Electronically unsaturated three-coordinate chloride and methyl complexes of iron, cobalt, and nickel

Article abstract of DOI:10.1021/ja025583m

Three-coordinate organometallic complexes are rare, especially with the prototypical methyl ligand. Using a hindered, rigid bidentate ligand (L), it is possible to create 12-electron methyliron(II) and 13-electron methylcobalt(II) complexes. These complex

Full text of DOI:10.1021/ja025583m

Published on Web 11/08/2002
Electronically Unsaturated Three-Coordinate Chloride and
Methyl Complexes of Iron, Cobalt, and Nickel
Patrick L. Holland,* Thomas R. Cundari,*,†,‡ Lanyn L. Perez, Nathan A. Eckert, and
Rene J. Lachicotte
Contribution from the Department of Chemistry, UniVersity of Rochester,
Rochester, New York 14627, and Department of Chemistry and the Computational Research on
Materials Institute, UniVersity of Memphis, Memphis, Tennesee 38152
Received January 14, 2002
Abstract: Three-coordinate organometallic complexes are rare, especially with the prototypical methyl ligand.
Using a hindered, rigid bidentate ligand (L), it is possible to create 12-electron methyliron(II) and 13-electron
methylcobalt(II) complexes. These complexes are thermally stable, and 1H NMR spectra suggest that the
low coordination number is maintained in solution. Attempts to create the 14-electron LNiCH3 led instead
to the three-coordinate nickel(I) complex LNi(THF). Single crystals of LMCH3 are isomorphous with the
new three-coordinate chloride complexes LNiCl and LCoCl. Along with the recently reported LFeCl (Smith,
J. M.; Lachicotte, R. J.; Holland, P. L. Chem. Commun. 2001, 1542), these are the only examples of three-
coordinate iron(II), cobalt(II), and nickel(II) complexes with terminal chloride ligands, enabling the systematic
evaluation of the effect of coordination number and metal identity on M-Cl bond lengths. Electronic structure
calculations predict the ground states of the trigonal complexes.
Introduction
protection often block the metal from the interesting inter-
molecular reactivity expected from the unusual orbital structure
Three-coordinate complexes of the transition metals are rare,
because it is difficult with only three ligand donors to reach
preferred configurations with 16 or 18 electrons in the metal
bonding and nonbonding orbitals. In all solution-phase three-
coordinate transition metal complexes, the metal is protected
with extremely bulky ligands. In this way, chemists have been
able to stabilize three-coordinate1 and even two-coordinate2
metal centers; the ground-breaking work of Power is especially
notable in the development of routes to three-coordinate late
transition metal complexes.3,4 However, the ligands that offer
of the low-coordinate metal. When reactivity is observed at
three-coordinate metal sites, it can be spectacular. Work by
Wolczanski and Cummins is illustrative, in which low-
coordinate complexes of group 4-6 metals cleave C-H bonds
of alkanes, the N-N bond of nitrous oxide, and triple bonds of
CO, NO, and N2.5-8 The small HOMO-LUMO gap, coinci-
dence of electrophilic and nucleophilic behavior, and attendant
large binding constants for the fourth ligand are factors that
contribute to this prodigious ability to activate organic and
inorganic substrates.
Low coordination may also be used in nature to perform
difficult bond-cleaving tasks: the iron-molybdenum cofactor
of nitrogenase has six iron atoms of roughly trigonal planar
geometry at its core (Figure 1).9 Although there is no conclusive
evidence that N2 binds or is transformed at these iron atoms,
biochemists, spectroscopists, and theoreticians have accumulated
evidence that points toward substrate binding one or more of
* To whom correspondence should be addressed. E-mail: holland@
chem.rochester.edu.
† University of Memphis.
‡ Current address: Department of Chemistry, University of North Texas,
Box 305070, Denton, TX 76203.
(1) Reviews: (a) Cummins, C. C. Prog. Inorg. Chem. 1998, 47, 685-836. (b)
Alvarez, S. Coord. Chem. ReV. 1999, 193-195, 13-41.
(2) (a) Power, P. P. Comments Inorg. Chem. 1989, 8, 177-202. (b) Power, P.
P. Chemtracts: Inorg. Chem. 1994, 6, 181-195. (c) Ellison, J. J.; Ruhlandt-
Senge, K.; Power, P. P. Angew. Chem., Int. Ed. Engl. 1994, 33, 1178-
1180. (d) Wehmschulte, R. J.; Power, P. P. Organometallics 1995, 14,
3264-3267.
(5) Neithamer, D. R.; LaPointe, R. E.; Wheeler, R. A.; Richeson, D. S.; Van
Duyne, G. D.; Wolczanski, P. T. J. Am. Chem. Soc. 1989, 111, 9056-
9072.
(6) (a) Cummins, C. C.; Baxter, S. M.; Wolczanski, P. T. J. Am. Chem. Soc.
1988, 110, 8731-8733. (b) Schafer, D. F.; Wolczanski, P. T. J. Am. Chem.
Soc. 1998, 120, 4881-4882.
(7) Cummins, C. C. Chem. Commun. 1998, 1777-1786 and references within.
(8) (a) Odom, A. L.; Cummins, C. C.; Protasiewicz, J. D. J. Am. Chem. Soc.
1995, 117, 6613-6614. (b) Veige, A. S.; Slaughter, L. M.; Wolczanski, P.
T.; Matsunaga, N.; Decker, S. A.; Cundari, T. R. J. Am. Chem. Soc. 2001,
123, 6419-6420.
(9) (a) There are short Fe-Fe contacts of 2.65 ( 0.04 Å in the FeMoco that
probably modify the behavior of iron: Mayer, S. M.; Lawson, D. M.;
Gormal, C. A.; Roe, S. M.; Smith, B. E. J. Mol. Biol. 1999, 292, 871-
891. (b) A very recent paper presents evidence that there is a fourth ligand
coordinated to these iron atoms: Einsle, O.; Tezcan, F. A.; Andrade, S. L.
A.; Schmid, B.; Yoshida, M.; Howard, J. B.; Rees, D. C. Science 2002,
297, 1696-1700.
(3) (a) Murray, B. D.; Power, P. P. Inorg. Chem. 1984, 23, 4584-4588. (b)
Murray, B. D.; Power, P. P. J. Am. Chem. Soc. 1984, 106, 7011-7015. (c)
Hope, H.; Olmstead, M. M.; Murray, B. D.; Power, P. P. J. Am. Chem.
Soc. 1985, 107, 712-713. (d) Olmstead, M. M.; Power, P. P.; Sigel, G.
Inorg. Chem. 1986, 25, 1027-1033. (e) Ellison, J. J.; Power, P. P.; Shoner,
S. C. J. Am. Chem. Soc. 1989, 111, 8044-8046. (f) Bartlett, R. A.; Ellison,
J. J.; Power, P. P.; Shoner, S. C. Inorg. Chem. 1991, 30, 2888-2894. (g)
Olmstead, M. M.; Power, P. P.; Shoner, S. C. Inorg. Chem. 1991, 30, 2547-
2551. (h) Power, P. P.; Shoner, S. C. Angew. Chem., Int. Ed. Engl. 1991,
30, 330-332. (i) Chen, H.; Olmstead, M. M.; Pestana, D. C.; Power, P. P.
Inorg. Chem. 1991, 30, 1783-1787.
(4) [Fe(SC6H2tBu3)3]-: (a) MacDonnell, F. M.; Ruhlandt-Senge, K.; Ellison,
J. J.; Holm, R. H.; Power, P. P. Inorg. Chem. 1995, 34, 1815-1822. (b)
Sanakis, Y.; Power, P. P.; Stubna, A.; Mu¨nck, E. Inorg. Chem. 2002, 41,
2690-2696. (c) Agostic interactions in some low-coordinate iron thiolate
complexes make higher coordinate descriptions more accurate: Evans, D.
J.; Hughes, D. L.; Silver, J. Inorg. Chem. 1997, 36, 747-748.
