Scandium alkyl complexes supported by a ferrocene diamide ligand

Article abstract of DOI:10.1021/om7007277

Synthesis of a scandium dimethylbenzyl complex supported by a ferrocene diamide ligand was accomplished by alkane elimination from Sc(CH ₂Xy-3,5)₃(THF)₂. The scandium dimethylbenzyl complex Sc(fc[NSi(t-Bu)Me₂]₂)(CH₂Xy-3,5)(THF), 2-(CH₂Xy-3,5)(THF), was used as a starting material for the synthesis of the corresponding chloride-bridged dimer, (2-Cl)₂, which, in turn, led to a scandium bis(neo-pentyl) ate salt, Li[2-Np₂]. Attempts to remove the coordinated THF molecule from 2-(CH₂Xy-3,5)(THF) with AlMe₃ led to the isolation of a scandium methyl complex with two coordinated AlMe₃ molecules, 2-Me(AlMe₃)₂. Compound 2-Me(AlMe₃)₂ led to a scandium methyl complex, 2-Me(THF)₂, by stirring in THF. All ferrocene diamido compounds were characterized by X-ray crystallography. DFT calculations on model compounds were used to explain the stability of the compounds synthesized and to probe the existence of an iron-scandium interaction. Compounds 2-(CH₂Xy-3,5) (THF) and 2-Me(AlMe₃)₂ polymerize L-lactide.

Full text of DOI:10.1021/om7007277

Organometallics 2008, 27, 363–370
363
Scandium Alkyl Complexes Supported by a Ferrocene Diamide
Ligand
Colin T. Carver, Marisa J. Monreal, and Paula L. Diaconescu*
Department of Chemistry & Biochemistry, UniVersity of California, Los Angeles, California 90095
ReceiVed July 20, 2007
Synthesis of a scandium dimethylbenzyl complex supported by a ferrocene diamide ligand was accomplished
by alkane elimination from Sc(CH2Xy-3,5)3(THF)2. The scandium dimethylbenzyl complex Sc(fc[NSi(t-
Bu)Me2]2)(CH2Xy-3,5)(THF), 2-(CH2Xy-3,5)(THF), was used as a starting material for the synthesis of the
corresponding chloride-bridged dimer, (2-Cl)2, which, in turn, led to a scandium bis(neo-pentyl) ate salt, Li[2-
Np2]. Attempts to remove the coordinated THF molecule from 2-(CH2Xy-3,5)(THF) with AlMe3 led to the
isolation of a scandium methyl complex with two coordinated AlMe3 molecules, 2-Me(AlMe3)2. Compound
2-Me(AlMe3)2 led to a scandium methyl complex, 2-Me(THF)2, by stirring in THF. All ferrocene diamido
compounds were characterized by X-ray crystallography. DFT calculations on model compounds were used
to explain the stability of the compounds synthesized and to probe the existence of an iron-scandium interaction.
Compounds 2-(CH2Xy-3,5)(THF) and 2-Me(AlMe3)2 polymerize L-lactide.
observed in a number of cases. In practice, a fine balance between
electronic and steric factors needs to be maintained in order to be
able to evaluate the reactivity observed and explore the electro-
philicity of the metal center.7
Introduction
Cyclopentadienyl complexes have dominated group 3 organo-
metallic chemistry for a long time.1–6 While astonishing results
have been achieved with this class of ligands, interest in noncy-
clopentadienyl complexes is increasing,7–10 with the goal of
expanding the reactivity of the metal centers toward challenging
substrates, such as saturated11 or unsaturated3,12 hydrocarbons and
polar substrates.13–15 Since the metal center is very electrophilic
even in neutral alkyl complexes, design of complexes with the goal
of σ-bond metathesis reactions in mind has focused on dianionic
ancillary ligand, monoalkyl complexes.16–23 Architectures such as
bis-cyclopentadienyl, ansa-metallocene, bis-amide have been ex-
plored2 and good reactivity in C-H bond activation24–27 has been
Our goal is to synthesize scandium alkyl complexes that
display reactivity behavior complementary to existing ones. We
are interested in exploiting ferrocene 1,1′-diamides as versatile
frameworks in supporting highly electrophilic group 3 metal
centers. Such ligands have a number of desirable characteristics:
(i) the 1,1′-disubstituted ferrocene enforces cis-coordination of
the two donors; thus only one side of the metal center is blocked;
(ii) the ferrocene backbone has the ability to accommodate
changes in the electronic density at the metal center, by varying
the geometry around iron; (iii) the other ligands and incoming
substrates coordinate in a plane perpendicular to the plane
formed by the two donors, iron, and the metal; thus the access
to the electrophilic center is unrestricted; (iv) a weak interaction
of donor–acceptor type could occur between iron and the metal,
possibly influencing the reactivity of the complex.28
* Corresponding author. E-mail: pld@chem.ucla.edu.
(1) Janiak, C. Coord. Chem. ReV. 2006, 250, 66–94.
(2) Gromada, J.; Carpentier, J.-F.; Mortreux, A. Coord. Chem. ReV. 2004,
248, 397–410.
(3) Hou, Z.; Wakatsuki, Y. Coord. Chem. ReV. 2002, 231, 1–22.
(4) Arndt, S.; Okuda, J. Chem. ReV. 2002, 102, 1953–1976.
(5) Evans, W. J.; Davis, B. L. Chem. ReV. 2002, 102, 2119–2136.
(6) Molander, G. A.; Romero, J. A. C. Chem. ReV. 2002, 102, 2161–
2186.
Results and Discussion
(7) Piers, W. E.; Emslie, D. J. H. Coord. Chem. ReV. 2002, 233–234,
131–155.
Synthesis and Characterization of Scandium Alkyl
Complexes. The ligand fc[NHSi(t-Bu)Me2]2 (fc ) 1,1′-ferro-
cenylene)29 has been chosen for initial studies because the
(8) Mountford, P.; Ward, B. D. Chem. Commun. 2003, 1797–1803.
(9) Edelmann, F. T.; Freckmann, D. M. M.; Schumann, H. Chem. ReV.
2002, 102, 1851–1896.
(10) Marques, N.; Sella, A.; Takats, J. Chem. ReV. 2002, 102, 2137–
2160.
(20) Gountchev, T. I.; Tilley, T. D. Organometallics 1999, 18, 2896–2905.
