Synthesis and reactivity of low-coordinate iron(II) fluoride complexes and their use in the catalytic hydrodefluorination of fluorocarbons

Article abstract of DOI:10.1021/ja042672l

Transition metal fluoride complexes are of interest because they are potentially useful in a multitude of catalytic applications, including C-F bond activation and fluorocarbon functionalization. We report the first crystallographically characterized examples of molecular iron(II) fluorides: [L^MeFe(μ-F)]₂ (1₂) and L^tBuFeF (2) (L = bulky β-diketiminate). These complexes react with donor molecules (L'), yielding trigonal-pyramidal complexes L^RFeF(L'). The fluoride ligand is activated by the Lewis acid Et₂O·BF₃, forming L^tBuFe(OEt₂)(η¹-BF₄) (3), and is also silaphilic, reacting with silyl compounds such as Me ₃SiSSiMe₃, Me₃SiCCSiMe₃, and Et ₃SiH to give new thiolate L^tBuFeSSiMe₃ (4), acetylide L^tBuFeCCSiMe₃ (5), and hydride [L ^MeFe(μ-H)]₂ (6₂) complexes. The hydrodefluorination (HDF) of perfluorinated aromatic compounds (hexafluorobenzene, pentafluoropyridine, and octafluorotoluene) with a silane R₃SiH (R₃ = (EtO)₃, Et₃, Ph ₃, (3,5-(CF₃)₂C₆H ₃)Me₂) is catalyzed by addition of an iron(II) fluoride complex, giving mainly the singly hydrodefluorinated products (pentafluorobenzene, 2,3,5,6-tetrafluoropyridine, and α,α,α,2, 3,5,6-heptafluorotoluene, respectively) in up to five turnovers. These catalytic perfluoroarene HDF reactions proceed with activation of the C-F bond para to the most electron-withdrawing group and are dependent on the degree of fluorination and solvent polarity. Kinetic studies suggest that hydride generation is the rate-limiting step in the HDF of octafluorotoluene, but the active intermediate is unknown. Mechanistic considerations argue against oxidative addition and outer-sphere electron transfer pathways for perfluoroarene HDF. Fluorinated olefins are also hydrodefluorinated (up to 10 turnovers for hexafluoropropene), most likely through a hydride insertion/β-fluoride elimination mechanism. Complexes 1₂ and 2 thus provide a rare example of a homogeneous system that activates C-F bonds without competitive C-H activation and use an inexpensive 3d transition metal.

Full text of DOI:10.1021/ja042672l

Published on Web 05/05/2005
Synthesis and Reactivity of Low-Coordinate Iron(II) Fluoride
Complexes and Their Use in the Catalytic Hydrodefluorination
of Fluorocarbons
Javier Vela,† Jeremy M. Smith,†,‡ Ying Yu,† Nicole A. Ketterer,†
Christine J. Flaschenriem,† Rene J. Lachicotte,† and Patrick L. Holland*,†
Contribution from the Department of Chemistry, UniVersity of Rochester,
Rochester, New York 14627
Received December 6, 2004; E-mail: holland@chem.rochester.edu
Abstract: Transition metal fluoride complexes are of interest because they are potentially useful in a
multitude of catalytic applications, including C-F bond activation and fluorocarbon functionalization. We
report the first crystallographically characterized examples of molecular iron(II) fluorides: [LMeFe(µ-F)]2
(12) and LtBuFeF (2) (L ) bulky â-diketiminate). These complexes react with donor molecules (L′), yielding
trigonal-pyramidal complexes LRFeF(L′). The fluoride ligand is activated by the Lewis acid Et2O‚BF3, forming
LtBuFe(OEt2)(η1-BF4) (3), and is also silaphilic, reacting with silyl compounds such as Me3SiSSiMe3,
Me3SiCCSiMe3, and Et3SiH to give new thiolate LtBuFeSSiMe3 (4), acetylide LtBuFeCCSiMe3 (5), and hydride
[LMeFe(µ-H)]2 (62) complexes. The hydrodefluorination (HDF) of perfluorinated aromatic compounds
(hexafluorobenzene, pentafluoropyridine, and octafluorotoluene) with a silane R3SiH (R3 ) (EtO)3, Et3, Ph3,
(3,5-(CF3)2C6H3)Me2) is catalyzed by addition of an iron(II) fluoride complex, giving mainly the singly
hydrodefluorinated products (pentafluorobenzene, 2,3,5,6-tetrafluoropyridine, and R,R,R,2,3,5,6-heptafluo-
rotoluene, respectively) in up to five turnovers. These catalytic perfluoroarene HDF reactions proceed with
activation of the C-F bond para to the most electron-withdrawing group and are dependent on the degree
of fluorination and solvent polarity. Kinetic studies suggest that hydride generation is the rate-limiting step
in the HDF of octafluorotoluene, but the active intermediate is unknown. Mechanistic considerations argue
against oxidative addition and outer-sphere electron transfer pathways for perfluoroarene HDF. Fluorinated
olefins are also hydrodefluorinated (up to 10 turnovers for hexafluoropropene), most likely through a hydride
insertion/â-fluoride elimination mechanism. Complexes 12 and 2 thus provide a rare example of a
homogeneous system that activates C-F bonds without competitive C-H activation and use an inexpensive
3d transition metal.
aliphatic6 fluorocarbons by homogeneous transition metal
complexes.7 Metal fluoride complexes8-10 are often formed
Introduction
Fluorine’s peculiar characteristics such as high electronega-
tivity, low polarizability, and small covalent radius render
fluorocarbons thermally stable, water repellent, and resistant to
chemical degradation.1 These unique properties, along with the
great strength of the C-F bond (120-129 kcal/mol for aliphatic
and olefinic C-F bonds, and up to 154 kcal/mol in C6F6),2 make
fluorocarbons valuable refrigerants, pesticides, and nonadhesive
polymers but also very environmentally persistent. Therefore,
methods for chemically manipulating fluorocarbons are sought
to either degrade or add value to these materials.
during these C-F activation reactions, and they can potentially
be recycled to make these reactions catalytic. Scheme 1 shows
a few examples of well-studied C-F activation reactions. Often,
an early transition metal drives the reaction through formation
of a very strong M-F bond, but the resultant compound is inert
(for example Cp*2ZrHF).7a To catalytically functionalize C-F
(3) Representative examples of intramolecular C-F activation: (a) Richmond,
T. G. Angew. Chem., Int. Ed. 2000, 39, 3241-3244. (b) Hughes, R. P.;
Zhang, D.; Zakharov, L. N.; Rheingold, A. L. Organometallics 2002, 21,
4902-4904. (c) Hughes, R. P.; Laritchev, R. B.; Zakharov, L. N.;
Rheingold, A. L. J. Am. Chem. Soc. 2004, 126, 2308-2309. (d) Hughes,
R. P.; Lindner, D. C.; Rheingold, A. L.; Liable-Sands, L. M. J. Am. Chem.
Soc. 1997, 119, 11544-11545. (e) Hughes, R. P.; Smith, J. M. J. Am. Chem.
Soc. 1999, 121, 6084-6085. (f) Albietz, P. J.; Houlis, J. F.; Eisenberg, R.
Inorg. Chem. 2002, 41, 2001-2003.
(4) Representative examples of intermolecular aromatic C-F activation: (a)
Jasim, N. A.; Perutz, R. N.; Whitwood, A. C.; Braun, T.; Izundu, J.;
Neumann, B.; Rothfeld, S.; Stammler, H.-G. Organometallics 2004, 23,
6140-6149. (b) Edelbach, B. L.; Jones, W. D. J. Am. Chem. Soc. 1997,
119, 7734-7742. (c) Cronin, L.; Higgitt, C. L.; Karch, R.; Perutz, R. N.
Organometallics 1997, 16, 4920-4928. (d) Harrison, R. G.; Richmond,
T. G. J. Am. Chem. Soc. 1993, 115, 5303-5304. (e) Bennett, B. K.;
Harrison, R. G.; Richmond, T. G. J. Am. Chem. Soc. 1994, 116, 11165-
11166. (f) Whittlesey, M. K.; Perutz, R. N.; Moore, M. H. Chem. Commun.
1996, 787-788.
Great advances have been made in recent years on the intra-3
and intermolecular activation of aromatic,4 olefinic,5 and even
† University of Rochester.
‡ Current address: Department of Chemistry & Biochemistry, New
Mexico State University, Las Cruces, NM 88003.
(1) (a) Lemal, D. M. J. Org. Chem. 2004, 69, 1-11. (b) Sanford, G.
Tetrahedron 2003, 59, 437-454. (c) Hiyama, T. Organofluorine Com-
pounds: Chemistry and Applications; Springer: Berlin, 2000.
(2) (a) Kerr, J. A. In CRC Handbook of Chemistry and Physics, 71st ed.; Lide,
R. L., Ed.; CRC Press: Boca Raton, FL, 1990; pp 9.65-9.98. (b) Smart,
B. E. Mol. Struct. Energ. 1986, 3, 141-191. (c) Smart, B. E. In The
Chemistry of Functional Groups, Supplement D; Pati, S., Rappoport, Z.,
Eds.; John Wiley & Sons: New York, 1983; Chapter 14.
9
10.1021/ja042672l CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 7857-7870
7857

A R T I C L E S
Scheme 1
Vela et al.
Figure 1. Diketiminate ligands used in this work.
report that these complexes serve as valuable synthetic precur-
sors as well as precatalysts in the catalytic hydrodefluorination
of fluorocarbons.
Results and Discussion
Synthesis and Behavior of Diketiminate Iron(II) Fluorides.
We previously have reported three-coordinate iron(II) hydro-
carbyl complexes based on the diketiminate ligands LMe and
LtBu (Figure 1).18 These complexes serve as precursors to
discrete iron(II) fluorides by reaction with trimethyltin fluoride
(Scheme 2).19 Because Me3SnF is poorly soluble in toluene,
alkyl complexes of LMe are treated with an excess of the
insoluble fluorinating agent, and unreacted Me3SnF is removed
from the reaction mixture by filtration. In a typical reaction,
[LMeFeF]2 (12) is isolated in 83% yield as a bright green powder
that readily dissolves in hydrocarbon solvents such as diethyl
ether, toluene, or THF. Its dimeric nature is shown in the X-ray
crystal structure (Figure 2).
(8) (a) Pagenkopf, B. L.; Carreira, E. M. Chem. Eur. J. 1999, 5, 3437-3442.
(b) Grushin, V. V. Chem. Eur. J. 2002, 8, 1007-1014. (c) Verdaguer, X.;
Lange, U. E. W.; Reding, M. T.; Buchwald, S. L. J. Am. Chem. Soc. 1996,
118, 6784-6785.
(9) (a) Grushin, V. V.; Marshall, W. J. J. Am. Chem. Soc. 2004, 126, 3068-
3069. (b) Grushin, V. V. Angew. Chem., Int. Ed. 1998, 37, 994-996.
(10) (a) Murphy, E. F.; Murugavel, R.; Roesky, H. W. Chem. ReV. 1997, 97,
3425-3468. (b) Doherty, N. M.; Hoffman, N. W. Chem. ReV. 1991, 91,
553-573.
(11) Heterogeneous catalytic C-F activation: (a) Burdeniuc, J.; Crabtree, R.
H. Organometallics 1998, 17, 1582-1586. (b) Stanger, K. J.; Angelici, R.
J. J. Mol. Catal. A: Chem. 2004, 207, 59-68. (c) Kiplinger, J. L.;
Richmond, T. G. J. Am. Chem. Soc. 1996, 118, 1805-1806. (d) Yang, H.;
Gao, H.; Angelici, R. J. Organometallics 1999, 18, 2285-2287. (e) Young,
R. J.; Grushin, V. V. Organometallics 1999, 18, 284-296.
(12) Homogeneous catalytic C-F activation with transition metal complexes:
(a) Aizenberg, M.; Milstein, D. Science 1994, 265, 359-361. (b) Aizenberg,
M.; Milstein, D. J. Am. Chem. Soc. 1995, 117, 8674-8675. (c) Ishii, Y.;
Chatani, N.; Yorimitsu, S.; Murai, S. Chem. Lett. 1998, 157-158. (d) Braun,
T.; Perutz, R. N.; Sladek, M. I. Chem. Commun. 2001, 2254-2255. (e)
Kuhl, S.; Schneider, R.; Fort, Y. AdV. Synth. Catal. 2003, 345, 341-344.
