Influence of ligand backbone flexibility in group 4 metal complexes of tetradentate mixed tertiary amine/alkoxide ligands

Abstract

Simple epoxide ring opening chemistry using the cyclic secondary amine 1,4-diazacycloheptane or the related linear species N,N′-dimethylethylenediamine, and racemic (±)−3,3-dimethyl-1,2-epoxybutane gives access to the pendant alcohol functionalised ditertiary amine pro-ligands [HOCH(^tBu)CH₂N(R)CH₂]₂ (H₂L¹: R₂ = CH₂CH₂CH₂; H₂L²: R₂ = Me₂). The contrasting reactions of H₂L¹ and H₂L² towards homoleptic group 4 alkoxides highlight the crucial role of ligand backbone flexibility in complex formation. Thus, the chemistry of the more conformationally rigid system (L¹)²⁻ appears to be constrained by the cyclic ligand core, such that it adopts a bridging (μ₂:η²,η²) mode of coordination towards Ti(IV), leading to dinuclear metal systems [e.g.L¹Ti₂(OⁱPr)₆]. By contrast, the more flexible linear system (L²)²⁻ binds to both Ti(IV) and Zr(IV) in a chelating fashion leading, for example, to the synthesis of the C₂ symmetric mononuclear complex rac-L²Ti(OⁱPr)₂. Thus, a simple synthesis of diastereomerically pure, C₂ symmetric, geometrically cis octahedral Ti(IV) complexes from racemic precursors is presented.

---

Introduction

The design and synthesis of ancillary ligand frameworks offering an alternative to cyclopentadienyl-based systems in early transition metal olefin polymerisation catalysts remains an area of intense research activity.¹ Within this field, ligands featuring anionic oxygen or nitrogen donors (e.g. alkoxides/aryloxides, amides or imides) are particularly attractive targets given their π-donor capabilities and the strong electrostatic component to the metal–ligand interaction typically found with electropositive metals. For anionic oxygen ligands, RO⁻, the avoidance of secondary bridging interactions has led to the exploitation of a number of strategies in complex synthesis, e.g. the incorporation of steric bulk, chelating ligand frameworks and/or additional neutral donor atoms.² Hence, for example chelating, sterically encumbered or heteroatom-functionalized bis(aryloxide) complexes of titanium and zirconium have been shown to be highly active olefin polymerisation catalysts,^3–5 and alkoxide systems bearing ancillary N-donors have also been the subject of a number of recent studies.⁶

The use of metal complexes in asymmetric transformations, e.g. C₂ symmetric (or pseudo-C₂ symmetric) species in the isotactic polymerisation of propylene, has led to the development of a number of strategies for the synthesis of such complexes. Tetradentate bis(aryloxide) ligands featuring a linear array of donor atoms (Scheme 1) have featured prominently among these, with examples of both pre-constructed C₂ symmetric pro-ligands^4b,f,g,5m and coordination induced C₂ symmetry having been reported.^{4d,k,u,5p,6b,7} Within this area we have been interested in developing synthetic routes to sterically encumbered bis(tertiary amine) bis(alkoxide) ligand frameworks, through the ring opening of (±)-3,3-dimethyl-1,2-epoxybutane (Scheme 2). It was envisaged that the use of racemic precursors and the separation of diastereomers by simple crystallization at either the pro-ligand or complex stage would provide a convenient route to C₂ symmetric species.

Scheme 1 Target C₂ symmetric, geometrically cis, octahedral complexes featuring tetradentate bis(donor) bis(aryl/alkoxide) ligands.

Scheme 2 Syntheses of ligand precursors H₂L¹ and H₂L². Reagents and conditions: excess (±)-3,3-dimethyl-1,2-epoxybutane, acetonitrile, sealed tube, 100 °C, 72 h, yield ca. 40% for H₂L¹ [obtained as the rac (R,R/S,S) diastereomers], 59% for H₂L² [obtained as a mixture of the rac (R,R/S,S) and meso (R,S) diastereomers].

Experimental

(i) General considerations

Unless otherwise stated, all manipulations were carried out under a nitrogen or argon atmosphere using standard Schlenk line or dry-box techniques. Solvents were pre-dried over sodium wire (hexanes, toluene) or molecular sieves (acetonitrile) and purged with nitrogen prior to distillation from the appropriate drying agent (hexanes: potassium, toluene: sodium, acetonitrile: calcium hydride). Benzene-d₆ and chloroform-d (both Goss) were degassed and dried over potassium (benzene-d₆) or molecular sieves (chloroform-d) prior to use. Ligand precursors [1,4-diazacycloheptane (Lancaster), 1,4-diazacyclohexane (Avocado), N,N′-dimethylethylenediamine (Aldrich) and (±)-3,3-dimethyl-1,2-epoxybutane (Lancaster)] and metal reagents [Ti(OⁱPr)₄ (Lancaster), Ti(OEt)₄ and Zr(OⁿPr)₄ (all Aldrich)] were used as received. NMR spectra were measured on a Bruker AM-400 or JEOL 300 Eclipse Plus FT-NMR spectrometer. Residual signals of solvent were used as reference for ¹H and ¹³C NMR. Infrared spectra were measured for each compound pressed into a disk with an excess of dried KBr or as a solution in an appropriate solvent on a Nicolet 500 FT-IR spectrometer. Mass spectra were measured by the EPSRC National Mass Spectrometry Service Centre, University of Wales, Swansea. Perfluorotributylamine was used as a standard for high resolution EI mass spectra. Elemental microanalysis was performed by Warwick Analytical Services or by the departmental service. Abbreviations: br = broad, s = singlet, d = doublet, t = triplet, q = quartet, sept = septet, m = multiplet.

