Reactions of free radicals with η3-allylpalladium(II) complexes: Phenyl and trityl radicals

Article abstract of DOI:10.1039/b307308d

The compounds (η³-allyl)PdCl(PPh₃) and [(η³-allyl)Pd(PPh₃)₂]Cl react with phenyl and trityl radicals generated from the thermal decomposition of phenylazotriphenylmethane (PhN=NCPh₃, PAT) in benzene at 60°C. The products are the palladium phenyl compounds, [PdPhCl(PPh₃)]₂ and trans-PdPhCl(PPh₃)₂, respectively, and 4,4,4-triphenyl-1-butene, the latter being the result of coupling of the trityl radical with the allyl ligands. In contrast, [(η³-allyl)PdCl]₂ reacts with phenyl and trityl radicals under the same conditions to form palladium metal, trityl chloride and 3-phenylpropene, which is subsequently catalytically isomerized to 1-phenylpropene. These disparate results are interpreted in all cases in terms of initial attack by phenyl radicals on the palladium(II) to give phenyl-palladium(III) intermediates, and it is the secondary reactions, influenced by the presence or absence of coordinated PPh₃ ligands, which provide variety in the products. The reaction of [(4-methoxy-1,3-η³-cyclohexenyl)PdCl]₂ gives a mixture of trans-3-methoxy-6-phenylcyclohexene and trans-4-methoxy-3-phenylcyclohexene, consistent with initial formation of a phenyl-palladium(III) intermediate followed by phenyl migration to the η³-cyclohexenyl ligand (reductive elimination).

Full text of DOI:10.1039/b307308d

View Article Online / Journal Homepage / Table of Contents for this issue
Reactions of free radicals with ꢀ3-allylpalladium(II) complexes:
phenyl and trityl radicals
Simon J. Reid and Michael C. Baird*
Department of Chemistry, Queen’s University, Kingston, ON, Canada K7L 3N6.
E-mail: bairdmc@chem.queensu.ca
Received 26th June 2003, Accepted 27th August 2003
First published as an Advance Article on the web 9th September 2003
The compounds (η3-allyl)PdCl(PPh3) and [(η3-allyl)Pd(PPh3)2]Cl react with phenyl and trityl radicals generated from
the thermal decomposition of phenylazotriphenylmethane (PhN᎐NCPh , PAT) in benzene at 60 ЊC. The products are
᎐
3
the palladium phenyl compounds, [PdPhCl(PPh3)]2 and trans-PdPhCl(PPh3)2, respectively, and 4,4,4-triphenyl-1-
butene, the latter being the result of coupling of the trityl radical with the allyl ligands. In contrast, [(η3-allyl)PdCl]2
reacts with phenyl and trityl radicals under the same conditions to form palladium metal, trityl chloride and
3-phenylpropene, which is subsequently catalytically isomerized to 1-phenylpropene. These disparate results are
interpreted in all cases in terms of initial attack by phenyl radicals on the palladium() to give phenyl–palladium()
intermediates, and it is the secondary reactions, influenced by the presence or absence of coordinated PPh3 ligands,
which provide variety in the products. The reaction of [(4-methoxy-1,3-η3-cyclohexenyl)PdCl]2 gives a mixture of
trans-3-methoxy-6-phenylcyclohexene and trans-4-methoxy-3-phenylcyclohexene, consistent with initial formation of
a phenyl–palladium() intermediate followed by phenyl migration to the η3-cyclohexenyl ligand (reductive
elimination).
ally useful reactions,12 and neutral compounds of the types
Introduction
[(η3-allyl)PdCl]2 (A) and (η3-allyl)PdCl(PPh3) (B) and cationic
Free radical additions to carbon-carbon double and triple
bonds provide a range of well developed and synthetically use-
complexes of the type [(η3-allyl)Pd(PPh3)2]ϩ (C) (all shown in
Fig. 2) have been studied intensively to date. While the two
types of neutral compounds A and B exhibit limited activity
ful organic transformations, and the subject has been reviewed
extensively.1 Free radicals have also been invoked as reactive
towards nucleophiles, cationic complexes of type C react readily
intermediates in a wide variety of reactions involving transition
as in eqn. (1) (NuϪ = various C-, O- and N-based nucleophiles).
metal compounds, e.g. oxidative addition reactions,2 SmI2
induced coupling of organic halides with carbonyl com-
[(η3-allyl)Pd(PPh3)2]ϩ ϩ NuϪ
pounds,3 reduction of organic halides by complexes of chrom-
allyl–Nu ϩ “Pd(PPh3)2” (1)
ium(),4 and hydrogenation and hydrometallation of aromatic
alkenes5 and conjugated dienes.6 Complementing this activity,
recent years have also seen increasing interest in addition reac-
tions of free radicals on complexes of a variety of unsaturated,
coordinated organic compounds such as arene,7 allylic,8 carb-
enoid9 and alkenyne10 ligands. Examples of such radical addi-
Fig. 2 Compounds present on the addition of triphenylphosphine to
[(η3-allyl)PdCl]2 (A).
tion reactions are shown in Fig. 1, where other steps in what
are all complex processes are omitted. The field of free radical
additions to organometallic compounds has been reviewed,11
and it is clear that there is as yet really no general knowl-
edge or understanding of the reactions of free radicals with
coordinated ligands.
Nucleophilic attack on complexes of type C has been thor-
oughly investigated and results in a range of extremely useful in
synthetic methodologies.12 Although there is a plethora of
nucleophiles that can be added to allyl–palladium complexes as
in eqn. (1),12 in fact all such reactions proceed via one of two
mechanisms. Of relevance here, there are those which involve
direct addition of the nucleophile to the η3-allyl ligand (exo
attack, Fig. 3(a)) and those which involve initial attack at the
palladium followed by migration of the coordinated nucleo-
phile to the η3-allyl ligand (endo attack, Fig. 3(b)).12
In view of the wide range of chemistry exhibited in reactions
of free radicals with unsaturated molecules, both free1 and co-
ordinated,11 coupled with the above mentioned, intense interest
in the synthetic utilization of allyl–palladium complexes, we
have begun an investigation of reactions of organic free radicals
with neutral and cationic allylic palladium compounds of types
A, B and C. It was anticipated that interesting and useful coup-
ling/addition reactions of free radicals with allylic ligands
would be observed, and that products not available by con-
ventional palladium–allyl chemistry might be formed.
