Salt effects on the Baylis-Hillman reaction

Article abstract of DOI:10.1016/S0040-4020(03)00760-9

The Baylis-Hillman reaction is shown to accelerate in salt solutions of water and the 'water-like' structured solvents, like formamide and N-methylformamide in the presence of DABCO. Ethylene glycol, another structured solvent and its salt solutions fail

Full text of DOI:10.1016/S0040-4020(03)00760-9

Tetrahedron 59 (2003) 5019–5026
Salt effects on the Baylis–Hillman reaction
*
Anil Kumar and Sanjay S. Pawar
Physical Chemistry Division, National Chemical Laboratory, Pune 411 008, India
Received 18 February 2003; revised 11 April 2003; accepted 8 May 2003
Abstract—The Baylis–Hillman reaction is shown to accelerate in salt solutions of water and the ‘water-like’ structured solvents, like
formamide and N-methylformamide in the presence of DABCO. Ethylene glycol, another structured solvent and its salt solutions fail to make
any impact on the reaction rates. The salts that are conventionally defined as salting-out or -in do not behave in a similar fashion, when
employed in the Baylis–Hillman reactions. The results are supported by solubility measurements. It seems that the cation, anion, nature of
solvent and of reactants together ascertain whether a salt will enhance or retard the Baylis–Hillman reaction. q 2003 Elsevier Science Ltd.
All rights reserved.
1. Introduction
phenomena. In fact, GnCl proved to be a more effective
rate-enhancing agent than LiCl for the reaction of
cyclohexenone with benzaldehyde. This effect in the
Baylis–Hillman reactions is in contrast with what is
observed in and benzoin condensation13,14 and Diels–
Alder reactions.15 Recently, we have described the salting
aspect in Diels–Alder reactions. In general, it is believed
that the salts that enhance the hydrophobic effect are the
salting-out agents. The hydrophobic effect is reduced in
the presence of the salting-in agents. In the case of Diels–
Alder reactions, rate enhancement and decrease in salt
solutions can be correlated with salting-out and -in
phenomena.13,15
The Baylis–Hillman reaction (Scheme 1), one of the most
important carbon–carbon bond-forming processes, is
known to be very sluggish.1–5 Several reports are available
in the literature describing various attempts made to
accelerate this reaction.6–11 The studies have shown that
water and other additives accelerate the reaction. In a very
recent study, Aggarwal and co-workers demonstrated that
the salt effects on the kinetics of these reactions could not be
explained in terms of hydrophobic phenomena using
salting-out and -in processes.12 They further concluded
that the hydrogen bonding in the ‘water-like’ solvents like
formamide and N-methylformamide (NMF) played a
dominant role over the hydrophobic effect. In order to
reach this conclusion, they carried out the reaction of
cyclohexenone with benzaldehyde (in amine catalyst) in the
presence of guanidinium chloride (GnCl) and LiCl and
noted that both GnCl and LiCl increased the reaction rates
as compared to those in 1,4-diazabicyclo[2.2.2]octane
(abbreviated as DABCO) and 3-hydroxyquinuclidine
(abbreviated as 3-HDQ). GnCl is known to reduce the
reaction rates owing to its salting-in tendency. On the other
hand, LiCl enhances the rates as a result of salting-out
As a part of our research program on delineating the origin
of forces responsible for rate enhancement of organic
reactions, we attempt to seek answers to the following
questions with regard to the Baylis–Hillman reactions: (1)
What is the effect of the salt solutions in water and other
structured solvents on the Baylis–Hillman reactions? (2)
How do the salts alter the rates of the Baylis–Hillman
reaction? (3) If the salting effect plays governing role in
controlling the Baylis–Hillman reactions, does the defi-
nition of salting effect remain uniform throughout as defined
conventionally? or more specifically, does the solvent
medium in conjunction with the nature of reactants decide
whether a salt will enhance or inhibit the rates of the
Baylis–Hillman reactions?, and (4) If the salting effect is
significant in the Baylis–Hillman reactions, is it supported
by solubility measurements?
2. Results and discussion
Scheme 1. Baylis–Hillman reaction.
We first tried the reaction of benzaldehyde with methyl
acrylate in the presence of DABCO. The results are shown
in Table 1. The reaction took 19 h to complete with 65%
Keywords: Baylis–Hillman reaction; salt effect; solubility.
*
Corresponding author. Tel./fax: þ91-20-5893044;
e-mail: akumar@ems.ncl.res.in
0040–4020/03/$ - see front matter q 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4020(03)00760-9

5020
A. Kumar, S. S. Pawar / Tetrahedron 59 (2003) 5019–5026
Table 1. The Baylis–Hillman reaction of benzaldehyde with methyl acrylate in different salt solutions
Entry
Solvent mediaa
DABCO
Water
Timeb (h)
Yieldc (%)
Entry
Solvent media
Timeb (h)
Yieldc (%)
1
2
3
4
5
6
7
8
9
19
12
11
12
15
10
8
65
68
70
70
55
66
67
75
74
10
11
12
13
LiCl–NMF (1 M)
LiCl–NMF (4 M)
LiCl–EG (1 M)
LiCl–EG (4 M)
LiClO4–water
LiClO4–water (1 M)
LiClO4–forma. (1 M)*
LiClO4 –NMF (1 M)
LiClO4 –EG (1 M)
8
6
13
9
22
9
10
10
12
77
75
57
59
61
74
72
72
73
Forma.d
NMF
EGe LiCl–water
LiCl–water (1 M)
(4 M)
14
15
16
17
LiCl–Forma. (1 M)
LiCl–Forma. (4 M)
7
5
a
Contains DABCO.
