Controlling hydrolysis reaction rates with binary ionic liquid mixtures by tuning hydrogen-bonding interactions

Article abstract of DOI:10.1021/jp211543v

The ability of a binary ionic liquid (IL) system consisting of a phosphonium transition state analogue (TSA) and 1-butyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide ([BMIM][NTf₂]) to accelerate the rate of the well-studied hydrolysis of a tert-alkyl picolinium salt by influencing the solvent structure was investigated. A significant rate enhancement was observed in the presence of the TSA; however, comparison with other cations illustrated that this enhancement was not unique to the chosen TSA. Instead, the rate enhancements were correlated with the dilution of hydrogen bonding by the added cations. This phenomenon was further examined by the use of 1-butyl-2,3-dimethylimidazolium bis(trifluoromethanesulfonyl)imide ([BMMIM][NTf₂]) as a cosolvent and the use of Reichardt's dye to measure the extent of hydrogen bonding on solutes in these systems. The rate increases are rationalized in terms of weaker hydrogen bonding from the solvent system to water. ? 2012 American Chemical Society.

Full text of DOI:10.1021/jp211543v

Article
pubs.acs.org/JPCB
Controlling Hydrolysis Reaction Rates with Binary Ionic Liquid
Mixtures by Tuning Hydrogen-Bonding Interactions
Cameron C. Weber, Anthony F. Masters, and Thomas Maschmeyer*
Laboratory of Advanced Catalysis for Sustainability, School of Chemistry, University of Sydney, Sydney 2006, Australia
S
* Supporting Information
ABSTRACT: The ability of a binary ionic liquid (IL) system consisting of a phos-
phonium transition state analogue (TSA) and 1-butyl-3-methylimidazolium bis(trifluoro-
methanesulfonyl)imide ([BMIM][NTf2]) to accelerate the rate of the well-studied hydrolysis
of a tert-alkyl picolinium salt by influencing the solvent structure was investigated. A
significant rate enhancement was observed in the presence of the TSA; however,
comparison with other cations illustrated that this enhancement was not unique to the
chosen TSA. Instead, the rate enhancements were correlated with the dilution of hydrogen bonding by the added cations. This
phenomenon was further examined by the use of 1-butyl-2,3-dimethylimidazolium bis(trifluoromethanesulfonyl)imide
([BMMIM][NTf2]) as a cosolvent and the use of Reichardt’s dye to measure the extent of hydrogen bonding on solutes in
these systems. The rate increases are rationalized in terms of weaker hydrogen bonding from the solvent system to water.
employed in the synthesis of structured materials for separa-
tions and catalysis,24 the generation of catalytic antibodies,25
and to probe and/or inhibit enzyme active sites.26 To the best
of our knowledge, no investigation has been conducted into
whether these structures are able to affect a reaction in an
ordered solvent, such as an IL.
Here, we investigate the effect of using a binary IL system
involving a TSA and [BMIM][NTf2], to determine whether
any evidence of structural reorganization to facilitate an increase in
reaction rate can be observed. The idea is that the reported
mesoscopic ordering of ILs would be affected by the TSA
resulting in transient solvent reorganization. This would most
likely require a charged TSA, as the predominant mechanism
for ordering in ILs is based on electrostatics. Our hypothesis is
that the use of a charged TSA would preorganize proximate
solvent molecules toward solvation of the transition state,
yielding a reaction rate enhancement.
In choosing a model reaction to test this hypothesis, several
criteria must be satisfied. The reaction must be well studied and
understood. It should also yield a simple product distribution
and be amenable to study in IL (i.e., desirable criteria include
solubility of reagents and products, and convenient experi-
mental and analytical protocols resulting in high quality data)
and involve relatively accessible materials. The hydrolysis of
trityl chloride in the presence of pyridine in nitromethane was
originally studied by Ingold and co-workers and determined to
proceed via the formation of the pyridinium cation ([Ph3CPyr]
[Cl]).27 The rate limiting step was proposed to be the
unimolecular dissociation of the pyridine, aided by a cosolvent
effect from the added substituting agent.27 A mechanism
change was observed in benzene, where pyridine was not found
INTRODUCTION
■
Ionic liquids (ILs) have been defined as salts that melt below
100 °C.1,2 They have been extensively investigated as replace-
ments for volatile organic compounds (VOCs) due to their
negligible vapor pressure.3,4 While some issues have been raised
subsequently regarding the green credentials of ILs, particularly
in terms of their syntheses and potential toxicity, they have
continued to attract interest due to the flexibility of physical
and chemical properties available through the judicious
variation of the cation and anion.5−7 Given that the potential
number of ionic liquids has been estimated at >1014, which
does not account for the possibility of forming binary mixtures
of varying composition, it would be valuable to possess both a
thorough understanding of ionic liquid structures and their
effect on reactivity, in addition to developing methods of rationally
manipulating those structures toward desirable outcomes.8
Attention has been drawn to the supramolecular nature of IL
solvents, particularly those based on 1,3-dialkylimidazolium
cations, which results from their vast array of intermolecular
interactions and the extent of their structural organization.2,9,10
Experimental and theoretical studies have demonstrated that
for 1-alkyl-3-methylimidazolium ILs, polar and nonpolar
nanostructured domains form in ILs possessing an alkyl chain
containing at least 4 carbons.11−14 The intermolecular forces,
which give rise to these domains, such as hydrogen bonding,
π−π stacking, and ion pairing, have also been found to lead to
inclusion compounds.10,15−18 The effects of this supramolecular
behavior can be found in changes to reaction rates and out-
comes for a range of processes including Diels−Alder cyclo-
additions,19 hydroformylation,20 the Baylis−Hillman reaction,21
and the Tsuji−Trost reaction.22 Despite these findings, the
successful use of inclusion compounds to influence reactions
has been largely serendipitous.
