Alcoholysis of acylpalladium(II) complexes relevant to the alternating copolymerization of ethene and carbon monoxide and the alkoxycarbonylation of alkenes: The importance of cis-coordinating phosphines

Published on Web 04/15/2003

Alcoholysis of Acylpalladium(II) Complexes Relevant to the

Alternating Copolymerization of Ethene and Carbon Monoxide

and the Alkoxycarbonylation of Alkenes: the Importance of

Cis-Coordinating Phosphines

Piet W. N. M. van Leeuwen,*^,†Martin A. Zuideveld,^†Bert H. G. Swennenhuis,^†

Zoraida Freixa,^†Paul C. J. Kamer,^†Kees Goubitz,^†Jan Fraanje,^†Martin Lutz,^‡and

Anthony L. Spek^‡

Contribution from the Institute of Molecular Chemistry, UniVersity of Amsterdam, Nieuwe

Achtergracht 166, 1018 WV, Amsterdam, The Netherlands and Department of Crystal and

Structural Chemistry, UniVersity of Utrecht, Padualaan 8, 3584 CH, Utrecht, The Netherlands

Received November 14, 2002; E-mail: pwnm@science.uva.nl

Abstract: The mechanism and kinetics of the solvolysis of complexes of the type [(L-L)Pd(C(O)-

CH₃)(S)]⁺[CF₃SO₃]^-(L-L ) diphosphine ligand, S ) solvent, CO, or donor atom in the ligand backbone)

was studied by NMR and UV-vis spectroscopy with the use of the ligands a-j: SPANphos (a), dtbpf (b),

Xantphos (c), dippf (d), DPEphos (e), dtbpx (f), dppf (g), dppp (h), calix-6-diphosphite (j). Acetyl palladium

complexes containing trans-coordinating ligands that resist cis coordination (SPANphos, dtbpf) showed no

methanolysis. Trans complexes that can undergo isomerization to the cis analogue (Xantphos, dippf,

DPEphos) showed methanolyis of the acyl group at a moderate rate. The reaction of [trans-(DPEphos)-

Pd(C(O)CH₃)]⁺[CF₃SO₃]^-(2e) with methanol shows a large negative entropy of activation. Cis complexes

underwent competing decarbonylation and methanolysis with the exception of 2j, [cis-(calix-diphosphite)-

Pd(C(O)CH₃)(CD₃OD)]⁺[CF₃SO₃]^-. The calix-6-diphosphite complex showed a large positive entropy of

activation. It is concluded that ester elimination from acylpalladium complexes with alcohols requires cis

geometry of the acyl group and coordinating alcohol. The reductive elimination of methyl acetate is described

as a migratory elimination or a 1,2-shift of the alkoxy group from palladium to the acyl carbon atom. Cis

complexes with bulky ligands such as dtbpx undergo an extremely fast methanolysis. An increasing steric

bulk of the ligand favors the formation of methyl propanoate relative to the insertion of ethene leading to

formation of oligomers or polymers in the catalytic reaction of ethene, carbon monoxide, and methanol.

how.²Chain initiation does not have to be left at chance either,

as methylpalladium complexes were found to be effective

Introduction

Since the introduction of well-defined and fast palladium

catalysts for copolymerization of ethene and carbon monoxide,

characteristically containing weakly coordinating anions and

bidentate ligands,^1-3chain initiation and termination reactions,

as compared to insertion reactions, have received relatively little

attention, although their importance is well recognized. Alcohols

are suitable solvents for the copolymerization, and not only do

they initiate the reaction, but they also function as chain transfer

agents. As a result, the chain ends contain a hydrogen atom (as

an ethyl group) and a methoxy group (as a methyl ester group).

In alcohols, several routes are available to form the two initiating

palladium species, hydrides, and carbomethoxy species, and the

conditions applied determine which species will be formed and

initiators for the polymerization reaction.^1,4The common chain

termination reactions comprise â-hydrogen elimination, nucleo-

philic attack at an acylpalladium complex, and protonation of

an alkylpalladium complex (see Scheme 1).

In the simplest case, each route leads to polymer chains or

oligomers containing one ester (E) and one ketone (K) end

group, KE polymers. When reactions 2 and 3 occur at

comparable rates, which accidentally seems to be the case in

several catalyst systems, there is formation of EE and KK

polymers in addition to KE polymers. This is so because the

growing polymer chain does not “remember” whether it started

from a hydride or a methoxy group, and thus, a statistical

distribution of chain ends within one molecule will be obtained

(EE:KE:KK ) 1:2:1).³Deviations from an average K:E ) 1:1

have been observed, because the intermediate palladium species

Pd²⁺or PdH⁺can be reduced or oxidized, respectively, to one

^†University of Amsterdam.

^‡University of Utrecht.

(1) (a) Sen, A.; Lai, T.-W. J. Am. Chem. Soc. 1982, 104, 3520. (b) Lai, T.-W.;

Sen, A. Organometallics 1984, 3, 866.

(2) Drent, E.; Budzelaar, P. H. M. Chem. ReV. 1996, 96, 663.

(3) (a) Drent, E. Eur. Pat. Appl. 121,965, 1984; Chem. Abstr. 1985, 102,

464223. (b) Drent, E.; van Broekhoven, J. A. M.; Doyle, M. J. J.

Organomet. Chem. 1991, 417, 235. (c) Mul, W. P.; Drent, E.; Jansens, P.

J.; Kramer, A. H.; Sonnemans, M. H. W. J. Am. Chem. Soc. 2001, 123,

5350.

(4) (a) Shultz, C. S.; Ledfort, J.; DeSimone, J. M.; Brookhart, M. J. Am. Chem.

Soc. 2000, 122, 6351. (b) Zudin, V. N.; Chinakov, V. D.; Nekipelov, V.

M.; Rokov, V. A.; Likholobov, Y. I.; Yermakov, Y. I. J. Mol. Catal. 1989,

52, 27. (c) Koide, Y.; Bott, S. G.; Barron, A. R. Organometallics 1996,

15, 2213.

9

J. AM. CHEM. SOC. 2003, 125, 5523-5539

5523

A R T I C L E S

Scheme 1

van Leeuwen et al.

Scheme 2 ^a

^aNumbers denote references.

As in alkene polymerizations, dihydrogen can also be used

as a chain transfer agent in polyketone formation. The reaction

is depicted as a hydrogenolysis (reaction 6 in Scheme 1), but it

can also be described as a sequence involving heterolytic

cleavage of dihydrogen (eqs 7, 8). As a result, only KK polymers

(or oligomers) will be obtained.

Water and carbon monoxide can play the same role as

hydrogen, as has been shown in the synthesis of diethyl ketone

(pentan-3-one).^4bEther chain ends have been observed in

polymerization reactions,⁸and low molecular weight products

can also contain an ether moiety at one end and an ester group

at the other end. Most likely, ether chain ends are formed by a

palladium-catalyzed attack of alcohols at enones resulting from

â-hydride elimination; the catalyzed Michael reaction has been

observed in separate experiments.⁹

When chain transfer is very fast, the reaction observed is the

alkoxycarbonylation of ethene, which is nothing but a perfect

chain transfer after the insertion of just two monomers. In recent

years, several very fast catalysts for this reaction have been

reported (Scheme 2, showing some of the ligands).^10-24,41

Catalyst systems and conditions for the two reactions are very

another leading to an excess of K or E chain ends or even

one type of chain end only.²Chain termination ratios also

depend on the phase in which the chain end resides, liquid phase

or solid phase, as higher oligomers precipitate.^3cFormally,

eq 2 involves a protonation of the alkyl chain, which is not a

very fast reaction at the cationic palladium center, as is also

apparent from the stability of methylpalladium cations.⁵At

room temperature, the half-life time of [(dppp)Pd(CH₂C(CH₃)₃)-

(CH₃CN)]⁺[CF₃SO₃]^-in CD₃OD with a 10-fold excess of CF₃-

COOH is 1 h;^6aand methyl platinum complexes containing

electron-withdrawing diphosphines survive strong acids for

several days.^6b,cFor a number of model compounds, it has been

shown^4b,7that reaction of compounds containing a γ-keto group

such as shown in reactions 4 and 5 with CH₃OD or D₂O form

a ketone having the deuterium atom in the â-position (relative

to palladium) instead of the expected R-position (relative to

palladium), which has been explained by an enolate mechanism

(Scheme 1, eqs 4-5).

(8) Jiang, Z.; Sen, A. Macromolecules 1994, 27, 7215.

(9) Miller, K. J.; Kitagawa, T. T.; Abu-Omar, M. M. Organometallics 2001,

20, 4403.

(10) (a) Clegg, W.; Eastham, G. R.; Elsegood, M. R. J.; Heaton, B. T.; Iggo, J.

A.; Tooze, R. P.; Whyman, R.; Zacchini, S. Organometallics 2002, 21,

1832. (b) Knight, J. G.; Doherty, S.; Harriman, A.; Robins, E. G.; Betham,

M.; Eastham, G. R.; Tooze, R. P.; Elsegood, M. R. J.; Champkin, P. A.;

Clegg, W. Organometallics 2000, 19, 4957. (c) Clegg, W.; Eastham, G.

R.; Elsegood, M. R. J.; Tooze, R. P.; Wang, X. L.; Whiston, K. Chem.

Commun. 1999, 1877.

(11) Papadogianakis, G.; Verspui, G.; Maat, L.; Sheldon, R. A. Catal. Lett. 1997,

47, 43.

(12) (a) Tilloy, S.; Monflier, E.; Bertoux, F.; Castanet, Y.; Mortreux, A. New J.

Chem. 1997, 21, 529. (b) Bertoux, F.; Monflier, E.; Castanet, Y.; Mortreux,

A. J. Mol. Catal. 1999, 143, 11.

(13) Schreuder Goedheijt, M.; Reek, J. N. H.; Kamer, P. C. J.; van Leeuwen,

P. W. N. M. Chem. Commun. 1998, 2431.

(14) Verspui, G.; Moiseev, I. I.; Sheldon, R. A. J. Organomet. Chem. 1999,

586, 196.

(15) Lin, Y.-S.; Yamamoto, A. Organometallics 1998, 17, 3466.

(16) Noskov, Yu. G.; Petrov, E. S. Kinet. Catal. 1997, 38, 568.

(17) (a) Ali, B. E.; Alper, H. J. Mol. Catal. 1992, 77, 7. (b) Ali, B.; Alper, H.

J. Mol. Catal. 1993, 80, 377.

(18) Del R´ıo, I.; Ruiz, N.; Claver, C.; van der Veen, L. A.; van Leeuwen, P. W.

N. M. J. Mol Catal. A: Chem. 2000, 161, 39.

(19) Jayasree, S.; Seayad, A.; Chaudhari, R. V. Chem. Commun. 2000, 1239.

(20) Elango, V.; Murphy, M. A.; Smith, B. L.; Davenport, K. G.; Mott, G. N.;

Moss, G. L. Eur. Pat. Appl. EP 284,310, 1988 (to Hoechst Celanese Corp.,

USA); Chem. Abstr. 1989, 110, 153916.

(21) (a) Drent, E.; Kragtwijk, E. Eur. Pat. Appl. EP 495,548, 1992 (to Shell).

(b) Pugh, R. I.; Drent, E. AdV. Synth. Catal. 2002, 344, 815.

(22) Eastham, G. R.; Tooze, R. P.; Wang, X. L.; Whiston, K. World Pat. 96/

19434, 1996 (to ICI).

(5) (a) Brookhart, M.; Rix, F. C.; DeSimone, J. M.; Barborak, J. C. J. Am.

Chem. Soc. 1992, 114, 5894. (b) Rix, F. C.; Brookhart, M. J. Am. Chem.

Soc. 1995, 117, 1137. (c) Schultz, C. S.; Ledford, J.; DeSimone, J. M.;

Brookhart, M. J. Am. Chem. Soc. 2000, 122, 6351.

(6) (a) Zuideveld, M. A.; Kamer, P. C. J.; van Leeuwen, P. W. N. M. To be

published. (b) Houlis, J. F.; Roddick, D. M. J. Am. Chem. Soc. 1998, 120,

11020. (c) Bennett, B. L.; Hoerter, J. M.; Houlis, J. F.; Roddick, D. M.

Organometallics 2000, 19, 615.

(7) Zuideveld, M. A.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Klusener,

P. A. A.; Stil, H. A.; Roobeek, C. F. J. Am. Chem. Soc. 1998, 120, 7977.

(23) Pugh, R. I.; Drent, E.; Pringle, P. G. Chem. Commun. 2001, 1476.

(24) Eastham, G. R.; Tooze, R. P.; Kilner, M.; Foster, D. F.; Cole-Hamilton,

D. J. J. Chem. Soc., Dalton Trans. 2002, 1613.

9

5524 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Alcoholysis of Acylpalladium(II) Complexes

A R T I C L E S

similar. The initial observations led to a rule of thumb that fast

polymerization catalysts required bidentate ligands with a 1,3-

propanediyl bridge and that carbomethoxylation was favored

using monodentate ligands.²Results of recent years have shown

that substituents on phosphorus that are more bulky than phenyl

groups lead to a high molecular weight polymer even when the

bridge comprises one carbon atom or one nitrogen atom

only.^25,26However, the most active catalyst for methyl pro-

panoate formation known to date contains bidentate cis

ligands,^10,21,23although one of these bulky ligands easily leads

to the formation of trans-coordinated, oligomeric diphosphine

complexes.^10cFormation of dimeric trans complexes had been

observed before for diphosphines containing a four-membered

carbon bridge.²⁷Since monophosphines also form trans

complexes^{2,10b-c,28,29}the latter were held responsible for the

alkoxycarbonylation reaction. For the bidentate ligand systems

mentioned, the possibility remains that one phosphine moiety

dissociates from the palladium center.

Sen et al have shown that changing the reaction medium of

the copolymerization reaction from methanol to higher alcohols

(ethanol, tert-butyl alcohol) increased the molecular weight of

the copolymer produced.^1bMilani has shown that the molecular

weight in styrene-CO copolymerization increases when 2,2,2-

trifluoroethanol is used instead of methanol.³³Thus, the alcohol

has a distinct effect on the rate of the termination reaction, at

least relative to the rate of propagation.

To´th and Elsevier showed that the reaction of an acetylpal-

ladium complex and sodium methoxide/methanol is very fast

and occurs already at low temperature³⁴to give methyl acetate

and a palladium(I) hydride dimer.^30a,35,36The reactivity of

cationic acylpalladium complexes toward alcohols under neutral

and acidic conditions has not been studied in great detail so

far.

In the present study, we have looked at the stoichiometric

reaction of several acetylpalladium species containing a variety

of cis and trans diphosphines with alcohols. We have studied a

series of ligands ranging from ones that exclusively form trans

complexes, via those that form both trans and cis complexes,

to those that form cis complexes only, having different electronic

properties. From previous literature, the impression^2,28,29had

grown that trans bis-phosphine complexes are responsible for

methoxycarbonylation, while cis coordination of bidentate

ligands leads to polymer formation via multiple insertions. The

pivotal questions to be answered are what are the steric and

electronic ligand requirements for the control of alkene/CO

insertion versus termination of the cycle by alcoholysis of the

palladium acyl species.

