 6009–6018.
[60] E. Xue, K. Seshan, J.R.H. Ross, Appl. Catal. B 11 (1996) 65–79.
[61] M. Salmeron, R.J. Koestner, E.B. Kollin, J.L. Gland, Surf. Sci. 172 (1986) 668–690.
[62] B.R. Strohmeier, D.M. Hercules, J. Catal. 86 (1984) 266–279.
[63] B.R. Strohmeier, D.E. Levden, R.S. Field, D.M. Hercules, J. Catal. 94 (1985) 514–
530.