Titanium alkoxides as initiators for the controlled polymerization of lactide

Article abstract of DOI:10.1021/ic026139n

Fourteen titanium alkoxides were synthesized for comparison of their catalytic properties in the bulk and solution polymerization of lactide (LA). In bulk polymerizations, they are effective catalysts in terms of polymer yield and molecular weight. Titanatranes gave polylactides with significantly increased molecular weight over more extended polymerization times, and those with five-membered rings afforded polymers in higher yields and with larger molecular weights than their six-membered ring counterparts. Steric hindrance of the rings was found to significantly affect polymer yields. Increased heterotactic-biased poly(rac-LA) was formed as the number of chlorine atoms increased in TiCl_x(O-i-Pr)_4-x. In solution polymerizations, titanium alkoxides catalyzed controlled polymerizations of LA, and end group analysis demonstrated that an alkoxide substituent on the titanium atom acted as the initiator. That polymerization is controlled under our conditions was shown by the linearity of molecular weight versus conversion. A tendency toward formation of heterotactic-biased poly(rac-LA) was observed in the solution polymerizations. The rate of ring-opening polymerization (ROP) and the molecular weight of the polymers are greatly influenced by the substituents on the catalyst, as well as by factors such as the polymerization temperature, polymerization time, and concentration of monomer and catalyst.

Full text of DOI:10.1021/ic026139n

Inorg. Chem. 2003, 42, 1437−1447
Titanium Alkoxides as Initiators for the Controlled Polymerization of
Lactide
Youngjo Kim, G. K. Jnaneshwara, and John G. Verkade*
Department of Chemistry, Iowa State UniVersity, Ames, Iowa 50011
Received October 29, 2002
Fourteen titanium alkoxides were synthesized for comparison of their catalytic properties in the bulk and solution
polymerization of lactide (LA). In bulk polymerizations, they are effective catalysts in terms of polymer yield and
molecular weight. Titanatranes gave polylactides with significantly increased molecular weight over more extended
polymerization times, and those with five-membered rings afforded polymers in higher yields and with larger molecular
weights than their six-membered ring counterparts. Steric hindrance of the rings was found to significantly affect
polymer yields. Increased heterotactic-biased poly(rac-LA) was formed as the number of chlorine atoms increased
in TiClx(O-i-Pr)4-x. In solution polymerizations, titanium alkoxides catalyzed controlled polymerizations of LA, and
end group analysis demonstrated that an alkoxide substituent on the titanium atom acted as the initiator. That
polymerization is controlled under our conditions was shown by the linearity of molecular weight versus conversion.
A tendency toward formation of heterotactic-biased poly(rac-LA) was observed in the solution polymerizations. The
rate of ring-opening polymerization (ROP) and the molecular weight of the polymers are greatly influenced by the
substituents on the catalyst, as well as by factors such as the polymerization temperature, polymerization time, and
concentration of monomer and catalyst.
Introduction
distributions. Despite the fact that some excellent initiators
have been reported for the polymerization of LA, the search
for new catalysts that generate well-defined polylactides
remains of keen interest. The roles of the structure of metal
alkoxide complexes in determining molecular weights and
molecular weight distributions, as well as the polymerization
pathway, are significant current research issues.
In view of the well-known significant number of similari-
ties in the chemistries of tin and titanium, we were somewhat
surprised to note that, until our entry into this area, no
Polylactide (PLA) polymers are biodegradable renewable
materials, and they are proving to be valuable, for example,
as packaging films, diapers, and a variety of medical implant
devices including matrices for the slow release of pharma-
ceuticals.1
The ring-opening polymerization (ROP) of cyclic esters
such as lactide (LA) and ꢀ-caprolactone (CL) with metal
complexes has been intensively studied over the past few
decades.2 Many types of metal alkoxides (e.g., of tin,3
aluminum,4 zinc,5 magnesium,5 iron,6 lanthanide,7 and lithium8
organometallic complexes) have been found to be active LA
polymerization catalysts, and many afford materials with
controlled molecular weights and narrow molecular weight
(3) (a) Nijenhuis, A. J.; Grijpma, D. W.; Pennings, A. J. Macromolecules
1992, 25, 6419. (b) Kricheldorf, H. R.; Lee, S.-R.; Bush, S.
Macromolecules 1996, 29, 1375. (c) Schwach, G.; Coudane, J.; Engel,
R.; Vert, M. J. Polym. Sci., Part A: Polym. Chem. 1997, 35, 3431.
(d) Atthoff, B.; Trollsås, M.; Claesson, H.; Hedrick, J. L. Macromol.
Chem. Phys. 1999, 200, 1333. (e) Duda, A.; Penczek, S.; Kowalsik,
A.; Libiszowski, J. Macromol. Symp. 2000, 153, 41. (f) Kricheldorf,
H. R. Macromol. Symp. 2000, 153, 55. (g) Kowalski, A.; Libiszowski,
J.; Duda, A.; Penczek, S. Macromolecules 2000, 33, 1964. (h)
Stridsberg, K.; Ryner, M.; Albertsson, A. C. Macromolecules 2000,
33, 2862. (i) Dove, A. P.; Gibson, V. C.; Marshall, E. L.; White, A.
J. P.; Williams, D. J. Chem. Commun. 2001, 283. (j) Mo¨ller, M.;
Nederberg, F.; Lim, L. S.; Kånge, R.; Hawker, C. J.; Hedrick, J. L.;
Gu, Y.; Shah, R.; Abbott, N. L. J. Polym. Sci., Part A: Polym. Chem.
2001, 39, 3529. (k) Ryner, M.; Finne, A.; Albertsson, A. C.;
Kricheldorf, H. R. Macromolecules 2001, 34, 7281. (l) Aubrecht, K.
B.; Hillmyer, M. A.; Tolman, W. B. Macromolecules 2002, 35, 644.
(m) Finne, A.; Albertsson, A. C. Biomacromolecules 2002, 3, 684.
* To whom correspondence should be addressed. E-mail: jverkade@
iastate.edu. Fax: 515-294-0105.
(1) (a) Chiellini, E.; Solaro, R. AdV. Mater. (Weinheim, Ger.) 1996, 8,
305-313. (b) Tullo, A. Chem. Eng. News 2000, 78, 13. (c) Bogaert,
J. C.; Coszach, P. Macromol. Symp. 2000, 153, 287. (d) Vert, M.
Macromol. Symp. 2000, 153, 333. (e) Drumright, R. W.; Gruber, P.
R.; Henton, D. E. AdV. Mater. (Weinheim, Ger.) 2000, 12, 1841. (f)
Dorgan, J. R.; Lehermeier, H. J.; Palade, L. I.; Cicero, J. Macromol.
Symp. 2001, 175, 55.
(2) See: (a) Kuran, W. Prog. Polym. Sci. 1998, 23, 919. (b) O’Keefe, B.
J.; Hillmyer, M. A.; Tolman, W. B. J. Chem. Soc., Dalton Trans. 2001,
2215 and references therein.
10.1021/ic026139n CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/11/2003
Inorganic Chemistry, Vol. 42, No. 5, 2003 1437

Kim et al.
investigations of titanium alkoxides as potential catalysts in
LA polymerization had been reported.9 Interestingly, such
compounds are well-known to homogeneously catalyze olefin
polymerization,10 and some Ti complexes have been found
to polymerize CL.11 Recently, we9 reported that several
titanium alkoxides showed reasonably good catalytic activity
in the bulk homopolymerization of l-LA and rac-LA at 130
°C. Harada et al.11d also reported the living polymerization
of LA by a Ti chloride complex, whose chloride apparently
plays the same role as an alkoxide.
We thus believed it would be interesting to test titanium
catalysts with well-defined ligand environments amenable
to systematic variation, thus potentially facilitating more
control of the molecular and physical properties of the PLA
produced. Here, we describe discrete titanium alkoxide
complexes for the study of the ROP of LA and CL under
bulk and solution polymerization conditions. We also
demonstrate that some of our titanium alkoxide catalysts
allow well-controlled polymerizations of LA and CL, and
that some of these catalysts give heterotactic-biased PLA
derived from rac-LA.
The strategy we employed for choosing candidate titanium
catalysts 1-14 is that, first of all, they should contain
alkoxide groups, because such metal substituents have proven
to be the initiation group of choice in the vast majority of
previous studies. Second, the initiating alkoxide group should
dissociate relatively easily from the titanium in the early stage
of polymerization so that it can be utilized to initiate LA
polymerization and provide a means of controlling the
molecular weight by functioning as an end group. Alkoxy
titanatranes seemed well-suited to these purposes because
they possess a transannular Ti-N bond that could potentially
labilize the trans axial OR group.
