Oxidation of H2 and CO over ion-exchanged X and Y zeolites

Article abstract of DOI:10.1021/ja0671520

Zeolites X and Y exchanged with Group IA cations were synthesized by aqueous ion exchange of NaX and NaY and used as catalysts in the oxidation of H₂ and CO at temperatures ranging from 473 to 573 K. The CsX zeolite was the most active material of the series for both reactions whereas HX was the least active. Moreover, the oxidation of CO in H₂ was very selective (～80%) over the alkali-metal exchanged materials. Isotopic transient analysis of CO oxidation during steady-state reaction at 573 K was used to evaluate the coverage of reactive carbon-containing intermediates that lead to product as well as the pseudo-first-order rate constant of the reaction. A factor of 4 enhancement in activity achieved by exchanging Cs for Na was attributed to a higher coverage of reactive intermediates in CsX because the pseudo-first-order rate constant was nearly same for the two materials (～0.7 s^-1). The number of reactive intermediates on both materials was orders of magnitude below the number of alkali metal cations in the zeolites but was similar to the number of impurity Fe atoms in the samples. Because the trend in Fe impurity loading was the same as that for oxidation activity, a role of transition metal impurities in zeolite oxidation catalysis is suggested.

Full text of DOI:10.1021/ja0671520

Published on Web 02/24/2007
Oxidation of H2 and CO over Ion-Exchanged X and Y Zeolites
Daniel G. Lahr, Junhui Li, and Robert J. Davis*
Contribution from the Department of Chemical Engineering, UniVersity of Virginia,
CharlottesVille, Virginia 22904-4741
Received October 5, 2006; E-mail: rjd4f@virginia.edu
Abstract: Zeolites X and Y exchanged with Group IA cations were synthesized by aqueous ion exchange
of NaX and NaY and used as catalysts in the oxidation of H2 and CO at temperatures ranging from 473 to
573 K. The CsX zeolite was the most active material of the series for both reactions whereas HX was the
least active. Moreover, the oxidation of CO in H2 was very selective (∼80%) over the alkali-metal exchanged
materials. Isotopic transient analysis of CO oxidation during steady-state reaction at 573 K was used to
evaluate the coverage of reactive carbon-containing intermediates that lead to product as well as the pseudo-
first-order rate constant of the reaction. A factor of 4 enhancement in activity achieved by exchanging Cs
for Na was attributed to a higher coverage of reactive intermediates in CsX because the pseudo-first-order
rate constant was nearly same for the two materials (∼0.7 s-1). The number of reactive intermediates on
both materials was orders of magnitude below the number of alkali metal cations in the zeolites but was
similar to the number of impurity Fe atoms in the samples. Because the trend in Fe impurity loading was
the same as that for oxidation activity, a role of transition metal impurities in zeolite oxidation catalysis is
suggested.
Introduction
balancing zeolite cations lowers the energy required to excite
charge transfer from the hydrocarbon to dioxygen. Therefore,
Frei recently summarized the rapid progress achieved on the
selective partial oxidation of hydrocarbons in faujasite-type
zeolites,1 which has been a growing area of research over the
past decade.2-15 One major difficulty encountered during partial
oxidation of hydrocarbons with molecular oxygen is the
formation of combustion side products formed at the elevated
temperatures typically required for the reaction. Moreover, the
higher reactivity of partially oxidized products compared to
hydrocarbons exacerbates the problem of low selectivity. New
routes to selectively oxidize molecules are clearly needed. Frei
et al. discovered that photoactivation of hydrocarbon and
dioxygen molecules loaded into a zeolite cage can yield selective
oxidation products at relatively low temperatures.1-6 They
suggest the high electric field associated with the charge-
excitation by visible light or by mild heating can initiate the
oxidation reaction to produce a radical hydrocarbon cation and
a superoxide (O2-) anion. The highly acidic radical cation then
transfers a proton to the superoxide to form a hydroperoxide
molecule, which decomposes to yield a selective oxidation
product.1 Although this novel approach to selective oxidation
avoids the high temperatures typically encountered in oxidation
reactions, the microporous nature of the zeolite prevents
desorption of the products to complete the catalytic cycle.
