Rhodium-Iodide Catalyzed Carbonylation of Methyl Formate into Acetaldehyde or Methyl Acetate: Mechanistic Aspects

Article abstract of DOI:10.1006/jcat.1997.1567

Under CO pressure, the rhodium/ionic-iodide system catalyzes either the reductive carbonylation of methyl formate into acetaldehyde or its homologation into methyl acetate. The influence of the reaction conditions on the selectivity of these two reactions was investigated and it was found that the former occurs selectively only in N-methylpyrrolidone (NMP) (or related solvents), for low I^-/Rh ratio, low substrate concentration, and high CO pressure, whereas methyl acetate is preferentially formed, in the same solvent, at higher I^- and substrate concentrations and under lower CO pressure. By using labeled methyl formate (H¹³CO₂CH₃) it was also shown that the carbonyl group of acetaldehyde or methyl acetate does not result from that of methyl formate. In situ IR studies conducted under catalytic conditions (high-pressure, high-temperature) have not enabled the identification of any other catalytic species than RhI₂(CO)^-₂ (which is also the active species in methanol or methyl acetate carbonylations), whatever the reaction conditions ([I^-], P_CO . . .). Plausible mechanisms are proposed for these reactions in which the essential role played by NMP in controlling the CH₃I content in the reaction medium is clarified and taking into account these experimental data.

Full text of DOI:10.1006/jcat.1997.1567

JOURNAL OF CATALYSIS 167, 324–336 (1997)
ARTICLE NO. CA971567
Rhodium–Iodide Catalyzed Carbonylation of Methyl Formate into
Acetaldehyde or Methyl Acetate: Mechanistic Aspects
Marc Fontaine,1 Yves Castanet,2 Andre´ Mortreux, and Francis Petit3
Laboratoire de Catalyse He´te´roge`ne et Homoge`ne, Groupe de Chimie Organique Applique´e, URA CNRS 402, ENSC Lille,
USTL, BP 108, 59652 Villeneuve d’Ascq Cedex, France
Received January 19, 1996; revised October 7, 1996; accepted November 26, 1996
iodide systems can catalyze, in N-methylpyrrolidone
(NMP) as solvent, the reductive carbonylation of methyl
formate into acetaldehyde (2) or its homologation into
methyl acetate (3) (Scheme 1). In these two unprecedented
reactions, we have shown that among the parameters which
influence the selectivity, the nature and content of the io-
dide promoter are the most important ones. On the other
hand, it is well known that in the absence of solvent and at
high promoter concentrations, acetic acid is preferentially
obtained (4). For this last reaction two mechanisms have
been proposed: decarbonylation of MF and subsequent car-
bonylation of methanol (5) (analogous to the Monsanto
process) or more recently direct carbonylation of the es-
ter to the mixed formic acetic anhydride with subsequent
decomposition to acetic acid and CO (4).
Under CO pressure, the rhodium/ionic-iodide system cata-
lyzes either the reductive carbonylation of methyl formate into
acetaldehyde or its homologation into methyl acetate. The influence
of the reaction conditions on the selectivity of these two reactions
was investigated and it was found that the former occurs selectively
only in N-methylpyrrolidone (NMP) (or related solvents), for low
Iꢀ/Rh ratio, low substrate concentration, and high CO pressure,
whereas methyl acetate is preferentially formed, in the same sol-
vent, at higher Iꢀ and substrate concentrations and under lower
CO pressure. By using labeled methyl formate (H13CO2CH3) it
was also shown that the carbonyl group of acetaldehyde or methyl
acetate does not result from that of methyl formate. In situ IR
studies conducted under catalytic conditions (high-pressure, high-
temperature) have not enabled the identification of any other cata-
lytic species than RhI2(CO)ꢀ2 (which is also the active species in
methanol or methyl acetate carbonylations), whatever the reaction
conditions ([Iꢀ], PCO . . .). Plausible mechanisms are proposed for
these reactions in which the essential role played by NMP in con-
trolling the CH3I content in the reaction medium is clarified and
c
In the same way, the hydrocarbonylation of methanol
catalyzed by cobalt–iodide systems and giving acetaldehyde
is also a well-known process (6).
taking into account these experimental data.
ꢁ 1997 Academic Press
[Co]ꢀCH3I
CH3OH + CO + H2 ꢀꢀꢀꢀꢀꢀꢀ→ CH3CHO + H2O. [2]
1. INTRODUCTION
The fact that reductive carbonylation of MF into ac-
During the past decade methyl formate (MF) has been etaldehyde presents a stoichiometry totally different from
the subject of great attention as a convenient C1 build- this last reaction and the great differences in selectivity re-
ing block (1). Indeed, it can be readily synthesized from sulting from slight changes in the catalytic system and/or
methanol and carbon monoxide and as the reaction is equi- conditions have led us to look at the mechanistic aspects of
librated, it can be regarded as a convenient liquid source of these reactions. We report here the results of our investiga-
its precursors:
tions.
CH3Oꢀ
HCO2CH3
CH3OH + CO.
[1]
In this context, we have previously reported that rhodium
1
Present address: Laboratory of Macromolecular Chemistry and Or-
ganic Catalysis, CERM, B6, Universite´ de Lie`ge, B-4000 Sart-Tilman,
Belgium.
2
To whom correspondence should be addressed. Fax: (33)0320436585.
Deceased December 30, 1992.
3
SCHEME 1. Methyl formate carbonylation.
324
0021-9517/97 $25.00
c
Copyright ꢁ 1997 by Academic Press
All rights of reproduction in any form reserved.

