Synthesis and decomposition of alkyl complexes of molybdenum(IV) that contain a [(Me3SiNCH2CH2)3N]3- ligand. Direct detection of α-elimination processes that are more than six orders of magnitude faster than β-elimination processes

Article abstract of DOI:10.1021/ja970697x

A variety of paramagnetic molybdenum complexes, [N₃N]MoR (N₃N^3- = [(Me₃SiNCH₂CH₂)₃N]^3-; R = Me, Et, Bu, Ch₂Ph, Ch₂SiMe₃, CH₂CMe₃, cyclopropyl, cyclobutyl, cyclopentyl, cyclohexyl, cyclopentenyl, phenyl), have been prepared from [N₃N]MoCl. The several that have been examined all follow Curie-Weiss S = 1 behavior with a magnetic moment in the solid state between 2.4 and 2.9 μ(B) down to 50 K. Below ~50 K the effective moments undergo a sharp decrease as a consequence of what are proposed to be a combination of spin-orbit coupling and zero field splitting effects. NMR spectra are temperature dependent as a consequence of 'locking' of the backbone into one C₃-symmetric conformation and as a consequence of Curie-Weiss behavior. The cyclopentyl and cyclohexyl complexes show another type of temperature-dependent fluxional behavior that can be ascribed to a rapid and reversible α-elimination process. For the cyclopentyl complex the rate constant for α-elimination is ~10³ s^-1 at room temperature, while the rate constant for α-elimination for the cyclohexyl complex is estimated to be ~200 s^-1 at room temperature. An isotope effect for α-elimination for the cyclohexyl complex was found to be ~3 at 337 K. Several of the alkyl complexes decompose between 50 and 120°C. Of the complexes that contain linear alkyls, only [N₃N]Mo(CH₂CMe₃) decomposes cleanly (but slowly by α,α-dehydrogenation to give [N₃N]Mo≡CCMe₃. [N₃N]MoMe is by far the most stable of the alkyl complexes; no [N₃N]Mo≡CH can be detected upon attempted thermolysis at 120°C. Other decompositions of linear alkyl complexes are complicated by competing reactions, including β-hydride elimination. β-Hydride elimination. β-Hydride elimination (to give [N₃N]MoH) is the sole mode of decomposition of the cyclopentyl and cyclohexyl complexes; the former decomposes at a rate calculated to be approximately 10x that of the latter at 298 K. β-Hydride elimination in [N₃N]Mo(cyclopentyl) to give (unobservable) [N₃N]Mo(cyclopentene)-(H) has been shown to be 6-7 orders of magnitude slower than α-hydride elimination to give (unobservable) [N₃N]-Mo(cyclopentylidene)(H). N₃N]Mo(cyclopropyl) evolves ethylene in a first-order process upon being heated to give [N₃N]Mo≡CH, while [N₃N]Mo(cyclobutyl) is converted into [N₃N]Mo≡CCH₂CH₂CH₃. [N₃N]MoH decomposes slowly and reversibly at 100°C to yield molecular hydrogen and [(Me₃SiNCH₂CH₂)₂NCH₂CH₂SiMe₂CH₂]Mo ([bitN₃N]Mo). X-ray structures of [N₃N]Mo(triflate), [N₃N]MoMe, [N₃N]MoMe,[N₃N]MoMe, [N₃N]Mo(cyclohexyl), and [bitN₃N]Mo show that the degree of twist of the TMS groups away from an 'upright' position correlates with the size of the ligand in the apical pocket and that steric congestion in the cyclohexyl complex is significantly greater than in the methyl complex. Relief of steric strain in the ground state in molecules of this general type to give a less crowded alkylidene hydride intermediate is proposed to be β-elimination in several circumstances.

Full text of DOI:10.1021/ja970697x

11876
J. Am. Chem. Soc. 1997, 119, 11876-11893
Synthesis and Decomposition of Alkyl Complexes of
Molybdenum(IV) That Contain a [(Me3SiNCH2CH2)3N]3-
Ligand. Direct Detection of R-Elimination Processes That Are
More than Six Orders of Magnitude Faster than â-Elimination
Processes
Richard R. Schrock,*,† Scott W. Seidel,† Nadia C. Mo1sch-Zanetti,† Keng-Yu Shih,†
Myra B. O’Donoghue,† William M. Davis,† and William M. Reiff‡
Contribution from the Departments of Chemistry, Massachusetts Institute of Technology,
Cambridge, Massachusetts 02139, and Northeastern UniVersity, Boston, Massashusetts 02115
ReceiVed March 4, 1997X
Abstract: A variety of paramagnetic molybdenum complexes, [N3N]MoR ([N3N]3- ) [(Me3SiNCH2CH2)3N]3-; R
) Me, Et, Bu, CH2Ph, CH2SiMe3, CH2CMe3, cyclopropyl, cyclobutyl, cyclopentyl, cyclohexyl, cyclopentenyl, phenyl),
have been prepared from [N3N]MoCl. The several that have been examined all follow Curie-Weiss S ) 1 behavior
with a magnetic moment in the solid state between 2.4 and 2.9 µΒ down to 50 K. Below ∼50 K the effective
moments undergo a sharp decrease as a consequence of what are proposed to be a combination of spin-orbit coupling
and zero field splitting effects. NMR spectra are temperature dependent as a consequence of “locking” of the backbone
into one C3-symmetric conformation and as a consequence of Curie-Weiss behavior. The cyclopentyl and cyclohexyl
complexes show another type of temperature-dependent fluxional behavior that can be ascribed to a rapid and reversible
R-elimination process. For the cyclopentyl complex the rate constant for R-elimination is ∼103 s-1 at room
temperature, while the rate constant for R-elimination for the cyclohexyl complex is estimated to be ∼200 s-1 at
room temperature. An isotope effect for R-elimination for the cyclohexyl complex was found to be ∼3 at 337 K.
Several of the alkyl complexes decompose between 50 and 120 °C. Of the complexes that contain linear alkyls,
only [N3N]Mo(CH2CMe3) decomposes cleanly (but slowly) by R,R-dehydrogenation to give [N3N]MotCCMe3.
[N3N]MoMe is by far the most stable of the alkyl complexes; no [N3N]MotCH can be detected upon attempted
thermolysis at 120 °C. Other decompositions of linear alkyl complexes are complicated by competing reactions,
including â-hydride elimination. â-Hydride elimination (to give [N3N]MoH) is the sole mode of decomposition of
the cyclopentyl and cyclohexyl complexes; the former decomposes at a rate calculated to be approximately 10× that
of the latter at 298 K. â-Hydride elimination in [N3N]Mo(cyclopentyl) to give (unobservable) [N3N]Mo(cyclopentene)-
(H) has been shown to be 6-7 orders of magnitude slower than R-hydride elimination to give (unobservable) [N3N]-
Mo(cyclopentylidene)(H). [N3N]Mo(cyclopropyl) evolves ethylene in a first-order process upon being heated to
give [N3N]MotCH, while [N3N]Mo(cyclobutyl) is converted into [N3N]MotCCH2CH2CH3. [N3N]MoH decomposes
slowly and reversibly at 100 °C to yield molecular hydrogen and [(Me3SiNCH2CH2)2NCH2CH2SiMe2CH2]Mo
([bitN3N]Mo). X-ray structures of [N3N]Mo(triflate), [N3N]MoMe, [N3N]Mo(cyclohexyl), and [bitN3N]Mo show
that the degree of twist of the TMS groups away from an “upright” position correlates with the size of the ligand in
the apical pocket and that steric congestion in the cyclohexyl complex is significantly greater than in the methyl
complex. Relief of steric strain in the ground state in molecules of this general type to give a less crowded alkylidene
hydride intermediate is proposed to be an important feature of the high rate of R-elimination relative to â-elimination
in several circumstances.
Introduction
transition metal complexes is [(C6F5NCH2CH2)3N]3- 1,5
. ) Tria-
midoamine ligands usually bind to a transition metal in a
tetradentate manner, thereby creating a sterically protected,
3-fold-symmetric “pocket” in which only three orbitals are
available to bond to additional ligands in that pocket, two
degenerate π orbitals (approximately dxz and dyz) and a σ orbital
Triamidoamine ligands, [(RNCH2CH2)3N],3- in which R is
a bulky substituent, can bind to a variety of transition metals1
(or main group elements2), especially in oxidation states 3+ or
higher. The synthesis of the pentane-soluble trilithium salt3 of
(Me3SiNHCH2CH2)3N4 allowed the chemistry of complexes
containing the [(Me3SiNCH2CH2)3N]3- ([N3N]3-) ligand to
develop relatively rapidly. (The only other ligand that has been
employed to prepare a significant number of triamidoamine
2
(approximately dz ). This frontier orbital picture contrasts
sharply with that for the ubiquitous bent metallocene in which
all three orbitals (two with A1 symmetry and one with B2
symmetry in the C2V point group) lie in the plane passing
between the two Cp rings.6 The fact that the two frontier π
orbitals in a C3-symmetric triamidoamine complex are strictly
degenerate and probably essentially pure d orbitals creates an
† Massachusetts Institute of Technology.
‡ Northeastern University.
X Abstract published in AdVance ACS Abstracts, November 15, 1997.
(1) Schrock, R. R. Acc. Chem. Res. 1997, 90, 99.
(2) Verkade, J. G. Acc. Chem. Res 1993, 26, 483.
(3) Cummins, C. C.; Schrock, R. R.; Davis, W. M. Organometallics 1992,
11, 1452.
(5) Kol, M.; Schrock, R. R.; Kempe, R.; Davis, W. M. J. Am. Chem.
Soc. 1994, 116, 4382.
(6) Albright, T. A.; Burdett, J. K.; Whangbo, M.-H. Orbital Interactions
in Chemistry; Wiley & Sons: New York, 1985.
(4) Gudat, D.; Verkade, J. G. Organometallics 1989, 8, 2772.
S0002-7863(97)00697-5 CCC: $14.00 © 1997 American Chemical Society

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11877
Table 1. Selected Distances (Å) and Angles (deg) for Structurally Characterized Compounds
[N3N]MoCla
[N3N]Mo(OTf)
[N3N]Mo(CD3)
[N3N]Mo(Cy)
[bitN3N]Mo
Mo-Lax
2.398(2)
1.976(3)
1.976(3)
1.976(3)
2.185(5)
128.2(2)
128.2(2)
128.2(2)
117.7(1)
117.7(1)
117.7(1)
176.6b
2.166(8)
1.974(8)
1.977(7)
1.980(8)
2.171(9)
129.8(5)
130.7(5)
126.2(5)
116.6(3)
117.9(3)
118.3(3)
136.1b
2.188(11)
1.984(9)
1.996(9)
1.973(10)
2.303(9)
128.0(5)
129.1(5)
127.1(5)
117.2(4)
117.6(4)
116.4(4)
162.4b
2.167(14)
1.991(7)
2.012(7)
1.987(7)
2.422(10)
128.2(4)
126.4(4)
131.9(4)
116.3(3)
113.9(3)
116.1(3)
129.1b
2.249(3)
1.948(2)
1.983(2)
1.994(2)
2.237(2)
102.30(10)
127.08(10)
125.87(11)
116.95(8)
116.83(8)
118.27(8)
176.1b
Mo-N(1)
Mo-N(2)
Mo-N(3)
Mo-N(4)
Mo-N(1)-Si(1)
Mo-N(2)-Si(2)
Mo-N(3)-Si(3)
N(1)-Mo-N(2)
N(2)-Mo-N(3)
N(1)-Mo-N(3)
N(4)-Mo-N(1)-Si(1)
N(4)-Mo-N(2)-Si(2)
N(4)-Mo-N(3)-Si(3)
N(1)-Mo-Lax
N(2)-Mo-Lax
N(3)-Mo-Lax
Mo-O(1)-S
176.6b
141.4b
163.3b
131.3b
178.9b
176.6b
142.6b
160.0b
99.9(4)
99.3(4)
100.7(4)
135.9b
105.8(3)
102.0(3)
99.6(3)
170.9b
76.13(9)
110.04(9)
110.68(11)
98.8b
103.4b
98.8b
97.7b
98.8b
95.8b
156.9(5)
Mo-C(101)-C(102)
Mo-C(101)-C(106)
Mo-C(11)-Si(1)
C(11)-Si(1)-N(1)
N(4)-Mo-C(11)
120.0(7)
116.9(7)
88.35(10)
92.93(11)
155.16(8)
a See ref 15. b Obtained from a Chem 3D model.
environment that is especially favorable for formation of a
metal-ligand triple bond.
We have been interested in the chemistry of Mo and W
triamidoamine complexes from several perspectives, including
dinitrogen fixation,5,7,8 terminal phosphido and arsenido
complexes,9-11 and new organometallic chemistry.7,12,13 In this
paper, we report the synthesis, characterization, and decomposi-
tion of a variety of alkyl complexes of molybdenum that contain
the [(Me3SiNCH2CH2)3N]3- ([N3N]3- ) ligand. Among the
unusual reactions reported here are dehydrogenation of [N3N]-
MoCH2R′ complexes to give [N3N]MotCR′ complexes and
cleavage of C-C bonds in cyclopropyl and cyclobutyl com-
plexes to give alkylidyne complexes. We also have been
presented with the opportunity to prove that R-elimination
reactions are faster than â-elimination reactions for cyclopentyl
and cyclohexyl complexes, even though the products of
R-elimination (alkylidene hydrides) and â-elimination (olefin
hydrides) are not observable. Some of these results have been
reported in a preliminary fashion.7,13 These results should be
compared with those obtained for related tungsten com-
plexes,12,13 the full details of which will be reported elsewhere.14
Figure 1. View of the structure of [N3N]Mo(triflate).
but so far without success. We suspect that loss of Me3SiCl or
reduction of Mo(IV) is a persistent problem, in part because
related reactions between (C6F5NHCH2CH2)3N and MoCl4(THF)2
in the presence of NEt3 give [(C6F5NCH2CH2)3N]MoCl in high
yield.5 The reaction between Li3[N3N] and MoCl3(THF)3 also
gives 1a in a yield of ∼25%. Similar results have been reported
by Verkade.15 Red [N3N]Mo(OTf) (1b) can be prepared in
∼62% yield by treating [N3N]MoCl with [Cp2Fe]OTf in
dichloromethane. We anticipated that the metal center in [N3N]-
Mo(OTf) would be more electrophilic than that in [N3N]MoCl,
possibly rendering it more suitable for some reactions. How-
ever, in reactions of the type discussed in this paper, [N3N]-
MoCl has been a reproducibly accessible and satisfactory
starting material.
It is instructive to compare the structure of [N3N]Mo(OTf)
(as determined in an X-ray study; Figure 1, Table 1) with the
published structure of [N3N]MoCl,15 selected bond distances
and angles for which are also listed in Table 1. The structure
of [N3N]MoCl shows that the TMS groups are all oriented
“upright,” i.e., the Nax-Mo-Neq-Si dihedral angles are all
Results
Chloride and Triflate Complexes. Addition of MoCl4-
(THF)2 to Li3[N3N] in THF yields paramagnetic [N3N]MoCl
(1a) in a maximum yield of ∼40%. Many variations of this
reaction have been attempted in an effort to increase the yield,
(7) Shih, K.-Y.; Schrock, R. R.; Kempe, R. J. Am. Chem. Soc. 1994,
116, 8804.
(8) O’Donoghue, M. B.; Zanetti, N. C.; Schrock, R. R.; Davis, W. M. J.
Am. Chem. Soc. 1997, 119, 2753.
(9) Zanetti, N. C.; Schrock, R. R.; Davis, W. M. Angew. Chem., Int. Ed.
Engl. 1995, 34, 2044.
(10) Johnson-Carr, J. A.; Zanetti, N. C.; Schrock, R. R.; Hopkins, M.
D. J. Am. Chem. Soc. 1996, 118, 11305.
(11) Wu, G.; Rovnyak, K.; Johnson, M. J. A.; Zanetti, N. C.; Musaev,
D. G.; Morokuma, K.; Schrock, R. R.; Griffin, R. G.; Cummins, C. C. J.
Am. Chem. Soc. 1996, 118, 10654.
(12) Shih, K.-Y.; Totland, K.; Seidel, S. W.; Schrock, R. R. J. Am. Chem.
Soc. 1994, 116, 12103.
(13) Schrock, R. R.; Shih, K.-Y.; Dobbs, D.; Davis, W. M. J. Am. Chem.
Soc. 1995, 117, 6609.
(14) Schrock, R. R.; Dobbs, D. A.; Seidel, S. W.; Mo¨sch-Zanetti, N. C.;
Shih, K.-Y.; Davis, W. M. Organometallics, in press.
(15) Duan, Z.; Verkade, J. G. Inorg. Chem. 1995, 34, 1576.