The mechanism of the alkoxycarbonylation reaction catalyzed

by palladium has been the topic of research for many years,

and it seemed fairly well understood at the end of the eighties.³⁰

Stepwise reactions had shown the feasibility of mechanistic

pathways, but kinetic studies and in situ observations on catalytic

systems were lacking. Besides, since the development of the

fast catalysts for polyketone formation, it became evident that

a lot of improvement could be made in alkoxycarbonylation.

The first publications on fast catalysts for methoxycarbonylation

originate from Drent.²¹The resting state of the new catalysts

may well be an acyl complex,^28-31while the attack of alcohol

at the acylpalladium complex is considered to be the rate-

determining step. It is probably more precise to say that fast

preequilibria exist between the acyl complex and other com-

plexes en route to it and that the highest barrier is formed by

the reaction of alcohol and acylpalladium complex. There is

general agreement that the catalytic cycle starts with palladium

hydride species, except for methyl methacrylate catalysis.³²

Results

Ligands Used in this Study. Scheme 3 depicts the ligands

a-j that have been used in this study arranged in the order of

their propensity to form trans complexes. It represents only a

small selection of the many ligands that have been used for the

catalytic studies in the literature, but they are representative, as

we will show. Their methylpalladium triflate complexes are

numbered 1, and their acetylpalladium complexes are designated

2.

Synthesis, Structure, and Reactivity of trans-Coordinated

Acetylpalladium Complexes. Acetylpalladium complexes were

synthesized by bubbling carbon monoxide through a dichloro-

methane solution of the corresponding methylpalladium precur-

sor at room temperature. Ligands that coordinate in a purely

trans fashion are rare, and the “trans” coordinating ligands

reported often also form cis complexes as a result of their

flexibility.³⁷We have recently reported on a new trans-spanning

ligand, SPANphos (a),³⁸characterized by a large phosphorus-

to-phosphorus distance and forming trans complexes with

platinum and palladium. When [trans-(SPANphos)PdCH₃(CH₃-

CN)]⁺[CF₃SO₃]^-(1a-CH₃CN) was dissolved in CD₂Cl₂and CO

was bubbled through for 15 min at room temperature, the acetyl

complex, 2a, was formed.

(25) Dossett, S. J. PCT Int. Appl. WO 9737765, 1997 (to BP Chemicals); Chem.

Abstr. 1997, 127, 319387.

(26) Dossett, S. J.; Gillon, A.; Orpen, A. G.; Fleming, J. S.; Pringle, P. G.;

Wass, D. F.; Jones, M. D. Chem. Commun. 2001, 699.

(27) (a) Dekker, G. P. C. M.; Elsevier, C. J.; Vrieze, K.; van Leeuwen, P. W.

N. M. Organometallics 1992, 11, 1598. (b) Sanger, A. R. J. Chem. Soc.,

Dalton Trans. 1979, 1971. (c) Scrivanti, A.; Botteghi, C.; Toniolo, L.;

Berton, A. J. Organomet. Chem. 1988, 334, 261.

(28) (a) Verspui, G.; Papadogianakis, G.; Sheldon, R. A. J. Chem. Soc., Chem.

Commun. 1998, 401. (b) Verspui, G.; Schanssema, F.; Sheldon, R. A. Appl.

Catal. 2000, 198, 5. (c) Verspui, G.; Schanssema, F.; Sheldon, R. A. Angew.

Chem., Int. Ed. 2000, 39, 804.

(29) Del Rio, I.; Claver, C.; Van Leeuwen, P. W. N. M. Eur. J. Inorg. Chem.

2001, 2719.

(30) (a) Portnoy, M.; Frolow, F.; Milstein, D. Organometallics 1991, 10, 3960.

(b) Milstein, D. Acc. Chem. Res. 1988, 21, 428. (c) Milstein, D.; Stille, J.

K. J. Am. Chem. Soc. 1979, 101, 4981. (d) Fenton, D. M. J. Org. Chem.

1973, 38, 3192. (e) Knifton, J. J. Org. Chem. 1976, 41, 2885. (f) Tsuji, J.

Organic Synthesis with Palladium Compounds; Springer-Verlag: New

York, 1980. (g) Heck, R. F. Palladium Reagents in Organic Synthesis;

Academic Press: New York, 1985.

(31) (a) Noskov, Y. G.; Petrov, E. S. Kinet. Catal. 1993, 34, 902. (b) Noskov,

Y. G.; Petrov, E. S. Kinet. Catal. 1994, 35, 672. (c) Noskov, Y. G.; Petrov,

E. S. Kinet. Catal. 1997, 38, 520. (d) Kron, T. E.; Noskov, Y. G.;

Terekhova, M. I.; Petrov, E. S. Zh. Fiz. Khim. 1996, 70, 82. (e) Tel′Naya,

Y. V.; Noskov, Y. G.; Petrov, E. S. Zh. Obshch. Khim. 1995, 65, 1391. (f)

Terekhova, M. I.; Kron, T. Y.; Noskov, Y. G.; Petrov, E. S. Petrol. Chem.

1996, 36, 331.

(33) Milani, B.; Anzilutti, A.; Vicentini, L.; Sessanta o Santini, A.; Zangrando,

E.; Geremia, S.; Mestroni, G. Organometallics 1997, 16, 5064.

(34) To´th, I.; Elsevier, C. J. J. Chem. Soc., Chem. Commun. 1993, 529.

(35) Portnoy, M.; Milstein, D. Organometallics 1994, 13, 600.

(36) Zudin, V. N.; Chinakov, V. D.; Nekipelov, V. M.; Likholobov, V. A.;

Yermakov, Y. I. J. Organomet. Chem. 1985, 289, 425.

(37) Bessel, C. A.; Aggarwal, P.; Maerschilok, A. C.; Takeuchi, K. J. Chem.

ReV. 2001, 101, 1031.

(38) Freixa, Z.; Beentjes, M. A.; Batema, G. D.; Dieleman, C. B.; van Strijdonck,

G. P. F.; Reek, J. N. H.; Kamer, P. C. J.; van Leeuwen, P. W. N. M. Angew.

Chem., Int. Ed. 2003, 42, 1284.

(32) (a) Drent, E.; Arnoldy, P.; Budzelaar, P. H. M. J. Organomet. Chem. 1993,

455, 247. (b) Kron, T. E.; Terekhova, M. I.; Noskov, Yu. G.; Petrov, E. S.

Kinet. Catal. 2001, 42, 182. (c) Drent, E.; Arnoldy, P.; Budzelaar, P. H.

M. J. Organomet. Chem. 1994, 475, 57. (d) Keijsper, J. J.; Arnoldy, P.;

Doyle, M. J.; Drent, E. Recl. TraV. Chim. Pays-Bas 1996, 115, 248. (e)

Scrivanti, A.; Beghetto, V.; Campagna, E.; Zanato M.; Matteoli, U.

Organometallics 1998, 17, 630.

9

J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5525

A R T I C L E S

Scheme 3

van Leeuwen et al.

Table 1. Selected NMR Data of Complexes 1a-e, 2a-e, 2g, 3d,

and 3e^a

complex

¹H NMR

³¹P NMR

Pd-CH₃

1a-CH₃CN

1b

1c

0.14 t ³J_PH) 6.3

1.70 t ³J_PH) 4.8

1.50 t ³J_PH) 5.7

0.73 br.d ³J_PH) 4.8

22.0 s

29.3 s

20.4 s

1d-CH₃CN

54.1 d ²J_PP) 21.3

32.7 d ²J_PP) 21.3

31.8 d ²J_PP) 31.7

7.7 d ²J_PP) 31.7

1e-CH₃CN

0.88 dd ³J_PH) 6.6, 2.9

Pd-C(O)CH₃

1.75 s

2a

2b

2c

2d

9.0 s

30.2 s

10.2 s

18.7 s

2.91 t ⁴J_PH) 1.5

2.20 t ⁴J_PH) 1.5

2.83 s

2e^b

2g-CO^c

2.38 t ⁴J_PH) 1.5

1.87 s

7.0 s

24.1 d ²J_PP) 62

12.1 d ²J_PP) 62

31.6 dd ²J_PP) 43.8, 228

10.8 dd ²J_PP) 43.8, 228

17.2 t ²J_PP) 43.8

26.5 dd ²J_PP) 46.0, 241

10.6 dd ²J_PP) 46.0, 241

-0.9 t ²J_PP) 46.0

3d^d

2.35 s

1.79 s

3e^e

^aCD₂Cl₂, 25 °C, δ (ppm), J (Hz). ^bT ) -90 °C. ^{c 13}C{¹H} NMR:

C(O)CH₃, d, 229.5 ppm, ²J_PC) 82 Hz; Pd(CO), dd 175.7 ppm, ²J_PC) 82

2

Hz and J_PC) 19 Hz. ^dT ) -30 °C. ¹³CO ³¹P{¹H} NMR: 10.8 ppm,

2

d²J_PC) 81.3 Hz. ¹³C{¹H} NMR: 230.0 ppm, d J_PC) 81.3 ppm. ^eT )

-90 °C. ¹³CO ³¹P{¹H} NMR: -0.9 ppm, d ²J_PC) 82.3 Hz. ¹³C{¹H} NMR:

Complex [trans-η³-(dtbpf)PdC(O)CH₃]⁺[CF₃SO₃]^-(2b) was

synthesized from [trans-η³-(dtbpf)PdCH₃]⁺[CF₃SO₃]⁺(1b).³⁹

The insertion of CO was slow (t_1/2≈ 15 min) at room tempera-

ture compared to the time for CO insertion into the ionic methyl-

palladium complexes containing Xantphos c and DPEphos e

(t_1/2< 1 min). The dtbpf ligand remained an η³terdentate P,Fe,P

ligand in complex 2b as shown by the large chemical shift

difference of the R and â-hydrogen atoms of the Cp-rings

(∆δ ) 0.82 ppm).⁴⁰

2

228.7 ppm, d J_PC) 82.3 ppm.

CH₂Cl₂. Conductivity measurements in CH₂Cl₂showed that

2e was ionic (Λ_m) 56 S mol^-1L^-1). The use of ¹³CO did

not result in additional P-C couplings in the ³¹P{¹H} NMR

spectrum. These data suggest that the phosphorus atoms are

coordinated in a trans fashion and both cis to the acetyl

group. At room temperature, the compound was unstable

toward decarbonylation, even in the solid state. The acetyl-

palladium complex (2e′) was synthesized using complex 1e′

containing the BARF anion ([B(p-C₆H₃(CF₃)₂)₄]^-, which

proved to be thermally more stable (i.e., less prone to deinsertion

of CO).

[trans-η³-(Xantphos)PdC(O)CH₃]⁺[CF₃SO₃]^-(2c) was syn-

thesized from [trans-η³-(Xantphos)PdCH₃]⁺[CF₃SO₃]^-(1c), and

its structure is trans, η³-P,O,P according to NMR spectroscopy.

Complex [trans-η³-(dippf)PdC(O)CH₃]⁺[CF₃SO₃]^-(2d) was

synthesized from [cis-(dippf)PdCH₃(CH₃CN)]⁺[CF₃SO₃]^-(1d-

CH₃CN). The CO insertion for 2d is faster than that for 2b, the

dtbpf complex (t_1/2≈ 5 min). In the methylpalladium complex

1d-CH₃CN, dippf acts as a cis bidentate ligand, but, in the

acetylpalladium complex, it coordinates as an η³-terdentate

P,Fe,P ligand as indicated by the ³¹P and ¹³C NMR spectra

(Table 1) and by the large chemical shift difference of the R-

and â-hydrogen atoms of the Cp-rings (∆δ ) 0.72 ppm). The

addition of an excess of triphenylphosphine to a solution of 2d

in CD₂Cl₂led to the formation of 3d (Scheme 3). This

demonstrates that the dippf ligand can coordinate in both a trans

fashion and a cis fashion and that the Pd-Fe dative bond can

be broken by the addition of a strongly coordinating ligand.

When [cis-(DPEphos)PdCH₃(CH₃CN)]⁺[CF₃SO₃]^-(1e-CH₃-

CN) was dissolved in CD₂Cl₂and CO was bubbled through,

the acetyl complex, trans-2e, was formed (Scheme 4). The

addition of silver triflate to a solution of cis-(DPEphos)Pd(C(O)-

CH₃)Cl in CD₂Cl₂also yielded trans-2e. The complex was

isolated by the addition of diethyl ether to a solution of 2e in

The DPEphos ligand was found to coordinate in an η³(planar

P,O,P) fashion in complex 2e even in the presence of additional

ligands such as CO and acetonitrile. The capacity of the

DPEphos ligand to coordinate in an η²P,P fashion to palladium

was demonstrated by the addition of triphenylphosphine to a

solution of 2e in CD₂Cl₂(Scheme 4). The NMR data (Table 1)

are in accordance with the structure [cis-(DPEphos)Pd(C(O)-

CH₃)(PPh₃)]⁺[CF₃SO₃]^-(3e, Scheme 4).The reaction of 2e

and methanol in a CH₂Cl₂solution under 1 bar of CO resulted

in the formation of a palladium(I) hydride dimer, [Pd₂(µ-H)-

(µ-CO){(DPEphos)}₂]⁺[CF₃SO₃]^-(4e), and methyl acetate

(Scheme 5). The IR spectrum in CH₂Cl₂showed a CO stretching

frequency at 1847 cm^-1

.

X-ray Crystal Structure Analysis of 2e′. The anion in 2e′

was highly disordered due to the rotation of the trifluoromethyl

groups, but the structure of the cation could be determined

accurately. The metal center has a square planar geometry with

all donor atoms and the metal center in one plane. The

phosphorus atoms are coordinated in a trans fashion with a

P1-Pd-P2 angle of 154.70(8)° (Τable 3). The oxygen atom

of the ligand backbone is coordinated to palladium (2.261(6)

Å, Figure 1), and there are no close contacts between the anion

and the palladium atom.

(39) Zuideveld, M. A.; Swennenhuis, B. H. G.; Kamer, P. C. J.; van Leeuwen,

P. W. N. M. J. Organomet. Chem. 2001, 637, 805.

(40) Sato, M.; Shigeta, H.; Sekino, M.; Akabori, S. J. Organomet. Chem. 1993,

458, 199.

