Group 4 complexes of salicylbenzoxazole ligands as effective catalysts for the ring-opening polymerization of lactides, epoxides and copolymerization of ε-caprolactone with L-lactide

Article abstract of DOI:10.1016/j.polymer.2016.08.007

A series of salicylbenzoxazole pro-ligand 2-(5-X-benzoxazol-2-yl)-6-R₁-4-R₂-phenol, L1H-L4H were subjected to reaction with titanium, zirconium and hafnium alkoxides. The substituents (R₁, R₂, X) present in the pro-ligand back bone are (H, H, H) for L1H; (H, H, Cl) for L2H; (Br, Br, H) for L3H; and (Cl, Cl, H) for L4H. The mononuclear Ti complexes, 1a-4a of [M(L)₂(OⁱPr)₂] formulation were prepared by combining the pro-ligands L1H-L4H with Ti(OⁱPr)₄ in dry toluene. However, with the addition of H₂O (0.5?M equiv.), the controlled hydrolysis of Ti complex 4a in a mixed solvent of THF/toluene produced oxo-bridged dinuclear Ti complex 5a having [M(μ-O)(L)₂(OⁱPr)]₂ formulation in good yield (≥75%) using L4H as the pro-ligand. The composition of the mononuclear Zr and Hf complexes of L1H-L4H are found to be quite similar to that of the Ti complexes. Additionally, mononuclear Zr complexes 1b-4b and Hf complexes 1c-4c were prepared. The oxo-bridged dinuclear Zr and Hf complexes 5b and 5c were isolated from the controlled hydrolysis reactions of mononuclear Zr and Hf complexes 4b and 4c. All these complexes were well characterized and crystal structure of six of the fifteen complexes were determined. These complexes were used as initiators for the ring-opening polymerization (ROP) of rac-LA, copolymerization of L-LA with ε-CL and homopolymerization of rac-epoxides. In particular, the dinuclear Zr complex (5b) was found to exhibit extremely high activity in these polymerization studies. Iso-enriched polymers were observed upon using the Zr and Hf complexes of the pro-ligands L3H or L4H as the initiator. DFT studies have been carried out on the ROP of lactides initiated by the Ti, Zr and Hf complexes 1a, 1b and 1c. The results indicated that the activation barrier to reach the transition state using Zr complexes is lower and hence facile as compared to the Ti and Hf complexes.

Full text of DOI:10.1016/j.polymer.2016.08.007

Accepted Manuscript
Group 4 complexes of salicylbenzoxazole ligands as effective catalysts for the ring-
opening polymerization of lactides, epoxides and copolymerization of ε-caprolactone
with L-Lactide
Sreenath Pappuru, Debashis Chakraborty, J. Vijaya Sundar, Sagnik K. Roymuhury,
Venkatachalam Ramkumar, Vinodh Subramanian, Dillip Kumar Chand
PII:
S0032-3861(16)30664-4
10.1016/j.polymer.2016.08.007
JPOL 18904
DOI:
Reference:
To appear in: Polymer
Received Date: 23 May 2016
Revised Date: 29 July 2016
Accepted Date: 2 August 2016
Please cite this article as: Pappuru S, Chakraborty D, Vijaya Sundar J, Roymuhury SK, Ramkumar V,
Subramanian V, Chand DK, Group 4 complexes of salicylbenzoxazole ligands as effective catalysts for
the ring-opening polymerization of lactides, epoxides and copolymerization of ε-caprolactone with L-
Lactide, Polymer (2016), doi: 10.1016/j.polymer.2016.08.007.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Group IV complexes of a series of salicylbenzoxazoles (L1H-L4H) are prepared. Alkoxide bearing mono
nuclear complexes are formed with these pro-ligands (L1H-L4H). However, control hydrolysis of L4H
supported mono nuclear complexes resulted in alkoxide bearing oxo-bridged dinuclear complexes. These
complexes are employed for homopolymerization of rac-LA, rac-epoxides and co-polymerization of L-LA with
ε
-CL. The complexes of L4H exhibited superior reactivity for the polymerization reactions.
Isotactic
D-LA L-LA
Highly reactive
complex
L-LA
ε
-CL
random

ACCEPTED MANUSCRIPT
Group 4 complexes of salicylbenzoxazole ligands as effective
catalysts for the ring-opening polymerization of lactides, epoxides
and copolymerization of ε-Caprolactone with L-Lactide
1
,1
2
Sreenath Pappuru, Debashis Chakraborty,* J. Vijaya Sundar, Sagnik K.
a
a
,2
Roymuhury, Venkatachalam Ramkumar, Vinodh Subramanian,* and
Dillip Kumar Chand*
,1
1
Department of Chemistry, Indian Institute of Technology Madras, Chennai-600 036, Tamil
Nadu, India
2
Chemical Laboratory, Central Leather Research Institute, Adyar, Chennai 600 020, Tamil
Nadu, India.
*
Tel: +91 44 22574223. Fax: +91 44 22574202. E-mail: dchakraborty@iitm.ac.in;
dillip@iitm.ac.in.
1

ACCEPTED MANUSCRIPT
Abstract
A series of salicylbenzoxazole pro-ligand 2-(5-X-benzoxazol-2-yl)-6-R -4-R -phenol, L1H-
1
2
L4H were subjected to reaction with titanium, zirconium and hafnium alkoxides. The
substituents (R , R , X) present in the pro-ligand back bone are (H, H, H) for L1H; (H, H, Cl)
1
2
for L2H; (Br, Br, H) for L3H; and (Cl, Cl, H) for L4H. The mononuclear Ti complexes, 1a-
i
4
a of [M(L) (O Pr) ] formulation were prepared by combining the pro-ligands L1H-L4H
2
2
i
with Ti(O Pr) in dry toluene. However, with the addition of H O (0.5 molar equiv.), the
4
2
controlled hydrolysis of Ti complex 4a in a mixed solvent of THF/toluene produced oxo-
i
bridged dinuclear Ti complex 5a having [M(ꢀ-O)(L) (O Pr)] formulation in good yield
2
2
(≥75%) using L4H as the pro-ligand. The composition of the mononuclear Zr and Hf
complexes of L1H-L4H are found to be quite similar to that of the Ti complexes.
Additionally, mononuclear Zr complexes 1b-4b and Hf complexes 1c-4c were prepared. The
oxo-bridged dinuclear Zr and Hf complexes 5b and 5c were isolated from the controlled
hydrolysis reactions of mononuclear Zr and Hf complexes 4b and 4c. All these complexes
were well characterized and crystal structure of six of the fifteen complexes were determined.
These complexes were used as initiators for the ring-opening polymerization (ROP) of rac-
LA, copolymerization of L-LA with ε-CL and homopolymerization of rac-epoxides. In
particular, the dinuclear Zr complex (5b) was found to exhibit extremely high activity in
these polymerization studies. Iso-enriched polymers were observed upon using the Zr and Hf
complexes of the pro-ligands L3H or L4H as the initiator. DFT studies have been carried out
on the ROP of lactides initiated by the Ti, Zr and Hf complexes 1a, 1b and 1c. The results
indicated that the activation barrier to reach the transition state using Zr complexes is lower
and hence facile as compared to the Ti and Hf complexes.
Keywords: Salicylbenzoxazoles; Group 4 metal; ROP; iso-enriched PLAs; Poly ethers
2