(ii) Ligand syntheses

N,N′-Bis(2-hydroxy-3,3-dimethylbutyl)-1,4-diazacycloheptane (H₂L¹). To a solution of 1,4-diazacycloheptane (0.90 g, 8.99 mmol) in acetonitrile (10 ml), contained in a pressure tube, was added excess (±)-3,3-dimethyl-1,2-epoxybutane (2.69 g, 26.86 mmol). The reaction mixture was stirred at 100 °C for 24 h, and the white crystals which formed on cooling of the reaction mixture were filtered off and washed with cold acetonitrile. Although single crystals suitable for X-ray diffraction could not be obtained, the ¹H and ¹³C NMR data for crystalline samples re-dissolved in chloroform-d imply that only one set of diastereomers is present; the crystal structures of nickel(II) and titanium(IV) complexes obtained by further reaction chemistry indicate that this is the rac (R,R/S,S) pair. Yield: 1.15 g, 43%. ¹H NMR (CDCl₃, 25 °C): δ 0.85 (s, 18 H, ^tBu), 1.85 (m, 2H, NCH₂CH₂CH₂N), 2.20 (m, 2H, C(H)(H)CHOH), 2.50 (m, 2H, C(H)(H)CHOH), 2.55 (m, 4H, NCH₂), 2.82 (m, 4H, NCH₂), 3.22 (m, 2H, CHOH), 3.82 (br s, 2H, OH). ¹³C NMR (CDCl_3, 25 °C): δ 25.7 (CH₃ of ^tBu), 28.2 (NCH₂CH₂CH₂N), 33.2 (^tBu quaternary), 54.4 (NCH₂), 56.2 (NCH₂), 59.5 (CH₂CHOH), 73.7 (CHOH). IR (KBr disk, cm⁻¹) 3417, 2955, 2855, 1653, 1467, 1410, 1362, 1244, 1154, 1088, 1013, 944, 863, 821. Mass spectrum (APCI): 301.6 (M + H)⁺. Elemental analysis. Calc. for C₁₇H₃₆N₂O₂: C, 67.95; H, 12.08; N, 9.32. Found: C, 68.31; H, 12.55; N, 8.99%. In an analogous manner, the corresponding reaction using 1,4-diazacyclohexane generates crystalline samples of N,N′-bis(2-hydroxy-3,3-dimethylbutyl)-1,4-diazacyclohexane which can be shown by a combination of NMR and X-ray crystallographic studies to contain solely the meso (R,S) isomer.⁸

N,N′-Bis(2-hydroxy-3,3-dimethylbutyl)dimethylethylenediamine (H₂L²). H₂L² was synthesized in an analogous manner to H₂L¹, using N,N′-dimethylethylenediamine (0.50 g, 5.67 mmol) and isolated after removal of volatiles from the reaction mixture in vacuo as a clear oil containing a mixture of rac (R,R/S,S) and meso (R,S) diastereoisomers at room temperature (waxy solid at −25 °C). Yield: 0.63 g, 59%. ¹H NMR (CDCl₃, 25 °C): δ 0.85 (s, 36 H, coincident ^tBu groups of both pairs of diastereomers), 2.23, 2.26 (s, each 6H, NCH₃), 2.31–2.68 (br overlapping m, 16H, overlapping NCH₂ groups of both diastereomers), 3.24, 3.27 (m, each 2H, CHOH), 4.31, 4.53 (br s, each 2H, OH). ¹³C NMR (CDCl₃, 25 °C): δ 25.87, 25.89 (CH₃ of ^tBu), 33.4 (coincident ^tBu quaternary carbons of both pairs of diastereomers), 42.5, 43.5 (NCH₃), 54.7, 55.8 (NCH₂), 58.6, 59.2 (CH₂CHOH), 74.6, 74.7 (CHOH). IR (KBr disk, cm⁻¹) 2957, 1463, 1363, 1217, 1090, 1017, 937. Mass spectrum (APCI): 289.1 (M + H)⁺. Exact mass. Calc. for C₁₆H₃₆N₂O₂: 289.2855. Found: 289.2859. Accurate elemental analysis proved impossible for this oil.

(iii) Complex syntheses

L¹Ti₂(OⁱPr)₆. Neat titanium tetrakis(isopropoxide) (0.50 ml, 1.76 mmol) was added to a H₂L¹ (0.50 g, 1.66 mmol) and the reaction mixture stirred until it became solid (ca. 4 h). Isopropanol formed during the reaction was then removed in vacuo and the resulting solid redissolved in dry hexanes. Colourless crystals suitable for X-ray diffraction were grown by cooling the hexanes solution to −30 °C. Variation in reaction conditions (stoichiometry of reaction, order of reagent addition, temperature) did not result in the isolation of alternative products containing a different Ti : L¹ ratio. The yield of L¹Ti₂(OⁱPr)₆ was optimised by consideration of the 2 : 1 Ti : L¹ ratio to 64% (ca. 0.75 g scale). ¹H NMR (C₆D₆, 25 °C): δ 1.00 (s, 18H, ^tBu), 1.37 (d, J = 6.0 Hz, 36H, CH₃ of ⁱPr), 1.85 (m, 2H, NCH₂CH₂CH₂N), 2.55 (m, 2H, CH(H)CHO), 2.84 (m, 2H, CH(H)CHO), 3.50 (br overlapping m, 8H, NCH₂ of diazacycloheptane), 4.08 (m, 2H, C(H)O), 4.92 (septet, J = 6.0 Hz, 6H, CH of ⁱPr). ¹³C NMR: δ 26.3 (CH₃ of ^tBu), 27.8 (CH₃ of ⁱPr), 31.1 (NCH₂CH₂CH₂N), 35.6 (^tBu quaternary), 58.1 (NCH₂CH₂CH₂N), 60.1 (NCH₂CH₂N), 62.3 (CH₂C(H)O), 77.4 (CH of ⁱPr), 83.1 (CHO). Mass spectrum (FAB): 747.3 (M − H)⁺. Elemental analysis. Calc. for C₃₅H₇₆N₂O₈Ti₂: C, 56.15; H, 10.23; N, 3.74. Found: C, 55.81; H, 10.01; N, 3.89%.

[L¹Ti(OEt)]₂(μ₂-O). [L¹Ti(OEt)]₂(μ₂-O) was prepared from titanium tetrakis(ethoxide) (0.40 ml, 1.75 mmol) and H₂L¹ (0.50 g, 1.66 mmol) using the method detailed above for L¹Ti₂(OⁱPr)₆. The product was isolated in very low yield (<5%) as colourless crystals obtained from hexanes solution at −30 °C. Complete characterization was frustrated in this case by the small amount of compound obtained. ¹H NMR (C₆D₆, 25 °C): δ 0.99 (s, 36H, ^tBu), 1.21 (t, J = 8.0 Hz, 6H, OCH₂CH₃), 1.88 (m, 4H, NCH₂CH₂CH₂N), 2.68 (br m, 4H, CH(H)CHO), 2.78 (m, 4H, CH(H)CHOH), 3.20–3.68 (br overlapping m, 16H, NCH₂ of diazacycloheptane), 3.91 (q, J = 8.0 Hz, 4H, OCH₂CH₃), 4.15 (m, 4H, CHO). Mass spectrum (FAB): 799.6 (M − H)⁺.

