Bis(permethylcyclopentadienyl)aluminum compounds: Precursors to [Cp*2Al]+ but not to Cp*3Al

Article abstract of DOI:10.1021/om000173x

Replacement of only two chloride ligands occurs when AlCl₃ is reacted with 3 equiv of either Cp*Na or Cp*K, forming Cp*₂AlCl (1) [Cp* = C₅Me₅]. This compound dissociates a chloride ion in polar solvents such as methylene chloride. Clean formation of decamethyl-aluminocenium salts could not be accomplished by abstraction of a chloride ligand from 2 with silver salts, LiB(C₆F₅)₄, or NaB(C₆F₅)₄. However, reaction of Cp*₂AlMe (2) with B(C₆F₅)₃ cleanly affords the salt [Cp*₂Al]⁺[B(C₆F₅) ₃Me]^- (3). X-ray crystal structures of both 2 and 3 are reported. In the solid state, 2 exhibits a bis-dihapto ring coordination similar to that of (η²-C₅H₅)₂AlMe. The molecular structure of 3 shows complete methyl anion abstraction from the aluminum. The decamethylaluminocenium salt is more stable in solution than its nonmethylated analogue; however, it is also considerably less active as an initiator for the cationic polymerization of isobutene.

Full text of DOI:10.1021/om000173x

Organometallics 2000, 19, 3361-3367
3361
Bis(p er m eth ylcyclop en ta d ien yl)a lu m in u m Com p ou n d s:
P r ecu r sor s to [Cp *2Al]+ bu t Not to Cp *3Al
Christopher T. Burns,† Pamela J . Shapiro,*,† Peter H. M. Budzelaar,‡
Roger Willett,§ and Ashwani Vij‡
Department of Chemistry, University of Idaho, Moscow, Idaho 83844-2343, Inorganic
Chemistry, University of Nijmegen, P.O. Box 9010, 6500 GL, Nijmegen, The Netherlands, and
Department of Chemistry, Washington State University, Pullman, Washington 99164-4630
Received February 23, 2000
Replacement of only two chloride ligands occurs when AlCl3 is reacted with 3 equiv of
either Cp*Na or Cp*K, forming Cp*2AlCl (1) [Cp* ) C5Me5]. This compound dissociates a
chloride ion in polar solvents such as methylene chloride. Clean formation of decamethyl-
aluminocenium salts could not be accomplished by abstraction of a chloride ligand from 2
with silver salts, LiB(C6F5)4, or NaB(C6F5)4. However, reaction of Cp*2AlMe (2) with B(C6F5)3
cleanly affords the salt [Cp*2Al]+[B(C6F5)3Me]- (3). X-ray crystal structures of both 2 and 3
are reported. In the solid state, 2 exhibits a bis-dihapto ring coordination similar to that of
(η2-C5H5)2AlMe. The molecular structure of 3 shows complete methyl anion abstraction from
the aluminum. The decamethylaluminocenium salt is more stable in solution than its
nonmethylated analogue; however, it is also considerably less active as an initiator for the
cationic polymerization of isobutene.
In tr od u ction
num and one of its rings to form (η1-C5Me4H)2Al{C-
(dNtBu)C(dNtBu)(C5Me4H)} (eq 1).8b Similar reactivity
Besides its importance in transition metal chemistry
and homogeneous catalysis, the cyclopentadienyl ligand,
in its various guises, has also played a prominent role
in the advancement of main group metal and nonmetal
chemistry. The ability of sterically encumbered cyclo-
pentadienyl ligands such as the pentamethylcyclopen-
tadienyl ligand and other polysubstituted cyclopenta-
dienyl ligands to support unusual oxidation states, low
molecularities, and low coordination numbers has al-
lowed chemists to demonstrate interesting, new modes
of bonding and reactivity among the main group
elements.1-6
(1)
In our work with cyclopentadienylaluminum com-
pounds, we have found that increasing the steric bulk
of the cyclopentadienyl rings through the introduction
of methyl substituents influences the manner in which
the rings coordinate to the aluminum.7 More intriguing,
however, is the effect of steric factors on the reactivity
of the cyclopentadienyl-aluminum linkages. Whereas
Cp3Al merely coordinates tert-butyl isocyanide,8a even
at elevated temperatures, (C5Me4H)3Al inserts two
molecules of tert-butyl isocyanide between the alumi-
between (1,2,4-Me3C5H2)3Al and tert-butyl isocyanide
1
has been identified by H NMR.9 Given the enhanced
(3) (a) J utzi, P.; Moehrke, A. Angew. Chem. 1990, 102, 913. (b) J utzi,
P.; Bunte, E. A. Angew. Chem., Int. Ed. Engl. 1992, 31, 1605. (c) J utzi,
P.; Kroos, R.; Mueller, A.; Bogge, H.; Penk, M. Chem. Ber. 1991, 124,
75. (d) J utzi, P.; Meyer, U.; Opiela, S.; Neumann, B.; Stammler, H.-G.
J . Organomet. Chem. 1992, 439, 279. (e) J utzi, P.; Bunte, E.-A.;
Holtmann, U.; Neumann, B.; Stammler, H.-G. J . Organomet. Chem.
