Thermal stability, decomposition paths, and Ph/Ph exchange reactions of [(Ph3P)2Pd(Ph)X] (X = I, Br, Cl, F, and HF2)

Article abstract of DOI:10.1021/om0001265

Complexes of the type [(Ph₃P)₂Pd(Ph)X], where X = I (1), Br (2), Cl (3), F (4), and HF₂ (5), possess different thermal stability and reactivity toward the Pd-Ph/P-Ph exchange reactions. While 1 decomposed (16 h) in toluene at 110 °C to [Ph₄P]I, Pd metal, and Ph₃P, complexes 2 and 3 exhibited no sign of decomposition under these conditions. Kinetic studies of the aryl-aryl exchange reactions of [(Ph₃P)₂Pd(C₆D₅)X] in benzene-de demonstrated that the rate of exchange decreases in the order 1 > 2 > 3, the observed rate constant ratio, k_I:k_Br:k_Cl, in benzene at 75 °C being ca. 100:4:1 for 1-d₅, 2-d₅, and 3-d₅. The exchange was facilitated by a decrease in the concentration of the complex, polar media, and a Lewis acid, e.g., Et₂O·BF₃. Unlike [Bu₄N]PF₆, which speeded up the exchange reaction of 2-d₅, [Bu₄N]-Br inhibited it due to the formation of anionic four-coordinate [(Ph₃P)Pd(C₆D₅)Br₂]^-. The latter and its iodo analogue were generated in dichloromethane and benzene upon addition of [Bu₄N]X or PPN Cl to [(Ph₃P)₂Pd₂(Ph)₂(μ-X) ₂] (X = I, Br, or Cl) and characterized in solution by ¹H and ³¹P NMR spectral data. The mechanism of the aryl-aryl exchange reactions of [(Ph₃P)₂Pd(C₆D₅)X] in noncoordinating solvents of low polarity may not require Pd-X ionization but rather involves phosphine dissociation, the ease of which decreases in the order X = I > Br > Cl, as suggested by crystallographic data. Two mechanisms govern the thermal reactions of [(Ph₃P)₂Pd(Ph)F], 4. One of them is similar to the aryl-aryl exchange and decomposition path for 1-3, involving a tight ion pair intermediate, [Ph₄P][(Ph₃P)PdF], within which two processes were shown to occur. At 75 °C, the C-P oxidative addition restores the original neutral complex (4). At 90 °C, reversible fluoride transfer from Pd to the phosphonium cation resulted in the formation of covalent [Ph₄PF] and [(Ph₃P)Pd], which was trapped by PhI to produce [(Ph₃P)₂Pd₂(Ph)₂(μ-I) ₂]. The other decomposition path of 4 leads to the formation of [(Ph₃P)₃Pd], Pd, Ph₂ , Ph₃PF₂, and Ph₂P-PPh₂ as main products. Unlike the aryl-aryl exchange, this decomposition reaction is not inhibited by free phosphine. The formation of biphenyl was shown to occur due to PdPh/PPh coupling on the metal center. Mechanisms accounting for the formation of these products are proposed and discussed. The facile (4 h at 75 °C) thermal decomposition of [(Ph₃P)₂Pd(Ph)(FHF)] (5) in benzene resulted in the clean formation of PhH, Ph₃PF₂, Pd metal, and [(Ph₃P)₃Pd].

Full text of DOI:10.1021/om0001265

1888
Organometallics 2000, 19, 1888-1900
Th er m a l Sta bility, Decom p osition P a th s, a n d P h /P h
Exch a n ge Rea ction s of [(P h 3P )2P d (P h )X] (X ) I, Br , Cl, F ,
a n d HF 2)
Vladimir V. Grushin
Central Research and Development, E. I. DuPont de Nemours and Co. Inc.,† Experimental
Station, Wilmington, Delaware 19880-0328
Received February 10, 2000
Complexes of the type [(Ph3P)2Pd(Ph)X], where X ) I (1), Br (2), Cl (3), F (4), and HF2 (5),
possess different thermal stability and reactivity toward the Pd-Ph/P-Ph exchange
reactions. While 1 decomposed (16 h) in toluene at 110 °C to [Ph4P]I, Pd metal, and Ph3P,
complexes 2 and 3 exhibited no sign of decomposition under these conditions. Kinetic studies
of the aryl-aryl exchange reactions of [(Ph3P)2Pd(C6D5)X] in benzene-d6 demonstrated that
the rate of exchange decreases in the order 1 > 2 > 3, the observed rate constant ratio,
kI:kBr:kCl, in benzene at 75 °C being ca. 100:4:1 for 1-d5, 2-d5, and 3-d5. The exchange was
facilitated by a decrease in the concentration of the complex, polar media, and a Lewis acid,
e.g., Et2O‚BF3. Unlike [Bu4N]PF6, which speeded up the exchange reaction of 2-d5, [Bu4N]-
Br inhibited it due to the formation of anionic four-coordinate [(Ph3P)Pd(C6D5)Br2]-. The
latter and its iodo analogue were generated in dichloromethane and benzene upon addition
of [Bu4N]X or PPN Cl to [(Ph3P)2Pd2(Ph)2(µ-X)2] (X ) I, Br, or Cl) and characterized in solution
1
by H and 31P NMR spectral data. The mechanism of the aryl-aryl exchange reactions of
[(Ph3P)2Pd(C6D5)X] in noncoordinating solvents of low polarity may not require Pd-X
ionization but rather involves phosphine dissociation, the ease of which decreases in the
order X ) I > Br > Cl, as suggested by crystallographic data. Two mechanisms govern the
thermal reactions of [(Ph3P)2Pd(Ph)F], 4. One of them is similar to the aryl-aryl exchange
and decomposition path for 1-3, involving a tight ion pair intermediate, [Ph4P][(Ph3P)PdF],
within which two processes were shown to occur. At 75 °C, the C-P oxidative addition
restores the original neutral complex (4). At 90 °C, reversible fluoride transfer from Pd to
the phosphonium cation resulted in the formation of covalent [Ph4PF] and [(Ph3P)Pd], which
was trapped by PhI to produce [(Ph3P)2Pd2(Ph)2(µ-I)2]. The other decomposition path of 4
leads to the formation of [(Ph3P)3Pd], Pd, Ph2, Ph3PF2, and Ph2P-PPh2 as main products.
Unlike the aryl-aryl exchange, this decomposition reaction is not inhibited by free phosphine.
The formation of biphenyl was shown to occur due to PdPh/PPh coupling on the metal center.
Mechanisms accounting for the formation of these products are proposed and discussed.
The facile (4 h at 75 °C) thermal decomposition of [(Ph3P)2Pd(Ph)(FHF)] (5) in benzene
resulted in the clean formation of PhH, Ph3PF2, Pd metal, and [(Ph3P)3Pd].
In tr od u ction
may be critical for successful use and further develop-
ment of the catalytic chemistry of aryl halides. In this
paper, we report on the thermal stability and decom-
position paths of palladium aryls containing the most
conventional Ph3P and different halogen ligands. Some
mechanistic aspects of the closely related aryl-aryl
exchange reactions of such complexes were also studied
and are presented in this work.
Complexes of the general formula [(R3P)2Pd(Ar)X]
(X ) halogen) and related organopalladium compounds
are key intermediates in numerous Pd-catalyzed reac-
tions of aryl halides, such as the Heck arylation of
olefins, the Kharasch-Fields, Stille, Suzuki, and Sono-
gashira coupling reactions, and various nucleophilic
displacement and carbonylation reactions.1 High tem-
peratures (100-200 °C) are frequently required for
conducting some of these reactions, e.g., when inert ArCl
substrates are employed.1c Quite often, however, pre-
cipitation of Pd metal and, as a consequence, fast
termination of the catalytic process are observed at
elevated temperatures needed for the reaction to occur
at a satisfactory rate. Identifying mechanisms that
govern the thermal decomposition of [(R3P)2Pd(Ar)X]
Resu lts a n d Discu ssion
Th er m a l Sta bility of [(P h 3P )2P d (P h )X] (X ) I, Br ,
a n d Cl). Surprisingly little has been reported on the
thermal stability and decomposition patterns of these
complexes.2 It has been shown that at temperatures
above 50 °C [(Ph3P)2Pd(Ar)I],3,4 [(Ph3P)2Pd(Ar)Br],5 and
[(Ph3P)2Pd(Ar)Cl]6 undergo the peculiar Pd-aryl/P-aryl
exchange (eq 1) that likely occurs via a tetraarylphos-
phonium intermediate involved in the C-P reductive
† Contribution No. 7972.
10.1021/om0001265 CCC: $19.00 © 2000 American Chemical Society
Publication on Web 04/13/2000