9
14416
J. AM. CHEM. SOC. 2002, 124, 14416-14424
10.1021/ja025583m CCC: $22.00 © 2002 American Chemical Society

Electronically Unsaturated Complexes of Fe, Co, Ni
A R T I C L E S
Figure 1. The iron-molybdenum cofactor of iron-molybdenum nitro-
genase. Coordinates derived from ref 9a.
Figure 3. Plot of LFeCH3 using 50% thermal ellipsoids. Hydrogen atoms
are omitted for clarity.
Figure 2. The â-diketiminate ligand used in this paper, L.
contribution fully discloses the structure and properties of this
12-valence-electron iron(II) complex. More importantly, we
show that the synthesis of stable trigonal-planar complexes
can be extended to analogous three-coordinate cobalt(II) and
nickel(II) chloride complexes and to a 13-electron cobalt(II)
methyl complex. Interestingly, all five of the above chloride
and methyl complexes have very similar solid-state structures,
although the ability to create methyl complexes drastically
changes through the series.
the coordinatively unsaturated iron atoms.10,11 Synthetic three-
coordinate iron compounds are expected to give insight into
the reactivity of these special iron sites,4 but only one synthetic
three-coordinate iron complex actually reacts with N2.12
A number of groups have reported that â-diketiminate ligands
are adept at stabilizing three-coordination in main group and
late transition metals.13,14 However, the low-coordinate chemistry
of iron, cobalt, and nickel with bulky â-diketiminates was
unknown prior to our work.15 The diketiminate ligand is ideal
because it is difficult to displace (chelate effect; anionic), easily
synthesized, and sterically enforces three-coordination with only
two donors. For example, sufficiently large diketiminate ligands
can stabilize a three-coordinate iron(II) complex with a syntheti-
cally versatile terminal chloride ligand.16 We recently reported
detailed electronic structure studies of this and other three-
coordinate iron complexes using Mo¨ssbauer and EPR spectros-
copy coupled with theoretical methods and mentioned the
synthesis of the first three-coordinate methyl complex of the
transition metals, LFeCH3 [L ) 2,2,6,6-tetramethyl-3,5-bis(2,6-
diisopropylphenylimido)hept-4-yl, Figure 2].17 The present
Results
Synthesis, NMR Spectroscopy, and Stability of a 12-
Electron Methyliron Complex. Treatment of ethereal solutions
of LFeCl with methyl Grignard reagents yields pale orange
crystalline LFeCH3. This complex was briefly reported, but its
solution behavior and solid-state structure have not been
discussed.17 Like its chloride precursor, LFeCH3 is highly
sensitive to air or moisture and was handled in a glovebox under
purified nitrogen. This complex is thermally stable in solution,
decomposing only after days in boiling C6D6. A representation
of its X-ray crystal structure is shown in Figure 3; the two halves
of the molecule are related by a crystallographic 2-fold axis
that renders the FeN2C unit rigorously planar. No agostic
interactions of ligand C-H bonds are evident (Fe‚‚‚H-C g
2.78 Å), showing that it is a bona fide4c 12-electron organo-
metallic complex. This trigonal planar metal geometry (ap-
proximately C2V) is shared by all of the three-coordinate chloride
and methyl complexes described below, because they all
crystallize in the same space group (C2/c) with the same packing
(Z ) 4, rotation axis along b and through the M-Cl or M-C
bond).
(10) Selected recent biochemical work that supports binding of N2 at Fe: (a)
Lee, H.-I.; Sorlie, M.; Christiansen, J.; Song, R.; Dean, D. R.; Hales, B. J.;
Hoffman, B. M. J. Am. Chem. Soc. 2000, 122, 5582-5587. (b) Christiansen,
J.; Dean, D. R.; Seefeldt, L. C. Annu. ReV. Plant Physiol. Plant Mol. Biol.
2001, 52, 269-295. (c) Benton, P. M. C.; Mayer, S. M.; Shao, J.; Hoffman,
B. M.; Dean, D. R.; Seefeldt, L. C. Biochemistry 2001, 40, 13816-13825.
(d) Krahn, E.; Weiss, B. J. R.; Kro¨ckel, M.; Groppe, J.; Henkel, G.; Cramer,
S. P.; Trautwein, A. X.; Schneider, K.; Mu¨ller, A. J. Biol. Inorg. Chem.
2002, 7, 37-45.
(11) Examples of theoretical work on N2 binding to the FeMoco: (a) Dance, I.
Chem. Commun. 1998, 523-530. (b) Siegbahn, P. E. M.; Westerberg, J.;
Svensson, M.; Crabtree, R. H. J. Phys. Chem. B 1998, 102, 1615-1623.
(c) Rod, T. H.; Nørskov, J. K. J. Am. Chem. Soc. 2000, 122, 12751-
12763.
(12) 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.
The proton NMR spectrum of LFeCH3 shows resonances for
all of the protons except the iron-bound methyl group.18 The
compound is paramagnetic (solution µeff ) 5.5 µB, S ) 2), and
the chemical shifts span a range of (150 ppm. The aryl groups
are locked on the NMR time scale, as shown by the inequiva-
lence of the two methyl groups of the isopropyl substituents.19
The diketiminate proton resonances could be assigned using
integration, leaving only two ambiguities: the inequivalent
(13) Bourget-Merle, L.; Lappert, M. F.; Severn, J. R. Chem. ReV. 2002, 102, in
press.
(14) (a) Holland, P. L.; Tolman, W. B. J. Am. Chem. Soc. 1999, 121, 7270-
7271. (b) Holland, P. L.; Tolman, W. B. J. Am. Chem. Soc. 2000, 122,
6331-6332. (c) Randall, D. W.; George, S. D.; Holland, P. L.; Hedman,
B.; Hodgson, K. O.; Tolman, W. B.; Solomon, E. I. J. Am. Chem. Soc.
2000, 122, 11632-11648. (d) Jazdzewski, B. A.; Holland, P. L.; Pink, M.;
Young, V. G., Jr.; Spencer, D. J. E.; Tolman, W. B. Inorg. Chem. 2001,
40, 6097-6107. (e) Spencer, D. J. E.; Aboelella, N. W.; Reynolds, A. M.;
Holland, P. L.; Tolman, W. B. J. Am. Chem. Soc. 2002, 124, 2108-2109.
(15) (a) Four-coordinate nickel complexes of the protonated diketimine ligand:
Feldman, J.; McLain, S. J.; Parthasarathy, A.; Marshall, W. J.; Calabrese,
J. C.; Arthur, S. D. Organometallics 1997, 16, 1514-1516. (b) Power’s
group has recently created some three-coordinate cobalt and iron complexes
using diketiminate ligands: Panda, A.; Stender, M.; Wright, R. J.; Olmstead,
M. M.; Klavins, P.; Power, P. P. Inorg. Chem. 2002, 41, 3909-3916.
(16) Smith, J. M.; Lachicotte, R. J.; Holland, P. L. Chem. Commun. 2001, 1542-
1543.
(18) Attempts to identify a resonance corresponding to the iron-bound methyl
group included the synthesis of LFeCD3, but no deuterium resonance could
be located between (1500 ppm in 2H NMR spectra. See: Reuben, J.; Fiat,
D. J. Am. Chem. Soc. 1969, 91, 1242. It is likely that the methyl protons/
deuterons are relaxed by the paramagnetic iron too quickly for observation.
(19) (a) Radzewich, C. E.; Guzei, I. A.; Jordan, R. F. J. Am. Chem. Soc. 1999,
121, 8673-8674. (b) Prust, J.; Stasch, A.; Zheng, W.; Roesky, H. W.;
Alexopoulos, E.; Uson, I.; Boehler, D.; Schuchardt, T. Organometallics
2001, 20, 3825-3828.
(17) Andres, H.; Bominaar, E.; Smith, J. M.; Eckert, N. A.; Holland, P. L.;
Mu¨nck, E. J. Am. Chem. Soc. 2002, 124, 3012-3025.
9
J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002 14417

A R T I C L E S
Holland et al.
