Preparation of protic ionic liquids with minimal water content and 15N NMR study of proton transfer

Article abstract of DOI:10.1039/b921432a

Low-molecular-weight Bronsted acids and amine bases were used to reproducibly prepare very dry, high-purity room-temperature protic ionic liquids (PILs). A series of eight amine bases and six Bronsted acids were combined to produce 48 mixtures, of which 18 were liquid at room temperature. The phase transitions and thermal decomposition temperatures were determined for each mixture; whereas viscosity, density and conductivity were determined for the room-temperature liquids. By utilising ¹⁵N NMR it was possible to distinguish between neutral and ionised amine bases (ammonia vs. ammonium-type ion), which indicated that the protic ionic liquids were completely ionised when made as a stoichiometric mixture. However, a Walden plot comparison of fluidity and molar conductivity indicated the majority of PILs had much lower conductivity than predicted by viscosity unless the base contained excess proton-donating groups. This disparity is indicative of protic ionic molecules forming neutral aggregates or non-Newtonian fluid hydrogen-bonded networks with a secondary Grotthuss proton-hopping mechanism arising from polyprotic bases. the Owner Societies 2010.

Full text of DOI:10.1039/b921432a

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/pccp | Physical Chemistry Chemical Physics
Preparation of protic ionic liquids with minimal water content
15
and N NMR study of proton transferw
a
b
c
d
Geoffrey L. Burrell,* Iko M. Burgar, Frances Separovic and Noel F. Dunlop
Received 13th October 2009, Accepted 27th November 2009
First published as an Advance Article on the web 12th January 2010
DOI: 10.1039/b921432a
Low-molecular-weight Brønsted acids and amine bases were used to reproducibly prepare very
dry, high-purity room-temperature protic ionic liquids (PILs). A series of eight amine bases and
six Brønsted acids were combined to produce 48 mixtures, of which 18 were liquid at room
temperature. The phase transitions and thermal decomposition temperatures were determined
for each mixture; whereas viscosity, density and conductivity were determined for the
1
5
room-temperature liquids. By utilising N NMR it was possible to distinguish between neutral
and ionised amine bases (ammonia vs. ammonium-type ion), which indicated that the protic ionic
liquids were completely ionised when made as a stoichiometric mixture. However, a Walden plot
comparison of fluidity and molar conductivity indicated the majority of PILs had much lower
conductivity than predicted by viscosity unless the base contained excess proton-donating groups.
This disparity is indicative of protic ionic molecules forming neutral aggregates or non-Newtonian
fluid hydrogen-bonded networks with a secondary Grotthuss proton-hopping mechanism arising
from polyprotic bases.
1
. Introduction
As PILs are formed by neutralisation of a molecular acid
and base species, a possible equilibrium may exist which can
return the components to the un-ionised state [eqn (1)]. The
reported low ionic conductivities of many PILs have suggested
that either proton transfer is incomplete and significant neutral
molecular species remain in the PILs or that neutral ion-pairs
The field of ionic liquids (ILs) has been rapidly expanding, in
concurrence with the number of applications in which ILs
1
have been used. However, for some of the more frequently
encountered ILs there are several aspects which are seen as
undesirable for certain applications. Such negative aspects
include trace halogen impurities from metathesis, the potential
requirements of a toxicological profile for new materials, and
are formed, possibly even as larger multi-ion aggregates
10
[eqn (1)]. Different techniques have been used to provide
qualitative information about the overall ionic nature of bulk
11
2
the cost of ILs relative to molecular solvents.
PILs from the transport properties, including NMR, IR
13
One method to bypass these potential problems is to use
protic ionic liquids (PILs), which are formed by a neutralisation
3
reaction of Brønsted acids and bases. As has been often
12
spectroscopy, and ionic conductivity via a Walden plot.
However, there has been no method to clearly distinguish
between the ionic states of a PIL. Hence, a method
of confirming proton transfer is necessary to identify the
ionisation states of species in solution rather than the overall
ionisation of the PIL.
stated, in 1914 Walden reported the synthesis of ethylammonium
nitrate by transferring a proton from a Brønsted acid to a
4
Brønsted base without the use of a solvent. The acid and base
typically used are commercially available Brønsted acids and
bases with a long history of use in industry, where the
neutralisation reactions are able to produce a final mixture
ꢀ
þ 0
ꢀ
þ
Acid ðHÞ þBase Ð ½Acid ;ðHÞBase ꢁ Ð Acid þðHÞBase
Molecular Neutral ion pair Dissociated ions
ð1Þ
5
free of contaminant species.
Much effort has been made to exploit these PILs in applications
6
7
8
such as reaction media, catalysis, and fuel cells. Although a
large number of PILs have been prepared and characterised to
date, there is still debate on the ionic nature of the PILs once
Finally, it is often assumed that PILs (and to a lesser extent,
aprotic ILs) are produced as an equal binary pair. In our
experience, when reagents were diluted in sufficient solvent
to facilitate removal of the exothermic energy from the
neutralisation reaction, removal of solvent in vacuo also
9
the neutralisation is complete.