(4) (a) Dubois, P.; Jacobs, C.; Jerome, R.; Teyssie, P. Macromolecules
1991, 24, 2266. (b) Spassky, N.; Wisniewski, M.; Pluta, C.; Le Borgne,
A. Macromol. Chem. Phys. 1996, 197, 2627. (c) Kowalski, A.; Duda,
A.; Penczek, S. Macromolecules 1998, 31, 2114. (d) Emig, N.;
Nguyen, H.; Krautscheid, H.; Re´au, R.; Cazaux, J.-B.; Bertrand, G.
Organometallics 1998, 17, 3599. (e) Cameron, P. A.; Jhurry, D.;
Gibson, V. C.; White, A. J. P.; Williams, D. J.; Williams, S. Macromol.
Rapid Commun. 1999, 20, 616. (f) Ovitt, T. M.; Coates, G. W. J. Am.
Chem. Soc. 1999, 121, 4072. (g) Eguiburu, J. L.; Fernandez-Berridi,
M. J.; Coss´ıo, F. P.; Roma´n, J. S. Macromolecules 1999, 32, 8252.
(h) Ko, B. T.; Lin, C. C. Macromolecules 1999, 32, 8296. (i) Bhaw-
Luximon, A.; Jhurry, D.; Spassky, N. Polym. Bull. 2000, 44, 31. (j)
Kitayama, T.; Yamaguchi, H.; Kanzawa, T.; Hirano, T. Polym. Bull.
2000, 45, 97. (k) Radano, C. P.; Baker, G. L.; Smith, M. R. J. Am.
Chem. Soc. 2000, 122, 1552. (l) Ovitt, T. M.; Coates, G. W. J. Polym.
Sci., Part A: Polym. Chem. 2000, 38, 4686. (m) Jhurry, D.; Bhaw-
Luximon, A.; Spassky, N. Macromol. Symp. 2001, 175, 67. (n) Huang,
C. H.; Wang, F. C.; Ko, B. T.; Yu, T. L.; Lin, C. C. Macromolecules
2001, 34, 356. (o) Chen, H. L.; Ko, B. T.; Huang, B. H.; Lin, C. C.
Organometallics 2001, 20, 5076. (p) Liu, Y. C.; Ko, B. T.; Lin, C. C.
Macromolecules 2001, 34, 6196. (q) Chisholm, M. H.; Navarro-Llobet,
D.; Simonsick, W. J., Jr. Macromolecules 2001, 34, 8851. (r) Ovitt,
T. M.; Coates, G. W. J. Am. Chem. Soc. 2002, 124, 1316. (s) Nomura,
N.; Ishii, R.; Akakura, M.; Aoi, K. J. Am. Chem. Soc. 2002, 124,
5938.
(5) (a) Chisholm, M. H.; Eilerts, N. W. Chem. Commun. 1996, 853. (b)
Cheng, M.; Attygalle, A. B.; Lobkovsky, E. B.; Coates, G. W. J. Am.
Chem. Soc. 1999, 121, 11583. (c) Chisholm. M. H.; Eilerts, N. W.;
Huffman, J. C.; Iyer, S. S.; Pacold, M.; Phomphrai, K. J. Am. Chem.
Soc. 2000, 122, 11845. (d) Chisholm, M. H.; Huffman, J. C.;
Phomphrai, K. J. Chem. Soc., Dalton Trans. 2001, 222. (e) Cham-
berlain, B. M.; Cheng, M.; Moore, D. R.; Ovitt, T. M.; Lobkovsky,
E. B.; Coates, G. W. J. Am. Chem. Soc. 2001, 123, 3229. (f) Chisholm,
M. H.; Gallucci, J. C.; Zhen, H. Inorg. Chem. 2001, 40, 5051. (g)
Chisholm, M. H.; Navarro-Llobet, D.; Gallucci, J. Inorg. Chem. 2001,
40, 6506.
(6) (a) Stolt, M.; Sodergard, A. Macromolecules 1999, 32, 6412. (b)
O’Keefe, B. J.; Monnier, S. M.; Hillmyer, M. A.; Tolman, W. B. J.
Am. Chem. Soc. 2001, 123, 339. (c) O’Keefe, B. J.; Breyfogle, L. E.;
Hillmyer, M. A.; Tolman, W. B. J. Am. Chem. Soc. 2002, 124, 4384.
(7) (a) Stevels, W. M.; Ankone, M. J. K.; Dijkstra, P. J.; Feijen, J.
Macromolecules 1996, 29, 3332. (b) Stevels, W. M.; Ankone, M. J.
K.; Dijkstra, P. J.; Feijen, J. Macromolecules 1996, 29, 6132. (c) Simic,
V.; Spassky, N.; Hubert-Pfalzgraf, L. G. Macromolecules 1997, 30,
7338. (d) Chamberlain, B. M.; Sun, Y.; Hagadorn, J. R.; Hemmesch,
E. W.; Young, V. G., Jr.; Pink, M.; Hillmyer, M. A.; Tolman, W. B.
Macromolecules 1999, 32, 2400. (e) Yuan, M.; Xiong, C.; Li, X.;
Deng, X. J. Appl. Polym. Sci. 1999, 73, 2857. (f) Simic, V.; Pensec,
S.; Spassky, N. Macromol. Symp. 2000, 153, 109. (g) Deng, X.; Yuan,
M.; Li, X.; Xiong, C. Eur. Polym. J. 2000, 36, 1151. (h) Spassky, N.;
Simic, V.; Montaudo, M. S.; Hubert-Pfalzgraf, L. G. Macromol. Chem.
Phys. 2000, 201, 2432. (i) Chamberlain, B. M.; Jazdzewski, B. A.;
Pink, M.; Hillmyer, M. A.; Tolman, W. B. Macromolecules 2000,
33, 3970. (j) Aubrecht, K. B.; Chang, K.; Hillmyer, M. A.; Tolman,
W. B. J. Polym. Sci., Part A: Polym. Chem. 2001, 39, 284. (k)
Giesbrecht, G. R.; Whitener, G. D.; Arnold, J. J. Chem. Soc., Dalton
Trans. 2001, 923. (l) Save, M.; Schappacher, M.; Soum, A. Macromol.
Chem. Phys. 2002, 203, 889.
(8) Ko, B.-T.; Lin, C.-C. J. Am. Chem. Soc. 2001, 123, 7973.
(9) (a) Kim, Y.; Verkade, J. G. Organometallics 2002, 21, 2395. (b) Kim,
Y.; Kapoor, P. N.; Verkade, J. G. Inorg. Chem. 2002, 41, 4834. (c)
Kim, Y.; Verkade, J. G. Macromol. Rapid Commun. 2002, 23, 917.
(d) Kim, Y.; Verkade, J. G. In preparation.
(10) Gladysz, J. A., Ed. Chem. ReV. 2000, 100, 1167-1682.
(11) (a) Okuda, J.; Rushkin, I. L. Macromolecules 1993, 26, 5530. (b)
Takeuchi, D.; Nakamura, T.; Aida, T. Macromolecules 2000, 33, 725.
(c) Takeuchi, D.; Aida, T. Macromolecules 2000, 33, 4607. (d)
Takashima, Y.; Nakayama, Y.; Watanabe, K.; Itono, T.; Ueyama, N.;
Nakamura, A.; Yasuda, H.; Harada, A.; Okuda, J. Macromolecules
2002, 35, 7538.
Results and Discussion
Syntheses. Except for minor modifications (see Experi-
mental Section), compounds 2-4 were obtained by the
reaction of 1 with the appropriate amount of TiCl4 in pentane
at room temperature.12 Unlike slightly viscous 2, compounds
3 and 4 precipitated as white solids within 20 min after
combination of the reactants. To avoid generation of byprod-
ucts and mixtures of 2-4, a solution of TiCl4 in dry pentane
was added dropwise to exactly the appropriate number of
equivalents of 1 in pentane while stirring rapidly at room
(12) Kamigaito, M.; Sawamoto, M.; Higashimura, T. Macromolecules 1995,
28, 5671.
