Cationic terminal gallylene complexes by halide abstraction: Coordination chemistry of a valence isoelectronic analogue of CO and N2

Article abstract of DOI:10.1021/ja806655f

While N₂ and CO have played central roles in developing models of electronic structure, and their interactions with transition metals have been widely investigated, the valence isoelectronic diatomic molecules EX (E = group 13 element, X = group 17 element) have yet to be isolated under ambient conditions, either as the free molecule or as a ligand in a simple metal complex. As part of a program designed to address this deficiency, together with wider issues of the chemistry of cationic systems [L _nM(ER)]⁺ (E = B, Al, Ga; R = aryl, amido, halide), we have targeted complexes of the type [L_nM(GaX)]⁺. Halide abstraction is shown to be a viable method for the generation of mononuclear cationic complexes containing gallium donor ligands. The ability to isolate tractable two-coordinate products, however, is strongly dependent on the steric and electronic properties of the metal/ligand fragment. In the case of complexes containing ancillary π-acceptor ligands such as CO, cationic complexes can only be isolated as base-trapped adducts, even with bulky aryl substituents at gallium. Base-free gallylene species such as [Cp*Fe(CO) ₂(GaMes)]⁺ can be identified only in the vapor phase by electrospray mass spectrometry experiments. With bis(phosphine) donor sets at the metal, the more favorable steric/electronic environment allows for the isolation of two-coordinate ligand systems, even with halide substituents at gallium. Thus, [Cp*Fe(dppe)(Gal)]⁺[BAr^f ₄]^- (9) can be synthesized and shown crystallographically to feature a terminally bound Gal ligand; 9 represents the first experimental realization of a complex containing a valence isoelectronic group 13/group 17 analogue of CO and N₂. DFT calculations reveal a relatively weakly bound Gal ligand, which is confirmed experimentally by the reaction of 9 with CO to give [Cp*Fe(dppe)-(CO)]⁺[BAr^f₄] ^-. In the absence of such reagents, 9 is stable for weeks in fluorobenzene solution, presumably reflecting (i) effective steric shielding of the gallium center by the ancillary phosphine and Cp* ligands; (ii) a net cationic charge which retards the tendency toward dimerization found for putative charge neutral systems; and (iii) (albeit relatively minor) population of the LUMOs of the Gal molecule through π overlap with the HOMO and HOMO-2 of the [Cp*Fe(dppe)]⁺ fragment.

Full text of DOI:10.1021/ja806655f

Published on Web 11/01/2008
Cationic Terminal Gallylene Complexes by Halide Abstraction:
Coordination Chemistry of a Valence Isoelectronic Analogue
of CO and N2
Natalie D. Coombs,†,‡ Dragoslav Vidovic,† Joanna K. Day,†,‡ Amber L. Thompson,†
Delphine D. Le Pevelen,† Andreas Stasch,§ William Clegg,| Luca Russo,|
Louise Male, Michael B. Hursthouse, David J. Willock,‡ and Simon Aldridge*,†
Inorganic Chemistry, UniVersity of Oxford, South Parks Road, Oxford, U.K. OX1 3QR, Cardiff
School of Chemistry, Main Building, Park Place, Cardiff, U.K. CF10 3AT, School of Chemistry,
Monash UniVersity, P.O. Box 23, Victoria 3800, Australia, School of Chemistry, Bedson
Building, Newcastle UniVersity, Newcastle upon Tyne, U.K. NE1 7RU, and School of Chemistry,
UniVersity of Southampton, Southampton, U.K. SO17 1BJ
Received August 22, 2008; E-mail: Simon.Aldridge@chem.ox.ac.uk
Abstract: While N2 and CO have played central roles in developing models of electronic structure, and
their interactions with transition metals have been widely investigated, the valence isoelectronic diatomic
molecules EX (E ) group 13 element, X ) group 17 element) have yet to be isolated under ambient
conditions, either as the “free” molecule or as a ligand in a simple metal complex. As part of a program
designed to address this deficiency, together with wider issues of the chemistry of cationic systems
[LnM(ER)]+ (E ) B, Al, Ga; R ) aryl, amido, halide), we have targeted complexes of the type [LnM(GaX)]+.
Halide abstraction is shown to be a viable method for the generation of mononuclear cationic complexes
containing gallium donor ligands. The ability to isolate tractable two-coordinate products, however, is strongly
dependent on the steric and electronic properties of the metal/ligand fragment. In the case of complexes
containing ancillary π-acceptor ligands such as CO, cationic complexes can only be isolated as base-
trapped adducts, even with bulky aryl substituents at gallium. Base-free gallylene species such as
[Cp*Fe(CO)2(GaMes)]+ can be identified only in the vapor phase by electrospray mass spectrometry
experiments. With bis(phosphine) donor sets at the metal, the more favorable steric/electronic environment
allows for the isolation of two-coordinate ligand systems, even with halide substituents at gallium. Thus,
[Cp*Fe(dppe)(GaI)]+[BArf4]- (9) can be synthesized and shown crystallographically to feature a terminally
bound GaI ligand; 9 represents the first experimental realization of a complex containing a valence
isoelectronic group 13/group 17 analogue of CO and N2. DFT calculations reveal a relatively weakly bound
GaI ligand, which is confirmed experimentally by the reaction of 9 with CO to give [Cp*Fe(dppe)-
(CO)]+[BArf4]-. In the absence of such reagents, 9 is stable for weeks in fluorobenzene solution, presumably
reflecting (i) effective steric shielding of the gallium center by the ancillary phosphine and Cp* ligands; (ii)
a net cationic charge which retards the tendency toward dimerization found for putative charge neutral
systems; and (iii) (albeit relatively minor) population of the LUMOs of the GaI molecule through π overlap
with the HOMO and HOMO-2 of the [Cp*Fe(dppe)]+ fragment.
first dinitrogen complexes in 1965;4 by contrast, the stronger
Introduction
metal-ligand bonds formed by the isoelectronic CO molecule
The fundamental changes in electronic structure which
accompany the interaction of a diatomic molecule with a
transition metal center have been exploited to great effect in
coordination chemistry,1 both in the activation of inherently inert
molecules (such as N2)2 and in the isolation and characterization
of otherwise labile species (e.g., CX, where X ) S, Se, Te).3
Thus, for example, relevance to the biological conversion of
N2 to NH3 was the driving force behind the isolation of the
underpin its widespread exploitation in low oxidation state
transition metal chemistry (ca. 35,000 structurally characterized
examples).5 Within this family of ten valence electron diatomic
molecules, species of the type EX (E ) group 13 element, X
) group 17 element), although predicted to be inherently less
stable in the “free” state (due to small HOMO/LUMO gaps),
(3) (a) Baird, M. C.; Wilkinson, G. Chem. Commun. (London) 1966, 267.
(b) Clark, G. R.; James, S. M. J. Organomet. Chem. 1977, 134, 229.
(c) Clark, G. R.; Marsden, K.; Roper, W. R.; Wright, L. J. J. Am.
Chem. Soc. 1980, 102, 1206. (d) Clark, G. R.; Marsden, K.; Rickard,
C. E. F.; Roper, W. R.; Wright, L. J. J. Organomet. Chem. 1988,
338, 393.
† University of Oxford.
‡ Cardiff School of Chemistry.
§ Monash University.
| Newcastle University.
University of Southampton.
(4) Allen, A. D.; Senoff, C. V. Chem. Commun. (London) 1965, 621.
(5) As determined from a survey of the Cambridge Structural Database
22/08/2008.
(1) See, for example: (a) Thorp, H. H. Science 2000, 289, 882.
(2) Laplaza, C. E.; Cummins, C. C. Science 1995, 268, 861.
9
10.1021/ja806655f CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 16111–16124 16111

