Convergent Synthesis of Branched Metathesis Polymers with Enyne Reagents

Abstract

Branched polymers have found utility in an array of fields due to the high density of functional groups combined with unique physical properties. Despite the abundant methods reported to synthesize various branched structures, controlling parameters such as the location of branch points and molecular weight distribution still remains a challenge. This report explores the ability of enyne-containing branching agents to synthesize star and miktoarm star polymers through a convergent synthesis pathway using ring-opening metathesis polymerization (ROMP). The branching agents contain an enyne metathesis terminator covalently linked to a norbornene monomer. When these agents are introduced into a living ROMP, macromonomers are generated in situ that continue to propagate via a grafting-through process with the remaining living chains. This strategy permits control over the degree of polymerization of the star arms, control of the number of star arms, and chain-extension with additional monomer to produce functional asymmetric miktoarm star polymers.

Graphical Abstract

INTRODUCTION

Branched polymers have attracted significant attention since the middle of the 20^th century due to their unique properties.^1–3 In comparison to linear polymers with the same molecular weight, branched polymers possess a higher density of functional groups, lower viscosities, and higher solubilities.⁴ Various complex polymer architectures have been reported to study structure-property relationships in application areas such as rheology modifiers, additives in coatings, and diagnostics.⁵ While star polymers and linear graft polymers have been extensively studied, the preparation of well-defined polymers with higher orders of branching remains a challenge if features such as molecular weight, degree of branching and branch position within a target architecture need to be precisely controlled.

To achieve this goal, a living chain-growth process is desirable. Ring-opening metathesis polymerization (ROMP) has emerged as the most versatile technique to synthesize high molecular weight bottle brush polymers via the “grafting through” of macromonomers.^6–8 Using this approach, well-defined block bottlebrush copolymers, Janus bottlebrush copolymers^9,10 and super-soft elastomeric bottlebrush networks^11–13 have readily been prepared. Despite the proclivity of ROMP to proceed in sterically constrained environments, exploration of additional branched architectures has only received limited attention. The synthesis of star polymers has been largely hindered by the inability to readily synthesize multifunctional ruthenium carbene initiators. Recently, Barnes and coworkers achieved this goal by adding the Grubbs 3^rd generation initiator (G3) to a norbornene-functionalized γ-cyclodextrin.¹⁴ Johnson and co-workers¹⁵ beautifully used a core cross-linking strategy^16–18 to generate highly functional brush-arm star polymers for use in a variety of biomedical applications.^19,20 Bates has developed an alternative approach to synthesize star and asymmetric miktoarm star polymers by graft-through polymerization of macromonomers with a low degree of polymerization (DP).²¹ It was found that these polymers have properties nearly identical properties to analogous star polymers with the same number of arms when the backbone DP is less than 12.²²

General synthetic strategy^23–29 for the preparation of highly branched polymers was pioneered by Knauss and coworkers in 2000 using anionic polymerization of styrene (Scheme 1A).^30–32 In this approach, a styrene monomer functionalized with a terminating silyl chloride is added to a living anionic polymerization of styrene in substoichiometric quantities. Upon reaction of the silyl chloride with an anionic polystyrene chain, a macromonomer is generated in situ that can react with the remaining active polymers in a graft-through fashion to form branches. By controlling the timing of addition and the quantity of branching agent added, a variety of architectures could be prepared including hyperbranched star-shape polystyrene, radially linked star-b-linear polystyrene, and star-b-linear-b-star triblock (Pom-Pom) triblock polystyrene.^33,34 In contrast to uncontrolled initiator-monomer chain-growth branching strategies, each living branched polymer contained a single active chain-end and permitted the generation of targetable molecular weights with low dispersities.

Scheme 1.

Approaches to synthesize branched polymers via convergent pathway.