1438 Inorganic Chemistry, Vol. 42, No. 5, 2003

Ti Alkoxides Initiating Lactide Polymerization
temperature for 10 min. After washing thoroughly with cold
pentane, these titanium compounds were used as catalysts
for making PLA or as starting materials for other titanium
alkoxides. Compounds 2-4 were soluble in toluene, di-
chloromethane, and ether. Displacement of the chloride from
2 with CpNa in toluene gave the desired product 14 as a
reddish-yellow oil.
Our group13 and Do’s group14 have synthesized several
types of titanatranes by an amine or chloride displacement
with an alkoxide or triethanolamine. Here, we synthesized
the new titanatranes 6-9 in very good yield in a one-pot
reaction containing Ti(O-i-Pr)4, triethanolamine, and the
appropriate phenol. These compounds were also synthesized
by a two-step reaction in which 5 is made from Ti(O-i-Pr)4
and triethanolamine, followed by reaction with the corre-
sponding phenol. However, yields were lower than from a
one-pot reaction (see Experimental Section). Although 6 and
9 showed good solubilities, 7 and 8 displayed limited
solubility in a wide variety of solvents. Despite the limited
solubility of 7 and 8, we were able to isolate these
compounds in analytically pure form. We previously showed
that 10 could be synthesized in 70% overall yield by reacting
2 equiv of (diethylamino)titanatrane with 1 equiv of pinacol
starting from tetrakis(diethylamino)titanium and triethanol-
amine.13 However, in the present work, we found that the
reaction between 2 equiv of 5 and 1 equiv of pinacol at room
temperature generated 10 in an overall yield of 84% with a
starting material (5) that is less expensive than tetrakis-
(diethylamino)titanium.
We first attempted to synthesize 11, 12, and 13 by trans-
esterifying 1 with the tris(phenol)s tris(2-hydroxyphenyl)-
amine (2,2′,2′′-nitrilotriphenol, THA),15 tris(2-hydroxy-3,5-
dimethylbenzyl)amine (THDA),16 and tris(2-hydroxy-3-tert-
butyl-5-methylbenzyl)amine (THBA),16 respectively. However,
11-13 obtained in this manner were contaminated with
i-PrOH despite extended drying under vacuum. For example,
11 consistently retained 0.5 equiv of i-PrOH after several
recrystallizations from THF/toluene. In searching for a new
synthetic route to 11-13, 4 was chosen as the starting
material. We had anticipated that the reaction of 4 with
THDA in the presence of triethylamine under mild conditions
could involve chloride displacement from the metal by
THDA. This prediction was based on a somewhat lower Ti-
Cl bond dissociation energy (430 kJ/mol for TiCl4,17)
compared with the analogous value for the Ti-O-i-Pr bond
[444 kJ/mol for Ti(O-i-Pr)417]. However, the product obtained
in the reaction of 4 with THDA was the corresponding
Figure 1. ORTEP drawing of 7 showing 50% probability thermal ellipsoids
with H atoms and solvent omitted for clarity. Selected bond distances (Å):
Ti1-O1 ) 1.977(2), Ti1-O2 ) 2.0588(19), Ti1-O3 ) 1.8261(19), Ti1-
O4 ) 1.8320(19), Ti1-N1 ) 2.284(2), Ti1-Ti1′ ) 3.3107(11), Ti1-O2′
) 1.9881(18). Selected bond angles (deg): O1-Ti1-N1 ) 128.95(8), O1-
Ti1-O2 ) 155.37(7), O1-Ti1-O3 ) 90.46(9), O1-Ti1-O4 ) 85.42(8),
O2-Ti1-O3 ) 100.87(9), O1-Ti1-O4 ) 138.29(9).
chlorotitanatrane, which was also reported recently18 to form
in the reaction of CpTiCl3 with THDA wherein the η5-Cp
ligand is displaced. Thus, as was reported earlier by others,19
we synthesized 12 in high yield by reacting 14 with THDA,
a reaction that also occurs by displacement of a η5-Cp ligand
(see Experimental Section). The C5-Ti bond dissociation
energy associated with the η5-Cp-Ti moiety is 335 kJ/mol
[an interpolated value between that of η5-CpTiCl3 and (η5-
Cp)2TiCl217]. The leaving ability of these groups in their
displacement by THDA decreases in the order Cp > O-i-Pr
> Cl in which the last two members appear reversed
considering only the bond dissociation energies involved.
Because the difference between these latter two energies is
only about 3%, greater relief of steric strain by departure of
an O-i-Pr compared with that of a Cl substituent may
dominate their leaving ability order. Compounds 11 and 13
were made starting from 1 (see Experimental Section). All
the compounds evaluated as catalysts in the present work
were prepurified by recrystallization or distillation.
Crystal and Molecular Structure of 7. The ORTEP
depiction20 of the structure is shown in Figure 1, and selected
interatomic distances and angles are provided in the caption
of Figure 1 and in Table 1. The X-ray analysis of 7 reveals
that its molecular structure features an oxo bridge as is
commonly observed for titanatranes in the solid state.13,21
As the ORTEP drawing in Figure 1 illustrates, the geometry
around each titanium in 7 can be viewed as a distorted
octahedron. The four oxygen atoms of the ethylene arms form
the equatorial plane, and the axial sites are occupied by a
nitrogen atom of the tripodal ligand and an oxygen atom of
the pentafluorophenoxy group.
Although the only difference between 5 and 7 is the Z
substituent of the alkoxide, the coordination geometry of 7
differs substantially from the one in compound 521 in that
the axial Z substituent in 5 is trans to a nitrogen, whereas 7
is derived from 5 by a twist indicated by the curved arrows
(13) (a) Menge, W. M. B. P.; Verkade, J. G. Inorg. Chem. 1991, 30, 4628.
(b) Naiini, A. A.; Menge, W. M. B. P.; Verkade, J. G. Inorg. Chem.
1991, 30, 5009. (c) Naiini, A. A.; Ringrose, S. L.; Su, Y.; Jacobson,
R. A.; Verkade, J. G. Inorg. Chem. 1993, 32, 1290.
(14) (a) Kim, Y.; Hong, E.; Lee, M. H.; Kim, J.; Han, Y.; Do, Y.
Organometallics 1999, 18, 36. (b) Kim, Y.; Han, Y.; Hwang, J. W.;
Kim, M. W.; Do, Y. Organometallics 2002, 21, 1127.
(15) Frye, C. L. Fr. Patent 1511256, 1968; Chem. Abstr. 70: 87303.
(16) (a) Chandrasekaran, A.; Day, R. O.; Holmes, R. R. J. Am. Chem. Soc.
2000, 122, 1066. (b) Timosheva, N. V.; Chandrasekaran, A.; Day, R.
O.; Holmes, R. R. Organometallics 2000, 19, 5614. (c) Timosheva,
N. V.; Chandrasekaran, A.; Day, R. O.; Holmes, R. R. Organometallics
2001, 20, 2331.
(18) Michalczyk, L.; de Gala S.; Bruno, J. W. Organometallics 2001, 20,
5547.
(19) Kol, M.; Shamis, M.; Goldberg, I.; Goldschmidt, Z.; Alfi, S.; Hayut-
Salant, E. Inorg. Chem. Commun. 2001, 4, 177.
(20) Johnson, C. K. ORTEP; Report ORNL-5138; Oak Ridge National
Laboratory, Oak Ridge, TN, 1976.
(17) Connor, J. A. Top. Curr. Chem. 1977, 71, 71.
(21) Harlow, R. L. Acta Crystallogr. 1983, C39, 1344.
Inorganic Chemistry, Vol. 42, No. 5, 2003 1439

Kim et al.
Table 1. Comparison of Bond Angles (deg) and Bond Distances (Å) in Dimeric titanatranes
structural param
5
7
15
16
Obridge-Ti-Z (deg)
Nbridgehead-Ti-Obridge (deg)
Ti-Obridge-Ti′ (deg)
Ti-Ti (Å)
113.20(5)
75.23(4)
109.66(5)
3.356(1)
2.333(1)
2.053(1)
1.854(1)
21
155.37(7)
145.40(8)
109.77(8)
3.3107(11)
2.284(2)
2.0235(19)
1.8784(19)
this work
171.03(6)
148.44(9)
107.40(9)
3.273(1)
2.269(3)
2.029(2)
1.875(1)
13c
179.46(6)
151.10(6)
105.01(5)
3.2547(6)
2.270(2)
2.050(1)
1.845(1)
13b
Ti-Nbridgehead (Å)
Ti-Obridge (Å)
Ti-Oterminal (Å)a
ref
a Average value.
in the transformation shown in Table 1. Previously, we
suggested that a more electron-donating substituent Z such
as S-i-Pr or NMe2 in a titanatrane (15 and 16, respectively,
in Table 1) would prefer to occupy a position trans to a
bridging oxygen rather than to the more electron-donating
tertiary bridgehead nitrogen atom.13 However, no investiga-
tions of an oxo-bridged dimeric titanatrane bearing a strongly
electron-withdrawing substituent Z have been reported. It
appears that influences such as subtle steric and/or crystal
packing effects may play a role in determining solid state
conformations of these compounds.
