X. Dai et al. / Journal of Molecular Liquids 219 (2016) 923–929
925
settle down at list 2 h until the upper layer was clear. About 1 mL upper
clear saturated solution was sampled by the pre-heated/cooled syringes
5 mL) with filters (0.2 μm) and extruded into a previously weighed vial
3. Results and discussion
(
3.1. Solubility and ILs structures
determined by an ESI200-4 A analytical balance (± 0.0001 g, Longteng
Electronics Co., Ltd., China). The vial with the sample solution was
closed quickly and weighed, and the mass of the sample solution deter-
mined accordingly. Then the sample solution was diluted by a factor
ranging from 1:10 to 1:100 (v:v) in alcohol which was used as solvent,
depending on the solubility of the ILs in the study. The concentration of
the ILs in ILs-saturated solution was determined by an UV2800S UV–Vis
spectrophotometer (Shunyuhengping Instrument Co., Ltd., China) using
calibration curves established previously. The maximum wavelength of
2 4
The mole fraction solubility values of [HOQu][H PO ],
[HOQu][HSO ], [HOQu][CH SO ], and [HOQu][p-TSA] in six alcohols
4 3 3
(methanol, ethanol, 1-propanol, 2-propanol, 1-butanol and 2-methyl-
1-propanol) in the temperature range from (278.15 to 348.15) K are
listed in Table 3, respectively. It could be found that the solubility of
ILs increased with increasing temperature. It was interesting to observe
that the obvious solubility variation could occur within a narrow tem-
perature range. For instance, the solubility of [HOQu][p-TSA] in 2-
propanol was rapidly doubled from 0.0127 to 0.0253 within the very
small temperature range from (323.15 to 333.15) K, that means the sol-
ubility is very sensitive to temperature change. This behavior refers to
“temperature-sensitive” property of solubility [17,18]. This may be
attributed to the temperature dependence of acid-dissociation and
ion-pair formation. Namely, the acid-dissociation and ion-pair forma-
tion of ionic liquids were increased and decreased with increasing tem-
perature, respectively. According to Table 3, the temperature-sensitive
region of the solubility behavior varies apparently in different alcohols
and [HOQu][p-TSA] has the most significantly “temperature-sensitive”
behavior in all investigated alcohols. The temperature-sensitivity
makes these ILs can be easily dissolved in solution as homogeneous cat-
alyst or extraction auxiliary reagent at high working temperature, then
they can easily crystallize or precipitate from the mixtures at low tem-
perature after reaction or extraction process.
[
HOQu][H
the UV–Vis spectrophotometric calibration curves were determined as
315, 309, 309 and 309) nm, respectively. The DSC Q20 differential scan-
2 4 4 3 3
PO ], [HOQu][HSO ], [HOQu][CH SO ] and [HOQu][p-TSA] at
(
ning calorimetry (TA Instruments, USA) was used to rapidly analyze the
ionic liquids individual samples before and after each experiment, and
no sign of degradation was observed.
The solubility results at each individual temperature were deter-
mined as the average of at least four independent measurements. The
i
mole fraction solubility of ILs in six lower alcohols (x ) was calculated
as follows:
m1=M1
xi ¼
ð1Þ
m1=M1 þ ðm−m1Þ=M2
As the electronegativity of anionic group increases, the hydrogen
bonding between the solvent molecule and the anionic group becomes
stronger [19–21]. The results show a large difference in the mole fraction
solubilities of the four [HOQu] -based ILs as with different anions, and
the tendence of the solubility of ILs in one solvent at the same tempera-
where m (g) and m
1
(g) represent the mass of the sample and the solute,
−
1
−1
respectively; M (g mol ) and M (g mol ) represent the molecular
1
2
mass of solute and the solvent, respectively.