Xu et al. followed the selective oxidation of propane to
acetone over various ion-exchanged Y zeolites in the absence
of irradiation and found the reaction rate increased in the order
BaY < MgY < SrY < CaY.12 In a follow-up study, they studied
the same reaction over various Ca-exchanged zeolites and found
that activity and selectivity were a function of zeolite type.13
For example, high selectivity to partial oxidation was observed
over CaY whereas deep oxidation of propane to CO and CO2
was observed over Ca-exchanged mordenite, presumably be-
cause of Bronsted acid sites present in the latter material. For
hydrocarbon oxidation, alkaline-earth exchanged zeolites are
generally found to be more active than alkali-metal exchanged
zeolites, presumably because of the high electric field associated
with divalent cations compared to monovalent cations.3 Indeed,
Xu et al. prepared a series of NaY zeolites containing various
levels of ion-exchanged Ca and studied these materials in the
selective oxidation of propane.14 The acetone formation rate
correlated well with Ca loading in the zeolite supercages.
(1) Frei, H. Science 2006, 313, 309.
(2) Blatter, F.; Frei, H. J. Am. Chem. Soc. 1993, 115, 7501.
(3) Blatter, F.; Sun, H.; Vasenkov, S.; Frei, H. Catal. Today 1998, 41, 297.
(4) Vasenkov, S.; Frei, H. J. Phys. Chem. B 1998, 102, 8177.
(5) Frei, H.; Blatter, F.; Sun, H. U.S. Patent 5,914,013, assigned to the Regents
of the University of California, June 22, 1999.
(6) Frei, H.; Blatter, F.; Sun, H. U.S. Patent 5,827,406, assigned to the Regents
of the University of California, Oct. 27, 1998.
(7) Tang, S. L. Y.; McGarvey, D. J.; Zholobenko, V. L. Phys. Chem. Chem.
Phys. 2003, 5, 2699.
(8) Panov, A. G.; Larsen, R. G.; Totah, N. I.; Larsen, S. C.; Grassian, V. H.
J. Phys. Chem. B 2000, 104, 5706.
(9) Myli, K. B.; Larsen, S. C.; Grassian, V. H. Catal. Lett. 1997, 48, 199.
(10) Xiang, Y.; Larsen, S. C.; Grassian, V. H. J. Am. Chem. Soc. 1999, 121,
5063.
(11) Larsen, R. G.; Saladino, A. C.; Hunt, T. A.; Mann, J. E.; Xu, M.; Grassian,
V. H.; Larsen, S. C. J. Catal. 2001, 204, 440.
(12) Xu, J.; Mojet, B. L.; van Ommen, J. G.; Lefferts, L. Phys. Chem. Chem.
Phys. 2003, 5, 4407.
(13) Xu, J.; Mojet, B. L. Lefferts, L. Micropor. Mesopor. Mater. 2006, 91, 187.
(14) Xu, J.; Mojet, B. L.; van Ommen, J. G.; Lefferts, L. J. Phys. Chem. B
2004, 108, 15728.
A recent quantum chemical study by Pidko and van Santen
explored the initial charge-transfer step proposed for selective
(15) Vanoppen, D. L.; De Vos, D. E.; Jacobs, P. A. J. Catal. 1998, 177, 22.
9
3420
J. AM. CHEM. SOC. 2007, 129, 3420-3425
10.1021/ja0671520 CCC: $37.00 © 2007 American Chemical Society

Oxidation of H2 and CO over Zeolites
A R T I C L E S
photo-oxidation reactions in zeolites.16 They studied the adsorp-
tion and reaction of 2,3-dimethyl-2-butene in a model pore of
the Y-zeolite supercage with two alkaline earth cations in
appropriate positions. In contrast to earlier studies, they suggest
that the electrostatic field of the zeolite pore is not the key
feature of the system. Instead, they found the relative orientation
and distance between 2,3-dimethyl-2-butene and O2 to be the
important factor. From their perspective, the zeolite pore
organizes the reagents into a pre-transition state arrangement
via interactions with the exchange cations. This model accounts
for CaY being more active that MgY and SrY, because the Ca
ions are of the appropriate size in the zeolite cage. The smaller
Mg and larger Sr cations did not produce as suitable a
configuration of 2,3-dimethyl-2-butene and O2 as did the Ca
cations.
at 363 K and calcined at 773 K for 5 h. An HX sample was prepared
by ion-exchanging NaX with ammonium nitrate (Acros, 99%) as
described above. However, after low-temperature drying, the material
was evacuated at room temperature for 2 h and heated under vacuum
at 723 K for 5 h to remove ammonia from the catalyst. Elemental
composition of these samples, determined by ICP analysis, was
performed by Galbraith Labs, Inc. (Knoxville, TN).
