Intramolecular insertion of alkenes into Pd-N bonds. Effects of substrate and ligand structure on the reactivity of (P-P)Pd(Ar)[N(Ar1)(CH 2)3CR=CHR′] complexes

Article abstract of DOI:10.1021/om200008t

Studies on the synthesis and reactivity of a series of (P-P)Pd(Ar) [N(Ar¹)(CH₂)₃CR=CHR′] complexes 3 are described. These complexes are transformed to observable (P-P)Pd(Ar)[pyrrolidin- 2-ylmethyl] complexes 4 via syn insertion of the pendant alkene into the Pd-N bond. Complexes 4 then undergo C-C bond-forming reductive elimination to yield N-aryl-2-benzylpyrrolidine derivatives 2. Kinetic studies indicate the rates of conversion of 3 to 4 and 4 to 2 are within 1 order of magnitude. The effects of phosphine ligand structure, alkene substitution, and the electronic properties of the Ar and Ar¹ groups on reaction rates are reported, as are the results of deuterium isotope effect studies. The mechanism of the aminopalladation step is discussed in detail, and the results of the experiments described in this paper are most consistent with conversion of 3 to 4 via rate-determining ligand displacement followed by fast aminopalladation. These transformations represent rare examples of syn migratory insertion of unactivated alkenes into Pd-N bonds.

Full text of DOI:10.1021/om200008t

Organometallics 2011, 30, 1269–1277 1269
DOI: 10.1021/om200008t
Intramolecular Insertion of Alkenes into Pd-N Bonds. Effects of
Substrate and Ligand Structure on the Reactivity of
(P-P)Pd(Ar)[N(Ar1)(CH2)3CRdCHR0] Complexes
Joshua D. Neukom, Nicholas S. Perch, and John P. Wolfe*
Department of Chemistry, University of Michigan, 930 N. University Avenue, Ann Arbor, Michigan
48109-1055
Received January 6, 2011
Studies on the synthesis and reactivity of a series of (P-P)Pd(Ar)[N(Ar1)(CH2)3CRdCHR0]
complexes 3 are described. These complexes are transformed to observable (P-P)Pd(Ar)[pyrrolidin-
2-ylmethyl] complexes 4 via syn insertion of the pendant alkene into the Pd-N bond. Complexes 4
then undergo C-C bond-forming reductive elimination to yield N-aryl-2-benzylpyrrolidine deriva-
tives 2. Kinetic studies indicate the rates of conversion of 3 to 4 and 4 to 2 are within 1 order of
magnitude. The effects of phosphine ligand structure, alkene substitution, and the electronic
properties of the Ar and Ar1 groups on reaction rates are reported, as are the results of deuterium
isotope effect studies. The mechanism of the aminopalladation step is discussed in detail, and the
results of the experiments described in this paper are most consistent with conversion of 3 to 4 via
rate-determining ligand displacement followed by fast aminopalladation. These transformations
represent rare examples of syn migratory insertion of unactivated alkenes into Pd-N bonds.
Scheme 1. Pd-Catalyzed Alkene Carboamination
Introduction
The syn insertion of alkenes into Pd-N bonds has been
implicated as a key step in many useful Pd-catalyzed reac-
tions. For example, Pd-catalyzed alkene carboaminations
between γ-aminoalkene derivatives 1 and aryl bromides are
believed to involve syn-aminopalladation of intermediate
palladium(aryl)amido complexes (e.g., 3), followed by re-
ductive elimination of the resulting (aryl)(pyrroldin-2-yl-
methyl)palladium complexes (e.g, 4) to yield substituted-
pyrrolidine products 2 (Scheme 1).1 The syn-aminopallada-
tion step leads to formation of a C-N bond and also leads to
the generation of two stereocenters, which are retained in the
pyrrolidine products. This mechanistic pathway is also be-
lieved to occur in Pd-catalyzed diaminations,2 oxidative
aminations,3 chloroaminations,4 aminoacetoxylations,5 and
hetero-Heck transformations.6,7
Despite the significance of syn-aminopalladation pro-
cesses, and the influence of this pathway on the stereochem-
ical outcome of synthetically useful reactions, documented
unambiguous examples of syn insertions of olefins into late-
transition-metal-nitrogen bonds are very rare,8 and cases
involving palladium complexes have only recently been de-
scribed by our group and Hartwig’s group.9,10 As such, little
is known about the effect of palladium amido complex struc-
ture on the facility of aminopalladation. However, information
*To whom correspondence should be addressed. E-mail: jpwolfe@
umich.edu.
(1) (a) Ney, J. E.; Wolfe, J. P. J. Am. Chem. Soc. 2005, 127, 8644.
(b) Bertrand, M. B.; Neukom, J. D.; Wolfe, J. P. J. Org. Chem. 2008, 73,
8851. (c) Ney, J. E.; Wolfe, J. P. Angew. Chem., Int. Ed. 2004, 43, 3605.
(8) (a) Villanueva, L. A.; Abboud, K. A.; Boncella, J. M. Organome-
tallics 1992, 11, 2963. (b) VanderLende, D. D.; Abboud, K. A.; Boncella,
J. M. Inorg. Chem. 1995, 34, 5319. (c) Cowan, R. L.; Trogler, W. C.
J. Am. Chem. Soc. 1989, 111, 4750. (d) Cowan, R. L.; Trogler, W. C.
Organometallics 1987, 6, 2451. (e) Casalnuovo, A. L.; Calabrese, J. C.;
Milstein, D. J. Am. Chem. Soc. 1988, 110, 6738. (f) Zhao, P.; Krug, C.;
Hartwig, J. F. J. Am. Chem. Soc. 2005, 127, 12066.
~
€
(2) (a) Muniz, K.; Hovelmann, C. H.; Streuff, J. J. Am. Chem. Soc.
2008, 130, 763. (b) Du, H.; Zhao, B.; Shi, Y. J. Am. Chem. Soc. 2007, 129,
762.
(3) (a) Liu, G.; Stahl, S. S. J. Am. Chem. Soc. 2007, 129, 6328.
(b) Brice, J. L.; Harang, J. E.; Timokhin, V. I.; Anastasi, N. R.; Stahl,
S. S. J. Am. Chem. Soc. 2005, 127, 2868.
€
(4) Helaja, J.; Gottlich, R. Chem. Commun. 2002, 720.
