Mechanistic investigations of palladium-catalyzed allylic fluorination

Article abstract of DOI:10.1021/om401240p

A computational and experimental approach was employed to study the mechanism of the palladium(0)-catalyzed fluorination of allylic chlorides with AgF as fluoride source. Our findings indicate that an allylpalladium fluoride is a key intermediate necessar

Full text of DOI:10.1021/om401240p

Article
pubs.acs.org/Organometallics
Mechanistic Investigations of Palladium-Catalyzed Allylic
Fluorination
Matthew H. Katcher,† Per-Ola Norrby,‡,§ and Abigail G. Doyle*,†
†Department of Chemistry, Princeton University, Princeton, New Jersey 08544, United States
‡Department of Chemistry and Molecular Biology, University of Gothenburg, Kemigarden 4, #8076, SE-412 96 Goteborg, Sweden
̊
̈
§Pharmaceutical Development, Global Medicines Development, AstraZeneca, Pepparedsleden 1, SE-431 83 Molndal, Sweden
̈
S
* Supporting Information
ABSTRACT: A computational and experimental approach was
employed to study the mechanism of the palladium(0)-catalyzed
fluorination of allylic chlorides with AgF as fluoride source. Our
findings indicate that an allylpalladium fluoride is a key intermediate
necessary for the generation of both the nucleophile and electrophile.
Evidence was also obtained to support a homobimetallic mechanism
in which C−F bond formation occurs by nucleophilic attack of a
neutral allylpalladium fluoride on a cationic allylpalladium electro-
phile (with fluoride as counterion). The high branched selectivity and
unusual ligand effects observed in the regioselective fluorination are assessed in light of this mechanism and calculated transition
states. These results may have important implications for the mechanism of other transition-metal-catalyzed fluorinations.
Scheme 1. Strategies for Pd-Catalyzed C−F Bond Formation
INTRODUCTION
■
Methods for carbon−fluorine bond formation have valuable
applications in the pharmaceutical, agrochemical, radiochem-
ical, and materials industries.1 Although these reactions have
been studied for over a century,2 selective C−F bond formation
enabled by catalysis has been the subject of intense research for
only the past decade.3 Palladium is a particularly attractive
catalyst for these transformations, as it generally exhibits high
functional group tolerance and a variety of ligands can be used
to rationally modulate its reactivity.4
Initial reports of palladium-catalyzed C−F bond formation
focused on the synthesis of aryl fluorides.5 Sanford,6 Yu,7 and
Ritter8 have illustrated that highly oxidizing, electrophilic
fluorinating reagents can allow access to high-valent Pd
intermediates from which C−F reductive elimination can
occur (Scheme 1a). Despite the requirement for stoichiometric
Pd or directing groups in the case of catalytic reactions, the
demonstration of this reactivity was a landmark achievement.9
An alternative strategy for Pd-catalyzed C−F bond formation
relies on the reduction of an organopalladium(II) intermediate
with a nucleophilic fluorine source (fluoride); this approach is
especially attractive due to the lower cost and higher availability
of fluoride-based reagents, in comparison to electrophilic
fluorine sources.10 For aryl fluoride synthesis, the Pd(0)/Pd(II)
catalytic cycle would parallel that for traditional cross-coupling
reactions,11 with C−F bond formation occurring by inner-
sphere reductive elimination from an arylpalladium fluoride
intermediate (Scheme 1b). This mechanism has been well
studied for other C−X reductive eliminations (X = C, N, O, S,
other halides, etc.), but extensive investigations by Grushin12
and Yandulov13 identified distinct complications for C−F bond
formation. Although C−F bond formation is thermodynami-
cally favored, a number of undesired side reactions (C−P, P−F,
P−P bond formation) were found to be more kinetically
favorable than C−F reductive elimination from an
arylpalladium(II) fluoride. To overcome this challenge,
Buchwald and co-workers utilized an electron-rich biarylphos-
phine ligand for the Pd(0)-catalyzed fluorination of aryl triflates
with CsF at high temperatures.14
As part of a program to develop catalytic C−F bond-forming
reactions with fluoride,15 we recently identified an alternative
strategy for fluorination with Pd(0)/(II) catalysis. To circum-
vent the aforementioned challenges associated with inner-
Received: December 31, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
sphere C−F reductive elimination from Pd(II), we envisioned
an outer-sphere pathway for C−F bond formation proceeding
by nucleophilic attack of fluoride on an organic ligand
coordinated to Pd (Scheme 1c). This mechanism could
potentially operate under conditions milder than those utilized
for Pd-catalyzed aryl C−F bond formation. Moreover, such an
approach would complement existing technologies for aryl
fluorination by providing entry into aliphatic fluorinated motifs.
Initially, we targeted the reaction of fluoride with an
allylpalladium intermediate to synthesize allylic fluorides,
since this motif is valuable in bioactive compounds and as a
synthetic intermediate.16 Our proposed mechanism had been
demonstrated for a number of other nucleophiles in Pd-
catalyzed allylic substitution.17 Furthermore, a theoretical study
had indicated that outer-sphere fluoride attack on an
allylpalladium intermediate is kinetically feasible.18 However,
this transformation had not been validated experimentally prior
to our work.
Notably, Togni and co-workers had extensively investigated
the viability of this approach to allylic fluoride synthesis.19
Using several different fluoride sources, the authors observed
no evidence for C−F bond formation in reactions with
stoichiometric allylpalladium complexes ligated by bidentate
P,N ligands. Instead, fluoride acted as a base to provide diene,
via elimination, as the major organic product in these studies.20
The authors also encountered an additional challenge in that
the desired allylic fluoride product was shown to be unstable to
their reaction conditions. In related work, Hazari, Gouverneur,
and Brown have also described the Pd(0)-catalyzed alkylation
of allylic fluorides with sodium malonate.21
(Scheme 2b).23 Recently, we have also extended this reactivity
to the carbofluorination of allenes via a cascade reaction that
also proceeds through an allylpalladium intermediate.24
After our initial publication, several other laboratories
reported transition-metal-catalyzed allylic fluorinations that
each rely on distinct combinations of catalyst, substrate, and
fluoride source. Using TBAF·(t-BuOH)4, Gouverneur, Brown,
and co-workers have described the fluorination of allylic
benzoates and carbonates under palladium and iridium
catalysis.25 Ir(I) catalysis has also been applied to the
fluorination of allylic trichloroacetimidates with Et3N·3HF by
Nguyen and co-workers.26 Additionally, Lauer and Wu have
discovered that AgF also serves as a fluoride source for the
Pd(0)-catalyzed fluorination of allylic phosphorothiorate
esters.27 Recently, several allylic fluorinations catalyzed by Rh,
Cu, and Pd have also been reported with Et3N·3HF as fluoride
source.28
Despite the successful development of the aforementioned
reactions for Pd(0)-catalyzed C−F bond formation, little
progress has been reported toward understanding the
mechanisms responsible for this reactivity.29 Herein, we present
detailed studies of Pd-catalyzed allylic fluorination with AgF
aimed at understanding the mechanism of C−F bond
formation from Pd(II). This work includes computational
and experimental investigations that implicate an allylpalladium
fluoride as a key intermediate in the Pd-catalyzed fluorination
of allylic chlorides. Although C−F bond formation occurs by an
outer-sphere mechanism, the proposed nucleophile and
electrophile are both derived from this palladium fluoride.
Furthermore, we suggest that this intermediate is responsible
for the remarkable reactivity and selectivity that is observed
under our optimized reaction conditions. As in other recent
reports,30 our work highlights the complex roles that transition-
metal fluorides play in catalytic fluorinations.
As a result of these studies, we recognized that distinct
conditions would be required to support a Pd-catalyzed allylic
fluorination. In initial investigations toward developing a
catalytic reaction, we therefore examined an extensive array of
substrates and metal fluoride salts. From this evaluation, the
unique combination of an allylic chloride as substrate and AgF
as fluoride source provided high chemoselectivity for C−F
bond formation. Application of Trost bisphosphine ligands L1
and L2 also afforded high enantioselectivity in the fluorination
of cyclic substrates (Scheme 2a,c).22 In further studies, we have
demonstrated that these conditions are suitable for the highly
regioselective synthesis of branched acyclic allylic fluorides
RESULTS AND DISCUSSION
■
Initial Mechanistic Investigations. In our initial report on
Pd-catalyzed allylic fluorination, the reaction mechanism was
intriguing due to several conflicting observations.22 As in our
original proposal, we found that C−F bond formation occurs
via an outer-sphere pathway wherein fluoride attacks the allyl
ligand with inversion of stereochemistry (Scheme 3). Although
Scheme 3. Inversion of Stereochemistry in Outer-Sphere
Fluoride Attack on an Allylpalladium Intermediate
Scheme 2. Palladium-Catalyzed Allylic Fluorination of
Allylic Chlorides
fluoride is generally classified as a hard nucleophile due to its
high charge density, this stereochemical outcome differs from
that obtained in allylic substitutions with hard nucleophiles
such as organomagnesium reagents and hydride. These
reactions are known to proceed by retention of stereochemistry
via an inner-sphere mechanism in which transmetalation of the
nucleophile precedes reductive elimination.17,31
We also documented an important difference in the substrate
requirements for allylic fluorinations and substitutions with
B
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
other nucleophiles. Although typical substrates for Pd-catalyzed
asymmetric allylic alkylations include allylic acetates and
carbonates, little to no C−F bond formation was observed
with these substrates under the conditions we identified for
allylic fluorination. Instead, allylic chlorides and bromides,
which are less common substrates for Pd-catalyzed allylic
substitution,32,33 were necessary for Pd-catalyzed allylic
fluorination with AgF. We initially proposed that the
precipitation of AgCl provides a driving force for C−F bond
formation but did not fully understand the role of this process
in the overall catalytic cycle.
