Rate Constants for Reaction of 1,2-Dimethylimidazole with Benzyl Bromide in Ionic Liquids and Organic Solvents

Article abstract of DOI:10.1002/kin.10162

The rate constant for the Menschutkin reaction of 1,2-dimethylimidazole with benzyl bromide to produce 3-benzyl-1,2-dimethylimidazolium bromide was determined in a number of ionic liquids and molecular organic solvents. The rate constants in 12 ionic liquids are in the range of (1.0-3.2) × 10 ^-3 L mol^-1 s^-1 and vary with the solvent anion in the order (CF₃SO₂)₂N^{-< PF} ₆^- < BF₄^-. Variations with the solvent cation (butylmethylimidazolium, octylmethylimidazolium, butyldimethylimidazolium, octyldimethylimidazolium, butylmethylpyrrolidinium, and hexyltributylammonium) are minimal. The rate constants in the ionic liquids are comparable to those in polar aprotic molecular solvents (acetonitrile, propylene carbonate) but much higher than those in weakly polar organic solvents and in alcohols. Correlation of the rate constants with the solvatochromic parameter E_T(30) is reasonable within each group of similar solvents but very poor when all the solvents are correlated together, Better correlation is obtained for the organic solvents by using a combination of two parameters, π* (dipolarity/polarizibility) and α (hydrogen bond acidity), while additional parameters such as δ (cohesive energy density) do not provide any further improvement.

Full text of DOI:10.1002/kin.10162

Rate Constants for Reaction
of 1,2-Dimethylimidazole
with Benzyl Bromide in
Ionic Liquids and Organic
Solvents
A. SKRZYPCZAK, P. NETA
Physical and Chemical Properties Division, National Institute of Standards and Technology, Gaithersburg, Maryland 20899
Received 14 May 2003; accepted 28 July 2003
DOI 10.1002/kin.10162
ABSTRACT: The rate constant for the Menschutkin reaction of 1,2-dimethylimidazole with ben-
zyl bromide to produce 3-benzyl-1,2-dimethylimidazolium bromide was determined in a num-
ber of ionic liquids and molecular organic solvents. The rate constants in 12 ionic liquids
are in the range of (1.0–3.2) × 10−3 L mol−1 s−1 and vary with the solvent anion in the
−
order (CF3SO2)2N− < PF6 < BF4−. Variations with the solvent cation (butylmethylimida-
zolium, octylmethylimidazolium, butyldimethylimidazolium, octyldimethylimidazolium, butyl-
methylpyrrolidinium, and hexyltributylammonium) are minimal. The rate constants in the ionic
liquids are comparable to those in polar aprotic molecular solvents (acetonitrile, propylene
carbonate) but much higher than those in weakly polar organic solvents and in alcohols. Cor-
relation of the rate constants with the solvatochromic parameter ET(30) is reasonable within
each group of similar solvents but very poor when all the solvents are correlated together. Better
correlation is obtained for the organic solvents by using a combination of two parameters, π∗
(dipolarity/polarizibility) and α (hydrogen bond acidity), while additional parameters such as δ
(cohesive energy density) do not provide any further improvement. ꢀ 2004 Wiley Periodicals,
C
Inc.∗ Int J Chem Kinet 36: 253–258, 2004
uids can be tuned by varying the structures of their ions
in order to optimize the solvent properties for specific
applications. Ionic liquids affect reaction rates and se-
lectivity. To understand the effects of ionic liquids on
chemical reactions, the rate constants for several ele-
mentary reactions in ionic liquids have been studied [2]
by the pulse radiolysis technique and compared with
those in other solvents. In a number of cases the rate
constants in ionic liquids were lower than those in water
and polar organic solvents either because of the high
viscosity of ionic liquids, which limits the diffusion
rate, or because of an apparent lower polarity.
INTRODUCTION
Room temperature ionic liquids [1] have been pro-
posed as solvents for green processing because of their
unique physical properties, such as nonvolatility and
nonflammability. The physical properties of ionic liq-
Correspondence to: P. Neta; e-mail: pedatsur.neta@nist.gov.