RHODIUM–IODIDE CATALYZED CARBONYLATION OF METHYL FORMATE
325
2. EXPERIMENTAL
General
reactor was enclosed in an electric furnace whose tempera-
ture was monitored and controlled by a thermocouple and
a PID temperature controller. In a standard experiment
RhCl3 ꢄ 3H2O (0.125 mmol), LiI (1.9 mmol), methyl formate
(80 mmol), and NMP (50 ml) were charged into the reactor
under nitrogen. The autoclave was purged and pressurized
with CO (10 bar) prior to stirring and heating. When the re-
action temperature was reached (180ꢃC), the pressure was
adjusted by addition of CO (80 bar) and the reaction per-
formed under vigorous agitation. The reaction medium was
sampled during the reaction for GC analyses. For kinetic
measurements the time corresponding to the last addition
of CO was considered as the beginning of the reaction. Af-
ter reaction, the autoclave was cooled to room temperature
and degassed. The exact volume of the off-gas was deter-
mined and the gas and liquid phases were analyzed by GC.
1H and 13C NMR spectra were recorded on Bruker
AM400 or AC300 spectrometers. High-pressure, high-
temperature infrared spectra were obtained on a Nicolet
510 spectrometer equipped with a MCT-A detector and in-
tegration, subtraction capabilities (NIC/IR Macintosh sta-
tion). Typically 250 scans were taken per spectrum with a
4 cmꢀ1 resolution. The optical bench attachment used the
design developed by Barnes Analytical-Spectra Tech. The
CIR autoclave (Parr modified reactor, 15 ml) equipped with
a mechanical stirring and with an internal silicon or germa-
nium reflection crystal was previously described by Moser
et al. [7] and was purchased from Spectra Tech. GC anal-
yses were performed on Chrompack 9000 or 9001 chro-
matographs equipped with a FID or a micro TCD (He
as carrier gas) as detector. Different fused silica capillary
columns were used: a 25 m ꢂ 0.53 mm Poraplot Q column
(Chrompack) for gas analyses (T = 25ꢃC, ꢀTCD). Liquids
were analyzed on a 50 m ꢂ 0.53 mm CP-Sil 5CB column
(Chrompack) or a 25 m ꢂ 0.32 mm FFAP-CB (Chrompack)
(T = 60–180ꢃC, ꢀTCD). The above Poraplot Q column or
a 30 ꢂ 0.25 m DB1 column (JW) were used for GC-mass
coupled analyses.
Procedure for Homologation of Methyl Formate
The same procedure as above was used for homologation
of methyl formate after adjustment of the reaction condi-
tions (amount of products, PCO and temperature).
Procedure for Studies with Labeled Methyl Formate
Due to the low reaction volume, the experiments con-
ducted from H13CO2Me were carried out in a 50-ml auto-
clave in order to obtain a better stirring. The autoclave was
equipped with a Teflon liner and a magnetic stirrer and was
heated in an oil bath. On the other hand, the reactions using
deuterated methyl formate were achieved in the standard
100-ml reactor.
Materials
All compounds except deuterated methyl formate were
commercial products and were generally used without pu-
rification except methyl formate and solvents which were
dried over molecular sieves (4A) and distilled under nitro-
gen. It should be noted that commercial methyl formate
contains about 3–4% of methanol and that classical distil-
lation does not improve its purity.
HCO2CD3 was synthesized by esterification of formic
acid (25 g, 0.54 mol) by CD3OD (11.5 g, 0.32 mol). The dis-
tillation of deuterated methyl formate formed during the
reaction shifts the equilibrium and allows one to achieve
the esterification with a high yield. A further distillation
of the crude HCO2CD3 on an efficient column (spinning
band) leads to a purity of about 99% .
DCO2CH3 was obtained by carbonylation of CH3OD
catalyzed by NaOCH3. The reaction was performed in a
100 ml stirred autoclave at 80ꢃC under 80 bar CO pressure
with 25 g (0.75 mol) of CH3OD and 0.5 g of NaOCH3. At
the end of the reaction, 1 ml of water was added to the reac-
tion mixture in order to destroy the catalyst and avoid the
decarbonylation of MF. A distillation as above then gave
deuterated methyl formate with the same purity.
3. RESULTS
3.1. Reaction Stoichiometry and By-products
Acetaldehyde
As previously reported (2), carbon dioxide is coproduced
during acetaldehyde synthesis from methyl formate. The
determination ofitsquantityin the gasphase hasshown that
1 mol of CO2 is formed per mole of acetaldehyde accord-
ing to Eq. [3]. In addition, the gas-phase analysis has also
proved that a small amount of methane (CH4/CO2 = 0.1) is
produced during the reaction.
[Rh]ꢀIꢀ
HCO2CH3 + CO ꢀꢀꢀꢀ→ CH3CHO + CO2
[3]
NMP
In the liquid phase, under optimized experimental condi-
tions for acetaldehyde synthesis, methanol and butanal are
the main by-products (see Table 1). Methanol is produced
by decarbonylation of methyl formate according to Eq. [1].
Butanal is only observed at high conversion. This obser-
vation suggests that it arises from aldolization of acetalde-
hyde followed by dehydration to butenal which is further
Typical Procedure for Reductive Carbonylation
of Methyl Formate
All reactions were carried out in a 100-ml stainless-steel
autoclave equipped with a mechanical stirrer. The batch

326
FONTAINE ET AL.
==
hydrogenated at the C C double bond.
TABLE 2
Homologation of Methyl Formate into Methyl Acetate: Influence
of the Nature of Catalytic Precursors
==
2 CH3CHO → CH3CH CHCHO + H2O
[4]
[5]
[6]
H2O + CO → CO2 + H2
CH3CH CHCHO + H2 → CH3CH2CH2CHO
Selectivitya (mol% )
Catalytic
precursor
t
Conversion
(% ) AcH AcOMe AcOH MeOH
==
No.
(h)
The hydrogen required for this latter reaction (Eq. [6]),
is probably formed via Eq. [5], by a water gas shift reac-
tion (WGSR), which is known to be catalyzed by rhodium
complexes (8).
1
2
3
4
5
6
7
CoI2
CoI2
PdCl2(PPh3)2
4
5
2
16
90
89
91
81
95
62
11
0
6
15
14
3
83
98
48
6
48
90
88
0
0
45
43
0
5
0
0
21
28
0
IrCl(CO)(PPh3)2 1.5
RuCl3 ꢄ 3H2O
1.5
5
5
RhCl3 ꢄ 3H2O
3
0
Methyl Acetate
RhCl(CO)(PPh3)2
7
0
According to the stoichiometry of the reaction 2 mol of
methyl formate are necessary to obtain 1 mol of methyl
acetate (Eq. [7]).
Note. Conditions: Catalytic precursor, 0.125 mmol, except No. 2; CoI2,
8 mmol; LiI, 7.5 mmol; HCO2Me, 250 mmol; NMP, 15 ml; T, 195ꢃC; PCO
,
10 bar.
a Selectivity in the liquid phase.
However, formic acid which is formally expected to be
coproduced (Eq. [7a]), has never been detected by gas
chromatography (using a microcatharometer as detector).
Moreover, when it was introduced into the autoclave during
the reaction course (with a high-pressure metering pump)
its decomposition was very slow. Therefore, it could be
deduced that methyl acetate does not arise from simple
transesterification between acetic acid (which may be the
primary product of the reaction) and unreacted methyl for-
mate. Hence, the overall homologation reaction of methyl
formate into methyl acetate should be expressed by Eq. [7b]
rather than Eq. [7a], consistent with the determination of
CO2 in the gas phase.
As shown in Table 2, the selectivities into methyl acetate
are high (ꢅ90% ) even at high conversion (Nos. 6 and 7),
the main side products under optimized conditions being
acetaldehyde, ethyl formate, ethyl acetate, and acetic acid.
It should be noted that a specific search for dimethyl ether
was also carried out, as suggested by a referee. Chromato-
graphic analysis on different columns showed the presence,
in the gas phase, of a compound that presented the same
retention time as an authentic sample of CH3OCH3 with
a ratio CH3OCH3/CO2 of about 0.05. On the other hand,
under acetaldehyde formation conditions, dimethyl ether
was detected only as a very small trace.
Rh/Iꢀ
2 HCO2CH3 ꢀꢀꢀꢀ→ CH3COOCH3 + HCO2H
[7a]
[7b]
3.2. Influence of Reaction Conditions on Activity
and Selectivity
NMPꢀCO
2 HCO2CH3 → CH3COOCH3 + CO2 + H2
(a) Catalytic Precursor
TABLE 1
We have been studying the reactivity of various cata-
lytic systems classically used in carbonylation reactions
(Tables 1 and 2). Among them, rhodium has given the best
results, although ruthenium and iridium can also catalyze
acetaldehyde synthesis with a high liquid phase selectiv-
ity (see Table 1, Nos. 5 and 6); however, in both of these
cases the amount of methane in the gas phase is higher
(CH4/CO2 = 1/4 to 1/2) than that observed with rhodium.
In addition, they are not selective toward methyl acetate
synthesis (Table 2, Nos. 4 and 5). On the other hand, the ex-
act nature of the rhodium precursor generally has only little
influence on the reaction course (see Table 1, Nos. 7–10 and
Table 2, Nos. 6 and 7).
Reductive Carbonylation of Methyl Formate into Acetaldehyde:
Influence of the Nature of Catalytic Precursors
Selectivitya (mol% )
Catalytic
precursor
t
Conversion
(% ) AcH AcOMe AcOH MeOH
No.
(h)
1
2
3
4
5
6
7
8
9
CoI2
PdCl2(PPh3)2
Fe2(CO)9
1
6
6
14
1
1
1
1
2
1
40
19
15
82
34
46
35
34
46
32
0
20
18
0
70
85
73
83
85
86
2
55
31
20
0
98
0
0
78
0
0
25
51
2
30
15
20b
14b
2b
NiI2
IrCl(CO)(PPh3)2
RuCl3 ꢄ 3H2O
RhCl3 ꢄ 3H2O
[RhCl(CO)2]2
RhCl(CO)(PPh3)2
0
0
0
0
0
0
0
0
(b) Role of the Solvent
10 Rh6(CO)16
0
0
13
Table 3 shows that in the absence of solvent the yields
of acetaldehyde and methyl acetate are very low (Nos. 1
and 11). In both cases MeOH resulting from decarbonyla-
tion of MF (Eq. [1]) or acetic acid are the main by-products
Note. Conditions: Catalytic precursor, 0.125 mmol; LiI, 1.9 mmol;
HCO2Me, 81 mmol; NMP, 50 ml; T, 180ꢃC; PCO, 80 bar.
a Selectivity in the liquid phase.
b Butanal is also formed.

