Reactions of free radicals with η3-allylpalladium(II) complexes: Cyclohexyl radicals

Article abstract of DOI:10.1016/j.jorganchem.2004.01.009

Allyl palladium complexes of the types [(η³-allyl)PdCl] ₂, (η³-allyl)PdCl(PPh₃) and [(η³-allyl)Pd(PPh₃)₂]Cl (allyl=C₃H₅, 1-MeC₃H₄, 2-MeC₃H₄, 1-PhC₃H₄, 2-PhC₃H₄) react with cyclohexyl radicals derived from the visible light photolysis of (c-hex)Co(DMG)₂(py). The reactions proceed via initial attack of the free radical at the metal center, followed by β-hydrogen elimination and subsequent reductive elimination of propene, 1-butene, isobutene, 3-phenylpropene and 2-phenylpropene, respectively. The 3-phenylpropene can be catalytically isomerized to the thermodynamically more stable 1-phenylpropene by either palladium metal or palladium(0) products, but the formation of 1-butene and 3-phenylpropene as primary products is unusual. A mechanism, differing in many ways from that proposed previously for analogous reactions of phenyl and trityl radicals, is proposed for the overall reaction and supported by use of the labeled cobaloxime, (2,2,6,6-D₄-c-hex)Co(DMG)₂(py).

Full text of DOI:10.1016/j.jorganchem.2004.01.009

Journal of Organometallic Chemistry 689 (2004) 1257–1264
www.elsevier.com/locate/jorganchem
Reactions of free radicals with g3-allylpalladium(II) complexes:
cyclohexyl radicals
*
Simon J. Reid, Michael C. Baird
Department of Chemistry, QueenÕs University, Kingston, Ont., Canada K7L 3N6
Received 2 December 2003; accepted 5 January 2004
Abstract
Allyl palladium complexes of the types [(g3-allyl)PdCl]2, (g3-allyl)PdCl(PPh3) and [(g3-allyl)Pd(PPh3)2]Cl (allyl ¼ C3H5,
1-MeC3H4, 2-MeC3H4, 1-PhC3H4, 2-PhC3H4) react with cyclohexyl radicals derived from the visible light photolysis of (c-hex)-
Co(DMG)2(py). The reactions proceed via initial attack of the free radical at the metal center, followed by b-hydrogen elimination
and subsequent reductive elimination of propene, 1-butene, isobutene, 3-phenylpropene and 2-phenylpropene, respectively. The 3-
phenylpropene can be catalytically isomerized to the thermodynamically more stable 1-phenylpropene by either palladium metal or
palladium(0) products, but the formation of 1-butene and 3-phenylpropene as primary products is unusual. A mechanism, differing
in many ways from that proposed previously for analogous reactions of phenyl and trityl radicals, is proposed for the overall re-
action and supported by use of the labeled cobaloxime, (2,2,6,6-D4-c-hex)Co(DMG)2(py).
Ó 2004 Elsevier B.V. All rights reserved.
Keywords: Palladium; Allyl; Free radicals; Alkenes
1. Introduction
In spite of this activity [11], there is as yet really no
general knowledge or understanding of the reactions of
free radicals with coordinated ligands and we recently
began an investigation of reactions of free radicals with
neutral and cationic g3-allylic palladium compounds of
types A, B and C (Fig. 1). Palladium compounds con-
taining g3-allylic ligands were chosen because of the
susceptibility of such compounds to nucleophilic attack
[12]. Indeed, cationic complexes of type C react readily
as in Eq. 1 with a wide variety of C-, O- and N-based
nucleophiles Nuꢀ although the neutral compounds A
and B exhibit relatively limited activities
Free radicals have been invoked as reactive interme-
diates in a wide variety of reactions involving transition
metal compounds, e.g., oxidative addition reactions [1],
SmI2 induced coupling of organic halides with carbonyl
compounds [2], reduction of organic halides by com-
plexes of chromium(II) [3], and hydrogenation and hy-
drometallation of aromatic alkenes [4] and conjugated
dienes [5]. In addition, there has in recent years been
increasing interest in addition reactions of free radicals
on complexes of a variety of unsaturated, coordinated
organic compounds such as arene [6], allylic [7], carbe-
noid [8] and alkenyne [9] ligands. The latter studies
complement to a growing extent many better developed
studies of free radical addition reactions to unsaturated
carbon–carbon bonds, which provide a range of syn-
thetically useful organic transformations [10].
½ðg3-allylÞPdðPPh3Þ ꢁþ þ Nuꢀ
2
! allyl–Nu þ \PdðPPh3Þ "
ð1Þ
2
Nucleophilic attack on complexes of type C has been
thoroughly investigated and results in a range of ex-
tremely useful synthetic methodologies; as a result, g3-
allylic palladium(II) complexes of type C have been
employed extensively in synthetic organic chemistry [12].
It was anticipated that equally interesting and useful
coupling/addition reactions of free radicals with allylic
ligands might be observed, and that products not
* Corresponding author. Tel.: +1-613-533-2616; fax: +1-613-533-
6669.
E-mail address: bairdmc@chem.queensu.ca (M.C. Baird).
0022-328X/$ - see front matter Ó 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.01.009

1258
S.J. Reid, M.C. Baird / Journal of Organometallic Chemistry 689 (2004) 1257–1264
Cl
Cl
PPh3
PPh3
+
PPh3
PPh3
Cl-
Pd
Pd
Pd
1/2
PPh3
2
A
B
C
Fig. 1. Compounds present on the addition of triphenylphosphine to [(g3-allyl)PdCl]2 (A).
available by conventional palladium allyl chemistry
might be formed since, e.g., base-sensitive functional
groups should not be affected.