In the regioselective fluorination of allylic halides, we also
noted several features unlike other Pd-catalyzed allylic
substitutions.23 In contrast to reactions with other nucleo-
philes,34 high regioselectivity for branched allylic fluorides was
obtained (Scheme 4). Although heteroatom nucleophiles
intermediate) or “naked” fluoride anion generated from one
of these complexes.
Unfortunately, efforts to further study the mechanism of Pd-
catalyzed allylic C−F bond formation in these reactions by
traditional spectroscopic methods and kinetic techniques were
hampered by the heterogeneity of the reaction mixtures.
Although reproducible yields and selectivities could be attained
with a high stir rate, reaction rates varied from run to run. We
were also unable to obtain high chemo-, regio-, and
enantioselectivity for fluorinations in solvents more polar
than toluene and with fluoride sources more soluble than AgF.
In this work, we therefore begin by evaluating plausible
reactive intermediates using theoretical methods. To support
the conclusions from these studies, experiments are then
performed with complexes that model potential intermediates.
As a simplification, we focus on the regioselective reaction with
PPh3 as ligand. Although this reaction is not as highly
regioselective as that with L2, it provides a more tractable
experimental and computational system.37
Scheme 4. Palladium-Catalyzed Regioselective Allylic
Fluorination
Computational Methods and Potential Intermediates.
DFT calculations were performed with Gaussian09,38 and the
methods were validated by comparing the solid-state structure
of 8 with the calculated geometry (full details are contained in
the Supporting Information). Notably, all geometry optimiza-
tions and single-point energy calculations were performed in
toluene with the PCM solvent model.39 Previous work has
demonstrated that geometry optimizations in solvent are
necessary to obtain transition states for the attack of fluoride
on an allylpalladium intermediate.18
typically provide moderate selectivity for the branched product,
the selectivities observed in allylic fluorination surpass those
reported for reactions of amines and phenols with similar
unbiased substrates (2:1 to 5:1).35 Additionally, allylic
substitutions with other nucleophiles are typically compatible
with a large array of bisphosphine ligands, including dppe, dppf,
and BINAP. However, good reactivity in regioselective allylic
fluorination was only observed with monodentate phosphines
such as PPh3 and bidentate phosphines with large bite angles
(Xantphos and L2).
To benchmark our calculations, allylpalladium complexes 8
and 9 were synthesized,40 and their reactivity with AgF was
examined (Scheme 6). In line with analogous catalytic
Scheme 6. Reactivity of Allylpalladium Chloride Complexes
a
with AgF
We made the most intriguing observation relating to the
mechanism of allylic C−F bond formation in the context of our
stoichiometric studies on the carbofluorination of allenes.24 In
this work, we demonstrated a Heck-fluorination cascade with
an aryl iodide, allene, and AgF. Mechanistically, outer-sphere
C−F bond formation occurs from an allylpalladium inter-
mediate, and stoichiometric studies indicated that a neutral
allylpalladium fluoride was generated after allene insertion into
the Pd−Ar bond (Scheme 5). Notably, the stoichiometric
reaction of 5 and 6 proceeds with or without AgF to afford 7 in
similar regioselectivity and yield. We therefore proposed that, in
the Ag-free reaction, the active fluorinating reagent could be a
palladium fluoride (536 or the allylpalladium fluoride
a
Regioselectivities were determined by GC (average of three
experiments). Reactions were conducted in the presence of dimethyl
fumarate (1 equiv).
reactions,41 both complexes afforded allylic fluorides with
regioselectivity in favor of the branched isomer. The differences
in experimental barriers leading to linear and branched
products can be determined from these regioselectivities
(reactions with complex 8, ΔΔG⧧ = 9 0.7 kJ/mol; reactions
with complex 9, ΔΔG⧧ = 6 0.5 kJ/mol).
Potential structures for the electrophile in the allylic
fluorination are shown in Chart 1. Neutral complex A would
be the immediate product of oxidative addition, since allylic
chlorides are known to react with Pd(0) to afford neutral
Scheme 5. Stoichiometric Carbofluorination of Allenes
a
Proceeding via an Allylpalladium Fluoride Intermediate
Chart 1. Structures of Potential Electrophiles
a
>20:1 b:l for both reactions.
C
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
intermediates with coordinated chloride: e.g., 1.32c Halide
exchange between A and AgF would result in another potential
electrophile: neutral allylpalladium fluoride B. Cationic complex
C is also a possible electrophile; similar cationic complexes have
been proposed in other Pd-catalyzed allylic substitutions (vide
infra).42
We also evaluated three possible nucleophiles for Pd-
catalyzed allylic fluorination. In toluene, monomeric AgF
would likely bind to at least one additional ligand for stability
and solubility.43 Therefore, we investigated (PPh3)AgF as one
potential nucleophile. Phosphine−Ag complexes are well
precedented:44 for example, Yamamoto has reported asym-
metric Sakurai allylations catalyzed by BINAP−AgF com-
plexes.45 In our reaction, only limited dissociation of PPh3 from
Pd would be necessary to ligate a catalytic amount of AgF.46
Inspired by our studies on carbofluorination, complex B and
“naked” fluoride in complex C were also included in the
computational study as possible nucleophiles.
Figure 2, and the calculated ΔΔG⧧ value for these pathways
(14 kJ/mol) is closer to the experimental value for reactions
with complex 8 (9 kJ/mol) in comparison to that for Figure 1.
With the computational methods applied here, this error is
probably not significant. Although the activation barriers for
these reactions are much lower than those in Figure 1, we limit
our comparisons to ΔΔG⧧ values. Previously, it has proven
difficult to compare calculated energies of charged and neutral
intermediates.49
We also computed pathways for the reaction of nucleophile
17 and cationic complex 16, and the calculated ΔΔG⧧ value is
11 kJ/mol (Figure S6, Supporting Information).48 This value is
within error of the experimental selectivity for the reaction of
AgF with complex 9 (6 kJ/mol). Furthermore, the calculations
reproduce the trend of lower regioselectivity with additional
substitution on the allyl terminus.50 Therefore, for two different
substrates, the computational data support cationic intermedi-
ate C, over neutral complex B, as the active electrophile.
Computational Studies of Additional Nucleophiles.
The data from the previous section are in reasonable agreement
with silver fluoride 17 as a potential nucleophile. Since a
palladium fluoride (B) could also serve as a nucleophile, we
computed reaction pathways for neutral allylpalladium fluoride
14 as the active nucleophile. To simplify calculations, the
palladium fluoride is modeled with an unsubstituted allyl ligand.
Using the methods described previously, we located
transition states for the reaction of palladium fluoride 14 as
nucleophile with cationic complexes 15 and 16. The energy
diagram for reaction pathways with electrophile 15 is shown in
Figure 3, and the analogous diagram for reactions with
electrophile 16 is presented in the Supporting Information
(Figure S7). The calculated ΔΔG⧧ values for 14 as nucleophile
are in excellent agreement with those obtained experimentally.
For 15, the calculated ΔΔG⧧ value is 8 kJ/mol (the
experimental value is 9 kJ/mol); for 16, calculated and
experimental ΔΔG⧧ values are both 6 kJ/mol.
We also studied the analogous reaction between palladium
fluoride 14 as nucleophile and neutral palladium fluoride 12 as
electrophile, but a transition state for the linear pathway could
not be located. However, the activation barrier for internal
attack was calculated as 127 kJ/mol (Figure S8, Supporting
Information),48 or 25 kJ/mol higher than the analogous
pathway with silver fluoride 17 as nucleophile (Figure 1).
Therefore, the bimetallic reaction with B as nucleophile and
electrophile is unlikely to compete with the other pathways
considered thus far.
Finally, we attempted to model nucleophilic attack by
“naked” fluoride as the counterion to cationic complex C.
Internal ion return has been proposed for cationic inter-
mediates with acetate and carbonate counterions,51 and a
similar process could occur from a tight ion pair with fluoride as
counterion.52 Our calculations on this mechanism were
inconclusive, since the solvent model did not provide verifiable
transition states. Furthermore, noncovalent interactions
between fluoride and nearby C−H bonds likely distort the
calculated energies. Although we cannot conclusively eliminate
the participation of unligated fluoride as nucleophile, the
observed chemoselectivity for C−F bond formation disfavors
this mechanism. For example, Pd-catalyzed allylic fluorinations
of allylic chlorides produce minimal diene (via E2 elimination)
and are tolerant of silyl ethers.22,23 This chemoselectivity does
not coincide with the reactivity of other sources of “naked”
fluoride that are known to be more basic53 and to readily
Computational Investigations of Electrophile Struc-
ture. We began our computational investigations by modeling
the attack of (PPh3)AgF (17) as nucleophile on neutral (12,
13) and cationic (15, 16) allylpalladium complexes with PPh3
as ligand(s) (Chart 2). All complexes were assumed to undergo
Chart 2. Relevant Intermediates for Computational Studies
rapid apparent rotation.47 In reactions with neutral complexes
12 and 13, we assumed that, due to the relative trans influences
of the anionic and phosphine ligands,11a the nucleophile would
preferentially attack the allylpalladium terminus trans to the
phosphine ligand (Figure S3, Supporting Information).46,48
For each electrophile, transition states leading to the linear
and branched products were located. The energy differences
between these transition states (ΔΔG⧧) were compared to
those calculated from the experimental regioselectivities. In a
smaller model system, structures approximating transition
states for both syn and anti allyl isomers were located by
identifying the highest point on a relaxed coordinate scan of the
C−F bond. In these calculations, the estimated activation
barriers for anti complexes were uniformly higher than those
calculated for syn complexes (Figure S4, Supporting
Information).48 Therefore, only syn complexes were considered
for the full system examined in this work.
