Mechanistic studies on the aryl-aryl interchange reaction of ArPdL2I (L = triarylphosphine) complexes

Article abstract of DOI:10.1021/ja972554g

The aryl-aryl interchange reaction of ArPdL₂I complex 1m was found to follow pseudo-first-order kinetics. A marked inhibition in the presence of excess phosphine and/or excess iodide was observed, suggesting that a dissociative pathway was involved, contrary to the analogous alkyl-aryl interchange reaction studied previously. Phosphine flooding experiments could not be performed due to a competing phosphonium salt formation reaction that occurred in the presence of excess phosphine. A deuterium labeling experiment indicated that the interchange reaction proceeded via the reductive elimination to form the phosphonium salt, suggesting that excess phosphine was acting as a trap for intermediate palladium(0) species preventing the generation of the interchanged palladium-(II) complex. Substituent effect studies of the interchange reaction indicated that it was inhibited by electron-withdrawing groups on both the phosphine and palladium-bound aryl groups and by increased steric bulk on both the phosphine and palladium- bound aryl groups. Under catalytic conditions, the distribution of phosphines formed from the aryl-aryl interchange during palladium-mediated cross- coupling reactions could be modeled by statistics. Various strategies for eliminating the formation of byproducts caused by the interchange during cross-coupling reactions were screened and optimized.

Full text of DOI:10.1021/ja972554g

J. Am. Chem. Soc. 1997, 119, 12441-12453
12441
Mechanistic Studies on the Aryl-Aryl Interchange Reaction of
ArPdL2I (L ) Triarylphosphine) Complexes
Felix E. Goodson,† Thomas I. Wallow,‡ and Bruce M. Novak*
Contribution from the Department of Polymer Science and Engineering and Materials Research
Science and Engineering Center, UniVersity of Massachusetts, Amherst, Massachusetts 01003
ReceiVed July 28, 1997X
Abstract: The aryl-aryl interchange reaction of ArPdL2I complex 1m was found to follow pseudo-first-order kinetics.
A marked inhibition in the presence of excess phosphine and/or excess iodide was observed, suggesting that a
dissociative pathway was involved, contrary to the analogous alkyl-aryl interchange reaction studied previously.
Phosphine flooding experiments could not be performed due to a competing phosphonium salt formation reaction
that occurred in the presence of excess phosphine. A deuterium labeling experiment indicated that the interchange
reaction proceeded via the reductive elimination to form the phosphonium salt, suggesting that excess phosphine
was acting as a trap for intermediate palladium(0) species preventing the generation of the interchanged palladium-
(II) complex. Substituent effect studies of the interchange reaction indicated that it was inhibited by electron-
withdrawing groups on both the phosphine and palladium-bound aryl groups and by increased steric bulk on both
the phosphine and palladium-bound aryl groups. Under catalytic conditions, the distribution of phosphines formed
from the aryl-aryl interchange during palladium-mediated cross-coupling reactions could be modeled by statistics.
Various strategies for eliminating the formation of byproducts caused by the interchange during cross-coupling reactions
were screened and optimized.
Introduction
carbon or carbon-heteroatom bonds, these reactions possess
the added benefits of proceeding through readily accessible
substrates, being tolerant of most functionalities, and preserving
the regiochemistry of the reactants involved. Under the
appropriate conditions, the yields of these reactions can be close
to quantitative, which makes them useful for condensation
polymerization reactions. Indeed, a large variety of polymeric
materials have been synthesized using the methodologies
provided by these reactions.13
Palladium-mediated cross-coupling reactions, such as the
coupling of an organic electrophile with an organoboron (Suzuki
coupling),1 -tin (Stille coupling),2 -aluminum,3 -silicon,4 -zinc,5
-zirconium,6 or -magnesium7 reagent; the coupling of an aryl
halide with an alkene (Heck coupling)8 or alkyne (Hagihara
coupling);9 or more recently the coupling of an aryl halide with
anionic amides,10 thiolates,11 or alkoxides,12 are very useful tools
in the arsenal of the synthetic chemist. In addition to their
obvious importance as a method for the formation of carbon-
One unfortunate drawback to these methodologies is the
occurrence of an interchange between phosphorus-bound aryl
moieties and palladium-bound aryl14 or alkyl15 groups in the
† Current address: Department of Chemistry, Yale University, New
Haven, CT 06520.
‡ Current address: IBM-Almaden Research Center, 650 Harry Rd., San
Jose, CA 95120.
(8) (a) Mizoroki, T.; Mori, K.; Ozaki, A. Bull. Chem. Soc. Jpn. 1971,
44, 581. (b) Heck, R. F.; Nolley, J. P., Jr. J. Org. Chem. 1972, 37, 2320.
(c) Mori, K.; Mizoroki, T.; Ozaki, A. Bull. Chem. Soc. Jpn. 1973, 46, 1505.
(d) Ziegler, C. B., Jr.; Heck, R. F. J. Org. Chem. 1978, 43, 2941. For
reviews, see: (e) Heck, R. F. Acc. Chem. Res. 1979, 12, 146. (f) de Meijere,
A.; Meyer, F. E. Angew. Chem., Int. Ed. Engl. 1994, 33, 2379. (g) Cabri,
W.; Candiani, I. Acc. Chem. Res. 1995, 28, 2.
(9) (a) Cassar, L. J. Organomet. Chem. 1975, 93, 253. (b) Heck, R. F.;
Dieck, H. A. J. Organomet. Chem. 1975, 93, 259. (c) Sonogashira, K.;
Tohda, Y.; Hagihara, N. Tetrahedron Lett. 1975, 4467. (d) Takahashi, S.;
Kuroyama, Y.; Sonogashira, K.; Hagihara, N. Synthesis 1980, 627.
(10) (a) Kosugi, M.; Kameyama, M.; Migita, T. Chem. Lett. 1983, 927.
(b) Guram, A. S.; Buchwald, S. L. J. Am. Chem. Soc. 1994, 116, 7901. (c)
Guram, A. S.; Rennels, R. A.; Buchwald, S. L. Angew. Chem., Int. Ed.
Engl. 1995, 34, 1348. (d) Louie, J.; Hartwig, J. F. Tetrahedron Lett. 1995,
36, 3609. (e) Widenhoefer, R. A.; Buchwald, S. L. Organometallics 1996,
15, 2755. (f) Hartwig, J. F.; Richards, S.; Baran˜ano, D.; Paul, F. J. Am.
Chem. Soc. 1996, 118, 3626.
(11) (a) Murahashi, S. -I.; Yamamura, M.; Yanagisawa, K.; Mita, N.;
Kondo, K. J. Org. Chem. 1979, 44, 2408. (b) Migita, T.; Shimizu, T.; Asami,
Y.; Shiobara, J.; Kato, Y.; Kosugi, M. Bull. Chem. Soc. Jpn. 1980, 53,
1385. (c) Kosugi, M.; Ogata, T.; Terada, M. Sano, H.; Migita, T. Bull.
Chem. Soc. Jpn. 1985, 58, 3657. (d) Reference 8c. (e) Dickens, M. J.; Gilday,
J. P.; Mowlem, T. S.; Widdowson, D. A. Tetrahedron 1991, 40, 8621. (f)
Baran˜ano, D.; Hartwig, J. F. J. Am. Chem. Soc. 1995, 117, 2937. (g) Rossi,
R.; Bellina, F.; Mannina, L. Tetrahedron 1997, 53, 1025.
X Abstract published in AdVance ACS Abstracts, December 15, 1997.
(1) (a) Miyaura, N.; Yamada, K.; Suzuki, A. Tetrahedron Lett. 1979,
20, 3437. (b) Miyaura, N.; Yanagi, T.; Suzuki, A. Synth. Commun. 1981,
11, 513. For selected reviews, compilations, and monographs, see: (c)
Suzuki, A. Acc. Chem. Res. 1982, 15, 178. (d) Miyaura, N.; Yamada, K.;
Suginome, H.; Suzuki, A. J. Am. Chem. Soc. 1985, 107, 972. (e) Suzuki,
A. Pure Appl. Chem. 1985, 57, 1749. (f) Suzuki, A. Pure Appl. Chem.
1991, 63, 419. (g) Martin, A. R.; Yang, Y. Acta Chem. Scand. 1993, 47,
221. (h) Suzuki, A. Pure Appl. Chem. 1994, 66, 213. (i) Miyaura, N.; Suzuki,
A. Chem. ReV. 1995, 95, 2457.
(2) (a) Milstein, D.; Stille, J. K. J. Am. Chem. Soc. 1979, 101, 4992. For
reviews see: (b) Stille, J. K. Pure Appl. Chem. 1985, 57, 1771. (c) Stille,
J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508. (d) Mitchell, T. N.
Synthesis 1992, 803.
(3) (a) Baba, S.; Nigishi, E. J. Am. Chem. Soc. 1976, 98, 6729. (b)
Negishi, E. Pure Appl. Chem. 1981, 53, 2333 and references therein. (c)
Negishi, E. Acc. Chem. Res. 1982, 15, 340 and references therein. (d)
Negishi, E.; Takahashi, T.; Baba, S.; Van Horn, D. E.; Okukado, N. J. Am.
Chem. Soc. 1987, 109, 2393.
(4) (a) Hatanaka, Y.; Hiyama, T. J. Org. Chem. 1988, 53, 918. (b)
Hatanaka, Y.; Hiyama, T. J. Org. Chem. 1989, 54, 268. (c) Hatanaka, Y.;
Hiyama, T. Tetrahedron Lett. 1990, 31, 2719.
(5) (a) Negishi, E.; King, A. O.; Okukado, N. J. Org. Chem. 1977, 42,
1823. (b) Reference 3d. (c) Erdik, E. Tetrahedron 1992, 48, 9577 and
references therein.
(12) (a) Mann, G.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 13109.
(b) Palucki, M.; Wolfe, J. P.; Buchwald, S. L. J. Am. Chem. Soc. 1997,
119, 3395. (c) Reference 11g.
(6) (a) Okukado, N.; Van Horn, D. E.; Klima, W. L.; Negishi, E.
Tetrahedron Lett. 1978, 1027. (b) Reference 3d.
(7) (a) Yamamura, M.; Moritani, I.; Murahashi, S.-I. J. Organomet. Chem.
1975, 91, C39. (b) Sekiya, A.; Ishikawa, N. J. Organomet. Chem. 1976,
118, 349. (c) Kumada, M. Pure Appl. Chem. 1980, 52, 669 and references
therein.
(13) For reviews, see: (a) Yamamoto, T. Prog. Polym. Sci. 1992, 17,
1153. (b) Schlu¨ter, A.-D.; Wegner, G. Acta Polym. 1993, 44, 59. (c) Tour,
J. M. AdV. Mater. 1994, 6, 190. (d) Percec, V.; Hill, D. H. ACS Symp. Ser.
1996, 624, 2.
S0002-7863(97)02554-7 CCC: $14.00 © 1997 American Chemical Society

