Carbon-centered strong bases in phosphonium ionic liquids

Article abstract of DOI:10.1021/jo701289d

(Chemical Equation Presented) Phosphonium ionic liquids (PhosILs), most notably tetradecyl(trihexyl)phosphonium decanoate (PhosIL-C₉H ₁₉COO), are solvents for bases such as Grignard reagents, isocyanides, Wittig reagents (phosphoranes), and N-heterocyclic carbenes (NHCs). The stability of the organometallic species in PhosIL solution is anion dependent. Small bases, such as hydroxide, react with the phosphonium ions and promote C-H exchange as suggested by deuterium-labeling studies. A method to dry and purify the ionic liquids is described and this step is important for the successful use of basic reagents in PhosIL. NHCs have been generated in PhosIL, and these persistent solutions catalyze organic transformations such as the benzoin condensation and the Kumada-Corriu cross-coupling reaction. Phosphoranes were generated in PhosIL, and their reactivity with various organic reagents was also tested. Inter-ion contacts involving tetraalkylphosphonium ions have been assessed, and the crystal structure of [(n-C₄H₉) ₄P][CH₃CO₂·CH₃-CO ₂H] has been determined to aid the discussion. Decomposition of organometallic compounds may also proceed through electron-transfer processes that, inter alia, may lead to decomposition of the IL, and hence the electrochemistry of some representative phosphonium and imidazolium ions has been studied. A radical derived from the electrochemical reduction of an imidazolium ion has been characterized by electron paramagnetic resonance spectroscopy.

Full text of DOI:10.1021/jo701289d

Carbon-Centered Strong Bases in Phosphonium Ionic Liquids
Taramatee Ramnial,† Stephanie A. Taylor,† Marissa L. Bender,‡ Brian Gorodetsky,†
Peter T. K. Lee,‡ Diane A. Dickie,† Brett M. McCollum,† Cory C. Pye,‡
Charles J. Walsby,*,† and Jason A. C. Clyburne*,‡
Department of Chemistry, Simon Fraser UniVersity, Burnaby BC, V5A 1S6, Canada, and Department of
Chemistry, Saint Mary’s UniVersity, Halifax NS, B3H 3C3, Canada
jason.clyburne@smu.ca
ReceiVed June 28, 2007
Phosphonium ionic liquids (PhosILs), most notably tetradecyl(trihexyl)phosphonium decanoate (PhosIL-
C9H19COO), are solvents for bases such as Grignard reagents, isocyanides, Wittig reagents (phosphoranes),
and N-heterocyclic carbenes (NHCs). The stability of the organometallic species in PhosIL solution is
anion dependent. Small bases, such as hydroxide, react with the phosphonium ions and promote C-H
exchange as suggested by deuterium-labeling studies. A method to dry and purify the ionic liquids is
described and this step is important for the successful use of basic reagents in PhosIL. NHCs have been
generated in PhosIL, and these persistent solutions catalyze organic transformations such as the benzoin
condensation and the Kumada-Corriu cross-coupling reaction. Phosphoranes were generated in PhosIL,
and their reactivity with various organic reagents was also tested. Inter-ion contacts involving
tetraalkylphosphonium ions have been assessed, and the crystal structure of [(n-C4H9)4P][CH3CO2‚CH3-
CO2H] has been determined to aid the discussion. Decomposition of organometallic compounds may
also proceed through electron-transfer processes that, inter alia, may lead to decomposition of the IL,
and hence the electrochemistry of some representative phosphonium and imidazolium ions has been
studied. A radical derived from the electrochemical reduction of an imidazolium ion has been characterized
by electron paramagnetic resonance spectroscopy.
Introduction
One of the biggest industrial concerns is the replacement of
volatile organic compounds (VOCs),6 particularly those that are
toxic, such as CH2Cl2, and those that are highly volatile and
flammable, such as ethers. Simply substituting one solvent for
another that is more environmentally friendly is often impos-
sible, due to the multifaceted functions of the solvent. Some of
these functions include behavior as a heat sink for exothermic
processes, as well as promotion of homogenization and the
mixing of reactants.1 Of particular importance to this paper are
ethereal solvents, primarily diethyl ether or tetrahydrofuran,
Environmental pressure to reduce waste and reuse materials
has prompted studies into “Green Chemistry”1,2 and several
reviews have been written concerning this emerging field.3-5
† Simon Fraser University.
‡ Saint Mary’s University.
(1) Williamson, P. T.; Anastas, C. T. Green Chemistry: Designing
Chemistry for the EnVironment; ACS Symp. Ser. No. 626; American
Chemical Society: Washington, DC, 1996.
(2) Rogers, R. D.; Seddon, K. R. Ionic Liquids: Industrial Applications
to Green Chemistry; American Chemical Society: Washington, DC, 2002.
(3) Zhao, D.; Wu, M.; Kou, Y.; Min, E. Catal. Today 2002, 74, 157.
(4) Ren, X. R.; Brenner, A.; Wu, J. X.; Ou, W. Proc. Electrochem. Soc.
2002, 2002 (19), 165.
(5) Freemantle, M. Chem. Eng. News 2004, 82 (45), 44.
(6) Lancaster, M. Green Chemistry: An Introductory Text; Royal Society
of Chemistry: Cambridge, UK, 2002.
10.1021/jo701289d CCC: $40.75 © 2008 American Chemical Society
Published on Web 01/04/2008
J. Org. Chem. 2008, 73, 801-812
801