With neutral complex 12 as the electrophile, the reaction
pathways for internal and terminal attack of nucleophile 17
were computed, and the energy diagram is shown in Figure 1.
Although the relative activation barriers favor production of the
branched allylic fluoride, the calculated ΔΔG⧧ value (24 kJ/
mol) is much higher than the experimental value for reactions
with 8 (9 kJ/mol). For closely related transition states, using a
dispersion-aware functional, we would expect the error in
relative energies to be within a few kJ/mol. The discrepancy
between these experimental and calculated values is outside the
expected margin of error; therefore, this reaction pathway
overestimates the regioselectivity.
We next considered the reaction of nucleophile 17 with
cationic electrophile 15. The full energy diagram is shown in
D
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 1. Energy diagram for the reaction of neutral electrophile 12 and nucleophile 17 (energies in kJ/mol).
deprotect silyl ethers.54 Therefore, the evidence suggests that
“naked” fluoride is unlikely to be the active nucleophile.
Comparing Reaction Pathways for Potential Active
Nucleophiles. The regioselectivities calculated for the attack
of nucleophiles 14 and 17 on cationic electrophiles 15 and 16
are both within the margin of error for the experimental values
(Table 1). However, the energies for palladium fluoride 14 as
nucleophile are in best agreement across the complete set of
calculations. Additionally, the activation barriers for the
branched pathways with 14 as nucleophile are calculated to
be consistently lower (by 2−5 kJ/mol) than those with silver
fluoride 17. Although we cannot definitively exclude 17 as
nucleophile, we propose that B is more likely to be the active
nucleophile in the catalytic reactions (Table 1, entries 4 and 7).
Therefore, the computational data support a homobimetallic
mechanism in which neutral allylpalladium fluoride B acts as a
nucleophile and cationic allylpalladium complex C acts as an
electrophile.
Experimental Investigations of Electrophile Structure.
Our calculations suggest that a cationic complex, with fluoride
as counterion, is the electrophile in palladium-catalyzed allylic
fluorination. Similarly, most Pd-catalyzed allylic substitutions
with outer-sphere nucleophiles proceed via cationic intermedi-
ates with dissociated counterions such as acetate.17,55
Conductivity studies have shown that, even in a polar solvent
such as DMF, ionization of an allylpalladium chloride to a
cationic complex is minimal.42,56 However, cationic allylpalla-
dium complexes with fluoride as counterion (C) have been
proposed in several Pd-catalyzed allylic substitutions. Using
NMR spectroscopy, Togni and co-workers have characterized a
cationic allylpalladium complex resulting from the treatment of
a dimeric allylpalladium chloride complex with a bidentate
ligand and TlF.57 Hazari, Gouverneur, and Brown have also
observed a cationic allylpalladium intermediate by ES-MS upon
treatment of an allylic fluoride with Pd(0) and PPh3.21
Nevertheless, these cationic complexes have only been
shown to be competent intermediates for allylic substitutions
with amines and malonates. We were also unable to observe
cationic complex C by NMR spectroscopy in catalytic or
stoichiometric fluorinations. To experimentally support the
intermediacy of a cationic electrophile in allylic fluorination, we
have found that allylic trifluoroacetates also undergo Pd-
catalyzed allylic fluorination with AgF (Table S1, Supporting
Information).58,48 Like allylic chlorides, allylic trifluoroacetates
are known to react with Pd(0) quantitatively and irreversibly,
but reactions with allylic trifluoroacetates afford cationic
allylpalladium complexes.59,60
To provide experimental evidence against a neutral electro-
phile such as A or B, we prepared crotylpalladium complex 20,
with a chelating o-diphenylphosphinobenzoate ligand, as a
model for A and B that cannot readily generate C (Scheme 7).
The geometry of 20, with the substituted allyl terminus trans to
phosphorus, was assigned in solution by examination of 1H−31P
NMR coupling constants.40 The structure was also confirmed
in the solid state by X-ray crystallography. Figure 4 presents an
overlay of the structures of 8 and 20, and Table 2 presents
selected metrical parameters and spectroscopic data. The
metrical parameters indicate that these complexes are
structurally similar; despite different anionic ligands, the C−C
and C−Pd bond lengths of the crotylpalladium fragment vary
by less than 2% (entries 1−4). The distinct conformations of
E
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 2. Energy diagram for the reaction of cationic electrophile 15 and nucleophile 17 (energies in kJ/mol).
Figure 3. Energy diagram for reaction of cationic electrophile 15 and neutral palladium fluoride 14 as nucleophile (energies in kJ/mol).
the phosphine substituents in each complex do not greatly
Furthermore, 13C NMR chemical shifts suggest that these
affect the bonding of and to the crotyl ligand.
complexes possess similar electronic characteristics. In partic-
F
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
where fluoride attack primarily occurs. Isolated allylpalladium
complexes with this phosphinobenzoate ligand have not been
previously described,61 but a catalyst generated from 18 and
Pd(OAc)2 has been utilized for allylic alkylation.62 In
stoichiometric studies, allylruthenium complexes possessing
this ligand react with amines and malonates,63 and reactions of
untethered allylpalladium carboxylate complexes with both
inner- and outer-sphere nucleophiles are known.64 On the basis
of this precedent, we expect that complex 20 should serve as a
reasonable model for a neutral allylpalladium complex in which
the anionic ligand does not readily dissociate.
Table 1. Summary of Experimental and Theoretical
Regioselectivities
entry
1
R
Z
electrophile
nucleophile
ΔΔG⧧ (kJ/mol)
a
Me
Cl
experimental reaction: 8 +
9
AgF
2
3
4
5
Me
Me
Me
Et
F
12
15
15
17
17
14
24
14
8
PPh3
PPh3
Cl
In the reaction between 20 and AgF, no C−F bond
formation was observed (Scheme 8). The lack of productive
a
experimental reaction: 9 +
6
AgF
a
Scheme 8. Attempted Allylic Fluorination of 20 with AgF
6
7
Et
Et
PPh3
PPh3
16
16
17
14
11
6
a
Determined from the regioselectivity of reactions in Scheme 6.
Scheme 7. Preparation of Crotylpalladium Complex 20
a
additive = none, 18, 19. Reactions were conducted in the presence of
dimethyl fumarate (1 equiv) and were analyzed by GC and 19F NMR.
reactivity with 20 may result from the absence of a ligand such
as PPh3 (in 8) that can dissociate and enhance the solubility of
AgF. To replicate this effect, the reaction of 20 and AgF was
also conducted in the presence of phosphines 18 and 19 as
catalytic additives, but allylic fluorides 10b and 10l were not
produced. C−F bond formation was observed in the presence
of catalytic PPh3, but 31P NMR control experiments
demonstrated that the mixture of PPh3 and 20 produces an
additional complex with no free PPh3 present.48 Therefore, the
reaction with exogenous PPh3 likely occurs via an intermediate
other than 20, and our experimental data are in agreement with
the computational studies indicating that the active electrophile
in Pd-catalyzed allylic fluorination is not a neutral intermediate.
Experimental Investigations of Nucleophile Structure.
To further corroborate the conclusions from our calculations,
we sought experimental evidence for the intermediacy and
viability of a palladium fluoride as a nucleophilic species. As
previously mentioned, experimental difficulties precluded
kinetic studies of this proposal.65 Furthermore, stoichiometric
reactions of AgF and allylpalladium chloride 8 (or an analogous
cationic complex with trifluoroacetate as counterion) were
confounded by ligand equilibria between Pd and Ag species.
Fortunately, Grushin has demonstrated in related work that
palladium fluoride 5 can participate in C−F bond-forming
reactions with dichloromethane.36c Similar Pd(II) and Ni(II)
fluorides are also known to fluorinate alkyl iodides and allyl
bromides.66 We found that fluorination of allylic chloride 21
with palladium fluoride 5 as fluoride source affords 22 with
regioselectivity identical with that in reactions with AgF (Figure
5a). Additionally, 5 does not react with 21 in the absence of a
Pd(0) catalyst, indicating that this reaction proceeds through an
allylpalladium intermediate. As in the reaction with AgF, 5 also
reacts with allylpalladium chloride cis-1 to provide trans-2 with
inversion of configuration (Figure 5b, compare to Scheme 3).
These studies indicate that a palladium fluoride is a
competent fluoride source for allylic fluorination and that it
reacts with selectivity analogous to that with AgF. However,
this outcome does not shed light on the intermediacy and
Figure 4. Overlay of solid-state structures of 8 (black) and 20 (gray)
(H atoms omitted for clarity).
Table 2. Selected Metrical Parameters and Spectroscopic
Data for 8 and 20
entry
parameter
8
20
a
1
2
3
4
5
6
C1−C2 bond
C2−C3 bond
Pd−C1 bond
Pd−C3 bond
1.426
1.389
2.117
2.237
56.59
100.13
1.400
1.388
2.115
2.277
46.04
100.37
a
a
a
b
b
δ (13C NMR) C1
δ (13C NMR) C3
a
b
Bond lengths in Å. Chemical shifts in ppm.
ular, the values for C3 differ by less than 1 ppm (Table 2, entry
6). Although there is an 11 ppm difference between the
chemical shifts for C1 in these complexes (entry 5),34a,40 this
disparity should not greatly affect the electrophilicity of C3,
G
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
the active nucleophile in both reactions is a palladium fluoride.
Although C−F bond formation with stoichiometric AgF as
nucleophile may seem more likely than the reaction of two
catalytic allylpalladium intermediates, AgF is also present in
catalytic quantities as a result of its poor solubility. Along with
our computational results, these experiments provide additional
support for allylpalladium fluoride B as the active nucleophile in
allylic fluorinations with AgF as fluoride source.
