Highly selective biaryl cross-coupling reactions between aryl halides and aryl Grignard reagents: A new catalyst combination of N-heterocyclic carbenes and iron, cobalt, and nickel fluorides

Article abstract of DOI:10.1021/ja9039289

Combinations of N-heterocyclic carbenes (NHCs) and fluoride salts of the iron-group metals (Fe, Co, and Ni) have been shown to be excellent catalysts for the cross-coupling reactions of aryl Grignard reagents (Ar¹MgBr) with aryl and heteroaryl halides (Ar²X) to give unsymmetrical biaryls (Ar¹-Ar²). Iron fluorides in combination with SIPr, a saturated NHC ligand, catalyze the biaryl cross-coupling between various aryl chlorides and aryl Grignard reagents in high yield and high selectivity. On the other hand, cobalt and nickel fluorides in combination with IPr, an unsaturated NHC ligand, exhibit interesting complementary reactivity in the coupling of aryl bromides or iodides; in contrast, with these substrates the iron catalysts show a lower selectivity. The formation of homocoupling byproducts is suppressed markedly to less than 5% in most cases by choosing the appropriate metal fluoride/NHC combination. The present catalyst combinations offer several synthetic advantages over existing methods: practical synthesis of a broad range of unsymmetrical biaryls without the use of palladium catalysts and phosphine ligands. On the basis of stoichiometric control experiments and theoretical studies, the origin of the unique catalytic effect of the fluoride counterion can be ascribed to the formation of a higher-valent heteroleptic metalate [Ar¹MF²]MgBr as the key intermediate in our proposed catalytic cycle. First, stoichiometric control experiments revealed the stark differences in chemical reactivity between the metal fluorides and metal chlorides. Second, DFT calculations indicate that the initial reduction of di- or trivalent metal fluoride in the wake of transmetalation with PhMgCl is energetically unfavorable and that formation of a divalent heteroleptic metalate complex, [PhMF²]MgCl (M ) Fe, Co, Ni), is dominant in the metal fluoride system. The heteroleptic ate-complex serves as a key reactive intermediate, which undergoes oxidative addition with PhCl and releases the biaryl cross-coupling product Ph-Ph with reasonable energy barriers. The present crosscoupling reaction catalyzed by iron-group metal fluorides and an NHC ligand provides a highly selective and practical method for the synthesis of unsymmetrical biaryls as well as the opportunity to gain new mechanistic insights into the metal-catalyzed cross-coupling reactions.

Full text of DOI:10.1021/ja9039289

Published on Web 07/29/2009
Highly Selective Biaryl Cross-Coupling Reactions between
Aryl Halides and Aryl Grignard Reagents: A New Catalyst
Combination of N-Heterocyclic Carbenes and Iron, Cobalt,
and Nickel Fluorides
Takuji Hatakeyama,† Sigma Hashimoto,†,‡ Kentaro Ishizuka,†,§ and
Masaharu Nakamura*,†,§
International Research Center for Elements Science, Institute for Chemical Research,
Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, and
Institute of Sustainability Science, Kyoto UniVersity, Uji, Kyoto 611-0011, Japan
Received May 14, 2009; E-mail: masaharu@scl.kyoto-u.ac.jp
Abstract: Combinations of N-heterocyclic carbenes (NHCs) and fluoride salts of the iron-group metals
(Fe, Co, and Ni) have been shown to be excellent catalysts for the cross-coupling reactions of aryl Grignard
reagents (Ar1MgBr) with aryl and heteroaryl halides (Ar2X) to give unsymmetrical biaryls (Ar1-Ar2). Iron
fluorides in combination with SIPr, a saturated NHC ligand, catalyze the biaryl cross-coupling between
various aryl chlorides and aryl Grignard reagents in high yield and high selectivity. On the other hand,
cobalt and nickel fluorides in combination with IPr, an unsaturated NHC ligand, exhibit interesting
complementary reactivity in the coupling of aryl bromides or iodides; in contrast, with these substrates the
iron catalysts show a lower selectivity. The formation of homocoupling byproducts is suppressed markedly
to less than 5% in most cases by choosing the appropriate metal fluoride/NHC combination. The present
catalyst combinations offer several synthetic advantages over existing methods: practical synthesis of a
broad range of unsymmetrical biaryls without the use of palladium catalysts and phosphine ligands. On
the basis of stoichiometric control experiments and theoretical studies, the origin of the unique catalytic
effect of the fluoride counterion can be ascribed to the formation of a higher-valent heteroleptic metalate
[Ar1MF2]MgBr as the key intermediate in our proposed catalytic cycle. First, stoichiometric control experiments
revealed the stark differences in chemical reactivity between the metal fluorides and metal chlorides. Second,
DFT calculations indicate that the initial reduction of di- or trivalent metal fluoride in the wake of
transmetalation with PhMgCl is energetically unfavorable and that formation of a divalent heteroleptic
metalate complex, [PhMF2]MgCl (M ) Fe, Co, Ni), is dominant in the metal fluoride system. The heteroleptic
ate-complex serves as a key reactive intermediate, which undergoes oxidative addition with PhCl and
releases the biaryl cross-coupling product Ph-Ph with reasonable energy barriers. The present cross-
coupling reaction catalyzed by iron-group metal fluorides and an NHC ligand provides a highly selective
and practical method for the synthesis of unsymmetrical biaryls as well as the opportunity to gain new
mechanistic insights into the metal-catalyzed cross-coupling reactions.
Introduction
polymers. Because of the significance and prevalence of this
class of compounds, numerous synthetic methods have been
Biaryls are important structural units for a wide range of
functional molecules,1 such as chiral ligands and catalysts,
drug intermediates, liquid crystals, physiologically active
natural products, organic electronic materials, and functional
developed for over a century.2 Reductive homocoupling of
aryl halides or pseudohalides i.e., the classical Ullmann
reaction, has been developed to give the desired symmetrical
biaryls using various transition metals,2a,c palladium catalysts
in combination with appropriate terminal reductants such as
zinc are currently the prevailing choice.2b Furthermore, the
cross-coupling of two different aryl halides under reducing
conditions has been also developed.3 The oxidative coun-
terpart of the biaryl synthesis has also enhanced its utility
† International Research Center for Elements Science (IRCELS), Institute
for Chemical Research, Kyoto University.
‡ Department of Energy and Hydrocarbon Chemistry, Kyoto University.
§ Institute of Sustainability Science, Kyoto University.
(1) (a) Demus, D., Goodby, J. W., Gray, G. W., Spiess, H.-W., Vill, V.,
Eds. Handbook of Liquid Crystals; Wiley-VCH: Weinheim, 1998. (b)
Mitschke, U.; Bauerle, P. J. Mater. Chem. 2000, 10, 1471–1507. (c)
Yamamoto, T. J. Organomet. Chem. 2002, 653, 195–199. (d) Baudoin,
O.; Gueritte, F. Stud. Nat. Prod. Chem., Part J 2003, 29, 355–417.
(e) Holder, E.; Langeveld, B. M. W.; Schubert, U. S. AdV. Mater.
2005, 17, 1109–1121. (f) Ambade, A. V.; Savariar, E. N.; Thayumana-
van, S. Mol. Pharm. 2005, 2, 264–272. (g) Doucet, H.; Hierso, J.-C.
Curr. Opin. Drug DiscoVery DeVel. 2007, 10, 672–690. (h) Abe, H.;
Harayama, T. Heterocycles 2008, 75, 1305–1320.
(2) (a) Cepanec, I., Ed. Synthesis of Biaryls; Elsevier Ltd: Oxford, 2004.
(b) Hassan, J.; Se´vignon, M.; Gozzi, C.; Schulz, E.; Lemaire, M. Chem.
ReV. 2002, 102, 1359–1469. (c) Nelson, T. D.; Crouch, R. D. Org.
React. 2004, 63, 265–555. (d) Corbet, J. P.; Mignani, G. Chem. ReV.
2006, 106, 2651–2710.
(3) Amatore, M.; Gosmini, C. Angew. Chem., Int. Ed. 2008, 47, 2089–
2092, and references cited therein.
9
10.1021/ja9039289 CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 11949–11963 11949

A R T I C L E S
Hatakeyama et al.
Scheme 1. Transition Metal-Catalyzed Biaryl Coupling Reaction
since the seminal work of Kharasch and co-workers who
discovered that aryl Grignard reagents undergo efficient
homocoupling in the presence of catalytic amounts of first-
row transition metal salts, such as the chloride salts of
chromium, manganese, iron, cobalt, nickel, and copper.4
Indeed, recent progress in this oxidative coupling chemistry
has extended their utility to the synthesis of various func-
tionalized biaryls.5 The advancement of both reductive and
oxidative biaryl coupling reactions notwithstanding, transition
metal-catalyzed cross-coupling reactions are the popular
choice, holding the synthetic advantages such as high
selectivity, broad substrate scopes, and mild reaction condi-
tions.6 In fact, a wide range of arylmetal compounds has been
successfully utilized to date as the nucleophilic partner in
unsymmetrical biaryl coupling reactions (Scheme 1). While
one may use aryllithium,7 magnesium,8 boron,9 silicon,10
copper,11 zinc,12 or tin compounds,13 arylmagnesium com-
pounds seem ideal for practical synthesis because they are
readily available, cheap and environmentally benign. Despite
these synthetic advantages of arylmagnesium compounds,
there is one serious drawback associated with their use: the
considerable generation of the symmetrical biaryls via
undesired homocouplings of the arylmagnesium compound
and/or the aryl electrophile. This issue often hampers their
general application for industrial production of biaryls.
We have been interested in the development of iron-
catalyzed reactions,14-16 and launched a thorough investiga-
tion of iron catalysis to overcome the above-mentioned
synthetic problem associated with the unwanted homocou-
(4) Kharasch, M. S.; Fields, E. K. J. Am. Chem. Soc. 1941, 63, 2316–
2320.
(5) Recent papers: (a) Nagano, T.; Hayashi, T. Org. Lett. 2005, 7, 491–
493. (b) Cahiez, G.; Chaboche, C.; Mahuteau-Betzer, F.; Ahr, M. Org.
Lett. 2005, 7, 1943–1946. (c) Cahiez, G.; Moyeux, A.; Buendia, J.;
Duplais, C. J. Am. Chem. Soc. 2007, 129, 13788–13789. Without metal
catalysts. (d) Krasovskiy, A.; Tishkov, A.; del Amo, V.; Mayr, H.;
Knochel, P. Angew. Chem., Int. Ed. 2006, 45, 5010–5014.
(6) (a) de Meijere, A., Diederich, F., Eds. Metal-Catalyzed Cross-coupling
Reactions, 2nd ed.; Wiley-VCH: Germany, 2004. (b) Miyaura, N.,
Ed.; Cross-Coupling Reactions: A Practical Guide; Springer: Berlin,
2002. (c) Stanforth, S. P. Tetrahedron 1998, 54, 263–303.
(7) (a) Snieckus, V. Chem. ReV. 1990, 90, 879–933. (b) Murahashi, S. J.