Scheme 1
Figure 4. Plot of the major disorder component of LCo(µ-Cl)2Li(ether)2
using 50% thermal ellipsoids. The minor component has two THF ligands
coordinated to the lithium atom.
isopropyl methyl groups each integrate to 12, and the isopropyl
methine and aryl protons meta to nitrogen each integrate to 4.
Fortunately, the distances to the paramagnetic iron center (from
the solid-state structure) differ greatly: the “front” (closer to
X) isopropyl methyl protons are much closer (av 3.99 Å; closest
2.78 Å) than the “back” (closer to the backbone) methyl protons
(av 5.53 Å; closest 5.14 Å), and the isopropyl methine protons
are much closer (av 3.29 Å; closest 3.08 Å) than the aryl protons
(av 5.18 Å; closest 5.14 Å). In each case, our assignments follow
from correlating the distances to iron with (a) the shift of the
peak and (b) the relaxation time. In general, the relaxation time
was estimated from the peak broadness;20 the accuracy of this
method was confirmed by the measurement of similar ((1 ms)
T1 and T2 values for each resonance of LFeCH3.
Synthesis and Structure of Low-Coordinate Cobalt Chlo-
ride Complexes. Addition of a lithium â-diketiminate complex
to a transition metal dihalide has been a general method for
creation of three-coordinate iron(II) and copper(II) com-
plexes.14,16 We have shown that the additional bulkiness of the
tert-butyl groups is necessary for restraining iron(II) to three-
coordination.16 However, addition of LLi(THF) to a slurry of
CoCl2(THF)1.5 in THF yielded not LCoCl but dark brown
LCo(µ-Cl)2Li(ether)2, in 48% yield. This “ate” compound was
characterized by X-ray crystallographic analysis (Figure 4). The
best crystallographic model included two disorder components,
one consisting of two THF ligands attached to lithium (minor)
and the other containing THF and diethyl ether donors (major).
The structure is well-ordered in the vicinity of the cobalt and
diketiminate and, like other tetrahedral â-diketiminate com-
plexes, has some folding of the MN2C3 ring into a “boat”
conformation.13,16,21
Figure 5. Plots of (a) LCoCl, (b) LCoCH3, and (c) LNiCl using 50%
thermal ellipsoids. Hydrogen atoms are omitted for clarity.
formation of the desired complex (Scheme 1). Accordingly,
treatment of CoCl2(THF)1.5 with LTl in THF gave the desired
three-coordinate product LCoCl. Its X-ray crystal structure
(Figure 5a) is virtually indistinguishable from that of LFeCl,
and its metrical parameters will be discussed below. Crystals
of LCoCl are strikingly dichroic, appearing purple or green
depending on the direction of the transmitted light.
Both cobalt(II) complexes are high-spin, with solution
magnetic moments of 4.7 µB that suggest an S ) 3/2 ground
state. The 1H NMR spectra show paramagnetically shifted peaks
that differ substantially by metal geometry. In tetrahedral LCo(µ-
Cl)2Li(ether)2, assignment of the broad, overlapping resonances
was troublesome. On the other hand, resonances corresponding
1
to all protons were reasonably sharp in the H NMR spectrum
Thallium(I) salts of ligands are often used in order to take
advantage of the insolubility of TlCl as a driving force for
of trigonal-planar LCoCl, and these could be assigned in the
same way as in LFeCl and LFeCH3 using integration, shift, and
relaxation time. Most of the paramagnetic shifts are in the same
direction as in the iron(II) analogues, with the most notable
exception being the backbone vinyl resonance, which is at very
high field in LCoCl and very low field in LFeCl.16,36 The NMR
analysis suggests that the solution structure of LCoCl is similar
to that observed in the solid state.
(20) Ming, L.-J. In Physical Methods in Bioinorganic Chemistry; Que, L., Ed.;
University Science Books: Sausalito, CA, 2000. In rapidly tumbling small
molecules, Τ2 ≈ T1.
(21) Analysis of â-diketiminate distortion: (a) Stender, M.; Eichler, B. E.;
Hardman, N. J.; Power, P. P.; Prust, J.; Noltemeyer, M.; Roesky, H. W.
Inorg. Chem. 2001, 40, 2794-2799. (b) Hitchcock, P. B.; Lappert, M. F.;
Layh, M. J. Chem. Soc., Dalton Trans. 2001, 2409-2416. (c) MacAdams,
L.; Kim, W.-K.; Liable-Sands, L. M.; Guzei, I. A.; Rheingold, A. L.;
Theopold, K. H. Organometallics 2002, 21, 952-960. (d) Bailey, P. J.;
Coxall, R. A.; Dick, C. M.; Fabre, S.; Parsons, S. Organometallics 2001,
20, 798-801. (e) Cheng, M.; Moore, D. R.; Reczek, J. J.; Chamberlain, B.
M.; Lobkovsky, E. B.; Coates, G. W. J. Am. Chem. Soc. 2001, 123, 8738-
8749.
A Three-Coordinate Methylcobalt Complex. The synthetic
route to LFeCH3 was also successful with Co(II): addition of
methyl Grignard reagent to either LCo(µ-Cl)2Li(ether)2 or
9
14418 J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002

Electronically Unsaturated Complexes of Fe, Co, Ni
A R T I C L E S
Figure 6. 1H NMR spectrum (400 MHz) of a solution of LCoCH3 in C6D6.
LCoCl gave good yields of the 13-electron organometallic
complex LCoCH3. Its solid-state structure is shown in Figure
5b, and metrical details are discussed below. The solution
magnetic moment (4.9 µB) and proton NMR spectra were similar
to those of LCoCl, strongly suggesting that it also has a high-
spin d7 configuration at cobalt. Again, all ligand hydrogen atoms
are at least 2.8 Å from the cobalt atom, indicating that no direct
electronic interaction between C-H bonds and the metal exists.
The stability of LCoCH3 is greater than conventionally
expected for an odd-electron organometallic complex,22 espe-
cially one that is so electronically unsaturated. Heating solutions
of LCoCH3 in benzene at 80 °C gave slow decomposition over
the course of a few days, but solutions were stable for long
periods at room temperature (with rigorous exclusion of air and
water).
Figure 7. (a) Plot of LNi(THF) using 50% thermal ellipsoids. Hydrogen
atoms are omitted for clarity. (b) X-band EPR spectrum of LNi(THF) in
2-methyltetrahydrofuran (77 K). Simulation yielded g values of 2.07, 2.11,
and 2.51.
1
It was possible to obtain well-resolved H NMR spectra of
LCoCH3 that have paramagnetically shifted resonances (Figure
6). Integration and peak broadness (T2) were again used to assign
the resonances to specific sets of protons. Again the CH3
resonance is not visible in LCoCH3, nor is any 2H resonance in
LCoCD3. As with LCoCl, the meta protons of the aryl
substituents resonate further downfield, and the backbone proton
of the diketiminate resonates much farther upfield, but the other
resonances have paramagnetic shifts in the same direction as
in LFeCH3.
A Three-Coordinate Nickel(II) Chloride Complex. In
analogy to the synthesis of LFeCl from FeCl2(THF)1.5 and
LLi(THF),16 NiCl2(THF)0.7 was treated with the lithium diketim-
inate to give a deep green solution. While the three-coordinate
iron(II) and cobalt(II) complexes were sensitive to oxidation
by CH2Cl2, the nickel(II) complex could be crystallized from
this solvent to give high yields of the three-coordinate complex
LNiCl. The solution magnetic moment was 3.1 µB, consistent
with an S ) 1 ground state. Crystals were isomorphous with
the three-coordinate iron and cobalt complexes, and a thermal-
ellipsoid plot of the crystallographic model is shown in Figure
5c. The metrical parameters will be discussed below.
resonance and between the 4-proton peak at δ 19 and the two
isopropyl methyl resonances. These observations allow us to
assign these two signals and confirm the assignments for the
other signals (see Experimental Section).
