Stereocontrolled acyclic diene metathesis polymerization

Abstract

The cis/trans geometry of olefins is known to dramatically influence the thermal and mechanical properties of polyalkenamers. Yet, polymerization methods that efficiently control this parameter are scarce. Herein, we report the development of a stereoretentive acyclic diene metathesis (ADMET) polymerization that capitalizes on the unique reactivity of dithiolate Ru carbenes combined with cis monomers. These Ru catalysts exhibit exquisite retention of the cis geometry and tolerate many polar functional groups, enabling the synthesis of all-cis polyesters, polycarbonates, polyethers, and polysulfites. Additionally, the stereoretentive ADMET is characterized by low catalyst loadings and tolerance toward trans impurities in the monomer batch, which should ease scalability. Modulation of the reaction temperature and time leads to an erosion of stereoretention, permitting a stereocontrolled synthesis of polyalkenamers with predictable cis:trans ratios. The impact of the cis:trans alkene content within the polymer backbone on the thermal properties was clearly demonstrated through differential scanning calorimetry and thermogravimetric analysis.

Perhaps best exemplified by the numerous modifications of the venerable Wittig and Horner–Wadsworth–Emmons reactions, controlling the cis or trans (Z or E) geometry of alkenes in organic molecules has motivated the development of a myriad of synthetic methods to access either isomer selectively.¹ Indeed, the configuration of alkenes dramatically impacts the shape, properties, and reactivity of compounds. Interestingly, despite the importance of this structural parameter on the thermal and mechanical properties of polymeric materials,² few polymerizations allow the practitioner to precisely dictate the cis:trans ratio within macromolecules and these processes are limited to a narrow scope of monomers.^3-6 Among the classic examples illustrating the influence of alkene geometry over material properties, trans-polyisoprene (gutta percha) is a hard, brittle semi-crystalline material, once used in the manufacturing of golf balls,⁷ while the more amorphous cis-polyisoprene is an elastic material found in latex gloves.^8,9

Ring-opening metathesis polymerization (ROMP) and acyclic diene metathesis (ADMET) polymerization are among the most powerful methods to access unsaturated polymers, but typically afford cis/trans mixtures or trans-rich linkages dictated by the thermodynamic stability of the products.¹⁰ The recent introduction of stereoselective Mo, W, and Ru catalysts has paved the way for kinetic control of the olefin metathesis process, overcoming the product preference for trans alkenes.^11-19 While careful catalyst design has afforded impressive cis selectivity in a variety of transformations based on olefin metathesis, only a limited number of polymers have been synthesized through stereoselective processes, all through ROMP. For example, dithiolate Ru carbenes uniquely preserve the configuration of the starting olefin throughout the metathesis process.^20-22 Hoveyda and coworkers demonstrated that ROMP of norbornene or cyclooctadiene with this family of stereoretentive catalysts leads to exquisite cis selectivity along the polymer backbone.¹² Choi and coworkers expanded the scope of norbornene-type monomers for stereoretentive ROMP,¹⁶ and reported that the shear stability of the resulting all-cis polymers was greatly enhanced relative to that of their all-trans counterparts. Recently, we harnessed the reactivity of this family of catalysts in combination with [2.2]paracyclophane diene monomers to synthesize all-cis poly(p-phenylene vinylene)s (PPVs) with living characteristics and unusually high molar masses.^23,24 While these examples of stereoretentive ROMP afforded extremely precise polymeric structures (narrow molar mass distributions, perfect cis stereoselectivity, etc.), the ROMP process typically requires highly strained monomers thereby narrowing the scope of accessible polymers. Further, these scarce examples only afforded strict cis selectivity. Herein, we report the implementation of stereoretentive olefin metathesis into the ADMET process as a versatile method to access all-cis polyalkenamers. Additionally, careful optimization of the reaction conditions provided unprecedented stereocontrol over the proportion of cis and trans alkenes throughout the polymer backbone. Several families of alkene-containing polymers were synthesized with predictable cis:trans ratios from 20:80 to >99:1, providing a valuable strategy to tune the thermal properties of several types of polymers.