No further increase in yield after the reported time.
Isolated yields.
Formamide.
Ethylene glycol.
b
c
d
e
isolated yield. The same reaction in water offered a yield of
68% in 12 h. However, when the reaction was carried out in
formamide, a yield of 70% was obtained in 11 h. Similarly,
70% yield in 12 h was obtained in NMF. On the other hand,
this reaction when carried out in ethylene glycol (EG)
offered 55% yield in 15 h. EG promoted the reaction as
compared to pure DABCO, but the yield in EG was lower
than in DABCO alone (Table 1, entries 1 and 5). The
reaction of benzaldehyde with acrylonitrile was reported to
be faster (7–8 h, 90% yield) in water, formamide and EG as
compared to in other solvents like methanol (34 h),
N-methylacetamide (48 h), DMSO, DMF, neat (3–5
days), THF and toluene (1 week).6 However, such a rate
acceleration is not observed for the reaction of benzal-
dehyde with methyl acrylate in EG (Table 1, entry 5),
though EG like water and formamide falls in the category of
the ‘structured solvents’. In fact, this reaction is slower in
EG with 20% less yield than achieved in water. The slowing
down of the reaction of benzaldehyde with methyl acrylate
as compared to that with acrylonitrile in EG suggests that
the effect of a solvent on reaction rates will depend not only
on the nature of solute and solvent, but also on the reactants.
Aggarwal and co-workers pointed out that the hydrophobic
effects and solvent polarity in such solvents might not be the
sole reason to account for the rate enhancement noted in the
Baylis–Hillman reaction and the hydrogen bonding in the
structured solvents played important role in accelerating
these reactions.12 The favorable conditions for the Baylis–
Hillman reactions in formamide were attributed to the
hydrogen bonding. We then examined the effect of salt
solutions prepared in the above mentioned solvents on this
reaction. This was guided by the observation that salts could
alter the hydrophobic effect in Diels–Alder reactions.13 The
reaction of benzaldehyde with methyl acrylate in 1 M
aqueous LiCl gave 66% yield in 10 h. The reaction time of
10 h was reduced to 8 h with 67% yield in a 4 M aqueous
LiCl, thereby accelerating the reaction. The use of 1 and
4 M LiCl solutions in formamide yielded 75 and 74% yields
in 7 and 5 h, respectively. It is clear that LiCl solution in
formamide is a powerful additive for achieving the rate
acceleration because of high isolated yields obtained in
relatively short time.
LiCl in NMF can accelerate the reaction by a factor of 2 as
compared to in NMF alone. However, we did not observe
any noticeable changes in the yields (variation by 2–4%
only) when the reactions were carried in the EG solutions of
LiCl. Though the reactions became faster (from 15 h in EG
to 9 h in 4 M LiCl solution of EG) it did not give any change
in the yields.
Later, we investigated the influence of LiClO4 (a salting-in
agent in water) on the above reaction. We obtained 61%
product in much higher time of 22 h in 1 M aqueous LiClO4.
This proves that aqueous LiClO4 acts as a salting-in agent
for this reaction by slowing it down as compared to in water
(68%, 12 h), as it does in many other Diels–Alder reactions,
like those of cyclopentadiene with methyl vinyl ketone16
and with methyl acrylate17 in aqueous environment. We
noted a 74% yield in 9 h in 1 M solution of LiClO4 in
formamide. The yields do not vary (72%) in the LiClO4
solutions of NMF and of EG. The reaction in the solution of
LiClO4 in NMF was, however, faster than that of LiClO4 in
EG (Table 1, entries 16 and 17).
