Promotion of the [PPN][Rh(CO)4]-catalysed carbonylation of nitrobenzene by 2-hydroxypyridine and related molecules: An apparent bifunctional activation

Article abstract of DOI:10.1016/S0022-328X(99)00502-1

2-Hydroxypyridine and related molecules have a large activating effect on the previously reported [PPN][Rh(CO)₄]-based catalytic system for the reductive carbonylation of nitrobenzene to methyl phenylcarbamate (PPN⁺=(PPh₃)₂N⁺). The effect is not due to the known 2-hydroxypyridine-2-pyridone tautomeric equilibrium, since 4-hydroxypyridine, for which the same tautomeric equilibrium exists, completely inhibits the reaction. A promoting effect of 2-hydroxypyridine is also observed in the reactions of a previously isolated metallacyclic complex, [PPN][Rh(CO)₂ON(Ar)C(O)O], believed to be an intermediate in the catalytic reactions. However, the dependence of the rate of the catalytic reactions on the aniline concentration indicates that the effect found for the stoichiometric reaction cannot be the one that is relevant for the acceleration of the catalytic reactions. Thus, two different effects are present, both of which appear to be due to the proximal positions of a basic and an acidic site in the promoter molecules.

Full text of DOI:10.1016/S0022-328X(99)00502-1

www.elsevier.nl/locate/jorganchem
Journal of Organometallic Chemistry 593–594 (2000) 109–118
Promotion of the [PPN][Rh(CO)4]-catalysed carbonylation of
nitrobenzene by 2-hydroxypyridine and related molecules: an
apparent bifunctional activation
Fabio Ragaini *, Emma Gallo, Sergio Cenini
Dipartimento di Chimica Inorganica, Metallorganica e Analitica and CNR Center, 6ia G. Venezian 21, I-20133 Milan, Italy
Received 12 May 1999; accepted 20 August 1999
Dedicated to Professor Fausto Calderazzo in recognition of his important contribution to organometallic chemistry.
Abstract
2-Hydroxypyridine and related molecules have a large activating effect on the previously reported [PPN][Rh(CO)4]-based
catalytic system for the reductive carbonylation of nitrobenzene to methyl phenylcarbamate (PPN+=(PPh3)2N+). The effect is
not due to the known 2-hydroxypyridine–2-pyridone tautomeric equilibrium, since 4-hydroxypyridine, for which the same
tautomeric equilibrium exists, completely inhibits the reaction. A promoting effect of 2-hydroxypyridine is also observed in the
¸¹¹¹¹¹¹¹¹¹¹¹¹º
reactions of a previously isolated metallacyclic complex, [PPN][Rh(CO)2ON(Ar)C(O)O], believed to be an intermediate in the
catalytic reactions. However, the dependence of the rate of the catalytic reactions on the aniline concentration indicates that the
effect found for the stoichiometric reaction cannot be the one that is relevant for the acceleration of the catalytic reactions. Thus,
two different effects are present, both of which appear to be due to the proximal positions of a basic and an acidic site in the
promoter molecules. © 2000 Elsevier Science S.A. All rights reserved.
Keywords: Homogeneous catalysis; Nitrobenzene; Carbonylation reactions; Rhodium; 2-Hydroxypyridine
1. Introduction
phosgene employed in the industrial process [5]. It
should be noted that the activity of the catalytic system
reported here is higher than that of all other rhodium-
or ruthenium-based catalytic systems for this reaction.
Only a few palladium-based catalytic systems showing
an even higher activity have been reported [5].
Previously reported data [4,6] that are relevant to the
present paper are the following:
1. Nitrogen organic bases such as 2,2%-bipyridine
(Bipy) or pyridine (Py) have a promoting effect on
the catalytic reaction.
2. It was shown that during the catalytic reactions
nitrobenzene is intermediately reduced to aniline
and only at a later stage of the reaction is this last
product carbonylated to the carbamate.
Bifunctional activation of reacting molecules is usu-
ally the rule in enzymatic catalysis, but it is an, as yet,
little explored field in traditional homogeneous catalysis
[1–3]. Several years ago, we reported on the use of
[PPN][Rh(CO)4] (PPN+=(PPh3)2N+) as a catalyst for
the carbonylation reaction of nitrobenzene to methyl
phenylcarbamate [4].
[PPN][Rh(CO) ]
4
ꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢀꢁ
CO
170−200°C
PhNO2+MeOH+3CO
PhNHCOOMe
THF or toluene
=40−80 bar+2CO2
(1)
P
Since carbamates can be thermolysed to yield the
corresponding arylisocyanates, this type of process rep-
resents a viable alternative for the synthesis of this last
type of product, which avoids the use of the toxic
3. A probable intermediate in the catalytic reaction,
¸¹¹¹¹¹¹¹¹¹¹¹¹º
[PPN][Rh(CO)2ON(Ar)C(O)O] (Ar=3,4-Cl2C6H3
1a, Ph 1b) was isolated and found to react with
methanol under an atmosphere of CO to yield
* Corresponding author. Tel.: +39-02-26680675; fax: +39-02-
2362748.
E-mail address: ragaini@csmtbo.mi.cnr.it (F. Ragaini)
0022-328X/00/$ - see front matter © 2000 Elsevier Science S.A. All rights reserved.
PII: S0022-328X(99)00502-1

110
F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
Scheme 1.
ArNH2. Under a higher CO pressure, the corre-
sponding methyl carbamate was also obtained.
