Using Kamlet-Taft solvent descriptors to explain the reactivity of anionic nucleophiles in ionic liquids

Article abstract of DOI:10.1021/jo0615302

In this paper, we report the effect of ionic liquids on substitution reactions using a variety of anionic nucleophiles. We have combined new studies of the reactivity of polyatomic anions, acetate, trifuoroacetate, cyanide, and thiocyanide, with our previous studies of the halides in [C₄C ₁Py][Tf₂N], [C₄C₁py][TfO], and [C₄C₁im][Tf₂N] (where [C₄C ₁im]⁺ is 1-butyl-3-methylimidazolium and [C ₄C₁py]⁺ is 1-butyl-1-methylpyrrolidinium) and compared their reactivities, k₂, to the same reactions in the molecular solvents dichloromethane, dimethylsulfoxide, and methanol. The Kamlet-Taft solvent descriptors (α, β, π*) have been used to analyze the rates of the reactions, which were found to have a strong inverse dependency on the α value of the solvent. This result is attributed to the ability of the solvent to hydrogen bond to the nucleophile, so reducing its reactivity. The Eyring activation parameters (ΔH? and ΔS?), while confirming the reaction mechanism, do not offer obvious correlations with the Kamlet-Taft solvent descriptors.

Full text of DOI:10.1021/jo0615302

Using Kamlet-Taft Solvent Descriptors To Explain the Reactivity of
Anionic Nucleophiles in Ionic Liquids
Lorna Crowhurst, Ruben Falcone, N. Llewellyn Lancaster, Veronica Llopis-Mestre, and
Tom Welton*
Department of Chemistry, Imperial College London, South Kensington Campus, London, SW7 2AZ, UK
t.welton@imperial.ac.uk
ReceiVed July 24, 2006
In this paper, we report the effect of ionic liquids on substitution reactions using a variety of anionic
nucleophiles. We have combined new studies of the reactivity of polyatomic anions, acetate, trifuoroacetate,
cyanide, and thiocyanide, with our previous studies of the halides in [C4C1py][Tf2N], [C4C1py][TfO],
and [C4C1im][Tf2N] (where [C4C1im]+ is 1-butyl-3-methylimidazolium and [C4C1py]+ is 1-butyl-1-
methylpyrrolidinium) and compared their reactivities, k2, to the same reactions in the molecular solvents
dichloromethane, dimethylsulfoxide, and methanol. The Kamlet-Taft solvent descriptors (R, â, π*) have
been used to analyze the rates of the reactions, which were found to have a strong inverse dependency
on the R value of the solvent. This result is attributed to the ability of the solvent to hydrogen bond to
the nucleophile, so reducing its reactivity. The Eyring activation parameters (∆Hq and ∆Sq), while
confirming the reaction mechanism, do not offer obvious correlations with the Kamlet-Taft solvent
descriptors.
Introduction
synthetic flexibility. However, this offers potential difficulties
as well as opportunities. Although the large number of
component ions promises ionic liquids to be adaptable solvents,
either by careful selection from those ionic liquids already
available or through the possibility of designing new ionic
liquids, it raises the question of how to identify the optimum
combination of cation and anion for any given process. Hence,
the study of how altering physicochemical properties of ionic
liquids affects solute species and their reactions is becoming
an increasingly active area of research.5 It is in this area that
much of our recent efforts have been focused. We are attempting
to answer a number of questions, primarily, which reactions
can be predicted to be accelerated or decelerated when using
an ionic liquid rather than a molecular solvent and thence in
which of these cases is the identity of the ions of the particular
ionic liquid used important. There have been many studies
published that, often on the basis of at best semiquantitative
data, claim that special “ionic liquid effects” are responsible
for changes in reactivities observed when ionic liquids are used
as solvents. We are also seeking to show that when a sufficiently
Interest in the use of ionic liquids as solvents for synthesis
continues unabated.1,2 Ionic liquids have various useful proper-
ties, being often nonflammable, noncorrosive, and nonvolatile
under atmospheric conditions. They have been used as solvents
for a wide range of organic reactions. Much of the attracted
interest has been based on the possibility that they might offer
an environmentally benign alternative to conventional VOC
solvents. However, it should be noted that this is a matter of
some current contention.3 Another point of interest in ionic
liquids is their potential to act as “designer solvents”.4
There have been a great variety of cations and anions used
in the preparation of ionic liquids, showing their considerable
* Corresponding author. Phone: +44 (0)2075945763.
(1) (a) Zhao, H.; Malhotra, S. V. Aldrichimica Acta 2002, 35, 75. (b)
Olivier-Bourbigou, H.; Magna, L. J. Mol. Catal. A 2002, 182, 419. (c) Zhao,
D.; Wu, M.; Kou, Y.; Min, E. Catal. Today 2002, 74, 157. (d) Sheldon, R.
Chem. Commun. 2001, 2399. (e) Gordon, C. M. Appl. Catal., A 2001, 222,
101. (f) Wasserscheid, P.; Keim, W. Angew. Chem., Int. Ed. 2000, 39, 3772.
(g) Welton, T. Chem. ReV. 1999, 99, 2071.
(2) Ionic Liquids in Synthesis; Welton, T., Wasserscheid, P., Eds.; VCH-
Wiley: Weinheim, 2002.
(3) Scammells, P. J.; Scott, J. L.; Singer, R. D. Aust. J. Chem. 2005, 58,
155.
(4) Freemantle, M. Chem. Eng. News 1998, 76, 32.
(5) Chiappe, C.; Pieraccini, D. J. Phys. Org. Chem. 2005, 18, 275 and
references therein.
