Kinetics of selective catalytic reduction of nitric oxide by ammonia over vanadia/titania

Article abstract of DOI:10.1006/jcat.1996.0342

A kinetic model based on a previously proposed reaction scheme was used to describe reaction kinetics measurements for the selective catalytic reduction of nitric oxide by ammonia over a 6 wt% vanadia/titania catalyst in the presence of oxygen (2 mol%) at nitric oxide and ammonia concentrations from 100 to 500 ppm and at temperatures of 523 and 573 K. This reaction scheme involves adsorption of ammonia on Bronsted acid sites (V⁵⁺-OH), followed by activation of ammonia via reaction with redox sites (V=O). This activated form of ammonia reacts with gaseous or weakly adsorbed NO, producing N₂ and H₂O, and leading to partial reduction of the catalyst. The V⁴⁺-OH species formed by the selective catalytic reduction (SCR) reaction combine to form water, and the catalytic cycle is completed by reaction of the reduced sites with O₂. Water adsorbs competitively with ammonia on acid sites. This reaction scheme can be used to describe the kinetics of the SCR reaction under laboratory as well as under industrially relevant reaction conditions.

Full text of DOI:10.1006/jcat.1996.0342

JOURNAL OF CATALYSIS 163, 409–417 (1996)
ARTICLE NO. 0342
Kinetics of Selective Catalytic Reduction of Nitric Oxide by Ammonia
over Vanadia/Titania
�
J. A. Dumesic, N.-Y. Topsøe,† H. Topsøe,† Y. Chen,† and T. Slabiak†
�
Department of Chemical Engineering, University of Wisconsin, Madison, Wisconsin 53706; and †Haldor Topsøe Research Laboratories,
DK-2800, Lyngby, Denmark
Received January 29, 1996; revised April 29, 1996; accepted May 1, 1996
(
10–13)). In the present paper, we report reaction kinetics
A kinetic model based on a previously proposed reaction scheme data for the selective reduction of nitric oxide by ammonia,
was used to describe reaction kinetics measurements for the collected on catalyst powdersto minimize possible effectsof
selective catalytic reduction of nitric oxide by ammonia over a
transport limitations. These kinetic data were obtained for
6
wt% vanadia/titania catalyst in the presence of oxygen (2 mol%)
the same vanadia/titania catalysts that were employed for in
situ Fourier transform infrared (FTIR) spectroscopic stud-
ies of the SCR reaction (12, 13). This selection of catalyst
enables one to use the information obtained in these pre-
vious spectroscopic studies to provide a more quantitative
description of the performance of vanadia/titania catalysts
under various reaction conditions.
at nitric oxide and ammonia concentrations from 100 to 500 ppm
and at temperatures of 523 and 573 K. This reaction scheme
5
+
involves adsorption of ammonia on Brønsted acid sites (V –OH),
followed by activation of ammonia via reaction with redox sites
(
V == O). This activated form of ammonia reacts with gaseous or
weakly adsorbed NO, producing N2 and H2O, and leading to
4
+
partial reduction of the catalyst. The V –OH species formed
by the selective catalytic reduction (SCR) reaction combine to
form water, and the catalytic cycle is completed by reaction of the
reduced sites with O2. Water adsorbs competitively with ammonia
on acid sites. This reaction scheme can be used to describe the
kinetics of the SCR reaction under laboratory as well as under
EXPERIMENTAL AND RESULTS
The catalysts used in this study are the same as those used
previously in our in situ FTIR studies (12, 13). These ma-
industrially relevant reaction conditions.
�c 1996 Academic Press, Inc.
terials were prepared by impregnating the titania support
2
(
anatase form, surface area of 90 m /g) with an oxalic acid
solution of ammonium metavanadate, followed by drying
at 375 K and calcination at 675 K for 1 h.
INTRODUCTION
The selective catalytic reduction (SCR) of nitric oxide by
Reaction kinetics measurements were carried out at 523
ammonia over vanadia/titania catalysts is an effective pro- and 573 K under ambient pressure in a glass-lined stainless
cess for NOx emissions control from stationary sources (e.g., steel reactor. Sieved fractions (75 to 150 �m) of the sample
(
1)). One of the major issues for the industrial utilization of consisting of small amounts of catalyst diluted in an inactive
this technology is to achieve high levels of nitric oxide con- titania matrix were loaded into a reactor (length of cata-
version and to minimize emissions of unreacted ammonia lyst bed equal to 30 mm) that was subsequently attached
(
ammonia slip). It is of importance to understand the fac- to an experimental apparatus that has been described
tors that control the reaction kinetics over a wide range of previously (14).
reaction conditions, since simplified kinetic treatments may
The reactant gas with the desired concentrations of NO,
have severe limitations in situations where removal of com- NH3, and O2 was prepared in a gas manifold equipped with
pounds to the ppm level is required, e.g., ultrapurification electronic mass flow meters (Brooks) and monitored by a
or zero emission devices (2–6).
