Liquid-phase isobutane/butene-alkylation using promoted Lewis-acidic IL-catalysts

Article abstract of DOI:10.1007/s10562-011-0675-2

The effect of different promoters on activity and selectivity of Lewis-acidic chloroaluminate ionic liquid catalysts was studied for isobutane/2-butene alkylation. When tert-butyl halides are used as promoters, the active species of the alkylation reaction, which is the tert-butyl cation, is directly generated whereas upon catalysis with Bronsted-acid supported ionic liquids, this species is indirectly provided through a hydride shift between protonated 2-butene and isobutane. Experimental results both from batch and continuously operated liquid phase alkylation reactors indicate, that tert-butyl halides are able to speed up the reaction rate significantly and shift the C₈-selectivity towards the desired high-octane trimethylpentanes (TMPs). However, secondary reactions like oligomerization and cracking could not be suppressed by the use of this additives and high deactivation rates in continuous opperation were observed. Suggestions are made, how the product composition is effected by the additive and how the promoted IL-catalyst system is deactivated with time on stream.

Full text of DOI:10.1007/s10562-011-0675-2

Catal Lett (2011) 141:1405–1419
DOI 10.1007/s10562-011-0675-2
Liquid-Phase Isobutane/Butene-Alkylation Using Promoted
Lewis-Acidic IL-Catalysts
•
Stephan Aschauer Lisa Schilder Wolfgang Korth
•
•
•
Susanne Fritschi Andreas Jess
Received: 30 April 2011 / Accepted: 23 July 2011 / Published online: 11 August 2011
Ó Springer Science+Business Media, LLC 2011
Abstract The effect of different promoters on activity
and selectivity of Lewis-acidic chloroaluminate ionic
liquid catalysts was studied for isobutane/2-butene alkyl-
ation. When tert-butyl halides are used as promoters, the
active species of the alkylation reaction, which is the tert-
butyl cation, is directly generated whereas upon catalysis
with Brønsted-acid supported ionic liquids, this species is
indirectly provided through a hydride shift between pro-
tonated 2-butene and isobutane. Experimental results both
from batch and continuously operated liquid phase alkyl-
ation reactors indicate, that tert-butyl halides are able to
speed up the reaction rate significantly and shift the
C8-selectivity towards the desired high-octane trimethyl-
pentanes (TMPs). However, secondary reactions like
oligomerization and cracking could not be suppressed by
the use of this additives and high deactivation rates in
continuous opperation were observed. Suggestions are
made, how the product composition is effected by the
additive and how the promoted IL-catalyst system is
deactivated with time on stream.
octane number and can be considered to be free of sulphur
and aromatics [1]. In future, the importance and global
production of alkylate may even grow. Catalytic reforming
of heavy straight-run naphtha, which is today still the most
important process for high octane gasoline production,
yields a gasoline blending stream containing huge amounts
of aromatic compounds, but the content of aromatics in
gasoline for sale is more and more limited. Hence, pro-
cesses like alkylation, isomerisation and production of
MTBE (methyl-tert-butyl-ether) may become increasingly
important.
Up to now the alkylation catalysts used in industry are
sulfuric acid or hydrogen fluoride [1]. Both catalysts show
some drawbacks from an environmental as well as toxi-
cological point of view. Although HF is a very efficient
catalyst, it is of course extremely toxic for humans and
animals. So plants using HF catalyst have to be provided
with strict and reliable safety equipment. However, the
governmental requirements became so strict that new
plants using HF are unlikely to be built. On the other hand
sulfuric acid is considerably consumed during the reaction
(about 100 kg/talkylate), very corrosive but environmentally
less hazardous compared to HF. So, new less toxic and
hazardous catalysts are of great interest for the alkylation
process.
Keywords Alkylation ꢀ Isobutane ꢀ Chloroaluminate
ionic liquid ꢀ tert-butyl halides
1 Introduction
Highly Brønsted acidic solid acids are considered as
alternative catalysts ([2–5]). Although substantial progress
for solid catalysts has been reported in literature, they still
show a fast deactivation [6, 7]. Some attempts to regenerate
the solid catalysts using hydrogen are reported, but the
initial activity can not be restored completely during the
regeneration process [8, 9]. Other alternatives described in
literature are highly Brønsted acidic ionic liquids, which
are prepared by mixing of Lewis acidic ionic liquids with
Brønsted acids in order to provide the catalytic active
The alkylation of isobutane with 2-butene forming highly
branched C8-paraffins is an important refinery process
since the products of this reaction provide a high research
S. Aschauer ꢀ L. Schilder ꢀ W. Korth ꢀ S. Fritschi A. Jess (&)
¨
University of Bayreuth, Universitatsstraße 30, 95445 Bayreuth,
Germany
e-mail: jess@uni-bayreuth.de
123

1406
S. Aschauer et al.
species [10–16]. Furthermore, ionic liquid systems are
described in which a strong Brønsted acid is dissolved [17,
18]. The results with ionic liquids published in literature
until now are very promising with regard to an alternative
to the ‘‘classical’’ catalysts for alkylation, especially one
where sulfuric acid in a Chinese alkylation plant with a
capacity of 65,000 tonnes per year was replaced by a
composite ionic liquid producing a comparable alkylate
quality [19]. Welton [20] mentioned that this is the almost
only industrial use of an ionic liquid as catalyst yet.
However, in all cases the alkylation catalysts are more or
less very strong Brønsted acids. Such acids are required
due to the generally accepted mechanism of alkylation
proposed by Schmerling [21].
Hal
AlCl3
+
-
AlCl3Hal
intermolecular
hydride-transfer
+
intramolecular
rearrangement
Here, the acid protonates the olefin which abstracts a
hydride from isobutane (Scheme 1) yielding n-butane and
the stable tert-butyl cation. Alternatively, the sec-butyl
cation formed by protonation of 2-butene isomerizes via an
intramolecular methyl shift. The tert-butyl cation formed
according to Scheme 1 can be regarded as the actual active
species of isobutane-butene-alkylation, since it initiates the
catalytic cycle by electrophilic addition to the olefin
yielding the 2,2,3-TMPyl-cation (see Scheme 2). As 2,2,3-
TMP is not the observed main alkylation product, it is
postulated that rapid intramolecular rearrangement of the
C8-carbocation precedes hydride transfer from isobutane
[1]. The intermolecular hydride-shift from isobutane is
generally accepted to be the rate determing step in the
reaction [22, 23].
Scheme 2 Catalytic circle of isobutane/2-butene alkylation
Br, I) to Lewis acidic ionic liquids as in classic Friedel
Crafts alkylation (Scheme 4).
However, the Lewis acid has to be acidic enough to
react with the alkyl halogenide. Direct formation of the
tert-butyl cation should reduce unwanted side reactions
like oligomerization of 2-butene and increase TMP selec-
tivity. Literature research confirms the idea since this
concept is described in patents disclosed by Chevron [24,
25]. The inventors describe alkylation of isoparaffins with
olefins using Lewis acidic chloroaluminate ionic liquids
promoted with alkyl halides as an example. In the cited
patent literature [24, 25] detailed examples on the product
distribution are found, but only regarding the alkylation of
isopentane with ethylene. Here, the best yield of the desired
C7-compounds was obtained with ethyl chloride as pro-
moter in comparison to the reaction promoted by HCl.
Apart from ionic liquids they also claimed compounds that
have Lewis acidic character. Furthermore, Rosenbach and
Mota proved a closely related concept to the one described
above with solid catalysts [26]. Rosenbach and Mota
obtained tertiary carbocations by reaction of metal
exchanged zeolites with alkyl halides. Since alkylation of
C4-isoparaffins with 2-butene is technically more impor-
tant, the reaction using Lewis acidic ionic liquids promoted
with tert-butyl halides was investigated in this work, in
order to gain more information about the catalytic perfor-
mance (activity, selectivity) and activation/deactivation
Since isobutane-butene alkylation becomes autocatalytic
by this hydrogen-transfer step, it is crucial to form a high
number of this activated isobutane-species at the beginning
of the reaction.
Lewis-acidic chloroaluminate ionic liquids inevitably
contain protic impurities like water. Since every water-
molecule releases two HCl-molecules according to Scheme
3, the initiation of alkylation starts with the same pathway,
which is taken by Brønsted acidic catalyts, e. g. Scheme 1.
However, according to the reaction mechanism it can be
concluded that highly acidic Brønsted acids are not
required for the alkylation reaction as long as a source for
tertiary cations is provided to the reaction system. The
formation of such tertiary carbenium ions is simply pos-
sible by adding a tertiary alkyl halogenide (e.g. hal. = Cl,
H
Cl
Al
Cl
Al
Cl
Al
Cl
Cl
Al
Cl
Al
Cl
Al
Cl
Cl
O
OH
Cl
Cl
+ HCl
H
Cl
Cl
Cl
O
Cl
Cl
Cl
Cl
Cl
intramolecular
rearrangement
Cl
H
H
H
+
Cl
Al
intermolecular
Cl
Al
Cl
Al
Cl
Cl
Cl
Cl
hydride-transfer
+
HCl
+
2
H
Cl
Cl
Cl
Cl
Scheme 1 Protonation and rearrangement of 2-butene
Scheme 3 Formation of Brønsted acid catalysts by protic impurities
123

