Ditantalum hydride complexes with bridging (2,6-iPr2C6H3)NSiHPh silanimine ligands resulting from PhSiH3-imido ligand coupling. A combined spectroscopic and theoretical investigation

Article abstract of DOI:10.1021/om000455x

The preparation and characterization of two novel dinuclear tantalum hydride complexes featuring bridging silanimine ligands are reported. The reaction of C_p*(ArN double bond)Ta[Si(SiMe₃)₃]H (C_p* = η⁵-C₅Me₅; Ar = 2,6-ⁱPr₂C₆H₃) with PhSiH₃ resulted in formation of [C_p*(ArN double bond)TaH(μ-H)]₂ (4% yield), yellow, paramagnetic Cp*₂(ArN double bond)Ta₂H₂(μ-ArNSiHPh) (1, 18% yield), and dark green, diamagnetic Cp*₂Ta₂H₂(μ-ArNSiHPh)₂ (2, 71% yield). For 1 and 2, X-ray structure determinations characterize the ArNSiHPh silanimine ligand as possessing a Si-N single bond. This is confirmed by molecular orbital calculations that provide an average bond order of 0.7 for the Si-N bond. The ArNSiHPh fragment is therefore best viewed as a bifunctional silyl-amido ligand. For diamagnetic complex 2, the X-ray analysis revealed a molecular structure possessing nearly exact 2-fold symmetry (the hydride ligands were not located), while NMR spectroscopy indicates that the two Cp*Ta(μ-ArNSiHPh) fragments in the molecule are chemically inequivalent. To analyze the structure and bonding in this compound, a theoretical study based on density functional theory and ab initio molecular dynamics was carried out. Calculations of the entire 140-atom dinuclear tantalum system confirm a structure with an asymmetric substitution of the two hydride ligands, with one terminal and one bridging. The paramagnetic compound 1 exhibits structural features that are similar to those for 2. For this complex, the spectroscopic data and density functional calculations are consistent with a structure featuring terminal and bridging hydride ligands.

Full text of DOI:10.1021/om000455x

3
830
Organometallics 2000, 19, 3830-3841
Dita n ta lu m Hyd r id e Com p lexes w ith Br id gin g
i
(
2,6- P r C H )NSiHP h Sila n im in e Liga n d s Resu ltin g fr om
2
6
3
P h SiH -Im id o Liga n d Cou p lin g. A Com bin ed
3
Sp ectr oscop ic a n d Th eor etica l In vestiga tion
Urs Burckhardt, Gary L. Casty, and T. Don Tilley*
Department of Chemistry, University of California at Berkeley,
Berkeley, California 94720-1460
Tom K. Woo and Ursula Rothlisberger*
Laboratory of Inorganic Chemistry, ETH Zentrum, Universit a¨ tsstrasse 6,
8
092 Z u¨ rich, Switzerland
Received May 30, 2000
The preparation and characterization of two novel dinuclear tantalum hydride complexes
featuring bridging silanimine ligands are reported. The reaction of Cp*(ArNd)Ta-
5
i
[
[
(
Si(SiMe
Cp*(ArNd)TaH(µ-H)]
1, 18% yield), and dark green, diamagnetic Cp*
3
)
3
]H (Cp* ) η -C
(4% yield), yellow, paramagnetic Cp*
Ta (µ-ArNSiHPh)
5
Me
5
; Ar ) 2,6- Pr
2
C
6
H
3
) with PhSiH
(ArNd)Ta
(2, 71% yield). For 1
3
resulted in formation of
2
2
2 2
H
(µ-ArNSiHPh)
2
2
H
2
2
and 2, X-ray structure determinations characterize the ArNSiHPh silanimine ligand as
possessing a Si-N single bond. This is confirmed by molecular orbital calculations that
provide an average bond order of 0.7 for the Si-N bond. The ArNSiHPh fragment is therefore
best viewed as a bifunctional silyl-amido ligand. For diamagnetic complex 2, the X-ray
analysis revealed a molecular structure possessing nearly exact 2-fold symmetry (the hydride
ligands were not located), while NMR spectroscopy indicates that the two Cp*Ta(µ-
ArNSiHPh) fragments in the molecule are chemically inequivalent. To analyze the structure
and bonding in this compound, a theoretical study based on density functional theory and
ab initio molecular dynamics was carried out. Calculations of the entire 140-atom dinuclear
tantalum system confirm a structure with an asymmetric substitution of the two hydride
ligands, with one terminal and one bridging. The paramagnetic compound 1 exhibits
structural features that are similar to those for 2. For this complex, the spectroscopic data
and density functional calculations are consistent with a structure featuring terminal and
bridging hydride ligands.
0
Molecular d metal chemistry has expanded rapidly
little is known about the influence of ancillary ligands
on reactivity. We have been particularly interested in
delineating electronic and steric factors that control the
course of σ-bond metathesis reactions involving d metal
centers. In this context, we have recently begun to
in recent years, as it has become increasingly apparent
that d centers can mediate a number of bond-forming
transformations despite being inactive in oxidative
addition-reductive elimination cycles. This chemistry
has allowed the development of various catalytic reac-
0
0
explore the chemistry of M-Si and M-H bonds in
1
7-9
tions (e.g., olefin polymerization, carbon-carbon cou-
complexes supported by ancillary imido ligands.
pling,2 olefin metatheses, hydrogenations,
3
2,4
hydro-
Early transition metal imido complexes have at-
5
6
10
silations, and dehydropolymerizations ) which feature
migratory insertions, cycloadditions, and/or σ-bond me-
tathesis. Despite the considerable attention that early
transition metal chemistry has received, surprisingly
tracted considerable recent attention, and such com-
plexes have led to the development of efficient alkene
3
0
metathesis catalysts. However, d imido complexes
(
5) (a) Fu, P.-F.; Brard, L.; Li, Y.; Marks, T. J . J . Am. Chem. Soc.
(
1) (a) Brintzinger, H. H.; Fischer, D.; Mulhaupt, R.; Rieger, B.;
1995, 117, 7157. (b) Molander, G. A.; J ulius, M. J . Org. Chem. 1992,
57, 6347. (c) Molander, G. A.; Dowdy, E. D.; Noll, B. C. Organometallics
1998, 17, 3754. (d) Carter, M. B.; Schiott, B.; Gutierrez, A.; Buchwald,
S. L. J . Am. Chem. Soc. 1994, 116, 11667.
Waymouth, R. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143. (b) J ordan,
R. F. Adv. Organomet. Chem. 1991, 32, 325. (c) Kaminsky, W. Catal.
Today 1994, 20, 257. (d) Piers, W. E. Chem. Eur. J . 1998, 4, 13.
(
2) Hoveyda, A. H.; Morken, J . P. Angew. Chem., Int. Ed. Engl. 1996,
5, 1262.
3) (a) Ivin, K. J . Olefin Metathesis and Metathesis Polymerization;
Academic Press: San Diego, 1997. (b) Grubbs, R. H.; Tumas, W. Science
989, 243, 907. (c) Schrock, R. R. Acc. Chem. Res. 1990, 23, 158. (d)
(6) (a) Gauvin, F.; Harrod, J . F.; Woo, H. G. Adv. Organomet. Chem.
1998, 42, 363. (b) Tilley, T. D. Acc. Chem. Res. 1993, 26, 22. (c) Corey,
J . In Advances in Silicon Chemistry; Larson, G., Ed.; J AI Press:
Greenwich, CT, 1991; Vol. 1, p 327.
(7) Casty, G. L.; Tilley, T. D.; Yap, G. P. A.; Rheingold, A. L.
Organometallics 1997, 16, 4746.
(8) Burckhardt, U.; Tilley, T. D. J . Am. Chem. Soc. 1999, 121, 6328.
(9) Burckhardt, U.; Casty, G. L.; Tilley, T. D. Manuscript in
preparation.
3
(
1
Schrock, R. R. Tetrahedron 1999, 55, 8141. (e) Schuster, M.; Blechert,
S. Angew. Chem., Int. Ed. Engl. 1997, 36, 2036.
(
4) (a) Lin, Z.; Marks, T. J . J . Am. Chem. Soc. 1987, 109, 7979. (b)
Lee, N. E.; Buchwald, S. L. J . Am. Chem. Soc. 1994, 116, 5985.
1
0.1021/om000455x CCC: $19.00 © 2000 American Chemical Society
Publication on Web 08/17/2000

