Expedient Synthesis and Ring-Opening Metathesis Polymerization of Pyridinonorbornenes

Abstract

Pyridine-containing polymers are promising materials for a variety of applications from the capture of contaminants to the self-assembly of block copolymers. However, the innate Lewis basicity of the pyridine motif often hampers living polymerization catalyzed by transition-metal complexes. Herein, we report the expedient synthesis of pyridinonorbornene monomers via a [4+2] cycloaddition between 2,3-pyridynes and cyclopentadiene. Well-controlled ring-opening metathesis polymerization was enabled by careful structural design of the monomer. Polypyridinonorbornenes exhibited high T_g and T_d, a promising feature for high-temperature applications. Investigation of the polymerization kinetics and of the reactivity of the chain ends shed light on the influence of nitrogen coordination on the chain-growth mechanism.

Graphical Abstract

Pyridinonorbornenes, a unique family of ROMP monomers, was synthesized via the intermediacy of 2,3-pyridynes. Modulation of the steric hindrance about the nitrogen allowed a living ROMP process and the production of polymers with high T_g and T_d.

Pyridine-containing polymers offer a wealth of applications stemming from their unique electronic features and the Lewis and Brønsted basicity of the nitrogen atom (Figure 1a).^1,2 Yet, few pyridine-containing monomers have been exploited in controlled polymerizations, in part due to the coordinative ability of pyridine toward transition-metal catalysts. 2- and 4-Vinyl pyridines (2VP and 4VP) remain the rare examples of monomers that allow for the precise synthesis of soft materials containing pyridine in the main chain (Figure 1a).^1–3 Poly(vinyl pyridine)s (P2VP or P4VP) have been investigated for the removal of organic contaminants from wastewater,⁴ for metal absorption,¹ and as support for recyclable reagents in oxidation, reduction, or halogenation reactions.^1,2 Vinyl pyridines have also emerged as monomers of choice for the self-assembly of diblock copolymers.³ Finally, nitrogen quaternization provides a practical entry into polyelectrolytes.^5,6 Interestingly, despite the well-documented functional-group tolerance of ring-opening metathesis polymerization (ROMP),^7–9 the rare pyridine-containing ROMP monomers reported in the literature (M1–M3) do not undergo well controlled polymerizations (Figure 1b).^10–13 Coordinating monomers are indeed known to detrimentally affect the rates of initiation and propagation through chelation to the Ru catalyst.^14–18 This issue is especially prevalent with monomers containing an unprotected amine in close proximity to the strained alkene.^19–21 Herein, we report the streamlined synthesis of a family of pyridinonorbornene monomers as analogs of 2VP for ROMP through the intermediacy of 2,3-pyridynes (Figure 1c). Optimization of the monomer structure and ROMP conditions allowed the synthesis of pyridine-containing polymers with predictable molar masses, functionalizable chain ends, as well as high thermal stability and glass-transition temperature (T_g), a desirable feature for a number of high-temperature applications.^22,23

Fig. 1.

(a) Poly(vinyl pyridine)s and some of their applications. (b) Prior examples of pyridine-containing norbornyl monomers. (c) Synthesis and ROMP of pyridinonorbornenes.

