Hydrogen effect on n-butane isomerization over sulfated zirconia-based catalysts

Article abstract of DOI:10.1006/jcat.1997.1595

Iron- and manganese-promoted sulfated zirconia (SFMZ) has been tested as an n-butane isomerization catalyst in the temperature range of 35 to 180°C. The catalytic activity exhibits an induction period whose length is dependent on the reaction conditions. The presence of H₂ in the feed stream strongly suppresses n-butane conversion over unpromoted sulfated zirconia (SZ) and over Pt-containing SFMZ (PtSFMZ). However, hydrogen had no effect on n-butane isomerization over SFMZ. These findings were interpreted on the basis of a bimolecular mechanism where unsaturated intermediates (carbenium ions and/or butene) are formed during the break-in period. The role of promoters (Fe and Mn) is not only facilitating the formation of hydrogen-deficient intermediates and their accumulation on the catalyst surface, but also enhancing their stability. The negative effect of hydrogen over PtSFMZ is attributed to the occurrence of atomic hydrogen via the dissociative adsorption of H₂ on Pt.

Full text of DOI:10.1006/jcat.1997.1595

JOURNAL OF CATALYSIS 167, 346–353 (1997)
ARTICLE NO. CA971595
Hydrogen Effect on n-Butane Isomerization over Sulfated
Zirconia-Based Catalysts
1
2
Abdelhamid Sayari, Yong Yang, and Xuemin Song
Department of Chemical Engineering and CERPIC, Universit e´ Laval, Ste-Foy, Qu e´ bec, Canada G1K 7P4
Received April 24, 1996; revised November 13, 1996; accepted December 20, 1996
molecules (e.g., benzene, acetonitrile, ammonia, pyridine)
Iron- and manganese-promoted sulfated zirconia (SFMZ) has during heating (6, 7, 11–13). By means of Fourier transform
been tested as an n-butane isomerization catalyst in the temperature infrared (FTIR) and nuclear magnetic resonance (NMR)
�
range of 35 to 180 C. The catalytic activity exhibits an induction
spectroscopies of adsorbed bases, Adeeva et al. (7) found
that both SZ and SFMZ are not superacids, and they have
similar acid strengths. This conclusion was also reached by
several other authors (8, 9, 12). It is important to mention
period whose length is dependent on the reaction conditions. The
presence of H2 in the feed stream strongly suppresses n-butane
conversion over unpromoted sulfated zirconia (SZ) and over Pt-
containing SFMZ (PtSFMZ). However, hydrogen had no effect on
that earlier UV–visible data have already indicated that SZ
n-butane isomerization over SFMZ. These findings were interpreted
on the basis of a bimolecular mechanism where unsaturated inter-
mediates (carbenium ions and/or butene) are formed during the
break-in period. The role of promoters (Fe and Mn) is not only
is not a superacid (14).
Adeeva et al. (7) explained the remarkable activity dif-
ference between SZ and SFMZ catalysts on the basis of
facilitating the formation of hydrogen-deficient intermediates and mechanistic considerations. These authors argued that the
their accumulation on the catalyst surface, but also enhancing their acidity of a catalyst is not the only factor that governs its
stability. The negative effect of hydrogen over PtSFMZ is attributed activity in acid-catalyzed reactions. The catalyst activity can
to the occurrence of atomic hydrogen via the dissociative adsorption
also be dependent on the nature of the intermediates and
of H2 on Pt.
�c 1997 Academic Press
the energetics associated with the reaction pathway. To ex-
plain the occurrence of disproportionation products from
�
n-butane isomerization over H-mordenite at about 350 C,
INTRODUCTION
Guisnet et al. (15–17) proposed that the reaction proceeds
via an intermolecular mechanism with the intermediacy of
a C8 species. The same mechanism was also suggested by
Gates and co-workers (18, 19) for n-butane isomerization
Sulfated zirconia-based materials are strong, noncorro-
sive solid acid catalysts targeted by many research groups as
potentially viable substitutes for currently used HF, H2SO4,
and chlorinated alumina catalysts. Recent progress in this
area has been reviewed (1, 2). Hsu et al. (3) found that at
low temperature, sulfated zirconia promoted with Fe and
Mn (referred to as SFMZ) is about three orders of magni-
tude more active in n-butane isomerization than the unpro-
moted sulfated zirconia (SZ) catalyst. Using temperature-
programmed desorption (TPD) of benzene, Lin and Hsu
�
over SFMZ at 40–225 C. Reaction studies by Adeeva et al.
(
20–22) using double 13C-labeled butane indicated that at
low temperature, n-butane isomerization over SFMZ, SZ,
and PtSZ catalysts proceeds predominantly via a bimolec-
ular mechanism involving (i) the formation of C4 carbe-
nium ion and butene, (ii) reaction of the carbenium ion with
butene to form a C8 intermediate, and (iii) rearrangement
and �-scission of the C8 intermediate to isobutane. This re-
action pathwayisenergeticallymore favorable compared to
a monomolecular n-butane isomerization mechanism (23).
