The selective catalytic reduction of NO2 by NH3 over HZSM-5

Article abstract of DOI:10.1006/jcat.2002.3542

We have studied the reduction of NO₂ by NH₃ over HZSM-5 and find that this reaction is several hundred times faster than the corresponding reduction of NO under similar conditions. In addition, the kinetics of the reduction of NO

Full text of DOI:10.1006/jcat.2002.3542

Journal of Catalysis 208, 100–105 (2002)
doi:10.1006/jcat.2002.3542, available online at http://www.idealibrary.com on
The Selective Catalytic Reduction of NO2 by NH3 over HZSM-5
Scott A. Stevenson1 and Jim C. Vartuli
ExxonMobil Research & Engineering Company, Corporate Strategic Research, Active Materials Laboratory/Materials
Synthesis Section, 1545 Route 22 East, Annandale, New Jersey 08801-3059
Received August 31, 2001; revised January 10, 2002; accepted January 10, 2002
product gas stream, we suggested that the rate-determining
step was the oxidation of NO by O2. We proposed that
the adsorption of ammonia on the active site inhibited
the reaction and, based on the extent of this inhibition at
400◦C and above, that the active site was extra-framework
aluminum.
In order to further test these conclusions, we have stud-
ied the reduction of NO2 by NH3 under similar conditions
over the same catalyst. In this paper, we report the results
of these studies and discuss their implications for the mech-
anism of the SCR reaction over zeolites.
We have studied the reduction of NO2 by NH3 over HZSM-5
and find that this reaction is several hundred times faster than
the corresponding reduction of NO under similar conditions. In
addition, the kinetics of the reduction of NO2 are strikingly dif-
ferent from the kinetics of NO reduction. Whereas the reduc-
tion of NO is inhibited by the adsorption of ammonia and is
first order in oxygen concentration, the rate of conversion of NO2
increases as NH3 concentration increases and is independent of
oxygen concentration. We believe that these results strongly sup-
port our earlier suggestion that the oxidation of NO is the rate-
determining step in the selective catalytic reduction of NO over
c
HZSM-5.
ꢀ 2002 Elsevier Science (USA)
Key Words: selective catalytic reduction; HZSM-5; zeolite; NO;
NO2; NOx.
METHODS
The catalyst used in this study was a Mobil commercial
preparation of HZSM-5. It was synthesized hydrothermally
at approximately 100◦C using an n-propylamine template.
INTRODUCTION
The selective catalytic reduction (SCR) of nitric oxide The Si/Al molar ratio in the synthesis mixture was 27 : 1;
by ammonia is the most widely used process for the reduc- elemental analysis suggested that the actual Si/Al ratio in
tion of NO emissions from combustion flue gas (e.g., 1). the final catalyst was roughly 22 : 1. The catalyst contained
Although a variety of materials show some catalytic activ- approximately 700 ppm sodium; this gives an Al/Na ratio of
ity for this reaction, catalysts based on mixtures of vanadia about 22 : 1. The average crystal size, as measured by TEM,
and titania are currently used in virtually all commercial was estimated to be 0.50 µm. Before testing, the catalyst
SCR units. There has been significant interest, however, in was dried in situ in flowing helium for 30 min at 100◦C,
developing zeolite-based catalysts. Zeolites offer a number 2 h at 200◦C, and 2 h at 300◦C; this pretreatment was pre-
of advantages over vanadia/titania catalysts: they are active viously found to be important in obtaining reproducible
over a wider temperature range, they are more resistant to results from run to run.
thermal excursions, and the spent catalyst can present less
of a disposal problem.
Rate data were collected using the same reactor system
described in our previous studies of NO reduction (2); how-
In a previous work (2), we reported kinetic data col- ever, due to the significantly higher rate of NO2 conver-
lected for the selective catalytic reduction of NO by NH3 sion, the space velocity was increased by using less cata-
over HZSM-5 under a variety of conditions and devel- lyst and, in some cases, a higher gas flow rate. A fixed-bed
oped a kinetic equation that accurately predicted the N2 reactor operated in a downflow configuration was used; a
formation rate given the temperature and inlet NO, NH3, bypass loop allowed the measurement of feed concentra-
and O2 concentrations. Because the rate was first order in tions. HZSM-5 (0.019 g), sized to 20/40 mesh, was mixed
NO and oxygen and negative order in ammonia, and be- with 0.091 g of HiSil silica, also sized to 20/40 mesh, and
cause only small amounts of NO2 were observed in the loaded into a 3.49-mm inner diameter quartz tube reactor,