4. The reaction of 1 with methanol was accelerated by
2. Results
2.1. Stoichiometric reactions
6
the addition of Bipy or Py, but these bases did not
react with the complex in the absence of methanol
(Scheme 1).
We have already reported [6] that complex 1a is
obtained in good yield by reaction of [PPN][Rh(CO)4]
with 3,4-Cl2C6H3NO2 in THF under CO at room tem-
perature. However, the work-up of the reaction mixture
was very delicate, because a dark polymeric impurity
was also formed. During this work, we found that the
reaction could be rendered more selective by working at
−78°C. In this way, no dark products were formed and
the isolated yield could be increased to 86%.
5. Methylation experiments showed that in the reac-
tion of 1 with methanol, the site of interaction
between the hydroxylic proton of methanol and 1
is neither of the two ‘CO2’ oxygen atoms, but it
could not be ascertained if the nitrogen or oxygen
atoms of the Ar–NO moiety was involved in this
reaction.
Reinvestigation of the reaction between 1a and
methanol showed that, quite unexpectedly, no reaction
at all occurs, at least at room temperature or slightly
above it, when it is performed under a dinitrogen
atmosphere, even under concentration conditions where
the reaction is very fast under CO. Interestingly, the
addition of pyridine to the reaction mixture under
dinitrogen allows the reaction to proceed, although, as
A simplified reaction sequence is reported in Scheme
2. Note that the precise identity of the carbalkoxy
complex is not known and the Scheme was drawn by
analogy with a ruthenium complex identified in related
mechanistic studies [7]. Note also that the formation
of the methyl carbamate from the carbalkoxy complex
and aniline is probably not straightforward. In the
case of the Ru(CO)3(DPPE) (DPPE=1,2-bis(-
diphenylphosphino)ethane) catalytic system, this step
was shown to involve the nucleophilic attack of aniline
on a coordinated CO (not on the ꢀCOOMe group),
the formation of isocyanate (PhNCO) and its trapping
by excess aniline to afford diphenylurea (PhN-
HC(O)NHPh), and finally the alcoholysis of the urea
to afford the carbamate and to regenerate an equiva-
lent of aniline.
Point (4) suggested to us that the simultaneous pres-
ence of a nitrogen base and an alcohol close to the
metal centre is a preferred situation for the reaction to
proceed. We investigated the effect of the addition of
molecules bearing both a basic nitrogen atom and an
ꢀOH group on both the catalytic reactions and the
stoichiometric decomposition of 1. At least, under cer-
tain conditions, a promotional effect was found for
both reactions, although some inhibiting effects were
also present that rendered the situation more complex.
Scheme 2.

F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
111
also previously reported [6], no reaction is observable
between 1a and pyridine in the absence of methanol. In
any case, the reaction is slower than the corresponding
reaction in the presence of CO. 3,4-Dichloroaniline was
formed as the organic product, analogously to what
was found for the reaction in the presence of CO
(Scheme 1), but [Rh(CO)4]− could not be reformed and
a mixture of complexes (apparently rhodium clusters),
which could not be identified, was obtained instead.
Although 1a was stable in the presence of either
methanol or pyridine alone and under a dinitrogen
atmosphere, 2-hydroxypyridine reacted with the metal-
lacycle even in the absence of both methanol and CO.
3,4-C6H3NH2 was formed, but no rhodium complexes
could be isolated.
As mentioned in Section 1, the hydroxylic proton of
methanol (and probably even that of hydroxypyridine)
appears to interact with either the oxygen or nitrogen
atom of the ArꢀNO moiety of 1, but which of the two
atoms is responsible for the initial interaction has not
been determined. In this work, we have examined the
reactivity of other electrophiles to gain a deeper insight
into this problem.
reaction, a complex could be isolated showing absorp-
tions in the IR spectrum (THF) at 2054, 1972 and 1621
cm−1. This complex was characterised by single-crystal
X-ray
diffractometry
and
shown
to
be
[PPN][Rh(CO)2(OAc)2] (2) [8]. The same complex has
been previously characterised by X-ray spectroscopy,
with Bu4N+ as a countercation [9].
Even the use of a bulkier anhydride, di-tert-butyl
anhydride, did not allow for the isolation of an inter-
mediate in which the nitrogen atom was still part of a
ligand moiety; a complex showing absorptions in the
IR spectrum identical to those of 2 was observed, which
is thus proposed to be [Rh(CO)2(ButCOO)2]−.
The reaction of 1a with aniline did not proceed at all
at room temperature under a dinitrogen atmosphere,
but, analogously to what was observed for methanol,
the reaction started immediately when the dinitrogen
atmosphere was replaced by CO. The complex
[Rh(CO)4]− was again obtained as the only
organometallic product, while diphenylurea and the
mixed urea PhNHC(O)NHC6H3Cl2 were isolated in a
41/9 molar ratio (Eq. (4)). A reaction was also observed
in the absence of CO if pyridine was also added, but in
this case a mixture of cluster complexes was formed
and no urea, but only 3,4-dichloroaniline, was detected
as the organic product.
Reaction of 1a with p-toluolyl chloride afforded
[Rh(CO)2Cl2]− and 4-MeC6H4C(O)NH(3,4-Cl2C6H3).
No intermediate could be isolated from this reaction
(Eq. (2)):
¸¹¹¹¹¹¹¹¹¹¹¹¹º
THF
[Rh(CO)2ON(Ar)C(O)O]− +PhNH2 “ [Rh(CO)4]−
(1a)
CO
¸¹¹¹¹¹¹¹¹¹¹¹¹º
[Rh(CO)2ON(Ar)C(O)O]−
+PhNHC(O)NHPh+ArNHC(O)NHPh
(4)
(1a)
+4-CH3C6H4C(O)ClT“HF[Rh(CO)2Cl2]−+CO2
where Ar=3,4-Cl2C6H3.
