Titanium pyridonates and amidates: novel catalysts for the synthesis of random copolymers

Abstract

A series of pyridonate and amidate supported titanium alkoxides have been isolated. These complexes can be readily prepared in high yield, under mild reaction conditions in only two steps from commercially available (Ti(NMe₂)₄). We have furnished one of the rare examples of discrete catalysts for random copolymer synthesis.

---

There have been great advances in the initiator systems available for the ring-opening polymerisation (ROP) of lactide (LA) and caprolactone (CL), including systems that produce very high molecular weight polymers and exhibit exquisite LA-stereocontrol under ambient conditions.¹ Titanium catalysts have not, as yet, shown such high levels of reactivity and control during polymerisation. Although Ti-mediated homopolymerisations cannot yet compete with alternative metal initiators, the low toxicity² of titanium makes it attractive for the synthesis of copolymers with potential biomedical application.³ For example, high molecular weight homopolymer for application in drug-delivery materials can result in poor biodegradation or detrimental “bursts” of drug release.⁴ Seppälä and co-workers have shown that random copolymers offering alternative mechanical properties to traditional homopolymers can be advantageous for such specialised applications.^4b,5 PLA, characterised as hard and brittle with low maximum strain, and PCL, known to be tough with high maximum strain, can be combined into copolymers offering high levels of strain yet a range of physical properties “from weak elastomers to tougher thermoplastics”.^5b As a result blends, block or random copolymers of PLA and PCL may offer a desirable, tunable combination of properties suitable for a range of applications in biomedical science. While copolymerisations are an intense area of investigation with a range of metal centres,^3,6 such reactivity with discrete titanium initiators has rarely been reported.^6a Indeed very few examples of random copolymer formation have been realised with designed catalysts.⁶ Exploring catalyst structure/polymer properties is of fundamental importance to our understanding of copolymerisations and realising their potential in real-life applications.

We have previously reported the synthesis of bis(pyridonate)- and bis(amidate)-metal-amido complexes of the Group 4 transition metals as excellent precatalysts for the catalytic synthesis of amines.^7–10 However, Group 4 pyridonate and amidate complexes have not yet been explored for their potential in ROP. Yttrium amidate complexes on the other hand have been exploited in this role,¹¹ whereby trisligated complexes have been shown to produce high molecular weight PCL.^11a However, the moisture sensitivity of such yttrium initiators is a limitation to their application toward scalable, functionalised polymer synthesis. Pyridonate or amidate titanium alkoxides could provide an alternative robust platform for investigating ligand stereoelectronic effects upon ROP, combined with the desirable bio-compatibility of titanium (Scheme 1).

Scheme 1 Synthesis of pyridonate- and amidate-supported titanium-alkoxides and their application in polymerisation.

Synthesis of bis(pyridonate)- or bis(amidate)-titanium-bis(dialkylamido) complexes is facile, by reaction of neutral amide or pyridone proligands with commercially available transition metal tetrakis(dialkyl)amido complexes at room temperature.⁷ However, synthesis and rigorous characterisation of the related alkoxides has thus far remained elusive, likely due to complex aggregate formation during attempted reactions of Ti(OⁱPr)₄ with pyridonate and amidate proligands. However, we were pleased to find that a sequential protonolysis approach, by reaction of Ti(NMe₂)₄ with proligand, followed by reaction with isopropanol in one pot, furnishes a range of pyridonate- or amidate-titanium-alkoxides in excellent yield.¹² To the best of our knowledge, these are the first examples of fully characterised and mononuclear pyridonate- and amidate-titanium-alkoxides (Fig. 1).

Fig. 1 Pyridonate- and amidate-titanium-alkoxides 1–5 were synthesised in high yield from Ti(NMe₂)₄. Right: ORTEP representation of the solid state molecular structure of 1, thermal ellipsoids set at 50%, selected bond lengths (Å) Ti–O1 2.003(1), Ti–O2 1.986(1), Ti–N1 2.219(1), Ti–N2 2.250(2), Ti–O3 1.785(1), Ti–O4 1.766(1), C1–O1 1.318(2), C1–N1 1.344(3), C7–O2 1.325(2), C7–N2 1.347(2) and angles (°) N1–Ti–O1 62.72(5), N2–Ti–O2 62.51(5), N1–C1–O1 111.8(2), N2–C7–O2 111.5(2).

