A transient TAP study of the adsorption of C4-hydrocarbons on sulfated zirconias

Article abstract of DOI:10.1016/j.molcata.2004.10.051

The interaction of C₄-hydrocarbons with the surface of two sulfated zirconias of considerable different catalytic activity has been investigated by using the temporal analysis of products (TAP) reactor. n-Butane and iso-butane as reactant and product molecules for the isomerization showed differences in their interaction with the surface of sulfated zirconia thus indicating slightly variable sorption features for both molecules. 1-Butene, as representative for a possible reaction intermediate or side product, had compared to the saturated C₄-molecules stronger adsorption abilities and desorbed slower from the surface of sulfated zirconia, which may also influence the sorption properties of n-butane and iso-butane during the catalytic cycle. Heats of adsorption from van't Hoff plots were modelled for n-butane and iso-butane and are comparable to calorimetry data for this system thus giving the opportunity to model the first step in a microkinetic model appropriately.

Full text of DOI:10.1016/j.molcata.2004.10.051

Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
A transient TAP study of the adsorption of
C -hydrocarbons on sulfated zirconias
4
∗
Cornelia Breitkopf
Universit a¨ t Leipzig, Institut f u¨ r Technische Chemie, Linnestr. 3, D-04103 Leipzig, Germany
Received 16 August 2004; received in revised form 26 October 2004; accepted 28 October 2004
Abstract
The interaction of C
4
-hydrocarbons with the surface of two sulfated zirconias of considerable different catalytic activity has been investigated
by using the temporal analysis of products (TAP) reactor. n-Butane and iso-butane as reactant and product molecules for the isomerization
showed differences in their interaction with the surface of sulfated zirconia thus indicating slightly variable sorption features for both
molecules. 1-Butene, as representative for a possible reaction intermediate or side product, had compared to the saturated C -molecules
4
stronger adsorption abilities and desorbed slower from the surface of sulfated zirconia, which may also influence the sorption properties
of n-butane and iso-butane during the catalytic cycle. Heats of adsorption from van’t Hoff plots were modelled for n-butane and iso-
butane and are comparable to calorimetry data for this system thus giving the opportunity to model the first step in a microkinetic model
appropriately.
©
2004 Elsevier B.V. All rights reserved.
Keywords: Temporal analysis of products; Sulfated zirconia; Alkane adsorption; Modelling of adsorption
1
. Introduction
This paper deals with the description of the adsorption
and desorption properties of the main reactant and product
molecules of the n-butane isomerization reaction. Isomeriza-
tion is studied over the surface of sulfated zirconia, which
is known to catalyze the isomerization at low temperatures
[6]. Moreover, two samples of considerable different activ-
ity towards iso-butane formation were chosen to compare the
adsorptive interaction of C4-hydrocarbons on them.
The activation of alkanes on the surface of sulfated
zirconias is still under discussion [7]. Adsorption is the
first and main important step in the microkinetics of this
process. Therefore, a detailed investigation of this activa-
tion step should give more insights regarding the general
mechanism of alkane activation on sulfated zirconias and
an adequate description of the adsorption step will be the
starting point for an appropriate kinetic model to come.
Heats of adsorption resulting from van’t Hoff plots will
be compared for the probe molecules n-butane and iso-
butane on the surface of sulfated zirconia with variable
activities.
Transient measurements in a temporal analysis of products
(TAP) reactor are a very fast and convenient way to determine
the adsorption and desorption properties of gaseous probe
molecules on catalytic active surfaces [1]. Sorption experi-
ments are used to determine diffusion constants, adsorption
and desorption constants as well as activation energies [2,3].
Moreover, transient studies allow a detailed view to microki-
netics as they are able to focus on the investigation of single
reaction steps. Only a limited number of transient studies ex-
ist which focus on the adsorption of alkanes on solid surfaces
[4,5]. So far no TAP study is known regarding the alkane
interaction with sulfated zirconias and the modelling of their
adsorption.
∗
Tel.: +49 341 9736307; fax: +49 341 9736349.
E-mail address: cbreit@chemie.uni-leipzig.de.
1
381-1169/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2004.10.051

2
70
C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
2
. Experimental
a quadrupole mass spectrometer (QMS) from HIDEN, where
single masses (m/e) of interest can be followed. All possible
fragments for n-butane and iso-butane were monitored. Frag-
mentation patterns of the pure substances were recorded sep-
arately in a reactor filled with corundum (dp = 0.1–0.3 mm).
Only their main mass fragments m/e = 27, 29, 41, 43, 58 will
be discussed in detail here. 1-Butene was mainly monitored
at m/e = 41, 56.