The considerable success of ROMP to synthesize high molecular weight bottlebrush polymers via grafting-through methods presents a promising opportunity to utilize Knauss’ convergent synthesis concept to metathesis-derived materials. Recognizing this potential, Kilbinger and co-workers³⁵ first explored the convergent branching strategy with ROMP in 2016 through the use of an exo-norbornene imide monomer linked to an enol ether terminator. Unfortunately, the internal olefin of the enol ether displayed slow rates of termination relative to the propagation of the norbornene monomer, resulting in an inseparable mixture of low molecular weight linear, three-arm star, and semidendritic branched polymers. Since this report, significant advancements in termination methods^36–40 for efficient termination of metathesis polymer has been developed,^41–43 prompting the convergent branching strategy to be revisited using ROMP. In 2018, our group reported the ability of functional enyne reagents to terminate of living metathesis polymers through a cascade metathesis sequence^44–46 initiated by rapid addition to the terminal alkyne (Scheme 1B).⁴⁷ In addition to introducing a range of small functional molecules, this method could also promote stoichiometric coupling to enyne-terminated polymers to directly afford block polymers. In this report, a norbornene-derived monomer covalently bound to an enyne terminator is studied to further expand the scope of ROMP using a convergent branching strategy through the accelerated rates of enyne-mediated termination (Scheme 1C).

EXPERIMENTAL SECTION

Materials:

All reactions were carried out under a nitrogen atmosphere with dry solvents using anhydrous conditions unless otherwise stated. Dry, degassed dichloromethane (CH₂Cl₂), dichloroethane (DCE), and tetrahydrofuran (THF) were obtained from a JC Meyer solvent purification system. CDCl₃ from Cambridge Isotopes was stored under 4Å molecular sieves to remove water and acid. Grubbs 3^rd generation catalyst (G3) was prepared following literature procedures.⁴⁸ Unless otherwise stated, all other reagents were purchased at the highest commercial quality and used without further purification. Yields refer to chromatographically and spectroscopically (¹H-NMR) homogeneous materials. Reactions were monitored by thin layer chromatography (TLC) carried out on 0.25 mm E. Merck silica gel plates (60F-254) using UV light as the visualizing agent and basic aqueous potassium permanganate (KMnO₄), and heat as developing agents. SiliCycle silica gel (60, particle size 0.043–0.063 mm) was used for flash column chromatography.

Instrumentation:

NMR spectra were recorded on Bruker Avance 400, 500 or 700 instruments and calibrated using residual undeuterated solvent as an internal reference (CHCl₃ @ 7.26 ppm ¹H NMR, 77.16 ppm ¹³C NMR). The following abbreviations (or combinations thereof) were used to explain the multiplicities: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, br = broad. Mass spectra (MS) were recorded on LC/MS (Agilent Technologies 1260 Infinity II/6120 Quadrupole) by ESI. Polymer samples were analyzed using a Tosoh EcoSEC HLC 8320GPC system with TSKgel SuperHZM-M columns eluting CHCl₃ containing 0.25% NEt₃ at a flow rate of 0.45 mL/min. All number-average molecular weights and dispersities were calculated from refractive index chromatograms using PStQuick Mp-M polystyrene standards.

General Procedure for Branching and Extension:

A degassed monomer (M1) solutions (250 μL, 25.3 mg, 0.1 mmol, 10 equiv. in DCE) was added to 8 ml vial, equipped with a stir bar under nitrogen. A solution of Grubbs third generation (G3) catalyst (250 μL, 7.26 mg, 0.01 mmol, 1.0 equiv in DCE) was rapidly transferred to the stirred monomer solution and control via using microliter syringe. After 5 min, a solution of branching agent exo-T1 and exo-T2 (400 μL, 0.009 mmol, 0.9 equiv. in DCE) was added to the reaction and stirred 4 hours. 100 μL reaction mixture was taken and quenched with ethyl vinyl ether (EVE) for SEC measurement and second monomer (M2) solution (40 mg in 500 μL DCE) was added to remaining solution for additional 2 hours. The reaction mixture was quenched with EVE and precipitated into THF and MeOH, separately. The purified polymer was characterized by using SEC and ¹H NMR.

