Probing the low temperature initiation sites in Fe-, Mn-promoted sulfated zirconia via CO and H2 adsorption

Article abstract of DOI:10.1006/jcat.1999.2549

Exposing freshly activated Fe-, Mn-promoted sulfated zirconia to CO at ≤ 50°C induced permanent loss of activity, while adding CO after butane isomerization had a reversible effect, regardless of whether the butane flow was interrupted or not. Similar experiments using dissociated hydrogen instead of CO led to irreversible poisoning in all cases. These findings were analyzed, based on the occurrence of initiation sites that are consumed stoichiometrically and very rapidly on exposure to butane, such initiation sites which are also consumed by CO or dissociated hydrogen, CO which competes effectively for adsorption sites without affecting accumulated reaction intermediates, and such intermediates that are removed in the presence of dissociated hydrogen.

Full text of DOI:10.1006/jcat.1999.2549

Journal of Catalysis 187, 186–190 (1999)
Article ID jcat.1999.2549, available online at http://www.idealibrary.com on
Probing the Low Temperature Initiation Sites in Fe-, Mn-Promoted
Sulfated Zirconia via CO and H2 Adsorption
Abdelhamid Sayari1 and Yong Yang
Department of Chemical Engineering and CERPIC, Universite´ Laval, Ste-Foy, Que´bec, Canada G1K 7P4
Received March 31, 1999; accepted May 3, 1999
sulfate ions were considered to be the oxidizing species.
However, Fe- and Mn-promoted SZ catalysts (SFMZ) are
not only, under some conditions, two to three orders of
magnitude more active than pristine SZ, but they are ac-
tive even at room temperature. This enhanced activity was
attributed to a new redox site, believed to be an iron oxy
group. Support for such a proposal was provided by Wan
et al. (10) based on a CO poisoning experiment. It was re-
ported that the addition of CO to the feed during butane
isomerization at 30 C poisoned the catalyst completely and
irreversibly. Moreover, contacting freshly activated SZ by
CO in a helium stream gave rise to CO2, the amount of
which was used to estimate the number of low-temperature
active sitesin the catalyst. More recently, Morterra etal. (11)
conducted a similar experiment at 50 C and reported con-
flicting findings. They observed that catalyst poisoning via
the addition of CO to the feed was reversible and no CO2
was formed. Thus, they associated the low-temperature ac-
tivity to Lewis acid sites. The current work dealt with the
nature of the sites responsible for near-room-temperature
catalytic activity for butane isomerization over SFMZ cata-
lysts. Particular attention was devoted to understanding the
discrepancy between literature data (10, 11).
Exposure of freshly activated Fe-, Mn-promoted sulfated zirco-
nia (SFMZ) to carbon monoxide at temperatures up to 50 C in-
duced permanent loss of activity, while the addition of CO after
the beginning of the reaction had a reversible effect, regardless of
whether the butane flow has been interrupted or not. Similar exper-
iments using dissociated hydrogen instead of CO led to irreversible
poisoning in all cases. These finding were interpreted based on (i)
the occurrence of initiation sites that are consumed stoichiomet-
rically and very rapidly upon exposure to butane, (ii) such initi-
ation sites which are also consumed by CO or dissociated hydro-
gen, (iii) CO which competes effectively for adsorption sites without
affecting accumulated reaction intermediates, and (iv) such inter-
mediates that are removed in the presence of dissociated hydro-
c
gen.
1999 Academic Press
Key Words: n-butane isomerization; active sites; sulfated zirco-
nia; CO poisoning.
INTRODUCTION
Owing to their exceptional ability to isomerize saturated
hydrocarbons at low temperature, sulfated zirconia-based
catalysts (SZ) attracted much attention (1). Careful charac-
terization studies played an essential role in the evolution
of our understanding of such catalysts. For example, it took
a long time before it was realized that SZ catalysts are not
superacidic (2, 3), which has significant implications on re-
action mechanisms. There was also a widespread belief that
Pt in Pt/SZ is difficult to reduce by hydrogen. Our XPS,
TPR, and XRD studies demonstrated that Pt in properly
calcined samples is essentially in the metallic state (4, 5),
and no hydrogen reduction is needed.