(9) A portion of the work reported in this article has been previously
communicated. See: Neukom, J. D.; Perch, N. S.; Wolfe, J. P. J. Am.
Chem. Soc. 2010, 132, 6276.
(5) Liu, G.; Stahl, S. S. J. Am. Chem. Soc. 2006, 128, 7179.
(6) Tsutsui, H.; Narasaka, K. Chem. Lett. 1999, 45.
(7) For reviews on metal-catalyzed reactions that involve syn-alkene
insertion into Pd-N bonds, see: (a) Wolfe, J. P. Synlett 2008, 2913.
~
(b) Minatti, A.; Muniz, K. Chem. Soc. Rev. 2007, 36, 1142. (c) Kotov, V.;
Scarborough, C. C.; Stahl, S. S. Inorg. Chem. 2007, 46, 1910.
(10) For a recent study on intermolecular alkene insertion into Pd-N
bonds of palladium amido complexes bearing a cyclometalated phos-
phine ligand, see: Hanley, P. S.; Markovic, D.; Hartwig, J. F. J. Am.
Chem. Soc. 2010, 132, 6302.
r
2011 American Chemical Society
Published on Web 02/03/2011
pubs.acs.org/Organometallics

1270 Organometallics, Vol. 30, No. 5, 2011
Neukom et al.
on the relationship between structural features and reactivity
could potentially be used to improve the efficiency of cata-
lytic processes or to guide the design of new catalysts for use
in challenging reactions or enantioselective transformations.11
In this article we describe detailed studies on the synthesis
and reactivity of the (P-P)Pd(Ar)[N(Ar1)(CH2)3CRdCHR0]
complexes 3.1,12 These complexes undergo syn migratory
insertion of the alkene into the Pd-N bond to provide
detectable (P-P)Pd(Ar)(pyrrolidin-2-ylmethyl) complexes
4, which undergo C-C bond-forming reductive elimination
to yield N-aryl-2-benzylpyrrolidine derivatives 2. The rates
of aminopalladation of 3 and reductive elimination of 4 are
influenced by several structural parameters, including the
electronic properties of the Ar and Ar1 groups, the degree of
alkene substitution, and the nature of the bis-phosphine
ligand. Our experiments suggest the alkene aminopallada-
tion occurs from a four-coordinate complex and illustrate
that ligand electronic properties can be tuned to have a
positive influence on the rates of both aminopalladation
and reductive elimination.
Scheme 2. Preliminary Studies
As such, we elected to generate the requisite amido complex
3a in situ from (dppf)Pd(p-C6H4F)(Br) (5a)15 and the potas-
sium anilide salt of N-(p-fluorophenyl)pent-4-enylamine (6a).
In our first series of experiments, a solution of 5a in THF
or THF-d8 in an NMR tube was treated with 6a (1.05 equiv)
in the presence of 2-fluorotoluene as internal standard and
dppf (2 equiv) as a trap for Pd(0) (Scheme 2). Upon mixing,
the solution underwent a rapid color change from orange to
bright red, and analysis of the mixture by 31P and 19F NMR
spectroscopy indicated the starting complex 5a had been
consumed in less than 90 s. The formation of amido complex
3a was evident by the presence of a pair of doublets at
24.9 ppm (JPP = 38.1 Hz) and 9.0 ppm (JPP = 35.5 Hz) in
the 31P NMR spectrum, which are comparable to data pre-
viously reported for (dppf)Pd(Ar)[N(Ar1)(R)] complexes.14,16
New signals at -123.7 and -137.3 ppm were also observed in
the 19F NMR spectrum of 3a.
Within 2 min amido complex 3a underwent reaction to
generate detectable amounts of a new intermediate complex
(A), which exhibited 19F NMR resonances at -124.1 and
-133.3 ppm and 31P NMR signals at 21.3 ppm (JPP = 24.1
Hz) and 16.6 ppm (JPP = 21.7 Hz). As the conversion of 3a to
A proceeded, pyrrolidine 2a and (dppf)2Pd were generated at
a rate that appeared to be roughly comparable to that of the
3a to A transformation. Overall, the conversion of 3a to 2a
proceeded in 86% NMR yield in 45 min at 24 °C. No
additional intermediates on the pathway from 3a to 2a were
detected, and no side products resulting from β-hydride
elimination were observed. Quantitative measurement of reac-
tion kinetics at 24 °C provided a data set consistent with
consecutive first-order reactions for the transformation of 3a
to A and the transformation of A to 2a (Figures 1, 2). Rate
constants were extracted for the two first-order reactions (3a
to A, k1 = (1.74 ( 0.02) ꢀ 10-3 s-1; A to 2a, k2 = (1.36 (
0.41) ꢀ 10-3 s-1),17,18 which occur with rates that are nearly
identical. Neither excess dppf nor excess 6a had an effect on
k1 or k2.
Results
Preliminary Studies on the Synthesis and Reactivity of
(dppf)Pd(p-C6H4F)[N(p-C6H4F)(CH2)3CHdCH2]. Exami-
nation of Reaction Rates and Identification of Key Intermedi-
ates. In our initial experiments on the intramolecular syn-
aminopalladation of (aryl)(amido)palladium complexes we
elected to examine the generation and reactivity of (dppf)-
Pd(p-C6H4F)[N(p-C6H4F)(CH2)3CHdCH2] (3a). The bis-
phosphine dppf was selected as the ligand for our preliminary
studies, as it provides acceptable yields of N-aryl-2-benzyl-
pyrrolidine products in catalytic reactions.1c,13 In addition,
(dppf)Pd(aryl)(amido) complexes have been described in the
literature,14 and the previously reported NMR data could
aid in the structural assignment of our complexes. In order to
allow for measurement of reaction rates by 19F NMR, we chose
to initially study (aryl)(amido)palladium complexes derived
from 1-bromo-4-fluorobenzene and N-(p-fluorophenyl)pent-4-
enylamine 1a.