Fig.
1
Addition reactions of free radicals with coordinated
unsaturated ligands.
We wished first to ascertain if free radicals would react at all
with the unsaturated allylic ligands present in these compounds
and, if so, what would be the products and the effects of the
various ligand environments and the positive charge. We have
η3-Allyllic palladium() complexes have been employed
extensively in synthetic organic chemistry. This is in large part
because of their ability to participate in a variety of synthetic-
begun our studies utilizing phenylazotriphenylmethane (PhN᎐
᎐
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
D a l t o n T r a n s . , 2 0 0 3 , 3 9 7 5 – 3 9 8 0
3975

View Article Online
Except as noted below, all chemicals were purchased from
Aldrich and were used without further purification. Deuterated
solvents were purchased from MSD isotopes and Cambridge
Isotopes Limited, while PdCl2 was obtained on loan from
Johnsonn Mathey. The following complexes were synthesized
via standard literature preparations: Pd(PPh3)4,16 [PdPh(PPh3)-
(µ-OAc)]2,17
[(η3-allyl)PdCl]2,18
[(4-methoxy-1,3-η3-cyclo-
hexenyl)PdCl]2,19 (η3-allyl)PdCl(PPh3),18d [(η3-allyl)Pt(PPh3)2]-
Cl,20 phenylazotriphenylmethane (PAT)13 and 1-diphenyl-
methylene-4-triphenylmethyl-2,5-cyclohexadiene (trityl dimer).14
PPh3 was recrystallized from methanol.
Reactions of (ꢀ3-allyl)PdCl(PPh3) and [(ꢀ3-allyl)Pd(PPh3)2]Cl
with PAT
In a typical procedure, a solution of 0.200 g [(η3-allyl)PdCl]2
(5.50 × 10Ϫ4 mol), 0.575 g PPh3 (2.2 × 10Ϫ3 mol), and 0.770 g
PAT (2.2 × 10Ϫ3 mol) (ratio Pd : PPh3 : PAT = 1 : 2 : 2) in 50 mL
benzene was stirred at 60 ЊC for 1.5 h, during which time the
solution changed from yellow to orange and a small amount of
a black solid formed. The solvent was removed under reduced
pressure to leave a gummy orange residue which was extracted
with pentane until the extracts were colourless (5 × 10 mL).
There resulted a pale yellow solution and a yellow–white resi-
due. The solid material was collected and recrystallized from di-
chloromethane–pentane to give 0.261 g trans-PdPhCl(PPh3)2
Fig. 3 (a) Exo nucleophilic attack on an allyl group; (b) endo
nucleophilic attack on an allyl group (L = PPh3).
NCPh3: henceforth PAT) as a very convenient source of phenyl
and triphenylmethyl (henceforth trityl) radicals.13 PAT
decomposes in inert solvents on heating by releasing phenyl and
trityl radicals in addition to molecular nitrogen (eqn. (2)).13
ؒ
ؒ
PhN᎐NCPh
Ph ϩ Ph3C ϩ N2
(2)
᎐
3
1
3
This reaction is known to be a two-step process which
involves initial homolytic dissociation of the weaker N–CPh3
(32% yield). H NMR (CDCl3): δ 6.23 (t, JH–H = 7.4 Hz, 2H,
3
m-H of Pd–Ph), 6.36 (t, JH–H = 7.2 Hz, 1H, p-H of Pd–Ph),
3
bond to produce a trityl radical and a phenyldiazenyl radical,
6.62 (d, JH–H = 7.6 Hz, 2H, o-H of Pd–Ph), 7.2–7.6 (m, 30H,
ؒ
PPh3). 13C{1H} NMR (CDCl3): δ 121.7 (s), 127.6 (s), 127.84
(vt), 129.7(s), 131.2 (t), 134.6 (vt), 136.4 (t), 154.0 (t). 31P NMR
(CDCl3): δ 24.2 (s). The 1H NMR data are identical to those of
a sample of the compound synthesized as in the literature.21
The solvent was removed from the pentane extract to give a
yellow, oily residue. From this 0.10 g 4,4,4-triphenyl-1-butene
(65% yield) was separated from byproducts arising from the
decomposition of PAT by chromatography on an alumina col-
PhN᎐N . The carbon–nitrogen bond of the latter then under-
᎐
goes homolysis to result in a phenyl radical and molecular
nitrogen. The phenyl radical is highly reactive but the trityl
radical is persistent and remains in solution in equilibrium with
its head-to-tail dimer, 1-diphenylmethylene-4-triphenylmethyl-
2,5-cyclohexadiene.14 Therefore, while it may appear that any
reactions involving PAT would involve predominantly the
initially formed trityl radical, the fact that the latter exists in
solution in equilibrium with its dimer means that radical
reactions derived from PAT may involve reactions with either
the phenyl or the trityl radical or both.
We now report in detail on the reactions of allylic palladium
compounds of types A, B and C with phenyl and trityl radicals.
We have earlier communicated some of our initial finding on
this and related chemistry;15 a subsequent paper will discuss
reactions of these same types of allyl–palladium compounds
with the cyclohexyl radical.
umn using a 2 : 1 pentane–benzene solution as eluent. 1H NMR
3
(C D ): δ 5.72 (ddt, –CH CH᎐CH , J
= 6.4, 19.6, 8.8 Hz),
᎐
5.04 (dq, –CH CH᎐CH(anti), JH–H H=–H 19.6, 1.4, 1.4 Hz),
6
6
2
2
3
᎐
2
3
4.97(dq, –CH CH᎐CH(syn), J
= 8.8, 1.4, 1.4 Hz), 3.32 (dt,
᎐
2
H–H
3
1
CH2 JH–H = 6.4, 1.4 Hz), 7.0–7.4 (m, 30H, Ph). The H NMR
data are identical to those of a sample of the compound
synthesized as in the literature.22
In a complementary experiment, an NMR-scale reaction of
[(η3-allyl)PdCl]2 with two equivalents each of PAT and PPh3 per
1
palladium was run in C6D6 at 60 ЊC. Monitoring by H and
31P NMR spectroscopy showed that trans-PdPhCl(PPh3)2 and
4,4,4-triphenyl-1-butene were formed essentially quantitatively
and that the reaction was complete within 1.5 h.