We began our studies utilizing phenylazotriphenyl-
methane, PhN@NCPh3 (PAT), as a source of phenyl
and triphenylmethyl (trityl) radicals [13]. PAT decom-
poses in inert solvents on heating by releasing phenyl
and trityl radicals in addition to molecular nitrogen (Eq.
2) [13]
class of so-called alkylcobaloxime compounds RCo-
(DMG)2(L) (R ¼ alkyl, L ¼ ligand) known to undergo
cobalt-carbon bond homolysis on irradiation with visi-
ble light at ambient temperatures to give a carbon cen-
tred radical and the cobalt(II) compound Co(DMG)2(L)
(as in Eq. 3 for R ¼ cyclohexyl) [15]
hv
ðc-hexyÞCoðDMGÞ ðpyÞ!CoðIIÞðDMGÞ ðpyÞ þ C6H1Å 1
2
2
ð3Þ
PhN@NCPh3 ! PhÅ þ Ph3CÅ þ N2
ð2Þ
We have earlier communicated some of our initial
finding on this chemistry [16] and have reported in
detail the chemistry of allyl-palladium compounds
with phenyl and trityl radicals [14]. We now discuss
an investigation of the analogous chemistry with cy-
clohexyl radicals, finding significant differences in re-
action pathways.
We found [14] that compounds of the types (g3-al-
lyl)PdCl(PPh3) and [(g3-allyl)Pd(PPh3)2]Cl react readily
with phenyl and trityl radicals in benzene at 60 °C, the
products being the palladium phenyl compounds,
[PdPhCl(PPh3)]2 and trans-PdPhCl(PPh3)2, respectively,
and 4,4,4-triphenyl-1-butene. In contrast, [(g3-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 catalyti-
cally isomerized to 1-phenylpropene. These disparate
results were interpreted generally in terms of initial at-
tack by phenyl radicals on the palladium(II) to give
phenyl-palladium(III) intermediates followed by a vari-
ety of secondary reactions, depending on the degree of
coordination of PPh3 ligands.
2. Experimental
Experiments were conducted under an inert atmo-
sphere of oxygen-free N2 or Ar, further purified by
passing through a column of BASF catalyst heated to
140 °C and a column of 5A molecular sieves. Manipu-
lations of air-sensitive materials employed standard
Schlenk line techniques, a Vacuum Atmosphere glove-
box and an Mbraun Labmaster glovebox. Solvents were
dried and distilled over sodium metal except dichlo-
romethane, which was dried over calcium hydride, and
methanol, which was dried over magnesium methoxide;
all were thoroughly deoxygenated prior to use by freeze–
thaw–pump cycles or by saturation with N2 or Ar.
NMR spectra were acquired on Bruker ACF 200,
Avance 300, AM 400 or Avance 500 NMR spectrome-
ters. The residual proton and the carbon resonances of
We have continued our studies by exploring the
chemistry of the same types of allyl-palladium com-
pounds A, B and C, with the cyclohexyl radical, chosen
because of its greater nucleophilic character relative to
the phenyl and trityl radicals [10]. As a source of cy-
clohexyl radicals, we have utilized the compound
cyclohexylbis(dimethylglyoximato)(pyridine)cobalt(III),
(c-hexyl)Co(DMG)2(py), D of Fig. 2, a member of a
1
deuterated solvents served as internal references for H
and 13C resonances, respectively, while 31P resonances
were referenced externally to 85% H3PO4. 2H NMR
spectra were run in non-deuterated solvents with the
naturally occurring abundance of deuterium in the sol-
vent serving as the internal reference. Mass spectra were
obtained on a Quatro Fisons Pro Quadropole mass
spectrometer in EIþ mode. GC–MS experiments were
carried out on a Fisons GC 8000 coupled to a Fisons
MD 800 spectrometer operating in EIþ mode or a
Varian GC–MS (CP-3800 GC and Saturn 2000 MS).
H
O
N
O
N
C6H11
H3C
H3C
CH3
CH3
Co
py
N
O
N
O
H
D
Fig. 2. Structure of (c-hexyl)Co(DMG)2(py).

S.J. Reid, M.C. Baird / Journal of Organometallic Chemistry 689 (2004) 1257–1264
1259
Except as noted below, all chemicals were purchased
from Aldrich and were used as received. Deuterated
solvents were purchased from MSD isotopes and
Cambridge Isotopes Limited, while PdCl2 was obtained
on loan from Johnsonn Mathey. The following com-
plexes were synthesized via standard literature prepa-
rations: [(g3-allyl)PdCl]2 [17], (g3-allyl)Pd(PPh3)Cl
[17d], [(g3-allyl)Pt(PPh3)2]Cl [18], [(g3-1-methylal-
lyl)PdCl]2 [17a,19], [(g3-2-methylallyl)PdCl]2 [17a,20],
[(g3-1-phenylallyl)PdCl]2 [17b,21], [(g3-2-phenylal-
lyl)PdCl]2 [22], trans-PtHCl(PPh3)2 [23] and (c-hex)-
Co(DMG)2(py) (D) [15a]. 1H NMR of D (CDCl3) d
18.05 (s, 2H, OH), d 8.50 (d, 2H, H-1, 3JH1–H2 ¼ 4:0 Hz),
0–10. There were also several resonances in the 31P NMR
spectrum, indicating that more than one palladium
compound had formed. GC–MS of the distillate showed
that only very low amounts of 3-cyclohexylpropene and
cyclohexenyl chloride were present. Similar results were
obtained using preformed (g3-allyl)Pd(PPh3)Cl.