Received: December 1, 2011
Revised: January 15, 2012
Published: January 20, 2012
Transition state analogues (TSAs) are stable mimics of high
energy transition state structures.23 TSAs have been successfully
© 2012 American Chemical Society
1858
dx.doi.org/10.1021/jp211543v | J. Phys. Chem. B 2012, 116, 1858−1864

The Journal of Physical Chemistry B
Article
to complex the trityl chloride, most likely due to benzene’s low
polarity and hence an inability to solvate ionic species.28 How-
ever, the apparent polarity of [BMIM][NTf2], as measured by
solvatochromic dyes, lies between those of protic and polar
Young’s valve. A 1 mL solution of either [BMIM][NTf2]
(0.1 M) or 1:1 [BMIM][NTf2] and the TSA, [PPh4][NTf2], or
[PBu4][NTf2] (0.1 M in each electrolyte) was independently
prepared in deuterated chloroform containing 0.03% (v/v)
TMS. The IL solution was added in known aliquots using a
aprotic solvents based on single parameter polarity scales such
1
as Kosower’s Z-value29 or Reichardt’s ET
.
30 Thus, it would be
N
microliter syringe to the NMR tube and a H NMR spectrum
recorded. The chemical shifts of peaks assigned to [BMIM]-
[NTf2] were then recorded and fitted to a dimerization
isotherm (see Supporting Information for fitting details).
Solvatochromic Dye Measurements. Solvatochromic
dye measurements were obtained as follows. A 250 μL aliquot
of an ethanolic stock solution of Reichardt’s dye (8.1 × 10−4 M)
was added to a 2 mL vial, and the bulk solvent was removed
under nitrogen flow. Residual solvent was removed by dry-
ing the vial contents in vacuo. The IL solution to be examined
(500 μL), previously dried in vacuo at 80 °C for several hours,
was added to the vial and the dye dissolved. The solution was
transferred to a minimum volume cuvette (1 cm path length)
and a UV−vis spectrum obtained at 25.0 °C, with the sample
maintained at this temperature by means of a Peltier cooling
anticipated that the pyridinium mechanism would operate
within this IL. The hydrolysis of a 4-picoline complexed trityl
chloride derivative was selected as a model reaction as this
enabled the monitoring of the reaction on a more convenient
time scale near room temperature relative to the corresponding
pyridine derivative.
EXPERIMENTAL SECTION
■
Materials. [BMMIM][NTf2], [BMIM][NTf2], p-fluorophenyl-
diphenylmethyl chloride (FDMC), and (p-fluorophenyl)-
diphenyl(p-tolyl)phosphonium iodide were prepared according
to literature procedures or modifications thereof31−33 and
characterized by NMR, ESI-MS, and melting point where appro-
priate. Tetramethylammonium bis(trifluoromethanesulfonyl)-
imide ([NMe4][NTf2]), tetrabutylphosphonium bis(trifluoro-
methanesulfonyl)imide ([PBu4][NTf2]), tetraphenylphosphonium
bis(trifluoromethanesulfonyl)imide ([PPh4][NTf2]), and (p-fluoro-
phenyl)diphenyl(p-tolyl)phosphonium bis(trifluoro-
methanesulfonyl)imide (the TSA) were all prepared by preci-
pitation from aqueous solution using LiNTf2 and the relevant
precursors and were characterized by NMR, ESI-MS, melting
point, and elemental analysis (see Supporting Information for
all experimental details and characterization). 4-Picoline
(Fluka) was purified by passage through a column of neutral
activated alumina and stored under nitrogen on activated 4 Å
molecular sieves. Nitromethane, benzene, and chloroform were
all distilled from drying agents and stored under nitrogen over
4 Å molecular sieves. IL samples to be used for kinetics experi-
ments were dried at 80 °C in vacuo for a minimum of 24 h
before the addition of the required amount of water using a
microliter syringe. The drying procedure afforded ILs with
water contents consistently below 200 ppm, as ascertained by
Karl Fischer titration (c.f., ∼7000 ppm for 0.55 M solution),
while the microliter syringe was found to be accurate and
reproducible to within 1% through the use of gravimetric
methods. The IL/water solution was stirred for at least 4 h
before the experiment to ensure thorough mixing.
N
system. ET values were calculated from the wavelength of the
lowest energy absorbance maximum (see Supporting Informa-
tion for calculation details).
RESULTS
■
To enable facile reaction monitoring, N-(p-fluorophenyldiphenyl-
methyl)-4-picolinium chloride ([Ar3CPic][Cl]), generated in
situ from the reaction of 4-picoline with p-fluorophenyldiphenyl-
methyl chloride (FDMC), was employed to allow the use of
19F NMR to follow the reaction, which is illustrated in Scheme 1.