The function of the promoters is still unclear. There is a
broad consensus now that Fe and Mn promoters, and pos-
sibly others such as Ni and Cr, do not increase the acid
strength of Brønsted and Lewis acid sites on the surface of
the catalyst. T a´ bora and Davis (8) found not only that SZ
and SFMZ catalysts exhibit similar FTIR spectra of sulfate
groups, adsorbed CO, and adsorbed pyridine, but, also that
EXAFS data indicate that the addition of promoters does
not alter the surface properties and the crystallographic
(
4) reported that SFMZ has more and stronger acid sites
than SZ which has long been considered as a superacid
5). The remarkable catalytic activity of SFMZ in n-butane
isomerization was later confirmed by several authors (6–
0). However, the validity of using TPD to characterize the
(
1
acid strength of SZ-based catalysts has been questioned
because of the oxidative decomposition of adsorbed probe
1
To whom all correspondence should be addressed. Fax: (418) 656
5
993. E-mail: sayari@gch.ulaval.ca.
2
Present address: University of Calgary, Department of Chemistry,
Calgary, AB, Canada T2N 1N4.
346
0021-9517/97 $25.00
Copyright �c 1997 by Academic Press
All rights of reproduction in any form reserved.

HYDROGEN EFFECT ON n-BUTANE ISOMERIZATION
347
phase of SZ. Furthermore since Ni (9) and Cr/Mn (24) ex- tion per gram of SFMZ catalyst. The resulting mixture was
�
hibit similar promoting effects as Fe/Mn, it is inferred that stirred at room temperature for 1 h and then dried at 110 C
they play the same role as Fe/Mn. Adeeva et al. (7, 20, 21) for several hours.
suggested that the role of Fe and Mn promoters is to facil-
n-Butane isomerization was carried out in a fixed bed
itate the reaction path by enhancing the formation of the downflow quartz reactor (10 mm i.d.) under atmospheric
unsaturated intermediates (butenes or carbenium ions) or pressure. The dried fresh catalyst (0.4–1.0 g) was loaded in
increasing their stability. Furthermore, Coelho et al. (9, 10) the reactor, and unless otherwise stated, it was pretreated in
�
stated that in addition to generating olefins, transition metal dry air (30 ml/min) at 650 C for 3 h, and then cooled to re-
�
promoters enhance the surface concentration of hydrogen- action temperature (35–180 C) in flowing He (30 ml/min).
deficient reaction intermediates. Several authors suggested The gas mixture consisted of n-butane and helium flow-
the occurrence of two (8) or even three (12) different active ing either at 0.5 and 30 ml/min, or at 1 and 4 ml/min, re-
sites for n-butane isomerization over SFMZ. Wan et al. (12) spectively. In some runs, helium was replaced by hydrogen.
�
proposed that at 30 C, SFMZ has a dual catalytic function n-Butane (CP grade), He (UHP), H2 (UHP), and com-
for n-butane isomerization, i.e., the promoters are redox- pressed air (industrial) used in this work were purchased
active and responsible for the oxidative dehydrogenation from Air Products, Inc.
of n-butane to butene which is then protonated over the
The reaction products were analyzed on-line using
acid sites, forming a carbenium ion. The isomerization pro- a HP-5890 II gas chromatograph (GC) equipped with
ceeds further through the formation, rearrangement, and a capillary column (crosslinked methyl silicone gum,
�
-scission of a C8 intermediate.
According to the bimolecular mechanism, the pres- tor. In experiments carried out at 100 C over SFMZ, the
50 m � 0.32 mm � 0.52 mm) and a flame ionization detec-
�
ence of hydrogen in the feedstock is expected to affect first analysis of the reactor effluent was undertaken after
n-butane conversion adversely since it inhibits the forma- 2 min on stream, and subsequent injections were made at
tion of hydrogen-deficient intermediate species (C4 carbe- 1-min intervals because of the rapid deactivation of the
nium ions and/or butenes). However, as reported briefly catalyst. Consequently, by-products (i.e., propane and pen-
in a previous paper (24), we did not observe any effect of tane) were most often not detected. In some runs, these
hydrogen on n-butane isomerization over SFMZ. We pro- by-products were measured, and their combined selectivity
posed that this is most likely due to the strong stabilization was found to be less than 5% . For reactions carried out at
�
of C4 carbenium ions by the promoters. The main objective temperature below 50 C, the first GC analysis was made af-
of this work is to shed further light on the role of the pro- ter 5 min on stream, and subsequent injections were made
moters by investigating the hydrogen effect over different every 5 min.
SZ-based catalysts.
