Hydrazine-Catalysed Ring–Opening Metathesis Polymerization Of Cyclobutenes

Abstract

Materials formed by the ring-opening metathesis polymerization (ROMP) of cyclic olefins are highly valued for industrial and academic applications but are difficult to prepare free of metal contaminants. Here we describe a highly efficient metal-free ROMP of cyclobutenes using hydrazine catalysis. Reactions can be initiated via in situ condensation of a [2.2.2]-bicyclic hydrazine catalyst with an aliphatic or aromatic aldehyde initiator. The polymerizations show living characteristics, achieving excellent control over molecular weight, low dispersity values, and high chain-end fidelity. Additionally, the hydrazine can be used in substoichiometric amounts relative to the aldehyde chain-end while maintaining good control over molecular weight and low dispersity values, indicating that a highly efficient chain transfer mechanism is occurring.

Graphical Abstract

A highly efficient metal-free ring-opening metathesis polymerization of cyclobutenes using hydrazine catalysis is described. The polymerizations proceed with high stereoselectivity and show living characteristics, achieving excellent control over molecular weight, and low dispersity values.

Ring-opening metathesis polymerization (ROMP) of cyclic olefins enables the synthesis of useful materials that are all but impossible to prepare by other means.¹ For example, a variety of commercial polymers, such as Vestenamer^®, Norsorex^®, Telene^®, Metton^®, Zeonex^®, and Zeonor^®2 are produced on large scale with ROMP, while more specialized materials are being investigated for applications ranging from drug delivery³ to photolithography.⁴ The chemistry that underwrites these polymerizations is the olefin metathesis reaction in which metal alkylidenes react with olefins, thereby shuffling their alkylidene fragments to generate new olefin products. Among the most important catalytic systems for olefin metathesis are those based on W,⁵ Ti,⁶ Mo,⁷ and Ru.⁸ While the value of ROMP has been apparent for many decades, there has been a strong desire to find ROMP catalysts that are based on more earth-abundant, less expensive, and less toxic elements. In this regard, intriguing recent developments in the use of iron catalysts have been reported,⁹ although the applicability of these systems is still an open question. On the other hand, there are some polymer applications for which the presence of metal contaminants is unacceptable, such as in the areas of drug delivery, biomedical implants, or microelectronics, which thus necessitates often difficult purification procedures.¹⁰ For this reason and others, it would be of great interest to use abundant main-group elements to achieve olefin ROMP. Unfortunately, designing functionality that effectively mimics the reactivity of metal alkylidenes is not trivial. Pioneering work from the Boydston group utilized a photoredox mediated system to generate polynorbornene via a radical cationic mechanism with living characteristics and temporal control of polymer growth.¹¹ This creative work notwithstanding, the development of a broadly applicable olefin ROMP strategy that uses simple, organic catalysts remains an open challenge.

We have developed a strategy for olefin ROMP that has the potential to fill this gap. Our approach replaces the classic “Chauvin mechanism”¹² for olefin metathesis involving [2+2] cycloadditions/cycloreversions of metal alkylidenes with a [3+2] manifold involving 1,3-dipoles, the so-called “Huisgen mechanism” (Figure 1a).¹³ This strategy makes use of readily available hydrazine catalysts that condense with aldehydes to form hydrazonium ions (protonated azomethine imines), which then achieve double bond metathesis via reversible 1,3-dipolar cycloadditions (Figure 1b). We have shown that this alternative metathesis strategy can effect a range of double-bond metathesis reactions including ring-opening and ring-closing carbonyl-olefin metathesis.¹⁴ Recently, we also demonstrated the use of this strategy for the ROMP of cyclopropenes.¹⁵ However, we observed significant monomer decomposition at the temperatures required for efficient polymerization, precluding our assessment of this polymerization as a controlled process. It was thus unclear whether hydrazine-catalyzed ROMP could be used to make larger, more well-defined polymers derived from less strained but more stable monomers. Here, we demonstrate that hydrazine-catalysis can be highly effective for the well-controlled ROMP of cyclobutene monomers.