L²Ti(OⁱPr)₂. L²Ti(OⁱPr)₂ was prepared from titanium tetrakis(isopropoxide) (0.50 ml, 1.76 mmol) and H₂L² (0.50 g, 1.74 mmol) using the method detailed above for L¹Ti₂(OⁱPr)₆, and isolated as colourless crystals after two recrystallizations from concentrated hexanes solution (yield: 44%, ca. 1 g scale). Although the oily ligand precursor H₂L² was used as a mixture of rac (R,R/S,S) and meso (R,S) isomers, X-ray diffraction analysis of the crystalline titanium complex reveals it has approximate (non-crystallographically imposed) C₂ symmetry and contains the rac (R,R/S,S) isomers of the ligand. The oily hexanes-soluble residue remaining after recrystallization can be shown by ¹H and ¹³C NMR to be the C₁ symmetric complex derived from the meso (R,S) ligand precursor. Attempts to obtain this second isomer in pure form for comparative structural and catalytic studies were frustrated by its oily nature and the difficulty in removing the last traces of the C₂ isomer. Characterizing data for crystalline product (R,R/S,S): ¹H NMR (C₆D₆, 25 °C): δ 1.15 (s, 18H, ^tBu), 1.32 (d, J = 6.3 Hz, 6H, CH₃ of ⁱPr), 1.38 (d, J = 9.0 Hz, 2H, NCH₂CH₂N), 1.50 (d, J = 6.0 Hz, 6H, CH₃ of ⁱPr), 2.14 (dd, J = 10.7, 4.1 Hz, 2H, NCH₂CHO), 2.22 (s, 6H, NCH₃), 2.74 (d, J = 9.0 Hz, 2H, NCH₂CH₂N), 3.09 (virtual t, J = 11.0 Hz, 2H, CH₂CHO), 3.84 (dd, J = 10.6, 4.1 Hz, 2H, NCH₂CHO), 5.14 (sept, J = 6.0 Hz, 2H, CH of ⁱPr). ¹³C NMR (C₆D₆, 25 °C): δ 26.0, 26.7 (CH₃ of ⁱPr), 26.4 (CH₃ of ^tBu), 36.5 (^tBu quaternary), 45.3 (NCH₃), 51.9 (NCH₂ backbone), 62.8 (NCH₂ ligand arm), 74.6 (CH of ⁱPr), 84.2 (CHO). Mass spectrum (EI): 452.4 M⁺. Exact mass. Calc. for TiC₂₂H₄₈N₂O₄ 452.3088: Found: 452.3079. Elemental analysis. Calc. for TiC₂₂H₄₈N₂O₄: C, 58.40; H, 10.69; N, 6.19. Found: C, 58.01; H, 10.14; N, 6.24%. ¹H and ¹³C NMR data for oily product (R,S): ¹H NMR (C₆D₆, 25 °C): δ 0.93 (s, 9H, ^tBu), 1.03 (s, 9H, ^tBu), 1.29 (d, J = 6.0 Hz, 3H, CH₃ of ⁱPr), 1.33 (d, J = 6.0 Hz, 3H, CH₃ of ⁱPr), 1.39 (d, J = 3.0 Hz, 6H, CH₃ of ⁱPr), 1.44 (d, J = 3.0 Hz, 6H, CH₃ of ⁱPr), 1.71 (dd, J = 12.4, 4.1 Hz, 1H, NCH₂CHO), 1.85 (dd, J = 14.0, 4.0 Hz, 1H, NCH₂CHO), 2.16 (m, 2H, NCH₂CH₂N), 2.41 (s, 3H, NCH₃), 2.42 (s, 3H, NCH₃), 2.75 (m, 2H, NCH₂CH₂N), 3.06 (virtual t, J = 11.0 Hz, 1H, CH₂CHO), 3.13 (virtual t, J = 11.0 Hz, 1H, CH₂CHO), 3.79 (dd, J = 10.8, 4.4 Hz, 1H, NCH₂CHO), 4.33 (dd, J = 9.8, 5.2 Hz, 1H, NCH₂CHO), 4.96 (sept, J = 6.0 Hz, 2H, coincident CH of ⁱPr). ¹³C NMR (C₆D₆, 25 °C): δ 26.0, 26.2 (CH₃ of ^tBu), 25.9, 26.2, 26.6, 26.9 (CH₃ of ⁱPr), 34.6 (coincident ^tBu quaternary carbons), 47.9, 50.2 (NCH₃), 56.7, 59.2 (NCH₂ backbone), 59.4, 66.0 (NCH₂ ligand arm), 74.7, 74.8 (CH of ⁱPr), 84.8, 85.1 (CHO).

(L²)₂Zr. (L²)₂Zr was prepared from zirconium tetrakis(propoxide) (0.78 ml of a 70% wt solution in propanol, 1.74 mmol) and H₂L² (0.50 g, 1.74 mmol) using the method detailed above for L¹Ti₂(OⁱPr)₆, and isolated as colourless crystals from a concentrated hexanes solution. In contrast to the crude reaction mixture, the crystalline product gives rise to relatively simple ¹H and ¹³C NMR spectra indicating only two distinct ^tBu and NMe resonances, with subsequent crystallographic analysis revealing both (L²)²⁻ ligands to be of meso (R,S) stereochemistry. The yield was subsequently optimised by consideration of the 1 : 2 Zr : L² ratio to 39% (ca. 1.00 g scale). ¹H NMR (C₆D₆, 25 °C): δ 1.08 (s, 18H, ^tBu), 1.10 (s, 18H, ^tBu), 1.81 (m, 4H, NCH₂CHO), 2.14 (s, 6H, NCH₃), 2.34 (m, 4H, NCH₂CH₂N), 2.80 (m, 4H, NCH₂CH₂N), 2.91 (s, 6H, NCH₃), 3.14 (t, J = 12.0 Hz, 2H, CH₂CHO), 3.29 (t, J = 11.6 Hz, 2H, CH₂CHO), 3.78 (dd, J = 2.6, 11.1 Hz, 2H, NCH₂CHO), 4.09 (dd, J = 5.1, 11.4 Hz, 2H, NCH₂CHO). ¹³C NMR (C₆D₆, 25 °C): δ 26.9, 27.2 (CH₃ of ^tBu), 35.2, 35.3 (^tBu quaternary), 43.4, 48.6 (NCH₃), 57.2, 58.9 (NCH₂CH₂N), 60.4, 64.5 (NCH₂), 78.9, 83.9 (CHO). Mass spectrum (FAB): 661.3 (M − H)⁺. Exact mass. Calc. for ZrC₃₂H₆₈N₄O₄: 661.4204. Found: 661.4187. Elemental analysis. Calc. for ZrC₃₂H₆₈N₄O₄: C, 57.87; H, 10.32; N, 8.44. Found: C, 57.41; H, 10.00; N, 8.59%.

(iv) Crystallographic method

Data for L¹Ti₂(OⁱPr)₆, L¹₂Ti₂(OEt)₂(μ-O), L²Ti(OⁱPr)₂ and L²₂Zr were collected on an Bruker Nonius Kappa CCD diffractometer. Data collection and cell refinement were carried out using DENZO and COLLECT; structure solution and refinement used SHELXS-97 and SHELXL-97, respectively; absorption corrections were performed using SORTAV.⁹ Details of each data collection, structure solution and refinement can be found in Table 1. Relevant bond lengths and angles are included in the figure captions and complete details of each structure have been deposited with the CCDC (numbers as listed in Table 1). In addition, complete details for each structure have been included in the supporting information. The two OEt ligands in L¹₂Ti₂(OEt)₂(μ-O) are disordered and were modeled over two orientations (70 : 30 and 53 : 47 occupancy factors) and refined isotropically. The structure was treated as a racemic twin with the BASF parameter refining to 0.4.