Organomet. Chem. 2002, 653, 27–33. (c) Anctil, E. J.-G.; Snieckus,
V. J. Organomet. Chem. 2002, 653, 150–160. (d) Na´jera, C.; Sansano,
J. M.; Yus, M. Tetrahedron 2003, 59, 9255–9303.
(13) (a) Stille, J. K. Angew. Chem., Int. Ed. Engl. 1986, 25, 508–523. (b)
Mitchell, T. N. J. Organomet. Chem. 1986, 304, 1–16. (c) Mitchell,
T. N. Synthesis 1992, 803–815. (d) Kosugi, M.; Fugami, K. J.
Organomet. Chem. 2002, 653, 50–53. (e) Espinet, P.; Echavarren,
A. M. Angew. Chem., Int. Ed. 2004, 43, 4704–4734. (f) Echavarren,
A. M. Angew. Chem., Int. Ed. 2005, 44, 3962–3965.
(14) Reviews: (a) Plietker, B. Ed. Iron Catalysis in Organic Chemistry;
Wiley-VCH: Weinheim, 2008. (b) Bolm, C.; Legros, J.; Le Paih, J.;
Zani, L. Chem. ReV. 2004, 104, 6217–6254. (c) Shinokubo, H.;
Oshima, K. Eur. J. Org. Chem. 2004, 2081–2091. (d) Fu¨rstner, A.;
Martin, R. Chem. Lett. 2005, 34, 624–629. (e) Correa, A.; Manchen˜o,
O. G.; Bolm, C. Chem. Soc. ReV. 2008, 37, 1108–1117. (f) Sherry,
B. D.; Fu¨rstner, A. Acc. Chem. Res. 2008, 41, 1500–1511.
(8) (a) Tamao, K. J. Organomet. Chem. 2002, 653, 23–26. (b) Banno,
T.; Hayakawa, Y.; Umeno, M. J. Organomet. Chem. 2002, 653, 288–
291. (c) Knochel, P.; Dohle, W.; Gommermann, N.; Kneisel, F. F.;
Kopp, F.; Korn, T.; Sapountzis, I.; Vu, V. A. Angew. Chem., Int. Ed.
2003, 42, 4302–4320. (d) Ila, H.; Baron, O.; Wagner, A. J.; Knochel,
P. Chem. Lett. 2006, 35, 2–7. (e) Terao, J.; Kambe, N. Acc. Chem.
Res. 2008, 41, 1545–1554.
(15) Iron-catalyzed cross-coupling reactions of sp2-carbon electrophiles: (a)
Tamura, M.; Kochi, J. K. J. Am. Chem. Soc. 1971, 93, 1487–1489.
(b) Fabre, J.-L.; Julia, M.; Verpeaux, J.-N. Tetrahedron Lett. 1982,
23, 2469–2472. (c) Molander, G. A.; Rahn, B. J.; Shubert, D. C.;
Bonde, S. E. Tetrahedron Lett. 1983, 24, 5449–5452. (d) Reddy, C. K.;
Knochel, P. Angew. Chem., Int. Ed. Engl. 1996, 35, 1700–1701. (e)
Cahiez, G.; Avedissian, H. Synthesis 1998, 1199–1205. (f) Fu¨rstner,
A.; Leitner, A.; Me´ndez, M.; Krause, H. J. Am. Chem. Soc. 2002,
124, 13856–13863. (g) Necas, D.; Kotora, M.; C´ısarova´, I. Eur. J.
Org. Chem. 2004, 1280–1285. (h) Duplais, C.; Bures, F.; Sapountzis,
I.; Korn, T. J.; Cahiez, G.; Knochel, P. Angew. Chem., Int. Ed. 2004,
43, 2968–2970. (i) Itami, K.; Higashi, S.; Mineno, M.; Yoshida, J.
Org. Lett. 2005, 7, 1219–1222. (j) Dunet, G.; Knochel, P. Synlett 2006,
407–410. (k) Cahiez, G.; Gager, O.; Habiak, V. Synthesis 2008, 2636–
2644. (l) Cahiez, G.; Habiak, V.; Gager, O. Org. Lett. 2008, 10, 2389–
2392. (m) Carril, M.; Correa, A.; Bolm, C. Angew. Chem., Int. Ed.
2008, 47, 4862–4865. (n) Hatakeyama, T.; Yoshimoto, Y.; Gabriel,
T.; Nakamura, M. Org. Lett. 2008, 10, 5341–5344. (o) Fu¨rstner, A.;
Martin, R.; Krause, H.; Seidel, G.; Goddard, R.; Lehmann, C. W.
J. Am. Chem. Soc. 2008, 130, 8773–8787.
(9) (a) Miyaura, N.; Suzuki, A. Chem. ReV. 1995, 95, 2457–2483. (b)
Lloyd-Williams, P.; Giralt, E. Chem. Soc. ReV. 2001, 30, 145–157.
(c) Miyaura, N. J. Organomet. Chem. 2002, 653, 54–57. (d) Suzuki,
A. J. Organomet. Chem. 2002, 653, 83–90. (e) Miyaura, N. Top. Curr.
Chem. 2002, 219, 11–59. (f) Kotha, S.; Lahiri, K.; Kashinath, D.
Tetrahedron 2002, 58, 9633–9695. (g) Bellina, F.; Carpita, A.; Rossi,
R. Synthesis 2004, 2419–2440. (h) Ishiyama, T.; Miyaura, N. Pure
Appl. Chem. 2006, 78, 1369–1375. (i) Molander, G. A.; Ellis, N. Acc.
Chem. Res. 2007, 40, 275–286. (j) Darses, S.; Genet, J.-P. Chem. ReV.
2008, 108, 288–325. (k) Doucet, H. Eur. J. Org. Chem. 2008, 2013–
2030.
(10) (a) Hiyama, T.; Hatanaka, Y. Pure Appl. Chem. 1994, 66, 1471–1478.
(b) Hiyama, T.; Shirakawa, E. Top. Curr. Chem. 2002, 219, 61–85.
(c) Denmark, S. E.; Sweis, R. F. Acc. Chem. Res. 2002, 35, 835–846.
(d) Hiyama, T. J. Organomet. Chem. 2002, 653, 58–61. (e) Denmark,
S. E.; Pan, W. J. Organomet. Chem. 2002, 653, 98–104. (f) Denmark,
S. E.; Ober, M. H. Aldrichimica Acta 2003, 36, 75–85. (g) Spivey,
A. C.; Gripton, C. J. G.; Hannah, J. P. Curr. Org. Synth. 2004, 1,
211–226. (h) Denmark, S. E.; Baird, J. D. Chem.sEur. J. 2006, 12,
4954–4963. (i) Denmark, S. E.; Regens, C. S. Acc. Chem. Res. 2008,
41, 1486–1499.
(16) Iron-catalyzed cross-coupling reactions of sp3-carbon electrophiles: (a)
Nakamura, M.; Matsuo, K.; Ito, S.; Nakamura, E. J. Am. Chem. Soc.
2004, 126, 3686–3687. (b) Nagano, T.; Hayashi, T. Org. Lett. 2004,
6, 1297–1299. (c) Martin, R.; Fu¨rstner, A. Angew. Chem., Int. Ed.
2004, 43, 3955–3957. (d) Bedford, R. B.; Bruce, D. W.; Frost, R. M.;
Goodby, J. W.; Hird, M. Chem. Commun. 2004, 2822–2823. (e)
Nakamura, M.; Ito, S.; Matsuo, K.; Nakamura, E. Synlett 2005, 1794–
1798. (f) Bedford, R. B.; Betham, M.; Bruce, D. W.; Danopoulos,
A. A.; Frost, R. M.; Hird, M. J. Org. Chem. 2006, 71, 1104–1110.
(g) Plietker, B. Angew. Chem., Int. Ed. 2006, 45, 1469–1473. (h)
Dongol, K. G.; Koh, H.; Sau, M.; Chai, C. L. L. AdV. Synth. Catal.
2007, 349, 1015–1018. (i) Cahiez, G.; Habiak, V.; Duplais, C.;
Moyeux, A. Angew. Chem., Int. Ed. 2007, 46, 4364–4366. (j) Cahiez,
G.; Duplais, C.; Moyeux, A. Org. Lett. 2007, 9, 3253–3254. (k)
Gue´rinot, A.; Reymond, S.; Cossy, J. Angew. Chem., Int. Ed. 2007,
46, 6521–6524. (l) Rao Volla, C.-M.; Vogel, P. Angew. Chem., Int.
Ed. 2008, 47, 1305–1307. (m) Bedford, R. B.; Huwe, M.; Wilkinson,
M. C. Chem. Commun. 2009, 600–602. (n) Hatakeyama, T.; Kondo,
Y.; Fujiwara, Y.; Takaya, H.; Ito, S.; Nakamura, E.; Nakamura, M.
Chem. Commun. 2009, 1216–1218.
(11) (a) Taylor, R. J. K. Ed.; Organocopper Reagents; Oxford University
Press: Oxford, 1994. (b) Lipshutz, B. H. Acc. Chem. Res. 1997, 30,
277–282. (c) Krause, N.; Gerold, A. Angew. Chem., Int. Ed. 1997,
36, 187–204. (d) Woodward, S. Chem. Soc. ReV. 2000, 29, 393–401.
(e) Krause, N. Ed.; Modern Organocopper Chemistry, Wiley-VCH,
Weinheim, 2002. (f) Nakamura, E.; Mori, S. Angew. Chem., Int. Ed.
2000, 39, 3751–3771. (g) Caprio, V. Lett. Org. Chem. 2006, 3, 339–
349.
(12) (a) Negishi, E. Acc. Chem. Res. 1982, 15, 340–348. (b) Knochel, P.,
Jones, P., Eds. Organozinc Reagents; Oxford University Press: Oxford,
1999. (c) Negishi, E. J. Organomet. Chem. 2002, 653, 34–40. (d)
Lessene, G. Aust. J. Chem. 2004, 57, 107. (e) Zvi, R., Marek, I. Eds.
Chemistry of Organozinc Compounds; John Wiley: Chichester, 2006.
9
11950 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Metal Fluoride/NHC To Catalyze Biaryl Cross-Coupling
A R T I C L E S
Scheme 2. Selective Biaryl Cross-Coupling with an Iron Fluoride
Catalyst
unsymmetrical biaryl synthesis with new catalyst combina-
tions consisting of NHC ligand and the iron-group metal
fluorides.
Results and Discussion
Iron Fluoride/NHC-Catalyzed Biaryl Cross-Coupling Reac-
tion. We began our investigation focusing on the iron-catalysts
because no general and practical method had theretofore been
developed for iron-catalyzed biaryl cross-coupling reactions.14-17
Whereas efficient homocoupling reactions of arylmagnesium
compounds have been achieved using iron catalysts,5a,b,d only
a few reports were reported for the iron-catalyzed cross-coupling
reactions before our study27 and a few more28 in the last two
years. We thus conducted a careful and detailed catalyst
screening for the reaction of a simple aryl chloride with an
arylmagnesium reagent. The benchmark coupling reaction was
performed by heating a THF solution of chlorobenzene 1,
p-tolylmagnesium bromide (p-TolMgBr, 2.5 equiv), an iron salt
(5 mol %), and an additive at 60 °C for 24 h (Table 1). Various
additives, including imidazolium and imidazolinium salts (as
shown in Chart 1), as well as typical phosphine ligands, were
studied in combination with a catalytic amount of iron fluorides.
The optimum yield (98%) of 4-methylbiphenyl 2 was
achieved with 5 mol % of FeF3 ·3H2O and 15 mol % of
SIPr·HCl (entry 1). The undesired homocoupling reaction
occurred only sluggishly and gave a negligible amount of
biphenyl 3 and a small amount of 4,4′-dimethylbiphenyl 4 (0.018
mmol, 4% yield, based on the amount of p-TolMgBr). As shown
in entry 2, when the less sterically demanding NHC was used,
a lower conversion of the starting chlorobenzene 1 was observed.
The unsaturated NHC precursors, IPr ·HCl and I-t-Bu·HCl, were
ineffective (entries 3 and 5). The counteranion of the NHC
precursors has a considerable effect on the reactivity (entry 4).
The use of 10 or 5 mol % of SIPr ·HCl resulted in a selective,
but slower, conversion (entries 6 and 7). The reaction was
sluggish without an NHC precursor (entry 8). Tricyclohexyl-
phosphine (PCy3) did not accelerate the reaction (entry 9). As
shown in entries 10-13, bidentate phosphine ligands and
nitrogen ligands were found to inhibit the reaction.
pling. We recently reported that the combination of iron
fluoride and N-heterocyclic carbene (NHC) ligand is a highly
selective and practical catalyst for biaryl cross-coupling
reaction of arylmagnesium compounds (Scheme 2).17 Metal
fluorides have been known to show unique reactivity and
selectivity in transition metal-catalyzed carbon-carbon bond-
forming reactions such as the asymmetric Mukaiyama-Aldol
reaction (copper),18 allylation reaction (titanium),19 as well
as carbon-heteroatom bond-forming reactions, hydroami-
nation reaction (iridium),20 and hydrosilylation reaction
(titanium).21 While such a “fluoride effect” has attracted
considerable interests in synthetic chemistry22 as well as
inorganic/organometallic chemistry, it remains unstudied in
the field of transition metal-catalyzed cross-coupling reac-
tions.23,24 We have continued careful investigation on the
“fluoride effect” in cobalt-25 and nickel-catalyzed26 cross-
coupling reactions to achieve markedly selective biaryl cross-
coupling of aryl halides with arylmagnesium compounds, and
herein, report on the full details of this synthetically useful
(17) Hatakeyama, T.; Nakamura, M. J. Am. Chem. Soc. 2007, 129, 9844–
9845.
