Mechanistic studies of alkene/CO polymerization with palladium complexes promoted by B(C6F5)3

Article abstract of DOI:10.1021/om990911l

The reaction of Pd(dppp)(OAc)₂ [dppp = 1,3-bis(diphenylphosphino)propane] with B(C₆F₅)₃ in situ gives an efficient catalyst for alkene/CO polymerization in aprotic media. The borane is consumed during the polymerization, and its fluoroaryl groups are incorporated into the polymer chain ends. In the absence of the monomers the catalyst components react to give palladium(II)-pentafluoroaryl complexes formulated as Pd(dppp)(C₆F₅){[B(C₆F₅)₃]_y OH} (y = 1, 3a; y = 2, 3b). Complex 3a can be isolated, albeit as an impure solid, and is itself a catalyst for the reaction. In light of these results a novel chain initiation process for the polymerization is proposed, involving insertion of monomers into a fluoroarylpalladium complex formed by aryl transfer from the borane to a palladium(II) complex. This facile initiation step, combined with the catalyst stability engendered by the presence of strong Bronsted acids, explains the effectiveness of this catalyst system in aprotic media.

Full text of DOI:10.1021/om990911l

1470
Organometallics 2000, 19, 1470-1476
Mech a n istic Stu d ies of Alk en e/CO P olym er iza tion w ith
P a lla d iu m Com p lexes P r om oted by B(C6F 5)3
Graham K. Barlow, J ane D. Boyle, Neil A. Cooley,*,† Talit Ghaffar, and
Duncan F. Wass
BP Amoco Chemicals, Chemicals Research and Engineering, Poplar House, Chertsey Road,
Sunbury-on-Thames, Middlesex, TW16 7LL, United Kingdom
Received November 19, 1999
The reaction of Pd(dppp)(OAc)2 [dppp ) 1,3-bis(diphenylphosphino)propane] with B(C6F5)3
in situ gives an efficient catalyst for alkene/CO polymerization in aprotic media. The borane
is consumed during the polymerization, and its fluoroaryl groups are incorporated into the
polymer chain ends. In the absence of the monomers the catalyst components react to give
palladium(II)-pentafluoroaryl complexes formulated as Pd(dppp)(C6F5){[B(C6F5)3]yOH} (y
) 1, 3a ; y ) 2, 3b). Complex 3a can be isolated, albeit as an impure solid, and is itself a
catalyst for the reaction. In light of these results a novel chain initiation process for the
polymerization is proposed, involving insertion of monomers into a fluoroarylpalladium
complex formed by aryl transfer from the borane to a palladium(II) complex. This facile
initiation step, combined with the catalyst stability engendered by the presence of strong
Brønsted acids, explains the effectiveness of this catalyst system in aprotic media.
In tr od u ction
lishing an accurate model of the alternating chain
growth process, the conclusions of the latter work being
supported by density functional calculations.6 A gener-
alized representation of the active species is shown in
Figure 1.
The synthesis of alternating copolymers of carbon
monoxide and alkenes, particularly ethene, using pal-
ladium-based catalysts is of considerable industrial and
academic interest.1 Industrial involvement in this area
stems fundamentally from the possibility of manufac-
turing high-performance polymers in one step from
inexpensive raw materials, with recent impetus coming
from the discovery by Drent of high-rate catalysts
incorporating chelating phosphine ligands.2 Since this
discovery, a wide range of alkene/CO copolymers have
been prepared, and the mechanism of the reaction has
been probed extensively. Early mechanistic work by Sen
et al.,3 predating the discovery of high-rate catalysts,
established that the active catalytic species are mono-
cationic palladium complexes and that polymer chain
growth proceeds by alternating migratory insertion
reactions of acyl and alkyl complexes. Mechanistic detail
regarding polymer chain initiation and termination
reactions was added later.4 Brookhart and co-workers5
have further defined the mechanism, studying in detail
the individual steps of the polymerization and estab-
We report here a study of palladium-based alkene/
CO polymerization catalysts incorporating the Lewis
acidic borane B(C6F5)3,7 where the presence of the
borane results in novel catalytic behavior.8 Our inves-
tigations of these catalysts have revealed that the aryl
groups from the borane are incorporated into the end
groups of the polymers produced and that the borane
is consumed during the polymerization. We also report
the results of mechanistic studies of this catalytic
system, which reveal that the use of the borane pro-
moter results in a novel chain initiation process involv-
ing aryl transfer from the borane to palladium. These
results largely explain the wide utility of this borane
in the catalysis of alkene/CO copolymerization. In
addition, the well-defined aryl transfer step has broader
relevance to the large body of polymerizations that use
a fluoroaryl boron component.9
Resu lts a n d Discu ssion
* To whom correspondence should be addressed.
†E-mail: cooleyna@bp.com.
1. Ca ta lytic Rea ction s w ith B(C6F 5)3. Initial stud-
ies showed that the in situ reaction of Pd(dppp)(OAc)2
(1) and B(C6F5)3 in dichloromethane gave a highly active
catalyst for alkene/CO polymerization. A typical cumu-
lative gas uptake profile for an alkene/CO polymeriza-
(1) For some leading references see: (a) Sen, A. CHEMTECH 1986,
48. (b) Drent, E.; Budzelaar, P. H. M. Chem. Rev. 1996, 96, 663. (c)
Gray, A. Chem. Br. 1998, March, 44, and references therein.
(2) Drent, E. Eur. Pat. Appl. 121,965, 1984; Chem Abstr. 1985, 102,
46423.
