Enhanced activity and selectivity in n-octane isomerization using a bifunctional SCILL catalyst

Article abstract of DOI:10.1016/j.jcat.2012.05.008

Bifunctional Solid Catalyst with Ionic Liquid Layer (SCILL) systems are presented and applied in the skeleton isomerization of n-octane in a slurry-phase reaction mode. It is demonstrated that Pt on silica, coated with a thin film of acidic chloroaluminate ionic liquid exhibits remarkable catalytic performance under mild conditions in the presence of hydrogen. Interestingly, both selectivity and activity of n-octane isomerization increase in these systems as a function of hydrogen pressure. This does not only suggest a hydrogenation activity of the catalytic Pt-centers embedded in the strongly Lewis acidic ionic liquid but also a significant increase in the proton acidity in this system as a function of the hydrogen pressure.

Full text of DOI:10.1016/j.jcat.2012.05.008

Journal of Catalysis 292 (2012) 157–165
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Enhanced activity and selectivity in n-octane isomerization using
a bifunctional SCILL catalyst
⇑
C. Meyer, V. Hager, W. Schwieger, P. Wasserscheid
Lehrstuhl für Chemische Reaktionstechnik, Friedrich-Alexander Universität Erlangen-Nürnberg, Egerlandstr. 3, 91058 Erlangen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Bifunctional Solid Catalyst with Ionic Liquid Layer (SCILL) systems are presented and applied in the skele-
ton isomerization of n-octane in a slurry-phase reaction mode. It is demonstrated that Pt on silica, coated
with a thin film of acidic chloroaluminate ionic liquid exhibits remarkable catalytic performance under
mild conditions in the presence of hydrogen. Interestingly, both selectivity and activity of n-octane isom-
erization increase in these systems as a function of hydrogen pressure. This does not only suggest a
hydrogenation activity of the catalytic Pt-centers embedded in the strongly Lewis acidic ionic liquid
but also a significant increase in the proton acidity in this system as a function of the hydrogen pressure.
Ó 2012 Elsevier Inc. All rights reserved.
Received 10 March 2012
Revised 7 May 2012
Accepted 9 May 2012
Available online 28 June 2012
Keywords:
Isomerization
n-Octane
Ionic liquid
Acidity
SCILL
Bifunctional
Hydrogen
1. Introduction
As known from previous literature, the isomerization of C7+ alkanes
is very different from the isomerization of light paraffins (e.g.,
Isomerization of alkanes is an important catalytic process in
refineries to improve the Research Octane Number (RON) of the
produced gasoline [1]. It is the state-of-the-art in industry to isom-
erize heterogeneously light naphtha streams with predominantly
C5/C6 alkanes using bifunctional catalysts, such as e.g. platinum
on chlorinated alumina, metal oxide, or zeolite. Carbon deposits
on the catalyst can be reduced if these reactions are operated
under hydrogen atmosphere. The chlorinated alumina-based cata-
lysts are active below 473 K. A strictly dry and oxygen-free feed as
well as a continuous addition of small amounts of organic chlorides
are required to maintain the catalyst activity in this process. In an
alternative system, metal oxide-based catalysts are applied for the
isomerization of light naphtha applying temperatures between 453
and 483 K. These systems are not deactivated irreversibly by water
or oxygen but are not as robust and tolerant to coordinating
poisons in the feedstock compared to isomerization catalysts
containing zeolites. However, the relevant processes using these
robust zeolite catalysts are run at significantly higher reaction
temperatures (523–643 K) [2].
n-C4–n-C6) as cracking becomes increasingly competitive to isom-
erization as the carbon chain length increases [3]. Moreover, it
would be desirable to decrease the reaction temperature of the
isomerization as the equilibria to the desired highly branched al-
kanes (leading to the highest RONs) become more favorable with
lower temperature [4] and the tendency for coke formation is
greatly reduced.
Important examples of recent research achievements in the het-
erogeneously catalyzed hydroisomerization/hydrocracking are
briefly presented in the following. The selected examples are
grouped according to the different catalyst types applied and they
deal with n-heptane and n-octane as model feedstock to allow di-
rect comparison with our results.
The hydroisomerization of n-octane with the hierarchical mes-
oporous zeolite Pt/ZSM-22 yielded 70% monobranched isomers, 5%
dibranched, and ca. 5% cracking products at 80% conversion
(T = 573 K, p = 50 bar, pH2 = 5 bar, W/F = 14–90 kg s molꢀ1) [5]. Li-
quid-phase n-octane hydroisomerization in a stirred semi-batch
microautoclave was investigated over beta, USY, and mordenite
zeolites loaded with 1 wt.% Pt. Best results regarding multi-
branched isomer selectivity (Smultibranched = 36.6 mol% and
Smonobranched = 55.7 mol%) were achieved with PtMOR (T = 543 K,
It would be desirable to extend the isomerization reaction to
heavier alkanes (larger than C6) provided that high isomerization
yields with minimum losses by hydrocracking could be achieved.
p = 90 bar, conversion = 77.4%, reaction time = 10.5 h, 7 gzeolite
/
moln-C8) [6]. Hydroisomerization of n-octane was studied kineti-
cally by Chica and Corma using different zeolitic structures with
⇑
Corresponding author.
E-mail address: wasserscheid@crt.cbi.uni-erlangen.de (P. Wasserscheid).
0021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.05.008

158
C. Meyer et al. / Journal of Catalysis 292 (2012) 157–165
large pore sizes (12 MR) [7]. It is evident that at temperatures at
which zeolites show catalytic activity, the cracking of mono- and
specially dibranched isomers is faster than the isomerization of
mono- to dibranched products. With the experimental conditions
of this study (T = 493–723 K, ptotal = 20 bar, H2/n-octane = 10,
WHSV = 5.13 hꢀ1 referred to n-octane), the thermodynamic equi-
librium of monobranched products was obtained, but this was
not the case for the dibranched products due to fast cracking.
The rate of cracking of dibranched products equaled almost the
isomerization of mono- to dibranched products. Consequently,
high levels of di- or tribranched C8 alkanes could only be reached
by introducing a distillation and recycling unit or by using mem-
brane reactors. Weitkamp et al. elucidated in model compound
studies the possibility of isomerizing the C7 cut obtained from
atmospheric distillation of crude oil as an alternative to its aroma-
tization [8]. The zeolites 0.27 Pd/LaNaY-72, 0.27 Pd/HMCM-41, and
0.27 Pd/HBeta were used for the hydroisomerization of n-heptane
(pH2 = 10 bar, pn-heptane = 0.10 bar, W/Fn-heptane = 400 g h molꢀ1 with
W = dry mass of catalyst). Isomerization of n-heptane was found to
be possible up to medium conversions without considerable
hydrocracking. However, the RON of the resulting isomerization
products was much too low (37–43).
product mixture included alkanes ranging from C1–C8 and small
amounts of aromatics.
Another catalyst system for alkane isomerization to mention is
the heteropolyacid H3PW12O40 supported on ZrO2, SiO2, or carbon
[15]. These catalyst systems were investigated in the hydrocrack-
ing and hydroisomerization of n-octane to gain fuel-grade gasoline
(C4–C7) (T = 573 K, p = 1 bar, WHSV = 1 hꢀ1, H2/n-C8 = 6). The sys-
tems were positively influenced by the addition of Pt with regard
to their activity and stability.
