Multielectron redox activity facilitated by metal - Metal interactions in early/late heterobimetallics: Co/Zr complexes supported by phosphinoamide ligands

Article abstract of DOI:10.1021/ic900552b

To assess the effect of dative M→M interactions on redox properties in early/late heterobimetallic complexes, a series of Co/Zr complexes supported by phosphinoamide ligands have been synthesized and characterized. Treatment of the Zr metalloligands (Ph₂PNⁱPr)₃ZrCl (1), (ⁱPr₂PNMes)₃ZrCl (2), and (ⁱPr ₂PNⁱPr)₃ZrCl (3) with Col₂ leads to reduction from Co^II to Co^I and isolation of the heterobimetallic complexes ICo(Ph₂PN⁽Pr)₃ZrCl (4), ICo(′Pr₂PNMes)₃ZrCl (5), and ICo-( ⁱPr₂PNⁱPr)3ZrCl (6), respectively. Interestingly, treatment of Col₂ with the phosphinoamine Ph ₂PNHⁱPr in the absence of a bound Zr center leads to the disubstituted Co^II complex (Ph₂PNHⁱPr) ₂Col₂ (7). The tris(phosphinoamine) Co^I complex (Ph₂PNH′Pr)₃Col (8) can only be generated in the presence of an added reductant such as Zn⁰, indicating that the reduction of Co^II to Co^I only occurs in the presence of Zr in the formation of complexes 4-6. Structural characterization of 4-6 reveals a Zr-Co interaction, with interatomic distances of 2.7315(5) A, 2.6280(5) A, and 2.6309(5) A, respectively. This distance appears to decrease as the phosphine donors on Co become more electron-releasing, strengthening the dative Co→Zr interaction. Cyclic voltammetry of 4-6 shows that all three compounds can be further reduced by two electrons at relatively mild reduction potentials (-1.65 V to -2.07 V vs Fc/Fc⁺). The potentials at which these reductions occur in each of these complexes are largely governed by the extent to which electron-density is donated to Zr, as well as the electron-donating ability of the phosphine substituents. Moreover, cyclic voltammetry of complex 8 reveals that in the absence of Zr, the Co center is significantly more electron rich, and thus more difficult to reduce. Chemical reduction of 5 leads to the isolation of the two-electron reduced dinitrogen complex [N₂Co(ⁱPr₂PNMes)₃ZrX] [Na(THF)₅] (9). X-ray crystallography of 9 reveals that two-electron reduction is accompanied by a significant contraction of the Co-Zr interatomic distance from 2.6280(5) A to 2.4112(3) A. These heterobimetallic complexes have been studied computationally using density functional theory to examine the nature of the metal-metal interactions in these species.

Full text of DOI:10.1021/ic900552b

Inorg. Chem. 2009, 48, 6251–6260 6251
DOI: 10.1021/ic900552b
Multielectron Redox Activity Facilitated by Metal-Metal Interactions in Early/Late
Heterobimetallics: Co/Zr Complexes Supported by Phosphinoamide Ligands
Bennett P. Greenwood, Scott I. Forman, Gerard T. Rowe, Chun-Hsing Chen, Bruce M. Foxman, and
Christine M. Thomas*
Department of Chemistry, Brandeis University, 415 South Street, Waltham, Massachusetts 02454
Received March 23, 2009
To assess the effect of dative MfM interactions on redox properties in early/late heterobimetallic complexes, a series of
Co/Zr complexes supported by phosphinoamide ligands have been synthesized and characterized. Treatment of the
Zr metalloligands (Ph2PNiPr)3ZrCl (1), (iPr2PNMes)3ZrCl (2), and (iPr2PNiPr)3ZrCl (3) with CoI2 leads to reduction from
CoII to CoI and isolation of the heterobimetallic complexes ICo(Ph2PNiPr)3ZrCl (4), ICo(iPr2PNMes)3ZrCl (5), and ICo-
(iPr2PNiPr)3ZrCl (6), respectively. Interestingly, treatment of CoI2 with the phosphinoamine Ph2PNHiPr in the absence of a
bound Zr center leads to the disubstituted CoII complex (Ph2PNHiPr)2CoI2 (7). The tris(phosphinoamine) CoI complex
(Ph2PNHiPr)3CoI (8) can only be generated in the presence of an added reductant such as Zn0, indicating that the
reduction of CoII to CoI only occurs in the presence of Zr in the formation of complexes 4-6. Structural characterization
˚
˚
˚
of 4-6 reveals a Zr-Co interaction, with interatomic distances of 2.7315(5) A, 2.6280(5) A, and 2.6309(5) A, respectively.
This distance appears to decrease as the phosphine donors on Co become more electron-releasing, strengthening the
dative CofZr interaction. Cyclic voltammetry of 4-6 shows that all three compounds can be further reduced by two
electrons at relatively mild reduction potentials (-1.65 V to -2.07 V vs Fc/Fc+). The potentials at which these reductions
occur in each of these complexes are largely governed by the extent to which electron-density is donated to Zr, as well as
the electron-donating ability of the phosphine substituents. Moreover, cyclic voltammetry of complex 8 reveals that in
the absence of Zr, the Co center is significantly more electron rich, and thus more difficult to reduce. Chemical reduction
of 5 leads to the isolation of the two-electron reduced dinitrogen complex [N2Co(iPr2PNMes)3ZrX][Na(THF)5] (9). X-ray
crystallography of 9 reveals that two-electron reduction is accompanied by a significant contraction of the Co-Zr
˚
˚
interatomic distance from 2.6280(5) A to 2.4112(3) A. These heterobimetallic complexes have been studied computa-
tionally using density functional theory to examine the nature of the metal-metal interactions in these species.
Introduction
metal-metal interactions in early/late heterobimetallics are
analogous to metalfborane interactions that have received
recent attention in metalloboratrane complexes.2 These types
of Z-type bonding interactions, in which a metal center datively
donates an electron-pair to a ligand atom, are intriguing in that
the effective oxidation state of the metal center is higher than
that predicted as the formal oxidation state, as result of electron
withdrawal by the Z-type ligand.3
In recent years, dinuclear complexes containing two differ-
ent transition metal centers have received increasing interest
owing to the presumption that their reactivity should vary
substantially from that of monometallic complexes or homo-
bimetallic complexes.1 In particular, early/late heterobimetallic
complexes featuring interactions between a hard, Lewis acidic
early metal center and a soft, Lewis basic late metal center are
particularly interesting targets in light of their two vastly
different reaction sites which lead to potential applications as
bifunctional catalysts for the activation of small molecule
substrates and homogeneous catalysis. While a number of
early/late heterobimetallic complexes have been synthesized,
the nature of the metal-metal interactions in these com-
plexes and the effect of these interactions on redox behavior
and reactivity remains largely unexplored. Moreover, the
Early/late heterobimetallic complexes linked by alkoxyalkyl-
phosphines [Ph2PCH2O]- were synthesized by Wolczanski and
(2) (a) Figueroa, J. S.; Melnick, J. G.; Parkin, G. Inorg. Chem. 2006, 45,
7056. (b) Hill, A. F.; Owen, G. R.; White, A. J. P.; Williams, D. J. Angew.
Chem., Int. Ed. 1999, 38, 2759. (c) Sircoglou, M.; Bontemps, S.; Bouhadir,
G.; Saffon, N.; Miqueu, K.; Gu, W.; Mercy, M.; Chen, C.-H.; Foxman,
B. M.; Maron, L.; Ozerov, O. V.; Bourissou, D. J. Am. Chem. Soc. 2008, 130,
16729. (d) Bontemps, S.; Sircoglou, M.; Bouhadir, G.; Puschmann, H.;
Howard, J. A. K.; Dyer, P. W.; Miqueu, K.; Bourissou, D. Chem.;Eur. J.
2008, 14, 731–740. (e) Sircoglou, M.; Bontemps, S.; Mercy, M.; Saffon, N.;
Takahashi, M.; Bouhadir, G.; Maron, L.; Bourissou, D. Angew. Chem., Int.
Ed. 2007, 46, 8583.
*To whom correspondence should be addressed. E-mail: thomasc@
brandeis.edu.
(1) (a) Stephan, D. W. Coord. Chem. Rev. 1989, 95, 41. (b) Wheatley, N.;
(3) (a) Green, M. L. H. J. Organomet. Chem. 1995, 500, 127–148.
Kalck, P. Chem. Rev. 1999, 99, 3379.
