Carbonylation of dinitrotoluene to dimethyl toluenedicarbamate; high efficiency of phosphorus acids as promoters for the palladium-phenanthroline catalytic system

Article abstract of DOI:10.1002/adsc.200404151

Phosphorus acids are excellent promoters for the palladium-phenanthroline catalyzed carbonylation of 2,4-dinitrotoluene to 2,4-toluenedicarbamate. For the first time, all intermediate nitrocarbamates and aminocarbamates have been independently synthesized and their amount after every catalytic reaction precisely quantified. An extensive optimization of all experimental variables has been carried out. The best acids are phenylphosphonic and 4-tolylphosphonic acids. The addition of 2,2-dimethoxypropane as an internal drying agent is highly beneficial. The addition of an amine derived from the starting dinitroarene increases both rate and selectivity of the carbonylation reaction. The complexes [Pd(Phen)₂] [SbF₆] and [Pd(Phen) ₂][BAr^F₄] [Ar^F = 3,5-(CF ₃)₂C₆H₃] have been prepared for the first time. The latter displays a markedly higher solubility than all other [Pd(Phen)₂]²⁺ complexes. The effect of several possible promoters has also been investigated. Under the optimized experimental conditions, a 77.6% selectivity in dicarbamate was obtained when working at a molar ratio dinitrotoluene/Pd = 2920. At the end of the reaction, the dicarbamate spontaneously precipitates out of the solution in high yields upon cooling, with no inclusion of the acid promoter or of phenanthroline. 2,6-Dinitrotoluene can also be efficiently carbonylated to the corresponding dicarbamate.

Full text of DOI:10.1002/adsc.200404151

FULL PAPERS
Carbonylation of Dinitrotoluene to Dimethyl
Toluenedicarbamate; High Efficiency of Phosphorus Acids as
Promoters for the Palladium-Phenanthroline Catalytic System
Michela Gasperini, Fabio Ragaini,* Chiara Cazzaniga, Sergio Cenini
Dipartimento di Chimica Inorganica, Metallorganica e Analitica, ISTM-CNR, and INSTM, UdR Milano, Via Venezian 21,
20133 Milano, Italy
Fax: (þ39)-2-5031-4405, e-mail: fabio.ragaini@unimi.it
Received: June 1, 2004; Accepted: October 6, 2004
Supporting Information for this article is available on the WWW under http://asc.wiley-vch.de/home/.
Abstract: Phosphorus acids are excellent promoters have been prepared for the first time. The latter dis-
for the palladium-phenanthroline catalyzed carbony- plays a markedly higher solubility than all other
lation of 2,4-dinitrotoluene to 2,4-toluenedicarba- [Pd(Phen)2]2þ complexes. The effect of several possi-
mate. For the first time, all intermediate nitrocarba- ble promoters has also been investigated. Under the
mates and aminocarbamates have been independent- optimized experimental conditions, a 77.6% selectivi-
ly synthesized and their amount after every catalytic ty in dicarbamate was obtained when working at a
reaction precisely quantified. An extensive optimiza- molar ratio dinitrotoluene/Pd¼2920. At the end of
tion of all experimental variables has been carried the reaction, the dicarbamate spontaneously precipi-
out. The best acids are phenylphosphonic and 4-tolyl- tates out of the solution in high yields upon cooling,
phosphonic acids. The addition of 2,2-dimethoxypro- with no inclusion of the acid promoter or of phenan-
pane as an internal drying agent is highly beneficial. throline. 2,6-Dinitrotoluene can also be efficiently
The addition of an amine derived from the starting di- carbonylated to the corresponding dicarbamate.
nitroarene increases both rate and selectivity of the
carbonylation reaction. The complexes [Pd(Phen)2] Keywords: carbamates; carbonylation; homogeneous
[SbF6] and [Pd(Phen)2][BArF ] [ArF ¼3,5-(CF3)2C6H3] catalysis; nitroarenes; palladium
4
Introduction
Moreover, carbonylation of dinitrotoluene is a difficult
task and many catalytic systems that show a good activ-
2,4-Toluenediisocyanate (TDI) is a key component for ity in the carbonylation of nitrobenzene fail completely
the production of polyurethanes and is actually pro- to carbonylate dinitrotoluene or afford at best very poor
duced worldwide on a more than 5 millions tons per yields.[1–3] One reason for this is that many catalytic cy-
year scale by a process employing the highly toxic phos- cles proceed through an intermediate reduction of the
gene. Carbonylation of nitroarenes in the presence of al- nitroarene to the corresponding aniline, which is only
cohols to afford carbamates [Eq. (1)], followed by ther- later carbonylated. In the case of dinitroarenes, the nitro
mal cracking to give the corresponding isocyanate and amino groups that are formed in this first step deac-
[Eq. (2)] has been considered as a safer alternative to- tivate each other; the nitro group decreasing the nucle-
wards these products.[1–3] However, no commercial ophilic power of the amino one and the latter decreasing
process has been developed up to now that can compete the oxidizing power of the first. In recent years the cata-
with the actual technology, mainly because of the insuf- lytic system based on palladium-phenanthroline com-
ficient turnover numbers achievable even with the best plexes has emerged as the most active and promising
catalytic systems and the associated cost of recycling for a possible industrial application.[4–35] This is also
the metal and the ligand.
one of the few catalytic systems that have been reported
to be successful even for dinitroarenes, although selec-
tivities are only moderate. Very recently, we reported
that phosphorus acids such as phenylphosphonic and es-
pecially phosphoric acid are excellent promoters for the
palladium-phenanthroline catalytic system and allowed
us to reach an unprecedented turnover frequency of
ð1Þ
ð2Þ
Adv. Synth. Catal. 2005, 347, 105–120
DOI: 10.1002/adsc.200404151
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
105

Michela Gasperini et al.
FULL PAPERS
6000/h and a total turnover number of 105 in the case of
nitrobenzene.[34,35] We have now extended the use of the
new promoters to the carbonylation of dinitrotoluene
and performed an optimization study of the experimen-
tal conditions.
ð3Þ
Results and Discussion
Reaction of 5 with methyl chloroformate in the presence
of ZnCl2 and triethylamine gave 3a almost exclusively
[Eq. (4)] and the pure product could be isolated by crys-
tallization free of any 3b.
Independent Synthesis of the Organic Intermediates or
Byproducts
The first problem which is encountered when dealing
with the carbonylation of 2,4-dinitrotoluene is the large
number of possible intermediates and byproducts.
Apart from the desired dicarbamate (1), there are two
nitrocarbamates (2a, b), two aminocarbamates (3a, b),
two nitrotoluidines (4a, b), diaminotoluene (5)
(Scheme 1) and a virtually unlimited number of azo-
ð4Þ
and azoxyarenes, and ureas. In this work we have inde- Compound 3b was obtained by reduction of the corre-
pendently synthesized all the nitro- and aminocarba- sponding nitrocarbamate by Fe3(CO)12, following our
mates and quantified all the products reported in modification[37] of the procedure originally reported by
Scheme 1 for all reactions.
Alper and coworkers.[38] This procedure is very selective
Despite the fact that the 2 and 3 are often mentioned and quite convenient on the laboratory scale [Eq. (5)].
as byproducts in carbonylation reactions of 2,4-dinitro- The carbamate group was not reduced or hydrolyzed
toluene, an independent synthesis and a full characteri- to any detectable extent under these conditions.
zation of these compounds has not been previously re-
ported in the literature.
The two nitrocarbamates 2 were obtained by reaction
of the corresponding nitroanilines 4 with methyl chloro-
formate in the presence of metallic zinc [Eq. (3)], fol-
lowing a protocol already shown in the literature to
give good results for the synthesis of carbamates in
ð5Þ
which the aryl group bears electron-withdrawing sub-
stituents.[36]
By analogy with the related carbonylation of nitroben-
zene, it is expected that the main byproducts will be
azoxy compounds. However, as previously mentioned,
many different such compounds may be formed and
we did not consider it worth targeting the synthesis of
any of them, even because no other clearly defined
peak could be observed in the HPLC analyses of most
of the catalytic reactions, apart for those attributable
to the compounds in Scheme 1. Only when a reaction
was performed in the absence of promoters, a red com-
pound could be isolated from the reaction mixture,
which was identified by its mass spectrum as
¼
(O2N)(CH3)C6H3-N N(O)C6H3(CH3)(NO2), but the
exact substitution pattern (nitro groups in the 2- or 4-po-
sition) could not be assessed.
Synthesis of the Catalysts
Throughout most of this study we employed
[Pd(Phen)2][BF4]2 as the catalyst. The non-coordinating
Scheme 1. Numbering scheme for the products derived from
2,4-dinitrotoluene.
106
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 105–120