Liquid-Phase Isobutane/Butene-Alkylation
1407
In a typical experiment 3.5 g [BMIM][AlCl4] was stir-
red with an equimolar amount of 2-olefin, iso-paraffin or
n-paraffin under inert atmosphere for 3 h at a temperature
of 25 °C. Solubility measurements were also performed for
mixtures with different molar isoalkane/2-alkene-ratios.
After saturation of the ionic liquid with the hydrocarbons
the biphasic system was allowed to settle down at least 3 h.
For reextraction, a 10 fold volumetric excess of cyclo-
hexane over ionic liquid was added to 1g of the saturated
ionic liquid-phase. The resulting mixture was stirred for
another 3 h at 25 °C and then left for further 3 h until
phase separation occured. Prior to GC-analysis, 2,4-
dimethylhexane as external standard was admixed with the
organic phase. With GC-analysis the amount of hydrocar-
bons in the reextracted phase were determined to calculate
the mass fraction-related partition coefficient. Additional
reactions (dimerization, cracking) could not be observed in
the reextracted phase.
+
AlCl3Hal
+
AlCl3
Hal
Scheme 4 Formation of tert-butyl cation
behaviour. The reaction was performed batch and contin-
uously as a liquid–liquid reaction.
2 Experimental
Feed gases cis-2-butene (purity grade 2.0) and isobutane
(purity grade 3.5) were purchased from Rießner gases and
used as obtained. The ionic liquid 1-butyl-3-methylimida-
zolium tetrachloroaluminate [BMIM][AlCl4] (95% purity),
cyclohexane (99.9% purity), tert-butyl chloride (99% pur-
ity), tert-butyl bromide (97% purity) and methyl-tert-butyl-
ether (99% purity) were obtained from Sigma Aldrich and
used without further purification. AlCl3 was obtained from
Fluka (99.9% purity) and used without further purification.
All experimental manipulations were conducted under inert
atmosphere (Argon) using standard Schlenk techniques or
in a nitrogen filled dry box.
The mass fraction-related partition coefficient Kw;i for
component i is defined as quotient of mass fraction wi;IL of i
eq
in the ionic liquid phase to mass fraction wi;org of i in the
eq
organic phase, both values holding for equilibrium situation:
w
i;ILeq
Kw;i
¼
ð2Þ
2.1 Synthesis of Ionic Liquid Catalyst
w
i;orgeq
Further details for calculation of Kw;i-values and the
extraction/reextraction procedure are described in [27].
Ionic liquid catalysts were prepared by slow addition of the
desired amount of anhydrous AlCl3 to liquid [BMI-
M][AlCl4] (xAlCl ¼ 0:5). The resulting Lewis acidic mix-
3
2.3 Product Analysis
ture was stirred for 24 h. After completion of the reaction a
yellow to brown liquid was obtained depending on the
amount of AlCl3 added. Lewis acidic catalysts containing
In general, appropriate analytics of alkylation products
have always been a considerable problem, because alkylate
is a complex mixture of up to 200 highly branched
hydrocarbons showing C-numbers from 4 to 15. Thus,
partial losses of light or heavy compounds during sampling
or storing of the samples and complete separation of all
product components are serious challenges to face in
alkylate analysis [28]. Furthermore, especially in case of a
deactivating catalyst, analytical data obtained from integral
samples, which are collected over an extended time on
stream, is highly problematic, since a change of product
distribution during time of sampling cannot be excluded
[28]. Weitkamp et al. [29] were the first who applied a
differential sampling system, where sampling frequencies
of up to 6 samples per minute are possible. The glass
ampoule smasher connected to a gaschromatographic sys-
tem used to analyze the samples has several disadvantages
like inefficient operation and broken ampoules remaining
in the smashing chamber. Furthermore, it remains ques-
tionable, if the involved, repetitive vaporization and liqu-
ification-procedure of the samples does not lead to a loss of
light or heavy end-products. So another, more elegant
sampling and analyzing system was designed and
mole fractions xAlCl of 0.64 were prepared. The mol
3
fraction x of AlCl3 in the ionic liquid is defined as:
n
AlCl3
xAlCl
¼
ð1Þ
3
nAlCl þ n
½BMIMꢁCl
3
2.2 Measurement of Isoalkane/2-Alkene-Solubilities
in Ionic Liquids via Extraction Method
Solubility measurements were performed by a consecutive
extraction- and reextraction procedure. Since Lewis acidic
chloroaluminate ionic liquids (xAlCl [ 0:5Þ catalyse addi-
3
tional reactions, like olefin-dimerization or paraffin-crack-
ing, the solubility measurements could only be performed in
a neutral chloroaluminate melt [BMIM][AlCl4] (xAlCl
¼
3
0.5). Due to the fact that isobutane and 2-butene solubilities
can not be measured by extraction- and reextraction proce-
dure, because of their high partial pressures (gaseous at
atmospheric pressure), solubility measurements were con-
ducted using different isoalkane/2-alkene-mixtures, which
are liquid ar ambient conditions. The solubilities of the iso-
butane/2-butene-system therefor can be extrapolated from
these data.
123

