Effect of alkyl and aryl substitutions on 1,2,4-triazolium-based ionic liquids for carbon dioxide separation and capture

Article abstract of DOI:10.1039/c2ra22646d

A series of 1,2,4-triazolium-based ionic liquids have been synthesized and evaluated for their use in supported ionic liquid membrane based CO₂ separations. The properties of these triazolium-based compounds have proven sensitive to isomeric substitutions, such as isopropyl and propyl groups, as well as ortho and para substitutions in the aryl derivative compounds. While physical properties such as viscosity did not vary significantly between structural isomers, the CO₂ permeability, selectivity, and solubility exhibited significant changes allowing for development of task-specific triazolium-based ionic liquids for separation applications. COSMOtherm studies were also completed to gain a better understanding of the ionic liquids which demonstrated a strong correlation between experimental and computational values for the alkyl bearing ionic liquids. Hence, 1,2,4-triazolium-based liquids comprise a class of compounds offering unique opportunities to examine how structural changes affect the physicochemical properties which are necessary for the continuous development of ionic liquids with enhanced adsorption capacity and selectivity in separations.

Full text of DOI:10.1039/c2ra22646d

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Effect of alkyl and aryl substitutions on 1,2,4-triazolium-
based ionic liquids for carbon dioxide separation and
capture3
Cite this: RSC Advances, 2013, 3,
3981
Patrick C. Hillesheim,a Joseph A. Singh,a Shannon M. Mahurin,*a Pasquale F. Fulvio,a
Yatsandra Oyola,a Xiang Zhu,b De-en Jianga and Sheng Dai*ab
A series of 1,2,4-triazolium-based ionic liquids have been synthesized and evaluated for their use in
supported ionic liquid membrane based CO2 separations. The properties of these triazolium-based
compounds have proven sensitive to isomeric substitutions, such as isopropyl and propyl groups, as well as
ortho and para substitutions in the aryl derivative compounds. While physical properties such as viscosity
did not vary significantly between structural isomers, the CO2 permeability, selectivity, and solubility
exhibited significant changes allowing for development of task-specific triazolium-based ionic liquids for
separation applications. COSMOtherm studies were also completed to gain a better understanding of the
ionic liquids which demonstrated a strong correlation between experimental and computational values for
the alkyl bearing ionic liquids. Hence, 1,2,4-triazolium-based liquids comprise a class of compounds
offering unique opportunities to examine how structural changes affect the physicochemical properties
which are necessary for the continuous development of ionic liquids with enhanced adsorption capacity
and selectivity in separations.
Received 25th October 2012,
Accepted 21st December 2012
DOI: 10.1039/c2ra22646d
www.rsc.org/advances
increased interest in the application of ILs for post-combus-
tion CO2 capture.14–23
Introduction
Ionic liquids (ILs) have drawn considerable attention due to
their favourable properties such as negligible vapour pressure,
wide electrochemical window, and high thermal stability.1
Additionally, ILs have been introduced into diverse scientific
and technological areas due to the many possible combina-
tions when modifying the structures of both the anion and
cation, and consequently the IL properties. To date, ILs have
found applications in catalysis,2,3 materials synthesis,4 and
green chemistry.5 Among all classes of ILs, those having an
imidazolium cation comprise the largest class of compounds
examined so far. While these imidazolium-based ILs offer a
wide range of modularity, other heterocyclic compounds
capable of forming ILs, such as thiazolium,6 1,2,3-triazolium,7
or 1,2,4-triazolium,8 have been less investigated. Among those,
triazolium-based ILs have been explored for biological
applications,9 catalysis,10 and in the development of energetic
materials.11–13 In addition to these uses, there has been an
There are numerous physicochemical properties governing
both the solubility and permeability of gases within ILs,
including free volume, viscosity, and cation–anion interac-
tions.24 These properties can readily be adapted as needed
given the ease of modulation for both the cation and anion
within ILs. For instance, in imidazolium-based systems, it has
been demonstrated that by adjusting the length of alkyl
substituting groups, the CO2 adsorption capacity can be
modified.5 While there exist theoretical studies examining
structure–property relationships in various ILs,25 studies
examining the direct effect of isomeric substitutions on
physicochemical properties of ILs are limited. The latter
limitations thus restrain the ability to design novel com-
pounds with enhanced adsorption capabilities, limiting our
understanding of anion–cation interactions, and their segre-
gation into 3-dimensional networks that affect CO2 perme-
ability, among other physicochemical properties.
The interaction of gases with ILs are complex, with
numerous physical properties affecting the overall solubility
and permeability of the gases within the liquids.22 The
structure of both the anionic and cationic species has been
shown to affect the separations properties of ILs, as modula-
tion of the intermolecular interactions, i.e. cation–cation and
cation–anion interactions, significantly affect both the solubi-
lity and permeability of gases.4,5,22,24,26 Steric or electronic
aChemical Sciences Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee
37831, USA. E-mail: mahurinsm@ornl.gov
bDepartment of Chemistry, University of Tennessee, Knoxville, TN 37996, USA.
E-mail: dais@ornl.gov
Electronic supplementary information (ESI) available: Full synthetic details,
3
TGA plots under ambient conditions, and complete DSC graphs are included in
the supplementary information. See DOI: 10.1039/c2ra22646d
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 3981–3989 | 3981

View Article Online
Paper
RSC Advances
the anion–cation interactions and the resulting effects on gas–
IL interfaces.