12442 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
Goodson et al.
Scheme 1. Simplified Mechanism for Palladium-Catalyzed
Cross-Coupling Reactions
attempt to elucidate the mechanism of this reaction. Recently,
Norton and co-workers presented a detailed account on the
mechanism of the alkyl-aryl interchange of RPdL2I com-
plexes.15 But, as we shall point out, there are a number of
interesting and fundamental differences between the two
exchange reactions despite their apparent similarity. Secondly,
we will present the results of several analytical experiments to
quantitatively show the impact of the aryl-aryl interchange on
a number of small-molecule Suzuki coupling reactions and, from
these, infer the possible consequences upon analogous poly-
merizations. Finally, we will suggest some possible strategies
that might be used by synthetic organic and polymer chemists
alike to minimize, if not eliminate, the interchange from
palladium-mediated cross-coupling reactions.
RPdL2X intermediates (where L ) triarylphosphine, X ) Br
or I, and R ) aryl or alkyl) of these reactions (eq 1). A
Results and Discussion
Kinetics Experiments. Complexes 1a-z were synthesized
as described elsewhere.21 Of these, complex 1m (R ) 4-meth-
oxyphenyl, Ar ) 4-fluorophenyl) was deemed the most suitable
for initial kinetics studies due to its comparative solubility and
simplified catalytic coupling cycle is depicted in Scheme 1.1i
The RPdL2X species is usually believed to be the resting state
of the catalyst in the cycle,16 and the occurrence of this
interchange would allow the ligand-bound aryl groups to enter
the cross-coupling in lieu of the aryl or alkyl halide derived
organic moieties. In fact, several recent investigations have
attributed the formation of phosphine-derived byproducts in
catalytic17 and stoichiometric18 cross-coupling reactions to this
exchange. Furthermore, our own work on water-soluble poly-
(p-phenylenes)19 suggests that incorporation of phosphine
ligands into cross-coupling polymerizations can not only result
in production of monofunctional aryl endcaps but also produce
branched network structures by inadvertent generation of
multifunctional phosphine monomers.20 Given these conse-
quences, it would be useful to gain a more detailed understand-
ing of the interchange process, with the end goal of eliminating
it altogether from the coupling reactions, thus improving the
efficiency of small-molecule couplings as well as providing a
means for synthesizing polymers with the intended linear
architecture.
1
stability and to the fact that the H NMR resonances for the
palladium-bound aryl group both before and after exchange were
clearly resolvable. The aryl-aryl interchange reaction appeared
to follow simple first-order kinetics with an observed rate
constant of (7.30 ( 0.12) × 10-4 s-1 at 50 °C and 0.042 M
initial concentration in THF-d8. Kinetics did not deviate from
linearity for the almost five half-lives over which the experi-
ments were run, and the observed rate constants were found to
be insensitive to the initial concentration of 1m in the range
0.021-0.164 M. As the interchange is a reversible process (Keq
≈ 20), the observed rate constant reflects the sum of the rate
constants for the forward and reverse reactions.22 Also, while
the interchange reaction is progressing, there is almost certainly
a rapid dissociation and association of the phosphine ligands
taking place, resulting in a statistical mixture of phosphines with
zero, one, two, and three exchanged aryl groups.14a,15 Conse-
quently, the kinetics observed are the loss of palladium-bound
methoxyphenyl groups and the growth of palladium-bound
fluorophenyl groups, independent of the specific identity of the
phosphine ligands involved. The fact that the kinetics remain
linear for the entire process suggests that the individual rate
constants for the different possible exchange reactions are not
significantly divergent.
In this paper, we present kinetics experiments performed on
the aryl-aryl interchange of ArPdL2I complexes 1a-z, in an
(14) (a) Kong, K.-C.; Cheng, C.-H. J. Am. Chem. Soc. 1991, 113, 6313.
(b) Herrmann, W. A.; Brossmer, C.; Priermeier, T.; O¨ fele, K. J. Organomet.
Chem. 1994, 481, 97. (c) Segelstein, B. E.; Butler, T. W.; Chenard, B. L.
J. Org. Chem. 1995, 60, 12. (d) Sakamoto, M.; Shimizu, I.; Yamamoto, A.
Chem. Lett. 1995, 1101.
(15) Morita, D. K.; Stille, J. K.; Norton, J. R. J. Am. Chem. Soc. 1995,
117, 8576.
(16) For aqueous Suzuki couplings with aryl bromides, we have found
the rate to be dependent upon concentration of catalyst and boronic acid,
suggesting that transmetalation is rate-determining, in agreement with results
for other coupling systems: (a) Reference 3d. (b) Brown, J. M. Cooley, N.
A. Organometallics 1990, 9, 353. (c) Farina, V.; Krishnan, B. J. Am. Chem.
Soc. 1991, 113, 9585. Smith et al. similarly found transmetalation to be
rate-determining when aryl iodides were used in organic Suzuki coupling
systems. However, when aryl bromides were used, oxidative addition was
rate-determining, suggesting that this generalization is not universal: (d)
Smith, G. B.; Dezeny, G. C.; Hughes, D. L.; King, A. O.; Verhoeven, T.
R. J. Org. Chem. 1994, 59, 8151.
(17) (a) O’Keefe, D. F.; Dannock, M. C.; Marcuccio, S. M. Tetrahedron
Lett. 1992, 33, 6679. (b) Reference 14c. (c) Herrmann, W. A.; Brossmer,
C.; O¨ fele, K.; Beller, M.; Fischer, H. J. Mol. Catal. A: Chem. 1995, 103,
133. (d) Herrmann, W. A.; Brossmer, C.; O¨ fele, K.; Beller, M.; Fischer, H.
J. Organomet. Chem. 1995, 491, C1.
To test for a pathway involving the predissociation of a
phosphine ligand, kinetics were followed in CDCl3 at 60 °C
and 0.042 M in initial concentration of 1m, in the presence of
varying amounts of excess phosphine; the results are presented
in Figure 1. As can be seen from this graph, there is a noticeable
inhibition of the rate of interchange with increased phosphine
concentration, again indicating that the loss of ligand is
somehow involved in the mechanism. Another noticeable
feature of this graph is the ample curvature of the plots,
indicating the presence of additional chemical processes.
Indeed, when kinetics were followed with excess phosphine in
THF-d8, a precipitate formed in the NMR tube. This was
1
isolated and characterized by H, 19F, and 31P NMR as well
as FAB mass spectrometry, which indicated that this solid was
an approximately 4:1 mixture of two different phosphonium
salts, tris(4-fluorophenyl)(4-methoxyphenyl)phosphonium ido-
dide and bis(4-fluorophenyl)bis(4-methoxyphenyl)phosphonium
iodide, as well as a trace of a third, tetrakis(4-fluorophenyl)-
(18) (a) Grushin, V. V.; Bensimon, C.; Alper, H. Organometallics 1995,
14, 3259. (b) Reference 15. (c) Reference 11f. (d) Driver, M. S.; Hartwig,
J. F. J. Am. Chem. Soc. 1995, 117, 4708.
(19) Wallow, T. I.; Novak, B. M. J. Am. Chem. Soc. 1991, 113, 7411.
(20) For preliminary disclosures, see: (a) Wallow, T. I.; Seery, T. A.
P.; Goodson, F. E.; Novak, B. M. Polym. Prepr.sAm. Chem. Soc., DiV.
Polym. Chem. 1994, 35 (1), 710. (b) Novak, B. M.; Wallow, T. I.; Goodson,
F. E.; Loos, K. Polym. Prepr.sAm. Chem. Soc., DiV. Polym. Chem. 1995,
36 (1), 693.
(21) Wallow, T. I.; Goodson, F. E.; Novak, B. M. Organometallics 1996,
15, 3708.
(22) Moore, J. W.; Pearson, R. G. Kinetics and Mechanism; John Wiley
and Sons: New York, 1981; p 304.

Aryl-Aryl Interchange Reaction of ArPdL2I
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12443
and co-workers, who observed that trans-[PdPhI(PPh3)2] pro-
duced ca. 30% PPh4I upon reflux in methylene chloride, again
without the presence of additional phosphine.14d Why the
phosphonium salt formation does not seem to occur in nonha-
logenated solvents without the addition of excess phosphine is
uncertain.
To investigate a possible pathway involving the predissocia-
tion of iodide, kinetics of the aryl-aryl interchange were
followed at 0.042 M in initial concentration of 1m in THF-d8
at 50 °C, first in the presence of 0.206 M LiI, then in the
presence 0.206 M LiPF6, and finally in the presence of 0.103
M in each (see the Supporting Information for linearized kinetic
plots). There was only a slight inhibition of the rate caused by
the addition of LiI with respect to kinetics with no added salt
((6.05 ( 0.13) × 10-4 s-1 vs (6.76 ( 0.05) × 10-4 s-1
,
respectively). However, as added salt significantly changes the
polarity of the solvent medium, a better comparison is with
kinetics followed in the presence of an equimolar amount of
added LiPF6. The kinetics followed in the presence of LiI were
significantly slower than those followed in the presence of LiPF6
((11.5 ( 0.2) × 10-4 s-1) with the kinetics in the presence of
one-half the amount of each falling in between. When kinetics
were followed in the presence of 0.206 M lithium trifluo-
romethanesulfonate, the observed rate constant ((20.3 ( 0.02)
× 10-4 s-1) was very close to that for kinetics observed in the
presence of 0.206 M LiPF6 ((19.1 ( 0.4) × 10-4 s-1), suggesting
that this salt effect is general. From these results it is apparent
that, in addition to a reaction pathway involving the predisso-
ciation of phosphine, there is a completely separate pathway
involving the predissociation of iodide.
Given these results, it is tempting to propose a mechanism
that requires a predissociation of either a phosphine or an iodide.
However, this mechanism is inconsistent with the observation
that an added equivalent of phosphine completely shuts down
the exchange reaction, since the additional phosphine should
have no effect on the iodide dissociation pathway.24 An
additional problem with this mechanism is the observation of
Hartwig, who found that the aryl-aryl interchange in (bis-
(diphenylphosphino)ethane)Pd(aryl)(thiolate) complexes was
inhibited by the presence of added triphenylphosphine.11f The
monodentate triphenylphosphine ligand cannot have an effect
on the dissociation equilibrium of (diphenylphosphino)ethane
due to the strong chelate effect of the latter ligand.
Figure 1. Disappearance of 1m in the presence of excess tris(p-
fluorophenyl)phosphine in CDCl3 at 60 °C.
phosphonium iodide. Palladium complexes are known to
catalyze the formation of phosphonium salts from aryl halides
and triarylphosphines,23 so it is not surprising that this side
reaction occurs here. Unfortunately, the presence of this
additional reaction prevents the use of flooding experiments that
would be required to obtain absolute rate constants for a
multistep chemical pathway. Finally, it should be noted that
the disappearance of 1m in the 1 equiv of additional phosphine
plot is almost entirely due to the competing phosphonium salt
formation. After 3 h at 60 °C, only about 3% of the exchanged
product 2m had grown in, in agreement with the observations
of Kong and Cheng.14a
Another aspect worth noting in Figure 1 is the curvature of
the plot for the run with no phosphine added. This feature was
not present in the experiments performed in THF-d8, but the
curvature was noticeable when kinetics were run in tetrachlo-
roethane-d2. Furthermore, for the experiments performed in
chlorinated solvents, there was as much as 30% less signal for
2m at the end of the run than there was for 1m at the beginning.
These results indicated a competing decomposition reaction in
chlorinated solvents. Upon comparison of the 1H NMR
spectrum of the isolated phosphonium salt mixture with spectra
obtained from the kinetics runs performed in chlorinated
solvents, we noticed that the very distinctive splitting pattern
(due to coupling with both 19F and 31P nuclei) at 7.78 ppm in
the phosphonium salt spectrum was present in the spectra from
the kinetics runs, as well as the methoxy resonance at 3.94 ppm.
Unfortunately, the 19F and 31P NMR spectra were not as clear
due to the fact that the phosphonium salts and the ArPdL2I
complexes have resonances in the same regions. This seems
to indicate that the competing decomposition seen in chlorinated
solvents is due to the phosphonium salt formation, even without
the presence of additional phosphine. Further evidence that this
might indeed be occurring was recently provided by Yamamoto
A possible solution to this conundrum was recently provided
by Chenard et al., who suggested that the interchange reaction
occurred first through a reductive elimination to form the
phosphonium salt, followed by an oxidative addition of a
different phosphorus-carbon bond to generate 2 (eq 2).14c
Evidence for this was given by the fact that phosphonium salts
were successfully coupled with aryl stannanes by a phosphine-
free catalyst. Soon afterward, Yamamoto and co-workers
showed that quaternary phosphonium salts could oxidatively
add to palladium(0) olefin species.14d These results would
explain the phosphine inhibition effect mentioned above in that
added phosphine could trap out the palladium(0) intermediate
(23) (a) Migita, T.; Nagai, T.; Kiuchi, K.; Kosugi, M. Bull. Chem. Soc.
Jpn. 1983, 56, 2869. (b) Kowalski, M. H.; Hinkle, R. J.; Stang, P. J. J.
Org. Chem. 1989, 54, 2783.
(24) Additional phosphine should have no effect on the iodide dissocia-
tion pathway unless the added phosphine was able to trap out the dissociated
palladium species as a cationic [ArPdL3]+ complex. However, if this were
occurring, a signal for this hypothetical complex should be present in the
31P NMR spectrum of an ArPdL2I compound in the presence of excess
phosphine. Garrou and Heck saw no evidence of peak broadening, chemical
shift changes, or coalescence when complex 1z was heated to 45 °C in the
presence of added phosphine: Garrou, P. E.; Heck, R. F. J. Am. Chem.
Soc. 1976, 98, 4113.
as PdL3, preventing the subsequent oxidative addition.
A
precedent for this was provided by Rubinskaya et al., who
isolated a zwitterionic palladium(0) compound upon heating an
(alkenyl)Pd(PPh3)4 complex.25 Here the phosphonium salt

