Isolation of ROMP-Generated Macrocyclic Polyalkenamers via Polar Monomer Incorporation and Chain-End Modification

Abstract

Macrocyclic polyalkenamers formed as byproducts during ring-opening metathesis polymerization (ROMP) provide valuable insight into ring–chain equilibria and secondary metathesis processes. However, the characterization of these macrocycles is often complicated by the presence of linear polymer products with identical chemical composition. Here, we present a strategy to facilitate the isolation of macrocycles generated during ROMP by modifying chain ends using enyne metathesis chemistry and introducing polar functional groups into linear polymer chains through the addition of polar monomers. The resulting polarity difference allows nonpolar macrocycles to be separated from linear polymers by using conventional silica gel chromatography. The isolated cyclic fractions were characterized by gel permeation chromatography (GPC), NMR spectroscopy, and matrix-assisted laser desorption-ionization time-of-flight mass spectrometry (MALDI-TOF MS). Using this approach, we systematically investigated how reaction parameters influence the yield and molecular weight of macrocycles in the ROMP of cyclooctadiene (COD). The methodology also enables comparative analysis of macrocycle formation in the ROMP of cyclooctene (COE), cyclopentene (CP), and norbornene (NB). These results provide a separation-based approach for studying the macrocyclic species formed during ROMP.

Introduction

Ring-opening metathesis polymerization (ROMP) can lead to secondary metathesis reactions including intrachain metathesis reactions, so-called backbiting reactions, and interchain metathesis reactions. These side reactions are often considered undesirable because they can broaden the product molecular weight distribution and cause deviations from the target molecular weight. , Despite being formed through side reactions, macrocycles generated by backbiting (Scheme ) exhibit unique chemical and physical properties relative to their linear counterparts. − As the synthesis of macrocycles continues to advance rapidly, there is increasing interest in exploring their potential applications in materials science and drug delivery. ,− For instance, cyclic polyalkenamers can serve as additives to their linear counterparts to tune their physical properties. In the commercialized polycyclooctene (PCOE)-based rubber additive VESTENAMER, the high percentage of macrocycles is responsible for the lower melting point and viscosity.

1. Secondary Metathesis Reactions for a Propagating PBD Chain in ROMP Initiated by Grubbs Catalyst.

Several strategies have been developed to synthesize cyclic polymers. Traditional approaches typically involve end-to-end coupling of telechelic linear polymers under highly dilute conditions. , More recently, ring-expansion metathesis polymerization (REMP) has emerged as an efficient route to cyclic polymers, in which monomer insertion occurs within a metallacyclic intermediate without generating chain ends. − Because polymer growth proceeds through repeated insertion and exchange reactions, cyclic polymers can be produced directly during propagation. Importantly, backbiting can also influence the molecular weight distribution of cyclic polymers in these systems. ,

In contrast to REMP, which is capable of producing high-molecular-weight cyclic polyalkenamers, ROMP can generate low-molecular-weight macrocycles as major products when the ring–chain equilibrium is appropriately controlled. For instance, Qiao and coworkers used a second-generation Hoveyda–Grubbs (HG2) catalyst to create functionalized macrocyclic oligo­(cyclooctene)­s, while Sampson and coworkers synthesized alternating cyclic polymers using a similar approach. , Weck and coworkers successfully synthesized macrocycles using a third-generation Grubbs (G3) catalyst from a cyclooctene derivative. , In many of these studies, specific reaction conditions such as dilute reaction concentrations (ranging from 0.1 to 0.5 M) with high loadings of catalyst (ranging from 0.8% to 4%) were employed to suppress the formation of high-molecular-weight linear polymers, ultimately because it is difficult to separate macrocycles from linear components.

Despite these advances, the characterization of macrocycles generated during ROMP remains challenging. Cyclic and linear polymers produced from the same monomer possess identical chemical compositions, making them difficult to distinguish by using conventional analytical techniques. Gas chromatography (GC) can be used to detect low-molecular-weight macrocycles but becomes ineffective at higher molecular weights. − Viscometry has been employed to estimate macrocycle content in polycyclopentene (PCP), but it does not provide information on the molecular weight of the macrocycles. Recently, liquid chromatography at the critical condition (LCCC) has shown promise in separating linear and cyclic polymers with similar chemical composition, but maintaining the critical conditions can be challenging, and even small changes in mobile phase composition or temperature can disrupt the delicate balance required for LCCC. − As an alternative, interaction chromatography can be used to distinguish linear and cyclic polymers chemically and isolate the macrocycles, providing greater insight into the ring–chain equilibrium.