In an NMR scale reaction, a solution of 0.010 g [(g3-
allyl)PdCl]2 (2.7 ꢄ 10ꢀ5 mol), 0.014 g PPh3 (5.33 ꢄ 10ꢀ5
mol), and 0.030 g D (6.7 ꢄ 10ꢀ5 mol) in 1 mL C6D6 in an
NMR tube was photolysed as above for 3 h. The solu-
tion turned from orange to deep red, but 1H NMR
monitoring revealed no resonances attributable to 3-
cyclohexylpropene; as D disappeared, propene and cy-
clohexene were formed in yields up to 80% in some
3
d 7.61 (t, 1H, H-3, JH3–H2 ¼ 6:1 Hz), d 7.20 (t, 2H, H2,
3
1
3JH1–H2 ¼ 4:0 Hz, JH3–H2 ¼ 6:1 Hz), d 2.07 (s, 12H,
cases. The H NMR spectrum was very similar to that
described above.
H(Me)), d 0.7–1.95 (m, 11H, C6H11). Literature [15a] 1H
NMR (CDCl3) d 8.56 (d, 2H, J ¼ 4 Hz), d 7.66 (m, 1H),
d 7.26 (m, 1H), d 2.14 (s, 12H), d 0.7–2.0 (m, 11H).
(2,2,6,6-D4-c-hex)Co(DMG)2(py) was prepared in
the same manner as for (c-hex)Co(DMG)2(py) except
that 2,2,6,6-D4-cyclohexyl bromide was used in place of
cyclohexyl iodide. 2,2,6,6-D4-cyclohexylbromide was
prepared via a literature preparation using the following
general route [24]. 2,2,6,6-D4-cyclohexanone was first
prepared via Na2CO3 catalyzed hydrogen/deuterium
exchange, at the 2,2,6 and 6 positions, of cyclohexanone
in D2O. 2,2,6,6-D4-cyclohexanone was then reduced to
2,2,6,6-D4-cyclohexanol with NaBH4 rather than
LiAlH4which was used in the literature. 2,2,6,6-D4-cyclo-
hexanol was finally converted to 2,2,6,6-D4-cyclohexyl
bromide with PPh3 and Br2. (2,2,6,6-D4-c-hex)-
Co(DMG)2(py) was then prepared using the standard
In a complementary NMR experiment, a solution of
0.010 g [(g3-allyl)PdCl]2 (2.7 ꢄ 10ꢀ5 mol), 0.014 g PPh3
(5.33 ꢄ 10ꢀ5 mol), and 0.030 g D (6.7 ꢄ 10ꢀ5 mol) in 1
mL C6D6 in an NMR tube was photolysed as above for
3 h. At this point, 2 lL 3-chloropropene (2.5 ꢄ 10ꢀ5
mol) were added to the mixture and the resonance of the
starting material, (g3-allyl)Pd(PPh3)Cl, immediately re-
appeared but no resonances of 3-chloropropene were
evident. A further 2.5 h of photolysis resulted in the
formation of more propene at the expense of the allyl
starting material. This procedure was repeated, dem-
onstrating that the reaction was catalytic in palladium to
at least a limited extent.
2.2. Reaction of [(g3-allyl)Pd(PPh3)2]Cl with cyclo-
hexyl radicals
1
literature procedure [15a]. Yield 1.05 g (36%) H NMR
(CDCl3) d 18.05 (s, 2H, OH), 8.49 (d, 2H, H-1,
3
3JH1ꢀH2 ¼ 4:0 Hz), 7.57 (t, 1H, H-3, JH3ꢀH2 ¼ 6:1 Hz),
A solution of 0.050 g [(g3-allyl)PdCl]2 (1.37 ꢄ 10ꢀ4
mol), 0.155 g PPh3 (5.91 ꢄ 10ꢀ4 mol) and 0.145 g D
(1.57 ꢄ 10ꢀ3 mol) in 15 mL benzene was photolysed as
above for 4 h. The reaction mixture was worked up as
above, and the 1H NMR spectrum of the resulting
reddish brown residue differed somewhat from that of
the residue resulting from the reaction of (g3-al-
lyl)Pd(PPh3)Cl in that weak multiplets in the region d
4.7–4.9 were present while the resonance at d )13 was
sometimes absent. An NMR reaction run as above in
C6D6 revealed also the formation of propene and cy-
clohexene in ꢃ25% yield.
3
3
7.19 (t, 2H, H-2, JH1ꢀH2 ¼ 4:0 Hz, JH3ꢀH2 ¼ 6:1 Hz),
2.06 (s, 12H, H(Me)), 0.7–1.95 (m, ꢂ 9–10 H, C6H11).
2H NMR (CDCl3) d 1.85, 1.68, 1.47, 1.23, 0.82.
Photochemical experiments were carried out in
Schlenk flasks and NMR tubes, maintained at 25 °C in a
water bath and placed ꢃ8 cm from a 300 W Sylvania
light bulb. The whole apparatus was wrapped in alu-
minum foil to allow maximum illumination.
2.1. Reaction of (g3-allyl)Pd(PPh3)Cl with cyclohexyl
radicals
A solution of 0.151 g [(g3-allyl)PdCl]2 (4.13 ꢄ 10ꢀ4
mol), 0.261 g PPh3 (8.24 ꢄ 10ꢀ4 mol) and 0.450 g D
(1.57 ꢄ 10ꢀ3 mol) in 15 mL benzene was photolysed as
above for 4 h, at which time the solvent was vacuum
2.3. Reactions of [(g3-allyl)PdCl]2 with cyclohexyl
radicals
A reaction as above but with no PPh3 added resulted
in the formation of palladium metal, propene and cy-
clohexene. A similar reaction of [(g3-1-phenylallyl)-
PdCl]2 also resulted in the formation of palladium metal
in addition to 1-phenylpropene.