Comparisons of pertinent bond angles and bond distances
in the dimeric titanatranes 5, 7, 15, and 16 are shown in
Table 1. As the substituent Z changes from 5 to 7, to 15,
and to 16, a widening of the Obridge-Ti-Z bond angles
[113.20(5)° in 5 to 179.46(6)° in 16] occurs with concomitant
opening of the Nbridgehead-Ti-Obridge bond angles [75.23(4)°
in 1 to 151.10(6)° in 16]. Even though 7 would be expected
to have the shortest Ti-Nbridgehead bond length and the longest
Ti-Ti distance, the actual values curiously fall between these
bond lengths for 5 and those for 15 or 16. Like the situation
for other oxo-bridged titanatrane dimers,13,21 substantial Ti-
Ti interaction was not indicated to be present in 7. Also
worthy of mention is that 7 has the shortest Ti-Obridge and
the longest Ti-Oterminal bond length of any oxygen-bridged
dimeric titanatranes reported so far.13,21 However, the average
Ti-O bond distance for all the oxygens of 7 is similar to
the average of this distance observed for other structurally
characterized titanatranes.13,14,21 The weaker interaction of
be described as being bound together by predominantly
alkoxy-type bridging bonds (similar to those in 15 and 16
in Table 1) while the analogous halves of 5 are bridged by
predominantly donor-acceptor-like linkages.
NMR Spectra. The chemical shift of the i-Pr group in
1
the H NMR spectra of 1-4 moves downfield in the same
order as expected on the basis of the electron-withdrawing
effect of the chlorides. Titanatranes 5-10 can be divided
1
into four categories on the basis of their H and 13C{1H}
1
NMR spectra. In the first category is 5, whose solution H
and 13C{1H} NMR spectra are quite temperature-independent,
displaying sharp resonances for two types of CH2CH2O
protons. The spectra are consistent with monomeric behavior
in solution, though it is a dimer in the solid state. Compounds
1
6-8 constitute a second category, for which the H and
13C{1H} NMR spectra are broadened at room temperature,
presumably owing to an exchange process that is relatively
slow on the NMR time scale. Because dilution of solutions
of these dimers does not affect the breadth of their 1H NMR
spectra, the exchange process can be envisioned as being
dominated by an intramolecular “gearing” type of fluxionality
around their Z-Ti-N axes rather than by dissociation into
monomers.13 This gearing motion requires the breakage of
only one bridge bond at a time with the subsequent formation
of a new one as opposite rotations about the Z-Ti-N axes
occur. Thus, 1H NMR spectroscopic data for 7 in CDCl3 are
consistent with retention of the dimeric unit in solution.
Room temperature 1H and 13C{1H} NMR spectra of 9 display
sharp resonances which are consistent with the presence of
two types of CH2CH2O groups in a 1:1 ratio. Because of
the bulky nature of the apical substituent Z, this compound
is monomeric in solution as well as in the solid state, thus
representing a third category of titanatranes. Dimeric com-
Oterminal with the titanium atom can be ascribed to repulsions
from the combined effect of robust Ti-Obridge bonding and
a strong transannular interaction from the bridgehead nitro-
gen. These observations suggest that the halves of 7 could
1440 Inorganic Chemistry, Vol. 42, No. 5, 2003

Ti Alkoxides Initiating Lactide Polymerization
Table 2. Bulk Polymerization of LA at 130 °C
pound 10 is a member of a trivial fourth category for which
1
the H and 13C{1H} NMR spectra are sharp over a wide
lactide
type
yield
(%)
b
b
entry catalysta
time
2 hr
Mw
Mn
PDIb
temperature range.
1
2
3
4
5
1
2
3
4
l-LA
75
71
79
75
92
90
94
93
20
29
37
46
69
66
92
90
71
75
91
88
92
88
98
93
89
84
96
95
84
82
94
91
95
91
95
94
55
53
26
24
97
94
0
35 700
41 100
40 700
38 000
36 900
34 900
79 500
97 600
c
16 000 2.24
19 600 2.10
27 600 1.47
23 800 1.60
19 900 1.85
28 400 1.23
60 900 1.31
68 600 1.42
The structures of 11-13 can be described as a 3-bladed
turbine of C3 symmetry.9a,18,19 Although the presence of a
transannular bond in 11 has not been verified, it is likely to
exist on the basis of its presence in a titanatrane18 and in
several silatranes22 containing the THA ligand. The 1H NMR
spectra of 11-13 display well-defined resonances possessing
expected integrations. Compounds 12 and 13 should have
pseudo-C3 symmetry on the NMR time scale, and there-
fore, the benzylic CH2 protons should not be equivalent.
Indeed, the proximity of these protons to the aromatic rings
may be responsible for the ca. 1.1 ppm chemical shift
difference displayed in the room temperature spectra. Upon
heating, compounds 12 and 13 assume pseudo-C3V symmetry
on the NMR time scale, and all six CH2 protons become
equivalent. Similar observations for (2,6-di-i-PrC6H3O)-
Ti(O-2,4-Me2C6H2CH2)3N (17) and i-PrOTi(O-2,4-tert-
Bu2C6H2CH2)3N (18) were reported recently by us9a and by
Kol et al.,19 respectively.
rac-LA 2 hr
l-LA 2 hr
rac-LA 2 hr
l-LA 2 hr
rac-LA 2 hr
l-LA 2 hr
rac-LA 2 hr
l-LA
l-LA
l-LA
l-LA
l-LA
rac-LA 2 h
l-LA
rac-LA 15 h
l-LA
rac-LA 4 h
l-LA
rac-LA 15 h
l-LA
rac-LA 4 h
l-LA
rac-LA 15 h
l-LA
rac-LA 4 h
l-LA
rac-LA 15 h
l-LA
rac-LA 4 h
l-LA
rac-LA 15 h
l-LA
rac-LA 15 h
l-LA 4 h
rac-LA 4 h
l-LA 4 h
rac-LA 4 h
6
7
8
9
1
2
3
4
5
30 min
c
c
c
c
c
c
c
c
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
30 min
30 min
30 min
2 h
c
c
c
135 400
132 600
270 300
303 600 119 200 2.55
73 500
38 900
188 200 101 100 1.86
136 500
91 100
80 000 1.69
78 200 1.70
73 500 3.68
15 h
6
7
8
9
4 h
39 500 1.86
21 800 1.78
15 h
76 900 1.78
63 200 1.44
39 500 1.56
94 500 2.90
93 200 2.63
66 500 1.69
25 500 1.72
4 h
61 000
15 h
273 700
245 000
112 700
44 000
4 h
Bulk Polymerization of LA. Bulk ROP of LA initiated
by 1-14 was carried out at 130 °C with the [LA]/[Ti] ratio
fixed at 300, and Table 2 shows that all of these compounds
were effective catalysts. The end groups of PLA produced
by 1-14 are the corresponding alkoxy ester units as indicated
15 h
223 100 110 900 2.01
142 500
65 800
30 900
131 400
115 500
98 200
72 800
161 800
194 400
76 100
51 400
38 400
43 400
38 300
31 200
82 700 1.72
43 900 1.50
19 600 1.57
82 600 1.59
67 900 1.70
46 700 2.10
34 700 2.10
80 800 2.00
96 000 2.02
52 000 1.46
38 000 1.35
28 400 1.35
30 200 1.44
29 200 1.31
20 000 1.56
4 h
15 h
1
by H NMR spectroscopy.23 Because chloride ion is also
10
11
15 h
known to function as an initiator in LA polymerization,11d
TiCl4 was also evaluated (Table 2, entry 43). Thus, initiation
occurs through the insertion of the alkoxy group from the
titanium catalyst into l-LA or rac-LA, consistent with a
polymerization process that proceeds via a coordination-
insertion mechanism. This was further supported by homo-
12
13
l-LA
rac-LA 14 h
l-LA
rac-LA 2 h
l-LA 2 h
14 h
14
2 h
1
nuclear decoupled H NMR spectroscopy. Such spectra of
PLA derived from rac-LA display the characteristic five
methine resonances, whereas spectra of PLA derived from
l-LA exhibit only one methine peak.23 However, we believe
at this stage that some transesterification occurred during
polymerization, because the PDI values of the resulting
polymers were somewhat higher than expected for a con-
trolled polymerization.