Thermal Stability of [(Ph3P)2Pd(Ph)X]
Organometallics, Vol. 19, No. 10, 2000 1889
Sch em e 1
elimination-oxidative addition reaction sequence
(Scheme 1).4,5 As a result of this exchange, the desired
product of a Pd-catalyzed reaction is often contaminated
with a similar compound derived from the aryl originally
bound to the phosphorus and eventually transferred
to the metal during the course of the catalytic run
(Scheme 1).4,5,7
(Ph)I], 1, in refluxing dichloromethane (40 °C).9 How-
ever, no formation of tetraarylphosphonium salts was
observed when 1 and similar complexes were heated in
THF-d8 or CDCl3 at higher temperatures (50-60 °C).3,4,6
Goodson, Wallow, and Novak4 pointed to the remarkable
ability of chlorinated solvents to promote the formation
of phosphonium salts from [(Ph3P)2Pd(Ar)I],10 but no
rationale for this effect was offered.4 Vicente and co-
workers9b,c also reported that the formation of phospho-
nium cations from cyclopalladated aromatics was very
facile in dichloromethane but not in acetone.
Sakamoto, Shimizu, and Yamamoto8 reported the
formation of [Ph4P]I in ca. 30% yield from [(Ph3P)2Pd-
(1) For selected recent reviews and monographs, see: (a) Negishi,
E.-I. In Organozinc Reagents; Knochel, P., J ones, P., Eds.; Oxford
University Press: Oxford, 1999; p 213. (b) Yang, B. H.; Buchwald, S.
L. J . Organomet. Chem. 1999, 576, 125. Wolfe, J . P.; Wagaw, S.;
Marcoux, J .-F.; Buchwald, S. L. Acc. Chem. Res. 1998, 31, 805. (c)
Grushin, V. V.; Alper, H. Topics Organomet. Chem. 1999, 3, 193. (d)
Stanforth, S. P. Tetrahedron 1998, 54, 263. (e) Herrmann, W. A.;
Reisinger, C.-P. In Aqueous-Phase Organometallic Catalysis; Cornils,
B., Herrmann, W. A., Eds.; Wiley-VCH: Weinheim, 1998; p 383. (f)
Farina, V.; Krishnamurthy, V.; Scott, W. J . The Stille Reaction;
Wiley: New York, 1998; Org. React. 1997, 50, 1. (g) Hartwig, J . F.
Acc. Chem. Res. 1998, 31, 852; Angew. Chem., Int. Ed. 1998, 37, 2046;
Curr. Org. Chem. 1997, 1, 287; Synlett 1997, 329. (h) Mitchell, T. N.
In Metal-Catalyzed Cross-Coupling Reactions; Diederich, F., Stang, P.
J ., Eds.; Wiley-VCH: Weinheim, 1998; p 167. (i) Browning, A. W.;
Greeves, N. In Transition Metals in Organic Synthesis; Gibson, S. E.,
Ed.; Oxford University Press: Oxford, 1997; p 35. (j) Malleron, J .-L.;
Fiaud, J .-C.; Legros, J .-Y. Handbook of Palladium-Catalyzed Organic
Reactions. Synthetic Aspects and Catalytic Cycles; Academic Press:
London, 1997. (k) Stephenson, G. R. In Advances in Asymmetric
Synthesis; Stephenson, G. R., Ed.; 1996; p 275. (l) Percec, V.; Hill, H.
H. ACS Symp. Ser. 1996, 624, 2. (m) Tsuji, J . Palladium Reagents and
Catalysis: Innovations in Organic Synthesis; Wiley: Chichester, 1995.
(n) Miyaura, N.; Suzuki, A. Chem. Rev. 1995, 95, 2457. (o) Cabri, W.;
Caudiani, I. Acc. Chem. Res. 1995, 28, 2. (p) Tsuji, J .; Mandai, T.
Angew. Chem., Int. Ed. Engl. 1995, 34, 2589. (q) Rossi, R.; Carpita,
A.; Bellina, F. Org. Prep. Proc. Int. 1995, 27, 129.
(2) (a) The thermal decomposition of ethyl Pd complexes, [(Me3P)2-
Pd(Et)X], involves â-hydrogen elimination.2b This decomposition path
is obviously impossible for Pd aryls. (b) Kawataka, F.; Kayaki, Y.;
Shimizu, I.; Yamamoto, A. Organometallics 1994, 13, 3517.
(3) Kong, K.-C.; Cheng, C.-H. J . Am. Chem. Soc. 1991, 113, 6313.
(4) Goodson, F. E.; Wallow, T. I.; Novak, B. M. J . Am. Chem. Soc.
1997, 119, 12441.
(5) (a) Segelstein, B. E.; Butler, T. W.; Chenard, B. L. J . Org. Chem.
1995, 60, 12. (b) Hartwig and co-workers observed similar aryl-aryl
exchange reactions when studying C-S reductive elimination from
[LPd(Ar)SR], where L ) bidentate phenylphosphine, e.g., dppe. See:
Baranano, D.; Hartwig, J . F. J . Am. Chem. Soc. 1995, 117, 2937. Mann,
G.; Baranano, D.; Hartwig, J . F.; Rheingold, A. L.; Guzei, I. A. J . Am.
Chem. Soc. 1998, 120, 9205.
We found that [(Ph3P)2Pd(Ph)I] slowly (16 h) decom-
posed in toluene under nitrogen at 110 °C (oil bath) to
give Pd metal, triphenylphosphine, and tetraphenylphos-
phonium iodide in high yield (eq 2).
(2)
As no dibenzyl was found among the reaction products
(GC-MS), a free-radical mechanism for the reaction can
be ruled out. Remarkably, no sign of decomposition was
observed after [(Ph3P)2Pd(Ph)Br], 2, and [(Ph3P)2Pd(Ph)-
Cl], 3, were heated at the same temperature for the
same period of time (1H and 31P NMR).11 Because the
thermal decomposition and aryl-aryl exchange reac-
tions of [(Ph3P)2Pd(Ar)X] are closely related (Scheme 1),
the significant difference in the thermal stability of the
iodo complex 1 and its bromo and chloro counterparts
2 and 3 (eq 2) prompted us to study the dependence of
aryl-aryl exchange rate (eq 1) on the nature of X.
(9) (a) Vicente et al.9b-d have reported the formation of tetraarylphos-
phonium cations as a result of exceedingly facile Ar-PPh3 reductive
elimination from cyclopalladated aromatic compounds. (b) Vicente, J .;
Arcas, A.; Bautista, D.; Tiripicchio, A.; Tiripicchio-Camellini, M. New
J . Chem. 1996, 20, 345. (c) Vicente, J .; Arcas, A.; Bautista, D.; J ones,
P. G. Organometallics 1997, 16, 2127. (d) Vicente, J .; Abad, J .-A.;
Frankland, A. D.; Ram´ırez de Arellano, M. C. Chem. Eur. J . 1999, 5,
3066.
(6) Herrmann, W. A.; Brossmer, C.; Priermeier, T.; O¨ fele, K. J .
Organomet. Chem. 1994, 481, 97.
(7) Herrmann, W. A.; Brossmer, C.; O¨ fele, K.; Beller, M.; Fischer,
H. J . Organomet. Chem. 1995, 491, C1; J . Mol. Catal. A: Chem. 1995,
103, 133. O’Keefe, D. F.; Dannok, M. C.; Marcuccio, S. M. Tetrahedron
Lett. 1992, 33, 6679. Kikukawa, K.; Takagi, M.; Matsuda, T. Bull.
Chem. Soc. J pn. 1979, 52, 1493. Kikukawa, K.; Yamane, T.; Ohbe, Y.;
Takagi, M.; Matsuda, T. Bull. Chem. Soc. J pn. 1979, 52, 1187.
Kawamura, T.; Kikukawa, K.; Takagi, M.; Matsuda, T. Bull. Chem.
Soc. J pn. 1977, 50, 2021. Fahey, D. R.; Mahan, J . E. J . Am. Chem.
Soc. 1976, 98, 4499. Yamane, T.; Kikukawa, K.; Takagi, M.; Matsuda,
T. Tetrahedron 1973, 29, 955. Asano, R.; Moritani, I.; Fujiwara, Y.;
Teranishi, S. Bull. Chem. Soc. J pn. 1973, 46, 2910. Kikukawa, K.;
Yamane, T.; Takagi, M.; Matsuda, T. J . Chem. Soc., Chem. Commun.
1972, 695.
(10) In the presence of extra phosphine, however, the formation of
tetraarylphosphonium salts from [(Ar3P)2Pd(Ar′)I] occurred in THF-
d8 at 50° C.4 For the Pd-catalyzed formation of tetraarylphosphonium
salts from haloarenes and triarylphosphine, see: Ziegler, C. B.; Heck,
R. F. J . Org. Chem. 1978, 43, 2941. Migita, T.; Nagai, T.; Kiuchi, K.;
Kosugi, M. Bull. Chem. Soc. J pn. 1983, 56, 2869.
(11) Herrmann et al.6 mentioned that [(Ph3P)2Pd(Ph)Cl] decomposed
very slowly in toluene at 120° C to produce small quantities of Pd
metal. The reaction time was not specified, and other products were
not reported.6
(8) Sakamoto, M.; Shimizu, I.; Yamamoto, A. Chem. Lett. 1995, 1101.

1890 Organometallics, Vol. 19, No. 10, 2000
Grushin
F igu r e 1. Aryl-aryl exchange reaction of [(Ph3P)2Pd(C6D5)Br] (2-d5) in benzene-d6 (C ) 1.1 × 10-2 M) at 75 °C, as monitored
1
by H NMR spectroscopy. Bottom: t ) 0 min. Top: t ) 540 min. Time to full Pd-C6D5/Pd-C6H5 equilibration ) 7 h. The
residual solvent peak appears as a singlet resonance at 7.1 ppm.
Obtaining this information was important because in
the catalytic reactions,1 intermediates [(Ph3P)2Pd(Ar)X]
are formed from the corresponding ArX, of which the
halogen X can be varied if it helps avoid or suppress
the undesirable aryl-aryl exchange processes.
Ar yl-Ar yl Exch a n ge Rea ction s of [(P h 3P )2P d -
(C6D5)X] (X ) I, Br , a n d Cl). Model compounds for
these studies were [(Ph3P)2Pd(C6D5)X], where X ) I (1-
d5), Br (2-d5), and Cl (3-d5). Complexes 1-d5 and 2-d5
were prepared by the oxidative addition reaction of
C6D5I and C6D5Br, respectively, to [(Ph3P)4Pd] (eqs 3
and 4).12,13 The chloro complex 3-d5 was synthesized via
the biphasic halide-halide exchange14 between 1-d5 and
KCl (eq 5).
All exchange reactions were run under nitrogen in
standard 5-mm NMR tubes with rigorously anhydrous,
oxygen-free solvents. After the tubes were placed in a
thermostated oil bath ((1 °C) or the VT probe of an
by integrating the signals at 6.3 ppm (2H, m-C6H5Pd)
and 7.7 ppm (12H, o-C6H5P). Results of these studies
are summarized in Table 1. It was found that the
exchange was more facile for dilute rather than con-
centrated solutions of the complexes; for example, at 75
°C in benzene-d6, the C6D5/C6H5 exchange in 2-d5 was
NMR spectrometer, the exchange was frequently moni-
tored by 1H NMR spectroscopy. In benzene-d6, the
proton Pd-C6H5 resonances appeared at 6.3 (2H, meta),
6.45 (1H, para), and 6.9 (2H, ortho) ppm, whereas two
multiplet signals at 7.0 (18H, meta and para) and 7.7
(12H, ortho) were from the phenylphosphine ligands. As
the exchange occurred, the originally absent Pd-C6H5
(12) The reaction of C6D5I occurred rapidly and cleanly at room
temperature. To avoid the aryl-aryl exchange during the addition of
less reactive13 C6D5Br to the Pd(0) complex at 75 °C, the reaction was
run in the presence of 2 equiv of Ph3P. It has been demonstrated3,4
that the addition of free phosphine ligand suppresses the aryl-aryl
exchange reactions of [(R3P)2Pd(Ar)I].
(13) Fitton, P.; Rick, E. A. J . Organomet. Chem. 1971, 28, 287.
(14) Flemming, J . P.; Pilon M. C.; Borbulevitch O. Ya.; Antipin M.
Yu.; Grushin, V. V. Inorg. Chim. Acta 1998, 280, 87.
resonances first appeared and then grew in intensity
until the Pd-C6D5 and Pd-C6H5 species equilibrated
(Figure 1). At that point, the integral intensity ratio of
the Ph-P to Ph-Pd resonances was 6, indicating
complete scrambling of the C6D5 group over the Pd and
P centers; most accurate measurements were obtained