A R T I C L E S
Coombs et al.
are thought to offer eVen more faVorable thermodynamics of
binding to metal centers.6-8 Such predictions stem from the
stronger σ donor properties (higher HOMO energies) expected
for group 13 donor ligands and the greater amplitude of the π
acceptor LUMO at E. Despite this, the complexes so formed
[LnM(EX)] are also expected to be extremely labile, due to the
buildup of positive charge at the group 13 element, E.6–8
Superficially, this reflects not only the inherent disparity in
electronegativities between group 13 and group 17 elements9but
also the removal of further electron density from E on
coordination to a transition metal. With this in mind, and
given the relative electronegativities of the group 17 elements,
iodine perhaps represents the optimal choice of halogen
substituent, X.
From a synthetic perspective, a further obstacle is the lack
of readily available sources of the diatomic EX molecule. Thus,
while N2 and CO are stable diatomic gases and their introduction
into metal coordination spheres has been achieved by a variety
of methods,10 BX (X ) F, Cl), AlX (X ) F, Cl, Br, I), and
GaX (X ) Cl, Br, I) are known (as donor-free species) only
under conditions of extreme temperature, with problems stem-
ming from disproportionation or aggregation inherent at (or close
to) room temperature.7a,11-14 Synthetic methodologies based on
coordination of the “free” EX molecule therefore seem likely
to flounder, and we have sought to exploit an alternative
methodology, that is abstraction of a halide anion (X-) from a
pre-existing metal complex of the type LnM(EX2).15-19 Such
an approach necessarily generates a cationic complex of the
type [LnM(EX)]+, thereby imposing an electrostatic barrier to
oligomerization processes such as those thought to be important
for charge-neutral analogues such as [(η5-C5H4Me)Mn
(CO)2BCl].7b That said, such cationic derivatives are also likely
to be highly electrophilic, and the steric/electronic properties
of peripheral substituents will therefore play key roles in
determining thermodynamic stability and/or kinetic lability.
Thus, for example, in the case of borylene complexes
[LnM(BR)]+, electronic structure has been shown to be strongly
dependent on the π-donor properties of the boron-bound
substituent, e.g. Fischer carbene-like MdB double bonds for
B(σ-aryl), simple M r B donor/acceptor interactions for B(η5-
Cp*), and intermediate behavior for B(NR2).15a,e,17 However,
if simple halogen-substituted derivatives, [LnM(EX)]+, are
sought, opportunities for variation in sterics/electronics lie
primarily with the metal/ligand fragment. We have therefore
targeted bulky, electron-rich systems LnM, which are known
to bind the diatomics CO and N2 (e.g., [(η5-C5R5)M(PR3)2]+,
where M ) a group 8 metal)20 and which are likely to offer
both steric and electronic protection of the coordinated EX
ligand.21 By utilizing this approach, we have been able to target
cationic compounds containing gallylene ligands {[LnM(GaR)]+;
R ) aryl, amido, halide}, including the successful isolation of
[Cp*Fe(dppe)(GaI)]+[BArf4]- [Arf ) C6H3(CF3)2-3,5]. This
complex features a terminally bound GaI ligand and, therefore,
represents the first experimental realization of a valence iso-
electronic group 13/group 17 analogue of CO and N2.16 Given
the heated debate concerning models of metal-metal bonding
in gallium-containing systems22 and, for example, the description
of superficially similar complexes as being bound via multiple
bonds (e.g., LnMtER) or via donor/acceptor interactions
(LnM r ER),23 we perceived such complexes as ideal platforms
on which to test bonding models by quantum chemical methods.
In addition, we have also sought to compare ligand properties with
related ten valence electron systems such as N2, CO, and BF.
(6) (a) Ehlers, A. W.; Baerends, E. J.; Bickelhaupt, F. M.; Radius, U.
Chem.sEur. J. 1998, 4, 210. (b) Radius, U.; Bickelhaupt, F. M.;
Ehlers, A. W.; Goldberg, N.; Hoffmann, R. Inorg. Chem. 1998, 37,
1080. (c) Frenking, G.; Fro¨hlich, N. Chem. ReV. 2000, 100, 717.
(7) Complexes containing a terminally bound EX ligand have been
postulated as labile intermediates. See, for example: (a) Timms, P. L.
Acc. Chem. Res. 1973, 6, 118. (b) Braunschweig, H.; Colling, M.;
Hu, C.; Radacki, K. Angew. Chem., Int. Ed. 2002, 41, 1359.
(8) Examples of three-coordinate bridging EX ligands simultaneously
bonded to two metal centers are known. See, for example: (a) Bunn,
N. R.; Aldridge, S.; Kays (ne´e Coombs) , D. L.; Coombs, N. D.; Day,
J. K.; Ooi, L.-L.; Coles, S. J.; Hursthouse, M. B. Organometallics
2005, 24, 5879. (b) Bissinger, P.; Braunschweig, H.; Seeler, F.
Organometallics 2007, 26, 4700.
Experimental Section
(i) General Considerations. All manipulations were carried out
under a nitrogen or argon atmosphere using standard Schlenk line
(16) For a preliminary account of part of this work, see: (a) Coombs, N. D.;
Clegg, W.; Thompson, A. L.; Willock, D. J.; Aldridge, S. J. Am. Chem.
Soc. 2008, 130, 5449. (b) Himmel, H.-J.; Linti, G. Angew. Chem.,
Int. Ed. 2008, 47, 6326 (highlight article)
(9) Emsley, J. The Elements; Oxford University Press: Oxford, 1991.
(10) See, for example: (a) Collman, J. P.; Hegedus, L. S.; Norton, J. R.;
Finke, R. G. Principles and Applications of Organotransition Metal
Chemistry; University Science Books: Sausalito, CA, 1987.
(17) (a) Braunschweig, H.; Radacki, K.; Uttinger, K. Angew. Chem., Int.
Ed. 2007, 46, 3979. (b) Braunschweig, H.; Kraft, K.; Kupfer, F.;
Radacki, K.; Seeler, F. Angew. Chem., Int. Ed. 2008, 47, 4931.
(18) Vidovic, D.; Findlater, M.; Reeske, G.; Cowley, A. H. Chem. Commun.
2006, 3786.
(11) (a) Timms, P. L. J. Am. Chem. Soc. 1967, 89, 1629. (b) Timms, P. L.
J. Am. Chem. Soc. 1968, 90, 4585.
(12) (a) Dohmeier, C.; Loos, D.; Schno¨ckel, H. Angew. Chem., Int. Ed.
1996, 35, 129. (b) Schnepf, A.; Schno¨ckel, H. Angew. Chem., Int.
Ed. 2002, 41, 3532. (c) Koch, K.; Burgert, R.; Schno¨ckel, H. Angew.
Chem., Int. Ed. 2007, 46, 5795.
(19) For a related approach utilizing protonation of a gallium-bound
hydrocarbyl substituent to generate a cationic complex, see: (a)
Cadenbach, T.; Gemel, C.; Zacher, D.; Fischer, R. A. Angew. Chem.,
Int. Ed. 2008, 120, 3487.
(13) With regard to gallium(I) iodide, the phosphine donor-stabilized cluster
Ga8I8(PEt3)6 has been reported: (a) Doriat, C.; Friesen, M.; Baum, E.;
Ecker, A.; Schno¨ckel, H. Angew. Chem., Int. Ed. Engl. 1997, 36, 1969.
In addition, a sparingly soluble reagent of uncertain composition
formulated as “GaI” has also been described. (b) Green, M. L. H.;
Mountford, P.; Smout, G. J.; Speel, S. R. Polyhedron 1990, 9, 2763.
(c) Baker, R. J.; Jones, C. Dalton Trans. 2005, 1341.
(20) See, for example:(a) Glaser, P. B.; Wanandi, P. W.; Tilley, T. D.
Organometallics 2004, 23, 693. (b) Paul, F.; Toupet, L.; Roisnel, T.;
Hamon, P.; Lapinte, C. C. R. Chimie 2005, 8, 1174.
(21) (a) Ueno, K.; Watanabe, T.; Tobita, H.; Ogino, H. Organometallics
2003, 22, 4375. (b) Muraoka, T.; Motohashi, H.; Hirotsu, M.; Ueno,
K. Organometallics 2008, 27, 3918.
(14) For a recent review encompassing aspects of group 13 sub-halides
isolated in inert gas matrices, see, for example:(a) Himmel, H.-J. Z.
Anorg. Allg. Chem. 2005, 631, 1551, and references therein.
(15) See, for example: (a) Coombs, D. L.; Aldridge, S.; Jones, C.; Willock,
D. J. J. Am. Chem. Soc. 2003, 125, 6356. (b) Bunn, N. R.; Aldridge,
S.; Coombs, D. L.; Rossin, A.; Willock, D. J.; Jones, C.; Ooi, L.-L.
Chem. Commun. 2004, 1732. (c) Kays, D. L.; Aldridge, S.; Day, J. K.;
Ooi, L.-L. Angew. Chem., Int. Ed. 2005, 44, 7457. (d) Bunn, N. R.;
Aldridge, S.; Kays, D. L.; Coombs, N. D.; Rossin, A.; Willock, D. J.;
Jones, C.; Ooi, L.-L. Organometallics 2005, 24, 5891. (e) Aldridge,
S.; Jones, C.; Gans-Eichler, T.; Stasch, A.; Kays (ne´e Coombs) , D. L.;
Coombs, N. D.; Willock, D. J. Angew. Chem., Int. Ed 2006, 45, 6118.
(22) See, for example: (a) Su, J.; Li, X.-W.; Crittendon, R. C.; Campana,
C. F.; Robinson, G. H. Organometallics 1997, 16, 4511. (b) Cotton,
F. A.; Feng, X. Organometallics 1998, 17, 128.
(23) For recent reviews of the coordination chemistry of diyl ligands, see,
for example: (a) Fischer, R. A.; Weiꢀ, J. Angew. Chem., Int. Ed. 1999,
38, 2830. (b) Linti, G.; Schno¨ckel, H. Coord. Chem. ReV. 2000, 206-
207, 285. (c) Cowley, A. H. Chem. Commun. 2004, 2369. (d) Gemel,
C.; Steinke, T.; Cokoja, M.; Kempter, A.; Fischer, R. A. Eur. J. Inorg.
Chem. 2004, 4161. (e) Aldridge, S.; Coombs, D. L. Coord. Chem.
ReV. 2004, 248, 535. (f) Braunschweig, H.; Rais, D.; Kollann, C.
Angew. Chem., Int. Ed. 2006, 45, 5254.
(24) Barry, S. T.; Richeson, D. S. Chem. Mater. 1994, 6, 2220.
9
16112 J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008

