Ionic liquids as designer solvents for nucleophilic aromatic substitutions

Article abstract of DOI:10.1021/ol702435f

(Chemical Equation Presented) Ionic liquids were designed to optimize the nucleophilic aromatic substitution reaction of an activated aniline with an activated arylhalide. The design was achieved by selecting the anions on the basis of calculations of the

Full text of DOI:10.1021/ol702435f

ORGANIC
LETTERS
2007
Vol. 9, No. 25
5247-5250
Ionic Liquids as Designer Solvents for
Nucleophilic Aromatic Substitutions
Ian Newington,† Juan M. Perez-Arlandis,‡ and Tom Welton*,‡
GE Healthcare Bio-Sciences, The GroVe Centre, White Lion Road Amersham,
HP7 9LL, and Imperial College of Science Technology and Medicine,
South Kensington, London SW7 2AZ, UK
t.welton@imperial.ac.uk
Received October 5, 2007
ABSTRACT
Ionic liquids were designed to optimize the nucleophilic aromatic substitution reaction of an activated aniline with an activated arylhalide. The
design was achieved by selecting the anions on the basis of calculations of the gas-phase basicity of their conjugate acids.
Ionic liquids have attracted much attention recently as
solvents for chemicals synthesis.1 They have various useful
properties, often being nonflammable, noncorrosive, and
nonvolatile under atmospheric conditions. Consequently, they
have been used as solvents for a wide range of organic
reactions. Although controversial,2 much of the work in the
area has been based on the possibility that they might offer
an environmentally benign alternative to conventional VOC
solvents. Our interest has largely focused on the possibility
that they may substantially alter the reactivity of dissolved
solutes. This potential has led to ionic liquids being described
as “designer solvents”.3 However, although there is extensive
literature describing the diversity of ionic liquids available,1
we know of no example to date of when this has been applied
to designing an ionic liquid for a particular reaction, which
we report here.
predictably determines the macroscopic outcome that one is
trying to control, then manipulate that property to achieve
the desired outcome. The aim of this paper is to demonstrate
how knowledge of a combination of experimental and
theoretical understandings of ion properties can be used to
design an ionic-liquid solvent to maximize the efficiency of
a nucleophilic aromatic substitution (SNAr) reaction.
Arylamines are versatile intermediates with widespread
use in polymers, pharmaceuticals, and photography. These
are usually prepared by nucleophilic aromatic substitution,
which is one of the fundamental reactions in traditional
synthetic organic chemistry.5 Classical methods for the
synthesis of arylamines typically require a large excess of
amine, addition of base and highly polar solvents such as
DMSO6 and DMF7 at high temperatures with highly activated
arylhalides.8,9 In this paper we report the use of the designer
approach to the selection of ions for ionic liquid solvents
for a model “base-free” nucleophilic aromatic substitution
reaction.
First it is necessary to define what we mean by solvent
design. This should be separated from the task-specific ionic-
liquid concept, in which a functional group that takes part
in a reaction is introduced to one of the ionic-liquid species.1,4
Also, for the process to be truly one of design, it is necessary
to identify some fundamental property of the solvent that
(4) Wierzbicki, A.; Davis, J. H., Jr. In Proceedings of the Symposium
on AdVances in SolVent Selection and Substitution for Extraction, March
5-9, 2000, Atlanta, Georgia; AIChE: New York, 2000.
(5) Smith, M. B.; March, J. AdVanced Organic Chemistry, 5th ed.; Wiley-
VCH: New York, 2001.
(6) Brown, G. R.; Foubister, A. J.; Ratcliffe, P. D. Tetrahedron Lett.
1999, 40, 1219.
† GE Healthcare Bio-Sciences.
‡ Imperial College of Science Technology and Medicine.
(1) Wasserscheid, P., Welton, T., Eds. Ionic Liquids in Synthesis, 2nd
ed.; VCH-Wiley: Weinheim, Germany, 2007.
(2) (a) Scammells, P. J.; Scott, J. L.; Singer, R. D. Aust. J. Chem. 2005,
58, 155. (b) Clark, J. H.; Tavener, S. J. Org. Process Res. DeV., 2007, 11,
149.
(7) Cho, Y. H.; Park, J. C. Tetrahedron Lett. 1997, 38, 8331.
(8) Shaw, J. E.; Kunerth, D. C.; Swanson, S. B. J. Org. Chem. 1976,
41, 732.
(3) Freemantle, M. Chem. Eng. News 1998, 76, 32.
(9) Kulagowsky, J. J.; Rees, C. W. Synthesis 1980, 215.
10.1021/ol702435f CCC: $37.00
© 2007 American Chemical Society
Published on Web 11/15/2007