11878 J. Am. Chem. Soc., Vol. 119, No. 49, 1997
Schrock et al.
approaches 5 K. However, a plot of the susceptibility above
50 K versus T can be fit to a curve defined by a Curie-Weiss
equation (ø ) µ2/8(T + θ)) to give values of µ ) 2.92 and θ )
6.4. (The error in µ is estimated to be (0.10 and in θ to be
(1 K.) These data suggest that [N3N]MoCl is a Curie-Weiss
paramagnet between room temperature and ∼50 K in the solid
state with µeff essentially the spin only value expected for d2
Mo(IV). We assume that the two electrons are in the set of
degenerate nonbonding orbitals of π symmetry (approximately
dxz and dyz if Nax-Mo-Cl is taken to be the z axis). The
susceptibility of [N3N]MoCl in the solid state between 6 and
300 K was mentioned in another publication,15 but only Curie-
Weiss behavior above 100 K (µeff ) 2.35, θ ) 0.148 K) was
noted.
The fairly abrupt decrease in the effective magnetic moment
below ∼50 K can be attributed to a combination of spin-orbit
effects and low-symmetry ligand field components that result
in zero field splitting of the d2 ground-state spin triplet.16 If
the d2 complex were an undistorted octahedron with no low-
symmetry ligand field components and a 3T1g ground state, the
ambient-temperature magnetic moment might reflect a substan-
tial orbital contribution above spin-only (λ-negative) or might
actually be somewhat less than the latter for λ-positive, where
the magnitude of these effects depends on the magnitude of λ
relative to kBT. In this context, the moments for undistorted d2
systems will generally gradually decrease with temperature with
3
depopulation of the nine spin-orbit states of T1g (A). In the
limit of low symmetry (e.g., the C3 or C3V environments of the
present complexes), the orbital degeneracy of the ground and
all excited states is largely removed, leading to ambient
temperature moments closer to spin-only, regardless of the sign
or magnitude of λ (at least for Mo). The temperature variation
of µ or ø can be described in terms of a formalism17,18 that
emphasizes the axial zero field splitting (D) of the ground spin
manifold. Using this formalism, the observed susceptibility of
[N3N]MoCl could be modeled (R ) 0.9995) over the entire
range of temperatures and a D value of +35 cm-1 obtained
assuming a value of E ) 0.1 (where E is the rhombic splitting
parameter) and gx ) gy ) gz ) 2 (Figure 2a). All d2 Mo
compounds of the [N3N]MoX type that we have examined here
show a variation of susceptibility with temperature similar to
that observed for [N3N]MoCl. Since λ(W4+) ∼ 1050 cm-1
while λ(Mo4+) ∼ 475 cm-1 16
, spin-orbit effects are larger for
the W4+ analogues.14 Consequently [N3N]WX species have
reduced moments of ∼2 µB near room temperature and a
susceptibility that changes relatively little (compared to that for
the analogous [N3N]MoX species) as the temperature is
lowered.14
Figure 2. (a) Plot of ø versus T for [N3N]MoCl. (b) Plot of ø versus
1/T for [N3N]MoCl. (c) Plot of µeff versus T for [N3N]MoCl.
176.6°, as a consequence of Câ,ax (see Figure 1) serving as the
“flap” in an “envelope” configuration of the MoN2C2 five-
membered ring. The only significant difference between the
structures of [N3N]Mo(OTf) and [N3N]MoCl is that the Nax-
Mo-Neq-Si dihedral angles in [N3N]Mo(OTf) are much smaller
than 180° (136-143°) as a consequence of Câ,eq serving as the
“flap” in the MoN2C2 five-membered ring (Figure 1). We
speculate that the steric bulk of the triflate ligand causes the
trimethylsilyl groups in [N3N]Mo(OTf) to “twist” and that this
is the most facile response to a sterically bulky ligand in the
apical pocket. A significant increase in the Mo-Neq-Si angles,
which is another possible way to relieve steric problems,7 is
likely to be more costly energetically. The structures of
compounds to be discussed later confirm that the degree of
“twist” of the trimethylsilyl groups correlates with the size of
the ligand in the apical pocket.
The proton NMR spectrum of [N3N]MoCl consists of three
broadened, shifted resonances, one for the TMS groups and two
for the backbone methylene protons of the [N3N]3- ligand,
consistent with a paramagnetic species that has C3V symmetry
on the NMR time scale. It is not known with certainty which
of the two methylene resonances observed at higher temperatures
can be ascribed to which [N3N]3- backbone methylene group.
As the temperature of the NMR sample is lowered the spectrum
changes in two ways. First, the methylene resonances shift
upfield (Figure 3a) and the TMS resonance shifts (to a lesser
extent) downfield (not shown). If the chemical shift of each
methylene proton set is plotted versus 1/T, then nearly linear
plots are obtained (Figure 3b), as expected for a Curie-Weiss
(16) Figgis, B. N.; Lewis, J. Prog. Inorg. Chem. 1964, 6, 37.
(17) Oosterhuis, W. T.; Dockum, B.; Reiff, W. M. J. Chem. Phys. 1977,
67, 3537.
(18) Reiff, W. M.; Wong, H.; Baldwin, J. E.; Huff, J. Inorg. Chim. Acta
1977, 25, 91.
The magnetic susceptibility of [N3N]MoCl (plotted versus T
in Figure 2a and versus 1/T in Figure 2b) approaches a constant,
and µeff decreases sharply (Figure 2c) as the temperature

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11879
Figure 3. (a) Variable-temperature 500 MHz proton NMR spectra of [N3N]MoCl in toluene-d8. (b) Plot of the chemical shift of the methylene
backbone resonances versus 1/T.
paramagnet in solution in this temperature range. (The average
of the two methylene resonances below the coalescence point
discussed below was employed whenever possible in Figure 3b.)
The second temperature-dependent process that is observed is
a “locking” of the three MoN2C2 rings in the ligand backbone
at the lowest temperatures to give a species that has one C3-
symmetric conformation, e.g., the one in which Câ,eq is the “flap”
of the envelope (as in [N3N]Mo(triflate); see above), the one in
which Câ,ax is the “flap” of the envelope (as in [N3N]MoCl15),
or some conformation between the two (as in [N3N]Mo(CD3;
see below). At higher temperatures the C3-symmetric species
becomes C3V-symmetric on the NMR time scale as a conse-
quence of the MoN2C2 rings “flipping” rapidly between the two
chiral forms. Using the chemical shift difference between the
two types of backbone protons at a temperature that is as close
as possible to the temperature of coalescence, an approximate
∆Gq for the C3V/C3 fluxional process was calculated to be 9.2
kcal/mol at ∼-30 °C and 9.0 kcal/mol at ∼-20 °C. (The errors
in these values are estimated to be at least ( 0.2 kcal mol-1.)
A C3V/C3 fluxional process is observed for all [N3N]MoX species
discussed in this paper. A C3V/C3 fluxional process is also
observed for a variety of other C3V symmetric metal complexes
that contain the [N3N]3- ligand, including diamagnetic species,1
and for complexes that contain the [(C6F5NCH2CH2)3N]3-
ligand.19 We assume that the apical nitrogen donor does not
have to dissociate from the metal for the C3V/C3 fluxional process
to occur. We have not attempted in any systematic way to
correlate ∆G‡ for the C3V/C3 fluxional process with the nature
of the ligand in the apical pocket.
Molybdenum-Alkyl Complexes. Several n-alkyl and cy-
cloalkyl complexes, a 1-cyclopentenyl complex, and a phenyl
complex were prepared by treating [N3N]MoCl with lithium
reagents, as shown in eq 1. Several complexes could be isolated
in only ∼50% yield as a consequence of their high solubility
in hydrocarbon solvents. Compounds 2k,l are also paramag-
netic, i.e., conjugation of the olefinic or aromatic π systems
with the metal is insufficient to break the dxz/dyz degeneracy.
Many of these compounds are purple with a relatively weak
transition near 500 nm (e.g., ꢀ ) 680 M-1 cm-1 at 480 nm for
[N3N]MoMe in toluene; see the Experimental Section for other
compounds). However, the benzyl, trimethylsilyl, and neopentyl
complexes have visible spectra that show an additional, increas-
ingly strong absorption shifted to lower energy (Figure 4). The
visible spectrum of the neopentyl complex shows the strongest
low-energy absorption at ∼688 nm. The neopentyl complex
consequently is green-black instead of purple. We suspect that
these absorptions arise from “d-d” transitions (e.g., dxz/dyz to
2
2
(19) Seidel, S. W.; Reid, S. M. Unpublished results.
dx -y /dxy) and speculate that increasingly significant distortions

11880 J. Am. Chem. Soc., Vol. 119, No. 49, 1997
Schrock et al.
Figure 6. View of the sttructure of [N3N]Mo(cyclohexyl).
Table 2. Values of µeff in the Solid State for [N3N]MoX
Compoundsa
Figure 4. UV/vis spectra of [N3N]MoR complexes in toluene: R )
(a) Me; (b) CH2Ph; (c) CH2SiMe3; (d) CH2CMe3.
compd
[N3N]MoCl
µeff
2.92
2.76 (2.61)
2.72 (2.78)
2.58 (2.60)
2.79
θ (K)
6.4
[N3N]MoMe
11.1 (4.8)
3.8 (7.5)
-2.2 (-2.4)
-2.5
[N3N]Mo(CH2SiMe3)
[N3N]Mo(cyclopentyl)
[N3N]MoD
[bitN3N]Mo
2.46 (2.78)
0.8 (2.12)
a Obtained from SQUID data between 5 and 300 K by fitting the
corrected susceptibility between 50 and 300 K to the equation ø )
µ2/8(T + θ). The error in µ is estimated to be (0.10 and in θ to be
(1. Numbers in parentheses refer to a second sample.
The Mo-C(101) distance (2.167(14) Å) is the same as Mo-
C(7) in [N3N]Mo(CD3). The Mo-C(101)-C(106) and Mo-
C(101)-C(102) angles (117-120°) are larger than tetrahedral,
consistent with a significant amount of steric pressure within
the trigonal pocket. The possibility of any concomitant R
agostic20,21 interaction between the C(101)-HR bond and the
metal seems remote in view of the fact that the only available
orbitals (dxz and dyz) each contain one electron. Interestingly,
the N(4)-Mo-Neq-Si dihedral angles are all dramatically
smaller than found in [N3N]Mo(CD3) (smaller even than in
[N3N]Mo(OTf)) and the Mo-N(4) distance is approximately
0.1 Å longer than in [N3N]Mo(CD3). We propose that both
are consistent with a significantly greater degree of steric
interaction between the cyclohexyl ring and the silyl groups on
the equatorial amido nitrogens. “Steric pressure” within the
pocket in [N3N]Mo(alkyl) complexes could sharply distinguish
one complex from another in terms of visible spectra (as
proposed above) or, as we shall see shortly, in terms of the
ease of some proton elimination processes.
Figure 5. View of the structure of [N3N]Mo(CD3).
of the MoN4 core when a bulky CH2R ligand is present in the
apical position lead to a splitting of orbitals that are degenerate
in C3V symmetry and to new lower energy absorptions. Detailed
studies clearly will be required to elucidate the origin of this
behavior fully and possibly to correlate any structural distortions
with altered magnetic behavior.
A view of the structure of [N3N]Mo(CD3), as determined in
an X-ray study, is shown in Figure 5, and bond distances and
angles are compared with distances and angles in other
compounds in Table 1. (The choice of the CD3 over the CH3
complex was circumstantial.) The structure of [N3N]Mo(CD3)
is most similar to that of [N3N]MoCl, except the TMS groups
are twisted slightly (N(4)-Mo-N-Si angles of 160-163°) and
both Câ,eq and Câ,ax deviate from a given MoNaxNeq plane. This
configuration of a MoN2C2 five-membered ring is between that
in the “upright” and “twisted” forms referred to above for [N3N]-
MoCl and [N3N]Mo(OTf), respectively. The long Mo-N(4)
distance (2.303(9) Å) could be ascribed to the metal being less
electrophilic in [N3N]Mo(CD3) than in [N3N]MoCl (Mo-N(4)
) 2.185(5) Å) or [N3N]Mo(OTf) (Mo-N(4) ) 2.171(9) Å).
We propose that the TMS groups are twisted to a greater degree
in [N3N]MoMe than in [N3N]MoCl as a consequence of the
slightly larger methyl ligand being present in the apical pocket.
A drawing of the structure of [N3N]Mo(cyclohexyl) is shown
in Figure 6. (See Table 1 for selected distances and angles.)
The cyclohexyl ring is oriented so that it lies roughly in a plane
that passes between N(1) and N(3) and between N(1) and N(2).
The susceptibilities of [N3N]MoMe, [N3N]Mo(CH2SiMe3),
and [N3N]Mo(cyclopentyl) in the solid state vary with temper-
ature in ways that are entirely analogous to the behavior of the
susceptibility of [N3N]MoCl versus temperature. Results for
[N3N]MoMe and [N3N]Mo(cyclopentyl) are shown, along with
the susceptibility of two compounds to be discussed later, in
(20) Brookhart, M.; Green, M. L. H.; Wong, L. Prog. Inorg. Chem. 1988,
36, 1.
(21) (a) Maus, D. C.; Copie´, V.; Sun, B.; Griffiths, J. M.; Griffin, R. G.;
Luo, S.; Schrock, R. R.; Liu, A. H.; Seidel, S. W.; Davis, W. M.; Grohmann,
A. J. Am. Chem. Soc. 1996, 118, 5665. (b) Bell, A.; Clegg, W.; Dyer, P.
W.; Elsegood, M. R. J.; Gibson, V. C.; Marshall, E. L. J. Chem. Soc., Chem.
Commun. 1994, 2547. (c) Casty, G. L.; Tilley, T. D.; Yap, G. P. A.;
Rheingold, A. L. Organometallics 1997, 16, 4746.

