Evolution of the Catalytic Activity in Pt/Sulfated Zirconia Catalysts: Structure, Composition, and Catalytic Properties of the Catalyst Precursor and the Calcined Catalyst

Article abstract of DOI:10.1006/jcat.1998.2104

A 3% Pt/sulfated zirconia catalyst was prepared and characterized before and after calcination at 900 K by XRD, XPS, EM, and in the catalytic hydroisomerization of n-hexane. The fresh sample exhibited small but definite catalytic properties. Calcination brought about a dramatic increase of the activity with practically constant high (90-100%) selectivity for hydroisomerization versus cracking. This increased activity was accompanied by the transformation of the predominantly amorphous support to predominantly tetragonal crystals and the wrapping up of most parts of surface Pt atoms into the bulk, as shown by the physical characterization methods. Hence metallic Pt particles exhibited mainly Pt-O rather than Pt-S interactions. S was present as sulfate. Pt-sulfated zirconia was different from traditional bifunctional metal catalysts on acidic supports. We attributed its higher catalytic activity and favorable isomerization selectivity to a few but very active centers, formed by interaction of Pt sites with sulfate groups on the high Miller-index surfaces of ZrO₂. Calcination must be essential to create these active sites. H₂ dissociating on Pt sites would provide the hydride species that are necessary for isomerization occurring on the acidic (sulfate-zirconia) part of that ensemble. We proposed the name compressed bifunctional sites for these centers of acid-metal cooperation. The assumption of such active sites, the maximum activity as a function of the hydrogen pressure, can also be explained in a consistent way.

Full text of DOI:10.1006/jcat.1998.2104

JOURNAL OF CATALYSIS 178, 338–351 (1998)
ARTICLE NO. CA982104
Evolution of the Catalytic Activity in Pt/Sulfated Zirconia Catalysts:
Structure, Composition, and Catalytic Properties of the Catalyst
Precursor and the Calcined Catalyst
,1
ꢀ
ꢀ
Jean-Marie Manoli, Claude Potvin, Martin Muhler,† Ute Wild,† Ga´bor Resofszki,‡
Thomas Buchholz,§,2 and Zolta´n Paa´l§,3
ꢀ
Laboratoire de Reactivite´ de Surface, URA 1106, casier 178, Universite´ P. & M. Curie, F-75252 Paris Cedex 05, France; †Fritz-Haber-Institut
der Max-Planck-Gesellschaft, Berlin, Faradayweg 4-6, D-14195 Germany; ‡Hungarian Oil and Gas Co., Sza´zhalombatta, P.O. Box 1,
H-2443 Hungary; and §Institute of Isotopes of the Hungarian Academy of Sciences, Budapest, P.O. Box 77, H-1525 Hungary
Received November 18, 1997; revised April 8, 1998; accepted April 8, 1998
peracidic properties (104 times stronger that 100% H2SO4),
the catalytic activity was no longer proportional to the acid
strength above a certain level of acidity (6). Hence su-
peracidic properties may not be directly related with iso-
merization activity. Sulfated zirconia catalysts deactivate
rapidly (7, 8). Adding platinum group metals (8–15) or
A 3% Pt/sulfated zirconia catalyst was prepared and character-
ized before and after calcination at 900 K by XRD, XPS, EM, and
in the catalytic hydroisomerization of n-hexane. The “fresh” sam-
ple exhibited small but definite catalytic properties. Calcination
brought about a dramatic increase of the activity with practically
constant high (90–100%) selectivity for hydroisomerization versus
cracking. This increased activity was accompanied by the transfor-
mation of the predominantly amorphous support to predominantly
tetragonal crystals and the wrapping up of most parts of surface
Pt atoms into the bulk, as shown by the physical characterization
–
Fe Mn (16) improved their long-term stability. The pres-
ence of hydrogen was indispensable for a sustainable iso-
merization (12, 13). A maximum curve was observed as a
function of hydrogen pressure for butane isomerization, in
a rather wide range of H2/butane ratio (12), analogous to
the H2 pressure dependence of other Pt based catalysts (17,
18). Acidity itself cannot explain the striking changes in the
catalytic properties brought about by the platinum content,
for example, temperature-programmed desorption (TPD)
curves of ammonia exhibited insignificant differences in the
presence of Pt.
There are counterclaims in the literature concerning
(i) the state of Pt (Pt metal, Pt sulfide, bulk or surface;
Pt oxide, bulk or surface), (ii) its possible effect in gen-
erating stable and very active isomerization sites by, e.g.,
hydrogen spillover from Pt generating protonic sites (19),
(iii) the nature of active sites in the presence of Pt (tra-
ditional bifunctional, superacidic), and (iv) the reason of
deactivation (coking, loss of acidity, loss of surface Pt, etc.).
Pt was assumed to be present mostly as Pt sulfide after
reduction (20), H2 at reduction temperature reducing sup-
posedly some of the sulfate groups. Earlier we questioned
that XPS results alone could adequately prove this state-
ment (21). Later studies (4, 14, 21–26) pointed rather to a
metallic Pt core surrounded by Pt oxide and/or sulfide as
surface compounds or just adlayers.
–
methods. Hence metallic Pt particles exhibited mainly Pt O rather
–
than Pt S interactions. S was present as sulfate. Pt-sulfated zirconia
was different from traditional bifunctional metal catalysts on acidic
supports. We attributed its higher catalytic activity and favorable
isomerization selectivity to a few but very active centers, formed by
interaction of Pt sites with sulfate groups on the high Miller-index
surfaces of ZrO2. Calcination must be essential to create these active
sites. H2 dissociating on Pt sites would provide the hydride species
that are necessary for isomerization occurring on the acidic (sulfate-
zirconia) part of that ensemble. We proposed the name “compressed
bifunctional sites” for these centers of acid–metal cooperation. The
assumption of such active sites, the maximum activity as a function
of the hydrogen pressure, can also be explained in a consistent way.
c
ꢁ 1998 Academic Press
INTRODUCTION
Sulfated zirconia based catalysts are very active and se-
lective in hydroisomerization of alkanes (1). The relatively
high number of papers appeared were summarized recently
(2–4). Although sulfated zirconia reportedly (5) exhibits su-
To our best knowledge, no papers have been published
monitoring the generation of the active catalyst during cal-
cination. Following our preliminary report (27), the present
paper will tackle the changes of the state of Pt in the pro-
cess of activating Pt/sulfated zirconia by calcination in air
1 Present address: Ruhr-Universita¨t Bochum, Lehrstuhl fu¨r technische
Chemie, Bochum, D 44780 Germany.
2 On leave from Fachhochschule Ostfriesland, Constantiaplatz 4, Em-
den, D-26723 Germany.
3 Corresponding author.
338
0021-9517/98 $25.00
c
Copyright ꢁ 1998 by Academic Press
All rights of reproduction in any form reserved.

