Retro-Diels-Alder Reaction in Aqueous Solution: Toward a Better Understanding of Organic Reactivity in Water

Article abstract of DOI:10.1021/jo9623200

The retro-Diels-Alder (RDA) reaction of anthracenedione 1a proceeds considerably faster in aqueous solutions than in organic solvents. Addition of organic solvents to water retards the reaction, whereas glucose induces a modest acceleration. SDS micelles induce a considerable retardation, but even at high concentrations of surfactant (complete micelle-substrate binding), the cycloreversion is not fully inhibited. Correlation with data for solvatochromic indicators strongly suggest that the origin of the water-induced acceleration involves primarily enhanced hydrogen bonding of water to the activated complex for the RDA reaction of 1a. Activation parameters support this view. A comparison of the present results with previous kinetic data for bimolecular and intramolecular Diels-Alder reactions provides insights into the contributions of hydrogen-bond and hydrophobic interactions to the aqueous accelerations of the latter two types of reactions.

Full text of DOI:10.1021/jo9623200

J . Org. Chem. 1997, 62, 2039-2044
2039
Retr o-Diels-Ald er Rea ction in Aqu eou s Solu tion : Tow a r d a Better
Un d er sta n d in g of Or ga n ic Rea ctivity in Wa ter
J an W. Wijnen and J an B. F. N. Engberts*
Department of Organic and Molecular Inorganic Chemistry, University of Groningen, Nijenborgh 4,
9747 AG Groningen, The Netherlands
Received December 11, 1996X
The retro-Diels-Alder (RDA) reaction of anthracenedione 1a proceeds considerably faster in aqueous
solutions than in organic solvents. Addition of organic solvents to water retards the reaction,
whereas glucose induces a modest acceleration. SDS micelles induce a considerable retardation,
but even at high concentrations of surfactant (complete micelle-substrate binding), the cyclo-
reversion is not fully inhibited. Correlation with data for solvatochromic indicators strongly suggest
that the origin of the water-induced acceleration involves primarily enhanced hydrogen bonding of
water to the activated complex for the RDA reaction of 1a . Activation parameters support this
view. A comparison of the present results with previous kinetic data for bimolecular and
intramolecular Diels-Alder reactions provides insights into the contributions of hydrogen-bond
and hydrophobic interactions to the aqueous accelerations of the latter two types of reactions.
In tr od u ction
hydrophobic hydration shells. Overlap of these hydration
shells leads to association of apolar species. However,
the strong focus on hydrophobic interactions often out-
shines other unique properties of water, which may also
affect organic reactivity. One of these properties is the
ability to act as a very efficient hydrogen-bond donor as
well as acceptor.8 The high acceptor number (AN), ET-
(30)-value, and Kamlet-Taft’s R-parameter show that
water is one of the best hydrogen-bond-donating solvents,
only surpassed by much more acidic solvents like formic
acid. Presumably, this quality is partly due to the small
size of the water molecule, which enables water to
interact effectively and multimolecularly with Lewis
bases. In water, organic reactions are often promoted
by these interactions.
The bimolecular Diels-Alder reaction has been thor-
oughly studied in aqueous solutions,4,5 and it has become
apparent that nearly all different types of this cycload-
dition benefit from an aqueous reaction medium. By now
it is widely accepted that two factors are primarily
responsible for the ‘aqueous acceleration’: (i) stabilization
of the transition state relative to the initial state due to
enhanced hydrogen bonding of water to the polarized
activated complex and (ii) substantial reduction of the
hydrophobic surface area of reactants during the activa-
tion process (‘enforced hydrophobic interactions’).4,9,10
Water is an unconventional solvent for organic reac-
tions. Besides obvious economical and environmental
advantages, water can have a surprisingly beneficial
effect on organic reactions, and this notion has popular-
ized water as a reaction medium.1 Presently almost
every type of organic reaction has been studied in water,
and a considerable number of these reactions are actually
promoted by an aqueous reaction medium, most notably
pericylic reactions like the Claisen rearrangement,2 1,3-
dipolar cycloadditions,3 and particularly the Diels-Alder
(DA) reaction.4,5 Successful attempts to catalyze organic
reactions in aqueous solutions have extended the poten-
tial of water as a reaction medium.6
By far the most intriguing issue associated with water
entails hydrophobic interactions.7 These rather complex
phenomena are governed by the limited ability of water
to dissolve apolar molecules through the formation of
X Abstract published in Advance ACS Abstracts, March 15, 1997.
(1) (a) Grieco, P. A. Aldrichim. Acta 1991, 24, 59. (b) Li, C. Chem.
Rev. 1993, 93, 2023. (c) Lubineau, A.; Auge´, J .; Queneau, Y. Synthesis
1994, 741.
(2) (a) Brandes, E.; Grieco, P. A.; Gajewski, J . J . J . Org. Chem. 1989,
54, 515. (b) Grieco, P. A.; Brandes, E. B.; McCann, S.; Clark, J . D. J .