A. Skrzypczak: On leave from Poznan University of Technology,
Poland.
2004 Wiley Periodicals, Inc. ∗This article is a US Government
c
ꢀ
work and, as such, is in the public domain of the United States of
America.

254
SKRZYPCZAK AND NETA
In the study of organic reactions in ionic liquids, the
emphasis has been on product yields and selectivity.
Few studies were devoted to measurements of reaction
rate constants. Welton and coworkers [3] compared the
rate constants for reaction of halide ions (Cl−, Br−, I−)
with methyl p-nitrobenzenesulfonate to give methyl
halide and p-nitrobenzenesulfonate anion and showed
that the relative nucleophilicities of the three halides
in different ionic liquids were dependent on the ionic
liquid cation. In a study of the Diels–Alder reaction
between cyclopentadiene and methyl acrylate in vari-
ous ionic liquids, Welton and coworkers [4] found that
both the rate and the selectivity increased with increas-
ing hydrogen bonding between the solvent cation and
the methyl acrylate, as is the case for molecular organic
solvents. The anion of the ionic liquid solvent affected
the rate and selectivity through competition for hydro-
gen bonding with the cation. The Diels–Alder reac-
tion involves neutral reactants and products and its rate
constant in a variety of organic solvent correlates [5]
with Hildebrand’s solubility parameter (cohesive en-
ergy density, δ) because it depends mainly on solvent
rearragement around the reactants and products and on
hydrogen bonding. In this study we measure the rate
constant for the reaction between a tertiary amine and
an organic halide to produce the quaternary ammonium
salt, the Menschutkin reaction. Because this reaction
involves a polar transition state and ionic products, its
rate constant is affected not only by hydrogen bonding
but also by solvent polarity. Its measurement in ionic
liquidis thus expected toshedsomelighton thepolarity
of these solvents.
vents do not absorb light to any significant extent at
wavelengths above 260 nm and that the disappearance
of benzyl bromide can be followed at 270–280 nm with
only minimal interference from the absorption of DMI.
EXPERIMENTAL SECTION∗
The ionic liquids used in this study were prepared by
literature methods. 1-Butyl-3-methylimidazolium hex-
afluorophosphate ([bmim][PF6]) and tetrafluoroborate
([bmim][BF4]) were prepared from 1-methylimida-
zole, 1-chlorobutane, and the inorganic salts for meta-
thesis, as described before [2a,b]. The other PF6 and
BF4 imidazolium ionic liquids were prepared by simi-
lar procedures using different combinations of starting
materials in addition to those mentioned above: 1,2-
dimethylimidazole, 1-bromobutane, or 1-bromooc-
tane. The bromide salts were prepared by a similar
method as the chloride [2a,b] except that the mixture
was refluxed for only 2 h instead of 2 days. Any remain-
ing reactants were removed by extraction with ethyl
acetate and the extracts were analyzed spectropho-
tometrically to ascertain that all of these impurities
were removed. The bis(trifluoromethylsulfonyl)imide
(NTf2) salt was prepared from the bromide by meta-
thesis with LiNTf2 in water and subsequent washing
with water and drying. Hexyltributylammonium
bis(trifluoromethylsulfonyl)imide ([HxBu3N][NTf2])
and N-methyl-N-butylpyrrolidinium bis(trifluoro-
methylsulfonyl)imide ([MeBuPyr][NTf2]) were pre-
pared as described before [2f]. In all preparations of
−
the hydrophobic ionic liquids (containing PF6 or
To measure the rate constant for such a reaction in
a homogeneous solution, both reactants and products
must be soluble in the liquid medium. The specific re-
action that was found to be measurable in a number
of ionic liquids and organic solvents is the reaction of
benzyl bromide with 1,2-dimethylimidazole (DMI) to
form 3-benzyl-1,2-dimethylimidazolium bromide.
−NTf2), the products were washed with pure water
(Millipore Super-Q water) at least six times and until
the water extract did not contain any halide ions or
any spectrophotometrically detectable impurities. The
water-soluble ionic liquids (containing BF4−) were
washed with ethyl acetate. All ionic liquids were dried
on a vacuum line at 80◦C for at least 2 h before use.