Carbonylation of Dinitrotoluene to Dimethyl Toluenedicarbamate
FULL PAPERS
nature of the BF4ꢀ is essential in order for high catalytic
activity to be observed. The use of the corresponding
hexafluorophosphate salt has also been reported in the
literature, but, in our previous work on the carbonyla-
tion of nitrobenzene, we showed this compound to
give slightly worse results in combination with phospho-
rus acids. In our quest for more reactive catalysts for the
carbonylation reactions of nitroarenes, we decided to
target the synthesis of the [Pd(Phen)2]2þ salts with the
SbF6ꢀ and BArF ꢀ [ArF ¼3,5-(CF3)2C6H3] anions. Hexa-
4
fluoroantimonate isꢀknown to be chemically very sta-
ble, whereas BArF is one of the least interacting
4
anions known to date. It generally imparts to its com-
pounds solubilities in low or medium polarity solvents
not achievable with any other anion and the corre-
sponding [Pd(Phen)2]2þ salt may become useful as cat-
alyst even for other reactions when this feature is rele-
ꢀ
vant. It should also not be forgotten that the BArF
4
shows a much lower tendency to transfer one of its
ꢀ
aryl groups as compared to the analogous BPh4
anion.[39,40]
Basically, three different syntheses have been report-
ed for complexes of the type [Pd(Phen)2][X]2 for which
the corresponding palladium salt is not available:
1) Anion exchange between [Pd(Phen)2][NO3]2 and
NaX, the first being directly obtained from palladi-
Scheme 2. Synthesis of [Pd(Phen)2][BArF ]2.
4
um nitrate and Phen.[7]
2) Reaction of PdCl2(CH3CN)2 with AgX in CH3CN,
to intermediately afford [Pd(CH3CN)4][X]2, fol- pared to the ones with other anions prevented its purifi-
lowed by reaction with Phen.[14]
cation from excess amounts of any of the two starting re-
3) Reaction of Pd(OAc)2 with the phenanthrolinium agents by crystallization.
salt [PhenH][X], obtained by reaction of Phen
We thus decided to follow a double exchange strategy
with the corresponding acid.[41] The [PhenH][X] (Scheme 2). First NaBArF [42,43] and [PhenH][Cl] were
4
salt can also be prepared in some cases by reaction independently synthesized. The two compounds were
of Phen with HCl in the presence of a salt of Xꢀ dissolved in methanol and the resulting solution was
(e.g., NH4X), provided the desired compound is dried under vacuum. Treatment of the residue with di-
less soluble than [PhenH][Cl].[41]
ethyl ether left NaCl as a solid residue, which was fil-
tered off. Evaporation of the ether solution afforded
The first method requires the desired [Pd(Phen)2][X]2 to the crude yellowish product, which was purified by
be less soluble (in acetone) than the starting nitrate an- Soxhlet extraction of the impurities with hexane to af-
alogue. The second requires the availability of the silver ford the colorless analytically pure [PhenH][BArF ].
4
salt of the final anion and the third necessitates of the
Treating Pd(OAc)2 with two equivalents of [PhenH]
corresponding acid or the solubility order mentioned [BArF ] in acetone gave the desired product and
4
should be fulfilled.
AcOH. It should be noted that the original recipe for
The conditions for the first method are met by the this last step employs a 50% molar excess of [PhenH][X]
ꢀ
SbF6 anion and the corresponding palladium-phenan- with respect to stoichiometry, and the insoluble product
throline complex was synthesized without difficulties is isolated by filtration. However, [Pd(Phen)2][BArF ] is
4 2
by this strategy (see Experimental Section for details). completely soluble in acetone under the conditions em-
However, the conditions for neither of the first two ployed. Thus it is essential to use the exact reagent
methods are met by BArF ꢀ. Moreover, the acid HBArF
amounts in order to obtain a pure product. The product
4
4
has been reported in the literature (as the etherate),[42] is isolated by drying the reaction solution under vacuum
but its synthesis is very delicate and the compound is sta- and washing the residue with hexane, to afford the ana-
ble only at low temperatures. Despite some attempts, we lytically pure product as a pale brown powder.
could not obtain analytically pure samples of [PhenH]
The solubility of [Pd(Phen)2][BArF ] is exceptional
4 2
[BArF ] by reaction of the so obtained acid with phenan- when compared to the one of the related complexes
4
throline, because the high solubility of this salt com- with other anions. It is well soluble in acetone, CH2Cl2
Adv. Synth. Catal. 2005, 347, 105–120
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
107

Michela Gasperini et al.
FULL PAPERS
and THF, although it is only slightly soluble in diethyl benzene as a substrate (Table 1). In the tables, the selec-
ether. On the contrary, the other complexes of the tivity in total carbamates (1þ2aþ2bþ3aþ3b) is also
[Pd(Phen)2][X]2 type are only sparingly soluble in reported. Compound 3b was obtained with
a
CH2Cl2, CHCl3, PhNO2 and DMSO and are essentially selectivity<1% in most cases and its exact quantifica-
insoluble in all of the other solvents, including THF tion was often problematic. Its amount has been includ-
and acetone. This high solubility is a very important fea- ed in the total carbamates value only when a reproduci-
ture from the point of view of catalytic applications, ble integration of its HPLC signal was observed.
since it considerably expands the range of solvents that
From the data reported it immediately emerges that in
can be employed with this kind of palladium-phenan- the absence of any additive, the conversion is low even at
throline complexes.
a relatively low catalytic ratio and only trace amounts of
dicarbamate are formed. We have previously shown[34,35]
that the addition of aniline strongly accelerates the car-
bonylation reaction of nitrobenzene, especially at short
reaction times. In the present system, addition of diami-
notoluene also markedly improves the dinitrotoluene
conversion, but the amount of dicarbamate remains
Catalytic Reactions
General Trends
A few general trends can be highlighted before discus- low. Upon addition of phosphoric acid, however, larger
sing the individual experiments. amounts of 1 are formed. Increasing the catalytic ratio
a) Dinitrotoluene is very reactive and is generally com- three-fold obviously results in a decreased conversion,
pletely consumed in the first part of the reaction, which can however be raised by the addition of
while nitrocarbamates (2) accumulate in solution.
DMOP (2,2-dimethoxypropane) as an internal drying
b) As may be expected, the methyl group on the arene agent. A positive effect of DMOP has been previously
exerts a moderate steric hindrance on the adjacent reported in the literature for these types of reac-
nitro group and 2a and 3a are always more abun- tions[20–24] and we also employed it in the case of nitro-
dant in solution than 2b and 3b, respectively, al- benzene.[34,35] The catalytic ratio can further be doubled,
though the ratio may be close to 1:1 in some cases. provided the reaction time is also increased to 3 h.
c) Diaminotoluene (5) is also very reactive and is nev-
Given the importance of the addition of an aniline, we
er detected at the end of the reaction even in those investigated it more deeply. In the case of dinitroto-
cases in which it was intentionally added to the re- luene, three different amines can be added: 4a, b and
action mixture since the beginning (see later).
5. A comparison among them (Table 2) shows that 4b
is the most effective, as may be expected based on the
observation that it is the least nucleophilic of the three,
being deactivated by both electronic and steric effects.
Indeed, if the rate determining step in the catalytic cycle
Influence of the Addition of an Aniline
The first set of reactions was run under the same exper- involves the nucleophilic attack of an amine on a palla-
imental conditions previously optimized by us for nitro- dium complex, 4b will be involved in the slowest reac-
Table 1. Effect of the addition of 2,4-diaminotoluene (5).[a]
Dinitro- Promoter
tol./Pd
Pro-
moter/
Pd
5/Pd DMPO t [h] Dinitrotol. 1 sel. 2a sel. 2b sel. 3a sel. Total
4a sel. 4b sel.
[mL]
conv.
[%][c] [%][c] [%][c] [%][c] carba-
mates[d]
[%][c] [%][c]
[%][b]
450
450
450
1350
1350
2700
2700
2700
2700
–
–
–
–
–
50
50
50
50
50
50
50
50
–
–
–
–
0.5
0.5
0.5
0.5
0.5
1.5
1.5
1.5
1.5
1.5
1.5
3
53.3
98.0
100
50.4
100
78.1
100
100
0.2
10.0
44.5
5.5
38.6
3.3
13.3
24.5
26.4
18.7
13.9
–
14.4
7.9
5.9
5.8
9.6
20.6
18.8
9.3
–
33.3
32.6
61.3
17.6
74.9
51.3
63.5
54.6
58.6
<1
1.1
<1
<1
<1
0.4
<1
<1
16.7
15.3
3.3
11.8
<1
4.1
3.9
–
0.8
10.9
–
4.6
–
1.5
1.7
1.9
H3PO4 85% 174
H3PO4 85% 174
H3PO4 85% 174
H3PO4 85% 174
H3PO4 85% 174
H3PO4 85% 174
H3PO4 85% 350
6.3
22.1
27.4
29.9
19.1
19.0
5
5
100
11.3
1.3
–
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼5.0 mg, 7.8ꢁ10ꢀ3 mmol, mol ratio Phen/catalyst¼20, DMPO¼2,2-
dimethoxypropane, PCO ¼60 bar, T¼1708C, in methanol (10 mL). Conversion of 5 was complete in all cases. Selectivity in
3b<1%.
Calculated with respect to the starting dinitrotoluene.
Calculated with respect to the sum of reacted 5 and dinitrotoluene.
Sum of the selectivities in 1–3.
[b]
[c]
[d]
108
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 105–120