To better understand macrocycle formation in ROMP, it would be valuable to selectively isolate macrocycles from reaction mixtures and analyze them independently. In this work, we introduce a strategy to facilitate the separation of macrocycles formed during ROMP by selectively introducing polar functional groups into the linear chains, which facilitates the separation of macrocycles from the reaction mixture by using silica gel chromatography. Our study focuses on the ROMP of cyclooctadiene (COD) as a prime example, examining how various reaction parameterssuch as feed monomer concentrations, reaction temperatures, reaction times, solvents, catalyst loadings, and catalyst structuresaffect the molecular weight and yield of macrocycles. We also examined the formation of macrocycles in the ROMP of cyclooctene (COE), cyclopentene (CP), and norbornene (NB).

Results and Discussion

Introducing polar groups at the chain ends of linear polymers during ROMP can facilitate the clean isolation of less polar macrocycles using silica gel chromatography because macrocycles have no chain end. Thin-layer chromatography (TLC) studies have shown that hydroxyl-terminated polystyrene (PS) and polybutadiene (PBD), within a certain molecular-weight range, exhibit stronger interaction with silica gel and have smaller retention factors (R _f) than hydrogen-terminated polystyrene or polybutadiene. This example highlights how a single polar end group can significantly alter adsorption behavior. ,

More recently, a series of metathesis reactions have been developed to efficiently modify the benzylidene initiator of G3 and terminate ROMP while attaching desired end-groups. − We used this chemistry to synthesize a ROMP initiator containing a polar hydroxymethyl functional group, G3-OH, and initiated ROMP of COD under “standard conditions” ([COD] = 1.1 M in C₆D₆; [COD]/[G3-OH] = 100, 30 min, and 20 °C). G3 was first reacted with 3 equiv of 1 for 15 min. ¹H NMR spectra showed >95% conversion from G3 to G3-OH, as determined from the integration of the benzylidene proton to anthracene as an internal standard (Figure S3a,b). A stock solution of G3-OH in C₆D₆ was added to COD in C₆D₆, resulting in a COD concentration of 1.1 M and [COD]/[G3-OH] molar ratio of 100. The solution was stirred at 20 °C for 30 min. G3-OH displayed activity similar to that of G3, producing crude PBD with a similar molecular weight distribution (M _n = 10.3 kDa, Đ = 1.48, Figure S4). After the monomer was consumed, 2 equiv of 2 were added as a polymerization quencher to introduce the polar functionality into the other chain terminus. ¹H NMR spectra showed full monomer conversion, no catalyst decomposition, and full conversion of the catalyst to the inert thioether chelate structure G3-thio (Figure S3c,d). Ideally, all linear species, including products arising from interchain metathesis, would bear hydroxyl groups at both chain ends (Scheme ).

2. Synthesis of Hydroxyl Terminated Polybutadiene Using ROMP and Structure of Products .

^a ROMP of COD initiated by G3 and terminated by ethyl vinyl ether.

^b ROMP of COD initiated by G3-OH and terminated by 2.

^c TLC analysis of product was developed by 10:3 (v/v) hexanes/dichloromethane and stained by cerium ammonium molybdate stain.

The reaction mixture was initially analyzed by TLC. With 10:3 (v/v) hexanes/dichloromethane (DCM) as the eluent, two clear spots were observedone polar component, which was tentatively assigned as linear PBD with an R _f close to 0, and the other oblong spot with larger R _f as cyclic PBD (cPBD). These two spots were readily separated by silica gel chromatography. By contrast, PBD initiated with G3 under the same conditions and terminated with ethyl vinyl ether produced a continuous streak on TLC (Scheme ). This strategy also provides a potential route for removing cyclic impurities from linear polymer products. In this work, however, the focus is on isolating cyclic fractions in order to analyze the macrocycle formation during ROMP.