Org. Chem. 1989, 54, 5849. (c) Severance, D. L.; J orgensen, W. L. J .
Am. Chem. Soc. 1992, 114, 10966. (d) Gao, J . J . Am. Chem. Soc. 1994,
116, 1563. (e) Gajewski, J . J .; Brichford, N. L. J . Am. Chem. Soc. 1994,
116, 3165.
(3) (a) Dignam, K. J .; Hegarty, A. F.; Quain, P. L. J . Org. Chem.
1978, 43, 388. (b) Inoue, Y.; Araki, K.; Shiraishi, S. Bull. Chem. Soc.
J pn. 1991, 64, 3079. (c) Rohloff, J . C.; Robinson, J .; Gardner, J . O.
Tetrahedron Lett. 1992, 33, 3113. (d) Wijnen, J . W.; Steiner, R. A.;
Engberts, J . B. F. N. Tetrahedron Lett. 1995, 35, 5389.
(4) (a) Blokzijl, W.; Blandamer, M. J .; Engberts, J . B. F. N. J . Am.
Chem. Soc. 1991, 113, 4241. (b) Blokzijl, W.; Engberts, J . B. F. N. J .
Am. Chem. Soc. 1992, 114, 5440. (c) Otto, S.; Blokzijl, W.; Engberts,
J . B. F. N. J . Org. Chem. 1994, 59, 5372. (d) Wijnen, J . W.; Zavarise,
S.; Engberts, J . B. F. N. J . Org. Chem. 1996, 61, 2001.
(5) (a) Rideout, D. C.; Breslow, R. J . Am. Chem. Soc. 1980, 102, 7816.
(b) Grieco, P. A.; Yoshida, K.; Garner, P. J . Org. Chem. 1983, 48, 3137.
(c) Waldmann, H.; Braun, M. Liebigs Ann. Chem. 1991, 1045. (d)
Lubineau, A.; Auge´, J .; Grand, E.; Lubin, N. Tetrahedron 1994, 50,
10265.
Since the debate on the origins of the water-induced
acceleration of DA reactions appears to be settled, the
discussion has shifted to the intriguing but rather
troublesome problem of separating and quantifying both
factors.4,9,10 This is a considerable challenge since nearly
all common DA reagents are hydrophobic and possess
activating substituents that are susceptible to hydrogen
bonding. The dimerization of cyclopentadiene looks like
a perfect example of a DA reaction in which hydrogen-
(8) Reichardt, C. Solvents and Solvent Effects in Organic Chemistry;
VCH: Cambridge, 1990.
(6) (a) Kalck, P.; Monteil, F. Adv. Organomet. Chem. 1992, 34, 219.
(b) Herrmann, W. A.; Kohlpainter, C. W. Angew. Chem., Int. Ed. Engl.
1993, 32, 1524. (c) Kobayashi, S. Synlett 1994, 689. (d) Papadogianakis,
G.; Sheldon, R. A. New J . Chem. 1996, 20, 175. (e) Otto, S.; Bertoncin,
F.; Engberts, J . B. F. N. J . Am. Chem. Soc. 1996, 118, 7002. (f) Loh,
T.-P.; Pei, J .; Lin, M. J . Chem. Soc., Chem. Commun. 1996, 2315.
(7) Blokzijl, W.; Engberts, J . B. F. N. Angew. Chem., Int. Ed. Engl.
1993, 32, 1545.
(9) (a) Blake, J . F.; J orgensen, W. L. J . Am. Chem. Soc. 1991, 113,
7430. (b) Blake, J . F.; Lim, D.; J orgensen, W. L. J . Org. Chem. 1994,
59, 803. (c) J orgensen, W. L.; Blake, J . F.; Lim, D.; Severance, D. L. J .
Chem. Soc. Faraday Trans. 1994, 90, 1727.
(10) Furlani, T. R.; Gao, J . J . Org. Chem. 1996, 61, 5492.
(11) (a) Sangwan, N. K.; Schneider, H.-J . J . Chem. Soc., Perkin
Trans. 2 1989, 1223. (b) Breslow, R.; Zhu, Z. J . Am. Chem. Soc. 1995,
117, 9923.
S0022-3263(96)02320-1 CCC: $14.00 © 1997 American Chemical Society

2040 J . Org. Chem., Vol. 62, No. 7, 1997
Wijnen and Engberts
Ta ble 1. F ir st-Or d er Ra te Con sta n ts a n d Gibbs En er gies
of Activa tion for th e RDA Rea ction of 1a in Or ga n ic
Solven ts a n d in Wa ter a t 40.0 °C
Sch em e 1
108 × k1
∆qGθ
entry
solvent
(s-1
)
(kJ mol-1
)
ET(30)a
ANb
1c
2c
3c
4c
5c
6
n-hexane
benzene
DMSO
2-PrOH
acetic acid
TFE
2.6
6.6
23
27
62
161
359
469
122.3
119.9
116.6
116.2
114.0
111.5
109.5
108.8
31.0
34.3
45.1
48.4
51.7
59.8
63.1
65.3
0
8.2
20.4
33.6
52.9
53.3
54.8
bond interactions can be neglected. Unfortunately ki-
netic studies are experimentally precarious, and indeed
there is some dispute about the exact rate constants.9c,11
Recently we have investigated another reaction in which
hydrogen-bond interactions are of minor importance: the
cycloaddition of cyclopentadiene to acridizinium bro-
mide.12 This reaction is only modestly accelerated in
water.