1408
S. Aschauer et al.
Table 1 Set up of the gas chromatograph
(a) Double-end Shutoff (DESO); stem+body;
http://www.swagelok.com
Carrier gas
Detector
Capillary-column
He 4.6
FID
CP Sil Pona CB
–
Parameter
Unit
Value
Oxygen-pressure
H2-pressure
bar
2.41
1.24
4.13
8
bar
(b) Vaporization system
He-pressure
bar
TI
column
Purgegas flow
Split flow
mL min-1
mL min-1
mL min-1
°C
He
PI
35
N2
sample
vapourizer
Carrier gas flow
Injector temperature
Detector temperature
30
250
260
°C
Column length
m
50
hood
Column inner diameter
Column outer diameter
Column pressure
lm
mm
bar
lL
0.5
0.21
1.41
100
Fig. 1 Sampling and analyzing system
Sample loop volume
assembled, that offers the opportunity to enhance quality of
analysis as well as its reproducibility. In order to keep all
C4 gases and C5 to C16 products in the liquid phase, the
samples were transferred into small containers connected
with DESO (Doubled End Shut Off) connectors to provide
quick coupling and operation (Fig. 1a). Subsequently,
liquid samples were quickly connected to a self constructed
vaporizer heated up to 200 °C, which results in complete
vaporization of product compounds and unconverted
feedstock (Fig. 1). When the pressure in the vaporizer has
reached a constant value of about 2 to 3 bar, the vessel is
opened until a system pressure of 0.5 bar is reached,
thereby flushing the sample loop with the gas-mixture prior
to switching the sampling valve for GC-analysis.
chromatographic analysis of the samples and based on
mass fraction. The conversion of cis-2-butene was deter-
mined by GC-analysis using cyclohexane as internal stan-
dard and was calculated according to Eqs. 3 and 4
m%cisꢂ2ꢂbutene;product
m%cisꢂ2ꢂbutene;feed
Xcisꢂ2ꢂbutene ¼ 1 ꢂ
ð3Þ
ð4Þ
m%cyclohexane ꢀ Area2ꢂbutene
with m%cisꢂ2ꢂbutene
¼
Areacyclohexane
2.5 Preparation of Feed Gas Mixture
Analysis of hydrocarbons was carried out by gas chro-
matography. GC-analyses were performed on a HP 5889
Series I gas chromatograph equipped with a flame ioniza-
tion detector (FID) and a CP Sil Pona CB-column
according to the following method: 2 min. isothermal at
40 °C, 10 °C min-1 heat up to 180 °C, 15 °C min-1 heat
up, 10 min isothermal at 250 °C. Further conditions are
shown in Table 1.
The product distribution of alkylation depends strongly on
the iso-paraffin to olefine ratio (P/O). In the same way, a
constant feed ratio is crucial in evaluation of alkylation
kinetics. Therefore a storage tank of about 20 L volume
was filled with isobutane, cis-2-butene, and cyclohexane in
a fixed ratio. Hence, experiments with identical feed ratios
were performed to gain reliable data on kinetics as well as
on product distribution of the reaction. Cyclohexane, cis-2-
butene and isobutane were filled into the tank in accor-
dance with their partial pressures. First, cyclohexane as
internal standard was transferred into the gas bottle. Sec-
ond, cis-2-butene was pumped into the bottle using an
annular gear pump controlled by a Bronkhorst cori-flow
device (CORI). Finally, isobutane liquefied by increasing
the pressure to 6.5 bar was transferred into the connected
storage tank by pressure difference. The amount of isobu-
tane was controlled by weighing from time to time. The
mixture was checked with gas chromatography (HP GC
5889 Series I). The composition of a 4.76 kg mixture is
shown in Table 2.
Peak correlation was accomplished comparing retention
times of a commercially available isoparaffins-standard
(Supelco) and PONA Kovats retention index.
2.4 Determination of Selectivity and Conversion
In alkylation, a multi component product is formed.
Therefore, the selectivity of an individual alkylation com-
pound is difficult to achieve which requires an enclosed
mass balance. The product composition and herefrom the
selectivity of the compounds of interest were calculated
from the integrated peak-areas obtained by gas
123

Liquid-Phase Isobutane/Butene-Alkylation
1409
nadditive
Table 2 Data of the mixture inside the storage bottle
n%additive
¼
ð5Þ
nadditive þ nisobutane þ n2ꢂbutene
Compound
Weight
(g)
Mass fraction
(wt%)
Mol
amount
(mol)
Molar
ratio(mol%)
The water sensitive ionic liquid catalyst was filled into a
transferable metal cylinder under inert gas atmosphere
(nitrogen) inside a dry box in order to avoid any contact
with moisture. If not noted otherwise, reaction parameters
summarised in Table 3 were chosen.
Cyclohexane
2-Butene
249
306
5.23
6.43
2.96
5.45
3.7
6.7
Isobutane
4203
88.34
72.32
89.6
The starting point of alkylation is set by pressure-
induced introduction of ionic liquid-catalyst from the
connected cylinder into the reactor. The sampling system
comprises quick connectors combined with a filter facili-
tating separation of both phases, catalytic and organic.
When the stirrer is stopped, phase separation takes place
instantaneously, which enables a direct sampling. The
pressure-gap between reactor (6 bar) and atmosphere cau-
ses a flow of liquid sample through an riser pipe into
sample containers. The use of quick connectors minimises
loss of volatile compounds by gas evolution. Since the
sampling volume is low (about 0.5 mL), an almost constant
reaction volume is ensured. So, a high sampling frequency
is obtained. The samples are analyzed by gas chromatog-
raphy as described in 2.3. In case of sulfuric acid-catalyzed
isobutane-butene alkylation a slightly modified sampling
procedure was used, since the liquid samples inevitably
contained small amounts of dispersed sulfuric acid. Thus
the reaction was quenched with small amounts of
butylamine.
The mixture was monitored regularly by GC to guar-
antee that no change in composition occurred.
2.6 Experimental Setup for Liquid Phase Alkylation
2.6.1 Batch Operation
Liquid phase alkylation batch experiments were carried out
in a 250 mL tank reactor with a glass inlet under a pressure
of 6 bar to keep the feed liquid (Fig. 2). The reactor was
provided with a stirrer operating at a constant agitation rate
of 1800 min-1 to guarantee effective dispersion of both
liquid phases. For a typical batch experiment, 70 g of
premixed feed (isobutane and cis-2-butene) including
cyclohexane as internal GC-standard were charged into the
reactor and cooled down to the desired temperature via a
cooling bath tempered by a cryostat (Lauda TK30D). If
additives were used, they were directly passed into the
reactor and cooled to reaction temperature with the feed
prior to catalyst addition. Mol fractions of the additives
always correspond to total amount of substances according
to Eq. 5.
2.6.2 Continuous Operation
The continuous alkylation device (Fig. 3a) consists of a
feed vessel (FV), a product vessel (PV) both made of
stainless steel (SS316), with a capacity of 1 L, and an
alkylation reactor. To charge the feed into the droplet
column reactor, a bronkhorst mini-cori-flow and a micro
annular gear pump from HNP (mzr-7255-fz) with a linear
electric-motor from Faulhaber are installed. Continuous
control of the feed mass was performed by a Bronkhorst
mini-cori flow and a Kern (Kb10K0.05N) balance. From
N2
Mixture
IL
Table 3 Basic reaction parameters for batch isobutane 2-butene
alkylation experiments
S
Parameter
Unit
Value
FV
TIR
Ionic liquid-catalyst
–
[BMIM]Cl–AlCl3
WIR
Mol fraction of AlCl3 (xAlCl
)
–
0.64
13.25
0.4
3
PIR
TIR
Molar paraffin/olefin-ratio (P/O)
Molar ionic liquid/olefin-ratio
Temperature
–
R
–
°C
bar
min-1
-20
6
Fig. 2 Experimental set-up for liquid-phase batch alkylation. S stirrer,
IL transferable container with ionic liquid-catalyst inside, TIR/PIR
temperature/pressure indication registration, FV feed vessel
Pressure
Agitation rate
1800
123