Attempted Alkylation of Nickel Gives Reduction to
Nickel(I). Treatment of LNiCl with methyllithium or Grignard
reagents did not give an alkylnickel(II) complex analogous to
the iron(II) and cobalt(II) examples described above. Instead, a
brown product was formed, the solid-state structure of which
is shown in Figure 7a. The crystallographic analysis clearly
shows that reduction occurred to give a nickel(I) diketiminate
complex with solvent THF as the second ligand. The coordina-
tion of tetrahydrofuran is not irreversibly lost in pentane or
diethyl ether solution, because crystals grown from these
solvents retained THF. Support for the nickel(I) oxidation state
comes from the solution magnetic moment (2.0 µB in THF-d8)
and a rhombic X-band EPR signal characteristic of an S ) 1/2
system (Figure 7b).
LNiCl + CH3Li + THF f LNi(THF) + LiCl + CH3• (1)
The proton NMR resonances were not as highly shifted as
the above cobalt and iron compounds, with only the backbone
resonance outside the range of +30 to -15 ppm. The relaxation
times were 10-20 ms and did not constitute a good criterion
for assigning peaks in the 1H NMR spectrum. These relaxation
times were too short to assign peaks using NOE spectroscopy,
but DQF-COSY (see Supporting Information) showed cor-
relations between the 4-proton peak at δ 29 and the other aryl
The reduction product LNi(THF) was observed even with
bulky alkyllithium reagents such as neopentyllithium. To
understand the mechanism of reduction, the reaction of LNiCl
with methyllithium was studied in greater detail. The reaction
1
is quite clean, with H NMR spectra in THF-d8 showing only
the very broad resonances of LNi(THF) and sharp singlets at δ
0.86 and 0.18 corresponding to ethane and methane. Methyl
radicals are expected as the product of methyl oxidation by
nickel, and it is reasonable to propose that the ethane and
(22) Poli, R. Chem. ReV. 1996, 96, 2135-2204.
9
J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002 14419

A R T I C L E S
Holland et al.
Table 1. Comparison of Calculated and Experimental Structures
â-diketiminate ligands. The complexes LMCl are distinctive
among three-coordinate complexes because of the combination
of the stable chelating diketiminate ligand and the easily
substituted terminal halide ligand. Each of the LMCl complexes
is the first characterized three-coordinate complex of its divalent
metal with a terminal chloride ligand. As such, they offer a
route into a systematic series of trigonal complexes with
differing electronic properties and spectroscopic parameters.17
for Three-Coordinate Metal Chloridesa
M−N (Å)
M−Cl (Å)
N−M−N (deg)
L′FeCl
LFeCl
L′CoCl
LCoCl
L′NiCl
LNiCl
(5A1)
(4A2)
(3A2)
2.005
1.948(2)
1.952
1.902(2)
1.922
1.815(3)
2.198
2.172(1)
2.178
2.140(3)
2.148
2.137(2)
93.1
96.35(11)
94.0
99.49(13)
91.8
97.3(2)
Solution susceptibility measurements suggest that all of the
complexes have maximum spin multiplicity at room tempera-
ture. Comparative calculations on L′MCl (L′ ) diketiminate
with alkyl and aryl groups replaced with hydrogen) using
restricted open-shell DFT under C2V symmetry support the
contention that high-spin configurations are lowest in energy.
These electronic configurations are shown in Figure 9 below,
with the z axis along the M-Cl bond. Interestingly, the two
a Calculated geometries (ground-state symmetry and multiplicity in
parentheses) determined at the B3LYP/6-311+G(d) level of theory.
methane arise from coupling and hydrogen atom abstraction,
respectively.
In an effort to trap a potential methylnickel complex through
carbonyl insertion, the above reaction was repeated with an
excess of CO or 13CO gas. The reaction was cleanly diverted
1
to a diamagnetic product. The H and 13C NMR spectra of an
2
2
2
a1-symmetry orbitals most closely approximate dx and dy -z
organic or organometallic CH313C(O)-containing product from
orbitals, in agreement with our description of three-coordinate
iron(II) complexes derived from spectroscopic measurements.17
The energy order of d orbitals is fairly consistent across the
13CO insertion would show 13C-13C coupling in the 13C
1
spectrum and a doublet for the acyl group in the H spectrum.
5
10
2
Neither feature was observed, indicating that no irreversible CO
insertion has taken place. The major product has two isopropyl
resonances, characteristic of a diketiminate complex with C2V
symmetry (only one isopropyl environment; the two methyl
groups are inequivalent, as discussed above). The high sym-
metry, diamagnetism, and the observation of a single isotopically
sensitive carbonyl band (1938 cm-1) in infrared spectra are most
consistent with the nickel(0) complex [LNi(CO)]- as an
assignment for the final product of the reaction of CH3Li and
LNiCl under CO, although conclusive characterization will
require further study.
Comparative Computations on LMCl. Density functional
B3LYP/6-311+G(d) calculations on L′MCl species (L′ )
C3N2H5-) reveal a large manifold of states due to the small
d-orbital splitting engendered by the low-coordination environ-
ment. Table 1 provides a comparison of the calculated L′MCl
and experimental LMCl geometries. In general, the level of
agreement between calculated and experimental structures is
reasonable, apart from the discrepancy between the calculated
(1.922 Å) and experimental (1.815(3) Å) Ni-N bond lengths.
As discussed below, this difference is due to distortion of the
ligand that was not possible in calculations constrained to the
C2V point group. The calculated N-M-N bite angles are
systematically too small by approximately 5°. Test calculations
on full models of L′FeCl (5A1) indicate that replacement of the
appropriate H substituents in L′ with 2,6-diisopropylphenyl and
tert-butyl increases the bite angle by 3°, and therefore, steric
effects appear to be the main contributor to this systematic effect.
This deviation should not detract from the trends that are
analyzed below. The calculated differences in bite angles are
small both in the computations as well as the experimental
structures, but the trends are the same, i.e. the bite angle
increases upon going from L′FeCl to L′CoCl and decreases from
L′CoCl to L′NiCl.
series of divalent metals from Mn (d ) to Zn (d ): dx , dxy
d
<
2
2
2
y -z , dxz < dyz. The dx and dxy orbitals are very close in energy,
making two states (5A1 and 5A2) energetically indistinguishable
for L′FeCl; in the comparisons below, the spectroscopically
5
determined A1 state (dx doubly occupied) is used.17 In L′NiCl,
2
the 4 kcal/mol energy difference favoring the 3A2 state suggests
2
2
double occupation of the dy -z rather than the dxz orbital. Thus,
the predicted ground-state terms for L′MCl are 6A1 (Mn2+), 5A1
(Fe2+), 4A2 (Co2+), 3A2 (Ni2+), 2B2 (Cu2+), and 1A1 (Zn2+). The
calculated ground state for L′CuCl (a close relative of the
recently synthesized LCuCl14e) agrees with that derived from
theory and calculation for (diketiminate)Cu(SCPh3),14c using
different axes.
Three-Coordinate Complexes of Iron, Cobalt, and Nickel:
Geometric Structures. Detailed metrical comparisons are
enabled by the fact that LFeCl, LCoCl, LNiCl, LFeCH3, and
LCoCH3 all have the same crystal packing. Therefore, geometric
differences between the complexes cannot come from the
ubiquitous “crystal packing effects,” and even small changes
should be interpretable. Table 2 compares key bond distances
and angles for the five LMX structures. Spencer et al. have
isolated the copper(II) analogue LCuCl,14e but its crystal
structure has not been reported.
The metal-ligand bond distances monotonically decrease
from iron to cobalt to nickel, as expected from the atomic radii
[rion(FeII-Td) ) 0.77 Å; rion(CoII-Td) ) 0.72 Å; rion(NiII-Td)
) 0.69 Å]. The M-N distances are 0.025-0.028 Å longer in
the methyl complexes, reflecting the greater polarizing ability
of chloride that leads to a more electrophilic metal.
