A new diamido-amine ligand based on three-carbon atom “arms”: synthesis, structures and polymerisation capability of zirconium derivatives of MeN(CH₂CH₂CH₂NSiMe₃)₂

Abstract

Zirconium compounds of the new diamido-amine ligand MeN(CH₂CH₂CH₂NSiMe₃)₂ feature significantly different molecular structures and considerably improved olefin polymerisation characteristics in comparison with analogous compounds based on the two-carbon arm homologues.

---

Early transition metal complexes of polydentate amide ligands^1a,b have been shown to be important in the polymerisation^1c and, more recently, hydroamination^1d of olefins. Most relevant to our present contribution are Group 4 (especially Zr) complexes of diamido-amine ligands (Chart 1).² Cloke^2a and Horton^2b first introduced the all-SiMe₃-substituted systems I (N₂N^C2,TMS) which were susceptible to intramolecular activation of the amide-bound SiMe₃ groups. Schrock has reported the living polymerisation of 1-hexene with the mesityl functionalised analogue II.^3a The SiMe₃-centred deactivation reactions of complexes of I and the success of II and related^3b non-SiMe₃ systems have led to the view^3c that SiMe₃ amide N-substituents are incompatible with cationic olefin polymerisation catalysts. The flexible ligand II can bind with fac or mer coordination. To force the apparently more favourable fac diamide-donor mode, a recent focus has been on the tripod-like ligand IV^4a which is a modification of Gade's original SiMe₃-functionalised ligand III.^4b Despite the attractiveness of these latter ligands, their multistep syntheses involve organic azide intermediates and a high temperature and pressure autoclave first step.^4b Here we report a new 3-carbon “arm” diamido-amine ligand that is available in multigram quantities from commercially available starting materials, and which exclusively affords fac coordination and exhibits promising olefin polymerisation behaviour, even with amide-SiMe₃ substituents.

Chart 1

Reaction of the commercially available MeN(CH₂CH₂CH₂NH₂)₂ with ClSiMe₃ (2 equiv.) in the presence of NEt₃ affords the protio ligand MeN(CH₂CH₂CH₂N(H)SiMe₃)₂ (H₂N₂N^C3, 1-H₂) as a colourless oil in 84% yield (> 15 g isolated quantities) after a simple pentane extraction.† Reaction with BuLi (2 equiv.) gives the structurally characterised⁵ dimer [MeN(CH₂CH₂CH₂N(Li)SiMe₃)₂]₂ (1-Li₂) in 91% yield. For the purposes of making a strict comparison of the 3-carbon arm N₂N^C3 ligand with 2-carbon arm analogues under otherwise identical conditions we selected the ligand N₂N^C2,MeV (Chart 1) reported by Bertrand in the context of main group chemistry.⁶

The synthesis and structures of zirconium complexes of N₂N^C3 and N₂N^C2,Me are summarised in Scheme 1. The reactions all proceed in good yields. Alternatively, the dichloride 2 can be prepared from 1-Li₂ and [ZrCl₄(THF)₂], and the dibenzyl 3 can be made from 2 and PhCH₂MgBr (2 equiv.). The analogous reaction of 2 with MeMgBr (2 equiv.) yields [Zr(N₂N^C3)Me₂] 6. The structures for 2–5 were confirmed by X-ray crystallography⁵ and that of 3 is shown in Fig. 1 by way of example.†‡ Unlike certain Group 4 complexes of the two-carbon atom ligand I,^2a the 300 MHz ¹H NMR spectra of 6 (toluene-d₈) show no evidence for dissociation of the amino NMe nitrogen on the NMR timescale up to 80 °C at which temperature thermal decomposition becomes significant.

Fig. 1 Molecular structure of [Zr(N₂N^C3)(CH₂Ph)₂] 3. Selected data: Zr(1)–N(1) 2.039(2), Zr(1)–N(2) 2.517(2), Zr(1)–N(3) 2.053(2), Zr(1)–C(14) 2.319(2), Zr(1)–C(21) 2.327(3) Å.

Scheme 1 (i) [ZrCl₂(NMe₂)₂(THF)₂], yields 52% (for 2) and 77% (for 4); (ii) [Zr(CH₂Ph)₄], yields 70% (for 3) and 45% (for 5).