The other solvents and reactants were of the purest
grade commercially available and were obtained from
Aldrich, Fluka, or Mallinckrodt.∗ They were used
without further purification, except for 1-octanol,
which was dried in vacuum at 60◦C before use.
(1)
The rate constant for reaction (1) was measured by
mixing known concentrations of the reactants and fol-
lowing the decay of the benzyl bromide spectrophoto-
metrically. Separate solutions of benzyl bromide and
DMI in the same solvent were freshly prepared before
Rate constants for the Menschutkin reaction have
been measured in many solvents by using conductiv-
ity detection of the ionic products. In ionic liquids,
however, it is not possible to observe the change in
conductivity due to reaction (1) because of the high
background conductivity of the solvent. Therefore, we
have used spectrophotometric detection. We found that
a wide selection of pure ionic liquids and organic sol-
∗The mention of commercial equipment or material does not im-
ply recognition or endorsement by the National Institute of Standards
and Technology, nor does it imply that the material or equipment
identified are necessarily the best available for the purpose.

RATE CONSTANTS FOR REACTION OF 1,2-DIMETHYLIMIDAZOLE WITH BENZYL BROMIDE
255
the experiment. Measured amounts of the two solutions
were combined in a quartz cuvette (0.2 or 1.0 cm opti-
cal path length), thoroughly mixed, and kept at (21
1)◦C. The optical absorption at 275 or 280 nm (mainly
due to benzyl bromide) was measured in a Cary 3 spec-
trophotometer at various times after mixing and the re-
action was followed until the absorbance decreased to a
plateaulevel(fromanhourtoseveraldays). Theplateau
level was due to the absorbance of DMI and was always
lower than the initial absorbance by at least a factor of 2.
In all experiments the concentration of benzyl bromide
was close to 9 × 10−3 mol L−1 and the concentration
of DMI was in large excess and was varied between
0.05 and 0.8 mol L−1. Kinetic plots of absorbance vs.
time provided the first-order rate constants (typical ex-
amples are in Fig. 1a), and from linear plots of these
rate constants as a function of DMI concentration the
second-order rate constants for reaction (1) (k1) were
derived (typical examples are in Fig. 1b). In addition
to reaction (1), benzyl bromide undergoes solvolysis in
protic solvents or by water present in aprotic solvents.
This side reaction, which can affect the measurement
of k1, was minimized by careful drying of all solvents
and its effect was examined in parallel experiments in
the absence of DMI. In particular, all the ionic liquids
were tested for reaction of benzyl bromide in the ab-
sence of DMI for periods (24–170 h) that are much
longer than those used in the kinetic experiments to
ascertain that there is no effect of water or any other
contaminant on the measured rate constant. The ef-
fect of solvolysis was significant only in methanol, and
was taken into account when deriving the rate constant,
but was practically negligible in all other solvents. The
rate constants have an estimated standard uncertainty
of 20%, taking into account the standard deviation of
the experimental kinetic results and estimated uncer-
tainties in concentrations.
RESULTS AND DISCUSSION
The Menschutkin reaction was chosen because it has
been studied extensively in many solvents [6] and is
a representative reaction between neutral reactants,
which form polar transition states and ionic products.
We have determined the rate constants for reaction (1)
in 12 ionic liquids (Table I) and 13 molecular solvents
(Table II) for comparison. The results show that the rate
constants in the molecular solvents vary by two orders
of magnitude while the rate constants in the ionic liq-
uids vary by less than a factor of 3 and are similar to
the highest two values in the molecular solvents, i.e.,
Table I Rate Constants for Reaction of PhCH2Br with
DMI in Ionic Liquids and ET(30) Values
Solventa
k (10−3 L mol−1 s−1 b
)
ET(30)c
[bmim][NTf2]
[bmim][PF6]
[bmim][BF4]
[omim][NTf2]
[omim][PF6]
[omim][BF4]
[bmmim][NTf2]
[bmmim][BF4]
[ommim][NTf2]
[ommim][BF4]
[MeBuPyr][NTf2]
[HxBu3N][NTf2]
1.4
1.9
2.8
1.0
1.8
2.3
1.9
3.2
1.5
2.8
2.3
1.2
51.5
52.3
52.5
51.1
51.2
48.6
47.7
48.3
Figure 1 (a) Kinetic plots of absorbance vs. time for the
reaction of benzyl bromide with DMI in [bmmim][NTf2]
using [DMI] = 0.48 mol L−1 ( ) and 0.70 mol L−1 ( ).