The fact that diphenylurea is strongly predominant
over the mixed urea supports the view that the ary-
lamine is an intermediate in the carbonylation reaction.
Indeed, once 3,4-Cl2C6H3NH2 is formed it must com-
pete with the more nucleophilic PhNH2 and the
product derived from this last amine prevails. We have
previously observed the same outcome for catalytic
reactions between 3,4-Cl2C6H3NO2 and PhNH2 in the
presence of either palladium or ruthenium complexes
[10,11].
N
2
+ArNHC(O)C6H4CH3
(2)
where Ar=3,4-Cl2C6H3.
Some non-carbonyl-containing rhodium complexes
must also be formed, but they could not be observed by
IR spectroscopy, since the reduction of the metal-
bound nitrosoarene requires the consumption of an
equivalent amount of CO, which can only derive from
a decomposition of 1a that does not afford a carbonyl
complex.
When a similar reaction was performed with acetic
anhydride, both the arylacetamide 3,4-Cl2C6H3NHC-
(O)Me and the bis-acetylated product Cl2C6H3N-
(C(O)Me)OC(O)CH3 were obtained in 37 and 30%
isolated yield, respectively (Eq. (3)). No product could
be detected in which only the oxygen atom of the
nitrosoarene moiety had been acylated. From the same
2.2. Catalytic reactions
We first compared the effect of the addition of
2,2%-bipyridine and 2-hydroxypyridine under the origi-
nally employed conditions [4] and the results are re-
ported in Table 11.
In agreement with the original results [4], the addi-
tion of 2,2%-bipyridine to the reaction carried out at
1 The results reported here for entries 1 and 2 of Table 1 are
slightly different from those reported in Ref. [4]. However, this is due
to the fact that the experiments have been repeated after several
years, using freshly distilled THF, instead of THF dried but then
stored for several weeks before use. Anyway, the trends are the same
as previously reported.
(3)

112
F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
Table 1
Effect of 2,2%-bipyridine and 2-hydroxypyridine on the [PPN][Rh(CO)4]-catalysed carbonylation of nitrobenzene to methyl phenylcarbamate in
THF a
Run
Promoter
Prom./Rh
mol ratio
PhNO2
Carbamate
select. (%) c
Aniline
Azobenzene
select. (%) c
Azoxybenzene
select. (%) c
conv. (%) b
select. (%) c
1
2
3
–
–
6
37.5
56.4
59.8
75.5
56.4
74.5
8.3
12.4
9.4
–
0.8
1.2
–
16.4
–
2,2%-Bipyridine
2-Hydroxypyridine 12
a Experimental conditions: [PPN][Rh(CO)4]=0.125 mmol, mol ratio PhNO2/Rh=300, T=170°C, PCO=60 bar, in THF (8 ml)+MeOH (3
ml), for 1.5 h.
b Calculated with respect to the starting nitrobenzene.
c Calculated with respect to converted nitrobenzene. The selectivity in azo- and azoxybenzene was lower than 1% in all cases where no value
is reported. Some diphenylurea was always present as a by-product, which could be detected by gas chromatography but not quantified in a
precise manner.
Table 2
Effect of several promoters on the [PPN][Rh(CO)4]-catalysed carbonylation of nitrobenzene to methyl phenylcarbamate in toluene a
Run
Promoter
PhNO2 conv.
(%) b
Carbamate
select. (%) c
Aniline select.
(%) c
Azobenzene
select. (%) c
Azoxybenzene
select. (%) c
1
2
3
4
5
6
7
8
Et3N
Pyrazole
66.1
ꢀ0
73.1
8.5
traces
10.3
9.6
10.5
9.3
12.1
12.4
traces
N-Methylimidazole
Pyridine
62.2
63.5
65.0
80.3
65.9
59.5
ꢀ0
79.7
76.3
74.5
87.2
69.8
81.0
traces
2,2%-Bipyridine d
0.8
1.9
1.6
2-Hydroxypyridine
2-Hydroxymethyl-pyridine
2(2-Hydroxyethyl)-pyridine
4-Hydroxypyridine
4-Hydroxyquinazoline
2-Hydroxypyrimidine
hydrochloride
9
10
11
ꢀ0
ꢀ0
12
2-Hydroxypyrimidine
hydrochloride + proton sponge
2-Pyridinethiol
ꢀ0
15.6
5.7
85.4
6.4
13
14
15
16
17
18
37.7
traces
88.0
10.0
71.9
77.1
45.7
traces
9
48.0
11.2
7.3
4.1
1.6
Picolinic acid
2-Hydroxyquinoline
8-Hydroxyquinoline
2-Hydroxypyridine e
2,2%-Bipyridine e
0.9
63.3
68.3
2.9
1.2
a Experimental conditions: [PPN][Rh(CO)4]=0.125 mmol, mol ratio PhNO2/Rh/promoter=300:1:6, T=200°C, PCO=60 bar, in toluene (8
ml)+MeOH (3 ml), for 1.5 h.
b Calculated with respect to the starting nitrobenzene.
c Calculated with respect to converted nitrobenzene. The selectivity in azo- and azoxybenzene was lower than 1% in all cases where no value
is reported. Some diphenylurea was always present as a by-product, which could be detected by gas chromatography but not quantified in a
precise manner.
d Mol ratio bipy/Rh=3.
e MeOH=1.7 ml.