9

5526 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Alcoholysis of Acylpalladium(II) Complexes

A R T I C L E S

Scheme 4

Scheme 5

Scheme 6

Table 3. Selected Bond Distances (Å) and Angles (deg) for

Non-Hydrogen Atoms of 2e′

Pd-P1

2.315(2)

2.261(6)

1.521(19)

154.70(8)

80.05(18)

100.9(3)

113.2(9)

Pd-P2

2.313(2)

1.936(10)

1.176(13)

78.77(15)

99.6(3)

Pd-O1

Pd-C1

C1-C2

C1-O2

P1-Pd-P2

P1-Pd-O1

P1-Pd-C1

Pd-C1-C2

P2-Pd-O1

P2-Pd-C1

Pd-C1-O2

123.6(8)

Table 2. Selected NMR Data of Complexes 4c-g

complex

¹H NMR

µ-H

³¹P NMR

¹³C NMR

µ-CO

4c

-8.58 q ²J_PH) 42

-8.58 q ²J_PH) 42,

-7.12 q ²J_PH) 39

-6.90 q ²J_PH) 42

5.5 s

217.7 q ²J_PC) 36

4c^a

4d

4e

5.5 d ²J_PC) 36 217.7 q ²J_PC) 36

35.3 s

10.8 s

233.7 q ²J_PC) 33

225.1 q ²J_PC) 33

4e^b

4g

4g^b

-6.90 qd ²J_PH) 42, ²J_CH) 2.7 10.8 d ²J_PC) 33 225.1 q ²J_PC) 33

-6.10 q ²J_PH) 42

20.3 s

232.0 q ²J_PC) 34

-6.10 qd ²J_PH) 42, ²J_CH) 4.5 20.3 d ²J_PC) 34 232.0 q ²J_PC) 34

^aCD₂Cl₂, 25 °C, δ (ppm), J (Hz). ^{b 13}CO.

palladium complexes were studied in methanol in the presence

and in the absence of CO. When 10 equiv of methanol were

added to [(dppf)PdC(O)CH₃(CH₃CN)]⁺[CF₃SO₃]^-(2g-CH₃CN)

in CH₂Cl₂in the presence of 1 bar of CO at room temperature,^27a

the palladium(I) hydride dimer, [Pd₂(µ-H)(µ-CO){(dppf)}₂]⁺-

[CF₃SO₃]^-(4g), and methyl acetate were formed (for NMR

data, see Table 2). The acetonitrile-free acetylpalladium

complex was synthesized in situ. When (dppf)Pd(CH₃)₂was

reacted with 1 equiv of triflic acid in CD₃OD at room tem-

perature, [(dppf)PdCH₃(CD₃OD)]⁺[CF₃SO₃]^-(1g-CD₃OD) was

formed. When CO was bubbled through this solution at -90

°C, this resulted in the formation of a mixture of [(dppf)Pd-

(C(O)CH₃)(CO)]⁺[CF₃SO₃]^-(2g-CO) and [(dppf)Pd(C(O)CH₃)-

(CD₃OD)]⁺[CF₃SO₃]^-(2g-CD₃OD) in a 1:2.5 ratio. When CO

was led through a solution of 1g-CD₃OD at -70 °C, 2g-CO

was formed in high yield (>85%, according to the ³¹P{¹H}

NMR spectrum). The structure of the complex was evidenced

by ¹³CO labeling. The acetylpalladium complex 2g-CD₃OD

containing CD₃OD as the fourth ligand was the minor product

(<15%). When the temperature was raised to room temperature,

an instantaneous reaction of 2g-CO and 2g-CD₃OD with CD₃-

OD was observed. The color changed from orange to dark-red

to give [Pd₂(µ-D)(µ-CO){(dppf)}₂]⁺[CF₃SO₃]^-and methyl

Figure 1. X-ray crystal structure of 2e′. The ellipsoids are drawn at the

50% probability level. Hydrogen atoms and anion ((3,5-(CF₃)₂-C₆H₃)₄)B)

have been omitted for clarity.

Synthesis, Structure, and Reactivity of Cis-Coordinated

Acetylpalladium Complexes. Three methods were used to

generate ionic cis-coordinated acetylpalladium complexes (meth-

ods A, B, and C; Scheme 6). In method A, carbon monoxide is

bubbled through a solution of an ionic methylpalladium

complex. The product is the ionic, solvent-coordinated acetylpal-

ladium complex.^27aAlternatively, the ionic methylpalladium

complex is synthesized in situ through the reaction of an acid

with a dimethylpalladium complex⁵and bubbling CO into the

solution (method B). In method C, the complex is synthesized

from the neutral methylpalladium compound by abstraction of

the chloride anion with silver triflate in a coordinating solvent

(method C).^5c,27a

The reactivity and stability of cis-chelated acetylpalladium

dppf complexes were studied. Cis-chelated acetylpalladium

complexes have already been shown to decarbonylate very easily

in the absence of CO.^27aTherefore, the cis-chelated acetyl-

9

J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5527

A R T I C L E S

van Leeuwen et al.

Table 4. Selected Bond Distances (Å) and Angles (deg) for

Non-Hydrogen Atoms of 5f-Me

Pd1-P1

2.2717(5)

2.4003(7)

2.069(2)

87.03(7)

78.19(16)

89.74(15)

Pd1-P2

2.4535(5)

2.431(5)

104.140(17)

80.53(7)

Pd1-Cl1^a

Pd1-Cl1A^b

P1-Pd1-P2

C1-Pd1-Cl1^a

Cl1^a-Pd1-P2

Pd1-C1

P1-Pd1-C1

C1-Pd1-Cl1A^b

Cl1A^b-Pd1-P2

90.84(2)

^aMajor disorder component. ^bMinor disorder component.

Scheme 7

Figure 2. X-ray crystal structure of 5f-Me. The ellipsoids are drawn at

the 50% probability level. The chlorine ligand is disordered over two

positions. Only the major disorder component is shown. Hydrogen atoms

have been omitted for clarity.

In Situ Synthesis of Silver-Palladium Complexes. The in

situ syntheses of ionic complexes 2 from chloro compounds 5g

and 5j at low temperatures gave rise to the formation of

intermediates 6 and 7 (Scheme 7). See Supporting Information

for details.

acetate. The ³¹P{¹H} NMR spectrum showed a 1:1:1 triplet at

20.3 ppm (²J_PD) 6 Hz).

[(dppp)Pd(C(O)CH₃)(CD₃OD)]⁺[CF₃SO₃]^-(2h-CD₃OD) was

synthesized in situ from (dppp)Pd(C(O)CH₃)Cl and 1 equiv of

AgCF₃SO₃in CD₂Cl₂/CD₃OD (1:1.33, v/v, [Pd] ) 3.3 × 10^-2

M) at -75 °C.^5cComplex 2h-CD₃OD is stable at this temper-

ature. Raising the temperature to room temperature led to

decarbonylation of complex 2h-CD₃OD and deposition of

metallic palladium. Complex [(dppp)Pd(C(O)CH₃)(CO)]⁺-

[CF₃SO₃]^-(2h-CO) was synthesized in situ in CD₃OD when

(dppp)Pd(CH₃)₂was reacted with 1 equiv of triflic acid and

subsequent bubbling through of CO at -60 °C ([Pd] ) 3.3 ×

10^-2M).^5cComplex 2h-CO gives deposition of palladium metal

instead of 4h upon methanolysis at room temperature.

(o-C₆H₄(CH₂P(t-Bu)₂)₂)Pd(CH₃)Cl (5f-Me) was synthesized

from (COD)Pd(CH₃)Cl and o-C₆H₄(CH₂P(t-Bu)₂)₂, f, in toluene.

Bubbling of CO through a solution of 5f-Me in CD₂Cl₂resulted

in the formation of (o-C₆H₄(CH₂P(t-Bu)₂)₂)Pd(C(O)CH₃)Cl (5f-

C(O)Me). Minor quantities of water resulted in the formation

of (o-C₆H₄(CH₂P(t-Bu)₂)₂)Pd(H)Cl (5f-H) instead of (o-C₆H₄-

(CH₂P(t-Bu)₂)₂)Pd(C(O)CH₃)Cl (5f-C(O)Me). For the metha-

nolysis studies, (o-C₆H₄(CH₂P(t-Bu)₂)₂)Pd(C(O)CH₃)(O₂CCF₃)

was synthesized in situ from (o-C₆H₄(CH₂P(t-Bu)₂)₂)Pd(C(O)-

CH₃)Cl (5f-C(O)Me) and silver trifluoroacetate.

Kinetics

Kinetics of the Solvolysis Reaction of Acetylpalladium

Complexes Containing Trans-Coordinating Ligands. The

reaction rate of the methanolysis of acetylpalladium complexes

2a-2e was determined in the absence and presence of CO. The

presence of CO had an effect neither on the rate of the

methanolysis reaction nor on the course of the reaction.

Complex 2a containing SPANphos did not react at all with

a 10-fold excess of methanol at room temperature during 20 h

1

according to H NMR in the presence of 1 bar of CO.

[(dtbpf)PdC(O)CH₃]⁺[CF₃SO₃]^-(2b) did not react with

methanol in CD₂Cl₂at room temperature. When [(Xantphos)Pd-

C(O)CH₃]⁺[CF₃SO₃]^-(2c) was dissolved in CD₃OD containing

3.3 M CF₃COOD ([Pd] ) 3.3 × 10^-2M) in the presence of

CO, the formation of the palladium(I) dimer and methyl acetate-

d₃was observed. A reaction rate of 7.3 × 10^-5L mol^-1s^-1

was found at 12 °C. Complex 2c reacted very slowly with

10 equiv of methanol in CH₂Cl₂at 25 °C (t_1/2≈ 3 h) to yield

in part the palladium(I) hydride dimer, [Pd₂(µ-H)(µ-CO)-

{(Xantphos)}₂]⁺[CF₃SO₃]^-(4c), and methyl acetate, but the

reaction was accompanied by deinsertion of CO to afford

complex 1c.

X-ray Structure Determination of 5f-Me. Single crystals

were grown from CH₂Cl₂/Et₂O, and an X-ray crystal structure

analysis was performed. The complex is severely distorted from

a square planar geometry, due to the very large P-Pd-P angle

(104.14(1)°), which is one of the largest observed for a cis-

chelating ligand in palladium(II) complexes (Figure 2).^10a,41Two

positions were found for the chloride atom in the crystal

structure. Characteristic data are shown in Table 4.

The crystal structure of the corresponding propionyl palladium

complex has been reported.^10aThe bite angle in the latter

complex is almost the same (103°), and the palladium-carbon

distance is somewhat smaller at 2.01 Å, as might be expected.

The difference in the Pd-P distances is large in both complexes

and is due to the difference in trans influences.

A solution of [(dippf)PdC(O)CH₃]⁺[CF₃SO₃]^-(2d) reacted

slowly with 10 equiv of methanol in dichloromethane (3.3 ×

10^-2M) at 25 °C (t_1/2) 20 min), but the reaction was much

faster than that of 2a,b (no reaction) and 2c. The ¹H NMR and

the ¹³C{¹H} NMR spectrum indicate the formation of the

palladium(I) hydride dimer, [Pd₂(µ-H)(µ-CO){(dippf)}₂]⁺[CF₃-

SO₃]^-(4d).

The isolated compound 2e was used to perform kinetics with

1

the use of H NMR spectroscopy. CO was bubbled through a

3.3 × 10^-2M solution of 2e and 3.3 × 10^-1M methanol (10

equiv) in CD₂Cl₂at -90 °C. The rate constants were determined

over a 30° temperature range (-30 through 0 °C). Five data

points were used to produce an Eyring plot (Figure 3) giving

the activation enthalpy (∆H^q) 33.8 ((3.5) kJ mol^-1) and the

activation entropy (∆S^q) ^-179 ((32) J K^-1mol^-1). The use

of different amounts of methanol (10-23 equiv) at -20 °C

(41) Doherty, S.; Robins, E. G.; Knight, J. G.; Newman, C. R.; Rhodes, B.;

Champkin, P. A.; Clegg, W. J. Organomet. Chem. 2001, 640, 182.

9

5528 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Alcoholysis of Acylpalladium(II) Complexes

A R T I C L E S

Figure 5. Eyring plot of the ethene insertion of [(DPEphos)PdC(O)-

CH₃]⁺[CF₃SO₃]^-(2e) from 213.0 to 233.0 K ([Pd] ) 2.21 × 10^-2M,

[ethene] ) 2.32 × 10^-1M).

Figure 3. Eyring plot of the methanolysis of [(DPEphos)PdC(O)-

CH₃]⁺[CF₃SO₃]^-(2e) from -30.0 to 0 °C. T is in K.

Figure 6. The rates of alcoholysis of 2e at 25 °C for various alcohols.

The rate of methanolysis was calculated from the Eyring plot in Figure 3.

[Pd] ) 3.26 × 10^-2M. [Alcohol] ) 3.26 × 10^-1M. The data points of

2,2,2-trifluoroethanol and tert-butyl alcohol overlap one another (no decrease

of the integrals of 2e is observed).

Figure 4. The methanol concentration dependency on the rate of metha-

nolysis of 2e ([Pd] ) 3.3 × 10^-2M, [CD₃OD] ) 3.3 × 10^-1to 7.6 × 10^-1

M, T ) -20 °C).

showed a first-order dependency of the methanolysis reaction

on the methanol concentration (Figure 4). The plot of the

observed rate constants k_obsversus the concentration of methanol

resulted in a straight line with an intercept with the y axis at

zero indicating that the kinetics obey the typical rate law k_obs

) k₂[CH₃OH].

Scheme 8

The rate of insertion of ethene into the acetylpalladium bond

of [(DPEphos)PdC(O)CH₃]⁺[CF₃SO₃]^-(2e) was determined

over a 20° temperature range (-60 through -40 °C). Ethene

(10.5 equiv) was injected via a gastight syringe into a 2.21 ×

10^-2M solution of 2e in CD₂Cl₂at -78 °C. Five data points

were used to produce an Eyring plot (Figure 5). From the Eyring

plot, we calculated the activation enthalpy (∆H^q) +36.1 ((2.5)

kJ mol^-1) and the activation entropy (∆S^q) -149.1 ((22) K^-1

mol^-1).

signal of the 2-propanol reagent overlapped with the C(O)CH₃

signal of the acetylpalladium complex in the ¹H NMR spectrum.

The solvolysis using 2-propanol was very slow and therefore

could be monitored by ³¹P{¹H} NMR spectroscopy.

Kinetics of the Solvolysis Reaction of Acetylpalla-

dium Complexes Containing Cis-Coordinating Diphosphine

Ligands. Monitoring of the alcoholysis of (o-C₆H₄(CH₂P(t-

Bu)₂)₂)Pd(C(O)CH₃)(O₂CCF₃) (2f-O₂CCF₃) was attempted by

NMR spectroscopy at -90 °C in CD₂Cl₂with the use of 10

equiv of MeOH. The reaction was instantaneous giving methyl

acetate and the tentatively assigned palladium products shown

in Scheme 8. The half-life time is well below 3 min.

Methanol, ethanol, 2-propanol, CF₃CH₂OH, and tert-butyl

alcohol were used to determine the dependency of the rate of

solvolysis on the alcohol (Figure 6). The alcohol (10 equiv)

was added to a 3.3 × 10^-2M solution of 2e in CD₂Cl₂. Neither

CF₃CH₂OH nor tert-butyl alcohol reacted with 2e at 25 °C and

40 °C. Methanol reacted too fast to determine the reaction rate

1

by H NMR spectroscopy at 25 °C. An extrapolation of the

The in situ synthesized compound [(dppf)Pd(C(O)CH₃)-

(CO)]⁺[CF₃SO₃]^-(2g-CO) in CD₃OD was used to study the

kinetics of the reaction ([Pd] ) 3.3 × 10^-2M). The reaction

Eyring plot in Figure 3 was used to estimate the rate of

solvolysis at 25 °C. ³¹P{¹H} NMR spectroscopy was used to

measure the rate of solvolysis using 2-propanol, since the OH

1

was monitored by H NMR spectroscopy between -40 and 0

9

J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5529

A R T I C L E S

van Leeuwen et al.