Carbonylation of Dinitrotoluene to Dimethyl Toluenedicarbamate
FULL PAPERS
Table 2. Influence of the addition of 4a, b and 5.[a]
Amine
Amine conv. [%][b]
1 sel. [%][c]
2b sel. [%][c]
2a sel. [%][c]
3a sel. [%][c]
Total
carbamates[d]
5
4a
4b
100
26.4
32.0
52.6
11.3
9.7
12.4
19.0
21.6
22.5
1.9
2.4
2.4
58.6
65.7
89.9
68.7
~ 0[e]
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼5.0 mg, 7.8ꢁ10ꢀ3 mmol, mol ratios 2,4-dinitrotoluene/added amine/
H3PO4/Phen/Pd¼2700: 50: 350: 20: 1, 2,2-dimethoxypropane¼0.5 mL, PCO ¼60 bar, T¼1708C, in methanol (10 mL), for
5 h. Conversion of dinitrotoluene was complete in all cases. Selectivity in 3b<1%. A very small amount (selectivity 0.9–
1.3%) was observed for 4a, b in those reactions in which they were not added since the beginning, but no 5 was ever de-
tected at the end of the reaction.
Calculated with respect to the starting amine.
Calculated with respect to the sum of reacted amine and dinitrotoluene.
Sum of the selectivities in 1–3.
The amount of 4b at the end of the reaction was indistinguishable within the experimental error from the added one.
[b]
[c]
[d]
[e]
Table 3. Comparison of carboxylic and phosphorus acids.[a]
Dinitrotol./Pd Promoter
1 sel. [%][b] 2a sel. [%][b] 2b sel. [%][b] 3a sel. [%][b] Total
4a sel. [%][b]
carbamates[c]
365
365
730
730
730
PhCOOH
85% H3PO4
PhCOOH
85% H3PO4
PhPO(OH)2 80.5
85% H3PO4 32.1
PhPO(OH)2 37.1
60.2
71.1
35.1
60.1
12.7
5.9
27.7
9.2
<1
28.4
20.5
6.7
<1
15.9
7.8
<1
14.4
9.5
8.5
11.8
2.7
4.4
<1
88.1
88.8
81.4
81.5
80.6
77.9
70.8
<1
2.2
<1
1.6
<1
1.2
1.7
1460
1460
3.0
3.7
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼12.8 mg, 0.020 mmol, mol ratios promoter/Phen/Pd¼100: 4: 1, 2,2-
dimethoxypropane¼0.5 mL, PCO ¼60 bar, T¼1358C, in methanol (10 mL), for 5 h. Conversion of dinitrotoluene was com-
plete in all cases. A selectivity in 3b and 4b ꢂ 1% was observed, but no 5 was ever detected.
Calculated with respect to the starting dinitroarene.
[b]
[c]
Sum of the selectivities in 1–3.
tion in all the cycle. Since 4b is both formed and con- the case of phosphoric acid. An even better 80% selec-
sumed during the reaction, its concentration will reach tivity is obtained with phenylphosphonic acid. The
a steady state anyway. However, the deliberate addition 60% selectivity in 1 is markedly higher than the one re-
of more 4b will alter the situation, resulting in a higher ported by van Leeuwen (31%) under very similar condi-
steady state concentration of this amine, thus increasing tions,[18,44] but in the absence of 2,2-dimethoxypropane,
the rate of both the slow step and the global catalytic re- and further highlights the importance of the addition
action.
of this dehydrating agent in the present system. Howev-
er, even under these conditions doubling the dinitroto-
luene amount causes a drop in the selectivity in 1.
Comparison with Other Experimental Conditions
Since the results were not fully satisfactory, we decided Optimization of the Amount of DMOP
to test also the experimental conditions optimized by
van Leeuwen for the related catalytic system promoted Results concerning the optimization of the amount of
by carboxylic acids[18] (Table 3), although we also added DMOP are reported in Table 4.
DMOP, which had been shown to have a positive effect.
The amount of 1 produced increases with the DMOP
At the previously reported dinitroarene/Pd mol ratio concentration up to a maximum and then decreases, but
of 365, both benzoic and phosphoric acid are effective, the total amount of carbamates increases steadily. This
with the second affording a somewhat higher yield in di- at first sight surprising behavior is due to the complex
carbamate. However, if the dinitrotoluene amount is role water and anilines play in this catalytic system.
doubled, the selectivity in 1 drops to 35% when benzoic An in-depth discussion of the reaction mechanism is
acid is employed, while it only marginally falls to 60% in outside the scope of this paper and only a brief descrip-
Adv. Synth. Catal. 2005, 347, 105–120
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
109

Michela Gasperini et al.
FULL PAPERS
Table 4. Effect of the amount of DMOP.[a]
DMOP [mL]
1 sel. [%][b]
2a sel. [%][b]
2b sel. [%][b]
3a sel. [%][b]
Total
4a sel. [%][b]
carbamates[c]
–
0.5
1
16.1
35.8
45.5
36.9
23.9
18.2
19.1
27.9
15.0
10.1
9.3
1.5
3.5
5.3
6.0
56.5
67.6
79.2
86.0
<1
1.3
<1
<1
2
18.2
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼12.8 mg, 0.020 mmol, mol ratios 2,4-dinitrotoluene/4b/85% H3PO4/
Phen/Pd¼1460: 50: 110: 4: 1. The amount of 4b at the end of the reaction was indistinguishable within the experimental
error from the initially added one. PCO ¼60 bar, T¼1508C, in methanol (10 mL), for 5 h. Dinitrotoluene conversion was
complete in all cases. A selectivity in 3b ꢂ 1% was observed, but no 5 was ever detected.
Calculated with respect to the sum of reacted dinitrotoluene and 4b.
[b]
[c]
Sum of the selectivities in 1–3.
tion of the reactions involved is given here. Studies con- Since DMOP reacts with water in competition to the
ducted in our group on nitrobenzene as a substrate have process in Eq. (6), the observed trend is exactly the
shown that at least three mechanisms are operating at one that may have been foreseen based on the preceding
the same time, but under the normal operating condi- discussion and comes as no surprise. In particular, the
tions one strongly prevails.[45] In this mechanism nitro- low amount of 1 at high DMOP concentration is due
benzene is intermediately reduced to aniline, which is to a slowing down of the reaction, which affects the sec-
in turn carbonylated to phenyl isocyanate, this step be- ond step in a more evident way, whereas the high selec-
ing rate-determining. When water is present, this can re- tivity in the formation of carbonylated products keeps
act with nitrobenzene and CO to afford aniline [Eq. (6)] the total carbamates amount very high.
and the increased steady state concentration of aniline
causes a global rate increase in the catalytic reaction.
Optimization of Other Experimental Conditions
ð6Þ
Results concerning the optimization of the promoter,
phenanthroline, solvent and 4b amounts are reported
During the reaction, water continuously reforms due to in Table 5. The following points are worthy of comment:
the hydrogen transfer reduction of nitrobenzene by the a) Increasing the solvent amount from 10 to 20 or
alcohol, a reaction which is slowest for methanol com-
pared with other alcohols, but cannot be completely sup-
pressed, Eq. (7).
30 mL improves both conversion of the intermedi-
ates 2 and selectivity in 1, the best volume depend-
ing also on the amount of promoter. A rate in-
crease upon dilution is not a common feature and
is unlikely to be due to any kinetic law. However,
Paul et al. have shown that when the carbonylation
reaction of nitrobenzene catalyzed by palladium-
phenanthroline complexes is run in neat nitroben-
zene, several complexes form which are catalytical-
ly inactive.[26] It is possible that similar inactive
complexes are formed in the case of dinitrotoluene
when its concentration is too high.
ð7Þ
However, it must also been considered that when the
amount of aniline in solution increases too much, the se-
lectivity in carbonylated products drops and large
amounts of azoxybenzene are formed.[16,45] The reason
for this change in selectivity is not known, but it may
be simply related to the basicity of anilines. Indeed, it b) As expected, the addition of 4b increases rate and
has been observed under various experimental condi-
tions that the addition of any base, including weak bases
such as carboxylates,[15] slows down the reaction and
shifts it towards the formation of azo- and azoxyarenes,
whereas acids increase both the rate of the reaction and
selectivity, but if too large amounts are added, se-
lectivity in 1 decreases (entries 6, 8–11). This trend
parallels, in the opposite sense, the effect of in-
creasing amounts of DMOP and is clearly connect-
ed to this.
the amount of carbonylated products. Thus it is clear c) Use of 3,4,7,8-tetramethylphenanthroline as a li-
that a moderate amount of water will lead to an amount
of aniline in solution that is beneficial for the reaction
rate, without compromising selectivity, but a larger
amount will cause the aniline amount to increase too
much, causing a decreased selectivity in carbamates.
gand in place of unsubstituted phenanthroline
gave worse results (runs 10 and 14). This is a bit sur-
prising, since usually this substituted phenanthro-
line gives better results in carbonylation reactions
of nitroarenes catalyzed by palladium. Since
110
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 105–120