◦
•
a The solvent contained between 0 and 0.9 mol L−1 DMI.
b The rate constants have an estimated standard uncertainty of
20%, taking into account the standard deviation of the experimental
kinetic results and estimated uncertainties in concentrations.
c Values of ET(30) from Ref. [7].
The lines are the first-order fits. (b) First-order rate constants
as a function of DMI concentration in [bmmim][BF ] ( ),
•
4
[bmim][BF ] (ꢀ), [bmmim][NTf ] ( ), and [omim][NTf ]
◦
4
2
2
(ꢁ).

256
SKRZYPCZAK AND NETA
Table II Rate Constants for Reaction of PhCH2Br with DMI in Various Solvents and Solvent Parameters
Solventa
k (L mol−1 s−1 b
)
ET(30)c
π∗d
αd
δe
Propylene carbonate
N-Methylformamide
Acetonitrile
Propionitrile
Butyronitrile
Ethylene glycol
2-Methoxyethanol
Methanol
Ethanol
1-Propanol
2-Propanol
1-Hexanol
2.6 × 10−3
7.1 × 10−4
1.2 × 10−3
6.0 × 10−4
5.0 × 10−4
1.8 × 10−4
7.9 × 10−5
5.3 × 10−5
2.9 × 10−5
1.7 × 10−5
4.1 × 10−5
1.8 × 10−5
2.0 × 10−5
46.6
54.1
45.6
43.7
43.1
56.3
52.3
55.4
51.9
50.7
48.4
48.8
48.3
0.83
0.92 f
0.75
0.71
0.71
0.92
0.71
0.60
0.54
0.52
0.48
0.52h
0.0
21.8
32.9
24.1
21.8
19.8
32.4
25.5
29.3
26.0
24.4
23.7
21.8
20.9
0.35 f
0.19
0.06g
0.0g
0.90
0.9
0.93
0.83
0.78
0.76
0.68h
1-Octanol
a The solvent contained between 0 and 0.9 mol L−1 DMI.
b The rate constants have an estimated standard uncertainty of 20%, taking into account the standard deviation of the experimental kinetic
results and estimated uncertainties in concentrations.
c Values of ET(30) from Ref. [8].
d Values of π∗ and α from Ref. [9] unless indicated otherwise.
e Values of δ from Ref. [10], except for the values for N-methylformamide, which was not reported by Marcus and was taken from the earlier
Ref. [11]. The two sets of values differ considerably for propylene carbonate and slightly for butyronitrile, ethylene glycol, and methoxyethanol.
Correlation using the values in Ref. [11] showed greater uncertainty in the coefficient for δ.
f
Estimated from the values for formamide and dimethylformamide.
g Estimated from the values for acetonitrile and other nitriles.
h From Ref. [12].
acetonitrile and propylene carbonate. Thus the ionic
liquids behave in this reaction like the polar aprotic sol-
vents. The variations within the ionic liquids indicate
measured rate constants with the ET(30) values for the
ionic liquids and the molecular solvents indicates that
the correlation is reasonable for each group of simi-
lar solvents but not for all solvents together (Fig. 2a).