Knudsen diffusion in the reactor during pulse experiments
was ensured by controlling the pulse intensity of the re-
sponses. The shape of the response curves has to be indepen-
dent from the pulse size. Pulse sizes were then kept constant.
Modelling was done by using the software by Schuurman
[14], which gives estimates for adsorption and desorption
rate constants. These values were used to calculate the equi-
librium constant K. van’t Hoff plots (ln K versus 1/T) gave
then estimates for heats of adsorption.
2
.1. Preparation of sulfated zirconia
The sulfated zirconias were prepared through precipita-
tion from zirconyl nitrate solution with ammonia to a final
pH of 8.4. The zirconias were aged either for 1 h at room
temperature (series 1) or for 24 h at 373 K (series 2). The
precipitates were washed and dried (24 h at 393 K) resulting
in two precursors referred to as P1 and P2. Both precursor
materials were sulfated by suspending them in a solution of
ammonium sulfate (theoretical Zr:S ratio of 1:5). The sulfated
materials were dried 24 h at 373 K. Calcination of the material
(2 g) was done in a vertical reactor with a constant flow of air
−
1
(30 ml min ) for 3 h at 873 K giving the catalyst materials
C1 and C2, respectively. For comparison an industrial ref-
erence material from MEL Chemicals (XZO-620) was used
and is referred here as C3. The unsulfated precursors were
also calcined under the same conditions and are named P1
and P2, respectively.
2.3. Characterization of the samples
Surface area measurements were performed with an
ASAP-2000 (Micromeritics). For X-ray diffraction the pow-
ders were investigated with XRD-7-diffractometer (Seifert)
using Cu K␣ radiation. The sulfur content was determined by
elementary analysis [11]. For temperature-programmed des-
orption (TPD) of NH , the catalysts were exposed to NH at
2
.2. TAP experiments
Temporal analysis of products (TAP) was conducted in a
TAP-2 system [8] in the temperature range 323–523 K. Sam-
ples were activated before use in TAP at 573 K for 2 h in
vacuum in the microreactor. Fresh samples (100 mg) were ex-
posed to single pulses of n-butane, iso-butane and 1-butene.
The gases employed were n-butane (99.5 vol.%, Messer), iso-
butane (99.95 vol.%, Fluka), 1-butene (99.0 vol.%, Messer).
Different inert reference gases were used (He, Ar) for com-
parison. In first measurements helium was used (323–523 K
for the precursors and catalysts). Also measurements with ar-
gon (373–523 K for catalyst samples only) were performed
to have a basis for a subsequent modelling procedure. It is
reported that modelling Knudsen diffusion coefficients for
helium is affected by its low molecular mass, which makes it
unable as basis for the evaluation of adsorption and desorp-
tion constants [9]. Thus, all modelling procedures were done
basedontheargonmeasurements. Ithastobenotedthatunder
the Knudsen diffusion regime the gases in a mixture do not in-
teract thus the use of a special inert gas does not influence the
measurement outcome at all. Pulses of a constant pulse width
of 130 ␮s were used and the reactant mixtures in the blend-
ing tank consisted of 50 mbar inert gas and 350 mbar reactant
gas. An experiment consisted of several cycles of pulses with
signal averaging (in general 10 pulses to improve the signal-
to-noise ratio). The reactor bed was built up from inert ma-
terial (corundum) of 400 mg, a catalyst section with 100 mg
followed by inert material of 600 mg. The particle diameter
for all materials was a sieve fraction of 0.1–0.315 mm as used
in our catalysis experiments in the fixed bed reactor at atmo-
spheric pressure. The temperature was controlled by a ther-
mocouple positioned in the centre of the bed. The catalyst was
heated to the measurement temperatures in vacuum at a rate
of 10 K/min. The gas outlet from the reactor was detected by
3
3
373 K with 300 ␮l NH . Then the samples were heated at a
3
−
1
rate of 10 K min and desorbing NH was detected by MS
3
(m/e = 15).