RHODIUM–IODIDE CATALYZED CARBONYLATION OF METHYL FORMATE
327
TABLE 3
Solvent Effect on the Selectivity in Methyl Formate Carbonylation
Selectivitya (mol% )
t
Conversion
(% )
No.
Conditions
Solvent
None
Toluene
Octane
DMF
Pyrrolidone
DMA
ꢁ -Butyrolactone
NMP
DMI
NMP + H2Ob
(h)
AcH
AcOMe
AcOH
MeOH
1
2
3
4
5
6
7
8
9
A
A
A
A
A
A
A
A
A
A
1.5
3
3
1
6
2
6
2
4
9
5
0
0
0
15
0
52
40
85
75
50
20
100
80
5
4
6
56
0
4
0
0
20
0
0
0
0
0
0
0
80
0
0
80
96
42
4
10
20
49
10
49
53
24
57
38
70
50
10
2
0
11
12
13
14
15
16
17
18
B
B
B
B
B
B
B
B
None
5
4
5
5
4
4
4
5
38
50
90
60
48
50
87
95
0
0
0
1
4
2
10
2
75
23
40
85
50
80
28
92
20
77
60
10
23
0
0
0
0
Toluene
Octane
ꢁ -Butyrolactone
DMF
DMA
DMI
3
13
12
60
0
0
3
NMP
Note. Conditions: RhCl3 ꢄ 3H2O, 0.125 mmol. A: solvent, 50 ml; LiI, 1.9 mmol; HCO2Me, 81 mmol; T, 180ꢃC; PCO
,
80 bar. B: solvent, 15 ml; RhCl3 ꢄ 3H2O, 0.125 mmol; LiI, 9.5 mmol; HCO2Me, 250 mmol; T, 195ꢃC; PCO, 10 bar.
a Selectivity in the liquid phase.
b H2O, 80 mmol.
whatever the amount of lithium iodide. To obtain high ac- been reported that ionic iodides (LiI, etc.) are much more
etaldehyde yield and selectivity it is necessary that the re- active promoters than covalent iodides (CH3I, I2, etc.) for
action takes place in N-methylpyrrolidone as solvent, or isomerization of methyl formate into acetic acid (4). Table 4
in similar solvents such as N,N-dimethylimidazole (DMI) shows, in the same way, that in NMP, acetaldehyde, and
and N-ethylpyrrolidone. In other amides such as pyrroli- methyl acetate are selectively obtained only in the pres-
done, dimethylformamide (DMF), and dimethylacetamide ence of ionic iodides. Other ionic halides (bromide or more
(DMA), the main reaction is again decarbonylation of MF, markedly chloride) are practically inactive, whereas with
which is favored by the solvent basicity or more proba- covalent iodides, the activities are not only poor, particu-
bly catalyzed by amines resulting from decomposition of larly at high dilution (Table 4, Nos. 5 and 6) but also the
amides at high temperature. The same behavior is also ob- selectivities drastically decrease (compare Table 4, Nos. 3
served in tertiary amines: triethylamine, N-ethylpiperidine, and 5 or 13 and 16). In that case, the addition of triaryl phos-
etc. (9).
phines or tertiary amines (one equivalent of phosphine or
Finally, in less polar or nonpolar solvents, activity and/or amine per iodide) which are able to generate phosphonium
selectivityare lowand the reaction givesa mixture ofmethyl or ammonium iodide salts according to Eqs. [8] and [9],
acetate and acetic acid (Table 3, Nos. 2, 3, 12, and 13).
NMP is also needed to obtain methyl acetate in high
yields; nevertheless, good results are still observed with
ꢁ -butyrolactone or DMA (Table 3, Nos. 14 and 16) and it
should also be noted that selective synthesis of methyl ac-
etate requires a lower solvent/substrate ratio than acetalde-
hyde (typically 1/1 (vol/vol) and 10/1, respectively).
respectively, restores the initial activities and selectivities.
CH3I + PR3 → CH3PR+3 Iꢀ
CH3I + NR3 → CH3NR+3 Iꢀ
[8]
[9]
The nature ofthe phosphonium or ammonium group aswell
as that of the alkaline cation in the case of ionic promoter
have only little influence on the results (see Table 4, Nos. 3,
4, 13, and 14).
(c) Influence of Halide Promoters
We refer first to the qualitative aspects. Numerous
Turning to quantitative aspects, classically the activity is
rhodium-based catalyzed carbonylations such as that of enhanced by increasing the promoter concentration; how-
methanol into acetic acid (Monsanto process) are carried ever, this parameter has a greater effect on the selectivity.
out in the presence of methyl iodide as promoter (10). It has Figure 1 shows the influence of the Iꢀ/Rh ratio on the

