Expedient iron-catalyzed coupling of alkyl, benzyl and allyl halides with arylboronic esters

Article abstract of DOI:10.1002/chem.201402174

While attractive, the iron-catalyzed coupling of arylboron reagents with alkyl halides typically requires expensive or synthetically challenging diphosphine ligands. Herein, we show that primary and secondary alkyl bromides and chlorides, as well as benzyl and allyl halides, can be coupled with arylboronic esters, activated with alkyllithium reagents, by using very simple iron-based catalysts. The catalysts used were either adducts of inexpensive and widely available diphosphines or, in a large number of cases, simply [Fe(acac)₃] with no added co- ligands. In the former case, preliminary mechanistic studies highlight the likely involvement of iron(I)-phosphine intermediates. More irons in the fire: Primary and secondary alkyl, benzyl and allyl halides were coupled with arylboronic esters by using very simple iron-based catalysts. These were either adducts of inexpensive and widely available diphosphines or, in a large number of cases, simply [Fe(acac)₃] with no added co-ligands (see scheme; acac=acetylacetonate). In the former case, preliminary mechanistic studies highlight the likely involvement of low-coordinate iron(I)-phosphine intermediates.

Full text of DOI:10.1002/chem.201402174

DOI: 10.1002/chem.201402174
Communication
&
Cross-Coupling
Expedient Iron-Catalyzed Coupling of Alkyl, Benzyl and Allyl
Halides with Arylboronic Esters
Robin B. Bedford,*[a] Peter B. Brenner,[a] Emma Carter,[b] Thomas W. Carvell,[a]
Paul M. Cogswell,[a] Timothy Gallagher,[a] Jeremy N. Harvey,[a] Damien M. Murphy,[b]
Emily C. Neeve,[a] Joshua Nunn,[a] and Dominic R. Pye[a]
pensive, readily available diphosphines (1,2-bis(diphenylphos-
Abstract: While attractive, the iron-catalyzed coupling of
arylboron reagents with alkyl halides typically requires ex-
pensive or synthetically challenging diphosphine ligands.
Herein, we show that primary and secondary alkyl bro-
mides and chlorides, as well as benzyl and allyl halides,
can be coupled with arylboronic esters, activated with al-
kyllithium reagents, by using very simple iron-based cata-
lysts. The catalysts used were either adducts of inexpen-
sive and widely available diphosphines or, in a large
number of cases, simply [Fe(acac)3] with no added co-
ligands. In the former case, preliminary mechanistic stud-
ies highlight the likely involvement of iron(I)–phosphine
intermediates.
phino)ethane (dppe) and 1,3-bis(diphenylphosphino)propane
(dppp)) or, in a large number of cases, phosphine-free systems
can catalyze the coupling of nucleophiles derived in situ from
arylboron pinacol esters and a wide range of primary and sec-
ondary alkyl halides.
Nakamura’s iron-catalyzed coupling of pinacol esters 1 (acti-
vated with alkyllithium reagents) with alkyl halides
2
(Scheme 1)[4a] is a particularly useful process, representing the
Iron-catalyzed cross-coupling reactions have recently enjoyed
a renaissance, having spent most of the seventy years since
their initial publication in relative obscurity.[1,2] One particularly
attractive aspect of iron-catalyzed cross-coupling processes,
aside from the obvious advantages of the low cost and benign
nature of iron, is that they can often tolerate both primary and
secondary alkyl halide substrates, for instance in coupling reac-
tions with aryl nucleophiles.
Scheme 1. Suzuki coupling of alkyl halides. Catalytic reaction conditions:[6]
[a] 408C, 3–4 h (1 mol% Fe); [b] 258C, 4 h (5 mol% Fe); [c] 408C, 3 h
(1 mol% Fe); [d] RT, 24 h (1 mol% Fe, 1 equiv dppp added); 408C, 1 h
(5 mol% Fe); [e] 408C, 1 h (5 mol% [Fe(acac)3]).
