Catalytic activity and selectivity of H-ZSM5 for the reduction of nitric oxide by propane in the presence of oxygen

Article abstract of DOI:10.1006/jcat.1996.0193

The reactions of NO, C₃H₈, and O₂ were studied over a H-ZSM5 catalyst at various conditions. The reaction can result in the formation of either NO₂ or N₂. The selectivity for N₂ mainly depends on the ratio of NO/C₃H₈ in the feedstock; the reaction temperature and the concentration of O₂ have only minor influence. When the concentration of NO is 1000 ppm (which is typical for automotive exhausts) and NO/C₃H₈ > 1, the oxidation of NO to NO₂ dominates, but the NO₂ does not react significantly with C₃H₈ at these reaction conditions. NO can be reduced by C₃H₈ selectively to N₂ at NO/C₃H₈ ≤ 1 when oxygen is present. Results suggest that the selective catalytic reduction (SCR) of NO can proceed via two different reaction pathways below and above 500°C. At low temperatures, the oxidation of NO by O₂ to NO₂ might be the initial reaction step. Below 500°C, this reaction is much faster than other possible reactions in the mono- or bimolecular mixtures of the three reactants. At 500°C, the reaction rate of the oxidation of C₃H₈ by O₂ is comparable to that of the NO oxidation. At 600°C, the oxidation of C₃H₈ is faster than other reactions in the mono- or bimolecular mixtures of reactants. Thus, probably the combustion of propane initiates the SCR process above 500°C. The rates of both initial reactions depend on the concentration of oxygen and the reaction temperatures, but contrary to common belief these parameters have little effect on the yield of N₂ in the SCR process. It appears that the active sites for the reactions of NO + O₂, C₃H₈ + O₂, and C₃H₈ + NO₂ are similar to each other, probably involving the Bronsted acidic bridging hydroxyls of H-ZSM5. When NO/C₃H₈ > 1, adsorbed NO_x species may prevent the adsorption of C₃H₈ since propane barely reacts with the other reactants. At NO/C₃H₈ ≤ 1, the SCR of NO to N₂ probably proceeds via the secondary reactions of NO₂ + C₃H₈ (below 500°C) or C_xH_yO_z + NO_x (above 500°C). Radicals are likely involved in both the initial and the secondary reaction steps. At 300°C, the probable overall stoichiometry of the SCR process is 2C₃H₈ + 6NO + 4O₂ ? 6CO + 8H₂O + 3N₂. At higher temperatures, CO₂ also appears in the products and the efficiency of propane to form N₂ decreases.

Full text of DOI:10.1006/jcat.1996.0193

JOURNAL OF CATALYSIS 161, 359–372 (1996)
ARTICLE NO. 0193
Catalytic Activity and Selectivity of H-ZSM5 for the Reduction of Nitric
Oxide by Propane in the Presence of Oxygen
ꢀ
ꢀ
Istvan Halasz, Alan Brenner, K. Y. Simon Ng,† and Ying Hou†
ꢀ
Departments of Chemistry and †Chemical Engineering, Wayne State University, Detroit, Michigan 48202
Received October 23, 1995; revised January 22, 1996; accepted January 29, 1996
elusive even over the most intensively investigated ZSM5
catalysts. Despite several attempts (1–9) to outline some
feasible general features of the mechanism, only few mech-
anistic issues have been settled.
The reactions of NO, C3H8, and O2 were studied over a H-ZSM5
catalyst at various conditions. The reaction can result in the for-
mation of either NO2 or N2. The selectivity for N2 mainly depends
on the ratio of NO/C3H8 in the feedstock; the reaction temperature
and the concentration of O2 have only minor influence. When the
concentration of NO is 1000 ppm (which is typical for automotive
exhausts) and NO/C3H8 > 1, the oxidation of NO to NO2 dominates,
but the NO2 does not react significantly with C3H8 at these reac-
tion conditions. NO can be reduced by C3H8 selectively to N2 at
NO/C3H8 ꢀ 1 when oxygen is present. Results suggest that the selec-
tive catalytic reduction (SCR) of NO can proceed via two different
reaction pathways below and above 500ꢁC. At low temperatures,
the oxidation of NO by O2 to NO2 might be the initial reaction step.
Below 500ꢁC, this reaction is much faster than other possible reac-
tions in the mono- or bimolecular mixtures of the three reactants.
At 500ꢁC, the reaction rate of the oxidation of C3H8 by O2 is compa-
rable to that of the NO oxidation. At 600ꢁC, the oxidation of C3H8
is faster than other reactions in the mono- or bimolecular mixtures
of reactants. Thus, probably the combustion of propane initiates
the SCR process above 500ꢁC. The rates of both initial reactions
depend on the concentration of oxygen and the reaction tempera-
tures, but contrary to common belief these parameters have little
effect on the yield of N2 in the SCR process. It appears that the ac-
tive sites for the reactions of NO + O2, C3H8 + O2, and C3H8 + NO2
are similar to each other, probably involving the Brønsted acidic
bridging hydroxyls of H-ZSM5. When NO/C3H8 > 1, adsorbed NOx
species may prevent the adsorption of C3H8 since propane barely
reacts with the other reactants. At NO/C3H8 ꢀ 1, the SCR of NO
to N2 probably proceeds via the secondary reactions of NO2 + C3H8
(below 500ꢁC) or CxHyOz + NOx (above 500ꢁC). Radicals are likely
involved in both the initial and the secondary reaction steps. At
300ꢁC, the probable overall stoichiometry of the SCR process is
2C3H8 + 6NO + 4O2 ⇒ 6CO + 8H2O + 3N2. Athighertemperatures,
CO2 also appears in the products and the efficiency of propane to
c
Little is known about the active sites of SCR catalysts.
Since most active ZSM5 zeolites contain exchangeable
multivalent cations, such as Cu2+, Co3+, Ce3+, Ga3+, or
Pt2+, it is expected that these cations are active sites
for the SCR reaction. The adsorption of NOx and Ox
species on Cu-ZSM5 catalysts has been well documented
(1). However, access to the multivalent cations is mostly
blocked by large extra lattice anions (e.g., OHꢁ, O2ꢁ, or
O2ꢁ) in these high Si/Al zeolites (10–12). Therefore, the
direct interaction of the reactant molecules (O2, CxHy, and
NO) and the cations is unclear. The role of the extra lattice
anionic ligands in the SCR process has barely been studied.
Likewise, little attention has been paid to other potential
active sites, such as the oxygen, aluminum, and silicon ions
of the zeolite lattice (1, 13–16). Moreover, the multivalent
metal cations usually occupy only 20 to 50% (occasionally
up to 75%) of the exchangeable positions in the ZSM5
catalysts. The rest is most frequently occupied by sodium
ions and protons. Since the catalytic activity of Na-ZSM5
for the SCR of NO to N2 is low (3, 7, 19), Na+ ions are
probably not active sites for this reaction. In contrast, the
possible catalytic role of protons has received increasing
experimental support recently (1, 2, 17–19). Although
some active ZSM5 samples may not contain protons
(20–22), the Brønsted acidic bridging hydroxyl (BA-OH)
groups are likely involved in the SCR process over the
proton exchanged H-ZSM5 catalysts (1, 8, 17–19, 23).
Recent studies indicate that the BA-OH groups of
H-ZSM5 are probably active sites for the oxidation of NO
by O2 to NO2 above 200ꢂC (23). Many researchers agree
that NO2 might be an important initial intermediate in the
SCR process (1–4, 7–9, 23–26). Others assume that this is
only valid for the acidic H-ZSM5 and Al2O3 catalysts and
the oxidation of hydrocarbons is the initial step over other
catalysts (5, 6, 21, 27). Many subsequent intermediates have
been proposed to participate in the SCR process, in which
the ultimate products are N2, COx, and H2O. Examples are
carbonaceousdeposits(1, 28–30), ammonia(31), N2O3 (32),
form N2 decreases.
ꢂ 1996 Academic Press, Inc.
INTRODUCTION
The selective catalytic reduction (SCR) of NO to N2 by
hydrocarbons at net oxidizing conditions is a complex pro-
cess which can be used for air pollution control. Although
much knowledge has been accumulated about this reaction
during the past several years (1), the reaction route is still
359
0021-9517/96 $18.00
c
Copyright ꢃ 1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

