DMSO is only slightly less acidic than water—see Table 4). In
DMSO, acetonitrile has a pKa of 31.3, compared with the value
of 31.2 for water,14 and the [CH2CN]Ϫ anion is the likely nucle-
ophile in condensations of 1 with acetonitrile; hence, this anion
may be the source of base in other reactions (e.g. epoxidation
using sulfonium salts in acetonitrile10,15). Consequently, DMSO
and acetonitrile also have potential as phase transfer catalysts
transporting base from KOH to an organic medium.6
Karl Fischer titrations were conducted using a Mettler-
Toledo DL 18 titrator and Hydranal 2 composite.
Kinetic methods
To a 100 ml three-necked flask, fitted with a Liebig condenser,
was added ketone (1, 0.02 mol) acetonitrile (40 ml), and water
(usually 0.1 ml), and the mixture was stirred magnetically (usu-
ally at a speed of 400 rpm) and was heated to reflux. The first
sample of the reaction mixture was removed (see below) before
KOH pellets (2.244 g, 0.04 mol) were added, and the reaction
was then sampled at regular intervals whilst heating and stirring
continued. Sampling involved using a 1 ml syringe with a 3Љ
needle to transfer of 2 or 3 drops of reaction mixture into a pre-
weighed volumetric flask, which was then reweighed quickly
before the reaction mixture was quenched with two drops of
water, diluted to 10 ml with acetonitrile, and stored at <5 ЊC
until duplicate HPLC analyses were made within 24 h.
Control experiments showed that rates were unaffected by a
CO2 trap, or by a drying tube fitted to the top of the condenser.
For ‘simultaneous’ HPLC and water analyses, water and benzo-
phenone were added to a refluxing mixture of acetonitrile and
KOH, and the Karl Fischer titrations were carried out without
significant delays. In reactions described using ‘conditioned’
KOH pellets, reaction mixtures were heated for predetermined
times with KOH and added water present but in the absence of
ketone, the ketone was then added, and monitoring was begun.
Experiments with furfural (4) were conducted as described
above, except at 60 1 ЊC.
Conclusions
The effectiveness of KOH pellets as base catalysts is increased
by the addition of small amounts of water (Fig. 3) to give
‘feebly-hydrated’10,15 KOH. In condensations such as formation
of nitriles (3) from ketones (1) in acetonitrile, a significant frac-
tion of the water produced during the reaction is absorbed by
the solid KOH, and water acts as an autocatalyst; the kinetic
order of the reaction changes from the expected first order (Fig.
2) to approximately zero order. As expected, the reaction is
catalysed by quaternary ammonium salts, and our results
(Tables 2 and 6) are not complicated by anion exchange which
occurs in PTC alkylations. Catalysis by cosolvent/cocatalysts
(Table 4) illustrate the potential role of solvents such as water,
acetonitrile, DMSO and PEG 200 as phase transfer catalysts in
reactions involving solid bases.6
Experimental
Materials
The concentration of carbonyl compound in each sample
could be calculated from the weight of each sample and the
average area of carbonyl signal, but the precision of the results
was improved by also taking into account the area of the nitrile
peak.32 From a plot of concentration of carbonyl compound vs.
time (e.g. Fig. 1), the appropriate linear region was selected,
from which zero order rate constants were obtained by least
squares analysis using Microsoft Excel.
Ketones (1, Z = H, Lancaster) and (1, Z = Me, Cl and F,
Aldrich) were used as supplied to synthesise the nitriles (3,
Z = Me, Cl and F) from acetonitrile and KOH;2 after column
1
chromatography and characterisation by H NMR spectros-
copy, 3, Z = H was recrystallised from 1:6 diethyl ether–
hexane, mp 46–47 ЊC (lit.,2,28 47–48 ЊC), 3, Z = Me was distilled
to give a yellow liquid, bp 210 ЊC (lit.,29 204–207 ЊC), 3, Z = Cl
was recrystallised from diethyl ether–hexane to give yellow
crystals, mp 53.5–55 ЊC (lit.,30 53 ЊC) and 3, Z = F was distilled
to give a yellow liquid. The hydroxynitrile (2, Z = H), obtained
by reacting 1, Z = H, with BunLi–acetonitrile in THF at
Ϫ78 ЊC,31 was recrystallised from ethanol, mp 142–144 ЊC
(lit.,31 141.5–143 ЊC). Furfural (4) was distilled under reduced
pressure and condensed with acetonitrile to give the hydroxy-
nitrile (5)31 and the acrylonitrile (6),2 as described elsewhere.