The interest in running the alkane isomerization reaction at
very low temperatures has triggered attempts in the past to apply
highly acidic chloroaluminate ionic liquids (IL) as catalysts for this
reaction. Our group demonstrated that catalyst systems of the
general type [cation]Cl/AlCl3 (xAlCl3 > 0.5) + H2SO4 are very effective
in n-octane isomerization at temperatures as low as room temper-
ature [16]. Interestingly, these systems operate in a liquid–liquid
biphasic reaction mode allowing for a straight forward catalyst
separation and recycling. Joni and coworkers developed
a
Supported Ionic Liquid Phase (SILP) catalyst based on these acidic
chloroaluminate ionic liquids. These authors showed the impor-
tance of support pretreatment if acidic chloroaluminate ionic liq-
uids are to be immobilized on high-surface silica materials and
the high level of acidity is to be maintained [17].
Sulfated zirconia was tested in the n-octane hydroisomerization
by Busto and coworkers [9]. This study aimed for a maximum
liquid product yield (C5–C8), which was obtained at 18.5% conver-
sion for the Pt/SZ catalyst (T = 573 K, p = 15 bar, WHSV = 4 hꢀ1, mo-
lar ratio H2/n-C8 = 6, time-on-stream = 30 min). Unfortunately,
selectivities to mono- and multibranched alkane isomers within
the liquid products were not reported in this publication.
In the present work, we apply for the first time the Solid Catalyst
with Ionic Liquid Layer (SCILL) catalyst concept to realize a bifunc-
tional catalytic system for the alkane hydroisomerization reaction.
In detail, we use Pt on silica (Pt/silica) coated with a highly acidic
chloroaluminate ionic liquid ([C4C1Im]Cl/AlCl3 = 1/2) for the
isomerization of n-octane in the presence of hydrogen. The
reaction is carried out under moderate temperature conditions
(Treaction = 373–423 K) and hydrogen pressures of up to 40 bar in
a slurry-phase reaction mode (liquid organic phase and solid SCILL
catalyst). The beneficial interaction of hydrogen, platinum, and
acidic ionic liquid is highlighted by comparing the new bifunc-
tional SCILL catalyst in the presence of hydrogen with the same
catalyst in the absence of hydrogen and with the Pt-free, acidic
ionic liquid on silica under hydrogen pressure.
Silica-supported tungsten zirconia catalysts were tested in the
simultaneous hydroisomerization and hydrocracking of n-octane
(T = 573 K, p = 1 bar, WHSV = 1 hꢀ1
, ) for
H2/n-C8 = 6 mol molꢀ1
the production of C4–C8 products with high octane number [10].
The catalysts applied in this study deactivated rapidly in the ab-
sence of Pt, which is required to hydrogenate coke precursors.
Pt(0,5%)W7.5Z1.0Si performed best in n-octane isomerization with
the following selectivities within the iso-octanes: Smonobranched
=
69.5%, Sdibranched = 23.3%, and Stribranched = 7.2%. Multicomponent
nanocomposite catalysts of the type Pt/WO3/M-ZrOx were also
tested in the n-heptane hydroisomerization (398–473 K) [11]. At
398 K under trickle bed conditions, 30% n-C7 was converted with
an isomerization selectivity of more than 99% using a Pt/WO3/
Al-ZrOx catalyst (p = 8 bar, H2/n-C7 = 1, volume hour space
velocity = 1 hꢀ1). Bifunctional Pt/WOx–ZrO2 (Pt/WZr, 12.7 wt.% W)
and Pt/Beta (Si/Al = 12) catalysts were studied for the simultaneous
n-alkane hydroisomerization and aromatic hydrogenation using a
n-heptane/benzene feed mixture (25 wt.% benzene) at 30 bar,
WHSV = 3.1 hꢀ1 and H2/hydrocarbon = 10 mol molꢀ1 by Arribas
and coworkers [12]. The Pt/WOx–ZrO2 (0.6% Pt) yielded 51.7%
iso-C7 at 493 K. The obtained heptane isomers contained 30.8% of
di-and tri-branched and 69.2% of monobranched products. The
experiments with Pt/Beta (1% Pt) resulted in a 49.9% iso-C7 yield
at 533 K with selectivities of 27.5% for the multibranched heptanes
and 72.5% for the monobranched heptanes.
For monofunctional catalysts, the SCILL concept has been al-
ready successfully demonstrated in a number of studies, mostly
for hydrogenation reactions [18–24]. In SCILL catalysis,
a
traditional heterogeneous catalyst is coated with a thin film of io-
nic liquid to modify the catalyst’s activity and/or selectivity. In par-
ticular, significant effects on selectivity have been reported. These
effects have been explained by either differential solubilities of the
reactants in the ionic liquid layer leading to modified ratios of feed-
stock and product molecules at the catalytic center or by direct
chemical interactions of the IL coating with the active sites of the
catalyst. This latter option has been supported by surface science
and model catalysis studies revealing distinct interactions of an
imidazolium [NTf2]ꢀ ionic liquid with different catalytic sites of
Pd or Pt on alumina [25].
2. Experimental
In a quite different approach, Ohno et al. studied the catalytic
properties of H2-reduced HxMoO3 in alkane isomerization in com-
parison with those of H2-reduced MoO3 and 0.01 wt.% Pt/MoO3
[13]. The latter had the highest activity in n-heptane hydroisomer-
ization (96.7% isomerization selectivity at 57.2% conversion; 76.7%
monobranched, 19.1% dibranched, and 4.2% tribranched heptanes)
at 523 K (atmospheric pressure, molar ratio H2/C7 = 10, flow rate of
C7 0.02 mol hꢀ1). n-octane hydroisomerization was also carried out
using the metal–acid bifunctional catalyst MoO2ꢀx(OH)y [14]. This
system achieved 73.5% isomerization selectivity at 86.3% conver-
sion, with 50.8% monobranched and 22.7% multibranched isooc-
tanes (T = 623 K, pH2 = 5 bar, LHSV = 0.8 hꢀ1, H2/n-C8 = 30.2). The
All steps in the preparation of the catalysts were carried out un-
der inert conditions in a PlexiglasÒ glovebox (GS GLOVEBOX Sys-
temtechnik GmbH) using argon (Linde AG, 4.6) as inert gas.
2.1. Catalyst preparation
2.1.1. Preparation of the Lewis acidic chloroaluminate ionic liquid
The Lewis acidic chloroaluminate ionic liquid [C4C1Im]Cl/
AlCl3 = 1/2 was prepared by mixing [C4C1Im]Cl/AlCl3 = 1/1 (Sig-
ma-Aldrich, 95%) with an equimolar amount of AlCl3 (Merck, sub-
limed, 98%). The mixture was stirred until complete dissolution of

C. Meyer et al. / Journal of Catalysis 292 (2012) 157–165
159
the AlCl3 was observed. This acidic chloroaluminate ionic liquid
was directly used as monofunctional catalyst in the liquid–liquid
biphasic n-octane isomerization experiments and for the prepara-
tion of the SILP and SCILL catalysts.
tem prior to use) and the resulting mixture was analyzed via gas
chromatography (GC).
In order to ensure reproducible data, reproducibility experi-
ments were conducted and the results are shown in the Supporting
information (Figs. S2–S7).