Ditantalum Hydride Complexes
Organometallics, Vol. 19, No. 19, 2000 3831
Sch em e 1
have not featured prominently in reactions involving
mechanical calculations on large systems with transi-
tion metal centers. In this work, we have applied density
functional theory and ab initio molecular dynamics
methods in combination with experimental studies to
examine the structure and bonding of compounds 1 and
2. The entire 140 atom diamagnetic complex 2, and the
whole 127-atom paramagnetic complex 1, have been
calculated at the density functional level on a parallel
computer.
σ-bond metathesis, and in fact imido complexes pos-
0
sessing reactive σ bonds (e.g., d M-H and M-Si) are
exceedingly rare. It is therefore uncertain whether imido
ligands would play an innocent role in the presence of
hydride or silyl ligands, which in principle could migrate
to the imido nitrogen atom to produce an amido group.
Tantalum imido hydrides of the type Cp*2Ta(dNR)H
5
t
(
Cp* ) η -C5Me5; R ) H, Ph, Bu) appear to form via
R-migration of hydrogen to tantalum in amido com-
plexes of the type Cp*2TaNHR.11 We have recently
Resu lts a n d Discu ssion
reported the synthesis of stable molybdenum and
Syn th esis of th e Dita n ta lu m Hyd r id es 1 a n d 2.
tungsten imido complexes, such as (ArNd)2Mo[Si-
Addition of phenylsilane (1 equiv) to a hexanes solution
i
(
SiMe3)3]CH2CMe3 (Ar ) 2,6- Pr2C6H3), containing a
9
of Cp*(ArNd)Ta[Si(SiMe3)3]H resulted in an initial
7
silyl ligand. However in this system the silyl hydrides
ArNd)2M[Si(SiMe3)3]H (M ) Mo, W) are highly reactive
color change from red to yellow, and then a dark green
color developed more slowly. Workup of the reaction
mixture after 12 h resulted in isolation of three products
by fractional crystallization (Scheme 1). The first prod-
uct crystallized from the reaction mixture, in 4% yield,
(
and could not be isolated. More recently, we have
succeeded in preparing stable tantalum imido complexes
containing hydride and silyl ligands, such as [Cp*-
(
)
ArNd)TaH(µ-H)]2 and Cp*(ArNd)Ta[Si(SiMe3)3]H (Ar
is a microcrystalline yellow solid identified as the pre-
i
9
2,6- Pr2C6H3). In this contribution, we describe
viously synthesized dihydride [Cp*(ArNd)TaH(µ-H)]2.9
results obtained from a search for σ-bond metathesis
chemistry in the reaction of Cp*(ArNd)Ta[Si(SiMe3)3]H
with PhSiH3. While this reaction is complex, it is
interesting in that it involves a dehydrogenative Si-N
coupling with production of a bridging ArNSiHPh
silanimine ligand. The structures of two products of this
coupling reaction, Cp*2(ArNd)Ta2H2(µ-ArNSiHPh) (1)
and Cp*2Ta2H2(µ-ArNSiHPh)2 (2), reveal unusual fea-
tures that have been investigated in detail via spectro-
scopic and theoretical methods.
Further concentration and cooling of the reaction solu-
tion induced the successive crystallizations of yellow,
paramagnetic Cp*2(ArNd)Ta2H2(µ-ArNSiHPh) (1, 18%
yield) and dark green, diamagnetic Cp*2Ta2H2(µ-ArN-
SiHPh)2 (2, 71% yield). Compounds 1 and 2 were
characterized by single-crystal X-ray diffraction, infra-
red and multinuclear NMR spectroscopy, combustion
analysis, and mass spectrometry. The spectroscopic
signatures for the hydride ligands in 1 and 2 are
consistent with those established for other dinuclear
Due to the enormous progress in computer power and
the development of efficient computational methods, it
has become possible to perform first-principles quantum
12
tantalum hydrides.
Interestingly, wide variations in the reaction stoichi-
ometry had virtually no effect on the observed distribu-
(
10) (a) Nugent, W. A.; Mayer, J . M. Metal-Ligand Multiple Bonds;
(12) (a) Belmonte, P. A.; Schrock, R. R.; Day, C. S. J . Am. Chem.
Soc. 1982, 104, 3082. (b) Scioly, A. J .; Luetkens, M. L., J r.; Wilson, R.
B., J r.; Huffman, J . C.; Sattelberger, A. P. Polyhedron 1987, 6, 741.
(c) Ting, C.; Messerle, L. J . Am. Chem. Soc. 1989, 111, 3449. (d) Fryzuk,
M. D.; McConville, D. H. Inorg. Chem. 1989, 28, 1613. (e) Profilet, R.
D.; Fanwick, P. E.; Rothwell, I. P. Polyhderon 1992, 11, 1559. (f) Miller,
R. L.; Toreki, R.; LaPointe, R. E.; Wolczanski, P. T.; van Duyne, G. D.;
Roe, D. C. J . Am. Chem. Soc. 1993, 115, 5570. (g) Cotton, F. A.; Daniels,
L. M.; Murillo, C. A.; Wang, X. J . Am. Chem. Soc. 1996, 118, 12449.
Wiley-Interscience: New York, 1988. (b) Wigley, D. E. Prog. Inorg.
Chem. 1994, 42, 239. (c) Gibson, V. C. J . Chem. Soc., Dalton Trans.
1
994, 1607.
11) (a) Parkin, G.; van Asselt, A.; Leahy, D. J .; Whunnery, L.; Hua,
(
N. G.; Quan, R. W.; Henling, L. M.; Schaefer, W. P.; Santarsiero, B.
D.; Bercaw, J . E. Inorg. Chem. 1992, 31, 82. (b) Antonelli, D. M.;
Schaefer, W. P.; Parkin, G.; Bercaw, J . E. J . Organomet. Chem. 1993,
4
62, 213.

3
832 Organometallics, Vol. 19, No. 19, 2000
Burckhardt et al.
Ta ble 1. Selected Bon d Len gth s (Å) a n d An gles
a
(
d eg) for Com p ou n d 1
Bond Length
Ta(1)-Ta(2)
Ta(1)-Si(1)
Ta(1)-N(2)
Ta(1)-Cp*(1)cent
Ta(2)-Si(1)
Ta(2)-N(1)
3.0190(3)
2.709(1)
1.812(4)
2.1605(2)
2.648(1)
1.963(4)
2.1088(2)
1.832(4)
N(1)-C(27)
1.423(6)
1.401(6)
1.95(5)
2.17(5)
1.79(5)
1.72(5)
1.42(5)
N(2)-C(39)
Ta(1)-H(71)
Ta(1)-H(72)
Ta(2)-H(71)
Ta(2)-H(72)
Si(1)-H(70)
Ta(2)-Cp*(2)cent
Si(1)-N(1)
Bond Angles
54.75(3) Si(1)-Ta(1)-H(71)
110.2(1) Si(1)-Ta(1)-H(72)
Ta(2)-Ta(1)-Si(1)
Ta(2)-Ta(1)-N(2)
75(1)
76(1)
82(1)
Ta(2)-Ta(1)-Cp*(1)cent 126.95(1) N(2)-Ta(1)-H(71)
Si(1)-Ta(1)-N(2)
Si(1)-Ta(1)Cp*(1)cent
N(2)-Ta(1)Cp*(1)cent
Ta(1)-Ta(2)-Si(1)
Ta(1)-Ta(2)-N(1)
Ta(1)-Ta(2)-Cp*(2)cent 137.10(1) Ta(2)-Si(1)-H(70)
Si(1)-Ta(2)-N(1) 43.8(1) N(1)-Si(1)-H(70)
95.1(1)
N(2)-Ta(1)-H(72)
140(1)
134.15(3) Cp*(2)cent-Ta(2)-H(71) 108(1)
118.5(1) Cp*(2)cent-Ta(2)-H(72) 111(2)
56.65(3) H(71)-Ta(2)-H(72)
97.9(1) Ta(1)-Si(1)-H(70)
71(2)
102(2)
131(2)
104(2)
F igu r e 1. ORTEP diagram of Cp*
SiHPh) (1).
2 2 2
(ArNd)Ta H (µ-ArN-
Si(1)-Ta(2)-Cp*(2)cent 163.50(3) Cp*(1)cent-Ta(1)-H(71) 136(1)
N(1)-Ta(2)-Cp*(2)cent
Ta(1)-Si(1)-Ta(2)
Ta(1)-Si(1)-N(1)
Ta(2)-Si(1)-N(1)
Ta(2)-N(1)-Si(1)
Ta(2)-N(1)-C(27)
Ta(1)-N(2)-C(39)
Ta(2)-Ta(1)-H(71)
Ta(2)-Ta(1)-H(72)
119.9(1)
68.60(3) H(71)-Ta(1)-H(72)
Cp*(1)cent-Ta(1)-H(72)
93(1)
59(2)
38(1)
45(2)
79(1)
85(2)
104(1)
127(2)
tion of products (by NMR spectroscopy of the reaction
mixture). Thus, the same product ratio was obtained
with substoichiometric quantities of PhSiH3 and when
the reaction was carried out in neat PhSiH3. However,
the product distribution does depend on the reaction
time. While [Cp*(ArNd)TaH(µ-H)]2 is generated im-
mediately upon addition of the silane, the hydrides 1
and 2 form more slowly and in an approximately
constant ratio, as shown by NMR analysis of the cooled
reaction solution at various time intervals.
113.1(1)
47.8(1)
88.4(2)
150.4(3)
170.4(3)
34(1)
Ta(1)-Ta(2)-H(71)
Ta(1)-Ta(2)-H(72)
Si(1)-Ta(2)-H(71)
Si(1)-Ta(2)-H(72)
N(1)-Ta(2)-H(71)
N(1)-Ta(2)-H(72)
34(1)
a
Cp*cent represents the average of the x, y, and z coordinates of
5
the η -C5Me5 ring carbons.
At this point, the detailed mechanism of the reaction
in Scheme 1 is unknown. As expected, the reaction
produces a quantitative amount of HSi(SiMe3)3 and
elemental hydrogen (by NMR spectroscopy and GC/MS),
but oligosilanes (which might be expected to form via
covalent bonds, as seen by comparisons to structurally
characterized tantalum silyls such as Cp*Cl TaSiMe
3
3
1
4
(Ta-Si ) 2.669(4) Å). This bridging interaction results
in a very acute Ta(1)-Si(1)-Ta(2) bond angle of 68.60(3)°.
The silanimine ligand appears to possess very little Sid
N double bond character, as indicated by the Si(1)-N(1)
bond length of 1.832(4) Å. For comparison, the SidN
6
silane dehydrocoupling ) were not observed. Reaction
of the deuteride Cp*(ArNd)Ta[Si(SiMe3)3]D with Ph-
SiH3 produced mainly DSi(SiMe3)3 (80%), but minor
quantities of PhSiH3-nDn (n ) 1-3) and HD were also
t
t
15
bond length in Bu SidNSi Bu is only 1.568(3) Å, and
2
3
the corresponding distance in the mononuclear silan-
1
2
2
t
observed (by H and H NMR spectroscopy). Upon
reaction of Cp*(ArNd)Ta[Si(SiMe3)3]D with PhSiD3, no
HSi(SiMe3)3 was observed as product. This finding
suggests the operation of competing pathways (including
reductive elimination-oxidative addition and/or σ-bond
metathesis) and is similar to observations made in the
reaction of the 16-electron hafnocene CpCp*Hf[Si-
imine complex Cp Zr(η -Me SidN Bu)(CO) is 1.661(4)
2
2
1
6
Å. In addition, the Ta(2)-N(1) bond length of 1.963(4)
Å is consistent with a Ta(IV)-N single covalent bond,
as seen (for example) by comparisons to the Ta-N
distances of 1.97-1.99 Å in the related ditantalum
1
2g
hydride [(Cy N) ClTa(µ-H)] .
2
2
2
Thus, the silanimine
ligand in 1 is best viewed as a bridging silyl-amido
ligand, with single covalent Ta-Si, Ta-N, and Si-N
bonds. The arylimido ligand in 1 is bound to Ta(1) in
the usual way, as indicated by the nearly linear Ta(1)-
N(2)-C(39) angle (170.4(3)°) and the short Ta(1)-N(2)
distance of 1.812(4) Å. For comparison, the Ta(2)-N(1)-
C(27) angle is 150.4(3)° and the Ta(2)-N(1) distance is
1.963(4) Å. The difference electron density map con-
tained two peaks that were assigned to the hydride
ligands. Both hydrides were characterized as asym-
metrically bridging the two tantalum atoms such that
they are closer to Ta(2): Ta(2)-H(71) ) 1.79(5) Å;
Ta(1)-H(71) ) 1.95(5) Å; Ta(2)-H(72) ) 1.72(5) Å;
Ta(1)-H(72) ) 2.17(5) Å. Note, however, that infrared
(
[
SiMe3)3]H with H2Si(SiMe3)2.13 Attempts to convert
Cp*(ArNd)TaH(µ-H)]2 and 1 to the main product 2 by
heating them in the presence of PhSiH3 failed; there
was no conversion even after refluxing in benzene for
several days. The addition of hydrogen, HSi(SiMe3)3, or
small amounts of AIBN did not initiate the reaction.
Ch a r a cter iza tion of 1. The molecular structure of
1
was determined by X-ray crystallography (Figure 1).
Important bond distances and angles are listed in Table
. The two tantalum atoms are separated by a distance
1
2
of 3.0190(3) Å and are bridged by the unusual µ,η -
ArNSiHPh silanimine ligand. Bonding of the latter
bridging ligand to the two tantalum centers results in
a “five-coordinate” silicon and similar Ta-Si bond
lengths (Si-Ta(1) ) 2.709(1) Å; Si-Ta(2) ) 2.648(1) Å).
These Ta-Si distances correspond to normal, single
(
14) Arnold, J .; Shina, D. N.; Tilley, T. D.; Arif, A. M. Organo-
metallics 1986, 5, 2037.
15) Wiberg, N.; Schurz, K.; Reber, G.; M u¨ ller, G. J . Chem. Soc.,
(
Chem. Commun. 1986, 591.
(16) Procopio, L. J .; Carroll, P. J .; Berry, D. H. Polyhedron 1995,
14, 45.
(
13) Casty, G. L.; Lugmair, C. G.; Radu, N. S.; Tilley, T. D.; Walzer,
J . F.; Zargarian, D. Organometallics 1997, 16, 8.