Experimental
Experiments were conducted under an inert atmosphere of
oxygen-free N2 or Ar, further purified by passing through a
column of BASF catalyst heated to 140 ЊC and a column of
5A molecular sieves. Manipulations of air-sensitive materials
employed standard Schlenk-line techniques, a Vacuum Atmos-
phere glovebox and an Mbraun Labmaster glovebox. Solvents
were dried and distilled over sodium metal except dichloro-
methane, which was dried over calcium hydride, and methanol,
which was dried over magnesium methoxide; all were thor-
oughly deoxygenated prior to use by freeze–thaw–pump cycles
or by saturation with N2 or Ar.
In a reaction in which [(η3-allyl)PdCl]2 was treated with two
equivalents of PAT and one equivalent of PPh3 per Pd, 4,4,4-
triphenyl-1-butene and [PdPh(PPh3)(µ-Cl)]2 (53% yield) were
obtained. The latter is a white compound which was identified by
preparing it as in the literature23 as well as by conversion to trans-
1
PdPhCl(PPh3)2 by reaction with PPh3 in benzene. H NMR of
[PdPh(PPh3)(µ-Cl)]2 (CDCl3): δ 7.51–7.17 (m, 15H, H (Ph)), 6.9
(br s, 2H), 6.6 (br s, 3H). 31P NMR (CDCl3): δ 31.9 (br s).23
Complementing this,
a solution of 0.15 g
(η3-allyl)-
PdCl(PPh3) (3.37 × 10Ϫ4 mol) and 0.236 g PAT (6.78 × 10Ϫ4
mol) in 0.75 mL benzene was heated at 60 ЊC for 1.5 h, and the
solvent was removed in vacuo to give an orange residue. The
residue was extracted with pentanes, giving an orange solution
and leaving a pale yellow solid. The solvent was removed from
NMR spectra were acquired on Bruker ACF 200, Avance
300, AM 400 or Avance 500 NMR spectrometers. The residual
proton and the carbon resonances of deuterated solvents served
1
as internal references for H and 13C resonances, respectively,
while 31P resonances were referenced externally to 85% H3PO4.
Mass spectra were obtained on a Quatro Fisons Pro Quad-
ropole mass spectrometer in EIϩ mode. GC-MS experiments
were carried out on a Fisons GC 8000 coupled to a Fisons MD
800 spectrometer operating in EIϩ mode or a Varian GC-MS
(CP-3800 GC and Saturn 2000 MS).
the extract and the H NMR spectrum showed the only allyl
1
product was 4,4,4-triphenyl-1-butene. The pale yellow solid left
after the extraction was collected and washed thoroughly with
benzene. It was identified as [PdPh(PPh3)(µ-Cl)]2 (40% yield)
by comparison to an authentic sample and by conversion to
trans-PdPhCl(PPh3)2.
D a l t o n T r a n s . , 2 0 0 3 , 3 9 7 5 – 3 9 8 0
3976

View Article Online
Reactions of [(ꢀ3-allyl)PdCl]2 with PAT
was extracted with pentane to give 4,4,4-triphenyl-1-butene,
1
identified by H NMR spectroscopy. The less soluble product
In an NMR-scale reaction, [(η3-allyl)PdCl]2 was treated at 60 ЊC
in C6D6 in an NMR tube with one equivalent of PAT in the
absence of PPh3. A small amount of a black precipitate
appeared within 5 min and a metallic mirror had formed on the
1
was shown by H, 13C and 31P NMR spectroscopy (CDCl3) to
be the known compound trans-PtPhCl(PPh3)2 (0.15 g, 94%
yield).26 A complementary NMR-scale reaction in C6D6 at
60 ЊC showed that trans-PtPhCl(PPh3)2 and 4,4,4-triphenyl-1-
butene were formed quantitatively.
1
side of the NMR tube within 2 h. The H NMR spectrum of
the reaction mixture was broad as a result, and the sample was
filtered to give a clear, orange solution. The 1H NMR spectrum
of the solution exhibited the resonances of 1-phenylpropene
(∼80% of the total allyl products), 3-phenylpropene (∼10% of
the total allyl products), and 4,4,4-triphenyl-1-butene (∼10% of
Reaction of [(ꢀ3-allyl)PdCl]2 with trityl dimer
A solution of 0.106 g [(η3-allyl)PdCl]2 (2.90 × 10Ϫ4 mol), 0.289 g
PPh3 (1.10 × 10Ϫ3 mol) and 1.40 g trityl dimer (2.90 × 10Ϫ3 mol)
in 50 mL C6H6 was refluxed for 1.5 h. The solvent was removed
and a 1H NMR spectrum showed that 4,4,4-triphenyl-1-butene
had formed. Some unreacted (η3-allyl)Pd(PPh3)2Cl was also
present, but no other palladium-containing product could be
identified.
1
the total allyl products), all identified by comparisons with H
NMR spectra of authentic samples. The relative proportions of
these products were confirmed by GC-MS, which also demon-
strated the presence of trityl chloride as a major product. The
NMR reaction was repeated, and in this case 3-phenylpropene
was added after the reaction was complete. After further heat-
1
ing at 60 ЊC for 2 h, a H NMR spectrum showed that the
Results and discussion
3-phenylpropene had been converted to 1-phenylpropene.
