Selective Reduction of NO by NH3 over Chromia on Titania Catalyst: Investigation and Modeling of the Kinetic Behavior

Article abstract of DOI:10.1006/jcat.1997.1499

The kinetics and the parametric sensitivity of the selective catalytic reduction (SCR) of NO by NH₃ were investigated over a chromia on titania catalyst. The chromium oxide phase was made up predominantly of X-ray amorphous Cr₂O3. High SCR activity and selectivity to N₂ was attained at low temperatures. The high selectivity is attributed to the absence of significant amounts of CrO₂ and crystalline α-Cr₂O₃ which favor N₂O formation. The selectivity to N₂O increased with higher temperature. Addition of up to 6% H₂O to the dry feed reduced the rate of NO conversion and decreased the undesired formation of N₂O. The effect of water on the catalytic behavior was reversible. In the absence of oxygen, the reaction between NO and NH₃ became marginal independently whether H₂O was present or not. Small amounts of oxygen were sufficient to restore SCR activity. Admission of SO₂ to the SCR feed resulted in a severe loss of activity. The poisoning of the catalyst by SO₂ was already notable for low SO₂ concentrations (30 ppm) and for temperatures up to 573 K. X-ray photoelectron and FTIR spectroscopy revealed the presence of sulfate species on the catalyst surface. Analysis of the kinetic data indicated that the SCR reaction is first order in NO and zeroth order in NH₃ for temperatures in the range 400-520 K. The estimated activation energies for dry and wet feed amounted to 60.0 ± 1.6 kJ/mol (95% confidence limits). For temperatures in the range 400-520 K, and for a SO₂ free feed, the steady-state kinetic data could be well described with a model based on an Eley-Rideal type reaction between activated ammonia surface species and gaseous or weakly adsorbed NO.

Full text of DOI:10.1006/jcat.1997.1499

JOURNAL OF CATALYSIS 166, 356–367 (1997)
ARTICLE NO. CA971499
Selective Reduction of NO by NH3 over Chromia on Titania Catalyst:
Investigation and Modeling of the Kinetic Behavior
R. Willi, M. Maciejewski, U. Go¨bel, R. A. Ko¨ppel, and A. Baiker1,2
Department of Chemical Engineering and Industrial Chemistry, Swiss Federal Institute of Technology, ETH-Zentrum, CH-8092 Zurich, Switzerland
Received August 28, 1996; accepted October 8, 1996
Various materials have been investigated for their appli-
cation as SCR catalysts. One of the most effective catalysts
is vanadia on titania, frequently in combination with tung-
sten and/or molybdenum oxide. Other oxides have received
less attention. Recently, chromia-containing catalysts have
been reported to exhibit interesting properties for the re-
duction of nitric oxide with ammonia in the presence of ex-
cess oxygen (2, 3), especially in the low-temperature range.
Bulk amorphous chromium oxide was shown to be highly
active at low temperatures affording nitrogen with high se-
lectivity, whereas crystalline chromia (ꢀ-Cr2O3) produces
substantial amounts of nitrous oxide and also exhibits sig-
nificant activity for ammonia oxidation (4, 5). Differences
in the catalyticbehavior ofthese unsupported chromia cata-
lysts are correlated with the higher density of labile oxygen
species available on the surface of the amorphous sample
under SCR reaction conditions (6).
In studies of supported chromium oxide catalysts it has
been demonstrated that the behavior in SCR is strongly in-
fluenced by the support as well as by the nature of the chro-
mia surface species. Wong et al. (7) reported that TiO2-sup-
ported chromia catalysts were generally more active than
Al2O3-supported catalysts and that formation of undesired
N2O was higher for the titania-supported catalyst. Previous
investigations performed in our group with CrOx/TiO2
catalysts have shown that the supported chromium species
are characterized by pronounced oxidation-reduction
chemistry. The nature and oxidation state of the chromium
oxide species depends on the Cr loading, preparation
procedure, and on the pretreatment conditions (8). It was
demonstrated that the various chromium oxide phases, i.e.,
CrO2, CrOOH, and Cr2O3 differ markedly in their activity
and selectivity (9). The highest activity combined with a
high selectivity to N2 was found for X-ray amorphous Cr2O3
supported on titania. Crystallization of Cr2O3 induced
at higher temperatures reduced the activity significantly.
Similar high conversion of NO is found with CrO2/TiO2, but
substantial amounts of undesired N2O are produced with
this catalyst. Supported CrOOH is oxidized to CrO2 under
typical SCR conditions above 570 K and consequently
exhibits similar catalytic behavior as CrO2 (10, 11).
The kinetics and the parametric sensitivity of the selective cata-
lytic reduction (SCR) of NO by NH3 were investigated over a chro-
mia on titania catalyst. The chromium oxide phase was made up
predominantly of X-ray amorphous Cr2O3. High SCR activity and
selectivity to N2 was attained at low temperatures. The high selec-
tivity is attributed to the absence of significant amounts of CrO2 and
crystalline ꢀ-Cr2O3 which favor N2O formation. The selectivity to
N2O increased with higher temperature. Addition of up to 6% H2O
to the dry feed reduced the rate of NO conversion and decreased the
undesired formation of N2O. The effect of water on the catalytic be-
havior wasreversible. In the absence of oxygen, the reaction between
NO and NH3 became marginal, independently whether H2O was
present or not. Small amounts of oxygen were sufficient to restore
SCR activity. Admission of SO2 to the SCR feed resulted in a severe
loss of activity. The poisoning of the catalyst by SO2 was already
notable for low SO2 concentrations (30 ppm) and for temperatures
up to 573 K. X-ray photoelectron and FTIR spectroscopy revealed
the presence of sulfate species on the catalyst surface. Analysis of
the kinetic data indicated that the SCR reaction is first order in NO
and zeroth order in NH3 for temperatures in the range 400–520 K.
The estimated activation energies for dry and wet feed amounted
to 60.0 ꢀ 1.6 kJ/mol (95% confidence limits). For temperatures in
the range 400–520 K, and for a SO2 free feed, the steady-state ki-
netic data could be well described with a model based on an Eley–
Rideal type reaction between activated ammonia surface species
c
and gaseous or weakly adsorbed NO.
ꢁ 1997 Academic Press
INTRODUCTION
NOx isa source ofphotochemicalsmogand contributesto
acid rain. For this reason several countries, above all Japan,
the United States, and Germany, established programs for
the reduction of the NOx emissions in the early 1970s. As a
consequence, various methods of combustion control and
flue gas treatment were developed. The selective catalytic
reduction of NOx with ammonia is the most proven and
effective method for the post-combustion control of the
NOx emission (1).
1 To whom correspondence should be addressed.
2 Fax: +41-1-632 1163. E-mail: Baiker@tech.chem.ethz.ch.
356
0021-9517/97 $25.00
c
Copyright ꢁ 1997 by Academic Press
All rights of reproduction in any form reserved.

