Identification of acidic species in chloroaluminate ionic liquid catalysts

Article abstract of DOI:10.1016/j.jcat.2014.09.004

Chloroaluminate ionic liquids (ILs), especially composite ionic liquids (CILs) which are ILs modified with copper(I) chloride, are highly active and selective catalysts for the alkylation of 2-butene with isobutane. The Lewis and Bronsted acidic species of these ILs were investigated by NMR spectroscopy. Pyridine is found to be a suitable indicator. Lewis acidity arises mainly from Al₂Cl₇^- having a chemical shift at 102 ppm in the ²⁷Al NMR spectrum, while Bronsted acidity arises from Al₂Cl₆OH^- (chemical shift at 97 ppm). The peak at 94 ppm in the ²⁷Al NMR spectrum is related to Al₂Cl₅O^-. These new insights have improved our understanding of the structure of the active species in chloroaluminate ionic liquid alkylation catalysts.

Full text of DOI:10.1016/j.jcat.2014.09.004

Journal of Catalysis 320 (2014) 26–32
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Identification of acidic species in chloroaluminate ionic liquid catalysts
a
a
a
Jia Cui a, Jan de With b, Peter A.A. Klusener b, , Xiaohong Su , Xianghai Meng , Rui Zhang ,
⇑
a
a
Zhichang Liu a, , Chunming Xu , Haiyan Liu
⇑
a State Key Laboratory of Heavy Oil Processing, China University of Petroleum, Beijing 102249, China
b Shell Global Solutions International B.V, Amsterdam 1031 HW, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
Chloroaluminate ionic liquids (ILs), especially composite ionic liquids (CILs) which are ILs modified with
copper(I) chloride, are highly active and selective catalysts for the alkylation of 2-butene with isobutane.
The Lewis and Brønsted acidic species of these ILs were investigated by NMR spectroscopy. Pyridine is
found to be a suitable indicator. Lewis acidity arises mainly from Al2Clꢀ7 having a chemical shift at
102 ppm in the 27Al NMR spectrum, while Brønsted acidity arises from Al2Cl6OHꢀ (chemical shift at
97 ppm). The peak at 94 ppm in the 27Al NMR spectrum is related to Al2Cl5Oꢀ. These new insights have
improved our understanding of the structure of the active species in chloroaluminate ionic liquid alkyl-
ation catalysts.
Received 16 June 2014
Revised 11 August 2014
Accepted 6 September 2014
Keywords:
Chloroaluminate ionic liquid
Lewis acidity
Brønsted acidity
NMR
Ó 2014 Elsevier Inc. All rights reserved.
Acidic species
Alkylation
1. Introduction
researchers [17,18]. Hence, there is ongoing interest in the role of
Brønsted acidity in chloroaluminate ILs.
Since the ninth decade of last century, room temperature ionic
liquids (ILs) especially chloroaluminate ILs have been discovered as
alternative alkylation catalysts [1–4]. Their adjustable acidity and
negligible vapor pressure made them attractive alternatives for
the current liquid acid catalysts, hydrofluoric acid, and sulfuric
acid, to produce alkylate for gasoline from isobutane and butene
[5–8].
The acidity study of chloroaluminate ILs has been reviewed by
Welton and Wasserscheid [9,10]. Yoo et al. [11] studied
[omim]Br–AlCl3 by FT-IR and 27Al NMR spectroscopy to explain
the good activity by the strongly Lewis acidic anion Al2Cl6Brꢀ.
Kou [12] used pyridine as probe in IR spectroscopy to determine
the acidity of [bmim]Cl–AlCl3 and concluded that chloroaluminate
ILs have both Lewis and Brønsted acidity. Other works described
that Brønsted acidity in chloroaluminate ILs may come from traces
of water or HCl, which could also improve the catalyst activity
[13–15]. HAlCl4 was proposed as a most likely Brønsted acidity
specie in chloroaluminate ILs [16]; although the potential exis-
tence and stability of HAlCl4 has been questioned by other
Our research group has developed a composite ionic liquid (CIL)
[19], which is a kind of chloroaluminate IL modified with CuCl and
showing excellent activity and selectivity of isobutane alkylation.
The research octane number of our alkylate is above 98, which is
better than that of the commercial catalysts. We later found that
the acidity can be increased or even restored by addition of HCl
or water [20]. The rejuvenation of deactivated CIL catalyst with
HCl can effectively recover the activity (Tables S1 and S2). The more
HCl dissolved in deactivated CIL, the longer lifetime in the alkyl-
ation reaction (Table S1). Similar findings have been reported by
Chevron [21] and Jess et al. [15,16,22]. The hydrolysis of chloroalu-
minate IL was found to be a minor but significant factor for alkyl-
ation activity, because Brønsted acidic proton could provide initial
supply of carbonium ion to initiate alkylation as follows (Fig. 1).