170°C increases the conversion, but causes the forma-
tion of a large amount of azoxybenzene (run 2). On the
other hand, the addition of 2-hydroxypyridine allows
for an increase in reactivity without any loss in selectiv-
ity, although the promoting effect on the conversion is
only slightly higher than that of 2,2%-bipyridine under
these conditions (see below).
Since we had in the mean time discovered that the
use of toluene as solvent and a higher temperature
(200°C) were beneficial to the reaction [4], a more
extensive comparison of several potential promoters
was performed under these last conditions; the results
are reported in Table 2. The structures of all but the
simplest promoters used are shown in Fig. 1.

F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
113
As can be seen from the data reported, the use of
2,2%-bipyridine or of a double molar amount of
pyridine, methylimidazole or triethylamine afforded
very similar results, whereas pyrazole was ineffective
(Table 2, runs 1–5). The use of 2-hydroxypyridine, on
the other hand, gave a significantly higher conversion
and selectivity (run 6), but its isomer 4-hydroxypyridine
completely inhibited the reaction (run 9).
Moving the hydroxy group further away from the
nitrogen atom had a detrimental effect. The promoting
efficiency of 2-hydroxymethylpyridine (run 7) was com-
parable to that of unsubstituted pyridine, whereas that
of 2-(2-hydroxyethyl)pyridine (run 8) was even lower,
probably because of the unfavourable position of the
two functional groups of this promoter.
Other promoters were tested, having the same, or
similar, geometrical features as 2-hydroxypyridine. Of
these, the more acidic 4-hydroxyquinazoline (run 10)
and 2-hydroxypyrimidine hydrochloride (run 11) com-
pletely suppressed the reaction. Addition of an equiva-
lent amount of proton sponge (1,8-bis(dimethyl-
amino)naphthalene) to the second promoter, in order
to neutralise the hydrochloric acid bound to the free
base, did not improve the conversion (run 12).
A very poor conversion was also observed when
2-pyridinethiol was employed as promoter (run 13) and
picolinic acid completely inhibited the reaction (run 14).
Of the tested promoters, only the use of 2-hydrox-
yquinoline (run 15) gave a somewhat higher conversion
than 2-hydroxypyridine, but its much higher price did
not justify its use. The isomeric 8-hydroxyquinoline was
not effective either (run 16).
As the hydroxy group in 2-hydroxypyridine was sup-
posed to enter the reaction at some stage in place of
methanol, we reasoned that the difference in promoting
activity between 2-hydroxypyridine and 2,2%-bipyridine
should have increased upon a decrease in the methanol
concentration in the reaction mixture. However, a com-
parison of the two reactions, performed with these two
promoters under the conditions of Table 2, but with the
addition of only 1.7 ml of methanol instead of 3 ml
(1.54 ml is the amount of methanol required by the
stoichiometry of the reaction to convert all of the
starting nitrobenzene into methyl phenylcarbamate)
shows that the nitrobenzene conversion is only slightly
altered in the case of 2,2%-bipyridine (it even increases a
little, run 18), but markedly decreases in the case of
2-hydroxypyridine (run 17). The effect is such that,
under these particular conditions, the use of 2,2%-
bipyridine affords better results than that of 2-hydrox-
ypyridine. Clearly, this observation cannot be
rationalised by any bifunctional effect and suggests
that, analogously to what is found for 4-hydroxy-
pyridine, even for its 2-isomer some inhibiting effects
on the reaction must exist and changes in the experi-
mental conditions may cause these negative effects to
prevail over the positive ones.

114
F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
As mentioned in Section 1, aniline is known to be
formed as an intermediate during reaction (1). Thus,
the effect of the addition of variable amounts of aniline
to the reaction mixture, with both 2,2%-bipyridine and
2-hydroxypyridine as promoters, was examined and the
results are reported in Table 3.
2-hydroxypiridine, so as to re-establish the 2-hydroxy-
pyridine/Rh ratio used in Table 2, leads to a higher
conversion (compare runs 8 and 4 in Table 3), but the
selectivity is only slightly altered.
In order to avoid reaching complete conversion, the
amount of rhodium was halved, but all of the other
concentrations were left unchanged. The data reported
clearly show that the addition of up to 60 equivalents of
aniline per equivalent of rhodium leads to a sharp
increase in reaction rate with both 2,2%-bipyridine and
with 2-hydroxypyridine as promoters, although this
effect levels off for higher amounts of aniline. Even the
selectivity in carbamate follows a similar trend, with a
small decrease at higher aniline concentrations being
due to the formation of larger amounts of diphenyl-
urea, a product that is anyway known to alcoholyse to
the carbamate under suitable conditions and can thus
be partly recycled into the desired product. Under these
3. Discussion
The investigation of the reaction between 1a and
methanol under a dinitrogen atmosphere led to the
unexpected observation that no reaction occurs if nei-
ther CO nor pyridine is present. Interaction of CO with
any part of 1a except for the metal is impossible. Thus
the only possible explanation for the observed effect is
that the tetracoordinated 1a is unreactive towards
methanol, but is in equilibrium with a very small and
unobservable amount of a pentacoordinated complex
that instead displays a high reactivity. Tetracoordinated
cationic rhodium(I) dicarbonyl complexes are known to
be in equilibrium with their pentacoordinated tricar-
bonyl analogues under CO and the equilibrium can
even be almost completely shifted towards the tricar-
bonyl complex [12]. However, to the best of our knowl-
edge, this is the first time that the onset of an analogous
equilibrium has been reported for an anionic complex.