SELECTIVE REDUCTION OF NO
357
A limited number of studies have focused so far on the investigated. The influence of different reaction parame-
parametric sensitivity, the mechanism, and the kinetics of ters was studied and a microkinetic model has been devel-
the SCR reaction over chromia based catalysts. For vanadia- oped, which includes the influences of all major exhaust
based catalysts it is well known from the literature that the components.
presence of oxygen in the reaction mixture markedly en-
hances the activity for the SCR reaction (12–14), whereas
water suppresses the reduction of NO (15). Duffy et al. (16)
investigated the influence of oxygen on the rate and the
selectivity of the SCR reaction over amorphous and crys-
talline chromia. A sharp rise in activity was observed for
concentrations up to 1000 ppm O2 with both forms of chro-
mia. For higher oxygen concentrations the effect leveled
off. Willey et al. (17) similarly found higher reaction rates
in the presence of oxygen for iron oxide–chromia–alumina
aerogel catalysts.
Although flue gases usually contain water, most of the
laboratory studies of chromia based SCR catalysts have
been performed under dry conditions. The effect of water
on the reaction of NO and NH3 over amorphous and crys-
talline chromia catalysts has been investigated by isotopic
labeling experiments by Duffy et al. (5). In the presence
of excess oxygen, the addition of 1.5% H2O decreased NO
and NH3 conversions and inhibited the formation of ammo-
nia oxidation products N2O and NO for temperatures up
to 523 K over both forms of chromia. The authors verified
that the effect of water on the activity and selectivity was
reversible. Ko¨hler et al. (8) observed significantly lower ac-
tivity as well as higher selectivity to N2O for chromium ox-
ide supported on titania when the SCR feed gas contained
water.
EXPERIMENTAL
Catalyst Preparation
The catalyst was prepared by wet impregnation of TiO2
(P25, specific surface area 49 m2/g, supplier Degussa) with
chromium(III)nitrate nonahydrate (Fluka) as described in
a preceding paper (11). A chromium content of 6.84 wt%
Cr, corresponding to 10 wt% Cr2O3 was used. After drying
at 1 ꢂ 104 Pa and 363 K for 2 h and at 413 K for 8 h
the catalyst was crushed and sieved to a grain size of 180
to 300 ꢁm. In order to avoid the formation of crystalline
CrO2 upon thermal decomposition of the chromia precur-
sor, the dried catalyst was reduced with pure hydrogen
at 523 K for 1 h without precalcination, thus producing
mainly CrOOH (11). Heating in argon at 773 K for 5 h
resulted in the decomposition of CrOOH to a mixture
of poorly crystalline Cr2O3 (>95 wt% ) besides of a mi-
nor amount of undecomposed CrOOH. The BET sur-
face area determined by N2-physisorption at 77 K us-
ing a Micrometrics ASAP 2000 instrument amounted to
49 m2/g.
Apparatus
Steady-state catalytic studies were carried out in a fully
Several authors (18, 19) suggested for SCR on supported computer controlled apparatus consisting of a gas dosing
chromia catalysts a first-order kinetics in nitric oxide and unit, a continuous tubular fixed-bed microreactor, and an
zeroth order in ammonia. Comparable reaction orders in analytical unit. For further details refer to Ref. (24). The
nitric oxide and ammonia were observed for vanadia-based flows of all reactant gases were controlled by means of
catalysts and several kinetic models were proposed (20– mass flow controllers (Brooks 5850). Water was fed by
23). However, to our knowledge no model including all a step-motor pump and evaporated into the preheated
relevant exhaust gas components has been reported so far feed stream. The reaction gas mixture consisted of 200–
for chromia-based catalysts.
1000 ppm NO (99.0 vol% , PanGas), 200–1000 ppm NH3
In the present work, a 10 wt% Cr2O3/TiO2 catalyst has (99.98 vol% , PanGas), 0–6 vol% H2O (bidistilled), and
been investigated. In order to decrease the amount of 0–12 vol% O2 (99.995 vol% , PanGas) in N2-balance
CrO2 and to minimize the formation of ꢀ-Cr2O3, which (99.995 vol% PanGas). The yield of the homogeneous reac-
both are considered responsible for N2O formation, the tion of NO and O2 to NO2 was smaller than 3% as evidenced
method of catalyst preparation of Maciejewski et al. (10) by measurements with an empty reactor. The gas composi-
was modified. By directly reducing the dried sample with- tion at the reactor inlet and outlet was monitored using a
out calcination, predominantly poorly crystalline Cr2O3 quadrupole mass spectrometer (GAM 400 Balzers) and a
(>95 wt% ) was obtained. Previous studies have concen- FTIR-spectrometer (Perkin Elmer System 2000) equipped
trated on chemical and structural properties as well as on with a heatable multi-pass gas cell (Infrared Analysis, 100 cc
the catalytic behavior of different chromium oxide phases volume, 2.4 m pathlength). Standard experiments were car-
(CrO2, CrOOH, Cr2O3) in the selective catalytic reduc- ried out with a feed containing 1000 ppm NO, 1000 ppm
tion of NO by NH3 in excess oxygen. Here we are focus- NH3, 10% O2 in N2-balance using 0.35 g of catalyst (180–
ing on the performance and stability of Cr2O3/TiO2 un- 300 ꢁm sieve fraction) and a reactant flow rate of 200–
der real SCR conditions, e.g., in the presence of water in 1000 ml (NTP) minꢃ1 (60,000–300,000 hꢃ1 (NTP) GHSV)
the feed. The influence of oxygen, water, and sulfur diox- at a constant pressure of 1.1 bar in the temperature range
ide on activity and selectivity and on catalyst stability is 400–520 K.