10.1021/jo0615302 CCC: $33.50 © 2006 American Chemical Society
Published on Web 10/19/2006
J. Org. Chem. 2006, 71, 8847-8853
8847

Crowhurst et al.
sophisticated description of solvent-solute interactions is used,
no such special effects need to be invoked.
Historically, nucleophilic substitutions have been used to
investigate the effects of solvents on chemical reactions, and
these effects have been rationalized as the Hughes-Ingold
rules.6,7 These predict that if charge separation increases as the
reaction system passes through the activated complex (e.g., an
SN2 reaction of neutral reagents), the effect of increased solvent
polarity is to increase the rate of reaction. Conversely, if charge
is destroyed (e.g., an SN2 reaction of oppositely charged
reagents) during the activation process, the effect of increased
solvent polarity is to reduce the rate of reaction. Finally, if the
charge becomes distributed over more atoms (e.g., an SN2
reaction of one charged with one neutral reagent) during the
activation process, the effect of increased solvent polarity is
also to reduce the rate of reaction. It should be noted that the
Hughes-Ingold rules are based upon an entirely electrostatic
model of solvation, with liquids being viewed as simple
dielectric media, and the effects of specific interactions, such
as hydrogen bonding, are ignored.
FIGURE 1. The ionic liquids used in this work (where [C4C1im]+ is
1-butyl-3-methylimidazolium and [C4C1py]+ is 1-butyl-1-methylpyr-
rolidinium).
SCHEME 1. Reaction of Methyl-p-nitrobenzenesulfonate
with an Anionic Nucleophile
We have previously reported quantitative kinetic studies of
nucleophilic substitutions by both halides8 and amines9 in a
range of ionic liquids, showing the effect of both cation and
anion variation. Chiappe has also made kinetic and mechanistic
studies of nucleophilic substitutions10 as well as other reac-
tions.11 Other investigators5,12 have studied reactions in ionic
liquids in depth, thus allowing quantitative comparison against
the same reactions in molecular solvents. This body of work
generally shows that, on one level, the general predictions of
reaction rates arising from the classical Hughes-Ingold rules
can be applied to ionic liquids, if they are treated as polar
solvents. However, closer inspection of the results reveals the
importance of specific ionic liquid ion-solute hydrogen-bonding
interactions on the reaction rates. For this reason, a study of
the nucleophilicity of polyatomic anions, that have been
deliberately selected to show different degrees of hydrogen
bonding, was made in both ionic liquids and molecular solvents,
using the same neutral substrate as in our previous work. It is
the results of this investigation that we report here.
temperature ionic liquids”, or ATIL’s. The reaction studied was
that of methyl-p-nitrobenzenesulfonate (Me-p-NBS) with the
appropriate nucleophile (Scheme 1). The use of alkylsulfonates
as substrates for studies of nucleophilicity is well established,7
and this substrate has been used in studies of nucleophilicity in
molecular solvents13 as well as by us in ionic liquids.
Results and Discussion
The rates of the reactions of the polyatomic anions acetate
([Ac]-), trifluoroacetate ([TFA]-), cyanide ([CN]-), and thio-
cyanide ([SCN]-) with methyl-p-nitrobenzenesulfonate were
measured in the ATIL’s [C4C1im][Tf2N], [C4C1py][Tf2N], and
[C4C1py][TfO] (see ESI), using the methodology previously
described.8 From these the bimolecular rate constants, k2, have
been determined. In addition, the reaction was performed in
the molecular solvents methanol (MeOH), dimethylsulfoxide
(DMSO), and dichloromethane (DCM). These results are
compared to halide nucleophilicity8 in the same reaction in Table
1.
Ionic liquids are by definition liquids that are composed
entirely of cations and anions. For experimental convenience,
all of the ionic liquids used in this work (see Figure 1) melt
below room temperature and so can be described as “ambient-
Referring first to the results in terms of the Hughes-Ingold
rules, it can be seen that although we would expect the reaction
to be slowed by increasing polarity, we actually see no
straightforward correlation between the dielectric constant of
the solvent and rate of reaction in the same medium. The
dielectric constants of the molecular solvents increase in the
order DCM, MeOH, and DMSO (8.93, 32.66, and 46.45,
respectively). Microwave dielectric spectroscopy has recently
been used to estimate static dielectric constants for some ionic
liquids.14 Although none of the ionic liquids for which values
are available were used in the study reported here, all of the
estimates available are in the range 11-15. Assuming that the
dielectric constants of our ionic liquids fall either within or close
to this range, then the polarity of the ionic liquids, as described
(6) (a) Ingold, C. K. Structure and Mechanism in Organic Chemistry,
2nd ed.; Bell: London, 1969. (b) Hughes, E. D.; Ingold, C. K. J. Chem.
Soc. 1935, 244. (c) Hughes, E. D. Trans. Faraday Soc. 1941, 37, 603. (d)
Hughes, E. D.; Ingold, C. K. Trans. Faraday Soc. 1941, 37, 657. (e) Cooper,
K. A.; Dhar, M. L.; Hughes, E. D.; Ingold, C. K.; MacNulty, B. J.; Woolf,
L. I. J. Chem. Soc. 1948, 2043.
(7) Reichardt, C. SolVents and SolVent Effects in Organic Chemistry,
2nd ed.; VCH (UK) Ltd.: Cambridge, 1998.
(8) (a) Lancaster, N. L.; Welton, T. J. Org. Chem. 2004, 69, 5986. (b)
Lancaster, N. L.; Salter, P. A.; Welton, T.; Young, G. B. J. Org. Chem.
2002, 67, 8855.
(9) Crowhurst, L.; Lancaster, N. L.; Pe´rez Arlandis, J. M.; Welton, T. J.