Support of a Homobimetallic Mechanism by Ligand
Effects. Whereas we have been unable to rationalize the ligand
effects observed in the regioselective allylic fluorination when
considering a mechanism in which ligated AgF serves as a
nucleophilic species,68 reaction pathways with B as nucleophile
and C as electrophile provide such a rationale. In this scenario,
the same ligand must be capable of supporting both
electrophilic and nucleophilic Pd species. Such a ligand should
therefore possess intermediate electronic properties and should
be capable of generating both monophosphine-ligated and
diphosphine-ligated intermediates. The reactivity observed with
monodentate phosphines such as PPh3 would derive from their
ability to ligate the nucleophile with a single phosphine and the
electrophile with two phosphines. To achieve similar
coordination environments with a bidentate phosphine, ligands
with large bite angles (Xantphos and L2) would be optimal, as
one arm could dissociate to access a monoligated complex.69 A
bidentate ligand with a small bite angle, such as dppe, strongly
favors cationic intermediates via the chelate effect,70 and the
necessary nucleophilic complex would be present in trace
quantities.
Figure 5. Allylic fluorination with 5 as fluoride source in (a) a catalytic,
regioselective reaction and (b) a stoichiometric reaction with inversion
of configuration (reaction conducted in the presence of dimethyl
fumarate, 1 equiv).
reactivity of allylpalladium fluoride B. Although we were unable
to observe B by NMR spectroscopy in these reactions, we
hypothesized that B could be generated by halide exchange
with allylpalladium chloride A. To model this exchange, we
examined reactions of AgF and 5 with arylpalladium chloride
23 (Scheme 9). Both fluoride sources react with 23 to afford
arylpalladium fluoride 24.67,48
Scheme 9. Halide Exchange with Arylpalladium Chloride 23
Proposed Catalytic Cycle and Mechanism for C−F
Bond Formation. Our proposed catalytic cycle for the Pd-
catalyzed fluorination of allylic chlorides with PPh3 as ligand is
shown in Figure 6. Coordination of the Pd(0) catalyst to the
substrate, followed by oxidative addition, affords a neutral
allylpalladium chloride. Halide exchange then produces a
neutral allylpalladium fluoride in equilibrium with a cationic
complex.71 These two intermediates then react to forge the new
This experimental evidence suggests that allylpalladium
fluoride B can be generated from the reaction of allylpalladium
chloride A and AgF. Furthermore, reactions with AgF and
palladium fluoride 5 both occur with identical regioselectivity.
The simplest hypothesis to explain these observations is that
Figure 6. Proposed catalytic cycle for Pd-catalyzed allylic fluorination with C−F bond formation via a homobimetallic pathway.
H
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Figure 7. Transition states for internal (a) and terminal (b) attack of palladium fluoride 14 on cationic electrophile 15 (distances in Å).
C−F bond in a homobimetallic transition state. The resulting
Pd(0) complex dissociates to release the allylic fluoride product
and regenerate the catalyst. Additionally, the cationic Pd(II)
complex remaining from the nucleophilic component re-enters
the catalytic cycle in equilibrium with the neutral allylpalladium
fluoride. With crotyl chloride as substrate, this transformation is
exergonic by approximately 90 kJ/mol from the reactants (AgF
and crotyl chloride) to the products (AgCl and the allylic
fluorides).48 The formation of AgCl and the exchange of a C−
Cl bond for a C−F bond therefore provide the thermodynamic
driving force for this reaction (Figure S11, Supporting
Information).
Rationalization of the Regioselectivity with Calcu-
lated Transition States. Optimized transition states for
reactions of cationic complexes 15 and 16 with nucleophiles 14
and 17 are late and product-like. For example, C−F bond
lengths are 1.816 and 1.782 Å for internal and terminal attack,
respectively, of palladium fluoride 14 on 15 (see reaction
pathways in Figure 3 and structures in Figure 7).72,48 The
geometries of the transition states resemble product-like η2-
olefin complexes, with the olefin significantly rotated away from
a reactant-like η3-allyl complex.73 By the Hammond postulate,
the lateness of the transition states suggest that factors which
influence the relative stability of the products should affect the
relative transition state energies. In fact, the calculated ΔΔG
values of the Pd(0)−product complexes (Figure S9, Supporting
Information;48 8 kJ/mol for 10 and 4 kJ/mol for 11) parallel
the calculated ΔΔG⧧ values for reactions of 14 with 15 (8 kJ/
mol) and 16 (6 kJ/mol).74
fluoride, in comparison with that for other nucleophiles, results
from a later transition state for C−X bond formation.
CONCLUSION
■
Our computational and experimental studies have elucidated
important features in the mechanism of the Pd-catalyzed
fluorination of allylic chlorides with AgF. As in Pd-catalyzed
allylic substitutions with soft nucleophiles, C−F bond
formation occurs by nucleophilic attack on a cationic
allylpalladium intermediate. Halide exchange between an
allylpalladium chloride and AgF generates both a cationic
intermediate and a neutral allylpalladium fluoride. Therefore,
the fluoride source must readily undergo halide exchange to
generate the active nucleophile and electrophile. We propose
that C−F bond formation occurs by nucleophilic attack of a
neutral allylpalladium fluoride on the cationic electrophile.
Since we initially anticipated challenges for C−F bond
formation by inner-sphere reductive elimination, our original
mechanistic proposal was devised to avoid the generation of a
palladium fluoride. Unexpectedly, the mechanistic studies
presented herein demonstrate that generation of a palladium
fluoride is, in fact, necessary to achieve selective C−F bond
formation via outer-sphere nucleophilic attack.
The proposed catalytic cycle (Figure 6) sheds some light on
the questions posed at the outset of this work. In contrast with
other hard nucleophiles, the reactivity of fluoride as an outer-
sphere nucleophile is presumably due to its association with a
transition metal. This fluorinating species is uniquely reactive,
exhibiting little basicity or reactivity with traditionally fluoride-
sensitive functional groups such as silyl ethers and alcohols.22,23
Allylic acetates and carbonates fail as substrates for these
fluorinations, since the allylpalladium complexes formed from
these substrates do not undergo counterion exchange with a
transition-metal fluoride.76 The ligand effects observed in the
regioselective reaction likely result from the requirement for a
ligand that can support both electrophilic (bisphosphine) and
nucleophilic (monophosphine) intermediates. Finally, in
Notably, the transition states we located for C−F bond
formation with 14 and 17 as nucleophiles are later than those
that have been calculated for allylic substitutions with other
nucleophiles. In transition states for the attack of stabilized
carbanions on cationic allylpalladium electrophiles, forming C−
C bonds are at least 2.0 Å. Furthermore, the transition states
are more reactant-like than those we have identified.37,49,75 We
therefore propose that the increased regioselectivity with
I
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
Synthesis and Characterization of Compounds. Compounds
1,22,32c 5,36b 8,40 21,32f and 2382 were prepared according to published
procedures.
comparison to Pd-catalyzed allylic substitutions with other
nucleophiles, the high regioselectivity for fluoride attack
originates from transition states that are later and more
product-like than those for other nucleophiles.
Although the mechanism for C−F bond formation was
studied in the context of a specific reaction, it is possible that
transition-metal fluorides are key intermediates in other
transition-metal-catalyzed allylic fluorinations.25−28 Notably,
Lauer and Wu’s data also suggest that fluoride attacks with
inversion of stereochemistry under their conditions.27 The
reactivity of palladium fluorides elucidated in our work may also
have implications for other Pd(0)-catalyzed C−F bond forming
reactions.
This article has focused on the factors governing reactivity
and regioselectivity in Pd-catalyzed allylic fluorination, but our
conclusions suggest that complex mechanisms of asymmetric
induction may operate in enantioselective allylic fluorinations
with L1 and L2.37 In future work, we intend to study these
asymmetric reactions and apply the unique reactivity of
palladium fluorides in other catalytic fluorinations.
Chloro(triphenylphosphine)[(1,2,3-η)-2-pentenyl]palladium (9).
A flame-dried flask was charged with palladium(II) trifluoroacetate
(500 mg, 1.504 mmol). After the flask was purged under high vacuum
and filled with nitrogen three times, acetone (11 mL) and 1-pentene
(165 μL, 105 mg, 1.50 mmol) were placed in it. The reaction mixture
was stirred at room temperature for 1 h before adding tetra-n-
butylammonium chloride (460 mg, 1.65 mmol) in acetone (4 mL).
After 15 min, the reaction mixture was filtered through Celite and
concentrated under reduced pressure. The residue was purified by
automated column chromatography (50 g silica gel, 0−15% ethyl
acetate in hexanes) to afford bis[chloro((1,2,3)-η-1-pentene)-
palladium] (226 mg, 0.536 mmol, 36% yield) as a yellow solid. A
flame-dried flask was then charged with bis[chloro((1,2,3-η)-1-
pentene)palladium] (100 mg, 0.237 mmol) and triphenylphosphine
(124 mg, 0.474 mmol). After the flask was purged and filled with
nitrogen three times, dichloromethane (5 mL) was added. The
reaction mixture was stirred at room temperature for 20 min before
adding hexanes (40 mL). The mixture was then stored overnight at
−30 °C before concentrating under reduced pressure. The residue was
recrystallized from ethyl acetate at −30 °C to afford the title
1
compound (178 mg, 0.376 mmol, 79% yield) as yellow blocks: H
EXPERIMENTAL SECTION
NMR (500 MHz, CDCl3) δ 7.50−7.72 (m, 6H), 7.34−7.48 (m, 9H),
5.38 (ddd, J = 12.6, 12.6, 6.8 Hz, 1H), 4.53 (dtd, J = 12.6, 8.7, 3.9 Hz,
1H), 2.87 (dd, J = 6.8, 2.0 Hz, 1H), 2.60−2.67 (m, 1H), 2.28−2.47
(m, 1H), 2.11−2.24 (m, 1H), 1.18 (t, J = 7.4 Hz, 3H); 13C NMR (125
MHz, CDCl3) δ 134.15 (d, J = 12.9 Hz), 132.85 (d, J = 41.2 Hz),
130.47 (d, J = 2.2 Hz), 128.68 (d, J = 10.3 Hz), 114.90 (d, J = 4.5 Hz),
106.39 (d, J = 26.5 Hz), 56.79 (d, J = 2.1 Hz), 24.81 (d, J = 3.9 Hz),
13.43 (d, J = 5.8 Hz); 31P NMR (121 MHz, CDCl3) δ 24.11. Anal.