Computational methods have been successfully applied
to investigate the origins of the ‘aqueous acceleration’ of
cycloadditions.9,10 They indicate that hydrogen-bond
interactions often are the dominant contribution to the
rate enhancements in water.
In this paper we present the results of a kinetic study
of a reaction in which contributions of hydrophobic effects
are negligible, thus providing an opportunity to quantify
the influence of hydrogen bonding on cycloadditions. The
retro-Diels-Alder (RDA) reaction of 1a seems a suitable
candidate13 (Scheme 1). According to frontier molecular
orbital theory (FMO) RDA reactions are mechanistically
comparable to DA reactions. Thus, RDA reactions are
accelerated by electron-withdrawing substituents on the
(future) dienophile and electron-donating substituents on
the (future) diene.14,15 RDA reactions are dominated by
the enthalpy of activation, the entropy of activation
usually being small.13,14,16,17 But compared to DA reac-
tions, the volume of activation (∆qVθ) of a RDA is
markedly smaller, in fact, close to zero.13,17 Furthermore,
due to the unimolecular nature of the RDA reaction, no
substantial dehydration of the reactant has to take place
during the activation process (in contrast to bimolecular
reactions in aqueous solutions) and also aggregation of
reactants is excluded. This means that contributions of
‘enforced hydrophobic interactions’ are negligible, since
no significant change of solvent-accessible surface area
takes place during the activation process.
7
8
water
HFIP
a
b
From ref 8. Acceptor number; see ref 8. c Calculated from
data in ref 16.
are a major drawback as they may reduce the selectivity
of the reaction. Therefore several techniques have been
developed that have rendered this cycloreversion more
applicable. Flash vacuum thermolysis19 is most fre-
quently used, but also examples of acid catalysis,20
antibody catalysis,21 silica gel22 or alumina23 catalysis,
and Lewis acid catalysis24 are known. Grieco25 has
reported efficient water-promoted retro-aza DA reactions
and demonstrated that N-alkyl-2-azanorbornenes readily
decompose in water, whereas in organic solvents rigorous
reaction conditions are required.
In this study we present kinetic data on a homo-RDA
reaction in water and in mixed aqueous solutions. It will
be shown that water facilitates this reaction, despite the
fact that hydrophobic interactions are of negligible
importance. The efficient hydrogen-bond-donating ability
of water appears to govern the increased reactivity by
stabilizing the polarized activated complex.
Resu lts a n d Discu ssion
Kin etics of th e Cyclor ever sion in Wa ter a n d
Or ga n ic Solven ts. Rate constants for the cyclorever-
sion of 1a were determined at 40.0 °C in water and in
several fluorinated alcohols (Table 1) and are compared
with rate constants calculated from previously reported
activation parameters16 (Table 1). The fluorinated alco-
hols were chosen because they are excellent hydrogen-
bond donors.
The RDA reaction proceeds exceptionally fast in water
compared to organic solvents, its rate only being faster
in 1,1,1,3,3,3-hexafluoroisopropanol (HFIP). These re-
sults are striking and important because they clearly
show that water can greatly accelerate pericylic reactions
through hydrogen-bond interactions with hydrophobic
effects playing no significant role. This water-induced
acceleration of the RDA reaction is similar to the promo-
tion of the Claisen rearrangement in aqueous solvents.2
The data confirm the efficiency of water as a hydrogen-
bond donating solvent and its ability to stabilize the
polar activated complex.
In organic solvents, medium effects on DA and RDA
reactions are small. Both reactions are modestly pro-
moted by protic solvents.8,18 Desimoni and co-workers16
reported reaction rates of the cycloreversion of an-
thracenedione 1a and elegantly demonstrated that in
organic solvents the DA and RDA reactions are domi-
nated by the same solvent properties.
The RDA reaction is an organic transformation that
allows the stereospecific formation or regeneration of an
unsaturated bond, but often high reaction temperatures
(12) Van der Wel, G. K.; Wijnen, J . W.; Engberts, J . B. F. N. J . Org.
Chem. 1996, 61, 9001.
(13) Sauer, J .; Sustmann, R. Angew. Chem., Int. Ed. Engl. 1980,
19, 779.
(14) (a) Nanjappan, P.; Czarnik, A. W. J . Org. Chem. 1986, 51, 2851.