Previous investigations of nucleophilic aromatic substitu-
tions in ionic liquids include the N-arylation of secondary
cyclic amines10 and the substitution of some 2-substituted-
5-nitrothiophenes with cyclic amines.11 In the latter, signifi-
cantly greater rates of the reactions were observed in the
ionic liquids [C4C1im][BF4], [C4C1im][PF6], and [C4C1C1-
im][BF4] (where [C4C1im]+ is the 1-butyl-3-methylimida-
zolium cation and [C4C1C1im]+ is the 1-butyl-2,3-dimeth-
ylimidazolium cation) than in methanol or benzene. The
effects of the ionic liquids were found to be complex and to
be dependent upon both the cation and anion.
Our initial studies involved the use of nine of the more
commonly used ionic liquids, four molecular solvents with
suitable boiling points for the established experimental
conditions, and reaction with no added solvent. The reactions
were monitored by measuring the yields after 24 h.
Clearly the nature of the solvent has a dramatic effect on
the reaction (Table 1). Toluene (nonpolar, aprotic solvent)
gives an exceptionally poor result, as does propylene
carbonate {PC, polar, moderate donor (i.e., hydrogen-bond
acceptor) solvent}. N,N′-Dimethylpropylene urea {DMPU,
polar, strong donor (i.e., hydrogen-bond acceptor) solvent}
provides the best solvent for this reaction. Hence, it can be
concluded that the yields of this reaction after 24 h are greatly
increased by using a solvent that is a strong donor. This is
in good agreement with previous investigations of solvent
effects and mechanistic predictions.12
materials, transition states, and/or the intermediate. In ad-
dition to simple dipole interactions between the transition
state and the solvent, it is important to consider the hydrogen-
bonding interactions that can occur between the amine/
solvent, transition state/solvent, and Meisenheimer complex/
solvent.
Solvents with hydrogen-bond accepting properties can
interact with the protons on the aniline, increasing the
electron density on the nitrogen atom and therefore its
nucleophilic character. This increase has been shown when
DMSO is used as the solvent for primary and secondary
amines.16 Also, the similarity between the transition state in
these SNAr reactions and the Meisenheimer complex inter-
mediate predicts that factors that contribute to the stabiliza-
tion of the transition state will also tend to stabilize the
complex. Investigations on the stability of Meisenheimer
complexes17 in various solvents have shown the formation
of more stable complexes in dipolar aprotic solvents relative
to protic solvents. Hence, this reaction should be accelerated
by polar, hydrogen-bond accepting solvents.
The reaction in DMF appears at first glance to contradict
this analysis. Although DMF is in many ways similar to
DMPU comparable yields were not obtained. However, in
this reaction it was observed that the amount of unreacted
p-fluoronitrobenzene was much lower than expected (only
30%) and the formation of p-dimethylaminonitrobenzene was
noted in HPLC results. Hence, formation of this byproduct
competes too well with the primary reaction for this, or
closely related solvents, to be useful.
Nucleophilic aromatic substitution reactions involving
activated substrates and good leaving groups are known to
proceed through an addition-elimination mechanism13
(Scheme 1) involving the formation of an intermediate
Meisenheimer complex.14,15
The effects of some of the most common ionic liquids
used in the literature on this reaction were also studied (Table
1). Since its rate determining step involves the formation of
Scheme 1
Table 1. Nucleophilic Aromatic Substitution of
p-Fluorobenzene and p-Anisidine in Molecular Systems and
Ionic Liquidsa
solvent
yield %
toluene
PC
DMF
DMPU
solventlessb
[C4C1pyrr][TfO]
[C4C1im][TfO]
[C4C1C1im][TfO]
[C4C1im][BF4]
[C4C1C1im][BF4]
[C4C1C1im][N(Tf)2]
[C4C1im][PF6]
[C4C1im][N(Tf)2]
[C4C1pyrr][N(Tf)2]
0.1
0.9
22.1
41.4
2.4
5.2
4.9
3.4
2.4
2.1
0.4
0.4
0.4
0.4
During the first step of this reaction, the nucleophilic
anisidine attacks the 1-fluoro-4-nitrobenzene and bonds to
the carbon that bears the halogen. This step, involving the
loss of aromaticity, generally proceeds slowly. In the second
step, the complex formed loses the halide in a fast step, where
the aromaticity of the ring is restored. The solvent can affect
the rates of these reactions by interaction with starting
a Conditions: slight excess p-anisidine (1:1.05 molar ratio), 100 °C, 24
h, under dry N2. Yield determined by HPLC using internal standard. b The
starting materials were mixed together and the reaction was heated to 100 °C
and stirred for 24 h. In this case the p-anisidine was added in 50% excess.
(10) Yadav, J. S.; Reddy, B. V. S.; Basak, A. K.; Venkat Narsaiah, A.
Tetrahedron Lett. 2003, 44, 2217.
(11) D’Anna, F.; Frenna, V.; Noto, R.; Pace V.; Spinelli, D. J. Org. Chem.
2006, 71, 5144.
(12) (a) Terrier, F. Nucleophilic Aromatic Displacement; VHC: Wein-
heim, Germany, 1991. (b) Nudelman, N. S.; The Chemistry of Amino,
Nitroso, and Related Groups; Patai S., Ed.; Wiley: Winchester, 1996;
Chapter 26.
a dipolar activated complex in which, compared to the initial
state, there are considerable differences in charge distribution,
5248
Org. Lett., Vol. 9, No. 25, 2007