One of the major advantages of ADMET over other polymerizations based on olefin metathesis is the simplicity and ubiquity of the monomer structure characterized by the presence of two terminal alkenes. These α,ω-dienes are polymerized following a step-growth mechanism, where removal of ethylene gas is the thermodynamic driving force (Fig. 1a).^25,26 Pioneered by Wagener and coworkers,²⁷ ADMET polymerization has been employed to synthesize an impressive variety of polymers including polyethylene with precise alkyl branching,²⁸ ionic polymers,²⁹ and conjugated polymers, albeit with high trans alkene content.³⁰ Developing a stereoselective ADMET would therefore open the door to a cornucopia of stereodefined polymers for a variety of applications, but several pitfalls were identified at the outset of this work. Since ADMET polymerization relies on iterative couplings between terminal dienes that are neither cis nor trans (Fig. 1b), typical ADMET monomers preclude the stereoretentive mechanism of dithiolate Ru catalysts. Additionally, the Ru methylidene intermediate generated upon reaction of the dithiolate carbene with a terminal alkene is known to readily decompose, likely through a 1,2-shift of the anionic sulfide ligand.¹⁴ Hoveyda and coworkers reported an elegant strategy to circumvent the premature deactivation of the catalyst in the cross metathesis of terminal alkenes through in situ methylene capping with gaseous cis-but-2-ene.³¹ This approach, however, is not compatible with ADMET, which requires constant removal of volatile alkene by-products to drive the step-growth process. We envisioned that synthesizing capped monomers with pre-installed cis geometry would obviate both catalyst decomposition and the need for cis-but-2-ene, as well as allow the desired stereoretention (Fig. 1c). While non-terminal dienes have previously been polymerized using ADMET,³² the retention of alkene configuration of the monomer was not observed with Grubbs’ second generation catalyst. The proposed mechanistic cycle of a stereoretentive ADMET is outlined in Fig. 1d. The [2+2] cycloaddition of carbene intermediate I with a capped monomer or a growing polymer would afford ruthenacyclobutane II and subsequent cycloreversion would lead to the formation of stabilized Ru ethylidene (R = Me) or propylidene (R = Et) intermediate III. In contrast to the bottom-bound approach typically observed with dichloro Ru carbenes, the dithiolate ligand is known to enforce a side-bound approach of the monomer with the alkene substituents pointing away from the bulky aryl groups of the NHC ligand (Fig. 1e).²¹ Following cycloreversion, Ru-ethylidene/propylidene III would further react with another capped alkene (from a monomer or a growing oligomer) to form ruthenacycle IV, whose cycloreversion would regenerate a reactive Ru carbene and expel volatile cis-but-2-ene (R = Me) or cis-hex-3-ene (R= Et). Removal of either byproduct would drive all equilibriums toward a productive cycle.

Fig. 1. Design of stereocontrolled ADMET.

a. Typical ADMET polymerization conditions with unselective dichloro Ru carbenes that favor high trans alkene content. b. Dithiolate Ru catalysts lead to highly unstable Ru methylidenes upon reaction with terminal diene monomers. c. This work: Design of a stereocontrolled ADMET modulated by the reaction temperature. d. Proposed stereoretentive ADMET cycle (the initial reaction between the Ru precatalyst and the monomer was omitted for clarity). e. Bottom-bound olefin approach with dichloro Ru catalysts compared to side-bound approach with dithiolate Ru catalysts that result in trans or cis olefin configuration, respectively.

