C-N bond formation on addition of aryl carbanions to the electrophilic nitrido ligand in TpOs(N)Cl2

Article abstract of DOI:10.1021/ja0028424

The osmium (VI) nitrido complex TpOs(N)Cl₂ (1) has been prepared from K[Os(N)O₃] and KTp in aqueous ethanolic HCl. It reacts rapidly with PhMgCl and related reagents with transfer of a phenyl group to the nitrido ligand. This forms Os(IV) metalla-analido complexes, which are readily protonated to give the analido complex TpOs(NHPh)Cl₂ (4). The nitrido-phenyl derivatives TpOs(N)PhCl and TpOs(N)Ph₂ react more slowly with PhMgCl and are not competent intermediates for the reaction of 1 with PhMgCl. Reactions of 1 with alkyl- and arylboranes similarly result in transfer of one organic group to nitrogen, leading to isolable borylamido complexes such as TpOs[N(Ph)(BPh₂)]Cl₂ (11). This is an unprecedented insertion of a nitrido ligand into a boron - carbon bond. Hydrolysis of 11 gives 4. Mechanistic studies suggest that both the Grignard and borane reactions proceed by initial weak coordination of Mg or B to the nitrido ligand, followed by migration of the carbanion to nitrogen. The hydrocarbyl group does not go to osmium and then move to nitrogen - there is no change in the atoms bound to the osmium during the reactions. It is suggested that there may be a general preference for nucleophiles to add directly to the metal - ligand multiple bond rather than binding to the metal first and migrating. Ab initio calculations show that the unusual reactivity of 1 results from its accessible LUMO and LUMO + 1, which are the Os≡N π* orbitals. The bonding in 1 and its reactivity with organoboranes are reminiscent of CO.

Full text of DOI:10.1021/ja0028424

J. Am. Chem. Soc. 2001, 123, 1059-1071
1059
C-N Bond Formation on Addition of Aryl Carbanions to the
Electrophilic Nitrido Ligand in TpOs(N)Cl2
Thomas J. Crevier, Brian K. Bennett, Jake D. Soper, Julie A. Bowman, Ahmad Dehestani,
David A. Hrovat, Scott Lovell,† Werner Kaminsky,† and James M. Mayer*
Contribution from the Department of Chemistry, Campus Box 351700, UniVersity of Washington,
Seattle, Washington 98195-1700
ReceiVed August 1, 2000
Abstract: The osmium(VI) nitrido complex TpOs(N)Cl2 (1) has been prepared from K[Os(N)O3] and KTp in
aqueous ethanolic HCl. It reacts rapidly with PhMgCl and related reagents with transfer of a phenyl group to
the nitrido ligand. This forms Os(IV) metalla-analido complexes, which are readily protonated to give the
analido complex TpOs(NHPh)Cl2 (4). The nitrido-phenyl derivatives TpOs(N)PhCl and TpOs(N)Ph2 react
more slowly with PhMgCl and are not competent intermediates for the reaction of 1 with PhMgCl. Reactions
of 1 with alkyl- and arylboranes similarly result in transfer of one organic group to nitrogen, leading to isolable
borylamido complexes such as TpOs[N(Ph)(BPh2)]Cl2 (11). This is an unprecedented insertion of a nitrido
ligand into a boron-carbon bond. Hydrolysis of 11 gives 4. Mechanistic studies suggest that both the Grignard
and borane reactions proceed by initial weak coordination of Mg or B to the nitrido ligand, followed by migration
of the carbanion to nitrogen. The hydrocarbyl group does not go to osmium and then move to nitrogensthere
is no change in the atoms bound to the osmium during the reactions. It is suggested that there may be a
general preference for nucleophiles to add directly to the metal-ligand multiple bond rather than binding to
the metal first and migrating. Ab initio calculations show that the unusual reactivity of 1 results from its
accessible LUMO and LUMO + 1, which are the OstN π* orbitals. The bonding in 1 and its reactivity with
organoboranes are reminiscent of CO.
Introduction
forming methodologies are of interest for both fine and
commodity chemicals.7
Metal-ligand multiple bonds are being utilized increasingly
both as supporting ligands and as reactive functionalities.1
Selective oxidations of organic compounds have long been
accomplished by metal-oxo complexes, such as osmium
tetroxide and permanganate.2 The use of metal-nitrogen
multiple bonds to oxidatively form C-N bonds is more recent,
as in the Sharpless osmium oxyamination reaction3 and the
heterogeneous ammoxidation of propylene to acrylonitrile.4
Metal nitrido compounds (LnMtN) are receiving increasing
attention because of their potential role in aziridination, nitrogen
fixation, and other processes.1,5,6 New selective C-N bond-
The nitrido ligand is an interesting functionality because its
reactivity is quite sensitive to the ancillary ligands on the metal
center. For example, the nitrido ligand of [Os(N)R4]- (R ) CH2-
SiMe3) is nucleophilic, being alkylated by MeI, while isoelec-
tronic [Os(N)Cl4]- is electrophilic, reacting with phosphines at
nitrogen.8 Complexes with electrophilic multiply bonded ligands
are good candidates to be oxidants because they can potentially
add nucleophiles directly at the ligand; this is now the favored
mechanism for OsO4 oxidations of alkenes.9
Reported here are reactions in which carbanions, especially
Grignard reagents and organoboranes, add to the electrophilic
nitrido ligand in TpOs(N)Cl2 (1) (Tp ) hydrotris(1-pyrazolyl)-
borate).10 The scope and mechanism of these new C-N bond-
forming reactions are described. The data indicate a mechanism
of direct addition to the nitrido ligand, without the involvement
of an osmium hydrocarbyl intermediate. The importance of the
electrophilicity of the nitrido ligand is indicated by experimental
† UW Chemistry departmental crystallographer.
(1) Nugent, W. A.; Mayer, J. M. Metal-Ligand Multiple Bonds; John
Wiley & Sons: New York, 1988. (b) Parkin, G. In Progress in Inorganic
Chemistry; Karlin, K. D., ed.; John Wiley & Sons: New York, 1998; Vol.
47, pp 1-167. (c) Dehnicke, K.; Stra¨hle, J. Angew. Chem., Int. Ed. Engl.
1992, 31, 955-978. (d) Wigley, D. E. Prog. Inorg. Chem. 1994, 42, 239-
482.
(2) Sheldon, R. A.; Kochi, J. K. Metal-Catalyzed Oxidations of Organic
Compounds; Academic Press: New York, 1981. (b) ComprehensiVe Organic
Synthesis, Vol. 7, Oxidation; Trost, B. M., Ed.; Pergamon: New York, 1991.
(c) Organic Syntheses by Oxidation with Metal Compounds; Mijs, W. J.,
de Jonge, C. R. H. I., Eds.; Plenum: New York, 1986. (d) Griffith, W. P.
Trans. Met. Chem. 1990, 15, 251-256.
(6) Brown, S. N. J. Am. Chem. Soc. 1999, 121, 9752-9753. (b) Brown,
S. N. Inorg. Chem. 2000, 39, 378-381.
(7) Recent reviews: (a) Mu¨ller, T. E.; Beller, M. Chem. ReV. 1998, 98,
675-703. (b) Dembech, P.; Seconi, G.; Ricci, A. Chem. Eur. J. 2000, 6,
1281-1286. (c) Hartwig, J. F. Angew. Chem., Int. Ed. 1998, 37, 2047-
2067. (d) Yang, B. H.; Buchwald, S. L. J. Organomet. Chem. 1999, 576,
125-146.
(8) Marshman, R. W.; Shapley, P. A. J. Am. Chem. Soc. 1990, 112,
8369-8378.
(9) DelMonte, A. J.; Haller, J.; Houk, K. N.; Sharpless, K. B.; Singleton,
D. A.; Strassner, T.; Thomas, A. A. J. Am. Chem. Soc. 1997, 119, 9907-
9908.
(10) Preliminarly accounts of a portion of this work have appeared. (a)
Crevier, T. J.; Mayer, J. M. J. Am. Chem. Soc. 1998, 120, 5595-5596. (b)
Crevier, T. J.; Mayer, J. M. Angew. Chem., Int. Ed. 1998, 37, 1891.
(3) Chong, A. O.; Oshima, K.; Sharpless, J. Am. Chem. Soc. 1977, 99,
3420.
(4) For a recent overview, see: Olah, G. A.; Molna´r, AÄ . Hydrocarbon
Chemistry; Wiley: New York, 1995; pp 372-374.
(5) Groves, J. T.; Takahashi, T. J. Am. Chem. Soc. 1983, 105, 2073-
2074. (b) Bottomley, L. A.; Neely, F. L. J. Am. Chem. Soc. 1988, 110,
6748-6752. du Bois, J.; Tomooka, C. S.; Hong, J.; Carreira, E. M. Acc.
Chem. Res. 1997, 30, 364-372. (c) Cummins, C. C. Chem. Commun. 1998,
1777-1786.
10.1021/ja0028424 CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/23/2001

1060 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
CreVier et al.
studies and by density functional calculations. There have been
few ab initio studies on transition metal nitrido complexes, and
few attempts to explain trends in reactivity using molecular
orbital arguments.11,12
Results
Synthesis of TpOs(N)Cl2 (1) and Related Compounds.
TpOs(N)Cl2 (1) is prepared in 35-50% yield by refluxing a
mixture of KOs(N)O3,13 KTp, and aqueous HCl in ethanol for
40 min (eq 1). This very closely follows the procedure for
preparation of the isoelectronic rhenium-oxo analogue, TpRe-
(O)Cl2, from KReO4.14 Reaction 1 initially turns purple before
becoming a brick-red color, from which 1 precipitates as an
orange-red solid. The reaction probably proceeds via purple
[Os(N)Cl5]2-, which is formed immediately upon treating KOs-
(N)O3 with HCl in ethanol in the absence of KTp.15 Complex
1 can also be obtained from K2[Os(N)Cl5] or [nBu4N][Os(N)-
Cl4] and KTp but in lower yield.16
Reaction 1 produces TpOsCl3 as a byproduct, particularly
when heating is continued beyond 40 min.17 Separation of 1
from TpOsCl3 is problematic; our best procedure involves
treatment with ferrocene, which removes TpOsCl3 as [Cp2Fe+]-
[TpOsCl3-].17 The X-ray crystal structure of 1 purified in this
fashion (Tables 1 and 2; Figure 1) shows N/Cl disorder at the
nitrido site, refined as 73% N/27% Cl, most likely due to
rotational disorder within the crystal.18 The elongated ellipsoid
for the nitrido ligand reflects the difficulty in resolving such
disorders, and the OstN distance is not accurately determined.19
The nitrido-diphenyl derivative, TpOs(N)Ph2 (3), is prepared
by adding KTp to a solution of [nBu4N][Os(N)Ph2I2], generated
in situ from [nBu4N][Os(N)Ph4] and aqueous HI (eq 2). This
2
procedure follows Koch and Shapley’s preparation of TpMe
-
Os(N)Ph2 from [Os(N)Ph2(py)2]BF4 (TpMe ) hydrotris(3,5-
dimethyl-1-pyrazolyl)borate).20 TpOs(N)PhCl (2) is formed on
treating 3 with HCl (1 M in Et2O) at 70 °C (eq 3). The selective
protonolysis of one phenyl group under fairly forcing conditions
illustrates the inertness of the Os-Ph bonds.
2
(11) Pandey, K. P.; Sharma, R. B.; Garg, J. J. Indian Chem. Soc. 1997,
74, 279-283.
(12) Lyne, P. D.; Mingos, M. P. J. Chem. Soc., Dalton Trans. 1995,
1635-1643.
(13) Griffith, W. P. Osmium. In ComprehensiVe Coordination Chemistry;
Wilkinson, G., Ed.; Pergamon: New York, 1987; Vol. 4, pp 519-633.
(14) Abrams, M. J.; Davison, A.; Jones, A. G. Inorg. Chim. Acta 1984,
82, 125-128. (b) Brown, S. N.; Mayer, J. M. Inorg. Chem. 1992, 31, 4091-
4100.
(15) Clifford, A. F.; Kobayahsi, C. S. Inorg. Synth. 1960, 6, 204-208.
(16) nBu4N][Os(N)Cl4] is the starting material for [(trpy)Os(N)Cl2]+ and
[(tpm)Os(N)Cl2]+. trpy ) 2,2′:6′,2′′-terpyridine: (a) Williams, D. S.; Coia,
G. M.; Meyer, T. J. Inorg. Chem. 1995, 34, 586-592. tpm ) tris(1-
pyrazolyl)methane: (b) Demadis, K. D.; El-Samanody, E.-S.; Meyer, T.
J.; White, P. S. Inorg. Chem. 1998, 37, 383-384. (c) Demadis, K. D.; El-
Samanody, E.-S.; Coia, G. M.; Meyer, T. J. J. Am. Chem. Soc. 1999, 121,
535-544.
(17) Bennett, B. K.; Crevier, T. J.; Mayer, J. M., manuscript in
preparation.
(18) The position of the nitrido ligand is independently indicated by the
large trans influence, with Os-Npz,trans (2.262(9) Å) almost 0.2 Å longer
than Os-Npz,cis (2.082, 2.059(9) Å). (b) For a related example, see: Lincoln,
S.; Koch, S. A. Inorg. Chem. 1986, 25, 1594-1602.
(19) Parkin, G. Chem. ReV. 1993, 93, 887-911.
(20) Koch, J. L.; Shapley, P. A. Organometallics 1997, 16, 4071-4076.