Cationic Terminal Gallylene Complexes
A R T I C L E S
or drybox techniques. Solvents were predried over sodium wire
(hexanes, toluene, diethyl ether) or molecular sieves (dichlo-
romethane, fluorobenzene) and purged with nitrogen prior to
distillation from the appropriate drying agent (hexanes, potassium;
toluene or diethyl ether, sodium; dichloromethane or fluorobenzene,
CaH2). d6-Benzene, d2-dichloromethane, and d5-fluorobenzene (all
Goss) were degassed and dried over the appropriate drying agent
(potassium or molecular sieves) prior to use. MesLi (Mes ) 2,4,6-
Me3C6H2), (Me3Si)2NGaCl2 ·THF,24 Na[BArf4] [Arf ) C6H3(CF3)2-
3,5],25 Na[(η5-C5R5)Fe(CO)2] (R ) H, Me),26 and [Cp*Fe(CO)2GaI2]2
(1)8a were synthesized by literature routes; the synthetic procedure for
[Cp*Ru(CO)2GaI2]2 (13) and details of its reactivity toward dppe are
included in the Supporting Information. Reagents dppe, carbon
monoxide, 4,4′-di-tert-butyl-2,2′-bipyridyl (dtbpy) and [nBu4N]I were
used as received.
9, 11, and 13; the remaining structures (4, 5, 10, 14, 15, and 17)
were solved with SIR-92,28e and refined using CRYSTALS.28f
Intensity data were processed and corrected for absorption effects
by the multiscan method, based on multiple scans of identical and
Laue equivalent reflections using the appropriate software
(DENZO,28a SADABS,28c or SORTAV28g). In general, coordinates
and anisotropic displacement parameters of all non-hydrogen atoms
were refined except where this was not possible due to the presence
of disorder. In most cases, the hydrogen atoms were visible in the
difference map, but in general, they were positioned geometrically
(then optimized by refinement with restraints for compounds 4, 5,
10, 14, 15, and 17) and refined using a riding model. The structure
of compound 13 (isomorphous with previously reported 1)8a and
those of 7, 10, 11, 13-15, and 17, which were obtained merely
for confirmation of composition, are included only in the Supporting
Information. For the remaining compounds (2-5 and 9), the details
of each data collection, structure solution, and refinement can be
found in Table 1. Relevant bond lengths and angles are included
in the figure captions, and complete details of each structure have
been deposited with the CCDC (numbers as listed in Table 1). In
addition, complete details for each structure (including CIF files)
have been included in the Supporting Information. In the case of
compound 9, the structural model has the iodine atom disordered
over three sites, with the two main sites being occupied in the ratio
3.8:1. The major site gives an almost linear Fe-Ga-I unit (ca.
171°), while the minor site is rather more bent (ca. 149°) and
features contacts between I′ and C(55)-C(58) of one aromatic ring
of the [BArf4]- anion which are within the sum of the van der
Waals radii for iodine and carbon. There is also a very small
contribution (1.85%) of the neutral starting material, substituted
on the cation site, with a corresponding slight deficiency on the
anion site. Details of minor disorder in other structures are included
in the Supporting Information.
(iii) Computational Method. The geometries of the model
complexes [CpFe(dmpe)(EX)]+ (EX ) GaI, BF, CO, N2) were
optimized using ADF (BLYP/TZP) using widely precedented and
previously described methods.29 In order to investigate quantita-
tively the ability of the Fe-Ga-I unit in complex 9 to adopt
nonlinear geometries, a series of further calculations was carried
out using the B3LYP functional within the Gaussian03 package
on a more realistic model system in which the dppe ligand was
reduced to Me(Ph)PCH2CH2P(Ph)Me with the Ph groups proximal
to the GaI ligand being retained; the methyl groups of the Cp*
ligand were also included in the calculation. Calculations of σ and
π contributions to bonding densities were then carried out as
reported previously for analogous investigations of transition metal
diyl and boryl complexes.29 To carry out the desired energy
decomposition analyses, the fully optimized structure was split into
two fragments: [CpFe(dmpe)]+ and the EX ligand. For each
fragment, molecular orbitals were calculated using a TZP basis set
in the geometry of the complete complex and these were then used
as the basis in a further SCF computation on the full structure.
This approach allows ADF to give a breakdown of the interaction
between [CpFe(dmpe)]+ and each of the diatomic ligands. An
alternative breakdown of the fragment interaction energy into
contributions from the Hamiltonian potentials was also carried out
(see Supporting Information).
NMR spectra were measured on Bruker AM-400, Jeol 300
Eclipse Plus, or Varian Mercury 300 FT-NMR spectrometers.
Residual signals of solvent were used for reference for 1H and 13
C
NMR spectroscopy. 11B, 19F, and 31P NMR spectra were referenced
with respect to Et2O·BF3, CFCl3, and 85% aqueous H3PO4,
respectively. Infrared spectra were measured for each compound
either pressed into a disk with excess dry KBr or as a solution in
the appropriate solvent, on a Nicolet 500 FT-IR spectrometer. Mass
spectra were measured on a Bruker MicroTOF-Q instrument with
direct sampling from a Braun LabMaster inert atmosphere box at
the University of Bath (compound 9)27 or by the EPSRC National
Mass Spectrometry Service Center, at the University of Wales
Swansea. Perfluorotributylamine was used as the standard for high-
resolution EI mass spectra. Elemental microanalysis data is reported
for all compounds, except in cases were either extreme air and
moisture sensitivity or the presence of volatile solvent within the
crystal lattice precluded reliable reproducible measurements. For
these compounds (2, 4, and 9), characterization is therefore based
upon multinuclear NMR, IR, and mass spectrometry data (including
accurate mass measurement), supplemented by single crystal X-ray
diffraction studies. In all cases, the purity of the bulk material was
established by multinuclear NMR to be >95%. Photolytic experi-
ments were carried out using a Spectral Energy mercury arc lamp
(1 kW) with samples contained within quartz Schlenk vessels.
Spectroscopic abbreviations: m ) multiplet, s ) singlet, d )
doublet, t ) triplet, sept ) septet, st ) strong.
(ii) Crystallographic Method. Low temperature single crystal
X-ray diffraction data for compounds 2-5, 7, 10, 13-15, and 17
were collected on an Nonius KappaCCD diffractometer; those for
9 and 11 were collected at the Daresbury Synchrotron Radiation
Source on station 9.8. For compounds 2-5, 7, 10, 13-15, and 17,
data collection and cell refinement were carried out using DENZO
and COLLECT,28a,b while, for 9 and 11, the Bruker APEX2
software (including SAINT) was used.28c Structure solution and
refinement were carried out with the SHELX software suite28c for
compounds 2 and 3 (solved using DIRDIF-99),28d together with 7,
(25) Reger, D. L.; Wright, T. D.; Little, C. A.; Lamba, J. J. S.; Smith,
M. D. Inorg. Chem. 2001, 40, 3810.
(26) (a) King, R. B.; Bisnette, M. B. J. Organomet. Chem. 1967, 8, 287.
(b) King, R. B. Acc. Chem. Res. 1970, 3, 417.
(27) For a full description of the experimental approach used, see: (a)
Lubben, A. T.; McIndoe, J. S.; Weller, A. S. Organometallics 2008,
27, 3303.
(iv) Syntheses. Cp*Fe(dppe)GaI2 (3). To a solution of
[Cp*Fe(CO)2GaI2]2 (1; 1.414 g, 1.240 mmol) in toluene (100 cm3)
was added a solution of dppe (1.110 g, 2.785 mmol) also in toluene
(200 cm3), and the reaction mixture was photolyzed for 86 h. The
resulting dark orange solution was filtered and concentrated in
Vacuo, and red crystals of 3 suitable for X-ray diffraction were
obtained from a concentrated toluene solution at -30 °C. Isolated
(28) (a) Denzo: Otwinowski, Z.; Minor, W. In Methods Enzymology; Carter,
C. W., Sweet, R. M., Eds.; Academic Press: New York, 1996; Vol.
276, p 307. (b) Collect: Nonius B.V.: Delft, The Netherlands, 1997-
1002. (c) Sheldrick, G. M. Acta Crystallogr., Sect. A 2008, 64, 112.
(d) DIRDIF-99: Beurskens, P. T.; Admiraal, G.; Beurskens, G.;
Bosman, W. P.; de Gelder, R.; Israel, R.; Smits, J. M. M. The DIRDIF-
99 Program System, Technical Report of the Crystallography Labora-
tory; University of Nijmegen: Nijmegen, The Netherlands, 1999. (e)
Sir-92: Altomare, A.; Cascarano, G.; Giacovazzo, C.; Guagliardi, A.;
Burla, M. C.; Polidori, G.; Camalli, M. J. Appl. Crystallogr. 1994,
27, 435. (f) CRYSTALS: Betteridge, P. W.; Carruthers, J. R.; Cooper,
R. I.; Prout, J.; Watkin, D. J. J. Appl. Crystallogr. 2003, 36, 1487. (g)
Sortav: Blessing, R. H. Acta Crystallogr., Sect. A 1995, 51, 33.
1
yield 0.789 g, 35%. H NMR (C6D6, 300 MHz): δH 1.33 (s, 15H,
(29) For previous DFT studies of related systems, see, for example: (a)
Dickinson, A. A.; Willock, D. J.; Calder, R. J.; Aldridge, S.
Organometallics 2002, 21, 1146. (b) Aldridge, S.; Rossin, A.; Coombs,
D. L.; Willock, D. J. Dalton Trans. 2004, 2649.
9
J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008 16113

A R T I C L E S
Coombs et al.
9
16114 J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008