Thermal Stability of [(Ph3P)2Pd(Ph)X]
Organometallics, Vol. 19, No. 10, 2000 1891
Ta ble 1. C6H5-C6D5 Exch a n ge Rea ction s of [(P h 3P )2P d (C6D5)X], w h er e X ) I (1-d 5), Br (2-d 5), a n d Cl (3-d 5),
a t 75 °C (C ) 1.1 × 10-2 M)
entry
no.
time to Pd-C6D5/Pd-C6H5
X
solvent
additive (equiv)
equilibration (h)
decomposition
1
2
I
benzene-d6
benzene-d6
DMF-d7
none
0.3
none
none
15%
none
none
10%
none
none
none
none
none
35%
5%
I
[Bu4N]+I- (1)
none
0.5
3
I
<0.1
4
Br
Br
Br
Br
Br
Br
Br
Br
Br
Br
Cl
Cl
benzene-d6
benzene-d6
benzene-d6
benzene-d6
benzene-d6
benzene-d6
benzene-d6
benzene-d6
DMSO-d6
DMF-d7
none
7
5
Et2O‚BF3 (0.01)
Et2O‚BF3 (1)
[Bu4N]+PF6- (0.5)
[Bu4N]+PF6- (1.5)
[Bu4N]+PF6- (3)
[Bu4N]+Br- (3)
[Bu4N]+I- (1.5)
none
0.1
6
exchanges at 25 °C
7
0.7
0.6
0.3
15
8
9
10
11
12
13
14
15
2
<1
0.2
30
none
benzene-d6
DMSO-d6
none
none
15%
none
<1
complete after 4 and 10 h for [2-d5] ) 3.6 × 10-3 and
4.6 × 10-2 M, respectively.15 Each of these and all other
experiments (Table 1) were successfully reproduced 2-6
times. Because of the concentration effect observed, all
exchange reactions listed in Table 1 were run with a
standard concentration of [[(Ph3P)2Pd(C6D5)X]] ) 1.1 ×
10-2 M.
Lewis acid, Et2O‚BF3 (Table 1, entries 4, 5, and 6). In
benzene-d6 at 75 °C, the reaction proceeded ca. 40 times
faster when only 0.01 equiv of Et2O‚BF3 was added. The
addition of an equimolar amount of Et2O‚BF3 resulted
in an even more pronounced effect; that is, 2-d5 began
to exchange its phenyls at ambient temperature. After
1 h at 25 °C, more than 50% of the C6D5 groups in 2-d5
were replaced with the phenyls of the PPh3 ligands. The
31P NMR spectrum of the sample exhibited singlets at
23.0, 23.1, and 23.2 ppm from different isotopomers of
deuterated 2 and two minor resonances at 21.9 and 22.0
ppm, presumably from two isotopomers of the orange
cationic complex formulated as [(Ph3P)2PdPh]+ (δ )
21.7).16 This spectral line assignment was corroborated
by the slow formation of a microcrystalline orange
precipitate during the experiment.
Consistent with the salt effect reported by Goodson,
Wallow, and Novak,4 the exchange reaction of 2-d5 in
benzene-d6 solutions speeded up considerably in the
presence of tetrabutylammonium hexafluorophosphate.17
Faster rates were observed with higher concentrations
of [Bu4N]PF6, as exemplified by entries 4 and 7-9 in
Table 1. For instance, in the presence of 3 equiv of
[Bu4N]PF6, the reaction was approximately 20 times
faster (entries 4 and 9). However, when this experiment
was repeated with [Bu4N]Br instead of [Bu4N]PF6
(entries 9 and 10), a retardation effect rather than
acceleration of the exchange was observed. In the
presence of [Bu4N]I, the exchange reaction of 1-d5 was
only slightly slower (entry 2). Importantly, tetrabutyl-
ammonium iodide also promoted the exchange reaction
of the bromo complex 2-d5, albeit not as efficiently as
the hexafluorophosphate salt (entries 4, 8, and 11).
Due to the multiple exchange (one Ph on Pd and six
on the P ligands) reliable and precise kinetic data for
this reaction can be obtained only at low conversions.
As it is clear that sufficiently accurate integration (1H
NMR) is impossible at such conversions, the alternative
“time to equilibration” technique was employed. Al-
though absolute rate constants cannot be obtained this
way, the design of all experiments (see above) allowed
for the reliable quantitative determination of the rate
constant ratio, a critical parameter needed to gain
insight into the mechanism of the thermal decomposi-
tion and aryl-aryl exchange reactions of the Pd aryls.
As seen from Table 1, the rate of exchange is depend-
ent on the nature of X, the observed rate constant ratio,
kI:kBr:kCl, in benzene being ca. 100:4:1 for 1-d5, 2-d5, and
3-d5, respectively (entries 1, 4, and 14). No sign of any
reaction other than the aryl-aryl exchange was noticed
in benzene at 75 °C. In contrast, after the iodo complex
1-d5 was heated in CDCl3 at 60 °C for 40 min, ca. 5-10%
of the complex decomposed to produce [Ph4P]+ and small
amounts of [(Ph3P)2PdI2] (31P NMR: 12.1 ppm) among
several other products. The initial exchange rate in
chloroform was faster than in benzene at the same
temperature (60 °C), but as the decomposition in CDCl3
took place, the exchange reaction decelerated consider-
ably. As a result, after 2 h 40 min, 1-d5 had not yet fully
equilibrated in chloroform, whereas complete Ph-d5
scrambling was observed for the benzene sample. The
exchange reactions in DMF-d7 and DMSO-d6 (Table 1,
entries 3, 12, 13, and 15) were considerably faster than
in benzene; side reactions always occurred when DMF
and DMSO solvents were used.
Mech a n ism of th e Ar yl-Ar yl Exch a n ge in Med ia
of Low P ola r it y. An ion ic Com p lexes [(P h 3P )P d -
(P h )X2]-. The aryl-aryl exchange reaction of [(Ar3P)2-
Pd(Ar′)X] (eq 1) involves C-P reductive elimination of
[Ar3PAr′]+ from the metal center, followed by oxidative
addition of the phosphonium cation back to the Pd(0)
species generated (Scheme 1).4,5 Evidence has been
presented for phosphine dissociation3,4 and/or Pd-X
ionization4 being the first key step of the exchange.
Either process would result in coordinative unsaturation
of the Pd(II) center, thus promoting C-P reductive
A significant acceleration of the exchange for the
bromo complex 2-d5 was observed in the presence of a
(15) It has been reported4 that the aryl-aryl exchange reaction of
{[(4-FC6H4)3P]2Pd(I)(4-C6H4OMe)} was first order in the complex. This
exchange reaction was run in THF-d8, a much more coordinating
solvent in comparison with benzene-d6 employed in our studies. In
addition, the use of dibromomethane as the internal standard4 may
not be justified for studying kinetics of reactions mediated by highly
reactive, electron-rich, coordinatively unsaturated Pd(0) species. Such
zerovalent complexes react with chlorinated solvents (see ref 4 and
main text above and below), which are less electrophilic than CH2Br2.
(16) Amatore, C. A.; Carre´, E.; J utand, A.; M’Barki, M. A.; Meyer,
G. Organometallics 1995, 14, 5605.
(17) [Bu4N]+PF6-, [Bu4N]+Br-, and [Bu4N]+I- are all moderately
soluble in benzene at 75 °C (5-10 mg/mL).