An important difference between the different groups
is their hydrogen bonding properties. Thus, the alco-
hols fit in one group whereas the aprotic nitriles and
propylene carbonate fit in another group; the slightly
hydrogen bonding N-methylformamide falls between
the lines for these two groups. The ionic liquids also
fall between the two lines and are divided into two dis-
tinct groups: the 1,3-dialkylimidazolium and the 1,2,3-
trialkylimidazolium salts. The latter ionic liquids have
lower hydrogen bonding capacity (absence of the rel-
atively acidic hydrogen at position 2) and are indeed
closer to the line of the aprotic solvents. The former
group of ionic liquids includes five members and shows
a trend of increasing reactivity with increasing ET(30)
value.
Correlation of the rate constant for the Menschutkin
reactionwithonesolventpolarityparameter, suchasdi-
electric constant [14] or ET(30) [8], has been shown be-
fore to give a good fit only within a group of similar sol-
vents and the above correlation with ET(30) is further
confirmation of the earlier findings. Correlation with
other single parameters, Kamlet and Taft’s dipolar-
ity/polarizability parameter (π∗) [9] and Hildebrand’s
a small effect of the solvent anion ((CF3SO2)2N−
<
PF6− < BF4−) but no clear effect of the solvent cation.
It appears that replacing the butyl group on the imida-
zolium cation with an octyl group lowers the rate con-
stant slightly and adding a methyl group at position 2
of the imidazolium increases the rate constant slightly.
These effects, however, although they are reasonable
for the expected effects of solvent polarity and hydro-
gen bonding capacity (see below), are almost within
the experimental uncertainty and must be considered
tentative.
Solvent effects on reaction rate constants are gener-
ally correlated with one or several solvent parameters
that represent polarity, hydrogen bonding, and other
properties. The only solvent polarity parameter that
was reported [7,13] for ionic liquids is that derived
from measurements of solvatochromic effects, such
as ET(30). The values of ET(30) derived from such
measurements are also summarized in Tables I and II.
They are quite similar for various alkylmethylimida-
zolium salts of several anions and are somewhat lower
for the corresponding alkyldimethylimidazolium salts,
where the slightly acidic hydrogen at position 2 is re-
placed with a methyl group. An attempt to correlate the

RATE CONSTANTS FOR REACTION OF 1,2-DIMETHYLIMIDAZOLE WITH BENZYL BROMIDE
257
a very large uncertainty ( 800%) while the depen-
dencies on π∗ and α have much smaller uncertainties
( 33%). This finding suggests that correlation with
only two parametrs, π∗ and α, may be sufficient. In-
deed, this correlation (Fig. 2b) fits the equation
log k = −(4.95 0.40) + (2.61 0.50)π∗
− (1.22 0.20)α
(3)
with a similar correlation coefficient, r2 = 0.92, and
the plot of Eq. (3) is indistinguishable from that of
Eq. (2) (Fig. 2b, ∇ vs. ꢀ). As found in the earlier re-
port, the hydrogen bonding (α) has an opposite effect
from that of π∗, and this is reflected in the findings that
the reaction is much slower in protic solvents than in
aprotic solvents of similar polarity.
The present results indicate that the effects of ionic
liquids on rates of reactions are different for different
ionic liquids. Although the slight variations in the val-
ues of ET(30) for different ionic liquids can be used
to rationalize some of the observed solvent effects, ad-
ditional solvent parameters have to be determined for
a variety of ionic liquids before full correlations with
reaction rates can be achieved. Determination of ad-
ditional solvent parameters will permit design of ionic
liquids, by proper choice of anion and cation structures,
for enhancement of the rates of specific reactions.
Figure 2 Correlation of the rate constant for reaction (1)
with solvent parameters. (a) Correlation with ET(30), show-
ing two separate lines for the alcohols (ꢁ) and for the aprotic
solvents (nitriles and propylene carbonate) (ꢀ). The values
BIBLIOGRAPHY
for N-methylformamide ( ), dialkylimidazolium ionic liq-
ꢂ
uids ( ), and trialkylimidazolium ionic liquids ( ) fall be-
•
◦
1. (a) Welton, T. Chem Rev 1999, 99, 2071–2083; (b)
Wasserscheid, P.; Keim, W. Angew Chem Int Ed 2000,
39, 3772–3789; (c) Earle, M. J.; Seddon, K. R. Pure
Appl Chem 2000, 70, 1391–1398; (d) Brennecke, J. F.;
Maginn, E. J. AIChE J 2001, 47, 2384–2389.
tween the two lines in distinct groups. (b) Multiparameter
correlation for the molecular solvents using π∗, α, and δ (ꢁ)
or only π∗ and α (∇) (see text).