ACCEPTED MANUSCRIPT
1
. Introduction
Synthesis of biodegradable polymers has attracted considerable attention in the recent years
because of the vast industrial, cosmetic, and pharmaceutical applications of such materials.
Biodegradable polymers possess good advantages in biomedical research such as drug
delivery, fabrication of surgical sutures, gene transfer, tissue engineering, and regenerative
medicine [1-4]. These polymers are synthesized via the ring-opening polymerization (ROP)
of cyclic monomers such as β-lactones, δ-lactones, ε-caprolactone, trimethylene carbonate
and cyclic diesters including lactide and glycolide. The ROP reactions have been usually
carried out using a variety of metal complexes [5-24], and recently organo-catalysts [25-26]
have also been explored. Among the synthetic polymers, poly(lactic acid) (PLA),
polycaprolactone (PCL), and their copolymers, are considered as the leading sustainable
materials [27]. In industry, PLA is synthesized by using Sn(Oct) as the initiator [28-31]. In
2
addition to the toxic nature of the Sn compound, the other associated disadvantages of this
industrial method include the lack of control in molecular weight (Mn(expt)), molecular weight
distribution (MWDs), and stereo selectivity [28-31]. In order to gain control over the
polymerization process and avoid toxicity, a variety of well characterized metal complexes
have been explored as initiators in the recent years [5-23]. In this context, the importance of
group 4 metal complexes is worth emphasizing, particularly for their low toxicity and good
control over the polymerization processes [7]. Polymerization of L-LA yields isotactic-PLA,
whereas polymerization of rac-LA, produces either atactic [28], isotactic [31-35], or
heterotactic [36-41] PLA depending on the catalyst. While the formation of atactic product is
most usual and there are limited reports on the production of heterotactic and isotactic PLA.
Preparation of isotactic PLA remains a significant challenge, at the same time it is rewarding
as there is a huge demand of such compounds in the market. Group 4 metal complexes of
bidentate ligands (chosen for the present work) are less explored for production of isotactic
3

ACCEPTED MANUSCRIPT
PLA in contrast to that of Salan, Salen and Salalen type tri and tetradenate ligands.
Preparation of block and random copolymers via one pot copolymerization of L-LA and ε-CL
using group 4 metal complexes has also been published recently [27, 42-43].
Coates and co-workers reported spectacular results on the homopolymerization of
rac-epoxides to develop enantiopure epoxides and stereoregular polyethers. They have used
designed bimetallic cobalt compounds as catalysts for the polymerization reactions [44-47].
Their finding motivated us to explore the catalytic activity of group 4 complexes for the
solvent free ROP of rac-epoxides. So far, limited number of reports are available that
describe the use of group 4 complexes towards the homo and co-polymerization of epoxides
[
48-57]. The known methods for the synthesis of isotactic polyethers by using isoselective
catalysts suffer from poor activity and low stereoselectivity. Most of the catalysts produced
stereo irregular polyethers by randomly incorporating both enantiomers into the polymer
chains. Group 4 transition metal complexes bearing [N,O]-donors at the ligand back bones,
though rare, have been utilized to improve cyclic esters and olefin polymerization activity and
also to control polymer properties such as stereoregularity, comonomer incorporation, and
microstructure. The variety of ligands that are explored for such purposes include imino
phenoxide [58-62], salicylpyridine [63-64], salicyloxazoline [65], and β-ketoimine [66].
In our ongoing research into group 4 metal chemistry and polymerization studies, our
group has previously reported a series of group 4 metal complexes with bi-dentate and tri-
dentate ligands for e.g., bis(imino)phenoxide ligands,[59] imino phenoxide ligands, [60]
[
NNO]-type tridentate imine functional salicylaldiminato ligands [53, 57], Mannich base
aminophenol ligands [61] and benzotriazole phenoxide ligands [52] for the ROP of cyclic
esters and lactides. Since the group 4 metal catalysts are very efficient and popular for
performing such polymerization, we observed excellent catalytic activity of these complexes
towards the ROP of cyclic esters and lactides. In the case of imino phenoxide and
4

ACCEPTED MANUSCRIPT
bis(imino)phenoxide ligands supported group 4 catalysts the major drawback was that the
observed M s were far higher than the theoretical values and also no control in stereo
n
selectivity during rac-LA polymerization was observed [59,60]. In the case of [NNO]-type
tridentate imine functional salicylaldiminato ligands supported mononuclear and binuclear
group 4 catalyst systems, most of the complexes have taken ≥60 min to convert 200 rac-LA
units into product and the resulting polymers were heterotactic enriched. The nuclearity in
these complexes not much affected the polymerization results [53]. In contrast group 4
catalyst systems with [NNO]-type tridentate ligands where the amine moiety has cyano
groups and central phenoxy goup has ortho steric t-butyl groups has a profound influence on
the polymerization results. Most of the complexes have taken ≤20 min to convert 200 rac-LA
units into product and produced heterotactic PLA [57]. Recently, Mannich base aminophenol
ligands supported group 4 chloride complexes were also applied as catalysts for the ROP L-
LA and rac-LA [61]. In these cases, benzyl alcohol was used as external initiator to enhance
the catalytic activity and also during rac-LA polymerization hetero selectivity was observed.
In the case of benzotriazole phenoxide ligand supported group 4 catalyst systems, we
reported benzotriazole phenoxide ligands with bulkier substituents at the ortho position (for
eg., tert-butyl and pentyl) for reacting with group 4 metals [52]. We never synthesized single
site complexes and were able to observe dimeric or coordinatively saturated complexes. All
these ligands have stronger conjugative effect than the above salicylaldimine functional
ligands, have shown significant influence on the metal center and the active center of the
catalyst during the catalytic reaction and good heterotacticity was observed. However iso-
enriched polymers were not observed by using the above catalyst systems. All these
aforementioned catalytic systems achieved heterotactic rich PLAs and also none of them
were applied for copolymerization of L-LA with ε-CL. With such knowledge, we were
specifically interested about the use of group 4 metal complexes stabilized by stronger
5

ACCEPTED MANUSCRIPT
conjugative benzoxazole phenoxide ligands with the hope of synthesizing single site
complexes active towards ROP. While mono nuclear Ti and Zr chloro complexes of
salicylbenzoxazoles have already been reported for ethylene polymerization by Jin et al., [67-
6
8]. Nevertheless, Ti, Zr and Hf mono as well as bi nuclear alkoxide complexes of
salicylbenzoxazole ligands for homopolymerization of rac-LA, copolymerization of L-LA
with ε-CL and homopolymerization of epoxides have not been reported until date.
Additionally, in this study DFT studies have been for the first time applied from our group
for the ROP of lactides initiated by the Ti, Zr and Hf complexes 1a, 1b and 1c and
successfully optimized intermediates and transition states which are used for the ROP of
lactide. By contrast to previous observations, Jones and co-workers reported a detailed study
of group 4 complexes with salalen ligands [33], showing that Zr and Hf complexes are the
O
best active catalysts, under solvent-free at 130 C, 300 eq. of rac-LA were converted to iso-
enriched PLA within 5 to 15min. However they achieved broad M and MWDs. The
n
binuclear Zr and Hf complexes in the present study can be comparable with these systems.
Nevertheless, we achieved good control in M and MWDs. Additionally these bidentate
n
ligands containing group 4 complexes are the rare examples that are producing iso-enriched
o
PLA under solvent-free at 140 C [31-35]. On the other hand group 4 complexes with
nitrogen-heterocycle containing phenolate ligands are well reported for producing
heterotactic or atactic PLA [10-12]. Very recently aluminum complexes containing
salicylbenzoxazole ligands have been demonstrated to be efficient initiators for the ROP of
cyclic esters [18].
In this work, we designed a series of group 4 metal complexes with less bulkier
salicylbenzoxazoles which have chloro and bromo substituents on the ligand back bone. The
coordination spheres of the complexes also contain coordinated alkoxide groups. The Ti, Zr
and Hf complexes are fully characterized and utilized for homopolymerization of rac-LA,
6