Addition of iron to zeolite X was accomplished by sublimation of
FeCl3 as described by Chen and Sachtler.21 Approximately 3 g of zeolite
(NaX and HX) was placed on one side of a U-tube and heated under
helium flow at 623 K for 2 h. The FeCl3 was placed in the other side
of the U-tube, and the entire system was heated to 623 K to sublime
iron chloride into the zeolite. The yellow zeolite product was then
washed with distilled deionized water and calcined at 773 K for
5 h.
Oxidation Reactions. The zeolite catalyst was loaded into a single-
pass fixed bed reactor system operating at atmospheric pressure. Prior
to reaction, the catalyst was heated in situ in flowing helium at 773 K
for 5 h. The reactant gases CO (BOC 99.997%) and O2 (BOC, 99.999%)
were both purified by passage through a silica gel trap immersed in
dry ice acetone, whereas H2 (BOC, 99.999%) was purified by a Supelco
OMI-2 filter. Argon (BOC, 99.999%) and helium (BOC, 99.999%) were
purified in sequence by a Supelco OMI-2 filter and a silica gel trap
held in dry ice acetone. The oxidation reactions involved 5 mol % H2
or CO, 2.5 mol % dioxygen, and the balance inert. To improve
sensitivity in the thermal conductivity detector, helium was used during
carbon monoxide oxidation reactions whereas argon was used during
dihydrogen reactions. In some cases, both dihydrogen and carbon
monoxide were fed to the oxidation reactor. These PROX conditions
were 40 mol % dihydrogen, 5 mol % carbon monoxide, 2.5 mol %
dioxygen, and the balance helium. The total flow rate in all cases was
54 mL min-1. Products were monitored with an online gas chromato-
graph equipped with a Alltech CTR I column and a thermal conductivity
detector using helium or argon as the reference gas depending on the
inert gas fed to the reactor.
Recent work in our lab revealed that both butane and butene
are catalytically oxidized over Cs and Na exchanged X zeolites
at moderate (474-573 K) temperatures.17 However, we ob-
served gas-phase CO2 without any significant quantity of CO
in the product stream. In an effort to better understand oxidation
catalysis in zeolites, we decided to explore the steady-state
carbon monoxide oxidation and dihydrogen oxidation over
alkali-metal and alkaline-earth exchanged zeolites. However,
we realize that these steady-state reactions require higher
temperatures than those typically used during photo-oxidation
reactions in zeolites. Because CO and H2 oxidation reactions
do not form multiple oxidation products, a direct comparison
with partial oxidation of hydrocarbons will be limited.
Carbon monoxide oxidation is a very well-studied reaction
because of its utilization in automotive emissions control. More
recently, preferential oxidation (PROX) of dilute carbon mon-
oxide in dihydrogen streams over zeolite-supported transition
metal catalysts has attracted attention for low-temperature PEM
fuel cell operations.18-20 Although PROX is effectively cata-
lyzed by an expensive transition metal such as Pt, a suitable
non-transition metal catalyst would be highly desirable.
In this study, we used isotopic transient analysis of the CO
oxidation reaction to probe the number of reactive intermediates
and the pseudo-first-order rate constant of the oxidation reaction
occurring at the steady state. In addition, the rates of dihydrogen
oxidation are compared directly to those of CO oxidation over
a series of ion-exchanged zeolites.
Isotopic Transient Measurements. Approximately 0.3 g of catalyst
was loaded into a fixed bed tubular reactor connected directly to a mass
spectrometer. The same gases, purification methods, and catalyst
pretreatments were used as described above. In addition, 13C-labeled
carbon monoxide, 13CO, (Cambridge Isotopes, 99.5% CO, 98+% 16O,
99+% 13C) was used to evaluate the number of reactive intermediates
and the relaxation time of the system. Isotopic transient experiments
involving ammonia synthesis and carbon monoxide oxidation have been
performed extensively in our lab,22-27 and the reactor system used in
this study is the same as the one used previously. Reaction conditions
were 2.5% carbon monoxide, 2.5% dioxygen, 94% helium, and an
additional 1% argon or helium. Volumetric flow rates ranged from 160-
300 mL min-1. A Balzers-Pfieffer Prisma 200 amu mass spectrometer
was used to monitor the concentrations of argon and the carbon dioxide
isotopomers, 12CO2 and 13CO2, in the effluent stream.