Pt/SULFATED ZIRCONIA CATALYSTS
339
at 900 K and intends to clarify its role in alkane hydroi- intensities were fitted by a pseudo-Voight function, assum-
somerization reactions. To this end, X-ray photoelectron ing a linear background. The standard error was ꢃꢄ5% .
spectroscopy (XPS), ion scattering spectroscopy (ISS), as
Electron Microscopy
well as X-ray diffraction (XRD) and electron microscopy
(EM) have been applied, along with monitoring the cata-
lytic properties of the dried impregnated catalyst precursor
as well as that of the calcined catalysts. Further papers will
tackle problems like the activity of sulfated zirconia with-
out Pt, the initial selectivity changes, and possible reasons
of temporary and final deactivation of Pt-sulfated zirconia,
The samples were ground, dispersed in ethanol in an ul-
trasonicbath, and deposited on a thin carbon film supported
on a standard copper grid. HRTEM studies were performed
on a JEOL apparatus (JEM 100 CXII) equipped with a top
entry device and operating at 100 kV. The camera length
was calibrated with polycrystalline gold powder.
–
as well as their likely Pt S interactions.
Catalytic Runs
EXPERIMENTAL
The activity tests were carried out in a laboratory closed-
loop glass circulation reactor (28–30), at an n-hexane pres-
sure of 10 Torr with the H2 pressure varying between 60 and
480 Torr. The sampling time was between 5 and 240 min.
Regeneration with circulating gases: oxygen (followed by
evacuation) and hydrogen took place between runs. A fused
quartz column coated with CP Sil-5 was used for GLC anal-
ysis, ensuring separation between methane and benzene
(28–30). Preliminary catalytic tests were carried out after
preparation in a single-pass reactor at atmospheric pres-
sure with 10 Torr n-hexane in N2.
Surface Analysis
A Leybold LHS 12 MCD instrument, equipped with
X-ray source (Mg Kꢀ or Al Kꢀ anode) and ion gun for
ISS was used for surface analysis (28–30). Dry powdered
samples placed on a stainless steel sample holder or sus-
pensions in methanol or n-hexane dropped to the sam-
ple holder and dried were equally suited. Samples could
also be subjected to in-situ reduction at near-atmospheric
pressure in the preparation chamber at 380ꢂC (653 K) and
the letter H is added to their code. The samples showed a
BE shift of 6–7 eV due to electrostatic charging. We used
the Si 2p peak of the minor SiO2 content for BE calibra-
tion, BE = 103.3 eV (31, 32). TUBUS experiments (32) with
45-V biaswere also carried out with one ofthe samplescaus-
ing a BE shift of ca 20–25 eV. After energy calibration, the
peak BE values agreed within a few tenths electron volts
with those measured without TUBUS, indicating that no
inhomogeneous charge occurred. For samples taken out of
the catalytic reactor, with a possible glass wool contamina-
tion (in which the Si peak may have carried a different elec-
trostaticcharge) the bindingenergyvalueswere normalized
for Zr in ZrO2, BE = 182.7 eV (34). The SciPlot software
RESULTS
Catalyst Preparation
ZrOCl2 ꢅ 8H2O was hydrolyzed by NH4OH to get zirco-
nium hydroxide. Gmelin’s Handbook of Inorganic Chem-
istry regards this as a hydrated zirconium oxide rather than
a stoichiometric compound. It was necessary to add more
than 2 moles of ammonia (equivalent to the Cl content
of the parent compound) to reach a basic final pH after
precipitation resulting in active catalysts (Table 1). Adding
ꢃ3 moles NH4OH per mole of ZrOCl2 was necessary to
c
(ꢁ M. Wesemann, Fritz-Haber-Institut) was used for spec-
TABLE 1
trum evaluation: satellite background subtraction, Shirley
background subtraction, and integration (28–30), using lit-
erature values of atomic sensitivity factors for quantita-
tive determination of surface composition (32). He+ ions of
2 kV were used for ISS under quasi-static conditions (28).
Ammonia Excess, Final pH at Precipitation of Zr Hydroxide,
and the Activity of Catalysts after Calcinationa
NH4OH excess
Final pH
Relative activityb
1.2
1.9
2.2
2.4
2.5
3
2.0
4.0
6.0
7.0
8.0
—
—
0.45
—
0.85
0.90
0.77
X-Ray Diffraction
A Siemens D500 automatic diffractometer with a Cu Kꢀ
monochromatized radiation source was used for determi-
nation of the X-ray diffraction (XRD) patterns of various
solid phases. The crystallite dimensions were determined
using the Scherrer equation: D(hkl) = 0.9 ꢁ/B cos ꢂ, where
D(hkl) is the crystal thickness in the hkl direction, ꢁ is the
wavelength of the X-ray diffraction, B is the peak width
at half-maximum corrected for instrumental broadening,
9.0
10.0
6
a Selected values from 60 parallel experiment. At least 3 and at most 8
parallel points have been measured for each line.
b Characterized by the overall conversion measured after 50 min time-
on-stream in a flow reactor at 423 K, with 10 Torr n-hexane in N2 carrier
and ꢂ is the Bragg angle of the diffraction peak. The peak gas over samples subjected to calcination at 900 K.

340
MANOLI ET AL.
reach a pH of 8–9, which is optimum for the subsequent program: the temperature was raised up to 573 K within 1 h,
catalytic activity of sulfated zirconia. A sixfold excess of kept at 573 K for 1 h, the subsequent heating up to 900 K
ammonia was necessary to reach the value of pH = 10 re- lasted for 5 h, followed by an isothermal period of 3 h.
ported by Yori et al. (7). This latter sample was less active Finally, the sample was left alone to cool. The absorption
in our case (Table 1).
and titration of SO3 leaving the sample during calcination
The support was then treated with H2SO4. An optimum showed the removal of 23% of the original SO3 content. The
SO3 content was reached when the concentration of H2SO4 weight loss indicated a water removal of 12 w/w% of the
was 0.5N. This 0.5N sulfuric acid was applied in a fivefold ex- original Zr oxide-hydrate; about 2/3 of the total water loss
cess related to Zr oxide hydrate. The sample was then dried occurred below 523 K. The drying temperature influenced
at 373 K. At this drying temperature, the H2SO4 uptake was the activity of the subsequently calcined sample: that dried
10 mol% accompanied by a loss of 3% Zr, as opposed to at 373 K had a higher activity by a factor of 1.2–1.5 than
an H2SO4 uptake of 17% after a drying at 323 K, which those dried at lower temperatures; drying at 403 K gave an
gave much higher Zr loss (10% ). Drying at 403 K, in turn, inactive sample after calcination.
resulted in an H2SO4 uptake as low as 2% .
Pt was introduced by suspending the dried sulfated zir-
conia with a 0.15N PtCl4 solution (25). It was then reduced
Electron Microscopy
in H2 at 373 K (code: “fresh”). The H2O/ZrO2 ratio in this
Electron microscopy showed some well-characterized
sample was 1.15 as opposed to the value of 1.4 after dry- Pt particles in the fresh state, together with smaller Pt
ing at 323 K and the value of 0.65 after drying at 403 K. It ones. Figure 1 shows a larger particle whose lattice fringes
contained about 3% Pt to facilitate detection of Pt by XPS. correspond to those of the Pt(111) plane (the magnification
To obtain an active catalyst, the dried and reduced sample applied in print cannot reproduce these). Pt particles show
was calcined in air at 900 K (code: “calc.”) according to a a somewhat weaker contrast after calcination (Fig. 2) but
FIG. 1. Electron micrograph of a “fresh” Pt/sulfated zirconia catalyst sample. Note the occurrence of lattice fringes on the large aggregate of Pt
particles. The interfringe distances correspond to that of Pt(111).