1
distilled from the reddish brown solution. A H NMR
spectrum of the residue exhibited resonances at d 41.7,
27.0, 21.3; 15.9, )3.3, )13 and )23.9, in addition to many
unassigned resonances, some quite broad, in the region d

1260
S.J. Reid, M.C. Baird / Journal of Organometallic Chemistry 689 (2004) 1257–1264
2.4. Reactions of (g3-1-methylallyl)Pd(PPh3)Cl, (g3-2-
methylallyl)Pd(PPh3)Cl, (g3-1-phenylallyl)Pd(PPh3)-
Cl, and (g3-2-phenylallyl)Pd(PPh3)Cl with cyclohexyl
radicals
showed there was a substantial percentage of the 3-
phenylpropene with a mass increase of one, corre-
sponding to mono-deuteration.
2.6. Reaction of [(g3-allyl)Pt(PPh3)2]Cl with cyclo-
hexyl radicals
The procedure outlined above for an NMR scale re-
action at 25 °C was followed in all cases. The reactions
were generally somewhat slower, and in all cases 1H
NMR spectroscopy showed that D disappeared while
cyclohexene was formed. In the cases of the 1-methyl-
allyl, 2-methylallyl and 2-phenylallyl compounds, 1-bu-
tene, isobutene and 2-phenylpropene, respectively, were
An orange slurry of 0.211 g [(g3-allyl)Pt(PPh3)2]Cl
(2.66 ꢄ 10ꢀ4 mol) and 0.232 g D (9.11 ꢄ 10ꢀ4 mol) in 50
mL C6H6 was irradiated as above for 4 h. The solvent
was removed from the resulting reddish brown solution
to give a brown residue, the 1H NMR spectrum of which
exhibited inter alia resonances at d 41.7, 27.0, 21.3 15.9,
)3.3, )23.9 and )16.32. The latter was a 1:6:1 triplet
(1JPtꢀH ¼ 1197 Hz). The reaction was also carried out
on an NMR scale in C6D6; the observed products were
propene, cyclohexene and PtHCl(PPh3)2 in high yields,
but no cyclohexane or bicyclohexyl.
1
obtained in high yields; all were identified by H NMR
spectroscopy. In the case of the 1-phenylallyl com-
pound, 1-phenylpropene and 3-phenylpropene were
obtained in a ꢃ1:1 ratio. In all cases, the 1H NMR
spectra also exhibited resonances at d 41.7, 27.0, 21.3
15.9, )3.3, )13 and )23.9. In a reaction of (g3-1-phe-
nylallyl)Pd(PPh3)Cl at 6 °C, a mixture of 3-phenylpro-
pene and 1-phenylpropene were formed in a 3:1 ratio.
In contrast, a similar reaction of [(g3-1-phenylallyl)Pd-
(PPh3)2]Cl at 6 °C produced only 3-phenylpropene.
3. Results and discussion
To extend our previous investigation of the reactions
of palladium allyl complexes of the types A, B and C
with phenyl and trityl radicals [14], we now report a
study of the reactions of the same types of allyl palla-
dium compounds with the cyclohexyl radical. Coba-
loximes, RCo(DMG)2(L), generally have rather weak
Co-C bonds which undergo homolytic cleavage on ir-
radiation by visible light to form alkyl radicals RÅ and
the cobalt(II) compound Co(DMG)2(L). Thus cyclo-
hexyl radicals are formed by visible light photolysis of
the cobaloxime, D [15], and have been shown to exhibit
chemical properties identical to those of cyclohexyl
radicals generated from other methods of radical pro-
duction [25].
2.5. Reaction of [(g3-1-phenylallyl)Pd(PPh3)2]Cl with
deuterated cyclohexyl radicals derived from (2,2,6,6-d4-c-
hex)Co(DMG)2(py)
This reaction was carried out as above at 6 °C. Again
3-phenylpropene was evident in the 1H NMR spectrum,
2
while a H NMR spectrum exhibited a resonance of 3-
phenylpropene at d 3.14 and of cyclohexene at d 5.68,
1.82 and 1.09 (Fig. 3). For purposes of comparison, a
2
natural abundance H NMR spectrum of a solution of
3-phenylpropene in benzene was also run. A GC–MS
was also run on the product of the above reaction, and
In view of our earlier work in which it was found that
attack by phenyl radicals on g3-allyl palladium com-
pounds generally occurred at the metal rather than at
the allylic ligand [14], we anticipated that cyclohexyl
radicals would also add preferentially at the metal.
What was very unclear was what would happen next. It
has been previously demonstrated that photolysis of
compounds of the type [(g3-allyl)Pd(PPh3)2]Cl (allyl ¼
propenyl, 1-methylpropenyl, 2-methylpropenyl, 1-phe-
nylpropenyl, 2-phenylpropenyl) using a high-pressure
Xe-Hg arc lamp results in the formation of the corre-
sponding 1,5-hexadienes [26] via allyl radical interme-
diates. However, control experiments with the 300 W
Sylvania bulb used in the present experiments show that
the palladium allyl complexes used in this study are
stable under the conditions employed here, involving
visible light photolysis.