Ramnial et al.
SCHEME 1. Deprotonation of an Imidazolium Ion in an
ImIL
FIGURE 1. Organometallic complexes of NHCs.
which are commonly used for reactions involving strong bases.
At present, few neoteric or novel solvents with advantageous
properties have been studied as alternative reaction media for
reactions involving electron-rich or highly basic reagents.
The use of ionic liquids (ILs)7 to replace or reduce VOCs is
a move that could have significant positive environmental
impact. There would also be safety benefits resulting from the
preferential use of ILs over VOCs, largely due to their low or
nonflammability with normal use.5,8 The most extensively
studied class of ILs is based on imidazolium cations and this
diverse class is generally referred to as imidazolium based ionic
liquids (ImILs). These solvents exhibit high thermal stability,
low volatility, and variation in cosolvent miscibility.9 Some
processes using these liquids are currently being commercialized.
For example, ImILs can serve as useful materials for transporting
reactive gases such as BF3 and PH3,10 which form complexes
with the anionic component of the IL.
ILs are inherently two-component systems, namely an anion
and a cation. The disparate nature of cations and anions,
specifically, intimate sites of high electron deficiency and high
electron richness, suggests the ionic liquids may be ideal solvents
to enable chemistry that is perhaps not possible in normal
molecular solvents. For these two-component mixtures, each
constituent of the solvent should ideally be inert. Otherwise,
either of these components could become concomitantly in-
volved in side-reaction pathways. Reactivity studies have shown
that both the cationic and/or anionic component can engage in
solvolysis reactions.11,12 Specific relevant examples include
activation of the C-H bonds in imidazolium ions, as well as
hydrolysis of [BF4] or [PF6] salts to generate HF in situ.
Imidazolium-based ionic liquids are known to support many
reactions that proceed well in acidic reaction conditions,13-15
but the track record for ImILs as solvents for reactions involving
strong bases is certainly less than stellar.16-18 Indeed, recent
reviews highlighted the rather unpredictable behavior of this
solvent type.19 Imidazolium ions have been shown to react under
basic conditions to produce N-heterocyclic carbenes (NHCs)20
as shown in Scheme 1. NHCs themselves are very reactive,
highly basic, neutral six-electron species possessing a dicoor-
dinate carbon atom with two nonbonding electrons.21 They have
strong σ-donor and poor π-acceptor characteristics, and have
recently attracted attention in numerous applications in syn-
thetic22,23 and transition metal chemistry (Figure 1).24,25
Even though imidazolium salts are reactive as solvent media,
they have been used for many organometallic reactions26 such
as the preparation of new metal carbonyl complexes,27 indium
and tin catalysts,28,29 and zinc reagents.30,31 Deprotonation of
the cationic component of ImILs can be important in some
catalytic reactions32-34 (i.e., it generates an active metal/NHC
complex) but in other cases, such as in the Baylis-Hillman
reaction (Scheme 2), the deprotonation results in a significant
decrease in reaction yields.17,18 ILs have been demonstrated as
catalytically active in addition and cross-coupling reactions,
respectively.35,36
Recently it has been reported that ionic liquids can serve as
solvents for basic organometallic reagents, such as Grignard
reagents. Specifically, ImILs can withstand the strong-base
(15) (a) Imidazolium-based ionic liquids (ImILs), specifically tetrachlo-
roaluminate (III) salts, have been classified as acidic or basic depending
on the relative amount of Cl-/AlCl3 in the materialsthe anion [Al2Cl7]- is
a source of AlCl3 which is a good Lewis acid for catalysis. We use the
term “acidic” to emphasize the high reactivity of the C-H fragment in
ImILs, which, with strong bases, behave as protic acids. (b) Holbrey, J. D.;
Seddon, K. R. Clean Prod. Processes 1999, 1, 223. (c) Earle, M. J.; Seddon,
K. R. Pure Appl. Chem. 2000, 72, 1391.
(16) Freemantle, M. Chem. Eng. News 2005, 83 (33), 31.
(17) Aggarwal, V. K.; Emme, I.; Mereu, A. Chem. Commun. 2002, 1612.
(18) Dupont, J.; Spencer, J. Angew. Chem., Int. Ed. 2004, 43, 5296.
(19) See, for example: Chowdhury, S.; Mohan, R. S.; Scott, J. L.
Tetrahedron 2007, 63, 2363. Nair, V.; Bindu, S.; Sreekumar, V. Angew.
Chem., Int. Ed. 2004, 43, 5130.
(20) Ott, L. S.; Cline, M. L.; Deetlefs, M.; Seddon, K. R.; Finke, R. G.
J. Am. Chem. Soc. 2005, 127, 5758.
(7) Wilkes, J. S. Green Chem. 2002, 4, 73.
(21) Bourissou, D.; Guerret, O.; Gabba¨ı, F. P.; Bertrand, G. Chem. ReV.
(8) (a) It should be noted that under specific conditions, some ILs are,
in fact, combustible. However, the bulk properties of an IL are dramatically
different than, for example, fine mists which in many ways are comparable
to normal solvent vapors. (b) Smiglak, M.; Reichert, W. M.; Holbrey, J.
D.; Wilkes, J. S.; Sun, L.; Thrasher, J. S.; Kirichenko, K.; Singh, S.;
Katritzky, A. R.; Rogers, R. D. Chem. Commun. 2006, 2554.
(9) Welton, T. Chem. ReV. 1999, 99, 2071.
(10) Tempel, Daniel Joseph Henderson Philip Bruce Brzozowski, Jeffrey
Richard Pearlstein, Ronald Martin, U.S. Patent 2006060817, 2006.
(11) Proton or hydrogen atom abstraction reactions are common in
aqueous and nonaqueous media. In systems such as water these reactions
lead to generation of O2, H2, or OH-, and H3O+. In ILs, such reactions can
also occur, but these lead to decomposition and contamination of the IL.
We recognize, however, that these decompositions can often result in the
formation of the truly active species in select reactions.
(12) (a) There have been numerous reports dealing with the hydrolysis
of anions, particularly tetrafluoroborate and hexafluorophosphate. These
reactions do cause significant problems and perhaps most important is the
evolution of HF which can damage glass and metal. (b) Scammells, P. J.;
Scott, J. L.; Singer, R. D. Aust. J. Chem. 2005, 58, 155.
(13) Qiao, K.; Yokoyama, C. Chem. Lett. 2004, 33, 472.
(14) Stark, A.; MacLean, B. L.; Singer, R. D. J. Chem. Soc., Dalton
Trans. 1999, 63.
2000, 100, 39.
(22) Arduengo, A. J., III Acc. Chem. Res. 1999, 32, 913.
(23) Cowley, A. H. J. Organomet. Chem. 2001, 617, 105.
(24) Herrmann, W. A.; O¨ fele, K.; Elison, M.; Ku¨hn, F. E.; Roesky, P.
W. J. Organomet. Chem. 1994, 480, C7.
(25) Abernethy, C. D.; Codd, G. M.; Spicer, M. D.; Taylor, M. K. J.
Am. Chem. Soc. 2003, 125, 1128.
(26) Dupont, J.; de Souza, R. F.; Suarez, P. A. Z. Chem. ReV. 2002,
102, 3667.
(27) Schottenberger, H.; Wurst, K.; Horvath, U. E. I.; Cronje, S.;
Lukasser, J.; Polin, J.; McKenzie, J. M.; Raubenheimer, H. G. Dalton. Trans.
2003, 4275.
(28) Gordon, C. M.; McCluskey, A. Chem. Commun. 1999, 1431.
(29) Gordon, C. M.; Ritchie, C. Green Chem. 2002, 4, 124.
(30) Kitazume, T.; Kasai, K. Green Chem. 2001, 3, 30.
(31) Law, M. C.; Wong, K-Y.; Chan, T. H. Green Chem. 2004, 6, 241.
(32) Xu, L.; Chen, W.; Xiao, J. Organometallics 2000, 19, 1123.
(33) Mathews, C. J.; Smith, P. J.; Welton, T.; White, A. J. P.; Williams,
D. J. Organometallics 2001, 20, 3848.
(34) Hasan, M.; Kazhevnikov, I. V.; Rafiq, M.; Siddiqui, H.; Femoni,
C.; Steiner, A.; Winterton, N. Inorg. Chem. 2001, 40, 795.
(35) Bica, K.; Gaertner, P. Org. Lett. 2006, 8, 733.
(36) Co´rdoba, R.; Plumet, J. Tetrahedron Lett. 2003, 44, 6157.
802 J. Org. Chem., Vol. 73, No. 3, 2008

C-Centered Strong Bases in Phosphonium Ionic Liquids
SCHEME 2. (A) The Baylis-Hillman Reaction and (B) the
Major Side Reaction that Occurs in ImIL between the
Solvent and Benzaldehyde, Dramatically Decreasing the
Yield of the Reaction
FIGURE 2. PhosILs used in this study.
of strong bases (Figure 2), namely, tetradecyl(trihexyl)phos-
phonium chloride (PhosIL-Cl) and tetradecyl(trihexyl)phospho-
nium decanoate (PhosIL-C9H19COO), which was prepared from
PhosIL-Cl via an anion exchange reaction. PhosIL-C9H19COO
is somewhat more viscous than PhosIL-Cl, but overall, de-
canoate has much stronger ligand donor properties than chloride
for the stabilization of metal sites. This effect will be described
later.
Purification of Phosphonium-Based Ionic Liquids. It was
found that commercially available PhosIL-Cl contains traces
of residual phosphines (observed through NMR studies), HCl,
and water. Also present are trace phosphine oxides as well as
trace amounts of isobutylnitrile, derived from the radical addition
step that generates the phosphines. Recognizing the sensitivity
of organometallic reagents to these species, a method was
designed to purify the PhosIL.
Excess HCl in the PhosIL is neutralized with aqueous sodium
bicarbonate. Care should be exercised since there can be
excessive foaming during this step. The ionic liquid layer is
then washed vigorously with water and extracted using
hexanes. Phosphine oxides can be easily removed by diluting
the PhosIL with small amounts of hexanes and passing the
solution through a short silica column. For the organo-
metallic reactions described below, it was important to remove
all traces of water by azeotropic distillation with toluene. The
PhosIL can also be further dried at this stage with a small
amount of solid potassium, but this treatment is not generally
required. Isobutylnitrile is removed in the organic washes and
distillation steps.
As noted above, the PhosIL can be dried with elemental
potassium,45 which suggests that unlike ImILs, PhosIL does not
readily react with this source of the simplest base: the electron.
This observation portends well for the stability of other bases
and electron-rich molecules in the PhosIL. A word of caution
is warranted: reaction of water with elemental potassium yields
hydroxide ion, which does react with the PhosIL cation.
Stability of Grignard Reagents in Phosphonium-Based
Ionic Liquids. The stability of the PhosIL in the presence of
several different nucleophilic reagents was examined. Perhaps
the most readily available carbon-based nucleophiles are com-
mercial solutions of Grignard reagents in tetrahydrofuran (THF)
(pKa ) 45-50). As a representative example of this important
class of reagent, phenylmagnesium bromide (PhMgBr) was
chosen for examination of its behavior in PhosIL. Anhydrous
samples of PhosIL-Cl form clear solutions with low viscosity
when mixed with commercially available 1 M PhMgBr in
THF.42 The solutions are air and moisture sensitive, but can be
stored under an inert atmosphere such as argon or nitrogen. The
reactivity of fresh PhosIL-Cl/Grignard reagent solutions with a
THF:PhosIL-Cl ratio of 1:3 was tested by the addition of
anhydrous bromine to give exclusive formation of bromoben-
zene, PhBr. To probe the persistence of the Grignard reagent
conditions of the Grignard solutions when the acidic hydrogen
site at the C2 position on the imidazolium or imidazolinium ion
has been protected or eliminated. This has been accomplished
with a methyl,37 isopropyl,38 or phenyl39 substituent at the
vulnerable site. In most cases, the Grignard was added to the
ImIL, but was not generated in situ. The choice of anion varied
although two contained at least one triflimide unit, and thus
contained an oxygen-based donor site, a feature that will be
identified as key for the stability of Grignard solutions in IL
solution.38,39 The third, with the methyl-protecting group, em-
ployed a tetrafluoroborate anion that was successful in support-
ing the Grignard itself, and its generation, although only at
elevated temperatures. A pyridinium-based ionic liquid was also
used to engender a Grignard in solution within the same study.37
Studies on basic reagents in ionic liquids have been performed
and will be presented here. These reactions can be achieved by
using related phosphonium-based ionic liquids especially since
the phosphonium cations are well-known to be more thermally
robust than imidazolium cations.40,41 In this paper it is shown
that phosphonium ionic liquids (PhosILs) (1) can be used as
solvents for Grignard reactions,42,43 (2) can dissolve other
important carbon centered ligands; (3) can be used as solvents
for generating NHCs and for preparing their metal complexes,
and (4) can be used as a solvent medium for the generation of
Wittig reagents. Preliminary reports dealing with select topics
have been the focus of earlier communications.43,44 We note
that in addition to acid/base deprotonation reactions, electron-
transfer reactions can result in the decomposition of both
organometallic species and the ionic liquid solvent. To this end,
we have examined the electrochemistry of imidazolium and
phosphonium ions and we have characterized by EPR spec-
troscopy a radical derived from the one-electron reduction of
1,3-bis(2,6-diisopropylphenyl)imidazolium chloride.
Results and Discussion
Phosphonium-Based Ionic Liquids. Two phosphonium-
based ionic liquids were chosen for the study of the chemistry
(37) Law, M. C.; Wong, K. W.; Chan, T. H. Chem. Commun. 2006,
2457.
(38) Handy, S. T. J. Org. Chem. 2006, 71, 4659.
(39) Vaclav, J.; Wilhelm, R. Green Chem. 2005, 7, 844.
(40) Kim, Y. J.; Varma, R. S. J. Org. Chem. 2005, 7882.
(41) Bradaric, C. J.; Downard, A.; Kennedy, C.; Robertson, A. J.; Zhou,
Y. Green Chem. 2003, 5, 143.
(42) McNulty, J.; Capretta, A.; Cheekoori, S.; Clyburne, J. A. C.;
Robertson, A. J. Chim. Oggi Chem. Today 2004, 22, 13.
(43) Ramnial, T.; Ino, D. D.; Clyburne, J. A. C. Chem. Commun. 2005,
325.
(44) Ramnial, T.; Hauser, M. K.; Clyburne, J. A. C. Aust. J. Chem. 2006,
59, 298.
(45) Gorodetsky, B.; Ramnial, T.; Branda, N. R.; Clyburne, J. A. C.
Chem. Commun. 2004, 1972.
J. Org. Chem, Vol. 73, No. 3, 2008 803