(3) (a) Sen, A.; Lai, T.-W., J . Am. Chem. Soc. 1982, 104, 3520. (b)
Lai, T.-W.; Sen, A., Organometallics 1984, 3, 866.
(4) (a) Drent, E.; Van Broekhoven, J . A. M.; Doyle, M. J . J .
Organomet. Chem. 1991, 417, 235. (b) Barasacchi, M.; Consiglio, G.;
Medici, L.; Petrucci, G.; Suter, U. W. Angew. Chem., Int. Ed. Engl.
1991, 30, 989. (c) Zhao, A. X.; Chien, J . C. W. J . Polym. Sci., Part A:
Polym. Chem. 1992, 30, 2735. (d) Sen, A. Acc. Chem. Res. 1993, 26,
303.
(6) Margl, P.; Ziegler, T. J . Am. Chem. Soc. 1996, 118, 7337.
(7) (a) Massey, A. G.; Park, A. J . J Organomet. Chem. 1964, 2, 245.
(b) Massey, A. G.; Park, A. J . J Organomet. Chem. 1966, 5, 218.
(8) For the first diclosure of alkene/CO copolymerization promoted
by B(C6F5)3 see: Cooley, N. A.; Kirk, A. P. Eur. Pat. Appl. 619335;
Chem. Abstr. 1995, 112, 550920.
(5) (a) Rix, F. C.; Brookhart, M. J . Am. Chem. Soc. 1995, 117, 1137.
(b) Rix, F. C.; Brookhart, M.; White, P. S. J . Am. Chem. Soc. 1996,
118, 4746.
(9) Piers, W. E.; Chivers, T. Chem. Soc. Rev. 1997, 26, 345, and
references therein.
10.1021/om990911l CCC: $19.00 © 2000 American Chemical Society
Publication on Web 03/17/2000

Alkene/ CO Polymerization with Pd Complexes
Organometallics, Vol. 19, No. 8, 2000 1471
F igu r e 1. Generalized active species for alkene/CO copo-
lymerization.
F igu r e 3. Polymer end groups observed in a typical
ethene/propene/CO terpolymerization reaction catalyzed by
1/B(C6F5)3.
rophenyl end groups. The pentafluorophenyl end groups
can be differentiated into the substructures c, d , and e
(Figure 3).
Under typical reaction conditions alkyl end groups
1
predominate, and from H and 13C NMR spectra it was
determined that these outnumber alkenyl groups in the
ratio 4:1. A combination of polymer analyses (Mn from
GPC and fluorine microanalysis) leads to the estimate
that, under typical reaction conditions, pentafluorophen-
yl end groups constitute about 15-25% of the total end
groups. The relative ratio of these pentafluorophenyl
end groups can be more accurately determined by 19F
NMR (see Experimental Section), and under represen-
tative conditions substructures c, d , and e were found
to constitute 5, 76, and 19%, respectively, of the total
pentafluorophenyl end groups. Fluorine microanalysis
and 19F NMR spectroscopy of a complete reaction
mixture showed that at the end of the reaction 47% of
the pentafluorophenyl groups were present as polymer
end groups; the remainder had formed pentafluoroben-
zene. The consumption of the borane accords with the
experimental observation that addition of B(C6F5)3 will
reactivate the catalyst (see above).
The above end group characterization shows that
incorporation of pentafluorophenyl groups is a signifi-
cant pathway in alkene/CO polymerization catalysis by
1/B(C6F5)3. Furthermore, the observation that these
catalysts can be reinitiated by the addition of further
B(C6F5)3 indicates strongly that this reagent is involved
in catalyst initiation.
2. Rea ction of 1 w ith B(C6F 5)3. This reaction was
studied by NMR spectroscopy in order to gain insight
into the formation of the active species. After initial
experiments gave complex mixtures of products, a series
of experiments was conducted in which water and
borane concentrations were varied systematically. These
experiments comprised two parallel sets of reactions,
one set in essentially dry CD2Cl2, used as received, and
one in CD2Cl2, to which water had been added. In each
set 0.5, 1, 2, and 5 molar equiv of B(C6F5)3 relative to 1
were added. NMR spectra across a range of nuclei were
then recorded; the 31P{1H} NMR spectra are shown in
Figure 4.
F igu r e 2. Cumulative gas uptake profile for a typical
ethene/propene/CO terpolymerization reaction catalyzed by
1/B(C6F5)3.
tion with the catalyst formed from 1 and B(C6F5)3, run
under constant pressure conditions, shows a steady rate
of reaction followed by a precipitous falloff in gas uptake
later in the reaction (Figure 2).
Reaction rates for 1/B(C6F5)3 are typically 10 kg g
Pd-1 h-1 for the production of ethene/propene/CO ter-
polymers with approximately 6 propene monomers per
100 monomers and number average molecular weights
(Mn) of approximately 20 000 (ref 10). Catalyst longevity
varies, but in general a high borane concentration leads
to longer catalyst lifetime. The catalyst can be reacti-
vated by addition of more B(C6F5)3, although the
catalyst does not usually reattain the rates seen ini-
tially. Addition of solutions of 1 to deactivated catalyst
solutions does not result in further polymerization.
Therefore, the observed decrease in activity is attribut-
able primarily to borane depletion. However, some
catalyst degradation is also evidently occurring, as
adding another charge of borane does not give the same
high rate obtained initially.