2.1.2. Preparation of the SILP and SCILL catalysts
Silica gel 100 (Merck, particle size 0.2–0.5 mm) was used as
support for the preparation of the SILP catalyst. The silica gel was cal-
cined prior to catalyst preparation (4 K minꢀ1 ? 423 K, 2 h isother-
mal at 423 K, 4 K minꢀ1 ? 873 K, 12 h isothermal at 873 K, under
N2). Platinum on silica (Pt/silica) (Sigma-Aldrich, 1 wt.% Pt) was used
for the preparation of the SCILL catalyst. The Pt/silica was ground,
2.3. Analytics
2.3.1. Gas chromatography (GC) analysis
All organic product samples were analyzed using a Shimadzu
2010 Plus GC equipped with a CP Sil PONA CS capillary column
(50 m ꢁ 0.21 mm ꢁ 0.5
lm) and a flame ionization detector. Con-
sieved (particle size 125–400 lm), and dried at room temperature
under high vacuum overnight prior to catalyst preparation.
version of n-octane (Xn-octane) and selectivity for component i (Si)
were calculated according to Eqs. (2), (3), or (4), respectively, with
Ai = peak area of component i.
As interactions between the acidic IL and the basic surface sites
of the silica support would result in a loss of acidity, the silica and
Pt/silica materials, respectively, were pretreated according to a
method developed by Joni et al. [17a]. In the pretreatment step, sil-
ica or Pt/silica, respectively, were impregnated with the acidic IL to
eliminate any hydroxyl group or any basic site on the support sur-
face. A defined amount of acidic IL was diluted in dichloromethane
(Akzo Nobel, 99.9%, water <20 ppm, solvent purification system
applied prior to use) and the silica or Pt/silica, respectively, was
added. After stirring this suspension for at least 1 h at 450 rpm,
the dichloromethane was removed by evaporation under vacuum.
In the next step, the impregnated support was washed with
dichloromethane to remove the excess of acidic IL that was not
covalently bound to the basic surface sites on the silica support.
The pretreated silica and pretreated Pt/silica were used for the
preparation of the SILP and SCILL catalyst by impregnating it with
Ai
jAj
P
xi ¼
ð2Þ
ð3Þ
ð4Þ
xn-octane
xn-octane;0
Xn-octane ¼ 1 ꢀ
xi ꢀ xi;0
Si ¼
xn-octane;0 ꢀ xn-octane
2.3.2. Gas chromatography–mass spectrometry (GC–MS)
To check the intactness of the IL layer of the SCILL catalyst after
reaction, the IL layer was removed in a washing procedure with
dichloromethane after the experiment. Ca. 50 ml distilled water
and 50 ml cyclohexane were added to the removed IL, stirred for
16 h followed by 5 h without stirring for phase separation. The or-
ganic phase was analyzed using a Varian 450-GC equipped with a
a defined amount of acidic IL to reach the targeted IL loading
e (Eq.
(1)). According to Eq. (1), the IL loading of the SILP and SCILL cat-
e
alyst is defined as the ratio of the used mass of acidic IL mIL and the
total mass mcat,tot of the SILP and SCILL catalyst, respectively. The
obtained catalyst was stored inside the glovebox until usage for
the catalytic isomerization experiments.
Varian VF-5 ms column (30 m ꢁ 0.25 mm ꢁ 0.25
lm) and Varian
220-MS (ion trap mass spectrometer) with electron ionization.
2.3.3. Electrospray ionization–mass spectrometry (ESI–MS)
mIL
mcat;tot
ESI–MS measurements of the organic product mixture were
taken on an esquire 6000 (Bruker Daltonics Inc.). The organic phase
was dissolved in methanol prior to analysis. Scans ranging from
50 m/z to 1500 m/z were achieved with N2 dry gas, 220 V at capil-
e
¼
ð1Þ
2.2. Isomerization experiments
lary exit, and 1562 ls accumulation time.
The n-octane isomerization experiments were carried out in a
300 ml Parr batch autoclave (see Supporting information for details,
Fig. S1). The reactor was equipped with a gas entrainment stirrer and
a glass liner. The desired amount of catalyst ([C4C1Im]Cl/AlCl3, SILP,
or SCILL catalyst) and n-octane (Merck, 99%) was filled into the glass
liner within the glovebox. The defined amount of 1-chlorooctane
(Merck, 98%)—if applied—was injected into the organic phase before
reaction. The filled glass liner was placed inside the reactor and the
reactor was flushed with argon. Prior to heating, the reactor was
charged with helium (Linde AG, 4.6) or hydrogen pressure (Linde
AG, 5.0) according to the desired reaction conditions (ptotal = phydro-
gen + phelium). An external heating jacket and internal cooling coil
were used for reactor heating, temperature control, and to ensure
isothermal reaction conditions. Upon reaching the reaction temper-
ature, the gas entrainment stirrer was set to 1000 rpm (slurry-phase
experiments with SILP or SCILL catalyst) or 1200 rpm (liquid–liquid
biphasic experiments), respectively. This moment was defined as
reaction start (treaction = 0). Samples of the organic phase were taken
during the experiment after 15, 30, 45, 60, 120, 180, and 240 min.
The organic product was filled in the sample line under reaction
pressure and valve V1 and V2 were closed. The sample line was
cooled to T = 268 K to avoid evaporation of light boiling components
while sampling. The organic product was released in excess solvent
cyclohexane (BASF, 99.9%, water <20 ppm, solvent purification sys-
2.3.4. Inductively coupled plasma atom emission spectroscopy
(ICP-AES)
The platinum and aluminum content in the organic product mix-
ture were analyzed by ICP-AES to determine the degree of metal and
IL catalyst leaching into the organic product. The acidic IL was ana-
lyzed for platinum to exclude dissolution of the platinum from the
silica support into the acidic IL. For this analysis, the IL layer was
washed off the SCILL catalyst (5 times) after use in catalysis using
dichloromethane. The amount of silicon, platinum, and aluminum
of the Pt on silica was analyzed as purchased, after pretreatment,
after IL impregnation, after reaction, and after removing of the IL
with dichloromethane. The analyses were carried out using a Ciros
CCD (SPECTRO Analytical Instruments GmbH). The latter was cali-
brated prior to measurement using plasma standard solutions of
platinum (1000
VWR Prolabo), aluminum (1000
Aesar) and silicon (1000
l
g mlꢀ1, Plasma emission standard (ICP) from
g mlꢀ1, SpecpureÒ from Alfa
g mlꢀ1, SpecpureÒ from Alfa Aesar).
l
l
2.3.5. H2 chemisorption
Pulse H2 chemisorption was performed on a Chembet 3000
(Quantachrome) at T = 273 K. Two independent investigations of
the Pt/silica catalyst (1 wt.%) (dp = 125–400
lm) were carried out,
after drying with argon (Rießner Gase, 4.6, T = 423 K, 10 K minꢀ1
,

160
C. Meyer et al. / Journal of Catalysis 292 (2012) 157–165
1 h) and after reduction in hydrogen (10% H2 in argon, H2: Rießner
Gase, 3.0, argon: Rießner Gase, 5.0; T = 473 K, t = 0.5 h; desorption
under argon flow at T = 473 K for 0.75 h). For analysis, pulses of
hydrogen in argon (10%) (Vpulse = 200 ll, Dtpulse = 3.5 min) were
injected.