1410
S. Aschauer et al.
polarity of the organic and catalytic phase. The tubular
reactor with a volume of about 30 mL is equipped with two
borosilicate-frits (P4, Robu Glas) on bottom and on top
(Fig. 3b). The ionic liquid catalyst is immobilized by the
bottom frit in the reactor, because of small pore diameters
(from 10 to 16lm) and high ionic liquid viscosities ( ꢄ
20 mPas). The bottom frit also guarantees a fine dispersion
and mixing of feed and catalyst. The upper frit prevents the
ionic liquid from being spouted out of the reactor, if high
feed rates are adjusted. The high reaction enthalpy (up to
-100 kJ mol-1) can be dissipated by an integrated water
cooling coil. For easy operation, the frits are mounted
between dividable connections allowing a facile fitting
exchange.
(a) Flow diagram
N2
Mixture
SI
PI
PI
FV
R
PV
TIR
TIR
TIR
WIR
FIR
3 Results and Discussion
(b) reactor concept
product
3.1 Solubility of Iso-Paraffins and 2-Olefins
in Chloroaluminate-Ionic Liquids
borsilikate-frits
As mentioned above, Langley and Pike [23] as well as
Simpson et al. [22] concluded that hydride transfer from
isobutane to a C?8 -cation leading to the correspondent C8-
isomers is the rate determining step in alkylation. So
contact time of the C?8 -species with isobutane or in other
words the concentration of isobutane in vicinity of the C?8 -
cation is the most important parameter for the catalytic
cycle to TMP-selectivity. Yoo et al. [12] found higher
conversions, when the length of the alkyl side chain of
imidazolium-ionic liquids was increased, but at the same
time TMP-selectivity was not affected. Yoo et al. [12]
concluded that longer alkyl side chains increase the solu-
bility of hydrocarbon reactants in the ionic liquid phase.
This data obtained by Yoo et al. [12] show that the reaction
does not only take place at the interface between ionic
liquid and organic phase but also inside the bulk phase of
the ionic liquid catalyst due to the solubility of hydrocar-
bons within the catalyst phase. The high content of
C8-olefins obtained by Yoo et al. [12], which is not
affected by increasing the alkyl chain on the imidazolium
cation, indicates that hydride transfer to C?8 -carbocations is
still too slow to produce TMPs. Hence, 2-butene solubility
in the catalyst phase is presumably greater than that of
isobutane. In order to prove this theory, solubility mea-
surements of different isoparaffins and 2-alkenes were
performed according to Sect. 2.2
cooling
catalyst-
solution
feed
Fig. 3 Experimental setup for the operation of continuous liquid
phase alkylation
the storage tank, the liquified hydrocarbon mixture was
filled into the feed vessel (200–400 g, depending on
experimental run) at ambient temperature and a pressure of
6.5 bar to keep the feed liquid. The required amount of
catalyst (10 mL) was poured into the reactor. Additives (2
or 4 mL ꢃ 0.18 or 0.36 mmol) were filled into the pipe
below the reactor. At a given set point, the complete feed is
passed into the reactor. The flux was controlled by the cori-
mass flow meter which can be tuned from 1 to 10 g min-1
which equates residence times from 354 to 35 s.
2.6.3 Reactor Concept
The design of the droplet column reactor is closely related
to the bubble column reactor. However, the crucial dif-
ference is that a liquid–liquid process is performed. Here,
products/unreacted feed are rapidly separated from the
ionic liquid catalyst due to the remarkable difference in
both density (Dq ¼ qionicliquid ꢂ qorganic [ 0:5 g cm-3) and
The Kw;i-values depicted in Fig. 4 show, that hydro-
carbon solubility in ionic liquid phase drops with increas-
ing carbon atom number the both for 2-alkene and for
isoalkane. Because the partition coefficients of olefines are
always higher than that of isoparaffins with the same
123

Liquid-Phase Isobutane/Butene-Alkylation
1411
ionic liquids showing high isoalkane solubility in combi-
nation with low olefin solubility is an important task for
research in Lewis acidic ionic liquid-catalyzed alkylation.
The importance of the isobutane/olefin ratio for yield
and alkylate quality in industrial alkylation processes has
already been emphasized by Corma and Martinez [1].
Similiar results for ionic liquid-catalyzed isobutane butene
alkylation were published by Liu et al. [14].
3.2 Batch Alkylation
As already mentioned in Sect. 1, two different initiation
pathways of alkylation are expected to take place
depending on the catalytic system (see Schemes 1, 3). With
Brønsted acids 2-butene is protonated resulting in a sec-
ondary alkyl cation, which starts the catalytic cycle by
hydride abstraction from isobutane. However, the second-
ary cation is also capable to react with another olefin
resulting in oligomerization as an unwanted side reaction.
According to the reaction mechanism, these side reactions
should be avoided at least in the beginning of the alkylation
reaction with tert-butyl halides resulting in different
product distribution. In order to show, if both pathways
influence the course of the reaction in different ways,
experiments with promoters opening either the Brønsted
pathway (H2O) or the Lewis pathway (tert-butyl bromide
and tert-butyl chloride) were performed. The results were
compared with experiments without any promoter (neat
Lewis acidic ionic liquid) as standard and with sulfuric-
acid as catalyst. The reference experiment with the neat
chloroaluminate melt without any promoters was done in
order to gain information about the basic activity of this
melt, brought upon by protic impurities incorporated dur-
ing ionic liquid-synthesis.
Fig. 4 Partition coefficients Kw;i defined as ratio of mass fraction in
ionic liquid to mass fraction in organic phase for iso-paraffins,
n-parrafins and 2-olefins versus the number of C-atoms for
[BMIM][AlCl4] (T = 25 °C)
number of carbon atoms, it is confirmed, that the solubility
of 2-alkenes in the catalyst phase is indeed higher than the
respective isoparaffins. Similiar experimental results were
found by Eichmann [27], who showed an increased solu-
bility of 1-olefins in 1-ethyl-3-methylimidazolium tetra-
chloroaluminates compared to the corresponding n-alkanes
with the same number of carbon atoms. It can therefore be
concluded, that the P/O-ratio in the organic phase does not
reflect the situation in the ionic liquid phase at all. In fact,
P/O-ratios in the reaction phase are only one third of the
original organic phase P/O-ratios. Hence, high P/O-ratios
in the organic phase from 13 drop to values of about 5 in
the catalytic phase. So it is no suprise, that in industrial
alkylation with ionic liquids feed-P/O values from 500 up
to 1000 to 1 are used, to suppress secondary and side
reactions [19]. It seems, that ionic liquid phase-P/O-values
of[100 are needed in order to prevent side and secondary
reactions and yield good alkylate quality. Isobutane solu-
bility seems to be the key value for chloroaluminate ionic
liquid catalyst alkylation. Therefore, the development of
3.2.1 Effect of Different Additives on Catalyst Activity
In Fig. 5, cis-2-butene conversions are shown as a function
of reaction time for sulfuric acid-catalyzed alkylation as
Fig. 5 2-butene conversion
versus reaction time for
[BMIM]Cl-AlCl3 ionic liquid
with different promoters
(P/O = 13/1, TReactor
=
-20 °C, mFeed = 70 g,
xAdd = 0.2 mol%. For sulfuric
acid: TReactor = 5 °C and
xAdd = 2.4 mol%.)
123