ACCEPTED MANUSCRIPT
copolymerization of L-LA with ε-CL and homopolymerization of epoxides. The epoxides
used are rac-cyclohexeneoxide (rac-CHO), rac-propyleneoxide (rac-PO) and rac-
styreneoxide (rac-SO). The effect of chloro and bromo substituents present at the oxazoline
and phenol ring were also investigated.
Scheme 1 Syntheses of the ligands L1H-L4H.
2
. Result and discussion
2
.1 Synthesis of the ligands, L1H-L4H
The pro-ligands L1H-L3H are reported [67-69] whereas L4H is a new compound. All the
ligands were synthesized by slight modifications of the reported procedures (see Scheme 1)
[
67-68]. In the first step of the synthesis, salicylaldehyde derivatives were reacted with o-
aminophenol to afford the corresponding Schiff bases in quantitative yields. In the second
step, the Schiff bases were oxidized to benzoxazole using 2,3-dichloro-5,6-dicyano-1,4-
benzoquinone (DDQ) under anhydrous conditions. The crude products were isolated by
aqueous work-up and then purified by column chromatography on silica gel by eluting with a
hexane/CH Cl gradient (100:0 → 90:10). The pure ligands were isolated as white or off-
2
2
1
white solids. All the pro-ligands (L1H-L4H) were characterized by recording their H NMR,
1
3
C NMR and ESI-MS spectra (see Supplementary Data for the pro-ligand L4H). The purity
of the ligands were assessed by elemental analysis.
7

ACCEPTED MANUSCRIPT
2
.2 Synthesis of the group 4 complexes, 1a-5a, 1b-5b and 1c-5c
i
The complexes were prepared by reacting the pro-ligands, L1H-L4H with Ti(O Pr)4
,
i
i
t
Zr(O Pr) ꢁ( P OH) and Hf(O Bu) in 2:1 molar ratios in dry toluene at room temperature.
4
r
4
Attempts to synthesize 1 : 1 stoichiometry products of the type [LM(OR) ] remained futile as
3
1
a complicated mixture of several products was seen in the H NMR of the crude reaction
mixture. The synthetic route for these complexes is shown in the Scheme 2. The Ti
complexes, 1a-4a corresponding to the pro-ligands L1H-L4H are mononuclear. The nature
of the Zr and Hf complexes are quite similar to that of Ti as shown in the Schemes 2. The
mononuclear Zr and Hf complexes of L1H-L4H are designated as 1b-4b and 1c-4c,
respectively. Next, we carried out controlled hydrolysis by using 0.5 equivalent of deionized
water in dry toluene and THF mixture (1:1) at room temperature for 4a-4c complexes
(Scheme 3). The mononuclear complexes (4a-4c) that are derived from pro-ligand L4H,
dimerized spontaneously through degradation reactions giving oxo-bridged binuclear
complexes (5a-5c). The same reaction with the other analogues gave intractable reaction
mixtures. Similar phenomenon was recently observed from our group with group 4 metals
[
53].
Scheme 2 Syntheses of the Ti, Zr and Hf complexes, 1a-1c, 2a-2c, 3a-3c, 4a-4c.
8

ACCEPTED MANUSCRIPT
Scheme 3 Syntheses of the Ti, Zr and Hf complexes, 5a-5c.
1
13
The H and C NMR spectra (in CDCl ) of the complexes exhibited downfield shifted
3
signals, corresponding to some of the H/C of the salicylbenzoxazole moieties close to the
metal binding sites, as compared to the free ligand. The downfield shift of signals indicated
the coordination of the ligand to the metal center. The signals corresponding to the
coordinated alkoxide moieties are detected in the aliphatic region (see the Supplementary
1
13
Data). The H and C NMR data recorded in CDCl for all these complexes showed only one
3
set of peaks corresponding to the coordinated ligands. A similar behaviour is reported for
group 4 complexes of a bidentate [N,O]-donors ligand where the complexes possess six
membered chelate rings [52,59-60]. The ESI-MS spectra of the complexes supported the
mononuclear nature of 1a-1c, 2a-2c, 3a-3c and 4a-4c and the binuclear nature of 5a-5c in
line with the structures shown in the Schemes 2 and 3. The purity of these complexes was
further confirmed through elemental analysis.
2
.3 X-ray diffraction studies of complexes 1a, 1b, 2a, 2c, 5a and 5b
The molecular structures of complexes 1a, 1b, 2a, 2c, 5a and 5b were unambiguously
confirmed by single-crystal X-ray diffraction analysis. The ORTEPs corresponding to the
compounds are depicted in Fig. 1, along with selected bond lengths and angles. The crystal
data are depicted in Table S1 (see the Supplementary Data). All these complexes were
9

ACCEPTED MANUSCRIPT
crystallized from saturated dry toluene solutions over a period of 15 days. The complexes
have hexa-coordinated metal centers with distorted octahedral coordination environment. The
molecular structure shows that mononuclear complexes 1a, 1b, 2a and 2c possess distorted
octahedral geometry, while the oxo-bridged binuclear complexes 5a and 5b are composed of
a pair of fused distorted octahedral entities. As depicted in Fig. 1, the phenolate-O donors of
the salicylbenzoxazole ligands are coordinated trans to each other (considered axial).
However, the overlap of bulky ligands around the metal center deviates the O-M-O bond
o
angles from the ideal value of 180 and this is the reason for distortion from octahedral
o
o
geometry. The values are O1-Ti1-O4 (156.86(8) ) in 1a, O1-Ti1-O3 (155.53(10) ) in 2a, O2-
o
o
o
Zr1-O4 (149.11(9) ) in 1b, O9-Hf1-O7 (149.72(12) ) in 2c, O1-Ti1-O3 (160.44(8) ) in 5a
o
and O1-Zr1-O3 (155.07(18) ) in 5b. This deviation is illustrative of the distortion from ideal
octahedral geometry. The metal and phenolate oxygen bond lengths are (1.9180(18) Å) in 1a,
(
1.920(2) Å) in 2a, (2.043(2) Å) in 1b, (2.035(3) Å) in 2c, (1.9326(18) Å) in 5a and (2.053(5)
Å) in 5b which are comparable to the reported values of some related group 4 complexes
52,59-60]. However O donors of the alkoxide coligands are cis to each other and trans to N
donors of the salicylbenzoxazole ligands. The O-M-O bond angles described by the two
[
o
o
nucleophilic alkoxide groups are (O6-Ti1-O3 (100.29(9) ) in 1a, O6-Ti1-O5 (100.72(12) ) in
o
o
2
a, O6-Zr1-O5 (100.28(10) ) in 1b, O2-Hf1-O3 (100.96(14) ) in 2c. Similarly, the O-M-O
o
angles described by alkoxide and the oxo groups are O6-Ti1-O5oxo (102.20(10) ) in 5a and
o
O6-Zr1-O5oxo (102.6(2) ) in 5b). The metal and nitrogen bond lengths (2.298(2) Å) in 1a,
(2.306(3) Å) in 2a, (2.417(2) Å) in 1b, (2.391(4) Å) in 2c, (2.352(2) Å) in 5a and (2.412(6)
Å) in 5b are comparable to the reported group 4 complexes [52,59-60]. It should be noted
that during controlled hydrolysis reactions in the complexes 5a-5c oxo-bridging happened by
the loss of one alkoxide group per metal center leading to dimerization. As depicted in
Figures 1, the complexes 1a, 2a, 1b and 2c each possess a C -symmetry axis along the
2
1
0