(b) Chung, Y.-S.; Duerr, B.; Nanjappan, P.; Czarnik, A. W. J . Org.
Chem. 1988, 53, 1334. (c) Chung, Y.-S.; Duerr, B. F.; McKelvey, T. A.;
Nanjappan, P.; Czarnik, A.W. J . Org. Chem. 1989, 54, 1018.
(15) J urczak, J .; Kawczynski, A. L.; Kozluk, T. J . Org. Chem. 1985,
50, 1106.
(16) Desimoni, G.; Faita, G.; Pasini, D.; Righetti, P. P. Tetrahedron
1992, 48, 1667.
(17) J enner, G.; Papadopoulos, M.; Rimmelin, J . J . Org. Chem. 1983,
48, 748.
(19) Lasne, M.-C.; Ripoll, J .-L. Synthesis 1985, 121.
(20) Bunnelle, W. H.; Shangraw, W. R. Tetrahedron 1987, 43, 2005.
(21) Bahr, N.; Gu¨ller, R.; Reymond, J .-L.; Lerner, R. A. J . Am. Chem.
Soc. 1996, 118, 3550.
(22) Chantarasiri, N.; Dinprasert, P.; Thebtaranonth, C.; Thebtara-
nonth, Y.; Yenjai, C. J . Chem. Soc., Chem. Commun. 1990, 286.
(23) Pagni, R. M.; Kabalka, G. W.; Hondrogiannis, G.; Bains, S.;
Anosike, P.; Kurt, R. Tetrahedron 1993, 49, 6743.
(24) (a) Marchand, A. P.; Vidyasagar, V. J . Org. Chem. 1988, 53,
4412. (b) Grieco, P. A.; Abood, N. J . Org. Chem. 1989, 54, 6008. (c)
Grieco, P. A.; Abood, N. J . Chem. Soc., Chem. Commun. 1990, 410.
(25) (a) Grieco, P. A.; Parker, D. T.; Fobare, W. F.; Ruckle, R. J .
Am. Chem. Soc. 1987, 109, 5859. (b) Grieco, P. A.; Clark, J . D. J . Org.
Chem. 1990, 55, 2271.
(18) Huisgen, R. Pure Appl. Chem. 1980, 52, 2283.

Retro-Diels-Alder Reaction in Aqueous Solution
J . Org. Chem., Vol. 62, No. 7, 1997 2041
Sch em e 2
reaction (2b + 3 f 1b) (Scheme 1) and the intramolecular
DA reaction (IMDA) of N-furfuryl-N-methylmaleamic
acid (Scheme 2) are compared. The different features of
these reactions enable an analysis of the factors that
govern the respective reactivities in water. In organic
solvents RDA reactions have a ∆qVθ close to zero.17 The
exact ∆qVθ for the cycloreversion of 1a is unknown, but
it appears likely that also in this case ∆qVθ is small. A
point of concern is the solvent effect on ∆qVθ. For the
bimolecular DA reaction it is known that ∆qVθ is more
negative in water than in organic solvents,27 which has
been attributed to electrostriction as a result of an
increased solvation of the activated complex (increased
hydration of the relatively polar activated complex). In
water, the ∆qVθ for the RDA reaction may be similarly
affected. However, we envisage that even in water the
∆qVθ for the RDA of 1a is much smaller than the ∆qV of
the bimolecular DA reaction. Furthermore it is obvious
that in the case of the bimolecular DA reaction both
reactants must be at least partly stripped of their
hydration shells before C-C bond formation can take
place, whereas this is not necessary in the case of the
RDA reaction, since in a geometrical sense the activated
complex resembles the reactant (1a ). Consequently the
∆qGθ of the RDA reaction will be hardly affected by
‘enforced hydrophobic interactions’, whereas this factor
does affect the DA and IMDA reactions since these
reactions are accompanied by a reduction of the solvent-
accessible surface area during the activation process.
The extent to which water can affect a reaction through
hydrogen-bond interactions naturally depends on the
number and the hydrogen-bond-accepting capability of
the substituents in the substrate. In the case of 1a the
two carbonyl groups enable protic solvents to exert this
stabilizing effect efficiently. A direct comparison of the
DA and RDA reactions is justified as follows: according
to the principle of microscopic reversibility, the DA and
the RDA reactions possess the same transition state and
the extent of polarization of all chemical bonds is identi-
cal. Therefore we contend that the hydrogen-bond-
induced reduction of the Gibbs energy of activation for
the RDA reaction of 1a affects the aqueous acceleration
of the bimolecular DA reaction of 2b with 3 similarly.
Also comparison with the IMDA reaction seems reason-
able because the dienophilic part of the IMDA probe
contains two activating groups.
F igu r e 1. Gibbs energy of activation for the RDA reaction of
1a vs the ET(30)-value of the solvents. Numbers correspond
to entries in Table 1.