Prior studies on the synthesis of LnPd(Ar)(NRR0) com-
plexes suggested that the high reactivity of these species would
preclude their isolation in most cases.14 (Aryl)(amido)pallad-
ium complexes are known to undergo relatively facile C-N
bond-forming reductive elimination to yield N-arylated
amine products, and complexes bearing β-hydrogen atoms
can also undergo competing β-hydride elimination. We antici-
pated the intramolecular alkene aminopalladation of com-
plexes 3 would occur even more rapidly than reductive elim-
ination or β-hydride elimination, as only small amounts of
side products resulting from these competing pathways were
observed in Pd-catalyzed alkene carboamination reactions.1c
On the basis of our prior studies on Pd-catalyzed alkene
carboamination reactions, the most likely candidates for inter-
mediate A are the five-coordinate alkene complex 7a (Figure3),
which would arise from intramolecular alkene binding of 3a, or
(11) For enantioselective reactions that are believed to occur via syn-
aminopalladation see: (a) Mai, D. N.; Wolfe, J. P. J. Am. Chem. Soc.
2010, 132, 12157.
(12) For a related intramolecular insertion of an unactivated alkene
into a Rh-O bond, see: Zhao, P.; Incarvito, C. D.; Hartwig, J. F. J. Am.
Chem. Soc. 2006, 128, 9642.
(13) Dpe-phos provides optimal results in catalytic transformations
of N-(phenyl)pent-4-enylamine, but this ligand displays somewhat
complicated coordination chemistry and can behave as either a cis- or
trans-chelating ligand. Moreover, in our hands the preparation of
(dppf)Pd(Ar)(Br) complexes proved more straightforward than synth-
esis of analogous (dpe-phos)Pd(Ar)(Br) complexes.
(15) Ligand definitions: dppf = 1,1-bis(diphenylphosphino)ferrocene;
dpp-benzene = 1,2-bis(diphenylphosphino)benzene; dppe = 1,2-bis-
(diphenylphosphino)ethane; dppp = 1,3-bis(diphenylphosphino)propane;
BINAP = 2,20-bis(diphenylphosphino)-1,10-binaphthyl; dpe-phos = bis-
(2-diphenylphosphino)phenyl ether; xantphos = 9,9-dimethyl-4,5-bis-
(diphenylphosphino)xanthene.
(16) Key data from ref 14 for (dppf)Pd(C6H4-p-CF3)[N(Me)(C6H4-p-
Me)]: 31P NMR (-45 °C) δ 9.3 (br), 24.3 (d, J = 38 Hz).
(17) Emanuel, N. M.; Knorre, D. G. Chemical Kinetics: Homoge-
neous Reactions ; Wiley: New York, 1973 (English translation by Kondor,
R.; Slutzkin, D.).
(14) Yamashita, M.; Cuevas Vicario, J. V.; Hartwig, J. F. J. Am.
Chem. Soc. 2003, 125, 16347.
(18) Kinetic data are reported as average values for k1 and k2 over two
or more separate runs.

Article
Organometallics, Vol. 30, No. 5, 2011 1271
Figure 3. Possible structures of intermediate A.
Scheme 3. 13C Labeling Experiments
Figure 1. Kinetic plot for the conversion 3a f A f 2a.
(Scheme 3). Analysis of the reaction by 13C and 31P NMR
indicated that intermediate A is the aryl(alkyl)palladium com-
plex 4a. The chemical shifts of the labeled carbon atoms in A
were not consistent with an alkene, and the chemical shift of Cb
indicated it was located adjacent to a heteroatom. Thus, these
data ruled out the possible intermediacy of 7a and 8a.21 More-
over, the 31P chemical shifts, coupling constants, and JCP values
for A/4a correlate well with data for (dppf)Pd(Ph)(Me).19,22
Effects of Palladium-Amido Complex Structure on Reac-
tivity. Following our initial experiments on the reactivity of
complex 3a, we sought to probe the effects of N-aryl group
structure, Pd-aryl group structure, and ligand structure on
the rate of the alkene aminopalladation process. To this end,
a series of (bis-phosphine)Pd(Ar)(Br) complexes were pre-
pared using standard routes23 and were treated with potas-
sium salts of N-(aryl)pent-4-enylamine derivatives in a
manner analogous to that described in Scheme 2. In all cases
the amido complexes (3a-k) were generated in <2 min at
24 °C and were characterized by diagnostic 31P NMR signals
with chemical shifts close to those observed for 3a.24 Reac-
tions were allowed to proceed at room temperature, and
kinetic data were collected by 19F NMR spectroscopy. Rate
constants for k1 (conversion of 3 to 4) and for k2 (conversion
of 4 to 2) were then determined as outlined above17,18 and are
provided in Table 1.
Figure 2. Plot of -ln([3a]/[3a]0) vs time.
the aryl(alkyl)palladium complex 4a, which derives from syn-
aminopalladation of 3a. In addition, although Pd-catalyzed
carboamination reactions have been shown to proceed through
aminopalladation pathways,1 we sought to exclude the possible
intermediacy of 8a, which would result from carbopalladation
of 3a, in the stoichiometric transformation. Unfortunately, the
data obtained in our initial experiments could not be used to
unambiguously assign the structure of A. For example, the
1H NMR alkene signals of 3a decreased as the reaction proceeded,
but this region of the spectrum was sufficiently complicated
that the presence of a new alkene-containing intermediate (7a)
could not be definitively confirmed or refuted. Similarly, the
complicated 1H NMR data also did not allow for differentia-
tion of 4a vs 8a. We observed that (dppf)Pd(C6H4-p-F)[CH2-
(cyclopentyl)] (9) generated in situ from 5a and (cyclopentyl)-
CH2MgBr underwent C-C bond-forming reductive elimination
in <5 min at room temperature,19 which seemed to argue
against the intermediacy of 4a. However, the reductive elim-
ination of 4a could be significantly slowed relative to 9 due to
the inductive electron-withdrawing effect of the nitrogen
atom in 4a.20 Thus, the identity of intermediate A could not
be ascertained without additional experimentation.
Hammett plots were constructed from the data shown in
Table 1. Clear trends were observed in transformations of
complexes 3a-f bearing various N-aryl substituents, and
linear plots of log(kR/kH) were obtained for both steps in the
conversion of 3a-f to 2a-f (Figure 4). Best fits were
In order to elucidate the structure of A, we prepared and
examined the reactivity of 13C-labeled amido complex 3a-13C3
(21) See the Supporting Information for a detailed description of the
structural assignment of 4a.
(22) Key data from ref 19 for (dppf)Pd(Ph)(Me): 31P NMR δ 17.8 (d,
J = 23 Hz), 21.3 (d, J = 23 Hz); 13C NMR JCP = 9 Hz (cis), 97 Hz
(trans).