1
4
removed more of one of the PIL component than the other.
a
School of Chemistry, The University of Melbourne, Melbourne,
VIC 3010, Australia
CSIRO Materials Science & Engineering, Clayton, VIC 3169,
Australia
Regardless of the preparative method and the Brønsted acid
and base used, if less than 100% yield were obtained, it is naive
to assume that the product is an equimolar mixture. Therefore,
a method to produce PILs of known and predictable
stoichiometry was developed so as to accurately examine
ionisation and proton transfer, with an emphasis on the use
of low-molecular-weight volatile reagents.
b
c
School of Chemistry, The University of Melbourne, Melbourne,
VIC 3010, Australia. E-mail: fs@unimelb.edu.au
Orica Ltd., Melbourne, VIC 3002, Australia
d
13
w Electronic supplementary information (ESI) available: C NMR
spectra of (HOEt)
NHꢂAcOH. See DOI: 10.1039/b921432a
2
This journal is ꢃc the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 1571–1577 | 1571

View Article Online
We have investigated 48 unique mixtures using eight
common substituted amine bases and six Brønsted acids to
better understand proton transfer behaviour and the extent of
ionisation in PILs.
Differential scanning calorimetry (DSC) was performed
using a PerkinElmer (Waltham, MA) PYRIS Sapphire system
in a sealed aluminium pan under a nitrogen atmosphere.
Samples of mass 10–20 mg were run using three temperature
cycles and examined from ꢀ100 1C at a heating rate of 10 1C
ꢀ
1
min.
Density was determined using an Anton Paar (Ostfildern,
Germany) DMA 4100 M density meter. Dynamic viscosity
was conducted using a TA Instruments (New Castle, DE)
AR-G2 controlled stress cone and plate rheometer.
2
. Experimental
2
.1 Materials and chemicals
Formic (HO CH) and glacial acetic (AcOH) acids were
obtained from Ajax Chemicals (Auckland, New Zealand).
2
The ionic conductivity of the liquid samples was obtained at
2
0
1
5
1C by measuring the complex impedance between
.1 Hz and 1 MHz with a Solartron (Farnborough, UK)
260 immediate response analyzer. locally designed
Di-n-Butylamine (Bu NH), tri-n-butylamine (Bu
2
3
N), diethylamine
(
Et NH), triethylamine (Et N), di-n-propylamine (Pr NH),
2
3
2
A
diethanolamine ((HOEt) NH), triethanolamine ((HOEt) N),
2
3
conductivity cell made from two platinum wires was used
for the measurements. The cell constant was determined by
calibration using an aqueous 0.01 M KCl solution.
pyrrolidine (Pyrr), ethylenediamine (EDA), di-n-butyl-
phosphate ((BuO) PO H), methylsulfonic acid (MsOH),
2
2
saccharin (Sacc), and sulfamic acid (OSA) were obtained from
Sigma-Aldrich (St Louis, MO). EDA was shaken over freshly
baked alumina before fractional distillation. (HOEt)
HOEt) N were fractionally distilled twice and the middle two
thirds retained in each step. Other bases were purified
by fractional distillation twice from CaH under an inert
atmosphere and the middle two thirds retained. HO CH was
2
NH and
3
. Results and discussion
(
3
3.1 Preparation of high-purity PILs
2
Previous work has demonstrated the negative effect of vacuum
on component ratios of a protic ionic liquid, particularly those
2
refluxed with boric anhydride and fractionally distilled. AcOH
1
4
containing volatile starting reagents. Attempts to remove
solvent and trace water under moderately low pressures
was heated at reflux with KMnO with 4% acetic anhydride
4
and distilled twice. OSA was recrystallized twice from
deionised (mQ) water and Sacc was recrystallized twice from
ethanol—both were then dried by heating up to 100 1C in vacuo.
MsOH was diluted with water, treated with 2.5% w/w
(
0.3–0.4 mbar) removed volatile PIL species, leaving an
excess of the relatively non-volatile component and a non-
stoichiometric mixture. Despite the common use of an in vacuo
step to dry ILs, significant amounts of water (B1% w/w) can
still remain. It is possible that as water concentrations
minimise, the remaining water becomes strongly hydrogen
bonded or ionised in the ILs and, thus, is less likely to be
removed compared to some other species.
Ba(OH)
2
, then filtered and concentrated in vacuo. The resulting
before distillation
1
5
solution was stirred over 4% w/v P
2
O
5
under vacuum. All reagents were degassed by nitrogen bubbling
and stored away from light under an inert atmosphere.
The water content of PILs was determined by Karl–Fischer
titration using an 831 Metrohm (Herisau, Switzerland)
Karl–Fischer Coulometer. For room-temperature PILs a neat
sample of the PIL was injected for analysis and measurements
were averaged over 5 values (data not shown)—standard
deviations were less than 5%. Room-temperature solid PILs
Hence, in order to produce high-purity PILs of known
stoichiometry a simple method was selected to remove the
need for solvent and introduction of incidental water. Thus,
similar to reports of others, the Brønsted acid and base
1
6
reagents were mixed as neat liquids. However, neutralisation
is well known to be an exothermic reaction. In extreme cases,
the localised heating from adding a neat base to acid
volatilised some PIL components (i.e., loss of formic acid),
again leading to non-stoichiometric ratios. The exothermic
reaction was also observed to initiate the formation of other
by-products, e.g. the condensation reaction between formic
acid and secondary amines, which produces ionically neutral
1
4
were examined as per previous reports. Water content for the
majority of PILs was in the approximate range of 50–125 ppm,
with the exception of OSA-containing compounds which were
in the approximate range of 225–275 ppm.
2
.2 Physiochemical characterisation procedures
1
13
15
1
H, C and N NMR measurements of neat liquid samples
were acquired with a Varian (Palo Alto, CA) INOVA
amides and water. H NMR spectra revealed a relatively large
proportion of amide compared to amine/ammonium a short
time after addition of Pyrr to AcOH commenced (data not
shown).