1993, 446, 139. (f) J utzi, P.; Brusdeilins, N.; Stammler, H. G.;
Neumann, B. Chem. Ber. 1994, 127, 997. (g) J utzi, P.; Eikenberg, D.;
Neumann, B.; Stammler, H.-G. Organometallics 1996, 15, 3659. (h)
J utzi, P.; Eikenberg, D.; Moehrke, A.; Neumann, B.; Stammler, H.-G.
Organometallics 1996, 15, 753. (i) J utzi, P.; Eikenberg, D.; Bunte, E.-
A.; Moehrke, A.; Neumann, B.; Stammler, H.-G. Organometallics 1996,
15, 1930. (j) J utzi, P.; Neumann, B.; Reumann, G.; Stammler, H.-G.
Organometallics 1998, 17, 1305. (k) J utzi, P.; Neumann, B.; Reumann,
G.; Stammler, H.-G. Organometallics 1999, 18, 2037.
(4) (a) Schulz, S.; Roesky, H. W.; Koch, H. J .; Sheldrick, G. M.;
Stalke, D.; Kuhn, A. Angew. Chem., Int. Ed. Engl. 1993, 32, 1729. (b)
Schulz, S.; Andruh, M.; Pape, T.; Heinze, T.; Roesky, H.; Haeming, L.;
Kuhn, A.; Herbst-Irmer, R. Organometallics 1994, 13, 4004. (c) Schulz,
S.; Ha¨ming; Herbst-Irmer, R.; Roesky, H. W.; Sheldrick, G. M. Angew.
Chem., Int. Ed. Engl. 1994, 33, 969.
† University of Idaho.
‡ University of Nijmegen.
§ Washington State University.
(1) J aniak, C.; Schumann, H. Adv. Organomet. Chem. 1991, 33,
291-393.
(2) (a) Dohmeier, C.; Schno¨ckel, H.; Robl, C.; Schneider, U.; Ahlrichs,
R. Angew. Chem., Int. Ed. Engl. 1993, 32, 1655. (b) Dohmeier, C.;
Krautscheid, H.; Schno¨ckel, H. Angew. Chem., Int. Ed. Engl. 1994,
33, 2482. (c) Sitzmann, H.; Lappert, M. F.; Dohmeier, C.; Uffing, C.;
Schno¨ckel, H. J . Organomet. Chem. 1998, 561, 203. (d) von Ha¨nisch,
C. K. F.; U¨ ffing, C.; J unker, M. A.; Ecker, A.; Kneisel, B. O.; Schno¨ckel,
H. Angew. Chem., Int. Ed. Engl. 1996, 35, 2875. (e) Loos, D.; Baum,
E.; Ecker, A.; Schno¨ckel, H. Angew. Chem., Int. Ed. Engl. 1997, 36,
860. (f) Uffing, C.; Baum, E.; Koppe, R.; Schno¨ckel, H. Angew. Chem.,
Int. Ed. Engl. 1998, 37, 2397. (g) Qing, Y.; Purath, A.; Donchev, A.;
Schno¨ckel, H. J . Organomet. Chem. 1999, 584, 94.
(5) Weiss, J .; Stetzkamp, D.; Nuber, B.; Fischer, R. A.; Boehme, C.;
Frenking, G. Angew. Chem., Int. Ed. Engl. 1997, 36, 70.
10.1021/om000173x CCC: $19.00 © 2000 American Chemical Society
Publication on Web 07/19/2000

3362 Organometallics, Vol. 19, No. 17, 2000
Burns et al.
reactivities exhibited by these polymethylated, homo-
leptic cyclopentadienylaluminum compounds, we were
interested in preparing and examining the properties
of the most highly methylated species possible, Cp*3Al
(Cp* ) C5Me5).
Since Cp*3Ga had been prepared previously by the
reaction of KCp* and GaCl3,10 we expected to be able to
prepare the aluminum analogue in a similar manner;
however, we have found aluminum to be resistant to
the coordination of three Cp* rings. At best, we have
coordinated two Cp* ligands to the aluminum in the
compounds Cp*2AlCl (1) and Cp*2AlMe (2). The reactiv-
ity of these Cp*2Al compounds offers some clues as to
why aluminum does not accommodate a third Cp* ring.
We believe that, rather than coordinating a third
cyclopentadienyl anion, the aluminocenium cation pre-
fers to adopt the more stable, bis-η5 sandwich structure
first observed in 1993 by Schno¨ckel and co-workers for
[Cp*2Al][Cp*AlCl3].2a The results that led us to this
conclusion are described herein along with the syntheses
of 1 and 2, an X-ray crystal structure determination of
2, and the synthesis and structure of the permethyla-
luminocenium salt [Cp*2Al][MeB(C6F5)3] (3), which is
considerably less active than its nonmethylated alumi-
nocenium analogue in the initiation of cationic olefin
polymerization.
Resu lts
Syn th esis a n d Solu tion Beh a vior of Cp *2AlCl.