6-Methyl-, 3-methyl- and 3-phenyl-pyridonate complexes (1, 2 and 3) have been synthesised in the first instance to provide a direct comparison of steric bulk adjacent to nitrogen versus the oxygen heteroatom. Crystals of 1 can be grown from hexanes at −30 °C and it can be seen that the ligand binds unsymmetrically through the N,O chelate.¹² The Ti–O1 and Ti–N1 bond lengths are 2.003(1) Å and 2.219(1) Å respectively. Such asymmetric binding modes have been previously reported for pyridonate group 4 bis(dialkylamido) complexes,¹³ but are in contrast to previously reported bis(amidate)-bis(dialkylamido) complexes (vide infra).⁸ The robust nature of these complexes is astonishing; heating complexes 1 to 3 in refluxing ethanol for up to 64 h does not result in ligand protonolysis and precipitation of Ti oxides as may be expected, instead the pyridonate ligand remains coordinated to the metal.¹² The highly sterically hindered amidate alkoxides 4 and 5 also form readily, but 5 shows broadened signals in the ¹H NMR spectrum at room temperature, associated with fluxional ligand binding on the NMR time-scale. Heating the sample to 95 °C results in a simplified NMR spectrum. Although X-ray quality crystals of these titanium amidates could not be obtained, characterisation by VT-NMR spectroscopy and HRMS suggest that discrete monomeric species form. This would be consistent with the previously reported monomeric bis(diethylamido) precursor.^8b Interestingly, in the solid state the diethylamido species is C₂ symmetric and exhibits bond lengths for Ti–O_amidate 2.146(1) Å and Ti–N_amidate 2.156(1) Å, indicating a far more delocalised N,O-chelate than that exhibited by the pyridonates described above. Considering the varying steric and electronic attributes of these related ligand sets, it is interesting to examine their different catalytic activities.

We first investigated the homopolymerisation of LA and CL using catalysts 1 to 5 in the melt state (Table 1). As anticipated M_n and PDIs are modest compared to alternative metal initiators,¹ but are competitive with leading Ti examples.^14,15 Differences between observed and theoretical M_n (M_ntheo) may be consistent with transesterification processes.¹² Some PLA chain-end stereocontrol is observed and it is interesting to note the shift in stereocontrol in moving from pyridonate initiators 2 and 3, where there is an isotactic bias, to amidates 4 and 5 where there is a slight preference for heterotacticity. Another intriguing facet is the increase in PLA M_n on moving from 6-methyl-pyridonate, 1, to 3-methyl-pyridonate, 2, which is further enhanced when 3, with a bulkier substituent in the 3-position, is employed. This trend is reversed during PCL synthesis (compare entries 6, 7 and 8): we questioned whether this tunability during homopolymerisations coupled with the unusual ability to afford efficient PLA and PCL formation could be exploited for the synthesis of random copolymers.

Table 1 Melt-phase homopolymerisation of rac-LA and ε-CL

Entry	Cat.	Monomer	Yield^a (%)	Mn^b (g mol^−1)	Mntheo^c^,12	PDI^b	Pm^d
PLA: 130 °C, 24 h, [LA]/[Ti] = 300, 0.5 g. PCL: 100 °C, 16 h, [CL]/[Ti] = 300, 0.5 ml.a Isolated yield.b Determined by GPC.c Mntheo = ([M]/2[Ti] × %Yield × MW).d ^1H{^1H} NMR spectrum.e Bimodal GPC trace.f 1 h.g 2 h.
1	1	LA	88	13 800	19 030	1.17	0.49
2	2		90	14 120	19 460	1.21	0.46
3	3		81	22 220	17 510	1.22	0.46
4	4		95	23 040	20 540	1.17	0.55
5	5		93	25 440	20 110	1.16	0.53
6	1	CL	98	39 400	16 780	1.29	—
7	2		98	22 810	16 780	1.38	—
8^e	3		96	21 360	16 440	1.28	—
9^f	4		98	20 860	16 780	1.48	—
10^e^,^g	5		85	27 470	14 550	1.33	—