The fact that the equilibrium is clearly almost com-
pletely shifted to the dicarbonyl complex side is proba-
bly responsible for the lack of earlier reports on similar
conditions, the negative effect of
a higher 2,2%-
bipyridine/Rh ratio is clearly evident and a large
amount of azoxybenzene is formed (run 1, Table 3)
even at this high temperature (compare with run 5,
Table 2). Even in this case, however, the addition of
aniline markedly increases the selectivity in carbamate.
Although we have not attempted to fully optimise all
the experimental conditions, halving the amount of
Fig. 1. Structures of the promoters used.

F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
115
the nitrogen atom of the metallacycle, but this is not
certain at this stage. To gain more information on this
point, several experiments were performed by using
some electrophiles other than protons, which could give
an irreversible interaction with the most nucleophilic
atom of 1a and allow the atom responsible for the
interaction with the hydroxylic group of 2-hydroxy-
pyridine and methanol to be identified. We recall that
previously reported experiments have excluded the two
‘CO2’ oxygen atoms as the reactive sites [6].
Reaction with p-toluolyl chloride afforded the N-acyl-
ated product (Eq. (2)). Although it may be argued that
the observed amide may derive from the intermediate
formation of aniline as an intermediate, which would
then be trapped by excess toluolyl chloride, the aprotic
conditions employed suggest that the nitrogen atom
was acylated as the first stage of the reaction. However,
the source of the nitrogen-bound hydrogen atom re-
mains uncertain, although a trace amount of p-methyl-
benzoic acid present even in the freshly distilled acyl
chloride is the most likely candidate. The analogous
reaction with acetic anhydride afforded both the N-
acylated and the N,O-diacylated products, but again no
O-monoacylated product was observed. Although we
could not isolate the acylated complexes formed as
intermediates, the identity of the isolated organic prod-
ucts represents a good indication, although not defini-
tive evidence, that the nitrogen atom of the metallacycle
is the most reactive site towards electrophiles.
Scheme 3.
The data reported in Tables 1 and 2 clearly show that
2-hydroxypiridine also accelerates the catalytic car-
bonylation reaction, but this does not a priori mean
that this is due to the acceleration of the decomposition
of 1.
2-Hydroxypyridine is known to be in tautomeric
equilibrium with its pyridonic form (Eq. (5)) and the
position of the equilibrium depends on the solvent [14].
The equilibrium is known to favour the pyridonic form
in polar solvent and at ambient temperature, but, to the
best of our knowledge, no experimental study on the
position of equilibrium (5) has been performed at high
temperatures and in solution, so that the relative
amounts of the two tautomers under the actual cata-
lytic conditions are not known.
Scheme 4.
equilibria. With this knowledge in mind, the most
reasonable explanation for the effect of pyridine, at
least in the present reaction, is that this base can also
coordinate to the fifth position of the rhodium complex
in place of CO (Scheme 3).
Interestingly, 2-hydroxypyridine reacts with 1a even
in the absence of both CO and methanol. Clearly, the
bifunctional nature of this reagent is essential in this
case. Based on what was previously said about the
interaction of methanol and pyridine with 1a, an inter-
action of the kind shown in Scheme 4 can be proposed,
where the possibility that the molecule is reacting in its
pyridonic form has also been considered.
(5)
Iridium complexes in which the nitrogen atom of a
2-hydroxypyridine is bound to the metal and the hy-
droxylic hydrogen is interacting with another ligand
have been isolated [13]. Independent of the separate use
of CO/Py and methanol or of 2-hydroxypyridine, it
appears that 1a needs simultaneous activation by a
coordinating group and an acid.
Moreover, 2-hydroxypyridine is more acidic than
methanol [15]. Given this knowledge, one may wonder
if either the existence of a tautomeric equilibrium or the
acidity of 2-hydroxypyridine (or both) is responsible for
the observed effect. To study this question, the promot-
ing effect of 4-hydroxypyridine, which has a similar
acidity and for which the same tautomeric equilibrium
It should be noted that the proton in the 2-hydroxy-
pyridine molecule has been shown as interacting with

116
F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
exists as for 2-OH isomer [15,16], was tested. However,
4-hydroxypyridine completely inhibited the reaction
(run 9, Table 2). The onset of a tautomeric equilibrium
by itself is unlikely to justify such a strong inhibiting
effect and the origin of the deactivation must lie in the
acidity of the phenolic group. This result implies that
the acidity even of the 2-OH isomer must severely
deactivate the catalytic system and that another positive
effect that overwhelms the first must be present in this
case. This effect can only be ascribed to the proximal
position of the basic and acidic sites and can be at-
tributed to some form of bifunctional activation. The
negative effect of an acidic proton is also indicated by
a comparison between the results obtained by using
2-hydroxypyridine and the ones obtained with the more
acidic 2-thiopyridine, 2-hydroxyquinazoline and hy-
droxypyrimidine. The same conclusion is supported by
the fact that picolinic acid completely inhibited the
reaction, despite its geometry being very close to that of
2-hydroxymethylpyridine, which promotes the reaction
to approximately the same extent as pyridine. Why
acids inhibit the reaction is not obvious. It should be
noted that acids have been found to promote, rather
than inhibit, the activity of the related palladium–
phenanthroline catalytic system [17], despite the fact
that the mechanism of the two catalytic cycles presents
many similarities [18]. One possibility is that acids
promote the aggregation of anionic mononuclear spe-
cies into catalytically inactive polynuclear ones by ren-
dering them neutral. We have previously shown that
the size of the countercations of [Rh(CO)4]− and
[Ir(CO)4]− has a dramatic effect in determining the
nuclearity of the products in reactions with organic
halides [19].
