Exceptionally high lactide polymerization activity of zirconium complexes with bridged diketiminate ligands

Article abstract of DOI:10.1039/c2dt31761c

A cyclohexanediyl-bridged, bis(N-xylyl) diketiminate ligand, (±)-C₆H₁₀(nacnac^XylH)₂, LH₂ (Xyl = 2,6-dimethylphenyl), was obtained from the reaction of [(2,6-dimethylphenyl)amino]-pent-3-en-2-one first

Full text of DOI:10.1039/c2dt31761c

View Online / Journal Homepage
Dynamic Article Links
Dalton
Transactions
Cite this: DOI: 10.1039/c2dt31761c
www.rsc.org/dalton
PAPER
Exceptionally high lactide polymerization activity of zirconium complexes
with bridged diketiminate ligands†
Ibrahim El-Zoghbi, Todd J. J. Whitehorne and Frank Schaper*
Received 1st August 2012, Accepted 14th September 2012
DOI: 10.1039/c2dt31761c
A cyclohexanediyl-bridged, bis(N-xylyl) diketiminate ligand, ( )-C6H10(nacnacXylH)2, LH2 (Xyl = 2,6-
dimethylphenyl), was obtained from the reaction of [(2,6-dimethylphenyl)amino]-pent-3-en-2-one first
with Meerwein’s salt, then with ( )-cyclohexanediamine. The reaction of the ligand with Zr(NMe2)4
yielded LZr(NMe2)2. Protonation of the remaining diamide ligands with EtOH or [H2NMe2]Cl yielded
LZr(OEt)2 and LZrCl2, respectively. The latter complex was also obtained by the reaction of LH2 first
with nBuLi and then with ZrCl4(THF)2. The dichloride complex yielded LZr(OEt)2 and LZrMe2 upon
reaction with NaOEt or MeLi/AlMe3, respectively. X-ray diffraction studies showed a trans-configuration
of the ancillary ligands in LZrCl2 and LZrMe2, and a cis-configuration in LZr(NMe2)2 and LZr(OEt)2.
LZr(OEt)2 was tested as a catalyst for the polymerization of rac-lactide. Kinetic investigations yielded a
rate law first order in catalyst and monomer and a rate constant k = 14(1) L mol−1 s−1, the latter being
orders of magnitude higher than typical activities for group 4 complexes in lactide polymerization.
Analyses of the obtained polymer revealed an atactic polymer and broad polymer molecular weight
distributions with sizeable fractions of cyclic oligomers. The influence of contaminants on the
polymerization activity was examined: while lactic acid deactivates the catalyst, addition of up to 1 equiv.
of water or para-toluenesulfonic acid revitalized catalysts not showing maximum activity.
of the chiral metal complex renders the coordination sites homo-
topic and allows stereoselectivity of the catalysed reaction. (iv)
Introduction
The increased popularity of β-diketiminate (“nacnac”) ligands
Last but not least, the N-substituents in an idealized cis-octa-
hedral structure are ideally placed to exert control on the reaction
(Scheme 1). Unfortunately, ortho-substituted N-aryl diketimines
are sterically too demanding to form bisdiketiminate complexes
and nacnacAr2ZrX2 complexes were limited to ligands where at
least one N-substituent lacks ortho-substituents.3–5,21 We recently
reported that diketimines with N-alkyl substituents form the
desired C2-symmetric bisdiketiminate complexes, nacnacR2ZrX2,
even with sterically demanding N-alkyl substituents such as
cyclohexyl or α-methylbenzyl.23 Application of these com-
pounds in catalytic reactions has been, however, prevented by
the high steric crowding, which shielded the active sites from
further reactions. Thus, nacnacR2ZrCl2 (R = benzyl, cyclohexyl,
α-methylbenzyl) failed to react either with a variety of alkylating
reagents such as MeLi, nBuLi, AlMe3, MeMgX, or ZnEt2, or
with sodium ethoxide. The corresponding dimethyl complex
nacnacBn2ZrMe2 reacted with B(C6F5)3, but the obtained cat-
ionic zirconium methyl complex reacted neither with diphenyl-
acetylene nor benzaldehyde, most likely due to steric blocking of
the metal center.23 In addition to the low reactivity at the “active
coordination sites”, the configuration of the complexes was not
stable. As observed for N-aryl complexes,3–5 Bailar-Twist iso-
merisation led to fast interconversion between Δ- and Λ-enantio-
mers at room temperature on the NMR time scale (Scheme 1),
which could not be prevented even by the use of a chiral
led to several investigations of their coordination chemistry with
zirconium in the last two decades. The obtained complexes can
1–10
be summarized in three types: nacnacZrX3L0–1
tetrahedral and
,
pseudo-
CpR(nacnac)ZrX2,3,5,10–19
octahedral
nacnac2ZrX2.3–5,20–23 The coordination mode of the diketiminate
ligand in all of these complexes can vary between simple N,N′-
κ2-coordination and higher coordination modes such as “η5-like”
or a κ2,η2-coordination.19,24 Octahedral zirconium bisdiketimi-
nate complexes, nacnac2ZrX2, or their respective cations,
[nacnac2ZrX]+, are very attractive targets as catalysts for coordi-
nation–insertion polymerization or related reactions: (i) They
show the least variation in diketiminate binding modes (in nearly
all complexes a simple κ2-coordination is observed). (ii) They
contain two reactive sites, which are forced in a cis-position with
each other, since the N-substituents disfavour an assembly of all
nitrogen atoms in the same plane.23 (iii) The fixed C2-symmetry
Département de chimie, Université de Montréal, Montréal, Québec H3C
3J7, Canada. E-mail: Frank.Schaper@umontreal.ca
†Electronic supplementary information (ESI) available: Tables S1–S6,
investigations regarding variations in catalyst activity, variable tempera-
ture NMR data, details of polymerization experiments, Crystallographic
Information Files (CIF). CCDC 894423–894426. For ESI and crystal-
lographic data in CIF or other electronic format see DOI:
10.1039/c2dt31761c
This journal is © The Royal Society of Chemistry 2012
Dalton Trans.

View Online
Scheme 1
N-substituent.23 To address both problems, we decided to
connect the two diketiminate ligands by a cyclohexanediyl
bridge (Scheme 1). The smaller bite angle enforced by the C2-
bridge should provide easier access to the reactive sites, while its
axial chirality should prevent the Δ–Λ-isomerisation. Similar bis-
diketiminate ligands with achiral (CH2)n-bridges (n = 2, 3) were
previously employed by Gong et al., but the respective zirco-
nium complexes still showed fast Δ–Λ-isomerisation in sol-
ution.22 Herein we report the preparation and structural analysis
of C6H10(nacnacXyl)2ZrX2 complexes (Xyl = 2,6-dimethylphe-
nyl), their reactivity compared to unbridged analogues and in
rac-lactide polymerization.
N-substituents and – in the best cases – to statistical product mix-
tures.25 However, due to the favoured formation of N-aryl imines
over N-alkyl imines, a more than statistical yield of a mixed
diketimine was obtained, when acetylacetone was reacted with
mixtures of aniline and amine.25 We thus decided to employ
N-xylyl substituents and obtained the bridged ligand C6H10(nac-
nacXyl)2H2, 2H2, in 76% yield (Scheme 2). Reaction conditions
were chosen to disfavour N-substituent scrambling. Thus, acetyl-
acetone was first reacted with aniline to form the respective
enamine. Instead of the acid-catalysed condensation we normally
employ for symmetric N-alkyl diketimines and which requires
long reaction times,25 we followed a synthetic protocol used for
closely related ligands26–28 and used Meerwein’s salt to activate
the enamine ([Me3O][BF4] worked slightly better than the
respective ethyl salt in our hands). Nevertheless, the reaction
outcome was very sensitive to the reaction times involving
amines. Either addition of NEt3 directly with the amine or pro-
longed reaction times (steps 2–4, Scheme 2) drastically reduced
isolated yields of 2H2.
Results and discussion
Ligand synthesis
As a close analogue to the N-alkyl bisdiketiminate complexes
studied before,23 we attempted to prepare the bridged ligand 1
with N-benzyl substituents (Scheme 2). However, independent
of the starting point of the reaction or the synthetic path used
(acid-catalysed water elimination, alkylation with Meerwein’s
salt, use of ethyleneketal), either no reaction occurred or insepa-
rable product mixtures were obtained. This was in agreement
with previous observations in our group: while several non-
symmetrically substituted N-aryl diketimines have been reported,
we were unable to isolate a mixed N-alkyl diketimine. Closer
investigations indicated that attack at the imine group was faster
than the attack at the keto group, leading to scrambling of the
Complex syntheses
Reactions of ligand 2H2 with ZrBn4 under various conditions
did not yield an isolable product. Zr(NMe2)4 proved to be more
reactive and yielded 50% conversion in solution (C6D6, 60 °C).
Reaction in the absence of a solvent at 120 °C then afforded the
diamido complex 3 (Scheme 3) in 71% crystallized yield. The
latter complex reacted with dimethylammonium chloride or
ethanol to yield the respective dichloride and diethoxide
Scheme 2
Dalton Trans.
This journal is © The Royal Society of Chemistry 2012