Ditantalum Hydride Complexes
Organometallics, Vol. 19, No. 19, 2000 3833
data and density functional calculations suggest that
one of the hydride ligands might best be described as
terminal (vide infra).
The infrared spectrum exhibits a weak Si-H stretch
-
1
at 2173 cm , and two broad bands at 1763 and 1653
-1
cm are assigned to the two hydride ligands. While the
higher frequency Ta-H vibration falls well within the
range expected for a terminal tantalum hydride, the
lower one lies somewhere between those for a terminal
and a bridged Ta-H stretch.1 For comparison, the
hydride [Cp*(ArNd)TaH(µ-H)]2 exhibits infrared νTaH
2
-1
-1
stretching bands at 1529 cm (bridging) and 1771 cm
9
(terminal).
F igu r e 2. Comparison of the X-ray and theoretical struc-
ture of the paramagnetic complex Cp* (ArNd)Ta H (µ-
ArNSiHPh) (1). Pendant groups including the dNAr moiety
have been removed for clarity. The bracketed values are
for the experimental X-ray structure.
Because of the paramagnetic nature of compound 1,
its NMR spectra feature only broad signals, which
sharpen somewhat with increases in temperature. The
2
2
2
1
H NMR spectrum contains two very broad signals at
1
0.5 and 0.2 ppm, which each integrate to one hydrogen
and are attributed to the two hydride ligands. The Si-H
resonance appears at 4.93 ppm (singlet) and displays
satellites reflecting a J SiH coupling constant of 220 Hz.
have been determined to be 2.73 and 2.18 Å from the
X-ray analysis. The calculated Ta(1)-H(72) distance of
2.94 Å is almost as long as the Ta-Ta distance of 3.02
Å. As previously mentioned, this asymmetry, although
in disagreement with the X-ray analysis, is in better
agreement with the observed infrared data. We also note
that in the calculated structure, the other bridging
hydride is found to lie closer to Ta(1), whereas from the
X-ray analysis both bridging hydrides were found to lie
closer to Ta(2). Besides the hydride positions, other
geometric parameters found in the calculated structure
are in excellent agreement with the X-ray structure.
Most notably, the Ta-Ta distances agree to within 0.01
Å. Additionally, parameters not shown in Figure 2 are
also in good agreement, such as the Ta(1)-N(2)-C(39)
and Ta(2)-N(1)-C(27) angles, which are calculated to
be 170° and 147° and are experimentally determined
to be 170° and 150°, respectively.
1
Resonances for the two Cp* ligands are observed at 2.14
and 1.95 ppm. Low signal intensities due to the para-
magnetism of 1 precluded direct measurements of the
1
3
29
C and Si NMR spectra; however, the heteroatoms
bound to hydrogen were conveniently observed via H,X
HMQC experiments. Thus, the methyl Cp* resonance
was observed at 12.6 ppm as a single peak, and the
chemical shift for the silicon atom was determined to
be 9.7 ppm.
The magnetic moment of 1 was determined to be µeff
2.0 µB (per dimer, by the Evans method17 at 24 °C).
)
This value is quite low and incompatible with two
independent Ta(IV) centers, and suggests the presence
of a weak antiferromagnetic interaction between the
tantalum atoms. This µeff value does not change with
temperature, up to 90 °C.
The nature of the Si-N and Ta-Ta bonds in 1 can
be analyzed within the framework of molecular orbital
Th eor etica l Ch a r a cter iza tion of 1. Although the
positions of the hydrides in paramagnetic complex 1
were determined by the X-ray diffraction analysis, there
is uncertainty in these assignments due to the proximity
of the hydrides to the heavy Ta atoms. Additionally, the
characterization of both hydrides as bridging the tan-
talum centers is at odds with what is suggested by the
infrared data. For these reasons we examined the
paramagnetic complex with density functional theory
1
8
theory. A Mayer bond order analysis provides a
qualitative measure of the covalent bond strength
between two atoms. As previously stated, the silanimine
ligand appears to possess very little SidN double bond
character. This is confirmed by the calculated bond
order of 0.75 between Si(1) and N(1). To compare, the
calculated bond orders in the same structure are found
to be 1.04 for the N(1)-C(aryl) bond, 0.79 for the Si(1)-
C(aryl) bond, and 1.36 for the Ta(1)dN(2) double bond.
The nature of the Ta-Ta bond is also of interest. The
calculated bond order between the Ta-Ta atoms is 0.13,
which suggests a very weak covalent bond. Mulliken
charges on the two Ta centers are almost identical at
(see Computational Details in the Experimental Sec-
tion). In these calculations the actual ligand system was
included and the calculations were performed within the
spin-unrestricted formalism assuming two unpaired
electrons. Full geometry optimization initiated from the
experimentally derived structure leads to the geometric
parameters given in Figure 2. With the exception of the
hydride positions the calculated geometry is in excellent
agreement with the experimental structure. The asym-
metric bridging arrangement of both hydrides, as de-
termined by the X-ray diffraction analysis, is approxi-
mately reproduced. However, the asymmetry involving
one of the bridging hydrides, H(72), is so severe that it
should be considered a terminal hydride. More specif-
ically, the calculated structure has distances of Ta(2)-
H(72) ) 2.77 and Ta(1)-H(72) ) 2.94 Å, whereas they
+
2.9 e, thereby eliminating the possibility of an ionic
type of interaction. From this analysis we conclude
that there is no significant Ta-Ta bond and that the
3
.02 Å interatomic distance is maintained by the bridg-
ing Ta-Si interactions.
Since the magnetic moment of 1 (2.0 µB) suggests the
presence of weak antiferromagnetic coupling, we have
also repeated the above unrestricted calculations start-
ing from the X-ray structure assuming a total spin
momentum of zero. This open-shell singlet state calcu-
lation also results in a structure that is in good agree-
ment with experiment, again with the exception of the
(
17) (a) Evans, D. F. J . Chem. Soc. 1959, 2003. (b) Sur, S. K. J .
Magn. Reson. 1989, 82, 169.
(18) Mayer, I. Chem. Phys. Lett. 1983, 97, 270-274.