RESULTS
EXPERIMENTAL
Figure 1 shows n-butane conversion versus time on
Zr(OH)4 was prepared as described earlier (24–26) by stream at different reaction temperatures. It is seen that in-
�
hydrolyzing a ZrOCl2 � 8H2O (98% , Aldrich) solution by creasing the reaction temperature in the range of 35–100 C
dropwise addition of aqueousammonia (28 wt% NH3, Ana- leads to a shorter induction period, a faster rate of deactiva-
chemina) under vigorous stirring at room temperature to tion, and a higher maximum rate of n-butane conversion.
pH 8. The precipitate was filtered and washed with deion- An apparent activation energy of 12.7 kcal/mol was de-
ized water until no chloride ions were detected by the termined from an Arrhenius plot based on the maximum
�
�
AgNO3 test. Then the solid was dried overnight at 110 C n-butane conversion over the temperature range 35–100 C
and crushed to below 100 mesh. SZ, SFZ, and SFMZ were (Fig. 1, insert). This value is in good agreement with the 11.4
prepared using the procedure of Jatia et al. (6); 0.15 M and 10.5 kcal/mol reported by Hsu et al. (3) and Wan et al.
Mn(NO3)2, 0.30 M Fe(NO3)3, and 3 wt% SO2 solutions (12), respectively.
�
4
were prepared by dissolving the appropriate amounts of
To investigate the contention that Fe and Mn promot-
Mn(NO3)2 � 6H2O (98% , Aldrich), Fe(NO3)3 � 9H2O (98% , ers enhance the surface concentration of butene (10), we
Aldrich), and (NH4)2SO4 (99% , Aldrich). Manganese, iron, undertook the following set of experiments. The first ex-
�
and sulfate were impregnated one at a time, in this order, periment was carried out at 35 C using a short residence
onto dried Zr(OH)4 by stirring a mixture of 1.5 ml per gram time over the same SFMZ catalyst as in Fig. 1 (n-butane =
of solid for 1 min and soaking for 14 min. After each im- 0.5 ml/min, He = 30 ml/min, SFMZ = 1 g). As shown in Fig.
pregnation, the solid was separated from the solution by 2 (curve A), under these conditions, the conversion of n-
filtration without washing and dried for several hours at butane was very low. In the second experiment, a longer res-
�
1
10 C. A PtSFMZ catalyst with 0.4 wt% Pt was prepared idence time was used (n-butane = 1 ml/min, He = 4 ml/min,
by incipient wetness using 1 ml of 1 wt% H2PtCl6 solu- SFMZ = 1 g), and a much higher n-butane conversion was

3
48
SAYARI, YANG, AND SONG
�
FIG. 1. n-Butane conversion over regenerated SFMZ vs time on
FIG. 3. n-Butane conversion over SFMZ vs time on stream at 35 C.
During the first 0.5 h on stream, the reaction conditions were the same
as in Fig. 2, curve B. The catalyst was purged by He (30 ml/min) for 0.5
h (᭹) or by H2 (30 ml/min) for 20 min followed by He (30 ml/min) for
stream at different reaction temperatures. Reaction conditions:1 g of cata-
�
lyst; n-C4 = 0.5 ml/min; He = 30 ml/min; reaction temperature: (A) 35 C,
� �
B) 50 C, (C) 100 C. Insert: Arrhenius plot.
(
1
0 min (� ). Then the reaction was resumed under the same conditions as
in Fig. 2, curve A.
observed (Fig. 2, curve B). In the third experiment (Fig.
2
, curves C1 and C2), the reaction was started under the
same conditions as for curve B. The n-butane conversion
thus obtained versus time on stream is shown by open cir-
cles in Fig. 2 (curve C1). At 0.4 h on stream, the reaction
was stopped and the catalyst was purged with helium (30
ml/min) for 0.5 h. Then the reaction was resumed under the
same conditions as for curve A. In this case, higher n-butane
conversions (curve C2) than those shown in curve A were
obtained.
Figure 3 summarizes two comparative experiments car-
ried out on a different batch of SFMZ catalyst and a differ-
ent experimental set-up. Data (full circles) were obtained
under the same pretreatment and reaction conditions as
those in Fig. 2, curves C1 and C2. The slight decrease in the
activity after the helium purge is probably related to analy-
sis timing during rapid deactivation. The other data (Fig. 3,
open circles) were obtained under identical conditions ex-
cept that the purge took place under flowing H2 (30 ml/min
for 20 min) followed by He (30 ml/min for 10 min).
Figure 4 shows the effect of hydrogen on n-butane iso-
�
FIG. 2. n-Butane conversion over SFMZ vs time on stream at 35 C. merization over the SZ catalyst. Similar pretreatment and
Reaction conditions:1gofcatalyst;(A) n-C4 = 0.5ml/min, He = 30ml/min;
reaction conditions as those reported by Adeeva et al. (21)
were used. The catalyst (1.5 g) was first calcined ex situ in
(
(
0
B) n-C4 = 1 ml/min, He = 4 ml/min. (C) During the first 0.4 h on stream
curve C1), the reaction conditions were the same as for curve (B). At
�
static air (620 C, 3 h), and then pretreated in situ in air
.4 h on stream the catalyst was purged by He (30 ml/min) for 0.5 h. Then
the reaction was resumed (curve C2) under the same conditions as for
curve A.