At this point, the problem of the identity of the
bifunctional effect responsible for the acceleration of
the catalytic reactions must be addressed. Although the
use of bifunctional promoters was originally suggested
by the results of the stoichiometric decomposition reac-
tions of 1, the data collected in order to determine the
role of this reaction in the catalytic cycle indicates that
the acceleration of the stoichiometric reaction is not
relevant to that of the catalytic reactions. Indeed, if the
role of 2-hydroxypyridine is to replace both CO and
methanol, a decrease in the concentration of this latter
reagent should maximise its effect, which is not the
case. Moreover, if the decomposition of 1a were rate
determining in the catalytic cycle, the reaction rate
should not be influenced by the addition of aniline,
which is again contrary to what is observed. This does
not mean that the interaction between 1 and 2-hydroxy-
pyridine does not occur under catalytic conditions, but
only that the decomposition of 1 is not rate determining
and its possible acceleration is not influential. Thus it
appears that there must be at least two bifunctional
effects of 2-hydroxypyridine. The data reported in
Table 3 indicate that the rate-determining step of the
cycle involves aniline and is likely to be its attack on a
carbomethoxy complex. There are several ways in
which this step may be accelerated by 2-hydrox-
ypiridine. Pyridine and substituted pyridines are known
to promote several carbonylation reactions. In the case
of the carbonylation of ethylene to methyl acetate,
catalysed by cobalt carbonyls, evidence has been re-
ported that the effect is due to the nucleophilic attack
of pyridine on the carboethoxy group of
(CO)4CoC(O)OEt, generating a reactive [PyC(O)OEt]+
intermediate [20]. Alternatively, the bifunctional nature
of 2-hydroxypyridine may allow for an assisted proton
transfer between the incoming aniline molecule and a
carbomethoxy group. Unfortunately, the failure to iso-
late any rhodium carbomethoxy complex in our reac-
tions prevented us from testing the validity of these
hypotheses and the detailed nature of this second effect
remains uncertain, although an increase in the rate of
the catalytic reactions is evident.
4. Conclusions
We have investigated the effect of several bifunc-
tional reagents on the [Rh(CO)4]−-catalysed carbonyla-
tion of nitrobenzene and some related stoichiometric
reactions. The results clearly indicate that some effects
are present that can only be ascribed to the bifunctional
nature of the reagents employed, but it has not been
possible to identify unequivocally the catalytically rele-
vant effect. The addition of 2-hydroxypyridine or 2-hy-
droxyquinoline to the catalytic system remarkably
increases the rates and selectivity under the suitable
conditions, but the acidity of these molecules also has
negative effects and these predominate under some
circumstances. On the other hand, since the first discov-
ery of the acceleration effects described here, we have
also tested the use of 2-hydroxypyridine as a promoter
in the related reduction reaction of 2-nitrostyrenes by
CO to afford arylindoles, again catalysed by
[PPN][Rh(CO)4]. In this last case, a positive effect on
both rate and selectivity was observed under all the
conditions employed [21]. Even in this case, 4-hydrox-
ypyridine had an inhibiting effect on the reaction.
Although a final clarification of the origin of the
effects observed must await for an in situ study under
catalytic conditions, it is interesting to note that 2-hy-
droxypyridine has also been found to promote the
rhodium-catalysed hydrogenation of CO to ethylenegly-
col and ethanol [22]. For these last reactions, it was
proposed that the promoting effect derived from the
ability of this reagent to supply an anion with a delo-
calised charge. However, the results reported here, indi-
cate that 2-hydroxypyridine has some more specific
properties that may be responsible for the acceleration
of even these other reactions.

F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
117
5. Experimental
1a (142 mg, 0.144 mmol) under a dinitrogen atmosphere.
The resulting brown solution was stirred at r.t. for 1 h
and 30 min and at 50°C for 1 h. During this time the
reaction mixture was analysed by IR spectroscopy. The
absorptions attributed to 1a (2050(s), 1973(s), 1638(m)
cm−1) disappeared and four new absorptions (2048(s),
2028(s), 2018(s), 1969(s) cm−1) were observed in the
carbonyl region of the spectrum. The solvent was evap-
orated to dryness in vacuo and n-hexane (20 ml) was
added to the black residue. The resulting black solid was
collected, washed with more n-hexane and dried in vacuo
(70 mg). Consistent elemental analyses could not be
obtained for this material. The formation of 3,4-
Cl2C6H3NH2 was evidenced by gas-chromatographic
analysis.
5.1. General procedure
All reactions were conducted under a dinitrogen or CO
atmosphere using either standard Schlenk apparatus or
an autoclave (see below). Solvents were dried by standard
procedures. [PPN][Rh(CO)4] was synthesised by the
method reported in the literature [23]. This complex is air
sensitive even in the solid state and has to be weighed
under an inert atmosphere. Nitrobenzene was purified by
shaking with 10% H2SO4, washing with water, and drying
with Na2SO4, followed by distillation under dinitrogen
and storage under an inert atmosphere. All other
reagents were commercial products and were used as
received. Gas chromatographic analyses were performed
on a Perkin–Elmer 8420 gas chromatograph (PS 255
capillary column) for quantitative analysis, or on a
Hewlett–Packard 5890 gas chromatograph equipped
with a 5971 A mass-selective detector for confirmation
of product identities. HPLC analyses were performed on
a Hewlett–Packard 1050 instrument equipped with a
Purospher® RP-18 e (5 mm) column. NMR spectra were
recorded on a Bruker AC 200 FT (200 MHz) or on a
Bruker AC 300 FT (300 MHz) spectrometer at room
temperature (r.t.). IR spectra were recorded on a FTS-7
Bio Rad FT-IR spectrometer. Elemental analyses and
mass spectra were recorded in the analytical laboratories
of Milan University.