360
HALASZ ET AL.
and alkyl (1, 2, 4), alkoxy (6, 7, 14, 33), nitoalkyl (1, 3, 7, and co-workers (21, 33, 41, 44), who first reported the
34), cyanide (35–37), and isocyanide (3, 36, 37) radicals. It SCR activity of H-ZSM5 (40), demonstrated that there is
appears that the rate determining step of the overall SCR no reaction between the two reactants at 400 and 500ꢂC
process is either the oxidation of NO to NO2 (1, 3) or the unless oxygen is also added to the feed. Since similar
dehydrogenation of hydrocarbons (2, 23, 38, 39).
observations were also made over other catalysts, the
There is a good chance that the different mechanisms do universal importance of oxygen to initiate the SCR process
not contradict each other. A large body of papers demon- seems to be well established (1).
strate that the ultimate reaction course may depend on
In the presence of 1% O2, about 50% of the NO and
several factors including the composition of catalysts, the 50% of the C3H8 was converted to N2, CO2, and H2O over
reaction temperature, the space velocity (SV), the relative the H-ZSM5 catalyst at 400ꢂC (21, 33, 41, 44). By reacting
concentrations of reactants, and the structure of the re- the same mixture at 500ꢂC, roughly 40% NO conversion
ductanthydrocarbons. Consequently, differentmechanisms and 90% C3H8 conversion was observed. These results in-
may be valid at different reaction conditions and a general dicate that higher temperature accelerates the combustion
mechanism for the SCR of NO by hydrocarbons may not of hydrocarbons and slightly slows the reduction of NO to
exist. In order to systematize the diversified and seemingly N2. An increase in the concentration of oxygen from 1 to
contradictory results, it is desirable to perform further stud- 10% did not significantly affect the production of N2 and
ies over carefully characterized catalysts at systematically COx (33, 44). By contrast, a similar change in the concen-
varied reaction conditions.
tration of oxygen considerably increased the yields of N2
Present experiments explore some details of the SCR of and COx over an Al2O3 catalyst (21, 41). This is surprising,
NO to N2 by propane over a H-ZSM5 catalyst at various because it is conjectured that the primary role of oxygen is
reactant concentrations and temperatures. Among the ac- to convert NO into more reactive NO2 over both catalysts
tive ZSM5 zeolites, H-ZSM5 is the simplest catalyst. Since (1, 33, 41, 44, 45, 48). Moreover, the reaction rate of the
protons are its only exchangeable cations, complications oxidation of NO by O2 to NO2 is much faster over H-ZSM5
associated with multivalent exchangeable cations can be than over Al2O3 (33, 41) and the production of NO2 can
avoided but the possible catalytic role of BA-OH sites can be considerably enhanced over H-ZSM5 by increasing the
be studied. A better understanding of the reaction path- concentration of oxygen (23). It is not known what happens
way over H-ZSM5 should facilitate the understanding of to the elevated NO2 production at high oxygen concentra-
the SCR process over more complex catalysts.
tions over H-ZSM5 when C3H8 is present. It is possible
Numerous papers have reported SCR studies over H- that the oxidation of NO by O2 is suppressed in the pres-
ZSM5 (1, 2, 8, 17–19, 21, 22, 33, 34, 40–48). Most of these ence of propane over this catalyst, or perhaps unreacted
studies concentrate on the comparison of the activities and NO2 remains despite the presence of the hydrocarbon. It
selectivities of H-ZSM5 samples with those of other cata- is also reasonable to assume that less than 1% O2 can ini-
lysts by varying the reductants, reaction temperatures, and tiate the SCR reaction, but how much less is not known.
space velocities. Hence, experiments are usually made at It is further unclear, what happens when the absolute and
constant concentrations of the three reactants, NO, CxHy, relative concentrations of NO and C3H8 are changed. For
and O2, typically near 0.1, 0.1, and 10%, respectively. While example, should oxygen be in excess compared to the ni-
such comparative experiments with constant reactant mix- tric oxide, to the hydrocarbon, or to both of them? These
tures are important, their significance is limited because the matters are examined in this study.
concentrations of reactants strongly impact the yields of N2
The effect of the hydrocarbon concentration on the N2
and COx over many catalysts (8, 17, 21, 39, 41–48). Not production over H-ZSM5 has been studied by using a feed
much is known about the reasons for these changes. It ap- containing about 1000 ppm NO, 10% O2, and different
pears that their magnitude varies with the composition of amounts of C3H8 (21, 33, 41–44). It was found that the NO
catalysts, space velocities, and reaction temperatures, but conversion to N2 roughly doubles (e.g., from 25 to 50%
there are insufficient data to elucidate why. To get more at 400ꢂC and SV = 18,000 hꢁ1) when the concentration of
information on this subject over the H-ZSM5 catalyst, a C3H8 increased from about 300 to 1300 ppm. This positive
wide range of reactant concentrations is used in the present effect of the hydrocarbon concentration is similar to those
work.
reported for Cu-ZSM5 (8) and Co-ZSM5 (48). In contrast,
SCR experiments at various oxygen and hydrocarbon the N2 production changed little when the concentrations of
concentrations have already been reported over H-ZSM5 hydrocarbons over a Ga-ZSM5 catalyst were changed (17,
(21, 33, 41–44). However, several aspects have not been 43, 46). The reasons for these differences are not clearly
investigated. For example, the effect of oxygen has only understood. There are several unanswered questions re-
been examined in a feed containing 1000 ppm NO and garding the concentration effect of hydrocarbons over the
323 ppm C3H8 (balance He). By passing this mixture H-ZSM5 catalyst as well. For example: (i) Is this effect still
through a H-ZSM5 catalyst (SV = 1800 hꢁ1), Hamada valid when the concentration of oxygen is lower than 10%?

CATALYTIC ACTIVITY AND SELECTIVITY OF H-ZSM5
361
(ii) What happens when NO is in excess compared to C3H8? 25% NH4OH to maintain pH > 7) in a glass flask at 80ꢂC
(iii) Is 300 ppm the minimum propane concentration that overnight. The sample was thoroughly washed five times
can initiate SCR reaction with 1000 ppm NO? According with 100 cm3 distilled water and dried in air. Then, 0.5 g
to the reaction C3H8 + 10NO ⇒ 5N2 + 3CO2 + 4H2O, even of the dried NH4-ZSM5 was placed in a quartz reactor and
300 ppm C3H8 represents a threefold stoichiometric excess heated in flowing N2 (ꢄ10 ml/min) to 600ꢂC at a rate of
compared to 1000 ppm NO. If, as very recently proposed (2, 10ꢂC/min. The effluent bubbled through two traps (each
48), the combustion of hydrocarbons by NO2 and/or NO is filled with 100 ml distilled water) to absorb evolving NH3.
a major driving force in the SCR process, it is expected that The solutions were continuously stirred by magnetic stirrers
excess NO2 or NO is desirable to accelerate the combustion. and neutralized with 0.1 N aqueous HCl by using methyl or-
Some of the above questions regarding the mechanism ange indicator. Acid consumption was detected only in the
of the SCR of NO by C3H8 over H-ZSM5 are addressed in first absorber, indicating full NH3 absorption. Repeated ex-
this paper. Several problems regarding to the stoichiometry periments with different zeolites proved that the described
of this reaction are addressed as well. To check what hap- method assures roughly 100% NH+4 -exchange and the er-
pens when one of the three reactants, NO, C3H8, and O2, is ror in determining the lattice aluminum (BA-OH) content
limiting or excess compared to the others, it is important to is less than ꢅ 5%.
know what is their stoichiometric ratio in the overall SCR
The reactants for the catalytic measurements were cer-
process. Since this information is not available, the deter- tified mastergases from Scott (0.1% C3H8 in He, and 0.1,
mination of the probable stoichiometry for the SCR of NO 1, and 10% O2 in He) and instrument grade gases from
by C3H8 was investigated.
Linde (NO), Liquid Carbonic (He, 1% NO in He, 1%
The compositions of the actual exhaust gases of spark ig- C3H8 in He), Union Carbide (C3H8), and Smith Welding
nition engines and the reaction conditions in the catalytic Co. (O2). Except for helium, which was passed through a
converters of automobiles are complex and can vary widely. General Electric Go-Getter purifier, all gases were used as
For example, various concentrations of CO, H2, N2, O2, received.
H2O, NO, soot, and several hydrocarbons can be simulta-
Activity Measurements and Analysis
neously present in a typical exhaust and the space velocity
A 1 cm3 (0.46 g) catalyst sample was placed in a fused
over a TWC catalyst randomly changes between 20,000 and
100,000 hꢁ1 (62). These and many other factors have to be
quartz reactor (10 mm i.d.) of the flow-through type. The
taken into account when practical applications are consid-
ered. It has also been shown that the SCR activity of ZSM5
catalysts is significantly suppressed in the presence of 10%
H2O which easily occurs in real exhausts (90). Hence, the
model reactions presented here of the three reactants, NO,
C3H8, O2, at roughly 10,000 hꢁ1 space velocity have been
selected to elucidate certain features of the SCR reaction
and the results cannot be directly used to define the best
operating conditions for an automotive SCR process.
catalyst was supported on a porous quartz frit and covered
with quartz wool. Prior to the first catalytic experiment, the
catalyst was evacuated at 250ꢂC for 2 h and pretreated in a
flow of He (100 ml/min) for 2 h at 600ꢂC. No further pre-
treatment was done during the catalytic tests. The catalyst
was kept in helium between runs.
For catalytic measurements, appropriate amounts of oxy-
gen, nitric oxide, and propane were mixed with a bal-
ance of helium to achieve a total flow rate of 160 ml/min
(SV = 9600 hꢁ1) at 1 atm (1.0 ꢆ 105 Pa). To cover the en-
tire range of possible stoichiometries (Table 1) and also
the reactant ratios used most frequently in the literature,
NO/C3H8 ratios of 0.5, 0.7, 1, 2, 4, and 20 and O2/NO ratios
EXPERIMENTAL
Catalysts and Materials
H-ZSM5 zeolite was prepared by calcining an NH+4 - of 0.13, 0.25, 0.5, 2.5, 5, 17.5, and 87.5 were used for the
exchanged ZSM5 sample from the PQ Corporation (Lot present studies. All permutations of these two ratios were
ZN-9; 80% crystallinity; BET surface area ꢄ460 m2/g; total studied, so the corresponding O2/C3H8 ratios varied from
pore volume 0.28 ml/g; 0.002 mmol/g sodium; Si/Al ratio, 0.065 to 1750. Typical NO/C3H8, O2/NO, and O2/C3H8 ra-
18) at 550ꢂC for 24 h in air. According to the weight loss up tios (they can vary widely) in an automotive exhaust are
to 800ꢂC (measured by TGA in nitrogen), the catalyst con- 1.4, 4.9, and 6.8, respectively (61). The concentrations of
tained 2.0 mmol/g (0.036 ml/g) adsorbed water. Li-ZSM5 O2 were varied from 125 ppm to 34%, that of C3H8 from
was prepared by solid-state reaction of H-ZSM5 and LiCl 50 ppm to 2%, and the concentration of NO was either 0.1
as described elsewhere (23). The powdered zeolites were or 2%. Experiments were carried out at reaction tempꢁe1r-
pressed into pellets, crushed, and the 0.3- to 0.6-mm frac- atures from 25 to 600ꢂC and at constant (SV = 9600 h
)
tion was used for catalytic measurements.
space velocity. Flow rates were controlled by Brooks mass
To measure the density of BA-OH groups in the H-ZSM5 flow controllers. The reaction system is highly leak tight and
sample, NH+4 -exchange was carried out three times by stir- is constructed of stainless steel with bellows seal valves. De-
ring 4 g zeolite with 100 ml aqueous 1 M NH4Cl (plus 5 ml tails of the system are described elsewhere (23, 34, 51–55).