View Online
Scheme 3
complexes 4 and 5 (Scheme 3). The protonation reaction,
however, was unselective and in both cases protonation of the
bisdiketiminate ligand occurred in significant amounts (up to
50%). Complex 4 was obtained, however, in good yields follow-
ing a salt metathesis route: deprotonation of bisdiketimine ligand
2H2 with nBuLi in THF yielded the ligand dilithium salt
2Li2(THF)2 (Scheme 3). Spectroscopic data and combustion
analysis suggest the coordination of one THF molecule per
lithium atom, as usually observed for diketiminate lithium salts
containing coordinated THF or diethyl ether. A further reaction
between 2Li2(THF)2 and ZrCl4(THF)2 yielded the zirconium
dichloride complex 4 in 80% yield (Scheme 3). Contrary to the
unbridged bis(diketiminate) zirconium complexes, nac-
nacR2ZrCl2, which did not react with MeLi even at higher temp-
eratures,23 reaction of dichloride complex 4 with MeLi at room
temperature in C6D6 afforded complete conversion of the
dichloride to the respective dimethyl complex 6 (contaminated
with significant amounts (35%) of the monomethylation
product) in less than 5 min (Scheme 3). Constraining the ligand
by introduction of a cyclohexanediyl bridge thus indeed resulted
in the intended increase of reactivity. Analytically pure dimethyl
complex 6 was isolated from large-scale reactions employing
MeLi and a catalytic amount of AlMe3 in 36% yield. Dichloride
4 also proved to be more reactive than unbridged bisdiketiminate
complexes towards NaOEt23 and cleanly yielded the diethoxy
complex 5 (Scheme 3), albeit under harsher reaction conditions
(3 d, 110 °C). Complex 5 can be prepared directly from
ZrCl4(THF)2 under milder conditions (12 h, 80 °C) and with
identical yields, if ZrCl4(THF)2 is first reacted with NaOEt2 and
then with 2Li2(THF)2. Formation of a putative ZrCl2(OEt)2 inter-
mediate is crucial. The direct reaction of ZrCl4 with a mixture of
NaOEt and 2Li2(THF)2 still yielded 5, but again required 3 d of
reaction time.
of the octahedral geometry. The bridgehead nitrogen atoms (N1
and N3) are severely tilted out of the ZrX2 plane (38–49° in
cis-C2-bridged, 8 5° in unbridged complexes, Table 1). The
reduced bite-angle enforced by the bridge (N1–Zr–N3 = 70–73°
in C2-bridged, 89 3° in unbridged complexes) further results in
a widening of the ZrX2 angles of the atoms trans to the bridge-
head nitrogen atoms (117–138° in cis-C2-bridged, 91 3° in
unbridged complexes) and slightly elongated (≈0.01 Å) Zr–N
distances for the non-bridgehead nitrogen atoms. A structural
effect of bridging favourable for potential applications is a slight
change in the position of the xylyl substituents. While the N-sub-
stituents in unbridged diketiminate complexes were positioned
between the ZrCl2-fragment (C–N–Zr–Cl torsion angles > 27°,
Table 1), they are eclipsed with one Zr–X bond in bridged com-
plexes (C–N–Zr–X torsion angles < 14°).
Of the possible structural isomers of C6H10(nacnacXyl)2ZrX2
complexes, only C2-symmetric isomers A and C were observed
in the solid state (Scheme 4). Complexes 3 and 5 show a cis-X2
configuration22 in the solid state, as was observed for all other
bisdiketiminate zirconium complexes reported (X–Zr–X angle =
89–117°).3,4,20–23 In both cases, the S,S,Δ/R,R,Λ-isomer (A,
Scheme 4) was observed, while the S,S,Λ/R,R,Δ-isomer (B,
Scheme 4) was absent. Complexes 4 and 6 crystallize as the
trans-isomer (C, Scheme 4). While variable temperature NMR
investigations have shown that for unbridged diketiminate com-
plexes a trans-Cl configuration is not obtained even as a short-
lived intermediate at elevated temperatures,23 the decreased N1–
Zr–N3 angle in 4 and 6, caused by the cyclohexanediyl bridge,
enables the placement of all four nitrogen atoms in the equatorial
plane. The preference for the trans-arrangement of the chloride
atoms in 4 is nevertheless surprising, since structures of cis-3
and cis-5 are nearly superposable with cis-C2H4(nacnacdipp)2-
ZrCl2 (dipp = 2,6-diisopropylphenyl, see also Fig. 5b),22 indicat-
ing that neither the cyclohexanediyl bridge nor the chloride
ligands should prevent formation of a cis-Cl complex. Most
likely, the trans-configuration is sterically slightly more favour-
able, but only for small ligands, such as chloride or methyl.
Space-filling diagrams of 4 and 6 indicate that increased steric
Solid state and solution structures. The solid state structures
of complexes 3–6 are shown in Fig. 1–4. Compared to unbridged
bisdiketiminate zirconium complexes, the constraint introduced
by bridging the diketiminate ligands results in a severe distortion
This journal is © The Royal Society of Chemistry 2012
Dalton Trans.

View Online
Fig. 3 X-ray structure of 5. Thermal ellipsoids are drawn at the 50%
probability level. Most hydrogen atoms and the disorder of one OEt
group were omitted for clarity.
Fig. 1 X-ray structure of 3. Thermal ellipsoids are drawn at the 50%
probability level. Most hydrogen atoms were omitted for clarity.
Fig. 2 X-ray structure of 4. Thermal ellipsoids are drawn at the 50%
probability level. Most hydrogen atoms and co-crystallized benzene
were omitted for clarity.
Fig. 4 X-ray structure of 6. Thermal ellipsoids are drawn at the 50%
probability level. Most hydrogen atoms were omitted for clarity. The
smaller fraction of the disordered cyclohexanediamine bridge is shown
in open lines without thermal ellipsoids or labelling.
demand, either of the ancillary ligand (as in 3 or 5) or of the
ortho-substituents on the phenyl ring (as in C2H4(nacnacdipp)2-
ZrCl2), would destabilize the trans-configuration. Additionally,
π-donation from the OEt or the NMe2 substituents in 3 and 5
might further favor a cis-geometry for these complexes.
glance. Since the diketiminate bite-angle of 74–78° (Table 1) is
significantly smaller than the ideal octahedral angle, nitrogen
atoms N1 and N3 trans to the ancillary ligands are slightly dis-
placed out of the ZrX2 plane (Fig. 1 and 3). In the case of the
Λ-isomer, an R,R-cyclohexanediyl bridge seems to accommodate
this deformation more readily than the S,S-configuration
The preference of cis-complexes 3 and 5 to form the S,S,Δ/
R,R,Λ-isomer (A, Scheme 4) seems intuitively correct on a first
Dalton Trans.
This journal is © The Royal Society of Chemistry 2012