While pyridine motifs have been successfully incorporated into the backbone of macromolecules via polycondensation,^24–27 very few controlled chain-growth polymerizations have been reported outside of PVPs. Buchmeiser and coworkers synthesized dipyridyl norbornene M1 that was polymerized via ROMP with relatively high dispersity (Đ) values (1.6–1.8) using a Mo-based Schrock catalyst (Figure 1b).^10,11 Wakatsuki and coworkers subsequently polymerized monopyridyl norbornene M2 using Grubbs 2^nd generation Ru catalyst in a non-controlled fashion.¹² In both cases, coordination of the pyridine nitrogen to the metal alkylidene was postulated to be a main factor in the lack of control over molar mass distribution. Similarly, Iacono and coworkers recently reported the uncontrolled ROMP of monomers M3.¹³ However, the development of the 3^rd generation of Grubbs catalysts that contains two pyridine ligands (typically 3-bromopyridine)²⁸ suggests that pyridine coordination is not necessarily detrimental to the ROMP process. Pyridine ligands have been shown to undergo a rapid association/dissociation equilibrium,²⁹ and Guironnet and coworkers recently demonstrated that the rate of polymerization is directly tied with their coordination strength.³⁰ We hypothesized that a fusion of the pyridine core to the norbornene as depicted in 1 would both increase the steric hindrance around the nitrogen and decrease the conformational freedom and flexibility of the pyridine core, thereby disfavoring coordination to the Ru center (Figure 1c). Additionally, the introduction of a variety of R groups at the C6 position of 1 would provide a practical handle to further increase the steric bulk around the nitrogen and to concomitantly increase solubility of the targeted polymers in organic solvents.

While pyridinonorbornenes 1 have never been reported as ROMP monomers, 1a has been previously synthesized by Tanida and Irie in 14 steps and 2.7% overall yield from norbornadiene.³² Since this synthetic route would be hardly amenable to a scalable production of polymers and would create a bottleneck for the diversification of the monomer structure, a streamlined alternative was targeted. Inspired by the synthesis of benzonorbornene monomers via a Diels-Alder reaction between cyclopentadiene and benzyne reported by Feast³³ and subsequently by Maynard and Garg,³⁴ we hypothesized that utilizing a 2,3-pyridyne^35–37 as the dienophile would provide an expedient and modular sequence toward the family of monomers 1a–c (Figure 2). Heterocyclic arynes such as 2,3-pyridynes have seen a resurgence in total syntheses of natural products^38–40 but remain largely absent from the field of polymer synthesis. Following a procedure by Carrol,⁴¹ commercially available and inexpensive pyridone 2a was converted to 3a in 62% yield via silylation and triflation. After optimization, it was found that upon in situ generation of 2,3-pyridine using CsF, [4+2] cycloaddition with cyclopentadiene delivered pyridinonorbornene 1a in 45% yield (28% yield from 2a). When the same sequence was applied to hexyl variant 2b, a drastic decrease in yield was observed in the preparation of 3b caused by competitive silylation of the benzylic-like methylene in the hexyl chain. This finding prompted us to synthesize tert-butyl congener 2c through a known two-step route.⁴² To our delight, 2c could be converted to pyridyne precursor 3c in a much improved 85% yield over two steps, and the following cycloaddition step provided 1c in 55% yield resulting in a 47% overall yield from 2c. The ring strain of pyridinonorbornenes 1a–c was estimated via an isodesmic analysis of the opening of 1a–c with ethylene.⁴³ The computed ring strain of each potential monomer was found to be higher than that of norbornene by ~7 kcal/mol (Supporting Information), which indicated that ROMP of 1a–c should be thermodynamically favorable. Unfortunately, initial polymerization attempts using monomer 1a and Grubbs 3^rd generation catalyst (Ru-1) resulted in virtually no conversion of monomer (Table 1, entry 1). Switching to hexyl monomer 1b afforded similar results at room temperature but increasing the temperature to 60 °C led to a polymer with relatively low Đ and a M_n reasonably close to the predicted value, albeit with low monomer conversion (48% after 8 h; entries 2 and 3). Following this encouraging result, the polymerization of tert-butyl monomer 1c was attempted. Gratifyingly, full monomer conversion could be reached in 5 h at room temperature at a 50:1 ratio of 1c:Ru-1. Poly-1c exhibited good solubility in organic solvents including THF, DCM, and chloroform. SEC analysis revealed a narrow Đ of 1.07 and an excellent match between the experimental M_n and the predicted value (9.7 vs 10.1 kg/mol; entry 4). Polymerization with Grubbs 2^nd generation catalyst (Ru-2) afforded poly-1c with a sightly broader dispersity (Đ = 1.21; entry 5).⁴⁴ Careful analysis of ¹H and ¹³C NMR spectra, in combination with HSQC, HMBC, and COSY experiments (Figures 3a and S38–41) suggests that poly-1c likely exhibits regio irregularities from concurrent head-to-head or head-to-tail pathways. Mixed tacticities and cis/trans olefin stereochemistry could also contribute to the complexity of the NMR spectra. MALDI-TOF analysis confirmed the mass of the repeating unit, as well as the installment of a methylene group at the chain end upon quenching with ethyl vinyl ether (EVE) (Figure 3b).

