Iron-phosphine, -phosphite, -arsine, and -carbene catalysts for the coupling of primary and secondary alkyl halides with aryl grignard reagents

Article abstract of DOI:10.1021/jo052250+

Simple catalysts formed in situ from iron chloride and a wide range of monodentate and bidentate phosphines and arsines have been screened in the coupling of alkyl halides bearing β-hydrogens with aryl Grignard reagents. The best of these show excellent activity, as do catalysts formed in situ with monodentate trialkyl and triaryl phosphite ligands. N-heterocyclic carbene-based precatalysts, either preformed or made in situ, also show excellent performance.

Full text of DOI:10.1021/jo052250+

Iron-Phosphine, -Phosphite, -Arsine, and -Carbene Catalysts for
the Coupling of Primary and Secondary Alkyl Halides with Aryl
Grignard Reagents
,†
†
‡
§
Robin B. Bedford,* Michael Betham, Duncan W. Bruce, Andreas A. Danopoulos,
|
⊥
Robert M. Frost, and Michael Hird
School of Chemistry, UniVersity of Bristol, Cantock’s Close, Bristol BS8 1TS, U.K.,
Department of Chemistry, UniVersity of York, Heslington, York YO10 5DD, U.K.,
Department of Chemistry, UniVersity of Exeter, Exeter EX4 4QD, U.K., School of Chemistry,
UniVersity of Southampton, Highfield, Southampton SO17 1BJ, U.K., and Kingston Chemicals,
UniVersity of Hull, Cottingham Road, Kingston upon Hull HU6 7RX, U.K.
r.bedford@bristol.ac.uk
ReceiVed October 28, 2005
Simple catalysts formed in situ from iron chloride and a wide range of monodentate and bidentate
phosphines and arsines have been screened in the coupling of alkyl halides bearing â-hydrogens with
aryl Grignard reagents. The best of these show excellent activity, as do catalysts formed in situ with
monodentate trialkyl and triaryl phosphite ligands. N-heterocyclic carbene-based precatalysts, either
preformed or made in situ, also show excellent performance.
Introduction
elimination, whereas the catalyzed reactions are plagued by
â-elimination (eq 2), which tends to give only the corresponding
alkene.
Coupling reactions leading to the formation of new C-C
2
bonds, typically catalyzed by ubiquitous palladium complexes,
1
form the bedrock of many contemporary syntheses. Despite
the undoubted usefulness of such processes, there are still holes
in the general methodologies available currently that can limit
applicability. Intense research is focused on addressing these
shortcomings, and the past few years have seen substantial
advances. One major class of substrate that has proved
particularly problematic in cross-coupling reactions are primary
and secondary alkyl halides bearing â-hydrogens (eq 1).
Recent studies show that the problem of â-elimination is
surmountable. For instance Ni and Pd complexes have been
shown to catalyze the coupling of primary alkyl halide substrates
2
3
4
with appropriate nucleophilic coupling partners, while Co, Ni,
5
,6
and Fe catalysts have all recently shown activity in coupling
reactions of both primary and secondary alkyl substrates,
typically without the formation of large amounts of â-eliminated
byproduct. Building on the seminal observations of Kochi that
7
iron catalysts can be employed in cross-coupling reactions, there
The uncatalyzed reactions are difficult or impossible due to
the exacting requirements of nucleophilic substitution versus
(2) For reviews see: (a) Frisch, A. C.; Beller, M. Angew. Chem., Int.
†
‡
§
|
Ed. 2005, 44, 674. (b) Netherton, M. R.; Fu, G. C. AdV. Synth. Catal. 2004,
346, 1525. (c) C a´ rdenas, D. J. Angew. Chem., Int. Ed. 2003, 42, 384. (d)
Luh, T.-Y.; Leung, M.-k.; Wong, K. T. Chem. ReV. 2000, 100, 3187.
(3) Tsuji, T.; Yorimitsu, H.; Oshima, K. Angew. Chem., Int. Ed. 2002,
41, 4137.
University of Bristol.
University of York.
University of Southampton.
University of Exeter.
⊥
Kingston Chemicals.
(
1) Reviews: (a) Metal-Catalyzed Cross-Coupling Reactions; Diederich,
(4) Zhou, J.; Fu, G. C. J. Am. Chem. Soc. 2004, 126, 1340.
(5) Nagano, T.;. Hayashi, T. Org. Lett. 2004, 6, 1297.
(6) Nakamura, M.; Matsuo, K.; Ito, S.; Nakamura, E. J. Am. Chem. Soc.
2004, 126, 3686.
F., Stang, P. J., Eds.; Wiley-VCH: Weinheim, Germany, 1998. (b) Cross-
Coupling Reactions; Miyaura, N., Ed.; Topics in Current Chemistry Vol.