Balzers QMG 420 quadrupole mass spectrometer equipped
Recent studies have shown that the kinetic rate expres- with a heated, continuous gas inlet. The mass spectrometer
sion for the SCR reaction over vanadia/titania catalysts is data were quantitatively analyzed using the fragmentation
complex, such that the observed reaction orders with re- patterns determined experimentally from calibration gases.
spect to nitric oxide and ammonia depend on reaction con-
The catalyst samples which were loaded in the reactor
ditions (e.g., (2, 7–9)). In addition, results from recent FTIR were treated for 16 h at 673 K in 8 mol% O2/Ar at a flow
spectroscopic studies have provided important information rate of 100 ml(NTP)/min. After cooling to the desired re-
about the nature and coverages of surface species on vana- action temperature, the reactor was by-passed and the gas
dia/titania catalysts under SCR reaction conditions (e.g., was exchanged with the reactant gas mixture at a flow rate
409
0021-9517/96 $18.00
Copyright �c 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

4
10
DUMESIC ET AL.
of 200 ml(NTP)/min. Upon stabilization of the intensities
Analyses of reaction kinetics data were conducted in this
measured with the mass spectrometer, the gas mixture was studyusinggeneralregression software (GREG) and a non-
passed through the reactor and the nitric oxide conversions linear equation solver (NNES) provided by W. E. Stewart
were determined.
and R. Bain (University of Wisconsin). The procedures
Table 1 presents nitric oxide conversions (XNO exp) mea- used in these analyses are described elsewhere (6).
sured over the 6 wt% V2O5/TiO2 catalyst at 523 and 573 K
for various NO and NH3 inlet concentrations from 100 to
DISCUSSION
5
00 ppm in 2 mol% O2 at the flow rate of 200 ml(NTP)/min.
Most of these data were collected using 2.5 mg of the 6%
V2O5/TiO2 catalyst mixed with 123 mg of TiO2. Additional
experiments were conducted at lower conversions by using
Survey of Catalytic Cycles Proposed in the Literature
Takagi et al. (15, 16) proposed a Langmuir–Hinshelwood
+
1
.2 mg of the 6% V2O5/TiO2 catalyst mixed with 123 mg of reaction between adsorbed NO and adsorbed NH . How-
2
4
TiO2. Experimental ammonia conversions were also mea- ever, Inomata et al. (17) did not observe oxidation of NO to
sured; however, these values are less reliable than XNO, NO by O in dilute gas mixtures (1000 ppm NO in 1 mol%
2
2
since long times were required to reach stable readings in O ) typical of industrial reaction conditions. Furthermore,
2
the mass spectrometer due to adsorption of ammonia on Topsøe (11, 18) did not detect NO on the catalyst sur-
2
various surfaces in the measuring chamber.
face under SCR conditions. Odriozola et al. (19) consid-
ered vanadia/titania as a bifunctional catalyst and suggested
a Langmuir–Hinshelwood reaction between NO adsorbed
TABLE 1
5+
Nitric Oxide Conversions (XNO) over 6% V2O5/TiO2 at 523 and on TiO and NH adsorbed on V . Spectroscopic stud-
2
3
5
73 K for Various NO and NH3 Inlet Concentrations in 2 mol% O2 ies have revealed, however, that significant amounts of ad-
at a Flow Rate of 200 ml(NTP)/min
sorbed NO are not present during the SCR reaction (11, 18,
0, 21). Therefore, reaction kinetics measurements showing
2
Temp.
K)
Cat. wt
(mg)
NH3
(ppm)
NO
(ppm)
XNO exp
(% )
XNO fit
(% )
that the SCR reaction order with respect to NO is less than
unity (e.g., (2, 7, 8)) cannot be explained by the blocking of
surface sites by adsorbed NO species. Bosch et al. (22) pro-
posed a redox mechanism, which is supported by Janssen
(
NH3/NO
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
23
23
23
23
23
23
23
23
23
23
23
23
23
23
73
73
73
73
73
73
73
73
73
73
23
23
23
23
23
73
73
73
73
73
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
100
103
103
192
195
196
93
504
404
404
504
505
502
204
204
504
504
104
503
104
503
505
502
204
204
504
504
104
503
104
503
500
504
105
104
104
500
504
105
104
104
0.20
0.25
0.25
0.38
0.39
0.39
0.46
0.47
0.79
0.80
0.99
1.00
1.00
1.06
0.39
0.39
0.46
0.47
0.79
0.80
0.99
1.00
1.00
1.06
0.57
1.02
1.11
1.13
1.25
0.57
1.02
1.11
1.13
1.25
18
19
19
20
20
19
24
24
20
20
29
20
30
20
37
35
39
39
40
39
46
40
46
40
5
18
20
20
19
19
19
24
24
19
19
28
20
28
20
35
35
39
40
38
38
44
38
44
38
10
10
14
14
14
20
20
24
24
24
et al. (23), where V5 is first reduced by NH3 and is subse-
+
quently reoxidized by NO.
Inomata et al. (17) suggested that NO reacts with NH3
adsorbed on dual sites comprised of V OH and an adja-
cent V == O species which assists in the activation of NH3.
They found that the surface concentration of V == O was
–
95
399
403
103
503
104
535
195
196
93
proportional to the rate of reaction between NH3 and NO.
A similar mechanism was suggested by Efstathiou and
Fliatoura (9). Gasior et al. (21) proposed that V OH was
the active site. This suggestion has since been supported by
several investigators (18, 24, 25). More recently, Ramis et al.
–
(
26) proposed an oxidation/reduction mechanism in which
95
the reaction takes place between strongly adsorbed ammo-
nia and gaseous or weakly adsorbed NO. Went et al. (27)
have suggested that the SCR reaction over vanadia/titania
occurs by reaction of adsorbed NO with an activated form
of ammonia, e.g., adsorbed NH2 species.