Here, the effects of the acidity of chloroaluminate IL, especially
CIL, on the catalytic performance prompted us to study these phe-
nomena on a molecular level mainly using NMR. This paper
describes our initial study on identification of the acidic species.
2. Experimental
⇑
Corresponding authors at: State Key Laboratory of Heavy Oil Processing, China
University of Petroleum, Changping District, Beijing 102249, China. Fax: +86 10
6972 4721 (Z. Liu). Shell Technology Centre Amsterdam, Grasweg 31, 1031 HW
Amsterdam, The Netherlands (P.A.A. Klusener).
2.1. Materials
Synthesis grade: Triethylamine hydrochloride (99.99%) was
purchased from Merck KGaA. Aluminum chloride (99.99%) and
E-mail addresses: peter.klusener@shell.com (P.A.A. Klusener), lzch@cup.edu.cn
(Z. Liu).
http://dx.doi.org/10.1016/j.jcat.2014.09.004
0021-9517/Ó 2014 Elsevier Inc. All rights reserved.

J. Cui et al. / Journal of Catalysis 320 (2014) 26–32
27
Table 1
Water content in Et3NHCl.
copper(I) chloride (P99.00%) were purchased from Sigma-Aldrich.
All these chemicals were stored in a glove box.
Technical grade: Triethylamine hydrochloride (98%), aluminum
chloride (anhydrous reagent grade), and copper(I) chloride
(90+%) were purchased from Alfa Aesar.
Et3NHCl
Water in Et3NHCl (mg/kg) Water in Et3NHCl (mmol/mol)
Synthesis grade
Technical grade 440.2
5.9
0.045
3.4
Water content in triethylamine hydrochloride was determined
by Karl Fischer method (Table 1).
Pyridine (P99.50%) and dichloromethane-d2 (99.8%) were pur-
chased from Merck KGaA, dried over molecular sieves 4 Å, and
stored in a glove box.
2.3. NMR analysis
All the samples were prepared in a glove box. An ionic liquid
sample (ca. 35 mg) was placed into an NMR tube (5 mm, borosili-
cate glass). Deuterated dichloromethane was added (ca. 1 mL).
The tube was closed with a standard cap and removed from the
glove box immediately prior to the measurement.
1H and 27Al NMR spectra were recorded at 25 °C on an Agilent
NMR spectrometer at 400 and 104 MHz, respectively.
Typical 1H NMR spectral data are as follows (400 MHz, CD2Cl2,
residual CDHCl2 referenced at 5.32 ppm):
BIL with AlCl3 molar fraction of 0.64: d (ppm) 1.43 (t, J = 7.6 Hz,
CH3, 9H), 3.32 (qd, J = 7.4 Hz, J = 5.2 Hz, CH2, 6H), and 5.68 (t,
J = 52.8 Hz, NH, 1H).
BIL + 10 mol% H2O: d (ppm) 1.42 (t, J = 7.4 Hz, CH3, 9H), 3.31 (qd,
J = 7.4 Hz, J = 5.2 Hz, CH2, 6H), 4.85 (s), and 5.69 (t, J = 53.3 Hz, NH,
1H).
2.2. Preparation of ILs
2.2.1. Basis ionic liquids (Et3NHCl–AlCl3, BILs)
BILs with different AlCl3 molar fractions were prepared accord-
ing to the typical procedure [23] under the protection of dry nitro-
gen using Schlenk techniques. A typical example is given for a BIL
with AlCl3 molar fraction of 0.64 (XAlCl3 = 0.64). Synthesis grade
Et3NHCl (13.77 g, 0.10 mol) was placed in a flask. AlCl3 (99.99%
grade, 24.00 g, 0.18 mol) was slowly added to the flask in 30 min
while keeping the reaction temperature below 80 °C. The mixture
was then heated to 80 °C and maintained at that temperature until
all solids ‘‘dissolved.’’ The BILs were generally formed within 3–4 h.
E (BIL + H2O): d (ppm) 1.41 (t, J = 7.4 Hz, CH3, 9H), 3.31 (qd,
J = 7.4 Hz, J = 5.2 Hz, CH2, 6H), 4.85 (s), and 5.71 (t, J = 52.0 Hz,
NH, 1H).
Dry CIL: d (ppm) 1.42 (t, J = 7.4 Hz, CH3, 9H), 3.31 (qd, J = 7.4 Hz,
J = 5.2 Hz, CH2, 6H), and 5.97 (t, J = 52.0 Hz, NH, 1H).