Experimental Procedures
Catalyst Preparation. Ion-exchanged zeolites were prepared ac-
cording to methods described previously.17 In summary, NaX (Union
Carbide, lot 07483-36) or NaY (Union Carbide, lot 955089001010-S)
was triply ion-exchanged with 1 M aqueous solutions (10 mL g-1
original zeolite) of the appropriate salt. Cesium acetate (Aldrich 99.9%),
potassium nitrate (Aldrich 99+%), sodium acetate (Acros 99+%),
barium acetate (Aldrich 99.999%), strontium nitrate (Acros ACS grade),
calcium nitrate tetrahydrate (Acros ACS grade), and magnesium acetate
hexahydrate (Acros 99%) were used as ion precursors. After the ion-
exchange steps, each 3 g sample was triply washed with 1 L of distilled
deionized water for 5 h. Each sample was then dried overnight in air
First, the carbon monoxide reaction was run at the steady state. Then,
an isotopic switch in one of the reactants, i.e., 12CO/13CO, is performed
while keeping all other variables constant. Care was taken to keep the
pressure constant at 2.6 bar throughout the switch. To account for the
gas-phase holdup, an inert tracer (Ar) of 1 mol % was also switched
with the carbon monoxide. For more information on the method, see
the excellent review of Shannon and Goodman.28
(21) Chen, H.-Y.; Sachtler, W. M. H. Catal. Lett. 1998, 50, 125.
(22) McClaine, B. C.; Davis, R. J. J. Catal. 2002, 210, 387.
(23) McClaine, B. C.; Davis, R. J. J. Catal. 2002, 211, 379.
(24) Siporin, S. E.; Davis, R. J. J. Catal. 2004, 222, 315.
(25) Calla, J. T.; Davis, R. J. J. Phys. Chem. B 2005, 109, 2307.
(26) Calla, J. T.; Bore, M. T.; Datye, A. K.; Davis, R. J. J. Catal. 2006, 238,
458.
(16) Pidko, E. A.; van Santen, R. A. J. Phys. Chem. B 2006, 110, 2963.
(17) Li, J.; Tai, J.; Davis, R. J. Catal. Today 2006, 116, 226.
(18) Igarashi, H.; Uchida, H.; Suzuki, M.; Sasaki, Y.; Watanabe, M. Appl. Catal.
A: Gen. 1997, 159, 159.
(19) Watanabe, M.; Uchida, H.; Ohkubo, K.; Igarashi, H. Appl. Catal. B: EnV.
2003, 46, 595.
(20) Rosso, I.; Galletti, C.; Saracco, G.; Garrone, E.; Specchia, V. Appl. Catal.
B: EnV. 2004, 48, 195.
(27) Calla, J. T.; Davis, R. J. J. Catal. 2006, 241, 407.
(28) Shannon, S. L.; Goodwin, J. G., Jr. Chem. ReV. 1995, 95, 677.
9
J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007 3421

A R T I C L E S
Lahr et al.
Table 1. Unit Cell Composition and Impurity Iron Content of Ion
Table 3. Rates of H2 and CO Oxidation over Alkaline-Earth
Exchanged X Zeolitesa
Exchanged X and Y Zeolitesa
H2 oxidation
reaction rate
CO oxidation
ion
exchange
(%)
iron
reaction rate
pore volume
loading
(ppm)
1
1
(µ
mol g-1 s-
)
(
µ
mol g-1 s-
)
1
sample
(cm3 g -
)
unit cell composition
sample
573 K
473 K
573 K
473 K
HX
NaX
KX
CsX
NaY
CsY
0.21
0.32
0.25
0.19
NDb
NDb
H50Na33Si109Al83O384
H3Na80Si109Al83O384
H13Na9K61Si109Al83O384
H6Na34Cs43Si109Al83O384
H5Na48Si139Al53O384
60
96
73
52
91
70
176
282
379
442
NDb
204
MgX
CaX
SrX
1.23
0.49
1.57
1.84
0.23
0.26
0.16
0.31
0.03
0.08
0.1
0.005
0.010
0.017
0.009
BaX
0.06
H2Na14Cs37Si139Al53O384
a Operating conditions: 54 mL min-1, 5% H2 or CO, 2.5% O2, balance
inert, 0.3 g catalyst.
a Protons were added to unit cell compositions to balance framework
charge. b ND: not determined.