In a complementary experiment, a solution of 0.01 g [(η3-
allyl)PdCl]2 (2.7 × 10Ϫ5 mol), 0.053 g PAT (1.5 × 10Ϫ4 mol) and
22.5 µL 3-chloropropene (2.75 × 10Ϫ4 mol) in 0.75 mL C6D6 was
heated at 60 ЊC for 2 h, by which time no metallic palladium had
formed but the solution had turned orange and an orange prod-
While nucleophilic addition reactions to allyl–palladium com-
plexes of types A, B and C have been thoroughly investigated,12
addition of organic radicals to these complexes has not been
explored previously. In this study, we have therefore investigated
the reactions of these same types of allyl–palladium complexes
A, B and C with phenyl and trityl radicals, formed via the
thermal decomposition of PAT. This was done with two major
objectives in mind, to gain an understanding of the nature
of radical interactions with allyl–palladium species, and, if
possible, to develop synthetically useful organic reactions. The
reactions studied here were carried out by reacting allylic com-
pounds of type A with phenyl and trityl radicals in the presence
of zero, one and two molar equivalents of PPh3 per palladium,
i.e. with compounds of types A, B and C, respectively, since we
find that PPh3 coordination occurs much more rapidly than
does PAT decomposition. As we shall show below, the nature
of the products varies as the PPh3 : palladium ratio changes,
demonstrating that different palladium substrates are involved.
Furthermore, coupling of the allylic fragments with the
radicals is shown to involve endo attack in one case, at least.
1
uct had begun to precipitate. The sample was filtered and a H
NMR spectrum was found to exhibit the resonances of 1-phenyl-
propene (∼10% of allyl products), 3-phenylpropene (∼80% of
allyl products) and 4,4,4-triphenyl-1-butene (∼10% of allyl prod-
ucts) as the major allyl products. The sample was analyzed using
GC-MS, which showed that 3-phenylpropene and trityl chloride
were the major products. The orange product was identified as
[(η3-trityl)PdCl]2 via conversion to the known (η3-trityl)Pd-
(acac).24 A repeat of the reaction in the presence of a 25 : 1 molar
excess of 3-chloropropene resulted in the formation predomin-
antly 3-phenylpropene in addition to [(η3-trityl)PdCl]2.24
Reactions of [(4-methoxy-1,3-ꢀ3-cyclohexenyl)PdCl2] with PAT
A yellow solution of 0.02 g [(4-methoxy-1,3-η3-cyclohexenyl)-
PdCl]2 19 (4.01 × 10Ϫ5 mol) and 0.03 g PAT (8.62 × 10Ϫ5 mol) in
0.75 mL C6D6 was heated at 60 ЊC for 10 min, forming a black/
green slurry. The slurry was filtered through alumina and the
Reactions in the presence of one and two equivalents of PPh3 per
palladium
1
filtrate was shown by H NMR spectroscopy to contain trans-
We shall address reactions of (η3-allyl)PdCl(PPh3) (B) and [(η3-
allyl)Pd(PPh3)2]Cl (C) first, thus making possible a comparison
with addition reactions of nucleophiles. The reaction of [(η3-
allyl)PdCl]2 with a two-fold excess of PAT in the presence of
two equivalents of PPh3 per palladium in C6H6 at 60 ЊC resulted
in the formation of the known compound trans-PdPhCl-
3-methoxy-6-phenylcyclohexene25a and trans-4-methoxy-3-
phenylcyclohexene25a in a ratio of ∼5 : 1. H NMR of trans-3-
1
methoxy-6-phenylcyclohexene (CDCl3): δ 7.2–7.8 (m, 5H, H
(Ph)), 5.8–5.95 (m, 2H, CH), 3.91 (br m, 1H, CH–OMe), 3.43
(m, 1H, CH–Ph), 3.41 (s, 3H, OCH3), 2.10 (m, 2H, CH2), 1.6
(m, 2H, CH2). 1H NMR of trans-4-methoxy-3-phenylcyclo-
hexene (CDCl3): δ 7.2–7.8 (m, 5H, H (Ph)), 5.90–5.55 (m, 2H,
CH), 3.43 (m, 1H, CH–Ph), 3.35 (br m, 1H, CH–OMe), 3.28 (s,
3H, OCH3), 1.9 (m, 4H, CH2).
These compounds were identified by comparison to a liter-
ature reference25a and by their synthesis via the reaction of a
solution of 0.065 g [(4-methoxy-1,3-η3-cyclohexenyl)PdCl]2
(1.30 × 10Ϫ4 mol) and 0.108 g PPh3 (4.12 × 10Ϫ4 mol) in 2.5 mL
THF at 0 ЊC with 0.15 mL of 3 M PhMgBr in ether (4.5 × 10Ϫ4
mol). The resulting greenish slurry was stirred for 30 min at 0 ЊC
and then filtered to give a yellow solution. The solvent was
1
(PPh3)2.21 Examination of the organic products by H NMR
spectroscopy suggested the presence of a single major allyl con-
taining product in addition to several byproducts resulting from
decomposition of PAT. The thermal decomposition of PAT in
benzene has been shown to produce biphenyl, triphenyl-
methane and 1-phenyl-4-triphenylmethyl-2–5-cylcohexadiene.13
The allylic product was separated from these and identified by
1H NMR spectroscopy as 4,4,4-triphenyl-1-butene.22
The same reaction was also carried out on an NMR-scale
1
1
and monitored by H NMR spectroscopy. The first H NMR
spectrum, run within 5 min of mixing of reagents, showed
the complete formation of [(η3-allyl)Pd(PPh3)2]Cl,18 and thus
phosphine coordination occurs much more rapidly than does
the decomposition of PAT. The resonances of trans-PdPh-
1
removed and the resulting yellow powder was shown by H
NMR spectroscopy to contain the above-mentioned trans-3-
methoxy-6-phenylcyclohexene and trans-4-methoxy-3-phenyl-
cyclohexene.
Cl(PPh ) and CH ᎐CHCH CPh were evident within 30 min
᎐
3
2
2
2
3
and these compounds had formed essentially quantitatively
within 1.5 h. Thus the overall the reaction is represented by
eqn. (3).