(18) Kru¨ger, J.; Carreira, E. M. J. Am. Chem. Soc. 1998, 120, 837–838.
(19) Gauthier, D. R.; Carreira, E. M. Angew. Chem., Int. Ed. 1996, 35,
2363–2365.
(20) Dorta, R.; Egli, P.; Zu¨rcher, F.; Togni, A. J. Am. Chem. Soc. 1997,
119, 10857–10858.
Entries 14-21 show a striking contrast in the reactivity of
various iron salts. Iron(II) fluoride tetrahydrate (FeF2 ·4H2O)
showed a comparable catalytic activity to afford 2 in 96% yield.
The reaction did not complete with anhydrous FeF3 or FeF2,
probably due to their lower solubility in THF (entries 15 and
(21) Verdaguer, X.; Lange, U. E. W.; Reding, M. T.; Buchwald, S. L. J. Am.
Chem. Soc. 1996, 118, 6784–6785.
(22) Reviews: (a) Pagenkopf, B. L.; Carreira, E. M. Chem. Eur. J. 1999,
5, 3437–3442. (b) Fagnou, K.; Lautens, M. Angew. Chem., Int. Ed.
2002, 41, 26–47.
(23) Concomitant halide effect on palladium- or nickel-catalyzed cross-
coupling and Heck reactions has been investigated for chloride,
bromide, and iodide ions but not for fluoride ions: Amatore, C.; Jutand,
A. Acc. Chem. Res. 2000, 33, 314–321, and references cited therein.
(24) Activation of organometal compounds with a stoichiometric amount
of fluoride ion in cross-coupling reactions. Boron: (a) Wright, S. W.;
Hageman, D. L.; McClure, L. D. J. Org. Chem. 1994, 59, 6095–6097.
(b) Darses, S.; Michaud, G.; Genet, J.-P. Eur. J. Org. Chem. 1999,
1875–1883. (c) Molander, G. A.; Biolatto, B. J. Org. Chem. 2003,
68, 4302–4314. Silane: (d) Hatanaka, Y.; Hiyama, T. J. Org. Chem.
1988, 53, 918–920. (e) Denmark, S. E.; Choi, J. Y. J. Am. Chem.
Soc. 1999, 121, 5821–5822. (f) Lee, H. M.; Nolan, S. P. Org. Lett.
2000, 2, 2053–2055. (g) Itami, K.; Nokami, T.; Ishimura, Y.; Mitsudo,
K.; Kamei, T.; Yoshida, J. J. Am. Chem. Soc. 2001, 123, 11577–11585,
Tin: (h) Littke, A. F.; Fu, G. C. Angew. Chem., Int. Ed. 1999, 38,
2411–2413. (i) Grasa, G. A.; Nolan, S. P. Org. Lett. 2001, 3, 119–
122. (j) Mee, S. P. H.; Lee, V.; Baldwin, J. E. Angew. Chem., Int. Ed.
2004, 43, 1132–1136.
(26) (a) Corriu, R. J. P.; Masse, J. P. J. Chem. Soc. Chem. Commun. 1972,
144. (b) Tamao, K.; Sumitani, K.; Kumada, M. J. Am. Chem. Soc.
1972, 94, 4374–4376. (c) Tamao, K.; Sumitani, K.; Kiso, Y.;
Zembayashi, M.; Fujioka, A.; Kodama, S.; Nakajima, I.; Minato, A.;
Kumada, M. Bull. Chem. Soc. Jpn. 1976, 49, 1958–1969. (d) Bo¨hm,
V. P. W.; Weskamp, T.; Gsto¨ttmayr, C. W. K.; Herrmann, W. A.
Angew. Chem., Int. Ed. 2000, 39, 1602–1604. (e) Bo¨hm, V. P. W.;
Gsto¨ttmayr, C. W. K.; Weskamp, T.; Herrmann, W. A. Angew. Chem.,
Int. Ed. 2001, 40, 3387–3389. (f) Dankwardt, J. W. Angew. Chem.,
Int. Ed. 2004, 43, 2428–2432. (g) Ackermann, L.; Born, R.; Spatz,
J. H.; Meyer, D. Angew. Chem., Int. Ed. 2005, 44, 7216–7219. (h)
Yoshikai, N.; Mashima, H.; Nakamura, E. J. Am. Chem. Soc. 2005,
127, 17978–17979. (i) Matsubara, K.; Ueno, K.; Shibata, Y. Orga-
nometallics 2006, 25, 3422–3427.
(27) (a) Quintin, J.; Franck, X.; Hocquemiller, R.; Figade´re, B. Tetrahedron
Lett. 2002, 43, 3547–3549. (b) Sapountzis, I.; Lin, W. W.; Kofink,
C. C.; Despotopoulou, C.; Knochel, P. Angew. Chem., Int. Ed. 2005,
44, 1654–1658. (c) Kofink, C. C.; Blank, B.; Pagano, S.; Go¨tz, N.;
Knochel, P. Chem. Commun. 2007, 1954–1956.
(25) (a) Korn, T. J.; Cahiez, G.; Knochel, P. Synlett 2003, 1892–1894. (b)
Ohmiya, H.; Yorimitsu, H.; Oshima, K. Chem. Lett. 2004, 33, 1240–
1241. (c) Korn, T. J.; Knochel, P. Angew. Chem., Int. Ed. 2005, 44,
2947–2951. (d) Korn, T. J.; Schade, M. A.; Wirth, S.; Knochel, P.
Org. Lett. 2006, 8, 725–728. (e) Korn, T. J.; Schade, M. A.; Cheemala,
M. N.; Wirth, S.; Guevara, S. A.; Cahiez, G.; Knochel, P. Synthesis
2006, 3547–3574.
(28) (a) Norinder, J.; Matsumoto, A.; Yoshikai, N.; Nakamura, E. J. Am.
Chem. Soc. 2008, 130, 5858–5859. (b) Yoshikai, N.; Matsumoto, A.;
Norinder, J.; Nakamura, E. Angew. Chem., Int. Ed. 2009, 48, 2925–
2928.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11951

A R T I C L E S
Hatakeyama et al.
Table 1. Effect of Iron Salts and Additives on the Cross-Coupling
of Chlorobenzene with p-TolMgBr
of anhydrous FeF3 and SIPr ·HCl (1:2 ratio), finely ground under
an inert atmosphere, showed the comparable reactivity with the
combination of FeF3 ·3H2O and SIPr ·HCl (data not shown). In
the presence of FeCl3 or Fe(acac)3, homocoupling was pre-
dominant with or without SIPr ·HCl (entries 18-20). Pretreat-
ment of FeCl3 with KF also generated a catalytically active
species, which gave the unsymmetrical biaryl with the same
efficiency of the catalyst prepared from the hydrates of iron
fluorides (entry 21). This experiment indicates that the involve-
ment of H2O or metal hydroxide to the catalytic C-C bond-
forming process is likely very minimal.
% yieldsb
RSMc bitolyld
entrya
iron salt
additive (mol %)
2
3
1
4
1
2
3
4
5
6
7
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
SIPr·HCl (15)
SIMes ·HCl (15)
IPr·HCl (15)
SIPr·HBF4 (15)
I-t-Bu·HCl (15)
SIPr·HCl (10)
SIPr·HCl (5)
98 trace
0
4
5
4
6
6
4
5
34 trace 64
25 trace 71
39 trace 48
51 trace 48
93 trace
61 trace 36
Cobalt- and Nickel Fluorides/NHC-Catalyzed Biaryl Cross-
Coupling Reactions. We next focused on cobalt30 and nickel31
catalysts. Although various salts and complexes of these metals
have been widely used as highly reactive metal catalysts for
various C-C bond-forming reactions, the corresponding metal
fluorides have only been minimally studied in the cross-coupling
field.32 We examined the reaction of chlorobenzene 1 with
p-TolMgBr in the presence of catalytic amounts of cobalt and
nickel fluorides with various NHCs and confirmed that high
selectivity for the unsymmetrical biaryl can also be attained with
these iron-group metal fluorides (Table 2).
5
8
9
10
11
12
13
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
FeF3 ·3H2O
none
PCy3 (10)
DPPE (10)
6
5
0
1
1
1
trace 93
4
4
3
4
6
8
0
0
0
0
0
87
100
99
95
95
DPPF (10)
phenanthroline (5)
TMEDA (2.5 equiv)
14
15
16
17e
18
19
20
21f
FeF2 ·4H2O
FeF3
SIPr·HCl (15)
SIPr·HCl (15)
SIPr·HCl (15)
96 trace
29 trace 69
18 trace 81
2
4
2
1
The reaction with CoF2 ·4H2O (3 mol %) and IPr ·HCl (6
mol %) gave 2 in a higher yield than 3 and 4 (95%, 3%, and
11% yields, respectively, entry 1), while the reaction with
CoCl2 ·6H2O gave 2 with a considerable amount of 3 and 4
(68%, 11%, and 15% yields, respectively, entry 2). CoF2 and
CoF3 showed lower catalytic activity that we attribute to their
low solubility, similar to the anhydrous iron fluorides (entries
3 and 4). IPr ·HCl emerged as the most effective NHC precursor
rather than SIPr ·HCl, as in entry 5. A similar “fluoride effect”
was observed using nickel catalysts. The reaction with NiF2 or
NiF2 ·4H2O (1 mol %) and IPr ·HCl (2 mol %) gave 2
selectively, but using NiCl2 or NiCl2 ·6H2O resulted in less
selectivity (entries 6-9). In the presence of a reduced amount
of NiF2 (0.5 mol %), the reaction was complete after 48 h,
affording 98% yield of 2 (entry 10). In contrast to the iron
catalyst, the counteranion of the NHC precursors did not affect
the product yield (entry 11). Decreasing the steric bulkiness of
the N-aryl substituent led to a decrease in product yield (entries
12 and 13). SIPr ·HCl was slightly less effective than IPr ·HCl
(entry 14). It is noteworthy that addition of KF to NiCl2
improved the cross-/homo-selectivity in the reaction of bro-
mobenzene with p-TolMgBr (entries 15-17). From these results,
we chose SIPr ·HCl for iron fluoride and IPr ·HCl for cobalt
FeF2
FeF3
FeCl3
Fe(acac)3
FeCl3
H2O (15), SIPr ·HCl (15) 40 trace 57
1
SIPr·HCl (15)
SIPr·HCl (15)
none
32
26
17
2
2
2
1
10
18
17
0
32
29
33
8
FeCl3
KF (20), SIPr ·HCl (15) 92
a Reactions were carried out on a 0.4-1.0 mmol scale. b The yield
was determined by GC analysis using undecane as an internal standard.
c Recovery of starting material. d Based on the amount of p-TolMgBr
used. e Iron salt and additives were mixed in THF for 10 min at room
temperature prior to the reaction. f FeCl3 was treated with KF in MeOH/
THF, which were then removed in vacuo.
Chart 1. Imidazolium and Imidazolinium Salts Examined as NHC
Ligand Precursors
(30) Recent reviews: (a) Shinokubo, H.; Oshima, K. Eur. J. Org. Chem.