Table 4. Activity and Selectivity of Alkali-Metal Exchanged X
Zeolites during Preferential Oxidation of CO in H2 at 573 Ka
Table 2. Rates of H2 and CO Oxidation over Group IA X and Y
Zeolitesa
CO oxidation rate
1)
H2 oxidation
reaction rate
CO oxidation
sample
PROX selectivityb
(
µ
mol g-1 s-
reaction rate
1
1
NaX
KX
CsX
81
77
85
0.08
0.07
0.35
(
µ
mol g-1 s-
)
(
µ
mol g-1 s-
)
sample
573 K
473 K
573 K
473 K
HX
NaX
KX
CsX
NaY
KY
0.2
1.71
2.86
10.43
0.4
0.10
0.55
0.92
1.31
0.14
0.18
0.22
0.08
0.14
0.1
0.44
0.04
0.06
0.12
0.012
0.022
0.018
0.039
0.006
0.009
0.010
a Reaction Conditions: 54 mL min-1, 40 mol % H2, 5 mol % CO, 2.5
mol % O2, balance helium. b Defined as (moles of CO2 produced/(2 × moles
of O2 consumed)).
more active than KY. In general, the relative rates between the
X and Y zeolites averaged to about 5 for dihydrogen oxidation
and about 3 for carbon monoxide oxidation. Apparently, the
number of zeolite cations influenced the overall oxidation rates
because Y zeolites contain fewer of them than X zeolites. The
number of accessible cations in Site III positions varies between
X-type and Y-type zeolites by a factor of 4.8, and after
accounting for Type II sites, the cation ratio is 1.8.
Table 3 summarizes the oxidation activity over alkaline-earth
zeolites. In this case, no clear trend in oxidation activity was
seen. Moreover, the alkaline-earth zeolites were less active for
both CO and H2 oxidation than the corresponding alkali metal
analogues, i.e., CsX > BaX, KX > CaX, and NaX > MgX.
Although water produced by H2 oxidation may hydroxylate the
multivalent cations during reaction, this should not happen
during CO oxidation. Prior work in the area of selective
hydrocarbon oxidation over zeolites generally reports increased
activity for alkaline-earth exchanged zeolites compared to alkali-
metal exchanged zeolites because of the larger local electric
field created by alkaline earth cations. If the local electric field
in the vicinity of the cation were the key feature in determining
the oxidation rate, we would expected the rates to follow the
trend CaX > NaX > CsX, which is opposite to what we found
in this work.
Reactivity results from the preferential oxidation of CO in
the presence of excess H2 (PROX) are reported in Table 4.
Evidently, the presence of CO greatly inhibits the H2 oxidation
rate because the selectivity of the PROX reaction is about 80%
to CO2 over the alkali-metal exchanged X zeolites at 573 K
and nearly 100% at 473K (not shown). The dihydrogen
oxidation activity regains its original level after removing CO
from the feed stream for 3 h. The apparent competition of CO
with H2 is consistent with a surface-mediated catalytic reaction
involving adsorbed CO.
In an effort to quantify the number of reactive intermediates
in the zeolite during CO oxidation, we analyzed the isotopic
transient of the product CO2 after a switch in the 13C content
of the reactant CO. The total flow rate and average CO2
concentration were both varied to explore the influence of
0.74
1.78
CsY
a Operating conditions: 54 mL min-1, 5% H2 or CO, 2.5% O2, balance
inert, 0.03 g catalyst for NaX, KX, CsX, and H2 at 573 K, otherwise 0.3 g
catalyst.
Results and Discussion
Elemental analysis of the zeolite samples was used to
calculate the unit cell compositions shown in Table 1 and
confirmed the expected levels of ion-exchange in the zeolites.
Because the Cs+ ion has a radius of 1.7 Å, it is too large to
exchange with sodium in the sodalite cages. Thus, it is not
surprising to see a fairly modest level of Cs exchange in the X
zeolite. Pore volumes, calculated as the amount of liquid
dinitrogen adsorbed in the zeolite pores at a relative pressure
(P/Po) of 0.3 are also shown in Table 1. The pore volume of
the X zeolites decreased as the cation size increased, with the
exception of HX. Although it is likely that HX partially degraded
during preparation, the reasonably high pore volume indicates
the zeolite was mostly intact.