Proceeding from 1 to 4, catalytic activity and molecular
weights generally increased, though the corresponding PDI
values generally decreased in that order (Table 2, entries
1-8). We were unable to distinguish between the relative
catalytic activities of 3 and 4 and also between those of 1
and 2 on the basis of yield at the end of 2 h. Therefore, the
polymerizations were terminated at the end of 30 min when
TiCl4
a Oil bath temperature: 130 ((3) °C, [LA]/[Ti] ) 300, 2 g of LA. b The
weight average molecular weights (Mw), the number average molecular
weights (Mn), and the polydispersity indices (PDI ) Mw/Mn) were
determined by GPC. c Not determined.
the trend in the yields for 1 to 4 was more apparent (Table
2, entries 9-12). With these catalysts, we also observed that
1
the homonuclear decoupled H NMR spectra of poly(rac-
LA) derived from 3 and 4, respectively, in Figure 2c and d
are quite different from the spectra in Figure 2a and b which
were derived from 1 and 2, respectively. This result is
consistent with that predicted from a Bernouillian analysis
of totally random poly(rac-LA).23 The methine region in the
homonuclear decoupled 1H NMR spectrum of poly(rac-LA)
derived from 3 and 4 displays rmr and mrm tetrads which
are much more intense than expected. These observations
are consistent with a heterotactic-biased poly(rac-LA) be-
cause the rmr microstructure can only arise from two
consecutive D-L or L-D interchanges. Each rmr tetrad is
accompanied by two mrm tetrads in agreement with the NMR
integration (Table 2, entries 6 and 8). The preference for
heterotacticity in our poly(rac-LA) is not as strong as
reported previously by Coates et al.5b,e and Kasperczyk et
(22) (a) Frye, C. L.; Vincent, G. A.; Hauschildt, G. L. J. Am. Chem. Soc.
1966, 88, 2727. (b) Boer, F. P.; Turley, J. W.; Flynn, J. J. J. Am.
Chem. Soc. 1968, 90, 5102. (c) Livant, P.; Northcott, J.; Webb, T. R.
J. Organomet. Chem. 2001, 620, 133.
(23) (a) Thakur, K. A. M.; Kean, R. T.; Hall, E. S.; Kolstad, J. J.; Lindgren,
T. A.; Doscotch, M. A.; Siepmann, J. I.; Munson, E. J. Macromolecules
1997, 30, 2422. (b) Thakur, K. A. M.; Kean, R. T.; Hall, E. S.;
Doscotch, M. A.; Munson, E. J. Anal. Chem. 1997, 69, 4303. (c)
Thakur, K. A. M.; Kean, R. T.; Hall, E. S.; Kolstad, J. J. Munson, E.
J. Macromolecules 1998, 31, 1487. (d) Thakur, K. A. M.; Kean, R.
T.; Zell, M. T.; Padden, B. E.; Munson, E. J. Chem. Commun. 1998,
1913.
Inorganic Chemistry, Vol. 42, No. 5, 2003 1441

Kim et al.
were observed in PLA prepared with 10 (which is a
dialkoxide-bridged bis-titanatrane) and by 11 (Table 2, entry
33-36).
Compound 10 contains a pinacolate ligand, which could
allow this catalyst to behave as a difunctional initiator in
the polymerization of LA, thus leading to a significant
increase in molecular weight than with other titanatranes.
However, no significant increase of molecular weight was
observed with the use of 10. Moreover, elemental analysis
revealed that there is less than 0.003% titanium in the
polymer obtained with this catalyst, which is far less than
the 0.049% expected had not the titanatrane structure at both
ends of the anticipated polymer chain been solvolyzed by
the excess MeOH used to isolate and purify the polymers.
In terms of yield and molecular weight, titanatrane 11 is
the most effective PLA bulk polymerization catalyst among
the 14 catalysts studied. According to an elemental analysis
of the product polymer obtained with this five-membered
ring catalyst, the residual titanium content is less than
0.001%. The question then arises as to whether six-membered
ring titanatranes with fused ortho-phenyls in the bridges may
be more active catalysts for making PLA than five-membered
ring 11, for example. Titanatranes such as 12 and 13 might
be expected to exhibit weak/long bridgehead-bridgehead
Ti-N interactions. Despite the absence of X-ray structural
data for 12 and 13, two similar compounds [(2,6-di-i-Pr-
PhO)Ti(O-2,4-Me2C6H2CH2)3N (17) and i-PrOTi(O-2,4-tert-
Bu2C6H2CH2)3N (18)] structured by us9a and by Kol et al.19
revealed Ti-N distances of 2.306(2)9a and 2.334(5) Å,19
respectively, which signify the presence of transannular
bonds that fall at the long end of the range between 2.264(3)
and 2.342(9) Å observed in other structurally characterized
titanium trialkanolamine derivatives.13,14 The data in entries
37-40 in Table 2 reveal that 12 and 13 are considerably
poorer catalysts than 5 or 11. We believe this is primarily
due to the greater steric protection of the titanium in 12 or
13, particularly in the region above the equatorial plane in
these molecules, where the methyl or tert-butyl substituent
on the six-membered rings can accentuate blockage of
titanium ligation in the coordination-insertion step. In accord
with this idea is the observation that less sterically hindered
12 shows higher activity and affords a greater polymer
molecular weight than 13. Although the PDI values associ-
ated with the PLA polymers produced by 12 and 13 were
smaller than those for the products provided by 5 and 11,
the polymer molecular weights are also considerably lower
for 12 and 13.
Figure 2. Methine region of homonuclear decoupled 1H NMR spectra
for poly(rac-LA) produced by (a) 1, (b) 2, (c) 3, and (d) 4 under bulk
polymerization conditions.
al.,24 but our tacticity bias is similar to that reported by
Gibson and co-workers.3i It is interesting that in proceeding
from 1 to 4 (Table 2, entries 2, 4, 6, and 8), the intensity of
heterotactic-biased poly(rac-LA) augments significantly for
reasons that are not obvious, especially because the precise
mechanism for preferred heterotacticity is currently unknown.
It is worth noting that titanatranes 5-9 provide PLA with
significantly increased molecular weights which are associ-
ated with the polymerization time (Table 2, entries 13-32).
The same result was also observed in the syndiospecific
polymerization of styrene using Cp*Ti(OCH2CH2)3N, in
which there is also a transannular bond from the bridgehead
nitrogen to the titanium.14 The unexpectedly high molecular
weights may be due to a faster rate of propagation than
initiation. This notion was supported by the observation of
bimodal GPC traces for some of the polymer samples. Thus,
although such traces of the polymers prepared using 5-9
with [LA]/[Ti] ) 300 were unimodal up to 80% conversion,
they began to show bimodal peaks at conversions greater
than 90%, which is consistent with the predictable effects
of transesterification processes in lactone polymerization.2
Thus, when transesterification effectively competes with
ROP, the PDI values of the resultant polymers should rise
with increasing conversion, and molecular weight distribu-
tions may be bimodal.25 The same bimodal GPC patterns
In the presence of methylaluminoxane, CpTi(OR)3 com-
pounds are well-known to catalyze the syndiospecific poly-
merization of styrene.26 Compound 14 also contains alkoxide
groups that could function as initiators of LA polymerization,
and entries 41 and 42 in Table 2 show that it too is a good
catalyst, giving PLA in high yields and with moderate PDI
and molecular weight values.
Solution Polymerization of LA and CL. Although 1-14
are catalysts for bulk polymerization of LA, the PDI values
(24) (a) Kasperczyk, J. E. Macromolecules 1995, 28, 3937. (b) Bero, M.;
Dobrzynski, P.; Kasperczyk, J. J. Polym. Sci., Part A: Polym. Chem.
1999, 37, 4038.
(25) Baran, J.; Duda, A.; Kowalski, A.; Szymanski, R.; Penczek, S.
Macromol. Rapid Commun. 1997, 18, 325.
(26) Ishihara, N.; Seimiya, T.; Kuramoto, M.; Uoi, M. Macromolecules
1986, 19, 2464.