In addition to the findingsthat adsorbed ammonia species
are predominant on vanadia/titania catalysts under SCR
reaction conditions (11, 14, 15, 17, 28), in situ IR measure-
ments (11, 28) revealed that ammonia is adsorbed on both
Brønsted and Lewis acid sites under reaction conditions.
Furthermore, ammonia adsorption was suggested (2, 13) to
be equilibrated under typical SCR reaction conditions. As
discussed above, significant amounts of adsorbed NO could
not be detected on the catalyst surface under reaction condi-
tions. The results of these spectroscopic studies and related
399
403
103
503
104
535
283
515
117
117
130
283
515
117
117
130
7
10
13
11
13
16
20
20
20

REDUCTION OF NO BY AMMONIA OVER VANADIA/TITANIA
411
kinetic information were shown by Dumesic et al. (2) to be species are the most abundant surface species under reac-
described by a catalytic cycle consisting of equilibrated am- tion conditions, (ii) Brønsted acid sites appear to be active
monia adsorption, activation of the adsorbed ammonia, and sites for the SCR reaction, (iii) the adsorption of ammonia
subsequent reaction of the activated ammonia species with on these sites appears to be an equilibrated process in the
gaseous or weakly adsorbed NO. This mechanism described SCR reaction, (iv) the activation of adsorbed ammonia by
quantitatively the NO conversion data as well as the am- reaction with surface vanadyl species appears to be neces-
monia slip for the vanadia/titania catalyst system obtained sary prior to reaction with NO, (v) significant amounts of
under industrially relevant conditions.
adsorbed NO are not present on the catalyst surface under
To obtain more detailed mechanistic insight, Topsøe reaction conditions, (vi) selective catalytic reduction of NO
et al. (12) recently employed in situ FTIR and on-line mass involves reaction of gaseous or weakly adsorbed NO with
spectrometry to study the surface species and the gaseous the activated ammonia species, producing N2 and H2O, and
reaction products when preadsorbed ammonia on vana- leading to partial reduction of the catalyst, and (vii) the
dia/titania catalysts was exposed to either NO + O2, NO, catalytic cycle is completed by reaction of the reduced sites
or O2 under temperature-programmed reaction conditions. with O2.
Lewis and Brønsted acid sites are present on the surfaces
of these catalysts, with the Brønsted acid sites being as-
Analysis of Experimental Results
sociated with V5 –OH species. Ammonia preadsorbed on
+
The spectroscopic and chemical information summa-
rized in the previous section has been used to formulate
a catalytic cycle for the SCR reaction (i.e., NO + NH3 +
vanadia/titania reacted in flowing NO + O2 or NO to form
N2. This reaction of preadsorbed ammonia with NO over
vanadia/titania produced new V–OH bands at higher fre-
1
4
3
2
O2 → N2 + H2O) in terms of acid and redox components.
quencies, related to some new reduced vanadia species.
These vanadyl bands on oxidized vanadia/titania catalysts
are converted to bands at lower frequencies upon ammonia
adsorption. The original vanadyl bands characteristic of the
oxidized catalyst are restored upon heating samples con-
taining preadsorbed ammonia in flowing NO + O2, while
this process was slower in flowing NO. These results suggest
that the surface becomes reduced upon reaction of pread-
sorbed ammonia with NO, and the catalyst is reoxidized
primarily by reaction with O2.
This cycle is shown as Scheme 1.
Step 1 is equilibrated under typical SCR reaction condi-
tions (9, 14). Step 2 is the ammonia activation step, and it is
not necessarily equilibrated. Gaseous or weakly adsorbed,
NO reacts in step 3 with surface species V–ONH3–V4 –OH
to form N2 and H2O. This step is probably not an elemen-
tary reaction, since the reverse of this step is very unlikely
to take place as written. Accordingly, we write step 3 as
an irreversible process, assuming that the reaction between
+
NO and species V–ONH3–V4 –OH involves the initial for-
mation and/or rearrangement of a highly reactive surface
complex. The removal of surface hydroxyl groups to form
water is described in step 4. In general, this may be a re-
versible step, especially at high water concentrations. At
industrial SCR reaction conditions, the catalyst surface is
essentially fully oxidized. Accordingly, the surface concen-
+
In a parallel study, Topsøe et al. (13) used in situ FTIR
and on-line mass spectrometry to monitor the surface and
catalytic properties of vanadia/titania under SCR reaction
conditions. The adsorption of ammonia was shown to be
an equilibrated step during typical SCR reaction condi-
tions. The SCR reaction rate was observed to depend on
the concentrations of surface vanadyl species and ammo-
nia adsorbed on Brønsted acid sites, suggesting that vanadyl
species react with ammonia on Brønsted acid sites to form
a surface species that subsequently reacts with nitric ox-
ide. This ammonia activation step was proposed to take
place by complete or partial H-transfer from ammonium
ions to surface vanadyl species. Again, observable amounts
of adsorbed nitric oxide species were not revealed under (1)
tration of reduced sites, V3 , should be sufficiently low to
make step 4 irreversible. Step 5 represents reoxidation of
the catalyst by O2. This step should be kinetically insignif-
icant, provided that it takes place at a rate sufficient to
+
5
+
OH → V–ONH
NH3 + V
–
4
�
reaction conditions. In agreement with the transient stud-
ies (12), it was found that the SCR reaction leads to the
production of new V–OH species associated with reduced
vanadium cations, and these species are present in signif-
icant concentrations on the catalyst under SCR reaction
conditions.
