New Coordination Modes for Modified Schiff Base Ti(IV) Complexes and Their Control over Lactone Ring-Opening Polymerization Activity

Article abstract of DOI:10.1021/acs.inorgchem.8b02271

A series of eight new bis(alkoxy)bis(phenoxy-imine)titanium(IV) catalysts, coordinated by Schiff base ligands derived from o-vanillin (2-hydroxy-3-methoxybenzaldehyde), show good activity and control for the ring-opening polymerization of ?μ -caprolactone

Full text of DOI:10.1021/acs.inorgchem.8b02271

Article
pubs.acs.org/IC
Cite This: Inorg. Chem. XXXX, XXX, XXX−XXX
New Coordination Modes for Modified Schiff Base Ti(IV) Complexes
and Their Control over Lactone Ring-Opening Polymerization
Activity
Christopher B. Durr† and Charlotte K. Williams*,†
†Chemistry Research Laboratory, Department of Chemistry, University of Oxford, 12 Mansfield Road, Oxford OX1 3TA, United
Kingdom
S
* Supporting Information
ABSTRACT: A series of eight new bis(alkoxy)bis(phenoxy-
imine)titanium(IV) catalysts, coordinated by Schiff base ligands
derived from o-vanillin (2-hydroxy-3-methoxybenzaldehyde),
show good activity and control for the ring-opening polymer-
ization of ε-caprolactone and ω-pentadecalactone. The new
complexes are easily prepared in two high-yield steps from
commercial reagents. The new ligands can all adopt two different
coordination modes, depending on the steric bulk on the imine: a
six-membered N−O chelate and/or a five-membered O−O
chelate. The complexes show three different structures, depend-
ing on the ligand coordination mode: type A (N−O:N−O), type
B (N−O:O−O), and type C (O−O:O−O). In all cases, the
structures were confirmed in solution using variable temperature NMR spectroscopy and in the solid state using X-ray
crystallography. The complex structure influences the polymerization rate, with the catalytic activities decreasing in the order:
type C > type B > type A for both monomers. Overall, the work demonstrates potential to use these new ligands to access
particular coordination modes, which allows enhancement of catalytic activity.
phenoxy-imine counterparts.24,25 A detailed study of the
coordination chemistry of bis(phenoxy-imine)Ti(IV) com-
plexes, reported by Johnson, Davidson, and coworkers, showed
INTRODUCTION
■
Metal complexes of Schiff base ligands are important for many
different catalytic transformations, including epoxidation,
hydrolysis, polymerization, and cross-coupling reactions.1−5
four different coordination types depending on the steric
hindrance of the imine substituent (Scheme 1).26 As its size
Catalysts featuring bidentate phenoxy-imine ligands are
increased, for example with a 2,6-iPr2(C6H3) substituent,
straightforward to synthesize and already offer a multitude of
sites for structure−activity exploration.6−13 Here, bis(Schiff
electronically disfavored modes were proposed to form as a
means to relieve steric hindrance. At the limit, with a
base)Ti(IV) catalysts are developed for lactone ring opening
2,4,6-tBu3(C6H2) substituent, dimeric complexes formed
polymerization (ROP).14 The process is a controlled polymer-
featuring a single monodentate phenoxide ligand without any
ization applicable to the production of various polyesters, some
of the target mononuclear complex, even using forcing reaction
of which are degradable and may be alternatives to polyolefins.
The polymerization is critically dependent on the catalyst
applied, not only in terms of activity but also for good control
conditions.
Although there have been many studies of bis(phenoxy-
imine) Group IV complexes, there are not yet any analogous
investigations using ligands derived from o-vanillin. The o-
vanillin derivatives are attractive as they feature an additional
ortho-methoxy substituent, adjacent to the phenolate, and as
such offer different coordination modes: either N−O or O−O
chelates should be accessible. There is precedent for complexes
and selectivity. In this field, titanium(IV) catalysts have
previously shown high activity, selectivity, and control.15−19
Additionally, titanium is earth-abundant and has a strong
precedent for deployment in industrial polymerization
catalysis. For example, bis(phenoxy-imine) Group IV com-
plexes are successful and highly active olefin polymerization
of o-vanillin derivatives of the well-known salen ligands
catalysts; their development was pioneered by Fujita and
coworkers working at Mitsui.20 Structure−activity studies
(tetradentate Schiff base ligands), and these complexes show
both chelate types.27−30 For example, these o-vanillin salen
revealed that some of the best rates are observed for cumyl
substituents at the ortho-phenolate site,10 while living polymer-
ligands formed dinuclear complexes with one metal coordi-
nated by two N−O chelates and the other by two O−O
ization of ethene or propene was observed for pentafluor-
ophenyl imine substituents.21−23 Bis(phenoxy-ether)Ti(IV)
complexes are also effective olefin polymerization catalysts,
Received: August 10, 2018
although activities are generally lower than those of their
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Scheme 1. Summary of the Coordination Types Observed for Bis(phenoxy-imine)Ti(IV) Complexes Where R1 = R2 = H, R3 =
Aryl, Alkyl (Top),26 Compared with Those Observed in This Work for o-Vanillin Derived Ligands (Bottom)
Scheme 2. Synthesis of New Ti(IV) Complexes Where Dotted Bonds Indicate the Two Coordination Modes, Depending on
a
Steric Hindrance at R
a
Reagents and conditions: (a) Ti(OiPr)4, toluene, −30 °C to RT, 24 h.
chelates.27 The different coordination modes also enable access
to heterodinuclear complexes which showed promising activity
in organo-catalysis and/or magnetic properties.28−30 In
contrast to the results for tetradentate ligands, here, bidentate
o-vanillin derived bis(phenoxy-imine) ligands form discrete
mononuclear Ti(IV) complexes but show three different
coordination types (Scheme 1, A−C).
solutions at −30 °C or by layering hexane over concentrated
THF or CHCl3 solutions. All of the complexes [(L1−8)2Ti-
(OiPr)2] are soluble in chloroform, THF, ε-caprolactone, and
toluene but only sparingly soluble in alkanes such as pentane
and hexane.