Scheme 1. Hydrolysis of [Ar3CPic][Cl] in [BMIM][NTf2],
the Model Reaction Employed in This Investigation
It was first necessary to confirm that 4-picoline dissociation was
indeed rate-limiting, and the results of these experiments are
summarized in Table 1. Excellent fits to first order rate laws
Hydrolysis Experiments. Kinetics experiments were
conducted using an NMR tube equipped with a Young’s
valve. The tube was charged with FDMC (15−25 mg; 0.05−
0.08 mmol) under a flow of nitrogen. 4-Picoline (50 or 100 μL)
was added to dissolve the substrate. Approximately 5 min
before the reaction was to be monitored, the solution was
diluted with ionic liquid containing H2O (0.98 mL of 0.55 M
H2O) and cooled on ice. The sample was mixed using a Vortex
mixer and a reference capillary (1-bromo-4-fluorobenzene in
acetone-d6 (0.50 M)) inserted coaxially before monitoring
the reaction using a Bruker Avance 300 NMR spectrometer
(300 MHz), maintained at 294.3 K, with 19F NMR spectra
(total acquisition time of 60 s) obtained every 65 s.
Experiments were performed in triplicate. Experimental details
for the reactions conducted in nitromethane are outlined in the
Supporting Information.
Table 1. Variation of the Observed First Order Rate
Constant with Water and 4-Picoline Concentrations for the
a
Hydrolysis of [Ar3CPic][Cl] in [BMIM][NTf2] at 294.3 K
water
concentration (M)
4-picoline
concentration (M)
rate constant
−1
(× 10−4 s )
0.51
0.51
1.02
0.49
0.94
0.49
10.6 0.2
10.5 0.1
19.3 0.5
a
Reported errors are standard deviations from replicate experiments.
were observed in all cases and no dependence on 4-picoline
concentration was found, while a positive dependence on the
amount of water existed. These results are consistent with those
of Ingold et al. and form the basis of the proposed reaction
transition state and TSA, depicted in Figure 1.
Dimerization Experiments. The dimerization experiments
were conducted using the following procedure adapted from
the literature.34 Deuterated chloroform containing 0.03% (v/v)
TMS (500 μL) was added to an NMR tube equipped with a
1859
dx.doi.org/10.1021/jp211543v | J. Phys. Chem. B 2012, 116, 1858−1864

The Journal of Physical Chemistry B
Article
The values of Δδ2H imply that the structure of the aggregate
formed is different when the template molecule is added.
Unfortunately, NOESY contacts could not be observed between
cations, so no direct evidence for heterodimers was obtained. The
apparent preferential association with aromatic cations is an
interesting result. The interaction of imidazolium cations with
neutral aromatic substrates has been well-documented10,16,18,35 as
has the formation of ionic cocrystals involving cation substitu-
tion.15 However, to the best of our knowledge, no preferential
association for aromatic cations with imidazolium ILs has ever
been shown.
Figure 1. Structure of (a) the proposed reaction transition state and
(b) the TSA.
Theoretical studies regarding the solubility of aromatic
compounds in ILs have suggested that the predominant
interactions between the IL and an aromatic solute are
electrostatic, arising from the strong quadrupole moment of
aromatic compounds.36 It was also discovered that for benzene,
the interaction with the anion dominated, while the addition of
electronegative substituents, for example, in hexafluorobenzene,
resulted in stronger interactions with the cation.37 Therefore,
the presence of a delocalized positive charge across the aro-
matic group, as in the TSA and [PPh4][NTf2], would probably
lead to an anion-dominated interaction with the aromatic
solute. The stronger anion/solute rather than cation/solute
interaction would account for the lack of any observable NOE
interactions between different cations as the upper limit for
detecting NOEs is approximately 4 Å,38 considerably less than a
cation−cation distance when the NTf2 anion is present. The
preferential formation of dimers, therefore, indicates a favorable
association between the IL and the added template molecules.
This would be a prerequisite if any structural reorganization
around the TSA was to occur.
To ascertain whether any potential rate increase was specific
to the TSA, alternative templates were also examined. The
templates chosen were [PPh4][NTf2] to determine the effect of
aryl substitution, [PBu4][NTf2] to examine the necessity of a
charge delocalized aromatic structure, and [NMe4][NTf2] to
probe the importance of the cation’s size. The molecular struc-
tures of all templates and ILs used in this study are depicted in
Figure 2.
In order to determine whether any rate increase could be
observed in the presence of the TSA, the NTf2 salt of the TSA
was dissolved in [BMIM][NTf2] in concentrations ranging
from 0.5 mol % to 15 mol % of the IL solution. As depicted
in Figure 3, the rate constant observed increases with TSA
Figure 2. Structure and acronyms of IL ions, the substrate, and
templates used in this investigation.
Prior to conducting the hydrolysis experiments, the potential
of each of these materials to induce solvent reorganization
1
was probed through a series of H NMR dilution experiments.
These experiments were conducted in CDCl3 utilizing 1:1
mixtures of the relevant template and [BMIM][NTf2]. Data
were fitted to a dimerization isotherm, which yielded both the
association constant (Ka) and limiting chemical shift values for
the ion-pair and dimer (δip, δd) (see Supporting Information for
model details). Values of Ka listed are weighted averages across
a minimum of 7 proton environments. As Table 2 illustrates, all
1
Table 2. H NMR Titration Data in CDCl3 at 300 K
a
template
Ka
ΔGa (kJ mol−1)
Δδ2H (ppm)
b
N/A
45
72
9
5
−9.5 0.5
−10.7 0.2
−12.2 0.1
−12.4 0.2
0.49
0.39
0.36
0.37
[PBu4][NTf2]
[PPh4][NTf2]
TSA
131
143
8
9
a
Limiting chemical shift variation for the C2 proton on dimerization.