Earlier work has also shown that reactions of g3-al-
lylpalladium compounds of type A with PPh3 to form
compounds of types B and C occur relatively rapidly
Fig. 3. 2H NMR spectrum showing the resonance at d 3.14 of
CH2@CH–CHD–Ph, the product of the reaction of (g3-1-phenylal-
lyl)Pd(PPh3)Cl with (2,2,6,6-D4-c-hex)Co(DMG)2(py).

S.J. Reid, M.C. Baird / Journal of Organometallic Chemistry 689 (2004) 1257–1264
1261
[14]. Thus many of the reactions carried out here in-
volved mixtures of A and PPh3, although in some cases
preformed PPh3 complexes were used. As anticipated,
similar results were found in both cases.
lyl)Pd(PPh3)2]Cl, produced products exhibiting similarly
1
complex, unassignable H NMR spectra.
A mechanism rationalizing the formation of propene
and cyclohexene formation from reaction of (g3-al-
lyl)Pd(PPh3)Cl is shown in Fig. 4. Following our pre-
vious study [14], it is likely that the initial step involves
reaction of the cyclohexyl radical directly with the pal-
ladium to form the palladium(III) intermediate (g3-al-
lyl)Pd(c-hex)Cl(PPh3). The latter would reasonably
undergo b-hydrogen elimination to form cyclohexene
and the palladium(III) hydride, (g3-allyl)PdHCl(PPh3),
which could subsequently reductively eliminate propene
to give PdI(PPh3)Cl as shown. The latter species would
then undergo further reactions as will be discussed
below.
As an alternative process, one might anticipate that
the palladium cyclohexyl intermediate, (g3-allyl)Pd(III)
(c-hex)Cl(PPh3) would undergo reductive elimination of
the cyclohexyl and allyl groups to form 3-cyclohexyl-
propene. Although product mixtures were examined
carefully for this product, it was not formed to a sig-
nificant extent.
While PdI(PPh3)Cl is proposed as a product in the
mechanism outlined above, this compound was not de-
tected and would be expected to react very rapidly,
possibly with a second cyclohexyl radical to form
PdII(C6H11) (PPh3)Cl as in Fig. 5. The latter might well
undergo b-hydrogen elimination to form PdIIHCl-
(PPh3), a species which could reductively eliminate HCl
to give a palladium(0) product, Pd0(PPh3).
In the reactions of D discussed above, cyclohexene is
formed but not cyclohexane or bicylohexyl. The latter
compounds would also be formed if free cyclohexyl
radicals achieved sufficient concentrations that bimo-
lecular self reactions could occur; cyclohexane and cy-
clohexene would be formed via disproportionation while
bicyclohexyl would be formed via combination reactions
of cyclohexyl radicals [29]. The absence of cyclohex-
ane and bicyclohexyl therefore provides evidence that
3.1. Reactions of (g3-allyl)Pd(PPh3)Cl and [(g3-al-
lyl)Pd(PPh3)2]Cl with cyclohexyl radicals
The photolytic reactions of [(g3-allyl)PdCl]2 and (c-
hex)Co(DMG)2(py), in benzene and benzene-d6 and in
the presence of one equivalent of PPh3 per palladium,
resulted in the formation of propene and cyclohexene in
high yields but essentially no products of coupling of the
cyclohexyl and allyl groups. The reaction of preformed
(g3-allyl)Pd(PPh3)Cl proceeded similarly, also resulting
in the formation of propene and cyclohexene. In all
cases, the reaction mixtures turned red during photolysis
and eventually yielded reddish brown residues on re-
moval of the solvent.
1
The H NMR spectra of the solid product mixtures
were rather complex, and included resonances with un-
usual chemical shifts at d 41.7, 27.0, 21.3, 15.9, )3.3, )13
and )23.9. The resonance at d )13 may alternatively be
attributable to a palladium hydride (see below) [28], but
most of these unusual resonances are reasonably as-
signed as contact shifted resonances [27] of one or more
cobalt(II) byproducts. Unfortunately, there appear to be
no NMR data such species in the literature, nor would
EPR spectroscopy be of use [27]. However, the solutions
are clearly paramagnetic, and while there were many
1
resonances in the d 0–10 part of the H NMR spectra,
many were quite broad and none could be assigned.
Furthermore, all attempts to isolate or identify any of
products, i.e., by extraction, chromatography, or re-
crystallization, proved unsuccessful. Several resonances
were also observed in the 31P NMR spectrum, consistent
with a mixture of products being formed. Reactions
carried out similarly but in the presence of two equiva-
lents of PPh3 per palladium, i.e., involving [(g3-al-
C6H11
Cl
H
Cl
- C6H10
PPh3
PdIII
Cl
PdIII
.
+
C6H11
Pd
PPh3
PPh3
-C3H6
PdI(PPh3)Cl
Fig. 4. Mechanism of the reaction of the cyclohexyl radical with (g3-allyl)PdCl(PPh3).
H
C6H11
Cl
Cl
.
Cl
C6H11
- C6H10
PdI
-HCl
Pd0(PPh3)
PdII
PdII
PPh3
PPh3
PPh3
Fig. 5. Reaction of Pd(PPh3)Cl with a second cyclohexyl radical.

1262
S.J. Reid, M.C. Baird / Journal of Organometallic Chemistry 689 (2004) 1257–1264
cyclohexene is not formed via disproportionation, but
rather via the mechanism proposed in Figs. 4 and 5 or
via photochemical decomposition of D. Control exper-
iments showed that photolysis of D in the absence of
allyl palladium compounds did indeed result in the
formation of some cyclohexene (but no cyclohexane
or bicyclohexyl), results which have been reported
previously [30].
sulted in the formation of isobutene and cyclohexene and
2-phenylpropene and cyclohexene, respectively. Thus
these reactions appear to proceed via the same mechanism
as does the reaction of (g3-allyl)Pd(PPh3)Cl.
