A new family of monocyclopentadienyl organoscandium bis-alkyls supported by a bulky trialkylphosphine oxide ancillary

Article abstract of DOI:10.1139/v03-157

Treatment of the oligomeric compound [Cp*ScCl₂] _n (Cp* = C₅Me₅) with tri-tert- butylphosphine oxide in THF leads to the monocyclopentadienyl scandium dichloride [Cp*Cf-Bu₃P=O)ScCl₂] as a monomeric, THF-free solid in 72% yield. This compound can be alkylated with MeLi to give the corresponding dimethyl derivative, [Cp*(t-Bu₃P=O)ScMe ₂], 1. The crystal structures of these two compounds have been done and reveal a three-legged piano stool geometry about the scandium center. The coordinated phosphine oxide ligand exchanges with free donor in both compounds via an associative mechanism, as indicated by the activation parameters (ΔH? = 6.4-8.4 kcal mol^-1-; ΔS? = -22.5 to -26.6 eu). Thus, the Cp*(t-Bu₃P=O) ligand set provides a stable platform for organoscandium chemistry. This was demonstrated by the activation of 1 by treatment with B(C₆F₅)₃ to form the fully characterized contact ion pair [Cp*(t-Bu₃P=O)ScMe] ⁺[MeB(C₆F₅)₃]^-, 2, which is an active ethylene polymerization catalyst.

Full text of DOI:10.1139/v03-157

162
A new family of monocyclopentadienyl
organoscandium bis-alkyls supported by a bulky
trialkylphosphine oxide ancillary1
Lee D. Henderson, Glen D. MacInnis, Warren E. Piers, and Masood Parvez
Abstract: Treatment of the oligomeric compound [Cp*ScCl2]n (Cp* = C5Me5) with tri-tert-butylphosphine oxide in
THF leads to the monocyclopentadienyl scandium dichloride [Cp*(t-Bu3P=O)ScCl2] as a monomeric, THF-free solid
in 72% yield. This compound can be alkylated with MeLi to give the corresponding dimethyl derivative, [Cp*(t-
Bu3P=O)ScMe2], 1. The crystal structures of these two compounds have been done and reveal a three-legged piano
stool geometry about the scandium center. The coordinated phosphine oxide ligand exchanges with free donor in both
compounds via an associative mechanism, as indicated by the activation parameters (∆H‡ = 6.4–8.4 kcal mol–1-; ∆S‡ =
–22.5 to –26.6 eu). Thus, the Cp*(t-Bu3P=O) ligand set provides a stable platform for organoscandium chemistry. This
was demonstrated by the activation of 1 by treatment with B(C6F5)3 to form the fully characterized contact ion pair
[Cp*(t-Bu3P=O)ScMe]+[MeB(C6F5)3]–, 2, which is an active ethylene polymerization catalyst.
Key words: organoscandium compounds, olefin polymerization, cationic complexes.
Résumé : Le traitement du produit oligomère [Cp*ScCl2]n (Cp* = C5Me5) avec de l’oxyde de tri-tert-butylphosphine,
dans le THF, conduit à la formation du dichlorure de monocyclopentadiénylscandium, [Cp*(t-Bu3P=O)ScCl2], isolé
avec un rendement de 72 %, sous la forme monomère et exempt de THF. Ce composé peut être alkylé avec du MeLi
pour donner le dérivé diméthylé correspondant, [Cp*(t-Bu3P=O)ScMe2], 1. On a déterminé les structures cristallines de
ces deux composés qui comportent une géométrie en banc de piano à trois pattes autour du centre scandium. Le ligand
oxyde de phosphine qui est coordiné s’échange avec un donneur libre dans les deux composés, par le biais d’un méca-
nisme associatif suggéré par les paramètres d’activation (∆H‡ = 6,4 à 8,4 kcal mol–1; ∆S‡ = –22,5 à –26,6 ue). Ainsi,
l’ensemble de ligand Cp*t-Bu3P=O fournit une plate-forme stable pour la chimie de composés organiques du scandium.
Ceci a été démontré par l’activation de 1 par traitement avec du B(C6F5)3 qui conduit à la formation de la paire d’ions
de contact complètement caractérisé, [Cp*(t-Bu3P=O)ScMe]+[MeB(C6F5)3]– (2), qui est un catalyseur actif pour la poly-
mérisation de l’éthylène.