2004, 2081–2091. (b) Blekkan, E. A.; Borg, O.; Froeseth, V.; Holmen,
A. Catalysis 2007, 20, 13–32. (c) Gosmini, C.; Be´gouin, J.-M.;
Moncomble, A. Chem. Commun. 2008, 3221–3233. (d) Hess, W.;
Treutwein, J.; Hilt, G. Synthesis 2008, 3537–3562. (e) Jeganmohan,
M.; Cheng, C.-H. Chem. Eur. J. 2008, 14, 10876–10886.
16). Note that the addition of 15 mol % of water into FeF3
increased the product yield to a limited extent (entry 17). We
assume that water or hydroxide can react with the solid surface
of FeF3 and make it partially soluble to promote the generation
of catalytically active species to some extent.29 In fact, a mixture
(31) Recent reviews: (a) Suginome, M.; Ito, Y. J. Organomet. Chem. 2003,
680, 43–50. (b) Montgomery, J. Angew. Chem., Int. Ed. 2004, 43,
3890–3908. (c) Netherton, M. R.; Fu, G. C. AdV. Synth. Catal. 2004,
346, 1525–1532. (d) Tamaru, Y., Ed.; Modern Organonickel Chem-
istry; Wiley-VCH: Weinheim, 2005. (e) Montgomery, J.; Sormunen,
G. J. Top. Curr. Chem. 2007, 279, 1–23. (f) Kimura, M.; Tamaru, Y.
Top. Curr. Chem. 2007, 279, 173–207. (g) Mori, M. Eur. J. Org.
Chem. 2007, 4981–4993. (h) Ng, S.-S.; Ho, C.-Y.; Schleicher, K. D.;
Jamison, T. F. Pure Appl. Chem. 2008, 80, 929–939. (i) Nakao, Y.;
Hiyama, T. Pure Appl. Chem. 2008, 80, 1097–1107. (j) Hirano, K.;
Yorimitsu, H.; Oshima, K. Chem. Commun. 2008, 3234–3241.
(32) Reviews for transition-metal fluoride: (a) Doherty, N. M.; Hoffmann,
N. W. Chem. ReV. 1991, 91, 553–573. (b) Murphy, E. F.; Murugavel,
R.; Roesky, H. W. Chem. ReV. 1997, 97, 3425–3468. (c) Mezzetti,
A.; Becker, C. HelV. Chem. Acta 2002, 85, 2686–2703.
(29) In FeF3 ·3H2O and FeF2 ·4H2O, each iron atom is surrounded not only
by fluorine atoms but also by water molecules in the form of a nearly
regular octahedron. It can account for higher solubility of FeF3 ·3H2O
and FeF2 ·4H2O than that of FeF3 and FeF2, in which an iron atom is
surrounded by six fluorine atoms: (a) Hepworth, M. A.; Jack, K. H.;
Peacock, R. D.; Westland, G. J. Acta Crystallogr. 1957, 10, 63–69.
(b) Penfold, B. R.; Taylor, M. R. Acta Crystallogr. 1960, 13, 953–
955. (c) Teufer, G. Acta Crystallogr. 1964, 17, 1480. (d) Jørgensen,
J.-E.; Smith, R. I. Acta Crystallogr. 2006, B62, 987–992.
9
11952 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Metal Fluoride/NHC To Catalyze Biaryl Cross-Coupling
A R T I C L E S
Table 2. Screening of Cobalt and Nickel Salts in the Presence of
Various NHC Ligands
Table 3. Fluoride Effect on Ni- and Co-Catalyzed Biaryl
Cross-Coupling in the Presence of Phosphine Ligands
% yieldsb
RSMc bitolyld
% yieldsb
entrya metal salt (mol %)
additive (mol %)
time (h)
2
3
1
4
RSM bitolyl
cPh-Br
entrya metal salt (mol %)
ligand (mol %) time (h)
2
3
d4
1
2
3
4
5
CoF2 ·4H2O (3)
CoCl2 ·6H2O (3)
CoF2 (3)
CoF3 (3)
CoF2 ·4H2O (3)
IPr ·HCl (6)
IPr ·HCl (6)
IPr ·HCl (6)
IPr ·HCl (6)
SIPr ·HCl (6)
48 95
3
0
2
11
15
2
1
2
3
4
5
6
7
8
9
NiF2 (1)
NiF2 (1)
NiF2 (1)
NiCl2 (1)
NiF2 (1)
NiF2 (1)
NiF2 (1)
NiCl2 ·dppp (1)
NiF2 (0.5)
none
PPh3 (1)
PPh3 (4)
PPh3 (4)
DPPE (1)
DPPP (1)
DPPB (1)
none
22
16
16
1
16
16
16
4
59
57
93
32
15
6
0
0
0
40
25
14
20
3
3
3
13
2
48 68 11
48 21 trace 79
48 33 58
1
5
5
84
8
7
48 39 trace 50
trace
trace
1
0
0
1
99
99
99
0
6
7
8
NiF2 (1)
NiF2 ·4H2O (1)
NiCl2 (1)
IPr ·HCl (2)
IPr ·HCl (2)
IPr ·HCl (2)
IPr ·HCl (2)
IPr ·HCl (1)
IPr·HBF4 (1)
I-t-Bu·HCl (1)
IMes·HCl (1)
SIPr·HCl (1)
24 98 trace trace
24 96 trace
15 64 18
3
1
23
22
2
1
5
92
7
IPr ·HCl (1)
3
99 trace
0
9
NiCl2 ·6H2O (1)
15 60 18 11
10
11
12
13
14
NiF2 (0.5)
48 98 trace
48 98 trace
48 95 trace
0
0
3
10
11
CoF2 ·H2O (1)
CoF2 ·4H2O (1)
none
DPPE (1)
16
16
3
12 trace
64
93
0
20
4
7
NiF2 (0.5)
NiF2 (0.5)
NiF2 (0.5)
NiF2 (0.5)
1
3
1
99
0
0
12 CoF2 ·4H2O (0.5) IPr ·HCl (1)
48
7
trace 92
2
48 94
2
4
10
a Reactions were carried out on a 1.0-3.0 mmol scale. b The yield
was determined by GC analysis using undecane as an internal standard.
c Recovery of starting material. d Based on the amount of p-TolMgBr
used.
15e
16e
17e
NiCl2 (1)
NiCl2 (1)
NiCl2 (1)
IPr ·HCl (2)
KF (2), IPr ·HCl (2)
KF (4), IPr ·HCl (2)
1
1
1
83 14
95
96
0
0
0
20
5
4
3
2
a Reactions were carried out on a 1.0-3.0 mmol scale. b The yield
was determined by GC analysis using undecane as an internal standard.
c Recovery of starting material. d Based on the amount of p-TolMgBr
used. e Bromobenzene was used instead of chlorobenzene.
yield of 2 (92%), albeit with a considerable by-production of 3
and 4 (7% and 13% yields, respectively, entry 8). As in entry
9, excellent yield and selectivity were attained by using 0.5 mol
% of NiF2 and 1 mol % of IPr ·HCl. Analogous to the nickel
fluoride catalysis, DPPE prevented the coupling reaction with
cobalt fluoride, while IPr dramatically accelerated the reaction
(entries 10-12). In combination with the data for the iron
fluoride catalysts (Table 1, entries 8-11), we conclude that the
prevailing chelating diphosphines are ineffective ligands for the
present metal fluoride-catalyzed cross-coupling reactions.
Preparation of Active Metal Catalyst by Pretreatment
with Alkyl Grignard Reagents. Having confirmed the unique
high selectivity of the cross-coupling catalysts consisting of an
iron-group metal fluoride and an NHC, we sought to demonstrate
the synthetic versatility and practicality of this novel biaryl cross-
coupling reaction. The aforementioned procedure requires an
excess amount of arylmagnesium bromide to achieve complete
conversion of aryl halide. For example, the cross-coupling
between chlorobenzene 1 and p-TolMgBr using a reduced
amount of iron fluoride catalysts (3 mol % vs 5 mol %) and
Grignard reagents (1.5 equiv vs 2.5 equiv) did not go to
completion and gave the cross-coupling product in only 72%
yield after 24 h at 60 °C (entry 1 in Table 4 vs entry 1 in Table
1). We assumed that insufficient basicity of the Grignard reagent
toward deprotonation of NHC precursors and hydrates of iron
or cobalt fluorides resulted in the slow conversion of substrate.34
After screening various bases, MeMgBr and EtMgBr were found
to be suitable for the activation of iron and cobalt composites
(entries 2-8, Table 4). The treatment of FeF3 ·3H2O and
SIPr·HCl with 18 mol % of MeMgBr to quench all protons
derived from catalyst precursors enhanced the reaction rate to
give 2 in 97% yield with 1% recovery of 1 (entry 2). Additional
MeMgBr slightly enhanced the reaction rate (entry 3). With
EtMgBr (18 mol %), the reaction completed to give 98% yield
using 1.2 equiv of p-TolMgBr (entry 4). When 27 mol % of
and nickel fluorides as the standard NHC ligand in the ensuing
investigations.
Biaryl Cross-Coupling Reactions in the Presence of
Phosphine Lignads. Since the seminal work reported by Kumada
and Tamao,26b,c,33 phosphine ligands such as chelating 1,2-
bis(diphenylphosphino)ethane (DPPE) and 1,3-bis(diphenylphos-
phino)propane (DPPP), have been known to be effective and
widely used in the cross-coupling reactions with nickel as well
as palladium catalysts. We therefore compared these orthodox
diphosphine and monophosphine ligands with the NHC ligands
in the nickel- or cobalt fluoride-catalyzed biaryl cross-coupling
reactions (Table 3).
Cross-coupling reaction between bromobenzene and p-
TolMgBr was catalyzed by NiF2 in the absence of an auxiliary
ligand to give a mixture of cross- and homocoupling products
(2, 3, and 4) as in entry 1. Triphenylphosphine (PPh3) did not
accelerate cross-coupling but reduced the amounts of both
homocoupling products 3 and 4 to some extent (entry 2).
Addition of 4 equiv of PPh3 to NiF2 improved the cross-/homo-
selectivity (entry 3). NiCl2 catalyzed the coupling reaction with
a slightly lower selectivity under the same conditions (entry
4). In sharp contrast to the earlier studies on the Kumada
coupling, the chelating diphosphine ligands, DPPE, DPPP, and
DPPB, prevented the cross-coupling reaction in the presence
of NiF2 (entries 5-7). These results can be explained by the
“fluoride-effect,” where the strong coordination of two fluoride
ions and bisphosphine to the nickel center occupies four
coordination sites, and therefore, one coordination site remains
as a reactive site, which would be not suitable for the cross-
coupling reaction (see the mechanistic studies below). Note that
the representative nickel catalyst, NiCl2 ·dppp, gave sufficient
(34) In situ generation of NHC ligand: Herrmann, W. A. Angew. Chem.,
Int. Ed. 2002, 41, 1290–1309, and references cited therein.
(33) Kumada, M. Pure. Appl. Chem. 1980, 52, 669–679.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11953

A R T I C L E S
Hatakeyama et al.