(13) Silyl cations can mediate catalytic C-F activation: (e) Scott, V. J.;
C¸ elenligil-C¸ etin, R.; Ozerov, O. V. J. Am. Chem. Soc. 2005, 127, 2852-
2853.
bonds,11-14 late transition metals may be more suitable because
their M-F bonds are not prohibitively strong and also because
late metals are more tolerant of functional groups. The work of
Milstein in particular has shown the effectiveness of rhodium
compounds in the catalytic hydrodefluorination (HDF) of
aromatic perfluorocarbons.12a,b However, the examples of
catalytic C-F activation by homogeneous metal complexes are
few, and research is needed to progress toward applications.
As part of our research program on low-coordinate late metal
chemistry,15-17 we were attracted to low-coordinate iron fluoride
complexes, anticipating high reactivity due to the unsaturated
metal and the exposed fluoride ligand. We present here the
synthesis and characterization of iron(II) fluoride complexes and
(14) Catalytic C-F activation through cross-metathesis reactions: (a) Widdow-
son, D. A.; Wilhelm, R. Chem. Commun. 1999, 2211-2212. (b) Bo¨hm, V.
P. W.; Gsto¨ttmayr, C. W. K.; Weskamp, T.; Herrmann, W. A. Angew.
Chem., Int. Ed. 2001, 40, 3387-3389.
(5) Representative examples of olefinic C-F activation: (a) Braun, T.; Noveski,
D.; Neumann, B.; Stammler, H.-G. Angew. Chem., Int. Ed. 2002, 42, 2745-
2748. (b) Kraft, B. M.; Jones, W. D. J. Am. Chem. Soc. 2002, 124, 8681-
8689. (c) Kirkham, M. S.; Mahon, M. F.; Whittlesey, M. K. Chem.
Commun. 2001, 813-814. (d) Noveski, D.; Braun, T.; Schulte, M.;
Neumann, B.; Stammler, H.-G. Dalton Trans. 2003, 4075-4083. (e)
Peterson, T. H.; Golden, J. T.; Bergman, R. G. Organometallics 1999, 18,
2005-2020. (f) Siedle, A. R.; Newark, R. A. Organometallics 1989, 8,
1442-1450. (g) Watson, L. A.; Yandulov, D. V.; Caulton, K. G. J. Am.
Chem. Soc. 2001, 123, 603-611. (h) Watson, P. L.; Tulip, T. H.; Williams,
I. Organometallics 1990, 9, 1999-2009. (i) Ferrando-Miguel, G.; Ge´rard,
H.; Eisenstein, O.; Caulton, K. G. Inorg. Chem. 2002, 41, 6440-6449.
(6) Aliphatic C-F activation: (a) Kraft, B. M.; Lachicotte, R. J.; Jones, W.
D. J. Am. Chem. Soc. 2000, 122, 8559-8560. (b) Kraft, B. M.; Lachicotte,
R. J.; Jones, W. D. J. Am. Chem. Soc. 2001, 123, 10973-10979.
(7) (a) Jones, W. D. Dalton 2003, 3991-3995. (b) Braun, T.; Perutz, R. N.
Chem. Commun. 2002, 2749-2757. (c) Burdeniuc, J.; Jedlicka, B.; Crabtree,
R. H. Chem. Ber. 1997, 130, 145-154. (d) Kiplinger, J. L.; Richmond, T.
G.; Osterberg, C. E. Chem. ReV. 1994, 94, 373-431. (e) Mazurek, U.;
Scwarz, H. Chem. Commun. 2003, 1321-1326. (f) Richmond, T. G. Metal
Reagents for Activation and Functionalization of Carbon-Fluorine Bonds.
In ActiVation of UnreactiVe Bonds in Organic Synthesis; Murai, S., Ed.;
Springer: Berlin, 1999; pp 243-269.
(15) (a) Eckert, N. A.; Smith, J. M.; Lachicotte, R. J.; Holland, P. L. Inorg.
Chem. 2004, 43, 3306-3321. (b) Eckert, N. A.; Bones, E. M.; Lachicotte,
R. J.; Holland, P. L. Inorg. Chem. 2003, 42, 1720-1725. (c) Holland, P.
L.; Cundari, T. R.; Perez, L. L.; Eckert, N. A. E.; Lachicotte, R. J. J. Am.
Chem. Soc. 2002, 124, 14416-14424. (d) Andres, H.; Bominaar, E. M.;
Smith, J. M.; Eckert, N. A. E.; Holland, P. L.; Mu¨nck, E. J. Am. Chem.
Soc. 2002, 124, 3012-3025. (e) Smith, J. M.; Lachicotte, R. J.; Pittard, K.
A.; Cundari, T. R.; Lukat-Rodgers, G.; Rodgers, K. R.; Holland, P. L. J.
Am. Chem. Soc. 2001, 123, 9222-9223. (f) Smith, J. M.; Lachicotte, R.
J.; Holland, P. L. Chem Commun. 2001, 1342-1343.
(16) Vela, J.; Stoian, S.; Flaschenriem, C. J.; Mu¨nck, E.; Holland, P. L. J. Am.
Chem. Soc. 2004, 126, 4522-4523.
(17) Smith, J. M.; Lachicotte, R. J.; Holland, P. L. J. Am. Chem. Soc. 2003,
125, 15752-15753.
(18) (a) Vela, J.; Vaddadi, S.; Cundari, T. R.; Smith, J. M.; Gregory, E. A.;
Lachicotte, R. J.; Flaschenriem, C. J.; Holland, P. L. Organometallics 2004,
23, 5226-5239. (b) Vela, J.; Smith, J. M.; Lachicotte, R. J.; Holland, P.
L. Chem. Commun. 2002, 2886-2887. (c) Smith, J. M.; Lachicotte, R. J.;
Holland, P. L. Organometallics 2002, 21, 4808-4814.
(19) Krause, E. Ber. Dtsch. Chem. Ges. 1918, 1447.
9
7858 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Low-Coordinate Iron(II) Fluoride Complexes
A R T I C L E S
Figure 2. Molecular structures of 12 (a) and 2 (b). Ellipsoids are shown at 50% probability. Hydrogen atoms, and isopropyl groups in 12, are omitted for
clarity.
Scheme 2
Table 1. Selected Interatomic Distances (Å) and Angles (deg) for
Diketiminate Iron(II) Fluoride Complexes
[LMeFe(
µ-F)]2 (12)
LtBuFeF (2)
Fe-F
1.9757(12), 1.9774(14)
2.0081(18), 2.0161(17)
119.92(18), 117.32(18)
93.27(7)
3.0831(6)
102.44(10), 102.56(9)
1.8079(15)
1.9609(14)
127.95(13)
95.66(8)
Fe-N
When Me3SnF is reacted with an alkyl complex of the more
C-N-C
N-Fe-N
Fe‚‚‚Fe
Fe-F-Fe
sterically demanding ligand LtBu (Figure 1), the pink iron product
shows low solubility in pentane, toluene, and diethyl ether.
Therefore, in this case, a slight excess of the highly soluble
alkyliron(II) complex is used, and pentane is used to wash out
the remaining alkyl complex. This leads to a 74% yield of
LtBuFeF (2), which was also characterized by X-ray diffraction
(Figure 2, discussed below). The low solubility of 2 is
attributable to the high dipole moment along the Fe-F bond
caused by the electronegative fluorine, as well as to exposure
of the hard fluoride ligand to the molecule’s surface. As a result,
2 dissolves only in coordinating solvents such as THF and
acetonitrile (see below for details). In contrast, the complex
[LMeFeF]2 (12) is very soluble in hydrocarbons because the
molecule has a low dipole moment and the fluoride ligands are
buried within the hydrocarbon ligand framework.
Cp*(dppe)FeIIF (an 18-electron complex) was recently synthe-
sized from [Cp*(dppe)FeII]+PF6 and CsF or CoCp2.21 This
-
fluoride complex was too unstable for X-ray diffraction analysis,
although a DFT-optimized structure yielded a theoretical Fe-F
distance of 1.927 Å. A tris(pyrazolyl)borate iron(II) fluoride
has also been isolated.23
LtBuFeF (2, formally a 12-electron complex) is to our
knowledge the only three-coordinate fluoride complex of any
transition metal. In 2, iron lies in a crystallographically required
planar geometry. Experimental distances for transition metal-
fluorine bonds average 1.91 Å (standard deviation 0.09 Å), and
thus the Fe-F distance in LtBuFeF (2), 1.8079(15) Å, is
exceptionally short. Likely reasons for the short Fe-F bond
length include the low coordination number at iron and the
fluoride’s ability to act as a π-donor toward the unsaturated
metal.24 This small terminal Fe-F distance is consistent with
the short FeII-OR bond distances observed in other three-
coordinate iron(II) complexes (examples: LRFeOtBu 1.761(10)
Å (R ) Me), 1.786(3) Å (R ) tBu) and LMeFeOCHPh2 1.8076-
(16) Å).15a,18a,25
These synthetic routes seem to be general in scope because
different alkyliron(II) compounds (methyl, isobutyl, and cyclo-
hexyl)18 give the iron(II) fluorides in similar yields. Noteworthy
is the use of Me3SnF to prepare iron(II) fluorides since this tin
reagent has mainly been used for making early and main group
but not late-metal fluoride complexes.10a,20
Characterization of Iron(II) Fluoride Complexes. Both 12
and 2 are paramagnetic at room temperature. For 12 the solution
magnetic moment is 6.2(3) µB (per dimer), whereas for 2 the
magnetic moment is 5.6(3) µB. The paramagnetism is evident
in their 1H NMR spectra, which have highly shifted resonances
reminiscent of those from other iron(II) diketiminate complexes.
The [Fe(µ-F)]2 rhomb in 12 has Fe-F-Fe and F-Fe-F
angles of 102.44-102.56° and 77.50(7)°, respectively. The Fe-
(20) (a) Herzog, A.; Roesky, H. W.; Ja¨ger, F.; Steiner, A. Chem. Commun. 1996,
29-30. (b) Herzog, A.; Liu, F.; Roesky, H. W.; Demsar, A.; Keller, K.;
Noltemeyer, M.; Pauer, F. Organometallics 1994, 13, 1251-1256.
(21) Tilset, M.; Fjeldahl, I.; Hamon, J.; Hamon, P.; Toupet, L.; Saillard, J.;
Costuas, K.; Haynes, A. J. Am. Chem. Soc. 2001, 123, 9984-10000.
(22) Cambridge Structural Database, ConQuest version 1.7, November 2004
release. (a) Allen, F. H. Acta Crystallogr. 2002, B58, 380-388. (b) Bruno,
I. J.; Cole, J. C.; Edgington, P. R.; Kessler, M.; Macrae, C. F.; McCabe,
P.; Pearson, J.; Taylor, R. Acta Crystallogr. 2002, B58, 389-397.
(23) Gorrell, I. B.; Parkin, G. Inorg. Chem. 1990, 29, 2452-2456.
(24) (a) Becker, C.; Kieltsch, I.; Broggini, D.; Mezzetti, A. Inorg. Chem. 2003,
42, 8417-8429. (b) Mezzetti, A.; Becker, C. HelV. Chim. Acta 2002, 85,
2686-2703. (c) Caulton, K. G. New J. Chem. 1994, 18, 25-41.
1
The assignments for the H NMR resonances are based on
relative integrations, similar to the previously characterized
diketiminate-iron(II) complexes,15 and are given in detail in the
Experimental Section. Neither 12 or 2 shows 19F resonances
between 500 and -500 ppm (vs CFCl3, THF-d8), presumably
because of very rapid relaxation of the 19F nuclei.
No discrete iron(II) fluoride has previously been struc-
turally characterized.10,21-23 The molecular iron(II) fluoride
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7859

A R T I C L E S
Vela et al.
Figure 3. Molecular structures of 2‚tBupy (a), 2‚ACN (b), 1‚tBupy (c), and 1‚CF3py (d). Ellipsoids are shown at 50% probability, and hydrogen atoms are
omitted for clarity.
