M. Falcon et al. / Journal of Organometallic Chemistry 629 (2001) 187–193
191
tion of a red insoluble polymer. As the catalytically
active rhodium fragment, in the absence of quenching,
is very likely bound to the insoluble polymeric chain,
the remaining solution is expected to contain only the
rhodium species which did not participate in the cata-
lytic reaction. The 31P-NMR spectrum of this solution
(Fig. 2b) shows 1 and 3 in unvaried amounts, indicating
that both compounds have not reacted with the alkyne;
moreover, compound 2 has disappeared completely,
whereas compound 4 was formed in an amount com-
parable to 1. One can therefore conclude that of the
three compounds, only 2, which was initially present,
has reacted with phenylacetylene partly to form 4 (vide
infra), and partly to catalyze the polymerization to give
a rhodium–polymer adduct insoluble in diethyl ether–
benzene.
Attempts have been made to isolate compound 2
from the reaction of [Rh(nbd)(OMe)]2 and dppb in
THF at 0°C: the solution obtained was treated with
cold methanol giving an orange solid which was filtered
under an inert atmosphere. However, upon standing
rapid decomposition of the precipitate occurred, with
the formation of an oil, which was recovered with light
petroleum. Such a product (6) has a 31P-NMR signal in
CDCl3 consisting of one broad singlet at l 32.1 ppm. A
CD2Cl2 solution of this compound shows no change in
signal width and multiplicity on lowering the tempera-
ture, and at −90°C it still appears as a broad singlet,
no coupling to rhodium becoming apparent even at
that temperature.
of dppb coordinated to the rhodium–polyphenyl-
acetylene species.
3. Discussion
The association of [Rh(nbd)(OMe)]2 and a bidentate
phosphine produces a catalytic system which promotes
the stereoselective polymerization of phenylacetylene.
The system employed in the present work is most
efficient in the absence of basic cocatalysts, in contrast
with all previously reported rhodium–phosphine cata-
lysts for such a reaction, with the exception of the
catalytic system [Rh(nbd)Cl]2/Ar2CꢀC(Ph)Li/PPh3 re-
ported by Masuda et al. [14a].
The li6ing nature of the catalyst is proved by moni-
toring the values of Mn and D with increasing conver-
sion, by the increase of the polymer molecular weight
with increasing initial monomer concentration, by the
results obtained by addition of a second monomer feed
to the reaction mixture after reaching 100% conversion
of the initial monomer feed, and finally by isolation of
a
catalytically active rhodium–polyphenylacetylene
adduct.
A few papers have been published on rhodium–di-
ene–phosphine catalysts which promote alkyne poly-
merization [10,11,14]. Such rhodium catalysts have
been reported to possess one monodentate phosphine
and the diene coordinated in bidentate fashion through-
out the catalytic cycle. Moreover, in the case of
rhodium catalysts the polymerization reaction has al-
ways been proposed — and in one case proven
[10a] — to proceed via a 2,1-insertion mechanism,
which requires the presence of a free coordination site
at rhodium, available for substrate coordination.
Substitution of a monodentate with a bidentate phos-
phine may in principle create a situation of coordina-
tive saturation which is not compatible to this type of
catalysis. Now we find that organorhodium–diphos-
phine catalysts efficiently promote the polymerization
of phenylacetylene. We must therefore assume that one
phosphorous atom goes out of the coordination sphere
of rhodium to generate the catalytically active species.
As reported in Section 2, the ligands PNP and PSP
do not form catalytically active systems with rhodium.
The lack of catalytic activity might be due, in this case,
to irreversible coordination of the ligand in a bidentate
or even tridentate fashion, the latter being not unlikely
especially for PNP [16].
Compound 6 was tested as a catalyst precursor and it
was found actually to promote polymerization of
phenylacetylene in THF; however, the reaction was two
orders of magnitude slower, and the polymer formed
had higher D and Mn (1.55 and 910 000, respectively, at
50% conversion), than the polymer formed by the cata-
lytic system prepared in situ.
Interestingly, a catalytically active rhodium complex
bound to a growing polymer chain was isolated by the
addition of methanol to the reaction mixture in THF:
1
such rhodium–polyphenylacetylene species, whose H-
NMR spectrum shows the typical resonances of cis-
transoid polyphenylacetylene, had Mn=170 000 and
D=1.08. When this adduct was dissolved in THF and
treated with excess phenylacetylene a new polymeriza-
tion was observed, with proportional increase of the
initial polymer chain length with time, until all the
monomer was consumed. This experiment provides fur-
ther evidence for the li6ing nature of the polymerization
reaction. Unfortunately, identification of the rhodium–
polyphenylacetylene complex was impossible, as its 31P-
NMR spectrum only consists of weak signals
attributable to traces of 4 and 6, which are present
probably as impurities. In fact, due to the low number
of rhodium active sites bound to the polymer chain,
one can hardly expect to detect the phosphorous atoms
An initially unexpected result, the lack of activity of
the Rh–dppb catalyst in CHCl3 and CH2Cl2, has been
explained actually by the spectroscopic investigations:
in such solvents only the rhodium compound 1 is
formed, which does not react with phenylacetylene.
In contrast, in other solvents such as THF, together
with unreactive 1, the catalytically active species 2 is
Falcon, Michele