328
FONTAINE ET AL.
TABLE 4
Influence of the Halogen Source as Promoter on the Carbonylation of Methyl Formate
Selectivity (mol% )
t
Conversion
(% )
No.
Conditions
Promoter
(h)
AcH
AcOMe
AcOH
MeOH
1
2
3
4
5
6
7
8
9
A
A
A
A
B
B
B
B
B
B
C
C
LiCl
LiBr
LiI
NaI
CH3I
3
2
2
6
12
22
20
20
26
38
38
34
30
30
64
40
64
86
82
18
15
84
82
78
84
60
81
0
0
1
2
12
0
0
0
6
0
0
0
0
0
0
0
0
0
0
0
0
0
60
36
11
15
64
77
2
10
13
12
26
19
2
15
15
2
3
1
1
6
6
I2
I2 + PPh3 (1/1)
CH3I + PPh3 (1/1)
PhI + PPh3 (1/1)
CH3I + NEt3 (1/1)
HI (aq.)
10
11
12
14
0
HI + PPh3 (1/1)
13
14
15
16
17
18
19
D
D
D
D
D
D
D
LiI
NaI
LiBr
CH3I
5
5
5
5
5
5
5
95
84
26
90
88
94
95
3
3
7
0
0
0
2
88
82
57
20
41
43
70
3
0
0
72
57
44
26
0
3
0
0
0
0
0
I2
CH3I + NEt3 (1/1)
LiI + CH3I (1/1)
Note. Conditions: A: RhCl3 ꢄ 3H2O, 0.125 mmol; promoter, 1.86 mmol; HCO2Me, 81 mmol; NMP, 50 ml; PCO, 80 bar; T,
180ꢃC. B: same conditions as A except Rh(CO)Cl(PPh3)2, 0.125 mmol; promoter, 0.3 mmol (Nos. 5 and 6) or 0.6 mmol. C:
same conditions as A except HCO2Me, 52 mmol; PCO, 100 bar; T, 170ꢃC. D: RhCl3 ꢄ 3H2O, 0.15 mmol; promoter, 7.5 mmol;
HCO2Me, 250 mmol; NMP, 15 ml; PCO, 10 bar; T, 195ꢃC.
acetaldehyde selectivity under other optimized conditions. in order to obtain a high activity and to preserve a high
The selectivity drastically decreases by formation of methyl selectivity.
acetate upon increasing the Iꢀ amount. However, a Iꢀ/Rh
As illustrated in Fig. 2, the same observations can be
ratio of about 10–15 seems to be the best compromise made in the case of the synthesis of methyl acetate, i.e.,
an increase of Iꢀ concentration leads to the formation of
acetic acid. The substitution of ionic iodide by covalent
iodide at the same concentration has the effect of shifting
FIG. 1. Selectivity versus Iꢀ/Rh at high CO pressure and high sol-
vent/substrate ratio. Conditions: RhCl3 ꢄ 3H2O, 0.125 mmol; HCO2Me, erate solvent/substrat ratio. Conditions: RhCl3 ꢄ 3H2O, 0.125; HCO2Me,
80 mmol (5 ml); NMP, 50 ml; PCO, 50 bar; T, 180ꢃC. 250 mmol (15 ml); NMP, 15 ml; PCO, 10 bar; T, 190ꢃC.
FIG. 2. Selectivity vs Iꢀ/Rh ratio at low CO pressure and mod-

RHODIUM–IODIDE CATALYZED CARBONYLATION OF METHYL FORMATE
329
ethyl formate (Table 5, No. 5). These by-products probably
arise via the hydrogenation of acetaldehyde which is the
primary by-product of the reaction followed by subsequent
transesterification:
CH3CHO + H2 → CH3CH2OH
[10]
CH3CH2OH + HCO2CH3 → HCO2Et + CH3OH [11]
(e) Influence of Temperature
A high activity of the Rh/Iꢀ catalytic system is achieved
at 180 and 195ꢃC for acetaldehyde and methyl acetate syn-
theses, respectively. As usual, the activity is enhanced when
the temperature increases; above 210–220ꢃC, however, the
formation of side-products significantly increases, whereas
the efficiency of the system is very low below 150ꢃC.
Under other standard conditions as described above for
FIG. 3. CO pressure effect on initial rate of reductive carbonylation of acetaldehyde or methyl acetate syntheses, liquid samples
methyl formate into acetaldehyde. Conditions: RhCl3 ꢄ 3H2O, 0.125 mmol;
taken at periodictimesand analyzed byGC allowusto draw
the curves depicting the product distribution as a function
of time at various temperature. From these curves, the ini-
tial rate of formation of acetaldehyde or methyl acetate vs
temperature can be determined over the range 170–210ꢃC
and the apparent activation energy (Ea) can be calculated
from a simple Arrhenius plot (Fig. 4). The Ea for acetalde-
hyde (21.4 kcal molꢀ1) is much higher than that for methyl
acetate (11.5 kcal molꢀ1).
LiI, 1.86 mmol; HCO2Me, 160 mmol; NMP, 30 ml; T, 180ꢃC.
the observed curves in Figs. 1 and 2 to the left and therefore
in those conditions acetaldehyde can never be obtained
selectively.
(d) Influence of CO Pressure
According to the stoichiometry of Eq. [3], 1 mol of CO
is consumed per mole of acetaldehyde, so a CO pressure is
necessary to achieve the reaction with a high yield. Figure 3
shows that the activity is directly related with the CO pres-
sure. In contrast, CO is not theoretically involved in the
formation of methyl acetate. However, a CO pressure near
to 10 bar gives the best yields (see Table 5). Below this
value, an improvement in activity was observed, but the
selectivity drastically decreases essentially by formation of
acetic acid. On the other hand, increasing the CO pressure
has not only a considerable inhibition effect on the rate but
the selectivity decreases again by formation of ethanol and
3.3. Mechanistic Considerations
3.3.1. Use of Labeled Substrates
In order to increase knowledge on the mechanism of
these reactions, the origin of the atoms of acetaldehyde,
TABLE 5
Effect of CO Pressure on the Synthesis of Methyl Acetate
Selectivity (mol% )
CO pressure
(bar)
t
Conversion
% AcH AcOMe AcOH Others
No.
(h)
1
2
3
4
5
0
5
10
20
60
1
3
5
5
5
95
98
95
74
70
0
0
3
6
19
16
21
88
83
37
84
79
3
0
0
0
0
6
10a
43b
Note. Conditions: NMP, 15 ml; HCO2Me, 250 mmol; RhCl3 ꢄ 3H2O,
0.125 mmol; LiI, 7.5 mmol; T, 195ꢃC.
a MeOH, HCO2Et, EtOH, and AcOEt, 2, 6, 1, and 1% , respectively.
b MeOH, HCO2Et, EtOH, and AcOEt, 9, 24, 6, and 4% , respectively.
FIG. 4. Arrhenius plot for acetaldehyde and methyl formate synthesis.