Cationic Terminal Gallylene Complexes
A R T I C L E S
CH3 of Cp*), 2.25 (m, 2H, CH2), 3.25 (m, 2H, CH2), 6.95 (m, 8H,
CH aromatic), 7.15 (m, 8H, CH aromatic), 7.69 (m, 4H, CH
aromatic). 13C NMR (C6D6, 75 MHz): δC 9.4 (CH3 of Cp*), 31.6
(dd, 1JPC ) 15, 2JPC ) 8 Hz, CH2 of dppe), 85.2 (quaternary carbon
of Cp*), 127.2 (aromatic CH), 127.3 (aromatic CH), 127.7 (aromatic
CH), 129.3 (aromatic CH), 131.6 (aromatic CH), 131.8 (aromatic
CH), 141.3 (d, 1JPC ) 31 Hz, ipso carbon of dppe), 141.7 (d, 1JPC
) 31 Hz, ipso carbon of dppe). 31P NMR (C6D6, 122 MHz): δP
103.9. EI-MS, m/z: 911.9 {1%, [Cp*Fe(dppe)GaI2]+}, 716.1 {22%,
[Cp*Fe(dppe)I]+}, 589.2 {11%, [Cp*Fe(dppe)]+}, 398.2 {100%,
[dppe]+}. Exact mass: calc for [Cp*Fe(dppe)GaI2]+ (i.e., M+)
911.9216, meas. 911.9217. Elemental microanalysis. Calc for
3·5/4(C7H8): C, 52.27; H, 4.80. Meas.: C, 51.91; H, 4.35.
manner using diethyl ether as the reaction solvent, albeit in lower
yield (ca. 15%).
Cp*Fe(CO)2Ga(Mes)I (5). To a solution of 1 (0.303 g, 0.266
mmol) in toluene (50 cm3) was added a solution/suspension of
MesLi (0.075 g, 0.595 mmol) also in toluene (40 cm3), and the
reaction mixture was stirred at 20 °C for 12 h. After filtration and
removal of volatiles in Vacuo, the resulting orange oil was extracted
into diethyl ether (40 cm3), concentrated (to ca. 15 cm3), and cooled
to -30 °C, yielding 5 as a pale yellow crystalline solid. Isolated
yield 0.078 g, 26%. Crystals suitable for X-ray diffraction were
obtained from a concentrated diethyl ether solution at -30 °C. 1H
NMR (CD2Cl2, 300 MHz): δH 1.13 (s, 15H, CH3 of Cp*), 1.88 (s,
3H, para-Me of Mes), 2.09 (s, 6H, ortho-Me of Mes), 6.39 (s, 2H,
CH of Mes). 13C NMR (CD2Cl2, 75 MHz): δC 10.0 (CH3 of Cp*),
21.5 (para-CH3 of Mes), 22.7 (ortho-CH3 of Mes), 95.2 (quaternary
carbon of Cp*), 138.6 (ortho-quaternary carbon of Mes), 139.8
(meta-CH of Mes), 155.6 (para-quaternary carbon of Mes), 215.9
(CO), ipso-quaternary carbon of Mes not observed. IR (CD2Cl2,
cm-1): ν(CO) 1984 (st), 1936 (st). EI-MS, m/z: 562.0 {weak,
[Cp*Fe(CO)2Ga(Mes)I]+}, 534.0 {69%, [Cp*Fe(CO)Ga(Mes)I]+},
506.0 {14%, [Cp*FeGa(Mes)I]+}, 435.1 {9%, [Cp*Fe(CO)2Ga-
(Mes)]+}, 310.1 {100%, [(OC)2FeGaI]+}. Exact mass: calc. for
[Cp*Fe(CO)2Ga(Mes)I]+ (i.e., M+) 561.9577, meas. 561.9581; calc.
for [Cp*Fe(CO)Ga(Mes)I]+ [i.e., (M - CO)+] 533.9628, meas.
533.9630. Elemental microanalysis. Calc. for 5: C, 44.81; H, 4.66.
Meas.: C, 49.22; H, 4.90.
Reaction of [Cp*Fe(CO)2GaI2]2 (1) with dppe: Isolation of
the Intermediate Cp*Fe(CO)(µ2-dppe)GaI2 (2). To a solution of
1 (1.255 g, 1.100 mmol of dimer) in toluene (200 cm3) was added
a solution of dppe (0.986 g, 2.474 mmol) also in toluene (200 cm3),
and the reaction mixture was photolyzed for 70 h with periodic
31P NMR monitoring. The resulting solution was filtered and
concentrated in vacuo (to ca. 15 cm3), and single crystals suitable
for X-ray diffraction were obtained by layering a concentrated
toluene solution with hexanes. Isolated yield 0.835 g, 40%. 1H NMR
(CD2Cl2, 300 MHz): δH 1.48 (s, 15H, CH3 of Cp*), 1.95 (m, 2H,
CH2 of dppe), 2.55 (m, 2H, CH2 of dppe), 7.10-7.49 (m, 16H,
ortho- and meta-CHs of dppe), 7.68, 7.88 (m, each 2H, para-CH
of dppe). 13C NMR (CD2Cl2, 75 MHz): δC 8.7 (CH3 of Cp*), 28.1
1
1
[CpFe(CO)2]2GaN(SiMe3)2 (6) and [Cp*Fe(CO)2]2GaN(SiMe3)2
(7). The two compounds were prepared in an analogous fashion,
illustrated for 6. To a slurry of Na[CpFe(CO)2] (0.442 g, 2.021
mmol) in diethyl ether was added a solution of (Me3Si)2-
NGaCl2 ·THF (0.377 g, 1.011 mmol) also in diethyl ether (30 cm3).
The reaction mixture was stirred at 20 °C for 2 h, after which
volatiles were removed in Vacuo. The resulting yellow oil was
dissolved in hexanes, filtered, concentrated (to ca. 20 cm3), and
cooled to -30 °C, yielding 6 as a pale yellow microcrystalline solid.
Isolated yield: 0.421 g, 71%. 1H NMR (C6D6, 300 MHz): δH 0.33
(s, 18H, CH3 of SiMe3), 4.21 (s, 10H, CH of Cp). 13C NMR (C6D6,
75 MHz): δC 4.2 (CH3 of SiMe3), 82.6 (Cp), 215.9 (CO). IR
(CD2Cl2, cm-1): ν(CO) 1993 (st), 1972 (st), 1926 (st). EI-MS, m/z:
555.0 {23%, [{CpFe(CO)}{CpFe(CO)2}GaN(SiMe3)2]+}, 527.0 {4%,
[{CpFe(CO)}2GaN(SiMe3)2]+}, 471.0 {17%, [(CpFe)2GaN(SiMe3)2]+},
406.0 {100%, [CpFe(CO)2GaN(SiMe3)2]+}. Exact mass: calc. for
[{CpFe(CO)}{CpFe(CO)2}GaN(SiMe3)2]+ [i.e., (M - CO)+]
554.9557, meas. 554.9554. Data for 7: crystals suitable for X-ray
diffraction obtained from a concentrated diethyl solution at -30
(d JPC ) 15 Hz, CH2 of dppe), 28.3 (d JPC ) 15 Hz, CH2 of
dppe), 91.0 (quaternary carbon of Cp*), 124.5 (aromatic CH), 127.3
(aromatic CH), 128.0 (aromatic CH), 128.4 (aromatic CH), 129.5
(aromatic CH), 133.2 (aromatic CH), 134.0 (d 1JPC ) 31 Hz, ipso
carbon of dppe), 134.2 (d 1JPC ) 31 Hz, ipso carbon of dppe), CO
carbon not observed. 31P NMR (CD2Cl2, 122 MHz): δP 65.0
(Fe-P), -41.0 (Ga-P). IR (CD2Cl2, cm-1): ν(CO) 1981 (st). EI-
MS, m/z: 939.7 {4%, [Cp*Fe(dppe)(CO)GaI2]+}, 541.8 {19%,
[Cp*Fe(CO)GaI2]+}, 398.2 {67%, [dppe]+}. Exact mass: calc. for
[Cp*Fe(dppe)(CO)GaI2]+ (i.e., M+) 939.9165, meas. 939.9169.
Reproducible microanalyses for crystalline samples of 2 proved
impossible to obtain, possibly due to the presence of toluene within
the crystal lattice.
Cp*Fe(dppe)Ga(Mes)I (4). To a solution/suspension of MesLi
(0.036 g, 0.285 mmol) in toluene (25 cm3) was added a solution
of 3 (0.150 g, 0.164 mmol) also in toluene (15 cm3), and the reaction
mixture was stirred for 16 h at 20 °C. The resulting orange solution
was filtered and concentrated (to ca. 10 cm3), and 4 was obtained
as an orange microcrystalline material on cooling to -30 °C.
Crystals suitable for X-ray diffraction were obtained from a
concentrated diethyl ether solution at -30 °C. Isolated yield 0.114
g, 71%. 1H NMR (C6D6, 300 MHz): δH 1.42 (s, 15H, CH3 of Cp*),
2.05 (m, 2H, CH2 of dppe), 2.10 (s, 6H, ortho-CH3 of Mes), 2.25
(s, 3H, para-CH3 of Mes), 3.95 (m, 2H, CH2 of dppe), 6.78 (s, 2H,
CH of Mes), 7.02 (m, 4H, aromatic CH of dppe), 7.14-7.26 (m,
16H, aromatic ortho- and meta-CH of dppe), 7.80 (m, 4H, aromatic
para-CH of dppe). 13C NMR (CD2Cl2, 75 MHz): δC 10.3 (CH3 of
Cp*), 20.9 (ortho-CH3 of Mes), 23.5 (para-CH3 of Mes), 30.8 (dd
1JPC ) 15.2, 2JPC ) 7.6 Hz, CH2 of dppe), 86.4 (quaternary carbon
of Cp*), 127.2 (aromatic CH of dppe), 128.0 (aromatic CH of dppe),
128.9 (aromatic CH of dppe), 132.5 (aromatic CH of dppe), 133.5
(aromatic CH of dppe), 136.3 (aromatic CH of dppe), 139.3 (ortho-
quaternary carbon of Mes), 139.4 (meta-CH of Mes), 141.0 (d 1JPC
1
°C. Isolated yield 0.543 g, 44%. H NMR (C6D6, 300 MHz): δH
0.52 (s, 18H, CH3 of SiMe3), 1.57 (s, 30H, CH3 of Cp*). 13C NMR
(C6D6, 75 MHz): δC 3.9 (CH3 of SiMe3), 9.2 (CH3 of Cp*), 93.0
(quaternary carbon of Cp*), carbonyl carbon not observed. IR
(CD2Cl2, cm-1): ν(CO) 1973 (st), 1955 (st), 1908 (st). EI-MS, m/z:
695.1 {2%, [{Cp*Fe(CO)}{Cp*Fe(CO)2}GaN(SiMe3)2]+}, 476.1
{41%, [Cp*Fe(CO)2GaN(SiMe3)2]+}, 420.1 {24%, [Cp*FeGaN-
(SiMe3)2]+}, 317.9 {31%, [Cp*Fe(CO)2Ga]+}. Exact mass: calc.
for [{Cp*Fe(CO)}{Cp*Fe(CO)2}GaN(SiMe3)2]+ [i.e., (M - CO)+]
695.1122, meas. 695.1124. Elemental microanalysis. Calc. for 7:
C, 49.75; H, 6.68. Meas.: C, 50.11; H, 6.77.
[Cp*Fe(CO)2Ga(Mes)(dtbpy)]+[BArf4]- (8). A solution of 5
(0.094 g, 0.167 mmol) and dtbpy (0.045 g, 0.168 mmol) in CD2Cl2
(3 cm3) was added to a slurry of Na[BArf4] (0.147 g, 0.166 mmol)
in CD2Cl2 (1 cm3) at -78 °C, and the reaction mixture was warmed
to 20 °C. After sonication for 30 min, the resulting orange solution
was layered with hexanes, yielding single crystals suitable for X-ray
1
) 31.1 Hz, ipso-carbon of dppe), 141.5 (d JPC ) 31.1 Hz, ipso-
carbon of dppe), 142.5 (para-quaternary carbon of Mes). 31P NMR
(C6D6, 122 MHz): δP 99.0. EI-MS, m/z: 904 {1%, [Cp*Fe(dppe)-
Ga(Mes)I]+}, 777.1 {1%, [Cp*Fe(dppe)Ga(Mes)]+}, 716.0 {4%,
[Cp*Fe(dppe)I]+}, 589.2 {9%, [Cp*Fe(dppe)]+}, 398.1 {49%,
[dppe]+}, 262.1 {42%, [Cp*FeGa]+}. Exact mass: calc. for
[Cp*Fe(dppe)Ga(Mes)I]+ (i.e., M+) 904.1032, meas. 904.1028.
Reproducible microanalyses for crystalline samples of 4 proved
impossible to obtain, possibly due to the presence of diethyl ether
within the crystal lattice. 4 can also be prepared in an analogous
1
diffraction. Isolated yield 0.080 g, 31%. H (300 MHz, CD2Cl2):
δH 1.34 (s, 18H, tBu of dtbpy), 1.76 (s, 15H, CH3 of Cp*), 2.07 (s,
3H, para-CH3 of Mes), 2.30 (s, 6H, ortho-CH3 of Mes), 6.66 (s,
2H, CH of Mes), 7.46 (s, 4H, para-CH of BArf4-), 7.62 (s, 8H,
ortho-CH of BArf4-), 7.72, 8.16, 8.75 (m, each 2H, CH of dtbpy).
13C (75 MHz, CD2Cl2): δC 10.2 (Me of Cp*), 20.3 (para-CH3 of
Mes), 26.2 (ortho-CH3 of Mes), 29.8 (tBuCH3 of dtbpy), 36.1 (tBu
9
J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008 16115