19; Springer: New York, 2002.
2
10.1021/jo052250+ CCC: $33.50 © 2006 American Chemical Society
1
104
J. Org. Chem. 2006, 71, 1104-1110
Published on Web 12/24/2005

Iron-Based Catalysts for Alkyl Halide Coupling
TABLE 1. Coupling of 4-Tolylmagnesium Bromide with Cyclohexyl Bromide Catalyzed by Iron Catalysts with Monodentate Phosphines,
a
Phosphites, and Arsines
b
conversion to given compound (%)
entry
iron chloride
ligand (2 equiv)
PPh3
1
2
3
4
5
1
2
3
4
5
6
7
8
9
1
1
1
1
FeCl3
FeCl2
FeCl3
FeCl2
FeCl3
72
81
87
68
53
27
35
82
67
82
83
69
83
2
3
6
0
1
2
3
0
10
7
0
0
0
0
0
0
0
1
0
0
0
0
0
1
7
3
3
0
0
1
0
8
3
1
6
2
6
7
8
16
6
26
14
8
16
8
13
9
PCy3
P(o-tolyl)3
PCy2(o-biphenyl)
t
P Bu2(o-biphenyl)
AsPh3
P(OPh)3
P(OC6H3-2,4- Bu2)3
t
0
1
2
3
P(OMe)3
P(OEt)3
0
5
4
i
P(O Pr)3
8
a
Conditions: FeCl3 (0.05 mmol); ligand (0.2 mmol); CyBr (1.0 mmol); MeC6H4MgBr (2.0 mmol); Et2O; reflux, 30 min. b Conversion to products 1-5
determined by GC (mesitylene internal standard).
are several particularly notable reports of the use of iron
precatalysts in the cross-coupling of both primary and secondary
alkyl halides with aryl Grignard reagents (eq 3).
This indeed turns out to be the case, and we report below the
use of simple iron-phosphine, -phosphite, -arsine, and
-carbene precatalysts for the coupling of primary and secondary
alkyl halides bearing â-hydrogens with aryl Grignard reagents.
Results and Discussion
Phosphine, Phosphite, and Arsine Ligands. For the initial
screening of catalyst performance, we chose the reaction outlined
in eq 4 as a typical example of aryl Grignard-secondary alkyl
coupling. Table 1 shows the conversions to the desired coupled
product 1 along with the formation of the â-elimination product,
cyclohexene (2); the hydrodehalogenated product, cyclohexane
Nagano and Hayashi showed that [Fe(acac)3] can be used to
good effect, while Martin and F u¨ rstner showed the ferrate
5
complex [Li(tmeda)2][Fe(C2H4)4] (tmeda ) N,N,N′,N′-tetram-
ethylethylenediamine) to be an effective precursor in the
coupling of a large range of alkyl halides with diverse
(3); and the two homo-coupled products, dicyclohexane (4) and
8
functionality. We found that simple iron-salen type complexes
4
,4′-bitolyl (5) with a range of monodentate phosphine and
9
can also be exploited. Nakamura and co-workers demonstrated
phosphite ligands with both iron(III) and iron(II) chloride. The
iron chloride and ligand were mixed in dichloromethane for 2
that iron(III) chloride can be employed in the presence of
6
appropriate amines, typically tmeda. While this latter method
11
min before addition to the reaction flask. The solvent was then
is particularly attractive due to the simplicity and low cost of
the catalyst and the excellent results obtained, as reported it
suffered from three major limitations. First, a greater than
stoichiometric amount of amine is required, which needs to be
added with the Grignard reagent. Second, the Grignard/amine
mixture must be added very slowly via the use of a syringe
pump and, third, the reactions must be cooled to low temper-
atures. We subsequently found that all these problems are
surmountable: amines can be used in catalytic quantities, the
reactions can be performed at elevated temperatures, and there
removed in vacuo, and diethyl ether was added as the solvent
for the catalytic reaction.
1
0
is no requirement for slow addition of the Grignard reagent.
The data we obtained with a variety of amines under our
conditions provide a significant contrast with the work of
Nakamura and co-workers, strongly suggesting that different
catalytic manifolds are operatiVe, despite the apparent similarity
in precatalyst composition.
Comparing entries 1 and 2, it can be seen that iron(II) chloride
gives a slightly higher conversion to the coupled product than
iron(III), but at the expense of selectivity; greater relative
amounts of homocoupled products 4 and 5 are produced. With
tricyclohexylphosphine the iron(III) chloride shows better
performance than iron(II) (entries 3 and 4), both in terms of
conversion to the desired product and relative amounts of side
products formed. For these reasons the rest of the studies with
in situ formed catalysts were performed using iron(III) chloride.