Table 4. Effect of MeMgBr and EtMgBr as an Activating Agent
Table 5. Leaving Group Capabilities
metal salt (mol %)
entrya (S) IPr·HCl (mol %)
% yieldb RSMc bitolyld
% yieldsb
RSMc bitolyld
Ph-X
R
X
Y
time (h)
2
1
4
entrya metal fluoride (mol %) (S) IPr·HCl (mol %)
X
time (h)
2
3
4
1
2
3
4
5
FeF3 ·3H2O (3)
SIPr·HCl (9)
FeF3 ·3H2O (3) Me 18 1.2
SIPr·HCl (9)
FeF3 ·3H2O (3) Me 27 1.2
SIPr·HCl (9)
FeF3 ·3H2O (3)
SIPr·HCl (9)
FeF3 ·3H2O (3)
SIPr·HCl (9)
-
0
1.5
4
24
4
24
4
24
4
12
4
32
72
52
97
78
97
92
98
89
96
64
25
45
1
16
trace
6
2
3
2
3
1
3
2
3
1
2
1
FeF3 ·3H2O (3)
SIPr ·HCl (9)
F
24
0
0
>99 trace
2
3
4
5
6e
7
8
FeF3 ·3H2O (3)
SIPr ·HCl (9) Br 24 28
1
0
0
0
0
0
0
0
36
7
2
4
2
CoF2 ·4H2O (0.5) IPr ·HCl (1) Br
3
3
3
3
99 trace
99 trace
NiF2 (0.5)
NiF2 (0.5)
NiF2 (0.5)
NiF2 (1.0)
NiCl2 ·dppp (1)
IPr·HCl (1) Br
IPr·HCl (0.5) Br
IPr·HCl (0.5) Br
98
2
Et 18 1.2
99 trace
0
9e
none
-
Br 22 59 32
40
13
Et 27 1.2
0e
Br
4
92
7
24
9
FeF3 ·3H2O (3)
SIPr ·HCl (9)
I
I
I
24 23
2
0
38
44
11
6
7
8
CoF2 ·4H2O (3)
IPr·HCl (6)
CoF2 ·4H2O (3) Me 18 1.2
IPr·HCl (6)
CoF2 ·4H2O (3) Et 18 1.2
IPr·HCl (6)
-
0
1.5
4
24
4
24
4
11
56
68
93
55
98
84
42
32
4
42
2
4
7
4
4
3
3
10 CoF2 ·4H2O (0.5) IPr ·HCl (1)
3
1
30 trace 10
92
11f
NiF2 (0.5)
IPr·HCl (1)
1
0
a Reactions were carried out on a 1.0-3.0 mmol scale. b The yield
was determined by GC analysis using undecane as an internal standard.
c Recovery of starting material. d Based on the amount of p-TolMgBr
used. e p-TolMgBr was added slowly over 3 h at 60 °C. f 3 mol %
MeMgBr was added and stirred for 3 h at 60 °C before addition of
iodobenzene and p-TolMgBr.
24
9
NiF2 (1)
IPr·HCl (2)
NiF2 (1)
-
0
2
1.2
1.2
4
24
4
65
98
60
94
33
1
39
2
1
1
2
2
10
Me
IPr·HCl (2)
24
Fluorobenzene was fully inert under these conditions (entry
1). While the reaction of bromobenzene with the iron catalyst
afforded predominantly the homocoupling product 4, the reac-
tion with cobalt or nickel catalyst gave 2 selectively (entries
2-4). As shown in entry 5, one equivalent of IPr ·HCl for NiF2
was ample for complete conversion, but not for complete
selectivity. Interestingly, the selectivity was improved by the
slow addition of p-TolMgBr to the reaction mixture (3 h at 60
°C, entry 6). The reaction without IPr ·HCl was found to be
slower and significantly less selective to give 2, 3, and 4 in
59%, 32%, and 40% yields, respectively (entry 7). Note that
NiCl2 ·dppp, a representative catalyst for the Kumada coupling,
gave 2 with lower selectivity under the same conditions (entry
8). These results indicate that the NHC ligand (IPr or SIPr) not
only accelerates the oxidative addition step due to its electron
donation toward the metal center but also suppresses the
homocoupling reaction. Bulky NHC ligands, such as IPr or SIPr,
would prevent the formation of Ar3M or rapid transmetalation
(and the following nonselective coupling reaction) due to its
steric hindrance. The reaction of iodobenzene showed the lowest
selectivity in all cases (entries 9-11). Iodobenzene is a stronger
oxidant than bromo- or chlorobenzene and can undergo oxida-
tive addition to form a low-valent metal species that does not
have an NHC ligand, which could subsequently form an Ar3M
species.
Regarding the substrate scope, the metal fluoride-catalyzed
cross-coupling reaction can be widely applicable to a broad
range of substrates (Table 6). We carried out coupling reactions
using iron, cobalt, and nickel catalysts according to the
procedure in entries 4, 8, and 9 in Table 4, respectively.
Electron-rich 4-chloroanisole reacted smoothly with p-TolMgBr
to give the desired product in 92%, 94%, and 88% yields,
respectively (entry 1). Fluorine-substituted biaryls, the repre-
sentative mesogen structure of liquid crystal molecules, can be
synthesized with 1-chloro-4-fluorobenzene, 1-chloro-3,4-di-
fluorobenzene, and 1-bromo-3,5-difluorobenzene in good to
excellent yields with proper choice of a metal catalyst (entries
a Reactions were carried out on a 0.4-2.0 mmol scale. b The yield
was determined by GC analysis using undecane as an internal standard.
c Recovery of starting material. d Based on the amount of p-TolMgBr
used. e Ethylbenzene was obtained in 3% yield.
EtMgBr was used for the generation of the active catalyst,
ethylbenzene was obtained in 3% yield via cross-coupling of
the residual EtMgBr and 1 (entry 5). Given the lower yield (2%)
of the homocoupling product 4, we assume that some of the
excess EtMgBr is consumed by the partial reduction of the iron
(III) to iron (II).35 Similarly, pretreatment of CoF2 ·4H2O (3 mol
%) and IPr ·HCl (6 mol %) with 18 mol % of MeMgBr or
EtMgBr afforded a considerable enhancement of the reaction
rate (entries 6-8). In contrast, NiF2 (1 mol %) did not require
the treatment with the alkyl Grignard reagents, and almost
complete conversion of 1 was observed either with or without
pretreatment (entries 9 and 10). Hence, we assume that EtMgBr
is basic enough to react with the water molecules of the metal
fluoride hydrates in the liquid-solid state, assisting in their
dissolution. In the following section, we display the scope and
limitations of the present coupling reactions according to the
procedures presented in entries 4, 8, and 9 respectively.
Selective Biaryl Cross-Coupling Reactions Catalyzed by
Iron-Group Metal Fluorides/NHC: Scope and Limitations. With
the optimal conditions in hand, we examined fluoro-, bromo-,
and iodobenzenes as electrophiles to determine the scope of
the leaving group in the present reactions, (Table 5). Reactions
with p-TolMgBr (1.2 equiv) were conducted at 60 °C in the
presence of FeF3 ·3H2O and SIPr ·HCl (3 and 6 mol %,
respectively), NiF2 and IPr·HCl (0.5 and 1 mol %, respectively),
or CoF2 ·4H2O and IPr ·HCl (0.5 and 1 mol %, respectively).
(35) Although alkyl Grignard reagents possessing b-hydrogens are known
to reduce some iron (III) and also iron (II) salts to zero or much lower
oxidation states (see ref 15o and references cited therein), we are
supposing the coordinating fluoride ion could hamper the full reduction
of the iron salts in the light of the stoichiometric control experiments
and theoretical studies described in the later part of the manuscript.
9
11954 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Metal Fluoride/NHC To Catalyze Biaryl Cross-Coupling
A R T I C L E S
Table 6. Biaryl Cross-Coupling Catalyzed by Combinations of an Iron-Group Metal Fluoride and an NHC Ligand
a Reactions were carried out on a 1.0-3.0 mmol scale following the procedure in entries 4 (Fe cat.), 8 (Co cat.), and 9 (Ni cat.) in Table 4 unless
d
otherwise noted. b Isolated yield. c The yield was determined by GC analysis using undecane as an internal standard. The yield was determined by H
1
NMR analysis using bromoform or 1,1,2,2-tetrachloroethane as an internal standard. e The reaction was carried out in toluene.
2-4). In the case of 1-chloro-3,4-difluorobenzene, the cobalt
catalyst gave a 97% yield of the desired product, but the nickel
catalyst gave a much lower yield with considerable generation
of defluorinated biaryl and tetraaryl compounds via C-F bond
cleavage. On the other hand, 1-bromo-3,5-difluorobenzene
selectively gave the desired product in the presence of the nickel
catalyst owing to much higher reactivity of the C-Br bond than
that of the C-F bond. The iron and cobalt catalysts gave 4,4′-
dimethylbiphenyl as the major product via the homocoupling
of p-TolMgBr (35% and 46%, respectively). Electron-deficient
4-fluorophenylmagnesium bromide gave the desired product in
a good yield with the iron and cobalt catalysts, but a much lower
yield with the nickel catalyst (entry 5). In the latter case, 40%
of 1-butyl-4-chlorobenzene was recovered, and 10% of 4-bu-
tylbiphenyl was also obtained via C-F bond cleavage. While
the reactions of o-tolyl- and mesitylmagnesium bromides were
rather slow because of their steric hindrance, elevated reaction
temperatures (80 and 120 °C) gave the corresponding products
in good to excellent yields (entries 6-8). In these cases, the
nickel catalyst showed higher catalytic activity than others.
1-Naphthylmagnesium bromide and 2-naphthylmagnesium bro-
mide took part in the iron-catalyzed coupling reaction (92%
and 96% yields, entries 9 and 10). The reactions of 2-naphth-
ylmagnesium bromide with the cobalt and nickel catalysts were
less selective, giving 70% and 82% yields of the desired product
and 24% and 17% of 2,2′-binaphthyl, respectively. As shown
in entry 11, the dimethylamino group seems to interfere with
the nickel-catalyzed coupling reaction, but not the iron-catalyzed
coupling reaction. Acetal functionality remained intact under
the reaction conditions (entry 12). Whereas the reaction of
1-bromo-2-(but-3-enyl)benzene and p-TolMgBr with the nickel
catalyst selectively gave the desired product, the reaction with
the iron catalyst gave only 18% yield with 68% yield of 4,4′-
dimethylbiphenyl via homocoupling reaction (entry 13).
Heteroaromatic nucleophiles, as well as electrophiles, took
part in the selective biaryl cross-coupling reaction (Table 7). In
the presence of the cobalt and nickel catalysts, 2-bromopyridine
reacted with p-TolMgBr at 60 °C to give the desired product in
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11955

A R T I C L E S
Hatakeyama et al.
Table 7. Biaryl Cross-Couplings of Heteroaromatic Substrates
a Reactions were carried out on a 1.0-3.0 mmol scale following the procedure in entries 4 (Fe cat.), 8 (Co cat.), or 9 (Ni cat.), Table 4 unless
otherwise noted. b The isolated yield. c The yield was determined by GC analysis using undecane as an internal standard. d The reaction was carried out
in toluene.
Table 8. Selective C-Cl Cleavage over C-S by Iron Catalyst
95% and 93% yields (entry 1). The reaction with the iron
catalyst was less selective, yielding 66% of the desired product
and 30% of 4,4′-dimethylbiphenyl via the homocoupling of
p-TolMgBr. In these cases, 2-bromopyridine was not recovered.
As shown in entry 2, the reaction between 3-bromopyridine and
p-TolMgBr took place selectively with the nickel catalyst, but
not with the iron or cobalt catalysts. The reaction with
p-anisylmagnesium bromide took place selectively with the
cobalt catalyst to give the desired product in 95% yield (entry
3). In the presence of the iron or nickel catalyst, the reaction
did not go to completion, and 2-chlorothiophene was recovered
(ca. 90% and 50%, respectively). 3-Bromothiophene gave the
corresponding coupling product in good yields in the presence
of the cobalt and nickel catalysts (entry 4). The reaction of
2-chloroquinoline with mesitylmagnesium bromide took place
at 100 °C to give the desired product in 82% yield (entry 5). In
the presence of the cobalt and nickel catalysts, the reaction of
2-bromopyridine with 2-thienylmagnesium bromide smoothly
took place at 80 °C to give 2-thiophen-2-yl pyridine in 99%
yield (entry 6).
% yields of producta
RSMb bitolylc
catalyst (X mol %)
6
7
8
2
1
5
4
FeF3 ·3H2O (6), SIPr ·HCl (18)
CoF2 ·4H2O (6), IPr ·HCl (12)
NiF2 (2), IPr ·HCl (4)
0
2
2
80
55
10 13
4
6
0
0
6
0
0
17
15
29
30
8
7
10
a The yield was determined by GC analysis using undecane as an
internal standard. b Recovery of starting material. c Based on the amount
of p-TolMgBr used.
the homocoupling of p-TolMgBr to give 4-chlorophenyl-
magnesium bromide. The nickel catalyst turned out to be the
best catalyst for the selective coupling to give a 4-chloro-
4′-methylbiphenyl 6 in 96% yield.37 In contrast, the cobalt
catalyst gave a moderate yield and selectivity.