While the most widely studied variants of these reactions
are based on aryl Grignard reagents, recently, there has been
a shift towards exploiting more substrate- and media-tolerant
nucleophiles. To this end, we[3] and Nakamura and co-workers[4]
have developed a range of diphosphine-based systems for the
iron-catalyzed coupling of organoboron-based nucleophiles.
Unfortunately, the high cost or the protracted syntheses of the
most effective chelating diphosphine ligands detract from the
synthetic appeal of the resultant catalytic processes. Herein,
we now report that very simple catalysts based on either inex-
current state of the art, since pinacol esters of arylboronic
acids are widely commercially available and easily synthesized
from the parent acids. Therefore, we were delighted to find
that the trivially simple complexes [FeCl2(dppe)] and [FeCl2-
(dppp)] (with an added equivalent of dppp) give excellent
yields of the cross-coupled product 3a,[5] not only with cyclo-
heptyl bromide (2a), but also with the more challenging chlo-
ride analogue 2b. Even more surprising was the observation
that excellent conversion to 3a could be obtained from 2a by
using 5 mol% [Fe(acac)3] (acac=acetylacetonate) with no added
phosphine ligand. [Fe(acac)3] is an inexpensive, free-flowing, air-
and moisture-stable solid, making it a particularly attractive
precatalyst.
[a] Prof. Dr. R. B. Bedford, Dr. P. B. Brenner, T. W. Carvell, P. M. Cogswell,
Prof. Dr. T. Gallagher, Prof. Dr. J. N. Harvey, Dr. E. C. Neeve, J. Nunn,
D. R. Pye
School of Chemistry, University of Bristol
Cantock’s Close, Bristol, BS8 1TS (UK)
E-mail: r.bedford@bristol.ac.uk
[b] Dr. E. Carter, Prof. Dr. D. M. Murphy
School of Chemistry, Cardiff University
Table 1 summarizes the catalytic reactions investigated in
the presence and absence of phosphine ligands. Gratifyingly,
under phosphine-free conditions, cyclic and open-chain secon-
dary alkyl bromides were well tolerated (entries 1, 4, 5, and 9),
Main Building, Park Place, Cardiff CF10 3AT (UK)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201402174.
Chem. Eur. J. 2014, 20, 1 – 5
1
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Communication
ence of either dppe or dppp (entries 3, 11, 13, and 14). Primary
alkyl chlorides remain a challenge, with only 50% conversion
to the cross-coupled product observed with n-octyl chloride in
the presence of dppp (entry 12). Esters and cyano groups were
well tolerated (entries 13, 14, and 20).
Table 1. Iron-catalyzed Suzuki cross-couplings without (methods A and
B) and with (methods C–F) phosphine ligands.
Entry Substrate
Method[a] Product
Spec. yield [%][b]
(isolated)
With regard to the mechanism of the catalytic processes, in
the presence of diphosphines, red homogeneous reaction mix-
tures were observed throughout the reaction. By contrast,
under phosphine-free conditions, black suspensions of iron
nanoparticles formed during the reactions[9] after a variable
length of time, dependent on the nature of the alkyl halide
substrate. Thus, with benzyl bromide 2m, 89% of the cross-
coupled product was obtained before nanoparticle formation
was observed,[10] which is suggestive of a homogeneous cata-
lytic manifold, but with alkyl bromides and allyl or benzyl
chlorides, the nanoparticles formed early in the coupling reac-
tion. Examining the reaction-time profile for the coupling of
cycloheptyl bromide (2a) with 1a, we observed an induction
period of around 2 min, which is coincident with the time nec-
essary for the reaction mixture to change from pale yellow to
black.[10] A black suspension was also formed on leaving out
the electrophile, however, when the magnesium salt was also
omitted, the reaction mixture remained as an orange homoge-
neous solution, suggesting that the salt is necessary for effi-
cient transmetalation.