3. Results and discussion
3.1. Characterization
Table 1 summarizes the results from nitrogen sorption
for the precursor and catalyst materials. The specific sur-
face areas were calculated according the BET equation
[16]. The pore size distributions were calculated by the
Barrett–Joyner–Halenda (BJH) method [17]. The desorption
branches were used for calculating the pore sizes. The unsul-
fated zirconias P1 and P2 exhibited only half of the surface
areas of C1 and C2. It is well known that sulfation has a
great influence on the formation of higher surface areas of
zirconias [10]. Calcination of the sulfated samples resulted
in a decrease in pore diameters and thus the surface areas
2
were increased. C1 had a surface area of 103 m /g, C2 one
2
of 150 m /g. The sulfate content for C1 was 9 wt.%, i.e. the
number of sulfur atoms on the surface can be estimated to be
Table 1
Nitrogen sorption data for precursors P1, P2 and catalyst samples C1–C3
Sample
P1
C1
P2
C2
C3
2
BET (m /g)
48
63
103
35
86
50
150
40
122
–
Pore diameter ( A˚ )

C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
271
Fig. 1. Ammonia-TPD for catalyst materials C1 (—), C2 (· · ·) (C1:
−
3
−2
−3
−2
5
.0 × 10 mmol m ; C2: 3.2 × 10 mmol m ).
1
8
2
5
.3 × 10 S-atoms/m . C2 had a comparable sulfate content
18
of 8.9 wt.%, thus resulting in sulfur coverage of 3.7 × 10 S-
2
atoms/m , which is therefore significantly smaller, compared
to C1. The reference sample C3 had a sulfur content of
1
8
6
.5 wt.% and the lowest sulfate coverage of 2.6 × 10 S-
2
atoms/m .
The XRD data of C1–C3 (not presented here) showed
typical diffraction peaks for the tetragonal phase of zirconia.
They were mainly tetragonal with only a small monoclinic
contribution [11].
The results from ammonia TPD are illustrated in Fig. 1.
They do not show a significant difference for both sam-
ples. C1 and C2 exhibited the typical broad desorption
profile for ammonia from sulfated zirconias. The concen-
Fig. 2. Response curves for single pulses of n-butane over C1 for 373–523 K,
inert reference argon: (a) m/e = 43; (b) m/e = 58.
−
3
tration of acid sites can be estimated as 5.0 × 10 and
−
3
−2
3
.2 × 10 mmol m for C1 and C2, respectively. As both
of interest was the investigation and comparison of samples
with differing porosity and catalytic activity.
samples had the same sulfur content this is a reasonable re-
sult.
The shapes of all curves indicated reversible adsorption
behaviour of n-butane on the surface of sulfated zirconia,
which is an important feature out of a TAP pulse experiment.
Fig. 2 shows the response curves for pulsing n-butane over
fresh C1. Argon was used as reference gas. From the response
curves a strong adsorption of n-butane can be concluded.
An extremely long desorption phase was observed for low
temperatures. Measurements with Helium (not shown here,
similar to the Argon outcomes) were first done also at 323 K.
Here it could be observed a rather long desorption phase up
to 30 s for the pulse response at 323 K. Therefore, a time
interval for pulsing of up to 30 s was chosen to desorb the
probe molecules totally.
Thecatalyticexperimentsaredescribedindetailinaprevi-
ous paper [11]. C1 had the highest activity towards iso-butane
formation (conversion is 20% at 423 K) followed by rapid de-
activation. C2 and C3 were not or less active and did not show
long term stability. This significantly different catalytic be-
haviour of the samples C1 and C2 was quite interesting as
they exhibited the same overall sulfur content, which should
be related to the isomerization activity. But due to their dif-
ferent specific surface areas the sulfate group density seems
to make the difference in the catalytic activity. The reason
for this dependence is still not clear and under consideration
[11]. These two samples were therefore an interesting start-
ing point to investigate adsorption, desorption and reaction
within a TAP experiment.
Especially the relations between the observed intensities
of m/e = 58 (parent peak) and m/e = 43 are shown in detail in
Fig. 2. For both masses (43, 58) the same curve shapes were
observed. The measured intensities of the response curves for
m/e = 43 were much higher compared to those from m/e = 58,
which corresponds to the intensity given by the general frag-
mentation pattern of C4-hydrocarbons in the MS which was
observed over inert material, too.
3
.2. TAP—interaction with n-butane
TAP experiments were carried out for the first time to
characterize the sorption abilities of sulfated zirconias while
interacting with reactant and product molecules. Especially

2
72
C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
In principle the trend of the response curves for the sam-
ples C2 and C3 with increasing temperature was similar (not
shown here). C3 as reference catalyst was, however only in-
vestigated at 423 and 523 K. The absolute measured intensi-
ties for C2 and C3 varied compared to C1 indicating different
surface interactions according to their catalytic activities.
It has to be noted that a TAP experiment is state defining,
due to the very small number of molecules, which are pulsed
during a cycle. Thus, the observed pulse responses for all
samples correlate to a nearly unchanged catalyst surface. This
would correspond to a time on stream in the very beginning
of a catalysis experiment under atmospheric conditions.