Am. Chem. Soc. 2004, 126, 11549.
(10) Chiappe, C.; Pieraccini, D.; Saullo, P. J. Org. Chem. 2003, 68, 6710.
(11) (a) Chiappe, C.; Leandri, E.; Pieraccini, D. Chem. Commun. 2004,
2536. (b) Chiappe, C.; Pieraccini, D. J. Org. Chem. 2004, 69, 6059. (c)
Chiappe, C.; Conte, V.; Pieraccini, D. Eur. J. Org. Chem. 2002, 2831. (d)
Chiappe, C.; Capraro, D.; Conte, V.; Pieraccini, D. Org. Lett. 2001, 3, 1061.
(12) (a) Skrzypczak, A.; Neta, P. Int. J. Chem. Kinet. 2004, 36, 253-
258. (b) Swiderski, K.; McLean, A.; Gordon, C. M.; Vaughan, D. H. Chem.
Commun. 2004, 590. (c) McLean, A. J.; Muldoon, M. J.; Gordon, C. M.;
Dunkin, I. R. Chem. Commun. 2002, 1880. (d) Grodkowski, J.; Neta, P.;
Wishart, J. F. J. Phys. Chem. A 2003, 107, 9794. (e) Skrzypczak, A.; Neta,
P. J. Phys. Chem. A 2003, 107, 7800.
(13) (a) Alluni, S.; Pica, M.; Reichenbach, G. J. Phys. Org. Chem. 2001,
14, 265. (b) Alluni, S.; Pero, A.; Reichenbach, G. J. Chem. Soc., Perkin
Trans. 2 1998, 1747. (c) Bunting, J. W.; Mason, J. M.; Heo, C. K. M. J.
Chem. Soc., Perkin Trans. 2 1994, 2291. (d) Dietze, P. E.; Hariri, R.;
Khattak, J. J. Org. Chem. 1989, 54, 3317. (e) Kurz, J. L.; Lu, Y.-W. J.
Phys. Chem. 1983, 87, 1444.
8848 J. Org. Chem., Vol. 71, No. 23, 2006

ReactiVity of Anionic Nucleophiles in Ionic Liquid
TABLE 1. Second-Order Rate Constants k2/M-1 s-1 for the Reaction of Anions with Methyl-p-nitrobenzenesulfonate in Ionic Liquids and
Molecular Solvents at 25 °Ca
k2/M-1 s-1
solvent
DMSO
[CN]-
17.7
[Ac]-
4.40
Cl-
Br-
I-
[TFA]-
[SCN]-
9.0 × 10-1
(1 × 10-2
1.07
3.91 × 10-1
(8 × 10-3
4.6 × 10-1
(1 × 10-1
6.0 × 10-2
(1 × 10-3
2.2 × 10-2
(1 × 10-3
1.9 × 10-2
(2 × 10-3
1.62 × 10-3
(6 × 10-5
1.92 × 10-1
(3 × 10-3
6.6 × 10-2
(2 × 10-3
4.1 × 10-2
(2 × 10-3
1.88 × 10-2
(8 × 10-4
2.3 × 10-2
(4 × 10-3
6 × 10-3
4.85 × 10-2
(6 × 10-4
6.4 × 10-3
(1 × 10-4
6.9 × 10-3
(1 × 10-4
2.50 × 10-3
(3 × 10-5
8.6 × 10-4
(5 × 10-5
5.14 × 10-5
(9 × 10-7
2.90 × 10-2
(0.4)
2.69
(0.1)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
)
(1 × 10-3
)
)
)
)
)
)
DCM
2.71 × 10-1
(3 × 10-3
9.0 × 10-2
(1 × 10-3
5.40 × 10-2
(6 × 10-4
9.3 × 10-3
(8 × 10-5
6.8 × 10-4
(2 × 10-5
2.27 × 10-2
(3 × 10-4
(5 × 10-2
)
)
)
)
)
)
)
)
)
)
(7 × 10-2
9.9 × 10-2
(5 × 10-3
3.9 × 10-2
(2 × 10-3
1.2 × 10-2
(1 × 10-3
4.0 × 10-4
(1 × 10-5
[C4C1py][TfO]
[C4C1py][Tf2N]
[C4C1im][Tf2N]
MeOH
1.91
6.1 × 10-3
(6 × 10-4
(6 × 10-2
1.56
4.61 × 10-3
(1 × 10-4
(5 × 10-2
3.86 × 10-1
(8 × 10-3
3.97 × 10-3
(8 × 10-5
1.62 × 10-2
(5 × 10-4
2.29 × 10-3
(4 × 10-5
(1 × 10-4
a Standard deviations for the values are given in parentheses.
TABLE 2. Second-Order Rate Constants Relative to Those in MeOH k2/k2(MeOH) for the Reaction of Anions with Me-p-NBS in Ionic
Liquids and Molecular Solvents at 25 °C
k2/k2(MeOH)
solvent
DMSO
[CN]-
[Ac]-
Cl-
Br-
I-
[TFA]-
[SCN]-
1093
166
118
96.3
23.8
1
6471
399
132
79.4
13.7
1
2250
2675
248
97.5
30.0
1
241
284
37.0
13.6
11.7
1
32.0
11.0
6.83
3.13
3.83
1
944
125
134
48.6
16.7
1
12.7
9.91
2.66
2.01
1.73
1
DCM
[C4C1py][TfO]
[C4C1py][Tf2N]
[C4C1im][Tf2N]
MeOH
by dielectric constant, lies between that of DCM and MeOH.
Hence, it is expected that the rates of the reactions should
decrease in that order (DCM > ATIL’s > MeOH > DMSO).