View Online
Table 1 Bond distances [Å] and bond angles [°] in the crystal structures of 3–6
a
b
3
4
5
6
nacnacR2ZrCl2
C2H4(nacnacdipp)2ZrCl2
Zr–N1/N3
Zr–N2/N4
Zr–Xc
2.252(1)
2.374(1)
2.068(1)
71.92(7)
159.47(7) 133.18(8)
74.36(5)
2.183(2), 2.211(2) 2.256(2), 2.226(2) 2.190(5) –2.250(7) 2.23 0.05
2.285(2), 2.277(2) 2.307(2), 2.344(2) 2.315(3), 2.344(3) 2.22 0.05
2.425(1), 2.438(1) 1.937(2), 1.941(2) 2.265(4), 2.283(3) 2.45 0.06
2.175(3)
2.318(4)
2.422(2)
72.68(11)
160.03(10)
N1–Zr–N3d
N2–Zr–N4d
N1–Zr–N2
N3–Zr–N4
X–Zr–Xc
72.26(8)
72.32(7)
156.14(7)
76.41(7)
74.82(7)
121.13(8)
49
70.1(2), 72.0(2)
138.4(1)
75.5(1), 75.9(2)
75.0(1), 76.2(2)
138.0(1)
89
165 13
3
77.97(8)
77.85(8)
124.93(8) 156.24(3)
91
8
3
5
7
117.40(4)
38
53
(Zr/N1/N3)–(Zr/X1/X2)c,d 49
80
12
—
80, 81
(Zr/N1/N3)–(Zr/N2/N4)c
C–N2/4-Zr–Xe
44
13
44
1, 10
9, 10
—
83
37 10
3, 6
a Taken from ref. 23. b Taken from ref. 22. c 4: X = Cl1, Cl2; 3: X = N5, N5A; 5: X = O1, O2; 6: X = C37, C38. d For 3: N3 = N1A, N4 = N2A. For
6: N1 = N1A/B, N3 = N3A/B. e Torsion angle between the N-substituent and the ancillary ligand on Zr.
Scheme 4
Fig. 5 (a) X-ray structure of (R,R,Λ)-5 with an idealized S,S-cyclohexanediamine fitted into the structure (hollow lines). (b) Best fit overlay of
C2H4(nacnacdipp)2ZrCl2 22 and 5. Ethyl groups, aryl substituents and hydrogen atoms were omitted for clarity.
(Scheme 5). To investigate in more detail this simplistic expla-
nation, which neglects the strong deviation from octahedral geo-
metry observed in 3 and 5, as well as the typical distortion of the
diketiminate ligand into a boat-like conformation,19,24 an ideal-
ized S,S-cyclohexanediyl bridge was fitted into the crystal struc-
ture of Λ-5 (Fig. 5a). The change of the configuration in the
cyclohexanediyl bridge requires a higher bending of the N–C3/
C23 bond out of the diketiminate mean plane (52° for the
hypothetical S,S,Λ-isomer depicted in Fig. 5a, 24 3° for the
Scheme 5
This journal is © The Royal Society of Chemistry 2012
Dalton Trans.

View Online
R,R,Λ-isomers 3 and 5). This latter value is very close to the
out-of-plane bending of the N-substituent in unbridged N-alkyl
bisdiketiminate complexes (16–24°)23 and in C2H4- and C3H6-
bridged bisdiketiminate complexes (24–33°).22 In fact, crystal
complexes by a Bailar-Twist would result in an isomer with the
bridgehead nitrogen atoms in an impossible trans-configuration.
Contrary to unbridged nacnacR2ZrX2 complexes, Δ–Λ-isomeri-
sation via a Bailar-Twist mechanism thus has to pass by a trans-
X2 complex (Scheme 6). Alternatively, based on the fact that iso-
merisation is faster for C2H4- than for C3H6-bridged complexes,
Dong et al. had proposed that Δ–Λ-isomerisation in bridged bis-
diketiminate complexes proceeds by dissociation of the bridging
nitrogen atoms and through a tetrahedral intermediate rather than
via a Bailar-Twist mechanism (Scheme 6).22 Although only two
structurally characterized examples of κ1-coordinated diketimi-
nate ligands were reported,29,30 such an intermediate does not
seem unreasonable, in particular considering the steric strain
introduced by the bridge. In the case of C2H4-bridged com-
plexes, the trans-X2 complex as well as the four-coordinated
intermediate easily achieve apparent C2v-symmetry by a ring
inversion of the metallacycle(s) and Δ–Λ-isomerisation would
thus lead to the observed equalization of the xylyl methyl groups
for C2H4-bridged complexes (Scheme 6). In 3–6, even fast inver-
sion of the metallacycle(s) of the same intermediates results only
in apparent C2-symmetry and the inequivalence of the xylyl
methyl groups is preserved. (In fact, if 4 and 6 retain the C1-sym-
metric trans-X geometry observed in the solid state, fast ring
inversion has to be present to result in the observed (apparent)
C2-symmetry of the NMR spectra.) While the equivalence of the
xylyl methyl resonances is thus a further indication of fast iso-
merisation of the complex geometry, the existence of two distinct
xylyl methyl peaks in 3–6 unfortunately does not indicate the
absence of these isomerisations.
To further investigate the possibility of a fast isomerisation,
proton NMR spectra of 5 were recorded at variable temperatures
down to −50 °C. While no splitting of resonances was observed,
significant broadening of all resonances occurred below −20 °C,
indicative of a dynamic process (Fig. 6 and ESI†). In summary,
while there is thus no clear evidence for dynamic processes in
3–6, the available data indicate that a fast isomerisation is indeed
present. Given that isomers A and C were observed in the solid
state, it seems reasonable to presume that the complexes retain
the geometry observed in the solid state and that a cis–trans iso-
merisation is responsible for the peak broadening at low temp-
eratures, without necessarily forming the wrong stereomatch B
or C1-symmetric species D.
structures
of
5
and
the
C2H4-bridged
complex
22
C2H4(nacnacdipp)2ZrCl2
are nearly superposable (Fig. 5b),
indicating the correct stereochemical match between an R,R-
configured cyclohexanediyl bridge and a Λ-configuration at the
metal center.
In solution, proton NMR spectra of 3–6 show one signal for
the methine protons of the cyclohexane ring, one signal for the
central CH resonance of the diketiminate ligand, two signals for
the diketiminate methyl groups and two signals for the xylyl
methyl groups. The C2-symmetry of the ligand is thus preserved
in the complex (either by symmetry or fast isomerisation). The
presence of two xylyl methyl resonances, of three aromatic reso-
nances in proton NMR spectra of 3–6, as well as of four reson-
ances for the aromatic ring in carbon NMR spectra of 3–6, and
the fact that two xylyl methyl resonances were observed also for
lithium complex 2Li2(THF)2 and the free ligand 2H2 (see the
Exp. section), all indicate that N-aryl rotation is slow on the
NMR time scale. In agreement with hindered N-aryl rotation,
NMR spectra of 5 did not indicate any coalescence of the xylyl
methyl groups up to 70 °C.
Unbridged diketiminate complexes have shown evidence for
fast Δ–Λ-isomerisation in solution (A ↔ B, Scheme 4) by a
Bailar-Twist mechanism.3–5,23 Despite the introduction of a brid-
ging group, the same fast isomerisation was observed by Dong
et al. in C2H4- or C3H6-bridged complexes: coalescence of the
CH2 protons in the bridge yielded averaged C2v-symmetric
NMR spectra,22 in the case of a C2H4-bridge even at −70 °C.
NMR spectra of 3–6 do not contain any group which would
coalesce in the case of a fast Δ–Λ-isomerisation and thus no
straightforward way to establish the absence or presence of Δ–
Λ-isomerisation. However, while spectra of C2H4(nacnacXyl)2-
ZrCl2 displayed only one resonance for the xylyl methyl group
even at low temperatures,22 two xylyl methyl resonances were
observed in spectra of 3–6. In the absence of N-xylyl rotation
(vide supra), the equivalence of the xylyl methyl groups in
C2H4(nacnacXyl)2ZrCl2 has to be caused by passage through a
C2v-symmetric intermediate associated with the Δ–Λ-isomerisa-
tion. Direct Δ–Λ-isomerisation of bridged bisdiketiminate
Scheme 6
Dalton Trans.
This journal is © The Royal Society of Chemistry 2012