The polymers produced using this catalyst have
molecular weight distributions in the expected region
for single-site catalysts with polydispersities (Mw/Mn)
generally between 2 and 3. Propene incorporation is
random as far as can be determined by NMR spectros-
copy and is predominantly 1,2-, although some 1,3-
incorporation is observed, as is the case for other
catalysts.4a The end groups of the polymers however
differ significantly from those seen previously. 19F and
13C NMR spectra of a representative polymerization
show the presence of alkyl, a , alkenyl, b, and pentafluo-
The product formed on adding 0.5 molar equiv of
B(C6F5)3 to 1 is the same both in the absence and
presence of added water. On the basis of its NMR
spectra this complex is formulated as Pd(dppp)(C6F5)-
(OAc), 2 (see Experimental Section). Addition of 1 molar
(10) Based on the assumption that Mn as determined by gel
permeation chromatography of a polymer solution in hexafluoro-2-
propanol relative to a poly(methyl methacrylate) standard is ap-
proximately 3 times the absolute value.

1472 Organometallics, Vol. 19, No. 8, 2000
Barlow et al.
F igu r e 4. 31P{1H} NMR spectra from reaction of 1 with 0.5, 1, 2, and 5 equiv of B(C6F5)3 in both wet and dry CD2Cl2.
equiv of B(C6F5)3 to 1 also results in the formation of 2,
and additionally, some precipitation from solution oc-
curs. The concurrent formation of pentafluorobenzene
in our experiments at this borane concentration both
with and without added water indicates that adventi-
tious water was present in all cases. All of our reactions
therefore contain strong Brønsted acids of the form
B(C6F5)3‚xH2O, formed from the reaction of B(C6F5)3
with adventitious H2O in the dichloromethane solvent.11
Addition of 2 molar equiv of B(C6F5)3 to 1 again gives
pentafluorobenzene in significant quantities. Only small
amounts of species with Pd-C6F5 groups were detected
however, and on the basis of the NMR spectra, it is
likely that dicationic species of the general formula
[Pd(dppp)L2]2+2A- are the major products. The likely
path for the formation of such species is the protonolysis
of the initially formed Pd-C6F5 complexes by strong
Brønsted acids, formed in the presence of water, to give
pentafluorobenzene. It should be noted that the borane
does not react significantly with water to give pen-
tafluorobenzene in the absence of the palladium complex
under these conditions.
Addition of 5 molar equiv of B(C6F5)3 to 1 results in
the immediate formation of two similar complexes, 3a
and 3b. In our experiments complex 3a predominates
in the presence of added water, and only 3b is formed
when only adventitious water was present. In addition
to 3a and 3b, a small amount of Pd(dppp)(C6F5)2 (4) was
also obtained. Complex 3a was readily converted to 3b
in solution by the addition of a further 5 equiv of
B(C6F5)3. By contrast, our attempts to convert 3b to 3a
by water addition resulted in a mixture of products. It
is clear from the 31P{1H} and 19F NMR spectra (see
Experimental Section) that both 3a and 3b contain the
diphosphine and a coordinated C6F5 group, leaving the
ligand at the fourth coordination site as the major point
of difference. Examination of the 19F NMR spectra of a
number of solutions of 3a shows that the complex
contains a single B(C6F5)3 moiety with resonances at
-135.1, -161.5, and -166.8 ppm. Similarly 3b contains
two equivalent B(C6F5)3 moieties with 19F NMR signals
at -135.1, -160.6, and -166.6 ppm. On the basis of
these NMR chemical shifts and the precedent of the
ligation of a B(C6F5)3OH moiety in platinum chemis-
try,12 we have formulated 3a as Pd(dppp)(C6F5)[B(C6F5)3-
OH]. It is therefore likely that the conversion of 3a to
3b upon addition of B(C6F5)3 is analogous to the recently
reported reaction13 of B(C6F5)3OH- with B(C6F5)3 to give
B(C6F5)3(µ-OH)B(C6F5)3-, and on this basis we have
formulated 3b as Pd(dppp)(C6F5)[B(C6F5)3(µ-OH)B-
(C6F5)3]. This structure is fully consistent with the
NMR data. Complex 3a can be obtained as an impure
solid by precipitation from solution with hexanes (see
Experimental Section). Redissolution of this solid in
CD2Cl2 and treatment with B(C6F5)3 again gives 3b,
confirming that 3b is best formulated as an adduct of
3a . The above results and structural proposals are
summarized in Scheme 1.
3. Releva n ce of Ar yl Tr a n sfer Rea ction s to Ca -
ta lysis of Alk en e/CO P olym er iza tion . The facile
transfer of a pentafluorophenyl group from boron to
palladium is surprising; compounds such as B(C6F5)3
-
and anions such as B(C6F5)4 are generally employed
in catalysis precisely because of their inertness to
reaction with electrophilic metal centers.9 Aryl transfer
from B(C6F5)3 has been observed previously, but the
reactions were slow.14 This reaction also has precedent
in the transfer of aryl groups from arylboranes to
platinum(II), where aryl transfer is believed to be
facilitated by bridged intermediates, formed by coordi-
nated water for example.15 It is also relevant that aryl
(12) Hill, G. S.; Manojlovic-Muir, L.; Muir, K. W.; Puddephatt, R.
J . Organometallics 1997, 16, 525.
(13) Danopoulos, A. A.; Galsworthy, J . R.; Green, M. L. H.; Caffer-
key, S.; Doerrer, L. H.; Hursthouse, M. B. Chem. Commun. 1998, 2529.