The reaction rates of dihydrogen oxidation and carbon
monoxide oxidation over the ion-exchanged zeolites are sum-
marized in Table 2. The reactor was operated at low conversion
(under 10%) to ensure the differential behavior. Although a
steady state was achieved within 30 min, the reactions were
performed for over 4 h.
Several interesting trends were observed in the oxidation of
dihydrogen and carbon monoxide over the ion-exchanged
zeolites. First, the dihydrogen oxidation rate was much higher
than the carbon monoxide reaction rate over all zeolites. The
reaction rates were also a strong function of zeolite cation.
Dihydrogen oxidation over CsX was over an order of magnitude
faster than over HX, whereas the same comparison for CO
oxidation shows only a 3-5-fold increase. Moreover, the trend
in the dihydrogen oxidation rate was CsX > KX > NaX >
HX. Although CsX was the most active catalyst for CO
oxidation, there was no clear trend with the other X zeolites.
Table 2 also reveals that CsX was about 6 times more active
for dihydrogen oxidation than CsY, and KX was about 4 times
9
3422 J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007

Oxidation of H2 and CO over Zeolites
A R T I C L E S
Figure 1. Example of an isotopic transient during CO oxidation over CsX
at 573 K.
interparticle and intraparticle readsorption on the observed
kinetics. The on-line mass spectrometer used to follow the
transient monitored the gas-phase concentration of Ar and the
two carbon dioxide isotopomers, 12CO2 and 13CO2. Because
X-type zeolites adsorb significant quantities of water at elevated
temperatures, isotopic transients during H2 oxidation were not
studied.
Figure 2. Effect of flow rate and average CO2 concentration in the reactor
on observed relaxation time, τ, over CsX at 573 K. The CO2 concentration
was the average of the inlet and outlet concentrations.
The average residence time of surface intermediates that lead
to reaction products is found by integrating the normalized
transient response after accounting for gas-phase hold-up. This
is shown in eq 1.
τ ) ∞ (F13CO- FAr)dt
(1)
∫
0
2
where τ is the mean residence time of reactive carbon-containing
intermediates leading to product and Fi are the normalized
transient responses of CO2 and Ar, respectively. Figure 1 shows
a typical normalized transient response with Ar and 13CO2.
Steady-state oxidation of CO over both CsX and NaX was
analyzed using the isotopic transient method. Three flow rates
(160, 210, and 300 mL min-1) and three average carbon dioxide
concentrations (0.75, 0.95, and 1.2 mol %) were used for nine
τ measurements. The overall conversion of CO in our experi-
ments was less than 5%, corresponding to 0.1-0.2 mol % CO2.
Thus, all conditions involved a significant amount of CO2 co-
fed to the reactor to reduce readsorption of the product by
competitive adsorption. The residence time τ was evaluated from
three measurements of the forward (replacing 12CO with 13CO)
and reverse (replacing 13CO with 12CO) switches in feed
composition.
A summary of the average τ values obtained from these
measurements is shown in Figure 2. The value of τ decreased
with increasing flow rate and increasing CO2 concentration, as
expected. The relationship between τ and inverse flowrate
indicates interparticle readsorption of CO2 is important. The
simplest way to account for interparticle readsorption is to
extrapolate the value of τ to infinite flow rate. However, this
cannot not correct for intraparticle readsorption of product
within the catalyst pores. To evaluate the importance of both
interparticle and intraparticle readsorption, a similar set of
transient experiments was performed without the oxidation
reaction. Instead of switching between 12CO with 13CO during
CO oxidation, we performed a 0.2 mol % step-change in the
CO2. Although carbon monoxide was not present during the
Figure 3. Effect of flow rate on τ at yCO2 ) 1.2 mol % at 573 K over
CsX. “with reaction” indicates data obtained during CO oxidation. “without
reaction” indicates CO2 transient data in the absence of CO oxidation.
switch, all other variables were the same as those in the reactive
isotopic transient experiment. The difference between the
transient observed during reaction (obtained by switching CO
isotopes) and the transient observed without reaction (obtained
by a step-change in CO2 concentration) was attributed to the
intrinsic kinetics of the CO oxidation reaction.
Figure 3 shows the results of the switching experiments over
CsX zeolite at 573 K, with and without CO oxidation. The upper
data set (with reaction) is derived from the three τ values found
at 1.2 CO2 mol % in Figure 2. The lower data set (without
reaction) is determined from the CO2 step-change experiment.
The average difference between these two data sets (eq 2) is
considered to be the average residence time of carbon-containing
reactive intermediates, τintr
.