1892 Organometallics, Vol. 19, No. 10, 2000
Grushin
elimination.18 While the addition of free phosphine
strongly inhibits the aryl-aryl exchange reactions of
[(Ar3P)2Pd(Ar′)I]3,4 and [(Ph3P)2Pd(C6D5)F] (see below),
the decelerating effect of extra halide is significantly less
pronounced (see ref 4 and Table 1) and insufficiently
understood.
expected for KD of 1, 2, and 3 in benzene or toluene
solutions. These considerations suggest that Pd-X
ionization is unlikely a requirement for the aryl-aryl
exchange reactions of 1-3 to occur in media of low
polarity, such as benzene or toluene. If so, the two
questions arise: (1) Why does the exchange rate depend
on the nature of halogen X? and (2) Why is the exchange
reaction of [(Ph3P)2Pd(Ph)X] in benzene speeded up in
the presence of [Bu4N]PF6 but inhibited by [Bu4N]X?
The mechanism shown in Scheme 2 does not involve
the Pd-X ionization step while readily accounting for
all observations made in this and previous3,4 works. It
is to be stressed that the concept of Pd-X ionization is
not challenged when the exchange is run in the presence
of a Lewis acid or in polar media, where the Pd-X bond
can ionize more easily. On the other hand, the catalytic
effect of BF3 observed in our studies (see above) might
be due to its complexation with both the halide and
phosphine.19c
Comparison of the P-Pd bond distances in the crystal
structures of 1 ([(Ph3P)2Pd(Ph)I]; P-Pd ) 2.342(1) and
2.338(1) Å), 2 ([(Ph3P)2Pd(Ph)Br]; P-Pd ) 2.327(2) and
2.322(2) Å), and 3 ([(Ph3P)2Pd(Ph)Cl]; P-Pd ) 2.324(1)
and 2.316(1) Å)14 predicts that ease of phosphine dis-
sociation from these complexes and hence the aryl-aryl
exchange efficiency would follow the trend 3 < 2 < 1,
other things being equal. This is exactly the order that
was observed (Table 1). The variation in the Pd-P bond
distances for complexes 1-3 is probably best rational-
ized in terms of different field effects (electronegativi-
ties) of the halogens.
Phosphine dissociation from [(Ph3P)2Pd(Ph)X] pro-
duces tricoordinate [(Ph3P)Pd(Ph)X], which then under-
goes reversible reductive elimination of tetraphenylphos-
phonium cation.4,5a The tricoordinate intermediate
probably exists in equilibrium with its dimer, [(Ph3P)2-
Pd2(Ph)2(µ-X)2]. It has been shown8 that [(Ph3P)2Pd2-
(Ph)2(µ-I)2], a better precursor for [(Ph3P)Pd(Ph)I],
decomposes in CH2Cl2 more readily than [(Ph3P)2Pd-
(Ph)I], to give [Ph4P]I in a much higher, almost quan-
titative yield. The formation of the key tricoordinate
intermediate [(Ph3P)Pd(Ph)X] should be favored at
lower concentrations of [(Ph3P)2Pd(Ph)X], as suggested
by the equilibria between the two and also between
[(Ph3P)Pd(Ph)X] and its dimer (Scheme 2). Indeed, we
observed that the exchange reaction of 2-d5 proceeded
more readily at lower concentrations of the complex (see
above).
We proposed that in the presence of [Bu4N]X the
reductive elimination of [Ph4P]X from the tricoordinate
complex [(Ph3P)Pd(Ph)X] might be inhibited owing to
its reaction with the halide of the ammonium salt, to
form saturated, square-planar [(Ph3P)Pd(Ph)X2]-.25 To
check this assumption, reactivity of [(Ph3P)2Pd2(Ph)2-
(µ-X)2] toward [Bu4N]X (X ) Br or I) was studied. We
found that while both [(Ph3P)2Pd2(Ph)2(µ-I)2] and [(Ph3P)2-
Pd2(Ph)2(µ-Br)2] are poorly soluble in benzene and
Our results demonstrate that the rate of the aryl-
aryl exchange reaction of [(Ph3P)2Pd(C6D5)X] is depend-
ent on the nature of X trans to the σ-aryl, decreasing in
the order I > Br > Cl (Table 1). Given the fact that polar
media promote the reaction, it would be tempting to
propose Pd-X ionization to be a requirement for the
exchange to occur.4 Indeed, the exchange reaction of 2-d5
was faster in the presence of a Lewis acid, Et2O‚BF3
(Table 1, entries 5 and 6), possibly due to the Pd-Br
bond ionization caused by the boron trifluoride.19 Fur-
thermore, it has been demonstrated that in polar
solvents such as DMF the Pd-X bond is ionized, with
the affinity of the metal center for X increasing in the
order I < Br < Cl.16,20 The same order for the binding
ability of the halide anions to [(Ph3P)2Pd(Ph)]+ (I < Br
< Cl < F) has been observed for significantly less polar
solvents (CH2Cl2 and CHCl3).14,21 However, in such
solvents and especially in benzene and toluene, which
were used in this work, substantial ionization of the
Pd-X bond is unlikely.22 The values of 5.9 × 10-4, 2.8
× 10-4, and 5 × 10-5 mol dm-1 have been obtained for
Pd-X dissociation constants (KD) of 1, 2, and 3, respec-
tively, in DMF at 293 K.20 In benzene, however, these
complexes show much less ionization.22 For comparison,
ionic tetrabutylammonium salts are 2-4 orders of
magnitude more dissociated in DMF23 than the Pd-X
bond of 1, 2, and 3.20 In benzene, however, [Bu4N]X are
strongly associated, forming large clusters consisting of
tight ion pairs.24 For instance, Everaert and Persoons24b
estimated KD for tetrabutylammonium picrate in ben-
zene to be ca. 6.5 × 10-18 at 298 K. Therefore, at least
a few orders of magnitude lower values might be
(18) Reductive elimination reactions are often facilitated by ligand
predissociation. See: Collman, J . P.; Hegedus, L. S.; Norton, J . R.;
Finke, R. G. Principles and Applications of Organotransition Metal
Chemistry; University Science Books: Mill Valley, CA, 1987; p 322.
Crabtree, R. H. The Organometallic Chemistry of the Transition Metals,
2nd ed.; J ohn Wiley and Sons: New York, 1994; p 151. Spessard, G.
O.; Miessler, G. L. Organometallic Chemistry; Prentice Hall: Upper
Saddle River, NJ , 1996; p 178.
(19) (a) It is noteworthy that Lewis acids can abstract a tertiary
phosphine from its Pd complexes. Kayaki, Shimizu, and Yamamoto19b
have demonstrated that the addition of 1 equiv of AgBF4 to [(Me3P)2-
Pd(R)X] furnishes [(Me3P)2Pd(R)(solvent)]+, whereas [(Me3P)Pd(R)-
(solvent)2]+ is formed when excess AgBF4 is employed. The formation
of a 1:1 complex of Ph3P with BF3 is well-documented.19c (b) Kayaki,
Y.; Shimizu, I.; Yamamoto, A. Bull. Chem. Soc. J pn 1997, 70, 1135.
(c) Spitsyn, V. I.; Kolli, I. D.; Sadykova, E. M.; Nesterenko, R. G. Izv.
Akad. Nauk SSSR, Ser. Khim. 1972, 45. Romm, I. P.; Sadykova, E.
M.; Gur’yanova, E. N.; Kolli, I. D. Zh. Obshch. Khim. 1973, 43, 728.
Fluck, E.; Weber, D. Z. Naturforsch. B 1974, 29, 603. Muylle, E.; Van
Der Kelen, G. P.; Claeys, E. G. Spectrochim. Acta A 1976, 32, 1149.
(20) Amatore, C.; Carre´, E.; J utand, A. Acta Chem. Scand. 1998,
52, 100.
(21) Grushin, V. V. Angew. Chem., Int. Ed. Engl. 1998, 37, 994.
(22) (a) J utand and Mosleh22b have reported that triflato complexes
[(Ph3P)2Pd(Ar)OTf] exhibited no conductivity in toluene, while being
ion-paired. (b) J utand, A.; Mosleh, A. Organometallics 1995, 14, 1810.
(23) (a) Szejgis, A.; Bald, A.; Gregorowicz, J . Phys. Chem. Liq. 1997,
35, 165, and references therein. (b) Islam, N.; Zaidi, S. B. A.; Ansari,
A. A. Bull. Chem. Soc. J pn 1989, 62, 309. (c) Tsentovskii, V. M.;
Mochalov, K. N.; Barabanov, V. P.; Tumasheva, N. A. Zh. Obshch.
Khim. 1974, 44, 1338. (d) Levitskaya, N. K.; Tkachenko, L. I.; Shkodin,
A. M. Elektrokhimiya 1974, 10, 580.
(24) (a) Ivashkevich, A. N.; Kostynyuk, V. P. Zh. Fiz. Khim. 1991,
65, 948, and references therein. (b) Everaert, J .; Persoons, A. J . Phys.
Chem. 1981, 85, 3930. (c) Barker, C.; Yarwood, J . Faraday Symp.
Chem. Soc. 1977, 136.
(25) (a) A closely related anionic Pd complex, [(Ph3P)PdCl3]-,25b-f
forms reversibly upon addition of [Bu4N]+Cl- to [(Ph3P)2Pd2(Cl)2(µ-
Cl)2]25b or [(Ph3P)2PdCl2].25f (b) Goodfellow, R. J .; Goggin, P. L.; Duddell,
D. A. J . Chem. Soc. A 1968, 504. (c) Dillon, K. B.; Waddington, T. C.;
Younger, D. J . Chem. Soc., Dalton Trans. 1975, 790. (d) Bardi, R.;
Piazzesi, A. M.; Del Pra, A.; Cavinato, G.; Toniolo, L. Inorg. Chim. Acta
1983, 75, 15. (e) Tollari, S.; Demartin, F.; Cenini, S.; Palmisano, G.;
Raimondi, P. J . Organomet. Chem. 1997, 527, 93. (f) Amatore, C.;
J utand, A.; Mottier, L. Eur. J . Inorg. Chem. 1999, 1081.

Thermal Stability of [(Ph3P)2Pd(Ph)X]
Organometallics, Vol. 19, No. 10, 2000 1893
Sch em e 2
dichloromethane, their solubility in these solvents in-
creased dramatically upon addition of [Bu4N]I and
[Bu4N]Br, respectively. This increase in the solubility
was accompanied with noticeable 31P and 1H NMR
spectral changes, indicating the reversible formation of
a single new species. For instance, as a solution of
[Bu4N]Br in CD2Cl2 was added portionwise to [(Ph3P)2-
Pd2(Ph)2(µ-Br)2] in CD2Cl2, the sharp 31P NMR singlet
from the dimer (30.4 ppm) gradually diminished in
intensity and eventually disappeared with the concomi-
tant appearance and growth of a new sharp singlet
of [Bu4N]X (X ) Br or I) were not due to the cis-trans
(syn-anti) isomerization of the dimer (eq 6)28,29 caused
by an increase in the polarity of the medium, but rather
due to the formation of the anionic complexes (eq 7).
The close 31P NMR chemical shifts of the [(Ph3P)Pd-
(Ph)X2]- and their dinuclear precursors [(Ph3P)2Pd2-
(Ph)2(µ-X)2] point to the cis geometry for the anions.
Square-planar Pd(II) complexes with an aryl ligand
trans to R3P are often unstable and therefore rare.9c
(26) When repeating the reaction between [Bu4N]+ I- and [(Ph3P)2-
1
Pd2(Ph)2(µ-I)2] in CD2Cl2, we observed, on
a few occasions, the
resonance at 29.9 ppm. The H NMR pattern exhibited
appearance of a weak singlet resonance at 22.3 ppm in the 31P NMR
spectrum and a very weak triplet proton resonance at 6.3 ppm, both
parameters being consistent with the formation of small quantities
(ca. 10%) of [(Ph3P)2Pd(Ph)I]. The yield of this product did not change
upon addition of larger quantities of [Bu4N]+I- to the reaction solution,
which suggested that it formed due to a side reaction. The reaction
between [Bu4N]+Br- and [(Ph3P)2Pd2(Ph)2(µ-Br)2], which was also
repeated several times, never gave rise to any side products.
by the resulting solution in the aromatic region was
clearly distinct from that of the starting dimer (Figure
2). Analogous changes were observed when this experi-
ment was repeated in benzene-d6. Likewise, the iodo
dimer underwent conversion to another species when
dissolved in CD2Cl2 or benzene-d6 in the presence of
[Bu4N]I (Table 2).26 Importantly, [Bu4N]PF6 was found
to have no influence on the solubility of the dinuclear
complexes nor NMR spectral parameters of their solu-
tions.27 Therefore, the changes observed in the presence
(27) (a) No reaction was observed when [(Ph3P)2Pd2(Ph)2(µ-Br)2] was
treated with [Bu4N]+PF6 in the absence or presence of catalytic
-
quantities (5%) of [Bu4N]+X- to promote the cis-trans isomerization.27b
(b) Anderson, G. K.; Cross, R. J . Chem. Soc. Rev. 1980, 9, 185.
(28) Anderson, G. K. Organometallics 1983, 2, 665.
(29) Grushin, V. V.; Alper, H. Organometallics 1993, 12, 1890.

