Ionic liquid-catalyzed alkylation of isobutane with 2-butene

Article abstract of DOI:10.1016/j.jcat.2003.11.018

A detailed study of the alkylation of isobutane with 2-butene in ionic liquid media has been conducted using 1-alkyl-3-methylimidazolium halides-aluminum chloride encompassing various alkyl groups (butyl-, hexyl-, and octyl-) and halides (Cl, Br, and I) on its cations and anions, respectively. The emphasis has been to delineate the role of both cations and anions in this reaction. The ionic liquids bearing a larger alkyl group on their cation ([C₈mim]) displayed relatively higher activity than a smaller one ([C₆ or C₄mim]) with the same anionic composition, due to the high solubility of reactants in the former. Among the ionic liquids with different halide groups, bromides ([C₈mim]Br-AlCl₃) showed outstanding activity, because of the higher inherent acidity relative to others. From the ²⁷Al NMR study, a major peak at ～99.5 ppm corresponding to [AlCl₃Br]^- (～99.5 ppm) was observed. Moreover, the anion showed a strong acidity based on FT-IR characterization; the largest peak related to acidity (1570 cm^-1) was detected. Under various composition conditions, catalytic activity and amount of TMPs increased with concentration of anion. This is mainly attributed to a higher amount of strong acid ions [Al₂Cl₆Br]^- which can react with hydrogen atoms at the 2-position of an imidazolium ion to form Bronsted acid. However, the ionic liquid with strong acidity (X=0.58) deactivated rapidly due to a higher sensitivity to moisture, causing decomposition. Under various reaction temperature conditions, optimum catalytic activity was observed at 80°C. The result is also attributed to the effect of anion composition. The strong acidic anion increased with temperature. However, at higher reaction temperatures (120°C), the ionic liquid showed a lower activity and TMP selectivity, since the solubility and Bronsted acid sites were reduced by decomposition of imidazolium ions. The selected ionic liquid sample ([C₈mim]Br-AlCl₃) was compared with one of the standard commercial catalysts, sulfuric acid. Under optimum experimental conditions, it was observed that both catalysts showed comparable catalytic behavior. However, ionic liquid showed higher activity, and lower TMP selectivity due to a more acidic nature and a lower amount of Bronsted acid sites, respectively.

Full text of DOI:10.1016/j.jcat.2003.11.018

Journal of Catalysis 222 (2004) 511–519
www.elsevier.com/locate/jcat
Ionic liquid-catalyzed alkylation of isobutane with 2-butene
a
b
b,∗
Kyesang Yoo, Vasudevan V. Namboodiri, Rajender S. Varma, and
a,∗
Panagiotis G. Smirniotis
a
Department of Chemical and Material Engineering, University of Cincinnati, Cincinnati, OH 45221-0012, USA
b
Clean Processes Branch, National Risk Management Research Laboratory, US Environmental Protection Agency, MS 443, Cincinnati, OH 45268, USA
Received 19 May 2003; revised 15 August 2003; accepted 6 November 2003
Abstract
A detailed study of the alkylation of isobutane with 2-butene in ionic liquid media has been conducted using 1-alkyl-3-methylimidazolium
halides–aluminum chloride encompassing various alkyl groups (butyl-, hexyl-, and octyl-) and halides (Cl, Br, and I) on its cations and anions,
respectively. The emphasis has been to delineate the role of both cations and anions in this reaction. The ionic liquids bearing a larger alkyl
group on their cation ([C mim]) displayed relatively higher activity than a smaller one ([C or C mim]) with the same anionic composition,
8
6
4
due to the high solubility of reactants in the former. Among the ionic liquids with different halide groups, bromides ([C mim]Br–AlCl )
8
3
2
7
showed outstanding activity, because of the higher inherent acidity relative to others. From the Al NMR study, a major peak at ∼ 99.5 ppm
−
corresponding to [AlCl Br] (∼ 99.5 ppm) was observed. Moreover, the anion showed a strong acidity based on FT-IR characterization;
3
−1
the largest peak related to acidity (1570 cm ) was detected. Under various composition conditions, catalytic activity and amount of TMPs
−
increased with concentration of anion. This is mainly attributed to a higher amount of strong acid ions [Al Cl Br] which can react with
2
6
hydrogen atoms at the 2-position of an imidazolium ion to form Brønsted acid. However, the ionic liquid with strong acidity (X = 0.58)
deactivated rapidly due to a higher sensitivity to moisture, causing decomposition. Under various reaction temperature conditions, optimum
◦
catalytic activity was observed at 80 C. The result is also attributed to the effect of anion composition. The strong acidic anion increased
◦
with temperature. However, at higher reaction temperatures (120 C), the ionic liquid showed a lower activity and TMP selectivity, since the
solubility and Brønsted acid sites were reduced by decomposition of imidazolium ions. The selected ionic liquid sample ([C mim]Br–AlCl )
8
3
was compared with one of the standard commercial catalysts, sulfuric acid. Under optimum experimental conditions, it was observed that
both catalysts showed comparable catalytic behavior. However, ionic liquid showed higher activity, and lower TMP selectivity due to a more
acidic nature and a lower amount of Brønsted acid sites, respectively.