As experimental and theoretical evidence that SZ-based
catalysts are not superacidic accumulated (2, 3, 6, 7), direct
protolysis as the initial step in the mechanism of hydrocar-
bon isomerization was abandoned (8). Attention was then
focused on the redox properties of SZ catalysts, and a one-
electron oxidation process was proposed as the initial step
for alkane activation (9). For nonpromoted SZ catalysts,
EXPERIMENTAL
The SFMZ catalyst used in the current study was pre-
pared as described elsewhere (12). The precursor Zr(OH)4
was first synthesized by dropwise addition of ammonium
hydroxide (28 wt% NH3, Anachemina) into a ZrOCl2
8H2O (98% , Aldrich) solution under vigorous stirring until
a pH of 8. Then, the precipitate was filtrated, washed by
distilled water to remove all chloride ions, then dried at
110 C for 12 h, and grinded to below 100 mesh. Sulfation
was performed by impregnation using 0.5 M sulfuric acid
solution (15 ml/g of solid). Then, 1.5% Fe and 0.5% Mn
were loaded on dried sulfated zirconium hydroxide by
the incipient wetness method using 0.3 M Fe(NO3)3 and
0.15 M Mn(NO3)2 solutions at the same time. The as-made
catalyst was dried at 110 C, calcined in static air at 620 C
for 6 h, and stored in ambient conditions.
1 Author to whom all correspondence should be addressed. Tel: 418-
656 3563, Fax: 418-656 5993, E-mail: sayari@gch.ulaval.ca.
186
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

Fe-, Mn-PROMOTED SULFATED ZIRCONIA SITES
187
Catalytic reactions were carried out in a fixed-bed down- during both the phase of increasing activity and the deac-
flow quartz microreactor with 10-mm inner diameter. Cal- tivation phase (filled symbols). It is seen that the activity
cined SFMZ (0.4 g) was loaded in the reactor between glass decreases to zero almost immediately after CO addition
wool layers. For all experiments, the catalyst was first acti- and recovers fully as soon as CO is removed. Similar re-
vated in flowing dry air (30 ml/min) at 500 C for 2 h, cooled sults were obtained at 35 C. Morterra et al. (11) introduced
to the reaction temperature (35–100 C), and then purged CO in the reaction feed during the deactivation phase and
in flowing helium (30 ml/min) for 15 min to remove air found the poisoning effect to be fully reversible. Our find-
from the reactor system. Butane isomerization was con- ings not only confirm this behavior but also show that the
ducted using a 1 : 4 mixture of n-butane and He at different poisoning is also reversible during the so-called induction
total flow rates, depending on the reaction temperature. period (Fig. 1).
Depending on the purpose of the experiment, pure carbon
In our first few attempts to run such CO-poisoning ex-
monoxide was either fed (30 ml/min for 20 min) through periments, we found that once the CO feed is switched off,
the freshly activated catalyst followed by helium purge and the activity recovers only after a long period of time, often
then an n-butane/helium mixture or fed (1 ml/min) dur- exceeding 2 h. This potentially misleading artifact was due
ing the isomerization reaction along with butane and he- to a “dead” volume in the experimental setup. A similar
lium. A third type of experiment was conducted as fol- artifact may be a possible reason CO poisoning seemed ir-
lows: the reaction mixture was stopped, CO was fed at reversible in Wan et al.’s investigation (10). Moreover, to
30 ml/min for 30 min, the catalyst was then purged with completely dismiss the possibility that the observed effect
He for 30 min, and the reaction resumed. Similar exper- of CO may be due to irreversible adsorption of CO at 30 C,
iments were carried out using H2 instead of CO. Quanti- Wan et al. purged the catalyst in flowing He at 200 C for 2 h.