Attempts to synthesize Cp*3Al by reacting 1.5 molar
equiv of decamethylmagnesocene with AlCl3 in a man-
ner analogous to that used for the synthesis of Cp3Al
(Cp ) C5H5) produced a mixture of products which
included unreacted Cp*2Mg, [Cp*2AlCl]x, and at least
two other unidentified Cp*-containing species. To have
a greater thermodynamic driving force for the exchange
of Cl with Cp* on the aluminum, Cp*Na and Cp*K were
tried, especially since Cp*K had been used by Schu-
mann and co-workers to prepare Cp*3Ga from GaCl3.10
Instead of forming Cp*3Al, 3 molar equiv of either
Cp*Na or Cp*K reacted with AlCl3 to produce [Cp*2-
AlCl]x, 1, exclusively, and 1 equiv of unreacted Cp*Na
or Cp*K was recovered. Compound 1 was prepared
previously by Roesky and co-workers by the reaction of
[Cp*AlCl2]2 with 2 equiv of Cp*K.11 A dimeric structure
of the compound was assumed by the authors, in
analogy with the structures of [Cp*AlCl2]2 and [Cp*-
(Ph)AlCl]2. We have so far been unable to establish the
molecularity of this compound either by molecular
weight determination or by crystallographic methods.
The tendency of 1 to dissociate a chloride ligand,
particularly in more polar solvents, is revealed in its
27Al NMR solution spectra in C6D6, CDCl3, and CD2Cl2
(Figure 1). Qualitatively, the intensity of the 27Al signal
F igu r e 1. (a) 27Al NMR spectrum of 1 in C6D6 at 297K.
(b) 27Al NMR spectrum of 1 in CDCl3 at 297 K. (c) 27Al NMR
spectrum of 1 in CD2Cl2 at 263 K. The asterisk marks the
signal from the NMR instrument probe.
at δ -114 due to [Cp*2Al]+ increases with increasing
solvent polarity. The 27Al NMR spectrum of a solution
of 1 in C6D6 at 297 K exhibits a strong, broad signal at
δ -3 (W1/2 ≈ 1400 Hz). A weak signal at δ -114 due to
[Cp*2Al]+ is also visible just above the baseline noise
(Figure 1a). By contrast, a solution of 1 in CD2Cl2
exhibits an intense 27Al signal at δ -114. The broad,
comparatively weaker, downfield signal, originally at δ
-3, broadens and disappears altogether into the shoul-
der of the signal from the NMR probe (Figure 1c). We
(6) Cowley, A. H.; Lomeli, V.; Voigt, A. J . Am. Chem. Soc. 1998,
120, 6401.
(7) Fisher, J . D.; Budzelaar, P. H. M.; Shapiro, P. J .; Staples, R. J .;
Yap, G. P. A.; Rheingold, A. L. Organometallics 1997, 16, 871.
(8) (a) Fisher, J . D.; Wei, M.-Y.; Willett, R.; Shapiro, P. J . Organo-
metallics 1994, 13, 3324-3329. (b) Shapiro, P. J .; Vij, A.; Yap, G. P.
A.; Rheingold, A. L. Polyhedron 1995, 14, 203.
(9) Burns, C. Unreported results.
(10) Schumann, H.; Nickel, S.; Weimann, R. J . Organomet. Chem.
1994, 468, 43.
(11) Schulz, S.; Roesky, H. W.; Noltemeyer, M.; Schmidt, H.-G. J .
Organomet. Chem. 1995, 493, 69.

Bis(permethylcyclopentadienyl)aluminum Compounds
Organometallics, Vol. 19, No. 17, 2000 3363
1
Solvent effects are also observed in H NMR spectra
of 1 in C6D6 and CD2Cl2. Whereas there is only one
resonance at δ 1.8 for a C6D6 solution of 1 at 297 K,
there are two signals, a major peak at δ 1.7, and a broad,
less intense signal at δ 2.1 (Figure 2a) in a ratio of 8:1,
respectively, for a solution of 1 in CD2Cl2. As the CD2-
Cl2 sample is cooled, the intensity of the major peak
decreases and the signal shifts upfield to δ 1.5 at 218 K
(Figure 2c). The minor peak also shifts upfield slightly,
and its intensity increases until at 218 K it has the same
area as the δ 1.5 resonance.
We attribute the δ 2.1 resonance to the aluminoce-
nium cation since its chemical shift is similar to that of
[Cp*2Al][MeB(C6F5)3] in CD2Cl2 (vide infra). As can be
seen in the 1H NMR spectrum at 248 K, this signal
splits (Figure 2b) and then collapses back to a single
1
peak at 218 K. This splitting is also observed in the H
NMR spectrum of [Cp*2Al][MeB(C6F5)3] described be-
low.
We attribute the upfield resonance to [Cp*2AlCl2]-
since a 1H NMR spectrum of a CD2Cl2 solution of 1
combined with an equivalent of [PPN]Cl (bis(triph-
enylphosphoranylidene)ammonium chloride) exhibits a
1
resonance at δ 1.6. Thus, the variable-temperature H
and 27Al NMR spectra of 1 in CD2Cl2 indicate the
ionization equilibrium shown in eq 2.
(2)
Ligand redistribution also takes place, as indicated
by our isolation of X-ray quality crystals of [Cp*2Al]-
[Cp*AlCl3] from a methylene chloride solution of 1
layered with pentane. The structure of this salt has been
reported previously by Schno¨ckel and co-workers.2a
Efforts to introduce the less bulky tetramethylcyclopen-
tadienyl ligand by reacting Cp*2AlCl with Na(C5Me4H)
produced an ill-defined mixture of species, most likely
resulting from cyclopentadienyl ligand redistribution.