1412
S. Aschauer et al.
well as for alkylation catalyzed by the neat Lewis acidic
The slightly higher activity of tert-butyl bromide over tert-
butyl chloride is to be explained by the difference of bond
dissociation energies of both tert-butyl halides. According
to literature, the bond dissociation energy of the tert-butyl
halide is 337.1 kJ mol-1 for the carbon–chlorine bond and
276.4 kJ mol-1 for carbon–bromine bond [31]. So due to
the weaker carbon–bromine bond tert-butyl bromide reacts
faster with the Lewis acidic ionic liquid than tert-butyl
chloride.
chloroaluminate ionic liquid [BMIM]Cl-AlCl3 (xAlCl
¼
3
0:64) and the same ionic liquid catalyst promoted with the
same molar amounts of water, tert-butyl chloride and tert-
butyl bromide (xAdditiv ¼ 0:2 mol%).
Fig. 5 depicts, that the activity of a pure chloroalumi-
nate melt [BMIM]Cl-AlCl3 limits the 2-butene conversion
at a maximum value of about 25%. This relative low
activity indicates that only traces of proton sources were
incorporated into the chloroaluminate catalyst during
preparation and reaction procedure, and thus only a small
amount of active species (tert-butyl cation) is generated
according to Scheme 1. This result is in agreement with the
finding, that neat chloroaluminate melts are catalytically
not active [30]. Addition of 0.12 mol% water to the feed,
which represents 0.24 mol% catalytically active protons
(corresponding to Scheme 3), results in an enhanced cat-
alytic activity of the ionic liquid-catalyst compared to the
standard Lewis acidic melt. For a better valuation, some
experiments were also performed with sulfuric acid as
standard catalyst for alkylation and for comparison of the
results with ionic liquid systems used in this publication.
Since under the conditions usually applied for isobutane-
butene alkylation (see Table 3) no reaction occurred in the
investigated time frame of 90 m, reactions with sulfuric
acid were performed under slightly changed reaction con-
ditions. The reaction temperature was increased to 5 °C
and the molar amount of the catalyst was 2.4 instead of
0.2 mol%.Other conditions such as pressure and P/O molar
ratio are identical to the promoted [BMIM]Cl-AlCl3 sys-
tems. Results displayed in Fig. 5 show, that despite of
higher reaction temperature sulfuric acid is not capable to
fully convert 2-butene in the investigated reaction time
range.
Figure 6 shows the effect of different amounts of tert-
butyl halogenide on the activity of the ionic liquid-catalyst.
By adding only small amounts (0.05 mol%) of this pro-
moter, 2-butene conversion plateaus on maximum values
of about 40% after a reaction time of 15 min. Compared to
isobutane/butene alkylation conducted with the neat ionic-
liquid catalyst, where only 25% conversion of 2-butene
was measured, activity of the catalyst could be increased
considerably upon tert-butyl bromide addition. The step-
wise increase of the tert-butyl bromide mol fraction to a
value of 0.17% yields a considerable acceleration of the
reaction rate reaching nearly 90% conversion in 4 m
(compare Fig. 6 and Table 4). Unsurprisingly, a further rise
of branched alkyl halogenide promoter to a molar amount
of 0.5 mol% speeds up the alkylation and further reduces
the time needed for complete 2-butene conversion to the
halve (Fig. 6). Since similar effects on the reaction rate
were observed for tert-butyl chloride, Fig. 6 only depicts
the experimental data for tert-butyl bromide. In order to
further illustrate the profound influence of different addi-
tives on the reaction rate, rate constants k were determined.
For a simplified kinetic analysis, a pseudo first order
approach was assumed for the olefin. This assumption is
also made in the literature [32–34]. For a first order, the
rate constant k can be determined by the following
equation:
When the active species of the TMP-production cycle is
generated by tert-butyl halogenide promoted ionic liquids,
2-butene conversion reaches 100% in less than 5 m (tert-
butyl bromide) and less than 10 min (tert-butyl chloride).
ꢂ lnð1 ꢂ Xcisꢂ2ꢂbutene
Þ
k ¼
ð6Þ
t
Fig. 6 2-butene conversion
versus reaction time for
different tert-butyl bromide
molar amounts (P/O = 13/1,
TReactor = -20 °C,
mFeed = 70 g)
123

Liquid-Phase Isobutane/Butene-Alkylation
1413
0.2 mol% promoted ionic liquid catalysts. The different
activities of tert-butyl halides and of water promoted ionic
liquid catalyst systems, are a result of the different initia-
tion pathways discussed in Sect. 1 We assume, that organic
promoters, initially dissolved in the organic phase, are in-
stantaneuosly converted into tert-butyl cations when con-
tacted with the Lewis-acidic ionic liquid phase.
Consequently, the active species are almost immediately
provided for the reaction. In case of the protic catalyst, we
assume that water as a polar molecule is partly absorbed
during the formation of highly acidic protons into the bulk
ionic liquid. So, only a fraction of active species is present
at the surface and available for the C4-feed. Since forma-
tion of the tert-butyl cation, according to Scheme 1
(Brønsted catalysis) requires an additional 2-butene-mole-
cule, generation of the active species is inevitably more
slowly compared to alkyl halogenide promoted alkylation,
where the tert-butyl cation is instantaneously liberated
from one haloalkane-molecule. In the case of sulfuric acid,
41 mol% of this Brønsted acid (together with a 25 °C
higher reaction temperature) are needed to reach rates
comparable to alkyl-halogenide promoted ionic liquid-
catalysis. Hence, regarding tert-butyl halides, the experi-
mental data clearly show that the reaction can be suc-
cessfully performed without highly Brønsted acidic
catalysts. Organic compounds, which act as tert-butyl
cation sources are potential alkylation catalysts as long as
they form the catalytically active species by reaction with a
suitable Lewis acid.
Table 4 Apparent rate constant k as function of molar tert-butyl
bromide-fraction
Furthermore values for tert-butyl chloride, water and sulfuric acid are
given
Figure 6 and Table 4 show the influence of different
amounts of tert-butyl bromide on the rate constant of iso-
butene/2-butene alkylation. For comparison, selected val-
ues for other additives and sulfuric-acid-catalyzed
alkylation are also shown in Table 4. The reaction with
tert-butyl halides (chlorides as well as bromides) as pro-
moter is accelerated by up to two orders of magnitude.
Concerning reaction rate, there is no need to increase tert-
butyl bromide-concentration above a value of about
0.2 mol%, since upon addition of higher tert-butyl bromide
amounts similar values of the reaction rate were obtained
(varying from 28.6 9 103 to 46.1 9 103 s-1, compare
Table 4). With Brønsted-acid-catalysis (water-addition or
catalysis by 2.4 mol% sulfuric acid), the reaction rate was
only quintupled from 2 9 104 s-1 (value with no additives;
pure Lewis-acidic system) to 1 9 103 s-1, though in the
case of sulfuric acid, a ten-fold higher concentration of
active species (protons) is present in the reaction. In order
to mimic industrial alkylation conditions even better, an
experiment with 41.4 mol% of sulfuric acid at a tempera-
ture of 5 °C was performed and the calculated rate constant
is also included in Table 4.
3.2.2 Effect of Different Additives on Catalyst Selectivity
With regard to the alkylation process, increasing the speed
of the reaction is not the main objective. The major concern
is the product distribution of the gained alkylate. Here
trimethylpentanes (TMPs) are of key interest because they
act as RON-boosters. Corresponding to mechanistic con-
siderations the TMP-content should be increased by the use
of tert-butyl halides (suppression of cracking and oligo-
merization reactions).
Figure 7 shows product distributions in wt% obtained
for three different promoters, a reference alkylate (without
promoter) and a sulfuric-acid catalyzed batch alkylation. In
the Lewis acidic catalytic reference system (ionic liquid
without additive) the C5 fraction dominates with isopen-
tane as the major component. But also, the product alkylate
contains relative high amounts of low ends and of heavy
ends with 10 carbon atoms or more. The C5-fraction as well
as the low and heavy ends strongly indicate that secondary
reactions like oligomerization and cracking take place
preferentially in this neat ionic liquid, whose activity relies
on water-impurities. The desired C8 fraction is only pro-
duced just under 10%. In this fraction, TMPs are obtained
Comparing the experimental data, tert-butyl halides
show a higher activity than the protic additive water as well
as sulfuric acid. This holds as long as similar catalytically
active concentrations are taken into account: The rate
constant in case of the alkylation experiment with
41.4 mol% H2SO4 can certainly not be compared with
123