Tri-o-tolylphosphine (entry 5) is less effective then either PPh3
or PCy3. The dialkyl o-biphenylphosphine ligands tested did
not prove to be particularly effective (entries 6 and 7); indeed
Nakamura et al. reported that phosphine ligands proved
6
ineffective under their conditions. Given the significant changes
in performance observed with the same amine ligands under
different reaction conditions, we wondered whether phosphine
ligands may actually prove effective under a modified protocol.
(
7) Selected reviews: (a) Kochi, J. K. J. Organomet. Chem. 2002, 653,
1
1. (b) Kochi, J. K. Acc. Chem. Res. 1974, 7, 351.
(8) Martin, R.; F u¨ rstner, A. Angew. Chem., Int. Ed. 2004, 43, 3955.
(
9) Bedford, R. B.; Bruce, D. W.; Frost, R. M.; Goodby, J. W.; Hird,
M. Chem. Commun. 2004, 2822.
10) Bedford, R. B.; Bruce, D. W.; Frost, R. M.; Hird, M. Chem.
Commun. 2005, 4161.
(11) We have found that adding the catalysts as a dichloromethane
solution can lead to an enhancement in activity, even for preformed catalysts.
See ref 9.
(
J. Org. Chem, Vol. 71, No. 3, 2006 1105

Bedford et al.
TABLE 2. Coupling of 4-Tolylmagnesium Bromide with Cyclohexyl Bromide Catalyzed by Iron Catalysts with Bidentate Phosphine and
a
Arsine Ligands
b
conversion to given compound (%)
entry
ligand (1 equiv)
Ph2PCH2PPh2
Ph2P(CH2)2PPh2
Ph2P(CH2)3PPh2
Ph2P(CH2)4PPh2
1
2
3
4
5
1
2
3
4
5
6
7
8
9
60
66
88
75
87
91
82
30
82
0
2
1
4
4
2
2
7
6
2
0
0
0
0
0
0
0
0
0
0
0
2
1
1
0
0
0
5
7
0
0
4
5
7
5
2
Ph2P(CH2)5PPh2
Ph2P(CH2)6PPh2
8
cis-Ph2P(CHdCH)PPh2
trans-Ph2P(CHdCH)PPh2
Ph2AsCH2AsPh2
14
17
10
5
c
1
0
DPPF
a
Conditions: As in Table 1. b Conversion to products 1-5 determined by GC (mesitylene internal standard). c DPPF ) 1,1′-bis(diphenylphosphino)ferrocene.
the activity is lower than that obtained with iron(III) chloride
in the absence of added ligand, which gives 39% conversion to
the desired product 1.10 Interestingly, triphenylarsine performs
better than triphenylphosphine under identical conditions (com-
pare entries 1 and 8). Looking across the data for all the mono-
dentate phosphine and arsine donors, there does not seem to be
a specific trend obvious from electronic or steric perspectives.
Phosphite ligands can also be employed to good effect.
Comparing entries 9 and 10, it can be seen that increasing the
bulk of triaryl phosphite ligand has a beneficial effect on both
activity and selectivity. By contrast there does not appear to be
such a trend with trialkyl phosphites; while both trimethyl
phosphite and triisopropyl phosphite perform well, lower
performance is seen with triethyl phosphite (entries 11-13).
Table 2 summarizes the data obtained for the coupling
outlined in eq 4 using bis-phosphine and arsine ligands. As can
be seen, increasing the chain length of the alkyl spacer in the
bidentate phosphines Ph2P(CH2)nPPh2 (n ) 1-6) leads to a
general increase in conversion to the desired coupled product
potentially interacting redox centers may conceivably act to
switch off such a process.
It is interesting to note that we observe essentially no
formation of cyclohexane (3) in any of the reactions with
13
phosphine, arsine or phosphite ligands. This is in contrast with
the use of either the catalysts formed in situ from FeCl with
3
amine ligands or preformed iron(III) salen-type precatalysts.9
,10
Having established that tricyclohexylphosphine, tris(2,4-di-
tert-butylphenyl)phosphite, and 1,6-bis(diphenylphosphino)-
hexane show essentially the best activity in the test reaction,
we next examined their performance in the coupling of a range
of primary and secondary alkyl halides with aryl Grignard
reagents. The results from this study are summarized in Table
3
. In some cases, isolated products could not be obtained pure,
even after two consecutive chromatographic separations.
Comparing entries 1-3, it can be seen that better conversion
to the desired product is typically obtained when cyclohexyl
bromide is used rather than the chloride or iodide counterparts,
irrespective of the choice of catalyst. The general trend appears
to be Br > I > Cl. This is the same pattern that we have
observed previously with Fe-salen systems under similar
(
entries 1-6), although 1,3-bis(diphenylphosphino)propane
shows higher activity than anticipated from the trend. The
structurally rigid ligand cis-Ph2P(CHdCH)PPh2 shows greater
activity than its more flexible alkyl counterpart Ph2P(CH2)2-
PPh2 (compare entries 2 and 7) but at the expense of selectivity.