(19) Brown has demonstrated that (dppf)Pd(Ph)(Me) undergoes
C-C bond-forming reductive elimination with a rate constant of
1.32 ꢀ 10-3 s-1 at 0 °C. See: Brown, J. M.; Guiry, P. J. Inorg. Chim.
Acta 1994, 220, 249.
(20) Prior studies have indicated that relative rates of C-C bond
forming reductive elimination from L2Pd(Ar)(CH2R) complexes dra-
matically decrease as the R group electron-withdrawing group ability
increases. See: Culkin, D. A.; Hartwig, J. F. Organometallics 2004, 23,
3398.
(23) (a) Widenhoefer, R. A.; Zhong, H. A.; Buchwald, S. L. J. Am.
Chem. Soc. 1997, 119, 6787. (b) Driver, M. S.; Hartwig, J. F. J. Am.
Chem. Soc. 1997, 119, 8232. (c) Zuideveld, M. Z.; Swennenhuis, B. H. G.;
Boele, M. D. K.; Guari, Y.; van Strijdonck, G. P. F.; Reek, J. N. H.;
Kamer, P. C. J.; Goubitz, K.; Fraanje, J.; Lutz, M.; Spek, A. L.; van
Leeuwen, P. W. N. M. Dalton Trans. 2002, 2308.
(24) See the Supporting Information for complete experimental de-
tails and characterization data.

1272 Organometallics, Vol. 30, No. 5, 2011
Table 1. Effect of N-Aryl Group and Pd-Aryl Group on Reaction Ratesa
Neukom et al.
a Conditions: all reactions were conducted in NMR tubes with [3] = 6.26 mM, [dppf] = 12.6 mM, [2-fluorotoluene] = 11.8 mM (internal standard),
THF, 24 °C. All values for k1 and k2 are the averages obtained over two or more runs. b C-N bond-forming reductive elimination from 3 to provide the
corresponding N-(C6H4-p-F)-N-(C6H4-p-R2)-pent-4-enylamine was the predominant reaction pathway observed.
obtained using the Hammett σp parameters, which gave F =
-2.5 ( 0.2 for step 1 (3 to 4) and F = -0.92 ( 0.06 for step 2
(4 to 2). The increased reactivity of complexes bearing
electron-rich N-aryl groups is consistent with trends pre-
viously reported by Hartwig for alkene insertion reactions of
cyclometalated [tBu2PCH2C6H4]Pd(NAr2) complexes.10
A similar analysis of data obtained in reactions of N-(p-
fluorophenyl)pentenylamine derived complexes 3a,g-k
bearing various R2 groups failed to provide clear trends.
Hammett plots derived from this series of experiments were
nonlinear (Figure 5), although all values of k1 for this series
were within a factor of 2.5 of each other and all values of k2
for this series were within a factor of 7 of each other. As such,
although the precise effect of Pd-aryl substituent on reactiv-
ity is unclear, it appears to be relatively small. We were
unable to obtain k1 and k2 values for reactions of complexes
derived from electron-poor aryl bromides (R = CN, CF3), as
these complexes underwent rapid C-N bond-forming re-
ductive elimination (full conversion was observed in <1 min
at room temperature) to yield N-(p-fluorophenyl)-N-(C6H4-
p-R2)pent-4-enylamines, rather than the desired aminopal-
ladation to afford 4.
The steric and electronic properties of the bis-phosphine
ligand also had a significant influence on reactivity in the
conversion of 3l-r to 2e. As shown in Table 2, the fastest
transformations were observed with wide bite angle ligands
N-methyl-nixantphos25 and xantphos.15 Amido complexes
3q,r, bearing these ligands, were rapidly converted to pyrro-
lidine 2e at 24 °C with rates too fast to accurately measure;
both reactions proceeded to completion in <1 min. In contrast,
complexes 3l-o, bearing ligands with relatively small bite
angles, failed to undergo the desired transformation. Com-
plexes 3m,o did not react at temperatures up to 60 °C, and
complexes 3l,n decomposed to afford complex mixtures of
products.26 The dpe-phos complex 3p was transformed to 2e
with an observed rate constant of 0.686 ꢀ 10-3 s-1; no
intermediate complex 4 was detected during this reaction.
The effect of ligand electronic properties on reaction rates
was examined through comparison of complexes bearing
differently substituted dppf-derived ligands. As shown in
Table 3, the presence of para-electron-withdrawing triflu-
oromethyl groups on the P-Ar substituents in complex 3s
led to acceleration of both steps of the transformation to 2d
relative to the analogous reaction of parent dppf complex 3d.
In contrast, decreased rates were observed for both steps in
the conversion of complex 3t bearing para-electron-donating
methoxy groups to 2d.
(25) nixantphos = 4,6-bis(diphenylphosphino)phenoxazine. An
N-methylated derivative of nixantphos was employed in these studies to
avoid acid/base side reactions between the ligand and the potassium N-
arylamide salts. The bite angle of this ligand is estimated to be similar to
that for N-benzyl nixantphos. See: Kamer, P. C. J.; van Leeuwen, P. W.
N. M.; Reek, J. N. H. Acc. Chem. Res. 2001, 34, 895–904.
(26) Hartwig has previously observed that (dppe)Pd(Ph)[N(p-tol)2]
decomposes to afford vinyldiphenylphosphine.23b

Article
Organometallics, Vol. 30, No. 5, 2011 1273
Figure 5. Hammett correlation for Pd-aryl group.