5
00 MHz NMR spectrometer. The neat liquids were trans-
ferred under an inert atmosphere to a standard 5 mm NMR
tube and equipped with a Wilmad (Buena, NJ) NMR coaxial
capillary insert of volume 60 ml containing 1% (w/w) sodium
It was noted that the reaction exotherms decreased as the
mixture approached neutrality. Therefore, an optimal method
was found when the mixture remained consistently near neutral
(stoichiometric rather than pH neutral) by simultaneously
adding both neat reagents to the reactor. Indeed, by
simultaneously adding equivalent amounts of both acid and
base in a steady stream of drops with vigorous stirring, the
exothermic reaction heat was readily dissipated and no amide
3
-(trimethylsilyl)propionic acid-d
4
acid (TSP) in D
13
2
O solution
1
as a reference/lock solution. H and
C spectra were
1
externally referenced to TSP as 0 ppm. N NMR neat PIL
5
samples were externally referenced to saturated aq. NH Cl.
4
1
5
Standard N measurements were performed with a 451 pulse,
inverse-gated decoupling, repetition time 3 s, 1000–2000
transients, and 3 Hz line broadening.
1
was detected in the final mixture by H NMR. Localised
1
572 | Phys. Chem. Chem. Phys., 2010, 12, 1571–1577
This journal is ꢃc the Owner Societies 2010

View Article Online
heating was avoided and the addition rate was only limited by
the removal of overall uniform reaction heat. The addition
rate scaled with respect to the reaction volume and ranged
By using the mass of reagents, PIL ratios of acid and base
(mol%) were able to be estimated to an accuracy of 0.1%.
1
Although H NMR indicated both the purity and approximate
ꢀ1
between 0.2–10 ml min . The rate followed an approximate
stoichiometry, the required precision was inadequate for
high-purity results. Often some reagent loss was observed
from the mass difference between the product PILs and the
starting reagents, but no mixture gained mass during the
reaction. The difference error was applied as though both
the reagents had lost that equivalent amount of mass, and
when converted to a molar scale, typical relative errors in the
ratio were 0.1–0.2%.
ꢀ1
rule-of-thumb of n/5 ml min , where n was the total volume
of reagents added to the reactor and 1 r n r 200. It has been
estimated that several litres of PIL could be obtained by this
method in less than an hour by simply cooling using
circulating water.
When the starting reagents were purified, dried and handled
under inert atmosphere, minimal water was introduced into
the final mixture. In all reactions observed, significantly less
water was detected when compared to the same PILs made via
solvent dilution/removal in vacuo according to a previous
In total, of the 48 possible component mixtures examined,
18 formed homogeneous liquids at room temperature and
12 remained as biphasic mixtures.
1
4
method (Table 2). Water content was in the range of
B100 ppm for the majority of PILs, almost an order of
magnitude lower than produced by solvent removal in vacuo.
An exception was OSA-containing PILs where the higher
water content was attributed to the acid purification involving
recrystallisation from water. Degassing liquid reagents by
bubbling nitrogen and storing PILs away from light prevented
formation of colour as a result of amine browning. Hence,
two of the most significant contaminants that can alter the
properties of PILs were minimised. All PILs made by this
method together with their thermal properties are detailed in
Table 1.
1
3.2 Determination of ionisation by N NMR
5
Empirical studies to distinguish between H-bonding and
proton transfer of a protonated nitrogen have been attempted
1
using UV/VIS and H NMR spectroscopy, but the required
solvation and concentration changes are not possible for
1
PILs. It is known that the NMR chemical shift is dependent
7
1
8
on both concentration and the solvent dielectric constant.
Concentration is also an important factor in ion-pairing.
1
9
Hence, PILs were examined as neat liquids in order to observe
the properties of the ionic liquid, rather than the dynamics of a
dilute acid–amine solution.
PILs with solid reagents were made by controlled addition
of the neat liquid reagent to the dried and purified solid
reagent. To reduce the amount of any contaminating water,
the reagent and pre-weighed vessel were dried in a 100 1C oven
overnight and the loaded vessel cooled under inert gas. The
slow uptake of the solid reagent into the liquid phase
sufficiently reduced the rate of reaction such that the strong
exothermic reaction observed when two liquids were mixed,
was not observed with the solid reagents.
1
5
Another benefit of using neat liquid samples is that
N
1
NMR data could be acquired in a reasonable time. N is a
5
1
spin- nucleus, which simplifies magnetic relaxation compared
1
2
4
15
to N, a quadrupolar spin-1 nucleus, but N is only present
at 0.37% natural abundance, which tends to limit applications
1
5
unless enrichment or site-specific N labelling is included. In
the case of neat PILs, the large sample concentration in the rf
coil made possible determination of the nitrogen substitution
at the exchange site and to distinguish between a protonated
and non-protonated nitrogen (i.e., a substituted ammonia vs.
ammonium ion).
A significant drawback to using neat reagents occurred
when the product was not a liquid at room temperature. The
lack of a significant fluid phase resulted in poor mixing and
typically produced an inhomogeneous amorphous solid.
Nevertheless, heating the product above the melting point
allowed for the melt to mix homogeneously and gave products
with ratios consistent with that of the added components as
1
5
N NMR spectroscopy has several desirable, but not
always practical, characteristics when compared to other
1
13
20
15
nuclei (e.g. H, C). Intermolecular effects on N are very
important due to the presence of a lone pair on nitrogen.