The new complexes [(L1−8)Ti(OiPr)2] were unambiguously
characterized by 1H and 13C{1H} NMR spectroscopy (Figures
S9−24), MALDI-ToF MS, elemental analysis, and single
crystal X-ray diffraction. These new complexes feature different
and independent coordination modes: either six-membered
N−O coordination and/or five-membered O−O coordination.
Thus, the complexes feature three different coordination
modes: type A with N−O:N−O coordination, type B with
N−O:O−O coordination, and type C with O−O:O−O
coordination (Scheme 1). For any single type, there are also
additional isomers that are theoretically possible, although the
eight catalysts synthesized in this series exhibit only four
distinct structures (in both solid and solution states), with type
A showing two different isomers. All of the new complexes
were characterized in the solid state by single crystal X-ray
diffraction experiments, and representative structures are
illustrated in Figure 1. Complexes [(L1)2Ti(OiPr)2] (Figure
1-I) and [(L2)2Ti(OiPr)2] (Figure S35) each adopt type A-I
coordination, where the imine nitrogen atoms are located at
positions cis to one another. Both complexes have imine
substituents with rather low steric hindrance (L1 = C6F5 and L2
= cyclohexyl) and adopt the lowest energy coordination mode
common to many other bis(phenoxy-imine)Ti(IV) com-
plexes.26 In the solid state structure of [(L1)Ti(OiPr)2], this
RESULTS AND DISCUSSION
■
Complex Synthesis. The o-vanillin derived Schiff base
ligands, HL1−HL8, were synthesized via a one-pot reaction
between o-vanillin and the desired primary amine in refluxing
acidified ethanol (see Supporting Information for experimental
1
details and Figures S1−8 for the H NMR spectra). The
amines were selected to enable understanding of the influence
of steric hindrance of the imine substituents over the resulting
coordination chemistry. The crude products were purified
through either crystallization from ethanol or by washing with
pentane (see Supporting Information for further details).
Yields, after purification, were generally excellent; the final
products were soluble in chloroform, THF and, with the
exception of HL8, toluene. The Ti(IV) complexes were
prepared by the reaction of two equivalents of HL1−8 with
an equivalent of Ti(OiPr)4; the reaction was conducted in dry
toluene at low temperature (−30 °C to room temperature)
and over 18−24 h (Scheme 2). After removal of the solvents,
the complexes [(L1−8)2Ti(OiPr)2] were isolated, as yellow or
orange powders, in fair to excellent yields (38−99%). All of the
complexes were crystallized either from concentrated THF
B
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 1. ORTEP representations of the structures of complexes [(L1)2Ti(OiPr)2] (I), [(L3)2Ti(OiPr)2] (II), [(L4)2Ti(OiPr)2] (III), and
[(L6)2Ti(OiPr)2] (IV). Ellipsoids drawn at 50% probability; hydrogen atoms and disorder are omitted for clarity. Green = titanium, blue =
nitrogen, red = oxygen, gray = carbon, lime = fluorine.
mode may be further enhanced by π-stacking interactions
between one set of the electron deficient C6F5 and the adjacent
phenyl ring. The average aromatic ring separation is 3.58 Å,
which is within range of those typically found in the
Cambridge Structural Database31 (CSD) for π−π stacking
between similar rings (Figure S34).
imine bond is 0.02 Å shorter than that for coordinated
analogues.
Complexes [(L6)2Ti(OiPr)2], [(L7)2Ti(OiPr)2], and
[(L8)2Ti(OiPr)2] all adopt type C coordination, where both
ligands show O−O chelation (Figures 1-IV, S37, and S38,
respectively). [(L6)2Ti(OiPr)2] has bond angles of [O(1)−Ti−
O(2)] = 73.67(5)° and [O(3)−Ti−O(4)] = 73.99(5)°; both
values are similar to those found in [(L4)2Ti(OiPr)2]. The two
OiPr moieties are arranged trans to the neutral OMe groups;
both Ti−OiPr distances are shorter than the analogous bonds
in type A and B complexes (Table S3). Both imine bonds are
ca. 1.27 Å and, as expected, are slightly shorter than
coordinated analogues. Interestingly, Johnson and coworkers
reported that an adamantyl imine substituted complex as the
least sterically congested type I coordination.26 In contrast,
here, when the adamantyl imine substituted complex adopts a
type C structure, it stabilizes the most sterically congested
ligands. Analysis of the CSD suggests that type B and C
coordination modes are the first examples of their kind
regardless of the metal center. While O−O chelation has been
previously observed, in general O−O chelated phenoxy-imines
required protonation of either the alkoxide or the imine
nitrogen, thus forming a zwitterionic or neutral ligand, rather
than the anionic ligands seen here.32
[(L3)2Ti(OiPr)2] (Figure 1-II) adopts a different coordina-
tion mode, type A-II, where the imine nitrogen atoms are
disposed in a mutually trans geometry. This coordination type
also correlates with that seen previously for bis(phenoxy-
imine)Ti(IV) complexes, featuring imine substituents of
intermediate steric hindrance. Because the coordination
geometry is not electronically favored, it is presumably adopted
to minimize steric interactions while still maintaining the N−
O:N−O chelates. In this type A-II mode, both the Ti−N bond
distances are shorter, and the Ti−O distances longer,
compared to equivalent bonds in type A-I. Complexes
[(L4)2Ti(OiPr)2] (Figure 1-III) and [(L5)2Ti(OiPr)2] (Figure
S36) both adopt type B coordination, with one nitrogen
located trans to an OiPr ligand and the other ligand featuring
O−O coordination with the nitrogen atom not coordinating to
Ti. Due to the formation of a five-membered ring, the O(1)−
Ti−O(2) angle is more acute at 72.92(8)° than the O(3)−Ti−
N(2) angle at 80.72(9)° with the latter value being similar to
that seen in [(L2)2Ti(OiPr)2]. Additionally, both Ti−OiPr
distances are slightly shorter than that in type A, with the
difference in bond lengths being ∼0.05 Å, and the unbound
To investigate whether the solid state structures are
maintained in solution, the complexes were characterized
using 1H NMR spectroscopy. Complex [(L1)2Ti(OiPr)2], type
C
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
1
Figure 2. Variable temperature H NMR spectra of selected regions for [(L4)2Ti(OiPr)2] in d8-THF (500 MHz).