Carbanion Addition to an Osmium Nitrido
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1061
Table 2. Selected Bond Lengths (Å) and Angles (Deg) for Os(VI) Nitrido Complexes: Crystallographic Results for 1, 3, and 17 and
Computational Results (in Italics) for 1 and 18
TpOs(N)Cl2 (1)
TpOs(N)Cl2 (1)
TpOs(N)Ph2 (3)a
TpOs(N)Me2 (18)
TpOs(NBAr3)Ph2 (17)b
Os-N(1)
2.262 (9)
2.082(9)
2.055(8)
1.70(2)
2.295(3)
2.317(4)
2.103
2.073
2.350 (4)
2.150(3)
2.143(3)
1.637(4)
2.110(4)
2.111(4)
2.216
2.209
2.200 (3)
2.168(3)
2.119(3)
1.688(3)
2.089(4)
2.129(4)
Os-N(3)
Os-N(5)
Os-N(7)
Os-Cl(1)/C(10)
Os-Cl(2)/C(16)
1.699
2.507
1.676
2.154
N(7)-Os-N(1)
170.9(10)
92.1(10)
93.8(11)
100.0(10)
99.7(10)
88.95(15)
176.4
95.3
174.44(16)
97.53(17)
100.22(16)
97.15(18)
95.56(18)
89.92(16)
178.2
99.9
169.56(14)
94.74(13)
106.84(13)
99.75(14)
92.90(14)
90.71(14)
N(7)-Os-N(3)
N(7)-Os-N(5)
N(7)-Os-Cl(1)/C(10)
N(7)-Os-Cl(2)/C(16)
Cl/C-Os-Cl/C
94.9
92.3
94.1
89.3
a Data from one of two independent molecules of 3. b Other important bond lengths and angles for 17: N(7)-B(2), 1.591(5); B(2)-C(22),
1.652(6); B(2)-C(28), 1.626(5); B(2)-C(34), 1.636(6); Os(1)-N(7)-B(2), 167.7(3); N(7)-B(2)-C(22), 107.8(3); N(7)-B(2)-C(28), 101.8(3);
N(7)-B(2)-C(34), 111.3(3).
Figure 1. ORTEP diagram of one of the two independent molecules
in the crystal of TpOs(N)Cl2 (1).
Figure 2. ORTEP diagram of one of the two independent molecules
in the crystal of TpOs(N)Ph2 (3).
1. HI (aq)
2. KTp
[nBu4N][Os(N)Ph4] acetone 8
(2)
(3)
TpOs(N)Ph
2
molecules, are typical of such linkages (1.525-1.703 Å, av
3
1.629 Å;24 TpMe Os(N)Ph2, 1.648 (3) Å).
2
Complexes 1-3, like most of the compounds described here,
are air stable and chromatograph well on silica gel. Complex 1
decomposes slowly in CD2Cl2 or CDCl3 over roughly a week
at 80 °C, yielding some TpOsCl3. 3 is somewhat more stable.
The chloride ligands in 1 are quite inert: no exchange is
observed with NaI in either acetone or acetonitrile after extended
heating, and metathesis with silver salts (e.g., silver acetate)
requires heating.25 Chloride substitution in the related rhenium-
oxo and -imido systems TpRe(E)Cl2 is much more facile.21,26
Complex 1 is unreactive with the electrophiles MeOTf (methyl
triflate), [CF3C(O)]2O, BF3‚Et2O, and [Ph3C][BF4] at ambient
temperatures, with complex reactions occurring over days at
∆
TpOs(N)Ph2
8 TpOs(N)PhCl
CH2Cl2/Et2O
HCl
2
3
1H and 13C NMR spectra of 1-3 are consistent with
diamagnetic complexes of Cs (1, 3) or C1 (2) symmetry and are
14,21
similar to the related TpRe(O)ClnPh(2-n)
.
The IR spectrum
of 1 shows peaks typical of octahedral Tp complexes22 and an
OstN stretch at 1066 cm-1 (Ost15N, 1036 cm-1; calcd, 1030
cm-1). Meyer et al. report that 1 undergoes an irreversible
reduction at -0.98 vs SSCE in MeCN, yielding [TpCl2OsII(N2)-
OsIIICl2Tp]-.16c We observe a similar reduction (-1.35 V vs
Cp2Fe+/0) and an oxidation wave for 1 at ∼2.05 V.23 The aryl
complexes are harder to reduce (-1.89 V, 2; -2.04 V, 3) and
easier to oxidize (1.60 V, 2; 1.36 V, 3). The X-ray crystal
1
70 °C (as determined by H NMR). This contrasts with the
reactivity of many nucleophilic metal nitrido complexes,1 such
as the clean reactions of CpOs(N)(CH2SiMe3)2 with MeOTf,
BF3‚Et2O, and Ag+.27 Shapley reports slow alkylation of TpMe
-
2
structure of 3 (Tables 1 and 2, Figure 2) is similar to those of
Os(N)Ph2 by MeOTf,28 while we find that TpOs(N)Ph2 (3) reacts
20
1, TpMe Os(N)Ph2, and especially TpRe(O)Ph2.21b The Ost
2
N bond lengths, av 1.634 (4) Å in the two independent
(24) From a search of the Cambridge Structural Database, January, 1998
release. Allen, F. H.; Kennard, O. Chem. Des. Autom. News 1993, 8, 31-
37.
(25) Dehastani, A.; Bennett, B. K.; Mayer, J. M., work in progress. (b)
Crevier, T. J. Ph.D. Thesis, University of Washington, 1998.
(26) McNeil, W. S.; DuMez, D. D.; Matano, Y.; Lovell, S.; Mayer, J.
M. Organometallics 1999, 18, 3715-3727.
(27) Marshman, R. W.; Shusta, J. M.; Wilson, S. R.; Shapley, P. A.
Organometallics 1991, 10, 1671-1676.
(28) Shapley, P. A., personal communication, 1998.
(21) Brown, S. N.; Mayer, J. M. J. Am. Chem. Soc. 1996, 118, 12119-
12133. (b) Brown, S. N.; Myers, A. W.; Fulton, J. R.; Mayer, J. M.
Organometallics 1998, 17, 3364-3374.
(22) Trofimenko, S. Chem. ReV. 1993, 93, 943-980 and references
therein.
(23) Bennett, B. K.; Crevier, T. J.; DuMez, D. D.; Matano, Y.; McNeil,
W. S.; Mayer, J. M. J. Organomet. Chem. 1999, 591, 96-103.

1062 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
CreVier et al.
Table 3. Selected Bond Lengths (Å) and Angles (Deg) for Os(IV) Complexes TpOs(NHPh)Cl2 (4), TpOs(NH2Et)Cl2 (8),
TpOs[N(Ph)BPh2]Cl2, (11), TpOs[NPh(BPhOBPh2)]Cl2 (12), and TpOs(N(C6F5)[B(C6F5)2])Cl2 (13)a
4b
8
11c
12d
13d
Os-Namide
1.919(6)
2.097(5)
2.052(6)
2.039(6)
2.363(2)
2.384(2)
1.384(9)
2.129(8)e
2.056(8)
2.051(5)
1.880(10)
2.103(9)
2.031(9)
2.065(10)
2.403(3)
2.355(3)
1.405(15)
1.602(16)
1.607(19)
1.60(2)
1.939(7)
2.074(7)
2.038(7)
2.038(7)
2.363(2)
2.372(2)
1.396(11)
1.507(12)
1.570(15)
1.344(13)
2.057(7)
2.075(7)
2.026(8)
2.036(8)
2.323(3)
2.308(2)
1.461(11)
1.408(14)
1.583(15)
1.632(18)
a
Os-Ntrans
a
Os-Ncis1
a
Os-Ncis2
Os-Cl(1)
2.3707(19)
1.501(13)e,f
Os-Cl(2)
Namide-C(10)
Namide-B(2)
B(2)-C(16)
B(2)-C(22)/O
B(2)-Cl(1)
B(3)-O
2.069(15)
1.406(13)
1.406(13)
Namide-Os-Ntrans
Namide-Os-Ncis1
Namide-Os-Ncis2
Namide-Os-Cl(1)
Namide-Os-Cl(2)
Cl(1)-Os-Cl(2)
Os-Namide-C(10)
Os-Namide-B(2)
C(10)-Namide-B(2)
178.9(2)
93.7(3)
94.6(2)
90.94(18)
89.88(19)
90.69(9)
133.8(5)
179.9(2)e
172.5(4)
97.4(4)
93.6(4)
178.9(3)
96.8(3)
93.5(3)
89.9(2)
93.1(2)
90.78(9)
127.6(6)
115.6(5)
116.7(7)
173.7(3)
91.0(3)
94.0(3)
93.0(2)
94.8(2)
89.31(9)
115.7(5)
130.6(7)
113.7(8)
93.4(2)e
88.06(15)e
75.7(3)
100.3(3)
89.24(12)
127.9(8)
112.0(7)
119.3(10)
91.21(9)
122.0(6)e,f
a
Ntrans and Ncis refer to the relationship to Namide.
b Data from one of two independent molecules. c Other angles for 11: N(7)-B(2)-C(16),
114.9(11); N(7)-B(2)-C(22), 115.3(10); N(7)-B(2)-Cl(1), 92.0(7); C(16)-B(2)-C(22), 116.3(11); C(16)-B(2)-Cl(1), 105.8(9); C(22)-B(2)-
Cl(1), 108.9(9); Os-Cl(1)-B(2), 80.2(4). d Other bond lengths and angles for 12: B(3)-C(22), 1.578(14); B(3)-C(28), 1.581(14); O-B(2)-
N(7), 121.7(9); O-B(2)-C(16), 118.7(8); N(7)-B(2)-C(16), 119.4(8); B(2)-O-B(3), 138.6(8); O-B(3)-C(28), 116.2(8); O-B(3)-C(22), 116.4(8);
C(22)-B(3)-C(28), 121.3(8). e Distance or angle involving Namine
.
f Distance or angle involving C(5).
with excess MeOTf in CDCl3 only upon heating at 70 °C for a
day, leading to decomposition. Changing chloride for phenyl
and changing Tp to TpMe both affect the reactivity of the nitrido
2
ligands, as well as the redox potentials of the complexes.23
Reactions with Grignard and Related Reagents. Addition
of 1 equiv of PhMgCl to an ether solution of TpOs(N)Cl2 (1)
at -78 °C gives a dark orange solution, which turns bright red
on exposure to air. Aerobic workup and chromatography on
silica gel give the red-purple osmium(IV) analido complex
TpOs(NHPh)Cl2 (4) (eq 4). Similar results are obtained in THF
or toluene, and using PhMgBr, PhLi, or “Ph3ZnLi” (3PhLi +
ZnCl2) as the arylating agents. Phenyl cuprates, however, gave
low and irreproducible yields of the Cl metathesis products 2
and/or 3.
Compound 4 has a mildly paramagnetically shifted 1H NMR
spectrum with sharp lines, as is typical for d4 octahedral Os
and Re complexes.29 Mass spectrometry shows a strong parent
ion cluster with the characteristic isotope pattern, along with
clusters for M+-Cl and M+-NHPh. The IR spectrum of 4
shows a somewhat broad peak at 3519 cm-1 due to the amido
proton. Workup with D2O instead of air (see below) gives 4-d1,
in which the characteristic NH resonance at δ 5.87 is absent in
the 1H NMR. The X-ray structure of 4 (Tables 1 and 3, Figure
3) shows two independent molecules with nearly identical bond
lengths and angles, and no evidence of intermolecular interac-
tions. The coordination geometry about osmium is roughly
octahedral, with the phenyl group of the analido ligand
interleaved between two pyrazoles.
Figure 3. ORTEP diagram of one of the two independent molecules
in the crystal of TpOs(NHPh)Cl2 (4).
Addition of tolylmagnesium bromide to 1 and workup gives
the tolyl derivative TpOs(NHTol)Cl2 (5). Interestingly, 5 is also
formed on reaction of 1 with p-toluidine (H2NTol) over 3 days
at 70 °C.25b Reaction of 1 with 2 or 3 equiv of phenyl anion
gives the analido-phenyl complexes TpOs(NHPh)PhCl (6) and/
or TpOs(NHPh)Ph2 (7) upon workup. In both cases, there is
formation of a dark orange solution which turns bright red upon
aerobic workup. Compounds 5-7 give mass spectra and sharp,
paramagnetically shifted 1H NMR spectra, consistent with their
formulation as analogous to 4. The X-ray structure of 5 is
described in the Supporting Information.
When the reaction of 1 and PhMgBr is kept anaerobic, the
initially formed dark orange intermediates can be observed,
(29) For example: Randall. E. W.; Shaw, D. J. Chem. Soc. A 1969,
2867-2872.