Based on the surface chemical insight provided by the
aforementioned spectroscopic studies, the following gen-
eral conclusions regarding the catalytic cycle for the SCR
reaction were proposed (10–13, 20): (i) adsorbed ammonia
→
4+
(
(
2)
V–ONH4 + V == O
V–ONH3–V –OH
�
4+
5+
3) NO + V–ONH3–V –OH → N2 + H2O + V –OH
4
+
+
V
–OH
2V4 OH
+
H2O + V + V O
==
3+
–
→
�
(4)
(5)
(6)
3+
O2 + 2V → 2V == O
5+
5+
–
H2O + V –OH
→
�
V
OH3O
SCHEME 1

4
12
DUMESIC ET AL.
maintain the catalyst in a fully oxidized state. Finally, step 6
involves the competitive adsorption of water on acid sites,
and this step is equilibrated under steady state SCR reac-
TABLE 2
Kinetic Parameters for Scheme 1 That Describe the
Kinetic Data of the Present Study
5+
tion conditions, since the species V –OH3O only appears
in this step. It should be noted that formal valence states
shown in Scheme 1 are provided for convenience to account
for reduction of the catalyst during the catalytic cycle. It is
Ei
E
� i
a
Step
Aia
(kcal/mol)
A� i
(kcal/mol)
6
13
11
1
2
3
4
5
6
8 � 10
1 � 10
0
1 � 10
1 � 10
—
20
3+
possible, for example, that a V center may react with a
11
21.6 � 2.2
5.5 � 0.1
26.3 � 0.6
0
32.1 � 5.3
neighboring V5 cation to give two V species; however,
this possible process is kinetically insignificant for the SCR
reaction in the presence of oxygen.
+
4+
4
3.6 � 0.1 � 10
—
—
—
—
1
2
1
1 � 10
8 � 10
—
—
—
—
—
In general, Scheme 1 involves 11 rate constants (all steps
are reversible with the exception of step 3), giving rise to
a
� 1
� 1
� 1
Units of sec for surface reactions and sec atm for
reactions involving gaseous (or weakly adsorbed) species.
2
2 kinetic constants (i.e., preexponential factors and activa-
tion energies). Clearly, our kinetic data are not sufficient to
allow estimates for all of these parameters, and assumptions
must be made. Most importantly, we assume reasonable val-
ues for all preexponential factors based on transition state
theory (6). These values were not adjusted further, with the
exception of the forward preexponential factor for step 3.
Since step 1 is an equilibrated process, it is only necessary to
constrain the difference between the forward and reverse
activation energies to be equal to the heat of adsorption.
Thisheat hasbeen estimated elsewhere (9, 14) to be approx-
imately equal to 20 kcal/mol for high loadings of vanadia
on titania, and this value was not adjusted further in our
analyses. Since our reaction kinetics data were collected at
low water concentrations, we can assume that step 4 is ir-
reversible and step 6 is kinetically insignificant. The rate of
O2 desorption is expected to be negligible at the low tem-
peratures of the present study, and step 5 can be assumed to
be irreversible. Finally, we assume that the surface concen-
tration of V == O sites is 10% of the surface concentration
experimental data in Table 1. The agreement between the
predicted and experimental values is very good. Impor-
tantly, the predicted values show the proper trend with inlet
NH3/NO ratio, as shown in Fig. 1. This figure presents the
experimental values of the nitric oxide conversion versus
NH3/NO ratio over 2.5 mg of catalyst at 523 and 573 K at
a constant NO concentration of about 500 ppm (circles)
and also at a constant NH3 concentration of about 100 ppm
(
squares). The predictions of Scheme 1 with the kinetic pa-
rameters of Table 2 are shown by the solid lines in Fig. 1. It
can be seen that these predictions reproduce well the ob-
served variations in nitric oxide conversion versus NH3/NO
of V5 –OH acid sites, since infrared spectroscopic results
+
indicate that the predominant surface species under SCR
reaction conditions is ammonia adsorbed on acid sites. This
assumption is arbitrary, and different values can be used
for the concentration of V == O sites, provided that the pre-
exponential factors are adjusted accordingly. The concen-
5+
tration of V –OH acid sites was assumed to be equal to
15
� 2
.
1
0
cm
In view of the above assumptions, the reaction kinetics
data were fitted using the following five adjustable para-
meters: A3, E2, E� 2, E3, and E4, where Ai, A� i, Ei, and E� i
are the forward and reverse preexponential factors (A) and
activation energies (E) for step i. The reactor was assumed
to operate as a plug-flow reactor, and the adjustment of pa-
rameters was accomplished by an optimization procedure
that minimized the sum of the squares of the differences
between the experimental and predicted nitric oxide con-
centrations in the reactor effluent. The kinetic parameters
resulting from this analysis are listed in Table 2.
FIG. 1. Nitric oxide conversion (XNO) versus inlet NH3/NO ratio over
.5 mg of 6% V2O5/TiO2 catalyst for inlet NO concentrations of approx-
imately 500 ppm at 523 K (᭹) and 573 K ( ) and for inlet NH3 concen-
trations of approximately 100 ppm at 523 K (� ) and 573 K ( ). Results
of simulations using Scheme 1 and the kinetic parameters of Table 2 are
shown as solid lines. Experimental and simulated nitric oxide conversions
2
The nitric oxide conversions predicted by Scheme 1 with
the kinetic parameters of Table 2 are compared with the are reported in more detail in Table 1.