Calcd for C23H24ClPPd: C, 58.37; H, 5.11. Found: C, 58.29; H, 5.13.
Silver 2-Diphenylphosphinobenzoate (19). A flask was charged
with 2-diphenylphosphinobenzoic acid (18; 306 mg, 1.00 mmol),
triethylamine (139 μL, 101 mg, 1.00 mmol), and ethanol (3 mL).
Another flask was charged with silver nitrate (170 mg, 1.00 mmol),
acetonitrile (0.3 mL), and ethanol (3 mL), and this flask was cooled to
0 °C. In the dark, the acid solution was added, dropwise, to the silver
nitrate solution. The reaction mixture was then warmed to room
temperature and stirred at room temperature in the dark for 2 h. The
mixture was then filtered, and the filter cake was washed with cold
ethanol and cold hexanes. The filter cake was then collected and dried
under high vacuum to afford the title compound (346 mg, 0.837
mmol, 84% yield) as a white powder: 1H NMR (500 MHz, CDCl3) δ
8.11−8.26 (m, 1H), 7.52 (t, J = 7.6 Hz, 1H), 7.30−7.48 (m, 11H),
6.97 (t, J = 8.3 Hz, 1H); 13C NMR (125 MHz, CDCl3) δ 134.00 (d, J
= 16.9 Hz), 133.80, 132.88 (d, J = 37.2 Hz), 131.94 (d, J = 10.5 Hz),
131.22, 130.62, 130.41, 130.01, 128.97 (d, J = 10.3 Hz); 31P NMR
(121 MHz, CDCl3) δ 13.44 (d, J = 744.2 Hz).
■
Computational Methods. Calculations were performed with
Gaussian09,38 with geometry optimizations using the B3LYP77
functional and a mixed-basis set. For this set (denoted as BS1), 6-
31G(d) was used on P, C, and H atoms, 6-31+G(d) was used on F
and Cl atoms, and LANL2DZ with accompanying effective core
potential (ECP) was used on Ag and Pd. All calculations, including
geometry optimizations, were performed in toluene using the PCM
solvent model.39 Stationary points were confirmed by obtaining the
correct number of imaginary frequencies. Additionally, transition states
were verified to be on the reaction pathway by performing IRC
calculations or by nudging the reaction coordinate along the imaginary
vibrational mode, followed by energy minimization (QRC).78 Since
B3LYP is known to neglect dispersion forces,79 we calculated single-
point energies for each stationary point using the M06 functional.80
For the single-point electronic energies, a larger mixed-basis set (BS2)
was used, consisting of cc-pVTZ on P, C, and H atoms, AUG-cc-pVTZ
on F and Cl atoms, and SDD with accompanying ECP on Ag and Pd.
Gibbs free energies were obtained by adding these M06 electronic
energies to thermal corrections obtained from frequency calculations
at 298 K (B3LYP/BS1).
The computational methods were validated by comparing the
calculated and experimental structures of allylpalladium chloride
phosphine 8. These structures, overlaid, with selected parameters are
presented in Table S2 (Supporting Information). The calculations
reproduce the geometry of 8, with the selected bond lengths within
5%. Although the calculated structure contains Pd−ligand bond
lengths that are slightly longer than those in the solid-state structure,
the agreement is reasonable, considering that the structure was
calculated in the solution state, while the experimental structure was
obtained in the solid state.
General Experimental Methods. Unless otherwise noted,
reactions were performed without the exclusion of air or moisture.
Reactions using silver(I) fluoride were performed with the exclusion of
light by wrapping reaction vessels in aluminum foil. Reactions were
monitored by thin-layer chromatography (TLC), visualizing with
fluorescence quenching, potassium permanganate (KMnO4), or ceric
ammonium molybdate (CAM). Organic solutions were concentrated
under reduced pressure using a rotary evaporator with an ice−water
bath for volatile compounds. Diethyl ether (Et2O), dichloromethane
(CH2Cl2), tetrahydrofuran (THF), toluene, 1,4-dioxane, and benzene
were dried by passing through activated alumina columns; acetonitrile
(CH3CN), N,N-dimethylformamide (DMF), and pyridine were dried
by passing through a column of activated molecular sieves.81 d8-
Toluene and d6-benzene were dried over activated 4 Å molecular
sieves before use.
[2-(Diphenylphosphino)benzoate-κ2P,O][(1,2,3-η)-2-butenyl]-
palladium (20). An oven-dried flask was charged with crotylpalladium
chloride dimer (118 mg, 0.3 mmol) and silver 2-diphenylphosphino-
benzoate (19; 248 mg, 0.600 mmol). After the flask was purged under
high vacuum and filled with nitrogen three times, dichloromethane (15
mL) was added. After it was stirred at room temperature for 30 min,
the reaction mixture was filtered through a short plug of Celite. After
concentration under reduced pressure, the residue was recrystallized
from dichloromethane/hexanes to afford the title compound (206 mg,
0.441 mmol, 74% yield) as an off-white powder in a 10:1 mixture of
trans and cis isomers (trans isomer: phosphine trans to the substituted
1
allyl terminus): H NMR (500 MHz, CDCl3; cis isomer indicated by
*) δ 8.31−8.36 (m, 1H*), 8.29 (ddd, J = 7.7, 4.5, 1.3 Hz, 1H), 7.26−
7.54 (m, 12H, 12H*), 6.91 (dd, J = 10.1, 7.7 Hz, 1H*), 6.75 (ddd, J =
10.4, 7.8, 1.3 Hz, 1H), 5.52−5.60 (m, 1H*), 5.48 (ddd, J = 12.3, 12.3,
6.8 Hz, 1H), 4.68−4.77 (m, 1H*), 4.67 (ddq, J = 12.8, 9.4, 6.5 Hz,
1H), 3.78 (dd, J = 13.8, 9.5 Hz, 1H*), 3.42−3.54 (m, 1H*), 3.10 (dd, J
= 6.7, 2.6 Hz, 1H), 2.60 (dm, J = 11.7 Hz, 1H), 1.78 (dd, J = 9.0, 6.3
Hz, 3H), 0.86 (dd, J = 8.6, 6.3 Hz, 3H*); 13C NMR (125 MHz,
CDCl3; cis isomer indicated by * when identifiable) δ 171.44 (d, J =
J
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
5.1 Hz), 170.74* (d, J = 5.0 Hz), 144.33 (d, J = 16.1 Hz), 143.52* (d, J
= 15.5 Hz), 133.81 (d, J = 14.0 Hz), 133.73* (d, J = 13.5 Hz), 133.40*
(d, J = 8.9 Hz), 133.35 (d, J = 9.2 Hz), 132.23* (d, J = 2.1 Hz), 132.13
(d, J = 2.3 Hz), 131.08 (d, J = 2.1 Hz), 130.60 (d, J = 13.3 Hz), 129.51
(d, J = 6.7 Hz), 129.17 (d, J = 21.9 Hz), 129.17, 127.23* (d, J = 40.0
Hz), 126.69 (d, J = 40.0 Hz), 120.89* (d, J = 5.6 Hz), 117.69 (d, J =
4.6 Hz), 100.37 (d, J = 26.2 Hz), 78.48* (d, J = 28.2 Hz), 65.82* (d, J
= 5.4 Hz), 46.04 (d, J = 3.7 Hz), 17.59*, 17.32 (d, J = 4.0 Hz); 31P
NMR (203 MHz, CDCl3; cis isomer indicated by *) δ 21.84, 19.18*.
Anal. Calcd for C23H21O2PPd: C, 59.18; H, 4.53. Found: C, 58.25; H,
4.35.
Halide Exchange with Arylpalladium Chloride. To oven-dried
1 dram borosilicate vials, trans-bis(triphenylphosphine)(p-tolyl)palla-
dium chloride (23; 15 mg, 0.02 mmol) and the appropriate fluoride
source (0.02 mmol) were added. Three times, the vials were purged
under high vacuum and filled with nitrogen. Then, the vials were
charged with d6-benzene (1 mL). The mixtures were stirred at 700
rpm, and at various time points aliquots from the reaction mixtures
were filtered through short plugs of Celite, with d6-benzene as eluent.
The reactions were then analyzed by 31P NMR to observe the
formation of trans-bis(triphenylphosphine)(p-tolyl)palladium fluoride
(24).36b
General Procedure for Reactions with Stoichiometric
Allylpalladium Complexes 8, 9, and 20. To oven-dried 1 dram
borosilicate vials, the appropriate palladium complex (8, 9, or 20; 0.05
mmol), dimethyl fumarate (7.2 mg, 0.05 mmol), and silver(I) fluoride
(19 mg, 0.15 mmol) were added (along with an additive, as
appropriate). Three times, the vials were purged under high vacuum
and filled with nitrogen. Then, each vial was charged with toluene (1
mL). After they were stirred at 700 rpm for 24 h (with the vials
covered in aluminum foil), the reaction mixtures were filtered through
short plugs of silica gel, with toluene as eluent. Regioselectivity was
determined by GC using a commercial column, and the reported
regioselectivities are the average of three runs. Spectral data for 10b,l
were in agreement with literature values.83
ASSOCIATED CONTENT
* Supporting Information
Text, figures, tables, and CIF files giving additional
experimental details, X-ray crystallographic data for 8 and 20,
additional energy diagrams, calculated structures and energies,
and the full ref 38. This material is available free of charge via
the Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
*E-mail for A.G.D.: agdoyle@princeton.edu.