where it was held in place with quartz wool. The total bed
volume was approximately 0.41 cm3; from previous den-
1 To whom correspondence should be addressed. Current address:
SABIC Technology Center, 1600 Industrial Boulevard, Sugar Land, TX
77478. E-mail: scott.stevenson@sabicusa.com.
sity measurements, we estimate the volume of zeolite to be
approximately 0.035 cm3. Oxygen, nitrogen dioxide, and
100
0021-9517/02 $35.00
c
ꢀ 2002 Elsevier Science (USA)
All rights reserved.

KINETICS OF NO2 REDUCTION BY NH3 OVER HZSM-5
101
TABLE 1
ammonia in balance helium, supplied from gas cylinders,
were metered using mass flow controllers and mixed with
helium to give the desired inlet concentrations. The NO2
contained a small amount of NO (NO2/NO ratio approxi-
mately 63); trace amounts of nitrogen (2–10 ppm) were also
present in the feed mixture and were properly accounted
for in the total mass balance. The total gas flow was set to
either 500 or 1000 sccm, giving a space velocity based on
Activity for NO2 Reduction Compared to NO Reduction
Conversion N2 formed N2O formed
(%) (ppm) (ppm)
NO2/NO rate
enhancement
T
(◦C) NO NO2 NO NO2 NO
NO2 Unadjusted Adjusted
300
350
3.8
5.6
400 10.3
31
55
57
14
26
54
98
190
220
4
5
7
86
135
126
490
520
270
570
870
500
zeolite of either approximately 870,000 or 1,740,000 h−1
respectively.
,
Note. Conditions: inlet NO + NO2 = inlet NH3 = 500 ppm; inlet O2 =
1%; GHSV = 36,000 h−1 for NO, 1,740,000 h−1 for NO2. Adjusted rates
calculated according to kinetics described in text.
Standard inlet conditions were 500 ppm NO + NO2,
500 ppm NH3, and 1% O2; variations of these concentra-
tions were used to study the reaction kinetics. The catalyst
activity was measured at 300, 350, and 400◦C.
verted per volume of catalyst per unit time (column labeled
“unadjusted” in Table 1) suggests that the reduction of NO2
is 250–500 times faster than the reduction of NO. However,
despite the high space velocity used for the NO2 reduc-
tion studies, nondifferential conversions were observed, so
that this calculation with integral reactor data underesti-
mates the rate enhancement due to NO2 if the reaction
order in NO2 or NH3 is nonzero. Using the kinetic assump-
tions discussed below, we calculate that the reduction of
NO2 is 500–900 times faster than the reduction of NO, de-
pending on the temperature (column labeled “adjusted” in
Table 1.)
Previous reports in the literature (3–5) have indicated
that the reduction of NO2 by NH3 over H–mordenite cata-
lysts is faster than the reduction of NO, although the mag-
nitude of the rate increase seen by these authors is less than
what we observe for our HZSM-5 catalysts.
The NO2 reduction rates implied by the data presented in
Table 1 are so large that they raise the question of whether
or not mass transport limitations may influence the ob-
served rate. In fact, halving the space velocity decreases
the measured rate by 25% at 300◦C and 50% at 350◦C,
suggesting that mass transfer is limiting the rate of reac-
tion. At 350◦C, inlet concentrations of 500 ppm NH3 and
NO2, and a space velocity of 1,740,000 h−1, and using dif-
fusion coefficients estimated according to the methods of
Satterfield (6), we calculate the Damko¨hler numbers for
interphase mass transport to be 0.24 and 0.33 for ammonia
and NO2, respectively; this implies that interphase mass
transport does not significantly limit the reaction rate (7).
However, we estimate the Damko¨hler number for mass
transport in the macropores to be in the range of 1.5–5,
The catalyst was left on stream at each set of conditions
for 90 min when the total flow rate was 1000 sccm and for 3 h
when the flow rate was 500 sccm. The concentration of all
products reached steady state in 20–60 min with the excep-
tion of the case where no oxygen was fed. In this case, the
N2 formed declined slowly but steadily during the 90-min
reaction period. Effluent gas samples were recorded every
5 min. An MTI 200 Gas Chromatograph with a molecu-
lar sieve column was used to determine O2, N2, and N2O
concentrations; calibrations were made using cylinders of
known concentration supplied by Matheson. Two Siemens
NDIR analyzers were used to measure NO and NH3 levels.
