Toward an Improved Understanding of the Unusual Reactivity of Pd 0/Trialkylphosphane Catalysts in Cross-Couplings of Alkyl Electrophiles: Quantifying the Factors That Determine the Rate of Oxidative Addition

Article abstract of DOI:10.1002/anie.200352858

What a difference a ligand makes! The reactivity of Pd ⁰/trialkylphosphane complexes is of great interest because of their unusual effectiveness as cross-coupling catalysts for alkyl electrophiles [Eq. (1)]. This study sheds light on various aspects of the critical oxidative-addition step of the catalytic cycle.

Full text of DOI:10.1002/anie.200352858

Angewandte
Chemie
Reaction Mechanisms
Toward an Improved Understanding of the
Unusual Reactivity of Pd0/Trialkylphosphane
Catalysts in Cross-Couplings of Alkyl
Electrophiles: Quantifying the Factors That
Determine the Rate of Oxidative Addition**
Ivory D. Hills, Matthew R. Netherton, and
Gregory C. Fu*
Palladium- and nickel-catalyzed cross-couplings are widely
used tools in synthetic chemistry.[1] No doubt, such processes
would be even more broadly employed if they were not
largely restricted to reactions of aryl and vinyl halides and
sulfonates. Recognizing the significance of this limitation,
several groups have recently turned their attention to the
development of methods for cross-coupling alkyl halides and
sulfonates.[2–7] Among the palladium catalysts that have been
described to date, the majority are based on sterically
demanding trialkylphosphanes (PCy3 or P(tBu)2Me).[8]
The catalytic cycle for most cross-coupling reactions
presumably includes a sequence of oxidative addition, trans-
metalation, and reductive elimination (Figure 1). With regard
to the challenge of coupling alkyl electrophiles, the difficulty
Figure 1. Generalized mechanism for a palladium-catalyzed cross-
coupling reaction.
[*] Prof. Dr. G. C. Fu, I. D. Hills, Dr. M. R. Netherton
Department of Chemistry
Massachusetts Institute of Technology
Cambridge, MA 02139 (USA)
Fax: (+1)617-324-3611
E-mail: gcf@mit.edu
[**] Support has been provided by the National Institutes of Health
(National Institute of General Medical Sciences, R01-GM62871 to
G.C.F. and R01-GM57034 supplement to I.D.H.), the American
Chemical Society (Organic Division Fellowship to I.D.H., sponsored
by Abbott Laboratories), the Natural Sciences and Engineering
Research Council of Canada (Postdoctoral Fellowship to M.R.N.),
Merck, and Novartis. Funding for the MIT Department of Chemistry
Instrumentation Facility has been furnished in part by NSF CHE-
9808061 and NSF DBI-9729592. We thank Johnson Matthey for
supplying palladium compounds.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2003, 42, 5749 –5752
DOI: 10.1002/anie.200352858
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5749

Communications
Table 1: Correlation between solvent polarity and the activation barrier
has generally been attributed to relatively slow oxidative
for oxidative addition.[a]
addition and to facile b-hydride elimination (Figure 1). To
complement our methods-development studies of Pd/trial-
kylphosphane catalysts for coupling alkyl electrophiles, we
have recently initiated mechanistic work on these systems. As
part of this new program, in this communication we describe a
wide-ranging investigation of the initial step of the catalytic
cycle, oxidative addition of an alkyl electrophile to a Pd/
trialkylphosphane complex.
Entry
Solvent
Polarity
DG°
[kcalmolÀ1
]
1
2
3
4
5
6
hexane
toluene
THF
0.68
1.66
2.08
2.46[c]
2.62
2.80
>23.0[b]
20.0
19.5
18.1
18.0
In an earlier study, we discovered that PdL2 (L = PCy3 or
À
P(tBu)2Me) can react with an alkyl electrophile (R X) to
tert-amyl alcohol
generate an oxidative-addition adduct that, in the absence of
a cross-coupling partner, can be isolated and even crystallo-
graphically characterized.[9] The relative stability of PdL2RX
toward b-hydride elimination was a pleasant surprise for us.