Ta ble 2. Gibbs En er gy of Activa tion for th e RDAa
Rea ction of 1a , th e IMDAb Rea ction of 4, a n d th e
Bim olecu la r DAc Rea ction of 2b w ith 3 in Or ga n ic
Solven ts a n d in Wa ter
∆qGθ
∆qGθ
∆qGθ
DA
RDA
IMDA
(kJ mol-1
)
(kJ mol-1
)
(kJ mol-1
)
solvent
(at 40.0 °C)
(at 25.0 °C)
(at 25.0 °C)
hexane
propanol
TFE
water
HFIP
122.3
116.2
111.5
109.5
108.8
94.6
93.2
87.6
82.0
90.5
83.2
77.9
69.4
a
b
Partly taken from ref 16. Taken from ref 26. c Taken from
refs 4c and 26.
Figure 1 shows a remarkable correlation between the
∆qGθ of the RDA reaction and the solvatochromic param-
eter ET(30) of the solvents. The latter parameter is an
accepted indicator of the polarity and hydrogen-bond-
donating capacity of solvents;8 therefore, this observation
supports the importance of the notion that the ∆qGθ of
this type of reaction is largely governed by these solvent
properties. Likewise, the ∆qGθ of the RDA reaction
decreases with increasing AN of the solvents, but the
correlation is less accurate.
No detailed experimental data on a RDA reaction in
water are available, but J orgensen9a and Furlani and
Gao10 simulated the cycloaddition of methyl vinyl ketone
(MVK) to 3 in water along the entire reaction path. The
studies also included a calculation of the change in Gibbs
energy of hydration, on going from the transition state
to the product, which in effect is the change of Gibbs
energy of activation of the RDA reaction of that system.
J orgensen computed a ∆∆qGθ of 13 kJ mol-1 for the RDA
reaction, and Furlani and Gao obtained a value of 2.5
The data in Table 2 show that on going from the apolar
and aprotic n-hexane to water the ∆qGθ of the RDA
reaction is reduced by nearly 12.8 kJ mol-1. The contri-
bution of hydrogen bonding to the bimolecular DA
reaction is assumed to be similar (as a matter of fact, it
seems likely that it is even more dominant, since water
can interact better with the transition state of 1b: the
methyl group of 1a surely hampers hydrogen bonding of
kJ mol-1
. These results seem quite inconsistent, al-
though a correction of J orgensen9c on the Gibbs energy
of hydration of the transition state somewhat reduces the
discrepancy. Comparison of these data with our experi-
mental results is obscured by the fact that in the former
case the substrate only possesses one activating carbonyl
group, but two in the latter case. Still, a water-induced
reduction of the ∆qGθ of the RDA reaction of 6-8 kJ mol-1
per carbonyl group seems reasonable.
(26) (a) Blokzijl, W. Ph.D. Thesis, University of Groningen, Gronin-
gen, The Netherlands, 1991. (b) Engberts, J . B. F. N. Pure Appl. Chem.
1995, 67, 823.
(27) Isaacs, N. S.; Maksimovic, L.; Laila, A. J . Chem. Soc., Perkin
Trans. 2 1994, 495.
Com p a r ison w ith Oth er DA Rea ction s. In Table
2 our kinetic data for the RDA reaction of 1a and
previously reported data4a,26 for the bimolecular DA

2042 J . Org. Chem., Vol. 62, No. 7, 1997
Wijnen and Engberts
Ta ble 3. F ir st/Secon d -Or d er Ra te Con sta n ts a n d
Rela tive Ra te Con sta n tsa for th e RDA Rea ction of 1a (a t
40.0 °C) a n d th e Bim olecu la r DA Rea ction b of 2b w ith 3
(a t 25.0 °C) in High ly Aqu eou s Solu tion sc
106 × k1
krel
(RDA)
k2
RDA
DA
solvent
(s-1
)
(M-1 s-1
)
krel(DA)
water
WM95
WM90
WE95
WE90
WP95
WP90
WB95
WB90
WU95
WU90
WG95
3.59
3.20
2.74
3.04
2.28
2.69
1.33
1.75
0.91
3.44
3.38
4.08
1
4.48
4.29
3.80
4.73
3.89
5.14
1.52
4.39
0.90
3.82
3.46
7.40
1
0.89
0.76
0.85
0.64
0.75
0.37
0.49
0.25
0.96
0.94
1.14
0.96
0.85
1.06
0.87
1.15
0.34
0.98
0.20
0.85
0.77
1.65
a
b
See text. Partly taken from ref 4a. c WM, WE, WP, WB, WU,
and WG: aqueous mixtures of methanol, ethanol, 1-propanol,
t-butyl alcohol, urea, and glucose, respectively. WM95 indicates
a water-MeOH solution containing 95 mol % of water.