Comparing all response curves in Fig. 2 it is obvious that
the interaction of the probe gas changed with temperature.
At low temperatures the adsorption of the probe molecule
n-butane was rather strong. It took some seconds for the to-
tal desorption of the pulsed/converted molecules from the
surface. Even at 423 K the adsorption was still strong. From
4
58 K up to 523 K the desorption time was shortened to less
than 1 s and the process of diffusion governed the curve re-
sponse shape. The curves tended to be narrower with increas-
ing temperature.
A quantitative calculation of conversion of n-butane to
iso-butane from TAP experiment is difficult because of the
similar fragmentation pattern for n-butane and iso-butane in
MS. It can be partly achieved by the comparison of frag-
ment ratios instead of single fragment masses for different
temperatures and samples. If the same processes, i.e. adsorp-
tion/reaction/desorption occur on the surface, similar frag-
ment ratios should be observed. To calculate these fragmen-
tation ratios the areas of the response curves were evaluated
for each temperature and special response curve. It has to be
noted that for masses with lower intensity the error in esti-
mating these areas is higher through a lower signal-to-noise
ratio.
For iso-butane especially m/e = 43 should be more pro-
nounced through stabilization of this fragment. If we had a
conversion of n-butane to iso-butane the general fragmenta-
tion ratio of m/e = 43 to 58 should be different in comparison
to an inert standard. In Table 2 the observed fragmentation
pattern for C1–C3 and those expected from pure n-butane
and iso-butane over inert material are listed for comparison.
Values for the range 373–423 K will be discussed in detail
only.
Fig. 3. Calculated fragmentation ratios: m/e = 27/m/e = 40; m/e = 41/
m/e = 40; m/e = 29/m/e = 40; m/e = 27/m/e = 41; m/e = 43/m/e = 29; m/e = 58/
m/e = 29.
be more pronounced for iso-butane this corresponds to the
observed isomerization activity known from catalysis, i.e. C1
is active and C2 and C3 are not or less active and there was
for the latter no change of the m/e = 43 to 58 ratios compared
to the inert material. An increase in the fragmentation ratio
with temperature, which would reflect a higher reaction rate
according to a higher temperature, may not be so pronounced
intheTAPresponsepulseasinacatalysisexperimentbecause
numerous simultaneous linked processes have to take into
account. However, there was a small increase for C1 and C2
observable with temperature.
To confirm a different product formation or reactant ad-
sorption/desorption on the surfaces of C1 and C2, additional
fragmentation ratios are shown in Fig. 3. It is obvious that all
fragment ratios of C1 and C2 varied in their absolute values
and also in their order of appearance which supports different
surface processes on C1 and C2 reflecting different activities
of the catalysts as observed in the catalytic experiments. The
fragments 27, 29 and 41 were chosen because they reflect
those fragments, which exhibit the highest changes between
n- and iso-butane when pulsed over inert reference material.
It is known that alkanes also adsorb on pure unsulfated
zirconia surfaces [12]. Thus, to investigate the sorption be-
haviour of n-butane on the surface of the porous and inac-
tive support material, P1 and P2 were investigated by puls-
ing reactant n-butane, too. As no conversion by pulsing n-
butane over these precursors is expected, the interaction of
the probe molecule should be only determined by the com-
bined processes of adsorption, desorption and diffusion. Thus
C1 showed clearly a ratio above the average value of eight
for pure n-butane measured with inert material. The values
for C2 and C3 lay below the value of C1. As m/e = 43 should
Table 2
Comparison of fragmentation pattern m/e = 43/m/e = 58 for C1–C3 with inert
reference argon with pure n-butane (43/58 = 8) and iso-butane (43/58 = 25)
measured over inert material
Sample
373 K
398 K
423 K
C1
C2
C3
11.5
6.6
–
11.6
8.3
–
14.2
7.5
11.3

C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
273
Table 3
Residence times (s) for pulses of n-butane for P1 and P2, C1 and C2 calcu-
lated from the first moments of the response curves, inert reference gas used
in brackets
Sample
K)
P1
(He)
P2
(He)
C1 (He)
C2 (He)
C1
(Ar)
C2
(Ar)
(
323
348
373
398
423
448
473
498
523
a
–a
2.7
1.8
0.83
0.40
0.23
0.19
0.18
0.17
9
4.9
1.5
0.78
0.39
0.27
0.19
0.15
0.13
18
8.5
5.5
1.54
0.53
0.28
0.17
0.14
0.12
13
8.9
2.7
0.88
0.40
0.24
0.16
0.14
0.12
–
–
–
1.05
0.50
0.28
0.19
0.15
0.13
–
–
–
1.08
0.45
0.26
0.17
–
0.12
Recording time was only 1 s, n-butane still desorbing, see Fig. 4.
all molecules in the response pulse can be attributed to n-
butane. The duration of the n-butane interaction with the pre-
cursor surface are can be estimated by the residence times.