Carbonylation of Dinitrotoluene to Dimethyl Toluenedicarbamate
FULL PAPERS
Table 5. Effect of the amount of promoters under the conditions of Tables 3 and 4.[a]
Run PhPO(OH)2/Pd 4b/Pd 4b conv. 4b sel. Phen/ MeOH 1 sel. 2a sel. 2b sel. 3a sel. Total
4a sel.
[%][b]
[%][c] cat.
[mL]
[%][c] [%][c] [%][c] [%][c] carbamates[d] [%][c]
1
2
3
4
5
6
7
8
9
110
220
220
220
330
165
165
165
165
165
165
50
50
50
–
50
50
50
–
75
100
200
50
100
100
200
ꢂ0
ꢂ0
38.4
–
ꢂ0
ꢂ0
ꢂ0
–
ꢂ0
22.6
60.9
ꢂ0
57.3
60.3
–
–
–
–
6.4
–
–
–
5.0
–
–
4
4
8
10
10
20
20
30
30
30
30
30
30
30
30
30
30
30
47.3
33.6
60.8
53.5
55.6
56.1
43.2
53.0
56.9
62.9
41.9
53.9
66.9
49.6
80.6
22.4
23.9
16.3
21.1
17.9
15.5
20.9
20.1
14.6
12.1
14.8
21.4
7.9
12.6
13.7
6.7
9.1
8.4
6.9
11.4
9.6
6.2
5.3
8.3
9.9
5.7
12.2
3.2
4.8
1.5
4.5
3.8
3.9
4.6
6.3
4.2
4.6
5.1
4.1
4.9
4.7
4.7
8.6
87.1
72.7
88.3
87.5
85.8
83.1
81.8
86.9
82.3
85.4
69.1
90.1
85.2
88.9
98.8
<1
<1
<1
8
2.1
12
12
24
12
12
12
12
12
12
12
24
<1
<1
<1
<1
<1
<1
<1
<1
10
11
–
<1
–
–
12[e] 165
13[f] 165
14[g] 165
15[h] 330
2.1
<1
<1
22.4
6.4
6.6
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼12.8 mg, 0.020 mmol, mol ratios 2,4-dinitrotoluene/Pd¼1460,
DMOP¼1 mL, PCO ¼60 bar, T¼1508C, in methanol (10 mL), for 5 h. In the cases where a conversion of 4bꢂ0 is indicated,
its amount at the end of the reaction was indistinguishable within the experimental error from the initially added one. Di-
nitrotoluene conversion was complete in all cases. A selectivity in 3b ꢂ 1% was observed, but no 5 was ever detected.
Calculated with respect to the initially added 4b.
Calculated with respect to the sum of reacted dinitrotoluene and 4b.
Sum of the selectivities in 1–3.
[b]
[c]
[d]
[e]
[f]
DMOP¼2 mL.
t¼7 h.
[g]
[h]
3,4,7,8-Tetramethylphenanthroline was employed instead of Phen.
[Pd(Phen)2][BF4]2 ¼6.4 mg, 0.010 mmol.
3,4,7,8-tetramethylphenanthroline is also much which is not commercially available and must be synthe-
more expensive than phenanthroline, its use was sized. Phosphoric acid and diphenylphosphinic acid
not investigated further.
d) Halving both palladium and dinitrotoluene
were less effective.
Given the importance of reaching high selectivities
amounts, so that the same molar ratio is main- while working at a high catalytic ratio, the amount of
tained, has a clear positive effect and a selectivity promoter, the temperature and the CO pressure were
in 1 of 80.6% is reached (run 15). This effect is further varied while keeping the molar ratio dinitroto-
probably connected both to the nitroarene dilution luene/Pd¼2920 constant (Table 7). As for the carbony-
effect discussed above and to the fact that palladi- lation of nitrobenzene,[35] an increase in the CO pressure
um catalysts deactivate more slowly when the pal- improves the carbamate yields (runs 1, 2 and 7, 8). The
ladium concentration is low.
last entry of the table shows that the best result at this
catalytic ratio (77.6% selectivity in 1) was reached at
1708C and 100 bar CO.
Effect of Other Acidic Promoters
At this stage, different phosphorus acids were tested Other Attempts to Improve the Catalytic System
(Table 6). The catalytic ratio was further increased in or-
der to better differentiate among the various promoters. Several attempts were performed in the aim of further
A comparison of the data in Table 6 shows that the intro- improving the efficiency of the catalytic system. They
duction of an electron-donating group on the phenyl are briefly described here, although none of them result-
ring of phenylphosphonic acid results in a small increase ed in a real improvement of the results.
in the amount of 1 formed, but the electron-withdrawing
a) Trimethyl orthoformate (TMOF) as the drying
chloride slows down the reaction and larger amounts of agent: Apart from 2,2-dimethoxypropane, trimethyl or-
the monocarbamates 2 are left unreacted. The positive thoformate is also a known internal drying agent, afford-
effect of the methyl group is within the experimental er- ing methyl formate and methanol by reaction with wa-
ror and it does not justify the use of this substituted acid, ter, Eq. (8).
Adv. Synth. Catal. 2005, 347, 105–120
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
111

Michela Gasperini et al.
FULL PAPERS
Table 6. Influence of the identity of the promoter.[a]
Promoter
4b conv. [%][b]
1 sel. [%][c]
2b sel. [%][c]
2a sel. [%][c]
3a sel. [%][c]
Total
carbamates[d]
PhPO(OH)2
71.8
56.8
46.4
28.5
57.9
53.3
44.0
45.8
30.7
30.0
23.6
28.9
14.4
9.8
21.1
20.2
19.9
20.4
21.5
18.9
29.0
26.2
26.7
27.0
5.2
4.9
3.5
6.5
4.1
3.7
85.1
79.4
84.3
82.9
74.3
80.0
4-CH3C6H4PO(OH)2
4-ClC6H4PO(OH)2
H3PO4 85%
H3PO4 100%
Ph2POOH
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼3.2 mg, 5.0ꢁ10ꢀ3 mmol, mol ratios 2,4-dinitrotoluene/4b/promoter/
Phen/Pd¼2920: 400: 660: 48: 1, 2,2-dimethoxypropane¼1.0 mL, PCO ¼60 bar, T¼1508C, in methanol (30 mL), for 5 h.
Conversion of dinitrotoluene was complete in all cases. Selectivity in 3b and 4a ꢂ 1%.
Calculated with respect to the starting amine.
Calculated with respect to the sum of reacted 4b and dinitrotoluene.
Sum of the selectivities in 1–3.
[b]
[c]
[d]
Table 7. Effect of CO pressure, PhPO(OH)2 amount, and temperature.[a]
Run
4b conv.
PhPO(OH)2/Pd
T [ 8C]
PCO
[bar]
1 sel.
2a sel.
2b sel.
3a sel.
Total
[%][b]
[%][c]
[%][c]
[%][c]
[%][c]
carbamates[d]
1
2
3
4
71.8
51.5
88.2
59.2
54.3
54.0
100
660
660
330
495
495
495
330
330
150
150
150
150
170
180
170
170
60
80
80
80
80
80
80
100
44.0
51.8
46.8
59.5
70.0
67.5
75.3
77.6
21.5
14.9
16.0
11.7
6.3
9.5
<1
1.1
14.4
7.5
8.8
6.1
5.8
8.3
<1
–
5.2
5.8
6.3
7.3
5.3
4.4
5.9
2.6
85.1
80.0
77.9
84.6
87.4
89.7
81.2
82.1
5
6
7[e]
8[e]
100
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼3.2 mg, 5.0ꢁ10ꢀ3 mmol, mol ratios 2,4-dinitrotoluene/4b/Phen/Pd¼
2920: 400: 48: 1, DMOP¼1.0 mL, in methanol (30 mL), for 5 h. Conversion of dinitrotoluene was complete in all cases. No
5 was observed. Selectivity in 3b and 4a ꢂ 1%.
[b]
[c]
[d]
[e]
Calculated with respect to the starting amine.
Calculated with respect to the sum of reacted 4b and dinitrotoluene.
Sum of the selectivities in 1–3.
Methanol¼20 mL.
ð8Þ
presence of compounds containing the amino group to
proceed, consumption of these by a side reaction is ex-
The effect of its addition in place of DMOP was investi- pected to slow down the reaction considerably. By com-
gated and the results are reported in Table S1, Support- parison, it should also be recalled that even DMOP can
ing Information. A positive effect of small amounts of react with amines and the reaction of 4-nitroaniline with
¼
TMOF is evident and the amount of formate affording DMOP to afford 4-O2NC6H4N C(CH3)2 has been pub-
the best results is somewhat lower than the best lished.[47] However, Schiff bases derived from aryl-
DMOP amount (0.3 mL against 1.0 mL), which would amines having electron-withdrawing substituents on the
be economically advantageous. However, overall the re- phenyl ring are well known to be hydrolyzed very easily
sults obtainable with TMOF are clearly inferior to the and, accordingly, the amounts of the intermediates con-
ones realizable with the use of DMOP and its use was taining the amino group do not decrease when the
abandoned. It may be noted that at the end of the reac- amount of DMOP is increased (see the data in Table 4).
tions to which TMOF was added, very low amounts of The resistance to hydrolysis of compounds of the gener-
¼
compounds containing the amino group were present. al formula Ar-N C(H)OR has not been studied exten-
It should be considered that trialkyl orthoformates can sively and no comparison data are available with imines
react with amines directly and the reaction of 4b with of related compositions. However, from general consid-
triethyl orthoformate to yield the corresponding erations it is clear that the mesomeric effect of the oxy-
¼
Ar-N C(H)OEt derivative has been reported in the lit- gen atom will make hydrolysis of these compounds more
erature.[46] Since the carbonylation reaction requires the difficult than that of normal Schiff bases. This increased
112
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 105–120

Carbonylation of Dinitrotoluene to Dimethyl Toluenedicarbamate
FULL PAPERS
Table 8. Effect of the modality of addition of 4b.[a]
Run
4b conv. [%][b]
t [h]
1 sel. [%][c]
2a sel. [%][c]
2b sel. [%][c]
3a sel. [%][c]
Total carbamates[d]
1[e]
2[f]
3[g]
56.0
49.7
51.9
2þ3
2þ3
2þ3
46.9
30.4
54.3
17.8
27.1
13.1
9.8
17.9
6.6
6.2
4.9
6.0
80.7
80.3
80.0
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼3.2 mg, 5.0ꢁ10ꢀ3 mmol, mol ratios 2,4-dinitrotoluene/4b/
PhPO(OH)2/Phen/Pd¼2920: 660: 400: 48: 1, DMOP¼1.0 mL, PCO ¼80 bar, T¼1508C, in methanol (30 mL), for 5 h. Con-
version of dinitrotoluene was complete in all cases. No 5 was observed. Selectivity in 3b and 4a ꢂ 1%.
Calculated with respect to the starting amine.
Calculated with respect to the sum of reacted 4b and dinitrotoluene.
Sum of the selectivities in 1–3.
[b]
[c]
[d]
[e]
[f]
Half of the amount of 4b was added at the beginning of the reaction and half after 2 h.
All 4b was added after the reaction had been run for 2 h.
[g]
All 4b was added at the beginning, but the reaction was anyway stopped and CO vented and recharged as for the first two
reactions in the table.
difficulty in hydrolysis can explain both the very low the overall process will slow down. We have thus inves-
amount of compounds containing the amino group at tigated the effect of adding 4b at different moments. To
the end of the reactions where TMOF had been added this purpose, a reaction was run in which half of the
and the inhibiting effect of even relatively small standard amount of 4b was initially added. The reaction
amounts of it.
was run for 2 h, after which it was stopped, the autoclave
b) Mode of addition of 4b: We have previously men- vented and the remaining aliquot of amine added. The
tioned that mechanistic studies in progress in our labora- reaction was then run for additional 3 h (Table 8, run
tories, employing nitrobenzene as substrate, indicate 1). Two comparison reactions were run with the same
that the principal mechanism for the reaction requires procedure. In the first (Table 8, run 2) all of 4b was add-
the intermediate reduction of nitrobenzene to aniline, ed after the first 2 hours, whereasin the second (run 3) all
which is only later carbonylated. Such a mechanism 4b was added at the beginning of the reaction. A com-
has several precedents[1] and is characterized by the parison is also worthwhile with the results of entry 2 in
fact that the amount of aniline consumed equals that Table 7, which is run under the same experimental con-
formed from nitrobenzene, so that any excess of aniline ditions, but with no interruption. Clearly there is no ad-
added to the reaction mixture will remain approximate- vantage in adding the amine stepwise and the best re-
ly constant during the reaction. However, a second stoi- sults are obtained when the amine is added all at once
chiometry for the carbonylation of nitroarenes [Eq. (9)] at the beginning. Stopping the reaction and recharging
is also possible as an alternative to the one described by CO appears to give a small increase in the amount of 1
Eq. (1). Such a process has been found to be operative in formed (from 51.8 to 54.3%), which is in line with the
several cases with palladium catalysts[1,19] and we have better results generally obtained when working at high-
also recently shown that it plays an important role er CO pressures. Thus irreversible consumption of 4b by
even in the presence of some ruthenium catalysts.[48] a reaction of the kind of that reported in Eq. (9) does not
When a process associated with this stoichiometry is op- appear to play a relevant role at least under the actual
erative, the additional aniline is irreversibly consumed, reaction conditions.
Eq. (9).
c) Presence of a second metal: In the literature, it has
been reported that the addition of azo- or azoxybenzene
to the reaction mixture retards the carbonylation reac-
tion of nitrobenzene by palladium-phenanthroline cata-
lysts. The addition of silver, copper or mercury salts was
said to restore the initial activity.[23] In the case of dinitro-
ð9Þ
At least in the case of the carbonylation of nitrobenzene, toluene, numerous azo and azoxy derivatives are possi-
the mechanistic studies in our group show that this stoi- ble and it is not obvious what their individual effects may
chiometry does indeed play a role under certain experi- be. At least under our reaction conditions, the deliberate
mental conditions, but is almost absent in the presence addition of up to 5 mol % azo- or azoxybenzene with re-
of acidic promoters. In the case of dinitrotoluene, how- spect to the starting dinitrotoluene did not result in any
ever, the contribution of this reaction stoichiometry to significant change in rate or selectivity. However, since
the overall process is not known and cannot be excluded other, possibly more noxious, azo derivatives can be
apriori. If such a process is occurring, then the amount of formed from dinitrotoluene, we tested in any case the ef-
4b initially added will decrease during the reaction and fect of the presence of silver triflate or copper acetate,
Adv. Synth. Catal. 2005, 347, 105–120
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
113