+
In order to check the accuracy of the measurement method, the sol-
ubility of benzoic acid in 2-propanol obtained by the measurement
method in this work was compared with the literature data [16] at at-
mosphere pressure and the results are listed in Table 2. The UV maxi-
mum wavelength used of determination of solubility of benzoic acid
in 2-propanol was 272 nm. The average relative deviation (ARD) be-
tween the experimental solubility and the literature data was 0.0033,
which is calculated by Eq. (2). This indicates that above experimental
procedure was reliable and accurate for solubility measurement.
2 4 4 3 3
ture is roughly [HOQu][H PO ] b [HOQu][HSO ] b [HOQu][CH SO ] b
[HOQu][p-TSA], which could be affected by the strength of hydrogen
bonding between alcohol and anionic group of ILs. For instance, Fig. 2
shows that the solubilities of the four ILs in 1-propanol were along
2 4 4 3 3
with the trend of [HOQu][H PO ] b [HOQu][HSO ] b [HOQu][CH SO ] b
[
HOQu][p-TSA], thus the order of the strength of hydrogen bonding
−
−
−
−
could be [H
2
PO
4
]
b [HSO
4
]
3 3
b [CH SO ]
b [p-TSA] .
Solvent polarity will decrease with the increase of the alkyl straight
chain length or carbon number of alcohols (e.g. methanol N ethanol N 1-
propanol N 2-propanol N 1-butanol N 2-methyl-1-propanol). Normally,
the hydrogen bonding interaction between solvent molecules and an-
ionic groups becomes weaker with the decreasing polarity of solvent,
which is not beneficial for the dissolution of ILs [22]. The results show
that the solubility of an ILs in different solvents decreases roughly
along with the decreasing polarity of solvents especially in the four
straight chain fatty alcohols (including methanol, ethanol, 1-propanol,
ꢀ
ꢀ
ꢀ
ꢂ
X
ꢁ
N
ꢀ
lit
i
exp
i
exp
i
ARD ¼ ð1=NÞ
ꢀχ −χ ꢀ=χ
ð2Þ
i¼1
where N is the number of experimental points, xlit i and xexp i repre-
sent the literature and the experimental mole fraction solubility,
respectively.
1
4
-butanol), except that the solubility of [HOQu][HSO ] and [HOQu][p-
Table 2
Comparison between the experimental mole fraction solubility (x
i
) of benzoic acid in
TSA] follows the order as 2-methyl-1-propanol N 2-propanol. For exam-
ple, as shown in Fig. Fig. 3, the exception of [HOQu][p-TSA] solubility
could be attributed to the branched-chain that exists in the structure
of alcohols.
2
-propanol of this work and those reported in the literature at atmosphere pressure
a,b
(
p = 0.1 MPa).
T/K
x
i
ARDd
Literature valuesc
Experimental values
3.2. Correlation of the experimental data
2
3
3
3
3
3
3
3
88.15
0.1595
0.2146
0.2525
0.2780
0.3251
0.3465
0.3735
0.4107
0.1587
0.2153
0.2516
0.2792
0.3245
0.3473
0.3745
0.4121
03.45
12.80
17.85
28.20
33.20
38.15
43.10
Considering the “temperature-sensitive” property is of great signifi-
+
0.0033
cance for the separation and reusage of [HOQu] -based ILs from reac-
tion mixtures. Therefore, correlation of the solubility data of ILs in
solvents is necessary for the prediction of the solubility of ILs at other
operation temperatures. Generally, the temperature dependence of
the solubility of a solute in solvents can be modeled in three ways, in-
cluding empirical correlation, non-activity coefficient method and activ-
ity coefficient method. In this work, according to the (solid + liquid)
phase equilibrium theory and supposing that the enthalpy of solution
varies with temperature linearly, the modified Apelblat equation [23,
a
Standard uncertainties u are u(T) = ± 0.05 K, u(x
Standard uncertainty u was calculated using standard deviation (SD), x is the mole
i
i
) = ± 0.01 x
i
and u(p) = ± 2 kPa.
b
fraction solubility of benzoic acid at the system temperature T.
c
Reference [16]. N
d
lit
exp
exp
ARD ¼ ð1=NÞ∑i¼1ðjχi −χi j=χi Þ.