Iron-catalyzed coupling of aryl grignard reagents with alkyl halides: A competitive hammett study

Article abstract of DOI:10.1002/chem.201100467

Competing for electrophiles: The elusive iron-catalyzed C-C coupling reaction was investigated and analyzed by a Hammett study of the nucleophilic partner. This required finding conditions in which the iron catalyst is stable in the presence of an excess of the Grignard reagent. The selectivity- determining step seems to be the transmetalation, occurring either before or after the oxidative addition step (see scheme).

Full text of DOI:10.1002/chem.201100467

COMMUNICATION
DOI: 10.1002/chem.201100467
Iron-Catalyzed Coupling of Aryl Grignard Reagents with Alkyl Halides: A
Competitive Hammett Study**
Anna Hedstrçm,[a] Ulla Bollmann,[b] Jenny Bravidor,[b] and Per-Ola Norrby*[a]
À
The use of transition-metal catalysis to form new C C
bonds is an important tool in organic synthesis. Palladium-
À
and nickel-catalyzed C C bond-forming reactions have been
extensively explored and are well understood.[1] The use of
iron as the catalyst has gained much less attention, despite
the innovative work of Kochi et al.[2] in the 1970s. Recently,
several groups have turned their attention towards iron-cat-
alyzed coupling reactions;[3] this growing interest in iron is
due to its environmentally benign character, low cost, and
non-toxicity. Furthermore, iron seems to allow all possible
combinations of carbon hybridization in the coupling reac-
tion.
Scheme 1. Competitive coupling of aryl Grignard reagents; I.S.=internal
standard.
We have recently published a mechanistic investigation
into the iron-catalyzed coupling reaction between an aryl
electrophile and an alkyl Grignard reagent.[4] A combination
of reaction monitoring, a Hammett competition study, and
DFT calculations indicated that the oxidation state of the
catalytically active iron species is FeI and that the oxidative
addition of the aryl halide is the rate-limiting step. The most
important factor for achieving high conversion was slow ad-
dition of the Grignard reagent; fast addition caused precipi-
tation of iron, presumably due to over-reduction.
Herein, we investigate the electronic effects on the nucle-
ophile by use of a competitive Hammett study (Scheme 1).
With the more weakly reducing aryl Grignards, it was found
that the catalyst is stable in diethyl ether without additives.[5]
Cyclohexyl bromide was added in aliquots to a mixture of
p-substituted and unsubstituted phenyl magnesium bromide
and consumed after each addition without adverse effects
on the catalytic efficiency. Product formation was followed
by GC, with samples taken before each addition of the elec-
trophile.
tected as a darkening of the solution, followed by precipita-
tion. In this study, no deactivation was observed, despite the
large excess of aryl Grignard present and the absence of sta-
bilizing additives, demonstrating the lower reducing power
of aryl Grignard reagents.
In analyzing the competition reaction data, we assumed
that the kinetic dependence on all reagents and catalysts is
the same for both substrates (X and H, for p-substituted and
unsubstituted phenyl magnesium bromides) and that the re-
action is first order in Grignard reagent. The relative rate
(krel =kX/kH) is then obtained as the slope of a plot of
ln([X]0/[X]) against ln([H]0/[H]), that is, the initial and in-
stantaneous concentrations of each Grignard reagent. Since
these cannot be measured directly, they were calculated by
comparing the instantaneous to the final product concentra-
tions after addition of excess cyclohexyl bromide. Note that
the analysis is insensitive to absolute concentration; only
relative concentrations need to be well described. All plots
gave straight lines with correlation coefficients r2 >0.99
(Figure 1), indicating both that the assumptions we made
were valid and that no significant side reactions occurred.
Inspection of the trends indicates that some curvature can
be detected in, for example, the case of the CF3-substituted
Grignard, introducing a minor uncertainty in krel for this
substituent, but it is clear from the plot that the possible de-
viation between the slopes in the initial and final phases of
the reaction are too small to affect the conclusions drawn in
this study.