2. (a) Behar, D.; Gonzalez, C.; Neta, P. J Phys Chem
A 2001, 105, 7607–7614; (b) Behar, D.; Neta, P.;
Schultheisz, C. J Phys Chem A 2002, 106, 3139–3147;
(c) Grodkowski, J.; Neta, P. J Phys Chem A 2002, 106,
5468–5473; (d) Grodkowski, J.; Neta, P. J Phys Chem
A 2002, 106, 9030–9035; (e) Grodkowski, J.; Neta, P. J
Phys Chem A 2002, 106, 11130–11134; (f) Skrzypczak,
A.; Neta, P. J Phys Chem A 2003, 107, 7800–7803.
3. (a) Lancaster, N. L.; Welton, T.; Young, G. B. J Chem
Soc, Perkin Trans 2 2001, 2267–2270; (b) Lancaster, N.
L.; Salter, P. A.; Welton, T.; Young, G. B. J Org Chem
2002, 67, 8855–8861.
solubility parameter (δ) [11], also gives two separate
lines for two distinct groups of solvents (graphs not
shown). To obtain a reasonable correlation with all sol-
vents it is necessary to use multiparameter correlations,
which can be done for the molecular solvents but not for
the ionic liquids. Following the suggestion of Abraham
[15], we correlated the rate constants with π∗, δ, and the
hydrogen bonding acidity α. The correlation (Fig. 2b)
fits the equation
4. Aggarwal, A.; Lancaster, N. L.; Sethi, A. R.; Welton, T.
Green Chem 2002, 4, 517–520.
5. Abraham, M. H. Prog Phys Org Chem 1974, 11,
1–87.
log k = −4.98 + 2.71π∗ − 1.18α − 8.3 × 10−5δ2
(2)
6. (a) Abraham, M. H.; Grellier, P. L. J Chem Soc, Perkin
Trans 2, 1976, 1735–1741; (b) Drougard, Y.; Decroocq,
D. Bull Soc Chim Fr 1969, 2972–2983.
with a correlation coefficient r2 = 0.92. The correla-
tion shows, however, that the dependence on δ2 has

258
SKRZYPCZAK AND NETA
7. Muldoon, M. J.; Gordon, C. M.; Dunkin, I. R. J Chem
Soc, Perkin Trans 2, 2001, 433–435.
8. Reichardt, C. Solvents and Solvent Effects in
Organic Chemistry, 2nd ed.; VCH: Weinheim,
1988.
9. Kamlet, M. J.; Abboud, J.-L. M.; Abraham, M. H.; Taft,
R. W. J Org Chem 1983, 48, 2877–2887.
10. Marcus, Y. Ion Solvation; Wiley: New York, 1985.
11. Barton, A. F. M. CRC Handbook of Solubility Parame-
ters and Other Cohesion Parameters; CRC Press: Boca
Raton, FL, 1983.
12. Calado, A. R. T.; Pinheiro, L. M. V.; Albuquerque, L. M.
P. C.; Goncalves, R. M. C.; Roses, M.; Rafols, C.; Bosch,
E. Collect Czech Chem Commun 1994, 59, 898–904.
13. (a) Carmichael, A. J.; Seddon, K. R. J Phys Org Chem
2000, 13, 591–595; (b) Aki, S. N. V. K.; Brennecke, J.
F.; Samanta, A. J Chem Soc, Chem Commun 2001, 413–
414; (c) Fletcher, K. A.; Storey, I. A.; Hendricks, A. E.;
Pandey, S.; Pandey, S. Green Chem 2001, 3, 210–215.
14. Wiberg, K. Physical Organic Chemistry; Wiley: New
York, 1964.
15. Abraham, M. H. Pure Appl Chem 1985, 51, 1055–1064.