362
HALASZ ET AL.
TABLE 1
Feasible Formal Stoichiometries for the SCR of NO by Propane When (i) Both NO and O2 Contribute a Minimum
of One Oxygen Atom to the Oxidation of C3H8, (ii) a Single Oxygen Atom Reacts with Only One Hydrocarbon
Molecule, and (iii) the Formation of CO and N2O Is Not Significant
Stoichiometric ratios
Expected
Possible reactions
O2/NO
NO/C3H8
O2/C3H8
N2/C3H8
1. 2C3H8 + 2NO + 9O2 ⇒ 6CO2 + 8H2O + 1N2
2. 2C3H8 + 4NO + 8O2 ⇒ 6CO2 + 8H2O + 2N2
3. 2C3H8 + 6NO + 7O2 ⇒ 6CO2 + 8H2O + 3N2
4. 2C3H8 + 8NO + 6O2 ⇒ 6CO2 + 8H2O + 4N2
5. 2C3H8 + 10NO + 5O2 ⇒ 6CO2 + 8H2O + 5N2
6. 2C3H8 + 12NO + 4O2 ⇒ 6CO2 + 8H2O + 6N2
7. 2C3H8 + 14NO + 3O2 ⇒ 6CO2 + 8H2O + 7N2
8. 2C3H8 + 16NO + 2O2 ⇒ 6CO2 + 8H2O + 8N2
9. 2C3H8 + 18NO + 1O2 ⇒ 6CO2 + 8H2O + 9N2
9/2 = 4.50
8/4 = 2.00
7/6 = 1.17
6/8 = 0.75
5/10 = 0.50
4/12 = 0.33
3/14 = 0.21
2/16 = 0.13
1/18 = 0.06
2/2 = 1
4/2 = 2
9/2 = 4.5
8/2 = 4.0
7/2 = 3.5
6/2 = 3.0
5/2 = 2.5
4/2 = 2.0
3/2 = 1.5
2/2 = 1.0
1/2 = 0.5
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
6/2 = 3
8/2 = 4
10/2 = 5
12/2 = 6
14/2 = 7
16/2 = 8
18/2 = 9
Note. The ratio N2/C3H8 is [mol produced N2/mol reacted C3H8] provided the combustion of C3H8 by O2 is negligible.
Separate measurements of the nitric oxide and total NOx actant mixture was low but measurable NH3 was not de-
contents in the reactor effluent were made using a Beckman tected at our reaction conditions.
Model 951 chemiluminescent NO-NOx gas analyzer and
were compared with the nitric oxide contents at the reactor
inlet to calculate the conversion of nitric oxide. The ana-
lyzer was calibrated with different concentrations of known
gases. Due mainly to the error of reading an analog meter, it
is estimated that the accuracy of the conversion is ꢅ 5% (re-
peatability plus linearity error). Repeated measurements
on the same catalyst (after miscellaneous intervening reac-
tions) gave a repeatability of ꢅ 10% (includes hysteresis of
catalyst). The ammonia content was analyzed by its reoxi-
dation to nitric oxide as described elsewhere (51). All other
products were analyzed by gas chromatography using a
GOW MAC 550P type chromatograph equipped with TCD
detectors using He carrier gas. A stainless steel column (2.5
RESULTS
Stoichiometry and Selected Reactant Ratios
There is substantial experimental experience that con-
siderable reduction of NO to N2 occurs only when all three
reactants, NO, O2, andCnH2n + x, aresimultaneouslypresent
in the feed (1). Hence, neglecting that some N2O and CO
may appear at certain reaction conditions, the general SCR
reaction can be described as
CnH2n+x + aNO + bO2 ⇒ (a/2)N2 + nCO2 + cH2O,
where b = [(6n + x ꢁ 2a)/4], c = [n + (x/2)], and a ꢇ 1;
b ꢇ 0.5. One can consider this reaction as the combustion
of hydrocarbons (2, 48). For this special combustion, a min-
imum of one oxygen atom must come from the nitric oxide
even when all other oxygens come from the molecular O2
and vice versa. Consequently, depending on the ratio of NO
to O2 molecules participating in the combustion, several
stoichiometric equations can exist for a single hydrocarbon
molecule.
Table 1 shows some feasible equations for the SCR of
NO by propane. Provided that a single oxygen atom (ei-
ther from NO or from O2) does not react with more than
one hydrocarbon molecule, one of the nine reactions in
Table 1 must be close to the overall stoichiometry of the
SCR process regardless of the catalyst used. Since there
are no chemical reasons to prefer any of these equations, all
corresponding reactant ratios must be considered as possi-
ble stoichiometry when experiments with different concen-
trations of reactants are designed. Note that the limits of
these stoichiometric ratios (Table 1) remain valid when the
possible production of some CO or N2O is also taken into
account.
1
feet ꢆ inch) packed with 5A molecular sieve at 60ꢂC sep-
arated8NO, N2, O2, and CO. Hydrocarbons (C1–C4), CO2,
N2O, and H2O were separated at 105ꢂC on a stainless steel
1
column (6.5 feet ꢆ 8 inch) packed with Poropack Q.
The NO and NOx concentrations were monitored con-
tinuously by passing the total effluent gas stream through
the NO–NOx analyzer. Steady-state NO conversion was at-
tained (as detected by the NO/NOx analyzer) before GC
analysisandanalysisofNH3 weredone. Thereactoreffluent
was then connected to the GC system and a gas sampling
valve (loop volume ꢄ0.52 cm3) was used to inject samples
into the gas chromatograph. The total product composition
was obtained by two injections on the two different GC
columns. To avoid the condensation of water, the sample
loop and the gas lines were heated to about 100ꢂC. Despite
this precaution, we were unable to separate and measure
quantitatively the water and propane content in products
when the hydrocarbon conversion exceeded roughly 30%.
Therefore, the total COx production was used to calculate
the conversion of propane. Possible NH3 formation was oc-
casionally checked whenever the oxygen content in the re-