5.4. Reaction of 1a with 2-hydroxypyridine
2-Hydroxypyridine (26 mg, 0.27 mmol) was added to
a CH2Cl2 (15 ml) solution of 1a (248 mg, 0.25 mmol) at
r.t. under a dinitrogen atmosphere. The resulting brown
solution was stirred at r.t. for 1 h. During this time the
reaction mixture was analysed by IR spectroscopy. The
absorptions attributed to 1a (2059(s), 1982(s), 1628(m)
cm−1) disappeared and two new absorptions (2057(s),
1984(s) cm−1) were observed in the carbonyl region of
the spectrum. The solvent was evaporated to dryness and
Et2O (20 ml) was added to the brown residue. The
resulting brown solid was collected, washed with more
n-hexane and dried in vacuo (160 mg). Consistent
elemental analyses could not be obtained for this mate-
rial. The formation of 3,4-Cl2C6H3NH2 was evidenced by
gas-chromatographic analysis.
5.2. Synthesis of [PPN][R¸h(¹C¹O¹)2¹O¹N¹(¹C¹6H¹3¹C¹l2¹)C¹(¹Oº)O]
(1a)
3,4-Cl2C6H3NO2 (350 mg, 1.82 mmol) was added at
−78°C to a solution of [PPN][Rh(CO)4] (905 mg, 1.26
mmol) in THF (15 ml) under a CO atmosphere. The
resulting yellow suspension was stirred for 30 min during
which time the temperature of the reaction was allowed
to increase to r.t. Then the reaction mixture was analysed
by IR spectroscopy and the spectrum showed that
[PPN][Rh(CO)4] had been completely consumed and, in
the carbonyl region, only the absorptions attributed to
1a (2050(s), 1973(s), 1638(m) cm−1) were present. The
flask was evacuated and filled with dinitrogen, n-hexane
(20 ml) was added to the yellow mixture to yield a yellow
microcrystalline solid that was collected by filtration,
washed with n-hexane (5 ml) and dried in vacuo (86%).
Contrary to the previously reported procedure [6], the
formation of a black solid was not observed. The
elemental analysis and spectroscopic data are in agree-
ment with those previously reported by us [6].
5.5. Reaction of 1a with p-toluolyl chloride
p-Toluolyl chloride (0.021 ml, 0.16 mmol) was added
to a THF (10 ml) solution of 1a (158 mg, 0.160 mmol)
at r.t. under a dinitrogen atmosphere. The resulting
brown solution was stirred at r.t. for 1 h. During this time
the reaction mixture was analysed by IR spectroscopy.
The absorptions attributed to 1a (2059(s), 1982(s),
1628(m) cm-1) disappeared and two new absorptions
(2054(s), 1974(s) cm−1) were observed in the carbonyl
region of the spectrum. The reaction mixture was concen-
trated to one third of its original volume in vacuo and
Et2O (20 ml) was added to precipitate a brown solid. The
mother liquor was analysed by GC–MS spectroscopy to
identify 4-CH3C6H4C(O)NH(3,4-Cl2C6H3), whereas the
brown precipitated solid was collected, solubilised in
THF and identified by IR spectroscopy as
[PPN][Rh(CO)2Cl2] [24].
5.3. Reaction of 1a with methanol and pyridine
5.6. Reaction of 1a with acetic anhydride
A total of 0.12 ml of CH3OH and 0.050 ml of pyridine
were added at r.t. to a THF (15 ml) solution of complex
Acetic anhydride (17 ml, 0.18 mmol) was added to a
THF (10 ml) solution of 1a (151 mg, 0.153 mmol) at

118
F. Ragaini et al. / Journal of Organometallic Chemistry 593–594 (2000) 109–118
−78°C under a dinitrogen atmosphere. The resulting
brown solution was stirred for 30 min during which
time the temperature of the reaction was allowed to
increase to r.t. During this time the reaction mixture
was analysed by IR spectroscopy. The absorptions at-
tributed to 1a (2059(s), 1982(s), 1628(m) cm−1) disap-
peared and new absorptions (2054(s), 1972(s), 1621(m)
cm−1) were observed in the carbonyl region of the
spectrum. The resulting brown solution was stirred at
r.t. for 1 h. The reaction mixture was evaporated to
dryness and Et2O (20 ml) was added to the residue. The
resulting yellow powder was collected and recrystalliza-
tion from THF–n-hexane gave crystals suitable for
X-ray analysis, showing it to be [Rh(CO)2-
(CH3COO)2][PPN]. The mother liquor was evaporated
to dryness and the organic products were separated by
flash chromatography (4:1 ethyl acetate–n-hexane) to
obtain Cl2C6H3NHC(O)Me (A) and Cl2C6H3N-
(C(O)Me)OC(O)CH3 (B) in 37 and 30% yield, respec-
ated and filled with dinitrogen three times. CO was then
charged at r.t. at the required pressure and the auto-
clave was immersed in an oil bath preheated to the
required temperature. At the end of the reaction, the
autoclave was cooled with an ice bath, vented and the
products were analysed by gas chromatography (naph-
thalene as internal standard).