(b) Parkin, G. Organometallics 2006, 25, 4744.
r
2009 American Chemical Society
Published on Web 06/05/2009
pubs.acs.org/IC

6252 Inorganic Chemistry, Vol. 48, No. 13, 2009
Greenwood et al.
co-workers several decades ago in an effort to model the
interactions between electron-rich late metals and early metal
oxide surfaces.4 More recently, Nagashima et al. have utilized
titanium and zirconium complexes supported by phosphino-
amides, [Ph2PNR]- (where R=iPr or tBu), as metalloligands
for late transition metal centers.5 The disubstituted titanium
complex (Ph2PNtBu)2TiCl2 was utilized as a bidentate ligand
in NiII, PdII, and PtII allyl complexes, and it was found that the
MfTi interactions in these complexes reduced the elec-
tron density at the late transition metal and, in turn,
greatly enhanced the electrophilicity of the M-bound allyl
moiety.5d In addition, the trisubstituted zirconium complexes
(Ph2PNR)2ZrCl were coordinated to Mo0(CO)3 and CuICl
fragments in a tridentate, tripodal fashion. Nagashima and
co-workers found that in the (CO)3Mo(Ph2PNiPr)3ZrCl com-
plex, the CO stretching frequencies are shifted 50-70 cm-1
higher than comparable (CO)3Mo(triphos) complexes as a
result of weaker back-bonding into CO π* orbitals.5c In light
of this, we were curious to ascertain what effect coordination
of a Lewis acidic early transition metal would have on a
potentially redox-active late transition metal. Such a dative
M1fM2 interaction would be expected to reduce the elec-
tron density at M1, and therefore could facilitate reduction
and small molecule activation at milder potentials. In a rel-
ated study, Berry and co-workers have examined hetero-
bimetallic interactions in CrtCr-Fe complexes, and found
that the metal-metal multiply bonded moieties donate
electron density to the Fe center, resulting in a trans influence
and significantly lowering the FeIII/II potential in these com-
plexes.6
Scheme 1
phosphine dissociation occurs in solution at room temp-
erature.5c Although only one 31P NMR signal is observed
at all temperatures, variable temperature H NMR spectra
1
of 2 and 3 confirm that these complexes exhibit the same
dynamic behavior, making these derivatives viable metalloli-
gands as well.8 Accordingly, treatment of 1, 2, and 3 with 1
equiv of CoI2 led to isolation of heterobimetallic Zr/Co
complexes as shown in Scheme 1. Surprisingly, the cobalt
centers in the green paramagnetic products ICo(Ph2-
PNiPr)3ZrCl (4), ICo(iPr2PNMes)3ZrCl (5), and ICo(iPr2P-
NiPr)3ZrCl (6), respectively, had been formally reduced from
CoII to CoI. The oxidation states of 4, 5, and 6 were confirmed
by their solution magnetic moments corresponding to S=1
(μeff = 2.92, 2.87, and 3.10 μB, respectively). Evidence for
CofZr interactions in 4-6 provided by structural char-
acterization using X-ray crystallography, computational stu-
dies using density functional theory (DFT), and dramatic
effects on redox activity will be discussed later in this
manuscript.
In attempts to synthesize derivatives featuring other
metal combinations, we found that reactions of 1-3 with
metal salts with a less accessible MII/I redox potential such
as FeCl2, FeI2, and ZnCl2 did not lead to heterobimetallic
complexes. In addition, CoCl2 also was an ineffective
starting material for the preparation of mixed metal
derivatives. In light of this halide dependence, it is likely
that the reducing agent is I-, and the resulting I2 is
subsequently consumed via reaction with the phosphin-
es in 1-3.9-11 Nonetheless, it is evident that the close
proximity of the Lewis acidic Zr center enforced by
the phosphinoamide ligand geometry necessitates reduc-
tion of the late transition metal center to a more Lewis
basic +1 oxidation state upon or prior to coordination.
To provide further evidence that the reduction to CoI in
Results and Discussion
Synthesis of Co/Zr Heterobimetallic Complexes 4-6.
For this investigation, we chose to explore heterobimetal-
lics with a series of phosphinoamide ligands of differ-
ing electronic properties. The phosphinoamine precursors
i
Ph2PNHiPr, Pr2PNHMes (Mes = 2,4,6-trimethylphenyl),
i
and Pr2PNHiPr were prepared using modifications of
literature procedures.7 In a procedure similar to that pre-
viously described, deprotonation of the phosphinoamines
with 1 equiv of nBuLi at -78 °C, followed by treatment with
1/3 equiv of ZrCl4, led to isolation of the zirconium metal-
loligands (Ph2PNiPr)3ZrCl (1), (iPr2PNMes)3ZrCl (2),
and (iPr2PNiPr)3ZrCl (3) in moderate to good yield.5c
Previously, Nagashima and co-workers showed through
X-ray crystallography and variable temperature 1H NMR
that the phosphines in 1 are unequivocally bound to the Zr
center at low temperature and in the solid state, while reversible
(8) Additional experimental details are included in the Supporting
Information for this manuscript.
(9) The potential for the 2I- f I2 + 2e- couple is -0.14 V vs ferrocene (in
MeCN) as reported in: Connelly, N. G.; Geiger, W. E. Chem. Rev. 1996, 96,
877–910.
(4) (a) Ferguson, G. S.; Wolczanski, P. T. Organometallics 1985, 4, 1601.
(b) Ferguson, G. S.; Wolczanski, P. T. J. Am. Chem. Soc. 1986, 108.
(10) I2 reacts stoichiometrically with LiN(PPh2)2 to effect P-P coupling:
Braunstein, P.; Hasselbring, R.; Tiripicchio, A.; Ugozzoli, F. J. Chem. Soc.,
Chem. Commun. 1995, 37–38. We have also confirmed that I2 is consumed
quantitatively by both our phosphinoamines and our zirconium phosphino-
amide complexes. Stoichiometric reaction of Ph2PNHiPr with I2 in C6D6
leads to complete consumption of both I2 (based on disappearance of the
ꢀ
ꢀ
(c) Ferguson, G. S.; Wolczanski, P. T.; Parkanyi, M. C.; Zonnevyille,
M. C. Organometallics 1988, 7, 1967. (d) Baxter, S. M.; Ferguson, G. S.;
Wolczanski, P. T. J. Am. Chem. Soc. 1988, 110, 4231.
(5) (a) Nagashima, H.; Sue, T.; Oda, T.; Kanemitsu, A.; Matsumoto, T.;
Motoyama, Y.; Sunada, Y. Organometallics 2006, 25, 1987. (b) Sunada, Y.;
Sue, T.; Matsumoto, T.; Nagashima, H. J. Organomet. Chem. 2006, 691,
3176. (c) Sue, T.; Sunada, Y.; Nagashima, H. Eur. J. Inorg. Chem. 2007,
2897. (d) Tsutsumi, H.; Sunada, Y.; Shiota, Y.; Yoshizawa, K.; Nagashima,
H. Organometallics 2009, 28, 1988.
(6) (a) Nippe, M.; Berry, J. F. J. Am. Chem. Soc. 2007, 129, 12684.
(b) Nippe, M.; Victor, E.; Berry, J. F. Eur. J. Inorg. Chem. 2008, 5569.
(7) (a) Sisler, H. H.; Smith, N. L. J. Org. Chem. 1961, 26, 611.
(b) Poetschke, N.; Nieger, M.; Khan, M. A.; Niecke, E.; Ashby, M. T.
Inorg. Chem. 1997, 36, 4087.
diagnostic purple color) and Ph2PNHiPr (based on disappearance of its 31
P
NMR signal) to form a highly insoluble red/orange material. Likewise,
stoichiometric reaction of 2 with I2 leads to complete consumption of both
materials and appearance of new phosphorus-containing products.
(11) Preliminary data suggests that the reaction of 2 with CoCl2 indeed
leads to a heterobimetallic product in the presence of an external reductant
such as NaBH4, based on the similarity of 1H NMR features to those of 5 in
the crude NMR of this reaction.