12444 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
Goodson et al.
Figure 2. (A) Pd-Aryl region of the 1H NMR spectrum of complex 1z in CDCl3. (B) Stack plot of the reaction of perdeuterated complex 1y with
10 equiv of Ph4P+I- in CDCl3. The signal at δ 5.08 ppm is due to the trioxane internal standard.
Table 1. Observed Rate Constants for the Disappearance of 1i in
the Presence of Varying Amounts of Added Phosphine, LiI, and
LiPF6
formed has its own intramolecular Pd(0) trap which prohibits
oxidative addition of a P-C bond to form the interchanged
Pd(II) species.
To verify the intermediacy of phosphonium salts in the aryl-
aryl interchange reaction, the perdeuterated complex 1y was
heated at 50 °C in CDCl3 in the presence of 10 equiv of
tetraphenylphosphinium iodide to see if signals due to pal-
ladium-bound [1H5]phenyl groups would appear in the 1H NMR
spectrum (eq 3).26 A stack plot of this experiment is presented
[PPh3-d15]
[LiI]
[LiPF6]
kobs (×104 s-1
)
0.400
0.300
0.200
0.100
0.040
0.020
0.400
0.400
0.400
0.400
0.200
0.200
0.200
0.200
0.200
0.200
0.160
0.120
0.080
0
0
1.10 ( 0.02
1.11 ( 0.02
1.09 ( 0.02
1.34 ( 0.02
curvature
0
0
0
0
0
curvature
0.040
0.080
0.120
0.200
1.19 ( 0.02
1.08 ( 0.02
1.22 ( 0.02
1.48 ( 0.02
sumed during this reaction. However, Yamamoto noted that
the dimeric [Pd(Ph)(I)(PPh3)]2 with only one phosphine per
palladium was more susceptible to the elimination of the
phosphonium salt in methylene chloride than the monomeric
trans-[PdPhI(PPh3)2], suggesting that ligand dissociation may
indeed play a role in the mechanism.14d Furthermore, there is
ample literature precedent for the dependence of reductive
elimination reactions upon ligand predissociation.27 To test for
this experimentally, the kinetics of phosphonium salt formation
for complex 1i were followed in THF-d8 at 60 °C in the presence
of various concentrations of added LiI, LiPF6, and tri-
phenylphosphine-d15 (kinetic plots can be found in the Sup-
porting Information). The derived pseudo-first-order rate
constants are listed in Table 1. The disappearance kinetics of
1i appears to be first order in complex as were the kinetics for
the aryl-aryl interchange. At the high phosphine concentrations
required for pseudo-first-order conditions, the inhibitory effect
in Figure 2B. One can clearly see the anticipated signals
growing in with time. Figure 2A shows the same region of a
spectrum of the nondeuterated derivative, PhPd(PPh3)2I (1z),
for comparison.
If this phosphonium cation intermediate were indeed involved
in the aryl interchange reaction, one would expect to see the
same effects of excess iodide and phosphine on the observed
rate constants of phosphonium salt formation as one sees on
the rate constants of aryl-aryl exchange. That the phosphonium
salt formation would be facilitated by a predissociation of
phosphine might seem contradictory since phosphine is con-
(25) Rybin, L. V.; Petrovskaya, E. A.; Rubinskaya, M. I.; Kuz’mina, L.
G.; Struchkov, Yu. T.; Kaverin, V. V.; Koneva, N. Yu. J. Organomet. Chem.
1985, 288, 119.
(26) Norton and co-workers (ref 15) performed a similar experiment to
rule out phosphonium salt intermediates in alkyl-aryl interchanges (Vide
infra).
(27) For example, see: (a) Ozawa, F.; Ito, T.; Nakamura, Y.; Yamamoto,
A. Bull. Chem. Soc. Jpn. 1981, 54, 1868. (b) Tatsumi, K.; Hoffmann, R.;
Yamamoto, A.; Stille, J. K. Bull. Chem. Soc. Jpn. 1981, 54, 1857. (c)
Reference 18d.

Aryl-Aryl Interchange Reaction of ArPdL2I
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12445
Scheme 2. Proposed Mechanism for Aryl-Aryl Interchange and Phosphonium Salt Formation
appears to be saturated. Kinetics were also monitored with
lower concentrations of added phosphine, and the effect was
found to be much more pronounced. The curvature of the plots
(due to the fact that the phosphine concentration was not
constant during the course of the reaction), however, prevented
accurate determination of the pseudo-first-order rate constants.
The dependence of observed rate constant on flooded iodide
concentration is not nearly as significant as it was for the aryl-
aryl interchange. Rate constants with 0.2 M LiI and 0.04 M
LiI/0.16 M LiPF6 are essentially identical, and the constant for
0.2 M LiPF6 is only marginally faster. This discrepancy may
be due to the fact that different phosphines were used in the
two different kinetics studies.
One unfortunate drawback to the use of the deuterated
phosphine is that any aryl-aryl interchange that is occurring,
possibly via a pathway that does not involve intermediate
phosphonium salts, is masked since the growth of phenyl-d5
fragments bound to palladium would not be visible by 1H NMR.
Furthermore, since it is only the disappearance of 1i that is being
measured, other first-order chemical phenomena may be present
that would similarly not be visible in the observed kinetics. To
investigate these possibilities, a THF solution 0.02 M in 1i and
0.4 M in triphenylphosphine was monitored by 31P NMR before
and after submersion in a 50 °C bath for 10 h. The resonance
at 22.6 ppm due to 1i had diminished considerably after the
thermal treatment to be replaced predominantly by a signal at
21.9 ppm assigned as (4-methoxyphenyl)triphenylphosphonium
iodide by comparison with an authentic sample (δ ) 22.0 ppm).
However, smaller (but not negligible) resonances at 23.2, 23.4,
and 25.3 were also noticeable, indicating that other chemical
pathways, possibly including aryl-aryl interchange, were oc-
curring despite the presence of a large excess of triphenylphos-
phine. Because of this, the observed rate constants obtained
reflect the sum of all first-order processes occurring, and
consequently, quantitative absolute rate constant data again could
not be obtained.
new proposed mechanism is presented in Scheme 2. The exact
nature of the extremely coordinatively unsaturated intermediates
is uncertain, but with the presence of flooded phosphine and
their apparent activity towards oxidative addition, their lifetime
is almost certainly very short. Furthermore, such exposed
palladium species with as few as one phosphine have been cited
as intermediates in other reaction pathways,30 and a number of
palladium(0) compounds with only two phosphines for ligands
have been isolated.31 In addition, halides, as well as phosphines,
are known to stabilize coordinatively unsaturated palladium-
(0)32 and palladium(II)33 complexes. Unfortunately, due to the
presence of a non-zero k4, plots of log(kobs) vs log([L]) (or
log[I-]) or 1/(kobs) vs [L] (or [I-]) would not provide useful
information. Similarly, due to the curvature in the kinetics plots
for phosphine concentration less than 0.1 M and to the apparent
insensitivity of kobs on phosphine or iodide concentrations under
less dilute conditions, plots of kobs vs 1/[L] (or 1/[I-]) do not
provide much meaning. Furthermore, with the apparent pres-
ence of additional first-order reactions, even k4, the rate constant
for when iodide and phosphine effects are both saturated, cannot
be quantitatively determined. As a result, this mechanism
cannot be proven beyond doubt from the data presented here,
although it is consistent with the deuterated labeling experiment
described earlier as well as the qualitative trends observed in
the kinetics for both the aryl-aryl interchange and the disap-
pearance of 1i in the presence of excess phosphine.
(28) For example, see: (a) Reference 24. (b) Ozawa, F.; Ito, T.;
Yamamoto, A. J. Am. Chem. Soc. 1980, 102, 6457. (c) Sustman, R.; Lau,
J. Chem. Ber. 1986, 119, 2531.
(29) For example, see: (a) Byers, P. K.; Canty, A. J.; Crespo, M.;
Puddephatt, R. J.; Scott, J. D. Organometallics 1988, 7, 1363. (b) Aye,
K.-T.; Canty, A. J.; Crespo, M.; Puddephatt, R. J.; Scott, J. D.; Watson, A.
A. Organometallics 1989, 8, 1518.
(30) (a) Grushin, V. V.; Alper, H. Organometallics 1993, 12, 1890. (b)
Paul, F.; Patt, J.; Hartwig, J. F. J. Am. Chem. Soc. 1994, 116, 5969. (c)
Grushin, V. V.; Alper, H. J. Am. Chem. Soc. 1995, 117, 4035. (d) Hartwig,
J. F.; Paul, F. J. Am. Chem. Soc. 1995, 117, 5373.
(31) (a) van der Linde, R.; de Jongh, R. O. J. Chem. Soc., Chem.
Commun. 1971, 563. (b) Musco, A.; Kuran, W.; Silvani, A.; Anker, M. W.
J. Chem. Soc., Chem. Commun. 1973, 938. (c) Immirzi, A.; Musco, A. J.
Chem. Soc., Chem. Commun. 1974, 400. (d) Matsumoto, M.; Yoshioka,
H.; Nakatsu, K.; Yoshida, T.; Otsuka, S. J. Am. Chem. Soc. 1974, 96, 3322.
(e) Mann, B. E.; Musco, A. J. Chem. Soc. Dalton Trans. 1975, 1673. (f)
Kuran, W.; Musco, A. Inorg. Chim. Acta 1975, 12, 187. (g) Yoshida, T.;
Otsuka, S. J. Am. Chem. Soc. 1977, 99, 2134. (h) Urata, H.; Suzuki, H.;
Yoshihiko, M.; Ikawa, T. J. Organomet. Chem. 1989, 364, 235. (i) ref.
32b. (j) Paul, F.; Patt, J.; Hartwig, J. F. Organometallics 1995, 14, 3030.
(32) (a) Negishi, E.; Takahashi, T.; Akiyoshi, K. J. Chem. Soc., Chem.
Commun. 1986, 1338. (b) Amatore, C.; Azzabi, M.; Jutand, A. J.
Organomet. Chem. 1989, 363, C41. (c) Amatore, C.; Azzabi, M.; Jutand,
A. J. Am. Chem. Soc. 1991, 113, 8375. (d) Amatore, C.; Jutand, A.; Khalil,
F.; Nielsen, M. F. J. Am. Chem. Soc. 1992, 114, 7076. (e) Amatore, C.;
Jutand, A.; Suarez, A. J. Am. Chem. Soc. 1993, 115, 9531.
One aspect that is apparent from these kinetics experiments
is that in addition to mechanistic pathways involving loss of
ligand, there is another route, which does not involve predis-
sociation, by which phosphonium salt formation (and presum-
ably thereby aryl-aryl exchange) can occur. This latter pathway
was masked in studies of the interchange reaction by the trapping
of Pd(0) intermediates in the presence of excess phosphine.
There is precedent for such multipathway behavior in similar
stoichiometric reactions involving palladium(II)28 and palladium-
(IV).29
With these results in mind, we can append the mechanism
depicted in Scheme 1 to incorporate the reversible phosphonium
salt formation reaction as well as the apparent trapping of the
Pd(0) intermediates in the presence of excess phosphine. This
(33) (a) Klabunde, K. J.; Low, J. Y. F. J. Am. Chem. Soc. 1974, 96,
7674. (b) Kavaliunas, A. V.; Taylor, A.; Reike, R. D. Organometallics 1983,
2, 377.