Scheme 9

at 80 °C. The ionic methylpalladium complexes containing

SPANphos (a),³⁸dtbpf (b), and Xantphos (c) were inactive.

[(DPEphos)Pd CH₃(CF₃CO₂)] showed an activity of 2500 mol

mol^-1h^-1. The only products observed were methyl propanoate

and the low molecular weight oligomers H₃COC(O)CH₂CH₂C-

(O)C₂H₅and H₃COC(O)CH₂CH₂C(O)OCH₃(in a 40:4:1 ratio).

In dichloromethane solution, a copolymer was formed at low

reaction rates, probably due to catalyst decomposition. Catalysis

with the use of f-j has been published previously (vide infra,

Table 5).

Figure 7. Eyring plot of the methanolysis of in situ synthesized [(syn-

(calix[6]arene)diphosphite)Pd(C(O)CH₃)(CD₃OD)]⁺[CF₃SO₃]^-(2j) from

215.0 to 228.0 K (-58 to -45 °C).

Discussion

Although the ligands of this study form a limited set, they

were selected on the basis of steric and electronic properties,

and we will show that it provides us with new and crucial

information. Our present findings will lead to an interpretation

of many old facts and adjustments of previous interpretations.

The histories of copolymerization of alkenes and carbon

monoxide and oxycarbonylation of alkenes are long, and it is

impossible to do justice to all contributions to chain transfer in

this one paper. In due time, an overview on the topic will be

published⁴²also describing chain transfer.⁴³

The key questions to be answered in alkoxycarbonylation and

alkene/CO copolymerization are what factors govern the chain

transfer versus chain growth, what is the mechanism of

alkoxycarbonylation, and how can we explain the low rate for

a number of palladium-catalyzed reactions. The focus of this

work is on the formation of an ester or acid as the chain end

(reaction 3, Scheme 1), but it should be borne in mind that in

polymerization systems the other chain transfer mechanisms play

a role as well or may even dominate. Alkoxycarbonylation of

alkenes using palladium complexes has been the subject of many

reviews and research papers.⁴⁴For convenience, we show in a

short-hand notation the two mechanisms, Scheme 9, the

“hydride” mechanism, and Scheme 10, the “carbomethoxy”

mechanism.

°C. A pseudo first-order behavior in [2g] was not observed. At

-30 °C, the observed half-life time of the reaction was 83 min.

The absence of CO in the reaction mixture of the in situ

generated cationic acetylpalladium complex [(dppf)Pd(C(O)CH₃)-

(CD₃OD)]⁺[CF₃SO₃]^-(CD₂Cl₂:CD₃OD ) 1:1.33, v/v; [Pd] )

3.3 × 10^-2M; [CD₃OD] ) ∼14 M) avoids the interference of

CO during kinetics. The methanolysis reaction is faster in the

absence of CO. Unfortunately, decarbonylation occurs at the

temperature at which methanolysis is observed (T > -40 °C),

and the kinetics remained complex.

The solvolysis of complexes containing the dppp ligand h

could not be studied by NMR spectroscopy due to a fast

decomposition of the acetylpalladium complex to yield metallic

palladium. Instead, the solvolysis reaction of in situ synthesized

[(dppp)Pd(C(O)CH₃)(CO)]⁺[CF₃SO₃]^-(2h-CO) in methanol

could be studied by UV-vis spectroscopy at 30 °C, although

lower palladium concentrations ([Pd] ) 2 × 10^-4M) had to be

applied. The absorption at 456 nm (palladium(I) hydride dimer

4h) was monitored during the reaction. The amount of acid

added to the reaction mixture had a large influence on the

reaction rate. Under neutral reaction conditions (0 equiv of CF₃-

SO₃H), methanolysis was not observed. The addition of excess

triflic acid resulted in methanolysis of [(dppp)Pd(C(O)CH₃)-

(CO)]⁺[CF₃SO₃]^-. The highest reaction rate was observed when

9 equiv of triflic acid were added to the reaction mixture. The

rate of reaction decreased again when more equivalents of triflic

acid were added to the solution.

All individual steps for each mechanism have been observed

in model compounds, and it is conceivable that both mechanisms

occur, depending on catalyst and conditions. Even kinetic studies

or in situ observations of catalyst resting states may not give a

definite answer to this question.⁴⁵There is consensus in the

literature that the fast catalysts reported in the past decade

operate via the “hydride” mechanism, Scheme 9.^10,29,45,46

Circumstantial evidence for a hydride mechanism is the absence

of acrylate (or its methanol adduct formed via a palladium-

[(syn-(Calix[6]arene)diphosphite)Pd(C(O)CH₃)(CD₃OD)]⁺-

[CF₃SO₃]^-(2j-CD₃OD) decarbonylates above -40 °C in CD₂-

Cl₂in the presence of 10 equiv of methanol, but its methanolysis

1

could be studied by H NMR spectroscopy in the temperature

range -58 through -45 °C (CD₂Cl₂:CD₃OD ) 1:1.33, v/v). In

this case, the rate of methanolysis is pseudo first order. The

reaction is extremely temperature dependent. From the data in

Figure 7, we calculated the activation enthalpy (∆H^q) 101.7

((10) kJ mol^-1) and the activation entropy (∆S^q) 151 ((37)

J K^-1mol^-1).

Catalytic CO and Ethene Copolymerization Reactions.

The methylpalladium complexes were tested in the catalytic CO

and ethene copolymerization reaction. The ionic methylpalla-

dium complexes were tested in methanol containing 5 equiv

of p-toluenesulfonic acid at 20 bar of CO and ethene (1:1, v/v)

(42) Sen, A., Ed. Catalytic Synthesis of Alternating Alkene-Carbon Monoxide

Copolymers. Catalysis by Metal Complexes; Kluwer Academic Publish-

ers: Dordrecht, 2003.

(43) Zuideveld, M. A. Solvolysis of Palladium-Carbon Bonds in Palladium-

(II) Complexes Containing Diphosphine Ligands. Thesis, University of

Amsterdam, 2001.

(44) (a) Kiss, G. Chem. ReV. 2001, 101, 3435. (b) Beller, M.; Tafesh, A. M. In

Applied Homogeneous Catalysis with Organometallic Compounds; Cornils,

B., Herrmann, W. A., Eds.; VCH: Weinheim, 1996; Chapter 2.1.2.6. (c)

Ali, B. E.; Alper, H. In Transition Metals for Organic Synthesis; Beller,

M., Bolm, C., Eds.; Wiley-VCH: Weinheim, 1998; Chapter 2.2.

(45) Sperrle, M.; Consiglio, G. Chem. Ber./Recueil 1997, 130, 1557.

(46) (a) Cavinato, G.; Toniolo, L. J. Organmetal. Chem. 1990, 398, 187. (b)

Dekker, G. P. C. M.; Elsevier, C. J.; Vrieze, K.; van Leeuwen, P. W. N.

M.; Roobeek, C. F. J. Organomet. Chem. 1992, 430, 357.

9

5530 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Alcoholysis of Acylpalladium(II) Complexes

A R T I C L E S

Table 5. Overview of Ligand Properties a-j, Structures and Reactivities of Compounds 1 and 2, and Catalysis Results^a

reaction of

estimated t₁

/2

X-ray angle

(deg) P

1, CO

methanolysis T (K),

[MeOH] (M)

2 with CH₃OH^c

at 243 K and

0.33 M MeOH

b

ligand

θ_1/

ø

−

M

−

P

1a

−j

insertion

2a

−j

t₁, ∆

/2

G^q(T in K)^c

catalysis results^a

2

a, SPANphos

b, dtbpf

c, Xantphos

162^d10.3 172^e

167

162

trans slow

trans 298, 0.33

trans 285, 24.7, 3.3 M 158′, 92 (285)

CF₃COOD

trans 298, 0.33

no reaction, 20 h >5 × 10⁶

no reaction

4.3 158⁶⁰, 104^f

10.3 150-165⁶⁰,

104, 108^g

no reaction, 24 h >5 × 10⁶

2.7 × 10⁵

d, dippf

e, DPEphos

155

162

6.7 99^h, 103ⁱ

10.3 101, 165⁶⁰

cis

moderate

fast

20′, 91 (298)

15′, 83 (273)

800′

102′

oligomers

methyl propanoate

10% oligomers

trans 243-273, 0.33

f, dtbpx

g, dppf

h, dppp

174

145

140

3.7 104^{10,this work}cis

very fast

fast^27a

cis

183, 0.33

243, 24.7

303, 24.7

<3′, <52 (183)

83′, 77 (243)

5′, 90 (303)

acid added 374′

4′, 66 (228)

<5 × 10^-6methyl propanoate¹⁰

3.5 × 10⁶,

2.6 × 10⁸

0.14′

13.0 99⁵⁸

10.9 91⁵⁸

cis

83′

oligomers⁹⁶

polymer²

very fast^4acis

j, calix-6-

diphosphite

145

30.2 90⁸⁷

cis

fast

cis

215-228, 14

no reaction

MeOH polymer

in CH₂Cl⁸⁷

^aCatalysis results were obtained typically in methanol, 20-60 bar ethene/CO (1:1), 60-90 °C. ^bTolman’s parameters θ and ø were estimated for half

(θ_1/2) the bidentate ligands using published methods.^{101a,b c}Not corrected for [MeOH]. ∆G^qis in kJ mol^{-1 d}Value for Ph₂P(o-anisyl).^{104a e}Value for

.

trans-(a)PtCl₂.³⁸^fValue for cis-(nor)(b)Rh⁺complex^{104b g}Value for cis-(c)Pd(allyl)⁺108°.^104cValue for cis-(c)Pd(TCNE) 104°.^{104d h}Value for (cod)(d)Ir⁺

complex.^104eⁱValue for (d)PdCl₂.^104f

Scheme 10

Scheme 11

catalyzed Michael addition)⁹and succinate in the product, the

putative byproducts of a carbomethoxy route, Scheme 10. Note

that after formation of these side products the catalysis continues

as a hydride cycle and oxidation is needed to turn the catalyst

back into the carbomethoxy cycle. Thus, in the absence of

oxidizing agents, any catalyst is bound to end up in the hydride

cycle. A best guess about rates will also point to the hydride

cycle as being the fastest one: (a) insertion of ethene into a

palladium hydride is much faster than insertion of ethene in a

palladium-carbomethoxy bond,⁴⁷(b) insertion of CO is fast in

any case, (c) protonation of alkylpalladium bonds is slow. In

the hydride mechanism, the highest barrier occurs at the

alcoholysis of the acyl species, the topic we will discuss in the

following. Under oxidative conditions, palladium-catalyzed

methoxycarbonylation of ethene does lead to dimethyl succinate,

but the reaction is much slower than methyl propanoate

formation.⁴⁸

ligand do not show methanolysis at all, but the other complexes

are reactive toward methanol. Carbonylation of cis-1d-CH₃CN

resulted in a trans complex trans-2d, due to the sp²character

and the larger trans influence of the acetyl carbon-to-palladium

bond. In the presence of CO, the cis-acetylpalladium complexes

contain coordinated CO,^4a,49but, in the absence of CO at room

temperature, they lose the coordinated CO and the acetyl groups

decarbonylate instantaneously (Scheme 10). This limits the

conditions at which the methanolysis of cis-acetylpalladium

complexes can be studied.

Synthesis of Acetylpalladium Complexes. The new acetyl-

palladium complexes we synthesized can be divided in trans

complexes and cis complexes. Part of the ligands coordinating

in a trans fashion can be forced to form cis complexes by adding

a strongly coordinating ligand, such as triphenylphosphine. The

trans-acetylpalladium(diphosphine) complexes are relatively

stable in the absence of CO at room temperature, which enabled

us to study the alcoholysis reaction. The acetylpalladium

complexes 2a and 2b containing SPANphos and dtbpf as the

The palladium-silver complexes 6 and 7 seem to have no

precedent. Presumably, they are intermediates in the common

anion exchange reactions. Similar platinum complexes have been

reported by Uso´n et al.⁵⁰

Mechanism of the Alcoholysis of Acetylpalladium Com-

plexes. The ionic acetylpalladium complexes 2a-j were used

to study the mechanism and kinetics of the alcoholysis reaction,

because they are involved in the common catalysts. We will

discuss several mechanistic proposals of the solvolysis of

acylpalladium bonds (Scheme 11).

(47) Cavinato, G.; Toniolo, L. J. Organomet. Chem. 1990, 398, 187.

(48) (a) Nefkens, S. C. A.; Sperrle, M.; Consiglio, G. Angew. Chem., Int. Ed.

Engl. 1993, 32, 1848. (b) Bianchini, C.; Mantovani, G.; Meli, A.;

Oberhauser, W.; Bru¨geller, P.; Stampfl, T. J. Chem. Soc., Dalton. Trans

2001, 690. (c) Drent, E.; van Broekhoven, J. A. M.; Doyle, M. J. J.

Organomet. Chem. 1991, 417, 235. (d) Bianchini, C.; Lee, H. M.; Meli,

A.; Oberhauser, W.; Peruzzini, M.; Vizza, F. Organometallics 2002, 21,

16.

Mechanisms A and B allow coordination of the two phos-

phines in trans positions and we have drawn them as such, but

(49) (a) Braunstein, P.; Frison, C.; Morise, X. Angew. Chem., Int. Ed. 2000,

39, 2867. (b) To´th, I.; Elsevier, C. J. J. Am. Chem. Soc. 1993, 115, 10388.

(50) Uso´n, R.; Fournies, J.; Tomas, M.; Ara, I. Inorg. Chem. 1994, 33, 4023.

9

J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5531

A R T I C L E S

van Leeuwen et al.

a cis coordination is not excluded for this type of mechanism.

The trans coordination has often been said to play a role in

creating a preference for chain termination rather than growth

reactions,^2,4b,14,51which by nature of the migratory insertion

reaction requires cis positions for the substrate and growing

chain and, hence, for the two phosphine ligands.^3bTrans acyl

complexes containing two monophosphines have been identified

by ³¹P{¹H} NMR spectroscopy in several instances^4b,14,51albeit

at temperatures lower than those at which the reaction takes

place, because at the reaction conditions exchange of triphenyl-

phosphine takes place on the NMR time scale leading to one

broad signal for the excess and coordinated phosphine.²⁹Thus,

we must assume that cis-trans isomerization is very fast in

monophosphine based catalysts and so is dissociation of a ligand.

Dissociation of one phosphine has also been suggested as a

possible step involved in hydroxycarbonylation of styrene.²⁹

Mechanism C is a classic reductive elimination reaction, and

mechanism D is a reductive elimination preceded by a migratory

step of the alkoxy moiety; both reactions require cis positioning

of the two groups.