Mots clés : composés organiques du scandium, polymérisation des oléfines, complexes cationiques.
[Traduit par la Rédaction] Henderson et al. 165
Introduction
ligand framework (A) and related structures (2), which sup-
ports the chemistry of base-free scandium dialkyls (3) and
cationic scandium alkyls (4) but cannot eliminate the need
for THF ligation in analogous organoyttrium chemistry (5).
Despite the success of this ligand ancillary and its deriva-
tives (6), new environments are of interest in the continued
expansion of this area of early transition-metal organo-
metallic chemistry.
Suitable ligand environments for group 3 metal bis-alkyl
functions free from occluded salts or coordinating solvent
molecules are rare; this lack of abundance has prevented de-
velopment of the chemistry of organo group 3 dialkyls to the
same degree as the extensively explored group 4 dialkyls (1).
A notable exception is the bulky β-diketiminato (nacnac)
Received 31 July 2003. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 12 January 2004.
Dedicated to Professor Edward Piers on the occasion of his
65th birthday and for his many contributions to chemistry in
Canada.
L.D. Henderson, G.D. MacInnis, W.E. Piers,2,3 and
M. Parvez. Department of Chemistry, University of Calgary,
2500 University Drive NW, Calgary, AB T2N 1N4, Canada.
1This article is part of a Special Issue dedicated to Professor
Ed Piers.
2S. Robert Blair Professor of Chemistry (2000–2005).
3Corresponding author (e-mail: wpiers@ucalgary.ca).
Can. J. Chem. 82: 162–165 (2004)
doi: 10.1139/V03-157
© 2004 NRC Canada

Henderson et al.
163
Fig. 1. Molecular structure of [Cp*(t-Bu3P=O)ScMe2], 1, with
Scheme 1.
hydrogen atoms omitted for clarity. Selected metrical parameters
are given below; numbers in parentheses are analogous distances
and angles for [Cp*(t-Bu3P=O)ScCl2]. Distances (Å): Sc(1)—
Cp*centroid, 2.21(1) (2.159(4)); Sc(1)—O(1), 2.072(1) (2.0373(1));
Sc(1)—C(11), 2.251(2) (Sc(1)—Cl(1), 2.3906(6)); Sc(1)—C(12),
2.252(2) (Sc(1)—Cl(2), 2.3995(5)); O(1)—P(1), 1.5225(13)
(1.5336(10)). Angles (°): Sc(1)-O(1)-P(1), 162.09(8) (161.12(6));
O(1)-Sc(1)-C(11), 103.22(7) (101.01(3)); O(1)-Sc(1)-C(12),
100.16(7) (100.02(3)); C(11)-Sc(1)-C(12), 104.48(8) (105.54(2)).
Inspiration for possible organo group 3 ligand environ-
ments can be found in successful areas of group 4 metal
chemistry. Stephan and co-workers (7) and others (8) have
demonstrated the efficacy of the bulky phosphinimide ligand
–
t-Bu3P=N– as a steric replacement for the Cp* ligand (C5Me5 ),
following a strategy originally tendered by Wolczanski and
co-workers (9) and embodied in the tri-tert-butylmethoxide
“tritox” ligand system. Thus, the organotitanium and zirco-
nium compounds B are sterically analogous to the more
well-known permethylmetallocenes but with a somewhat
more open environment in the immediate vicinity of the
metal center, which confers high reactivity. Since dialkyl
group 3 chemistry requires a monoanionic ancillary ligand
set, either the Cp* or the phosphinimide ligand must be neu-
tralized to transfer this general ligand environment to scan-
dium or yttrium. Herein we describe the preparation and
some preliminary chemistry of [Cp*(t-Bu3P=O)ScMe2], 1, a
new THF-free organoscandium dimethyl complex incorpo-
rating a bulky phosphine oxide ligand as a Cp* equivalent.
The synthesis of 1 is straightforward and shown in
Scheme 1.4 Treatment of [Cp*ScCl2]n (10), which possesses
an ill-defined oligomeric structure, with t-Bu3P=O (11)
leads to deoligomerization and formation of [Cp*(t-
Bu3P=O)ScCl2] in good yield on a multigram scale.