(µ-F) distances of about 1.97 Å are between the range of M-(µ-
F) distances reported for nickel (1.89-1.91 Å) and manganese
(1.95-2.05 Å) complexes with bridging fluoride ligands.10a The
difference in coordination number and nuclearity between 12
and 2 has a steric origin. The ligand LtBu is substantially more
hindering than LMe: we have previously quantified this steric
effect using the C-N-C angle at the imine nitrogen, which is
ridyl)(Cl) do not show any absorptions between 430 and 600
cm-1 4c
The IR spectrum of the dimeric fluoride complex 12
(KBr pellet) has no bands between 450 and 1000 cm-1 that
could be attributed to the Fe-F bond. This is consistent with
elongation and overall weakening of the Fe-F bond to bridging
fluoride ligands.
The visible spectrum (Figure 4) of the free three-coordinate
fluoride 2 has a single absorption maximum at 530 nm (730
M-1cm-1). This band is very similar to that seen in LtBuFeCl,
which has λmax ) 560 nm (520 M-1cm-1).15f We tentatively
assign this band to an iron-to-diketiminate (MLCT) transition,
reasoning that the more polarized M-F bond leads to a more
electrophilic metal and lower d-orbital energies.27
Monomeric Four-Coordinate Iron(II) Fluorides. Treatment
of the fluoride complex 12 or 2 with 1 equiv of pyridines or
excess acetonitrile is accompanied by a substantial shift in the
.
roughly 8° larger in LtBu than LMe 18a,15f
This trend holds in
.
chloride15f and alkyl18a complexes of iron(II) bearing the
diketiminates LMe and LtBu (Figure 1). Similarly, in the fluoride
complexes the C-N-C angle is 127.95° for 2 (LtBu) and
119.92-117.32° for 12 (LMe) (Table 1).
The infrared (IR) spectrum of 2 (KBr pellet, 450-4000 cm-1
)
shows a strong band at 596 cm-1. This band lies in the range
reported for terminal M-F bond stretching vibrations (νM-F
and is absent in the IR spectrum of
:
500-750 cm-1 10a,26
)
LtBuFeCl15f (νM-Cl range: 200-400 cm-1).26 Therefore, we
assign this band to the iron-fluorine stretching vibration in 2.
For comparison, Perutz and co-workers have reported that the
Ni-F stretching vibrations in trans-Ni(Et3P)2(C6F5)(F) and
(25) (a) Kornev, A. N.; Chesnokova, T. A.; Semenov, V. V.; Zhezlova, E. V.;
Zakharov, L. N.; Klapshina, L. G.; Domrachev, G. A.; Rusakov, V. S. J.
Organomet. Chem. 1997, 547, 113-119. (b) Gibson, V. C.; Marshall, E.
L.; Navarro-Llobet, D.; White, A. J. P.; Williams, D. J. J. Chem. Soc.,
Dalton Trans. 2002, 4321-4322. (c) Chen, H.; Power, P. P.; Shoner, S.
C. Inorg. Chem. 1991, 30, 2884-2888. (d) Shannon, R. D. Acta Crystallogr.
1976, A32, 751-767.
trans-Ni(Et3P)2(2-C5F4N)(F) appear at 535 and 530 cm-1
,
respectively, while the chloride complexes trans-Ni(Et3P)2-
(C6F5)(Cl) and trans-Ni(Et3P)2(5-chloro-2,4,6-trifluoro-3-py-
(26) Nakamoto, K. Infrared and Raman Spectra of Inorganic Compunds, part
B, 5th ed.; Wiley: New York, 1997; pp 180-190.
9
7860 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Low-Coordinate Iron(II) Fluoride Complexes
A R T I C L E S
Table 2. Structural Parameters of Four-Coordinate Iron(II)
Fluoride Complexes
Cmpd(Lapical
)
Fe
−F (Å)
Fe
−X (Å)
τa
LtBuFeF(4-tBu-py) (2‚tBupy)
LMeFeF(4-tBu-py) (1‚tBupy)
LMeFeF(4-CF3-py) (1‚CF3py)
LtBuFeF(NCCH3) (2‚ACN)
LtBuFeOEt2(η1-BF4) (3)
1.8700(14)
1.8393(23)
1.871(3)
1.946(2)
2.0672(10)
1.9757(12)
2.1190(19)
2.1137(30)
2.115(4)
2.099(4)
2.0404(11)
1.9774(14)
0.45
0.43
0.45
0.34
0.13
0.08
[LMeFe(µ-F)]2 (12)
a τ ) [∑(Lbasal-Fe-Lbasal) - ∑(Lbasal-Fe-Laxial)]/90. See ref 16.
recently in related sulfido,16 hydrido,17 and amido iron(II)15a
complexes. The extent of this pyramidal distortion can be
quantified using τ, which has ideal values of τ ) 1 and τ ) 0
for perfect trigonal pyramidal (where the metal sits in the plane
of the three basal ligands) and tetrahedral geometries, respec-
tively.16 Table 2 shows the Fe-F distances and τ values for a
series of four-coordinate fluoride complexes. We recently have
described the tendency for π-donor ligands to prefer the basal
position and explained it in terms of a crystal-field model in
which the only doubly occupied d orbital has the appropriate
symmetry to undergo π-interactions with the axial ligand, but
not the basal ligand.15a Accordingly, the π-donor fluoride is
expected to occupy the basal position, and the π-acceptor
pyridines the axial position. This idea is further supported here
by the trend in axial distortions: acetonitrile induces a lower
distortion from tetrahedral geometry (τ ) 0.34)16 than the
stronger π-acceptor pyridines (τ ) 0.43-0.45).
Figure 4. Electronic absorption spectra of three- and four-coordinate
iron(II) fluoride complexes of LtBu
.
Scheme 3
The four-coordinate adducts of 2 have a strong band in the
IR spectrum attributable to νFe-F at 544 cm-1 in 2‚tBupy (Fe-F
1.8700(14) Å) and 538 cm-1 in 2‚ACN (Fe-F 1.946(2) Å),
lower than that observed for the free three-coordinate fluoride
complex (2: 596 cm-1, Fe-F 1.8079(15) Å). Interestingly, the
fluoride complexes of LMe have a somewhat higher Fe-F
stretching frequency: the monomeric adducts of 1 have νFe-F
at 717 cm-1 in 1‚tBupy (Fe-F 1.8393(23) Å), 704 cm-1 in
1‚py, 690 cm-1 in 1‚Phpy, and 669 cm-1 in 1‚CF3py (Fe-F
1.871(3) Å). These data show that the Fe-F bond elongates
and weakens for complexes of a given diketiminate ligand as
the σ-donating ability of the pyridine ligand decreases.
In contrast to the three-coordinate fluoride complex 2 (λmax
) 530 nm), the visible spectra of the four-coordinate complexes
in toluene (including the dimeric fluoride 12) show absorption
bands near 460 nm (930-2110 M-1 cm-1) (Figure 4). Addition
of 1 equiv of pentafluoropyridine causes no change in the visible
spectrum of 2, indicating that it does not coordinate. Conversely,
tetrahydrofuran and acetonitrile solutions of 2 show the 460
nm maximum indicative of adduct formation (Figure 4).
Therefore, we conclude that solubilization of 2 in THF is
accompanied by formation of the four-coordinate fluoride
LtBuFeF(THF) (similar to the structurally characterized 2‚ACN).
A series of pyridine adducts with the fluoride complex of
LMe (1‚L′) show two absorption bands in the visible spectrum:
one very similar to the four-coordinate complexes derived from
2 (λmax ) 400-420 nm, 1630-2180 M-1cm-1) and another
one that shifts to higher energy when the electron-donating
ability of the pyridine (L′) increases (λmax ) 512 nm (670 M-1
cm-1) for CF3py; λmax ) 465 nm (1260 M-1 cm-1) for tBupy).
This latter band may thus be assigned as a MLCT transition
between the iron and the pyridine ligand.
1H NMR resonances as well as a color change (see below),
indicative of adduct formation (Scheme 3a). Slow concentration
of these solutions leads to crystallization of the four-coordinate
complexes LtBuFeF(4-tBu-py) (2‚tBupy), LtBuFeF(NCCH3) (2‚
ACN), LMeFeF(4-tBu-py) (1‚tBupy), and LMeFeF(4-CF3-py) (1‚
CF3py). The molecular structures of these adducts are shown
in Figure 3. These complexes are also paramagnetic, having
solution magnetic moments of 5.0(4) µB, consistent with a high-
spin FeII center.
In the new four-coordinate fluoride complexes, the coordina-
tion sphere of iron substantially deviates from a tetrahedral
geometry toward a trigonal pyramidal structure, as observed
All four coordinate fluoride complexes derived from 12 and
2 also possess an electronic absorption between 900 and 980
(27) An alternative assignment is a halide-to-metal LMCT.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7861

A R T I C L E S
Vela et al.
Figure 5. Molecular structures of 3 (a), 62 (b), 4 (c), and 5 (d). Ellipsoids are shown at 50% probability. Carbon-bound hydrogen atoms, and isopropyl
groups in 62, are omitted for clarity.
nm. A similar band is observed for the parent fluoride complexes
12 and 2 at 791 and 715 nm, respectively. Because this
absorption occurs at a lower energy and has a lower intensity
(90-210 M-1 cm-1) than the absorptions mentioned above, we
assign it as a d-d transition.
compounds to generate new complexes, in reactions that are
driven by the formation of very strong Si-F bonds. Thus, the
reaction of equimolar amounts of [LMeFeF]2 and hexamethyl-
disilathiane (HMDS) proceeds smoothly in a couple of hours
at 60 °C in toluene to give the diiron(II) sulfide LMeFe(µ-S)-
FeLMe (Scheme 3c), which we have synthesized independently.16
Under similar reaction conditions, the fluoride complex 2 reacts
partially with 0.5 equiv of HMDS to give a 1:1 mixture (as
observed by 1H NMR in C6D6) of a product tentatively assigned
as the analogous diiron(II) sulfide LtBuFe(µ-S)FeLtBu, along with
the new compound LtBuFeSSiMe3 (4) (Scheme 3c). The tri-
methylsilylthiolate complex 4 is isolated in 86% yield, and its
X-ray crystal structure is shown in Figure 5c (see also Table
3).
The reaction of LtBuFeF with 2 equiv of bis(trimethylsilyl)-
acetylene (Me3SiCCSiMe3) provides complex LtBuFeCCSiMe3
(5) in 61% isolated yield (Scheme 3d). The molecular structure
of 5 is shown in Figure 5d (see also Table 3). New routes to
paramagnetic complexes of transition metals containing sp-
hybridized hydrocarbyl ligands are of interest due to the potential
ability of such materials to act as molecular wires.29
Disruption of the Fe-F Bond by Boron and Silicon
Reagents. Treatment of a slurry of the three-coordinate fluoride
complex 2 in diethyl ether with 1 molar equiv of Et2O‚BF3
results in dissolution of the material (Scheme 3b). Slow
evaporation of solvent gives crystalline LtBuFe(OEt2)(η1-BF4)
(3), for which the X-ray crystal structure is shown in Figure
-
5a. The structure of 3 shows that the BF4 anion lies in the
pseudoaxial position and the molecule has a relatively low
pyramidal distortion with τ ) 0.13 (Table 2). The Fe-F distance
in 3 is 2.0672(10) Å, whereas the B-Fbound distance is 1.488-
(2) Å, roughly 10% longer than the mean B-Ffree distance of
-
1.349(5) Å (Table 2). Therefore the BF4 anion is strongly
activated based on the definition of Roesky et al.10a Still, the
BF4 anion dissociates readily in solution: the 19F NMR
-
spectrum of 3 in THF shows a single resonance at 155.1 ppm
-
(vs CFCl3), virtually the same as that of Na+BF4 (see
Experimental Section). The complete abstraction of the fluoride
ligand in 2 suggests that although the FeII-F bond is Very short,
it can be actiVated by Lewis acids.