3
834 Organometallics, Vol. 19, No. 19, 2000
Burckhardt et al.
Ta ble 2. Selected Bon d Len gth s (Å) a n d An gles
2
a
(
d eg) for Com p ou n d 2‚ /
3
C
6
H
14
Bond Length
Ta(1)-Ta(2)
Ta(1)-Si(1)
Ta(1)-Si(2)
Ta(1)-N(1)
Ta(1)-Cp*(1)cent
Ta(2)-Si(1)
Ta(2)-Si(2)
Ta(2)-N(2)
2.8472(5)
2.695(3)
2.614(3)
1.957(7)
2.1275(3)
2.644(3)
2.697(2)
1.967(7)
Ta(2)-Cp*(2)cent
Si(1)-N(1)
Si(2)-N(2)
2.1287(3)
1.817(8)
1.824(7)
1.43(1)
1.45(1)
1.553
N(1)-C(7)
N(2)-C(35)
Si(1)-H(81)
Si(2)-H(80)
1.429
Bond Angles
56.92(6) Ta(1)-Ta(2)-Cp*(2)cent 135.39(2)
Ta(2)-Ta(1)-Si(1)
Ta(2)-Ta(1)-Si(2)
Ta(2)-Ta(1)-N(1)
Ta(1)-Si(2)-Ta(2)
Ta(1)-Si(2)-N(2)
59.00(5) Si(1)-Ta(2)-Si(2)
99.2(2) Si(1)-Ta(2)-N(2)
88.45(8)
115.1(2)
64.82(6) Si(1)-Ta(2)-Cp*(2)cent 113.80(6)
111.6(2) Si(2)-Ta(2)-N(2) 42.5(2)
Ta(2)-Ta(1)-Cp*(1)cent 135.51(2) Si(2)-Ta(2)-Cp*(2)cent 157.74(5)
Si(1)-Ta(1)-Si(2)
Si(1)-Ta(1)-N(1)
Si(1)-Ta(1)-Cp*(1)cent 159.36(6) Ta(1)-Si(1)-N(1)
Si(2)-Ta(1)-N(1) 115.3(2) Ta(2)-Si(1)-N(1)
Si(2)-Ta(1)-Cp*(1)cent 111.40(5) Ta(2)-Si(2)-N(2)
N(1)-Ta(1)-Cp*(1)cent 121.2(2) Ta(1)-N(1)-Si(1)
89.12(8) N(2)-Ta(2)-Cp*(2)cent 120.6(2)
42.4(2) Ta(1)-Si(1)-Ta(2)
64.44(6)
46.5(2)
F igu r e 3. ORTEP diagram of Cp*
2).
2 2 2 2
Ta H (µ-ArNSiHPh)
110.9(3)
46.8(2)
91.1(3)
(
hydride positions. Similar to the triplet state structure,
a complex with one bridging hydride and one hydride
that can be considered to be terminal or highly asym-
metrically bridging is found. Compared to the triplet
state structure, the singlet state structure possesses a
slightly longer Ta-Ta bond of 3.14 Å and a longer
Ta(1)-H(72) distance of 3.14 Å. (Thus, the hydride in
the singlet state structure can be considered more
terminal in nature.) Interestingly, the total energy of
the singlet complex was only 0.02 kcal/mol higher than
that of the triplet complex. The total unpaired spin
density, η as given by η ) ∫|F (r ) - F (r )|dr , was found
to be 2.03 e, indicating the presence of an antiferro-
magnetically coupled state. Considering the very small
energy difference between the two states, it is possible
that the experiment refers to a mixed state (multi-
determinantal), which could give rise to the observed
intermediate value of the magnetic moment.
Ta(1)-Ta(2)-Si(1)
Ta(1)-Ta(2)-Si(2)
Ta(1)-Ta(2)-N(2)
58.64(6) Ta(1)-N(1)-C(7)
56.18(5) Ta(2)-N(2)-C(35)
98.7(2)
150.7(6)
148.8(5)
a
Cp*cent represents the average of the x, y, and z coordinates of
5
the η -C5Me5 ring carbons.
the butterfly wings or terminally bound at each tanta-
lum atom. Surprisingly, however, the spectroscopic data
are not consistent with such a symmetric structure.
1
The H NMR spectrum of 2 contains individual, well-
separated sets of resonances for the two halves of the
molecule. Most striking is the large separation in shift
for the two hydride resonances, which appear at 11.63
ppm (doublet of doublets) and -0.99 ppm (triplet) and
exhibit an H-H coupling constant of 2.3 Hz. Resonances
for the inequivalent Si-H hydrogens appear at 5.23 (as
a doublet of doublets) and 5.09 ppm (as a singlet), with
R
â
1
observed satellites that correspond to J SiH coupling
constants of 215 and 193 Hz, respectively. Interestingly,
only the low-field Si-H group is coupled to the hydride
ligands (J HH ) 5.5 and 2.3 Hz). Two sets of ligand
resonances are also observed in the C and Si NMR
spectra of 2, and the latter spectrum exhibits individual
resonances at -7.6 and 14.2 ppm, respectively.
From these calculations, we conclude that the para-
magnetic complex possesses a structure in which one
hydride is terminal in nature and the other is bridging.
1
3
29
Ch a r a cter iza tion of 2. The molecular structure of
2
was determined by a single-crystal X-ray diffraction
analysis (Figure 3), and selected bond lengths and
angles are listed in Table 2. The tantalum atoms are
The infrared spectrum of 2 contains several weak
-
1
2
bands between 1535 and 1792 cm , in the regions
expected for both terminal and bridging hydride ligands
in tantalum complexes, and the Si-H stretching bands
linked by two bridging µ,η -ArNSiHPh ligands, which
give the molecule a butterfly conformation such that the
two “wings” (defined by the planar Ta2NSi rings)
intersect at a dihedral angle of 67.8°. The interatomic
Ta-Ta distance of 2.8472(5) Å is significantly shorter
than the corresponding distance in 1 (by 0.17 Å). As can
be seen in Figure 3, the molecule possesses approximate
-
1
for 2 appear at 2145 and 2181 cm
.
Given the X-ray structural analysis of 2, the spectro-
scopically observed asymmetry in the molecule must be
generated from the two hydride ligands. However, the
data presented above does not allow determination of
the hydride ligand positions. A potentially complicating
factor in the interpretation of spectroscopic data would
be the presence of isomers in the sample. For example,
the asymmetric silicon centers in 2 suggest the possible
existence of diastereomers. To help address these ques-
(noncrystallographic) 2-fold symmetry. As in 1, the
silicon atoms are pentacoordinate and roughly evenly
positioned between the two tantalum atoms (Ta-Si
distances range from 2.61 to 2.70 Å). In general, the
distances and angles associated with the silanimine
ligation are very similar to the corresponding values
observed for 1. Probably due to peripheral noise associ-
ated with disordered hexane solvent of crystallization
in the structure, the hydride ligands could not be
located. Given the observed molecular symmetry, it
would seem likely that they adopt symmetrical positions
in the complex, either as bridging ligands underneath
1
tions, we recorded the H NMR spectrum of a single
crystal of 2. This spectrum was identical to those of
other samples of 2, indicating that 2 forms as one isomer
with a high degree of diastereoselectivity.
With the aid of 2D TOCSY, NOESY, and 1H, 13
HMQC spectra, all proton and carbon resonances for 2
C

Ditantalum Hydride Complexes
Organometallics, Vol. 19, No. 19, 2000 3835
6
F igu r e 4. (a) Sections of the 2D NOESY spectra (500 MHz, benzene-d , room temperature, mixing time 0.8 s) of 2. NOE
signals to both hydrides are observed for the lower-field Si-H hydrogen only. (b) Representation of the structure of 2,
illustrating the labeling scheme used to indicate various hydrogens characterized by NOE experiments.
have been assigned. Furthermore, careful evaluation of
the NOE signals allowed a determination of the ap-
proximate location of the hydride ligands in the mol-
ecule. Both hydrides appear to be close to the Si-H
hydrogen resonating at 5.23 ppm (to which they also
couple), but no NOE signals to the other Si-H reso-
nance were observed (Figure 4). It would appear that
the low-field hydride (δ ) 11.63 ppm) is in a terminal
scan of the relevant configurational space, we performed
elevated temperature ab initio molecular dynamics
1
9
simulations. With the aim of performing as much as
possible an unbiased search, both hydride and dihydro-
gen complexes have been examined. This includes sym-
metric structures that are not consistent with the two
distinct TaH chemical shifts observed experimentally.
To more efficiently evaluate the relative energies of
selected conformers, we have constructed a model
system, Cp2Ta2H2(µ-HNSiH2)2 (3), whereby the aryl and
phenyl groups of the full system have been replaced with
1
position and bonded to Ta (Figure 4), oriented away
2
i
1
from Ta and toward the upper Pr group (as evidenced
1
i
1
by a strong NOE to Si -H and Pr -Hup and weak
NOEs to o-Ph -H, Cp* , and Pr -Meup,out). The high-
field hydride (δ ) -0.99 ppm) is located in a region
between the two Cp* ligands (weak NOEs to Cp* , Cp* ,
Si -H, and Pr -Hdown), possibly bridging the Ta atoms
Figure 4). Complex 2 is stereochemically rigid over the
temperature range of -60 to 95 °C, as indicated by H
NMR spectra (in toluene-d8).
1
1
i
1
*
hydrogen atoms and the Cp ligands have been replaced
with Cp ligands. Using the results from the model
system as a guide, we then have examined the actual
1
2
1
i
2
i
complex whereby the Cp* and (2,6- Pr2C6H3)NSiHPh
(
ligands are included.
1
All candidate structures represented in Figure 5
have been examined with the model system. Structures
in which the two hydride ligands are positioned ap-
proximately endo to the two Cp* rings are shown in
Figure 5a, those in which the hydride ligands are both
exo to the Cp* rings are displayed in Figure 5b, and
candidate structures in which one hydride is positioned
endo while the other is positioned exo to the Cp* rings
are depicted in Figure 5c. In the following discussion
we have adopted an alphanumeric labeling scheme
whereby the numeric component refers to either the
actual system, 2, or the model system, 3, and the
alphabetic component refers to the conformation of the
hydride ligands as displayed in Figure 5.
Thus, the experimental data for 2 are consistent with
a structure involving asymmetric substitution of the
hydride ligands in the otherwise C2-symmetric molec-
ular fragment Cp*2Ta2(µ-ArNSiHPh)2. Given this sur-
prising result and the uncertainty in the hydride
positions of 2, the theoretical study presented below was
undertaken.
Th eor et ica l Ch a r a ct er iza t ion of 2: Gen er a l
Con sid er a tion s. To determine the most probable posi-
tion of the tantalum-bound hydrogens in the diamag-
netic complex 2, the geometries of various candidate
isomers were optimized at the density functional level,
and their relative energies were compared. As a further
(19) Parrinello, M. Solid State Commun. 1997, 102, 107-120.