However, for no reaction studied is this true. Indeed, the
reactions are usually fastest in DMSO. This failure of the
Hughes-Ingold approach is because it assumes an entirely
electrostatic model of solvent-solute interactions and does not
take into account the effect of specific interactions such as
hydrogen bonding. Clearly, such specific interactions are
dominating solvent effects in these reactions. It can be seen
that not only do the absolute rates of the reactions change, but
the relative nucleophilicities and the ordering of the nucleophi-
licities of the different anions in the various solvents also
changes.
Viewing the data as a whole, the general trend in k2 values
is DMSO > DCM > [C4C1py][TfO] > [C4C1py][Tf2N] >
[C4C1im][Tf2N] > MeOH. However, it is clear that this trend
is not adhered to by all nucleophiles. Two examples of this can
be seen with the Cl- and I- data; the chloride reaction is faster
in DCM than in DMSO, and when using iodide as a nucleophile
[C4C1im][Tf2N] gives a greater k2 value than [C4C1py][TfO],
both being contrary to the general trend.
From this it can be seen that in all of the solvents studied
[CN]- is the strongest nucleophile, while [TFA]- is the poorest
in MeOH, DCM, [C4C1im][Tf2N], and [C4C1py][Tf2N], and
[SCN]- is the weakest nucleophile in DMSO and [C4C1py]-
[TfO].
Finally, the range in solvent dependency of the reaction, and
therefore the effective nucleophilicity, varies dramatically
between the different nucleophiles, with [Ac]- k2 values varying
dramatically between the different solvents, whereas [SCN]-
k2 values show very little change. It is informative, therefore,
to compare the variation in k2 values for each nucleophile upon
changing solvents (Table 2). From Table 1, it can be seen that
the lowest k2 values for all of the nucleophiles are found in
methanol. Consequently, all k2 values are shown relative to
MeOH.
Table 2 clearly shows that there is little solvent dependency
in k2 when using either [SCN]- or I- as the nucleophile, with
only 13- and 32-fold increases in rate, respectively, on going
from the slowest to the fastest solvent. In contrast to this, a
6471-fold increase is seen when using [Ac]- as the nucleophile.
To understand these effects, it is necessary to consider how
the different solvents might interact with each of the nucleo-
philes. A particularly successful approach when attempting to
quantitatively understand solvent-dependent data is the linear
solvation energy relationship (LSER).7 The equation, developed
by Kamlet and Taft, explains the variation of any solute property
in terms of three microscopic properties (R, â, and π*). R is a
quantitative scale of the hydrogen-bond acidity of a solvent, or
Another way of viewing the data is that the relative nucleo-
philicity of each anion differs, depending on the solvent used.
In fact, a different order of nucleophilicity is found in each of
the solvents investigated. These are:
DMSO
[CN]- > [Ac]- > Cl- > Br- > I- > [TFA]- > [SCN]-
[CN]- > Cl- > Br- > [Ac]- > I- > [SCN]- > [TFA]-
[CN]- > Cl- > [Ac]- > Br- > I- > [TFA]- > [SCN]-
[CN]- > [Ac]- > Cl- > Br- > I- > [SCN]- > [TFA]-
[CN]- > I- > Br- > Cl- > [Ac]- > [SCN]- > [TFA]-
[CN]- > I- > [SCN]- > Br- > [Ac]- > Cl- > [TFA]-
DCM
(14) (a) Wakai, C.; Oleinikova, A.; Ott, M.; Weinga¨rtner, H. J. Phys.
Chem. B 2005, 109, 17028. (b) Bonhoˆte, P.; Dias, A.-P.; Papageorgiou,
N.; Kalyanasundaram, K.; Gra¨tzel, M. Inorg. Chem. 1996, 35, 1168. (c)
Baker, S. N.; Baker, G. A.; Kane, M. A.; Bright, F. V. J. Phys. Chem. B
2001, 105, 9663. (d) Angelini, G.; Chiappe, C.; De Maria, P.; Fontana, A.;
Gasparrini, F.; Pieraccini, D.; Pierini, M.; Siani, G. J. Org. Chem. 2005,
70, 8193.
[C4C1py][TfO]
[C C py][Tf N]
4
1
2
[C C im][Tf N]
4
1
2
MeOH
J. Org. Chem, Vol. 71, No. 23, 2006 8849

Crowhurst et al.
TABLE 3. Kamlet-Taft Parameters for the Solvents Used in This
TABLE 4. LSER Correlations for ln k2 Obtained for the
Investigationa
Nucleophiles Useda
solvent
DMSO
R
â
π*
nucleophile used
R2
correlation obtained
0.00
0.04
0.40
0.42
0.61
1.05
0.76
-0.01
0.46
0.25
0.24
1.00
0.79
1.02
0.95
0.98
0.73
Cl-
0.99
ln k2 ) 0.21 - 7.56R
DCM
(â ) 1.00)
(S ) 0.34; F ) 369)b
[C4C1py][TfO]
[C4C1py][Tf2N]
[C4C1im][Tf2N]
MeOH
Br-
0.97
0.88
0.95
1.00
1.00
0.90
0.97
0.89
0.86
0.99
ln k2 ) -0.87 - 5.38R
(S ) 0.41; F ) 131)
(â ) 0.67)
I-
(â ) 0.30)
ln k2 ) -2.21 - 2.85R
(S ) 0.46; F ) 29)
0.61
ln k2 ) -2.57 - 3.05R + 1.16â
(S ) 0.33; F ) 30)
a All values were obtained using procedures previously described in the
literature, using only Reichardt’s dye, N,N-diethyl-4-nitroaniline, and
4-nitroaniline.16
Ac-
(â ) 1.49)
ln k2 ) -0.69 - 7.79R + 2.90â
(S ) 0.26; F ) 325)
ln k2 ) -2.37 - 7.60R + 2.65â + 1.83π*
(S ) 0.20; F ) 373)
its ability to donate a hydrogen bond; â is a scale of the
hydrogen-bond basicity of a solvent, or its ability to accept a
hydrogen bond; and π* is the solvent dipolarity/polarizability,
which is a scale of the ability of the solvent to stabilize a charge
or dipole. Each of the parameters is empirically obtained and
has been measured for a wide range of solvents,15 including
ionic liquids.16 These scales have been used in multi-parameter
equations to fit a number of different solvent-dependent
observations, with the most useful form shown in eq 1.