Ramnial et al.
SCHEME 3. Reactions Performed in PhosIL-C9H19COO
without any Ethereal Cosolventa
FIGURE 3. A phosphorane, the anticipated product resulting from
deprotonation of a PhosIL cation.
in PhosIL, the tests were repeated on month-old solutions. When
the aged solution was treated with Br2, 5% of biphenyl was
detected along with bromobenzene. The presence of benzene
was not observed in either case, indicating that a deprotonation
reaction did not occur. Most importantly, deprotonation of the
PhosIL to produce a phosphorane was not observed, as shown
in Figure 3.
Complete removal of THF cosolvent from the PhosIL-Cl/
1
Grignard solutions, as determined by H NMR spectroscopic
a A few drops of either toluene or hexanes was added to reduce the
viscosity of the IL. Reaction conditions: (i) DMF, stirred for 3 h; (ii)
acetone, stirred for 3 h; (iii) benzaldehyde, stirred for 2 h; (iv) 2,6-dibromo-
4-methyl-iodobenzene, stirred for 3 h, quenched with Br2; (v) Br2, stirred
for 1 h; (vi) CuCl2, stirred for 2 h.
studies, led to the formation of biphenyl and a variety of
products that could be traced to the decomposition of PhosIL-
Cl, including tetradecyl(dihexyl)phosphine and hexene. These
products are formed by the reduction of the phosphonium ion
presumably via an electron-transfer reaction. Also, we note here
that similar preparations involving PhMgBr/THF solutions with
PhosIL dicyanamide or PhosIL bis(trifluoromethylsulfonyl)imide
resulted in immediate decomposition of the solution after mixing
the reagents. Since the cations and reaction conditions were
otherwise identical, this suggested that the anion of the ionic
liquid plays a central role in the stabilization of the Grignard
reagent (see below).
has been shown that magnesium metal can form a protective
surface coating, which impeded reactions when exposed to
PhosIL.48
It is interesting to note at this point that analogous reactions
using phenyllithium produced a variety of PhosIL decomposition
products including tetradecyl(trihexyl)phosphine and hexene, as
identified by GC-MS studies. A similar observation was made
by Handy when he noted the inability of imidazolium-based
ionic liquids to support butyllithium.38 The reasons for this
difference in reactivity between the Grignard reagent and
organolithium reagents are not obvious. For the Grignard
reagents, there may be an outer-sphere mechanism where the
Schlenk equilibrium49 might be favorably influenced by the large
excess of anions (strong donors) in PhosIL solution. Recently,
we reported that the mode of reactivity for Grignard reagents,
either as nucleophiles or as reducing agents, is strongly
influenced by the solvent, whether it is ionic (PhosIL) or
molecular (THF).50 Finally, organolithium reagents tend to
aggregate in solution and form clusters and we speculate that
organolithium reagents may act as strong reducing agents
through inner-sphere electron-transfer reactions, rather than the
organolithium reagent behaving as a discrete hydrocarbyl
transfer species.
Reactivity of Grignard Reagents in Phosphonium-Based
Ionic Liquids. To test the tolerance of PhosIL to strong bases,
a survey of Grignard reactions at room temperature was
performed. Reactions included addition to carbonyl compounds
(Scheme 3, i-iii), benzyne reactions (iv), halogenation (v), and
coupling reactions (vi) in PhosIL-C9H19COO. The most interest-
ing result came from the benzyne reaction (iv) because it
suggests that the postulated benzyne intermediate is feasible in
the PhosIL. GC-MS studies show that the benzyne intermediate
does not abstract a hydrogen atom or proton from the PhosIL
to form 3-methylbiphenyl. An added benefit for this system is
the high heat capacity of PhosIL; therefore it is not necessary
To test this hypothesis, solutions Grignard solutions in PhosIL
devoid of ether were prepared and the absence of ether was
indicated by spectroscopic studies. After adding a few drops of
THF or toluene to tetradecyl(trihexyl)phosphonium decanoate
to lower its freezing point and decrease viscosity, the ionic liquid
was cooled to -78 °C and PhMgBr in THF was added to make
a 1 M solution. The new PhosIL-C9H19COO/Grignard reagent
solutions were allowed to warm to room temperature and stirred
for 15 min. The color changed from pale yellow to orange-red
with an intermediate faint blue color. This unusual blue color
has previously been observed when two neighboring carbonyl
groups of benzil make a stable intermediate radical anion with
a Grignard reagent,46 but no significant effect of the trace
transient blue species was observed for the reactivity of the
solution. THF or toluene was completely removed in vacuo to
yield an ether-free Grignard solution, and this new solution
showed no sign of decomposition over several months. It is
proposed that the stability of these new solutions is the result
of coordination between the magnesium atom of the Grignard
reagent and the oxygen donor site of the decanoate anion46 of
PhosIL-C9H19COO, which replaces a similar metal-oxygen
interaction when using ethereal solvent.47
All attempts to generate Grignard reagents in PhosIL failed,
for reasons not clear at this time. It is speculated that it may be
because the oxidative addition reaction of the magnesium metal
into the carbon halide bond is slow, or an additional cosolvent
must be added to the reaction mixture to facilitate this reaction.
Recent reports, however, have shown that the oxidative addition
step does proceed in pyridinium ionic liquids.37 Adjunctly, it
(48) Birbilis, N.; Howlett, P. C.; MacFarlane, D. R.; Forsyth, M. Surf.
Coat. Technol. 2007, 201, 4496.
(49) Elschenbroich, C.; Salzer, A. Organometallics: A Concise Introduc-
tion; VCH: Weinheim, Germany, 1992.
(46) Maruyama, K.; Katagiri, T. J. Am. Chem. Soc. 1986, 108, 6263.
(47) Jure`´ık, V.; Wilhelm, R. Green Chem. 2005, 7, 844.
(50) Ramnial, T.; Taylor, S. A.; Clyburne, J. A. C.; Walsby, C. J. Chem.
Commun. 2007, 2066.
804 J. Org. Chem., Vol. 73, No. 3, 2008