ACCEPTED MANUSCRIPT
Fig. 1. ORTEP (30% probability) of 1a, 1b, 2a, 2c, 5a and 5b. Hydrogen atoms are omitted for clarity. Selected
bond distances and angles (Å, deg) for the complex 1a : Ti1-O4 = 1.9180(18), Ti1-N1 = 2.298(2), Ti-O3 =
1
=
.7927 (19); O6-Ti1-O3 = 100.29(9), O1-Ti1-O4 = 156.86(8); for the complex 1b : Zr1-O2 = 2.043(2), Zr1-N1
2.417(2), Zr1-O6 = 1.925 (2); O6-Zr1-O5 = 100.28(10). for the complex 2a : Ti1-O3 = 1.920(2), Ti1-N1 =
2
2
.306(3), Ti-O5 = 1.795 (2); O6-Ti1-O5 = 100.72(12), O1-Ti1-O3 = 155.53(10); for the complex 2c: Hf1-O7 =
.035(3), Hf1-N1 = 2.391(4), Hf1-O3 = 1.905 (3); O2-Hf1-O3 = 100.96(14), O9-Hf1-O7 = 149.72(12). for the
complex 5a : Ti1-O6 = 1.7848(18), Ti1-N1 = 2.289(2), Ti-O = 1.8075 (5), O1-Ti1-O5 = 93.66(6), O1-Ti1-O3
oxo
i
=
2
160.44(10), Ti1-O1 = 1.9189(19), Ti1 -O = 1.8074 (5); for the complex 5b: Zr1-O6 = 1.916(5), Zr1-N1 =
.452(6), Zr1-Ooxo = 1.9360 (8), O1-Zr1-O5 = 94.07(13), Zr1-O1 = 2.053(5), Zr1 -O = 1.9360 (8).
oxo
i
oxo
1
1

ACCEPTED MANUSCRIPT
approximate N O plane described by the two alkoxide oxygen, and two benzoxazole
2
2
nitrogen atoms, however, bisecting the N-M-N or O-M-O) angle. The complexes 5a and 5b
also have a C symmetry axis along the approximate O plane described by the phenoxide
2
5
oxygen, oxide oxygen atoms, however, bisecting the M-oxide-M angle. For all these
complexes the bond lengths and bond angles are well supported by the literature data [52,59-
6
2
0].
.4. Ring-opening homopolymerization of rac–LA
All complexes were tested towards the polymerization of rac-LA under the industrially
ᴼ
preferred melt condition (140 C) at 200:1 ratio of monomer to catalyst (457.7 mg, i.e. 3.18
mmol :15.9 ꢀ mol, e.g. 10 mg of 1b). The results are depicted in Table 1. All these complexes
are found to be excellent catalysts for the controlled ROP of rac-LA under solvent free
conditions, where the coordinated alkoxide group of the catalyst is the proposed initiating
group. It was observed that the conversion of monomer is faster and the microstructure
control is better for the reactions catalysed by one set of catalysts that are the Zr and Hf
complexes of the pro-ligands L3H and L4H (i.e. 3b, 3c, 4b, 4c, 5b and 5c) as compared to
the other set derived from the pro-ligands L1H-L4H (i.e. 1a, 2a, 3a, 4a, 5a, 1b, 2b, 1c
and2c). The superior behavior of the first set of catalysts is ascribed to the presence of
electron withdrawing substituent (i.e. R1) near the metal center of the complexes (at ortho-
position to the phenolic oxygen) as well as facile approach of the monomer towards Zr and
1
Hf complexes due to the bigger size of the metal center. The H NMR spectra of the PLA
obtained from all these reactions were recorded (Fig. 2). The percentage conversion of rac-
1
LA to product was determined by integration of methine region of monomer HNMR peaks
vs polymer peaks. From this conversion predicted M values (M (NMRpredict)) were calculated
n
n
1
2

ACCEPTED MANUSCRIPT
1
(
see Table 1). However exact M values were not determined by H NMR analysis because
n
the end group signals were not always observed/identified.
Table 1 Raing-opening polymerization of rac–LA catalyzed by complexes 1a
-
1c
,
2a
-
2c
,
3a
-
3c
,
4a-4c
and 5a-5c
b
c
d
e
f
[
rac-LA]o/
time
Conv M (GPC)
M (calcd) M (NMRpredict
)
n
n
n
d
g
h
Entry Cat
M /Mn
Pm
Tg
w
[
Cat] /[BnOH]o (min)
(%)
92
94
90
92
95
95
95
96
96
98
98
95
98
99
99
99
98
97
92
98
98
97
98
(kg/mol)
22.26
20.45
4.21
(kg/mol)
(kg/mol)
26.58
27.16
5.30
o
1
2
3
4
5
6
7
8
1a
2a
2a
3a
4a
5a
1b
2b
2b
3b
3b
3b
4b
5b
5b
5b
5b
1c
1c
2c
3c
4c
5c
25
20
6
28.89
28.89
5.87
1.36
1.25
1.30
1.16
1.13
1.09
1.33
1.26
1.19
1.25
1.14
1.12
1.17
1.07
1.16
1.03
1.12
1.28
1.19
1.12
1.19
1.16
1.09
0.50 51.28
0.52 42.21
2
2
2
00/1/0
00/1/0
00/1/5
-
-
-
200/1/0
200/1/0
200/1/0
200/1/0
200/1/0
200/1/0
200/1/0
200/1/5
200/1/0
200/1/0
100/1/0
200/1/0
400/1/0
800/1/0
200/1/0
200/1/5
200/1/0
200/1/0
200/1/0
200/1/0
18
12
8
21.58
30.76
30.38
41.42
42.38
34.96
31.56
7.10
28.89
28.89
28.89
28.89
28.89
28.89
28.89
5.87
26.60
27.44
27.44
27.44
27.73
27.73
28.31
5.75
0.55
0.53 42.47
0.53 45.26
10
7
0.50
-
0.53 51.89
i
9
8
-
-
1
0
1
5
0.61 54.65
1
2
-
-
-
-
i
1
2
6
29.69
31.42
15.85
32.39
62.79
132.67
42.41
6.98
28.89
28.89
14.47
28.89
57.71
115.36
28.90
5.87
27.46
28.31
14.33
28.60
nd
1
1
1
1
1
1
1
2
2
2
2
3
4
5
6
7
8
9
0
1
2
3
3
0.63 49.58
0.63 27.34
0.63 48.85
0.63 54.46
0.63 55.87
1
2
4
9
nd
8
28.03
5.40
0.50
-
-
-
4
7
30.58
40.29
30.98
32.41
28.90
28.90
28.90
28.90
28.32
28.32
28.03
28.32
0.53 51.25
0.60 54.38
5
4
0.65
0.64
-
-
2
a
b
Reaction conditions: Unless otherwise noted, [LA] /[Cat.] = 200:1, 140 ˚C, solvent-free.
Time of polymerization measured by quenching the polymerization reaction when all monomer was found
0
0
consumed.
c
1
Obtained from H NMR determination.
d
Measured by GPC at 27
°
C in THF relative to polystyrene standards with Mark-Houwink corrections for Mn
.
e
(calcd)
Mn
group, (in presence of BnOH) at 100 % conversion = [M] /[BnOH] × mol wt (monomer) + mol wt (BnOH).
at 100 % conversion (in absence of BnOH) = [M] /[C] × mol wt (monomer) + molecular weight of end
o
o
o
o
f
1
Obtained from H NMR analysis of monomer and PLA.
The probability for isotactic enchainment calculated from homonuclear decoupled H NMR analysis (Bernoullian
statistics).
g
1
h
Tg (glass transition temperature) determined from the second heating cycles of DSC.
i
L-LA as the monomer, [LA] /[Cat.] = 200:1, 130 ˚C, solvent-free.
0 0
1
3