Article
Inorganic Chemistry, Vol. 48, No. 13, 2009 6253
Scheme 2
in reported covalent radii for Zr and Co exist in the
literature,12 the Zr-Co interatomic distances in 4-6 are
shorter than the sum of the covalent radii of Zr and Co
determined using most methods. More explicit evidence for
a bond comes from comparison of the Zr-Co distances in
4-6 with those in the only two Zr-Co bonds that have
been structurally determined to date, 2.705(1) and 2.617(1)
A in complexes [SiMe2N(4-CH3C6H4)]3Zr-Co(CO)3L (L
= CO and PPh3, respectively), indicate that the Zr-Co
distances in 4-6 are within the range of legitimate Zr-Co
bonds.13 It is worth noting that these previously character-
ized Zr-Co bonds, unlike those in 4-6, are not supported
or enforced by a ligand framework. In complexes 4-6, it is
possible that the close proximity of the Zr and Co atom is
enforced by ligand geometry and is not the result of orbital
overlap between Zr and Co; however, electrochemical
consequences and DFT calculations, discussed later in this
manuscript, suggest that Zr-Co interactions exist in 4-6
and are dative in nature, with the electron-rich, formally
CoI center donating electrons to the Lewis-acidic ZrIV
center. The shorter Co-Zr interatomic distances in 5 and
6 compared to 4 can be attributed to the more electron-rich
iPr phosphine substituents that impart more electron den-
sity on the cobalt center, strengthening its donation to Zr.
Assessment of the Redox Behavior of Heterobimetallic
Co/Zr Complexes 4-6 Using Cyclic Voltammetry. As
interatomic distances determined by X-ray crystallography
suggest a CofZr interaction which is expected to decrease
the electron density at Co in 4, 5, and 6, the effect of this
interaction on the redox properties of cobalt in these com-
plexes was probed using cyclic voltammetry (CV). Inter-
estingly, the CVs of 4-6 showed reductive events, although
the nature of these reductions was different for each species
(Figure 2B). These reductions are assigned as cobalt-based,
since CVs of the zirconium complexes 1-3 did not contain
any features within the tetrahydrofuran (THF) solvent
window. Complex 4 showed a fully reversible reduction
at -1.65 V. Although this appears to be a single reduction,
even by differential pulse voltammetric (DPV) techniques,
bulk electrolysis of a sample of 4 at -1.77 V required
∼2 electrons per molecule of 4 (1.94 e-). Thus, the rever-
sible reduction in the CV of 4 is assigned as a 2e- process, as
confirmed by bulk chemical reduction (vide infra).
complexes 4-6 is unequivocally a result of the Zr’s
presence, the reaction of CoI2 with the phosphinoamine
Ph2PNHiPr was examined in the absence of a tethered
Zr center (Scheme 2). Indeed, treatment of CoI2 with
excess Ph2PNHiPr led exclusively to the red disubstituted
CoII product (Ph2PNHiPr)2CoI2 (7), as confirmed struc-
turally and spectroscopically.8 The CoI trisubstituted
complex (Ph2PNHiPr)3CoI (8) could be synthesized, but
only in the presence of elemental Zn as a reducing agent.
Much like 4-6, complex 8 is green and paramagnetic,
with a solution magnetic moment indicative of high spin
CoI ( μeff=2.89 μB).
Structural Characterization of Co/Zr Heterobimetallic
Complexes 4-6. Single crystals of complexes 4, 5, and 6
were grown and subjected to X-ray diffraction. The
molecular structures of 4-6 are shown in Figure 1 and
relevant interatomic distances and angles are tabulated in
Table 1. All three complexes have C3v symmetry with an
essentially tetrahedral geometry at the Co center. The
P-Co-P angles in all three complexes are within 5° of the
expected 109.5° angle between ligands in an ideal tetra-
hedron. Furthermore, the average P-Co-P angles in
4-6 are only slightly expanded (∼2-4°) from the P-
Co-P angles in monometallic complex 8. Thus, there is
no significant structural distortion from tetrahedral geo-
metry to trigonal bipyramidal geometry at Co in com-
plexes 4-6 as a result of interactions between Co and Zr.
However, the geometry about Zr clearly indicates a
structural distortion toward trigonal bipyramidal geome-
try in all three complexes. The average N-Zr-N angles
in 4-6 are very close to 120° (118.6°, 119.5°, and 119.5°,
respectively) and the sum of the three N-Zr-N angles is
356-359°, as would be expected for the angles in an ideal
trigonal bypyramid. In addition, the average N-Zr-Co
angles in 4-6 are close to 90° (83.2°, 86.2°, and 86.3°,
respectively), consistent with Co occupying the axial
position of the trigonal bipyramidal Zr coordination
sphere. Notably, a search of the Cambridge Structural
Database (CSD) indicated that this sort of geometry
about Zr, (3 L-Zr-L angles of 120° ( 2°) is unprece-
dented in the absence of a fifth donor ligand.
The CV of complex 5 also possessed a reversible reduc-
tion at -1.64 V; however, in this case the reduction
was followed by an immediate, irreversible reduction at
-1.87 V. This can best be explainedby anEEC mechanism,
in which the first one-electron reduction is reversible, but
(12) Examples of discrepancies in covalent radii reported in the literature:
(a) 1.75 A +1.26 A=3.01 A (for l.s. Co) or 1.75 A + 1.50 A=3.25 A:
ꢀ
ꢀ
Cordero, B.; Gomez, V.; Platero-Prats, A. E.; Reves, M.; Echeverrıa,
´
ꢀ
J.; Cremades, E.; Barragan, F.; Alvarez, S. Dalton Trans. 2008, 2832.
::
(b) 1.54 A + 1.11 A=2.65 A: Pyykko, P.; Atsumi, M. Chem.;Eur. J.
2009, 15, 186. (c) 1.45 A + 1.16 A=2.61 A: Pauling, L. The Nature of
the Chemical Bond, 3rd ed.; Cornell University Press: Ithaca, NY, 1960.
(d) 1.56 A + 1.33 A=2.89 A: Cambridge Structural Database, http://www.
ccdc.cam.ac.uk/products/csd/radii/table.php4#symbol. (e) If the Co-Co
(2.506 A) and Zr-Zr (3.179 A) distances in elemental cobalt and zirconium
metal are divided by 2 to extract a typical covalent radius in a metal-metal
bond, we obtain 1.59 A + 1.25 A=2.84 A: In Table of Interatomic Distances
and Configuration in Molecules and Ions, Supplement 1956-1959, Special
publication No. 18; Sutton, L. E., Ed.; Chemical Society: London, U.K.,
1965.
As shown in Figure 1 and Table 1, complexes 4, 5, and
6 have Zr-Co interatomic distances of 2.7315(5), 2.6280
(5), and 2.6309(5) A, respectively. While large variations
(13) Jansen, G.; Schubart, M.; Findeis, B.; Gade, L. H.; Scowen, I. J.;
McPartlin, M. J. Am. Chem. Soc. 1998, 120, 7239.

6254 Inorganic Chemistry, Vol. 48, No. 13, 2009
Greenwood et al.
Figure 1. Displacement ellipsoid (50%) representations of 4, 5, and 6. All hydrogen atoms and substituents on all but one phosphinoamide ligand have
been omitted for clarity.
Table 1. Relevant Interatomic Distances (A) and Angles (deg) in Complexes 4-6
and 8
4
5
6
8
Zr-Co
Zr-N
2.7315(5)
2.107(2)
2.100(2)
2.101(2)
2.2933(7)
2.3116(7)
2.2964(7)
2.4226(7)
2.5434(4)
2.6280(5)
2.129(2)
2.131(2)
2.6309(5)
2.109(3)
2.120(7)
2.133(2)
2.096(8)
Co-P
2.3645(10)
2.3613(9)
2.3806(8)
2.4450(7)
2.5837(4)
2.3418(8)
2.334(10)
2.369(10)
2.4508(10)
2.5497(4)
2.2109(5)
2.2109(5)
2.2109(5)
Zr-Cl
Co-I
Cl-Zr-Co 178.94(2)
I-Co-Zr
P-Co-P
2.5455(3)
178.95(2)
171.331(15) 174.86(2)
177.70(5)
177.63(18)
107.4(2)
111.5(2)
108.4(5)
118.6(3)
122.4(3)
117.72(12)
86.37(8)
86.0(3)
109.46(3)
109.07(3)
104.89(3)
117.82(8)
119.80(8)
118.31(8)
83.74(5)
83.45(6)
82.60(6)
106.42(3)
111.78(3)
107.07(3)
117.97(9)
121.5(9)
119.15(9)
86.38(7)
85.70(6)
86.51(6)
105.296(13)
105.296(13)
105.296(13)
N-Zr-N
N-Zr-Co
86.5(3)
upon addition of a second electron, a chemical process
occurs to generate a new species that cannot be reoxidized.
