Expression of chirality in salicyloxazolinate complexes of zirconium

Article abstract of DOI:10.1016/j.jorganchem.2005.10.055

Treatment of chiral non-racemic salicyloxazoline proligands HL with tetrabenzylzirconium(IV) gave species [L₂Zr(CH₂Ph)₂]. Reactions of KL with ZrCl₄(THF)₂ gave similar chloro complexes. One example of

Full text of DOI:10.1016/j.jorganchem.2005.10.055

Journal of Organometallic Chemistry 691 (2006) 2228–2236
www.elsevier.com/locate/jorganchem
Expression of chirality in salicyloxazolinate complexes of zirconium
1
Ian Westmoreland , Ian J. Munslow, Adam J. Clarke,
*
Guy Clarkson, Robert J. Deeth, Peter Scott
Department of Chemistry, University of Warwick, Gibbett Hill Road, Coventry CV4 7AL, UK
Received 12 October 2005; received in revised form 30 October 2005; accepted 30 October 2005
Available online 9 December 2005
Abstract
Treatment of chiral non-racemic salicyloxazoline proligands HL with tetrabenzylzirconium(IV) gave species [L2Zr(CH2Ph)2]. Reac-
tions of KL with ZrCl4(THF)2 gave similar chloro complexes. One example of a benzyl complex was shown to exist as the K-trans,cis,cis
diastereomer by X-ray crystallography. DFT calculations showed the observed isomer to be the most stable by 35 kJ molꢀ1, indicating
that thermodynamic diastereoselection for this species is excellent. Examination of the chiral environment about the benzyl co-ligands
indicates however that the degree of expression of the chirality of the structure in what would be the site of metal-based reactions is poor
1
in comparison to related systems. Variable temperature H NMR data are consistent with this in that the low temperature spectrum
exhibits a very small chemical shift difference between the chemically inequivalent benzylic CH atoms, and at higher temperatures dis-
sociation of the portion of the ligand that contains the chiral information, i.e., the oxazoline unit, leads to apparent equivalence.
ꢀ 2005 Elsevier B.V. All rights reserved.
Keywords: Zirconium; Oxazoline; Diastereoselection; Fluxionality
1. Introduction
bonds are principally electrostatic and the rates of
exchange are greater, very high levels of thermodynamic
discrimination are probably needed in order to make sys-
tems that are of practical value.
Octahedral coordination complexes may contain an ele-
ment of chirality by virtue of an enantiomeric arrangement
of, for example, three bidentate ligands about the metal
[1,2]. Using chiral nonracemic ligands, one enantiomer
may be formed preferentially as a result of kinetic or ther-
modynamic diastereoselection [3–7]. The situation is com-
plicated by the great structural variety possible in
organometallic and coordination compounds and also the
range in kinetic stabilities of coordination environments.
Accordingly, most studies on chiral metal complexes have
involved later transition elements where d-orbitals tend to
control the structure of the inner coordination sphere and
the rates of topographical exchange are generally slow
enough to be studied on convenient experimental time-
scales. For the early transition metals and f elements, where
The problem is deepened by the need for free or poten-
tially free coordination sites where useful reactions might
take place [6]. The synthesis of single diastereomeric chi-
ral-at-metal complexes for which this criterion is satisfied
has been achieved hitherto through the use of multidentate
ligands, although the design and synthesis of such coordi-
nation environments also presents synthetic challenges
[8]. Bidentate ligands are more readily available and more
diverse in nature than multidentate systems, and if the
aim is to produce chiral metal catalysts then, for e.g. group
4 metals, this points us toward octahedral complexes of the
type [M(AB)2X2], where AB is a chiral bidentate monoan-
ionic ligand. While this class of complex can in principle
form eight diastereomers we have shown that it is quite
possible to design highly diastereoselective systems for
zirconium using the bidentate N–N ligands aminopyridi-
nate[9,10] and aminooxazolinate [11]. Surprisingly high
kinetic stability and excellent diastereoselection was found
*
Corresponding author.
E-mail address: peter.scott@warwick.ac.uk (P. Scott).
Present address: College of Chemistry, University of California at
1
Berkeley, Berkeley, CA 94720-1460, USA.
0022-328X/$ - see front matter ꢀ 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2005.10.055

I. Westmoreland et al. / Journal of Organometallic Chemistry 691 (2006) 2228–2236
tBu
2229
tBu
results were obtained when the mixture was stirred for 48 h
following completion of this addition. We also found that
reaction performance was improved dramatically under
anhydrous conditions; the pre-drying of commercial I being
particularly beneficial. Experimentation with different sol-
vents tended to favour acetonitrile over dichloromethane,
while the effect of adding pyridine to the reaction mixtures
as described by Vorbruggen was not studied [16a]. We fur-
ther investigated reports by Sigman and Rajaram [18] that
di-iso-propylethylamine can provide results preferential to
those obtained with triethylamine, but were unable to dis-
cern any difference in this system.
R
OH
H2N
OH
tBu
O
tBu
3 PPh3, 4 Et3N, 10 CCl4
CH3CN 48 h
OH
N
OH
O
R
I
(S)-HL1, R = tBu
(S)-HL2, R = iPr
(S)-HL3, R = Ph
(i) KH/THF
(ii) ZrCl4(THF)2
Zr(CH2Ph)4
tBu
tBu
2.2. Synthesis of complexes
tBu
tBu
Treatment of the potentially bidentate proligand
(S)-HL1 with slightly less than half an equivalent of tetra-
benzylzirconium(IV) in toluene at ꢀ78 ꢁC gave an orange
solution. Isolation of the crude material by drying in vacuo
followed by recrystallisation from pentane at ꢀ35 ꢁC gave
the species ½L12ZrðCH2PhÞ2ꢁ as yellow microcrystals in 92%
yield (Scheme 1). The isopropyl-substituted (S)-HL2 was
treated with Zr(CH2Ph)4 at ꢀ78 ꢁC in a similar manner.