Let us now comment on the effect of solvents and their salt
solutions on Diels–Alder reactions with an objective to
examine the similarities between the Baylis–Hillman
reactions and Diels–Alder reactions. The rate enhancement
of Diels–Alder reactions in formamide is possibly due to
solvophobic interactions, as formamide is a structured
solvent with higher dielectric constant (111) as compared to
that of water (78.4).16 The cycloaddition of 1,3-cyclo-
hexadiene with nitrosobenzene was faster in formamide and
EG than in other organic solvents.16 Water, of course is
proven to be the most effective rate-promoting solvents. The
reaction of cyclopentadiene with methyl vinyl ketone was
accelerated in solvents in the order of water.ethylene
glycol.formamide.methanol.isooctane.16 The rates of
this reaction decreased with the addition of LiCl in
formamide, while aqueous LiCl enhanced the reaction
rate. On the other hand, the formamide solutions of LiClO4
promoted the reaction. Similar effects could be seen if the
reaction was carried in the solutions of these salts in EG. We
wish to mention here that the reaction of cyclopentadiene
with methyl vinyl ketone was faster in the EG solutions of
LiCl, LiClO4 and GnCl, though GnCl was less effective salt
than LiClO4. It should be emphasized here that again a salt
plays a dual role of either salting-out or -in depending upon
The reaction reached completion in 8 h (77% yield) and 6 h
(75% yield) in 1 and 4 M solutions of LiCl in NMF,
respectively. It is interesting to note that a 4 M solution of

A. Kumar, S. S. Pawar / Tetrahedron 59 (2003) 5019–5026
5021
the solvents. Similar results were observed in these salt
solutions prepared in formamide. The commonly observed
salting-in agents (LiClO4 and GnCl) that inhibit the rates of
Diels–Alder reactions in water promote the reaction when
carried in their solutions prepared in EG and formamide by
smaller magnitude than water. In terms of salting phenom-
ena, it seems that the solubilities of diene and dienophile
remain unaltered both in these solvents and their salt
solutions. The relevant theory of salting phenomena is
described elsewhere.18–20
agents for this reaction. To support these findings, we
decided to examine the solubility of benzaldehyde in the salt
solutions without and with DABCO (Table 2). The
solubility of benzaldehyde (in the absence of DABCO) is
120, 63, 135 and 84 mM in formamide, water, NMF and
EG, respectively. In aqueous LiCl containing DABCO, the
solubility of benzaldehyde reduces to 35 mM resulting in
the salting out of benzaldehyde. There is a marginal increase
in the solubility of benzaldehyde in aqueous LiClO4
(60 mM) as compared to that in water (54 mM) all in the
presence of DABCO. Benzaldehyde is precipitated out in
the presence of formamide solution of LiCl, as the solubility
of benzaldehyde is lowered from 101 to 85 mM. The
lowering of solubility of benzaldehyde in formamide
solution of LiClO4 from 101 to 60 mM is another example
of the salting-out effect by LiClO4. The solubility of
benzaldehyde also reduces in the solutions of LiCl in NMF.
However, we did not notice any change in the solubility of
benzaldehyde in the LiCl–EG solutions, as it was in the
range of 81–85 mM. Neither LiCl nor LiClO4 in spite of
strong solvating power of Liþ or large sized ClO42 fail to
alter the solubility of benzaldehyde in EG thus indicating
insensitivity of EG towards LiCl and LiClO4. This suggests
that the solubility can be a vital probe in determining
whether the rate variation in the Baylis–Hillman reaction is
an outcome of the salting-out or -in process, as the rate
appears is inversely proportional to the reactant solubility.
The salting phenomena in reactions are directly related to
the solubility of reactants in salt solutions. The salts that
increase the reaction rates also promote the hydrophobic
effect. Breslow and Connors called these salts as pro-
hydrophobic and those, which inhibit the rates as anti-
hydrophobic ones.21 Thus, the salting phenomena, solubility
of reactants and hydrophobic effect, are interrelated. At
molecular level, ion-solvent interactions in conjunction with
the nature of reactant molecule determine whether a salt will
enhance or lower the reactant solubility. If a salt is dissolved
in water, there is a volume contraction called electro-
striction, as water collapses around the ions to solvate
them.17 This results in less empty space for a reactant. The
energy cost to create space for a reactant is greater. We can
imagine this as an energy cost of cavitation producing a hole
of the reactant size in water. As the energy cost of cavitation
increases due to electrostriction, the reactant precipitates out
due to its lower solubility in the salt solution as compared to
in water resulting into the rate enhancement. The salting-in
effect is opposite in its physical process. In this case, the
reactant becomes more soluble in the presence of a salt.
These salts break the structural arrangement of water
molecules making cavitation easier. This improves the
solvation of the reactant in salt solutions. The reactant
therefore, is not available for the reaction and hence the rate
is lowered in the presence of such salts. The role of a salt in
aqueous environment can also be extended to non-aqueous
solvents in a similar fashion. The free energy of solution can
be computed from solubility of solute in a solvent. The
solubility of solute changes on addition of salt and thereby
the Gibbs free energy of solution. The solubility of a
reactant in the presence of salt over that in the absence is
directly proportional to the change in Gibbs free energy in
this process.22
We earlier noted that this reaction in EG remained sluggish
and offered comparatively poor yield of about 55%. This
yield did not improve in both LiCl and LiClO4. These
experiments in EG show that the salt solutions in EG cannot
improve the reaction conditions.