Perhaps unsurprisingly, given its inability to form chelates, the
alternate isomer trans-Ph2P(CHdCH)PPh2 performs particularly
badly (entry 8). Again the activity here is lower than that
observed in the absence of added ligand.10 Interestingly, the
bidentate arsine ligand Ph2AsCH2AsPh2 fares somewhat better
than its phosphine counterpart (compare entries 1 and 9) in line
with the result obtained with triphenylarsine.
9
conditions. Hayashi observed a similar trend using [Fe(acac)3]
5
(
acac ) acetylacetonato), while the use of Fe-amine systems
tends to give the trend I > Br > Cl irrespective of Fe/amine
stoichiometry or conditions.
6,10
Increasing the nucleophilicity
of the Grignard reagent (entry 4) leads to a slight decrease in
conversion; in this case the least electron-donating, triaryl
phosphite ligand shows a marked improvement in performance
over the other two systems.
Increasing the steric bulk of the Grignard reagent is deleteri-
ous to the reaction with o-tolylmagnesium bromide, giving
significantly reduced conversion to the coupled product (entry
The catalyst formed in situ from FeCl3 and DPPF (1,1′-bis-
(diphenylphosphino)ferrocene) shows no activity (entry 10). It
5). Again the triaryl phosphite ligand proves to be most effective
is interesting to note that while all of the other reactions turn
very dark brown to black on addition of the Grignard, there is
no significant color change in this case, other than a slight
darkening of the yellow color obtained on mixing the ligand
and the iron chloride. The lack of activity with DPPF is perhaps
surprising considering that this ligand is particularly effective
in the palladium- and nickel-catalyzed coupling of aryl Grignard
reagents with aryl halides.12 This inactivity may be a conse-
quence of the fact that in this case coupling probably proceeds
via a radical pathway rather than a classical oxidative-addition/
reductive elimination manifold (vide infra); the presence of two
in this instance. When the steric hindrance is increased further
by the use of (1,3-dimethylphenyl)magnesium bromide, then
no reaction is observed (entry 6). We have previously found
both Fe-amine and Fe-salen systems to be ineffective in this
reaction, and to the best of our knowledge there are no reports
of the coupling of such sterically hindered, di-ortho-substituted
aryl Grignard reagents with secondary alkyl halides.14 This is
an area that obviously needs further attention in the future.
The coupling of 4-methylcyclohexyl bromide with 4-tolyl-
magnesium bromide (entry 7) leads to the formation of both
cis and trans isomers of the coupled product 10, with a similar
10,9
(
12) See, for example: Togni, A.; Hayashi, T. FerrocenesHomogeneous
Catalysis, Organic Syntheses and Materials Science; VCH: Weinheim,
Germany, 1995; Parts 1 and 2.
(13) At most we see trace amounts by GC, representing much less than
1% conversion.
1106 J. Org. Chem., Vol. 71, No. 3, 2006

Iron-Based Catalysts for Alkyl Halide Coupling
TABLE 3. Coupling of Alkyl Halides with Aryl Grignards Using PR3 as Liganda
a
b
Conditions: Alkyl halide (2.0 mmol), ArMgBr (4.0 mmol), FeCl3 (0.1 mmol), ligand (0.1 or 0.2 mmol), Et2O, reflux. Conversion to coupled product
determined by H NMR spectroscopy (mesitylene internal standard). Isolated by column chromatography. d Product could not be separated from bicyclohexyl
side product. Product contains 2,4-di-tert-butylphenol, tris(2,4-di-tert-butylphenyl)phosphite, and 3-bromopentane. Contains 2,4-di-tert-butylphenol.
1
c
e
f
ratio in all cases; the trans isomer is the preferred product. The
use of open-chain alkyl bromides leads to a decrease in
conversion to the desired coupled products, with secondary alkyl
bromides faring worse than primary substrates (entries 8-10).
Carbene Ligands. Palladium complexes with N-heterocyclic
carbene ligands 14 and 15, have proved useful in palladium-
based coupling reactions of alkyl halide substrates.15 While such
catalysts can be preformed, they are often formed in situ by
deprotonation of the corresponding imidazolium or related salts
Given the success enjoyed with phosphine, phosphite, and
arsine ligands, we were keen to see whether carbene ligands
would prove useful in the coupling of alkyl halides with aryl
Grignard reagents. The results from a survey of activity with
varying carbene ligands in the reaction outlined in eq 4 (above)
are collected in Table 4. Preformed carbene adducts of iron-
halides remain rare; one recently published example is the
“CNC”-pyridyl bis(carbene) pincer complex 18. This shows
excellent activity in the test reaction (entry 1), comparable with
some of the best phosphine-, phosphite-, and arsine-containing
systems outlined above. The N,N-dicylcohexyl-substituted car-
bene formed in situ from the salt 17a shows a good conversion
to the coupled product (entry 2) and the di-tert-butyl analogue
formed from 17b displays excellent activity (entry 3). The
conversion obtained here is even higher than those obtained
using the best phosphine-, phosphite-, and arsine-containing
systems. The carbenes formed from the N,N-diaryl-substituted
salts 17c,d show somewhat lower activity (entries 4 and 5) than
their alkyl-substituted counterparts.
