3764 Organometallics, Vol. 16, No. 17, 1997
Ferna´ndez-Gala´n et al.
the alkene rotation. The only way to interconvert these
signals is by amino group inversion after the rupture
of the Pd-N bond. However, PPFA must remain
coordinated to the palladium center via the phosphorus
atom. Otherwise, a unique alkene proton signal would
ture range studied and, hence, the Pd-N bond rupture
is not observed for complex 5. (The corresponding ∆Gq
c
values are listed in Table 5.) Considering possible
mechanisms to account for the observed isomer inter-
conversion, an allyl rotation in its own plane may be
proposed. Although this mechanism is often put for-
ward to account for some isomerizations in complexes
of metals such as Mo, W, and Fe, it is not widely
accepted for square-planar palladium complexes. Or-
bital considerations suggest a high activation barrier
for complexes with square-planar geometries.25 Two
consecutive Berry pseudorotations in a pentacoordinate
activation state could also explain allylic interconver-
sions.25,26 We might also consider an “apparent π-allyl
rotation” mechanism similar to that proposed by other
authors15 and also by us24 for certain allylpalladium
complexes with N-donor ligands. That implies, how-
ever, a partial dissociation of the bidentate ligand.
Since from the experimental data a Pd-N bond rupture
can be rejected, such a mechanism would involve a
Pd-P bond rupture, a process which is very unlikely,
considering the strong Pd-P bond. In any case, both
mechanisms can be rejected for an additional reason:
they would not show the observed exchange of the
q
be observed. Surprisingly, the ∆G c value calculated for
this coalescence (see Table 5) fits properly with the
linear plot obtained for the other coalescences noted for
this system. Therefore, the activation barrier for this
bond rupture process is comparable to that of the
rotamer interconversion phenomenon.
For complex 4, an isomer interconversion is not
observed within the temperature range studied. For
example, the unsubstituted cyclopentadienyl signals
show only a small shift difference (δν 24 Hz) but do not
coalesce, at least up to the maximum temperature
recorded (378 K). At this temperature they are slightly
broadened. Thus, on the NMR time scale, neither an
intermolecular alkene interchange nor an N-P ligand
dissociation is observed. However, an interchange of
the alkene signals (alkene protons and ester methyls)
of each individual isomer is observed. At the highest
temperature recorded, only one alkene proton as well
as one ester methyl signal is seen for each isomer. In
principle, this exchange of alkene resonances can be
explained by an olefin rotation process. Besides, as for
3, the two diastereotopic amino methyls of each indi-
vidual isomer exchange. Above the coalescence tem-
peratures corresponding to this interchange, two broad
singlets are seen that narrows when the temperature
is raised (see Figure 1). These observations are consis-
tent with a Pd-N bond rupture. This process must be
of lower energy than that of an olefin decoordination, a
reasonable process for an isomer interconversion. In
order to confirm this statement, we have monitored the
transformation of pure 4M into 4m in toluene-d8 at 243
H
syn,trans P(isomer 5M) T Hsyn,trans P(isomer 5m ) signals
but would lead to an Hsyn,trans P(isomer 5M) T
Hsyn,cis P(isomer 5m ) interchange. A η3-η1-η3 rear-
rangement of the allyl group could explain the observed
interchange, but with such a mechanism operating, the
typical Hsyn-Hanti interconversion should also be present.
Because not all allyl hydrogen interchanges were identi-
fied unequivocally from the monodimensional study, we
have performed an EXSY experiment.27 The allyl
methyl groups of the two isomers show exchange cross-
peaks, evidencing the isomer interconversion. Exami-
nation of the region corresponding to the allyl protons
cis to phosphorus reveals the following exchange
cross-peaks: Hsyn(5M) T Hanti(5m ) and Hanti(5M) T
K until an equilibrium ratio was reached. A ∆Gq
243
value of 78 ( 2 kJ mol-1 is obtained, significantly higher
than the values listed in Table 5. Consequently, an
isomer interconversion via a decoordination-coordina-
tion mechanism takes place but is of higher energy than
those processes observed by our NMR experiments.
H
syn(5m ) (see Figure 2). The region of the allyl protons
trans to phosphorus is not resolved sufficiently due to
very similar chemical shifts of some allyl protons and
in addition due to overlap with some cyclopentadienyl
resonances. In any case, the Hsyn(5M) T Hsyn(5m )
interconversion has already been observed in the mon-
odimensional study. All these exchanges, depicted in
Scheme 1, are consistent with the formation of a
selective σ-allyl intermediate cis to phosphorus as shown
in Scheme 2(i). In this intermediate, the Pd-C and
C-C bond rotation (ii) and re-formation of the π form
(iii) allows the H1-H2 interchange and the isomer
interconversion. A σ-allyl intermediate trans to phos-
phorus can be excluded for complex 5, because the
Allyl Com p lexes. The variable temperature 1H
NMR studies for the allyl compounds 5 and 6 have been
recorded at first in acetone-d6 because these complexes
are quite insoluble in toluene (see Tables 1 and 3 for
the assignment of the signals at 233 K). When the
temperature was increased, all the signals of 5 and 6
broadened but coalescence could not be achieved. The
relatively low boiling point of the solvent prevents a
complete study. Hence, further variable-temperature
studies were carried out in 1,1,2,2-tetrachloroethane-
d2. For both complexes, the isomer ratio at room
temperature is similar to that observed in acetone-d6
(see Table 1); when the temperature is increased, an
isomer interconversion is observed unambiguously (see
below).
interchange Hsyn(5M) T Hsyn(5m ) and Hanti(5M) T Hanti
-
(5m ) is not observed in the spectral region correspond-
ing to the allyl protons cis to phosphorus. Such a
selective η3-η1-η3 isomerization has also been reported
by Pregosin et al. for chiral allyl complexes with
For 5, an interconversion between the two isomers
was observed and coalescence temperatures have been
measured for the nonfunctionalized cyclopentadienyl,
the Hsyn,trans P, and also the amino methyl resonances.
The two resulting signals obtained in the amino methyl
region after the coalescence narrow to 413 K. This
means that for like isomers an exchange of the diaste-
reotopic amino methyls does not occur in the tempera-
(25) Vrieze, K. In Dynamic Nuclear Magnetic Resonance Spectros-
copy; J ackman, L. M., Cotton, F. A., Eds.; Academic Press: New York,
1975.
(26) (a) Crociani, B; Di Bianca, F.; Giovenco, A.; Boschi, T. Inorg.
Chim. Acta 1987, 127, 169. (b) Hansson, S.; Norrby, P. O.; Sjo¨gren, M.
P. T.; A° kermark, B.; Cucciolito, M. E.; Giordano, F.; Vitagliano, A.
Organometallics 1993, 12, 4940.
(27) Bodenhausen, G.; Ernst, R. R. J . Am. Chem. Soc. 1982, 104,
1304.