2004 Elsevier Inc. All rights reserved.
Keywords: Alkylation; Ionic liquid; Sulfuric acid; 1-Alkyl-3-methylimidazolium halides–aluminum chloride
1
. Introduction
a rapid catalyst decay [2]. The mechanism of deactivation
has been attributed to carbonaceous materials depositing in-
side solid catalysts [3]. The removal of these materials from
porous catalysts would be difficult due to their high molec-
ular weights and low diffusion coefficients. Consequently,
there is ongoing interest in replacing conventional liquid
acids with newer catalysts which have little diffusion limi-
tation.
Ionic liquids have received widespread attention as an
alternative to conventional hazardous solvents with poten-
tial applications in electrochemistry [4,5], liquid/liquid sep-
arations [6,7], extractions [8], catalysis [9–11], biocataly-
sis [12,13], and polymerization [14,15]. Most of these ionic
salts are good and stable solvents for a wide range of organic
and inorganic materials [16,17]. The important attributes of
Alkylation of isobutane with 2-butene is an industri-
ally important process, since the ensuing products, alky-
lates, have high octane numbers and exhibit clean burning
characteristics [1]. However, the capacity of the alkylation
process is restricted due to constraints imposed by the use
of highly corrosive and environmentally detrimental sulfuric
and hydrofluoric acid as commercial catalysts. Even though
solid acid catalysts have been investigated, they have yet to
achieve commercial use as alkylation catalysts because of
*
Corresponding authors.
E-mail addresses: Varma.Rajender@epa.gov (R.S. Varma),
Panagiotis.Smirniotis@uc.edu (P.G. Smirniotis).
0021-9517/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2003.11.018

512
K. Yoo et al. / Journal of Catalysis 222 (2004) 511–519
earlier [18,21–23]. GC-MS analyses of the products were
carried out by using a Hewlett-Packard 6890 gas chro-
matograph using a HP-5, 5% phenyl methyl siloxane cap-
illary column (30 m × 320 µm × 0.25 µm) and a quadruple
mass filter equipped HP 5973 mass selective detector un-
der a temperature-programmed heating from 313 to 473 K
Fig. 1. Structures of ionic liquids employed in this study. R = C H :
4
9
+
+
+
−
−
−
[
C4MIM] ; C H : [C6MIM] ; C H : [C8MIM] ; X = Cl , Br , I
.
6
13
8 17
◦
at 10 /min.
these ionic liquids include negligible vapor pressure, poten-
tial for recycling, compatibility with various organic com-
pounds and organometallic catalysts, and ease of separation
of products from reactions [18]. From this point of view,
ionic liquids with strong acidity can be potential alterna-
tives to conventional alkylation catalysts without the usual
diffusion limitation drawback. Moreover, they are known to
possess interesting tunable physical and chemical proper-
ties depending on the organic cations and inorganic anions
employed. The organic cation determines the solubility, the
density, and the viscosity of the ionic liquids. Even while
working with a single cation, the properties of the liquid ion
are highly variable through the different possibilities of alkyl
groups. The acidity of the ionic liquid is dependent on the
metal halide employed and the ratio of metal halide to or-
ganic base [18,19]. The possibility of modifying the acidity
of the ionic liquids is very interesting from a catalytic point
of view in the alkylation of isobutane with 2-butene, since
the salient features of a good homogeneous catalyst and re-
action media are the acid strength and the solubility of reac-
tants in the catalyst [1]. However, the alkylation of isobutane
with butene has been barely studied with the exception of
the use of 1-butyl-3-methylimidazolium chloride–aluminum
chloride ([C4mim]Cl–AlCl3) as a catalyst [20].
In a typical representative method, anhydrous aluminum
chloride (107.0 mmol) was slowly added to a round-bottom-
ed flask containing dry 1-butyl-3-methylimidazolium chlo-
ride (100.0 mmol) under a nitrogen atmosphere in a glove
box. The reaction being exothermic and the flask was cooled
during addition and mixing.
2.2. Characterization of ionic liquids
Infrared spectra were obtained using a Bio-Rad high-
resolution FT-IR spectrometer (FTS-40) to investigate the
nature of acidity quantitatively. All spectra were acquired at
−1
2-cm resolution with a total 4 scans acquired per spec-
trum. The preheated samples were placed in a portable,
stainless-steel path-length cell (ThermoNicolet) with NaCl
windows (50 × 25 × 6 mm) suitable for FT-IR spectroscopy.
All spectra were manipulated by Bio-Rad Win-ir software
(ver. 4.14). All the 27Al NMR measurements were carried
out using a Bruker WB-400 AMX spectrometer to inves-
tigate the species of aluminum complex qualitatively. The
2
7
Al NMR chemical shifts were referenced to an external
3
+
Al(H O)6 standard.
2
2.3. Characterization of dissolved hydrocarbons
In the present study, we have investigated the perfor-
mance of various ionic liquids by varying the component
of anions and cations for alkylation of isobutane with 2-
butene. A number of 1-alkyl-3-methylimidazolium halides–
aluminum chloride [Cnmim]X-AlCl3 (where n = 4, 6,
and 8) have been prepared. The catalytic activity and se-
lectivity were carefully determined with prepared samples
under identical conditions. The acidity of ionic liquids was
characterized thoroughly to identify the catalytic perfor-
mance by FT-IR and 27Al NMR spectroscopy. Furthermore,
the selected ionic liquid is compared with one of commercial
catalyst, sulfuric acid, under various experimental conditions
to determine its potential as an alternative catalyst to liquid
acids.
