β-diketiminato scandium chemistry: Synthesis, characterization, and thermal behavior of primary amido alkyl derivatives

Article abstract of DOI:10.1021/om034345c

Treatment of the β-diketiminato-supported scandium dichlorides {[ArNC(R)CHC(R)NAr]-ScCl₂}_n (Ar = 2,6-ⁱPr ₂-C₆H₃; R = CH₃, 1a, n = 2; R = ^tBu, 1b, n = 1) with 1 equiv of a lithium amide reagent LiN(H)R′ (R′ = ^tBu, 2,6-ⁱPr₂-C ₆H₃) gave scandium amido derivatives. For 1a, use of LiN(H)^tBu leads to the bis-amido derivative (6a) regardless of the equivalency of amide reagent employed, suggesting that facile ligand redistribution processes are operative when the ligand is the less bulky methyl-substituted example. For 1b, mono-amido chlorides 2b (R′ = ^tBu) and 3b (R′ = 2,6-ⁱPr₂-C ₆H₃) are obtained in good yields, and these compounds can be alkylated with MeLi to provide mono-amido methyl compounds 4b (R′ = ^tBu) and 5b (R′ = 2,6-ⁱPr₂-C ₆H₃). All four of these compounds were characterized crystallographically. The amido ligand occupies the exo coordination site exclusively, and there is no evidence in solution for a diastereomer with the amido group in the endo site. DFT calculations suggest that there is a strong steric preference and a slight electronic bias for the amido ligand to assume the exo position. Thermolysis of the amido methyl complex 4b leads to loss of CH₄ and production of a scandacylic product, 7, formed via metalation with one of the N-aryl isopropyl methyl groups. This compound was characterized crystallographically. Deuterium labeling experiments suggest that 7 is produced via direct metalation and does not form via a scandium imido intermediate.

Full text of DOI:10.1021/om034345c

Organometallics 2004, 23, 2087-2094
2087
â-Dik etim in a to Sca n d iu m Ch em istr y: Syn th esis,
Ch a r a cter iza tion , a n d Th er m a l Beh a vior of P r im a r y
Am id o Alk yl Der iva tives
Lisa K. Knight,† Warren E. Piers,*,†,1 Paul Fleurat-Lessard,†,2
Masood Parvez,† and Robert McDonald‡
Department of Chemistry, University of Calgary, 2500 University Drive, N.W,
Calgary, Alberta, Canada T2N 1N4, and X-Ray Structure Laboratory,
Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2G2
Received December 3, 2003
Treatment of the â-diketiminato-supported scandium dichlorides {[ArNC(R)CHC(R)NAr]-
t
ScCl2}n (Ar ) 2,6-iPr2-C6H3; R ) CH3, 1a , n ) 2; R ) Bu, 1b, n ) 1) with 1 equiv of a
lithium amide reagent LiN(H)R′ (R′ ) tBu, 2,6-iPr2-C6H3) gave scandium amido derivatives.
For 1a , use of LiN(H)tBu leads to the bis-amido derivative (6a ) regardless of the equivalency
of amide reagent employed, suggesting that facile ligand redistribution processes are
operative when the ligand is the less bulky methyl-substituted example. For 1b, mono-amido
t
chlorides 2b (R′ ) Bu) and 3b (R′ ) 2,6-iPr2-C6H3) are obtained in good yields, and these
compounds can be alkylated with MeLi to provide mono-amido methyl compounds 4b (R′ )
tBu) and 5b (R′ ) 2,6-iPr2-C6H3). All four of these compounds were characterized crystallo-
graphically. The amido ligand occupies the exo coordination site exclusively, and there is no
evidence in solution for a diastereomer with the amido group in the endo site. DFT
calculations suggest that there is a strong steric preference and a slight electronic bias for
the amido ligand to assume the exo position. Thermolysis of the amido methyl complex 4b
leads to loss of CH4 and production of a scandacylic product, 7, formed via metalation with
one of the N-aryl isopropyl methyl groups. This compound was characterized crystallo-
graphically. Deuterium labeling experiments suggest that 7 is produced via direct metalation
and does not form via a scandium imido intermediate.
In tr od u ction
the LnMdNR species is a critical first step in the cycle
involving LnMR2 precursors; the imido intermediate is
rapidly captured by substrate via a 2+2 cycloaddition,
which is the key C-N bond forming step in the process.
Imido ligands dNR play an important role in early
transition metal organometallic chemistry as both spec-
tator ligands and reactive intermediates.3 For example,
bulky imido donors have been used in an ancillary role
as a Cp equivalent in group 5 and 6 metal-based olefin
polymerization catalysts,4 providing molecular frag-
ments that are isolobal with group 4 metallocenes.
Furthermore, highly reactive group 4 imido complexes,
LnMdNR, have been shown to be capable of activating
C-H bonds by 1,2 addition across the MdN linkage in
a remarkable reaction that both activates and function-
alizes the alkane in one step.5 Such intermediates are
also implicated in catalytic cycles for the hydroamina-
tion of both allenes and alkynes by titanium- and
zirconium-based catalyst precursors.6 Extensive kinetic
and experimental evidence suggests that generation of
Given the rich mosaic of group 4, 5, and 6 imido
chemistry, it is surprising that group 3 imido chemistry
is to date extremely limited. To our knowledge, no
examples of discrete, monomeric compounds LnMdNR
for a group 3 metal have been prepared.7 Recently,
Hessen et al. have reported a dimeric scandium imido
that implicates the intermediacy of a monomeric imido
species,8 but clearly the tendency to dimerize9 in these
very Lewis acidic complexes is a challenge to address
in their preparation and in their potential competence
as intermediates in catalytic cycles.
(5) Walsh, P. J .; Hollander, F. J .; Bergman, R. G. J . Am. Chem. Soc.
1988, 110, 8729. (b) Walsh, P. J .; Hollander, F. J .; Bergman, R. G.
Organometallics 1993, 12, 3705. (c) Cummins, C. C.; Baxter, S. M.;
Wolczanski, P. T. J . Am. Chem. Soc. 1988, 110, 8731. (d) Schaller, C.
P.; Cummins, C. C.; Wolczanski, P. T. J . Am. Chem. Soc. 1996, 118,
591. (e) Wolczanski, P. T.; Bennett, J . L. J . Am. Chem. Soc. 1994, 116,
2179.
(6) Bytschkov, I.; Doye, S. Eur. J . Org. Chem. 2003, 935.
(7) Emslie, D. J . H.; Piers, W. E. Coord. Chem. Rev. 2002, 233-34,
129.
(8) Beetstra, D. J .; Meetsma, A.; Hessen, B.; Teuben, J . H. Orga-
nometallics 2003, 22, 4372.
(9) Other related examples: (a) Gordon, J . C.; Giesbrecht, G. R.;
Clark, D. L.; Hay, J .; Keogh, D. W.; Poli, R.; Scott, B. L.; Watkin, J .