we proposed that ionic liquids could be suitable solvents
in which to conduct the reaction of p-fluoronitrobenzene
and p-anisidine. We have also previously noted the ability
of ionic liquids to accelerate nucleophilic substitution
reactions of amines.19 However, none of the ionic liquids
gave a synthetically useful yield for the reaction under
the conditions used. Even with these very low yields some
trends could be identified. Between the ionic liquids, the
nature of the anion appears to be the factor that has more
effect on the yields of the reaction. We have previously
shown that the ability of these ionic liquids to act as hydro-
gen bond acceptors is a property of the anions.20 Hence, this
is in accord with the results for the molecular solvents.
Higher yields were achieved for triflate ([TfO]-) ionic
liquids, while those based on the bis(trifluoromethylsulfonyl)-
imide ([N(Tf)2]-) anion and [PF6]- provided the lowest. For
ionic liquids based on [TfO]- and [N(Tf)2]-, [C4C1im]+ gave
slightly lower yields than [C4C1pyrr]+ (where [C4C1pyrr]+
is the 1-butyl-1-methylpyrrolidinium cation) ionic liquids.
However, [C4C1C1im]+ gave the lowest yield of the [TfO]-
ionic liquids, whereas for the [N(Tf)2]- ionic liquids it gave
the highest. Given the very low yields of these reactions, it
is possible that these differences arise from experimental
error.
reaction. For this reason the [CH3SO3]- and [CF3COO]-
versions of the [C4C1im]+ and [C4C1pyrr]+ ionic liquids were
prepared.18
A considerable improvement was achieved through the
use of these more basic anions (see Table 2). Both conver-
Table 2. Nucleophilic Aromatic Substitution of
p-Fluorobenzene and p-Anisidine in “Designed” Ionic Liquidsa
solvent
yield %
conversion %
difference %
[C4C1pyrr][CH3SO3]
[C4C1im][CH3SO3]
[C4C1pyrr][CF3COO]
[C4C1im][CF3COO]
76.0
76.7
69.1
67.6
77.8
77.6
74.3
73.0
1.8
0.9
5.2
5.4
a Conditions slight excess p-anisidine (1:1.05 molar ratio), 100 °C, 24
h, under dry N2. Yield determined by HPLC using an internal standard.
sions and yields were increased, with a yield of 78% for the
best ionic liquid, which is almost twice as good as the best
molecular solvent tested (DMPU) for the same reaction time
and temperature. The [CF3COO]--based ionic liquids gave
marginally lower conversions than the [CH3SO3]- ionic
liquids. Although this difference is small, it is consistent with
the expectations from the ∆GH values. The difference was
more pronounced when comparing the yields of the reactions.
For the [CH3SO3]- ionic liquids the difference between the
yields and conversions was small, but it was much larger
for the [CF3COO]- ionic liquids. The most likely explanation
for this is that the [CF3COO]- anion is sufficiently nucleo-
philic that at the high concentrations that it is present (as a
component of the solvent) it reacts with the 1-fluoro-4-
nitrobenzene.22
So, the reaction yield after 24 h in ionic liquids is in-
creased by more hydrogen bond accepting anions. The
question is, can an ionic liquid be designed such that it
provides a higher yielding solvent for the reaction than any
of the molecular solvents that have been used? The huge
range of potential ionic liquids means that this may be
possible, but the task of ion selection is not straight-
forward.
We have shown previously that good correlations were
obtained between the theoretical Gibbs free energy change
upon deprotonation of the conjugate acid of the anion
in the gas phase (∆GH) and the hydrogen bond basicity of
the ionic liquids.20 Koppel et al.21 have published an extended
list of calculated ∆GH values for some neutral Brønsted
superacids. Therefore, we opted to select candidate anions
from this list. The highest values of ∆GH in this study were
obtained for CH3SO3H and CF3COOH. This led us to the
prediction that ionic liquids based upon [CH3SO3]- and
[CF3COO]- anions should be effective solvents for this
The small cation effect for those ionic liquids sharing the
same anion was also noted. The yields and conversions were
found to be slightly higher for those based in the pyrroli-
dinium cation. This is probably attributable to the slowing
of the reaction by the better hydrogen-bond donor cation,
[C4C1im]+, interacting more strongly with the nitrogen of
the anisidine. We have seen similar behavior in our studies
of SN2 reactions in ionic liquids.19
Traditional methods for synthesising nitropheny-
lamines often require severe conditions and yields are
often unsatisfactory.22 An improved preparation for
2-nitrodiphenylamines was reported by Rees et al. in 1980.9
The substrates for this improved reaction involved the
use of o-fluoronitrobenzene and substituted aniline. The
starting materials were mixed in a ratio 1:2, respectively,
and were allowed to react at 180 °C for 48 h in the pres-
ence of 1 equiv of solid anhydrous potassium fluoride.
Improvements in these reactions were attributed to the
presence of KF, which was a particularly good base
when hydrogen fluoride was generated. In an attempt to
improve the reaction in ionic liquids still further,
we investigated the effect of the addition of F- and other
(13) Brunnet, J. F.; Zahler, R. E. Chem. ReV. 1951, 49, 273.
(14) Meisenheimer, J. Justus Liebigs Ann. Chem. 1902, 323, 205.
(15) For a review on rate and mechanism involving the formation of
Meisenheimer complexes see: Terrier, F. Chem. ReV. 1982, 62, 78.
(16) (a) Mancini, P. M. E.; Terenzani, A.; Adam, C.; Vottero, L. R. J.
Phys. Org. Chem. 1997, 10, 849. (b) Crampton, M.; Robotham, I. A. J.
Chem. Res., Synop. 1997, 22.
(17) Cox, B. G.; Parker, A. J. Am. Chem. Soc. 1973, 95, 408.
(18) First these ionic liquids were heated to 100 °C for 24 hours to ensure
that there was no evidence of decomposition.
(19) Crowhurst, L.; Lancaster, N. L.; Pe´rez Arlandis, J. M.; Welton, T.
J. Am. Chem. Soc. 2004, 126, 11549.
(20) Crowhurst, L.; Mawdsley, P. R.; Perez-Arlandis, J. M.; Salter, P.
A.; Welton, T. Phys. Chem. Chem. Phys. 2003, 5, 2790.
(21) Koppel, I. A.; Burk, P.; Koppel, I.; Leito, I.; Sonoda, T.; Mishima.
M. J. Am. Chem. Soc. 2000, 122, 5114.
(23) (a) John, H. G. J. Chem. Soc., Perkin Trans. 1 1988, 1331. (b)
Lindley, J. M.; McRobbie, I. M.; Cohn, O. M.; Suschitxky, H. Chem. Soc.,
Perkin Trans. 1 1977, 2194.
(22) Lewis, E. S.; Douglas, T. A.; McLaughlin, M. L. AdV. Chem. 1987,
215, 35.
Org. Lett., Vol. 9, No. 25, 2007
5249