Notes
■
( )-3-Fluoropent-1-ene (11b) and (E)-1-fluoropent-2-ene (11l):
1H NMR (500 MHz, CDCl3, linear isomer indicated by *) δ 5.50−
The authors declare no competing financial interest.
5.94 (m, 1H, 2H*), 5.26−5.35 (m, 1H), 5.22 (dt, J = 10.7, 1.4 Hz,
1H), 4.65−4.92 (m, 1H, 2H*), 2.00−2.16 (m, 2H*), 1.61−1.82 (m,
2H), 1.02 (t, J = 7.5 Hz, 3H*), 0.96 (t, J = 7.5 Hz, 3H); 13C NMR
ACKNOWLEDGMENTS
■
3
(125 MHz, CDCl3, linear isomer indicated by *) δ 139.29* (d, J =
We thank Phil Jeffrey for X-ray crystallographic structure
determination and Julia Kalow for helpful discussions.
Calculations were performed at the TIGRESS high perform-
ance computer center at Princeton University, which is jointly
supported by the Princeton Institute for Computational Science
and Engineering and the Princeton University Office of
Information Technology. Financial support provided by
Merck, Princeton University, the NSF (CAREER-1148750),
and a fellowship from Eli Lilly to M.H.K. are gratefully
acknowledged. P.-O.N. acknowledges support from the
Swedish Research Council. A.G.D. is an Alfred P. Sloan
Foundation Fellow, Eli Lilly Grantee, an Amgen Young
Investigator, and a Roche Early Excellence in Chemistry
Awardee.
12.0 Hz), 136.44 (d, 2J = 19.7 Hz), 123.47* (d, 2J = 16.5 Hz), 117.04
(d, 3J = 11.9 Hz), 94.91 (d, 1J = 166.8 Hz), 83.87* (d, 1J = 159.6 Hz),
28.22 (d, 2J = 22.6 Hz), 25.25* (d, 4J = 2.3 Hz), 13.01*, 8.96 (d, 3J =
6.0 Hz); 19F NMR (282 MHz, CDCl3, linear isomer indicated by *) δ
−177.87 (m), −207.24* (m).
Reactions with Palladium Fluoride 5 as Fluoride Source. A
0.05 M stock solution was prepared in toluene (1 mL/reaction),
containing (E)-1-chlorodec-2-ene (21; 5.2 mg, 0.03 mmol/reaction, 1
equiv) and dodecane (2.7 μL/reaction, 40 mol %). An aliquot was
reserved for measurement of the initial ratio by GC. To an oven-dried
1 dram borosilicate vial, tris(dibenzylideneacetone)dipalladium (1.4
mg, 0.0015 mmol), triphenylphosphine (2.4 mg, 0.009 mmol), and
trans-bis(triphenylphosphine)phenylpalladium fluoride (5; 24 mg,
0.033 mmol) were added. Three times, the vial was purged under
high vacuum and filled with nitrogen. Then, the vial was charged with
the stock solution (0.6 mL). After it was stirred at 700 rpm for 48 h
(with the vial covered in aluminum foil), the reaction mixture was
filtered through a short plug of silica gel, with 10% ether in hexanes as
eluent. Conversion, yield, and regioselectivity were determined by GC
using a commercial column (a response factor was calculated using 1H
NMR). Spectral data for 22b,l were in agreement with literature
values.84
REFERENCES
■
(1) (a) Banks, R. E., Smart, B. E., Tatlow, J. C., Eds. Organofluorine
Chemistry: Principles and Commercial Applications; Plenum Press: New
York, 1994. (b) Muller, K.; Faeh, C.; Diederich, F. Science 2007, 317,
̈
1881−1886. (c) Purser, S.; Moore, P. R.; Swallow, S.; Gouverneur, V.
Chem. Soc. Rev. 2008, 37, 320−330. (d) Jeschke, P. ChemBioChem
2004, 5, 570−589. (e) Cai, L.; Lu, S.; Pike, V. W. Eur. J. Org. Chem.
2008, 2853−2873.
For the determination of background reaction between 21 and 5,
the same procedure was followed, omitting tris-
(dibenzylideneacetone)dipalladium and PPh3. No allylic fluorides
were detected by GC.
(2) For the synthesis of benzoyl fluoride from benzoyl chloride and
potassium bifluoride, see: Borodine, A. Justus Liebigs Ann. Chem. 1863,
126, 58−62.
To an oven-dried
1 dram borosilicate vial, chloro-
(triphenylphosphine)[cis-5-carbomethoxy-(1,2,3-η)-cyclohexenyl]-
palladium (1; 14 mg, 0.025 mmol), dimethyl fumarate (3.6 mg, 0.025
mmol), and trans-bis(triphenylphosphine)phenylpalladium fluoride
(5; 36 mg, 0.05 mmol) were added. Three times, the vial was purged
under high vacuum and filled with nitrogen. Then, the vial was charged
with d8-toluene (0.5 mL). After it was stirred at 700 rpm for 24 h (with
the vial covered in aluminum foil), the reaction mixture was charged
with a solution of an internal standard (1,4-difluorobenzene, 40 mol
%) in d8-toluene. The reaction mixture was then filtered through a
short plug of Celite, with d8-toluene as eluent. Diastereoselectivity
(>20:1 dr) and yield (72%) were determined by 19F NMR. Spectral
data for 2 were in agreement with literature values.21
(3) (a) Ma, J.-A.; Cahard, D. Chem. Rev. 2008, 108, PR1−PR43.
(b) Lectard, S.; Hamashima, Y.; Sodeoka, M. Adv. Synth. Catal. 2010,
352, 2708−2732. (c) Furuya, T.; Kamlet, A. S.; Ritter, T. Nature 2011,
473, 470−477. (d) Hollingworth, C.; Gouverneur, V. Chem. Commun.
2012, 48, 2929−2942.
(4) Negishi, E.-i., Ed. Handbook of Organopalladium Chemistry for
Organic Synthesis; Wiley: New York, 2002.
(5) For a review on aryl fluoride synthesis, see: Furuya, T.; Klein, J. E.
M. N.; Ritter, T. Synthesis 2010, 1804−1821.
(6) (a) Hull, K. L.; Anani, W. Q.; Sanford, M. S. J. Am. Chem. Soc.
2006, 128, 7134−7135. (b) Ball, N. D.; Sanford, M. S. J. Am. Chem.
Soc. 2009, 131, 3796−3797.
K
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(7) (a) Wang, X.; Mei, T.-S.; Yu, J.-Q. J. Am. Chem. Soc. 2009, 131,
7520−7521. (b) Chan, K. S. L.; Wasa, M.; Wang, X.; Yu, J.-Q. Angew.
Chem., Int. Ed. 2011, 50, 9081−9084.
(27) Lauer, A. M.; Wu, J. Org. Lett. 2012, 14, 5138−5141.
(28) (a) Zhu, J.; Tsui, G. C.; Lautens, M. Angew. Chem., Int. Ed. 2012,
51, 12353−12356. (b) Zhang, Z.; Wang, F.; Mu, X.; Chen, P.; Liu, G.
Angew. Chem., Int. Ed. 2013, 52, 7549−7553. (c) Braun, M.-G.; Doyle,
A. G. J. Am. Chem. Soc. 2013, 135, 12990−12993. (d) Zhang, Q.;
Nguyen, H. M. Chem. Sci. 2014, 5, 291−296.
(8) (a) Furuya, T.; Kaiser, H. M.; Ritter, T. Angew. Chem., Int. Ed.
2008, 47, 5993−5996. (b) Furuya, T.; Ritter, T. J. Am. Chem. Soc.
2008, 130, 10060−10061. (c) Furuya, T.; Benitez, D.; Tkatchouk, E.;
Strom, A. E.; Tang, P.; Goddard, W. A.; Ritter, T. J. Am. Chem. Soc.
2010, 132, 3793−3807. For related applications of this chemistry to
18F fluorination, see: (d) Lee, E.; Kamlet, A. S.; Powers, D. C.;
Neumann, C. N.; Boursalian, G. B.; Furuya, T.; Choi, D. C.; Hooker, J.
M.; Ritter, T. Science 2011, 334, 639−642. (e) Kamlet, A. S.;
Neumann, C. N.; Lee, E.; Carlin, S. M.; Moseley, C. K.; Stephenson,
N.; Hooker, J. M.; Ritter, T. PLoS ONE 2013, 8, e59187.
(9) For examples of aliphatic C−F bond formation from high-valent
Pd, see: (a) Qiu, S.; Xu, T.; Zhou, J.; Guo, Y.; Liu, G. J. Am. Chem. Soc.
2010, 132, 2856−2857. (b) Qiu, S.; Xu, T.; Zhou, J.; Guo, Y.; Liu, G. J.
Am. Chem. Soc. 2010, 132, 2856−2857. (c) Racowski, J. M.; Gary, J. B.;
Sanford, M. S. Angew. Chem., Int. Ed. 2012, 51, 3414−3417.
(d) McMurtrey, K. B.; Racowski, J. M.; Sanford, M. S. Org. Lett.
2012, 14, 4094−4097. For a related example of Pd-catalyzed vinyl
fluoride synthesis, see: (e) Peng, H.; Liu, G. Org. Lett. 2011, 13, 772−
775.
(29) For mechanistic studies on C−F bond formation from Pd(IV)
intermediates, see refs 6b, 8c, and 9c.
(30) For examples, see: (a) Reference 15b. (b) Liu, W.; Huang, X.;
Cheng, M.-J.; Nielsen, R. J.; Goddard, W. A.; Groves, J. T. Science
2012, 337, 1322−1325. (c) Barthazy, P.; Togni, A.; Mezzetti, A.