Residence times are available over momentum analysis of
the response curves and are calculated as the first moments
here [13]. For the experiments with the precursor materials
Helium was used as reference gas, exceptionally. Also lower
temperatures of 323 and 348 K were investigated in this case.
Table 3 summarizes all residence times, calculated as first
moments from the response curves [14]. To get a systematic
overview the values calculated for C1 and C2 are listed, too.
The range of the response curve, which was considered forthe
calculation (integration), varies depending on the desorption
behaviour. This is especially important for the evaluation of
the low temperature response curves because of the observed
strong adsorption. The error made by an integration of curves
with a low signal-to-noise ratio is around 5%. The principal
trend was not affected at all. The slight differences between
the C1 and C2 values for the Helium and Argon measure-
ments lie within the error range and have no influence on the
interpretation at all.
Fig. 5. Height normalized intensities for (a) P1 and (b) C1 at different tem-
peratures.
peratures high residence times were observed corresponding
to strong adsorption of n-butane with the surface. The resi-
dence times decreased with higher temperatures showing less
interaction of the probe molecule with the surface. It is inter-
esting to compare the pairs P1/C1 and P2/C2. In both cases
n-butane was stronger adsorbed on the catalyst material in
comparison to the precursor. This is evidently the influence
of the surface sulfate groups. Corresponding to this in Fig. 5
the height normalized response curves are given for C1 show-
ing the significant changes of the maxima of the m/e = 43 re-
sponse curves compared to the precursor P1 (C2 changed as
well, but not shown here).
It can be seen that the residence times of n-butane on the
catalysts changed dramatically with temperature. At low tem-
The changes in the residence times, going from P1 to C1,
were more pronounced than those of the pair P2/C2. As both
catalyst samples have the same sulfate contents this points
again to the different surface processes (activity) occurring
on C1 and C2. The differences were especially strong at low
temperature where the highest activity for the isomerization
is expected. At higher temperatures (>473 K) however there
were no significant differences between the samples, i.e. the
residence times reached a constant value for all investigated
samples.
To illustrate the influence of porosity and surface activity
on the observed response intensities, a nonporous inert ma-
Fig. 4. Response curve for n-butane over P1 at 323 K.

2
74
C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
Table 4
Residence times (s) for pulses of iso-butane for P1 and P2, C1 and C2
calculated from the first moments of the response curves, inert reference
gases in brackets
Sample (K)
P1 (He)
P2 (He)
C1 (He)
C2 (He)
3
3
3
3
4
4
4
4
5
23
48
73
98
23
48
73
98
23
3.2
2.4
1.7
0.69
0.36
0.20
0.14
0.12
–
7.6
4.3
1.4
0.62
0.34
0.22
0.17
0.15
–
7.8
5.5
1.7
0.65
0.31
–
0.14
0.11
0.10
13.5
7.0
2.5
0.73
0.39
0.22
0.16
0.13
0.11
Fig. 6. Response curves for single pulses of n-butane over corundum, P1
and C1 at 373 K for m/e = 43 (1 g for corundum, 100 mg of sample for P1,
C1; inert reference helium) (intensities are corrected to reference helium for
corundum).
ened compared to those of n-butane. Second, an important
difference was found when comparing the pairs P1/C1 and
P2/C2. Iso-butane showed higher residence times for the pair
P2/C2, which was the opposite behaviour compared to that of
n-butane.
In Fig. 7, the response curves of m/e = 43 for three selected
temperatures (373, 398, and 423 K) for C1 and C2 are com-
pared, which represent the most interesting temperature in-
terval. For this comparison it was necessary to adjust slightly
different overall intensities measured on different days by
the MS taking the argon reference gas response of C1 as
terial (corundum) was compared with the porous inactive P1
and porous active C1 after pulsing n-butane. Fig. 6 shows the
changes in the intensities of the response curves for one se-
lected temperature (373 K). All data sets are corrected to the
intensity of helium, measured over corundum. Porosity in the
sample decreased the intensity due to an increased influence
of diffusion processes and also due to adsorption and desorp-
tion processes in the inner surface. The intensity of m/e = 43
was further lowered for the active material due to the en-
hanced interaction of the probe molecules with the sulfate
groups and by reaction. The differences in these response be-
haviours of precursor and catalyst will be further investigated
and may contribute at least to an estimation of the reaction
part in an isomerization reaction within a TAP experiment.