Arenes bearing more than one leaving group can be
selectively cross-coupled with a suitable catalyst (Tables 8
and 9). In the presence of the iron catalyst, the reaction of
4-chlorothioanisole 5 with p-TolMgBr took place via a
selective C-Cl bond cleavage to give 4-methyl-4′-methyl-
sulfanylbiphenyl 7 in 80% yield. The cobalt- and nickel-
catalyzed reactions were less selective to afford 7 in 55%
and 10% yields, respectively, with considerable amounts of
side products (6, 8, 2, and 1) via C-S bond cleavage.36 Next,
we carried out the reaction of 1-bromo-4-chlorobenzene with
p-TolMgBr, as shown in Table 9. The reaction with the iron
catalyst stopped after initial turnovers, with formation of a
considerable amount of homocoupling product 4 and chlo-
robenzene 1. This suggests that 9 served as an oxidant for
The present method is also useful for multiple arylation
reactions of polychloroarenes. In the presence of the nickel
catalyst prepared from nickel fluoride (6 mol %), IPr ·HCl (12
mol %), and MeMgBr (24 mol %), the reaction between 1,2,4,5-
tetrachlorobenzene 11 and phenylmagnesium bromide (PhMgBr)
at 60-80 °C afforded 1,2,4,5-tetraphenylbenzene 12 in 71%
yield and 1,2,4-triphenylbenzene in 20% yield (eq 1). Note that
(37) (a) Ikoma, Y.; Taya, F.; Ozaki, E.-i.; Higuchi, S.; Naoi, Y.; Fuji-i, K.
Synthesis 1990, 147–148. (b) Littke, A. F.; Dai, C. Y.; Fu, G. C. J. Am.
Chem. Soc. 2000, 122, 4020–4028.
(36) Tiecco, M.; Testaferri, L.; Tingoli, M.; Chianelli, D.; Wenkert, E.
Tetrahedron Lett. 1982, 23, 4629–4632.
9
11956 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Metal Fluoride/NHC To Catalyze Biaryl Cross-Coupling
A R T I C L E S
Table 9. Selective C-Br Cleavage over C-Cl by Nickel Catalyst
Scheme 3. Reaction of Iron Salts with p-TolMgBr in the Presence
of SIPr ·HCl
% yields of producta
RSMb bitolylc
catalyst (X mol %)
6
10
8
2
1
9
4
Table 10. Reaction of Cobalt and Nickel Salts with p-TolMgBr in
the Presence of IPr ·HCl
FeF3 ·3H2O (6), SIPr ·HCl (18)
4
0
0
0
3
3
2
5
2
0
16
11
0
62
0
0
28
13
3
CoF2 ·4H2O (6), IPr ·HCl (12) 78
NiF2 (2), IPr ·HCl (4)
96
a The yield was determined by GC analysis using undecane as an
internal standard. b Recovery of starting material. c Based on the amount
of p-TolMgBr used.
bitolylb
4
entrya
metal salt
conditions
the reaction with 6 mol % of NiCl2 ·dppp was sluggish (12%
yield) under the same conditions.38,39
1
2
3
CoCl2 ·6H2O
NiCl2
CoF2 ·4H2O
0 °C, 1 h
0 °C, 1 h
0 °C, 1 h
60 °C, 1 h
0 °C, 1 h
60 °C, 1 h
60 °C, 4 h
99
98
0
8
0
4
NiF2
7
11
a Reactions were carried out on a 0.1 mmol scale. b The yield was
based on the amount of metal salts used and determined by GC analysis
using undecane as an internal standard.
temperature in the presence of an aryl halide. Table 10
summarizes the results of similar control experiments for cobalt
and nickel halides. Treatment of cobalt and nickel chlorides with
p-TolMgBr resulted in the rapid homocoupling of the Grignard
reagents (0 °C, 1 h) to give 4 in quantitative yield as in entries
1 and 2. On the other hand, the corresponding fluorides of these
metals did not react with the Grignard reagent under the same
conditions. Homocoupling product 4 was obtained in 8-11%
yield even at 60 °C.
In light of the thermal instability of homoleptic tetraphenyl-
ferrate species, such as [Ph4Fe]Li2,41 we considered that the
formation of the homoleptic metalate complex via the trans-
metalation and addition of an arylmagnesium reagent is
dominant in the reaction of the metal chlorides. On the other
hand, the sharp contrast observed in the reactivities of metal
fluorides (no significant biaryl formation) suggests that the
fluoride counterion may interfere with the formation of the fully
arylated metalate complex presumably due to its strong coor-
dination of fluoride to the iron-group metal center and high
electronegativity.
The results of the control experiments using the iron-group
metal(II) fluorides/NHC ligands (SIPr or IPr), and chlorobenzene
1 are summarized in Scheme 4: chlorobenzene 1 was recovered
in 95-99% without any formation of byproducts after 24 h at
60 °C. Notably, oxidative addition of the aryl halide to the
divalent metal fluorides does not occur when a Grignard reagent
is not added, even in the presence of NHC ligands (not the NHC
precursors). These results show that the generation of reactive
intermediates requires the reaction of corresponding metal
fluoride with an arylmagnesium reagent.
Mechanistic Considerations on the Catalytic “Fluoride
Effect”. To gain mechanistic insights into the origin of the high
cross-/homo-selectivity of the metal fluoride catalysts, we
conducted a series of control experiments and theoretical studies.
The results of control experiments using a variety of metal salts,
NHC ligand precursors (SIPr·HCl or IPr·HCl), p-TolMgBr are
summarized in Scheme 3 and Table 10. The reactions were
carried out in the absence of an aryl halide, which has been
reported to accelerate the reductive elimination of neutral
organonickel compounds,40 and we observed a stark difference
between the reactivities of the metal chloride and fluoride toward
the Grignard reagent. As shown in Scheme 3, treatment of 0.1
mmol of FeCl2 with p-TolMgBr (20 equiv) at 0 °C in the
presence of SIPr ·HCl (3.0 equiv) gave 0.096 mmol of 4,4′-
dimethylbiphenyl 4 (96% yield based on the amount of FeCl2).
Similarly, treatment of FeCl3 with p-TolMgBr under the same
conditions gave 1.5 times as much of the stoichiometric amount
of 4 as when FeCl2 was treated with the Grignard reagent. These
results clearly indicate that p-TolMgBr reduces the metal
chlorides to give the corresponding metal (0) species and the
biaryl product simultaneously, which corresponds to the initial
activation step in the cross-coupling reaction. On the contrary,
iron (II or III) fluorides did not afford 4 under the same
conditions. Only a small amount of 4 was formed at 60 °C (with
FeF2 and FeF3 in 6% and 11% yields, respectively), while the
biaryl cross-coupling reaction took place smoothly at the same
(38) An improved yield (58%) has been achieved using 7.5 mol % of
Ni(acac)2 and a hydroxyphosphine. See ref 26h.
(39) Palladium-catalyzed tetraarylation of 1,2,4,5-tetrabromobenzene has
been reported. (a) Berthiol, F.; Kondolff, I.; Doucet, H.; Santelli, M.
J. Organomet. Chem. 2004, 689, 2786–2798. (b) Li, Z. H.; Wong,
M. S.; Tao, Y. Tetrahedron 2005, 61, 5277–5285.
(41) Fu¨rstner et al. reported that a divalent tetraphenylferrate complex,
[Ph4Fe]Li2, is prone to decompose thermally to give biphenyl in the
presence or absence of an oxidant. See ref 15o.
(40) Uchino, M.; Yamamoto, A.; Ikeda, S. J. Organomet. Chem. 1970,
24, C63-C64.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11957

A R T I C L E S
Hatakeyama et al.
Scheme 4. Reaction of Metal Fluorides with Chlorobenzene in the
Presence of NHC Ligands
high-valent intermediate B, thereby, prevailing over undesirable
transmetalation or formation of -ate complexes carrying excess
aryl groups (such as Ar1Ar2 M, G).48 The (0)-(II) and (II)-(IV)
2
catalytic cycles have identical rate expressions and, thus, are
kinetically indistinguishable as we assumed that the oxidative
addition step is irreversible and the rate-limiting step.
In order to access the above-mentioned mechanistic consid-
erations, we carried out computational studies to evaluate
reaction pathways in the proposed mechanism starting from a
metal (II)-ate complex. We chose a chemical model consisting
of PhMgCl and PhCl, as the nucleophilic and electrophilic
counterparts, and NiF2 and I(Me) as the metal catalyst and the
auxiliary ligand (eq 2). A DFT calculation method widely used
for the mechanistic studies of the nickel catalyzed C-C bond
formations were employed. The theoretical model is described
as B3LYP/6-31G*.49
On the basis of these control experiments and the previously
suggested reaction mechanisms involving the organometalate
complexes of iron,15o cobalt,42 and nickel43 in cross-coupling
reactions, we propose several plausible mechanisms shown in
Figure 1. A metalate mechanism depicted in the left catalytic
cycle starts with the formation of a heteroleptic metal (II)-ate
complex A from a divalent metal fluoride and an arylmagnesium
reagent (Ar1MgX). The resulting heteroleptic metal (II)-ate
complex A undergoes oxidative addition with an aryl halide to
give an elusive higher-valent (formally IV oxidation state)
species (B) carrying Ar1 and Ar2.44 Subsequent reductive
elimination to give the unsymmetrical biaryl (Ar1-Ar2) gener-
ates a metal (II) complex C bearing two fluorides and one
halogen ligand derived from Ar2X on the metal center. The
reaction of C with Ar1MgX regenerates the reactive intermediate
A. A radical-type (II)-(III) mechanism has been reported for
iron-catalyzed cross-coupling reaction of alkyl halides with
arylmagnesium reagents.45 The catalytic cycle depicted on the
right side of Figure 1 shows a canonical “(0)-(II) mechanism”,
which consists of oxidative addition of an aryl halide to the
metal (0) intermediate D, transmetalation between arylmetal
halide E and Ar1MgX, and reductive elimination of Ar1-Ar2
from diarylmetal(II) F.46 We believe that the present reaction
based on the metal fluoride catalyst does not take place via the
popular (0)-(II) mechanism, but more likely via the higher-
valent metalate mechanism, which is analogous to those of
cuprate-mediated substitution reactions and catalytic cross-
coupling reactions.47 The final reductive elimination process in
the “metalate mechanism” is presumed to be much faster than
that of the “(0)-(II) mechanism” owing to instability of the
We could locate a single reaction pathway that connects the
formation of a σ-complex, the oxidative addition of chloroben-
zene (Ph-Cl), and the following reductive elimination of the
biphenyl (Ph-Ph). Figure 2 displays the energy profile for the
reaction pathway and Figure 3 summarizes the obtained
structural information of the equilibrium and transition structures
on the reaction coordinate. Nickelate complex (PhNiF2) interacts
with chlorobenzene to form σ-complex (CP1), which is 9.7 kcal/
mol higher in energy (all electron energy, ∆∆E) than the sum
of PhNiF2 and Ph-Cl. The higher (uphill) energetics is likely
due to the instability of the distorted square pyramidal structure
of CP1 (angles: F1-Ni-F2 ) 76.3°, Cl-Ni-F2 ) 75.0°,
Cl-Ni-C1 ) 175.4° in Figure 3). The bond lengths indicate a
weak electrostatic interaction between the nickel center and the
lone electron pair of the chloride ligand (Ni-Cl ) 2.690 Å,
C2-Cl ) 1.773 Å, cf. C-Cl of Ph-Cl ) 1.760 Å). The
oxidative addition of Ph-Cl via TS1 (C2-Cl ) 2.119 Å, Ni-C2
) 2.107 Å, Ni-Cl ) 2.313 Å) requires a reasonable activation
energy (∆Eq ) +18.3 kcal/mol) to form an octahedral Ni(IV)
intermediate, CP2 (Ni-C2 ) 1.945 Å, Ni-Cl ) 2.352 Å). This
transformation is endothermic (∆∆E ) +6.2 kcal/mol), mir-
roring the elusive nature of a tetravalent organonickel species.