1
A
>99 (75)
2
3
A
E
42
>99
4
5
6
A
A
C
>99 (65)
>99 (75)
>99 (71)
7
C
99 (85)
8
9
C
A
98 (77)
90 (67)
10
11
12
A
C
F
50
>99 (83)
50
13
14
D
C
84 (60)
97 (60)
To simplify the mechanistic investigations, we focused on
the dppe-containing catalyst system for the rest of the study,
because this remains homogeneous and shows a faster rate in
the coupling of 2a with the preformed borate Li[pinBPh(tBu)]
(5a)[11] than the catalyst derived from [Fe(acac)3] (see Figure S4
in the Supporting Information).[10] The EPR spectrum (recorded
at 140 K, Figure 1) of a catalytic reaction between 2a and 5a,
catalyzed by [FeBr2(dppe)] and one equivalent of dppe,[10] re-
vealed the presence of the previously reported iron(I) complex
[FeBr(dppe)2] (6a, Scheme 2).[3b] Similarly, 6a or the related
iron(I) chloride complex 6b were observed on reaction of
[FeX2(dppe)] (X=Cl or Br) and dppe with excess 5a/MgCl2.[10]
Interestingly, 6b was also produced in the absence of the
15
16
A
B
93 (58)
98 (67)
17
A
96 (66)
18
19
B
>99
A
92 (89)
20
21
22
A
C
C
>99 (89)
89 (58)
88 (45):11
[a] Conditions: A) RX (1 mmol), Li[pinBAr(tBu)] (formed in situ from pinBAr
and tBuLi at 08C, 1.4 mmol), MgBr2 (0.2 mmol), [Fe(acac)3] (5 mol%), RT,
1 h. B) As A, 408C. C) As A with [FeCl2(dppe)] (1 mol%), 408C, 3 h. D) As C
with [FeCl2(dppe)] (3 mol%). E) As
A with [FeCl2(dppp)] and dppp
(1 mol% each), RT, 24 h. F) As E with [FeCl2(dppp)] and dppp (3 mol%
1
each). [b] Determined by H NMR spectroscopy (1,3,5-(MeO)3C6H3 as inter-
nal standard).
giving good to excellent conversions to the desired cross-cou-
pled products. While [Fe(acac)3] was routinely exploited, FeCl3
also worked reasonably well.[7]
To the best of our knowledge, the iron-catalyzed coupling of
allyl halides with arylboron nucleophiles has not been reported
previously,[8] but was achieved readily under phosphine-free
conditions (entries 15 and 16). Benzyl bromides and chlorides
also performed well (entries 17–20) under phosphine-free con-
ditions, but, in general, the reactions with open-chain primary
alkyl bromides or cycloalkyl chlorides fared better in the pres-
Figure 1. EPR spectra (140 K) of a) a sample removed from the catalytic reac-
tion between 2a and 5a catalyzed by [FeBr2(dppe)] at 408C and b) a genu-
ine sample of [FeBr(dppe)2] 6a.
&
&
Chem. Eur. J. 2014, 20, 1 – 5
www.chemeurj.org
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communication
While so far we have not been able to observe the putative
low-coordinate FeI intermediate 7a directly, owing to its high
lability,[14] we reasoned that the bulkier, more conformationally
rigid diphosphine ligand 4a may allow the observation of anal-
ogous but more stable low-coordinate FeI species. This indeed
appears to be the case: the EPR spectra recorded of reaction
mixtures of [FeBr2(4a)] with either 20 equivalents of borate 5a
and four equivalents of MgBr2 (comparable with catalytic con-
ditions but in the absence of an electrophile) or with excess
PhMgBr[15] showed signals consistent with the formation of
Scheme 2. Observation of iron(I).