�
(
30 ml/min) at 450 C for 2 h and cooled to the reaction
temperature (180 C) under flowing He. The feed stream
�

HYDROGEN EFFECT ON n-BUTANE ISOMERIZATION
349
FIG. 5. n-Butane conversion rate over SZ vs time on stream at
�
FIG. 4. n-Butane conversion rate over SZ vs time at different H2/n-C4
100 C. Reaction conditions: 0.4 g of SZ calcined ex situ in static
�
�
�
ratio. Reaction conditions: 1.5 g of catalyst; reaction temperature, 180 C. air at 620 C for 3 h and in situ in flowing air at 500 C for 2 h;
H2/n-C4 (mol/mol): (A) 60, (B) 45, (C) 30, (D) 20, (E) 9, (F) 4, (G) 0.
(᭹) n-C4 = 0.5 ml/min, He = 30 ml/min; (� ) n-C4 = 0.5 ml/min, H2 =
Insert: Dependence of n-butane conversion on hydrogen partial pressure. 30 ml/min.
n-Butane conversion over Pt-containing SFMZ was
(
120 ml/min) consisted of n-C4 (0.75 ml/min), hydrogen,
strongly affected by hydrogen. Three reactions were car-
and helium. The flow rates of H2 and He were adjusted
according to the desired H2/He ratio. As seen in Fig. 4, at
high H2/n-C4 ratios, the catalyst activity was stable and the
conversion rate increased as more hydrogen was replaced
by helium. As the H2/n-C4 ratio reached 4 and then 0, the
deactivation rate increased dramatically so that data had
to be collected immediately after changing the feed com-
position in order to measure the initial (or intrinsic) ac-
tivity of the catalyst. It is shown clearly in Fig. 4 that the
rate of n-butane conversion increases significantly (from
�
�
ried out at 100 C over 0.4-g batches of PtSFMZ calcined
in situ at 650 C for 3 h using n-C4, He, and H2 mix-
tures flowing at 30.5 ml/min. The feed compositions were
n-C4 :He :H2 = 0.5 :30 :0, 0.5 :29 :1, and 0.5 :0 :30. The
�
8
� 7
� 1
ca. 2 � 10 to 1.5 � 10 mol/g cat. s ) with decreasing
partial pressure of hydrogen (from H2/n-C4 = 60 to 0). Ki-
netic analysis of the hydrogen effect (Fig. 4, insert) shows
that the reaction order in H2 as calculated from the slope of
the straight line ln(conv.% ) vs hydrogen pressure is � 1.2.
This result is in excellent agreement with that reported re-
cently (� 1.1 to � 1.3) by Liu et al. (22).
Figures 5–7 depict the effect of hydrogen on n-butane
isomerization over three different catalysts (SZ, SFZ, and
SFMZ) under the same reaction conditions (0.4 g of cata-
lyst; 0.5 ml/min of n-butane; 30 ml/min of He or H2; at
�
1
00 C). Again, a strong negative effect of hydrogen was
observed over SZ (Fig. 5). In addition, isomerization of bu-
tane in helium exhibits an induction period followed by
deactivation. When Fe was added to SZ, the hydrogen ef-
fect became very weak (Fig. 6). Further, when both Fe and
Mn promoters were present on SZ, the hydrogen effect dis-
appeared (Fig. 7).
FIG. 6. n-Butane conversion rate over SFZ vs time on stream at
00 C; 0.4 g of SFZ calcined ex situ at 600 C for 16 h. All other condi-
tions are the same as in Fig. 5.
�
�
1

3
50
SAYARI, YANG, AND SONG
In this mechanism, the formation of butene takes place
most likely through a redox process (Eqs. [1] and [2]). The
butene thus formed can react with a C4 carbenium ion to
form a C8 carbenium ion (Eq. [3]) which then undergoes
skeletal rearrangement (Eq. [4]) and �-scission (Eq. [5]) to
generate a tertiary C4 carbenium ion. Finally, isobutane is
+
formed as a result of hydride transfer between the t-C and
4
a reactant molecule (Eq. [6]). Compared to a monomolec-
ular n-butane isomerization process, the bimolecular path-
way avoids the formation of primary carbenium ions, and
thus requires a lower activation energy to initiate the reac-
tion. Experimental evidence for such a mechanism is based
mainly on the binomial distribution of 13C in the isomer-
ization products of double-labeled n-butane (20–22), the
negative effect of H2 observed over SZ-based catalysts (10,
2
1, 22), and also the formation of C8 disproportionation
products (30).
Based on the above-mentioned investigations, in the
following discussion of our results, we concur that n-butane
isomerization at low temperature over SZ-based catalysts
takes place via the bimolecular mechanism. Moreover, on
the basis of our experimental data on hydrogen effect, we
believe, as proposed by Coelho et al. (8, 10), that the role
of the promoters is not only to facilitate the formation of
FIG. 7. n-Butane conversion rate over SFMZ vs time on stream at
00 C; 0.4 g of SFMZ calcined ex situ at 600 C for 16 h. All other conditions
are the same as in Fig. 5.