1894 Organometallics, Vol. 19, No. 10, 2000
Grushin
While readily establishing within the time of mixing,
equilibrium 7 is slow on the NMR time scale at room
temperature, allowing for equilibrium constant mea-
surements, using 31P NMR spectroscopy. Thus, in CD2-
Cl2 at 25 °C, Keq was estimated to be 10 and 0.2 mol-1
dm3 for X ) Br and I, respectively, with the initial
concentrations of the reagents used being [(Ph3P)2Pd2-
(Ph)2(µ-Br)2] ) 1.2 × 10-3, [Bu4NBr] ) 9.5 × 10-3
,
[(Ph3P)2Pd2(Ph)2(µ-I)2] ) 4.3 × 10-3, and [Bu4NI] ) 6.5
× 10-2 M. Similar observations were made with solu-
tions of [(Ph3P)2Pd2(Ph)2(µ-Cl)2] and PPNCl in dichlo-
romethane; Keq ) 11 mol-1 dm3 for X ) Cl was
measured, using [(Ph3P)2Pd2(Ph)2(µ-Cl)2] and PPNCl in
CD2Cl2 at the initial concentrations of 1.7 × 10-3 and
8.0 × 10-2 M, respectively. Since the complexes [Bu4N]X
(X ) Br, I) and PPNCl involved in eq 7 are strongly
associated in low polar solvents, the equilibrium con-
stants were calculated in the approximation that all
activity coefficients were close to unity. In accord with
the order of halide anion affinity for Pd(II) in media of
low polarity,14,21 the inertness of the anionic complex
increases when going from [(Ph3P)Pd(Ph)I2]- (Keq ) 0.2)
to [(Ph3P)Pd(Ph)Br2]- (Keq ) 10), and further to [(Ph3P)-
Pd(Ph)Cl2]- (Keq ) 11).
It is clear that in the presence of [Bu4N]X or PPNCl
the tricoordinate intermediate [(Ph3P)Pd(Ph)X], which
is prone to reductive elimination, would equilibrate with
[(Ph3P)Pd(Ph)X2]-, which is a coordinatively saturated
anion and is therefore less inclined to reductively
eliminate [Ph4P]+ (Scheme 2).30 A transition state with
substantial separation of charges for the reductive
elimination from neutral [(Ph3P)Pd(Ph)X] accounts for
the solvent effect observed. Indeed, the aryl-aryl ex-
change occurs faster in media of higher polarity, which
stabilizes the polar transition state, thus lowering the
activation energy barrier to the formation of [Ph4P]+ and
[PdX]-. The latter (or more likely [(Ph3P)PdX]- emerg-
ing from the phosphine addition, as shown in Scheme
2) and the phosphonium cation form a tight ion pair,
within which the reverse process, oxidative addition,
occurs. We obtained strong evidence for the halide
ligand remaining on the metal during the exchange
when conducted at 75 °C in benzene and toluene (see
below). The X ligand not only stabilizes the Pd(0) species
but also enhances the reactivity of the metal center
F igu r e 2. 1H NMR spectra (CD2Cl2, 20 °C) of [(Ph3P)2-
Pd2(Ph)2(µ-Br)2] (bottom) and [(Ph3P)2Pd(Ph)Br2]- gener-
ated by the addition of [Bu4N]Br (50 equiv per Pd) to
[(Ph3P)2Pd2(Ph)2(µ-Br)2] (top).
Ta ble 2. 31P a n d 1H NMR Da ta for Com p lexes
[(P h 3P )2P d 2(P h )2(µ-X)2] a n d [(P h 3P )P d (P h )X2]-,
Wh er e X ) I, Br , Cl, Obta in ed for Solu tion s in
CD2Cl2 a n d C6D6 a t 25 °C (see text)
[(Ph3P)2Pd2(Ph)2(µ-X)2]
[(Ph3P)Pd(Ph)X2]-
nucleus
δ (CD2Cl2)
δ (C6D6) δ (CD2Cl2) δ (C6D6)
31P
X ) Cl
X ) Br
X ) I
31.6
30.4
27.3
n.a.a
30.7
27.6
30.9
29.9
28.2
n.a.a
30.5
28.0
1Hb
3,4-C6H5Pd
2-C6H5Pd
3-C6H5P
6.7
7.0
7.3
7.4
6.6
7.3
6.4
6.9
7.15
6.5
7.4
4-C6H5P
7.25
3,4-C6H5P
2-C6H5P
6.9
7.6
7.0
7.9
7.5
7.5
a
b
Not available. 1H NMR chemical shift and signal shape
parameters of these complexes are identical for X ) I and Br. No
1H NMR data are available for [(Ph3P)Pd(Ph)Cl2]-.
toward oxidative addition,31 in this case of the tetraphe-
nylphosphonium cation.
At higher temperatures and/or in polar solvents, the
Pd-X bond of the anionic zerovalent Pd complexes
ionizes to produce [Ph4P]X and unstable low-ligated
neutral Pd(0) species (Scheme 2). In chlorinated sol-
vents, such as dichloromethane, 1,2-dichloroethane, and
chloroform, the Pd-X ionization should be more facile
than in benzene or toluene. Moreover, the C-Cl bonds
of these solvents may react with coordinatively unsatur-
ated neutral Pd(0) species, thus driving the equilibrium
between [(Ph3P)PdX]- and [(Ph3P)Pd] (eq 8) toward the
latter. This is why [(Ph3P)2Pd(Ph)I] and especially
[(Ph3P)2Pd2(Ph)2(µ-I)2] readily decompose in chlorinated
(30) There may be an alternative path25a,f for the formation of
[(Ph3P)Pd(Ph)X2]- that involves nucleophilic attack of the halide on
[(Ph3P)2Pd(Ph)X], followed by phosphine elimination from the resulting
five-coordinate anionic complex [(Ph3P)2Pd(Ph)X2]-. As the aryl-aryl
exchange reactions occurred, neither [(Ph3P)Pd(Ph)X2]- nor [(Ph3P)2-
Pd(Ph)X2]- were observed by the 31P NMR technique.
(31) (a) Negishi, E.; Takahashi, T.; Akiyoshi, K. J . Chem. Soc., Chem.
Commun. 1986, 1338. (b) Amatore, C.; Azzabi, M.; J utand, A. J . Am.
Chem. Soc. 1991, 113, 8375. (c) Amatore, C.; J utand, A. J . Organomet.
Chem. 1999, 576, 254, and references therein.

Thermal Stability of [(Ph3P)2Pd(Ph)X]
Organometallics, Vol. 19, No. 10, 2000 1895
solvents to give [Ph4P]I4,8 (see also our results described
above).
Sch em e 3
An increase in Keq for eq 8 in the order X ) Cl < Br
< I (for solvents of low polarity but not for polar
solvents, such as DMF31b,c) is perfectly consistent with
the observed order of thermal stability of [(Ph3P)2Pd-
(Ph)X], X ) I < Br, Cl. Bromo and chloro complexes 2
and 3 are less prone to decompose to [Ph4P]X and also
to exchange its aryls for a number of reasons. First, as
we already mentioned, the Pd-P bond weakens when
going from 3 to 2 and further to 1. As a result, the ability
of [(Ph3P)2Pd(Ph)X] to eliminate a phosphine ligand to
produce the reactive tricoordinate intermediate, [(Ph3P)-
Pd(Ph)X], increases in the order X ) Cl < Br < I.
Second, the capability of [(Ph3P)Pd(Ph)X] to reductively
eliminate [Ph4P]+ should increase in the same sequence
due to the well-established order of the π-donating
ability of halogens F > Cl > Br > I.32 While symmetry
forbidden for square-planar Pd(II) complexes,32 pπ f dπ
interactions between X and Pd in tricoordinate [(Ph3P)-
Pd(Ph)X] are allowed, resulting in partial electron/
coordinative saturation of the metal center and conse-
quently a higher activation energy barrier to reductive
elimination. On the contrary, the reverse process, C-P
oxidative addition within the tight ion pair (Scheme 2),
is expected to be faster for Pd(0) species ligated with
chloride rather than bromide and iodide.31b,c
Th er m a l Decom p osit ion of [(P h 3P )2P d (P h )X]
(X ) F a n d HF 2). These and similar organopalladium
fluoro complexes have been recently synthesized in our
laboratories,33-35 primarily in order to explore the
possibility of reductive elimination of ArF as a key step
for the metal-catalyzed synthesis of fluoroarenes. Thus,
thermal decomposition reactions of [(Ph3P)2Pd(Ph)F], 4,
[(Ph3P)2Pd(Ph)(FHF)], 5, and analogous compounds
have been of special interest to us. All experiments
involving 4 and 5 were conducted in a nitrogen atmo-
sphere under rigorously anhydrous conditions, in Teflon
liners placed inside standard 5-mm glass NMR tubes.
For heat-transfer purposes, the space between the liner
and the glass tube was filled with the pure solvent that
was used for the reaction.
(19F NMR) and triplet (31P NMR) with the same coupling
constant were from Ph3PF2.36 To assign the 31P NMR
singlet at 21.3 ppm, iodobenzene was added to the
sample. As a result, the yellow solution decolorized
within 15 s, and the signal at 21.3 ppm disappeared.
Two new singlet resonances were observed instead, at
-5.5 ppm (Ph3P) and 22.3 ppm ([(Ph3P)2Pd(Ph)I]) in a
1:2 integral ratio. Bubbling CO through the colorless
solution converted the [(Ph3P)2Pd(Ph)I] to [(Ph3P)2Pd-
(COPh)I] (18.9 ppm).37 These experiments allowed for
the unambiguous assignment of the singlet at 21.3 ppm
to [(Ph3P)3Pd].
Equation 9 describes the stoichiometry of the decom-
position of 5. The mechanism of this reaction likely
involves the formation of an unstable Pd(IV) hydride38
via oxidative addition of HF, followed by reductive
elimination of benzene and the formation of [(Ph3P)2-
PdF2]. The latter is also unstable, decomposing39 im-
mediately to Ph3PF2 and a zerovalent palladium com-
plex, [(Ph3P)Pd], which disproportionates to [(Ph3P)3Pd]
and Pd metal (Scheme 3).
3
3
3
Bifluoride 5 was found to be substantially less ther-
mally stable than its fluoro counterpart 4. After a
toluene solution of 5 was kept at 75 °C for 4 h, the
decomposition of the complex was complete, resulting
in a yellow solution and a black precipitate of Pd metal.
Benzene was found in the liquid phase (GC-MS). Only
one resonance, a doublet at -38 ppm with J P-F ) 666
Hz, was present in the 19F NMR spectrum of the sample.
The 31P NMR spectrum exhibited two signals, a triplet
at -56.7 ppm (J P-F ) 666 Hz) and a singlet at 21.3 ppm,
their integral intensity ratio being 1:3. Both the doublet
Fluoro complex 4 did not decompose under these
conditions. At 110-120 °C, however, a slow (16 h)
decomposition of 4 in toluene was observed, leading to
biphenyl, Pd metal, [(Ph3P)3Pd], Ph3PF2, and Ph2P-
PPh2 (eq 10) and also small (5-15%) amounts of a few
other P-containing products that were not identified.
In a separate series of similar experiments carried
out in toluene at 110 °C, it was found that conducting
the decomposition in the presence of 2.5 equiv of Ph3P
had no influence on the reaction rate. The reaction
(36) Reddy, G. S.; Schmutzler, R. Z. Naturforsch. B 1970, 25, 1199.
Doxsee, K. M.; Hanawalt, E. M.; Weakley, T. J . R. Inorg. Chem. 1992,
31, 4420.
(37) Garrou, P. E.; Heck, R. F. J . Am. Chem. Soc. 1976, 98, 4115.
(38) Grushin, V. V. Chem. Rev. 1996, 96, 2011.
(39) Mason, M. R.; Verkade, J . G. Organometallics 1990, 9, 864;
1992, 11, 2212.
(32) Caulton, K. G. New J . Chem. 1994, 18, 25.
(33) Fraser, S. L.; Antipin, M. Yu.; Khroustalyov, V. N.; Grushin,
V. V. J . Am. Chem. Soc. 1997, 119, 4769.
(34) Pilon, M. C.; Grushin, V. V. Organometallics 1998, 17, 1774.
(35) Marshall, W. J .; Thorn, D. L.; Grushin, V. V. Organometallics
1998, 17, 5427.