Since electrochemical experiments were performed under
an N2 atmosphere, we propose that the features in the CV
of 5 are attributed to rapid, irreversible N2 binding upon
reduction of 5 by 2e- (this is further confirmed by bulk
chemical reduction, vide infra).
Interestingly, we find that while the phosphine substit-
uents on 4 and 5 should vary in electron-donating ability
enough to affect the reduction potential of these Co
centers, the reduction potentials for the first reduction
of 4 and 5 are identical. It seems likely that this is a direct
consequence of the stronger Zr-Co interaction in com-
plex 5 based on the ∼0.1 A shorter Zr-Co interatomic
distance. The CV of complex 6, which has a Zr-Co
distance identical to that of 5, shows quasi-reversible
reductions at -1.86 and -2.07 V. In this case, the reduc-
tions are presumably shifted to more negative potentials
than those of 5 as a result of the more electron-releasing
iPr amide substituents.
Figure 2. Full CV of complex 4 (A) and reductive portion of the CV of 4,
5, 6, and8(B)(2mMin0.4M[nBu4N][PF6] in THF; scan rate = 100 mV/s).
two irreversible reductive events (Figure 2B). However,
these reductions occurred at -2.46 and -2.65 V, nearly
1 V more negative than the reduction potentials of the
corresponding Co/Zr complex 4. In light of this, it can
be concluded that in complex 4 the electron-density on
cobalt is significantly diminished with respect to 8, thermo-
dynamically favoring the reduced product, and raising
the reduction potential. Thus, dative donation from
Co to Zr has a dramatic effect on the redox activity of
cobalt, allowing further reduction of the CoI center at
milder potentials.
To compare the reduction potentials of 4-6 with
a similar compound lacking an electron-withdrawing
Zr center, the CV of complex 8 was obtained.8 As obser-
ved for the heterobimetallic complexes, the CV of 8 showed

Article
Inorganic Chemistry, Vol. 48, No. 13, 2009 6255
Table 2. Oxidation and Reduction Potentials of Complexes 4, 5, 6, and 8 as
Determined Using Cyclic Voltammetrya
4
5
6
8
CoIII/CoII
CoII/CoI
CoI/Co0
0.53 Vd
0.32 Vd
0.60 Vd
0.47 Vd
0.25 Vd
-0.07 Vd
-1.86 Vc
-2.07 Vc
0.57d
-2.49c
-2.66c
-1.65 Vb
-1.65 Vb
-1.64 Vb
-1.87 Vc
Co0/Co-I
a 2 mM analyte in 0.4 M [nBu4N][PF6] in THF; scan rate = 100 mV/s.
c
d
b E1/2 (reversible). Epc (irreversible or quasi-reversible). Epa (irrever-
sible).
Scheme 3
Figure 3. Diplacement ellipsoid (50%) representation of 9. Hydrogen
atoms have been omitted for clarity. Relevant bond distances (A) and
angles (deg): Zr1-Co1, 2.4112(3); Co1-N4, 1.8186(16); N4-N5, 1.120(2);
Zr1-I1, 2.9722(7); Zr1-Cl1, 2.690(7); Na1-I1, 3.2695(12); Co1-P1,
2.1209(5); Co1-P2, 2.2153(5); Co1-P3, 2.2098(5); Zr1-N1, 2.1524;
Zr1-N2, 2.1559(15); Zr1-N3, 2.1639(15); I1-Zr1-Co1, 178.560(17);
N4-Co1-Zr1, 179.33(5); N1-Zr1-N2, 117.80(6); N2-Zr1-N3, 121.16(6);
N1-Zr1-N3, 119.82(6); N1-Zr1-Co1, 96.35(4); N2-Zr1-Co1,
86.73(4); N3-Zr1-Co1, 85.93(4); P1-Co1-P2, 112.404(19); P2-Co1-
P3, 113.80(2); P3-Co1-P1, 113.24(2).
In addition, it is worth noting that the CVs of 4-6
in THF showed only irreversible CoII/CoI oxidation
events, consistent with the inability to isolate ZrIV/CoII
species prior to reduction (Table 2).8 For example, the
CV of 4 has only irreversible oxidative features at 0.32
and 0.53 V, implying that a CoII species is unstable
(Figure 2A).
Two-Electron Reduction of Heterobimetallic Co/Zr
Complex 5 and Binding of Dinitrogen to the Reduced
Species. As predicted by cyclic voltammetry, bulk sam-
ples of complexes 4-6 can be chemically reduced with
excess Na/Hg in THF solution, resulting in a color change
from green to red/brown. Notably, in all three complexes,
addition of just 1 equiv of Na/Hg led to a 1:1 mixture of
starting material and reduced product, indicating that
reduction by two electrons is favored. As shown in
Scheme 3, the product generated upon treatment of
5 with excess Na/Hg is diamagnetic and contains a
dinitrogen moiety bound to cobalt, as indicated by IR
spectroscopy (ν(N2)=2023 cm-1).
(avg. P-Co-P angles are 108° and 113° in 5 and 9,
respectively), indicative of a distortion toward a trigonal
bipyramid.
Both the relatively short N-N distance of 1.120(2) A
(in free N2, N-N=1.0975 A) and the relatively high ν(N2)
(2023 cm-1 vs 2331 cm-1 in free N2) suggest very little acti-
vation of the N2 unit in complex 9 despite the seemingly low
formal oxidation state (Co-I).14 Moreover, in comparing the
N-N distance and ν(N2) stretch for 9 with parameters for
previously reported Co-N2 complexes, it is found that the
degree of N2 activation in 9 is most consistent with CoI-N2
complexes, which have N-N distances ranging from 1.08 A
to 1.15 A and ν(N2) bands in the 2024-2125 cm-1 range.15
In contrast, Co-N2 complexes in the formal -1 or 0
oxidation state typically have lower ν(N2) stretching frequen-
cies (1830-1910 cm-1) and longer N-N distances (1.15 A to
1.19 A).15a,16 We attribute this to the involvement of the filled
Co d orbitals in both σ and π-donation to Zr, disfavoring
backbonding into the N-N antibonding orbitals of coordi-
nated N2, as supported by our DFT calculations (vide infra).
Bourissou and co-workers note a similar shift to higher
stretching frequency as a result of a dative RhfB interaction
The structure of this complex, [N2Co(iPr2PN-
Mes)3ZrX][Na(THF)5] (9, where X = a mixture of Cl
and I) was determined via X-ray crystallography and is
shown in Figure 3. Upon reduction, the halide remaining
on Zr exchanges with the NaI in solution, leading to
significant crystallographic disorder between Cl- and I-
which has been modeled adequately with occupancies of
73% I and 27% Cl in the structure of 9. In addition to
halide loss and dinitrogen binding, some noteworthy
structural changes occur upon reduction of 5 to 9. Most
interestingly, the Zr-Co distance in 9 is 2.4112(3) A,
more than 0.2 A shorter than the 2.6280(5) A distance in
5. This contraction of the intermetallic distance suggests
that the increase in electron density upon reduction of Co
leads to a stronger dative donation from Co to the Lewis
acidic Zr center. The geometry about the Zr center
continues to be trigonal bipyramidal, as was observed in
the structure of 5. While the Co center is still best
described as tetrahedral, the P-Co-P angles are slightly
expanded upon strengthening the Co-Zr interaction
(14) Mackay, B. A.; Fryzuk, M. D. Chem. Rev. 2004, 104, 385.
(15) (a) Betley, T. A.; Peters, J. C. J. Am. Chem. Soc. 2003, 125, 10782. (b)
Enemark, J. H.; Davis, B. R.; McGinnety, J. A.; Ibers, J. A. Chem. Commun.
1968, 96. (c) Davis, B. R.; Payne, N. C.; Ibers, J. A. Inorg. Chem. 1969, 2719.
(d) Bianchini, C.; Peruzzini, M.; Zanobini, F. Organometallics 1991, 10,
3415. (e) Bianchini, C.; Mealli, C.; Meli, A.; Peruzzini, M.; Zanobini, F.
J. Am. Chem. Soc. 1988, 110, 8725. (f) Egan, J. W.; Haggerty, B. S.;
Rheingold, A. L.; Sendlinger, S. C.; Theopold, K. H. J. Am. Chem. Soc.
1990, 112, 2445. (g) Fout, A. R.; Basuli, F.; Fan, H.; Tomaszewski, J.;
Huffman, J. C.; Baik, M.-H.; Mindiola, D. J. Angew. Chem., Int. Ed. 2006,
45, 3291.