�
�
1
�
7
� 7
maximum conversion rates were 1.4 � 10 , 0.3 � 10 , and
+
unsaturated intermediates (C4H9 and butene), but also
�
8
� 1
0
.2 � 10 mol/g cat. s . The induction period in the two
to stabilize them on the catalyst surface. As a result of
this stabilization effect, the surface concentration of the
unsaturated species may be increased.
first experiments was ca. 5 min.
The bimolecular mechanism is consistent with many ex-
perimental observations associated with low-temperature
n-butane conversion over SZ-based catalysts. For exam-
ple, a common feature reported in the literature dealing
with n-butane isomerization over SFMZ under inert atmo-
sphere is the occurrence of an induction period of increas-
ing conversion, followed by deactivation. The induction
period corresponds to the time needed for the formation
and accumulation of unsaturated intermediates (butenes).
The activity decrease following the induction period is as-
cribed to the formation of coke precursors. Deactivation
by coke deposition was suggested by several workers (25,
DISCUSSION
Reaction Mechanism
There are essentially two proposed mechanisms for
n-butane isomerization over acid catalysts (23, 27). The first
one is a monomolecular mechanism involving abstraction
of a hydride followed by isomerization of the secondary car-
benium ion into a tertiary carbenium ion via methyl shift,
hydride transfer, and isobutane desorption. This mecha-
nism was reported by at least two research groups (28, 29)
to occur in the presence of PtSZ. However, the majority
of authors (9, 10, 12, 18–22, 24) have proposed that low-
temperature n-butane conversion over SZ-based catalysts
proceeds through a bimolecular mechanism. This mecha-
nism can be represented as follows:
3
1) and confirmed recently by ESR and by direct observa-
tion by in situ Raman and UV–visible diffuse reflectance
spectroscopies (32).
Based on the above reasoning, the effect of reaction tem-
perature on the length of the induction period and the rate
[1] of deactivation (Fig. 1) can be rationalized as follows. The
formation of C4 carbenium ions and butenes is acceler-
nC4H10 + SO4 *) C4H9 + OH� + SO 32 �
2
�
+
C4H *) C4H8 + H+
+
[2]
9
ated at higher temperature. The impact of the enhanced
formation of butenes on the catalyst surface is twofold.
On the one hand, it favors the bimolecular process, and
therefore increases the conversion of n-butane. This leads
C4H8 + C4H+9 *) C8H+
[3]
17
+
+
C8H *) t-C8H
[4]
1
7
17
t-C8H17 *) C4H8 + t-C4H+
+
to a shorter induction period and a higher maximum con-
[5]
9
version at higher reaction temperature (Fig. 1). On the
[6] other hand, the increased butene concentration enhances
t-C4H9 + n-C4H10 *) i-C4H10 + C4H9+
+

HYDROGEN EFFECT ON n-BUTANE ISOMERIZATION
351
the polymerization of the olefins, leading to a more rapid while the conversion was still increasing (curve C1). It was
formation of coke precursors and catalyst deactivation. assumed that during this period butene forms and accumu-
Obviously, butene plays an important role in the bi- lated on the catalyst surface. Then, helium (30 ml/min) was
molecular process. Liu et al. (22) found that the n-butane introduced through the catalyst bed for 0.5 h, and the re-
conversion rate over a PtSZ catalyst decreased when the action was resumed under the same short residence time
feed which contained 33 ppm butene was preequilibrated as in curve A. As shown in curve C2, a much higher con-
in hydrogen over a Pt/SiO2 catalyst. Coelho et al. (10) found version than under conditions (A) was observed. Clearly,
that preadsorption of a small amount of 1-butene on the this conversion difference is due to the presence of a higher
catalyst resulted in a shorter induction period, a higher ac- surface concentration of butenes generated during the first
tivity, and a faster deactivation.
reaction period represented by curve C1. These butenes
As mentioned in the Introduction, the role of the promot- remained on the SFMZ catalyst surface even after a 0.5-h
�
ers is still unclear. Several authors proposed that the pro- purge with helium at 35 C. In agreement with Coelho et al.
moters help generate butenes (7, 12, 20–22). Coelho et al. (9, 10), it is inferred that in the presence of SFMZ, butene
(
10) argued that if the only role of the promoters is the for- is strongly stabilized on the catalyst surface. Moreover, as
mation of butene, addition of 1-butene in the feed should shown in Fig. 3, if the 30-min He purge were replaced by a
generate an activity level over SZ catalyst comparable to 20-min H2 treatment followed by a 10-min He purge at the
�
that observed over Fe- and Mn-promoted SZ. Since this same temperature (35 C), the activity obtained when the
was not the case they suggested that the role of the pro- reaction is resumed at high residence time is significantly
moters is not only to generate butene, but also to enhance lower. This indicates that the accumulated intermediates
its concentration on the catalyst surface. To further exam- are not stable in a pure hydrogen atmosphere, most likely
ine the ability of the promoters to retain butene (and other because of their hydrogen-deficient character. In summary,
hydrogen-deficient species) on the catalyst surface, we de- these simple experiments provide convincing evidence of
signed the experiments shown in Fig. 2. In the experiment the occurrence of hydrogen-deficient intermediates species
represented by curve A, low n-butane conversions were and the ability of the promoters to stabilize them.