358
WILLI ET AL.
The selectivitiesto N2 and N2O were calculated according limitation occurred in the investigated temperature range.
to Eq. [1].
Si =
The calculated maximum adiabatic temperature increase
for 1000 ppm NO in the feed was about 13 K. The mea-
sured experimental increase of the gas temperature for full
conversion at the highest space velocity amounted to 1.5 K.
Due to the negligible increase of the temperature and the
small particle size, temperature gradients in the catalyst bed
and catalyst particles were negligible.
2 ꢄ Fi
,
[1]
FNOin + F
ꢃ FNOout ꢃ F
NH3in
NH3out
where Fi is the molar flow rate of component i and in and
out refer to the reactor inlet and outlet, respectively.
Kinetic Modeling
Stability of the catalyst. The effect of time on stream
on the catalytic behavior of chromia/titania in the selec-
tive reduction of NO by NH3 with and without water in
the feed is depicted in Fig. 1 for a temperature of 473 K.
Starting with a dry feed, NO conversion and the selectivity
to N2O reached steady state within 30 min. Changing to a
feed containing 5% water resulted in an immediate sharp
drop in NO conversion from 87 to 42% . Simultaneously a
pronounced decrease of the undesired formation of N2O
from 1.9 to 0.5% was observed. NO conversion further de-
creased with time on stream to reach a new steady-state
value of 31% after 2 h, while N2O selectivity remained con-
stant. The results manifest that the formation of N2O (73%
decrease) is slightly more inhibited in the presence of wa-
ter than the reduction of NO to N2 (64% decrease). The
previous activity and selectivity of the catalyst was reestab-
lished within several minutes if water was removed from
the feed. For a catalyst being on stream for 250 h under
varying reaction conditions a decrease in catalytic activity,
combined with an increase of the formation of N2O was ob-
served. Note that the same catalyst was used for consecutive
measurements at times in the following experiments thus
leading to slightly different conversion and selectivity val-
ues for similar reaction conditions. The decrease in activity
The model parameters were estimated from the exper-
imental data by linear and nonlinear regression analysis.
For the nonlinear regression analysis the simulation soft-
ware “Simusolv” (version 3.0-120, Dow Chemical Co.) was
used. As objective function the maximum likelihood func-
tion of the NO and N2O partial pressures was chosen. The
fourth-order Runge–Kutta method was used for the numer-
ical integration and the procedure of Nelder–Mead direct
search method for maximizing the objective function.
X-Ray Photoelectron Spectroscopy (XPS)
XPS analysis of the catalyst samples was performed in a
Leybold–Heraeus LHS 11 MCD instrument using MgKꢀ
radiation (240 W) to excite photoelectrons. The analyzer
was operated at 37.8 eV constant pass energy at an energy
scale calibrated versus the Au 4f7/2 signal at 84.0 eV. Under
these conditions the full-width at half-maximum (FWHM)
ofthe Ag3d5/2 line was0.9eV. To compensate for the steady-
state charging effects, binding energies have been normal-
ized with respect to the position of the C 1s signal, resulting
from the adsorbed hydrocarbons.
Transmission FTIR Spectroscopy
The transmission FTIR spectra were recorded on a
Perkin Elmer System 2000spectrometer. Mixturesof0.5mg
catalyst sample and KBr (Fluka) were finely ground and
agglomerated under pressure (15 MPa, 180 s). The trans-
parent wafers were mounted on a special sample holder in
an environmental chamber and heated at 473 K in a flow
of dried nitrogen for 1 h. Before measurements the sample
was cooled to 323 K to avoid a broadening of the absorption
bands. Spectra were recorded with a resolution of 4 cmꢃ1 in
the range of 4000–450 cmꢃ1 accumulating 500 scans. Back-
ground spectra were recorded under identical conditions
with a pure KBr wafer.
RESULTS
Catalytic Behavior
Due to the small catalyst particle size in comparison to
the reactor diameter and a quartz wool layer in front of
the catalyst bed, plug flow was established over the whole
catalyst bed. Theoretical calculations based on the criterion
FIG. 1. NO conversion and selectivity to N2O as a function of time on
stream for a chromia/titania catalyst being alternately exposed to dry and
wet feed conditions. (ꢀ) NO conversion, (᭹) selectivity to N2O, (᭝) H2O
concentration, 0 or 5% H2O; gas flow rate = 300 ml/min(NTP), Treactor
=
of Weisz and Prater (25) indicated that no mass transfer 473 K.

SELECTIVE REDUCTION OF NO
359
FIG. 2. Conversion of NO (filled symbols) and selectivity to N2O (open symbols) as a function of temperature for different gas flow rates. (᭡,
᭝) 300 ml(NTP)/min, (᭹, ᭺) 500 ml(NTP)/min, (ꢀ, ꢁ) 700 ml(NTP)/min. Symbols indicate measured values, whereas the conversion and selectivity
curves (—) were calculated using the model described by Eqs. [4–7] and the preexponential factors k10 = 1.86E-2 m3/kg s and k20 = 2.4E-4 m3/kg s.
is thereby reflected by the variation of the corresponding and measured conversions was achieved over the whole ex-
preexponential factors k10 and k20.
perimental range with the model based on an Eley–Rideal
mechanism.
Influence of temperature. Figure 2 demonstrates the de-
pendence of the NO conversion and the selectivity to N2O
on the temperature for different gas flow rates. The sym-
bols in this and the subsequent figures indicate measured
values, whereas the curves represent the predicted conver-
sions and selectivities from an Eley–Rideal model. A de-
tailed description of the model is presented in a following
part of this paper. The temperature dependence of the NO
conversion shows that 90% NO are converted at 459 K for
a flow rate of 300 ml(NTP)/min (Fig. 2). Increasing the flow
rate resulted in an increase of the temperature to obtain
50% NO conversion from 424 K (300 ml/min) to 448 K
(700 ml/min). Note that the amounts of NO and NH3 con-
verted were equal within the experimental error for all ex-
periments (not shown). From Fig. 2 it is also apparent that
the selectivity to N2O is only influenced by the temperature
and is independent of the gas flow rate. With raising tem-
perature an increase of the formation of the undesired N2O
is observed.
Effect of water. As revealed by Fig. 1, the addition
of water substantially influenced the conversion of NO.
Figure 4A depicts the decrease in activity caused by water
concentrations of 1–6 vol% for temperatures in the range
Influence of gas flow rate. Figure 3 shows the depen-
dence of NO conversion on the ratio catalyst mass to gas
flow (W/F) for different temperatures. Increasing the W/F
ratio led to higher NO conversions for all temperatures.
Integral analysis of the data presented in Fig. 3 revealed
reaction orders of one with respect to NO and of zero with
respect to NH3. The reaction orders were confirmed by vari-
ation of the inlet concentrations of NO and NH3 in the
range of 200 to 1000 ppm. Good agreement of predicted
FIG. 3. Influence of the catalyst weight/flow ratio (W/F) on the NO
conversion for different temperatures. (ꢀ) 402 K, (᭺) 413 K, (᭡) 422 K,
(᭞) 430 K, (᭜) 443 K, (ꢁ) 452 K, (᭹) 463 K, (᭝) 473 K, (᭢) 483 K;
catalyst weight = 0.35 g, gas flow rates = 200 ml/min–800 ml(NTP)/min.
Symbols indicate measured values, whereas the conversion and selectivity
curves (—) were calculated using the model described by Eqs. [4–7] and
the preexponential factors k10 = 1.86E-2 m3/kg s and k20 = 2.4E-4 m3/kg s.