In the previous study employing strongly reducing alkyl
Grignard reagents, catalyst deactivation could be visibly de-
[a] A. Hedstrçm, Prof. P.-O. Norrby
Department of Chemistry, University of Gothenburg
Kemigꢀrden 4, 412-96, Gothenburg (Sweden)
Fax : (+46)31-772-38-40
E-mail: pon@chem.gu.se
[b] U. Bollmann, J. Bravidor
Institute of Inorganic and Analytical Chemistry
Friedrich Schiller University Jena
Lessingstrasse 8, 07743, Jena (Germany)
An alternative method of estimating the relative concen-
tration of the Grignard reagent would be to analyze the
amount of protonation product (benzene and substituted
benzene) after workup. The protonation products are also
[**] NMP (N-methyl-2-pyrrolidone), the most commonly employed addi-
tive combined with alkyl Grignard reagents, inhibits product forma-
tion in this reaction.
Chem. Eur. J. 2011, 17, 11991 – 11993
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11991

stituents (chloride, fluoride, and trifluoromethyl) inhibit the
coupling reaction. This is in line with studies of transmetala-
tion reactions in related palladium-catalyzed couplings,[7]
and in agreement with the hypothesis of mechanistic similar-
ity between the metals.[4]
Figure 2 shows that the fit to s is less than perfect, and in
particular that the halide substituents deviate. The latter ob-
servation is not uncommon in Hammett studies.[8] We tested
several other s scales (s+, sÀ, s·, as well as Swain–Lupton
parameters[9]), alone and in combination, but no statistically
significant improvement could be obtained (as judged by
using F tests). Fitting to the standard s values gives 1ꢀ
À0.5, varying slightly depending on whether the halides are
included or not.
This Hammett study gives a weakly negative 1 value,
which is in good agreement with previous studies of trans-
metalation reactions.[7] We note that the reductive elimina-
tion step is the same here as in our previous study; at this
point the identity of the nucleophile and electrophile has
been scrambled.[4] Since the Hammett study of the electro-
phile gave a positive 1 value, the reductive elimination
cannot be the rate- or selectivity-determining step. In com-
bination with the earlier computational study of reductive
elimination from different iron species,[4] this gives strong
support for an FeI–FeIII cycle. We note that a possible ex-
planation for the unexpectedly low selectivity with the
NMe2 substituent would be a change in the selectivity-deter-
mining step with this reagent; it is entirely possible that the
high electron density here retards the reductive elimination
enough to allow further transmetalation and thus negate the
acceleration expected for oxidative addition in the presence
of the NMe2 substituent.
Our previous study indicated that TM can occur either
before or after OA with an sp2-hybridized electrophile, with
very little difference in the two calculated energy barriers.
In this case, OA to an sp3-hybridized electrophile is expect-
ed to be quite different and could potentially include either
single-electron[5c,10] or atom transfer, as recently indicated
for Cu-catalyzed couplings.[11] A catalytic cycle of that type
should produce an alkyl radical that could later couple.
However, one good indicator of radical mechanisms is that
all substituents are expected to provide stabilization and
therefore should accelerate the reaction (i.e., most s· values
are positive). In this case, however, we can exclude radical
involvement in the coupling step, since we observe a de-
creased preference for nucleophiles containing electron-
withdrawing groups.
~
^
Figure 1. Plots of ln([X]0/[X])=krelln([H]0/[H]); =OMe; & =Me;
=
&
*
NMe2; ꢂ =CF3; =F; =Cl.
produced in the initiation phase and therefore cannot be
used directly as a measure of the Grignard concentration,
but the concentrations should correlate after the initiation.
We have therefore verified that the concentration of the
protonation product at each point correlates with the con-
centration of the Grignard reagents estimated from product
formation, providing a correlation coefficient of at least
98%.