TFA-
(no â available)
ln k2 ) -3.62 - 5.66R
(S ) 0.82; F ) 36)
ln k2 ) -9.18 - 4.94R + 5.76π*
(S ) 0.48; F ) 57)
ln k2 ) -3.87 - 2.46R
(S ) 0.89; F ) 31)
ln k2 ) -2.03 - 5.75R
(S ) 1.00; F ) 25)
ln k2 ) -3.16 - 5.07R + 5.78π*
(S ) 0.99; F ) 128)
SCN-
(â ) 0.33)
CN-
(â ) 1.37)
a Where the best correlation was not obtained using R alone, the better
equation is shown beneath. The â values for the nucleophiles, shown in
brackets, are estimates taken from the literature.15 b Where R2 is the
proportion of variation in the response data explained by the descriptors; F
is a ratio of the mean square of the factor and the mean square of the error,
that is, the greater is the F value, the better is the fit; and S is an estimate
of the standard deviation in the model, that is, the smaller is the value, the
better is the fit.
XYZ ) XYZ0 + sπ* + aR + bâ
(1)
XYZ represents the solvent-dependent property under study
(e.g., the logarithm of a rate or equilibrium constant, or a λmax
value in a UV-vis or IR spectrum), and a, b, and s are solvent-
independent coefficients characteristic of the process and are
indicative of the sensitivity of the property under study to the
associated solvent property. This has been used in the analysis
of numerous reaction rates and equilibria, spectroscopic data,
and various other solvent-dependent processes.17 Here, we are
using the LSER model to quantify solvent effects on the rate
constant of a reaction. Since the relationship between the activa-
tion energy, Ea, and k is natural logarithmic in the Arrehnius
equation, ln k, rather than log k, must be used in LSER corre-
lations for the outcome to be physically meaningful (see
Table 4).
“ionic liquid effect” and that all significant interactions between
the ionic liquids and these solutes are adequately described by
an appropriate combination of their Kamlet-Taft parameters.
R appears in all of the LSER’s generated and always has a
negative value. For three of the nucleophiles, Cl-, Br-, and
[SCN]-, LSER’s based on R alone provide the best fit. Two of
the nucleophiles, I- and [Ac]-, have best fit LSER’s with â in
the correlations. π* appears in the best-fit LSER’s for three of
the nucleophiles, [CN]-, [Ac]-, and [TFA]-.
The Hughes-Ingold interpretation of polarity effects on this
reaction would lead us to expect a significant π* effect. Linear
solvation energy relationships based on the Kamlet-Taft param-
eters have been applied extensively to the solvolysis of tertiary
chlorides,19 but seldom to SN2 processes. An analysis of the
reaction of benzyl chloride and aniline showed a strong positive
correlations of log k2 with both π* and R.20 More recent studies
of the cyclization of 4-hydroxy-4-methyl-1-pentyl-p-toluene-
sulfonate (also an SN2 process) also showed positive π*
effects.21 We only find π* effects in the LSER’s for [CN]-,
[Ac]-, and [TFA]-. However, ionic liquids have a very narrow
range of π*. It was not possible to extend the range of π* values
by the selection of molecular solvents, because attempts to do
so led to failure due to lack of solubility of the ionic starting
materials. The consequence of this restriction is that the narrow
range of π* values may make this experiment unable to deter-
mine any effect that might otherwise have been seen. However,
in all of the cases where a correlation was seen, the s value is
positive, as expected from the Hughes-Ingold interpretation.
We have previously reported R, â, and π* values for a number
of ionic liquids,16 including those used here {with the exception
of [C4C1py][TfO]}. These values and those of the molecular
solvents investigated are given in Table 3.
To apply the Kamlet-Taft LSER (eq 1), data obtained with
each of the nucleophiles were first separated, and the solvent
dependency of each process was analyzed by correlating ln k2
with eq 1 in Minitab.18 The error associated with each parameter
was appraised in terms of the F-test, and the deviation associated
with the equation and unnecessary terms in the correlation were
eliminated. The results of these fits, along with the associated
statistical data, are shown in Table 4.
The first point is that acceptable correlations were achieved
using eq 1 with all of the nucleophiles, using data sets from
both ionic liquids and molecular solvents. Hence, the data clearly
demonstrate that, in this reaction at least, there is no special
(15) Marcus, Y. J. Phys. Chem. 1991, 95, 8886.
(16) (a) Crowhurst, L.; Mawdsley, P. R.; Perez-Arlandis, J. M.; Salter,
P. A.; Welton, T. Phys. Chem. Chem. Phys. 2003, 5, 2790. (b) Huddleston,
J. G.; Broker, G. A.; Willauer, H. D.; Rodgers, R. D. ACS Symp. Ser. 2002,
818, 270. (c) Muldoon, M. J.; Gordon, C. M.; Dunkin, I. R. J. Chem. Soc.,
Perkin Trans. 2 2001, 433.