It is interesting to note that the solubility of benzaldehyde is
lower in LiCl with DABCO than without DABCO
suggesting that LiCl with DABCO is a salting-out agent
and hence the reaction rate increases in LiCl. It is surprising
to observe that GnCl, which is a salting-in agent in water for
several Diels–Alder reactions,13,15,17 behaves in an entirely
opposite manner in this case. GnCl in conjunction with
DABCO suppresses the solubility of benzaldehyde, when
compared without DABCO. Thus, GnCl for this reaction
acts as a salting-out agent rather than salting-in one. We
have recently shown in connection with Diels–Alder
reaction that a given salt changes its salting-in property to
the salting-out or vice versa in conjunction with the solvent
in which its solution is made.23 We confirmed the dual role
of a given salt in two different solvents possessing very large
difference in their dielectric constants by carrying out
solubility measurements and calculations of salting coeffi-
The recent results on the reaction of cyclohexenone with
benzaldehyde with GnCl and LiCl show that both these salts
act as rate-enhancing agents, though in principle, the former
is expected to inhibit the reaction rates and latter enhance
them.12 Both these salts therefore, act as the salting-out
Table 2. Solubility of benzaldehyde in different media; concentration of salt in each solution¼1 M; DABCO¼1 mmol
Entry
Mediuma
Solubility (mM)
Entry
Mediuma
Solubility (mM)
1
2
3
4
5
6
Water
LiCl–water
LiClO4–water
Formamide
LiCl–formamide
LiClO4–formamide
54 (1)b
35 (0.4)
60 (1)
101 (1)
85 (0.5)
60 (1)
7
8
9
10
11
12
NMF
LiCl–NMF
LiClO4–NMF
Ethylene glycol
LiCl–ethylene glycol
LiClO4–ethylene glycol
110 (1.2)
97 (1)
91 (0.8)
85 (0.4)
83 (0.4)
81 (0.5)
a
Contains DABCO.
Value in parenthesis is the average deviation in solubility calculated from triplicate measurements.
b

5022
A. Kumar, S. S. Pawar / Tetrahedron 59 (2003) 5019–5026
influence of MgCl2 and CaCl2 on the reaction rates. The
same reaction took 11 (in 1 M) to 23 h (in 4 M) of LiCl with
80% yield in good agreement with the literature values.6
The rate-retarding effects were noticed in NaCl, LiBr and
NaBr. CsI was a poor rate-inhibiting salt (from 1 to 4 M),
where the reaction gave 88% yield in 8–11 h. Both KCl and
KI did not exhibit any effect on the reaction; the reaction
went in about 8 h with 90% product comparable to in water
alone.
On the other hand, we obtained 93% yield in 4 M salts of
LiI, LiClO4 and NaI with rate almost doubled (reaction time
3–5 h) in 4 M salts. The reaction in a 2 M Na2SO4 solution
gave 94% yield in about 5 h. The rate-enhancing capability
of these salts varies in the following order: Na2SO4.
LiI.LiClO4.NaI.KCl<KI.CsI.NaBr.LiBr.NaCl.
LiCl.CaCl2.MgCl2.
Figure 1. Time (h) as a function of the salt concentration, [salt] in water for
the reaction of benzaldehyde and acrylonitrile, (a) (B) MgCl2, (W) CaCl2,
(A) LiCl, (O) NaCl, (X) LiBr, (L) NaBr, (£) CsI; (b) (K) KI, (W) KCl, (P)
NaI, (B) LiClO4, (S) LiI, ( p ) Na2SO4.
cients. For example, LiClO4 in water is a salting-in agent,
but becomes a salting-out when dissolved in diethyl ether.
This shows the significance of the solvent (water or diethyl
ether) in which LiClO4 was dissolved to make its solution.
An example of the salting-in effect (rate-inhibiting) by
LiClO4–water was shown by Breslow and co-workers13 in
the case of Diels–Alder reactions. On the other hand,
Grieco and his group have established the rate enhancement
of several sluggish Diels–Alder reactions in LiClO4–
diethyl ether.24 This can be attributed to the salting-out
effect. Other effects, like viscosity, clatherates and Lewis
acid catalysis are also possible.25
The reaction goes faster in the salts having a common
cation, but with increasing size of their anions. For example,
the rate varies as: LiCl,LiBr,LiI; NaCl,NaBr,NaI.
In this connection, the work of Rizzo on the Diels–Alder
reaction of anthracene-9-carbinol with N-ethylmaleimide in
2 M solutions of sodium and guanidinium salts with
different anions is of relevance.26 The reaction rates,
which decreased with the increasing anionic radii, followed
Hofmeister series.27,28 The rates of the Baylis–Hillman
reaction between benzaldehyde and acrylonitrile, however
increases with the increase in anionic radii as seen above.
This is an important point of the current investigation. It is
therefore, clear that a classification of salts into salting-out
and -in additives cannot be made on universal basis. For this
reaction, MgCl2, CaCl2, LiCl, NaCl, LiBr, NaBr and CsI are
the salting-in agents, while NaI, LiClO4, LiI and Na2SO4 the
salting-out ones. Otherwise, MgCl2, CaCl2, LiCl, LiI, NaCl,
NaBr, NaI and Na2SO4 are the conventional salting-out
agents in water. LiBr, CsI and LiClO4 are known to be the
salting-in agents in water.
Encouraged by the above mentioned salt effects on the rates
of the Baylis–Hillman reaction between benzaldehyde and
methyl acrylate, we then examined the reaction of
benzaldehyde with acrylonitrile in the presence of salts.
Auge and co-workers showed that this reaction was faster in
LiI and NaI as compared to in water and other organic
solvents.6 LiCl and CsI slowed the reaction, while the
reaction remained unaffected by KI. The behavior of these
salts on the reaction is not uniform considering the nature of
ions. We therefore, decided to examine this effect in detail.