View Online
Fig. 7 PLA concentration vs. time for rac-lactide polymerization with
5: [lactide] = 0.6 M; [5] = 2 mM; THF; ambient temperature. The inset
shows the linearised plot of monomer consumption assuming a first-
order rate law; kobs = 0.030 s−1
.
Fig. 6 1H NMR spectra (aliphatic region) of 5 at different tempera-
tures. See ESI† for complete spectra.
Lactide polymerization
Attempts to obtain
a
cationic alkoxide complex
[C6H10(nacnacXyl)2ZrOEt]+, which would be an attractive cata-
lyst for lactide polymerization, from the reaction of 5 with either
para-toluenesulfonic acid, triflic acid or [PhN(H)Me2][OTf]
under a variety of conditions yielded only complex product mix-
tures, in which the protonated ligand was the most prevalent
species. The reaction of 4 with one equiv. of AgOTf, followed
by the reaction with NaOEt, likewise did not afford the desired
cationic complex. We thus investigated octahedrally coordinated
5 directly as a catalyst for lactide polymerization. While six-
coordinate zirconium alkoxide complexes have been shown to be
active for lactide polymerization, they typically display only
moderate activity, requiring several hours at temperatures of
Fig. 8 Dependence of apparent rate constants kobs on catalyst concen-
tration; k = 14(1) L mol−1 s−1
.
catalyst concentration (Fig. 8), yielding a second-order rate con-
stant at room temperature of k298 K = 14(1) L mol−1 s−1. To the
best of our knowledge, this is the highest activity, by several
orders of magnitude, reported for any zirconium or other group 4
complexes in lactide polymerization. Highly active Zr catalysts
include a salalen Zr complex, described as “incredibly active”
and which reached 99% conversion in 6 h (k298 K ≈ 5 × 10−2
>70 °C in solution31–34 or polymerization in
a molten
polymer,35–37 which is not surprising for a sterically saturated
octahedral complex. We were thus surprised to note that rac-
lactide polymerization with 5 at room temperature in a dichloro-
methane solution reached 40% conversion already after one
minute. Conversion did not exceed 50%, however, indicative of
catalyst decomposition in this solvent (see ESI†). Better catalyst
stability was obtained in toluene or diethyl ether, but the lactide
monomer has only limited solubility in these solvents. Poly-
merization in THF finally led to fast and complete (95%)
conversion of the monomer in 5 min (see ESI†). For comparison,
the unbridged bisdiketiminate complex nacnacBn2Zr(OEt)2
was only reactive in the molten monomer at 130 °C and
still required 30 min to reach >95% conversion.23 As observed
for the substitution of chloride ligands in 4, bridging of the
diketiminates thus drastically increased the reactivity of the
complex.
L mol−1 s−1),40 a carbene Zr complex (95% after 15 h, k298 K
≈
5 × 10−3 L mol−1 s−1),41 a sulfonamide Zr complex (5 h at
60 °C to reach completion, k333 K ≈ 1 × 10−2 L mol−1 s−1),42
and a dithiodiolate Hf complex reported in 2010 to show the
“highest activity of any group 4 metal catalyst”, which reached
complete conversion in 1 min in a molten monomer (k403 K
≈
4 L mol−1 s−1).
While 5 thus displays an extremely high activity for group 4
complexes and compares well with the most active catalyst
systems reported so far for the polymerization of lactide, its
remaining characteristics, such as stereoselectivity, complex
stability and polymer molecular weight control, are much less
desirable. Stereoselectivity: all obtained polymers are essentially
atactic with a slight heterotactic bias (Pr = 0.56–0.64). The lack
of isoselectivity imposed by the C2-symmetric complex might be
due to thermodynamic, kinetic or mechanistic reasons: (i) in the
crystal structure of 3 and 5 we note that the introduction of the
bridge forces the central atom of the diketiminate ligand
(3: Fig. 1, C12; 5: Fig. 3, C6/C36) towards the ancillary ligands,
occupying the “empty” part of the catalytic pocket. (ii) The
observed dynamic process in the NMR might prevent stereoche-
mical stability of the catalytic site. (iii) The active species might
not be a C2-symmetric cis-isomer, but rather a species for which
More detailed investigations of the reaction kinetics showed
the expected first-order dependence on monomer concentration
(Fig. 7). Reactions typically reached completion in less than
5 min, but conversions did not exceed ≈95%. Incomplete con-
version is typical for lactide polymerizations and attributed to a
reversible polymerization.38,39 Apparent rate constants at catalyst
concentrations of 0.5–2 mM show a linear dependence on
This journal is © The Royal Society of Chemistry 2012
Dalton Trans.

View Online
less stereocontrol by the chiral bridge would be expected, such
as the trans-isomer or a tetracoordinated species (Scheme 6).
Complex stability: despite rigorous drying of solvents and
repeated recrystallisation of lactide, the stability of 5 under
polymerization conditions is limited. All kinetic analyses show
slight deviations from first order behaviour, which can be
ascribed to catalyst decomposition. No further polymerization is
observed, when a second portion of lactide is added after 15 min
polymerization time. In fact, polymerizations either reached
completion in less than 5 min or, in the case of lower activities,
such as with catalyst concentrations below 0.5 mM, did not
reach completion at all.
combinations thereof, as an initiator did not result in any activity
of the same order of magnitude as the activity observed for 5
(see ESI†). Although these polymerizations were not investi-
gated in detail, it should be noted that sodium ethoxide in the
presence of free ligand 2H2 showed surprisingly high activities
(94% conversion in 1 h at ambient temperature).
In an alternative approach, we added selected contaminants to
rac-lactide polymerizations with 5slow, i.e. stock solutions of 5
which showed less than the normally observed activity (approxi-
mately 20–35% of the maximum activity observed, cf.
Table S3†). Surprisingly, addition of protic contaminants, such
as water or para-toluenesulfonic acid, slightly increased the
polymerization activity of 5slow, albeit never surpassing the
maximum activity observed for 5. Addition of lactic acid
quenched the polymerization (Table S3†). While the slight rate-
enhancing effect of water remains unexplained at the moment, it
seems improbable that activation by protic substances is respon-
sible for the high activities observed (see ESI†). Water (or
TsOH) might rather be involved in the re-activation of deacti-
vated species, for example after the reaction with lactic acid
impurities, although no mechanism for this can be proposed at
the moment.
Polymer molecular weight. Resulting polymers showed
broad, sometimes bimodal polymer molecular weight distri-
butions with polymer molecular weights much lower than
expected even for two chains produced per zirconium center and
despite high conversions of >90%. Polymerizations in solvents
other than THF or at different lactide : Zr ratios yielded the same
results (see ESI†). The obtained polymers contained a sizeable
fraction of oligomers with polymerization degrees lower than 10.
Maldi-MS analysis of the oligomeric fraction showed a series of
peaks with m/z ratios of n·72 + m(Na+) (Fig. 9). Combined with
the overall to high number of polymer chains per catalyst, this
indicates that intramolecular transesterification leads to the for-
mation of cyclic oligomers next to linear polymers.
Conclusion
Intrigued by the variations of activity on aging of catalyst
stock solutions, as well as by the unusual high activity of 5 in
general, we investigated the effects of potential contaminants on
polymerization activity. Due to the possibility of fast alkoxide
exchange, the observed polymer molecular weights do not
exclude that small quantities of a highly active species are
responsible for the observed polymerization activity. Polymeriz-
ations using possible contaminants from the catalyst preparation,
i.e. free ligand 2H2, sodium ethoxide, ZrCl4(THF)2 or
Bridging of two diketiminate ligands by a cyclohexanediyl
bridge increased, as intended, the reactivity of bisdiketiminate
zirconium complexes, which are now able to undergo ligand
exchange and show vastly higher activities in lactide polymeriz-
ation. The chiral bridge also shows a clear stereochemical impact
on the conformation at the metal center. Unfortunately, the steric
constraints introduced permit the formation of undesired trans-
X2 species and variable temperature NMR spectra showed
evidence for a dynamic process, which is most likely a fast
Fig. 9 Maldi-mass spectrum of the oligomeric fraction of PLA obtained from 5 (2 mM, THF, [lactide]/[5] = 100/1).
This journal is © The Royal Society of Chemistry 2012
Dalton Trans.