Organometallics 2001, 20, 3472−3477.
(31) (a) Keinan, E.; Roth, Z. J. Org. Chem. 1983, 48, 1769−1772.
(b) Fiaud, J. C.; Legros, J. Y. J. Org. Chem. 1987, 52, 1907−1911.
(32) For examples, see: (a) Sheffy, F. K.; Godschalx, J. P.; Stille, J. K.
J. Am. Chem. Soc. 1984, 106, 4833−4840. (b) Goliaszewski, A.;
Schwartz, J. Organometallics 1985, 4, 417−419. (c) Kurosawa, H.;
Kajimaru, H.; Ogoshi, S.; Yoneda, H.; Miki, K.; Kasai, N.; Murai, S.;
Ikeda, I. J. Am. Chem. Soc. 1992, 114, 8417−8424. (d) Matsumoto, Y.;
Ohno, A.; Hayashi, T. Organometallics 1993, 12, 4051−4055.
́
́
(e) Albeniz, A. C.; Espinet, P.; Martın-Ruiz, B. Chem. Eur. J. 2001,
7, 2481−2489. (f) Brozek, L. A.; Ardolino, M. J.; Morken, J. P. J. Am.
Chem. Soc. 2011, 133, 16778−16781.
(10) Furuya, T.; Kuttruff, C. A.; Ritter, T. Curr. Opin. Drug Discovery
Dev. 2008, 11, 803−819.
(11) (a) Hartwig, J. Organotransition Metal Chemistry; University
Science Books: Sausalito, CA, 2010. (b) Johansson Seechurn, C. C. C.;
Kitching, M. O.; Colacot, T. J.; Snieckus, V. Angew. Chem., Int. Ed.
2012, 51, 5062−5085.
(33) These substrates are more common in Cu-catalyzed allylic
substitution: Falciola, C. A.; Alexakis, A. Eur. J. Org. Chem. 2008,
3765−3780.
(34) (a) Åkermark, B.; Zetterberg, K.; Hansson, S.; Krakenberger, B.;
Vitagliano, A. J. Organomet. Chem. 1987, 335, 133−142. (b) Trost, B.
M.; Ariza, X. J. Am. Chem. Soc. 1999, 121, 10727−10737.
(35) (a) Goux, C.; Lhoste, P.; Sinou, D. Synlett 1992, 725−727.
(b) Trost, B. M.; Toste, F. D. J. Am. Chem. Soc. 1999, 121, 4545−
4554. (c) Dubovyk, I.; Watson, I. D. G.; Yudin, A. K. J. Org. Chem.
2013, 78, 1559−1575.
(12) (a) Grushin, V. V. Chem. Eur. J. 2002, 8, 1006−1014.
(b) Grushin, V. V. Acc. Chem. Res. 2010, 43, 160−171.
(13) Yandulov, D. V.; Tran, N. T. J. Am. Chem. Soc. 2007, 129,
1342−1358.
́
(14) (a) Watson, D. A.; Su, M.; Teverovskiy, G.; Zhang, Y.; Garcıa-
(36) (a) Fraser, S. L.; Antipin, M. Y.; Khroustalyov, V. N.; Grushin,
V. V. J. Am. Chem. Soc. 1997, 119, 4769−4770. (b) Pilon, M. C.;
Grushin, V. V. Organometallics 1998, 17, 1774−1781. (c) Grushin, V.
V. Angew. Chem., Int. Ed. 1998, 37, 994−996.
(37) Butts, C. P.; Filali, E.; Lloyd-Jones, G. C.; Norrby, P.-O.; Sale, D.
A.; Schramm, Y. J. Am. Chem. Soc. 2009, 131, 9945−9957.
(38) Frisch, M. J., et al. Gaussian 09, revision C.01; Gaussian, Inc.,
Wallingford, CT, 2010.
Fortanet, J.; Kinzel, T.; Buchwald, S. L. Science 2009, 325, 1661−1664.
(b) Noel, T.; Maimone, T. J.; Buchwald, S. L. Angew. Chem., Int. Ed.
̈
̈
2011, 50, 8900−8903. (c) Wannberg, J.; Wallinder, C.; Unlusoy, M.;
̈
Skold, C.; Larhed, M. J. Org. Chem. 2013, 78, 4184−4189. (d) Lee, H.
̈
G.; Milner, P. J.; Buchwald, S. L. Org. Lett. 2013, 15, 5602−5605.
(15) (a) Kalow, J. A.; Doyle, A. G. J. Am. Chem. Soc. 2010, 132,
3268−3269. (b) Kalow, J. A.; Doyle, A. G. J. Am. Chem. Soc. 2011, 133,
16001−16012. (c) Shaw, T. W.; Kalow, J. A.; Doyle, A. G. Org. Synth.
2012, 89, 9−18. (d) Kalow, J. A.; Schmitt, D. E.; Doyle, A. G. J. Org.
Chem. 2012, 77, 4177−4183. (e) Kalow, J. A.; Doyle, A. G.
Tetrahedron 2013, 69, 5702−5709.
(39) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999−3094.
(40) Åkermark, B.; Krakenberger, B.; Hansson, S.; Vitagliano, A.
Organometallics 1987, 6, 620−628.
(16) Pacheco, M. C.; Purser, S.; Gouverneur, V. Chem. Rev. 2008,
108, 1943−1981.
(41) In the catalytic reaction with (E)-crotyl chloride as substrate, an
18:1 b:l ratio was obtained.
(17) (a) Trost, B. M.; Van Vranken, D. L. Chem. Rev. 1996, 96, 395−
422. (b) Lu, Z.; Ma, S. Angew. Chem., Int. Ed. 2008, 47, 258−297.
(c) Trost, B. M.; Zhang, T.; Sieber, J. D. Chem. Sci. 2010, 1, 427−440.
(18) Hagelin, H.; Åkermark, B.; Norrby, P.-O. Chem. Eur. J. 1999, 5,
902−909.
(42) (a) Åkermark, B.; Åkermark, G.; Hegedus, L. S.; Zetterberg, K. J.
́
Am. Chem. Soc. 1981, 103, 3037−3040. (b) Cantat, T.; Genin, E.;
Giroud, C.; Meyer, G.; Jutand, A. J. Organomet. Chem. 2003, 687,
365−376.
(43) 19F NMR studies indicated that the solubility of AgF in toluene
is in the picomolar to nanomolar range.
(19) Hintermann, L.; Lang, F.; Maire, P.; Togni, A. Eur. J. Inorg.
̈
Chem. 2006, 1397−1412.
(44) Although we did not observe 31P NMR evidence for PPh3−Ag
binding, the NMR signals of a related complex are broad, due to rapid
(20) Clark, J. H. Chem. Rev. 1980, 80, 429−452.
(21) Hazari, A.; Gouverneur, V.; Brown, J. M. Angew. Chem., Int. Ed.
2009, 48, 1296−1299.
́
ligand exchange: Letinois-Halbes, U.; Pale, P.; Berger, S. Magn. Reson.
Chem. 2004, 42, 831−834.
(22) Katcher, M. H.; Doyle, A. G. J. Am. Chem. Soc. 2010, 132,
17402−17404.
(45) (a) Yanagisawa, A.; Kageyama, H.; Nakatsuka, Y.; Asakawa, K.;
Matsumoto, Y.; Yamamoto, H. Angew. Chem., Int. Ed. 1999, 38, 3701−
3703. (b) Wadamoto, M.; Yamamoto, H. J. Am. Chem. Soc. 2005, 127,
14556−14557.
(23) Katcher, M. H.; Sha, A.; Doyle, A. G. J. Am. Chem. Soc. 2011,
133, 15902−15905.
(24) Braun, M.-G.; Katcher, M. H.; Doyle, A. G. Chem. Sci. 2013, 4,
1216−1220.
(46) Svensen, N.; Fristrup, P.; Tanner, D.; Norrby, P.-O. Adv. Synth.
Catal. 2007, 349, 2631−2640.
(25) (a) Hollingworth, C.; Hazari, A.; Hopkinson, M. N.; Tredwell,
M.; Benedetto, E.; Huiban, M.; Gee, A. D.; Brown, J. M.; Gouverneur,
V. Angew. Chem., Int. Ed. 2011, 50, 2613−2617. (b) E. Benedetto, E.;
Tredwell, M.; Hollingworth, C.; Khotavivattana, T.; Brown, J. M.;
Gouverneur, V. Chem. Sci. 2013, 4, 89−96.
(47) Hansson, S.; Norrby, P. O.; Sjogren, M. P. T.; Åkermark, B.;
̈
Cucciolito, M. E.; Giordano, F.; Vitagliano, A. Organometallics 1993,
12, 4940−4948.
(48) See the Supporting Information for details.
(49) Fristrup, P.; Ahlquist, M.; Tanner, D.; Norrby, P.-O. J. Phys.
Chem. A 2008, 112, 12862−12867.
(26) Topczewski, J. J.; Tewson, T. J.; Nguyen, H. M. J. Am. Chem.
Soc. 2011, 133, 19318−19321.
L
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX

Organometallics
Article
(50) Notably, 16 has several more accessible conformations than 15
(Figure S5, Supporting Information); ΔΔG⧧ was calculated for the
transition states with the lowest energies of all reaction pathways
examined. Although a transition state was found for internal attack of
silver fluoride 17 on the neutral electrophile 13, a transition state for
the analogous terminal attack could not be located.
(51) Evans, L. A.; Fey, N.; Harvey, J. N.; Hose, D.; Lloyd-Jones, G.
C.; Murray, P.; Orpen, A. G.; Osborne, R.; Owen-Smith, G. J. J.;
Purdie, M. J. Am. Chem. Soc. 2008, 130, 14471−14473.