Nitrogen-containing bonds often exhibit a high polarity,
which gives a chemical shift range in organic molecules that
spans up to 400 ppm. This makes changes in electronic density
1
13
confirmed by H and C NMR.
Some liquid phase acid/amine products remained inhomo-
geneous when made as a stoichiometric mixture (Table 1). On
the other hand, biphasic mixtures were not observed when
mixing occurred in a solvent medium. For instance, the
1
3
on the nitrogen atom much more sensitive than either
1
C
(B200 ppm) or H (B12 ppm) and has been used to estimate
the empirical order of ion pairing in amine–polyfluorinated
21
acid mixtures.
trialkylamines, triethylamine (Et N) and tributylamine
3
(
Bu N), when mixed stoichiometrically with AcOH formed
3
Due to fast intermolecular proton exchange, a single
5
discrete liquid layers, whereas the dialkylamines, diethylamine
Et NH) and dibutylamine (Bu NH), formed homogeneous
1
isotropic N signal was obtained which was an average of
protonated and molecular bases [Fig. 1(a)]. The neat base
(
2
2
mixtures. However, compounds with hydroxyl groups formed
homogeneous liquids regardless of the degree of nitrogen
substitution. The biphasic mixtures were characterised by
NMR to have a bottom acid layer with an excess of acid
1
5
d
measurements,
N measurement was used as a reference for subsequent
2
R NHþ
2
¼ K0. The maximum chemical shift was
R2NH
taken from a dilute solution of the base in excess neat acid,
R2NH2
R NH
þ
ꢄ K1. Despite the excess acid, it is possible that not all
(
typically 2 : 1 or 3 : 1 acid:base). The top layer consisted
2
almost entirely of un-reacted base, indicating that a mobile
proton may be required for neutralisation and that the neat
reagent was not sufficiently reactive.
of the base may have been ionised, consequently, this value
was taken to be the upper limit of proton transfer. Hence, the
concentration of acid required for ionisation of a neat base
This journal is ꢃc the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 1571–1577 | 1573

View Article Online
Table 1 Phase transitions of PILs upon heating from ꢀ100 1C at 10 1C minꢀ
1
Protic ionic liquid
Base:acid (%)
T
g
/1C
T
m
/1C
Phase transition/1C
(
(
(
(
(
(
(
(
(
(
(
(
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
Bu
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
2
2
2
2
2
2
3
3
3
3
3
3
NHꢂ(BuO)
NHꢂOSA
NHꢂSacc
2
PO
2
H
50.7
50.1
67.9
50.0
49.6
49.1
49.9
50.1
50.6
48.7
50.0
49.6
50.5
52.5
56.2
50.3
66.5
52.0
50.0
50.7
52.6
39.0
50.5
49.1
49.9
50.3
49.9
51.2
50.3
51.7
49.6
50.6
54.7
50.0
50.3
49.6
48.7
49.4
49.4
49.6
50.4
51.7
49.1
49.5
50.9
49.4
50.1
50.1
49.3
49.9
32.1
50.0
50.4
50.9
50.1
49.9
49.4
51.3
50.0
50.4
50.0
47.5
43.8
49.7
33.5
48.0
50.0
49.3
47.4
61.0
49.5
50.9
50.1
49.7
50.1
48.8
49.7
48.3
50.4
49.4
45.3
50.0
49.7
50.4
51.3
50.6
50.6
50.4
49.6
48.3
50.9
50.5
49.1
50.6
49.9
49.9
ꢀ71.4
ꢀ56.9
—
ꢀ58.5
ꢀ71.9
ꢀ62.2
ꢀ59.7
ꢀ58.6
ꢀ52.4
—
28.8
—
ꢀ39.0
—
—
a
NHꢂAcOH
—
—
29.0
43.2
48.4
—
—
NHꢂHO
NHꢂMsOH
Nꢂ(BuO) PO
NꢂAcOH
NꢂHO CH
2
CH
—
ꢀ11.1, 24.3
2
2
H
ꢀ14.2
2
63.3
—
—
a
NꢂSacc
NꢂMsOH
NꢂOSA
ꢀ78.6
69.9
—
—
a
—
—
2
2
2
2
2
2
3
3
3
3
3
3
NHꢂ(BuO)
2
PO
2
H
63.4
45.0
27.8
72.4
87
ꢀ65.1, ꢀ10.3, 31.4
NHꢂAcOH
—
ꢀ31.1, ꢀ12.3
NHꢂHO
2
CH
ꢀ76.6
ꢀ46.9
ꢀ38.9
ꢀ24.7
ꢀ82.9
—
—
26.2, 59.1, 68.8
NHꢂMsOH
NHꢂOSA
NHꢂSacc
26.5
16.0, 67.8
—
—
—
29.1
—
—
24.9
—
—
—
—
—
—
—
88.3
—
Nꢂ(BuO)
2
PO
2
H
a
NꢂAcOH
a
NꢂHO
2
CH
—
ꢀ25.5
NꢂMsOH
NꢂOSA
NꢂSacc
70.8
a
—
—
a
EDAꢂ(BuO)
2
PO
2
H
ꢀ70.4
ꢀ71.1
ꢀ85.7
ꢀ82.4
—
ꢀ64.6
ꢀ91.0
—
ꢀ18.3
—
—
EDAꢂAcOH
EDAꢂHO
2
CH
EDAꢂMsOH
EDAꢂOSA
EDAꢂSacc
54.9
a
—
—
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
3
3
3
3
3
3
2
2
2
2
2
2
Nꢂ(BuO)
2
PO
2
H
a
NꢂAcOH
a
NꢂHO
2
CH
—
ꢀ78.9
—
ꢀ31.6
NꢂMsOH
NꢂOSA
NꢂSacc
2 2
PO H
24.3
a
—
—
ꢀ27.4
ꢀ89.9
ꢀ62.1
—
—
—
ꢀ22.2
ꢀ93.9
—
72.0
ꢀ6.8
47.5
35.8
a45.1
36, 59.4
ꢀ18.4
ꢀ14.5, 21.7
—
—
—
66.4
ꢀ43.7
—
—
—
NHꢂ(BuO)
NHꢂAcOH
NHꢂHO
2
CH
NHꢂMsOH
NHꢂOSA
NHꢂSacc
2 2
PO H
77.9
ꢀ3.7
30.7
10.6
a122.1
Pyrrꢂ(BuO)
PyrrꢂAcOH
PyrrꢂHO
2
CH
—
ꢀ15.1
PyrrꢂMsOH
PyrrꢂOSA
PyrrꢂSacc
—
ꢀ62.1
—
ꢀ42.2
33.4
a
Inhomogeneous.