Table 1. Ring-Opening Polymerization Data of CL from [(L1‑8)2Ti(OiPr)2] in Toluene Solution
a
c
d
e
entry
catalyst
time (h)
CL conv (%)
Mn (exp)
Mn (calcd)
Đ
coordination type
1
2
3
4
5
6
7
8
[(L1)2Ti(OiPr)2]
[(L2)2Ti(OiPr)2]
[(L3)2Ti(OiPr)2]
[(L4)2Ti(OiPr)2]
[(L5)2Ti(OiPr)2]
[(L6)2Ti(OiPr)2]
[(L7)2Ti(OiPr)2]
[(L8)2Ti(OiPr)2]
[(L4)2Ti(OiPr)2]
[(L7)2Ti(OiPr)2]
[Ti(OiPr)4]
24
24
9
4
4
2
2
2
99
99
90
99
73
99
99
99
90
90
95
9460
8780
8890
13 100
10 900
16 100
13 100
12 200
12 100
13 600
11 300
11 300
10 300
11 300
8330
11 300
11 300
11 300
10 300
10 300
1.13
1.20
1.27
1.27
1.10
1.41
1.09
1.44
1.63
1.46
2.02
A-I
A-I
A-II
B
B
C
C
C
B
C
b
9
10
11
0.10
0.12
0.05
b
f
19 800
a
b
c
Polymerization conditions: [cat]:[CL] = 1:200, [CL] = 1 M, toluene, 80 °C. [cat]:[CL] = 1:200, unpurified CL, neat, 140 °C, in air. Calculated
d
from the 1H NMR spectrum of the crude polymer by comparison of the integrals assigned to CL (4.20 ppm) and PCL (4.07 ppm); Measured by
GPC in THF and calibrated using PS standards; Mn values are reported after correction (multiplication by 0.56).33
M
n(calcd) = (conversion/100) ×
e
f
loading/[2] × RMM(CL) From ref 36. Conditions: [cat]:[CL] = 1:212, CL, neat, 100 °C, in air.
1
A-I in the solid state, shows resonances at lower chemical shifts
compared to the free ligand. In particular, upon coordination,
the imine CH resonance is observed at lower chemical shift
(8.23 ppm) and broadens significantly (Figure S25). The peak
broadening is resolved at low temperature (−80 °C), and a
single set of well-defined peaks are observed (Figure S28).
[(L4)2Ti(OiPr)2], type B, shows a single, broadened CH imine
peak (8.55 ppm) at higher chemical shift compared to the free
ligand (8.34 ppm) (Figure S26). The broadening is attributed
to rapid equilibration between two asymmetrically bound
ligands, vide infra. Indeed, previous studies of bis(phenoxy-
imine)Ti(IV) complexes showed a similar resolution of
fluxional processes at low temperatures.26 The imine
substituent signals (2,6-di-isopropyl-phenyl) show broadened
isopropyl resonances (3.0 ppm), which suggests restricted
rotation in solution. Complexes [(L6−L8)2Ti(OiPr)2], type C,
all display the same general features in their H NMR spectra.
In each case, the imine CH resonance is at higher chemical
shift compared to the free ligand, but the OMe resonance is at
lower chemical shift (Figure S27). Over the series, the imine
proton chemical shift is indicative of the coordination type,
with values decreasing in the order: type C > type B > type A.
VT NMR spectroscopy was used to characterize complexes
[(L1)2Ti(OiPr)2] (type A-I), [(L4)2Ti(OiPr)2] (type B), and
[(L7)2Ti(OiPr)2] (type C). Complexes of types A-I and C,
both C2 symmetric, showed little change to the spectra upon
cooling (Figures S28 and S30). In contrast, the spectrum for
structure type B underwent significant changes on cooling
(Figure 2). At room temperature, a single broadened set of
signals was observed, but at −80 °C, two imine signals were
observed at ∼8.75 and 8.3 ppm, integrating to one proton
each. These peaks are assigned to the different imine
D
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
resonances for the N−O and O−O chelates. The assignment is
supported by a similar chemical shift for the N−O imine
resonance to that observed for type A complexes, ∼8.3 ppm,
and by the value for the O−O chelate being close to those
observed for type C complexes, ∼8.7 ppm. Furthermore, the
reduced complex symmetry at low temperature results in a
significantly more complex aromatic region as well as the
splitting of the methoxy and isopropyl peaks. Overall, the VT-
NMR data indicate that type B coordination is retained in
solution, and at room temperature, averaged signals arise from
dynamic ligand exchange. It should also be noted that at low
temperature, several additional low intensity signals are
observed throughout the spectrum, and these may be
attributed either to restricted N−Ar bond rotation, residual
equilibration, or to minor populations of other structures (e.g.,
types A-II or C) (Figure S29).