Triethylsilane (Et3SiH) also reacts with the iron(II) fluoride
complexes, providing a convenient synthetic route to the hydride
complexes [LMeFeH]2 (62) and [LtBuFeH]2 (72) (Scheme 3e). The
Transition metal fluorides have been used as precursors to
other species by exploiting the silaphilicity of the fluoride
ligand.10,28 In agreement with this idea, 12 and 2 react with silyl
(28) (a) Hoffman, N. W.; Prokopuk, N.; Robbins, M. J.; Jones, C. M.; Doherty,
N. M. Inorg. Chem. 1991, 30, 4177-4181. (b) Doherty, N. M.; Critchlow,
S. C. J. Am. Chem. Soc. 1987, 109, 7906-7908.
9
7862 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Low-Coordinate Iron(II) Fluoride Complexes
A R T I C L E S
Table 3. Structural Characterization of Complexes Prepared from Iron(II) Fluorides and Silyl Compounds: Bond Distances (Å) and Angles
(deg)
LtBuFeSSiMe3 (4)
LtBuFeCCSiMe3 (5)
[LMeFe(
µ-H)]2 (62)
Fe-S
2.2487(12), 2.2460(12)
1.964-1.981
2.1100(16), 2.0991(16)
118.29(6), 117.90(6)
95.30(13), 94.32(12)
126.1-126.7
Fe-C
1.961(6)
1.953(4), 1.964(4)
1.226(6)
141.07(18), 124.11(18)
94.81(16)
Fe-H
1.40(9), 1.45(8), 1.55(8), 1.56(9)a
1.971(4), 1.975(4), 1.978(4), 1.988(4)
2.4638(11)
Fe-N
Fe-N
Fe-N
S-Si
C-C
Fe‚‚‚Fe
Fe-S-Si
N-Fe-N
C-N-C
N-Fe-C
N-Fe-N
C-N-C
Fe-H-Fe
N-Fe-N
C-N-C
111(4), 113(4)a
95.0(2), 95.0(2)
117.8(4), 117.9(4), 119.3(4), 119.5(4)
127.9(4), 128.1(4)
a Iron-bound hydrogen atoms (62) were located in the electron density map and refined with isotropic thermal parameters.
1
as monitored by 19F and H NMR. Nevertheless, when these
Scheme 4
reactions are repeated in the presence of triethylsilane, C-F
actiVation takes place, giVing in each case the product of
monohydrodefluorination (HDF): pentafluorobenzene (C6F5H)
or heptafluorotoluene (p-H-C6F4CF3), respectively. The forma-
tion of fluorotriethylsilane (FSiEt3) is also observed (19F NMR,
GC-MS). Thus, while direct reaction of fluorinated aromatics
with the iron(II) hydride complexes does not occur, C-F
actiVation becomes feasible in the presence of silane as a
fluoride trap.
latter complex has been synthesized independently, and its
properties and some of its reactivity have been reported.17
Compounds 62 (µeff ) 4.0(3) µB) and 72 (µeff ) 3.8(3) µB)17 are
among the very few characterized paramagnetic hydride com-
plexes in the literature.17,30, The new compound 62 is isolated
in 89% yield, and its X-ray crystal structure is shown in Figure
5b (see also Table 3). Its properties and reactivity will be the
subject of future publications.
Although Scheme 4 shows a simple conceptual explanation
for HDF catalysis, it is incomplete because the hydride complex
does not directly react with fluorinated aromatic compounds.
The way in which the silane assists the hydride complex is
uncertain. However, the intermediacy of the hydride complex
cannot be ruled out since both fluoride and hydride complexes
are able to carry out the HDF reaction in the presence of silane.
Catalytic Perfluoroarene Hydrodefluorination. The hy-
drodefluorination (HDF) of aromatic perfluorocarbons is achieved
when a stoichiometric mixture of triethylsilane and an aromatic
perfluorocarbon is heated in the presence of 0.2 molar equiv of
12, 2, 62, or 72 (Table 4).32 Conversions are higher when starting
with fluorides rather than hydrides (see below) and higher with
the monomeric fluoride 2 than with the dimeric 12, as shown
for hexafluorobenzene (Table 4, entries 1 and 2). The highest
activities among the aromatic substrates are observed for
octafluorotoluene (TON ) 4.5) and perfluoropyridine (TON )
3.6), and HDF activity decreases as a function of the degree of
fluorination on the substrate. The extent of HDF activity
correlates with the electron affinity of the perfluoroarene (Figure
6),33 a feature that has been identified previously in some C-F
activation schemes.7c,33 Other aromatic substrates with low
fluorine substitution such as para-fluorotoluene, R,R,R-trifluo-
rotoluene, or 1,3-bis(trifluoromethyl)benzene do not undergo
C-F activation under the reaction conditions described above.
However, only substrates containing sp2 carbons undergo C-F
activation (fluorinated alkenes are discussed below). No benzylic
C-F activation is observed in octafluorotoluene. Similarly,
perfluorinated methylcyclohexane (containing only aliphatic
C-F bonds) does not undergo hydrodefluorination despite
having one of the highest electron affinities known for a
perfluorocarbon (1.06 eV).33b Therefore, a high electron affinity
does not suffice for HDF to occur in the presence of the iron
catalyst, and only unsaturated substrates are activated.
Similarly to the fluoride complexes 12 and 2, the four-
coordinate fluoride complexes such as 1‚tBupy or 2‚tBupy (see
above) react with triethylsilane to give the corresponding four-
coordinate hydride complexes (complete conversion by 1H
NMR), which are similar to the recently reported LtBuFeH(4-
tBu-py).17 The driving force for all the reactions between the
fluoride complexes and silyl reagents is clearly the formation
of a very strong Si-F bond (159 kcal/mol),2,31 which renders
the reactions enthalpically favorable. The FeII-F bond is thus
silaphilic, making the fluoride complexes valuable synthetic
precursors for new molecules with a variety of ligands.
Envisioning a Catalytic Cycle for Hydrodefluorination.
We have previously observed that low-coordinate iron hydride
complexes are capable of breaking strong bonds to electroneg-
ative elements,17 suggesting that they might be active toward
C-F activation reactions. Because the iron(II) fluoride com-
plexes react with Et3SiH to give iron(II) hydrides, this could
lead to catalytic C-F bond activation (Scheme 4). However,
there is no evidence of reaction when the hydride complexes
62 and 72 are heated at 120 °C for one week in C6D6 or THF-d8
with hexafluorobenzene (C6F6) or octafluorotoluene (C6F5CF3),
(29) (a) Rosenthal, U. Angew. Chem., Int. Ed. 2003, 42, 1794-1798. (b) Roue´,
S.; Lapinte, C.; Bataille, T. Organometallics 2004, 23, 2558-2567. (c)
Berry, J. F.; Cotton, F. A.; Murillo, C. A. Organometallics 2004, 23, 2503-
2506. (d) Mironov, V. S.; Chibotaru, L. F.; Ceulemans, A. J. Am. Chem.
Soc. 2003, 125, 9750-9760. (e) Stang, S. L.; Paul, F.; Lapinte, C.
Organometallics 2000, 19, 1035-1043. (f) Gu, X. Organometallics 1998,
17, 5920-5923.
(30) (a) Poli, R. Chem. ReV. 1996, 96, 2135-2204. (b) Bruckhardt, U.; Casty,
G. L.; Tilley, T. D.; Woo, T. K.; Rothlisberger, U. Organometallics 2000,
19, 3830-3841. (c) Kupfer, V.; Thewalt, U.; Horacek, M.; Petrusova, L.;
Mach, K. Inorg. Chem. Commun. 1999, 2, 540-544. (d) Jewson, J. D.;
Liable-Sands, L. M.; Yap, G. P. A.; Rheingold, A. L.; Theopold, K. H.
Organometallics 1999, 18, 300-305. (e) Hessen, B.; Van Bolhuis, F.;
Teuben, J. H.; Petersen, J. L. J. Am. Chem. Soc. 1988, 110, 295-296. (f)
Bianchini, C.; Mealli, C.; Meli, A.; Sabat, M. J. Chem. Soc., Chem.
Commun. 1986, 777-779. (g) Raynor, J. B. Sattelberger, A. P.; Luetkens,
M. L. Inorg. Chim. Acta 1986, 113, 51-54. (h) Luetkens, M. L. Elcesser,
W. L. Huffman, J. C. Inorg. Chim. Acta 1984, 23, 1718-1726.
(31) Brook, M. A. Silicon in Organic, Organometallic and Polymer Chemistry;
Wiley: New York, 2000; pp 28-38.
(32) In the absence of iron fluoride or hydride there was no sign of reaction ()
1%) between any of the substrates and trisubstituted silanes.
(33) (a) Dillow, G. W.; Kebarle, P. J. Am. Chem. Soc. 1989, 111, 5592-5596.
(b) Kebarle, P.; Chowdhury, S. Chem. ReV. 1987, 87, 513-534. (c)
Wentworth, W. E.; Limero, T.; Chen, E. C. M. J. Phys. Chem. 1987, 91,
241-245. (d) Chowdhury, S.; Grimsrud, E. P.; Heinis, T.; Kebarle, P. J.
Am. Chem. Soc. 1986, 108, 3630-3635.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7863

A R T I C L E S
Vela et al.
Table 4. Catalytic Hydrodefluorination (HDF) of Fluorocarbons with Iron(II) Diketiminate Complexes
precatalyst
conc, M
substrate
conc, M
reagent
conc, M
selectivitya
(%)
solvent
conditions
product distribution
TONb
1
2
3
12, 0.021
2, 0.021
none
C6F6, 0.11
THF-d8
Et3SiH, 0.11
Et3SiH, 0.11
KHBEt3, 0.11
45 °C, 4 days
C6HF5 (20%)
p-C6H2F4 (0.2%)
C6HF5 (50%)
p-C6H2F4 (0.8%)
C6HF5 (79%)
99
98
84
1.0
C6F6, 0.11
C6F6, 0.11
THF-d8
THF-d8
45 °C, 4 days
2.5
RT, 5 min
p-C6H2F4 (14%)
C6HF5 (24%)
C6HF5 (6%)
4
5
6
7
8
9
2, 0.021
2, 0.021
2, 0.021
2, 0.021
2, 0.021
12, 0.010
C6F6, 0.11
C6F6, 0.11
C6F5H, 0.11
C6F5CF3, 0.11
C5F5N, 0.11
CF2dCFCF3, 0.11c
THF-d8
C6D6
THF-d8
THF-d8
THF-d8
THF-d8
Et3SiH, 0.11
Et3SiH, 0.11
Et3SiH, 0.11
Et3SiH, 0.11
Et3SiH, 0.11
Et3SiH, 0.11
80 °C, 4 days
45 °C, 4 days
45 °C, 4 days
45 °C, 12 h
45 °C, 4 days
100 °C, 3 h.
100
100
100
100
100
67
1.2
0.3
0.2
4.5
3.6
9.8
p-C6H2F4 (4%)
p-C6HF4CF3 (90%)
p-C5HF4N (71%)
CHFdCFCF3 (E: 60%, Z: 27%),
CF2dCHCF3 (2%)
CF2dCHCH3 (11%)
10
12, 0.010
CH2dCHCF3, 0.11c
THF-d8
Et3SiH, 0.11
100 °C, 4 days
100
1.2
a Selectivity for the main hydrodefluorination product. b TON ) moles of products (overall)/moles of catalyst. Precision limits are (4% of the given
value. c Total concentration (see Experimental Section for details).
(entries 2 and 4). When phenylsilane (PhSiH3) is used, complete
decomposition ensues at room temperature and no C-F
activation is observed. Changing the electronic environment at
the silicon center affects both the catalyst stability and the HDF
rate. The following order of reactivity is evident for different
trisubstituted silanes, expressed in terms of conversion of
hexafluorobenzene (0.11 M) in THF-d8 at 60 °C with 20 mol
% catalyst loading: (EtO)3SiH (41%, 5 min), Ph3SiH (22%, 12
h), 3,5-(CF3)2C6H3-SiHMe2 (92%, 4 days), Et3SiH (complete
in 4 days). Thus, while triethoxysilane reacts at room temper-
ature, it also leads to the most rapid catalyst deactivation
(accompanied by precipitate formation).32 H2 gas is not an
effective reductant at pressures of 250-1500 psi, even with
Figure 6. Plot of hydrodefluorination activity vs perfluoroarene electron
affinity.
added traps for HF such as Et3N, py, and NaF.