Recognizing that there are no systematic, quantitative inves-
tigations of oxidative addition of alkyl electrophiles to
palladium(o),[10] we decided to remedy this deficiency.
NMP[d]
DMF[e]
17.8
[a] All data are the average of two runs. [b] No reaction at 0–608C. DG°
was calculated for 608C. [c] Value for tert-butanol. [d] NMP=N-methyl-
pyrrolidinone. [d] DMF=dimethylformamide.
Table 2: Effect of the leaving group on the rate of oxidative addition.[a]
Through kinetics studies, we have determined that, for the
oxidative addition of n-nonyl bromide to Pd(P(tBu)2Me)2
[Eq. (1)],[11] the activation parameters are: DG° = 20.8 kcal-
molÀ1 at 208C; DH° = 2.4 kcalmolÀ1; DS° = À63 eu. The large
Entry
X
t1/2
1
2
3
4
5
I
Br
Cl
2.2 h at À608C
2.3 h at 08C
2.0 d at 608C
<2% reaction after 43 h at 608C
10.4 h at 408C
F
OTs[b]
negative DS° is consistent with the associative SN2-type
pathway suggested by our earlier stereochemical investiga-
tions.[12]
[a] All data are the average of two runs. [b] Ts=toluenesulfonyl.
The addition of P(tBu)2Me (e.g., 2 or 4 equiv) to the
reaction illustrated in Equation (1) does not affect the rate,
suggesting that the alkyl bromide oxidatively adds to PdL2,
not PdL or PdL3.[13] In contrast, in the case of cross-couplings
of aryl bromides that are catalyzed by Pd/P(tBu)3, the
available data indicate that a palladium–monophosphane
adduct undergoes oxidative addition.[14]
We have examined the relationship between solvent
polarity and the rate of oxidative addition. As illustrated in
Table 1, we have determined that an increase in polarity[15]
leads to a decrease in the activation energy, consistent with
unactivated secondary alkyl electrophiles. Naturally, we were
interested in whether this shortcoming may be attributable to
a reluctance of these compounds to undergo oxidative
addition. We therefore examined the relationship between
the rate of oxidative addition to Pd(P(tBu)2Me)2 and the
steric demand of the electrophile.
As illustrated in Table 3, the bulk of the electrophile has a
large impact on the reaction rate. Branching in the g-position
leads to a roughly fivefold decrease in reactivity relative to an
unbranched substrate (entry 2 vs. entry 1), and branching in
the b-position attenuates the reactivity by a factor of ~ 20
(entry 3 vs. entry 1). Consistent with the failure of
Pd(P(tBu)2Me)2 to catalyze cross-couplings of secondary
alkyl electrophiles, no oxidative addition occurs if the
branch is in the a-position (entry 4).
the expectation for an SN2-type nucleophilic attack of PdL2 on
[16]
À
R X.
We have also investigated the impact of the leaving group
À
X on the rate of oxidative addition of alkyl electrophiles R X
to Pd(P(tBu)2Me)2 (Table 2). We have established that an
alkyl iodide rapidly oxidatively adds to Pd(P(tBu)2Me)2 even
at À608C (entry 1), whereas the corresponding bromide
requires a temperature of 08C to achieve a comparable rate
(entry 2). Upon heating to 608C, an alkyl chloride (entry 3),
but not a fluoride (entry 4), reacts with Pd(P(tBu)2Me)2.
Finally, we have determined that an alkyl tosylate (entry 5) is
intermediate between a bromide and a chloride in reactivity.