The structure of complex [(L4)2Ti(OiPr)2] was also
investigated at higher temperatures by heating in d2-1,1,2,2-
tetrachloroethane (TCE) from room temperature to 100 °C.
Although the peaks sharpened slightly there was not any
significant change in the chemical shifts (Figure S31).
Additionally, the spectrum remained unchanged both after
being held at 100 °C for 24 h in TCE (Figure S32) or after
being kept at 70 °C for 5 h in d8-THF (Figure S33). The
temperature and solvent stability assessment is important
because polymerizations are conducted at elevated temper-
atures.
Polymerization Catalysis. The series of complexes were
tested as initiators for the ring-opening polymerization of ε-
caprolactone (CL). Polymerizations were conducted both in
toluene solution (1:200 [cat]:[CL], 1 M [CL]; Table 1) and
using neat conditions (1:200 [cat]:[CL] or 1:1000 [cat]:[CL];
Table S1). Catalysts [(L1)2Ti(OiPr)2] and [(L2)2Ti(OiPr)2]
were both able reach high conversions under either set of
conditions but are comparably slow (taking 24 h to reach
complete conversion in toluene). Catalysts [(L4−8)2Ti(OiPr)2]
are also active and, in some cases, are several times faster than
[(L1,2)2Ti(OiPr)2]. For example, [(L4)2Ti(OiPr)2] resulted in
complete conversion within 4 h, while [(L6−8)2Ti(OiPr)2]
reached full conversion in just 2 h.
All of the initiators produced PCL showing a monomodal
molar mass distribution and with relatively narrow dispersity
values. The experimentally determined molar mass values
(GPC) are in good agreement with values calculated on the
basis of two initiating sites per complex (i.e., both isopropoxide
ligands initiate). The complexes also enable controlled
polymerization as evidenced by a linear evolution of molar
mass vs conversion (Figures S39−41).
On the basis of the point kinetic measurements, the activity
order is [(L8)2Ti(OiPr)2] ∼ [(L7)2Ti(OiPr)2] ∼ [(L6)2Ti-
(OiPr)2] > [(L5)2Ti(OiPr)2] ∼ [(L4)2Ti(OiPr)2] > [(L3)2Ti-
(OiPr)2] > [(L2)2Ti(OiPr)2] ∼ [(L1)2Ti(OiPr)2] or, in terms
of coordination chemistry, type C > type B > type A-II > type
A-I. The rates are generally slower than the homoleptic
Ti(OiPr)4,36 although comparisons are complicated by differ-
ent reaction conditions (temperature) and uncertainties
regarding number of initiating groups.
Figure 3. Plots of ln(Mo/Mt) vs time for three different catalysts, each
selected as an example of a particular coordination type. Polymer-
ization conditions: [cat]:[CL] = 1:200, [CL] = 1.0 M, toluene, 80 °C.
dispersity throughout the reaction (Figures S42−44). All three
catalysts showed rates that were first order in monomer
concentration, and the reactivity trends are retained with type
C > type B > type A. Specifically, type C is twice as fast as type
B, which is three times faster than type A-II. A sample of low
molar mass PCL from each polymerization was analyzed by
MALDI-ToF and showed only isopropoxide chain end groups
(Figures S45−47). Thus, all catalysts enchain by a coordina-
tion−insertion mechanism, and the particular coordination
chemistry significantly influences the polymerization rate.
Polymerizations were also conducted in neat CL, and the
reactivity trends were similar to those obtained in toluene,
however, under these conditions the increased viscosity results
in diffusion limitations. As such, the molar mass values are
higher than expected, particularly at higher conversions.
Selected catalysts were also tested under conditions more
analogous to those used industrially, i.e. using neat CL, at
elevated temperature and using lower catalyst loading (1:1000,
[cat]:[CL]): all catalysts maintained activity and control
(Table S1). One advantage of these titanium catalysts may
be an improved resistance to decomposition in air, particularly
compared to other organometallic initiators. Indeed, the high
stability of other bis(phenoxy-imine) Ti(IV) alkoxide com-
plexes toward hydrolysis was already described.34,35 To test the
resilience of the new catalysts, polymerizations were conducted
using unpurified CL preheated to 140 °C for 15 min to remove
residual water.36 The catalyst, dissolved in minimal toluene,
was added, and the polymerization vessel was open to the
atmosphere. Complete monomer conversion was achieved
within minutes and produced PCL with predictable molar
mass and narrow dispersity (Table 1, entries 9 and 10).
PCL is applied in the preparation of new medical devices
and potentially for future applications in elastomer, coating,
and sealant areas.14,37 In contrast, macrolactone ROP provides
access to polyesters with thermal properties approaching those
of some grades of polyethene.38−43 As such, these long-chain
aliphatic polyesters may be substitutes for pervasive petro-
polymers, and an attractive method to prepare them is by the
ring-opening polymerization of macrolactones.38,39,44−46 So
far, the range of catalysts investigated for macrolactone ROP is
more limited, and activities tend to be significantly lower than
equivalent reactions with CL.44 Thus, it was of interest to
To gain insight into the influence of the coordination type,
polymerization kinetic evaluations were undertaken using
selected examples of each structure, i.e. complexes [(L3)2Ti-
(iOPr)2], [(L4)2Ti(OiPr)2], and [(L8)2Ti(OiPr)2] (Figure 3).