observed because the residence time of the feed over the
Direct comparison of this behavior with SZ was unfor-
catalyst was short. With a longer residence time, the conver- tunately not possible because on the one hand SZ is not
�
sion increased significantly (curve B). It is inferred that at active at 35 C, and on the other hand if comparison is made
the same time on stream the amount of butene accumulated at a temperature where SZ exhibits a reasonable activity,
�
under the conditions in curve B is higher than that accumu- e.g., 150 C, the reaction intermediates over SFMZ are no
�
lated under the conditions in curve A. In the third experi- longer stable under He. Moreover, at 150 C the reaction
ment (curves C1 and C2), we started the reaction using the over SFMZ becomes extremely fast and the catalyst is de-
same long residence time as in curve B, and then stopped it activated very quickly (Table 1).
TABLE 1
n-Butane Isomerization over SFMZ under Different Conditions
Reaction
conditions
Reaction
temp. ( C)
Maximum
conv. (% )
Time
(min)
�
a
No.
1
Pretreatment
�
b
Air, 650 C, 3 h
1 g cat., 0.5 ml/min n-C4,
35
1.2
48
3
0 ml/min He
or H2
Same
2
3
4
Same
Same
Air, 600 C, 16 h ; air,
50
100
100
3.4
29.5
18
19
3.5
3.2
Same
�
c
d
0.4 g cat., 10.9 ml/min
n-C4, 20.6 ml/min
�
b,d
6
00 C, 2 h
d
He or H2
�
�
b,e
c
5
6
7
Air, 500 C, 2 h
Air, 620 C, 3 h
Same as experiment (4)
Same
Same
Same
100
100
180
14
35
23.8f
4.1
3.0
t.s.
g
a
b
c
d
e
f
Time for maximum conversion.
In situ treatment.
Ex situ treatment.
Same conditions as in Ref. (10).
Regenerated catalyst.
First data point.
g
Too short.

3
52
SAYARI, YANG, AND SONG
Hydrogen Effect
the hydrogenation reactor and then through the isomeriza-
tion reactor. The data obtained under these conditions were
According to the bimolecular mechanism of n-butane the same as those found without using the hydrogenation
isomerization, hydrogen is supposed to display a nega- reactor. This confirms that no negative effect of hydrogen
tive effect on the reaction rate, because the formation occurs over our SFMZ catalysts.
of hydrogen-deficient species is obviously inhibited in the
The absence of H2 effect on n-butane isomerization
presence of hydrogen. This is clearly demonstrated in the over SFMZ catalyst is not necessarily inconsistent with the
case of SZ catalyst (Fig. 4). Several authors (21, 23, 32– occurrence of a bimolecular mechanism. A reasonable
3
4) have also observed that hydrogen suppresses the ini- explanation rests on the strong stabilization effect of Fe
tial (intrinsic) n-butane conversion over SZ, although it in- and Mn promoters on unsaturated intermediates (carbe-
creases the catalyst lifetime by reducing the accumulation nium ions and butenes) on the catalyst surface, as already
of carbonaceous species on the catalyst surface. As shown inferred from Figs. 2 and 3. The surface concentration of
in Fig. 4, the deactivation rate increases as the H2/n-C4 ra- the unsaturated species depends not only on the partial
tio decreases, and it was necessary to use the initial activity pressure of hydrogen in the gas phase, but also on the
data in order to have a satisfactory Arrhenius plot. In some ability of the catalyst to retain them on its surface. The
of the earlier investigations (22, 35), H2 was found to have a more stable the species on the surface, the weaker the effect
positive effect on n-butane isomerization over SZ and PtSZ. of H2 on the rate of their formation or transformation. It is
However, this conclusion was drawn from steady state data understood that if the stability of the reaction intermediate
or from data obtained after more than 30 min on stream. is high enough, the limited range of hydrogen pressure
In our opinion, it reflects a strong decrease in the rate of explored here may not be sufficient for the catalyst to
deactivation in H2 rather than a higher intrinsic activity of experience a measurable hydrogen effect. It is inferred that
+
the catalyst.