Wet CIL: d (ppm) 1.42 (t, J = 7.4 Hz, CH3, 9H), 3.30 (qd, J = 7.4 Hz,
J = 5.2 Hz, CH2, 6H), 4.81 (s), and 5.96 (t, J = 51.6 Hz, NH, 1H). The
precise integral ratio was CH3: CH2: NH: AlOH = 54.11: 36.72:
5.94: 0.17.
ECIL: d (ppm) 1.42 (t, J = 7.4 Hz, CH3, 9H), 3.30 (qd, J = 7.4 Hz,
J = 5.2 Hz, CH2, 6H), and 6.03 (t, J = 51.6 Hz, NH, 1H).
ECIL + HCl: d (ppm) 1.42 (t, J = 7.4 Hz, CH3, 9H), 3.31 (qd,
J = 7.3 Hz, J = 5.2 Hz, CH2, 6H), 4.81 (s), and 5.99 (t, J = 51.6 Hz,
NH, 1H).
2.2.2. Addition of water to BILs
BIL (ca. 10.00 g, precisely weighed) was placed in a Schlenk ves-
sel, and specific molar equivalent of water was added via syringe.
The molar equivalent was defined as mol H2O per mol of N in
the ionic liquid.
2.2.3. Wet composite ionic liquid (Et3NHCl–AlCl3-CuCl, wet CIL)
CIL was synthesized according to patent application
US20040133056A1 [19]: Anhydrous aluminum chloride was
slowly added to a round-bottomed flask containing technical grade
Et3NHCl under atmospheric conditions at 80 °C. After the forma-
tion of chloroaluminate IL (see the procedure for the BILs), cuprous
chloride was introduced. The reaction mixture was stirred at
100 °C for 2 h, until the complete homogenization of the resulting
ionic liquid.
Pyridine: d (ppm) 7.25 (t, J = 1.6 Hz, meta, 2H), 7.66 (t, J = 8.0 Hz,
para, H), and 8.58 (d, J = 4.0 Hz, ortho, 2H).
ECIL + pyridine (extra signals): Lewis acid complex: –CHa–
(d = 8.91 ppm), –CHb– (d = 8.34 ppm), and –CHc– (d = 7.87 ppm),
ratio of Ha:Hb:Hc = 0.77:0.39:0.79 (Fig. 11).
2.2.4. Dry composite ionic liquid (Et3NHCl–AlCl3–CuCl, dry CIL)
The procedure was repeated as for wet CIL using synthesis
grade feedstock and under the protection of nitrogen atmosphere.
Wet CIL + pyridine (extra signals): Pyridinium: –CH –
a
(d = 8.85 ppm), –CHb– (d = 8.15 ppm), –CH – (d = 8.71 ppm), and –
c
NHd (d = 12.79 ppm, t, J = 68.0 Hz), the ratio of H :Hb:H :Hd was
a
c
2.2.5. Evacuated wet CIL (ECIL)
almost at 2:2:1:1 (Fig. 11).
Some typical 1H NMR spectra of BIL and CIL before and after dif-
ferent treatments are shown in Fig. S1. Other NMR spectral data
(all in CD2Cl2 unless indicated otherwise) are given in the results
and discussion part.
Wet CIL was evacuated with a vacuum pump (P < 0.1 mbar) at
50 °C for 12 h with stirring. Lots of bubbles were generated at
the surface of CIL indicating the loss of gas (HCl).
2.2.6. Addition of HCl to ECIL (ECIL + HCl)
Hydrogen chloride was introduced into ECIL (ca. 10 g) via a tube
filled with anhydrous calcium chloride.
2.4. IR analysis
Infrared spectra were recorded in KBr disks by means of a Nico-
let FTIR spectrophotometer. In the glove box, IL samples (0.5 mL)
were put in between the KBr disks and removed from the glove
box immediately prior to measurement. FT-IR spectra were
recorded on a Nicolet 380/FT-IR Nexus infrared spectrometer in
the typical KBr range of the 4000–400 cmꢀ1 at room temperature,
with a resolution of 4 cmꢀ1 in 32 parallel scans.
2.5. Alkylation test with dry, wet CIL and ECIL and product analysis
In a typical experiment, 50 mL of ionic liquid catalyst was
placed in a 250 mL semi-continuous glass autoclave (Fig. 2). An
Fig. 1. Protonation of butene, formation of tert-butyl cation, and alkylation of 2-
butene.

28
J. Cui et al. / Journal of Catalysis 320 (2014) 26–32
isobutane/2-butene (a mixture of cis- and trans-isomers) mixture
(molar ratio isobutane/butenes of 20) was introduced into the
reactor at 5 mL/min flow rate. Stirring was applied at 500 rpm,
generating a well-mixed zone in the bottom and a non-mixed
hydrocarbon phase in the top of the reactor. The reaction temper-
ature was controlled by a thermostatic oil bath at 23 °C.