The reaction mixture was loaded onto a silica gel column and eluted with 10:1 to 1:10 (v/v) hexanes/DCM. A notable feature of the separation was the elution order, with macrocycles of smaller molecular weight eluting first, followed by macrocycles of higher molecular weight. Subsequently, longer linear chains were eluted, followed by shorter linear chains (Table ). Similar elution orders for macrocycles were reported previously. ,, Adsorption of linear chains on the stationary phase was primarily attributed to the hydroxyl end groups. As the molecular weight increased, the polarity of the chains was progressively diminished by the hydrocarbon backbone, resulting in larger R _f values for longer chains. By contrast, macrocycles lack polar end groups, and their interactions with the stationary phase became weaker as the molecular weight decreased, leading to earlier elution of the smaller macrocycles. McCarthy and coworkers have previously reported similar differences in adsorption behavior for hydrogen-terminated PS and hydroxyl-terminated PS.

1. Elution Order of Cyclic and Linear PBD .

	M n (kDa)	Đ	Yield (%)
F1 (cyclic)	0.7	2.86	16.1
F2 (cyclic)	1.4	1.27	4.3
F3 (cyclic)	1.9	1.04	7.0
F4 (cyclic)	4.3	1.03	2.0
F5 (linear)	14.1	1.33	18.0
F6 (linear)	10.9	1.49	26.8
F7 (linear)	8.8	1.33	24.2

^aThe table shows M _n, Đ, and yield of each fraction.

All four cPBD fractions were combined to form C1, with a yield of 29%. GPC analysis of C1 revealed a high dispersity of 2.75 and a small M _n of 0.8 kDa. The differential refractive index (DRI) trace showed a distinct peak at the elution time of trans,trans,trans-1,5,9-cyclododecatriene (ttt-CDT), followed by a monotonic decrease in the RI response at shorter elution times corresponding to larger macrocycles (Figure a). This behavior is consistent with a decreasing probability of Ru reacting with more distant double bonds along the backbone. Attempts to synthesize a linear analog with molecular weight and dispersity similar to those of C1 were unsuccessful, but the retention time of synthesized linear PBD L2 (M _n = 1.3 kDa, Đ = 1.47) was significantly shorter than that of fractionated cPBD F2 (M _n = 1.4 kDa, Đ = 1.27) of similar molecular weight (Figure b). Fractionated cPBD (acquired from another experiment) with a retention time similar to that of L2 exhibited higher molecular weight (M _n = 2.4 kDa, Đ = 1.15). All cPBD fractions F1–F4 from C1 exhibited longer retention times and thus smaller hydrodynamic volumes than their linear counterparts (Figure S5), consistent with C1 being predominantly cyclic. While end-group signals of high-molecular-weight linear PBD can be too weak to observe by ¹H NMR, high-molecular-weight linear species were not evident in the GPC trace of C1. Conversely, low-molecular-weight linear PBD would be expected to give clearly observable end-group signals, but none were detected in the ¹H NMR spectrum of C1 (Figure c). Together, these data support substantial enrichment of cyclic species and suggest that any residual linear impurities were below the detection limits of the methods used here.

1.

Characterization of cPBD. (a) GPC DRI traces showing the product distribution of C1 and the elution time of ttt-CDT. (b) GPC DRI traces comparison between cPBD and linear PBD. (c) ¹H NMR spectrum (CDCl₃) of C1 with an expansion of the 6.00–6.50 ppm region. (d) MALDI-TOF MS for F1 obtained in reflection mode. The inset shows the plot of the m/z values (y) versus the number of C₄H₆ repeat units (x) for the peak series.

The most straightforward method for the analysis of polymer end groups, MALDI-TOF MS, is challenging for the (nearly) pure hydrocarbon composition because of poor ionization. Despite these challenges, the MALDI-TOF MS for the lower molecular weight fraction F1 was obtained, revealing a single series of peaks corresponding to cPBD (Figure d, see Supporting Information for MALDI-TOF MS obtained in linear mode). Plotting the m/z values for the peaks versus the number of PBD repeating units yielded a relationship of y = 54.11x + 106.25, corresponding to the structure of [(C₄H₆)_n + Ag⁺]. Peaks corresponding to linear PBD of different end groups (Scheme ) with a degree of eight repeating units (1041.19, 910.19, and 779.18) were absent. A peak interval of 54.11 suggests that the macrocycles were formed via backbiting rather than cyclooligomerization of COD, which would have resulted in a peak interval of 108.22 or increased intensities for peaks with even numbers of repeating units in the MS.