C-Centered Strong Bases in Phosphonium Ionic Liquids
SCHEME 4. Halogen/Magnesium Exchange Reaction
between Ethylmagnesium Bromide and Bromobenzene in
PhosIL-C9H19COO
FIGURE 4. Calculated lowest energy structures for the reaction of
diphenylmagnesium/acetate complex [Ph2Mg-O2CCH3]. The net en-
ergy released for this reaction is 303 kJ/mol.
to cool the reaction solutions to the low temperatures often
required for ethereal solutions.
The addition of water and hexanes to the reaction mixtures,
as described above, results in the formation of a three-phase
system, with the organic layer on top, ionic liquid in the middle,
and the aqueous layer on the bottom. The products were isolated
from the organic layer and analyzed by GC-MS. In some cases,
the low yields reported in Scheme 3 reflect partitioning between
the ionic liquid and the organic phase. Isolated yields can be
markedly improved by successive extractions. Alternatively, in
some cases, due to the high thermal stability of the PhosIL and
the volatility of the products, distillation could be used to remove
the products from the reaction mixtures. In all cases, PhosIL-
C9H19COO can be washed with water and hexanes, dried, and
re-used, demonstrating the potential of these solvents to
participate in recyclable, “green” processes.
Halogen/magnesium exchange reactions (Scheme 4) were also
carried out in PhosIL-C9H19COO. For example, ethylmagnesium
bromide in THF was added to cold PhosIL-C9H19COO and the
THF was removed under vacuum. Bromobenzene was added
and the solution was stirred for 6 h. To probe the Mg/Br
exchange reaction, benzaldehyde was then added and the
mixture was stirred for an additional 16 h. Following an aqueous
quench and extraction with dichloromethane, diphenylmethanol
was obtained in 83% yield. A very small amount of 1-phenyl-
propan-1-ol (8%) was also obtained due to incomplete halogen/
magnesium exchange reactions.
studies, evidence was obtained that suggested that the Schlenk
equilibrium was strongly shifted so that in ionic liquid solvent
Ph2Mg was the dominant species, whereas in molecular solvent
PhMgBr was present in significant amounts.50 With these results
in hand we decided to probe computationally the complexation
of Ph2Mg with carboxylate anions to appreciate more fully the
structure of the apparently dominant species in solutions of
Grignard reagents in PhosIL.
Calculations on the interactions of arylmagnesium species
with acetate ions (as a model for decanoate) were performed.
Computations were carried out on the individual components
of the reaction mixtures, specifically diphenylmagnesium,
acetate, phenylmagnesium bromide, and related complexes. The
computational studies indicate that diphenylmagnesium prefers
the twisted D2d form, although it should be noted that the barrier
to rotation about the C-Mg-C axis to give the D2h transition
structure is less than 3 kJ/mol, and is clearly likely to occur at
room temperature. The calculated Mg-C bond length is 2.098
Å. The barrier to C-C rotation in the acetate ion (AcO-) is
much smaller, 0.05 kJ/mol, and neither of the two structures is
clearly preferred. The C-O distance was calculated to be 1.234
Å. The barrier to phenyl rotation in the phenylmagnesium
bromide-acetate complex is estimated to be 10 kJ/mol. For
details of calculations consult the Experimental Section.
Halogen/magnesium exchange reactions appear to occur
readily in PhosIL. A recent report on highly efficient reagents
for Br/Mg exchange sheds some light on this kind of reaction.51
It now appears likely that the highly polar (i.e., discrete ions)
PhosIL media may stabilize polar “ate” complexes and, hence,
also facilitate both the metal exchange reaction and increase
the reactivity of the PhosIL/Grignard reagent solutions. The
formation of the “ate” complex, although speculative, is
consistent with both the Lewis acidity of diorganomagnesium
species, reactivity studies, and the computational studies reported
in the following section.
A Computational Study on the Complexation of Aryl-
magnesium Species with RCOO-. As we reported recently,
the reactivity of Grignard reagents in molecular and ionic liquids
solvents is dramatically different. In molecular solvents Grignard
reagents react dominantly as nucleophiles, whereas in ionic
solvents electron-transfer reactions are preferred.50 Recall that
Grignard solutions are complex mixtures, with ArMgX, Ar2-
Mg, and MgX2 all possible species in solution. During our recent
The binding of the first acetate to diphenylmagnesium is
highly exothermic, at ca. ∆H ) -303 kJ/mol. The structure of
this complex, shown in Figure 4, featuring a bidentate acetate
ion, exhibits two Mg‚‚‚O interactions (2.120 Å). This evidently
makes the magnesium metal formally tetracoordinate, as perhaps
is anticipated for a magnesium center. Note, also, that the second
acetate binds endothermically with a value estimated to be ∆H
) 120 kJ/mol.
We propose that the strong interaction between the Mg center
and the anion of the ionic liquid may be required for the stability
of the solution and hence may suggest a reason for the apparent
decomposition of the PhosIL/Grignard solution where the
counterion is bis(trifluoromethylsulfonyl)imide or chloride.
These anions are notoriously nonbasic and a weakly coordinating
anion and, hence, may not stabilize Ph2Mg in ionic liquid
solution.
PhosIL and Neutral Bases of Interest to Organometallic
Chemistry. Some of the most basic neutral ligands are the
carbenes with pKa values in the range of 22 to 24.52,53 Highly
basic solutions containing NHCs dissolved in PhosIL can be
prepared by mixing the NHC with the PhosIL in the presence
(51) Krasovskiy, A.; Straub, B. F.; Knochel, P. Angew. Chem., Int. Ed.
2006, 45, 159.
J. Org. Chem, Vol. 73, No. 3, 2008 805

Ramnial et al.
SCHEME 7. Kumada-Corriu Cross-Coupling Reaction in
SCHEME 5. Preparation and Reactivity of the Carbene
1,3-Bis(2,4,6-trimethylphenyl)imidazol-2-ylidene (NHC 1),
Synthesized in PhosIL-Cl
PhosIL-C9H19COO
TABLE 1. Coupling of Aryl-Grignard Reagents with Aryl Halides
in the Presence of a Ni(0) Complex of NHC 1 in PhosIL-C9H19COO
aryl halide
MePh-Ph (%)
Me2Ph2 (%)
4-fluorotoluene
4-chlorotoluene
4-bromotoluene
4-iodotoluene
42
88
73
74
0
0
25
22
SCHEME 6. Benzoin Condensation of Benzaldehyde
Catalyzed by NHC 1 in PhosIL-Cl
SCHEME 8. Reaction of the Wittig Reagent with Carbonyl
Compounds To Form an Alkene
of several drops of benzene, toluene, or THF. The addition of
the cosolvent facilitates dissolution of the NHC, after which
the cosolvent can be removed under vacuum with no observable
effect on the stability of the resulting solution. Other strong
neutral Lewis bases, such as triphenylphosphine and tert-
butylisocyanide, were examined and were found by IR spec-
troscopic studies to be persistent in PhosIL.
Although difficulties were encountered during the attempted
synthesis of Grignard reagents in PhosIL, the carbene 1,3-bis-
(2,4,6-trimethylphenyl)imidazol-2-ylidene (NHC 1) was readily
synthesized by two different routes in PhosIL: (1) by reduction
of 1,3-bis(2,4,6-trimethylphenyl)imidazolium chloride using
metallic potassium45 and (2) by the treatment of the aforemen-
tioned imidazolium chloride with PhMgBr dissolved in a PhosIL
(Scheme 5). Imidazolium ions labeled at the carbeneic carbons
with 13C were used to provide unambiguous evidence for the
formation of NHC 1 by the observation of the 13C NMR signal
for the carbeneic carbon at 216 ppm. Further evidence was
supplied by reactivity studies (see below). The PhosIL/NHC
solutions are viscous and light brown in color. Surprisingly,
these basic solutions are stable in excess of 1 month and are
active for organic transformations. For example, NHC 1 in
PhosIL catalyzes the benzoin condensation of benzaldehyde in
40% yield as shown in Scheme 6.54
The stability of the NHC in PhosIL was tested by surveying
some well-established NHC chemistry with the p-block55 and
the d-block elements (Scheme 5). The products of these reactions
were characterized in PhosIL-Cl using analytical techniques such
as NMR, IR, and GC/MS. The NHC 1 solutions prepared in
PhosIL-Cl behave as normal NHC solutions. Treatment of NHC
1 with S8 produces 1,3-bis(2,4,6-trimethylphenyl)imidazol-2-
thione as indicated by 13C NMR spectroscopy and mass
spectrometry (m + 1 ) 336.3 amu).56 NHCs coordinated to
transition metal sites have attracted interest in catalysis, so the
reactivity of NHC 1 in PhosIL-Cl was illustrated by reacting
the solution with Cr(CO)6. Displacement of one carbonyl occurs
to afford IMesCr(CO)5 as identified by IR spectroscopy.
The Kumuda-Corriu reaction was examined to illustrate the
ability of Grignard reagents in PhosIL-C9H19COO to support
carbene reactions. These reactions involve the coupling of aryl-
Grignard reagents with aryl halides in the presence of a Ni(0)
complex of NHC 1, as shown in Scheme 7.57 The coupling
reaction proceeds with high yield for 4-bromotoluene and
4-iodotoluene, although halogen/magnesium exchange reactions
also occur with concomitant generation of ca. 25% 4,4′-
dimethylbiphenyl as shown in Table 1. The reaction works best
for chlorotoluene, and surprisingly even facilitates C-F bond
activation in modest yield. The activation of C-F bonds is of
great current interest as they are perhaps one of the most
challenging bonds to activate.58
Generation and Use of Wittig Reagents in PhosIL.
Synthetically, one of the most valuable classes of carbon-based
nucleophiles are the phosphoranes, also known as Wittig
reagents.59 These molecules react readily with aldehydes and
ketones to produce C-C double bonds (Scheme 8), from which
other important reactions proceed. Wittig reagents, ranging from
weakly basic, “stabilized” ylides (pKa of the conjugate acid ca.
8-11) to highly basic derivatives (pKa of [Ph3P-CH3] ca. 22.5
in DMSO), are available, but their generation in ILs has not
been reported.60
(56) Arduengo, A. J., III; Calabrese, J. C.; Cowley, A. H.; Rasika Dias,
H. V.; Goerlich, J. R.; Marshall, W. J.; Riegel, B. Inorg. Chem. 1997, 36,
2151.
(57) Bo¨hm, V. P. W.; Gsto¨ttmaryr, C. W. K.; Weskamp, T.; Herrmann,
W. A. Angew. Chem., Int. Ed. 2001, 40, 3387.
(58) Terao, J.; Ikumi, A.; Kuniyasu, H.; Kanbe, N. J. Am. Chem. Soc.
2003, 125, 5646.
(59) Smith, M. B.; March, J. AdVanced Organic Chemistry: Reactions,
Mechanism, and Structure; John Wiley & Sons, Inc.: New York, 2001.
(60) For chemistry of stabilized Wittig reagents in ILs see: Le Boulaire,
V.; Gree, R. Chem. Commun 2000, 2195-2196.
(52) Alder, R. W.; Allen, P. R.; Williams, S. J. J. Chem. Soc., Chem.
Commun. 1995, 1267.
(53) Kim, Y.-J.; Streitwieser, A. J. Am. Chem. Soc. 2002, 124, 5757.
(54) Enders, D.; Kallfass, U. Angew. Chem., Int. Ed. 2002, 41, 1743.
(55) Carmalt, C. J.; Cowley, A. H. AdV. Inorg. Chem. 2000, 50, 1.
806 J. Org. Chem., Vol. 73, No. 3, 2008