CATALYTIC ACTIVITY AND SELECTIVITY OF H-ZSM5
363
By determining the ratio of [produced N2/reacted C3H8]
(Table 1) when the direct combustion of C3H8 by O2 is neg-
ligible, the stoichiometry of the SCR process can be elu-
cidated. This ratio also describes the efficiency of propane
to produce N2, which is an important practical factor (1).
According to the calculated values in Table 1, the minimum
efficiency of propane must be near 0.5 when significant side
reactions do not occur. Higher efficiencies can be expected
when more than one nitric oxide molecule can react with
one hydrocarbon. Hence, Table 1 suggests that maintaining
high NO/C3H8 and low O2/C3H8 and O2/NO ratios could
be favorable to improve the efficiency of propane. In con-
trast, considerable N2 yields over H-ZSM5 have only been
reported by selecting O2/C3H8 ꢈ 10 and O2/NO ꢈ 10 ra-
tios (2, 8, 17–19, 21, 22, 33, 34, 40–48). The corresponding
NO/C3H8 ratios were mostly close to 3 or 0.8 (2, 21, 33, 40–
44) and in a few experiments near 2, 1.5, and 1 (22, 41, 43).
FIG. 1. Conversions of NO at various reaction temperatures and O2/
NO molar ratios with feeds containing 1000 ppm NO and variable O2; total
P = 1 ꢆ 105 Pa; SV = 9600 hꢁ1. Cross-hatched line is the stoichiometric
ratio for 2NO + O2 ⇒ 2NO2.
BA-OH Coverage
Based on the HCl consumption, the NH+4 -exchange ca-
pacity of H-ZSM5 was 0.424 mmol/g. Since the Si/Al ra-
tio of 18 (calculated from the total Si and Al content of
the zeolite sample) would correspond to 0.85 mmol/g lat-
tice aluminum atoms, the NH+4 -exchange capacity of H-
ZSM5 indicates that only one-half of the total aluminum
content is associated with exchangeable protons (BA-OH
groups).
sion of NO vanishes at high temperatures when the protons
are replaced by Li+ ions in this zeolite.
Figure 3 shows the oxidation of C3H8 by O2. When only 50
ppm propane is in the feed (Fig. 3a), the combustion starts
above 400ꢂC. By increasing the concentration of propane to
1000 ppm (Fig. 3b), the “light-off” temperature decreases to
300ꢂC and the conversions (at identical O2/C3H8 ratios) go
up. The light-off temperature remained above 300ꢂC when
the concentration of propane was increased to 2%.
Figure 4 shows an example for the effect of the concentra-
tion of C3H8 on the reaction rates of propane combustion
in the presence of 10% O2. (BA-OH groups are assumed
active sites, vide infra.) This oxygen concentration would
Reactions with One or Two Reactants
To get information about the possible initial steps of the
SCR process, reactions with feeds containing only one or
two of the three reactants, NO, O2, and C3H8, were studied
first. The reaction parameters of these experiments were
the same as those used for the SCR reactions.
In agreement with prior findings (34, 41), the conversions
of NO or C3H8 remained below 3% in all possible reactions
of the mono- or bimolecular mixtures except for the oxida-
tions of NO and C3H8 by O2. Only the oxidation of NO
to NO2 proceeds with measurable conversion below 400ꢂC.
Figure 1 shows the effects of oxygen concentrations and re-
action temperatures on the conversion of NO in reaction
mixtures containing 0.1% NO. The shape of the activity
curves for feeds containing 2% NO are similar to those in
Fig. 1, but, due to a second-order dependence of the rate
of reaction on the nitric oxide partial pressure (56), the
conversions are somewhat higher than those in Fig. 1. De-
tails of these processes have been discussed elsewhere (23).
Figure 2 illustrates two earlier observations that are rele-
vant here (the homogeneously formed NO2 (23) has been
subtracted): (i) the conversion of NO declines after reach-
ing a maximum at around 350ꢂC because the equilibrium
NO + 1/2O2 = NO2 is attained; (ii) the BA-OH groups of
H-ZSM5 are probably active sites for this reaction above
200ꢂC(butnotnearroomtemperature)becausetheconver-
FIG. 2. Comparison of the catalytic activities of H-ZSM5 and Li-
ZSM5 to the equilibrium conversion for the oxidation of 2% NO by 5%
O2 at various temperatures; total P = 1 ꢆ 105 Pa; SV = 9600 hꢁ1
.

364
HALASZ ET AL.
FIG. 3. Conversions of C3H8 to COx at various reaction temperatures and O2/C3H8 molar ratios with feeds containing 50 ppm (a) and 1000 ppm
(b) C3H8 and variable O2; total P = 1 ꢆ 105 Pa; SV = 9600 hꢁ1. Cross-hatched line is the stoichiometric ratio for C3H8 + 5O2 ⇒ 3CO2 + 4H2O.
correspond to O2/C3H8 ratios of 2000 and 100 in Figs. 3a described above. Figure 6 illustrates that the conversion is
and 3b, respectively.
also much lower when Li-ZSM5 catalyst is used instead of
Similar to other reports (48), significant CO production H-ZSM5.
was observed at all reaction conditions studied. Figure 5
Figure 7 compares the reaction rates of the oxidation of
shows an example for the CO and CO2 distributions at the C3H8 and NO by O2 at various temperatures. The combus-
reaction conditions of Fig. 3b. For feeds containing from tion of propane is about 10 times slower at 400ꢂC and about
0.05 to 2% C3H8, the CO/CO2 ratios were similar to those 10 times faster at 600ꢂC than the oxidation of NO. Data in
in Fig. 5 at identical O2/C3H8 ratios and reaction tempera- Figs. 1, 3, and 7 indicate that excess oxygen compared to the
tures. The combustion of propane over H-ZSM5 is a hetero- stoichiometric ratios of reactants (checked lines in Figs. 1
geneously catalyzed reaction because less than 3% propane and 3) can significantly enhance the reaction rates in both
conversion was measured when blank experiments were reactions.
carried out in an empty reactor at the reaction condition
FIG. 4. Effect of the concentration of C3H8 on the turnover frequency
in the oxidation of propane by oxygen with feeds containing 10% O2 at
500 and 600ꢂC; total P = 1 ꢆ 105 Pa; SV = 9600 hꢁ1. The number of active
FIG. 5. CO ( ) and CO2 ( ) production in the oxidation of C3H8 by
O2. The reaction conditions are the same as in Fig. 3b.
sites was determined from the BA-OH density (0.424 mmol/g).

CATALYTIC ACTIVITY AND SELECTIVITY OF H-ZSM5
365
FIG. 6. Comparison of the catalytic activities of H-ZSM5 and Li-
ZSM5 in the oxidation of 0.1% C3H8 by 9% O2 at various temperatures;
total P = 1 ꢆ 105 Pa; SV = 9600 hꢁ1
.
FIG. 8. Effect of the concentration of oxygen on the total conversions
of NO in the presence of C3H8 at various reaction temperatures. Feed
composition: 1000 ppm NO, 1000 ppm C3H8, and variable O2. Total P =
Reactions with Three Reactants
1 ꢆ 105 Pa; SV = 9600 hꢁ1
.
Figure 8 shows the total conversions of NO for a reactant
mixture of 0.1% NO, 0.1% C3H8, and 0.01–9% O2 at various
temperatures. Except that the feed also contained propane,
the reaction conditions in Fig. 8 are the same as those in
Fig. 1. There are only minor differences between the two
figures: the high temperature maxima are ꢄ50ꢂC higher in
Fig. 8 and the conversions in Fig. 8 are somewhat lower than
those in Fig. 1 at temperatures below 400ꢂC but somewhat
higher at temperatures above 400ꢂC.
Figure 9 shows the conversions of NO to NO2 and N2 at
the same reaction conditions as Fig. 8. As expected, mainly
N2 is formed in the presence of propane and considerable
NO2 (the only product in Fig. 1) only appears when the
reaction temperature is low. Experiments in Fig. 9 were
carried out at a constant ratio of NO/C3H8 = 1. According
to Table 1, this is the lowest stoichiometric value for the
SCR process. Therefore, SCR reaction was also carried out
at different NO/C3H8 ratios. For these measurements, the
highest oxygen concentration (8.75%) was selected from
Fig. 9 because this corresponds to the maximum N2 pro-
duction.
The rather surprising results of these experiments are
summarized in Fig. 10. Despite many suggestions to the con-
trary, this graph indicates that the formation of NO2 alone
is not enough to initiate the SCR of NO to N2 regardless of
the temperature. Although ample NO2 is present to react
with the hydrocarbon when NO/C3H8 > 1, noticeable N2
production occurs only when NO/C3H8 ꢉ 1. Above 100ꢂC,
FIG. 7. Effect of the concentration of oxygen on the turnover fre-
quencies in the oxidation of 0.1% NO (solid lines) and 0.1% C3H8 (dashed
lines) by O2 at 400, 500, and 600ꢂC; total P = 1 ꢆ 105 Pa; SV = 9600 hꢁ1
.
The number of active sites was determined from the BA-OH density
(0.424 mmol/g).
FIG. 9. ConversionsofNOtoNO2 ( )andN2 ( )atthesamereaction
conditions as those in Fig. 8.