Figure 1.

Ring-opening metathesis polymerization (ROMP). a. Chauvin vs Huisgen mechanism for olefin metathesis. b. General design of a hydrazine-catalyzed ROMP reaction. c. Scalable synthesis of TDD monomers. d. Hydrazine-catalyzed TDD ROMP.

Cyclobutene ROMP is relatively underexplored in comparison to other strained olefins, such as norbornene, likely due to the paucity of strategies to synthesize cyclobutenes on scale. However, endo-tricyclo[4.2.2.0^2,5]deca-3,9-dienes (TDDs) are a class of thermally stable cyclobutenes that are easily synthesized via the one-pot 6-π electrocyclization/[4+2]-cycloaddition between 1,3,5,7-cyclooctatetraene and various dienophiles (Figure 1c). This simple reaction provides ready access to TDD monomers suitable for ROMP.¹⁶ As described herein, we have found that hydrazine-catalyzed ROMP of TDD monomers proceeds with high efficiency, high functional group and moisture tolerance, and results in polymers with excellent control over molecular weight, low dispersity values, and high chain-end fidelity (Figure 1d). Importantly, these materials are inherently metal-free, and so require no special purification to remove such impurities.

For our initial explorations, we selected TDD 1¹⁷ derived from N-2-ethylhexyl succinimide (Table 1), due to the increased solubility and decreased melting point typically seen in monomers containing a 2-ethylhexyl substituent. Our initial attempt entailed the reaction of hydrazonium 3 with TDD 1 (100 equiv) neat at 140 °C. While the resulting polymer 2 exhibited a low dispersity value (Đ) as well as excellent matching of its number-average molecular weight (M_n^exp) with the theoretical value (M_n^theo), these conditions led to only modest conversion of monomer to polymer (entry 1). Discoloration of the reaction mixture along with ¹H NMR analysis led us to conclude that the poor performance was due to oxidation of the hydrazine to the corresponding diazene with adventitious oxygen. Because such oxidations are known to be facile with free-base hydrazines, we examined the addition of exogenous acid to help ensure protonation of the hydrazine. Indeed, the addition of one equivalent of TFA¹⁸ resulted in 96% conversion of monomer, producing a polymer with excellent agreement between M_n^exp and M_n^theo and a low Đ (entry 2). We also found that the ROMP of 1 proceeds well by mixing hydrazine salt 4 and one equivalent of an aldehyde (e.g. 5, entry 3). Again, one equivalent of TFA was necessary for high conversion, but increasing TFA loading beyond one equivalent had no additional benefit (entry 4). We were able to target a range of molecular weights by altering the cyclobutene to aldehyde ratio (entries 4–6), as long as we kept the amount of TFA at one equiv per 100 equiv of monomer. In each case, the molecular weight distribution exhibited a low Đ, and the M_n^exp matched well with M_n^theo, suggesting that one aldehyde molecule leads to one polymer chain. Importantly, no reaction occurred in the absence of hydrazine 4 or aldehyde 5, demonstrating that this is not a Brønsted acid-mediated process (entries 7 and 8).

Table 1.

Optimization of hydrazine-catalyzed ROMP of TDD 1.^[a]

Entry	Initiator	Monomer (equiv)	TFA (equiv)	Conv (%)	Mntheo (kg/mol)	Mnexp (kg/mol)	Đ
1	3	100	–	48	15.1	14.8	1.08
2	3	100	1	96	30.2	31.5	1.07
3	4 + 5	100	1	96	30.2	32.3	1.04
4	4 + 5	100	2	96	29.6	33.2	1.04
5[b]	4 + 5	200	2	98	58.4	62.2	1.07
6[c]	4 + 5	300	3	94	90.0	90.7	1.09
7	4	100	1	0	–	–	–
8	5	100	1	0	–	–	–

^[a]Conditions: aldehyde 5 (1.0 equiv), monomer 1, and catalyst 4 (1.0 equiv) were heated in a sealed vial under an N₂ atmosphere for 4 h. Conversion determined by ¹H NMR analysis. M_n^exp and Đ determined on purified products via SEC-MALS. The olefins generated by the polymerization were formed exclusively as the E-isomers with an anti relative configuration in all cases.