The single organometallic product of this reaction
½L22ZrðCH2PhÞ2ꢁ was isolated as an orange solid in 79%
yield after recrystallisation from pentane. Reaction of
two equivalents the phenyl-substituted proligand (S)-HL3
with Zr(CH2Ph)4 was less straightforward. Initial NMR-
scale experiments at ambient temperature appeared to
show the formation of mixtures of [L3Zr(CH2Ph)3] and
½L32ZrðCH2PhÞ2ꢁ, although the latter was the major prod-
uct. The slow addition of a solution of Zr(CH2Ph)4 to
slightly more than two equivalents of HL3 at ꢀ78 ꢁC in tol-
uene was more successful. Recrystallisation from pentane
O
O
O
O
N
R
R
N
CH Ph
2
R
R
H
Zr
Cl
Cl
H
CH2Ph
Zr
N
N
O
O
O
O
t
Bu
t
Bu
tBu
R = tBu, iPr, Ph
tBu
R = tBu, iPr
Scheme 1. Synthesis of bidentate proligands.
for a pyridine/alcoholate N–O system [12]. Another N–O
ligand system, the salicyloxazolines (such as Ln, Scheme
1), has been used to create chiral-at-metal complexes of
later transition metals, particularly ruthenium [13]. Some
group 4 metal complexes have been reported also, but the
issue of diastereoselection in these compounds was not
addressed [14]. We report here a range of complexes of sal-
icyloxazolinates with zirconium, and a combined struc-
tural, computational and NMR spectroscopic study in
respect of the stereogenic element at the metal.
of crude material thus obtained gave ½L32ZrðCH2PhÞ ꢁ as
2
an orange microcrystalline solid in 83% yield.
Treatment of the proligands (S)-HL1–3 with potassium
hydride in THF provided the salts (S)-KL1–3 Æ (THF)n.
These compounds were isolated as white or pale yellow sol-
ids and the value of n was calculated from integration of
1
2. Results and discussion
the relevant H NMR resonances from spectra recorded
in d5-pyridine. The isolation of these compounds is not a
necessary step prior to salt metathesis with ZrCl4(THF)2,
it usually being more convenient to carry out the latter
reaction in situ.
2.1. Proligand synthesis
The preparations of (S)-HL1–3 were accomplished in ca.
70% yield via reaction of 3,5-di-tert-butylsalicylic acid I
with the appropriate amino alcohol after the method of
Vorbruggen and Krolikiewicz [15,16] (Scheme 1). Treat-
ment of the starting materials with triphenylphosphine, car-
bon tetrachloride and a base was found to be most effective
using the quantities of these reagents as detailed in Scheme
1. Optimal conditions for this type of synthesis appear to be
highly substrate dependant [17]. In order to maintain low
concentrations of the chlorophosphonium salt and thus
prevent side reactions, either the CCl4 or PPh3 is added to
the reaction mixture over 4 h; it generally being more con-
venient to administer CCl4 via syringe pump [18]. Our best
Salt metathesis of ZrCl4(THF)2 with prepared (S)-
KL1 Æ (THF)n was straightforward, and resulted in the for-
mation of a yellow product. Sublimation of this material at
260 ꢁC, 10ꢀ6 mmHg, gave a pale yellow microcrystalline
solid, which was shown to be ½L12ZrCl2ꢁ. The salt metathesis
reaction of ZrCl4(THF)2 with (S)-KL2 Æ (THF)n was rather
capricious. Ultimately however, the treatment of an in situ
prepared solution of (S)-KL2 with ZrCl4(THF) at 0 ꢁC did
lead to the isolation of a pale yellow solid. Sublimation of this
material at 265 ꢁC/10ꢀ6 mmHg gave a pale yellow microcrys-
talline solid. Analysis of the sublimated compound con-
firmed the molecular formula as ½L22ZrCl2ꢁ. Although a

2230
I. Westmoreland et al. / Journal of Organometallic Chemistry 691 (2006) 2228–2236
variety of reaction conditions were tested we were unable to
angle for the six-membered chelates of ca. 77ꢁ (Table 1).
The ligand nitrogen donor atoms also take up mutually
cis coordination sites at the metal centre, the N–Zr–N bond
angle being 99.1(3)ꢁ. Each nitrogen atom is trans to a
benzyl ligand. The remaining mutually trans sites are occu-
pied by the aryl oxygens at a O–Zr–O bond angle of
167.1(3). These and other metrical parameters are compa-
rable to related Schiff base [19,20] and oxazoline [14]
complexes.
develop a synthesis of adequately pure ½L32ZrCl2ꢁ.
2.3. Molecular structure of L12Zr CH2Ph
2
Single crystals of ½L21ZrðCH2PhÞ2ꢁ suitable for X-ray
structural determination were obtained by cooling a con-
centrated solution in pentane to ꢀ30 ꢁC. In the solid state,
the K-trans,cis,cis diastereomer is observed exclusively
(Figs. 1 and 2(a)).
The complex is essentially C2-symmetric with a six-coor-
dinate zirconium centre at which the benzyl coligands
occupy mutually cis positions. The geometry at the metal
is distorted from octahedral due to the N–Z–O ligand bite
2.4. Computational investigation of the structure of
L12Zr CH2Ph
2
We set out to investigate the relative energies of the six-
coordinate diastereomers of ½L12ZrðCH2PhÞ2ꢁ via DFT [21];
predicted diastereoselectivities from this approach consis-
Table 1
1
˚
Selected bond lengths (A) and angles (ꢁ) for ½L2ZrðCH2PhÞ2ꢁ
Zr(1)–O(1)
Zr(1)–O(10)
Zr(1)–C(100)
Zr(1)–C(200)
Zr(1)–N(1)
Zr(1)–N(10)
1.981(5)
2.010(5)
2.319(11)
2.319(12)
2.403(7)
2.417(8)
O(1)–Zr(1)–O(10)
O(1)–Zr(1)–C(100)
O(10)–Zr(1)–C(100)
O(1)–Zr(1)–C(200)
O(10)–Zr(1)–C(200)
C(100)–Zr(1)–C(200)
O(1)–Zr(1)–N(1)
O(10)–Zr(1)–N(1)
C(100)–Zr(1)–N(1)
C(200)–Zr(1)–N(1)
O(1)–Zr(1)–N(10)
O(10)–Zr(1)–N(10)
C(100)–Zr(1)–N(10)
C(200)–Zr(1)–N(10)
N(1)–Zr(1)–N(10)
C(13)–O(1)–Zr(1)
C(130)–O(10)–Zr(1)
167.1(3)
94.6(3)
95.1(3)
97.0(4)
91.9(4)
87.3(5)
77.5(2)
94.0(2)
168.8(3)
85.9(4)
95.2(3)
76.4(3)
89.3(4)
167.5(3)
99.1(3)
146.4(6)
143.2(7)
Fig. 1. Molecular structure of ½L12ZrðCH2PhÞ ꢁ viewed along the crystal-
2
lographic C2 axis with benzyl ligands at the rear.
tBu
tBu
tBu
tBu
tBu
tBu
tBu
tBu
O
O
O
Zr
O
O
O
O
O
N
N
N
tBu
N
N
Zr
O
Zr
O
O
tBu
Bn
tBu
Bn
Bn
Bn
Bn
Bn
tBu
N
tBu
tBu
O
tBu
tBu
tBu
a
b
c
Fig. 2. Selected diastereomers of ½L1ZrðCH2PhÞ ꢁ: (a) K-trans,cis,cis observed in solid state rel. energy = 0 kJ molꢀ1, (b) K-trans,cis,cis overlap of groups
2
2
indicated, (c) K-cis,cis,cis rel. energy = 35 kJ molꢀ1
.