1442 Inorganic Chemistry, Vol. 42, No. 5, 2003

Ti Alkoxides Initiating Lactide Polymerization
Table 3. Solution Polymerization of LA
Td (°C)
time
yielde (%)
Mw
Mn
PDIg
g
g
entry
catalyst
monomer type
[M]/[Ti]
1
2
3
4
5
6
7
8
9
1
l-LAa
300
300
200
200
200
200
300h
125
175
200
200
200
300
300
300
300
300
300
300
300
300
200
200
300
400
200
300
300
300
300
300
300
300
300
70
70
70
70
70
70
70
70
70
70
70
70
130
130
70
70
70
70
70
130
50
70
130
130
130
130
130
130
130
130
130
130
130
130
24 h
24 h
4 h
85
70
28 700
16 900
4600
14 300
9 200
4200
2.01
1.83
1.10
1.06
1.08
1.10
1.12
1.13
1.06
1.07
1.10
1.08
1.20
1.20
1.09
1.15
1.08
1.07
1.03
1.35
1.09
1.10
1.31
1.39
1.23
1.29
1.49
1.45
1.66
1.34
rac-LAa
l-LA b
2
15 (20)f
55 (58)f
65 (69)f
81 (85)f
78
l-LAb
16 h
20 h
24 h
36 hh
24 h
24 h
24 h
6 h
13 600
16 300
20 200
30 700
12 000
16 700
17 600
6200
12 800
15 100
18 400
27 300
10 600
15 800
16 300
5600
l-LAb
l-LAb
l-LAb
ꢀ-caprolactonec
ꢀ-caprolactonec
rac-LAb
l-LAa
52
77
87
26
28
53
48
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
3
4
5
rac-LAa
l-LAa
6 h
5600
5200
24 h
24 h
3 h
61 500
44 500
3000
51 100
37 000
2800
rac-LAa
l-LAb
7 (8)f
21 (25)f
29 (31)f
28 (35)f
58 (70)f
81
16
89
43
58
51
17
26
24
68
64
0
0
l-LAb
10 h
14 h
17 h
36 h
24 h
15 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
24 h
12 h
12 h
11 500
13 500
14 200
28 900
34 300
14 400
19 400
12 300
25 800
43 200
13 600
25 800
23 000
18 500
17 500
10 000
12 500
13 300
28 000
25 400
13 200
17 600
9400
18 600
35 200
10 500
17 400
15 800
11 100
13 100
l-LAb
l-LAb
l-LAb
l-LAb
rac-LAb
ꢀ-caprolactoneb
l-LAb
6
7
8
9
10
l-LAb
l-LAb
l-LAb
l-LAa
rac-LAa
l-LAa
11
rac-LAa
l-LAa
12
13
14
l-LAa
l-LAa
63
59
8000
9000
6700
8300
1.20
1.09
rac-LAa
a
Solvent: 40 mL of toluene, 2 g of LA. b Solvent: 30 mL of toluene, 2 g of LA. c Solvent: 30 mL of toluene, 2 g of ꢀ-caprolactone. d Oil bath
temperature: 130 ((3) °C. e Isolated yield. f Conversion determined via integration of the methine resonances of LA and PLA (CDCl3). g See Table 1,
footnote b. h Continuation of polymerization after equilibrium was established (see Results and Discussion section).
(1.31-3.68) of the polymers obtained are higher than the
ideal value of 1.00 for a completely controlled polymeriza-
tion. Therefore, we investigated the possibility that better
polymerization control might be achieved under solution
conditions. In Table 3 are shown the results of our experi-
ments carried out in toluene solution with different [LA]/
[Ti] ratios, with various polymerization times, and at
temperatures higher than 50 °C (owing to a lack of sufficient
solubility of the monomer and polymers at lower tempera-
tures27). The yields and molecular weights of the polymers
obtained using 1-14 in toluene were inferior compared with
those synthesized in the bulk polymerization experiments.
The PDI values of the polymers obtained using catalysts other
than 1 and 10-13 in Table 3 are quite small, indicating a
substantial degree of molecular weight control. Even though
the polymerizations were carried out in toluene, 1 gave rise
to PLA with a high PDI value, which may be associated
with the dissociation of more than one O-i-Pr group from 1,
thus generating more than one initiating site (Table 3, entries
1-2). On the other hand, 2 and 3, which contain more than
one chlorine atom, generated PLA with very narrow PDI
values (entries 3-12, Table 3). This suggests that the
presence of chlorine atoms in chlorotitanium alkoxides may
permit only one O-i-Pr group to dissociate. In keeping with
this suggestion, the methine region in the homonuclear
decoupled 1H NMR spectrum of poly(rac-LA) derived from
4 displayed rmr and mrm tetrads that were much more
intense than expected [as was also observed under bulk
polymerization conditions for this catalyst (Figure 3d)].23 In
progressing from 1 to 4 (Table 3, entries 2, 10, 12, and 14),
the intensity in the 1H NMR spectra of heterotactic-
biased poly(rac-LA) increased, although to a lesser degree
than in such spectra of the bulk polymerization product
(Figure 3).
For a further investigation of the degree of control in these
polymerizations, we selected catalysts 2 and 5. The PDI
values of PLA obtained with 2 ranged from 1.06 to 1.12.
These values vary linearly with Mn and with conversion as
shown in Figure 4 (Table 3, entries 3-6), implying a very
substantially controlled polymerization process. The con-
trolled nature of these polymerizations was further confirmed
by a polymerization resumption experiment that resulted in
further ROP of LA. In this experiment (Table 3, entry 7),
an additional 100 equiv of LA monomer was added to the
reaction medium corresponding to that of entry 6 in Table
3. The GPC traces in Figure 5 show that the molecular weight
increased for the final polymer (peak e, Mn ) 27 300, PDI
) 1.12) relative to the initial product (peak d, Mn ) 18 400,
PDI ) 1.10).
(27) Although these compounds did catalyze the polymerization of LA at
room temperature in toluene and also in THF, yields, molecular
weights, and PDI values were not sufficiently reproducible.
Inorganic Chemistry, Vol. 42, No. 5, 2003 1443

Kim et al.
Figure 5. GPC traces of isolated PLA produced with 2 ([LA]/[Ti] ) 200)
at (a) 4 h, (b) 16 h, (c) 20 h, (d) 24 h, and (e) 36 h. The corresponding PDI
values are 1.10, 1.06, 1.08, 1.10, and 1.12, respectively.
back-biting reactions do not occur to any appreciable extent
under our conditions. This conclusion was further verified
by the following observations. First, the homonuclear de-
coupled 1H NMR spectrum reveals only one resonance and
five resonances in the methine region for poly(l-LA) and
poly(rac-LA), respectively.23 Thus, if back-biting reactions
had occurred, side peaks would have been observed. Second,
if intermolecular cyclization reactions had taken place during
polymerization, the PDI values of the resulting PLA would
have been substantially higher than the nearly ideal values
we observed. Interestingly, epimerization of the chiral centers
in poly(l-LA) apparently does not occur to a detectable extent
Figure 3. Methine region of the homonuclear decoupled 1H NMR spectra
for poly(rac-LA) produced by (a) 1, (b) 2, (c) 3, and (d) 4 under solution
polymerization conditions.
1
according to the homonuclear decoupled H NMR spectra
for the methine region.
In the case of 5, PLA polymers with narrow PDI values
were obtained from reactions conducted with a fixed [LA]/
[Ti] ratio of 300 at 70 °C. The linear relationship between
Mn and the conversion shown in Figure 7 (Table 3, entries
15-19) implies that polymerization was substantially con-
trolled. To aid in understanding the initiating process in PCL
1
formation initiated by 5, H NMR studies were carried out
(Table 3, entry 22). These spectra indicate that initiation
occurs through the insertion of an O-i-Pr group from 5 into
the CL molecule, giving a titanium alkoxide intermediate.
This observation accords with our expectation that the
polymer chain should possess an i-Pr ester and a hydroxy
end group.
Figure 4. Plot of Mn and PDI values (employing polystyrene standards)
for PLA as a function of conversion at 70 °C in toluene with [LA]/[Ti] )
200 using 2 as the catalyst. Mn (theory) was calculated from the formula
(Mw of LA) × ([LA]/[Ti]) × (conversion).
In an effort to better understand the initiating process, 1H
NMR studies on PCL [poly(ꢀ-caprolactone)] formation
initiated by 2 were carried out as shown in Figure 6. The 1H
NMR spectrum of PCL indicates that initiation occurs
through the insertion of an O-i-Pr group from compound 2
to CL, giving a titanium alkoxide intermediate that further
reacts with excess CL (ꢀ-caprolactone) giving PCL (Table
3, entries 8 and 9). This result is in agreement with our
expectation that the polymer chain should terminate with one
i-Pr ester and one hydroxy end group, the latter arising from
methanolysis of the metal alkoxide terminus present during
polymerization.