Table 1 Details of data collection, structure solution and refinement for L¹Ti₂(OⁱPr)₆, L¹₂Ti₂(OEt)₂(μ-O), L²Ti(OⁱPr)₂ and L²₂Zr (refinement method: full-matrix least squares (F²))

Compound	L^1Ti2(O^iPr)6	L^1 2Ti2(OEt)2(μ-O)	L^2Ti(O^iPr)2	L^2 2Zr
Empirical formula	C35H76N2O8Ti2	C38H78N4O7Ti2	C22H48NO4Ti	C32H68N4O4Zr
M r	748.78	798.84	452.52	664.12
T/K	150(2)	120(2)	120(2)	120(2) K
CCDC no.	286 915	286 913	286 916	286 914
λ/Å	0.71073	0.71073	0.71073	0.71073
Crystal system	Triclinic	Orthorhombic	Triclinic	Triclinic
Space group	P1̄	Pca21	P1̄	P1̄
a/Å	11.0190(5)	19.7850(10)	8.830(5)	10.762(2)
b/Å	14.3144(6)	11.7253(4)	9.726(5)	12.115(2)
c/Å	15.7604(8)	19.6134(7)	17.047(5)	15.963(3)
α/°	70.240(3)	90	93.801(5)	76.68(3)
β/°	88.145(2)	90	98.098(5)	72.59(3)
γ/°	71.787(2)	90	112.844(5)	68.58(3)
V/Å^3	2214.83(18)	4550.0(3)	1323.9(11)	1831.5(6)
D c/Mg m^−3	1.123	1.166	1.135	1.204
Z	2	4	2	2
μ/mm^−1	0.404	0.397	0.349	0.337
F(000)	816	1736	496	720
Crystal size/mm	0.20 × 0.18 × 0.15	0.20 × 0.15 × 0.05	0.30 × 0.20 × 0.02	0.20 × 0.20 × 0.20
θ Range/°	3.05–25.39	2.92–25.02	2.98–25.03	2.92–27.45
Index ranges hkl	−13 to 13, −17 to 17, −18 to 19	−23 to 18, −10 to 13, −23 to 21	−10 to 10, −11 to 11, −20 to 20	−13 to 13, −15 to 15, −10 to 20
Reflections collected	31 559	22 572	13 694	36 431
Independent reflections (Rint)	8020 (0.1319)	7502 (0.0809)	4583 (0.0501)	8322 (0.0562)
Completeness to θmax (%)	98.5	99.2	97.8	99.4
Absorption correction	SORTAV	Semi-empirical from equivs	Semi-empirical from equivs	SORTAV
Max., min. transmission	0.9419, 0.9236	0.9804, 0.9249	0.9930, 0.9025	0.9356, 0.8956
Data/restraints/parameters	8020/0/442	7502/9/471	4583/0/275	8322/0/407
Goodness-of-fit on F^2	1.002	1.008	1.030	1.032
Final R indices [I > 2σ(I)]	R1 = 0.0684, wR2 = 0.1524	R1 = 0.0607, wR2 = 0.1248	R1 = 0.0416, wR2 = 0.0881	R1 = 0.0373, wR2 = 0.0851
R Indices (all data)	R1 = 0.1451, wR2 = 0.1758	R1 = 0.1033, wR2 = 0.1410	R1 = 0.0538, wR2 = 0.0932	R1 = 0.0458, wR2 = 0.0886
Δρmax, min/e Å^−3	0.451, −0.427	0.618, −0.314	0.209, −0.366	1.117, −0.520

---

CCDC reference numbers 286913–286916.

For crystallographic data in CIF or other electronic format see DOI: 10.1039/b608680b

Results and discussion

(i) Ligand precursors

The ligand precursors H₂L¹ and H₂L² have been synthesised via the nucleophilic ring opening of racemic (±)-3,3-dimethyl-1,2-epoxybutane by the cyclic secondary amine 1,4-diazacycloheptane, or the related linear species N,N′-dimethylethylenediamine (to give H₂L¹ and H₂L², respectively; Scheme 2). Forcing conditions (excess epoxide, acetonitrile solution at 100 °C in a pressure tube) are required to drive the reaction to completion, the analogous chemistry in ethanol at room temperature leading to incomplete conversion. Similar chemistry can be also applied to 1,4-diazacyclohexane to yield the related compound N,N′-bis(2-hydroxy-3,3-dimethylbutyl)-1,4-diazacyclohexane.⁸ H₂L¹ (like its diaminocyclohexane analogue) is a white crystalline solid which has been characterised by multinuclear NMR and IR spectroscopies, mass spectrometry and elemental analysis. In both cases, the ¹H and ¹³C NMR spectra (in chloroform-d) of crystalline ligand samples obtained by direct cooling of the acetonitrile reaction mixture, show single resonances for the tert-butyl methyl groups, indicating the presence of only one pair of diastereomers [i.e. either rac (R,R/S,S) or meso (R,S)]. In the case of N,N′-bis(2-hydroxy-3,3-dimethylbutyl)-1,4-diazacyclohexane, a crystallographic study reveals a centrosymmetric space group, in which each molecule lies on a centre of inversion coincident with the centre of the 1,4-diazacyclohexane ring. Consequently, this ligand necessarily features the meso (R,S) stereochemistry.⁸ Although crystalline H₂L¹ can also be obtained by direct cooling of the acetonitrile reaction mixture, the crystals so obtained are not suitable for X-ray diffraction. In this case, however, crystallographic studies (i) of alkoxo-titanium complexes formed by reaction of H₂L¹ with Ti(OR)₄ (R = ⁱPr, Et) (vide infra); and (ii) of the complex [trans-(H₂L¹)Ni(OH₂)₂]²⁺ isolated as the bis(perchlorate) salt from the subsequent reaction of crystalline H₂L¹ with Ni(ClO₄)₂·6H₂O in ethanol,¹⁰ confirm that H₂L¹ crystallises from the reaction mixture as the rac (R,R/S,S) pair of diastereomers. Yields of crystalline rac-H₂L¹ and meso-N,N′-bis(2-hydroxy-3,3-dimethylbutyl)-1,4-diazacyclohexane are typically of the order of 40%, reflecting the fact that in each case the alternative pair of diastereomers remains in solution. In contrast, H₂L² can only be obtained as an oily liquid at room temperature (cooling to a waxy solid at −25 °C). ¹H and ¹³C NMR spectra clearly indicate the presence of both pairs of diastereoisomers, the separation of which by fractional crystallisation proves impossible. Consequently, separation of diastereomeric species at the metal complex stage, making use of the differential solubilities of more tractable derivatives was thought to offer a more practical methodology (vide infra).