Taken together, these characterization data support the successful isolation of the cyclic fractions from the reaction mixture. This approach nevertheless has two main limitations. First, as the degree of polymerization increases, the effect of polar end groups on longer linear PBDs would not be sufficient to separate them from the macrocycles. For instance, when the [COD]/[G3-OH] ratio was increased to 1000, linear PBD with higher molecular weight was mixed with cPBD of higher molecular weight (Figure S6). Second, the enyne metathesis chemistry that was adopted to modify the ruthenium catalyst and quench polymerization was not efficient for catalysts other than G3. , To investigate the distribution of cPBD for these more challenging cases, we modified our approach by adding exo-cis-5-norbornene-2,3-dicarboxylic acid (NDA) as a polar second monomer after the polymerization of the COD and before the reaction was quenched with 2 ([COD]/[NDA] = 10). The polymerization can be quenched shortly (within 30 s) after the addition of NDA because of its fast propagation, thus minimizing the effect on the reaction equilibrium of ROMP of COD (Figure S7). , The resulting PBD-b-PNDA, which has a much higher polarity, greatly improves the separation of cPBD through interaction chromatography.

Analysis of Macrocycles in ROMP of COD

Relative to the original method, the extra step of adding NDA did not alter the yield or molecular weight of cPBD in the “standard conditions” (Table , entry 2–1). Using this modified method, we investigated the formation of macrocycles during the ROMP of COD under various conditions. A decreased catalyst loading resulted in an increased molecular weight for linear PBD, whereas the molecular weight of cPBD remained nearly unchanged relative to the standard conditions (Table , entries 2–1 and 2–2). In order to obtain a larger cPBD, the propagating Ru should bite a farther backbone olefin from the chain end, the reactivity of which is independent of the catalyst loading. The yield of cPBD decreased slightly. This behavior may arise from fewer backbiting events at lower Ru concentration, a pre-equilibrium state caused by the slower reaction rate, or a combination of these effects.

2. Molecular Weight of Crude PBD and Yield and Molecular Weight of Macrocycles from ROMP of COD.

Entry	Change from the “standard conditions”	Crude PBD M n(M w) (kDa)	cPBD yield (%)	cPBD M n(M w) (kDa)
2–1	None	10.7(15.9)	29	0.8(2.3)
2–2	[COD]/[Ru] = 400	22.0(34.8)	28	1.0(1.7)
2–3	[COD]/[Ru] = 1000	35.6(63.0)	24	0.7(1.7)
2–4	T = 0 °C	11.7(15.6)	20	0.7(1.1)
2–5	T = 50 °C	10.3(14.0)	34	1.4(3.7)
2–6	T = 70 °C	4.5(5.5)	25	1.7(2.1)
2–7	[COD]0 = 0.27 M	3.3(5.3)	64	0.3(1.5)
2–8	[COD]0 = 0.1 M	1.2(5.2)	92	0.3(0.9)
2–9	t = 5 min	8.2(12.5)	22	1.6(3.2)
2–10	t = 300 min	9.1(15.7)	31	2.5(4.4)
2–11	THF instead of benzene	9.4(15.4)	25	0.5(1.1)
2–12	hexanes/benzene (v/v = 1/6) instead of benzene	10.9(16.1)	26	0.7(2.1)
2–13	[Ru] = G1	6.2(7.3)	8	2.7(4.3)
2–14	[Ru] = G2	11.1(17.9)	26	2.1(4.5)
2–15	[Ru] = HG2	17.3(26.9)	37	0.4(1.3)
2–16	[Ru] = G3	10.8(16.1)	42	0.6(3.1)
2–17	[COD]/[Ru] = 1000, [COD]0 = 0.27 M	12.2(22.2)	34	0.5(1.3)
2–18	[COD]/[Ru] = 1000, [COD]0 = 0.27 M	11.4(20.7)	29	0.4(1.7)
2–19	[COD]/[Ru] = 1000, [COD]/[MA] = 10, THF instead of benzene	2.3(3.8)	21	0.4(1.6)

^aStyrenic protons (ca. δH 6.23 and 6.37 ppm) were identified in the ¹H NMR spectra of separated cPBD products.