Pt/SULFATED ZIRCONIA CATALYSTS
341
FIG. 2. Restructuring of a “calc.” Pt/sulfated zirconia catalyst under the electron beam: (a) first electron micrograph after placement under the
beam; (b) last electron micrograph after having kept the sample under the beam for several minutes. Arrows denote crystallites where reconstruction,
disappearance (a), reappearance (b) of Pt particles were observed.

342
MANOLI ET AL.
FIG. 3. Electron micrographs of areas showing mainly support. (a) support with finer crystal structure, exhibiting excellent diffraction rings (inset)
corresponding to tetragonal ZrO2, in particular, a d-spacing of 0.209 nm corresponding to the (112) plane missing in the cubic form; (b) area with a
large amorphous support grain exhibiting no diffraction pattern at all.

Pt/SULFATED ZIRCONIA CATALYSTS
343
their size—as appeared in the electron microscope—was not
TABLE 2
very different from those observed in the fresh sample. The
particles were too small to be identified by lattice fringes.
Some dark spots that may correspond to Pt particles un-
dergo changes during exposure to the electron beam: some
disappear, some gain a sharper contrast, and some appear
Surface Composition of Pt/Sulfated ZrO2 Catalyst Samplesa
Composition, %
Sample
O
Zr
S
C
Si
Pt
anew, as illustrated by the comparison of the same area Freshb
57.4
59
55.4
57
68.4
65.5
18.2
21.7
17
19.1
22.1
23.1
3.3
2.7
3.3
3
4.5
3.5
12.3
7.2
13
9.6
1.4
4.2
5.4
6.0
9.3e
9e
3.1
3.3
2.5
3.3
2
2.3
0.42
0.4
Fresh/Hc
depicted in Figs. 2a and 2b.
Fresh/nHd
Fresh/nH/Hc,d
“Calc.”
Smaller ZrO2 particles were seen, together with larger
ones in the “calc.” catalyst (Fig. 3). The small particles
showed sharp electron diffraction rings, corresponding to
well-crystallized tetragonal ZrO2, whereas some amor-
phous larger particles were also seen.
“Calc.”/Hc
a As determined from the areas of the O 1s, Zr 3d, S 2p, C 1s, Si 2p, and
Pt 4f lines, using sensitivity factors taken from the literature (32).
b 0.9% Cl impurity.
X-Ray Diffraction
c H means an in-situ hydrogen treatment at 653 K.
d
The X-ray diffraction pattern (Fig. 4) of the “fresh” cata-
nH means a catalyst used for several subsequent catalytic runs in
lyst shows well-defined narrow lines of Pt, together with n-hexane–hydrogen mixtures (summarized length of runs up to 8 h).
e Glass wool contamination possible.
broad and indistinct lines for zirconia, indicating its low
crystallinity. After calcination, intense zirconia lines cor-
responding to tetragonal zirconia (JPCD file No. 17-0923)
XPS of “Fresh” and “Calcined” Catalysts
appear. No lines corresponding to monoclinic zirconia were
Composition. The composition of both samples was
seen and the presence of cubic zirconia could also be ex-
similar—except for the values for Pt (Table 2). Calcina-
cluded. The most intense ZrO2 line in the spectrum of the
tion removed a large fraction of carbon impurity, which ad-
sorbed, likely, from pollution present in surrounding air. An
in-situ hydrogen treatment of both the “fresh” and “calc”
samples removed a small fraction of carbon, increased
“calc” sample (line b, Fig. 4) at 2ꢂ–30ꢂ is clearly seen to be
superimposed to a broad, poorly defined amorphous back-
ground (line a) and this confirms the coexistence of amor-
phous and crystalline zirconia. The Pt lines after calcination
somewhat the Pt concentration; otherwise it did not influ-
were broader and less intense. The fitting of the Pt line at
ence markedly the overall composition. After having used
the “fresh” sample in catalytic reactions, its carbon content
increased (Table 2), leaving the other components almost
2ꢂ–40ꢂ gave a peak area about twice as high with the “fresh”
sample than with the “calc.” one. Thus, the redispersion of
Pt was more clearly evidenced from XRD than observed
unchanged. Hydrogen treatment removed some of this car-
by electron microscopy.
bon, too. The small Cl impurity of the “fresh” sample orig-
inating from impregnation with hydrogen chloroplatinate
disappeared even after in-situ hydrogen treatment at 650 K.
The Pt content of “fresh” catalyst corresponded roughly
to its nominal composition of ca 3% and this remained so
after n-hexane reactions, too. One of the most dramatic
changes brought about by calcination in air at 900 K was
the drop of its surface concentration by nearly an order
of magnitude. The structural rearrangements accompany-
ing this composition change must have been necessary for
creation of active sites.
Oxygen. The O 1s peak showed in all cases two com-
ponents (Fig. 5). One at ca 530.5 eV should correspond to
ZrO2, along with a shoulder at ca 532.5 eV. This latter can
be attributed to a possible zirconium hydroxide species (34)
and/or to oxygen in sulfate (31, 35). Our O 1s spectra had
a similar shape as reported by Milburn et al. (35), although
their different BE scale makes quantitative comparison dif-
ficult. This higher BE component was ca equal to the ZrO2
peak in the “fresh” catalyst. Any heat treatment (even hy-
drogenation at 653 K) decreased its intensity (Fig. 5, dotted
spectra). The higher BE peak was restored to some extent
after n-hexane reactions (Fig. 5, dashed spectrum).
FIG. 4. X-ray diffraction pattern of Pt/sulfated zirconia catalysts in
the “fresh” (a) and “calc.” (b) states. The recrystallization after calcination
(curve b) is clearly shown by the appearance of lines for tetragonal ZrO2
(JPCD 17-0923) depicted as vertical lines with lengths proportional to the
corresponding intensities. The diffraction lines of Pt metal (JPCD 4-0802),
denoted by asterisks are also clearly distinguishable in both cases, but they
are more marked with the “fresh” catalyst.