ACCEPTED MANUSCRIPT
1
Fig. 2 H NMR spectrum (500 MHz, CDCl ) of the low-molecular-weight PLA obtained from a
3
o
reaction between rac–LA and 2b in 5:1 ratio at 140 C.
Furthermore the number-average molecular weight (Mn(expt)) and molecular weight
distributions (MWDs) values of polymers were measured by GPC analysis (Supplementary
Data, Fig.S49-50). For 4a, 5a, 3b, 4b, 5b, 2c, 4c and 5c the observed molecular weights
(Mn(expt) GPC method) are close to those predicted on the basis of the monomer loading
(200:1 i.e. 28.8 kg/mol of initiator) and consistent with one polymer chain per metal.
Molecular weight distributions (M /M ) are narrow, indicative of well controlled
w
n
polymerization. For 1b, 2b, 1c, and 3c the observed molecular weights (Mn(expt)) are higher
than predicted, probably due to the relatively slower initiation rate compared to the
propagation [43, 70] and involvement of one of the alkoxides. The MWDs are slightly
broader indicative of only minor amount of transesterification and/or formation of cyclic
oligomers during propagation and also indicative of well controlled polymerization. In
contrast, the molecular weights of the products obtained by using the complexes 1a, 2a and
3
a as catalyst are lower than the predicted values. As anticipated, all the Ti complexes are
less active and required longer time (10-30 min) for rac-LA polymerization as compared to
the corresponding Zr and Hf complexes. This can be explained by the fact that the metal
1
4

ACCEPTED MANUSCRIPT
centers in titanium complexes are smaller in size as compared to the corresponding Zr and Hf
complexes. The approach of the monomer is facile in Zr and Hf complexes due to the bigger
size of the metal center. The complexes 5b and 5c showed good activity as compared to all
other complexes towards ROP of rac-LA. The proposed reason is the augmented
electrophilic nature of the metal centers in the complexes 5b and 5c due to the presence of
chloro substituents at otho and para positions in the phenolic ring. In addition, the co-
ordination environment around the metal center is more opens susceptible to monomer
coordination. The catalytic performance of these complexes is comparable with reported
group 4 catalysts for rac-LA polymerization, in terms of activity and control over
polymerization process [35,55,42]. For 5a-5c, the proximity of the observed M to theoretical
n
M indicates that only one metal center participates in the polymerization. This is consistent
n
with the previous observations [7, 52 and 53].
To understand the thermal behaviour of these polymers we collected data using
differential scanning calorimetry (DSC) technique. According to DSC measurements, all
ᴼ
these polymers exhibited glass transition temperatures, T = 42-54 C (Fig. 3), which indicates
g
a transition of the polymer from the glassy state to an extremely super cooled liquid state. It
is because of the rigid nature of the molecular chain of polymers [72]. The M as well as Tg
n
values increased with the increase of monomer ratio as demonstrated from the polymers
prepared using 5b (See Table 1 entry 14-17) (see Supplementary Data, Fig. S51).
In a complementary series of experiments, the initiator (2b
was kept constant (10 mg, i.e. 14.3 ꢀ mol for 2b, 7.0 ꢀ mol for 5b and 6.1 ꢀ mol for
) and the monomer amount was varied to study the effect of monomer to catalyst
ratio on the Mn and MWD values. The study was performed in the range of 100:1 to
00:1 ([M]o [Cat] ) where upon a linear relationship was observed for the plots of Mn
versus [rac-LA] : [Cat] also MWDs versus [rac-LA] : [Cat] (Fig. 4). This indicates
, 5b and 5c) amount
5
c
8
/
o
t
t
1
5

ACCEPTED MANUSCRIPT
that the values of Mn and MWD’s remain controlled (1.05 – 1.12) which suggest that
the growth of polymer chain happened in a controlled fashion.
nd
Fig. 3 DSC traces (heating rate: 10 °C/min, 2 scan) of homopolymers of rac-LA with complexes 2a, 2b, 2c, 3c
and 5a at 200 : 1 monomer initiator ratio.
Fig. 4. Plot of Mn and Mw/M vs.[M]ₒ/[Cat]ₒ for rac-LA polymerization at 140 °C using 2b, 5b and 5c.
n
In the case of 2a, 3b and 1c we tested the polymerization rate by the addition of
excess benzyl alcohol (BnOH) as co-initiator during the ROP of rac-LA. Gratifyingly, the co-
1
6

ACCEPTED MANUSCRIPT
initiator acted as a chain-transfer agent (Table 1) while the observed M and expected M are
n
n
still comparable and the MWDs are also reasonably narrow.
Since the complex 5b showed good activity and selectivity towards ROP of rac-LA,
we carried out kinetics studies with complex 5b for the polymerization of rac-LA in the ratio
o
of [rac-LA] /[Cat] = 200, 400 and 600 at 140 C and showed that there is a linear
o
o
relationship, between initiator/monomer ratios and time. Under solvent free conditions,
1
conversions of rac-LA with time at various initiator/monomer ratios were monitored by H
NMR spectroscopy until monomer consumption was complete (Supplementary Data, Fig.
S52). In each case, first-order kinetics in monomer was observed. The appropriate semi
logarithmic plots for these polymerizations are shown in Fig. 5. From the slope of the plots
(
Fig. 5), the first order rate constants (K ) for rac-LA polymerizations at 400:1, 600:1 and
app
-1
-1
-1
-1
8
00:1 monomer/initiator ratios are 6.2×10 , 3.6×10 and 3.1×10 min respectively.
Fig. 5 Semilogarithmic plots of rac–LA, conversion in time initiated by 5b at various initiator/monomer ratios
[M:Cat] = 400; I, [M:Cat] = 600; II, [M:Cat] = 800; III).
(
2
.5 MALDI-TOF MS analysis of oligomeric PLA
Analysis of the oligomeric PLA produced with these compounds (at 20:1 ratio) under solvent
free condition through MALDI-TOF mass spectrometry indicated iso-propoxy or tertiary-
1
7