(16) (a) Yamamoto, A.; Miura, Y.; Ito, T.; Chen, H. L.; Iri, K.; Ozawa, F.;
Miki, K.; Sei, T.; Tanaka, N.; Kasai, N. Organometallics 1983, 2, 1429. (b)
Hammer, R.; Klein, H.-F.; Schubert, U.; Frank, A.; Huttner, G. Angew.
Chem., Int. Ed. Engl. 1976, 15, 612. (c) Cecconi, F.; Ghilardi, C. A.;
Midollini, S.; Moneti, S.; Orlandini, A.; Bacci, M. J. Chem. Soc., Chem.
Commun. 1985, 731.

6256 Inorganic Chemistry, Vol. 48, No. 13, 2009
Greenwood et al.
Figure 4. Calculated molecular orbital diagram of the frontier orbitals of 5 with orbital visualizations derived from natural orbital population analysis.
in their complex RhCl(CO)(diphosphine-borane).2d It is also
noteworthy that the Zr-X distance (in both the case of Cl
Table 3. Selected Bond Lengths (A) and Angles (deg) of 5 as Determined by
Crystallography and Theoretical Calculations
experimental (X-ray)
calculated (DFT)a
and I) is significantly elongated from that observed in typical
Zr halide complexes. While this is due, at least in part, to the
strong interaction with the Na+ countercation, it is also
likely that the electron rich Co center exerts a strong
trans influence on Zr. Reduction of complexes 4 and 6 with
Na/Hg also leads to diamagnetic red/brown products with
Zr-Co
Zr-Cl
2.6280(5)
2.4450(7)
2.3645(10)
2.3613(9)
2.3806(8)
2.129(2)
2.131(2)
2.133(2)
178.95(2)
2.666
2.440
2.378
2.384
2.380
2.136
2.147
2.132
178.75
Co-P1
Co-P2
Co-P3
Zr-N1
Zr-N2
Zr-N3
Cl2-Zr-Co
dinitrogen bound to Co (ν(N2) = 2019 and 1992 cm-1
,
respectively), and the structures and characterization of these
products will be reported in a future work.
Theoretical Investigation of Metal-Metal Interactions
in Co/Zr Heterobimetallics using DFT. To better under-
stand the interaction between the cobalt and zirconium
atoms in this series of complexes, theoretical calculations
were performed on an analogue of compound 5 and the
reduced dinitrogen adduct, 9. Geometry optimization
and frequency calculations were carried out using the
BP86 functional and a mixed basis set consisting of
LANL2TZ(f) (Co,Zr), 6-311+G(d) (Cl, N, P), and D95V
(C, H) basis sets. Although the full molecules were used
for our computational analyses starting from coordinates
determined via X-ray crystallography, the I atom in 5 was
replaced with Cl for simplicity.
The results of geometry optimization of 5 using DFT are
in excellent agreement with experiment, as shown in
Table 3. In accordance with our assignment of the cobalt
atom as existing in the +1 oxidation state, the unpaired
spin density resides almost exclusively on the cobalt atom
(Supporting Information, Figure S14) and Mulliken po-
pulation analysis predicts a spin density of 1.87 on the
a Gaussian ’03: BP86/LANL2TZ(f)/6-311+G(d)/D95 V.
cobalt atom in 5. Visualization of the frontier molecular
orbitals of 5 reveals several interesting features relating to
the interaction between the cobalt and zirconium atoms
(Figure 4). Rather than the three-over-two d-orbital
splitting pattern that would be predicted for a tetrahedral
cobalt center, theory predicts a significant lowering of the
energy of the Co dz2 orbital in 5. This stabilization appears
to result from substantial orbital overlap between the Co
and Zr dz2 orbitals. Thus, donation of electron density
from the filled Co dz2 orbital to the empty Zr dz2 orbital
results in a net bonding interaction. In addition to this
sigma bonding orbital, close inspection of the frontier
orbitals also reveals two orbitals corresponding to the
Co dxz and dyz orbitals that exhibit π-type bonding inter-
actions with the Zr center. Furthermore, Mayer popula-
tion analysis reveals a Co-Zr bond index of 0.73 as a result
of these three interactions. Thus, electron density is clearly

Article
Inorganic Chemistry, Vol. 48, No. 13, 2009 6257
+
Figure 5. Calculated molecular orbital diagram of the frontier orbitals of 9 with visualizations of the corresponding molecular orbitals. The Na(THF)5
unit has been omitted for clarity.
drawn away from the Co center by the Lewis acidic Zr
atom, and it is to this dative-like interaction that we
attribute the facile two-electron reduction of 5.
As evident in the molecular orbital surfaces in Figure 5,
the molecular orbitals that house the Co-N2 backbond-
ing interactions are also involved in the π-type inter-
actions between Co and Zr, and it appears that this
diminishes the amount of electron density available for
back-bonding into the N2 π* orbitals.
In computational studies of 9, it was necessary to
include the THF-coordinated Na+ countercation to ob-
tain a geometry consistent with that determined using
X-ray diffraction data. The calculated Co-Zr inter-
atomic distance agrees well with the experimental value
(2.41 A vs. 2.42 A, respectively), as does the N-N bond
distance of the ligated dinitrogen moiety (1.12 A vs.
1.14 A, respectively). Moreover, the dinitrogen stretching
frequency of 2025 cm-1 derived from the frequency
calculation is nearly identical to the 2023 cm-1 frequency
observed in the infrared spectrum of 9.
The 0.23 A reduction in the Co-Zr distance that accom-
panies the two-electron reduction of 5 to form 9 is intriguing,
as it suggests that a stronger bonding interaction is forming
between the two metal centers. Accordingly, the Co-Zr
Mayer bond index is increased to 1.23 in 9. As with complex
5, visualization of the frontier orbitals reveals a discrete σ-
bonding interaction between the two metal centers mediated
by the Co and Zr dz2 orbitals as well as two π-type bonding
interactions originating from the dxz and dyz orbitals on Co
(Figure 5). In addition to these interactions, examination of
the highest occupied molecular orbitals (HOMOs), to which
the two additional electrons have been added, reveals a
noticeably greater bonding contribution between Co and
Zr which was not present in 5. We attribute the significantly
higher Zr-Co bond order and shorter Zr-Co distance to
this new bonding contribution.
Conclusion
In summary, we have shown that tethering a Lewis acidic
early transition metal such as Zr in close proximity to an
electron-rich late transition metal leads to metal-metal
interactions which are dative in nature. Evidence for CofZr
interactions in 4-6 is provided by structural characterization
data and computational studies using DFT. DFT calcula-
tions reveal that the interaction between Co and Zr occurs
through both σ and π overlap of the metal d orbitals.
Donation of electron density from Co to Zr has dramatic
effects on the redox activity of Co, causing shifts of nearly
1 V to milder reduction potentials. The Zr center in these
complexes plays the role of a Z-type metalloligand,3 render-
ing these complexes analogous to recently reported “metal-
laboratranes” featuring MfB dative bonds. Much like
transition metal boranes, the assignments of a formal oxida-
tion to the metal centers in 4-6 and 9 is not informative. This
is particularly evident in complex 9, which would tradition-
ally be assigned an oxidation state of Co-I but has structural
and spectroscopic features more consistent with a CoI center.
The polarized nature of dative metal-metal interactions such
as those presented in this manuscript may prove advantage-
ous in the activation of small molecule substrates. Further-
more, the ability of dative interactions in heterobimetallic
complexes to facilitate multielectron redox activity at signi-
ficantly milder potentials is a promising lead in the design of
catalysts for redox transformations. Future studies will focus
on the utility of these particular heterobimetallic complexes
for small molecule activation reactions and also on new more
flexible ligand systems to provide more opportunity for
reversible metal-metal bond formation.
With respect to the bonding interactions between
Co and N2, the elongation and diminished stretching
frequency of the dinitrogen bond with respect to gaseous
N2 suggests activation of the N-N bond upon binding to
the metal center, but to a far lesser extent than is typically
seen in such low valent late transition metal dinitrogen
complexes. The results of the theoretical investigation of
9 give some insight into the nature of this phenomenon.

6258 Inorganic Chemistry, Vol. 48, No. 13, 2009
Greenwood et al.