330
FONTAINE ET AL.
No. 4). This observation is probably connected with the
TABLE 6
enolization of acetaldehyde and proton exchange with
methanol resulting from decarbonylation of methyl for-
mate according to the following reactions:
Reductive Carbonylation of Labeled Methyl Formate
into Acetaldehyde
Isotopic distributiona (% )
DCO2CH3 → CH3OD + CO
[14]
Acetaldehyde
Substrate
CO2
44
+CH3OD
No.
(isotopic purity % )
43
44
45
46
47
45
,OH
,OD
1
2
3
4
HCO2CH3
35
35
0
64
63
0
2
3
0
0
0
34
25
0
0
62
9
99
26
—
—
1
74
—
—
*
)
==
2
==
CH2 C
CH CDO
CH
C
3
ꢀCH3OH
lD
lD
O
H
13CO2CH3 (92)
HCO2CD3 (99)
DCO2CH3 (99)
0
28
39
,
,
*
)
–
CH D C
[15]
2
Note. Conditions: RhCl3 ꢄ 3H2O, 0.125 mmol; LiI, 1.86 mmol; substrate,
l D
81 mmol; NMP, 50 ml; PCO, 80 bar; T, 180ꢃC except No. 2: Rh6(CO)16
,
0.6 mmol; LiI, 5 mmol; substrate, 8 mmol; NMP, 15 ml.
a The isotopic distribution was determined from the molecular peak of
mass spectra and verified by NMR.
The fact that this process does not occur with CD3CHO (re-
sulting from the carbonylation of HCO2CD3) is consistent
with the large isotope effect generally observed in enoliza-
tion reactions (11).
methyl acetate, and coproducts has been determined by us-
ing isotopically enriched reactants.
Methyl acetate synthesis. Table 7 shows the results ob-
served in the homologation of labeled methyl formate
(H13CO2CH3, HCO2CD3, and DCO2CH3) into methyl ac-
etate.
Acetaldehydesynthesis. When the synthesisofacetalde-
hyde was carried out from methyl formate with a totally
13C-labeled carbonyl group, the acetaldehyde formed con-
tained a 13C/12C ratio corresponding strictly to the ra-
tio observed in the natural product. Moreover, the car-
bon dioxide coproduced exhibited more than 75% of 13C
(Table 6, No. 2).
With regard to the initial amount of H13CO2CH3, about
one-third of 13C is found in methyl acetate, and the main
part is distributed among its coproducts CO and CO2
(Table 7, No. 2).
These results indicate that the carbonyl group of MF is
not retained in acetaldehyde. The finding that CO2 is not to-
tally labeled can be explained by the existence of a WGSR
between CO and water present in the solvent and substrate.
As the amount of labeled MF used in this experiment com-
pared with that in the solvent is much smaller than in a
standard reaction, the part of CO2 resulting from WGS,
and consequentlly unlabeled, is more important.
Deuterated methylformateswere also tested. Thus, when
the reaction wasconducted with HCO2CD3 assubstrate, the
methylgroup ofacetaldehyde waspracticallytotallylabeled
and no transfer of CD3 to NMP was observed (Table 6,
No. 3). From this result, we can deduce that the solvent does
not intervene in a reaction where it could have participated
via ammonium cation exchange reactions (Eq. [13]) after
cleavage of HCO2CD3 by ionic iodide (vide infra).
RhꢀI
2 H13CO2CH3 + CO ꢀꢀꢀ→
NMP
CH313CO2CH3 + 13CO + 13CO2 + H2 [16]
|
{z
}
1/3 initial 13C
2/3 initial 13C
Although it is difficult to conclude with certainty, this
result seems again rather in favor of the nonretention of
the methoxycarbonyl group of MF. The fact that an im-
portant part of the methyl acetate was labeled may be
due to the reincorporation of a part of 13CO released in
TABLE 7
Homologation of Labeled Methyl Formate into Methyl Acetate
Methyl acetate isotopic
distributiona (% )
ꢀ
HCO2CD3 + Iꢀ
CD I + HCO
[12]
Substrate
*
)
3
2
No.
(isotopic purity% )
74
75
76
77
80
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
1
2
3
4
HCO2CH3
96
80
—
95
3
20
—
4
—
—
—
—
—
—
1.5
—
—
—
98
—
Iꢀ
+ CH3I
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
*
)
*
)
ꢀ
CD3I +
H13CO2CH3 (46)
HCO2CD3 (99)
DCO2CH3 (99)
l
l
l
+
,l
,l
,l
l
l
l
O
O
O
N
N
N
|
|
,
l
CH3
CH3
CD3
CD3
Note. Conditions: RhCl3 ꢄ 3H2O, 0.125 mmol; LiI, 7.5 mmol; substrate,
250 mmol; NMP, 15 ml; PCO, 10 bar; T, 190ꢃC except No. 2: substrate,
17 mmol; NMP, 5 ml.
[13]
a
On the other hand, acetaldehyde obtained from
DCO2CH3 showed a wider isotopic distribution (Table 6,
The isotopic distribution was determined from the molecular peak
of mass spectra and verified by NMR.

RHODIUM–IODIDE CATALYZED CARBONYLATION OF METHYL FORMATE
331
FIG. 5. High-pressure FTIR spectra (ꢂCO region) presented as transmission vs wavenumber (cmꢀ1) for Rh catalyst under different reaction
conditions. Spectra A: RhCl3 ꢄ 3H2O, 0.25 mmol; LiI, 2 mmol; HCO2Me, 25 mmol; NMP, 10 ml; PCO, 50 bar. Spectra B: Same conditions as spectra A
except LiI, 15 mmol; HCO2Me, 100 mmol; NMP, 6 ml.
the liquid phase during the last step of the reaction (see was conducted utilizing a high-pressure CIR reactor (see
Scheme 4).4
Experimental).
Finally, methyl acetate resulting from HCO2CD3 was
Both catalytic reactions were followed in situ by FTIR
completely deuterated, whereas when DCO2CH3 was used and the results are presented in Fig. 5. Spectra A show the
as substrate, methyl acetate exhibited no label at all.
evolution of the system with temperature leading to the
formation of acetaldehyde. For this system, a weak amount
of lithium iodide is required. At 50ꢃC the infrared spec-
trum showed two main absorption bands in the carbonyl
region at 2052 and 1979 cmꢀ1. By comparison with liter-
ature data and an authentic sample prepared separately,
these absorptions can be assigned to the square planar RhI
anion [RhI2(CO)2]ꢀ 1. Upon heating to 150ꢃC, no other
absorption band appears or disappears.
Now, if we compare the series of infrared spectra A with
those of spectra B (Fig. 5) for which the amount of lithium
salt is present in large excess and thus favours the synthe-
sis of methyl acetate, we can observe at 50ꢃC, besides the
peaks in spectra A, a supplementary absorption band at
2085 cmꢀ1. When the temperature increases, this last peak
disappears (around 80ꢃC). Furthermore, coming back to
the initial temperature restores this absorption.
We have tried to attribute this absorption band at
2085 cmꢀ1. For that we ran some experiments for which the
rhodium precursor, solvent or substrate were changed. Thus
the reactions which were realized with RhI3 alone have the
same behavior as those using RhCl3 ꢄ 3H2O. Indeed in the
absence of lithium iodide (that is to say when the unique
source of iodide is the complex itself), only the typical sig-
nal due to [RhI2(CO2)]ꢀ is observed. When the lithium salt
is added in excess to the starting solution, we observed the
peak at 2085 cmꢀ1 and the influence of the temperature is
similar to the RhCl3 based system. Thus we can eliminate
3.3.2. High-Pressure in Situ Infrared Spectroscopic Study
It is interesting to compare our systems with that de-
scribed for the well-known methanol carbonylation re-
action for which a great deal of work has been done
from a mechanistic point of view. The [RhI2(CO)2]ꢀ ion
1 was early identified as the active species (12). Later,
acyl and alkyl rhodium complexes, [Rh(CO)(COMe)I3]ꢀ
and [MeRh(CO)2I3]ꢀ, were spectroscopically detected and
characterized as reactive intermediates but not under cata-
lytic conditions (13). Recently rhodium (III) hydrides,
[HRh(CO)I3]ꢀ and [HRh(CO)2I2], were detected by H
1
and 13C NMR spectroscopy and also proposed as in-
termediates (14). Using rhodium/iodide-based catalytic
precursor, it has to be pointed out that the presence of
N-methylpyrrolidone as solvent and the Rh/Iꢀ ratio are cru-
cial to obtain selective carbonylations of methyl formate. In
order to gain insight into the nature of the species involved
in the catalytic cycles, an in situ HP-FTIR spectral study
12
13
4
Calculations of CO consumed and CO released during the reac-
tion according to the mechanism outlined in Scheme 4 and considering
the gas as perfect, indicates that with a 50-ml reactor used in this experi-
ment and for 10 bar initial CO pressure, about 50 to 60% of CO in the gas
phase should be labeled at the end of the reaction for 80% methyl formate
conversion.