(i) C4H9 and butene intermediates are held more strongly
Figure 5 shows the n-butane conversion rate over SZ dur- on the surface of SFMZ than on the SZ surface, and (ii)
ing the early stages of the reaction. A clear induction period the strong stabilizing effect of Fe and Mn promoters on the
was observed when the reaction was carried out in helium, reaction intermediates is at the origin of the remarkable
which is consistent with the occurrence of a bimolecular increase in butane conversion over SFMZ compared to its
reaction mechanism. When the reaction was performed unpromoted SZ counterpart. Lee et al. (36–38) studied the
in hydrogen, the conversion was dramatically suppressed. interactions of sulfated iron oxide with pyridine, ammonia,
However, it is interesting to observe that in the presence 1-butene, and butane by TPD and mass spectrometry. They
of Fe-promoted catalyst, the negative effect of hydrogen found that contrary to butane, the other three molecules ad-
became much weaker (Fig. 6). Furthermore, over SFMZ sorb strongly on the catalyst by direct interaction between
catalyst the negative effect of hydrogen disappeared com- the nitrogen electron pair or the �-bond with coordina-
pletely (Fig. 7). At variance with this finding, Coelho et al. tively unsaturated iron. A similar stabilization process may
(
10) recently reported a strong negative effect of hydrogen occur during n-butane isomerization at low temperature
on n-butane conversion over SFMZ catalysts.
over iron-containing catalysts.
In order to make sure that the data shown in Fig. 7
Although no hydrogen effect was observed on SFMZ,
are representative of the behavior of our SFMZ catalyst, we did observe a strongly negative effect of H2 over Pt con-
we compared n-butane isomerization in helium and in hy- taining SFMZ catalysts. Furthermore, we found that the
�
drogen over a wide range of temperature (35 to 180 C), presence of even a small amount of hydrogen (3.3 mol% )
feed compositions (H2/n-C4 = 2 to 60), and flow rates (5 to in the feed causes a sharp decrease in n-butane conversion
3
1.5 ml/min). Table 1 provides typical data obtained un- rate. In our previous reports we found that PtSZ catalyst
�
der different pretreatment and reaction conditions. In all calcined at 600 C exhibits metallic Pt (26, 39), and cata-
cases, no hydrogen effect was observed. Even though no lyzes propene hydrogenation at 35 C (24). This indicates
�
butene impurity was reported by the supplier of the bu- that under these conditions PtSZ is capable of adsorbing
tane used in this study, we considered the possibility that H2 dissociatively. It is inferred that this may also be the
trace amounts of butene present in the feed participate in case for calcined PtSFMZ. Furthermore, because of the in-
the dimerization-cracking process (Eqs. [3]–[6]), leading to timate contact between the Pt particles and the promoter
an apparent lack of hydrogen effect. In one experiment, a species on the catalyst surface, it is reasonable to assume
hydrogenation reactor was placed upstream of the isomer- that the highly active atomic hydrogen will react readily on
ization reactor. The hydrogenation reactor contained 0.5 g the catalyst surface with butenes. Under these conditions,
of 5 wt% Pt/Al2O3 catalyst (Strem Chemicals, Inc.) which hydrogenation of butenes will prevail over their associa-
�
+
was prereduced in situ in flowing H2 (30 ml/min) at 300 C tion with C4H9 to form a C8 carbenium ion. As a result,
for 1 h and then cooled to room temperature. The feed the bimolecular pathway is blocked, and the n-butane con-
stream consisting of n-butane and H2 flowed first through version decreases. It should be noticed that the occurrence

HYDROGEN EFFECT ON n-BUTANE ISOMERIZATION
353
of hydrogen spillover over prereduced PtSZ has already
been documented (35, 40). In this case, higher tempera-
tures were needed because the catalyst was prereduced and
consequently Pt was poisoned by sulfur during reduction.
van de Ven, L. J. M., Sachtler, W. M. H., and van Santen, R. A.,
J. Catal. 151, 364 (1995).
. T a´ bora, J. E., and Davis, R. J., J. Chem. Soc. Faraday Trans. 91, 1825
8
9
(
1995).
. Coelho, M. A., Resasco, D. E., Sikabwe, E. C., and White, R. L., Catal.
Lett. 32, 253 (1995).
CONCLUSION
10. Coelho, M. A., Alvarez, W. E., Sikabwe, E. C., White, R. L., and
Resasco, D. E., Catal. Today (1996), in press.
1
1. Sikabwe, E. C., Coelho, M. A., Resasco, D. E., and White, R. L., Catal.
Lett. 34, 23 (1995).
Based on the fact that n-butane isomerization over sul-
fated zirconia and related catalysts takes place via an in-
termolecular mechanism, the following conclusions were
reached. The induction period of the catalytic activity cor-
1
2. Wan, K. T., Khouw, C. B., and Davis, M. E., J. Catal. 158, 311 (1996).
3. Srinivasan, R., Keogh, R. A., Ghenciu, A., F a˘ rca s¸ iu, D., and Davis,
B. H., J. Catal. 158, 502 (1996).
1
responds to the accumulation of unsaturated intermediates 14. Umansky, B., Engelhardt, J., and Hall, W. K., J. Catal. 127, 128 (1991).
1
5. Guisnet, M., Avendano, F., Bearez, C., and Chevalier, F., J. Chem. Soc.
Chem. Commun. 336 (1985).