The reductive elimination of Ph-Ph from CP2 successively
takes place via TS2 (C1-C2 ) 2.215 Å, Ni-C1 ) 1.983 Å,
Ni-C2 ) 1.978 Å) with a very small activation energy (∆Eq )
(42) Wakabayashi, K.; Yorimitsu, H.; Oshima, K. J. Am. Chem. Soc. 2001,
123, 5374–5375.
(43) (a) Terao, J.; Watanabe, H.; Ikumi, A.; Kuniyasu, H.; Kambe, N. J. Am.
Chem. Soc. 2002, 124, 4222–4223. (b) Terao, J.; Todo, H.; Watanabe,
H.; Ikumi, A.; Kambe, N. Angew. Chem., Int. Ed. 2004, 43, 6180–
6182.
(48) [Ar4FeIII][Li(THF)3]: (a) Alonso, P. J.; Arauzo, A. B.; Fornie´s, J.;
Garcia-Monforte, M. A.; Martin, A.; Martinez, J. I.; Menjo´n, B.; Rillo,
C.; Sa´iz-Garitaonandia, J. J. Angew. Chem., Int. Ed. 2006, 45, 6707–
6711[Me4FeII] [Li(OEt2)]2: (b) Fu¨rstner, A.; Krause, H.; Lehmann,
C. W. Angew. Chem., Int. Ed. 2006, 45, 440–444. Fe-F complex: (c)
Vela, J.; Smith, J. M.; Yu, Y.; Ketterer, N. A.; Flaschenriem, C. J.;
Lachicotte, R. J.; Holland, P. L. J. Am. Chem. Soc. 2005, 127, 7857–
7870.
(44) Organometal (IV) compounds: (a) Bower, B. K.; Tennent, H. G. J. Am.
Chem. Soc. 1972, 94, 2512–2514. (b) Byrne, E. K.; Theopold, K. H.
J. Am. Chem. Soc. 1989, 111, 3887–3896. (c) Dimitrov, V.; Linden,
A. Angew. Chem., Int. Ed. 2003, 42, 2631–2633. (d) Carnes, M.;
Buccella, D.; Chen, J. Y.-C.; Ramirez, A. P.; Turro, N. J.; Nuckolls,
C.; Steigerwald, M. Angew. Chem., Int. Ed. 2009, 48, 290–294.
(45) Noda, D.; Sunada, Y.; Hatakeyama, T.; Nakamura, M.; Nagashima,
H. J. Am. Chem. Soc. 2009, 131, 6078–6079.
(49) The DFT calculations were performed using the B3LYP hybrid
functional and the 6-31G* basis set. All stationary points were
optimized without any symmetry assumptions, and characterized all
by normal coordinate analysis at the same level of the geometry
optimization. Details are described in the Supporting Information.
Theoretical studies using DFT method on (0)-(II) mechanism for
nickel-catalyzed cross-coupling and related reactions: Reinhold, M.;
McGrady, J. E.; Perutz, R. N. J. Am. Chem. Soc. 2004, 126, 5268–
5267, and references cited therein. See also a recent paper: Yoshikai,
N.; Matsuda, H.; Nakamura, E. J. Am. Chem. Soc. 2008, 130, 15258–
15259.
(46) Although the canonical mechanism has been widely accepted for Pd-
catalyzed cross-coupling reactions, several alternatives have been
proposed for the cross-coupling reactions based on the first-row
transition metal catalysts including Fe, Co, and Ni. See refs 14f, 15o,
26c, 42, and 43.
(47) (a) Nakamura, E.; Mori, S.; Morokuma, K. J. Am. Chem. Soc. 1998,
120, 8273–8274. (b) Mori, S.; Nakamura, E.; Morokuma, K. J. Am.
Chem. Soc. 2000, 122, 7294–7307.
9
11958 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Metal Fluoride/NHC To Catalyze Biaryl Cross-Coupling
A R T I C L E S
Figure 1. Plausible mechanism: metal-fluoride-ate complex as a reactive intermediate.
Figure 2. Energy profiles for the metalate mechanism based on the DFT calculation (B3LYP/6-31G*). Relative electron energies based on PhNiF2 plus
Ph-Cl (∆E, kcal/mol) are shown in parentheses.
+3.5 kcal/mol) to form a square-planar Ni(II) complex PD1
with a stabilization energy of 49.4 kcal/mol. The calculated low
activation barrier of the reductive elimination and the large
exothermicity suggests that the reductive elimination pathway
should be more favorable than an alternative transmetalation
pathway of CP2 with arylmagnesium reagents (see below). In
light of the low activation barrier, one may expect that the
tetravalent nickel intermediate would not be an equilibrium
structure (a stationary point) when the NHC ligand carries
sterically demanding aryl groups on the nitrogen atoms as in
the real system.
We next examined the reaction of metal fluorides and
arylmagnesium reagents using a chemical model consisting of
a phenylmagnesium reagent, an NHC, and either nickel salts
of NiF2 or NiCl2. This system corresponds to the control
experiments described in Scheme 3 and Table 10. We adopted
I(Me)NiX2 (X ) Cl or F) as the models of the catalyst precursors
and examined the reaction with phenylmagnesium chloride
solvated with two molecules of dimethyl ether
(PhMgCl·2Me2O) at the same level of theory (eq 3).
In each case of I(Me)NiCl2 and I(Me)NiF2 (denoted as NiCl2
and NiF2, respectively), a single reaction pathway was obtained
for transmetalation process with PhMgCl ·2Me2O (Path-A and
Path-B, respectively) to form I(Me)NiPh2 (denoted as NiPh2).
The energy profiles for Path-A and Path-B are shown in black
and blue colors, respectively, in Figure 4. Shown as Path-C is
the subsequent reaction pathway from the common intermediate,
NiPh2 via a reductive elimination to give the biphenyl (Ph-Ph)
and I(Me)Ni(0)-biphenyl η4 complex (denoted as Ni0).50
The starting NHC/nickel chloride complex NiCl2 and
PhMgCl·2Me2O formed a nickelate complex (PhNiCl2) after
releasing one molecule of a solvent, Me2O, in a highly
exothermic manner (∆∆E ) -42.7 kcal/mol). The formation
of the neutral phenylnickel species I(Me)PhNiCl (denoted as
PhNiCl), upon the dissociation of MgCl2 ·2Me2O, was found
to be 10.0 kcal/mol greater in energy.51 The second molecule
of PhMgCl ·2Me2O and PhNiCl formed the nickelate intermedi-
ate carrying two phenyl groups (denoted as Ph2NiCl) in the
same manner as the preceding formation of PhNiCl2. This
process is also exothermic (∆∆E ) -28.7 kcal/mol). Diphe-
nylnickel coordinated by the NHC ligand, (denoted Ph2Ni)
formed upon the dissociation of MgCl2 ·2Me2O in an endother-
(50) Molecular structures and all electron energies of the stationary points
are described in the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11959

A R T I C L E S
Hatakeyama et al.
Figure 3. Reaction pathway for nickel fluoride-catalyzed cross-coupling based on the DFT calculation (B3LYP/6-31G*). Relative electron energies based
on PhNiF2 plus Ph-Cl (∆E, kcal/mol) are shown in parentheses. Bond lengths are given in Å. Hydrogen atoms are omitted for clarity.
Figure 4. Energy profiles for the transmetalation and reductive elimination processes based on the DFT calculation (B3LYP/6-31G*). Relative electron
energies based on Ph2Ni (∆E, kcal/mol) are shown in parentheses (Path-A, X ) Cl: black; Path-B, X ) F: blue; Path-C: black).
mic manner (∆∆E ) +10.9 kcal/mol). The diphenylnickel
species underwent reductive elimination to afford a biphenyl
complex of the nickel (0) species Ni0 via a three-centered
transition structure TS3 with a reasonable activation barrier and
exothermicity (∆Eq ) +15.0 kcal/mol and ∆∆E ) -10.7 kcal/
mol, respectively). The whole process starting from NiCl2 to
Ni0 is exothermic (∆∆E ) -61.2 kcal/mol) and the transforma-
tion from the initial nickelate complex PhNiCl2 into Ni0 is also
found to be exothermic (∆∆E ) -18.5 kcal/mol). Taking the
whole reaction pathway into consideration, the DFT study
suggests that the reduction of NiCl2 with phenyl Grignard
reagent should be facile in the presence of an NHC ligand, which
agrees well with the experimental results described above.
The DFT calculations for the NiF2 system provided us with
a similar sequence of transformations from the starting nickel
fluoride into the zero-valent nickel species Ni0, but its energy
profile was notably different from the one obtained for the NiCl2
system (Path-A vs Path-B). Transformation from NiF2 into
Ph2Ni as in Path-B thus consists of four steps: (1) formation of
the monophenyl nickelate complex PhNiF2 (∆∆E ) -71.2 kcal/
(51) Other geometrical isomers are less stable than PhNiX2. Details are
described in the Supporting Information.
9
11960 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Metal Fluoride/NHC To Catalyze Biaryl Cross-Coupling
A R T I C L E S
mol); (2) formation of a divalent phenyl nickel halide PhNiF
upon the dissociation of MgClF ·2Me2O (∆∆E ) +37.0 kcal/
mol); (3) formation of the diphenylnickelate complex Ph2NiF
(∆∆E ) -43.3 kcal/mol); (4) formation of diphenylnickel
species Ph2Ni upon the dissociation of MgClF ·2Me2O (∆∆E
) +17.6 kcal/mol). After the formation of Ph2Ni, each reaction
pathway of NiCl2 and NiF2 merged into Path-C, resulting in
the formation of Ph-Ph and Ni0 from Ph2Ni. Although the
overall process from NiF2 to Ni0 (Path-B and Path-C) is
exothermic in total (∆∆E ) -70.2 kcal/mol), energy minima
of Path-B are at the stages of the organonickelate intermediates,
PhNiF2 and Ph2NiF (∆∆E ) -1.0 and -6.9 kcal/mol,
respectively). The most significant difference between Path-A
and Path-B is the highly endothermic nature of the formation
of the phenylnickel halide intermediates PhNiF from the
nickelates PhNiF2 (∆∆E ) +37.0 kcal/mol vs +10.0 kcal/mol
for the formation of PhNiCl from PhNiCl2). Cleavage of a
strong Ni-F bond is energetically unfavorable and is presum-
ably the origin of the large endothermicity. The large electron
negativity of fluorine may account for the stabilization of the
fluorinated nickelate complex and also contribute in part to the
enhancement of the uphill energy difference. It is noteworthy
that the activation barrier of the reaction of PhNiF2and Ph-Cl
is much lower (∆Eq ) +28.0 kcal/mol as in Figure 2) than the
endothermicity of the PhNiF formation (∆∆E ) +37.0 kcal/
mol). The results of the DFT calculations show that PhNiF2 is
a reactive intermediate, which undergoes oxidative addition with
an aryl halide substrate, and hence, a prime candidate of
catalytically active species in the NiF2-NHC-catalyzed selective
biaryl cross-coupling.
Figure 5. Energy profiles for the oxidative-addition and transmetalation
processes based (B3LYP/6-31G*). Relative electron energies based on
PhFeF2_q plus Ph-Cl (∆E, kcal/mol) are shown in parentheses (singlet
state, S ) 0: green; triplet state, S ) 1: black; quintet state, S ) 2: blue).