^bThe reaction was first equilibrated at 1.1 M for 15 min, then diluted to 0.27 M and equilibrated for another 15 min.

^cNDA was not added.

The influence of the reaction temperature is more complex. Among the cyclic products accessible from COD, the formation of ttt-CDT is thermodynamically the most favorable. As a result, changes in temperature alter its relative abundance more strongly than those of the larger macrocycles. , At 50 °C, the molecular weight and yield of cPBD were increased (Table , entry 2–5). This trend is consistent with the reduced relative formation of ttt-CDT and the increased formation of larger macrocycles, which may reflect the greater entropic favorability of the latter. ,, At 0 °C, the molecular weight and yield of cPBD were decreased (Table , entry 2–4). When the temperature was further increased to 70 °C, G3-OH underwent decomposition during polymerization, leading to lower-molecular-weight linear PBD and linear impurities in the isolated cPBD, which was determined by ¹H NMR analysis with characteristic signals of styrenic protons (Table , entry 2–6, see Figure S27 for ¹H NMR spectrum).

Initial monomer concentration is a key parameter governing the ring–chain equilibrium. According to the Jacobson–Stockmayer theory, a critical monomer concentration [M] _c can be established, below which only cyclic oligomers are formed. When the monomer concentration is higher than [M] _c , the equilibrium concentration of the cyclic fraction remains constant, and linear polymer chains start to grow. , When [COD]₀ was decreased to 0.27 and 0.1 M, the yields of cPBD were increased to 64% and 92%, respectively. However, the molecular weights of cPBDs were decreased (Table , entries 2–7 and 2–8) due to the absence of sufficiently long linear PBDs that can generate high-molecular-weight cPBD.

For the ROMP of COD, prolonged reaction times of up to 8 h are required to approach equilibrium. The yield of macrocycles, especially ttt-CDT, gradually increases during this time. ,, When the reaction was cut off at 5 min, the molecular weight of cPBD increased because of the lower yield of ttt-CDT but the total yield of cPBD decreased (Table , entry 2–9). However, when the reaction was run for 300 min, linear impurities which were generated by catalyst decomposition were identified in the separated cPBD (Table , entry 2–10, Figure S28).

Previous reports suggested that solvent quality has a negligible influence on the kinetic and thermodynamic distribution of macrocycles by comparing toluene, DCM, chloroform, and a theta solvent (3-pentanone/2-pentanone). , By contrast, some earlier studies reported a slightly larger fraction of macrocycles in heptane than in benzene for the ROMP of COD and COE. , Theoretical models also suggest a decreased yield of macrocycles in good solvents, particularly in a diluted solution, because the expanded chains would impede chain-end encounter. When hexanes/benzene (1:6 v/v) was used as the solvent, the molecular weight and yield of cPBD were nearly identical with those obtained under the standard conditions. During polymerization, PBD temporarily precipitated but redissolved after the addition of the NDA solution (Table , entry 2–12). Replacement of benzene with tetrahydrofuran (THF) decreased both the molecular weight and yield of cPBD. This behavior may be due to the coordination of THF to the active Ru species, which could suppress secondary metathesis processes, including backbiting. Both benzene and THF are good solvents for PBD, indicating that this effect does not arise from solvent quality. ,

The structure of the catalyst can significantly affect the reaction equilibrium in ROMP. For example, the first-generation Grubbs (G1) catalyst has been reported to initiate ROMP with minimal secondary metathesis. , When G1 was used in place of G3-OH, the yield of cPBD decreased to 8%. The product distribution also differed significantly from the standard conditions (Figure S20), with minimal formation of ttt-CDT (Table , entry 2–13). This product distribution appears more consistent with kinetic control than with thermodynamic control, possibly because G1 is less effective at polymerizing ttt-CDT. , However, ¹H NMR analysis revealed linear contaminants eluting with cPBD (Figure S29). These species may arise from interchain metathesis products, potentially including linear species bearing two styrene end groups that were not fully separated chromatographically. When the second-generation Grubbs (G2) catalyst was used, ROMP yielded cPBD with a higher molecular weight than the standard conditions, but the product also contained linear impurities (Table , entry 2–14, Figure S30). When G3 was used instead of G3-OH, the yield of nonpolar species increased to 42%, with 13% attributed to the styrene-terminated linear PBD (Table , entry 2–16, Figure S32). HG2 contains a coordinating isopropoxyaryl ligand, which may increase the propensity for backbiting and thereby favor the formation of lower-molecular-weight macrocycles. In contrast to G1, G2, and G3, no evidence for interchain-metathesis-derived linear impurities was detected for HG2 under the conditions examined here, consistent with the absence of detectable end-group signals in the ¹H NMR spectrum of the separated nonpolar product (Table , entry 2–15, Figure S31).