344
MANOLI ET AL.
of Pt intensity upon calcination was concomitant to a very
slight BE shift. The BE was shifted to ca 71.3–71.6 eV after
in-situ hydrogen treatment as well as after n-hexane reac-
tions. This shift was somewhat more pronounced with the
calcined catalyst. This indicates that Pt can be reduced by
hydrogen at 650 K and its state approached that of a clean
metallic Pt (BE = 71.1 eV in the same apparatus (29, 30)).
Carbon. The main carbon peak was at ca BE 284.8 eV,
corresponding to hydrocarbonaceous and/or graphitic de-
posits (32). In the “fresh” sample, also a smaller peak ap-
peared at ca 289 eV indicating heavily oxidized residues
(COOH groups). This disappeared upon hydrogen treat-
ment. Calcination reduced the carbon content drastically
(Table 2). More carbon was present after hydrogen treat-
ment of the “calc.” catalyst, pointing to a possible reap-
pearance of carbonaceous entities encapsulated in deeper
layers during its restructuring. Carbon aggregation to three-
dimensional islands could also take place under hydrogen
effect (38). These contiguous carbonaceous islands were
found earlier to give rise to inhomogeneous electrostatic
charge (30) and could have been the reason of a BE shift of
ca 0.5 eV as compared to all components but Si (corrected
in the figures).
FIG. 5. O 1s region of “fresh” and “calc.” catalysts in the “as received”
state and after in-situ hydrogen treatments at 650 K (denoted by “ +H2”).
as well as after n-hexane reactions on the “fresh” catalysts. Inset: Zr 3d
region of “fresh” and “calc.” catalysts.
Zirconium. Literature reported various BE(max) val-
ues for the Zr 3d line in ZrO2: 182.8 eV (34); 182.9 eV (36);
or 182.75 eV (37). Our values of BE = 182.7 eV measured
with “fresh” and “calc.” catalysts were closer to those val-
ues after energy calibration (Fig. 5, inset) than the value
of BE = 182.3 eV published by other sources (20, 32). The
shape of the doublet in the catalyst samples was similar after
each treatment.
Sulfur. The state of sulfur (Fig. 6) corresponded in both
“fresh” and “calc.” catalysts to sulfate, BE = 169.5 eV (32).
No changes appeared after n-hexane reactions. The tailing
at BE ꢃ 165 eV may have indicated a minor reduction to
S4+. Thus, the redox reaction of S assumed during hydro-
carbon reactions may have been stopped at this valence
state. In-situ hydrogen reduction of the samples had given
rise to some sulfide close to the detection limit, not sepa-
rated reliably from spectral noise (Fig. 6, dotted spectra).
Platinum. The higher Pt content of “fresh” samples
gave well-shaped Pt 4f doublet, similar to what would be
expected from a supported catalyst of not too high disper-
sion (Fig. 7). n-Hexane reactions did not change the inten-
sity or the BE position of Pt 4f. The BE of Pt 4f 7/2 in the
“as received” state was around 71.8–72 eV, i.e., higher than
FIG. 6. S 2p region of “fresh” and “calc.” samples measured in “as
received” states and after in-situ hydrogen treatment (denoted by “ +H2”).
Spectra taken after n-hexane reactions on the “fresh” catalysts were very
that corresponding to metallic Pt (32). The marked drop similar to the “fresh” spectra; hence they are not shown.

Pt/SULFATED ZIRCONIA CATALYSTS
345
FIG. 8. Ion scattering spectra (ISS) of the fresh and calcined catalysts.
He + ions of 2 keV were used. Cumulated results obtained in 30 subsequent
scans are shown. The abrasive effect of subsequent scans wore off some of
the sample and signals from the underlying sample holder became visible.
540 K upwards in “skeletal” reactions of alkanes: isomer-
ization, C5-cyclization, hydrogenolysis, aromatization. The
“fresh” Pt/sulfated zirconia catalyst showed some activity
already at ca 480 K. The “calc.” sample, in turn, produced
isomers already at 420 K. A much smaller mass of “calc.”
catalyst gave higher conversions than a higher mass of the
“fresh” one (Table 3). Figure 9 shows this in terms of spe-
cific activity per unit mass of Pt. No signs of irreversible
deactivation was observed even after runs up to 50 min.
Sulfated zirconia without Pt showed, in turn, a short-living
catalytic activity.
FIG. 7. Pt 4f region of “fresh,” “fresh” + n-hexane, and “calc.” cata-
lysts, in the “as received” state and after in-situ hydrogen treatments at
650 K (denoted by “ +H2”). The intensity of the sample after n-hexane
reactions has been measured another occasion as the other ones, and was
normalized to the Pt intensity of the “fresh” catalyst by a factor of 0.67.
The selectivities as a function of the temperature
(Table 3) show a similar abundance of skeletal isomers on
both the “fresh” and the “calc.” sample up to ca 573 K. Be-
low 540 K, skeletal isomers (mostly methylpentanes, along
with minor amounts of dimethylbutanes) were the only
products. Traces of fragments and olefins also appeared at
Silicon was regarded as an impurity and has, very likely,
nothing to do with the other catalyst components.
Ion Scattering Spectroscopy (ISS)
The spectra of both “fresh” and “calc.” samples gave
an independent experimental evidence of their very differ-
ent surface Pt concentrations (Fig. 8). The “fresh” catalyst
showed a small but clearly discernible Pt signal, whereas
that from the “calc.” catalyst was hardly separated from
the background. The spectra shown were obtained after
several scans accumulated. The sputtering effect of the
He+ ions as well as the repulsion between insulating pow-
der grains charged electrostatically under a constant im-
pact of positive ions (which could be only partly compen-
sated by the use of electron flood gun) thinned the sample
layer so that signals from the underlying steel holder plate
appeared.
Catalytic Results
Comparison of “fresh” and “calc.” catalysts at vari-
ous temperatures. Various supported (39–41) and unsup-
ported (28, 29) Pt catalysts started to be active from 520–
FIG. 9. Activity (mol n-hexane transformed per mass unit Pt per unit
time) of “fresh” and “calc.” Pt/sulfated zirconia catalysts as a function of
the temperature: p(nH) = 10 Torr, p(H2) = 120 Torr, sampling time, 5 min.