1
6
(16 and 17).
(14) Hayashi and co-workers demonstrated that the similarly sized
mesitylmagnesium bromide is able to react with a primary alkyl halide.
See ref 5.
In the examples listed in entries 2-4 we are relying on in
situ deprotonation of the ligand precursor to yield the free
carbene. We could use the free carbenes themselves, but this
can lead to problems with handling due to their air and moisture
(15) (a) Hadei, N.; Kantchev, E. A. B.; O’Brien, C. J.; Organ, M. G.
Org. Lett. 2005, 7, 3805. (b) Arentsen, K.; Caddick, S.; Cloke, F. G. N.;
Herring, A. P.; Hitchcock, P. B. Tetrahedron Lett. 2004, 45, 351. (c) Zhou,
J.; Fu, G. C. J. Am. Chem. Soc. 2003, 125, 12527. (d) Eckhardt, M.; Fu, G.
C. J. Am. Chem. Soc. 2003, 125, 13642. (e) Frisch, A. C.; Rataboul, F.;
Zapf, A.; Beller, M. J. Organomet. Chem. 2003, 687, 403. (f) Kirchhoff, J.
H.; Dai, C.; Fu, G. C. Angew. Chem., Int. Ed. 2002, 41, 1945.
(16) Danopoulos, A. A.; Tsoureas, N.; Wright, J.; Light, M. Organo-
metallics 2004, 23, 166
J. Org. Chem, Vol. 71, No. 3, 2006 1107

Bedford et al.
TABLE 4. Coupling of 4-Tolylmagnesium Bromide with Cyclohexyl Bromide Catalyzed by Iron Catalysts with Carbene Ligandsa
a
Conditions: As in Table 1. b Conversion to products 1-5 determined by GC (mesitylene internal standard).
TABLE 5. Coupling of Alkyl Halides with Aryl Grignards Using Carbenes as Ligandsa
a
Conditions: As for Table 3. b Trans:cis ) 69:31. c Trans:cis ) 70:30.
sensitivity. Instead we opted to examine the use of neutral
carbene precursors that form carbenes by thermal decomposition
under mild conditions rather than deprotonation. Waymouth and
co-workers very recently showed that the 2-(pentafluorophenyl)-
imidazolidine 19a can be used as a simple, air-stable precursor
for the synthesis of both the free carbene and a carbene adduct
of allylpalladium chloride under mild thermolytic conditions,
via loss of pentafluorobenzene.17 Subsequently we showed that
both 19a,b can be used to form carbene adducts of phosphite-
based palladacycles.18 The catalysts formed in situ from 19a
show enhanced performance compared with that formed from
the salt 17c (compare entries 4 and 5). Increasing the steric bulk
of the N-aryl substituent leads to a substantial increase in
performance, with 19b showing excellent activity. Interestingly,
it appears that the carbene precursors 19a,b react with iron(III)
chloride even at room temperature in dichloromethane; the
reactions are accompanied by a very rapid color change from
1
9
yellow to orange-red.
The preformed complex 18 and the catalyst formed in situ
from iron(III) chloride and 17b were then singled out for further
brief testing with selected substrates; the results from this study
are presented in Table 5. In the coupling of 4-tolylmagnesium
bromide with either cyclohexyl chloride or 4-methylcyclohexyl
bromide carbene catalysts are significantly more active than the
best phosphine or phosphite systems (compare with Table 3),
(
17) Nyce, G. W.; Csihony, S.; Waymouth, R. M.; Hedrick, J. L. Chem.
Eur. J. 2004, 10, 4073.
18) Bedford, R. B.; Betham, M.; Blake, M. E.; Frost, R. M.; Horton, P.
N.; Hursthouse, M. B.; L o´ pez-Nicol a´ s, R.-M. Dalton Trans. 2005, 2774.
(
(19) Investigations into the structures of the resultant complexes are
ongoing within our group.