Results and discussion

Capitalizing on the inexpensive and readily available cis-hex-4-en-1-ol and cis-oct-5-en-1-ol,³³ a variety of monomers were synthesized in one or two steps in high yields (Fig. 2a and Supplementary Information General Procedures A-E).^34-37 These two groups of monomers (1–4 and 5–8) were selected to interrogate the tolerance of dithiolate catalysts to polar monomers containing Lewis basic functional groups, as well as the effect of increasing the size of the capping group (Et vs Me) on the polymerization. As a benchmark of cis/trans selectivity, terminal diene 14 was polymerized under standard ADMET conditions using Ru-1 (Fig. 2b) in 1,2,4-trichlorobenzene at 50 °C under high vacuum (~100 mTorr) for 16 h. The resulting polymer, poly-1 (M_n = 13.1 kg/mol, Đ = 2.61), was isolated with 23% cis content as established by ¹H NMR (Table 1, entry 1 and Supplementary Fig. 4). As expected, the polymerization of 14 with dithiolate catalysts Ru-3 and Ru-4 (Fig. 2b) did not afford any polymers (Table 1, entries 2 and 3). Switching to methyl-capped monomer 1 did not result in any increase of cis content with Ru-1 (Table 1, entries 4 and 5) nor with Ru-2 (Table 1, entry 6). Of note, increasing the temperature with Ru-1 allowed the production of poly-1 with higher molar masses, but did not significantly affect the cis content. By contrast, methyl-capped monomer 1 in combination with Ru-3 afforded poly-1 with 54% cis content (Table 1, entry 7). Interestingly, when the reaction time was decreased to 4 h or 2 h, the molar masses of poly-1 did not significantly decrease (M_n = 12.6 kg/mol and 11.8 kg /mol, respectively), but the cis selectivity improved to 62% and 73% (Table 1, entries 8 and 9). To our delight, running the polymerization at room temperature to further increase kinetic control over the polymerization and at lower catalyst loading (0.5 mol%) afforded perfect cis selectivity (M_n = 13.5 kg/mol, Đ = 1.88, Table 1, entry 10). The low catalyst loading permitted by the high reactivity of Ru-3 is promising for large-scale applications. Notably, Ru-4 furnished an all-cis polymer of similar size (M_n = 13.0 kg/mol, Ð = 1.75) when subjected to the same conditions (Table 1, entry 11), which is consistent with the step-growth mechanism in Fig. 1 in which only the activation of the precatalyst differs between Ru-3 and Ru-4. Finally, the tolerance to trans impurities in the monomer batch was probed. Indeed, trans alkenes are known to react slower with stereoretentive dithiolate catalysts than the cis congeners due to the steric clash between one substituent of the ruthenacycle and the NHC ligand,²² which would potentially allow the use of less expensive, stereochemically impure monomers. Impressively, polymerization of 95% cis 1 led to poly-1 with >99% cis content and a slightly diminished molar mass (M_n = 7.7 kg/mol, Đ = 1.63, Table 1, entry 12). This finding should open the door to the synthesis of a variety of monomers through diverse synthetic routes (e.g., hydrogenation of alkynes with Lindlar’s catalyst, Wittig olefination, etc.) without the need for perfect diastereoselectivities.

Fig. 2. Monomers and Ru catalysts.

a. Readily accessible methyl- and ethyl-capped cis,cis-monomers with ester, carbonate, sulfite, and ether functional groups from commercially available reagents cis-hex-4-en-1-ol and cis-oct-5-en-1-ol. b. Catalysts screened in stereocontrolled ADMET including dichloro (Ru-1 and Ru-2) and dithiolate (Ru-3 and Ru-4) Ru carbenes.

Table 1.