At the higher polymerization temperature of 130 °C and
at the higher [LA]/[Ti] ratio of 300, the PDI value for PLA
using initiator 5 increases (Mn ) 25 400, PDI ) 1.35, Table
3, entry 20), but a significant increase in polymer yield did
not occur. This tendency toward increased PDI values at this
temperature was also observed for titanatranes 6-11 (Table
3, entry 23-30) along with dramatically decreased polymer
yields. At the higher polymerization temperatures, ti-
tanatranes increasingly initiate transesterification reactions,
which could account for the augmented PDI values. Interest-
ingly, 12 and 13 did not show any catalytic activities for the
solution polymerization of LA though they showed moderate
catalytic activity in bulk polymerizations. In addition, 14 gave
This experiment was also carried out for PLA obtained
with catalyst 2 with the analogous result, suggesting that
1444 Inorganic Chemistry, Vol. 42, No. 5, 2003

Ti Alkoxides Initiating Lactide Polymerization
Figure 6. 1H NMR spectrum of PCL synthesized in toluene at 70 °C along with a typical GPC trace (PDI ) 1.06).
two categories: simple tetracoordinate titanium complexes
(1-4 and 14) and pentacoordinate complexes (5-13) derived
from tetradentate trisalkoxy- or trisaryloxyamine podands.
Additionally, pentacoordinate 5-13 break down into a set
of three five-membered ring titanatranes (5-11) and three
six-membered ring analogues (12 and 13). Although the
different ring sizes result in only minor structural changes
in these compounds, they give rise to a major effect on their
polymerization activity and the characteristics of the resulting
polymers, with five-membered ring systems affording poly-
mers in higher yields and with larger molecular weights than
their six-membered ring counterparts.
Increased heterotactic-biased poly(rac-LA) was formed as
the number of chlorine atoms increased in TiClx(O-i-Pr)4-x
.
Figure 7. Plot of Mn and PDI values (employing polystyrene standards)
for PLA as a function of conversion at 70 °C in toluene with [LA]/[Ti] )
300 using 5 as the catalyst. Mn (theory) was calculated from the formula of
(Mw of LA) × ([LA]/[Ti]) × (conversion).
In solution polymerizations, titanium alkoxides catalyzed
controlled polymerizations of LA, and end group analysis
demonstrated that an alkoxide group acted as the initiator.
That polymerization is controlled under our conditions was
shown by the linearity of molecular weight versus the
conversion of LA into PLA.
PLA of very low molecular weight when compared with
other titanium compounds, although the PDI values are quite
narrow.
Experimental Section
Summary and Conclusions
General Considerations. All reactions were carried out under
an argon atmosphere using standard Schlenk and glovebox tech-
niques.28 All chemicals were purchased from Aldrich and were used
as supplied unless otherwise indicated. Pentane, dichloromethane,
We have synthesized a series of titanium alkoxides from
readily available starting materials by simple procedures in
high yields. These compounds showed remarkably high
catalytic activity in bulk and solution polymerizations of LA,
revealing interesting correlations of catalyst structure with
polymer activity. The catalysts can be roughly divided into
(28) Shriver, D. F. The Manipulation of Air-SensitiVe Compounds; McGraw-
Hill: New York, 1969.
Inorganic Chemistry, Vol. 42, No. 5, 2003 1445

Kim et al.
THF, and toluene (Fischer HPLC grade) were dried and purified
under a nitrogen atmosphere in a Grubbs-type nonhazardous two-
column solvent purification system29 (Innovative Technologies) and
were stored over activated 3 Å molecular sieves. All deuterium
solvents were dried over activated molecular sieves (3 Å) and were
used after vacuum transfer to a Schlenk tube equipped with a J.
Young valve. CL was distilled under reduced pressure (90 °C/7
micron Hg pressure) from calcium hydride and stored in vacuo
over 4 Å molecular sieves. l-LA and rac-LA were purified twice
by sublimation at 70 °C under 7 µm Hg pressure before use.
Measurements. 1H, 13C{1H}, and 19F NMR spectra were
recorded at ambient temperature on a Varian VXR-300 or VXR-
400 NMR spectrometer using standard parameters. The chemical
product 6 was isolated as orange-yellow crystals after the solution
remained at -15 °C in a refrigerator for several days. Yield )
62% (0.62 g).
Compound 7. Characterization data for 7 include the follow-
ing: colorless crystals; yield, 90% (method 1), 66% (method 2).
1H NMR (CDCl3, 400.147 MHz): δ 4.72 (br s, 6H, CH2O), 3.44
(br s, 6H, NCH2). 19F NMR (CDCl3, 376.479 MHz): δ -9.74 (s,
2F), -14.05 (s, 2F), -20.43 (s, 2F). 13C{1H} NMR (CDCl3, 100.626
MHz): δ 74.37 (CH2O), 62.01 (CH2N). Aromatic carbons could
not be observed because of the low solubility of 7 in CDCl3. Anal.
Calcd for C12H12F5NO4Ti: C, 38.22; H, 3.21; N, 3.75. Found: C,
38.35; H, 3.08; N, 3.75.
Compound 8. Characterization data for 8 include the follow-
1
1
shifts are referenced to the peaks of residual CDCl3 (δ 7.24, H
ing: yellow powder; yield, 88% (method 1), 61% (method 2). H
1
NMR; δ 77.0, 13C{1H} NMR) and acetone-d6 (δ 2.05, H NMR).
NMR (acetone-d6, 400.147 MHz): δ 8.13 (d, J ) 11 Hz, 2H, aryl-
H), 7.01 (d, J ) 12 Hz, 2H, aryl-H), 4.62 (br s, 6H, CH2O), 3.49
(br s, 6H, NCH2). Anal. Calcd for C12H16N2O6Ti: C, 43.39; H,
4.86; N, 8.43. Found: C, 43.40; H, 5.22; N, 8.30.
Elemental analyses were performed by Desert Analytics Laboratory.
Molecular weights of polymers were determined by gel permeation
chromatography (GPC), and the measurements were carried out at
room temperature with THF as the eluent (1 mL/min) using a
Waters 510 pump, a Waters 717 Plus Autosampler, four Polymer
Laboratories PLgel columns (100, 500, 104, 105 Å) in series, and
a Wyatt Optilab DSP interferometric refractometer as a detector.
The columns were calibrated with polystyrene standards.
Syntheses. Compounds 2-4,12 5,13,21 10,13c 11,15 12,19 THBA,16
and 1430 were synthesized using literature procedures except that
these procedures were modified for the synthesis of 2-4, 10, and
11. In the case of 2-4, pentane and a shortened reaction time (1
h) were used. Compound 10 was obtained by a reaction between 5
and pinacol at room temperature instead of the reaction between
Et2NTi(OCH2CH2)3N with pinacol. Compound 11 was synthesized
by mixing Ti(O-i-Pr)4 and THA in dichloromethane instead of THF.
Compounds 7-9 were made by procedures analogous to both of
those given for 6.
Synthesis of 6. Method 1: In a 250 mL Schlenk flask containing
a stirring bar, phenol (0.471 g, 5.00 mmol), triethanolamine (0.746
g, 5.00 mmol), and 1 (1.42 g, 5.00 mmol) were charged in the
order given. Then, 50 mL of THF was added, and the reaction
mixture was refluxed overnight. After cooling to room temperature,
volatiles were evaporated under vacuum, leaving an orange-yellow
solid to which was added 15 mL of toluene. The orange solution
was filtered, and the desired product 6 was isolated as orange-
yellow crystals after the solution remained at -15 °C in a
refrigerator for several days (1.21 g, 84%). 1H NMR (CDCl3,
400.147 MHz): δ 7.13 (t, J ) 7.4 Hz, 2H, aryl-H), 7.01 (s, 1H,
aryl-H), 6.76 (t, J ) 6.6 Hz, 2H, aryl-H), 4.60 (br s, 6H, CH2O),
3.27 (br s, 6H, NCH2). 13C{1H} NMR (CDCl3, 100.626 MHz): δ
165.0, 128.7, 119.5, 119.0, 129.4 (aryl), 72.43 (CH2O), 61.10
(NCH2). Anal. Calcd for C12H17NO4Ti: C, 50.20; H, 5.97; N, 4.88.