Syn th esis a n d Molecu la r Str u ctu r e of Cp *2AlMe.
In analogy with the synthesis of 1, MeAlCl2 was reacted
with 2 equiv of Cp*Na to afford Cp*2AlMe (2) as a white
microcrystalline solid. An X-ray crystal structure de-
termination revealed that 2 adopts a bis-dihapto coor-
dination geometry in the solid state akin to the less
crowded Cp2AlMe.8a Crystallographic data for 2 are
listed in Table 1, and selected bond distances and angles
are listed in Table 2. The molecule lies on two crystal-
lographic mirror planes, and a disordering of the rings
about one of the mirror planes was indicated by very
large atomic displacement parameters on the cyclopen-
tadienyl methyl groups and an unlikely C-C distance
of 1.19 Å between the cyclopentadienyl carbons coordi-
nated to the aluminum when an ordered model was
used. Introduction of a disorder consisting of rotation
of the rings by (16° improved the refinement, although
the displacements of the ring methyl groups remained
quite large, as can be seen from the ORTEP drawing of
one orientation in the disorder in Figure 3a. The
disorder of the rings appears to be driven by steric
factors arising from nonbonded interactions between
methyl groups on carbons C(4) and C(5) on opposite Cp*
rings. This can best be seen from the ball-and-stick view
of the structure in Figure 3b. The steric strain experi-
enced by those methyl groups is also manifested in their
degree of bending away from the cyclopentadienyl ring
F igu r e 2. 1H NMR spectra of 1 in CD2Cl2 at (a) 297 K,
(b) 258 K, (c) 218 K. The asterisked peak is a toluene
impurtity in the sample, which was used as a internal
standard to ensure that the net integration of the Cp*
peaks remained constant as the temperature of the sample
was lowered.
believe that this shifting of the downfield signal is
associated with the dynamics of chloride ion exchange
between [Cp*2AlCl]x and [Cp*2AlCl2]- (eq 2) and that
the 27Al NMR resonance for [Cp*2AlCl2]- must be
obscured by the probe signal.

3364 Organometallics, Vol. 19, No. 17, 2000
Burns et al.
η3, and η1 ring coordination geometries in the solid
state4b,11-14 and in solution.15
As is typical with cyclopentadienylaluminum com-
1
pounds, very simple H and 13C NMR solution spectra
are exhibited by 2 due to its fluctionality. Two signals
1
are observed in the H NMR spectrum of 2 in C6D6 at δ
1.92 and -1.58 for the Cp* ring methyls and the Al-
bound methyl group, respectively. The slight upfield
shift of the Al-Me in 2 relative to Cp2AlMe can be
attributed to the greater electron-releasing properties
of Cp* relative to Cp. Signals for only the quarternary
and primary cyclopentadienyl ring carbons are observed
in the 13C NMR spectrum of 2. As with Cp2AlMe,16
a
13C signal for the methyl group of 2 is not observed,
presumably due to extreme broadening by the quadru-
polar Al nucleus (I ) 5/2).
Syn th esis a n d Molecu la r Str u ctu r e of [Cp *2Al]-
[B(C6F 5)3Me]. Given the tendency of 1 to dissociate a
chloride ligand, we sought to convert it to a permethyl-
aluminocenium salt by replacement of the chloride
ligand with a noncoordinating anion. Although the
permethylaluminocenium cation can be generated by
17
reacting 1 with either LiB(C6F5)4 or NaB(C6F5)4, the
reaction is accompanied by considerable decomposition,
and we have so far been unable to isolate the perm-
ethylaluminocenium salt cleanly in this manner. Similar
problems with decomposition were encountered with the
silver salts AgBF4 and AgOTf.
As an alternative approach to the permethylalumi-
nocenium cation, we followed Bochmann’s example for
generating Cp2Al+ from Cp2AlMe16 and reacted 2 with
B(C6F5)3 (eq 3). The reaction proceeds cleanly to afford
F igu r e 3. (a) ORTEP drawing of the molecular structure
of (η2-C5Me5)2AlCH3 (2). Thermal ellipsoids are at 30%
probability. (b) Ball-and-stick view of 2 showing staggered
arrangement of Cp* rings.
(3)
plane, with C(4)-C(5)-C(4′)-C(5′) forming a dihedral
angle of 18.6° to C(4)-C(5)-C(3)-C(6). A dihedral angle
of 60.4° between the two cyclopentadienyl ring planes
is created by their bis-dihapto geometry. The cyclopen-
tadienyl rings are not completely planar, however, with
a maximum deviation of 0.081 Å from the plane. The
distortion in the ring plane arises from a folding of C(2)
away from the Al-CH3 region, forming an angle of 12°
with a fold along C(3) and C(6). There are also large
deviations among the C-C bond lengths within the ring,
C(2)-C(3) (1.37(3) Å) and C(5)-C(6) (1.37(2) Å) being
substantially shorter than the other C-C bonds in the
ring as if to form a more localized 1,3-diene structure.