(21) Görlitzer, H. W.; Spiegler, M.; Anwander, R. Eur. J. Inorg. Chem.
1998, 1998, 1009–1014.
(22) Liu, B.; Cui, D.; Ma, J.; Chen, X.; Jing, X. Chem.-Eur. J. 2007,
13, 834–845.
(23) Cameron, T. M.; Gordon, J. C.; Scott, B. L.; Tumas, W. Chem.
Commun. 2004, 1398–1399.
(24) Shilov, A. E.; Shul’pin, G. B. Chem. ReV. 1997, 97, 2879–2932.
(25) Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Peterson, T. H.
Acc. Chem. Res. 1995, 28, 154–162.
(26) Labinger, J. A.; Bercaw, J. E. Nature 2002, 417, 507–514.
(27) Goldberg, K. I.; Goldmanv, A. S. ActiVation and Functionalization
of C-H Bonds. In ACS Symp. Ser. 2004, 885.
(11) Piers, W.; Shapiro, P.; Bunel, E.; Bercaw, J. Synlett. 1990, 74–84.
(12) Hitzbleck, J.; Beckerle, K.; Okuda, J.; Halbach, T.; Mülhaupt, R.
Macromol. Symp. 2006, 236, 23–29.
(13) Zhong, Z.; Dijkstra, P. J.; Feijen, J. J. Biomat. Sci., Polym. Ed.
2004, 15, 929–946.
(14) O’Keefe, B. J.; Hillmyer, M. A.; Tolman, W. B. J. Chem. Soc.,
Dalton Trans. 2001, 2215.
(15) Yasuda, H. J. Organomet. Chem. 2002, 647, 128–138.
(16) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.; Nolan,
M. C.; Santarsiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am. Chem.
Soc. 1987, 109, 203–219.
(17) Fontaine, F. G.; Tilley, T. D. Organometallics 2005, 24, 4340–
4342.
(28) Seyferth, D.; Hames, B. W.; Rucker, T. G.; Cowie, M.; Dickson,
R. S. Organometallics 1983, 2, 472–474.
(18) Sadow, A. D.; Tilley, T. D. J. Am. Chem. Soc. 2003, 125, 7971–
7977.
(29) Monreal, M. J.; Carver, C. T.; Diaconescu, P. L. Inorg. Chem. 2007,
46, 7226–7228.
(19) Bercaw, J. E. Pure Appl. Chem. 1990, 62, 1151–4.
10.1021/om7007277 CCC: $40.75
2008 American Chemical Society
Publication on Web 01/04/2008

364 Organometallics, Vol. 27, No. 3, 2008
CarVer et al.
Scheme 1. Synthesis of Scandium ate Complex 1-K(THF)
synthesis of a close analogue has already been reported30 and
it is amenable to alteration of the nitrogen substituent. Salt
metathesis reactions between ScCl3(THF)3 and the potassium
salt of the ferrocene diamide ligand, [K(OEt2)]2fc[NSi(t-
Bu)Me2]2 (Scheme 1), were initially carried out. Regardless of
the conditions used, a potassium ate complex, [K(solvent)x]-
[Sc(fc[NSi(t-Bu)Me2]2)2], 1-K(solvent) (solvent ) THF, x )
2 and solvent ) toluene, x ) 1), was always isolated; two ligands
are coordinated to the metal (Figure 1). The single-crystal X-ray
structures are informative with respect to the relative orientation
of the two ligands: the N-Sc-N angles are 108° regardless of
whether the nitrogen atoms belong to the same ligand, indicating
a tetrahedral geometry around the scandium center. The ferrocene
backbones adopt an eclipsed conformation and the scandium-
nitrogen distances are ca. 2.14 Å, similar to other scandium-amide
distances reported.31–35
Figure 1. Thermal ellipsoid (35% probability) representation of
[K(THF)2][Sc(fc[NSi(t-Bu)Me2]2)2], 1-K(THF). The potassium
cation is not shown. Hydrogen atoms were omitted for clarity.
Scheme 2. Synthesis of Scandium Tris(dimethylbenzyl) Com-
plex Sc(CH2Xy-3,5)3(THF)3
Since salt metathesis reactions resulted in ate complex
formation, a common occurrence7 in the chemistry of group 3
and lanthanide elements, we decided to carry out acid–base
reactions. Inspired by a recent report36 by Hessen et al. on the
synthesis of LaBz3(THF)3 (Bz ) benzyl) we set out to isolate
the analogous scandium benzyl complex. Although ScBz3(THF)3
did form when employing a procedure similar to the one for
LaBz3(THF)3, its isolation was not trivial because of its limited
solubility in hydrocarbon solvents. It was reasoned that increas-
ing the hydrophobicity of the benzyl substituent might solve
this problem. Indeed, the isolation of Sc(CH2Xy-3,5)3(THF)2
(Scheme 2) was facile, and the desired compound could be
purified by recrystallization from toluene/hexanes. The yield of
the pure product is consistently 60%. During the preparation
of the manuscript, the synthesis of a scandium tris(dimethyl-
aminobenzyl) complex was reported;37,38 this complex features
dimethylamine coordinated as an internal Lewis base. Other
scandium tris(alkyl) starting materials39 have been known and
used in syntheses of scandium alkyl complexes; in most cases,
the alkyl chlorides used to prepare the alkali/alkaline earth alkyl
compound are expensive. Additionally, in acid–base reactions,
1 or 2 equiv of the corresponding alkane is eliminated and
therefore lost during the preparation of the final scandium
complex. It is advantageous that the byproduct of the alkane
elimination reactions is mesitylene since it is volatile; contami-
nation of the desired product with mesitylene is not likely. When
all these factors are considered, Sc(CH2Xy-3,5)3(THF)2 is a
practical entry into scandium organometallic chemistry.
With Sc(CH2Xy-3,5)3(THF)2 in our hands, we carried out
its reaction with fc[NHSi(t-Bu)Me2]2 and isolated Sc(fc[NSi(t-
Bu)Me2]2)(CH2Xy-3,5)(THF), 2-(CH2Xy-3,5)(THF), in 80%
yield after crystallization from pentane/Et2O (Scheme 3). Single
crystals were grown from pentane/toluene, and the crystal
structure of 2-(CH2Xy-3,5)(THF) (Figure 2) was determined.