Experimental Section
Table 4. X-ray Diffraction Experimental Details for ICo(Ph2PNiPr)3ZrCl (4),
ICo(iPr2PNMes)3ZrCl (5), and ICo(iPr2PNiPr)3ZrCl (6)
General Considerations. All syntheses reported were carried
out using standard glovebox and Schlenk techniques in the
absence of water and dioxygen, unless otherwise noted. Ben-
zene, pentane, diethyl ether, tetrahydrofuran, dichloromethane,
and toluene were degassed and dried by sparging with N2 gas
followed by passage through an activated alumina column.
Ethanol was dried over CaH2 and distilled, followed by sparging
with N2 prior to use. All solvents were stored over 3 A molecular
sieves. Deuterated benzene and toluene were purchased from
Cambridge Isotope Laboratories, Inc., degassed via repeated
freeze-pump-thaw cycles, and dried over 3 A molecular sieves.
THF-d8 was dried over Na/K alloy, vacuum-transferred, and
degassed via repeated freeze-pump-thaw cycles. Solvents were
frequently tested using a standard solution of sodium benzophe-
none ketyl in tetrahydrofuran to confirm the absence of oxygen
and moisture. Ph2PNHiPr,7 and complex 15c were synthesized
using literature procedures. All other chemicals were purchased
from Aldrich, Strem, or Alfa Aesar and used without further
purification. NMR spectra were recorded at ambient tempera-
ture unless otherwise stated on a Varian Inova 400 MHz
4
5
6
chemical
formula
fw
T (K)
λ (A)
a (A)
C
59H67Cl-
CoIN3P3Zr
1223.63
120
0.71073
10.8403(14)
14.4482(19)
18.620(2)
74.452(5)
82.637(5)
86.324(6)
2785.1(6)
P-1
C
45H75Cl-
CoIN3P3Zr
1063.54
120
0.71073
27.8496(8)
53.0394(14)
15.0536(4)
90
C
27H63Cl-
CoIN3P3Zr
835.25
120
0.71073
15.1913(8)
18.1727(10)
13.4668(7)
90
b (A)
c (A)
R (deg)
β (deg)
γ (deg)
V (A3)
space group
Z
90
90
90
90
22236.1(10)
Fdd2
16
1.271
12.05
0.0332, 0.0867
3717.7(3)
Pnma
2
1.459
4
1.492
Dcalc (g/cm3)
μ (cm-1
)
12.13
0.0299, 0.0778
17.78
0.0300, 0.0620
R1, wR2a
(I > 2σ(I))
P
P
P
P
a R1 = ||Fo| - |Fc||/ |Fo|, wR2 = { w(Fo2 - Fc2)2/ w(Fo2)2}1/2
.
1
instrument. H and 13C NMR chemical shifts were referenced
to residual solvent. 31P NMR chemical shifts were referenced to
85% H3PO4. IR spectra were recorded on a Varian 640-IR
spectrometer controlled by Resolutions Pro software. UV-vis
spectra were recorded on a Cary 50 UV-vis spectrophotometer
using Cary WinUV software. Elemental microanalyses were
performed by Atlantic Microlab, Inc., Norcross, GA. Solution
magnetic moments were measured using Evans’ method.17
X-ray Crystallography Procedures. All operations were per-
formed on a Bruker-Nonius Kappa Apex2 diffractometer, using
graphite-monochromated Mo KR radiation. All diffractometer
manipulations, including data collection, integration, scaling,
and absorption corrections were carried out using the Bruker
Apex2 software.18 Preliminary cell constants were obtained
from three sets of 12 frames. Structures were solved using
SIR-92,19 and refined (full-matrix-least-squares) using the Ox-
ford University Crystals for Windows program.20 All ordered
non-hydrogen atoms were refined using anisotropic displace-
ment parameters; hydrogen atoms were fixed at calculated
geometric positions and allowed to ride on the corresponding
carbon atoms. Crystallographic parameters for all structures
reported are summarized in Tables 4 and 5. Further crystal-
lographic details may be found in the Supporting Information
file and the accompanying CIF files.
Table 5. X-ray Diffraction Experimental Details for (Ph2PNHiPr)2CoI2 (7),
(Ph2PNHiPr)3CoI (8), and [N2Co(iPr2PNMes)3ZrX][Na(THF)5] (9)
7
8
9
chemical
formula
fw
T (K)
λ (A)
a (A)
C
30H36-
CoI2N2P2
C45H54
CoIN3P3
915.70
-
C65H115Cl0.27CoI0.73
N5NaO5P3Zr
1415.34
120
-
799.32
120
120
0.71073
20.6816(13)
17.3232(11)
19.5909(12)
90
113.823(3)
90
6420.8(7)
P 1 21/c 1
8
0.71073
14.8343(4)
14.8343(4)
74.260(2)
90
0.71073
14.7610(9)
24.6060(15)
19.1884(11)
90
90.284(3)
90
6969.3(7)
P 1 21/n 1
4
b (A)
c (A)
R (deg)
β (deg)
γ (deg)
V (A3)
space group
Z
90
120
14152.0(7)
R-3c
12
1.289
11.51
Dcalc (g/cm3) 1.654
1.349
8.47
μ (cm-1
)
25.79
0.0207, 0.0469 0.0227, 0.0542 0.0330, 0.0721
R1, wR2a
(I > 2σ(I))
P
P
P
P
a R1 = ||Fo| - |Fc||/ |Fo|, wR2 = { w(Fo2 - Fc2)2/ w(Fo2)2}1/2
.
Electrochemistry. Cyclic voltammetry measurements were
carried out in a glovebox under a dinitrogen atmosphere in a
one-compartment cell using a CH Instruments electrochemical
analyzer. A glassy carbon electrode and platinum wire were used
as the working and auxiliary electrodes, respectively. The re-
ference electrode was Ag/AgNO3 in THF. Solutions (THF) of
electrolyte (0.40 M [nBu4N][PF6]) and analyte (2 mM) were also
prepared in the glovebox.
A standard H-type bulk electrolysis cell was used for bulk
electrolysis measurements. The cathodic chamber was charged
with crystalline complex 4 (20.6 mg, 0.0168 mmol) and both
chambers were filled with 10 mL of a 0.1 M [nBuN4][PF6]
solution in THF. Controlled potential electrolysis commenced
at -1.77 V until the solution current was less than 0.5% of the
initial value (3.14 C passed, in 62 min). This corresponds to 1.94
electron equivalents if you take into account the three toluene
solvent molecules per complex in crystals of 4.
Computational Details. All calculations were performed
using Gaussian03-E.0121 for the Linux operating system. DFT
calculations were carried out using a combination of Becke’s
1988 gradient-corrected exchange functional22 and Perdew’s
1986 electron correlation functional23 (BP86). For open shell
systems, unrestricted wave functions were used in energy calcu-
lations. A mixed-basis set was employed, using the LANL2TZ(f)
triple-ζ basis set with effective core potentials for cobalt and
zirconium,24-27 Gaussian03’s internal 6-311+G(d) for atoms
bonded directly to the metal centers (nitrogen, phosphorus,
(21) Frisch, M. J. et al. Gaussian, Inc.: Wallingford, CT, 2004.
(22) Becke, A. D. Phys. Rev. A 1988, 38, 3098.
(23) Perdew, J. P. Phys. Rev. B 1986, 33, 8822.
(24) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299–310.
(25) Roy, L. E.; Hay, P. J.; Martin, R. L. J. Chem. Theory Comput. 2008,
4, 1029–1031.
(17) (a) Sur, S. K. J. Magn. Reson. 1989, 82, 169. (b) Evans, D. F. J. Chem.
Soc. 1959, 2003.
(18) Apex2, Version 2 User Manual, M86-E01078; Bruker Analytical
X-ray Systems: Madison, WI, 2006.
::
::
(19) Altomare, A.; Cascarano, G.; Giacovazzo, G.; Guagliardi, A.; Burla,
M. C.; Polidori, G.; Camalli, M. J. Appl. Crystallogr. 1994, 27, 435.
(20) Betteridge, P. W.; Carruthers, J. R.; Cooper, R. I.; Prout, K.;
Watkin, D. J. J. Appl. Crystallogr. 2003, 36, 1487.
(26) Ehlers, A. W.; Bohme, M.; Dapprich, S.; Gobbi, A.; Hollwarth, A.;
::
Jonas, V.; Kohler, K. F.; Stegmann, R.; Veldkamp, A.; Frenking, G. Chem.
Phys. Lett. 1993, 208, 111–114.