(19) McManus, S. P.; Somani, S.; Harris, J. M.; McGill, R. A. J. Org.
Chem. 2004, 69, 8865. Catalan, J.; Diaz, C.; Garcai-Blanco, F. J. Org. Chem.
1999, 64, 6512.
(17) Kamlet, M. J.; Abboud, J. L. M.; Abraham, M. H.; Taft, R. W. J.
Org. Chem. 1983, 48, 2877.
(18) For information, see www.minitab.com.
(20) Kamlet, M. J.; Taft, R. W. J. Chem. Soc., Perkin Trans. 2 1979,
350.
(21) Shimizu, N.; Tsutsumi, T.; Tsuno, Y. Chem. Lett. 1991, 2065.
8850 J. Org. Chem., Vol. 71, No. 23, 2006

ReactiVity of Anionic Nucleophiles in Ionic Liquid
TABLE 5. Eyring Activation Parameters for the Reaction of
Chloride with Methyl-p-nitrobenzenesulfonate in Ionic Liquids and
Molecular Solvents
For all of the nucleophiles, a is negative, meaning that an
increase in the hydrogen-bond donating ability of the solvent
leads to a decrease in the rate of reaction. This can be attributed
to slowing of the reaction by the solvent hydrogen bonding to
the nucleophile, stabilizing it with respect to the activated
complex, which is a less strong hydrogen-bond acceptor than
is the reagent. It can be seen that the stronger the hydrogen-
bond acceptor character of the nucleophile (represented by the
nucleophile â values listed in Table 4), the larger this effect.
This is particularly emphasized when anions of similar structure
are compared {i.e., Cl- > Br- > I-; [Ac]- > [TFA]-; [CN]-
> [SCN]-}.
q
q
∆Hq
∆Sq
T∆S298K
∆G298K
solvent
DMSO
DCM (free ion)
DCM (ion pair)
[C4C1py][TfO]
[C4C1py][Tf2N]
[C4C1im][Tf2N]
MeOH
(kJ mol-1
)
(J K-1 mol-1
)
(kJ mol-1
)
(kJ mol-1
)
65.75
54.4
79.5
70.62
68.5
71.8
-25.48
-58.6
7.9
-25.92
-43.9
-42.2
-27.85
-7.6
-17.5
2.35
-7.73
-13.1
-12.6
-8.31
73.35
71.9
77.2
78.35
81.6
84.4
83.94
92.25
When comparing all of the nucleophiles, the effect of
changing the R value of the solvent on the reaction, as described
in eq 1, broadly splits into three groups: Cl- ≈ [Ac]- > [CN]-
≈ [TFA]- ≈ Br- > I- ≈ [SCN]-. This again fits with the
qualitative order that would be predicted from the â values of
the anionic nucleophiles alone, with the exception of [CN]-
(see below). This is an excellent demonstration of the intimacy
of the solvent-solute relationship. While a solvent may have
the potential to behave as a hydrogen-bond donor to a solute
species, this potential is only realized when the solute is a good
hydrogen-bond acceptor.
TABLE 6. Eyring Activation Parameters for the Reaction of
[CN]- with Methyl-p-nitrobenzenesulfonate in Ionic Liquids and
Molecular Solvents
q
q
∆Hq
∆Sq
T∆S298K
∆G298K
solvent
DMSO
(kJ mol-1
)
(J K-1 mol-1
)
(kJ mol-1
)
(kJ mol-1
)
)
)
)
28.89
49.15
66.47
58.03
39.23
58.68
-125.27
-73.25
-20.24
-47.44
-121.85
-82.02
-37.35
-21.82
-6.03
-14.13
-36.33
-24.45
66.24
70.97
72.50
72.16
75.56
83.13
DCM
[C4C1py][TfO]
[C4C1py][Tf2N]
[C4C1im][Tf2N]
MeOH
The apparent anomaly of the [CN]- result, for which a higher
a would be expected on the basis of its â value alone, is
consistent with the need to consider the hard/soft nature of the
bases in addition to a simple strong/weak interpretation. While
[CN]- can be considered to be a strong base (â ) 1.37), it is a
significantly softer base than Cl-, Br-, or either of the acetate
ions. Consequently, its nucleophilicity toward carbon centers
should be more difficult to reduce through hydrogen-bond
donation, which can be considered to be a hard interaction.
Hence, the ionic liquids used here are acting as hard solvents.
Both I- and [SCN]- are soft bases as well as weak bases.
Consequently, they show the lowest ranges of solvent effects
in these reactions, with significantly lower a values than the
other nucleophiles used. As far as we are aware, this is the first
time that such a hard/soft classification of ionic liquids has been
made, and it is certainly worthy of further study.
In the two cases where the use of â leads to an improvement
in the LSER correlations, b has a positive value. The value of
b is significantly lower than a, suggesting that it is a secondary
effect. We have previously shown that the hydrogen-bond
donation of ionic liquids of a particular cation can be moderated
by the hydrogen-bond acceptor ability of the ionic liquid anion.16
Stronger hydrogen-bond acceptor anions hydrogen bond to the
ionic liquid cation more strongly and reduce the availability of
the cation hydrogen bond to the solute. This competition for
the hydrogen-bond donor site is probably the effect that is seen
in these results.
In our previous studies of halide nucleophilicity in ionic
liquids, we noted that a clear effect of changing the cation could
be seen, but no systematic effect of changing anion could be
elucidated.8 This can now be explained. The hydrogen-bond
donor ability, R, of ionic liquids to anionic hydrogen-bond
acceptors has been shown to be largely a property of the
cation.9,22 Consequently, changing the cation will affect the rate
of the reaction. The principal effect of changing the anion of
the ionic liquid on the Kamlet-Taft parameters is to change â.