C-Centered Strong Bases in Phosphonium Ionic Liquids
TABLE 2. Wittig Reactions in PhosIL-C9H19COOa
PhosIL and base
Wittig precursor
RR′CdO
alkene
yield (%)
PhosIL-C9H19COO/PhMgBr
PhosIL-C9H19COO + KOBut
PhosIL-C9H19COO/PhMgBr
PhosIL-C9H19COO + KOBut
[Ph3PCH2CH3][Br]
[Ph3PCH2CH3][Br]
[Ph3PCH3][Br]
benzaldehyde
benzaldehyde
benzaldehyde
benzaldehyde
PhCHdCHCH3
PhCHdCHCH3
PhCHdCH2
96
88
86
94
[Ph3PCH3][Br]
PhCHdCH2
a The yields reported for PhCHdCHCH3 include both E/Z isomers.
FIGURE 5. Possible interactions between a phosphonium ion and a
carbonyl compound: (A) interaction between the quaternary phospho-
nium ion and oxygen of the carbonyl and (B) hydrogen bonding.
Generation of Wittig reagents was found to be feasible in
PhosIL without observable decomposition of the solvent. For
example, [Ph3PCH2CH3]+[Br]- can be deprotonated by PhosIL-
C9H19COO/PhMgBr solutions or other bases such as potassium
tert-butoxide suspended in PhosIL-C9H19COO to form the
nonstabilized phosphorane Ph3PdCHCH3. The product was
identified by the appearance of a distinctive 31P{1H} resonance
at 15 ppm.61 A similar resonance is observed for samples of
pure Ph3PdCHCH3 dissolVed in PhosIL-C9H19COO. The result-
ing ylide is synthetically useful, and was used in Wittig reactions
with aldehydes and ketones to generate an alkene as shown in
Table 2. The byproduct of Wittig reactions is triphenylphosphine
oxide, which was obtained as a white residue after the reaction,
and was characterized by mass spectrometry (M+ ) 278 amu).
Interaction of PhosIL with Basic Reagents. Phosphonium
ions have been used as catalysts in several reaction types62,63
including phase transfer catalysis,64,65 and quarternary phos-
phonium cations have been reported to activate carbonyl
compounds.66 It seems probable that this catalytic activity arises
from interactions between the cation and basic substrates. To
probe this interaction, some simple experiments were performed.
There are two probable types of interaction between the solute
(carbonyl compound) and the solvent (phosphonium cation)
(Figure 5), namely coordination to the quaternary cationic
phosphorus site through a [P‚‚‚O] interaction (Mode A) or
through hydrogen bonding to the hydrogen R to the phospho-
nium cation [C-H‚‚‚O] (Mode B).
FIGURE 6. X-ray structure of tetrabutyl phosphonium acetate
monoacetic acid. Selected bond lengths (Å) and angles (deg): O(1)-
C(20) 1.229(2), O(2)-C(20) 1.286(3), O(3)-C(30) 1.287(2), O(4)-
C(30) 1.235(2), O(2)‚‚‚H(2C) 1.2245(14), O(2)‚‚‚O(2) 2.449(3),
O(3)‚‚‚H(3C) 1.2256(14), O(3)‚‚‚O(3) 2.451(3), O(1)‚‚‚H(5B) 2.388,
O(1)‚‚‚C(5) 3.377, O(4)‚‚‚H(1A) 2.452, O(4)‚‚‚C(1) 3.437;
O(2)‚‚‚H(2C)‚‚‚O(2) 180.00(15), O(3)‚‚‚H(3C)‚‚‚O(3) 180.00(15), C(5)-
H(5B)‚‚‚O(1) 178.25, C(1)-H(1A)‚‚‚O(4) 173.19.
the ionic solvent and solute molecules, NMR titration experi-
ments were performed. It was found that the addition of carbonyl
compounds, such as benzaldehyde or propionaldehyde, to
PhosIL-Cl had little effect on the 31P NMR spectrum. This result,
although not conclusive, suggests that the local coordination
sphere of the phosphorus site has not been affected. Note that
the interaction would be dynamic, but even in dynamic systems,
such a change should be observable even as a small movement
of the 31P NMR resonance.
More interestingly, 1H NMR studies indicate a downfield shift
of the proton R to the phosphorus with increased concentration
of the carbonyl compound. This observation suggests that the
mode of association between solute and solvent likely involves
a hydrogen bonding interaction, consistent with Mode B
coordination. The shortest reported [cation‚‚‚anion] contact for
a phosphonium ion, as determined by a survey of the Cambridge
crystallographic database, was a [C-H‚‚‚O] hydrogen bond
(1.935(5) Å) reported for the simple phenoxide salt [Ph3PCH3]-
[2,6-Ph2C6H3O].67 The aforementioned salt is not a good model
for the systems under investigation, due to the differences
between arylphosphonium and alkylphosphonium cations, and
hence a crystalline sample of a salt composed of tetrabutylphos-
phonium acetate monoacetic acid was produced, as shown in
Figure 6.
The existence of [P‚‚‚O] coordination would be consistent
with the well-known strength of the P-O bond but formation
of this interaction would be accompanied by a reorganization
of the P-coordination sphere to a crowded, distorted, trigonal
bipyramidal geometry, which should cause significant changes
to the 31P NMR spectrum. To probe the association between
(61) Schmidbaur, H.; Buchner, W.; Scheutzow, D. Chem. Ber. 1973,
106, 1251.
(62) Kaufmann, D. E.; Nouroozian, M.; Henze, H. Synlett 1996, 11, 1091.
(63) Gerritsma, D. A.; Robertson, A.; McNulty, J.; Capretta, A.
Tetrahedron Lett. 2004, 45, 7629.
(64) de Bellefon, C.; Pollet, E.; Grenuillet, P. J. Mol. Catal. A: Chem.
1999, 145, 121.
(65) Landini, D.; Maia, A.; Montanari, F. J. Chem. Soc., Chem. Commun.
1977, 4, 112.
X-ray crystallographic studies revealed the expected tetra-
hedral phosphonium cation. The anion, however, is more
interesting: the acetate and acetic acid fragments are indistin-
guishable, with a proton present at 50% occupancy on both O(2)
and O(3). This proton is located equidistant between the
(66) McNulty, J.; Dyck, J.; Larichev, V.; Capretta, A.; Roberston, A. J.
Lett. Org. Chem. 2004, 1, 137.
(67) Davidson, M. G.; Goeta, A. E.; Howard, J. A. K.; Lamb, S.; Mason,
S. A. New J. Chem. 2000, 24, 477.
J. Org. Chem, Vol. 73, No. 3, 2008 807