tative analysis of the reaction products was performed by Such a catalyst was found to have no activity at 30 C. How-
an online gas chromatograph (HP 5890 series II) equipped ever, treatment in He at 200 C is likely to affect the cata-
with a capillary column (crosslinked methyl silicone gum, lyst surface more than merely remove potentially adsorbed
50 m 0.32 mm 0.52 mm) and a flame ionization detector. CO. The following experiment was designed to investigate
such effects. After activation, the catalyst was exposed to a
He–butane mixture at 50 C for 14 min, CO at 50 C for
10 min, and then He at 200 C for 2 h. In agreement with
RESULTS AND DISCUSSION
Wan et al. (10), when cooled to 50 C, this catalyst exhib-
ited no isomerization activity. Furthermore, even when the
step involving CO is skipped, the catalyst showed no ac-
tivity. In an early investigation from this laboratory (12), it
was reported that interruption of the n-butane flow at 35 C
followed by He purge at the same temperature did not af-
fect the reaction intermediates as the same conversion was
obtained upon resumption of n-butane isomerization. In-
dependent work by Alvarez et al. (13) showed that such in-
termediates are stable in He up to a temperature between
100 and 150 C. Therefore, He purge at 200 C after inter-
rupting the flow of butane leads to the complete removal of
reaction intermediates. As will be shown later, the so-called
low-temperature active sites are actually initiation sites that
can be used only once. Therefore, regardless of the effect
of CO, no activity can be recovered at 35 C after interrup-
tion of the butane flow and elimination of adsorbed reactive
intermediates.
Figure 1 shows the conversion rate vs time on stream for
standard n-butane isomerization over SFMZ at 50 C (open
symbols) and the effect of CO addition to the reaction feed
More importantly, the types of CO-poisoning experi-
ments reported by Wan et al. (10) and Morterra et al. (11)
have no direct bearing on catalyst sites responsible for the
isomerization activity at low temperature. Indeed, recent
mechanistic proposals show that butane isomerization over
SZ-based catalysts at low temperature is a chain reaction
with the usual three steps, initiation, propagation, and ter-
mination (8, 14). At high temperature, the initiation step
may be dehydrogenation over a Zr–O site in the sulfate
FIG. 1. Effect of CO on the catalytic activity of SFMZ at 50 C. Reac-
tion conditions: 0.4 g of precalcined catalyst (620 C, 6 h); total flow rate, 5
ml/min; (ᮀ) n-C4 :He = 1 :4; (᭿) n-C4 :He :CO = 1 :3 :1.

188
SAYARI AND YANG
cluster (8) or a one-electron oxidation process involving
sulfate species (9). In the current situation of butane iso-
merization over SFMZ at near room temperature, both of
these possibilities are unlikely since the temperature is too
low for dehydrogenation to occur and for sulfate species
to oxidize butane. Otherwise, contrary to experimental evi-
dence, nonpromoted sulfated zirconia would also be active
at room temperature. Assuming, as proposed by Wan et al.
(10), that we are dealing with an iron oxy species, the ini-
tiation step would be a noncatalytic (i.e., stoichiometric)
reaction (14) involving abstraction of active oxygen. This
reaction is likely to be very fast with all oxy species being
eliminated at once upon exposure of the catalyst to the re-
action feed. If this is the case, then contacting the catalyst
with CO (or other reactive species) before the reaction feed
within the temperature range where no other active sitesare
triggered would eliminate any isomerization activity com-
pletely. Our findings, described hereafter, fully support this
contention.