Experimental section
General
1-methyl-1,2,4-triazole, 4-methoxytetrafluorobenzyl bromide,
4-(trifluoromethoxy)benzyl bromide, 2-(trifluoromethoxy)ben-
zyl bromide were purchased from Matrix Scientific. Lithium
bis(trifluoromethanesulfonimide) was purchased from 3M. All
other chemicals were purchased from Sigma–Aldrich in the
highest available purity and used as received. 1H, 13C NMR, 31
P
spectra were recorded on a Bruker Avance 400, at 400 MHz,
100 MHz, and 162 MHz respectively. A JEOL (Peabody, MA)
JMS-T100LC (AccuTOFTM) orthogonal time-of-flight (TOF)
mass spectrometer was used to characterize the compound.
3-butyl-1-methyl-1H-imidazol-3-ium bis((trifluoromethyl)sulfo-
nyl)amide) [Bmim][TF2N],30 3-benzyl-1-methyl-1H-imidazol-3-
ium bis((trifluoromethyl)sulfonyl)amide [BzMim][TF2N],31 and
3-hexyl-1-methyl-1H-imidazol-3-ium bis((trifluoromethyl)sulfo-
nyl)amide) [Hmim][TF2N]32 were synthesized according to
established literature procedures.
Fig. 1 Ionic liquids synthesized and examined in this work.
changes in the ion-pair interactions should then have an
influence on the gas separations properties of ILs. As such,
1,2,4-triazolium-based ILs provide a unique opportunity to
examine the separation properties imparted by the addition of
a nitrogenous lone-pair located in the cationic core of the IL as
directly compared with the more commonly examined
imidazolium ILs. Given the limited publications currently
available for 1,2,4-triazole based ILs, the data acquired from
the present studies are necessary to help expand the general
knowledge base of ILs,27 increase the data available for
accurate computational studies,28 and to add candidates
useful in the search for superior materials for separation
applications.29
Herein we report the synthesis and characterization of a
series of twelve 1,2,4-triazolium-based ILs with alkyl (com-
pounds 1a–1g), and N-benzyl groups (compounds 2a–2e),
which are summarized in Fig. 1. The effects of systematic
ancillary nitrogen substitutions on the overall properties of the
ILs and their use in supported ionic liquid membrane (SILM)
based gas separation was investigated. Of the twelve com-
pounds, seven bear N-alkyl substitutions to examine the effects
of increasing chain length and of isomers on physicochemical
properties. The remaining five ILs bear N-benzyl substituents
to examine the effect of aryl substitutions in addition to
examining the effects of ortho, meta, and para aryl substitu-
tions. Our results demonstrate that optimization of the
separations properties of triazolium ILs can be achieved
through the simple tailoring of isomeric and locational
substitutions of ancillary groups. Additional studies using
COSMOtherm were completed for the triazolium-based ILs
which show a strong correlation between the experimental and
computational values for CO2 solubility of the alkyl-bearing
ILs. Hence, these new class of triazolium-based cations are
important model systems for the understanding of shifts in
Physical properties
Thermal stabilities were measured using a Thermal Advantage
2950 Analyzer with platinum crucibles under a nitrogen
atmosphere (10 uC min21 heating to 750 uC). The instrument
was calibrated prior to measurement using
a calcium
carbonate standard. Melting points were obtained by dynamic
scanning calorimetry (DSC) on a Thermal Advantage Q50
instrument using hermetically sealed aluminum pans. The ILs
were subjected to three consecutive cycles from 290 uC to 200
uC (5 uC min21 rate).
Gas solubility measurements were obtained using a gravi-
metric microbalance (Hiden Isochema, IGA). A known mass of
the ionic liquid (generally of the order of 70 mg) was loaded
into a quartz sample boat and sealed in the stainless steel
chamber. To remove any residual water, the ionic liquid was
then dried under vacuum at a temperature of approximately 60
uC until the mass no longer decreased (minimum of four
hours) and the dry mass was recorded. The ionic liquid mass
was then measured for various CO2 pressures up to 10 atm at a
constant temperature of 25 uC, which was maintained using a
constant-temperature recirculating water bath. Volume-based
solubility values were then calculated using the slope of the
best fit line.
Gas permeabilities through the ionic liquids were measured
by loading the ionic liquid onto a porous polyethersulfone
support with a pore diameter of 100 nm and measuring gas
flux through the ionic liquid membrane. The permeability
measurement system which has been previously described in
detail33 consisted of two chambers separated by the supported
ionic liquid membrane. After initially evacuating both cham-
bers, one chamber was filled to a pressure of 35 kPa with
either CO2 or N2 and the gas flux through the membrane was
obtained by recording the pressure rise in the second chamber
as a function of time. The permeability was then calculated
3982 | RSC Adv., 2013, 3, 3981–3989
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
using eqn (1):
Compounds 1a–g displayed similar decomposition profiles
characterized by a single decomposition step under both air
t
V
dP0
w RTADP0 dt
(see ESI Fig. S1 and S2) and nitrogen, whereas compound 1d
3
P~
(1)
shows two overlapping decomposition steps. The peak max-
ima of the DTG curve for compound 1d shows a main
decomposition at 343 uC, with the second step happening at
354 uC. All compounds, except 1e and 1g, exhibit similar initial
decomposition temperatures under both ambient and inert
atmospheres.
where t is the tortuosity, w is the membrane porosity, V is the
permeate volume, R is the ideal gas constant, T is the absolute
temperature, A is the membrane area, DP0 is the pressure
difference, and dP99/dt is the rate of pressure increase on the
permeate side of the membrane. The tortuosity and porosity
were obtained using a standard ionic liquid, [Emim][Tf2N].