G.; Organometallics 2002, 21, 4726. (b) Chan, H.-S.; Li, H.-W.; Xie, Z.
Chem. Commun. 2002, 652.
† University of Calgary.
‡ University of Alberta.
(1) Phone: 403-220-5746. E-mail: wpiers@ucalgary.ca. S. Robert
Blair Professor of Chemistry (2000-2005).
(2) Current address: Laboratoire de Chimie, EÄ cole Normale Su-
pe´rieure de Lyon, 46, Alle´e d’Italie, 69364 Lyon Cedex 07, France.
(3) (a) Gade, L. H.; Mountford, P. Coord. Chem. Rev. 2001, 216-
217, 65. (b) Duncan, A. P.; Bergmann, R. G. Chem. Rec. 2002, 2, 431.
(4) (a) Chan, M. C. W.; Cole, J . M.; Gibson, V. C.; Howard, J . A. K.;
Lehmann, C.; Poole, A.; Siemeling, U. J . Chem. Soc., Dalton Trans.
1998, 103. (b) Gibson, V. C.; Redshaw, C.; Clegg, W.; Elsegood, M. R.
J .; Siemeling, U.; Tu¨rk, T. J . Chem. Soc., Dalton Trans. 1996, 4513.
(c) Gibson, V. C. Mod. Coord. Chem. 2002, 140.
10.1021/om034345c CCC: $27.50 © 2004 American Chemical Society
Publication on Web 03/26/2004

2088 Organometallics, Vol. 23, No. 9, 2004
Knight et al.
Sch em e 1
t
Recently, we have reported a family of â-diketiminato
(“nacnac”)-supported organoscandium compounds [ArN-
C(R)CHC(R)NAr]ScR′2 (Ar ) 2,6-iPr2-C6H3; R ) CH3,
series a , R ) tBu, series b; R′ ) Me, Et, Bn, CH2SiMe3,
CH2CMe3).10 The bulky character of these particular
nacnac ancillaries11 allows for access to base-free scan-
dium dichloride starting materials and bis-alkyl deriva-
tives. We have recently observed that some of the bis-
alkyl compounds are viable catalyst precursors for the
intramolecular hydroamination of alkynes and alkynes.12
While organo group 3 and lanthanide-based catalysts
for this reaction have been explored in detail by Marks
and co-workers, most are based on mono-alkyl or amido
precursors13 and utilize a mechanism whose foundation
is the σ bond metathesis reaction.14 Given that the bis-
alkyl functionality of our catalyst precursors offers the
opportunity for an imido-based mechanism similar to
that seen in group 4 catalysts, we were interested in
exploring the (nacnac)Sc fragment as a platform for
imido chemistry.15 Herein we report our initial inves-
tigations with this goal in mind, using a strategy
pioneered by Bergman5a,b and Wolczanski5c-e in their
seminal group 4 studies involving the thermolysis of
primary amido methyl complexes of general formula
(nacnac)Sc(CH3)(NHR).
a since the larger Bu groups push the N-aryl groups
forward by 5-7° and hold them more perpendicular to
the N2C3 ligand plane.11 Thus, the scandium dichloride
starting materials 1a and 1b, whose syntheses have
been described previously,10 differ in that 1a exists as
a mono-THF adduct upon initial preparation, while the
more sterically congested 1b is monomeric and THF-
free. The THF can be removed from 1a ‚THF under high
vacuum, yielding a dimeric dichloride that is used to
preclude inclusion of THF into subsequent products.16
Amido derivatives can be prepared from dichlorides
1 and LiNHR (R ) tBu, 2,6-iPr2-C6H3). The steric
differences between the ligands are apparent in their
reactivity toward LiNHtBu (Scheme 1). 1b reacts cleanly
with 1 equiv of either lithium tert-butyl amide or the
bulky lithium anilide LiNH-2,6-iPr2C6H3 to give the
amido-chlorides 2b and 3b; these can be subsequently
alkylated with MeLi to give the methyl amido complexes
4b and 5b. While Mindiola et al. have shown that
reaction between 1a ‚THF and 1 equiv of LiNH-2,6-i-
Pr2C6H3 gives the amido chloride 3b‚THF,17 reaction
between base-free 1a and LiNHtBu affords mixtures of
the starting dichloride and the bis-amido complex 6a .
Compound 6a can be separately prepared by using 2
equiv of LiNHtBu. Although we have not studied this
reaction in detail, it appears that putative amido
chloride 2a is not stable toward ligand redistribution
to 1a and 6a , probably due to the fact that the less
sterically demanding methyl-substituted ligand a allows
for formation of the dimeric transition states necessary
for ligand exchange.
Resu lts a n d Discu ssion
Syn th etic Ch em istr y. The nacnac ligands employed
in this study differ only in the substituent on the
backbone of the ligand, one incorporating methyl groups
(a series), the other the more bulky tBu group (b series).
Although the structural change is somewhat remote
This tendency to redistribute ligands in the a series
of compounds can be used to advantage for the prepara-
tion of the tert-butylamido methyl complex 4a (Scheme
2). Equimolar mixtures of 1a and the dimeric dimethyl
complex shown16 slowly redistribute the methyl and
chloro ligands at room temperature to give pure samples
of the methyl chloride complex. The process cannot be
encouraged using heat, since this leads to metallative
decomposition of the dimethyl complex, as reported
earlier.10 Treatment of the methyl chloride species with
LiNHtBu gives the desired amido methyl derivative 4a .
Unfortunately, the thermolysis of this compound is also
dominated by ligand redistribution processes; heating
t
from the locus of coordination, Bu-substituted ligand
b is significantly more sterically imposing than ligand
(10) Hayes, P. G.; Lee, L. W. M.; Knight, L. K.; Piers, W. E.; Parvez,
M.; Elsegood, M. R. J .; Clegg, W.; MacDonald, R. Organometallics 2001,
20, 2533.
(11) Budzelaar, P. H. M.; van Oort, A. B.; Orpen, A. G. Eur. J . Inorg.
Chem. 1998, 1485.
(12) Lauterwasser, F.; Hayes, P. G.; Piers, W. E.; Brase, S.; Schafer,
L. L. Organometallics 2004, 23, in press.
(13) Ryu, J .-S.; Li, G. Y.; Marks, T. J . J . Am. Chem. Soc. 2003, 125,
12584.
(14) For a recent paper on mono-ligand bis-amido and alkyl precur-
sors, see: Hong, S.; Tian, S.; Metz, M. V.; Marks, T. J . J . Am. Chem.
Soc. 2003, 125, 14768.
(15) These bulky â-diketiminato ligands have been used to generate
monomeric group 13 imido complexes: (a) Cui, C. M.; Roesky, H. W.;
Schmidt, H. G.; Noltemeyer, M. Angew. Chem., Int. Ed. 2000, 39, 4531.