Carbanion Addition to an Osmium Nitrido
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1063
Scheme 1
Figure 4. ORTEP diagram of TpOs(NH2Et)]Cl2 (8).
identified as the osmium(IV) amide complex TpOs(NHEt)Cl2
(10; see below). Most likely, all of these reactions proceed by
addition of R- to the nitrido ligand, analogous to the aryl anion
additions, but the alkyl carbanion sources cause further reduction
to osmium(III). Consistent with the involvement of electron
transfer, treatment of 1 with PhCH2MgBr forms substantial
bibenzyl and toluene.
showing paramagnetic 1H NMR spectra. With less than 1 equiv
of Grignard reagent, a red-orange solid can be isolated that
contains a mixture of 1 and a new species that has Cs symmetry
and contains one Tp and one phenyl group. When excess
Grignard reagent is used, the intermediate has a Cs-symmetric
Tp ligand and two sets of phenyl resonances in a 2:1 ratio. Both
species are very rapidly protonated by stoichiometric degassed
H2O, yielding 4 or 7 (as happens on aerobic workup). The
intermediates are therefore tentatively formulated as osmium-
(IV) metalla-analido complexes TpOs[N(Ph)MgBr]Cl2 (from 1
equiv of PhMgBr) and TpOs[N(Ph)MgBr]Ph2 (from excess
PhMgBr) (Scheme 1). The solubility of these species in benzene
suggests that there is tight ion pairing or coordination of the
imido nitrogen to the magnesium counterion, and there is likely
to be oligomerization about the cation. Dark orange materials
The nitrido-phenyl complexes 2 and 3 react very slowly with
PhMgBr in THF solution at room temperature. This contrasts
with the reaction of 1 with PhMgBr, which occurs instantly at
20 °C and over the course of a few hours at -78 °C. Workup
of the dark brown solution from 2 + excess PhMgBr after 1
day gives ∼20% 7, ∼20% of unreacted 2, and NMR-silent
1
material. The use of C6D5MgBr gives 7-d10, whose H NMR
spectra show that the analido phenyl ring is >95% deuterated,
and the OsPh2 resonances have half the intensity seen for
protio-7 (eq 6). The analogous reaction of 3 with PhMgBr over
1
formed from PhLi have slightly different H NMR spectra,
further supporting tight bonding with the cation. Various
attempts to isolate these materials, including cation exchange,
have yielded only impure materials. This is due in part to their
very high basicity, similar to other M-N(R)Li complexes30 (as
described elsewhere31).
Reactions of 1 with the alkyl carbanion sources EtMgBr,
EtAlCl2, ZnEt2, AlEt3, MeMgBr, and PhCH2MgBr give a
mixture of products, predominantly H NMR-silent materials.
In two cases, with ZnEt2 and MeMgBr, osmium(III) amine
complexes TpOs(NH2R)Cl2 (R ) Et, 8; Me, 9) have been
isolated in ∼30% yields (eq 5). 8 and 9 have been characterized
1
1 day gives ∼10% 7, ∼80% unreacted 3, and uncharacterized
materials. Reaction with C6D5MgBr gives 7-d5 (eq 6), where
only the analido phenyl group is deuterated (as determined by
1H NMR). The use of TolMgBr in reaction 6 gives somewhat
higher yields of Os(IV) analido products, apparently with
analogous substitution patterns.
Reactions with Organoboranes. Anaerobic addition of BPh3
to a toluene solution of 1 results in an immediate color change
from pale orange to dark, vivid orange. Layering the solution
with pentanes yields the borylamido complex TpOs[N(Ph)-
(BPh2)]Cl2 (11) as dark, air-sensitive orange crystals (eq 7). A
by elemental analysis, IR spectroscopy, mass spectrometry, and
X-ray crystal structures (8, Tables 1 and 3, Figure 4; 9 (quite
similar), see Supporting Information). The structures of 8 and
9 are very close to those of 4 and 5, except that the OsIII-
NH2R distances of 2.129(8) (8) and 2.139(8) (9) Å are ∼0.2 Å
longer than the OsIV-NHAr bonds of 1.919(6) (4) and 1.922-
(7) (5) Å. Some reactions of 1 with 1 equiv of EtMgBr show,
on protonation, small NMR signals for what is tentatively
slower reaction between 1 and diphenylboronic anhydride (Ph2-
BOBPh2) forms TpOs[NPh(BPhOBPh2)]Cl2 (12) after stirring
overnight. Complex 1 reacts very rapidly with the potent
electrophile B(C6F5)3 in benzene to form TpOs(N(C6F5)-
[B(C6F5)2])Cl2 (13). Complexes 11-13 exhibit the sharp,
(30) For example: (a) Dewey, M. A.; Stark, G. A.; Gladysz, J. A.
Organometallics 1996, 15, 4798-4807. (b) Powell, K. R.; Pe´rez, P. J.; Luan,
L.; Feng, S. G.; White, P. S.; Brookhart, M.; Templeton, J. L. Organome-
tallics 1994, 13, 1851-1864.
(31) Soper, J.; Bennett, B. K.; Lovell, S.; Mayer, J. M. Inorg. Chem.
2001, in press.
32

1064 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
CreVier et al.
Chart 1.a
a X ) H, B(Y)Z.
-40 °C, indicating either that there is no B-Cl interaction in
solution or that the boron is rapidly “hopping” from one chloride
to the other in solution. In the structures of 12 and 13 there are
no B‚‚‚Cl interactions, despite the higher Lewis acidity of the
B(C6F5)2 group.
The B(2)-N(7) distances of 1.602(16) Å in 11 and 1.509-
(13) Å in 12 correspond to long and average B-N single bonds,
respectively (average B-N, 1.52 Å24), while the distance of
1.408(14) Å in 13 is quite short. The osmium-amido [Os-N(7)]
bond lengths vary in the reverse order: 1.884(10), 1.937(7),
and 2.057(7) Å for 11-13 (compare Os-N7 in 4 of 1.919(6)
Å). In [Os(en-H)2(en)]Br2 (en-H ) H2NCH2CH2NH-), the Os-
N(amide) bonds of 1.896(7) Å are much shorter than the Os-
N(amine) distances (cis, 2.113 (9); trans, 2.194(7) Å), and this
has been attributed to Os-N π bonding.33 On this basis, there
is likely to be significant Os-N π bonding in 11 and 12, with
a large contribution of resonance form B in Chart 1. In 11, the
B-Cl interaction reduces the contribution of C, leading to
shorter Os-N and longer N-B distances. Complex 13 is best
described by C, a result of electron-withdrawing C6F5 groups.
C is also the best description of the bonding for the two previous
reports of borylamido complexes, bulky M[N(Ar)(BAr2)]2
compounds reported by Power et al.34 and bis(borylamide)
complexes described by Schrock et al.35
Exposure of 11-13 to air, or addition of water under oxygen-
free conditions, results in hydrolysis of the B-N bonds to form
the analido complexes 4 or TpOs(NHC6F5)Cl2 (14; eq 8). In
Figure 5. ORTEP diagrams of (a) TpOs[N(Ph)BPh2]Cl2 (11) and (b)
TpOs(N(C6F5)[B(C6F5)2])Cl2 (13).
paramagnetically shifted 1H NMR spectra indicative of Os(IV),
with a significantly larger range of chemical shifts for 13 (for
instance, one of the pyrazole peaks appears at -23.31 ppm).
The paramagnetic 19F NMR spectrum of 13, though broadened
such that only two peaks have resolvable F-F coupling, shows
the presence of C6F5 groups in a 2:1 ratio.
the case of 11, the boron hydrolysis product is Ph2BOBPh2, as
determined by 1H NMR. 11 and 13 are very sensitive to
atmospheric moisture, even in the solid state, while solid 12
1
can be handled briefly in the air. H and 19F NMR spectra of
The structures of all three compounds have been confirmed
by X-ray diffraction (Tables 1 and 3, Figure 5 [for 12, see
Supporting Information]). All three show the roughly octahedral
coordination about osmium typical of this group of compounds.
In 11, the boron atom forms single bonds to the nitrogen and
to two phenyl groups and has an intramolecular interaction with
one of the chloride ligands. At 2.078(14) Å, this is the longest
B-Cl distance in the Cambridge Crystallographic Database (av
1.822 ( 0.049 Å).24 Still, this interaction is enough to pull
chloride toward the boron (N(7)-Os-B(2) ) 75.7(3)°) and to
lengthen the Os-Cl(1) bond (2.403(3) Å, vs Os-Cl(2), 2.355-
(3) Å). The boron is almost in the plane of the amide nitrogen
and the ipso carbons (Σ[X-Β-Y angles] ) 359.2°), further
14 are paramagnetically shifted, with pyrazoles in the expected
2:1 ratio and shifted but unbroadened fluorine resonances.
Complex 1 reacts rapidly with tri(tolyl)borane to give TpOs-
[N(Tol)(BTol2)]Cl2 (as determined by 1H NMR), which hydro-
lyzes to TpOs(NHTol)Cl2 (5). Reacting 1 with a 10-fold excess
of both BTol3 and BPh3 gave a 1:1.4 ratio of TpOs[N(Tol)-
(BTol2)]Cl2 to 11 and roughly the same ratio of 5:4 on hydrolytic
workup. 1H NMR spectra of the reaction mixture of 1 plus BTol3
and BPh3 showed no evidence of the crossover products TpOs-
[N(Ph)(BTol2)]Cl2 and TpOs[N(Tol)(BPh2)]Cl2 (<10%). Reac-
(33) Lay, P. A.; Sargeson, A. M.; Skelton, B. W.; White, A. H. J. Am.
Chem. Soc. 1982, 104, 6161-6162.
(34) Chen, H.; Bartlett, R. A.; Olmstead, M. M.; Power, P. P.; Shoner,
S. S. J. Am. Chem. Soc. 1990, 112, 1048-1055.
(35) Warren, T. H.; Schrock, R. R.; Davis, W. M. Organometallics 1996,
15, 562-569.
1
indicating the weakness of the B-Cl interaction. H NMR
spectra of 11 show Cs symmetry in CD2Cl2 solution down to
(32) Piers, W. E.; Chivers, T. J. Chem. Soc. ReV. 1997, 26, 345-354.