RESULTS AND DISCUSSION

In the convergent strategy, the overall degree of branching is determined by the quantity of terminating monomer introduced into the polymerization. When a branching agent is added, a portion of the living polymers are converted into macromonomers, which then continue to polymerize with the remaining living chains in a graft-through process. This requires a substoichiometric amount of branching agent to be successful, with the number of new branch points formed increasing as the number of branching agents approaches the number of living polymers in the system. A visual depiction of the process is shown in Scheme 2 for a representative polymerization using ROMP in which 9 molecules of branching agent (orange circular sector designates terminator) are added to a polymerization mixture containing 10 living polymers (P1, green triangle represents ruthenium alkylidene). In an ideal sequence of events, this would result in a short bottlebrush with a number average degree of polymerization (DP) of 10 (P1₁₀). Bates and coworkers^21,22 have shown that bottlebrushes of this length behave like star polymers with the same number of arms, and these branched polymers will be referred to as star polymers in this work. To achieve this ideal course of events using enyne termination, the rate of addition to the terminal alkyne by the living polymer will need to sufficiently outcompete addition to the norbornene. While branching can still occur in a graft-to process, this also introduces the possibility of undesired termination reactions.

Scheme 2.

Visual depiction of convergent branching using ROMP.

To investigate this concept, the initial branching agent exo-T1 was prepared through an EDC coupling between exo-N-hydroxyethyl norbornene imide and a carboxylic acid modified enyne terminator (Figure 1, see Supporting Information for synthesis). Initial branching experiments were performed using living oligomeric arms (DP 10) prepared from the addition of Grubbs 3^rd Generation initiator (G3) to exo-benzyl norbornene imide (M1) in dichloroethane (DCE). After stirring for 5 minutes, 0.9 equivalent of exo-T1 was added to the reaction mixture, stirred at room temperature for four hours, and quenched with an excess of ethyl vinyl ether (Figure 2A). Analysis of the resulting polymer by size-exclusion chromatography (SEC) revealed a substantial increase in molecular weight, though a minor peak remained at elution times similar to the oligomer and implied incomplete conversion of DP 10 arms (P1) (Figure 2B). The percentage of oligomer converted was calculated by the integration of the two elution peaks after deconvolution to determine 56% of the oligomeric arms were incorporated into the final star product. The observation of unreacted oligomers was surprising, since undesired initial ring-opening of the norbornene could still lead to higher molecular weight branched products. Further consideration of the branching agent structure with molecular models revealed the possibility of intramolecular macrocyclization after ring-opening of the norbornene (Figure 2C). Given the relatively dilute reaction conditions (enyne concentration of 10 mM), this was proposed as a possible explanation. To test this hypothesis, the experiment was repeated with a final enyne concentration of 100 mM to promote intermolecular branching reactions. This alteration led to increase in conversion of the oligomeric arms to 67%, supporting macrocyclization, though a substantial low molecular weight shoulder still remained. (Figure 2B).

Figure 1.

Branching agents, monomers and initiators used in convergent branching.

Figure 2.

(A) Convergent branch process with 0.9 equiv. exo-T1. (B) Branched polymer formation with exo-T1; (C) Proposed macrocyclic side-reaction.

To prevent the possibility of intramolecular cyclization immediately after norbornene ring-opening, a 2^nd generation branching agent (exo-T2) was prepared that replaces the flexible ester linker with a norbornene imide directly attached to the sulfonyl arene (Figure 1). In addition to rigidity, studies by Grubbs have shown that the N-phenyl norbornene imide monomers have a lower polymerization rate than N-alkylated imide monomer,⁴⁹ potentially offering additional selectivity for the alkyne over the norbornene. Repeating the same branching experiment in Figure 3A with exo-T2 resulted in complete consumption of the starting oligomer P1 after four hours and the generation of a new monomodal elution peak by SEC analysis with a number average molecular weight of 10.6 kDa and dispersity of 1.4 (Figure 3B). While this new branching agent appeared to resolve the issue of immediate macrocycle formation, termination events could still occur through formation of larger macrocyclic loops upon further propagation (Figure 3C). It is difficult to detect these loops through ¹H NMR and SEC analysis, so 23 equivalents of exo-methyl norbornene imide monomer (M2) was added relative to the initially added G3 to form an asymmetric miktoarm star polymer (P1₁₀-P2). This quantity of monomer was found to be sufficient to provide resolution of the living stars from the deactivated stars by SEC, and indeed a portion of the star-polymers did not reinitiate. The different solubilities of the methyl and benzyl norbornene blocks could be utilized to isolate the pure miktoarm star polymer after precipitation and permit additional characterization of the branched product.

Figure 3.