Downfield shifts in the 1H and 31P NMR spectra (by
0.18 ppm and 24.5 ppm, respectively) are observed for the
coordinated t-Bu3P=O relative to the free ligand, mirroring
trends observed in other R3P=O coordination compounds of
scandium (12). Notably, the material is isolated free of THF
even though the reaction is performed in this normally per-
sistent solvent, attesting to the steric satisfaction provided by
the Cp* – phosphine oxide combination to the Lewis-acidic
scandium center. The dichloride is readily methylated when
treated with MeLi in toluene, giving 1 in 45% isolated yield
(1H NMR (Sc—Me) = –0.36 ppm; 31P NMR = 82.2 ppm).
Care must be taken not to expose this compound to excess
MeLi, to avoid the formation of a side product that we be-
lieve to be the scandate complex [Cp*ScMe3]–[Li]+.
Both 1 and its dichloride precursor are monomeric in the
solid state as determined by X-ray crystallography; the
structure of 1 is shown in Fig. 1, along with selected metri-
cal parameters for both compounds for comparative pur-
poses. The structures are grossly similar to that reported for
[Cp*(t-Bu3P=N)TiCl2] (7a), with the phosphine oxide ligand
occupying a roughly equivalent region of coordination space
to that of the Cp* ligand in a tetrahedral environment. The
4 Full experimental details on the synthesis and characterization of new compounds can be found in the Supporting information. Supplemen-
tary data may be purchased from the Depository of Unpublished Data, Document Delivery, CISTI, National Research Council Canada, Ot-
tawa, ON K1A 0S2, Canada (http://www.nrc.ca/cisti/irm/unpub_e.shtml for information on ordering electronically). CCDC 215982–215984
contain the supplementary data for this paper. These data can be obtained, free of charge, via www.ccdc.cam.ac.uk/conts/retrieving.html (or
from the Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge, U.K.; fax +44 1223 336033; or deposit@ccdc.cam.ac.uk).
© 2004 NRC Canada

164
Can. J. Chem. Vol. 82, 2004
Fig. 2. Molecular structure of [Cp*(t-Bu3P=O)ScMe]+[MeB(C6F5)3]–, 2, with hydrogen atoms omitted for clarity. Selected bond dis-
tances (Å): Sc(1)—Cp*centroid, 2.155(1); Sc(1)—O(1), 2.0370(13); Sc(1)—C(11), 2.201(2); Sc(1)—C(12), 2.529(11); O(1)—P(1),
1.536(13); B(1)—C(12), 1.658(7). Selected bond angles (°): Sc(1)-O(1)-P(1), 167.49(8); Sc(1)-C(12)-B(1), 170.00(11); O(1)-Sc(1)-
C(11), 102.10(7); O(1)-Sc(1)-C(12), 95.06(5).
(15) using variable temperature 1H NMR spectra, was
Scheme 2.
11.9(5) kcal mol–1 (253 K) for [Cp*(t-Bu3P=O)ScCl2] and
17.7(5) kcal mol–1 (363 K) for the less Lewis-acidic
dimethyl derivative 1. The activation parameters were deter-
mined by analysis of the rate constants at various tempera-
tures using EXSY methods (16) and the ∆S‡ values are
–22.4(8) and –26.6(6) eu for the dichloride and 1, respec-
tively. These values are consistent with an associative mech-
anism for ligand exchange and indicate that the compounds
are stable towards dissociation of the ligand under ambient
conditions. It should be noted, however, that compound 1
decomposes to unknown products when heated to tempera-
Sc(1)-O(1)-P(1) angles are about the same for both com-
pounds and are bent slightly away from the Cp* ligand. The
Sc(1)—O(1) distances are comparable with those of other
tures above 80 °C in toluene, a process that is accelerated in
phosphine oxide complexes of scandium (12) and intermedi-
the donor solvent THF.
ate between the typical distances for scandium alkoxides
(≈1.90 Å (13)) and scandium ethers (≈2.20 Å (14)). In 1, the
Sc(1)—O(1) distance is elongated relative to the dichloride,
reflecting the lower Lewis acidity of the Cp*ScMe2 frag-
ment.