Optimization of stereoretentive ADMET with dithiolate Ru catalysts

entry	monomer	[Ru] (mol%)	Temperature (°C)	Time (h)	Mn (kg/mol)a	Đ b	cis (%)c
1	14	Ru-1 (5)	50	16	13.1	2.61	23
2	14	Ru-3 (5)	50	16	–	–	–
3	14	Ru-4 (5)	50	16	–	–	–
4	1	Ru-1 (5)	50	16	4.3	1.50	29
5	1	Ru-1 (5)	80	16	11.8	2.76	26
6	1	Ru-2 (5)	50	16	13.4	2.00	21
7	1	Ru-3 (5)	50	16	13.1	2.13	54
8	1	Ru-3 (5)	50	4	12.6	1.74	62
9	1	Ru-3 (5)	50	2	11.8	1.81	73
10	1	Ru-3 (0.5)	23	2	13.5	1.88	>99
11	1	Ru-4 (0.5)	23	2	13.0	1.75	>99
12	1 (95% cis)	Ru-3 (0.5)	23	2	7.7	1.63	>99

^aM_n determined through size exclusion chromatography (SEC) in THF against polystyrene standards (RI detection).

^bM_w/M_n.

^cCalculated using ¹H NMR analysis.

Under optimized conditions (Fig. 3a), all-cis poly-2, poly-3, and poly-4 were also obtained from monomers 2, 3, and 4, respectively, displaying the excellent functional group tolerance of stereoretentive ADMET (Fig. 3b). These ADMET conditions are also milder compared to traditional ADMET processes that are generally conducted at higher temperatures for several hours or even days. After determining suitable conditions for the polymerization of methyl-capped monomers, the ethyl-capped monomers were found to only require moderate increases in reaction time (4 h instead of 2 h) and catalyst loading (1–2 mol% instead of 0.5 mol%) to access similarly sized all-cis polymers, poly-5–8 (Fig. 3c) (see Supplementary Tables S5-S8). The ability to use different capping groups with similarly high cis selectivity should broaden the scope of monomers for the stereoretentive ADMET.

Fig. 3. Polymer scope of stereocontrolled ADMET.

a. Reaction scheme to prepare polymers from either Ru-1 (left) and Ru-3 (right) to produce trans-rich and cis-rich polymers, respectively. b. Polymers from methyl-capped monomers (poly-1–poly-4). c. Polymers from ethyl-capped monomers (poly-5–poly-8).

The lower cis-selectivity observed at 50 °C with Ru-3 prompted an investigation of temperature modulation as a means to dictate the content of cis and trans olefins in polyalkenamers. Decreasing amounts of cis olefins (from ~70% to ~30%) were obtained when the ADMET polymerization was performed at 50 °C or 80 °C for increased reaction times (6–16 h) (Fig. 4a and Supplementary Tables S1-S8). Olefin isomerization has been reported in ADMET processes with Ru metathesis catalysts containing an NHC ligand such as Ru-2. Several mechanisms have been postulated for this isomerization, including a Ru–H species and metal allyl complexes derived from decomposed catalyst.^38-40 Notably, these pathways generally lead to concomitant alkene migration, which was indeed observed using Ru-2 (Supplementary Fig. 5). However, hydrolysis of poly-1 revealed that olefin migration does not take place with stereoretentive Ru-3, neither at room temperature, nor at higher temperatures (Fig. 4b). Hydrolytic degradation of cis-rich poly-1, poly-2, poly-3, poly-5, poly-6, and poly-7 under basic conditions cleanly delivered the cis isomer of oct-4-ene-1,8-diol (15) or dec-5-ene-1,10-diol (S8). While the exact mechanism of isomerization remains unclear, modulation of the cis/trans content without olefin migration was successfully performed with all monomers thereby providing access to a library of polyalkenamers with tailored backbone geometries from single building blocks. Importantly, the lack of olefin migration engenders more reproducible polymer properties.⁴¹ All polymers were soluble in several organic solvents including dichloromethane, chloroform, and tetrahydrofuran, which should facilitate the processability of these materials that are structurally analogous to industrially-relevant polymers.

Fig. 4. Selectivity in stereocontrolled ADMET.

a. Preparation of polyalkenamers with different cis/trans content via ADMET. Reactions run at 80 °C with either Ru-1 or Ru-3 led to more trans alkenes being formed while polymerizations carried out at room temperature with Ru-3 cis-rich polyalkenamers. b. Hydrolysis of poly-1 (99% cis from Table 1, entry 10 and 54% cis from Table 1, entry 7; see General hydrolysis procedure in the Methods section for conditions) to exclusively obtain cis or trans isomers of diol 15 indicates that olefin migration does not take place with Ru-3.