All polymerizations showed good control with molar mass
values increasing linearly with conversion and showing narrow
E
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 2. PDL Ring-Opening Polymerization Using Catalysts [(L1‑8)2Ti(OiPr)2]
a
b
c
d
entry
catalyst
time (h)
conv (%)
Mn (exp)
Mn (calcd)
Đ
coordination type
1
2
3
4
5
6
7
[(L1)2Ti(OiPr)2]
[(L3)2Ti(OiPr)2]
[(L4)2Ti(OiPr)2]
[(L5)2Ti(OiPr)2]
[(L6)2Ti(OiPr)2]
[(L7)2Ti(OiPr)2]
[(L8)2Ti(OiPr)2]
24
24
48
24
24
24
72
5
A-I
A-II
B
B
C
27
72
32
96
47
7680
21 700
3430
36 500
14 800
6960
3250
8650
3850
11 500
5650
11 400
1.49
1.79
1.89
1.69
1.70
2.15
C
C
95
a
b
[cat]:[PDL] = 1:100, [PDL] = 1 M in toluene, 100 °C, N2. Calculated from the 1H NMR spectrum of the crude polymer by comparison of the
c
integrals assigned to PDL (4.14 ppm) and PPDL (4.07 ppm). Measured by GPC, in CHCl3 and calibrated using PS standards and uncorrected.
d
Mn(calcd) = (conversion/100) x loading/[2] x RMM(PDL)
Scheme 3. Proposed Dissociative Coordination−Insertion Polymerization Pathway for New Ti(IV) Catalysts. Where R = iso-
a
propyl (initiator) and polymer chain (propagation); m = 2 (CL) or 11 (PDL)
a
Where R = iso-propyl (initiator) and polymer chain (propagation).
evaluate these Ti(IV) catalysts for the ring opening polymer-
ization of ω-pentadecalactone (PDL). Polymerizations were
conducted either in toluene or in the melt at 100 °C (Tables 2
and S2). In all cases, the activities were lower than in CL ROP,
but the overall activity trend remained the same: type C > type
B > type A-II > type A-I.
The resulting polyesters showed molar mass values higher
than expected, although it should be noted that the molar mass
values determined by GPC are calibrated with polystyrene and
are uncorrected. Previous studies into PDL ROP also yielded
polyesters showing higher molar mass values than expected,
and this was rationalized by limited catalyst solubility under
the polymerization conditions.46 In contrast, the Ti(IV)
catalysts remain entirely soluble both in toluene and in the
PDL melt. Catalyst deactivation is also unlikely because a
polymerization in the melt using freshly distilled PDL gave
similarly high molar mass values to unpurified PDL (Table S2,
entries 5 and 6). Furthermore, these catalysts are more tolerant
to low levels of water, even at elevated temperatures, than
other organometallic ROP catalysts. Thus, the discrepancy
between experimental and calculated molar masses appears
most likely to result from GPC calibration methods, with
future work necessary to determine the Mark−Houwink
parameters for PPDL.
Following the successful ROP of both CL and PDL,
terpolymerization reactions were investigated as a means to
form poly(CL-co-PDL). Because the rates of CL ROP are
much faster than those of PDL ROP, a gradient or block
copolymer might be expected. Nonetheless, previous inves-
tigations into terpolymerizations using these monomers
resulted in the formation of random copolymers.44
A
terpolymerization reaction using [(L6)2Ti(OiPr)2] was con-
ducted in toluene at 100 °C (1:100:100 [cat]/[CL]/[PDL],
[lactones] = 1 M). The polymerization was monitored by
aliquots removed after 2.5 and 24 h of reaction. After 2.5 h,
64% conversion was achieved, and the polymer was mostly
comprised of PCL, as evident from the relative intensity of the
α-methylene junction units in 13C{1H} NMR spectrum (65−
F
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
63 ppm, Figure S50). After 24 h, all monomers were fully
converted, and a random polymer was indicated. The
copolymer was analyzed using DSC and showed a melting
point at 74.6 °C (Figure S51), which was slightly higher than
those in previous reports perhaps due to molar mass
differences.44
substituents, influences the coordination chemistry in the
complexes. The complexes adopt three major coordination
modes, each featuring N−O and/or O−O chelation. These
different coordination modes influence the catalytic activity in
lactone ring-opening polymerizations with the most active
catalysts showing O−O chelation by both ligands. It is
proposed that increased steric hindrance at the imine nitrogen
substituent favors formation of five-membered O−O chelation,
which in turn affects the lability of the ligand around the
Ti(IV) active site, thereby resulting in faster rates. Overall, this
work highlights the importance and relevance of ligand
coordination modes as a viable means to control polymer-
ization activity. In general, other Group IV catalysts show
outstanding performance in catalysis, and exploration of these
catalysts for other processes is warranted, in particular for
alkene polymerization catalysis. Given the ease of o-vanillin
ligand and complex preparation, and the means to qualify
coordination mode, studies into coordination chemistry at
metal centers are recommended. In particular, using these
ligands to prepare complexes with other transition metals and
main group elements, particularly those in higher oxidation
states M(III/IV), should provide access to new ring-opening
polymerization catalysts and allow better understanding of how
to exploit coordination geometry to enhance rates.