360
WILLI ET AL.
FIG. 5. Influence of oxygen concentration in the feed gas on the NO
conversion for a dry feed. (ꢀ) 420 K, (᭹) 431 K, (᭡) 440 K, (᭢) 444 K,
(᭜) 456 K; gas flow rate = 500 ml(NTP)/min. Symbols indicate measured
values, whereas the conversion and selectivity curves (—) were calculated
using the model described by Eqs. [4–7] and the preexponential factors
k10 = 1.76E-2 m3/kg s and k20 = 5.8E-4 m3/kg s.
or wet conditions were applied. Addition of a low amount
of oxygen (1500 ppm) to a dry feed already substantially
increased NO conversion, as revealed by the measurements
at 456 K (Fig. 5). The increase in NO conversion was most
pronounced for oxygen concentrations up to 2 vol% for all
temperatures, whereasfor concentrationsexceeding2vol%
O2 the catalyst exhibited an almost linear raise in activity.
No significant dependence of the selectivity to N2O on the
oxygen concentration was found.
FIG. 4. NO conversion (A) and selectivity to N2O (B) as a function
In the presence of 5 vol% water in the feed gas a
similar behavior was observed (Fig. 6), although a linear
of the water concentration in the feed gas for different temperatures. (ꢀ)
471 K, (᭹) 481 K, (᭡) 492 K, (᭢) 500 K, (᭜) 512 K; gas flow rate =
500 ml(NTP)/min. Symbols indicate measured values, whereas the con-
version and selectivity curves (—) were calculated using the model de-
scribed by Eqs. [4–7] and the preexponential factors k10 = 1.2E-2 m3/kg s
and k20 = 3.5E-4 m3/kg s.
471 to 512 K. The effect of water on NO conversion is
stronger for low temperatures and low concentrations of
H2O. The experimental data presented in Fig. 4B show that
water has not only an impact on the conversion of NO,
but also influences substantially the selectivity to nitrous
oxide. Addition of water in the range 1–6 vol% strongly
suppressed the undesired formation of N2O, with the effect
being more pronounced at higher temperatures where com-
parably high selectivities to N2O would be attained in the
absence of water. The deviation of measured values from
the model predictions is in the range of the experimental
error for both NO conversion and N2O selectivity.
FIG. 6. Influence of oxygen concentration in the feed gas on the NO
conversion for a feed containing 5% H2O. (ꢀ) 489 K, (᭹) 497 K, (᭡) 505 K,
(᭢) 520 K; gas flow rate = 500 ml(NTP)/min. Symbols indicate measured
values, whereas the conversion and selectivity curves (—) were calculated
using the model described by Eqs. [4–7] and the preexponential factors
k10 = 0.7E-2 m3/kg s and k20 = 7.5E-4 m3/kg s.
Influence of oxygen concentration. The effect of oxygen
on the catalytic behavior is presented in Figs. 5 and 6. In the
absence of oxygen, conversion of NO was negligible in the
temperature range investigated, irrespective whether dry

SELECTIVE REDUCTION OF NO
361
FIG. 7. Influence of feed gas composition on the formation of N2O and on conversion of NO and NH3, respectively. (ꢁ) NH3 conversion, (᭝) NO
conversion, (᭺) N2O concentration; (feed A): 1000 ppm NO, 1000 ppm NH3, 10% O2, 5% H2O, balance N2; (feed B): 1000 ppm NH3, 10% O2, 5% H2O,
balance N2; (feed C): 1000 ppm NO, 10% O2, 5% H2O, balance N2, gas flow rate = 500 ml(NTP)/min, Treactor = 505 K.
dependence of NO conversion on oxygen concentration decrease of the NO conversion. The degree of deactivation
was only found for higher O2 levels. Similar slopes of the lin- is a function of the SO2 concentration and exposure time.
ear raise were observed for oxygen concentrations higher During the deactivation SO2 is completely consumed. Af-
than 6% with both dry and wet feeds.
ter changing to a feed without SO2 the activity remained
stable. Increasing the SO2 concentration in steps to 60 ppm
and further to 90 ppm led to a continuous decrease in NO
conversion, while switching off SO2 resulted in stable activ-
ities. No SO3 was detected in the product stream over the
whole experimental period. Running the catalyst at 503 K
using the same feed stream without SO2 for 3 h revealed
no change in catalytic activity and the conversion remained
on the low level. Attempts were made to regenerate the
catalyst by heating at 573 K in a stream of pure nitrogen
for 3 h. During this treatment compounds such as ammo-
nium sulfates or ammonium bisulfates formed in the lower
temperature range should decompose. However, neither
SO2 nor SO3 were detected in the effluent gas stream and
a subsequent conversion measurement at 503 K showed no
increase in activity. If SO2 was added at 573 K a further
deactivation was observed. This indicated that the forma-
tion of ammonium sulfates or ammonium bisulfates cannot
exclusively explain for the poisoning effect. Moreover, no
loss of specific surface area and no significant change in
porosity occurred during the SO2 poisoning experiments.
Additional catalytic tests with the partly poisoned catalyst
in the temperature range 523 to 573 K showed that not only
the reduction of NO to N2 but also the undesired forma-
tion of N2O was affected by the poisoning. At 573 K and a
gas composition of 10% O2, 1000 ppm NO and 1000 ppm
NH3 in N2 balance 100% conversion and a selectivity of
98% to N2 were obtained. Based on the assumption of dif-
ferent sites being responsible for the formation of N2 and
Influence of the feed gas composition. The dependence
of the formation of N2O on varying the feed gas composi-
tion is illustrated in Fig. 7 together with the corresponding
conversions of NO and NH3. Feed A had a composition
of 1000 ppm NO, 1000 ppm NH3, 10% O2, and 5% H2O.
Feed B contained no nitric oxide, and feed C no ammonia,
respectively. Starting with feed A the conversions of NO
and NH3 were comparable within the experimental error.
After switching off NO (feed B), the conversion of NH3
dropped from 86 to 10% and the N2O concentration de-
creased simultaneously from 88 to 10 ppm, corresponding
to a constant selectivity to N2O of 10% . The finding shows
that the direct oxidation of ammonia by oxygen is com-
parably small at 505 K and moreover results in a similar
selectivity to N2O. In the absence of ammonia in the feed
gas (feed C) NO conversion decreased to zero and the for-
mation of N2O was completely inhibited. Presented results
obtained with feed A and feed B demonstrate that nitrogen
and nitrous oxide are predominantly produced by the reac-
tion of NH3 with NO at 505 K and that the contribution by
direct oxidation of ammonia is negligible.
Effect of SO2. The activity of the catalyst was irre-
versibly affected by the presence of sulfur dioxide in the
feed gas even at low concentrations of SO2. The poisoning
effect of sulfur dioxide as a function of the SO2 concen-
tration is illustrated in Fig. 8. At a temperature of 503 K
30 ppm of SO2 in the feed gas are sufficient for a distinct