(27) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270–283.

Article
Inorganic Chemistry, Vol. 48, No. 13, 2009 6259
and chlorine), and Gaussian03’s internal LANL2DZ basis
set (equivalent to D95 V28) for carbon and hydrogen. Using
crystallographically determined geometries as a starting point,
when available, the geometries were optimized to a minimum,
followed by analytical frequency calculations to confirm that no
imaginary frequencies were present. Mayer bond order analy-
sis29 was performed with the routines included in the Gaussi-
an03 package.
(161.8 MHz, C6D6): δ=9.49 (s). 13C NMR (100.5 MHz, C6D6):
δ=148.7, 133.3, 131.9, 129.3, 34.5, 21.9, 21.7, 20.8. Anal. Calcd
for C45H75ClN3P3Zr: C, 61.58; H, 8.61; N, 4.79. Found: C,
61.09; H, 8.47; N, 4.61.
i
(iPr2PNiPr)3ZrCl (3). A solution of Pr2PNHiPr (2.589 g,
10.06 mmol) in Et2O (100 mL) was cooled to -78 °C. To this was
added nBuLi (6.33 mL, 1.6 M in hexanes, 10.1 mmol) dropwise
over 10 min. The resulting yellow/orange solution was warmed
to room temperature and stirred for 3 h. The solution was then
cooled back to -78 °C, and ZrCl4 (0.800 g, 3.43 mmol) was
added portionwise as a solid. This reaction mixture was then
warmed to room temperature and stirred for 12 h. Volatiles were
removed from the solution in vacuo, and the resulting solids
were extracted with CH2Cl2 (20 mL) and filtered through a pad
of Celite to remove LiCl. The volume of the filtrate was reduced
to 5 mL in vacuo. The resulting supersaturated solution was
layered with pentane and cooled to -35 °C for 12 h. The mother
liquor was decanted to yield analytically pure colorless crystals
of 3. The decanted mother liquor was removed in vacuo, and the
resulting solid was repeatedly washed with pentane to precipi-
tate a pure colorless solid (0.727 g, 32.7%). 1H NMR (400 MHz,
C6D6): δ=3.95 (m, 3H, NCH), 2.06 (m, 6H, PCH), 1.60 (d, 18H,
NC(CH3)2), 1.17 (m, 36H, PC(CH3)2). 31P{1H} NMR (161.8
MHz, C6D6): δ=1.5 (br, s). 13C NMR (100.5 MHz, C6D6): δ=
52.7, 29.0, 26.9, 21.6, 19.1.Owing to the increased oxygen
sensitivity of this complex, repeated elemental analysis samples
analyzed as the mono-oxidized compound (iPr2P(=O)NiPr)
(iPr2PNiPr)2ZrCl. Anal. Calcd for C27H63ClN3P3ZrO: C,
48.74; H, 9.54; N, 6.31. Found: C, 48.41; H, 9.69; N, 6.15.
ICo(Ph2PNiPr)3ZrCl (4). Solid 1 (1.502 g, 1.76 mmol) and
solid CoI2 (0.551 g, 1.76 mmol) were combined in THF (15 mL)
and stirred for 4 h at room temperature. The resulting dark
green reaction solution was filtered through Celite, and solvent
was removed from the volatiles in vacuo. The remaining green
solids were extracted with toluene (10 mL) and filtered through
Celite. The filtrate was cooled to -35 °C overnight to yield 4 as
dark green crystalline solids. (1.137 g, 62.2%). Crystals for
X-ray diffraction were grown by cooling a toluene/pentane
solution to -35 °C. 1H NMR (400 MHz, C6D6): δ = 14.22
(br s, Ph), 3.45 (br, (CH(CH3)2), 1.66 (br s, (CH(CH3)2), -5.57
(br, Ph), -7.01 (br s, Ph). UV-vis (C6H6) λmax, nm (ε): 588
(190), 689 (140), 709 (130), 800 (240), 872 (610). Evans’ method
(C6D6): 2.92 μB. Anal. Calcd for C45H51ClCoIN3P3Zr: C, 52.00;
H, 4.95; N, 4.04. Found: C, 51.17; H, 4.99; N, 3.94.
MesNHPiPr2. A solution of 2,4,6-trimethylaniline (5.00 g,
37.0 mmol) in Et2O (100 mL) was cooled to 0 °C in an ice bath.
To this was added nBuLi (25.0 mL, 1.6 M in hexanes, 40.7 mmol)
dropwise over 20 min. The resulting solution was allowed to
warm slowly to room temperature, resulting in the formation of
a white precipitate, then refluxed for 30 min to ensure complete
reaction. The resulting suspension was added dropwise to
i
a stirring solution of Pr2PCl (5.88 mL, 37.0 mmol) in Et2O
(100 mL) at -78 °C. Upon warming slowly to room tempera-
ture, the mixture became slightly yellow with a white LiCl
precipitate. After stirring for 1 h, the reaction mixture was
filtered through a pad of Celite, and solvent was removed from
the filtrate in vacuo. The resulting oily residue was redissolved in
Et2O and filtered through a plug of silica gel. Removal of the
solvent in vacuo yielded analytically pure product as a yellow oil
1
(7.64 g, 82.2%). H NMR (400 MHz, C6D6): δ=6.79 (s, 2H,
Mes), 3.16 (d, 1H, 2JH-P=9.2 Hz, NH), 2.38 (s, 6H, Mes), 2.21
(s, 3H, Mes), 1.57 (m, 2H, CH(CH3)2), 1.04 (m, 12H, CH-
(CH3)2). 31P{1H} NMR (161.8 MHz, C6D6): δ=57.7 (br, s).
13C NMR (100.5 MHz, C6D6): δ=142.0 (2JC-P=9.9 Hz), 130.4,
129, 127.8, 28.3 (1JC-P=16.8 Hz), 20.4, 19.0, 18.8, 17.3. Anal.
Calcd for C15H26NP: C, 71.68; H, 10.43; N, 5.57. Found: C,
71.75; H, 10.66; N, 5.55.
iPr2PNHiPr. A solution of isopropylamine (4.29 g, 72.5 mmol)
in benzene (100 mL) was cooled to 0 °C. To thiswas added iPr2PCl
(4.25 mL, 29.0 mmol) dropwise while stirring. The resulting
solution was allowed to warm to room temperature over 1 h to
ensure completion of the reaction. Upon warming to room
temperature the solution became viscous. After completion, the
viscous solution was filtered through a plug of silica gel. Removal
of the solvent in vacuo yielded analytically pure product as a
clear colorless liquid (4.43 g, 87.2%). 1H NMR (400 MHz, C6D6):
δ 2.99 (m, N-iPr, 1H), 1.33 (m, P-iPr, 2H), 0.95 (d, N-iPr, 3H), 0.92
(m, P-iPr, 6H), 0.56 (s, NH, 1H). 31P{1H} NMR (161.8 MHz,
C6D6): δ 57.6 (br, s). 13C NMR (100.5 MHz, C6D6): δ 49.0 (2JC-P
=
34 Hz), 26.4 (1JC-P=15.1), 26.2 (2JC-P=7.1), 19.4 (3JC-P=25.18),
17.2 (2JC-P=10.07). Owing to the increased oxygen sensitivity of this
ligand, repeated elemental analysis samples analyzed as the oxidized
form (iPr2P(=O)NHiPr). Anal. Calcd for C9H22NPO: C, 56.52; H,
11.59; N, 7.32. Found: C, 56.29; H, 11.76; N, 7.15.
ICo(iPr2PNMes)3ZrCl (5). Solid 2 (1.641 g, 1.87 mmol) and
solid CoI2 (0.585 g, 1.87 mmol) were combined in CH2Cl2 and
stirred for 48 h at room temperature. The resulting bright green
solution was filtered through Celite, and solvent was removed
from the filtrate in vacuo. The remaining green solids were
washed with copious amounts of pentane and dried in vacuo to
yield analytically pure 5 as a bright green solid (1.571 g, 79.0%).
Crystals suitable for X-ray diffraction were grown via slow
evaporation of a concentrated CH2Cl2 solution. 1H NMR
(400 MHz, C6D6): δ=12.98 (br s, iPr-Me), 7.01 (s, Mes-Me),
2.54 (br s, iPr-Me), 1.95 (s, Mes-Me), -1.92 (br, Mes-Ar). UV-
vis (C6H6) λmax, nm (ε): 654 (190), 894 (350). Evans’ method
(C6D6): 2.87 μB. Anal. Calcd for C45H75ClCoIN3P3Zr: C, 50.82;
H, 7.11; N, 3.95. Found: C, 49.69; H, 7.09; N, 3.79.