NO2 concentrations were estimated from an overall mass
balance.
Some measurable conversion of NO2 and NH3 to N2 was
observed even when the gas flow bypassed the catalyst.
We believe that this conversion is due to the homogeneous
reaction of NO2 and ammonia in the heated lines leading
to the ammonia analyzer. Simple bypass tests showed that
this background conversion is approximately first order
in NO2 and half order in NH3; these results were used to
correct the amount of N2 formed over the catalyst during
kinetic testing. This correction was typically small, only
1–10% of the N2 formation rate over the catalyst. The
HiSil used to dilute the zeolite bed was also tested for
NO2 and NH3 conversion activity; its contribution to the
observed conversion was found to be negligible.
RESULTS
Activity for Reduction of NO2
Our measurements of the NO2 reduction activity of depending on the value used for the tortuosity. This sug-
HZSM-5 demonstrate that this reaction is much faster than gests that mass transport in the macropores is limiting the
the reduction of NO over the same catalyst. Table 1 com- rate (7); we estimate effectiveness factors (8) in the range
pares conversion as well as N2 and N2O formed during of 0.85–0.95 under these conditions. It is also possible that
NO2 reduction with that observed during NO reduction (2). mass transport limitations in the zeolite pores also affects
Even though the space velocity is approximately 50 times the reaction rate; however, without a reasonable value for
greater for NO2 reduction, the conversions and amounts of the rate of diffusion of NO2 in HZSM-5 it is difficult to
N2 and N2O formed are 5–10 times higher. A simple rate estimate the importance of this process. If mass transport
calculation based solely on the amount of NO2 or NO con- processes are indeed limiting the NO2 reduction rate, the

102
STEVENSON AND VARTULI
rate-enhancement factors presented in Table 1 will under- obtained from fitting the adjusted data. Because we must
estimate the increase in rate when NO2 is fed in place of NO. also contend with the fact that the ammonia and NO2 con-
One interesting difference between the reduction of NO2 centrations do not decrease in a 1:1 stoichiometry, the rate
and the reduction of NO is the much larger amount of N2O adjustments were made using the approximate form
formed when NO2 is fed. Under normal SCR conditions
ꢀ
ꢁ
2rinlet
rinlet +rexit
little NO is converted to N2O; in our previous work (2)
N2O concentrations did not exceed 10 ppm. When NO2 was
fed, however, much larger amounts of nitrous oxide were
formed, with concentrations often greater than 100 ppm.
We observed N2O/N2 ratios of approximately 0.9 at 300◦C;
this ratio decreases as the temperature increases, falling to
approximately 0.55 at 400◦C. It is interesting that this ratio
is not strongly dependent on the space velocity, suggest-
ing that little N2 is formed from the decomposition of gas
phase N2O.
Ammonia conversion was invariably higher than NO2
conversion, as based on a comparison of the amounts
of NO + NO2 and NH3 at the reactor exit. Additionally,
the amount of ammonia converted always exceeded the
amount of N2 + N2O formed by 10–20%. This may suggest
that some ammonia is oxidized to NO and NO2 under these
conditions.
NO (1–15 ppm) was observed in the product gas stream,
depending on the conditions. Comparisons with the small
amounts of inlet NO present in the NO2 feed show that
at 300◦C there is a net formation of NO amounting to
1–3 ppm, while at 350◦C there is net conversion of NO
of 1–5 ppm. Thermodynamic equilibrium for 1% O2 and
500 ppm NO + NO2 would be 165 ppm NO at 300◦C and
280 ppm at 350◦C. However, using the rate of NO oxidation
reported in our earlier work and the known equilibrium
constant we can estimate the rate of conversion of NO2 to
NO at the high space velocities used for the NO2 + NH3
reaction. We calculate that when 500 ppm NO2 is fed,
roughly 3 ppm NO2 should be converted to NO at 350◦C,
suggesting that some NO may be converted even at 300◦C.
rtrue = rmeas
,
[1]
where rtrue is the actual reaction rate at the inlet conditions,
rmeas is the measured rate, rinlet is the rate calculated from
the kinetic expression using the inlet NO2 and NH3 con-
centrations, and rexit is the rate calculated from the kinetic
expression using the NO, NO2, and NH3 concentrations
measured at the reactor exit. This expression, while not ex-
act, is a reasonably good approximation and allows us to
account for the nondifferential conversions in our integral
reactor.