References
[1] M. Sawamura, Y. Ito, Chem. Rev. 92 (1992) 857.
[2] M. Shibasaki, H. Sasai, T. Arai, Angew. Chem. Int. Ed. Engl. 36
(1997) 1236.
[3] E.K. van der Beuken, B.L. Feringa, Tetrahedron 54 (1998)
12985.
[4] F. Ragaini, S. Cenini, A. Fumagalli, C. Crotti, J. Organomet.
Chem. 428 (1992) 401.
[5] S. Cenini, F. Ragaini, Catalytic Reductive Carbonylation of
Organic Nitro Compounds, Kluwer, Dordrecht, 1997.
[6] F. Ragaini, S. Cenini, F. Demartin, Organometallics 13 (1994)
1178.
1
tively. H-NMR (CDCl3, 200 MHz, 298 K) l, ppm: (A)
7.74 (s, 1H); 7.34 (s, 1H, NH); 7.33 (m, 2H); 2.18 (s,
3H). (B) 7.59 (s, 1H); 7.48 (d, 1H, J=8.6 Hz); 7.34 (d,
1H, J=8.6 Hz); 2.17 (s, 3H); 2.12 (s, 3H).
[7] J.D. Gargulak, W.L. Gladfelter, J. Am. Chem. Soc. 116 (1994)
3792.
[8] F. Demartin, personal communication.
[9] A. Fulford, N.A. Bailey, H. Adams, P.M. Maitlis, J. Organomet.
Chem. 417 (1991) 139.
5.7. Reaction of 1a with aniline
[10] F. Ragaini, M. Macchi, S. Cenini, J. Mol. Catal. A. 127 (1997)
33.
Aniline (74 ml, 0.82 mmol) was added to a THF (10
ml) solution of 1a (203 mg, 0.226 mmol) at r.t. under a
CO atmosphere. The resulting brown solution was
stirred for 4 h, during which time the reaction mixture
was analysed by IR spectroscopy. The absorptions at-
tributed to 1a (2059(s), 1982(s), 1628(m) cm−1) disap-
peared and an absorption (1896(s) cm−1) due to
[Rh(CO)4]− was observed in the carbonyl region of the
spectrum. PhNHC(O)NHPh and 3,4-Cl2C6H3NHC-
(O)NHPh were also present in the reaction mixture in a
41:9 ratio (HPLC analysis).
[11] F. Ragaini, A. Ghitti, S. Cenini, Organometallics in press.
[12] G. Mestroni, A. Camus, G. Zassinovich, J. Organomet. Chem.
65 (1974) 119.
[13] E. Peris, J.C. Lee Jr., J.R. Rambo, O. Eisenstein, R.H. Crabtree,
J. Am. Chem. Soc. 117 (1995) 3485.
[14] J.M. Rawson, R.E.P. Winpenny, Coord. Chem. Rev. 139 (1995)
313.
[15] A. Albert, J.N. Phillips, J. Chem. Soc. (1956) 1294.
[16] J. Frank, A. Katritzky, J. Chem. Soc. Perkin Trans. 2 (1976)
1428.
[17] (a) P. Wehman, L. Borst, P.C.J. Kamer, P.W.N.M. van
Leeuwen, J. Mol. Catal. A 112 (1996) 23. (b) P. Wehman, P.C.J.
Kamer, P.W.N.M. van Leeuwen, Chem. Commun. (Cambridge)
(1996) 217.
[18] (a) F. Ragaini, S. Cenini, 11th Int. Symp. Homogeneous Cataly-
sis, St. Andrews, UK, 1998. (b) E. Gallo, F. Ragaini, S. Cenini,
F. Demartin, J. Organomet. Chem. 586 (1999) 190.
[19] (a) F. Ragaini, F. Porta, A. Fumagalli, F. Demartin,
Organometallics 10 (1991) 3785. (b) F. Ragaini, F. Porta, F.
Demartin, Organometallics 10 (1991) 185. (c) F. Porta, F. Ra-
gaini, S. Cenini, M. Pizzotti, F. Demartin, Organometallics 9
(1990) 929.
[20] N.S. Imyanitov, N.M. Bogoradovskaya, T.A. Semenova, Kinet.
Katal. 19 (1978) 573 [Engl. transl. in Kinet. Catal. 19 (1978)
452].
[21] F. Ragaini, S. Tollari, S. Cenini, E. Bettetini, J. Mol. Catal. A
111 (1996) 91.
5.8. Catalytic reactions
In a typical reaction, nitrobenzene and, if required,
aniline and the organic promoter (see the Tables for
amounts) were weighed in a glass liner. The liner was
placed inside a Schlenk tube with a wide mouth, frozen
at dry ice temperature, evacuated and filled with dini-
trogen, after which [PPN][Rh(CO)4] (separately
weighed under dinitrogen) and the solvents were added.
After the solvent was also frozen, the liner was closed
with a screw cap having a glass-wool-filled open mouth
that allows for exchange of gaseous reagents and
rapidly transferred to a 200 ml stainless steel autoclave
with magnetic stirring. The autoclave was then evacu-
[22] B.D. Dombek, Adv. Catal. 32 (1983) 325.
[23] L. Garlaschelli, R. Della Pergola, S. Martinengo, Inorg. Synth.
28 (1990) 211.
[24] L.M. Vallarino, Inorg Chem. 4 (1965) 161.