1896 Organometallics, Vol. 19, No. 10, 2000
Grushin
mixtures were analyzed as described above,40 the diphos-
phine being identified by its characteristic 31P NMR
chemical shift of -14.3 ppm.41 Fluorobenzene (19F NMR;
GC-MS) and dibenzyl (GC-MS) were not found among
the reaction products, which indicated that neither C-F
reductive elimination42 nor free-radical processes took
place during the thermal decomposition of 4. It was
tempting to propose that 4 decomposed via a mechanism
involving its disproportionation to a 1:1 mixture of
[(Ph3P)2Pd(Ph)2] and [(Ph3P)2PdF2]. The former would
rapidly undergo reductive elimination of biphenyl,43
small amount (<10%) of Ph3PF2 being produced. The
dimer (31P NMR: 27.6 ppm) eventually precipitated out
of the reaction mixture and was found identical with
an authentic sample.29 Although the phosphorane,
Ph4PF (the sublimable covalent form of tetraphenyl-
phosphonium fluoride), has been reported,44 its NMR
parameters in benzene and toluene have not been. We
generated Ph4PF in solution by sonicating a suspension
of [Ph4P]I and AgF in anhydrous toluene under N2 at
60-90 °C for 6 h (eq 12). The 19F and 31P NMR spectra
39
while the latter would decompose to Pd(0) and Ph3PF2
(see Scheme 3). To test this hypothesis, [(Ph3P)2Pd(4-
C6H4NO2)F],34,35 6, was decomposed under similar
conditions (toluene, 110° C, N2, 16 h). The choice of 6
was dictated by the inability3,4,6 of complexes of the type
[(Ph3P)2Pd(4-C6H4Y)X] to produce [(Ph3P)Pd(Ph2P(4-
C6H4Y))(Ph)X] via aryl-aryl exchange when Y is a
strong electron acceptor (eq 1). GC-MS analysis of the
reaction mixture obtained upon decomposition of 6
revealed the presence of only the monosubstituted
biphenyl 4-NO2C6H4-C6H5 and absence of (4-NO2C6H4)2.
Two conclusions can be drawn from this observation.
First, the proposed mechanism that involved the dis-
proportionation of 4 to the diphenyl and difluoro com-
plexes should be ruled out. Second, the biphenyl emerged
from the thermal decomposition of 4 must have been
formed via coupling of the σ-phenyl with a phenyl of
the Ph3P ligands, somehow occurring on the metal
center.
In an attempt to trap Pd(0) intermediates, the ther-
mal decomposition of 4 was carried out in the presence
of iodobenzene, which is known13 to readily undergo
oxidative addition to zerovalent Pd complexes. When 4
was decomposed at 90 °C in the presence of 1.5 equiv
of PhI, a totally different set of products formed (eq 11),
namely, an organopalladium dimer [(Ph3P)2Pd2(Ph)2(µ-
I)2] and fluorotetraphenylphosphorane, with only a
of the resulting solution exhibited doublets with J P-F
) 557 Hz at 45 ppm (19F) and -61.9 ppm (31P).
Consistent with the formulation Ph4PF, these doublet
resonances were identical with those observed in the
NMR spectra of the reaction mixture obtained by the
thermolysis of 4 in benzene-d6 in the presence of PhI.
Ar yl-Ar yl Exch a n ge Rea ction s of [(P h 3P )2P d -
(C6D5)F ]. We also studied the aryl-aryl exchange
reaction of [(Ph3P)2Pd(C6D5)F], 4-d5, which we synthe-
sized via the ultrasound-promoted I/F exchange33,34
between 1-d5 and AgF in benzene (eq 13). At 75 °C in
(40) Misleading data may be obtained when GC or GC-MS tech-
niques are used for analysis of solutions containing [(Ph3P)nPd] for
benzene and biphenyl. We found that when a toluene solution of pure
[(Ph3P)4Pd] was injected in
a GC-MS instrument, the complex
decomposed in the injector port (250 °C) to produce Ph3P along with
considerable quantities of benzene, biphenyl, and o-terphenyl. Hence,
before performing the analysis of such solutions for benzene/biphenyl,
all [(Ph3P)nPd] species present shoud be somehow converted to stable
products that are neither PhH/Ph2 nor compounds capable of producing
PhH/Ph2 upon injection in the instrument. In a series of control tests
it was found that air oxidation of the zerovalent Pd complex did not
provide a solution for the problem. Satisfactory results were obtained
after toluene solutions of [(Ph3P)4Pd] had been shaken with excess 30%
H2O2 under nitrogen. After 3-10 h at room temperature, the complex
was completely oxidized to an inorganic form of Pd and Ph3PO which
did not decompose in the injector port.
benzene-d6, the deuterated phenyl on the metal in 4-d5
equilibrated with the phenyls on the P ligands after 9
h. This exchange reaction appeared to be autocatalytic,
exhibiting an induction period of ca. 0.5-1 h at 75 °C.
This was not the case for the aryl-aryl interchange
reactions of 1-d5, 2-d5, and 3-d5 (see above). The
observed time to equilibration for the exchange reaction
of 4-d5 was noticeably longer than for the iodo complex
1-d5, slightly longer than for the bromide 2-d5, but
shorter than for the chloride 3-d5 (see Table 1). Unlike
(41) Kuchen, W.; Buchwald, H. Chem. Ber. 1958, 91, 2871. McFar-
lane, H. C. E.; McFarlane, W. J . Chem. Soc., Chem. Commun. 1972,
1189.
(42) No formation of PhF took place when 4 was decomposed in the
presence of Ar3P (Ar ) Ph and C6F5) or R2PCH2CH2PR2 (R ) Ph, C6F5,
and Cy). In all these cases P-F rather than C-F bond formation was
observed. It is noteworthy that after the reaction the
F always
appeared bound to a P atom of the bidentate phosphine, not Ph3P.
(43) Negishi, E.; Takahashi, T.; Akiyoshi, K. J . Organomet. Chem.
1987, 334, 181. Ozawa, F.; Kurihara, K.; Fujimori, M.; Hidaka, T.;
Toyoshima, T.; Yamamoto, A. Organometallics 1989, 8, 180, and
references therein.
(44) Brown, S. J .; Clark, J . H. J . Chem. Soc., Chem. Commun. 1983,
1256. Brown, S. J .; Clark, J . H.; Macquarrie, D. J . J . Chem. Soc., Dalton
Trans. 1988, 277.