12446 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
Goodson et al.
It is very interesting to note that the observations presented
here are in stark contrast to those reported by Norton on the
related alkyl-aryl interchange reaction (eq 1, R ) alkyl).15 This
reaction was completely unaffected by the addition of excess
phosphine. Furthermore, the authors were able to eliminate the
possibility of phosphonium salt intermediates by monitoring the
exchange of the perdeuterated complex, Pd(P(C6D5)3)2CH3I, in
the presence of methyltriphenylphosphonium triflate. The fact
that no undeuterated phenyl groups were incorporated into the
palladium phosphine complexes eliminated the possibility of
phosphonium cation intermediates. This disparity would seem
to suggest that, despite the apparent similarity of the two
interchange reactions, they proceed Via entirely different mech-
anisms. The novelty of this discrepancy can perhaps be
mollified if one considers that, although the aryl-aryl inter-
change reaction has generated a lot of recent interest due to
Kong and Cheng’s report,14a the problem of P-C cleavage,34-40
mediated by a variety of transition metals in both stoichio-
metric34-39 and catalytic34,40 reactions, is not a new one. In
these examples, authors have provided evidence for P-C
cleavage mechanisms involving oxidative addition of the P-C
bond,35f,h-j,l,o;37c;40a,c,e ortho-metalation of the phosphorus-bound
aryl group,35p,38a,b radicals,35n benzyne,38c and nucleophilic attack
at phosphorus,35b,d-e,k,39e as well as the afore-mentioned reduc-
tive elimination to form intermediate phosphonium salts.41
ortho-Metalation and benzyne intermediates can be ruled out
for the aryl-aryl interchange since the para regiochemistry on
the aryl groups is maintained throughout. Hartwig has found
byproduct formation due to aryl-aryl interchange to be unaf-
fected by the presence of radical traps, suggesting that radical
mechanisms are not involved.11f The phosphonium salt forma-
tion does bear a striking resemblance to the cleavage of
triphenylphosphine by Pd(OAc)2 studied by Matsuda.35b,d-e,k
However, this reaction is retarded by electron-donating groups
on the phosphine (supporting the nucleophilic attack at phos-
phorus mechanism) while aryl-aryl interchange (Vide infra) and
phosphonium salt formation are facilitated by electron-donating
groups on the phosphine.23a An additional aryl interchange
pathway via oxidative addition to form Pd(IV) intermediates
(as suggested by Kong and Cheng14a), however, cannot be ruled
out. (Indeed, it is possible that the additional first-order
processes indirectly observed by 31P NMR as described above
may be due to an alternative pathway for aryl-aryl interchange.)
Considering that there is convincing evidence for the interme-
diacy of Pd(IV) species in the reductive elimination of
(alkyl)2PdL2 compounds42 and that there is also strong evidence
against Pd(IV) intermediates in the reductive elimination of
RR′PdL2 (R,R′ ) aryl, alkenyl) complexes,43 it does not seem
too unreasonable to suggest that aryl-aryl interchange may
proceed predominately via a reductive elimination pathway
involving predissociation of phosphine or iodide, while alkyl-
aryl interchange may proceed via an oxidative addition pathway
in which predissociation is not involved. This interesting
disparity warrants further study.
Substituent effects of the aryl-aryl interchange reaction were
investigated by monitoring 0.021 M solutions of complexes
1a-x in CDCl3 before and after exposure to ambient and then
elevated temperatures for specified amounts of time. The results
of this experiment are listed in Tables 2 and 3. For complexes
of the same phosphine, it is clear that electron-donating groups
on the palladium-bound aryl group accelerate the interchange
reaction while electron-withdrawing groups appear to inhibit it
completely. This agrees well with the trends observed by Migita
on the substituent effects of phosphonium salt formation.23a
Similarly, for complexes with the same palladium-bound aryl
moiety, the exchange reaction is accelerated for complexes with
phosphines having electron-donating groups relative to those
having electron-withdrawing groups. Since a positive charge
is formed on the reductively eliminated phosphonium species,
a rate enhancement due to electron-donating substituents on the
phosphine is not surprising. Unfortunately, the equilibrium
effects are somewhat clouded by the competing decomposition
reaction (particularly with electron-rich species), but Kong and
Cheng found that electron-donating substituents on the pal-
ladium-bound aryl group drive the equilibrium to the right,14a
and our work, as well as that of others,14b-c,15,17c corroborates
this. One interesting comparison is with complexes e, i, and
m; although e approaches equilibrium the fastest (due to its more
electron-donating phosphine), the final equilibrium values of
percent 2 for i and m are higher. We believe that this is due to
the greater discrepancy between the electronics of the phosphine-
bound aryl groups and those bound to the metal; the larger the
difference between the free energies of 1 and 2, the larger the
value of Keq. In this table, complex q had not been allowed to
come to full equilibrium. Upon heating this sample for 24 h at
(34) For a review, see: Garrou, P. E. Chem. ReV. 1985, 85, 171.
(35) For examples of P-C cleavage in palladium compounds, see: (a)
Coulson, R. D. J. Chem. Soc., Chem. Commun. 1968, 1530. (b) Kikukawa,
K.; Yamane, T.; Takagi, M.; Matsuda, T. J. Chem. Soc., Chem. Commun.
1972, 695. (c) Asano, R.; Moritani, I.; Fujiwara, Y.; Teranishi, S. Bull.
Chem. Soc. Jpn. 1973, 46, 2910. (d) Yamane, T.; Kikukawa, K.; Takagi,
M.; Matsuda, T. Tetrahedron 1973, 29, 955. (e) Kawamura, T.; Kikukawa,
K.; Takagi, M.; Matsuda, T. Bull. Chem. Soc. Jpn. 1977, 50, 2021. (f)
Kikukawa, K.; Yamane, T.; Ohbe, Y.; Takagi, M.; Matsuda, T. Bull. Chem.
Soc. Jpn. 1979, 52, 1187. (g) Kikukawa, K.; Takagi, M.; Matsuda, T. Bull.
Chem. Soc. Jpn. 1979, 52, 1493. (h) Milstein, D.; Stille, J. K. J. Am. Chem.
Soc. 1979, 101, 4981. (i) Nishiguchi, T.; Tanaka, K.; Fukuzumi, K. J.
Organomet. Chem. 1980, 193, 37. (j) Gillie, A.; Stille, J. K. J. Am. Chem.
Soc. 1980, 102, 4933. (k) Kikukawa, K.; Matsuda, T. J. Organomet. Chem.
1982, 235, 243. (l) Abatjoglou, A. G.; Bryant, D. R. Organometallics 1984,
3, 932. (m) Goel, A. B. Tetrahedron Lett. 1984, 25, 4599. (n) Goel, A. B.
Inorg. Chim. Acta 1984, 84, L25. (o) Ortiz, J. V.; Havlas, Z.; Hoffman, R.
HelV. Chim. Acta 1984, 67, 1. (p) Bumagin, N. A.; Bumagina, I. G.;
Beletskaya, I. P. Zh. Org. Khim. 1984, 20, 457. (q) Reference 14.
(36) For examples of P-C cleavage in nickel compounds, see: (a) Green,
M. L. H.; Simth, M. J.; Felkin, H.; Swierczewski, G. J. Chem. Soc., Chem.
Commun. 1971, 158. (b) Nakamura, A.; Otsuka, S. Tetrahedron Lett. 1974,
463. (c) Reference 35f. (d) Reference 35l.
(37) For examples of P-C cleavage in cobalt compounds, see: (a)
Reference 35f. (b) Reference 35l. (c) Sakatura, T.; Kobayashi, T.; Hayashi,
T.; Kawabata, Y.; Tanaka, M.; Ogata, I. J. Organomet. Chem. 1984, 267,
171.
(38) For examples of P-C cleavage in osmium compounds, see: (a)
Bradford, C. W.; Nyholm, R. S.; Gainsford, G. J.; Guss, J. M.; Ireland, P.
R.; Mason, R. J. Chem. Soc., Chem. Commun. 1972, 87. (b) Bradford, C.
W.; Nyholm. R. S. J. Chem. Soc., Dalton Trans. 1973, 529. (c) Deeming,
A. J.; Kimber, R. E.; Underhill, M. J. Chem. Soc., Dalton Trans. 1973,
2589. (d) Reference 35l.
(39) For examples of P-C cleavage in compounds of other metals, see:
(a) Blickensderfer, J. R.; Kaesz, H. D. J. Am. Chem. Soc. 1975, 97, 2681.
(b) Reference 35i. (c) Reference 37c. (d) Reference 35l. (e) Chakravarty,
A. R.; Cotton, F. A. Inorg. Chem. 1985, 24, 3584.
(40) For examples of phosphine-derived byproducts in catalytic reactions,
see: (a) Fahey, D. R.; Mahan, J. E. J. Am. Chem. Soc. 1976, 98, 4499. (b)
Mitchell, R. H.; Chaudhary, M.; Dingle, T. W.; Williams, R. V. J. Am.
Chem. Soc. 1984, 106, 7776. (c) Dubois, R. A.; Garrou, P. E.; Lavin, K.
D.; Allcock, H. R. Organometallics 1984, 3, 649. (d) Dubois, R. A.; Garrou,
P. E.; Lavin, K. D.; Allcock, H. R. Organometallics 1986, 5, 460. (e) Dubois,
R. A.; Garrou, P. E. Organometallics 1986, 5, 466. (f) Brenda, M.; Greiner,
A.; Heitz, W. Makromol. Chem. 1990, 191, 1083. (g) Reference 17.
(41) (a) Reference 40b. (b) Pietrusiewicz, K. M.; Kuznikowski, M.
Phosphorus, Sulfur, and Silicon 1993, 77, 57. (c) Reference 14c. (d)
Reference 14d.
(42) (a) Stille, J. L.; Lau, K. S. Y. J. Am. Chem. Soc. 1976, 98, 5841.
(b) Reference 35h. (c) Reference 35j. (d) Moravskiy, A.; Stille, J. K. J.
Am. Chem. Soc. 1981, 103, 4182. (e) Loar, M. K.; Stille, J. K. J. Am. Chem.
Soc. 1981, 103, 4174. (f) Kurosawa, H.; Emoto, M.; Urabe, A. J. Chem.
Soc., Chem. Commun. 1984, 968. (g) Kurosawa, H.; Emoto, M.; Kawasaki,
Y. J. Organomet. Chem. 1988, 346, 137.
(43) (a) Reference 35p. (b) Ozawa, F.; Fujimori, M.; Yamamoto, T.;
Yamamoto, A. Organometallics 1986, 5, 2144. (c) Scott, W. J.; Stille, J.
K. J. Am. Chem. Soc. 1986, 108, 3033. (d) Ozawa, F.; Kurihara, K.;
Fujimori, M.; Hidaka, T.; Toyoshima, T.; Yamamoto, A. Organometallics
1989, 8, 180. (e) Reference 16b.

Aryl-Aryl Interchange Reaction of ArPdL2I
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12447
Table 2. Substituent Effects of Aryl-Aryl Interchange
Scheme 3. Aryl-Aryl Interchange under Catalytic
Conditions
2a (%)
1.5 h,
25 °C
8 h,
25 °C
1 h,
50 °C
3 h,
50 °C
facilitate the iodide-loss pathway since the charged intermediates
formed would be stabilized.
X
Y
a
b
c
d
e
f
g
h
i
j
k
l
m
n
o
p
q
r
s
OCH3
OCH3
OCH3
OCH3
CH3
CH3
CH3
CH3
H
H
H
H
F
F
F
F
CF3
CF3
CF3
CF3
CH3
H
F
CF3
OCH3
H
20
2
2
50
6
3
71
28
12
0
93
31
NAb
0
66
31
21
0
93
59
NAb
0
100
95
13
0
96
56
6
Time-Course Experiments. To evaluate the impact of aryl
interchange on both catalysis and polymerization, organic
emulsion cross-couplings in conjunction with deuterium labeling
and GC-MS analysis yielded time-course data on both cross-
coupling and aryl interchange. It is worth noting that the
concentration regimes of the kinetics experiments described
earlier and the catalytic procedures presented here are drastically
different. With this in mind, slight deviations in the results
obtained from the two types of experiments are to be expected.44
Kong and Cheng used deuterium labeling to confirm that
multiple aryl interchanges can occur on a single phosphorus
center under stoichiometric conditions, apparently due to rapid
intermolecular exchange of phosphines between ArPdL2I
complexes.14a The simplest modification of this approach to
studying catalytic cross-couplings is outlined in Scheme 3.
Bromobenzene-d5 was used to introduce phenyl-d5 fragments
into the cross-coupling reaction manifold. The fate of these
fragments could be determined semi-quantitatively by GC-MS
analysis. Catalysts based on nondeuterated phosphines were
employed; the evolution of the aryl component of these
phosphines could be compared with the incorporation of
phosphine-derived [1H5]phenyl fragments in the 4-methoxybi-
phenyl cross-coupling byproduct. Results of these studies are
compiled in Table 4. Relatively high loadings (1.0 mol %) of
catalyst precursors were used to obtain useful signal intensities
from the MS detector. Under these conditions, aryl interchange
was extensive for ArPdL2I complexes and less severe for Pd-
(PPh3)4 (compare entries 1-2 and 3-5) when bromobenzene-
d5 was the substrate. Different samples of Pd(PPh3)4 gave
different results. Material stored for some months at -30 °C
in a drybox underwent significantly greater interchange than a
freshly recrystallized sample. These materials were identical
0
0
41
trace
NAb
0
13
6
0
0
5
0
trace
0
trace
0
0
0
82
6
F
NAb
0
CF3
OCH3
CH3
F
CF3
OCH3
CH3
H
CF3
OCH3
CH3
H
52
22
trace
0
13
8
trace
0
7
99
64
5
0
64
22
2
0
27
6
0
69
13
7
trace
trace
trace
4
3
t
F
8
a {[2]/([1] + [2])} × 100. 1 refers to complex, regardless of
phosphine, where aryl interchange has not yet occurred; 2 refers to
complex, regardless of phosphine, where exchange has occurred.
b Resonances overlapped.
Table 3. Substituent Effects for Aryl-Aryl Interchange
2a (%)
1.5 h,
25 °C
8 h,
25 °C
1 h,
50 °C
3 h,
50 °C
Ar
L
u
v
j
w
i
p-tolyl
p-tolyl
p-tolyl
o-tolyl
p-anisyl
o-anisyl
DPPP
EtPPh2
PPh3
PPh3
PPh3
0
0
6
0
13
0
0
0
22
0
52
trace
8
5
15
20
95
trace
100
19
64
0
99
11
1
by visual inspection and H, 13C, and 31P NMR and possessed
x
PPh3
very similar catalytic behaviors. This discrepancy remains
unexplained.
a {[2]/([1] + [2])} × 100. 1 refers to complex, regardless of
phosphine, where aryl interchange has not yet occurred; 2 refers to
complex, regardless of phosphine, where exchange has occurred.
Observations made at 0.021 M in CDCl3.
Quantitation of all phosphines allowed a determination of the
exchanged fraction of the initial complement of phosphine-
bound aryls. Observed distributions of exchanged phosphines
measured using this method could be compared with probabi-
listic distributions. In general, a set of n nonordered independent
events with binary outcomes (i.e., heads/tails, true/false, inter-
change/no interchange, etc.) has n + 1 possible outcomes. If
the individual events are truly independent, then the most
probable distribution of outcomes for a large ensemble of sets
of events is given by the binomial frequency function (B(x))
(eq 4), where n represents the number of independent trials in
50 °C, it too had been completely converted to 2. Comparison
of complexes u and v indicate that chelating ligands have little
effect on the interchange reaction. However, the drastic contrast
of the results for j with w and i with x suggest that ortho
substituents on the metal-bound aryl group have a strong
retardation effect on aryl-aryl exchange. This also agrees with
the results of others for both the interchange14b and phosphonium
salt formation reactions.23a
In benzene-d6, resonances for 1m and 2m overlapped. In
chlorinated solvents, the competing decomposition prevented
the determination of observed rate constants for aryl-aryl
exchange. In DMSO-d6, the exchange was complete in the time
it took to shim the NMR and start the automated acquisition
program (ca. 5 min). However, qualitatively, it appears that
the interchange is accelerated by more polar solvents. This trend
would support the formation of a charged intermediate in the
rate-determining step of the reaction. Polar solvents would also
a set, x represents the number of positive outcomes, and p
represents the relative frequency of success (i.e., the statistical
bias in the ratio of yes:no answers for an individual event).45
(44) For a discussion on possible mechanistic differences between
catalytic and stoichiometric cross-coupling reactions, see ref 34e.