(65) Like AgF, palladium complexes 5 and 8 are also sparingly
soluble in benzene or toluene.
(66) (a) Breyer, D.; Braun, T.; Klaring, P. Organometallics 2012, 31,
̈
́
1417−1424. (b) Martınez-Prieto, L. M.; Melero, C.; del Río, D.;
́
́
pora, J.; Alvarez, E. Organometallics 2012, 31, 1425−
Palma, P.; Cam
1438.
(67) The reaction of 23 and 5 affords 24 within minutes, while hours
are required for the production of 24 from AgF and 23. The greater
solubility of 5 may account for the rate difference.
(68) Although silver is known to adopt varying coordination modes,
it is not clear how these coordination environments would affect the
nucleophilicity of AgF. In contrast, we would expect the nucleophil-
icity of a palladium fluoride to vary in neutral and cationic
intermediates.
(52) Ion pairs with anions as outer-sphere nucleophiles have also
been invoked in the asymmetric allylic alkylation of enol carbonates:
Trost, B. M.; Xu, J.; Schmidt, T. J. Am. Chem. Soc. 2009, 131, 18343−
18357.
(53) (a) Christe, K. O.; Wilson, W. W.; Wilson, R. D.; Bau, R.; Feng,
(69) Cationic allylpalladium complexes of L1 have been shown to
equilibrate between a variety of coordination modes. See: Eastoe, J.;
J. A. J. Am. Chem. Soc. 1990, 112, 7619−7625. (b) Schwesinger, R.;
Link, R.; Thiele, G.; Rotter, H.; Honert, D.; Limbach, H.-H.; Mannle,
̈
́ ́
Fairlamb, I. J. S.; Fernandez-Hernandez, J. M.; Filali, E.; Jeffery, J. C.;
F. Angew. Chem., Int. Ed. Engl. 1991, 30, 1372−1375.
Lloyd-Jones, G. C.; Martorell, A.; Meadowcroft, A.; Norrby, P.-O.;
Riis-Johannessen, T.; Sale, D. A.; Tomlin, P. M. Faraday Discuss. 2010,
145, 27−47.
(54) (a) Holton, R. A.; Kim, H. B.; Somoza, C.; Liang, F.; Biediger,
R. J.; Boatman, P. D.; Shindo, M.; Smith, C. C.; Kim, S. J. Am. Chem.
Soc. 1994, 116, 1599−1600. (b) Wender, P. A.; Badham, N. F.;
Conway, S. P.; Floreancig, P. E.; Glass, T. E.; Houze, J. B.; Krauss, N.
E.; Lee, D.; Marquess, D. G.; McGrane, P. L.; Meng, W.; Natchus, M.
G.; Shuker, A. J.; Sutton, J. C.; Taylor, R. E. J. Am. Chem. Soc. 1997,
119, 2757−2758. (c) Carpenter, J.; Northrup, A. B.; Chung, D. M.;
Wiener, J. J. M.; Kim, S.-G.; MacMillan, D. W. C. Angew. Chem., Int.
Ed. 2008, 47, 3568−3572.
(70) (a) Amatore, C.; Jutand, A.; Meyer, G.; Mottier, L. Chem. Eur. J.
1999, 5, 466−473. (b) Amatore, C.; Gamez, S.; Jutand, A. Chem. Eur.
J. 2001, 7, 1273−1280.
(71) Calculations indicate that halide exchange is exergonic by 3 kJ/
mol. See Figure S10 (Supporting Information).
(72) Calculated C−F bond lengths in the product-coordinated Pd(0)
complexes are 1.45−1.50 Å.
(55) (a) Åkermark, B.; Hansson, S.; Krakenberger, B.; Vitagliano, A.;
Zetterberg, K. Organometallics 1984, 3, 679−682. (b) Amatore, C.;
Gamez, S.; Jutand, A.; Meyer, G.; Mottier, L. Electrochim. Acta 2001,
46, 3237−3244.
(73) The dihedral angle between the allyl or olefin plane and the
Pd−P plane is a quantitative measure of the position of the transition
state relative to the reactant or product. These values are presented in
Table S3 (Supporting Information). Also see: Oslob, J. D.; Åkermark,
B.; Helquist, P.; Norrby, P.-O. Organometallics 1997, 16, 3015−3021.
(74) In other Pd-catalyzed allylic alkylations, enantioselectivities of
reactions with late transition states have been rationalized by
calculating relative product energies: (a) Svensson, M.; Bremberg,
(56) In this work, we assume that the active electrophile is an η3-allyl
rather than an η1-allyl palladium complex. This assumption is
supported by other studies indicating that η1-allyl palladium
intermediates are usually nucleophilic. See: (a) Nakamura, H.; Bao,
M.; Yamamoto, Y. Angew. Chem., Int. Ed. 2001, 40, 3208−3210.
U.; Hallman, K.; Csoregh, I.; Moberg, C. Organometallics 1999, 18,
̈
(b) Solin, N.; Kjellgren, J.; Szabo,
́
K. J. J. Am. Chem. Soc. 2004, 126,
4900−4907. (b) Hallman, K.; Frolander, A.; Wondimagegn, T.;
̈
́
7026−7033. (c) Garcıa-Iglesias, M.; Bunuel, E.; Cardenas, D. J.
́
̃
Svensson, M.; Moberg, C. Proc. Natl. Acad. Sci. U.S.A. 2004, 101,
5400−5404.
(75) Kleimark, J.; Johansson, C.; Olsson, S.; Hakansson, M.;
Hansson, S.; Åkermark, B.; Norrby, P.-O. Organometallics 2011, 30,
230−238.
(76) Pd(0)-catalyzed reactions of the acetate and carbonate
analogues of 21 afford <5% yields of the allylic fluoride products
when 5 is used as a fluoride source.
Organometallics 2006, 25, 3611−3618.
(57) Burckhardt, U.; Baumann, M.; Togni, A. Tetrahedron:
Asymmetry 1997, 8, 155−159.
̊
(58) For examples of Pd(0)-catalyzed allylic substitutions of allylic
trifluoroacetates, see: (a) Obora, Y.; Ogawa, Y.; Imai, Y.; Kawamura,
T.; Tsuji, Y. J. Am. Chem. Soc. 2001, 123, 10489−10493. (b) Obora, Y.;
Nakanishi, M.; Tokunaga, M.; Tsuji, Y. J. Org. Chem. 2002, 67, 5835−
5837. (c) Grote, R. E.; Jarvo, E. R. Org. Lett. 2008, 11, 485−488.
(59) Vitagliano, A.; Åkermark, B.; Hansson, S. Organometallics 1991,
10, 2592−2599.
(60) In contrast, oxidative addition of Pd(0) to allylic acetates and
carbonates occurs reversibly to establish an equilibrium between the
allylpalladium intermediate and the reactant. See: (a) Reference 55.
(77) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 5648−5652. (b) Becke,
A. D. J. Chem. Phys. 1993, 98, 1372−1377. (c) Lee, C.; Yang, W.; Parr,
R. G. Phys. Rev. B 1988, 37, 785−789.
(78) Goodman, J. M.; Silva, M. A. Tetrahedron Lett. 2003, 44, 8233−
8236.
(79) Kruse, H.; Goerigk, L.; Grimme, S. J. Org. Chem. 2012, 77,
10824−10834.
(b) Amatore, C.; Gamez, S.; Jutand, A.; Meyer, G.; Moreno-Manas,
̃
M.; Morral, L.; Pleixats, R. Chem. Eur. J. 2000, 6, 3372−3376.
(61) For the synthesis and stoichiometric reactions of an
allylpalladium complex with a similar phosphinocarboxylate ligand,
(80) Zhao, Y.; Truhlar, D. G. Acc. Chem. Res. 2008, 41, 157−167.
(81) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518−1520.
see: Knuhl, G.; Sennhenn, P.; Helmchen, G. J. Chem. Soc., Chem.
̈
̈
(82) Herrmann, W. A.; Broßmer, C.; Priermeier, T.; Ofele, K. J.
Commun. 1995, 1845−1846.
Organomet. Chem. 1994, 481, 97−108.
(62) Kawatsura, M.; Ikeda, D.; Komatsu, Y.; Mitani, K.; Tanaka, T.;
Uenishi, J. Tetrahedron 2007, 63, 8815−8824.
́ ́
(83) Boukerb, A.; Gree, D.; Laabassi, M.; Gree, R. J. Fluorine Chem.
1998, 88, 23−27.
(63) (a) Kawatsura, M.; Ata, F.; Wada, S.; Hayase, S.; Uno, H.; Itoh,
T. Chem. Commun. 2007, 298−300. (b) Kawatsura, M.; Ata, F.;
Hirakawa, T.; Hayase, S.; Itoh, T. Tetrahedron Lett. 2008, 49, 4873−
4875. (c) Sundararaju, B.; Achard, M.; Demerseman, B.; Toupet, L.;
Sharma, G. V. M; Bruneau, C. Angew. Chem., Int. Ed. 2010, 49, 2782−
2785.
(84) Umemoto, T.; Singh, R. P.; Xu, Y.; Saito, N. J. Am. Chem. Soc.
2010, 132, 18199−18205.
(64) (a) Yamamoto, T.; Saito, O.; Yamamoto, A. J. Am. Chem. Soc.
1981, 103, 5600−5602. (b) Tsuji, Y.; Kusui, T.; Kojima, T.; Sugiura,
Y.; Yamada, N.; Tanaka, S.; Ebihara, M.; Kawamura, T. Organometallics
1998, 17, 4835−4841. (c) Chang, C.-W.; Norsikian, S.; Guillot, R.;
Beau, J.-M. Eur. J. Org. Chem. 2010, 2010, 2280−2294.
M
dx.doi.org/10.1021/om401240p | Organometallics XXXX, XXX, XXX−XXX