ACCEPTED MANUSCRIPT
butoxy end capped polymer chains (Fig. S53 & S54, see the Supplementary Data). However,
all the fragments are adducts of a potassium ion or a sodium ion or acetonitrile. The presence
1
of peaks at 4.4 ppm (methine) and 1.3 ppm (methyl) in the H NMR spectrum of the product
also indicated isopropoxy end capped polymer chains (Fig. 2). Careful observation of
magnified MALDI-TOF spectra suggested that the major signals appear as the potassium
adduct with the mass differences of 72.06 corresponding to the molar mass of lactide half
unit(M = 72.06×2). In addition, less intensity signals with a molecular weight interval of
7
2.06 were also observed. However, all the fragments are adducts of a potassium ion or a
sodium ion or acetonitrile. The exact mass of the each signal is the sum of the mass of the
iso-propoxy or tertiary-butoxy initiating group (M = 59.05 or 73.07), the respective number
of lactide repeating units, and the counterion. This data suggests that the expected
coordination–insertion mechanism operated during the ROP. The sharpness of the base line
indicates that there is minimal transesterification.
2
.6 Microstructure analysis of polylactic acid (PLA)
The Zr and Hf complexes bearing electron withdrawing substituents ortho and para to the
phenolic oxygen (4b, 5b, 4c, and 5c) resulted in good control in stereoselectivity and
1
produced isotactic rich polymers as observed from homonuclear decoupled H NMR spectra
(Fig 6 and Supplementary Data, Fig.S55-63) [71]. The peaks were assigned to appropriate
tetrads in accordance with the shifts reported in the literature [9] and the tetrad distribution
1
values were evaluated from deconvolution of the homodecoupled HNMR spectrum (Fig. 6)
and P are calculated using Bernoullian statistics based on the assignments made by Coates
m
and co-workers (where P +P = 1) [9]. The spectra showed intense peak corresponding to iii
m
r
tetrad suggesting repetition of the same stereocenter in the polymer chain. However in the
case of the other complexes (1a, 2a, 3a, 1b, 1c and 2c) atactic polymers were formed.
1
8

ACCEPTED MANUSCRIPT
1
Fig. 6. Deconvolution of the methine region (δ =5.10–5.25 ppm) of the homodecoupled H NMR spectrum (500
MHz, CDCl , 298 K) of a PLA prepared with complex 5c (Pm = 0.64) (entry 23, Table 1), showing the
3
1
respective integrations for each tetrad [14]. Pink, recorded homonuclear decoupled H NMR spectrum.
Furthermore the effect of co-catalyst on stereoselectivity of PLA was also examined. We
o
carried out the reactions under the same conditions (solvent-free, 140 C) with complex 5b
for the polymerization of rac-LA in the ratio of [LA] /[Cat] /[BnOH]
o
=
o
o
1
5
00:1:5 and 1000:1:10. From the observed homodecoupled H NMR results, the control in
stereoselectivity was slightly reduced by increasing the concentration of monomer and
BnOH (Supplementary Data, Fig. S61, S62). Hence it is inferred that the complex 5b suffered
from partial removal of ligand moiety at high concentration of BnOH. In L-LA
polymerization using any of the catalysts a single resonance peak was observed at the
1
methine region in the homonuclear decoupled H NMR spectra (Supplementary Data, Fig.
S64). The L-LA polymerization data for the 2b and 3b is provided in the Table 1, entry 9 and
1
2. This indicates that practically no epimerization occurred at the chiral center of the
polymer.
1
9

ACCEPTED MANUSCRIPT
2
.7 Density functional theory (DFT) calculations
In order to understand the difference in reactivity between Ti, Zr and Hf complexes towards
ROP of lactide, DFT investigations were carried out at M06-2X/6-31g* level of theory. The
free energy profiles and optimized structures obtained for the initiation step of the ROP of LA
with both Ti, Zr and Hf are presented in Fig. 7, S65, S66 and S67 (see the Supplementary
Data). The monomer coordination at metal center is important to open the monomer ring of
LA. We represented the carbonyl oxygen and carbon coordinated to the metal center as O1
and C1 respectively and the oxygen of one of the alkoxide near to the monomer as O3
(Supplementary Data, Fig. S65, S66, and S67). In the complex formation, EC, the bond
-1
Fig. 7 Calculated free energy profiles (kcal mol ) of the initiation step in the ROP of LA mediated by Ti and Zr
complexes, 1a and 1b.
distances between the metal center and O1 of LA are 2.35, 2.44 and 2.40 Å for Ti, Zr and for
Hf, respectively (Supplementary Data, Fig. S65, S66 and S67). The metal centers are lodged
2
0

ACCEPTED MANUSCRIPT
in octahedral coordination environments even when the monomer approaches the metal
center due to partial ligand dissociation (Fig. S65, see the Supplementary Data). The
intermediates and transition states used for the ROP of LA were similar to those reported in
the literature [73-75]. The ROP proceeds by the S 2 attack of alkoxide oxygen (O3) on the
N
carbonyl carbon of monomer (C1) via a transition state with barrier height of 14.2, 13.7 and
9
.8 kcal/mol for Ti, Zr and Hf complexes respectively and leads to the intermediate IM1. In
IM1, the alkoxide group is bonded to C1 whereas O1 is bonded to the metal center. The
-1
formation of the first intermediate IM1 is endergonic (+5.2 kcal/mol ) (Fig. S65, see the
Supplementary Data) with respect to the Zr . Rotation around the O1-C1 bond afforded IM2
EC
(
Fig. S65, see the Supplementary Data), in which the O2 atom of LA approaches the metal
Zr····O2 = 2.34 Å , Hf····O2 = 4.13 Å and Ti····O2 = 4.16 Å). Here it should be noted that
(
the longer bond distances for Ti and Hf complexes indicates lesser activation of O2 and the
same is reflected in TiTS2 and Hf TS2 barrier height. However Hf TS1 barrier height is
comparatively lower than Ti and Zr. Which may influence the Hf complex for high reactivity
in these reactions. From IM2 breaking of the acyl-oxygen bond takes place and as a
consequence the ring will open by passing through the low-lying transition state TS2, TS2
comprises of an eight-membered ring with a shorter M−O2 distance (Zr ···· O2= 2.174 Å)
and a longer M − O1 donor interaction (Zr ···· O1= 2.230 Å) (Fig. S65, see the
-1
Supplementary Data). The activation energy for the TS2 is calculated to be 15.4 kcal mol for
-1
-1
Ti, 3.1 kcal mol for Zr and 9.3 kcal mol for Hf with respect to the IM2. This transition state
led to the product in which the monomer was fully opened (Fig. S65, see the Supplementary
Data). The energies for the conversions of the reactant to product were exergonic by −21.5
-1
-1
-1
kcal mol for Ti, −9.5 kcal mol for Zr and −19.8 kcal mol for Hf based models,
respectively. Our DFT calculations suggest that the polymerization reactions are metal
dependent and the ring-opening transition state TS2 is more facile for Zr complex. This may,
2
1