Michela Gasperini et al.
FULL PAPERS
Table 9. Experiments run under COþCO2 pressure.[a]
Run
4b conv.
Dinitrotol.
PCO
t [h]
1 sel.
2a sel.
2b sel.
3a sel.
Total
4a sel. [%][d]
2
[%][b]
conv. [%][c]
[bar]
[%][c]
[%][d]
[%][d]
[%][d]
carbamates[e]
1[f]
2
3
100
100
51.0
46.7
100
100
76.7
85.8
10
–
10
–
5
5
1
1
68.8
71.7
2.9
2.5
2.2
37.4
37.3
1.9
4.5
34.0
30.5
1.9
2.3
<1
75.1
80.7
74.3
70.5
<1
<1
2.3
2.1
4
2.7
<1
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼3.2 mg, 5.0ꢁ10ꢀ3 mmol, mol ratios 2,4-dinitrotoluene/PhPO(OH)2/
4b/Phen/Pd¼2920: 330: 400: 32: 1, DMOP¼1.0 mL, PCO ¼80 bar, T¼1708C, in methanol (20 mL). Selectivity in 3b <1%.
No 5 was observed after runs 2–4.
Calculated with respect to the starting 4b.
Calculated with respect to the starting dinitrotoluene.
Calculated with respect to the sum of reacted 4b and dinitrotoluene.
Sum of the selectivities in 1–3.
A small amount (sel. 0.8%) of 5 was formed.
[b]
[c]
[d]
[e]
[f]
two of the best promoters according to the literature. tral reagents. Thus it is surely associated with a strongly
The results (Table S2, Supporting Information), howev- negative DS and is disfavored by high temperatures. The
er, show that under our experimental conditions the reported results suggest that the equilibrium is almost
presence of this second metal always depresses the reac- completely shifted to the left side in the case of the reac-
tion, rather than promoting it. The effect is larger when tions (and the experimental conditions) studied in this
the amount of metal is increased.
paper.
d) Effect of CO2: Carbonic anhydride is formed during
e) Effect of the counteranion: It is well known that the
the reaction and may influence the reactivity. Specifical- counteranion of [Pd(Phen)2]2þ must be non-interacting
ly, it is well known that amines react with CO2 to afford for maximum activity to be observed. However, it
carbamate salts, Eq. (10).
should also be very stable if the catalyst is to remain ac-
tive for prolonged periods of time. During our studies,
ꢀ
we have mostly employed BF4 as a counteranion. We
ð10Þ
had earlier tested PF6ꢀ with nitrobenzene as a substrate
and found it to afford slightly worse results, despite the
It has been shown that the formation of the carbamate fact that Mestroni and co-workers found it to be a little
salts plays an important and beneficial role in the syn- superior to BF4ꢀ with the same catalytic system, but un-
thesis of methyl phenylcarbamate by reaction of aniline der different experimental conditions.[7] In this work, we
and dimethyl carbonate,[49] the effect being due to the decided to try two different counteranions, the use of
higher nucleophilicity of the carbamate ion with respect which in association with the palladium-phenanthroline
ꢀ
to aniline. We have no evidence for dimethyl carbonate system has not been reported before: SbF6ꢀ and BArF
.
4
formation under our experimental conditions, but a nu- The former is known for its remarkable chemical stabil-
cleophilic attack of an arylamine on an intermediate pal- ity, whereas the second is among the least interacting
ladium complex is surely a key step of the catalytic cycle anions known. The synthesis of the two corresponding
and formation of carbamate salts may influence it. We complexes is described earlier in this paper. A compar-
thus investigated the effect of charging 10 bar CO2 ison between the results obtained with these complexes
from the beginning of the reaction. Only a slightly neg- and with the “standard” BF4ꢀ complex is reported in Ta-
ative effect is observable (Table 9). Especially signifi- ble 10. The use of [Pd(Phen)2][SbF6]2 ascatalyst gives re-
cant are the experiments run for a short reaction time, sults which perfectly match those obtainable with the
ꢀ
ꢀ
where any effect of CO2 should be enhanced since little BF4 salt, whereas the one of BArF affords a slightly
4
CO2 has at that time been produced by the reduction re- poorer performance.
action itself.
The almost identity of the results obtained with the
Without questioning the importance of carbamate BF4ꢀ and SbF6ꢀ complexes indicates that no specific de-
ꢀ
salts formation when the reactions in the presence of di- composition of BF4 can be responsible for the catalyst
methyl carbonate are performed, we note that the tem- deactivation that is discussed later.
perature at which our reaction is run is much higher. Al-
The lower amount of 1 formed with the last catalyst is
though we do not have precise thermodynamic data, the unlikely to be due to an intrinsically lower reactivity of
reaction in Eq. (10) consumes one equivalent of a gas- the BArF ꢀ compound and it is possible that a faster de-
4
eous compound and produces ionic products from neu- activation is occurring in this case. Indeed, despite the
114
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 105–120