Reaction of [(ꢀ3-allyl)Pt(PPh3)2]Cl with PAT
A mixture of 0.152 g [(η3-allyl)Pt(PPh3)2]Cl (1.88 × 10Ϫ4 mol)
and 0.135 g PAT (3.76 × 10Ϫ4 mol) in 50 mL benzene was heated
at 60 ЊC for 2 h to give an orange–green solution. The solvent
was removed in vacuo and the resulting whitish yellow residue
[(η3-allyl)Pd(PPh3)2]Cl ϩ PAT
trans-PdClPh(PPh ) ϩ CH ᎐CHCH CPh ϩ N (3)
᎐
3
2
2
2
3
2
D a l t o n T r a n s . , 2 0 0 3 , 3 9 7 5 – 3 9 8 0
3977

View Article Online
There are two reasonable mechanistic routes to the observed
products. These involve either initial addition of the trityl radi-
cal to the allyl group to directly form 4,4,4-triphenyl-1-butene
(Fig. 4(a)), or initial attack of the phenyl radical on the
palladium (Fig. 4(b)). The initial intermediate in Fig. 4(a) is a
palladium() alkene complex, from which the π bonded ligand
would presumably dissociate to give the free 4,4,4-triphenyl-1-
butene which is observed. The resulting palladium() species
[PdI(PPh3)2]ϩ would then couple with a phenyl radical to form
the coordinatively unsaturated [PdII(PPh3)2Ph]ϩ, which would
readily coordinate the chloride ion to form the observed
product, trans-PdPhCl(PPh3)2.
butene while the palladium product was [PdPh(PPh3)(µ-Cl)]2,
identified spectroscopically, by its synthesis via a known route,23
and by its reaction with PPh3 to give trans-PdPhCl(PPh3)2. The
same products were also obtained using preformed (η3-allyl)-
PdCl(PPh3), and thus the overall reaction is represented by
eqn. (4).
(η3-allyl)PdCl(PPh3) ϩ PAT
½[PdPh(PPh )(µ-Cl)] ϩ CH ᎐CHCH CPh ϩ N (4)
᎐
3
2
2
2
3
2
These reactions proceed as readily as do the reactions in the
presence of two equivalents of PPh3 per palladium, demon-
strating that the cationic bisphosphine complex is not required
and thus that the radical addition/coupling processes are not
subject to the same controls as are nucleophilic addition reac-
tions. In fact the great similarity in the allyl-radical coupling
products from the radical reactions with (η3-allyl)PdCl(PPh3)
and [(η3-allyl)Pd(PPh3)2]Cl suggests essentially identical mech-
anisms for the two types of allylic compounds.
Reactions in the absence of PPh3
Quite different results were obtained, however, when [(η3-allyl)-
PdCl]2 was reacted with phenyl and trityl radicals in the absence
of PPh3. The most immediately apparent difference was the
rapid appearance of palladium metal although, after two hours,
GC-MS and 1H NMR monitoring in C6D6 showed that 1-
phenylpropene was the major allylic product (∼80%), in addi-
tion to 3-phenylpropene (∼10%) and 4,4,4-triphenyl-1-butene
(∼10%). GC-MS also showed that trityl chloride was also
formed as a major product during the reactions. The reaction
of [(η3-allyl)PdCl]2 with PAT may broadly be represented by
eqn. (5).
[(η3-allyl)PdCl]2 ϩ PAT
PhCH᎐CHCH ϩ PhCH CH᎐CH ϩ
᎐
᎐
3
2
2
4,4,4-triphenyl-1-butene ϩ N2 ϩ Pd(0) ϩ Ph3CCl (5)
Fig.
4
Possible mechanism for the reaction of [(η3-allyl)-
It seemed possible in this case that direct attack by the highly
reactive phenyl radical was occurring on the allyl ligand. There-
fore, in order to ascertain whether the coupling between the
phenyl radical and the allyl ligand proceeds via net exo or endo
attack, as defined in Fig. 3, we utilized [(4-methoxy-1,3-η3-
cyclohexenyl)PdCl]2 (D).19 Attack on the allylic ligand of this
compound by a phenyl radical, either directly (exo) or indirectly
via an initially formed phenyl-palladium intermediate (endo), is
expected to occur at C-1 and C-3 and to give 3-methoxy-6-
phenylcyclohexene and 4-methoxy-3-phenylcyclohexene.
Pd(PPh3)2]ϩClϪ with trityl and phenyl radicals (a) assuming initial
reaction of the trityl radical (exo), (b) assuming initial reaction of the
phenyl radical (endo).
The alternative mechanism, shown in Fig. 4(b), involves
initial reaction of the phenyl radical with the palladium. This
seems more plausible since, if the initial reaction does indeed
involve direct addition of the trityl radical to the allyl group,
then one would also expect at least some attack of phenyl radi-
cal to give 3-phenylpropene, which is not observed. Reaction of
the phenyl radical with the palladium centre would result in the
formation of an allylpalladium() intermediate which appar-
ently couples with trityl radical to form 4,4,4-triphenyl-1-
butene. Following dissociation of the latter, the palladium
product would be the same intermediate [PdII(PPh3)2Ph]ϩ pro-
posed in Fig. 4(a). Again coordination of chloride ion would
lead to the formation of the observed palladium product.
To complement this study, preformed [(η3-allyl)Pt(PPh3)2]-
Cl20 was also reacted with trityl and phenyl radicals via the
decomposition of PAT. This reaction was expected to proceed
similarly to the reaction of the analogous palladium complex,
and indeed it did although [(η3-allyl)Pt(PPh3)2]Cl is only mini-
mally soluble in benzene and the reaction began as a slurry. The
solution resulting at the end of the reaction was worked up
similarly to the analogous palladium reaction, and the products
were found to be 4,4,4-triphenyl-1-butene and trans-PtPhCl-
(PPh3)2.26 Thus the palladium and platinum allyl complexes
behave similarly.