366
HALASZ ET AL.
FIG. 12. Effect of the concentrations of propane and oxygen on the
conversions of NO (total , to NO2 and to N2 ) at 600ꢂC. Except for
the temperature, conditions are the same as in Fig. 11.
FIG. 10. Effect of the concentration of propane in the presence of
O2 on the conversions of NO to NO2 ( ) and N2 ( ) at various reaction
temperatures. Feed composition: 0. 1% NO, 9% O2, and variable C3H8.
Total P = 1 ꢆ 105 Pa; SV = 9600 hꢁ1
.
full SCR mixture. At NO/C3H8 > 1, the total conversion of
NO (in the presence of C3H8) is higher than in the NO + O2
reaction (in the absence of C3H8, line “Cat.”) but still NO2 is
the main product. At NO/C3H8 ꢉ 1, the total conversion of
NO drops and N2 forms selectively. Similar to the results of
Fig. 9 at different reaction temperatures, the concentration
of oxygen (O2/NO ratio) affects the total conversion of NO
but has no significant influence on the selectivity for N2.
Results at 400ꢂC were very similar to those shown in
Fig. 11. Results at 500ꢂC were rather similar to those at
600ꢂC, which are shown in Fig. 12. Although the NO/C3H8
ratio still determines whether NO2 or N2 is produced at
600ꢂC, the trends in Fig. 12 are different from those in
Fig. 11. In Fig. 12, the conversion of NO monotonically in-
creases with decreasing NO/C3H8 ratios (increasing C3H8
concentrations).
the reaction temperature seems not to affect significantly
the selectivity for N2.
Figures 11–13 show the effects of the ratios O2/NO and
NO/C3H8 on the total conversion of NO and on the selec-
tivity to NO2 or N2 at various temperatures and NO concen-
trations. For comparison, some data regarding the NO + O2
reaction are also included in these figures.
Figure 11 summarizes results for the reaction of 0.1% NO
with different amount of O2 and C3H8 at 300ꢂC. Similar to
Fig. 10, the ratio of NO/C3H8 = 1 clearly divides results into
two groups. Whenever the concentration of oxygen is high
enough to generate noticeable reaction between the NO
and the O2 (line “Cat.”), NO conversion also occurs in the
Figure 13 shows that the conversion of NO increases
when the reactant mixture contains 2% NO instead of 0.1%
FIG. 11. Effect of the concentrations of propane and oxygen on the
conversions of NO (total , to NO2 , and to N2 ) at 300ꢂC. Feed com-
position: 0.1% NO, variable C3H8 and O2. Total P = 1 ꢆ 105 Pa; SV = 9600
h
ꢁ1. The first three lines are the conversions of NO in the absence of C3H8:
Eq., equilibrium conversion of NO for the reaction NO + 1/2 O2 ⇒ NO2;
Cat., conversion of NO over H-ZSM5 catalyst; Ho., blank conversion of
NO in an empty reactor.
FIG. 13. Effect of the concentrations of propane and oxygen on the
conversions of NO (total , to NO2 , and to N2 ) at 300ꢂC. Except for
the concentration of NO being 2%, conditions are the same as in Fig. 11.

CATALYTIC ACTIVITY AND SELECTIVITY OF H-ZSM5
367
when NO/C3H8 > 1, no matter how much oxygen is present
at 500ꢂC. Similar trends were observed at all temperatures
from 300 to 600ꢂC. This result confirms that NO2, which is
the chief product at these conditions (Figs. 10–12), does not
react with C3H8 until the partial pressure of C3H8 attains or
exceeds the partial pressure of NO. Figure 15 also illustrates
that, akin to the bimolecular oxidation of propane by O2
(Fig. 5), a significant amount of CO is formed when the
ratio of NO/C3H8 is ꢉ1. At the highest concentrations of
oxygen (8.75% or 87.5 O2/C3H8 ratio) the average ratio of
CO2/CO was roughly 1/100, 1/2, 1/1, and 2/1 at 300, 400,
500, and 600ꢂC, respectively.
Figure 16 shows the total conversions of C3H8 to COx in
the SCR reaction at temperatures and O2/NO ratios sim-
ilar to those in Fig. 3. Comparison of Fig. 16a and Fig. 3a
indicates that NO inhibits the combustion of propane at
all reaction temperatures when NO/C3H8 ꢈ 1. In contrast,
comparison of Fig. 16b and Fig. 3b indicates that the com-
bustion of propane is significantly accelerated between 200
and 500ꢂC (but not at 600ꢂC) when NO/C3H8 = 1.
FIG. 14. Effect of the concentrations of propane and oxygen on the
conversions of NO (total , to NO2 , and to N2 ) in an empty reactor.
Conditions are same as in Fig. 13.
(Fig. 11). This increase is largely due to a significant ho-
mogeneous (or wall catalyzed) NO2 formation (line “Ho.”)
(23). Despite this possibility, the NO/C3H8 ratio still primar-
ily controls the selectivity. At these high concentrations of
reactants, NO2 is able to activate propane even at ratios
NO/C3H8 > 1. However, the production of N2 becomes se-
lective only when the partial pressure of C3H8 approximates
the partial pressure of NO. Figure 14 demonstrates that at
NO/C3H8 < 20 less N2 forms homogeneously in an empty
reactor at the same reaction conditions as in Fig. 13. Thus,
the reaction of NO2 and C3H8 is mainly heterogeneously
catalyzed even at these high partial pressures of reactants.
As demonstrated before (34), the process is largely homo-
geneous at these high reactant concentrations near 600ꢂC.
Figures 15 and 16 illustrate results for the combustion of
propane by O2 in the presence of NO. Figure 15 shows that
the NO/C3H8 ratio has a strong impact on the conversion of
propane to COx: there is no significant propane combustion
Hydrocarbon Efficiency and SCR Stoichiometry
The number of N2 molecules formed per consumed C3H8
(N2/C3H8 in Table 1) was calculated from the total COx
production. Since the measurements have an accuracy of
ꢅ 2.5% in determining the conversions of NO or C3H8, ef-
ficiency data were only calculated when the conversions of
both NO and C3H8 exceeded 5%. Figure 17 shows the ef-
ficiency ratio, N2/C3H8, at various reaction conditions. As
expected, the efficiency decreases with increasing temper-
atures and oxygen concentrations. The efficiency also de-
creases when the NO/C3H8 ratio goes below 1 because it
is unlikely that a single NO can react with more than one
C3H8 molecule. Figure 17 also indicates that the best effi-
ciency for the SCR of NO by C3H8 to N2 over the H-ZSM5
catalyst at the reaction conditions studied is about 1.5; i.e.,
one C3H8 forms 1.5 N2, which indicates that it has reduced
three NO molecules.
DISCUSSION
Experiments with mixtures of one and two of the three
reactants, NO, O2, and C3H8, indicate that the oxidation of
NO by O2 and the oxidation of C3H8 by O2 are the only sig-
nificant reactions at temperatures from 25 to 600ꢂC. Since
concerted trimolecular reactions have low probability, one
of these oxidations probably initiates the SCR reaction. At
300 and 400ꢂC, the combustion of propane is slow com-
pared to the oxidation of NO by O2 (Figs. 1, 3, and 7).
Therefore, the oxidation of NO to NO2 may initiate the
SCR reaction at these low temperatures. The reaction rates
of propane burning approximate the rates of the oxidation
of NO at 500ꢂC and substantially exceed them at 600ꢂC
(Fig. 7). Consequently, the C3H8 + O2 reaction may be the
initial reaction step for the SCR process near 600ꢂC.
FIG. 15. Effect of the concentrations of propane and oxygen on the
conversions of C3H8 (total , to CO , and to CO2 ) at 500ꢂC. Feed
composition: 0.1% NO, variable C3H8 and O2. Total P = 1 ꢆ 105 Pa; SV =
9600 hꢁ1
.

368
HALASZ ET AL.
FIG. 16. Total conversion of C3H8 to COx at various reaction temperatures and O2/C3H8 molar ratios with feeds containing 50 ppm (a) and 1000
ppm (b) C3H8, 1000 ppm NO, and variable O2; total P = 1 ꢆ 105 Pa; SV = 9600 hꢁ1
.
These observations coincide with the recently suggested sidered to be an important intermediate in the SCR of NO
possibility that the SCR reaction can proceed via two dif- over H-ZSM5 because it can react directly with C3H8 both
ferent reaction pathways below and above 500ꢂC even in the presence or in the absence of oxygen (1–4, 7–9, 23–
over a single H-ZSM5 catalyst (34). It was proposed that 26, 41–48). The remarkable similarity of Figs. 1 and 8 also
the low temperature reactions are possibly homogeneous– suggests that the oxidation of NO to NO2 may have an im-
heterogeneous processes involving NO2 as the key reac- portant role in the SCR process. Only recently has it been
tant to form radical surface intermediates. At high tem- noted that the formation of NO2 may be a necessary but
peratures, homogeneous radical reactions involving both not sufficient requirement for the SCR of NO to N2 (1, 24).
NO and NO2 were suggested to be rate determining. The Results in Figs. 10–14 demonstrate that the presence of NO2
present results suggest that different reaction pathways be- alone is not sufficient to initiate the SCR process. In con-
low and above 500ꢂC may also exist when the SCR feed- trast to common belief, the reaction temperature and the
stock contains as little as 0.1% NO and 0.1% C3H8, which concentration of oxygen have only little effect on the inter-
are close to their concentrations in automotive exhausts action of the NO2 with propane. Figures 11, 12, and 13 show
(62). It also appears that the initial reaction step is different that both NO2 and N2 can be selectively formed in the re-
in the low and high temperature SCR reactions.
action of NO, O2, and C3H8 over H-ZSM5 at very different
At low temperatures, the formation of NO2 is a likely reaction conditions and mainly the NO/C3H8 ratio controls
initial reaction. As mentioned before, NO2 is widely con- whether NO2 or N2 is formed.
The disappearance of the initially formed NO2 to selec-
tively yield N2 largely depends on the NO/C3H8 ratio in
the SCR feedstock. Figures 10–13 show that this change
occurs when the partial pressure of C3H8 approximates or
exceeds the partial pressure of NO, i.e., when NO/C3H8 ꢉ 1.
Figures 15 and 16 indicate that, simultaneous with the ap-
pearance of N2, a significant propane combustion begins
at NO/C3H8 ꢉ 1. Several papers have emphasized that the
NO/CxHy ratio can affect the conversion of NO to N2 over
different catalysts (8, 62). However, this effect has never
been associated with a switch from the selective formation
of NO2 to the selective production of N2 near NO/C3H8 = 1.
The relation between the NO/C3H8 ratio and the conver-
sion of propane has also not been previously noted.
Since NO or O2 alone do not react with C3H8 at 300
or 400ꢂC, it is unlikely that C3H8 totally suppresses the
formation of NO2 at NO/C3H8 ꢉ 1. In agreement with the
FIG. 17. Efficiency of C3H8 to produce N2 (N2/C3H8) over H-ZSM5
catalyst at various reaction conditions.
prior opinions (1–4, 7–9, 23–26), it is likely that NO2 or

