Simplifying iron-phosphine catalysts for cross-coupling reactions

Article abstract of DOI:10.1002/anie.201207868

Any old iron? Iron catalysts based on the widely available diphosphine ligand bis(diphenylphosphino)ethane have not previously fared particularly well in iron-catalyzed cross-coupling processes. However, this turns out not to be due to any inherently poor performance associated with the ligand, but rather the need to form a bis-chelate complex, either before or during the formation of an active Fe^I species. Copyright

Full text of DOI:10.1002/anie.201207868

Angewandte
Chemie
DOI: 10.1002/anie.201207868
Iron catalysis
Simplifying Iron–Phosphine Catalysts for Cross-Coupling Reactions**
Robin B. Bedford,* Emma Carter, Paul M. Cogswell, Nicholas J. Gower, Mairi F. Haddow,
Jeremy N. Harvey, Damien M. Murphy, Emily C. Neeve, and Joshua Nunn
Iron–phosphine complexes are emerging as an excellent
correct choice of pre-catalysts, iron complexes of this simple
À
À
choice of catalysts in a range of C C bond-forming process-
ligand can show excellent activity across a range of C C
bond-forming processes.
es,[1] and yet the “rules” for determining selection of the most
appropriate phosphine ligands are, at present, far from clear.
For example, complexes 1 and 2, which are based on
phenylene bis(phosphine) ligands, seem to occupy a privileged
position, displaying, to date, unsurpassed performance in the
cross-coupling of alkyl halides in terms of the variety of
nucleophilic substrates that they tolerate.[2–6] But are these
systems genuinely unique or can simpler catalysts based on
cheaper, more widely available bisphosphines be exploited
instead? We now show that an understanding of the likely
nature of the active catalyst allows the development of far
simpler catalysts based on the widely available and inex-
pensive ligand bis(diphenylphosphino)ethane (dppe).
Iron(I) was proposed as a possible active oxidation state in
iron-catalyzed cross-coupling reactions by Kochi and Tamura
over 40 years ago,[7] whereas more recent mechanistic and
computational evidence in favor of FeI was reported by
Norrby and coworkers.[8] We recently showed that FeI is the
most kinetically reasonable lowest oxidation state in the
catalytic cycle for the coupling of aryl zinc reagents with alkyl
halides using pre-catalyst 1.[9] We isolated the FeI complexes 3
and 4, and demonstrated that 3b is likely to be an on-cycle
intermediate.[10] In contrast with pre-catalyst 1, dppe does not
readily form a bis-chelate complex, [FeCl2(dppe)2],[11] thus the
formation of any putative iron(I) intermediates of the form
[FeX(dppe)2] (5) would require not only reduction of an
[FeX2(dppe)] (6) precursor but also the concurrent coordi-
nation of another equivalent of dppe. Other iron-based
species formed in the meantime may show either poor
activity or decreased selectivity towards the cross-coupled
product or both. We reasoned that if this is the case, then
facilitating the formation of bis-chelate complexes prior to the
catalytic reaction may allow good activity.
The data in Table 1 provide support for this hypothesis. In
the absence of added dppe, 78% of the cross-coupled product
9a was obtained, along with significant amounts of competing
homo-coupled products 10 and 11. Pre-treating 6a with the
diarylzinc reagent prior to addition of the benzyl bromide led
to a significant worsening of performance. Employing a mix-
ture of 6a and one equivalent of dppe gave significantly
improved performance and, in this case, pre-reducing the
mixture with the diarylzinc reagent improved activity and
selectivity further still.
The best performance was seen with the pre-formed
Fe(I)–bis(dppe) complex 5a, which gave excellent conversion
to the cross-coupled product 9a.[12] Unfortunately, the air-
sensitivity of complex 5a and the relative complexity of its
synthesis, detract from its utility. Accordingly, we next
examined the use of the easily handled iron(II) bis-dppe
complex 12,[13] which was readily prepared in one pot from
commercially available [Fe(OH2)6][BF4]2,[14] and showed no
sign of decomposition under air over several days. Complex
12 gave a good amount of 9a, albeit with slightly reduced
selectivity and the need for more forcing reaction conditions
(85 vs. 458C). The latter may in part be due to the incomplete
solubility of the pre-catalyst under the reaction conditions at
the lower temperature. Using the formation of bitolyl as
a proxy for average oxidation state, the reaction of complex
12 with 8a at 858C revealed that it is rapidly reduced to FeI
under the reaction conditions.[15]
Previous screening approaches revealed dppe to be
a thoroughly unremarkable ligand in iron-catalyzed cross-
coupling reactions,[1,2a,b] but we now show that, given the
[*] Prof. Dr. R. B. Bedford, P. M. Cogswell, N. J. Gower,
Dr. M. F. Haddow, Prof. Dr. J. N. Harvey, E. C. Neeve, J. Nunn
School of Chemistry, University of Bristol
Cantock’s Close, Bristol, BS8 1TS (UK)
E-mail: r.bedford@bristol.ac.uk
Dr. E. Carter, Dr. D. M. Murphy
School of Chemistry, Cardiff University
Main Building, Park Place, Cardiff, CF10 3TB (UK)
[**] We thank the EPSRC and Pfizer for funding under both the
collaborative EPSRC Programme for Synthetic Organic Chemistry
with AZ-GSK-Pfizer (E.C.N.) and for funding from Bristol Chemical
Synthesis Doctoral Training Centre (PMC) and the EPSRC for
student (J.N.) and postdoctoral (EC) support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201207868.
Angew. Chem. Int. Ed. 2013, 52, 1285 –1288
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1285