In a series of experiments, the freshly activated catalyst
FIG. 3. Effect of H2 on the catalytic activity of SFMZ at 50 C. Reac-
was contacted with pure CO (30 ml/min) for 30 min at dif- tion conditions: 0.4 g of precalcined catalyst (620 C, 6 h); total flow rate,
5 ml/min; n-C4 :He = 1 :4; (ᮀ) H2 (30 ml/min, 30 min) was added before
the beginning of the reaction followed by He purge (30 ml/min, 30 min);
(᭿) similar treatment was performed after interruption of the reaction.
ferent temperatures and purged with helium before con-
ducting butane isomerization at the same temperatures. Ex-
perimental data obtained at 35 C are shown in Fig. 2. As
seen, exposure of the catalyst to CO before reaction had a
deleterious effect on its activity. At this stage, two differ-
ent interpretations may be considered: (i) CO irreversibly
destroys the sites responsible for low-temperature catalytic
activity and (ii) CO was not removed by the helium purge
at 35 C, i.e., the active sites remained blocked by adsorbed
CO, leading to complete deactivation. The latter possibil-
ity was dismissed on the basis of the following experiment.
After CO introduction at 35 C, we switched to flowing He,
increased the temperature to 100 C, and purged with He
for 1 h before cooling to 35 C and conducting butane iso-
merization. The catalyst did not recover any activity, even
though the temperature of 100 C was not high enough for
He to remove reaction intermediates if they were present
on the catalyst surface (13). It was then clear that, regard-
less of its nature, the low-temperature initiation site is irre-
versibly eliminated by CO. Using mass spectrometry, Wan
et al. (10) detected the formation of CO2 upon exposure
of freshly activated SFMZ to CO in He at 30 C. They sug-
gested that CO2 was formed via the oxidation of CO by iron
oxy species. Our current results are consistent with such a
proposal.
Similar experimentsusingH2 instead ofCO, either before
or during the reaction, did not affect the catalyst behavior
(Fig. 3). This is consistent with our early finding that H2
has no effect on n-butane isomerization over SFMZ at at-
mospheric pressure (12, 15). It may thus be concluded that
the proposed labile oxygen part of the initiation site does
not react with molecular hydrogen. However, this oxygen
is likely to react readily with dissociated hydrogen. The fol-
lowing experiment was designed to support this proposal.
The SFMZ catalyst (0.4 g) was activated at 500 C, cooled
to 100 C in air, and exposed to flowing He. Freshly reduced
0.5 wt% Pt/Al2O3 (0.2 g) was then poured on top of SFMZ.
FIG. 2. Effect of CO on the catalytic activity of SFMZ at 35 C. Reac-
tion conditions: 0.4 g of precalcined catalyst (620 C, 6 h); total flow rate,
5 ml/min; n-C4 :He = 1 :4; (ᮀ) activated SFMZ; (᭿) CO-poisoned SFMZ. The two-component catalyst was shaken and cooled to

Fe-, Mn-PROMOTED SULFATED ZIRCONIA SITES
189
TABLE 1
Effect of Added Species before or after the Beginning
of n-Butane Isomerization at 50 C
After butane
During reactionc interruptiond
Species added Before reactionb
CO
Irreversible effect Reversible effect No effect
H2
No effect
Irreversible effect Irreversible
effect
No effect
No effect
Irreversible
effect
Dissociated
Ha
a A mixture of reduced 0.5% Pt/Al2O3 and activated SFMZ was used.
Added hydrogen was dissociated on Pt (see text).
b
The species added (30 ml/min, 30 min) was purged with He (30 ml/
min, 30 min) before switching to the 1 : 4 C4 : He mixture.
c The 1 :4 C4 : He mixture was replaced by a 1 :3 :1 C4 :He :X mixture
for about 15 min, X being CO or H2.
d
The reaction was stopped and pure CO, He, or H2 (30 ml/min) was
used for 30 min and then purged with He when necessary.
35 C, exposed to flowing H2 (30 ml/min) for 30 min, and
purged with He. Such a catalyst was found to be com-
pletely inactive for n-butane isomerization at 35 C, indi-
cating that the initiation sites were consumed by atomic
hydrogen which is formed via dissociative adsorption on
Pt/Al2O3 and spilled over the SFMZ catalyst. Addition of
hydrogen after the beginning of the reaction and inter-
ruption of n-butane also poisons the catalyst irreversibly.