(14) (a) Yang, X.; Stern, C. L.; Marks, T. J . J . Am. Chem. Soc. 1994,
116, 10015. (b) Scollard, J . D.; McConville, D. H.; Rettig, S. J .
Organometallics 1997, 16, 1810. (c) Pindado, G. J , Lancaster, S. J .;
Thornton-Pett, M.; Bochmann, M. J . Am. Chem. Soc. 1998, 120, 6816.
(11) Siedle, A. K.; Lamanna, W. M., Newmark, R. A., Stevens, J .,
Richardson, D. E., Ryan, M. Makromol. Chem., Makromol. Symp. 1993,
66, 215.

Alkene/ CO Polymerization with Pd Complexes
Organometallics, Vol. 19, No. 8, 2000 1473
Sch em e 1. P r op osed Rea ction s of 1 w ith B(C6F 5)3
the idea that the aryl transfer is not a conventional
electrophilic aromatic substitution.18
The above observations firmly establish the possibility
that aryl transfer to a palladium(II) precursor is a step
in the catalytic cycle. Furthermore, the presence of
water sets up the well-precedented protonolysis of the
initially formed arylpalladium(II) complex, and this
provides a potential route to pentafluorobenzene,19 the
major byproduct of the alkene/CO polymerization reac-
tion using catalysts containing B(C6F5)3.
To further probe the possible intermediacy of the
above palladium(II) arylation reaction in alkene/CO
copolymerization, further polymerizations were con-
ducted using 3a as the procatalyst. In the absence of
additional B(C6F5)3 the reaction gave an initial rate of
4 kg g Pd-1 h-1 but quickly deactivated. With added
B(C6F5)3 the polymerization gives an activity similar to
the 1/B(C6F5)3 catalyst, cf. 8.5 kg g Pd-1 h-1 and 10 kg
g Pd-1 h-1, respectively, and a longer catalytic lifetime.
In the absence of additional B(C6F5)3 the end group
distribution obtained with 3a is very similar to that
obtained with 1/B(C6F5)3, with substructures c, d , and
e constituting 9, 77, and 14%, respectively, of the total
pentafluorophenyl end groups. These reactions are fully
consistent with the proposal that aryl transfer to a
palladium(II) complex is the route by which the polym-
erization is initiated.
4. F or m a tion of Alk yl a n d Alk en yl En d Gr ou p s.
In the closely related palladium- and nickel-based
alkene polymerization catalysts reported by Brookhart
and co-workers20 chain termination results from â-hy-
dride elimination followed by associative ligand substi-
tution of the alkene coordinated polymer chain by the
substrate, giving unsaturated polymer end groups. It
is likely that a similar chain termination operates in
the present case to give an intermediate palladium
hydride (eq 2).
and alkyl transfer reactions from organoboron reagents
are part of the catalytic cycles for many palladium com-
plex-catalyzed cross-coupling reactions.16 Here again
bridged species, particularly those formed via alkoxide
complexes, are thought to be important in aiding
transfer from boron.17 To test whether similar interme-
diates are involved in aryl transfer in this case, the
reactions of [Pd(dppp)(PhCN)2][BF4]2, 5, with a variety
of organoboron compounds have been studied. The
reaction of 5 with B(C6F5)3 is facile, even under es-
sentially anhydrous conditions, giving [Pd(dppp)(PhCN)-
(C6F5)]BF4, 6, identified from its NMR spectra. In
contrast a solution of 5 and B(p-C6H4F)3 in dichlo-
romethane, under analogous conditions, is stable with
respect to aryl transfer (approximately 4% formation of
aryl transfer product after 1 h). However, when a
dichloromethane solution of 5 is treated with Na[B(p-
C6H4F)3OH], a rapid reaction occurs to give a mixture
comprising mainly (64%) a monocationic arylated prod-
uct 7 (eq 1).
(1)
(2)
The contrasting reactivity of B(p-C6H4F)3 and Na[B(p-
C6H4F)3OH] with 5 confirms that aryl transfer is
facilitated by the presence of groups capable of bridging
boron and palladium (eq 1). In addition, it is notable
that transfer from B(C6F5)3 is more facile than transfer
from B(p-C6H4F)3, and this observation lends weight to
As well as producing alkenyl end groups, insertion of
ethene into the intermediate hydride species results in
a alkyl chain end. Therefore, a â-elimination chain
transfer mechanism, operating alone, would produce
equal numbers of alkenyl and alkyl polymer end groups.
Given that alkyl end groups substantially outnumber
(15) Siegmann, K.; Pregosin, P. S.; Venanzi, L. M. Organometallics
1989, 8, 2659.
(16) (a) Littke, A. F.; Gregory, C. F. Angew. Chem., Int. Ed. 1998,
37, 3387. (b) Old, D. W.; Wolfe, J . P.; Buchwald, S. L. J . Am. Chem.
Soc. 1998, 120, 9722. For selected reviews and monographs, see: (c)
Suzuki, A. Acc. Chem. Res. 1982, 15, 178. (d) Miyaura, N., Yamada,
K., Suginome, H. Suzuki, A., J . Am. Chem. Soc. 1985, 107, 972. (e)
Suzuki, A. Pure Appl. Chem. 1985, 57, 1749. (f) Suzuki, A. Pure Appl.
Chem. 1991, 63, 419. (g) Suzuki, A. Pure Appl. Chem. 1994, 66, 213.
(h) Miyyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457.
(17) Suzuki, A. In Metal-Catalysed Cross Coupling Reactions;
Diedrich, F., Stang, P. J ., Eds.; Wiley-VCH: Weinheim, Germany,
1998; Chapter 2.