Figure 4. Hammett correlations for the N-aryl group.
tion of the aryl group and the N atom across the C-C double
bond. This supports a mechanism involving syn migratory
insertion of the alkene into the Pd-N bond, rather than
amide dissociation, alkene coordination, and outer-sphere
attack of the pendant nucleophile. This result is also con-
sistent with the stereochemical outcome of Pd-catalyzed
carboamination reactions between γ-aminoalkene deriva-
tives and aryl bromides.1
Determination of Activation Parameters. The experiments
outlined above indicated that the presence of electron-with-
drawing substituents on the N-aryl group decreased the rate
of both steps in the conversion of 3 to 2. Given this infor-
mation, we elected to examine the activation parameters for
this transformation using a complex bearing an electron-
poor amido group, which would react at a sufficiently slow
rate as to allow for collection of data across a reasonable
range of temperatures. Thus, the activation parameters for
the conversion of amido complex 3f to 2f (by way of 4f) were
determined by Eyring plot analysis over a temperature range
of 25-60 °C (Scheme 4).27 For the conversion of 3f to 4f
ΔHq = 24.8 ( 0.6 kcal/mol and ΔSq = 4.6 ( 1.8 eu. For the
reductive elimination of 2f from 4f ΔHq = 23.3 ( 0.8 kcal/mol
and ΔSq =4.6(2.5 eu. The reaction enthalpies are comparable
to those observed for insertion of alkenes into late-metal-
carbon bonds28a-c and for C-C bond-forming reductive
elimination processes.28d
Scheme 4. Activation Parameters
Kinetic measurements were acquired for the two-step
conversion of 3u to 2u (k1 = (0.528 ( 0.041) ꢀ 10-3 s-1
;
k2 = (0.701 ( 0.023) ꢀ 10-3 s-1) andwerecomparedto values
obtained for the analogous nondeuterated complex 3e. This
comparison indicated no significant isotope effect for step 1
(27) The N-(C6H4-p-CN) derivative 3f was employed for these stud-
ies, as it reacted at a slower rate than 3a, which simplified experimental
setup and allowed for rates to be measured over a range of temperatures
above room temperature.
(28) (a) Perch, N. S.; Widenhoefer, R. A. J. Am. Chem. Soc. 2004, 126,
6332. (b) Rix, F. C.; Brookhart, M. J. Am. Chem. Soc. 1995, 117, 1137.
(c) Ermer, S. P.; Struck, G. E.; Bitler, S. P.; Richards, R.; Bau, R.; Flood,
T. C. Organometallics 1993, 12, 2634. (d) Moravskiy, A.; Stille, J. K.
J. Am. Chem. Soc. 1981, 103, 4182.
Stereochemistry of Alkene Aminopalladation and Deuter-
ium Isotope Effects. The stereochemistry of the aminopalla-
dation reaction was determined through the reaction of deu-
terated amido complex 3u. As shown in eq 1, this complex
was cleanly transformed to pyrrolidine 2u with net syn addi-

1274 Organometallics, Vol. 30, No. 5, 2011
Table 2. Effect of Ligand Bite Angle on Reactivitya
Neukom et al.
starting complex
intermediate complex
ligand
dppe
dpp-benzene
dppp
(()-BINAP
dppf
dpe-phos
xantphos
nixantphos-Me
bite angle (deg)
result
3l
not obsd
not obsd
not obsd
not obsd
4e
not obsd
not obsd
not obsd
86
87
91
93
dec of complexb
no reacnc
3m
3n
3o
3e
3p
3q
3r
dec of complexb
no reacnc
99
k1 = (0.56 ( 0.02) ꢀ 10-3 s-1; k2 = (0.90 ( 0.02) ꢀ 10-3 s-1
104
108
114
kobs = 0.686 ꢀ 10-3 s-1d
too fast to measuree
too fast to measuree
a Conditions: all reactions were conducted in NMR tubes with [3] = 6.26 mM, [ligand] = 12.6 mM, [2-fluorotoluene] = 11.8 mM (internal
standard), THF, 24 °C. All values for k1 and k2 are the averages obtained over two or more runs. b Decomposition to afford a complex mixture of
products was observed. The expected pyrrolidine 2e was not detected in significant amounts. c No reaction was observed at temperatures up to 60 °C.
d No intermediate was detected in this reaction. e Complete conversion to 2e was observed within 1 min of mixing 3q,r and the potassium
N-arylamide salt.
Table 3. Ligand Electronic Effectsa
the rate of reductive elimination from intermediate 4w to
yield 2w is comparable to that for the transformation of 4d to
2d. Complexes 3x,y failed to undergo aminopalladation at
temperatures up to 60 °C.
Discussion
Plausible Mechanistic Scenarios for the Conversion of
Amido Complexes 3 to (Aryl)(pyrrolidin-2-ylmethyl)palladium
starting
complex
intermediate
complex
k1
k2
ligand
(10-3 s-1
)
(10-3 s-1
)
Complexes 4. The conversion of (bis-phosphine)Pd(Ar)[N-
(Ar1)(CH2)3C(R)dC(R)(R0)] complexes 3 to (bis-phosph-
ine)Pd(Ar)(pyrrolidin-2-ylmethyl) complexes 4 presumably
does not occur via a single step but instead likely involves (a)
intramolecular coordination of the alkene to palladium and
(b) syn-aminopalladation. As such, the conversion of 3 to 4
could potentially proceed through four different reasonable
pathways (Scheme 5). Two scenarios would involve syn-
aminopalladation from the five-coordinate complex 7. The
first would entail rate-limiting alkene coordination of 3 to
provide 7, which could then undergo rapid aminopallada-
tion to afford 4 (path A). Alternatively, fast and reversible
intramolecular alkene coordination of 3 would yield 7,
which could undergo rate-limiting aminopalladation to 4
(path B).
Two other possibilities would involve ligand substitu-
tion of alkene for one arm of the chelating bis-phosphine
ligand to give four-coordinate alkene complex 10 (pre-
sumably via an associative mechanism; 7 would be a
transient intermediate en route to 10).29 One of these
pathways (path C) would proceed via rate-limiting associa-
tive substitution of 3 to give 10, followed by fast aminopallada-
tion of 10 to yield 4. Finally, fast and reversible associative
ligand substitution of 3 to 10 followed by rate-limiting
aminopalladation from 10 to 4 is also a reasonable possibility
(path D).
3s
3d
3t
4s
4d
4t
dppf-p-CF3
dppf
dppf-p-
OMe
4.08 ( 0.03
2.44 ( 0.12
0.77 ( 0.01
14.6 ( 2.4
1.88 ( 0.07
0.59 ( 0.20
a Conditions: all reactions were conducted in NMR tubes with [3] =
6.26 mM, [ligand] = 12.6 mM, [2-fluorotoluene] = 11.8 mM (internal
standard), THF, 24 °C. All values for k1 and k2 are the averages obtained
over two or more runs.
of the transformation (k1H/k1D = 1.06 ( 0.09), but a small
isotope effect was observed for the reductive elimination
(step 2, k2H/k2D = 1.25 ( 0.05). Related experiments con-
ducted with substrate 3v, which contains a deuterium atom at
the internal alkene carbon, indicated no significant isotope
effect (eq 2).