The monodentate chloride and methyl ligands are expected
to accurately illustrate the effect of coordination number on the
metal’s covalent radius. Compared to literature complexes, the
Co-Cl, Ni-Cl, Fe-C, and Co-C distances are amazingly
short. Each M-Cl distance is about two standard deviations
below the mean M-Cl distance in the Cambridge Structural
Database.23 The metal-carbon distances, at 2.01 Å or less, are
Discussion
Three-Coordinate Complexes of Iron, Cobalt, and Nickel:
Synthesis and Electronic Structures. Straightforward metath-
esis reactions with divalent metal halide salts appear to be a
general route into three-coordinate complexes supported by
(23) Cambridge Structural Database and VISTA, October 2001 release. Allen,
R. H.; Kennard, O. Chem. Design Automation News 1993, 8, 31-37.
Histograms generated by VISTA are in the Supporting Information.
9
14420 J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002

Electronically Unsaturated Complexes of Fe, Co, Ni
A R T I C L E S
Table 2. Bond Length and Angle Comparison. X ) Metal-bound C, Cl, or O
dist/angle
M-N (Å)
LFeCl
LFeCH
LCoCl
LCoCH
LNiCl
LNi(THF)
LCo(µ-Cl)2Li(ether)2
3
3
1.948(2)
1.973(1)
2.009(3)
1.902(2)
1.930(1)
1.963(3)
1.815(3)
1.876(1)
1.882(1)
2.000(1)
1.961(2)
1.968(2)
2.2942(9)
2.3023(9)
99.4(1)
M-X (Å)
2.172(1)
2.140(1)
2.137(2)
N-M-N (deg)
N-M-X (deg)
96.35(11)
131.83(5)
94.85(8)
132.57(4)
99.49(13)
130.25(6)
98.24(7)
130.88(3)
97.3(2)
131.4(1)
100.46(6)
133.31(6)
126.21(6)
126.4(1)
129.6(1)
1.16(9)
111.7-118.3
C-N-C (deg)
128.4(2)
128.56(1)
128.0(2)
128.6(1)
130.3(4)
124.8(3)
125.8(3)
16.1(2)
3.00
fold/twista (deg)
N‚‚‚N (Å)
0.49(9)
2.90
0.49(6)
2.91
0.27(10)
2.90
0.38(6)
2.92
7.15(1)
2.72
2.89
a Angle between the least-squares N-C-C-C-N and N-M-N planes of the six-membered diketiminate-metal ring.
trend is also followed in the optimized theoretical models,
suggesting that it is not a crystallographic artifact. There are
two reasonable models through which one can explain the
relative bite angles.
First, a close look at the crystal structures shows that, except
for LNiCl, the distance between the two diketiminate nitrogen
atoms (2.888-2.919 Å) is quite consistent in the LMX
structures, despite the substantial changes in the bond distances.
It is particularly notable that the N‚‚‚N distance is 2.904 ( 0.001
Å in LFeCl and LCoCl, even though the Co-N bonds are
shorter by 0.046 Å. Using a simple trigonometric model with a
rigid diketiminate ligand, one can readily rationalize the change
in bite angle because shorter bonds to a rigid chelating ligand
must increase the angle (see Figure S-1). However, using a rigid
diketiminate and the observed Ni-N distance of 1.815 Å, one
predicts a N-Ni-N angle of about 106°. Presumably the
complex avoids the poor metal-ligand orbital overlap in this
planar geometry by distorting the diketiminate ring. The
observed “twist” distortion maintains the crystallographically
observed C2 symmetry while bringing the nitrogen atoms about
0.018 Å closer together. In this model, the large N-Co-N angle
represents maximum strain on the rigid planarity of the
diketiminate backbone before it succumbs to distortion. This
model also explains the exceptional ability of LCoCl to bind
LiCl (see above): this strain is relieved by folding in the four-
coordinate LCo(µ-Cl)2Li(ether)2.
Figure 8. Mean metal-terminal chloride bond lengths for four- to six-
coordinate complexes compared to the M-Cl bond lengths in three-
coordinate LMCl. Mean bond lengths from VISTA analysis of the
Cambridge Structural Database (Oct 2001).
especially notable: there is one shorter Fe-CH3 bond,24 and
the only shorter Co-CH3 bonds are in cobalt(III) complexes.
The bond lengths to the small monodentate ligand can be used
to illustrate the effect of coordination number on bond distances,
unfettered by steric or chelate effects. For example, Figure 8
shows the trend in mean MII-(terminal Cl) distances for iron,
cobalt, and nickel in various coordination numbers, compared
with our complexes, which are the first three-coordinate terminal
chloride complexes of each metal in the +2 oxidation state.
Despite the extraordinarily low coordination number at the
metal, the Fe-N and Co-N distances are only about one
standard deviation shorter than the mean metal-N distance from
the Cambridge Structural Database. However, the Ni-N
distance is about two standard deviations shorter than the mean.23
This suggests that the rigid diketiminate ligand is causing
anomalous behavior in the iron and cobalt complexes; the
reasons for this behavior will become clear below.
An interesting quantity is the “bite angle” of the ligand,
defined as the N-M-N angle. There is no monotonic trend in
this angle, which interestingly is maximum in the cobalt
complex. The large N-Co-N angle implies that the diketimi-
nate is more sterically congesting in LCoCl, which seems at
odds with the greater tendency of LCoCl to attract lithium
chloride in the synthesis of LCo(µ-Cl)2Li(ether)2. The bite angle
Reproduction of the bite angle trend in calculations in which
C2V symmetry is enforced (out-of-plane distortions are impos-
sible) indicates that other factors may also affect the bite angle.
In particular, an analysis of the frontier orbitals suggests that
the metal d orbital configuration can play a role in determining
the bite angle of â-diketiminate ligands. The d orbital manifold
of L′MCl derived from the density functional calculations is
depicted in Figure 9. From L′FeCl (5A1) to L′CoCl (4A2), an a2
orbital that is primarily comprised of the metal dxy is filled; from
4
3
2
2
L′CoCl ( A2) to L′NiCl ( A2), an a1 orbital (∼dy -z ) is filled.
Although these orbitals are primarily (∼90%) metal in character,
they do contain some ligand character, in particular that derived
from the diketiminate nitrogen atoms. The ligand a2 orbital is
N‚‚‚N antibonding, while the a1 orbital is N‚‚‚N bonding.
Therefore, occupation of the a2 and a1 orbitals leads to an
increase and decrease, respectively, in the bite angle.
Stability of High-Spin Methyl Complexes of the Late
Metals. Low-coordinate alkyl complexes have been postulated
as unobserved, reactive intermediates in C-H and C-C bond
cleaving reactions.25,26 Thus, it is surprising that the complexes
LMCH3 (Fe, 12-electron complex; Co, 13-electron complex)
(24) Balch, A. L.; Olmstead, M. M.; Safari, N.; St. Claire, T. N. Inorg. Chem.
1994, 33, 2815-2822.
9
J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002 14421

A R T I C L E S
Holland et al.
thought to be an intermediate in methanogenesis by methyl-
coenzyme M reductase,31 a very recent computational study
suggests a mechanism with no methyl-nickel bond formation.32
These considerations spurred us to attempt formation of LNiCH3
by reactions analogous to those used to create LFeCH3 and
LCoCH3, but ethane, methane, and a nickel(I) product were
observed.
In one mechanistic possibility, LNiCH3 is formed and then
•
decomposes through homolysis to LNi and CH3 . Observation
of ethane as a byproduct strongly supports the formation of
methyl radicals. Homolysis of NiII-CH3 bonds is well-known,33
and this facile decomposition pathway for high-spin organo-
metallic complexes presumably is responsible for the absence
of trigonal or tetrahedral nickel(II) methyl complexes in the
1
literature.28,33,34 Low-temperature H NMR spectroscopy has
been useful in observing nickel-containing intermediates,35 but
LNiCl is practically insoluble in THF at low temperatures. By
briefly warming the solution and reintroducing it to the NMR
probe at -80 °C, it was possible to observe an intermediate
with resonances at δ 66 and -47 that builds up to very low
concentration before converting to LNi(THF). Its peaks are
relatively narrow but paramagnetically shifted, suggesting a
high-spin Ni(II) formulation, but it could not be identified
conclusively, and its concentration is low enough that other
spectroscopic methods have not been pursued.