The spent ionic liquid (1 g) or the sulfuric acid (1 g)
were added to water (10 mL) and the product was extracted
with ether. The ether layer was washed with water to remove
traces of catalyst and dried over sodium sulfate. The filtrate
was then analyzed with TLC and GC-MS. For spent sulfuric
acid, the remaining acid-soluble materials in water were ex-
tracted with ether after neutralization. The amount weighed
after evaporation of ether was 0.4 g.
2.4. Catalytic experiment
In a typical reaction, liquid catalyst samples (15 mL) were
placed in a 100-mL double-walled glass reactor (Pyrex).
Dead volume in this reactor was decreased using glass beads
(
Kimble Glass Inc., 3 mm diameter). The isobutane/2-butene
2
. Experimental
mixture (molar ratio of 20) was then introduced into the re-
actor with various flow rates to adjust the liquid hourly space
velocity (LHSV) of system. The flow rate of feed was con-
trolled by a gas-flow controller; thus, all experiments have
been performed precisely. The residence time of feed was
determined by LHSV of feed stream with no stirring in the
reactor. The reaction temperature was controlled by a ther-
mostatic water bath with temperature controller. The reactor
effluent was sent to a 16-port sampling valve (Vici-Valco)
2
.1. Preparation of ionic liquids
The 1-methylimidazole (mim) and alkyl halides were ob-
tained from Aldrich Chemical Company and were used as
such. The 1-alkyl-3-methylimidazolium halides–aluminum
chloride ([Cnmim]X-AlCl3 where X = Cl, Br, I) (see Fig. 1)
were prepared and characterized using protocols described

K. Yoo et al. / Journal of Catalysis 222 (2004) 511–519
513
Table 1
The list of IR spectra of 1-alkyl-3-methylimidazolium halides–aluminum
chloride
−1
Band assignment
Complex of AlCl
Peak frequency (cm
1568
)
3
Ring-stretching symmetrical
Me C–H-bending asymmetrical
Me C–H-bending symmetrical
N–H in-plane bending
Ring C–H in-plane bending
N–H/C–H in-plane bending
C–H in-plane bending
1455, 1366, 1326, 1159
1390
1336
1246
1108
956
832
Ring out of plane asymmetrical
740
Fig. 2. Catalytic conversion of 2-butene versus reaction time over
various ionic liquids: [C mim]Br–AlCl (F), [C mim]Cl–AlCl ("),
8
3
8
3
[
C mim]I–AlCl (2), [C mim]Br–AlCl (E), [C mim]Cl–AlCl (!),
8 3 6 3 6 3
[
C mim]Br–AlCl (!), and [C mim]Cl–AlCl (Q). Reactions were per-
4
3
4
3
−1
◦
formed at LHSV = 24 h and 80 C.
and the product samples were taken periodically and stored
in heated loops. After completion of the reaction, the sam-
ples were sent to a gas chromatograph (Hewlett-Packard,
6
890). The GC column used for separation is a Supelco
Petrocol DH capillary column (50 m × 0.1 mm × 0.1 µm).
Initial qualitative identifications of the products were ac-
complished with the help of a gas chromatograph (Hewlett-
Packard, 5890 Series II) equipped with a mass spectrometer
(
Hewlett-Packard, 5972 Series II).
3
. Result and discussion
3
.1. Comparison of ionic liquids
Fig. 3. FT-IR spectra of ionic liquids with different cations obtained at
◦
80 C.
Catalytic activities determined for the alkylation of isobu-
tane with 2-butene using various ionic liquids are depicted
in Fig. 2. The ionic liquids with the same chemical com-
position, defined as X = [anion]/([anion] + [cation]), were
used in this study. The ionic liquids bearing a larger alkyl
group on their cation ([C8mim]) exhibit higher activity than
a smaller one ([C6 or C4mim]) with the same anionic com-
position, either Cl–AlCl3 or Br–AlCl3 for the entire reaction
time. This is mainly attributable to the high solubility of re-
actants in ionic liquid which can be adjusted by variation of
the alkyl group on the cation. This study corroborates the
earlier report that with increasing length of the alkyl group,
the solubility of hydrocarbons increases significantly [24]. It
has been observed that the solubility of 1-octene increases
with increasing length of the alkyl group on the cation at
peaks representing “ring-stretching symmetrical” increased
with length of the alkyl group on the cation. However, the
−1
peak at 1571 cm representing a complex of AlCl3 [26]
was barely changed. Consequently, the difference of these
samples in terms of catalytic properties is dependent mainly
on the anion part. This trend was confirmed from 27Al NMR
characterization illustrated in Fig. 4. The major signal of
these samples was located around 99.2 ppm representing the
−
presence of [AlCl3Br] species [27].
The influence of the anion on the ionic liquid was inves-
tigated by adjusting halide with the same alkyl group on its
cation. Among the ionic liquids with different halides (Cl,
Br, and I), the sample with bromide as a halide ([C8mim]Br–
AlCl3) has outperformed other ionic liquids under identical
experimental conditions. The higher activity is believed to
be a result of the enhanced acidic nature of the catalyst
determined by an anion group. According to the accepted
mechanism [1], the primary alkylation reaction is facilitated
by a strong acid strength that is crucial for maintaining
a stable activity for this reaction. From the FT-IR results
(see Fig. 5), it was observed that the area of peak indicat-
◦
8
0 C. This entails that the ionic liquid bearing longer alkyl
groups on its cation could catalyze larger amount of reac-
tants under identical conditions, thus, enhancing its catalytic
activity.