.
Angewandte
Communications
Table 2: Cross-coupling reactions catalyzed by Fe–dppe complexes.
Table 1: Optimization of the dppe-based catalyst.
Entry
RX
7a
R’M
[Fe]
Product, yield [%][a]
Zn(tol)2
[5a]
1[b]
2[c]
Zn(tol)2
[12]
9a, 93%
9a, 91%
Entry [Fe], conditions
Products [mmol][a]
10 11 7a
Zn(tol)2
[12]
3[d]
9a
1
2
3
4
5
6
6a, no pre-reduction
6a, pre-reduction[b]
0.78 0.15 0.23 0.07
0.31 0.03 0.10 0.53
Zn(tol)2
[5a]
4[c]
5[c]
6a+dppe (1 equiv), no pre-reduction 0.87 0.12 0.18
0
0
0
6a+dppe (1 equiv), pre-reduction[b]
0.91 0.04 0.08
0.96 0.03 0.15
5a
12
7c
7d
7e
Zn(tol)2
[12]
9b, 91%
9c, 94%
9d, 78%
0.93 0.07 0.10 0[c]
[a] Determined by 1H NMR spectroscopy (1,3,5-trimethoxybenzene
internal standard). [b] Pre-reduction with 8a. [c] 858C, 4 h.
Zn(tol)2
[5a]
6[b]
7[c]
Zn(tol)2
[12]
Zn(tol)2
[5a]
8[c]
9[c]
Zn(tol)2
[12]
Na[BPh4]
[5a]
10[e]
11[e]
7a
7a
Next, we briefly explored the application of both the pre-
formed FeI complex 5a and the FeII complex 12 in a selected
range of cross-coupling reactions of benzyl and alkyl halides
with various organometallic partners, many of which have
previously been shown to be amenable to catalysis using
phenylene bis(phosphine)-based catalysts,[2,3a,4a] and these
results are summarized in Table 2. As can be seen, 5a is
highly effective in the coupling of alkyl bromides and
chlorides with zinc, aluminium, boron, and indium-based[16]
nucleophiles, with good to excellent spectroscopic yields of
the desired cross-coupled products, in most cases. The
tolerance of the reactions to sensitive functional groups,
such as esters, cyanides, and aryl bromides, is noteworthy. The
more easily prepared and handled complex 12 performs at
least as well as 5a in most of the reactions selected for
comparison.
Na[BPh4]
[12]
9e, 93%
Na[BPh4]
[5a]
12[e]
13[f]
14[f]
15[c]
16[c]
7d
7a
7a
7a
7a
MgCl[Al(tol)4]
[5a]
MgCl[Al(tol)4]
[12]
MgCl[In(tol)4]
[5a]
MgCl[In(tol)4]
[12]
9a, 90%
9a, 92%
9a, 73%
9a, 81%
Zn(C6H3F2)2
[5a]
17[g]
These catalytic data demonstrate that dppe-based cata-
lysts can show excellent activity in reactions where previously
dpbz-based catalysts (dpbz = 1,2-bis(diphenylphosphino)ben-
zene) were found to be optimal. Does this demonstrate that
the active iron(I) species are electronically comparable in
both cases, or is the similarity in performance merely
coincidental? To address this question, we next turned our
attention to ascertaining the molecular and electronic struc-
tures of FeI–dppe species 5a and the related bromide
complex, 5b.
Complex 5a was reported nearly 40 years ago, but no
structural information was provided.[17] We prepared both 5a
and its bromide analogue 5b in good yields by the reaction of
the corresponding [FeX2(dppe)] complex and dppe with two
equivalents of benzylmagnesium chloride and bromide,
respectively, in THF.[14] The X-ray crystal structures of both
MgCl[Al(tol)4]
[5a]
18[f]
19[f]
7 f
7 f
MgCl[Al(tol)4]
[12]
9h, 91%
MgCl[Al(tol)4]
[5a]
20[f]
MgCl[Al(tol)4]
[5a]
21[f]
22[f]
bromodecane, 7h
7h
MgCl[Al(tol)4]
[12]
9j, 78%
[a] Yield determined by 1H NMR spectroscopy (1,3,5-MeO3C6H3 internal
standard). [b] 458C, 4 h. [c] 858C, 4 h. [d] 858C, 16 h. [e] Zn(tol)2 co-
catalyst (10 mol%), 858C, 4 h. [f] 808C, 48 h. [g] 608C, 3 h.
1286
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1285 –1288