Table 2 Water content from Karl–Fischer titration produced by two
different PIL mixing methods
The sample with a base composition of 75% required only
28 pulses to obtain satisfactory signal-to-noise (S/N) ratio.
1
Water content (ꢅ5%)/ppm
However, the 20% base sample required greater than 2000
pulses (B 2 h) and had a comparatively poor S/N ratio
a
PIL
Vacuum
Neat addition
[
Fig. 1(a)]. While this was partially due to a change in
(
(
(
(
Bu
Et
Et
HOEt)
HOEt)
HOEt)
HOEt)
2
2
2
2
NHꢂOSA
9408
2033
5371
1200
5539
12102
1205
227
67
119
102
94
15
concentration of N active nuclei in the sample, the altered
dynamics of the ionised nitrogen also contributed. The
increased linewidth near equimolar concentrations also
suggested that the molecular relaxation times had altered
significantly, possibly due to broadening from increased
proton-exchange rate or interaction with a strong dipole from
a close-contact ion, such as would be found in a short-lived
aggregation of ionised species.
NHꢂAcOH
NHꢂHO
2
CH
NHꢂMsOH
CH
CH
NꢂMsOH
2
NHꢂHO
2
3
NꢂHO
2
124
106
3
a
14
Data from Zhao et al.
was able to be determined by examination of a range of
concentrations between K and K [Fig. 1(b)].
In the case of (HOEt) NHꢂAcOH, K had a chemical shift
2
N
0
N
of B10 ppm. A shift of similar magnitude was also present at
1
574 | Phys. Chem. Chem. Phys., 2010, 12, 1571–1577
This journal is ꢃc the Owner Societies 2010

View Article Online
15
Fig. 1
N NMR spectra of diethanolamine samples with different acetic acid concentrations.
the neutralisation point [Fig. 1(b)], indicating the nitrogen
species in equimolar solution were mostly in the ionised state.
Concomitantly, at stoichiometric concentrations all of the acid
The density of the PILs was relatively constant and close to
unity with only a few exceptions. Bases with an excess
of hydrogen bond (H-bond) donors (diethanolamine
1
3
species had formed anions, which was confirmed by C NMR
2
((HOEt) NH), ethanediamine (EDA)) were more dense than
(
see the ESIw), albeit with a lesser change as the carbon nuclei
the N-substituted alkyl bases. The two sulfur-containing acids
(methanesulfonic acid (MsOH), sulfamic acid (OSA)) also
gave an increased density over the carboxylic acids (AcOH,
were not directly involved in the ion exchange. Comparison of
both strong and weak acids showed that proton transfer was
complete for stoichiometric PILs.
HO CH). Thus, predictably the least dense compounds had a
2
combination of N-substituted alkyl amines and formic acid
3
.3 Transport properties
(
i.e., Bu NHꢂHO CH), whereas PILs with the highest density
2
2
As ILs are proposed to be composed entirely of ions, they are
considered to be among the most concentrated electrolytic
fluids with large numbers of charge carriers per unit volume.
In particular, ion transport properties can be crucial in
electrochemical devices and for reaction kinetics in a synthetic
had multiple possible H-bond exchange sites and a sulfur-
containing acid ((HOEt)
2
NHꢂOSA).
Viscosity was correlated with density results—PILs with
excess H-bond donors had higher viscosity than PILs that
only had exchangeable hydrogens at the presumed ionic
2
2
process. Very high conductivities are possible when charge
carriers are mobile and, hence, conductivity provides important
insight into the molecular transport properties of PILs.
Therefore, the density, viscosity and ionic conductivity of
the 18 room-temperature liquid mixtures were examined to
characterise the PIL transport properties (Table 3).