In the FT-IR characterization results (Fig. 3), it was ob-
−1
served that most of peaks (1500–700 cm ) correspond to
cationic components (see Table 1) [25] in ionic liquid sam-
ples with various lengths of alkyl groups. The height of

514
K. Yoo et al. / Journal of Catalysis 222 (2004) 511–519
◦
Fig. 5. FT-IR spectra of ionic liquids with different anions obtained at 80 C.
27
Fig. 4. Al NMR spectra of ionic liquids with different cations.
ing acidity (∼ 1568 cm 1) varied with halide ion on the
anion part. The peak for ionic liquid with Br as a halide
was the largest among the samples. This means that the
Al anion formed in [C8mim]Br–AlCl3 is a stronger acidic
anion. Various Al complex anions were identified in ionic
liquid samples with different halides in the NMR charac-
−
−
terization (Fig. 6). As described earlier, [AlCl3Br] anion
was dominant in the [C8mim]Br–AlCl3. For the [C8mim]Cl–
−
AlCl3 sample, AlCl4 (103.5 ppm) ion was detected as the
−
main anion species. The linewidth of the AlCl4 peak indi-
cates very little Al2Cl7 present [28]. For [C8mim]I–AlCl3,
three peaks were detected as the anions equilibrated. Each
−
−
peak represents AlCl4 (103.1 ppm), AlCl3I (86.8 ppm),
−
and AlCl2I2 (59.4 ppm) ions. As we discussed above, the
AlCl3Br] ion is stronger than other major anions among
−
[
the various ionic liquids used in this study. It is also reported
that the bromide-containing aluminates are more acidic than
the corresponding chloride ones [18]. The present study also
shows that the activity of the bromide-based aluminates is
higher than the corresponding tetrachloroaluminates. In ad-
dition, it should be noted that the cationic part of the ionic
liquid could play a role in the acidity of the system. The im-
idazolium ion itself shows a weak Lewis acid [29] and the
hydrogen atom in the 2-position of an imidazolium ion dis-
plays Brønsted acidity [30]. However, the effect of cations
on acidity could be ignored since the same imidazolium ion
was used in this case.
2
7
Fig. 6. Al NMR spectra of ionic liquids with different anions.
sults in the highest octane yield. Moreover, this ratio de-
scribes the relative extent of real alkylation in compari-
son to the oligomerization reaction [31]. The TMP/DMH
ratio for all ionic liquids was almost constant during en-
tire reaction time (not shown) and the trend of selectivity
(TMPs/DMHs) was comparable with that of catalytic ac-
tivity (see Fig. 7). Furthermore, it should be noted that the
ionic liquids with the same anion group yielded compara-
ble values of TMPs/DMHs regardless of the chain length
of the cation on the ionic liquid. From the product distrib-
ution described in Table 2, it is obvious that the length of
cation barely affected the entire selectivity of C8 product.
A small change of cracking product distribution is observed
The ratio of trimethylpentanes (TMPs) to dimethyl-
hexanes (DMHs) is considered as an important criterion
for examining the performance of alkylation catalysts.
A high TMPs/DMHs ratio is desirable as it typically re-

K. Yoo et al. / Journal of Catalysis 222 (2004) 511–519
515
Fig. 7. Catalytic results of ionic liquid samples at time on stream 200 min.
Fig. 8. Catalytic conversion of 2-butene versus reaction time over
C mim]Br–AlCl with various concentrations of anion species. Reactions
were performed at LHSV = 240 h and 80 C.
[
8
3
−1
◦
Table 2
Product distribution obtained using various ionic liquid samples (TOS =
30 min)
[
C mim] [C mim] [C mim] [C mim] [C mim]
4 6 8 8 8
Br–AlCl Br–AlCl Br–AlCl Cl–AlCl I–AlCl
3
3
3
3
3
Olefin conversion (%) 32.1
73.2
15.2
22.7
15.7
45.9
0.5
91.0
18.9
20.1
14.7
45.4
0.9
69.5
11.2
22.5
16.2
50.3
0.4
52.4
8.7
24.3
15.5
50.5
1.0
C5 (wt%)
C6 (wt%)
C7 (wt%)
C8 (wt%)
13.8
23.4
17.4
45.1
0.3
+
C9 (wt%)
C8 distribution
TMPs (wt%)
8.8
0.0
5.4
0.9
2.5
18.8
0.0
2.2
6.1
9.9
0.0
0.6
17.5
9.4
0.0
5.8
1.0
2.6
18.0
0.0
2.1
6.2
9.4
0.0
0.3
18.5
9.3
0.0
5.6
1.1
2.6
18.1
0.0
2.0
5.8
9.5
0.1
0.5
18.2
6.9
0.0
0.3
2.2
4.4
22.7
0.0
3.4
8.1
10.0
0.0
2.3
0.0
1.0
0.5
0.8
13.6
0.0
1.6
4.6
6.9
0.0
0.5
34.6
2,2,3-TMP
2,2,4-TMP
2,3,3-TMP
2,3,4-TMP
DMHs (wt%)
2,2-DMH
2,3-DMH
2,4-DMH
2,5-DMH
3,3-DMH
3,4-DMH
1.2
20.7
Fig. 9. FT-IR spectra of [C mim]Br–AlCl with various concentrations of
C8 olefins (wt%)
8
3
◦
anion species obtained at 80 C.
◦
−1
Reactions were performed at 80 C (LHSV
ratio = 20).