(18) Electron-withdrawing substutents on the aryl groups of the
borane would be expected to slow the rate of aryl transfer for a reaction
pathway analogous to, for example, the proton-induced cleavage of aryl
borates (see ref 15). The opposite substituent dependency is observed
in the present case.
(19) Kurosawa, H.; Urabe, A.; Miki, K.; Kasai, N. Organometallics
1986, 5, 2002.
(20) J ohnson, L. K.; Killian, C. M.; Brookhart, M. J . Am. Chem. Soc.
1995, 117, 6414.

1474 Organometallics, Vol. 19, No. 8, 2000
Barlow et al.
Sch em e 2. P r op osed Mech a n ism for Alk en e/CO P olym er iza tion w ith P a lla d iu m Com p lexes P r om oted by
B(C6F 5)3
alkenyl end groups, there must be another route to
saturate chain ends. It is likely therefore that alkyl end
groups are also formed by protonolysis: a well-prece-
dented termination mechanism in alkene/CO polymer-
ization.21 This seems especially likely given the avail-
ability of hydrated B(C6F5)3 during catalysis, providing
the necessary Brønsted acid. The molecular weight of
the polymer is however largely insensitive to the
concentration,22 or even the type, of triarylborane used.
This is best explained by protonolysis taking place after
a relatively slow isomerization of an alkyl palladium
complex into an enolate species (eq 3).21
is usually inserted first. Following chain propagation,
some termination occurs via â-hydrogen elimination/
displacement. However, protonolysis is the main ter-
mination route, leaving the catalyst ready for reinitia-
tion by aryl transfer. The pentafluorobenzene byproduct
results from protonolysis of the palladium aryl inter-
mediates. Thus we have two polymer-producing cata-
lytic cycles (A and B) and a catalytic cycle (C) producing
pentafluorobenzene (Scheme 2).
6. Ca ta lysis in th e Absen ce of Bor a n es. The minor
polymer-producing catalytic cycle (B) does not result in
the consumption of the promoter, and it would be de-
sirable if the reaction could proceed through this cycle
alone in an aprotic diluent. This has been achieved
previously with low-activity catalysts under mild condi-
tions.23 We have employed a number of catalysts in
order to achieve this goal, but have not matched the
catalytic performance of 1/ B(C6F5)3; the use of 3a
without added B(C6F5)3 as a procatalyst, for example,
falls into this category. Our best results have come from
using 1 in combination with [B(C6F5)4][(CH3)2NH(C6H5)]
in dichloromethane,23 which forms a mildly active
catalyst, prone to deactivation.24 It is noteworthy that
products formed with this system show no evidence for
aryl transfer.
(3)
5. P r op osed Mech a n ism . The above evidence has
led to a working hypothesis in which polymerization is
initiated by aryl transfer from boron to palladium; this
gives a Pd-aryl complex which undergoes migratory
insertion of a monomer. Analysis of pentafluorophenyl
polymer end groups indicates that an ethene molecule
Con clu sion s
In summary, the combination of a facile and efficient
means of catalyst initiation combined with increased
catalyst stability with respect to reduction25 afforded by
(21) Zuideveld, M. A.; Kramer, P. C. J .; Van Leewen, P. W. N. M.;
Klusener, P. A. A.; Stil, H. A.; Roobek, C. F. J Am. Chem. Soc. 1998,
120, 7977.
(22) Initial polymerization rates and intrinsic viscosities (a measure
of molecular weight) of the products from ethene/propene/CO poly-
merizations with 1/B(C6F5)3 and 1/B(p-C6H4Cl)3 under standard condi-
tions (see Experimental Section) were found to be identical, within
experimental error.
(23) See ref 3 for example.
(24) The highest rate reaction we have seen with the system gave
an initial polymerization rate of approximately 4000 g g Pd-1 h-1, with
steady deactivation to a rate of approximately 300 g g Pd-1 h-1 over a
period of 30 min. The conditions were similar to those detailed in the
Experimental Section except that [B(C6F5)4][(CH3)2NH(C6H5)] was used
in place of B(C6F5)3 and higher concentrations of both catalyst
components were employed.

Alkene/ CO Polymerization with Pd Complexes
Organometallics, Vol. 19, No. 8, 2000 1475
alternating units originating in CO and alkene, with 6.2 per
100 monomers originating in propene. Fluorine microanaly-
sis: 1600 ppm. Palladium microanalysis: 36 ppm. Alkyl and
alkenyl chain ends were quantified using 1H NMR spectros-
copy: ethyl ends δ ) 1.10 (t, J HH ) 7.6 Hz, CH3CH2C(O)-);
propyl ends δ ) 1.58 (m, J HH ) 7.2 Hz, CH3CH2CH2C(O)-);
propenyl ends δ ) 7.03 (dq, J HH ) 15.3, 6.8 Hz, CH3CHdCHC-
(O)-), 1.94 (dd, J HH ) 6.8, 1.5 Hz, CH3CHdCHC(O)-); other
signals are obscured by overlap with main-chain resonances.
Fluorinated chain ends were quantified using 19F NMR
spectroscopy: C6F5C(O)- δ ) -140.4 (ortho-F), -148.7 (para-
F), -160.6 (meta-F); C6F5CH2CH2C(O)- δ ) -144.2 (ortho-
F), -157.8 (para-F), -163.2 (meta-F); C6F5CH2CH(CH3)C(O)-
δ ) -143.6 (ortho-F), -157.8 (para-F), -163.2 (meta-F).