Effect of Alkene Substitution on Reactivity. In order to
probe the effect of alkene substitution on reactivity, com-
plexes 3w-y, bearing 1,1- or 1,2-disubstituted alkenes, were
prepared in a manner analogous to that for the amido
complexes described above. As shown in Table 4, complex
3w, which contains a substituent at the internal alkene
carbon, undergoes aminopalladation to give intermediate
4w at a rate that is 10-fold slower than for the analogous
conversion of unsubstituted derivative 3d to 4d. However,
(29) Ligand substitution reactions of d8-Pd(II) complexes generally
occur via associative pathways. See: (a) Qian, H.; Widenhoefer, R. A.
J. Am. Chem. Soc. 2003, 125, 2056. (b) Shultz, L. H.; Tempel, D. J.;
Brookhart, M. J. Am. Chem. Soc. 2001, 123, 11539and references cited
therein. For rare exceptions, which involve very large, sterically bulky
ligands, see: (c) Bartolome, C.; Espinet, P.; Martin-Alvarez, J. M;
Villafane, F. Eur. J. Inorg. Chem. 2004, 2326. (d) Louie, J.; Hartwig,
J. F. J. Am. Chem. Soc. 1995, 117, 11598.

Article
Organometallics, Vol. 30, No. 5, 2011 1275
Table 4. Alkene Substituent Effectsa
starting complex
intermediate complex
product
R
R1
R2
k1 (10-3 s-1
)
k2 (10-3 s-1
)
3d
3w
3x
3y
4d
4w
not obsd
not obsd
2d
2w
H
H
Me
H
H
H
H
Me
H
Me
H
2.44 ( 0.12
1.88 ( 0.07
0.25 ( 0.09
1.58 ( 0.16
b
b
b
b
H
a Conditions: all reactions were conducted in NMR tubes with [3] = 6.26 mM, [dppf] = 12.6 mM, [2-fluorotoluene] = 11.8 mM (internal standard),
THF, 24 °C. All values for k1 and k2 are the averages obtained over two or more runs. b No reaction was observed up to 60 °C.
As shown in Scheme 5, paths B and D for the conversion of
3 to 4 both involve rapid formation of an alkene-bound
complex (7 or 10) followed by rate-limiting aminopallada-
tion (from either 7 or 10). The fact that neither 7 nor 10 is a
detectable intermediate argues against paths B and D but
cannot be used to rule out these pathways, as it is possible the
equilibrium between 3 and 7 or 3 and 10 is fast but lies far to
the left, favoring 3. In contrast, the results obtained in reac-
tions with deuterated substrates 3u,v provide good evidence
that neither path B nor path D is in operation.30 The trans-
formations of 7 to 4 and 10 to 4 both involve rehybridization
of the alkene carbon atoms from sp2 to sp3. As such, if this
step were rate limiting, a significant deuterium isotope effect
should be observed at both alkene carbon atoms. However,
the conversions of deuterated complexes 3u,v to 4u,v proceed
at the same rate as the transformation of all-protio complex
3e to 4e. Finally, the observed effect of ligand bite angle on
reaction rate (Table 2) provides additional evidence against path
D, as the bite angle should not influence the rate of aminopall-
dation from 10 to 4 if the ligand is not bound to the metal by
both phosphine groups in the rate-determining step.31
The effect of ligand and amine properties on reaction rate
can also be used to differentiate between paths A and C. If
transformations proceed via path C, the reaction rate should
be strongly influenced by factors that favor phosphine dis-
placement.32 In contrast, the rate of reactions that proceed
by way of path A should be insensitive to factors that favor
ligand substitution and instead should only be affected by
parameters that influence initial alkene binding to the metal.
The effect of phosphine ligand properties on reaction rate is
most consistent with reaction via path C. As illustrated in
Scheme 4, complex 3s, which contains electron-withdrawing
p-CF3-C6H4 groups on the phosphines, reacts ca. 5 times
faster than the related complex 3t, which bears p-MeO-C6H4
phosphine substituents. The displacement of one arm of the
electron-poor p-CF3-dppf ligand should be more facile than
for the relatively electron rich p-MeO-dppf ligand. In addi-
tion, although we were unable to obtain quantitative rate
data for ligands with very large or very small bite angles,
qualitatively it is clear that the transformation is facilitated
by wide bite angle ligands and impeded by ligands with small
bite angles. This effect is also consistent with rate-limiting
associative ligand substitution (path C).33
Paths A and C both involve rate-limiting alkene coordina-
tion to the metal center but differ in the nature of the
intermediate complex that undergoes aminopalladation. In
path A, aminopalladation would occur directly from the
five-coordinate intermediate 7, whereas path C involves
substitution of alkene for phosphine followed by insertion
from four-coordinate complex 10. Several pieces of evidence
indicate that the mechanism of conversion of 3 to 4 does not
proceed via path A. First of all, the positive entropy values
measured for the conversion of 3f to 4f suggest that path A is not
operating, as the conversion of 3 to 7 should have a fairly large
negative entropy of activation due to the increase in organization
in the transition state between complex 3 with a single chelate
(P-P) and 7, which contains two chelates (P-P and alkene-N).
In contrast, the measured entropy of þ4.6 eu is consistent with the
conversion of 3 to 10 via intermediate 7 (path C), as monoche-
lated complex 10 is less ordered than doubly chelated complex 7.
The reactions of complexes bearing electron-rich N-aryl
groups are considerably faster than those bearing electron-
poor N-aryl groups. For example, the conversion of com-
plexes 3b,c, which contain electron-donating p-tBu and
p-OMe groups on the N-aryl moiety, to 4b,c is 2 orders of
magnitude faster than the conversion of 3f to 4f (N-aryl =
p-CN-C6H4). This electronic effect also suggests that the
conversion of 3 to 4 proceeds via path C rather than path A.
The electron-rich amido groups should increase the elec-
tron density of the metal center, which should in turn increase
the ease of phosphine displacement. In contrast, if path A
were operating, coordination of the relatively electron-rich
alkene should be facilitated by a less electron-rich, more
Lewis acidic metal center,34,35 and rates should be faster with
relatively electron-poor N-aryl groups.