3
836 Organometallics, Vol. 19, No. 19, 2000
Burckhardt et al.
F igu r e 5. Sketches of candidate isomers of the diamagnetic complex 2 that have been examined by density function
theory calculations. (a) Isomers in which both hydrogen atoms are positioned endo to the Cp rings. (b) Isomers that have
been investigated in which both hydrogen atoms are positioned exo to the Cp rings. (c) Isomers considered in which one
hydrogen atom is positioned exo to the two Cp rings. For each isomer, two orientations are shown and pendant groups
have been removed for clarity.
position of the cyclopentadienyl rings. Compared to the
Ta ble 3. Rela tive DF T-BLYP En er gies for
Ca n d id a te Com p lexes in th e Mod el (3) a n d F u ll (2)
System s
X-ray structure, the Ta-silanimine ring system is more
puckered in the optimized geometry of the model
structure. This is evidenced by the contracted Ta-Ta
bond distance, the diminished N-Ta-Si angles, and the
enlarged centroid-Ta-Ta angles. The most severe
distortion is the N-Ta-Si angle, which is calculated to
be 96° in the model system but is 115° in the experi-
mentally determined structure.
b
relative energies (kcal/mol)
is optimized
model system
full system
structure consistent
with NMR data?
structurea
(3)
(2)
a
0.0
converges to 3a not investigated
+7.9 +3.2
converges to 3f not investigated
0.0
no
N/A
no
N/A
possiblec
N/A
yesd
N/A
yes
b-e
f
g
h
i
j
k
l
m
n
Of the isomers with hydride ligands exo to the
cyclopentadienyl rings (Figure 5b), only isomers 3f and
3j, which lie 7.9 and 5.8 kcal/mol above 3a , respectively,
are regarded as energetically significant. Geometry
optimization of isomer 3f resulted in two terminal
hydrides, which are homotopic. In isomer 3j, the two
terminal hydrides are bound to the same Ta center and
are not equivalent. Thus, if no rapid exchange occurs
between the hydrides, two distinct Ta-H NMR shifts
would be observed. Shown in Figure 6b are selected
geometric parameters for the optimized structure of the
model system, 3j. Despite the asymmetric hydride
binding in 3j, the Ta2N2Si2 six-membered ring also
remains nearly symmetric. Also notable is that the ring
puckering observed in 3j is less severe than in 3a .
+45
ejected
+5.8
not investigated
not investigated
+15
H
2
converges to 3f not investigated
+7.3 -0.7
converges to 3l converges to 2l
+34.7 not investigated
N/A
yes
a
See Figure 5. b DFT BLYP energies relative to isomer a . c The
hydrogens in this dihydrogen complex are inequivalent, but rapid
rotation is likely to make them equivalent on the NMR time scale.
d
Inequivalent terminal hydrides are on the same Ta center.
Ca lcu la tion s on th e Mod el System . With the
model system, optimization of all structures shown in
Figure 5a lead to generation of the bis-µ-dihydride
compound 3a . As it turns out, 3a is the lowest energy
structure that we have located on the potential energy
surface of the truncated model system. Thus, in Table
Of the structures in Figure 5c only 3l, which lies 7.3
kcal/mol above 3a , is considered energetically relevant.
Structure 3l is interesting because it contains one
terminal hydride and one bridging hydride and is
therefore consistent with the observed large difference
in Ta-H NMR shifts (11.63 and -0.99 ppm). The
optimized structure of 3l is displayed in Figure 6c. The
asymmetric bonding of the two hydrides in 3l induces
a noteworthy asymmetry in the Ta-silanimine ring
system, which is highlighted by a 0.1 Å difference in
the length of the two Ta-Si bonds.
3
we report the energies of all other optimized structures
relative to it. We note that the two Ta-bound hydrogens
in 3a are homotopic, and therefore only one NMR
chemical shift would be observed for this isomer. In
Table 3 we also indicate whether the optimized struc-
ture is consistent with the experimental observation of
two distinct Ta-H NMR shifts.
Characteristic geometrical parameters of the opti-
mized structure 3a , compared with that of the X-ray
structure 2, are displayed in Figure 6a. The geometry
of the Ta-silanimine six-membered-ring system is
reasonably reproduced. The most notable geometric
distortions involve the puckering of the ring and the
Ab initio molecular dynamics simulations of the model
complexes 3a , 3f, and 3l have been performed at
elevated temperatures in order to scan the potential

Ditantalum Hydride Complexes
Organometallics, Vol. 19, No. 19, 2000 3837
F igu r e 6. Comparison of selected geometric parameters from the X-ray structure of 2 and the calculated geometries for
the model system (3) and the actual system (2). The comparison is made for isomers a , j, and l in parts a, b, and c,
respectively. The right- and left-hand structures are two different views of the same complex. Some pendant groups have
been removed for clarity. Plain text values are of the calculated model system, underlined values correspond to the calculated
system with the full ligand set, and the bracketed values are of the experimental X-ray structure. For isomer a values
2
displayed for the experimental X-ray structure are averages of equivalent parameters, assuming C symmetry.
surface for additional structures that may have been
overlooked with the local optimizations. Three 1.5 ps
simulations at 600 K initiated from the optimized
structures 3a , 3f, and 3l, respectively, revealed no new
isomers. Starting from isomer 3a , an extended simula-
tion totaling 4.0 ps at 900 K also revealed no new
isomers. Within these relatively short time scales, the
molecular dynamics simulations suggest that the struc-
tures are stable and represent well-defined complexes
even at elevated temperatures.
Ca lcu la tion s In volvin g th e Actu a l Liga n d Sys-
tem . We have examined the potential surface of the full
diamagnetic system (2), with no approximations being
made for the pendant ligands. The four lowest energy
conformers of the model system, a , f, j, and l (see Figure
Instead, the asymmetric isomer 2l is slightly lower in
energy on the BLYP-DFT potential energy surface. Also
noteworthy is that isomer j, which was found to be the
second most stable isomer with the model ligand system,
lies 15 kcal/mol above isomer a with the actual ligand
system. Compared to the model system, the geometries
obtained with the real ligands tend to be in much better
agreement with the X-ray structure. Most significantly,
the severe puckering of the Ta2N2Si2 rings in model
isomers a and l (as evidenced by the contracted N-Ta-
Si and enlarged centroid-Ta-Ta angles) is not observed
in the full system. Whereas these angles differ by as
much as 19° compared with the X-ray structure with
the model system, they differ by no more than 4° with
the full system. These results demonstrate that the
BLYP plane wave computational scheme that was
utilized is capable of reproducing the experimental
geometries accurately. Moreover, the geometric distor-
tions observed with the model complexes are due to the
truncation of the ligands and not to the computational
scheme adopted.
5
), were fully optimized, and a summary of the relative
energies of the isomers is displayed in the third column
of Table 3. The resulting geometric parameters of 2a ,
2
j, and 2l are also depicted in Figures 6a, 6b and 6c,
respectively, and compared with the model system and
the X-ray structure.
With the authentic ligand set, the symmetric bis-µ-
hydride isomer, a , is no longer the most stable complex.
Although complex 2l is calculated to be energetically
most stable, it lies less than 1 kcal/mol lower than