We carried out the same DFT calculations adopting similar
chemical models for the iron and cobalt fluorides-NHC-
catalyzed biaryl coupling reactions. Because the difficulty of
the treatment of multiple spin-state systems of these metals has
been well-documented using the DFT method (a divalent iron
can take S ) 0, 1, and 2 spin states, and a divalent cobalt can
take S ) 1/2 and 3/2 spin states),52 we carefully conducted
computational studies on the critical early steps of the above-
presented metalate mechanism of these metals. We thus
compared the energies associated with the oxidative addition
of chlorobenzene to the heteroleptic metalate complexes (Ph-
MF2, M ) Fe or Co as shown in Figures 5 and 6, respectively)
and the transmetalation from the metalates. As shown in Figure
5, we could locate the reaction coordinates of the quintet state
(S ) 2) and the triplet state (S ) 1), but failed to obtain that of
the singlet state (S ) 0) for the iron catalyst. The oxidative
addition of Ph-Cl to the ferrate complex in the quintet state,
PhFeF2_q, took place via TS4_q with a reasonable activation
barrier (∆E‡ ) +29.5 kcal/mol) to give the tetravalent inter-
mediate CP3_q. The activation energy is almost comparable
to the endothermicity associated with the formation of the
divalent phenyliron fluoride intermediate PhFeF_q by the
Figure 6. Energy profiles for the oxidative-addition and transmetalation
processes (B3LYP/6-31G*). Relative electron energies based on PhCoF2_q
plus Ph-Cl (∆E, kcal/mol) are shown in parentheses (doublet state, S )
1/2: black; quartet state, S ) 3/2: blue).
dissociation of MgClF ·2Me2O (∆∆E ) +29.6 kcal/mol). In
the reaction coordinate of the triplet state (shown in black line
in Figure 5), it was found that the tetravalent intermediate CP3_t
is slightly more stable than the one in the quintet state (CP3_q,
∆∆Et-q ) -5.5 kcal/mol), whereas the other stationary points
(PhFeF_t, PhFeF2_t, and TS4_t) are more unstable than those
in the quintet state (∆∆Et-q ) +20.1, +9.6, and +0.1 kcal/
mol, respectively). We could locate only two stationary points
PhFeF_s and PhFeF2_s for the singlet state and confirmed the
relative energy of these structures were much less stable than
the corresponding structures in the triplet and quintet states.53
From these results, it is supposed that the most stable PhFeF2_q
should be a reactive intermediate for the oxidative addition and
the cross-coupling reaction should proceed in the quintet state
to form the elusive iron (IV) intermediates tentatively, which
may be prone to undergo a rapid reductive elimination to give
the cross-coupling product (Ph1-Ph2).
(52) Because the B3LYP functional have been known to overestimate the
stability of high spin states, we examined several functionals including
the one without HF exchange term (BP91) and the other with modified
admixture of HF term (B3LYP*) to obtain the virtually same results.
Further computational study on the present catalytic cross-coupling
reactions waits for the development of multiconfigurational treatment
such as CASSCF calculation using realistic chemical models, and of
course, it has been so far impractical. Zein, S.; Borshch, S. A.; Fleurat-
Lessard, P.; Casida, M. E.; Chermette, H. J. Chem. Phys. 2007, 126,
014105. For B3LYP* and the coefficient of exact exchange admixture
in DFT calculation affects relative energies of states of different
multiplicity; see: Reiher, M.; Salomon, O.; Hess, B. A. Theor. Chem.
Acc. 2001, 107, 48-55.
We could locate reaction coordinates of the quartet state
(S ) 3/2) and the doublet state (S ) 1/2) for the cobalt
catalyst to notice that the energy profiles are more compli-
cated. As shown in Figure 6, the potential energy surfaces
of the quartet state and the doublet state cross over during
(53) No reaction pathway for oxidative addition process was obtained in
singlet state.
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11961

A R T I C L E S
Hatakeyama et al.
the oxidative-addition process (the blue lines and the black
lines, respectively). The cobaltate complex in the quartet state,
PhCoF2_q, is found to be more stable than the one in the
doublet state, PhCoF2_d (∆∆Eq-d ) -14.6 kcal/mol). On
the other hand, at the transition structure of the oxidative
addition, the quartet state TS5_q is much less stable than
the doublet state TS5_d (∆∆Eq-d ) +16.6 kcal/mol).
Analogously, at the tetravalent cobaltate intermediate, the
quartet state CP4_q is much less stable than the doublet state
CP4_d (∆∆Eq-d ) +19.9 kcal/mol). These results indicate
that, during the oxidative addition process, a crossover of
reaction pathways between the quartet and doublet states may
take place with the C-Cl bond cleavage occurring via TS5_d
and leading the formation of CP4_d in the doublet state. The
energy difference between PhCoF2_q and TS5_d (∆∆E )
+28.7 kcal/mol) is substantially smaller than the endother-
micity of the formation of divalent phenylcobaltale fluoride
intermediate PhCoF_q by dissociation of MgClF · 2Me2O
(∆∆E ) +37.5 kcal/mol). This energy profile suggests that
the initial transmetalation of cobalt fluoride precatalyst is
sluggish and the cobaltate complex PhCoF2_q can be a
reactive intermediate for the cobalt fluoride-catalyzed biaryl
coupling reaction.
The resulting high-valent heteroleptic metalate complex
catalyzes the desired biaryl cross-coupling with high selectiv-
ity over the undesired homocoupling. We believe that the
concept of the synergistic effect of a fluoride ion and an
auxiliary ligand will further advance the development of
cross-coupling technologies, as well as relevant carbon-carbon
bond-forming reactions, which are currently under investiga-
tion in this laboratory.
Experimental Section
Synthesis of 4-Methylbiphenyl (2): A Representative Proce-
dure for the Iron-Catalyzed Reaction Shown in Tables 2, 4,
and 6-9. A THF solution of ethylmagnesium bromide (5.00 mL,
1.08 M, 5.40 mmol) was added to FeF3 ·3H2O (0.150 g, 0.90 mmol)
and 1,3-bis(2,6-diisopropylphenyl)imidazolinium hydrochloride
(1.15 g, 2.70 mmol) at 0 °C. THF (1.0 mL) was added to rinse the
vessel. After stirring at room temperature for 5 h, chlorobenzene
(3.38 g, 30.0 mmol) and a THF solution of p-tolylmagnesium
bromide (38.7 mL, 1.02 M, 36.0 mmol) was added at 0 °C. The
reaction was carried out at 60 °C for 24 h. After cooling to ambient
temperature, aqueous sodium potassium tartrate (saturated, 60.0 mL)
was added. The aqueous layer was extracted five times with hexane.
The combined organic extracts were filtered with a pad of Florisil.
GC analysis was carried out (98% yield) using undecane as an
internal standard. After the solvent was removed in Vacuo, the crude
product was purified by chromatography on silica gel (pentane) to
obtain the title compound (4.97 g, 98% yield, > 98% pure on GC
analysis) as a white solid.
Synthesis of 3,4-Difluoro-4′-methoxybiphenyl: A Representa-
tive Procedure for the Cobalt-Catalyzed Reaction Shown in
Tables 3-9. A THF solution of ethylmagnesium bromide (0.55
mL, 1.08 M, 0.60 mmol) was added to CoF2 · 4H2O (17.0 mg,
0.10 mmol) and 1,3-bis(2,6-diisopropylphenyl)imidazolium hy-
drochloride (85.4 mg, 0.20 mmol), and THF (1.5 mL) at 0 °C.
THF (0.5 mL) was added to rinse the vessel. After stirring at
room temperature for 4 h, 4-chloro-1,2-difluorobenzene (0.297
g, 2.0 mmol) and a THF solution of 4-methoxyphenylmagnesium
bromide (3.4 mL, 0.88 M, 3.0 mmol) was added at 0 °C. The
reaction was carried out at 60 °C for 12 h. After cooling to
ambient temperature, aqueous ammonium chloride (saturated,
2.0 mL) was added. The aqueous layer was extracted five times
with Et2O. The combined organic extracts were filtered with a
pad of Florisil. After the solvent was removed in Vacuo, the
crude product was purified by chromatography on silica gel (0%,
5% toluene in hexane) to obtain the title compound (0.428 g,
97% yield, > 98% pure on GC analysis) as a white solid.
Synthesis of 3,5-Difluoro-4′-methylbiphenyl: A Representa-
tive Procedure for the Nickel-Catalyzed Reaction Shown in
Tables 3-9. A THF solution of p-tolylmagnesium bromide (2.12
mL, 1.13 M, 2.4 mmol) was added to NiF2 (0.97 mg, 0.010 mmol)
and 1,3-bis(2,6-diisopropylphenyl)imidazolium hydrochloride (8.52
mg, 0.020 mmol), and 1-bromo-3,5-difluorobenzene (0.386 g, 2.0
mmol) at 0 °C. After stirring at room temperature for 0.5 h, the
reaction was carried out at 40 °C for 18 h. After cooling to ambient
temperature, aqueous ammonium chloride (saturated, 2.0 mL) was
added. The aqueous layer was extracted five times with Et2O. The
combined organic extracts were filtered with a pad of Florisil. After
the solvent was removed in Vacuo, the crude product was purified
by chromatography on silica gel (pentane) to obtain the title
compound (0.388 g, 95% yield, > 98% pure on GC analysis).
On the basis of the results of the control experiments and
the theoretical studies, we conclude that strong coordination of
the fluoride ion to the metal center suppresses the initial
transmetalation and following reduction processes and a result-
ing heteroleptic metalate (II) complex such as PhMF2 can
catalyze a biaryl cross-coupling reaction. A low activation
barrier for the reductive elimination in the present metalate
mechanism can account for the characteristically high cross-/
homo-selectivity.
Conclusion
We developed a selective and practical biaryl cross-
coupling reaction catalyzed by iron-group metal catalysts
based on the synergistic modulation by the fluoride counterion
and an NHC ligand. Experimental control reactions and
theoretical calculations suggest that the catalytic “fluoride
effect” underlies the observed high selectivity of cross-
coupling over homocoupling in the reaction. Each of the iron-
group metal catalysts possesses a characteristic catalytic
reactivityprofile;theappropriatechoiceofametalfluoride-NHC
ligand combination allows for the efficient coupling of a wide
range of aryl nucleophile and electrophiles in excellent yield
and high selectivity. First, the iron catalyst can achieve highly
selective coupling reactions using various aryl chlorides.
Second, the cobalt catalyst is particularly effective in
heteroaromatic couplings. And third, the nickel catalyst shows
high catalytic activity and a broad substrate scope when aryl
bromides and sterically hindered substrates are employed.
The present biaryl coupling features several synthetic ad-
vantages: (1) The catalytic method can achieve high-yielding
unsymmetrical biaryl synthesis with an excellent cross-
coupling selectivity. (2) The simple and scalable procedures
of this coupling method are suitable for large-scale produc-
tion. (3) This catalyst system is free of both palladium and
phosphine. As for the origin of the catalytic “fluoride effect”,
a noncanonical mechanism is proposed, in which strong
coordination of the fluoride ion to the divalent metal center
suppresses the reduction of the metal via the conventional
transmetalation reduction elimination process, and promotes
the formation of a high-valent heteroleptic metalate complex.
Acknowledgment. We thank Japan Society for the Promotion
of Science (JSPS) and the Ministry of Education, Culture, Sports,
Science and Technology of Japan (MEXT) for financial supports;
Grants-in-Aid for Scientific Research on Priority Areas “Synergistic
Effects for Creation of Functional Molecules” (M.N., 18064006)
and “Chemistry of Concerto Catalysis” (T.H., 20037033), and
9
11962 J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009

Metal Fluoride/NHC To Catalyze Biaryl Cross-Coupling
A R T I C L E S
Grant-in-Aids for Young Scientists from JSPS (S, M.N., 2075003;
B, T.H., 2175098). Financial supports from the Uehara Memorial
Foundation, Tosoh Finechem and Tosoh Corporations are gratefully
acknowledged. This work has been supported in part by Global
COE Program “International Center for Integrated Research and
Advanced Education in Materials Science” and the Joint Project
of Chemical Synthesis Core Research Institution from MEXT. We
thank Dr. J. Hong for helpful discussion on the reaction mechanism
and proofreading. We also thank Prof. H. Nagashima and Dr. Y.
Sunada for valuable discussion. We are also thankful for helpful
comments from reviewers and thorough proofreading by Dr. H.
Seike.
Supporting Information Available: Experimental procedure
and characterization data. This material is available free of
charge via the Internet at http://pubs.acs.org.
JA9039289
9
J. AM. CHEM. SOC. VOL. 131, NO. 33, 2009 11963