The thermal properties of the synthesized polyalkenamers with varying cis/trans ratios were analyzed via thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) (Fig. 5a for selected polymers and supporting information). While the range of molar masses (~7–25 kg/mol) resulting from the ADMET step-growth process renders the deconvolution of the impact of olefin stereochemistry challenging in some cases, comparison of polyalkenamers of similar sizes revealed enhanced stability with an increase in cis content (Fig. 5b-d). For poly-5 with 94% cis double bonds (M_n = 15.9 kg/mol), a decomposition temperature (T_d) recorded at 5% mass loss was observed at 361 °C compared to 357 °C for 62% cis (M_n = 23.1 kg/mol) and 330 °C for 31% cis (M_n = 19.0 kg/mol) (Fig. 5b). The T_d of poly-6 decreased by 46 °C when going from 99% cis to 16% cis for almost identical molar masses (Fig. 5c). Poly-8 exhibited a T_d of 390 °C for the 93% cis polymer (M_n = 13.9 kg/mol) as opposed to 374 °C for a 67% cis polymer (M_n = 16.9 kg/mol, Fig. 5d). Analogously, all other synthesized polyalkenamers had thermal stabilities that varied with cis content (Supplementary Tables 9-16). DSC revealed a clear influence of cis-alkene content on thermal transitions as depicted in the thermograms in Fig. 5d-f, which is in line with reports by Buchard⁴² and Dove and Becker.^3,43 In addition to small variations of glass transition temperatures (T_g), striking differences in crystalline behavior were observed. For example, cis-rich poly-5, poly-6, and poly-8 appeared amorphous via DSC, while the same polyalkenamers with a majority of trans olefins exhibited both a melting (T_m) and a crystallization temperature (T_c). Poly-5 with only 31% and 21% cis olefins displayed a T_m at 67 and 86 °C, respectively (Fig. 5e). The only sample of poly-6 to present a T_m (38 °C) contained only 16% cis linkages (Fig. 5f). Meanwhile, all other samples of poly-6 had no observable thermal transitions. Poly-8 showcased the most diverse thermal profile as a function of alkene stereochemistry. No thermal transitions could be detected with the 93% and 67% cis forms, but T_ms and T_cs were measured for samples with decreased cis content (38% and 20%) (Fig. 5g). DSC analysis of the two trans-rich poly-8 samples revealed sharp T_ms at −6 °C (20% cis) and −19 °C (38% cis). Other synthesized polyalkenamers such as poly-7 displayed amorphous characteristics regardless of cis content with no thermal transitions detected in the DSC (Supplementary Fig. 22). While the nature of the polar groups and the number of methylenes in the repeating unit clearly impacts the phase transitions of the polyalkenamers, controlling the geometry of alkenes throughout the backbone provides an additional handle to tune the thermal properties of a variety of polymers.

Fig. 5. Relationship between polymer stereochemistry and thermal properties.

(a) Selected polymers poly-5, poly-6, and poly-8 (see TGA and DSC data for all other polymers reported in the Thermal Characterization section in the Supplementary information). TGA thermograms of poly-5 (b), poly-6 (c), and poly-8 reveal a relationship between cis content and thermal stability with trans-rich polymers degrading at lower temperatures (d). DSC thermograms of poly-5 (e), poly-6 (f), and poly-8 (g) that illustrate dependence of crystalline character on the alkene geometry in the polymer backbone with only trans-rich polymers of poly-5, poly-6, and poly-8 exhibiting a T_m, while cis rich polymers have either no thermal transitions observed (poly-6 and poly-8) or display a T_g (poly-5).