Scheme 1 clearly shows how the 3-carbon arm N₂N^C3 ligand favours fac-coordination. It also illustrates that, with this ligand, dichloride 2 remains monomeric whereas the otherwise identical 2-carbon arm ligand N₂N^C2,Me leads to binuclear 4. The binuclear structure of 4 parallels that formed with the all-SiMe₃ ligand I;² the mer-coordination found for N₂N^C2,Me in 5 is paralleled by the coordination of the N-mesityl ligand II in its dialkyl zirconium derivatives.^3a Points of note in the structure of 3 include the well-defined trigonal bipyramidal geometry at Zr with the longer ligand arms allowing the metal to be more fully embraced. The conformations of the two C₃ arms are different such that one SiMe₃ group (Si(1)) orientates towards the equatorial benzyl ligand and the other one is orientated away. The Zr–N distances for 3 are comparable to those reported for the compounds of the diamido-pyridine ligands III and IV.

With regard to olefin polymerisation catalysis, evidence of well-defined alkyl cations is essential. Dimethyl [Zr(N₂N^C3)Me₂] 6 reacts cleanly with [CPh₃][B(C₆F₅)₄] (TB, 1 ∶ 1 ratio) in C₆D₅Br to form [Zr(N₂N^C3)Me]⁺7⁺. There is no evidence for SiMe₃ activation at room temperature, and all data point to 7⁺ being fully solvent-separated from the anion. We note that the analogous reaction with ligand IV gives a self-trapped binuclear μ-methyl cation with a {Zr₂Me₂(μ-Me)} unit,^4a highlighting again how the new ligand N₂N^C3 helps enforce the formation of mononuclear species.

All three compounds [Zr(N₂N^C3)X₂] (X = Cl 2, CH₂Ph 3, Me 6) are active for the polymerisation of ethylene (Table 1) with very favourable polydispersities as indicated by the gel permeation chromatography (GPC) data (Fig. 2). Selected data for [Zr(N₂N^C2,Me)X₂]_n (X = Cl, n = 2 4; X = CH₂Ph, n = 1 5) and the previously reported² [Zr(N₂N^C2,TMS)Cl₂]₂8 under the same conditions are presented for comparison.

Fig. 2 GPC traces for polyethylene produced by [Zr(N₂N^C3)X₂] (X = Cl 2, CH₂Ph 3, Me 6), [Zr(N₂N^C2,Me)Cl₂]₂4 and [Zr(N₂N^C2,TMS)Cl₂]₂8.

Table 1 Polymerisation activities for [Zr(N₂N^C3)X₂] (X = Cl 2, CH₂Ph 3, Me 6), [Zr(N₂N^C2,Me)X₂]_n (X = Cl 4, CH₂Ph 5) and [Zr(N₂N^C2,TMS)Cl₂]₂8^a

Dichloride pre-catalyst^c	Activity^b (avg. Mw)	Dialkyl pre-catalyst^d	Activity^b (avg. Mw)
a Conditions: 10 or 20 µmol precatalyst, 250 cm^3 toluene; 5 bar ethylene; run time 60 min; To 22 ± 3 °C. b In kg(PE)/mol(Zr)/h/bar. c 1500 equiv. MAO. d 1 equiv. TB and 250 equiv. Al^iBu3.
2	110 (1.81 × 10^5)	3	164 (2.48 × 10^5)
4	1.3 (7.52 × 10^5)	5	0.8 (not measured)
8	4.0 (1.46 × 10^6)	6	47 (1.29 × 10^4)

---

On MAO activation (Al ∶ Zr ratio = 1500 ∶ 1), the dichloride 2 (3-carbon arm) has an activity that is about two orders of magnitude higher than that for the 2-carbon arm analogue 4. A very similar position emerges for the previously reported 8. With TB activation, the dibenzyl 3 is again considerably more active than the 2-carbon analogue 5 under identical conditions. The dimethyl compound 6 has a lower activity than that of the dibenzyl 3 (but similar polymer is formed, Fig. 2). This perhaps points to an activator effect^4a where, in the poorly polar toluene medium used for the polymerisation, there may in fact be some self-trapping of the 7⁺ prior to enchainment of monomer. Alternatively, 7⁺ could interact more strongly with the AlⁱBu₃ scavenger present than the cation derived from 3 does. Such factors will be the subject of future studies.

The ¹H and ¹³C NMR data (100 °C, C₆D₄Cl₂) for the soluble portion of the solid, free-flowing polymers formed by the new catalysts suggest they are non-branched and without detectable vinyl end-groups. No significant levels of hydrocarbon or other impurity were detected. The GPC data (Fig. 2) show that all three catalyst systems 2/MAO, 3/TB and 6/TB produce rather similar polymers containing low and high molecular weight fractions. Under identical conditions the polymers formed for the 2-carbon chain analogues are very broad and multimodal. Although the polymers formed by 2/MAO and 3/TB each have a high molecular weight fraction, the prominent low molecular weight components have polydispersity indices (PDIs) between 1.5 (6/TB; M_w = 2.6 × 10³) and 1.8 (2/MAO; M_w = 890). We believe that the larger PDIs for the more active 2/MAO and 3/TB systems can be attributed to the non-isothermal experimental conditions (ΔT_max between 20 and 50 °C were recorded). Indeed, preliminary results show that diluting the catalyst solution (and reducing the exotherm output) for 2/MAO (at constant Zr ∶ Al ratio) retains the lower molecular weight component and significantly reduces the higher molecular weight material. Further work on these aspects are in progress as well as extending the studies to α-olefins.