Ideal CO2/N2 selectivity values were obtained by separately
measuring the CO2 and N2 permeability and calculating the
ratio of these values to obtain the selectivity. Mixed gas
measurements were not acquired.
Temperature dependent viscosity measurements were
acquired using a cone/plate viscometer (Brookfield, DVII+
Pro), which allowed the use of small sample volumes of
approximately 1 mL. The sample was maintained at the
desired temperature with a recirculating water bath (VWR) for
at least 15 min prior to measurement and five separate
measurements at each temperature were obtained and
averaged. Viscosity standards (Cannon) were used to calibrate
the instrument before ionic liquid viscosities were measured.
To better understand the thermal stabilities of the latter
compounds, the imidazolium based congener of compound
1e, namely [Hmim][TF2N], was subjected to the same tests in
order to provide a direct comparison of properties between the
two compounds. Unlike compound 1e, [Hmim][TF2N] dis-
played the same onset temperature for decomposition under
either air (377 uC) or N2 (378 uC). However, compound 1f
displayed a higher stability under inert atmosphere (333 uC) as
compared to the values reported within the literature for
[Omim][Tf2N] (325 uC).41 Similarly, compound 1g displayed
lower decomposition temperatures in air than in N2, and
significantly lower thermal stability compared to its imidazo-
lium analog [Bmim][PF6] (y360 uC).42 In general, these results
indicate a decreased thermal stability of triazolium-based ILs
as compared to imidazolium ILs. This observation was further
examined by the isothermal thermogravimetric measurements
for compounds 1d and [Bmim][TF2N] within 150 uC to 250 uC
for 6 h at each selected temperature (see Table 2). After 30 h,
compound 1d exhibited a weight loss of y23 wt%, nearly 2.5
times that of [Bmim][TF2N]. It is possible that this decrease in
stability arises from the unique reactivity and stability
associated with certain triazolium-based cations.43–47 For
example, in previous studies the methylene groups of alkyl
chains directly bound to triazolium cations were found to be
particularly reactive.43 It is plausible that this increased
reactivity would lead to additional decomposition pathways
under high temperature, accounting for the observed decrease
in stability.
Calculations
The COSMOtherm program was used to predict gas solubility
in the ionic liquids. The COSMOtherm program34 is based on
the COSMO-RS (COnductor-like Screening MOdel for Real
Solvents)35 theory which links the quantum mechanically
calculated, solvated molecular surfaces from the COSMO
method to interacting molecular surfaces and macroscopic
properties such as solubility via a statistical thermodynamics
model. Full COSMO optimization of the molecular structure
was obtained with Turbomole V6.0 at the level of density
functional theory (DFT) with the B-P86 functional for electron
correlation and exchange36 and using the resolution-of-
identity (RI) technique and the large TZVP basis set.
Furthermore, the five benzyl-bearing ILs exhibited much
lower thermal stabilities than the alkyl-substituted com-
pounds, with almost 100 uC difference in the onset of the
decomposition curves. Among the benzyl-containing ILs, the
lowest Tdec was observed for 2d (253 uC) while the highest was
recorded for 2c (294 uC). In general, those compounds with
fluorine in the substituting groups were more thermally stable
than the methylated compounds 2d and 2e. In addition, an
ortho-substitution in the benzyl ring resulted in slightly higher
decomposition temperature than the two other fluorinated
para-substituted compounds. Finally, as with the compound 1
series, benzyl-substituted triazolium ILs show lower thermal
stability than imidazolium congeners. For instance, com-
pound 3-benzyl-1-methyl-1H-imidazol-3-ium bis((trifluoro-
methyl)sulfonyl)amide, [BzMim][TF2N], was synthesized and
found to decompose at 367 uC, which is nearly 100 uC above
the decomposition temperature found for compound 2a. Also,
the isothermal thermogravimetric tests for compound 2a and
[BzMim][TF2N] showed a total weight loss of 82.2 wt% for the
triazolium compound, compared to only 4.09 wt% for its
imidazolium congener (see Table 2).
Synthesis of ionic liquids, ESI
3
Synthesis of 1,2,4-triazolium ionic liquids were carried out
using standard literature procedures for imidazolium based
ILs.29 Full synthetic details are presented in the ESI section.
3
Briefly, the appropriate bromo alkane was added dropwise
to 1-methyl-1,2,4-triazole, and the mixture stirred for a
minimum of 24 h. The resultant solid was washed, dissolved
in water and decolorized with carbon. The aqueous solution
was filtered and lithium bis(trifluoromethanesulfonimide) was
added, resulting in the formation of two phases. The formed
IL was separated from the aqueous phase and purified.
Discussion
Thermal properties
Thermogravimetric (TG) and calculated derivative curves
(DTG) for compounds 1a–g and 2a–e recorded in flowing
nitrogen atmosphere are shown in Fig. 2, and a summary of
the measured physicochemical properties is given in Table 1.
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 3981–3989 | 3983

View Article Online
Paper
RSC Advances
Fig. 2 TG and DTG curves in nitrogen for the alkyl substituted triazolium ILs (A) and (B), and benzyl substituted triazoliums (C) and (D), respectively.