NMR Sca le Rea ction s of 1 w ith B(C6F 5)3 (in Situ
P r ep a r a tion of 2, 3a , a n d 3b): Gen er a l Con sid er a tion s.
All manipulations were carried out in a drybox, the NMR
tubes used having a Youngs’ tap fitting. Experiments in “dry”
CD2Cl2 used this solvent as received in 10 mL ampules from
Aldrich; experiments in “wet” CD2Cl2 used a stock solution of
the same solvent to which deionized water had been added
(1.8 µL of H2O in 10 mL of CD2Cl2). Spectra were initially
recorded within 30 min, although later experiments showed
that in all cases reactions were essentially complete by the
time spectra could be recorded, and no changes were observed
in the reaction products even after extended periods (up to 1
week). The mole percent (mol %) of reaction products are given
as percentages of the total soluble palladium species as
determined by 31P{1H} NMR spectroscopy.
the presence of B(C6F5)3 and its hydrates provides a
convenient means of generating high-activity alkene/
CO polymerization catalysts in aprotic diluents. This
adds flexibility to the conditions under which such
polymerizations can be conducted, but at the cost of the
consumption of the borane promoter.
Integral to this catalysis is the facile initiation by aryl
transfer from boron to palladium(II). This is surprising
given the strong electron-withdrawing substituents on
B(C6F5)3 and may be aided by the formation of bridged
intermediates which lower the activation barrier to
electrophilic aromatic substitution. This result is also
more generally important as arylboranes and borates
are widely used as cocatalysts, particularly for polym-
erization catalysis. A wide variety of polymerization-
active systems have been reported, many of which are
based on late transition metals,26 and it would be
prudent to be aware of possible aryl transfer processes
taking place when these catalysts are used in combina-
tion with aryl boranes or borates.
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All operations were performed
in a dinitrogen-filled Vacuum Atmospheres glovebox or using
Schlenk line techniques. Solvents were purified by distillation
from an appropriate drying agent: dichloromethane (CaH2)
and n-hexane (Na/Ph2CO). B(C6F5)3 was purchased from
Boulder Scientific and used without further purification for
polymerization experiments and preparative scale reactions.
For NMR scale reactions this reagent was sublimed prior to
use. Ethene, carbon monoxide, and propene were purchased
from Air Products. The starting materials [Pd(dppp)(OAc)2],
1, and [Pd(dppp)(PhCN)2](BF4)2, 5, were synthesized by slight
modifications of literature procedures for close analogues.27
Rea ction s in “Dr y” CD2Cl2: (i) Rea ction of 1 w ith 0.5
equ iv of B(C6F 5)3. CD2Cl2 (1 mL) was added to 1 (6.4 mg,
0.01 mmol) and B(C6F5)3 (2.6 mg, 0.005 mmol) in an NMR tube
to give a colorless solution. A mixture of unreacted 1 (60 mol
%) and 2 (40 mol %) was obtained. 1: 1H NMR δ ) 7.70 (m,
8H, ArH), 7.43 (m, 4H, ArH), 7.36 (m, 8H, ArH), 2.48 (m, 4H,
-CH2CH2CH2-), 2.05 (m, 2H, -CH2CH2CH2-), 1.26 (s, 6H,
-O(O)CCH3); 31P{1H} NMR δ ) 10.7 (s). 2: 1H NMR δ ) 1.47
(br, O(O)CCH3); 19F NMR δ ) -117.1 (m, 2F, ortho-C6F5),
-162.8 (br t, 1F, para-C6F5), -164.1 (t, 2F, meta-C6F5);
1
Other reagents were obtained from commercial suppliers. H
(400 MHz), 19F (376 MHz), and 31P (162 MHz) NMR spectra
were acquired on a Bruker Avance 400 MHz spectrometer.
NMR spectra of polymer samples were obtained from solutions
in hexafluoro-2-propanol/CD2Cl2. Polymer molecular weights
were determined using a Waters 150CV GPC system with
hexafluoro-2-propanol diluent.
2
31P{1H} NMR δ ) 15.8 (br d), -1.5 (m, J PP ) 43.5 Hz).28
(ii) Rea ction of 1 w ith 1 equ iv of B(C6F 5)3. CD2Cl2 (1
mL) was added to 1 (6.4 mg, 0.01 mmol) and B(C6F5)3 (5.1 mg,
0.01 mmol) in an NMR tube to give a colorless solution. The
formation of a small amount of white precipitate was observed
in this case. A mixture of unreacted 1 (24 mol %) and 2 (76
mol %) was obtained, as well as pentafluorobenzene.
P r oced u r e for E t h en e/P r op en e/CO P olym er iza t ion
Ca ta lyzed by 1/B(C6F 5)3. Propene (25.7 g, 0.61 mol) and a
solution of B(C6F5)3 (0.086 g, 0.17 mmol) in CH2Cl2 (100 mL)
were charged to a 300 mL autoclave. The stirred contents of
the autoclave were heated to 70 °C and pressurized to 48 bar
G with a 1:1 mixture of carbon monoxide and ethene. A
solution of [Pd(dppp)(OAc)2] (6.0 mg, 0.009 mmol) in dichlo-
romethane (10 mL) was injected into the autoclave, and the
pressure was adjusted to 50 bar G. The reaction pressure was
maintained during the polymerization by the addition of a 1:1
mixture of carbon monoxide and ethene on demand from
ballast vessels of known volume. After 3 h the polymerization
was terminated by releasing the pressure and cooling the
reaction mixture. The polymer was collected by filtration,
washed with acetone, and dried under reduced pressure. The
polymer was obtained as a white powder. Yield ) 9.30 g. The
rate of polymerization was determined by fitting of a first-
order curve to the cumulative gas uptake profile (for example,
Figure 2). NMR analysis showed the polymer to contain
(iii) Rea ction of 1 w ith 2 equ iv of B(C6F 5)3. CD2Cl2 (1
mL) was added to 1 (6.4 mg, 0.01 mmol) and B(C6F5)3 (10.2
mg, 0.02 mmol) in an NMR tube to give a yellow solution. The
formation of a small amount of gray precipitate was observed.