magnesium salt, but to a lesser extent, suggesting that the salt
facilitates transmetalation.[10]
a low-spin S=1= species, with non-equivalent P environments,
2
Isolated 6b is not only catalytically competent, but it also
appears to show no induction period when used as a catalyst
in the coupling of 2a and 5a and shows the same initial rate
of reaction as [FeCl2(dppe)] with one equivalent of dppe.[10]
The appearance of significant amounts of 6a in the catalytic
reaction is consistent with it being involved in the catalytic
cycle, prior to the rate-determining step. If this is the case,
then 6a could be either the iron-containing intermediate im-
mediately before the rate-determining step, or it could be in
pre-equilibrium with such a species. To differentiate between
these two possibilities, kinetic studies were undertaken.
very tentatively assigned as [FePh(4a)] (7b).[10,16]
Once generated, there are at least two potential fates for
the organic radical produced in the rate-determining step
shown in Scheme 3: it could react with either [FeX2(dppe)] to
generate an organoiron(III) intermediate or with a second
equivalent of 6 to generate an organoiron(II) intermediate,
either of which could then react with the nucleophile and ulti-
mately close the catalytic cycle.[17] At present, it is not possible
to distinguish between these (or other)[18] possibilities and
work is ongoing to establish the full details of the catalytic
manifold.
The rate of reaction of 2a with 5a catalyzed by [FeCl2(dppe)]
was found to show a first-order dependence on the iron com-
plex and an inverse first-order dependence on added dppe, up
to about eight equivalents per Fe, after which point the effect
began to tail off.[10] This suggests that the active catalyst con-
tains only one dppe ligand in the rate-determining step of the
catalytic reaction. The observed rate showed a zero-order de-
pendence on [5a], but an approximately first-order depend-
ence on [2a],[12] indicating that the rate-determining step is as-
sociated with the reaction of the alkyl halide with the iron
In summary, we have developed a highly attractive, expedi-
ent coupling of arylboronic esters with alkyl, benzyl and allyl
halides. When chelating diphosphines are employed, the evi-
dence suggests that the reaction proceeds via iron(I) as the
lowest oxidation state in the manifold.[19] We are currently fur-
ther investigating the full scope and mechanism of the reac-
tion under both phosphine-containing and phosphine-free
conditions.
center. The reaction was accelerated by the addition of up to Acknowledgements
0.20 equivalents (with respect to 2a) of MgBr2, after which
point it was retarded.[13]
We thank the EPSRC and Pfizer for funding under both the col-
We propose that the rate-determining step in the catalytic
cycle is associated with the sequence outlined in Scheme 3,
wherein a pre-equilibrium between the five-coordinate FeI spe-
laborative EPSRC Programme for Synthetic Organic Chemistry
with AZ-GSK-Pfizer (E.C.N) and for funding from Bristol Chemi-
cal Synthesis Centre for Doctoral Training (P.M.C. and D.R.P.)
and the EPSRC for student (J.N.) and post-doctoral (P.B.B., E.C.)
support and for funding from the Bristol Centre for Functional
Nanomaterials (T.W.C.).
Keywords: alkyl halides · catalysis · cross-coupling · iron ·
Suzuki coupling
[1] For early examples, see: a) G. Vavon, P. Mottez, C. R. Hebd. Seances Acad.
Sci. 1944, 218, 557; b) W. C. Percival, R. B. Wagner, N. C. Cook, J. Am.
Chem. Soc. 1953, 75, 3731; c) M. Tamura, J. K. Kochi, J. Am. Chem. Soc.
1971, 93, 1487; d) M. Tamura, J. Kochi, Synthesis 1971, 303; e) S. M. Neu-
mann, J. K. Kochi, J. Org. Chem. 1975, 40, 599; f) R. S. Smith, J. K. Kochi,
J. Org. Chem. 1976, 41, 502.
[2] For reviews. see: a) C. Bolm, J. Legros, J. Le Paih, L. Zani, Chem. Rev.