It is interesting and surprising that the reaction of
(g3-1-methylallyl)Pd(PPh3)Cl resulted in the formation
of 1-butene but not cis- or trans-2-butene. Similarly the
reaction of [(g3-1-phenylallyl)PdCl]2 produced a mix-
ture of 3-phenylpropene and trans-1-phenylpropene at
25 °C but predominantly the thermodynamically less
stable isomer, 3-phenylpropene, at 6 °C. When the re-
action was carried out with [(g3-allyl)PdCl]2, only the
thermodynamically more stable trans-1-phenylpropene
was formed, in addition to palladium metal, and we
suspected that 3-phenylpropene was the primary prod-
uct formed from b-hydrogen elimination and that the it
was subsequently isomerized by Pd0(PPh3) or palladium
metal to trans-1-phenylpropene. We find that the reac-
tion of [(g3-allyl)PdCl]2 and (c-hex)Co(DMG)2(py), in
the absence of PPh3 resulted in formation of palladium
metal, propene, and cyclohexene, and we have previ-
ously observed that 3-phenylpropene undergoes palla-
dium metal-catalyzed isomerism to 1-phenylpropene,
probably via an allylic intermediate [14].
It has been reported that (g3-allyl)palladium(II) hy-
dride species normally undergo reductive elimination to
form 2-alkenes rather than the thermodynamically less
stable 1-alkene analogues [12c]. Thus the coordinated
hydride migrates to the less hindered terminus of the g3-
allylic ligand, and the 2-alkene is obtained from the re-
ductive elimination process. While it is not at all clear
why the same regiochemistry is not observed here, the
result does imply a mechanism other than reductive
elimination from palladium(II) and may be taken ten-
tatively as implicit evidence for an alternative mecha-
nism, such as is shown in Fig. 4.
3.2. Reaction of (g3-allyl)Pd(PPh3)Cl with cyclohexyl
radicals in the presence of 3-chloropropene
If a palladium(0) species such as ‘‘Pd0(PPh3)’’ is in-
deed formed during the reaction of (g3-allyl)Pd(PPh3)Cl
with cyclohexyl radicals derived from D, it should be
possible to render the reaction catalytic by carrying it
out in the presence of added 3-chloropropene which
would oxidatively add to reform the starting (g3-al-
lyl)Pd(PPh3)Cl. To lend support to the hypothesis con-
cerning the intermediacy of palladium(0) species, we
therefore added 3-chloropropene to a reaction mixture
in which essentially all of the (g3-allyl)Pd(PPh3)Cl had
reacted. The resonances of this compound reappeared as
those of the 3-chloropropene disappeared, and further
photolysis resulted in the formation of more propene.
This process was repeated a second time, and thus the
reaction can be rendered catalytic to at least a limited
extent.
3.3. Reactions of (g3-1-methylallyl)Pd (PPh3)Cl, (g3-2-
methylallyl) Pd(PPh3)Cl and (g3-2-phenylallyl)
Pd(PPh3)Cl with cyclohexyl radicals
In the photolytic reaction of (g3-1-methylal-
1
lyl)Pd(PPh3)Cl with D, a H NMR spectrum of the re-
action mixture showed that the (g3-1-methylallyl)Pd
(PPh3)Cl disappeared while 1-butene and cyclohexene
were formed. Similarly the reactions of (g3-2-methylal-
lyl)Pd(PPh3)Cl and (g3-2-phenylallyl)Pd(PPh3)Cl re-
As a final test of the mechanism proposed above, the
photolysis experiment of [(g3-1-phenyllallyl)PdCl]2was
repeated using (2,2,6,6-D4-c-hex)Co(DMG)2(py). If re-
D
D
D
Cl
D
Cl
Ph
Pd
+
D
D
D
Ph
Pd
PPh3
.
D
PPh3
-
D
D
D
D
Ph
Cl
Ph
Pd
D
PPh3
H
Fig. 6. Formation of CH2@CHCHDPh via the photolytic reaction of (g3-1-phenyllallyl)Pd(PPh3)Cl and (2,2,6,6-D4-c-hex)Co(DMG)2(py).

S.J. Reid, M.C. Baird / Journal of Organometallic Chemistry 689 (2004) 1257–1264
1263
ductive elimination of propene follows b-elimination
from a cyclohexyl-palladium intermediate as in Fig. 5,
then placement of deuterium atoms on the C-2 atoms
should result in the formation of 3-deutero-3-phenyl-
propene, as in Fig. 6.
triphenylphosphine ligands, react with cyclohexyl radi-
cals derived from (c-hex)(py)Co(DMG)2. These reac-
tions proceed via initial attack of the radical at the metal
center, followed by various secondary processes. The
reactions of palladium and platinum allyl complexes
with the cyclohexyl radical involve an initial attack of
the radical with the palladium center, followed by b-
hydrogen elimination and subsequent reductive elimi-
nation of RCH2CH@CH2. Since compounds of the type
[(g3-RC3H4)PdCl]2 (R ¼ alkyl, aryl) are readily pre-
pared from, e.g., alkenes RCH@CHCH3 and allylic
acetates RCH@CHCH2Oac, the chemistry involving
hydrogen transfer from the cyclohexyl radical would
seem in effect to provide a route for both hydrogenolysis
of alkenes or allylic acetate, via g3-allylpalladium
compounds, to terminal alkenes RCH2CH@CH2.