The relative rate, krel, for each p-substituted phenyl
Grignard reagent was fitted to the literature s values[6] by
using the Hammett expression logACTHNUTRGNE(UNG kX/kH)=1s. The results
are shown in Table 1 and depicted in Figure 2.
As seen in Table 1, coupling is favored for Grignard re-
agents with electron-donating substituents (dimethylamine,
methoxy, and methyl), whereas electron-withdrawing sub-
Table 1. Relative rates and s values[6] for different para substituents.
para Substituent
s
krel
logAHCTUNGRTENGUN(krel)
NMe2
OMe
Me
F
À0.83
À0.27
À0.17
0.06
1.99
2.13
1.22
0.39
0.39
0.51
0.30
0.33
0.09
À0.41
À0.41
À0.29
Cl
0.23
CF3
0.54
The two possible TM steps shown in Scheme 2 are expect-
ed to show quite different behavior with respect to the elec-
tron density on the nucleophile, since one occurs at FeI and
the other at FeIII. This could help rationalize the low correla-
tion observed in Figure 2, as it is plausible that the prefer-
ence displayed by the two iron species could change with
electron density and therefore result in a change in the pre-
ferred reaction pathway. In particular, the deviation in selec-
tivity for the CF3-substituted reactant could be a shift in
preference from one OA pathway to the other. The local V
Figure 2. Hammett plot of logACHTNUTRGNEUNG(krel) versus s.
11992
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 11991 – 11993

Competitive Hammett Study
COMMUNICATION
Keywords: catalysis · coupling reactions · Hammett plots ·
iron · reaction mechanisms
Scheme 2. Proposed mechanistic pathways; oxidative addition (OA) fol-
lowed by transmetalation (TM), or TM followed by OA. Reductive elimi-
nation (RE) gives the product and regenerates the iron catalyst.
[1] a) Handbook of Organopalladium Chemistry for Organic Synthesis
(Eds.: E. Negishi), Wiley, New York, 2002; b) J. Tsuji, Palladium Re-
agents and Catalysts, 2nd ed., Wiley, New York, 2004; c) Metal-Cata-
lyzed Cross-Coupling Reactions, 2nd ed. (Eds.: A. de Meijere, F.
Diederich), Wiley-VCH, Weinheim, 2004; d) Transition Metals for
Organic Synthesis, 2nd ed. (Eds.: M. Beller, C. Bolm), Wiley-VCH,
Weinheim, 2004.
[2] a) M. Tamura, J. Kochi, J. Organomet. Chem. 1971, 31, 289; b) M.
Tamura, J. Kochi, Syntheses 1971, 303; c) M. Tamura, J. K. Kochi, J.
Am. Chem. Soc. 1971, 93, 1487; d) S. M. Neumann, J. K. Kochi, J.
Org. Chem. 1975, 40, 599.
[3] a) A. Fꢃrstner, R. Martin, Chem. Lett. 2005, 34, 624; b) C. Bolm, J.
Legros, J. Le Paih, L. Zani, Chem. Rev. 2004, 104, 6217; c) G.
Cahiez, H. Avedissian, Synthesis 1998, 1199; d) C. Duplais, F. Bures,
I. Sapountzis, T. J. Korn, G. Cahiez, P. Knochel, Angew. Chem. 2004,
116, 3028; Angew. Chem. Int. Ed. 2004, 43, 2968; e) A. Fꢃrstner, A.
Leitner, M. Mendez, H. Krause, J. Am. Chem. Soc. 2002, 124, 13856;
f) J. F. Hartwig, Organotransition Metal Chemistry From Bonding to
Catalysis, University Science, Sausalito, 2010.
shape in a Hammett plot is generally associated with a
change in reaction pathway.
In conclusion, this study supports a common mechanism
for the two types of iron-catalyzed aryl–alkyl coupling reac-
tion, based on an FeI–FeIII cycle. For both reactions, the rela-
tive ordering of the OA and TM steps is uncertain and path
selection may be substrate dependent. Radical intermediates
in the coupling step can be excluded, but we cannot draw
any conclusions about the involvement of radicals in the oxi-
dative addition step[10] as long as any radicals formed recom-
bine with iron before the actual coupling occurs.