The relative propensity of alkyl electrophiles to oxidatively
add to Pd(P(tBu)2Me)2 (Table 2) is consistent with an SN2
pathway for oxidative addition.[17]
For couplings of primary alkyl electrophiles catalyzed by
Pd/PR3, in earlier work we observed a remarkable depen-
dence of the efficiency of the reaction on the structure of the
phosphane.[3d,e] We have now determined that cross-coupling
efficiency correlates with the propensity of PdL2 to undergo
oxidative addition (Table 4). Thus, adducts of P(tBu)2Me or
PCy3 readily add n-nonyl bromide (Table 4, entries 1 and 2),
in keeping with their effectiveness in a range of coupling
processes. In contrast, complexes of P(tBu)2Et or P(tBu)3 are
relatively unreactive toward oxidative addition (entries 3 and
4), as well as ineffective at cross-coupling alkyl electrophiles.
The divergent reactivity of PdL2 complexes of P(tBu)2Me
(cone angle: 1618) and P(tBu)2Et (cone angle: 1658),[18] which
Unfortunately, no catalyst, including Pd(P(tBu)2Me)2, has
yet been reported to be effective for cross-couplings of
5750
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5749 –5752

Angewandte
Chemie
Table 3: Correlation between the steric demand of the electrophile and
the activation barrier for oxidative addition.[a]
Entry
krel
DG°
À
R Br
[kcalmolÀ1
]
1
2
1.0
19.5
20.3
0.19
3
4
0.054
21.0
<0.0001
>24.0[b]
[a] All data are the average of two runs. [b] Extrapolated from a reaction
run at 608C.
Figure 2. A rationale for the divergent reactivity of Pd(P(tBu)2Me)2 and
Pd(P(tBu)2Et)2.
Table 4: Effect of the phosphane on the activation barrier for oxidative
addition.[a]
have quantified the impact of the following parameters on the
À
rate of oxidative addition of R X to PdL2: the solvent, the
leaving group (X = I, Br, Cl, F, OTs), and the steric demand of
the electrophile (R) and the ligand (L). Furthermore, based
on DFT calculations, we have provided a rationale for the
remarkable difference in reactivity between Pd(P(tBu)2Me)2
and Pd(P(tBu)2Et)2. Additional mechanistic studies of these
intriguing cross-coupling catalysts are underway.
Entry
L
DG° [kcalmolÀ1 [b]
]
1
2
3
4
P(tBu)2Me
PCy3
P(tBu)2Et
P(tBu)3
19.5 (08C)
20.0 (08C)
25.4 (608C)
>28.4 (608C)
[a] All data are the average of two runs. [b] The temperature at which DG°
was measured is noted in parentheses.
Received: September 12, 2003 [Z52858]
Keywords: cross-coupling · homogeneous catalysis · palladium ·
.
phosphane ligands · reaction mechanisms
at first glance appear to be nearly isosteric, is particularly
striking (Table 4, entries 1 and 3; krel > 103). However, com-
putation of the energy of the various conformers of palla-
dium–phosphane adducts leads to a reasonable explanation
for this initially surprising difference. Through B3LYP/
LanL2DZ DFT calculations, we have determined that con-
former 1, in which Me and Pd are eclipsed, is ~ 3.5 kcalmolÀ1
lower in energy than conformer 2, in which Me and Pd are
anti, due to the presence of additional syn-pentane-like
interactions in 2 (Figure 2a).[19] Figure 2b shows that, in the
illustrated conformation of Pd(P(tBu)2Et)2 (3), approach of
an electrophile to Pd would be impeded significantly by the
Me groups. In contrast, higher energy conformation 4
presents a steric environment around Pd that resembles that
of Pd(P(tBu)2Me)2.[20] We believe that this analysis provides
an adequate rationalization for the difference in reactivity
between Pd(P(tBu)2Et)2 and Pd(P(tBu)2Me)2.