1108 J. Org. Chem., Vol. 71, No. 3, 2006

Iron-Based Catalysts for Alkyl Halide Coupling
SCHEME 1. Simplified Radical Coupling Pathway
SCHEME 2. Coupling of PhMgBr with 14 Using
Representative Catalystsa
a
Conditions: PhMgBr (4.0 mmol), 14 (2.0 mmol), Fe catalyst (5 mol
b
%
), Et2O/THF (3:2), 45 °C (external temp), 30 min. Conversion to 16
determined by H NMR spectroscopy (mesitylene internal standard).
1
SCHEME 3. Coupling of PhMgBr with 17 Using
Representative Catalystsa
while the activity shown by 18 in the coupling with octyl
bromide is comparable with the best phosphine (PCy3). Interest-
ingly, both the preformed pincer complex and the in situ formed
precatalyst show very similar trans selectivities in the coupling
of 4-methylcyclohexyl bromide with 4-tolylmagnesium bromide.
Mechanistic Considerations. For the coupling of alkenyl
halides and Grignard reagents, Kochi proposed a “classical”
coupling cycle based on oxidative addition of the alkenyl halide
to an iron(I) center which generates an iron(III) alkenyl species,
followed by transmetalation and reductive elimination of the
2
0
product. More recently F u¨ rstner and co-workers have invoked
a
Conditions: PhMgBr (4.0 mmol), 17 (2.0 mmol), Fe catalyst (5 mol
b
an Fe(0)/Fe(II) couple in the reaction of aryl halides with
%), Et2O/THF (3:2), 45 °C (external temp), 30 min. Conversion to 18 and
Grignard reagents.21 Hayashi and co-workers also favor a
19 determined by 1H NMR spectroscopy (mesitylene internal standard).
classical coupling mechanism for the coupling of alkyl halides
with aryl Grignard reagents, but evidence has been presented
5
the ring-opened product 4-phenylbutene, 16, is obtained, lending
support to a radical pathway.
Further evidence in favor of an alkyl radical intermediate is
provided by the reaction of phenylmagnesium bromide with
2
3
by both Nakamura and F u¨ rstner to suggest that in this case the
reaction may in fact proceed via a radical process.6,8 In both
cases this includes the observation that the coupling of resolved
6
-bromohexene, 17 (Scheme 3); this predominantly yields the
ring-closed product 18 as well as the simple coupled product
9.
In summary we have shown that iron catalysts with phos-
2
-bromooctane with PhMgBr leads to the formation of racemic
product.
Scheme 1 shows a highly simplified representation of a
1
possible radical-based coupling mechanism. The active iron
species in oxidation state n reacts with the alkyl halide by the
transfer of a single electron to generate an alkyl radical (via
phine, phosphite, arsine, and carbene ligands are all active
catalysts for the coupling of aryl Grignard reagents with primary
and secondary alkyl halide substrates bearing â-hydrogens.
Many of the catalysts examined show excellent activity.
Representative examples of the catalysts give similar results in
coupling experiments designed to highlight the intermediacy
of radical species, implying similar manifolds in all cases. We
are currently investigating the mechanism in greater depth and
also probing whether common catalyst species are formed with
differing ligandssfor instance nanoparticulate iron speciess
and the results from this study will be published in due course.
(n+1)
the intermediate formation of a radical anion) and an [Fe
X]
species. It is possible that the alkyl radical is not free but rather
associated with the iron center.6 Transmetalation with the
Grignard reagent generates an iron-aryl complex which is then
attacked by the alkyl radical to give the product and active
catalyst.
We have previously presented evidence that suggests that the
coupling of alkyl halides with aryl Grignard reagents catalyzed
by iron-amine-based systems proceeds via a radical pathway.
To see whether a similar manifold is adopted by the systems
reported here, we examined the use of representative catalysts
Experimental Section
t
formed in situ from FeCl3 and PCy3 or P(OC6H2-2,4- Bu2)3 in
All catalytic reactions were performed on a Radleys Carousel
reactor. This consists of 12 ca. 45 mL tubes which are fitted with
screw-on Teflon caps that are equipped with valves for the
introduction of inert gas and septa for the introduction of reagents.
The 12 reaction tubes sit in two stacked aluminum blocks; the lower
one fits on a heater-stirrer and can be maintained at a constant
temperature with a thermostat, while the upper block has water
the coupling of phenylmagnesium bromide with (bromomethyl)-
cyclopropane, 14 (Scheme 2). If an oxidative addition pathway
is operative, then it would be expected that the simple coupled
product 15 would form.22 However, this is not the case; instead
(20) Scott Smith, R.; Kochi, J. K. J. Org. Chem. 1976, 41, 502.
(
21) F u¨ rstner, A.; Leitner, A.; M e´ ndez, M.; Krause, H. J. Am. Chem.
Soc. 2002, 124, 13856.
22) Terao, J.; Watanabe, H.; Ikumi, A.; Kuniyasu, H.; Kambe, N. J.
Am. Chem. Soc. 2002, 124, 4222.