bases to this reaction system. Owing to it excellent
performance above we selected the [C4C1pyrr][CH3SO3]
ionic liquid for this experiment. In all cases, the base was
added in 10% molar ratio relative to the starting materials.
We used a variety of F- sources (KF, CsF, Bu4NF), strong
neutral bases (1,8-diazabicyclo[5.4.0]undec-7-ene, 1,1,3,3-
tetramethylguanidine), Cs2CO3, tBuONa, and NaOH, but in
no case was the yield of the reaction greater than that for
the reaction without added base. This is presumably due to
the ionic-liquid anion being sufficiently basic to play this
role.
In conclusion, we have used calculated theoretical para-
meters, ∆GH for the conjugate acid of the anion of the ionic
liquid, to design an optimum ionic liquid to be used as a
solvent for a reaction, namely the nucleophilic aromatic
substitution of p-fluorobenzene and p-anisidine. This has
provided dramatic increases in yield above that possible to
achieve in the best available molecular solvents. It has also
provided reaction conditions that can be used without the
addition of a base. This is the first demonstration of ionic
liquids as truly “designer solvents”.
Acknowledgment. We would like to thank the Kodak
Foundation for a scholarship (J.M.P.A.) and Merck for the
kind donation of the trifluoroacetate ionic liquids.
Supporting Information Available: Detailed experi-
mental procedures and spectroscopic data (1H and 13C NMR,
MS). This material is available free of charge via the Internet
at http://pubs.acs.org.
OL702435F
5250
Org. Lett., Vol. 9, No. 25, 2007