80
70
60
50
40
30
20
10
0
phydrogen = 0 bar (nCl-octane / nAlCl3 = 0.036)
phydrogen = 15 bar (nCl-octane / nAlCl3 = 0.036)
phydrogen = 40 bar (nCl-octane / nAlCl3 = 0.036)
phydrogen = 0 bar (nCl-octane / nAlCl3 = 0.36)
phydrogen = 15 bar (nCl-octane / nAlCl3 = 0.36)
phydrogen = 40 bar (nCl-octane / nAlCl3 = 0.36)
2.3.6. N2 adsorption
Specific surface area and pore volume of porous catalyst samples
were determined by physisorption of nitrogen using a QUADRA-
SORB SI (Quantachrome). Before nitrogen adsorption was carried
out at 77 K, the samples were heated under vacuum (Pt/silica and
pretreated Pt/silica: 423 K, 4 h). The specific surface area was calcu-
lated according to BarrettꢀJoynerꢀHalenda (BJH) method and the
pore volume with the help of the isotherm at p p0-1 nearby 1.
0
10
20
30
40
50
60
-1
modified reaction time / min gIL
g
n-octane
Fig. 1. Influence of hydrogen partial pressure on n-octane conversion for the
monofunctional, Lewis acidic IL catalyst using different amounts of the promoter
1-chlorooctane; conditions: [C4C1Im]Cl/AlCl3 = 1/2, Treaction = 393 K.
3. Results and discussion
3.1. Bifunctional SCILL material and its characterization
n-octane isomerization. No metal catalyst and porous support were
applied in these experiments. Fig. 1 indicates the influence of
hydrogen partial pressure (phydrogen = 0–40 bar) on n-octane con-
version at 393 K for two different amounts of the promoter 1-chlo-
rooctane (nCl-octane/nAlCl3 = 0.036 and 0.36). From the results, it is
evident that the presence of hydrogen and the level of hydrogen
pressure do not result in significantly higher catalytic activity.
The higher conversions obtained for the higher level of the 1-chlo-
rooctane promoter (compare results with nCl-octane/nAlCl3 = 0.36 vs.
nCl-octane/nAlCl3 = 0.036 in Fig. 1) can be rationalized by the higher
concentration of carbenium ions in the system.
The carbenium ions are generated directly by the reaction of 1-
chlorooctane with the Lewis acidic IL [C4C1Im]Cl/AlCl3 = 1/2 and
represent the active species in the isomerization/cracking reaction
sequence. The promoting effect of chloroalkanes in acidic chloroa-
luminate ionic liquids observed here was already demonstrated by
our group in a previous publication [26] (see also Supporting infor-
mation for details, Fig. S11) and was first described by the group of
Jess for the alkylation of isobutane and 2-butene [27].
The hydrogen uptake and dispersion of platinum on the porous
silica were determined by H2 chemisorption of the dried Pt/silica
and the reduced one (see Supporting information for details,
Table S1). The dispersion of the platinum was found to be in a sim-
ilar range for both, the dried and reduced Pt/silica. We concluded
from this result that the commercial Pt/silica was already in its re-
duced form. Therefore, the material was applied without further
reduction step prior to SCILL preparation.
The bifunctional SCILL catalyst was prepared by taking the com-
mercial Pt/silica material and modifying it with the acidic IL
[C4C1Im]Cl/AlCl3 = 1/2 after silica pretreatment to neutralize basic
surface groups of the silica [17]. The ICP-AES analysis (Pt and Si)
of the Pt/silica resulted in a mean Pt value of 0.566 wt.%, which
was used for all further calculations (see Supporting information
for details, Table S2).
As expected, both the specific surface area ABJH, as determined
with BJH method (desorption), and the pore volume Vpore decrease
if the Pt/silica is pretreated according to the method described by
Joni et al. [17] (Table S3, Figs. S8–S10).
Selectivity to iso-octanes was found to be a function of n-octane
conversion but independent on the hydrogen partial pressure for
this monofunctional, Lewis acidic IL catalyst (Fig. 2). Thus, hydro-
gen does not act as cracking inhibitor in the case of the monofunc-
tional Lewis acidic IL catalyst. According to Weitkamp [3], the rate
of b-scission of carbenium ions is drastically increased if the
number of C atoms in the respective alkane is P8. This conclusion
derived from hydroisomerization studies using zeolite catalysts
The so obtained pretreated Pt/silica was coated with a defined
amount of the Lewis acidic ionic liquid [C4C1Im]Cl/AlCl3 = 1/2.
The applied amount of IL was chosen to obtain a catalyst material
with defined IL pore filling degree
a (Eq. (5)) and catalyst loading e
(Eq. (1)). The -value calculation is based on the volume of Lewis
a
acidic IL [C4C1Im]Cl/AlCl3 = 1/2 and the pore volume of the pre-
treated Pt/silica as calculated from the nitrogen adsorption
isotherm nearby p p0-1 = 1.0.
VIL
Vpore
45
a
¼
ð5Þ
phydrogen = 0 bar (nCl-octane/ nAlCl3 = 0.036)
40
phydrogen = 15 bar (nCl-octane/ nAlCl3 = 0.036)
For the evaluation of all catalytic experiments, a modified reac-
phydrogen = 40 bar (nCl-octane/ nAlCl3 = 0.036)
phydrogen = 0 bar (nCl-octane/ nAlCl3 = 0.36)
phydrogen = 15 bar (nCl-octane/ nAlCl3 = 0.36)
phydrogen = 40 bar (nCl-octane/ nAlCl3 = 0.36)
35
30
25
20
15
10
5
tion time tmod. (Eq. (6)) was applied, which includes the reaction
time treaction, mass of n-octane mn-octane, and the mass of catalyti-
cally active Lewis acidic IL [C4C1Im]Cl/AlCl3 = 1/2 mIL (mass of IL
used in liquid–liquid biphasic experiments or used for the impreg-
nation of the SILP and SCILL catalyst).
mIL
mn-octane
tmod ¼ treaction
ꢂ
ð6Þ
0
0
10
20
30
40
50
60
70
80
3.2. Catalytic experiments with monofunctional, Lewis acidic IL in
liquid–liquid biphasic reaction mode under various hydrogen
pressures
X
n-octane / %
Fig. 2. Influence of hydrogen partial pressure and n-octane conversion on selec-
tivity to iso-octanes for the monofunctional Lewis acidic IL catalyst using different
amounts of the promoter 1-chlorooctane; conditions: [C4C1Im]Cl/AlCl3 = 1/2,
Treaction = 393 K.
The first set of catalytic experiments applied a monofunctional,
Lewis acidic IL catalyst in liquid–liquid biphasic reaction mode for

C. Meyer et al. / Journal of Catalysis 292 (2012) 157–165
161
seems also to be valid for the highly acidic IL and is indicative for
45
40
35
30
25
20
15
10
5
an ionic mechanism of cleavage. Once the carbenium ions are gen-
erated, consecutive cracking reactions of mono- and dibranched
octane isomers are faster than hydride transfer from molecular
hydrogen to generate the branched high-octane products. The de-
crease of selectivity to branched octanes with increasing n-octane
conversion is in excellent agreement with this mechanistic
interpretation.
phydrogen= 0 bar, ptotal= 15 bar
phydrogen= 0 bar, ptotal= 40 bar
phydrogen= 15 bar
phydrogen= 40 bar
3.3. Pt/silica catalyst for n-octane isomerization
As expected, the untreated and uncoated Pt/silica catalysts
(e = 0, a = 0) show no catalytic activity for n-octane conversion un-
0
0
10
20
30
40
50
60
70
80
der hydrogen atmosphere (Treaction = 393 K, phydrogen = 40 bar, mPt/
silica = 10.4 g, nCl-octane/nAlCl3 = 0.36, treaction = 4 h). Due to the lack
of any sufficiently acidic site in this system, no carbenium ions
are generated and no alkane isomerization is observed.