Because turnovers are higher than 1 and because the catalyst
is introduced in the form of a fluoride complex, these reactions
are truly catalytic. The activity observed is dependent on the
solvent, with THF (ꢀ ) 7.6)34 giving higher conversions than
the less polar solvent toluene (ꢀ ) 2.3).34 The reactions also
proceeded in dry pyridine as solvent (ꢀ ) 12.4),34 but the
conversions achieved were intermediate between those in THF
and benzene.35
The perfluoroarene HDF reactions catalyzed by 2 are regi-
oselective, with C-F activation at the position para to the most
electron-withdrawing group always observed. Interestingly,
attempts to carry out HDF of p-HC6F4CF3 (heptafluorotoluene)
fail even after prolonged heating. The HDF reactions are highly
chemoselective as well, since mainly mono-hydrodefluorination
products and only traces of double HDF products are observed.
For example, the iron-catalyzed HDF of C6F6 gives C6HF5 with
98% selectivity, while direct reaction of C6F6 with the strong
reducing agent KHBEt336 gives a lower selectivity of the mono-
HDF product, 84% (entries 2 and 3 of Table 4).
The hydride complex 72 also catalyzes HDF, although the
conversion was lower. Catalyst decay was faster in this case,
offering a likely explanation for the difference in catalytic
activity. Control experiments show that heating a solution of
hydride complex 62 or 72 in THF-d8 in the presence of the silane
leads to decomposition, and multiple unidentified paramagneti-
1
cally shifted peaks are evident in the H NMR spectra.
Kinetics of Perfluoroarene Hydrodefluorination. Because
of catalyst decomposition, we are not able to follow the HDF
reaction kinetics to completion and turned to kinetic inquiry
using the initial rate method.37 In a typical kinetics experiment,
the initial concentration of substrate (octafluorotoluene), silane,
or 2 is varied while the other two are kept constant relative to
a reference experiment ([Et3SiH]0 ) [C6F5CF3]0 ) 0.11 M, [2]0
) 0.02 M in THF-d8 at 341 K). The disappearance of substrate
and formation of a hydrodefluorination product are monitored
by 19F NMR. Figure 7 and Table 5 show the initial rates from
these experiments. The initial rate has a first-order dependence
on silane concentration and on iron concentration, but it is
independent of the concentration of octafluorotoluene.
Choice of Silane and Catalyst Deactivation. Some differ-
ences are evident when the temperature of reaction and the
fluoride acceptor are varied. When triethylsilane is used, the
best conversion is observed at a temperature of 45 °C, with
higher temperatures resulting in faster catalyst decomposition
d[p-C6HF4CF3]/dt ) k[Et3SiH]1[LtBuFeF]1[C6F5CF3]0 (1)
The second-order rate constant derived from this rate equation
is k ) 1.4(1) × 10-3 M-1 s-1. There is a small normal kinetic
isotope effect with deuteration of the silane, kSiH/kSiD ) 1.4(1)
(entry 5 in Table 5 and Figure 7). Finally, the rate of HDF is
(34) Dielectric constant: Kerr, J. A. In CRC Handbook of Chemistry and Physics,
71st ed.; Lide, R. L., Ed.; CRC Press: Boca Raton, FL, 1990; p 8-44.
(35) The HDF reaction does not take place in more polar solvents such as DMF
or HMPA. This may be due to reaction between the iron(II) hydride
complex and these solvents.
(36) Fryzuk, M. D.; Lloyd, B. R.; Clentsmith, G. K. B.; Rettig, S. J. J. Am.
Chem. Soc. 1994, 116, 3804-3812.
(37) Espenson, J. H. Chemical Kinetics and Reaction Mechanisms, 2nd ed.;
McGraw-Hill: New York, 1995.
9
7864 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Low-Coordinate Iron(II) Fluoride Complexes
A R T I C L E S
Viewed as only a working model. Unfortunately, the apparent
competition between catalyst decomposition and catalytic HDF
hinders our ability to query the mechanism in more detail with
this system. While a more complete explanation will await a
more robust catalyst, it is possible to narrow down the
mechanistic possibilities based on the currently available data.
All of the perfluoroarene HDF reactions occur without an
induction period, suggesting that autocatalysis by a reaction
byproduct such as fluoride anion4b is unlikely. Repeated
distillation of the starting materials, different batches of iron
catalyst, addition of a drop of mercury, or different reaction
vessel materials (glass or Teflon) had little effect on the reaction
rate, also arguing against catalysis by a trace impurity. Neither
Figure 7. Time course of octafluorotoluene hydrodefluorination, showing
changes in the initial observed rates vs a reference experiment with [silane]0
) [C6F5CF3]0 ) 0.11 M, [Fe]0 ) 0.02 M, using THF-d8 at 341 K.
LtBuFeCl,15f LtBuFeCH2 Bu,18c nor LtBuFeMe15c catalyzed the
t
HDF reaction.
Hydrodefluorination of octafluorotoluene under standard
conditions (60 °C, 19 h) with and without dihydroanthracene
(DHA), a common radical trap, gives nearly the same extent of
conversion to p-C6HF4CF3 (46% and 54%, respectively).39 In
an analogous experiment with 20 mM 2, 110 mM Et3SiH, and
47 mM DHA, the derived rate constant is k ) 1.2(1) × 10-3
M-1 s-1, statistically the same as that in the absence of the
radical trap (Table 5).40 Other workers have ruled out a radical
path for C-F activation reactions based on similar observa-
tions.4a,b,1
Table 5. Hydrodefluorination of Octafluorotoluene with
Triethylsilane in THF-d8, Initial Ratesa
initial concentrations (M)
initial rate
1
entry
[Et3SiH]0
[C6F5CF3]0
[LtBuFeF]0
(10-6 M s-
)
1
2
3
4
5
0.11
0.11
0.22
0.11
Et3SiD: 0.11
0.11
0.11
0.22
0.11
0.11
0.11
0.11
0.021
0.021
0.021
0.0072
0.021
0.021
3.38(4)
3.30(4)
5.53(4)
1.2(1)
2.43(5)
2.9(1)
6, added
47 mM DHA
A pathway involving oxidative addition to 72 or another FeII
complex would form an iron(IV) complex. Such a high-valent
intermediate is unprecedented in diketiminate-iron complexes.15-18
In addition, a five-coordinate intermediate would suffer from
severe steric congestion as a consequence of the hindering
diketiminate ligand. An oxidative addition pathway is also
inconsistent with the fact that H2 gas (possibly the most active
substrate toward oxidative addition)42 does not function as a
competent reductant, nor does H2 add to other iron(II) diketimi-
nate complexes.15-18,43
a Data for first 100 min (e39% conversion).
only slightly affected by addition of an excess amount of
dihydroanthracene (DHA, entry 6 in Table 5).
During the course of the experiments, the only iron species
1
detected by H NMR is 2, indicating the fluoride complex is
the resting state of the catalyst. It is possible that the active
species is a form of the hydride complex, which quickly reacts
with the relative excess of perfluorocarbon and silane under
the reaction conditions.
The dependence of catalytic conversion on the substrate
electron affinity suggests the possibility of a mechanism
involving rate-limiting electron transfer from iron(II) to fluo-
roarene.44,45 Therefore, the redox properties of the putative
intermediate iron-hydride complexes are relevant. The hydride
complexes 72 and 62 are irreversibly oxidized at a potential of
about -0.6 V vs ferrocene. Because perfluoroarenes have
reduction potentials of -2.5 to -3.0 V vs ferrocene,4b,46 electron
transfer from iron to perfluoroarene is thus uphill by at least
1.8 V. Using this energy as the minimum value for ∆GqET, a
HDF of Aromatic Perfluorocarbons: Mechanistic Con-
siderations. The apparent rate law, the H/D kinetic isotope
effect, and the observation of 2 as the resting state each suggests
that the rate-limiting step in the HDF of octafluorotoluene is
reaction of the iron fluoride complex with silane. The reaction
of LtBuFeF with silane is expected to form 72, a hydride complex
that is known to be very reactive (it even cleaves the NdN
bond of azobenzene).17 It seems reasonable that this low-
coordinate iron hydride could also reduce C-F bonds. This is
supported by the fact that the hydride complex 72 can be used
instead of 2 as a catalyst, although conversions are lower due
to more rapid catalyst decomposition.
However, some observations are inconsistent with the simple
pathway shown in Scheme 4. First, neither 72 nor 62 stoichio-
metrically hydrodefluorinates fluoroaromatics in the absence of
silane (see above). This is not due to an unfavorable equilibrium
constant, because aromatic compounds do not react with the
iron(II) fluoride complexes. This suggests that the hydride alone
is not directly responsible for C-F activation and the actual
catalytic species is more complicated, perhaps a silane adduct
of the hydride.38 Additionally, unless the mechanism changes
or step (i) is not rate-limiting for all substrates, then there should
be identical rates with different fluorinated substrates, contrary
to our observations. Therefore, the simple model in Scheme 4
is not a satisfactory mechanistic explanation and should be
(38) Another possibility is that the active species is a silyl complex. Unfortu-
nately, our several attempts to isolate a diketiminate iron(II) silyl complex
independently have so far failed. The intermediacy of silyl cations (cf. ref
13) or silyl radicals is inconsistent with the regioselectivity of C-F
activation and with the first-order dependence of the initial rate on iron
concentration.
(39) A similar result is obtained when triphenylmethane (Ph3CH, 0.47 molar
equiv) is used as the radical trap (48% conversion under the same
conditions).
(40) Separate experiments show that the hydride complex 72 slowly yields an
undefined mixture of products (20-30% yield after 24 h at 60 °C) in the
presence of DHA or Ph3CH. Therefore, the slight decreases in HDF yield
and rate caused by these radical scavengers may be attributed to a secondary
reaction with the iron hydride complex. For example, hydride 6 can
deprotonate acetonitrile, forming a dimeric alkyl complex [LMeFe(η1-CH2-
CN)]2.18a
(41) Whittlesey, M. K.; Perutz, R. N.; Moore, M. H. Chem. Commun. 1996,
787-788.
(42) (a) Heinekey, D. M.; Lledo´s, A.; Lluch, J. M. Chem. Soc. ReV. 2004, 33,
175-182. (b) Kubas, G. J. Metal Dihydrogen and σ-Bond Complexes:
Structure, Theory and ReactiVity; Kluwer Academic/Plenum Publishers:
New York, 2001.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7865

A R T I C L E S
Vela et al.
Scheme 5
simple analysis using the Eyring equation suggests that the rate
for electron transfer from 72 to substrate would be no faster
than 10-13 M-1 s-1 47
. Therefore, outer-sphere electron transfer
does not appear to be kinetically competent for perfluoroarene
HDF. A similar conclusion was reached by Edelbach and Jones
for the stoichiometric C-F activation of perfluoroarenes by
Cp*Rh(PMe3)H2.4b
Our experiments do not distinguish between other potential
mechanisms. One such mechanism is nucleophilic aromatic
substitution by a hydride ligand. This is consistent with the
regioselectivity of HDF, which occurs para to electron-
withdrawing groups. Alternatively, the reaction could proceed
through an asynchronous σ-bond metathesis reaction involving
a four-center transition state. To account for the effect of
substrate electron affinity on the HDF, the C-F activation may
involve the development of partial charges within the reaction
pathway. We stress again that conclusive determination of the
mechanism for HDF of fluoroaromatics will only be possible
in a future system where the catalyst lifetime is improved.
Catalytic Hydrodefluorination of Fluoroolefins. The iron-
(II) fluoride 12 is also a precatalyst for the hydrodefluorination
of fluorinated olefins (9.8 turnovers for hexafluoropropene and
1.2 turnovers for 3,3,3-trifluoropropene, Table 4). At 100 °C
in THF, hexafluoropropene (0.11 M) quickly reacts with
triethylsilane (0.11 M) in the presence of 12 (0.1 molar equiv),
giving a mixture of C1 (87%, E/Z 2:1) and C2 (2%) hydrode-
fluorination products (Table 4, entry 9, Scheme 5). To our
surprise, even 3,3,3-trifluoropropene undergoes HDF to give
1,1-difluoropropene as the main product (11% conversion in 4
days under similiar reaction conditions, entry 10).
lectivity observed for the activation of aromatic substrates. (5)
The hydrodefluorination of 3,3,3-trifluoropropene occurs through
formal allylic C-F activation and despite this substrate having
a lower degree of fluorination.