346
MANOLI ET AL.
TABLE 3
Selectivity of n-Hexane Reactions on “Fresh” and “Calc.” Pt/Sulfated ZrO2 Catalystsa Together with Results Obtained
on Other Pt Catalysts on Supports of Various Acidity
Selectivity, % b
Sample
T, K
Conv., %
<C6
DMB
MP
MCP
Hex.
Bz
“Fresh”
483
513
543
573
603
0.07
1.3
2.2
3.6
2.6
0
0
0
4.4
19.0
0
0
9
11.9
10.4
100
100
90.2
92.2
67.3
0
0
0.8
0.4
0.6
0
0
0
0.6
2.0
0
0
0
0.5
0.7
“Calc.”
423
453
483
513
543
573
603
0.08
0.3
1.5
0
0
0
6.7
9.4
10.6
11.7
11.9
10.9
100
0
0
0
0
0
0.3
—
0.5
2.6
0
0
0
0
0.4
0.4
0.9
93.3
89.8
87.7
85.6
82.5
74.8
0.4
1.0
1.8
4.2
10.1
0.4
0.4
0.5
0.5
0.6
4.9
10.6
12.3
12.9
c
Pt/SiO2
543
603
6.7
12.8
7
13.5
ꢃ1
37
23
49
42.5
—
9
6
12
Pt/NaHYd
Pt/HYd
603
603
4.2
5.1
5
17
3
6
48
47
9
1
20
29
15
0
a 135 mg “fresh” catalyst and 29 mg “calc” catalyst.
b
<C6: fragments; DMB: dimethylbutanes; MP: methylpentanes; MCP: methylcyclopentane; Hex.: n-hexenes; Bz: benzene. Calculated for C6 units
transformed.
c 6% Pt on neutral SiO2 support; Ref. (42).
d 0.8% Pt on highly acidic (Pt/HY) or slightly acidic (Pt/NaHY) support; Ref. (40).
higher temperatures. Their selectivity increased abruptly with respect to the complementary C2 and C1 fragment
above 570 K on the “fresh” sample, whereas their amount was not too marked even at the highest temperature.
was much lower over the “calc.” catalyst. Traces of ben-
zene and methylcyclopentane (MCP) appeared at highest
temperatures. The formation of methylpentanes was fa-
vored far above thermodynamical equilibrium (Table 4).
Catalyst stability. The XPS regions of “fresh” catalysts
showed only minor structural changes after n-hexane reac-
tions, due to their heating, to its interaction with n-hexane,
or with regenerating O2 and/or H2 gases. This could be seen
The ratio 2-methylpentane/3-methylpentane (2MP/3MP)
in the O 1s and Pt 4f spectra (Figs. 5 and 8), the other peaks
was around 1.50, almost independently of the temperature.
They were in good agreement with values obtained over Pt
on slightly acidic zeolites (40, 41) but were lower than those
found with Pt on neutral supports (39, 40, 42).
TABLE 4
Isomer Ratios as a Function of Temperature
and Hydrogen Pressure
Fragment distribution can give information on the
possible pathway of their formation. Hydrogenolysis on
Pt sites gave mostly propane with less C5 + C1 fragments
and even less C2 + C4 products (39, 42). Acid catalyzed
cracking, in turn, gave no C1 and C2 fragments at all
and propane prevailed (43). The fragmentation patterns
obtained over Pt on zeolites of various acidity indicated
that the predominance of the metal or the acid function
depended strongly on the conditions, e.g., on the hydrogen
excess (40, 41). Individual fragment concentrations are
shown in Table 5. The “fresh” catalyst gave no methane and
ethane at lower temperatures. Enhanced fragmentation
at higher temperatures was due to the increase of C3–C5
products, including isobutane and isopentane. The “calc.”
sample produced small amounts of C1–C2 fragments. The
excess of C4 and C5 (containing also branched fragments)
DMB/MP
2MP/3MP
Catalyst
“Fresh”
p(H2), Torr
543 K
603 K
543 K
603 K
60
120
240
360
480
0.14
0.13
0.12
0.09
0.09
0.16
0.15
0.15
0.15
0.14
1.65
1.57
1.55
1.49
1.49
1.69
1.59
1.53
1.50
1.50
“Calc.”
60
120
240
360
480
0.14
0.13
0.12
0.11
0.10
0.15
0.15
0.14
0.14
0.13
1.57
1.52
1.49
1.46
1.45
1.60
1.54
1.50
1.49
1.48
Equilibrium
0.800a
0.635
2.02a
1.95
a Interpolated value.

Pt/SULFATED ZIRCONIA CATALYSTS
347
FIG. 10. Yields of individual products as a function of the hydrogen pressure: (a) “fresh” catalyst, 543 K; (b) “calc.” catalyst (543 K); (c) “fresh”
catalyst, 603 K; (d) “calc.” catalysts, 603 K. p(nH) = 10 Torr, sampling time, 5 min., T = 603 K.
(Zr, S) remained almost unchanged. Some carbon accu- samples show shallow maxima in yields, hexene being no-
mulation occurred (Table 2). The “calc.” sample showed ticeable at the lowest hydrogen pressure only. The higher
a remarkable stability of activity and selectivity but the activity of the “calc.” sample is obvious in the whole range.
fresh sample exhibited a “running-in” behavior, the overall Traces of MCP and (for the “calc.” sample, of benzene)
conversion increasing from Run 1 to Run 5 by a factor of accompanied the main products. At 603 K (Figs. 10c and
2.5 with the isomerization selectivity increasing from 83 to d), low hydrogen pressures favor fragmentation and de-
92% .
hydrogenation to olefins in both cases. The effect is much
more marked with the “fresh” catalyst. Isomerization pre-
vailed at higher hydrogen pressures. The isomer production
showed maxima in all four cases as a function of p(H2).
The sharp increase of yields at 603 K on the “fresh” sam-
ple above p(H2) = 360 Torr is worth mentioning. The shift
of the maxima towards higher hydrogen pressures with in-
creasing temperatures appeared, here, too, as with other
Pt catalysts (17, 18, 39). Increasing hydrogen pressure de-
Hydrogen pressure effects. The yields of skeletal hydro-
carbon reactions showed, as a rule, maxima as a function
of the hydrogen pressure (17, 18, 39, 40). The same was
reported for butane isomerization of Pt/sulfated zirconia
(12). Figure 10 illustrates characteristic differences between
the yields obtained over “fresh” and “calc.” samples. At
543 K (Figs. 10a and b), with mostly isomerization, both