2004, 104, 6217; b) A. Fꢁrstner, R. Martin, Chem. Lett. 2005, 34, 624;
c) B. D. Sherry, A. Fꢁrstner, Acc. Chem. Res. 2008, 41, 1500; d) W. M. Cza-
Scheme 3. The proposed sequence incorporating the rate-determining step.
cies 6 and a three-coordinate FeI species 7a is followed by the
comparatively slow reaction of 7a with the alkyl halide. Provid-
ing that the pre-equilibrium favors 6, the proposed process is
consistent with: a) the observation of 6a under catalytic condi-
tions, b) the inverse order dependence on added dppe, c) the
approximately first-order dependence on the alkyl halide, and
d) the intermediacy of organic radicals, as supported by the
products observed from the radical probe experiments in en-
tries 21 and 22 of Table 1.
ˇ
plik, M. Mayer, J. CvengroS, A. J. von Wangelin, ChemSusChem 2009, 2,
396.
[3] a) R. B. Bedford, M. A. Hall, G. R. Hodges, M. Huwe, M. C. Wilkinson,
Chem. Commun. 2009, 6430; b) R. B. Bedford, E. Carter, P. M. Cogswell,
N. J. Gower, M. F. Haddow, J. N. Harvey, D. M. Murphy, E. C. Neeve, J.
Chem. Eur. J. 2014, 20, 1 – 5
www.chemeurj.org
3
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Communication
Nunn, Angew. Chem. 2013, 125, 1323; Angew. Chem. Int. Ed. 2013, 52,
1285.
[15] In this case, we also saw signals in the EPR spectrum for a second
minor component that was also S=1= , but showed no hyperfine cou-
2
pling to phosphorus. The spectrum was very similar to one recorded
previously for the reaction of FeCl3 with 4-tolMgCl (see: R. B. Bedford,
P. B. Brenner, E. Carter, P. M. Cogswell, M. F. Haddow, J. N. Harvey, D. M.
Murphy, J. Nunn, C. H. Woodall, Angew. Chem. 2014, 126, 1835–1839;
Angew. Chem. Int. Ed. 2014, 53, 1804–1808) and identical to a spectrum
obtained on reacting FeBr2 with 20 equivalents PhMgBr, suggesting
that the ligand 4a can be displaced by a large excess of Grignard to
give a phosphine-free iron(I) complex. Interestingly, the last experiment
also revealed that a soluble iron(I) species is accessible under condi-
tions in which Fe nanoparticles are formed.
[4] a) T. Hatakeyama, T. Hashimoto, Y. Kondo, Y. Fujiwara, H. Seike, H.
Takaya, Y. Tamada, T. Ono, M. Nakamura, J. Am. Chem. Soc. 2010, 132,
10674; b) T. Hatakeyama, T. Hashimoto, K. K. A. D. S. Kathriarachchi, T.
Zenmyo, H. Seike, M. Nakamura, Angew. Chem. 2012, 124, 8964; Angew.
Chem. Int. Ed. 2012, 51, 8834.
[5] As observed previously (Ref. [4a]), the boron-containing by-product is
tBuBPin, as determined by GC-MS and 11B NMR analysis in comparison
with a genuine sample. No reaction was observed in the absence of
MgBr2. Similarly, replacing MgBr2 with ZnCl2 proved ineffective.
[6] 1a and tBuLi were premixed in THF (À40–08C) to yield Li[pinBPh(tBu)]
in situ; then, this mixture was added to the THF solution of the precata-
lyst, alkyl halide and magnesium bromide.
[7] Repeating the reaction in Table 1, entry 1 with FeCl3 gave 87% of 3a.
The equivalent reaction with high-purity FeCl3 (99.99%) gave 89% of
3a. Leaving the iron precatalyst out gave no activity, while using
[Fe(acac)3], tBuLi, magnesium bromide diethyl etherate and substrates
from more than one supplier in each case did not give an appreciable
change in performance.
[16] The inequivalence of the P-donors could be due to the adoption of T-
shaped geometry, but is more likely to be due to either coordination of
solvent or a secondary interaction with one of the ligands. For a review
of related effects in 3-coordinate PtII complexes, see: M. A. OrtuÇo, S.