After the autoclave was filled, samples were taken from the
hydrocarbon leaving the top of the reactor when 500 mL and
1000 mL of C4 feed was fed into the system. Samples were analyzed
with a gas chromatograph (Hewlett–Packard, 6890) equipped with
a HP-1 methyl silox Agilent column (50 m ⁄ 320
l
m ⁄ 0.25
lm).
The olefin conversion data are given in Fig. 3. Detailed results
are given in the supporting material in Tables S1 and S2.
3. Results and discussion
Fig. 3. Butene conversion after 500 and 1000 mL of isobutane/butene feed at 20 M
ratio over 50 mL of catalyst at 23 °C.
3.1. Alkylation test results
Alkylation tests were done by wet CIL (prepared from technical
grade starting materials and without nitrogen protection) and
evacuated wet CIL (ECIL) and dry CIL (prepared from starting mate-
rials at synthesis grade and under nitrogen protection). The results
showed that wet CIL kept its activity after 1000 mL of C4 feed,
while ECIL and dry CIL started losing activity already after
500 mL of feed (Fig. 3). Apparently, the presence of trace amounts
of water has a dramatic influence on the catalyst activity. Accord-
ing to the literature, [24,25], water has most likely reacted with
chloroaluminate ILs to form HCl, which is assumed to be the active
Brønsted acid. Evacuation leads to removal of HCl and apparently
also to loss of catalyst activity. By these tests, we were even more
triggered to study the structure and effect of acidity of our CIL
alkylation catalyst.
27Al NMR spectra of various BILs with different AlCl3 molar fractions
were recorded and only one peak at about 102 ppm was observed: a
sharp peak in case of AlCl3 molar fraction of 0.5 and a broad signal
of which the peak width increased with increasing AlCl3 molar frac-
tion (Fig. 4A). The changes in chemical shift as well as the peak
broadening show an almost linear trend with the AlCl3 molar frac-
tion (Fig. 4B). The broadening was explained by the rate of the inter-
change reactions between those anions taking place at the NMR
time scale, resulting in the more broadening with more Al2Cl7ꢀ
and Al3Clꢀ10 present. These observations and conclusions are consis-
tent with the findings of Takahashi et al. [28] who studied the NMR
behavior of 1-ethyl-3-methylimidazolium chloroaluminate and
measured the lifetime of the chloroaluminate anions to be fast on
the 27Al NMR timescale.
The chemical shift did only change a little to upfield with
changing the AlCl3 molar fraction as shown in Fig. 4. It is therefore
concluded that the chemical shifts of Al2Clꢀ7 , Al3Cl1ꢀ0, etc. were close
to that of AlClꢀ4 . The signal at 102 ppm (width and intensity) is thus
indicative for Lewis acidic species in ILs with AlCl3 molar fraction
of above 0.5.
3.2. 27Al NMR analysis of BIL
To simplify the structural study, we started with BIL. This sys-
tem is also known to be catalytically active, but with lower selec-
tivity [23] for alkylation of isobutane and 2-butene.
Previous studies [12,26] concluded that, when the molar frac-
tion of AlCl3 in BIL is equal or below 0.5, the major chloroaluminate
anion is AlClꢀ4 . Additional Lewis acidic ionic species, Al2Clꢀ7
(mainly) and Al3Clꢀ10, had been proposed for BILs with AlCl3 molar
fractions of above 0.5 [27] Eqs. (1)–(3). These Lewis anions are
interrelated according to equilibria shown in Eqs. (4) and (5).
3.3. 27Al NMR spectroscopy of neat CIL and diluted CIL
The study was intended to investigate the structure of CIL,
which in reality is applied in pure form as alkylation catalyst. How-
ever, NMR analysis is normally being done in deutero solvent, e.g.
CD2Cl2. Therefore, 27Al NMR analysis of CIL solution was compared
with the 27Al NMR analysis of neat CIL.
Clꢀ þ AlCl3 ꢀ AlClꢀ
ð1Þ
ð2Þ
ð3Þ
ð4Þ
ð5Þ
4
AlClꢀ4 þ AlCl3 ꢀ Al2Clꢀ
7
Al2Clꢀ þ AlCl3 ꢀ Al3Clꢀ
7
10
AlꢁClꢀ4 þ Al2Clꢀ ꢀ AlꢁAlClꢀ þ AlClꢀ4
Neat CIL was analyzed using dimethylsulfoxide-d6 (DMSO) as
external lock agent, using an insert filled with CIL. At room temper-
ature, a very broad signal at ca. 102 ppm was obtained and it
sharpened with increasing temperature (Fig. 5). The broad signals
of neat CIL clearly showed that it will be difficult to get structural
information from those 27Al NMR spectra. When this sample was
heated to 75 °C, the peaks sharpened and became close to that of
the CIL sample diluted in CD2Cl2 which was measured at room
temperature (Fig. 5). Thus, the exchange between the anion species
in solution is faster than that in neat CIL, while the structures of
CIL, the species as observed in CD2Cl2 solution, are similar to those
in neat CIL.