The solid-state structure of 2-(CH2Xy-3,5)(THF) is informative
with respect to the scandium-iron distance of 3.16 Å. This
distance is longer than the titanium-iron distance in
[Ti(fc[NSiMe3]2)Me][MeBPh3] (3.07 Å),40 but it is shorter than
in the corresponding titanium dimethyl compound (3.32 Å),
consistent with an electrophilic scandium center in a neutral
compound (Sc(III) ionic radius is 0.87 Å, Ti(IV) ionic radius is
0.74 Å).41 The benzyl ligand coordinates in an η2-fashion: the
distance to the aryl carbon is 2.74 Å and the Sc-C-C angle is
92.1°. The THF molecule and the benzyl ligand coordinate in
a plane perpendicular to the one formed by scandium and the
nitrogen atoms. Although the O-Sc-C angle is 111.3°, the
O-Sc-N angles are 96.3° and the N-Sc-N angle is 140.9°,
indicating a relatively distorted tetrahedral environment around
scandium.
(30) Shafir, A.; Power, M. P.; Whitener, G. D.; Arnold, J. Organome-
tallics 2001, 20, 1365–1369.
(31) Knight, L. K.; Piers, W. E.; Fleurat-Lessard, P.; Parvez, M.;
McDonald, R. Organometallics 2004, 23, 2087–2094.
(32) Ward, B. D.; Dubberley, S. R.; Maisse-Francois, A.; Gade, L. H.;
Mountford, P. J. Chem. Soc., Dalton Trans. 2002, 4649–4657.
(33) Roussel, P.; Alcock, N. W.; Scott, P. Chem. Commun. 1998.
(34) Fryzuk, M. D.; Giesbrecht, G. R.; Rettig, S. J. Can. J. Chem. 2000,
78, 1003–1012.
(35) Eppinger, J.; Spiegler, M.; Hieringer, W.; Herrmann, W. A.;
Anwander, R. J. Am. Chem. Soc. 2000, 122, 3080–3096.
(36) Bambirra, S.; Meetsma, A.; Hessen, B. Organometallics 2006, 25,
3454–3462.
Thermal Decomposition of Scandium Benzyl Complex.
The thermal stability of 2-(CH2Xy-3,5)(THF) was investigated
by heating a C6D6 solution gradually to 60 °C for 24 h; under
these conditions, 2-(CH2Xy-3,5)(THF) was relatively robust.
(37) Jaroschik, F.; Shima, T.; Li, X.; Mori, K.; Ricard, L.; LeGoff, X. F.;
Nief, F.; Hou, Z. Organometallics 2007, 26, 5654–5660.
(38) Li, X.; Nishiura, M.; Mori, K.; Mashiko, T.; Hou, Z. Chem.
Commun. 2007, 4137–4139.
(40) Shafir, A.; Arnold, J. J. Am. Chem. Soc. 2001, 123, 9212.
(41) Shannon, R. Acta Crystallogr. A 1976, 32, 751–767. See also http://
abulafia.mt.ic.ac.uk/shannon/.
(39) Emslie, D. J. H.; Piers, W. E.; Parvez, M.; McDonald, R.
Organometallics 2002, 21, 4226–4240.

Scandium Alkyl Complexes
Organometallics, Vol. 27, No. 3, 2008 365
Scheme 3
Decomposition was observed upon prolonged heating at 80 °C,
but the products could not be identified. Heating a C7D8 solution
of 2-(CH2Xy-3,5)(THF) at 120 °C for 45 min also led to
decomposition and intractable mixtures. Since aromatic hydro-
carbons might be involved in C-H activation reactions with
alkyl groups, a control experiment was carried out, in which
2-(CH2Xy-3,5)(THF) was heated at 95 °C for 16 h in c-C6D12;
a similar reaction mixture was obtained during the experiment.
On the basis of the results of the thermal decomposition
experiments we can conclude that complex 2-(CH2Xy-
3,5)(THF) is relatively stable at moderate temperatures (60–80
°C); at higher temperatures its thermal decomposition is too
complicated, but it is not solvent dependent.
2-(CH2Xy-3,5)(THF) in a scandium chloride complex would
allow the isolation of a THF-free compound since the chloride
ion is a small ligand, biasing the metal center toward dimer
formation. If a dimer were formed, then the steric crowding
around the metal centers would prevent THF coordination.
Indeed, the reaction42 between 2-(CH2Xy-3,5)(THF) and excess
Me3SiCl (Scheme 3) resulted in the isolation of [Sc(fc[NSi(t-
Bu)Me2]2)(µ-Cl)]2, (2-Cl)2, a chloride-bridged dimer without
THF coordinated to the scandium center, as evidenced by its
X-ray crystal structure (Figure 3). The scandium-iron distance
in (2-Cl)2 (2.80 Å) is significantly shorter than in 2-(CH2Xy-
3,5)(THF) (3.16 Å), a trend similar to that observed when
comparing titanium-iron distances in [Ti(fc[NSiMe3]2)Me]-
[MeBPh3] and [Ti(fc[NSiMe3]2)Cl]2[B(C6F5)4] (3.07 vs 2.49 Å).
Unlike in the ate complex 1-K(THF), the two ferrocene ligands
are not tilted with respect to each other. One interesting feature
is the fact that the scandium atom sits out of the plane formed
by the two nitrogens and the two carbons connected to them;
such a geometrical characteristic has been correlated to π-bond-
ing involving the nitrogen atoms in other chelating diamide
complexes.43,44
Synthesis of THF-Free Chloride Scandium Complex. The
stability of 2-(CH2Xy-3,5)(THF) was not unexpected consider-
ing that Lewis base coordination (THF) to an electrophilic metal
center is known to increase the stability of the complex and
decrease its reactivity. We reasoned that the transformation of
With (2-Cl)2 isolated, the formation of a scandium alkyl,
THF-free complex was investigated. The reaction between (2-
Cl)2 and 2 equiv of LiNp (Np ) CH2-t-Bu) carried out in
hexanes gave Li[Sc(fc[NSi(t-Bu)Me2]2)Np2], Li[2-Np2], in
which two neo-pentyl ligands coordinate to scandium (Scheme
3). The Fe-Sc distance of 3.19 Å is only slightly longer than
in 2-(CH2Xy-3,5)(THF) (3.16 Å). The X-ray crystal structure
of Li[2-Np2] (Figure 4) features a lithium center disordered over
two equivalent positions. This symmetrical structure is supported
by the solution structure of Li[2-Np2], since the 1H NMR
spectrum is consistent with a C2V arrangement of the ligands
(see Supporting Information). Compound Li[2-Np2] decomposes
when heated at 60 °C in C6D6, but it is relatively stable to 40
°C. Other attempts to synthesize a scandium monoalkyl complex
supported by the ferrocene diamide ligand discussed here from
(42) Voskoboynikov, A. Z.; Parshina, I. N.; Shestakova, A. K.; Butin,
K. P.; Beletskaya, I. P.; Kuz’mina, L. G.; Howard, J. A. K. Organometallics
1997, 16, 4041–4055.