332
FONTAINE ET AL.
the formation of a mixed halogenorhodium complex such
as [RhCl(I)(CO)2]ꢀ. In the same way the use of hexade-
cacarbonylhexarhodium, Rh6(CO)16, with or without LiI
in excess, led in both cases only to [RhI2(CO)2]ꢀ. Only its
formation rate is different (slowest under these conditions).
It was also noticed that in the absence of NMP (i.e. in pure
methyl formate or in a nonpolar solvent such as toluene)
the peak at 2085 cmꢀ1 is the only one observed in the ꢂCO
region and remains present until 120ꢃC. Moreover, when
methyl formate was replaced in the absence of solvent by
another ester such asmethylor ethylacetate under the same
conditions, the IR spectrum exhibits the same absorption
band at 2085 cmꢀ1, which seems to indicate that this signal is
not characteristic of a catalytic intermediate. This assertion
was confirmed by an experiment conducted in pure toluene
(without methyl formate) in which the peak at 2085 cmꢀ1
was also observed. However, under these conditions, due
to the very weak solubility of LiI in toluene, its appearance
was very slow (3 h vs 10 min in the other cases) and occurred
only at a temperature up to 80ꢃC.
From these results, as the signal at 2085 cmꢀ1 is not re-
lated to an intermediate in the catalytic cycle and as it is
only observed at high iodide content, we can assume that
it corresponds to a Rh(III) species, probably to the trans-
[RhI4(CO)2]ꢀ anion which is known to absorb in CH2Cl2
solution in the ꢂCO region as a single stretch at 2091 cmꢀ1
(15, 16). The monocarbonyl anion [RhI4(CO)]ꢀ cannot be
totally excluded as this compound exhibits in the carbonyl
vibration region a single band at 2076–2078 cmꢀ1 in solu-
tion (8a, 15) and as its synthesis was previously reported,
although with moderate yield, by Vallarino (17) working
under nearly identical conditions (Rh(III)/Iꢀ/Pco). How-
ever, as these two Rh(III) species are easily interconvert-
ible under CO atmosphere (15), under our conditions the
presence of [RhI4(CO)2]ꢀ seems more probable.
SCHEME 2. Catalytic cycle for methanol or methyl acetate carbonyla-
tion.
bonyl rhodium species is the complex [RhI2(CO)2]ꢀ, what-
ever the iodide content or the nature of solvent.
As previously mentioned, this complex 1 is also the ac-
tive species in the carbonylation of methanol into acetic
acid (12) (Monsanto process) or of methyl acetate into
acetic anhydride (19) (Eastman Kodak process), catalyzed
by rhodium/iodide systems. In both cases, the currently ac-
cepted mechanisms involve oxidative addition of CH3I to
1, followed by CO insertion into the alkyl–rhodium bond.
The subsequent reductive elimination gives acetyl iodide.
The main differences between these two mechanisms con-
sist in the formation of CH3I (which is the real substrate
of these carbonylations) and in the last step of the cata-
lytic cycle, namely hydrolysis of CH3COI into acetic acid
(methanol carbonylation) or its reaction with AcOꢀ giving
acetic anhydride (Eastman Kodak process) (see Scheme 2).
In the same way, the changes in selectivity observed in
the reactions leading to the acetaldehyde or methyl ac-
etate synthesis, could rather be due to a change in reaction
conditions than as a result of an evolution of the catalytic
species.5
Finally, as [RhI2(CO)2]ꢀ was the sole species observed
under catalytic conditions at low or high Iꢀ contents, in or-
der to obtain more spectroscopic information on the mech-
anism and on the basis of the known reaction cleavage of
esters by ionic iodides (see Eq. [15]), different experiments
were carried out in which methyl formate was replaced by
sodium formate. Thus the peaks at 2052 and 1979 cmꢀ1
initially formed when RhCl3 and LiI were stirred in NMP
under CO pressure, quickly disappeared after addition of
NaHCO2 and a strong absorption band at 1898 cmꢀ1 was
the sole one observed in the ꢂCO region when the reac-
tion medium was heated. This signal can be assigned to the
presence of [Rh(CO)4]ꢀ (18).
In this context, the concentration value of methyl iodide
in the reaction medium is probably a key parameter of the
selectivity. Indeed, the cleavage of esters by ionic iodides
4. DISCUSSION
5
Preliminary Remarks
Although infrared spectra of reaction medium exhibit no other bands
in the CO absorption region than those corresponding to the complex 1,
The in situ IR studies have shown that in our reaction con-
the existence of a very active catalytic species, in too low concentration to
ditions (T > 150ꢃC, PCO > 5 bar) the only observable car- be detected, cannot be totally excluded.