Polymerization Mechanism. The catalysts’ reactivity
trend (type C > type B > type A) appears to correlate with
reduced steric hindrance at the Ti(IV) active site. As steric
bulk of the substituents attached to the imine increases, so
complexes tend to adopt structures with O−O chelation, thus
placing the sterically hindered substituents further from the
active site. In an attempt to qualify this notion, Solid-g
analysis47,48 of the crystal structures of [(L1)2Ti(OiPr)2],
[(L4)2Ti(OiPr)2], and [(L6)2Ti(OiPr)2] was undertaken but
regardless of the specific coordination mode, the Ti(IV) center
is always 96−98% shielded by ligands. While this analysis does
not rule out the importance of steric hindrance, particularly
surrounding the growing polymer chain, it does indicate that
other factors warrant consideration. In terms of the polymer-
ization mechanism, lactone ROP generally occurs by the well-
known coordination−insertion process; here, polymer chain
end-group analysis supports this pathway. An interesting
consideration is the influence the ancillary ligands may exert
over monomer coordination. Previously, Chen and coworkers
invoked mechanisms resulting in formation of seven-coordi-
nated lactone intermediates in ROP using bis(alkoxide)bis-
(phenoxy-imine)Ti(IV) catalysts.49 The proposal was based on
structure−activity results, but there was no direct evidence for
the associated seven-coordinate intermediate. In contrast, we
propose a dissociative-type coordination insertion mechanism,
whereby one of the neutral donor atoms from the Schiff base
ligand is decoordinated from the metal to form a vacant site
prior to monomer coordination (Scheme 3 for an illustration
of the pathway). Considering the structure−activity data,
higher rates are observed for ligands adopting the more labile
five membered O−O chelates, which show weaker Ti−O
bonds. Accordingly, lower rates are observed from ligands
adopting the six membered N−O chelates, which have
stronger Ti−N bonds. Thus, the most active complexes may
show greater ligand dissociation (Ti−O bond cleavage), which
accelerates monomer coordination and formation of the
activated complex. Although there is no direct spectroscopic
evidence for such ligand hemilability, it could be occurring
more rapidly than NMR time scales and/or accelerated in the
presence of coordinating monomer. Additionally, the polymer
molar mass studies indicate that both alkoxide groups initiate
and that two polymer chains grow per Ti(IV) center. Such a
ligand dissociation pathway would be much more likely to
accommodate the steric demands of two propagating sites with
associated dissociation/coordination and insertion reactions,
compared with higher coordinate associative type intermedi-
ates. We wish to emphasize that prior to a complete kinetic
evaluation, it would be premature to speculate further upon the
intimate details of chain growth. Rather, the overall structure−
activity data demonstrate the potential to exploit ligand
dissociation and coordination modes to enhance catalytic
activity.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.8b02271.
Complete experimental details, full 1H and 13C{1H}
NMR assignments of [(L1−8)2Ti(OiPr)2], polymer
MALDI-ToF MS data, and representative GPC
chromatograms (PDF)
Accession Codes
CCDC 1856548−1856555 contain the supplementary crys-
tallographic data for this paper. These data can be obtained
free of charge via www.ccdc.cam.ac.uk/data_request/cif, by
emailing data_request@ccdc.cam.ac.uk, or by contacting The
Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
AUTHOR INFORMATION
Corresponding Author
*E-mail: Charlotte.williams@chem.ox.ac.uk.
■
ORCID
Christopher B. Durr: 0000-0001-7569-6856
Charlotte K. Williams: 0000-0002-0734-1575
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
The authors acknowledge Siam Cement Group (SCG)
Chemicals Co., Ltd. for funding.
■
REFERENCES
■
(1) Decortes, A.; Castilla, A. M.; Kleij, A. W. Salen-Complex-
Mediated Formation of Cyclic Carbonates by Cycloaddition of CO2
to Epoxides. Angew. Chem., Int. Ed. 2010, 49 (51), 9822−9837.
(2) Cozzi, P. G. Metal−Salen Schiff base complexes in catalysis:
practical aspects. Chem. Soc. Rev. 2004, 33 (7), 410−421.
CONCLUSION
■
Eight new o-vanillin derived phenoxy-imine ligands and
bis(phenoxy-imine)titanium(IV) complexes were prepared.
The steric bulk of the ligand, in particular the imine nitrogen
G
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(3) Darensbourg, D. J. Making Plastics from Carbon Dioxide: Salen
Metal Complexes as Catalysts for the Production of Polycarbonates
from Epoxides and CO2. Chem. Rev. 2007, 107 (6), 2388−2410.
(4) Gupta, K. C.; Sutar, A. K. Catalytic activities of Schiff base
transition metal complexes. Coord. Chem. Rev. 2008, 252 (12), 1420−
1450.
(5) Jacobsen, E. N. Asymmetric Catalysis of Epoxide Ring-Opening
Reactions. Acc. Chem. Res. 2000, 33 (6), 421−431.
(6) Jones, D. J.; Gibson, V. C.; Green, S. M.; Maddox, P. J.; White,
A. J. P.; Williams, D. J. Discovery and optimization of new chromium
catalysts for ethylene oligomerization and polymerization aided by
high-throughput screening. J. Am. Chem. Soc. 2005, 127 (31), 11037−
11046.
(7) Kim, I.; Kwak, C. H.; Kim, J. S.; Ha, C. S. Ethylene
oligomerizations to low-carbon linear alpha-olefins by structure
modulated phenoxy-imine nickel(II) complexes combined with
aluminum sesquichloride. Appl. Catal., A 2005, 287 (1), 98−107.
(8) Kinoshita, S.; Kawamura, K.; Fujita, T. Early-Transition-Metal
Catalysts with Phenoxy−Imine-Type Ligands for the Oligomerization
of Ethylene. Chem. - Asian J. 2011, 6 (2), 284−290.
(9) Kirillov, E.; Roisnel, T.; Razavi, A.; Carpentier, J. F.
Chromium(III) Complexes of Sterically Crowded Bidentante
{ONR} and Tridentate {ONNR} Naphthoxy-Imine Ligands:
Syntheses, Structures, and Use in Ethylene Oligomerization. Organo-
metallics 2009, 28 (8), 2401−2409.
(10) Matsui, S.; Mitani, M.; Saito, J.; Tohi, Y.; Makio, H.;
Matsukawa, N.; Takagi, Y.; Tsuru, K.; Nitabaru, M.; Nakano, T.;
Tanaka, H.; Kashiwa, N.; Fujita, T. A Family of Zirconium Complexes
Having Two Phenoxy−Imine Chelate Ligands for Olefin Polymer-
ization. J. Am. Chem. Soc. 2001, 123, 6847−6856.