Table 1 Physicochemical parameters for the triazolium based ionic liquids and calculated gas absorption data
a
b
c
d
e
g
h
D
Tg
Tmp
(uC)
DHmp
Tdec
g
PermCO2
(barrer)
SolCO2
SelCO2/N2 (mol L21 atm21
KH,px
c
f
Compound
(g cm23
)
(uC)
(J g21
)
N2 (uC)
Tdec Air (uC) (cP)
)
(atm)
1a
1b
1c
1d
1e
1f
1g
2a
2b
2c
2d
2e
—
269.4
268.1
271.1
265.2
270.6
271.2
—
246.0
244.6
245.7
237.3
234.4
287
50.0 59.8
358
290
351
343
344
338
308
271
288
294
253
290
423
367
359
287
348
347
327
333
303
270
285
293
252
289
—
—
—
—
—
0.0778
0.0813
0.0713
0.0766
0.0755
—
1.484
1.485
1.489
1.372
1.314
—
1.488
1.590
1.598
1.420
—
—
—
107 ¡ 1 641 ¡ 11 24 ¡ 2
102 ¡ 1 786 ¡ 14 28 ¡ 1
246 ¡ 3 483 ¡ 15 22 ¡ 1
130 ¡ 3 804 ¡ 18 22 ¡ 1
45.0
41.9
45.4
37.8
35.5
—
44.4
34.1
38.8
43.4
52.8
—
31.7 58.9
—
—
—
—
17.4 41.6
55.3 32.3
154 ¡ 5
859 ¡ 9 17 ¡ 1
—
—
—
—
—
—
—
—
—
23
—
—
—
—
—
—
—
—
317 ¡ 10 304 ¡ 10 21 ¡ 3
530 ¡ 33 407 ¡ 10 22 ¡ 1
573 ¡ 23 277 ¡ 11 19 ¡ 3
839 ¡ 34 223 ¡ 19 19 ¡ 2
0.0679
0.0819
0.0689
0.0647
0.0492
0.090
2048 ¡ 51
108 ¡ 7 12 ¡ 3
Bmim[TF2N]i
BzMim[TF2N]i
1.436
1.42
52
61
1344
528
19.7
31.4
—
—
0.080
—
a
b
c
Measured density. Glass transition temperature from DSC. Decomposition temperature taken from the maximum of the first peak on the
d
e
f
g
DTG curves. Viscosity. Carbon dioxide permeability. CO2 selectivity calculated from the permeability ratios of CO2 to N2. CO2 solubility.
h
i
Henry’s Law constant calculated according to KH,XP = p/x, where p is the pressure and x the mol fraction of CO2 in a given IL. Data taken
from ref. 37–40
3984 | RSC Adv., 2013, 3, 3981–3989
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
Table 2 Results from long-term thermal stability testa on compounds 1d and 2a as compared with imidazole based ionic liquids
150 uC
180 uC
200 uC
220 uC
250 uC
Total weight lossb
1d
2a
0.015
0.051
0.20
0.20
1.6
0.084
0.019
0.65
6.8
0.26
0.085
2.9
4.9
1.2
0.58
23.6
82.2
9.36
4.09
0.074
Bmim[TF2N]
BzMim[TF2N]
2.2 6 1023
1.2 6 1023
0.022
1.6 6 1023
a
b
(Units given in wt% h21). Total weight lost for entirety of test (wt%).
The DSC results for the observed melting transitions and
glass transition temperatures are summarized in Table 1.
While only compounds 1a, 1c, 1f and 1g exhibited melting
transitions (see Fig. 3), glass transitions were observed for all
the highest for 2e at 234.4 uC. For the latter series of
compounds, the lowest glass transition temperatures were
comparable to the unsubstituted benzyl ring, with or without
the addition of an –OCF3 in either the ortho or para positions.
Shifts towards higher Tg values were found for fluorinated and
protonated rings with methoxy and methyl moieties in para
positions, respectively. For instance, compounds 2b and 2c,
showed minimal shift in Tg despite the change in position of
the perfluoromethoxy group. Similarly, compounds 1b and 1c,
differing by the substitution of a propyl chain, displayed very
small differences in the Tg, despite the significant decrease in
thermal stability associated with steric changes in alkyl chains
discussed previously, i.e. isopropyl (1b) to propyl (1c). In
addition to glass transitions, compounds 1a, 1c, and 1f
displayed cold crystallization transitions attributed to mole-
cular rearrangements.51
ILs (see ESI Fig. S3–S13). The melting points for the
3
investigated triazolium compounds were higher than those
for their corresponding imidazolium based congeners. For
instance, compound 1a displayed an average melting point of
50.0 uC, well above the previously reported melting point of
approximately 216 uC for [Emim][Tf2N].48
The melting point for compound 1c was 31.7 uC, whereas
[Prmim][TF2N] exists as
a free flowing liquid at room
temperature.48 The melting point for compound 1f was as
low as 217.4 uC, y4 uC above that reported for [Omim][TF2N]
(223.3 uC).49 Compound 1g was a solid at room temperature
and had an observed melting point of 55.3 uC, which is also
significantly higher than [Bmim][PF6] (11 uC).50 Interestingly,
the increasing length of the substituting alkyl chain did not
dramatically change the glass transition temperatures for this
series of compounds, being observed within the range of 271
uC to 265 uC. Those were far lower than the glass transition
temperatures observed for the benzyl containing ILs, for which
the lowest Tg was observed in compound 2a at 246.0 uC and
Viscosity
Temperature dependent viscosities for the RTILs within 25 uC
to 60 uC are shown in Fig. 4. Compound 1c had the lowest
viscosity at 25 uC (102 cP), while 2e displayed the highest at the
same temperature (2047 cP). In general, the alkyl bearing ILs
displayed lower viscosity than those containing aryl groups,
whereas this parameter increased within the alkyl containing
ILs with increasing chain length, with a maximum value for
the butyl substituted compound 1d. Compounds 2b and 2c
varying only by the location of the perfluoromethoxy group,
displayed nearly identical viscosities, thus indicating that an
ortho or para substitution on the aryl ring did not affect this
property. In addition, the viscosity of compound 2d, varying by
addition of methyl groups, was more than twice that of
compound 2a. Similarly, the fluorinated methoxy derivative,
compound 2e, exhibited more than six times the viscosity of
compound 2a.