I. Westmoreland et al. / Journal of Organometallic Chemistry 691 (2006) 2228–2236
2231
343 K
313 K
283 K
253 K
223 K
(ppm)
4.0
3.8
3.6
3.4
3.2
3.0
2.8
4.2
Fig. 3. Variable temperature 1H NMR spectra of ½L1ZrðCH2PhÞ ꢁ.
2
2
tently agree well with experimental observations in related
systems [10–12].
as the ligands do not become detached, the benzyl methy-
lene groups remain diastereotopic and would be expected
to give rise to a pair of AB doublets in the H NMR spec-
1
Starting from the observed molecular structure of
½L12ZrðCH2PhÞ2ꢁ we attempted to construct models of the
remaining seven possible six-coordinate diastereomers
using the MM2 package of Chem3D and standard bond
lengths. In most cases these structures were found to be
infeasible for steric reasons. For example, compare the
observed K-trans,cis,cis diastereomer (Fig. 2(a)) with its
partner of D helicity (b). In the K isomer, the oxazoline
tert-butyl groups point into sterically unencumbered
regions at the back of the complex as drawn, while for
the inaccessible D isomer, these groups would necessarily
be oriented towards the front, and overlap (rather than
merely interact sterically) with the phenoxy tert-butyl sub-
stituents on neighbouring salicyloxazoline ligand. The only
feasible alternative diastereomer was K-cis-cis-cis (c). This
structure and the crystallographically observed K-trans,
cis,cis isomer were investigated using DFT calculations.
The minimised structure of the K-cis,trans,cis isomer (a)
was essentially superimposable on that observed in the
solid state. The D-cis,cis,cis diastereomer (c) was found to
be 35 kJ molꢀ1 less stable; the most obvious unfavourable
steric interaction, that between an oxazoline tert-butyl
group and benzyl CH2 is indicated in Fig. 2. In any event,
we expect to observe only the K-trans,cis,cis isomer (a) in
the gas phase.
trum. It is not easy to quantify the effect of the presence
of the asymmetry on the magnitude of the apparent chem-
ical shift difference between HA and HB, i.e., how well the
chirality of the system is expressed in this physical property
[22]. Experience dictates however that in a given chiral-at-
metal system with diastereotopic co-ligands, the magnitude
of the splitting (Dd ppm) is affected strongly by the degree of
fluxionality; as temperature is increased and exchange pro-
cesses become more rapid, Dd tends to fall [9,10,12].
Fig. 3 shows portions of the 1H NMR spectra of
½L12ZrðCH2PhÞ2ꢁ recorded in d8-toluene at various tempera-
tures. At 343 K the oxazoline CH2 proton trans to the CH
hydrogen appears as an AB doublet at d 4.23 ppm. The
other CH2 proton appears as a triplet at d 3.40 ppm, and
shows a vicinal coupling of 8 Hz to the CH, to which it
is also cis. The low chemical shift of this signal can be
explained by reference to the crystal structure (Fig. 1);
the hydrogen from which it arises sits directly above the
phenyl group of the Zr-benzyl, the shift to high field being
a result of ring current. The second AB doublet at 3.92 is
attributed to the CH itself, which also shows a coupling
of 8 Hz to the two vicinal protons, but has a dihedral angle
of about 90ꢁ with that to which it is trans resulting in the
observed fine splitting. The zirconium-bound benzyl CH2
protons, despite being chemically inequivalent (vide supra),
give rise to a singlet resonance observed at d 2.73 ppm.2 A
2.5. NMR spectroscopic investigations
The complex ½L12ZrðCH2PhÞ2ꢁ contains three stereogenic
centres, two at C and one at Zr. No matter what the mech-
anism of epimerisation at the stereogenic Zr atom, so long
2
In dichloromethane-d2 at ambient temperature the Zr–CH2 groups
give rise to a pair of AB doublets with apparent Dd of only 0.05 ppm.

2232
I. Westmoreland et al. / Journal of Organometallic Chemistry 691 (2006) 2228–2236
very small amount of a minor species, most likely an impu-
rity, though possibly also a different isomer, is observed at
d 2.97 and 4.08 ppm. Cooling of the sample leads to
changes in the chemical shift for all of these resonances.
By 253 K the chemical shifts of the benzyl CH2 protons
are sufficiently different that the expected second order pair
of AB doublets is observed (ca. d 2.93 ppm). Conversely,
the oxazoline CH2 proton resonates at lower frequency as
the temperature is decreased, although by 253 K there is lit-
tle discernable change in the shape of signal. By 223 K the
resonances for these protons overlap, but Dd for the benzyl
CH2 group has risen to ca. 0.06 ppm. Much below this tem-
perature, increased viscosity of the solvent causes extreme
broadening. The complexes ½Ln2ZrðCH2PhÞ2ꢁ (n = 2,3),
and indeed ½Ln2ZrCl2ꢁ (n = 1,2) behave in a very similar
manner.
between the six-coordinate K-trans,cis,cis isomer and unob-
served five-coordinate species via N-dissociation. The
appearance of the spectra as temperature is increased is
consistent with this; the reduction in Dd for the Zr–CH2
signals probably arises as a result of the increased average
distance between these units and the stereogenic centres on
the oxazoline ligands, and simultaneously poorer expres-
sion of chirality at metal. This behaviour contrasts with
that displayed by [(NO*)2Zr(CH2Ph)2] (Fig. 4) using the
same solvent and magnetic field strength [12]. This system
is clearly in the slow exchange regime below 273 K showing
only one diastereomer (Dd for benzyl CHABB doublets ca.