Figure 1 shows the effects of varying the inlet NO2 con-
centration on the N2 + N2O formation rate. The rate of
N2 + N2O formation is positive order in NO2 concentra-
tion; the power-law rate expression used to fit the data sug-
gests that the rate is proportional to the NO2 concentration
to the 0.68 power at 300◦C and the 0.93 power at 350◦C.
This is less than the first-order dependence in NO observed
in our previous work (2). Although it is possible that the
NO2 reaction order observed at 350◦C is decreased by dif-
fusional limitations or small errors in the adjustments made
for nondifferential conversions, it appears that the rate at
300◦C is intrinsically less than first order in NO2.
Figure 2 shows the dependence of the N2 + N2O forma-
tion rate on inlet ammonia concentration. The power law fit
gives rate orders of 0.13 and 0.82 at 300 and 350◦C, respec-
tively. This result stands in marked contrast to our earlier
work using NO (2), where the rate was found to be negative
order in NH3. This suggests that although ammonia is still
adsorbed on the active site, it is no longer passively blocking
Kinetics of NO2 Reduction
In order to better understand the selective catalytic re-
duction of NO2 over HZSM-5, we performed some limited
kinetic experiments. Unfortunately this reaction is so fast
that even at the high space velocities used the NO2 and NH3
conversions were large; to accurately analyze these data,
we would need to know the true form of rate expression so
that the observed activities could be corrected to account
for the high conversions. Lacking this, the best we can do is
to make some simple assumptions concerning the approx-
imate form of the rate equation, adjust the observed rates
to compensate for the nondifferential conversions, and see
if the resulting rates justify the original assumptions. Best
results were obtained by assuming a power-law rate equa-
tion; the power-law coefficients were adjusted by a trial-
and-error process until the coefficients used to correct the
measured integral rate data to account for the nondiffer-
ential conversions were approximately the same as those
FIG. 1. Variation of N2 + N2O formation rate with inlet NO2 concen-
tration. Inlet NH3 = 500 ppm; inlet O2 = 1%; GHSV = 1,740,000 h−1
.

KINETICS OF NO2 REDUCTION BY NH3 OVER HZSM-5
103
FIG. 2. Variation of N2 + N2O formation rate with inlet NH3 concen-
FIG. 4. Variation of N2 + N2O formation rate when inlet NO2 and
tration. Inlet NO2 = 500 ppm; inlet O2 = 1%; GHSV = 1,740,000 h−1
.
NH3 are varied together. Inlet O2 = 1%; GHSV = 1,740,000 h−1
.
the active site, but participates in the rate-determining step this apparent change in order; possibilities include competi-
or in some step leading up to the rate-determining step.
tive adsorption, inaccuracies in the conversion adjustments,
The effect of oxygen concentration on the rate at 350◦C mass transfer limitations, or oversimplification of the rate
is shown in Fig. 3. We observe no change in the reaction expression.
rate as the inlet oxygen concentration is varied between
0 and 1%. This observation is in agreement with the work
of Hirsch (3), who reported that O2 is not needed to reduce
NO2 with NH3 over H–mordenite. This behavior is very
different from the effect of oxygen on the reduction of NO,
where the reaction was observed to be first order in O2 (2).
Figure 4 shows the variation in rate observed at 350◦C
when the inlet NO2 and NH3 concentrations are changed
equally together. The rate changes almost linearly with the
inlet NO2 and NH3 concentration. This behavior is some-
what unexpected from the observed NO2 and NH3 kinet-
ics discussed above; from those results we might have ex-
pected the rate to increase as the 1.7 power of the inlet NO2
and NH3 concentrations. It is not clear what the cause is of
DISCUSSION
In our previous work on the selective catalytic reduction
of NO by NH3 (2), we suggested that the rate-determining
step was the oxidation of NO to NO2. This conclusion was
based on several observations. First, the rate of reduction
of NO was first order in NO and O2 but was negative order
in ammonia, suggesting that ammonia did not participate
in the reaction until after the rate-determining step. Sec-
ond, the gas exiting the reactor contained little NO2, even
though the catalyst was shown to convert NO and O2 to a
near-equilibrium mixture of NO and NO2 under the same
conditions but in the absence of ammonia. Third, the kinet-
ics of NO oxidation to NO2 in the absence of NH3 were ob-
served to be similar to the kinetics of the SCR reaction, i.e.,
first order in NO and positive order in O2. And finally, the
forward rate constant calculated for NO oxidation agreed
well with the rate constant calculated for the SCR reaction.