2 2
exchange site. Dibutylphosphate ((BuO) PO H) was the acid
that produced the largest number of room-temperature liquid
PILs, increasing in viscosity approximately 30-fold as the
number of protons attached to base heteroatoms increased
from zero (Et
(EDA, (HOEt)
N, Bu
N) or one (Pyrr, Et NH) to multiple
3 2
3
NH). Also, HO
CH gave a viscosity 3–8 times
2
2
Table 3 Physical properties of room-temperature PILs
ꢀ1
ꢀ1
Protic ionic liquid
Density/g ml
Viscosity/cP
Conductivity/mS cm
(
(
(
(
(
HOEt)
HOEt)
HOEt)
HOEt)
HOEt)
2
2
2
2
2
NHꢂ(BuO)
2
2
PO H
1.110
1.430
1.181
1.220
1.344
0.920
0.941
1.068
1.103
1.107
1.009
1.007
0.992
1.135
1.056
1.067
1.050
1.101
2409
449
5647
762
1590
113
105
2945
958
112
0.06
0.14
0.11
0.93
0.32
1.05
0.07
0.08
0.59
9.83
0.08
0.23
8.85
1.91
0.28
3.02
20.05
0.91
NHꢂOSA
NHꢂAcOH
NHꢂHO
NHꢂMsOH
CH
PO
PO
2
CH
Bu
Bu
2
NHꢂHO
2
3
Nꢂ(BuO)
2
2
H
EDAꢂ(BuO)
EDAꢂAcOH
CH
NHꢂ(BuO)
Nꢂ(BuO) PO
NꢂHO CH
NꢂMsOH
PO
2
2
H
EDAꢂHO
2
Et
Et
Et
Et
2
3
3
3
2
2
PO H
201
2
2
H
94.4
17.7
100
2
Pyrrꢂ(BuO)
PyrrꢂAcOH
PyrrꢂHO CH
PyrrꢂSacc
2
2
H
105
36.3
15.4
499
2
This journal is ꢃc the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 1571–1577 | 1575

View Article Online
lower than that of AcOH with the same base; PILs with
MsOH were also close to an order of magnitude more viscous.
Most surprising was the high viscosity of (HOEt) NHꢂAcOH,
2
which may be explained by the formation of acetic acid dimers
and trimers that could lead to large species more resistant to
2
3
flow.
The high viscosity had a major impact on the measured
conductivity of the PILs. This was foremost with the relative
fluid formic acid mixtures, which had conductivity values an
order of magnitude higher than any other ion combination.
Although in general there is a correlation between viscosity
and conductivity, the viscosity alone cannot account for the
conductivity behaviour. For instance, pyrrolidine (Pyrr) PILs,
PyrrꢂSacc and PyrrꢂAcOH had similar conductivity values but
the measured viscosity differed by almost two orders of
magnitude. Furthermore, EDAꢂHO CH and Bu NHꢂHO CH
Fig. 2 Walden plot of PILs fluid at room temperature. PILs are
numerically denoted as: (1) (HOEt)
2
NHꢂAcOH, (2) EDAꢂ(BuO)
H, (4) (HOEt)
NHꢂMsOH, (5) EDAꢂAcOH,
CH, (7) PyrrꢂSacc, (8) EDAꢂHO CH, (9)
NHꢂOSA, (10) Et NHꢂ(BuO) PO H, (11) Bu NHꢂHO CH,
12) Pyrrꢂ(BuO) PO H, (13) Et Nꢂ(BuO) PO H, (14) Bu Nꢂ(BuO) PO H,
(15) PyrrꢂAcOH, (16) Et NꢂMsOH, (17) Et NꢂHO CH, and (18)
2 2
PO H,
2
2
2
(
(
(
3) (HOEt)
6) (HOEt)
HOEt)
2
NHꢂ(BuO)
2
PO
2
2
had a similar viscosity but a large difference in conductivity.
The diverse results suggest the possibility that multiple
modes of proton transport may be a significant process in
PIL conductance, i.e., Grotthus proton-hopping and ion
2
NHꢂHO
2
2
2
2
2
2
2
2
(
2
2
3
2
2
3
2
2
3
3
2
2
4
transport. Many other factors, which can contribute to the
conductivity (i.e., ion size, density and ion aggregation),
however, were not examined.
2
PyrrꢂHO CH.
EDAꢂ(BuO)
PO H, all roughly the same distance from the
2
2
One way to understand the correlation between ionic
conductivity and viscosity is to use the qualitative approach
ideal line. This suggests ion transport mechanisms and inter-
molecular interactions were constant for a given base, with the
1
3,25
of Angell et al. based on the classical Walden rule:
acid size determining transport properties. (HOEt)
follows a similar pattern (ZHO PO
CH o MsOH o (BuO)
with notable exceptions. (HOEt)
NHꢂAcOH was the most
viscous PIL examined, possibly due to AcOH forming larger
2
NH
2
2
2
H)
LZ = k
(2)
2
2
3
where L is the molar conductivity, Z is the viscosity, and k is a
temperature-dependent constant. Dilute aqueous solutions,
molten salts and ionic liquids have been observed to approxi-
mately follow the Walden rule with conductivity inversely
multi-ion aggregates; (HOEt)
2
NHꢂOSA was similar to PILs
without exchangeable hydrogens, possibly as OSA can form a
zwitterion depending on the proton solvation and may not be
involved in ionic conduction. There was an overall implication
that a proton-hopping mechanism increased molar conductivity
when a base had more than one proton capable of inter-
2
6
related to viscosity. A log–log plot of L vs. Z predicts a
straight line that passes through the origin; 0.01 M KCl, which
exists as dissociated ions, provides a reference ‘ideal’ line for
the plot. Compounds that lie close to the ideal line are said to
be ionic, except for one or two notable exceptions, most ILs
2
9
molecular exchange.
It has been suggested that the relative difference in aqueous
pK between the acid and base of PILs can provide some
measure of proton transfer. A large DpK (aq) PIL, i.e., Et
Nꢂ
MsOH, would be predicted to be more ‘ionic’ than the small
a
2
7
fall below, suggesting neutral species in the liquid. Improved
fits to the Walden plot are possible by compensating for the
size of the conductive species although the Walden plot
a
3
2
6
DpK
However, it must be remembered that aqueous DpK
values are not necessarily the same in non-aqueous solutions.