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11881
is shown in Figure 8a. As the temperature is lowered the two
ligand methylene resonances shift to higher field in a linear
fashion versus 1/T (Figure 8b). Values for ∆Gq for the C3V/C3
fluxional process that becomes slow on the NMR time scale at
∼-60 °C in [N3N]Mo(CH3) can be calculated as described for
[N3N]MoCl; values thus calculated are 8.6 and 8.2 kcal mol-1
.
(We estimate the error to be at least (0.2 kcal mol-1 for each.)
These values for ∆Gq are close to those found for [N3N]MoCl
(9.0 and 9.2 kcal/mol). Therefore it does not appear, according
to these data, that activation energies for the C3/C3V fluxional
processes in different compounds in this general family of [N3N]-
Mo complexes differ dramatically.
So far we have not been able to observe an R proton
resonance in any [N3N]MoR complex. To be certain that we
2
did not overlook an R-proton resonance, we searched for a H
resonance in [N3N]Mo(CD3) between +400 and -400 ppm.
(We chose 2H NMR because resonances in 2H NMR spectra of
paramagnetic compounds are often significantly narrower than
those in 1H NMR spectra and, therefore, are more readily
observed in paramagnetic complexes.22) However, we could
observe no 2H resonance that could be ascribed to the CD3 group
in [N3N]Mo(CD3), even after hours of collection time at 76.7
MHz. We conclude that either the CD3 resonance in [N3N]-
Mo(CD3) is outside the +400 to -400 ppm region or its
intensity is virtually nil using the standard methods of data
collection that we have employed so far, or both. In no other
[N3N]Mo(alkyl) species discussed here have we been able to
Figure 7. Plot of ø versus T for [N3N]MoMe, [N3N]Mo(cyclopentyl),
[N3N]MoD, and [bitN3N]Mo.
Figure 7. In Table 2 are shown values for µ and θ determined
from a fit of the data above 50 K to a Curie-Weiss law. The
magnetic moment for [N3N]MoMe in C6D6 at 25 °C was found
to be 2.9 µB by the Evan’s method, consistent with the results
obtained in the solid state, and with an S ) 1 spin ground state.
NMR spectra of [N3N]MoR species show broadened and
shifted ligand resonances and variable-temperature behavior that
is analogous to that noted for N3N]MoCl (Figure 3a). For
example, the high-field portion of the spectrum of [N3N]MoMe
that contains resonances for the backbone methylene protons
1
2
observe any H or H resonance for protons or deuterons on a
carbon directly bound to Mo.
Figure 8. (a) Variable-temperature 500 MHz proton NMR spectra of the ligand backbond resonances in [N3N]MoMe. (b) Plot of the chemical
shift of the methylene backbone resonances versus 1/T.

11882 J. Am. Chem. Soc., Vol. 119, No. 49, 1997
Schrock et al.
Figure 9. (a) 500 MHz 1H NMR spectrum of [N3N]Mo(C5H9) at 22 °C. (b) Variable-temperature 76.7 MHz 2H NMR spectra of [N3N]Mo(C5H8D).
Resonances for protons on the â-carbon of an alkyl ligand
can be observed, however. For example, the 1H NMR spectrum
of [N3N]Mo(CH2CH3) shows a broad resonance at -51.5 ppm
(at 22 °C), in addition to temperature-dependent resonances that
can be ascribed to the TMS groups and two types of ligand
backbone methylene protons, as described earlier for analogous
species. â-Proton resonances in other alkyl complexes (see
below) are also typically found in the range -50 to -60 ppm.
The position of the â-proton resonances also depends on
temperature, as expected for a Curie-Weiss paramagnetic
species, moving to higher field at lower temperatures.
The variable-temperature 1H NMR spectrum of [N3N]Mo(CH2-
CH2CH2CH3) between 60 and -60 °C shows (in addition to
the TMS and backbone methylene resonances) three resonances
(at 22 °C) at -48, +2.9, and +5.4 ppm. The resonance at -48
ppm can be assigned to the â methylene protons, while the
resonances at +2.9 and +5.4 ppm can be assigned to the γ and
δ proton resonances, or vice versa. The only temperature-
dependent processes observed in the range +60 to -60 °C are
the Curie 1/T chemical shift dependence and “locking” of the
ligand backbone to give a species with C3 symmetry at low
temperatures.
The 1H NMR spectrum of [N3N]Mo(cyclopentyl) shows four
broadened and shifted resonances for cyclopentyl ligand protons
at 21.3, 15.6, -48.2, and -58.3 ppm at 22 °C (Figure 9a). On
the basis of chemical shift, the resonances at 21.3 and 15.6 ppm
can be ascribed to exo and endo γ-protons (without specifying
which is which), endo protons being defined as those on the
same side of the ring as the metal, while those at -48.2 and
-58.3 ppm can be ascribed to exo and endo â-protons (again
without being able to specify which is endo and which is exo).
The HR proton resonance is not observable, and at low
temperatures, the ligand backbone becomes locked in a C3-
symmetric configuration.
2
The H NMR spectrum of [N3N]Mo(cyclopentyl-R-d) (pre-
(22) NMR of Paramagnetic Molecules; LaMar, G. N., Horrocks, W. D.,
Jr., Holm, R. H., Ed.; Academic: New York, 1973.
pared by adding Li(cyclopentyl-R-d) to [N3N]MoCl and isolated

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11883
Figure 10. 46.02 MHz 2H NMR spectra (at -20 °C) of [N3N]Mo(R-C5H8D) as a function of the time the sample was kept at 25 °C. (* A resonance
2
for natural abundant H in the TMS group of [N3N]MoH, a decomposition product; see text.)
in a normal workup procedure at room temperature; Figure 9b)
revealed four broad resonances for 2Hγ,exo and 2Hγ,endo (or vice
2
2
versa) and Hâ,exo and Hâ,endo (or vice versa) at the same
chemical shifts (at a given temperature) as found in the 1H NMR
spectrum of [N3N]Mo(cyclopentyl) (Figure 9a). They are best
observed when the sample is cooled to 0 °C for reasons that
will be discussed below. These data confirm that the one
deuterium has washed into all exo and endo â and γ cyclopentyl
elimination would yield the species in which D appears on a
γ-carbon in an exo position. However, if only â-hydride
elimination were taking place, D would appear only in exo
positions (eq 3). Therefore another faster process must inter-
2
positions. (We assume at this stage that residual HR is also
present but is not observable.) When [N3N]Mo(cyclopentyl-
R-d) was prepared by adding Li(cyclopentyl-R-d) to [N3N]MoCl
in toluene at -20 °C and the sealed sample was monitored by
2H NMR, the results shown in Figure 10 were obtained. Initially
only a resonance for the deuterium in Li(cyclopentyl-R-d) was
observed at 0.6 ppm. (Alkylation is relatively slow at -20 °C
in toluene.) The sample was then warmed for brief periods to
convert exo and endo protons on â- and γ-carbons. (This faster
2
25 °C and cooled to -20 °C to record the H NMR spectrum.
process is discussed below.) We can estimate from Figure 10
2
After a total of ∼10 min at 25 °C all Li(cyclopentyl-R-d) had
been consumed. However, at this stage, only weak 2Hâ
resonances were observed (at -20 °C) at -58 and -72 ppm.
No γ-resonances were observed near 20 ppm. The intensity of
the â-resonances increased with time, reaching a maximum after
the sample had stood for a total of 1-2 h at room temperature.
After 2 h some γ-deuteron intensity was observed near 20 ppm.
After 24 h the spectrum clearly revealed resonances at 18 and
that scrambling of H from the R carbon to the â carbons in
both exo and endo positions is complete (∼5 half-lives) in 60-
120 min. Therefore the half-life for appearance of D in a
â-position is between 12 and 24 min at 22 °C or kâ ) ∼5 ×
10-4 s-1 to ∼10-3 s-1. No significant cyclopentene is lost from
the cyclopentene hydride intermediate under these conditions.
(We show later that at 66 °C t1/2 ) 17 min for cyclopentene
elimination.) At equilibrium one-ninth of one deuterium should
be found at each of the nine cyclopentyl positions, ignoring
any equilibrium isotope effects and assuming that D is scrambled
relatively rapidly between exo and endo positions by some other
process.
2
27 ppm (at -20 °C) for H in γ-positions (Figure 10). The
resonances at 18 and 27 ppm are slightly less intense than those
at -58 and -72 ppm, probably as a consequence of the rapidly
decreasing rate of incorporation into γ-positions for statistical
2
2H NMR spectra of [N3N]Mo(cyclopentyl-R-d) from 0 to 50
°C (Figure 9b) reveal that exo and endo deuterons equilibrate
rapidly on the NMR time scale at the higher temperatures, where
decomposition to give [N3N]MoH (see later) is still slow. The
four broad resonances observed at low T coalesce to two average
resonances that shift toward the diamagnetic region of the
spectrum as a consequence of their Curie dependence on 1/T.
This behavior is entirely reversible. (The sharp resonances
between 4 and 9 ppm in Figure 9b are due to naturally abundant
2H in the toluene solvent and the small amount of cyclopentene-
d1 formed upon decomposition of [N3N]Mo(cyclopentyl-d1) at
the higher temperatures.) This fluxional process cannot be
observed unambiguously by 1H NMR for several reasons, among
reasons, as only one H total is present. (The γ-resonances in
the sample whose 2H NMR spectrum is shown in Figure 9b are
also much less intense than the â-resonances, as this solid sample
had been kept at -30 °C after being isolated at 25 °C.)
Therefore we can say confidently that R-D in [N3N]Mo-
(cyclopentyl-R-d) washes into the â,exo and â,endo positions
of the cyclopentyl ring before it washes into the γ,exo and
γ,endo positions.
The mechanism by which D appears on a â-carbon in an exo
position is proposed to be the well-known â-hydride elimination
(eq 2). (We will show later that loss of cyclopentene from
intermediate [N3N]Mo(cyclopentene)(H) is relatively slow and
irreversible under these conditions.) Subsequent â-hydride