In the absence of direct spectroscopic evidence, we sought
to trap LNiCH3 by thawing the THF solution under an
atmosphere of carbon monoxide. Insertion of CO into the Fe-C
bond of the analogue LFeCH3 is rapid.36 However, attempted
alkylation of LNiCl under carbon monoxide yielded a product
tentatively assigned as [LNi(CO)]-. With no evidence to support
transient formation of LNiCH3, we must conclude that the
simplest reasonable mechanism, electron transfer from CH3Li
to LNiCl followed by LiCl formation and THF coordination to
nickel, is operative.
Figure 9. Frontier orbital diagram for L′MCl derived from density
functional calculations. L′ ) parent diketiminate ligand with no alkyl or
aryl substituents.
are thermally stable in solution. In his review of open-shell
organometallic complexes, Poli notes that many decomposition
pathways are invalidated by the absence of empty d-orbitals in
high-spin complexes,22 a contention supported by the analogous
stability of Parkin’s 14-electron methyliron complex (PhTptBu)-
FeCH3.27 Tris(pyrazolyl)borate and related ligands also lead to
tetrahedral 15-electron cobalt(II)-methyl complexes.28,29
The reactivity of paramagnetic methylnickel complexes is also
relevant to nickel-containing enzymes. Methylnickel intermedi-
ates have been postulated in the A cluster of acetyl-coenzyme
A synthase.30 Although a methylnickel intermediate is generally
Conclusions
This report shows that â-diketiminate ligands offer a general
route to robust three-coordinate complexes of group 8, 9, and
10 metals. The three-coordinate complexes are unusual in that
one of the ligands can be small, and this ligand may be
substituted easily. The steric protection even stabilizes highly
electronically unsaturated organometallic complexes, and thus
it has been possible to synthesize the first two three-coordinate
methyl complexes of the transition metals. Detailed structural
comparison of five isomorphous three-coordinate complexes,
in combination with theory, has given us preliminary insights
into their electronic structure, a subject of future detailed
spectroscopic studies. Finally, systematic comparison has shown
that paramagnetic nickel(II)-alkyl complexes are not formed
with the method that generates stable iron(II) and cobalt(II)
(25) Recent examples: (a) Plutino, M. R.; Scolaro, L. M.; Romeo, R.; Grassi,
A. Inorg. Chem. 2000, 39, 2712-2720. (b) Gandelman, M.; Vigalok, A.;
Konstantinovski, L.; Milstein, D. J. Am. Chem. Soc. 2000, 122, 9848-
9849. (c) Kanzelberger, M.; Singh, B.; Czerw, M.; Krogh-Jespersen, K.;
Goldman, A. S. J. Am. Chem. Soc. 2000, 122, 11017-11018. (d) Crumpton,
D. M.; Goldberg, K. I. J. Am. Chem. Soc. 2000, 122, 962-963. (e) Fekl,
U.; Goldberg, K. I. J. Am. Chem. Soc. 2002, 124, 6804-6805. (f) Review:
Stahl, S. S.; Labinger, J. A.; Bercaw, J. E. Angew. Chem., Int. Ed. Engl.
1998, 37, 2180-2192.
(26) Recent evidence suggests that associative mechanisms without three-
coordinate intermediates may be more reasonable in some cases: (a) Cao,
Z.; Hall, M. B. Organometallics 2000, 19, 3338-3346. (b) Johansson, L.;
Tilset, M. J. Am. Chem. Soc. 2001, 123, 739-740. (c) Zhong, H. A.;
Labinger, J. A.; Bercaw, J. E. J. Am. Chem. Soc. 2002, 124, 1378-1399.
(d) Procelewska, J.; Zahl, A.; van Eldik, R.; Zhong, H. A.; Labinger, J.
A.; Bercaw, J. E. Inorg. Chem. 2002, 41, 2808-2810.
(31) (a) Ermler, U.; Grabarse, W.; Shima, S.; Goubeaud, M.; Thauer, R. K.
Science 1997, 278, 1457-1462. (b) Horng, Y.-C.; Becker, D. F.; Ragsdale,
S. W. Biochemistry 2001, 40, 12875-12885.
(27) Kisko, J. L.; Hascall, T.; Parkin, G. J. Am. Chem. Soc. 1998, 120, 10561-
(32) Pelmenschikov, V.; Blomberg, M. R. A.; Siegbahn, P. E. M.; Crabtree, R.
H. J. Am. Chem. Soc. 2002, 124, 4039-4049.
10562.
(28) Schebler, P. J.; Mandimutsira, B. S.; Riordan, C. G.; Liable-Sands, L. M.;
Incarvito, C. D.; Rheingold, A. L. J. Am. Chem. Soc. 2001, 123, 331-
332.
(33) (a) Jolly, P. W.; Jonas, K.; Krueger, C.; Tsay, Y. H. J. Organomet. Chem.
1971, 33, 109-122. (b) D’Aniello, M. J.; Barefield, E. K. J. Am. Chem.
Soc. 1976, 98, 1610-1611.
(29) Jewson, J. D.; Liable-Sands, L. M.; Yap, G. P. A.; Rheingold, A. L.;
Theopold, K. H. Organometallics 1999, 18, 300-305.
(30) Recent discussions: Lindahl, P. A. Biochemistry 2002, 41, 2097-2105.
(b) Seravalli, J.; Kumar, M.; Ragsdale, S. W. Biochemistry 2002, 41, 1807-
1819.
(34) Diamagnetic square-planar complexes of nickel(II) are common.
(35) A recent example: Shultz, C. S.; DeSimone, J. M.; Brookhart, M. J. Am.
Chem. Soc. 2001, 123, 9172-9173.
(36) Smith, J. M.; Lachicotte, R. L.; Holland, P. L. Organometallics 2002, 21,
4808-4814.
9
14422 J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002

Electronically Unsaturated Complexes of Fe, Co, Ni
A R T I C L E S
mmol) in THF (5 mL) and stirred for 4 h. Volatile materials were
removed from the dark green mixture under vacuum, and the residue
was extracted with 2:1 diethyl ether/pentane (17 mL), filtered, and
concentrated to 8 mL. Cooling to -35 °C gave brown crystals of
LCo(µ-Cl)2Li(ether)2 (458 mg, 48% yield). 1H NMR (400 MHz,
C6D6): δ 50 (br), 22.2, 20 (br), 12.0, 8.8, 3.8, 2.1, 1.5, 0.4, -7, -40,
-43. µeff (C6D6, 295 K)) 4.7 µB. IR (toluene): 1619 (w), 1535 (m),
1492 (s), 1432 (m), 1384 (vs), 1362 (vs), 1317 (vs), 1217 (m), 1102
(w), 1049 (m) cm-1. UV-vis (toluene, ꢀ in mM-1 cm-1): 338 (12),
512 (0.50), 538 (0.47), 689 (0.40) nm. Anal. Calcd for C43H71N2O2-
CoLiCl2: C, 65.81, H, 9.12, N, 3.57. Found: C, 65.86, H, 9.02, N,
3.35.
analogues, an observation that may be relevant to the mechanism
of enzymatic methanogenesis.
Experimental Section
General Considerations. All 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 150 °C overnight. Proton NMR data were
recorded on Bruker Avance 400 or Bruker AMX-400 spectrometers
(400 MHz) at 22 °C. Shifts are reported in ppm, relative to residual
protiated solvent in C6D6 (δ 7.15), CH2Cl2 (δ 5.31), or THF-d8 (δ 1.73,
3.58). In parentheses are listed: integrations, T2 values in ms calculated
LTl. LLi(THF) (742 mg, 1.28 mmol) was added to a stirred solution
of thallium ethoxide (Aldrich, 319 mg, 1.28 mmol) in pentane (13 mL).