= 24 h
and i/o
molar
total
and then deactivated gradually with time on stream (see
Fig. 8). This is mainly attributed to the moisture sensitivity
of ionic liquids and the decomposition of the organic com-
ponents [18]. From the FT-IR spectra illustrated in Fig. 9, the
using these samples with the same anionic part. For ionic liq-
uids with different anion compositions, significant changes
in product selectivity occurred. The higher amount of crack-
ing products (C5–C7) and alkylation products (TMPs) was
obtained over [C8mim]Br–AlCl3 possessing strong acidic
anions. The acidity influences the selectivity of products and
essentially the anionic part on the ionic liquid is of signifi-
cance.
−1
peak representing an anionic component (∼ 1570 cm ) is
displayed as two different anion species in this composition,
−
−
such as [AlCl3Br] and [Al2Cl6Br] presumably. The ex-
−
istence of stronger acidic species, [Al2Cl6Br] , is attributed
to an excess of anion compound also confirmed from Al
NMR. The appearance of a broad peak (98.7 ppm) in Al
NMR (Fig. 10) is supportive of the fact that the concentration
of [Al2Cl6Br] increased under strong acidic conditions.
2
7
2
7
3
.2. Effect of composition
−
Experiments were conducted with various compositions
For ionic liquids with a lower amount of anionic compo-
nent (X = 0.48), the catalytic performance was lower than
other samples. This is mainly attributed to a lower amount
of available acidic sites, as is obvious from the FT-IR spec-
of ideal ionic liquid ([C8mim]Br–AlCl3), which showed the
most desirable performance profile to critically examine the
effect of anion species. For a more acidic sample (X = 0.58),
it was found that the catalytic activity was high initially
−1
tra; a weak peak (∼ 1570 cm ) was detected.

516
K. Yoo et al. / Journal of Catalysis 222 (2004) 511–519
Fig. 11. Catalytic conversion of 2-butene versus reaction time over
C mim]Br–AlCl at various reaction temperatures. Reactions were per-
[
8
3
−1
formed at LHSV = 240 h
.
increased with the amount of anionic component of the
ionic liquid. This means that higher numbers of Brønsted
acid sites responsible for producing TMPs exist in more
acidic ionic liquid (X = 0.58) as corroborated by results
obtained earlier [20]. It is well known that strong Lewis
2
7
Fig. 10. Al NMR spectra of [C mim]Br–AlCl with various concentra-
−
8
3
acid, [Al Cl Br] , can react with hydrogen atoms at the 2-
2 6
tions of anion species: (A) X = 0.58, (B) X = 0.52, (C) X = 0.48.
position of an imidazolium ion to form Brønsted acid [30,32]
as described in the reaction:
Table 3
−
+
−
+
[
Al2Cl6Br] + H ↔ [AlCl3Br] + [AlHCl3] .
(1)
Product distribution obtained over [C mim]Br–AlCl with various compo-
sitions
8
3
Moreover, it was observed that a small amount of TMP
about 0.7 wt%) was formed using ionic liquid with a lower
Composition
0.52
(
0.48
0.58
300
35.1
anionic composition (X = 0.48). This is attributed to the fact
−
TOS (min)
Olefin conversion (%)
C5 (wt%)
C6 (wt%)
C7 (wt%)
10
18.0
4.5
10.3
14.1
30.3
40.8
300
17.4
4.3
10.2
15.0
31.4
39.1
10
300
39.6
10
that there is a paucity of strong acidic anion, [Al2Cl6Br] ,
40.5
18.9
20.1
14.7
45.2
1.1
70.4
32.5
26.6
16.6
23.7
0.6
which can be transformed to Brønsted acid sites in this sam-
ple. This result supports the aforementioned discussion.
17.9
20.1
13.7
45.4
2.9
24.5
22.2
18.1
22.7
12.5
C8 (wt%)
3.3. Effect of temperature
+
C9 (wt%)
C8 distribution
TMPs (wt%)
In terms of olefin consumption, the maximum activity
was observed at 80 C (Fig. 11), thereby suggesting that
0.7
0.0
0.2
0.2
0.3
7.3
0.0
1.1
2.6
3.4
0.0
0.2
22.3
0.7
0.0
0.2
0.2
0.3
7.2
0.0
1.0
2.6
3.5
0.0
0.1
23.5
9.4
0.0
5.8
1.2
2.5
18.2
0.0
2.0
5.7
9.6
0.3
0.6
17.8
9.2
0.0
5.6
1.1
2.5
18.2
0.0
2.0
5.9
9.7
0.1
0.5
18.0
8.7
0.2
4.7
1.2
2.6
8.3
0.0
1.1
3.0
3.9
0.0
0.3
6.7
5.7
0.3
3.0
1.0
1.4
8.9
0.0
1.3
3.6
3.6
0.0
0.4
8.1
◦
2,2,3-TMP
2,2,4-TMP
2,3,3-TMP
2,3,4-TMP
the acidity of ionic liquid, as determined by the formation
of anions, is influenced by the reaction temperature. In the
FT-IR results illustrated in Fig. 12, the peak heights repre-
DMHs (wt%)
−1
senting acidity (∼ 1570 cm ) under varying temperatures
were similar for the anion composition (see Fig. 9). Ap-
parently, the concentration of strong acidic anion increases
with the temperature as has been observed by others [28,
2,2-DMH
2,3-DMH
2,4-DMH
2,5-DMH
3,3-DMH
3,4-DMH
3
3]. However, under favorable reaction conditions in terms
◦
of acidity, the catalytic activity was low at 120 C. This is
mainly attributed to decomposition of the cation part of the
ionic liquid [18]. In the FT-IR data shown in Fig. 12, it was
seen that the diminution of peaks corresponded with cations
C8 olefins (wt%)
◦
−1
and i/o
Reactions were performed at 80 C (LHSV
ratio = 20).