---

To date, few examples of well-characterised initiators are known to facilitate the copolymerisation of LA and CL,⁶ and the present state-of-the-art is Al Schiff-base or salen complexes.^6d,f Examples of CL/LA random copolymers where there is a 1 : 1 incorporation of both monomers are rare simply because there is a propensity for LA to polymerise first, followed by CL incorporation at a lower rate.¹⁶ Alternatively, an initiator which facilitates randomisation through transesterification can be used, thus avoiding this formation of block-like polymers, instead resulting in short average sequence lengths¹⁷ (an ideal random copolymer will have L_CL = L_LL = 2^16a,6d). Complexes 1, 2 and 3 form random copolymers in excellent yield using a low catalyst loading compared to leading salan/salen examples.^6b,d We also observe excellent levels of CL and LA incorporation without the need for an excess CL feed. M_n is competitive with previously reported values using well-defined initiators with a 1 : 1 monomer feed (e.g. 22 320 using initiator 3vs. 13 900,^6b 16 300,^6c and 21 600^6d), Table 2. An initial attempt to increase the incorporation of CL into the random copolymer, by moving to reaction conditions which favour CL homopolymerisation i.e. lower reaction temperature, did not affect the polymer composition as hoped, but in fact led to predominantly PLA with only 12% CL (entry 2). The change from initiator 1 to 2, resulted in a decrease in PDI and more importantly a reduction in L_LL and a decrease in the LA content (48%). A T_g of −3.3 °C is obtained for copolymer synthesised using 2, where the theoretical value is −20.8 °C,¹⁶ while T_m was not observed. Pushing the reactivity trends further, to 3-phenyl-pyridonate 3, results in an increase in M_n, but a return to higher levels of LA incorporation. Although the yield with initiator 4 is modest, an unusual preference for CL incorporation is observed. In contrast, PLA dominates when bulkier amidate complex 5 is employed.

Table 2 Melt-phase copolymerisation of ε-CL and rac-LA

Entry	Catalyst	Yield^a (%)	CL/LA^b (mol%)	LCL/LLL^c	Mn (g mol^−1)^d	Mntheo^e^, ^12 (g mol^−1)	PDI^d
Conditions: 130 °C, 24 h, [monomer]/[Ti] = 600, 0.5 g (3.45 mmol) LA, 0.38 ml (3.45 mmol) CL.a Isolated yield.b Ratio of CL/LA determined by ^1H NMR.c Average chain length determined by ^13C{^1H} NMR.d Values determined by GPC analysis.e Mntheo = ([CL]/2[Ti] × %CL × 114.14) + ([LA]/2[Ti] × %LA × 144.13).f 100 °C, 18 h, then 130 °C, 7 h.
1	1	82	45/55	1.8/3.4	18 750	19 600	1.41
2	1^f	60	29/71	—	28 780	20 320	1.36
3	2	83	52/48	1.9/2.9	19 070	19 280	1.29
4	3	86	43/57	1.7/3.5	22 320	19 690	1.38
5	4	68	64/36	3.1/2.4	19 190	18 740	1.37
6	5	63	5/95	—	23 660	21 400	1.38

---

Investigation of the carbonyl region of the ¹³C{¹H} NMR spectrum reveals that all of the copolymers undergo transesterification.¹⁷ Reaction monitoring of polymerisation catalysed by 3 shows rapid polymerisation of LA to form longer chains of PLA (6 h: L_CL 1.5; L_LL 5.7) with CL reacting modestly (Fig. 2a). CL/LA ratios remain static after 10 h, where there is also a concomitant increase in the presence of heterodiads (CL–LA bonds), thereby reiterating the randomising effect of transesterification (Fig. 2b). After 10 h L_LL and L_CL reach their final observed values, however, M_n is 12 550 g mol⁻¹ and the PDI is 1.48 (compared to 22 320 g mol⁻¹ and 1.38 in the final isolated polymer). This data suggests that the polymer chain continues to grow after the 10 h time period through transesterification.^18,19 Methanol soluble (i.e. low mass) fractions of the quenched reaction solutions were subjected to qualitative MALDI-TOF analysis. Low mass cyclic species and short linear chains composed of half lactide units can be seen. The methyl ester peak, resulting from methanol-quenched transesterifications, is also present in the ¹H NMR spectrum.¹²