12448 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
Table 4. Aryl Exchange During Cross-Couplinga
Goodson et al.
final composition of triphenylphosphinesc
entry
[C6D5Br] (mM)
cat.
10
10
PdL4
PdL4
PdL4
temp, °C
final exchb
d0
d5
d10
d15
1
2
3
4
5
75
75
75
37.5
75
50
65
50
50
65
52
65
23.4
12.0
5.3
7.1 (11.0)
3.6 (4.2)
44.0 (44.9)
67.2 (68.0)
84.0 (84.8)
38.9 (35.9)
24.3 (23.9)
42.5 (41.2)
29.4 (29.0)
16.0 (14.6)
43.6 (38.9)
44.7 (44.4)
12.8 (12.6)
3.4 (3.8
10.4 (14.1)
27.5 (27.5)
0.8 (1.3)
0 (0.2)
d
e
e
0 (0.8)
0 (0)
a All reactions went to 95% completion or better, as determined by GC-MS. b Percent of all phosphine-bound [1H5]phenyl moeities replaced by
phenyl-d5 moieties as determined by GC-MS. Values are the average of several measurements. (See the Experimental Section.) c Expressed as
percentage of total composition. Values in parentheses are expected distributions calculated from the measured percent exchange. d L ) PPh3.
e Results obtained using freshly recrystallized Pd(PPh3)4.
Figure 4. Phosphine distributions for 65% exchange (tris(dibenzyli-
deneacetone)dipalladium(0) at 65 °C, entry 2 in Table 4).
Figure 3. Cross-coupling and aryl interchange in the presence of tris-
(dibenzylideneacetone)dipalladium(0) at 50 °C (Table 4, entry 1).
specifics require further investigation. The suppression of aryl
interchange with concurrent decrease in catalytic activity
suggests either that an excess of phosphine relative to homo-
geneous palladium centers is generated (i.e., through decom-
position of the catalyst) or that phosphines are decoupled not
only from aryl interchange but from any homogeneous catalytic
process. In the latter case, the catalytically active species may
be a “poisoned” ligandless colloid analogous to Lindlar’s
catalyst. Since phosphines are known to produce homogeneous
complexes by reacting with supported heterogeneous pal-
ladium,47 the second possibility appears to be the more distant
one.
In the presence of Pd(PPh3)4, cross-coupling occurs in an
almost linear manner and aryl interchange is largely suppressed
during the initial 75% conversion. As the reaction nears
completion, aryl interchange becomes more competitive. One
possible explanation is that this behavior is due to the competing
phosphonium salt formation. Early in the reaction, the excess
phosphine suppresses transfer (Vide supra) and inhibits cross-
coupling relative to the ArPdL2I species.48c However, as the
reaction progresses and the excess phosphine is consumed, the
catalyst gradually becomes more active (giving rise to the zero-
order-like behavior) and the aryl-aryl interchange is allowed
to proceed. The evolution of the phosphine distribution can be
monitored throughout the reaction; experimental distributions
closely mirror the most probable distribution in all cases (Figure
5).
Experimentally measured distributions allowed us to determine
the fraction of all phosphine-bound aryl rings interchanged.
Predicted distributions for independent exchange processes could
then be determined using eq 4 with n ) 3 (the maximum number
of exchanges per phosphine), x ) 0-3 (the actual number of
exchanges per phosphine), and p equal to the experimentally
measured fraction of all aryl rings interchanged. Experimental
and predicted distributions were generally in excellent agree-
ment. This result provides further evidence of reversibility in
the interchange process, since it indicates that interchange of
any given aryl moiety bound to a phosphine is statistically
independent of any other interchange on the same phosphine.46
Aryl rings can thus be expected to continue to interchange in
catalytic reactions until the equilibration pathway is deactivated
by catalyst decomposition, other side reactions that consume
phosphines, or completion of the reaction.
Time-course experiments revealed remarkable differences in
the behavior of cross-couplings with the two catalyst precursors.
In the presence of the ArPdL2I precursor, cross-coupling occurs
very rapidly in early portions of the reaction; the conversion
then tapers off. All aryl interchange occurs during the initial
burst of catalytic activity; the phosphines become completely
decoupled from the interchange process following this initial
burst (Figure 3). Nevertheless, the aryl interchange process
remains statistically well-behaved (Figure 4). This behavior
strongly suggests a change in catalytic mechanism whose
(45) Sokolnikoff, I. S.; Redheffer, R. M. Mathematics of Physics and
Modern Engineering, 2nd ed.; McGraw-Hill: New York, 1966; pp 617-
620.
Screening of Possible Strategies for the Elimination of the
Aryl-Aryl Interchange Reaction in Palladium-Mediated
Cross-Couplings. Given the information gained from the above
mechanistic studies, the simplest method to eliminate the aryl-
aryl interchange reaction from palladium-mediated cross-
couplings would be to eliminate phosphine ligands altogether
from the cross-coupling reactions. In an earlier paper,48c we
(46) This argument holds only for electronically and sterically equivalent
interchanges. It also assumes that deuterium isotope effects for the
interchange of phenyl-d5 and [1H5]phenyl rings are negligible. A negligible
isotope effect is in fact expected due to the at least two-bond separation
between the reaction center and a perturbing isotope, as well as the geometry
of the aryl ring, and is strongly confirmed by the fidelity of the experimental
interchange distributions to the predicted distributions. For a discussion of
secondary isotope effects, see: Carpenter, B. K., Determination of Organic
Reaction Mechanisms; Wiley: New York, 1984; pp 96-100.
(47) De la Rosa, J. M. A.; Velarde, E.; Guzman, A. Synth. Commun.
1990, 20, 2059.

Aryl-Aryl Interchange Reaction of ArPdL2I
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12449
Figure 5. Evolution of phosphine distribution in cross-coupling reactions (Pd(PPh3)4 at 50 °C; Table 4, entry 3). Boxed numbers indicate total
percent exchange.
expanded upon the work of Beletskaya48b and showed that
simple palladium precursors such as palladium(II) acetate, tris-
(dibenzylideneacetone)dipalladium(0), and allylpalladium(II)
chloride dimer are effective catalysts for the Suzuki coupling
reaction with acetone as the organic solvent, without the addition
of phosphines or other ligands. For most small-molecule
applications, we believe that this is the best way to eliminate
the formation of undesired side products caused by the
incorporation of the ligands into the coupling cycle. However,
in the case of our poly(p-phenylene) polymerizations, this system
was not compatible due to the insolubility of the polymers in
the acetone/water solvent medium. Furthermore, for cross-
coupling reactions with other organometallic reagents, this
medium may not be compatible due to reactivity concerns. Due
to these factors, we sought alternative strategies to eliminate
the aryl-aryl exchange.
Unfortunately, deviation from the acetone/water solvent
system led to almost universal failure. In general, the use of
more polar solvents led to more improved results, but the failure
of DMSO to promote quantitative coupling suggests that this
generalization is not universal. With the failed reactions, a
substantial amount of coupling did occur (generally 80% or
better), but as condensation polymerizations require quantitative
conversion to produce polymers of adequate molecular weight,
the results obtained from the “ligandless” systems were unac-
ceptable for our poly(p-phenylene) polymerizations.
phines (e.g., tricyclohexylphosphine) under the rationale that,
since the alkyl-aryl exchange was found to be irreversible,15
there should be no migration of alkyl groups from the phos-
phines to the palladium center under catalytic conditions. In
addition to employing the above three ligands, pyridine, 2,2′-
bipyridine, triphenylarsine, and 1,2,5-triazaphosphaadamantane
were also examined. With the exception of triphenylarsine,
which unfortunately is also known to undergo the aryl-aryl
interchange reaction in palladium complexes,14c,15 none of these
ligand systems promoted the Suzuki coupling reaction to
quantitative conversion.
As increased steric bulk was found to inhibit both the aryl-
aryl interchange reaction (Vide supra) and the phosphonium salt
formation reaction,23 the use of bulky triarylphosphines should
minimize the occurrence of these side reactions under catalytic
conditions. Indeed, tris(o-methoxyphenyl)phosphine has seen
use in suppressing byproduct formation in Stille couplings,14a
and tris(o-tolyl)phosphine has been shown to inhibit phospho-
nium salt formation8d and aryl-aryl exchange in Heck and
Suzuki couplings.14c,d Products derived from the exchange
reaction are indeed significantly reduced but not eliminated.51
Exchange percentages of 0.194 ( 0.023 and 1.80 ( 0.32% were
measured using the bulky tris(o-methoxyphenyl)phosphine and
tris(o-tolyl)phosphine ligands, respectively. However, further
increase of the steric bulk on the phosphine (i.e., tris(2-
trifluorophenyl)phosphine, tris(2,6-dimethylphenyl)phosphine,
and tris(2,6-dimethoxyphenyl)phosphine) resulted in the deac-
tivation of the catalyst as evidenced by incomplete conversion
of starting materials.
The next logical strategy was to use non-phosphine-based
ligands under typical Suzuki coupling conditions. For example,
Elsevier has had excellent success in utilizing acenaphtho-
quinone-based bidentate imine ligands49 in palladium-mediated
cross-couplings between aryl or benzyl halides and organotin
or organomagnesium reagents.50 Other possibilities were to use
phosphite ligands (e.g., triphenyl phosphite), or trialkylphos-
With these results in consideration, the tris(o-tolyl)phosphine
catalyst system appears to be optimal for minimizing the aryl-
aryl interchange during palladium-mediated cross-coupling
reactions. We then undertook additional experiments to try and
optimize the system further. As can be seen in Table 5, for
(48) (a) Sekiya, A.; Ishikawa, N. J. Organomet. Chem. 1977, 125, 281.
(b) Beletskaya, I. P. J. Organomet. Chem. 1982, 250, 551 and references
therein. (c) Wallow, T. I.; Novak, B. M. J. Org. Chem. 1994, 59, 5034.
(49) van Asselt, R.; Elsevier, C. J.; Smeets, W. J. J.; Spek, A. L.; Benedix,
R. Recl. TraV. Chim. Pays-Bas 1994, 113, 88.
(50) (a) van Asselt, R.; Elsevier, C. J. Organometallics 1992, 11, 1999.
(b) van Asselt, R.; Elsevier, C. J. Organometallics 1994, 13, 1972. (c) van
Asselt, R.; Elsevier, C. J. Tetrahedron 1994, 50, 323.
(51) Heitz et al. also found a trace amount of tris(o-tolyl)phosphine-
derived byproducts when this ligand was used in Heck couplings: ref 40f.