A R T I C L E S
Coombs et al.
Scheme 1. Contrasting Halide Abstraction Products from Three-Coordinate Boryl, Gallyl, and Indyl Ligand Systemsa
t
a Conditions: (i) Na[BArf4], dichloromethane, E ) B; R ) Me; (ii) Na[BArf4], dichloromethane, E ) Ga, In; R ) Bu.
quaternary carbon of dtbpy), 95.5 (quaternary carbon of Cp*), 117.3
of 9 proved impossible to obtain, possibly due to the extreme
sensitivity of this compound to air and moisture.
(CH of dtbpy), 117.4 (para-CH of BArf4-), 119.5 (CH of dtbpy),
Reaction of 9 with CO: Formation of [Cp*Fe(dppe)(CO)]+
[BArf4]- (12). A solution of 9 (0.050 g, 0.030 mmol) in d5-
fluorobenzene (2 cm3) was exposed to CO gas (1 atm) at 20 °C for
a period of 12 h, leading to a violet to yellow color change and to
the formation of a gray precipitate. At this point, monitoring by
31P NMR revealed quantitative conversion of 9 to a single
phosphorus-containing species giving rise to a sharp signal at δP
1
122.7 (q JCF ) 273 Hz, CF3 of BArf4-), 124.7 (CH of dtbpy),
126.7 (quaternary carbon of dtbpy), 128.6 (q 2JCF ) 29 Hz, meta-
carbon of BArf4-), 134.8 (ortho-CH of BArf4-), 138.5 (ortho-
quaternary carbon of Mes), 142.8 (meta-CH of Mes), 147.8
(quaternary carbon of dtbpy), 154.3 (para-quaternary carbon of
1
Mes), 161.2 (q JCB ) 49 Hz, ipso-carbon of BArf4-), 168.2
(quaternary carbon of dtbpy), 217.2 (CO), ipso-carbon of Mes not
observed. 19F (283 MHz, CD2Cl2): δF -62.9. 11B (96 MHz,
CD2Cl2): δB -8.4. IR (CD2Cl2, cm-1): ν(CO) 1971 (st), 1919 (st).
ES-MS, m/z: (10 V cone voltage) 703.2 {100%, [Cp*Fe(CO)2-
Ga(Mes)(dtbpy)]+}; (50 V cone voltage) 703.2 {35%, [Cp*Fe-
(CO)2Ga(Mes)(dtbpy)]+}, 435.0 {100%, [Cp*Fe(CO)2Ga(Mes)]+}.
Exact mass: calc. for [Cp*Fe(CO)2Ga(Mes)(dtbpy)]+ 701.2519,
meas. 701.2518. Elemental microanalysis. Calc. for 8: C, 54.40;
H, 3.99; N, 1.79. Meas.: C, 54.54; H, 4.12; N, 1.66.
1
88.0. Further analysis by H and 13C NMR, IR spectroscopy, and
electrospray mass spectrometry, and comparison with literature data
confirmed the presence of the [Cp*Fe(dppe)(CO)]+ cation.15
Reaction of 9 with [nBu4N]I: Formation of 3. To a solution of
3 (0.082 g, 0.050 mmol) in d5-fluorobenzene (2 cm3) was added a
solution of [nBu4N]I (0.022 g, 0.06 mmol) also in d5-fluorobenzene
with immediate formation of an orange solution and quantitative
conversion of 9 to a single phosphorus-containing compound (δP
104.1). The identity of the product 3 was confirmed by comparison
of 1H, 13C, and 31P NMR data with those obtained from an authentic
sample.
[Cp*Fe(dppe)(GaI)]+[BArf4]- (9). To a suspension of Na-
[BArf4] (0.053 g, 0.060 mmol) in fluorobenzene (1 cm3) at -30
°C was added an orange-red solution of 3 (0.049 g, 0.054 mmol)
also in fluorobenzene (2 cm3), and the reaction mixture was warmed
to 20 °C over a period of 20 min. Monitoring of the reaction by
31P{1H} NMR spectroscopy revealed quantitative conversion of 3
(δP 103.9) to a single phosphorus containing species giving rise to
a broad resonance at δP 87.0. The resulting deep violet solution
was filtered and concentrated in Vacuo, and purple crystals of 9
suitable for X-ray diffraction were obtained by layering with
hexanes and storage at -30 °C. Analogous chemistry carried out
in dichloromethane solution leads to the formation of a single
species giving rise to a similar 31P NMR resonance but which has
a half-life of ca. 30 min at 20 °C. Isolated yield 0.031 g, 35%. 1H
NMR (C6D5F, 300 MHz): δH 1.18 (s, 15H, CH3 of η5-C5Me5), 1.88
(m, 2H, CH2 of dppe), 2.04 (m, 2H, CH2 of dppe), 7.14-7.37
(overlapping m, 20H, aromatic CH of dppe), 7.54 (s, 4H, para-CH
of anion), 8.27 (s, 8H, ortho-CH of anion). 13C NMR (C6D5F, 75
MHz): δC 9.2 (CH3 of η5-C5Me5), 31.5 (m, CH2 of dppe), 86.5
(quaternary carbon of η5-C5Me5), 117.0 (sept, J ) 4.1 Hz, para-
CH of anion), 124.3 (1:3:3:1 q, J ) 272.1 Hz, CF3 of anion), 128.0
(pseudo t, J ) 4.6 Hz, meta-CH of dppe), 128.5 (1:3:3:1 q, J )
29.8 Hz, meta-carbon of anion), 129.1 (pseudo t, J ) 5.1 Hz, ortho-
CH of dppe), 130.0 (pseudo t, J ) 4.6 Hz, meta-CH of dppe), 130.4
(para-CH of dppe), 130.8 (para-CH of dppe), 131.8 (pseudo t, J
) 5.1 Hz, ortho-CH of dppe), 134.4 (ortho-CH of anion), 161.9
(1:1:1:1 q, J ) 49.8 Hz, ipso-carbon of anion), ipso-carbons of
dppe not observed. 11B NMR (C6D5F, 96 MHz): δB 1.9. 19F NMR
(C6D5F, 282 MHz): δF -62.7. 31P NMR (C6D5F, 122 MHz): δP
87.0. FT-Raman (C6D5F solution): ν(Ga-I) 186 cm-1. UV/vis
(C6H5F solution): λmax ) 549 nm; ε ) 754 cm-1 mol-1 dm3. ES-
MS (positive ion mode), m/z: 785.0 {100%, [(η5-C5Me5)Fe-
(dppe)(GaI)]+}, correct isotope pattern for C36H39FeGaIP2. Exact
mass: calc. for [(η5-C5Me5)Fe(dppe)(GaI)]+ (i.e., M+) 785.0173,
meas. 785.0204. Reproducible microanalyses for crystalline samples
Results and Discussion
(i) Synthetic and Reactivity Studies. In previous work, we
have shown that halide abstraction represents a valid synthetic
approach to transition metal complexes containing low-
coordinate group 13 ligands.15–19 Extending the methodology
pioneered for boron donors to the heavier elements of group
13, however, presents additional challenges. Thus, the greater
size and metallic character of gallium and indium (cf. boron)
are presumably responsible for the formation of halide-bridged
dinuclear systems of the type [(η5-C5R5)Fe(CO)2E(Mes*)-
(µ-X)(Mes*)EFe(CO)2(η5-C5R5)]+ (E ) Ga, In; X ) Cl, Br; R
) H, Me; Mes* ) 2,4,6-tBu3C6H2) by halide abstraction from
(η5-C5R5)Fe(CO)2E(Mes*)X;15d by contrast, the two-coordinate
terminal borylene system [Cp*Fe(CO)2(dBMes)]+ is obtained
from the analogous reaction of Cp*Fe(CO)2B(Mes)Br with
Na[BArf4] (Scheme 1).15a
In the case of cationic borylene complexes, the tendency of
the unsaturated group 13 center to become trigonal by coordina-
tion of an additional two-electron donor can be reduced by
incorporating sterically bulky π electron-donating substituents
at boron (e.g., amino groups).15e While such an approach is also
conceivable for the synthesis of analogous two-coordinate
gallium-based ligand systems, an attractive alternative, which
has achieved notable success in related group 14 chemistry, is
the exploitation of sterically encumbered electron-rich metal/
ligand fragments, LnM. Synthetic studies designed to assess the
relative merits of these two approaches are outlined below.
The dimeric species [Cp*Fe(CO)2GaI2]2 (1) proves to be a
versatile starting material for the exploration of cationic
9
16116 J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008

Cationic Terminal Gallylene Complexes
A R T I C L E S
Scheme 2. Syntheses of Precursor Complexes 3-5 from 1 via
to be unnecessary), with attention focusing instead on manipula-
tion of the steric/electronic properties of the metal fragment,
LnM.
Gallium- and Iron-Centered Substitution Chemistriesa
Electron-rich group 8 metal fragments containing bulky
cyclopentadienyl and tertiary phosphine ligands have been
shown to be compatible with highly electrophilic ligands such
as base-free dialkyl silylenes.32 On this basis, we targeted similar
metal/ligand combinations to provide the steric and electronic
protection necessary to isolate a cationic terminal gallylene
complex. Furthermore, we identified photolytic carbonyl sub-
stitution chemistry at the gallyl precursor stage as a potential
route to the desired metal/ligand combinations.33 In the event,
and somewhat counterintuitively, dppe [1,2-bis(diphenylphos-
phino)ethane] proves to be an ideal phosphine ancillary ligand
(on the basis of optimal steric shielding of the gallium fragment
by the pendant phenyl rings, Vide infra) and iron (rather than
ruthenium) the metal of choice.34 Reaction of the readily
available dicarbonyl substituted precursor 1 with dppe in toluene
can be shown by a combination of 31P NMR and X-ray
crystallography to proceed via a monocarbonyl intermediate (2)
in which the dppe ligand bridges in an intramolecular fashion
between the iron and gallium centers (Scheme 2). Particularly
diagnostic are the two 31P resonances measured for 2; that at
δP -41 is similar to other gallyl/dppe Lewis acid/base com-
plexes {cf. δP -28.0 for [(η7-C7H7)Mo(CO)2GaI2]2(µ-dppe)},33
while that at δP 65 is consistent with dppe bound to iron in a
nonchelating fashion [cf. δP 70.3 and 75.9 for CpFe(CO)-
(COMe)(µ-dppe)Fe(CO)2(η4-C5H4Me)].35 Prolonged photolysis
results in the disappearance of the signal due to 2, with
accompanying growth of the single resonance (δP 104) associ-
ated with the final chelating dppe product.36 Multinuclear NMR,
mass spectrometry, analytical, and X-ray crystallographic data
a Key reagents and conditions: (i) dppe (2.25 equiv per mol of dimeric
1), toluene, 20 °C, UV photolysis for 70 h, 40% isolated yield; (ii) (from
1) dppe (2.25 equiv per mol of dimeric 1), toluene, 20 °C, UV photolysis
for 86 h, 35% isolated yield; (iii) MesLi (1.73 equiv), toluene, 20 °C, 16 h,
77% isolated yield; (iv) MesLi (2.24 equiv per mol of dimeric 1), toluene,
20 °C, 12 h, 26% isolated yield.
gallium-donor complexes (Scheme 2). 1 (and its ruthenium
analogue 13) is available in a one-pot process in multigram
quantities from [Cp*M(CO)2]2 (M ) Fe, Ru), gallium metal,
and iodine, and it is amenable to further functionalization at
either the gallium or iron centers. Controlled functionalization
at gallium to give unsymmetrically substituted gallyl derivatives
such as the (monomeric) iodo(mesityl) complex Cp*Fe(CO)2-
Ga(I)Mes (5) can be achieved in moderate yield, by substitution
of a single iodide substituent with mesityllithium. Attempted
substitution reactions with amido nucleophiles, [NR2]-, using
1 [or Cp*Fe(dppe)GaI2 (3), Vide infra] as the gallium-containing
substrate, however, proved unsuccessful.30 Moreover, alternative
approaches to complexes of the type LnMGa(X)NR2 exploiting
the reactions of dihalo(amino)gallanes, R2NGaX2, with orga-
nometallic nucleophiles, generate instead disubstituted products
of the type [LnM]2GaNR2. Thus, the reactions of [(η5-C5R5)-
Fe(CO)2]- (R ) H, Me) with (Me3Si)2NGaCl2 ·THF lead to
the formation of the bridging aminogallylene complexes [(η5-
C5R5)Fe(CO)2]2GaN(SiMe3)2 (6, R ) H; 7, R ) Me), irrespec-
tive of reaction stoichiometry and conditions.31 Further attempts
to manipulate the substituent at gallium, beyond simple halo or
aryl derivatives, were therefore ignored (and ultimately proved
(32) See, for example: (a) Grumbine, S. K.; Tilley, T. D.; Arnold, F. P.;
Rheingold, A. L. J. Am. Chem. Soc. 1994, 116, 5495. For a discussion
of the use of related electron rich metal/ligand frameworks in
stabilizing dihalocarbene ligands, see, for example: (b) Brothers, P. J.;
Roper, W. R. Chem. ReV. 1988, 88, 1293.
(33) Potential syntheses, such as insertion of gallium(I) reagents into the
metal-halogen bonds of bis(phosphine) metal halides [e.g. of the type
(η5-C5R5)Ru(PR3)2X] have previously been shown to generate alterna-
tive tetrahalogallate products of the type (η5-C5R5)Ru(PR3)2(GaX4)
containing no direct Ru-Ga bonds: (a) Coombs, N. D.; Stasch, A.;
Aldridge, S. Inorg. Chim. Acta 2008, 361, 449.
(34) Although enhanced back-bonding to the putative EX ligand might
identify ruthenium (or osmium) containing systems as more viable
targetssin line with classical bonding modelssthis same factor is
presumably responsible for the much less facile carbonyl ligand
substitution chemistry observed for gallyl precursors containing
ruthenium. The reaction chemistry between dppe and [Cp*Ru-
(CO)2GaI2]2 (13) is consistent with significantly higher barriers to
carbonyl substitution than observed for [Cp*Fe(CO)2GaI2]2 (1). Thus,
although a compound with 31P NMR signals at δP 44 and-40, similar
to those measured for Cp*Fe(CO)(µ2-dppe)GaI2 (2; δP 65 and-41),
is observed in low concentration at short reaction times, Cp*Ru-
(dppe)GaI2 is not identified among the reaction components at any
time. The dppe-bridged dinuclear species [Cp*Ru(CO)2GaI2]2(µ-dppe)
(14) is the only Ru/Ga/P containing product isolated from the toluene
reaction mixture after 60 h, while the formation of appreciable
quantities of Cp*Ru(dppe)I (15) and the known cation [Cp*Ru-
(dppe)(CO)]+ (16) at extended reaction times testifies to the lability
of the gallyl ligand under conditions which are sufficiently forcing to
effect carbonyl substitution chemistry (see Supporting Information for
synthetic,spectroscopic,andcrystallographicdata).[Cp*Ru(dppe)(CO)]+:
(a) Bruce, M.; Ellis, B. G.; Skelton, B. W.; White, A. H. J. Organomet.
Chem. 2005, 690, 792.
(30) Work on related systems has shown that metallated trihalogallates,
[LnMGaX3]-, are among the products of such reactions with bulky
amide reagents, typically isolated as the [R2NH2]+ salts. See, ref 8a
and: Aldridge, S.; Kays, D. L.; Bunn, N. R.; Coombs, N. D.; Ooi,
L.-L. Main Group Met. Chem. 2005, 28, 201.
(35) Luh, L.-S.; Liu, L.-K. Organometallics 1995, 14, 1514.
(36) A similar mechanism has been proposed by Ueno and co-workers for
related dmpe [1,2-bis(dimethylphosphino)ethane] chemistry on the
basis of spectroscopic data, albeit without structural characterization:
(a) Ueno, K.; Hirotsu, M.; Hatori, N. J. Organomet. Chem. 2007, 692,
88.
(31) Such reactivity mirrors analogous chemistry toward the sterically
encumbered dichloroborane (Me3Si)2NBCl2, which generates a bridg-
ing aminoborylene complex, irrespective of reaction conditions:
Braunschweig, H.; Kollann, C.; Englert, U. Eur. J. Inorg. Chem. 1998,
465.
9
J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008 16117