362
WILLI ET AL.
FIG. 8. Effect of SO2 on the conversion of NO. (᭝) SO2 inlet concentration, (᭺) SO2 outlet concentration, (ꢀ) NO conversion; feed composition:
1000 ppm NO, 1000 ppm NH3, 10% O2, 5% H2O, 0–90 ppm SO2, balance N2, gas flow rate = 300 ml(NTP)/min, Treactor = 503 K.
N2O, we presume that the sites which are involved in the analysis revealed a decrease of the sulfur surface concen-
undesired formation of nitrous oxide are much stronger af- tration from 1.6 to 1.2 at.% and a concomitant increase of
fected by SO2 than sites which catalyze the reduction of NO the chromium surface concentration from 7.7 to 9.3 at.% ,
to N2.
thus indicating that the detected sulfate species could partly
originate from soluble ammonium sulfates or ammonium
bisulfates.
XPS Analysis
In order to elucidate the nature of SO2 poisoning, XPS
measurements were made either with fresh catalyst sam-
ples or with catalysts exposed to SCR conditions with and
without SO2 in the feed. Part of the samples were washed
three times with bidistilled water at a temperature of 353 K
for 8 h and subsequently dried under vacuum at 400 K.
XPS measurements were made with the washed and un-
washed samples. Figure 9 shows the sulfur 2p signal which
was only observed for the samples exposed to feeds con-
taining SO2. The binding energy of 169.1 eV indicates that
sulfur is present as a sulfate species (26, 27), but whether it is
an adsorbed species or part of an ammonium or chromium
complex is not clear. Dickinson et al. (28) reported a sulfur
2p binding energy of 169.1 eV with an asymmetric compo-
nent on the higher binding energy side for chromium sul-
fate. The binding energy of the S 2p orbital of (NH4)2SO4 is
found at 168.3 eV (29), which is substantially lower than
the value observed in the present work. Moreover, no
significant changes in the signals for the other elements
(O 1s, Cr 2p, C 1s), and especially of the N 1s signal, were
observed. In addition, the signal due to sulfur is still present
after treatment of the catalyst with water (Fig. 9B), which
should eliminate soluble ammonium sulfate. This suggests
that predominantly a sulfate type complex of chromium (or
titanium) is present on the catalyst surface. Upon treatment
FIG. 9. XP spectra of S 2p core level of (A) fresh catalyst, (C) catalyst
after SCR reaction in a feed containing 90 ppm SO2, and (B) the same
catalyst after subsequent treatment with water at 353 K for 8 h to remove
of the catalyst with water at 353 K, the semiquantitative soluble sulfate species.

SELECTIVE REDUCTION OF NO
363
FIG. 10. Transmission FTIR-spectra of a catalyst after SCR reaction with a SO2 free feed and with a feed containing 90 ppm SO2, respectively.
Transmission FTIR-Spectroscopy
typical SCR conditions over Cr2O3/TiO2.
The transmission FTIR measurements were made with
catalyst samples after use in SCR with and without SO2 in
the feed. Only one weak additional band at 1115 cmꢃ1 was
perceptible (Fig. 10) upon exposure of the catalyst to a feed
containing 90 ppm SO2. No other significant differences in
the spectra were observed. The reported ranges of the char-
acteristic absorptions of sulfates (1000–1420 cmꢃ1) and sul-
fites (980–1225 cmꢃ1) (30) include the observed band. This
is a further indication that a sulfate or sulfite type species
was formed on the catalyst surface during the exposure to
the SO2 containing SCR-feed.
4 NO + 4 NH3 + O2 → 4 N2 + 6 H2O
4 NO + 4 NH3 + 3 O2 → 4 N2O + 6 H2O
[2]
[3]
Different approaches based on Langmuir–Hinselwood
and Eley–Rideal models were investigated for setting up
the kinetic equations. The number of kinetic parameters
was minimized in order to prevent an overfitting of the
obtained results. The best description of the kinetics was
achieved with the following approach:
The reaction rate is first order in NO and zeroth order
in ammonia for the main and the side reaction.
The inhibition of the main and side reaction by water
is caused by a Langmuir adsorption of water on the
active sites.
The reoxidation of the catalyst determines the number
of the active sites and can occur with adsorbed oxygen
as well as with oxygen from the gas phase.
Water inhibits the reoxidation by a competitive Lang-
muir type adsorption.
Kinetic Modeling
The kinetic model used to describe the experimental re-
sults was based on the overall stoichiometric reactions ex-
pressed by the Eqs. [2] and [3]. Equation [2] represents the
main reaction of the selective catalytic reduction of NO for
an oxygen containing flue gas, whereas Eq. [3] describes
the side reaction responsible for the undesired formation
of nitrous oxide. In all experiments the amounts of NO and
NH3 consumed during the reaction were comparable within
the experimental error, indicating that NO reacts stoichio-
The reaction orders of NH3 and NO were established
metrically with NH3 under the reaction conditions inves- by varying the inlet concentrations of NO and NH3 in the
tigated. Regarding the stoichiometry of the side reaction, range of 200 to 1000 ppm and by integral analysis, which
the experimental results depicted in Fig. 7 suggest that in involves a comparison of the experimental datal with val-
the presence of water N2O is predominantly formed by the ues predicted by the model. The zeroth order in ammonia
stoichiometric reaction of NO with NH3 in the investigated suggests a strong adsorption of ammonia. The assumption
temperature range. However, substantial amounts of am- that the main and the side reaction have the same reaction
monia are oxidized directly to N2 and N2O in the absence orders implies that a variation of the NO or NH3 concentra-
of NO (Fig. 7). These findings support the assumption of a tion has no effect on selectivity. Figure 2 indeed shows that
reaction stoichiometry according to Eqs. [2] and [3] under the same selectivity is observed for different conversions