Carbonylation of Dinitrotoluene to Dimethyl Toluenedicarbamate
FULL PAPERS
Table 10. Effect of the counteranion.[a]
4b conv. [%][b]
Anion
1 sel. [%][c]
2a sel. [%][c]
2b sel. [%][c]
3a sel. [%][c]
3b sel. [%][c]
Total
carbamates[d]
ꢀ
100
100
100
BF4
68.1
68.2
63.5
1.4
3.2
3.3
2.6
3.4
3.7
2.3
2.4
2.2
0.8
0.7
0.7
75.2
77.9
73.4
ꢀ
SbF6
BArF
ꢀ
4
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2 ¼3.2 mg, 5.0ꢁ10ꢀ3 mmol, mol ratios 2,4-dinitrotoluene/4b/
PhPO(OH)2/Phen/Pd¼2920: 330: 400: 32: 1, DMOP¼1.0 mL, PCO ¼100 bar, T¼1708C, in methanol (20 mL), for 3.5 h.
Conversion of dinitrotoluene was complete in all cases. No 5 was observed. Selectivity in 4a<0.5%.
Calculated with respect to the starting amine.
Calculated with respect to the sum of reacted 4b and dinitrotoluene.
Sum of the selectivities in 1–3.
[b]
[c]
[d]
fact that the fluorinated substituents impart a much bet- conditions, it has been shown that a major path for cata-
ꢀ
ter stability towards the breaking of the B C bond with lyst deactivation is due to palladium hydride formation,
respect to the unsubstituted BPh4ꢀ, a decomposition re- which then evolves to inactive compounds.[51] This deac-
action of this derivative has been reported in the litera- tivation can be avoided by the addition of benzoqui-
ture,[50] which occurs under conditions that are milder none, which oxidizes the palladium hydride complex.
than those employed in the present reaction.
In the present reaction, an oxidant is already present,
f) Addition of benzoquinone: The palladium complex the nitroarene, but it is not clear how efficient it will
employed in this paper is also commonly employed to be in scavenging possible palladium hydride complexes.
catalyze another reaction, namely the styrene-CO co- We thus investigated the addition to the reaction mix-
polymerization.[41] This reaction shares with the one in- ture of variable amounts of benzoquinone (Table S3,
vestigated here the presence of an alcohol as solvent Supporting Information). Under the usual operational
and the use of CO, although the olefin and nitroarene conditions, no effect of the presence of benzoquinone
clearly differentiate the two. Under co-polymerization is detectable even when it is added in a large amount
Table 11. Catalyst deactivation.[a]
Run Cat. amount Dinitrotol./ PhPO(OH)2/ 4b
4b/Pd Phen/ T
t [h]
1 sel. 2a sel. 2b sel. 3a sel. Total
[%][c] [%][c] [%][c] [%][c] carba-
mates[d]
[10ꢀ3 ꢁmmol] Pd
Pd
conv.
Pd
[ 8C]
[%][b]
1
2
3
4
5
6
7
8
9
5.0
5.0
5.0
5.0
5.0
2.5
2.5
2.5
5.0
2920
2920
2920
2920
2920
5840
5840
5840
495
495
495
495
495
990
990
990
660
660
660
(495)
495ꢁ2
25.9 400
55.9 400
77.7 400
54.3 400
48
48
48
48
48
96
96
96
48
170
170
170 3.5
170
170
170 3.5
170
170
150
1
2
2.4 36.8
16.5 33.6
64.0 10.2
35.0 2.4
26.8 2.9
7.3 6.8
5.8 5.3
9.4 4.9
29.9 0.9
14.6 3.3
13.5 2.4
7.5 5.8
12.2 0.7
14.8 0.6
3.9 5.9
76.6
79.8
88.3
87.4
87.4
68.5
73.0
69.1
80.0
64.7
64.1
81.5
67.6
5
7
70.0
69.3
6.3
3.8
100
38.1 800
84 800
400
7.2 30.5
33.8 21.3
32.6 20.6
51.8 14.9
39.1 12.7
5
7
5
87.3 800
51.5 400
2920
10[e] 5.0
2920ꢁ2
2920ꢁ3
(2920)
2920
29.4 400ꢁ2 48
150 5þ5
11[f] 5.0
100
100
100
400
(400) (48) 170 3.5þ1.5 68.0
400 48
170 3.5þ1.5 60.9
48
170 1þ1þ7 24.9 23.3
12[g] 5.0þ5.0
13[h] 5.0
3.7
3.3
<1
3.4
[a]
Experimental conditions: catalyst [Pd(Phen)2][BF4]2, DMOP¼1.0 mL, PCO ¼100 bar, in methanol (20 mL). Conversion of
dinitrotoluene was complete in all cases. No 5 was observed. Selectivity in 3b and 4a ꢂ 1%. In the cases in which the cat-
alytic solution was recycled, the results reported refers to the finally obtained amounts.
Calculated with respect to the added 4b.
[b]
[c]
[d]
Calculated with respect to the sum of reacted 4b and dinitrotoluene.
Sum of the selectivities in 1–3. [e] After a standard reaction, more dinitrotoluene and 4b were charged, in amounts equal to
the initial ones, and the reaction was repeated.
[f]
The catalytic solution was recycled two more times, by adding only dinitrotoluene after the first two runs.
After 3.5 h the reaction was stopped and more catalyst (5.0ꢁ10ꢀ3 mmol) added, before running the reaction for a further
1.5 h. The ratios in parentheses refer to the initial amount of catalyst.
[g]
[h]
After 3.5 h the reaction was stopped and more PhPO(OH)2 was added, before running the reaction for a further 1.5 h.
Adv. Synth. Catal. 2005, 347, 105–120
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
115

Michela Gasperini et al.
FULL PAPERS
(runs 1–4). At a short reaction time, a small inhibiting and 10), but dinitrotoluene was anyway completely con-
effect is even present (runs 5 and 6), but, at a higher cat- sumed and, noteworthy, complete dinitrotoluene con-
alytic ratio, dinitrotoluene/Pd¼5840, and for a long re- version was observed even when the same solution
action time a positive effect is observed (runs 7 and 8). was recycled two times, with no other addition apart
However, it was also noticed that when large amounts from dinitrotoluene and, inevitably, CO (run 11).
of benzoquinone are added, the solution becomes vis-
Two more experiments were also performed, in which
cous during the reaction and some polymerization is the reaction was stopped after 3.5 h and a further aliquot
clearly occurring that will make work-up more difficult. of either the catalyst (run 12) or phenylphosphonic acid
Thus addition of benzoquinone is not a viable way of im- (run 13) was added before running the reaction for a fur-
proving the performances of the catalytic system for the ther 1.5 h. A comparison with the data for entry 4 in the
carbonylation of dinitrotoluene.
table, which was run for 5 h continuously without any
In any case, it should be noted that a catalytic ratio di- addition, shows that there is no improvement in effect-
nitrotoluene/Pd¼5840 is by far the highest reported in ing either of the additions. The addition of more acid
the literature for this reaction. The highest ratio previ- even decreased the amount of 1 produced, which con-
ously reported (at least in the free literature) is 500.[24]
firms that the initially added amount is close to the opti-
mal value.
We recall that when nitrobenzene is employed as sub-
strate, no catalyst deactivation was observed up to 105
turnovers. To test if an intrinsically much lower reactiv-
Catalyst Deactivation
To test the stability of the catalytic system, the products ity of 2 with respect to nitrobenzene was responsible for
amounts were investigated as a function of reaction the slowing down of the reaction rate, a reaction was run
time. The results in runs 1–5 of Table 11 show that the under the same experimental conditions of entry 6 in Ta-
catalytic reaction proceeds for the first five hours, but in- ble 11, but for only 1 h, employing a mixture of nitroben-
creasing the reaction time by two more hours has no ef- zene and 2b (2920 times each with respect to palladium).
fect, in spite of the fact that unreacted 2 and 3 are still At the end of the reaction, 61.7% of nitrobenzene and
present (see also Figure 1). It should be noted that the 43.8% of 2b had been converted. Thus nitrocarbamates
low amount of 1 formed in the first hour does not imply are indeed less reactive than nitrobenzene, but the dif-
that an induction time is present. It is simply due to the ference is small and is unlikely to be responsible for
two-steps nature of the reaction itself and the previously any major deactivation pathway.
mentioned fact that carbonylation of 2 only begins when
almost all dinitrotoluene has reacted.
Taken together, the results of these experiments imply
that:
Worthy of note is that the very reactive dinitrotoluene a) Catalyst deactivation is neither due to palladium
was completely consumed in the first reaction hour and
was not present at the end even ofthe first reaction ofthe
series. When the catalyst amount was halved, catalyst
deactivation occurred in approximately the same time
(Table 11, runs 6–8; Figure 1) and about half of the
amount of 1 was formed with respect to that obtained
with the standard amount of catalyst.
black formation (which is indeed not visually ob-
served, although the very small amount of palladi-
um present may render its detection problematic)
nor to the formation of large inactive metal clus-
ters. Indeed, decreasing the palladium concentra-
tion generally affords higher turnover numbers
when its aggregation is the cause for catalyst deac-
tivation. Moreover, addition of fresh catalyst after
3.5 h does not restore the initial catalytic activity.
b) Deactivation is not due to amine consumption, a
result also in line with the results of the reactions
in which 4b was not added all at the beginning of
the reaction (Table 8).
Adding more dinitrotoluene and 4b to the solution af-
ter a reaction results in a decreased amount of 1 (runs 9
c) Even a consumption of the acidic promoter is not
involved in the deactivation. We have previously
reported that esterification of phenylphosphonic
acid only occurs to a limited extent under the reac-
tion conditions.[34, 35]
At this stage, the most likely explanation for the ob-
served deactivation is that some organic compounds
are formed during the reaction which interact with the
palladium catalyst and decrease its reactivity. Azo- and
azoxyarenes have been reported to have an inhibiting
effect on these reactions, but, as already discussed earli-
Figure 1. Selectivity in 1 as a function of reaction time. Cata-
lyst¼5.0ꢁ10ꢀ3 mmol (circles), catalyst¼2.5ꢁ10ꢀ3 mmol
(squares). Data in Table 11.
116
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 105–120

Carbonylation of Dinitrotoluene to Dimethyl Toluenedicarbamate
FULL PAPERS
er in this paper, azo- and azoxybenzene do not deacti-
vate the catalyst under the conditions of this study.
Whether any of the azo- or azoxy compounds which
can be formed from dinitrotoluene can deactivate the
catalyst more efficiently remains an open question. In
any case, the remedy suggested in the literature, addi-
tion of a silver or copper salt, is ineffective under the
present conditions.
Whichever the cause, it must be stressed that the deac-
tivation is only partial. Indeed the data clearly show that
the “deactivated” catalyst is still able to carbonylate the
very reactive dinitrotoluene to yield 2 (complete conver-
sion of the starting material is observed even in run 11,
Scheme 3. Numbering scheme for the products derived from
2,6-dinitrotoluene.
where the ratio dinitrotoluene/Pd¼8760), but it is no describing its carbonylation to the corresponding dicar-
longer active enough to react with 2.
bamate,[52] and only a 30% yield was obtained although
the catalytic ratio was low [PtCl2(PPh3)2 1 mol %, FeCl3
20 mol %]. By carbonylating 2,6-dinitrotoluene under
the same conditions of the last entry in Table 7 (but em-
ploying 2-nitro-6-aminotoluene in place of 4b) we ob-
Separation of 1 from the Reaction Mixture
Easy separation of the reaction product is an important served complete conversion of the dinitroarene and ob-
issue for an industrial process. In the actual technology, tained the desired dicarbamate, 6 with 73.5% selectivity,
the diisocyanate is purified by distillation, which is not only slightly lower than the one obtainable with the 2,4-
feasible in the case of the dicarbamate. Fortunately, 1 isomer, accompanied by 7 (5.1%) and 8 (9.9%). Since 6 is
is poorly soluble in methanol at room temperature. To less soluble in methanol than 1, its separation from the
test the efficiency of a separation based on this property, reaction solution should be even easier in an industrial
a reaction was run under the conditions of the last entry process. Analogously to the case of 2 and 3, an independ-
in Table 7. At the end of the reaction, the glass liner (see ent synthesis of 7 and 8 has not been reported in the lit-
Experimental Section) was cooled down to 08C and, af- erature. We obtained them by the same technique em-
ter 1 h, the precipitate was filtered off. HPLC analysis of ployed for 2 and 3b, respectively (see the Experimental
the solid showed it to be 90% pure 1 (52.7% yield, calcu- Section for details). A pure sample of 6 was best ob-
lated on the pure product), contaminated by 2a and 2b. tained by crystallization of the solid precipitated after
No phenanthroline or phosphorus-containing products a catalytic reaction.
could be detected by NMR even in very concentrated
solutions. A single recrystallization of the solid from
methanol afforded 1 with 98.8% purity. Since the re-
maining catalytic solution is now saturated with 1, it is
Conclusions
expected that further recycling should afford even high-
In conclusion, we have shown that phosphorus acids are
er yields of isolated product, although we did not test
not only excellent promoters for the palladium-phenan-
this aspect in any more detail. The methanol employed
throline-catalyzed carbonylation of nitrobenzene, but
in the recrystallization may easily be recycled as the re-
the strongly positive effect also extends to the difficult
action solvent, taking into consideration that some
but economically important dinitrotoluenes and activi-
methanol is also consumed as a reactant during the proc-
ties and selectivities very close to the limits for commer-
cial applicability have been reached for the first time.
ess. The co-precipitated 2a and 2b wouldalso be recycled
by this procedure. Thus an easy and selective separation
of the product is possible.
Experimental section
Use of 2,6-Dinitrotoluene as Feedstock
General Procedure
Finally, it must be considered that the dinitrotoluene
All solvents were dried by standard procedures, distilled and
employed as feedstock in the industrial process contains
stored under dinitrogen before use. Reactions were generally
ca. 10% of the 2,6-isomer and this also needs to be car-
conducted under a dinitrogen atmosphere, with standard
bonylated efficiently in order for the process to be com-
Schlenk and cannula techniques (for catalytic reactions see
mercially viable. This compound (Scheme 3) is the most
the dedicated paragraph). However, the exchange reactions
difficult to carbonylate, since both nitro groups are in a
sterically hindered position. To the best of our knowl-
between [Pd(Phen)2][NO3]2 and NaSbF6 and the synthesis of
[PhenH][Cl] can be conducted in the air. Phenanthroline mon-
edge, only one report has appeared in the open literature ohydrate was dried by dissolving it in CH2Cl2 and drying the re-
Adv. Synth. Catal. 2005, 347, 105–120
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
117