X
n-octane / %
Fig. 4. Influence of the hydrogen partial pressure on the selectivity to iso-octanes
over n-octane conversion; conditions: mSCILL = 20 g, [C4C1Im]Cl/AlCl3 = 1/2, = 0.5,
= 0.85, Treaction = 393 K, nCl-octane/nAlCl3 = 0.036.
e
a
3.4. SCILL catalyst ([C4C1Im]Cl/AlCl3 on Pt/silica) for n-octane
isomerization
of 40 bar with the hydrogen-free system, this difference increases
to an absolute value of up to 15%. Like in the case of n-octane con-
version, the effect of total pressure on the selectivity has been
checked by pressurizing the system to the same values with he-
lium (15 and 40 bar) but in this case, the selectivity effect was
negligible.
3.4.1. Influence of hydrogen partial pressure on activity and selectivity
of the bifunctional SCILL catalyst
In the next set of experiments, the influence of hydrogen partial
pressure (phydrogen = 0, 15, and 40 bar) on activity and selectivity of
the bifunctional SCILL catalyst was investigated in the hydroisomer-
ization of n-octane (T = 393 K and nCl-octane/nAlCl3 = 0.036). The
results are depicted in Fig. 3. They reveal a remarkable dependence
of the catalyst’s activity on the hydrogen partial pressure. Obviously,
higher hydrogen partial pressures result in a significant increase in
n-octane conversion and iso-octane selectivity. At a hydrogen partial
pressure of 40 bar, the n-octane conversion was 73.7% after 4 h
reaction time, while in the absence of hydrogen, the conversion
was less than 20% under otherwise identical conditions. Comparing
these results with the outcome of the reactions with the same acidic
3.4.2. Influence of the reaction temperature on activity and selectivity
of the SCILL catalyst
To investigate the influence of the reaction temperature on the
activity and the selectivity of the SCILL catalyst, hydroisomeriza-
tion experiments were carried out between 373 K and 423 K.
Fig. 5 shows the conversion of n-octane as a function of the mod-
ified reaction time. As expected from the Arrhenius law, the reac-
tion rate of the SCILL-catalyzed reaction and thus the conversion
of n-octane over time increases with higher reaction temperature.
Remarkably, the selectivity toward iso-octanes was found to be
nearly identical for all three temperatures at comparable n-octane
conversions (see Fig. 6). This indicates that, at least within the con-
sidered temperature range, the selectivity toward iso-octanes is
independent on the applied reaction temperature.
ionic liquid in liquid–liquid biphasic mode (see Fig. 1, nCl-octane
/
nAlCl3 = 0.036), it is evident that the bifunctional SCILL system
provides a much more active and selective catalyst system. A rise
in total pressure from15 to 40 bar heliumin the absence of hydrogen
shows no significant effect on n-octane conversion.
Apart from the strongly enhanced activity, the almost parallel
shift of the iso-octane selectivity in the Siso-octane/Xn-octane-plot with
rising hydrogen partial pressures is highly remarkable (Fig. 4). At
the same level of n-octane conversion, a hydrogen partial pressure
of 15 bar increases the selectivity toward iso-octanes by an abso-
lute value of 10% compared to the reaction in the absence of
hydrogen. Comparing the reaction at a hydrogen partial pressure
3.4.3. Comparison of bifunctional SCILL and monofunctional SILP
catalysis in n-octane isomerization
To further elucidate the remarkable performance of the bifunc-
tional SCILL catalyst under hydrogen pressure and to underline the
necessity of all three essential components of the catalytic
100
100
90
80
70
60
50
40
30
20
10
0
Treaction= 373 K
Treaction= 393 K
Treaction= 423 K
phydrogen = 0 bar, ptotal = 15 bar
phydrogen = 0 bar, ptotal = 40 bar
phydrogen = 15 bar
90
80
70
60
50
40
30
20
10
0
phydrogen = 40 bar
0
10
20
30
40
50
60
0
10
20
30
40
50
60
-1
-1
modified reaction time / min gIL
g
modified reaction time / min gIL
g
n-octane
n-octane
Fig. 3. Influence of the hydrogen partial pressure on n-octane conversion over
modified reaction time; conditions: mSCILL = 20 g, [C4C1Im]Cl/AlCl3 = 1/2, = 0.5,
= 0.85, Treaction = 393 K, nCl-octane/nAlCl3 = 0.036.
Fig. 5. Influence of reaction temperature on n-octane conversion over modified
e
reaction time; conditions: mSCILL = 20 g, [C4C1Im]Cl/AlCl3 = 1/2,
phydrogen = 40 bar, nCl-octane/nAlCl3 = 0.036.
e = 0.5, a = 0.85,
a

162
C. Meyer et al. / Journal of Catalysis 292 (2012) 157–165
45
40
35
30
25
20
15
10
5
SCILL catalysts under investigation were loaded with the same
mass of Lewis acidic IL after the support pretreatment [17] and
exhibited the same pore filling degree . (For details of the silica
support characterization, see Supporting information, Table S4,
Fig. S12–S14.)
Interestingly, the catalytic results showed very similar results
for the SILP catalyst under hydrogen atmosphere as for the SCILL
catalyst under helium. This behavior suggests that the bifunctional
SCILL catalyst—with platinum as differentiating factor—acts like
the monofunctional SILP catalyst in the absence of hydrogen. Thus,
the presence of Pt becomes only relevant if the reaction is carried
out under hydrogen pressure and only the presence of hydrogen,
platinum on support, and acidic ionic liquid leads to a significant
increase in catalytic activity.
Treaction = 373 K
Treaction = 393 K
Treaction = 423 K
a
0
0
10
20
30
40
50
60
70
80
X
n-octane / %
The fact that the SCILL system shows significantly higher n-
octane conversion and selectivity to iso-octanes in the presence of
hydrogen in comparison with the SCILL catalyst under helium atmo-
sphere further confirms the importance of hydrogen activation by
the catalytic Pt-center for the observed reactivity. According to the
mechanism of isomerization, the n-octane molecule is converted
first to an unbranched octanyl cation by hydride abstraction
through a highly acidic proton or by another carbenium ion. The
octanyl cation undergoes isomerization to form a branched octanyl
cation. The latter can react via hydride transfer with any alkane of
the reaction mixture to yield the corresponding branched octane
isomer and a new carbenium ion. However, besides the desired
hydride transfer, the branched octanyl cation can also undergo
undesired side reactions. These include cracking reactions via b-
scission or alkylation reactions of carbenium ions with certain
amounts of olefins that are formed e.g. by deprotonation of carbe-
nium ions (Scheme 1) [28] or cracking [3]. Both reactions decrease
the selectivity to branched iso-octanes significantly.