We believe the data presented in this paper on the re-
duction of NO2 by NH3 support our conclusion that NO
oxidation to NO2 is the rate-determining step for the selec-
tive catalytic reduction of NO to N2. We observe that the
rate of reduction of NO2 is roughly 500 times faster than
the rate of reduction of NO or the rate of oxidation of NO
to NO2. This implies that once NO2 is formed under normal
SCR conditions it will be rapidly converted to N2.
A second important difference between the reduction
of NO and NO2 can be seen in the kinetics. The rate of NO
reduction was observed to be first order in NO but was
strongly inhibited by ammonia, suggesting that ammonia
blocked the active site by adsorption and did not participate
in any step up to and including the rate-determining step.
FIG. 3. Variation of N2 + N2O formation rate with inlet O2 concen-
tration. Inlet NO2 = NH3 = 500 ppm; GHSV = 1,740,000 h−1
.

104
STEVENSON AND VARTULI
TABLE 2
Stoichiometry of NO2 Reduction
Ratio NH3/NOx
converted
T (◦C)
Conditions
Observed stoichiometry
300
350
350
350
350
400
500 NO2 + 500 NH3 + 1% O2
500 NO2 + 500 NH3 + 1% O2
500 NO2 + 500 NH3 + 0% O2
500 NO2 + 200 NH3 + 1% O2
200 NO2 + 500 NH3 + 1% O2
500 NO2 + 500 NH3 + 1% O2
1.55 NOx + 2.21 NH3 = 1.0 N2 + 0.88 N2O
1.45 NOx + 1.97 NH3 = 1.0 N2 + 0.71 N2O
1.49 NOx + 1.89 NH3 = 1.0 N2 + 0.69 N2O
1.55 NOx + 2.14 NH3 = 1.0 N2 + 0.84 N2O
1.04 NOx + 2.09 NH3 = 1.0 N2 + 0.58 N2O
1.28 NOx + 1.86 NH3 = 1.0 N2 + 0.57 N2O
1.42
1.36
1.26
1.38
2.02
1.45
Note. GHSV = 1,740,000 h−1
.
In contrast, however, NO2 reduction is positive order in or the oxidation of ammonia to NO by NO2,
ammonia. In fact, at 350◦C the rate of N2 and N2O forma-
tion is 0.8 order in ammonia, suggesting that the activation
5NO2 + 2NH3 → 7NO + 3H2O,
[8]
[9]
of ammonia plays an important role in the reaction during
or before the rate-determining step. Likewise, the effect of
or by O2,
4NH3 + 5O2 → 4NO + 6H2O.
oxygen on the two reactions also suggests differences in the
Of course linear combinations of Eqs. [6]–[9] with reactions
[3] and [4] are also possible.
mechanism; while the reduction of NO is nearly first order
in oxygen concentration, the rate of reduction of NO2 is
unaffected by the amount of oxygen present. It appears
that the primary role of oxygen in SCR is to oxidize NO;
in the case of NO2 this function is no longer needed. The
fact that feeding NO2 instead of NO changes the reaction
kinetics suggests that it also changes the rate-determining
step by removing the primary barrier to NO reduction.
We would propose that the rate-determining step for NO
reduction is the oxidation of NO, while for the reduction of
NO2 the rate-determining step involves the combination
of adsorbed NH3 and NO2 species in some form.