(ii) Complexes of group 4 metals

The coordination chemistries of ligands (L¹)²⁻ and (L²)²⁻ with respect to group 4 metal centres have been investigated via the reactions of H₂L¹ and H₂L² with homoleptic titanium and zirconium tetrakis(alkoxide) precursors (Schemes 3 and 4).

Scheme 3 Syntheses of dinuclear titanium complexes L¹Ti₂(OⁱPr)₆ and L¹₂Ti₂(OEt)₂(μ-O). Reagents and conditions: (i) Ti(OⁱPr)₄ (2 equiv.), neat reagents, 20 °C, 4 h, 64%; (ii) Ti(OEt)₄ (1.05 equiv.), neat reagents, adventitious water, 20 °C, 4 h, <5%.

Scheme 4 Syntheses of mononuclear group 4 complexes L²Ti(OⁱPr)₂ and L²₂Zr. Reagents and conditions: Ti(OⁱPr)₄ or Zr(OⁿPr)₄ (1 equiv.), neat reagents, 20 °C, 4 h, separation of isomers by recrystallization from hexanes, 44 and 39%, respectively (for crystalline products).

(a) Coordination chemistry of (L¹)²⁻. The reactions of H₂L¹ with titanium alkoxides lead to the formation of dinuclear complexes in which the tetradentate ligand bridges between two Ti(IV) centres. Thus, the reaction of rac-H₂L¹ with titanium tetrakis(isopropoxide) at room temperature in the absence of solvent gives a white crystalline solid, for which the integration of ^tBu and ⁱPr ¹H NMR signals (1 : 2) implies an L¹ : OⁱPr ratio of 1 : 6. The formulation L¹Ti₂(OⁱPr)₆ is given further credence by the results of mass spectrometry experiments and has been confirmed crystallographically (Fig. 1 and Table 1).

Fig. 1 Structure of L¹Ti₂(OⁱPr)₆; hydrogen atoms have been omitted for clarity and ORTEP ellipsoids set at the 30% probability level. Important bond lengths (Å) and angles (°): Ti(1)–O(1) 1.853(3), Ti(1)–O(2) 1.784(3), Ti(1)–O(3) 1.821(3), Ti(1)–O(4) 1.820(3), Ti(1)–N(1) 2.378(3), Ti(2)–N(2) 2.391(3); O(1)–Ti(1)–O(3) 119.3(1), O(1)–Ti(1)–O(4) 118.4(1); O(3)–Ti(1)–O(4) 115.9(1), O(2)–Ti(1)–N(1) 171.0(1), O(6)–Ti(2)–N(2) 170.38(12).

The X-ray crystal structure of L¹Ti₂(OⁱPr)₆ shows that the complex contains two Ti(OⁱPr)₃ fragments linked by a single bridging (L¹)²⁻ ligand which binds to each metal centre through one alkoxide and one tertiary amine donor. Each titanium centre is thus five coordinate, with three isopropoxide ligands being retained from the starting material. The geometry around each titanium atom appears to be a distorted trigonal bipyramid with the L¹ amine donor and one isopropoxide ligand occupying the axial sites [O(2)–Ti(1)–N(1) 171.0(1)°, O(6)–Ti(2)–N(2) 170.4(1)°]. The L¹-derived alkoxide donor and the two remaining isopropoxides occupy the three equatorial sites [O(1)–Ti(1)–O(3) 119.3(1)°, O(1)–Ti(1)–O(4) 118.4(1)°; O(3)–Ti(1)–O(4) 115.9(1)°]. In general terms, the Ti–O distances are within the bounds expected for five-coordinate titanium complexes of this type, with the Ti–N distance [2.385 Å (mean)] being, if anything, slightly longer than those found in comparable systems. Thus, for example, a useful comparison can be made with the corresponding bond lengths found in the complexes N[CH₂(3,5-Me₂C₆H₂)O]₃TiOR [R = C₆H₃ⁱPr₂-2,6, d(Ti–O) = 1.833 Å (mean), d(Ti–N) = 2.305(2) Å; R = ⁱPr, d(Ti–O) = 1.831 Å (mean), d(Ti–N) = 2.295(3) Å] and N[CH₂(3,5-Me₂C₆H₂)O]₂(CH₂CH₂O)TiOR [R = C₆H₃ⁱPr₂-2,6, d(Ti–O) = 1.822 (mean), d(Ti–N) = 2.288(3) Å], each of which features an analogous (approximately trigonal bipyramidal) Ti(OR)₄(NR₃) unit, with the amine donor occupying one of the axial positions.¹¹

In a bid to obtain a complex of the desired composition, i.e.L¹₂Ti(OR)₂, the reaction stoichiometry was varied (using a large excess of H₂L¹ under more forcing conditions); the same dinuclear complex was isolated. Reduction in the steric bulk of the alkoxide co-ligands was therefore investigated in order to probe whether this would allow for the coordination of more than one (L¹)²⁻ moiety. Reaction of H₂L¹ with titanium tetrakis(ethoxide) under similar conditions yields colourless crystals in low yield after recrystallisation from hexane. Integration of the ^tBu and Et ¹H NMR signals implies an L¹ : OEt ratio of 1 : 1, and the formulation L¹₂Ti₂(OEt)₂(μ-O) demonstrated by mass spectrometry has been confirmed crystallographically (Fig. 2 and Table 1).

Fig. 2 Structure of L¹₂Ti₂(OEt)₂(μ-O); hydrogen atoms have been omitted for clarity and ORTEP ellipsoids set at the 30% probability level. Important bond lengths (Å) and angles (°): Ti(1)–O(7) 1.826(4), Ti(2)–O(7) 1.818(4), Ti(1)–O(1) 1.860(4), Ti(1)–O(2) 1.860(4), Ti(1)–O(3) 1.835(3), Ti(2)–O(4) 1.852(3), Ti(2)–O(5) 1.840(4), Ti(2)–O(6) 1.827(4), Ti(1)–N(1) 2.506(4), Ti(1)–N(2) 2.757(4), Ti(2)–N(3) 2.443(4), Ti(2)–N(4) 2.962(4); O(6)–Ti(2)–N(3) 165.3(2), O(4)–Ti(2)–O(5) 109.4(2), O(4)–Ti(2)–O(7) 116.6(2), O(5)–Ti(2)–O(7) 125.0(2).