12450 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
Goodson et al.
Table 5. Analytical Experiments for the Optimization of
Conditions for Suzuki Coupling Reactions
take place. This effect is evident in the general trend (most
visible in the triphenylphosphine-d15 entries) that the ratio of
percent observed transfer to percent possible transfer decreases
with increased amount of catalyst.
In the experiments presented in this section, the catalyst was
formed in situ by adding 2 equiv of ligand to tris(dibenzyli-
deneacetone)dipalladium(0). Although not apparent from the
experiments performed here, some polymerization results that
will be presented elsewhere53 suggest that the dibenzylidene-
acetone released from the in situ catalyst formation may be
involved with the catalytic cycle as well. To eliminate this
possibility, we suggest the use of bis[tris(o-tolyl)phosphine]-
palladium(0)30b,31j as a catalyst for cross-coupling reactions when
the “ligandless” methodology48 is not suitable. Control experi-
ments analogous to those presented in Table 4 showed com-
parable results for this catalyst as for that generated in situ from
tris(dibenzylideneacetone)dipalladium(0) and tris(o-tolyl)phos-
phine.
Ar
solvent % cat.
% excha
% possible exchb
C6D5
C6D5
C6D5
C6D5
C6D5
2-MePh THF
2-MePh THF
2-MePh THF
2-MePh THF
2-MePh THF
2-MePh CH2Cl2
2-MePh CH2Cl2
2-MePh CH2Cl2
2-MePh CH2Cl2
2-MePh CH2Cl2
THF
THF
THF
THF
THF
0.2
0.5
1.0
2.0
5.0
0.2
0.5
1.0
2.0
5.0
0.2
0.5
1.0
2.0
5.0
1.10 ( 0.12
2.76 ( 0.13
4.94 ( 0.05
91.7 ( 10.0
92.0 ( 4.3
82.3 ( 0.8
88.1 ( 2.3
55.9 ( 1.7
28.3 ( 2.9
7.27 ( 0.80
5.18 ( 0.80
1.03 ( 0.28
0.647 ( 0.077
0.51 ( 0.12
10.57 ( 0.29
16.76 ( 0.49
0.340 ( 0.035
0.218 ( 0.024
0.311 ( 0.048
0.124 ( 0.034
0.194 ( 0.023
0.0061 ( 0.0014
0.00281 ( 0.00014 0.094 ( 0.005
0.0081 ( 0.0010 0.14 ( 0.02
0.01066 ( 0.00029 0.0889 ( 0.0025
0.0276 ( 0.0012 0.092 ( 0.004
Conclusions
a (Amount of phosphine-derived biphenyl divided by the total amount
of biphenyl) × 100 as determined by GC-MS. b (% exchange found
divided by theoretical % exchange if all phosphine-bound aryl groups
were incorporated into the coupling cycle) × 100. Values given are
the average of five trials. The listed errors are the standard deviations
for those trials.
To summarize, evidence presented in this paper supports the
suggestion of Chenard14c and Yamamoto14d that, unlike the
documented alkyl-aryl exchange reaction,15 the aryl-aryl
interchange reaction of ArPdL2X proceeds first through a
reductive elimination to form a phosphonium salt followed by
an oxidative addition of a different phosphorusn-carbon bond.
The interchange and phosphonium salt formation reactions alike
are at least facilitated by predissociation of either phosphine or
iodide. Furthermore, like the phosphonium salt reaction,23 the
aryl-aryl interchange is promoted by more polar solvents and
inhibited by increased steric bulk. Under catalytic conditions,
the distribution of phosphines formed from the aryl-aryl
interchange can be modeled by statistics. For most synthetic
applications, probably the best strategy to eliminate phosphine-
derived byproducts is to use the “ligandless” methodology
described elsewhere.48 However, if this protocol is unsuitable
due to reasons of insolubility, reactivity, etc., we suggest the
use of bis[tris(o-tolyl)phosphine]palladium(0)30b,31j as a catalyst
and a nonhydrophilic organic solvent such as CH2Cl2 to
minimize the effects of the aryl-aryl interchange reaction.
couplings carried out in THF, reactions utilizing the tris(o-tolyl)-
phosphine ligand resulted in the formation of much smaller
amounts of phosphine-derived byproducts as compared to those
utilizing triphenylphosphine-d15. (The deuterated phosphine was
employed to distinguish between biphenyl derived from the
phosphine and biphenyl derived from homocoupling of phen-
ylboronic acid, an unrelated side reaction ubiquitous in pal-
ladium-mediated cross-couplings.52) This difference, however,
is less pronounced for the trials utilizing smaller amounts of
catalyst (which are more typical conditions for actual synthetic
reactions). Byproduct formation can be suppressed further by
switching to a less hydrophilic solvent such as methylene
chloride. We believe that this is due to a difference in polarity.
Although pure methylene chloride has a higher dielectric
constant than pure THF, the latter is more water-miscible.
Consequently, more water is able to partition into the organic
phase of the emulsion, resulting in a more polar environment
which is favorable for the aryl-aryl interchange. The optimal
conditions appear to be with 0.5% catalyst, tris(o-tolyl)-
phosphine, and methylene chloride as solvent, resulting in a
product that was contaminated by only 0.003% of the phosphine-
derived 2-methylbiphenyl byproduct. It is interesting to note
that attempts at performing the triphenylphosphine-d15 trials in
CH2Cl2 resulted in incomplete conversions. This was likely
due to the decomposition (Via phosphonium salts) of ArPdL2I
intermediates in chlorinated solvents that was observed earlier.
The effect of catalyst concentration on the amount of byproduct
formed seems to be a compromise between two competing
factors. Larger amounts of catalyst require larger amounts of
phosphine, which can in turn supply larger amounts of aryl
groups available for the interchange reaction. However, larger
amounts of catalyst also lead to faster rates of cross-coupling.
As a result, there is less time for the interchange reaction to
Experimental Section
General Procedure. Schlenk-line or drybox techniques were used
for all air-sensitive manipulations. 1H NMR spectra were acquired at
200, 300, 400, or 500 MHz using Bruker AC-series, AM-series, MSL
series, and AMX-series spectrometers; proton-decoupled 13C, 19F, and
31P spectra were obtained at corresponding frequencies. 1H and 13C
chemical shifts are reported relative to internal TMS; 31P chemical shifts
are reported relative to an external standard of triphenyl phosphite (δ
) 127.0 ppm); 19F chemical shifts are reported relative to an external
standard of fluorobenzene (δ ) -113.1 ppm). THF, toluene, and
hexanes were purified by distillation from sodium/benzophenone and
used immediately. CH2Cl2, CDCl3, tetrachloroethane, tetrachloroethane-
d2, and dibromomethane were dried over CaH2 and vacuum-transferred
prior to use. THF-d8 and benzene-d6 were vacuum-transferred from a
sodium/benzophenone ketyl. Triphenylarsine, triphenylphosphine, and
triphenylphosphine-d15 were purchased from Aldrich, recrystallized from
degassed ethanol, and sublimed under vacuum proir to use. Authentic
samples of 4-methylbiphenyl, 2-methylbiphenyl, and 2-methoxybiphe-
nyl were obtained from Aldrich and sublimed or distilled prior to use,
as were the aryl halides and trioxane. Lithium trifluoromethane-
sulfonate was purchased from Aldrich and dried for 48 h under vacuum
at 100 °C. The benzene adduct of tris(dibenzylideneacetone)di-
(52) A detailed understanding of how scrambled products arise remains
elusive despite careful investigation: (a) Negishi, E.; Takahashi, T.;
Akiyoshi, K. J. Organomet. Chem. 1987, 331, 334. In Stille couplings,
homocoupling of arylstannanes occurs in the presence of adventitions
oxygen: (b) Farina, V.; Krishnan, B.; Marshall, D. R.; Roth, G. P. J. Org.
Chem. 1993, 58, 5434. Homocoupled products predominate in certain Suzuki
couplings. This can be synthetically useful: (c) Song, Z. Z.; Wong, H. C.
J. Org. Chem. 1994, 59, 33.
(53) Goodson, F. E.; Wallow, T. I.; Novak, B. M. Manuscript in
preparation.