Ramnial et al.
FIGURE 7. Three-dimensional network formed by [O-H‚‚‚O] and [C-H‚‚‚O] hydrogen bonds, viewed along the crystallographic b and c axes,
respectively. For clarity, only the R carbon and hydrogens of each butyl are shown.
symmetry related C-O oxygen atoms, forming a strong68 O-H‚
‚‚O hydrogen bond as shown in Figure 6. The CdO oxygen
atoms each form a weak [C-H‚‚‚O] hydrogen bond to the
hydrogen R to the phosphonium cation, consistent with a Mode
B interaction, as described above. Upon examination of the
extended structure (Figure 7), it became apparent that all four
alkyl chains are involved in this type of hydrogen bonding. This
leads to the formation of an extensive three-dimensional,
hydrogen-bonded network.
Deprotonation and Proton Exchange: Occurrence and
FIGURE 8. Space-filling diagram of 1,3-bis(2,4,6-trimethylphenyl)-
imidazolium ion (left) and tetradecyl(trihexyl)phosphonium ion (right).
Reasoning. Two, often competing, reactions can result in the
decomposition of solvents, and of importance to this paper, to
the decomposition of ions. These reactions are deprotonation
reactions and electron-transfer reactions. It was initially thought
that the hydrogen sites in the PhosIL were resistant to depro-
tonation, and a simple test was performed to examine this
hypothesis. A solution of the PhosIL was prepared by dissolving
a few drops of 50/50 D2O in ethanol-d containing a small
amount of NaOD. Over a period of days, the 1H signal for the
R-CH2 protons (ca. 2.3 ppm) decreased relative to the nonex-
changeable protons on the PhosIL. This indicates that the
resistance to deprotonation of the PhosIL is couched in terms
other than the acidity of the CH2 groups. This experiment clearly
indicates that the proton on the phosphonium ion, although
crowded, and contrary to our first expectation,43 is indeed
accessible by small bases.
There also appear to be steric reasons why deprotonation
reactions occur more readily in ImILs than in PhosIL. The
imidazolium ring is rigid whereas the alkyl chains on the
phosphonium ions are flexible and thus provide more protection
to the reactive proton. As shown in the space-filling diagram
of 1,3-bis(2,4,6-trimethylphenyl)imidazolium ion (Figure 8, left)
and tetradecyl(trihexyl)phosphonium ion (Figure 8, right), it is
difficult to sterically shield the carbeneic site in the imidazolium
ion, whereas in the tetradecyl(trihexyl)phosphonium ion there
is considerable steric congestion and flexibility and hence the
access to the reactive C-H site is diminished, although possible
for small bases as observed for hydroxide.
Highlighted (*) are the reactive C-H fragments of the ions.
the IL, it is largely anion dependent. For example, cosolvent-
free solutions of Grignard reagents dissolved in ILs are persistent
with O-donor anions, such as decanoate, whereas chloride and
bistriflamide appear to be insufficiently strong enough bases to
stabilize the metal-containing species in solution. Although it
would be reasonable to expect that deprotonation of a phos-
phonium ion to produce phosphorane and a salt would be
thermodynamically favored, evidence of this reaction was not
observed with these bases. In the case of poor donor anions,
such as chloride, when the cosolvent THF is stripped away, an
electron-transfer reaction occurs producing phosphines and
alkenes as products rather than the phosphorane. In IL solution
where the Grignard reagent is persistent, such as in a decanoate
solution, a bulky complex is formed between Ph2Mg and
decanoate, an ion that possesses substantial steric bulk. This
anion may not have access to the reactive protic site on PhosIL
and hence bases such as Grignard reagents dissolved in
PhosIL fail to react with the phosphonium component. Further
support for this kinetic argument is provided by noting that
[Ph3PCH2CH3]+[Br]- is deprotonated to form a phosphorane
by PhosIL-C9H19COO/PhMgBr solutions or other bases such
as potassium tert-butoxide, as shown by 31P{1H} NMR studies.
These solutions exhibit a single 31P{1H} resonance at 15 ppm,
consistent with the presence of Ph3PdCH(CH3),69 and also
consistent with a known sample of the phosphorane dissolved
in the phosphonium-based ionic liquid. In other words, for
phosphonium ions, small bases are susceptible to reacting with
The inertness of PhosIL toward reactions of bases therefore
appears to have primarily a kinetic basis and for Grignards in
(69) Albright, T. A.; Gordon, M. D.; Freeman, W. J.; Schweizer, E. E.
J. Am. Chem. Soc. 1976, 98, 6249.
(68) Steiner, T. Angew. Chem., Int. Ed. 2002, 41, 48.
808 J. Org. Chem., Vol. 73, No. 3, 2008

C-Centered Strong Bases in Phosphonium Ionic Liquids
TABLE 3. Reduction Potential of 1.5 mM Imidazolium Ions and
Phosphonium Ions Measured by Using a Platinum Working
Electrode in DMSO (Scan Rates ) 300 mV s-1), Pt Coil Counter
Electrode, and Ag/AgCl Reference Electrode
reduction
salt
potential (V)
1,3-bis(2,4,6-trimethylphenyl)imidazolium chloride
1,3-bis(2,6-diisopropylphenyl)imidazolium chloride
tetraethylphosphonium bromide
tetradecyl(trihexyl)phosphonium decanoate
(PhosIL-C9H19COO)
-2.23
-2.28
-2.06
-2.82
tetradecyl(trihexyl)phosphonium chloride
(PhosIL-Cl)
tetra-n-octylphosphonium bromide
-2.93
-3.70
FIGURE 9. Cyclic voltammogram of 1,3-bis(2,6-diisopropylphenyl)-
imidazolium chloride (1.66 mM) referenced to ferrocene/ferrocenium
redox couple (internal standard) taken to be 0.475 vs SCE.
FIGURE 10. Experimental (bottom) and simulated (top) EPR spectra
of 1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene neutral radical
intermediate produced by reductive electrolysis: -2.3 V vs SCE of
1.66 mg/mL solution of 1,3-bis(2,6-diisopropylphenyl)imidazolium
chloride in acetonitrile/0.1 M tetrabutylammonium hexafluorophosphate.
Experimental spectral parameters: ν ) 9.71 GHz, scan time ) 21 s,
time constant ) 5.12 ms, modulation amplitude ) 0.05 mT, receiver
gain ) 2 × 105, average of 50 scans. Simulation parameters: a(14N ×
2) ) 0.395 mT, a(1H) ) 0.800 mT, a(1H) ) 0.135 mT, a(1H) ) 0.0923
mT, line width ) 0.05 mT.
the PhosIL through deprotonation reactions whereas large bases
are resistant to this reaction.
Thus, the inertness of the phosphonium cations in the PhosIL-
Grignard solutions appears to have manifold reasons/causes.
First of all, and most obvious is the bulky and flexible nature
of the cation, which limits approach from external bases to the
eight acidic hydrogen sites. Second, the bases, namely a
decanoate complex of Ph2Mg, is itself bulky and has difficulty
approaching the R-hydrogen atoms of the phosphonium cation.
Furthermore, the phosphonium ions are electron precise, i.e.,
eight electrons in four σ bonds, and hence are more electro-
chemically robust than the unsaturated ions, specifically imi-
dazolium ions. From a steric (kinetic) point of view, the acidic
C2 hydrogen is chemically more accessible than the phospho-
nium C-H. It also engages in electrochemical reactions that
appear to be deprotonation reactions. These three features, the
bulky tetrahedral cation, the bulky anion, and its electrochemical
robustness, all collude to minimize detrimental side reactions
from occurring in basic PhosIL solutions.
Reported here is the comparison of the reduction potentials of
imidazolium salts and phosphonium salts with different lengths
of alkyl chains. The cyclic voltammograms (CVs) using standard
concentrations (1.5 mM) of the imidazolium and phosphonium
salts in anhydrous and anoxic DMSO were carried out versus
a standard calomel electrode (SCE) at 300 mV s-1 with
ferrocene as internal reference.
It was observed that the reduction of imidazolium salts
generally occurs at a less negative potential than that of
phosphonium salts and phosphonium salts possessing longer
alkyl chains are harder to reduce. For example, tetra-n-
octylphosphonium bromide has a more negative potential (Epred
) -3.70 V) compared to tetradecyl(trihexyl)phosphonium
chloride (Epred ) -2.93 V) as shown in Table 3. Hence, since
phosphonium ions have a higher reduction potential compared
to ImILs, reduction of imidazolium salts using potassium metal
can be carried out in PhosIL without decomposition of PhosIL.
It is important to note that there is no reaction between the
PhosIL and elemental potassium.
Electrochemical Comparison of ImILs and Phosphonium
Ions. As shown by reactivity studies, phosphonium ionic liquids
are more robust than ImILs toward bases and certain reducing
agents and the difference in reactivity can be explained by the
reduction potential of the ILs. Cyclic voltammetry of imidazo-
lium and phosphonium salts has been studied before.70,71
(70) Brooks, C. A.; Doherty, A. P. J. Phys. Chem. B 2005, 109, 6276.
(71) Buzzeo, M. C.; Evans, R. G.; Compton, R. G. ChemPhysChem 2004,
5, 1106.
J. Org. Chem, Vol. 73, No. 3, 2008 809