(A) Illustration of graft-through branch and extension. (B) SEC traces of polymers by using exo-T2. (C) Proposed mechanism for large cyclic formation. (D) ¹H NMR spectrum of P1₁₀-P2 showing chain-end in CDCl₃.

The ¹H NMR spectrum of purified P1₁₀-P2 contained a number of unique proton environments to enable further characterization of the structure and determine the efficiency of the branching process (Figure 3D). The styryl protons from each initial oligomer (red circle) were observed at 6.59 ppm and the peaks at 4.45 ppm could be assigned to the methylene protons (green triangle) at the chain-extended terminus. The relative integration of these peaks indicated the number of N-benzyl norbornene arms in the polymer to be 7.4, slightly below the theoretical 10 arms expected from the amount of added enyne. Integration of the tosyl protons at 8.0 ppm (blue square) relative to the initiating styryl protons confirmed that 0.9 equiv exo-T2 was incorporated into the polymer, and implied that a small quantity (0.05 equiv) of exo-T2 copolymerized but did not result successfully terminate. These results support the hypothesis that the rigid branching agent is able to suppress initial termination reactions but does not prevent all instances of initial ring-opening at the norbornene center. The incomplete consumption of the terminators suggests that the enynes remain geometrically inaccessible or kinetically slow to react when embedded in the branched polymer core. This NMR data can be further combined with the relative abundance of methyl and benzyl monomers in the miktoarm structure to determine the total percent of living stars after initial branching is 42% (see S10 for full analysis).

Several strategies were pursued to improve the convergent branching process by increasing the selectivity between enyne and norbornene additions, and these results are summarized in Table 1. Attempts to increase or decrease the concentration of the reaction did not lead to significant improvements (Table 1, entry 1–3). When a higher enyne concentration of 90 mM was employed, the number of star arms increased to 10.5, which was closer to the target value but the living star percentage was not improved. The more dilute reaction at 5 mM slightly reduced the degree of branching and decreased the amount of living star to 38%. The introduction of additional pyridine ligand has been shown to improve stability of ruthenium carbene in cascade yne-yne metathesis polymerization and also reduces overall rates of propagation.⁵⁰ The addition of 3.0 equiv of pyridine relative to initial G3 led to a slight increase in amount of living star to 46%, though concomitantly decreased the average number of arms to 4.9 in the miktoarm product (Table 1, entry 4). Since the terminal alkyne is less sterically hindered than the norbornene, the bulkier initiator H₂Dipp-G3 bearing isopropyl substituents was examined in hopes of improving the selectivity for initial reaction at the enyne (Figure 1, Table 1, entry 5). This improved the overall number of star arms to the theoretical value of 10 and the percentage of living stars slightly increased to 45%, comparison to entry 1. Lastly, an endo-T2 isomer was synthesized and examined under the standard reaction conditions (Figure S12). Guironnet determined that the ruthenium alkylidene cycloaddition onto endo-norbornenes has a higher transition state energy⁵¹, potentially leading to higher selectivity for an enyne-first pathway. Unfortunately, the overall reduction in norbornene reactivity suppressed complete propagation in the graft-through stage to give incompletely converted star intermediates.

Table 1.

Branching Condition Optimization Data

Entry	Branching Agent (equiv.)	Concentration (branching agent)	Living Star%a	Number of star-arm (narm)b
1	exo-T2 (0.9)	0.09M	42%	10.5/10
2	exo-T2 (0.9)	0.01M	42%	7.4/10
3	exo-T2 (0.9)	0.005M	38%	7.0/10
4 c	exo-T2 (0.9)	0.01M	46%	4.9/10
5 d	exo-T2 (0.9)	0.01M	45%	11.2/10
6	endo-T2 (0.9)	0.01M	53%e	n.d

^aBranching conversion is calculated based on SEC deconvolution technique.

^bNumber of star arm is estimated by ¹H NMR of purified miktoarm star polymer.

^cBranching with 3.0 equiv. pyridine.

^dBranching with H₂Dipp-G3 catalyst.

^eCalculated from star formation step based on deconvolution analysis of SEC.