Outer-Sphere Attack. Mechanism A is an outer-sphere

attack of the alcohol or alkoxyde nucleophile at the acetyl-carbon

atom, similar to alcoholysis of acyl chlorides, and this mech-

anism was assumed to be the most likely one.³⁰In the absence

of base, one molecule of alcohol attacks the acyl-carbon atom,

while a second molecule of alcohol may serve as a general base,

abstracting the proton. Therefore, acyl chloride alcoholysis often

gives a higher order in alcohol concentration than one.⁵²In our

experiments, only first-order dependency on the methanol

concentration was observed, but, in general alcoholysis, is a

too complex a reaction to draw firm conclusions. As in ester

hydrolysis under acidic conditions, protonation of the acyl

oxygen may also occur. The reactivity order found MeOH >

EtOH > i-PrOH > t-BuOH > CF₃CH₂OH is not the order of

nucleophilicity, except for the nonreactivity of trifluoroethanol,

but this order is usually encountered in alcoholysis reactions of

acyl chlorides, organyl halides, and the alcoholysis of chloro-

silanes.⁵³The order is ascribed to the predominance of steric

effects when the reaction is of an S_N2 type and also to inductive

and entropic effects.⁵⁴Capture of bulky vinylic cations by

alcohols is also dominated by steric effects of the alcohols.⁵⁵

In the following, we will discard the mechanism of direct

attack.⁵⁶

We have recently reported on a new trans-spanning ligand,

SPANphos (a), which forms trans complexes with platinum and

palladium and does not form cis complexes. In the crystal

structure of the free ligand, the P-P distance is 4.99 Å, and in

the platinum dichloride trans-adduct, the P-P distance is 4.59

Å, which shows that even for the formation of a trans complex

some distortion of the ligand is needed and a cis coordination

cannot be achieved.³⁸Interestingly, this bis-triarylphosphine

ligand was completely inactiVe in the methoxycarbonylation

reaction of ethene (or copolymerization) under the standard

conditions with the use of several ways of initiation. After the

reaction, the palladium complexes were still intact and no

palladium metal was formed.

Ligand a forms trans complexes 1-2a. For ligands b (dtbpf)

and c (Xantphos), it was known that they form ionic trans-

methylpalladium complexes,⁶⁰but the i-propyl derivative of

ferrocene diphosphine d formed a cis complex under the same

conditions. The trans configuration of 1c was explained by the

wide bite angle and the stabilizing coordination of the ether

oxygen atom, while the trans configuration of 1b was ascribed

to the steric repulsion of the tert-butyl groups. Trans-acetyl

complexes 2a and 2b showed no methanolysis with a 10-fold

excess of methanol at room temperature, while the methanolysis

of 2c had a half-life time of almost 3 h at 12 °C. Since the acyl

carbon atoms in the cationic acetyl complexes 2a-c all have

weak donor atoms in trans positions, we assume that they are

highly electrophilic, and yet they do not (or very slowly) react

with methanol. From this we conclude that alcoholysis of

acylpalladium complexes in neutral to acidic media does not

involve a direct attack by alcohol at the acyl carbon atom.

An outer-sphere attack of alkoxide ions has been discussed

as a possible mechanism for the related methoxycarbonylation

reaction of organic halides, which typically occurs in highly

basic media and where the nucleophile is the alkoxide an-

ion.^30,61,62Yamamoto et al. reported in their study of the

reductive elimination of aryl acetates from (aryloxy)(acetyl)-

palladium complexes that addition of free alkoxides as the base

had no effect on the rate.⁶¹Buchwald also excluded an outer-

(57) (a) Venanzi, L. M. Appl. Chem. 1980, 52, 1117. (b) Boron-Rettore, P.;

Grove, D. M.; Venanzi, L. M. HelV. Chim. Acta 1984, 67, 65. (c) Johnson,

D. K.; Pregosin, P. S.; Venanzi, L. M. HelV. Chim. Acta 1976, 59, 2691.

(d) Barrow, M.; Burgi, H. B.; Camalli, M.; Caruso, F.; Fischer, E.; Venanzi,

L. M.; Zambonelli, L. Inorg. Chem. 1983, 22, 2356. (e) Camalli, M.; Caruso,

F.; Chaloupka, S.; Venanzi, L. M. HelV. Chim. Acta 1988, 71, 703. (f)

Barrow, M.; Burgi, H. B.; Johnson, D. K.; Venanzi, L. M. J. Am. Chem.

Soc. 1976, 98, 2356.

Since for trans diphosphine complexes direct attack is the

only available mechanism (after having excluded oxidative

addition mechanisms, vide infra), we studied complexes with

an enforced trans configuration, that is, a P-Pd-P angle close

to 180°. Ligands that coordinate in a purely trans fashion are

rare, and it is often observed that trans-coordinating ligands form

cis complexes as well, like some of the ligands in this

work.^37,57-59

(58) Dierkes, P.; van Leeuwen, P. W. N. M. J. Chem. Soc., Dalton Trans. 1999,

1519.

(59) (a) Sawamura, M.; Hamashima, H.; Ito, Y. Tetrahedron: Asymmetry 1991,

2, 593. (b) Sawamura, M.; Hamashima, H.; Ito, Y. J. Am. Chem. Soc. 1992,

114, 8295. (c) Sawamura, M.; Hamashima, H.; Ito, Y. Tetrahedron 1994,

50, 4439. (d) Kuwano, R.; Sato, K.; Kurokawa, T.; Karube, D.; Ito, Y. J.

Am. Chem. Soc. 2000, 122, 7614. (e) Kuwano, R.; Sawamura, M.; Ito, Y.

Bull. Chem. Soc. Jpn. 2000, 73, 2571. (f) Kuwano, R.; Sawamura, M.;

Shirai, J.; Takahashi, M.; Ito, Y. Bull. Chem. Soc. Jpn. 2000, 73, 485. (g)

Sawamura, M.; Hamashima, H.; Shinoto, H.; Ito, Y. Tetrahedron Lett. 1995,

36, 6479. (h) Sawamura, M.; Sudoh, M.; Ito, Y. J. Am. Chem. Soc. 1996,

118, 3309. (i) Kuwano, R.; Miyazaki, H.; Ito, Y. Chem. Commun. 1998,

71.

(51) (a) Seayad, A.; Jayasree, S.; Damodaran, K.; Toniolo, L.; Chaudhari, R.

V. J. Organomet. Chem. 2000, 601, 100. (b) Bardi, B.; Del Pra, A.; Piazzesi,

A. M.; Toniolo, L. Inorg. Chim. Acta 1979, 35, L345.

(52) Ross, S. D. J. Am. Chem. Soc. 1970, 92, 5998.

(60) Zuideveld, M. A.; Swennenhuis, B. H. G.; Boele, M. D. K.; Guari, Y.; van

Strijdonck, G. P. F.; Reek, J. N. H.; Kamer, P. C. J.; Goubitz, K.; Fraanje,

J.; Lutz, M.; Spek, A. L.; van Leeuwen, P. W. N. M. J. Chem. Soc., Dalton

Trans. 2002, 2308.

(61) (a) Komiya, S.; Akai, Y.; Tanaka, K.; Yamamoto, T.; Yamamoto, A.

Organometallics 1985, 4, 1130. (b) Kubota, Y.; Hanoka, T.; Takeuchi, K.;

Sugi, Y. J. Mol. Catal. A: Chem. 1996, 111, L187. (c) Moser, W. R.;

Wang, A. W.; Kjeldahl, N. K. J. Am. Chem. Soc. 1988, 110, 2816. (d)

Schoenberg, A.; Bartoletti, I.; Heck, R. F. J. Org. Chem. 1974, 39, 3318.

(62) (a) Ma¨gerlein, W.; Indolese, A.; Beller, M. Angew. Chem., Int. Ed. 2001,

40, 2856, (b) Ma¨gerlein, W.; Indolese, A.; Beller, M. J. Organomet. Chem.

2002, 641, 30.

(53) Yasuhara, S.; Shimizu, N.; Inazu, T. Kyushi UniVersity Ser. C (Engl) 1993,

19, 129; Chem. Abstr. 1995, 122, 187842.

(54) Hine, J. Physical Organic Chemistry, 2nd ed.; McGraw-Hill: New York,

1962.

(55) Biali, S. E.; Rappoport, Z. J. Org. Chem. 1986, 51, 964.

(56) (a) Bentley, T. W. AdV. Chem. Ser. 1987, 215. Harris, J. M., McManus, S.

P., Eds.; Nucleophilicity; American Chemical Society: Washington, DC;

p 255. (b) Bentley, T. W.; Shim, C. S. J. Chem. Soc., Perkin Trans. 2

1993, 1659.

9

5532 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Alcoholysis of Acylpalladium(II) Complexes

A R T I C L E S

sphere attack of alkoxide at the ipso-carbon of an aryl-

palladium bond as the mechanism for palladium-catalyzed aryl

ether formation.⁶³

in a cis fashion. Mechanism C is a concerted reductive

elimination in which the new carbon-oxygen bond is formed

while the palladium-carbon and palladium-oxygen bonds are

being broken. Three pathways have been proposed and experi-

mentally verified for this reaction when both groups are

hydrocarbyls: elimination from a reactive, 14-electron, three-

coordinate species after dissociation of one of the ligands,^69,70

elimination from four-coordinate species,^61a,71and elimination

from five-coordinate species.^61aDissociation of a monophos-

phine prior to reductive elimination has been established

experimentally, and it was found to occur also for cis

complexes.^61aThe intermediacy of three-coordinate species in

hydroxycarbonylation of alkenes has been invoked to explain²⁹

the differences between monophosphines and diphosphines in

these reactions, in terms of rates and selectivities.^17,72As they

show similar behavior as monophosphines, the selectivities for

low molecular weight products of bulky diphosphines have been

assigned to the assumed propensity to dissociate one arm of

the bidentate ligand.^2,10b,16A direct elimination of esters from

a four-coordinate palladium species has been observed,^61aand

also in related C-X bond forming reactions, it has been

demonstrated that this is the most likely pathway.⁷¹Reductive

elimination from a five-coordinate intermediate often involves

a π-acceptor ligand (such as benzoquinone or acrylonitrile) as

the fifth ligand, which accelerates the reductive elimination,

especially for dialkyl nickel complexes and also, if only slightly,

for palladium complexes.^61a,69a,73Addition of benzoquinone to

a copolymerization system of palladium sometimes leads to a

reduction of the molecular weight of the polymer, which might

indicate that benzoquinone induces chain termination by reduc-

tive elimination.⁷⁴

Thus, trans-acyl complexes resist alcoholysis, and apparently

cis complexes are required for alcoholysis. The question then

arises why did the trans complexes 1a-c undergo insertion of

carbon monoxide, which definitely requires cis coordination of

the methyl group and carbon monoxide and, hence, also of the

two phosphine donors, assuming they remain bonded to the

metal. There are several solutions to this, but there are no

experimental data in support of any mechanism. CO is a

relatively strong ligand, and it may replace one of the phosphines

in 1a-c. Monocoordination of a has been observed in rhodium

complexes under CO.³⁸For ligand c, we know it can also

function as a cis ligand. Complex 2b could under the influence

of CO break the palladium-iron interaction and act as a short-

lived cis bidentate. A second general possibility is the formation

of 5-coordinate 18-electron species, iso-electronic with the

rhodium species involved in rhodium hydroformylation, which

could perform an insertion reaction as has been proposed for

platinum and nickel complexes.⁶⁴Thus, carbon monoxide can

enter the coordination sphere more easily than methanol, and

routes to its insertion are accessible.

Oxidative Addition of Alcohols to Palladium(II). Mecha-

nism B (Scheme 11) involves an oxidative addition of methanol,

which results in a cationic Pd(IV) complex that undergoes a

reductive elimination to form methyl acetate. This mechanism

seems unlikely for cationic palladium(II) complexes containing

phosphine ligands, although palladium(IV) intermediates have

also been proposed for Heck reactions⁶⁵and, in compounds

containing nitrogen donor atoms, their occurrence is abundant.⁶⁶

Palladium(IV) formation does not comply with the low reactivity

of CF₃CH₂OH, as it is known that especially the acidic alcohols

such as phenols and fluoro-substituted alkanols will oxidatively

add to palladium(0)⁶⁷and methanol will add when the pal-

ladium(0) center is very electron-rich,³⁵but oxidative addition

to palladium(II) has not been observed. The low reactivity of

some of the alkyl-substituted phosphines argues against oxida-

tive addition as the mechanism. An oxidative addition mecha-

nism has been proven as the pathway for the protonolysis of

the Pt-C bond in neutral Pt(II) complexes using strong acids,⁶⁸

but the reactivity of cationic complexes toward oxidative

addition is much lower, while strong acids are much stronger

oxidizing species in such reactions than alcohols. Thus, most

likely alcoholysis of acylpalladium species does not proceed

via an oxidative addition pathway.

Mechanism D is a reductive elimination that due to the

unlikeness of the two groups starts off as a migratory reaction,

in which one of the reactants acts as a nucleophile migrating to

the electrophilic neighbor. In recent years, a lot of evidence

has been presented for this reaction mode in the palladium-

catalyzed C-X bond formation by the groups of Buchwald and

Hartwig.^63,75An early theoretical study about migratory hydro-

carbyl 1,2-shifts between transition metals and coordinated main

group atoms was published by Hoffmann and co-workers.⁷⁶

(69) (a) Ozawa, F.; Ito, T.; Nakamura, Y.; Yamamoto, A. Bull. Chem. Soc. Jpn.

1981, 54, 1868. (b) Ozawa, F.; Ito, T.; Yamamoto, A. J. Am. Chem. Soc.

1981, 102, 6457. (c) Milstein, D.; Stille, J. K. J. Am. Chem. Soc. 1979,

101, 4981. (d) Tatsumi, K.; Hoffmann, R.; Yamamoto, A.; Stille, J. K.

Bull. Chem. Soc. Jpn. 1981, 54, 1857. (e) Alcazar-Roman, L. M.; Hartwig,

J. F. J. Am. Chem. Soc. 2001, 123, 12905. (f) Hartwig, J. F. Angew. Chem.,

Int. Ed. 1998, 37, 2046. (g) Littke, A. F.; Dai, C.; Fu, G. C. J. Am. Chem.

Soc. 2000, 122, 4020. (h) Gillie, A.; Stille, J. K. J. Am. Chem. Soc. 1980,

102, 4933. (i) Driver, M. S.; Hartwig, J. F. J. Am. Chem. Soc. 1995, 117,

4708. (j) Stockland, R. A., Jr.; Anderson, G. K.; Rath, N. P. J. Am. Chem.

Soc. 1999, 121, 7945. (k) Reid, S. M.; Mague, J. T.; Fink, M. J. J. Am.

Chem. Soc. 2001, 123, 4081.

(70) (a) Beare, N. A.; Hartwig, J. F. J. Org. Chem. 2002, 67, 541. (b) Stambuli,

J. P.; Stauffer, S. R.; Shaughnessy, K. H.; Hartwig, J. F. J. Am. Chem.

Soc. 2001, 123, 2677. (c) Netherton, M. R.; Fu, G. C. Org. Lett. 2001, 3,

4295. (d) Aranyos, A.; Old, D. W.; Kiyomori, A.; Wolfe, J. P.; Sadighi, J.