(43) Boncella, J. M.; Wang, S.-Y. S.; Vanderlende, D. D.; Leigh Huff,
R.; Abboud, K. A.; Vaughn, W. M. J. Organomet. Chem. 1997, 530, 59–
70.
Figure 2. Thermal ellipsoid (35% probability) representation of
Sc(fc[NSi(t-Bu)Me2]2)(CH2Xy-3,5)(THF), 2-(CH2Xy-3,5)(THF).
Only one of the two molecules in the unit cell is shown. Hydrogen
atoms were omitted for clarity.
(44) Ketterer, N. A.; Ziller, J. W.; Rheingold, A. L.; Heyduk, A. F.
Organometallics 2007, 26, 5330–5338.

366 Organometallics, Vol. 27, No. 3, 2008
CarVer et al.
one methyl group and one amide nitrogen. The 1H NMR
spectrum shows broad peaks for the ancillary ligand protons
and the coordinated AlMe3 protons, and it is consistent with
interchange of the methyl groups on the NMR spectroscopy
time scale.
When the stoichiometry for the formation of the bis(trimethyl
aluminum) adduct was 3 equiv of AlMe3 to 1 equiv of
2-(CH2Xy-3,5)(THF), the yield of the desired product increased
as expected. We propose that 1 equiv of AlMe3 is necessary
for the transmetalation of the benzyl group since AlMe3 is a
known alkylation reagent;46–48 it is likely that AlMe2(CH2Xy-
3,5)(THF) is formed and eliminated during the work up of the
reaction mixture. If less than 3 equiv of AlMe3 is used, then
starting material is still present at the end of the reaction.
Compound 2-Me(AlMe3)2 is thermally stable, as indicated by
heating it in C6D6 at 82 °C, which led to no decomposition
after several hours.
When 2-Me(AlMe3)2 was stirred in THF (Scheme 4), a new
compound formed after 1 h at room temperature. On the basis
1
of its H NMR spectrum, we assigned this product to be the
scandium methyl THF compound Sc(fc[NSi(t-Bu)Me2]2)-
Me(THF)2, 2-Me(THF)2. This assignment is supported by an
X-ray crystal structure determination (Figure 6). The iron-
scandium distance in 2-Me(THF)2 is 3.26 Å, longer by 0.1 Å
than in 2-(CH2Xy-3,5)(THF), consistent with the presence of
two coordinated THF molecules and decreased electrophilicity
of the scandium center. The geometry around scandium is best
described as a distorted trigonal bipyramid; although the
N-Sc-CMe angles deviate from 120° (108° and 113°), the other
angles around scandium are consistent with this interpretation:
O-Sc-CMe angles are 85° and 88° and the O-Sc-O angle is
173°. The scandium-carbon distance, Sc-CMe, of 2.28 Å is
similar to that found in Cp*2ScMe (2.24 Å).16
Figure 3. Thermal ellipsoid (35%) representation of [Sc(fc[NSi(t-
Bu)Me2]2)(µ-Cl)]2, (2-Cl)2. Hydrogen atoms were omitted for
clarity.
DFT Calculations. In order to understand the stability of
the scandium alkyl complexes reported here and to probe the
scandium-iron interaction, we set out to investigate computa-
tionally models for the various scandium ferrocene diamido
complexes synthesized (Table 1, see Supporting Information
for further details). Calculations were also performed on a model
featuring a THF-free scandium center; such calculations could
point to the preferred geometry of the scandium center in this
THF-free complex and might suggest explanations for the
observed coordination of more than one ligand besides the
ancillary ligand.
Figure 4. Thermal ellipsoid (35% probability) representation of
Li[Sc(fc[NSi(t-Bu)Me2]2)Np2], Li[2-Np2]. Hydrogen atoms and
atoms obtained from modeling thermal disorder were omitted for
clarity.
The agreement between the scandium-iron distances in the
model complexes and those obtained from crystal structure data
was good. In the case of [fc(NSiH3)2]Sc(CH3)2-, the Fe-Sc
calculated distance is larger by 0.08 Å than that observed in
the X-ray crystal structure of Li[2-Np2]; this difference might
arise from the fact that the calculations were carried out on an
anionic compound, with no contribution from the corresponding
cation. As can be seen in Figure 7 and Table 1, the compound
for which the molecular orbital shows the greatest contribution
from scandium to the interaction between iron and scandium is
[fc(NSiH3)2]Sc(CH3)(OMe2)2. In this model the alkyl group is
methyl and the coordinated THF molecules were replaced by
Me2O. The contribution from the iron orbital though is around
10%, much smaller than the iron contribution for the benzyl
the scandium chloride dimer (2-Cl)2 failed when starting with
the most common Grignard or lithium reagents, regardless of
reaction conditions.
Attempts to Remove the Coordinated THF Molecule.
Since the chloride route did not lead to a monoalkyl scandium
complex, other avenues for the removal of coordinated THF
were considered. Trimethyl aluminum has been used to abstract
coordinated THF molecules.45 The reaction between 2-(CH2Xy-
3,5)(THF) and an excess of AlMe3 (Scheme 4) led to the
isolation of
a scandium methyl complex, Sc(fc[NSi(t-
Bu)Me2]2)Me(AlMe3)2, 2-Me(AlMe3)2, in which two molecules
of AlMe3 are present (Figure 5): one molecule is connected to
scandium through two bridging methyl groups, the other through
(46) Vidyaratne, I.; Scott, J.; Gambarotta, S.; Duchateau, R. Organo-
metallics 2007, 26, 3201–3211.
(47) Yue, N.; Hollink, E.; Guerin, F.; Stephan, D. W. Organometallics
2001, 20, 4424–4433.
(45) Zimmermann, M.; Törnroos, K. W.; Anwander, R. Angew. Chem.,
Int. Ed. 2007, 46, 3126–3130.