Table 2 shows the average stoichiometries we observed
under several conditions. From these stoichiometries, it
seems probable that more than one reaction pathway is
involved. The ratio of N2O to N2 formed falls from 0.88
at 300◦C to 0.57 at 400◦C. Because the N2/N2O ratio is
not greatly affected by space velocity, declining only 6–
8% as the space velocity was doubled from 870,000 or
1,740,000 h−1, we do not believe that the decomposition of
N2O, Eq. [6], plays a significant role in this process; instead,
it appears that the N2/N2O ratio is determined primarily
by the kinetics of NO2 reduction. The observed decrease in
this ratio as the inlet NH3/NO2 ratio decreases is consistent
with the stoichiometries of Eqs. [3] and [4]. The fact that the
N2/N2O ratio is unaffected by the presence or absence of
molecular oxygen is in agreement with the conclusion that
O2 plays no direct role in the reaction of NH3 and NO2.
When equal amounts NO2 and NH3 are fed, approxi-
mately 1.4 mol of ammonia are consumed for every 1 mol
of NO2 consumed. Hirsch (3) observed that the conversion
of NO2 over H–mordenite was highest with an inlet ratio of
NH3/NO2 of approximately 1.3. As might be expected, the
ratio of NH3/NO2 consumed that we observed increases as
the inlet NH3/NO2 ratio increases. The amount of ammonia
consumed is too large to be accounted for by Eqs. [3] and [4]
and suggests that ammonia oxidation is taking place. In our
earlier work (2), we determined the rate of ammonia oxi-
dation occurring during the reduction of NO over the same
catalyst; this rate is much too slow to account for the ex-
cess ammonia consumed during the reduction of NO2. This
wouldseemtosuggest, therefore, thatNO2 initiatesthisoxi-
dation, although the stoichiometry indicates that even if this
is so, then molecular oxygen must be involved in some role,
perhaps to remove hydrogen from the catalyst surface or to
activate NO2 toward ammonia oxidation. It is noteworthy
While there is general agreement that the selective cata-
lytic reduction of NO occurs with a one-to-one NO/NH3
stoichiometry, i.e.,
4NO + 4NH3 + O2 → 4N2 + 6H2O,
[2]
the appropriate stoichiometry for the reduction of NO2 is
not clear. If O2 is neither consumed nor formed, there are
two independent forms (and an infinite number of linear
combinations) for the reduction of NO2 by NH3:
6NO2 + 8NH3 → 7N2 + 12H2O,
8NO2 + 6NH3 → 7N2O + 9H2O.
[3]
[4]
If combined to give a 1 : 1 ratio of ammonia and NO2 con-
sumption, a 1 : 1 ratio of N2 and N2O is also obtained:
2NO2 + 2NH3 → N2 + N2O + 3H2O.
[5]
If O2 can be activated, then other possibilities must be con-
sidered, including the decomposition of nitrous oxide,
2N2O → 2N2 + O2,
the oxidation of ammonia to N2 by O2,
4NH3 + 3O2 → 2N2 + 6H2O,
[6]
[7]

KINETICS OF NO2 REDUCTION BY NH3 OVER HZSM-5
105
that the ratio of NH3/NO2 converted falls when no oxygen Experiments with labeled NO2 and NH3 provide support
is fed. It is also interesting that this was the only feed condi- for this hypothesis (10). This reaction is probably occurring
tion where steady state was not reached quickly, suggesting under our conditions, but the amount of NO involved is so
perhaps that under these conditions the reaction was slowly small that it makes little difference to the overall rate or
removing small amounts of oxygen from the catalyst.
While the reduction of NO2 is nearly zero order in am-
monia at 300◦C, suggesting that ammonia coverage on the
active site is high, it is not far from first order at 350◦C,
suggesting that at this temperature ammonia coverage on
the active site is fairly low. This contrasts sharply with the
temperature dependence of ammonia inhibition on the re-
duction of NO, where it was inferred that ammonia cov-
erage on the active site was high even at 450◦C. This may
imply that the active site is different for the reduction of
NO2 than for the reduction of NO. Although the kinetic
data presented above are not sufficient to extract accurate
adsorption constants, estimates made from the ammonia
reaction order data, assuming that the rate is proportional
to a simple Langmuir–Hinshelwood term for adsorbed am-
monia, suggest that the ammonia coverage is roughly 90%
at 300◦C and 30% at 350◦C; this is much lower than the 85%
coverage estimated for the active site for NO reduction at
350◦C (2). For comparison, calorimetric data (9) suggest
that the ammonia coverages on the framework Brønsted
sites in the presence of 500 ppm ammonia would be 97 and
68% at 300 and 350◦C, respectively. This suggests the possi-
bility that the active site for NO2 reduction is the framework
Brønsted acid site. However, in the absence of spectro-
scopic evidence or a fully developed kinetic model, such a
suggestion is only speculation. In contrast, our earlier work
(2) indicated that the active site for the reduction of NO is
highly acidic extra-framework alumina.