Fig. 2.

Synthesis of pyridinonorbornenes 1a–c via 2,3-pyridynes. Reagents and conditions: 1) LDA (2.2 equiv), THF, 0 °C, 1 h, then TMSCl (1.2 equiv), 0 °C, 15 min, then rt, 12 h. 2) Tf₂O (1.1 equiv), pyridine, 0 °C, 20 min, then rt, 12 h. 3) CsF (5.0 equiv), C₅H₆ (5.0 equiv), MeCN, 30 °C, 3 h. ^a4:1 MeCN:DCM.

Table 1.

ROMP of Pyridinonorbornenes

Entrya	M	Ru	t (h)	Conv. (%)	Mntheo (kg/mol)	Mnexp (kg/mol)b	Đ b
1	1a	Ru-1	22	trace	–	–	–
2	1b	Ru-1	22	16	–	–	–
3c	1b	Ru-1	8	48	5.6	3.4	1.38
4	1c	Ru-1	5	>99	10.1	9.7	1.07
5	1c	Ru-2	5	>99	10.1	9.9	1.21

^aMonomer (M, 0.10 mmol) was polymerized by Ru catalysts (0.002 mmol) in DCE at rt.

^bM_n’s and Đ’s were determined by SEC (THF) using polystyrene standards (RI detection).

^cReaction temperature = 60 °C.

Fig. 3.

(a) ¹H NMR spectrum of poly-1c in CDCl₃ (*). (b) MALDI-TOF analysis of poly-1c (M_n^SEC = 5.0 kg/mol, Đ = 1.12).

In order to thoroughly characterize the livingness of the polymerization, the kinetics of the ROMP of 1c with Ru-1 were studied via ¹H NMR in CD₂Cl₂. The measured linear relation between ln([M₀]/[M_t]) and time confirmed the first-order dependence in concentration of monomer 1c as expected for a living chain-growth process (Figure 4a). The rate of polymerization of 1c with Ru-1 is notably lower than that of typical norbornene monomers.³⁰ M_n was found to grow linearly with conversion of 1c, and dispersity values remained below 1.2 throughout the ROMP (Figure 4b). Of note, the quenching with EVE of the terminal Ru-carbene was noticeably slower with 1c when compared to more common ROMP monomers, which rendered accurate determination of the conversion of 1c over time challenging (Table S5). Sequential addition of the strongly Lewis basic ligand 4-dimethylaminopyridine (DMAP)^45,46 followed by EVE helped circumvent this issue. Poly-1c with higher degrees of polymerization were obtained by varying the ratio of monomer 1c to catalyst Ru-1, however a small shoulder was detected for DPs ≥ 100 corresponding to polymeric impurities roughly twice as large as the targeted M_n (Figure S2a). This shoulder was tentatively ascribed to adventitious cross metathesis between methylene and Ru-bond chain ends taking place during the slow quenching process. Fortunately, this issue was minimal with Ru-2, which afforded poly-1c with predictable M_n’s up to 77.6 kg/mol (Table 2, entries 1–4; Figure S2b).

Fig. 4.

Polymerization of 1c with Ru-1 ([1c]:[Ru-1] = 50:1): (a) Determination of the rate of propagation (R² = 0.998) at 32 °C in CD₂Cl₂; (b) M_n (blue squares) and Đ (green triangles) vs conversion.

Table 2.