Modeled as 1:1 [BMIM][NTf2] to ensure a comparable ionic
b
Figure 3. Variation of the observed rate constant with TSA loading.
Reported errors are the standard deviations of replicate experiments.
strength with the templated solutions.
templates investigated resulted in a more favorable ΔGa when
compared to that of [BMIM][NTf2] self-dimerization. Unfortu-
nately, [NMe4][NTf2] was unable to be examined due to its
poor solubility in CDCl3.
concentration; however, to determine the specificity of this rate
increase, the other templates (i.e., the NTf2 salts of the PPh4,
PBu4, and NMe4 cations) were also examined at 1 and 10 mol %
concentrations. The results are shown in Figure 4.
1860
dx.doi.org/10.1021/jp211543v | J. Phys. Chem. B 2012, 116, 1858−1864

The Journal of Physical Chemistry B
Article
geometric rearrangement of the solvent but that they dilute the
more hydrogen bond acidic [BMIM] cation. As the hydrogen
bond acidity of the medium is reduced, it would be expected
that water would interact less strongly with the solvent and
hence be better able to solvate the cationic transition state. To
ascertain whether this hypothesis was valid, the hydrolysis was
conducted both in neat [BMMIM][NTf2], to determine the
effect of definitively lowering the hydrogen bond acidity of
the medium, and in 15 mol % [BMMIM][NTf2] in [BMIM]
[NTf2], to see if a low concentration of a more weakly hydrogen-
bonding cation could give rise to the rate enhancements
observed. As the [BMMIM] cations are comparable in size to
[BMIM], this would also remove the geometric considerations.
Any increase in rate can, therefore, be attributed solely to
reducing the hydrogen bond acidity and not to templating
through a molecular size effect. These results are illustrated in
Table 3, with comparisons to the corresponding rates obtained
Table 3. Comparison of the Observed First Order Rate
Constants for Hydrolysis in Different Ionic Liquid Solvent
Systems
Figure 4. Observed first order hydrolysis rate constants using different
templates in [BMIM][NTf2] at 1 (light gray) and 10 (dark gray) mol %
loadings (X = NTf2). The untemplated result is depicted for comparison.
Reported errors are standard deviations of replicate experiments.
a
−1
ionic liquid solution
[BMIM][NTf2]
rate constant (× 10−4 s )
From Figure 4, at 1 mol %, it appears that all templates gave
a slightly faster rate although some values, in particular those
for [NMe4][NTf2], lie only marginally outside the observed
experimental error. At the 10 mol % loading, it was found that
[NMe4][NTf2] yielded a rate consistent with that in the untem-
plated IL, while the other templates resulted in a significantly
faster reaction. Unfortunately, because of its poor solubility,
[PPh4][NTf2] could not be investigated at 10 mol %.
10.6 0.2
13.9 0.7
14.5 0.6
12.6 0.7
18.8 0.4
b
15% TSA
b
15% [PBu4][NTf2]
b
15% [BMMIM][NTf2]
[BMMIM][NTf2]
a
Reported errors are standard deviations of replicate experiments.
Mol % in [BMIM][NTf2].
b
The similar results obtained for both [PBu4][NTf2] and the
TSA, with no rate increase relative to the untemplated solu-
tion observed for the [NMe4] cation, indicates that the rate
enhancement could be a result of the size of the template and
is not a result of the template’s precise electronic structure.
Further experiments were conducted with [PBu4][NTf2] and at
each concentration, the observed rate constant was similar to
with those obtained in the presence of the TSA at the same
concentration (see Supporting Information).
One factor that cannot be ignored when examining the
concept of molecular templating is that of the apparent solvent
polarity, as the addition of the template molecule may vary the
composition of the solvent cage surrounding the substrate.
Solvent polarity has been shown to be of limited importance
when pyridine and its derivatives form the leaving group for
tertiary carbocations in neutral solvents due to the lack of
charge development between the reactants and the transition
state.39 It should be noted, however, that these reactions
utilized the solvent as the nucleophile. Therefore, the solvent−
nucleophile interactions that may be experienced in the present
system would have been inherently unobservable in the
literature study. When nitromethane was used as a solvent for
the nucleophilic substitution of [Ph3CPyr][Cl], it was found
that the rate of substitution increased in the order phenol <
water < ethanol, an effect ascribed to the increasing order of
nucleophilicity.27 In light of the lack of participation of pyridine
in the rate determining step, it can be concluded that it is the
ability of the substituting agent to solvate the forming
carbocation which is important, i.e., its hydrogen bond basicity
rather than its nucleophilicity.
in [BMIM][NTf2], 15 mol % TSA and 15 mol % [PBu4]-
[NTf2].
As Table 3 illustrates, the observed rate constant increases
quite considerably when the reaction is conducted in
[BMMIM][NTf2] rather than [BMIM][NTf2]. Further, when
only 15 mol % [BMMIM][NTf2] is used in [BMIM][NTf2], a
rate enhancement of similar magnitude to that for similar
concentrations of the TSA and of [PBu4][NTf2] is observed.