Thermal Stability of [(Ph3P)2Pd(Ph)X]
Organometallics, Vol. 19, No. 10, 2000 1897
Sch em e 4
its iodo, bromo, and chloro analogues, the fluoro com-
plex, 4-d5, partially decomposed as the exchange oc-
curred. Thus, by the time the PdC6D5/PdC6H5 equilib-
rium was reached (9 h at 75 °C), ca. 5-7% of the fluoride
had decomposed to compounds that gave rise to a singlet
at 23.2 ppm (31P NMR) and a group of multiplets at 6.9,
7.5, and 7.8 ppm (1H NMR).
When the reaction of 4-d5 at 75 °C was repeated in
the presence of 2.5 equiv of Ph3P, no sign of aryl-aryl
interchange was observed after 13 h. However, some
minor decomposition of the complex must have taken
place, as was judged by changes in line shape of the free
phosphine 31P NMR signal over the course of time.
Before the heating, both components of the mixture
exhibited sharp lines in the 31P NMR spectrum, a
doublet at 18.0 ppm (J P-F ) 12.5 Hz)33,34 and a singlet
at -5.5 ppm (∆ν1/2 ) 4 Hz at LB ) 1.8 Hz) from the
fluoro complex and free phosphine, respectively.45 As the
sample was kept at 75 °C, the singlet resonance from
the free Ph3P broadened noticeably, while undergoing
a slight downfield shift. Thus, after 2 h 10 min and 13
h of reaction, the free phosphine signal was measured
at -5.1 ppm (∆ν1/2 ) 18 Hz) and -4.6 ppm (∆ν1/2 ) 30
Hz), respectively.45b At the same time, no visible changes
were observed for the doublet resonance from the
complex. This change in the spectral pattern was
consistent with the formation of small quantities of
triphenylphosphine complexes of zerovalent Pd, which
are known46 to undergo fast (on the NMR time scale)
exchange with free Ph3P.
The aryl-aryl exchange reaction of 4-d5 in the pres-
ence of iodobenzene (10 equiv) occurred at slower rate.
Thus, after 9 h at 75° C, only ca. 15% of the deuterated
phenyls on the metal had been displaced by the C6H5
groups coming from the phosphine ligands. In the
absence of PhI, however, the same conversion was
reached after 2 h, and after 9 h the exchange was
already complete (see above). The fact that no noticeable
Ph-I activation was observed (31P NMR) constitutes
strong evidence for the F ligand remaining on the metal
during the exchange in benzene at 75 °C.
Mech a n ism s of th e Th er m a l Decom p osition a n d
P h /P h Exch a n ge Rea ction s of [(P h 3P )2P d (P h )F ].
Our results reveal that two clearly distinct reaction
paths take place when [(Ph3P)2Pd(Ph)F] is thermally
decomposed. One of these two (Scheme 4) is similar to
the mechanism that governs the aryl-aryl exchange
and thermal decomposition reactions of the chloro,
bromo, and iodo complexes 1, 2, and 3, as outlined in
Scheme 2. After predissociation of a phosphine ligand
from 4 (4-d5), followed by reductive elimination of Ph4P+,
the F ligand remains in the inner coordination sphere
of the anions [PdF]- and [(Ph3P)PdF]- (at 75 °C). These
Pd(0) complexes are negatively charged, forming tight
ion pairs with the phosphonium cation. Because of that,
oxidative addition of the Ph4P+ to its ion-paired Pd(0)
counterions occurs at least 2 orders of magnitude faster
than oxidative addition of the C-I bond when the
exchange is conducted in the presence of iodobenzene
(see above). In contrast, neutral phosphine complexes
of Pd(0) are known to be much more reactive toward
PhI13 than toward Ph4P+.8 At 90 °C and above, however,
the F ligand on the Pd(0) anion may transfer to the
phosphonium cation to form Ph4PF and neutral [(Ph3P)-
Pd], which readily adds PhI to give [(Ph3P)2Pd2(Ph)2(µ-
I)2] (Scheme 4). The lower thermal stability of 4 stems
(at least in part) from the ability of the P atom in Ph4P+
to abstract fluoride but not other halide anions from
[(Ph3P)PdX]-.
There is another decomposition path for [(Ph3P)2Pd-
(Ph)F] that is characteristic of only the fluoro Pd
complexes and is clearly distinct from any of the paths
described above (Schemes 2 and 4). While occurring
independently and simultaneously with the reversible
reductive elimination of Ph4P+ and Ph4PF (Scheme 4),
this irreversible process is considerably slower and is
not inhibited by free phosphine, giving rise to Pd(0), Ph3-
PF2, Ph2, and Ph2P-PPh2 as main products. Two
mechanisms are conceivable that would account for the
formation of these products (Scheme 5, paths A and B).
One of them (path A) is similar to the mechanism
reported by Morita, Stille, and Norton47 for the alkyl-
aryl exchange reaction of σ-methyl palladium complexes
containing aromatic phosphines. This mechanism in-
volves R-elimination of a phenyl group on the phospho-
rus atom, a process that is not retarded upon addition
of free phosphine,47 followed by reductive elimination
(45) (a) This demonstrates that at room temperature45b phosphine
exchange between 4 and Ph3P is slow on the NMR time scale. (b) All
exchange reactions of 4-d5 were run in a thermostated oil bath. The
NMR spectra to follow the reaction were all taken at 25 °C, after the
sample was removed from the bath and brought to ambient temper-
ature. After that the heating was resumed if necessary.
(47) Morita, D. K.; Stille, J . K.; Norton, J . R. J . Am. Chem. Soc. 1995,
117, 8576.
(46) Mann, B. E.; Musco, A. J . Chem. Soc., Dalton Trans. 1975, 1673.

1898 Organometallics, Vol. 19, No. 10, 2000
Grushin
Sch em e 5
to produce biphenyl and [(Ph3P)Pd(Ph2P)F]. The latter
may equilibrate with [(Ph3P)Pd(Ph2P)2] and [(Ph3P)-
PdF2], which would reductively eliminate Ph2P-PPh2
and Ph3PF2, respectively, to produce Pd(0). It is also
conceivable that [(Ph3P)Pd(Ph2P)F] rearranges to [(Ph3P)-
Pd(Ph2PF)]. In fact, as the decomposition of 4 occurred,
the formation and eventual disappearance of two species
was observed (major, 19F NMR, doublet, 2 ppm; 31P
NMR, doublet, -68.5 ppm; J P-F ) 630 Hz) and (minor,
19F NMR, doublet, 3 ppm; 31P NMR, doublet, -69.7 ppm;
J P-F ) 648 Hz), which can be assigned to Pd complexes
containing Ph2PF ligand(s).48 It has been shown48b,c,49
that Ph2PF readily disproportionates below 100 °C to
give a mixture of Ph2P-PPh2 and Ph2PF3. The latter
may be reduced by the Pd(0) species to Ph2PF via a
series of oxidative addition and reductive elimination
reactions.
Path B (Scheme 5) differs from path A in that the
biphenyl reductively eliminates from a Pd(II) complex
which is formed via P-F reductive elimination of [Ph3-
PF]+,50 followed by oxidative addition of one of its Ph-P
bonds to the Pd(0) complex. Both paths A and B lead to
coordinatively unsaturated neutral Pd(0) species which
trap avidly the phosphine molecules emerging from the
predissociation step that is required for the aryl-aryl
exchange to occur (Schemes 2 and 4). As a result, the
aryl-aryl exchange, reaction of 4-d5 is autocatalytic,
exhibiting an induction period needed to produce the
coordinatively unsaturated Pd species (Scheme 5) that
catalyze the aryl-aryl exchange process. This autocata-
lytic effect disappears, or at least diminishes substan-
tially, when the exchange reaction is run in the presence
of iodobenzene (see above), an efficient scavenger for the
Pd(0) complexes which readily add PhI, thus gaining
coordinative saturation. For instance, PhI readily reacts
with “[(Ph3P)Pd]” generated in situ, to give [(Ph3P)2Pd2-
(Ph)2(µ-I)2],29 the latter possessing much lower affinity
for Ph3P than the former.
Con clu d in g Rem a r k s
Results obtained in this work bear several implica-
tions for catalysis. The leaving group X of the aromatic
substrate, ArX, used for a Pd-catalyzed reaction has
been long known1,13 to determine the reactivity of the
C-X bond toward catalytically active zerovalent pal-
ladium. Oxidative addition of aryl halides or triflates
to Pd(0) is the first key step which frequently but not
always determines the overall rate of the catalytic
reaction, the well-established order of reactivity being
ArI > ArBr ≈ ArOTf . ArCl . ArF. At the same time,
little is known about the influence of the ligand X on
successive steps of the catalytic cycle. It had been
thought that after the oxidative addition that results
in the formation of [(Ph3P)2Pd(Ar)X] the halogen X
becomes more or less a spectator rather than an active
player for the rest of the catalytic cycle. In recent years,
however, information has begun to accrue, suggesting
that the nature of X can sometimes strongly control such
critical catalytic steps as olefin or CO coordination,
ligand exchange, migratory insertion, reductive elimina-
tion, and so forth.1c Our data clearly demonstrate that
catalytically active Pd species forming in a reaction
using iodoarenes will be less thermally stable and more
prone to undergo P-aryl/Pd-aryl exchange reactions
than analogous complexes mediating the same catalytic
process that employs similar bromoarenes. To suppress
the highly undesired aryl-aryl exchange reactions, one
might consider conducting the process in media of low
polarity in the presence of extra phosphine ligand and/
(48) (a) NMR parameters of free Ph2PF: 19F NMR ) -202 ppm;
31P NMR ) 168 ppm; J P-F ) 905 Hz;48b 19F NMR ) -195.9 ppm; 31P
NMR ) 162.3 ppm; J P-F ) 883 Hz.48c (b) Brown, C.; Murray, M.;
Schmutzler, R. J . Chem. Soc. C 1970, 878. (c) Riesel, L.; Haenel, J . Z.
Anorg. Allg. Chem. 1991, 603, 145.
(49) Riesel, L.; Haenel, J . J . Fluorine Chem. 1988, 38, 335.
(50) (a) 19F NMR ) -129.1 ppm; 31P NMR ) -93.7 ppm; J P-F
)
998 Hz. (b) Seel, F.; Bassler, H.-J . Z. Anorg. Allg. Chem. 1975, 418,
263.