REDUCTION OF NO BY AMMONIA OVER VANADIA/TITANIA
TABLE 3
413
ratio at constant NO concentration, at constant NH3 con-
centration, and at two different temperatures.
Kinetic Parameters for Scheme 1 That Describe the Kinetic
Data Collected under Industrially Relevant SCR Reaction
Conditions (2)
Under the reaction conditions of the present study con-
taining 2 mol% O2, the surface is essentially oxidized and
the rate constant of step 5 is kinetically insignificant. How-
ever, the rate of the SCR reaction begins to decrease when
the O2 concentration decreases below approximately 1
mol% (e.g., see Ref. (1)). Therefore, we have set the value
of the rate constant for step 5 to an appropriate value for
which the kinetic model shows a decrease in the SCR rate
at these lower O2 concentrations.
Ei
E� i
(kcal/mol)
a
Step
Aia
(kcal/mol)
A� i
6
13
11
1
2
3
4
5
6
8 � 10
1 � 10
0
1 � 10
1 � 10
—
20
32.1
—
0
—
1
1
21.6
5.5
26.3
0
4
1.3 � 0.1 � 10
1
2
7
1
4
1 � 10
8 � 10
8 � 10
1.9 � 10
—
13
0
1 � 10
16.5
Extrapolation to Industrially Relevant SCR
Reaction Conditions
a
� 1
� 1
� 1
Units of sec for surface reactions and sec atm for reactions
involving gaseous (or weakly adsorbed) species.
In a previouswork (2), we reported SCR reaction kinetics
data at two different space velocities over a 6% V2O5/TiO2
catalyst at 623 K in 4 mol% O2 and 5 mol% H2O for an inlet
NO concentration of 300 ppm and for NH3/NO ratios from
approximately 0.7 to 1.1. Under these experimental condi-
tions, the SCR reaction is partially limited by internal mass
transfer within the catalyst particles, and the effectiveness
factor was shown to be near the value of 0.3.
A simplified three-step reaction sequence, shown below
as Scheme 2, was found to describe satisfactorily the reac-
tion kinetics data of the previous study, where M is an acid
site onto which NH3 adsorbs in step 1 and S is a site on
which NH3 is activated in step 2 for subsequent reaction
with NO in step 3. Step 1 was assumed to be equilibrated,
step 2 was allowed to be reversible, and step 3 was defined
to be irreversible.
reactor. The kinetic parameters resulting from this analysis
are listed in Table 3.
The nitric oxide conversions predicted by Scheme 1 with
the kinetic parameters of Table 3 are compared with the
experimental data in Fig. 2. This figure presents the experi-
mental values of the nitric oxide conversion versus NH3/NO
ratio at space velocities of 181 l(NTP)/h/g (filled circles) and
5
8.7 l(NTP)/h/g (filled squares). The predicted values of the
nitric oxide conversion are shown by the solid lines. It can
be seen that the agreement between the predicted and ex-
perimental values is very good.
A more sensitive test of whether Scheme 1 can be used to
describe the SCR kinetics data obtained under industrially
It is now interesting to test whether Scheme 1, with ki-
netic parameters similar to those in Table 2, can be used
to describe the SCR kinetics data obtained under industri-
ally relevant conditions. In order to account for the effects
of water in steps 4 and 6, normal values of preexponential
factors were assumed for step 6, with a heat of water ad-
sorption equal to 16.5 kcal/mol (14); these values were not
adjusted further. The reverse of step 4 is also assumed to
be a nonactivated process, with a value of the preexponen-
tial factor that will be addressed further in the next section.
Finally, for simplicity, the effectiveness factor is assumed to
be equal to 0.2 at the higher space velocity and 0.3 at the
lower space velocity.
Using the above assumptions, the reaction kinetics data
of our previous study were fitted by a small adjustment
of A3. The reactor was assumed to operate as a plug-flow
→
�
(1)
(2)
(3)
NH3 + M
NH3M
FIG. 2. Nitric oxide conversion (XNO) and ammonia slip concen-
trations versus inlet NH3/NO ratio in 4 mol% O2 and 5 mol% H2O at
23 K over 6% V2O5/TiO2 catalyst for space velocities of 181 Nl/h/g (᭹)
and 58.7 Nl/h/g ( , ). Filled symbols give nitric oxide conversions and
open symbols give ammonia slip concentrations from Ref. 2. Results of
simulations using Scheme 1 and the kinetic parameters of Table 3 are
shown as solid lines.
→
�
NH3M + S
M + NH3S
6
NO + NH3S → Products
SCHEME 2

4
14
DUMESIC ET AL.
relevant conditions is provided in Fig. 2 by comparing
the effluent NH3 concentration (i.e., the ammonia slip)
from the reactor operating at the lower space velocity. The
experimental values of the ammonia slip (open squares)
are described well by the solid line representing the val-
ues predicted by Scheme 1 with the kinetic parameters of
Table 3.
Analysis of Results from Lintz and Turek
Lintz and Turek (7) conducted an extensive study of the
kinetics of nitric oxide reduction by ammonia over a vana-
dia/titania catalyst containing a high loading (20% ) of vana-
dia. In particular, thisstudyvaried the concentrationsofNO
and NH3 from approximately 100 to 1500 ppm, at water
concentrations from 0 to 10 mol% , for an oxygen concen-
tration of 4 mol% , and at temperatures of 523 and 623 K.
