3270 J . Org. Chem., Vol. 63, No. 10, 1998
Selva et al.
chemisorption processes of the substrate and of the
intermediates can be significantly altered by the nature
of the PTC. Different PTCs presumably alter the metal
catalyst by causing distinct and selective interactions in
the binding of reactant functionalities over the active
surface sites.
At present, a relationship between the structure and
the effect for PTCs is hardly inferable; however, our
preliminary results seem to support the fact that for
nonionic surfactants, the higher the HBL number is, the
better the selectivity toward alcohol 4 is as well. That
is, the examined chemoselectivity seems to be improved
by increasing the hydrophilicity of the added Brij. On
this point, back-tracking to the case of onium salts, it
happens that also the water-soluble TEBA is more
effective than TBBA in promoting the formation of 4. On
the whole, the situation suggests that PTC-induced
partitioning phenomena between the aqueous and or-
ganic phases may promote chemoselectivity by mass-
transfer processes where the basic aqueous phase may
be involved.
F igu r e 3. Hydrodechlorination of 4-chloropropiophenone with
Pt/C under multiphase conditions in the presence of Brij 35
as PT agent.
agglomeration of Pt/C occurs with its subsequent deac-
tivation due to the loss of the active surface.
As for the use of Brijs as PT agents, they are well-
known surfactants whose different amphipathic struc-
tures determine HLB (hydrophile-lipophile balance21a
)
As aforementioned, hydrodehalogenation of chlorinated
aromatic ketones may have a high synthetic potential if
aromatic or aliphatic ketones and alcohols can be afforded
selectively, these products being often encountered in the
preparation of fine chemicals and speciality intermedi-
ates.17,19 Therefore, the applicability of the multiphase
conditions with Pt/C as the catalyst has been tested over
different chlorinated ketones. This was also in order to
comprehend (i) the effects (if any) induced by the presence
of either a bulky aromatic ring or an alkyl substituent
on the reacting molecule, and (ii) the behavior of poly-
chloro- or polyhalo-substrates toward multiphase deha-
logenation. Results are reported in Table 4.
numbers of 17, 5, 13, and 15, for Brij 35, 52, 56, and 58,
respectively. Accordingly, they are placed in detergent-
solubilizer (35, 56, and 58) and water-in-oil emulsifier
(52) classes.21b
In the case of multiphase hydrodehalogenation of 1,
Brij 35 is the most effective in promoting chemoselectivity
toward the alcohol 4 which is obtained in a 97% amount,
after 7 h of reaction (entry 3); whereas, Brij 52 is the
least efficacious since sizable quantities (23%) of ring
reduction byproducts (compounds 5 and 6) are observed
(entry 4). An intermediate behavior is exhibited by Brij
56 and 58: the former prevents the formation of second-
ary products but the complete chlorine removal is rather
slow due to the accumulation of the halogenated alcohol
3 (71%, after 6 h; entry 5). The latter allows a fairly good
selectivity toward 4 (82%, after 3.5 h; entry 6), though
the formation of 3 is still not negligible (13%).
The reaction run in the presence of Brij 35 exemplifies
how the PT agent can affect the reaction rate, and it may
provide further insight into the mechanism of the PTC
action. Figure 3 shows the outcome of the hydrodechlo-
rination reaction of 4-chloropropiophenone in the pres-
ence of Brij 35: after the substrate 1 is totally consumed,
the disappearance rate of 2 exhibits a sharp flex point,
as does also the related rate of formation of 4. This may
indicate that the halo ketone 1 has a higher affinity for
the Brij 35-modified catalyst surface than the dehaloge-
nated ketone 2; therefore, as long as 1 is present, the
reaction proceeds rapidly, allowing a sufficiently high
accumulation of 2 on the catalyst surface (path “a” of
Scheme 2). However, when 1 is totally consumed, 2 is
chemisorbed much less efficiently, and both its reaction
rate and the formation of 4 are concurrently slowed. From
that point on, the formation of 4 is controlled by the
disappearance of 2 and 3 which proceeds with comparable
rate constants (see also path “b” in Scheme 2): i.e., the
Brij-modified catalyst shows similar affinity for 2 and 3.
In the case of Aliquat 336, this trend is much less
evident, if at all present (Figure 1b), indicating that the
p-Chloroacetophenone shows a slightly different be-
havior than p-chloropropiophenone (1). In the case of 1,
the hydrogenation of the carbonyl group is never totally
prevented before the dechlorination is complete; and the
hydrogenation also takes place on the dechlorinated
ketone 2 which forms from the direct dechlorination of
the substrate. As a consequence, a mixture of compounds
2-4 is observed during the reaction (Table 2); in the long
run, only the alcohol 4 can be obtained with up to 95%
selectivity (entry 1, Table 4). Instead, the C-Cl bond
cleavage can occur selectively on p-chloroacetophenone
yielding the corresponding dehalogenated ketone (C6H5-
COCH3) in a quantitative yield (100% after 0.25 h of
reaction, entry 2); then, in a second step, the hydrogena-
tion takes place giving the dechlorinated alcohol (C6H5-
CH(OH)CH3: 100% after 2 h of reaction, entry 2).
Sterically hindered molecules are also effectively de-
halogenated to the corresponding alcohols. For instance
p-chlorobenzophenone affords diphenylcarbinol in a 95%
yield after 3 h reaction (entry 3), being cyclohexylphenyl
ketone (5%) the sole byproduct. In the case of polychlo-
rinated substrates, a lower selectivity is observed: for
instance, 4,4′-dichlorobenzophenone gives the correspond-
ing dechlorinated alcohol (diphenylcarbinol) in a 86%
yield (after 2.5 h reaction, entry 4) along with different
byproducts such as dicyclohexyl ketone, cyclohexylphenyl
ketone, 1-cyclohexyl-1-phenylmethanol (1%, 1%, and 6%,
respectively) and phenyl(p-methoxyphenyl)methane (6%,
structure assigned through GC/MS analysis). In this
case, the reaction was carried out by halving the con-
centration of the organic solution (from 0.07 to 0.035 M)
(21) (a) In the case of simple ethoxylated alcohols such as Brijs, HLB
is defined as E/5 where E is the weight % of the ethylene oxide in the
molecule. (b) Kirk-Othmer, Encyclopedia of Chemical Technology, 3rd
ed.; J . Wiley & Sons: 1983; Vol. 22, pp 360-365.