Conclusion

Control over the molecular structure and physical properties of polyalkenamers—including polyesters, polycarbonates, polysulfites, and polyethers—has been achieved through stereocontrolled ADMET. This method capitalizes on methyl- or ethyl-capped diene monomers to impart both the stereochemical information and catalyst stability critical to the polymerization process. Over 99% cis selectivity was achieved for a variety of polymers including unsaturated polyesters, polycarbonates, polysulfites, and polyethers, which is a testament to the versatility of this method. The high reactivity of stereoselective dithiolate Ru catalysts towards cis monomers led to a low catalyst loading (0.5 mol%), rapid reaction at room temperature (2–4 h), and tolerance toward trans monomer impurities, which are all attractive features for large-scale implementation. Stereoretention during the polymerization process was found to be sensitive to the reaction conditions, which provided a functional handle to modulate the ratio of cis:trans alkene units within the polymer backbone. This unique stereocontrol was harnessed to produce a variety of polyalkenamers with predictable cis:trans ratios from 20:80 to >99:1 and to study the influence of alkene geometry over the thermal properties of the macromolecules. Increased thermal stability was generally correlated to increased cis olefin content and thermal transitions such as T_g, T_m, and T_c were greatly affected by the stereochemistry of the alkenes in the backbone. Of note, polyesters, polycarbonates, and polysulfites were prepared and exhibited tunable phase transition temperatures and efficient depolymerization via hydrolysis. This process provides a practical and efficient handle to control material properties for a large variety of olefin-containing polymers, which aligns well with the polymer field’s pursuit of precise macromolecular structures for soft materials with designable properties and functionality.⁴⁴

Methods

General ADMET polymerization procedure.

An oven-dried vial was charged with a Ru catalyst (Ru-1–4) and dissolved in 1,2,4-trichlorobenzene (0.1 mL) in an N₂-filled glovebox. The solution of catalyst was added to another vial containing the pre-weighed monomer (0.1–1.0 mmol). The reaction mixture was then transferred by a syringe to a Schlenk flask equipped with a stir bar and a red rubber septum wrapped with electrical tape. The flask was then removed from the glovebox and attached to a high vacuum manifold where the contents of the flask were put under vacuum (~100 mTorr) and the reaction was allowed to stir until completion at room temperature or in an oil bath preheated to the desired temperature. After the reaction was complete, the flask was placed under static vacuum and ethyl vinyl ether (50 μL) was added at room temperature and the mixture was stirred for an additional 5 minutes. The polymer was then precipitated with ice-cold methanol and isolated via centrifugation (8500 rpm, 10 minutes). The polymers were then dried under high vacuum for 16–24 h. Conversions of monomers were determined by ¹H NMR spectra of the crude reaction mixture and the cis content was quantified by deconvolution of cis and trans olefin peaks in the ¹H NMR spectra. Molar masses and dispersity were characterized by size exclusion chromatography. The configuration of olefins on polymers was further confirmed by the isolation of the corresponding cis or trans diols (15 or S8) via basic hydrolysis of the polymer (poly-1–3 and poly-5–7) (procedure described in general hydrolysis procedure) or IR spectra (poly-4 and poly-8) (Supplementary Fig. 2 and 3).

General hydrolysis procedure.

~3 mg of polymer was placed in a reaction tube equipped with a stir bar, and solution of NaOH (100 mg) in a EtOH:H2O mixture (1:1, 8 mL) was added to the tube. The polymer suspension was heated to 80 °C for 2 h (16 h for poly-2 and 40 h for poly-6). The reaction was cooled to room temperature and diluted with dichloromethane (10 mL). The organic layer was washed with H₂O (5 mL), dried over sodium sulfate, and then the solvent was removed in vacuo affording diol 15 (from poly-1–poly-3) or diol S8 (from poly-5–poly-7). In the case of poly-1 and poly-5, the aqueous layer was acidified with 3M aq. HCl and then extracted with dichloromethane twice. The combined organic fraction was dried over sodium sulfate and concentrated in vacuo to obtain terephthalic acid (9).