hydride mechanism has been shown to operate in copolymeris-
ation of CO and ethene,1 as well as in the methoxycarbonyl-
ation of ethene catalysed by Pd/PPh3 complexes7 or in the
synthesis of 3-pentanone catalysed by rhodium complexes of
triethylphosphine.8
what had been calculated as above. An example of the
comparison between the measured and simulated spectra is
shown in Fig. 3.
Given the extremely high activity and unexpected selectivity
for the production of methyl propanoate (MeP) when com-
plexes derived from palladium precursors and DBPMB are
activated with acids such as methanesulfonic acid, we were
interested in which mechanism might be operating in this case.
We, therefore, carried out the reaction in CH3OD to investigate
whether the labelling pattern of the product would give any
information. Labelling studies have been used successfully
to derive mechanisms for the formation of oligoketones and
small amounts of methyl propanoate form CO and ethene,4 as
well as for the formation of 3-pentanone from hydrocarbonyl-
ation of ethene9 or from reaction of ethene with CO and
methanol, where the methanol is the source of the extra H
atoms.6,8 A multinuclear NMR study, in which all the possible
intermediates were identified, strongly suggests that a hydride
mechanism operates for the methoxycarbonylation catalysed by
Pd/DBPMB complexes.10
Fig. 3 Experimental (black) and simulated (grey) mass spectra of the
mixture of isotopomers of dissolved ethene after 15 min in a reaction
carried out in reactor C, stirring at 100 rpm. The values used for the
simulation are in the Experimental section. (Note the analysis was
carried out several days after the reaction and the D incorporation
largely arises from post reaction exchange).
Experimental
GCMS data were collected on a Hewlett-Packard HP 6890 gas
chromatograph with an HP 5973 mass selective detector.
13C{1H} and 13C{1H, 2H} spectra were recorded on a Varian 500
MHz spectrometer operating in the Fourier transform mode.
Quantitative analysis of the labelling pattern in the mixtures
of CH2DCH2CO2Me and CH3CHDCO2Me obtained from
reactor A, below, was carried out by integration of a 13C{1H}
NMR spectrum accumulated with a 10 s pulse delay. Quantit-
ative analysis of the labelled products from reactors B and C
was carried out using the parent ion peak in the GCMS since
for pure methyl propanoate this consists of a single peak at m/e
88. This is flanked by two small peaks at m/e 87 and 89 with
intensities ca. 10% and 4% of that of the major peak. These
were ignored in the analysis for the products from reactor B
but not from reactor C. The identity of the isotopomers was
confirmed by 13C{1H, 2H} NMR spectroscopy.11
Mixtures of partially deuteriated ethenes were analysed by
GCMS. The spectrum of each isotopomer was calculated based
on that of ethene and making the assumptions (i) that loss of H
or D from the parent ion or fragment depends only on the
number of H or D atoms present, i.e. that there is no isotope
effect on fragmentation; (ii) that [M Ϫ 2]ϩ in ethene is [HCCH]ϩ
not [H2CC]ϩ and (iii) that Z- and E-CHDCHD give identical
fragmentation patterns. The peak at 32 amu arises only from
C2D4 and that at 31 amu only from C2D3H so they were used to
calculate the relative amounts of these two isotopomers and
their contribution to the peak at 30 amu, which was subtracted
from the total, leaving a peak corresponding to all isomers of
C2D2H2. The relative contributions to the peak at 29 amu from
all isotopomers with > 1 D atom were subtracted leaving a peak
corresponding to C2DH3. In principle a similar process can be
repeated to obtain the contribution of the peak at 28 amu from
C2H4. In practice this is difficult since the loss of 2 H/D atoms
from CH2CD2 leads only to [CHCD]ϩ, whilst from CHDCHD,
[C2H2]ϩ, [C2HD]ϩ and [C2D2]ϩ are expected in a 1 : 2 : 1 ratio.
Since the relative amounts of these isomers are not known,
there is ambiguity about the contribution of these to the peak at
28 amu. In practice, we used a simulation program, written
in-house, in which the height of each mass spectral peak was
calculated knowing the mass spectra of the individual
isotopomers and their relative abundance. The amounts for
C2D4, C2D3H and C2DH3, calculated as described above were
put in and variations were then made in the amounts of
the other three isotopomers until the best fit was achieved,
always ensuring that the sum of CH2CD2 and CHDCHD was
Catalytic experiments
Reactor A. A catalyst solution consisting of CH3OD (120
cm3), [Pd(DBPMB)(DBA)] (DBA = dibenzylideneacetone)
(30 mg, 4.08 × 10Ϫ5 mol) and MeSO3H (53 µL, 8.2 × 10Ϫ4 mol)
was prepared under argon and sucked into a stainless steel
autoclave (volume 300 cm3), which had previously been evac-
uated, fitted with a paddle stirrer. The solution was heated to 80
ЊC and then the stirrer started (1000 rpm). The autoclave was
opened to a supply of ethene/CO (1 : 1) at 10 bar (t = 0) and
the pressure maintained at 10 bar as the reaction proceeded by
feeding the same gas mixture from a supply vessel through a
constant pressure valve. Samples were taken through a liquid
sampler after 5, 10, 15, 30, 45, 60, 90 and 120 min by removing 5
cm3 to flush the sampling pipe (for all samples except the first),
followed by 5 cm3 for analysis. The samples for analysis were
stored in screw top glass vials in which diffusion of gas between
the headspace and the atmosphere could occur. These were
analysed by GCMS and all contained only d1-methyl propan-
oate. The mass spectrum of the methanol was also measured
and shown to be the same in all samples. (The analysis suggests
that significant amounts (35%) of MeOH are present, but we
have shown that this arises from post reaction exchange). The
2
products were further analysed by 13C{1H} and 13C{1H, H}
NMR11 spectroscopy and shown to contain both CH2DCH2-
CO2Me and CH3CHDCO2Me.
Reactor B. A glass Buchi designed autoclave, fitted with
a magnetic stirrer, was charged via syringe with a deep
red–orange catalyst solution consisting of CH3OD (ca. 30 cm3,
see Table 1), [Pd(DBPMB)(DBA)] (100 mg, 1.36 × 10Ϫ4 mol)
and MeSO3H (44 µL, 6.8 × 10Ϫ4 mol). The reaction was then
initiated as described above except at 90 ЊC. After the desired
reaction time (Table 1), the reactor was sealed and cooled
in a cold water bath. The excess pressure was released, the
contents of the autoclave poured into a sample bottle, and
the mass of the product solution measured to give the
weight gain and hence the yield of methyl propanoate. As with
the samples taken from reactor A there was no barrier to
exchange between the sample bottle headspace and the atmos-
phere in this system. The reaction products, pale yellow–green
solutions, were analysed as described above and the results are
collected in Table 1. GCMS analysis of the methanol showed
1614
J. Chem. Soc., Dalton Trans., 2002, 1613–1617
Eastham, Graham R.