Additionally, the knowledge of the influence of porosity on
the response curve shape has to be known exactly when ex-
perimental outcomes will be used for modelling. It has also to
be noted that there may different processes of diffusion (like
pore diffusion) take place in porous solids besides the Knud-
sendiffusionhere. Asourporesystemconsistedofmesopores
only, an influence of pore diffusion on the experimental out-
come could not be observed here. All normalized response
curves showed only Knudsen behaviour as described in
Section 2.
100% and correct the responses of C2 to this value. As no
conversion of iso- to n-butane must be assumed due to ther-
modynamics, the m/e = 43 response should be in contrast to
the n-butane case unique for the probe molecule iso-butane.
Therefore, the higher intensities observed for C2 can be at-
tributed to a decreased adsorption of iso-butane on C2. C1
clearly adsorbed more iso-butane than C2, thus the response
intensities were lower for C1 than for C2. These results corre-
late with results from microcalorimetry [11] on this system.
Also other groups reported differences in n-butane and iso-
butane adsorption for SZ. Gonzalez et al. described two SZ
samples, where similar adsorption of n-butane was observed
while iso-butane adsorption differed for those two samples
[15].
As can be seen in Fig. 7 the response curves at one temper-
ature showed the same shape for both samples. This is in con-
trast to the n-butane interaction where the shape of the curves
was slightly different (not shown) and this again excludes any
contribution of reaction to the shape of the response curve
for the iso-butane responses. With increasing temperature
the shape of the curves was narrowed, too. However, there
were observed higher residence times for the pair P2/C2. If
C2 adsorbed less iso-butane, as could be shown from the
intensity profiles, the only explanation for this behaviour is
the higher surface area of the samples P2/C2 compared to
P1/C1.
The comparison of the nonporous and porous materials
for pulsing iso-butane is shown in Fig. 8. The differences in
the observed overall intensity between P1 and C1 were much
smaller compared to that for n-butane also indicating that
there is no influence of reaction here.
3
.3. TAP—interaction with iso-butane
In case of reaction iso-butane was formed and had to
be desorbed from the surface of the sulfated zirconia as
well as the not converted n-butane. Therefore, in a next
step it was investigated the interaction of the isomeriza-
tion product iso-butane with the porous active and inac-
tive samples. The calculated residence times are listed in
Table 4. Two effects can be observed. First, for all tem-
peratures the interaction times of iso-butane with surfaces
of unsulfated and sulfated zirconia was significantly short-

C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
275
Fig. 8. Response curves for single pulses of iso-butane over corundum, P1
and C1 at 373 K m/e = 43 (1 g for corundum, 100 mg of sample for P1, C1;
inert reference helium) (intensities are corrected to reference helium for
corundum).
The adsorptive interaction of 1-butene with the surface
of zirconia and sulfated zirconia was strong. No responses
were detected below 423 K. At these temperatures 1-butene
was irreversibly adsorbed. Further experiments are necessary
to investigate whether a saturation phenomenon can be ob-
served with an increasing number of pulses, which would
give a break through effect. Also the strong adsorption even
at 423 K has to be considered when the isomerization reaction
is investigated in more detail in that temperature range. The
formation of unsaturated side products may not be observable
in a TAP experiment.
The interaction of 1-butene differed for the precursors and
the catalyst materials. P1 and P2 showed the same trend with
temperature and had comparable residence times although
both differ in their specific surface areas. Thus, the interac-
tion of 1-butene with a plain zirconia surface was the same
for both samples. In comparison to that the interaction of
1-butene with the sulfated zirconias was stronger and even
for higher temperatures long residence times (>20 s) were ob-
served. And both catalyst samples differed in their interaction
with1-butene. Theobservedresponseintensitiesafterpulsing
1-buteneat473, 498, and523 KforC1andC2werecompared
(see Fig. 9). The adsorption of 1-butene was much stronger
on C2 than on C1. The curve shapes differed drastically for
all investigated temperatures. Fig. 9a exhibits especially for
C2 the typical response curve shape with a hump [13] for
a fast adsorption and long desorption process. The addition
Fig. 7. Corrected intensities of m/e = 43 (iso-butane) for C1 and C2 at dif-
ferent temperatures (a) 373 K, (b) 398 K, and (c) 423 K.