Table 1 summarizes the effect of different species added
either before or after the beginning of the isomerization at
temperatures below ca. 50 C.
The effect of CO exposure, before reaction, on the cata-
lyst activity was found to be strongly dependent on the tem-
perature. For example, at 100 C, the effect was significantly
reduced (Fig. 4). The maximum activity of the catalyst con-
tacted with CO was more than 80% of its original value,
i.e., 4.9 10 6 vs 6.0 10 6 mol/(g of catalyst s 1). This indi-
cates that, at such a temperature, other types of active sites
which are not irreversibly destroyed by interaction with CO
are triggered. Additionaldata shown in Table 2indicate that
the threshold temperature for these sites to be catalytically
active is ca. 50 C. Indeed, at this temperature, the cata-
lyst that has been exposed to CO before reaction exhibited
some isomerization activity (1.5% conversion). Consistent
with Wan et al.’s proposal (10), there must be another kind
of nonacid site which is responsible for n-butane isomeriza-
tion in the temperature range of 50–100 C since, under the
same conditions, unpromoted sulfated zirconia has much
lower activity, and yet promoted and nonpromoted cata-
lysts exhibit comparable acidity. These active sites are most
likely comprised of Fe(III) oxide formed during precalci-
nation in air at high temperature.
FIG. 4. Effect of CO on the catalytic activity of SFMZ at 100 C. Re-
action conditions: 0.4 g of precalcined catalyst (620 C, 6 h); total flow rate,
50 ml/min; n-C4 :He = 10 : 40; (ᮀ) activated SFMZ; (᭿) CO-poisoned
SFMZ.
all reaction may be represented as a three-step process.
The initiation step is a noncatalytic reaction between the
reactant (R) and the initiation sites (IS), presumably Fe
oxy species. The obtained intermediate I1 transforms into a
second intermediate I2 (step 2), most likely butene, which
undergo propagation and eventually termination (step 3).
Because of the high reactivity of the iron oxy species,
step 1islikelyto be veryfast. To accumulate I2 duringthe so-
called induction period, its consumption should be slower
than its formation, i.e., r3 < r2. As suggested earlier (12),
such intermediates could be stabilized by the promoters.
TABLE 2
Catalytic Data for n-Butane Isomerization over SFMZ with or
without CO Treatment at Different Temperatures
Feed comp.
(ml/min)
Reaction
temperature
( C)
Maximum rate
(mol/(g of
catalyst s 1))
Maximum
conv. (% )
Entry
n-C4 He
7
1
2
3
4
5
6
7
8
35
35
50
50
75
75
100
100
1
1
1
1
5
5
10
10
4
4
9
6
0
17
1.5
22.5
18
20.2
15.8
1.8 10
0
7
5.0 10
2.0 10
3.3 10
2.7 10
6.0 10
4.9 10
8
9
6
20
20
40
40
6
6
As mentioned earlier, n-butane isomerization at near
room temperature over SFMZ is a surface chain reaction;
the initiation step is a fast redox process involving Fe oxy
6
Note. Reaction entries 1, 3, 5, and 7 are measured on CO-free SFMZ;
as an oxidizing species. As shown in Scheme 1, the over- entries 2, 4, 6, and 8 are measured on CO-poisoned SFMZ.

190
SAYARI AND YANG
r1
r2
r3
In summary, literature discrepancies regarding the effect
IS + R
→ I1
→ I2
→ P
step 2
step 3
step 1
ofCO on n-butane isomerization over SFMZ at lowtemper-
ature were lifted. It was found that the CO-poisoning effect
maybe irreversible or reversible, dependingon whether CO
addition took place before or after the start of n-butane
isomerization. The irreversible effect was associated with
the destruction of the initiation sites, while the reversible
effect was attributed to competitive adsorption of CO on
the catalyst surface. Experiments using other species, i.e.,
molecular and atomic hydrogen, provided strong support
to this interpretation. The strong and irreversible negative
effect of dissociated hydrogen (Table 1) was due to either
the destruction of the initiation sites or to the removal of
reaction intermediates, depending on whether it was gen-
erated before or after the beginning of the reaction.
r1
r2 > r3
SCHEME 1. n-Butane isomerization over SFMZ at near room temper-
ature. IS = initiation site, R = reactant, I1 and I2 = reaction intermediates,
P = final product, ri = overall reaction rate of step i.