7
7
2Al2Clꢀ ꢀ AlClꢀ þ Al3Clꢀ
7
4
10
The 27Al NMR spectrum of CIL in CD2Cl2 showed three peaks at
102, 97, and 94 ppm, respectively (Fig. 5 bottom). The broad signal
at 102 ppm is similar to that in the 27Al NMR spectrum of BIL and
ꢀ
can be assigned to the mixture of AlClꢀ4 , Al2Cl7ꢀ, e.g. AlxCl3x+1 anions.
Thus, the intensity and broadening of the peak at 102 ppm indi-
cated a high content of Lewis acidic species in CIL.
Fig. 2. Apparatus for isobutane alkylation.

J. Cui et al. / Journal of Catalysis 320 (2014) 26–32
29
Fig. 4. Change of 27Al NMR spectra of BILs with various AlCl3 molar fractions at room temperature.
Fig. 5. 27Al NMR spectra of neat CIL (DMSO-d6 as external lock agent) at various
temperatures and of CIL in CD2Cl2 solution at 25 °C.
3.4. Brønsted acidic species
Brønsted acidic species in chloroaluminate ILs were investi-
gated by addition of water and HCl to BIL and CIL. Water
(10 mol%) added to BIL (XAlCl3 = 0.64) yielded a new peak at
97 ppm (Fig. 6B), while the peak at 102 ppm sharpened. Obviously,
the 97 ppm peak is related to the product formed by the addition of
water. The same signal was found in wet CIL that was prepared
with wet feedstock (Fig. 6E). Takahashi et al. [28] also concluded
earlier that this 97 ppm signal was attributed to a contaminant
derived from reaction of chloroaluminate with water. The reaction
to be expected to occur is given in Eq. (6).
Fig. 6. 27Al NMR spectra of BIL and CIL before and after different treatments.
ECIL resulted in an increase of the 97 ppm peak and a slight
decrease of the 94 ppm peak (Table S3). As expected, the 27Al
NMR spectrum of ECIL with the addition of HCl was almost
returned to the one of CIL before evacuation. This phenomenon
verified the hypothesis that the 97 ppm peak is related to chloro-
hydroxy aluminate species and the 94 ppm peak is related to chlo-
rooxy aluminates. In addition, the chloro(hydr)oxy aluminates
species Al2Cl6OHꢀ and Al2Cl5Oꢀ have been identified by electro-
spray ionization mass spectrometry [3,29].
Al2Clꢀ þ H2O ! Al2Cl6OHꢀ þ HCl "
ð6Þ
ð7Þ
7
Al2Cl6OHꢀ ꢀ Al2Cl5Oꢀ þ HCl "
The 1H NMR spectra of BIL + 10 mol% H2O, wet CIL and
ECIL + HCl showed a small peak at ca. 4.8 ppm (Section 2.3) that
was not observed in the 1H NMR spectra of BIL, dry CIL, and the
evacuated ILs. This extra peak is most likely related to the Brønsted
acidic proton. This observation is consistent with the study of Ber-
enblyum et al. [14]. The integral of the signal at 4.8 ppm in the 1H
NMR spectrum of wet CIL was about 3 mol%, while the integral of
the 97 ppm peak in 27Al NMR spectrum was ca. 7 mol%. Therefore,
the peaks at 97 ppm in the 27Al NMR spectra and at ca. 4.8 ppm in
the 1H NMR spectra both can be assigned to Al2Cl6OHꢀ, which
plays a Brønsted acidic role in CIL via the equilibrium in Eq. (7).
Although Al2Cl6OHꢀ might have two different Al centers, we
assume that in the 27Al NMR spectrum only one signal (at
97 ppm) was observed due to a fast internal exchange (Eq. (8)).
Hence, it explains that the addition of water into chloroaluminate
ILs can generate Brønsted acid HCl.
Berenblyum et al. [14] found that the HCl solubility in chloroa-
luminate ILs depends on the system pressure and the AlCl3 molar
fraction. Therefore, CIL was evacuated in order to remove any dis-
solved HCl in wet CIL. Fig. 6F shows the 27Al NMR spectrum of wet
CIL after evacuation (ECIL). The peak at 97 ppm almost disap-
peared, while the intensity of peak at 94 ppm increased. Such
behavior was also observed at evacuating BIL after 10 mol% of
water had been added (cf. spectra B and C in Fig. 6). It seemed that
the aluminate anion with signal at 97 ppm was converted into a
species resonating at 94 ppm. Most likely, the evacuation not only
removed dissolved HCl, which would not lead to such a change in
the 27Al NMR spectrum, but also induced a chemical conversion as
shown in Eq. (7) [24,26].