These two possibilities are illustrated in Figs. 5(a) and (b),
where it is seen that endo attack would give the trans isomeric
products while exo attack would give the cis isomers. The com-
pounds trans-3-methoxy-6-phenylcyclohexene and trans-4-
1
methoxy-3-phenylcyclohexene are known and their H NMR
spectra have been reported.25a They have been prepared via the
reaction of compound D, activated with chelating, neutral
diimine ligands,25a with a source of phenyl carbanion in a reac-
tion known to give endo products via initial attack at the metal
followed by carbanion migration to the allylic ligand.25a To
obtain the two trans isomers for purposes of comparison, we
therefore reacted [(4-methoxy-1,3-η3-cyclohexenyl)Pd(PPh3)2]ϩ
with phenylmagnesium bromide, obtaining products exhibiting
1H NMR spectra consistent with literature data for the two
trans isomers.
In order to compare the reactions of compounds of types B
and C of Fig. 2, we also carried out a series of experiments in
which [(η3-allyl)PdCl]2 was reacted with PAT as above but in the
presence of only one molar equivalent of PPh3 per palladium.
In these cases, the only allyl product was 4,4,4-triphenyl-1-
D a l t o n T r a n s . , 2 0 0 3 , 3 9 7 5 – 3 9 8 0
3978

View Article Online
system, it is likely that the planar trity radical cannot abstract
the chlorine atom from the palladium when coordinated PPh3 is
present because of steric hindrance by the bulky PPh3 ligands.
Also, of course, in [(η3-allyl)Pd(PPh3)2]Cl the chloride is not
available for abstraction as it is probably ion paired to the com-
plex cation, not covalently bonded to the palladium. In any
case, phenyl radical migration from the palladium to the allyl
group results in the formation of 3-phenylpropene as the kinetic
product.
Catalytic isomerization of 3-phenylpropene
To assess the possibility that the major, thermodynamically
more stable product 1-phenylpropene was formed by palladium
metal catalyzed isomerization of the kinetic product, 3-phenyl-
propene, a reaction of [(η3-allyl)PdCl]2 with PAT was carried
out for 2 h. At this point 1-phenylpropene was the major allyl
product, and 3-phenylpropene was added and the mixture was
1
heated for a further 2 h at 60 ЊC. A H NMR spectrum then
showed that all of the 3-phenylpropene had been converted to
1-phenylpropene. The mechanism probably involves an allylic
intermediate as in Fig. 6.29
Fig.
cyclohexenyl)PdCl]2 (a) via a mechanism involving initial attack at the
palladium, (b) via mechanism involving initial attack at the
coordinated allyl group.
5
Reaction of the phenyl radical with [(6-methoxy-1,3-η3-
a
We then reacted D with PAT, much as above although over
shorter time periods (10, 20 min) because D itself is somewhat
thermally unstable under these conditions. We also wished to
minimize possible palladium metal-catalyzed isomerization of
the product (see below). The products of the two phenylation
reactions of D, utilizing carbanionic and radical sources, were
then compared (1H NMR spectroscopy, GC-MS) and found to
be identical, consistent with the formation of the trans endo
products in both cases. Therefore, unless the corresponding cis
isomers exhibit 1H NMR spectra identical to those of the trans
isomers, the reaction of (4-methoxy-1,3-η3-cyclohexenyl)PdCl]2
with PAT appears to proceed as illustrated in Fig. 5(a) rather
than as in Fig. 5(b). While the alternative cis isomeric com-
pounds have not been reported, isomeric pairs of several other
disubstituted cyclohexenyl compounds are known and gener-
Fig. 6 Proposed mechanism for the catalytic isomerization of 3-
phenylpropene to 1-phenylpropene.
It seemed unfortunate that the 3-phenylpropene, formed as
the kinetic product of the reaction of [(η3-allyl)PdCl]2 with
phenyl and trityl radicals, is isomerized catalytically by the pal-
ladium metal also formed, and we wondered if we might be able
to intercept the palladium metal by carrying out the reaction in
the presence of 3-chloropropene. It has long been known that
allylic halides undergo oxidative addition reactions to Pd-
(PPh3)4 to give compounds of the type [(η3-allyl)Pd(PPh3)2]Cl,12
and we realized that if 3-chloropropene would react similarly
with the palladium metal formed in eqn. (5), it would regenerate
[(η3-allyl)PdCl]2. This would result in shutting down the isomer-
ization of 3-phenylpropene by removing the palladium metal
catalyst, and might possibly also result in the coupling reaction
becoming catalytic in palladium.
1
ally exhibit both different H chemical shifts and different GC
retention times.19,25b,c Thus the above conclusions are very
reasonable.
It is probable that the analogous reaction of [(η3-allyl)PdCl]2
(eqn. (5)) also proceeds via initial attack at the palladium to
form a phenylpalladium() intermediate, as with (η3-allyl)-
PdCl(PPh3) and [(η3-allyl)Pd(PPh3)2]Cl, rather than via direct
addition of a phenyl radical to the coordinated allyl ligand.
However, in contrast to the analogous reactions of the phos-
phine containing complexes, here the phenyl group sub-
sequently couples with the allyl ligand to give 3-phenylpropene
rather than remaining bonded to the palladium. The process by
which 3-phenylpropene is formed presumably involves phenyl
ligand migration to the allyl ligand in what is essentially a
reductive elimination process. For trityl chloride formation to
occur, the trityl radical must abstract the chlorine as the “PdCl”
is formed. The phenyl–allyl coupling and the chlorine abstrac-
tion processes are probably concerted as a simple halogen
abstraction reaction by the trityl radical from the metal would
almost certainly be rather endothermic.27
We therefore carried out a reaction of [(η3-allyl)PdCl]2
with PAT (C6D6, 60 ЊC) in the presence of five equivalents of
3-chloropropene per palladium. In pleasant contrast to the
analogous reaction carried out in the absence of 3-chloro-
propene, palladium metal did not precipitate but rather an
orange solution and subsequently an orange precipitate were
formed. 1H NMR spectroscopy and GC-MS analyses again
showed that 1-phenylpropene, 3-phenylpropene and 4,4,4-
triphenyl-1-butene were formed, but in this case 3-phenyl-
propene was the major allyl product (∼80% yield). A reaction
with 25 equivalents of 3-chloropropene per palladium also
resulted in 3-phenylpropene being formed exclusively, while a
control experiment showed that PAT does not react directly
with 3-chloropropene under these conditions.