Despite the accessibility of these new organoscandium
compounds, the dative nature of the phosphine oxide – scan-
dium linkage raises concerns about the kinetic stability of
this ligand and its suitability as an ancillary. Indeed, for both
the dichloride and 1, the coordinated t-Bu3P=O ligand ex-
changes with free ligand on the NMR timescale. The barri-
ers to exchange, as measured by the coalescence method
Treatment of 1 with 1 equiv of B(C6F5)3 in toluene leads
to a soluble contact ion pair, [Cp*(t-Bu3P=O)ScMe]+[Me-
B(C6F5)3]–, 2 (Scheme 2), demonstrating that the Cp*–t-
Bu3P=O ligand set can deliver a stable platform for organo-
metallic chemistry. Although phosphine oxide adducts of
B(C6F5)3 are known (17), the 11B NMR chemical shift for 2
(–15.2 ppm) and the small difference between the meta and
para fluorine signals in the 19F NMR spectrum (∆m,p
=
4.1 ppm (18)) clearly establishes 2 as a contact ion pair deriv-
ing from methide abstraction (19). Separate, nonexchanging
signals for Sc-Me (–0.07 ppm) and B-Me (1.09 ppm, br) in-
© 2004 NRC Canada

Henderson et al.
165
(2002); (c) P.G. Hayes, W.E. Piers, and M. Parvez. J. Am.
dicate that this abstraction is irreversible, at least on the
NMR timescale; a similar lack of Sc-Me–B-Me exchange is
observed in the organoscandium methyl cation of complex A
(4b). The structure was confirmed by X-ray analysis of
colourless crystals of 2 grown from toluene; an ORTEP dia-
gram is shown in Fig. 2, along with selected distances and
angles. The carbon-boron distance in the anion is 1.658(7) Å,
comparable with that observed in the contact ion pair
formed from the β-diketiminato complex A (4b). In contrast
to the nacnac system, however, there are no close Sc···F con-
tacts, and the Sc(1)···C(12) contact of 2.529(11) Å is much
closer than the 2.703(6) Å observed in the ion pair formed
from A. In addition, the B(1)-C(12)-Sc(1) angle of
170.00(11)° is more characteristic of a metallocenium ion
pair (19). The decreased steric congestion about scandium
relative to 1 allows for a slightly more linearly coordinated
phosphine oxide ligand (Sc(1)-O(1)-P(1) = 167.49(8)°).
Compound 2 is relatively thermally stable in solution, and
preliminary experiments indicate that 2 is a highly active
ethylene polymerization catalyst at room temperature and
1 atm (1 atm = 101.325 kPa) of ethylene (1.3 × 105 g mol–1 h–1
of polyethylene). However, polymerizations carried out at
higher temperatures indicated a decline in activity presum-
ably due to catalyst decomposition, either thermally or via
catalyst interaction with the solvent and with monomer feed
scrubbers present in the medium.
Chem. Soc. 125, 5622 (2003).
5. (a) G.C. Welch, P.G. Hayes, D.J.H. Emslie, C.L. Noack, W.E.
Piers, and M. Parvez. Organometallics, 22, 1577 (2003).
6. (a) A.M. Neculai, H.W. Roesky, D. Neculai, and J. Magull.
Organometallics, 20, 5501 (2001); (b) A.M. Neculai, D.
Neculai, H.W. Roesky, J. Magull, M. Baldus, O. Andreonesi,
and M. Jansen. Organometallics, 21, 2590 (2002).
7. (a) D.W. Stephan, J.C. Stewart, F. Guerin, S. Courtenay, J.
Kickham, E. Hollink, C. Beddie, A. Hoskin, T. Graham, P.
Wei, R.E.v. H. Spence, W. Xu, L. Koch, X. Gao, and D.G.
Harrison. Organometallics, 22, 1937 (2003); (b) N. Yue, E.
Hollink, F. Guerin, and D.W. Stephan. Organometallics, 20,
4424 (2001); (c) F. Guerin, J.C. Stewart, C. Beddie, and D.W.
Stephan. Organometallics, 19, 2994 (2000); (d) D.W. Stephan,
J.C. Stewart, F. Guerin, R.E.v.H. Spence, W. Xu, and D.G.
Harrison. Organometallics, 18, 1116 (1999); (e) D.W. Stephan,
F. Guerin, R.E.v. H. Spence, L. Koch, X. Gao, S.J. Brown,
J.W. Swabey, Q. Wang, W. Xu, P. Zoricak, and D.G. Harrison.
Organometallics, 18, 2046 (1999).