The crystal structure of L¹₂Ti₂(OEt)₂(μ-O) confirms that a 1 : 1 ratio of (L¹)²⁻ to Ti(IV) can be achieved, albeit with the ligand still adopting a bridging, rather than chelating mode of binding. Noteworthy is the linking of the two titanium centres via a symmetrically bridging oxo ligand [Ti(2)–O(7) 1.818(4) Å, Ti(1)–O(7) 1.826(4) Å], which is presumably derived from conversion of titanium-bound OEt ligands to OH (by adventitious water) followed by a condensation step. The coordination sphere at each titanium also features one ethoxide ligand remaining from the starting material and two alkoxide linkages (one from each L¹ ligand). Each metal centre is also engaged in two disparate Ti–N interactions [e.g. Ti(2)–N(3) 2.443(4), Ti(2)–N(4) 2.962(4) Å]. The shorter Ti–N distance is relatively long for a N→Ti donor/acceptor interaction featuring a Ti(OR)₄ type Lewis acid,¹¹ whereas the longer Ti–N distance falls outside the sum of conventional covalent radii for N and Ti.¹² The geometry at each titanium centre can therefore probably best be considered as intermediate between trigonal bipyramidal and octahedral, being formally derived from a regular trigonal bipyramid by approach of the weakly bound sixth donor in the equatorial plane (Scheme 5). The axial positions are occupied by the more tightly bound amine donor and the ethoxide ligand [O(6)–Ti(2)–N(3) 165.3(2)°]. The two remaining L¹ alkoxide donors and the bridging oxygen constitute the trigonal plane (sum of the O–Ti–O angles 351.0°) in which distortions from 120° angles presumably reflect, at least in part, the approach of the second, weakly-bound amine donor [O(4)–Ti(2)–O(5) 109.4(2)°; O(4)–Ti(2)–O(7) 116.6(2)°; O(5)–Ti(2)–O(7) 125.0(2)°].

Scheme 5 Schematic representation of the coordination environment at titanium in L¹₂Ti₂(OEt)₂(μ-O).

Preliminary results therefore imply that the relatively rigid ligand backbone in (L¹)²⁻ prevents it from binding in a chelating fashion to small transition metal centres such as Ti(IV). Furthermore, although the formation of a Ni(II) complex containing the diprotonated ligand H₂L¹ has demonstrated the possibility of chelation with larger metals, the ancillary water ligands in [(H₂L¹)Ni(OH₂)₂][ClO₄]₂ are coordinated in the undesirable trans orientation.¹⁰ The Ti(IV) and Zr(IV) coordination chemistries of the more flexible system (L²)²⁻ were therefore targeted.

(b) Coordination chemistry of (L²)²⁻. In the case of ligand (L²)²⁻, the precursor H₂L² is obtained as an inseparable mixture of rac and meso diastereomers and utilised as such for reactions with Ti(OⁱPr)₄ and Zr(OⁿPr)₄ (see Scheme 5). The reaction of H₂L² with titanium tetrakis(isopropoxide) gives two products, as determined from the ¹H and ¹³C NMR spectra of the crude reaction mixture. Careful recrystallization from hexanes (twice) allows the isolation of a colourless crystalline solid, for which ¹H and ¹³C NMR and mass spectrometry indicate a formulation as L²Ti(OⁱPr)₂. The ¹³C NMR spectrum (see ESI†) shows only nine signals, indicating that in this complex the two halves of the ligand are symmetry related. This inference is confirmed by the results of a single-crystal X-ray diffraction study (Fig. 3 and Table 1) which confirms local (non-crystallographic) C₂ symmetry and that the coordinated ligands are of rac stereochemistry (both R,R and S,S enantiomers being present in the crystal lattice). Overall, the titanium centre is six-coordinate (a distorted octahedron) and bound to a single L² ligand, with two mutually cis isopropoxide ligands remaining from the starting material. Thus it would appear that the greater flexibility of the linear (L²)²⁻ ligand [cf. cyclic analogue (L¹)²⁻] allows for a chelating mode of coordination, even for Ti(IV), and that the simple synthesis/isolation of rac-L²Ti(OⁱPr)₂ from wholly racemic precursors offers a convenient route to C₂ symmetric geometrically cis complexes.

Fig. 3 Structure of L²Ti(OiPr)₂; hydrogen atoms have been omitted for clarity and ORTEP ellipsoids set at the 30% probability level. Important bond lengths (Å) and angles (°): Ti(1)–O(1) 1.906(2), Ti(1)–O(2) 1.901(2), Ti(1)–O(3) 1.836(2), Ti(1)–O(4) 1.834(2), Ti(1)–N(1) 2.325(2), Ti(1)–N(2) 2.313(2); O(1)–Ti(1)–O(2) 163.03(6), O(3)–Ti(1)–O(4) 107.70(7), N(1)–Ti(1)–N(2) 76.20(6).

The molecular structure of rac-L²Ti(OⁱPr)₂ contains general structural features characteristic of Ti(IV) complexes of chelating bis(tertiary amine) bis(phenolate) ligands,⁴ and closely resembles the complexes L^CF₃MX₂ (M = Ti, X = Cl; M = Zr, X = CH₂Ph) recently synthesized by Carpentier and co-workers containing a diaminodialkoxo ligand system derived from the related but achiral diol [HOC(CF₃)₂CH₂N(Me)CH₂]₂ (H₂L^CF₃).^6b Thus the transoid O(1)–Ti(1)–O(2) and cisoid N(1)–Ti(1)–N(2) angles associated with the L²Ti unit [163.03(6) and 76.20(6)°, respectively] and the related Ti–O and Ti–N distances [1.903 (mean) and 2.319 Å (mean)] are essentially identical to the corresponding parameters for L^CF₃TiCl₂.^6b Encouraged by the reported synthesis of complexes of the type L^CF₃ZrX₂ we also examined the reactivity of H₂L² towards Zr(IV) precursors. Reaction with zirconium tetrakis(propoxide), however, yields the 1 : 2 complex, Zr(L²)₂ irrespective of reaction stoichiometry, which crystallizes from hexane solution with both L² ligands of meso (R,S) stereochemistry (Fig. 4). Similar reactivity to generate eight coordinate Zr(IV) complexes of the type ZrL₂ is well precedented,¹³ with slight lengthening in the Zr–O [2.056(2)–2.073(1) Å] and Zr–N bonds [2.595(2)–2.621(2) Å] with respect to L^CF₃Zr(CH₂Ph)₂ presumably reflecting increases in steric crowding at the metal centre.^6b

Fig. 4 (a) Structure of L²₂Zr; hydrogen atoms have been omitted for clarity and ORTEP ellipsoids set at the 30% probability level. Important bond lengths (Å) and angles (°): Zr(1)–O(1) 2.060(2), Zr(1)–O(2) 2.057(2), Zr(1)–O(3) 2.073(2), Zr(1)–O(4) 2.056(2), Zr(1)–N(1) 2.620(2), Zr(1)–N(2) 2.621(2), Zr(1)–N(3) 2.595(2), Zr(1)–N(4) 2.616(2); O(1)–Ti(1)–O(2) 88.99(6), O(3)–Ti(1)–O(4) 87.69(6), N(1)–Ti(1)–N(2) 67.43(6), N(3)–Ti(1)–N(4) 67.62(6); (b) structure of L²₂Zr emphasizing the coordination geometry at the metal centre.