The yields and molecular weights of cPBD in entries 2–17 and 2–18 were similar despite different reaction conditions. Entry 2–17 was performed with reduced catalyst loading ([COD]/[Ru] = 1000) and concentration ([COD]₀ = 0.27 M), while entry 2–18 began with the same catalyst loading at a higher concentration (1.1 M) for 15 min before being diluted to 0.27 M to reach equilibrium. These experiments support the view that the ROMP of COD is strongly influenced by ring–chain equilibrium and that the product distribution is largely governed by the final reaction conditions. ,,,

Molecular weight control in the ROMP of medium-ring-strain monomers can be achieved using acyclic internal olefins as chain-transfer agents (CTAs). − In this study, maleic acid (MA) was used as a CTA ([COD]/[MA] = 10) to end-cap a linear PBD with carboxylic acid groups. Linear PBD was synthesized with low molecular weight, and cPBD was still separated with a yield of 21% (Table , entry 2–19). These results indicate that under the conditions examined here the introduction of MA as a CTA did not eliminate macrocycle formation.

Analysis of Macrocycles in ROMP of COE, CP, and NB

Compared with COD, COE ROMP has been less extensively studied in terms of the product distribution. Overall, the yield of cyclic PCOE (cPCOE) was lower than that of cPBD. This lower yield is consistent with the lower density of double bonds in the COE backbone, which provides fewer potential sites for backbiting. Since the ROMP of COE does not produce a specific thermodynamically favored product like ttt-CDT, the reaction quickly reached equilibrium in as little as 5 min and generated macrocycles with higher molecular weight, compared to the ROMP of COD (Table , entries 3–1 and 3–6). The reaction temperature barely influenced the formation of cPCOE for the same reason (Table , entry 3–3). When G3-OH was substituted with G2 or HG2, the reaction mixture gelled rapidly, likely because slow initiation led to the formation of high-molecular-weight linear PCOE, which remained insoluble even after extended reaction times. Accordingly, reactions using G2 and HG2 were conducted under more dilute conditions. Table , entry 3–8 shows nearly full conversion to cPCOE (Figure S44), which is consistent with the findings of Qiao and coworkers. The comparison between Table , entries 3–1 and 3–9 revealed that the yield of linear PCOE resulting from interchain metathesis reactions was calculated to be 7% (Table , entry 3–9).

3. Molecular Weight of Crude PCOE and Yield and Molecular Weight of Macrocycles from ROMP of COE.

Entry	Change from the “standard conditions”	Crude PCOE M n(M w) (kDa)	cPCOE yield (%)	cPCOE M n(M w) (kDa)
3–1	None	12.3(16.6)	12	2.1(4.6)
3–2	[COE]: [Ru] = 1000	77.8(107.5)	8	3.0(4.0)
3–3	T = 50 °C	12.3(16.8)	11	2.1(6.8)
3–4	[COE]0 = 0.27 M	6.2(8.7)	42	2.4(4.2)
3–5	[COE]0 = 0.1 M	4.7(5.4)	81	2.2(2.7)
3–6	t = 5 min	12.5(15.3)	12	2.4(4.8)
3–7	[Ru] = G2, [COE]0 = 0.1 M	1.2(11.0)	92	3.7(5.8)
3–8	[Ru] = HG2, [COE]0 = 0.1 M	0.6(1.5)	99	1.2(3.0)
3–9	[Ru] = G3	10.8(15.9)	19	3.0(5.6)

^aStyrenic protons (ca. δH 6.23 and 6.37 ppm) were identified in the ¹H NMR spectra of separated cPCOE products.