Thus, in comparison to fluoroaromatics, the HDF of fluo-
roolefins with the iron fluoride is more efficient with the smaller
diketiminate, it is less regioselective, and it defluorinates more
completely. All of these data are most consistent with a hydride
insertion/â-fluoride elimination mechanism for the HDF of
perfluoroolefins such as shown in Scheme 5.48 We have
extensively studied â-hydrogen elimination in diketiminate iron-
(II) alkyl complexes.18 This work showed that hydride insertion
into olefins rapidly generates alkyl complexes (step (i) in
Scheme 5) and that â-hydrogen elimination occurs reversibly
(step (ii)).18 We have also shown that because of the higher
steric hindrance of the larger diketiminate LtBu, the alkyl com-
plexes derived with this ligand undergo reversible insertion and
â-hydrogen elimination much more slowly than complexes with
the smaller ligand LMe. The reproduction of these trends in the
HDF of fluorinated alkenes indicates that an analogous mech-
anism is operative for hydrodefluorination of these substrates.
The intermediacy of fluorohydrocarbyl complexes is sup-
ported by experiments in which the putative intermediate
LMeFeCH2CH2CF318a is treated with triethylsilane. Interestingly,
small amounts (up to 5% based on the alkyl complex) of
CF2dCHCH3 are produced: this is the same product seen in
the HDF of 3,3,3-trifluoropropene (Table 4, entry 10).49
Although we previously have not observed â-F elimination in
iron(II) diketiminates, the activation of C-F bonds in metal-
bound fluoroalkyl ligands has been well documented.3,50
There are many interesting differences between the HDF of
fluoroaromatics and fluoroolefins by the iron fluorides: (1) The
resting state observed by NMR during HDF of olefinic substrates
1
has H and 19F resonances analogous to diketiminate iron(II)
alkyl complexes (we have previously reported LMeFeCH2CH2-
CF3).18a In contrast, the fluoride complex (2 or 12) is the
observed resting state during HDF of perfluoroarenes. (2) The
temperature required for olefinic C-F activation (100 °C) is
considerably higher than the optimal temperature for the HDF
of perfluoroarenes (45 °C, see above). (3) Only the LMe-based
fluoride 12 is active for HDF of perfluoroolefins, and 2 is
inactive. This contrasts with aromatic HDF, in which 12 is less
active than 2. (4) The lack of selectivity for C1-F and C2-F
activation in perfluoropropene contrasts with the high regiose-
(43) Perfluorocarbon C-F bond oxidative addition is commonly preceded by
well-defined and isolable fluorocarbon metal complexes. For example,
Perutz and co-workers have characterized η2-complexes of perfluoronaph-
thalene and perfluoropyridine (see ref 7b). In these nickel systems, C-F
activation of perfluoropyridine occurred at C2 rather than at C4 as observed
in the iron-catalyzed process demonstrated here. In our system, electronic
absorption spectra show no substantial coordination of perfluorocarbons
to the iron(II) diketiminate complex (see text and Figure 4).
(44) The selectivity for para C-F activation in C6HF5 has been explained by
the relative radical character at this position in the radical anion.41
(45) (a) Yim, M. B.; Wood, D. E. J. Am. Chem. Soc. 1976, 98, 2053. (b)
Shchegoleva, I. I.; Bilkis, I. I.; Schastnev, P. V. Chem. Phys. 1983, 82,
343.
(46) Marsella, J. A.; Glicinski, A. G.; Coughlin, A. M.; Pez, G. P. J. Org. Chem.
1992, 57, 2856-2860.
(48) A similar hydride insertion/â-fluoride elimination mechanism is unlikely
to operate in the HDF of perfluoroarenes. Such a mechanism would involve
breaking the aromaticity of the substrate and therefore is expected to
increase the activation barrier relative to olefinic C-F activation. This is
inconsistent with HDF of perfluoroarenes occurring at lower temperatures
than for olefinic substrates.
(49) Diketiminate iron(II) complexes without fluorinated hydrocarbyl ligands
(e.g., LMeFeiBu)18 do not react with triethylsilane at 120 °C for 2 weeks.
In a separate experiment, [LMeFeF]2 does not react with 1-hexene: we
conclude that hydride formation must occur before catalysis proceeds.
(47) There is a large disparity between the ability of perfluoroarenes to be
reduced in the gas and solution phases. For example, the gas phase electron
affinities of hexafluorobenzene and octafluorotoluene are favorable with
values of 0.52(1) and 0.94(1) eV, respectively, but both have highly negative
irreversible reduction potentials in THF solution (-3.0 and -2.95 V vs
ferrocene; see refs 48 and 4b). In our mechanistic analysis, we use the
solution values and assume that the overpotentials are not excessive (more
than a few hundred millivolts).
9
7866 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Low-Coordinate Iron(II) Fluoride Complexes
A R T I C L E S
performed by Desert Analytics (Tucson, AZ). Vibrational spectra were
recorded (450-4000 cm-1) on KBr pellet samples in a Shimadzu
Fourier transform infrared spectrophotometer (FTIR-8400S). A total
of 64 scans at a 2 cm-1 resolution were collected in each case.
Absorptions that were common among a pair ([LMeFeF]2 and
[LMeFeCl]2, LtBuFeF and LtBuFeCl) or series (LtBuFeF(L′) or LMeFeF(L′))
of complexes were rejected on the grounds that they likely originate
from the LMeFe- fragment, rather than the Fe-F vibrations. Bands in
the 450-1000 cm-1 region are thus reported according to their intensity
(s ) strong, m ) medium, w ) weak). Electronic spectra were recorded
between 400 and 1100 nm with a Cary 50Bio UV-visible spectro-
photometer, using quartz cuvettes of 1 cm optical path length. Pentane,
diethyl ether, tetrahydrofuran (THF), toluene, and acetonitrile were
purified by passage through activated alumina and “deoxygenizer”
columns from Glass Contour Co. (Laguna Beach, CA). Deuterated
benzene, tetrahydrofuran, and pyridine were dried over CaH2, then over
Na, and then vacuum distilled into storage containers or directly into
an NMR tube. Hexamethyldisilathiane (HMDT), trimethyltin chloride,
potassium fluoride (99+%), 3,3,3-trifluoropropene (99%), hexafluo-
ropropene (99+%), and 4-phenylpyridine were purchased from Aldrich
and used as received. Hexafluorobenzene, octafluorotoluene, pentafluo-
robenzene, pentafluoropyridine, p-heptafluorotoluene, triethylsilane,
triethylsilane-d, triphenylsilane, dimethyl-3,5-bis(trifluoromethylphe-
nyl)silane, triethoxysilane, 4-tertbutylpyridine, pyridine, 4-trifluorom-
ethylpyridine, and boron trifluoride diethyl etherate were dried over
activated molecular sieves or vacuum distilled prior to use. Super-
Hydride (KHBEt3),36 LtBuFeCl,15f [LMeFeCl]2,15a LMeFe(µ-Cl)2Li(THF)2,15f
Conclusions
Unprecedented three- and four-coordinate iron(II) fluorides
can be isolated using the â-diketiminate ligands LMe and LtBu
.
These represent some of the first examples of discrete iron(II)
fluoride complexes and the first example of a three-coordinate
transition metal complex containing the fluoride ligand. The
FeII-F bond in these complexes is short, and they are very
thermally robust. However, with suitable fluoride acceptors the
compounds become reactive. For example, trialkylsilyl-substi-
tuted compounds lead to iron complexes with thiolate, acetylide,
and hydride ligands, and BF3‚OEt2 removes the fluoride ligand.
Combining (a) the ability to transfer fluoride to silicon and
(b) the high reactivity of the corresponding hydride complexes,
it is possible to perform catalytic hydrodefluorinations. This
process gives HDF of fluoroarenes and fluoroalkenes. There
are few catalysts for homogeneously catalyzed C-F cleavage
reactions,11-13 and many in the literature that perform HDF are
based on more expensive rhodium complexes.12 Because the
rhodium complexes are able to follow an oxidative addition
mechanism, C-H activation competes with C-F activation. In
contrast, the iron(II) complexes described here do not react with
C-H bonds, perhaps because the first-row metal has a prefer-
ence for M-F over M-H bonds.50,51
Despite the significance of the homogeneous iron-catalyzed
HDF process described here, it is clear that major improvements
are necessary. High catalyst loadings (20 mol %) are needed
and the turnover numbers observed are low due to catalyst
degradation at the temperatures needed for catalysis. This
difficulty has also hindered mechanistic studies on the HDF
reactions of perfluoroarenes. Our current mechanistic investiga-
tions indicate that the rate-limiting step for HDF of a perfluo-
roarene (octafluorotoluene) is the regeneration of the active
hydrodefluorinating species from the iron fluoride and the silane.
Although oxidative addition, as well as free-radical or other outer
sphere electron transfer pathways are unlikely for perfluoroarene
HDF, ruling out some of the remaining mechanisms will await
a more robust or more rapid catalyst based on the discoveries
demonstrated here. At this point, our working model is based
on nucleophilic attack of hydride (in an unknown active species)
on aromatic fluorocarbons. On the other hand, the evidence
strongly supports a hydride insertion/â-fluoride elimination
mechanism for olefinic HDF.
and all alkyl complexes such as LMeFe-iBu and LMeFeCH2CH2CF3
18
were prepared by known procedures.
Synthesis of Trimethyltin Fluoride, Me3SnF. Details reported on
the preparation of Me3SnF date back to the year 1918, and the reference
is not widely available.19 Thus we describe here our synthetic procedure.
To a stirred solution of trimethyltin chloride (2 g, 10 mmol) in absolute
ethanol (4 mL) was added dropwise a solution of KF (870 mg, 15
mmol) in deionized water (3 mL). The white precipitate was collected
by filtration through a frit and washed with deionized water (1 mL),
ethanol (1 mL), and diethyl ether (1 mL). The solid product was dried
under high vacuum overnight before use (1.1 g, 61%).
[LMeFe(µ-F)]2 (12). Trimethyltin fluoride (240 mg, 1.32 mmol),
LMeFe-iBu (600 mg, 1.2 mmol), and toluene (8 mL) were placed in a
resealable Schlenk bomb with a stir bar. The mixture was stirred and
heated to 80 °C overnight. After cooling the mixture to room
temperature and allowing the precipitate to settle, the soluble fraction
was filtered through Celite in a frit. After solvent removal under vacuum
a green solid was obtained, which was recrystallized from diethyl ether
(490 mg, 83% yield). Anal. Found (calcd): C, 70.75 (70.72), H, 8.63
1
(8.39), N, 5.62 (5.69). µeff (per dimer; C6D6, 21 °C) ) 6.2(3) µB. H
Experimental Section
NMR (C6D6, 21 °C): 14.0 (1H, R-CH), 5.2 (4H, m-CH), 2.0 (12H,
General Considerations. Manipulations were performed under a
nitrogen atmosphere by standard Schlenk techniques or in an M. Braun
Unilab N2-filled glovebox maintained at or below 1 ppm of O2 and
H2O. Glassware was dried at 130 °C overnight. 1H and 19F NMR data
were recorded on a Bruker Avance 400 MHz spectrometer at the
specified temperature. 1H shifts are reported in ppm, relative to residual
protiated solvent in C6D6 (7.13 ppm) or THF-d8 (3.58 ppm); relative
iPr-CH3), -10 (4H, iPr-CH), -11.7 (6H, CH3-L), -46.3 (2H, p-CH),
i
-51.0 (12H, Pr-CH3). The 19F NMR of 12 showed no resonances
between 500 and -500 ppm. No IR bands corresponding to the Fe-F
bond vibrations were observed, and the IR spectra of 12 and [LMeFeCl]2
were nearly identical. Vis (toluene): 397 nm (3600 M-1 cm-1), 455
nm (560 M-1 cm-1), 791 nm (90 M-1 cm-1).
LtBuFeF (2). Trimethyltin fluoride (300 mg, 1.6 mmol), LtBuFeMe15c
(1.02 g, 1.8 mmol), and toluene (20 mL) were placed in a resealable
flask and heated to 80 °C while stirring overnight. Then the solvent
was evaporated under vacuum, and the solid residue washed with
pentane (5 × 5 mL). After vacuum-drying a pink powder was obtained
(760 mg, 74%). Crystals for X-ray diffraction were grown by cooling
a saturated toluene solution from 80 °C to room temperature. Anal.