We performed this reaction in several concentrations of
aqueous LiCl, LiBr, LiI, LiClO4, NaCl, NaBr, NaI, KCl, KI,
CsI, MgCl2, CaCl2 and Na2SO4. DABCO was used in each
reaction. The results are shown in Figure 1(a) and (b), where
time is plotted as a function of the salt concentration. This
reaction in water gave 90% yield in 8 h. The reaction in
MgCl2 and CaCl2 was very slow with low yields. In a 2 M
solution each of MgCl2 and CaCl2, for example, it took 19
and 16 h, respectively with ,70% product. We were
surprised to see that the reaction became slow with the
increase in the salt concentration showing negative
We thought it appropriate to investigate the effect of salts on
the Baylis–Hillman reaction between cyclohexenone and a
deactivated and hindered aldehyde. o-Anisaldehyde is one
of such aldehydes. The reaction of o-anisaldehyde with
cyclohexenone in the presence of DABCO gave 68% yield
in 27 h as compared to in 3-HDQ (74%, 24 h).12 Table 3
summarizes these results.
This reaction in water gave 68% product in 27 h (Table 3,
entry 1). However, a 68% product was obtained in 21 h in
Table 3. The reaction of cyclohexenone with o-anisaldehyde in 1 M salts in different solvents
Entry
Solvent mediaa
Timeb (h)
Yieldc (%)
Entry
Solvent media
Timeb (h)
Yieldc (%)
1
2
3
4
5
6
Water
27
21
27
32
20
18
68
68
67
65
69
71
7
8
9
10
11
12
KCl–forma
CaCl2–forma.d
EG
NaI–EG
KCl–EG
CaCl2–EG
26
22
28
29
28
30
68
69
60
62
60
59
NaI–water
KCl–water
CaCl2–waterd
Forma.
NaI–forma.
a
Contains DABCO.
No further increase in yield after the reported time.
Isolated yields.
0.34 M salt.
b
c
d

A. Kumar, S. S. Pawar / Tetrahedron 59 (2003) 5019–5026
5023
NaI–water accelerating the reaction. Aqueous KCl did not
exhibit any effect on the course of the reaction (Table 3,
entry 3), as both the yield and time remained unaltered with
respect to those noted in water alone. This reaction was
further slowed down in aq. CaCl2 and gave 65% yield in
32 h. Thus, NaI acted like the salting-out agent, while CaCl2
as the salting-in. Formamide, another structured solvent like
water accelerated the reaction giving 69% yield in 20 h. The
formamide solution of NaI was also effective in speeding
the reaction with 71% product in 18 h. The performance of
the CaCl2–formamide solution was better (22 h, 69%)
(Table 3, entry 8) as compared to that in aqueous CaCl2.
Again the reaction rates did not improve in the solution of
KCl in formamide (Table 3, entry 7). NaI is an effective
salting-out agent both in water and formamide. However,
aqueous CaCl2 which is a salting-in agent, changed into a
salting-out in the formamide solution. This reaction was
quite slow in EG with 60% yield in 28 h. It is important to
note that there were no apparent changes in reaction times
and yields in the solutions of NaI, KCl and CaCl2 in EG
(Table 3, entries 10–12).
affected by EG or by the solutions of LiCl, LiClO4 and NaI
prepared in EG (Table 4, entries 11–14).
As pointed earlier, the salting-out and -in phenomena are
related to ion-solvent interactions suggesting that the
magnitude of ion-solvent interactions can help in rationaliz-
ing the rate variations in reactions. The ions and their salts
can be classified as the structure-makers or the structure-
breakers. One of the most important criteria to estimate
these interactions is the analysis of viscosity data, which can
be employed to extract the ion-solvent interaction parameter
popularly called as the ‘B-coefficients’. The B-coefficient of
a salt can be divided into corresponding ionic B values. If a
B value for a species (ion or its salt) is positive, the salt is
termed as structure-maker. The structure-making and
-breaking terms are meant to describe the changes in
arrangement of water molecules around an ion and it
depends upon the ionic radius, ionic charge and dielectric
constant of water or in more general sense of a solvent. The
B-coefficient of salt is an additive property and is the sum of
the ionic B values. This term does not imply a making or
breaking of a chemical species by chemical method. The
structure-making and -breaking phenomena are purely
physical in nature. The solubility changes in salt solutions
can also rationalized in terms of the B-coefficients of salts.
The B-coefficients of these salts in water as taken from the
literature29 are: Na2SO4 (20.034), LiI (þ0.067), LiClO4
(þ0.087), NaI (þ0.006), KCl (20.002), KI (20.089), CsI
(20.127), NaBr (þ0.054), LiBr (þ0.115), NaCl (þ0.079),
LiCl (þ0.140), CaCl2 (þ0.284) and MgCl2 (þ0.371).