ð8Þ
In order to support this assumption, anhydrous HCl was intro-
duced into ECIL under nitrogen protection. The addition of HCl to

30
J. Cui et al. / Journal of Catalysis 320 (2014) 26–32
The formation of the 97 ppm peak in the 27Al NMR spectrum
In comparison with BIL, the signals of wet CIL showed much
sharper signals in the 27Al NMR spectrum indicating that cuprous
chloride might have a positive effect on stabilizing chlorooxy alu-
minate species. In our previous work [3], a new composite anion
AlCuClꢀ5 was proposed based on ESI-MS and 27Al NMR spectros-
copy, and the chemical shift at 97 ppm was assigned to this anion.
However, at present, there is no evidence of AlCuClꢀ5 resonating at
97 ppm, and this peak is now proved further to be from Al2Cl6OHꢀ.
Bui [16] speculated the formation of super acids by interaction of
CuCl with the Et3NH+ cation, forming Et3NCu+ and HCl. The 1H
NMR analysis of wet CIL, showing the CH3, CH2, and NH signals
in the expected 9:6:1 ratio (see experimental part), does not sup-
port the proposal of CuCl reacting with Et3NH+. The role of copper
still remains unclear.
was followed for BILs with different AlCl3 molar fractions as a func-
tion of amount of water added. Several phenomena were observed.
The peak at 102 ppm sharpened with the addition of water, which
was explained by the consumption of Al2Clꢀ7 to form Al2Cl6OHꢀ
(Fig. 7). The sharpening stopped and solids were formed when all
Al2Clꢀ7 was consumed (Fig. 8). For XAlCl3 = 0.64, this happened at a
H2O/IL molar ratio of about 0.8; for XAlCl3 = 0.58, this happened at
a H2O/IL molar ratio of about 0.4. This can be explained by the fol-
lowing stoichiometries:
XAlCl3 ¼ 0:64 ! Al=N ¼ 1:8 ! ðAl2ClꢀÞ ðAlCl4ꢀÞ NHEt3þ
7
XAlCl3 ¼ 0:58 ! Al=N ¼ 1:4 ! ðAl2ClꢀÞ0:8ðAlClꢀ4 Þ0:2NHEtþ3
7
0:4
0:6
During the water addition, another small signal at 89 ppm
appeared. It is thus far unknown to which species this signal
belongs. It might be a dihydroxy chloroaluminate, which is only
slightly soluble.
The experimental integral ratio of the 102 and 97 ppm peaks
was compared with the calculated value assuming the stoichiom-
etry of Eq. (6). The experimental data showed a good fit with the
calculated data (Fig. 8), being another confirmation of the forma-
tion of Al2Cl6OHꢀ from Al2Cl7ꢀ.
3.5. The determination of acidic species by pyridine
Pyridine is a well-known indicator to determine the acidity of
ionic liquid by IR spectroscopy [12,16]. Pyridine added to CIL
resulted in absorption bands at 1538 and 1454 cmꢀ1 (Fig. 9), which
can be assigned to pyridine molecules interacting with Brønsted
and Lewis acid sites, respectively.
NMR spectroscopy was used to evaluate the structures of Lewis
and Brønsted acidic species in the wet CIL interacting with pyri-
dine. Pyridine was added in 0.1 and 0.3 M equivalent to wet CIL.
The 27Al NMR spectra are shown in Fig. 10B, and C, respectively.
Compared with the 27Al NMR spectrum of wet CIL in Fig. 6E, a
new peak at 105 ppm appeared, which was also found in the 27Al
NMR spectrum of ECIL with pyridine addition. When the Py/CIL
ratio was 0.3, the 97 ppm peak almost disappeared while the
102 ppm signal sharpened. These changes indicated that pyridine
had reacted with the aluminate anions resonating at 97 and
102 ppm forming a new aluminate complex with a signal at
105 ppm. The latter is considered to be the Lewis acidic aluminate
species coordinating to pyridine. When the Py/CIL ratio was 0.35,
the 97 ppm signal, belonging to the Brønsted acid aluminate, was
completely disappeared.