3
838 Organometallics, Vol. 19, No. 19, 2000
Burckhardt et al.
Ta ble 4. Best F it r m s Devia tion betw een th e X-r a y
Cr ysta l Str u ctu r e a n d th e Ca lcu la ted Str u ctu r es
isomer
best fit rms deviation (Å)a
2
2
2
2
a
f
j
0.08
0.16
0.17
0.09
l
a
Best mass weighted fit rms deviation between the position of
the heavy atoms in the X-ray structure and the calculated
structures in Cartesian coordinates. The positions of the hydrogens
were not included in the fitting procedure. Equally weighted values
agree within 0.01 Å to the mass weighted values displayed here.
F igu r e 7. Best fit superposition of the X-ray structure and
the calculated gas-phase structure of isomer 2l. Hydrogens
are not shown and were not included in the fit.
structure 2a . This energy difference is within the error
limits of the methodology employed. (Arguably the
energy difference of 4 kcal/mol between 2f and 2l is
also within this range.) Another density functional
Ta ble 5. Com p a r ison of Exp er im en ta l NOE Sign a l
In ten sities to Bon d Dista n ces in th e Ca lcu la ted
Str u ctu r e of 2l
such as B3LYP2
0-22
or BP86
23-25
may yield slightly
different orderings of the isomers examined. Rather
than reexamine the structures with different theoretical
methodologies, we turn our attention to reconciling the
present results with the experimental data.
Com p a r ison w ith Exp er im en ta l Da ta . In this
section we compare the four candidate structures 2a ,
calculated
H-H
distance (Å)
atom paira
intensityb
TaH (bridging, δ ) -0.99 ppm) to:
TaH(terminal)
1
1
3/3
3.22
3.18
1
Si -H
both Cp* methyl protons
minimum )
2
f, 2j, and 2l with the spectroscopic data. Of the four
2
.60/3.02
isomers that we have examined with the authentic
ligand framework, only two possess hydrides in distinct
environments, namely, 2j and 2l. Isomer 2j contains two
terminal hydrides bound to the same Ta center, whereas
isomer 2l contains one bridging and one terminal
hydride. Even without considering the calculated rela-
tive energies of the two isomers (which reveal that
isomer 2l is most energetically favorable), isomer 2l is
the most promising candidate for two reasons. First, in
complex 2j, the two hydrides are both terminally bound
to the same Ta center and might be expected to
exchange on the NMR time scale. Second in 2j, both
hydrides are terminal in nature, whereas 2l possesses
hydrides in two distinct bonding schemes, one terminal
and one bridging. This mixed terminal and bridging
bonding scheme is more consistent with the two rather
distinct 1H NMR shifts observed at 11.63 and -0.99
ppm.
i
2
Pr -Hdown
2
3.18
TaH (terminal,δ ) -11.63 ppm) to:
TaH(bridging)
1
4
3
3
5
2
3.22
2.69
2.72
2.68
2.41
3.88
1
Si -H
1
Cp* methyl-H
1
o-Ph -H
iPr1-H
up
i
1
Pr -Hup,out
a
Refer to Figure 4b. b The signal intensity was assigned a value
of 1 through 5, where 1 represents a weak signal corresponding
to an interatomic distance of around 3 Å or sightly above and 5
represents a strong signal corresponding to an interatomic distance
of 2-2.4 Å.
in Figure 7, which depicts the best fit superposition of
the X-ray structure and the calculated structure 2l. The
Ta-Ta bonding distance, which has been determined
to be 2.85 Å from the X-ray analysis, is calculated to be
2
.88 Å for 2l. This is in better agreement than the
calculated 2.78 Å Ta-Ta bond distance for the sym-
metric isomer 2a . As previously stated, the X-ray
structure revealed a symmetrical Ta2N2Si2 ring with
approximate C2-symmetry. However, due to the asym-
metric binding of the two hydrides in 2l, the symmetry
of this ring is necessarily broken. Although the ring
asymmetry is not severe, the two Ta-Si bonds differ
by 0.08 Å and the two Si-N bonds differ by 0.1 Å.
It is of interest to compare the NOE data of Figure 4
with results from the theoretical calculations. Table 5
compares the intensity of the NOE signals to the
calculated H-H distances in the best candidate (2l). In
this table we have assigned the low-field hydride (δ )
The calculated geometries of the four candidate
isomers have been quantitatively compared with the
experimental X-ray structure. The mass-weighted best
fit rms deviation between Cartesian positions of the
heavy atoms in the X-ray structure and each of the
calculated structures is presented in Table 4. The
symmetric isomer 2a provides the best-fit geometry with
an rms deviation of 0.08 Å. The asymmetric bridging/
terminal structure 2l provides a fit that is almost as
good with an rms deviation of 0.09 Å. We note that the
candidate isomers in which both hydrogens are bound
exo to the cyclopentadienyl rings, 2f and 2j, have struc-
tural rms deviations roughly twice that of 2a and 2l.
Since 2l is our best candidate, both in terms of relative
energy and its agreement with the NMR data, we will
compare its structure with the experimental structure
in more detail. The agreement is striking, as illustrated
1
1.63 ppm) to the terminal position and the high-field
hydride (δ ) -0.99 ppm) to be bridging. Further, the
signal intensities have been assigned values of 1 through
5
, where 1 represents a weak signal corresponding to
an interatomic distance of around 3 Å or slightly above
and 5 represents a strong signal intensity corresponding
to an interatomic distance of <2.4 Å. Table 5 reveals
that the calculated H-H distances indeed fit the
experimental NOE data very well. The only exception
(
(
(
(
(
(
20) Becke, A. D. Chem. Phys. 1996, 104, 1040.
21) Becke, A. D. Chem. Phys. 1993, 98, 5648.
22) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785.
23) Becke, A. Phys. Rev. A 1988, 38, 3098.
24) Perdew, J . P. Phys. Rev. B 1986, 33, 8822-8824.
25) Perdew, J . P. Phys. Rev. B 1986, 34, 7406.
i
1
is that the coupling of protons on Pr -Meup,out (Figure

Ditantalum Hydride Complexes
Organometallics, Vol. 19, No. 19, 2000 3839
Ta ble 6. Selected Ma yer Bon d Or d er s2 of th e
5
a
Ca lcu la ted Isom er s
model system
full system
bondb
3a
3l
2a
2l
2
1
Ta -Ta
0.44
0.33
0.30
0.33
0.30
0.80
0.80
0.39
0.35
0.24
0.37
0.37
0.81
0.87
0.41
0.41
0.24
0.41
0.24
0.84
0.84
0.38
0.43
0.20
0.42
0.30
0.84
0.88
2
2
Ta -Si
2
1
Ta -Si
1
1
2
Ta -Si
1
Ta -Si
2
2
Ta -N
1
1
Ta -N
a
Bond orders calculated from Kohn-Sham wave functions in
the plane wave basis that are projected on an atom-centered basis.
Labeling scheme is defined in Figure 4b.
b
i
will examine the bonding in the symmetric isomer a and
the asymmetric isomer l. Although we have concluded
that the asymmetric isomer l is the best candidate, the
two isomers are close in energy and the symmetry of
isomer a simplifies the bonding analysis. Furthermore,
variations in the bonding that arise from changes in
ligand substitutions and hydride ligand positions are
of interest.
Our analysis suggests that the bond between the Ta
centers can be characterized as a weak σ-type bond. The
calculated Ta-Ta bond orders for the complexes 2a , 2l,
3a , and 3l are all approximately 0.4. This compares to
calculated values of approximately 0.8 for the Ta-N
bonds, 0.3-0.4 for the Ta-Si bonds, 0.7 for the Si-N
bonds, and 1.0 for the N-C(aryl) bonds. On the basis
of this qualitative measure, we conclude that there is a
weak Ta-Ta covalent bond. What can be characterized
as the Ta-Ta bonding orbital is the HOMO for each of
these complexes (see Supporting Information).
F igu r e 8. H-H distances of terminal hydride with Pr-
Me protons during a 400 K ab initio molecular dynamics
simulation of 2l.
4
b) with the terminal hydride of the calculated structure
yields a distance of 3.88 Å. Such an interatomic distance
is too large to provide the signal observed. Possible
explanations for the discrepancy include solvation ef-
fects (for which these calculations do not account) and
finite temperature effects. To investigate the latter
possibility, we performed a short (∼1 ps) ab initio
molecular dynamics simulation of the full complex (2l)
at 400 K. Displayed in Figure 8 is the interatomic
distance between the terminal hydride to each of the
three protons of the Pr -Meup,out group. Figure 8 shows
that during the course of the dynamics simulation, the
H-H distance does fall within the 3 Å regime necessary
to give rise to the observed NOE signal. This suggests
that finite temperature effects may account for the
weak NOE signal assigned to the terminal hydride-
i
1
We now turn our attention to the Ta-Si bonding.
Each silicon atom is bonded to both tantalum atoms,
with nearly equivalent Ta-Si bond distances (2.61-2.79
Å). There are actually two types of Ta-Si bonds: one
that can be considered as part of the Ta2NSi ring system
and one that “crosses” the ring. The “ring-crossing” Ta-
Si bonds are depicted as the dashed bonds in Figure 4b.
With the labeling scheme of Figure 4b, the ring-crossing
i
1
Pr -Meup,out group interaction.
Of the lowest energy candidate isomers, 2a , 2f, 2j,
and 2l, only isomers 2l and 2j have nonequivalent
hydrides. However, the calculated structure for candi-
date 2j violates the NOE assignments in several key
H-H distances. First and foremost, the two hydrides
in structure 2j are calculated to be only 1.8 Å apart,
which is not in agreement with the weak NOE signal
for this pair. Additionally, as previously stated, the high-
field hydride appears to be in the area between the two
Cp* ligands. In candidate 2j, neither hydride is located
between the Cp* ligands.
To summarize, the present calculations suggest that
isomer 2l, in which there is a terminal and bridging
hydride, is the best candidate in terms of its relative
energy and agreement with the available spectroscopic
data. It is also the only isomer considered that is
consistent with the NOE data, and its calculated
geometry is in excellent agreement with the X-ray
structure.
Th eor et ica l An a lysis of t h e Ta -Ta a n d Ta -Si
Bon d in g. In this section the bonding within the
Ta2NSi ring is examined. In particular we describe the
Ta-Ta bonding and the Ta-Si bonding from the
standpoint of molecular orbital theory based on the
calculated Kohn-Sham DFT wave functions. A quali-
tative measure of the strength of the covalent bond
between two atoms can be provided by the Mayer
bond order.1 These are displayed in Table 6 for isomers
a and l for both the model and actual systems. We
1
2
2
1
bonds involve the Ta ,Si and Ta ,Si atom pairs.
The calculated Mayer bond orders given in Table 6
show that there exists a weak ring-crossing Ta-Si bond.
For the symmetric model complex 3a , the calculated
ring and ring-crossing Ta-Si bond orders are nearly
equivalent at 0.33 and 0.30, respectively. This is also
true for the asymmetric model complex 3l, albeit with
somewhat more variance. For the “full” complexes with
i
Cp* and (2,6- Pr2C6H3)NSiHPh ligands there is more
2
2
2
1
discrepancy between the Ta -Si and Ta -Si bond
orders, with the ring-crossing Ta-Si bond being much
2
2
weaker. For example, in complex 2a the Ta -Si and
2
1
Ta -Si bond orders are 0.41 and 0.24, respectively.
Interestingly, the Ta-Si ring-crossing bond distance is
longer in the model complex than in the actual complex,
which is opposite of what might be expected based on
the calculated bond orders. This suggests that the short
ring-crossing Ta-Si distance results from geometrical
constraints of the ring system.
Con clu sion s
8
As expected, the Ta-H and Ta-Si bonds in Cp*(Ar-
Nd)Ta[Si(SiMe3)3]H are active toward σ-bond metath-