Conclusions

Epoxide ring opening chemistry using 1,4-diazacycloheptane (or its cyclic six-membered analogue 1,4-diazacyclohexane), or the related linear species N,N′-dimethylethylenediamine, and racemic (±)-3,3-dimethyl-1,2-epoxybutane gives single-step access to pendant alcohol functionalised ditertiary amine pro-ligands. Thus, [HOC(H)^tBuCH₂N(R)CH₂]₂ [R₂ = CH₂CH₂CH₂ (H₂L¹) and CH₂CH₂] can be isolated as diastereomerically pure crystalline materials from the reaction mixture, while oily H₂L² (R₂ = Me₂) is obtained as a mixture of rac and meso isomers. The contrasting reactions of H₂L¹ and H₂L² towards homoleptic group 4 alkoxides highlight the crucial role of ligand backbone flexibility in complex formation. Thus, the chemistry of the more conformationally rigid system (L¹)²⁻ appears to be constrained by the cyclic ligand core, such that it adopts a bridging (μ₂:η²,η²) mode of coordination towards Ti(IV), leading to dinuclear metal systems [e.g.L¹Ti₂(OⁱPr)₆]. By contrast, the more flexible linear system (L²)²⁻ binds to both Ti(IV) and Zr(IV) in a chelating fashion leading, for example, to the syntheses of the mononuclear 1 : 1 complex L²Ti(OⁱPr)₂ and the 1 : 2 Zr(IV) complex Zr(L²)₂. Although H₂L² is necessarily employed as a mixture of rac and meso diastereomers in its reaction with Ti(OⁱPr)₄, simple crystallization yields solely the C₂ symmetric isomer of L²Ti(OⁱPr)₂, featuring the rac form of the ligand. A simple procedure for the synthesis and isolation of diastereomerically pure C₂ symmetric Ti(IV) complexes of cis geometry from racemic precursors is therefore presented.

Acknowledgements

We acknowledge the EPSRC and Johnson Matthey for funding this project, and the assistance of the EPSRC National Mass Spectrometry Service Centre (Swansea).