0.3 ppm). On increasing the temperature, the signals for
HA and HB broaden but do not coalesce, then finally re-
sharpen at new chemical shifts in the fast exchange regime
above 353 K. Notably N-dissociation in this system does
not increase the distance of the stereogenic C centres from
the inner coordination sphere. A similar comparison is pos-
sible between the present system and our recently reported
chiral aminopyridinates [10] where the portion of the
ligand remaining bonded to the metal contains the chiral
information.
The DFT calculations detailed above predict the exis-
tence of a single diastereomer for this compound. It is evi-
dent however that the system is undergoing structural
change down to 223 K and perhaps lower on the NMR
chemical shift timescale at that temperature, presumably
A space-filling model of the computed structure of
K-trans; cis; cis-½L12ZrðCH2PhÞ2ꢁ is shown in Fig. 5(a), with
approximately the same orientation as that shown in
Fig. 2(a) (i.e., viewed along the C2-axis with the two
CH2Ph groups toward the front). The analogous structure
of D-trans,cis,cis-[(NO*)2Zr(CH2Ph)2] is shown for com-
parison [12]. For the L1 complex, the benzyl H atoms
are in rather similar environments. For the NO* complex
the chirality of the system is expressed very strongly, lead-
ing to an S-shaped conformation of the Zr(CH2Ph)2 units
Me
iPr
O
Me
[ZrBn4]
N
N
N
Pri
Zr
O
iPr
Bn
Bn
OH
H(NO*)
which places the
environments.
H
atoms in distinctly different
Me
Δ isomer
3. Conclusion
The salicyloxazoline proligands (S)-HL1–3 give C2-sym-
metric compounds of the form [L1–32ZrX2]. Examination
Fig. 4. A pyrdininealcoholate complex of zirconium, [(NO*)2Zr(CH2Ph)2]
[12].
Fig. 5. Space-filling models of (a) K-trans; cis; cis-½L1ZrðCH2PhÞ ꢁ and (b) D-trans,cis,cis-[(NO*)2Zr(CH2Ph)2]; benzyl C atoms marked with
.
*
2
2

I. Westmoreland et al. / Journal of Organometallic Chemistry 691 (2006) 2228–2236
2233
of the solid state structure of ½L12ZrðCH2PhÞ2ꢁ, and DFT
4.2. Proligand synthesis
calculations on this and the only other feasible diastereo-
mer point toward the existence of only one six-coordinate
species at equilibrium. Nevertheless the degree of expres-
sion of the chirality of the structure in what would be the
site of metal-based reactions is poor in comparison to other
Zr systems, as assessed from the molecular structure and
The following represents our refined general procedure
for the synthesis of proligands HL1–3: Acetonitrile (50 ml)
was added to a Schlenk vessel charged with 3,5-di-tert-
butylsalicylic acid (2.0 g, 8.0 mmol), appropriate amino
alcohol (8.0 mmol) and triphenylphosphine (6.30 g,
24.0 mmol). To the resulting white suspension was added
triethylamine (4.5 ml, 32.0 mmol) with stirring, and a col-
ourless solution was obtained after about 10 min. Carbon
tetrachloride (7.8 ml, 80 mmol) was added dropwise to
the reaction mixture over 4 h. During the course of the
addition the solution became cloudy, and changed colour;
first to yellow, then orange, and ultimately dark red, or
even dark brown in some instances. The mixture was stir-
red for a further 48 h following completion of the addition.
The solution was then filtered, and diethyl ether (30 ml)
was added to the red filtrate. Shaking of this mixture
resulted in the precipitation of a white solid which was also
removed by filtration. The filtrate was concentrated in
vacuo a leave a dark red residue. This material was
extracted with hexane (100 ml) over 4 h. The extract was
filtered, and concentrated under reduced pressure to yield
an orange solid. This material was eluted through a column
of silica gel using a hexane:ethyl acetate (20:1) mobile
phase. Product containing fractions were identified by thin
layer chromatography and collected. Removal of the sol-
vent under reduced pressure gave the product as a white
microcrystalline solid.
1
variable temperature H NMR data.
4. Experimental
4.1. General details
Where necessary, procedures were carried out under
an inert atmosphere of argon by using a dual manifold
vacuum/argon line and standard Schlenk techniques, or
in an MBraun dry box (<1 ppm O2/H2O). Solvents were
dried by refluxing for three days under dinitrogen over
the appropriate drying agents (sodium for toluene; potas-
sium for THF, benzene and heptane; sodium–potassium
alloy for diethyl ether, petroleum ether, and pentane; cal-
cium hydride for carbon tetrachloride, dichloromethane,
pyridine and acetonitrile) and were degassed before use.
Solvents were stored in glass ampoules under argon.
All glassware and cannulae and Celite were stored in
an oven at >373 K. Deuterated solvents were freeze–thaw
degassed and dried by refluxing over potassium (for ben-
zene, toluene and THF) or calcium hydride (for dichloro-
methane, pyridine and acetonitrile) for three days before
trap-to-trap distillation and storage in the dry box. Deu-
terated chloroform was dried in the bottle over molecular
4.2.1. (S)-HL1
˚
sieves (4 A). Hydrated 3,5-di-tert-butylsalicylic acid was
From (S)- tert-leucinol (936 mg, 8.0 mmol). Yield
dried thoroughly and recrystalised prior to use. Triethyl-
amine was dried by distillation from calcium hydride.
Sodium hydride dispersion in mineral oil was placed in
a Schlenk vessel under an inert atmosphere and washed
three times with diethyl ether to remove the oil. The solid
was subsequently dried thoroughly and stored in the dry
box. All other purchased chemical reagents were used as
received.
Flash chromatography was performed either in stan-
dard glassware or with a FlashMaster Personal chromatog-
raphy system and a selection of pre-packed disposable
columns. Thin-layer chromatography was performed using
Merck 0.25 mm silica layer foil-backed plates.