(48) Gibson, V. C.; Newton, C.; Redshaw, C.; Solan, G. A.; White,
A. J. P.; Williams, D. J. Eur. J. Inorg. Chem. 2001, 2001, 1895–1903.

Scandium Alkyl Complexes
Organometallics, Vol. 27, No. 3, 2008 367
Scheme 4
model, [fc(NSiH3)2]Sc(CH2Ph)(OMe2). In an analogous com-
plex, [fc(NSiH3)2]Sc(CH3)(OMe2), in which the benzyl group
was replaced with a methyl group, the contributions from both
ironandscandiumtothemolecularorbitalwiththeiron-scandium
interaction are reduced. These contributions increase slightly
when the coordinated ether fragment is omitted and decrease
again for the bis(methyl) scandium anion [fc(NSiH3)2]Sc-
-
(CH3)2
.
Since we could not correlate conclusively the calculated
iron-scandium distance and the iron-scandium interaction,
NBO (natural bond order) charges were investigated (Table 1).
These charges might be used as an indication of the electro-
philicity of the metal center.49 The charge of the scandium center
(1.35) in [fc(NSiH3)2]Sc(CH2Ph)(OMe2) is slightly smaller than
the one in Cp2ScMe (1.39) and in [fc(NSiH3)2]Sc(CH3)(OMe2)2
(1.41). The methyl complexes with no or one coordinated ether
molecule, [fc(NSiH3)2]Sc(CH3) and [fc(NSiH3)2]Sc(CH3)-
(OMe2), show even higher scandium charges (1.57 and 1.51,
respectively). Based on this data, some conclusions can be
drawn: (1) the scandium charge (indicative of electrophilicity)
can be correlated with the iron-scandium distance in complexes
where the alkyl ligand is the same and (2) the scandium center
in ferrocene diamide alkyl complexes is highly electrophilic,
so the stability of these complexes is probably a consequence
of Lewis base coordination.
An explanation for the fact that an alkyl complex with no
additional ligand coordinating the scandium center has not been
synthesized can be advanced based on the geometry optimization
results for [fc(NSiH3)2]Sc(CH3); regardless of the geometry of
the starting complex, the final geometry showed an Fe-Sc-CMe
angle of ∼135°; the methyl group does not sit in the plane that
bisects the Cp rings, as observed for Cp2ScMe.50 As a
consequence, the unoccupied d orbitals of π symmetry for
[fc(NSiH3)2]Sc(CH3) point away from the methyl group, while
for Cp2ScMe they point toward the methyl group. The avail-
ability of multiple d orbitals for interaction with an incoming
ligand as opposed to just one for Cp2ScMe (the orbital of σ
symmetry) might explain the facility with which additional
ligands coordinate to the scandium center in ferrocene diamide
complexes.
Figure 5. Thermal ellipsoid (35% probability) representation of
Sc(fc[NSi(t-Bu)Me2]2)Me(AlMe3)2, 2-Me(AlMe3)2. Hydrogen at-
oms were omitted for clarity.
It is difficult to understand the contribution of the scandium-
iron interaction to the stability and/or electrophilicity of the
scandium center. Such donor–acceptor interactions51 have been
identified in zirconium or thorium-platinum complexes.52–54
In those cases, contributions from different metal orbitals were
taken as an indication of the strength of the interaction. If the
same criterion is applied to the present complexes, then
[fc(NSiH3)2]Sc(CH2Ph)(OMe2) and [fc(NSiH3)2]Sc(CH3)-
(OMe2)2 show the strongest iron-scandium interaction. As a
consequence, the electrophilicity of the scandium center in these
complexes is lower than in the other scandium ferrocene diamide
complexes modeled by DFT.
(49) Barros, N.; Eisenstein, O.; Maron, L. Dalton Trans. 2006, 3052–
3057.
(50) Barros, N.; Eisenstein, O.; Maron, L.; Tilley, T. D. Organometallics
Figure 6. Thermal ellipsoid (35% probability) representation of
Sc(fc[NSi(t-Bu)Me2]2)Me(THF)2, 2-Me(THF)2. Hydrogen atoms
were omitted for clarity.
2006, 25, 5699–5708.
(51) Bursten, B. E.; Novo-Gradac, K. J. J. Am. Chem. Soc. 1987, 109,
904–5.

368 Organometallics, Vol. 27, No. 3, 2008
CarVer et al.
Table 1. Orbital Contributions of Fe and Sc to the Fe-Sc Interaction and Charges of Scandium Determined by NBO Analysis
Sc-Fe distance
compound
(Å) calcd (exptl)
orbital
Sc + Fe (%)
Sc charge
[fc(NSiH3)2]ScBz(OMe2)
[fc(NSiH3)2]ScMe(OMe2)2
[fc(NSiH3)2]ScMe(OMe2)
[fc(NSiH3)2]ScMe
3.13 (3.16)
3.23 (3.26)
3.09
2.76
3.27 (3.19)
HOMO-3
HOMO-3
HOMO-2
HOMO-2
HOMO
8.8 + 27.2
13.9 + 9.44
3.4 + 5.6
5.0 + 10.0
2.8 + 44.5
1.35
1.41
1.51
1.57
-
[fc(NSiH3)2]ScMe2
Table 2. L-Lactide Polymerization Data for Reactions Catalyzed by
Ring-Opening Polymerization of L-Lactide. Although lan-
thanide complexes are widely used in the ring-opening poly-
merization of cyclic esters,14,55 reports of active scandium
catalysts are scarce.56,57 On the basis of studies performed with
the lanthanides and extension to scandium, the decreased
reactivity of its complexes is attributed to its smaller size
compared to those of the lanthanides.57 Given the results of the
DFT calculations that indicate electrophilic metal centers in
2-(CH2Xy-3,5)(THF) and 2-Me(THF)2 and the fact that access
to the scandium center is relatively unrestricted (coordination
of additional ligands/substrates happens in a plane perpendicular
to the plane bisecting the ferrocene ligand), we decided to
investigate whether these characteristics translate into good
activity for the ring-opening polymerization of cyclic esters.