After several hours, the yellow mixture was pumped down, extracted
with toluene (60 mL), and filtered. Volatile materials were removed
from the filtrate, and the residue was washed with cold pentane to give
550 mg (61% yield) of yellow powder. This powder appeared pure by
NMR spectroscopy, but sensitivity to trace air or to light prevented
trustworthy elemental analysis (the material turned black in transit).
1H NMR (400 MHz, C6D6): δ 7.03(m, 6, aryl), 5.31(s, 1, backbone),
3.34 (septet, 4, iPr), 1.29(d, 12, iPr), 1.27(s, 18, tBu), 1.2(br s, 12, iPr).
13C NMR (400 MHz, C6D6): δ 169.7, 140.3, 123.6, 33.2, 27.7, 23.2.
IR (toluene): 1533 (m), 1390 (s), 1366 (m), 1327 (w), 1218 (w), 1146
(w), 1096 (w) cm-1. UV-vis (toluene, ꢀ in mM-1 cm-1): 376(18) nm.
A preliminary X-ray crystal structure showed monomeric LTl but had
apparent twinning problems that prevented accurate solution of the
structure.
-1 20
as (π∆ν1/2
)
,
and assignments. Carbon-13 NMR spectra of the
paramagnetic compounds showed substantially fewer peaks than
expected and were not pursued. T1 measurements and DQF-COSY
were recorded on a Varian INOVA-500 spectrometer (500 MHz) at
30.0 °C; T1 values were calculated using a curve-fitting procedure.37
IR spectra were recorded on a Mattson Instruments 6020 Galaxy Series
FTIR; solution spectra used a cell with CsF windows. EPR spectroscopy
at 77 K used a quartz Dewar suspended in the cavity of a Bruker ESP-
300 spectrometer; parameters are given in the Supporting Information.
UV-vis spectra were measured on a Cary 50 spectrophotometer, using
cuvettes sealed to a Schlenk-type stopcock. Solution magnetic suscep-
tibilities were determined by Evans’ method,38 and are considered
accurate to (0.3 µB. Microanalysis was performed by Desert Analytics
(Tucson, AZ).
Pentane, diethyl ether, methylene chloride, tetrahydrofuran (THF),
and toluene were purified by passage through activated alumina and
“deoxygenizer” columns from Glass Contour Co. (Laguna Beach, CA).
Deuterated benzene, THF-d8, and 2-methyltetrahydrofuran were dried
over CaH2 and then over Na and then vacuum distilled into a storage
container or directly into the NMR tube. Carbon monoxide (CP grade)
was obtained from Air Products, and 13CO (99 atom %) was from
Aldrich Chemical Co. Anhydrous metal salts were prepared by treatment
of the hydrates with thionyl chloride.39 The MCl2(THF)n complexes
were prepared using the method of Kern,40 and NiCl2 was converted
to a THF adduct using an analogous method.41 Celite was dried at 200
°C under vacuum. The lithium salt LLi(THF) was prepared by treating
a solution of LH42 in THF with 1.0 equiv of butyllithium, removing
solvent, and crystallizing from pentane. Methyllithium (Aldrich, 1.4
M in diethyl ether) was transferred in a 8-10 mL portion to a 20-mL
scintillation vial, treated with several drops of dioxane to precipitate
halide impurities,43 filtered through Celite, and pumped down to a white
solid for use. Methylmagnesium chloride (Aldrich, 3 M in THF) was
used as received. The preparation of LFeCl and LFeCH3 has been
described previously.16,17 A solution of CD3MgI for the preparation of
LFeCD3 and LCoCD3 was generated by stirring a solution of CD3I
(Aldrich) in diethyl ether with Mg metal. Methyl-d3 compounds
synthesized from this Grignard reagent solution were spectroscopically
identical to their protiated analogues.
LCoCl. A solution of LTl (126 mg, 0.178 mmol) in THF (5 mL)
was added to a slurry of CoCl2(THF)1.5 (42 mg, 0.176 mmol) in THF
(5 mL), stirred for 2.5 h, and then filtered. Volatile materials were
removed to leave a brown residue that was dissolved in diethyl ether
(2 mL). Crystals formed upon standing at room temperature for several
hours. Total yield: 42 mg (40%). 1H NMR (400 MHz, C6D6): δ 60.8
(4, 4, m-aryl), 27.6 (18, 5, tBu), -2.8 (12, 6, iPr “back” methyl), -46.0
i
i
(2, 6, p-aryl), -57.4 (4, 1, Pr methine), -84.4 (12, 2, Pr “front”
methyl), -90.7 (1, 3, backbone). µeff (C6D6, 295 K) ) 4.7 µB. UV-
vis (toluene, ꢀ in mM-1 cm-1): 358(6.8), 510 (sh, ∼0.2), 652 (0.10)
nm. Mp: 240 °C (dec). Anal. Calcd for C35H53N2CoCl: C, 70.51, H,
8.96, N, 4.70. Found: C, 69.30, H, 9.12, N, 4.81.
LCoCH3. Methylmagnesium chloride (0.12 mL, 0.36 mmol) was
added to a solution of LCo(µ-Cl)2Li(ether)2 (216 mg, 0.276 mmol) in
diethyl ether (7 mL), leading to immediate formation of a white
precipitate. After 0.5 h, the mixture was filtered, volatile materials were
removed from the filtrate under vacuum, and the residue was extracted
with 1:1 diethyl ether/pentane (20 mL). The solution was concentrated
to 8 mL, filtered, and cooled to -35 °C to give brown crystals of
1
LCoCH3. H NMR (400 MHz, C6D6): δ 78.1 (4, 11, m-aryl), 27.2
i
(18, 23, tBu), 4.1 (12, 27, Pr “back” methyl), -27.6 (2, 18, p-aryl),
i
i
-53.8 (4, 2, Pr methine), -67.1 (1, 6, backbone), -118 (12, 4, Pr
“front” methyl). µeff (C6D6, 295 K) ) 4.9 µB. IR (toluene): 1511 (m),
1433 (m), 1383 (vs), 1363 (vs), 1319 (s), 1220 (m), 1198 (w), 1098
(w) cm-1. UV-vis (toluene, ꢀ in mM-1 cm-1): 335(15), 565(0.13),
725(0.16) nm. Anal. Calcd for C36H56CoN2: C, 75.10, H, 9.80, N, 4.87.
Found: C, 74.78, H, 9.68, N, 4.80.
Assignment of 1H NMR Signals in LFeCH3.17 [shift in ppm
(integration, T1 in ms, T2 in ms, assignment)] δ 133.2 (1, <1, 0.4,
backbone), 42.8 (18, 2, 2, tBu), -3.1 (4, 6, 5, m-aryl), -28.3 (12, 6,
5, “back” iPr methyl), -108.1 (2, 4, 3, p-aryl), -118.2 (4, <1, 0.2, iPr
i
methine), -129.5 (12, <1, 0.4, “front” Pr methyl).
LNiCl. A Schlenk flask was loaded with NiCl2(THF)0.7 (1.75 g, 10.2
mmol), LLi(THF) (3.72 g, 6.41 mmol), and THF (30 mL) and heated
to 70 °C overnight. Volatile materials were removed from the dark
green mixture under vacuum, and the residue was extracted with CH2Cl2
(70 mL), filtered, and concentrated to 15 mL. Addition of diethyl ether
(10 mL) and cooling to -35 °C gave green crystals of LNiCl. A second
crop of crystals was collected to give a total yield of 2.5 g (66%).1H
LCo(µ-Cl)2Li(ether)2. A solution of LLi(THF) (722 mg, 1.24 mmol)
in THF (5 mL) was added to a slurry of CoCl2(THF)1.5 (291 mg, 1.22
(37) Evaluation of T1 from an inversion-recovery experiment is described:
Levy, G. C.; Peat, I. R. J. Magn. Reson. 1975, 18, 500-521.
(38) Schubert, E. M. J. Chem. Educ. 1992, 69, 62.
(39) Pray, A. R. Inorg. Synth. 1990, 28, 321-323.