1.91 g, 72%. 1H NMR dH (400 MHz, 298 K, CDCl3)
4
12.64 (1H, bs, OH), 7.44 (1H, d, JHH = 2 Hz, Ar–H),
4
2
7.36 (1H, d, JHH = 2 Hz, Ar–H), 4.24 (1H, dd, JHH
=
3
3
9 Hz, JHH = 2 Hz, CH2), 4.11 (1H, t, JHH = 8 Hz,
2
3
CH2), 4.03 (1H, dd, JHH = 9 Hz, JHH = 2 Hz, CH),
t
t
t
1.37 (9H, s, Bu) 1.23 (9H, s, Bu) 0.88 (9H, s, Bu) ppm.
13C{1H} NMR dC (100.6 MHz, 298 K, CDCl3) 166.0,
157.1, 139.9, 136.4, 127.9, 122.1, (Ar), 109.8 (C@N), 75.1
(CH), 67.8 (CH2), 35.2, 34.3, 33.8 (CMe3), 31.5, 29.5,
25.9 (C Me3) ppm.
MS (EI+) m/z: 331 (100%, M+), 316 (85%, M+ꢀCH3),
t
274 (60%, M+ꢀ Bu).
IR (Golden gate) m cmꢀ1: 2961, 2906, 2867, 1633, 1596,
1437, 1392, 1368, 1362, 1287, 1254, 1219, 1204, 1189, 1106,
1028, 979, 892, 822, 784, 70, 727.
NMR spectra were recorded on Bruker ACF-250, DPX-
1
300, DPX-400, AV-400 and DPX-500 spectrometers. H
and 13C spectra were referenced internally using residual
protio solvent resonances relative to tetramethylsilane
(d = 0 ppm). Proton and carbon NMR assignments were
routinely confirmed by 1H–1H (COSY) or 1H–13C
(HMQC) experiments. Infra-red spectra were obtained
either as Nujol mulls using a Perkin–Elmer Paragon 1000
FTIR spectrometer, or directly using an Avatar 320 FTIR
instrument. EI and CI mass spectra were obtained on a VG
Autospec mass spectrometer. Elemental analyses were per-
formed by Warwick Analytical Services.
EA Calc. for C21H33NO2: C, 76.09; H, 10.03; N, 4.23.
Found: C, 76.29; H, 10.11; N, 4.17%.
4.2.2. (S)-HL2
From (S)-(+)-2-amino-3-methyl-1-butanol (825 mg,
8.0 mmol). Yield 1.7 g, 67%. 1H NMR dH (400 MHz,
298 K, CDCl3) 12.63 (1H, bs, OH), 7.45 (1H, d,
4JHH = 2 Hz, Ar–H), 7.36 (1H, d, JHH = 2 Hz, Ar–H),
4
4.32 (1H, m, CH2), 4.06 (2H, m, CH2 and CH), 1.78 (1H,

2234
I. Westmoreland et al. / Journal of Organometallic Chemistry 691 (2006) 2228–2236
m, CHMe2), 1.37 (9H, s, tBu) 1.28 (9H, s, tBu), 0.96 (3H, s,
Me), 0.88 (3H, s, Me) ppm.
4JHH = 3 Hz, Ar–H), 4.15 (1H, m, CH2), 3.95 (2H, m,
CH and CH2), 3.60 (m, THF), 1.85 (1H, m, CHMe2),
1.74 (9H, s, Bu), 1.60 (m, THF), 1.43 (9H, s, Bu), 0.83
(3H, s, CH3), 0.73 (3H, s, CH3).
t
t
13C{1H} NMR dC (100.6 MHz, 298 K, CDCl3) 166.0,
157.0, 139.9, 136.4, 127.9, 122.1 (Ar), 109.8, (C@N), 71.7
(CH), 69.6 (CH2), 35.2, 34.3, (CMe3), 33.1, (CHMe2),
31.5, 29.5 (C Me3), 18.9, 18.7 (CHMe2) ppm.
4.3.3. (S)-KL3 Æ (THF)0.80
MS (EI+) m/z: 317 (100%, M+), 302 (85%, M+ꢀCH3),
Yield 210 mg, 92%. 1H NMR (400 MHz, 298 K, d5-pyr-
t
4
260 (45%, M+ꢀ Bu).
idine): dH 8.14 (1H, d, JHH = 3 Hz, Ar–H), 7.66 (1H, d,
IR (Golden gate) m cmꢀ1: 2959, 2870, 1633, 1595, 1468,
1437, 1391, 1362, 1288, 1274, 1253, 1215, 1188, 1103, 1044,
1030, 980, 822, 783, 772, 765, 728.
4JHH = 3 Hz, Ar–H), 7.32 (2H, m, Ar–H), 7.17 (3H, m,
Ar–H), 5.35 (1H, m, CH2), 4.33 (1H, m, CH), 3.91 (1H,
t
m, CH2), 3.60 (m, THF), 1.79 (9H, s, Bu), 1.60 (m,
t
EA Calc. for C20H31NO2: C, 75.67; H, 9.84; N, 4.41.
Found: C, 75.66; H, 9.87; N, 4.18%.
THF), 1.49 (9H, s, Bu).
4.4. Zirconium benzyls
4.2.3. (S)-HL3
4.4.1. L12Zr CH2Ph
From (S)-(+)-2-phenylglycinol (1.10 g, 8.0 mmol). Yield
2
1
2.1 g, 75%. H NMR dH (400 MHz, 298 K, CDCl3) 12.48
A solution of Zr(CH2Ph)4 (153 mg, 0.336 mmol) in tol-
uene (10 ml) was cooled to ꢀ78 ꢁC with stirring. To this
was added a solution of (S)-HL1 (222 mg, 0.670 mmol) in
toluene (10 ml). The resulting yellow solution was stirred
at this temperature for 1 h in the absence of light. The cold
bath was then removed and the reaction mixture was
allowed to warm to ambient temperature. Stirring was con-
tinued for a further 15 h. After this time the solvent was
removed in vacuo, and an orange residue was obtained.
The residue was re-dissolved in pentane and filtered. The
filtrate was reduced in volume to ca. 5 ml, and cooled to
ꢀ30 ꢁC. After cooling for 12 h at this temperature, the
product was obtained as fine orange crystals. Yield
331 mg, 92%.
4
(1H, bs, OH), 7.56 (1H, d, JHH = 2 Hz, Ar–H), 7.42
4
(1H, d, JHH = 2 Hz, Ar–H), 7.31 (2H, m, Ar–H), 7.26
2
3
(3H, m, Ar–H), 5.40 (1H, dd, JHH = 8 Hz, JHH = 2 Hz,
2
3
CH2), 4.72 (1H, dd, JHH = 8 Hz, JHH = 2 Hz, CH),
4.18 (1H, t, JHH = 8 Hz, CH2), 1.40 (9H, s, Bu) 1.28
3
t
t
(9H, s, Bu) ppm.