RHODIUM–IODIDE CATALYZED CARBONYLATION OF METHYL FORMATE
333
from 1 (see Eqs. [18] and [19]). This species could result
from the substitution of an iodide of 1 by the formate anion
giving complex 2 which at high temperature decomposes
into 3 (Scheme 3).
TABLE 8
Solvent Effect on the Cleavage of Methyl Formate by LiI
Concentration (mol% )
a
No.
Solvent
HCO2Me
CH3I
K ꢂ 103
ꢀ
ꢀ
RhI2(CO)ꢀ + HCOꢀ
(HCO )RhI(CO) + I
[18]
*
)
2
2
2
2
1
2
3
NMP
nC8H18
HCO2Me
96.2
92.6
96.1
0.32
2.9
1
4.3
356
42
1
2
,Ol
ꢀ
RhI(CO)ꢀ2
–
(HCO2)RhI(CO)2 ꢆ H C
Note. Conditions: HCO2Me, 250 mmol (15 ml); LiI, 7.5 mmol; solvent,
15 ml; T, 195ꢃC; t, 1 h; P, 35 bar N2.
lO,
a K = [CH3I] ꢄ [HCO2ꢀ]/[HCO2CH3] ꢄ [Iꢀ].
ꢀ
–
ꢀꢀ→ H RhI(CO)2 + CO2 [19]
3
is a well-known reaction (20) which leads in this case to
methyl iodide and formate anion by reaction of an ionic
iodide on methyl formate.
This assumption is in accordance with the fact that alkali
formates have been reported as efficient precursors of tran-
sition metal hydrides (22) and moreover with the fact that
the decarboxylation of formate complexes has also been re-
ported several times (23). From complex 3, the steps lead-
ing to acetaldehyde are classical (Scheme 3). Thus oxida-
tive addition of CH3I to 3 followed by CO insertion to the
metal–alkyl bond gives the acylrhodium 5. Reductive elim-
ination produces acetaldehyde and regenerates the initial
active species 1.
*
)
HCO CH + MI
CH I + HCO M,
[17]
2
3
3
2
where M is alkali metal. This reaction is an equilibrium and
as shown in Table 8 the equilibrium constant K is greatly
solvent-dependent. Although the values listed above are
essentially indicative, it clearly appears that for a given ini-
tial amount of ionic iodide, the cocentration of CH3I at
equilibrium is much smaller in NMP than in pure methyl
formate or in a nonpolar solvent. Hence the role played
by NMP on the selectivity could be to regulate the CH3I
content in the reaction medium and to allow the presence
of a sufficient quantity of ionic iodide to easily generate
anion 1.
In the absence of CH3I and under high CO pressure
(FTIR experiments carried out with NaHCO2 instead
methyl formate), reductive elimination of HI from 3 can
occur leading to [Rh(CO)4]ꢀ according to the following
On the basis of these mechanistic considerations and
from the results obtained with labeled methyl formate,
plausible mechanisms can be drawn up for the synthesis
of these compounds, as presented below.
Mechanism for Acetaldehyde Production
Although anion 1 was found as the main rhodium species,
the hypothesisofa direct reaction ofCH3I with thiscomplex
as an effective way of synthesis of acetaldehyde can be ruled
out if we consider that this reaction should lead to acetyl
iodide as in methanol or methyl acetate carbonylation (see
Scheme 2). From the fact that the presence ofwater doesnot
affect the synthesis of acetaldehyde and does not give any
acetic acid (see Table 3, No. 10), we can exclude CH3COI
as intermediate.
Moreover, CH3COI should also react with formate ions
to give the mixed formic–acetic anhydride which is rela-
tively stable in the cold but very unstable at high tem-
perature (21). Actually, when this anhydride is intro-
duced into the autoclave during the reaction course with
a high-pressure pump, it decomposes quickly into acetic
acid.
Therefore, the key point of acetaldehyde synthesis is
SCHEME 3. Proposed catalytic cycle for reductive carbonylation of
probably the formation of a hydride rhodium complex, 3 methyl formate into acetaldehyde.

334
FONTAINE ET AL.
reaction:
TABLE 9
Influence of [Iꢀ] on the Decarboxylation of Formate Salts
[RhI2(CO)2]ꢀ + HCO2ꢀ
CO
Gas phase
→→ [HRhI(CO)2]ꢀ ꢀꢀ→ Rh(CO)ꢀ4 + HI [20]
composition (% )
LiI
(mmol)
3
No.
Iꢀ/Rh
CO
CO2
In the same way, methane identified in the gas phase may
result from reductive elimination of methyl and hydride lig-
andsfrom the complexintermediate 4before CO insertion.6
Itsimportant formation observed with Ir-based catalyst (see
Section 3.2) can be attributed to a more difficult insertion
of CO into the metal–alkyl bond with Ir than with Rh (25).
The mechanism depicted above can be related to that
given for the synthesis of aromatic aldehydes by carbony-
lation of aryl halides, under CO pressure, in the presence
of palladium and stoichiometric amount of alkali formate.
(26)
1
2
12
0.45
80
3
99.4
97.6
0.6
2.4
Note. Conditions: HCO2Na, 0.4 mmol; RhCl3 ꢄ 3H2O, 0.15 mmol;
NMP, 4 ml; PCO, 10 bar; T, 180ꢃC; t, 2h.
The last part of this catalytic cycle (i.e., the steps follow-
ing the addition of CH3I to complex 7) is supported by the
work of Bernard and Atwood (27) in which the reaction of
acyl chloride with alkoxyiridium complex giving selectively
the corresponding ester is described. On the other hand, the
first step (i.e., the oxidative addition of methyl formate to
–
1 by cleavage of the C O bond) is much less documented,
[Pd]
but a similar reaction has been invoked with nickel complex
ArX + NaHCO2 ꢀꢀꢀ→ ArCHO + NaX + CO2.
[21]
CO
–
(28). Moreover, activation of the C O bond of several es-
ters by various transition metal complexes including Rh has
been reported (29). In addition, the formation of dimethyl
ether, identified in the gas phase, gives further evidence in
favor of this mechanism as complex 8 can undergo reductive
elimination before CO insertion, giving CH3OCH3.
This is also in agreement with the isotopic distribution
observed in experiments conducted with labeled substrates.
Mechanism for Methyl Acetate Formation
Considering that the ligand exchange reaction (Eq. [16])
involved in the first step ofthiscatalyticcycle isequilibrated,
an increase of the Iꢀ concentration must result in a shift of
the equilibrium to the left and to reduce the formation of
complex 2. Thus, although we are not able to detect com-
plexes 2 or 3 by infrared studies, the influence of Iꢀ on the
stoichiometric decarboxylation of formate salts by rhodium
complex, in the absence of methyl formate is proved by the
results given in Table 9.
Consequently at higher Iꢀ contents than those needed
for acetaldehyde synthesis complex 3 does not form and a
new mechanism must be taken into account to explain the
formation of methyl acetate.
As the direct carbonylation of CH3I into CH3COI must
still be dismissed (for the same reasons as those aforemen-
tioned), methyl acetate synthesis could proceed as previ-
ously reported (3) according to the mechanism outlined in
Scheme 4, by oxidative addition of methyl formate to 1,
reductive elimination of HCOI followed by oxidative addi-
tion of CH3I to complex 7, CO insertion and reductive elim-
ination of AcOMe. In a parallel reaction, HCOI reacts with
a formate anion to give the unstable formicanhydride which
at the reaction temperature immediately decomposes into
CO, CO2, and H2, without formation of formic acid.
6
Methane might also be formed by direct decarboxylation of methyl
formate according to HCO CH → CH + CO as reported by Pruett and
2
3
4
2
Kacmarick (24). However, when a blank experiment was performed in the
absence of catalyst the methane content in the gas phase was much lower.
SCHEME 4. Proposed catalytic cycle for homologation of methyl for-
mate into methyl acetate.

