Mechanism of Hydrothermal Cleavage of Bisphenol A
Further, the portion of the methylene bridge-contain-
ing diaryl that is cleaved from the methylene group
develops a negative charge after cleavage, and hence the
reactivity of the methylene group is related to the
stability of the cleaved anion. This is analogous to the
concept of the leaving group in nucleophilic substitution
reactions. Hence, as the basicity of the cleaved anion
decreases, the reactivity of the diaryl increases.
than the benzene anion, bisphenol A should be, and is,
more reactive than 4-cumylphenol through base catalysis
in HTW.
Con clu sion
The synthesis of p-isopropenylphenol via bisphenol A
cleavage in HTW is first order in bisphenol A. BPA
cleavage occurs by specific acid catalysis, specific base
catalysis, and general catalysis with water acting as a
general catalyst. The primary mechanism for BPA cleav-
age in neutral HTW is general catalysis by water. The
existence of specific acid- and base-catalyzed mechanisms
suggests that water can act as both a general acid and a
general base for the cleavage reaction. General base
catalysis by water is more significant than general acid
catalysis by water and therefore the dominant mecha-
nism in neutral HTW. A detailed kinetics model built
upon a water-catalyzed mechanism was quantitatively
consistent with the experimental data and accurately
predicted the experimental acetone yields.
The results of this work allow one to predict the
reactivity of methylene bridge-containing diaryls under
hydrothermal conditions. If a hydroxyl group is present
on an aromatic ring, then hydrothermal cleavage may
be possible. Base-catalyzed cleavage is possible when the
hydroxyl group is located ortho or para to the bridging
carbon. Additionally, the portion of the molecule cleaved
from the methylene bridge must be a good leaving group,
such as the phenoxide anion, or base-catalyzed cleavage
is not likely to occur. Cleavage may occur via an acid-
catalyzed path when the aromatic ring is hydroxy-
substituted at the ortho or para position and there exists
a hydrogen â to the bridging carbon. Acid-catalyzed
cleavage of a meta-hydroxy-substituted diaryl containing
a methylene cross-link may also be possible if the
â-hydrogen is present, provided that the reaction is
initiated through protonation of the hydroxyl group. In
neutral HTW, where both acid and base catalysis is
possible, we expect base-catalyzed bridge cleavage to be
more significant than acid-catalyzed cleavage for com-
pounds that contain hydroxyl groups in the ortho or para
positions, have good leaving groups, and are alkyl-
substituted at the methylene bridge.
Additionally, there is no â-hydrogen required for base-
catalyzed cleavage. Thus, both carbene bridges and true
methylene bridges may be cleaved according to Scheme
1, whereas only carbene bridges react via acid catalysis.
The mechanisms presented in Schemes 1-3 can ac-
count for the reactivity observed for different methylene
bridges under hydrothermal conditions. If an aromatic
hydroxyl group is not present and the aromatic rings are
unsubstituted, methylene bridge-containing diaryls will
be stable under hydrothermal conditions. For example,
diphenylmethane and 1-benzylnaphthalene were stable
after 1 h in water at 460 °C.18 Similarly, di-p-tolyl-
methane did not react after 1 h in water at 430 °C.19 The
hydroxyl substituent is required for both acid- and base-
catalyzed cleavage, and its presence may enable meth-
ylene bridge cleavage in HTW. In addition to bisphenol
A, 4,4′-dihydroxydiphenylmethane and 2,2′-dihydroxy-
diphenylmethane are cleaved in HTW at temperatures
below 350 °C.19
The effect of added acid and base on the hydrothermal
reactivity of methylene bridges can also be explained by
Schemes 1-3. For example, cleavage of 4,4′-dihidroxy-
diphenylmethane in HTW is accelerated by added base
but not by added acid.19 Acid-catalyzed cleavage would
yield an energetically unfavorable cresol primary cation.
Despite the presence of the hydroxyl group, the instability
of the primary carbocation that needs to form and the
lack of a fast reaction path for this intermediate (no
â-hydrogen) militate against acid-catalyzed cleavage.
Thus, one would not expect added acid to accelerate the
reaction. Base catalysis, however, is permitted by the
presence of the para hydroxyl group and the phenoxide
leaving group. Hence, added base should and does
accelerate the reaction.
This explanation can also be used to account for the
reactivity of phenol resin prepolymers in HTW. These
prepolymers are structurally similar to 4,4′-dihydroxy-
diphenylmethane, as they consist of several substituted
phenols linked by true methylene bridges. Again, added
base accelerated the decomposition of these prepolymers,
but added acid did not.32
Exp er im en ta l Section
We performed experiments using 0.6 mL batch reactors.
Reactors were loaded with reactants and degassed water in a
helium-filled glovebox and then sealed. The sealed reactors
were then immersed in a preheated temperature-controlled
fluidized sand bath for the desired reaction time. After the
reaction time had elapsed, the reactors were removed from
the sandbath and submersed in a bath of ambient temperature
water to quench the reaction. Reactors were then further
cooled for 30 min in a refrigerator and the contents subse-
quently collected. Reaction products were identified using GC-
MS, and quantified using GD-FID. Full details of the general
experimental method are available in our earlier paper.8
The effect of the leaving group is demonstrated when
considering that BPA is readily cleaved in HTW at 250
°C, whereas 4-cumylphenol did not react even after 1 h
at this temperature. The fragment cleaved from BPA
through base catalysis is an isomeric phenoxide, which
rearranges to give the phenoxide anion. The fragment
that would be cleaved from 4-cumylphenol via base
catalysis is the benzene anion. Benzene is clearly a much
weaker acid than phenol, and hence the benzene anion
is a much stronger base than the phenoxide anion. Thus,
because the phenoxide anion is a better leaving group
Ack n ow led gm en t. We thank Prof. M. R. DeCamp
for providing mechanistic advice. Additionally, we grate-
fully acknowledge the NSF for providing a graduate
fellowship to S.E.H. Financial support was also provided
by the ACS-PRF (37603-AC9) and NSF (CTS-0218772).
(32) Suzuki, Y.; Tagaya, H.; Asou, T.; Kadokawa, J .; Chiba, K. Ind.
Eng. Chem. Res. 1999, 38, 1391.
J O0356964
J . Org. Chem, Vol. 69, No. 14, 2004 4731