Ramnial et al.
a(1H) ) 0.800, a(1H) ) 0.093, and a(1H) ) 0.137. The 14N
1
couplings are from the ring nitrogen atoms and the largest H
coupling is assigned to the proton on C2, the primary site of
unpaired electron density. The two smaller proton couplings
are assigned to the allylic protons on C4 and C5. Their distinct
hyperfine couplings are consistent with nonplanarity of the ring
due in large part to pyramidalization at the primary radical
center. Overall, these observations show that the bulk of the
electron density is positioned on the C2-carbeneic carbon with
delocalization onto the neighboring nitrogen heteroatoms and
minor density on the backbone of the imidazole, thus consistent
with the presence of the 1,3-bis(2,6-diisopropylphenyl)imida-
zolyl radical. Comparable results were obtained in a previous
study where similar carbene reagents were reacted with muo-
nium, µ, a lighter “isotope” of hydrogen, to generate the neutral
radical.72 As shown in Figure 11, further resolution of the main
EPR lines is possible. Simulation of this hyperfine structure
shows couplings to 4 protons with a(1H) ) 0.015 mT and 2
with a(1H) ) 0.021 mT consistent with spin delocalization onto
the phenyl rings of the diisopropylphenyl groups, and on the
basis of the number of each type of proton these are assigned
to the meta and para positions, respectively. The quality of the
EPR data allow us to say with certainty that the species initially
produced from the electrochemical reaction is the 1,3-diarylimi-
dazolyl radical. We note that Nehta previously reported the
formation of 1,3-dialkylimidazolyl radicals during the pulse
radiolysis of ionic liquids and detected them by transient
absorption spectroscopy (UV).73
Conclusions
FIGURE 11. Close up of the central triplet in the EPR spectrum of
1,3-bis(2,6-diisopropylphenyl)imidazol-2-ylidene neutral radical inter-
mediate showing coupling with the phenyl hydrogens. Experimental
spectral parameters: ν ) 9.71 GHz, scan time ) 10.5 s, time constant
) 5.12 ms, modulation amplitude ) 0.01 mT, receiver gain ) 2 ×
105, average of 100 scans. Simulation parameters: a(1H) ) 0.135 mT,
a(1H) ) 0.0923 mT, a(1H × 2) ) 0.021 mT, a(1H × 4) ) 0.015 mT,
line width ) 0.015 mT.
In this report it has been shown that some strongly basic
carbon-centered reagents are compatible with PhosILs. Ether-
free solutions of Grignard reagents can be prepared in PhosIL
and they are stable for over 1 month at room temperature under
an inert atmosphere with no observable decomposition. These
Grignard solutions in PhosIL behave as normal Grignard
reagents, i.e., undergo addition reaction to carbonyls. As well,
1,3-bis(2,4,6-trimethylphenyl)imidazolium chloride, which has
a less negative potential than PhosIL, can be synthesized by
reduction as well as deprotonation in PhosIL, and these solutions
are persistent. The carbenes dissolved in the PhosIL can be used
to prepare complexes of both main group and transition metals.
Wittig reagents were also synthesized in the PhosIL and reacted
with carbonyl compounds to make alkenes without any decom-
position of the ionic liquid. These observations open up the use
of PhosIL as a reliable reaction media for a wide variety of
basic reagents. The problems associated with C-H activation
in imidazolium ions by highly reactive bases are not observed
for the PhosIL. Finally, a radical derived from the reduction of
an imidazolium ion was detected and characterized by EPR
spectroscopy.
A qualitative coulometric analysis was obtained by comparing
the peak currents of the ferrocene redox couple to the imida-
zolium cation reduction peak and this suggests that the reduction
is a one-electron process. It is postulated that a neutral radical
intermediate formed, which then collapsed to give the NHCs.
Cyclic voltammetry and EPR studies were used to aid in the
structural assignment of the radical intermediate.
CV studies of different imidazolium salts were carried out
and it was observed that the reduction of 1,3-bis(2,6-diisopro-
pylphenyl)imidazolium chloride was slightly reversible (Figure
9), suggesting that the radical formed has a longer lifetime
compared to those of other imidazolium salts, which allowed
for EPR studies.
A sample of 1,3-bis(2,6-diisopropylphenyl)imidazolium chlo-
ride was electrolyzed by applying a potential of -2.3 V for 5
min in acetonitrile (1.66 mM) under an inert atmosphere, with
reference to a standard calomel electrode. The acetonitrile
solution was then rapidly transferred to a quartz flat cell and
the EPR spectrum was measured.
The main 21-line spectrum, shown in Figure 10, was
reproduced very accurately by simulation and demonstrates that
the electrochemically generated species contains radical-spin
coupling to two equivalent nitrogen atoms (14N, I ) 1) and three
inequivalent hydrogen atoms (1H, I ) 1/2). The hyperfine
couplings (mT) measured were as follows: a(14N × 2) ) 0.395,
In any solvent, there are two reactions one must be concerned
with: the first is deprotonation and the second electron transfer.
In the case of a PhosIL, a phosphorane (or Wittig reagent) would
be the product of deprotonation and for an ImIL, a carbene.
With respect to electron transfer, a phosphoryl or imidozyl
radical would be expected dependent on the choice of IL;
however, these are both reactive and thus transient. The absence
(72) McKenzie, I.; Brodovitch, J.-C.; Percival, P. W.; Ramnial, T.;
Clyburne, J. A. C. J. Am. Chem. Soc. 2003, 125, 11565.
(73) Behar, D.; Gonzalez, C.; Neta, P. J. Phys. Chem. A 2001, 105, 7607.
810 J. Org. Chem., Vol. 73, No. 3, 2008

C-Centered Strong Bases in Phosphonium Ionic Liquids
mL) to give a clear colorless solution. This stability of triph-
enylphosphine in PhosIL-Cl was analyzed by 31P NMR (162
MHz): δ -4.8. The same chemical shift was observed after 2
months consistent with 31P NMR of triphenylphosphine in CDCl3:
δ -4.9.
2.7. Stability of tert-Butylisocyanide in PhosIL-Cl. A few drops
of tert-butylisocyanide was added to PhosIL-Cl, and an IR study
was carried out on the mixture. No degradation was observed (1
month) as indicated by the presence of a diagnostic stretch at 2133
(vs) (NtC).
of phosphoranes and phosphoryl in the PhosILs indicates that
they are a suitable choice as solvent for strongly basic, carbon-
centered reagents.
Experimental Section
1. General Experimental Procedures. The experimental pro-
cedures are described in the Supporting Information.
2. Synthetic Procedures. 2.1 Purification of Phosphonium
Ionic Liquids. Saturated aqueous sodium hydrogen carbonate (20
mL) was added to the PhosIL (120 mL) and the solution was stirred
for 15 min. Vigorous foaming occurred. The solution was then
washed with water (500 mL). The ionic liquid layer obtained was
extracted using hexane (120 mL) and water (120 mL) in 3 × 40
mL aliquots. The ionic liquid was then dried by azeotropic
distillation using toluene (20 mL), followed by exhaustive evacu-
2.8. Preparation of a Solution of 1,3-Bis(2,4,6-trimethylphe-
nyl)imidazol-2-ylidene, NHC 1, in PhosIL-Cl. 1,3-Bis(2,4,6-
trimethylphenyl)imidazol-2-ylidene, NHC 1 (0.20 g, 0.66 mmol),
was dissolved in PhosIL-Cl (2 mL) with the addition of toluene
drops to aid the dissolution of the NHC. The toluene was then
1
removed under vacuum. H NMR (THF-d8, 400 MHz) δ 2.02 (s,
12 H), 2.30 (s, 6 H), 6.94 (s, 4 H), 7.04 (s, 2 H); 13C NMR (THF-
d8, 101 MHz) δ 18.0 (s,), 21.0 (s), 121.3 (s), 129.7 (s), 135.7 (s),
137.6 (s), 139.7 (s), 219.7 (s).
1
ation. H NMR spectroscopy showed the absence of water in the
ionic liquid and 31P NMR spectroscopy (single peak at 33 ppm)
showed the presence of only one type of phosphorus site and no
residual phosphines present.
2.9. Reduction of 1,3-Bis(2,4,6-trimethylphenyl)imidazolium
Chloride with Potassium in PhosIL-Cl To Give 1,3-Bis(2,4,6-
trimethylphenyl)imidazol-2-ylidene, NHC 1. 1,3-Bis(2,4,6-trim-
ethylphenyl)imidazolium chloride74 (2.00 g, 5.87 mmol) and an
excess of potassium metal (0.35 g, 8.75 mmol), previously washed
with anhydrous THF, were added to PhosIL-Cl (10 mL). The
reaction mixture was heated at 80 °C under nitrogen for 24 h.
Hexanes (10 mL) were added to the resulting suspension and the
solution was filtered through Celite to remove the side product,
potassium chloride. Evaporation to remove hexanes gave a reddish
brown viscous residue that was characterized as a solution of
PhosIL-Cl and NHC 1. 1H NMR (THF-d8, 400 MHz) δ 2.08 (s, 12
H), 2.31 (s, 6 H), 6.96 (s, 4 H), 7.14 (s, 2 H); 13C NMR (THF-d8,
101 MHz) δ 18.6 (s), 21.6 (s), 122.1 (s), 129.8 (s), 136.2 (s), 138.2
(s), 139.9 (s), 215.8 (s). These spectroscopic data are consistent
with NHC 1 dissolved in PhosIL-Cl.
2.2. Preparation of Stock Solutions of Ethereal Grignards in
PhosIL-Cl. Stock solutions of reaction media comprising Grignard
reagents in PhosIL-Cl were prepared by mixing commercially
available ethereal (i.e., in diethyl ether, tetrahydrofuran, etc.)
solutions of Grignard reagents with PhosIL. Five milliliters of
PhMgBr (1 M in THF) was added dropwise to 5 mL of PhosIL-Cl
under an atmosphere of nitrogen and the solution was cooled at
-78 °C. Ethereal solutions of Grignard reagents dissolved in
PhosIL-Cl are air and moisture sensitive. These solutions show no
significant sign of degradation after 1 month as shown by reactivity
studies; however, upon complete removal of THF, decomposition
of the phosphonium is observed.
2.3. Reaction of PhosIL-Cl/Grignard with Bromine. To an
aliquot (5 mL) of 1-month-old 1 M PhosIL-Cl/Grignard reagent
solution was added 1 equiv of Br2 and the mixture was stirred for
1 h. The reaction was quenched with water (20 mL) and extracted
with hexanes (20 mL). The extract was analyzed by GC-MS and
the formation of phenylbromide was observed in addition to a small
amount (5%) of biphenyl. Benzene was not detected indicating little
degradation of the PhosIL-Cl/Grignard reagent solution.
2.4. Generation of an Ether-Free Solution Composed of a
Grignard Reagent Dissolved in PhosIL-C9H19COO. To PhosIL-
C9H19COO (5 mL) with a few drops of THF was added com-
mercially available 1 M PhMgBr (5.0 mL) in THF at -78 °C. The
mixture was stirred and warmed to room temperature. Tetrahydro-
furan was removed in vacuo leaving a viscous pale yellow or orange
solution to which a few drops of hexanes was added to reduce
viscosity. Treatment of this solution with 1 equiv of either bromine
or N,N′-dimethylformamide followed by quenching with saturated
aqueous ammonium chloride and addition of water, followed by
extraction with dichloromethane resulted in the formation of
bromobenzene (98%) and benzaldehyde (99%), respectively, as
determined by GC-MS studies.
2.5. Addition Reaction of Grignard Reagents in PhosIL-
C9H19COO. A stock solution of 1 M PhosIL-C9H19COO and
PhMgBr was prepared by mixing equal volumes of commercially
available 1.0 M PhMgBr in THF with the ionic liquid, followed
by removal of the THF under vacuum. Aliquots (5 mL) of the
PhosIL-C9H19COO/Grignard reagent solution were treated with 1
equiv of reagents (e.g., aldehydes, ketone, DMF, etc.). The solutions
were stirred under N2 for 1-3 h. The reaction mixtures were
quenched with saturated aqueous ammonium chloride (2 mL)
followed by the addition of water (25 mL). Hexanes (20 mL) were
added to establish a three-phase system and the products were
extracted into the hexanes layer. Products were identified using
NMR spectroscopy and the extracts were analyzed by GC-MS
studies. Yields are reported in Scheme 3.
2.10. Reaction of 13C-Labeled 1,3-Bis(2,4,6-trimethylphenyl)-
imidazolium Chloride with Phenylmagnesium Bromide in
PhosIL-Cl. 1,3-Bis(2,4,6-trimethylphenyl)imidazolium chloride
(0.10 g, 0.29 mmol) was added to PhosIL-Cl (2 mL) to afford a
creamy suspension to which 1.0 M PhMgBr (0.35 mL, 0.29 mmol)
was added dropwise with stirring. The mixture turned to a clear
1
yellow solution after being stirred for 15 min. H NMR (THF-d8,
400 MHz) δ 2.18 (s, 12 H), 2.41 (s, 6 H), 7.15 (s, 4 H), 7.41 (s, 2
H); 13C NMR (THF-d8, 101 MHz) δ 15.8 (s,), 21.3 (s), 130.2 (s),
130.6 (s), 136.9 (s), 137.1 (s), 143.6 (s), 219.7 (s). These spec-
troscopic data are consistent with NHC 1 dissolved in PhosIL-Cl.
2.11. Preparation of 1,3-Bis(2,4,6-trimethylphenyl)imidazol-
2-thione, IMes(S). To an NHC/PhosIL solution prepared according
to the method described above in subsection 2.8 (0.50 g, 1.64 mmol)
in PhosIL-Cl (10 mL) was added sulfur (0.05 g, 1.64 mmol)
followed by an amount of hexanes to decrease viscosity. The
resulting mixture was stirred for 4 h at room temperature. More
hexanes were then added (5 mL) and the mixture was filtered
through Celite and the filtrate evaporated to remove hexanes leaving
1
a viscous yellow residue. H NMR (C6D6, 400 MHz) δ 2.08 (s, 6
H), 2.32(s, 12 H), 6.66 (s, 2 H), 6.78 (s, 4 H); 13C NMR (C6D6,
101 MHz) δ 20.9 (s), 22.4 (s), 124.9 (s), 128.4 (s), 133.4 (s), 138.9
(s), 158.1 (s), 162.9 (s); MS (CI) m/z 336.3. These 13C NMR data
for the CdS are consistent with previously reported data for IMesd
S.56
2.12. Preparation of Pentacarbonyl(1,3-bis(2,4,6-trimeth-
ylphenyl)imidazol-2-ylidene)chromium, IMesCr(CO)5 in Pho-
sIL-Cl. A solution of 1,3-bis(2,4,6-trimethylphenyl)imidazol-2-
ylidene was prepared by dissolving NHC 1 (0.20 g, 0.91 mmol) in
PhosIL-Cl (10 mL). Toluene (2 mL) was added to dissolve the
NHC in PhosIL-Cl. The toluene was then removed under vacuum.
2.6. Stability of Triphenylphosphine in PhosIL-Cl. Triph-
enylphosphine (0.20 g, 0.76 mmol) was dissolved in PhosIL-Cl (2
(74) Arduengo, A. J., III; Krafczyk, R.; Schmutzler, R. Tetrahedron 1999,
55, 14523.
J. Org. Chem, Vol. 73, No. 3, 2008 811