References

1. For recent reviews, see: (a) P. Corradini, G. Guerra and L. Cavallo, Acc. Chem. Res., 2004, 37, 231; (b) J. Gromada, A. Mortreux and J.-F. Carpentier, Coord. Chem. Rev., 2004, 248, 397; (c) Y. Suzuki, H. Terao and T. Fujita, Bull. Chem. Soc. Jpn., 2003, 76, 1493; (d) V. C. Gibson and S. K. Spitzmesser, Chem. Rev., 2003, 103, 283; (e) G. W. Coates, J. Chem. Soc., Dalton Trans., 2002, 467; (f) G. J. P. Britovsek, V. C. Gibson and D. F. Wass, Angew. Chem., Int. Ed., 1999, 38, 428.
2. (a) L. G. Hubert-Pfalzgraf, Coord. Chem. Rev., 1998, 178–180, 967; (b) R. M. Mehrotra and A. Singh, Prog. Inorg. Chem., 1997, 46, 239; (c) D. C. Bradley, R. M. Mehrotra, I. P. Rothwell and A. Singh, in Alkoxo and Aryloxo Derivatives of Metals, Academic Press, London, 2001.
3. For examples of sterically encumbered chelating bis(aryloxide) ligands, see: (a) S. Fokken, T. P. Spaniol, J. Okuda, F. G. Sernetz and R. Mullhaupt, Organometallics, 1997, 16, 4240; (b) A. van der Linden, C. J. Schaverien, N. Meijboom, C. Ganter and A. G. Orpen, J. Am. Chem. Soc., 1995, 117, 3008.
4. For recent examples of chelating bis(aryloxide) ligands featuring additional donor centres, see: (a) A. J. Chmura, M. G. Davidson, M. D. Jones, M. D. Lunn and M. F. Mahon, Chem. Commun., 2006, 887; (b) C. Capacchione, R. Manivannan, M. Barone, K. Beckerle, R. Centore, L. Oliva, A. Proto, A. Tuzi, T. P. Spaniol and J. Okuda, Organometallics, 2005, 24, 2971; (c) C. L. Boyd, T. Toupance, B. R. Tyrrell, B. D. Ward, C. R. Wilson, A. R. Cowley and P. Mountford, Organometallics, 2005, 24, 309; (d) S. Segal, I. Goldberg and M. Kol, Organometallics, 2005, 24, 200; (e) L. S. Natrajan, C. Wilson, J. Okuda and P. L. Arnold, Eur. J. Inorg. Chem., 2004, 3724; (f) P. D. Knight, I. J. Munslow, P. N. O’Shaughnessy and P. Scott, Chem. Commun., 2004, 894; (g) C. Capacchione, F. De Carlo, C. Zannoni, J. Okuda and A. Proto, Macromolecules, 2004, 37, 8918; (h) S. Groysman, I. Goldberg, M. Kol, E. Genizi and Z. Goldschmidt, Organometallics, 2003, 22, 3013; (i) S. Groysman, I. Goldberg, M. Kol, E. Genizi and Z. Goldschmidt, Inorg. Chim. Acta, 2003, 345, 137; (j) V. Reimer, T. P. Spaniol, J. Okuda, H. Ebeling, A. Tuchbreiter and R. Mülhaupt, Inorg. Chim. Acta, 2003, 345, 221; (k) C. Capacchione, A. Proto, H. Ebeling, R. Mülhaupt, K. Möller, T. P. Spaniol and J. Okuda, J. Am. Chem. Soc., 2003, 125, 4964; (l) Y. Nakayama, H. Bando, Y. Sonobe, H. Kaneko, N. Kashiwa and T. Fujita, J. Catal., 2003, 215, 171; (m) E. Y. Tshuva, S. Groysman, I. Goldberg, M. Kol and Z. Goldschmidt, Organometallics, 2002, 21, 662; (n) S. Fokken, F. Reichwald, T. P. Spaniol and J. Okuda, J. Organomet. Chem., 2002, 663, 158; (o) M. C. W. Chan, K.-H. Tam, Y.-L. Pui and N. Zhu, J. Chem. Soc., Dalton Trans., 2002, 3085; (p) S. Matsui and T. Fujita, Catal. Today, 2001, 66, 61; (q) E. Y. Tshuva, I. Goldberg, M. Kol and Z. Goldschmidt, Organometallics, 2001, 20, 3017; (r) S. Y. Bylikin, D. A. Robson, N. A. H. Male, L. H. Rees, P. Mountford and M. Schröder, J. Chem. Soc., Dalton Trans., 2001, 170; (s) E. Y. Tshuva, I. Goldberg, M. Kol and Z. Goldschmidt, Inorg. Chem. Commun., 2000, 3, 611; (t) E. Y. Tshuva, I. Goldberg, M. Kol, H. Weitman and Z. Goldschmidt, Chem. Commun., 2000, 379; (u) E. Y. Tshuva, I. Goldberg and M. Kol, J. Am. Chem. Soc., 2000, 122, 10706.
5. For recent examples of aryloxide ligands featuring Schiff base or related pendant donors, see: (a) I. Westmoreland, I. J. Munslow, A. J. Clarke, G. Clarkson, R. J. Deeth and P. Scott, J. Organomet. Chem., 2006, 691, 2228; (b) R. K. J. Bott, M. Mammond, P. N. Horton, S. J. Lancaster, M. Bochmann and P. Scott, Dalton Trans., 2005, 3611; (c) D. A. Pennington, S. J. Coles, M. B. Hursthouse, M. Bochmann and S. J. Lancaster, Chem. Commun., 2005, 3150; (d) M. Sanz, T. Cuenca, M. Galakhov, A. Grassi, R. K. J. Bott, D. L. Hughes, S. J. Lancaster and M. Bochmann, Organometallics, 2004, 23, 5324; (e) C. Cuomo, M. Strianese, T. Cuenca, M. Sanz and A. Grassi, Macromolecules, 2004, 37, 7469; (f) A. F. Mason and G. W. Coates, J. Am. Chem. Soc., 2004, 126, 16326; (g) D. C. H. Oakes, B. S. Kimberley, V. C. Gibson, D. J. Jones, A. J. P. White and D. J. Williams, Chem. Commun., 2004, 2174; (h) P. D. Knight, P. N. O’Shaughnessy, I. J. Munslow, B. S. Kimberley and P. Scott, J. Organomet. Chem., 2003, 683, 103; (i) D. Owiny, S. Parkin and F. T. Lapido, J. Organomet. Chem., 2003, 678, 134; (j) S. Reinartz, A. F. Mason, E. B. Lobkovsky and G. W. Coates, Organometallics, 2003, 22, 2542; (k) R. Furuyama, J. Saito, S. Ishii, M. Mitani, S. Matsui, Y. Tohi, H. Makio, N. Matsukawa, H. Tanaka and T. Fujita, J. Mol. Catal. A: Chem., 2003, 200, 31; (l) P. D. Hustad and G. W. Coates, J. Am. Chem. Soc., 2002, 124, 11578; (m) P. D. Knight, A. J. Clarke, B. S. Kimberley, R. A. Jackson and P. Scott, Chem. Commun., 2002, 352; (n) S. Matsui, M. Mitani, J. Saitoo, Y. Tohi, H. Makio, N. Matsukawa, Y. Takagi, K. Tsuru, M. Nitabaru, T. Nakano, H. Tanaka, N. Kasiwa and T. Fujita, J. Am. Chem. Soc., 2001, 123, 684; (o) M. Wang, H. Zhu, K. Jin, D. Dai and L. Sun, J. Catal., 2001, 220, 392; (p) J. Balsells, P. J. Carroll and P. J. Walsh, Inorg. Chem., 2001, 40, 5568; (q) P. R. Woodman, N. W. Alcock, I. J. Munslow, C. J. Sanders and P. Scott, J. Chem. Soc., Dalton Trans., 2000, 3340.
6. For recent examples of alkoxide ligands featuring pendant donors, see: (a) M. E. G. Skinner, T. Toupance, D. A. Cowhig, B. R. Tyrrell and P. Mountford, Organometallics, 2005, 24, 5586; (b) L. Lavanant, T.-Y. Chou, Y. Chi, C. W. Lehmann, L. Toupet and J.-F. Carpentier, Organometallics, 2004, 23, 5450; (c) I. J. Munslow, A. J. Clarke, R. J. Deeth, I. Westmoreland and P. Scott, Chem. Commun., 2002, 1868; (d) R. Manivannan and G. Sundararajan, Macromolecules, 2002, 35, 7883; (e) Y. Kim, Y. Han and Y. Do, J. Organomet. Chem., 2001, 634, 19; (f) D. A. Robson, S. Y. Bylikin, M. Cantuel, N. A. H. Male, L. H. Rees, P. Mountford and M. Schröder, J. Chem. Soc., Dalton Trans., 2001, 157; (g) D. A. Robson, L. H. Rees, P. Mountford and M. Schröder, Chem. Commun., 2000, 1269; (h) R. Manivannan, G. Sundararajan and W. Kaminsky, J. Mol. Catal. A: Chem., 2000, 160, 85; (i) R. M. Gauvin, J. A. Osborn and J. Kress, Organometallics, 2000, 19, 2944; (j) P. Shao, R. A. L. Gendron, D. J. Berg and G. W. Bushnell, Organometallics, 2000, 18, 509.
7. P. D. Knight and P. Scott, Coord. Chem. Rev., 2003, 242, 125.
8. R. R. Strevens and I. A. Fallis, manuscript in preparation.
9. (a) DENZO in: Z. Otwinowski and W. Minor, Methods Enzymol., 1996, 276, 307; (b) Collect: Data Collection Software, R. Hooft, Nonius B. V., 1998; (c) G. M. Sheldrick, Acta Crystallogr., Sect. A, 1990, 46, 467; (d) G. M. Sheldrick, University of Göttingen, 1997; (e) Sortav: R. H. Blessing, Acta Crystallogr., Sect. A, 1995, 51, 33.
10. R. E. Baghurst, PhD thesis, Cardiff University, 2005.
11. (a) S. D. Bull, M. G. Davidson, A. L. Johnson, D. E. J. E. Robinson and M. F. Mahon, Chem. Commun., 2003, 1750; (b) M. Kol, M. Shamis, I. Goldberg, Z. Goldschmidt, S. Alfi and E. Hayut-Salant, Inorg. Chem. Commun., 2002, 4, 177; (c) Y. Kim and J. G. Verkade, Organometallics, 2002, 21, 2395.
12. J. Emsley in, The Elements, Oxford University Press, Oxford, 1995.
13. For a recent example featuring a bis(N-donor) bis(aryloxide) ligand system, see: T. Toupance, S. R. Dubberley, N. H. Rees, B. R. Tyrrell and P. Mountford, Organometallics, 2002, 21, 1367.

---

Footnote

† Electronic supplementary information (ESI) available: Synthetic and characterizing data for N,N′-bis(2-hydroxy-3,3-dimethylbutyl)-1,4-diazacyclohexane; full details of the crystal structures of L¹Ti₂(OⁱPr)₆, L¹₂Ti₂(OEt)₂(μ-O), L²Ti(OⁱPr)₂ and L²₂Zr; and ¹³C NMR spectra relevant to the separation of rac- and meso-L²Ti(OⁱPr)₂. See DOI: 10.1039/b608680b

---