(b) Hardman, N. J .; Cui, C. M.; Roesky, H. W.; Fink, W. F.; Power, P.
P. Angew. Chem., Int. Ed. 2001, 40, 2172.
(16) Hayes, P. G.; Piers, W. E.; Parvez, M. J . Am. Chem. Soc. 2003,
125, 5622.
(17) Basuli, F.; Tomaszewski, J .; Huffman, J . C.; Mindiola, D. J .
Organometallics 2003, 22, 4705.

â-Diketiminato Scandium Chemistry
Organometallics, Vol. 23, No. 9, 2004 2089
Sch em e 2
to 100 °C yields samples of the bis-amido complex 6a
and the thermal decomposition products of the dimethyl
species. Given the propensity of the amido complexes
supported by the less sterically demanding a ligand to
undergo ligand redistribution, we subsequently focused
on the full characterization and thermolysis of the
compounds 2-5b, supported by the tBu-substituted
ligand system.
Solid St a t e a n d Solu t ion Ch a r a ct er iza t ion of
2-5b. All four of the amido compounds of ligand b
reported here have been characterized crystallographi-
cally; the molecular structures of the tert-butyl amido
complexes 2b and 4b are given in Figure 1, while those
of the N-aryl amido compounds 3b and 5b are given in
Figure 2. Selected metrical parameters for the four
compounds are collected in Table 1, and crystal data
and refinement details can be found in Table 3; full
experimental details for all structure determinations are
included as Supporting Information. Data associated
with the ligand are similar to those found in a range of
(nacnac)ScR2 compounds;10 in particular, the bite angle
of the ligand defined by N(1)-Sc-N(2) is within the
expected range for this ligand, and the Sc-Nligand bond
lengths of ∼2.1 Å are as anticipated. The Sc-N(3)
distance to the amido nitrogen, however, is significantly
shorter in all cases, and the distance of 1.986(2) Å for
2b is to our knowledge the shortest Sc-N bond to date.
While this might be attributable in part to the low
coordination number of the compound, there is also
likely a strong π-donation component to the bonding
with the N(3) nitrogen in these compounds.
As for most four-coordinate complexes of scandium
and other metals supported by the bulky nacnac ligand
b, the metal dips out of the plane defined by the N2C3
atoms of the ligand backbone. This is primarily to allow
the two other substituents on the metal to minimize
interactions with the N-aryl isopropyl groups. For the
tert-butyl amido compounds 2b and 4b the deviation of
Sc from this plane (∼1.2 Å, Table 1) is more severe than
that found for the N-aryl amido compounds (∼0.9 Å).
This arrangement creates chemically distinct environ-
ments for the two coordination sites at the metal not
occupied by the ligand. The endo position lies under-
neath the N2C3 ligand plane, while the exo site points
away from this plane. As can be seen from Figures 1
and 2, in all four compounds 2-5b, the amido ligand
occupies the exo position, while the chloride or methyl
groups are found in the endo coordination site.
F igu r e 1. Crystalmaker diagrams of the tert-butyl amido
complexes 2b (A, above) and 4b (B, below). In B, the N-aryl
isopropyl groups and the methyl groups of the tert-butyl
groups of the ligand backbone have been removed for
clarity.
We have shown that the endo and exo positions are
in fast exchange in solution for compounds (nacnac)-
ScR2 and that this exchange can be frozen out at low
temperature on the NMR time scale.10 Likely, this
exchange takes place via a transition state of C2v
symmetry where the Sc atom lies in the N2C3 ligand
plane; we thus term this exchange process the “ligand
flip” mechanism. For compounds where the two ligands
on Sc are different, this ligand flip process exchanges
two diastereomers in which the two groups occupy
different sites in each isomer. We have observed this
behavior in the ion pairs formed from (Ligb)ScMe2 and

2090 Organometallics, Vol. 23, No. 9, 2004
Knight et al.
For compounds 2-5b, which should also exhibit this
sort of behavior, only one set of signals is observed in
1
the H NMR spectrum down to 185 K and up to 363 K,
suggesting that the exo-amido isomer found in the solid
state is significantly more stable than the endo isomer
or that the barrier to exchange of the two diastereomers
is high enough to make exchange on the NMR time scale
slow. The latter explanation is unlikely, given that a
ligand flip exchange of the two tert-butyl amido groups
in 6a is quite readily frozen out at 195 K (∆Gq ) 13.6
kcal mol-1), indicating that the presence of π donor
amido ligands does not increase the barrier to ligand
flipping significantly.20 However, the fact that the mixed
alkyl species [ArNC(tBu)CHC(tBu)NAr]ScCH2SiMe3-
(CH3), which is sterically similar to 4b, exhibits ex-
changing diastereomers, present in a 2:1 ratio at 185
K,21 suggests that the π donating amido ligand may be
playing a role in stabilizing the exo-amido isomer over
the isomer with the amido group in the endo position.
To probe this question further, a DFT computational
study was conducted.
Using coordinates obtained from the X-ray structure
of 2b, the relative energies and Sc-N bond orders were
calculated for the exo and endo isomers of (Ligb)ScCl-
(NHR) (Table 2). To minimize steric factors, calculations
for R ) H were performed in addition to the genuine
molecule where R ) tBu. In addition, for the calculations
on the isomers of 2b, the orientation of the tert-butyl
group relative to Sc-Cl was varied. In the crystal
structure, the tert-butyl group is located trans to the
Sc-Cl vector (Cl-Sc-N(3)-C(36) ) 179.3(3)°), compris-
ing the exo-trans isomer. The other permutations are
depicted in the heading of Table 2. Note that as the
amido substitutent gets larger, deviations from pure
trans and cis orientations are quite severe, as demon-
strated by the E-Sc-N(3)-C(36) torsion angles of 49.8-
(2)° and 16.8(2)° for 3b and 5b, respectively. Therefore,
calculations were not performed on models of these
compounds.
F igu r e 2. Crystalmaker diagrams of the 2,6-iPr2-C6H3
anilido complexes 3b (A, above) and 5b (B, below). In A,
the N-aryl isopropyl groups of the ligand have been
removed for clarity.
As can be seen in Table 2, in the sterically unencum-
bered NH2 compounds, the exo isomer is calculated to
be more stable than the endo isomer by 3.0 kcal mol-1
.