These observations strongly suggest that the rate enhancements
do not arise from molecular size but from polarity effects.
In order to quantify the relevant change in apparent polarity
resulting from the addition of a nonpolar template, Reichardt’s
dye was used to investigate the ILs and templated solutions,
with the results summarized in Table 4 (see Supporting
Information for calculation details). Reichardt’s dye has been
used extensively to quantify the polarity of solvents and has
been widely used in the IL literature due to its large
solvatochromic range (357 nm spectroscopic shift from water
to diphenyl ether).40 The wavelength of the lowest energy
absorbance maximum has been shown to correlate with the
hydrogen bond acidity, Lewis acidity, and dipolarity of the
solvent, with the hydrogen bond acidity representing the most
important contribution.41
N
As is evident from Table 4, no significant change in the ET
polarity parameter can be observed in the templated solutions
using the solvatochromic probe, except for the 10 mol %
[NMe4][NTf2] solution, which exhibited an increase. The lack
N
of a discernible ET decrease for the other templates could be
due to preferential solvation of Reichardt’s dye by the more
strongly hydrogen-bonding [BMIM] cation, a phenomenon
that has been shown for ILs in binary mixtures with neutral
solvents such as dichloromethane.42 Unfortunately, this could
An alternative hypothesis for the observed reaction rate
increase is, therefore, not that the templates participate in the
1861
dx.doi.org/10.1021/jp211543v | J. Phys. Chem. B 2012, 116, 1858−1864

The Journal of Physical Chemistry B
Article
of the mole fraction of one solvent (x2) and the ETN values for
each neat solvent (ETN(1) and ETN(2), respectively) as shown
in eq 2.43
N
Table 4. Comparison of ET Values Obtained in Templated
a
Solutions with Those of the Neat ILs at 25.0 °C
N
sample
ET (lit.)
N
N
[BMIM][NTf2]
0.647 (0.644)50
0.547 (0.541)50
0.706
Kx (E (2) − E (1))
N
2
T
T
N
E
=
+ E (1)
T
T
[BMMIM][NTf2]
10% [NMe4][NTf2]
1 − x + Kx
(2)
2
2
b
b
10% TSA
0.657
As shown in Figure 5, this model for the solvent exchange
process gave an excellent fit to the experimental data (R2 =
0.996), which indicates that a more complicated model is not
necessary. Table 5 summarizes the results obtained from the fit.
b
10% [PBu4][NTf2]
0.645
b
15% TSA
0.658
b
15% [PBu4][NTf2]
0.658
b
15% [BMMIM][NTf2]
0.639
Table 5. Parameters Obtained from Fitting the ETN Isotherm
to the One-Step Preferential Solvation Model
a
b
Literature values are provided for comparison, where available. Mol
a
% in [BMIM][NTf2].
binary solvent system
K
ΔG (kJ mol−1)
not be confirmed for the TSA nor for [PBu4][NTf2], as both
are solids, and therefore, the concentrations amenable to
examination at 25.0 °C are limited by their solubilities. The
apparent polarity of [BMMIM][NTf2], however, could be
examined in [BMIM][NTf2] in compositions from 0 mol % to
100 mol % to investigate this proposition, and these measure-
ments are depicted in Figure 5.
[BMIM][NTf2]:[BMMIM][NTf2]
0.243 0.013
3.5 0.2
a
[BMMIM][NTf2] was used as solvent 2 in the model.
As many recent publications42,45−47 report preferential
solvation data in terms of the more comprehensive, albeit
more complex, two-step model proposed by Buhvestov et al.,48
these data were also fitted to that model, which yielded good
fits although no additional insight (see Supporting Information
for details and fitting parameters).
The preferential solvation of Reichardt’s dye within the
binary IL mixture would appear to suggest that a polar solute
capable of accepting hydrogen bonds would be primarily
solvated by [BMIM] cations at low mole fractions of added
template. Given that statistically significant rate increases are
observed at low template loadings (<10 mol %), it would seem
unlikely that preferential solvation of water is occurring. To
examine this more thoroughly, the hydrolysis reaction was
conducted at a variety of mole fractions of [BMMIM][NTf2] in
[BMIM][NTf2]. The results of this investigation are illustrated
in Figure 6.
Figure 5. Variation of ETN with the mole fraction of [BMMIM][NTf2]
in [BMIM][NTf2] at 25.0 °C. The bold line represents the ideal
polarity curve, and the thin line represents the fitted curve to the
preferential solvation model.
As illustrated in Figure 5, preferential solvation of Reichardt’s
dye occurs in mixtures of [BMMIM][NTf2] and [BMIM]
[NTf2]. The shape of the curve suggests, as would be expected,
that the stronger hydrogen bond donor ([BMIM]) is
interacting more strongly with the dye than [BMMIM] is.
Models for the preferential solvation of Reichardt’s dye have
been proposed and fitted to a wide variety of binary solvent
mixtures involving neutral solvents and binary mixtures of ILs
and neutral solvents.42−44 The simplest of these is illustrated in
eq 1, which models the process as a one step solvent exchange,
where RSi is Reichardt’s dye solvated by solvent i, and Si
indicates free solvent i.
Figure 6. Variation of the observed rate constant with the mole fraction
of [BMMIM][NTf2] in [BMIM][NTf2] with fitted regression line.