CATALYTIC ACTIVITY AND SELECTIVITY OF H-ZSM5
369
its adsorbed precursors initiate the reaction of propane at found to be a homogeneous process at these high reactant
these low temperatures. Once the propane is activated, NO2 concentrations near 600ꢂC (34).
is totally used up to produce N2. Figures 1 and 7–9 support
NO2 can activate the C3H8 either by nitration (1, 3, 63) or
the idea that NO2 can be a reactant in the SCR process be- by oxidation (2, 24, 48). Over H-ZSM5, both reactions are
cause they suggest that the reaction temperature and the heterogeneous–homogeneous processes involving BA-OH
concentration of oxygen can control the amount of NO2 active sites to form HNO2, NO+2 , C3H+, OHꢊ, and other
produced in the initial step, and therefore they also affect radical intermediates (59, 63). The NH+4 9exchange capacity
the production of N2.
of our H-ZSM5 sample indicates that about one-half of the
The comparison of Figs. 3b and 16b also suggests that lattice aluminum atoms are associated with BA-OH sites.
NO2 can be an intermediate in the SCR process because This result suggests that roughly 25% of the total aluminum
the accelerated combustion of C3H8 in Fig. 16b is prob- content is in extra lattice position and blocks the exchange
ably promoted by NO2. It has been reported that small positions on another 25% of the aluminum ions. It is known
amounts of NO can promote the homogeneous combus- that similar neutralization can occur even when the zeolite
tion of hydrocarbons by converting unreactive HO2 rad- is only physically mixed and calcined with a binder oxide,
icals into chain propagating OH radicals by the reaction e.g., with boehmite at 350ꢂC (50).
NO + HO2 ⇒ NO2 + OH (67). However, this possibility
In agreement with prior observations (2, 23, 24, 48), the
seems unlikely for the increased propane conversion in low activity of Li-ZSM5 for the oxidations of NO and C3H8
Fig. 16b for several reasons: (i) NO does not react with C3H8 by O2 compared to that of H-ZSM5 (Figs. 2 and 6) suggests
at the studied reaction conditions; (ii) there is no promotion that the BA-OH sites of H-ZSM5 are involved in the active
when NO is in excess compared to C3H8 (Fig. 16a); (iii) the sites for both reactions because the only difference between
promotion of the homogeneous combustion of C3H8 by NO the two zeolites was that the exchangeable protons of H-
does not result in N2 formation (67), whereas Fig. 9 shows ZSM5 were substituted with Li+ ions in Li-ZSM5. Hence,
that significant N2 formation can occur at the reaction con- it is possible that the competition of NO and C3H8 for these
ditions of Fig. 16b; (iv) the promotion effect is negligible sites can explain the crucial role of the NO/C3H8 ratio in
at 600ꢂC (Fig. 16b), where only minimum NO2 can form the SCR process: until the partial pressure of C3H8 is low
(Figs. 1, 12) (23).
compared to that of NO (NO/C3H8 > 1), the activated ad-
Figure 11 indicates that the hydrocarbon can promote sorption of C3H8 (probably in the form of C3H+9 (34, 64–66))
the oxidation of NO, and the conversion of NO to NO2 can be inhibited by adsorbed NOx species that preferably
approaches equilibrium in the presence of propane (at adsorb on the BA-OH sites of zeolites (63, 65, 86). Hence,
NO/C3H8 > 1) but not in its absence (line “Cat.”) It is well C3H8 cannot react with NO2 at these conditions if the acti-
known that low levels of hydrocarbons can promote the vated adsorption of C3H8 is a prerequisite for this reaction.
homogeneous oxidation of NO to NO2 (67, 68). Although Consistent with this description, the total conversion of NO
data in Fig. 11 unambiguously indicate that the oxidation in Fig. 11 probably decreases at NO/C3H8 ꢉ 1 because ad-
of NO is mainly a heterogeneously catalyzed process over sorbed intermediates from the C3H8 + NO2 reaction (or-
H-ZSM5 (there is no significant NO + O2 reaction in an ganic nitro deposits are likely participants in forming N–N
empty reactor, designated as “Ho.”), the promoting effect bonds (1–3, 34, 48)) partly cover the active sites for the
of C3H8 suggests that the mechanisms of the homogeneous oxidation of NO to NO2. The reported periodic rate os-
and heterogeneous oxidations may include similar reaction cillations near 400ꢂC (34) may also be associated with the
steps. Fast interaction of NO with alkylperoxy radicals alternative accessibility of the active sites for the reactions
(RO2 + NO ⇒ RO + NO2) is a conceivable key reaction of NO + O2 and C3H8 + NO2. Olefins may not require BA-
for the hydrocarbon promoted homogeneous oxidation of OH sites to react with NO2. For instance, N2 production was
NO by O2 (67). Thus, the catalytic process may also involve observed in the direct reaction of NO2 and C3H6 (but not
radicals.
C3H8) over a Na-ZSM5 catalyst which is probably free of
As proposed before (34), the interaction of C3H8 BA-OH (3, 71). AlthoughsomeBA-OHsitemay bepresent
and NO2 (a free radical) is probably a homogeneous– on the surface of Na-ZSM5 (72, 73), it is likely that the easy
heterogeneous process over the H-ZSM5 catalyst involving adsorption of olefins on Lewis sites is more important here.
the formation of radical surface intermediates. Figures 11 to It is known that olefins readily adsorb on the Lewis acidic
16 demonstrate that this reaction requires a catalyst at our sites of oxides and are able to create Brønsted acidic sites
reaction conditions because even excess NO2 did not initi- from their own hydrogens for further reactions (74, 75). It
ate significant homogeneous reaction with C3H8. By com- is also possible that base catalysis occurs which is typical for
paring Figs. 13 and 14, one can conclude that the selective the alkaline exchanged zeolites (76, 77).
N2 production is mainly a catalytic process even at high
Near 600ꢂC, the oxidation of C3H8 by O2 is probably an
concentrations of reactants when the reaction temperature initial reaction step in the SCR process over H-ZSM5. As
is below 500ꢂC. In contrast, the SCR of NO by C3H8 was shown in Fig. 12, NO2 is an unlikely initial catalytic product