The pyridine regions of the 1H NMR spectra of ECIL and wet CIL
after adding pyridine are shown in Fig. 11. The signals in the
Et3NHCl region did not change with the addition of pyridine (Sec-
tion 2.3). With the addition of pyridine to ECIL, new signals
appeared in the 7.5–9.0 ppm region (a, b, and c in Fig. 11A), which
were different from pure pyridine (Section 2.3). With the addition
of pyridine to wet CIL, a second set of new pyridine signals
Fig. 7. Change of 1/2 peak width of 102 ppm signal in 27Al NMR at the addition of
water to BILs with XAlCl3 = 0.64 (Al/N = 1.8) and XAlCl3 = 0.58 (Al/N = 1.4).
Fig. 8. Integral ratio of 102/97 peaks in 27Al NMR with the addition of water to BILs
for XAlCl3 = 0.64 (Al/N = 1.8) and XAlCl3 = 0.58 (Al/N = 1.4).
Fig. 9. FT-IR spectra of pure Py, wet CIL, and ECIL interacting with 10 mol% of Py.

J. Cui et al. / Journal of Catalysis 320 (2014) 26–32
31
Fig. 12. Formation pathways and identification of acidic species of chloroaluminate
ILs with 27Al NMR.
97 signal in the 27Al NMR spectrum was disappeared, and then, in
the 1H NMR spectrum, the amount of the PyH+ triplet signal at
12.8 ppm relative to the amount of the N–H of Et3NH+ was deter-
mined. Thus, the amount of Al2Cl6OHꢀ in wet CIL was determined
to be 5.3 mol%, which was much higher than the water content in
the triethylamine hydrochloride. Besides that, Al2Cl5Oꢀ was also
present, which was also originating from the reaction of the alumi-
nate with water. Apparently, lots of extra water was introduced
during the wet CIL synthesis by not working under inert conditions.
Fig. 10. 27Al NMR spectra of ECIL and wet CIL after addition of Py.
4. Conclusions
Lewis and Brønsted acidic species were found in chloroalumi-
nate ILs by 1H NMR, 27Al NMR, and IR analysis. Lewis acidic species
were mainly provided by Al2Clꢀ7 represented by a broad signal at
102 ppm in the 27Al NMR spectrum. Lewis acidic species were fur-
thermore demonstrated by using pyridine as indicator giving a
pyridine-Lewis acid complex that was identified by 1H NMR, 27Al
NMR, and IR spectroscopy. The Brønsted acidic species were
related to chlorohydroxy aluminate Al2Cl6OHꢀ resonating at the
97 ppm in the 27Al NMR spectrum. This anion could lose HCl, e.g.
by evacuation, to form non-reactive chlorooxy aluminate Al2Cl5Oꢀ
(at 94 ppm). The reappearance Brønsted acidity, e.g. of the 97 peak,
was demonstrated by rejuvenation of evacuated CIL with HCl.
Brønsted acidity was furthermore demonstrated by using pyri-
dine as indicator with 1H NMR, 27Al NMR, and IR spectroscopy,
showing the disappearance of the Brønsted acid Al2Cl6OHꢀ and
the formation of a pyridium hydrogen salt.
Fig. 11. 1H NMR spectra of ECIL and wet CIL after addition of Py (only pyridine and
pyridinium signals are shown).
27Al and 1H NMR spectroscopy has been proven to be a strong
and efficient tool to characterize the acidity of chloroaluminate
ILs. This study has provided a new insight into Brønsted acidity
formed in chloroaluminate ILs. However, the effects of CuCl, HCl,
and H2O are still not fully understood, especially that of CuCl on
selectivity, and are being investigated further.
appeared in the 1H NMR spectrum (
a, b, and c in Fig. 11B). The co-
appearance of the latter signals with a new triplet d at 12.8 ppm
with a 14N–H coupling constant of 68.0 Hz is a clear identification
of a hydrogen pyridium salt PyH+ derived from the reaction with
Brønsted acid. The other set of pyridine signals (a, b, and c) is iden-
tified as pyridine-AlCl3 formed from the reaction with Lewis acid
(Eq. (9)). Thus, pyridine is a good indicator of the different acidic
species in chloroaluminate ILs.
Acknowledgments
The experimental work was carried out at the Shell Technology
Centre Amsterdam. The support by the Shell colleagues A.D. Hor-
ton, B. van Oort, and H.H. Meurs was highly appreciated. Financial
support was provided by Shell Global Solutions International B.V.,
The Netherlands, the National Science Foundation of China (Grant
Nos. 21036008, 21276275, 21206193, and 20976194), the Program
for New Century Excellent Talents in the University of China (No.
NCET-12-0970), and the Science Foundation of China University
of Petroleum, Beijing (Nos. KYJJ2012-03-23 and KYJJ2012-03-25).