(32) The lability of a chelating ligand has previously been observed to
influence the rate of carbopalladation in (L-L)Pd(Me)[PPh2(C6H4-o-
CHdCH2)] complexes. See: Cavell, K. J.; Jin, H. J. Chem. Soc., Dalton
Trans. 1995, 4081.
(33) Delis, J. G. P.; Groen, J. H.; Vrieze, K.; van Leeuwen, P. W. N.
M. Organometallics 1997, 16, 551.
(30) Prior studies by Hartwig have shown that cyclometalated four-
coordinate [tBu2PCH2C6H4]Pd(NAr2)(ethylene) complexes undergo
very fast insertion of the alkene into the Pd-N bond at temperatures
as low as -40 °C.10
(31) Although the wide-bite-angle ligands dpe-phos, xantphos, and
N-Me-nixantphos are more electron rich than dppf, the trends observed
with para-substituted dppf derivatives (complexes 3s,t) suggest that the
large bite angle is responsible for the rate enhancement rather than
electronic properties. The electron-rich complex 3t reacts at only a
slightly slower rate than 3s. In addition, dpe-phos and xantphos have
similar electronic properties, but the complexes bearing these ligands
have considerably different reactivities.
(34) Tanaka, D.; Romeril, S. P.; Myers, A. G. J. Am. Chem. Soc.
2005, 127, 10323.
(35) The opposite trend has been observed in reactions between
cationic methylpalladium complexes and electron-poor alkenes. This
effect has been ascribed to the influence of π back-bonding interactions
between a relatively electron rich metal center and the electron-poor
alkene. See: Wu, F.; Jordan, R. F. Organometallics 2006, 25, 5631–5637.

1276 Organometallics, Vol. 30, No. 5, 2011
Scheme 5. Possible Mechanistic Pathways for Conversion of 3 to 4
Neukom et al.
No clear trend was observed for the influence of Pd-aryl
group electronics on the rate of conversion of 3 to 4. As such,
these data cannot be used to refute any of the possible mech-
anistic pathways. The origin of these electronic effects is
unclear, but the differences in relative rates of aminopallada-
tion for complexes 3a,g-i are small (within a factor of ca. 2).
However, our data do indicate that the rate of C-N bond-
forming reductive elimination dramatically increases relative
to the rate of aminopalladation in complexes 3j,k, which
contain strong electron-withdrawing substituents on the Pd-
Ar group.
Influence of Structural Features on Carbon-Carbon Bond-
Forming Reductive Elimination of 4 To Afford 2. The rate of
C-C bond-forming reductive elimination from (dppf)Pd-
(Ar)(pyrrolidin-2-ylmethyl) complexes 4 is also affected by
the structural features of the complexes. For example, the
electronic properties of the N-aryl group have a significant
influence on this transformation, as complexes 4 bearing
electron-rich N-aryl groups undergo reductive elimination 5
times faster than electron-poor derivatives. In addition, the
rate of reductive elimination of the (dppf)Pd(Ar)(pyrrolid-
ine-2-ylmethyl) complexes is considerably slower than the
analogous reaction of the (dppf)Pd(Ar)(alkyl) derivative
(dppf)Pd(C6H4-p-F)[CH2(cyclopentyl)] (9). These trends
are likely due to inductive effects that slow the relative rate
of reductive elimination as the electron-withdrawing power
of the nitrogen atom increases in derivatives of 4.20 The
possibility that the rate of reductive elimination is slowed by
binding of the nitrogen atom in 4 to the metal center appears
less likely, given the fact that electron-poor N-aryl groups
should disfavor N-coordination, but rates are slowest with
these groups.
state for C-C bond formation.37 The observed deuterium
isotope effect at the carbon undergoing bond formation is
consistent with rate-limiting C-C bond formation in the
conversion of 4 to 2, rather than rate-limiting phosphine
dissociation.
Conclusions
In conclusion, our experiments on the conversion of
(P-P)Pd(Ar)[N(Ar1)(CH2)3CRdCHR0] complexes 3 to N-
aryl-2-benzylpyrrolidine derivatives 2 indicate that the trans-
formations proceed via syn insertion of the alkene into the
Pd-N bond. This alkene syn-aminopalladation pathway has
rarely been observed in well-characterized palladium com-
plexes but plays a key role in catalytic reactions. These studies
illustrate that ligand structure and heteroatom basicity/
nucleophilicity10 have a large impact on the rate of amino-
palladation, and the observed trends could potentially be
used in the design of new catalysts for reactions involving
aminopalladation. Finally, our data suggest that insertion
occurs from a four-coordinate alkene complex, rather than a
five-coordinate species. This mechanistic information pro-
vides insight into previously observed trends in asymmetric
Pd-catalyzed alkene carboamination reactions. Use of chiral
bis-phosphine ligands provides poor enantioselectivity in these
transformations,11 which is likely due to dissociation of one
arm of the bis-phosphine ligand prior to aminopalladation.
Experimental Section
In Situ Formation of Pd-Amido Complexes 3 and Conversion to
2. Representative Procedure. In a nitrogen-filled glovebox,
(dppf)Pd(C6H4-p-F)(Br) (5a; 6.3 mg, 0.0075 mmol) and dppf
(4.7 mg, 0.0085 mmol) were placed into a small vial. THF-d8
(550 μL) was added, and the resulting orange solution was
transferred to an NMR tube. 4-Fluorotoluene or 2-fluoroto-
luene (0.3 μL, 0.0027 mmol) was added as an internal 19F stan-
dard, and the tube was sealed with a septum. The tube was
cooled to -60 °C in the probe of an NMR spectrometer, and 1H,
19F, and 31P spectra were obtained. A solution of potassium (4-
fluorophenyl)(pent-4-enyl)amide (6a; 2.7 mg, 0.012 mmol) in
200 μL of THF-d8 was prepared in the glovebox, and 121 μL (1
equiv) of that solution was loaded into a gastight syringe and
injected into the NMR tube containing the Pd complex. The
tube was inverted several times to ensure complete mixing, and a
rapid color change from orange to red was observed. The tube
was returned to the cold NMR probe and allowed to re-
equilibrate at -60 °C, and a 19F spectrum was obtained. The
solution was then warmed to -20 °C and 1H, 19F, and 31P
spectra were obtained.