1H and 19F NMR data were complex and could not be fully
assigned in this case; however 31P NMR data are consistent
with a symmetric dicationic palladium species (42 mol %);
31P{1H} NMR δ ) 18.3 (s). In addition pentafluorobenzene,
complex 1, and small amounts (5 mol %) of 3b were observed.
(iv) Rea ction of 1 w ith 5 equ iv of B(C6F 5)3. CD2Cl2 (1
mL) was added to 1 (6.4 mg, 0.01 mmol) and B(C6F5)3 (25.6
mg, 0.05 mmol) in an NMR tube to give a pale yellow solution.
A mixture of 3b (92 mol %) and 4 (8 mol %) was obtained. 3b:
19F NMR δ ) -117.2 (br m, 2F, ortho-Pd-C6F5), -135.1 (m,
12F, ortho-[B(C6F5)3]2OH), -159.7 (br t, 2F, meta-Pd-C6F5),
-161.5 (m, 12F, meta-[B(C6F5)3]2OH), -165.8 (br t, 1F, para-
Pd-C6F5), -165.8 (m, 6F, para-[B(C6F5)3]2OH); 31P{1H} NMR
2
δ ) 27.0 (br d), -5.2 (m, J PP ) 34.8 Hz).28 4: 19F NMR δ )
-114.8 (br d, 4F, ortho-Pd-C6F5), -158.3 (br m, 2F, meta-
Pd-C6F5), -164.5 (br t, 1F, para-Pd-C6F5); 31P{1H} NMR δ
) 4.2 (s).
(v) R ea ct ion of P r od u ct s fr om (iv) w it h Ad d it ion a l
Wa ter . Water (0.9 µL, 0.05 mmol) was added to reaction (iv).
(25) Catalyst reduction is the principal means of deactivation of most
polyketone catalysts.1b
(26) For a leading reference see: Britovsek, G. J . P.; Gibson, V. C.;
Wass, D. F. Angew. Chem., Int. Ed. 1999, 38, 428.
(27) (a) For 1: Stephenson, T. A.; Morehouse, S. M.; Powell, A. R.;
Heffer, J . P.; Wilkinson, G. J . Chem. Soc. 1965, 3632. (b) For 5: Davies,
J . A.; Hartley, F. R.; Murray, S. G. J . Chem. Soc., Dalton Trans. 1980,
2246.
(28) P-F coupling was highly complex in this case, and J PF coupling
constants could not be determined.

1476 Organometallics, Vol. 19, No. 8, 2000
Barlow et al.
A complex mixture of products was obtained, consisting of 2
(46 mol %), 4 (24 mol %), and a characteristic signal for a
dicationic palladium species (30 mol %), 31P{1H} NMR δ )
18.3.
pale yellow solution shows characteristic peaks for a monoary-
lated complex, which accounts for over 95 mol % conversion
to this product; 31P{1H} NMR (CH2Cl2) δ ) 19.0 (dd), -2.0 (m,
P trans to -C6F5), J 2 ) 40.0 Hz.
PP
Rea ction s in “Wet” CD2Cl2: (i) Rea ction of 1 w ith 0.5
equ iv of B(C6F 5)3. CD2Cl2 (1 mL) was added to 1 (6.4 mg,
0.01 mmol) and B(C6F5)3 (2.6 mg, 0.005 mmol) in an NMR tube
to give a colorless solution. A mixture of unreacted 1 (31 mol
%) and 2 (69 mol %) was obtained.
Reaction of [P d(dppp)(NCP h )2](BF4)2 with B(p-C6H4F)3.
A solution of [Pd(dppp)(NCPh)2](BF4)2 (50 mg, 0.06 mmol) and
B(p-C6H4F)3 (20 mg, 0.07 mmol) in anhydrous CH2Cl2 (1 mL)
was stirred for 10 min. The 31P{1H} NMR spectrum of the
resulting pale yellow solution shows characteristic peaks for
a monoarylated complex, which account for 4 mol % conversion
to this product after 1 h; 31P{1H} NMR (CH2Cl2) δ ) 23 (dd),
(ii) Rea ction of 1 w ith 1 equ iv of B(C6F 5)3. CD2Cl2 (1
mL) was added to 1 (6.4 mg, 0.01 mmol) and B(C6F5)3 (5.1 mg,
0.01 mmol) in an NMR tube to give a colorless solution. The
formation of a small amount of white precipitate was observed
in this case. Again, a mixture of unreacted 1 (50 mol %) and
2 (50 mol %) was obtained, as well as pentafluorobenzene.
(iii) Rea ction of 1 w ith 2 equ iv of B(C6F 5)3. CD2Cl2 (1
mL) was added to 1 (6.4 mg, 0.01 mmol) and B(C6F5)3 (10.2
mg, 0.02 mmol) in an NMR tube to give a yellow solution. The
formation of a small amount of gray precipitate was observed.