Although the data in Table 1 and Fig. 2 show that the silylated system N₂N^C3 has considerable merit and promise for future development (with the added benefit of the inexpensive and facile introduction of different SiR₃ amide N-substituents), it is clear from the literature that one should also have access to N-arylated homologues. Therefore we also report here our preliminary results that arylation of MeN(CH₂CH₂CH₂NH₂)₂ with mesityl bromide using standard procedures⁷ affords MeN(CH₂CH₂CH₂N(mesityl)H)₂ in ca. 50% isolated yield.† Complexation reactions of this ligand are underway, together with polymerisation studies of the compounds so formed. We will report on this work in due course.

In conclusion, we have introduced a new, simple and readily-available diamido-donor ligand, established its capability in areas of current interest and have demonstrated how its chemistry and applications may be developed in the future.

We thank the EPSRC and Leverhulme Trust for support and Albermarle for samples of MAO.

Notes and references

1. (a) L. H. Gade, Chem. Commun., 2000, 173; (b) R. Kempe, Angew. Chem., Int. Ed., 2000, 39, 468; (c) V. C. Gibson and S. K. Spitzmesser, Chem. Rev., 2003, 103, 283; (d) K. C. Hultzsch, F. Hampel and T. Wagner, Organometallics, 2004, 23, 2601.
2. (a) F. G. N. Cloke, P. B. Hitchcock and J. B. Love, J. Chem. Soc., Dalton Trans., 1995, 25; (b) A. D. Horton, J. de With, A. J. van der Linden and H. van de Weg, Organometallics, 1996, 15, 2672.
3. (a) L.-C. Liang, R. R. Schrock, W. M. Davis and D. H. McConville, J. Am. Chem. Soc., 1999, 121, 5797; (b) R. Baumann, W. M. Davis and R. R. Schrock, J. Am. Chem. Soc., 1997, 119, 3830; (c) R. R. Schrock, L.-C. Liang, R. Baumann and W. M. Davis, J. Organomet. Chem., 1999, 591, 163.
4. (a) P. Mehrkhodavandi, R. R. Schrock and L. L. Pryor, Organometallics, 2003, 22, 4569; (b) S. Friedrich, M. Schubart, L. H. Gade, I. J. Scowen, A. J. Edwards and M. McPartlin, Chem. Ber.-Recueil, 1997, 130, 1751.
5. R. A. Whannel, B. D. Ward, T. A. Lowes, A. R. Cowley, S. R. Dubberley and P. Mountford, unpublished results.
6. N. Emig, H. Nguyen, H. Krautscheid, R. Réau, J.-B. Cazaux and G. Bertrand, Organometallics, 1998, 17, 3599.
7. J. P. Wolfe, S. Wagaw and S. L. Buchwald, J. Am. Chem. Soc., 1996, 118, 7215.
8. P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout and D. J. Watkin, J. Appl. Crystallogr., 2003, 36, 1487.

---

Footnotes

† Electronic supplementary information (ESI) available: characterising data for the new compounds. See http://www.rsc.org/suppdata/cc/b4/b412381f/

‡ Crystal data for [Zr{MeN(CH₂CH₂CH₂NSiMe₃)₂}(CH₂Ph)₂] (3): C₂₇H₄₇N₃Si₂Zr₁, M_w = 561.09, orthorhombic, Pna 21, a = 20.9927(4), b = 12.0115(2), c = 12.1022(2) Å, α = 90.00, β = 90.00, γ = 90.00°, U = 3051.6(1) ų, Z = 4, F(000) = 1180.68, T = 150 K, Nonius Kappa CCD, Mo-Kα radiation, 2.91 ≤ 2θ ≤ 27.48°, 6677 independent reflections, 5525 reflections I > 3σ(I), R = 0.0273, R_w = 0.0262. The structure was solved using the CRYSTALS software suite.⁸ Notes on refinement: the refined Flack parameter of 0.50(3) is indicative of an intimately twinned structure since Friedel pairs were collected but not merged. CCDC 247514. See http://www.rsc.org/suppdata/cc/b4/b412381f/ for crystallographic data in .cif or other electronic format.

---