A R T I C L E S
Coombs et al.
Scheme 3. Formation of the Base-Stabilized Cationic Gallylene 8
V) a “flagpole”mass spectrum corresponding to the molecular
ion of the dtbpy adduct [Cp*Fe(CO)2Ga(Mes)(dtbpy)]+is ob-
served, while at higher voltages (ca. 50 V) the predominant
isotopic envelope (centered at m/z 435.0) corresponds to the
base-free cationic gallylene [Cp*Fe(CO)2(GaMes)]+. While
such spectra do not necessarily offer any encouragement for
the isolation of CO-ligated species in macroscopic quantities,
they do at least provide the first spectroscopic evidence for
the cationic gallium-donor analogue of [Cp*Fe(CO)2-
(dBMes)]+ 15a and an indication that the Fe-Ga bond is
retained, at least for the isolated compound in the gas phase.
by Halide Abstraction in the Presence of dtbpy Trapping Agenta
a Key reagents and conditions: (i) Na[BArf4] and dtbpy (1.0 equiv of
each), d2-dichloromethane, 20 °C, 30 min of sonication, 31% isolated yield.
In terms of isolating the desired base-free terminal gallylene
systems in quantities suitable for spectroscopic and structural
characterization, the use of more electron rich and sterically
shielded metal/ligand fragments appears to be crucial. Thus,
although the reaction of Cp*Fe(dppe)Ga(Mes)I (4) with
Na[BArf4] does not lead to the isolation of any tractable Fe-Ga
containing products, reaction of Cp*Fe(dppe)GaI2 (3) with the
same halide abstraction agent generates [Cp*Fe(dppe)-
(GaI)]+[BArf4]- (9) in 35% isolated yield (see Scheme 4). In
the case of 3, the reaction proceeds quantitatively in fluoroben-
are consistent with this species being the expected product
Cp*Fe(dppe)GaI2 (3) isolated from toluene solution in ca. 35%
yield. As with 1, further substitution chemistry at the gallium
center in 3 can be achieved cleanly with MesLi (ca. 2 equiv) in
diethyl ether or with improved yields (ca. 70%) in toluene;
Cp*Fe(dppe)Ga(Mes)I (4) has been characterized by standard
spectroscopic techniques, and its structure in the solid state has
been confirmed by single crystal X-ray diffraction studies.
The success of subsequent halide abstraction chemistry in
delivering tractable terminal gallylene complexes from halogallyl
precursors 3-5 can readily be demonstrated but is strongly
dependent on the nature of the metal- and gallium-bound
substituents, the halide abstraction agent, and the solvent used
(Vide infra).37 Given the appropriate choice of counterion
(typically the weakly coordinating anion [BArf4]-), the nature
of the metal-bound ancillary ligands appears to be the key factor
in determining complex stability. In previous studies, we have
shown that halide abstraction, even from relatively bulky
halogallyl ligands such as -Ga(Mes*)X, invariably leads to the
formation of halide-bridged dinuclear products if partnered with
a transition metal fragment of the type [(η5-C5R5)Fe(CO)2].38
Indeed, in our hands, the isolation of mononuclear cationic
species can only be achieved by carrying out such abstraction
reactions in the presence of a suitable trapping agent, such as a
chelating bipyridyl ligand (Scheme 3). Thus, iodide abstraction
from Cp*Fe(CO)2Ga(Mes)I (5) in the presence of dtbpy leads
to the formation of [Cp*Fe(CO)2Ga(Mes)(dtbpy)]+[BArf4]- (8)
as yellow acicular crystals in ca. 30% isolated yield. The
composition of 8 has been established by standard spectroscopic
and analytical techniques and by comparison with related species
reported by Ueno and Ogino.39 Most informative, though, are
the results obtained by positive-ion electrospray mass spec-
trometry (Figure 1). Thus, at a relatively low cone voltage (10
1
zene solution (as evidenced by 31P NMR monitoring or by H
NMR monitoring in d5-fluorobenzene) to a single phosphorus-
and Cp*-containing compound; the lower isolated yield presum-
ably reflects losses associated with the recrystallization of this
extremely air- and moisture-sensitive material. Crystallization
from a mixture of fluorobenzene and hexanes (ca. 1:10) at -30
°C leads to the isolation of violet crystals of 9, the constitution
of which has been unambiguously established by a combination
of multinuclear (1H, 11B, 13C, 19F, and 31P) NMR, UV/vis, and
FT-Raman spectroscopies, electrospray mass spectrometry
(including exact mass measurement), and single crystal X-ray
diffraction. Particularly informative are the positive ion elec-
trospray mass spectra obtained for 9 in fluorobenzene solution
(see Figure 2). In addition to providing definitive identification
of the molecular ion by isotopic profiling (Figure 2a), fragmen-
tation data reveal the presence of the coordinated GaI ligand.
Thus, MS-MS experiments performed on the molecular ion are
consistent with ready fragmentation generating the [Cp*Fe-
(dppe)]+ cation by loss of GaI (Figure 2b). Facile dissociation
of the GaI ligand from 9 is also consistent with the results of
chemical and computational experiments (Vide infra).
Attempts to isolate 9 from the corresponding reaction in
dichloromethane lead instead to the formation of paramagnetic
products. Monitoring of the reaction mixture by 31P NMR
reveals initial conversion of 3 to a single species which gives
rise to a broad resonance at a very similar chemical shift (δP
87.0) to that observed for the analogous reaction in fluoroben-
zene solution (together with a similar red to violet color change).
In dichloromethane solution this signal has a half-life of ca. 30
min at 20 °C; after several hours the solution is typically brown
in color and exhibits no discernible 31P resonances. Tilley has
previously shown that cationic osmium silylene complexes react
with chlorocarbons via chlorine atom abstraction,40 and we were
successful in isolating and structurally characterizing the 17-
electron [Cp*Fe(dppe)Cl]+ cation by layering of the dichlo-
romethane reaction mixture with hexanes {as the [BArf4]- salt
(11); see the Supporting Information}. In common with related
cationic borylene complexes, 9 also shows evidence of both
gallium- and iron-centered reactivity toward nucleophiles. Thus,
(37) The use of commercially available but more reactive counterions such
as [BPh4]- or [BF4]- typically leads to products containing Fe-Ph, or
E(µ-F)E moieties (E ) Ga, In); the [BArf4]- anion is typically found
to be more weakly interacting, although under appropriate conditions
it too can act as a source of fluoride toward very strongly electrophilic
cations: (a) Kays, D. L.; Rossin, A.; Day, J. K.; Ooi, L.-L.; Aldridge,
S. Dalton Trans. 2006, 399. (b) Coombs, D. L.; Aldridge, S.; Rossin,
A.; Jones, S.; Willock, D. J. Organometallics 2004, 23, 2911. (c)
Buschmann, W. E.; Miller, J. S. Chem.sEur. J. 1998, 4, 1731. (d)
Hughes, R. P.; Husebo, T. L.; Maddock, S. M. J. Am. Chem. Soc.
1997, 119, 10231. (e) Ferraris, D.; Cox, C.; Anand, R.; Lectka, T.
J. Am. Chem. Soc. 1997, 119, 4319. (f) Bahr, S. R.; Boudjouk, P.
J. Am. Chem. Soc. 1993, 115, 4514.
(38) Reaction of Na[BArf4] with [Cp*M(CO)2GaI2]2 (1, M ) Fe; 13, M )
Ru) can be shown to proceed to the generation of the iodide-bridged
cations [{Cp*M(CO)2}2(µ-I)]+ (10, M ) Fe; 17, M ) Ru) as the
[BArf4]- salts (see Supporting Information for details).
(39) Ueno, K.; Watanabe, T.; Ogino, H. Appl. Organomet. Chem. 2003,
17, 403.
(40) Wanandi, P.; Glaser, P. B.; Tillet, T. D. J. Am. Chem. Soc. 2000,
122, 972.
9
16118 J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008

Cationic Terminal Gallylene Complexes
A R T I C L E S
Figure 1. Positive ion electrospray mass spectra of 8: (a) cone voltage 10 V; (b) cone voltage 50 V.
Scheme 4. Synthesis and Reactivity of Cationic Iodogallylene Complex 9a
a Key reagents and conditions: (i) Na[BArf4] (1.11 equiv), fluorobenzene, -30 to 20 °C, 20 min, 35% isolated yield; (ii) [nBu4N]I (1.2 equiv), d5-
fluorobenzene, 20 °C, 5 min, quantitative by NMR; (iii) CO (1 atm), d5-fluorobenzene, 20 °C, 12 h, quantitative by NMR.
reaction with [nBu4N]I in fluorobenzene solution proceeds
rapidly via addition of iodide at the gallium center and simple
(quantitative) regeneration of the diiodogallyl precursor complex
3. Reaction with carbon monoxide, on the other hand, can be
shown by IR and multinuclear NMR spectroscopies to proceed
via ligand displacement at the iron center, with the correspond-
ing carbonyl complex [Cp*Fe(dppe)(CO)]+[BArf4]- being
generated in quantitative yield (by NMR) over a period of ca.
20 h; gallium metal is precipitated during the course of the
reaction. The course of this reaction can be understood by
recourse to DFT calculations carried out for 9 and [Cp*Fe-
(dppe)(CO)]+ (Vide infra) which reveal a significantly stronger
metal ligand bond for CO vs GaI; gallium metal presumably
results from the disproportionation of the ejected GaI moiety.
(ii) Structural Studies. The structures of precursor iodogallyl
complexes 2-5 and of cationic iodogallylene 9 have been
determined in the solid state by X-ray crystallography (see
Figures 3 and 4), and they imply a crucial role for ancillary
ligand steric shielding in the stabilization and geometric
properties of low-coordinate gallium centers. Thus, in contrast
to the dimeric structure adopted by the corresponding dicarbonyl
complex [Cp*Fe(CO)2GaI2]2 (1),8a the structure of Cp*Fe-
(dppe)GaI2 (3) is monomeric and features a trigonal gallium
center [∑(angles at gallium) ) 360°, within the standard 3σ
9
J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008 16119