Ab initio calculations at the RHF/6-31G* level on the
model complexes Cp*AlH2 and Cp*AlMe2 indicate η2
ground-state geometries, with shortest Al-C(ring) dis-
tances of 2.128 and 2.132 Å for the former compound
and 2.140 and 2.146 Å for the latter, in close agreement
with corresponding Al-C(4) distance of 2.143(13) Å in
the molecular structure of 2. Nevertheless, as has been
previously shown for the (C5H5)Al complexes, the pref-
erence for an η2 structure is very slight, and calculations
at the RHF/3-21G(*) level show that slippage of the Cp*
ring to other hapticities should be extremely facile, with
barriers from 2 to 6 kcal/mol.7 Indeed, other peralkyl-
cyclopentadienylaluminum(III) compounds exhibit η5,
[Cp*2Al][MeB(C6F5)3] (3), which is more stable in solu-
tion than its (C5H5) analogue. By contrast, reaction of
2 with [Ph3C][B(C6F5)4]18 produces Ph3CH, presumably
via hydrogen abstraction from a Cp* ring. The aluminum-
containing product of this reaction has not been identi-
fied. Over time it converts to the aluminocenium cation;
however, the reaction is not clean.
The 500 MHz 1H NMR spectrum of 3 in CDCl3 at 297
K exhibits a splitting in the signal for the Cp* hydro-
gens. This persists as the sample is heated to 50 °C but
collapses to a single peak when the sample is cooled to
-20 °C, behavior that is similar to that of the alumi-
(12) Schonberg, P. R.; Paine, R. T.; Campana, C. F.; Duesler, E. N.
Organometallics 1982, 1, 799.
(13) Koch, H.-J .; Schulz, S.; Roesky, H. W.; Noltemeyer, M.; Schmidt,
H.-G.; Heine, A.; Herbst-Irmer, R.; Stalke, D.; Sheldrick, G. M. Chem.
Ber. 1992, 125, 1107.
(14) J utzi, P.; Dalhaus, J .; Neumann, B.; Stammler, H.-G. Organo-
metallics 1995, 15, 747.
(15) Scherer, M.; Kruck, T. J . Organomet. Chem. 1996, 513, 135.
(16) Bochmann, M.; Dawson, D. M. Angew. Chem., Int. Ed. Engl.
1996, 35, 2226.
(17) Rhodes, B.; Chien, J . C. W.; Rausch, M. D. Organometallics
1998, 17, 1931.
(18) Chien, J . C. W.; Tsai, W.-M.; Rausch, M. D. J . Am. Chem. Soc.
1991, 113, 8570.

Bis(permethylcyclopentadienyl)aluminum Compounds
Organometallics, Vol. 19, No. 17, 2000 3365
Å. No interactions between the fluorine atoms and the
aluminum centers are apparent either, the shortest
Al-F distance being 5.07 Å for Al(2)-F(30).
Compound 3 can be stored for months at -17 °C
under an N2 atmosphere with no apparent decomposi-
tion. Over days at room temperature it turns dark
purple, and there is evidence for the elimination of Cp*H
in the 1H NMR spectrum. Unlike its Cp analogue, which
decomposes readily in CD2Cl2 solution upon warming
to 20 °C, 3 is stable at room temperature in CDCl3 for
months, exhibiting a gradual purple discoloration over
time but little accompanying decomposition. In addition
to being more stable than its Cp analogue, compound 3
is also considerably less active as an initiator for the
cationic polymerization of isobutene and isoprene.
Whereas [Cp2Al][MeB(C6F5)3] is most active at initiating
isobutene polymerization in CH2Cl2 at -78 °C and is
less active at higher temperatures due to decomposition
of the catalyst,16 3 is essentially inactive at -78 °C but
initiates cationic olefin polymerization at room temper-
ature. Reaction of 10 mL of isobutene with approxi-
mately 50 µg of 3 in 5 mL of CH2Cl2 overnight at room
temperature afforded 2.1 g of polymer with Mn ) 8000
and Mw/Mn ) 2.2. Catalyst decomposition during the
course of the polymerization reaction was indicated by
the development of a purple tinge to the reaction over
time. Essentially no polymer was obtained when the
reaction was quenched after 20 min. Carrying out the
reaction under similar conditions, but instead with
approximately 40 mg of [Cp*2AlCl]x (1) as catalyst,
afforded only 0.04 g of polymer, consistent with fact that
1 dissociates only partially to the permethylaluminoce-
nium cation in methylene chloride.
F igu r e 4. ORTEP drawing of the molecular structure of
[(η5-C5Me5)2Al]+[MeB(C6F5)3]-. Thermal ellipsoids are at
30% probability. Hydrogen atoms are omitted for clarity.
nocenium cation generated from the dissociation of
[Cp*2AlCl]x in CD2Cl2 (Figure 2b, vide supra). No
changes are observed in the borate methyl resonance
at δ 0.48 in this temperature range; nor are there any
peculiarities or changes in the 19F and 27Al NMR spectra
of the sample over this temperature range. We are
currently investigating solvent effects on this mysteri-
1
ous splitting in the Cp* H NMR signal of 3. There is
no evidence in the variable-temperature 19F NMR
spectra for ion pairing or coupling to a fluorine atom of
the borate anion. We suspect that the splitting is caused
by slippage of one of the Cp* rings from its η5-hapticity
and that at lower temperature the aluminocenium
cation assumes the more stable bis-pentahapto sand-
wich geometry exhibited in its crystal structure (vide
infra). Further experiments are underway to explore
this hypothesis.