Despite the lack of significant success in these optimization experiments for improving living star percentages, the ability to easily isolate pure miktoarm products from this process encouraged further studies with exo-T2 following conditions from Entry 2 (Table 1) to determine the impact of number of enyne equivalents (Figure 4A–C) and molecular weight of initial P1 arms (Figure 4D–F) on the convergent branching process. As the equivalents of branching agent exo-T2 increases from 0.7 equiv to 0.9 equiv, the average number of molecular weight (M_n) gradually increases from 7.5 k Da to 11 kDa, suggesting star products with increasing numbers of arms (Figure 4A, Table 2). The experimental M_n for P1₅ and P1₁₀ is lower than the calculated value, which is expected in SEC due to the lower hydrodynamic volume of star polymers relative to linear analogs. Further ¹H NMR analysis confirms this trend with 0.7 equiv and 0.8 equiv of enyne giving 2.7 arms and 3.8 arms, respectively, and the percentage of living star decrease from 56% to 42%. An extreme condition for this system is adding 1.0 equiv of exo-T2, in which all intermediate polymers should be theoretically terminated. After chain-extension with M2, a very high molecular weight (145 kDa) polymer resulted (Figure 4C). This suggests a small number of initiation sites remained and supporting the hypothesis that some terminators remain kinetically inaccessible after being incorporated into a polymer chain. As the molecular weight of initial P1 arms increased from DP 10 to DP 30 (Figure 4D - F), the number average molecular weights of the intermediate star polymers similarly increase from 10 kDa to 28 kDa. The number of star arms decreased to 7.5 and 6.5 for DP 20 and DP 30, respectively, and the percentage of living stars slightly decreased from 38% to 37%. In all cases, the final asymmetric miktoarm polymers could be cleanly isolated via precipitation to give monomodal and low dispersity materials by SEC analysis.

Figure 4.

(A-C) Enyne equivalents effect: star, miktoarm star and purified miktoarm star polymers. (D-F) Initial arm length effect: star, miktoarm star and purified miktoarm star polymers.

Table 2.

Branching with Various Enyne Equivalents and DP of Initial Star Polymers

Entry	Branching Agent (equiv.)	Initial DP	Living Star%a	Number of star-arm (narm)b
1	exo-T2 (0.7)	10	56%	2.7/3.3
2	exo-T2 (0.8)	10	49%	3.8/5.0
3	exo-T2 (0.9)	10	42%	7.4/10
4	exo-T2 (1.0)	10	n.d.	n.d.
5	exo-T2 (0.9)	20	38%	7.5/10
6	exo-T2 (0.9)	30	37%	6.5/10

One of the desirable features of branched polymers is the high degrees of functionality provided by the large number of chain-ends. Since all of the polymers synthesized thus far contain non-functional phenyl initiating end groups, a non-terminating enyne reagent containing a methyl ester was used to prepare a functional MeO₂C-G3 initiator^52,53 for use in the convergent branching process (Figure 5). The initiator cleanly formed after reaction with G3 for 15 minutes at room temperature and was directly used to form DP 10 oligomers of M1 without further purification. Addition of 0.9 equiv of exo-T2 to these oligomers, followed by chain extension with M2 produced the arm-functionalized miktoarm star polymer MeO₂C-P1₁₀-P2. Analysis of the purified polymer by ¹H NMR and SEC analysis verified the generation of a methyl ester functionalized product in similar efficiencies to the G3 initiated system with 7.9 arms (M_n = 116 kDa, Ɖ = 1.2) and an overall star livingness of 39%.

Figure 5.

Convergent branching ROMP for MeO₂C-P1₁₀-P2

CONCLUSION

A new enyne-mediated branching agent is reported and applied to convergent branching in ring-opening metathesis polymerization for the synthesis of low dispersity star and asymmetric miktoarm star polymers. Tunable control over the average number of arms and overall arm lengths was achieved by varying the quantity of branching agent employed and degree of polymerization of macroinitiators used. Detailed analysis of the purified polymers by ¹H NMR and SEC analysis permitted an understanding of branching efficiency, characterization of side reactions, and highlighted the importance of a rigid branching agent structure. These results also highlight the limitations of the current synthetic strategy and the need for new metathesis termination methods with increased reaction rates to fully realize the potential of convergent branching processes for ROMP materials. This method also demonstrates the compatibility with in situ initiator modification to construct chain-end modified miktoarm star polymers, which opens the path to generate functional asymmetric miktoarm structures use in drug delivery and therapeutic applications.