Aryl-Aryl Interchange Reaction of ArPdL2I
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12451
palladium(0),54 tetrakis(triphenylphosphine)palladium(0),55 phenylboric
anhydride,48c 2-(4′-methoxyphenyl)-1,4,3-dioxaborolane,56 tris(2,6-di-
methylphenyl)phosphine,57 tris[2-(trifluoromethyl)phenyl]phosphine,58
and bis(p-tolylimino)acenaphthene49 were all prepared via literature
procedures. Other chemcals were used as received from commercial
suppliers. The preparation of ArPdL2I complexes except 1y,z is
reported elsewhere.21 FAB Mass spectra were performed by the U.C.
Berkeley Mass Spectrometry Laboratory. Analytical data were obtained
by the elemental analysis facilities at the University of California at
Berkeley and the University of Massachusetts at Amherst. GC-MS
measurements were performed on either a Hewlett-Packard 5890A gas
chromatograph in line with a 5970 Series mass-selective detector or a
Hewlett-Packard 5890 Series II gas chromatograph in line with a 5972
Series mass-selective detector. In either case, the instrument was
equipped with a polysiloxane capillary column and helium was used
as the carrier gas.
(Phenyl-d5)Pd[P(C6D5)3]2I (1y). A procedure exactly analogous to
the syntheses of 1a-x21 was followed to give 1y in 82% yield as a
pale greenish-yellow powder: IR (neat) 2956 (w), 2856 (w), 1528 (m),
1308 (s), 1046 (m), 1006 (w), 956 (w), 835 (s); 13C NMR (75 MHz,
CDCl3) no identifiable resonances due to multiple coupling with
deuterium and phosphorus atoms; 31P NMR (121 MHz, CDCl3) δ 22.0.
Anal. Calcd for C42D35IP2Pd (assuming 98% D from iodobenzene-d5,
99% D from triphenylphosphine-d15): C, 58.30; D, 8.05. Found: C,
58.49; D, 8.06.
(Phenyl)Pd[P(C6H5)3]2I (1z).59 A procedure exactly analogous to
the syntheses of 1a-x21 was followed to give 1z in 85% yield as a
pale greenish-yellow powder: 1H NMR (200 MHz, CDCl3) δ 6.20 (t,
J ) 7.0 Hz, 2H), 6.32 (tm, Jt ) 7.0 Hz, 1H), 6.59 (dm, Jd ) 6.7 Hz,
2H), 7.17-7.35 (m, multiple resonances, 18H), 7.49 (m, 12H) [lit.34d
(250 MHz, CDCl3) δ 6.25 (d, J ) 7.5 Hz, 2H), 6.36 (t, J ) 7.5 Hz,
1H), 6.63 (d, J ) 7.5 Hz, 2H), 7.3 (m, 18H), 7.54 (m, 12H)]; 13C NMR
(75 MHz, CDCl3) δ 127.2 (apparent t, Japp ) 22.5 Hz), 129.1 (apparent
t, Japp ) 22.0 Hz), 132.0 (apparent t, Japp ) 22.0), 134.4 (apparent t,
Japp ) 19.9 Hz), other resonances not resolved; 31P NMR (202 MHz,
4:1 CH2Cl2:CDCl3) δ 22.2 [lit.24 (solvent and field strength not listed)
δ 22.8].
2,3,4,5-Pentadeuteriobiphenyl. This compound was synthesized
by following a known protocol.48c A 100 mL Schlenk flask was charged
with 1.93 g (11.9 mmol) of bromobenzene-d5, 1.36 g (13.1 mmol as
phenylboric acid) of phenylboronic anhydride, 4.91 g (29.7 mmol) of
potassium carbonate sesquihydrate, 25 mL of acetone, 25 mL of HPLC
grade water, and a stirbar. This mixture was then degassed via several
freeze-pump-thaw cycles, and 0.185 mL (2.39 µmol, 0.02%) of a
0.0129 M (in Pd) solution of 9 was added via syringe. The flask was
degassed once more and then heated in a 50 °C bath overnight. The
reaction started out as a tan, biphasic mixture and finished as an off-
white, triphasic suspension. The reaction was transferred to a separatory
funnel, and the water phase was isolated, and washed with 3 × 50 mL
of ether. The organic layers were combined and dried over MgSO4,
and the solvent was removed with a rotary evaporator to give an orange,
greasy solid. This was repeatedly recrystallized from acetone/water
until colorless and then sublimed at 50 °C, dynamic 5 mTorr vacuum,
to give large, flaky, white crystals of 2,3,4,5-pentadeuteriobiphenyl:
1.53 g (9.62 mmol) recovered, 80.7% yield: mp 69-71 °C (acetone/
H2O) (lit.60 mp 63-66 °C); 1H NMR (200 MHz, CDCl3) δ 7.27-7.45
(m, multiple resoncances, 3H), 7.57 (dm, Jd ) 6.8 Hz, 2H); MS (m/z)
159.
1m were weighed into four separate 2 mL volumetric flasks on which
the 1.80 mL level (calibrated by weight of toluene) had been marked.
These were each dissolved in 1 mL of dry THF-d8 prechilled to -40
°C and diluted to the 1.80 mL mark, yielding stock solutions of 0.0206,
0.0411, 0.0822, and 0.164 M in concentration. Aliquots (600 µL) of
these solutions were then filtered through 0.2 µm syringe filters into 5
mm NMR tubes which were then capped with septa and removed from
the glovebox. Dibromomethane (1 µL) was then added to each sample,
and the tubes were degassed via three freeze-pump-thaw cycles (or
until completely degassed) and sealed under vacuum. The tubes were
stored at -40 °C prior to the kinetics experiments.
Kinetics were observed via 1H NMR at 400 MHz on a Bruker AM-
400 spectrometer. The temperature of the probe was maintained at
50.0 ( 0.5 °C calibrated with 100% ethylene glycol and monitored
with an internal thermocouple. Single scans were then taken at a 90°
pulse approximately every 5-6 min. The exact time was monitored
via stopwatch. The aryl-aryl exchange reaction was monitored by
the decrease of the palladium-bound aryl signals of 1m and the growth
of the corresponding signals for 2m, integrated relative to the
dibromomethane standard. The results were fit to an exponential
function via least-squares analysis with Igor, a curve-fitting program
for the Apple Macintosh computer. This analysis yielded observed
rate constants (kobs) that were equal to the sum of the rate constants for
the forward (kF) and reverse (kR) exchange reactions.22 Alternatively,
plots of ln[(C - Cinf)/(C0 - Cinf)] vs time (where C, Cinf, and C0
correspond to integrations ant time ) t, infinity, and 0, respectively)
yielded straight lines with slopes of kobs. Experimental errors were
estimated from the least-squares analyses. The observed rate constant
values thus determined were (7.57 ( 0.09) × 10-4, (7.60 ( 0.17) ×
10-4, (7.30 ( 0.09) × 10-4, and (7.08 ( 0.09) × 10-4 s-1 for initial
concentrations of 1m of 0.164, 0.082, 0.041, and 0.021 M, respectively.
(b) Phosphine Inhibition Experiment. In a glovebox, 0.0411,
0.004 11, and 0.008 22 M stock solutions of tris(p-fluorophenyl)-
phosphine in dry, degassed CDCl3 were prepared in 2.00 mL volumetric
flasks. These solutions were chilled to -40 °C, while samples of 72.0
mg (74.0 µmol) of 1m were weighed into four new vials. Three of
these samples were dissolved in 1.80 mL of the respective phosphine
stock solutions, while the fourth was dissolved in 1.80 mL of neat
CDCl3. In this manner, sample solutions of approximately 0.0411 M
in 1m with 1, 0.2, 0.1, and 0 equiv of added phosphine were obtained.
Aliquots (600 µL) of these solutions were then filtered through 0.2
µm syringe filters into 5 mm NMR tubes which were then capped with
septa and removed from the glovebox. Tetrachlorethane (1 µL) was
added via syringe to each tube, and the samples were sealed as before
and stored at -40 °C prior to the kinetics experiments. Kinetics were
observed via 1H NMR at 400 MHz on a Bruker AM-400 spectrometer
at 60.0 ( 0.5 °C by following the procedure described above. The
samples with added phosphine showed a marked inhibition in rate (see
Figure 1), but due to the complexity of the kinetics observed, no
meaningful rate constants could be derived.
An identical experiment was attempted with 0.4 equiv of added
phosphine in THF-d8. Again, a marked inhibition was observed, but
the kinetics were too complex to derive meaningful rate constant data.
Also, a small amount white precipitate formed during the kinetics run.
Upon analysis by 1H, 19F, and 31P NMR, as well as FAB mass
spectrometry, this material was found to be an approximately 4:1
mixture of two major products, as well as a trace of a third, that were
not separated. Major product (3): 1H NMR (400 MHz, CDCl3) δ 3.96
(s, 3H), 7.28 (td, Jt ) 7.0 Hz, Jd ) 2.4 Hz, 2h), 7.47 (m, 6H), 7.62
(dd, Jd ) 10.0 Hz, Jd ) 7.1 Hz, 2 H), 7.76 (m, 6H); 19F NMR (376
MHz, CDCl3) δ -98.4; 31P NMR (162 MHz, CDCl3) δ 21.2; MS (m/
z) 423. Second product (4): 1H NMR (400 MHz, CDCl3) δ 7.57 (dd,
Jd ) 9.9 Hz, Jd ) 7.1 Hz, 2H), 7.71 (m, 2H), three additional resonances
not resolved; 19F NMR (376 MHz, CDCl3) δ -99.2; 31P NMR (162
MHz, CDCl3) δ 20.9; MS (m/z) 435. Third product (5): MS(m/z) 411.
(c) Iodide Inhibition Experiment. In a drybox, 0.206 M stock
solutions of LiI and LiPF6 in dry THF-d8 were prepared in 2.00 mL
volumetric flasks. To each of these was added 1 mg of trioxane, and
the resulting solutions were chilled to -35 °C. 1m (48.0 mg, 49.3
µmol) was then weighed into three separate vials. The first of these
samples was dissolved in 1.20 mL of the LiI stock solution, the second
was dissolved in 1.20 mL of the LiPF6 stock solution, and the third
Kinetics Experiments for Aryl-Aryl Exchange. (a) Concentration
Profile. In a glovebox, 36.0 mg (37.0 µmol), 72.0 mg (74.0 µmol),
144.0 mg (0.1480 mmol), and 288.0 mg (0.2960 mmol) samples of
(54) Ukai, T.; Kawazura, H.; Ishii, Y. J. Organomet. Chem. 1974, 65,
253.
(55) Cotton, F. A. Inorg. Synth. 1990, 28, 342.
(56) Kaminski, J. J.; Lyle, R. E. Org. Mass Spectrom. 1978, 13, 425.
(57) Stepanov, B. I.; Matyuk, V. M.; Bokanov, A. I.; Karpova, E. N.
Zh. Obshch. Khim. 1977, 45, 2096.
(58) Eapen, K. C.; Tamborski, C. J. Fluorine Chem. 1980, 15, 239.
(59) Howell, J. A. S.; Palin, M. G.; Yates, P. C.; McArdle, P.;
Cunningham, D.; Goldschmidt, Z.; Gottlieb, H. E.; Hezroni-Langerman,
D. J. Chem. Soc., Perkin Trans. 2 1992, 1769.
(60) Cooks, R. G.; Howe, I.; Tam, S. W.; Williams, D. H. J. Am. Chem.
Soc. 1968, 90, 4064.