11884 J. Am. Chem. Soc., Vol. 119, No. 49, 1997
Schrock et al.
them the factor of 6.5 difference between 1H and 2H frequencies.
Note that the fluxional process that leads to interconversion of
is possible to heat it to higher temperatures before it decomposes
to [N3N]MoH and cyclohexene (see below). Nevertheless,
resonances for [N3N]MoH complicate the study, as do other
resonances for [N3N]Mo(cyclohexyl). We assign the resonances
at 15.8 and 6.1 ppm to exo and endo γ-protons, those at 1.9
and -2.0 ppm to exo and endo δ-protons, and those at -40.9
and -49.9 ppm to exo and endo â-protons. Interference by
other resonances and the breadth of the cyclohexyl â-, γ-, and
δ-resonances in the 1H NMR spectrum of [N3N]Mo(cyclohexyl)
makes these results less convincing than those for [N3N]Mo-
(cyclopentyl-d1). Nevertheless, we can observe coalescence of
the δ-proton resonances at 64 ( 2 °C (shown in Figure 11a)
and the â-resonances at ∼100 °C (less accurately and not
shown). Unfortunately, the γ-resonances coalesce under the
broad resonance for the TMS group in [N3N]Mo(cyclohexyl)
at ∼10 ppm. From coalescence of exo and endo δ-protons we
can obtain an approximate value for kR,C6 (2 × 103 s-1) at 64
°C. This value should be compared with the value for kR in
the cyclopentyl complex, kR,C5 ≈ 103 s-1 at 22 °C. At 64 °C,
kR,C5 perhaps would be ∼16 × 103, or approximately eight times
larger than kR,C6. (Another estimate at a different temperature
is possible using another approach below.) Note that in the
cyclohexyl case it is possible to distinguish a rapid R elimination
from a rapid â-elimination on the basis of the number of average
resonances alone (three ideally), regardless of assignment.
2
exo and endo H cannot inVolVe any R-2H in the averaging
process. If 2HR were scrambling rapidly with γ- or â-protons,
then the resulting average resonance would be either shifted
significantly or decreased in intensity dramatically, or both.
2
2
2
Instead, Hâ,exo averages with Hâ,endo, Hγ,exo averages with
2Hγ,endo, and the average Hâ and Hγ resonances shift toward
the diamagnetic region as the temperature approaches 50 °C.
The fluxional process is not consistent with any slowed rotation
of the cyclopentyl ring within the trigonal pocket, as the
backbone still has pseudo C3V symmetry at a temperature where
2
2
2
the cyclopentyl H ring resonances are resolved (0 °C in the
1H NMR spectrum.)
We propose that the fluxional process that leads to averaging
of exo and endo 2H on Câ and exo and endo 2H on Cγ consists
of a rapid and reVersible R-elimination to give an intermediate
[N3N]Mo(cyclopentylidene)(H) complex (eq 4). This unusual
The 2H NMR spectrum of [N3N]Mo(C6D11) at 46.02 MHz is
shown in Figure 11b. The expected six C6D11 resonances in
the ratio of 2(γ):2(γ):1(δ):1(δ):2(â):2(â) are observed at 10 °C.
Upon heating the sample these resonances broaden and coalesce
pairwise. The coalescence temperature varies for each set of
resonances as a consequence of the different chemical shift
difference between the respective exo and endo 2H resonances
(γ > â > δ). But since a single physical process (R-elimination)
is responsible for all coalescence phenomena, we can obtain a
rate for that process at three different temperatures from the
spectrum shown in 11b. The results for a ∼0.1 M sample in
toluene are kR ) 378 s-1 at 324 K, 755 s-1 at 339 K, and 955
s-1 at 343 K. Virtually identical values were obtained for a
sample that was one-half the concentration (0.05 M). By
performing a similar experiment at 76.7 MHz three additional
points were obtained: kR ) 606 s-1 at 331 K, 1273 s-1 at 346
K, and 1555 s-1 at 352 K. A plot of ln(k/T) versus 1/T for
these six points gave ∆Hq ) 11(1) kcal mol-1 and ∆Sq ) -14-
(3) e.u. (Exact values may be found in the Experimental
Section.) From the data shown in Figure 11a a value for the
rate constant for R-elimination in [N3N]Mo(C6H11) at 337 K
was found to be 2155 s-1. Since the rate of R-elimination in
[N3N]Mo(C6D11) at 337 K can be calculated to be 696 s-1, kH/
kD for R-elimination at 337 K is 3.1. (We estimate the error
(σ) in kH/kD to be at least (0.5.) Observation of a substantial
isotope effect confirms that a C-H bond cleavage is involved
in the process that gives rise to interconversion of exo and endo
protons and, therefore, rules out a possible, if implausible,
alternative equilibration by C-C bond cleavage and formation
of a metallacycloheptene ring. We also can compare the rate
constant for R-elimination at 22 °C for [N3N]Mo(C6D11)
(calculated to be ∼64 s-1) and [N3N]Mo(C5H9) (estimated to
be ∼1000 s-1). The kR at 22 °C for [N3N]Mo(C6H11) should
be ∼200 s-1 on the basis of kH/kD ≈ 3. The rate of
R-elimination in [N3N]Mo(C5H9) is approximately 5 times faster
than the rate of R-elimination in [N3N]Mo(C6H11) at 22 °C (cf.
the estimate of 8 times faster at 64 °C above).
proposal is sensible in view of the fact that an attempt to prepare
[N3N]W(cyclopentyl) yielded crystallographically characterized
[N3N]W(cyclopentylidene)(H) instead.13 The rate constant for
R-elimination (kR) at 22 °C can be estimated to be 103 (from
NMR experiments), while the half-life for â-elimination can
be estimated to be between 12 (k ) 5 × 10-4 s-1) and 24 (k )
10-3 s-1) min at 22 °C (see above). Therefore we can say that
the rate constant for R-elimination is at least 6 orders of
magnitude larger than the rate constant for â-elimination at room
temperature, i.e., kR > 106 kâ (eq 4). (An accurate comparison
of the rate constants for R- and â-elimination per C-H bond
would have to include a factor of 2 for statistical reasons, but
we need not consider that adjustment here in view of the
magnitude of the rate differences in question.) We assume that
[N3N]Mo(cyclopentylidene)(H), like [N3N]W(cyclopentylidene)-
(H),13 is diamagnetic and that [N3N]Mo(cyclopentene)(H) also
would be diamagnetic. There is no evidence that suggests that
any significant amount of either is present, i.e., the equilibria
for both R- and â-elimination lie well toward [N3N]Mo-
(cyclopentyl). If we assume that the equilibrium constants for
forming [N3N]Mo(cyclopentyl) (kR,rev/kR and kâ,rev/kâ) are both
∼102 (or greater), then kâ,rev ≈ 10-1 s-1 (or greater) and kR,rev
≈ 105 s-1 (or greater). An alternative explanation of the
phenomenon of interconversion of exo and endo cyclopentyl
protons involves C-C bond cleavage and formation of a
1-molybdacyclohexene complex that has mirror symmetry, but
rapid and reversible C-C bond cleavage would seem to be much
less likely than rapid and reversible C-H bond cleavage. (See
also the Discussion.) It should be noted that a significant
primary H/D kinetic isotope effect might be expected for an
R-elimination process,23 but since only one 2H is present among
eight 1H nuclei in the sample whose 2H NMR spectrum is shown
in Figure 9b, we are observing a virtually pure 1HR elimination
in this case.
The 1H NMR spectrum of [N3N]Mo(cyclohexyl) (Figure 11a)
shows six resonances for the cyclohexyl ring exo and endo
protons that average upon heating the sample. [N3N]Mo-
(cyclohexyl) is more stable than [N3N]Mo(cyclopentyl), so it
Both [N3N]Mo(cyclobutyl) and [N3N](cyclopropyl) have been
1
studied by variable-temperature H NMR to determine if a
(23) Schrock, R. R. In Reactions of Coordinated Ligands; P. R.
Braterman, Ed.; Plenum: New York, 1986.
similar rapid R-elimination process is occurring. The endo and

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11885
Figure 11. (a) 300 MHz 1H NMR spectra of [N3N]Mo(C6H11) at two temperatures. Peaks labeled â, γ, and δ are due the cyclohexyl ring protons.
Peaks labeled with * are due to [N3N]MoH, which is formed relatively rapidly at 70 °C. Peaks labeled # are due to cyclohexene, the other product
2
resulting from thermolysis of [N3N]Mo(C6H11). The & indicates the residual protonated toluene resonances. (b) The 46.02 MHz variable-T H
NMR spectrum of [N3N]Mo(C6D11).
exo â-protons in the cyclopropyl complex are separated by >110
ppm. A value of 110 ppm would require too high a temperature
to observe coalescence, since the cyclopropyl complex decom-
poses to [N3N]MotCH with a half-life of ∼1 h at 50 °C (see
below). In [N3N]Mo(cyclobutyl), the endo and exo γ-protons
are too close to each other and broad to evaluate whether they
indeed coalesce. (As the sample is warmed, the resonances
simply shift to give one broad resonance, presumably as a
consequence of Curie 1/T dependence.) Therefore we cannot
confirm or exclude a reversible R-elimination process in [N3N]-
Mo(cyclobutyl) and [N3N](cyclopropyl) at this stage.
Decomposition of Molybdenum-Alkyl Complexes. When
solutions of [N3N]Mo(CH2-t-Bu) in toluene are heated, dihy-
drogen is lost and [N3N]MotC-t-Bu (3f) is formed quantita-
tively in a first-order reaction (eq 5). This conversion can be
followed conveniently by visible spectroscopy by monitoring
the decrease in intensity of the absorption at either 480 or 688
nm (Figure 4), since the alkylidyne complex does not absorb
significantly in this region. The rate was measured in toluene
at four temperatures from 90 to 121 °C (see the Experimental
Section) and from a plot of ln(k/T) versus 1/T ∆Hq was found
to be 26.0 kcal mol-1 and ∆Sq to be -7 eu (Table 3). A

11886 J. Am. Chem. Soc., Vol. 119, No. 49, 1997
Schrock et al.
Table 3. ∆Hq and ∆Sq Values for Decomposition of Alkyls
of tungsten complexes of this type.12 It is proposed that [N3N]-
Mo(CH2R′) is converted into [N3N]Mo(H)(CHR′) and that
dihydrogen is lost from [N3N]Mo(H)(CHR′) by a type of “R-
hydrogen abstraction” reaction (eq 6).23 We have shown that
Obtained via UV/Vis Kinetic Techniques
∆Hq
∆Sq k (341 K) k (298 K)
a
c
c
(kcal mol-1
)
(eu)b (10-4 s-1
)
(10-6 s-1
0.014
)
[N3N]Mo(neopentyl)
(2f) in toluene
[N3N]Mo(cyclopropyl)
(2g) in toluene
[N3N]Mo(cyclopropyl)
(2g) in THF
[N3N]Mo(cyclobutyl)
(2h) in toluene
[N3N]Mo(cyclopentyl)
(2i) in toluene
[N3N]Mo(cyclohexyl)
26.0
21.5
21.4
24.4
21.6
24.5
-7
-9
-10
0
0.042
10.0
8.8
7.0
6.9
5.3
0.40
7.7
14.5
6.2
-10
-5
the rates of forming a cyclopentylidene hydride from [N3N]-
Mo(cyclopentyl) and a cyclohexylidene hydride from [N3N]-
Mo(cyclohexyl) can be relatively fast. Since the difference
between the rates of R-elimination in [N3N]Mo(cyclopentyl) and
[N3N]Mo(cyclohexyl) is already significant (∼10×), we suspect
that the rates of forming [N3N]Mo(H)(CHR′) complexes from
various [N3N]Mo(CH2R′) complexes also will differ dramati-
cally. (See the Discussion.)
[N3N]Mo(CH2CH2CH2CH3) decomposes at 80 °C to give a
mixture of paramagnetic [N3N]MoH (see later) and [N3N]-
MotCPr (see later), while [N3N]Mo(CH2CH3) decomposes to
what we presume to be C3-symmetric [N3N]MotCMe, in both
cases as a consequence of competitive R,R-dehydrogenation (eq
6) and â hydride elimination followed by loss of olefin (eq 7).
0.86
(2j) in toluene
a Error ≈ 0.2 kcal. b Error ≈ 1 eu. c Calculated from exact values
for ∆Hq and ∆Sq given in the Experimental Section.
characteristic low-field resonance for the alkylidyne carbon
atom24 in [N3N]MotC-t-Bu was observed at 316.1 ppm.
[N3N]Mo(CH2SiMe3) decomposes much more slowly than
[N3N]Mo(CH2CMe3), and the decomposition is more compli-
cated. After 2 days in toluene-d8 at 100 °C a sample of
[N3N]Mo(CH2SiMe3) had only partially decomposed to two
diamagnetic products, a C3-symmetric species and an Cs-
symmetric species that contains characteristic resonances at 12.1
(triplet) and 4.55 (doublet) ppm in a ratio of 1:2. After 5 days
at 100 °C only the C3-symmetric species was clearly visible in
solution in a yield (versus an internal standard) of ∼30%.
Although the low yield prevented confirmation by 13C NMR
that the C3-symmetric species is [N3N]MotCSiMe3, the chemi-
cal shifts of the proton NMR resonances in the [N3N]3- ligand
are virtually identical with those for well-characterized alkyli-
dyne complexes of this type ([N3N]MotCH and [N3N]MotCPr
below). Evidently the unidentified Cs-symmetric species is not
stable at 100 °C, while [N3N]MotCSiMe3, like other alkylidyne
complexes in this category, is stable at 100 °C. Although the
complex nature of the decomposition of [N3N]Mo(CH2SiMe3)
to [N3N]MotCSiMe3 prevented measurement of a meaningful
half-life, we estimate it to be ∼48 h at 100 °C. For comparison,
the half-life of [N3N]Mo(CH2CMe3) at 100 °C is ∼1.4 h. We
prefer not to speculate on the nature of the unstable Cs-
symmetric species formed in this decomposition reaction.
[N3N]Mo(CH2Ph) decomposes even more slowly than
[N3N]Mo(CH2SiMe3), and the yield of what we presume to be
[N3N]MotCPh (according to its C3-symmetry and [N3N]3-
ligand chemical shifts in the proton NMR spectrum) is again
low. After 5 days at 100 °C in toluene-d8 a sample contained
13% [N3N]Mo(CH2Ph) and 40% [N3N]MotCPh (versus an
internal standard); the fate of the remaining material could not
be determined.
Heating [N3N]MoMe to 120 °C for 24 h in toluene-d8 did
not yield any significant amount of [N3N]MotCH, a compound
that can be prepared by another route (see later); [N3N]MoMe
simply slowly disappeared, and no decomposition products could
be identified. This result should be compared with the
decomposition of [N3N]WMe to [N3N]WtCH at 120 °C, for
which the calculated half-life is ∼10 s.12 Therefore [N3N]-
MoMe is approximately 4 orders of magnitude more stable than
[N3N]WMe at 120 °C with respect to decomposition to give a
methylidyne complex. It is interesting to note that [N3N]MoMe
does not lose methane by C-H activation in the TMS group to
give a complex that is stable under these conditions (see later).
Decomposition of [N3N]Mo(CH2R′) to [N3N]MotCR′ is
analogous to the faster and better documented decompositions
After heating a solution of [N3N]MoBu for 2 days at 80 °C the
ratio of [N3N]MoBu to [N3N]MoH to [N3N]MotCPr was
67:21:12; after 7 days it was 17:54:29. After heating a solution
of [N3N]Mo(CH2CH3) for 16 h at 80 °C the ratio of [N3N]-
MoEt to [N3N]MoH to [N3N]MotCMe was 60:28:12. [N3N]-
MotCPr can be made more practically from [N3N]Mo-
(cyclobutyl) (see below) and [N3N]MoH from [N3N]-
Mo(cyclopentyl) (see below). Qualitatively the rate of formation
of [N3N]MoH from [N3N]MoBu or [N3N]MoEt is significantly
less than from the analogous cyclopentyl or cyclohexyl com-
plexes (see below), perhaps in part because a monosubstituted
olefin is not lost as readily from the metal in a [N3N]Mo(olefin)-
(H) intermediate as is a cyclic olefin. It is clear that for ethyl
and butyl complexes the overall rate of forming [N3N]MoH via
the two-step process shown in eq 7 is approximately the same
as the rate of forming an alkylidyne complex via the two-step
process shown in eq 4.
[N3N]Mo13CH2CH2CH2CH3 was prepared and its decomposi-
tion monitored by 13C NMR in a sealed NMR tube.
A
resonance was observed at 298.3 ppm, the same position as the
R-carbon atom in [N3N]MotCCH2CH2CH3 prepared by another
method (see below), consistent with formation of virtually
entirely [N3N]Mot13CCH2CH2CH3. On the basis of relative
intensities of resonances in the 13C NMR spectrum, we estimate
that <10% [N3N]MotCCH2CH213CH3 (δ13C ) 14.6 ppm) is
(24) Murdzek, J. S.; Schrock, R. R. Carbyne Complexes; VCH: New
York, 1988.