P.; Buchwald, S. L. J. Am. Chem. Soc. 1999, 121, 4369.

(71) (a) Brown, J. M.; Cooley, N. A. Organometallics 1990, 9, 353. (b) Culkin,

D. A.; Hartwig, J. F. J. Am. Chem. Soc. 2001, 123, 5816. (c) Driver, M.

S.; Hartwig, J. F. J. Am. Chem. Soc. 1997, 119, 8232.

(72) del Rio, I.; Ruiz, N.; Claver, C.; van der Veen, L. A.; van Leeuwen, P. W.

N. M. J. Mol. Catal. A: Chem. 2000, 161, 39.

(73) (a) Ozawa, F.; Fujimori, M.; Yamamoto, T.; Yamamoto, A. Organometallics

1986, 5, 2144. (b) Yamamoto, T.; Yamamoto, A.; Ikeda, S. J. Am. Chem.

Soc. 1971, 93, 3360.

Reductive Elimination in Cis Complexes. In mechanisms

C and D, the methoxy group (or molecule of methanol) and

the acetyl group should occupy cis positions relative to one

another and therefore the diphosphine ligand coordinates also

(63) (a) Widenhoefer, R. A.; Zhong, H. A.; Buchwald, S. L. J. Am. Chem. Soc.

1997, 119, 6787. (b) Widenhoefer, R. A.; Buchwald, S. L. J. Am. Chem.

Soc. 1998, 120, 6504.

(64) (a) Bryndza, H. E. Organometallics 1985, 4, 406. (b) Shultz, C. S.;

DeSimone, J. M.; Brookhart, M. J. Am. Chem. Soc. 2001, 123, 9172.

(65) Shaw, B. L. Chem. Commun. 1998, 1863.

(66) Canty, A. J. Acc. Chem. Res. 1992, 25, 83.

(67) (a) Di Bugno, C.; Pasquali, M.; Leoni, P.; Sabatino, P.; Braga, D. Inorg.

Chem. 1989, 28, 1390. (b) Perez, P. J.; Calabrese, J. C.; Bunel, E. E.

Organometallics 2001, 20, 337.

(74) (a) Sperrle, M.; Consiglio, G. J. Am. Chem. Soc. 1995, 117, 12130. (b)

Nefkens, S. C. A.; Sperrle, M.; Consiglio, G. Angew. Chem., Int. Ed. 1993,

32, 1719.

(75) Mann, G.; Baranano, D.; Hartwig, J. F.; Rheingold, A. L.; Guzei, I. A. J.

Am. Chem. Soc. 1998, 120, 9205.

(68) (a) Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 1995,

117, 9371. (b) Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. J. Am. Chem.

Soc. 1996, 118, 5961. (c) Holtcamp, M. W.; Labinger, J. A.; Bercaw, J. E.

J. Am. Chem. Soc. 1997, 119, 848.

9

J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5533

A R T I C L E S

Scheme 12

van Leeuwen et al.

Scheme 13

They studied reductive elimination/oxidative addition of P-C

fragments to various transition metals including the Pd(II)-

(IV) couple in relation to P-C bond cleavage.⁷⁷The migration

barrier of a phenyl group turned out to be lower than that of a

methyl group due to the π-stabilization.

The cross-coupling reaction is particularly effective when one

of the partners of the reductive elimination has an sp²-hybridized

carbon atom bonded to the metal, and usually this is the (former)

electrophile stemming from the organic halide. The reductive

elimination of ethene from a vinyliridium hydride complex was

best described as a migration of the hydride anion to the vinyl

group.⁷⁸Calhorda et al. analyzed the reductive elimination of

propene from cis-(PH₃)₂Pd(CH₃)(CHdCH₂) and they found that

the best description for this reaction is a migration of the methyl

group to the sp²carbon of the vinyl unit.⁷⁹

The formation of the C-X bond in hetero-cross-coupling

reactions studied in the past decade is thought to proceed via a

migration of the heteroatom to the aryl group, which develops

a negative charge, which is π-stabilized by mesomeric interac-

tion with acceptor substituents. π-Stabilization of 1,2-phenyl

carbanion shifts has been analyzed many years ago.⁸⁰The

resonance-stabilized Meisenheimer complex proposed by Buch-

wald for the reductive elimination of aryl ethers is depicted in

Scheme 12.

For the reverse reaction, oxidative addition of substituted

chlorobenzenes to zerovalent palladium, Portnoy and Milstein

proposed that the Hammett correlation obtained could be best

explained by the assumption of an intermediary Meisenheimer

(or similar) complex having a late transition state.⁸¹As early

as 1971, Fitton and Rick described the oxidative addition of

aryl halides in a similar fashion.⁸²

3a,10,27a,49b,^83-85but their reactivity toward methanol has received

little attention. When S is replaced in cis-(L-L)PdC(O)CH₃-

(S)⁺by methanol, one of the fast reactions observed is formation

of methyl acetate, even in the absence of a base. When less

nucleophilic alcohols were used, such as phenols, both the cis

and trans complexes L₂PdC(O)CH₃(OAr) could be isolated

containing even the aryloxide anions instead of the alcohols,

and the reactivity of these aryloxide-palladium complexes has

been studied.^61aFor Ar ) p-NCC₆H₄, a slow formation of aryl

acetate was observed (t_1/2) 24 h, 35 °C), while, after addition

of p-CH₃C₆H₄OH to this complex, rapid formation of p-tolyl

acetate was found (t_1/2) 1 h, 25 °C) indicating a rapid exchange

of aryl oxides and a much faster elimination reaction for the

stronger nucleophile. The same reduction in t_1/2was observed

when dppe was added to a solution of trans-(PEt₃)₂PdC(O)-

CH₃(OAr), which also led to the suggestion that cis complexes

are required for ester formation. The low reactivity of the

acetylpalladium(aryloxide) complexes is in accord with the

nonreactivity of CF₃CH₂OH toward ionic acetylpalladium, and

it illustrates the enormous range of reactivities.

Complexes cis-(L-L)PdC(O)CH₃(CH₃OH)⁺(L-L is a diphos-

phine) react extremely fast to give methyl acetate, but we hardly

succeeded in collecting any data on this process. The kinetics

turned out to be very complex, and the cationic species are

highly prone to decarbonylation. Thus, ironically, one can only

study the elimination reaction for the most reactive alcohol, that

is, methanol! Part of the complexity of the kinetics becomes

evident from Scheme 13. In the cis complexes, a competition

exists between the coordination of CO and methanol, and the

rate of methyl acetate formation depends on the concentration

of CO. In the absence of CO, decarbonylation occurs, and this

reaction is fast.^3a,27a,49bSecond, like esterification reactions, there

may be a general acid and base catalysis, as probably the

coordinated methanol is deprotonated before it migrates to the

acyl carbon. Water, alcohol, the anion, and also palladium(0)

may serve as a general base. During the reaction, acid is released

because only one out of two protons ends up in the palladium

dimer product. Since the rate of methyl acetate formation is

dependent on the acid concentration, this rate will change during

the alcoholysis process. It is expected that hydrogen bond

formation, protonation and deprotonation, and methanol ex-

change may constitute an important step in the process, but since

they are all occurring at a very high rate, they can be dealt with

as fast preequilibria (Curtin-Hammett conditions).

We are strongly in favor of a migratory elimination to form

esters from acyl-alkoxy palladium complexes for reasons given

in the following. First, the acyl carbon atom is sp²hybridized,

much more electrophilic than an aryl carbon atom, and highly

stabilized by the structure where the negative charge is on the

oxygen atom (Scheme 12). The acyl oxygen atom may, as in

acid-catalyzed alcoholysis of esters, be protonated, before or

after the formation of the new carbon-oxygen bond. Second,

for the reverse reaction, the oxidative addition of esters to

palladium(0), the atom at which the nucleophilic palladium

center will attack the ester clearly is the electrophilic ester carbon

atom, rather than the oxygen atom.

Cationic acetylpalladium complexes cis-(L-L)PdC(O)CH₃-

(S)⁺have been reported (S ) nitrile, thf, CO, PPh₃),

(76) Ortiz, J. V.; Havlas, Z.; Hoffmann, R. HelV. Chim. Acta 1984, 67, 1.

(77) (a) Green, M. L. H.; Smith, M. J.; Felkin, H.; Swierczewski, G. J. Chem.

Soc., Chem. Commun. 1971, 158. (b) Coulson, R. D. J. Chem. Soc., Chem.

Commun. 1968, 1530. (c) Garrou, P. E. Chem. ReV. 1985, 85, 171. (d)

Goodson, F. E.; Wallow, T. I.; Novak, B. M. J. Am. Chem. Soc. 1997,

119, 12441. (e) Van Leeuwen, P. W. N. M. Appl. Catal., A 2001, 212, 61.

(78) Silvestre, J.; Calhorda, M. J.; Hoffmann, R.; Stoutland, P. O.; Bergman,

R. G. Organometallics 1986, 5, 1841.

(79) Calhorda, M. J.; Brown, J. M.; Cooley, N. A. Organometallics 1991, 10,

1431.

(80) Zimmerman, H. E.; Zweig, A. J. Am. Chem. Soc. 1961, 83, 1196.

(81) Portnoy, M.; Milstein, D. Organometallics 1993, 12, 1665.

(82) Fitton, P.; Rick, E. A. J. Organomet. Chem. 1971, 28, 287.

For the study of ligand f, complex (o-C₆H₄(CH₂P(t-Bu)₂)₂)-

Pd(C(O)CH₃)(O₂CCF₃) was synthesized with the aim to slow

(83) Dekker, G. P. C. M.; Buijs, A.; Elsevier, C. J.; Vrieze, K.; van Leeuwen,

P. W. N. M.; Smeets, W. J. J.; Spek, A. L.; Wang, Y. F.; Stam, C. H.

Organometallics 1992, 11, 1937.

(84) Fryzuk, M. D.; Clentsmith, G. K. B.; Retting, S. J. J. Chem. Soc., Dalton

Trans. 1998, 2007.

(85) Nozaki, K.; Sato, N.; Tonomura, Y.; Yasutomi, M.; Takaya, H.; Hiyama,

T.; Matsubara, T.; Koga, N. J. Am. Chem. Soc. 1997, 119, 12779.

9

5534 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Alcoholysis of Acylpalladium(II) Complexes

A R T I C L E S

slightly the rate of alcoholyisis by the use of a coordinating

anion, but the reaction was too fast, even at -90 °C. While

this manuscript was in preparation, Clegg et al. reported the

same result for the related complex (o-C₆H₄(CH₂P(t-Bu)₂)₂)-

Pd(C(O)C₂H₅)(THF)⁺.^10a

The in situ synthesized compound [(dppf)Pd(C(O)CH₃)-

(S)]⁺[CF₃SO₃]^-(2g; S ) MeOD, CO) did not yield a pseudo

first-order behavior. At -30 °C, the observed half-life time of

the reaction was 83 min when CO was present, while, in the

absence of CO, decarbonylation took place simultaneously with

methanolysis (>-40° C).

it was thought that bite angle effects were dominant in many

palladium-, nickel-, and rhodium-catalyzed reactions,^58,88-91but

evidence is growing that the steric bulk, measured as cone angle,

solid angle, or pocket angle, determines the relative rates of

insertion reactions and reductive elimination reactions.²⁶If the

substituents at phosphorus are all phenyl groups for the ligands

we are comparing, clearly the backbone induced bite angles run

parallel with the steric properties of the (bidentate) ligands. High

rates for reductive elimination, be it a symmetric type (C,

Scheme 11) or a migratory type of reaction (D, Scheme 11),

are well documented for a large number of palladium-catalyzed

reactions.^63,78,79,81Thus, when complexes are used containing

ligands having a high steric bulk, a fast elimination takes place,

and for ligands that are less bulky, the insertion reactions become

dominant (e.g., dppp). For still smaller ligands, the overall

productivity decreases, but the rate of the elimination reaction

is affected more than the insertion reactions. The first insertion

taking place in methyl propanoate formation is insertion of

ethene in a palladium hydride bond, and clearly this is little

affected by the steric bulk of the cis diphosphine as is the

insertion of CO. Insertion of ethene into acyl-palladium bonds

is slowed drastically by sterically encumbered ligands such as

f.

[(dppp)Pd(C(O)CH₃)(CO)]⁺[CF₃SO₃]^-, 2h-CO, gave only

little solvolysis at room temperature under neutral reaction

conditions in pure methanol. The reaction was accelerated by

the addition of triflic acid. The role of the acid might be to

activate the acetyl group and enhance the electrophilicity of the

carbonyl carbon. The experiments showed that the reaction is

slower when a large excess of acid is used. Because of the

complexity of this system, no conclusions can be drawn,

although it seems justified to say that half-life times are in the

order of tens of minutes at 30 °C, which is fairly slow compared

to those of other cis ligands, and are even in the range of several

“trans” ligands.

The only cis complex that neither decarbonylated nor

solvolyzed too fast was [(syn-(calix[6]arene)diphosphite)Pd-

(C(O)CH₃)(CD₃OD)]⁺[CF₃SO₃]^-, 2j. The extremely large posi-

tive entropy of activation (∆S^q) 151 ((37) J K^-1mol^-1) can

be explained by the loss of the proton and/or the reaction of

ions with opposite charges. The conversion of the ionic complex

into a neutral complex, which is less solvated and less highly

structured, can give rise to a high positive entropy of activation.

Similar effects have been reported in proton transfer studies.⁸⁶

The value found for ∆S^qremains relatively high for such

phenomena, and perhaps more preequilibrium steps are involved

in the overall expression of the rate. At -45 °C, [CD₃OD] )

∼14 M, the value for t_1/2amounts to 13′. In view of the fast

termination reaction, it would have been interesting to test the

activity of catalyst 2j in methanol, but unfortunately, it is

inactive in methanol/dichloromethane. Presumably, it is inactive

because after the fast termination reaction, a stable, zerovalent

palladium complex is formed, which cannot be reprotonated

by acid. In dichoromethane, 2j gave a polymer containing in

part acid groups as chain ends indicating the presence of traces

of water.⁸⁷Proton release from water coordinated to an electron-

poor palladium center should be facile compared to the other

ligand systems discussed here, and a reductive elimination step

should be fast in view of the electron-withdrawing character of

a triaryl phosphite. The P-Pd-P angle is small at 90°, which

is not attractive for reductive elimination, but we propose that

the dominant factor is not a steric one in this instance but an

electronic one. This results in a fast reductive elimination

irrespective whether it proceeds via mechanism C or D.

The rate of methanolysis decreases in the series of cis ligands

reported here as follows: dppp < dppf < calix-6, j < dtbpx, f.

The series dppp < dppf < dtbpx represents a series of steric

bulk and bite angles, but the former is more important. Initially,

The critical bite angle seems to be 103-104°; the bite angle

found in (o-C₆H₄(CH₂P(t-Bu)₂)₂)Pd(CH₃)Cl (11) amounted to

104°, and Clegg¹⁰found 103° for the -C(O)C₂H₅analogue.