370
HALASZ ET AL.
at this temperature, although homogeneously formed NO2 the SCR process. At 600ꢂC (Fig. 12), excess C3H8 enhances
can participate in the SCR reaction. After correcting for the production of N2 because more C3H8 can produce more
the homogeneous (or wall-induced) formation of NO2 at CxHyOz to initiate reaction with NOx.
high O2/NO ratios (line “Ho.”), it is seen that significant
Since both the low and the high temperature SCR pro-
catalytic oxidation of NO to NO2 does not occur in the cesseslikelyinvolveradicals, itisunderstandablethatavari-
absence of propane (line “Cat.”). Thus, the conversion of ety of intermediates could be detected at different reaction
NO to N2 probably does not require the preliminary for- conditions as mentioned in the Introduction. Radical type
mation of NO2 at the reaction conditions of Fig. 12 (how- mechanisms have been recently proposed for several metal
ever, the homogeneously formed NO2 is also used up at exchanged ZSM5 zeolites (1–3, 48). Some of the metal ex-
NO/C3H8 ꢉ 1). This observation agrees with the earlier sug- changed ZSM5 samples may be more active for the SCR
gestion that both NO and NO2 may directly participate in of NO than H-ZSM5. By replacing the protons of H-ZSM5
the SCR process at 600ꢂC (34). Although significant CO with suitable multivalent metal cations, enhanced N2 pro-
production was observed even at this high reaction tem- duction can be observed when the metal cation can accel-
perature (average CO2/CO ratio is 1/2), it is well docu- erate the oxidative dehydrogenation of hydrocarbons (10,
mented that CO is not a good reductant for NO at net 19, 26, 27, 31, 34) or the oxidation of NO to NO2 (24) (both
oxidizing conditions (1, 21, 44, 51–53, 59, 78). Large CO reactions have been suggested to be rate determining (1–3,
production is typical when the oxidation of hydrocarbons 23, 38, 39, 48) or can increase the capacity of zeolites to ad-
involves the decomposition of alkoxy of alkyl peroxy inter- sorb hydrocarbons (69). Accelerated hydrocarbon burning
mediates (CxHyOz) to CO because the oxidation of CO is is probably not helpful because it will reduce the concen-
a relatively slow catalytic process (57, 58). CxHyOz deriva- tration of CxHyOz intermediates that may be required for
tives have been suggested to be possible intermediates in the subsequent interaction with NOx (87).
the SCR process over many catalysts (1, 6, 21, 40, 60, 61).
According to Fig. 17, the best efficiency for the SCR of
Although this possibility has rarely been considered for the NO by C3H8 to N2 is about 1.5 N2/C3H8. Since the direct
H-ZSM5 zeolite (1, 40), present results suggest that for- oxidation of C3H8 by O2 to COx is not significant when N2
mation of CxHyOz can be a viable initial step in the SCR /C3H8 ꢄ 1.5, this value can be used to calculate the stoi-
reaction over this catalyst above 400ꢂC.
chiometry of the SCR reaction. Hence, the third equation
The combustion of hydrocarbons by O2 is a typical radical in Table 1 is the closest to describe this stoichiometry at
process which can proceed homogeneously at temperatures 300ꢂC. Recalling that CO is virtually the only product from
much lower than 600ꢂC(57). However, thisprocess requires C3H8 at this temperature, the equation is
a catalyst at our reaction conditions. Figure 6 demonstrates
2C3H8 + 6NO + 4O2 ⇒ 6CO + 8H2O + 3N2.
that the BA-OH sites of H-ZSM5 are probably involved
in the active sites for this reaction. Since some catalytic
At higher temperatures, where CO2 formation was also ob-
activity remained after replacing the protons of H-ZSM5
served, the direct combustion of C3H8 by O2 (without N2
by Li+ ions, it is likely that the oxidation of propane can
production) is probably significant.
proceed on the Lewis acidic and basic sites of the zeolite.
This is reasonable, because such sites are capable of gen-
Although the H-ZSM5 catalyst itself can be more active
for certain SCR reactions than many other ZSM5 catalysts
erating oxygen and hydrocarbon radicals and an exchange
(2, 42–46), its hydrothermal stability above 600ꢂC (an im-
between gas phase and lattice oxygen can also occur (13–
portant practical consideration) is lower than that of the
16, 48, 59, 79–81). Lewis sites and lattice oxygens may also
metal exchanged ZSM5 zeolites (70, 88, 89). Our H-ZSM5
participate in the oxidation of NO(1, 12, 14, 79, 80, 82–85).
sample was used for more than 1.5 years. Despite large
The comparison of Figs. 3a and 16a indicates that the com-
amounts of H2O produced during the experiments at differ-
bustion of propane is inhibited in the presence of NO when
NO/C3H8 > 1 (excess NO). Consistent with the previous de-
ent reaction conditions, the activity of the catalyst for the
SCR of NO to N2 proved to be quite stable. Repeated mea-
scription, the strong adsorption of NO (probably mainly on
surements at the beginning and at the end of the 1.5-year
the BA-OH sites (65, 86)) may prevent the adsorption of
period showed less than 5% difference in the N2 production
propane at these conditions and, hence, its ignition by O2.
when the SCR experiments were carried out at identical re-
A similar inhibition effect of NO in the aromatization of
action conditions. However, the conversion of propane in
propane has been recently reported (65).
the oxidation of C3H8 by O2 decreased by about 30% during
The different trends in the N2 production when
this time. The XRD pattern of the sample did not change.
NO/C3H8 ꢉ 1 shown in Figs. 11 and 12 coincide with the
proposed different reaction routes at temperatures below
CONCLUSIONS
and above 500ꢂC. At 300ꢂC (Fig. 11), excess C3H8 (decreas-
ing NO/C3H8 ratio) has no significant effect on the conver-
sion of NO to N2 because the formation of NO2 controls
(1) The reaction of NO + C3H8 + O2 over H-ZSM5 cata-
lyst can result in the selective production of either NO2 or

CATALYTIC ACTIVITY AND SELECTIVITY OF H-ZSM5
371
N2. The selectivity for N2 largely depends on the ratio of
NO/C3H8 in the feedstock. The concentration of oxygen
and the reaction temperature have only minor impact on
the selectivity for N2.
(2) In agreement with prior findings, the SCR of NO by
C3H8 toN2 probably proceedsviatwodifferent mechanisms
at temperatures below and above 500ꢂC over H-ZSM5. It
is also likely that the initial reactions are different in these
temperature regimes. Homogeneous (or wall catalyzed)
and heterogeneous reactions of radicals can be involved
in both the initial and the subsequent reaction steps.
5. Burch, R., and Scire, S., Appl. Catal. B 3, 295 (1994).
6. Cho, B. K., J. Catal. 142, 418 (1993).
7. Iwamoto, M., and Mizuno, N., J. Automotive Eng. 207, 23 (1993).
8. Petunchi, J. O., Sill, G., and Hall, W. K., Appl. Catal. B 2, 303 (1993).
9. Ansell, G. P., Diwell, A. F., Golunski, S. E., Hayes, J. W., Rajaram,
R. R., Truex, T. J., and Walker, P., Appl. Catal. B: Environmental 2, 81
(1993).
10. Larsen, S. C., Aylor, A., Bell, A. T., and Reimer, J. A., J. Phys. Chem.
98, 11533 (1994).
11. Moretti, G., Catal. Lett. 23, 135 (1994).
12. Lei, G. D., Adelman, B. J., Sarkany, J., and Sachtler, W. M. H., Appl.
Catal. B 5, 245 (1995).
13. Loktev, M. I., and Slinkin, A. A., Russ. Chem. Rev. 45, 807 (1976).
(3) Oxidation of C3H8 by O2 is a likely initial step in the 14. Rabo, J. A., and Gajda, G. J., Catal. Rev. 31, 385 (1989–1990).
SCR process near 600ꢂC. Both NO and NO2 may directly
participate in the SCR process at these high temperatures
and the interaction of NOx and CxHyOz is a possible reac-
tion step.
15. Volodin, A. M., Bolshov, V. A., and Panov, G. I., J. Phys. Chem. 98,
7548 (1994).
16. Valyon, J., Millman, W. S., and Hall, W. K., Catal. Lett. 24, 215 (1994).
17. Li, Y., and Armor, J. N., J. Catal. 145, 1 (1994).
18. Petunchi, J. O., and Hall, W. K., Appl. Catal. B 3, 239 (1994).
19. Nishizaka, Y., and Misono, M., Chem. Lett. 2237 (1994).
(4) The SCR process may begin with the oxidation of NO
to NO2 below 500ꢂC. There is no significant homogeneous
20. Li, Y., and Armor, J. N., Appl. Catal. B 2, 239 (1993).
interaction between NO2 and C3H8 at these temperatures. 21. Hamada, H., Kintaichi, Y., Yoshinari, T., Tabata, M., Sasaki, M., and
Ito, T., Catal. Today 17, 111 (1993).
22. Centi, G., Perathoner, S., and Dall’Olio, L., Appl. Catal. B 4, L275
When the partial pressure of NO exceeds the partial pres-
sure of C3H8 in the SCR feed, the catalytic interaction of
NO2 and C3H8 is also inhibited.
(1994).
23. Halasz, I., Brenner, A., and Ng, S. K., Catal. Lett. 34, 151 (1995).
(5) The oxidation of NO by O2 and the oxidation of C3H8
by O2 are both catalytic processes and the active sites of H-
ZSM5 probably involve BA-OH groups. These groups may
also be involved in the active sites for the reaction of NO2
with C3H8.
24. Shelef, M., Montreuil, C. N., and Jen, H. W., Catal. Lett. 26, 277
(1994).
25. Shimokawabe, M., Ohi, A., and Takezawa, N., React. Kinet. Catal. Lett.
52, 393 (1994).
26. Bethke, K. A., Li, C., Kung, M. C., Yang, B., and Kung, H. H., Catal.
Lett. 31, 287 (1995).
(6) The concentration of oxygen and the reaction tem-
perature affect the rates of the reactions C3H8 + O2 and
NO + O2, therefore these parameters also affect the pro-
duction of N2 in the SCR process.
(7) Since significant coke formation was not observed in
the absence of oxygen, it is an unlikely intermediate in the
SCR process. The direct interaction of NO and C3H8 is also
an unlikely reaction step.
(8) The efficiency of propane to produce N2 varies
from about 0.1 to 1.8 N2/C3H8. The efficiency decreases
with increasing temperatures as well as with increasing
O2 and C3H8 concentrations in the feedstock. The likely
overall stoichiometry of the SCR process at 300ꢂC is
2C3H8 + 6NO + 4O2 ⇒ 6CO + 8H2O + 3N2. Athighertem-
peratures, formation of CO2 also occurs.
27. Chajar, Z., Primet, M., Praliaud, H., Chevrier, M., Gauthier, C., and
Mathis, F., Catal. Lett. 28, 33 (1994).
28. Bell, V. A., Feeley, J. S., Deeba, M., and Farrauto, R. J., Catal. Lett. 29,
15 (1994).
29. Obuchi, A., Ogata, A., Mizuno, K., Ohi, A., Nakamura, M., and
Ohuchi, H., J. Chem. Soc. Chem. Commun. 247 (1992).
30. Obuchi, A., Nakamura, M., Ogata, A., Mizuno, K., Ohi, A., and
Ohuchi, H., J. Chem. Soc. Chem. Commun. 1150 (1992).
31. Poignant, F., Saussey, J., Lavalley, J.-C., and Mabilon, G., J. Chem. Soc.
Chem. Commun. 89 (1995).
32. Adelman, B. J., Lei, G.-D., and Sachtler, W. M. H., Catal. Lett. 28, 119
(1994).
33. Sasaki, M., Hamada, H., Kintaichi, Y., and Ito, T., Catal. Lett. 15, 297
(1992).
34. Halasz, I., Brenner, A., Shelef, M., and Ng, S. K., J. Phys. Chem. 99,
17186 (1995).
35. Radtke, F., Koeppel, R. A., and Baiker, R. A., Catal. Lett. 28, 131
(1994).
36. Hayes, N. W., Gru¨nert, W., Hutchings, G. J., Joyner, R. W., and Shpiro,
E. S., J. Chem. Soc. Chem. Commun. 531 (1994).
37. Radtke, F., Koeppel, R. A., and Baiker, A., J. Chem. Soc. Chem. Com-
mun. 427 (1995).
ACKNOWLEDGMENTS
The authors thank the Ford Motor Company and the Institute for Man-
ufacturingResearchatWayneStateUniversityforsupportofthisresearch. 38. Bennett, C. J., Bennett, P. S., Golunski, S. E., Hayes, J. W., and Walker,
We also thank Mr. Michael Lukitsch for determining the NH+4 exchange
capacity of the H-ZSM5 sample.
A. P., Appl. Catal. A 86, L1 (1992).
39. Cowan, A. D., Du¨mpelmann, R., and Cant, N. W., J. Catal. 151, 356
(1995).
REFERENCES
40. Hamada, H., Kintaichi, Y., Sasaki, M., Ito, T., and Tabata, M., Appl.
Catal. 64, L1 (1990).
1. Shelef, M., Chem. Rev. 95, 209 (1995).
41. Hamada, H., Kintaichi, Y., Sasaki, M., Ito, T., and Tabata, M., Appl.
Catal. 70, L15 (1991).
42. Hamada, H., Kintaichi, Y., Tabata, M., Sasaki, M., and Ito, T., “67th
CATSJ Meeting Abstracts,” No. A1, Vol. 33, p. 59. 1991.
2. Witzel, F., Sill, G. A., and Hall, W. K., J. Catal. 149, 229 (1994).
3. Yokoyama, C., and Misono, M., J. Catal. 150, 9 (1994).
4. Li, Y., Slager, T. L., and Armor, J. N., J. Catal. 150, 388 (1994).