Because â is not found to correlate with the reaction rates of
Cl- and Br- and only has a small effect on the reaction rate of
TABLE 7. Eyring Activation Parameters for the Reaction of
[SCN]- with Methyl-p-nitrobenzenesulfonate in Ionic Liquids and
Molecular Solvents
q
q
∆Hq
∆Sq
T∆S298K
∆G298K
solvent
DMSO
(kJ mol-1
)
(J K-1 mol-1
)
(kJ mol-1
)
(kJ mol-1
69.54
66.77
58.57
45.34
65.51
67.94
-41.86
-51.83
-89.62
-137.47
-71.22
-68.05
-12.48
-15.45
-26.71
-40.97
-21.23
-20.29
82.02
82.22
85.29
86.31
86.74
88.23
DCM
[C4C1py][TfO]
[C4C1py][Tf2N]
[C4C1im][Tf2N]
MeOH
TABLE 8. Eyring Activation Parameters for the Reaction of
Acetate with Methyl-p-nitrobenzenesulfonate in Ionic Liquids and
Molecular Solvents
∆Hq
∆Sq
T∆S298K
∆G298K
q
q
solvent
DMSO
(kJ mol-1
)
(J K-1 mol-1
)
(kJ mol-1
)
(kJ mol-1
42.40
52.80
62.15
48.31
25.23
83.57
-89.65
-79.48
-55.21
-107.22
-199.14
-25.33
-26.73
-23.68
-16.46
-31.95
-59.35
-7.55
69.13
76.48
78.61
80.26
84.58
91.12
DCM
[C4C1py][TfO]
[C4C1py][Tf2N]
[C4C1im][Tf2N]
MeOH
TABLE 9. Eyring Activation Parameters for the Reaction of
Trifluoroacetate with Methyl-p-nitrobenzenesulfonate in Ionic
Liquids and Molecular Solvents
q
q
∆Hq
∆Sq
T∆S298K
∆G298K
solvent
DMSO
(kJ mol-1
)
(J K-1 mol-1
)
(kJ mol-1
)
(kJ mol-1
65.20
72.82
72.93
68.08
73.89
65.10
-51.65
-41.57
-41.13
-66.04
-56.65
-107.62
-15.4
80.60
85.21
85.19
87.76
90.78
97.19
DCM
-12.39
-12.26
-19.68
-16.88
-32.09
[C4C1py][TfO]
[C4C1py][Tf2N]
[C4C1im][Tf2N]
MeOH
I-, changing the anion does not have a noticeable systematic
effect on the nucleophilicities of the halides.
Eyring Activation Parameters
To gain greater insight into the reaction process, the activation
(22) Hunt, P.; Kirchner, B.; Welton, T. Chem.-Eur. J. 2006, 12, 600.
enthalpy (∆Hq) and entropy (∆S q) were both determined using
J. Org. Chem, Vol. 71, No. 23, 2006 8851

Crowhurst et al.
SCHEME 2
q
TABLE 10. LSER Correlations for ∆G298K Obtained for the
the Eyring equation. These values are reproduced in Tables 5-9.
In our previous studies of the halides,8 we did not study the
activation parameters for iodide, because the data were not of
the same high quality as those for chloride. Furthermore, in
[C4C1im][BF4] and in [C4C1im][PF6] the activation parameters
for bromide and chloride were largely similar. Hence, only the
reactions of chloride were studied in detail.
Nucleophiles Used in This Investigationa
nucleophile used
R2
correlation obtained
Cl-
0.98
∆G298Kq ) 72.3 + 19.2R
(â ) 1.00)
(S ) 1.38; F ) 164)
[Ac]-
0.95
1.00
0.97
0.91
0.95
0.98
∆G298Kq ) 71.4025 + 19.9R
(S ) 1.86; F ) 70)
(â ) 1.49)
∆G298Kq ) 90.0 + 21.6R - 3.2â - 18.5π*
(S ) 0.44; F ) 767)
SN2 reactions are associative, and negative activation entro-
pies are expected. However, in DCM, the reaction of Cl- has
been shown to proceed through both the discrete anion and the
ion pair.13 ∆Sq for the ion-paired chloride is slightly positive
and, therefore, atypical for this reaction. When studying the
chloride reaction in ionic liquids, it was noted that ∆Sq for the
reaction was more similar to the reaction of free solvated ions,
but ∆Hq was more similar to that of the ion-paired chloride
nucleophile. This led us to consider the nature of ion pairing in
ionic liquids. In a pure ionic liquid, there are no molecules
available to separate ions, and the ions of the solute will be
repelled by the ions of the ionic liquid bearing the same charge
(cation-cation or anion-anion). Hence, in an ionic liquid, a
solute ion will always be closely associated with ions of opposite
charge. Studies of the compound [C1C1im]Cl (where [C1C1im]+
is 1,3-dimethylimidazolium) by neutron diffraction showed that
the anion is coordinated by 6 cations within 6.5 Å.23 This is
not the same as an ion pair, but it is certainly not a free ion.
We used this to propose a mechanism for the reaction, which
included the need to remove a cation from the nucleophile to
liberate an active site for reaction before the activation process
itself occurred, which also involved the reduction hydrogen
bonding (see Scheme 2). Although the number of cations
surrounding the other anions used as nucleophiles in this study
may be different, it is likely that they are still maximally solvated
by cations. The activation parameters for all of the other
nucleophiles used in this project are also consistent with reaction
by the same mechanism.