Discu ssion a n d Con clu sion s
The 27Al NMR spectrum of 3 in CDCl3 exhibits a
single, sharp signal at δ -114, and the 11B NMR
spectrum exhibits a signal at δ -34 vs H3BO3, consis-
tent with the formulation of the compound as an ion
pair. X-ray quality crystals of 3 were obtained from a
methylene chloride solution of the compound layered
with pentane and cooled at -78 °C. Crystallographic
data are listed in Table 1, and selected bond lengths
and angles are listed in Table 3. The asymmetric unit
consists of two independent aluminocenium cations,
with each aluminum atom lying on an inversion center,
and a borate counteranion. An ORTEP drawing of the
structure of one ion pair is shown in Figure 4. The
structure of the permethylaluminocenium cation, with
its 180° (ring-centroid)-Al-(ring-centroid) angle and
36° staggering of the Cp* rings, is similar to that found
by Schno¨ckel and co-workers for [Cp*2Al][Cp*AlCl3] and
is of higher symmetry than that found for the cation
in [Cp*2Al]+[{Ph(Me)B(η5-C5H4)2}ZrCl2]-, another per-
methylaluminocenium salt that we recently reported.19
The B(1)-C(21) bond length of 1.650(7) Å is comparable
to that of other free [MeB(C6F5)3]- counteranions that
have been structurally characterized.20 There is no
evidence for residual Al- - - Me interaction in the struc-
ture, the closest Al-C(21) methyl) contact being 5.95
We have commented previously on the delicate bal-
ance between ionicity, π-bonding, and steric repulsion
that is manifested in the preferred hapticities of cyclo-
pentadienylaluminum complexes.7,21 Changing the num-
ber of methyl substituents on the cyclopentadienyl ring
or changing the electron-withdrawing/donating proper-
ties of additional ligands on the aluminum has been
shown to tip the balance from primarily covalent
σ-bonding (η1), to covalent π-bonding (η2, η3, η5), to a
primarily ionic interaction between the aluminum atom
and its cyclopentadienyl ring (η5). Even more remark-
able are the effects of methyl substituents on the
cyclopentadienyl ring on the reactivity of the cyclopen-
tadienyl-aluminum linkage as well as its ability to even
form.
It appears that our inability to prepare Cp*3Al using
methods that were successful for the synthesis of (C5-
Me4H)3Al and Cp*3Ga is associated with the preference
of [Cp*2Al]+ for a bis-pentahapto sandwich geometry,
making addition of a third permethylcyclopentadienyl
anion thermodynamically unfavorable in the case of
Cp*2AlCl. Consistent with this idea is Schno¨ckel’s
reaction of Cp*Al with AlCl3 to produce [Cp*2Al]-
[Cp*AlCl3], presumably via a Cp*3Al intermediate.2a
Further support comes from chemistry recently reported
by J utzi and co-workers, in which decamethylsilicocene
(19) Burns, C. T.; Stelck, D. S.; Shapiro, P. J .; Vij, A.; Kunz, K.; Kehr,
G.; Concolino, T.; Rheingold, A. L. Organometallics 1999, 18, 5432.
(20) Yang, X.; Stern, C. L.; Marks, T. J . J . Am. Chem. Soc. 1994,
116, 10015.
(21) (a) Shapiro, P. J . Coord. Chem. Rev. 1999, 189, 1. (b) Fisher, J .
D.; Shapiro, P. J .; Budzelaar, P. H. M.; Staples, R. J . Inorg. Chem.
1998, 37, 1295.

3366 Organometallics, Vol. 19, No. 17, 2000
Burns et al.
clear yellow solution with a suspended white solid. More
(Cp*)2AlCl precipitated out as a white solid upon cooling the
flask to -78 °C for 1 h. The white solid was isolated by cold
filtration and dried under vacuum (yield 1.24 g, 74%). 1H NMR
(300 MHz, 297 K, C6D6): 1.79 (s, 30H, C5(CH3)5). 13C{1H}
NMR: (118.8 (C5(CH3)5), 12.3 (C5(CH3)5). 27Al NMR (vs exter-
nal H3AlO3): -3.0. Anal. Calcd for C20H30AlCl: C, 72.16; H,
9.08. Found: C, 72.01; H, 9.02.
Cp *2AlMe, 2. Solid dichloromethylaluminum (0.9 g, 8
mmol) was added to a solution of Cp*Na (2.5 g, 16 mmol) in
50 mL of toluene at room temperature, and the reaction
mixture was stirred overnight. The NaCl was removed from
the reaction mixture by filtration and rinsed with toluene. The
volatiles were removed from the filtrate under reduced pres-
sure, leaving a yellow-tinged, white residue. The residue was
dissolved in 25 mL of peteroleum ether and cooled to -78 °C
to afford a white, microcrystalline solid, which was isolated
by cold filtration (yield 1.3 g, 50%). 1H NMR (300 MHz, C6D6):
δ 1.92 (s, 30H, C5(CH3)5, -1.58 (s, 3H, AlCH3). 13C{1H} NMR:
118.7 (C5(CH3)5, 11.3 (C5(CH3)5. 27Al NMR (vs external H3-
AlO3): δ 70. Anal. Calcd for C21H33Al: C, 80.72; H, 10.64.