Though we have demonstrated use of this coefficient in
correlating the kinetic data of Diels–Alder reactions, we are
unable to trace such a correlation in the case of the Baylis–
Hillman reaction. In the present study, it seems that the salts
with very high B-coefficients (for example, MgCl2 and
CaCl2) slow down the reaction, though with exception of
NaCl, LiCl and LiBr. The salts with negative B values in
general, promote the reaction. KCl is the only salt studied in
this work which has its B value as 20.002 suggesting KCl
cannot disturb the arrangement of solvent molecules with its
cationic charge. Conventionally, KCl with a small negative
B value is considered as a poor structure breaker. From the
experimental results obtained on the above reactions, it is
clear that irrespective of nature of reactants and solvents,
KCl fails to influence the kinetics of the Baylis–Hillman
reactions. In non-aqueous solvents like formamide, which
accelerates the Baylis–Hillman reaction (Table 1, entry 3),
the B-coefficient of LiCl from the literature viscosity data is
20.043. A change in the B-coefficient of LiCl from water
It is of interest to check how the salts can influence a
Baylis–Hillman reaction containing a more reactive
aldehyde than o-anisaldehyde. We therefore tested the salt
effects for the reaction between p-anisaldehyde with ethyl
acrylate, as p-anisaldehyde is more reactive than o-anisal-
dehyde. The results are summarized in Table 4. This
reaction gave 38% product in 80 h in water alone. When this
reaction was carried in aqueous LiClO4 and LiCl, the
reaction further slowed down with lower yields to be 19 and
15%, respectively in about 100 h as compared to in water
alone. NaI showed very interesting effect by accelerating the
reaction to give 50% product in 48 h (Table 4, entry 4),
when compared to those in LiCl and LiClO4 salts. Like LiCl
and LiClO4, NaCl did not improve the kinetics of this
reaction (Table 4, entry 5). As compared to in formamide,
its solutions with LiClO4 and LiCl improved the reaction
conditions by reducing the times to complete the reactions.
This reaction did not go well in diethyl ether giving 2%
product in 96 h. However, a 5 M LiClO4–diethyl ether
(LPDE) was noted to make a difference by yielding 50%
product in 48 h (Table 4, entry 10). Use of LPDE in
accelerating the Baylis–Hillman reaction of benzaldehyde
with methyl acrylate has recently been demonstrated in
which 70 mol% of LiClO4 was used with 15 mol% of
DABCO for the molar ratio of benzaldehyde/methyl
acrylate as 0.83.9 This reaction condition offered 81%
yield. Once again, we noted that the reaction was not
Table 4. The reaction of p-anisaldehyde with ethyl acrylate in 1 M salts in different solvents
Entry
Solvent mediaa
Timeb (h)
Yieldc (%)
Entry
Solvent media
Timeb (h)
Yieldc (%)
1
2
3
4
5
6
7
Water
80
110
96
48
85
38
19
15
50
30
35
45
8
9
LiCl–forma
Diethyl ether
LPDEd
58
96
48
101
98
105
95
48
2
LiClO4–water
LiCl–water
NaI–water
NaCl–water
Forma.
10
11
12
13
14
50
22
23
24
22
EG
LiCl–EG
LiClO4–EG
NaI–EG
72
62
LiClO4–forma
a
Contains DABCO.
b
c
d
No further increase in yield after the reported time.
Isolated yields.
LiClO4–diethyl ether, 5 M.

5024
A. Kumar, S. S. Pawar / Tetrahedron 59 (2003) 5019–5026
(þ0.140) to that in formamide (20.043) indicates that LiCl
which was a structure-maker changed to structure-breaker
when solvent was changed from water to formamide. LiCl
in NMF has a value of B-coefficient as 20.029 showing
again a strong change in the structure-altering ability of LiCl
in water to NMF.30 Analysis of viscosity data of LiCl in EG
does not represent a noticeable change in its B value
(þ0.045). This suggests that LiCl act as a structure-maker in
EG as it does in water, but its tendency to alter the solvent
structure reduces by about 3 times. The above observations
indicate that the structure-maker salts inhibit the reaction
rates of the Baylis–Hillman reactions, the strong structure
breakers enhance them. The salts that have poor or very
weak influence on the arrangement of the solvent molecules
seem to exhibit negligible effect on the progress of these
reactions. We wish to state here such a classification of salts
is still vague when seen in the perspective of the Baylis–
Hillman reactions and calls for more physical-organic
investigation.
Figure 2. Time versus the mixture composition (% v/v) of mixed salts. (B)
NaCl–NaI, (K) LiCl–LiClO4.
(8 h, 90% yield). This finding confirms that the changes in
the ion–water interactions due to NaCl were neutralized by
those due to NaI, when mixed in appropriate combination.
This is also true in the case of LiCl–LiClO4 system. These
experiments suggest that it is possible to mimic the water-
like solvent conditions even in aqueous mixed salt solutions
as a result of neutralization of the ion–water interactions of
the opposite nature.
We have recently shown that Naþ, a hydrophilic cation
neutralizes the effect produced by guanidinium (Gnþ) cation
in aqueous environment.31 We studied this aspect to probe
the origin of forces responsible for the rate enhancement and
inhibition of Diels–Alder reaction in NaCl and GnCl by
appropriate manipulation of ionic concentrations of Naþ
and Gnþ for producing a water-like reaction medium.32 For
example, the rate of reaction of cyclopentadiene with
methyl acrylate is reduced in the presence of 1 M GnCl
relative to that in water. The rate inhibiting effect of GnCl
can be neutralized by NaCl as shown in the case of Diels–
Alder reactions and further supported by the solution data. If
we want to neutralize the rate-inhibiting effect of GnCl, how
much NaCl will be required so that the rate does not reduce
and becomes equal to that noted in water alone? We
discussed these results in terms of salting and structure
altering properties of these salts. In Diels–Alder reactions,
these effects can be easily ascribed to ion–water inter-
actions among other factors. In nutshell, the hydrophilic
tendency of Naþ is neutralized by the hydrophobic nature of
Gnþ species. We examined this issue for the reaction of
benzaldehyde with acrylonitrile in the mixtures of salts that
gave opposite effects. As seen above, this reaction goes
faster in NaI (4 h) and LiClO4 (3 h) and becomes slow in
NaCl (19 h) and LiCl (23 h). We performed this reaction in
several compositions of the NaCl–NaI and LiCl–LiClO4
mixtures. The results are shown in Figure 2.