τintr ) τw/COrxn - τw/outCOrxn
(2)
Equation 3 describes the relationship among τintr, the number
of surface intermediates leading to reaction products, NCO2 (mol
9
J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007 3423

A R T I C L E S
Lahr et al.
Table 5. Relative Surface Coverages during Carbon Monoxide
Oxidation over Ion-Exchanged X Zeolite as Determined by Isotopic
Transient Analysis
g-1), and the overall reaction rate, RCO2 (mol g-1 s-1).
NCO ) τintr × RCO
(3)
2
2
parameter
CsX
NaX
τ
intr (s)
1.3
0.77
1.7
0.59
The pseudo-first-order rate constant, k, of the CO oxidation
reaction is simply the inverse of τintr. Table 5 compares the
residence time, rate constant and number of reactive intermedi-
ates determined from CO oxidation over CsX and NaX.
Although the global rate of CO oxidation varied by almost a
factor of 4 between the two catalysts, the intrinsic rate constant
was nearly independent of the zeolite cation. The higher rate
over CsX is attributed to a higher number of reactive intermedi-
ates in the zeolite compared to NaX.
The elemental analysis reported in Table 1 can be used to
normalize the coverage of reactive intermediates to the total
cation loading of the zeolites or to the iron impurity content in
the sample. As shown in Table 5, the number of reactive
intermediates is 4-5 orders of magnitude lower than the number
of zeolite cations in the samples. An extremely low coverage
is also realized when normalizing to only those cations present
in the supercages. Interestingly, the coverage of reactive
intermediates is between 1 and 10%, when based on the iron
impurity level in the zeolite. Moreover, the iron impurity level
as provided by Galbraith Labs increases from HX < NaX <
KX < CsX, which is the same trend with the activity for H2
and CO oxidation.
To explore the reactivity of iron in the zeolite, we sublimed
FeCl3 into HX and NaX. Iron chloride reacts with protons in
the zeolite to form HCl. Because elemental analysis of the NaX
catalyst did not show a perfect stoichiometry of Na:Al equal to
1:1, some protons are likely to be present in the sample. After
exposure to FeCl3, the NaX turned light yellow-brown in color
whereas HX turned very bright yellow, which is consistent with
the higher proton content of the latter sample. After washing
and calcining the samples, the rate of CO oxidation was
evaluated. The oxidation rate of both catalysts was only 30%
of the parent material. This decrease in activity could be
attributed to pore blockage or to partial deactivation of the active
sites by the sublimed Fe. If impurity Fe atoms were the active
species in the zeolites, they were in a form that is not readily
reproduced by simple iron addition by sublimation. Evidently,
Fe must be located in the supercages in an appropriate chemical
state to facilitate oxidation catalysis. This conclusion is sup-
ported by earlier work with Fe-loaded zeolites. For example,
Petunchi and Hall performed a detailed kinetic study of Fe-
loaded NaY zeolite containing 6.13 × 1020 Fe atoms per gram
of zeolite.29 From the data presented in that work, the rate of
CO oxidation is calculated to be approximately 0.1 µmol g-1
s-1 at 573 K, which is an order of magnitude greater than the
rate over NaY in our Table 2, but less than that over either
NaX or CsX. Moreover, the turnover frequency (TOF) of the
reaction based on the Fe content over the Fe-NaY prepared
by Petunchi and Hall was estimated to be 1 × 10-4 s-1 at 573
K, which is several orders of magnitude lower than the value
of k reported in Table 5. Because the pseudo-first-order rate
constant determined from isotopic transient analysis is consid-
ered to be an intrinsic rate at which the catalytic cycle turns
over, the low TOF reported by Petunch and Hall was likely the
result of very few of the Fe atoms participating in the reaction.
k (s-1
)
N (µmol g-1
θion
θFe
)
0.69
0.23
2.9 × 10-4
3.9 × 10-5
a
0.09
0.03
a Based on total cesium or sodium content in the zeolite.
The results of our study suggest that transition metal
impurities may play an important role in zeolite oxidation
catalysis. This is reminiscent of the findings by Parmaliana et
al., who reported that isolated Fe impurities account for the
catalytic activity and high selectivity of commercial silica
samples in the partial oxidation of methane to formaldehyde.30,31
Although there is abundant evidence that hydrocarbon oxidation
in zeolites, whether photoassisted or not, is strongly influenced
by the presence of alkali metal and alkaline earth cations, a
similar conclusion cannot be drawn unambiguously for the
oxidation of small molecules such as H2 and CO. The lower
reactivity of alkaline-earth exchanged zeolites compared to
alkali-metal exchanged zeolites together with a strong competi-
tion of CO with H2 for active sites during PROX are consistent
with a normal surface redox mechanism for the reactions.