View Online
X-ray diffraction studies
isomerisation between cis- and trans-X2 complexes. In
lactide polymerization, the unprecedented high activity of
diethoxy complex 5 provides valuable starting points in optimiz-
ing group 4-based lactide polymerization catalysts. Its lack
in stereoselectivity, its low stability under polymerization
conditions, and the formation of notable amounts of cyclic
oligomers drastically reduce its value as a catalyst for cyclic
ester polymerizations, however. We are currently investigating
other applications of the chiral C6H10(nacnacXyl)2ZrX2
system, in particular hydroamination and carbozirconation
reactions.
Diffraction data were collected on a Bruker Smart 6000 with
Helios MX mirror optics and Cu Kα radiation (rotating anode),
using the APEX2 software package.46 Data reduction was per-
formed with SAINT,47 absorption corrections with SADABS.48
Structures were solved with direct methods (SHELXS97).49 All
non-hydrogen atoms were refined anisotropically using full-
matrix least-squares on F2 and hydrogen atoms refined with
fixed isotropic U using a riding model (SHELXL97).49 One
ethyl group in 5 was disordered and refined anisotropically with
refined occupation factors of 0.73 and 0.27. The cyclohexanedi-
amine bridge in 6 was found to be disordered and was refined,
partially isotropically, with appropriate restraints. For further
details, please see Table 2 or the ESI.†
Experimental section
All reactions, except ligand synthesis, were carried out using
Schlenk and glove box techniques under a nitrogen atmosphere.
ZrCl4(THF)2,43 Zr(NMe2)4,44 and 4-[(2,6-dimethylphenyl)-
amino]-pent-3-en-2-one45 were prepared according to literature
procedures. ZrCl4, NaOEt, LiNMe2 and other chemicals were
purchased from common commercial suppliers. 1H and 13C
NMR spectra were acquired on a Bruker AMX 400 or Bruker
AV 400 spectrometer. 19F NMR spectra were acquired on a
Bruker AV 400 spectrometer. Chemical shifts were referenced to
the residual signals of the deuterated solvents (C6D6: 1H: δ
( )-C6H10(nacnacXyl)2H2,
7.3 mmol) was added to a yellow dichloromethane solution of
4-[(2,6-dimethylphenyl)amino]-pent-3-en-2-one (1.49 g,
2H2. [Me3O][BF4]
(1.08
g,
7.3 mmol, 10 mL). The resulting suspension was stirred for
approximately 2 h until complete dissolution of [Me3O][BF4]. A
dichloromethane solution of ( )-trans-cyclohexane-1,2-diamine
(0.21 g, 1.8 mmol, 1.0 mL) was added, after which stirring was
continued for 30 min to yield an orange solution. A mixture of
( )-trans-cyclohexane-1,2-diamine (0.24 mL, 1.8 mmol) and tri-
ethylamine (0.76 mL, 5.4 mmol) was added. After short
additional stirring (5 min), the volatiles were removed in vacuo.
Diethyl ether (100 mL) was added to the residue. The obtained
suspension was treated with an excess of Et3N (0.11 mol,
15 mL) and stirred for 5 min. From the yellow phase separated
by decantation, solvents were removed in vacuo and the residual
yellow solid was crystallized from ethanol at −20 °C to afford
1.35 g (76%) of yellow crystals.
1
7.16 ppm, 13C: δ 128.38 ppm; CDCl3: H: δ 7.26 ppm, 13C: δ
77.00 ppm). Solvents were dried by passage through activated
aluminum oxide (MBraun SPS) and de-oxygenated by repeated
extraction with nitrogen. Polymerization solvents were addition-
ally passed through a column of activated molecular sieves.
C6D6 was dried over sodium, CDCl3 was dried over CaH2 and
both were distilled under reduced pressure, and then degassed by
three freeze–pump–thaw cycles. Racemic lactide was sublimed
and recrystallized twice from dry ethyl acetate. Elemental ana-
lyses were performed by the Laboratoire d’analyse élémentaire
(Université de Montréal).
1H NMR (CDCl3, 400 MHz): δ 10.94 (bs, 2H, NH), 7.01 (d, J
= 2 Hz, 4H, meta Ar), 6.85 (t, J = 2 Hz, 2H, para Ar), 4.56 (bs,
2H, CH(CvN)), 3.14 (m, 2H, Cy CH), 2.17 (s, 6H, Me2Ar),
Table 2 Details of X-ray diffraction studies
3
4
5
6
Formula
C36H54N6Zr
662.07; 1.27
0.14 × 0.14 × 0.18
200; 1408
Monoclinic
C2/c
C32H42Cl2N4Zr·C6H6
722.92; 1.33
0.03 × 0.03 × 0.08
150; 756
C36H52N4O2Zr
664.04; 1.27
0.08 × 0.08 × 0.08
150; 704
C34H48N4Zr
603.98; 1.26
0.1 × 0.1 × 0.3
150; 1280
Monoclinic
P21/n
Mw (g mol−1); dcalcd (g cm−3
Crystal size (mm)
T (K); F(000)
Crystal system
Space group
)
Triclinic
Triclinic
ˉ
ˉ
P1
P1
Unit cell:
a (Å)
b (Å)
c (Å)
α (°)
β (°)
18.5065(3)
11.7278(2)
17.2506(4)
9.7217(4)
10.9326(11)
11.2187(12)
16.3787(16)
82.459(5)
71.522(5)
65.213(5)
1729.7(3); 2
2.8–69.7; 0.98
66389; 0.032
6426; 0.084
2.88; multi-scan
0.037; 0.096
0.039; 0.102
1.107
11.4759(9)
14.9142(12)
18.6288(14)
11.9708(5)
16.5433(7)
100.423(2)
104.614(2)
96.408(2)
1807.0(1); 2
2.8–69.6; 0.97
25219; 0.060
6640; 0.049
4.09; multi-scan
0.033; 0.070
0.047; 0.073
0.915
112.842(1)
91.042(3)
γ (°)
V (Å3); Z
3450.5(1); 4
4.6–72.6; 0.99
22047; 0.014
3393; 0.031
2.85; multi-scan
0.027; 0.073
0.027; 0.073
1.098
3187.9(4); 4
3.8–69.8; 1.00
116139; 0.024
5989; 0.083
3.02; multi-scan
0.045; 0.117
0.047; 0.118
1.156
θ range (°); completeness
Collected reflections; Rσ
Unique reflections; Rint
μ (mm−1); abs. corr.
R1(F); wR(F2) (I > 2σ(I))
R1(F); wR(F2) (all data)
GoF(F2)
Residual electron density
0.36, −0.41
0.58, −0.34
0.97; −0.58
1.39; −0.68
This journal is © The Royal Society of Chemistry 2012
Dalton Trans.