3. Conclusion
The findings of the above investigation based on the chosen
reactions are summarized below:
1. The salt solutions have pronounced effect on the rates of
the Baylis–Hillman reactions in the presence of
DABCO. Water, formamide and NMF and their salt
solutions promote the Baylis–Hillman reactions.
2. EG and its salt solutions fail to influence the Baylis–
Hillman reactions.
3. The rate enhancement and inhibition in the presence of
salts can be ascribed to the salting-out and -in
phenomena.
4. The salting effect in the Baylis–Hillman reactions does
not conform to conventional definitions of the salting-out
and -in phenomena as noted in the case of Diels–Alder
reactions. The cause of dual role of the salts on rates of
the Baylis–Hillman reactions is yet to be ascertained.
5. The salting effect has been supported by the experimental
solubility data.
This reaction becomes faster in the NaCl–NaI mixtures rich
in NaI. There is an acceleration of the reaction with increase
in NaI in NaCl. Similarly, a monotonous rate enhancement
is witnessed in the LiCl–LiClO4 mixtures. Again the
reaction is faster in the LiClO4-rich mixtures. As noted
above, this reaction is carried in water in 8 h. From Figure 2,
it is clear that a time of 8 h required to complete this reaction
can be obtained by the mixtures of 70% (v/v) NaI–30%
(v/v) NaCl. Similarly, a mixture of 55% (v/v) LiCl–45%
(v/v) LiClO4 offers a time of 8 h. In order to check this, we
carried out this reaction in two different mixtures: (1) 70%
NaI and (2) 55% LiClO4 mixtures with the specified
compositions. It was interesting to find that the reactions
completed in 8 h with 89% product in both the mixed salts,
which agrees well with the reaction carried in water alone
As a matter of fact, the Baylis–Hillman reactions are
accompanied by negative activation volumes.6 This
suggests that the rate variation in these reactions should
be related to the hydrophobic or solvophobic effect.
However, the results from the present studies do not
conform to this view in full agreement of the literature.12
How to quantify rates of the Baylis–Hillman reactions in
terms of these ion-interactions arising from different types
of ions is being currently investigated in our laboratory and
will be communicated in the near future. We believe that the
above study will be helpful in the synthetic aspects of

A. Kumar, S. S. Pawar / Tetrahedron 59 (2003) 5019–5026
5025
Baylis–Hillman reactions by manipulating solvent
conditions.
reported in this work is an average of three measurements
with deviations not exceeding 1.7% from mean solubility
value (see Table 2).
4. Experimental
Acknowledgements
4.1. General remarks
The authors acknowledge financial support received from
Director, NCL to carry out this work.
All the aldehydes and cyclohexenone were distilled before
use, while methyl acrylate, ethyl acrylate and acrylonitrile
were used as such. DABCO obtained commercially was
purified prior to use. Analytical reagents grade salts were
used in preparing their solutions. (Caution: LiClO4 is
potentially explosive and must be handled with care.)33 All
other solvents were of high purity and used as purchased.
Deionized water was used throughout the work.
References
1. Baylis, A. B.; Hillman, M. E. D. German Patent 2,155,113,
1972; Chem. Abstr. 1972, 77, 34174q.
2. Drewes, S. E.; Roos, G. H. P. Tetrahedron 1988, 44, 4653.
3. Basavaiah, D.; Rao, P. D.; Hyma, R. S. Tetrahedron 1996, 52,
8001.
The reaction procedures for carrying out these reactions
were followed as described in the literature. Table 5 gives
the reaction conditions used in this work.
4. Ciganek, E. Org. React. 1997, 51, 201.
5. Langer, P. Angew. Chem. Int. Ed. 2000, 39, 3049.
6. For using water and salts, see: Auge, J.; Lubin, N.; Lubineau,
A. Tetrahedron Lett. 1994, 35, 7947. This report also gives
references to other methods employed to speed up the Baylis–
Hillman reactions.
Table 5. General conditions for the Baylis–Hillman reactions
Reaction
Reactant
Reactant
1
2
3
4
Benzaldehyde (1 mmol)
Benzaldehyde (1 mmol)
o-Anisaldehyde (1 mmol)
p-Anisaldehyde (1 mmol)
Methyl acrylate (1.2 mmol)
Acrylonitrile (1.2 mmol)
Cyclohexenone (2 mmol)
Ethyl acrylate (1.2 mmol)
7. For the combined use of metal catalyst and co-ligand see:
Aggarwal, V. K.; Mereu, A.; Tarver, G. J.; McCaugue, R.