1414
S. Aschauer et al.
Fig. 7 Alkylate- and
C8-composition in wt% for
different promoters and
promoter-concentrations
(P/O = 13/1, TReactor
=
-20 °C, mFeed = 70 g; final
composition of products
comprising 5 to 16 carbon
atoms at full or maximum
2-butene conversion)
as minor compounds (14%) and within the TMP-fraction
2,3,4-TMP is the only isomer present. The main C8-com-
ponent formed in a significant amount is 3,4-dimethyl-
hexane (DMH). The appearance of this C8-isomer is
attributed to the reaction of cis-2-butene with sec-butyl
cations according to Scheme 5. Such secondary cations are
easily formed due to highly acidic protons which are
present because of small water impurities in chloroalumi-
nate ionic liquids (compare Scheme 3).
depending on the alkyl bromide content. However, the main
fraction is still C9 comprising about 30 wt%. When the
promoter content is enhanced, the results illustrate that the
DMH-fraction is reduced whereas TMP generation starts to
be the favorite reaction pathway in C8-formation. The
alkylation leads to 2,2,4-TMP as the predominant TMP iso-
mer. On the other hand, the both 2,3,3- and 2,3,4-TMP are not
affected in terms of their fraction. The increase of 2,2,4-TMP
is assumed to be the consequence of self alkylation corre-
sponding to Scheme 6.
As pictured in Fig. 7, addition of low amounts of water
(0.12 mol% = 0.24 mol% H?) increases the amount of
heavy ends drastically (C9, C10?) to over 50% (compared
to about 15 wt% with neat chloroaluminate ionic liquids)
while the content of the desired C8-fraction is approxi-
mately doubled. However within the C8-fraction, the dis-
tribution is shifted from DMHs towards the well desired
TMPs. Again, oligomerisation and cracking reactions are
favored in the course of the reaction time. The product
distribution indicates that the Brønsted catalyst accelerates
side and secondary reactions. The shift to TMPs in the C8-
fraction can be ascribed to a higher degree of the actual
alkylation reaction between isobutane and olefin due to the
enhanced concentration of active catalyst. The distribution of
TMP isomers is a consequence of intermolecular hydride and
methyl shifts within the 2,2,3-TMPyl-cation which are well
known in literature [35]. However, self alkylation has also to
be considered for the yield of 2,2,4-TMP. For tert-butyl
halide as promoter, the product distribution as a result of
varying the promoter content was studied. Figure 7 also
shows the composition of the C8-fraction as function of the
molar tert-butyl bromide-content. Using tert-butyl bro-
mide as promoter, the C8-fraction increases up to 26 wt%
Here, 2,2,4-TMP is generated by the reaction of tert-
butyl cation and isobutene, which is supposed to be the
species formed during deactivation of the promoted cata-
lyst system (compare Scheme 10). However, consecutive
internal hydride and methyl shifts in the course of the
alkylation of 2-butene have to be taken also into account as
another source of this isomer. tert-butyl bromide promoted
ionic liquid catalysis leads to a significant formation of C9-
fraction. This increase results from preferential formation
of the C9-isomer 2,2,5-trimethylhexane (2,2,5-TMH) as
dominating compound within the C9-fraction. We suppose
that 2,2,5-TMH is formed due to the reaction of isopentyl
cation with 2-butene followed by a consecutive hydride
transfer. This is confirmed by high amounts of isopentane
observed in the C5-fraction. Cracking of C8-compounds
could be responsible for the generation of isopentyl cations
(yielding propene as by-product) required to form 2,2,5-
TMH. Such a pathway is described by Zuazo [36]
(Schemes 7 and 8).
Furthermore in 1985, Kohlheim showed that alkylation
of isopentane with 2-butene in HF yields 2,2,5-TMH as
major product [37].
+
+
Scheme 5 Formation of 3,4-dimethylhexyl cation
Scheme 6 Formation of 2,2,4-trimethylpentyl cation
123

Liquid-Phase Isobutane/Butene-Alkylation
1415
reason for the unwanted shift of the product composition
towards C10? can be found in the high reactivity of iso-
butene, which is produced during deactivation of the cat-
alysts and which is very prone to polymerization [1, 38].
+
Scheme 7 Cracking pathway to propene and isopentyl cation
3.3 Continuous Liquid-Phase Alkylation
+
3.3.1 Effect of Additive
Scheme 8 Formation of 2,2,5-trimethylhexyl cation
In continuous operation, the same types of tert-butyl
sources (tert-butyl bromide, tert-butyl chloride and MTBE
additionally) and water were used as in batch operation.
The promoters were added one-time at the beginning of
each experiment in order to investigate the activity and
deactivation behaviour of the catalytic system. Here,
0.5 mol% of promoter was charged to the process, since
the data from the batch experiments showed the best result
for a high TMP content with this amount of additive. Yet in
all experiments the same molar amounts of promoter were
added in order to generate the same number of catalytically
active centres.
In tert-butyl halogenide supported ionic liquid-catalysis
the reaction rate is so fast, that high amounts of initially
formed 2,2,3- TMPyl-cations (which rearrange) cannot all
be saturated by a hydride transfer from isobutane. These
highly reactive C?8 -carbocations, which are not converted
into TMPs, thus react with another 2-butene-molecule
leading to high ends, which are cracked into low ends in
the course of the reaction. Furthermore, the low and heavy
ends confirm the results reported by Langley and Pike [23]
and Simpson et al. [22] for H2SO4 catalyzed alkylation,
that the hydride transfer from isobutane to C?8 yielding the
correspondent C8 paraffin is the rate determining step in
ionic liquid catalyzed alkylation as well. The rapid reaction
of C?8 -carbocations with 2-butene-molecules is enhanced
by the high solubility of 2-butene in the ionic liquid, which
is significantly higher than the solubility of isobutane in the
catalyst phase (compare Sect. 3.1). Our experiments show,
that tert-butyl halogenide promoted IL-catalysts produce
TMPs and C10?-compounds simultaneously from the first
moment of reaction, so that these compounds are found
already in the first samples (taken usually after 30 or 70 s).
In sulfuric acid or H2O-supported reactions, the synthesis
of C10? and TMPs is delayed. Here the formation of TMPs
precedes the C10?-synthesis, indicating that in this system
the hydride transfer can cope with C?8 -synthesis. Another
In Fig. 8, the experimental data are summarized with
regard to conversion of 2-butene with time on stream. In all
experiments deactivation occurs fast. The Lewis acidic
ionic liquid catalyst without any promoter, yielded a
maximum conversion of about 25 wt%, which is observed
only at 5 m time on stream. The conversion dropped rap-
idly to 5% and remained constant with time on stream.
Addition of water showed a very similar behaviour com-
pared to promoter free catalysis. The conversion drops also
quickly with time on stream. Switching to tert-butyl halides
results in a fast deactivation as well. But conversions are
significantly higher in the beginning, independent from the
tert-butyl halide used. This indicates that the generated
tert-butyl cations are directly available for alkylation,
XinitialꢂXfinal
Fig. 8 2-butene conversion in continuous operation versus reaction
time for different promoters (P/O = 13/1, Treactor = 22 °C,
Þ; tinitial ¼ 5 m except for water (tinitial = 10 min) (with-
out promoter = 2%/min; water = 1.3%/min; MTBE = 1.9%/min; tert-
butyl chloride = 2.1%/min; tert-butyl bromide = 2.2%/min)
Xinitialꢀðtinitialꢂtfinal
m_ Feed ¼ 4 g min-1, Vionic
= 10 mL, xAlCl ¼ 0:65; Vpromoter
¼
liquid
3
2 mL). Relative decrease of conversion with time on stream
123