364
WILLI ET AL.
upon varying the gas flow rate at a constant temperature. respectively, are described by the equations
As different conversions involve different concentrations
ꢀ
ꢁ
ꢁ
KO pO
of NO and NH3, this finding confirms our assumption.
The inhibition by water could be caused by the com-
petitive adsorption of water and ammonia, but no suffi-
cient fitting was achieved under this assumption. Good
fitting was obtained on condition that part of the active
sites are covered independently by water. The tempera-
ture dependencies of the adsorption equilibrium constants
KH O and KH O were significant which allowed us to es-
2
2
r1 = k1 npO
+
2
1 + KO pO + KW pH O
2
2
2
1
ꢂ
pNO
[4]
1 + KH O pH O
2
(1)
2
ꢀ
KO pO
2
2
r2 = k2 npO
+
2
1 + KO pO + KW pH O
2
2
2
2
(1)
2
(2)
timate the heats of adsorption (1HH O ) and the preexpo-
1
2
ꢂ
pNO
[5]
[6]
nential factors (KH0 O ) for the main and the side reaction
1 + KH O pH O
2
2
(2)
2
(Table 1).
ꢀ
ꢀ
ꢁꢁ
For the reoxidation of the catalyst gaseous as well as ad-
sorbed oxygen were taken into account. This can be justi-
fied by the observed sharp increase of NO conversion in
the range of low oxygen concentrations and the linear de-
pendence in the higher O2 concentration range. Worse fit-
ting was obtained if only one type of reoxidation was con-
sidered. An empirical weighing factor was introduced to
account for the different influence on reoxidation by ei-
ther gaseous or adsorbed oxygen. The promoting effect
of oxygen has to be the same for the main and the side
reaction because the selectivity was independent of the
ꢃEA
1
1
i
ki (T ) = ki0(Tref.) exp
ꢃ
R
T
Tref.
KH O (T ) = KH0 O (Tref.
)
2
(i)
2
(i)
ꢀ
ꢀ
ꢁꢁ
ꢃ1HH O
1
1
2
(i)
ꢂ exp
ꢃ
Tref.
[7]
R
T
Tref. = 450 K.
To reduce the correlation between the preexponential
oxygen concentration. The observed inhibiting effect of factors (ki0(Tref.), K0 (Tref.)), the activation energies (EA ),
i
H2O
water on the adsorption of oxygen was considered in the and the enthalpies of adsorption (1Hi), the reaction rate
model by taking into account competitive adsorption of O2 constants (ki) and the equilibrium constant of adsorption
and H2O.
(KH O ) were expressed as proposed by Himmelblau (31).
2 (i)
In order to prevent any distortion of the parameter esti- All estimated parameters of the kinetic model are listed in
mation by the observed slow decrease of catalytic activity Table 1.
and the simultaneous slight increase of the selectivity to
N2O with time on stream, the preexponential factors were
adjusted according to the aging of the catalyst. Separate
DISCUSSION
parameter estimation with used and fresh catalysts showed
The kinetic behavior of the 10 wt% Cr2O3/TiO2 catalyst
that only the preexponential factors k10(Tref.) and k20(Tref.
)
is comparable with findings of Wong and Nobe (19) for
titania-supported vanadia and chromia catalysts for which
a reaction rate first order in NO and zeroth order in NH3
is proposed. The integral analysis of the experimental data
revealed a zeroth order reaction with respect to NH3, sug-
gesting that the catalyst is completely covered by ammonia
under reaction conditions, which is reasonable for low tem-
peratures. At higher temperatures and for low ammonia
concentrations the coverage is not complete and ammonia
adsorption has to be considered in the model (32). Based
on temperature-programmed desorption and in situ diffuse
reflectance FTIR measurements, Schneider et al. (33) pro-
posed that the presence of Brønsted bound ammonia is a
necessary requirement for the reduction of NO to N2 over
CrOx/TiO2.
changed. Based on this assumption the reaction rates for
the reduction to N2 (r1) and for the formation of N2O (r2),
TABLE 1
Estimated Parameters for Kinetic Model Described by Eqs. [4–7]
Parameter
Value
Standard deviations
Units
k10
EA
k20
EA
KH0
1.9E-2–0.7E-2
60
2.4E-4–7.5E-4
115
0.01E-2–0.06E-2
m3/kg s
kJ/mol
m3/kg s
kJ/mol
Paꢃ1
0.8
0.04E-4–0.6E-4
1.6
1
2
6.0E-4
0.7E-4
O
2
(1)
1HH
ꢃ43
4
kJ/mol
Paꢃ1
O
2
(1)
(2)
KH0
1.5E-2
0.3E-2
The observed first-order reaction with respect to NO in-
dicates that gaseous NO is involved in the rate-limiting step.
This finding would support and Eley–Rideal mechanism, as
proposed similarly for vanadia-based catalysts. The same
reaction order in NO would be expected if the adsorption
O
2
(2)
1HH
n
ꢃ77
4
kJ/mol
Paꢃ1
O
2
8.9E-5
4.8E-3
1.2E-3
0.7E-5
0.6E-3
0.3
KO
Paꢃ1
2
Kw
Paꢃ1