(11) Parssinen, A.; Luhtanen, T.; Klinga, M.; Pakkanen, T.; Leskela,
M.; Repo, T. Bis(salicylaldiminato)titanium complexes containing
bulky imine substituents: Synthesis, characterization and ethene
polymerization studies. Eur. J. Inorg. Chem. 2005, 2005 (11), 2100−
2109.
(12) Parssinen, A.; Luhtanen, T.; Klinga, M.; Pakkanen, T.; Leskela,
M.; Repo, T. Alkylphenyl-substituted bis(salicylaldiminato) titanium
catalysts in ethene polymerization. Organometallics 2007, 26 (15),
3690−3698.
(13) Suzuki, Y.; Tanaka, H.; Oshiki, T.; Takai, K.; Fujita, T.
Titanium and Zirconium Complexes with Non-Salicylaldimine-Type
Imine−Phenoxy Chelate Ligands: Syntheses, Structures, and Ethyl-
ene-Polymerization Behavior. Chem. - Asian J. 2006, 1 (6), 878−887.
(14) Labet, M.; Thielemans, W. Synthesis of polycaprolactone: a
review. Chem. Soc. Rev. 2009, 38 (12), 3484−3504.
(15) Chmura, A. J.; Davidson, M. G.; Jones, M. D.; Lunn, M. D.;
Mahon, M. F.; Johnson, A. F.; Khunkamchoo, P.; Roberts, S. L.;
Wong, S. S. F. Group 4 Complexes with Aminebisphenolate Ligands
and Their Application for the Ring Opening Polymerization of Cyclic
Esters. Macromolecules 2006, 39 (21), 7250−7257.
(16) Gendler, S.; Segal, S.; Goldberg, I.; Goldschmidt, Z.; Kol, M.
Titanium and Zirconium Complexes of Dianionic and Trianionic
Amine−Phenolate-Type Ligands in Catalysis of Lactide Polymer-
ization. Inorg. Chem. 2006, 45 (12), 4783−4790.
(17) Kim, Y.; Jnaneshwara, G. K.; Verkade, J. G. Titanium Alkoxides
as Initiators for the Controlled Polymerization of Lactide. Inorg.
Chem. 2003, 42 (5), 1437−1447.
(21) Mitani, M.; Furuyama, R.; Mohri, J.-i.; Saito, J.; Ishii, S.; Terao,
H.; Nakano, T.; Tanaka, H.; Fujita, T. Syndiospecific Living
Propylene Polymerization Catalyzed by Titanium Complexes Having
Fluorine-Containing Phenoxy−Imine Chelate Ligands. J. Am. Chem.
Soc. 2003, 125 (14), 4293−4305.
(22) Junji, S.; Makoto, M.; Jun-ichi, M.; Yasunori, Y.; Shigekazu, M.;
Sei-ichi, I.; Shin-ichi, K.; Norio, K.; Terunori, F. Living Polymer-
ization of Ethylene with a Titanium Complex Containing Two
Phenoxy-Imine Chelate Ligands. Angew. Chem., Int. Ed. 2001, 40
(15), 2918−2920.
(23) Tian, J.; Hustad, P. D.; Coates, G. W. A New Catalyst for
Highly Syndiospecific Living Olefin Polymerization: Homopolymers
and Block Copolymers from Ethylene and Propylene. J. Am. Chem.
Soc. 2001, 123 (21), 5134−5135.
(24) Suzuki, Y.; Inoue, Y.; Tanaka, H.; Fujita, T. Phenoxy-ether
ligated Ti complexes for the polymerization of ethylene. Macromol.
Rapid Commun. 2004, 25 (3), 493−497.
(25) Lamberti, M.; Mazzeo, M.; Pellecchia, C. Group 4 bis(chelate)
metal complexes of monoanionic bidentate [E,O−] ligands (E = O,
S): synthesis and application as α-olefin polymerization catalysts.
Dalton Trans. 2009, 41, 8831−8837.
(26) Johnson, A. L.; Davidson, M. G.; Lunn, M. D.; Mahon, M. F.
Synthesis, Isolation and Structural Investigation of Schiff-Base
Alkoxytitanium Complexes: Steric Limitations of Ligand Coordina-
tion. Eur. J. Inorg. Chem. 2006, 2006, 3088−3098.
(27) Thevenon, A.; Garden, J. A.; White, A. J.; Williams, C. K.
Dinuclear Zinc Salen Catalysts for the Ring Opening Copolymeriza-
tion of Epoxides and Carbon Dioxide or Anhydrides. Inorg. Chem.
2015, 54 (24), 11906−15.
(28) Xie, Q.-W.; Wu, S.-Q.; Liu, C.-M.; Cui, A.-L.; Kou, H.-Z.
Synthesis, crystal structure and magnetic properties of dinuclear
NiIILnIII complexes based on a flexible polydentate ligand. Dalton
Trans. 2013, 42 (31), 11227−11233.
(29) Xie, Q.-W.; Wu, S.-Q.; Shi, W.-B.; Liu, C.-M.; Cui, A.-L.; Kou,
H.-Z. Heterodinuclear MII−LnIII single molecule magnets con-
structed from exchange-coupled single ion magnets. Dalton Trans.
2014, 43 (29), 11309−11316.
(30) Matsunaga, S.; Shibasaki, M. Recent advances in cooperative
bimetallic asymmetric catalysis: dinuclear Schiff base complexes.
Chem. Commun. 2014, 50 (9), 1044−1057.
(31) Allen, F. The Cambridge Structural Database: a quarter of a
million crystal structures and rising. Acta Crystallogr., Sect. B: Struct.
Sci. 2002, 58 (3), 380−388.
(32) Upadhyay, A.; Das, C.; Shanmugam, M.; Langley, S. K.;
Murray, K. S.; Shanmugam, M. Electronic and Magnetic Properties of
a Gadolinium(III) Schiff Base Complex. Eur. J. Inorg. Chem. 2014,
2014 (26), 4320−4325.