In the presence of hydrogen, Lewis acidic ionic liquid, and active
platinum centers, most of the olefin formed by deprotonation is
likely to be hydrogenated prior to the undesired alkylation reac-
tion. This hydrogenation activity explains the higher selectivity
to iso-octanes. The role of hydrogenation reactions in the bifunc-
tional SCILL system can also be recognized by monitoring the total
reactor pressure during the different isomerization experiments
(Fig. 9). The hydrogen pressure of the experiment conducted with
Fig. 6. Influence of reaction temperature on selectivity to iso-octanes over n-octane
conversion conditions: mSCILL = 20 g, [C4C1Im]Cl/AlCl3 = 1/2,
phydrogen = 40 bar, nCl-octane/nAlCl3 = 0.036.
e = 0.5, a = 0.85,
system—platinum, Lewis acidic chloroaluminate IL, and hydro-
gen—comparison experiments with a Pt-free, monofunctional,
solid supported system (SILP catalyst) were carried out. All exper-
iments were conducted at 373 K. For both the bifunctional SCILL
system and the monofunctional SILP system, the effect of hydrogen
vs. helium pressure was investigated (Figs. 7 and 8). The SILP and
80
SILP (phydrogen= 40 bar)
SCILL (phelium = 40 bar)
SCILL (phydrogen = 40 bar)
70
60
50
40
30
20
10
0
0
10
20
30
40
50
60
-1
modified reaction time / min gIL
g
n-octane
Fig. 7. Comparison of n-octane conversion over modified reaction time for the SILP
catalyst under hydrogen atmosphere (phydrogen = 40 bar) vs. the SCILL catalysts
under helium (phelium = 40 bar) and hydrogen atmosphere (phydrogen = 40 bar);
conditions: [C4C1Im]Cl/AlCl3 = 1/2, Treaction = 373 K; SILP: mSILP = 23 g,
e = 0.4,
a
= 0.85; SCILL: mSCILL = 20 g,
e
= 0.5,
a
= 0.85.
Scheme 1. Equilibrium reaction of carbenium ion and olefin in acidic media [28].
45
40
35
30
25
20
15
10
5
60
55
50
45
40
35
SILP (phydrogen= 40 bar)
SCILL (phelium = 40 bar)
SCILL (phydrogen = 40 bar)
30
25
20
15
SCILL (hydrogen)
SCILL (helium)
SILP (hydrogen)
0
0
10
20
30
40
50
60
70
80
0
10
20
30
40
50
X
n-octane / %
-1
modified reaction time / min gIL
g
n-octane
Fig. 8. Comparison of selectivity to iso-octanes as function of n-octane conversion
for the SILP catalyst under hydrogen atmosphere (phydrogen = 40 bar) vs. the
SCILL catalysts under helium (phelium = 40 bar) and hydrogen atmosphere
(phydrogen = 40 bar); conditions: [C4C1Im]Cl/AlCl3 = 1/2, Treaction = 373 K; SILP:
Fig. 9. Reactor pressure over modified reaction time for SILP (phydrogen = 40 bar) and
SCILL catalysts under hydrogen and helium atmosphere (phydrogen = 40 bar and
phelium = 40 bar); conditions: [C4C1Im]Cl/AlCl3 = 1/2, Treaction = 373 K; SILP:
mSILP = 23 g,
e
= 0.4,
a
= 0.85; SCILL: mSCILL = 20 g,
e
= 0.5,
a
= 0.85.
mSILP = 23 g, e = 0.4, a = 0.85; SCILL: mSCILL = 20 g, e = 0.5, a = 0.85.

C. Meyer et al. / Journal of Catalysis 292 (2012) 157–165
163
the SCILL catalyst showed a notable continuous decrease, while the
pressure in the two comparison runs—monofunctional SILP under
hydrogen and bifunctional SCILL under helium—remained almost
constant (the little jags in the graphs are due to reactor sampling).
Only in the presence of Pt and hydrogen, the hydrogenation of ole-
fins that are formed via b-scission of carbenium ions or by carbe-
nium ion deprotonation takes place and results in steady
hydrogen consumption throughout the reaction time. Note that
no net hydrogen consumption should be observed if olefins are
hydrogenated that are formed by deprotonation of a carbenium
ion, as the released acidic proton will typically react with an alkane
to yield a new carbenium ion and hydrogen molecule (see Support-
ing information for details, Scheme S1).
While the hydrogenation activity gives a good explanation for
the improved selectivity of the bifunctional SCILL system, it does
not explain yet its much higher activity. The fact that the catalytic
activity correlates also strongly to the level of hydrogen partial
pressure (see Fig. 3) suggests that the concentration of acidic pro-
tons and thus the acidity of the catalytic system increases if hydro-
gen dissolved in the Lewis acidic ionic liquid is in contact with the
active platinum center. It is known that hydrogen can homolytical-
ly split on platinum [29]. The catalytic result is in line with the
interpretation that in the strongly Lewis acidic environment of
the chloroaluminate melt, electron transfer may lead to the forma-
tion of Pt-hydride and additional protons in the IL. This interpreta-
tion is supported by the fact that 1-chlorooctane added to the
bifunctional SCILL system causes no additional enhancement of
the catalytic activity (see Supporting information for details,
Figs. S15 and S16). Obviously, the amount of carbenium ions
formed from hydrogen at the catalytic Pt-center is dominating
the catalytic activity of the system.
Fig. 11. Composition of by-products (carbon fraction in the respective by-product
fraction) as a function of n-octane conversion for the monofunctional IL and the
bifunctional SCILL catalyst. The fractions contain n- and iso-alkanes. The sum of all
fractions adds up to 100%; conditions: [C4C1Im]Cl/AlCl3 = 1/2, Treaction = 393 K,
phydrogen = 40 bar; SCILL catalyst: mSCILL = 20 g,
e
= 0.5,
a = 0.85, nCl-octane/
nAlCl3 = 0.036; IL catalyst: mIL = 10 g, nCl-octane/nAlCl3 = 0.36.
bifunctional SCILL catalyst and the liquid–liquid catalysis with the
acidic IL can be observed. In Fig. 11, the selectivities for products of
different chain lengths are given as their fractions in total carbon.
For the SCILL catalyst, the fraction of propane and butanes within
the by-products increases, while the fraction of pentanes, hexanes,
and heptanes decreases compared to the monofunctional Lewis
acidic IL catalyst. The fraction of the by-products with a chain
length higher than C8 is unaffected. The differences in the distribu-
tions of the by-products can again be explained by the hydrogena-
tion activity of the Pt-containing catalyst. While the shorter chain
by-products (C3–C4) are mainly formed by b-scission, the formation
of by-products with a higher chain length requires a combination of
alkylation and cracking reactions. The alkylation reactions are
dependent on the existence of olefines, which are formed by
cracking reactions via b-scission or by deprotonation of carbenium
ions. In the presence of platinum and hydrogen, a significant part of
the olefins is hydrogenated. With decreasing concentration of
olefins, the alkylation rate decreases leading to a lower fraction of
C5–C7 products. For both the monofunctional and the bifunctional
catalyst, the formed by-products consist mainly of branched
alkanes (around 90% iso-alkanes and 10% n-alkanes).
