8584 J. Am. Chem. Soc., Vol. 118, No. 36, 1996
Ritter et al.
In all types, the three cobalt atoms form the corners of a
Scheme 2
triangle capped by a carbon atom. In type A, all of the nine
CO ligands are terminally bonded and have an average metal-
metal distance of 246.8 pm. This compares well with those in
1
(245.46(14)-247.58(12) pm). This type of terminally bonded
CO ligand is exclusively found in the two clusters bonded to
Si1 and Si3 of the molecule. In type B, which is bonded to
Si4, three of the nine CO ligands are arranged in a bridging
mode. The average metal-metal distance is 245.1 pm. While
the Co3 triangle in A and B is highly symmetric, this unit in
cluster C is highly distorted. Here, six of the nine CO ligands
are terminally bonded as in the case of type B. Two other CO
isolation of substantial amounts of internal olefins at the end
of catalytic runs (17-37% after 18 h). In order to test this
supposition, we performed hydroformylation of 2-hexene under
similar conditions with 4 and found little or no conversion of
the olefin to aldehyde. Hence the results indicate that 2 and 3
are equally efficient for the hydroformylation of both terminal
and branched olefins. On the other hand, compound 4 catalyzes
the hydroformylation of terminal olefins with a faster rate.
Furthermore, the selectivity of the formation of terminal
aldehyde 1-heptanal is considerably high in the case of Al
catalyst 2 (Table 1).
These observations put together can be explained by a simple
reaction scheme involving the individual rate constants of the
various processes as shown in Scheme 2. It appears from the
data presented in Table 1 that the rate constants k2 and k3 are
higher compared to k1 in the case of Al and Ga catalysts 2 and
2
ligands are present in a µ -bridged position and the remaining
CO ligand (C4′3, Figure 6) is surprisingly semibridged (Co-C
bond distance 191.7 pm). This semibridging ligand shows a
weak interaction with the adjacent Co4 (Co-C bond distance,
1
98.2 pm).
The late transition metals with higher atomic numbers show
2
an increasing tendency to form µ -bridged carbonyl clusters.
This is attributed to the larger atomic radii resulting in reduced
electron density in these atoms. The occurrence of bridging
ligands in 2 indicates an electron-withdrawing effect of the
aluminosiloxane framework on the cobalt methylidyne units.
The electron-withdrawing nature of the cubic Si8O12 siloxane
framework has recently been experimentally demonstrated by
3
6
Feher and co-workers for a series of spherosilicates. Hence,
it might be expected that the E4O12Si4 (E ) Al, Ga, In)
frameworks in the heterosiloxanes 2-4 would also have similar
electron-withdrawing effects on the Co3(CO)9C substituents,
resulting in bridging and semibridging CO ligands.
3
, while the reverse is true for In compound 4.
In order to have a better understanding of the observed
parameters listed in Table 1, we have performed additional
experiments. The rate of formation of the various products in
each case has been followed as a function of time by drawing
small amounts of the catalytic mixture and analyzing them using
a gas chromatograph. The results of these experiments for all
three catalysts along with the observed n/iso ratio of the
aldehydes are shown in Figure 7. In all the cases there is no
observable induction period for the hydroformylation reactions
to set in. This contrasts with the observation of the earlier
workers with compounds Co3(CO)9CR as the catalysts.38
Another observation that is apparent in this figure is the rate of
formation of the 2-hexene or internal olefins. The formation
of these internal olefins in the case of Al catalyst 2 is very low
and remains more or less constant as a function of time. In the
case of 3, the isomerization increases with time initially up to
Hydroformylation Studies. The parent silanetriol 1 and the
new heterosiloxanes 2-4 show catalytic activity in hydroformyl-
ation reactions involving 1-hexene. The use of 1 and related
polyethylene glycol ethers of 1 for this purpose has already been
described in a preliminary communication.22 The summary of
the present hydroformylation investigations using 1-hexene as
the substrate is presented in Table 1. The experiments were
performed at an initial CO/H2 pressure of 70-80 bar. It should
be noted that this pressure is significantly lower compared to
the pressure of about 200 bar at which hydroformylation
reactions with Co2(CO)8 are usually performed.37
As can be seen from Table 1, the Al and Ga catalysts 2 and
3
show a better catalytic activity than the In analogue 4. For
example, at a 0.02 mmol concentration of the catalysts 2 and
, the observed turnover numbers are 6700 and 6650, respec-
3
2
h, indicating a considerable amount of isomerization. How-
tively. The corresponding value for In catalyst 4 after 18 h of
the catalytic run is only 840. It is also interesting to note that
the numbers observed for 2 and 3 are close to those observed
for silanetriol 1.
The significant observation in all these catalytic runs is the
amount of 1-hexene converted into the aldehydes. While values
closer to 90% are observed irrespective of the concentration of
the catalysts for compounds 2 and 3, in the case of In catalyst
ever, the ability of Ga catalyst 3 in hydroformylating the internal
olefins decreases the amount of internal olefins present in the
solution. In the case of In catalyst 4, the rates of formation of
both aldehydes and internal olefins increase as a function of
time and the process is not complete even after 5 h. Also there
is a drop in the n/iso ratio as a function of time in all the three
cases (Table 1).
Since the effect of catalyst concentration on the catalytic
process is likely to throw insight into the retention or decom-
4
this value is considerably lowered (Table 1). However, in
the case of In catalyst 4, a large amount of olefin isomerization
is found to take place. This implies that the isomerization and
hydroformylation reactions are taking place simultaneously. The
former process results in the formation of internal olefins which
are ultimately converted into aldehydes in the case of 2 and 3.
However, the hydroformylation process of the internal olefins
is apparently less effective with In catalyst 4. This leads to the
39
position of the catalytic structure, we performed a series of
catalyst concentration dependent hydroformylation runs. The
results are pictorially represented in Figure 8. The turnover
numbers quoted are those obtained after 3 h of the catalytic
40
process. It is evident from this figure that the increase in the
concentration of catalysts 2 and 3 lowers the turnover numbers
(38) Withers, H. P.; Seyferth, D. Inorg. Chem. 1983, 22, 2931.
(
36) Feher, F. J.; Budzichowski, T. A. J. Organomet. Chem. 1989, 379,
(39) Laine, R. M. J. Mol. Catal. 1982, 14, 137.
3
3.
(
(40) A comparison of these values with that listed in Table 1 for catalytic
runs carried out for 18 h indicates that the hydroformylation reaction is
almost complete after 3 h in the case of 2 and 3, while the reaction with 4
is only half-way through.
37) Cornils, B. ReactiVity and structure concepts in organic chemistry
1
1: New synthesis with carbon monoxide; Falbe, J., Ed.; Springer Verlag:
Berlin, Heidelberg, New York, 1980.