Found: C, 80.67; H, 10.75.
reacts with AlX3 (X ) Cl, Br) to form the permethyl-
aluminocenium salts [Cp*2Al]+[AlX4]- exclusively.22
Also noteworthy is the isolation of the side product [Al-
(η5-C5Me5)2][Li(η5-C5Bz5)2] by Dohmeier et al. in the lig-
and exchange reaction between Li(C5Bz5) and Al(C5-
Me5).23
Failure of the aluminum to undergo chloride for
cyclopentadienide metathesis has been encountered by
us in another reaction in which Al(OiPr)Cl2 undergoes
only a single metathesis with (Me4CH)2MgCl to form
[Cl(C5Me4H)Al]2(µ-OiPr)2 and (C5Me4H)MgCl, even
though a similar reaction with Cp2Mg affords the double
metathesis product, [Cp2Al]2(µ-OiPr)2.24 In this case,
steric blocking of the Al-Cl sites in the dimer by the
tetramethylcyclopentadienyl ligands is the likely deter-
rent to further metathesis.
The contrasting behavior of gallium and aluminum
with respect to gallium’s ability to form Cp*3Ga and
aluminum’s inability to form Cp*3Al is noteworthy and
reflects the greater tendency of Al(III) to form π-interac-
tions with its cyclopentadienyl rings as compared with
Ga(III), which exhibits a preference for localized σ-bond-
ing.25 This difference in bonding is undoubtedly associ-
ated with the greater electronegativity of Ga(III) relative
to Al(III).26
The effect of steric factors on increasing the stability
of [Cp*2Al]+ relative to [Cp2Al]+ and decreasing its
effectiveness at activating cationic olefin polymerization
is not surprising. It highlights the possibility of varying
the ring substituents of the aluminocenium cation to
fine-tune its activity in initiating cationic olefin polym-
erization and in other Lewis-acid-catalyzed reactions.
[Cp *2Al][MeB(C6F 5)3], 3. At room temperature, 2 (0.359
g, 1.15 mmol) was added via a sidearm addition tube to a
solution of B(C6F5)3 (0.589 g, 1.15 mmol) in 50 mL of methylene
chloride, and the reaction mixture was stirred for 30 min. The
methylene chloride was removed under vacuum, leaving a tan
solid. Addition of petroleum ether to the solid afforded a white
precipitate and a gold-colored mother liquor. The white solid,
3, was isolated by filtration and washed twice with petroleum
ether (yield 0.792 g, 83%). 1H NMR (200 MHz, CDCl3): δ 2.13,
2.10 (2s, 30H, C5(CH3)5, 0.48 (s, 3H, BCH3). 13C{1H} NMR: 118
(C5(CH3)5, 9.0 (C5(CH3)5. 27Al NMR (vs external H3AlO3): δ
-114. 11B NMR (C6D6, vs external H3BO3): δ -34. 19F NMR
(470.6 MHz, CDCl3, vs external CFCl3): δ -133.1 (d, o-F, 3J FF
3
) 19.8 Hz), -165.3 (t, p-F, J FF ) 20.7 Hz), -167.8 (m, m-F).
Anal. Calcd for C39H33AlBF15: C, 56.82; H, 4.03. Found: C,
56.42; H, 3.89.
X-r a y Cr ysta l Str u ctu r e Deter m in a tion s. Crystals of 2
were grown from a saturated hexamethyldisiloxane solution,
and a suitable crystal was mounted in a glass capillary under
nitrogen.
Exp er im en ta l Section
Gen er a l Con sid er a tion s. All manipulations were per-
formed using a combination of glovebox, high-vacuum, and
Schlenk techniques as described elsewhere.2 Aluminum trichlo-
ride (Aldrich) was sublimed prior to use. Dichloromethylalu-
minum was handled as a solid, which was obtained by
removing the solvent from a 1.0 M hexane solution (Aldrich).
Pentamethylcyclopentadiene was prepared from 2,3,4,5-tet-
ramethyl-2-cyclopentenone (Aldrich) as described by Marks
and co-workers. Tris(pentafluorophenyl)borane was both pre-
pared and purchased (Boulder Scientific). Cp*K was prepared
by deprotonating Cp*H with potassium hexamethyldisilazide.
Cp *2AlCl, 1. Aluminum trichloride (0.67 g, 5.0 mmol) was
added to a solution of Cp*Na (1.58 g, 10 mmol) in 50 mL of
toluene at -78 °C, and the reaction mixture was slowly
warmed to room temperature and stirred overnight. The clear
yellow solution was filtered to remove NaCl. The NaCl was
rinsed repeatedly with toluene, and the toluene was removed
under reduced pressure, leaving behind a white solid with
traces of a yellow oil. Petroleum ether (25 mL) was vacuum
transferred into the reaction flask. The yellow oil dissolved
upon warming the flask to room temperature, resulting in a
Data were collected using a Siemens (Bruker) SMART CCD
(charge coupled device) based diffractometer equipped with an
LT-2 low-temperature apparatus operating around -54 °C. A
total of 1271 frames of data were collected using ω scans with
a scan width of 0.3° per frame for 45 s. Additional parameters
are available in the cif file. The first 50 frames were recollected
at the end of data collection to monitor for decay. No decom-
position of the crystals during data collection was indicated.