3.5. Comparison of the product distributions for the bifunctional SCILL
catalyst vs. liquid–liquid biphasic catalysis
Fig. 10 represents the composition of the octane isomers as a
function of n-octane conversion for the slurry-phase experiment
using the bifunctional SCILL catalyst and the liquid–liquid biphasic
experiment using the monofunctional Lewis acidic IL alone
(phydrogen = 40 bar, T = 393 K). Both catalyst systems yield only
monobranched and dibranched octane isomers in very similar
distributions. The absence of even higher branched octane isomers
is due to their high cracking rate.
When comparing the composition of the alkane products with
carbon chain length other than C8, differences between the
3.6. Catalyst stability
An important aspect for the further development of these
bifunctional SCILL systems toward technical applications is the sta-
bility of the catalyst. ICP-AES analyses of the product proved that
no platinum and aluminum leach into the organic phase (Table 1).
As no aluminum leaching is observed, we conclude that no Lewis
acidic IL is lost to the product phase during reaction. This indicates
that the very unpolar feedstock/product phase is not able to dis-
solve the ionic liquid off the support. Moreover, the results indicate
that the ionic liquid physical adsorption on the support is strong
enough to avoid removal of ionic liquid droplets from the support.
In addition, the platinum leaching into the immobilized acidic IL
was analyzed. The question whether the Pt stays on the support
100
bifunctional
mono-branched C8
monofunctional
mono-branched C8
90
80
70
60
50
40
30
20
10
0
di-branched C8
di-branched C8
Table 1
0
10
20
30
40
50
60
70
80
ICP-AES (Pt and Al) results of the organic and removed IL phase.
X
n-octane / %
Element
Analyzed
phase
Leaching (wt.%)
(referred to total Pt, Al)
Detection limit (wt.%)
(referred to total Pt, Al)
Fig. 10. Composition of the formed branched octane isomers as a function of n-
octane conversion for the monofunctional IL catalyst and for the bifunctional SCILL
catalyst; conditions: [C4C1Im]Cl/AlCl3 = 1/2, Treaction = 393 K, phydrogen = 40 bar; SCILL
catalyst: mSCILL = 20 g,
nCl-octane/nAlCl3 = 0.36.
Pt
Al
Pt
Organic
Organic
Ionic liquid
–
–
0.21
0.03
0.03
0.03
e = 0.5, a = 0.85, nCl-octane/nAlCl3 = 0.036; IL catalyst: mIL = 10 g,

164
C. Meyer et al. / Journal of Catalysis 292 (2012) 157–165
Table 2
ICP-AES (Pt, Si, and Al) results of the applied Pt on silica material, SCILL catalyst before
reaction (e = 0.5, a = 0.85), SCILL catalysts after reaction, and removal of acidic IL.
Entry
Sample
Pt/Si ratio
(mol molꢀ1
)
1
Pt/silica (mean value)
1.76 ꢁ 10ꢀ3
1.03 ꢁ 10ꢀ4
1.36 ꢁ 10ꢀ3
1.52 ꢁ 10ꢀ3
Standard deviation (for mean value calculation)
SCILL catalyst, before reaction
SCILL catalyst, IL washed off
2
3
(the bifunctional SCILL concept) or dissolves in the acidic ionic
liquid during reaction (leading to a bifunctional SILP concept) is
of conceptual importance. In the SCILL case, the support material
is of direct relevance for the nature of the catalytic Pt-center, while
in the SILP case, only the embedding ionic liquid defines the reac-
tivity of the dissolved Pt-complexes. The platinum leaching into
the ionic liquid layer was checked in two independent ways: (a)
the acidic IL was removed after reaction from the support by a
washing procedure with dichloromethane and the washing solu-
tion was checked for its Pt-content by ICP-AES; (b) the Pt/Si ratio
of the applied Pt on silica catalyst was determined by ICP-AES prior
to the IL coating and after catalytic reaction and removal of the io-
nic liquid. The results are given in Table 2. The deviation of the Pt/
Si ratio of SCILL catalyst and Pt/silica (entry 1, 2 and entry 1, 3,
respectively) is higher than the standard deviation of the Pt/silica
mean value calculation (Table 2). This fact indicates small amount
of platinum leaching from the silica into the Lewis acidic IL, which
is proved by the ICP-AES results (0.21 wt.% of the total Pt) of the
washed-off acidic IL. The Pt leaching can take place during pre-
treatment or during catalysis. However, it is important to note that
by far the largest part of the platinum remains on support, making
the whole catalyst a true bifunctional SCILL system.
Fig. 12. Comparison of n-octane conversion over modified reaction time for three
catalytic runs with the same SCILL catalyst; conditions: mSCILL = 12 g, [C4C1Im]Cl/
AlCl3 = 1/2,
e = 0.5, a = 0.85, Treaction = 373 K, phydrogen = 40 bar.
To exclude the theoretical possibility of chlorine insertion from
the chloroaluminate IL into the organic products, the organic phase
(experiments without 1-chlorooctane addition) was analyzed by
ESI–MS. The chlorine concentration was found to be clearly below
the detection limit as no characteristic chlorine peaks (m1/
z + 2 = m2/z) were found in the mass spectrum (see Supporting
information for details, Fig. S17).
Fig. 13. Comparison of selectivity to iso-octanes as function of n-octane conversion
for three catalytic runs with the same SCILL catalyst; conditions: mSCILL = 12 g,
[C4C1Im]Cl/AlCl3 = 1/2,
e = 0.5, a = 0.85, Treaction = 373 K, phydrogen = 40 bar.
stability of SCILL catalysts as no Al and Pt can be detected in the or-
ganic product phase after reaction. Consequently, the recyclability
of the SCILL catalysts under hydrogen atmosphere was tested in
two recycling runs (see Figs. 12 and 13). For the recycling experi-
ments, most of the organic phase was drawn off with a syringe
from the SCILL catalyst after settling time at ambient reaction con-
ditions and replaced with new n-octane. A small amount of organic
matter left in the glass liner acted as protective liquid layer for the
SCILL catalyst during loading of n-octane. It was unavoidable that
some amount of the solid SCILL catalyst sticking to the stirrer
and cooling coil were shortly exposed to air while the organic
phase was changed in the glass liner.
Conversion of n-octane decreases slightly with every catalyst
recycling from 9.6% (recycling run 0) to 8.4% (recycling run 1)
and finally 6.6% (recycling run 2) after 2 h reaction time. In con-
trast, the curve of iso-octanes selectivity over n-octane conver-
sion is almost identical for all three catalytic runs (Fig. 13).
These results indicate the recyclability of the SCILL catalyst.
Losses in catalytic activity can be ascribed to the deactivation
of SCILL catalyst that came in contact with air during the re-
charging procedure.
Typical heterogeneous, state-of-the-art isomerization/cracking
catalysts operate at much higher temperatures than the here pre-
sented SCILL systems. Therefore, the deposition of heavy products
on the catalytic surface and coking is a major deactivation path in
these high temperature systems. To demonstrate the advantage of
the SCILL systems operating at mild temperatures in this respect,
the ionic liquid was tested after reaction for any accumulation of
high boiling, unsaturated, or cyclic by-products. The removed IL
from the SCILL catalyst (washing procedure with dichloromethane)
was contacted with an excess water to hydrolyze the IL anion. The
resulting aqueous solution was extracted with cyclohexane, and
the organic extraction phase was investigated for all organic prod-
ucts accumulated in the IL phase during reaction. The highest m/z
ratio found in the mass spectrum in this analysis was 164, which
might originate from cyclic unsaturated hydrocarbons. However,
this mass and the respective size of the molecule are still small
enough to expect this species to be in extraction equilibrium with
the organic product phase. Thus, accumulation of carbon-rich
species in the SCILL system or coke formation can be excluded
under the conditions applied in this study. Catalyst stability includ-
ing coke formation under extended operation or in a continuous
reaction mode has not been investigated up to now.