The effect of ligand electronic properties and bite angle on
the rate of reductive elimination from 4 to 2 is also consistent
with prior observations on the rates of C-C bond formation
from Pd(II) complexes.36,37 In our system complex 4s, which
bears a relatively electron poor ligand, undergoes reductive
elimination 25 times faster than the related complex 4t,
which is ligated by a more electron-rich phosphine. This is
likely due to the destabilizing effect of the electron-poor
phosphine on the Pd(II) oxidation state.36 The reductive
elimination processes also appear to be most facile with wide-
bite-angle ligands, which both destabilize the ground state of
(P-P)Pd(Ar)(R) complexes and also stabilize the transition
(36) (a) Ariafard, A.; Yates, B. F. J. Organomet. Chem. 2009, 694,
2075. (b) Ananikov, V. P.; Musaev, D. G.; Morokuma, K. Eur. J. Inorg.
Chem. 2007, 5390. (c) Hartwig, J. F. Inorg. Chem. 2007, 46, 1936.
(37) Zuidema, E.; van Leeuwen, P. W. N. M.; Bo, C. Organometallics
2005, 24, 3701.
(dppf)Pd(C6H4-p-F)[N(C6H4-p-F)(CH2CH2CH2CHdCH2)]
(3a). 1H NMR (400 MHz, THF-d8): δ 7.95-7.87 (m, 4 H), 7.76

Article
Organometallics, Vol. 30, No. 5, 2011 1277
(t, J = 8.6 Hz, 2 H), 7.59-7.40 (m, 10 H), 7.20-7.12 (m, 4 H),
7.06-7.01 (m, 2 H), 6.60 (q, J = 7.4 Hz, 2 H), 6.36 (t, J = 8.4 Hz,
2 H), 6.09 (t, J = 8.6 Hz, 2 H), 5.62 (tdd, J = 6.8, 10.0, 16.8 Hz,
1 H), 4.88-4.81 (m, 2 H), 2.57-2.51 (m, 1 H), 2.22-2.14
(m, 1 H), 1.70 (m, obscured by THF), 1.42-1.30 (m, 1 H),
1.20-1.08 (m, 1H). 19F NMR (376 MHz, THF-d8): δ -123.7 (m,
Pd-C6H4F), -137.3 (s, N-C6H4F); 31P NMR (162 MHz, THF-
d8): δ 24.9 (d, J = 38.1 Hz), 9.0 (d, J = 35.5 Hz).
The solution of the palladium-amido complex 3a was warmed
to 15 °C, with monitoring at 2 min intervals by 19F spectroscopy for
the appearance of peaks at -124.1 and -133.3 ppm (attributed to
4a), along with diminishment of the peaks at -123.7 and -137.3
ppm. When the new peaks were near their maximum, the solution
was cooled to -20 °C and 1H, 19F, and 31P spectra were obtained.
19F and 31P data are reported; the 1H spectrum for this and related
compounds could not be extracted from the combined spectra of the
species present in the reaction mixture.
quenched with saturated aqueous NH4Cl (2 mL), and diluted
with ethyl acetate (10 mL). The layers were separated, and the
aqueous layer was extracted with ethyl acetate (2 ꢀ 10 mL).
The combined organic extracts were dried over anhydrous
sodium sulfate, filtered, and concentrated in vacuo. The crude
product was then purified by flash chromatography on silica
gel to afford 53 mg (68%) of the title compound as a colorless
oil. 1H NMR (400 MHz, CDCl3): δ 7.19-7.11 (m, 2 H),
7.03-6.93 (m, 4 H), 6.60-6.53 (m, 2 H), 3.94-3.85 (m, 1 H),
3.42-3.32 (m, 1 H), 3.19-3.08 (m, 1 H), 2.94 (dd, J = 3.2, 13.6
Hz, 1 H), 2.58 (dd, J = 8.8, 13.6 Hz, 1 H), 1.95-1.74 (m, 4 H).
13C NMR (100 MHz, CDCl3): δ 161.8 (d, J = 245 Hz), 155.0
(d, J = 234 Hz), 143.9 (d, J = 1.6 Hz), 135.1 (d, J = 3.1 Hz),
130.9 (d, J = 7.6 Hz), 115.9 (d, J = 22.2 Hz), 115.4 (d, J = 20.7
Hz), 112.4 (d, J = 6.9 Hz), 60.2, 49.2, 38.0, 29.9, 23.4. 19F
NMR (376 MHz, CDCl3): δ -117.0 (m), -130.7 (m). IR (film):
1225 cm-1. MS (ESI): m/z 274.1413 (274.1407 calcd for
C17H17F2N, M þ Hþ).
(dppf)Pd(C6H4-p-F){CH2[CHCH2CH2CH2N(C6H4-p-F)]} (4a).
19F NMR (376.9 MHz, THF-d8): δ -124.1 (s), -133.3 (s). 31P NM-
R (162 MHz, THF-d8):δ21.3(d,J= 24.1 Hz), 16.6 (d, J= 21.7 Hz).
Synthesis of Authentic Samples of Pyrrolidine Products For-
med in Kinetic Runs. Representative Procedure.1a Synthesis of
1-(4-Fluorophenyl)-2-(4-fluorobenzyl)pyrrolidine (2a). An oven-
or flame-dried Schlenk tube was cooled under a stream of argon
or nitrogen and charged with Pd2(dba)3 (2.6 mg, 2.8 μmol), dppf
(3.1 mg, 5.6 μmol), NaOtBu (27 mg, 0.28 mmol), and 1-bromo-
4-fluorobenzene (47 μL, 0.42 mmol). The tube was purged with
argon or nitrogen, and a solution of 4-fluoro-N-(pent-4-en-
yl)aniline (51 mg, 0.28 mmol) in toluene (1 mL) was added.
The mixture was heated to 80 °C with stirring until the starting
material had been consumed, as judged by GC analysis (3 h).
The reaction mixture was cooled to room temperature,
Acknowledgment. We thank the National Science
Foundation (Grant No. CHE-0705290) for financial
support of this work. Additional funding was provided
by GlaxoSmithKline, 3M, Amgen, and Eli Lilly. We also
thank Pfizer for a generous gift of Pd2(dba)3. J.D.N. was
supported by a GAANN fellowship.
Supporting Information Available: Text, tables, and figures
giving experimental procedures, characterization data for all
new compounds, copies of 1H, 31P, 19F, and 13C NMR spectra
for selected compounds, and descriptions of stereochemical
assignments. This material is available free of charge via the
Internet at http://pubs.acs.org.