There are probably several reasons for the quite different
secondary processes observed in the presence and absence of
PPh3. Tertiary phosphines have long been known for their abil-
ity to stabilize alkyl– and aryl–metal compounds with respect to
a variety of thermal decomposition processes,28 and therefore,
in a PPh3-free compound, it seems reasonable that reductive
elimination via initial phenyl migration is intrinsically more
facile. Complementing this higher lability of the PPh3-free
The same reaction was then carried out in the presence of 50
equivalents of 3-chloropropene and 25 equivalents of PAT. The
volatile products were collected by vacuum distillation and
shown (GC-MS) to be exclusively 3-phenylpropene. Also
1
obtained was an orange residue, the H NMR spectrum of
which exhibited only phenyl resonances. This product therefore
D a l t o n T r a n s . , 2 0 0 3 , 3 9 7 5 – 3 9 8 0
3979

View Article Online
seemed likely to be [(η3-trityl)PdCl]2, formed via oxidative addi-
tion of Ph3Cl to palladium(0).24 To characterize the compound,
it was reacted with Tl(acac) to form the known (η3-trityl)Pd-
(acac).24 The 1H NMR spectrum of the product from this reac-
tion contained singlets from the acac group at δ 5.09, 1.76 and
1.84, consistent with the literature spectrum of (η3-trityl)-
Pd(acac).
The results imply that a catalytic cycle is present, quite likely
as shown in Fig. 7. On the basis of the GC-MS experiments, it
appears as though approximately 15 turnovers in 1.5 h were
achieved with a 50 : 1 ratio of 3-chloropropene to Pd.
5 (a) R. L. Sweaney, D. S. Comberrel, M. F. Dombourian and
N. A. Peters, J. Organomet. Chem., 1981, 216, 57; (b) F. Ungváry and
L. Markó, Organometallics, 1982, 1, 1120; (c) T. M. Bockman, J. F.
Garst, R. B. King, L. Markó and F. Ungváry, J. Organomet. Chem,
1985, 279, 165; (d ) R. M. Bullock and E. G. Samsel, J. Am. Chem.
Soc., 1990, 112, 6886.
6 (a) J. W. Connolly, Organometallics, 1984, 3, 1333; (b) F. Ungváry
and L. Markó, Organometallics, 1984, 3, 1466; (c) B. Wassink, M. J.
Thomas, S. C. Wright, D. J. Gillis and M. C. Baird, J. Am. Chem.
Soc., 1987, 109, 1995; (d ) T. A. Shackleton, S. C. Mackie, S. B.
Fergusson, L. J. Johnston and M. C. Baird, Organometallics, 1990, 9,
2248.
7 (a) E. Baciocchi, B. Floris and E. Muraglia, J. Org. Chem., 1993, 58,
2013; (b) H.-G. Schmalz, S. Siegel and J. W. Bats, Angew. Chem.,
Int. Ed. Engl., 1995, 34, 2383; (c) H.-G. Schmalz, S. Siegel and
A. Schwarz, Tetrahedron Lett., 1996, 17, 2947.
8 (a) C. A. G. Carter, R. McDonald and J. M. Stryker, Organo-
metallics, 1999, 18, 820; (b) S. Ogoshi and J. M. Stryker, J. Am.
Chem. Soc., 1998, 120, 3514; (c) G. L. Casty and J. M. Stryker,
J. Am. Chem. Soc., 1995, 117, 7814.
9 C. A. Merlic and D. Xu, J. Am. Chem. Soc., 1991, 113, 9855.
10 G. G. Melikyan, O. Vostrowsky, W. Bauer, H. J. Bestmann, M. Khan
and K. M. Nicholas, J. Org. Chem., 1994, 59, 222.
11 K. E. Torraca and L. McElwee-White, Coord. Chem. Rev., 2000,
206–207, 469.
12 (a) S. A. Godleski, in Comprehensive Organic Synthesis, ed. B. M.
Trost and I. Fleming, Pergamon Press, Oxford, 1991, vol. 4, p. 585;
(b) F. Guibe, Tetrahedron, 1998, 54, 2967; (c) J. Tsuji, Palladium
Reagents and Catalysis: Innovations in Organic Synthesis, Wiley,
Chichester, 1995; (d ) P. J. Harrington, Comprehensive Organo-
metallic Chemistry, II, vol. 12 (ed. E. W. Abel, F. G. A. Stone and
G. Wilkinson, Elsevier, New York, 1995, p. 797.
13 (a) S. G. Cohen and C. H. Wang, J. Am. Chem. Soc., 1953, 75, 5504;
(b) R. R. Bridger and G. A. Russell, J. Am. Chem. Soc., 1963, 85,
3754; (c) R. G. Kryger, J. P. Lorand, N. R. Stevens and N. R. Herron,
J. Am. Chem. Soc., 1977, 99, 7589.
14 T. H. Colle, P. S. Glaspie and E. S. Lewis, J. Org. Chem., 1978, 43,
2722.
15 S. J. Reid, N. T. Freeman and M. C. Baird, Chem. Commun., 2000,
1777.
16 D. R. Coulson, Inorg. Synth., 1972, 13, 121.
17 V. V. Grushin, C. Bensimon and H. Alper, Organometallics, 1995, 14,
3259.
18 (a) W. T. Dent, R. Long and A. J. Wilkinson, J. Chem. Soc., 1964,
1585; (b) M. Sakakibara, Y. Takahashi, S. Sakai and Y. Ishii, Chem.
Commun., 1969, 396; (c) G. L. Statton and K. C. Ramey, J. Am.
Chem. Soc., 1966, 88, 1327; (d ) K. C. Ramey and G. L. Statton,
J. Am. Chem. Soc., 1966, 88, 4837.
19 J.-E. Bäckvall, R. E. Nordberg, K. Zetterberg and B. Åkermark,
Organometallics, 1983, 2, 1625.