Ligands having similar bite angles but carrying the smaller

phenyl groups instead of the tert-butyl groups at phosphorus,

or a phosphole group, give polymerization catalysts rather than

methyl propanoate catalysts, although sometimes a bimodal

product distribution has been observed.⁹²

Reductive Elimination in Trans Complexes. Previously, we

concluded that the trans complexes should rearrange to the cis

complexes, during which the Z atom (Fe, O) decoordinates and

methanol enters the coordination sphere (Scheme 13). The large

negative entropy of activation (∆S^q) -179 ((32) J K^-1mol^-1),

which has been observed for the methanolysis of [trans-

(DPEphos)PdC(O)CH₃]⁺[CF₃SO₃]^-, can be accounted for by

the associative nature of the reaction involving the coordination

of methanol. The kinetic data for ethene insertion in this complex

are almost the same (Figures 4 and 5). The present tridentate

ligands are indeed capable of rearranging to cis-chelating

bidentate ligands, as can be inferred from the addition of ligands,

such as triphenylphosphine, to a solution of acetylpalladium

complexes 2e and 2d giving complexes 3e and 3d, respectively.

Furthermore, the related methyl derivative cis-(DPEphos)-

PdCH₃(CH₃CN)]⁺[CF₃SO₃]^-(1d-CH₃CN) has a cis structure

with a P-Pd-P bite angle of 103°.⁶⁰A similarly large negative

entropy of activation was reported for the association/insertion

reactions in N,N,N tridentate acetylpalladium complexes.⁹³The

large negative value is in contrast to the large positive value

(88) (a) Goertz, W.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Vogt, D.

Chem.sEur. J. 2001, 7, 1614.

(89) Carbo´, J. J.; Maseras, F.; Bo, C.; van Leeuwen, P. W. N. M. J. Am. Chem.

Soc. 2001, 123, 7630.

(90) van Leeuwen, P. W. N. M.; Kamer, P. C. J.; Reek, J. N. H.; Dierkes, P.

Chem. ReV. 2000, 100, 2741.

(91) Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Reek, J. N. H. Acc. Chem

Res. 2001, 34, 895.

(92) Moulton, C. J.; Shaw, B. L. J. Chem. Soc., Dalton 1976, 1020.

(93) Groen, J. H.; de Zwart, A.; Vlaar, M. J. M.; Ernsting, J. M.; van Leeuwen,

P. W. N. M.; Vrieze, K.; Kooijman, H.; Smeets, W. J. J.; Spek, A. L.;

Budzelaar, P. H. M.; Xiang, Q.; Thummel, R. P. Eur. J. Inorg. Chem. 1998,

1129.

(86) (a) Oki, M.; Ohira, M.; Yoshioka, Y.; Morita, T.; Kihara, H.; Nakamura,

N. Bull. Chem. Soc. Jpn. 1984, 57, 2224. (b) Parker, V. D.; Tilset, M. J.

Am. Chem. Soc. 1991, 113, 8778.

(87) Parlevliet, F. J.; Zuideveld, M. A.; Kiener, C.; Kooijman, H.; Spek, A. L.;

Kamer, P. C. J.; van Leeuwen, P. W. N. M. Organometallics 1999, 18,

3394.

9

J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5535

A R T I C L E S

van Leeuwen et al.

found for 2j, and this raises the question if the same steps are

involved in the rate equation. Since the bite angle of DPEphos

is rather wide, we expect (vide infra) that the methanolysis

reaction will be extremely fast, once the cis complex has been

formed. It might well be therefore that the rate-determining step

for the methanolysis of 2e is the formation of the cis complex

via association of methanol. This would also explain the neat

first-order dependency in methanol we have found for this

reaction, which one would not expect for a complex process

such as acyl-palladium bond rupture and ester formation, similar

to alcoholysis of reactive acyl halides. The first-order depen-

dency is also in accord with the noninterference of carbon

monoxide with the alcoholysis process, because the rate of

alcohol complexation does not change when no substantial

amount of trans-2e is converted in a CO adduct.

Extrapolation of the present kinetic data of DPEphos com-

plexes shows that the expected rate of methanolysis is 35 mol

mol^-1h^-1at 80 °C in 0.33 M methanol. Since the reaction rate

is first order in methanol concentration, this corresponds to 3200

mol mol^-1h^-1in pure methanol, which is the maximum rate

to be expected for the catalytic process. This is considerably

lower than the value for the most effective catalysts (40 000

mol mol^-1h^-1at 80-100 °C for “fast” ligands such as f),¹⁰

because the DPEphos (e) system resides in the inactive trans

configuration for a great deal of the time. The observed reaction

rate in the CO and ethene co-oligomerizations reaction (2500

mol mol^-1h^-1) corresponds very well with this number.

Extrapolation of the data for the insertion rates of ethene in the

acetylpalladium complex shows a rate of 45 mol mol^-1h^-1at

80 °C at an ethene concentration of 0.23 M. Assuming a first-

order dependency in ethene concentration, we would correspond

this to 900 mol mol^-1h^-1at 20 bar of ethene pressure, which

is the correct order of magnitude. On the basis of the Flory-

Schulz constant found (0.1), one would expect 270 mol mol^-1

h^-1, but we did not and cannot take into account the competing

coordination of CO at higher pressures, which will lead to a

lower rate than the one calculated (900 mol mol^-1h^-1).

Furthermore 20% of the oligomer is a diester (EE) product; this

proves that for this catalyst part of the reaction occurs via the

carbomethoxy cycle, which may lead to a lower productivity,

as part of the palladium content is tied up as a less reactive

carbomethoxy complex.

tert-butyl alcohol, CF₃CH₂OH. This is the order of nucleo-

philicity toward the acyl-palladium center, but also toward the

palladium cation, or the order of stability of the alcohol solvate

complexes to palladium. CF₃CH₂OH is much less reactive

toward the acetylpalladium complex than ethanol, due to its

reduced nucleophilicity. The very low rate of alcoholysis

observed for CF₃CH₂OH is in agreement with the results of

the copolymerization of CO and styrene reported by Milani et

al.³³who observed that the molecular weight of the copolymer

obtained in CF₃CH₂OH was higher than that in methanol. This

result and ours seems to be in contrast to the findings of

Yamamoto et al. who compared the selectivity of acylpalladium

chloride complexes in a competition experiment between

diethylamine and various alcohols. They found that the more

acidic alcohols resulted in a higher selectivity for ester instead

of acid amide.¹⁵In these competition experiments, the more

acidic alcohols such as CF₃CH₂OH may react via the alkoxide,

because a replacement of the chloride ion by alkoxide may be

required and thus the position of CF₃CH₂OH in the series is

different from that in our series. Another factor we did not

mention so far is the strong ionizing power of CF₃CH₂OH, but

also that of methanol, compared to those the other alcohols.

These solvents cause a relative stabilization of the ionic species,

which are the catalytically active ones, compared to the neutral

palladium zero resting states of the catalyst. Even though we

presume that especially the anions will be stabilized by the

alcohols with a high ionizing power⁹⁵rather than the large

cations constant, these effects can be very large.

The order of the rate of methanolysis of complexes containing

the trans-coordinating diphosphine ligands is SPANphos ≈ dtbpf

(no reaction) < Xantphos , dippf < DPEphos. The slow

reaction of [(Xantphos)PdC(O)CH₃]⁺[CF₃SO₃]^-with methanol

can be explained by the rigidity of Xantphos compared to

DPEphos. The DPEphos ligand can rearrange to a cis ligand

more easily than the rigid Xantphos ligand. The complex

containing the dtbpf ligand, 2b, does not react at all with

methanol. The cationic acetylpalladium complex is too sterically

crowded to form a cis complex that can undergo methanolysis.

The less crowded dippf complex, 2d, does react with methanol

faster than the Xantphos complex, 2c, but more slowly than

the DPEphos complex, 2e. A comparison of the steric properties

between ferrocene based ligands and the other three ligands is

not straightforward, but the order within each group {(SPAN-

phos, Xantphos, DPEphos) versus (dtbpf, dippf)} is as expected.

The ferrocene range can be extended with two examples from

recent literature, dppf and octamethyl-dppf, carrying eight

methyl substituents at the ferrocene rings.⁹⁶It is known that

dppf occasionally forms trans complexes,⁹⁷but the acetyl-

palladium(dppf) complex has a cis structure in the presence of

acetonitrile.^27aThe product of the co-oligomerization with the

use of dppf is indeed a mixture of oligomers (rate 5000 mol

mol^-1h^-1, at 85 °C). Octamethyl-dppf is sterically more

crowded, although not via substitution directly at the phosphorus

atoms, as appears from the P-Pd-P angle, which is 101° as

It is interesting to compare the product distributions obtained

with DPEphos and PPh₃under similar conditions in methanol.

The Flory-Schulz constant (growth factor) for DPEphos is

∼0.1, and that for PPh₃was reported to be ∼0.33.¹The resting

state for PPh₃is also the trans complex, which has to rearrange

to the cis complex before it can undergo insertion reactions of

substrates or undergo a chain termination. The respective bite

angles of the two complexes are 103° (DPEphos) and 100.5°

(cis-(PPh₃)₂),⁹⁴and since the aryl substituents are almost the

same, the pocket angle for the DPEphos complex will be

smaller. In line with our reasoning, this will result in a lower

Flory-Schulz constant for the DPEphos-based catalyst.

As mentioned previously, the order of the rate of alcoholysis

for the several alcohols is methanol > ethanol > i-propanol >

(95) Bentley, T. W. AdVances in Chemistry Series; American Chemical

Society: Washington, DC, 1987; Vol. 215, Chapter 18. ibid. Peterson, P.

E. Chapter 21.

(96) Gusev, O. V.; Kalsin, A. M.; Peterleitner, M. G.; Petrovskii, P. V.; Lyssenko,

K. A.; M. G.; Akhmedov, N. G.; Bianchini, C.; Meli, A.; Oberhauser, W.

Organometallics 2002, 21, 3637.

(97) Sato, M.; Shigeta, H.; Sekino, M.; Akabori, S. J. Organomet. Chem. 1993,

458, 199.

(94) A search in the CSD gave 39 structures of monometallic, square-planar,

divalent palladium complexes containing two triphenylphophine ligands

in cis positions with an average P-Pd-P angle of 100.46°. Typically the

ones with angles > 100° contain small groups such as allylic or

cyclopropenyl groups trans to the triphenylphosphine ligands. See Sup-

porting Information for a listing.

9

5536 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Alcoholysis of Acylpalladium(II) Complexes

A R T I C L E S

compared to 96° for dppf in the dicationic palladium diaqua

adducts.⁹⁶Indeed, octamethyl-dppf gives methyl propanoate in

the palladium-catalyzed reaction with ethene, CO, and MeOH,

albeit at a modest rate (600 mol mol^-1h^-1, at 85 °C). It should

be noted that the dppf system is also highly susceptible to the

“enolate” chain termination reaction, which contributes to the

formation of short chains.^96,98

insertion of isolated CO molecules in a polyethylene chain is

thermodynamically uphill and, hence, kinetically determined!).

The increased steric bulk also destabilizes the alcohol complex,

facilitating the “migratory elimination” and thus causing chain

transfer. The importance of steric bulk in this type of elimination

reactions has been reported many times in recent years.^63,71c,102

Quantitative Overview. Before we set out to try to present

a quantitative comparison of the ligands, a comment on the rate-

determining step seems in place. Several authors have sug-

gested^{10,11,29,31,51}that the solvolysis of the acylpalladium species

is the rate-determining step of the reaction sequence. In a system

like the present one, this often leads to confusion, because the

barrier for the alcoholysis for the fast catalyst does not seem

very high at all when one looks at the stoichiometric reaction.

Indeed, neither we nor Clegg et al.¹⁰succeeded in measuring

the rate of the methanolysis of acylpalladium complexes

containing ligand f, not even when the fourth site was occupied

by a coordinating trifluoroacetic acid anion. The half-life time

for the reaction of the latter complex with methanol is below

180 s at -90 °C. When we extrapolated to the reaction

conditions of the catalysis (80-90 °C), the rate could be 5-10

orders of magnitude higher, depending on the value we assign

to the entropy of activation. At both low and high temperatures,

it seems to be the fastest step of all steps considered! Perhaps

the statement should be reconsidered or modified. Under the

actual conditions, there is competition between the coordination

of methanol and CO for the cis complexes. Furthermore, all

species involved are very close in energy, and an equilibration

between all species occurs as is shown in a shorthand notation,

with omission of dimeric species and the various coordination

states that the complexes may assume.

When the tert-butyl groups are replaced by the smaller

i-propyl groups in (t-Bu)₂P(CH₂)₃P(t-Bu)^21aor ligand f,^10cboth

systems give oligomers as the product instead of methyl

propanoate at high rates. When the 1,3-propanediyl bridge in

(t-Bu)₂P(CH₂)₃P(t-Bu)₂is replaced by a 1,2-ethanediyl bridge,

the accessiblity of the catalyst for ethene increases such that,

in the reaction of ethene, CO, MeOH, and H₂, pentan-3-one

was formed at extremely high rates (showing the fast insertion

of ethene) instead of methyl propanoate, the product of the more

bulky ligand.^21b

Qualitative Overview. Previously, we have reported on the

bidentate ligand effect on the rates of migration reactions, and

we found that, in accord with the overall rates in the catalytic

experiments of Drent et al.,³the rates of the stoichiometric

insertion reactions followed the sequence dppe < dppp < dppb,

which we explained using the model developed by Thorn and

Hoffmann.⁹⁹According to their EH calculations, the ligand

sitting next to the migrating group will “follow” the migrating

group, thus enlarging the bite angle of the bidentate ligand. This

led Thorn and Hoffmann to the statement that dppe would be

a bad choice as a ligand in catalysts for migration reactions.

Ab initio results for platinum confirmed this picture.^99bIn the

sequence tested, both the bite angle and the flexibility increase,

and until recently, there was little reason to abandon this

model.¹⁰⁰In IMOMM calculations on alkene insertion in

rhodium hydride bonds, it was found that the bite angle of the

diphosphine hardly changes during the migration process.⁸⁹In

recent years, however, several small bite angle catalysts lacking

flexibility have been reported that gave fast copolymerization

catalysts.^25,26,34,60Thus, the picture emerging from current

literature is that the steric bulk of the complex enhances the

migratory insertion reactions and that the effect is not one of

flexibility or bite angle, although bite angle and steric properties

are related within certain series of substituted ligands. The steric

properties of bidentate ligands have been described in a number

of ways in the literature, but we will not go into detail of that

here (cone angle,^101a,bAMS,^101csolid angle,^101dpocket angle^4c).