Conejero, A. Lledꢂs, Beilstein J. Org. Chem. 2013, 9, 1352; for an exam-
ple of a secondary interaction in a low-coordinate iron(I) complex, see:
C. Ni, B. D. Ellis, T. A. Stich, J. C. Fettinger, G. J. Long, R. D. Britta, P. P.
Power, Dalton Trans. 2009, 5401.
[17] For a recent discussion of these possibilities in similar nickel-catalyzed
processes, see: J. Breitenfeld, J. Ruiz, M. D. Wodrich, X. Hu, J. Am. Chem.
Soc. 2013, 135, 12004.
[18] For instance, the cross-coupled product could be produced by reaction
of the organic radical directly with the nucleophile.
[8] For the coupling of aryl Grignards with allyl halides, see: A. Fꢁrstner, R.
Martin, H. Krause, G. Seidel, R. Goddard, C. W. Lehmann, J. Am. Chem.
Soc. 2008, 130, 8773.
[9] For TEM image and EDX analysis, see the Supporting Information.
[10] For details, see the Supporting Information.
[19] For recent suggestions of iron(I) in aryl-nucleophile cross-couplings, see
Refs. [3b,14] and: a) A. Hedstrçm, E. Lindstedt, P.-O. Norrby, J. Organo-
met. Chem. 2013, 748, 51; b) C. J. Adams, R. B. Bedford, E. Carter, N. J.
Gower, M. F. Haddow, J. N. Harvey, M. Huwe, M. A. Cartes, S. M. Mansell,
C. Mendoza, D. M. Murphy, E. C. Neeve, J. Nunn, J. Am. Chem. Soc. 2012,
134, 10333; c) A. Hedstrçm, U. Bollmann, J. Bravidor, P.-O. Norrby, Chem.
Eur. J. 2011, 17, 11991; d) J. Kleimark, A. Hedstrçm, P.-F. Larsson, C. Jo-
hansson, P.-O. Norrby, ChemCatChem 2009, 1, 152.
[11] Nakamura and co-workers demonstrated the formation of the borates
Li[pinBR(Ar)] 5 from 1 and lithium reagents under catalytically relevant
conditions, see Ref. [4].
[12] Rate dependence on [2a] determined to be ꢀ0.7 order.
[13] As yet, we have not been able to establish why the [MgBr2] (or any
Grignard reagent produced in situ from it) has an accelerating influence
on the rate-determining step, at least at lower loadings.
[14] [FeX(dppe)] species appear to be unstable with respect to ligand re-
distribution. Thus, the reaction of [FeCl2(dppe)] with 5a shows [FeCl-
(dppe)2] (6b) and not [FeCl(dppe)] in the EPR spectrum (see Figure S7
in the Supporting Information).
Received: February 13, 2014
Published online on && &&, 0000
&
&
Chem. Eur. J. 2014, 20, 1 – 5
www.chemeurj.org
4
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communication
COMMUNICATION
More irons in the fire: Primary and sec-
ondary alkyl, benzyl and allyl halides
were coupled with arylboronic esters by
using very simple iron-based catalysts.
These were either adducts of inexpen-
sive and widely available diphosphines
or, in a large number of cases, simply
[Fe(acac)3] with no added co-ligands
(see scheme; acac=acetylacetonate). In
the former case, preliminary mechanistic
studies highlight the likely involvement
of low-coordinate iron(I)–phosphine
intermediates.
&
Cross-Coupling
R. B. Bedford,* P. B. Brenner, E. Carter,
T. W. Carvell, P. M. Cogswell, T. Gallagher,
J. N. Harvey, D. M. Murphy, E. C. Neeve,
J. Nunn, D. R. Pye
&& – &&
Expedient Iron-Catalyzed Coupling of
Alkyl, Benzyl and Allyl Halides with
Arylboronic Esters
Chem. Eur. J. 2014, 20, 1 – 5
www.chemeurj.org
5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