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11887
present. Resonances for 1-butene also were consistent with
formation of >90% 13CH2dCHCH2CH3. This is the result to
be expected if the metal cannot “walk” to the other end of the
butyl chain (by forming [N3N]Mo(sec-butyl) and [N3N]Mo(2-
butene)(H) intermediates) before 1-butene is lost. If an isomer-
ization of this type were rapid relative to the rate of loss of
1-butene, then a 1:1 mixture of 13CH2dCHCH2CH3 and
CH2dCHCH213CH3 would be obtained. Formation of only
[N3N]Mot13CCH2CH2CH3 is also consistent with no
[N3N]Mo(CH2CH2CH213CH3) being formed before R,R-dehy-
drogenation.
When 2g is heated it is converted into pale yellow [N3N]-
MotCH (3a; eq 8), while 2h is converted into [N3N]MotCCH2-
CH2CH3 (3c; eq 9). Both reactions are quantitative and can be
followed conveniently by UV/vis spectroscopy at 540 and 558
nm, respectively, in toluene or THF. The reactions are first
order through several half-lives and independent of concentration
over a 40 °C temperature range. Rates in THF are within
experimental error the same as those in toluene, as shown in
Table 3 for 2g. The rates of decomposition of 2g and 2h are
comparable and more than 2 orders of magnitude faster than
the rate of decomposition of [N3N]Mo(neopentyl) to [N3N]-
MotCCMe3. (See calculated k values at 298 K in Table 3.)
The reaction shown in eq 8 is the only practical way to prepare
conditions where incorporation would be easily ascertained.
Therefore the decomposition of 2i is essentially irreversible.
The ∆Hq and ∆Sq values for decomposition of 2i are virtually
the same as those for [N3N]Mo(cyclopropyl) (Table 3). The
hydride resonance could not be located in the NMR spectrum
of [N3N]MoH, but an M-H stretch could be found in the IR
spectrum at 1681 cm-1
. [N3N]MoΗ is best prepared by
preparing 2i in situ and heating the solution thereafter. [N3N]-
MoΗ shows the same type of temperature-dependent suscep-
tibility as other compounds of this type, as shown in Figure 7
for [N3N]MoD. A fit of these data to an equation17,18 involving
g () 2), E () 0.1 cm-1), and D yielded a value of D ) 22
cm-1 (R ) 0.9993).
[N3N]Mo(cyclohexyl) also decomposes cleanly to give [N3N]-
MoH, as determined by NMR and by following the reaction in
toluene by UV/vis at 562 nm. The first-order conversion of
[N3N]Mo(cyclohexyl) into [N3N]MoH was followed at five
temperatures between 323 and 363 K, and the values of ∆Hq
and ∆Sq from an Eyring plot are listed in Table 3. The
decomposition of [N3N]Mo(cyclohexyl) is ∼10× slower than
that of [N3N]Mo(cyclopentyl). (At 298 K the rate constants
are calculated to be 4.0 × 10-7 s-1 for [N3N]Mo(cyclohexyl)
and 5.3 × 10-6 s-1 for [N3N]Mo(cyclopentyl).) At 343 K [N3N]-
Mo(C6D11) was found to decompose with k ) 3.0 × 10-5 and
3.3 × 10-5 s-1. Since the rate constant for decomposition of
[N3N]Mo(C6H11) at 343 K is calculated to be 10.7 × 10-5 s-1
,
the isotope effect at that temperature is 3.4(2). A deuterium
isotope effect of this magnitude is consistent with cleavage of
a CH bond.
At 100 °C [N3N]MoH begins to evolve dihydrogen to give
the paramagnetic “[bitN3N]Mo” complex shown in eq 11.
3a, since [N3N]MoMe will not lose dihydrogen to give 3a. The
reaction shown in eq 9 is also the only practical way to prepare
3c, since [N3N]Mo(n-butyl) decomposes largely to [N3N]MoH
(see above). As mentioned earlier, rapid and reversible C-C
bond cleavage in [N3N]Mo(cyclopentyl) and [N3N]Mo(cyclo-
hexyl) is an alternative way to explain interconversion of exo
and endo protons, but the relatively slow rate of C-C bond
cleavage in 2g,h would seem to preclude rapid C-C bond
cleavage in less strained five- and six-membered rings in
analogous compounds. Compound 3a, like high oxidation state
methylidyne complexes of the type W(CH)(OR)3(donor),24 is
relatively stable toward bimolecular decomposition reactions,
whereas W(CH)(OR)3 complexes are not.25
When [N3N]MoH is heated to 105 °C in toluene-d8 in an
evacuated, sealed NMR tube, a 50/50 mixture of [N3N]MoH
and [bitN3N] is observed after 22 h. Removal of the H2 above
this solution and further heating for another 22 h yields a 70/
30 mixture of [bitN3N] and [N3N]MoH. Heating for another
24 h gave a 80/20 mixture. [bitN3N]Mo reacts with D2 slowly
to give largely [N3N]MoD in which a second deuterium is
located in a TMS group, as shown by 2H NMR spectra, thereby
confirming that loss of hydrogen (eq 11) is reversible. However,
both loss of hydrogen from [N3N]MoH and reaction of [bitN3N]-
Mo with hydrogen are relatively slow. [bitN3N]Mo has mirror
symmetry, according to its 1H NMR spectra. Its susceptibility
versus temperature (Figure 7) is similar to other d2 complexes
discussed here, a fact that is consistent with magnetic behavior
that does not depend on any significant structural change. A
fit of these data to an equation17,18 involving g () 2), E () 0.1
cm-1), and D yielded a value of D ) 31 cm-1 (R ) 0.9993).
The most convenient laboratory synthesis of [bitN3N]Mo is
by reduction of [N3N]MoCl with excess magnesium powder in
Upon heating a solution of the cyclopentyl complex (2i),
cyclopentene is lost and yellow, paramagnetic [N3N]MoΗ (eq
10) is formed quantitatively, according to NMR spectra. It can
be isolated in ∼60% yield on a scale of ∼1 mmol. When [N3N]-
MoD (see below) is dissolved in neat cyclopentene and the
solution is heated to 50 °C for 18 h, no deuterium is incorporated
2
into the cyclopentene, according to H NMR acquired under
(25) Chisholm, M. H.; Folting, K.; Hoffman, D. M.; Huffman, J. C. J.
Am. Chem. Soc. 1984, 106, 6794.

11888 J. Am. Chem. Soc., Vol. 119, No. 49, 1997
Schrock et al.
Likewise, no [bitN3N]Mo is formed when [N3N]MoMe decom-
poses, as noted earlier.
Discussion
The phenomenon of â-hydride elimination to give an olefin
hydride has dominated transition metal alkyl chemistry com-
pared to R-elimination to give an alkylidene hydride.30 Evidence
for an equilibrium between an alkyl complex and alkylidene
hydride has usually involved high oxidation state tungsten31-33
or tantalum34-38 alkyl complexes. In the Cp*2Ta system
observable Cp*2Ta(CH2)(H) was found to be in equilibrium with
unobservable Cp*2Ta(CH3).37 It was also shown that
Cp*2(H)TadCdCH2 decomposed via intermediate Cp*(η5-C5-
Me4CH2CH2CH2)Ta to the kinetic product, Cp*(η5-C5Me4CH2-
CH2CH)Ta(H), by R-elimination, and to the thermodynamic
product, Cp*(η5-C5Me4CH2-η2-CHdCH2)Ta(H), by â-elimina-
tion.38 It was noted that “the rate of R-H elimination is 108
times greater than the rate of â-H elimination” in this circum-
stance. However, it was also noted that “the generality of faster
R-H elimination versus â-H elimination is questionable, since
the transition state for â-H elimination in this system is highly
strained, whereas that for R-H elimination is much less so.”
Finally, it was argued that “for a general alkyl ligand in the
permethyltantalocene system, R-H elimination cannot be kineti-
cally favored over â-H elimination by more than... a factor of
∼30 in rate at 25 °C.” Similar statements could be made about
many of the alkyl complexes reported here, although the validity
and generality of such statements are limited by the experimental
data available. It is generally assumed that an agostic interac-
tion20 between the metal and CHâ or CHR is on the pathway
for â- or R-elimination, respectively, even though definitive
evidence for R-agostic interactions in group VI d0 complexes
is still sparse.21
R-Abstraction in a dialkyl complex (to give alkane and an
alkylidene complex)23 is closely related to R-elimination, and
an R-agostic activation of an R-proton in one of the alkyls is
thought to precede R-abstraction. It has long been known that
neopentyl complexes are most prone to R-abstraction, with
(trimethylsilyl)methyl, benzyl, and especially methyl being
progressively less so.23 It was also noted in early studies that
R-abstraction is faster in sterically crowded environments, e.g.,
when additional ligands coordinate to the metal.23 Recently it
has been shown that R-abstraction and â-abstraction can
compete in [(Me3SiNCH2CH2N)3N]Ta(alkyl)2 species,39 that the
rate of R abstraction can be enhanced by increasing the size of
the alkyls in the apical “pocket”,40 and that R-abstraction to
give an alkylidene can be forced to be virtually the only
Figure 12. View of the structure of [bitN3N]Mo.
THF in an evacuated vessel over a period of 7 days. The
product is proposed to arise via loss of a hydrogen radical from
B (eq 12), which is proposed to be in equilibrium with the
“monopyramidal” complex A, the product of reducing [N3N]-
MoCl by one electron followed by loss of chloride. One
possible fate of the hydrogen radical is to combine with A to
give [N3N]MoH. In fact, [N3N]MoH is a significant impurity
in the reduction, but [bitN3N]Mo can be separated from [N3N]-
MoH by crystallization. Another possible fate of the hydrogen
radical is formation of dihydrogen, which we know reacts slowly
with [bitN3N]Mo to give [N3N]MoH. Therefore the precise
origin of [N3N]MoH in the reduction reaction is not known.
An X-ray study showed [bitN3N]Mo to have the structure
shown in eq 11 (Figure 12, Table 1). The Mo-C(11) bond is
somewhat longer than the Mo-Lax bonds in [N3N]MoMe (2a)
and [N3N]Mo(cyclohexyl) (2j), but the M-Neq bond lengths
and Neq-Mo-Neq and Mo-Neq-Si (except to Si(1)) bond
angles are remarkably close to those found in 2a,j. The Mo-
N(4) bond length is significantly shorter than that in 2a,j, but
bond lengths and angles in general are consistent with a
molecule whose coordination sphere is relatively uncrowded
and undistorted. The absence of distortion in the MoN4 core is
somewhat surprising in view of the presence of the Mo-C-
Si-N ring and relatively small N(4)-Mo-C(11) angle (155.16-
(8)°). Although C-H activation in TMS amido complexes is
relatively well-known,26-28 X-ray structures of complexes that
contain a MNSiC ring are rare. One such species is Zr[CH2-
SiMe2N(SiMe3)]2(dmpe).29 It contains two planar four-
membered metallacyclic rings analogous to that found in
[bitN3N]Mo (C-Zr-N ) 76°; Zr-N-Si - 97°; N-Si-C -
99°; Si-C-Zr ) 87°).
(30) 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.
(31) Cooper, N. J.; Green, M. L. H. J. Chem. Soc., Chem. Commun.
1974, 209.
(32) Cooper, N. J.; Green, M. L. H. J. Chem. Soc., Chem. Commun.
1974, 761.
(33) Cooper, N. J.; Green, M. L. H. J. Chem. Soc., Dalton Trans. 1979,
1121.
(34) Turner, H. W.; Schrock, R. R. J. Am. Chem. Soc. 1982, 104, 2331.
(35) Turner, H. W.; Schrock, R. R.; Fellmann, J. D.; Holmes, S. J. J.
Am. Chem. Soc. 1983, 105, 4942.
(36) Fellmann, J. D.; Schrock, R. R.; Traficante, D. D. Organometallics
1982, 1, 481.
(37) van Asselt, A.; Burger, B. J.; Gibson, V. C.; Bercaw, J. E. J. Am.
Chem. Soc. 1986, 108, 5347.
(38) Parkin, G.; Bunel, E.; Burger, B. J.; Trimmer, M. S.; Asselt, A. v.;
Bercaw, J. E. J. Mol. Catal. 1987, 41, 21.
[bitN3N]Mo is not formed upon heating [N3N]MoR com-
plexes that are relatively stable toward decomposition by other
pathways. For example, [N3N]MoPh is quite stable thermally
(decomposition is slow at 120 °C), and there is no evidence
that [bitN3N]Mo forms over a period of 24 h at 120 °C.
(26) Simpson, S. J.; Andersen, R. A. Inorg. Chem. 1981, 20, 2991.
(27) Turner, H. W.; Simpson, S. J.; Andersen, R. A. J. Am. Chem. Soc.
1979, 101, 2782.
(39) Freundlich, J. S.; Schrock, R. R.; Davis, W. M. J. Am. Chem. Soc.
1996, 118, 3643.
(28) Simpson, S. J.; Andersen, R. A. Inorg. Chem. 1981, 20, 3627.
(29) Planalp, R. P.; Andersen, R. A. Organometallics 1983, 2, 1675.
(40) Fruendlich, J. S.; Schrock, R. R.; Davis, W. M. Organometallics
1996, 15, 2777.