Carbanion Addition to an Osmium Nitrido
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1065
tion with the unsymmetrical BPh(Tol)2 suggested that the
crossover products would have been resolvable. There is thus
no scrambling of N-aryl or B-aryl groups either during or after
the reaction. The reaction of 1 with BPh(Tol)2 gave, after
hydrolysis, a 7:1 ratio of 5:4. In a similar competition study,
B(C6F5)3 and BPh3 were mixed 1:1 in CDCl3 (there was no
evidence of exchange over 3 h), and 0.33 equiv of 1 was added.
1
Prompt analysis by H and 19F NMR showed only 11 and
unreacted BPh3 and B(C6F5)3, and aerobic workup gave 4 in
good yield.
Complex 1 also reacts rapidly with BPhCl2 to give a complex
1
similar to 11 and 12, as determined by H NMR. Hydrolysis
1
yielded primarily 4, with small amounts of H NMR-silent
materials. Complex 1 does not react with 2-phenyl-1,3,2-
dioxaborinane, Ph[BOCH2CH2CH2O], even after 3 days at 70
°C in benzene solution. Heating a solution of 1 and excess
NaBPh4 in acetonitrile results in a very slow reaction with some
formation of 4. This suggests the presence of trace water and
that the reaction could have proceeded by initial hydrolysis of
-
the BPh4 to BPh3.
The alkylboranes BEt3 and BEt2(OMe) also add to 1, forming
TpOs[N(Et)B(Et)X]Cl2 (X ) Et, 15; OMe, 16; eq 9). BEt3 reacts
Figure 6. ORTEP diagram of TpOs[NB(C6F5)3]Ph2 (17).
1
the course of several days at room temperature, and H NMR
spectra of the resulting purple solution show only the disap-
pearance of starting material. No products analogous to 11 or
17 are observed.
Complex 17 has a 1H NMR spectrum indicative of a
diamagnetic complex, consistent with the Os(VI) formulation.
The equivalence of pairs of pyrazole and phenyl groups shows
Cs symmetry. Only the chemical shifts of the trans pyrazole
are dramatically shifted from 3, as expected from perturbation
of the nitrido ligand. The similarity of the spectra shows that
the phenyl groups remain on the osmium. 19F NMR indicates
the presence of only one type of C6F5 group, and the chemical
shift of -81.6 ppm for the para fluorine atoms is strongly
suggestive of an adduct.36 The X-ray structure of 17 (Tables 1
and 2, Figure 6) confirms the formulation. As is typical of
nitrido-borane adducts,1,37 the Os-N(7) bond (1.688(3) Å) is
only slightly lengthened from that in 3 (1.634(4) Å). The B(2)-
N(7) bond length (1.591(5) Å) is a long B-N single bond (see
above). There is a π stacking interaction (∼3.2 Å) between one
of the C6F5 groups on the boron and one of the C6H5 groups on
the metal center. The Os-N-B angle (167.7(3)°) is slightly
bent, apparently to accommodate this interaction. While B(C6F5)3
and TpOs(N)Ph2 (3) could both be considered fairly bulky, the
presence of the intervening nitrogen leads to an Os-B distance
of 3.26 Å and little steric interaction.
rapidly, while BEt2(OMe) requires over 40 min at room
temperature to form 16. The same reactivity pattern is observed
for arylboranes: BAr3 reacts faster than Ph2BOBPh2. The
1
products are assigned on the basis of their H NMR spectra,
both showing standard Cs symmetry pyrazole resonances and
two types of ethyl groups. The ethyl groups bound to boron
are broadened by the quadrupolar boron nucleus (as observed
in free ethyl boranes), but the ethyl groups bound to nitrogen
show a clear triplet/quartet pattern. A few resonances are shifted
out of the diamagnetic region. The FAB/MS of 15 shows a weak
parent ion with stronger signals due to M+-BEt2 and M+-
BEt2Cl. Brief atmospheric exposure of 16 in CDCl3 gives clean
conversion to TpOs(NHEt)Cl2 (10), which has been observed
in low yield in the reaction of 1 with EtMgCl (see above). It
has been characterized by its mildly paramagnetically shifted
1H NMR spectrum with Cs symmetry and one ethyl group, and
its FAB mass spectrum (M+, M+ - Cl). In contrast, 15 decays
in the air without formation of new 1H NMR-active Tp species.
The diphenyl nitrido complex 3 rapidly adds B(C6F5)3,
forming a green complex even as the methylene chloride solvent
is thawing. The product is the Lewis acid adduct TpOs[NB-
(C6F5)3]Ph2 (17; eq 10), in which the nitrido ligand simply binds
to the boron without insertion into a B-C bond. Exposure of
Ab Initio Studies of TpOs(N)Cl2 (1) and TpOs(N)Me2 (18).
Ab initio calculations have been performed on 1 and the
hypothetical dimethyl analogue TpOs(N)Me2 (18). The calcula-
tions were done prior to our obtaining an X-ray structure of 1,
38
so the coordinates of the related TpTc(O)Cl2 were used as a
(36) Massey, A. G.; Park, A. J. J. Organomet. Chem. 1966, 5, 218-
225.
(37) Ritter, S.; Abram, U. Inorg. Chim. Acta 1995, 231, 245-248. (b)
Ritter, S.; Abram, U. Z. Anorg. Allg. Chem. 1996, 622, 965. (c) Datona,
R.; Schweda, E.; Stra¨hle, J. Z. Naturforsch. B 1984, 39, 733-736. (d)
Yoshida, T.; Adachi, T.; Yabunouchi, N.; Ueda, T.; Okamoto, S. J. Chem.
Soc., Chem. Commun. 1994, 151-152. (e) Bishop, M. W.; Chatt, J.;
Dilworth, J. R.; Dahlstrom, P.; Hyde, J.; Zubieta, J. J. Organomet. Chem.
1981, 213, 109-124.
17 to the atmosphere results in rapid conversion back to 3. In
contrast, BPh3 reacts with TpOs(N)Ph2 (3) in benzene only over
(38) Thomas, R. W.; Estes, G. W.; Elder, R. C.; Deutsch, E. J. Am. Chem.
Soc. 1979, 101, 4581.

1066 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
CreVier et al.
Figure 8. Plots of the LUMO (a) and LUMO + 1 (b) of TpOs(N)Cl2
(1), from XR/LANL2DZ/RHF/LANL2MB calculations.
Figure 7. Partial molecular orbital diagram for TpOs(N)Cl2 (1) and
TpOs(N)Me2 (18), from XR/LANL2DZ/RHF/LANL2MB calculations.
Energies are given relative to the HOMO of 1.
This is illustrated by the 0.7 eV lower HOMO (dxy) orbital
energy in 1. The predominantly nitrogen lone pair orbital in 1
(the HOMO-8) is 1 eV lower in energy and has less frontier
orbital character than the equivalent orbital (the HOMO-3) in
18. The difference is even larger for the LUMO/LUMO + 1,
with those in 1 being 1.3 eV lower. This means that the HOMO/
LUMO gap is smaller in 1 than in 18 (1.6 vs 2.2 eV). Even
larger differences in orbital energies are observed when compar-
ing [Os(N)Cl4]- vs [Os(N)Me4]- because four ligands are
different instead of two, as in 1 vs 18 (see Supporting
Information). These energetic differences are not apparent in
the composition of the LUMOs: even for 18 the π* orbitals
have substantial nitrogen character.
starting point. The molecular geometry was optimized at the
RHF/LANL2MB level of theory, with imposed Cs symmetry.
For 18, the computed structure of 1 was used to start, with Os-C
bonds of 2.15 Å. There is reasonably good agreement between
the computed structure and the X-ray results (Table 2), although
the computations underestimate the trans effect of the nitrido
ligand and give an overly long Os-Cl distance. Calculations
on the simpler [Os(N)X4]- complexes (X ) Cl,39 Me40) show
similar discrepancies at this level of theory but better agreement
upon optimization at the XR/LANL2DZ level (see Supporting
Information). The differences between the computational and
crystallographic structures of 1 and 18 should have little effect
on the orbital energies and compositions. These orbital properties
were determined by single-point calculations at the optimized
geometries at the XR/LANL2DZ level.
The HOMO in both 1 and 18 is a mixture of the Os dxy orbital
and Cl or C p orbitals. It is of δ symmetry with respect to the
nitrido ligand and roughly nonbonding, as drawn in Figure 7.
The orbital energies in this figure are given relative to the
HOMO of 1; similar energy spacings (though not absolute
energies) were observed in RHF calculations. The orbitals below
the HOMO are primarily ligand centered; notable are the
nitrogen lone pair orbital and the OstN π orbitals (the HOMO-8
and HOMO-15,16 in 1). The LUMO and nearly degenerate
LUMO + 1 in both 1 and 18 are the Os-N π* orbitals, one in
the xz plane and one in the yz plane. Each is primarily an out-
of-phase combination of an osmium π-symmetry d orbital, dxz
or dyz, and the corresponding nitrogen p orbital, px or py
(illustrated for 1 in Figure 8). These results fit the widely
accepted Ballhausen and Gray d-orbital splitting pattern for
pseudo-octahedral complexes with one multiply bonded ligand.1a,41
The remarkable feature is the large amount of nitrogen character
in the LUMO and LUMO + 1, roughly comparable to the
amount of osmium d character. Thus, it is not surprising that
an entering nucleophile will attack at nitrogen. The orbital
pattern, with accessible π* LUMOs and low-energy lone pair
and π orbitals, resembles that of CO.
Discussion
The additions of Grignard reagents and aryl boranes to the
Os(VI) nitrido complex TpOs(N)Cl2 (1) are novel. The nitrido
ligand is formally a closed-shell anion and would normally be
expected to add electrophiles, not nucleophiles. There is much
interest in exploiting such a reversal of nitrogen reactivity.7
I. The Mechanism of Grignard Addition to the Nitrido
Ligand. The reaction of 1 with 1 equiv of PhMgBr initially
generates a dark orange material, indicated to be the osmium-
(IV) complex [TpOs(NPhMgCl)Cl2] (Scheme 1). It is a more
reactive version of the borylamido complexes 11-13. Both can
be viewed as having a very basic TpCl2OsIVdNR- fragment
+
that is strongly stabilized by a Lewis acidic MgCl+ or BX2
fragment. They are better described as metalla-analido rather
than imido complexes. Osmium(IV) imido complexes without
an associated cation would be rare examples of d4 octahedral
imido complexes,42 in which two electrons formally occupy OsN
π* orbitals (the LUMOs in 1).1 The antibonding interaction
should be minimized on decreasing the Os-N-Ph angle and
binding a cation to this electron pair.
The key issue is how the C-N bond is formed. Three possible
mechanisms have been considered (Scheme 2). Path I involves
an initial electron transfer to generate a phenyl radical and
[TpOs(N)Cl2•-], followed by trapping of the phenyl radical to
give the observed intermediate. Grignard reagents often react
by initial electron transfer, and rapid radical trapping is well
The frontier orbital energies of 1 are lower than those of 18,
a result of the lesser donating character of chloride vs methyl.
(42) Wigley, D. E. Prog. Inorg. Chem. 1994, 42, 239-482. (b) Coia, G.
M.; Devenney, M.; White, P. S.; Meyer, T. J.; Wink, D. A. Inorg. Chem.
1997, 36, 2341-2351 and references therein. (c) Arndtsen, B. A.; Sleiman,
H. F.; Chang, A. K.; McElwee-White, L. J. Am. Chem. Soc. 1991, 113,
4871-4876. (d) Au, S.-M.; Huang, J.-S.; Yu, W.-Y.; Fung, W.-H.; Che,
C.-M. J. Am. Chem. Soc. 1999, 121, 9120-9132. (e) Reference 31. (f) d4-
Rhenium imido-alkyne complexes have nonoctahedral structures with a
different electronic structure: Williams, D. S.; Schrock, R. R. Organome-
tallics 1993, 12, 1146-1160.
(39) Phillips, F. L.; Skapski, A. C. J. Cryst. Mol. Struct. 1975, 5, 83-
92.
(40) Chosen as a computational model for [Os(N)(CH2SiMe3)4]-: (a)
Belmonte, P. A.; Own, Z.-Y. J. Am. Chem. Soc. 1984, 106, 7493-7496.
(b) Shapley, P. A.; Own, Z.-Y.; Huffman, J. C. Organometallics 1986, 5,
1269-1271.
(41) Balhausen, C. J.; Gray, H. B. Inorg. Chem. 1962, 1, 111.