CP has a low ring strain of 28.4 kJ mol^–1 and the entropic penalty for its polymerization is large (−ΔS _p ≈ 70 ± 6 J K^–1 mol^–1). , At a [CP]₀ of 0.27 M, no polymerization occurred at 23 °C (Table , entries 4–3 and 4–4). Increasing [CP]₀ to 2.8 M resulted in a 5% yield of cyclic PCP (cPCP) with M _n of 1.6 kDa (Table , entry 4–2). The yield of cPCP was improved to 11% when [CP]₀ was decreased to 1.1 M, with a comparable M _n (Table , entry 4–1). Using a more reactive Mo-based catalyst, Register and coworkers reported the synthesis of higher molecular weight cPCP at lower catalyst loadings. In an attempt to produce higher molecular weight cPCP, we used a high [CP]₀ of 2.8 M and a [CP]/[G3-OH] of 10000, but the resulting gelled solution made it impossible to introduce the NDA block into the linear chain, preventing the separation of the product.

4. Molecular Weight of Crude PCP and Yield and Molecular Weight of Macrocycles from the ROMP of CP.

Entry	Change from the “standard conditions”	Crude PCP M n(M w) (kDa)	cPCP yield (%)	cPCP M n(M w) (kDa)
4–1	None	5.3(6.8)	11	1.8(2.2)
4–2	[CP]0 = 2.8 M	17.2(27.4)	5	1.6(2.1)
4–3	[CP]0 = 0.27 M	-	-	-
4–4	[Ru] = HG2, [CP]0 = 0.27 M	-	-	-

In the ROMP of NB, a high-ring-strain monomer, no macrocycles were detected under the conditions examined here. This remained true even at lower initial monomer concentration, longer reaction time, and with HG2 as the catalyst (see Supporting Information for details). The steric bulkiness of the polynorbornene (PNB) backbone results in a decreased reactivity toward backbiting. Cyclic PNB formed via ROMP has been reported only with WCl₆/Sn­(CH₃)₄ or more active Mo-based catalyst. − In other cases, a tether structure in the catalyst is required to synthesize cyclic PNB via REMP. ,− Under the conditions examined here, PNB showed no detectable macrocycle formation, consistent with a low propensity for backbiting in ROMP with Grubbs catalysts lacking a tether structure.

Conclusions

In summary, we have developed a strategy to facilitate the isolation of macrocyclic polyalkenamers formed during ROMP by selectively introducing polar functional groups into linear polymer chains. The resulting polarity difference enables the efficient separation of nonpolar macrocycles from linear species using conventional silica gel chromatography, allowing the direct characterization of cyclic fractions.

Using this approach, we systematically investigated the influence of reaction parameters on macrocycle formation in the ROMP of cyclooctadiene. The results support a major role for ring–chain equilibrium in governing macrocycle formation, with temperature, monomer concentration, reaction time, and catalyst structure exerting distinct effects on the yield and molecular weight of the cyclic products. In particular, dilution favors macrocycle formation, while, under the conditions examined here, 50 °C and shorter reaction times were associated with the formation of higher-molecular-weight macrocycles.

Extension of this method to cyclooctene and cyclopentene further highlights the role of backbone structure and double-bond density in determining the extent of backbiting, whereas norbornene was found to resist macrocycle formation under the conditions studied.

Although several limitations remainincluding reduced effectiveness for high-molecular-weight systems, limited applicability to G1- and G2-mediated reactions, linear impurities arising from catalyst decomposition, and gelationthis strategy provides a practical platform for the isolation and analysis of macrocycles generated during ROMP. More broadly, it enables detailed investigation of ring–chain equilibria and secondary metathesis processes in ROMP and related olefin-metathesis polymerization systems.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.macromol.6c00118.

• Experimental details, synthetic procedures, GPC chromatograms, NMR spectra, and MALDI-TOF MS spectrum (PDF)

Q.G.: Conceptualization, Investigation, Methodology, Validation, Formal analysis, Visualization, Writingoriginal draft. K.-Y.Y.: Investigation, Methodology, Writingreview and editing. Y.X.: Investigation, Methodology, Writingreview and editing. J.W.: Methodology, Writingreview and editing. R.T.: Writingreview and editing. M.J.R.: Supervision, Writingreview and editing. R.H.G.: Conceptualization, Supervision, Project administration, Funding acquisition.

This work is financially supported by the National Science Foundation (CHE#1807154).

The authors declare no competing financial interest.

#.

R.H.G. deceased on December 19, 2021.