4. Conclusion
3.7. Catalyst recycling
Bifunctional Solid Catalyst with Ionic Liquid Layer (SCILL) sys-
tems composed of Pt on porous silica coated with a thin layer
of acidic chloroaluminate ionic liquid exhibit surprising catalytic
The recycling of the catalyst is an issue of practical relevance.
The results discussed in chapter 3.6 already demonstrate the

C. Meyer et al. / Journal of Catalysis 292 (2012) 157–165
165
performance in the isomerization of n-octane in the presence of
References
hydrogen at moderate reaction conditions. Our investigations
reveal that all components, the platinum on the solid support,
the Lewis acid ionic liquid coating, and the hydrogen pressure
in the system are critical to realize the observed enhancements
in catalytic activity and iso-octane selectivity. While the selectiv-
ity effect is understandable from the fact that the Pt-centers in
the system act as hydrogenation catalysts to decrease the
concentration level of olefinic intermediates in the system (thus
reducing consecutive alkylation steps), the increasing catalytic
activity with increasing hydrogen partial pressure is very
remarkable. Based on the known fact that catalytic activity in
skeleton isomerization of alkanes is primarily a function of acid-
ity, it is reasonable to assume that the special combination of
the immobilized Pt nanocrystals embedded in the highly acidic
ionic liquid may lead to a heterolytic cleavage of dissolved
hydrogen forming surface bound Pt-hydride and additional pro-
tons during reaction. At least, the obtained catalytic results indi-
cate a gradual increase of the systems acidity with increasing
hydrogen pressure. This observation alone may open new and
interesting research perspectives for the near future. In addition,
the general industrial interest in efficient and durable alkane
isomerization catalysts working at moderate reaction conditions
for the upgrading of fuels makes the here reported bifunctional
SCILL catalysts a highly attractive object for further research
and development.
[1] (a) H. Weyda, E. Köhler, Catal. Today 81 (2003) 51–55;
(b) R.P. Silvy, Appl. Catal. A 261 (2004) 247–252.
[2] N.A. Cusher, in: R.A. Meyer (Ed.), Handbook of Petroleum Refining Processes,
third ed., McGraw-Hill, 2003. 9.29–9.39, 9.41–9.45.
[3] J. Weitkamp, P.A. Jacobs, J.A. Martens, Appl. Catal. 8 (1983) 123–141.
[4] R. Glaeser, J. Weitkamp, Springer Series, in: Chemical Physics, Springer-Verlag,
Heidelberg, 2004, pp. 161–212.
[5] D. Verboekend, K. Thomas, M. Milina, S. Mitchell, J. Pérez-Ramírez, J.-P. Gilson,
Catal. Sci. Technol. (2011) 1, 1331–1335.
[6] A. Fúnez, J.W. Thybaut, G.B. Marin, P. Sánchez, A. De Lucas, J.L. Valverde, Appl.
Catal. A 349 (2008) 29–39.
[7] A. Chica, A. Corma, Chem. Ing. Tech. 79 (6) (2007) 857–870.
[8] S. Ernst, C. Bischof, M. Hartmann, J. Weitkamp, in: Proc. of the 16th World
Petroleum Congress, vol. 3, 2000, p. 27.
´
[9] M. Busto, K. Shimizu, C.R. Vera, J.M. Grau, C.L. Pieck, M.A. DAmato, M.T. Causa,
M. Tovar, Appl. Catal. A 348 (2008) 173–182.
[10] M. Busto, M.E. Lovato, C.R. Vera, K. Shimizu, J.M. Grau, Appl. Catal. A 355 (2009)
123–131.
[11] J. Xu, J.Y. Ying, Angew. Chem. Int. Ed. 45 (2006) 6700–6704.
[12] M.A. Arribas, F. Márquez, A. Martínez, J. Catal. 190 (2000) 309–319.
[13] T. Ohno, Z. Li, N. Sakai, H. Sakagami, N. Takahashi, T. Matsuda, Appl. Catal. A
389 (2010) 52–59.
[14] H. Al-Kandari, F. Al-Kharafi, A. Katrib, Appl. Catal. A 383 (2010) 141–148.
[15] J.C. Yori, J.M. Grau, V.M. Benítez, J. Sepúlveda, Appl. Catal. A 286 (2005) 71–
78.
[16] C. Meyer, P. Wasserscheid, Chem. Commun. 46 (2010) 7625–7627.
[17] (a) J. Joni, M. Haumann, P. Wasserscheid, Adv. Synth. Catal. 351 (2009) 423–
431;
(b) J. Joni, M. Haumann, P. Wasserscheid, Appl. Catal. A 372 (1) (2010) 8–15.
[18] U. Kernchen, B. Etzold, W. Korth, A. Jess, Chem. Eng. Technol. 30 (8) (2007)
985–994.
[19] N. Wörz, J. Arras, P. Claus, Appl. Catal. A 391 (2011) 319–324.
[20] J. Arras, E. Paki, C. Roth, J. Radnik, M. Lucas, P. Claus, J. Phys. Chem. C 114 (2010)
10520–10526.
[21] J. Arras, D. Ruppert, P. Claus, Appl. Catal. A 371 (2009) 73–77.
[22] J. Arras, M. Steffan, Y. Shayeghi, D. Ruppert, P. Claus, Green Chem. 11 (2009)
716–723.
Acknowledgments
Dr. R.A. Rakoczy and Dr. A. Geisbauer (Süd-Chemie) are
acknowledged for fruitful discussions. We thank the DFG for sup-
porting some parts of this research (biphasic and SILP systems)
within the Erlangen Cluster of Excellence ‘‘Engineering of Ad-
vanced Materials.’’ The European Community supported the SCILL
part of the here reported studies through an ERC Advanced Grant
to P.W.
[23] R. Knapp, A. Jentys, J.A. Lercher, Green Chem. 11 (2009) 656–661.
[24] Y. Gu, C. Ogawa, J. Kobayashi, Y. Mori, S. Kobayashi, Angew. Chem. Int. Ed. 45
(2006) 7217–7220.
[25] H.-P. Steinrück, J. Libuda, P. Wasserscheid, T. Cremer, C. Kolbeck, M. Laurin, F.
Maier, M. Sobota, P.S. Schulz, M. Stark, Adv. Mater. 23 (2011) 2571–2587.
[26] C. Meyer, V. Hager, M. Haumann, P. Wasserscheid, in: Presentation at
International DGMK Conference, Catalysis
Petrochemistry and Refining, Dresden, 2011.
– Innovative Applications in
[27] S. Aschauer, L. Schilder, W. Korth, S. Fritschi, A. Jess, Catal. Lett. 141 (2011)
1405–1419.
[28] H. Pines, N.E. Hoffman, Isomerization of Saturated Hydrocarbons, in: G.A. Olah
(Ed.), Friedel Crafts and Related Reactions, vol. 2, Interscience Publishers, New
York, 1964, pp. 1211–1217.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2012.05.008.
[29] K. Christmann, G. Ertl, T. Pignet, Surf. Sci. 54 (1976) 365–392.