This appears to be primarily due to an increased bond
order between Sc and N in comparison to the endo
structure. In general, the bond order between Sc and
the amido nitrogen is indicative of some multiple bond
character, while the bonds to the nacnac nitrogens more
closely approximate Sc-N single bonds. This is in
keeping with the observed differences in bond lengths
to the two types of nitrogens. A similar but less
pronounced trend of higher bond orders for the exo
isomers is observed in the calculated structures for the
tert-butyl amido compounds. In both the trans and cis
oriented exo isomers, the bond order to the amido
nitrogen is slightly higher than that observed in the
corresponding endo isomers. However, in these com-
pounds, steric factors play a much larger role in
determining the relative energies of the isomers. The
exo-trans species, corresponding to the solid state struc-
Ta ble 1. Selected Metr ica l P a r a m eter s for
Sca n d iu m Mixed -Am id o Com p ou n d s 2-5b
parameter
2b
4b
3b
5b
Selected Bond Distances (Å)
Sc-N(1)
2.125(2) 2.131(2) 2.062(1) 2.097(1)
2.115(2) 2.148(1) 2.165(1) 2.173(1)
1.986(2) 2.000(2) 2.013(1) 2.040(1)
2.125(2) 2.229(2) 2.3692(5) 2.212(2)
1.265(3) 1.207(2) 0.923(2) 0.920(2)
Sc-N(2)
Sc-N(3)
Sc-Ea
N2C3 plane-Sc
Selected Bond Angles (deg)
N(1)-Sc-N(2)
C(1)-N(2)-C(24)
C(3)-N(1)-C(12)
Sc-N(3)-C(36)
E-Sc-N(3)
92.82(8) 93.40(6) 91.90(5) 91.55(5)
126.3(2) 126.2(2) 126.6(1) 126.0(1)
127.6(2) 126.0(1) 124.8(1) 124.4(1)
140.3(2) 139.4(2) 149.5(1) 144.8(1)
105.60(8) 108.72(9) 125.63(4) 119.68(8)
Selected Torsion Angle (deg)
Ea-Sc-N(3)-C(36) 179.3(3) 175.1(2) 49.8(2)
16.8(2)
a
E ) Cl for 2b and 3b, C(40) for 4b, and C(48) for 5b.
(18) Hayes, P. G.; Piers, W. E.; McDonald, R. J . Am. Chem. Soc.
2002, 124, 2132.
(19) Knight, L. K.; Piers, W. E.; McDonald, R. Chem. Eur. J . 2000,
6, 4322.
B(C6F5)3, where two isomers are observed in an ap-
proximately 2:1 ratio at low temperature,18 as well as
some telluride and tellurolate derivatives.19
(20) The barrier to ligand flip in (Liga )Sc(CH2CMe3)2 is 11.4 kcal
mol-1; see ref 10.
(21) Hayes, P. G.; Piers, W. E., unpublished results.

â-Diketiminato Scandium Chemistry
Organometallics, Vol. 23, No. 9, 2004 2091
Ta ble 2. Ca lcu la ted Bon d Or d er s a n d Rela tive En er gies of Va r iou s Am id e Ch lor id e Com p ou n d s
of Liga n d b
∆E°rel
(kcal/mol)
compound
Sc-N(1)
Sc-N(2)
Sc-N(3)
LbSc(NHH)Cl-exo
0.84
0.83
0.81
0.82
0.81
0.81
0.82
0.81
0.83
0.83
0.83
0.83
1.35
1.30
1.27
1.29
1.24
1.24
0.0
3.0
0.0
25.9
72.4
11.1
LbSc(NHH)Cl-endo
LbSc(NHtBu)Cl-exo-trans
LbSc(NHtBu)Cl-exo-cis
LbSc(NHtBu)Cl-endo-trans
LbSc(NHtBu)Cl-endo-cis
Ta ble 3. Su m m a r y of Da ta Collection a n d Str u ctu r e Refin em en t Deta ils for Com p lexes
2b
3b
4b
5b
7
formula
fw
C
39H63N3ScCl
C50H78N3ScCl
801.56
C40H66N3Sc
633.92
C48H74N3Sc
738.06
C39H62N3Sc
617.88
654.33
cryst syst
a, Å
b, Å
monoclinic
18.3145(15)
17.6376(15)
12.6483(11)
triclinic
10.9568(5)
12.9352(6)
19.6028(9)
77.8167(9)
73.7643(9)
65.6747(9)
2416.02(19)
P1h
triclinic
11.2559(8)
12.4490(9)
15.4106(11)
88.2825(14)
89.3523(15)
65.4512(15)
1963.3(2)
P1h
orthorhombic
13.5396(2)
18.1218(2)
18.5560(2)
monoclinic
12.147(3)
17.490(5)
18.028(6)
c, Å
R, deg
â, deg
γ, deg
V, Å3
105.0938(19)
96.116(12)
3944.7(6)
P21/c
4
1.102
0.282
0.0553
0.1662
1.171
4552.94(10)
P212121
4
1.077
0.195
0.045
0.105
1.02
3808.3(19)
P21/n
4
1.078
0.221
0.0459
0.1098
1.033
space group
Z
2
2
d
calc, mg m-3
1.102
0.242
0.0396
0.1098
1.072
0.216
0.0481
0.1352
µ, mm-1
R1
wR2
gof
1.020
1.024
Sch em e 3
methine multiplets and seven methyl doublets for the
iPr groups is clearly indicative of the outcome of this
reaction. Furthermore, two doublets of doublets at 0.95
3
and 0.44 ppm (2J HH ) 12.4 Hz; J HH ) 5.4 Hz) for the
diastereotopic protons of the metalated methyl group
are a hallmark of these metalated compounds. The
structure of 7 was also confirmed by X-ray crystal-
lography, and two views of the molecule are given in
Figure 3, along with selected metrical data. While some
disorder was present in the molecular core, it was
successfully modeled (see Supporting Information) and
the metalated nature of the compound is clearly ex-
posed. From the front view of the molecule shown in
Figure 3A, it is clear that a consequence of the meta-
lation is a more “upright” set of N-aryl rings, which
allows the Sc atom to slip back into the ligand’s
coordination plane (deviation of the Sc atom from the
N2C3 plane is only 0.438(8) Å). Notably, the amido group
again occupies the exo coordination site, and the Sc-
C(35) distance of 2.251(4) Å is only slightly longer than
the analogous bonds in 4b and 5b.
ture, was found to be the lowest in energy, with the
other orientations of the tert-butyl amido ligand signifi-
cantly destabilized relative to this isomer. Thus, for the
substituted tert-butyl amido derivative, there is both an
electronic and a steric preference for the amido ligand
to occupy the exo coordination site, and equilibration to
the endo-trans isomer via a ligand flip mechanism is
not thermodynamically favored.
Th er m olysis of Meth yl Am id o Com p ou n d 4b. The
exo-trans geometry of 4b seems poised to eliminate
methane in chemistry analogous to that reported by
Bergman5a,b and Wolczanski,5c-e and generation of a
scandium imido complex. Heating NMR scale samples
of 4b in C6D6 or C7D8 to 60 °C indeed led to loss of CH4
over the course of several hours, cleanly producing
Two potential mechanisms for the formation of 7 are
depicted in Scheme 4. One involves direct metalation
i
via σ bond metathesis between the C-H bond of the Pr
1
compound, 7, whose H NMR spectrum was consistent
group and the Sc-CH3 bond,22 while the other invokes
the intermediacy of a reactive scandium imido species.