As discussed above, we see a small net increase (1–3 ppm)
in the NO concentration across the reactor at 300◦C and a
small net decrease (1–5) at 350◦C. From our data on the
rate of NO oxidation and the known equilibrium constant
for NO oxidation, we have further estimated that under
these conditions approximately 3 ppm NO2 should be cata-
lytically converted into NO; the extent of this reaction is
limited by the very high space velocity used. Given the low
conversion of NO2 to NO expected, we cannot support the
hypothesis of Andersson et al. (4) that this reaction is rate
determining for NO2 reduction. Under our conditions, it is
clear that some NO is reduced to N2 or N2O at 350◦C, and
possibly also at 300◦C. Hirsch (3) observed that in the pres-
ence of excess NH3 mixtures of NO and NO2 are converted
with a 1 : 1 stoichiometry of NO2/NO. Likewise, it has been
reported (4, 5) that the reduction of NOx over H–mordenite
was faster when the NO2/NO ratio was approximately 1 : 1
than it was when either NO or NO2 was fed alone, leading to
the suggestion that NO and NO2 combined with ammonia
to give N2:
kinetics of the process.
SUMMARY
The reduction of NO2 by NH3 over HZSM-5 is two to
three orders of magnitude faster than the reduction of
NO by NH3 under the same conditions. The reduction
of NO2 is so fast that high conversions are obtained
even at gas hourly space velocities of 800,000 h−1 and
greater and that mass transfer may limit the observed
rate. The reaction order in NO2 is 0.68 and 0.93 at 300
and 350◦C, respectively; the reaction order in NH3 is
0.13 and 0.82 at 300 and 350◦C, respectively. The reaction
is zeroth order in oxygen. Large amounts of N2O are
formed under all conditions studied; the N2O/N2 ratio
decreases from approximately 0.9 at 300◦C to 0.55 at
400◦C. The stoichiometry of NO2 reduction is complicated
and suggests that multiple reaction pathways are present.
Approximately 1.4 mol of ammonia is converted for every
1 mol of NO2 converted. Part of this higher conversion
of ammonia is most likely due to the stoichiometry of
NO2 reduction, while some appears to be due to ammonia
oxidation, possibly initiated by NO2. The large increase in
rate coupled with the change in kinetics strongly supports
our earlier suggestion that the rate-determining step in the
reduction of NO by ammonia is the oxidation of NO. We
believe that the rate-determining step in NO2 reduction
involves the combination of adsorbed NH3 and NO2
species, possibly on a framework Brønsted alumina site.
ACKNOWLEDGMENTS
We acknowledge Sanjay Sharma, Carlton Brooks, Dave Shihabi, Al
Lang, and Bill Bundens for their helpful assistance.
REFERENCES
1. Bosch, H., and Janssen, F., Catal. Today. 2, 1 (1987).
2. Stevenson, S. A., Vartuli, J. C., and Brooks, C. F., J. Catal. 190, 228
(2000), doi:10.1006/jcat.1999.2747.
3. Hirsch, P. M., Environ. Prog. 1, 24 (1982).
4. Andersson, L. A. H., Brandin, J. G. M., and Odenbrand, C. U. I., Catal.
Today 4, 173 (1989).
5. Brandin, J. G. M., Andersson, L. A. H., and Odenbrand, C. U. I., Catal.
Today 4, 187 (1989).
6. Satterfield, C. N., “Mass Transfer in Heterogeneous Catalysis,”
Krieger, Malabar, FL, 1981.
7. Mears, D. E., Ind. Eng. Chem. Process Des. Dev. 10, 541 (1971).
8. Levenspiel, O., “Chemical Reaction Engineering.” Wiley, New York,
1972.
9. Sharma, S. B., Meyers, B. L., Chen, D. T., Miller, J., and Dumesic,
J. A., Appl. Catal. A 102, 253 (1993).
10. Eng, J., and Bartholomew, C. H., J. Catal. 171, 27 (1997).
NO + NO2 + 2NH3 → 2N2 + 6H2O.
[10]