SELECTIVE REDUCTION OF NO
365
of NO is involved in the rate-limiting step. As opposed to reaction. However, based on our results it cannot be ex-
vanadia-based catalysts, where the adsorption of NO has cluded that NH3 is oxidized to NO in a first step, followed
only been reported under reducing conditions (34, 35), the by subsequent reaction to N2 and N2O under certain condi-
adsorption of NO on chromia was observed by several au- tions. The fact that differences in the selectivity of used and
thors. Kugler et al. (36) suggested that nitric oxide adsorbs fresh catalysts have only an influence on the preexponential
on reduced chromia surface in the form of a cis-N2O2 dimer factors indicates that the reaction to N2 and N2O, respec-
and as an NO2 chelate surface complex. The same surface tively, occur on different sites in accordance with findings by
species were observed on ꢀ-Cr2O3 but not on amorphous Schneider et al. (9, 33). The estimated adsorption enthalpy
chromia (37). On ꢀ-Cr2O3, these two species are discussed of water amounted to ꢃ43 kJ/mol for the reaction to N2, and
to be formed by oxidation of adsorbed NH3 molecules and ꢃ77 kJ/mol for the reaction to N2O, respectively, which is
are intermediates on the reaction pathway to N2O (37). a further indication of different sites being responsible for
Schneider et al. (9) observed by FTIR bands due to surface the main and the side reaction. Taking into account the es-
bound NO under reaction conditions on CrO2/TiO2, but not timated adsorption enthalpies and the observed inhibition
on Cr2O3/TiO2. The question whether adsorbed or gaseous of the NO conversion, it can be assumed that water adsorbs
NO reacts with ammonia on the surface of chromium-based stronger on a Lewis site than on a Brønsted site.
catalysts is still unresolved and no conclusive statement can
be given based on experimental evidence.
The role of water in the reversible inhibition of the SCR
reaction is not fully understood yet. Water can be involved
In analogy with supported vanadia catalysts, Schneider in the rate-limiting step or block the active sites. If the inhi-
et al. (9) observed Brønsted sites even at higher tempera- bition by water is due to a competitive adsorption between
tures for Cr2O3/TiO2 catalysts. During SCR, N2O is formed water and ammonia, the reaction order of ammonia should
byoxidation ofLewis-bound ammonia with NO and/or oxy- not equal zero in the presence of water. In the model we as-
gen present in the feed gas. In the case of Cr2O3, with no sumed kinetics of first order in NO and zeroth order in NH3
adsorbed NO, oxidation of ammonia to N2O occurs at tem- in the presence and absence of water. These assumptions
peratures higher than 450 K (9). In the absence of NO only necessitate a linear temperature dependencyofthe reaction
Lewis bound ammonia is observed above 420 K, which is rate in the Arrhenius-type plot. The data depicted in Fig. 11
oxidized to N2 and N2O (33). At higher temperatures am- confirm our assumption of the reaction orders for NO and
monia oxidation to NO is favored in the presence of oxy- NH3 in presence and absence of water in the feed gas.
gen. In our experiments we observed in the presence of
The variation of ammonia concentration in the feed gas
NO that N2O was mainly formed by the reaction of NH3 led to small changes of the NO conversion, which were
with NO (Fig. 7). The direct oxidation of NH3 by O2 to N2, in the range of the experimental error and could also be
N2O and H2O was negligible in comparison to the main caused by the observed slow deactivation of the catalyst in
FIG. 11. Arrhenius-type plot for kinetic tests under standard conditions with a feed containing 5% H2O; gas flow rate = 500 ml(NTP)/min.

366
WILLI ET AL.
the presence of water at higher temperatures. An estima- soning may be different for the sites responsible for N2 and
tion of the adsorption constant of ammonia in a statistically N2O formation, respectively. The sites which produce N2O
significant way was not possible due to the lack of accurate are mainly poisoned by SO2. XPS as well as FTIR measure-
data at low NH3 concentrations. No successful fitting was ments revealed the presence of sulfate species on the cata-
achieved with models based on an assumption of a compet- lyst surface. The affinity of these species to the Lewis acid
itive adsorption between water and ammonia.
sites seems to be stronger than to the Brønsted sites. This is
Oxygen plays a crucial role in SCR for vanadia- and in accordance with the stronger inhibition by water and the
chromia-based catalysts (38, 16). Schneider et al. (33) demo- postulated stronger adsorption of ammonia on Lewis acid
nstrated by TPD experiments that CrOx/TiO2 reduced by sites (35). Lower SCR activity upon deposition of sulfate
ammonia is inactive for the selective catalytic reduction of ions on amorphous chromia was attributed by Zhang et al.
NO. They suggested that a partially oxidized state of the (40) to strong adsorption of NH3 which prevents chromia
surface must be maintained for SCR. A marked increase from reacting with NO.
in activity and a concomitant decrease in the selectivity to
N2O upon addition of 0.1% O2 to the SCR feed was also
reported for amorphous chromia by Duffy et al. (16). In ac-
cordance with these results we found that in the absence of
oxygen the reaction rate collapsed. Below 1 vol% oxygen
in the dry feed, and below 2 vol% oxygen for a feed con-
taining 5 vol% water, a sharp raise in activity was observed
with increasing oxygen concentration. For higher oxygen
concentrations, a smaller linear increase was found (Figs. 5
and 6). The description of the kinetic conversion curves by a
power law approach, implementing a Freundlich isotherm,
was not satisfactory. By using a Temkin isotherm good fit-
ting was only achieved for the range between 2 and 10 vol%
oxygen and by assuming a pure Langmuir isotherm, only
the concentration range below 5 vol% could be fitted prop-
erly. The linear dependence of NO conversion on oxygen
concentration led us to the assumption that an additional
path of reoxidation of the catalyst takes place, which in-
volves gas-phase oxygen and which is significant for higher-
oxygen partial pressures. This assumption in combination
with the model based on a Langmuir-type adsorption re-
sulted in the best description of the oxygen dependency for
the investigated experimental range. Water in the feed gas
inhibited the path of the reoxidation via adsorbed oxygen.
The inhibition was significant and was taken into account
by a competitive adsorption of oxygen and water. The ad-
sorption equilibrium constant was about four times smaller
for water than for oxygen.
The nature of SO2 poisoning of the chromia on tita-
nia catalyst differs from fouling effects caused by ammo-
nium sulfate deposition observed on vanadia-based cata-
lysts, where the effect is reversible (39). Ammonium sul-
fate or bisulfate decomposes at higher temperatures and no
fouling is expected for reaction temperatures above 530 K.
In our case the catalyst was also poisoned at temperatures
of 573 K, as well as in the presence of low SO2 concentra-
tions. Attempts to reactivate the catalyst either by heating
in pure nitrogen at 573 K, or by washing with water failed.
Note that no significant change of the specific surface area
and the pore size distribution was perceptible after the SO2
poisoning experiments. Observed changes of the selectivity
after the exposure to SO2 indicate that the process of poi-
CONCLUSIONS
Chromia supported on titania exhibits high activity and
selectivity to N2 in the catalytic reduction of NO in the
temperature range 400–520 K. Applying a modified prepa-
ration procedure resulted in a catalyst containing >95 wt%
poorly crystalline Cr2O3, which shows significantly lower
tendency for N2O formation compared to CrO2 or crys-
talline ꢀ-Cr2O3.
The addition of water decreases NH3 and NO conversion
and increases the selectivity to N2. The effect of water on
activity and selectivity is reversible. Similar as with vanadia-
based catalysts, the activity collapsed in the absence of O2
in the feed gas. Low amounts of oxygen resulted in a sharp
increase of the reaction rate. The addition of SO2 led to
irreversible poisoning of the catalyst, which restricts appli-
cation of this catalyst to SO2 free waste gases.
With a model based on an Eley–Rideal mechanism good
prediction of the kinetic behavior was obtained for the tem-
perature range 400–520 K and for a SO2 free feed gas.
APPENDIX: NOTATION
mcat.
V ꢅ
ri
ki
KA
catalyst mass, kg
gas flow (NTP), m3/s
reaction rate, Pa m3/kg s
reaction rate constant, m3/kg s
adsorption equilibrium constant of
component A, Paꢃ1
ki0
pre-exponential factor of the reaction rate
constant, m3/kg s
KH0 O
pre-exponential factor of the adsorption
equilibrium constant of water, Paꢃ1
adsorption enthalpie of water, kJ/mol
empirical weighting factor for the reoxidation
2
(i)
1HH O
n
2
(i)
ACKNOWLEDGMENTS
Financial support of this work by the Nationaler Energie-
Forschungsfond (NEFF-Project 569) is gratefully acknowledged. The au-
thors thank H. Schneider and K. Ko¨hler for their helpful support in the
preparation of the catalyst samples.