ROMP of 1c by Ru-2

Entrya	[1c]:[Ru]	t (h)	Conv. (%)	Mntheo (kg/mol)	Mnexp (kg/mol)b	Đ b
1	100:1	11	85	17.0	16.7	1.17
2	200:1	25	90	36.0	29.7	1.12
3	300:1	37	93	55.7	49.8	1.11
4	400:1	42	87	69.4	77.6	1.12

^a1c (0.10 mmol) was polymerized with Ru-2 in DCE at rt for specified amounts of time.

^bM_n’s and Đ’s were determined by SEC (THF) using polystyrene standards (RI detection).

Chain-extension of poly-1c with additional 1c provided evidence that the propagating chain ends are preserved after monomer depletion. Full conversion of 1c was observed following the addition of a second batch of monomer concomitant with a uniform shift of the SEC trace (Figure 5a). However, when chain extension was attempted with exo-norbornene diimide 4, a bimodal distribution was observed clearly indicating partial chain extension (Figure 5b top). Taken together with the relatively slow rates of polymerization and quenching with EVE, this finding suggests that the Lewis basic nitrogen of the terminal pyridine motif binds to the Ru center (Figure 5b) thereby hampering reinitiation with a non- or poorly coordinating monomer such as 4.^14–18 A similar issue was encountered by Xia and coworkers in the synthesis of block copolymers containing a coordinating diester unit and solved through addition of 3-bromopyridine to facilitate the reinitiation.⁴⁷ This strategy allowed the isolation of poly-1c-b-4 with minimal chain termination as shown by SEC and DOSY NMR (Figures 5b bottom and S44). Notably, the polymerization of the second block reached full conversion in about 1 h, while homopolymerization of 4 in the same conditions (catalyst, solvent, concentration, temperature, etc.) was complete in only a few minutes (Table S7). The slow rate of propagation of 4 when used for chain extension likely indicates that nitrogens in the main chain of the poly-1c segment also coordinate to the catalyst. Random copolymerization of 1c and 4 (~4:1) similarly led to a relatively slow polymerization of 4. A simultaneous and almost quantitative incorporation of both monomers was observed over 10 h, while the molar mass distribution remained narrow and monomodal (Table S8 and Figure S6). These experiments point toward a dynamic and complex interplay between the Ru center and the Lewis basic nitrogens of the chain end, backbone, and monomer. Future work will investigate the influence of the coordination of each species over the ROMP process.

Fig. 5.

(a) Synthesis of poly-1c₁₀₀ via chain extension. (b) Synthesis of diblock copolymer poly-1c-b-4 using 3-bromopyridine to facilitate reinitiation of the chain ends.

Lastly, the thermal properties of poly-1c were investigated. Due to the structural rigidity of its backbone, it was hypothesized that this material would possess a higher glass transition temperature than P2VP. Indeed, poly-1c exhibited a T_g of 157 °C via differential scanning calorimetry (DSC), which is significantly higher than the reported T_g of P2VP (~104 °C).^3,48 DSC however did not reveal any melting, nor crystallization transitions. Poly-1c demonstrated good thermal stability through thermogravimetric analysis with a decomposition temperature (T_d) recorded at 5% mass loss, T_d = 281 °C. These thermal properties are promising for high-temperature applications and could be further manipulated through functionalization of the olefinic backbone.

Conclusions

In summary, this study demonstrates the potential of heterocyclic arynes as building blocks toward unique monomers and macromolecules. A key [4+2] cycloaddition between 2,3-pyridynes and cyclopentadiene delivered a streamlined access to pyridinonorbornenes, a unique class of ROMP monomers. Fine tuning of the pyridinonorbornene structure led to a ROMP process of 1c with living characteristics despite the postulated coordination of the pyridine nitrogen to the Ru center. This addition to the ROMP toolbox opens the door to the precise synthesis of pyridine-containing polymers as alternatives to PVPs with higher T_g. Potential applications range from metal capture to polymer self-assembly. Finally, this work demonstrates that modulation of the steric hinderance is a viable and protecting-group-free strategy for controlled ROMP of Lewis-basic monomers.