Carbanion Addition to an Osmium Nitrido
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1067
Scheme 2
Schrock et al. have suggested that an unobserved intermediate
[N3N]MsP(Li)Ph complex eliminates LiPh to give the terminal
phosphido complexes, [N3N]MtP ([N3N] ) (Me3SiNCH2-
CH2)3N; M ) Mo, W).48 Hydride addition to an electrophilic
imido ligand occurs via initial attack at a carbonyl ligand.49
II. The Mechanism of Arylboron Addition to the Nitrido
Ligand. The insertion of the nitrido ligand of 1 into B-C bonds
(reactions 7 and 9) is, to our knowledge, without precedent for
any metal-ligand multiple bond. A number of organo- and
haloboranes form Lewis acid adducts with nitrido complexes,
such as Re(NBPh3)(Et2dtc)2(PMePh2) (Et2dtc ) diethyldithio-
carbamate) and TpOs[NB(C6F5)3]Ph2 (17).1,37 But these reactions
do not transfer a boron substituent to nitrogen. Close analogies
to reactions 7 and 9 are the insertion of nitrogen into B-C bonds
by chloramines and related reagents (eq 11),50 and C f N
migration of R groups in the Hoffman rearrangement.51 Reaction
precedented for metal-oxo complexes.43 Complex 1 can be
reduced, albeit at low potentials (Ered ) -0.98 vs SSCE).16c,23
However, this pathway can be easily ruled out for the present
reactions because the reaction proceeds in high yield in THF
solvent, a good trap for Ph• (k ) 4.8 × 106 M-1 s-1).44 An
osmium species of concentration e10-4 M would have to react
11 occurs by [1,2] migration of R- from boron to nitrogen, with
displacement of the N-X bonding electrons to give X- (Cl,
OSO3H, etc.). The relative rates of arylborane addition to 1
support a similar mechanism of R- migration, with two electrons
displaced from an Os-N π bond to the Os center. The reaction
of BPh(Tol)2 shows that, in a direct intramolecular competition,
tolyl is 3.5 times more likely to be transferred to nitrogen than
phenyl (after statistical correction). Preferential transfer of the
more electron-rich group is characteristic of carbanion migra-
tions to an electron-deficient center.51 Almost all organic and
organometallic [1,2] migrations are of this type.50,51 [1,2]
Migrations of R- are symmetry allowed, while R+ moving to
an electron pair is formally forbidden.51
In contrast to the preference for tolyl over phenyl migration
with BPh(Tol)2, the intermolecular competition of BPh3 with
B(Tol)3 shows a small preference for phenyl over tolyl migra-
tion. This shows that the products are not simply determined
by migratory aptitudes and indicates a multistep mechanism.
The same conclusion is derived from the large variation in rates
of phenylborane additions: BPh3 (within seconds) > Ph2-
with Ph• with an impossible rate constant of g1012 M-1 s-1
,
faster than the diffusion limit, to compete with trapping by 12.3
M THF solvent.
Path II involves initial phenyl addition to the osmium center
to form 2, followed by [1,2] phenyl migration to nitrido ligand.
Phenyl migration to an oxo ligand has been observed for a
related rhenium(VII) oxo complex.21a However, this pathway
involves trapping a coordinatively unsaturated intermediate with
Cl-, and no bromide incorporation is observed when PhMgBr
is used. To more rigorously test for the intermediacy of metal-
aryl complexes, TpOs(N)(Ph)Cl (2) and TpOs(N)(Ph)2 (3) were
synthesized. Both species are stable, even at elevated temper-
atures, with no evidence for aryl migration. Reactions of 2 or
3 with excess PhMgBr-d5 give partially deuterated 7 in low
yield, but the reactions are much slower than the reaction of 1
under the same conditions. In addition, the protio phenyl group
remains on the osmium, indicating that aryl migration does not
occur in the formation of 7. Complexes 2 and 3 are not
kinetically or mechanistically competent to be intermediates,
ruling out path II. Intermediacy of a seven-coordinate species
[TpOs(N)Ph3-xClx-] (x ) 0-2) seems unlikely,45 and it would
have to have nonexchanging phenyl groups.
BOBPh2 (stirring overnight) > Ph[O(BOCH2CH2CH2O] (un-
reactive over 3 days at 70 °C). This is too big an effect to solely
reflect changes in the nucleophilicity of the Ph-B bonds. It
indicates that the Lewis acidity of the boron plays a key role.
This accounts for the rapid reaction of 1 with B(C6F5)3, almost
as fast as that with BPh3: the very high Lewis acidity of
B(C6F5)3 compensates for C6F5 being a much poorer migrator
than C6H5.
The data indicate a stepwise mechanism (Scheme 3) involving
reversible formation of an initial borane adduct A, followed by
[1,2] migration of R- to an electrophilic nitrogen, similar to eq
11. The reaction with BPh(Tol)2 reflects the relative rates only
The data indicate a mechanism of direct addition of the
carbanion to the nitrido ligand (path III). We know of no clear
precedent for such a carbanion addition to a nitrido, oxo, imido,
or sulfido ligand. This reaction is, however, well known for
carbene and carbonyl ligands.46 A recently reported reaction
proceeds by direct attack of LiMe at a selenido ligand bound
to tantalum.47 In the microscopic reverse of this mechanism,
(43) Cook, G. K.; Mayer, J. M. J. Am. Chem. Soc. 1994, 116, 1855-
1868 and references therein.
(44) Scaiano, J. C.; Stewart, L. C. J. Am. Chem. Soc. 1983, 105, 3609-
3614.
(48) Mo¨sch-Zanetti, N. C.; Schrock, R. R.; Davis, W. M.; Wanninger,
K.; Seidel, S. W.; O’Donoghue, M. B. J. Am. Chem. Soc. 1997, 119,
11 037-11 048.
(49) Luan, L.; Brookhart, M.; Templeton, J. L. Organometallics 1992,
11, 1433-1435.
(50) Brown, H. C. Organic Synthese Via Boranes; Wiley: New York,
1975. (b) Matteson, D. S. Stereodirected Synthesis with Organoboranes;
Springer: New York, 1995; pp 48-57. (c) Onak, T. Organoborane
Chemistry; Academic Press: New York, 1975; pp 100-101, 106. (d)
Kabalka, G. W.; Wang, Z. Organometallics 1989, 8, 1093-1095.
(51) March, J. AdVanced Organic Chemistry, 4th ed.; Wiley: New York,
1992; discussions of [1,2] migrations, p 1051ff (see also ref 21a); Hoffman
rearrangment, pp 1090-1091.
(45) While there are precedents for seven-coordinate transition metal Tp
complexes, examples seem to be restricted to larger metal centers, such as
OsI and OsII (Trofimenko, S. Chem. ReV. 1993, 93, 943-980. Burns, I. D.;
Hill, A. F.; White, A. J. P.; Williams, D. J.; Wilton-Ely, J. D. E. T.
Organometallics 1998, 17, 1552-1557), and these are often fluxional. For
related ReV-oxo complexes with the B(pz)4- ligand, there is a preference
for dissociating a pyrzole rather going seven-coordinate (Paulo, A.;
Domingos, A.; Santos, I. Inorg. Chem. 1996, 35, 1798-1807).
(46) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles
and Applications of Organotransition Metal Chemistry; University Science
Books: Mill Valley, CA, 1987; pp 401-406.
(47) Shin, J. H.; Parkin, G. Organometallics 1995, 14, 1104.