Fig. 2 (a) Reaction profiling the consumption of LA () and CL () during copolymerisation; (b) monitoring the presence of LA homodiads (), CL homodiads () and heterodiads (×) over time.

In conclusion, novel bis(pyridonate)- and bis(amidate)-titanium-alkoxides have been isolated in excellent yield. All can be prepared under mild reaction conditions from a common precursor, Ti(NMe₂)₄. The complexes have been used in the polymerisation of LA and CL and are the first examples of polymerisation initiated by this class of titanium compound. Most importantly, these titanium initiators can be used for copolymer synthesis with short sequence lengths, close to 1 : 1 monomer incorporation and good M_n due to transesterification. We plan to further exploit our synthesis of random copolymers with these readily modified ligand sets to access a family of polymerisation catalysts affording a large scope of M_n, sequence length and level of transesterification. Such synthetic results, polymer characterization and rheological investigations will be reported shortly.

NSERC and NOVA Chemicals Corporation are acknowledged for financial support of this work. RLW thanks the Government of Canada for a Commonwealth PDRF. We thank Derek P. Gates for use of GPC instrumentation and Parisa Mehrkhodavandi for helpful discussion. This research was undertaken, in part, thanks to funding from the Canada Research Chairs program.

Notes and references

1. For an overview see: (a) M. Labet and W. Thielemans, Chem. Soc. Rev., 2009, 38, 3484; (b) R. H. Platel, L. M. Hodgson and C. K. Williams, Polym. Rev., 2008, 48, 11; (c) J. Wu, T.-L. Yu, C.-T. Chen and C.-C. Lin, Coord. Chem. Rev., 2006, 250, 602; (d) B. J. O'Keefe, M. A. Hillmyer and W. B. Tolman, J. Chem. Soc., Dalton Trans., 2001, 2215.
2. D. M. Brunette, P. Tengvall, M. Textor and P. Thomsen, Titanium in Medicine: Materials Science, Surface Science, Engineering, Biological Responses and Medical Applications, Springer, 2001.
3. (a) Polymers for Biomedical Application, ed. A. Mahapataro and A. S. Kulshrestha, American Chemical Society, 2008; (b) L. S. Nair and C. T. Laurencin, Prog. Polym. Sci., 2007, 32, 762.
4. (a) J. Rich, T. Karjalainen, L. Ahjopalo and J. Seppälä, J. Appl. Polym. Sci., 2002, 86, 1; (b) T. Karjalainen, J. Rich and J. Seppälä, J. Appl. Polym. Sci., 2001, 81, 2118; (c) N. Kunou, Y. Ogura, T. Yasukawa, H. Kimura, H. Miyamoto, Y. Honda and Y. Ikada, J. Controlled Release, 2000, 68, 263.
5. (a) N. Ahola, J. Rich, T. Karjalainen and J. Seppälä, J. Appl. Polym. Sci., 2003, 88, 1279; (b) M. Hiljanen-Vainio, T. Karjalainen and J. Seppälä, J. Appl. Polym. Sci., 1996, 59, 1281.
6. (a) D. Dakshinamoorthy and F. Peruch, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 2161; (b) Y. Wang and H. Ma, Chem. Commun., 2012, 48, 6729; (c) D. J. Darensbourg and O. Karroonnirum, Macromolecules, 2010, 43, 8880; (d) N. Nomura, A. Akita, R. Ishii and M. Mizuno, J. Am. Chem. Soc., 2010, 132, 1750; (e) D. Pappalardo, L. Annunziata and C. Pellecchia, Macromolecules, 2009, 42, 6056; (f) M. Florczak and A. Duda, Angew. Chem., Int. Ed., 2008, 47, 9088.