[4] J. Kleimark, A. Hedstrçm, P.-F. Larsson, C. Johansson, P.-O. Norrby,
ChemCatChem 2009, 1, 152.
Experimental Section
[5] a) T. Nagano, T. Hayashi, Org. Lett. 2004, 6, 1297; b) M. Nakamura,
K. Matsuo, S. Ito, E. Nakamura, J. Am. Chem. Soc. 2004, 126, 3686;
c) G. Cahiez, V. Habiak, C. Duplais, A. Moyeux, Angew. Chem.
2007, 119, 4442; Angew. Chem. Int. Ed. 2007, 46, 4364; d) R. B. Bed-
ford, M. Betham, D. W. Bruce, A. A. Danopoulos, R. M. Frost, M.
Hird, J. Org. Chem. 2006, 71, 1104.
[6] a) C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165; b) X.
Creary, M. E. Mehrsheikhmohammadi, S. Mcdonald, J. Org. Chem.
1987, 52, 3254.
[7] a) V. Farina, B. Krishnan, D. R. Marshall, G. P. Roth, J. Org. Chem.
1993, 58, 5434; b) T. Nishikata, Y. Yamamoto, N. Miyaura, Organo-
metallics 2004, 23, 4317; c) D. Zim, S. M. Nobre, A. L. Monteiro, J.
Mol. Catal. A 2008, 287, 16; d) Z. B. Dong, G. Manolikakes, L. Shi,
P. Knochel, H. Mayr, Chem. Eur. J. 2010, 16, 248.
[8] D. A. Culkin, J. F. Hartwig, Organometallics 2004, 23, 3398.
[9] a) C. G. Swain, E. C. Lupton, J. Am. Chem. Soc. 1968, 90, 4328;
b) C. G. Swain, S. H. Unger, N. R. Rosenquist, M. S. Swain, J. Am.
Chem. Soc. 1983, 105, 492.
General Procedure: A round bottomed flask (50 mL) was sealed with a
rubber septum, then evacuated and refilled with nitrogen four times. The
flask was charged with diethyl ether (DEE) (30 mL), dodecane (225 mL,
1 mmol), phenyl magnesium bromide (400 mL, 1.2 mmol, 3m in DEE),
and p-substituted phenyl magnesium bromide (approx. 1.2 mmol). An ali-
quot (0.5 mL) was taken from the mixture, quenched with saturated
NH4Cl (0.5 mL), filtered through a small silica plug, diluted with DEE
and analyzed by GC (dodecane was used as the internal standard). A so-
lution of FeACHTUNGTRENNUNG(acac)3 (18 mg, 0.05 mmol) and cyclohexyl bromide (0.2 mL,
0.1 mmol, 0.5m in DEE) in DEE (5 mL) was added to the reaction
vessel. After stirring for 5 min, an aliquot (0.5 mL) was collected and an-
alyzed as described above. The procedure was repeated by adding cyclo-
hexyl bromide (0.5m in diethyl ether) in portions of 0.2 mL (a total of
nine portions). Two extra portions of 1 or 2 equivalents of cyclohexyl bro-
mide (122 mL, 1 mmol or 244 mL, 2 mmol) were added to react with any
remaining Grignard reagent.
[10] A. Guꢄrinot, S. Reymond, J. Cossy, Angew. Chem. 2007, 119, 6641;
Angew. Chem. Int. Ed. 2007, 46, 6521.
[11] G. O. Jones, P. Liu, K. N. Houk, S. L. Buchwald, J. Am. Chem. Soc.
2010, 132, 6205.
Acknowledgements
Received: February 11, 2011
Revised: August 8, 2011
The current project was supported by the Swedish Research Council.
Published online: September 16, 2011
Chem. Eur. J. 2011, 17, 11991 – 11993
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11993