(iPr2PNMes)3ZrCl (2). A solution of MesNHPiPr2 (1.046 g,
4.162 mmol) in Et2O (50 mL) was cooled to -78 °C. To this was
added BuLi (2.6 mL, 1.6 M in hexanes, 4.2 mmol) dropwise
n
over 10 min. The resulting yellow/orange solution was warmed
to room temperature and stirred for 2 h. The solution was then
cooled again to -78 °C, and ZrCl4 (0.323 g, 1.39 mmol) was
added portionwise as a solid. The reaction mixture was warmed
to room temperature and stirred for 12 h. Volatiles were
removed from the solution in vacuo, and the resulting solids
were extracted with CH2Cl2 (30 mL) and filtered through a
pad of Celite to remove LiCl. The volume of the filtrate was
reduced to 10 mL in vacuo. The resulting supersaturated solu-
tion was layered with pentane and cooled to -35 °C to yield
analytically pure colorless crystals of 2 (0.772 g, 63.4%).
1H NMR (400 MHz, C6D6): δ = 6.84 (s, 6H, Mes), 2.48
(s, 18H, Mes), 2.34 (m, 6H, CH(CH3)2), 2.19 (s, 9H, Mes), 1.30
(m, 18H, CH(CH3)2), 1.16 (m, 18H, CH(CH3)2). 31P{1H} NMR
ICo(iPr2PNiPr)3ZrCl (6). Solid 3 (0.637 g, 9.81 mmol) and
solid CoI2 (0.307 g, 9.81 mmol) were combined in CH2Cl2 and
stirred for 15 h at room temperature. The resulting green
solution was filtered through Celite, and the solvent was re-
moved from the filtrate in vacuo. The remaining green
solids were repeatedly washed with diethyl ether and dried in
vacuo to yield analytically pure product as a green solid (0.382 g,
46.5%). Crystals suitable for X-ray diffraction were grown via
slow diffusion of pentane into a concentrated toluene solu-
(28) Dunning, T. H.; Hay, P. J. In Modern Theoretical Chemistry;
Schaefer, H. F., Ed.; Plenum: New York, 1976; Vol. 3, p 1-28.
(29) Mayer, I. Int. J. Quantum Chem. 1986, 29, 477–483.
1
tion. H NMR (400 MHz, C6D6): δ=70.6 (s br, P-CH), 8.05
(s br, P-iPrMe), 3.26 (s br, N-CH), 1.92 (s br, N-iPrMe), -1.68

6260 Inorganic Chemistry, Vol. 48, No. 13, 2009
Greenwood et al.
(s br, P-iPrMe). UV-vis (C6H6) λmax, nm (ε): 691.0 (180), 724.9
(210), 744.0 (230), 767.9 (290), 873.9 (240). Evans’ method
(C6D6): 3.10 μB. Anal. Calcd for C27H63ClCoIN3P3Zr: C,
38.83; H, 7.60; N, 5.03. Found: C, 38.77; H, 7.62; N, 5.00.
(Ph2PNHiPr)2CoI2 (7). Solid CoI2 (0.176 g, 0.562 mmol)
and Ph2PNHiPr (0.410 g, 1.69 mmol) were combined in THF
(15 mL), and the resulting brick red solution was stirred for 4 h
at room temperature. Solvent was removed from this solution
in vacuo, and the resulting red solids were extracted into benzene
(5 mL) and filtered through Celite. The brick red filtrate was
then layered with pentane (15 mL) and allowed to sit for 12 h at
room temperature to obtain 7 as a pure red crystalline product
(0.182 g, 40.4%). Crystals of 7 suitable for X-ray diffraction
were grown via vapor diffusion of pentane into a concentrated
CH(CH3)2) 2.96 (br s, CH(CH3)2), 0.64 (br s, Ph). UV-vis
(C6H6) λmax, nm (ε): 611 (130), 683 (150), 955 (280). Evans’
method (C6D6): 2.89 μB. Anal. Calcd for C45H54CoIN3P3: C,
59.02; H, 5.94; N, 4.59. Found: C, 58.94; H, 5.94; N, 4.58.
[N2Co(iPr2PNMes)3ZrX][Na(THF)5] (9). A 0.5% Na/Hg
amalgam was prepared from 0.0057 g Na (0.25 mmol) and
1.1 g Hg. To this vigorously stirred amalgam in 10 mL of
THF was added a solution of 5 (0.1056 g, 0.0993 mmol) in
THF (5 mL). The solution immediately began to change color
from green to red. After 2 h, the resulting red solution was
filtered away from the amalgam, and the solvent was removed
from the filtrate in vacuo. Solvent was extracted back into THF
and filtered through Celite. Layering the resulting concentrated
red solution with pentane and cooling to -35 °C resulted in red
crystals of 9 (0.0879 g, 61.5%). Crystals suitable for X-ray
diffraction were grown via vapor diffusion of pentane into a
concentrated THF solution of 9 at -35 °C. 1H NMR (400 MHz,
C6D6): δ=6.77 (s, 6H, Mes), 3.36 (br m, 20H, THF), 3.03 (m,
6H, CH(CH3)2), 2.51 (s, 18H, Mes-Me), 2.13 (s, 9H, Mes-Me),
1.79 (m, 18H, CH(CH3)2), 1.57 (m, 18H, CH(CH3)2), 1.24 (br m,
20H, THF). 31P{1H} NMR (162 MHz, C6D6): δ = 43 ppm
(br s). IR (KBr solution cell, THF): 2024 cm-1. UV-vis (C6H6)
1
benzene solution of 7. H NMR (400 MHz, C6D6): δ=15.94
(br s, Ph), 4.66 (br s, Ph), 4.86 (br s, CH(CH3)2), -10.0 (br, CH-
(CH3)2), -10.8 (br s, Ph). UV-vis (C6H6) λmax, nm (ε): 486
(1600), 677 (1040), 715 (1390), 782 (1100). Evans’ method
(C6D6): 4.38 μB. Anal. Calcd for C30H36CoI2N2P2: C, 45.08;
H, 4.54; N, 3.50. Found: C, 45.34; H, 4.46; N, 3.48.
(Ph2PNHiPr)3CoI (8). Solid CoI2 (0.582 g, 1.86 mmol) and
Ph2PNHiPr (1.583 g, 6.51 mmol) were combined in THF
(15 mL) at room temperature, initially generating a brick red
solution of 7. To this vigorously stirred mixture was added
Zn powder (0.608 g, 9.30 mmol). After 4 h of vigorous stirring,
the resulting green solution was filtered away from the ex-
cess Zn. The solvent was removed from the filtrate in vacuo,
and the resulting green residue was extracted into toluene
(10 mL) and filtered to remove Zn salts. The green toluene
solution was concentrated to 5 mL and layered with pentane
(15 mL). After storage for 12 h at -35 °C, 8 was isolated as green
crystals (1.129 g, 66.4%). Note: Over time in solution, 8
disproportionates to form 7 even at low temperature. Thus,
the yield reported above corresponds to material that is only
∼85-90% pure, with red/brown powder of 7 present. Spectro-
scopic and elemental analysis data was collected on crystals
of 8 that were mechanically separated from this mixture. Crys-
tals suitable for X-ray diffraction were grown via layering a
concentrated benzene solution with pentane. 1H NMR (400
MHz, C6D6): δ = 79.5 (br, Ph), 11.17 (br s, Ph), 7.03 (br s,
λmax, nm (ε): 509 (410), 676 (sh). The lability of the dinitrogen
ligand and the overall instability of 9 precluded the collection
of satisfactory combustion analysis data. Repeated samples
consistently analyzed low for %N, as well as %C.
Acknowledgment. This work was supported by the gracious
startup funds of Brandeis University. The authors are grateful to
Nate Szymczak and Jonas C. Peters for their assistance with
bulk electrolysis experiments. We thank D. Kellenberger and
R. Rotstein for their assistance with X-ray crystallography.
Supporting Information Available: Experimental details, vari-
able temperature H NMR data for 2 and 3, complete cyclic
1
voltammetry data for 4-6 and bulk electrolysis data for 4,
crystallographic data for 4-9 (in CIF format) and additional
computational detials. This material is available free of charge
via the Internet at http://pubs.acs.org.