In conclusion, we have described the first systematic
investigation of the oxidative addition of alkyl electrophiles to
Pd0, specifically, adducts of trialkylphosphanes. The reactivity
of these complexes is of high current interest due to their
unusual effectiveness as cross-coupling catalysts for alkyl
electrophiles. We have determined that the species that
undergoes oxidative addition is PdL2 and that the activation
parameters for this SN2 process are: DG° = 20.8 kcalmolÀ1 at
208C; DH° = 2.4 kcalmolÀ1; DS° = À63 eu. In addition, we
[1] For reviews of metal-catalyzed cross-coupling reactions, see:
a) Metal-Catalyzed Cross-coupling Reactions (Eds.: F. Dieder-
ich, P. J. Stang), Wiley-VCH, New York, 1998; b) “Cross-
Coupling Reactions: A Practical Guide”: Top. Curr. Chem.
2002, 219 (Ed.: N. Miyaura); c) Handbook of Organopalladium
Chemistry for Organic Synthesis (Ed.: E.-i. Negishi), Wiley
Interscience, New York, 2002.
[2] For an overview of the difficulty of achieving coupling reactions
3
À
of C(sp ) X electrophiles, see: D. J. Cµrdenas, Angew. Chem.
1999, 111, 3201 – 3203; Angew. Chem. Int. Ed. 1999, 38, 3018 –
3020; D. J. Cµrdenas, Angew. Chem. 2003, 115, 398 – 401; Angew.
Chem. Int. Ed. 2003, 42, 384 – 387. See also: T.-Y. Luh, M.-k.
Leung, K.-T. Wong, Chem. Rev. 2000, 100, 3187 – 3204.
[3] Suzuki reaction: a) T. Ishiyama, S. Abe, N. Miyaura, A. Suzuki,
Chem. Lett. 1992, 691 – 694; b) M. R. Netherton, C. Dai, K.
Neuschütz, G. C. Fu, J. Am. Chem. Soc. 2001, 123, 10099 –
10100; c) J. H. Kirchhoff, C. Dai, G. C. Fu, Angew. Chem.
2002, 114, 2025 – 2027; Angew. Chem. Int. Ed. 2002, 41, 1945 –
1947; d) M. R. Netherton, G. C. Fu, Angew. Chem. 2002, 114,
4066 – 4068; Angew. Chem. Int. Ed. 2002, 41, 3910 – 3912; e) J. H.
Kirchhoff, M. R. Netherton, I. D. Hills, G. C. Fu, J. Am. Chem.
Soc. 2002, 124, 13662 – 13663.
[4] Negishi reaction: a) A. Devasagayaraj, T. Stüdemann, P. Kno-
chel, Angew. Chem. 1995, 107, 2952 – 2954; Angew. Chem. Int.
Ed. Engl. 1995, 34, 2723 – 2725; b) R. Giovannini, T. Stüdemann,
G. Dussin, P. Knochel, Angew. Chem. 1998, 110, 2512 – 2515;
Angew. Chem. Int. Ed. 2003, 42, 5749 –5752
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5751

Communications
Angew. Chem. Int. Ed. 1998, 37, 2387 – 2390; c) R. Giovannini, P.
Knochel, J. Am. Chem. Soc. 1998, 120, 11186 – 11187; d) R.
Giovannini, T. Stüdemann, A. Devasagayaraj, G. Dussin, P.
Knochel, J. Org. Chem. 1999, 64, 3544 – 3553; e) M. Piber, A. E.
Jensen, M. Rottländer, P. Knochel, Org. Lett. 1999, 1, 1323 –
1326; f) A. E. Jensen, P. Knochel, J. Org. Chem. 2002, 67, 79 – 85;
g) J. Zhou, G. C. Fu, J. Am. Chem. Soc. 2003, 125, 12527 – 12530.
[5] Kumada reaction: a) J. Terao, H. Watanabe, A. Ikumi, H.
Kuniyasu, N. Kambe, J. Am. Chem. Soc. 2002, 124, 4222 – 4223;
b) A. C. Frisch, N. Shaikh, A. Zapf, M. Beller, Angew. Chem.