Cell parameters were retrieved using SMART software (V.
4.050, Bruker Analytical X-ray Systems, Madison, WI, 1995)
and refined using SAINT (V. 4.050, Bruker Analytical X-ray
Systems, Madison, WI, 1995) on all observed reflections. Data
reduction was performed using the SAINT software, which
corrects for Lp and decay. Absorption corrections were applied
using SADABS (Siemens Area Detector Absortion Correction
Program). The structure was solved in the space group P421m
by direct methods using the SHELXS-97 program (Sheldrick,
G. M., University of Go¨ttingen, Germany, 1997). In this space
group, the Al and C(1) methyl groups sit on the intersection
of two mutually perpendicular mirror planes, with the cyclo-
pentadienyl ring sitting athwart one of the mirror planes. With
an ordered model, there were two factors that pointed toward
the presence of disorder: (1) very large atomic displacement
parameters on the cyclopentadienyl methyl groups indicative
of disorder, and (2) a very short ring C-C distance of 1.19 Å
between the two carbon atoms bonded to the aluminum.
Hence, a disorder of the cyclopentadienyl rings was introduced,
and refinement proceeded smoothly by the least-squares
method on F2 using SHELXL-97, which is incorporated in
(22) Holtmann, U.; J utzi, P.; Ku¨hler, T.; Neumann, B.; Stammler,
H.-G. Organometallics 1999, 18, 5531-5538.
(23) Dohmeier, C.; Baum, E.; Acker, A.; Ko¨ppe, R.; Schno¨ckel
Organometallics 1996, 15, 4702-4706.
(24) Fisher, J . D.; Golden, J . T.; Shapiro, P. J .; Yap, G. P. A.;
Rheingold, A. L. Main Group Met. Chem. 1996, 19, 521.
(25) (a) Beachley, O. T., J r.; Getman, T. D.; Kirss, R. U.; Hallock,
R. B.; Hunter, W. E.; Atwood, J . L. Organometallics 1985, 4, 751. (b)
Beachley, O. T.; Hallock, R. B.; Zhang, H. M.; Atwood, J . L. Organo-
metallics 1985, 4, 1675. (c) Cowley, A. H.; Mehrotra, S. K.; Atwood, J .
L.; Hunter, W. E. Organometallics 1985, 4, 1115.
(26) Huheey, J . E.; Keiter, E. A.; Keiter, R. L. Inorganic Chemistry,
Principle of Structure and Reactivity; 4th ed.; Harper Collins: New
York, 1993; p 188.

Bis(permethylcyclopentadienyl)aluminum Compounds
Organometallics, Vol. 19, No. 17, 2000 3367
SHELXTL-PC V 5.10 (PC/UNIX-Version, Bruker Analytical
X-ray Systems, Madison, WI, 1995). The disorder consisted of
a rotation by (16° about the normal to the cyclopentadienyl
ring. No disorder of the Al or C(1) position was introduced.
The cyclopentadienyl methyl groups were constrained to ideal
C3v symmetry.
Crystals of 3 were grown by slow diffusion of pentane into
a methylene chloride solution of the compound cooled at -78
°C. A suitable crystal was mounted on a pin with silicon grease,
and the pin was mounted on a goniometer head.
Data collection was similar to that for 2, except that the ω
scans were collected with a scan width of 0.3° per frame for
30 s. Data reduction and absorption corrections were per-
formed using the methods listed above. The structure was
solved by the direct method.
All non-hydrogen atoms were refined anisotropically. Hy-
drogen positions were calculated by geometrical methods and
refined as a riding model.
from the solvent. The polymer was washed twice with 10%
nitric acid, then with distilled water. It was then reprecipitated
as a white powder from a methylene chloride solution at -78
°C and dried in a 60 °C oven. Yield of polymer was 2.1 g.
Ack n ow led gm en t. Support of this work by the
University of Idaho Seed Grant program is gratefully
acknowledged. The establishment of a Single-Crystal
X-ray Diffraction Laboratory and the purchase of a 500
MHz NMR spectrometer were supported by the M. J .
Murdock Charitable Trust of Vancouver, WA, the
National Science Foundation, and the NSF-Idaho EP-
SCoR Program. We thank Mr. Zheng Hao for experi-
mental contributions and Dr. Emilio Bunel (DuPont) for
GPC analyses. We also thank Dr. Gary Knerr for his
assistance with the NMR experiments.
Isobu ten e P olym er iza tion Exp er im en ts. In a typical
experiment, 50-60 µmol of 3 was dissolved in 5 mL of
methylene chloride in a thick-walled, glass vessel. Approxi-
mately 10 mL of isobutene was condensed into the reaction
vessel at -78 °C. The reaction was allowed to warm to room
temperature and stirred overnight. Methanol (0.5-1 mL) was
added to quench the reaction. Additional methanol precipitated
the polymer as a white, gummy material which was separated
Su p p or tin g In for m a tion Ava ila ble: Tables of crystal-
lographic data, positional parameters, bond lengths and
angles, thermal parameters, and atomic parameters of hydro-
gen atoms for 2 and 3. This material is available free of charge
via the Internet at http://pubs.acs.org.
OM000173X