Al2Clꢀ þ Py ! AlCl4ꢀ þ Py ꢀ AlCl3
ð9Þ
7
The 105 ppm signal in the 27Al NMR spectrum was related to the
pyridine-Lewis acid complex observed in 1H NMR spectrum, while
the disappearance of the 97 ppm signal in the 27Al NMR spectrum
was related to the formation of pyridinium cation from pyridine
and Brønsted acid. Thus, Al2Clꢀ7 is the main Lewis acid in our cata-
lytic system, and Al2Cl6OHꢀ is the main Brønsted acid that is in
equilibrium with HCl. The overall pathway of identification of Lewis
and Brønsted acidic species in chloroaluminate ILs is shown in
Fig. 12.
Appendix A. Supplementary material
Finally, pyridine was used to determine the amount of Brønsted
acid present in wet CIL. Wet CIL was titrated with pyridine until the
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2014.09.004.

32
J. Cui et al. / Journal of Catalysis 320 (2014) 26–32
[14] A.S. Berenblyum, E.A. Katsman, Y.Z. Karasev, Appl. Catal. A: Gen. 315 (2006)
References
128.
[15] T.L.T. Bui, W. Korth, S. Aschauer, A. Jess, Green Chem. 11 (2009) 1961.
[16] T.L.T. Bui, W. Korth, A. Jess, Catal. Commun. 25 (2012) 118.
[17] H.C. Brown, H. Pearsall, J. Am. Chem. Soc. 73 (1951) 4681.
[18] C. Sikorska, S. Freza, P. Skurski, J. Phys. Chem. A 114 (2010) 2235.
[19] Z.C. Liu, C.M. Xu, C.P. Huang, US 2004/0133056 A1, 2004.
[20] Z.C. Liu, C.M. Xu, R. Zhang, X.H. Meng, US 2012/0283500 A1, 2012.
[21] E. Saleh, T.V. Harris, US 2007/0142676 A1, US 2007/0142217 A1, 2007.
[22] F. Pöhlmann, L. Schilder, W. Korth, A. Jess, ChemPlusChem 78 (2013) 570.
[23] C.P. Huang, Z.C. Liu, C.M. Xu, B.H. Chen, Y. Liu, Appl. Catal. A: Gen. 277 (2004)
41.
[1] G.A. Olah, T. Mathew, A. Goeppert, B. Török, I. Bucsi, X.Y. Li, Q. Wang, E.R.
Marinez, P. Batamack, R. Aniszfeld, G.K.S. Prakash, J. Am. Chem. Soc. 127 (2005)
5964.
[2] Y. Chauvin, A. Hirschauer, H. Olivier, J. Mol. Catal. 92 (1994) 155. US 5750455
(1998).
[3] Z.C. Liu, R. Zhang, C.M. Xu, R.A. Xia, Oil Gas J. 104 (2006) 52.
[4] P. Kumar, W. Vermeiren, J.P. Dath, W.F. Hoelderich, Appl. Catal. A: Gen. 304
(2006) 131.
[5] L.F. Albright, Ind. Eng. Chem. Res. 48 (2009) 1409.
[6] R.D. Rogers, K.R. Seddon, Science 302 (2003) 792.
[7] A. Corma, A. Martínez, Catal. Rev. Sci. Eng. 35 (1993) 483.
[8] S. Tang, A.M. Scurto, B. Subramaniam, J. Catal. 268 (2009) 243.
[9] T. Welton, Chem. Rev. 99 (1999) 2071.
[24] A.K. Abdul-Sada, A.M. Greenway, K.R. Seddon, T. Welton, Org. Mass Spectrom.
28 (1993) 759.
[25] P.C. Trulove, R.A. Osteryoung, Inorg. Chem. 31 (1992) 3980.
[26] T.A. Zawodzinski, R.A. Osteryoung, Inorg. Chem. 29 (1990) 2842.
[27] L.G. Boxall, H.L. Jones, R.A. Osteryoung, J. Electrochem. Soc. 120 (1973) 223.
[28] S. Takahashi, M.L. Saboungi, R.J. Klingler, M.J. Chen, J.W. Rathke, J. Chem. Soc.,
Faraday Trans. 89 (1993) 3591.
[10] P. Wasserscheid, W. Keim, Angew. Chem. Int. Ed. 39 (2000) 3772.
[11] K. Yoo, V.V. Namboodiri, R.S. Varma, P.G. Smirniotis, J. Catal. 222 (2004) 511.
[12] Y.L. Yang, Y. Kou, Chem. Commun. (2004) 226.
[13] G.P. Smith, A.S. Dworkin, R.M. Pagni, S.P. Zingg, J. Am. Chem. Soc. 111 (1989)
525.
[29] Y. Liu, R. Hu, C.M. Xu, H.Q. Su, Appl. Catal. A: Gen. 346 (2008) 189.