348
MANOLI ET AL.
creased the DMB/MP and the 2MP/3MP ratios on both must have been the major component and the sulfate the
catalysts (Table 4). The fragment distribution—normalized minor component. The O 1s region of the “fresh” catalyst
to 100% fragments (Table 5)—showed propane, butane(s), after n-hexane reactions (Fig. 5) may have indicated
and pentane(s) as the main products. The “calc.” catalyst some rehydroxylation during repeated oxygen–hydrogen
produced more methane and ethane at lower temperatures. regenerations between catalytic runs.
Increasing p(H2) gave rise to less butane plus pentane and No detectable reduction of sulfate to give Pt sulfide (20,
more propane on both catalysts but isobutane and isopen- 23, 24) occurred. EM showed no signs of PtS particles, as
tane prevailed within the C4–C5 fraction.
opposed to Pt/MoOx/Al2O3 (50). Thus, S was present ex-
clusively as sulfate in the active catalyst.
Although electron microscopy cannot monitor the pres-
ence of ultrafine Pt particles, the decrease of particle size
is shown by XRD (Fig. 4). Had these smaller Pt particles
been evenly distributed on the surface or had they obtained
a raft-like surface structure, the XPS intensity of Pt would
have increased. Instead, the marked loss of Pt intensity may
indicate that a considerable fraction of Pt particles were
“embedded” into sulfated zirconia; i.e., they were partially
covered by the support, as reported for Pt/ZrO2 (51) as well
as for Pt/sulfated zirconia (52). This was confirmed also by
present ISS results (Fig. 8) as well as by ISS of similar sam-
ples (27). The in-beam reconstruction has been evidenced
also by EM indicating the appearance or disappearance of
Pt crystallites without influencing dramatically their shape
or size (see the twin in the middle of the Fig. 2 which kept
its shape although its exposition changed). As far as we
know, the effect of this dramatic solid-state rearrangement
of Pt/sulfated zirconia on the state and distribution of plat-
inum is being reported for the first time.
DISCUSSION
Structure of Active Sites in the “Fresh”
and Calcined Catalysts
Sulfated zirconia is regarded as a strongly acidic catalyst,
exhibiting Bronsted-acid sites such as OH/H2O attached
to sulfate groups and Lewis-acid sites as coordinatively
unsaturated Zr ions surrounded by sulfate groups (4,
11, 13, 44–47). There seems to be an optimum degree
of hydration (around 1.1 mol H2O/ZrO2) and sulfation
(ca 10 mol% ). Calcination of our samples seem to have
transformed most of the zirconia content to catalytically
active (45, 46) tetragonal ZrO2 (Fig. 4). The O 1s doublet,
similar to that reported for ZrO2 reduced in hydrogen
plasma (48), may indicate the presence of a large number
of ZrOH groups (BE(max.) ꢃ533 eV (49)), together with
sulfate groups (35). The atomic ratio of O/S being much
higher than unity (Table 2) indicates that Zr hydroxide
TABLE 5
Fragment Distribution over “Fresh” and “Calc.” Pt/Sulfated ZrO2 Catalysts as a Function of the Temperature and Hydrogen
Pressure,a Together with Results Obtained on Other Pt Catalystsb
Fragment composition, % c
Sample
“Fresh”
T, K
p(H2), Torr
C1
C2
C3
C4
C5
(18)
(14)
(9)
Ref.
543
573
603
120
120
60
120
480
51
45
46
48
31
33
32
31
10
(31)
18
22
14
18
This work
(24)
(15)
(19)
(5)
4
—
11
4
3
6
(11)
(3.5)
66.5
6.5
“Calc.”
543
573
603
120
120
60
120
480
10.5
7
1.5
2.5
8
6
5
3
4
54
60
62
69
15.5
17
16
10.5 (10)
3
(7.5)
(11)
(12)
14
11
(7)
(8)
This work
17.5 (13)
14
9
(10)
(7)
5.5
74.5
(3)
d
Pt/SiO2
603
603
603
120
120
120
480
120
35
46
21
25
11
20
20
15
16
22
45
13
12
14
13
27
(0)
(0)
(1)
(0)
(21)
12
21
39
30
14
(0)
(0)
(6)
(0)
(8)
42
40
40
40
40
Pt/NaYd
Pt/NaHYd
6
10
10
3
Pt/HYd
603
a Sampling time, 5 min., p(nH) = 10 Torr.
b Conversion values, X, see Table 3 for Pt/sulfated zirconia catalysts. X = 8% for Pt/SiO2 and around 4–5% for the zeolite supported ones.
P
c Normalized to (fragments) = 100% . In brackets: branched fragment, isobutane or isopentane (in the same units) within the total amount of C4
and C5 displayed.
d Recalculated values after original data serving as a basis for the references indicated.

Pt/SULFATED ZIRCONIA CATALYSTS
349
The chemical state of Pt did not correspond to that of
pure Pt0 (Fig. 7). The model of metallic Pt particles sur-
rounded by an adsorbed overlayer (14, 22–26) would agree
–
–
better with our results (27). Pt O, rather than Pt S inter-
actions are likely both in the “fresh” and “calc.” samples
(Fig. 7) in the “as received” state. The structural changes
shown by EM and especially by XRD (Figs. 1 to 4) indicate
that these are brought about by Pt and support oxygen ions
in atomic closeness (27), similarly to recent excellent EX-
AFS studies of supported Pt catalysts (53, 54), where the
–
“short” Pt O bonds mean a direct Pt–support interaction.
–
After calcination, several such “short” Pt O bonds must
be present. In-situ hydrogen reduction brought the Pt 4f
7/2 BE closer to that of the metallic state, because it could
be inserted between the Pt and support interface, splitting
–
some of these Pt O bonds. This explains why the BE of Pt
0
–
was closer to Pt in the “calc.” sample, the remaining Pt O
bonds causing still some shift (Fig. 7). Most of the Pt sur-
face being surrounded by support species may explain the
apparent hindered hydrogen chemisorption (13, 19). This
–
structure could explain the presence of Pt O bonds demon-
strated by EXAFS in a paper which appeared after submit-
ting the first draft of the present study (55). It attributed
this structure to Pt with an O overlayer. Figure 11 offers
an alternative explanation for the structure of Pt-sulfated
zirconia ensembles after calcination showing Pt particles
surrounded almost totally by the support. It may have been
just the “fresh” sample where some Pt was interacting with
support OH groups and/or H2O which were found to be
stable on Pt surface, too (29, 38). Our model explains the
drastic drop of Pt intensity upon calcination (Figs. 7, 8) and
does not contradict to the facts of Ref. (55).
Redispersion at the same time gives rise to few but very
active sites, asshown bythe increase ofactivityper unit mass
of Pt (Fig. 9). These sites may be analogous to the small Pt
ensembles anchored by a proton to a zeolite framework
(41, 56). These entities may have catalyzed bifunctional re-
actions without shuttling the intermediates in the gas phase
(40, 41). Some of the catalytic properties of the zeolitic sam-
ples (40–42) have also been included in Tables 3 and 5.
With a stronger acid like sulfated zirconia, a concerted
metal–acid catalysis may prevail. Sachtler et al. (57) called
those combined active ensembles “collapsed bifunctional
sites.” We prefer the word “compressed bifunctional sites”
to avoid the meaning of “ruin” or “decay” for “collapse”—
these sites are not ruined; on the contrary they get new force
by being compressed. These sites do need neighboring Pt
and sulfate (and/or sulfate-activated Zr) sites. Not all sul-
fate species may fulfill this requirement, all the more that
the surface abundance of sulfate groups is higher than that
of Pt (Table 2). SO24ꢆ was shown to stabilize small tetragonal
ZrO2 crystallites on well-faceted particles and even induc-
ing their faceting in the (110) direction (58). Thus, most of
SO24ꢆ groups would only be spectator species and a small
FIG. 11. Proposed structure of the Pt-sulfated zirconia catalysts be-
fore calcination (upper panel); and after calcination (lower panel). The
drawing is largely symbolic, and not to scale, apart from indicating the de-
crease of the size of Pt particles in the “calc.” sample. Patches of adsorbed
OH (partly migrating from the strongly hydrated zirconia) are assumed
on the larger Pt particles of the “fresh” sample. The lower panel illustrates
the wrapping most Pt under the zirconia. The sulfated character of the sup-
port is not indicated in the figure, since Fig. 4 gives only evidence on ZrO2
recrystallization. The shift of BE(max) values of Pt 4f spectra (Fig. 7) can
be attributed to atomic interaction of most of the interfacial Pt atoms with
support O atoms in those particles (shown by shading, not to scale) rather
than to covering Pt by an O adlayer (55). The inset of the lower panel
gives a very rough illustration on the atomic scale for the ”compressed bi-
functional sites” emerging at the metal–support borderline. They should
involve a neighboring sulfate entity as shown by “Sul,” without attempting
to reproduce the disputed (2, 3, 46, 58) molecular structure of the latter.
number of ZrO2 crystallites with high Miller-index surfaces
would give rise to a few highly active acidic sites. If they
meet Pt sites in their close vicinity, combined active ensem-
bles arise, as illustrated schematically in Fig. 11. Pt may also
stabilize these highly active sulfate species. They need spe-
cific ZrO2 crystal planes which are not present on the mainly
amorphous zirconia in the “fresh” state exhibiting poor
activity although the amount of sulfate was not very differ-
ent from that found after calcination (Table 1, Fig. 6). The
validityofthe modeliscorroborated byfailingto prepare an
active isomerization catalyst via introducing platinum into
sulfated zirconia after its calcination. Pt and sulfated zirco-
nia come to the necessary vicinity by recrystallization and
the necessary price to pay for creating such entities is the
loss of a considerable fraction of Pt. This may be one of the
reasons why the accessibility (59) and reducibility (60) of
Pt becomes lower upon calcination. Our idea on wrapping