= 240 h
total
molar
−1
(
1000–1100 and 1300–1400 cm ). As a result of this re-
From the product distribution summarized in Table 3,
a trend was observed. With increasing acidity, the amount
of cracking products (C5–C7) increased with a concomi-
tant decrease in oligomerization products (olefins and heavy
products). It should be noted that the ratio of TMPs/DMHs
duction, the solubility of reactant deceased. Thus, the lower
catalytic performance was observed at this temperature.
From the product distribution point of view, it was ob-
served that the nature of acidity could be altered with reac-
tion temperature as shown in Table 4. Predominant yields

K. Yoo et al. / Journal of Catalysis 222 (2004) 511–519
517
Fig. 13. Catalytic conversion of 2-butene versus reaction time over
C mim]Br–AlCl (1) and H SO (E). Reactions were performed at
[
8
3
2
4
−
1
◦
◦
LHSV = 240 h and 80 C over ionic liquid and 18 C over H SO .
2
4
and oligomerization usually occur over Brønsted and Lewis
+
acids, respectively. Hence, the existence of [AlHCl3] as a
Brønsted acid could maintain the optimum acidity of ionic
liquid determining catalytic performance at this temperature.
Fig. 12. FT-IR spectra of [C mim]Br–AlCl obtained at various tempera-
tures.
8
3
3
.4. Ionic liquid versus sulfuric acid
Table 4
Product distribution obtained under various temperature conditions for both
The change in olefin conversion with time on stream for
samples (TOS = 300 min)
H2SO4 and the ideal ionic liquid ([C8mim]Br–AlCl3) is il-
lustrated in Fig. 13. For a fair comparison, the reaction with
sulfuric acid was performed under optimum conditions, de-
lineated in the literature [1,34–36] such as sulfuric acid of
Sample:
[C mim]Br–AlCl
8
3
◦
Temperature ( C):
50
80
120
Olefin conversion (%)
C5 (wt%)
C6 (wt%)
30.2
10.1
21.1
18.8
46.2
3.8
39.6
17.9
20.1
13.7
45.4
2.9
27.8
15.8
18.7
17.6
44.2
3.7
◦
◦
9
5% purity and at 18 C. At temperatures above (25 C),
oligomerization of olefin becomes significant and the ac-
tivity decreases with temperature. It can be seen that both
catalysts showed comparable catalytic behavior under op-
timum reaction conditions. For the sulfuric acid system,
the olefin conversion gradually decreased and the color of
the catalyst changed with time on stream. It is mainly at-
tributed to the formation of acid-soluble oil (ASO) also
referred to as conjunct polymers or red oil. These mate-
rials have been described to have empirical compositions
of about (C1.0H1.75)n, where n averages from about 10 to
C7 (wt%)
C8 (wt%)
+
C9 (wt%)
C8 distribution
TMPs (wt%)
1.8
0.0
1.4
0.0
0.4
4.7
0.0
0.7
1.7
2.3
0.0
0.0
39.7
9.2
0.0
5.6
1.1
2.5
18.2
0.0
2.0
5.9
9.7
0.1
0.5
18.0
4.1
0.0
3.7
0.0
0.4
3.5
0.0
0.3
2.7
0.5
0.0
0.0
36.6
2,2,3-TMP
2,2,4-TMP
2,3,3-TMP
2,3,4-TMP
DMHs (wt%)
2,2-DMH
2,3-DMH
2,4-DMH
2,5-DMH
3,3-DMH
3,4-DMH
3
0 [34]. These highly unsaturated cyclic hydrocarbons rela-
tively rapidly deactivate the acid catalyst [35].
In the case of bromide-bearing ionic liquid, the olefin
conversion barely changed and remained higher than sul-
furic acid during the entire reaction time. The higher ac-
tivity of ionic liquid is attributable its higher acid strength
which is higher than 100% sulfuric acid [24]. Moreover,
the catalytic stability can be explained by the tunable co-
C8 olefins (wt%)
−1
Reactions were performed at LHSV
ratio = 20.
= 240
h
and i/o
total
molar
◦
of olefins (85%) were observed when the reaction was per-
ordinating ability of the anions at this temperature (80 C).