We have taken the empirical rate expression reported by
these authors to estimate the rate of the SCR reaction at 36
sets of experimental conditions covered in the study. Using
these results we examine whether Scheme 1 can be used to
describe their kinetic data with kinetic parameters similar
to those given in Table 3. Moreover, since the experimental
study by Lintz and Turek varied the concentration of water,
it is possible to use these data to probe the reverse rate of
step 4 and to test the validity of the equilibrium constant
estimated above for adsorption of water in step 6.
FIG. 3. Comparison of experimental SCR rates measured by Lintz
and Turek (7) over a range of NO, NH3, and H2O concentrations and
temperatures with SCR rates predicted by Scheme 1 with the kinetic pa-
rameters of Table 4.
atures studies by Lintz and Turek. The agreement between
the experimental and predicted values is seen to be satis-
factory at all conditions. In addition, it can be seen that the
kinetic parameters given in Table 4 to fit the data of Lintz
and Turek are very similar to the values given in Table 2 to
describe the kinetic data of the present study and the val-
ues given in Table 3 to describe kinetic data obtained under
industrially relevant SCR reaction conditions.
Starting with the kinetic parameters of Table 2 and the
equilibrium constant for step 6 given in Table 3, the SCR
reaction rates estimated from the work of Lintz and Turek
were fitted using the following six adjustable parameters:
A3, A� 4, E2, E� 2, E3, and E4. The concentration of acid
sites was assumed to be the same as for the high-loading
vanadia/titania catalyst employed in the present study (i.e.,
Simulation of Temperature-Programmed Reaction Studies
15
� 2
1
0 cm ). The kinetic parameters resulting from this anal-
In a previous study, we used temperature-programmed
desorption (TPD) and temperature-programmed reaction
ysis are listed in Table 4.
Figure 3 presents a comparison of the experimental and (TPR) to study NH3 desorption and the SCR reaction over
predicted rates of nitric oxide reduction by ammonia over a 6% V2O5/TiO2 catalyst (14). As mentioned above, we have
the range of NO, NH3, and H2O concentrations and temper- already used the TPD result from this previous study that
the heat of ammonia adsorption on a 6% V2O5/TiO2 cata-
lyst is approximately 20 kcal/mol. However, it is also impor-
TABLE 4
tant to test whether Scheme 1 can be used to describe the
TPR results using the kinetic parameters of Table 2 with
the water adsorption parameters given in Table 3.
Kinetic Parameters for Scheme 1 That Describe the Kinetic
Data Collected by Lintz and Turek (7)
The TPR studies involved preadsorption of NH3 on the
6% V O /TiO catalyst, followed by heating in a flow-
Ei
E� i
2
5
2
a
Step
Aia
(kcal/mol)
A� i
(kcal/mol)
ing NO/Ar gas mixture (50 ml(NTP)/min, 3890 ppm NO)
at a heating rate of 2 K/min. The concentration of NO
in the reactor effluent passed through a minimum at ca.
430 K, corresponding to the maximum rate of the SCR
reaction in this temperature-programmed experiment. In
addition, the concentration of H2O passed through a max-
imum at about 430 K, corresponding to the maximum
rate of water production and desorption in this transient
experiment.
6
13
1
2
3
4
5
6
8 � 10
1 � 10
0
21.8 � 2.2
5.5 � 0.1
26.8 � 0.6
0
1 � 10
20
1
1
11
1 � 10
31.6 � 5.3
4
6.2 � 0.1 � 10
—
—
0
1
2
7
1
4
1 � 10
8 � 10
8 � 10
1.9 � 0.1 � 10
—
—
16.5
13
0
1 � 10
a
� 1
� 1
� 1
Units of sec for surface reactions and sec atm for reactions
involving gaseous (or weakly adsorbed) species.

REDUCTION OF NO BY AMMONIA OVER VANADIA/TITANIA
415
Simulations of these TPR experiments have been con- step 4 becomes irreversible. In particular, we assume that
ducted by solving the differential equations that describe steps 1 and 6 are equilibrated, steps 3–5 are irreversible,
the changes in the gaseous pressures and the changes in the and step 2 is reversible. Furthermore, we neglect the con-
4+
surface coverages by adsorbed species that accompany the tribution of V–ONH3–V –OH species in the site balance
heating of the catalyst in a flowing gas stream (6). The re- for the V5 –OH sites, since the surface concentration of
+
actor was assumed to be well mixed in these simulations. V == O sites is much smaller than the surface concentration
Using Scheme 1 with the kinetic parameters of Table 2 and of V5 –OH acid sites. Under these conditions, the follow-
the water adsorption parameters of Table 3, the TPR simu- ing rate expression can be used to describe the rate of NO
lations predict that the concentration of NO in the reactor conversion during the SCR reaction,
effluent passes through a minimum at 450 K, while the con-
+
2
�
centration of H2O passes through a maximum at 470 K.
These temperatures are in good agreement with the exper-
imental value of approximately 430 K. Thus, Scheme 1 is
again shown to be able to capture the essential behavior of
the SCR reaction under transient reaction conditions.
2
1/2 2
Rate =
[� � (� + �S/�)
]
�
k k P K P
NH3
2
3
NO
1
�
= (
k� 2 + k3 PNO )(1 + K6 PH O + K1 PNH ) + k2 K1 PNH
2
3
3
�
�
= 16k4k5 PO
1/2
2
1/2
Approximate Rate Expression for the SCR Reaction
= (2k5 PO2 ) + (k4)
5
+
An important aspect of the experimental SCR kinetics
data of the present study and the results of Lintz and Turek
is that the reaction shows a fractional order (e.g., 0.7) with
respect to the nitric oxide pressure. This behavior is de-
scribed well by Scheme 1 and the kinetic parameters of
Tables 2–4. In the present analyses, steps 1 and 6 are equi-
librated, steps 3 and 5 are irreversible, step 2 is reversible
but not necessarily equilibrated, and step 4 is irreversible
in the absence of water and reversible in the presence of
water.