3
840 Organometallics, Vol. 19, No. 19, 2000
Burckhardt et al.
esis, as evidenced by the reaction with phenylsilane.
Currently, however, the mechanism by which these
bonds are cleaved in this reaction is unknown. Given
the complexity of this reaction, we intend to address
this issue in related tantalum compounds possessing
only one active σ-bond (e.g., Cp*(ArNd)Ta(R)Cl; R ) H,
benzophenone ketyl. Benzene-d
from Na/K alloy. PhSiH was dried over molecular sieves
4 Å) and distilled before use. The compound Cp*(ArNd)Ta-
6 8
and toluene-d were distilled
3
(
[
8
,9
3 3
Si(SiMe ) ]H was prepared according to a known procedure.
Elemental analyses were performed by the Microanalytical
Laboratory in the College of Chemistry at the University of
+
California, Berkeley. Mass spectra (EI ) were recorded on a
9
Si(SiMe3)3 ).
Micromass VG ProSpec instrument (ionization energy: 70 eV).
FT-infrared spectra were recorded as KBr pellets (solids) on
Of perhaps greater interest is the observed coupling
of a silane with an imido ligand, to produce a Si-N
bond. This coupling, and the accompanying cleavage of
two Si-H bonds, ultimately results in the formation of
a silanimine as a bridging ligand. The structural and
theoretical data indicate that this ligand possesses very
little (if any) Si-N double bond character and is
therefore best viewed as a bifunctional silyl-amido
ligand. Various mechanisms can be envisioned for this
multistep process. One possible scheme involves σ-bond
metathesis and/or oxidative addition-reductive elimi-
nation to introduce a -SiH2Ph ligand. This silyl ligand
could migrate to the imido nitrogen atom, to produce
an amido ligand of the type -N(Ar)SiH2Ph. To our
knowledge, this type of migration has not been observed,
but it seems quite conceivable, especially given obser-
1
2
13
29
a Mattson FTIR 3000 instrument. H, H, C, and Si NMR
spectra were recorded on Bruker AMX-300, AMX-400, or DRX-
500 instruments, and chemical shifts are reported in ppm
1
downfield from internal SiMe
4
. For H NOESY spectra, a
1
13
mixing time of 0.8 s was applied. 2D H, C HMQC spectra
were recorded as 2048 × 256 files using the BIRD sequence
1
13
28
29
[
π/2( H) ) 14.0 µs; π/2( C) ) 6.4 µs]. 1D Si NMR spectra
were obtained using a refocused INEPT pulse sequence with
or without 1H decoupling; delays were optimized for J SiH
)
.4 Hz [π/2( H) ) 14.0 µs; π/2( Si) ) 6.6 µs]. 2D H, 29Si
1
29
1
6
HMQ(B)C spectra were acquired as 2048 × 128 files with 16
transients accumulated per t increment and were optimized
1
for an assumed 4 Hz (125 ms) long-range J SiH. Bruker
XWINNMR software (ver. 2.1) was used for all processing.
X-r a y Cr ysta llogr a p h y. Single-crystal analyses of 1 and
2
were carried out by Dr. Fred J . Hollander and Dr. Dana L.
1
1
Caulder at the UC Berkeley CHEXRAY crystallographic
facility. Measurements were made on a Bruker SMART CCD
area detector with graphite-monochromated Mo KR radiation
vation of a corresponding hydrogen migration. A Ta-
N(Ar)SiH2Ph complex could then undergo an inter-
molecular Si-H activation to produce the bridging
silanimine ligand. Alternative schemes are based on the
direct attack of silane on the TadNAr linkage, resulting
in addition of the Si-H bond across the TadN double
bond to give an (H)Ta-N(Ar)SiH2Ph intermediate. This
addition could proceed in a stepwise manner, as we have
recently observed in the reaction of PhSiH3 with a bent
tantalum imido complex,26 or in a concerted way, as has
(λ ) 0.71069 Å). Data were integrated by the program SAINT,
corrected for Lorentz and polarization effects, and analyzed
for agreement and possible absorption using XPREP. Empiri-
cal absorption corrections were made using SADABS. Struc-
tures were solved by direct methods and expanded using
Fourier techniques. All calculations were performed using the
teXsan crystallographic software package.
Rea ction of Cp *(Ar Nd)Ta [Si(SiMe
3 3
) ]H w ith P h en yl-
sila n e. To a freshly prepared solution of Cp*(ArNd)Ta[Si-
been proposed for the activation of hydrocarbons by
9
group 4 imido complexes.27
(SiMe
and (THF)
mL) was added PhSiH
3
)
3
2
]H (from [Cp*(ArNd)TaCl(µ-H)] (400 mg, 0.38 mmol)
3
LiSi(SiMe
3
)
3
(360 mg, 0.76 mmol) in hexanes (30
(83 mg, 0.76 mmol). An instant color
As suggested by the spectroscopic data and confirmed
by the DFT calculations, the ditantalum dihydride
complex 2 appears to be a rare example of a dinuclear
metal complex possessing an asymmetrical distribution
of hydride ligands, in a complex that would otherwise
possess chemically equivalent metal centers (we are
aware of no other examples).
3
change from deep red to yellow was observed. The mixture
was stirred at room temperature for 12 h while the color of
the reaction solution gradually darkened to deep green. The
solution was filtered, concentrated to half its original volume,
and cooled to -30 °C. Three pure products were isolated from
the reaction mixture by fractional crystallization: [Cp*(ArNd)-
TaH(µ-H)] as a yellow microcrystalline precipitate (15 mg, 4%
2
based on Ta), paramagnetic 1 as yellow crystals (75 mg, 18%
based on Ta), and diamagnetic 2 as dark green crystals in two
The results from the DFT calculations on the entire
1
40-atom diamagnetic complex and the 127-atom para-
magnetic complex are a convincing demonstration of
the power of current state-of-the-art computational
methods in the structure elucidation of novel transition
metal complexes. Of particular note is the observation
that the truncated ligand system (3) was inadequate to
probe certain key aspects of the bonding in this system.
This is particularly interesting given that simplified
crops (322 mg, 71% based on Ta).
1
Da ta for 1. H NMR (500 MHz, benzene-d
6
, 73 °C): δ 10.5
(
9
br s, 1 Ta-H), 7.65 (d, 2 H, o-Ph-H, J ) 6.5), 7.16-6.82 (m,
Ar-H), 4.93 (s, 1 H, Si-H, J SiH ) 220 Hz), 3.98, 3.60 (2 × br
i
i
s, 2 Pr-H), 3.41 (br s, 2 Pr-H), 2.14, 1.95 (2 × s, 2 × 15 H,
i
C
5
Me
5
), 1.48, 1.38, 1.24, 1.14 (4 × d, 4 × 3 H, Pr-Me, J )
i
13
6.8), 1.3-0.8 (br s, 12 H, Pr-Me), 0.2 (br s, 1 Ta-H). C NMR
(101 MHz, benzene-d
, shift values from 1H, C HMQC spec-
13
“surrogate” ligands are often used in theoretical inves-
6
tigations of inorganic and organometallic complexes.
trum): δ 137.0 (o-Ph H), 124.0, 123.7, 122.3, 121.6 (Ar H),
i
29
2
7.5-23.0 ( Pr), 12.6 (C
5
Me
5 6
). Si NMR (99 MHz, benzene-d ,
1
29
Exp er im en ta l Section
shift value from H, Si HMQC spectrum): δ 9.7 (br). IR (KBr,
-
1
cm ): 3051 w, 2958 s, 2910 m, 2866 m, 2173 br w, 1763 br
m, 1653 br m, 1587 w, 1458 m, 1425 s, 1379 m, 1340 s, 1317
m, 1290 m, 1246 m, 1200 m, 1109 m, 1028 w, 966 w, 933 m,
Gen er a l Com m en ts. All experiments were performed
under dry nitrogen using standard Schlenk or drybox tech-
niques. Solvents were distilled under nitrogen from sodium
8
29 sh, 773 m, 798 m, 754 m, 739 m, 722 w, 702 m, 474 w.
+
MS: m/z 1090 (95, [M] ), 1086 (100), 1041 (19), 980 (28), 935
(
26) Gountchev, T. I.; Tilley, T. D. J . Am. Chem. Soc. 1997, 119,
2831.
27) (a) Cummins, C. C.; Baxter, S. M.; Wolczanski, P. T. J . Am.
(
µ
10), 591 (13), 552 (15), 541 (16), 489 (23), 466 (12), 444 (12).
1
eff ) 2.0 (Evans’ method).17 Anal. Calcd for C50
H
N
SiTa
(
72
2
2
Chem. Soc. 1988, 110, 8731. (b) Cummins, C. C.; Schaller, C. P.; Van
Duyne, G. D.; Wolczanski, P. T.; Chan. A. W. E.; Hoffmann, R. J . Am.
Chem. Soc. 1991, 113, 2985. (c) Walsh, P. J .; Hollander, F. J .; Bergman,
R. G. J . Am. Chem. Soc. 1988, 110, 8729. (d) Cundari, T. R. J . Am.
Chem. Soc. 1992, 114, 10557.
(1091.12): C, 55.04; H, 6.65; N, 2.57. Found: C, 54.75; H, 6.42;
N, 2.51.
Str u ctu r a l Deter m in a tion of 1. Suitable crystals were
grown from pentane solution at -30 °C. A fragment (0.25 ×