The effect of CO addition at different stages of n-
butane isomerization can be rationalized on the basis of
Scheme 1. Exposure of the catalyst to CO before the in-
troduction of n-butane destroys the initiation sites; thus,
the catalyst will not exhibit any activity. Complete and ir-
reversible inhibition was also observed upon exposure of
the catalyst to spilled-over atomic hydrogen under other-
wise similar conditions. Addition of CO during either the
induction or the deactivation period essentially puts the re-
action on hold because, at low temperature, CO competes
efficiently for adsorption sites that are required for steps 2
and 3. Such sites may be acidic in nature (8, 11). Earlier
work from this laboratory (16) showed that the compet-
itivity of CO for adsorption sites during isomerization of
hydrocarbons over SZ catalysts increases as the reaction
temperature decreases. In addition, our findings (Fig. 1)
are consistent with the contention that the surface concen-
trations of I1 and I2 are hardly affected by flowing CO at
near room temperature, even if the n-butane flow is in-
terrupted. Similar behavior was observed in the presence
of flowing He or H2. Once CO is removed, the reaction
resumes unperturbed. However, it is important to men-
tion that the intermediates are removed in the presence
of dissociated hydrogen, leading to irreversible loss of acti-
vity.
REFERENCES
1. Song, X., and Sayari, A., Catal. Rev.-Sci. Eng. 38, 329 (1996).
2. Kustov, V., Kazansky, V., Figueras, F., and Tichit, D. J., J. Catal. 150,
143 (1994).
3. Adeeva, V., de Haan, J. W., Ja¨nchen, J., Lei, G. D., Schu¨nemann, V.,
van de Ven, L. J. M., Sachtler, W. M. H., and van Santen, R. A., J. Catal.
151, 364 (1995).
4. Sayari, A., and Dicko, A., J. Catal. 145, 561 (1994).
5. Dicko, A., Song, X., Adnot, A., and Sayari, A., J. Catal. 150, 254 (1994).
6. Haase, F., and Sauer, J., J. Am. Chem. Soc. 120, 13503 (1998).
7. Adeeva, V., Lei, G. D., and Sachtler, W. M. H., Catal. Lett. 33, 135
(1995).
8. Hong, Z., Fogash, K. B., Watwe, R. M., Kim, B., Masqueda-Jime´nez,
B. I., Natal-Santiago, M. A., Hill, J. M., and Dumesic, J. A., J. Catal.
178, 489 (1998).
9. Ghenciu, A., and Faˇrcas¸iu, D., J. Mol. Catal. 158, 311 (1996).
10. Wan, K. T., Khouw, C. B., and Davis, M. E., J. Catal. 158, 311 (1996).
11. Morterra, C., Cerrato, G., Di Ciero, S., Signoretto, M., Minesso, A.,
Pinna, F., and Strukul, G., Catal. Lett. 49, 25 (1997).
12. Sayari, A., Yang, Y., and Song, X., J. Catal. 167, 346 (1997).
13. Alvarez, W. E., Liu, H., and Resasco, D. E., Appl. Catal. A 162, 103
(1997).
14. Ta´bora, J. E., and Davis, R. J., J. Catal. 162, 125 (1996).
15. Song, X., Reddy, K. R., and Sayari, A., J. Catal. 161, 206 (1996).
16. Song, X., and Sayari, A., Energy Fuels 10, 561 (1996).
The effect of CO exposure at 100 C before the reaction
was not as dramatic as that observed at 35 or 50 C because
another active site (as opposed to initiation site), presum-
ably Fe (III), was triggered. Such an active site was not
affected by CO.