A R T I C L E S
Coombs et al.
Figure 2. (a) Positive ion electrospray mass spectrum of 9; (inset) calculated and measured isotopic envelopes for the molecular ion. (b) (upper) Positive
ion MS/MS spectrum of the molecular ion of 9 showing formation of the [Cp*Fe(dppe)]+ fragment by loss of GaI (mass 195.8459); (lower) calculated
isotopic envelope for [Cp*Fe(dppe)]+.
limit]. While such three-coordinate complexes (of the type
LnMEX2) are more common for E ) B, 3 represents a very
rare example of such a ligand system incorporating one of the
heavier group 13 elements.41 The steric bulk of the chelating
dppe ligand is clearly of key importance, with the consequence
that the Fe-Ga distances for 3 [2.3236(14) and 2.3201(14) Å,
for the two distinct molecules in the asymmetric unit] are actually
slightly longer than that measured for 1 [2.314(1) Å] despite the
lower coordination number at gallium (three vs four);8a further
lengthening of the Fe-Ga bond is observed for Cp*Fe(dppe)Ga-
(Mes)I [4; 2.355(1) Å]. Moreover, an informative illustration of
the relative importance of steric and electronic factors can be
obtained by comparison of the structures of iodo(mesityl) com-
plexes Cp*Fe(L)2Ga(Mes)I [4: L2 ) dppe; 5: L2 ) (CO)2] which
9
16120 J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008

Cationic Terminal Gallylene Complexes
A R T I C L E S
Figure 3. Molecular structures of halogallyl complexes (a) 2, (b) 3, (c) 4, and (d) 5 with hydrogen atoms omitted for clarity and atomic displacement
ellipsoids set at the 50% probability level. Key bond lengths (Å) and angles (deg): (for 2) Fe(1)-Ga(1) 2.3558(10), Fe(1)-P(1) 2.2107(13), Ga(1)-I(1)
2.6855(7), Ga(1)-I(2) 2.6416(8), Ga(1)-P(2) 2.5288(12), I(1)-Ga(1)-I(2) 97.80(2); (for 3) Fe(1)-Ga(1) 2.3236(14), Fe(1)-P(1) 2.193(2), Fe(1)-P(2)
2.208(2), Ga(1)-I(1) 2.6294(11), Ga(1)-I(2) 2.6323(11), P(1)-Fe(1)-P(2) 86.88(9), Fe(1)-Ga(1)-I(1) 131.67(5), Fe(1)-Ga(1)-I(2) 134.02(5),
I(1)-Ga(1)-I(2) 94.30(4), angle between least-squares planes defined by [Cp* centroid, Fe(1), Ga(1)] and [Ga(1), I(1), I(2)] 16.78; (for 4) Fe(3)-Ga(2)
2.355(1), Fe(3)-P(4) 2.192(2), Fe(3)-P(19) 2.186(2), Ga(2)-I(1) 2.726(1), Fe(3)-Ga(2)-C(42) 140.0(2), Fe(3)-Ga(2)-I(1) 127.54(4), I(1)-Ga(1)-C(42)
91.3(2), angle between least-squares planes defined by [Cp* centroid, Fe(3), Ga(2)] and [Ga(2), I(1), C(42)] 14.14, angle between least-squares planes
defined by [Ga(2), I(1), C(42)] and [C(42) to C(47)] 82.87; (for 5) Fe(1)-Ga(12) 2.3113(12), Ga(12)-I(22) 2.6073(9), Fe(1)-Ga(12)-C(13) 137.20(19),
Fe(1)-Ga(12)-I(22) 116.86(4), C(13)-Ga(12)-I(22) 105.16(18), angle between least-squares planes defined by [Cp* centroid, Fe(1), Ga(12)] and [Ga(12),
C(13), I(22)] 13.11, angle between least-squares planes defined by [Ga(12), C(13), I(22)] and [C(13) to C(18)] 89.71.
much longer Fe-Ga bond [2.355(1) vs 2.3113(12) Å] and a
significantly wider Fe-Ga-Cipso angle [140.0(2) vs 137.20(19)°]
are found for dppe-ligated 4 implies that the extra steric demands
of the chelating phosphine predominate over any potential increase
in Fe f Ga back-bonding.
The structure of 9 determined crystallographically in the solid
state is shown in Figure 4. Disorder within the structure is
successfully modeled in terms of two cationic species; the major
component (79%) features discrete [Cp*Fe(dppe)(GaI)]+ and
[BArf4]- ions, with no short secondary interactions involving
the Fe-Ga-I unit (within standard van der Waals contacts).
Figure 4. (a) Molecular structure of the major (79%) component of
Key structural features are the essentially linear arrangement
[Cp*Fe(dppe)(GaI)]+[BArf4]- (9) as determined by single crystal X-ray
of the iron, gallium, and iodine atoms [ Fe-Ga-I ) 171.37(3)°]
diffraction studies. Counterion and hydrogen atoms omitted for clarity, and
atomic displacement ellipsoids drawn at the 50% probability level. Key
bond lengths and angles: Fe-Ga, 2.2221(6) Å; Ga-I, 2.4436(5) Å; Fe-P(1),
typical of a terminally bound diatomic ligand {cf. Fe-C-O
) 175.8(5)° for [Cp*Fe(dppe)(CO)]+[PF6]-},20b and the ex-
2.2068(9) Å; Fe-P(2), 2.2199(10) Å; Fe-Ga-I, 171.37(3)°; P(1)-Fe-P(2),
86.60(3)°; P(1)-Fe-Ga, 86.16(3)°; P(2)-Fe-Ga, 91.02(3)°. (b) Space
filling diagram of the cationic component of 9. Color key: gray (carbon),
white (hydrogen), blue (phosphorus), red (gallium), purple (iodine).
tremely short Fe-Ga and Ga-I distances [2.2221(6) and
2.4436(5) Å, respectively]. The gallium-iodine distance is the
shortest yet reported; likewise, the metal-gallium distance, is
among the shortest yet reported involving any transition metal
feature similar alignments of the gallyl ligand and which differ
only in the ancillary ligand framework at the iron center. That a
(41) Buchin, B.; Gemel, C.; Kempter, A.; Cadenbach, T.; Fischer, R. A.
Inorg. Chim. Acta 2006, 359, 4833.
9
J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008 16121

A R T I C L E S
Coombs et al.
Table 2. Energy Partitioning Analyses for the Model Complexes [CpFe(dmpe)(EX)]+ (EX ) N2, CO, BF, and GaI)48
N2
CO
BF
GaI
dipole moment (Debye)a
0 (0)
-1.036
374
-232
-262
-120
38
0.047 (0.138)
-1.459
623
0.884 (0.971)
-3.533
842
1.866 (3.818)
-13.327
367
-234
-236
-103
33
quadrupole moment (au)
Pauli repulsion (∆EPauli, kJ mol-1
)
electrostatic interaction (∆Eelectrostat, kJ mol-1
)
-439
-658
orbital interaction (∆Eorb, kJ mol-1
total interaction energy (∆Eint, kJ mol-1
π contribution to covalent bonding density (%)
)
-397
-469
)
-213
-285
39
42
a Dipole moments quoted are for the ligands in the geometries they have in the complex; those in parentheses are for the isolated ligands.
(and the shortest involving iron).5,42 With due allowance made
for the differing radii of Fe(II) and Ni(II) (0.82 and 0.78 Å),9
the M-Ga distance measured for 9 is comparable to that found
by Uhl and co-workers for the homoleptic diyl system Ni[GaC-
(SiMe3)3]4 [2.1700(4) Å].42 Structural evidence for 9 also points
to ready bending deformation of the Fe-Ga-I bond, viz. large
thermal ellipsoid amplitude for the iodine atom perpendicular
to the Fe-Ga-I axis. Secondary off-axis electron density has
been modeled as a minor (21%) cationic component featuring
a markedly more bent Fe-Ga-I unit [148.92(5)°] and contacts
between I′ and C(55)-C(58) of one of the [BArf4]- aromatic
rings which fall within the sum of the van der Waals radii of
iodine and carbon. Large librational amplitudes at oxygen in
related metal carbonyl complexes are often associated with
analogous Fe-C-O bending motions, and the ready deforma-
tion of the linear Fe-Ga-I fragment in 9 (and a small calculated
energy difference between the linear and bent geometries) is
consistent with the smaller absolute magnitude of directional
covalent contributions to the metal ligand bond (Vide infra).
A contributory factor to the short bond lengths in 9 is the
low coordination number at gallium. Thus, short FesGa bonds
are also associated with the two-coordinate gallium centers in
(OC)4FeGaAr [Ar ) C6H3(C6H2iPr3)2-2,6; d(FesGa) ) 2.2248(7)
Å],22a Cp*Fe(dppe)GaFe(CO)3L [e.g., d(FesGa) ) 2.248(1) for
L ) CO],21 and [{Cp*Fe(CO)2}2Ga]+ [d(FesGa) ) 2.272(1)
and 2.266(1) Å],15b while longer bonds are measured for the
three-coordinate precursor 3 [d(FesGa) ) 2.322 Å (mean);
d(GasI) ) 2.631 Å (mean)] and for the four coordinate system
[Cp*Fe(CO)2GaCl(phen)]+ [d(FesGa) ) 2.3047(4) Å; phen )
1,10-phenanthroline].39,43 Potentially, a second factor underlying
these short bonds is the presence of off-axis electronic contribu-
tions to the bonding, involving gallium-based orbitals of π
symmetry. The contraction of the FesGa bond on halide
abstraction (ca. 4.3% for 9 compared to 3) is markedly less than
that for analogous boron-containing systems (typically
9-10%),15 for which descriptions incorporating FedB π bonds
have been advanced for the cationic products. That said, smaller
changes in bond length as a function of bond order are typically
found for the heavier main group elements,44 and the FesGa
contraction between 9 and 3 mirrors that found between double
and single bonds involving the adjacent group 14 element
germanium (e.g., 4.7% between MnsGe and MndGe bonds).45
In order to better understand the bonding in the unprecedented
ligand system present in 9 and to provide comparison of group
13/17 EX ligands with group 14/16 and group 15/15 counter-
parts, an in-depth computational investigation of the bonding
in 9 and related complexes was undertaken.
(iii) Computational Studies. Density functional theory (DFT)
analyses of electronic structure were carried out using the
computationally efficient model systems [CpFe(dmpe)(EX)]+
(EX ) GaI, BF, CO, and N2; dmpe ) 1,2-bis(dimethylphos-
phino)ethane, Me2PCH2CH2PMe2), bonding analyses for which
are discussed below. For all four ligands, essentially linear
minimum energy geometries are obtained, with that for EX )
GaI ( Fe-Ga-I ) 174.4°) being consistent with the major
component of the solid state structure. Moreover, in order to
rationalize the soft librational deformation implied by crystal-
lographic measurements, calculations were also carried out on
the larger system [Cp*Fe{Me(Ph)PCH2CH2P(Ph)MePh}(GaI)]+,
in which the phenyl groups proximal to the GaI ligand and the
methyl groups of the Cp* ligand are retained, so as to provide
more reliable modeling of the GaI ligand environment in 9.46
A very shallow potential energy surface is found for the
Fe-Ga-I bending deformation (∆E < +3.5 kJ mol-1 for 159
< ϑ < 179°). Consistent with this, a relatively small energy
increase (+11 kJ mol-1) is associated with a Fe-Ga-I angle
(149°) analogous to that found for the “bent” conformer of 9 in
the solid state (see the Supporting Information for details of
the energy profile). This energy difference is comparable to
binding energies typically associated with halogen · · ·arene
interactions (e.g., 12.6-20.9 kJ mol-1 for “complexes” between
halogens and various substituted benzenes, as determined from
charge transfer bands),47 and the existence of such contacts in
the solid state structure of 9 [signaled by distances from I′ to
C(55)-C(58) (3.437-3.665 Å) which are within the sum of
the respective van der Waals radii]9 then offers a rationale for
the bent structure of the minor (21%) component found in the
crystal.
From a bonding perspective, a breakdown of the covalent
(orbital) components of the metal-ligand bonds of each of the
four model compounds reveals notable similarities (see Table
2);48 in each case, partitioning of the bonding density implies
that orbital interactions of π symmetry represent a significant
fraction of the covalent bond [i.e., 33 (GaI), 42 (BF), 39 (CO),
and 38% (N2) of the total covalent bonding density]. However,
taken in isolation, such an analysis is somewhat misleading,
since it merely compares the relative σ/π breakdowns for each
complex, without quantifying the absolute magnitude of the
orbital (and other) contributions to the metal-ligand bond. To
enable a more in-depth comparison, the fragment correlation
(42) Uhl, W.; Benter, M.; Melle, S.; Saak, W.; Frenking, G.; Uddin, J.
Organometallics 1999, 18, 3778.
(46) Analogous calculations on the “full” cation [Cp*Fe(dppe)(GaI)]+ at
an acceptable level of theory did not satisfy all of the convergence
criteria.
(43) For other examples of complexes containing base-stabilized EX
fragments as ligands, see, for example: (a) Fischer, R. A.; Schulte,
M. M.; Weiss, J.; Zsolnai, L.; Jacobi, A.; Huttner, G.; Frenking, G.;
Boehme, C.; Vyboishchikov, S. F. J. Am. Chem. Soc. 1998, 120, 1237.
(44) Power, P. P. Chem. ReV. 1999, 99, 3463.
(47) Rosokha, S. V.; Kochi, J. K. Struct. Bonding (Berlin) 2007, 126, 137.
(48) Under the energy decomposition analysis employed, the interaction
energy ∆Eint between two fragments is given by the sum of attractive
orbital (∆E orb) and electrostatic (∆E electrostat) and repulsive Pauli (∆E
(45) Melzer, D.; Weiss, E. J. Organomet. Chem. 1984, 263, 67.
Pauli) terms.6c ∆Eint ) ∆Eorb + ∆Eelectrostat + ∆EPauli
.
9
16122 J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008