12452 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
Goodson et al.
was dissolved in 600 µL of each, resulting in solutions of the same
overall salt concentration that were approximately 0.041 M in complex
with 5, 0, and 2.5 equiv of excess iodide, respectively. Samples were
then prepared and stored as described above. Kinetics were observed
via 1H NMR at 500 MHz on a Bruker AMX-500 spectrometer at 50.0
( 0.5 °C by following the procedure described above. As a check,
kinetics were also observed for a solution 0.206 M in lithium
trifluoromethanesulfonate and 0.041 M in 1m. The observed rate
constant ((2.03 ( 0.02) × 10-3 s-1) was almost identical to that
observed for the kinetics monitored in the presence of 0.206 M LiPF6
((1.91 ( 0.04) × 10-3 s-1), suggesting that the latter salt is not
interfering with the aryl-aryl interchange reaction.
every 20 min, and the progress of the reaction was evidenced by the
growth of proton signals due to Pd-bound phenyl-h5 groups as shown
in Figure 3.
Monitoring Phosphonium Salt Formation by Phosphorus NMR.
In a drybox, 84.8 mg (98.0 µmol) of 1i and 514.1 mg (1.96 mmol) of
triphenylphosphine were dissolved in 4.8 mL of THF. Half of this
solution was placed in a scintillation vial and stored overnight in the
dark at -35 °C (the “before” sample), while the other half was
transferred into a reaction tube with a Teflon stopcock and sidearm. A
stirbar was added, and the tube was sealed, removed from the drybox,
and placed in a 60 °C oil bath. After the mixture was stirred at 60 °C
for 10 h, during which time a precipitate had formed and the reaction
had turned from a pale greenish-yellow to a bright yellow, the tube
was returned to the drybox. The solution was transferred to a 10 mm
NMR tube to which a sealed capillary containing triphenylphosphite
(external standard) and 500 µL of CDCl3 (for lock) had been added.
An NMR sample of the “before” solution prepared previously was made
analogously. The NMR tubes were capped with septa and removed
from the drybox. 31P NMR spectra were then acquired at 121 MHz
on a Bruker MSL-300 spectrometer.
Time-Course Experiments. (a) Example: Pd(PPh3)4 as Catalyst.
A specially designed reaction flask with an isolable sampling port was
flushed with argon and charged with 243 mg of bromobenzene-d5 (1.50
mmol), 285 mg of 2-(4′-methoxyphenyl)-1,4,3-dioxaborolane (1.60
mmol), 1.322 g of potassium carbonate sesquihydrate (8.0 mmol), and
18.0 mg of 4,4′-dimethylbiphenyl (0.1 mmol). Water (20 mL) and
THF (10 mL) were added, and the mixture was degassed with two
freeze-pump-thaw cycles. A solution of Pd(PPh3)4 in degassed THF
(1.73 mg/mL; 10 mL gives 1.50 × 10-5 mol, 1% vs bromobenzene-
d5) was introduced, and the mixture was subjected to an additional two
freeze-pump-thaw cycles. The flask was immersed in a glycerol bath
held at 50 °C and allowed to stir. Aliquots (200 µL) were removed
periodically and immediately quenched into prepared Ar-flushed,
septum-capped vials containing 200 µL of dilute aqueous KCN (ca.
50 mM). Ether (200 µL) was added to the vials, and the aqueous phase
was removed via syringe. The organic layer was rinsed twice with
200 µL portions of degassed water, then stored on ice or at -20 °C
prior to GC-MS analysis. The data obtained via GC-MS analysis
were converted to concentrations (in the case of 4-methoxybiphenyl-
d0 and -d5) to monitor conversion or used as is (in the case of
triphenylphosphine mixtures) to assess the degree of aryl interchange
as portrayed in Figures 4 and 5. The example described here
corresponds to entry 3 in Table 4 as well. Other runs were conducted
in an analogous manner by modifying conditions as specified in Table
4.
(d) Kinetics of Phosphonium Salt Formation. (1) Phosphine
Profile. In a drybox, a THF-d8 stock solution 0.200 M in LiI, 0.500
M in triphenylphosphine-d15, and 6 µM in trioxane in was prepared in
a 5.00 mL volumetric flask (stock solution “A”). A second THF-d8
stock solution that was 0.200 M in LiI, 0.100 M in triphenylphosphine-
d15, and 6 µM in trioxane was similarly prepared (stock solution “B”).
1i (21.2 mg, 0.0244 mmol) was then weighed out into four separate
vials. Into the first of these was syringed 0.900 mL of stock solution
“A” and 0.300 mL of stock solution “B”, resulting in 1.200 mL of a
kinetics solution that was 0.200 M in LiI, 0.400 M in phosphine, and
approximately 0.020 M in 1i. Kinetics solutions (all 0.200 M in LiI
and 0.020 M in 1i) that were 0.300, 0.200, and 0.100 M in phosphine
were analogously prepared. Samples were prepared and stored at -40
°C as described above. Additional samples that were 0.040 and 0.020
M in phosphine were similarly prepared from two stock solutions, the
first 0.0400 M in phosphine and 0.200 M in LiI, the second 0.200 M
in LiI with no phosphine. Kinetics were observed via 1H NMR at 200
MHz on a Bruker AC-200 spectrometer at 60.0 ( 0.5 °C analogous to
the protocol described above. The phosphonium salt formation was
monitored by the disappearance of palladium-bound aryl signals of 1i
integrated relative to the trioxane internal standard. (2) Iodide Profile.
Kinetics solutions (all 0.400 M in phosphine and approximately 0.020
M in 1i) that were 0.160 M in LiI, 0.040 M in LiPF6; 0.120 M in LiI,
0.080 M in LiPF6; and 0.080 M in LiI, 0.120 M in LiPF6 were prepared
from stock solution “C” (0.200 M in LiI, 0.400 M in triphenylphos-
phine-d15, 6 µM in trioxane) and stock solution “D” which was 0.040
M in LiI, 0.160 M in LiPF6, 0.400 M in triphenylphosphine-d15, and 6
µM in trioxane. An additional solution (2.0 mL) that was 0.200 M in
LiPF6, 0.400 M in triphenylphosphine-d15, and 6 µM in trioxane was
also prepared for the “no iodide” runs. NMR samples were then
prepared and kinetics monitored as described above.
(e) Substituent Effect Experiment. In a drybox, 24.7 µmol samples
of complexes 1a-x were weighed into separate vials. Into each of
these was added 1.20 mL of dry, degassed CDCl3 prechilled to -40
°C, resulting in stock solutions of approximately 0.0201 M in each
complex. Samples were then prepared and stored at -40 °C as
described above. Eight scan spectra were then taken at -10 °C prior
to the experiment, after submerging the samples in a 25 °C oil bath for
1.5 h, after submerging an additional 6.5 h, after heating the samples
in a 50 °C bath for 1 h, and after submerging in the 50 °C bath for an
additional 2 h. During the time intervals between acquisition of the
spectra and submersion in the oil baths, the samples were chilled in
ice. The scans were taken on a 90° pulse with a relaxation delay of 2
min between scans to ensure accurate relative intensities. For all
complexes except 1g, the extent of transfer was determined by the loss
of signal due to the palladium-bound aryl group for 1 relative to the
total signal due to the palladium-bound aryl groups for both 1 and 2.
The resonances for 1g and 2g overlapped, preventing the determination
of percent exchange for this complex.
(b) Control Experiment: Recovery of Triphenylphosphine from
(1l). An argon-flushed round-bottomed flask was charged with 53.6
mg of 1l and 10 mL degassed DME. KCN (36 mg, 5.5 × 10 -4 mol,
ca. 10 equiv vs Pd) dissolved in 10 mL of degassed water was added
to the DME solution, and the mixture was stirred for 5 min. The
mixture was extracted into 50 mL of ether; the ethereal layer was
washed with 3 × 25 mL of water, then concentrated under vacuum.
The residue was sublimed to yield 26.2 mg of triphenylphosphine (1.0
1
× 10-4 mol, 90% recovery) with H and 31P NMR spectra indistin-
guishable from a commercial sample.
Screening of Possible Strategies for the Elimination of the Aryl-
Aryl Interchange Reaction in Palladium-Mediated Cross-Couplings.
(a) “Ligandless” Catalyst Systems. A thick-walled glass tube
equipped with a sidearm and Teflon stopcock was charged with 1 mmol
of aryl halide, 0.106 g of phenylboric anhydride (0.35 mmol, 1.05 equiv
as phenylboric acid), 0.450 g of potassium carbonate sesquihydrate (2.72
mmol), 2.5 mL of water, 2.0 mL of organic solvent, and a stirbar. The
tube was sealed, and the contents were degassed via several freeze-
pump-thaw cycles. Under an argon backflow, the stopcock was
replaced with a septum and 0.5 mL of a 4.0 mM catalyst stock solution
(0.002 mmol) for the 0.2% catalyst entries, or 0.5 mL of a 20.0 mM
catalyst stock solution (0.01 mmol) for the 1.0% catalyst entries was
added via syringe. The tube was then resealed, degassed once more,
backfilled with argon, and heated in a 65 °C oil bath for the specified
amount of time. For the 4-iodotoluene entries, the presence or absence
of residual aryl halide was determined by thin layer chromatography
(petroleum ether) of the final reaction solutions; the reactions were
Reaction of 1y with Tetraphenylphosphonium Iodide. In a
drybox, 186.5 mg (0.400 mmol) of tetraphenylphosphonium iodide and
1 mg of trioxane were weighed into a 2.00 mL volumetric flask which
was then diluted to the mark with CDCl3. 1y (21.3 mg, 0.0245 mmol)
was then weighed into a separate vial and then dissolved in 1.2 mL of
the above solution, yielding a sample solution that was 0.2 M in
tetraphenylphosphonium iodide and approximately 0.02 M in 1y. NMR
samples were prepared and stored at -40 °C as described above. The
1
reaction was monitored at 50.0 ( 0.5 °C via H NMR at 200 MHz
with a Bruker AC-200 spectrometer. Four 90° pulse scans were taken

Aryl-Aryl Interchange Reaction of ArPdL2I
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12453
not worked up. For the other entries, the crude reactions were extracted
into ether (4 × 10 mL), washed with brine, and dried over MgSO4.
The solvent was then removed with a rotary evaporator, and the
of choice in 10.0, 2.0, and 2.0 mL volumetric flasks, respectively. These
solutions were then transferred to glass tubes equipped with sidearms
and Teflon stopcocks. A separate, similarly equipped tube (the 5.0%
catalyst entry) was then charged with 24.8 mg of tris(dibenzylidene-
acetone)dipalladium(0) (0.0500 mmol of Pd), 0.100 mmol of phosphine,
and a stirbar. The tubes were then sealed, removed from the drybox,
and fitted to dual-manifold vacuum lines, along with four additional
tubes (the 0.2-2.0% entries) which had been equipped with stirbars,
evacuated, and backfilled with argon. On the benchtop, stock solutions
of potassium carbonate sesquihydrate (1.089 M, 50 mL, in HPLC grade
water) and 4-iodotoluene (1.176 M, 10 mL) were similarly prepared
and degassed via three freeze-pump-thaw cycles. The stopcocks were
all replaced with septa and to each of the five reaction tubes were added
0.850 mL (1.00 mmol) of the 4-iodotoluene stock, 0.850 mL (1.10
equiv) of the phenylboric anhydride stock, and 2.50 mL (2.72 mmol)
of the K2CO3 solution, all via syringe. The reaction tubes were then
similarly charged with the following amounts of phosphine stock,
catalyst stock, and additional degassed organic solvent: 0.2% catalyst
entry with 40.0 µL (0.004 mmol) of phosphine stock, 40.0 µL (0.002
mmol) of palladium stock, and 0.720 mL of additional solvent; 0.5%
catalyst entry with 100.0 µL (0.010 mmol) of phosphine stock, 100.0
µL (0.005 mmol) of palladium stock, and 0.600 mL of additional
solvent; 1.0% catalyst entry with 200.0 µL (0.020 mmol) of phosphine
stock, 100.0 µL (0.010 mmol) of palladium stock, and 0.400 mL of
additional solvent; 2.0% catalyst entry with 400.0 µL (0.040 mmol) of
phosphine stock, 400.0 µL (0.020 mmol) of palladium stock, and no
additional solvent; and the 5.0% catalyst entry with 0.800 mL of
additional solvent. At this point, each tube contained 2.5 mL of water
and 2.5 mL of total organic solvent. The reaction tubes were then all
resealed, degassed once more, backfilled with argon, and heated in a
65 °C bath for 24 h.
1
resulting product mixture was analyzed by H NMR.
(b) Ligand Screening, 0.2% Catalyst Systems. A thick-walled
glass tube equipped with a sidearm and Teflon stopcock was charged
with 218.0 mg of 4-iodotoluene (1.00 mmol), 0.106 g of phenylboric
anhydride (0.350 mmol, 1.05 equiv as phenylboric acid), 0.450 g of
potassium carbonate sesquihydrate (2.72 mmol), 2.5 mL of water, 2.0
mL of THF, and a stirbar. The tube was sealed, and the contents were
degassed via several freeze-pump-thaw cycles. Under an argon
backflow, the stopcock was replaced with a septum and 0.250 mL of
a 16.0 mM ligand stock solution (0.004 mmol) and 0.250 mL of a 4.0
mM solution of the benzene adduct of tris(dibenzylideneacetone)-
dipalladium(0) (0.001 mmol, 0.002 equiv of Pd) was added via syringe.
The tube was then resealed, degassed once more, backfilled with argon,
and heated in a 65 °C oil bath for 12 h. The presence or absence of
residual 4-iodotoluene was then determined by thin layer chromatog-
raphy (petroleum ether). For the 4-iodoanisole entry, the crude reaction
was worked up as above, and the resulting product mixture was
1
analyzed by H NMR.
(c) Ligand Screening, 5.0% Catalyst Systems. A thick-walled
glass tube equipped with a sidearm and Teflon stopcock was charged
with 218.0 mg of 4-iodotoluene (1.00 mmol), 0.106 g of phenylboric
anhydride (0.350 mmol, 1.05 equiv as phenylboric acid), 0.450 g of
potassium carbonate sesquihydrate (2.72 mmol), 0.100 mmol of ligand,
0.05 mmol of palladium species, and a stirbar. The tube was evacuated
and backfilled with argon three times, and the stopcock was replaced
with a septum. Under an argon backflow, 2.5 mL of degassed THF
and 2.5 mL of degassed water were added via syringe. The tube was
then resealed, degassed once more, backfilled with argon, and heated
in a 65 °C oil bath for 12 h. The two phases were then separated via
pipette, and the aqueous phase was washed with 3 × 5 mL of ether.
The organic rinsings were combined and shaken vigorously with 10
mL of 10% KCN in a 50 mL separation funnel until the color faded to
a pale yellow. The aqueous layer was carefully separated, and the
organic phase was washed with 1 × 10 mL of water and then with 1
× 10 mL of brine. The resulting solution was then dried over MgSO4,
and the solvent was removed with a rotary evaporator. For the tris-
(o-tolyl)phosphine entry, the residue was dissolved in CDCl3 and
analyzed by 31P {1H} NMR (121 MHz, CDCl3). Only signals for
residual phosphine (δ ) -29.9, authentic sample δ ) -30.2) and
phosphine oxide (δ ) 37.1, authentic sample61 δ ) 36.3) were present.
For all the entries, the crude residue was distilled at 100 °C (oven
temperature) in a bulb-to-bulb apparatus under a dynamic 5 mTorr
vacuum, yielding a clear, solid distillate (150-180 mg), which was
then dissolved in 10 mL of ether. This solution (25 µL) was further
diluted to 10 mL with ether, and this final solution was analyzed by
GC-MS to determine the presence of residual 4-iodotoluene and
products resulting from the aryl-aryl exchange reaction. The relative
amounts of products formed were determined from integrations of the
exact ion chromatograms obtained (m/z ) 184, 168) from these
solutions normalized to predetermined response factors.
Each reaction was then worked up as follows. The two phases were
separated via pipette, and the aqueous phase was washed with 3 × 5
mL of methylene chloride. The organic washings were combined, dried
over molecular sieves, and the solvent was removed with a rotary
evaporator. The crude residue was then distilled and analyzed as
described above.
Acknowledgment. We gratefully thank the National Science
Foundation (U.S.), the Office of Naval Research, and the
Materials Research Science & Engineering Center at the
University of Massachusetts for funding this research. T.I.W.
thanks the U.S. Department of Education, Dow Chemical, and
IBM for fellowship support. Similarly, F.E.G. thanks the U.S.
Department of Education for a graduate fellowship. In addition,
we acknowledge Professor Thomas A. P. Seery and Dr. Michael
L. Greenfield for their assistance in interpreting the kinetics data,
and we thank Professor Donald J. Darensbourg for the sample
of 1,2,5-triazaphosphaadamantane.
Supporting Information Available: Linearized first-order
decay plots for aryl-aryl exchange in the presence and absence
of salt and excess phosphine ligands (3 pages). See any current
masthead page for ordering and Internet access instructions.
(d) Optimization of Cross-Coupling Protocol. In a drybox, stock
solutions of phenylboronic anhydride (1.294 M in phenylboric acid),
phosphine (0.100 M), and the benzene adduct of tris(dibenzylidene-
acetone)dipalladium(0) (0.0500 M in Pd) were prepared in the solvent
JA972554G