PILs have also been observed to alter the pK of acids
depending on the paired base. For this work, the DpK (aq)
a
(aq) (HOEt)
2
NHꢂAcOH—the opposite of observations.
2
8
remains a qualitative analysis.
a
(aq)
3
0
Of the 18 binary liquid mixtures examined via Walden plots,
two main types of behaviour were observed (Fig. 2). Firstly,
eight PILs lie close to the ideal line; secondly, the majority of
the room-temperature PILs were significantly below the ideal
line, suggesting between 10–20% ionic species by the Walden
definition. PILs close to the ideal line typically were relatively
high viscosity—conversely, highly fluid and conductive
PILs were typically much below the ideal line. The only
compound slightly above the ideal line, PyrrꢂSacc, had a large
freezing/melting point hysteresis and remained liquid at room
a
3
1
a
approach was too simple to account for all intermolecular
phenomena related to ‘good’ or ‘poor’ PILs, particularly with
regard to hydrogen bonding or a mobile proton.
Regardless, the majority of PILs appeared to be poorly
ionised as predicted by conductivity and were clustered
approximately an order of magnitude below the ideal line
1
5
(Fig. 2). However, N NMR indicated that proton transfer
was complete and the molecules in PILs were ionic. Therefore,
a significant amount of the ionic molecules must have ordered
temperature below its T
property may influence the mechanism of conduction.
PILs close to the ideal line were (HOEt) NH- or EDA-type,
m
, suggesting some thermal phase
1
0,32
in such a way as to be electrically neutral, i.e., an ion-pair,
or the viscosity was greater than predicted for a Newtonian
2
bases with exchangeable protons away from the proposed
3
3
ion-forming proton transfer position (i.e.,–NH , –OH). In a
2
electrolyte fluid, i.e., a liquid crystalline order. The large
number of PILs in about the same region on the Walden
plot suggests that all PILs may be mostly ion-paired or
given series, EDAꢂHO CH was an order of magnitude
2
less viscous than EDAꢂAcOH and B2 orders less than
1
576 | Phys. Chem. Chem. Phys., 2010, 12, 1571–1577
This journal is ꢃc the Owner Societies 2010

View Article Online
liquid-phase ordered—a Grotthuss proton-hopping mechanism
then promotes conductivity if appropriate functional groups
were included.
9 D. R. MacFarlane and K. R. Seddon, Aust. J. Chem., 2007, 60,
3
1
–5.
0 (a) K. J. Fraser, E. I. Izgorodina, M. Forsyth, J. L. Scott and
D. R. MacFarlane, Chem. Commun., 2007, 3817–3819;
(
6
b) H. Weinga
54–670.
1 (a) A. Noda, K. Hayamizu and M. Watanabe, J. Phys. Chem. B,
001, 105, 4603–4610; (b) H. A. Every, A. G. Bishop,
D. R. MacFarlane, G. Oradd and M. Forsyth, J. Mater. Chem.,
001, 11, 3031–3036.
¨
rtner, Angew. Chem., Int. Ed., 2008, 47,
4
. Conclusion
1
2
A series of protic ionic liquids has been prepared and their
physical properties characterised. PILs produced by simultaneous
addition of pre-dried reagents were at least an order of
magnitude drier than comparative methods and component
stoichiometry was known to better than 0.1 mol%. The
physical characteristics of the PILs were strongly influenced
by the functional groups of the base: those with protons
capable of undergoing exchange led to higher viscosity PILs
than similar bases without exchangeable hydrogens (protons).
2
12 B. Nuthakki, T. L. Greaves, I. Krodkiewska, A. Weerawardena,
M. I. Burgar, R. J. Mulder and C. J. Drummond, Aust. J. Chem.,
2007, 60, 21–28.
3 W. Xu and C. A. Angell, Science, 2003, 302, 422–425.
1
14 C. Zhao, G. Burrell, A. A. J. Torriero, F. Separovic, N. F. Dunlop,
D. R. MacFarlane and A. M. Bond, J. Phys. Chem. B, 2008, 112,
6
5 T. L. Greaves, A. Weerawardena, C. Fong, I. Krodkiewska and
923–6936.
1
C. J. Drummond, J. Phys. Chem. B, 2006, 110, 22479–22487.
16 (a) A. Noda, M. A. B. H. Susan, K. Kudo, S. Mitsushima,
K. Hayamizu and M. Watanabe, J. Phys. Chem. B, 2003, 107,
1
5
N NMR of neat acid–amine PILs provided a straight-
forward method to show that proton transfer from acid to
amine had occurred in individual molecules of a PIL.
However, for many PILs, conductivity did not match theoretical
values based on viscosity and indicated that significant neutral
species (ion aggregates) or liquid-phase order were present in
PILs. Possible reasons for neutral species formed from ionised
molecules may include neutral ion pairs or increased viscosity
from hydrogen-bond networks.
4
024–4033; (b) C. Wang, L. Guo, H. Li, Y. Wang, J. Weng and
L. Wu, Green Chem., 2006, 8, 603–607; (c) I. Cota, R. Gonzalez-
Olmos, M. Iglesias and F. Medina, J. Phys. Chem. B, 2007, 111,
12468–12477; (d) M. Anouti, M. Caillon-Caravanier, Y. Dridi,
H. Galiano and D. Lemordant, J. Phys. Chem. B, 2008, 112,
13335–13343.