J. Org. Chem. 1998, 63, 7183.
8. For the use of more effective catalysts than DABCO, see:
Aggarwal, V. K.; Mereu, A. Chem. Commun. 1999, 2311.
9. For using LiClO4 –diethyl ether see: Kawamura, M.;
Kobayashi, S. Tetrahedron Lett. 1999, 40, 1539.
10. For the use of DABCO see: Yu, C.; Liu, B.; Hu, L. J. Org.
Chem. 2001, 66, 5413.
DABCO 1 mmol; solvent or salt solution 1 mL.
As a general procedure, DABCO was added to the stirred
mixture of reactants (Table 5) dissolved in a solvent. Diethyl
ether was used to stop the reaction. The mixture was washed
with 2 M HCl and then by water. The mixture was then
dried and handled in a usual way. The product was purified
by silica gel column chromatography and characterized by
1H NMR.6,9,12 All the products reported here have been
successfully characetrized in the literature and hence details
are not given here. The reaction mixture was heterogeneous
in water and its salt solutions. The reaction mixtures were
therefore stirred vigorously. In the case of LiClO4–diethyl
ether, 1N HCl was added to quench the reaction.
Dichloromethane was used to extract the aqueous layer.
The reported yields throughout the work are the isolated
yields.
11. For the use of catalyst see: Ameer, F.; Drewes, S. E.; Freese,
S.; Kaye, P. T. Synth. Commun. 1988, 18, 495.
12. For an exhaustive use of catalysts and optimum choice of
3-hydroxyquinuclidine in the Baylis–Hillman reactions, see:
Aggarwal, V. K.; Dean, D. K.; Mereu, A.; Williams, R. J. Org.
Chem. 2002, 67, 510.
13. Breslow, R. Acc. Chem. Res. 1991, 24, 159.
14. Garner, P. P. In Organic Synthesis in Water. Diels–Alder
Reactions in Aqueous Media; Grieco, P. A., Ed.; Blackie:
Glassgow, 1998; Chapter 1.
15. Kumar, A. Chem. Rev. 2001, 101, 1, and references cited
therein.
The solubilities were determined by equilibrating benzal-
dehyde in a solvent or its salt solutions for 3 h in a Julabo-
made constant temperature bath set at 25^0.18C with
vigorous agitation during the first 3 min. The lower phase
was diluted 20 times with solvent (50 mL to 1 mL of
solvent). The concentration of benzaldehyde was deter-
mined using a Varian UV–Visible spectrophotometer at
248 nm (in water), 255 nm (in formamide), 252 nm (in
NMF) and 242 nm (in EG). The concentration of benzal-
dehyde was measured relative to a prepared standard in the
same solvent. For example, the standard solution in the case
of water–salt–benzaldehyde system was water–benzal-
dehyde with identical definitions for other systems. When
the solubility of benzaldehyde was measured in the presence
of DABCO, the standard solution also contained DABCO.
The salt was noted to have negligible effect on the
absorbance of benzaldehyde in solvents. The solubility
16. Breslow, R.; Guo, T. J. Am. Chem. Soc. 1988, 110, 5613.
17. Pawar, S. S.; Phalgune, U. D.; Kumar, A. J. Org. Chem. 1999,
64, 7055.
18. Debye, P.; McAulay, J. Phys. Z. 1925, 26, 22.
19. Long, F. A.; McDevitt, F. W. Chem. Rev. 1952, 52, 119.
20. McDevitt, F. W.; Long, F. A. J. Am. Chem. Soc. 1952, 74,
1773.
21. Breslow, R.; Connors, R. V. J. Am. Chem. Soc. 1995, 117,
6601.
22. Breslow, R.; Connors, R. V.; Zhu, Z. Pure Appl. Chem. 1996,
68, 1527.
23. Kumar, A.; Phalgune, U. D.; Pawar, S. S. J. Phys. Org. Chem.
2001, 14, 577.
24. Grieco, P. A.; Nunes, J. J.; Gaul, M. D. J. Am. Chem. Soc.
1990, 112, 4595.
25. Kumar, A.; Pawar, S. S. J. Org. Chem. 2001, 66, 7646, and
references cited therein.

5026
A. Kumar, S. S. Pawar / Tetrahedron 59 (2003) 5019–5026
26. Rizzo, C. J. J. Org. Chem. 1992, 57, 6382.
27. Hofmeister, F. Ach. Exptl. Pathol. Pharmakol. 1888, 24, 247.
28. For details see: Collins, K. D.; Washabaugh, M. W. Q. Rev.
Biophys. 1985, 18, 323.
31. Kumar, A. J. Phys. Chem. B 2000, 104, 9505.
32. Kumar, A.; Phalgune, U. D.; Pawar, S. S. J. Phys. Org. Chem.
2002, 15, 131.
33. Schumacher, J. C. Perchlorates—Their Properties,
Manufacture and Uses; ACS Monographs Series, American
Chemical Society: Washington DC, 1960.
29. Marcus, Y. Ion Solvation; Wiley/Interscience: New York,
1986.
30. Kumar, A. Can. J. Chem. Engng 1993, 71, 948.