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11889
sterically tenable process by increasing the size of the silyl
substituent in [(R3SiNCH2CH2N)3N]Ta(alkyl)2 species from R
) Me to R ) Et. In a more crowded “pocket”, the â-agostic
interaction that precedes â-abstraction in the alkyl becomes
sterically disfavored, while the R-agostic interaction that
precedes R-abstraction at the same time is encouraged.
The preceding discussion would suggest that R-processes
(abstraction or elimination) can be dramatically accelerated at
the expense of â-processes (abstraction or elimination) and that
steric (or “conformational”) factors play a key role in determin-
ing the relative rates of R- versus â-processes in a given
complex. Therefore, it is becoming increasingly clear that broad
statements concerning the relative rates of R- versus â-processes
simply cannot be justified, even for a given metal with the same
ligand coordination sphere. Nevertheless, it seems relatively
safe to say that competition between R- and â-processes will
continue to be found among the earlier, heavier metals (Ta, Mo,
W, Re) where multiple metal-carbon bonds (alkylidenes and
alkylidynes) are relatively common.23 Whether R-elimination
is faster than â-elimination for later metals (e.g., iridium41) has
yet to be investigated thoroughly.
We have observed that R-elimination is fast relative to
â-elimination in two [N3N]Mo(cycloalkyl) systems, even though
no alkylidene hydride complex is observed in either case. To
our knowledge any consequence of rapid and reversible
R-elimination has never been observed directly before, perhaps
largely because no net change in the alkyl itself occurs. Rapid
R-elimination could be documented here in part because of the
paramagnetic nature of the [N3N]Mo(cycloalkyl) complexes and
the resulting dispersion of all cycloalkyl resonances (including
exo and endo resonances) in NMR spectra, a phenomenon that
has proved useful in the coordination chemistry of paramagnetic
coordination complexes for some time.22 We suspect that
R-elimination in some [N3N]Mo(n-alkyl) complexes can also
be “rapid” (e.g., 10-1 s-1 or greater), but perhaps only when
the Mo-CR-Câ angle is opened up as a consequence of steric
interactions (e.g., in [N3N]Mo(neopentyl)). Unfortunately, we
have not been able to think of a way to measure the rate of
conversion of, for example, [N3N]Mo(CH2CMe3) to [N3N]Mo-
(CHCMe3)(H) to compare the rate of R-elimination with that
found in [N3N]Mo(cyclopentyl) and [N3N]Mo(cyclohexyl). For
alkyls in general the situation is complicated by competitive
R-abstraction reaction (R,R-dehydrogenation) and â-hydride
elimination processes. A plausible scenario is that the rate of
R-elimination in [N3N]Mo(CH2CMe3) could be of the same
order of magnitude as it is in [N3N]Mo(cyclopentyl) or [N3N]-
Mo(cyclohexyl) (i.e., 10 to 103 s-1 at 22 °C), but [N3N]Mo-
(CCMe3) does not form readily because [N3N]Mo(CHCMe3)(H)
does not lose hydrogen rapidly enough to compete with the back
reaction to reform [N3N]Mo(CH2CMe3). Therefore a plausible
explanation as to why other [N3N]Mo(CH2R) complexes
decompose to [N3N]MotCR complexes progressively less
readily as the size of R decreases is that the equilibrium between
[N3N]Mo(H)(CHR) and [N3N]Mo(CH2R) complexes becomes
prohibitively large, i.e., virtually no [N3N]Mo(H)(CHR) is
present. The equilibrium between [N3N]Mo(H)(CHR) and
[N3N]Mo(CH2R) could become large either as a consequence
of a precipitous drop in the rate of R-elimination or as a
consequence of a dramatic increase of the rate of forming an
alkyl complex from an alkylidene hydride complex. If it is the
former, then R-elimination could even become rate limiting in
the reaction in which an alkylidyne complex is formed from an
alkyl complex, e.g., R,R-dehydrogenation in [N3N]MoMe to
form [N3N]MotCH. It is interesting to note that the overall
relative ease of forming an alkylidyne complex via loss of H2
from the alkyl (CH2CMe3 > CH2SiMe3 > CH2Ph . Me) is
analogous to what has been observed qualitatively for R-hy-
drogen abstraction in tantalum dialkyl complexes to give
alkylidene complexes.23 Data for formation of alkylidyne
complexes from alkylidene complexes24 by R-hydrogen abstrac-
tion is more sketchy but in accord with a general trend toward
more facile R-abstraction from bulkier alkylidenes.
The X-ray structures of the methyl and cyclohexyl [N3N]-
MoR species illustrate semiquantitatively the differences in the
degree of steric congestion within the trigonal pocket. These
differences suggest that a greater rate of R-elimination in part
could be traced to a relief of steric strain within the pocket.
However, so many other factors must be taken into account
that we can only in a very general sense feel secure in
concluding that steric congestion increases the rate of R-elim-
ination. In fact, the rate of R-elimination in the cyclohexyl
complex is significantly slower than the rate of R-elimination
in the cyclopentyl complex at 22 °C, even though one would
expect steric congestion in the ground state of the cyclohexyl
species to be greater. However, a greater degree of steric
congestion in [N3N]Mo(cyclohexylidene)(H) (versus [N3N]Mo-
(cyclopentylidene)(H)) should oppose that trend and lead to a
slower rate of R-elimination. Correlating the rates of â-elimina-
tion with steric congestion in either the ground state or in the
olefin hydride intermediate species in these systems is also
virtually impossible for similar reasons at this stage.
In contrast to the equilibria observed here, the equilibrium
between [N3N]W(cycloalkyl) and [N3N]W(cycloalkylidene)(H)
complexes has been observed to lie toward the latter for
cyclopentyl13 and cyclohexyl.14 However, [N3N]WMe is ob-
servable.12 These results are consistent with a substantially
slower rate of R-elimination in the methyl complex and/or a
relatively small equilibrium constant for forming [N3N]W(CH2)-
(H). The observation of tungsten(VI) alkylidene hydride
complexes, but molybdenum(IV) alkyl complexes, could be
explained using oxidation state and/or bond strength arguments.
The fact that R,R-dehydrogenation reactions in [N3N]W(CH2R)
complexes are orders of magnitude faster than in the analogous
[N3N]Mo(CH2R) complexes also could be explained in terms
of a higher concentration of alkylidene hydride in a W system
versus an analogous Mo system. A more detailed discussion
and relevant results may be found in a paper concerning
[N3N]W(alkyl) chemistry.14
It may be worth pointing out that we have assumed
throughout this work that the [N3N] ligand is not directly
involved in proton migration reactions, i.e., a proton does not
migrate from carbon to nitrogen to produce a [(Me3SiNCH2-
CH2)2N(CH2CH2NHSiMe3)]2- ligand. However, such a ligand
has been observed in the analogous C6F5-substituted ligand
system.42 Therefore the possibility of forming an intermediate
via proton migration to an amido nitrogen (e.g., eq 13) must be
considered in a totally objective analysis. There is no change
in oxidation state of the metal in such a reaction if we assume
(41) Burk, M. J.; McGrath, M. P.; Crabtree, R. H. J. Am. Chem. Soc.
1988, 110, 620.
(42) Rosenberger, C.; Schrock, R. R.; Davis, W. M. Inorg. Chem. 1997,
36, 123.

11890 J. Am. Chem. Soc., Vol. 119, No. 49, 1997
Schrock et al.
that the alkylidene is a dianion. A rapid process of this nature
could explain interconversion of exo and endo protons in [N3N]-
Mo(cyclopentyl), if the alkylidene could rotate readily by 180°
before the amido proton migrates back to the alkylidene carbon
atom.
The carbon-carbon bond cleavage reactions that have been
observed here for [N3N]Mo(cyclopropyl) and [N3N]Mo-
(cyclobutyl) are surprisingly facile alternatives to CH bond
cleavage reactions and, in fact, take place more readily than
the overall rate of R,R-dehydrogenation in [N3N]Mo(neopentyl).
Carbon-carbon bond cleavage in cyclopropane rings that are
external to the coordination sphere has much precedent in the
literature,30 and in some cases, C-C bond cleavage in a
cyclopropyl ligand has been observed.43 However, probably
for thermodynamic reasons,30 C-C bond cleavage in hydro-
carbons or hydrocarbyl ligands with less π character in the C-C
bond is rare.44,45 We could find no example in the literature of
C-C cleavage of a cyclobutyl ligand by a transition metal.
Surprisingly, both the cyclopropyl and cyclobutyl systems
manage to find a pathway to alkylidyne complexes. We
speculate that a highly strained 1-metallacyclobutene complex
(A) is the result of C-C cleavage in the cyclopropyl complex
(eq 14). A also could be formed via reaction of the methylidyne
It is perhaps largely a consequence of the relative stability
of the [N3N]3- ligand system to various other decomposition
reactions, especially intermolecular decomposition reactions and
reactions involving ligand dissociation, that such a variety of
intramolecular reactions can be observed. However, it is also
somewhat surprising that the trimethylsilyl substituents do not
undergo CH activation more readily than they do (e.g., to form
[bitN3N]Mo), as carbon-hydrogen bond activation in (usually
hydride) complexes that contain one or more trimethylsilyl-
substituted ligands has been observed in several systems, 26-28
including [N3N]TiH.3 Other possible decomposition reactions,
e.g., loss of Me3SiH39 or degradation of the tren backbone by
cleavage of an Nax-C bond,39,40 so far have not been identified
in [N3N]Mo alkyl complexes.
A potentially important feature of the alkyl systems studied
here is the fact that they are all 16-electron, high-spin d2 species.
Many of the reactions discussed here, R-hydride elimination
in particular, would seem to require an empty orbital in order
to activate a C-H bond through an agostic interaction.
However, so far there is no evidence that the rate of any of the
reactions discussed here is limited as a consequence of the high
spin ground-state alone, a possibility that was proposed in a
preliminary communication.13 Arguments concerning the de-
gree to which spin state alone can alter the rate of an
organometallic reaction have been in the literature for some
time,30,46-49 and the conclusion in every case so far has been
that it cannot. However, since the vast majority of examples
in the literature concern metals in lower oxidation states (usually
carbonyl complexes), we will keep an open mind concerning
the possibility that accessibility of some 1A spin state (with an
accompanying structural change) is an important feature of some
of the chemistry of d2 complexes of the type described here.
complex with ethylene, but that hypothetical reaction is still
unknown for any high oxidation state alkylidyne complex.24
Therefore it would not be surprising to find that A would eject
ethylene to form the methylidyne complex. It is not known
whether reversible C-Hâ bond cleavage competes with C-C
bond cleavage reactions in the cyclopropyl complex, as an
appropriate 13CR labeling experiment has not been carried out.
How a butylidyne complex forms from a cyclobutyl complex
is more speculative. The reaction between [N3N]WCl and
cyclobutyllithium has been shown to yield the 1-tungstacyclo-
pentene complex shown in eq 15,14 and this species has been
shown to rearrange cleanly to the butylidyne complex by what
could be viewed as an R-hydrogen abstraction from an alky-
lidene ligand to give an alkylidyne ligand. Therefore we assume
that rearrangement of the molybdenum cyclobutyl complex
follows a similar pathway. However, there is no evidence that
suggests that the [N3N]Mo(cyclobutyl) complex interconverts
rapidly and reversibly with the 1-molybdacyclopentene complex
(eq 16), with the equilibrium lying well toward the cyclobutyl
complex, and indeed, one would not expect C-C bonds to be
cleaved and reformed nearly as readily as C-H bonds. The
equilibrium shown in eq 16 could be observed in the NMR
spectrum of [N3N]Mo(cyclobutyl) if it were fast enough, as endo
and exo protons would exchange in the process. Although we
did not observe rapid exo/endo exchange of cyclobutyl protons
in [N3N]Mo(cyclobutyl), we cannot exclude the possibility,
however unlikely it might be, that the equilibrium between
[N3N]M(cyclobutyl) and the 1-metallacyclopentene complex (eq
16) is fast on the chemical time scale, lies on the side of
[N3N]M(cyclobutyl) when M ) Mo, and lies on the side of the
1-metallacyclopentene complex when M ) W.
Experimental Section
General Details. All experiments were conducted under nitrogen
in a Vacuum Atmospheres drybox, using standard Schlenk techniques,
or on a high-vacuum line (<10-4 Torr). Pentane was washed with
HNO3/H2SO4 (5/95 v/v), sodium bicarbonate, and H2O, stored over
CaCl2, and then distilled from sodium benzophenone under nitrogen.
Regent grade ether, tetrahydrofuran, and benzene were distilled from
sodium benzophenone under nitrogen. Toluene was distilled from
molten sodium. Methylene chloride was distilled from CaH2. All
solvents were stored in the drybox over activated 4 Å molecular sieves.
(46) Detrich, J. L.; Reinaud, O. M.; Rheingold, A. L.; Theopold, K. H.
J. Am. Chem. Soc. 1995, 117, 11745.
(47) Gadd, G. E.; Upmacis, R. K.; Poliakoff, M.; Turner, J. J. J. Am.
Chem. Soc. 1986, 108, 2547.
(48) Janowicz, A. H.; Bryndza, H. E.; Bergman, R. G. J. Am. Chem.
Soc. 1981, 103, 1516.
(49) Parshall, G. W. Catalysis. A Specialist Periodical Report; The
Chemical Society: London, 1977; Vol. 1, p 335.
(43) Periana, R. A.; Bergman, R. G. J. Am. Chem. Soc. 1984, 106, 7272.
(44) Halpern, J. Acc. Chem. Res, 1970, 3, 386.
(45) Bishop, K. C. Chem. ReV. 1976, 76, 461.