F igu r e 2. First-order rate constant (×106; 40 °C) of the RDA
reaction of 1a vs mole fraction of organic cosolvents: MeOH
(0), EtOH (0), 1-PrOH (4), t-BuOH (3), formamide (b),
acetonitrile (9), 1-cyclohexyl-2-pyrrolidinone (×), urea (+), and
glucose (x).
water), which would mean that hydrophobic interactions
are responsible for a reduction of the ∆qGθ by 8.3 kJ
(RDA) and krel(DA) indicate the relative acceleration/
deceleration of the reactions compared to pure water as
the solvent. The data show that nearly all organic
cosolvents decrease the rate of the RDA reaction. This
is opposite to the effect of small concentrations of these
cosolvents on the DA reaction. Hydrophobic cosolvents
induce the most significant retardation of the RDA
reaction, the impact of 1-cyclohexyl-2-pyrrolidinone being
most dramatic. The decelerating effect of the alcohols
parallels their hydrophobicity. Also acetonitrile induces
a substantial rate inhibition, whereas the effect of
formamide (which shares some characteristics of water)
is quite modest.
mol-1
.
Going from hexane to water, the ∆qGθ for the
IMDA reaction is reduced by 12.6 kJ mol-1, and this
suggests that also in this case hydrogen-bond interactions
play a pivotal role.
Using computational techniques, J orgensen9 has also
quantified the contributions of hydrogen-bond and hy-
drophobic interactions to the rate enhancements of the
‘aqueous DA reaction’. In the case of the DA reaction of
3 with MVK, enhanced hydrogen bonding of water to the
carbonyl substituent is shown to be largely responsible
for the accelerations in water. Depending on the geom-
etry of the MVK-water complex, this interaction may
account for up to 8.2 kJ mol-1. Hydrophobic interactions
are thought to account for 4-6.5 kJ mol-1. In a subse-
quent calculation the hydrogen-bond stabilization is
computed to be less efficient, which suggests that both
contributions are roughly similar. Furlani and Gao
conclude that hydrogen bonding contributes about 4.2-
6.7 kJ mol-1 to the acceleration of the same reaction in
water.10 Taking into account the fact that 1a contains
two carbonyl groups whose interactions with water are
hampered by steric hindrance, the computed hydrogen-
bond-induced reduction of the ∆qGθ and our results are
reasonably consistent.
Urea and glucose behave quite differently as addi-
tives: apparently substantial concentrations of these
compounds do not interfere with the hydrogen-bond-
donating capacity of water. This is in accord with
previous studies.28 Even 10 mol % of urea hardly affects
the RDA reaction in water, whereas addition of 5 mol %
of glucose slightly accelerates the reaction. In this case
the effects of urea and glucose on the RDA reaction and
the bimolecular DA reaction are very similar. Also the
latter reaction is insensitive to large concentrations of
urea but is accelerated by glucose.
E ffect of Cosolven t s on t h e R DA R ea ct ion in
Wa ter . A curious feature of the DA reaction in water is
the fact that small concentrations of hydrophobic organic
cosolvents lead to an additional acceleration.4,26 The
extent of this acceleration and the specific mole fraction
where maximum acceleration is observed depend on the
hydrophobicity of the cosolvent, e.g., addition of small
amounts of methanol hardly affects the rate constant,
whereas about 5 mol % of propanol or about 2-3 mol %
of t-butyl alcohol actually provides additional accelera-
tions. Previously, this phenomenon has been attributed
to increased ‘enforced hydrophobic interactions’.4,26
We examined the effect of small mole fractions of
cosolvents on the rate constant of the RDA reaction in
water, and the results are compiled in Table 3 and Figure
2. The second-order rate constants for the bimolecular
Under the assumption that the water-induced ac-
celerations of bimolecular DA reactions are caused by
both a hydrogen-bond effect and enforced hydrophobic
interactions, we can now discuss the effects that cosol-
vents exert on bimolecular DA reactions. The kinetic
data of the RDA reaction reveal that the hydrogen-bond
activation of water is decreased upon addition of alcohols
to water. Consequently the observed acceleration of the
bimolecular DA reaction in the presence of small amounts
of alcohol is predominantly caused by enhanced enforced
hydrophobic interactions. Addition of urea to water
hardly affects the RDA and bimolecular DA reactions;4a
therefore, we conclude that urea does not alter the
hydrogen-bond capacity of water (as the RDA results
show) and only slightly reduces the hydrophobic interac-
tions.28 Finally, the glucose-induced acceleration of bi-
molecular DA reactions4a seems to be partly caused by
enhanced hydrogen bonding of water to the electron-
withdrawing substituents of the dienophile, although
DA reaction (2b + 3 f 1b) are also included, while krel
-
(28) (a) Finney, J . L.; Soper, A. K. Chem. Soc. Rev. 1994, 1. (b)
Posthumus, W.; Engberts, J . B. F. N.; Bijma, K.; Blandamer, M. J . J .
Mol. Liq., in press.