1416
S. Aschauer et al.
Fig. 9 Product distribution at
different reaction times for
different promoters
(P/O = 13/1, TReactor = 22 °C,
m_ Feed ¼ 4 g min-1
,
Vionic liquid = 10 mL,
xAlCl ¼ 0:65; Vpromoter ¼ 2 mL)
3
otherwise the tert-butyl cations must be formed through
Scheme 1 involving more reaction steps and byproducts
(n-butane). The influence of the tested additives to product-
distribution with time on stream are displayed in Fig. 9.
Without promoter, the TMPs increased with time on
stream and decreasing conversion resulting in a slightly
higher alkylate quality. But, the different promoters
enhance the selectivity to TMPs at the beginning of the
reaction in the order tert-butyl bromide [ tert-butyl chlo-
ride[MTBE[water, which indicates a higher amount of
active species. This corresponds to a decrease in the
binding energy of the tert-butyl cations in the order tert-
butyl bromide \ tert-butyl chloride \ MTBE. We assume
that the drop of TMPs is the consequence of deactivation,
in which the branched alkyl cation reacts to isobutene.
However, if this deactivation pathway is true, the formation
of isobutene inherently releases a proton, which has to be
regarded as highly Brønsted acidic, since the corresponding
anion ½AlCl4ꢁꢂ has to be considered as extremely weak
base in the ionic liquid catalyst. Therefore, the activity
maintained at a certain level meaning that the non-Brønsted
catalysts, responsible for the alkylation in the beginning of
the process, transforms into a Brønsted catalyst in the
course of the reaction. Because of TMP formation, we
suppose that the Brønsted acidic catalyst shows the same
acidity range as sulfuric acid. In all experiments a high
amount of low-ends (above 40 wt%) was detected. Within
this C5 to C7 fraction isopentane is by far the dominant
compound. Cracking of C8-products by b-scission seems to
be likely in accordance to the results obtained by Zuazo
[36]. But propene, which should be formed as co-product,
was only found to a minor degree. Since propene is a very
reactive compound under alkylation conditions one should
assume that propene is released in the ionic liquid and
reacts rapidly with isobutane to form C7 or dimerizes to C6.
The isopentyl cation formed through cracking of C?8 is
converted to isopentane via hydride transfer. This pathway
is supported by the high amount of 2,2,5-TMH (around 10
wt%) in the alkylate and by Kohlheim and Mortikov et al.
who independently showed the formation of 2,2,5-TMH as
main product by secondary alkylation of isopentane and
2-butene [37, 39]. However, high isopentane values may
also be explained by cracking of the C10? fraction as
described by Smirnova et al. [4]. The formation of these
products is considered to be a result of multiple additions
of olefin to the carbocation, which is enabled by slow
hydride transfer kinetics compared to the rate of oligo-
merisation [22]. Scheuch [40] described the formation of
isopentane as an indicator for self-alkylation. Following
this conclusion and according to the high selectivity for
2,2,4-TMP within the TMP fraction, self-alkylation plays a
major role in tert-butyl halide promoted alkylation and
therefore product formation (Scheme 6).
Also formation of 2,3,4-TMP, which shows the second
highest content within the TMP-fraction, is attributed to a
methyl-shift from 2,2,4-TMP?, whereas 2,2,3-TMP and
2,3,3-TMP are formed via 2,2,3-TMP? (Scheme 9) [35].
Since Scheuch [40] and Kohlheim [37] argue that in
micelle systems hydride shift ability of isobutane is
increased, we suppose a micellar environment in the con-
tinuous reactor. Kohlheim also observed a high selectivity
~ H
~ CH3
~ CH3
~ H
~ CH3
Scheme 9 Reaction pathway to different trimethylpentyl cations
123

Liquid-Phase Isobutane/Butene-Alkylation
1417
to isopentane, C9 and other light ends in micelle activated
HF-systems.
product gas. Indeed, a white AgCl-precipitate was formed
when the gaseous product was bubbled through an aqueous
AgNO3-solution. The detection of isobutene in the liquid
alkylate and hydrochloric acid in the gaseous reactor
effluent strongly indicates deactivation of the promoted
ionic liquid-catalyst by Schemes 10 and 11.
3.3.2 Effect of Liquid Feed-Flow
As expected, conversion increases with decreasing mass
flow and increasing residence time (Fig. 10). This is
attributed to higher contact times and smaller organic
droplets in the ionic liquid. But also deactivation occurs
more rapidly with increasing mass flow (4 g min-1). Pre-
liminary experiments with a magnetic impeller showed that
the rate of deactivation is also mainly connected to the
droplet size, because smaller droplets decrease the rate of
deactivation significantly. At low interfacial area (larger
droplets) between organic and catalyst phase, the catalytic
cycle (Scheme 2) maybe could not be completed because
of slow hydride-transfer.
3.3.3 Influence of the Mode of Promoter Addition
Since rapid deactivation is observed in experiments where
the additive is added only once at the beginning, the
influence of different promoter concentrations added con-
tinuously into the ionic liquid mixture was studied
(Fig. 11a). While conversion of 2-butene dropped very
quickly to a level around 5–10 wt% with time on stream in
case of one time addition, in continuous mode the con-
version dropped slower and kept constant for 0.4 mol% at
about 55–60% and for 0.2 mol% promoter concentration
about 20–30% conversion.
When relative high amounts of promoter were used,
isobutene was observed in considerable amounts. Isobutene
results from deprotonation of the isobutyl cation, which is
supported by Sievers [41] (see Scheme 10). Now, isobutene
is also an olefin source and is able to react with isobutyl-
cations to form 2,2,4-TMP? directly. This self-alkylation is
often observed and explains the high amount of 2,2,4-TMP
within the TMP-fraction mentioned in almost every pub-
lication on alkylation, especially for HF [1].
With time on stream, the TMP content within alkylate
shows very similar behaviour as 2-butene conversion
(compare Fig. 11a, b). In contrast to one time promoter
addition, the decrease of TMP in continuous promoter
addition is significantly slower indicating that the active
branched cations are still formed due to halide abstraction
caused by the Lewis acidic ionic liquid. However, the
decline in activity gives hints to a steady-state like satu-
ration of the catalytic system. Whether a stable alkylation
catalytic system is really obtained has to be proved by
experiments with higher time on stream.
Deprotonation releases a highly acidic proton, which
must either be recaptured by a basic organic compound in
the reaction mixture (alkene) or react with ½AlCl4ꢁꢂ
according to Scheme 11, since a separate proton is non-
existing. The resulting, hypothetical HAlCl4 acid [42]
would decompose instantaneously upon its formation,
yielding HCl and AlCl3.
3.3.4 Industrial Application Concept
If this assumption is true catalyst deactivation is caused
by release of gaseous hydrochloric acid together with the
The retrofitted PetroChin (Ionikylation) refinery [19]
demonstrates that classic alkylation catalysts (HF, H2SO4)
can easily be substituted by Lewis acidic chloroaluminate
ionic liquids. As shown in this paper, the use of ionic
liquids in combination with tert-butyl halide promoters
leads to higher reaction rates and TMP-production may
increase significantly.
Figure 12 shows the respective concept of an industrial
alkylation plant including continuous addition of the tert-
+ H
Scheme 10 Formation of isobutene
Cl
Al
Cl
Cl
Al
Cl
Fig. 10 2-butene conversion versus reaction time for different feed-
rates varying from 2 to 4 g min-1 (P/O = 13/1, TReactor = 22 °C,
+ HCl
H
Cl
Cl
Cl
Vionic
tert-butyl chloride)
= 10 mL, xAlCl ¼ 0:65; Vpromoter ¼ 2 mL, promoter =
liquid
3
Scheme 11 Decomposition of hypothetical H[AlCl4]
123