ACCEPTED MANUSCRIPT
therefore, account for the greater potency of zirconium compounds for these polymerization
reactions. However the reactions are also dependent on the steric and electronic nature of the
ligands. Also from Fig. S68A in Supplementary Data the highest occupied molecular orbital
(HOMO) of Ti and Zr compound, it was observed that the eletrons from ligand are tranfered
to metal and facilitating to cleave the isopropoxide bond [76]. Additionally MPA (Mulliken
i
Population Analysis) method was carried out and confirmed that oxygen atoms of O Pr
moieties(–0.610 net charge) are more electron rich than the oxygen atoms of phenoxide
groups (-0.312 net charge) in the complex 1b (Fig. S68B in Supplementary Data).
2
.7 Copolymerization of L-LA with ε–CL
Since these group 4 complexes showed good activity and selectivity towards ROP of rac-
lactide, we decided to focus our studies on ring-opening copolymerization of L-lactide and ε-
caprolactone by using 100 equivalents of both monomers in dry toluene at 100 °C. The
number-average molecular weight (Mn(expt)) and molecular weight distributions (MWDs)
values of copolymer products were measured by GPC analysis. We investigated all these
complexes towards ring-opening copolymerization studies and the results are summarized in
Table 2. Complexes 4a, 5a, 1b-5b, and 1c-5c showed good reactivity and the obtained
1
copolymers are incorporated with more of lactide than ε-CL. From the H NMR analyses, the
percent of lactide was found to be 72-98%, while that of the conversion of ε-caprolactone was
5
5 to 68 % (Table 2). Under the prolonged reaction time with complex 5b and 5c we
achieved 100 % conversion for both the monomers (see Table 2, entry 16, 23). The presence
1
of multiple resonances at about 4.2-2.2 ppm in H NMR (Fig. 8 & Fig. S69, in
Supplementary Data) and carbonyl resonances, methine carbon resonances at about 170-
13
1
69ppm, 70-63ppm in C NMR (Fig. 9 & 10) confirms the random distribution of both
1
monomers (L-LA and CL) along the polymer chain to afford random copolymers. The H and
2
2

ACCEPTED MANUSCRIPT
Table 2. Copolymerization of
L
-LA and
ε
-CL with 1a
-
1c
,
2a
-
2c
,
3a-3c, 4a-4c and 5a-5c
b
d
e
Conv (%)
M (calcd)
n
M (GPC)
n
c
c
e
f
f
Entry
Cat
1a
2a
3a
4a
5a
1b
2b
2b
2b
3b
3b
3b
4b
5b
5b
5b
1c
2c
3c
4c
5c
5c
5c
LA
42
CL
22
35
38
45
58
57
64
76
82
60
77
87
63
68
78
100
55
65
59
60
65
75
100
LLA
1.9
3.2
2.8
2.4
2.8
3.1
4.7
3.8
4.1
3.4
3.6
3.2
4.4
5.1
4.8
5.3
4.5
4.8
4.2
3.8
5.5
4.8
5.2
LCL
(kg/mol)
8.56
(kg/mol)
5.65
M /Mn
Tg
-
Tm
w
1
2
3
4
5
6
7
1.4
1.7
2.1
1.9
2.4
1.3
2.0
2.2
2.5
2.5
3.1
2.9
4.1
3.7
3.5
4.4
2.4
2.8
3.2
3.4
3.0
2.9
4.1
1.34
1.31
1.19
1.21
1.19
1.27
1.20
1.25
1.28
1.24
1.25
1.25
1.19
1.17
1.22
1.23
1.30
1.27
1.22
1.17
1.21
1.18
1.21
-
68
59
13.79
12.84
14.07
17.72
16.88
19.12
21.42
22.54
17.66
21.96
23.54
19.87
21.88
23.38
25.90
16.38
20.68
18.41
18.81
21.11
23.03
25.90
8.52
-
-
14.96
16.25
19.63
19.65
22.84
23.81
25.91
21.95
25.12
28.60
20.54
24.63
25.87
27.51
19.54
24.67
22.87
21.78
24.41
21.31
22.40
-
-
62
-
-
77
-
--
72
-
-
82
26.43
123.63
g
8
88
-
-
h
9
91
-
-
1
0
75
22.52
125.61
g
1
1
2
91
-
-
h
1
94
-
-
1
3
88
-
-
1
4
98
25.78
123.47
g
1
5
6
100
100
70
-
-
h
1
-
-
1
7
-
-
1
1
2
2
8
9
0
1
92
24.36
122.42
81
25.41
123.12
83
-
-
95
84.05
110.60
g
2
2
3
100
100
-
-
-
-
h
2
a
Each reaction was performed in 2.5 mL of dry toluene, [L-LA]ₒ : [ ε-CL]ₒ : [Cat.]ₒ = 100 : 100 : 1, at 100 ºC.
b
1
conversion determined by H NMR spectroscopy.
Average sequences length of the caproyl unit and of the lactidyl unit as determined by C NMR analysis.
c
13
d
M (calcd) calculated using [(M/I) × (% conversion of L-LA) × (mol wt of L-LA)] + [(M/I) × (% conversion of ε-CL) ×
n
(
mol wt of ε-CL)].
e
M (GPC) measured by GPC at 27 °C in THF relative to polystyrene standards with Mark-Houwink corrections for M .
n
n
f
T and T determined from the second heating cycles of DSC
g
m
g
h
reaction was carried out upto 24h under the same condition.
reaction was carried out upto 48h under the same condition.
1
3
C NMR multiple sequences were assigned according to the literature [42-43, 77-78]. The
average length of the caproyl (L ) and lactidyl (L ) sequences can be calculated from the
CL
LA
integrals of the triads sequences signals following previously reported methods [79-81]. An
ideal random copolymer will have L = L = 2. In recent reports random copolymers were
LA
CL
2
3

ACCEPTED MANUSCRIPT
achieved by using bis-thioetherphenolate group 4 complexes [43], bis(pyridonate) and
bis(amidate) titanium complexes [42] and block copolymers were also achieved in the
presence of ferrocene based ligands containing group 4 complexes [27].
1
Fig. 8 H NMR spectra (500 MHz, CDCl ) of the copolymer showing the
ε
- and α-methylene ranges in the
3
copolymers of entry 14 (a) entry 10 (b) entry 18 (c) and entry 4 (d) from Table 2.
13
Fig. 9 C NMR spectra (125 MHz, CDCl ) of PLA and PCL homopolymers (a, b) and random poly(LA-co-
3
CL) copolymers (b, c) prepared using 5b and 5c, representing the corresponding linkages between units for
random poly(LA-co-CL).
All these reported complexes are rare examples of group 4 catalysts for producing random
and block copolymers. The thermal behavior of P-LLA, PCL and their copolymers were
analysed by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC)
and depictedin Table 2 and Fig. 11 and Fig. S70, in Supplementary Data. As shown in fig. 11
two melting endotherms were observed. The double melting behavior is usually caused by the
2
4

ACCEPTED MANUSCRIPT
melt-recrystallization process during heating trace, due to unstable defective crystals [82].
Next we subjected MALDI-ToF-MS analysis for the low molecular weight copolymer
prepared using zirconium complex 5b and observed major signals corresponding to linear
polymers with isopropoxide end group see Fig. S71, in Supplementary Data.
1
3
Fig. 10 Carbonyl range of C NMR spectra (125 MHz, CDCl ) of copolymer of entry 14 (a) entry 10 (b) entry
3
7
(c) and entry 4 (d) from Table 2.
nd
Fig. 11 DSC traces (heating rate: 10 °C/min, 2 scan) of copolymers prepared using 2b and 5b.
2
5