(23) For the use of (bromomethyl)cyclopropane as a probe of radical
pathways in coupling reactions see: Ikeda, Y.; Nakamura, T.; Yorimitsu,
H.; Oshima, K. J. Am. Chem. Soc. 2002, 124, 6514, and references therein.
(
J. Org. Chem, Vol. 71, No. 3, 2006 1109

Bedford et al.
circulating which cools the top of the tubes, allowing reactions to
be performed at reflux temperature.
53.5, 125.3, 125.4, 126.0, 130.1, 135.0, 145.8. HRMS (EI). Calcd
+
for C13
H18 [M ]: 174.140 851. Found: 174.140 662. Product
General Method for the Coupling of 4-Tolylmagnesium
Bromide with Bromocyclohexane (Tables 1, 2, and 4). The
appropriate amount of phosphine, phosphite, arsine, or carbene
contaminated with bicyclohexyl.
1-Methyl-4-(4-methylcyclohexyl)benzene, 10 (Table 3, Entry
7). Cyclohexane eluent. Colorless oil, 0.27 g (72%). HRMS (EI).
+
ligand precursor in CH
0.05 mmol) in a Radleys Carousel reaction tube, and then after
standing (2 min) the solvent was removed in vacuo. In the reaction
with complex 18, the precatalyst was introduced as a CH Cl (2
mL) solution and the solvent removed in vacuo. Et O (3 mL) was
2
Cl
2
(2 mL) was added to anhydrous FeCl
3
Calcd for C14
H
20 [M ]: 188.156 501. Found: 188.156 196.
1
(
(a) Trans Isomer. H NMR (270 MHz, CDCl ) δ 0.94 (d, 3H,
3
3J ) 6.4 Hz, Me), 1.51 (m, 9H, CH of Cy), 2.32 (s, 3H, Me),
HH
2
1
3
2
2
2.42 (m, 1H, Cy), 7.11 (s, 4H); C NMR (68 MHz, CDCl ) δ
3
2
21.0, 22.8, 32.5, 34.5, 35.8, 44.0, 126.7, 129.0, 135.3, 145.0.
1
added, and the solution was stirred (∼2 min). CyBr (1.0 mmol)
was added, and the solution stirred for 5 min and then heated to
reflux temperature (external temperature 45 °C; reaction temperature
(b) Cis Isomer. H NMR (270 MHz, CDCl ) δ 1.02 (d, 3H,
3
3J ) 7.2 Hz, Me), 1.70 (m, 9H, CH of Cy), 2.32 (s, 3H, Me),
HH
2
13
2.51 (m, 1H, Cy), 7.11 (s, 4H); C NMR (68 MHz, CDCl ) δ
3
∼
36-38 °C), and 4-MeC
6
H
4
MgBr (1.0 M solution in Et
2
O, 2.0
18.3, 21.0, 27.6, 28.8, 32.0, 43.9, 126.7, 129.0, 135.3, 145.0.
mL) was added in one portion. The reaction was then heated for
1-Methyl-4-(pentan-3-yl)benzene, 11 (Table 3, Entry 8).
Cyclohexane eluent. Colorless oil. 0.14 g (43%); H NMR (400
1
3
5
0 min, quenched with H
mL), and dried (MgSO
Cl , 1.00 mL) was added, and the conversion to products
-6 was determined by GC analysis.
General Method for the Coupling of Aryl Grignard Reagents
2
O (5 mL), extracted with CH
). Mesitylene (internal standard, 0.1439
2
Cl
2
(3 ×
4
MHz, CDCl ) δ 0.81 (t, 6H, CH CH ), 1.68 (m, 4H, CH ), 2.33
3
2
3
2
3
M in CH
2
2
2
(m, 4H including a singlet at 2.38), 7.08 (d, J ) 8.3 Hz, 2H,
HH
+
Ar), 7.14 (d, 8.3 Hz, 2H, Ar). HRMS (EI). Calcd for C H [M ]:
1
2
18
162.140 851. Found: 162.140 668.
Product contains 2,4-di-tert-butylphenol, tris(2,4-di-tert-butylphe-
nyl)phosphite, and 3-bromopentane.
1-(2-Cyclohexylethyl)-4-methylbenzene, 12 (Table 3, Entry
9). Cyclohexane eluent. Colorless oil, 0.27 g (72%); H NMR (400
with Alkyl Halides (Tables 3 and 5). The reactions were performed
as above with appropriate alkyl halide (2.0 mmol), ArMgBr (4.0
mmol), and catalyst (5 mol % Fe). Reactions were quenched (H
mL), extracted with CH Cl
(3 × 5 mL), and dried (MgSO
Mesitylene (internal standard, 0.667 M CH Cl , 1.00 mL) was
2
O,
1
5
2
2
4
).