Angewandte
Chemie
calculations) are listed in Table S3.[14] The agreement between
experiment and theoretical values is excellent, and both
reveal a rhombic g tensor, which is consistent with a low spin
(S = 1= ) FeI center. Slightly larger Dg shifts were observed in
2
the EPR spectrum of 5b compared to 5a (giso = 2.093 vs.
2.076), and this trend was reflected in the DFT calculations.
The EPR spectra are complicated considerably by the
numerous and overlapping superhyperfine features originat-
ing from the 31P (I = 1/2) and 35,37Cl/79,81Br (I = 3/2) nuclei.
According to the DFT calculations, the two axial 31P nuclei
are equivalent and produce large hyperfine couplings and the
two remaining equatorial 31P nuclei possess smaller couplings.
These results were confirmed by the EPR spectra for both
complexes (Table S4).[14] Although hyperfine couplings to the
Cl or Br nuclei were observed experimentally (Figure 2), the
quadrupole couplings could not be determined. However, it
appears that the underlying quadrupole interactions may be
responsible for the considerable linewidths and distortions
observed in the EPR spectra.
In summary, the structural, electronic, DFT, and prelimi-
nary catalytic data for the iron(I) dppe complexes bear a very
close similarity to the analogous phenylenebis(phosphine)-
containing species, thus belying any suggestion that the latter
ligands are in any way electronically privileged. We attribute
the relatively poor activity associated with dppe in previous
reports to the way that the catalysts are formed in situ in
screening methods, and caution that such approaches to
catalyst discovery may lead to excellent candidates being
inadvertently overlooked. This highlights the need to develop
a more thorough understanding of the basic mechanisms of
iron catalyzed cross-coupling reactions, as this will act as
a powerful aid in the rational development of new catalysts
and catalyzed processes.
Figure 1. Left: Single crystal X-ray structure of complex 5a, hydrogen
atoms and THF solvate omitted for clarity, thermal ellipsoids set at
50% probability. Right: Calculated SOMO of low-spin complex 5a,
with the phosphine ligand residues shown in wireframe.
complexes were determined and the structure of 5a is shown
in Figure 1 (the isostructural complex 5b is shown in the
Supporting Information, Figure S3).[14] Both complexes adopt
a distorted trigonal bipyramidal structure, which is very
similar to those observed for the analogous dpbz-containing
complexes 3a and 3b,[9] although in the present molecules, the
halide atoms lie much closer to the approximate symmetry
axis of the system.
The solid state magnetic moments for complexes 5a and b
were determined to be 1.8 and 1.9 BM at room temperature,
respectively, which is consistent with a single unpaired
electron, thus suggesting that the complexes are low spin in
the solid state.
DFT calculations,[14] assuming a low-spin (S = 1= ) charac-
2
ter, return optimized structures for both complexes that are in
close agreement with the crystal structure data. Figures 1 and
S4[14] show the calculated SOMOs for 5a and 5b, respectively.
In each case, the Mulliken spin density corresponding to the
unpaired electron is mostly located on the iron (0.87 and 0.88
for 5a and 5b, respectively), with only very small contribu-
tions from the ligating P and halide atoms, and the other
ligand atoms.
Received: September 28, 2012
Revised: October 26, 2012
Published online: December 6, 2012
The low temperature X-band EPR spectra of 5a and 5b
are shown in Figure 2. The experimental g values (extracted
by simulation) and calculated values (determined by DFT
Keywords: alkyl halides · catalysis · cross-coupling · iron ·
phosphine ligands
.
[1] a) R. B. Bedford, M. Betham, D. W. Bruce, A. A. Danopoulos,
R. M. Frost, M. Hird, J. Org. Chem. 2006, 71, 1104; b) X. Lin, F.
Zheng, F.-L. Qing, Organometallics 2012, 31, 1578.
[2] For use with arylzinc reagents, see Ref. [3a] and: a) R. B.
Bedford, M. Huwe, M. C. Wilkinson, Chem. Commun. 2009,
600; b) T. Hatakeyama, Y. Kondo, Y. Fujiwara, H. Takaya, S. Ito,
E. Nakamura, M. Nakamura, Chem. Commun. 2009, 1216.
[3] For use with arylboron reagents, see: a) R. B. Bedford, M. A.
Hall, G. R. Hodges, M. Huwe, M. C. Wilkinson, Chem.
Commun. 2009, 6430; b) 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; c) T. Hashimoto,
T. Hatakeyama, M. Nakamura, J. Org. Chem. 2012, 77, 1168;
d) 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.
[4] For Al-based nucleophiles, see: a) S. Kawamura, K. Ishizuka, H.
Takaya, M. Nakamura, Chem. Commun. 2010, 46, 6054; b) S.
Kawamura, T. Kawabata, K. Ishizuka, M. Nakamura, Chem.
Commun. 2012, 48, 9376.
Figure 2. Continuous wave X-band EPR spectra (140 K) of complexes
a) 5a and b) 5b recorded in a THF/toluene (15:1) solution. The
corresponding simulations are shown in a’, b’.
Angew. Chem. Int. Ed. 2013, 52, 1285 –1288
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1287