Retro-Diels-Alder Reaction in Aqueous Solution
J . Org. Chem., Vol. 62, No. 7, 1997 2043
F igu r e 4. Activation parameters of the RDA reaction of 1a
in water/1-PrOH mixtures vs mole fraction of water: ∆qGθ (O),
∆qHθ (9), and -T∆qSθ (3). The T∆qSθ plot has been displaced
upward by 110 kJ mol-1 for clarity.
F igu r e 3. Relative rate constant of the RDA reaction (O) and
the DA reaction (9) in SDS solutions (see text).
Ta ble 4. F ir st-Or d er Ra te Con sta n ts for th e RDA
Rea ction of 1a (a t 40.0 °C) a n d Secon d -Or d er Ra te
Con sta n ts for th e Bim olecu la r DA Rea ction of 2b w ith 3
(a t 25.0 °C) in SDS Solu tion s
Ta ble 5. Activa tion P a r a m eter s for th e RDA Rea ction of
1a (a t 40.0 °C) in Wa ter /P r op a n ol Mixtu r es
∆qGθ
∆qHθ
-T∆qSθ
solvent
water
SDS 5.4 mM
10.3 mM
15.6
106 × k1(RDA) (s-1
)
k2(DA) (M-1 s-1
)
solvent
(kJ mol-1
)
(kJ mol-1
)
(kJ mol-1
)
3.59
3.61
2.59
1.86
1.49
1.32
1.11
4.48
5.21
5.09
5.00
4.83
4.54
3.78
water
109.5 ((0.1) 109.8 ((1.9) -0.3 ((1.8)
water/1-PrOH (Xw ) 0.95)a 110.8 ((0.2) 105.1 ((2.2) 5.7 ((2.0)
(Xw ) 0.90)a 112.5 ((0.1) 108.2 ((2.1) 4.3 ((2.0)
(Xw ) 0.80)a 113.5 ((0.1) 111.5 ((1.7) 3.6 ((1.6)
23.6 mM
36.6 mM
58.9 mM
2-PrOHb
116.2 ((0.1) 116.2 ((4.3) 0 ((4.2)
a
b
Xw is mole fraction of water. Taken from ref 16.
enhanced hydrophobic interactions again seem to be the
dominant cause of the acceleration of the bimolecular DA
reaction.
trations of SDS is a decrease of the rate constant
observed. These results are in agreement with previously
reported kinetic data on the effect of micelles on aqueous
DA reactions. Both the cycloaddition of 3 to MVK4a,26
and the cycloaddition of 3 to fumaronitrile33 are retarded
at relatively high concentrations of SDS, although the
10-20% decrease of the rate constant (similar to our
results) is certainly not dramatic.
Combining the results of the RDA and DA reactions,
it appears that micelles substantially reduce the hydrogen-
bond activation by water. However, in the case of the
bimolecular DA reaction this ‘deactivation’ is largely
compensated, presumably by the local increase of sub-
strate concentrations in the micelle, leading to the modest
deceleration of the bimolecular DA reaction.
Activa tion P a r a m eter s. Activation parameters shed
more light on the origin of the acceleration of the RDA
reaction of 1a in water. These parameters were deter-
mined in pure water and in water/1-propanol mixtures
and may be compared with previously reported values
for this reaction16 (Table 5 and Figure 4). Even in pure
water the entropy of activation is close to zero, supporting
the view that hydrophobic effects are of minor importance
for this reaction and also that solvent rearrangement is
rather small. The higher reactivity of 1a in water is
completely accounted for by a reduction of the enthalpy
of activation (relative to organic solvents), most likely
reflecting an increased interaction of water with the
activated complex. Addition of 1-propanol leads to en-
thalpy-entropy compensating behavior.34 When 5 mol
% of 1-propanol is added to water, the modest increase
of the ∆qGθ of the RDA reaction is the result of an
Effect of Micelles on th e RDA a n d DA Rea ction s
in Wa ter . In Table 4 and Figure 3 is recorded the effect
of SDS micelles on the RDA (at 40.0 °C) and DA (at 25.0
°C) reactions. In Figure 3, krel is defined as the ratio of
the (first- or second-order) rate constants in SDS solution
to that in pure water (krel ) kSDS/k0). The cycloreversion
is gradually retarded above the CMC of this surfactant
(8.56 mM at 40 °C29 ), but even at [SDS] ) 58.9 mM it
still proceeds rather fast. Note that in these SDS
solutions the rate constant is still much larger than that
in propanol. This is of interest because (under our
experimental conditions) at this [SDS] all solute mol-
ecules (1a ) certainly interact with micelles (the ratio SDS/
1a is more than 200 while the aggregation number of
SDS is about 8830). Apparently this does not prevent
water from activating 1a , which is in accord with the
notion that water can still interact with molecules that
are bound to micelles.31
Contrary to the above results, the bimolecular DA
reaction is slightly accelerated by SDS. This already
occurs at 5.4 mM SDS, suggesting that 2-3 mmol of 3
(or the propanol which is used as the solvent for the stock
solution) induces the formation of some sort of mixed
SDS/3 aggregates, which promote the reaction. This is
a well-documented phenomenon.32 Only at high concen-
(29) Moroi, Y.; Nishikido, N.; Uehara, H.; Matuura, R. J . Colloid
Interface Sci. 1975, 50, 254.