Mindiola has reasonably speculated on the involvement
of such an imido species to rationalize some deuterium
scrambling processes in the related compound (Liga )-
i
with metalation of one of the Pr methyl groups of the
N-aryl nacnac substitutents (Scheme 3). This process
is the primary thermal decomposition pathway for the
neutral dialkyl compounds LScR2;10 formation of 7 is
qualitatively much slower than the metalation processes
in LScR2. However, a characteristic pattern of four
(22) Fekl, U.; Goldberg, K. I. J . Am. Chem. Soc. 2002, 124, 6804.

2092 Organometallics, Vol. 23, No. 9, 2004
Knight et al.
of 7 involves the imide intermediate trapped by addition
i
of an Pr C-H bond across the reactive ScdNR frag-
ment. Heating samples of d1-4b lead exclusively to
2
production of CH4, and both 1H and H NMR spectro-
scopy indicated that the deuterium label remains on the
amido nitrogen, with no significant scrambling of the
deuterium into other positions of the molecule. This
observation argues against the intermediacy of a scan-
dium imido species in this chemistry.
Su m m a r y a n d Con clu sion s
While the (nacnac)Sc fragment supports primary
amido alkyl complexes, these compounds do not serve
as precursors to scandium imido derivatives in the same
way that amido alkyls of the group 4 metals do. Rather,
metalation of the nacnac ligand is the low-energy
pathway for CH4 loss. Use of more metalation-resistant
nacnac ligands or other approaches based on deproto-
nation strategies are currently being explored in at-
tempts to realize well-defined scandium imido com-
plexes. Our results do, however, suggest that scandium
imido compounds are relatively high energy species and
raise questions as to the viability of such intermediates
in hydroamination cycles catalyzed by LnMR2 com-
pounds,12,14 where M is a group 3 or lanthanide metal.
Exp er im en ta l Section
Gen er a l P r oced u r es. An MBraun or Vacuum Atmo-
spheres argon-filled glovebox was employed for manipulation
and storage of oxygen- and moisture-sensitive complexes.
Reactions were performed utilizing a double-manifold argon/
vacuum line and modified Schlenk line techniques. Matheson
Oxisorb-W gas purification cartridges were used to remove
residual moisture and oxygen in the argon stream. Hexane,
toluene, and THF were dried and deoxygenated as they passed
through a Grubbs/Dow purification system23 and were stored
in evacuated bombs over titanocene24 (hexane and toluene) or
sodium/benzophenone ketyl (THF).
F igu r e 3. Crystalmaker diagrams of two views of the
metalated amido complex 7. Selected bond lengths (Å): Sc-
N(1), 2.149(3); Sc-N(2), 2.140(3); Sc-N(3), 2.016(2); Sc-
C(35), 2.251(4); N2C3 plane-Sc, 0.438(8). Selected bond
angles (deg): N(1)-Sc-N(2), 83.6(1); C(1)-N(2)-C(24),
130.6(3); C(3)-N(1)-C(12), 127.8(2); Sc-N(3)-C(36), 143.5-
(2); C(35)-Sc-N(3), 116.0(1). Selected torsion angles
(deg): C(3)-N(1)-C(12)-C(13), 91.0(6); C(1)-N(2)-C(24)-
C(25), 116.0(4); C(3)-N(1)-C(12)-C(17), -96.8(6); C(1)-
N(2)-C(24)-C(29), -75.6(5).
NMR spectra were obtained on Bruker AC-200 MHz (1H and
13C{1H}), AMX-300 MHz (1H, 2H, 1H-1H EXSY, and 1H-1H
ROESY), and DRX-400 MHz (1H, 13C{1H}, 1H-13C HMQC, 1H-
1
1H EXSY, and H-1H COSY) instruments. All NMR spectral
Sc(NHAr)HBEt3,15 but no definitive evidence for its
competence as an intermediate was put forward. We
probed this question by monitoring the loss of CH4/
CH3D from d1-4b, selectively labeled in the ND position.
As shown in Scheme 4, direct metalation should produce
CH4, while loss of CH3D would indicate that formation
data are reported in ppm, and NMR spectra of 1H, 2H, and
13C were internally referenced to the residual solvent peak.
Temperature calibration for the NMR spectra was achieved
1
by monitoring the H NMR spectrum of methanol or ethylene
glycol.25 Elemental analyses were performed by Dorothy Fox
or Roxanna Smith using a Control Equipment Corporation
Sch em e 4

â-Diketiminato Scandium Chemistry
Organometallics, Vol. 23, No. 9, 2004 2093
(CEC) 440 elemental analyzer. The electrospray ionization
mass spectrometry was undertaken by Qiao Wu with a Bruker
Esquire 3000, and the high-resolution mass spectrometry was
performed by Dorothy Fox with a Kratos MS80RFA.
mL). Removal of hexane in vacuo gave crude product, and after
recrystallization from the same solvent at -36 °C, pure 2b
was obtained. Yield: 0.222 g, 72%. 1H NMR: 7.08-6.97 (m,
6H, C6H3), 5.82 (s, 1H, CH), 4.32 (br s, 1H, NH), 3.62, 3.03
(m, 2 × 2H, CH(CH3)2), 1.67, 1.38, 1.29, 1.23 (d, 4 × 6H, CH-
(CH3)2, J H-H ) 6.6 or 6.8 Hz), 1.15 (s, 18H, NCC(CH3)3), 0.81
(br s, 9H, NHC(CH3)3. 13C{1H} NMR: 175.2 (NCC(CH3)3),
144.8, 142.7, 140.8, 126.5, 124.8, 124.3 (C6H3), 93.1 (CH), 54.5
(NHC(CH3)3), 45.3 (NCC(CH3)3), 34.4 (NHC(CH3)3), 32.7 (NCC-
(CH3)3), 29.6 and 29.1 (CH(CH3)2), 27.5, 26.1, 25.2, and 25.0
((CH(CH3)2). Anal. Calcd for C39H63ClN3Sc: C, 71.77; H, 9.71;
N, 6.43. Found: C, 71.92; H, 9.54; N, 6.42.
All materials were obtained from Sigma-Aldrich and puri-
fied according to standard procedures. ScCl3‚THF3 was pro-
duced by refluxing Sc2O3 (purchased from Boulder Scientific)
in concentrated HCl, and after the removal of the excess HCl,
the reaction mixture was dried with thionyl chloride, SOCl2,
in refluxing THF.26 The ligands a and b were synthesized and
lithiated according to literature procedures. All amines/
anilines were dried over CaH2 and freshly distilled to another
flask before lithiation with 1 equiv of nBuLi to produce the
LiNRR′ salts. For the deuterium labeling studies, H2NtBu was
mixed with 20 equiv of D2O and stirred overnight to produce
D2NtBu and dried as above.