Ditantalum Hydride Complexes
Organometallics, Vol. 19, No. 19, 2000 3841
0
.23 × 0.15 mm) was cut from a yellow platelike crystal,
The valence orbitals are expanded in a basis of plane waves
with a kinetic energy cutoff of 70 Ry. A face-centered cubic
(fcc) super cell of edge 18 and 22 Å was utilized for the model
diamagnetic complex, 3, and the full diamagnetic complex, 2,
respectively. An fcc cell of length 23 Å was used for the
calculations of the paramagnetic complex, 1. For C, N, and Si
only the valence electrons are treated explicitly. The nonlocal,
mounted on a quartz fiber using Paratone N hydrocarbon oil,
and cooled under a nitrogen stream in the diffractometer. Non-
hydrogen atoms were refined anisotropically. The silicon
hydrogen (H70) and the hydrides (H71, H72) were located in
the difference Fourier map and refined with fixed thermal
parameters. All other hydrogen atoms were included in
calculated idealized positions but not refined. The final refine-
3
3
norm-conserving pseudopotentials of Martins and Trouiller
have been applied to represent the effects of the ionic cores.
The following radial cutoffs were used: r ) r ) 1.23 au for
carbon, r ) r ) 1.12 au for nitrogen, and r ) 1.9; r ) r )
ment cycle converged at R ) 0.025 and R
w
) 0.028.
6
1
Da ta for 2. H NMR (500 MHz, benzene-d , 25 °C): δ 11.63
s
p
1
1
(dd, 1 H, Ta -Hterminal, J ) 5.5, 2.3), 7.63 (m, 2 H, o-Ph H),
s
p
s
p
d
2
1
7
1
.55 (m, 2 H, o-Ph -H), 7.22 (dd, 1 H, m-Ar -Hdown, J ) 7.6,
2.1 au for silicon. For tantalum we constructed a semi-core
pseudopotential such that the 5s, 5p, 5d, and 6s electrons were
treated explicitly. Using this pseudopotential with a kinetic
energy cutoff of 70 Ry, the calculated bond distance of 1.677
Å in TaO compares well with the experimental gas-phase
2
.8), 7.19 (dd, 1 H, m-Ar -Hdown, J ) 7.6, 1.8), 7.12 (m, 1 H,
1
2
2
p-Ar -H), 7.10 (m, 2 H, m-Ph -H), 7.09 (m, 1 H, p-Ph -H),
7
1
2
.08 (m, 1 H, m-Ar -Hup), 7.05 (m, 1 H, p-Ar -H), 7.00 (m, 1
1 2 1
H, p-Ph -H), 6.99 (m, 1 H, m-Ar -Hup), 6.98 (m, 2 H, m-Ph -
1
1
34
H), 5.23 (dd, 1 H, Si -H, J ) 5.5, 2.9, J Si,H ) 215), 5.09 (s, 1
electron diffraction value of 1.6875 Å. We have examined the
2
1
i
2
H, Si -H, J Si,H ) 193), 3.59 (sept., 1 H, Pr -Hdown, J ) 6.7),
convergence properties of the Ta pseudopotential by repeating
the TaO calculations at 60, 80, 90, and 120 Ry. Both the total
energy and Ta-O bond distance show smooth and steady
convergence. At 70 Ry, the TaO bond distance is only 0.004 Å
larger than the distance obtained with a high energy cutoff of
120 Ry. A spin-restricted formalism was utilized for all
calculations of the diamagnetic complexes, whereas the spin-
unrestricted formalism has been utilized for all calculations
of the paramagnetic species, 1. Ab initio molecular dynamics
simulations were performed with a time step of 5 au and
fictitious mass for the electronic degrees of freedom of µ ) 500
i
2
i
1
3
.10 (sept., 1 H, Pr -Hup, J ) 6.7), 3.06 (sept., 1 H, Pr -Hup,
i 1
J ) 6.7), 3.03 (sept., 1 H, Pr -Hdown, J ) 6.7), 2.01 (s, 15 H,
2
1
i
2
Cp* ), 1.91 (s, 15 H, Cp* ), 1.68 (d, 3 H, Pr -Medown,out, J )
i
2
i
1
6
.7), 1.67 (d, 3 H, Pr -Medown,in, J ) 6.7), 1.53 (d, 3 H, Pr -
i 1
Medown,in, J ) 6.7), 1.44 (d, 3 H, Pr -Medown,out, J ) 6.7), 1.13
i
2
i
1
(
6
d, 3 H, Pr -Meup,out, J ) 6.7), 0.90 (d, 3 H, Pr -Meup,out, J )
i
1
i
2
.7), 0.56 (d, 3 H, Pr -Meup,in, J ) 6.7), 0.24 (d, 3 H, Pr -
Meup,in, J ) 6.7), -0.99 (pseudo-t, 1 H, Ta-H-Ta, J ) 2.7).
1
3
1
C{ H} NMR (126 MHz, benzene-d
6
): δ 151.11, 150.44,
1
1
44.33, 141.74, 141.66, 140.84, 139.09, 138.53 (Ar/Ph C),
2
1
1
18
39.29 (o-Ph CH), 137.75 (o-Ph CH), 129.12 (m-Ar CHup, and
au. To calculate the Mayer bond orders, the one-electron
2
1
2
m-Ar CHup), 127.94 (m-Ph CH), 127.66 (m-Ph CH), 124.47
Kohn-Sham orbitals in the plane wave basis were projected
onto a minimal, atom-centered basis consisting of the atomic
pseudo wave functions.
2
2
1
(p-Ph CH), 124.28 (m-Ar CHdown, and p-Ar CH), 124.22 (m-
1
1
2
Ar CHdown), 124.10 (p-Ph CH), 123.76 (p-Ar CH), 113.69,
i
2
i
1
1
2
10.52 (C
5
1
Me
5
), 29.17 ( Pr Medown,in), 28.87 ( Pr Medown,in),
All calculations were performed on the IBM-SP (160 MHz
nodes) of the Competence Center for Computational Chemis-
i
i
2
i
2
8.65 ( Pr CHdown), 28.02 ( Pr CHdown), 27.82 ( Pr CHup), 27.52
Pr CHup), 27.02 ( Pr Meup,in), 25.94 and 25.92 ( Pr Meup,out
i
1
i
2
i
1
4
(
,
try, C , at the ETH in Z u¨ rich. For calculations on the
and Pr Medown,out), 25.57 and 25.54 ( Pr Meup,out, and Pr2
i
1
i
2
i
diamagnetic model system 3, which consists of 34 atoms and
102 electrons, each molecular dynamics time step required
about 22 s on 16 nodes. For the full diamagnetic complex 2,
which contains 140 atoms and 346 electrons, each time step
required about 77 s on the same number of nodes.
i
1
2
5 5 5
), 12.92 (C Me
1).
Medown,out), 25.17 ( Pr Meup,in), 13.96 (C
5
Me
2
9
1
Si{ H} NMR (99 MHz, benzene-d
6
): δ 14.2, -7.6. IR (KBr,
-
1
cm ): 3049 m, 2959 s, 2904 s, 2869 s, 2724 w, 2181 m, 2145
sh, 1950 vw, 1906 vw, 1792 w, 1648 vw, 1622 vw, 1588 w, 1535
vw, 1459 m, 1425 s, 1380 m, 1360 w, 1314 m, 1238 m, 1191
m, 1105 m, 1027 m, 918 m, 858 sh, 839 m, 806 m, 786 m, 756
m, 733 m, 700 m, 606 w, 549 w, 526 m, 474 w, 432 m. MS:
m/z 1196 (100, [M] ). UV-vis (THF): λmax. 392 nm. Anal. Calcd
for C56
Found: C, 55.95; H, 6.31; N, 2.44.
Ack n ow led gm en t. We thank the National Science
Foundation for their generous support of this work. Dr.
Fred J . Hollander and Dr. Dana L. Caulder are grate-
fully acknowledged for the determination of the crystal
structures, and Tomislav I. Gountchev is acknowledged
for recording the high-temperature NMR spectra of
compound 2. U.B. is indebted to the Schweiz. National-
fonds and the Novartis Stiftung for postdoctoral fellow-
ships. T.K.W. would like to thank the Natural Sciences
and Engineering Council of Canada for a postdoctoral
fellowship. U.R. acknowledges an ETH research grant
under No. TH 23/97-4 Fonds Nr. 0-28472-99.
+
78 2 2 2
H N Si Ta (1197.32): C, 56.18; H, 6.57; N, 2.34.
2
Str u ctu r a l Deter m in a tion of 2‚ /
were grown from hexane solution at -30 °C. A fragment (0.41
0.30 × 0.05 mm) was cut from a green blade-shaped crystal,
3 6
C H14. Suitable crystals
×
mounted on a quartz fiber using Paratone N hydrocarbon oil,
and cooled under a nitrogen stream in the diffractometer. Two
regions of highly disordered solvent molecules (hexane) were
modeled as three half-occupancy carbons and as 10 quarter-
occupancy carbons, respectively, giving a total of 2/3 of a
hexane molecule. Non-hydrogen atoms were refined anisotro-
pically, except for those that were part of the disordered
solvent region, which were refined isotropically. Hydrogen
atoms were included in calculated idealized positions but not
refined. The final refinement cycle converged at R ) 0.036 and
Su p p or tin g In for m a tion Ava ila ble: Tables of crystal,
data collection and refinement parameters, atomic coordinates,
bond distances, bond angles, and anisotropic displacement
2
parameters for compounds 1 and 2‚ /
3
C
6
H
14. A plot of the
R
w
) 0.046.
Com p u ta tion a l Deta ils. Kohn-Sham density functional
HOMO of isomer 2l. This material is available free of charge
2
9
via the Internet at http://pubs.acs.org.
theory calculations as embodied in the Car-Parrinello ab
31
3
0
initio molecular dynamics method and pseudopotential theory
have been performed with the program CPMD.32 The BLYP
2
3
OM000455X
density functional was utilized as developed by Becke for the
exchange component and by Lee, Yang, and Parr22 for the
correlation component.
(32) Hutter, J .; Ballone, P.; Bernasconi, M.; Focher, P.; Fois, E.;
Goedecker, S.; Parrinello, M.; Tuckerman, M. CPMD; Max-Planck-
Institut f u¨ r Festk o¨ rperforschung, Stuttgart, Germany, and IBM Z u¨ rich
Research Laboratory, 1998.
(33) Trouiller, N.; Martins, J . L. Phys. Rev. B 1991, 43, 1993.
(34) Lide, D. R. CRC Handbook of Chemistry and Physics, 76th ed.;
CRC Press: Boca Raton, FL, 1996.
(28) Bax, A.; Subramanian, S. J . Magn. Reson. 1986, 67, 565.
(29) Kohn, W.; Sham, L. J . Phys. Rev. 1965, A140, 1133.
(30) Car, R.; Parrinello, M. Phys. Rev. Lett. 1985, 55, 2471.
(31) Kleinman, L.; Bylander, D. M. Phys. Rev. Lett. 1982, 48, 1425.