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11891
Deuterated solvents were freeze-pump-thaw degassed and vacuum
transferred from an appropriate drying agent, or sparged with argon
and stored over sieves. NMR spectra are recorded in C6D6 unless noted
otherwise. 1H and 13C data are listed in parts per million downfield
from tetramethylsilane and were referenced using the residual protonated
solvent peak. 19F NMR are listed in parts per million downfield of
CFCl3 as an external standard. 2H NMR spectra usually were obtained
at 76.7 MHz and are referenced to external C6D6 (7.15 ppm). Coupling
constants are given in hertz, and routine couplings are not listed.
Elemental analyses (C, H, N) were performed on a Perkin-Elmer 2400
CHN analyzer in our own laboratory.
All organic compounds were received from commercial suppliers
and were used as received. MoCl4(THF)2,50 MoCl3(THF)3,50 cyclo-
hexyllithium,51 and ethyllithium52 were prepared according to published
procedures. Benzyllithium was prepared from tribenzyltin chloride and
methyllithium by the literature procedure.53 Cyclopentyllithium and
cyclobutyllithium were prepared from the chloride or bromide and
lithium powder in pentane. Cyclopropyllithium was prepared from the
bromide and lithium powder in diethyl ether. Perdeuteriocyclohexyl
chloride was prepared from commercially available C6D11OH and DCl
as described in the literature.54 1-Deuteriocyclopentyllithium was
prepared by reduction of cyclopentanone with lithium aluminum
deuteride followed by treatment with Ph3PBr2 in DMF.55 The bromide
was converted to the lithium reagent by reaction with lithium powder
in refluxing hexane. The lithium reagent was recrystallized from
pentane before use. Li13CH2CH2CH2CH3 was prepared by treating
13CO2 with n-PrMgCl, reducing the acid with LiAlH4 to the alcohol,
chlorinating the alcohol with SOCl2 in hexane, and finally treating the
chloride with lithium wire in the usual manner; overall yield ∼20%.
In UV/vis studies the following equation was employed to follow
the disappearance of an alkyl species: ln[(λo - λ∞)/(λ - λ∞)] ) kt,
where λo ) absorbance at wavelength λ at time 0, λ∞ ) absorbance at
wavelength λ at infinite time, and λ ) absorbance at wavelength λ at
time t.
[N3N]Mo(CH3) (2a). A solution of 1a (0.20 g, 0.41 mmol) in ∼25
mL of THF was treated with LiMe (291 µL, 0.41 mmol, 1.4 M in
ether) at 25 °C. The reaction mixture was stirred for 3 h, and the solvent
was removed in vacuo. Pentane (∼30 mL) was added to the reaction
residue, and the extract was filtered through Celite to remove LiCl.
The volume of the filtrate was then reduced to ∼3 mL in vacuo, and
the concentrated solution was chilled to -40° to yield red crystals of
the product. The product was collected by decantation of the mother
liquor and dried in vacuo: yield 0.17 g in two crops (91%); 1H NMR
δ +7.77 (SiMe3), -25.00 (NCH2), -83.24 (NCH2); λmax ) 486 nm in
toluene, ꢀ ) 680 M-1 cm-1; µ ) 2.9 µB at 25 °C (Evan’s method).
Anal. Calcd for C16H42N4MoSi3: C, 40.82; H, 8.99, N, 11.90.
Found: C, 40.87; H, 9.17; N, 12.11.
The chemical shifts for the methylene groups in [N3N]MoMe (in
toluene-d8) are as follows (T (K), shift 1, shift 2): 363.16, -19.76,
-62.29; 353.16, -20.57, -64.82; 343.16, -21.40, -67.52; 333.16,
-22.23, -70.40; 323.16, -23.12, -73.58; 313.16, -24.06, -76.93;
303.16, -24.97, -80.34; 293.16, -26.11, -84.79; 283.16, -27.16,
-89.27; 273.16, -28.18, -94.25; 263.16, -29.45, -99.52; 253.16,
-30.67, -105.54; 243.16, -32.15, -112.24. These data were used
to generate the plots shown in Figure 8b.
[N3N]Mo(CH2CH3) (2b). Ethyllithium (21 mg) was added to a
solution of [N3N]MoCl (134 mg, 0.27 mmol) in 10 mL of ether. The
color changed immediately to deep purple. The reaction mixture was
stirred for 2 h, after which the solvents were removed in vacuo and
the resulting solid was extracted with a minimum amount of pentane.
The solution was cooled to - 40 °C to yield purple crystals: yield
1
100 mg (76%); H NMR (toluene-d8) 9.3 (SiMe3), -25.05 (NCH2),
-51.5 (ethyl CH3), -60.0 (NCH2).
[N3N]Mo(CH2CH2CH2CH3) (2c). Compound 2c was synthesized
as described for 2b from n-butyllithium in hexane (1.1 equiv) and [N3N]-
MoCl (120 mg, 0.24 mmol) in 5 mL of ether; yield 88 mg (70%): 1H
NMR (toluene-d8) 9.3 (SiMe3), -26 (NCH2), - 47 (butyl CH2), -59
(NCH2).
[N3N]MoCl (1a). Method (a). MoCl4(THF)2 (2.53 g, 6.63 mmol)
was added to a THF (∼150 mL) solution of Li3[N3N] (3.50 g, 9.13
mmol), and the reaction mixture was stirred for 6 h. The resulting
dark-brown solution was evaporated to dryness, and the residue was
repeatedly extracted with pentane (500 mL). The pentane extract was
filtered through Celite, and the volume was reduced in vacuo until
orange crystals formed. The mixture was then chilled at -40 °C for
1 h, and the orange crystals were collected by filtration and washed
quickly with cold pentane (2 × 10 mL). The crude product was
recrystallized from diethyl ether to give bright orange crystals: yield
1.34 g (41%); 1H NMR δ 5.73 (br s), -23.6 (br s), -99.2 (br s). Anal.
Calcd for C15H39N4ClMoSi3: C, 36.68; H, 8.00; N, 11.41. Found: C,
36.63; H, 7.97; N, 10.90.
Method (b). A mixture of MoCl3(THF)3 (1.10 g, 2.63 mmol) and
Li3[N3N] (1.10 g, 2.63 mmol) in 80 mL of diethyl ether was stirred at
room temperature for 4 h. The reaction mixture was filtered and the
solid washed with 10 mL of diethyl ether. The volume of the filtrate
was reduced to ∼5 mL in vacuo, and 10 mL of pentane was added.
The solution was chilled to -40 °C overnight, and the orange-brown
solid was filtered off and recrystallized twice from a mixture of diethyl
ether and pentane: yield 0.31 g (24%).
[N3N]Mo(CH2Ph) (2d). Compound 2d was synthesized as described
for 2b from 80 mg (0.16 mmol) of [N3N]MoCl and 38 mg of
LiCH2Ph(ether)1.25: yield 84 mg (94%); H NMR (toluene-d8) 10.92
(s, 1, Hp), 10.53 (s, 2, Hm), 9.18 (SiMe3), -24.8 (NCH2), - 63 (NCH2);
λmax ) 504 nm in toluene.
1
[N3N]Mo(CH2SiMe3) (2e). Compound 2e was prepared as described
for 2a from 108 mg (0.21 mmol) of [N3N]MoCl and 34 mg (0.36 mmol)
of LiCH2TMS: yield 91 mg (76%); 1H NMR (toluene-d8) 8.5 (TMS),
1.28 (TMS), - 20.1 (NCH2), -34.7 (NCH2); λmax(1) ) 496 nm, λmax(2)
) 640 nm in toluene or THF. Anal. Calcd for C19H50N4Si4Mo: C,
42.03; H, 9.28, N, 10.32. Found: C, 42.05; H, 9.56; N, 10.20.
[N3N]Mo(CH2CMe3) (2f). This complex was prepared from 1a
(0.50 g, 1.02 mmol) in THF (∼50 mL) and LiCH2CMe3 (0.12 g; 1.52
mmol) at 25 °C and was isolated as described for 2a: yield 0.35 g
1
(65%); H NMR 8.8 (SiMe3), -0.5 (s, t-Bu), -15.3 (NCH2), -20.6
(NCH2); λmax(1) ) 480 nm, λmax(2) ) 688 nm in toluene. Anal. Calcd
for C20H50N4MoSi3: C, 45.60; H, 9.57, N, 10.63. Found: C, 45.51;
H, 9.68; N, 10.69.
The conversion of [N3N]Mo(CH2CMe3) into [N3N]MotCCMe3 was
followed by visible spectroscopy in toluene at 480 and 688 nm through
at least three half-lives. (Identical results were obtained at each
wavelength.) The rate constants (in units of 10-4 s-1) obtained at
temperature T (in K) were 0.50 (363), 1.70 (374), 4.38 (385), and 9.41
(394). A plot of ln(k/T) versus 1/T gave ∆Hq ) 25 973 cal mol-1 and
∆Sq ) -6.94 eu with an R value of 0.9988. The exact values were
used to calculate k at temperature T.
[N3N]Mo(OTf) (1b). [N3N]MoCl (0.50 g, 1.02 mmol) and Cp2-
FeOTf (0.375 g, 1.12 mmol) were added to 25 mL of dichloromethane,
and the resulting solution was stirred for 6 h at 22 °C. The volume
was reduced by one-third in vacuo, and 40 mL of diethyl ether was
added. The mixture was chilled to -40 °C for several hours, and pink
1
crystals were filtered off and dried in vacuo: yield 0.38 g (62%); H
[N3N]Mo(cyclopropyl) (2 g). This compound was prepared from
126 mg (0.26 mmol) of [N3N]MoCl in 10 mL of diethyl ether and a
-10 °C solution of 64 mg (0.38 mmol) of cyclopropyllithium in 5 mL
of diethyl ether. The reaction mixture was stirred at -10 °C overnight,
during which the color changed from orange to deep purple. The
solvent was evaporated, and the residue was extracted with cold (-40
°C) pentane. The volume of the filtrate was reduced to ∼1 mL in
vacuo, and the solution was stored at -40 °C. After 2 days purple
crystals were collected: yield 70 mg (55%); 1H NMR δ 21 (cyclopropyl
CH2), 10.9 (SiMe3), -27.9 (NCH2), -47.6 (NCH2), -93.8 (cyclopropyl
CH); λmax ) 540 nm in toluene. Anal. Calcd for C18H44N4MoSi3: C,
43.52; H, 8.93; N, 11.28. Found: C, 43.58; H, 9.11; N, 11.15.
NMR δ 11.0 (s, TMS), -42 (br s), -60 (br s). Anal. Calcd for
C16H39N4F3MoO3SSi3: C, 31.78; H, 6.50, N, 9.26. Found: C, 31.78;
H, 6.28; N, 9.78.
(50) Dilworth, J. R.; Richards, R. L. Inorg. Synth. 1990, 28, 33.
(51) Zerger, R.; Rhine, W.; Stucky, G. J. Am. Chem. Soc. 1974, 96, 6048.
(52) Bryce-Smith, D.; Turner, E. E. J. Chem. Soc. 1953, 861.
(53) Seyferth, D.; Suzuki, R.; Murphy, C. J.; Sabet, C. R. J. Organomet.
Chem. 1964, 2, 431.
(54) Perlman, D.; Davidson, D.; Bogert, M. T. J. Org. Chem. 1936, 1,
288.
(55) Wiley: G. A.; Hershkowitz, R. L.; Rein, B. M.; Chung, B. C. J.
Am. Chem. Soc. 1964, 86, 964.