Syn th esis of d 1-2b. This compound was prepared in a
manner identical to the one previously described for 2b, with
the exception that LiNDtBu was used in place of LiNHtBu.
1
The H NMR spectrum of d 1-2b matched the spectrum of 2b,
except that no resonance was observed for the proton attached
to the nitrogen of the amido group.
Com p u ta tion a l P r oced u r es. All computational analyses
were performed on the MACI cluster (Multimedia Advanced
Computational Infrastructure) at the University of Calgary.
All of the calculations have been carried out using the
Amsterdam Density Function program package developed by
Barends et al.27 and vectorized by Ravenek.28,29 The Becke-
Perdew exchange correlation function30 has been used through-
out. The numerical integration scheme applied was developed
by te Velde et al.31 Auxialiary32 s, p, d, f, and g STO functions
centered on all nuclei were used to fit the Coulomb and
exchange potentials during the SCF process. For the scandium
atom, a standard triple-ú STO basis set from the ADF database
TZP was employed with 1s-3p electrons treated as frozen core.
For the nonmetal elements, a standard double-ú basis set with
one set of polarization functions (ADF database DZP) was
used, with 1s electron treated as frozen core for non-hydrogen
atoms. The geometries were taken directly from the X-ray
structures and were not further optimized. The geometries of
the endo isomers were derived from those of the exo isomers
by exchanging the amido ligand with the Cl or Me ligands.
The geometries of the cis and trans isomers were derived by
rotating the amido ligand around the Sc-N bond by 180° with
no further adjustment. The bond orders were calculated
following the scheme developed by R. F. Nalewajski and co-
workers.33
Syn th esis of 3b. The method to prepare 3b using 1b (0.607
g, 0.984 mmol) and LiNH-2,6-iPr2-C6H3 (0.215 g, 1.17 mmol)
is similar to the procedure for 2b above except that the reaction
mixture was sonicated for approximately 5 min prior to stirring
to increase the solubility of the reagents. On account of the
reduced solubility of 3b in comparison to 2b, the precipitate
was washed five times with hexane to extract any residual
product. 3b was recrystallized from hexane at -36 °C. Yield:
0.356 g, 48%. 1H NMR: 7.02-6.90 and 6.75 (m, 9H, 3 × C6H3),
6.24 (br s, 1H, NH), 6.09 (s, 1H, CH), 3.75 and 3.09 (m, 2 ×
2H, CH(CH3)2), 1.98 (m, 2H, NH-2,6-CH(CH3)2C6H3), 1.50,
1.29, 1.15, and 1.07 (d, 4 × 6H, CH(CH3)2, J H-H ) 6.7 or 6.8
Hz), 1.11 (s, 18H, NCC(CH3)3), 1.04 (d, 12H, NH-2,6-CH-
(CH3)2C6H3), J H-H ) 6.6 Hz ). 13C{1H} NMR: 177.3 (NCC-
(CH3)3), 149.1, 143.7, 143.2, 142.0, 134.6, 127.4, 125.8, 124.8,
124.6, 124.2, 123.5, and 118.8 (3 × C6H3), 96.8 (CH), 45.0
(NCC(CH3)3), 32.5 (NCC(CH3)3), 32.2 (NH-2,6-CH(CH3)2C6H3),
31.0, 27.3, 25.6 and 25.5 (CH(CH3)2), 29.2 and 29.1 (CH(CH3)2),
25.1 (NH-2,6-CH(CH3)2C6H3). Anal. Calcd for C47H71N3ScCl:
C, 74.42; H, 9.44; N, 5.54. Found: C, 74.58; H, 9.12; N, 5.30.
Syn t h esis of 4b. Toluene (10 mL) was added to a flask
charged with 2b (0.796 g, 1.22 mmol) and MeLi (0.034 g, 1.55
mmol) and stirred overnight at 35 °C. The toluene was
removed in vacuo and replaced with hexane (10 mL). Prior to
filtration, the deep yellow reaction mixture was warmed to
approximately 40 °C; after filtration, the residual product was
extracted from the LiCl precipitate with hexane (3 × 5 mL).
Upon removal of the solvent, a yellow powder was obtained.
Crude 4b was recrystallized from hexane at -36 °C, generat-
Syn th esis of 2b. Toluene (25 mL) was condensed into an
evacuated flask containing 1b (0.291 g, 0.472 mmol) and
LiNHtBu (0.049 g, 0.619 mmol). The yellow reaction mixture
was warmed to room temperature with a hot water bath and
stirred for 1 h before the removal of the solvent under reduced
pressure. Hexane (25 mL) was added to the flask, and the
mixture was heated to approximately 40 °C to increase the
solubility of the compound in hexane. The LiCl was removed
by filtration, and the solid was washed with hexane (3 × 10
1
ing yellow crystals. Yield: 0.394 g, 51%. H NMR: 7.10-6.99
(m, 6H, C6H3), 5.72 (s, 1H, CH), 3.68 and 3.09 (m, 2 × 2H,
CH(CH3)2), 3.54 (br s, 1H, NH), 1.67, 1.40, 1.28 and 1.26 (d, 4
× 6H, CH(CH3)2, J H-H ) 6.7 Hz), 1.16 (s, 18H, NCC(CH3)3),
0.80 (s, 9H, NHC(CH3)3), -0.07 (s, 3H, ScCH3). 13C{1H}
NMR: 174.6 (NCC(CH3)3), 145.1, 142.5, 141.0, 126.1, 124.7,
and 124.3 (C6H3), 93.2 (CH), 53.7 (NHC(CH3)3), 45.3 (NC-
C(CH3)3), 34.9 (NHC(CH3)3), 32.9 (NCC(CH3)3), 30.6 (ScCH3)
29.5 and 29.0 (CH(CH3)2), 27.3, 26.2, 25.2, and 25.0 (CH(CH3)2).
Anal. Calcd for C40H66N3Sc: C, 75.78; H, 10.49; N, 6.63.
Found: C, 75.72; H, 10.31; N, 6.58.
(23) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J . Organometallics 1996, 15, 1518.
(24) Marvich, R. H.; Brintzinger, H. H. J . Am. Chem. Soc. 1971,
93, 2046.
(25) Ammann, C.; Meier, P.; Merbach, A. E. J . Magn. Reson. 1982,
46, 319.
(26) Manzer, L. Inorg. Synth. 1982, 21, 135.
(27) (a) Baerends, E. J . Ph.D. Thesis, Free University, Amsterdam,
The Netherlands, 1973. (b) Baerends, E. J .; Ellis, D. E.; Ros, P. Chem.