2
The H NMR spectrum of the product mixture ex-
hibited resonances at d 5.8, 1.85 and 1.4, attributable to
partially deuterated cyclohexene, and at d 3.14, attrib-
utable to the deuterium on C-3 of CH2@CH–CHD–Ph
(Fig. 3). There were no vinylic resonances in the region d
4.5–6.5, observed in a natural abundance 2H NMR
spectrum of 3-phenylpropene. The formation specifi-
cally of CH2@CH–CHD–Ph supports the mechanism
proposed for this reaction as it is consistent with a b-
deuterium elimination followed by reductive elimination
(Fig. 6). The sample was also analyzed using GC–MS,
and the mass spectrum obtained showed that 3-phenyl-
propene was formed predominantly with one hydrogen
atom replaced by a deuterium atom.
Acknowledgements
3.4. Reaction of [(g3-allyl)Pt(PPh3)2]Cl with cyclo-
hexyl radicals
We thank the Natural Sciences and Engineering Re-
search Council (Research Grant to MCB) and the
Government of Ontario (Ontario Government Schol-
arship to S.J.R.) for financial support. We also thank
Johnson Matthey for a loan of palladium(II)chloride.
We have pointed out above that a product resonance
at d )13 provides tentative evidence for a hydrido-pal-
ladium species, although the resonance could not in fact
be attributed specifically to a known compound. An
analogous experiment was therefore carried out using
[(g3-allyl)Pt(PPh3)2]Cl, as the anticipated hydride in this
case, PtHCl(PPh3)2, is a stable compound which would
be isolable [23]. The reaction of [(g3-allyl)Pt(PPh3)2]Cl
with (c-hex)Co(DMG)2(py) was therefore carried out,
the initially formed slurry forming a red solution.
Evaporation of the solvent and analysis of the solid
residue by 1H NMR spectroscopy showed that no allylic
products were present, but the spectrum did exhibit a
resonance in the hydride region (d )15), attributable to
PtHCl(PPh3)2 [23] although spin–spin coupling to the
phosphorus nuclei was not observed. The subsequent
addition of free PPh3 to a solution of PtHCl(PPh3)2
prepared as in the literature [23] showed that spin-spin
coupling phosphorus does indeed disappear under these
conditions, presumably because of exchange between
free and coordinated PPh3. When the reaction of [(g3-
allyl)Pt(PPh3)2]Cl with (c-hex)Co(DMG)2(py) was per-
formed on an NMR scale, propene and cyclohexene
were formed in addition to PtHCl(PPh3)2. Thus the
reaction of [(g3-allyl)Pt(PPh3)2]Cl with (c-hex)-
Co(DMG)2(py) appears to proceed via a mechanism
similar to that of the palladium system.
References
[1] (a) J.K. Stille, in: F.R. Hartley, S. Patai (Eds.), The Chemistry of
the Metal–Carbon Bond, 2, Wiley and Sons, Chichester, 1985
(Chapter 9);
(b) L.M. Rendina, R.J. Puddephatt, Chem. Rev. 97 (1997) 1735.
[2] A. Krief, A.M. Laval, Chem. Rev. 99 (1999) 745.
[3] (a) J.K. Kochi, J.W. Powers, J. Am. Chem. Soc. 92 (1970) 137;
(b) R. Sustmann, R. Altevogt, Tetrahedron Lett. 22 (1981) 5167.
[4] (a) R.L. Sweaney, D.S. Comberrel, M.F. Dombourian, N.A.
Peters, J. Organomet. Chem. 216 (1981) 57;
ꢀ
ꢀ
(b) F. Ungvary, L. Marko, Organometallics 1 (1982) 1120;
ꢀ
ꢀ
(c) T.M. Bockman, J.F. Garst, R.B. King, L. Marko, F. Ungvary,
J. Organomet. Chem. 279 (1985) 165;
(d) R.M. Bullock, E.G. Samsel, J. Am. Chem. Soc. 112 (1990)
6886.
[5] (a) J.W. Connolly, Organometallics 3 (1984) 1333;
ꢀ
ꢀ
(b) F. Ungvary, L. Marko, Organometallics 3 (1984) 1466;
(c) B. Wassink, M.J. Thomas, S.C. Wright, D.J. Gillis, M.C.
Baird, J. Am. Chem. Soc. 109 (1987) 1995;
(d) T.A. Shackleton, S.C. Mackie, S.B. Fergusson, L.J. Johnston,
M.C. Baird, Organometallics 9 (1990) 2248.
[6] (a) E. Baciocchi, B. Floris, E. Muraglia, J. Org. Chem. 58 (1993)
2013;
(b) H.-G. Schmalz, S. Siegel, J.W. Bats, Angew. Chem. Int. Ed.
Engl. 34 (1995) 2383;
(c) H.-G. Schmalz, S. Siegel, A. Schwarz, Tetrahedron Lett. 17
(1996) 2947.
[7] (a) C.A.G. Carter, R. McDonald, J.M. Stryker, Organometallics
18 (1999) 820;
4. Conclusions
(b) S. Ogoshi, J.M. Stryker, J. Am. Chem. Soc. 120 (1998) 3514;
(c) G.L. Casty, J.M. Stryker, J. Am. Chem. Soc. 117 (1995) 7814.
[8] C.A. Merlic, D. Xu, J. Am. Chem. Soc. 113 (1991) 9855.