350
MANOLI ET AL.
up of most Pt would give an adequate explanation for these sample even at lower temperature (Table 5) indicates some
results, too, and would agree with the proposed chemical metal catalyzed hydrogenolysis along with the predominant
interaction (60) between Pt and support.
cracking. However, considering its low activity, we do not
believe that the alleged function of Pt to hydrogenolyze
coke precursors (47, 61) can be important. Even higher hy-
drogen pressure (603 K, 480 Torr) could not bring about
any metal catalyzed hydrogenolysis, as opposed to Pt/NaY
(40)—see Table 5. Instead, the complementary fragments
become almost equal, indicating that abundant hydrogen
may have shifted the decomposition of n-hexane fully to-
wards the intramolecular mechanism.
Effect of Hydrogen Pressure on the Catalytic Properties
Literature reported positive (11, 13, 61) as well as nega-
tive (47, 62) effect of increasing hydrogen pressure on the
catalytic activity. Garin et al. (12) were the only ones who
studied the H2 effect in a wide pressure range and found
maximum activity as a function of p(H2). Hydride trans-
fer agents are necessary to start the isomerization reaction
(13); in our case this is H2 dissociated on metal sites. To
act so, they should preserve their metallic character. Our
ideas on the structure of Pt-acidic combined active ensem-
bles agree with this assumption. The reason why the yields
should have maxima as a function of hydrogen pressure are
the same as was earlier proposed for metals (17, 18, 63)
even if the actual steps that are promoted and hindered,
respectively, by surface hydrogen may not be the same. At
low H2 pressure, the availability of hydrogen is scarce. Near
maximum, there is an optimum lifetime of surface species
which, in turn, is governed by the accessibility of hydrogen,
as well as by the acid strength (6, 13, 57). The shift of max-
ima toward higher p(H2) values as temperature increases
is due to the fact that higher H2 pressures are necessary to
maintain the same optimum coverage (17, 18). Too much
hydrogen, in turn, competes with the intermediates on the
active sites and may cause their too rapid desorption (13).
Apart from the maximum in the methylpentane (and a
much shallower maximum for dimethylbutane) yields, the
behavior of the catalysts changes profoundly at 603 K with
changing hydrogen pressures. At low p(H2), the catalysts
behave like a traditional metal supported on acidic sup-
port; dehydrogenation to olefins and cracking on acidic
sites appeared, in addition to isomerization (Fig. 10).
The peculiar character of metal–acid cooperation started
above ꢃ120 Torr H2 .These active sites produced mostly
methylpentanes from n-hexane. The surface residence time
of carbenium ions must have been insufficient (13) to form
isomers with quaternary C atoms as shown by the DMB/MP
ratios (Table 4) being far below equilibrium. The hydrogen
pressure dependence of this ratio is in agreement with the
importance of optimum surface residence time; with more
hydrogen present, the residence time becomes shorter and
less time is available for a second isomerization step.
ACKNOWLEDGMENTS
Z.P. is pleased to thank Deutsche Akademische Austauschdienst
(DAAD) for a fellowship to support his stay in Berlin during which the fi-
nal experiments were carried out and the manuscript could be completed.
The kind help of Ms. A. Hornung during this period is gratefully acknowl-
edged. The assistance of M. Lavergne in TEM measurements as well as
that of Dr. C. Ne´meth and Ms. A. Ko¨ke´ny for preparing Fig. 11 is greatly
appreciated.
REFERENCES
1. Olah, G. A., and Molna´r, A., “Hydrocarbon Chemistry,” Wiley, New
York, 1995.
2. Davis, B. H., Keogh, R. A., and Srinivasan, R., Catal. Today 20, 219
(1994).
3. Song, X., and Sayari, A., Catal. Rev.-Sci. Eng. 38, 329 (1996).
4. Tanabe, K., and Hattori, H., in “Handbook of Heterogeneous Catal-
ysis” (G. Ertl, H. Kno¨zinger, and J. Weitkamp, Eds.), Vol. 1, p. 404.
Verlag Chemie, Weinheim, 1997.
5. Hino, M., Kobayashi, S., and Arata, K., J. Am. Chem. Soc. 101, 5439
(1979).
6. Adeeva, V., de Haan, J. W., Ja¨nchen, J., Lei, G. D., Schu¨nemann,
G., van de Ven, L. J. M., Sachtler, W. M. H., and van Santen, R. A.,
J. Catal. 151, 364 (1995).
7. Yori, J. C., Luy, J. C., and Parera, J. M., Appl. Catal. 46, 103 (1989).
8. Hosi, T., Shimadzu, T., Ito, S., Baba, S., Takaoka, H., Imai, T., and
Yokoyama, N., in “Prepr. Sym. Div. Petr. Chem. Am. Chem. Soc.,
1988,” p. 562.
9. Ebitani, K., Konishi, J., Horie, A., Hattori, H., and Tanabe, K., in
“Acid-Base Catalysis” (K. Tanabe et al., Eds.), p. 491, Kodansha,
Tokyo, 1989.
10. Wen, M. Y., Wender, I., and Tierney, J. W., Energy Fuels 4, 372 (1990).
11. Ebitani, K., Konishi, J., and Hattori, H., J. Catal. 130, 257 (1991).
12. Garin, F., Andriamasinoro, D., Abdulsamad, A., and Sommer, J.,
J. Catal. 131, 199 (1991).
13. Iglesia, E., Soled, S. L., and Kramer, G. M., J. Catal. 144, 238 (1993).
14. Sayari, A., and Dicko, A., J. Catal. 145, 561 (1994).
15. Vaudagna, S. R., Comelli, R. A., and Figoli, N. S., React. Kinet. Catal.
Lett. 58, 1 (1995).
16. Hsu, C.-Y., Heimbuch, C. R., Armes, C. T., and Gates, B. C., J. Chem.
Soc. Chem. Commun. 1645 (1992).
17. Paa´l, Z., and Menon, P. G., Catal. Rev.-Sci. Eng. 25, 223 (1983).
18. Paa´l, Z., Catal. Today 12, 297 (1992).
19. Ebitani, K., Tsuji, J., Hattori, H., and Kita, H., J. Catal. 135, 609(1992).
20. Ebitani, K., Konishi, J., Tanaka, T., and Hattori, H., J. Catal. 135, 60
(1992).
Although cracking was a minor reaction, yet, fragment
composition may serve as still another indicator of the pos-
sible reaction routes. It revealed (Table 5) a predominant
acid-catalyzed cracking. The dimeric intermediates (12, 13,
61, 62) to give C4 and C5 products (of these, branched ones)
may be more important on the “fresh” catalysts containing
less “compressed bifunctional sites.” The bimolecular path-
way should not be important for C6 alkane isomerization
(57). The appearance of C1 and C2 fragments on the “calc.”
21. Paa´l, Z., Muhler, M., and Schlo¨gl, R., J. Catal. 143, 318 (1993).
22. Tanaka, T., Shishido, T., Hattori, H., Ebitani, K., and Yoshida, S.,
Physica B 208/209, 649 (1995).
23. Ebitani, K., Tanaka, T., and Hattori, H., Appl. Catal. A 102, 79 (1993).