View Online
2.04 (s, 6H, Me2Ar), 2.02 (s, 6H, Me(CvN)), 1.95–1.86 (m,
8H, CH2), 1.59 (s, 6H, Me(CvN)). 13C{1H} NMR (CDCl3,
101 MHz): δ 165.9 (MeCvN), 155.2 (MeCvN), 149.5 (ipso
Ar), 128.1 (Ar), 127.7 (Ar), 127.6 (Ar), 127.5 (Ar), 121.7 (para
Ar), 92.6 (CH(CvN)), 58.0 (Cy CH), 33.5 (Cy CH2), 24.7 (Cy
CH2), 21.2 (MeCvN), 19.6 (MeCvN), 18.5 (Me2Ar), 18.2
(CvN)), 21.3 (Me2Ar), 18.9 (Me2Ar). Anal. calcd for
C36H54N6Zr: C, 65.31; H, 8.22; N, 12.69; found C, 64.68; H,
8.86; N, 12.42.
( )-C6H10(nacnacXyl)2ZrCl2, 4. A suspension of ZrCl4(THF)2
(0.50 g, 1.34 mmol) in toluene (10 mL) was added drop by drop
to a solution of 2Li2(THF)2·0.5 THF (0.86 g, 1.34 mmol) in
toluene (10 mL) while stirring. After 10 min at room tempera-
ture, the yellow suspension obtained was filtered and the remain-
ing residue extracted with toluene (5 mL). The toluene solution
was evaporated to dryness yielding a yellow powder (0.69 g,
80%). Crystals suitable for elemental analysis and X-ray struc-
ture determination were obtained by slow evaporation of C6D6.
1H NMR (CDCl3, 400 MHz): δ 6.94–6.85 (m, 6H, Ar), 5.45
(bs, 2H, CH(CvN)), 4.19 (m, 2H, Cy CH), 2.18 (s, 6H, Me
(CvN)), 2.14 (s, 6H, Me2Ar), 2.11 (m, 2H, Cy CH2), 2.06 (s,
6H, Me2Ar), 1.87–1.72 (m, 4H, Cy CH2), 1.64 (s, 6H, Me
1
(Me2Ar). H NMR (C6D6, 400 MHz): δ 11.09 (bs, 2H, NH),
7.10 (d, J = 2 Hz, 4H, meta Ar), 6.96 (t, J = 2 Hz, 2H, para Ar),
4.61 (bs, 2H, CH(CvN)), 2.90 (m, 2H, Cy CH), 2.20 (s, 6H,
Me2Ar), 2.17 (s, 6H, Me2Ar), 1.87 (s, 6H, Me(CvN)), 1.67 (m,
2H, Cy CH2), 1.58 (s, 6H, Me(CvN)) 1.3–0.80 (m, 6H, Cy
CH2). 13C{1H} NMR (C6D6, 101 MHz): δ 166.2 (MeCvN),
155.2 (MeCvN), 150.3 (ipso Ar), 128.3 (Ar), 128.2 (Ar), 128.1
(Ar), 127.9 (Ar), 122.5 (para Ar), 93.6 (CH(CvN)), 58.2 (Cy
CH), 33.6 (Cy CH2), 24.6 (Cy CH2), 21.3 (MeCvN), 19.7
(MeCvN), 18.8 (Me2Ar), 18.7 (Me2Ar). Anal. calcd for
C13H26N2: C, 79.29; H, 9.15; N, 11.56; found C, 79.05; H, 9.18;
N, 11.56.
1
(CvN)), 1.41 (m, 2H, Cy CH2). H NMR (C6D6, 400 MHz): δ
6.95–6.84 (m, 6H, Ar), 5.39 (bs, 2H, CH(CvN)), 4.19 (m, 2H,
Cy CH), 2.37 (s, 6H, Me2Ar), 2.29 (s, 6H, Me2Ar), 1.86 (m, 2H,
Cy CH2), 1.83 (s, 6H, Me(CvN)), 1.56 (s, 6H, Me(CvN)),
1.53–1.46 (m, 4H, Cy CH2), 1.03 (m, 2H, Cy CH2). 13C{1H}
NMR (C6D6, 101 MHz): δ 66.2 (MeCvN), 161.2 (MeCvN),
133.1 (ortho Ar), 132.6 (ortho Ar), 128.5 (ipso Ar), 128.4 (meta
Ar), 128.3 (meta Ar), 125.1 (para Ar), 105.7 (CH(CvN)), 69.0
(Cy CH), 32.6 (Cy CH2), 25.3 (MeCvN), 23.3 (MeCvN), 22.5
(Me2Ar), 19.6 (Me2Ar), 19.53 (Cy CH2). Anal. calcd for
C32H42N4ZrCl2·C6D6: C, 62.61; H, 6.64; N, 7.69; found C,
62.79; H, 6.94; N, 7.57 (X-ray structure shows the presence of
one C6D6).
( )-C6H10(nacnacXylLi(THF))2·0.5
THF,
2Li2(THF)2·0.5
THF. To a yellow THF solution of ligand 1a (0.3 g,
0.63 mmol), nBuLi (2.9 M, 0.43 mL, 0. 13 mmol) was added
gradually at room temperature. The obtained orange solution was
stirred for 5 min. All volatiles were removed in vacuo to give an
orange oil. Addition of hexanes (5 mL) gave an off-white pre-
cipitate, which was isolated by decantation and dried under
vacuum to yield an off-white powder (0.32 g, 80%).
1H NMR (C6D6, 400 MHz): δ 7.10 (d, J = 2 Hz, 2H, meta
Ar), 7.03 (d, J = 2 Hz, 2H, meta Ar), 6.90 (t, J = 2 Hz, 2H, para
Ar), 4.67 (bs, 2H, CH(CvN)), 3.40 (m, 10H, THF), 3.29 (m,
2H, Cy CH), 2.32 (s, 6H, Me2Ar), 2.17 (s, 6H, Me2Ar), 2.12 (m,
2H, Cy CH2), 1.99 (s, 6H, Me(CvN)), 1.78 (s, 6H, Me(CvN)),
1.40–1.32 (m, 6H, Cy CH2), 1.14 (m, 10H, THF). 13C{1H}
NMR (C6D6, 101 MHz): δ 164.5 (MeCvN), 161.8 (MeCvN),
153.6 (ipso Ar), 131.2 (ortho Ar), 130.7 (ortho Ar), 128.13
(meta Ar), 127.8 (meta Ar), 121.4 (para Ar), 93.5 (CH(CvN)),
68.3 (THF), 65.9 (Cy CH), 35.9, 19.6 (Cy CH2), 26.5, 23.0
(MeCvN), 25.2 (THF), 19.6 (Me2Ar), 19.0 (Me2Ar). Anal.
calcd for C40H58Li2N4O2·0.5 C4H8O: C, 74.53; H, 9.23; N,
8.28; found C, 74.14; H, 9.22; N, 8.68. (NMR shows the pres-
ence of 2.5 THF).
( )-C6H10(nacnacXyl)2Zr(OEt)2, 5. Method A: a yellow solu-
tion of 4 (0.32 g, 0.50 mmol) in toluene (10 mL) was added
drop by drop to a suspension of NaOEt (0.86 g, 1.34 mmol) in
toluene (5 mL). The obtained red suspension was refluxed for 3
days. The suspension was filtered while hot and the obtained
filtrate was evaporated to dryness to yield a red oil (0.30 g, 91%,
90% purity determined by NMR). Addition of hexane (2 mL)
gave a red solution which yielded orange crystals separated upon
standing (200 mg, 60%).
Method B: to a suspension of ZrCl4(THF)2 (222 mg,
0.59 mmol) in toluene (5 mL) was added a suspension of NaOEt
(80 mg, 1.2 mmol) in toluene (20 mL). The suspension was
heated for 3 h at 80 °C. The resulting mixture was added to a
solution of 2Li2THF2 (400 mg, 0.59 mmol) in toluene (10 mL).
The orange suspension was heated overnight at 80 °C, then
filtered. Recrystallisation of the residue yielded 5 as orange crys-
tals (55%).
( )-C6H10(nacnacXyl)2Zr(NMe2)2, 3. Zr(NMe2)4 (0.14 g,
0.53 mmol) and ligand 2H2 (0.25 g, 0.53 mmol) were mixed
and heated, in the absence of a solvent, to 120 °C under a N2
atmosphere for 1 h. The obtained brown oil was dissolved in
toluene (1 mL) and crystallized by slow evaporation at room
temperature (0.25 g, 71%).
1H NMR (C6D6, 400 MHz): δ 7.10 (d, J = 2 Hz, 2H, meta
Ar), 7.00 (d, J = 2 Hz, 2H, meta Ar), 6.93 (t, J = 2 Hz, 2H,
p-Ar), 5.04 (bs, 2H, CH(CvN)), 3.79 (m, 2H, Cy CH), 2.80 (s,
6H, NMe2), 2.50 (s, 6H, Me2Ar), 2.13 (m, 2H, Cy CH2), 2.09 (s,
6H, Me2Ar), 1.90 (s, 6H, Me(CvN)), 1.77 (s, 6H, NMe2), 1.58
(s, 2H, Cy CH2), 1.56 (s, 6H, Me(CvN)), 1.29–1.25 (m, 4H,
Cy CH2). 13C{1H} NMR (C6D6, 101 MHz): δ 166.1 (Me
(CvN)), 163.0 (Me(CvN)), 153.7 (ipso Ar), 132.1 (ortho Ar),
131.9 (ortho Ar), 128.7 (meta Ar), 127.7 (meta Ar), 123.7 (para
Ar), 105.7 (CH(CvN)), 70.2 (Cy CH), 43.2 (Cy CH2), 41.5 (Cy
CH2), 35.3 (NMe2), 25.7 (NMe2), 24.8 (Me(CvN)), 22.6 (Me
1H NMR (C6D6, 400 MHz): δ 7.06 (d, J = 8 Hz, 2H, meta
Ar), 7.04 (d, J = 8 Hz, 2H, meta Ar), 6.93 (t, J = 8 Hz, 2H, para
Ar), 5.10 (bs, 2H, CH(CvN)), 3.75 (m, 2H, Cy CH), 3.20 (m,
4H, OCH2), 2.37 (s, 6H, Me2Ar), 2.20 (s, 6H, Me2Ar), 2.07 (m,
2H, Cy CH2), 1.96 (s, 6H, Me(CvN)), 1.55–1.48 (m, 8H, Cy
CH & Me(CvN)), 1.30 (m, 4H, Cy CH2), 0.77 (t, J = 7 Hz, 6H,
OCH2Me). 13C{1H} NMR (C6D6, 101 MHz): δ 63.8 (MeCvN),
162.8 (MeCvN), 152.8 (ipso Ar), 132.6 (ortho Ar), 130.9
(ortho Ar), 128.6 (meta Ar), 127.9 (meta Ar), 123.9 (para Ar),
104.1 (CH(CvN)), 68.8 (Cy CH), 63.8 (OCH2Me), 34.8 (Cy
CH2), 25.8 (OCH2Me), 22.7 (Me2Ar), 22.3 (Cy CH2), 19.7
Dalton Trans.
This journal is © The Royal Society of Chemistry 2012