1068 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
CreVier et al.
Scheme 3a
V vs Cp2Fe+/0 in MeCN].23 The pattern of higher electrophilicity
for halide vs organometallic derivatives has been observed
previously, such as in electrophilic [Os(N)Cl4]- vs nucleophilic
[Os(N)(CH2SiMe3)4]- (which is harder to reduce by a remark-
able 1.45 V).8 Calculations on these compounds are in general
agreement with the results given here.12
It is interesting in this context to consider the analogy between
1 and CO, advanced when 1 was observed to act as a π acid
ligand for late transition metals.56 Both 1 and CO have low-
lying LUMOs of π symmetry and low nucleophilicity. CO
inserts into B-C bonds of organoboranes much like 1, and a
similar mechanism is proposed, with alkyl group migration from
B to C occurring in an initial CO-borane adduct.57 It seems
likely that the reactions of 1 with BR3, ArMgX, and ArLi all
proceed through an initial N-M adduct. Binding of the Lewis
acid to the nitrido ligand increases its electrophilicity, facilitating
the migration. The strategy of enhancing the electrophilic
reactivity of nitrido ligands by binding Lewis acids has been
employed by a number of groups1,5a,b,58 and is also well
established for carbon monoxide.
a Rate ) Keqkmig[TpOs(N)Cl2][BR3].
of the migration step. In the intermolecular competition,
however, the faster tolyl migration step is more than balanced
by the stronger pre-equilibrium binding of the more Lewis acidic
BPh3. An alternative mechanism of concerted, direct attack of
the B-C bond at nitrogen does not account for the role of the
boron Lewis acidity. Initial interaction of the electrophilic
nitrogen with the aromatic ring is conceivable for the arylboranes
but is inconsistent with the facile reaction with ethylboranes.
(Initial phenyl transfer from boron to osmium is unlikely, given
the experiments described above with Grignard reagents and
the low reactivity of 3.)
The characterization of the borane adduct TpOs[NB(C6F5)3]-
Ph2 (17) provides further support for Scheme 3, as it is an
excellent model for the initial borane adduct A. Such adducts
are not observed for 1, both because they react further and
because 1 is an unusually weak nucleophile (it is inert to BF3‚
Et2O, MeOTf, and [Ph3C][BF4] at ambient temperatures). TpOs-
(N)Ph2 (3) is a much better nucleophile, favoring the initial
equilibrium formation of 17. And the lower electrophilicity of
3 inhibits the aryl migration step, allowing observation of 17.
In a related reaction, an adduct has been observed between
B(C6F5)3 and a bis(amino)silylene prior to migration of C6F5
from B to Si.52
III. Electrophilic and Nucleophilic Character of the
Nitrido Ligand. Nitrido ligands can show nucleophilic and/or
electrophilic character.1,53,54 The dominant reactivity of 1 is as
an electrophile, adding Ph- from PhMgX and BPh3 (and adding
PPh325). Meyer and co-workers have elegantly shown the
electrophilicity of 1 and related Os(VI) nitrido complexes, for
instance by their addition of amines and phosphines.55 In
contrast, the diphenyl derivative TpOs(N)Ph2 (3) is a poor
electrophile, reacting only slowly with PhMgCl and PPh3.
The DFT calculations provide insight into the reactivity of
the nitrido ligand and how it is tuned by the ancillary ligands.
In 1, the nearly degenerate LUMO and LUMO + 1 (Os-N π*
orbitals) are low in energy and have substantial nitrogen
character. These orbitals are higher in energy in the dimethyl
complex 18, used as a computational model for the diphenyl
complex 3. This is the reason for the higher electrophilicity of
1 versus 2 and 3 (and 18). Conversely, the nitrogen lone pair is
higher and closer to the HOMO in 18 (Figure 7), consistent
with 3 forming a stable adduct with B(C6F5)3, while 1 has very
low nucleophilicity. These trends are mirrored in the redox
potentials: adding hydrocarbyl ligands makes reduction sub-
stantially more difficult [-1.35 (1), -1.89 (2), -2.04 (3) V]
and oxidation much more facile [∼2.05 (1), 1.60 (2), 1.36 (3)
IV. Implications: Reactions at Ligands. The reactions of
1 occur predominantly at the nitrido ligand and not at osmium.
In a direct comparison, external addition of Ph- to 1 is facile,
while [1,2] phenyl migration from Os to N in 2 and 3 does not
occur. Both pathways involve adding Ph- to the multiply bonded
ligandswhy is external attack favored? One reason is that the
phenyl group in PhMgX or PhLi is a better nucleophile than an
osmium-bound phenyl group. Perhaps more importantly, the
presence of a phenyl ligand on osmium strongly decreases the
electrophilicity of the nitrido ligand (see above). And in the
cases described here, binding of a Lewis acid enhances the
electrophilicity of the multiple bond.
These ideas suggest that direct addition of carbanions and
other good nucleophiles to metal-ligand multiple bonds (eq
12) is likely to be more facile than initial addition to the metal
followed by migration. Preference for direct attack is frequently
seen in the more developed literature on metal carbonyl
complexes. For instance, CpRe(CO)(NO)Cl reacts with hard,
nucleophilic sources of Ph- directly at the electrophilic carbonyl
to produce an anionic acyl complex. The related phenyl complex
CpRe(CO)(NO)Ph, prepared by chloride metathesis using the
softer phenylcuprate, does not undergo phenyl migration to
CO.59 This system is closely analogous to the reactivity of 1
with Ph- sources. If the preference for direct attack over
metathesis/migration is general, re-examination of the reactions
of V(O)Cl3 with arylmercury compounds60 and Cr(Ntbutyl)2-
(OSiMe3)2 with ZnPh261 may be warranted. In both cases [1,2]
migration was suggested, but direct attack at the multiple bond
is a good alternative. Direct attack may be occurring in many
other systems, such as the alkylation of OsO4 with (Me3SiCH2)2-
(56) Crevier, T. J.; Lovell, S.; Mayer, J. M. J. Chem. Soc., Chem.
Commun. 1998, 2371-2372.
(57) Onak, T. Organoborane Chemistry; Academic Press: New York,
1975. (b) Hillman, M. E. D. J. Am. Chem. Soc. 1962, 84, 4715.
(58) Shapley, P. A.; Shusta, J. M.; Hunt, J. L. Organometallics 1996,
15, 1622-1629.
(52) Metzler, N.; Denk, M. Chem. Commun. 1996, 2657-2658.
(53) Cf., Mo(N)(tetraphenylporphyrin): Kim, J. C.; Lee, B. M.; Shin, J.
I. Polyhedron 1995, 14, 2145-2149.
(54) Electrophilic reactivity of oxo ligands: (a) Holm, R. H. Chem. ReV.
1987, 87, 1401-1449. (b) Mayer, J. M. AdV. Trans. Met. Coord. Chem.
1996, 1, 105-157. (c) Seymore, S. B.; Brown, S. N Inorg. Chem. 2000,
39, 325-332.
(59) Sweet, J. R.; Graham, W. A. G. J. Organomet. Chem. 1983, 241,
45-51.
(60) Reichle, W. T.; Carrick, W. L. J. Organomet. Chem. 1970, 24, 419-
426. (b) Thiele, V. K. H.; Schumann, W.; Wagner, S.; Bru¨ser, W. Z. Anorg.
Allg. Chem. 1972, 390, 280-288.
(55) Huynh, M. H. V.; El-Samanody, E.-S.; Demadis, K. D.; White, P.
S.; Meyer, T. J. Inorg. Chem. 2000, 39, 2825-2830 and references therein.
(b) Reference 16.
(61) Nugent, W. A.; Harlow, R. L. J. Am. Chem. Soc. 1980, 102, 1759-
1760.

Carbanion Addition to an Osmium Nitrido
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1069
Mg to give Os(O)(CH2SiMe3)4.62 Finally, it should be noted
that the preference for direct attack observed here is consistent
with the recent conclusion that the cis-hydroxylation of olefins
by OsO4 proceeds via olefin attack at oxygen, rather than
involving a 1,2 alkyl migration.9
The insertion of a nitrido ligand into B-C bonds (eqs 7 and
9) is an interesting example of a reaction that occurs at a ligand.
This reaction resembles oxidative addition and σ-bond metath-
esis processes in that an X-Y bond is cleaved and the X and
Y fragments are added to the metal complex. In oxidative
addition, both X and Y bind to the metal center with concomitant
oxidation of M. In σ-bond metathesis, X is added to the metal
and Y is added to a ligand, without redox change at the metal.
In reactions 7 and 9, both the R and BX2 pieces are added to
the nitrido ligand, and the osmium is formally reduced by two
electrons. Current work is exploring the scope of such reactions
that occur at ligands, without change in the atoms bound to the
metal center.
peaks: δ (multiplicity, coupling constant, number of protons, assign-
ment). All pyrazole JHH couplings are 2 Hz. 13C NMR spectra were
recorded on Bruker AC-200 and AF-300 spectrometers at ambient
temperatures and referenced to the solvent. 19F chemical shifts were
recorded at 188 MHz and are reported in ppm with CF3COOH used as
an external reference. IR spectra were obtained as thin films or Nujol
mulls on NaCl plates using a Perkin-Elmer 1600A spectrometer and
are reported in wavenumbers. Each complex contains a number of
stretches associated with the Tp ligand that typically vary by less than
3 wavenumbers: 3112 (m), 1503 (m), 1404 (s), 1312 (s), 1216 (s),
1118 (s), 1075 (m), 1046 (vs), 982 (w), 768 (s), 715 (s), 657 (m), 615
(m). EI/MS spectra were obtained on a Kratos Analytical mass
spectrometer at 70 eV using electron impact ionization; samples were
introduced as solids using a direct inlet probe techniques and were
typically heated to 150 °C. FAB mass spectra were taken on a Joel
VG 70 SEQ tandem hybrid instrument of EBqQ geometry. Samples
were quickly introduced into the matrix (4-nitro-benzyl alcohol) as
methylene chloride solutions.
KOs(N)O3 was prepared from OsO4 by literature methods15 or
purchased from Colonial and used without further purification. KTp
was synthesized by literature methods.63 Triphenylphosphine was
purchased from Strem and used as received. The reported syntheses of
[nBu4N][Os(N)Ph4] and [nBu4N][Os(N)Ph2Cl2]20 were modified to use
an aerobic, aqueous workup. B(C6F5)3 was generously supplied by
Professor K. I. Goldberg. All other reagents were purchased from
Aldrich and used without further purification. Solvents were degassed
and dried according to standard procedures.64 Chromatography was on
silica gel unless otherwise noted.
Conclusions
The osmium(VI) nitrido complex TpOs(N)Cl2 (1) is electro-
philic at nitrogen but is a poor nucleophile. PhMgCl and related
reagents deliver carbanions directly to the nitrido ligand to form
reactive metalla-analido complexes which are readily protonated
to analido complexes such as TpOs(NHPh)Cl2 (4). These
reactions proceed by direct addition of R- to the nitrido ligand,
and not via osmium-carbon bonded intermediates such as
TpOs(N)PhCl (2). Independently prepared 2 and TpOs(N)Ph2
(3) add external C6D5MgCl at nitrogen, without migration of
osmium-bound phenyl groups. A mechanism of electron transfer
from PhMgCl to 1, forming Ph•, is ruled out by the lack of
trapping of the radical by THF solvent. Reactions of 1 with
alkyl- and arylboranes similarly result in transfer of one organic
group to nitrogen, leading to isolable borylamido complexes
such as TpOs[N(Ph)(BPh2)]Cl2 (11). This is an unprecedented
insertion of a nitrido ligand into a boron-carbon bond. The
mechanism proposed involves pre-equilibrium formation of a
nitrido-borane adduct, followed by a 1,2 shift of an organic
group from boron to nitrogen (Scheme 3). A related adduct,
TpOs[NB(C6F5)3]Ph2 (17), has been structurally characterized.
The mechanism is supported by the strong influence of both
the electrophilicity of the borane and the nucleophilicity of the
migrating group. Similar mechanisms have been proposed in
related reactions of organoboranes with chloramines, azides, and
CO. The unusual reactivity of 1, and the much diminished
reactivity of 3, are consistent with the orbital picture derived
from ab initio calculations. The electrophilicity of 1 derives from
its accessible LUMO and LUMO + 1, which are Os-N π*
orbitals with significant density on the nitrogen. The lone pair
on the nitrido ligand in 1 is the low-lying HOMO-8. This
bonding scheme and reactivity have similarities to those of CO.
The reactions described here are interesting examples of metal-
mediated chemistry that occurs with bond making and bond
breaking at ligands, without change in the atoms bound to the
metal center.
TpOs(N)Cl2 (1). A 250 mL round-bottom flask was charged with
7.00 g (27.8 mmol) of KTp, 120 mL of dry ethanol, and 32.0 mL of
concentrated aqueous HCl. KOsNO3 (2.00 g, 6.87 mmol) was added,
resulting in a deep purple-colored solution. The solution was refluxed
for 30 min, turning a reddish orange color, and then it was cooled and
placed in a freezer overnight. The chilled solution was filtered, and a
mixture of white and orange-red solids was collected. The solids were
rinsed with ∼100 mL of H2O and ∼20 mL of ethanol and air-dried for
several hours, yielding 1.50 g of 1 (3.07 mmol, 45% yield) as a red-
orange power. To remove the TpOsCl3 impurity, 250 mg of 1 (0.51
mmol) containing 5 mol % TpOsCl3 (as determined by 1H NMR) was
dissolved in CH2Cl2, and ferrocene (4.83 mg, 0.026 mmol, 1.05 equiv
based on TpOsCl3) was added. After the solution was stirred for 1 h,
the solvent was removed, and the resulting solid was purified by
chromatography and two recrystallizations, affording red plates suitable
1
for X-ray crystallography. H NMR (CDCl3): 6.03 (t), 7.42 (d), 7.52
(d) (each 1H, pz); 6.55 (t), 7.88 (d), 8.38 (d) (each 2H, pz′). 13C{1H}
NMR (CDCl3): 106.3, 134.5, 143.9 (pz trans to N), 108.6, 138.1, 146.5
(pz trans to Cl). EI/MS: 489 (M+), 455 (M+ - Cl). IR: 2516 (BsH),
1066 (OstN). Anal. Calcd (found) for C9H10BCl2N7Os: C, 22.14
(22.35); H, 2.06 (2.09); N, 20.09 (19.89).
TpOs(15N)Cl2 was prepared on the same scale from KOs(15N)O3
(from OsO4 + 15NH4OH). IR: 1036 (υOst N).
15
TpOs(N)(Ph)Cl (2). A glass bomb was loaded with 3 (35 mg, 0.060
mmol), 5 mL of CH2Cl2, and then 1 M HCl in Et2O (0.435 mL, 0.435
mmol; 7.25 equiv) and then heated at 60 °C for 2 days. Column
chromatography using toluene as an eluant yielded 15 mg (0.028 mmol,
47%) of orange powder. 1H NMR (CDCl3): 5.94, 6.43, 6.53 (each 1H,
t, pz), 7.00, 7.47, 7.79, 7.88, 7.92, 8.50 (each 1H, d, pz), 7.15 (1H, t,
7 Hz; para), 7.25 (2H, t, 7 Hz; meta), 7.35 (2H, d, 7 Hz; ortho). 13C-
{1H} NMR: 106.7, 108.5, 109.1, 135.1, 137.9, 138.7, 145.0, 146.4,
148.4 (all pz), 127.3, 128.1, 142.8 (all -Ph, p, m, o, ipso not observed).
EI/MS: 531 (M+), 495 (M+ - Cl).
TpOs(N)Ph2 (3). Concentrated aqueous HI (1 mL) in air was added
to a stirred solution of [nBu4N][Os(N)Ph4] (400 mg, 0.530 mmol) in
100 mL of acetone, resulting in a rapid color change from yellow to
bright orange. After 5 min, KTp (200 mg, 0.794 mmol) was added,
and the solution gradually darkened. This mixture was heated briefly
to 60 °C and then allowed to stir for 4 h at room temperature.
Chromatography using toluene as an eluant gave 3 (100 mg, 0.175
Experimental Section
General Considerations. All manipulations were conducted in the
absence of air and water using dried reagents and glovebox and high-
vacuum techniques, except where noted. 1H NMR spectra were recorded
on Bruker AC-200, AF-300, AM-499, and WM-500 spectrometers at
ambient temperatures and referenced to Me4Si or residual solvent
(63) Trofimenko, S. J. Am. Chem. Soc. 1966, 88, 1842-1844.
(64) Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory
Chemicals 3rd ed.; Pergamon Press: New York, 1988.
(62) Alves, A. S.; Moore, D. S.; Andersen, R. A.; Wilkinson, G.
Polyhedron 1982, 1, 83-87.