Reported errors are the standard deviations of replicate experiments.
RS + S ⇌ RS + S
1
(1)
1
2
2
As can be observed from Figure 6, a good linear fit is found
upon varying the proportion of [BMMIM][NTf2] in [BMIM]
[NTf2]. The linearity is in stark contrast with the preferential
From the above equilibrium, an expression can be derived for
the corresponding constant of solvent exchange (K), in terms
1862
dx.doi.org/10.1021/jp211543v | J. Phys. Chem. B 2012, 116, 1858−1864

The Journal of Physical Chemistry B
Article
solvation exhibited by Reichardt’s dye and indicates that the
composition of the solvation sphere of water is representative
of the proportions present in the bulk solvent, unlike the
cybotactic sphere of Reichardt’s dye, which differs significantly
from that of the bulk.
Harper et al.51−53 In their system, the IL stabilized the
transition state electrostatically as the charge developed upon
dissociation of the carbon chlorine bond. This stabilization was
offset by the entropy cost of solvent reorganization upon
ionization of the substrate.51−53 It should be noted that for the
alkyl chloride methanolysis, it is the solvation of the carbo-
cation and the chloride nucleofuge that is of primary
significance, as has been well-established for SN1 reactions in
neutral solvents.54 In the hydrolysis of [Ar3CPic][Cl], however,
we have found that the rate-determining factor is the solvation
of the nucleophile, not the nucleofuge, by the IL, thereby
indicating a clear difference between these two processes.
Hence, the varying extent of charge development, as has pre-
viously been discussed, is responsible for the different behaviors
observed.
Given that the apparent polarity of the medium toward water
appears to be responsible for the rate enhancement, the
linearity observed upon varying the proportion of [BMMIM]
[NTf2] in [BMIM][NTf2] is extremely interesting in light of
the preferential solvation of Reichardt’s dye. Recent reports
have suggested that water diffuses several times faster than
either the cation or anion in imidazolium based ILs.55−57
Conversely, larger molecules have been found to diffuse more
slowly, limited by the formation of suitably sized cavities.2,58
These factors may result in the water molecules encountering a
more representative solvation environment than Reichardt’s
dye does, with the latter molecule diffusing slower than the
general rate of solvent reorganization. This hypothesis is
currently being further investigated within our laboratory.
DISCUSSION
■
On the basis of the data obtained, it is apparent that the rate
enhancement observed upon the addition of the TSA to the IL
can be ascribed to polarity effects. This accounts for the similar
N
apparent behavior, both in terms of observed ET value and
rate enhancement, of 15 mol % [PBu4][NTf2], TSA, and
[BMMIM][NTf2]. The inability to observe these polarity shifts
using Reichardt’s dye as a probe can be rationalized from the
perspective of preferential solvation, which has been shown to
occur for [BMMIM][NTf2] and would account for the lack of
any detectable variation in ETN for the [PBu4][NTf2] and TSA
solutions.
Furthermore, the slightly larger magnitude of the rate en-
hancement for [PBu4][NTf2] with respect to [BMMIM][NTf2N]
can also be rationalized in terms of polarity effects as the ET
value for [PBu4][Cl] has been measured as 0.3849 (c.f., 0.5450
for [BMMIM][NTf2]). Although the former value was mea-
sured at elevated temperature (125 °C vs 25 °C) and using a
different anion (making direct comparisons complicated), the
values are sufficiently different to suggest that at 15 mol % of
each template, the average hydrogen bond acidity of the
medium would be lower for the [PBu4] cation.
[NMe4][NTf2] was the only template not to yield a signifi-
cantly faster rate at the 10 mol % concentration and also
yielded a higher observed value of ETN. It can be concluded, in
light of the other evidence, that the polarity increase results
from preferential solvation of Reichardt’s dye by [NMe4],
which presumably is more capable of hydrogen bonding than
the [BMIM] cation is, due to the short alkyl chains of the [NMe4]
cation. Therefore, [NMe4] does not dilute the [BMIM] cation’s
ability to hydrogen bond, if anything, it enhances it, resulting in
no rate enhancement and perhaps a marginal rate depression at
10 mol %.
To provide a thorough discussion of this system, examination
of the reactions conducted in IL and reactions in neutral sol-
vents should be compared. To enable this, hydrolysis experi-
ments were conducted within nitromethane. These suggested
that in a neutral solvent, ion-pairing between the chloride ion
and [Ar3CPic] cation had a more significant effect on the rate
than even the concentration of water. This is most likely due to
the chloride acting as a water carrier as chloride is a strong
hydrogen bond acceptor, meaning that an ion-paired substrate is
more likely to contain a cybotactic water molecule than a free
cation. Further to this, in more weakly hydrogen-bonding
solvents, such as chloroform or benzene, [Ar3CPic][Cl] was not
formed, and only the neutral FDMC was observed in solution.
A direct comparison between the results observed in IL and in
neutral solvent, therefore, cannot be made as the underlying
mechanisms in all cases are quite different. However, for
completeness, the rate data obtained in nitromethane and the
speciation data in benzene and chloroform are available in the
Supporting Information, along with a more thorough discussion
of those results.
CONCLUSIONS
■
The presence of a TSA in [BMIM][NTf2] did lead to a rate
enhancement for the hydrolysis of [Ar3CPic][Cl]. This
enhancement, however, could also be achieved via the use of
other cations, including [PBu4][NTf2] and [BMMIM][NTf2].