7. R. K. Thomson, J. A. Bexrud and L. L. Schafer, Organometallics, 2006, 25, 4069.
8. (a) Z. Zhang, D. C. Leitch, M. Lu, B. O. Patrick and L. L. Schafer, Chem.–Eur. J., 2007, 13, 2012; (b) C. Li, R. K. Thomson, B. Gillon, B. O. Patrick and L. L. Schafer, Chem. Commun., 2003, 2462; (c) Z. Zhang and L. L. Schafer, Org. Lett., 2003, 5, 4733.
9. R. O. Ayinla and L. L. Schafer, Inorg. Chim. Acta, 2006, 359, 3097.
10. J. A. Bexrud, P. Eisenberger, D. C. Leitch, P. R. Payne and L. L. Schafer, J. Am. Chem. Soc., 2009, 131, 2116.
11. (a) L. J. E. Stanlake, J. D. Beard and L. L. Schafer, Inorg. Chem., 2008, 47, 8062. Bis-ligated examples: (b) F. Zhang, J. Zhang, H. Song and G. Zi, Inorg. Chem. Commun., 2011, 14, 72; (c) Q. Wang, F. Zhang, H. Song and G. Zi, J. Organomet. Chem., 2011, 696, 2186.
12. See ESI†.
13. J. A. Bexrud and L. L. Schafer, Dalton Trans., 2010, 39, 361.
14. Selected rac-LA examples include: (a) A. D. Schwarz, K. R. Herbert, C. Paniagua and P. Mountford, Organometallics, 2010, 29, 4171; (b) A. J. Chmura, M. G. Davidson, M. D. Jones, M. D. Lunn, M. F. Mahon, A. F. Johnson, P. Khunkamchoo, S. L. Roberts and S. S. F. Wong, Macromolecules, 2006, 39, 7250; (c) S. Gendler, S. Segal, I. Goldberg, Z. Goldschmidt and M. Kol, Inorg. Chem., 2006, 45, 4783; (d) C. K. A. Gregson, I. J. Blackmore, V. C. Gibson, N. J. Long, E. L. Marshall and A. J. P. White, Dalton Trans., 2006, 3134; (e) M. H. Chisholm, C. C. Lin, J. C. Gallucci and B. T. Ko, Dalton Trans., 2003, 406; (f) Y. Kim, G. K. Jnaneshwara and J. G. Verkade, Inorg. Chem., 2003, 42, 1437.
15. Selected ε-CL examples include: (a) D. Dakshinamoorthy and F. Peruch, J. Polym. Sci., Part A: Polym. Chem., 2012, 49, 5176; (b) A. D. Schwarz, A. L. Thompson and P. Mountford, Inorg. Chem., 2009, 48, 10442; (c) J. Cayuela, V. Bounor-Legare, P. Cassagnau and A. Michel, Macromolecules, 2006, 39, 1338; (d) A. J. Chmura, M. G. Davidson, M. D. Jones, M. D. Lunn and M. F. Mahon, Dalton Trans., 2006, 887.
16. (a) Y. Gnanou and M. Fontanille, Organic and Physical Chemistry of Polymers, Wiley, USA, 2008; (b) J. M. Vion, R. Jerome, P. Teyssie, M. Aubin and R. E. Prudhomme, Macromolecules, 1986, 19, 1828.
17. J. Kasperczyk and M. Bero, Makromol. Chem., 1993, 194, 913.
18. (a) H. R. Kricheldorf, K. Bornhorst and H. Hachmann-Thiessen, Macromolecules, 2005, 38, 5017; (b) P. Vanhoorne, P. Dubois, R. Jerome and P. Teyssie, Macromolecules, 1992, 25, 37.
19. It can be assumed that transesterification continues to proceed after 10 h because there is no significant change in the L_LL and L_CLvalues.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental details and full spectroscopic analysis data. CCDC 882572. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c2cc37201k

---