13C{1H} NMR dC (100.6 MHz, 298 K, CDCl3) 167.2,
157.1, 141.8, 140.1, 136.6, 128.8, 128.3, 127.8, 126.6,
122.3 (Ar), 109.6 (C@N), 73.8 (CH2), 69.0 (CH), 35.2,
34.3 (CMe3), 31.5, 29.5 (C Me3) ppm.
MS (EI+) m/z: 351 (100%, M+), 336 (85%, M+ꢀCH3),
t
294 (30%, M+ꢀ Bu).
IR (Golden gate) m cmꢀ1: 2958, 1632, 1439, 1390, 1367,
1344, 1276, 1252, 1187, 1141, 1103, 980, 894, 820, 786, 754,
717.
1H NMR dH (400 MHz, 298 K, CD2Cl2) 7.54 (2H, d,
4JHH = 2 Hz, Ar–H), 7.49 (2H, d, JHH = 2 Hz, Ar–H),
4
6.66 (4H, t, 3JHH = 7 Hz, Ar–H), 6.51 (4H, d, 3JHH = 7 Hz,
EA Calc. for C23H29NO2: C, 78.59; H, 8.32; N, 3.99.
Found: C, 78.50; H, 8.46; N, 3.95%.
3
Ar–H), 6.40 (2H, d, JHH = 7 Hz, Ar–H), 4.09 (2H, dd,
3
2JHH = 8 Hz, JHH = 2 Hz, oxazoline CH2), 4.00 (2H, dd,
3
4.3. Potassium salts of proligands
2JHH = 8 Hz, JHH = 2 Hz, oxazoline CH), 3.28 (2H, t,
3JHH = 8 Hz, oxazoline CH2), 2.25 (2H, d, JHH = 10 Hz,
2
2
THF (15 ml) was added to a Schlenk vessel charged with
proligand (0.55 mmol) and potassium hydride (44 mg,
1.10 mmol). The resulting yellow suspension was stirred
for 15 h under reduced pressure. After this time the stirring
was ceased and unreacted KH was allowed to settle. The
solution was filtered via cannula, and the yellow filtrate
was concentrated in vacuo to yield a yellow microcrystalline
benzyl CH2), 2.20 (2H, d, JHH = 10 Hz, benzyl CH2),
1.61 (18H, s, tBu) 1.26 (18H, s, tBu), 0.50 (18H, s,
tBu) ppm.
13C{1H} NMR dC (100.6 MHz, 298 K, CD2Cl2) 172.8,
160.7, 151.8, 141.1, 138.0, 130.2, 127.8, 125.3, 124.5,
119.3 (Ar), 116.7 (C@N), 73.5, (CH), 69.3 (oxazoline
CH2), 67.9 (benzyl CH2), 35.8, 35.4, 34.7 (CMe3), 31.6,
30.7, 25.5 (C Me3) ppm.
1
solid. Analysis by H NMR spectroscopy in d5-pyridine
enabled the composition of the product to be established.
MS (CI+) m/z: 934 (20%, M+), 843 (100%,
M+ꢀCH2Ph).
4.3.1. (S)-KL1 Æ (THF)0.75
IR (Nujol) m cmꢀ1: 1594, 1556, 1461, 1377, 1301, 1259,
1222, 1206.
Yield 241 mg, 93%. 1H NMR (400 MHz, 298 K, d5-pyr-
4
idine): dH 8.04 (1H, d, JHH = 3 Hz, Ar–H), 7.61 (1H, d,
EA Calc. for C56H78N2O4Zr: C, 71.98; H, 8.41; N, 3.00.
Found: C, 71.68; H, 8.36; N, 2.95%.
4JHH = 3 Hz, Ar–H), 4.07 (3H, m, CH and CH2), 3.67
t
(m, THF), 1.74 (9H, s, Bu), 1.63 (m, THF), 1.45 (9H, s,
t
tBu), 0.88 (9H, s, Bu).
4.4.2. L22Zr CH2Ph
2
Yield 205 mg, 79%. 1H NMR dH (400 MHz, 298 K,
4.3.2. (S)-KL2 Æ (THF)1.25
C6D6) 7.81 (2H, d, JHH = 2 Hz, Ar–H), 7.69 (2H, d,
4
Yield 260 mg, 93%. 1H NMR (400 MHz, 298 K, d5-pyr-
4JHH = 2 Hz, Ar–H), 6.97 (4H, d, JHH = 7 Hz, Ar–H),
3
4
idine): dH 8.07 (1H, d, JHH = 3 Hz, Ar–H), 7.60 (1H, d,
6.88 (4H, t, 3JHH = 8 Hz, Ar–H), 6.60 (2H, t, 3JHH = 7 Hz,

I. Westmoreland et al. / Journal of Organometallic Chemistry 691 (2006) 2228–2236
2235
Ar–H), 4.03 (2H, dd, 2JHH = 8 Hz, 3JHH = 3 Hz, oxazoline
was sublimated at 260 ꢁC, 10ꢀ6 mmHg, to yield a pale yel-
low solid. Yield 250 mg, 56%.
2
3
CH2), 3.60 (2H, dd, JHH = 8 Hz, JHH = 3 Hz, oxazoline
CH), 3.16 (2H, m, oxazoline CH2), 3.03 (2H, d,
1H NMR dH (400 MHz, 298 K, C6D6) 8.09 (2H, d,
2
4
2JHH = 10 Hz, benzyl CH2), 2.73 (2H, d, JHH = 10 Hz,
4JHH = 2 Hz, Ar–H), 7.69 (2H, d, JHH = 2 Hz, Ar–H),
t
2
3
benzyl CH2), 1.78 (18H, s, Bu) 1.35 (2H, m, CHMe2),
5.54 (2H, dd, JHH = 8 Hz, JHH = 2 Hz, CH2), 4.33 (2H,
t
3
2
3
1.17 (18H, s, Bu), 0.18 (6H, d, JHH = 8 Hz, CH3), 0.15
dd, JHH = 8 Hz, JHH = 2 Hz, CH), 4.25 (2H, t,
3
t
t
(6H, d, JHH = 8 Hz, CH3) ppm.