Found (calcd): C, 72.37 (72.90), H, 8.96 (9.26), N, 4.71 (4.86). µeff
integrations of peaks (and assignment when solved) are also given. 19
F
shifts were referenced to R,R,R-trifluorotoluene (δ -63.73 ppm) and
are reported against CFCl3 (0 ppm). Solution magnetic susceptibilities
were determined at 294 K by the Evans method.52 Microanalyses were
(50) (a) Kraft, B. M.; Lachicotte, R. J.; Jones, W. D. Organometallics 2002,
21, 727-731. (b) Clot, E.; Me´gret, C.; Kraft, B. M.; Eisenstein, O.; Jones,
W. D. J. Am. Chem. Soc. 2004, 126, 5647-5653. (c) Ge´rard, H.; Eisenstein,
O.; Dalton Trans. 2003, 839-845. (d) Strazisar, S. A.; Wolczanski, P. T.
J. Am. Chem. Soc. 2001, 123, 4728-4740.
(51) (a) Reinhold, M.; McGrady, J. E.; Perutz, R. N. J. Am. Chem. Soc. 2004,
126, 5268-5276.
(52) (a) Schubert, E. M. J. Chem. Educ. 1992, 69, 62. (b) Evans, D. F. J. Chem.
Soc. 1959, 2003-2005.
1
(toluene-d8, 21 °C) ) 5.6(3) µB. H NMR (toluene-d8, 21 °C): 42.1
(18H, (CH3)3C-L), -27.0 (2H, m-CH), -31.1 (12H, iPr-CH3), -98.5
i
i
(2H, p-CH), -113.5 (16H, Pr-CH3, Pr-CH). The 19F NMR of 2 in
toluene-d8 or THF-d8 showed no resonances between 500 and -500
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7867

A R T I C L E S
Vela et al.
ppm. IR (KBr pellet, cm-1): 596 (s, νFe-F). Vis (toluene): 530 nm
of solvent under vacuum, the material was dissolved in diethyl ether
(3 mL) and cooled to -38 °C to give 4 (98 mg, 86%). Anal. Found
(calcd): C, 68.83 (68.85), H, 9.72 (9.43), N, 3.97 (4.23). µeff (C6D6,
(730 M-1 cm-1), 715 nm (100 M-1 cm-1).
Formation of Four-Coordinate Fluoride Complexes. The substi-
tuted pyridine (0.2 mmol) or acetonitrile (1.0 mmol) was added to the
fluoride complex 12 (100 mg, 0.1 mmol) or 2 (115 mg, 0.2 mmol) in
diethyl ether (5 mL). Each individual product was isolated by
crystallization from these solutions at -38 °C. LMeFeF(4-tBu-py)
(1‚tBupy): 87% yield. Anal. Found (calcd): C, 72.19 (72.71), H, 8.93
(8.67), N, 7.12 (6.69). µeff (C6D6, 21 °C) ) 5.1(3) µB. 1H NMR (C6D6,
21 °C): 36.4 (4H, m-CH), 20.8 (2H, o-CH-py), 16.8 (2H, m-CH-py),
1.5 (12H, iPr-CH3), -2.5 (12H, iPr-CH3), -6.8 (4H, iPr-CH), -12.2
(9H, (CH3)3C-py), -38.5 (2H, p-CH), -64.0 (1H, R-CH), -86.3 (6H,
(CH3)2-L). IR (KBr pellet, cm-1): 717 (m, νFe-F). Vis (toluene): 425
nm (2180 M-1 cm-1), 465 nm (1260 M-1 cm-1), 946 nm (150 M-1
cm-1). LMeFeF(4-CF3-py) (1‚CF3py): 68% yield. Anal. Found (cal-
cd): C, 65.96 (65.73), H, 7.12 (7.09), N, 6.61 (6.57). µeff (C6D6, 21
1
21 °C) ) 5.5(3) µB. H NMR (C6D6, 21 °C): 115 (1H, R-CH), 70.7
(9H, (CH3)3Si), 43.8 (18H, (CH3)3C-L), -26.1 (14H, iPr-CH3, m-CH),
i
i
-106.0 (4H, Pr-CH), -120.0 (12H, p-CH, Pr-CH3, p-CH).
Spectroscopic Observation of LtBuFe(µ-S)FeLtBu. When 2 and
HMDS are reacted in equimolar amounts, this sulfide complex was
1
formed along with 4 in a 1:1 ratio. The H NMR spectrum is very
similar to that of the related sulfide complex LMeFe(µ-S)FeLMe 16 1H
.
NMR (C6D6, 21 °C): 23.9 (1H, R-CH), 12.4 (18H, (CH3)3C-L), 6.2
i
i
(4H, m-CH), -2.3 (12H, Pr-CH3), -7.0 (4H, Pr-CH), 18.5 (2H,
i
p-CH), 19.0 (12H, Pr-CH3).
Reaction of [LMeFe(µ-F)]2 (12) with Hexamethyldisilathiane. The
reaction of 12 (9.8 mg, 10 µmol) and HMDS (4.3 µL, 20 µmol)
proceeded cleanly in C6D6 at 60 °C for 4 h to give LMeFe(µ-S)FeLMe
1
(100% by H NMR).16
1
°C) ) 5.1(3) µB. H NMR (C6D6, 21 °C): 21.0 (2H, o-CH-py), 20.3
LtBuFeCCSiMe3 (5). Prepared in a similar way to 3 from bis-
(trimethylsilyl)acetylene (140 µL, 0.62 mmol) and 2 (180 mg, 0.31
mmol): 124 mg, 61%. Anal. Found (calcd): C, 73.17 (73.36), H, 9.32
(9.54), N, 4.36 (4.28). µeff (C6D6, 21 °C) ) 5.8(3) µB. 1H NMR (C6D6,
21 °C): 112 (1H, R-CH), 57.4 (9H, (CH3)3Si), 43.1 (18H, (CH3)3C-
i
(4H, m-CH), 10.0 (2H, m-CH-py), 1.1 (12H, Pr-CH3), -9.3 (12H,
iPr-CH3), -39.2 (2H, p-CH), -51.0 (4H, iPr-CH), -69.0 (1H, R-CH),
-87.8 (6H, (CH3)2-L). IR (KBr pellet, cm-1): 717 (w), 669 (s, νFe-F),
609 (m), 557 (m). Vis (toluene): 401 nm (1930 M-1 cm-1), 512 nm
(670 M-1 cm-1), 898 nm (120 M-1 cm-1). LtBuFeF(4-tBu-py)
(2‚tBupy): 95% yield. Anal. Found (calcd): C, 74.45 (74.24), H, 8.68
(9.35), N, 5.82 (5.90). µeff (C6D6, 21 °C) ) 5.0(3) µB. 1H NMR (C6D6,
21 °C): 38.0, 21.8, 7.9, 2.0, 1.1, -2.3, -46.5, -88.0. IR (KBr pellet,
cm-1): 544 (s, νFe-F). Vis (toluene): 462 nm (2110 M-1 cm-1), 968
nm (150 M-1 cm-1). LtBuFeF(NCCH3) (2‚ACN): 61% yield. Anal.
Found (calcd): C, 71.44 (71.94), H, 9.56 (9.14), N, 7.02 (6.80). µeff
i
L), -27.7 (12H, Pr-CH3), -32.5 (2H, m-CH), -116 (18H, p-CH,
i
iPr-CH3, Pr-CH). IR (KBr pellet, cm-1): 2092 (νCtC).
[LMeFe(µ-H)]2 (62). Prepared in a similar way to 3 and 4, from
triethylsilane (61 µL, 0.38 mmol) and 12 (188 mg, 0.19 mmol): 160
mg, 89%. Anal. Found (calcd): C, 73.14 (73.41), H, 8.39 (8.92), N,
1
5.98 (5.90). µeff (C6D6, 21 °C) ) 4.0(3) µB. H NMR (C6D6, 21 °C):
i
1
13.2 (12H, Pr-CH3), 3.2 (2H, p-CH), 1.1 (4H, m-CH), -24.6 (18H,
(C6D6, 21 °C) ) 4.9(3) µB. H NMR (C6D6, 21 °C): 21.4, 17.4, 12.0,
iPr-CH3, CH3-L), -57.6 (4H, iPr-CH). Virtually identical 1H
resonances and chemical shifts were observed at room temperature in
THF-d8, indicating that in contrast to the LtBu analogue 72,17 62 does
not dissociate into a monomer in the absence of strong donor ligands
(e.g., pyridine).
7.9, 1.0, -4.3, -13.9, -49.8, -85.1. IR (KBr pellet, cm-1): 538 (s,
ν
Fe-F). Vis (toluene-acetonitrile, 10:1 v/v): 457 nm (1040 M-1 cm-1),
920 nm (150 M-1 cm-1). The following adducts were observed
spectroscopically: [LMeFeF(py)] H NMR (C6D6, 21 °C): 35.0 (5H,
1
o-, p-, m-CH-py), 19.0 (4H, m-CH), 2.1 (12H, iPr-CH3), -10.0 (16H,
iPr-CH, iPr-CH3), -38.7 (2H, p-CH), -68.1 (1H, R-CH), -87.2 (6H,
(CH3)2-L). IR (KBr pellet, cm-1): 704 (s, νFe-F). Vis (toluene): 432
nm (1630 M-1 cm-1), 468 nm (1110 M-1 cm-1), 944 nm (130 M-1
cm-1). [LMeFeF(4-Ph-py)] 1H NMR (C6D6, 21 °C): 36.0 (4H, m-CH),
21.0 (2H, o-CH-py), 17.3 (2H, m-CH-py), 11.5 (1H, p-CH-Phpy), 10.7
(2H, m-CH-Phpy), 3.3 (2H, o-CH-Phpy), 1.3 (12H, iPr-CH3), -6.6
(4H, iPr-CH), -12.1 (12H, iPr-CH3), -38.6 (2H, p-CH), -66.0 (1H,
R-CH), -86.8 (6H, (CH3)2-L). IR (KBr pellet, cm-1): 733 (w), 690
(m, νFe-F), 623 (w), 544 (m). Vis (toluene): 411 nm (1840 M-1 cm-1),
480 nm (1960 M-1 cm-1), 943 nm (210 M-1 cm-1). [LtBuFeF(py)] 1H
NMR (C6D6, 21 °C): 39.5, 20.0, 7.3, 2.5, 1.2, -3.6, -47.3, -87.2. IR
(KBr pellet, cm-1): 540 (s, νFe-F). Vis (toluene): 467 nm (2070 M-1
Reaction of LtBuFeF (2) with Et3SiH. A J. Young NMR tube was
loaded with 2 (10 mg, 17 µmol), Et3SiH (2.8 µL, 17 µmol), and C6D6
(0.4 mL). The tube was sealed and heated to 45 °C for 12 h. Complete
conversion to 72 (1H NMR) and Et3SiF (19F NMR, R,R,R-trifluoro-
toluene used as internal standard, see below) was observed.
HDF of Fluoroarenes. A J. Young NMR tube was loaded with a
solution of substrate ([ArF] ) 0.11 M), a trisubstituted silane ([R3SiH]
) 0.11 M), and one of 12, 62, 72 (0.01 M), or 2 (0.02 M). The tube was
heated to the specified temperature in an oil bath, and NMR spectra
(19F, 1H) were recorded periodically. A capillary containing a solution
of R,R,R-trifluorotoluene was used as an internal standard for chemical
shift and integration purposes. HDF products were identified by a
combination of 19F NMR and GC-MS data. In all cases, fluorotrieth-
ylsilane was observed by 19F NMR (-176.7 ppm) and by GC-MS
analysis of the reaction mixture (F-SiEt3, m/z ) 134).
Kinetics of Perfluoroarene HDF. The above procedure was
repeated with octafluorotoluene while varying the initial concentrations
of reactants (one at a time).37 The sample was placed on a previously
equilibrated and temperature-calibrated53 NMR probe, and 19F spectra
were recorded periodically. Monitoring was continued up to 10-20
mol % substrate conversions. When triethylsilane-d was used, deuterium
incorporation into the HDF product was observed by GC-MS (DC6F4-
CF3 m/z ) 219).