Phys. 1973, 2, 41.
(28) Ravenek, W. In Algorithms and Applications on Vector and
Parallel Computers; te Riele, H. J . J ., Dekker, T. J ., van de Horst, H.
A., Eds.; Elsevier: Amsterdam, The Netherlands, 1987.
(29) Boerrigter, P. M.; te Velde, G.; Baerends, E. J . Int. J . Quantum
Chem. 1988, 33, 87.
(30) (a) Becke, A. Phys. Rev. A 1988, 38, 3098. (b) Perdew, J . P. Phys.
Rev. B 1986, 34, 7406. (c) Perdew, J . P. Phys. Rev. B 1986, 33, 8822.
(31) te Velde, G.; Baerends, E. J . Comput. Chem. 1992, 99, 84.
(32) Krijn, J .; Baerends, E. J .; Vernoijs, P. At. Nucl. Data Tables
1982, 26, 483.
(33) (a) Nalewajski, R. F.; Mrozek, J . Int. J . Quantum Chem. 1994,
51, 187. (b) Nalewajski, R. F.; Mrozek, J .; Mazur, G. Can. J . Chem.
1996, 74, 1121. (c) Mrozek, J .; Nalewajski, R. F.; Michalak, A. Polish
J . Chem. 1998, 72, 1779.
Syn th esis of d 1-4b. This compound was prepared in a
manner identical to the one previously described for 4b, with
1
the exception that d 1-2b was used. The H NMR spectrum of
d 1-4b matched the analogous spectrum of 4b, except that no
resonance was observed for the proton attached to the nitrogen
of the amido group. The synthesis of d 1-4b was also confirmed
with 2H NMR: 3.54 (ND).
Syn th esis of 5b. This synthesis was similar to that
described for 4b, except that the reaction mixture of 3b (0.728
g, 0.961 mmol) and MeLi (0.024 g, 1.09 mmol) was heated for
1
a full 24 h at 35 °C. Yield: 0.330 g, 47%. H NMR: 7.02-6.89
and 6.74 (m, 9H, 3 × C6H3), 5.99 (s, 1H, CH), 5.44 (br s, 1H,
NH), 3.68 and 3.02 (m, 2 × 2H, CH(CH3)2), 2.18 (m, 2H, NH-

2094 Organometallics, Vol. 23, No. 9, 2004
Knight et al.
2,6-CH(CH3)2C6H3), 1.45, 1.30, 1.16, and 1.05 (d, 4 × 6H, CH-
(CH3)2, J H-H ) 6.6 Hz or 6.8 Hz), 1.12 (s, 18H, NCC(CH3)3),
1.03 (d, 12H, NH-2,6-CH(CH3)2C6H3), J H-H ) 6.8 Hz ), 0.13
(s, 3H, ScCH3). 13C{1H} NMR: 177.1 (NCC(CH3)3), 149.8,
143.6, 143.4, 142.1, 134.2, 127.0, 125.5, 124.8, 123.4, and 117.7
(10 of 12 signals were observed for 3 × C6H3), 96.1 (CH), 45.0
(NCC(CH3)3), 32.7 (NCC(CH3)3), 30.7 (NH-2,6-[CH(CH3)2]-
C6H3), 29.2 and 29.0 (CH(CH3)2), 27.1, 25.7, 25.3, and 25.0 (CH-
(CH3)2), 24.8 (NH-2,6-CH(CH3)2C6H3), 23.4 (ScCH3). Anal.
Calcd for C48H74N3Sc: C, 78.11; H, 10.11; N, 5.69. Found: C,
77.88; H, 9.94; N, 5.66.
mixture was warmed to room temperature, and the white solid
was removed by filtration. The product was extracted from the
precipitate with hexane (3 × 10 mL); the hexane was removed
under reduced pressure from the filtrate, and the remaining
yellow solid was recrystallized from dilute hexane at -36 °C.
1
Yield: 0.072 g, 34%. H NMR: 7.18-7.09 (m, 6H, C6H3), 5.03
(s, 1H, CH), 3.93 (br s, 1H, NH), 3.59 and 3.08 (m, 2 × 2H,
CH(CH3)2, J H-H ) 6.6 Hz), 1.62 (s, 6H, NCCH3), 1.56, 1.29,
1.25, and 1.06 (d, 4 × 6H, CH(CH3)2, J H-H ) 6.6 Hz), 0.87 (s,
9H, NHC(CH3)3), and 0.03 (s, 3H, ScCH3). 13C{1H} NMR: 167.8
(NCCH3), 143.50, 143.45, 142.1, 126.9, 124.8, and 124.5 (C6H3),
95.5 (CH), 53.2 (NHC(CH3)3, 35.0 (NHC(CH3)3), 29.3 and 29.0
(CH(CH3)2), 26.5, 25.0, 24.8, and 24.7 (CH(CH3)2), 24.4 (NCCH3),
and 21.4 (ScCH3). Anal. Calcd for C34H54N3Sc: C, 74.28; H, 9.90;
N, 7.64. Found: C, 74.20; H, 9.42; N, 7.29.
Syn th esis of 7. Toluene (20 mL) was transferred into a
round-bottom flask of a frit assembly containing 4b (0.437 g,
0.690 mmol); the reaction mixture was heated overnight at
90 °C. After heating, the mixture became an opaque orange
solution. The solvent was removed under reduced pressure,
yielding an orange foam. Hexane (12 mL) was added to the
flask, and the mixture was stirred at room temperature for
30 min. The hexane was removed in vacuo, affording an orange
powder. The powder was suspended in hexamethyldisiloxane
(7 mL) and isolated by filtration. Yield: 0.158 g, 37%. 1H
NMR: 7.31-6.95 (m, 6H, C6H3), 5.55 (s, 1H, CH), 3.48 and
2.99 (m, 1H, CH(CH3)2, J H-H ) 6.9 Hz ), 3.22 (m, 2H, CH(CH3)2
and CH2CH(CH3)), 1.96 (br s, 1H, NH), 1.56, 1.50, 1.35, 1.34,
1.26 and 1.20 (d, 6 × 3H, CH(CH3)2, J H-H ) 6.8 or 6.4 Hz ),
1.18 (br s, 12H, CH(CH3)2 and NCC(CH3)3), 1.17 (br s, 9H,
Th er m olysis of 4a . The thermolysis was performed only
1
on an NMR tube scale, and the reaction was monitored by H
NMR spectroscopy. Upon heating of the scandium complex 4a
at 100 °C in toluene-d8, the compound redistributed to 6a ,
which was stable at 100 °C, and La ScMe2, which decomposed
as a mixture of species at this temperature.