S = ratio of V == O to V –OH sites on the clean, fully
oxidized catalyst (assumed to be equal to 0.1 in
the present study),
where ki and k� i are forward and reverse rate constants,
Ki are equilibrium constants, and Pi are pressures (in atm).
Substitution of the kinetics parameters from Tables 2–4 into
these expressions gives the rate of NO consumption in units
of molecules per acid site per second.
It is not possible to derive an analytical rate expression
that is valid over the wide range of reaction conditions ex-
plored in this paper, including laboratory reactor studies
and measurements under industrially relevant SCR reac-
Dynamic Aspects of the Catalytic Cycle
for the SCR Reaction
It is useful to express Scheme 1 as a catalytic cycle in-
tion conditions. Instead, it is necessary to solve simulta- volving participation of acid sites (V5 –OH) and redox sites
neously the appropriate equations that describe the reac- (V == O). The surface chemical aspects of this catalytic cycle
tor performance and the surface steady state relations for were described elsewhere (10, 13), based on insight pro-
adsorbed species (6). However, it is possible to derive a vided by spectroscopic studies. It is now possible to use the
quite simple rate expression that is valid for the specific results of the present study to address the dynamic aspects
case where the water concentration is sufficiently low that of this catalytic cycle.
+
FIG. 4. Schematic representation of catalytic cycle for the SCR reaction over acid sites (V5+–OH) and redox sites (V == O). The corresponding steps
from Scheme 1 are indicated by the respective numbers on this diagram. Steps 1 and 6 involving the acid sites are equilibrated, steps 2 and 4 on the
redox sites are reversible (but not generally equilibrated), and steps 3 and 5 on the redox sites are irreversible.

4
16
DUMESIC ET AL.
Figure 4 presents a schematic diagram of the catalytic cy- of the present study collected over a 6% V2O5/TiO2 cata-
cle for the SCR reaction. Steps 1 and 6 take place over the lyst in a laboratory reactor. The kinetic parameters listed in
acid sites (V5 –OH), steps 3–5 occur over the redox sites Table 3 can be used to describe the nitric oxide conversion
+
(
V == O), and step 2 provides coupling between the two types and ammonia slip data collected over a 6% V2O5/TiO2 cat-
of sites. Steps 1 and 6 are equilibrated over the acid sites, alyst under industrially relevant SCR reaction conditions.
while the steps that occur over the redox sites are either The SCR reaction kineticsdata collected byLintzand Turek
irreversible (steps 3 and 5) or reversible (step 4), but they on high loadings of vanadia on titania over a wide range of
are not generally equilibrated. In addition, the process that experimental conditions can be described using the kinetic
couples the functions of the two sites (step 2) is reversible parameters listed in Table 4. Finally, the TPD and TPR data
but not equilibrated. Thus, the surface coverages on the collected over 6% V2O5/TiO2 catalyst can be described with
acid sites are controlled by surface thermodynamics (e.g., the kinetic parameters of Table 2 combined with the water
the equilibrium constants of steps 1 and 6), while the surface adsorption parameters of Table 3. The kinetic parameters
coverages on the redox sites are controlled by surface dy- in Tables 2–4 have similar values, and we conclude that
namics (e.g., the rate constants for steps 2–5). In this latter Scheme 1 provides an excellent description of the kinetics
respect, it is not appropriate to define a rate-determining of the SCR reaction.
step on the redox sites.
The results from the kinetic simulations of the present
study suggest that no single surface species is most abun-
ACKNOWLEDGMENTS
dant on the catalyst over the wide range of SCR reaction
We thank Eric T o¨ rnqvist, Bjerne Clausen, and Per Morsing (Haldor
Topsøe A/S) for their valuable help throughout this project. In addition,
we thank Randy Cortright (University of Wisconsin) for critical comments
during analysis of the results. Finally, we thank W. E. Stewart and R. Bain
for providing us with their computer software.
conditions explored. In particular, the surfaces coverages
on the acid sites by V–ONH4 and V5+–OH3O become sig-
nificant at high partial pressures of the respective gaseous
species. Moreover, the surface coverages on the redox sites
by V–ONH3–V4 –OH and V –OH appear to be signifi-
cant under all reaction conditions examined in this study.
The only surface coverage that is low under the conditions
of this study, involving gaseous O2 concentrations higher
than 1 mol% , is V3 associated with the redox sites.
+
4+
REFERENCES
1. Bosch, H., and Janssen, F. J. J. G., Catal. Today 2, 369 (1988).
2. Dumesic, J. A., Topsøe, N.-Y., Slabiak, T., Morsing, P., Clausen, B. S.,
T o¨ rnqvist, E., and Topsøe, H., New Frontiers in Catalysis, “Proceed-
ings of the 10th International Congress on Catalysis, Budapest” (L.
Guczi, F. Solymosi, and P. Tetenyi, Eds.). Akad e´ miai Kiad o´ , Budapest,
+
1
993. p. 1325.