Destabilization of the starting hydrocarbyl-metal complex with

the alkene by steric hindrance causes a higher reaction rate of

migration. At a certain point though while the steric bulk of

the complex increases, the complex can no longer fulfill the

steric demands for alkene coordination and only CO coordina-

tion or alcohol coordination remains. Insertion of a CO in a

metal-acyl bond does not take place, as it is thermodynamically

unfavorable (note that in the overall thermodynamics even the

Pd(0) h PdH⁺h PdC₂H₅⁺h PdD(O)C₂H₅⁺h

(CH₃OH)PdC(O)C₂H₅⁺f C₂H₅CO₂CH₃

Only the last step is irreversible (the overall reaction has a

free energy gain of ∼92 kJ mol^-1, and this seems to be gained

mainly in the last step), and oxidative addition of nonactivated

esters to palladium(0) cannot occur.¹⁰³The equilibrium con-

centration of the acyl species can be low as the acyl species

need not be the “resting” state of the catalyst. Even “dormant”

states may be involved that require reactivation to form

palladium(0). Improvement of catalyst systems are related not

only to changing the rates of individual steps but also perhaps

more so to the optimization of the positions of the equilibria

involved.

(102) (a) Brown, J. M.; Guiry, P. J. Inorg. Chim. Acta 1994, 220, 249. (b) Huang,

J.; Haar, C. M.; Nolan, S. P.; Marcone, J. E.; Moloy, K. G. Organome-

tallics 1999, 18, 297. (c) Marcone, J. E.; Moloy, K. G. J. Am. Chem. Soc.

1998, 120, 8527. (d) Kranenburg, M.; Kamer, P. C. J.; van Leeuwen, P.

W. N. M.; Vogt, D.; Keim, W. J. Chem. Soc., Chem. Commun. 1995,

2177. (e) Goertz, W.; Kamer, P. C. J.; Van Leeuwen, P. W. N. M.; Vogt,

D. J. Chem. Soc., Chem. Commun. 1997, 1521. (f) Aranyos, A.; Old, D.

W.; Kiyomori, A.; Wolfe, J. P.; Sadighi, J. P.; Buchwald, S. L. J. Am.

Chem. Soc. 1999, 121, 4369. (g) Mann, G.; Baran˜ano, D.; Hartwig, J. F.;

Rheingold, A. L.; Guzei, I. A. J. Am. Chem. Soc. 1998, 120, 9205. (h)

Hayashi, T.; Konishi, M.; Kobori, Y.; Kumada, M.; Higuchi, T.; Hirotsu,

K. J. Am. Chem. Soc. 1984, 106, 158. (i) Wolfe, J. P.; Wagaw, S.;

Buchwald, S. L. J. Am. Chem. Soc. 1996, 118, 7215. (j) Driver, M. S.;

Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7217. (k) Hartwig, J. F.;

Richards, S.; Baran˜ano, D.; Paul, F. J. Am. Chem. Soc. 1996, 118, 3626.

(l) Dyer, P. W.; Dyson, P. J.; James, S. L.; Martin, C. M.; Suman, P.

Organometallics 1998, 17, 4344. (m) Fu, G. C. Pure Appl. Chem. 2002,

74, 33. (n) Littke, A. F.; Schwarz, L.; Fu, G. C. J. Am. Chem. Soc. 2002,

124, 6343.

(98) Zuideveld, M. A.; Kamer, P. C. J.; van Leeuwen, P. W. N. M. Submitted.

(99) (a) Thorn, D. L.; Hoffmann, R. J. Am. Chem. Soc. 1978, 100, 2079. (b)

Koga, N.; Morokuma, K. J. Am. Chem. Soc. 1986, 108, 6136.

(100) In IMOMM calculations on alkene insertion in rhodium hydride bonds,

it was found that the bite angle of the diphosphine hardly changes during

the migration process.⁸⁹

(101) (a) Tolman, C. A. Chem. ReV. 1977, 77, 313. (b) Bartik, T.; Himmler, T.

J. Organomet. Chem. 1985, 293, 343. (c) Angermund, K.; Baumann, W.;

Dinjus, E.; Fornika, R.; Go¨rls, H.; Kessler, M.; Kru¨ger, C.; Leitner, W.;

Lutz, F. Chem. Eur. J. 1997, 3, 755. (d) White, D.; Tavener, B. C.; Leach,

P. G. L.; Coville, N. J. J. Organomet. Chem. 1994, 478, 205.

(103) Aryl trifluoroacetates oxidatively added to (Me₃P)₂Pd(styrene) with acyl-O

bond cleavage: Nagayama, K.; Shimizu, I.; Yamamoto, A. Bull. Chem.

Soc. Jpn. 1999, 72, 799.

9

J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5537

A R T I C L E S

van Leeuwen et al.

Since the rates of methanolysis differ by many orders of

magnitude for the complexes studied, a comparison of ligand

properties and the performance of their acetylpalladium com-

plexes in the methanolysis reaction is a risky enterprise. For

one trans complex, we have found a highly negative entropy of

activation and for one cis complex, a highly positive entropy

of activation has been measured. This makes that extrapolation

of all data to the same temperature may lead to large errors, if

one uses a wrong number for the entropy. Besides, at the

temperature of extrapolation, another reaction might have

become the dominant one. Nevertheless, to avoid vague

descriptors, we have calculated the estimated half-life times for

complexes 2a-j in the reaction with methanol at 243 K. For

the trans complexes 2c-e, we have assumed that the reaction

is first order in methanol and ∆S^qis taken as -179 J mol^-1

K^-1(as found for 2e). For the cis complexes, it was assumed

that the reaction is zero order in methanol, the effects of CO

and acid were not taken into account, and ∆S^qis taken as +151

J mol^-1K^-1(as found for 2j). Table 5 gives the results of these

estimates together with a few properties such as the Tolman

values θ_1/2and ø, P-M-P angles taken from X-ray studies as

a measure for the bite angle.⁵⁸It can be seen that, for the fastest

and the slowest systems (2a vs 2f), the rates may differ by as

much as 12 orders of magnitude, and even if one would assume

a highly negative entropy of activiation for 2f, a difference of

still 8 orders of magnitude would remain. Ligand h, dppp, is

the best ligand for making polymers in this list. Using the

assumptions mentioned previously, we find a half-life times for

methanolysis at 353 K of milliseconds and tenths of seconds in

the presence of acid. These values seem too low because the

time for the growth of one polymer chain is in the order of

minutes, but the presence of CO will lead to much higher values.

Clearly, more data are needed, but the present set of ligands

containing many extreme cases can probably serve as a good

starting point for further research.

6-diphosphite complex showed a large positive entropy of

activation, which was ascribed to the loss of solvation in the

transition state. On the basis of electronic properties of the

ligand, one expects a high rate of reductive elimination (methyl

acetate formation), and indeed the reductive elimination is

relatively fast. For the phosphine series, the steric effects

dominate. In a series of increasing steric bulk of bidentate

ligands, initially the rates of insertions increase giving more

productive polymerization catalysts due to destabilization of the

alkene complexes formed prior to insertion; at one point, this

breaks down as ethene coordination becomes unfavorable and

only carbon monoxide or alcohol can coordinate cis to an acyl

group. In its turn, elimination is also accelerated by an increasing

steric bulk, and thus, the bulkiest ligands such as dtbpx, f, give

the fastest catalysts for making methyl propanoate. The two

insertion reactions involved in this process can remain fast

because neither ethene insertion into a palladium hydride bond

nor CO insertion into an ethylpalladium bond requires much

space.

Experimental Section

Kinetics of the Protonolysis Reaction of [(dppf)PdCH₃-

(CH₃CN)]⁺[CF₃SO₃]^-(1g-CH₃CN). An NMR tube was charged with

1g-CH₃CN and 0.700 mL of CD₃OD. CF₃COOH (10 equiv) was added

to this solution at -78 °C. The tube was shaken and immediately

transferred to the NMR probe, which was set to the required temper-

ature, and the spectra were acquired after the temperature of the probe

was stabilized. The integrals of [PdCH₃]⁺were measured to follow

the decay of 1g to give [Pd]²⁺

.

Kinetics of the Alcoholysis Reaction of [(DPEphos)PdC(O)-

CH₃]⁺[CF₃SO₃]^-(2e). The reaction was followed by ¹H NMR

spectroscopy (methanol, ethanol, 1,1,1-trifluoroethanol, and tert-butyl

alcohol) or ³¹P NMR spectroscopy (i-propanol). NMR probe temper-

atures were calibrated using an anhydrous methanol sample.¹⁰⁵An NMR

tube with a Young valve was charged with 0.700 mL of a 3.3 × 10^-2

M CD₂Cl₂solution of [(DPEphos)PdC(O)CH₃]⁺[CF₃SO₃]^-. Alcohol

(ROH, 10 equiv) was added to this solution at -78 °C. The tube was

shaken and immediately transferred to the NMR probe, which was set

to the required temperature, and the spectra were acquired after the

temperature of the probe was stabilized. The integrals of PdC(O)CH₃

and ROC(O)CH₃were measured to follow the decay of [(DPEphos)-

PdC(O)CH₃]⁺[CF₃SO₃]^-. The mass balance was checked using the

CHDCl₂peak.

Conclusions

The alcoholysis of a variety of acetylpalladium complexes

has been studied, and it was found that cis coordination of

alcohol and acetyl is needed to form an ester. Trans complexes

containing SPANphos, a, or dtbpf, b, do not react at all.

Complexes that have a trans structure but can temporarily adopt

a cis structure do react with methanol to form methyl acetate.

This proves that an outer-sphere attack of alkoxides/alocohols

is not a viable mechanism for alcoholysis. The reaction of

[(DPEphos)Pd(C(O)CH₃)]⁺[CF₃SO₃]^-with methanol shows a

large negative entropy of activation as the association with

methanol is rate determining; the kinetics of this reaction are

similar to that of the insertion reaction of this complex with

ethene. The pressure of carbon monoxide has no effect on the

rate of the methanolysis reaction. The mechanism is thought to

be an intramolecular, migratory, nucleophilic attack of alcohol/

alkoxide at the acyl group, which is in accord with the rates of

reaction found for the series of alcohols (methanol > ethanol

> i-propanol > tert-butyl alcohol, CF₃CH₂OH), the less

sterically hindered alcohols reacting faster as is commonly

observed in solvolysis reactions.

Kinetics of the Alcoholysis Reaction of [(syn-Calix[6]arene

diphosphite)PdC(O)CH₃(CH₃OD)]⁺[CF₃SO₃]^-, 2j-CH₃OD. The re-

1

action was followed by H NMR spectroscopy. NMR probe tempera-

tures were calibrated using an anhydrous methanol sample. An NMR

tube with a Young valve was charged with 0.3 mL of a 5.13 × 10^-2

M solution of (syn-calix[6]arene diphosphite)Pd(C(O)CH₃)Cl in CD₂-

Cl₂. A solution of CD₃OD (0.4 mL) containing 1 equiv of TlCF₃SO₃

was added to the solution at -90 °C. The NMR tube was shaken and

immediately transferred to the NMR probe, which was set to the

(104) (a) Hirsivaara, L.; Guerricabeitia, L.; Haukka, M.; Suomalainen, P.;

Laitinen, R. H.; Pakkanen, T. A.; Pursiainen, J. Inorg. Chim. Acta 2000,

307, 48. (b) Cullen, W. R.; Kim, T.-J.; Einstein, F. W. B.; Jones, T.

Organometallics 1983, 2, 714. (c) van Haaren, R. J.; Goubitz, K.; Fraanje,

F.; van Strijdonck, G. P. F.; Oevering, H.; Coussens, B.; Reek, J. N. H.;

van Leeuwen, P. W. N. M. Inorg. Chem. 2001, 40, 3363. (d) Kranenburg,

M.; Delis, J. G. P.; Kamer, P. C. J.; van Leeuwen, P. W. N. M.; Vrieze,

K.; Veldman, N.; Spek, A. L.; Goubitz, K.; Fraanje, J. J. Chem. Soc.,

Dalton Trans. 1997, 1839. (e) Avent, A. G.; Bedford, R. B.; Chaloner, P.

A.; Dewa, S. Z.; Hitchcock, P. B. Dalton Trans. 1996, 4633. (f) Elsagir,

A. R.; Gaâner, F.; Go¨rls, H.; Dinjus, E. J. Organomet. Chem. 2000, 597,

139. Boyes, A. L.; Butler, I. R.; Quayle, S. C. Tetrahedron Lett. 1998,

39, 7763.

Cis acetyl complexes underwent competing decarbonylation

and methanolysis, and for these complexes, the rate depends

strongly on acid and carbon monoxide concentration. The calix-

(105) Van Geet, A. L. Anal. Chem. 1970, 42, 679.

9

5538 J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003

Alcoholysis of Acylpalladium(II) Complexes

A R T I C L E S

required temperature, and the spectra were acquired after the temper-

ature of the probe was stabilized. The integrals of PdC(O)CH₃and

CH₃C(O)OCD₃were measured to follow the decay of [(syn-calix[6]-

arene diphosphite)Pd(C(O)CH₃)(CD₃OD)]⁺[CF₃SO₃]^-. The mass bal-

ance was checked using the CHDCl₂peak.

Acknowledgment. We are indebted to the Shell Research

and Technology Centre, Amsterdam (M.A.Z.) and the National

Research School Combination Catalysis (Z.F.) for financial

support. We thank Drs. W. P. Mul, J. N. H. Reek, and E. Drent

for stimulating discussions. This work was supported in

part (M.L., A.L.S.) by The Netherlands Foundation for Chem-

ical Sciences (CW) with financial aid from the Netherlands

Organization for Scientific Research (NWO).

Catalytic CO/Ethene Copolymerization Reactions. The experi-

ments were carried out in a 180 mL stainless steel autoclave. The

autoclave was charged with 20 mL of methanol and 0.100 mmol

of p-toluenesulfonic acid and was pressurized with 10 bar of ethene.

The autoclave was then heated to 80 °C. A dichloromethane solution

(1 mL) containing 0.20 mmol (L-L)Pd(CH₃)(CF₃CO₂) or [(L-L)-

PdCH₃]⁺[CF₃SO₃]^-was introduced into the autoclave, and the autoclave

was further pressurized to 20 bar of CO and ethene (1:1 ratio). The

autoclave was depressurized after 1 h. The methanol soluble fractions

Supporting Information Available: Experimental details,

Synthesis and NMR spectra of all compounds, UV spectra of

compounds 2e, the list of complexes containing cis-(PPh₃)₂-

PdX₂, and CIF files for the two crystal structures (compounds

2e′ and 5f-Me). This material is available free of charge via

the Internet at http://pubs.acs.org.

1

were analyzed by GC, GC-MS, and H and ¹³C NMR spectroscopy

(CDCl₃). The products with higher boiling points than that of methyl

propanoate were analyzed by ¹H and ¹³C NMR spectroscopy in CD₂Cl₂/

CF₃CO₂D (90:10, v/v).

JA029341Y

9

J. AM. CHEM. SOC. VOL. 125, NO. 18, 2003 5539

Article Doi

Alcoholysis of acylpalladium(II) complexes relevant to the alternating copolymerization of ethene and carbon monoxide and the alkoxycarbonylation of alkenes: The importance of cis-coordinating phosphines

DOI: 10.1021/ja029341y

Source and publish data:

Authors:

Article abstract of DOI:10.1021/ja029341y

Full text of DOI:10.1021/ja029341y

Products guided by the article

R&D Labs maybe for 540536-69-4

Relevant to this article

Hot Product