Ramnial et al.
The PhosIL-Cl solution of NHC 1 was treated with chromium
hexacarbonyl (0.20 g, 0.91 mmol). A yellow precipitate was
obtained. The mixture was warmed at 60 °C for 2 h. An IR study
on the yellow precipitate contained diagnostic stretches at 2056
(s) and 1923 (vs) (CO) consistent with IMesCr(CO)5.75
verify the nature of the stationary point. Z-matrix coordinates
constrained to the appropriate symmetry were used for efficiency,
as any problems would manifest themselves by an imaginary mode
orthogonal to the spanned Z-matrix space. The Hessian was also
evaluated at the starting STO-3G geometry to aid convergence.
2.13. Generation of Ethyltriphenylphosphorane in PhosIL.
Ethyltriphenylphosphonium bromide (0.20 g, 0.58 mmol) was
dissolved in PhosIL-Cl and to it was added 1.0 M PhMgBr in
THF (0.58 mL, 0.58 mmol) dropwise. The color of the solution
changed from white to brick orange. 31P NMR (CD2Cl2, 162
MHz): δ 15.6. These spectroscopic data are consistent with the
formation of ethyltriphenylphosphorane in PhosIL-Cl. Similarly,
ethyltriphenylphosphonium bromide is deprotonated in PhosIL-
C9H19COO Grignard stock solutions to produce ethyltriphenylphos-
phorane.
2.14. Wittig Reactions: Formation of Methylstyrene in
PhosIL-C9H19COO. To a cold (-78 °C) sample of tetradecyl-
(trihexyl)phosphonium decanoate (5 mL) was added PhosIL-C9H19-
COO, 1 M PhMgBr in THF (1.2 mL, 1.2 mmol), and the solution
was allowed to slowly warm to room temperature. The THF was
removed under vacuum, and to the resulting solution several drops
of hexanes was added to reduce viscosity. Triphenylethylphospho-
nium bromide (0.4 g, 1.10 mmol) was then added to the solution.
A color change from white to reddish orange was observed and
the mixture was stirred under nitrogen for 1 h. 31P{1H} NMR
showed a distinctive peak at 15.3 ppm for the deprotonation of
triphenylethylphosphonium bromide to give ethyltriphenylphos-
phorane. Benzaldehyde (0.11 g, 1.10 mmol) was then added to the
mixture and an instant color change from yellow to white was
observed. The mixture was stirred for 2 h and then quenched with
water. The product was extracted with dichloromethane, which was
dried using anhydrous magnesium sulfate to give methylstyrene
(96%) analyzed by GC-MS. Formation of Ph3PO was confirmed
by mass spectrometry (278 amu).
Acknowledgment. Funding for this research was provided
by the Natural Sciences and Engineering Council of Canada.
We also acknowledge the Donors of the Petroleum Research
Fund, administered by American Chemical Society. We are
grateful to CYTEC for their donation of PhosILs, and to Al
Robertson (Cytec) for encouraging this research. We extend our
thanks to Dr. Michael Jennings (University of Western Ontario)
for collection of X-ray data. J.A.C.C. acknowledges support
from the Canada Research Chairs Program, the Canadian
Foundation for Innovation, and the Natural Sciences and
Engineering Research Council of Canada. The authors also
acknowledge the Atlantic Region Magnetic Resonance Centre
(ARMRC) at Dalhousie University for NMR data, and especially
Dr. K. Robertson and Dr. M. Lumsden for their help in obtaining
NMR spectra.
Supporting Information Available: General experimental
procedures and atomic coordinates for diphenylmagnesium, acetate
and the acetate anion complex of diphenylmagnesium, as well as
crystallographic data and a CIF file for tetrabutylphosphonium
acetate monoacetic acid (CCDC 650492). This material is available
free of charge via the Internet at http://pubs.acs.org.
JO701289D
(76) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. B.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.; Rega,
N.; Salvador, P.; Dannenberg, J. J.; Malick, D. K.; Rabuck, A. D.;
Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A.
G.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.;
Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.;
Replogle, E. S.; Pople, J. A. Gaussian 98, Revision A.11.3; Gaussian, Inc.:
Pittsburgh, PA, 2002.
3. Computational Studies. Calculations were performed using
Gaussian 98,76 using a stepping stone approach in which the
geometries at the levels HF/STO-3G, HF/3-21G, HF/6-31G*, and
HF/6-31+G* were sequentially optimized using default specifica-
tions. After each level, a frequency calculation was performed to
(75) O¨ fele, K.; Herrmann, W. A.; Mihalios, D.; Elison, M.; Herdtweck,
E.; Scherer, W.; Mink, J. J. Organomet. Chem. 1993, 459, 177.
812 J. Org. Chem., Vol. 73, No. 3, 2008