The mechanism of the rate enhancement was therefore
attributed to the dilution of the more strongly hydrogen-
bonding [BMIM] cation by a weaker hydrogen bond donor.
This illustrates that dissolved salts within ILs can modulate the
polarity of the medium in a similar fashion to conventional
binary solvent systems. These findings also indicate that
reaction engineering through the rational use of binary ILs
consisting of a salt with desirable solvation properties, such as
strong hydrogen bond donating or accepting groups, and
another IL with relevant physical properties, such as low
viscosity or melting point, is feasible and may enable new
classes of chemical transformations to proceed more selectively
and/or under milder conditions.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details for the synthesis and characterization of all
compounds, calculation details for dimerization experiments,
Reichardt’s dye polarity measurements and two-step prefer-
ential solvation model, comparison of [PBu4][NTf2] and the
TSA at various loadings, and rate and speciation data for neutral
solvent experiments. This material is available free of charge via
the Internet at http://pubs.acs.org.
It is also interesting to consider the results of this
investigation in light of other results obtained for unimolecular
solvolysis reactions within ILs. The primary such example is the
methanolysis of a tertiary alkyl chloride, extensively studied by
1863
dx.doi.org/10.1021/jp211543v | J. Phys. Chem. B 2012, 116, 1858−1864

The Journal of Physical Chemistry B
Article
(30) Chiappe, C.; Pomelli, C. S.; Rajamani, S. J. Phys. Chem. B 2011,
115, 9653−9661.
AUTHOR INFORMATION
■
Corresponding Author
*Phone: +612 9351 2581. Fax: +612 9351 3329. E-mail:
th.maschmeyer@chem.usyd.edu.au.
(31) Burrell, A. K.; Del Sesto, R. E.; Baker, S. N.; McCleskey, T. M.;
Baker, G. A. Green Chem. 2007, 9, 449−454.
(32) Bowden, S. T.; Thomas, T. L. J. Chem. Soc. 1940, 1242−1249.
(33) Marcoux, D.; Charette, A. B. J. Org. Chem. 2008, 73, 590−593.
(34) Hunter, C. A.; Low, C. M. R.; Rotger, C.; Vinter, J. G.; Zonta, C.
Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 4873−4876.
(35) Holbrey, J. D.; Reichert, W. M.; Nieuwenhuyzen, M.; Sheppard,
O.; Hardacre, C.; Rogers, R. D. Chem. Commun. 2003, 1636−1637.
(36) Hanke, C. G.; Johansson, A.; Harper, J. B.; Lynden-Bell, R. M.
Chem. Phys. Lett. 2003, 374, 85−90.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We would like to acknowledge Professor Christopher Hunter
for his assistance with dimerization modelling, Dr Ian Luck for
his assistance with NMR experiments, and the University of
Sydney for funding. C.C.W. would like to thank the University
of Sydney for a Vice Chancellor’s Research Scholarship.
(37) Harper, J. B.; Lynden-Bell, R. M. Mol. Phys. 2004, 102, 85−94.
(38) Mele, A. Chim. Oggi 2010, 28, 48−55.
(39) Katritzky, A. R.; Brycki, B. J. Am. Chem. Soc. 1986, 108, 7295−
7299.
(40) Reichardt, C. Green Chem. 2005, 7, 339−351.
(41) Taft, R. W.; Kamlet, M. J. J. Am. Chem. Soc. 1976, 98, 2886−
2894.
(42) Khupse, N. D.; Kumar, A. J. Phys. Chem. B 2011, 115, 711−718.
(43) Skwierczynski, R. D.; Connors, K. A. J. Chem. Soc. Perkin Trans.
2 1994, 467−472.
(44) Bosch, E.; Roses, M. J. Chem. Soc. Faraday Trans. 1992, 88,
3541−3546.
REFERENCES
■
(1) Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39,
3772−3789.
(2) Weingartner, H. Angew. Chem., Int. Ed. 2008, 47, 654−670.
(3) Earle, M. J.; Esperanca, J. M. S. S.; Gilea, M. A.; Lopes, J. N. C.;
Rebelo, L. P. N.; Magee, J. W.; Seddon, K. R.; Widegren, J. A. Nature
2006, 439, 831−834.
(4) Deetlefs, M.; Seddon, K. R. Green Chem. 2010, 12, 17−30.
(5) Welton, T. Chem. Rev. 1999, 99, 2071−2083.
(6) Hallett, J. P.; Welton, T. Chem. Rev. 2011, 111, 3508−3576.
(7) Stark, A. Top. Curr. Chem. 2009, 290, 41−81.
(8) Chiappe, C.; Pieraccini, D. J. Phys. Org. Chem. 2005, 18, 275−
297.
(9) Dupont, J. J. Braz. Chem. Soc. 2004, 15, 341−350.
(10) Leclerq, L.; Schmitzer, A. R. Supramol. Chem. 2009, 21, 245−
263.
(45) Salari, H.; Khodadadi-Moghaddam, M.; Harifi-Mood, A. R.;
Gholami, M. R. J. Phys. Chem. B 2010, 114, 9586−9593.