1070 J. Am. Chem. Soc., Vol. 123, No. 6, 2001
CreVier et al.
1
mmol, 33%) as an orange powder. H NMR (CDCl3): 5.99 (t, 1H, t,
dark orange 9 (35.9 mg, 0.071 mmol, 35% yield) as dark orange plates.
IR (KBr): 3566, 3244, 2954, 2497 (B-H), 1713, 1592, 1499, 1408,
1306, 1210, 1118, 1074, 1049, 763, 708, 655, 618. DIP/MS: 506 (M+),
475 (M+ - H2NEt). Anal. Calcd (found) for C10H15BCl2N7Os: C, 23.78
(24.11); H, 2.99 (3.22); N, 19.41 (19.07).
pz), 6.38 (2H, t, pz′), 6.86, 7.55 (each 1H, d, pz), 7.80, 7.89 (each 2H,
d, pz′), 7.02 (2H, t, 7 Hz; para), 7.15 (4H, t, 7 Hz; meta), 7.36 (4H, d,
7 Hz; ortho). 13C{1H} NMR: 105.4, 134.5, 144.8 (pz trans to N), 107.1,
136.8, 146.2 (pz trans to Ph), 125.1 (para), 127.4 (meta), 141.4 (ortho),
155.1 (ipso). EI/MS: 572 (M+), 495 (M+ - Ph). Anal. Calcd (found)
for C21H20BN7Os: C, 44.14 (44.74); H, 3.53 (3.69); N, 17.16 (17.24).
TpOs(NHPh)Cl2 (4). PhMgBr in THF (200 mg in ∼10 mL) was
added dropwise over 5 min to a -78 °C solution of 1 (100 mg, 0.204
mmol) in 20 mL of THF, resulting in a darkening of the orange solution.
The solution was stirred at -78 °C for 1 h, and then at room temperature
for 1 h, and then was exposed to the atmosphere, resulting in a rapid
color change from dark orange to deep red. Chromatography using 50:
50 CH2Cl2/hexanes allowed the isolation of 4 (94 mg, 0.165 mmol,
TpOs(NHEt)Cl2 (10). A solution of 16 (generated in situ, see below)
was exposed to air for approximately 10 min, and 2 µL of H2O was
1
added. The solution slowly turned a pale yellow. H NMR (CD2Cl2):
6.27 (t, 1H, pz), 6.98, 6.55 (d, 1H, pz), 6.50 (t, 2H, pz′), 6.27, 7.17 (d,
2H, pz′), 19.24 (q, 6.7 Hz, 2H, NCH2CH3), 1.24 (t, 6.7 Hz, 3H,
NCH2CH3), 3.4 (br, NH + H2O). FAB/MS: 519 (st, M+), 483 (v st,
M+ - Cl). Attempts to work up 10 by column chromatography
following the procedures for 4-7 yielded only 1H NMR-silent yellow
materials.
TpOs[N(Ph)(BPh2)]Cl2 (11). A flask charged with 180 mg of 1
(0.368 mmol), 108 mg of BPh3 (0.446 mmol, 1.21 equiv), and 10 mL
of C6H6 was stirred for 5 min. Filtering, layering the filtrate with 60
mL of hexanes, and it allowing to stand overnight yielded dark orange
11 (138 mg, 52%). 1H NMR (CDCl3): 6.29 (t, 1H, pz), 6.94, 7.47 (d,
1H, pz), 6.12 (t, 2H, pz′), 6.05, 7.00 (d, 2H, pz′), 4.50 (d, 7 Hz, 2H,
NPho), 7.26 (t, 7 Hz, 2H, NPhm), 6.90 (t, 7 Hz, 1H, NPhp), 7.63 (d, 7
Hz, 4H, BPh2ortho), 7.32 (t, 7 Hz, 4H, BPh2meta), 7.43 (t, 7 Hz, 2H,
BPh2para). Anal. Calcd (found) for OsC27H25N7B2Cl2: C, 44.40 (44.03);
H, 3.45 (3.42); N, 13.43 (13.42).
1
81%). H NMR (CDCl3): -3.50 (d, 7.0 Hz, 2H, NPho), 8.69 (t, 7.0
Hz, 2H, NPhm), 0.01 (t, 7.0 Hz, 1H, NPhp), 5.50 (br s, 1H, NH), 6.06
(t, 1H, pz), 5.38, 6.19 (each d, 1H, pz), 6.62 (t, 2H, pz′) 6.82, 4.59
(each d, 2H, pz′). IR: 3519 (NH), 2508 (BH). EI/MS: 567 (M+). Anal.
Calcd (found) for C15H16BCl2N7Os: C, 31.81 (31.87); H, 2.85 (2.98);
N, 17.33 (16.87). UV/vis (CH2Cl2) [nm (ꢀ, Μ-1 cm-1)]: 256 (12 300),
339 (8200), 420 (7700), 504 (12 800).
TpOs(NHTol)Cl2 (5). A bomb was charged with toluidine (15.5
mg, 145 µmol), 1 (63.6 mg, 130 µmol), and 5 mL of CH2Cl2. The
bomb was heated at 70 °C for 1 day and then opened in the atmosphere,
and another 15.0 mg (141 µmol) of toluidine was added. The bomb
was heated for another 6 days at 70 °C. Column chromatography with
CH2Cl2 as the eluant yielded 32 mg of red/violet powder 5 (55 µmol,
42%) in addition to ∼20 mg of 1. 1H NMR (CDCl3): 6.14 (t, 1H, pz),
6.56, 5.78 (each d, 1H, pz), 6.58 (t, 2H, pz′), 6.68, 5.05 (each d, 2H,
pz′), 8.02, -2.39 (each d, 2H, 7 Hz, tolyl), 10.26 (s, 3H, CH3), 7.59
(br s, 1H, NH). EI/MS: 581 (M+), 545 (M+ - Cl).
TpOs(NHPh)(Ph)Cl (6). A procedure analogous to that for 4 was
used, with 2 equiv of PhMgBr. H NMR (CDCl3): -2.49 (d, 7.5 Hz,
2H, NPho), 8.57 (t, 7.5 Hz, 2H, NPhm), 1.00 (t, 7.5 Hz, 1H, NPhp),
2.27 (d, 7.5 Hz, 2H, OsPho), 8.54 (t, 7.5 Hz, 2H, OsPhm), 3.93 (t, 7.5
Hz, 1H, OsPhm), 11.5 (br, 1H, NH), 6.92, 6.70, 5.88 (t, 1H, pz), 7.62,
6.36, 6.21, 5.63, 4.99, 4.73 (d, 1H, pz). EI/MS: 609 (M+).
TpOs(NHPh)Ph2 (7). A procedure analogous to that for 4 was used,
with 3 equiv of PhMgBr. H NMR (CD2Cl2): -3.66 (d, 7.5 Hz, 2H,
NPho), 8.69 (t, 7.5 Hz, 2H, NPhm), -0.69 (t, 7.5 Hz, 1H, NPhm), 1.55
(d, 7.5 Hz, 4H, OsPho), 9.17 (t, 7.5 Hz, 4H, OsPhm), 3.50 (t, 7.5 Hz,
2H, OsPhm), 12.1 (br, 1H, NH), 5.84 (t, 1H, pz), 5.62, 4.08 (d, 1H,
pz), 6.91 (t, 2H, pz′), 7.61, 5.46 (d, 2H, pz′). IR: 3526 (NH), 2514
(B-H). EI/MS: 650 (M+).
TpOs[N(Ph)(BPhOBPh2)]Cl2 (12). Compound 12 was prepared as
1
for 11 except with stirring overnight. Isolated yield: 34%. H NMR
(CDCl3): 5.96 (1H, t, pz), 5.73, 4.72 (each 1H, d, pz), 6.69 (2H, t,
pz′), 6.96, 4.12 (each 2H, d, pz′); Ph, 9.20 (2H, t), 7.70 (4H, t), 7.47
(1H, t), 4.10 (2H, t), 1.91 (1H, t), 1.27 overlapping (3H, d), 0.89 (2H,
t), -4.10 (4H, t). Repeated submittals of the same sample gave
substantially different elemental analyses, indicating analytical dif-
ficulties.
1
B(Tol)3 was prepared by addition of LiTol in pentane (from PhCH2I,
n
2.18 g, 10 mmol + BuLi, 6.25 mL/1.6 M in hexanes, 10 mmol) to a
pentane solution of BBr3 (0.32 mL, 3.4 mmol). Filtration and removal
1
of the volatiles gave 0.75 g (2.6 mmol, 76%) of white B(Tol)3. H
NMR (CDCl3): 7.71, 7.25 (d, 7 Hz, 2H each), 2.43 (s, 3H). EI/MS:
284 (M+), 193 (M - Tol+).
1
BPh(Tol)2 was prepared similarly from PhCH2I (3.26 g, 15 mmol),
nBuLi (9.4 mL/1.6 M in hexanes, 15 mmol), and PhBCl2 (0.97 mL,
7.5 mmol). The white product (0.50 g, 1.85 mmol, 25%) was sublimed
1
and recrystallized from pentane. H NMR (CDCl3): 7.54, 7.27 (d, 2
Hz, 4H, Tol), 7.60 (d, 2H), 7.52 (t, 1H), 7.43 (t, 2H, each 8 Hz, Ph o,
p, m).
TpOs(NPh)Ph2MgCl. A screwtop NMR tube was charged with 50
µL of 2 M PhMgCl in THF (0.100 mmol). The solvent was removed
in vacuo, and 0.50 mL of C6D6 was condensed in. The tube was shaken,
and then the solvent was stripped off in vacuo. The tube was brought
into the box, where it was charged with 14.0 mg (0.028 mmol, 0.28
equiv) of 1. The contents were dissolved in 1 mL of C6D6, resulting in
TpOs(N(C6F5)[B(C6F5)2])Cl2 (13). A solution of 1 (78 mg, 160
µmol), B(C6F5)3 (78 mg, 152 µmol), and 40 mL of benzene was stirred
for 10 min. The solvent was removed, and the solids were extracted
with 100 mL of pentane. Removal of the pentane yielded 13 (80 mg,
1
80 µmol, 53% yield) as a reddish orange powder. H NMR (CDCl3):
0.80 (t, 1H, pz), -3.28 (d, 1H, pz), -18.62 (br, 1H, pz), 5.32 (t, 2H,
pz′), 8.66 (d, 2H, pz′), -23.31 (br, 2H, pz′). 19F NMR: 3.9 (br s, 2F),
-19.5 (br s, 2F), -39.4 (br s, 2F), -76.8 (br s, 1F), -84.0 (br s, 4F),
-85.9 (br s, 4F).
TpOs[NH(C6F5)]Cl2 (14). A methylene chloride solution of 13 was
exposed to air, resulting in an immediate color change from dark orange
to vivid purple. Purification of the solution by short column chroma-
tography yielded 14 in 90% yield. 1H NMR (CDCl3): 4.99 (t, 1H, pz),
1.46, 1.34 (d, 1H, pz), 6.61 (t, 2H, pz′), 7.76, -1.19 (d, 2H, pz′). 19F
NMR (CDCl3): 0.2 (d, 6 Hz, 2F, ortho), -120.4 (t, 6 Hz, 2F, meta),
36.9 (t, 6 Hz, 1F, para). FAB/MS: 658 (strong, M+), 623 (strong, M+
- Cl), 475 (very strong, M+ - NH(C6F5)).
1
a dark orange solution of TpOs(NPh)Ph2MgCl as determined by H
NMR: 2.86 (d, 7.5 Hz, 2H, NPho), 8.21 (t, 7.5 Hz, 2H, NPhm), 4.25 (t,
7.5 Hz, 1H, NPhm), 5.75 (d, 7.5 Hz, 4H, OsPho), 7.88 (t, 7.5 Hz, 4H,
OsPhm), 6.02 (t, 7.5 Hz, 2H, OsPhm), 5.88 (t, 1H, pz), 6.14, 6.78 (d,
1H, pz), 6.12 (t, 2H, pz′), 6.81, 6.85 (d, 2H, pz′).
TpOs(H2NEt)Cl2 (8). Et2Zn (0.200 mmol in 20 mL of THF) was
added dropwise over 5 min to a -78 °C solution of 1 (0.100 mg, 0.204
mmol) in 35 mL of THF. The solution immediately turned dark purple.
After being stirred for 3 h at -78 °C and 5 h at room temperature, the
solution was exposed to air, affording a dark brown solution. Purifica-
tion by silica gel chromatography and recrystallization in air from
acetone/heptane gave NMR-silent 8 (27.5 mg, 0.053 mmol, 26%) as
red plates. IR (KBr): 3241, 2967, 2505 (B-H), 1704, 1500, 1408,
1308, 1213, 1115, 1047, 838, 782, 711, 620. DIP/MS: 520 (M+), 475
(M+ - H2NEt). Anal. Calcd (found) for C11H17BCl2N7Os: C, 25.45
(25.63); H, 3.30 (3.47); N, 18.88 (19.01).
TpOs[N(Et)BEt2]Cl2 (15). A screwtop NMR tube was loaded with
1 (4.3 mg, 8.8 µmol) and 0.5 mL of methylene chloride. Addition of
BEt3 (10 µL, 1 M in THF, 10 µmol) resulted in an immediate color
change to vibrant yellow. The solvent was removed in vacuo, and 0.5
1
mL of CDCl3 was condensed in. H NMR (CDCl3): 6.33 (t, 1H, pz),
TpOs(H2NMe)Cl2 (9). Following the procedure for 8, MeMgCl
(0.200 mmol in 20 mL of THF) was added to 1 (0.100 mg, 0.204 mmol
in 35 mL of THF) and stirred at -78 °C for 1 h, and then at 20 °C for
2 h. Exposure to air afforded a blue solution. Purification by silica gel
chromatography and recrystallization in air from CH2Cl2/heptane gave
7.50, 7.26 (d, 1H, pz), 6.34 (t, 2H, pz′), 7.14, 7.05 (d, 2H, pz′), 0.91 (t,
7.4 Hz, 3H, NCH2CH3), 7.100 (obscured q, 2H, NCH2CH3), 1.29 (br
s, 10H, BCH2CH3). FAB/MS: 587 (w, M+), 572 (vw, M+ - Me), 558
(w, M+ - Et), 552 (w, M+ - Cl), 537 (M+ - MeCl), 519 (st, M+
BEt2), 483 (v st, M+ - BEt2Cl).
-