(40) Kern, R. J. J. Inorg. Nucl. Chem. 1962, 24, 1105-1109.
(41) Eckert, N. A.; Bones, E. M.; Lachicotte, R. L.; Holland, P. L., submitted.
(42) Budzelaar, P. H. M.; van Oort, A. B.; Orpen, A. G. Eur. J. Inorg. Chem.
1998, 1485-1494.
i
NMR (400 MHz, CD2Cl2): δ 29.3 (4, 14, m-aryl), 18.5 (4, 11, Pr
methine), 5.6 (12, 19, iPr methyl), 4.1 (12, 18, iPr methyl), 2.5 (18, 21,
tBu), -14.7 (2, 14, p-aryl), -123.5 (1, 11, backbone). µeff (CD2Cl2,
295 K) ) 3.1 µB. IR (CH2Cl2): 2967 (vs), 2869 (m), 1587 (w), 1521
(43) Holland, P. L.; Smith, M. E.; Andersen, R. A.; Bergman, R. G. J. Am.
Chem. Soc. 1997, 119, 12815-12823.
9
J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002 14423

A R T I C L E S
Holland et al.
(m), 1464 (m), 1385 (m), 1366 (s), 1315 (s), 1268 (vs), 1221 (w), 1099
(w), 1057 (w) cm-1. UV-vis (CH2Cl2, ꢀ in mM-1 cm-1): 372 (8.4),
467 (sh, 1.7), 838 (1.4) nm. Anal. Calcd for C35H53N2NiCl: C, 70.54,
H, 8.96, N, 4.70. Found: C, 69.25, H, 9.02, N, 4.63.
System equipped with a normal focus molybdenum-target X-ray tube
operated at 2.0 kW (50 kV, 40 mA). A total of 1321 frames of data
(1.3 hemispheres) were collected using a narrow frame method with
scan widths of 0.3° in ω and exposure times of 30-60 s/frame using
a detector-to-crystal distance of 5.09 cm (maximum 2θ angle of 56.6°).
Peaks were integrated to a maximum 2θ angle of 56.5° with the Siemens
SAINT program. The final unit cell parameters (at -80 °C) were
determined from the least-squares refinement of three-dimensional
centroids of up to 8192 intense reflections. The data were corrected
for absorption with the SADABS program.
Space groups were assigned using the XPREP program. The
structures were solved using direct methods and refined employing full-
matrix least-squares on F2. All nonhydrogen atoms were refined with
anisotropic thermal parameters. Hydrogen atoms were included in
idealized positions with riding thermal parameters. Details on each
structure may be found in the Supporting Information.
LNi(THF). A mixture of LNiCl (295 mg, 0.495 mmol) and CH3Li
(14 mg, 0.64 mmol) was dissolved in THF (10 mL). After stirring the
orange solution for 2 h, volatile materials were removed. The residue
was extracted with pentane (10 mL), filtered, and reduced to 6 mL
under vacuum. Cooling to -35 °C gave brown crystals of LNi(THF).
A second crop was also collected, for a total yield of 0.18 g (58%). 1H
NMR (400 MHz, THF-d8): δ 24 (sh), 21.5, 20 (sh), 12.5, 3.9, -1.8,
-11.5, -50 (sh); there was also a noticeable lump in the baseline from
about +60 to -60 ppm, suggesting that other resonances are too broad
to be observed as peaks. µeff (THF-d8, 295 K) ) 2.0 µB. EPR (9.425
GHz, 77K, MeTHF): g ) 2.068, 2.11, 2.51. IR (Nujol): 1504 (s),
1427 (s), 1402 (vs), 1366 (s), 1322 (s), 1220 (m), 1152 (m), 1097 (m)
cm-1. UV-vis (THF, ꢀ in mM-1 cm-1): 310 (25), 412 (5.8), 428 (5.5),
485 (1.8), 520 (sh, 1.5) nm. Anal. Calcd for C39H61N2NiO: C, 74.05,
H, 9.72, N, 4.43. Found: C, 73.66, H, 9.53, N, 4.76.
Computational Methods. Calculations were carried out using the
Gaussian program.44 The all-electron 6-311+G(d) basis set was used
in conjunction with the B3LYP hybrid functional for all calculations
described herein. The restricted Kohn-Sham formalism was used.
Geometry optimizations were carried out within C2V symmetry.
Reaction of LNiCl with CH3Li. Roughly 15 mg of LNiCl and 2
mg of CH3Li were placed in a J. Young NMR tube, and THF-d8 (0.8
mL) was condensed into the tube at 77 K. At -80 °C, the chloride
complex does not dissolve in THF, but brief warming caused the liquid
phase to change to a light orange color. The tube was immediately
Acknowledgment. P.L.H. thanks the University of Rochester
and the National Science Foundation (CHE-0134658) for
funding. L.L.P. thanks REACH and NSF-REU (CHE-9987991)
for summer support. T.R.C. acknowledges the National Science
Foundation (CHE-9983665) for partial support of this research.
We thank Phil Power for sharing ref 15b prior to publication,
Emile Bominaar for insightful discussions, and Jeremy Smith
for synthesizing NiCl2(THF)0.7 as well as helpful criticism. We
are indebted to Brandy Russell and Kara Bren for help with T1
measurements, to Sandip Sur for assistance with NMR spec-
troscopy, and to an anonymous reviewer for suggesting methyl-
tetrahydrofuran as a solvent for EPR spectroscopy.
1
cooled to -80 °C for several minutes. H NMR spectroscopy at -80
°C showed the presence of starting materials, products, and additional
peaks at δ 66 and -47 ppm. Warming the solution to room temperature
for several minutes caused these two peaks to disappear, and in the
final mixture only LNi(THF), excess CH3Li, ethane, and methane (see
1
below) were detectable by H NMR.
Spectroscopic Yield and Detection of Byproducts. A mixture of
solid LNiCl (22 mg, 0.037 mmol) and solid CH3Li (0.8 mg, 0.04 mmol)
was placed in a J. Young NMR tube, and THF-d8 (0.6 mL) was
condensed into the tube at 77 K. Warming gave a deep orange solution.
The 1H NMR spectrum of this solution was indistinguishable from that
of LNi(THF), except for singlets at δ 0.86 and 0.18 ppm. The volatile
materials were vacuum-distilled into another NMR tube. The 1H NMR
spectrum of the colorless distillate consisted of solvent peaks, trace
water, and singlets at δ 0.86 and 0.18 ppm.
Supporting Information Available: Bond length comparisons
from the Cambridge Structural Database, EPR spectrum of
LNi(THF), DQF-COSY spectra of LNiCl, DFT optimized
geometries for L′MCl, and crystallographic information files.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Trapping with CO. The reaction of LNiCl and CH3Li was
performed as above, but 1 atm of carbon monoxide was admitted into
the tube before thawing the mixture. Warming gave an orange-yellow
solution that was analyzed spectroscopically. DANGER: A small
amount of Ni(CO)4 may be produced in this reaction, and the reaction
mixture should be handled with care in a fume hood. IR (THF
solution): 1968, 1938 cm-1 (with 13CO: 1966, 1893 cm-1). 1H NMR
(400 MHz, THF-d8): δ 6.91 (d, 4, m-aryl), 6.80 (t, 2, p-aryl), 4.94 (s,
backbone), 3.24 (septet, 4, iPr methine), 2.27 (s, 2-3, unassigned), 1.26
JA025583M
(44) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A. Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Cui, Q., Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; Nanayakkara, A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. GAUSSIAN; Gaussian, Inc.:
Pittsburgh, PA, 1998.
i
i
(d, 12, Pr methyl), 1.14 (s, 18, tBu), 1.01 (d, 12, Pr methyl).
X-ray Crystallography. Crystals or shards of crystals were mounted
under Paratone-8277 on a glass fiber and placed in a cold nitrogen
stream at -80 °C on the X-ray diffractometer. The X-ray intensity
data were collected on a standard Siemens SMART CCD Area Detector
9
14424 J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002