.
Angewandte
Communications
[5] For the alkynylation of alkyl halides, see: T. Hatakeyama, Y.
Okada, Y. Yoshimoto, M. Nakamura, Angew. Chem. 2011, 123,
11165; Angew. Chem. Int. Ed. 2011, 50, 10973.
[12] Complex 5a has recently been proposed to be one of the
catalytic intermediates in the reductive radical cyclization of
organic halides in the presence of NaBH4, see: A. Ekomiꢀ, G.
Lefꢁvre, L. Fensterbank, E. Lacꢂte, M. Malacria, C. Ollivier, A.
Jutand, Angew. Chem. 2012, 124, 7048; Angew. Chem. Int. Ed.
2012, 51, 6942.
[6] For use with aryl Grignard reagents, see: T. Hatakeyama, Y.-i.
Fujiwara, Y. Okada, T. Itoh, T. Hashimoto, S. Kawamura, K.
Ogata, H. Takaya, M. Nakamura, Chem. Lett. 2011, 40, 1030.
[7] a) M. Tamura, J. K. Kochi, J. Am. Chem. Soc. 1971, 93, 1487;
b) M. Tamura, J. Kochi, J. Organomet. Chem. 1971, 31, 289; c) M.
Tamura, J. K. Kochi, Bull. Chem. Soc. Jpn. 1971, 44, 3063.
[8] J. Kleimark, A. Hedstrçm, P.-F. Larsson, C. Johansson, P.-O.
Norrby, ChemCatChem 2009, 1, 152.
[9] 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.
[13] D. L. DuBois, A. Miedaner, Inorg. Chem. 1986, 25, 4642.
[14] See the Supporting Information.
[15] Lower oxidation states are formed, but at too low a rate to be
significant in the catalytic cycle. See Figure S2 and the associated
discussion in the Supporting Information for details.
[16] To the best of our knowledge, the iron catalyzed cross-coupling
of organoindium reagents has not been previously reported,
although the iron-catalyzed annulation of arylindium reagents
and alkynes was recently described, see: L. Adak, N. Yoshikai,
Tetrahedron 2012, 68, 5167.
[10] Complex 4 appears to be off-cycle, see Ref. [9].
[11] For a discussion, see: J. E. Barclay, G. J. Leigh, A. Houlton, J.
Silver, J. Chem. Soc. Dalton Trans. 1988, 2865.
[17] M. Gargano, P. Giannocarro, M. Rossi, A. Sacco, J. Chem. Soc.
Chem. Commun. 1973, 233b.
1288
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1285 –1288