We have chosen L-lactide because of its cost and availability
and the fact that it is more difficult to polymerize than other
cyclic esters (ꢀ-caprolactone, for example).55 Reactions with 1
mol % 2-(CH2Xy-3,5)(THF) or 2-Me(THF)2 were conducted
both in toluene and in THF. Surprisingly, the reaction in THF
mediated by 2-(CH2Xy-3,5)(THF) gave no polymer after 15 h
at 60 °C, although THF was reported to have a beneficial effect
on scandium lactide ring-opening reactions.56 On the other hand,
a
2-Me(THF)2
b
L-lactide:2-Me(THF)2
Mc
Mn
PDI
100
200
300
400
500
13 724
27 448
41 171
54 895c
68 619
52 583
43 655
52 151
69 0828c
80 012
1.33
1.63
1.59
1.46c
1.52
a All reactions were carried out in toluene at 70 °C and allowed to
reach 95.2% conversion before workup. b Mc was calculated based on
95.2% conversion. c A high molecular weight peak was also found: Mn
) 1 062 853; PDI ) 1.03.
the reaction in C6D6 reached 91% completion after 6 h at 60
°C and 88% after 15 h at room temperature for 2-(CH2Xy-
3,5)(THF). The polymer obtained had a PDI (polydispersity
index, Mw/Mn) of 1.4, but the value of Mn was 3 times that of
the calculated molecular weight, indicating possible decomposi-
tion of the active species during polymerization. When
2-Me(THF)2 was used as an initiator, the conversion was 92%
in C6D6 after only 1 h at 70 °C. Again, experimental molecular
weights were larger than those calculated (Table 2, 1.2 to 3.8
times larger) although the PDI values were around 1.5. Catalyst
decomposition was confirmed by the presence of free ligand in
1H NMR spectra. Although the polymerization seems somewhat
controlled, it obviates a kinetic analysis, since no comparison
with living-polymerization processes can be undertaken. In order
to understand the cause for this phenomenon, the following
experiment was performed: a polymerization reaction was
carried out with a 100:1 monomer to catalyst ratio; at the end
of this first polymerization, the reaction was sampled and a
second 100 equiv of L-lactide (with respect to initial catalyst
amount) was added. The polymers from the sample and at the
end of the second reaction were analyzed by GPC. The ratio
between the molecular weights (Mn) of the two polymer samples
was 1.5 and a unimodal molecular weight distribution was found
for the second polymer. We reasoned that if impurities are the
cause of the higher molecular weights obtained, then the same
amount of catalyst destroyed at the end of the first reaction will
be destroyed at the end of the second reaction and the Mn of
the second polymer will be higher than twice the Mn of the
first polymer. Since the molecular weight of the second polymer
is less (and not more) than twice as high as the molecular weight
of the first polymer, we propose that the catalyst decomposition
observed is related to the reaction progress and not to the
presence of adventitious impurities. It is likely that the process
by which an active catalyst is generated from the scandium alkyl
complexes is complicated; once the active species is generated,
polymerization proceeds in a rather controllable manner,
explaining the relatively low PDIs observed.
(52) Hay, P. J.; Ryan, R. R.; Salazar, K. V.; Wrobleski, D. A.;
Sattelberger, A. P. J. Am. Chem. Soc. 1986, 108, 313–15.
(53) Sternal, R. S.; Brock, C. P.; Marks, T. J. J. Am. Chem. Soc. 1985,
107, 8270–2.
Figure 7. Various computational models and molecular orbitals
that show a scandium-iron interaction for scandium alkyl com-
plexes (L ) fc(NSiH3)2).
(54) Sternal, R. S.; Marks, T. J. Organometallics 1987, 6, 2621–3.
(55) Dechy-Cabaret, O.; Martin-Vaca, B.; Bourissou, D. Chem. ReV.
2004, 104, 6147.

Scandium Alkyl Complexes
Organometallics, Vol. 27, No. 3, 2008 369
SiCH3). 13C NMR (126 MHz, C4D8O): δ 106.5, 69.1, 63.8, 29.1,
26.9, 0.65, -3.95.
Although there are two THF molecules coordinated to
scandium in 2-Me(THF)2 and only one in 2-(CH2Xy-
3,5)(THF), the first compound is a more active mediator for
L-lactide polymerization. Given the different coordination
geometries in the two compounds, it is difficult to interpret the
role that steric factors play. If only the electrophilicity of the
scandium center is considered, then the polymerization results
follow the DFT findings of scandium charges.
In conclusion, we have shown that Sc(CH2Xy-3,5)3(THF)2
is a useful starting material for the synthesis of ferrocene diamide
scandium complexes. In addition, a scandium dimethylbenzyl
compound was transformed into the corresponding THF-free
scandium chloride-bridged dimer, which was used subsequently
as a precursor in a reaction with neo-pentyl lithium. As a
consequence of the ligand architecture, the isolation of a salt-
free scandium neo-pentyl complex was not possible. Attempts
to remove the coordinated THF molecule from 2-(CH2Xy-
3,5)(THF) with AlMe3 led to the isolation of a scandium methyl
complex with two coordinated AlMe3 molecules; the formation
of a scandium methyl complex with two THF molecules
coordinated was observed when the trimethyl aluminum com-
plex was stirred in THF. DFT calculations indicate that if the
iron-scandium interaction becomes stronger, then the electro-
philicity of the scandium center decreases. We are currently
exploring ligand frameworks that are more conducive to the
formation of scandium monoalkyl, Lewis base-free complexes,
which would complement existing scandocene examples.
Synthesis of Sc(CH2Xy-3,5)3(THF)2. ScBr3(THF)3 (1.5 g, 2.99
mmol) was placed in a 100 mL round-bottom flask along with 80
mL of 2:1 THF/hexanes; this mixture was cooled to near freezing.
K(CH2Xy-3,5) (1.42 g, 8.98 mmol) was divided into two equal
portions; the first portion was added immediately and the solution
was warmed to 0 °C. After the orange color of the K(CH2Xy-3,5)
had faded to pale yellow (ca. 1 h), the second portion of K(CH2Xy-
3,5) was added and the mixture was stirred a further 2 h at 0 °C.
After this time the reaction mixture was filtered through Celite and
the solvent was removed. The resulting tan solid was dissolved in
toluene (20 mL), and hexanes (20 mL) was added. The solution
was cooled to -35 °C overnight, and feathery crystals were
observed in the morning. Yield: 0.962 g, 59%. Note: this reaction
does not scale up well, but reactions run in parallel can be combined
at the filtration step with no decrease in yield. When precipitated
1
from toluene with hexanes, H NMR spectroscopy shows that the
product is Sc(CH2Xy-3,5)3(THF)2.
