Angewandte
Chemie
indeed occurring, however, this working hypothesis has
proved fruitful in allowing a marked improvement in the
efficiency of the catalytic system, in terms of both reaction
rate and catalyst lifetime.
As candidates for the promotion of bifunctional activa-
tion, we initially investigated 2-hydroxypyridine,[ which had
21]
previously been shown by us to act as a promoter for the
À
reaction in Equation 1 when [Rh(CO) ] was employed as the
4
[
22]
catalyst,
its 5-chloro analogue, and a phosphorus acid:-
diphenylphosphinic acid.[
21,23]
Van Leeuwen and co-workers
have reported on the promoting efficiency of carboxylic acids
in conjunction with [Pd(phen) ][TfO] (phen = 1,10-phenan-
2
2
À
À
[12,14]
throline; TfO = CF SO ) as a catalyst at 1358C.
How-
3
3
ever, in another paper,[ the same authors reported that the
11]
related [Pd(phen) ][BF ] complex affords better results. We
have compared both of these complexes in the presence of the
aforementioned promoters, also including benzoic acid and
2
4
2
Figure 1. Dependence of PhNO conversion on reaction time. Experi-
2
mental conditions: [Pd(phen) ][BF ] (2.8 mg, 4.410 mmol) in
2 4 2
methanol (30 mL) and 2,2-dimethoxypropane (1.0 mL); molar ratios
À3
Pd/phen/Ph POOH/PhNO =1:100:682:10000; T=1708C;
[
phenH][PF ] for a comparison, and in a broader range of
2
2
6
PCO =60 bar. The linear regression affords a line with the equation:
temperatures. The results are reported in Table S1 (see
Supporting Information). The main conclusions that can be
drawn from the data reported are the following:
2
Conv.%=23.49tÀ18.56 (R =0.999).
1
) We confirm that benzoic acid increases both conversion
and selectivity when employed at 1358C with [Pd(phen)2]
cause the reaction to be only 1.3 times slower in the first hour
than in the following time.
[
TfO] as catalyst. However, with the more active tetra-
Since aniline is known to be formed as an intermediate in
several catalytic carbonylation reactions of nitroarenes,[
we investigated the effect of its addition from the beginning of
such reactions. As shown in Table S2 (see Supporting
Information), the addition of aniline (4% with respect to
the starting nitrobenzene) accelerates the reaction in the first
hour by a factor of approximately four. Thus it is clear that the
induction time is, for the most part, due to the required
generation of aniline, which is slow in the dry solvent
employed. The addition of an equimolar amount of aniline
with respect to nitrobenzene has been previously reported to
accelerate the rate of the reaction with a palladium–phenan-
2
1–3,15]
fluoborate catalyst, no effect on conversion is observed at
either 135 or 1508C, although an increase in selectivity is
still observed. A positive effect on the reaction rate is only
observed at 1708C with [Pd(phen) ][BF ] .
) 2-Hydroxypyridine and its chlorinated analogue show
some promoting effects but, at least for the triflate
catalyst, the effect is smaller than that of benzoic acid
2
4 2
2
3
and thus they were not investigated further.
) Mestroni's promoter[
16–18,20]
([phenH][PF ]) is also effec-
6
tive with the tetrafluoborate catalyst, in accordance with
the literature. Under these conditions it affords better
conversions, but lower selectivities with respect to benzoic
acid.
) Diphenylphosphinic acid is clearly the best promoter, and
increases rate and selectivity with both catalysts at all
temperatures.
[
13]
throline catalyst,
but in the presence of such an high
amount of aniline the selectivity of the reaction is much
decreased and large amounts of azoxybenzene were formed.
At this stage other phosphorus acids were tested. A
comparison of the data in Table 1 shows that the introduction
of an electron-donating group on the phenyl rings of
diphenylphosphinic acid has a small positive effect, but the
electron-withdrawing chloride is not tolerated. Phenylphos-
phonic acid and its substituted analogues are a better class of
promoters, but phenylphosphinic acid, which can act as a
reductant towards palladium, completely deactivated the
catalytic system, apparently because of the fast formation of
inactive palladium metal. Even the more acidic dithiodiphe-
nylphosphinic acid was not a suitable promoter, and it is noted
that this acid is known not to be a good bifunctional
4
Working under a pressure of CO (60 bar), the reaction
temperature and the amounts of phen and diphenylphos-
phinic acid were optimized. During the optimization study it
was realized that, in accordance with data reported in the
[
16–20]
literature,
the addition of a small amount of 2,2-
dimethoxypropane as an internal drying agent has a beneficial
effect on the selectivity. This optimization of the conditions
allowed us to work at the unprecedented catalytic ratio
PhNO :Pd = 10 . Under these conditions, the reaction shows
an induction time of about one hour, after which it proceeds
4
2
[
21]
with a zero-order reaction rate with respect to PhNO without
catalyst. However, it was very interesting to observe that
the best conversion was obtained with commercial 85%
phosphoric acid! This is an ideal promoter in terms of cost,
absence of toxicity, and ease of separation from the reaction
products. The 1:1 mixture of mono- and dimethylphosphate,
obtained by reaction of P O with the stoichiometric amount
2
showing any rate decrease due to possible decomposition of
the catalytic system (Figure 1). The reaction is 5.4 times
slower during the first hour than at any following time. It
should be noted that the autoclave takes about 15 min to
reach the final temperature when immersed in the preheated
oil bath (see Experimental Section), but the reaction certainly
starts well before full equilibration has been achieved.
However, a delay of 15 min in the first hour would, by itself,
4
10
of methanol, was less efficient, but still retained a high
promoting efficiency, which shows that a partial esterification
of the acid does not block catalytic activity. Such esterification
Angew. Chem. Int. Ed. 2003, 42, 2886 – 2889
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2887
Ragaini, Fabio