3JHH = 8 Hz, CH2), 1.50 (18H, s, Bu) 1.34 (18H, s, Bu),
13C{1H} NMR dC (100.6 MHz, 298 K, C6D6) 169.5,
161.2, 150.9, 141.1, 138.7, 130.0, 128.1, 125.9, 124.2,
119.7 (Ar), 115.5 (C@N), 69.8, (CH), 68.3 (benzyl CH2),
67.1 (oxazoline CH2), 36.0, 34.5 (CMe3), 31.5, 30.6 (C
Me3), 30.0 (CHMe2), 19.2, 14.8 (CH3) ppm.
1.12 (18H, s, Bu) ppm.
t
13C{1H} NMR dC (100.6 MHz, 298 K, C6D6) 173.3,
161.0, 141.4, 138.0, 131.1, 123.8 (Ar), 114.3 (C@N), 79.8
(CH), 70.2 (CH2), 36.2, 35.1, 34.5 (CMe3), 31.5, 30.0,
26.4 (C Me3) ppm.
MS (CI+) m/z: 905 (20% MH+), 813 (50%,
M+ꢀCH2Ph), 723 (30%, M+ꢀ2 · CH2Ph).
MS (CI+) m/z: 822 (100%, M+), 785 (70%, M+ꢀCl), 765
(15%, M+ꢀ Bu).
t
IR (Nujol) m cmꢀ1: 1613, 1580, 1462, 1377, 1260, 1206,
1103, 1018.
IR (Nujol) m cmꢀ1: 1606, 1556, 1544, 1461, 1378, 1310,
1281, 1259, 1228, 1200.
EA Calc. for C54H74N2O4Zr: C, 71.56; H, 8.23; N, 3.09.
Found: C, 71.36; H, 8.08; N, 3.12%.
EA Calc. for C42H64Cl2N2O4Zr: C, 61.29; H, 7.84; N,
3.40. Found: C, 61.21; H, 7.87; N, 3.25%.
4.4.3. L32Zr CH2Ph
4.5.2. L22ZrCl2
2
Yield 190 mg, 83%. 1H NMR dH (400 MHz, 298 K,
Yield 186 mg, 63%. 1H NMR dH (400 MHz, 298 K,
4
C6D6) 7.87 (2H, d, JHH = 3 Hz, Ar–H), 7.73 (2H, d,
4
C6D6) 7.91 (2H, d, JHH = 2 Hz, Ar–H), 7.73 (2H, d,
4JHH = 3 Hz, Ar–H), 7.04 (6H, m, Ar–H), 6.93 (3H, m,
Ar–H), 6.80 (5H, m, Ar–H), 6.70 (6H, m, Ar–H), 4.95
4JHH = 2 Hz, Ar–H), 3.95 (2H, dd, 2JHH = 9 Hz,
3JHH = 3 Hz, CH2), 3.79 (2H, dd, 2JHH = 9 Hz,
2
3
(2H, dd, JHH = 9 Hz, JHH = 4 Hz, oxazoline CH2), 3.60
3JHH = 3 Hz, CH) 3.25 (2H, t, JHH = 8 Hz, oxazoline
3
2
3
(2H, dd, JHH = 9 Hz, JHH = 4 Hz, oxazoline CH), 3.46
CH2), 1.89 (18H, s, tBu) 1.73 (2H, m, CHMe2) 1.48
3
4
(2H, dd, JHH = 8 Hz, JHH = 1 Hz, oxazoline CH2), 3.33
t
3
(18H, s, Bu), 0.85 (6H, d, JHH = 7 Hz, CH3), 0.80 (6H,
2
(2H, d, JHH = 10 Hz, benzyl CH2), 2.90 (2H, d,
3
d, JHH = 7 Hz, CH3) ppm.
2JHH = 10 Hz, benzyl CH2), 2.05 (18H, s, Bu) 1.51 (18H,
t
13C{1H} NMR dC (100.6 MHz, 298 K, C6D6) 173.8,
161.4, 142.0, 140.5, 131.5, 123.5 (Ar), 113.7 (oxazoline
C@N), 72.5 (CH), 69.9 (CHMe2), 67.0 (CH2), 35.9, 34.6
(CMe3), 31.4, 30.1 (C Me3), 19.3, 18.8 (CH3) ppm.
MS (CI+) m/z: 794 (100%, M+), 759 (70%, M+ꢀCl).
IR (Nujol) m cmꢀ1: 1602, 1458, 1377, 1260, 1024, 798,
721.
t
s, Bu) ppm.
13C{1H} NMR dC (100.6 MHz, 298 K, C6D6) 169.6,
161.4, 150.7, 140.5, 140.2, 137.9, 129.6, 129.3, 128.5,
125.7, 124.7, 119.7 (Ar), 115.9 (C@N), 74.8, (oxazoline
CH2), 68.5 (CH), 67.8 (benzyl CH2), 36.0, 34.5 (CMe3),
31.8, 30.7 (C Me3) ppm.
MS (CI+) m/z: 897 (35%, M+-Ph), 883 (100%,
M+ꢀCH2Ph).
EA Calc. for C40H60Cl2N2O4Zr: C, 60.43; H, 7.61; N,
3.52. Found: C, 60.19; H, 7.62; N, 3.41%.
IR (Nujol) m cmꢀ1: 1610, 1574, 1456, 1377, 1299, 1261,
1222, 1204, 1112, 1026.
EA Calc. for C60H70N2O4Zr: C, 73.96; H, 7.24; N, 2.87.
Found: C, 73.95; H, 7.16; N, 2.83%.