CONCLUSIONS
3
. Rostrup-Nielsen, J. R., Schoubye, P. S., Christiansen, L. J., and Nielsen,
P. E., Chem. Eng. Sci. 49, 3995 (1994).
The present study shows that the recently proposed cata-
lytic cycle (10–13) describes quantitatively the kinetics of
the selective catalytic reduction of nitric oxide by ammonia
over vanadia/titania catalysts under a wide range of reac-
tion conditions. The catalytic cycle involves adsorption of
ammonia on acid sites (V5+–OH), activation of adsorbed
ammonia byinteraction with redoxsites(V == O), reaction of
4
. Boudart, M., Ind. Eng. Chem. Res., presented at Spring ACS Meeting,
San Diego, 1994, Symposium on Catalytic Reaction Engineering for
Environmentally Benign Processing.
5
6
. Rostrup-Nielsen, J. R., “Elementary Reaction Steps in Heterogeneous
Catalysis” (R. W. Joyner and R. A. van Santen, Eds.), pp. 441–460.
Kluwer Academic, Dordrecht, 1993.
. Dumesic, J. A., Rudd, D. F., Aparicio, L. M., Rekoske, J. E., and
Trevino, A. A., “The Microkinetics of Heterogeneous Catalysis,” Am.
Chem. Soc. Washington, D.C., 1993.
activated ammonia with gaseous, or weakly adsorbed nitric
oxide, recombination ofsurface hydroxylgroups(V4 –OH)
+
7
8
. Lintz, H. G., and Turek, T., Appl. Catal. A 85, 13 (1992).
. Robinson, W. R. A. M., van Ommen, J. G., Woldhuis, A., and Ross,
J. R. H., 10th International Congress on Catalysis, Budapest (L. Guczi,
F. Solymosi, and P. Tetenyi, Eds.). Akad e´ miai Kiad o´ , Budapest, 1993.
p. 2673.
to form water, and reoxidation ofreduced vanadium cations
3
+
(
V
) by O2. Water competes with ammonia for adsorp-
5+
tion on acid sites (V –OH). The adsorption of ammonia
on acid sites (V5 –OH) is an equilibrated process, whereas
+
9. Efstathiou, A. M., and Fliatoura, K., Appl. Catal. B 6, 35 (1995).
0. Topsøe, N.-Y., Science 265, 1217 (1994).
1. Topsøe, N.-Y., and Topsøe, H., Catal. Today 2, 77 (1991).
2. Topsøe, N.-Y., Topsøe, H., and Dumesic, J. A., J. Catal. 151, 226
1
1
1
the activation of adsorbed ammonia by reaction with re-
dox sites (V == O) is reversible but not necessarily equi-
librated. The recombination of surface hydroxyl groups
(
1995).
4+
(
V
–OH) to form water isirreversible at lowwater concen-
1
1
1
1
3. Topsøe, N.-Y., Dumesic, J. A., and Topsøe, H., J. Catal. 151, 241
(1995).
4. Srnak, T. Z., Dumesic, J. A., Clausen, B. S., T o¨ rnqvist, E., and Topsøe,
N.-Y., J. Catal. 135, 246 (1992).
trations, but it becomes reversible at higher water concen-
trations. Reaction ofnitricoxide with activated ammonia on
the surface and catalyst reoxidation by O2 are irreversible
processes.
Use of the kinetic parameters listed in Table 2 with
Scheme 1 gives a good description of the SCR kinetics data
5. Takagi, M., Kawai, T., Soma, M., Onishi, T., and Tamaru, K., J. Catal.
50, 441 (1977).
6. Takagi-Kawai, M., Kawai, T., Soma, M., Onishi, T., and Tamaru, K.,
J. Catal. 57, 528 (1979).

REDUCTION OF NO BY AMMONIA OVER VANADIA/TITANIA
417
1
1
1
7. Inomata, M., Miyamoto, A., and Murakami, Y., J. Catal. 62, 140 (1980). 23. Janssen, F. J. J. G., van den Kerkhof, F. M. G., Bosch, H., and Ross,
8. Topsøe, N., J. Catal. 128, 499 (1991).
J. R. H., J. Phys. Chem. 91, 5921 (1987).
9. Odriozola, J. A., Heinemann, H., Somorjai, G. A., Garcia de la Banda,
J. F., and Pereira, P., J. Catal. 119, 71 (1989).
24. Rajadhyaksha, R. A., and Kn o¨ zinger, H., Appl. Catal. 51, 81 (1989).
25. Chen, J. P., and Yang, R. T., J. Catal. 125, 411 (1990).
2
2
2
0. Miyamoto, A., Yamazaki, Y., Hattori, T., Inomata, M. and Murakami, 26. Ramis, G., Busca, G., Lorenzelli, V., and Forzatti, P., Appl. Catal. 64,
Y., J. Catal. 74, 144 (1982).
243 (1990).
1. Gasior, M., Haber, J., Machej, T., and Czeppe, T., J. Mol. Catal. 43, 359
27. Went, G. T., Leu, L.-J., Rosin, R. R., and Bell, A. T., J.Catal. 134, 492
(
1988).
(1992).
2. Bosch, H., Janssen, F. J. J. G., Oldenziel, J., van den Kerkhof, F. M. G., 28. Takagi, M., Kawai, T., Soma, M., Onishi, T., and Tamaru, K., J. Phys.
van Ommen, J. G., and Ross, J. R. H., Appl. Catal. 25, 239 (1986). Chem. 80, 430 (1976).