(30) Hayter, J . B.; Penfold, J . Colloid Polym. Sci. 1983, 261, 1022.
(31) Lissi, E.; Abuin, E. in Solubilization in Surfactant Aggregates;
Christian, S. D., Scamehorn, J . F., Eds.; Marcel Dekker: New York,
1995; p 297.
(33) Hunt, I.; J ohnson, C. D. J . Chem. Soc., Perkin Trans. 2 1991,
1051.
(34) Lumry, R.; Rajender, S. Biopolymers 1970, 9, 1125.
(32) Magid, L. In Surfactants in Solution; Mittal, K. L., Ed.; Plenum
Press: New York, 1977; Vol. 1, p 427.

2044 J . Org. Chem., Vol. 62, No. 7, 1997
Wijnen and Engberts
one (1a ) was synthesized according to a literature procedure16
and crystallized several times from cyclohexane, mp 97 °C
(lit.16 mp 96 °C). On a synthetic scale the RDA reaction was
carried out as follows: 50 mg of 1a was dissolved in 5 mL of
acetonitrile, and this solution was added dropwise to 200 mL
of water. The slightly turbid solution was stirred overnight
at 70-80 °C. After cooling, the yellow solution was extracted
several times with chloroform, the organic layer dried with
sodium sulfate, and finally the solvent removed. 1H NMR
analysis only revealed the presence of 2a .
unfavorable ∆qSθ, which is largely compensated by a more
favorable ∆qHθ. On increasing the concentration of
propanol, ∆qSθ drops back to zero and the reaction is
again completely enthalpy-controlled. This compensation
effect also strongly indicates that water directly partici-
pates in the reaction under study.34
The activation parameters for the bimolecular DA
reaction of 3 with MVK also exhibit this curious com-
pensating effect. In this case and when 1,4-naphtho-
quinone is used as dienophile, the reduction of the ∆qGθ
(going from 1-propanol to water) is almost equally ac-
counted for by the more favorable entropy and enthalpy
of activation.4c
Kin etic Exp er im en ts. Rate constants were determined
using UV/vis spectroscopy. Water was twice distilled in an
all-quartz distillation unit. Solvents were of the best available
quality and were distilled before use. The RDA reaction was
monitored at 340 nm, and the first-order rate constant was
determined using the initial rate method.35 This method is
less accurate than conventional pseudo-first-order kinetics, but
it enables determination of a large number of rate constants
for slow reactions. The technique requires determination of
the extinction coefficient of both 1a and 2a . In the highly
aqueous solutions these were nearly identical to those in water.
A few microliters of a stock solution of 1a dissolved in propanol
was added to the cuvettes; initial concentrations of 1a were
0.2-20 mM. The first-order rate constants are the average
of at least five independent experiments and were reproducible
to within 4%. Activation parameters were calculated from rate
constants at four different temperatures in the range 30-49
°C.
Con clu d in g Rem a r k s
The results of this study illustrate how water can
increase the reactivity of vinylic compounds by exerting
an activating effect on their substituents. It therefore
seems reasonable that water may also promote other
organic reactions in a similar fashion. It is indeed
striking that a large number of the water-promoted
organic reactions have been carried out with reagents
that possess polarizable, electron-withdrawing substit-
uents.1 Undoubtedly all these reactions are affected by
hydrophobic interactions, which can provide a unique
driving force for organic reactivity, but the hydrogen-bond
activation by water which we have studied in detail in
this paper seems to provide at least an equally important
contribution to the beneficial effect of water as a reaction
medium.
The DA reaction (2a + 3 f 1a ) does not affect the kinetics
of the RDA reaction: addition of a large excess of acrylonitrile
(to scavenge 3) does not change the observed rate constants.
The rate constants of the bimolecular DA reaction were
determined as described previously.4
Exp er im en ta l Section
Syn th esis a n d P r od u ct An a lysis. 1,4,4a,9a-Tetrahydro-
4a-methyl-(1R,4R,4aR,9aR)-1,4-methanoanthracene-9,10-di-
Ack n ow led gm en t. Mr. Evert van Rietschoten and
Mr. Frans Wessels are gratefully acknowledged for
carrying out the initial experiments of this work and
for valuable comments and discussion.
(35) (a) Frost, A. A.; Pearson, R. G. Kinetics and Mechanism;
Wiley: New York, 1961; p 45. (b) Scheiner, P.; Schomaker, J . H.;
Deming, S.; Libbey, W. J .; Nowack, G. P. J . Am. Chem. Soc. 1965, 87,
306.
J O9623200