2
2
MHz, CDCl ) δ 1.02 (m, 2H, CH of Cy), 1.30 (m, 4H, CH of
3
2
2
added; an aliquot (2 mL) was removed from which the solvent was
removed at room temperature under reduced pressure. The residue
2 2
Cy), 1.58 (m, 2H, CH of Cy), 1.81 (m, 5H, CH & CH of Cy),
3
2.40 (s, 3H, Me), 2.67 (t, 2H, J ) 8.3 Hz CH CH Ar), 7.16 (s,
HH
2
2
3
(∼0.7 mL), and the conversion to coupled
br, 4H, Ar); 13C NMR (100 MHz, CDCl ) δ 21.1, 26.5, 26.9, 32.9,
3
was dissolved in CDCl
1
product was determined by H NMR spectroscopy. For selected
examples of each reaction, the organic phases were recombined,
the solvent removed in vacuo and the coupled product isolated by
column chromatography (silica).
33.5, 37.4, 39.7, 128.4, 129.1, 135.0, 140.3. HRMS (EI). Calcd
+
for C H [M ]: 202.172 151. Found: 202.171 664. Contains 2,4-
15
22
di-tert-butylphenol.
1-Methyl-4-octylbenzene, 13 (Table 3, Entry 10). Cyclohexane
1
-Cyclohexyl-4-methylbenzene, 6 (Table 3, Entry 3). Cyclo-
1
3
eluent. Colorless oil. (51%); H NMR (270 MHz, CDCl ) δ 1.03
1
3
hexane eluent. Colorless oil, 0.178 g (51%); H NMR (270 MHz,
CDCl ) δ 1.27 (m, 5H, CH of Cy), 1.81 (m, 5H, CH of Cy), 2.30
s, 3H, Me); 2.45 (m, 1H, CH of Cy); 7.08 (s, br, 4H); C NMR
68 MHz, CDCl ) δ 21.1, 26.4, 27.2, 34.8, 44.4, 126.9, 129.1, 135.3,
(t, 3H, J ) 6.9 Hz, CH ), 1.44 (m, 10H, CH of alkyl chain),
HH
3
2
3
2
2
2
1.74 (m, 2H, CH of alkyl chain), 2.45 (s, 3H, Me), 2.70 (t, 2H,
13
3
13
(
(
1
JHH ) 7.9 Hz), 7.41 (s, br, 4H); C NMR (68 MHz, CDCl ) δ
14.3, 22.9, 27.1, 29.5, 29.6, 29.8, 31.9, 32.1, 35.8, 128.4, 129.1,
135.0, 140.0. HRMS (EI). Calcd for C H : 204.187 801. Found:
3
3
+
45.3. HRMS (EI). Calcd for C13
74.140 762.
-Cyclohexyl-4-methoxybenzene, 7 (Table 3, Entry 4). Toluene
H18 [M ]: 174.140 851. Found:
15
24
1
204.186 998.
1
1
eluent. Colorless oil, 0.282 g (74%); H NMR (270 MHz, CDCl
δ 1.31 (m, 5H, CH of Cy), 1.81 (m, 5H, CH of Cy), 2.44 (m, 1H,
CH of Cy), 3.79 (s, 3H, OMe), 6.84 (d, 2H, JHH ) 8.6 Hz); 7.13
d, 2H, JHH ) 8.6 Hz); C NMR (68 MHz, CDCl
3.8, 55.3, 55.4, 113.7, 127.7, 140.5, 157.7. HRMS (EI). Calcd
3
)
Acknowledgment. We thank the EPSRC for the provision
of an Advanced Research Fellowship for R.B.B. and a PDRA
for M.B., the COMIT Faraday Partnership, Kingston Chemicals,
and EPSRC for support to R.M.F., and Professor John Goodby
for helpful discussions.
2
2
3
3
13
(
4
3
) δ 26.3, 34.8,
+
for C13H18O [M ]: 190.135 765. Found: 190.135 215.
1
-Cyclohexyl-2-Methylbenzene, 8 (Table 3, Entry 5). Cyclo-
Supporting Information Available: 1H and 13C NMR spectra
of the coupled products. This material is available free of charge
via the Internet at http://pubs.acs.org.
1
hexane eluent. Colorless oil, 0.17 g (49%); H NMR (270 MHz,
CDCl of Cy), 1.53 (m, 2H, CH of Cy), 2.15
of Cy), 2.33 (s, 3H, Me), 2.70 (m, 1H, CH of Cy),
.14 (m, 4H); 13C NMR (68 MHz, CDCl
) δ 19.3, 26.3, 33.6, 37.5,
3
) δ 1.39 (m, 3H, CH
m, 5H CH
2
2
(
7
2
3
JO052250+
1110 J. Org. Chem., Vol. 71, No. 3, 2006