11892 J. Am. Chem. Soc., Vol. 119, No. 49, 1997
Schrock et al.
Decomposition of [N3N]Mo(cyclopropyl) was followed in toluene
by UV/vis spectroscopy at 540 nm; the values for k (×10-4 s-1) at
temperature T (K) were 0.608 (313), 1.88 (323), 4.99 (333), 14.0 (343).
A plot of ln(k/T) versus 1/T gave ∆Hq ) 21 506 cal/mol and ∆Sq )
-9.20 eu with an R value of 0.9998. In THF the values for k (×10-4
s-1) at temperature T (K) were 0.147 (303), 0.490 (313), 0.838 (318),
2.44 (328), 3.98 (333). A plot of ln(k/T) versus 1/T gave ∆Hq ) 21 400
cal mol-1 and ∆Sq ) -10.02 eu with an R value of 0.9999.
variable-T NMR as shown in Figure 11b and described in the text.
Values for k are 606 s-1 at 331 K, 1273 s-1 at 346 K, and 1555 s-1 at
352 K. Virtually identical values were obtained for a sample that was
one-half the concentration (0.05 M). By performing a similar experi-
ment at 46.02 MHz three different points were obtained: 378 s-1 at
324, 755 s-1 at 339 K, and 955 s-1 at 343 K. A plot ln(k/T) versus 1/T
for these six points gave ∆Hq ) 10 588 cal mol-1 and ∆Sq ) -14.19
eu with R ) 0.9880. Using these values, k at 337 K was calculated to
be 696 s-1. From the data shown in Figure 11a the rate constant for
R-elimination in [N3N]Mo(C6H11) was found to be 2155 s-1 at 337 K.
The isotope effect at 337 K is therefore kH/kD ) 3.1(2).
[N3N]Mo(phenyl) (2k). Phenyllithium (10 mg, 0.1 mmol) was
added to a cold solution of 40 mg (0.08 mmol) of [N3N]MoCl in 5 mL
of diethyl ether. The solution was stirred overnight. The solvents were
removed from the reaction mixture in vacuo, and the residue was
extracted with pentane. The volume of the pentane extract was reduced
in vacuo to 2 mL, and after 2 h at -30 °C, 30 mg (70%) of orange
crystals was collected: 1H NMR δ 64.2 (br s, Ph), 10.8 (SiMe3), -23.9
(NCH2), -67.0 (NCH2). Anal. Calcd for C21H44N4MoSi3: C, 47.34;
H, 8.32; N, 10.52. Found: C, 47.55; H, 8.18; N, 10.58.
[N3N]Mo(cyclobutyl) (2b). Cyclobutyllithium (14 mg, 0.23 mmol)
was added to a solution of 93 mg (0.19 mmol) of N3NMoCl in 10 mL
of diethyl ether at -40 °C. The solution turned purple immediately.
The solution was warmed to room temperature and was stirred for 1 h.
The solvent was evaporated, and the residue was extracted with pentane
and the extract was filtered. The volume of the extract was reduced to
∼1 mL in vacuo. Purple crystals were obtained after standing the
1
solution overnight at - 40 °C: yield 30 mg (31%); H NMR at -10
°C δ 33.9 (cyclobutyl), 13.1 (SiMe3), -29.3 (NCH2), -39.6 (br s,
cyclobutyl CH2), -40.4 (cyclobutyl), -107.8 (cyclobutyl); λmax ) 558
nm in toluene. Anal. Calcd for C19H46N4MoSi3: C, 44.68; H, 9.08;
N, 10.97. Found: C, 44.40; H, 8.68; N, 10.79.
Conversion of [N3N]Mo(cyclobutyl) into [N3N]MotCCH2CH2CH3
was followed in toluene by UV/vis spectroscopy at 558 nm; the values
for k (×10-4 s-1) at temperature T (K) were 0.110 (303), 0.443 (313),
1.54 (323), 5.03 (333), 13.8 (343). A plot of ln(k/T) versus 1/T gave
∆Hq ) 24 368 cal/mol and ∆Sq ) -0.07 eu with R ) 0.9998. In
THF the values for k (×10-4 s-1) at temperature T (K) were 0.349
(313) and 3.98 (333).
[N3N]Mo(cyclopentyl) (2i). A solution of cyclopentyllithium (155
mg, 2.02 mmol) dissolved in 10 mL of pentane was added to a solution
of 1a (498 mg, 1.01 mmol) in 25 mL of ether in a 100 mL flask. The
solution changed color immediately to purple. After 1 h at room
temperature, the volatile materials were removed in vacuo and the
residue was extracted with minimum (∼5 mL) pentane. This extract
was concentrated to ∼2 mL and cooled to -40 °C to give purple
crystals (250 mg, 47%): 1H NMR δ 21.2 (cyclopentyl), 15.3 (cyclo-
pentyl), 10.2 (SiMe3), -24.0 (NCH2), -35.6 (NCH2), -48.8 (cyclo-
pentyl), -58.8 (cyclopentyl); λmax ) 562 nm in toluene. Adequate
analytical data could not be obtained, we believe as a consequence to
the facile decomposition of this compound to [N3N]MoH.
The conversion was followed in toluene by UV/vis at 562 nm. The
values for k (×10-4 s-1) at temperature T (K) are 0.363 (313), 1.188
(323), 3.197 (333), and 8.465 (343). A plot of ln(k/T) versus 1/T gave
∆Hq ) 21 631 cal/mol and ∆Sq ) -9.78 eu with R ) 0.9997.
[N3N]Mo(1-cyclopentenyl) (2j). Cyclopentenyllithium (446 µL of
a 1.7 M solution, 0.759 mmol) was added to a solution of 250 mg
(0.506 mmol) of [N3N]MoCl in 40 mL of ether. After 1 h, the volatile
components were removed in vacuo and the yellow solid was extracted
with 10 mL of pentane. The extract was filtered through Celite and
cooled to -40 °C to yield 175 mg (0.334 mmol, 66%) of deep red
crystals: 1H NMR δ 18.5 (s, 2, cyclopentenyl CH2), 11.1 (s, 27, TMS),
-24.8 (s, 6, NCH2), -60.5 (s, 6, NCH2). Anal. Calcd for C20H46N4-
Si3Mo: C, 45.95; H, 8.84; N, 10.72; Found: C, 45.90; H, 8.78; N,
10.75.
[N3N]Mo(cyclohexyl). Cyclohexyllithium (19 mg (0.21 mmol) was
added to a solution of 92 mg (0.19 mmol) of 1a in 10 mL of ether.
The solution turned deep purple immediately. After 1 h, the solvent
was removed in vacuo and the residue was extracted with pentane.
Cooling the extract to -40 °C afforded 81 mg (80%) of purple
crystals: 1H NMR δ 15.58 (cyclohexyl CH2), 10.04 (SiMe3), 6.14
(cyclohexyl CH2), 1.70 (cyclohexyl CH2), -1.73 (cyclohexyl CH2),
-21.70 (NCH2), -34.97 (NCH2), -39.81 (cyclohexyl CH2), -48.51
(cyclohexyl CH2).
[N3N]MotCH (3a). A solution of [N3N]Mo(cyclopropyl) in ether
was stirred at room temperature for 16 h. The color changed from
deep purple to light brown. The solvent was evaporated, and the
resulting powder was dissolved in a minimum amount of pentane.
1
Crystallization at -30 °C gave pale yellow crystals: yield 75%; H
NMR 5.54 (CH), 3.37 (NCH2), 2.15 (NCH2), 0.48 (SiMe3); 13C NMR
(toluene-d8) δ 284.3 (CH), 53.2 (CH2), 52.4 (CH2), 4.5 (TMS). Anal.
Calcd for C16H40N4MoSi3: C, 41.00; H, 8.60; N, 11.95. Found: C,
41.26; H, 8.65; N, 12.16.
[N3N]MotCPr (3c). [N3N]Mo(cyclobutyl) (124 mg, 0.243 mmol)
was dissolved in 5 mL of toluene, and the solution was heated in a
sealed tube to 60 °C for 2 h. The toluene was removed in vacuo, and
the residue was dissolved in a minimum amount of pentane. The
pentane solution was cooled to -40° to give brown crystals of the
product after 24 h: yield 107 mg in two crops (86%); 1H NMR δ 3.50
(m, 2, CH2), 3.41 (t, 6, NCH2N), 2.13 (t, 6, NCH2N), 2.05 (m, 2, CH2),
0.86 (t, 3, CH3), 0.50 (s, 27, TMS); 13C{1H} NMR δ 4.6 (TMS), 14.6
(CH3), 23.4 (CH2), 52.5 (NCH2N), 53.8 (CH2), 298.3 (MotC). Anal.
Calcd for C19H46N4Si3Mo: C, 44.68; H, 9.08; N, 10.97. Found: C,
44.39; H, 9.16; N, 11.06.
[N3N]MotCCMe3 (3f). Compound 3f was identified in solution
as the sole product of decomposition of [N3N]Mo(CH2CMe3) (vide
supra): 1H NMR (toluene-d8) 3.2 (NCH2), 2.18 (NCH2), 1.57 (t-Bu),
0.37 (TMS); 13C NMR (toluene-d8) 316.08 (C-t-Bu), 56.88 (NCH2),
54.16 (CCMe3), 51.56 (NCH2), 32.74 (t-Bu), 3.84 (TMS).
[N3N]MotCMe (3b), [N3N]MotCPh (3d), and [N3N]MotCSiMe3
(3e) were observed only in solution by proton NMR. Although their
identity does not seem to be in doubt, proof in the form of a 13C NMR
spectrum was not practical for experimental reasons, primarily the low
percentage of alkylidyne complex in the product mixture.
[N3N]MoH. A solution of [N3N]Mo(C5H9) (150 mg, 0.286 mmol)
in 10 mL of benzene was placed in a 100 mL Pyrex vessel which was
sealed and heated at 45 °C for 24 h. The volatile materials were
removed in vacuo, and the resulting yellow crystalline material was
extracted with minimum pentane (∼5 mL). The extract was filtered
through a 1 cm plug of Celite, and the volume of the solution was
reduced by one-half in vacuo. The solution was stored at -35 °C for
1 day to produce yellow-orange needles: yield 75 mg (0.164 mmol,
58%); 1H NMR δ 15.3 (SiMe3, w1/2 ) 67), -16.9 (NCH2, w1/2 ) 306),
-117.6 (NCH2, w1/2 ) 462); IR (Nujol) cm-1 1681 (m, νMoH). Anal.
Calcd for C15H40N4Si3Mo: C, 39.45; H, 8.83 N; 12.27. Found: C,
39.95; H, 8.46; N, 12.06.
The conversion of [N3N]Mo(C6H11) into [N3N]MoH was followed
in toluene by UV/vis at 562 nm. The first-order rate constants (in units
of 10-4 s-1) at temperature (T, K) at 323 K and k were found to be
0.127 (323), 0.428 (333), 1.07 (343), 3.68 (353), and 9.64 (363). A
plot of ln(k/T) versus 1/T gave ∆Hq ) 24 501 cal/mol and ∆Sq ) -5.27
eu with R ) 0.9989. At 343 K the rate constant for decomposition of
[N3N]Mo(C6H11) was calculated to be 1.07 × 10-4 s-1, while [N3N]-
Mo(C6D11) was found in two experiments to be 3.0 × 10-5 and 3.3 ×
10-5 s-1 for a kH/kD ) 3.4(2).
[bitN3N]Mo. (Me3SiNCH2CH2)3N]MoCl (500 mg, 1.02 mmol) was
dissolved in 13 mL THF and placed in a bomb. Magnesium powder
(30 mg, 1.23 mmol) was added, and the bomb was sealed. The vessel
was subjected to three freeze-pump-thaw cycles to remove any
dinitrogen present, and the solution was stirred under vacuum. After
7 days, the solvent was removed and the residue extracted with 15 mL
of pentane and filtered to give a blood-red solution. The pentane was
removed to give the crude product as a red solid (420 mg) that was
(according to its NMR spectrum) a 1:4 mixture of [N3N]MoH and
[bitN3N]Mo. The crude yield of [bitN3N] (in the mixture) therefore is
The rate of R-elimination in [N3N]Mo(C6D11) was determined by

Synthesis and Decomposition Alkyl-Mo Complexes
J. Am. Chem. Soc., Vol. 119, No. 49, 1997 11893
72%. Pure [bitN3N]Mo was obtained by recrystallization of the crude
product from hexamethyldisiloxane: 1H NMR δ 18.58 (CΗ2), 14.81
(SiMe3), 1.10 (SiMe2), -18.95 (CΗ2), -20.26 (CΗ2), -98.65 (CΗ2),
-103.55 (CΗ2), -125.06 (CΗ2). Anal. Calcd for C15H38N4Si3Mo:
C, 39.62; H, 8.42; N, 12.32. Found: C, 39.74; H, 8.73; N, 12.39.
[N3N]MoD. [bitN3N]Mo (95 mg, 0.21 mmol) were dissolved in 5
mL of THF and placed in a glass bomb with a stirring bar. The vessel
was sealed and subjected to two freeze-pump-thaw cycles. D2 (1
atm) was introduced and the solution stirred for 2 days, during which
time the color of the solution changed from blood-red to yellow. The
solvent was removed and the solid recrystallized from pentane as yellow
needles; the yield was quantitative: 2D NMR (C6H6) δ 16.48 (Si(CH3)2-
CH2D, line width ) 6 Hz).
N2-filled drybox. A gelatin capsule and 2.2 × 1.9 cm piece of parafilm
were weighed. The capsule was then loaded with the sample, and the
parafilm was folded and packed on top using plastic tongs. The capsule
was closed and weighed again to determine the sample mass. It was
then suspended in a straw. The straw was placed in a plastic bottle
with a screw cap and the bottle was tightly sealed. At the instrument
the straw was quickly attached to the sample rod and transferred to the
helium atmosphere. Measurements were taken in 1° intervals from 5
to 10 K, 2° intervals from 12 to 20 K, 3° intervals from 23 to 50 K, 5°
intervals from 55 to 100 K, 10° intervals from 110 to 200 K, and 20°
intervals from 220 to 300 K. A background measurement of an empty
gel capsule, parafilm square, and straw was taken over the entire
temperature range and subtracted from the experimental values at each
temperature. The susceptibility at each temperature also was corrected
for the diamagnetic contribution by the ligands using Pascal’s constants.
X-ray Structure of [N3N]Mo(triflate). The structure was solved
at 188 K using a Siemens SMART/CCD diffractometer (see the
Supporting Information): empirical formula C16H39F3N4O3SSi3Mo, FW
) 604.78, space group Pna21, a ) 20.951(4) Å, b ) 10.999(2) Å, c )
12.050(2) Å, V ) 2776.7(10) Å,3 Z ) 4, Dcalc ) 1.447 mg/m3.
X-ray Structure of [N3N]Mo(CD3). The structure was solved at
188 K using a Siemens SMART/CCD diffractometer (see the Sup-
porting Information): empirical formula C16H39D3N4Si3Mo, FW )
473.77, space group Cc, a ) 17.352(10) Å, b ) 9.6851(2) Å, c )
Acknowledgment. We thank the National Science Founda-
tion for research support (CHE 91 22827) and for funds to help
purchase a departmental Siemens SMART/CCD diffractometer.
N.C.M-Z. thanks Ciba-Geigy for a Ciba-Geigy Jubila¨ums
Stiftung and Jake Parrott (RSI program) for assistance in UV/
vis measurements. S.W.S. thanks M. G. Fickes and R.R.S.
thanks D. Nocera for helpful discussions.
9.445(3) Å, â ) 110.6210(10)°, V ) 2508.00(7) Å,3 Z ) 4, Dcalc
)
1.255 mg/m3.
X-ray Structure of [N3N]Mo(cyclohexyl). The structure was solved
at 188 K using a Siemens SMART/CCD diffractometer (see the
Supporting Information): empirical formula C21H50N4Si3Mo, FW )
538.86, space group P6Å, a ) 20.820(6) Å, b ) 20.820(6) Å, c )
11.680(5) Å, V ) 4385(3) Å,3 Z ) 6, Dcalc ) 1.224 mg/m3.
X-ray Structure of [bitN3N]Mo. The structure was solved at 188
K using a Siemens SMART/CCD diffractometer (see the Supporting
Information): empirical formula C15H38N4Si3Mo, FW ) 454.70, space
group P21/n, a ) 8.972(3) Å, b ) 17.308(4) Å, c ) 15.398(3) Å, â )
100.61(3)°, V ) 2350.2(11) Å,3 Z ) 4, Dcalc ) 1.285 mg/m3.
Solid-State Magnetic Susceptibility Measurements. SQUID
experiments were performed at 5 kG on a Quantum Design 5.5 T
instrument running MSRP2 software. Samples were prepared in an
Supporting Information Available: Labeled ORTEP draw-
ing, crystal data and structure refinement, atomic coordinates,
bond lengths and angles, and anisotropic displacement param-
eters, for [N3N]Mo(triflate) (1b), [N3N]Mo(CD3) (2a-d3), [N3N]-
Mo(cyclohexyl) (2j), and [bitN3N]Mo, and solid-state suscep-
tibility data for selected compounds (27 pages). An X-ray
crystallographic file, in CIF format, is available through the
Internet only. See any current masthead page for ordering and
Internet access instructions.
JA970697X