SELECTIVE REDUCTION OF NO
367
REFERENCES
22. Tufano, V., and Turco, M., Appl. Catal. B: Environmental 2, 133
(1993).
1. Bosch, H., and Janssen, F., Catal. Today 2, 369 (1988).
2. Curry-Hyde, E., and Baiker, A., Ind. Eng. Chem. Res. 29, 1985 (1990).
3. Engweiler, J., Nickl, J., Baiker, A., Ko¨hler, K., Schla¨pfer, C. W., and
von Zelewsky, A., J. Catal. 145, 141 (1994).
23. Dumesic, J. A., Topsøe, N.-Y., Slabiak, T., Morsing, P., Clausen,
B. S., To¨ rnqvist, E., and Topsøe, H., “Proceedings, 10 International
Congress on Catalysis, Budapest, 1992” (L. Guzi et al., Eds.), p. 1325.
Akade´miai Kaido´, Budapest, 1993.
4. Curry-Hyde, H. E., Musch, H., and Baiker, A., Appl. Catal. 65, 211
(1990).
24. Willi R., Ko¨ppel R. A., and Baiker, A., Ind. Eng. Chem. Res., in press.
25. Weisz, P. B., and Prater, C. D., Adv. Catal. Rel. Subj. 6, 144 (1954).
5. Duffy, B. L., Curry-Hyde, H. E., Cant, N. W., and Nelson, P. F., J. Catal. 26. Briggs, D., and Seah, M. P., “Practical Surface Analysis by Auger and
154, 107 (1995). X-Ray Photoelectron Spectroscopy,” Wiley, Chichester, 1985.
6. Schraml-Marth, M., Wokaun, A., Curry-Hyde, H. E., and Baiker, A., 27. Wagner, C. D., et al., “Handbook of X-Ray Photoelectron Spec-
J. Catal. 133, 415 (1992).
7. Wong, W. C., and Nobe, K., Ind. Eng. Chem. Prod. Res. Dev. 25, 179
(1986).
8. Ko¨hler, K., Engweiler, J., Viebrock, H., and Baiker, A., Langmuir 11,
3423 (1995).
9. Schneider, H., Maciejewski, M., Ko¨hler, K., Wokaun, A., and Baiker,
A., J. Catal. 157, 312 (1995).
troscopy,” Perkin Elmer (Physical Electronics Division), Eden Prairie,
1978.
28. Dickinson, T., Povey, A. F., and Sherwood, P. M. A., J. Chem. Soc.
Faraday Trans. I 72, 686 (1976).
29. Barbaray, B., Contour, J. P., and Mouvier, G., Env. Sci. Technol. 12,
1294 (1978).
30. Pretsch, E., Seibl, J., Simon, W., and Clerc, P. D., “Struktur-
aufkla¨rung organischer Verbindungen mit spektroskopischen Metho-
den,” Springer-Verlag, Berlin, 1981.
10. Maciejewski, M., Ko¨hler, K., Schneider, H., and Baiker, A., J. Solid
State Chem. 119, 13 (1995).
11. Ko¨hler, K., Maciejewski, M., Schneider, H., and Baiker, A., J. Catal. 31. Himmelblau, D. M., “Process Analysis by Statistical Methods,” Wiley,
157, 301 (1995). New York, 1970.
12. Bauerle, G. L., Wu, S. C., and Nobe, K., Ind. Eng. Chem. Prod. Res. 32. Efstathiou, A. M., and Fliatoura, K., Appl. Catal. B: Environmental 6,
Dev. 17, 117 (1978).
35 (1995).
13. Inomata, M., Miyamoto, A., and Murakami, Y., J. Catal. 62, 140 (1980).
14. Jung, J., and Panagiotidis, T., Chemie im Kraftwerk 1990, 1 (1990).
15. Odenbrand, C. U. I., Gabrielsson, P. L. T., Brandin, J. G. M., and
Andersson, L. A. H., Appl. Catal. 78, 109 (1991).
33. Schneider, H., Scharf, U., Wokaun, A., and Baiker, A., J. Catal. 147,
545 (1994).
34. Odriozola, J. A., Heinemann, H., Somorjai, G. A., la Banda, J. F. G. d.,
and Pereira, P., J. Catal. 119, 71 (1989).
16. Duffy, B. L., Curry-Hyde, H. E., Cant, N. W., and Nelson, P. F., Appl.
Catal. B: Environmental 5, 133 (1994).
35. Srnak, T. Z., Dumesic, J. A., Clausen, B. S., and E., J. Catal. 135, 246
(1992).
17. Willey, R. J., Lai, H., and Peri, J. B., J. Catal. 130, 319 (1991).
18. Niiyama, H., Murata, K., and Echigoya, E., J. Catal. 48, 201 (1977).
19. Wong, W. C., and Nobe, K., Ind. Eng. Chem. Prod. Res. Dev. 25, 179
(1986).
36. Kugler, E. L., Kadet, A. B., and Gryder, J. W., J. Catal. 41, 72
(1976).
37. Schraml-Marth, M., Wokaun, A., Curry-Hyde, H. E., and Baiker, A.,
J. Catal. 134, 75 (1992).
20. Nam, I. S., Eldridge, J. W., and Kittrell, J. R., Ind. Eng. Chem. Prod.
Res. Dev. 25, 186 (1986).
38. Janssen, F. J. J. G., den Kerkhof, F. M. G. v., Bosch, H., and Ross,
J. R. H., J. Phys. Chem. 91, 5921 (1987).
21. Robinson, W. R. A. M., van Ommen, J. G., Woldhuis, A., and Ross, 39. Kittrell, J. R., Eldridge, J. W., and Conner, W. C., Catalysis 9, 126
J. R. H., “Proceedings, 10th International Congress on Catalysis,
Budapest, 1992” (L. Guzı´ et al., Eds.), p. 2673. Akade´miai Kaido´,
Budapest, 1993.
(1992).
40. Zhang, G., Buckingham, S., Curry-Hyde, H. E., and Cant, N., APCChE
3, 31 (1993).