Syn th esis of 6a . A round-bottom flask was charged with
1a (0.440 g, 0.825 mmol) and 2 equiv of LiNHtBu (0.135 g,
1.71 mmol); after the combination of the two solids, toluene
(20 mL) was transferred into the flask. The yellow, opaque,
reaction mixture was heated overnight at 50 °C. Upon cooling
to room temperature, the solution was filtered and the
precipitate was washed with toluene (2 × 15 mL). The toluene
was removed in vacuo from the filtrate, and hexane (15 mL)
was added to the flask; the solution was warmed to ambient
temperature and stirred for 20 min. A yellow powder resulted
from the removal of hexane under reduced pressure, and it
was used without further purification. Yield: 0.492 g, 98%.
1H NMR (T ) 298 K): 7.14-7.10 (m, 6H, C6H3), 5.02 (s, 1H,
CH), 3.50 (br s, 2H, NHC(CH3)3), 3.44 (m, 4H, CH(CH3)2, J H-H
) 6.8 Hz), 1.63 (s, 6H, NCCH3), 1.38, and 1.18 (d, 2 × 12H,
2
NCC(CH3)3), 0.97, 0.91 (dd, 1H, CH2CH(CH3), J H-H ) 12.2
3
Hz, J H-H ) 5.4 Hz ), 0.78 (s, 9H, NHC(CH3)3), 0.44 (br t, 1H,
CH2CH(CH3)). 13C{1H} NMR: 174.2 and 173.9 (2 × NCC-
(CH3)3), 147.9, 144.5, 143.8,142.4, 139.9, 139.4, 127.2, 126.9,
125.3, 125.1,124.3, and 123.2 (2 × C6H3), 98.1 (CH), 54.2
(ScCH2) 52.9 (NHC(CH3)3), 43.55 and 43.51 (2 × NCC(CH3)3),
35.0 (NHC(CH3)3), 33.1 and 32.3 (2 × NCC(CH3)3), 39.5, 28.9,
28.7, and 28.5 (3 × CH(CH3)2 and CH2CH(CH3)), 26.7, 26.6,
26.33, 26.29, 25.2, 24.0, and 23.2 (7 × CH(CH3)2); mass
spectrum, m/z (relative intensity, %) 617 (M+, 1), 544 (M+
-
H2NtBu, 3), 502 (M+ - (H2NtBu + C3H6), 3), 9 (100); exact mass
calcd for C39H62N3Sc 617.4503, found 617.4495.
Syn th esis of d 1-7. In a sealed NMR tube, the deuterated
scandium amido methyl complex d 1-4b was heated overnight
at 60 °C in C6D6. 1H NMR spectroscopy was employed to
determine the products, and d 1-7 was assigned on the basis
of the noticeable absence of a peak at 1.98 ppm for the amido
proton of the metalate complex, although all of the other peaks
were present in the spectrum. In addition, only CH4 at 0.14
ppm was observed in the reaction mixture; a resonance for
CH3D was not present in the NMR tube.
Syn th esis of [Ar NC(Me)CHC(Me)NAr ]Sc(Cl)Me. Equimo-
lar ratios of [ArNC(Me)CHC(Me)NAr]ScMe2 (0.308 g, 0.625
mmol) and 1a (0.335 g, 0.628 mmol) were combined in toluene
(70 mL), and the reaction was stirred for 3 days. The reaction
mixture was filtered and washed once with toluene (10 mL),
and the filtrate was dried under reduced pressure, yielding a
pale, yellow powder. Yield: 0.199 g, 62%. 1H NMR: 7.13-7.00
(m, 6H, C6H3), 5.09 (s, 1H, CH), 3.42 and 2.73 (m, 2 × 2H,
CH(CH3)2, J H-H ) 6.6 Hz), 1.63 (s, 6H, NCCH3), 1.24, 1.22,
1.17, and 0.89 (d, 4 × 6H, CH(CH3)2, J H-H ) 6.6 Hz), 0.29 (s,
3H, ScCH3). 13C{1H} NMR (C7D8): 168.3 (NCCH3), 145.6,
143.0, 141.1, 126.2, 124.7, and 123.6 (C6H3), 98.1 (CH), 39.6
(ScCH3), 29.1 and 28.0 (CH(CH3)2), 26.3, 25.0, 24.9, 24.7, and
24.3 (CH(CH3)2 and NCCH3).
Syn th esis of 4a . The scandium complex LaSc(Me)Cl (0.199
g, 0.388 mmol) was weighed into a two-neck round-bottom
flask and dissolved in toluene (20 mL). A separate round-
bottom flask was charged with LiNHtBu (0.031 g, 0.392 mmol),
and it was also dissolved in toluene (5 mL). The toluene
solution of LiNHtBu was added slowly via syringe to the
scandium solution. The reaction mixture was stirred at ambi-
ent temperature for 30 min, resulting in a yellow solution with
a white precipitate. The toluene was removed in vacuo, and
hexane (25 mL) was transferred into the flask. The reaction
1
CH(CH3)2, J H-H ) 6.8 Hz), and 1.03 (s, 18H, NHC(CH3)3); H
NMR (T ) 195 K, toluene-d8): 7.14-7.03 (m, 6H, C6H3), 4.84
(s, 1H, CH), 3.98 and 3.24 (br s, 2 × 1H, NHC(CH3)3), 3.98
and 3.04 (br m, 2 × 2H, CH(CH3)2), 1.65, 1.31, 1.11, and 0.98
(br d, 2 × 12H, CH(CH3)2), 1.52 (s, 6H, NCCH3), 1.44 and 0.82
(br s, 2 × 9H, NHC(CH3)3). 13C{1H} NMR: 166.7 (NC(CH3)3),
144.0, 142.0, 125.9, and 124.1 (C6H3), 94.9 (CH), 52.5 (NH-
C(CH3)3, 34.8 (NHC(CH3)3, 28.1 (CH(CH3)2), 25.1, and 24.6
(CH(CH3)2), and 24.0 (NCCH3); high-resolution mass spectrum,
m/z (relative intensity, %): 534 (M+ - NHtBu, 7), 460 (M+
-
2H2NtBu, 9), 418 (M+ - (2H2NtBu + C3H6), 24), 58 (100); exact
mass calcd for C33H51N3Sc (M+ - NHtBu) 534.3642, found
534.3618. Electrospray ionization mass spectrum, m/z: 629 (M
+ Na+) and 557 (M + Na+ - H2NtBu).
Ack n ow led gm en t. Funding for this work came
from the Natural Sciences and Engineering Research
Council of Canada in the form of a Discovery Grant (to
W.E.P.). P.F.-L. thanks Dr. A. Michalak for providing
the software for bond order calculations.
Su p p or tin g In for m a tion Ava ila ble: Tables of atomic
coordinates, anisotropic displacement parameters, and com-
plete bond distances, angles, and torsion angles for 2b, 3b,
4b, 5b, and 7. This material is available free of charge via the
Internet at http://pubs.acs.org.
OM034345C