When compared to imidazolium-based ILs, triazolium ILs
display significantly higher viscosities. With the exception of
1d, viscosity values of the alkyl triazolium ILs are approxi-
mately double that of their reported imidazolium congeners
(see Table 1 for a comparison).48,52,53 Given the small increase
in molecular mass from the substitution of one nitrogen for a
carbon in these heterocycles, it is likely that the increases in
viscosity arises from an increase in the intra- and inter-
molecular interactions.
Gas separations
CO2 solubility curves for the ionic liquids are displayed in
Fig. 5, and calculated parameters are summarized in Table 1.
Permeability for the entire series of ILs ranged from 108
Fig. 3 DSC curves for selected compounds 1a, 1c, 1f, and 1g, with the first three
exhibiting cold crystallization (exothermic peaks), and all four showing well-
defined melting transitions (endothermic peaks).
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 3981–3989 | 3985

View Article Online
Paper
RSC Advances
Fig. 5 High pressure CO2 solubility isotherms at 25 uC for the alkyl-bearing (A)
and the aryl-bearing (B) ionic liquid.
Selectivity of the ILs ranged from a low of 12 for compound
2e to 28 for 1c, placing the selectivity of triazolium-based ILs in
the range typically observed for the standard imidazolium
based compounds.54 The two compounds with the lowest
viscosity, namely 1b and 1c, display the highest selectivity.
Between those compounds, the n-propyl substitution not only
resulted in higher selectivity, but also increased CO2 solubility
compared to the isopropyl substituted compound. Similarly,
the para-substituted –OCF3 aryl compound, 2b, displayed
higher selectivity, and solubility than the ortho-substituted
–OCF3 compound 2c. Among the aryl-bearing ILs, compounds
2b and 2c have the highest solubility values as compared with
the others. This increase in solubility can be attributed to the
presence of perfluoromethyl groups which have been shown to
have increased affinity to CO2.55
For the case of these triazolium ILs, solubilities of the alkyl-
bearing ILs display a correlation with viscosity, where the
highest viscosity compound, 1d, had a lower solubility (0.0713
mol L21 atm21) than the less viscous samples. Furthermore, a
y6% increase in solubility was achieved by replacing the butyl
chain with an octyl chain (compound 1f), which in turn is
nearly 37% less viscous than compound 1d. Finally, the
viscosity of compound 2e, which is too high to readily facilitate
Fig. 4 Temperature dependent viscosity for compounds 1b–1f (A) and
compounds 2a–e (B)
barrer, for 2e, to 859 barrer for 1f. In general, the alkyl-bearing
ILs, compounds 1b–f, showed significantly higher permeabil-
ities as compared to the aryl-bearing ILs, which likely results
from the lower viscosities of the former compounds. Within
both sets of compounds, the CO2 permeability decreased with
increasing viscosity. In general, for imidazole-based ILs with
the same anion, increasing the substituting alkyl chain length
leads to increases in viscosity, and consequently lower CO2
permeability.37 Such a trend within compound 1 series is less
obvious, with a sudden increase in viscosity for the butyl
derivative. Interestingly, despite the similar viscosity values
measured for compounds 1b and 1c, with isopropyl and
n-propyl chains, respectively, the latter exhibited 23% higher
permeability than the former. Similarly, compounds 2b and
2c, which vary only by locational substitution of an O–CF3
group and having similar viscosity values, reveal that the ortho-
substituted IL (2b) has y1.5 times higher CO2 permeability
than the para-derivative (2c).
3986 | RSC Adv., 2013, 3, 3981–3989
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
Table 3 Physical parameters of triazolium ionic liquids as predicted by COSMOtherm
a
b
c
d
Compound Mol. weight (g mol21
)
D (g cm23
)
Molar volume (Å3)
Vcosmo (Å3)
Fractional free volume
SolCO2 (mol L21 atm21
)
1a
1b
1c
1d
1e
1f
1g
2a
2b
2c
2d
2e
392.20
406.33
406.33
420.35
448.41
476.46
285.17
454.37
538.37
538.37
482.42
556.36
1.571
414.6
443.9
445.0
471.4
525.9
583.2
341.3
491
561
563
551
569
369.5
392.8
393.8
413.6
456.8
500.9
296.8
445
501
504
489
509
0.109
0.115
0.115
0.123
0.131
0.141
0.130
0.103
0.107
0.105
0.113
0.105
0.073
0.080
0.080
0.081
0.083
0.081
0.072
0.082
0.071
0.077
0.086
0.068
1.520
1.516
1.481
1.416
1.357
1.387
1.522
1.594
1.588
1.454
1.624
a
b
d
calculated density. Volume enclosed by COSMO surface.57 c FFV = (molar volume 2 Vcosmo)/molar volume. Calculated CO2 solubility.