Table 5
3
.4. Interaction with 1-butene
-Butene is assumed to be a possible reaction intermedi-
Residence times (s) for P1 and P2, C1 and C2, calculated from the first
moments of the response curves pulses of 1-butene (m/e = 41)
1
Sample (K)
P1 (He)
P2 (He)
C1 (He)
C2 (He)
ate, side product or deactivating reactant of the isomerization
reaction. Therefore, the role of surface sulfate groups for the
binding of 1-butene was investigated by pulsing 1-butene.
Table 5 shows the calculated residence times for 1-butene
over porous inactive and porous active samples.
4
448
473
23
3.3
2.5
2.2
1.8
0.9
3.4
2.4
2.0
1.9
1.2
–
10
6.9
4.3
2.2
–
13
14
7.8
4.4
4
5
98
23

2
76
C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
Fig. 10. Response curves for single pulses of 1-butene over corundum, P1
and C1 at 473 K m/e = 41 (100 mg of sample for P1, C1, inert reference
helium) (intensities are corrected to reference helium for corundum).
It has to be noted that pulsing of 1-butene directly on the
surface may differ from a situation in the catalytic cycle,
where it may be formed as side product within the reac-
tion sequence. No comparable data are available to address
this question. However, the basic adsorption behaviour of 1-
butene should be reflected by these experiments.
3.5. Modelling—heats of adsorption
The response curves for pulsing n-butane and iso-butane,
measured at different temperatures, were used to model the
adsorption and desorption rate constants on the basis of a
simple first order Langmuir adsorption model without assum-
ing reaction. Reaction was excluded in these first attempts
of modelling, because the primary aim was to concentrate
on adsorption and desorption, only. The modelled curves re-
produced the measured response curves quite well. For low
temperatures (below 373 K) the modelled curves deviated
slightly from the measured ones. This effect has to be inves-
tigated in more detail in the future. It is known that Langmuir
adsorption may not be adequate if strong site heterogeneity
exists [4]. Also a contribution from reaction may lead to this
slight deviation. Despite this the basic trends of the mod-
elling was not affected at all. Now kinetic parameters can
be extracted out of the fit and the rate constants for adsorp-
tion and desorption and their ratios were derived. As long
as the underlying model does not contain precise values for
a special catalyst (i.e. number of adsorption sites or surface
area) the single calculated rate constants are not of physical
meaning. The ratio of both rate constants gives us however
the estimate for the equilibrium constant.
Fig. 9. Corrected intensities of m/e = 41 for C1 and C2 at different temper-
atures (a) 473 K, (b) 498 K, and (c) 523 K.
of sulfate groups enhanced the interaction of 1-butene with
the surface and the effect of the different surface area, which
increases the residence times, comes into play here.
The different response curves of 1-butene for the non-
porous and porous samples (at 473 K) are shown in Fig. 10.
The heats of adsorption from van’t Hoff plots were calcu-
lated then for C1 and C2. For the catalytically active C1 a heat
of adsorption of 53 kJ/mol was obtained as slope of the linear
fit (see Fig. 11). For the inactive sample C2 a quite similar
heat of adsorption of 52 kJ/mol was calculated. These values
are in very good agreement with calorimetry data [11,15].
1
-Butene is a strongly adsorbing agent on the surface of sul-
fated zirconia and compared to a nonporous solid also dif-
fusion was much more hindered compared to n-butane and
iso-butane.

C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
277
It could be shown that the adsorption of the reactant n-
butane is similar on the catalytically active or inactive sul-
fated zirconias. The TAP pulse experiments showed that
especially at low temperatures there was a strong adsorp-
tion of the reactant with extreme long desorption times.
The whole process of adsorption was reversible as can be
seen from the shape of the TAP pulse responses. These
performed TAP experiments refer to a nearly unchanged
catalyst surface. Further TAP experiments should be per-
formed on definitely altered surfaces of zirconia to be able
to describe the process of activation and deactivation more
precisely.
The results obtained on the inactive precursor materials
showed that also the unsulfated zirconia surface adsorbed
hydrocarbons to a certain amount. This may also influence
the measured adsorption on the sulfated zirconias which are
not totally covered by sulfate.
Fig. 11. van’t Hoff-plot for n-butane over C1 (᭹ modelled values, linear fit).
Thus both catalysts exhibited the same heat of adsorption for
n-butane. These values are not influenced by a change of the
actual catalyst coverage because single pulse experiments are
state-defining and should represent the heat of adsorption at
low coverages.
The product iso-butane showed different adsorption be-
haviour on catalytically active or inactive material. The ad-
sorption on the partially sulfate covered zirconia, which was
catalytically inactive, was less pronounced than on the zirco-
nia with a monolayer sulfate coverage.