As in the dry case, a complex mixture was obtained; however
31P NMR data are consistent with a symmetric dicationic
palladium species being the majority species (52 mol %). In
addition, pentafluorobenzene was observed.
-2 (m, P trans to -C6F5), J 2 ) 50.4 Hz
PP
Rea ction of [P d (d p p p )(NCP h )2](BF 4)2 w ith Na [B(p-
C6H4F )3OH]. A suspension of Na[B(p-C6H4F)3OH] (22 mg,
0.066 mmol) in anhydrous CH2Cl2 (1 mL) was treated with
[Pd(dppp)(NCPh)2](BF4)2 (50 mg, 0.060 mmol) and stirred for
10 min. The 31P{1H} spectrum of the resulting pale orange
solution shows characteristic peaks for a monoarylated com-
plex, which accounts for 64 mol % conversion to this product;
31P{1H} NMR (CH2Cl2) δ ) 20.6 (dd), -4.3 (m, P trans to
-C6F5), J 2 ) 50.4 Hz. 1H and 19F NMR spectra of this
PP
reaction contained a large number of overlapping peaks, which
could not be unequivocally assigned.
(iv) Rea ction of 1 w ith 5 equ iv of B(C6F 5)3. CD2Cl2 (1
mL) was added to 1 (6.4 mg, 0.01 mmol) and B(C6F5)3 (25.6
mg, 0.05 mmol) in an NMR tube to give a colorless solution. A
mixture of 3a (65 mol %), 3b (27 mol %), and 4 (7 mol %) was
obtained. 3a : 19F NMR δ ) -118.4 (m, 2F, ortho-Pd-C6F5),),
-135.1 (m, 6F, ortho-B(C6F5)3OH), -160.2 (br t, 1F, para-Pd-
C6F5), -160.6 (m, 6F, meta-B(C6F5)3OH), -162.8 (br t, 2F,
meta-Pd-C6F5), -166.6 (m, 3F, para-B(C6F5)3OH); 31P{1H}
NMR δ ) 21.8 (dt, J PF ) 7.8 Hz), -3.8 (m, J PP ) 41.0 Hz).
(v) Rea ction of P r od u cts fr om (iv) w ith a n Ad d ition a l
5 equ iv of B(C6F 5)3. B(C6F5)3 (25.6 mg, 0.05 mmol) was added
to reaction (iv). Clean conversion of 3a to 3b was observed (91
mol %) as well as small remaining amounts of 4 (9 mol %).
La r ger Sca le P r ep a r a tion of P d (d p p p )(C6F 5)-
[B(C6F 5)3OH] (3a ). [Pd(dppp)(OAc)2] (0.2 g, 0.32 mmol) was
added to a solution of B(C6F5)3 (1.0 g, 2.0 mmol) in CH2Cl2 (20
mL). The reaction was stirred vigorously at room temperature
for 45 min and then concentrated under reduced pressure to
10 mL. Addition of hexane gave a yellow/orange oil, which was
decanted, washed with further portions of hexane (3 × 10 mL),
and dried in vacuo to give an orange solid residue. NMR
spectroscopy data of this sample were consistent with the
mixture of reaction products of the analogous NMR scale
reaction described above. The impure nature of this compound
precluded elemental analysis.
Eth en e/P r op en e/CO P olym er iza tion Ca ta lyzed by 3a /
B(C6F 5)3. This was performed as detailed for 1/B(C6F5)3 using
propene (23.4 g, 0.61 mol), CO/C2H4 (1:1, 50 bar G), B(C6F5)3
(0.102 g, 0.20 mmol), and 3a (0.017 g, 0.01 mmol, prepared
according to a larger scale procedure described above) at 70
°C for 3 h. The polymer was obtained as a white powder. Yield
) 9.8 g. NMR analysis showed the polymer to contain
alternating units originating in CO and alkene, with 5.0 per
100 monomers originating in propene. GPC analysis (hexafluoro-
2-propanol, relative to a poly(methyl methacrylate) stan-
dard): Mn ) 60 000; Mw ) 116 000; Mw/Mn ) 1.9; Mpk ) 91 000.
Eth en e/P r op en e/CO P olym er iza tion Ca ta lyzed by 3a .
This reaction was carried out in the absence of additional
B(C6F5)3; otherwise the same procedure was performed as
detailed for 1/B(C6F5)3 using propene (25.7 g, 0.61 mol), CO/
C2H4 (1:1, 50 bar G), and 3a (0.006 g, 0.009 mmol, prepared
according to a larger scale procedure described above) at 70
°C for 3 h. The polymer was obtained as a white powder. Yield
) 0.8 g. NMR analysis showed the polymer to contain
alternating units originating in CO and alkene, with 4.3 per
100 monomers originating in propene. GPC analysis (hexafluoro-
3
2
2-propanol, relative to polystyrene): Mn ) 63 000; Mw
144 000; Mw/Mn ) 2.3.
)
Ack n ow led gm en t. Dr. Stephen Dossett (BP Amoco
Chemicals) and Dr. Paul Pringle (University of Bristol)
are thanked for useful discussions.
Rea ction of [P d (d p p p )(NCP h )2](BF 4)2 w ith B(C6F 5)3. A
solution of [Pd(dppp)(NCPh)2](BF4)2 (50 mg, 0.06 mmol) and
B(C6F5)3 (35 mg, 0.07 mmol) in anhydrous CH2Cl2 (1 mL) was
stirred for 10 min. The 31P{1H} NMR spectrum of the resulting
OM990911L