It has been shown that palladium allyl and platinum
allyl complexes, containing zero, one or two coordinated

1264
S.J. Reid, M.C. Baird / Journal of Organometallic Chemistry 689 (2004) 1257–1264
[9] G.G. Melikyan, O. Vostrowsky, W. Bauer, H.J. Bestmann, M.
Khan, K.M. Nicholas, J. Org. Chem. 59 (1994) 222.
[10] (a) D.P. Curran, in: Comprehensive Organic Synthesis, 4,
Pergamon Press, New York, 1991, p. 715;
(d) For a recent review, see M. Tada, Reviews on Heteroatom
Chemistry 20 (1999) 97.
[16] S.J. Reid, N.T. Freeman, M.C. Baird, J. Chem. Soc. Chem.
Commun. (2000) 1777.
(b) B. Giese, Radicals in Organic Synthesis: Formation of
Carbon–Carbon Bonds, Pergamon Press, Oxford, 1986;
(c) D.P. Curran, N.A. Porter, B. Giese, Stereochemistry of
Radical Reactions, VCH, Weinhem, 1996;
[17] (a) W.T. Dent, R. Long, A.J. Wilkinson, J. Chem. Soc. (1964)
1585;
(b) M. Sakakibara, Y. Takahashi, S. Sakai, Y. Ishii, J. Chem.
Soc., Chem. Commun. (1969) 396;
(d) J.A.M. Simoes, A. Greenberg, J.F. Liebman, Energetics of
Organic Free Radicals, Blackie Academic and Professionals, 1996;
(e) J.E. Leffler, An Introduction to Free Radicals, Wiley, 1993;
(f) H. Fischer, L. Radom, Angew. Chem., Int. Ed. Engl. 40 (2001)
1340.
(c) K.C. Ramey, G.L. Statton, J. Am. Chem. Soc. 88 (1966) 1327;
(d) K.C. Ramey, G.L. Statton, J. Am. Chem. Soc. 88 (1966) 4837.
[18] H.C. Volger, K. Vrieze, J. Organomet. Chem. 9 (1967) 527.
[19] V.N. Sokolov, V.M. Khvostic, I.Ya. Poddubyni, G.P. Kondra-
tenkov, J. Organomet. Chem. 54 (1973) 361.
[11] For a recent review, see K.E. Torraca, L. McElwee-White, Coord.
Chem. Rev. 206–207 (2000) 469.
[20] R.L. Batstone-Cunningham, H.W. Dodgen, J.P. Hunt, D.M.
Roundhill, J. Organomet. Chem. 289 (1985) 431.
[21] M.C. Ashraf, T.G. Burrowes, W.R. Jackson, Aust. J. Chem. 29
(1976) 2643.
[12] (a) S.A. Godleski, in: B.M. Trost, I. Fleming (Eds.), Compre-
hensive Organic Synthesis, vol. 4, Pergamon Press, Oxford, 1991,
p. 585;
[22] R.O.C. Norman, C.B. Thomas, G. Watson, J. Chem. Soc. Perkin
Trans. II (1980) 1099.
(b) F. Guibe, Tetrahedron 54 (1998) 2967;
(c) J. Tsuji, Palladium Reagents and Catalysis: Innovations in
Organic Synthesis, Wiley, Chichester, 1995;
[23] (a) J. Chatt, B.L. Shaw, J. Chem. Soc. (1962) 5075;
(b) A.F. Clemmit, F. Glockling, J. Chem. Soc. A (1969) 2163;
(c) W. McFarlane, J. Chem. Soc., Dalton Trans. (1974) 325.
[24] S. Ternieden, D. Muller, K. Muller, Liq. Cryst. 26 (1999) 759.
(d) P.J. Harrington, in: E.W. Abel, F.G.A. Stone, G. Wilkinson
(Eds.), Comprehensive Organometallic Chemistry II, 12, Elsevier,
New York, 1995, p. 797.
€
[25] B. Giese, J. Hartung, J. He, O. Huter, A. Koch, Angew. Chem.
[13] (a) S.G. Cohen, C.H. Wang, J. Am. Chem. Soc. 75 (1953) 5504;
(b) R.R. Bridger, G.A. Russell, J. Am. Chem. Soc. 85 (1963) 3754;
(c) R.G. Kryger, J.P. Lorand, N.R. Stevens, N.R. Herron, J. Am.
Chem. Soc. 99 (1977) 7589.
Int. Ed. Eng. 28 (1989) 325.
[26] G. Vermeersch, J. Marko, J. Muzart, J. Chem. Soc., Perkin Trans.
II (1986) 383.
[27] I. Bertini, C. Luchinat, G. Parigi, Solution NMR of Paramagnetic
Molecules, Elsevier, Amsterdam, 2001.
[14] S.J. Reid, M.C. Baird, J. Chem. Soc., Dalton Trans. (2003) 3975.
[15] (a) B.P. Branchaud, M.S. Meier, M.N. Malekzadeh, J. Org.
Chem. 52 (1987) 212;
[28] H. Munakata, M.L.H. Green, Chem. Commun. (1970) 881.
[29] Z.B. Alfassi, in: Z.B. Alfassi (Ed.), Chemical Kinetics of Small
Organic Radicals, 1, CRC Press, Boca Raton, FL, 1988, p. 129.
[30] (a) B. Giese, J. Hartung, Chem. Ber. 124 (1991) 387;
(b) J. Hartung, B. Hertel, F. Trach, Chem. Ber. 126 (1993)
1187.
(b) M. Rangel, T. Arcos, B. de Castro, Organometallics 18 (1999)
3451;
(c) G.N. Schrauzer, R.J. Windgassen, J. Am. Chem. Soc. 88
(1966) 3738;