(33) Save, M.; Schappacher, M.; Soum, A. Controlled Ring-Opening
Polymerization of Lactones and Lactides Initiated by Lanthanum
Isopropoxide, 1. General Aspects and Kinetics. Macromol. Chem. Phys.
2002, 203 (5−6), 889−899.
(34) Peri, D.; Meker, S.; Manna, C. M.; Tshuva, E. Y. Different
ortho and para Electronic Effects on Hydrolysis and Cytotoxicity of
Diamino Bis(Phenolato) “Salan” Ti(IV) Complexes. Inorg. Chem.
2011, 50 (3), 1030−1038.
(35) Tzubery, A.; Tshuva, E. Y. Cytotoxic Titanium(IV) Complexes
of Salalen-Based Ligands. Eur. J. Inorg. Chem. 2017, 2017 (12), 1695−
1705.
(18) Bakewell, C.; Fateh-Iravani, G.; Beh, D. W.; Myers, D.;
Tabthong, S.; Hormnirun, P.; White, A. J. P.; Long, N.; Williams, C.
K. Comparing a series of 8-quinolinolato complexes of aluminium,
titanium and zinc as initiators for the ring-opening polymerization of
rac-lactide. Dalton Trans. 2015, 44 (27), 12326−12337.
(19) Saha, T. K.; Mandal, M.; Chakraborty, D.; Ramkumar, V. Imino
phenoxide complexes of group 4 metals: synthesis, structural
characterization and polymerization studies. New J. Chem. 2013, 37
(4), 949−960.
(36) Parssinen, A.; Kohlmayr, M.; Leskela, M.; Lahcini, M.; Repo, T.
Catalytic polymerization of epsilon-caprolactone in air. Polym. Chem.
2010, 1 (6), 834−836.
(37) Ikada, Y.; Tsuji, H. Biodegradable polyesters for medical and
ecological applications. Macromol. Rapid Commun. 2000, 21 (3),
117−132.
(38) Myers, D.; Witt, T.; Cyriac, A.; Bown, M.; Mecking, S.;
Williams, C. K. Ring opening polymerization of macrolactones: high
conversions and activities using an yttrium catalyst. Polym. Chem.
2017, 8 (37), 5780−5785.
(20) Fujita, T.; Kawai, K. FI Catalysts for Olefin Oligomerization
and Polymerization: Production of Useful Olefin-Based Materials by
Unique Catalysis. Top. Catal. 2014, 57 (10−13), 852−877.
H
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(39) Pepels, M. P. F.; Hansen, M. R.; Goossens, H.; Duchateau, R.
From Polyethylene to Polyester: Influence of Ester Groups on the
Physical Properties. Macromolecules 2013, 46 (19), 7668−7677.
(40) Stempfle, F.; Ortmann, P.; Mecking, S. Long-Chain Aliphatic
Polymers To Bridge the Gap between Semicrystalline Polyolefins and
Traditional Polycondensates. Chem. Rev. 2016, 116 (7), 4597−4641.
(41) Stempfle, F.; Ritter, B. S.; Mulhaupt, R.; Mecking, S. Long-
̈
chain aliphatic polyesters from plant oils for injection molding, film
extrusion and electrospinning. Green Chem. 2014, 16 (4), 2008−2014.
̈
(42) Trzaskowski, J.; Quinzler, D.; Bahrle, C.; Mecking, S. Aliphatic
Long-Chain C20 Polyesters from Olefin Metathesis. Macromol. Rapid
Commun. 2011, 32 (17), 1352−1356.
(43) Vilela, C.; Silvestre, A. J. D.; Meier, M. A. R. Plant Oil-Based
Long-Chain C26 Monomers and Their Polymers. Macromol. Chem.
Phys. 2012, 213 (21), 2220−2227.
(44) Bouyahyi, M.; Duchateau, R. Metal-Based Catalysts for
Controlled Ring-Opening Polymerization of Macrolactones: High
Molecular Weight and Well-Defined Copolymer Architectures.
Macromolecules 2014, 47, 517−524.
(45) Pepels, M. P. F.; Hermsen, I.; Noordzij, G. J.; Duchateau, R.
Molecular Structure−Catalytic Activity Relationship in the Ring-
Opening Polymerization of (Macro)lactones. Macromolecules 2016,
49 (3), 796−806.
(46) van der Meulen, I.; Gubbels, E.; Huijser, S.; Sablong, R.;
Koning, C. E.; Heise, A.; Duchateau, R. Catalytic Ring-Opening
Polymerization of Renewable Macrolactones to High Molecular
Weight Polyethylene-like Polymers. Macromolecules 2011, 44 (11),
4301−4305.
(47) Guzei, I. A.; Wendt, M. An improved method for the
computation of ligand steric effects based on solid angles. Dalton
Trans. 2006, No. 33, 3991−3999.
(48) Guzei, I. A. W. M. Program Solid-G; UW-Madison: Wisconsin,
2004.
(49) Chen, H.-Y.; Lu, W.-Y.; Chen, Y.-J.; Hsu, S. C. N.; Ou, S.-W.;
Peng, W.-T.; Jheng, N.-Y.; Lai, Y.-C.; Wu, B.-S.; Chung, H.; Chen, Y.;
Huang, T.-C. Synthesis, characterization, and catalytic activity of
titanium iminophenoxide complexes in relation to the ring-opening
polymerization of L-lactide and ε-caprolactone. J. Polym. Sci., Part A:
Polym. Chem. 2013, 51 (2), 327−333.
I
DOI: 10.1021/acs.inorgchem.8b02271
Inorg. Chem. XXXX, XXX, XXX−XXX