4.6. Crystallography
Crystals of K-trans; cis; cis-½L12ZrðCH2PhÞ2ꢁ were coated
in an inert oil prior to transfer to a cold nitrogen gas stream
on Bruker-AXS SMART three circle area detector diffrac-
tometer system equipped with Mo Ka radiation
4.5. Zirconium chlorides
4.5.1. L12ZrCl2
A Schlenk vessel was charged with (S)-HL1 (362 mg,
1.09 mmol) and NaH (52 mg, 2.18 mmol). THF (20 ml)
was added, and the resulting white suspension was stirred
under reduced pressure for 15 h. After this time the stirring
was ceased and unreacted NaH was allowed to settle. The
solution was passed through a filter cannula into a Schlenk
vessel charged with ZrCl4(THF)2 (206 mg, 0.545 mmol) at
0 ꢁC. After stirring for 45 min, the ice bath was removed
and the reaction vessel was allowed to warm to ambient
temperature. Stirring was continued for 15 h. After this
time the solution was filtered, and the filtrate was evapo-
rated in vacuo to yield a pale yellow solid. This material
(k = 0.71073 A). Data were collected with narrow (0.3ꢁ in
˚
x) frame exposures. Intensities were corrected semi-empiri-
cally for absorption, based on symmetry-equivalent and
repeated reflections (SADABS). Structures were solved by
direct methods (SHELXS [23]) with additional light atoms
found by Fourier methods. All H atoms were constrained
with a riding model; U(H) was set at 1.2 (1.5 for methyl
groups) times Ueq for the parent atom. Programs used were
Bruker AXS SMART (control), SAINT (integration) and
SHELXTL [23] for structure solution, refinement, and molecu-
lar graphics. Yellow block, C56H78N2O4Zr, Monoclinic,
˚
P2(1), a = 10.7948(7), b = 18.3340(13), c = 13.4520(9) A,

2236
I. Westmoreland et al. / Journal of Organometallic Chemistry 691 (2006) 2228–2236
3
˚
[4] A. von Zelewsky, O. Mamula, J. Chem. Soc., Dalton Trans. (2000)
219.
[5] H. Brunner, Angew. Chem., Int. Ed. Engl. 38 (1999) 1195.
b = 92.1260(10)ꢁ, V = 2660.5(3) A , Z = 2, l = 0.250
mmꢀ1, total reflections = 14069, independent reflections =
7711 [Rint = 0.0.0817], R1, wR2[I > 2(I)] 0.0864, 0.2002.
[6] K. Muniz, C. Bolm, Chem. Eur. J. 6 (2000) 2309.
˜
[7] (a) H. Brunner, Eur. J. Inorg. Chem. (2001) 905;
(b) C. Ganter, Chem. Soc. Rev. 32 (2003) 130;
(c) J.W. Faller, A.R. Lavoie, New. J. Chem. 27 (2003) 899.
[8] P.D. Knight, P. Scott, Coord. Chem. Rev. 243 (2003) 125.
[9] E.J. Crust, I.J. Munslow, C. Morton, P. Scott, J. Chem. Soc., Dalton
Trans. (2004) 2257.
[10] E.J. Crust, A.J. Clarke, R.J. Deeth, C. Morton, P. Scott, J. Chem.
Soc., Dalton Trans. (2004) 4050.
[11] I.J. Munslow, A.R. Wade, R.J. Deeth, P. Scott, Chem. Commun.
(2004) 2596.
[12] I.J. Munslow, A.J. Clarke, R.J. Deeth, I. Westmoreland, P. Scott,
Chem. Commun. (2002) 1868.
[13] A.J. Davenport, D.L. Davies, J. Fawcett, D.R. Russell, Dalton
Trans. (2004) 1481.
[14] P.G. Cozzi, C. Floriani, A. Chiesi-Villa, C. Rizzoli, Inorg. Chem. 34
(1995) 2921;
P.G. Cozzi, E. Gallo, C. Floriani, A. Chiesi-Villa, C. Rizzoli,
Organometallics 14 (1995) 4994;
R.K.J. Bott, M. Hammond, P.N. Horton, S.J. Lancaster, M.
Bochmann, P. Scott, Dalton Trans. 22 (2005) 3611.
[15] M.J. Miller, P.G. Mattingley, M.A. Morrison, J.F. Korwin, J. Am.
Chem. Soc. 102 (1980) 7026.
[16] (a) H. Vorbruggen, K. Krolikiewicz, Tetrahedron Lett. 22 (1981)
4471;
(b) H. Vorbruggen, K. Krolikiewicz, Tetrahedron 49 (1993)
9353.
[17] A. Scheurer, P. Mosset, W. Bauer, R.W. Saalfrank, Eur. J. Org.
Chem. 0 (2001) 3067.
4.7. DFT calculations
All DFT calculations employed the Amsterdam density
functional (ADF) program version 2005.01. Geometries
were optimised in the gas phase at the local density approx-
imation level using a triple-f + p STO basis set on Zr (ADF
basis TZP), double-f + polarisation bases on the ligand
donor atoms (ADF basis DZP) and double-f bases on all
remaining atoms (ADF basis DZ). Single-point energies
were then computed with the PW91 gradient corrected
functional. A 3d frozen core was used for Zr and 1s frozen
cores for O, N and C. Default convergence criteria were
employed throughout.
5. Supplementary material
Crystallographic data for the structural analysis of
½L12ZrðCH2PhÞ2ꢁ have been deposited with the Cambridge
Crystallographic Data Centre, CCDC No. 286257. Copies
of this information may be obtained free of charge from
The Director, CCDC, 12 Union Road, Cambridge CB2
1EZ, UK (fax: +44 1223 336 033; e-mail: deposit@ccdc.
cam.ac.uk or www: http://www.ccdc.cam.ac.uk).
[18] S. Rajaram, M.S. Sigman, Org. Lett. 4 (2002) 3399.
[19] P.R. Woodman, P.B. Hitchcock, P. Scott, Chem. Commun. (1996)
2735.
Acknowledgements
[20] P.D. Knight, P.N. OꢀShaughnessy, I.J. Munslow, B.S. Kimberley, P.
Scott, J. Organomet. Chem. 683 (2003) 103.
[21] G. Milano, L. Cavallo, G. Guerra, J. Am. Chem. Soc. 124 (2002)
13368.
P.S. thanks the University of Warwick for a Postgradu-
ate Research Fellowship (to I.W.) and EPSRC for support.
[22] For a discussion of this concept, see M. Asakawa, G. Brancato,
M. Fanti, D.A. Leigh, T. Shimizu, A.M.Z. Slawin, J.K.Y. Wong,
F. Zerbetto, S.W. Zhang, J. Amer. Chem. Soc. 124 (2002)
2939.
References
[1] A.P. Smirnoff, Helv. Chim. Acta 3 (1920) 177.
[2] A. Werner, A.Z. Vilmos, Allorg. Allg. Chem. 21 (1989) 145.
[3] U. Knof, A. von Zelewsky, Angew. Chem., Int. Ed. Engl. 38 (1999) 302.
[23] G.M. Sheldrick, SHELXTL Ver 5.1, Bruker Analytical X-ray Systems,
1997.