372
HALASZ ET AL.
43. Yogo, K., Tanaka, S., Ihara, M., Hishiki, T., and Kikuchi, E., Chem. 68. Hori, M., Matsunaga, N., Malte, P. C., and Marinov, N. M., “Proceed-
Lett. 1025 (1992).
ings of the Twenty-Fourth International Symposium on Combustion
Sydney, Australia,” p. 899. Organized by The Combustion Institute,
Pittsburgh, Pennsylvania, 1992.
44. Hamada, H., Kintaichi, Y., Tabata, M., Sasaki, M., and Ito, T., Sekiyu
Gakkaishi 36, 149 (1993).
45. Yogo, K., Umeno, M., Watanabe, H., and Kikuchi, E., Catal. Lett. 19,
131 (1993).
46. Yogo, K., Ihara, M., Terasaki, I., and Kikuchi, E., Appl. Catal. B 2, L1
(1993).
47. Satsuma, A., Yamada, K., Mori, T., Niwa, M., Hattori, T., and
Murakami, Y., Catal. Lett. 31, 367 (1995).
48. Lukianov, D. B., Sill, G., d’Itri, J. L., and Hall, W. K., J. Catal. 153, 265
(1995).
69. Jen, H. W., and Otto, K., Catal. Lett. 26, 217 (1994).
70. Grinsted, R. A., Jen, H. W., Montreuil, C. N., Rokosz, M. J., and Shelef,
M., Zeolites 13, 602 (1993).
71. Yokoyama, Ch., Yasuda, H., and Misono, M., 71th CATSJ Meeting
Abstracts: No. 1B6 35, 125 (1993).
72. d’Itri, J. L., and Sachtler, M. H., Appl. Catal. B 2, L7 (1993).
73. Parrillo, D. J., Dolonec, D., Gorte, R. J., and McCabe, R. W., J. Catal.
142, 708 (1993).
49. Li, Y., and Armor, J. N., Appl. Catal. B 1, L31 (1992).
50. Martin, A., and Berndt, H., React. Kinet. Catal. Lett. 52, 405 (1994).
74. Irvine, E. A., John, C. S., Kemball, C., Pearman, A. J., Day, M. A., and
Sampson, R. J., J. Catal. 61, 326 (1980).
51. Halasz, I., Brenner, A., Shelef, M., Appl. Catal. B: Environmental, 2, 75. Gati, Gy., and Halasz, I., J. Catal. 82, 223 (1983).
131 (1993).
76. Garces, J. M., Vrieland, G. E., Bates, S. I., and Scheidt, F. M., Stud.
52. Halasz, I., Brenner, A., and Shelef, M., Catal. Lett. 22, 147 (1993).
53. Halasz, I., Brenner, A., and Shelef, M., Catal. Lett. 16, 311 (1992).
54. Halasz, I., Brenner, A., Shelef, M., and Ng, K. Y. S., J. Catal. 126, 109
(1990).
55. Halasz, I., Brenner, A., Shelef, M., and Ng, K. Y. S., Catal. Lett. 6, 349
(1990).
Surf. Sci. Catal. 20, 67 (1985).
77. Martens, L. R. M., Grobet, P. J., and Jacobs, P. A., Nature 315, 568
(1985).
78. d’Itri, J. L., and Sachtler, W. M. H., Catal. Lett. 15, 289 (1992).
79. Kasai, P. H., and Bishop, R. J., Jr., ACS Monogr. 171, 350 (1976).
80. Poutsma, M. L., ACS Monogr. 171, 529 (1976).
81. Chang, Y.-F., Somorjai, G. A., and Heinemann, H., J. Catal. 154, 24
(1995).
56. Brandin, J. G. M., Andersson, L. A. H., and Obenbrand, C. U. I., Acta
Chem. Scand. 44, 784 (1990).
57. Benson, S. W., Prog. Energy Combust. Sci. 7, 125 (1981).
58. Krylov, O. V., and Shub, B. R., “Nonequilibrium Processes in Cataly-
sis.” CRC Press, Boca Raton, Ann Arbor, London, Tokyo, 1994.
82. Ve´drine, J. C., Auroux, A., Bolis, V., Dejaifve, P., Naccache, C., Wierzh-
chowski, P., Derouane, E. G., Nagy, J. B., Gilson, J. P., van Hooff,
J. H. C., van den Berg, J. P., and Wolthuizen, J., J. Catal. 59, 248 (1979).
59. Yu. Sinev, M., Ya. Margolis, L., and Korchak, V. N., Russ. Chem. Rev. 83. Witzel, F., Karge, H. G., Gutsze, A., and Hartel, U., Chem. Ing. Tech.
64, 349 (1995).
63, 744 (1991).
60. Iwamoto, M., and Hamada, H., Catal. Today 10, 57 (1991).
61. Montreuil, C. N., and Shelef, M., Appl. Catal. B. 1, L1 (1992).
62. Taylor, K. C., Catal. Rev. Sci. Eng. 35, 457 (1992).
63. Malysheva, L. V., Paukshtis, E. A., and Ione, K. G., Catal. Rev. Sci.
Eng. 37, 179 (1995).
64. Giannetto, G., Monque, R., and Galiasso, R., Catal. Rev. Sci. Eng. 36,
271 (1994).
65. Buckles, G., and Hutchings, G. J., J. Catal. 151, 33 (1995).
66. Meusinger, J., and Corma, A., J. Catal. 152, 189 (1995).
84. Anderson, L. A. H., Brandin, J. G. M., and Odenbrand, C. U. I., Catal.
Today 4, 173 (1989).
85. Brandin, J. G. M., Anderson, L. A. H., and Odenbrand, C. U. I, Catal.
Today 4, 187 (1989).
86. Hoost, T. E., Laframboise, K. A., and Otto, K., Catal. Lett. 33, 105
(1995).
87. Nishizaka, Y., and Misono, M., Chem. Lett. 1295 (1993).
88. Suzuki, K., Sano, T., Shoji, H., Murakami, T., Ikai, S., Shin, S., Hagi-
wara, H., and Takaya, H., Chem. lett. 1507 (1987).
67. Bromly, J. H., Barnes, F. J., Mandyczewsky, R., Edwars, T. J., and 89. Iwamoto, M., Yahiro, H., Shin, H. K., Watanabe, M., Guo, J., Konno,
Haynes, B. S., “Proceedings of the Twenty-Fourth International Sym- M., Chikahisa, T., and Murayama, T., Appl. Catal. B 5, L1 (1994).
posium on Combustion Sydney, Australia,” p. 899. Organized by the 90. Ciambelli, P., Corbo, P., Gambino, M., Minelli, G., Moretti, G., and
Combustion Institute, Pittsburgh, Pennsylvania, 1992. Porta, P., Catal. Today 26, 33 (1995).