ACCEPTED MANUSCRIPT
2
.8 Ring opening polymerization of rac-epoxides
After completion of our initial reactivity studies on lactide, we focused on
homopolymerization of rac-PO, rac-CHO and rac-SO using these group 4 complexes
as catalysts. The polymerization reactions were performed under solvent free
condition with a monomer to catalysts molar ratio of 1000:1 (15.9 mmol :15.9 ꢀ mol,
e.g. 10 mg of 1b). However, the temperatures of the reactions were chosen on the
basis of observed reactivity of monomers as shown in Table 3. As depicted in Fig. 12,
the Zr and Hf complexes 5b and 5c showed unique reactivity towards ROP of rac-PO,
-1
-1
-1
rac-CHO and rac-SO (TOF = 495 min for CHO, 241 min for PO and 128 min for
SO). The activity of these complexes were comparable to the activity of other group 4
compounds reported from our group [52,53,55,57]. In our studies CHO conversion
was found to be faster than PO and SO monomers (see Table 3). The difference in
reactivity of these epoxides is due to the bicyclic nature of CHO that has higher steric
strain compared to other epoxides. In order to release this extra strain CHO undergo
faster polymerization than the other two epoxides (PO and SO) [83]. The complexes
5
b
and 5c showed high conversion within two minutes under solvent free condition at
room temperature for CHO polymerization. The number-average molecular weight
Mn(expt) ) and molecular weight distributions (MWDs) values of polymer products were
(
measured by GPC analysis (Fig. S72 & 73, in Supplementary Data). As observed in lactide
polymerization, here also Ti complexes showed less activity than Zr and Hf complexes
1
13
(
hafnium was almost similar to zirconium, see Table-3). For all polymer samples H and C
NMR was recorded (Fig. 13 and Fig. S74-76, in Supplementary Data). Some of these
polymers are amorphous in nature and others are oily. All these complexes produced stereo-
1
3
irregular (atactic) polyethers as confirmed from C NMR.
2
6

ACCEPTED MANUSCRIPT
Table 3. Polymerization data for epoxides catalyzed by complexes 1a
-
1c
,
2a
-
2c
,
3a
-
3c
,
4a
-
4c and 5a
-
5
c
a
b
(GPC)c
Time
min)
Yield
TOF
Mn
d
Entry
Cat
monomer
Temp
-1
M /Mn
w
Tg
(
(%)
(min ) (kg/mol)
1
2
3
4
5
6
7
8
9
1a
2a
3a
4a
5a
1b
2b
3b
4b
5b
1c
2c
3c
4c
5c
1a
2a
3a
4a
5a
1b
2b
3b
4b
5b
1c
2c
3c
4c
5c
1a
2a
3a
4a
5a
1b
2b
3b
4b
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
CHO
PO
24.2
11.2
18.4
7.0
100
100
100
24
90
94
94
90
95
91
95
93
99
99
90
95
92
91
98
72
75
70
70
82
85
81
75
82
84
75
80
78
77
82
70
72
72
74
74
75
81
75
81
37.26
84.05
51.20
128.00
228.86
78.00
124.50
105.75
247.00
495.00
128.91
105.00
154.00
267.65
425.00
11.05
33.04
10.97
22.00
54.39
33.00
73.00
33.05
117.00
241.17
18.04
38.00
88.08
96.25
127.14
12.69
22.80
09.75
22.15
32.54
18.00
32.00
18.83
81.00
72.65
66.37
62.20
77.41
81.32
73.07
141.65
92.41
81.57
89.21
79.95
124.54
90.75
97.55
120.41
22.81
28.74
35.41
33.85
40.83
37.62
61.64
51.41
48.34
52.37
44.64
47.04
50.21
52.81
68.41
91.24
108.41
147.65
122.61
110.51
81.60
88.31
121.84
117.30
1.09
1.21
125
66.2
66.5
67.1
68.4
67.9
67.8
70.8
66.1
70.1
70.7
70.5
65.7
69.4
68.1
70.4
nd
1.14
1.11
1.01
1.38
1.04
1.19
1.15
1.13
1.27
1.10
1.18
1.27
1.31
1.25
1.23
1.24
1.12
1.21
1.29
1.17
1.19
1.12
1.20
1.28
1.17
1.22
1.26
1.33
1.27
1.23
1.28
1.22
1.31
1.24
1.21
1.25
4.2
24
11.8
7.6
60
55
8.8
55
4.0
24
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
2.0
24
6.9
24
9.0
24
5.9
24
3.4
24
2.0
24
65.2
22.7
63.8
32.0
15.1
28.0
11.0
22.7
7.0
45
PO
45
nd
PO
45
nd
PO
45
nd
PO
45
nd
PO
45
nd
PO
45
nd
PO
45
nd
PO
45
nd
PO
3.5
45
nd
PO
41.6
21.0
8.9
45
nd
PO
45
nd
PO
45
nd
PO
8.0
45
nd
PO
6.5
45
nd
SO
55.2
31.6
73.9
33.4
22.7
40.0
25.0
39.8
10.0
100
100
100
100
100
100
100
100
100
nd
SO
nd
SO
nd
SO
nd
SO
nd
SO
nd
SO
nd
SO
nd
SO
nd
2
7

ACCEPTED MANUSCRIPT
4
4
4
4
4
4
0
1
2
3
4
5
5b
1c
2c
3c
4c
5c
SO
SO
SO
SO
SO
SO
6.6
100
100
100
100
100
100
85
70
75
75
82
80
128.36
8.83
138.74
100.42
145.22
111.68
121.71
147.62
1.20
1.25
1.32
1.19
1.20
1.24
nd
nd
nd
nd
nd
nd
79.3
40.0
34.2
24.7
14.3
19.00
21.96
33.20
55.90
-
3
Polymerization conditions: [M] /[C] = 1000, [M] = 14.3 mM, [C] = 14.3×10 mM, under sovent-free.
Based on crude polymer weight and H NMR analysis.
TOFs calculated as (mol of epoxide consumed)/(mol of catalyst × reaction time).
Measured by GPC at 27 °C in THF relative to plyostyrene standards.
Tg determined from the second heating cycles of DSC. nd = not determined for the polyethers from PO and SO.
o
o
o
o
a
b
1
c
d
1
3
In the C NMR of PPO, multiple resonances in the regions of 77-70 ppm assignable to
methine and methylene carbons were observed which confirms the irregularity (Fig. S75 in
Supplementary Data). MALDI-ToF-MS analysis of low molecular weight oligomer prepared
using zirconium complex 5b showed major signals corresponding to linear polymers. The
expected isopropoxide or tert-butoxide end-capped polymers were observed from MALDI
spectra (Fig. S77 in Supplementary Data).
Fig. 12 Turnover frequencies (TOF) for the homopolymerization of various epoxides catalysed by group 4
complexes.
Thermal analysis of a polymer sample by differential scanning calorimetry (DSC) showed the
o
glass-transition temperature (T ) at around 67 to 70 C (Fig. S78, in Supplementary Data).
g
The polyethers with low Mns (below 3 kg/mol) have different applications from those
with high Mns (over 10 kg/mol) [84]. We screened the complexes 5b and 5c from low
monomer ratio to high, varying the initial [rac-epoxide] : [cat] ratio from 200 : 1 to
1
000 : 1 that gave a linear relationship between Mn and [rac-epoxide] : [cat]ₒ ratio
t
(
Fig. S79, in Supplementary Data).
2
8