1418
S. Aschauer et al.
(a) Convnvnersion
Alkyl chlorides and HCl in the effluent stream of the
alkylation reactor are separated by caustic wash (NaOH),
yielding about 9 kg NaCl. The current price of HCl is
around 0.1€/kg, which equals 5.5€/t of alkylate. Hence, the
additional production costs per liter of alkylate (density
700 kgm-3) are only 0.4 cents, which should be reasonably
low to justify the use of tert-butyl chloride.
4 Conclusion
Our measurements confirm, that it is possible to conduct
isobutane/butene-alkylation not only with strong Brønsted-
acid catalysts, but also with pure Lewis-acidic melts sup-
ported by tert-butyl halides. Comparing the experimental
data, tert-butyl halides show a higher activity than the
protic additive water as well as sulfuric acid. However, the
alkyl halide promoted chloroaluminate ionic liquid catalyst
can only be considered as an ideal Lewis-acidic catalyst in
the first instance. With increasing deactivation of the sys-
tem, e. g. with increasing decomposition of the tert-butyl
cation into H? and isobutene, a Brønsted-acidic catalyst
system is formed to some extent. So that for the description
of the real catalyst system two regimes (a Lewis-acidic and
a Brønsted-acidic one) have to be considered. Protons,
which are released upon deactivation of tert-butyl haloge-
nide supported chloroaluminate ionic liquid-catalyst, are
discharged via HCl-formation.
(b) TMP-amount in wt.-%
Since acceleration of the alkylation in haloalkane sup-
ported IL-catalysis leads to a rapid formation of C?8 -carb-
ocations, the slow hydride shift from isobutane to the
C?8 -cation is not able to quench the C?8 -cations. Further
addition of olefines will occur which results in a product
distribution dominated by high ends (C9 and C10?).
Because deactivation of the catalytic system observed in
continuous alkylation leads to isobutene release, polymer-
ization of this molecule could further account for the
observed selectivities. The TMP-selectivity, however,
could be increased continuously upon increasing the pro-
moter concentration. The high fraction of high ends is
indicative of low isobutane solubility within the catalyst-
phase, which was also confirmed experimentally.
Fig. 11 2-butene conversion and TMP-selectivity with reaction time
and different modes of promoter addition (P/O = 13/1, TReactor
=
22 °C, m_ Feed ¼ 4 min-1, Vionic
= 10 mL, xAlCl ¼ 0:65; pro-
liquid
3
moter=tert-butyl chloride)
Acknowledgment We would like to thank the German Research
Foundation for supporting the project (JE257/14-1).
Fig. 12 Industrial alkylation concept using tert-butyl chloride as
promoter (1 tert-butyl chloride generation; 2 alkylation; 3 caustic
wash; 4 distillation)
References
butyl chloride promoter. A small isobutene stream is fed
into a pre-reactor, where it is converted with neat HCl to
tert-butyl chloride. A similiar process for promoter-syn-
thesis can be found in [24]. For the production of 1 t of
alkylate, only 14 kg of promoter (5.5 kg HCl) are needed.
1. Corma A, Martinez A (1993) Catal Rev Sci Eng 35(4):483
2. Weitkamp J, Traa Y (1999) Catal Today 49(1–3):193
¨
3. Feller A, Lercher JA (2004) In: Gates BC, Knozinger H (eds)
Advances in catalysis, vol 48. Academic Press, Amsterdam,
pp 229–295. http://dx.doi.org/10.1016/S0360-0564(04)48003-1
123

Liquid-Phase Isobutane/Butene-Alkylation
1419
4. Smirnova MY, Toktarev AV, Ayupov AB, Echevsky GV (2010)
Catal Today 152(1–4):17
5. Costa BOD, Querini CA (2010) Appl Catal A Gen 385(1–2):144
6. Feller A, Barth JO, Guzman A, Zuazo I, Lercher JA (2003) J
Catal 220(1):192
26. Rosenbach NJ, Mota CJA (2005) J Braz Chem Soc 16(4):691
27. Eichmann M (1999) Zweiphasige Dimerisierung von Propen und
1-Buten mit ionischen Flu¨ssigkeiten. Ph.D. thesis, RWTH
Aachen, Aachen
¨
28. Weitkamp J, Traa Y (1997) In: Ertl G, Knozinger H, Weitkamp J
7. Sievers C, Zuazo I, Guzman A, Olindo R, Syska H, Lercher JA
(2007) J Catal 246(2):315
¨
8. Klingmann R, Josl R, Traa Y, Glaser R, Weitkamp J (2005) Appl
(eds) Handbook of heterogeneous catalysis. Wiley-VCH, Wein-
heim, pp 2039–2069
29. Weitkamp J (1980) In: Imelik B, Naccache C, Ben Taarit Y,
Vedrine JC, Coudurier G, Praliaud H (eds) Studies in surface
science and catalysis: catalysis by zeolites. Elsevier, Amsterdam,
pp 65–75
30. Nenitzescu CD, Cantuniari IP (1933) Berichte der deutschen
chemischen Gesellschaft (A and B Series) 66(8):1097
31. Bai H, Ault BS (1991) J Phys Chem 95(8):3080
32. Simpson MF, Wei J, Sundaresan S (1996) Ind Eng Chem Res
35(11):3861
33. Langley RJ (1969) The kinetics of alkylation of isobutane with
propylene using a sulfuric acid catalyst. Ph.D. thesis, Louisiana
State University and Agricultural & Mechanical College, Baton
Rouge
Catal A Gen 281(1–2):215
9. Josl R, Klingmann R, Traa Y, Glaser R, Weitkamp J (2004) Catal
¨
Commun 5(5):239
10. Chauvin Y, Hirschauer A, Olivier H (1994) J Mol Catal
92(2):155
11. Huang CP, Liu ZC, Xu CM, Chen BH, Liu YF (2004) Appl Catal
A Gen 277(1-2):41
12. Yoo K, Namboodiri VV, Varma RS, Smirniotis PG (2004) J Catal
222(2):511
13. Kumar P, Vermeiren W, Dath JP, Hoelderich WF (2006) Energy
Fuels 20(2):481. http://dx.doi.org/10.1021/ef050264c
14. Liu Y, Hu R, Xu CM, Su H (2008) Appl Catal A Gen
346(1–2):189
34. Lee LM (1973) Alkylation of isobutane with butene-1 in sulfuric
acid: a kinetic study. Ph.D. thesis, Cornell University, Ithaca
35. Feller A (2003) Reaction mechanism and deactivation pathways
15. Zhang J, Huang CP, Chen BH, Li J, Li Y (2008) Korean J Chem
Eng 25(5):982
¨
¨
in zeolite. Ph.D. thesis, Technischen Universitat Munchen,
16. Bui TTL, Korth W, Aschauer SJ, Jess A (2009) Green Chem
11(12):1961
Mu¨nchen
36. Zuazo I (2004) Deactivation routes in zeolite catalyzed isobutane/
2-butene alkylation and regeneration procedures. Ph.D. thesis,
¨ ¨
17. Olah GA, Mathew T, Goeppert A, Torok B, Bucsi I, Li XY,
Wang Q, Marinez ER, Batamack P, Aniszfeld R, Prakash GKS
(2005) J Am Chem Soc 127(16):5964
¨
¨
¨
Technischen Universitat Munchen, Munchen
18. Tang S, Scurto AM, Subramaniam B (2009) J Catal 268(2):243
19. Liu ZC, Zhang R, Xu CM, Xia R (2006) Oil Gas J 40(104):52
20. Welton T (2010) Ion pairs or ion solvation: dissolving salts in
ionic liquids. In: Talk at Conference: 2nd Asia Pacific Conference
on Ionic Liquids and Green Processes; Bayshore Hotel, Dalian,
China; 7. September 2010
37. Kohlheim KE (1985) Zur Wirkungsweise amphiler Verbindungen
als Phasen-Transfer-bzw. Micellarkatalysatoren bei der Isoalkan/
Olefin-Alkylierung in wasserfreiem Fluorwasserstoff. Ph.D. the-
¨
sis, Universitat Greifswald, Greifswald
38. Albright LF, Spalding MA, Kopser CG, Eckert RE (1988) Ind
Eng Chem Res 27(3):386
21. Schmerling L (1953) Ind Eng Chem 45(7):1447
22. Simpson MF (1996) The zeolite-catalyzed alkylation of isobutane
with butene. Ph.D. thesis, Princeton University, Princeton
23. Langley RJ, Pike RW (1972) AIChE J 18(4):698
24. Lacheen H, Timken HkC (2009) Production of low sulphur
alkylate gasoline fuel. Patent (WO/2009/058517); Chevron
U.S.A. Inc
25. Elomari S (2009) Alkylation process using an alkyl halide pro-
moted ionic liquid catalyst. Patent (US 7495144 B2); Chevron
U.S.A. Inc
39. Mortikov ES, Zen’kovskii SM, Mostovoi NV, Kononov NF,
Golomshtok LI, Minachev KM (1974) Russ Chem Bull 23(7):
1469
40. Scheuch E (1991) Micellaktivierte Alkylierungen von Butenen
bzw. Buten/Propen-Gemischen mit Isobutan in wasserfreiem Flu-
¨
orwasserstoff. Ph.D. thesis, Universitat Greifswald, Greifswald
41. Sievers C, Liebert JS, Stratmann MM, Olindo R, Lercher JA
(2008) Appl Catal A Gen 336(1–2):89
42. Sikorska C, Freza S, Skurski P (2010) J Phys Chem A 114(5):
2235
123