CO2 diffusion, displayed the lowest selectivity and the lowest
CO2 solubility among all the compounds investigated. Of
course, viscosity is not the only property that affects perme-
ability and solubility. It has been observed that molar free
volume has an influence on gas permeability in ionic liquids.54
Hence, the CO2 separations properties of the triazolium ILs are
likely influenced by a combination of many physical proper-
ties, such as changes in cation–anion interactions and the
segregation of the charged species forming porous networks.24
crepancies between the experimental values as compared to
the values found for the alkyl-bearing compounds.
Interestingly, the calculated solubility values decrease when
comparing the structurally related compounds 2b and 2c,
which is the opposite of experimental results, leading to the
conclusion that further refinement of computational methods
is required.
As has been previously described in the literature, predicting
the solubility of gases in ILs poses significant challenges as
there are multiple inter- and intra-molecular interactions
which effect the solubility. As reported by Bara et al.,58
fractional free volume (FFV) of ILs has been shown to correlate
inversely to the solubility and selectivity of gases within
imidazolium based IL systems. In an effort to further examine
the solubility changes imparted by the locational substitutions
within the triazolium ILs, fractional free volumes for all the
triazolium compounds were calculated using COSMOtherm,
with a summary of relevant values shown in Table 3. With
respect to the alkyl-bearing ILs, calculated CO2 solubility
remained approximately the same despite seeing increases in
the FFV. Correlation between the calculated and experimental
values for the aryl-containing ILs proves to be difficult as no
exact trends are readily observed. The constitutional isomers,
namely compounds 2b and 2c, were predicted to have nearly
identical FFV values of 0.107 and 0.105 respectively, yet as
discussed previously show significantly different solubility
values. The two methyl substituted aryl ring systems, namely
2a and 2d, were calculated to have the lowest and highest FFV
values at 0.103 and 0.113 respectively, a change of approxi-
mately 10%. As expected, the experimental solubility values are
inversely proportional to the FFV values, with compound 2d
having the lower solubility values than 2a.
COSMOtherm calculations
In an effort to better understand the changes in CO2 solubility
associated with the triazolium-based ILs, the compounds were
examined via computational methods using COSMO-RS,
which has been widely employed in previous studies to
examine solubility and selectivity of gases in ILs.25,56,57 The
calculated parameters for all compounds are summarized in
Table 3. Several structural aspects are known to affect CO2
solubility in ionic liquids, such as increasing alkyl chain
length as well as addition of electronegative atoms.25,56,57 As
can be seen from the values in Tables 1 and 3, the
experimental solubility values of the alkyl bearing ILs show
good correlation with the calculated values. There is a good
correlation between the calculated and experimental values for
the alkyl bearing ILs, with calculated CO2 solubility values
being slightly higher than the experimental ones. Such
differences are acceptable, given the existence of errors
associated to experimental measurements. Compounds 1g
and 1d exist as solids at room temperature, precluding the
possibility of obtaining experimental solubility data. COSMO-
RS calculations, however, predict a solubility of 0.072 mol L21
atm21 for 1g and 0.081 mol L21 atm21 for 1d. Based on these
data we can conclude that the PF6 anion in compound 1g
lowers CO2 solubility as compared with the TF2N based ionic
liquids. This is in agreement with previous literature reports
which demonstrate that PF6 containing ionic liquids have
lower CO2 solubility as compared to their TF2N congeners.40
The calculated values for CO2 solubility of the aryl-contain-
ing ILs show the expected trends,25 with the alkyl substituted
benzene rings having higher CO2 solubility (2a and 2d) than
the fluorine-bearing benzene rings (2b, 2c, 2e). In general, the
calculated values for the aryl-bearing ILs show larger dis-
Overall, the use of COSMOtherm as a method for predicting
the solubility of CO2 within triazolium-based ILs is a viable
method, with calculated values showing reasonable agreement
with experimental ones in respect to alkyl containing ILs. The
aryl-containing ILs show some correlation between experi-
mental and calculated results, with the simpler aliphatic
bearing arene systems appearing to show higher correlation
than the more complex fluorinated systems. It is likely that
further enhancements to current computational methods are
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 3981–3989 | 3987

View Article Online
Paper
RSC Advances
5 J. E. Bara, T. K. Carlisle, C. J. Gabriel, D. Camper,
A. Finotello, D. L. Gin and R. D. Noble, Ind. Eng. Chem.
Res., 2009, 48, 2739–2751.
required in order to more accurately predict solubility
selectivity values for these systems.
6 J. H. Davis and K. J. Forrester, Tetrahedron Lett., 1999, 40,
1621–1622.
7 S. Sanghi, E. Willet, C. Versek, M. Tuominen and E.
B. Coughlin, RSC Adv., 2012, 2, 848–853.
Conclusions
A series of twelve novel 1,2,4-triazolium-based ionic liquids
have been synthesized and their physicochemical properties
studied. Ten of the compounds remained as free flowing
liquids at room temperature, effectively expanding the avail-
able library of room-temperature ILs. The ILs displayed high
thermal stability under both ambient and inert atmospheres.