HPLC grade acetonitrile (Fisons) contained 0.1% water;
acetonitrile was dried by distillation from phosphorus pent-
oxide and then contained 0.015% water. Pelleted KOH (15.3%
water), sodium hydroxide (3.0% water) and potassium carbon-
ate (0.1% water) from Fisons were used as supplied.
Acknowledgements
This work was supported by the EPSRC and Zeneca through a
CASE award to A. H. L. and an industrially-funded student-
ship to S. J. L. We are also grateful for research grants from
SERC for HPLC equipment and from EPSRC for the Karl
Fischer titrator, and we thank G. Llewellyn (Swansea) for tech-
nical assistance.
References
1 M. Fedorynski, K. Wojciechowski, Z. Matacz and M. Makosza,
J. Org. Chem., 1978, 43, 4682.
2 S. A. DiBiase, B. A. Lipisko, A. Haag, R. A. Wolak and G. W.
Gokel, J. Org. Chem., 1979, 44, 4640.
3 S. Caddick, Tetrahedron, 1995, 51, 10 403.
4 C. Einhorn, J. Einhorn and J. L. Luche, Synthesis, 1989, 787.
5 O. Arrad and Y. Sasson, J. Am. Chem. Soc., 1988, 110, 185.
6 T. W. Bentley, R. V. H. Jones, A. H. Larder and S. J. Lock, J. Chem.
Soc., Chem. Commum., 1994, 2309.
Quaternary ammonium salts (Aldrich) were also used as
supplied, so probably contained traces of water. Methanol
and acetonitrile (Fisons HPLC grade), tert-butyl alcohol and
dimethyl sulfoxide, PEG 200 and PEG 6000 were used without
further purification.
Analytical methods
HPLC analyses were made on a 15 cm × 0.25Љ Spherisorb S5
ODS2 stainless steel column, eluted with 55% methanol–water
for analyses of 1–3 (λ set to 225 nm) and with 30% acetonitrile–
water for analyses of 4 and the (clearly separated) E–Z mixture
6 (λ set to 275 nm); to complete analyses within 10 min under
our isocratic conditions, flow rates (LDC Milton Roy Consta-
metric 3200 pump) of 2 ml minϪ1 for 1–3 and of 1ml minϪ1 for 4
and 6 were required; the absorbance range was usually set to 0.2
(Cecil CE2112 monitor). Samples were injected in 10 µl aliquots
using a Perkin Elmer ISS 101 autosampler, and integration was
done using a Hewlett Packard 3395 integrator plotter. A range
of standard solutions of pure samples in acetonitrile (ca. 10Ϫ3
) were used to obtain linear response/concentration plots for
each compound of interest.
7 E. V. Dehmlow and S. S. Dehmlow, Phase Transfer Catalysis, 3rd
edn., VCH, Weinheim, 1993.
8 C. M. Starks, C. L. Liotta and M. Halpern, Phase Transfer
Catalysis, Fundamentals, Applications, and Industrial Perspectives,
Chapman and Hall, New York, 1994, pp. 113–119.
9 (a) Y. Sasson, O. Arrad, S. Dermiek, H. A. Zahalka, M. Weiss and
H. Weiner, Mol. Cryst. Liq. Cryst., 1988, 161, 495; (b) C. L. Liotta,
J. Berkner, J. Wright and B. Fair, in Phase Transfer Catalysis;
Mechanisms and Syntheses, ACS Symposium Series, 1996, 659, 29;
(c) E. V. Dehmlow and H.-C. Raths, J. Chem. Res., 1988, 384.
10 E. Borredon, F. Clavellinas, M. Delmas, A. Gaset and J. V.
Sinisterra, J. Org. Chem., 1990, 55, 501.
11 R. N. Lewis and J. R. Wright, J. Am. Chem. Soc., 1952, 74, 1257.
12 S. J. Lock, Ph.D. thesis, University of Wales, 1993.
13 Determined by M. Boras (Swansea, 1997).
14 F. G. Bordwell, Acc. Chem. Res., 1988, 21, 456.
J. Chem. Soc., Perkin Trans. 2, 1998
93