1
cm-1), 965 nm (160 M-1 cm-1). [LtBuFeF(THF)] H NMR (THF-d8,
adduct, 21 °C): 28.0 (1H, CH3-L), 27.4 (18H, (CH3)3C-L), -4.0 (12H,
i
iPr-CH3), -25.4 (4H, m-CH), -50.6 (12H, Pr-CH3), -57.5 (2H,
i
p-CH), -69.7 (4H, Pr-CH). Vis (THF): 464 nm (930 M-1 cm-1),
979 nm (120 M-1 cm-1).
LtBuFeOEt2(η1-BF4) (3). F3B‚OEt2 (20 µL, 0.16 mmol) was added
to a stirred suspension of 2 (92 mg, 0.16 mmol) in diethyl ether (3
mL), causing the immediate dissolution of the pink powder to form a
yellow solution. Upon standing of this solution at -38 °C, 3 was
obtained as yellow crystals (53 mg in two crops, 46%). Anal. Found
(calcd): C, 65.01 (65.19), H, 8.17 (8.84), N, 3.99 (3.90). µeff(THF-d8,
21 °C) ) 4.6(3) µB. 1H NMR (THF-d8, 21 °C): 23.2, 15.2, 9.8, -23.0,
-47.1, -51.3, -48.0, -70.0, -106.0. The 19F NMR of 3 in THF-d8
Radical Scavenger Experiments. The octafluoroarene HDF was
repeated as described above and in parallel with and without one of
the radical traps dihydroanthracene (0.047 M, 0.43 equiv) or triph-
enylmethane (0.044 M, 0.40 equiv).
shows a single resonance that is identical to that seen for Na+BF4
-
-
(155.1 ppm), suggesting that complete dissociation of the BF4 anion
HDF of Fluoroolefins. The above procedure was repeated, but the
gaseous olefinic substrate (hexafluoropropene or 3,3,3-trifluoropropene)
occurs in solution.
Reaction of LtBuFeF (2) with Hexamethyldisilathiane. Hexa-
methyldisilathiane (60 mg, 0.34 mmol, caution: STENCH!), 2 (100
mg, 0.17 mmol), and toluene (5 mL) were placed in a resealable flask
and heated to 80 °C while stirring overnight. After thorough evaporation
(53) (a) Ammann, C.; Meier, P.; Merbach, A. E. J. Magn. Reson. 1982, 46,
319-321. (b) Kaplan, M. L.; Bovey, F. A.; Cheng, H. N. Anal. Chem.
1975, 47, 1703-1705.
9
7868 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005

Low-Coordinate Iron(II) Fluoride Complexes
A R T I C L E S
Table 6. Details of X-ray Crystal Structures
[LMeFe(
(12)
µ
-F)]2
LtBuFeF(4-tBu-py)
(2
tBupy)
LtBuFeCCSiMe3
LtBuFeF (2)
‚
LtBuFeSSiMe3 (4)
(5)
empirical formula
fw
C58H82F2Fe2N4
984.98
C35H53FFeN2
576.64
C44H66FFeN3
711.85
C38H62FeN2SSi
662.90
C40H62FeN2Si
654.86
cryst syst
space group
a (Å)
b (Å)
c (Å)
monoclinic
C2/c
monoclinic
C2/c
monoclinic
P21/c
monoclinic
P21/n
monoclinic
P21/c
15.2552(10)
16.7611(11)
21.8794(14)
91.1460(10)
5593.3(6)
4
1.170
0.563
0.0543, 0.1057
1.076
23.480(3)
8.5555(11)
17.105(2)
108.491(3)
3258.7(7)
4
1.175
0.493
0.0409, 0.0886
1.030
13.9028(6)
20.4954(9)
18.8040(6)
128.169(2)
4212.5(3)
4
1.122
0.394
0.0523, 0.1150
1.016
15.4030(8)
26.3684(15)
19.7988(11)
97.3040(10)
7976.1(8)
8
1.104
0.486
0.0479, 0.0779
0.899
16.941(2)
12.5476(16)
19.408(3)
90.781(2)
4125.3(9)
4
1.054
0.421
0.0453, 0.0953
1.044
â (deg)
V (Å3)
Z
F (g/cm3)
µ (mm-1
)
R1, wR2 (I > 2σ(I))
GOF
LMeFeF(4-CF3-py)
(THF)
‚
LtBuFeF(NCCH3)
(CH3CN)
‚
LMeFeF(4-tBu-py)
LtBuFeOEt2(
η
1-BF4)
[LMeFe(
(OEt2) (62
µ
‚
-H)]2
‚
(1‚
tBupy)
(1‚
CF3py
‚
THF)
(2‚
ACN
‚ACN)
(3)
OEt2)
empirical formula
C38H54FFeN3
(C38H51F4FeN3)‚(C4H8O)
C39H59FFeN
C39H63BF4FeN2O
(C58H84Fe2N4)‚
(C4H10O)
1023.11
monoclinic
P21/c
11.5959(15)
18.315(2)
28.590(4)
90
98.870(2)
90
5999.4(13)
4
fw
627.69
triclinic
P1h
753.77
triclinic
P1h
658.75
triclinic
P1h
718.57
triclinic
P1h
cryst syst
space group
a (Å)
b (Å)
c (Å)
8.6321(10)
12.4465(14)
18.430(2)
92.217(2)
91.025(2)
99.486(2)
1951.0(4)
2
1.068
0.417
0.0956, 0.2932
1.281
8.9028(8)
12.6570(11)
18.2746(15)
93.3230(10)
101.575(2)
94.635(2)
2004.9(3)
2
1.249
0.430
0.0860, 0.2304
1.039
9.8640(9)
12.2956(11)
17.9941(16)
96.695(2)
97.672(2)
112.544(2)
1963.7(3)
2
1.114
0.418
0.0944, 0.1910
1.390
10.0195(11)
12.4908(13)
18.648(2)
107.879(2)
103.667(2)
91.076(2)
7976.1(8)
2
1.111
0.397
0.0399, 0.1157
1.122
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
F (g/cm3)
1.133
0.524
0.0791, 0.1954
1.048
µ (mm-1
)
R1, wR2 (I > 2σ(I))
GOF
was condensed from a calibrated volume bulb into a resealable NMR
tube containing a THF-d8 solution of 12 (0.01 M) and triethylsilane
(0.11 M). The individual HDF products, 1,2,3,3,3-pentafluoropropene
(CHFdCFCH3, both E and Z isomers),54 1,1,3,3,3-pentafluoropropene
(CF2dCHCF3),55 or 1,1-difluoropropene (CF2dCHCH3),5b,56 were
identified by 19F NMR by comparison to the reported data on chemical
shifts and coupling constants for each of these compounds, as well as
by GC-MS (m/z ) 132). Relative concentrations of reactant and
products were obtained from integration of 19F resonances, assuming
that all gases had similar solubility. During the HDF of R,R,R-
trifluoropropene, the only species observed by NMR was the previously
reported alkyl complex LMeFeCH2CH2CF3.18a 1H NMR (THF-d8, 21
°C): 159.0 (2H, p-CH), 10.7 (4H, m-CH), 0.5 (12H, iPr-CH3), -18.4
(1H, R-CH-L), -30.7 (12H, Pr-CH3), -45.1 (4H, Pr-CH), -45.7
(6H, CH3-L). 19F NMR (THF-d8, 21 °C): 46 ppm (3H, CF3 terminus).
A similarly distinctive paramagnetically broadened signal at 53 ppm
(presumably attributable to LMeFeCF2CF2CF3) was observed by 19F
NMR during the HDF of hexafluoropropene. Neither fluoride nor
hydride species were observed by NMR during catalysis.
0.2 M NBu4OTf supporting electrolyte (Aldrich, 99%) and 0.003 M
analyte. The scan rate in all experiments was 0.1 V/s. Both hydride
complexes showed an irreversible oxidation at -0.4 V vs the Ag
electrode (ca. -0.6 V vs ferrocene (Fc); ferrocene could not be used
directly as an internal reference because it showed noninnocent behavior
during the electrochemical experiments). Chemical study. To substanti-
ate the results of the electrochemical experiments, the hydride
complexes were exposed to two chemical oxidants of varying strength
in THF-d8 and the fate of the reaction followed by 1H NMR. A
stoichiometric amount of either Cp2Fe+B(ArF)- (E°rel ) 0 V; B(ArF)-
) tetrakis(3,5-bis(trifluoromethyl)phenyl)borate)57 or even the weaker
oxidant Cp*2Fe+B(ArF)- (E°rel ) -0.59 V vs Fc)57 instantaneously
reacts with the hydride complexes. Formation of Cp2Fe (4.05 ppm)
and Cp*2Fe (1.42 ppm) was observed by 1H NMR spectroscopy, along
with paramagnetic products that were not further characterized.
X-ray Structures. Crystalline samples were grown in the glovebox
from pentane or ether solutions at -38 °C. Each sample was rapidly
mounted under Paratone-8277 onto a glass fiber and immediately placed
in a cold nitrogen stream at -80 °C on the X-ray diffractometer. X-ray
intensity data were collected on a standard Bruker-axs SMART CCD
area detector system equipped with a normal focus molybdenum-target
X-ray tube operated at 2.0 kW (50 kV, 40 mA). A total of 1121 frames
(for 12, 2‚ACN, 4, and 62) or 2424 frames (for 2, 3, 5, 1‚tBupy, 1‚
CF3py, and 2‚tBupy) of data were collected using a narrow frame
method with scan widths of 0.3° in ω. Frames were integrated to a
maximum 2θ angle of 56.6° with SAINT. The final unit cell parameters
(at -80 °C) were determined from the least-squares refinement of three-
dimensional centroids of >4000 reflections for each crystal. Data were
i
i
Oxidation of Hydride Complexes. Electrochemical Study. Cyclic
voltammetry measurements on [LMeFeH]2 62 and [LtBuFeH]2 72 were
carried out under an inert atmosphere (e1 ppm O2) with a Cypress
Systems Potentiostat/Electroanalytical System CS-1200 inside an argon
drybox. A glassy graphite electrode was used as the working electrode,
and two silver wires were used as auxiliary and reference electrodes.
The CV experiments were performed in THF at room temperature with
(54) (a) Koroniak, H.; Palmer, K. W.; Dolbier, W. R.; Zhang, H. Magn. Reson.
Chem. 1993, 31, 748-751. (b) Burton, D. J.; Spawn, T. W.; Heinze, P. L.;
Bailey, A. R.; Shin-Ya, S. J. Fluorine Chem. 1989, 44, 167-174.
(55) Fields, R.; Germain, M. M.; Haszeldine, R. N.; Wiggans, P. W. J. Chem.
Soc. (A) 1970, 1961-1974.
(56) Shaler, T. A.; Morton, T. H. J. Am. Chem. Soc. 1991, 113, 6771-6779.
(57) Connelly, N. G.; Geiger, W. E. Chem. ReV. 1996, 96, 877-910.
9
J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005 7869

A R T I C L E S
Vela et al.
corrected for absorption with the SADABS58 program. The space groups
were assigned using XPREP, and the structures were solved by direct
methods and refined employing full-matrix least-squares on F2 (Bruker-
axs, SHELXTL-NT,59 version 5.10). All non-H atoms were refined with
anisotropic thermal parameters. Iron-bound hydrogen atoms in 62 were
located in the electron density map and refined with isotropic thermal
parameters. All other hydrogen atoms were included in idealized
positions. The structures refined to goodness of fit values and final
residuals found in Table 6.
Acknowledgment. Funding for this work was provided by
the National Science Foundation (CHE-0112658), the Petroleum
Research Fund (38275-G to P.L.H.), the A. P. Sloan Foundation
(Research Fellowship to P.L.H.), and the University of Rochester
(Hooker and Messersmith fellowships to J.V.). We are indebted
to Bill Jones, Brad Kraft, Joe Dinnocenzo, and an anonymous
reviewer for helpful comments, and to Karin Ruhlandt for the
use of an X-ray diffractometer.
Supporting Information Available: Additional spectroscopic
and crystallographic data. This material is available free of
charge via the Internet at http://pubs.acs.org.
(58) The SADABS program is based on the method of Blessing; see Blessing,
R. H. Acta Crystallogr. A 1995, 51, 33-38.
(59) SHELXTL NT: Structure Analysis Program, version 5.10; Bruker-AXS:
Madison, WI, 1995.
JA042672L
9
7870 J. AM. CHEM. SOC. VOL. 127, NO. 21, 2005