Overall, while triazolium ILs showed higher viscosities, these
displayed CO2 solubility and selectivity values comparable to
imidazolium-based ILs. Previous reports on imidazolium-
based IL showed a direct correlation between increasing CO2
selectivity and solubility properties with viscosity and length of
substituting alkyl groups. In the present study, the triazolium
compounds displayed more complex correlations. For
instance, the CO2 adsorption properties greatly changed for
isomeric compounds having similar viscosity values. This
effect was clearly seen for isopropyl and n-propyl substituted
triazolium compounds, as well as for ortho and para-substitu-
tions of the –OCF3 group in aryl-derivatives. Also, addition of
perfluoro groups appears to have a beneficial effect on the
solubility of CO2, as compared to fully protonated, and to
fluorinated benzyl groups. COSMOtherm calculations showed
significant agreement between calculated and experimental
values for the alkyl-bearing ILs. Triazolium-bearing aryl
substitutions proved to be a challenge to model properly
using the currently established parameters in COSMOtherm.
The latter results demonstrate that further refinement is
needed for accurately modeling more complex systems. We are
currently using the experimental values acquired for the ILs in
this study to develop new methods to increase the validity of
computational parameters for 1,2,4-triazolium ILs. Therefore,
triazolium-based cations are important model systems for
developing a more complete understanding of small changes
in the ionic volume of cations, as well as shifts in the anion–
cation interactions which are crucial for the development of
ILs with enhanced separation properties.
8 M. W. Schmidt, M. S. Gordon and J. A. Boatz, J. Phys. Chem.
A, 2005, 109, 7285–7295.
9 J. Z. Vlahakis, S. Mitu, G. Roman, E. P. Rodriguez, I.
E. Crandall and W. A. Szarek, Bioorg. Med. Chem., 2011, 19,
6525–6542.
10 N. Rios-Lombardia, R. Porcar, E. Busto, I. Alfonso,
J. Montejo-Bernardo, S. Garcia-Granda, V. Gotor, S.
V. Luis, E. Garcia-Verdugo and V. Gotor-Fernandez,
ChemCatChem, 2011, 3, 1921–1928.
11 P. J. Carlson, S. Bose, D. W. Armstrong, T. Hawkins, M.
S. Gordon and J. W. Petrich, J. Phys. Chem. B, 2012, 116,
503–512.
12 Q. H. Lin, Y. C. Li, Y. Y. Li, Z. Wang, W. Liu, C. Qi and S.
P. Pang, J. Mater. Chem., 2012, 22, 666–674.
13 P. K. Swain, H. Singh and S. P. Tewari, J. Mol. Liq., 2010,
151, 87–96.
14 R. E. Baltus, R. M. Counce, B. H. Culbertson, H. M. Luo, D.
W. DePaoli, S. Dai and D. C. Duckworth, in 13th Symposium
on Separation Science and Technology for Energy Applications,
Gathinburg, TN, 2003, pp. 525–541.
15 D. Camper, J. E. Bara, D. L. Gin and R. D. Noble, Ind. Eng.
Chem. Res., 2008, 47, 8496–8498.
16 J. H. Davis, Chem. Lett., 2004, 33, 1072–1077.
17 B. E. Gurkan, J. C. de la Fuente, E. M. Mindrup, L. E. Ficke,
B. F. Goodrich, E. A. Price, W. F. Schneider and J.
F. Brennecke, J. Am. Chem. Soc., 2010, 132, 2116–2117.
18 M. J. Muldoon, S. Aki, J. L. Anderson, J. K. Dixon and J.
F. Brennecke, J. Phys. Chem. B, 2007, 111, 9001–9009.
19 H. B. Nulwala, C. N. Tang, B. W. Kail, K. Damodaran,
P. Kaur, S. Wickramanayake, W. Shi and D. R. Luebke,
Green Chem., 2011, 13, 3345–3349.
20 M. B. Shiflett, D. W. Drew, R. A. Cantini and A. Yokozeki,
Energy Fuels, 2010, 24, 5781–5789.
21 C. M. Teague, S. Dai and D. E. Jiang, J. Phys. Chem. A, 2010,
114, 11761–11767.
22 C. M. Wang, H. M. Luo, D. E. Jiang, H. R. Li and S. Dai,
Angew. Chem., Int. Ed., 2010, 49, 5978–5981.
23 D. Wappel, G. Gronald, R. Kalb and J. Draxler, Int. J.
Greenhouse Gas Control, 2010, 4, 486–494.
24 Y. F. Hu, Z. C. Liu, C. M. Xu and X. M. Zhang, Chem. Soc.
Rev., 2011, 40, 3802–3823.
25 K. Z. Sumon and A. Henni, Fluid Phase Equilib., 2011, 310,
39–55.
26 C. Cadena, J. L. Anthony, J. K. Shah, T. I. Morrow, J.
F. Brennecke and E. J. Maginn, J. Am. Chem. Soc., 2004, 126,
5300–5308.
Acknowledgements
This work was fully sponsored by the Division of Chemical
Sciences, Geosciences, and Biosciences, Office of Basic Energy
Sciences, U. S. Department of Energy.
References
27 S. Aparicio, M. Atilhan and F. Karadas, Ind. Eng. Chem. Res.,
2010, 49, 9580–9595.
28 M. Diedenhofen and A. Klamt, Fluid Phase Equilib., 2010,
294, 31–38.
29 L. Vidal, M. L. Riekkola and A. Canals, Anal. Chim. Acta,
2012, 715, 19–41.