2002, 114, 4218 – 4221; Angew. Chem. Int. Ed. 2002, 41, 4056 –
4059; c) J. Terao, A. Ikumi, H. Kuniyasu, N. Kambe, J. Am.
Chem. Soc. 2003, 125, 5646 – 5647.
[6] Stille reaction: a) K. Menzel, G. C. Fu, J. Am. Chem. Soc. 2003,
125, 3718 – 3719; b) H. Tang, K. Menzel, G. C. Fu, Angew. Chem.
2003, 115, 5233 – 5236; Angew. Chem. Int. Ed. 2003, 42, 5079 –
5082.
[7] Hiyama reaction: J.-Y. Lee, G. C. Fu, J. Am. Chem. Soc. 2003,
125, 5616 – 5617.
[8] For exceptions, see refs. [3a] and [6b].
[9] a) Ref. [3e]; b) I. D. Hills, M. R. Netherton, unpublished results.
[10] For an excellent review of oxidative addition, including the
chemistry of Pd0, see: J. K. Stille in The Chemistry of the Metal–
Carbon Bond (Ed.: S. Patai), Wiley, New York, 1985, chap. 9. See
also: M. Huser, M.-T. Youinou, J. A. Osborn, Angew. Chem.
1989, 101, 1427 – 1430; Angew. Chem. Int. Ed. Engl. 1989, 28,
1386 – 1388.
[11] We chose to focus on complexes of P(tBu)2Me because it
generally furnishes the most active catalyst.
[12] a) Ref. [3d]; b) I. D. Hills, M. R. Netherton, unpublished results.
[13] When Pd(P(tBu)2Me)2 is dissolved in THF, the only phosphorus-
containing species that is present according to 31P NMR
spectroscopy is the PdL2 adduct (e.g., no PdL1), and there is
no change in the spectrum in the presence of excess P(tBu)2Me
(e.g., no PdL3).
[14] a) A. F. Littke, G. C. Fu, Angew. Chem. 1998, 110, 3586 – 3587;
Angew. Chem. Int. Ed. 1998, 37, 3387 – 3388; A. F. Littke, C. Dai,
G. C. Fu, J. Am. Chem. Soc. 2000, 122, 4020 – 4028; b) J. P.
Stambuli, R. Kuwano, J. F. Hartwig, Angew. Chem. 2002, 114,
4940 – 4942; Angew. Chem. Int. Ed. 2002, 41, 4746 – 4748, and
references therein; M. Prashad, X. Y. Mak, Y. Liu, O. Repic, J.
Org. Chem. 2003, 68, 1163 – 1164. See also: E. Galardon, S.
Ramdeehul, J. M. Brown, A. Cowley, K. K. Hii, A. Jutand,
Angew. Chem. 2002, 114, 1838 – 1841; Angew. Chem. Int. Ed.
2002, 41, 1760 – 1763.
[15] The values for solvent polarity are according to the S’ scale. For
an overview, see: J. Zou, Q. Yu, Z. Shang, J. Chem. Soc. Perkin
Trans. 2 2001, 1439 – 1443.
[16] In our methods-development studies, we have consistently found
nonpolar solvents such as hexane and toluene to be unsuitable
for cross-coupling reactions of alkyl electrophiles.
[17] Consistent with the possibility that oxidative addition is the
turnover-limiting step of the catalytic cycle, the same order of
reactivity (Br> OTs> Cl) is observed for Suzuki reactions
catalyzed by Pd/PR3.[3b–d]
[18] These values were calculated by the method of Tolman: C. A.
Tolman, Chem. Rev. 1977, 77, 313 – 348.
[19] Notes: a) The only other energy minimum that was located is
the conformation in which Me and Pd are gauche. This lies
~ 0.6 kcalmolÀ1 higher in energy than the eclipsed conformation.
b) These calculations are, of course, imperfect, since they do not
account for effects such as solvation.
[20] The PdL2 structures in Figure 2b are not the direct result of
calculations—they are derived from computed structures 1 and
2.
5752
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5749 –5752