However, in comparison to n-butane the heats of adsorp-
tion for iso-butane on active C1 and inactive C2 were more
different. The calculated heat of adsorption for iso-butane on
C1 was 48 kJ/mol, whereas for C2 it was 55 kJ/mol. This
again coincided with findings from calorimetry measure-
ments on these two samples cited above.
These experimental and modelling findings can be dis-
cussed as follows. The adsorption of n-butane on the surface
of SZ was similar for the reactive and the nonreactive sample
here and seems to be determined only by the overall sul-
fate content, which is quite similar for the two investigated
samples here. Also this seems to be independent from the
density of the sulfate groups on the surface of SZ, which
was in very good agreement with calorimetry data on this
special material [11]. These calorimetry measurements re-
vealed that the majority of sites on the two catalysts were
similar.
Temperature dependent measurements allowed the esti-
mation of heats of adsorption for n- and iso-butane from tran-
sient experiments and revealed comparable heats of adsorp-
tion as from calorimetry experiments. These findings will be
used to extend the adsorption and desorption model with an
appropriate reaction part. The correlation between TAP and
calorimetry is also very important regarding the transferabil-
ity of data from different investigation techniques to bridge
the existing pressure gap.
Acknowledgements
The DFG is gratefully acknowledged for support within
the Priority program 1091 “Bridging the gaps between ideal
and real systems in heterogeneous catalysis” with the subpro-
ject “Acid–base catalyzed alkane activation”. Thanks to Dr.
F.E. Jentoft (FHI Berlin) and Dr. B o¨ hm (University Leipzig)
for fruitful discussions.
The heats of adsorption of iso-butane differed more for
the active and the nonactive samples. Nevertheless, the dif-
ferences between the samples were small and the catalytic
activity may not be accounted on this behaviour alone. How-
ever, product adsorption and desorption seems to influence
the catalytic activity.
References
The finding that the adsorption step is so well reproduced
by the existing model will be the basis for further modelling
on this system which will include a detailed reaction part.
[1] T.A. Nijhuis, L.J.P. van den Broeke, M.J.G. Linders, M. Makkee, F.
Kapteijn, J.A. Moulijn, Catal. Today 53 (1999) 189.
[2] J.T. Gleaves, J.R. Ebner, T.C. Kuechler, Catal. Rev. Sci. Eng. 30
(1988) 49.
[
[
3] M. Rothaemel, M. Baern, Ind. Eng. Chem. Res. 35 (1996) 1556.
4] H.D. Do, D.D. Do, I. Praetyo, AIChe J. 47 (2001) 2515.
4
. Conclusions
[5] Y. Schuurmann, J.T. Gleaves, Catal. Today 33 (1997) 25.
[
[
[
6] M. Hino, S. Kobayashi, K. Arata, J. Am. Chem. Soc. 101 (1979)
6439.
Transient measurements with the TAP reactor were per-
7] E. Garcia, M.A. Volpe, M.L. Ferreira, E. Rueda, J. Mol. Catal.: A
Chem. 201 (2003) 263.
8] J.T. Gleaves, G.S. Yablonski, Y. Schuurman, Appl. Catal. A 160
(1997) 55.
formedforthefirsttimetocharacterizethesorptionbehaviour
of C4-hydrocarbons over sulfated zirconias of different cat-
alytic activity.

2
78
C. Breitkopf / Journal of Molecular Catalysis A: Chemical 226 (2005) 269–278
[9] J.F. Haw, In situ Spectroscopy in Heterogeneous Catalysis, Wi-
[13] P. Phanawadee, PhD Thesis, Washington University, 1997.
ley/VCH, Weinheim, 2002.
[14] Modelling software to TAP-2, Leipzig.
[
[
10] H.K. Mishra, K.M. Parida, Appl. Catal. A: Gen. 224 (2002) 179.
11] M. Standke, C. Breitkopf, H. Papp, S. Wrabetz, B.S. Klose, X. Yang,
F.E. Jentoft, X. Li, L.J. Simon, J.A. Lercher, Proceedings of the
DGMK-Conference, Berlin, October 9–11, 2002, ISBN 3-931850-
[15] M.R. Gonzalez, K.B. Fogash, J.M. Kobe, J.A. Dumesci, Catal. Today
33 (1997) 303.
[16] S. Brunauer, P.H. Emmett, E. Teller, J. Am. Chem. Soc. 73 (1938)
309.
9
8-6, pp. 23–30.
[17] E.P. Barret, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951)
374.
[
12] B.S. Klose. Diploma Thesis, TU M u¨ nchen, 2001.