8. (a) K. Dehnicke, S. Anfang, K. Harms, F. Weller, O.
Borgmeier, H. Lueken, and H. Schilder. Z. Anorg. Allg. Chem.
624, 159 (1998); (b) M.J. Sarsfield, M. Said, M. Thornton-
Pett, L.A. Gerrad, and M. Bochmann. J. Chem. Soc. Dalton
Trans. 822 (2001); (c) K. Dehnicke, M. Krieger, and W.
Massa. Coord. Chem. Rev. 182, 19 (1999).
9. T.V. Lubben, P.T. Wolczanski, and G.D. Van Duyne. Organo-
metallics, 3, 977 (1984).
10. W.E. Piers, E.E. Bunel, and J.E. Bercaw. J. Organomet. Chem.
407, 51 (1991).
11. H. Schmidbaur and G. Blaschke. Z. Naturforsch. 33B, 1556
(1978).
12. (a) L. Deakin, W. Levanson, M.C. Popham, G. Reid, and M.
Webster. J. Chem. Soc. Dalton Trans. 2439 (2000); (b) J.
Fawcett, A.W.G. Platt, and D.R. Russell. Polyhedron, 21, 287
(2002); (c) N.J. Hill, W. Levason, M.C. Popham, G. Reid, and
M. Webster. Polyhedron, 21, 1579 (2002).
13. (a) P.B. Hitchcock, M.F. Lappert, and A. Singh. J. Chem. Soc.
Chem. Commun. 1499 (1983); (b) G.B. Deacon, P.E. Fanwick,
A. Gitlits, I.P. Rothwell, B.W. Skelton, and A.H. White. Eur. J.
Inorg. Chem. 1505 (2001).
In summary, we have prepared a new dimethyl scandium
derivative with a ligand environment that mimics the
metallocene set (20) but provides two reactive hydrocarbyl
ligands with which to develop new organoscandium chemis-
try. The facile preparation of these novel compounds offers
an opportunity to open new areas of organoscandium chem-
istry.
Acknowledgments
Funding for this work came from NOVA Chemicals
(Calgary, Alberta) and from the Natural Sciences and Engi-
neering Research Council of Canada (NSERC) in the form
of a CRD grant. LDH also thanks NSERC for scholarship
support (PGSA and PGSB).
14. (a) J.L. Atwood and K.D. Smith. J. Chem. Soc. Dalton Trans.
921 (1974); (b) B. Campion and T.D. Tilley. Organometallics,
12, 2584 (1993).
15. J. Sandström. Dynamic NMR spectroscopy. Academic Press,
References
Toronto. 1982. pp. 93–97.
16. C.L. Perrin and T.J. Dwyer. Chem. Rev. 90, 935 (1990).
17. (a) M.A. Beckett, D.S. Brassington, S.J. Coles, and M.B.
Hursthouse. Inorg. Chem. Commun. 3, 530 (2000); (b) M.A.
Beckett, D.S. Brassington, M.E. Light, and M.B. Hursthouse.
J. Chem. Soc. Dalton Trans. 1768 (2001).
18. A.D. Horton and J. de With. Organometallics, 16, 5424 (1997).
19. E.Y.-X. Chen and T.J. Marks. Chem. Rev. 100, 1391 (2000).
20. M.E. Thompson, S.M. Baxter, A.R. Bulls, B.J. Burger, M.C.
Nolan, B.D. Santarsiero, W.P. Schaefer, and J.E. Bercaw. J.
Am. Chem. Soc. 109, 203 (1987).
1. D.J.H. Emslie and W.E. Piers. Coord. Chem. Rev. 233–234,
129 (2002).
2. L. Bourget-Merle, M.F. Lappert, and J.R. Severn. Chem. Rev.
102, 3031 (2002).
3. P.G. Hayes, L.W.M. Lee, L.K. Knight, W.E. Piers, M. Parvez,
M.R.J. Elsegood, W. Clegg, and R. McDonald. Organo-
metallics, 20, 2533 (2001).
4. (a) L.W.M. Lee, W.E. Piers, M.R.J. Elsegood, W. Clegg, and
M. Parvez. Organometallics, 18, 2947 (1999); (b) P.G. Hayes,
W.E. Piers, and R. McDonald. J. Am. Chem. Soc. 124, 2132
© 2004 NRC Canada