◦
formed at 50 and 120 C possibly due to oligomerization
A highly electrophilic metal center involved as the active
sites can be stabilized in the solvent without loss of their
activity [37]. Moreover, the ionic liquid employed in this
study can be considered as a biphasic system. In the case
of a biphasic catalysis it is assumed that the active cat-
alytic species is soluble in the solvent but reaction prod-
ucts, especially heavier hydrocarbons, are not [38]. The
hydrocarbon materials dissolved in ionic liquid were iden-
tified to support the above discussion. Most of the ex-
activity that dominates over Lewis acids under these con-
ditions. The fragmentation of the cationic component de-
creases the number of Brønsted acid sites, thus affecting
the selectivity as discussed above. In contrast, a relatively
large amount of primary product (TMPs) was obtained in
◦
alkylation activity at 80 C. Based on a well-known alky-
lation mechanism [1], both alkylation and oligomerization
are strongly dependent on the nature of acidity. Alkylation

518
K. Yoo et al. / Journal of Catalysis 222 (2004) 511–519
Table 5
Product distribution obtained with various LHSV conditions over both sam-
ples (TOS = 300 min)
Sample:
H SO
2
[C mim]Br–AlCl
8 3
4
LHSV:
24
60
240
24
60
240
Olefin conversion (%)
C5 (wt%)
C6 (wt%)
74.6
13.6
16.3
12.5
42.8
14.8
40.1
9.7
12.8
13.1
51.9
12.5
24.5
5.8
11.4
16.7
52.4
13.7
84.1
18.1
19.3
13.5
46.1
3.0
57.1
17.3
20.0
14.8
44.6
3.3
39.6
17.9
20.1
13.7
45.4
2.9
C7 (wt%)
C8 (wt%)
+
C9 (wt%)
C8 distribution
TMPs (wt%)
2,2,3-TMP
28.1
0.7
11.5
7.4
8.5
12.4
0.0
2.6
3.7
5.5
0.0
0.6
2.3
30.8
1.0
12.8
8.7
8.3
17.2
0.0
2.9
6.6
6.8
0.0
0.9
3.9
28.5
1.0
13.5
7.1
6.9
18.8
0.0
2.8
7.1
7.9
0.0
1.0
5.1
9.2
0.0
6.0
0.8
2.4
18.1
0.0
1.8
5.7
9.8
0.0
0.8
18.8
9.4
0.0
5.4
1.3
2.7
18.7
0.0
1.8
6.1
10.2
0.0
9.2
0.0
5.6
1.1
2.5
18.2
0.0
2.0
5.9
9.7
0.1
0.5
18.0
(A)
2,2,4-TMP
2,3,3-TMP
2,3,4-TMP
DMHs (wt%)
2,2-DMH
2,3-DMH
2,4-DMH
2,5-DMH
3,3-DMH
3,4-DMH
0.6
16.5
C8 olefins (wt%)
◦
Reactions were performed at 80 C (i/omolarratio = 20).
The product selectivity over various LHSV is shown
in Table 5. Both catalysts showed different patterns with
LHSV. For sulfuric acid, cracking (C5–C7) and alkylation
(TMPs/DMHs) products formed by strong acid sites de-
creased with increasing of LHSV. Moreover, an increase
of olefin products by oligomerization was observed. Us-
ing ionic liquid, even though olefin conversion varied with
LHSV, the distribution of products was comparable. This
means that the nature of acidity, determined by the anion
part on ionic liquid, hardly changed regardless of reaction
conditions at this temperature. This result supports the stabil-
ity of ionic liquid as well. However, the amount of primary
products (TMPs) obtained is much smaller using ionic liquid
than sulfuric acid. This is because of the lower concentration
of Brønsted acid sites in the case of ionic liquid. Therefore,
ionic liquid with a higher amount of Brønsted acid sites is
more feasible as a commercial alkylation catalyst.
(
B)
Fig. 14. Catalytic conversion of 2-butene versus reaction time with various
LHSV over (A) H SO at 18 C and (B) [C mim]Br–AlCl at 80 C.
◦
◦
2
4
8
3
tracted carbon from spent ionic liquid are smaller than C8
hydrocarbon, e.g., 2,3-dimethylbutane, 3-methylpetane, 2-
methyl-2-pentene, n-hexane, 3-methylhexane, and isohep-
tane. This means that larger hydrocarbons causing deacti-
vation of acidic sites, such as ASO, were barely formed in
ionic liquid and hence, ionic liquid can be used as a stable
catalyst for this reaction.
The influence of liquid hourly space velocity (LHSV)
based on feed mixture on olefin conversions with respect to
the time on stream is illustrated in Fig. 14 for sulfuric acid as
well as ionic liquid. Both catalysts displayed a quite compa-
rable catalytic pattern. In lower LHSV conditions (24 h ),
the catalytic activity was very high without further deactiva-
tion. With increasing LHSV, the catalytic activity decreased
gradually. This reduction of catalytic activity is mainly at-
tributed to the fact that resident time of reactants decreased
with LHSV. At higher LHSV condition, some of reactant
passed over acidic sites without undergoing any reaction. In-
terestingly, using ionic liquid, higher than 240 h of LHSV,
the activity remained at the same level (see Fig. 14B). Thus,
the ionic liquid invariably yields higher than 40% of olefin
−1
4
. Conclusion
Alkylation of isobutane with 2-butene using various
-alkyl-3-methylimidazolium halides–aluminum chloride
1
catalysts has been successfully demonstrated. Among theses
ionic liquids, [C8mim]Br–AlCl3 displayed the best perfor-
mance in terms of activity and selectivity for this reaction.
From various parametric studies, such as anion compositions
and temperature, optimum catalytic activity was observed at
−
1
◦
◦
27
conversion at this reaction temperature (80 C). For sulfuric
80 C and X = 0.52. Based on FT-IR and Al NMR analy-
ses, it was found that the formation of anions was crucial for
determining the catalytic performance. Especially, the ex-
acid, however, olefin conversion still decreased with LHSV.
Therefore, the activity of both catalysts can be determined
by changing reaction conditions.
−
istence of strong acidic anion, [Al2Cl6Br] , maintains the

K. Yoo et al. / Journal of Catalysis 222 (2004) 511–519
519
higher activity and TMP selectivity. In the comparison with
[11] D. Zhao, M. Wu, Y. Kon, E. Min, Catal. Today 74 (2002) 157.