Michela Gasperini et al.
FULL PAPERS
sulting solution with Na2SO4. The solution was then filtered,
evaporated under vacuum, and the obtained phenanthroline
was stored under a dinitrogen atmosphere. Dry Phen can be
quickly weighed in the air without problems, but long storing
in an undried atmosphere converts it quantitatively to the mon-
minimum amount of methanol (~ 5 mL) and the solution stir-
red for 30 min, after which it was evaporated under vacuum.
Treatment of the residue with diethyl ether left NaCl as a solid
residue, which was filtered off. Evaporation of the ether solu-
tion under vacuum afforded the crude yellowish product,
which was purified by the Soxhlet extraction of the impurities
with hexane for 10 h, to afford the colorless analytically pure
ohydrate. [Pd(Phen)2][BF4]2,[7] [Pd(Phen)2][NO3]2,[7] and
[42,43]
NaBArF
were synthesized as reported in the literature.
4
For the last compound, a slight modification of the two report-
ed procedures, taking the advantages of both, was employed.
Phenanthrolinium chloride monohydrate was prepared by
adding the stoichiometric amount of 36% HCl to a toluene sol-
ution of dry phenanthroline (phenanthroline monohydrate is
poorly soluble in toluene) and was isolated by filtration. Inclu-
sion of one molecule of water per molecule of phenanthrolini-
[PhenH][BArF ]; yield: 600 mg (75%). The compound was
4
shown to be free from chloride impurities by the Agþ test. El-
emental analysis: calcd. for C42H21N2BF24: C 50.60, H 2.03, N
1
2.68; found: C 50.64, H 1.73 N 2.73; H NMR (CD3COCD3):
d¼9.44 (dd, J2,3 ¼4.9, J2,4 ¼1.3 Hz, 2H, H2, H9), 9.09 (dd,
J4,3 ¼8.3, J4,2 ¼1.3 Hz, 2H, H4, H7), 8.40 (s, 2H, H5, H6), 8.80
(dd, J3,2 ¼4.9, J3,4 ¼8.3 Hz, 2H, H3, H8), 7.80 [t(br), J1H,11B
¼
1
um chlorohydrate was evidenced by both H NMR and ele-
2.3 Hz, 8H, Hortho ArF], 7.68 (s, 4H, Hpara ArF).
mental analysis. Use of gaseous HCl in this reaction to avoid
the co-crystallization of water is not advisable, since a mixture
of mono- and dichlorohydrate (with the second prevailing) was
obtained with this reagent. 4-MeC6H4P(O)(OH)2 and
4-ClC6H4P(O)(OH)2 were synthesized by reaction of triethyl
phosphite with 4-MeC6H4Br and 4-ClC6H4Br, respectively, in
the presence of NiBr2, following the method of Tavs,[53] fol-
lowed by hydrolysis of the ester with 37% HCl as reported.[54]
They were characterized by comparison of their analytical
data with those reported in the literature.[55] 2,4-Dinitrotoluene
was recrystallized from methanol to remove the water present
in the commercial product. The anhydrous product was stored
under dinitrogen. Compounds 4a, b and 5 were purified by sub-
limation. The diamine 5 was then stored under dinitrogen,
whereas 4a, b can be stored in the air. All other reagents except
for those mentioned in the following were commercial prod-
ucts and were employed as received. IR spectra were recorded
on a Bio-Rad FTS-7 spectrophotometer; 1H NMR spectra
were recorded on a Bruker DPX 300 (300 MHz) spectrometer;
elemental analyses were performed in the analytical laborato-
ries of the Milan University. HPLC analyses were performed
on a Hewlett Packard series 1050 chromatograph. Gas chroma-
tographic analyses were performed on a Dani 8610 capillary
gas chromatograph (see later for details of the analysis condi-
tions).
Synthesis of [Pd(Phen)2][BArF ]2
Pd(OAc)2 (19.4 mg, 0.086 mmol) and [PhenH][BArF ]
4
4
(180.1 mg, 0.172 mmol) were separately dissolved in acetone
(~ 2.5 mL for each). The palladium acetate solution was fil-
tered through filter paper with the aid of a cannula to remove
very small amounts of insoluble impurities and added to the
phenanthrolinium solution. After stirring at room temperature
for 2 h the still homogeneous solution was evaporated under
vacuum and the resulting sticky solid treated with hexane
(3 mL). After stirring the suspension at room temperature
overnight the solid residue was collected by filtration and
washed with more hexane (2 mL) to give the analytically
pure product; yield: 265 mg (70%). Elemental analysis: calcd.
(%) for C84H40N4B2F48Pd: C 48.19, H 1.84, N 2.55; found: C
1
48.46, H 1.86, N 2.48; H NMR (CD3COCD3): d¼9.34 (dd,
J2,3 ¼5.0, J2,4 ¼1.2 Hz, 4H, H2, H9), 8.99 (dd, J4,3 ¼8.2, J4,2
¼
1.2 Hz, 4H, H4, H7), 8.39 (s, 4H, H5, H6), 8.07 (dd, J3,2 ¼5.0,
J3,4 ¼8.2 Hz, 4H, H3, H8), 7.79 [t(br), 16H, Hortho ArF], 7.67 (s,
8H, Hpara ArF).
Catalytic Reactions
In a typical catalytic reaction, the catalyst, Phen, dinitroto-
luene, and, when required, an amine (4a, 4b, or 5) and the acidic
promoter were weighed in a glass liner. The liner was placed in-
side a Schlenk tube with a wide mouth under dinitrogen and
was frozen at ꢀ788C with dry ice, evacuated and filled with di-
nitrogen, after which the solvent was added. After the solvent
was also frozen, the liner was closed with a screw cap having a
glass wool-filled open mouth which allows gaseous reagents to
exchange and rapidly transferred to a 200-mL stainless steel
autoclave with magnetic stirring. The autoclave was then evac-
uated and filled with dinitrogen three times. CO was then
charged at room temperature at the required pressure and
the autoclave was immersed in an oil bath preheated at the re-
quired temperature; the autoclave took about 15 min to fully
equilibrate at the final temperature. Other experimental con-
ditions are reported in the captions to the tables and figures.
At the end of the reaction the autoclave was cooled with an
ice bath and vented. The products were analyzed by HPLC
(RP-18 column, methanol/water 55:45 as eluent, benzophe-
none as the internal standard; for 2,4- and 2,6-dinitrotoluene,
Synthesis of [Pd(Phen)2][SbF6]2
The procedure is adapted from ref.[7] To a flask were added
[Pd(Phen)2][NO3]2 (300 mg, 0.51 mmol), NaSbF6 (527.7 mg,
2.04 mmol) and methanol (30 mL). The suspension was stirred
at room temperature for 2 h, after which the pale yellow solid
was collected by filtration, washed with methanol and dried un-
der vacuum; yield: 384 mg (79.9%). The solid was pure from re-
sidual nitrate ion, as evidenced by the lack of nitrate absorp-
tions at 825 and 667 cmꢀ1 in its IR spectrum (in nujol)ꢀ, while
the new absorptions at 884 and 653 cmꢀ1 due to SbF6 were
present. Elemental analysis: calcd. for C20H16N4F12PdSb2: C
30.72, H 1.72, N 5.97%; found: C 30.57, H 1.73, N 6.00%.
Synthesis of [PhenH][BArF ]
4
To a Schlenk flash under a dinitrogen atmosphere were added
[PhenH][Cl]·H2O (179.8 mg, 0.766 mmol) and NaBArF
4
(679.6 mg, 0.766 mmol). The reagents were dissolved in the
118
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 105–120