7 V. Amendola, M. Boiocchi, L. Fabbrizzi and A. Palchetti,
Chem.–Eur. J., 2005, 11, 120–127.
8 J. Akitt, NMR and Chemistry: An Introduction to Modern NMR
Spectroscopy, Kluwer Academic Publishers, 3rd edn, 1992.
9 P. S. Pregosin, Prog. Nucl. Magn. Reson. Spectrosc., 2006, 49,
1
1
1
2
2
2
2
61–288.
Acknowledgements
0 G. C. Levy and R. L. Lichter, Nitrogen-15 Nuclear Magnetic
Resonance Spectroscopy, Wiley, New York, 1979, p. 221.
1 C. Iojoiu, P. Judeinstein and J. Sanchez, Electrochim. Acta, 2007,
This study was financed by the Australian Research Council
Linkage Grant LP0668123 and Orica Ltd, Australia. GLB is
the recipient of an APA(I) scholarship. The authors thank
Douglas R. MacFarlane, David Dunstan and Alan Bond for
use of equipment and helpful discussions.
5
3, 1395–1403.
2 (a) J. N. Barisci, G. G. Wallace, D. R. MacFarlane and
R. H. Baughman, Electrochem. Commun., 2004, 6, 22–27;
(
1
b) A. K. Chakraborti and S. R. Roy, J. Am. Chem. Soc., 2009,
31, 6902–6903.
2
2
3 K. M. Johansson, E. I. Izgorodina, M. Forsyth, D. R. MacFarlane
and K. R. Seddon, Phys. Chem. Chem. Phys., 2008, 10, 2972–2978.
4 T. Ueki and M. Watanabe, Macromolecules, 2008, 41, 3739–3749.
References
1
(a) P. Wasserscheid and T. Welton, Ionic Liquids in Synthesis,
Wiley-VCH, 2007, p. 776; (b) Z. Fei, D. Zhao, T. J. Geldbach,
R. Scopelliti and P. J. Dyson, Chem.–Eur. J., 2004, 10, 4886–4893;
25 M. Yoshizawa, W. Xu and C. A. Angell, J. Am. Chem. Soc., 2003,
125, 15411–15419.
26 C. A. Angell, N. Byrne and J. Belieres, Acc. Chem. Res., 2007, 40,
1228–1236.
(c) M. J. Earle and K. R. Seddon, Pure Appl. Chem., 2000, 72,
1391–1398; (d) H. Ohno, Electrochemical Aspects of Ionic Liquids,
John Wiley & Sons, Hoboken, NJ, 2005.
N. V. Plechkova and K. R. Seddon, Chem. Soc. Rev., 2008, 37,
123–150.
27 J. Belieres and C. Angell, J. Phys. Chem. B, 2007, 111, 4926–4937.
28 D. R. MacFarlane, M. Forsyth, E. I. Izgorodina, A. P. Abbott,
G. Annat and K. Fraser, Phys. Chem. Chem. Phys., 2009, 11,
4962–4967.
2
3
T. L. Greaves, A. Weerawardena, I. Krodkiewska and
C. J. Drummond, J. Phys. Chem. B, 2008, 112, 896–905.
P. Walden, Bull. Acad. Imp. Sci. (St. Petersburg), 1914, 1800.
(a) A. Stark, P. Behrend, O. Braun, A. Muller, J. Ranke,
¨
B. Ondruschka and B. Jastorff, Green Chem., 2008, 10,
29 M. Meuwly, A. Bach and S. Leutwyler, J. Am. Chem. Soc., 2001,
123, 11446–11453.
4
5
30 (a) I. Kaljurand, A. Ku
I. Leito and I. A. Koppel, J. Org. Chem., 2005, 70, 1019–1028;
(b) A. Kutt, I. Leito, I. Kaljurand, L. Soovali, V. M. Vlasov,
¨
¨ ¨
tt, L. Soovali, T. Rodima, V. Maemets,
¨
¨
1
2
152–1161; (b) T. L. Greaves and C. J. Drummond, Chem. Rev.,
008, 108, 206–237.
L. M. Yagupolskii and I. A. Koppel, J. Org. Chem., 2006, 71,
2829–2838.
6
7
8
F. D’Anna, S. L. Marca and R. Noto, J. Org. Chem., 2008, 73,
397–3403.
W. Wang, L. Shao, W. Cheng, J. Yang and M. He, Catal.
Commun., 2008, 9, 337–341.
A. Martinelli, A. Matic, P. Jacobsson, L. Borjesson, A. Fernicola,
S. Panero, B. Scrosati and H. Ohno, J. Phys. Chem. B, 2007, 111,
31 M. Anouti, M. Caillon-Caravanier, C. L. Floch and
D. Lemordant, J. Phys. Chem. B, 2008, 112, 9406–9411.
32 L. Zhang, H. Li, Y. Wang and X. Hu, J. Phys. Chem. B, 2007, 111,
11016–11020.
33 (a) T. Ichikawa, M. Yoshio, A. Hamasaki, T. Mukai, H. Ohno and
T. Kato, J. Am. Chem. Soc., 2007, 129, 10662–10663; (b) R. Atkin
and G. G. Warr, J. Phys. Chem. B, 2008, 112, 4164–4166.
3
12462–12467.
This journal is ꢃc the Owner Societies 2010
Phys. Chem. Chem. Phys., 2010, 12, 1571–1577 | 1577