Indeed, Aparicio and Dumesic have modeled the oxidation of
CO over a variety of transition metal-loaded zeolites according
to a simple redox mechanism.32 Therefore, a strong correlation
of oxidation activity with alkali metal or alkaline earth cations
in zeolites does not by itself justify the role of field effects in
oxidation catalysis because the level of transition metal impuri-
ties can also vary with cation type.
It should be noted that an Fe impurity level of 300 ppm in
zeolite X corresponds to about 1% of the supercages having an
associated Fe atom. Interestingly, Vasenkov and Frei used time-
resolved FT-IR spectroscopy to study the visible-light-induced
oxidation of 2,3-dimethyl-2-butene in NaY.4 Their elegant
experiments utilized a short (150 ms) photolysis pulse to
determine the number of supercages involved in the oxidation
reaction without the complication of cage-to-cage diffusion of
the reactants. They found a minimum of 1% of Y zeolite
supercages are involved in the photoreaction, which is far above
the estimated level of defects in the zeolite. It is, however, near
the transition metal impurity level of some faujasite zeolites as
shown in Table 1. Thus, although other evidence supports the
idea that zeolite countercations influence hydrocarbon oxidation
activity through either an electric field effect or a preorganization
of the transition state, there might still be an important role of
impurity transition metal ions in other oxidation systems.
Conclusions
The oxidation of H2 and CO was catalyzed by ion-exchanged
X and Y-type zeolites at temperatures ranging from 473 to 573
K. For H2 oxidation, the reactivity trend was CsX > KX >
NaX > HX. Likewise for CO oxidation, CsX was the most
active sample and HX was the least active of the Group IA
exchanged zeolites. Moreover, oxidation activities of CsX and
(30) Parmaliana, A.; Arena, F.; Frusteri, F.; Martinez-Arias, A.; Lopez Granados,
M.; Fierro, J. L. G. Appl. Catal. A: Gen. 2002, 226, 163.
(31) Arena, F.; Parmaliana, A. Acc. Chem. Res. 2003, 36, 867.
(32) Aparico, L. M.; Dumesic, J. A. J. Mol. Catal. 1989, 49, 205.
(29) Petunchi, J. O.; Hall, W. K. J. Catal. 1982, 78, 327.
9
3424 J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007

Oxidation of H2 and CO over Zeolites
A R T I C L E S
KX were greater than those of the corresponding Y zeolites,
CsY and KY. Comparison of Group IA exchanged zeolite X to
Group IIA exchanged samples, i.e., CsX vs BaX, KX vs CaX,
and NaX vs MgX, revealed higher activity of the alkali metal
forms. The PROX behavior of the Group IA-exchanged X
zeolites exhibited a high selectivity (∼80%) of CO oxidation
over H2 oxidation at 573 K. Isotopic transient analysis during
steady-state CO oxidation over CsX and NaX suggests the
higher activity of CsX results mainly from a higher concentration
of reactive intermediates in the sample. Although the coverage
of reactive intermediates was many orders of magnitude below
the cation loading in the zeolites, it was similar to the iron
impurity level in the samples. Because the overall trend in the
iron impurity content is similar to the reactivity trend for the
reactions studied here, a critical role of iron impurities in the
oxidation catalysis cannot be ruled out. It is quite possible,
however, that favorable molecular interactions of the reactants
with heavy alkali metal cations are responsible for the observed
catalysis. This study suggests that future work with oxidation
in zeolites should always address a possible role of redox
impurities in the catalytic cycle.
Acknowledgment. We acknowledge the financial support of
the Chemical Sciences, Geosciences, and Biosciences Division,
Office of Basic Energy Sciences, Office of Science, U.S.
Department of Energy (Grant No. DE-FG02-95ER14549). We
also acknowledge Chee Seng Neo for his help with some of
the data collection. Finally, we are grateful to Dr. Jason Calla
for his aid with the isotopic transient experiment.
JA0671520
9
J. AM. CHEM. SOC. VOL. 129, NO. 11, 2007 3425