1H NMR (300 MHz, C6D6): δ, ppm 6.52 (s, 6H, o-Ar-CH), 6.45
(s, 3H, p-Ar-CH), 3.49 (br s, 8H, OCH2CH2), 2.22 (s, 18H, Ar-
CH3), 2.10 (s, 6H, Ar-CH2), 1.16 (br s, 8H, OCH2CH2). 13C NMR
(75 MHz, C6D6): δ, ppm 151.3, 123.2, 120.9, 70.3, 25.4, 21.8. Anal.
Calcd for C35H49O2Sc: C, 76.89; H, 9.03. Found: C, 76.70, H, 8.73.
Synthesis of Sc(fc[NSi(t-Bu)Me2]2)(CH2Xy-3,5)(THF), 2-(CH2
Xy-3,5)(THF). Sc(CH2Xy-3,5)3(THF)2 (285 mg 0.521 mmol) and
fc[NHSi(t-Bu)Me2]2 (232 mg, 0.521 mmol) were each dissolved
in 5 mL of toluene and cooled to -35 °C. The solutions were
combined and stirred for 3 h at 0 °C. The solvent was removed,
hexanes was added, and the mixture was filtered through Celite.
After solvent removal the resulting solid was recrystallized from
Et2O/pentane. Yield: 283 mg, 80%.
1H NMR (500 MHz, C6D6): δ 6.94 (s, 2H, o-Ar-CH), 6.49 (s,
1H, p-Ar-CH), 4.09 (s, br, 2H, fc-CH), 3.85 (s, br, 2H, fc-CH),
3.76 (br, 4H, OCH2CH2), 3.29 (s, br, 2H, fc-CH), 3.15 (s, br, 2H,
fc-CH), 2.51 (s, 2H, benzyl), 2.32 (s, 6H, Ar-CH3), 1.32 (br, 4H,
OCH2CH2), 1.03 (s, 18H, SiC(CH3)3), 0.34–0.21 (br, 12H SiCH3).
13C NMR (126 MHz, C6D6): δ 149.2, 121.7, 103.6, 70.9, 69.9,
68.0, 67.8, 66.7, 55.6, 27.9, 25.1, 20.3, -1.8, -3.0. Anal. Calcd
for C35H57FeN2OScSi2: C, 61.93; H, 8.46. Found: C, 61.83, H, 8.15.
Synthesis of [Sc(fc[NSi(t-Bu)Me2]2)(µ-Cl)]2, (2-Cl)2. Compound
2-(CH2Xy-3,5)(THF) (415 mg, 0.611 mmol) and Me3SiCl (332
mg, 3.06 mmol) were each dissolved in 5 mL of toluene and cooled
to -78 °C. The two solutions were combined and stirred for 16 h
at room temperature. The solvent was removed by vacuum and
the resulting oily mixture was washed with hexanes, yielding a
yellow solid, which was toluene soluble. This solid was dissolved
in toluene:pentane and crystals were observed after 3 days. Yield:
crude, 315 mg, 99%; crystals, 197 mg, 61%.
1H NMR (300 MHz, C6D6): δ 4.06 (s, br, 8H, fc-CH), 3.87 (s,
4H, fc-CH), 1.09 (s, 36H, SiC(CH3)3), 0.41 (s, 24H, SiCH3). 13C
NMR (126 MHz, C7D8): δ 136.8, 106.7, 70.2, 27.8, 20.0, -2.1.
Anal. Calcd for C30H54FeN2O2ScClSi2 (this compound was submit-
ted for analysis as the THF adduct, since the adduct is more
thermally stable than the chloride dimer): C, 54.00; H, 8.16. Found:
C, 54.00, H, 7.88.
Synthesis of Li[Sc(fc[NSi(t-Bu)Me2]2)(Np)2], Li[2-Np2]. NpLi
(11.5 mg, 0.147 mmol) and (2-Cl)2 (70.0 mg, 0.067 mmol) were
each cooled to freezing in 2 mL of hexanes and combined
immediately upon thawing. The mixture was stirred at -78 °C for
3 h. The solution was filtered in hexanes and crystals were grown
from a concentrated pentane solution. Yield: 15 mg, 32%. Reactions
run with freshly recrystalized (2-Cl)2 at -78 °C constantly yielded
one major product.
Experimental Section
All experiments were performed under a dry nitrogen atmosphere
using standard Schlenk techniques or an MBraun inert-gas glove-
box. Solvents were purified using a two-column solid-state purifica-
tion system by the method of Grubbs58 and transferred to the
glovebox without exposure to air. NMR solvents were obtained
from Cambridge Isotope Laboratories, degassed, and stored over
activated molecular sieves prior to use. Scandium oxide was
purchased from Stanford Materials Corporation (Aliso Viejo, CA)
and used as received. KBz59 and fc[NHSi(t-Bu)Me2]2 (fc ) 1,1′-
ferocenylene)29 were prepared following published procedures.
Other chemicals were used as received. 1H NMR spectra (this
material is based on work supported by NSF grant CHE-9974928)
were recorded on Bruker 300 or Bruker 500 spectrometers at room
temperature in C6D6 or CDCl3 unless otherwise specified. Chemical
shifts are reported with respect to internal solvent, 7.16 ppm (C6D6).
CHN analyses were performed by Schwarzkopf Microanalytical
Laboratory (Woodside, NY), Desert Analytics (Tucson, AZ), and
UC Berkeley Micro-Mass Facility (Berkeley, CA).
Synthesis of [K(solvent)][Sc(fc[NSi(t-Bu)Me2]2)2], 1-K(sol-
vent) (solvent ) THF or toluene). ScCl3(THF)3 (50.8 mg, 0.1382
mmol) and K2fc[NSi(t-Bu)Me2]2(OEt2)2 (185 mg, 0.276 mmol) were
each dissolved in ca. 2 mL of THF and cooled to freezing. The
solutions were combined slowly, and the resulting mixture was
stirred for 3 h. The reaction mixture was filtered through Celite,
and the THF was removed by vacuum. The remaining solid was
dissolved in toluene, layered with pentane, and cooled to -35 °C.
Crystals were isolated after 2 days. Yield: 0.080 g, 56%.