Found: C, 49.15; H, 6.14; N, 5.13. Method 2: To a solution of
triethanolamine (0.746 g, 5.00 mmol) in 10 mL of THF was added
dropwise at room temperature a solution of 1 (1.42 g, 5.00 mmol)
in 10 mL of THF. After stirring overnight at room temperature, all
volatiles were evaporated under vacuum, leaving a yellow solid 5
(1.19 g, 94%). To a THF solution of 5 (1.00 g, 3.95 mmol) in 10
mL of THF was added dropwise at room temperature a solution of
phenol (0.371 g, 3.95 mmol) in 10 mL of THF. The reaction mixture
was refluxed overnight, and then, the volatiles were evaporated
under vacuum, leaving an orange-yellow solid, to which was added
15 mL of toluene. The orange solution was filtered, and the desired
Compound 9. Characterization data for 9 include the follow-
ing: yellow crystals; yield, 81% (method 1), 56% (method 2). 1H
NMR (CDCl3, 400.147 MHz): δ 6.70 (s, 2H, aryl-H), 4.51 (t, J )
5.6 Hz, 6H, CH2O), 3.25 (t, J ) 5.6 Hz, 6H, NCH2), 2.31 (s, 6H,
aryl-Me), 2.17 (s, 3H, aryl-Me). 13C{1H} NMR (CDCl3, 100.626
MHz): δ 160.3, 128.9, 128.0, 127.0 (aryl), 71.20 (CH2O), 57.00
(CH2N), 20.65, 17.02 (aryl-Me). Anal. Calcd for C15H23NO4Ti: C,
54.72; H, 7.04; N, 4.36. Found: C, 54.49; H, 7.67; N, 4.36.
Synthesis of N[CH2(ButMeC6H2)O]3TiOiPr (13). A solution
of 1 (5 mmol, 1.42 g) in dichloromethane (30 mL) was added to a
solution of THBA (5 mmol, 2.73 g) in dichloromethane (40 mL)
with stirring at room temperature over a period of 10 min, and
then, the solution was stirred for an additional 14 h. The solvent
was removed from this solution in vacuo, and the residue was
recrystallized from dichloromethane/hexane (1:2, 30 mL). The
yellowish crystalline product 13 was washed with cold pentane and
dried (yield, 2.76 g, 85.2% before recrystallization; yield, 2.12 g,
65.4% after recrystallization). Mp 366-368 °C. 1H NMR (299.94
MHz, CDCl3, ppm): δ 6.95 (d, J ) 1.7 Hz, 1H, aryl), 6.74 (d, J
) 1.5 Hz, 1H, aryl), 5.20 (m, 1H, OCHMe2), 3.93 (d, J ) 13 Hz,
3H, NCH2), 2.82 (d, J ) 13 Hz, 3H, NCH2), 2.24 (s, 9H, aryl-Me),
1.49 (d, J ) 6.1 Hz, 6H, OCHMe2), 1.42 (s, 27H, tert-butyl).
13C{1H} NMR (75.43 MHz, CDCl3, ppm): δ 160.5, 135.7, 128.8,
127.9, 126.6, 124.8 (aryl), 79.53 (CHMe2), 58.40 (NCH2), 34.72
(CHMe2), 29.50, 20.95 (aryl-Me). Anal. Calcd for C39H55NO4Ti:
C, 72.09; H, 8.53; N, 2.16. Found: C, 71.96; H, 8.66; N, 2.37.
X-ray Crystallography for 7. Crystallographic measurements
for 7 were performed at 173 K using a Bruker CCD-1000
diffractometer with Mo KR (λ ) 0.71073 Å) radiation and a
detector-to-crystal distance of 5.03 cm. Specimens of suitable
quality and size (0.2 × 0.08 × 0.05 mm3) were selected and
mounted onto glass fibers with silicon grease. The initial cell
constants were obtained from three series of ω scans at different
starting angles. Each series consisted of 30 frames collected at
intervals of 0.3° in a 10° range about ω with an exposure time of
50 s per frame. All the intensity data were corrected for Lorentz
and polarization effects. The structures were solved by the direct
method and were refined by a full-matrix anisotropic approximation.
All hydrogen atoms were placed at idealized positions in the
structure factor calculation and were allowed to ride on the
neighboring atoms with relative isotropic displacement coefficients.
Final least-squares refinement of 208 parameters against 2752
independent reflections converged to R (based on F2 for I > 2.0 σ)
and Rw (based on F2 for I > 2.0σ) values of 0.0408 and 0.0986,
respectively. Further details are listed in Table 4.
(29) Pangborn A. B.; Giardello, M. A.; Grubbs, R. H.; Rosen, R. K.;
Timmers, F. J. Organometallics 1996, 15, 1518.
(30) Kucht, A.; Kucht, H.; Barry, S.; Chien, J. C. W.; Rausch, M. D.
Organometallics 1993, 12, 3075.
1446 Inorganic Chemistry, Vol. 42, No. 5, 2003

Ti Alkoxides Initiating Lactide Polymerization
Table 4. Crystallographic Data for Compound 7
then, the appropriate amount of toluene was added to the flask at
the desired polymerization temperature. Polymerization began with
the addition of a stock toluene solution of the titanium compound.
After the appropriate time, the reaction was terminated by the
addition of 5 mL of methanol. The precipitated polymers were
dissolved in a minimum amount of methylene chloride, and then,
excess methanol was added. The reprecipitated polymers were
collected, washed with 3 × 50 mL of methanol, and dried in vacuo
empirical formula
fw
C24H24F10N2O8Ti2
377.13
temp (K)
cryst syst
space group
a (Å)
173(2)
triclinic
P1h
7.0985(15)
b (Å)
7.2229(15)
c (Å)
13.930(3)
1
at 50 °C for 12 h. H and 13C{1H} NMR spectra of PLA samples
R (deg)
â (deg)
γ (deg)
V (Å3)
103.976(4)
93.096(4)
101.600(4)
675.0(2)
were recorded in CDCl3.
A polymerization resumption experiment was carried out after
the desired polymerization time had been reached. The toluene
reaction mixture was transferred via cannula to a 50 mL Schlenk
flask containing a stirring bar, and then, another portion of LA that
had been heated to the desired polymerization temperature was
added. The remaining workup steps were the same as those
Z
D
1
calcd (Mg/m3)
1.856
0.715
380
abs coeff (mm-1
F(000)
)
θ range (deg)
2.97-28.25°
reflections collected
independent reflections
no. of parameters refined
GOF
3461
1
described earlier in this section. H and 13C{1H} NMR spectra of
2752
208
0.975
PLA samples were recorded in CDCl3.
final R indicesa [I > 2σ(I)]
R indices (all data)
R1 ) 0.0408, wR2 ) 0.0986
R1 ) 0.0550, wR2 ) 0.1040
0.425 and -0.409
Acknowledgment. This work was generously supported
by the National Science Foundation through a grant. We
thank Dr. Arkady Ellern for carrying out the X-ray structure
determination of 7 and Mr. Erik C. Hagberg in Prof. Valerie
Sheares-Ashby’s group in this department for obtaining the
GPC data.
largest diff. peak and hole (e Å-3
)
a R1 ) ∑||Fo| - |Fc||/∑|Fo| and wR2 ) {∑[w(Fo2 - Fc2)2]/∑[w(Fo ) ]}1/2
.
2 2
Polymerization Procedure. LA bulk polymerizations were
carried out by charging a stirring bar, 2.00 g of LA, and then the
appropriate amount of catalyst precursor to a 10 mL Schlenk flask.
The flask was then immersed in an oil bath at 130 °C, and after
the appropriate time, the reaction was terminated by the addition
of 5 mL of methanol. The precipitated polymers were dissolved in
a minimum amount of methylene chloride, and then, excess
methanol was added. The resulting reprecipitated polymers were
collected, washed with 3 × 50 mL of methanol, and dried in vacuo
at 50 °C for 12 h.
Supporting Information Available: GPC traces, homonuclear
1
decoupling H NMR spectra of the methine region for PLA, and
tables of crystal data, atomic coordinates, thermal parameters, bond
distances, bond angles and ORTEP diagrams for compound 7.
This material is available free of charge via the Internet at
http://pubs.acs.org.
Solution polymerizations of LA were carried out by charging a
stirring bar and LA to a 50 mL Schlenk flask in the glovebox, and
IC026139N
Inorganic Chemistry, Vol. 42, No. 5, 2003 1447