Carbanion Addition to an Osmium Nitrido
J. Am. Chem. Soc., Vol. 123, No. 6, 2001 1071
TpOs{N(Et)[BEt(OMe)]}Cl2 (16). A screwtop NMR tube was
loaded with 1 (8.0 mg, 16.4 µmol) and 0.5 mL of CD2Cl2. Et2B(OMe)
(2.5 µL, 19 µmol) was added, and the color of the solution changed
SHELXL-97.65 All heavy atoms were located in the Patterson map and
refined anisotropically; hydrogen atoms were placed in calculated
positions and refined using a riding model. Crystals of 13 were grown
by slow evaporation of a benzene solution; the structure is described
in the Supporting Information.
Crystal Structure of TpOs[NB(C6F5)3]Ph2 (17). A green, prismatic
crystal of 17 was immersed in oil and mounted on a glass capillary
with epoxy under a stream of nitrogen. Data were collected on a Nonius
KappaCCD with two sets of 10 s exposures using a κ offset of 100
between the two sets. An empirical correction for absorption anisotropy
was applied before structure solution.66 Solution by Patterson methods
produced a complete phasing model consistent with the proposed
structure.67 Two benzene molecules of crystallization were located by
difference Fourier synthesis.68 All hydrogen atoms were placed with
idealized geometry and were refined isotropically with a riding model
with Uiso values fixed at 1.1Ueq of their parent atom. All non-hydrogen
atoms were refined anisotropically.
1
from orange to dark yellow over 40 min. H NMR (CD2Cl2): 6.19 (t,
1H, pz), 6.21, 6.68 (d, 1H, pz), 6.56 (t, 2H, pz′), 5.99, 7.32 (d, 2H,
pz′), 4.38 (s, 3H, OMe), 1.52 (t, 7.3 Hz, 3H, BCH2CH3), 1.44 (d, 8.3
Hz, 2H, NCH2CH3), 0.67 (t, 8.3 Hz, 3H, NCH2CH3), 15.9 (br s, 2H,
BCH2CH3). 13C{1H} NMR (CD2Cl2): 149.7, 136.9, 102.8 (pz, 1C),
150.0, 135.0, 102.8 (pz′, 2C).
TpOs[NB(C6F5)3]Ph2 (17). An NMR tube was charged with 3 (12.4
mg, 22 µmol), B(C6F5)3 (12.3 mg, 24 µmol), and then CD2Cl2 (0.5
1
mL). Upon thawing, a green solution was observed, whose H NMR
did not change on heating for 8 h at 80 °C. Removal of the CD2Cl2
and recrystallization from benzene gave air-sensitive green crystals
suitable for X-ray analysis. 1H NMR (CD2Cl2): 5.89 (t, 1H, pz), 6.33,
∼7.00 (d, 1H, pz), 6.39 (t, 2H, pz′), 7.71, 8.00 (d, 2H, pz′), 6.98 (d, 7
Hz, 4H, OsPho), 7.18 (t, 7 Hz, 4H, OsPhm), 7.05 (t, 7 Hz, 2H, OsPhm).
19F NMR (CD2Cl2): -54.4 (d, 18 Hz, 6F, ortho), -81.6 (t, 19 Hz, 3F,
para), -88.3 (t, 18 Hz, 6F, meta).
Computational Details. All computations were performed using the
GAUSSIAN-94 package.69 The geometries of all molecules were
optimized using idealized molecular symmetries at the RHF/LANL2MB
level of theory. Complexes 4 and 5 were also optimized at the XR/
LANL2DZ level of theory. Single-point calculations were then
performed on all molecules using XR/LANL2DZ.
Crystal Structures of 3, 4, and 8. Crystals were grown by slow
evaporation of hexane (3), hexane/ethyl acetate (4), or acetone/heptane
solutions (8). Crystals were mounted on a glass fiber with epoxy (3, 4)
or in a glass capillary in oil (8). Data were collected on a Nonius
KappaCCD diffractometer, by rotation about the φ axis in 1° increments
over 360°. The exposure times per frame and the crystal-detector
distances were 60 s and 27 mm for 3, 30 s and 30 mm for 4, and 30
s and 27 mm for 8. The merging R factors of 0.052 (3), 0.037 (4), and
0.056 (4) indicated the data were all of good quality. Solutions by direct
methods produced a complete phasing model for 3 and 8, and an almost
complete phasing model for 4, with the rest located by difference
Fourier synthesis. In 3 and 4 there are two independent molecules in
the asymmetric unit. 8 crystallizes with one molecule of acetone; both
the Os complex and the solvent reside on a mirror plane. All heavy
atoms were refined anisotropically by full matrix least-squares. For 3,
hydrogen atoms H1A, 2A, 3, 6, 7, 8, 9, 11, 12, 13, 15, 19, 20, 22, 23,
24, 25, 28, 32, 33,35, 36, 39, 41, and 42 were located in the difference
maps; the remaining ones were placed in calculated positions. For 4
and 8, all hydrogen atoms were placed in calculated positions, except
for the boron hydrogen in 8 (H1B). Hydrogens were given isotropic
parameters of 1.1Ueq of their parent atom (3) or fixed isotropic
parameters of 0.05 (4) and refined with a riding model. The min/max
residual electron densities were located near the Os atom and were not
of chemical significance. Empirical absorption corrections were applied
before structure solution.
Acknowledgment. We gratefully acknowledge the NSF for
support to J.M.M. and for support of an upgrade of our 500
MHz NMR spectrometers (Grant No. 9710008).
Supporting Information Available: Crystallographic data
for 1, 3, 5, 8, 9, 11, 12, 13, and 17 (for 4, see ref 10a) (PDF).
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA0028424
(65) Sheldrick, G. M. SHELXL93. In Crystallographic Computing 6;
Flack, P., Parkanyi, P., Simon, K., Eds.; IUCr/Oxford University Press:
Oxford, U.K., 1993.
(66) Blessing, R. H. Acta Crystallogr. 1995, A51, 33.
(67) Beurskens, P. T.; Admirall, G.; Beurskens, G.; Bosman, W. P.;
Garcia-Granda, S.; Gould, R. O.; Smits, J. M. M.; Smykalla, C. DIRDIF,
University of Nijmegen, Nijmegen, The Netherlands.
(68) Sheldrick, G. M. SHELXL97, Program for the Refinement of Crystal
Structures, University of Gottingen, Gottingen, Germany.
(69) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.;
Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.; Keith, T.; Petersson, G.
A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski,
V. G.; Ortiz, J. V.; Foresman, J. B.; Peng, C. Y.; Ayala, P. Y.; Chen, W.;
Wong, M. W.; Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.;
Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J. P.; Head-
Gordon, M.; Gonzalez, C.; Pople, J. A., Gaussian 94, Revision B.3;
Gaussian, Inc.: Pittsburgh, PA, 1995.
Crystal Structures of 11 and 12. Crystals of 11 and 12, obtained
on slow evaporation from CH2Cl2/Me3SiOSiMe3 solutions, were
immersed in oil and mounted under a stream of N2. Data were collected
on an Enraf-Nonius CAD4 Diffractometer. Semiempirical absorption
corrections used ψ-scans. Full-matrix least-squares refinement used