Thermal Stability of [(Ph3P)2Pd(Ph)X]
Organometallics, Vol. 19, No. 10, 2000 1899
or a halide source.51 Running the process at the highest
possible Pd catalyst concentration may also help avoid
the aryl-aryl interchange reactions. Although no pana-
cea is offered in this work to cure the aryl-aryl
exchange “disease”, it is hoped that our data and
mechanistic considerations may help develop some
“preventive measures” that will broaden the scope and
reduce the limitations of the catalytic chemistry of
aromatic halides.
hexanes (10 mL). The white crystals were separated and dried
under vacuum. The yield of spectroscopically pure 4-d5 was
0.15 g (86%). The complex was brought into a drybox, where
it was recrystallized from anhydrous dichloromethane-ether
1
and dried. H NMR (C6D6, 20 °C), δ: 7.0 (m, 3H, 3,4-C6H5P),
7.9 (m, 2H, 2-C6H5P). 31P NMR (C6D6, 20 °C), δ: 18.7 (d, J P-F
) 13 Hz).
Syn th esis of [(P h 3P )2P d 2(P h )2(µ-Br )2]. This complex was
prepared and isolated in air. To a clear colorless solution of
[(Ph3P)2Pd2(Ph)2(µ-OAc)2]52 (0.095 g; 0.09 mmol) and [Bu4N]-
Br (0.076 g; 0.24 mmol) in CH2Cl2 (1 mL) were added acetone
(10 mL) and water (0.2 mL). After 3 h the pale lemon-yellow
crystalline solid was separated, thoroughly washed with
acetone and dichloromethane, and dried under vacuum. The
yield of the bromo dimer was 0.093 g (94%). Anal. Calcd for
Exp er im en ta l Section
All manipulations were done in a glovebox under nitrogen,
unless otherwise noted. Solvents used for the decomposition
and exchange studies were dried and stored over freshly
activated molecular sieves (4 Å). Teflon liners for standard
5-mm glass NMR tubes were supplied by the Wilmad Com-
pany. All chemicals were purchased from Aldrich, Kodak,
Strem, and Organometallics Chemical Companies. Complexes
1,13 2,13 3,14 4,33,34 5,34 6,34,35 and [(Ph3P)2Pd2(Ph)2(µ-I)2]29 were
prepared as described in the literature. NMR spectra were
obtained with a Bruker Avance DPX 300 spectrometer. A
Hewlett-Packard 6890 Series GC-MS instrument was used for
GC-MS analysis.
C
48H40Br2P4Pd2: C, 54.8; H, 3.8; Br, 15.2. Found: C, 55.1; H,
3.7; Br, 15.6. For NMR parameters of the complex, see Table
2.
Syn th esis of [(P h 3P )2P d 2(P h )2(µ-Cl)2]. This complex was
also prepared and isolated in air. To a clear colorless solution
of [(Ph3P)2Pd2(Ph)2(µ-OAc)2]52 (0.170 g; 0.16 mmol) and PPN
Cl (0.203 g; 0.34 mmol) in CH2Cl2 (5 mL) were added acetone
(20 mL) and water (1 mL). After 15 min the white crystals of
[(Ph3P)2Pd2(Ph)2(µ-Cl)2] were separated, thoroughly washed
with acetone, and dried under vacuum. The yield of the
spectroscopically pure28 chloro dimer was 0.140 g (86%).
Th er m a l Decom p osition of 1-3. A mixture of 1 (0.16 g;
0.19 mmol) and toluene (10 mL) was stirred under reflux (N2)
for 16 h. A palladium mirror formed. 31P NMR analysis of the
brown liquid phase indicated the presence of Ph3P (br s, -0.1
ppm), small quantities (ca. 10%) of unreacted 1 (s, 22.4 ppm),
and trace amounts (ca. 1%) of [(Ph3P)2PdI2] (s, 12.5 ppm). Both
the shape (broadened) and the chemical shift of the Ph3P
resonance suggested exchange with phosphine Pd(0) complexes
present in low concentration. The solids were extracted with
CH2Cl2 (3 × 5 mL). The combined extracts were filtered,
reduced in volume to ca. 3 mL, and treated with ether (30 mL).
The white precipitate of [Ph4P]I was separated, washed with
ether, and dried under vacuum. The product (0.075 g; 84%)
was identical with a commercial sample of tetraphenylphos-
phonium iodide. Under similar conditions, complexes 2 and 3
exhibited no sign of decomposition (1H and 31P NMR).
P r ep a r a tion of Sa m p les for Ar yl-Ar yl Exch a n ge a n d
Th er m a l Decom p osition Stu d ies. NMR tubes and Teflon
liners were kept under vacuum (5 × 10-3 mm) for 12 h prior
to use. All complexes, tetrabutylammonium salts, and anhy-
drous solvents were kept under nitrogen in a drybox. Teflon
liners were used to prepare samples for reactions of bifluoro
or fluoro Pd complexes. After charged NMR tubes were capped
with 5-mm rubber septum stoppers and sealed with Parafilm,
the samples were taken out of the drybox. To speed up
dissolution of solid complexes and tetrabutylammonium salts
in benzene-d6, sonication and/or gentle heating (25-40 °C)
were applied. Once all solids dissolved, the sample was placed
in a preheated oil bath or VT probe of the NMR spectrometer.
Gen er a tion of [(P h 3P )P d (P h )X2]- fr om [(P h 3P )2P d 2-
(P h )2(µ-X)2] a n d [Bu 4N]X (X ) I a n d Br ) or P P N Cl. (a) A
mixture of a dinuclear complex, [Bu4N]X (or PPN Cl), and a
dry deuterated solvent was prepared in the drybox, as de-
scribed above. The mixture was shaken and/or slightly heated
until the solids dissolved (3-5 min), after which the sample
was analyzed by NMR (Table 2). (b) A solution (suspension)
of a dinuclear complex was placed in an NMR tube capped
Syn th esis of 1-d 5. Iodobenzene-d5 (Aldrich, 98% D; 0.17 g;
0.81 mmol) was added to a stirring mixture of [(Ph3P)4Pd]
(Strem, 99.9%; 0.85 g; 0.74 mmol) and toluene (15 mL). After
the mixture was stirred at room temperature for 18 h, the
product was isolated in air. The white precipitate was sepa-
rated, washed with toluene (3 × 1 mL) and hexanes (5 × 10
mL), and dried under vacuum. This gave 0.55 g of the product.
Combining the mother liquor with the toluene and hexane
washings resulted in precipitation of 0.04 g of pure 1-d5. Total
1
yield: 0.59 g (96%). H NMR (CD2Cl2, 20 °C), δ: 7.3 (m, 2H,
3-C6H5P), 7.4 (m, 1H, 4-C6H5P), 7.6 (m, 2H, 2-C6H5P). 31P NMR
(CD2Cl2, 20 °C), δ: 22.2 (s).
Syn th esis of 2-d 5. A mixture of [(Ph3P)4Pd] (Strem, 99.9%;
0.55 g; 0.48 mmol), bromobenzene-d5 (Aldrich, 99.5% D; 0.55
g; 3.4 mmol), Ph3P (0.20 g; 0.76 mmol), and toluene (5 mL)
was stirred at 75 °C (oil bath) for 5 h. The mixture was cooled
to room temperature. After 2 h the product was isolated in
air. The solid was separated, washed with hexanes (5 × 5 mL),
recrystallized from dichloromethane-ether, and dried under
vacuum. The yield of the white crystalline 2-d5 was 0.31 g
1
(83%). H NMR (CD2Cl2, 20 °C), δ: 7.3 (m, 2H, 3-C6H5P), 7.4
(m, 1H, 4-C6H5P), 7.6 (m, 2H, 2-C6H5P). 31P NMR (CD2Cl2, 20
°C), δ: 22.8 (s).
Syn th esis of 3-d 5. This complex was prepared under
biphasic conditions from 1-d5 and KCl in CH2Cl2-H2O, ac-
cording to the literature method for the synthesis of 3.14 The
reaction, isolation, and purification of the product were carried
out in air. The complex was purified by three recrystallizations
from dichloromethane-ether. 1H NMR (CD2Cl2, 20 °C), δ: 7.3
(m, 2H, 3-C6H5P), 7.4 (m, 1H, 4-C6H5P), 7.6 (m, 2H, 2-C6H5P).
31P NMR (CD2Cl2, 20 °C), δ: 23.9 (s).
Syn th esis of 4-d 5. A mixture of 1-d5 (0.20 g; 0.24 mmol),
AgF (0.04 g; 0.31 mmol), PhI-d5 (Aldrich, 98% D; 0.01 g; 0.05
mmol), and benzene (7.5 mL) was sonicated under N2 at 20-
25 °C for 3.5 h (monitored by 31P NMR spectroscopy). The
product was isolated in air. The benzene solution was filtered
through a Celite plug. The solids were thoroughly extracted
with warm (45 °C) benzene. After filtration of the extracts
through the same Celite plug, all benzene solutions were
combined, reduced in volume to ca. 3 mL, and treated with
with
a rubber septum. Separately, a solution of known
concentration of [Bu4N]X or PPNCl was prepared in a GC vial.
Both the tube and the vial were taken out of the drybox. The
salt solution was added, portionwise, to the NMR tube
containing the complex with a microsyringe. NMR spectra
(51) Running such reactions in the presence of extra halide might
suppress the formation of cationic Pd intermediates, which can be
crucial for successful catalysis. See: Yamamoto, A. J . Organomet.
Chem. 1995, 500, 337. Kayaki, Y.; Yamamoto, A. Yuki Gosei Kagaku
Kyokaishi 1998, 56, 96. Yamamoto, A. J . Chem. Soc., Dalton Trans.
1999, 1027.
(52) Grushin, V. V.; Bensimon, C.; Alper, H. Organometallics 1995,
14, 3259.

1900 Organometallics, Vol. 19, No. 10, 2000
Grushin
were taken after each addition. A representative experiment
is described below.
be 0.8 × 10-3, 8.7 × 10-3, and 0.8 × 10-3 M, respectively (Keq
≈ 10 mol-1 dm3).
Ultr a sou n d -P r om oted Rea ction of [P h 4P ]I w ith AgF
in Tolu en e. A suspension of [Ph4P]I (0.06 g; 0.13 mmol) and
AgF (0.09 g; 0.71 mmol) in dry toluene (0.3 mL) was sonicated
in a Teflon liner placed in a standard 5-mm NMR tube under
N2 at 60 °C for 2 h. The sample was analyzed by 31P and 19F
NMR: δP ) -61.9 ppm (d, J P-F ) 557 Hz); δF ) 45 ppm (d,
J P-F ) 557 Hz). After sonication for an additional 4 h at 90 °C
the conversion to Ph4PF somewhat increased.
A solution of [Bu4N]Br (92 mg) in CD2Cl2 (0.5 mL) was
prepared in a standard GC vial. Using a microsyringe, 10 µL
of this solution was added to a septum-capped NMR tube
containing [(Ph3P)2Pd2(Ph)2(µ-Br)2] (3 mg) in CD2Cl2 (0.5 mL).
After the tube was shaken for 5 min at 25 °C, a small amount
of the undissolved dinuclear complex still remained in the
mixture. Both 31P and 1H NMR spectra of the mixture
indicated that the liquid phase contained the unreacted dimer
(δP ) 30.4 ppm) and [Bu4N][(Ph3P)Pd(Ph)Br2] (δP ) 29.9 ppm)
in a 1:2 ratio. Because counterion exchange between the latter
and excess [Bu4N]Br is fast on the NMR time scale, one type
of butyl group was observed in the 1H NMR spectrum.
Integration of the butyl proton resonances against the aromatic
signals allowed for calculation of the concentrations of the
dimer, [Bu4N]Br, and the anionic complex at equilibrium to
Ack n ow led gm en t. I thank Dr. Soley Kristjansdottir
for a stimulating discussion. Drs. Alexander Shtarev,
Elisabeth Hauptman, Christian Lenges, and Karen
Goldberg are acknowledged for reading the manuscript
and for valuable comments.
OM0001265