[
[
[
[
12] S.G. Cull, J.D. Holbrey, V. Vargas-Mora, K.R. Seddon, G. Lye,
sulfuric acid, the selected ionic liquid showed higher cat-
alytic activity but lower TMP selectivity. Consequently, an
ionic liquid possessing a larger amount of Brønsted acid sites
would be ideally suited as a commercial alkylation catalyst.
J. Biotechnol. Bioeng. 69 (2000) 227.
13] R.A. Sheldon, R. Maderia Lau, M.J. Sorgedrager, F. van Rantwijk,
K.R. Seddon, Green Chem. 4 (2002) 147.
14] C. Hardacre, J.D. Holbrey, S.P. Katdare, K.R. Seddon, Green Chem. 4
(2002) 143.
15] A.J. Carmichael, D.M. Haddleton, S.A.F. Bon, K.R. Seddon, Chem.
Commun. (2000) 1237.
Acknowledgments
[
[
16] S.J. Nara, J.R. Harjani, M.M. Salunkhe, J. Org. Chem. 66 (2001) 8616.
17] L. Green, I. Hemeon, R.D. Singer, Tetrahedron Lett. 41 (2000) 1343.
The authors thank Dr. Elwood Brook for the NMR analy-
ses. V.V.N. is a postgraduate research participant at the
National Risk Management Research Laboratory adminis-
tered by the Oak Ridge Institute for Science and Educa-
tion through an interagency agreement between the US De-
partment of Energy and the US Environmental Protection
Agency.
[18] T. Welton, Chem. Rev. 99 (1999) 2071.
[19] K.R. Seddon, J. Chem. Technol. Biotechnol. 68 (1997) 351.
[
[
20] Y. Chauvin, A. Hirschauer, H. Olivier, J. Mol. Catal. 92 (1994) 155.
21] J.G. Huddleston, A.E. Visser, W.M. Reichert, H.D. Willauer, G.A.
Broker, R.D. Rogers, Green Chem. 3 (2001) 156.
[22] R.S. Varma, V.V. Namboodiri, Chem. Commun. (2001) 643.
[23] V.V. Namboodiri, R.S. Varma, Org. Lett. 4 (2002) 3161.
[24] P. Wasserscheid, W. Keim, Angew. Chem. Int. Ed. 39 (2000) 3773.
[
[
25] S. Tait, R.A. Osteryoung, Inorg. Chem. 23 (1984) 4352.
26] S. Csihony, A. Bodor, J. Rohonczy, I.T. Hovath, J. Chem. Soc. Perkin
Trans. 1 (2002) 2861.
References
[
[
27] H. Haraguchi, S. Fujiwara, J. Phys. Chem. 73 (1969) 3467.
28] J.L. Gray, G.E. Maciel, J. Am. Chem. Soc. 103 (1981) 7147.
[
[
[
[
[
1] A. Corma, A. Martinez, Catal. Rev. 35 (1993) 483.
2] J. Weitkamp, Y. Traa, Catal. Today 49 (1999) 193.
3] J. Weitkamp, S. Maxiner, Zeolites 7 (1987) 6.
4] C.L. Hussey, Molten Salt Chem. 5 (1983) 185.
5] J.S. Wilkes, J.A. Levisky, R.A. Wilson, C.L. Hussey, Inorg. Chem. 21
[29] J. Howarth, K. Hanlon, D. Fayne, P. McCormac, Tetrahedron Lett. 38
(1997) 3097.
[30] A.J. Arduengo, R.L. Harlow, M. Kline, J. Am. Chem. Soc. 113 (1991)
361.
[31] A. Corma, A. Martinez, C. Martinez, J. Catal. 146 (1994) 185.
[32] P.R. Gifford, J.B. Palmisano, J. Electrochem. Soc. 134 (1987) 610.
[33] K. Ichikawa, M. Izumi, T. Jin, T. Matsumoto, Chem. Lett. (1992) 1651.
[34] L.F. Albright, Ind. Eng. Chem. Res. 41 (2002) 5627.
[35] L.F. Albright, M.A. Spalding, C.G. Kopser, R.E. Eckerty, Ind. Eng.
Chem. Res. 27 (1988) 386.
(1982) 1263.
[
[
[
[
6] R.P. Swatloski, A.E. Visser, W.M. Reichert, G.A. Broker, L.M. Farina,
J.D. Holbrey, R.D. Rogers, Green Chem. 4 (2002) 81.
7] L.A. Blanchard, D. Hancu, E.J. Beckmann, J.F. Brennecke, Nature 399
(1999) 28.
8] J.G. Huddleston, H.D. Willauer, R.P. Swatloski, A.E. Visser, R.D.
Rogers, Chem. Commun. (1998) 1765.
9] R.P. Swatloski, S.K. Spear, J.D. Holbrey, R.D. Rogers, J. Am. Chem.
Soc. 124 (2002) 4974.
[36] K.W. Li, R.E. Eckert, L.F. Albright, Ind. Eng. Chem. Process Des.
Dev. 9 (1970) 434.
[37] H. Olivier-Bourbigou, L. Magna, J. Mol. Catal. A: Chem. 182–183
(2002) 419.
[
10] A.C. Cole, J.L. Jensen, I. Ntai, K. Loan, K.L. Tran, T.K.J. Weaver,
[38] P.B. Hitchcock, T.J. Mohammed, K.R. Seddon, J.A. Zora, C.L.
Hussey, E.H. Ward, Inorg. Chem. Acta 113 (1986) L25.
D.C. Forbes, J.H. Davis Jr., J. Am. Chem. Soc. (2002) 5962.