Pt/SULFATED ZIRCONIA CATALYSTS
351
24. Dicko, A., Song, X., Adnot, A., and Sayari, A., J. Catal. 150, 254
(1994).
25. Zhao, J., Huffman, G. P., and Davis, B. H., Catal. Lett. 24, 385 (1994).
26. Comelli, R. A., Catal. Lett. 40, 67 (1996).
27. Paa´l, Z., Buchholz, T., Resofszki, G., Muhler, M., Wild, U., Manoli,
J.-M., and Potvin, C., “11 Internat. Congr. Catalysis, Baltimore, 1996,”
Poster 029.
44. Nascimento, P., Akratopoulou, C., Oszagyan, M., Coudurier, G.,
Travers, C., Joly, J.-F., and Vedrine, J. C., Stud. Surf. Sci. Catal. 75,
1185 (1993).
45. Morterra, C., Cerrato, G., Pinna, F., and Signoretto, M., J. Catal. 157,
109 (1995).
46. Spielbauer, D., Mekhemer, G. A. H., Zaki, M. I., and Kno¨zinger, H.,
Catal. Lett. 40, 71 (1996).
28. Paa´l, Z., Muhler, M., and Matusek, K., Appl. Catal. A 149, 113 (1997).
29. Paa´l, Z., and Zhan, Z., Langmuir 13, 3752 (1997).
30. Paa´l, Z., Xu, X. L., Paa´l-Luka´cs, J., Vogel, W., Schlo¨gl, R., and Muhler,
M., J. Catal. 152, 252 (1995).
47. Signoretto, M., Pinna, F., Strukul, G., Chies, P., Cerrato, G., Di Ciero,
S., and Morterra, C., J. Catal. 167, 522 (1997).
48. Wong, P. C., Li, Y. S., Zhou, M. Y., and Mitchell, K. A. R., Appl. Surf.
Sci. 89, 255 (1995).
31. Moulder, J. F., Stickle, W. F., Sobol, P. E., and Bomben, K. D., “Hand-
book of X-Ray Photoelectron Spectroscopy,” Perkin–Elmer, Eden
Prairie, MN, 1992.
32. Briggs, D., and Seah, M. P. (Eds.), “Practical Surface Analysis,” Ap-
pendix 6, Vol. 1, p. 635. Wiley, Chichester, 1990.
49. Li, Y. S., Wong, P. C., and Mitchell, K. A. R., Appl. Surf. Sci. 89, 263
(1995).
50. Paa´l, Z., Te´te´nyi, P., Muhler, M., Manoli, J.-M., and Potvin, C.,
J. Chem. Soc. Faraday Trans. 94, 459 (1998).
51a. Hoang, D.-L., and Lieske, H., Catal. Lett. 27, 33 (1994).
33. Muhler, M., Paa´l, Z., and Schlo¨gl, R., Appl. SurfaceSci. 47, 281(1991). 51b. Hoang, D.-L., Berndt, H., and Lieske, H., Catal. Lett. 31, 165 (1995).
34. Morant, C., Sanz, J. M., Gala´n, L., Soriano, L., and Rueda, F., Surf.
Sci. 218, 331 (1989).
52. Xue, E., Seshan, K., Mercera, P. D. L., van Ommen, J. G., and Ross,
J. R. H., ACS Symp. Series 552, 250 (1994).
35. Milburn, D. R., Keogh, R. A., Srinivasan, R., and Davis, B. H., Appl.
Catal. A 147, 109 (1996).
36. Wang, Y. M., Li, Y. S., Wong, P. C., and Mitchell, K. A. R., Appl. Surf.
Sci. 72, 237 (1993).
53. Miller, J. T., Meyers, B. C., Modica, F. S., Lane, G. S., Vaarkamp, M.,
and Koningsberger, D. C., J. Catal. 143, 395 (1993).
54. Munoz-Pa´ez, A., and Koningsberger, D. C., J. Phys. Chem. 99, 4193
(1995).
37. Maurice, V., Salmeron, M., and Somorjai, G. A., Surf. Sci. 237, 116
(1990).
38. Paa´l, Z., Schlo¨gl, R., and Ertl, G., J. Chem. Soc. Faraday Trans. 88,
1179 (1992).
55. Shishido, T., Tanaka, T., and Hattori, H., J. Catal. 172, 24 (1997).
56. Bai, X., and Sachtler, W. M. H., J. Catal. 129, 121 (1991).
57. Liu, H., Lei, H., and Sachtler, W. M. H., Appl. Catal. A 137, 167
(1996).
39. Bond, G. C., and Paa´l, Z., Appl. Catal. A 86, 1 (1992).
40. Zhan, Z., Manninger, I., Paa´l, Z., and Barthomeuf, D., J. Catal. 147,
333 (1994).
41. Paa´l, Z., Zhan, Z., Manninger, I., and Sachtler, W. M. H., J. Catal.
155, 43 (1995).
42. Paa´l, Z., Groeneweg, H., and Paa´l-Luka´cs, J., J. Chem. Soc. Faraday
Trans. 86, 3159 (1990).
43. Haag, W. O., and Dessau, R. M., in “Proceedings, 8th International
Congress on Catalysis, Berlin, 1984,” Vol. 2, p. 305. Verlag Chemie,
Weinheim, 1984.
58. Benaissa, M., Santiesteban, J. G., Diaz, G., Chang, C. D., and Jose´-
Yacaman, M., J. Catal. 161, 694 (1996).
59. Coq, B., Walter, C., Brown, R., McDougall, G., and Figueras, F., Catal.
Lett. 39, 197 (1996).
60. Ivanov, A. V., Kustov, L. M., Vasina, T. V., Kazanskii, V. B., and
Zeuthen, P., Kinet. Catal. (Eng. ed.) 38, 403 (1997).
61. Comelli, R. A., Finelli, R., Vaudagna, S. R., and Figoli, N. S., Catal.
Lett. 45, 227 (1997).
62. Sayari, A., Yang, Y., and Song, X., J. Catal. 167, 346 (1997).
63. Bond, G. C., and Slaa, J. C., J. Mol. Catal. 98, 81 (1995).