1H NMR (500 MHz, C4D8O): δ, ppm: 3.94 (br s, 8H, fc-CH),
3.65 (br s, 8H, fc-CH), 0.643 (s, 36H, SiC(CH3)3), 0.416 (s, 24H,
(56) Ma, H.; Spaniol, T. P.; Okuda, J. Dalton Trans. 2003, 4770–4780.
(57) Ma, H.; Spaniol, T. P.; Okuda, J. Angew. Chem., Int. Ed. 2006,
45, 7818–7821.
(58) Pangborn, A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518–1520.
(59) Bailey, P. J.; Coxall, R. A.; Dick, C. M.; Fabre, S.; Henderson,
L. C.; Herber, C.; Liddle, S. T.; Lorono-Gonzalez, D.; Parkin, A.; Parsons,
S. Chem.-Eur. J. 2003, 9, 4820–4828.
1H NMR (300 MHz, C6D6): δ 3.95 (s, 4H, fc-CH), 3.23 (s, 4H,
fc-CH), 1.19 (s, 18H, SiC(CH3)3), 0.92 (s, 18H, CH2C(CH3)3), 0.89

370 Organometallics, Vol. 27, No. 3, 2008
CarVer et al.
(s, 4H, CH2C(CH3)3), 0.51 (s, 12H, SiCH3). 13C NMR (75 MHz,
C6D6): δ 104.3, 68.3, 67.2, 35.9, 31.6, 27.7, 20.7, -1.9.
Synthesis of Sc(fc[NSi(t-Bu)Me2]2)Me(AlMe3)2, 2-Me-
(AlMe3)2. Compound 2-(CH2Xy-3,5)(THF) (400 mg, 0.589 mmol)
was divided into two equal portions and each portion was dissolved
in hexanes (10 mL). To each portion was added AlMe3 (100 mg,
1.38 mmol, 4.68 equiv). The reaction mixture was stirred 3 h at
room temperature. The solutions were combined, reduced in volume
until precipitate began to form, and filtered through Celite. The
filtrate was cooled to -35 °C overnight, and yellow crystals were
observed in the morning. Yield: 314 mg, 82%.
Computational Details. The Amsterdam Density Functional
(ADF) package (version ADF2006.01) was used to do a full
geometry optimization on Cartesian coordinates of the model
compounds specified in the text. For the scandium, iron, and silicon
atoms, standard triple-ꢀ STA basis sets from the ADF database
TZP were employed with 1s-3p (Sc, Fe) and 1s-2s (Si) electrons
treated as frozen cores. For all the other elements, standard double-ꢀ
STA basis sets from the ADF database DZP were used, with the
1s electrons treated as frozen core for non-hydrogen atoms. The
local density approximation (LDA) by Vosko, Wilk, and Nusair
(VWN) was used together with the exchange and correlation
corrections that are used by default by the ADF2006.01 program
suite.
1H NMR (300 MHz, C6D6): δ 3.91 (br, 4H, fc-CH), 3.58 (br,
4H, fc-CH), 0.95 (s, br, 18H, SiC(CH3)3), 0.30–0.50 (br, 21H), 0.25
(s, br, 12H, SiCH3). Anal. Calcd for C29H59Al2FeN2Si2Sc.: C, 53.85;
H, 9.19; N, 4.33. Found: C, 53.47; H, 9.49%; N, 4.26.
For the complex [fc(NSiH3)2]ScMe two different geometry
optimizations gave the same result: in the first one, the coordinates
were imported from the optimization of [fc(NSiH3)2]ScMe(OMe2)
geometry and then the OMe2 fragment was removed; for the latter,
the CScFe angle was modified to start at 90°.
L-Lactide Polymerization Details. In the glovebox, L-lactide
(0.2506 g, recrystallized from toluene) was transferred into a 25
mL Schlenk tube equipped with a magnetic stirrer. The L-lactide
was dissolved in the proper volume of benzene to make the final
L-lactide concentration 0.223 M and the final volume 7.8 mL. Then,
the proper amount of catalyst (in a benzene solution) was rapidly
added and the Schlenk tube was removed from the glovebox and
placed in a vigorously stirring 70 °C bath. After the reaction was
completed, the polymer was precipitated twice from dichlo-
romethane with cold methanol and dried overnight under vacuum.
Synthesis of Sc(fc[NSi(t-Bu)Me2]2)Me(THF)2, 2-Me(THF)2.
Compound 2-Me(AlMe3)2 (289.2 mg, 0.4471 mmol) was dissolved
in THF (5 mL) and stirred 1 h at room temperature. The solvent
was removed and the reaction mixture was extracted with pentane
and filtered through Celite. Crystals were grown from Et2O/pentane.
Yield: 232.8 mg, 80%.
1H NMR (300 MHz, C6D6): δ 4.02 (s, 4H, fc-CH), 3.69 (s, br,
8H, OCH2CH2), 0.95 (s, 4H, fc-CH), 1.24 (s, br, 8H, OCH2CH2),
1.06 (s, 18H, SiC(CH3)3), 0.31 (s, 12H, SiCH3), -0.38 (s, 3H,
ScCH3). Anal. Calcd for C31H57FeN2O2Si2Sc: C, 57.56; H, 8.88;
N, 4.33. Found: C, 57.72; H, 8.78; N, 4.60.
X-ray Crystal Structures. General Considerations. X-ray
quality crystals were obtained from various concentrated solutions
placed in a -35 °C freezer in the glovebox. Inside the glovebox,
the crystals were coated with oil (STP Oil Treatment) on a
microscope slide, which was brought outside the glovebox. The
X-ray data collections were carried out on a Bruker AXS single-
crystal X-ray diffractometer using MoKR radiation and a SMART
APEX CCD detector. The data are reduced by SAINTPLUS and
an empirical absorption correction is applied using the package
SADABS. The structure was solved and refined using SHELXTL
(Brucker 1998, SMART, SAINT, XPREP, and SHELXTL, Brucker
AXS Inc., Madison, WI). Tables with atomic coordinates and
equivalent isotropic displacement parameters, with all the bond
lengths and angles, and with anisotropic displacement parameters
are listed in the cif files.
Acknowledgment. The authors thank Dr. Saeed Khan for
help with crystallography and Prof. Heather Maynard for the
use of the GPC instrument. This work was supported by the
UCLA and the UC Energy Institute (EST grant).
Supporting Information Available: Experimental details for
compound syntheses, full crystallographic descriptions (as cif), and
computational details are available free of charge via the Internet
at http://pubs.acs.org.
OM7007277