[TFA]-
∆G298Kq ) 94.2 + 12.4R - 12.8π*
(S ) 1.29; F ) 47)
(no â available)
[SCN]-
∆G298Kq ) 82.5 + 6.19R
(S ) 0.85; F ) 40)
(â ) 0.33)
[CN]-
(â ) 1.37)
∆G298Kq ) 66.8 + 15.7R
(S ) 1.437; F ) 72.91)
∆G298Kq ) 79.9 + 12.3R - 12.8π*
(S ) 0.99; F ) 80)
a Where the best correlation was not obtained using R alone, the better
equation is shown beneath. The â values for the nucleophiles, shown in
brackets, are estimates taken from the literature.15
the case. That ∆Gq was derived from ∆Hq and ∆Sq and still
gives LSER’s of the same form as those for ln k2, also confirms
that the forms of these LSER’s are not arising as artifacts of
the calculations. Because the values for ∆Gq and ln k2 cover
very different ranges (66-97 kJ mol-1 and 4 × 10-4 to 17.7,
respectively), the values of the coefficients in the LSER’s are,
of course, expected to be different.
It can be seen that, as for ln k2, changes in ∆Gq with solvent
can primarily be attributed to the changing ability of the solvent
to act as a hydrogen-bond donor, R, with the other Kamlet-
Taft parameters only occasionally being used in the LSER. This
again confirms the proposed mechanism (Scheme 2), in which
the nucleophile becomes a poorer hydrogen-bond acceptor as
it enters the activated complex.
Conclusion
It was found that applying the Kamlet-Taft LSER approach
to kinetic data from several anionic nucleophilic substitution
reactions of methyl-p-nitrobenzenesulfonate in both molecular
and ionic liquids yielded consistent results. This approach shows
that this reaction does not display an “ionic liquid effect”. The
correlations can be explained mainly in terms of the solvent
donating a hydrogen bond to the nucleophilic anion and thus
reducing its capacity as a nucleophile. The effect of using a
range of multi-atom nucleophiles, rather than just the simple
halides that had been studied previously, was also demonstrated.
It was shown that the degree of correlation between the rate
constant and R was dependent on the nucleophile itself, with
more basic anions showing a higher degree of solvent depen-
dency in their reaction with methyl-p-nitrobenzenesulfonate.
Furthermore, the hard/soft nature of the nucleophile was also
shown to be significant, with ionic liquids with the [C1C4im]+
cation being shown to act as hard solvents, which therefore
interact more strongly with hard solutes (e.g., Cl-) rather than
soft solutes (e.g., [CN]-).
To understand more about the effects of changing solvent
on these reactions, LSER’s of these reactions were attempted.
However, all attempts to achieve statistically reliable LSER
correlations failed. Subtle differences in ostensibly similar
reactions can have significant effects on the balance of the
activation parameters. It is likely that this is preventing a
successful LSER analysis for ∆Sq and ∆Hq of these reactions.
∆Gq for the reactions was calculated from ∆Sq and ∆Hq. ∆Gq
is directly proportional to ln k, and the LSER correlations for
∆Gq might be expected to be almost identical to those seen
with ln k2. These are reproduced in Table 10.
The best fit LSER’s for ∆Gq are superficially similar to those
for ln k2, but very different in detail. For ∆Gq, the best
correlations are achieved with R alone for Cl- and [SCN]-, as
they were for ln k2. Ac- has best fit LSER’s, for both ∆Gq and
ln k2 include both R and â. Finally, the best fit ∆Gq LSER’s
for [Ac]-, [TFA]-, and [CN]- include π*, again as they do for
ln k2. Given the relationship between ∆Gq and ln k2, their
LSER’s should, at least, have the same form, as is seen to be
Experimental Section
Materials. All syntheses were performed under anaerobic
(23) Hardacre, C.; Holbrey, J. D.; McMath, S. E. J.; Bowron, D. T.;
Soper, A. K. J. Chem. Phys. 2003, 118, 273.
conditions using standard Schlenk techniques. All heterocycles were
8852 J. Org. Chem., Vol. 71, No. 23, 2006

ReactiVity of Anionic Nucleophiles in Ionic Liquid
distilled from potassium hydroxide, alkyl halides from phosphorus
pentoxide, and solvents from standard drying agents before use.
Methyl p-nitrobenzenesulfonate and lithium bis(trifluoromethyl-
sulfonyl)imide were purchased from commercial sources and used
as received. The preparations and spectral data of the ionic liquids
have been described elsewhere.24
were recorded at regular time intervals. UV/vis spectra were
recorded using a spectrophotometer fitted with a thermostated
sample holder.
Acknowledgment. We wish to thank the EPSRC (V.L.-M.
and L.C.) and GSK (L.C.) for studentships, and Fundacion
Antorchas for a research fellowship (R.F.).
Kinetic Procedure. To the ionic liquid (1.5 cm3, freshly
outgassed in Vacuo) was weighed a solution of [cat][nucleophile]
(0.2 mmol, where cat ) the cation of the ionic liquid) into a 0.5
cm path length UV/vis quartz cuvette under anaerobic conditions.
At known time, an aliquot of methyl p-nitrobenzenesulfonate (5 ×
10-7 mol) in dichloromethane (0.1 cm3) was added, and spectra
Supporting Information Available: Tables containing the
observed rates of reactions of the nucleophiles discussed in all of
the solvents used at various temperatures. This material is available
free of charge via the Internet at http://pubs.acs.org.
(24) Cammarata, L.; Kazarian, S. G.; Salter, P. A.; Welton, T. Phys.
Chem. Chem. Phys. 2001, 3, 5192.
JO0615302
J. Org. Chem, Vol. 71, No. 23, 2006 8853