Cationic Terminal Gallylene Complexes
A R T I C L E S
Figure 5. Molecular orbital energy level diagram for the model complexes [CpFe(dmpe)(CO)]+ and [CpFe(dmpe)(GaI)]+ showing correlation with the
[CpFe(dmpe)]+ and CO/GaI fragments. Local σ symmetry interactions involving the LUMO of [CpFe(dmpe)]+ and the HOMO of the CO/GaI ligand are
shown in blue; π symmetry interactions originating in the HOMO and HOMO-2 of [CpFe(dmpe)]+ and the degenerate pair of LUMOs of CO/GaI are shown
in red.
diagrams for [CpFe(dmpe)(CO)]+ and [CpFe(dmpe)(GaI)]+
were determined and are reproduced in Figure 5. The HOMOs
for both compounds are dominated by a metal centered d-orbital
with δ-symmetry along the Fe-C/Ga vector; this orbital is
therefore nonbonding for the CO/GaI ligands. The HOMO/
HOMO-2 and LUMO states for the [CpFe(dmpe)]+ fragment
present π- and σ-symmetry,49 respectively, at the vacant
coordination site, and so bonding between this fragment and
CO/GaI has the potential to include both σ-donation from the
ligand to the metal center and π-back-donation. The primary
contribution to bonding in each case involves the states of
σ-symmetry, with the CO ligand showing a much stronger
interaction with the metal center than GaI. Indeed, for [CpFe-
(dmpe)(CO)]+, a number of other molecular orbitals (involving
interactions of the metal center with the phosphine and
cyclopentadienyl ligands) are actually present at energies higher
than that of the primary Fe-CO σ-bonding orbital. This orbital
in the CO case is the HOMO-10 (at -11.72 eV), whereas the
corresponding orbital for the GaI complex is the HOMO-5 (at
-9.55 eV; Figure 6).
For each complex there are also two sets of orbitals which
could conceivably be characterized by metal to CO/GaI π
bonding; the two degenerate π* antibonding orbitals of the
ligand can interact with complementary metal d-orbitals, i.e.
the mutually orthogonal HOMO and HOMO-2 of [CpFe-
(dmpe)]+. The bonding MOs so formed are lower in energy
for [CpFe(dmpe)(CO)]+ (HOMO-1 and HOMO-2 at -8.44 and
-8.83 eV, respectively) than for [CpFe(dmpe)(GaI)]+ (-7.95
and -8.34 eV), despite the higher energy LUMO states for the
CO ligand (-2.24 eV cf. -2.70 eV for GaI). Thus, as in the
Figure 6. Orbital 41 (HOMO-5) for [CpFe(dmpe)(GaI)]+: the metal ligand
σ bonding MO.
case of the σ bonding manifold, significantly smaller perturba-
tions of the bonding MOs are observed for the π symmetry
interactions of the GaI complex (cf. CO). As such, it seems
likely that the relatively similar σ and π ratios calculated for
the GaI and CO complexes (67:33 and 61:39, respectively; Table
2) are reflective of marked reductions in both the σ and π orbital
interactions for the GaI ligand. This supposition is corroborated
by explicit calculation of the orbital, electrostatic, and Pauli
energetic contributions to the metal-ligand bond (Vide infra).48
Moreover, in terms of the π symmetry atomic orbitals at gallium,
Figure 7 gives an idea of the MOs to which these ultimately
contribute in the model complex [CpFe(dmpe)(GaI)]+. Thus,
the HOMO-3 and HOMO-1 orbitals [at -9.06 and -7.95 eV;
Figure 7a and b], which feature in-phase combinations of Ga
4px with the I 5px and Fe 3dxy orbitals, respectively, feature 4.3
and 6.2% contributions from the gallium orbital. By contrast,
the antibonding LUMO+3 (-4.94 eV) features a 71.6%
contribution from Ga 4px, a finding which also points to the
(49) Schilling, B. E. R.; Hoffmann, R.; Lichtenberger, D. J. Am. Chem.
Soc. 1979, 101, 585.
9
J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008 16123

A R T I C L E S
Coombs et al.
-285 (BF), -213 (CO), and -120 kJ mol-1 (N2)] reveal
significantly weaker binding of the GaI ligand,7–9 and cor-
roborate its experimentally observed (quantitative) displacement
from 9 by the addition of carbon monoxide to give
[Cp*Fe(dppe)(CO)]+[BArf4]-.
Conclusions
Halide abstraction has been shown to be a viable method for
the generation of mononuclear cationic complexes containing
gallium donor ligands. The ability to isolate tractable two-
coordinate products, however, is strongly dependent on the steric
and electronics properties of the metal/ligand fragment. In the
case of complexes containing ancillary π-acceptor ligands such
as CO, cationic complexes can only be isolated as base-trapped
adducts, even with bulky aryl substituents at gallium. Base-
free gallylene species such as [Cp*Fe(CO)2GaMes]+ can be
identified only in the vapor phase by electrospray mass
spectrometry experiments. With bis(phosphine) donor sets at
the metal, the more favorable steric/electronic environment
allows for the isolation of two-coordinate ligand systems, eVen
with halide substituents at gallium. Thus, [Cp*Fe(dppe)-
(GaI)]+[BArf4]- (9) can be synthesized and shown crystallo-
graphically to feature a terminally bound GaI ligand; 9 represents
the first experimental realization of a complex containing a
valence isoelectronic group 13/group 17 analogue of CO and
N2. DFT calculations reveal a relatively weakly bound GaI
ligand, which is confirmed experimentally by the reaction of 9
with CO to give [Cp*Fe(dppe)(CO)]+[BArf4]-. In the absence
of such reagents, 9 is stable for weeks in fluorobenzene solution,
presumably reflecting (i) effective steric shielding of the gallium
center by the ancillary phosphine and Cp* ligands; (ii) a net
cationic charge which retards the tendency toward dimerization
found for putative charge neutral systems; and (iii) (albeit
relatively minor) population of the LUMOs of the GaI molecule
through π overlap with the HOMO and HOMO-2 of the
[Cp*Fe(dppe)]+ fragment.
Figure 7. MOs for [CpFe(dmpe)(GaI)]+ featuring contributions from the
Ga 4px orbital: (a) the HOMO-3; (b) the HOMO-1; and (c) the LUMO+3.
Acknowledgment. We thank the EPSRC for funding, including
the National Crystallography and Mass Spectrometry Services;
STFC for access to synchrotron facilities; and A. S. Weller, A.
Lubben, D. L. Kays, and C. Tang for assistance with mass
spectrometry and other compound characterization.
relatively weakly bonding nature of any interactions in which
this atomic orbital is involved.
In energetic terms, the magnitude of the covalent (orbital)
bonding component for the GaI complex (-236 kJ mol-1) can
be put into context by values of -469, -397, and -262 kJ
mol-1 for the corresponding BF, CO, and N2 complexes and
by a value of -234 kJ mol-1 for the electrostatic contribution
to the (significantly polar) Fe-GaI bond. Presumably, despite
the higher energy of the HOMO for GaI (-6.08 eV cf. -9.03
eV for CO, Figure 5) and the greater localization of the LUMO
at the donor atom (plus its lower energy), the weaker orbital
contribution for GaI reflects (at least in part) the more diffuse
nature of the 4s/4p derived orbitals at gallium and less effective
interaction with the fragment orbitals of [CpFe(dmpe)]+.49 Thus,
overall metal-ligand interaction energies [∆Eint ) -103 (GaI),
Supporting Information Available: Crystallographic data for
[Cp*Fe(CO)2]2GaN(SiMe3)2 (7), [{Cp*M(CO)2}2(µ-I)]+[BArf4]-
(10, M ) Fe; 17, M ) Ru), and [Cp*Fe(dppe)Cl]+[BArf4]- (11);
spectroscopic and crystallographic data for [Cp*Ru(CO)2GaI2]2 (13)
and details of its reactivity toward dppe; spectroscopic and
crystallographic data for [Cp*Ru(CO)2GaI2]2(µ-dppe) (14) and
Cp*Ru(dppe)I (15); and details of DFT calculations. This material
is available free of charge via the Internet at http://pubs.acs.org.
JA806655F
9
16124 J. AM. CHEM. SOC. VOL. 130, NO. 47, 2008