(
butenes and carbenium ions). The length of the induction
period is dependent on the reaction conditions (temper-
ature, flow rate, and feed composition). The negative re-
action order in hydrogen over SZ observed in this work
1
1
6. Guisnet, M. R., Acc. Chem. Res. 23, 392 (1990).
7. Bearez, C., Chevalier, F., and Guisnet, M., React. Kinet. Catal. Lett.
22, 405 (1983).
and others is consistent with the intermolecular mechanism. 18. Zarkalis, A. S., Hsu, C.-Y., and Gates, B. C., Catal. Lett. 29, 235 (1994).
1
2
9. Cheung, T. K., D’Itri, J. L., and Gates, B. C., J. Catal. 151, 464 (1995).
0. Adeeva, V., Lei, G. D., and Sachtler, W. M. H., Appl. Catal. A 118, L11
The addition of promoters (Fe and Mn) to SZ dramati-
cally decreases the negative effect of hydrogen. This find-
ing indicates that transition metal oxides not only favor the
formation of olefinic intermediates but also stabilize them
(
1994).
1. Adeeva, V., Lei, G. D., and Sachtler, W. M. H., Catal. Lett. 33, 135
1995).
2
(
strongly. The strong negative effect of H2 on n-butane iso- 22. Liu, H., Adeeva, V., Lei, G. D., and Sachtler, W. M. H., J. Mol. Catal.
1
00, 35 (1995).
merization over PtSFMZ was attributed to the dissociative
adsorption of hydrogen and the fast hydrogenation of un-
saturated intermediates by atomic hydrogen.
2
3. Garin, F., Andriamasinoro, D., Abdulsamad, A., and Sommer, J.,
J. Catal. 131, 199 (1991).
4. Song, X., Reddy, K. R., and Sayari, A., J. Catal. 161, 206 (1996).
5. Song, X., and Sayari, A., Appl. Catal. A 110, 121 (1994).
2
2
ACKNOWLEDGMENTS
26. Dicko, A., Song, X., Adnot, A., and Sayari, A., J. Catal. 150, 254 (1994).
27. Asuquo, R. A., Eder-Mirth, G., and Lercher, J. A., J. Catal. 155, 376
This work was supported by the Natural Sciences and Engineering Re-
search Council of Canada (NSERC) and by le Fonds pour la Formation
de Chercheurs et l’Aide a` la Recherche (FCAR, Qu e´ bec). We thank Pro-
fessor D. E. Resasco for a preprint of his work.
(1995).
28. Chao, K.-J., Wu, H.-C., and Leu, L.-J., J. Catal. 157, 289 (1995).
29. Garin, F., Seyfried, L., Girard, P., Maire, G., Abdulsamad, A., and
Sommer, J., J. Catal. 151, 26 (1995).
3
3
0. Zarkalis, A. S., Hsu, C.-Y., and Gates, B. C., Catal. Lett. 37, 1 (1996).
1. Tanabe, K., and Yamaguchi, Y., Stud. Surf. Sci. Catal. 44, 99 (1989).
REFERENCES
32. Spielbauer, D., Mekhemer, G. A. H., Bosch, E., and Kn o¨ zinger, H.,
Catal. Lett. 36, 59 (1996).
1
2
3
. Song, X., and Sayari, A., Catal. Rev. Sci. Eng. 38, 329 (1996).
. Song, X., and Sayari, A., CHEMTECH Aug. 27 (1995).
. Hsu, C.-Y., Heimbuch, C. R., Armes, C. T., and Gates, B. C., J. Chem. 34. Yaluris, G., Larson, R. B., Kobe, J. M., Gonz a´ lez, M. R., Flogash, K. B.,
Soc. Chem. Commun. 1645 (1992). and Dumesic, J. A., J. Catal. 158, 336 (1996).
. Lin, C.-H., and Hsu, C.-Y., J. Chem. Soc. Chem. Commun. 1479 35. Ebitani, K., Konishi, J., and Hattori, H., J. Catal. 130, 257 (1991).
33. Chen, F. R., Coudurier, G., Joly, J.-F., and V e´ drine, J. C., J. Catal. 143,
616 (1993).
4
(
1992).
36. Lee, J. S., and Park, D. S., J. Catal. 120, 46 (1989).
5
6
. Arata, K., Adv. Catal. 37, 165 (1990).
37. Lee, J. S., Yeom, M. H., and Park, D. S., J. Catal. 126, 361 (1990).
. Jatia, A., Chang, C., MacLeod, J. D., Okubo, T., and Davis, M. E., 38. Park, D. S., and Lee, J. S., React. Kinet. Catal. Lett. 40, 101 (1989).
Catal. Lett. 25, 21 (1994). 39. Sayari, A., and Dicko, A., J. Catal. 145, 561 (1994).
. Adeeva, V., de Haan, J. W., J a¨ nchen, J., Lei, G. D., Sch u¨ nemann, V., 40. Ebitani, K., Tsuji, J., Hattori, H., and Kita, H., J. Catal. 135, 609 (1992).
7