(46) Navarro, A. M.; Garcia, B.; Hoyuelos, F. J.; Penacoba, I. A.; Leal,
J. M. J. Phys. Chem. B 2011, 115, 10259−10269.
(47) Fortunato, G. G.; Mancini, P. M.; Bravo, M. V.; Adam, C. G.
J. Phys. Chem. B 2010, 114, 11804−11819.
(48) Buhvestov, U.; Rived, F.; Rafols, C.; Bosch, E.; Roses, M. J. Phys.
Org. Chem. 1998, 11, 185−192.
(49) Harrod, W. B.; Pienta, N. J. J. Phys. Org. Chem. 1990, 3, 534−
544.
(11) Canongia Lopes, J. N. A.; Padua, A. A. H. J. Phys. Chem. B 2006,
110, 3330−3335.
(50) Crowhurst, L.; Mawdsley, P. R.; Perez-Arlandis, J. M.; Salter,
P. A.; Welton, T. Phys. Chem. Chem. Phys. 2003, 5, 2790−2794.
(51) Man, B. Y. W.; Hook, J. M.; Harper, J. B. Tetrahedron Lett. 2005,
46, 7641−7645.
(12) Triolo, A.; Russina, O.; Bleif, H.-J.; Di Cola, E. J. Phys. Chem. B
2007, 111, 4641−4644.
(13) Wang, Y.; Voth, G. A. J. Am. Chem. Soc. 2005, 127, 12192−
(52) Yau, H. M.; Barnes, S. A.; Hook, J. M.; Youngs, T. G. A.; Croft,
A. K.; Harper, J. B. Chem. Commun. 2008, 3576−3578.
(53) Yau, H. M.; Chan, S. J.; George, S. R. D.; Hook, J. M.; Croft,
A. K.; Harper, J. B. Molecules 2009, 14, 2521−2534.
(54) Bateman, L. C.; Church, M. G.; Hughes, E. D.; Ingold, C. K.;
Taher, N. A. J. Chem. Soc. 1940, 979−1011.
12193.
(14) Schroder, U.; Wadhawan, J. D.; Compton, R. G.; Marken, F.;
Suarez, P. A. Z.; Consorti, C. S.; de Souza, R. F.; Dupont, J. New J.
Chem. 2000, 24, 1009−1015.
(15) Leclerq, L.; Suisse, I.; Nowogrocki, G.; Agbossou-Niedercorn, F.
Green Chem. 2007, 9, 1097−1103.
(16) Kowsari, M. H.; Alavi, S.; Ashrafizaadeh, M.; Najafi, B. J. Chem.
Phys. 2010, 132, 044507.
(55) Stark, A.; Zidell, A. W.; Hoffman, M. M. J. Mol. Liq. 2011, 160,
166−179.
(56) Menjoge, A.; Dixon, J.; Brennecke, J. F.; Maginn, E. J.;
Vasenkov, S. J. Phys. Chem. B 2009, 113, 6353−6359.
(57) Rollet, A.-L.; Porion, P.; Vaultier, M.; Billard, I.; Deschamps, M.;
Bessada, C.; Jouvensal, L. J. Phys. Chem. B 2007, 111, 11888−11891.
(58) Taylor, A. W.; Licence, P.; Abbott, A. P. Phys. Chem. Chem. Phys.
2011, 13, 10147−10154.
(17) Holbrey, J. D.; Reichert, W. M.; Nieuwenhuyzen, M.; Johnston,
S.; Seddon, K. R.; Rogers, R. D. Chem. Commun. 2003, 1636−1637.
(18) Lachwa, J.; Bento, I.; Duarte, M. T.; Lopes, J. N. A. C.; Rebelo,
L. P. N. Chem. Commun. 2006, 2445−2447.
(19) Aggarwal, A.; Lancaster, N. L.; Sethi, A. R.; Welton, T. Green
Chem. 2002, 4, 517−520.
(20) Leclerq, L.; Suisse, I.; Agbossou-Niedercorn, F. Chem. Commun.
2008, 311−313.
(21) Silva Santos, L.; Neto, B. A. D.; Consorti, C. S.; Pavam, C. H.;
Almeida, W. P.; Coelho, F.; Dupont, J.; Eberlin, M. N. J. Phys. Org.
Chem. 2006, 19, 731−736.
(22) Ross, J.; Xiao, J. Chem.Eur. J. 2003, 9, 4900−4906.
(23) Schramm, V. L. J. Biol. Chem. 2007, 282, 28297−28300.
(24) Chen, L.; Xu, S.; Li, J. Chem. Soc. Rev. 2011, 40, 2922−2942.
(25) Xu, Y.; Yamamoto, N.; Janda, K. D. Bioorg. Med. Chem. 2004,
12, 5247−5268.
(26) Lillelund, V. H.; Jensen, H. H.; Liang, X.; Bols, M. Chem. Rev.
2002, 102, 515−553.
(27) Gelles, E.; Hughes, E. D.; Ingold, C. K. J. Chem. Soc. 1954,
2918−2929.
(28) Swain, C. G.; Pegues, E. E. J. Am. Chem. Soc. 1958, 80, 812−819.
(29) Lui, M. Y.; Crowhurst, L.; Hallett, J. P.; Hunt, P. A.;
Neidermeyer, H.; Welton, T. Chem. Sci. 2011, 2, 1491−1496.
1864
dx.doi.org/10.1021/jp211543v | J. Phys. Chem. B 2012, 116, 1858−1864