20 H. C. Volger and K. Vrieze, J. Organomet. Chem., 1967, 9, 527.
21 (a) D. R. Coulson, Chem. Commun., 1968, 1530; (b) P. Fitton, M. P.
Johnson and J. E. McKeon, Chem. Commun., 1968, 6; (c) A. Mentes,
R. D. W. Kemmitt, J. Fawcett and D. R. Russeli, Polyhedron, 1999,
18, 1141.
Fig. 7 Catalytic cycle of reaction of [(η3-allyl)PdCl]2, PAT and 3-
chloropropene, showing also the role of trityl chloride in subverting the
catalysis.
The reaction of [(ꢀ3-allyl)Pd(PPh3)2]Cl with the trityl dimer
The reaction of [(η3-allyl)PdCl]2 with trityl dimer in the pres-
ence of two equivalents of PPh3 at 60 ЊC was carried out as a
test of the ability of the trityl radical to react directly with a
coordinated allyl group. Although the reaction was allowed to
run for 1.5 h, much longer than is required for complete reac-
tions of PAT, only a small amount of 4,4,4-triphenyl-1-butene
was produced and a substantial amount of unreacted [(η3-
allyl)Pd(PPh3)2]Cl was also isolated. All attempts to identify a
palladium product from this reaction were unsuccessful.
While this result suggests that the trityl radical can react
directly with the allyl group of [(η3-allyl)Pd(PPh3)2]Cl, it is clear
that the reaction proceeds at a much lower rate than the reac-
tion of [(η3-allyl)Pd(PPh3)2]Cl with the combination of phenyl
and trityl radicals derived from the thermal decomposition of
PAT.
Acknowledgements
22 (a) E. Moret, F. Fürrer and M. Schlosser, Tetrahedron, 1988, 44,
3539; (b) J.-P. Dau-Schmidt and H. Mayr, Chem. Ber., 1994, 127,
205; (c) K. Nozaki, T. Nanno and J. Takaya, J. Organomet. Chem.,
1997, 527, 103.
We thank the Natural Sciences and Engineering Research
Council (Research Grant to M. C. B.) and the Government of
Ontario (Ontario Government Scholarship to S. J. R.) for
financial support.
23 V. V. Grushin, Organometallics, 2000, 19, 1888.
24 (a) A. Sonoda, B. E. Mann and P. M. Maitlis, J. Chem. Soc., Chem.
Commun., 1975, 108; (b) B. E. Mann, A. Keasey, A. Sonoda and
P. M. Maitlis, J. Chem. Soc., Dalton Trans., 1979, 338; (c) A. Sonoda,
P. M. Bailey and P. M. Maitlis, J. Chem. Soc., Dalton Trans., 1979,
346; (d ) S. J. Reid and M. C. Baird, Organometallics, 1997, 16, 2481.
25 (a) B. Crociani, S. Antonaroli, F. Di Bianca and A. Fontana,
J. Organomet. Chem., 1993, 450, 21; (b) B. M. Trost and T. R.
Verhoeven, J. Am. Chem. Soc., 1980, 102, 4730; (c) H. Kurosawa,
H. Kajimaru, S. Ogoshi, H. Yoneda, K. Miki, N. Kasai, S. Murai
and I. Ikeda, J. Am. Chem. Soc., 1992, 114, 8417.
26 (a) G. K. Anderson, H. C. Clark and J. A. Davies, Organometallics,
1982, 1, 64; (b) C. Eaborn, K. J. Odell and A. Pidcock, J. Chem. Soc.,
Dalton Trans., 1978, 357; (c) C. Eaborn, A. Pidcock and B. R. Steele,
J. Chem. Soc., Dalton Trans., 1976, 767; (d ) M. C. Baird and
G. Wilkinson, J. Chem. Soc. A, 1967, 865; (e) M. C. Baird, J. Inorg.
Nucl. Chem., 1967, 29, 367.
References
1 (a) D. P. Curran, in Comprehensive Organic Synthesis, Pergamon
Press, New York, 1991, vol. 4, p. 715; (b) B. Giese, Radicals in
Organic Synthesis: Formation of Carbon–Carbon Bonds, Pergamon
Press, Oxford, 1986; (c) D. P. Curran, N. A. Porter and B. Giese,
Stereochemistry of Radical Reactions, VCH, Weinhem, 1996;
(d ) J. A. M. Simoes, A. Greenberg and J. F. Liebman, Energetics of
Organic Free Radicals, Blackie Academic and Professionals, 1996;
(e) J. E. Leffler, An Introduction to Free Radicals, John Wiley and
Sons, New York, 1993; ( f ) H. Fischer and L. Radom, Angew.
Chem., Int. Ed., 2001, 40, 1340.
2 (a) J. K. Stille, in The Chemistry of the Metal-Carbon Bond,
ed. F. R. Hartley and S. Patai, Wiley and Sons, Chichester, 1985, vol.
2, ch. 9; (b) L. M. Rendina and R. J. Puddephatt, Chem. Rev., 1997,
97, 1735.
27 M. C. Baird, Chem. Rev., 1988, 88, 1217.
28 M. C. Baird, J. Organomet. Chem., 1974, 64, 289.
3 A. Krief and A. M. Laval, Chem. Rev., 1999, 99, 745.
4 (a) J. K. Kochi and J. W. Powers, J. Am. Chem. Soc., 1970, 92, 137;
(b) R. Sustmann and R. Altevogt, Tetrahedron Lett., 1981, 22,
5167.
29 (a) I. I. Moiseev, T. A. Stromnova and M. N. Vargaftik, J. Mol.
Catal., 1994, 86, 71; (b) R. Touroude, L. Hilaire and F. G. Gault,
J. Catal., 1974, 32, 279; (c) N. R. Davies, Aust. J. Chem., 1964, 17,
212.
D a l t o n T r a n s . , 2 0 0 3 , 3 9 7 5 – 3 9 8 0
3980