RHODIUM–IODIDE CATALYZED CARBONYLATION OF METHYL FORMATE
335
SCHEME 5. Reductive carbonylation of methyl formate into acetaldehyde. Alternative reaction mechanism via oxidative addition of methyl formate
to Rh.
At very high iodide contents acetic acid is probably complex 10 requires its decarboxylation (see Scheme 5).
formed, as recently reported (4), by carbonylation of CH3I The fact that the latter is unlikely with Rh complexes (28)
via its oxidative addition to RhI2(CO)ꢀ2 (see Scheme 2) and has induced us to discard this possibility. In the case of
subsequent reaction of CH3COI with HCOꢀ2 .
methyl acetate synthesis, such a mechanism could be plau-
sible (see Scheme 6). It would nevertheless lead to the for-
mation of HCO2H which has been found stable under our
reaction conditions; moreover, when the reaction was con-
ducted with H13CO2CH3 the methyl acetate obtained ac-
cording to Scheme 6 should be totally labeled. The fact that
HCO2H has never been observed in catalytic reactions giv-
ing AcOMe (see Section 3.1) and that the methyl acetate
resulting from H13CO2CH3 was largely nonlabeled, lead us
to the conclusion that the mechanism outlined in Scheme 4
is more credible.
Rhl2(CO)ꢀ
+HCOꢀ2
2
CH3I ꢀꢀꢀꢀꢀ→ CH3COI ꢀꢀꢀ→ HCO2COCH3
CO
1
HCO2COCH3 ꢀꢀ→ CH3COOH + CO.
As CH3I adds to 1 probably much more easily than
HCO2Me, the methyl acetate vs acetic acid selectivity is
certainly connected with the relative amounts of CH3I and
HCO2Me in the medium. From this point of view, the use
of NMP as the solvent is crucial as it allows one to limit
the CH3I concentration when an ionic iodide is used as
promoter. On the other hand, with covalent iodides, NMP
cannot play this role and acetic acid is always obtained as
the major product.
5. CONCLUSION
These results illustrate the great versatility of methyl for-
mate. Thus, starting from this compound, with the same
Rh-based catalyst precursor and the same ionic iodide pro-
moter, LiI, it is possible to have access selectively to very
different products such as acetaldehyde, methyl acetate and
–
A mechanism involving the activation of the H C bond
of methyl formate giving complex 10, could also have been
considered (30). However, acetaldehyde synthesis from
SCHEME 6. Alternative mechanism for homologation of methyl formate into methyl acetate via its oxidative addition to Rh.

336
FONTAINE ET AL.
acetic acid by a judicious control of the reaction conditions.
Mechanistic studies seem to indicate that a single species,
RhI2(CO)ꢀ2 , 1 is involved in all these reactions. According
to our experimental data diffꢀerent catalytic cycles in which
9. Ramirez-Vega, F., Clement, J. C., and Des Abbayes, H., Tetrahedron
Lett. 34, 8117 (1993).
10. Forster, D., Adv. Organomet. Chem. 17, 255 (1979).
11. March, J., “Advanced Organic Chemistry,” 3rd ed., p. 528. Wiley, New
York, 1985; Francalanci, F., Gardano, A., Abis, L., Fiorani, T., and Foa,
M., J. Organometal. Chem. 243, 87 (1983).
–
the anion 1 reacts with HCO2 , HCO OMe, or CH3I in the
initial steps of the catalytic reaction are proposed to explain 12. Forster, D., J. Am. Chem. Soc. 98, 846 (1976); Forster, D., and Dekleva,
T. W., Adv. Catal. 34, 81 (1986).
13. Haynes, A., Mann, B., Gulhiver, D., Morris, G., and Maitlis, P., J. Am.
Chem. Soc. 113, 8567 (1991).
the formation of the final products.
ACKNOWLEDGMENTS
14. Roe, D. C., Sheridan, R. E., and Bunel, E. E., J. Am. Chem. Soc. 116,
1163 (1994).
15. Forster, D., Inorg. Chem. 8, 2556 (1969).
16. Daly, J. J., Sanz, F., and Forster, D., J. Am. Chem. Soc. 97, 2551 (1975).
We thank the CNRS for financial support (GDR Formiate de me´thyle)
and for a Postdoctoral Fellowship to M.F. as Associated Researcher.
17. Vallarino, L., Inorg. Chem. 4, 161 (1965).
REFERENCES
18. Garlaschelly, L., Della Pergola, R., and Martinengo, S., Inorg. Synth.
28, 211 (1990).
1. Ro¨per, M., Erd. Kohle-Erd. Petrochem. 37, 506 (1984) and references 19. Polichnowski, S. W., J. Chem. Ed. 63, 206 (1986) and references therein.
therein; Lee, J. S., Kim, J. C., and Kim, Y. G., Appl. Catal. 57, 1 (1990)
and references therein.
20. March, J., “Advanced Organic Chemistry,” 3rd ed., p. 386. Wiley, New
York, 1985.
2. Vanhoye, D., Melloul, S., Castanet, Y., Mortreux, A., and Petit, F., 21. Muramatsu, I., Murakami, M., Yoneda, T., and Hagitani, A., J. Org.
Angew. Chem., Int. Ed. Engl. 27, 683 (1988). Chem. 38, 244 (1965).
3. Melloul, S., Seuillet, B., Castanet, Y., Mortreux, A., and Petit, F., 22. Brothers, P. J., and Roper, W. R., J. Organomet. Chem. 258, 73
J. Mol. Catal. 60, L15 (1990); Seuillet, B., Castanet, Y., Mortreux, A.,
and Petit, F., Appl. Catal. A: Gen. 93, 219 (1993).
4. Schreck, D. J., Busby, D. C., and Wegman, R. W., J. Mol. Catal. 47, 117
(1988); Bub, G., and Ho¨g, H. U., J. Mol. Catal., A: Chem. 95, 45 (1995).
(1983).
23. Roper, W. R., and Wright, L. J., J. Organomet. Chem. 234, C5
(1982).
24. Pruett, R. L., and Kacmarcik, R. T., Organometallics 1, 1693 (1982).
5. Briant, F. J., Johnson, W. R., and Singleton, T. C., “General Papers: 25. Ellis, P. R., Pearson, J. M., Haynes, A., Adams, H., Baily, N. A., and
Petrochemicals,” American Chemical Society, Meeting, Dallas, TX,
April 1973.
6. Gauthier-Lafaye, J., and Perron, R., “Methanol et Carbonylation,”
Maitlis, P., Organometallics 13, 3215 (1994).
26. Pri-Bar, J., and Buchman, O., J. Org. Chem. 53, 624 (1988); Okano, T.,
Harada, N., and Kiji, J., Bull. Chem. Soc. Jpn. 67, 2329 (1994).
p. 39. Rhoˆne-Poulenc Recherche, Ed Technip, Paris 1986, For a use- 27. Bernard, K. A., and Atwood, J. D., Organometallics 6, 1133 (1987).
ful summary see Gauthier-Lafaye, J., Perron, R., Leconte, P., and
28. Yamamoto, T., Ishizu, J., and Yamamoto, A., Bull. Chem. Soc. Jpn. 55,
Colleuille, Y., Bull. Soc. Chim. Fr. 353 (1985).
623 (1982).
7. Moser, W. R., Cnossen, J. E., Wang, A. W., and Krouse, S. A., J. Catal. 29. Yamamoto, T., Miyashita, S., Naito, Y., Komiya, S., Ito, T., and
95, 21 (1985). Yamamoto, A., Organometallics 1, 808 (1982).
8. Baker, E. C., Hendriksen, D. E., and Eisenberg, R., J. Am. Chem. Soc. 30. Milstein, D., J. Am. Chem. Soc. 108, 3525 (1986); Bianchini, C.,
102, 1020 (1980); Yoshida, T., Okano, T., Ueda, Y., and Otsuka, S.,
J. Am. Chem. Soc. 103, 3411 (1981); Pardey, A., and Ford, P., J. Mol.
Catal. 53, 247 (1989).
Peruzzini, M., Vizza, F., and Zanobini, F., J. Organomet. Chem. 346,
C9 (1989); Legrand, C., Castanet, Y., Mortreux, A., and Petit, F., J.
Chem. Soc., Chem. Commun. 1173 (1994).