View Online
(Me2Ar), 19.2 (Me(CvN)), 19.0 (Me(CvN)). Anal. calcd for
C36H52N4O2Zr: C, 65.11; H, 7.89; N, 8.44; found C, 64.92; H,
8.04; N, 8.14.
Acknowledgements
Floriane Cuenca and Arnaud Parrot contributed during their
internship to the work presented here. We thank Sylvie Bilodeau
for variable temperature NMR measurements, Erik Fournaise for
MS measurements and Francine Bélanger for help with X-ray
crystallography. I. E. and T. W. received a doctoral stipend from
the Fonds de recherche du Québec – Nature et technologies
(FQRNT). This work was supported by the Canadian National
Research Council (NSERC) and the Centre in Green Chemistry
and Catalysis (CGCC).
( )-C6H10(nacnacXyl)2ZrMe2·0.5 toluene, 6. A mixture of
MeLi (10 mg, 0.45 mmol) in toluene (2 mL) and AlMe3 in
hexane (0.20 M, 0.12 mL, 0.024 mmol) was gradually added at
−78 °C to a solution of 4 (150 mg, 0.23 mmol) in toluene
(10 mL) and stirred for 5 min. The cold bath was removed and
stirring was completed in another 5 min. The obtained yellow
suspension was filtered cold and the remaining residue was
extracted with toluene (5 mL). The combined toluene filtrates
were concentrated to 2 mL. Slow diffusion of hexane into a
toluene solution of 6 at −35 °C gave yellow crystals (50 mg,
References
1
36%). H NMR of the obtained crystals showed the presence of
1 P. B. Hitchcock, M. F. Lappert and D.-S. Liu, Chem. Commun., 1994,
2637–2638.
0.5 equiv. of toluene.
1H NMR (C6D6, 400 MHz): δ 7.10–6.93 (m, 6H, Ar), 5.27
(bs, 2H, CH(CvN)), 3.69 (m, 2H, Cy CH), 2.34 (s, 6H,
Me2Ar), 2.29 (s, 6H, Me2Ar), 2.07–1.93 (m, 4H, Cy CH2), 1.86
(s, 6H, Me(CvN)), 1.66 (s, 6H, Me(CvN)), 1.50 (m, 2H, Cy
CH2), 1.04 (m, 2H, Cy CH2), −0.06 (s, 6H, ZrMe). 13C NMR
(C6D6, 101 MHz): δ 164.0, 156.2 (Me(CvN)), 153.1 (ipso Ar),
132.2 (ortho Ar), 131.7 (ortho Ar), 129.6 (meta Ar), 128.8
(meta Ar), 124.7 (para Ar), 103.8 (CH(CvN)), 67.4 (Cy CH),
57.7, 33.9 (Cy CH2), 26.0 (Me2Ar), 23.9 (Me2Ar), 22.7 (Me
(CvN)), 19.9 (Me(CvN)), 19.8 (ZrMe). Anal. calcd for
C34H48N4Zr·12C7H8. C, 69.29; H, 8.06; N, 8.62; found C, 68.87;
H, 8.22; N, 8.26.
2 B. Qian, W. J. Scanlon, M. R. Smith and D. H. Motry, Organometallics,
1999, 18, 1693–1698.
3 M. Rahim, N. J. Taylor, S. Xin and S. Collins, Organometallics, 1998,
17, 1315–1323.
4 L. Kakaliou, W. J. Scanlon, B. Qian, S. W. Baek, M. R. Smith and
D. H. Motry, Inorg. Chem., 1999, 38, 5964–5977.
5 R. Vollmerhaus, M. Rahim, R. Tomaszewski, S. Xin, N. J. Taylor and
S. Collins, Organometallics, 2000, 19, 2161–2169.
6 X. Jin and B. M. Novak, Macromolecules, 2000, 33, 6205–6207.
7 F. Basuli, U. J. Kilgore, D. Brown, J. C. Huffman and D. J. Mindiola,
Organometallics, 2004, 23, 6166–6175.
8 H. Hamaki, N. Takeda and N. Tokitoh, Organometallics, 2006, 25,
2457–2464.
9 R.-V. Fortuné, E. Verguet, P. O. Oguadinma and F. Schaper, Acta Crystal-
logr. Struct, Rep. Online, 2007, 63, m2822–m2823.
10 A. K. Mittal, A. Shukla, P. R. Shukla, A. K. Pathak and N. Ahmad,
Synth. React. Inorg. Met.-Org. Chem., 1995, 25, 739–759.
11 E. Shaviv, M. Botoshansky and M. S. Eisen, J. Organomet. Chem., 2003,
683, 165–180.
rac-Lactide polymerization
12 R. Andrés, E. de Jesús, F. J. de la Mata, J. C. Flores and R. Gómez,
J. Organomet. Chem., 2005, 690, 939–943.
13 E. Verguet, R.-V. Fortuné, P. O. Oguadinma and F. Schaper, Acta Crystal-
logr. Struct, Rep. Online, 2007, 63, m2539–m2540.
14 E. Verguet, P. O. Oguadinma and F. Schaper, Acta Crystallogr. Struct,
Rep. Online, 2007, 63, m2541–m2542.
15 A. R. Cabrera, Y. Schneider, M. Valderrama, R. Fröhlich, G. Kehr,
G. Erker and R. S. Rojas, Organometallics, 2010, 29, 6104–6110.
16 R. S. Rojas, B. C. Peoples, A. R. Cabrera, M. Valderrama, R. Fröhlich,
G. Kehr, G. Erker, T. Wiegand and H. Eckert, Organometallics, 2011, 30,
6372–6382.
17 R. S. Rojas, A. R. Cabrera, B. C. Peoples, K. Spannhoff, M. Valderrama,
R. Frohlich, G. Kehr and G. Erker, Dalton Trans., 2012, 41, 1243–
1251.
18 H. Hamaki, N. Takeda, M. Nabika and N. Tokitoh, Macromolecules,
2012, 45, 1758–1769.
19 I. El-Zoghbi, E. Verguet, P. O. Oguadinma and F. Schaper, Inorg. Chem.
Commun., 2010, 13, 529–533.
20 P. L. Franceschini, M. Morstein, H. Berke and H. W. Schmalle, Inorg.
Chem., 2003, 42, 7273–7282.
21 S. Gong, H. Ma and J. Huang, J. Organomet. Chem., 2008, 693, 3509–
3518.
22 S. Gong, H. Ma and J. Huang, Dalton Trans., 2009, 8237–8247.
23 I. El-Zoghbi, S. Latreche and F. Schaper, Organometallics, 2010, 29,
1551–1559.
In a well closed vial, a toluene solution of 5 (200 μL, 10 mM,
2.0 mmol) was added to a THF solution of rac-lactide (87 mg,
0.60 mmol, 0.8 mL). Reaction mixtures were quenched at the
desired polymerization time by addition of a dichloromethane
solution of acetic acid (5 mM). Samples for kinetic investi-
gations were taken at the desired intervals and added to vials
already containing a dichloromethane solution of acetic acid
(5 mM). In both cases, volatiles were immediately evaporated
afterwards. Solid polymer samples were stored at −80 °C. Con-
1
version was determined from H NMR in CDCl3 by comparison
to remaining lactide. Pr values were determined from homo-
decoupled 1H NMR spectra. Molecular weight analyses were
performed on a Waters 1525 gel permeation chromatograph
equipped with three Phenomenex columns and a refractive index
detector at 35 °C. THF was used as the eluent at a flow rate of
1.0 mL min−1 and polystyrene standards (Sigma-Aldrich, 1.5 mg
mL−1, prepared and filtered (0.2 mm) directly prior to injection)
were used for calibration. Obtained molecular weights were cor-
rected by a Mark–Houwink factor of 0.58.50 Mass spectroscopic
analyses were performed on a MALDI TOF/TOF Ultraflextreme
mass spectrometer equipped with a SmartBeam II Nd:YAG/
355 nm laser operating at 1 kHz and providing a laser focus
down to 20 μm in diameter for the “minimum” focus setting
(Bruker Daltonics, Billerica, MA). The matrix used contained
10 mg mL−1 of 2,5-dihydroxybenzoic acid (DHB) in MeOH :
deionised water (1 : 1).
24 L. Bourget-Merle, M. F. Lappert and J. R. Severn, Chem. Rev., 2002,
102, 3031–3066.
25 I. El-Zoghbi, A. Ased, P. O. Oguadinma, E. Tchirioua and F. Schaper,
Can. J. Chem., 2010, 88, 1040–1045.
26 D. V. Vitanova, F. Hampel and K. C. Hultzsch, J. Organomet. Chem.,
2005, 690, 5182–5197.
27 D. V. Vitanova, F. Hampel and K. C. Hultzsch, Dalton Trans., 2005,
1565–1566.
28 D. V. Vitanova, F. Hampel and K. C. Hultzsch, J. Organomet. Chem.,
2011, 696, 321–330.
This journal is © The Royal Society of Chemistry 2012
Dalton Trans.