1 J. S. Wilkes, Green Chem., 2002, 4, 73–80.
2 V. I. Parvulescu and C. Hardacre, Chem. Rev., 2007, 107,
2615–2665.
3 T. Welton, Chem. Rev., 1999, 99, 2071–2083.
4 M. Smiglak, A. Metlen and R. D. Rogers, Acc. Chem. Res.,
2007, 40, 1182–1192.
3988 | RSC Adv., 2013, 3, 3981–3989
This journal is ß The Royal Society of Chemistry 2013

View Article Online
RSC Advances
Paper
30 H. Jin, B. O’Hare, J. Dong, S. Arzhantsev, G. A. Baker, J.
F. Wishart, A. J. Benesi and M. Maroncelli, J. Phys. Chem. B,
2008, 112, 81–92.
31 T. Fujimori, K. Fujii, R. Kanzaki, K. Chiba, H. Yamamoto,
Y. Umebayashi and S. I. Ishiguro, J. Mol. Liq., 2007, 131,
216–224.
46 G. G. Surpateanu, G. Vergoten and G. Surpateanu, J. Mol.
Struct., 2000, 526, 143–150.
47 Z. Zeng, B. Twamley and J. M. Shreeve, Organometallics,
2007, 26, 1782–1787.
48 U. L. Bernard, E. I. Izgorodina and D. R. MacFarlane, J.
Phys. Chem. C, 2010, 114, 20472–20478.
32 S. Pitula and A. V. Mudring, Phys. Chem. Chem. Phys., 2010,
12, 7056–7063.
33 S. M. Mahurin, J. S. Lee, G. A. Baker, H. M. Luo and S. Dai,
J. Membr. Sci., 2010, 353, 177–183.
34 F. Eckert and A. Klamt, COSMOlogic GmbH & Co. KG,
Leverkusen, Germany, 2011.
35 F. Eckert and A. Klamt, AIChE J., 2002, 48, 369–385.
36 R. Ahlrichs, F. Furche and S. Grimme, Chem. Phys. Lett.,
2000, 325, 317–321.
37 P. Scovazzo, J. Membr. Sci., 2009, 343, 199–211.
38 T. Buffeteau, J. Grondin, Y. Danten and J. C. Lassegues, J.
Phys. Chem. B, 2010, 114, 7587–7592.
39 S. M. Mahurin, T. Dai, J. S. Yeary, H. M. Luo and S. Dai, Ind.
Eng. Chem. Res., 2011, 50, 14061–14069.
49 T. J. Hughes, T. Syed, B. F. Graham, K. N. Marsh and E.
F. May, J. Chem. Eng. Data, 2011, 56, 2153–2159.
50 C. P. Fredlake, J. M. Crosthwaite, D. G. Hert, S. Aki and J.
F. Brennecke, J. Chem. Eng. Data, 2004, 49, 954–964.
51 J. G. Huddleston, A. E. Visser, W. M. Reichert, H.
D. Willauer, G. A. Broker and R. D. Rogers, Green Chem.,
2001, 3, 156–164.
52 D. Almantariotis, T. Gefflaut, A. A. H. Padua, J. Y. Coxam
and M. F. C. Gomes, J. Phys. Chem. B, 2010, 114, 3608–3617.
53 A. Muhammad, M. I. A. Mutalib, C. D. Wilfred,
T. Murugesan and A. Shafeeq, J. Chem. Thermodyn., 2008,
40, 1433–1438.
54 P. Luis, T. V. Gerven and B. V. d. Bruggen, Prog. Energy
Combust. Sci., 2012, 38, 419–448.
40 J. L. Anthony, J. L. Anderson, E. J. Maginn and J.
F. Brennecke, J. Phys. Chem. B, 2005, 109, 6366–6374.
41 N. Papaiconomou, N. Yakelis, J. Salminen, R. Bergman and
J. M. Prausnitz, J. Chem. Eng. Data, 2006, 51, 1389–1393.
42 D. Fang, H. Cheng, K. Gong, Q. R. Shi, X. L. Zhou and Z.
L. Liu, J. Fluorine Chem., 2008, 129, 108–111.
43 I. Badea, G. G. Surpateanu, P. Woisel, G. Vergoten and
G. Surpateanu, New J. Chem., 2002, 26, 1658–1661.
44 C. Cadena and E. J. Maginn, J. Phys. Chem. B, 2006, 110,
18026–18039.
55 J. E. Bara, C. J. Gabriel, T. K. Carlisle, D. E. Camper,
A. Finotello, D. L. Gin and R. D. Noble, Chem. Eng. J., 2009,
147, 43–50.
56 R. Babarao, S. Dai and D. E. Jiang, J. Phys. Chem. B, 2011,
115, 9789–9794.
57 K. N. Marsh, J. A. Boxall and R. Lichtenthaler, Fluid Phase
Equilib., 2004, 219, 93–98.
58 M. S. Shannon, J. M. Tedstone, S. P. O. Danielsen, M.
S. Hindman, A. C. Irvin and J. E. Bara, Ind. Eng. Chem. Res.,
2012, 51, 5565–5576.
45 G. G. Surpateanu, F. Delattre, P. Woisel, G. Vergoten and
G. Surpateanu, J. Mol. Struct., 2003, 645, 29–36.
This journal is ß The Royal Society of Chemistry 2013
RSC Adv., 2013, 3, 3981–3989 | 3989