Carbonylation of Dinitrotoluene to Dimethyl Toluenedicarbamate
FULL PAPERS
1, 2a, b, 3a, b, 6, 7, and 8) and gas chromatography (PS 255 col-
umn, naphthalene as the internal standard; for 4a, b and 5).
[19] F. Ragaini, S. Cenini, J. Mol. Catal. A 1996, 109, 1–25,
and references cited therein.
[20] R. Santi, A. M. Romano, P. Panella, G. Mestroni, It. Pat.
Appl. MI96A 002071, 1996.
Supporting Information Available
[21] R. Santi, A. M. Romano, P. Panella, G. Mestroni, It. Pat.
Appl. MI96A 002072, 1996.
[22] R. Santi, A. M. Romano, R. Garrone, P. Panella, It. Pat.
Appl. MI97A 000433, 1997.
Tables S1-S3; experimental details for the synthesis of com-
pounds 1–8 and their characterization data.
[23] R. Santi, A. M. Romano, P. Panella, C. Santini, J. Mol.
Catal. A 1997, 127, 95–9.
Acknowledgements
[24] R. Santi, A. M. Romano, P. Panella, G. Mestroni, A. Ses-
santa o Santi, J. Mol. Catal. A 1999, 144, 41–45.
[25] P. Leconte, F. Metz, A. Mortreux, J. A. Osborn, F. Paul,
F. Petit, A. Pillot, J. Chem. Soc. Chem. Commun. 1990,
1616–1617.
[26] a) F. Paul, J. Fischer, P. Ochsenbein, J. A. Osborn, An-
gew. Chem. Int. Ed. Engl. 1993, 32, 1638–1640; b) F.
Paul, J. Fischer, P. Ochsenbein, J. A. Osborn, C. R. Chi-
mie 2002, 5, 267–287.
[27] A. Sessanta o Santi, B. Milani, G. Mestroni, E. Zangran-
do, L. Randaccio, J. Organomet. Chem. 1997, 545–546,
89–91.
[28] N. Masciocchi, F. Ragaini, S. Cenini, A. Sironi, Organo-
metallics 1998, 17, 1052–1057.
[29] F. Paul, J. Fisher, P. Ochsenbein, J. A. Osborn, Organo-
metallics 1998, 17, 2199–2206.
[30] E. Gallo, F. Ragaini, S. Cenini, F. Demartin, J. Organo-
met. Chem. 1999, 586, 190–195.
[31] A. Sessanta o Santi, B. Milani, E. Zangrando, G. Mestro-
ni, Eur. J. Inorg. Chem. 2000, 2351–2354.
[32] M. Gasperini, F. Ragaini, S. Cenini, E. Gallo, J. Mol. Cat-
al. A 2003, 204–205, 107–114.
[33] F. Ragaini, E. Gallo, S. Cenini, J. Organomet. Chem.
2000, 593–594, 109–118.
[34] F. Ragaini, C. Cognolato, M. Gasperini, S. Cenini, An-
gew. Chem. Int. Ed. 2003, 42, 2886–2889; Angew.
Chem. 2003, 115, 2992–2995.
[35] F. Ragaini, M. Gasperini, S. Cenini, Adv. Synth. Catal.
2004, 346, 63–71.
[36] J. S. Yadav, G. S. Reddy, M. M. Reddy, H. M. Meshram,
Tetrahedron Lett. 1998, 39, 3259–3262.
[37] F. Ragaini, Organometallics 1996, 15, 3572–3578.
[38] a) H. des Abbayes, H. Alper, J. Am. Chem. Soc. 1977, 99,
98–101; b) H. Alper, D. Des Roches, H. des Abbayes,
Angew. Chem. Int. Ed. Engl. 1977, 16, 41–41; c) H. Alp-
er, H.-N. Paik, Nouv. J. Chim. 1978, 2, 245–247; d) H.
Alper, L. C. Damude, Organometallics 1982, 1, 579–
581; e) H. Alper, M. Gopal, J. Chem. Soc. Chem. Com-
mun. 1980, 821–821.
[39] H. Nishida, N. Takada, M. Yoshimura, T. Sonoda, H. Ko-
bayashi, Bull. Chem. Soc. Jpn. 1984, 57, 2600–2604.
[40] H. Iwamoto, T. Sonoda, H. Kobayashi, Tetrahedron Lett.
1983, 24, 4703–4706.
We thank Enichem for financial support and The Dow Chemi-
cal Company for the permission to publish these results. We also
thank Dr. C. Querci (Polimeri Europa, Istituto Donegani) for
helpful discussions.
References and Notes
[1] S. Cenini, F. Ragaini, Catalytic Reductive Carbonylation
of Organic Nitro Compounds, Kluwer Academic Press,
Dordrecht, 1997, and references cited therein.
[2] F. Paul, Coord. Chem. Rev. 2000, 20, 269–323, and refer-
ences cited therein.
[3] A. M. Tafesh, J. Weiguny, Chem. Rev. 1996, 96, 2035–
2052, and references cited therein.
[4] E. Drent, P. W. N. M. Van Leeuwen, US Patent 4,474,978,
1984; Eur. Pat. Appl. EP 86,281, 1983; Chem. Abstr. 1984,
100, 6109x.
[5] E. Drent, Eur. Pat. Appl. EP 231,045, 1987; Chem. Abstr.
1988, 109, 128605n.
[6] J. Stapersma, K. Steernberg, Eur. Pat. Appl. EP 296,686,
1988; Chem. Abstr. 1989, 111, 133802v.
[7] A. Bontempi, E. Alessio, G. Chanos, G. Mestroni, J. Mol.
Catal. 1987, 42, 67–80.
[8] E. Alessio, G. Mestroni, J. Mol. Catal. 1984, 26, 337–340.
[9] E. Alessio, G. Mestroni, J. Organomet. Chem. 1985, 291,
117–127.
[10] E. Drent, Pure Appl. Chem. 1990, 62, 661–9.
[11] S. Cenini, F. Ragaini, M. Pizzotti, F. Porta, G. Mestroni,
E. Alessio, J. Mol. Catal. 1991, 64, 179–190.
[12] P. Leconte, F. Metz, Eur. Pat. Appl. EP 330,591, 1989;
Chem. Abstr. 1990, 112, 78160c.
[13] P. Wehman, G. C. Dol, E. R. Moorman, P. C. J. Kamer,
P. W. N. M. van Leeuwen, J. Fraanje, K. Goubitz, Orga-
nometallics 1994, 13, 4856–4869.
[14] P. Wehman, V. E. Kaasjager, W. G. J. de Lange, F. Hartl,
P. C. J. Kamer, P. W. N. M. van Leeuwen, J. Fraanje, K.
Goubitz, Organometallics 1995, 14, 3751–3761.
[15] P. Wehman, L. Borst, P. C. J. Kamer, P. W. N. M. van
Leeuwen, J. Mol. Catal. A 1996, 112, 23–36.
[16] P. Wehman, L. Borst, P. C. J. Kamer, P. W. N. M. van
Leeuwen, Chem. Ber./Recueil 1997, 130, 13–21.
[17] P. Wehman, H. M. A. van Donge, A. Hagos, P. C. J. Kam-
er, P. W. N. M. van Leeuwen, J. Organomet. Chem. 1997,
535, 183–193.
[41] B. Milani, A. Anzilutti, L. Vicentini, A. Sessanta o Santi,
E. Zangrando, S. Geremia, G. Mestroni, Organometallics
1997, 16, 5064–5075.
[42] M. Brookhart, B. Grant, A. F. Volpe Jr., Organometallics
1992, 11, 3920–3922.
[18] P. Wehman, P. C. J. Kamer, P. W. N. M. van Leeuwen,
Chem. Commun. 1996, 217–218.
Adv. Synth. Catal. 2005, 347, 105–120
asc.wiley-vch.de
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
119

Michela Gasperini et al.
FULL PAPERS
[43] S. R. Bahr, P. Boudjouk, J. Org. Chem. 1992, 57, 5545–
5547.
[50] W. V. Konze, B. L. Scott, G. J. Kubas, Chem. Commun.
1999, 1807–1808.
[44] 4-Chlorobenzoic acid was employed in ref.[18] instead of
benzoic acid, but this is unlikely to make any large differ-
ence because the two promoters gave almost indistin-
guishable results when employed in the corresponding
carbonylation of nitrobenzene.[15]
[51] a) E. Drent, P. H. M. Budzelaar, Chem. Rev. 1996, 96,
63–81, and references cited therein; b) C. Querci, R.
DꢁAlosio, R. Bortolo, M. Ricci, D. Bianchi, J. Mol. Catal.
A 2001, 176, 95–100; c) A. Macchioni, A. Romani, C.
Zuccaccia, G. Guglielmetti, C. Querci, Organometallics
2003, 22, 1526–1533.
[45] F. Ragaini, M. Gasperini, S. Cenini, unpublished results.
[46] J. Bergman, P. Sand, Tetrahedron 1990, 46, 6085–6112.
[52] Y. Watanabe, Y. Tsuji, R. Takeuchi, Bull. Chem. Soc.
Jpn. 1984, 57, 3011–3012.
˜
[47] M. C. Carreno, J. M. Cuerva, M. Ribarorda, A. M. Echa-
varren, Angew. Chem. Int. Ed. 1999, 38, 1449–1452.
[48] a) F. Ragaini, A. Ghitti, S. Cenini, Organometallics 1999,
18, 4925–4933; b) F. Ragaini, S. Cenini, J. Mol. Catal. A
2000, 161, 31–38.
[49] M. Aresta, E. Quaranta, Tetrahedron 1991, 47, 9489–
9502.
[53] P. Tavs, Chem. Ber. 1970, 103, 2428–36.
[54] C. Yuan, H. Feng, Synthesis 1990, 140–141.
[55] M. M. Kabachnik, M. D. Solntseva, V. V. Izmer, Z. S.
Novikova, I. P. Beletskaya, Russ. J. Org. Chem. 1998,
34, 93–97; Zh. Org. Khim. 1998, 34, 106–111.
120
ꢀ 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2005, 347, 105–120