View Online
29 M. M. Melzer, S. Mossin, X. Dai, A. M. Bartell, P. Kapoor, K. Meyer
and T. H. Warren, Angew. Chem., Int. Ed., 2010, 49, 904–907.
30 N. Carrera, N. Savjani, J. Simpson, D. L. Hughes and M. Bochmann,
Dalton Trans., 2011, 40, 1016–1019.
39 J. Mosnáček, A. Duda, J. Libiszowski and S. Penczek, Macromolecules,
2005, 38, 2027–2029.
40 E. L. Whitelaw, M. G. Davidson and M. D. Jones, Chem. Commun.
(Cambridge, UK), 2011, 47, 10004–10006.
31 A. L. Zelikoff, J. Kopilov, I. Goldberg, G. W. Coates and M. Kol, Chem.
Commun. (Cambridge, UK), 2009, 6804–6806.
32 J.-C. Buffet and J. Okuda, Chem. Commun. (Cambridge, UK), 2011, 47,
4796–4798.
41 C. Romain, B. Heinrich, S. B. Laponnaz and S. Dagorne, Chem.
Commun. (Cambridge, UK), 2012, 48, 2213–2215.
42 A. D. Schwarz, A. L. Thompson and P. Mountford, Inorg. Chem., 2009,
48, 10442–10454.
33 S. L. Hancock, M. F. Mahon, G. Kociok-Kőhn and M. D. Jones,
43 L. E. Manzer, Inorg. Synth., 1982, 21, 135–136.
44 D. C. Bradley and I. M. Thomas, J. Chem. Soc., 1960, 3857.
45 J. Kim, J.-W. Hwang, Y. Kim, M. Hyung Lee, Y. Han and Y. Do, J. Orga-
nomet. Chem., 2001, 620, 1–7.
46 APEX2, Bruker AXS Inc., Madison, USA, 2006.
47 SAINT, Bruker AXS Inc., Madison, USA, 2006.
48 G. M. Sheldrick, SADABS, Bruker AXS Inc., Madison, USA, 1996 &
2004.
Eur. J. Inorg. Chem., 2011, 2011, 4596–4602.
34 A. J. Chmura, D. M. Cousins, M. G. Davidson, M. D. Jones, M. D. Lunn
and M. F. Mahon, Dalton Trans., 2008, 1437–1443.
35 T. K. Saha, V. Ramkumar and D. Chakraborty, Inorg. Chem., 2011, 50,
2720–2722.
36 T. K. Saha, B. Rajashekhar and D. Chakraborty, RSC Adv., 2012, 2, 307–
318.
37 T. K. Saha, B. Rajashekhar, R. R. Gowda, V. Ramkumar and
D. Chakraborty, Dalton Trans., 2010, 39, 5091–5093.
38 A. Kowalski, A. Duda and S. Penczek, Macromolecules, 1998, 31, 2114–
2122.
49 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystalloger., 2008,
64, 112–122.
50 M. Save, M. Schappacher and A. Soum, Macromol. Chem. Phys., 2002,
203, 889–899.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2012