^[b]Reaction time = 6 h.

^[c]Reaction time = 8 h.

When monitoring the reaction, we observed a linear relationship between M_n^exp and monomer conversion, which demonstrates that this polymerization proceeds through a chain-growth process (Figure 2a). Furthermore, plotting the natural log of monomer depletion versus time showed a linear relationship corresponding to well-behaved first-order kinetics (Figure 2b). Together, these results support that this reaction is characteristic of a well-controlled polymerization.

Figure 2.

Kinetic profile of hydrazine ROMP of cyclobutene 1. a. The molec1.09ular weight of 2 grows linearly with conversion, indicating a chain-transfer mechanism. b. The semilogarhythmic plot of monomer conversion versus time shows first-order reaction kinetics.

We found that a variety of TDD monomers, including those with imide (1 and 6) or ester functionalities (7), result in efficient polymerization (Table 2). In addition, polymerization could be initiated from different types of aldehydes, including aromatic (5), aliphatic (8), and linear aliphatic dialdehydes (9). Notably, dialdehyde 9 leads to the productive ROMP of monomer 1, forming a telechelic polymer.

Table 2.

Scope investigations for hydrazine-catalyzed cyclobutene ROMP.^[a]

Entry	Monomer	Aldehyde	TFA (equiv)	Conv (%)	Mntheo (kg/mol)	Mnexp (kg/mol)	Đ
1	1	5	1	96	30.2	32.3	1.04
2[b]	1	8	1	97	30.5	31.8	1.05
3[c]	1	9	2	99	31.2	35.3	1.06
4[d]	6	5	1	97	25.1	33.9	1.14
5	7	5	2	98	24.5	38.6	1.06

^[a]Conditions: aldehyde (1.0 equiv), monomer (100.0 equiv) and catalyst 4 (1.0 equiv) were added to a sealed vial and heated for 4 h under an N₂ atmosphere. Conversion determined via ¹H NMR analysis. The olefins generated by the polymerization were formed exclusively as the E-isomers with an anti relative configuration in all cases. M_n^exp and Đ determined on purified products via SEC-MALS.

^[b]Reaction performed using 1.0 equiv of a preformed initiator via condensation of catalyst 4 with aldehyde 8.

^[c]Reaction performed using 1.0 equiv of a preformed initiator via condensation of catalyst 4 with aldehyde 9.

^[d]Reaction run in 1,2-dichlorobenzene (1.0 M) for 12 h.

The low Đs observed for these polymerizations suggest that initiation is faster than propagation. Furthermore, because the hydrazine/aldehyde condensation step is reversible, we hypothesized that chain transfer by hydrolysis of the growing chain-end followed by recondensation of the hydrazine with another polymer should enable the use of catalytic loadings of hydrazine 4. To test this hypothesis, we combined aldehyde 5, cyclobutene monomer 1, TFA, and 25 mol% of hydrazine catalyst 4 (Table 3, entry 1). Under these conditions, we observed efficient polymerization to 89% conversion in 6 h, generating a polymer with a narrow Đ of 1.16 and good agreement between M_n^exp and M_n^theo. Importantly, the addition of 20 equivalents of water resulted in better agreement between M_n^exp and M_n^theo (entry 2), further supporting our hypothesis that chain transfer is likely occurring by a hydrolysis/recondensation mechanism. The addition of water did not impact the kinetics of the polymerization.

Table 3.

Demonstration of catalytic hydrazine.^[a]

Entry	hydrazine (mol%)	Monomer (equiv)	H2O (equiv)	Conv (%)	Mntheo (kg/mol)	Mnexp (kg/mol)	Đ
1	25	50	–	89	14.1	18.4	1.16
2	25	50	20	78	12.3	15.6	1.11

^[a]Conditions: aldehyde 5 (1.0 equiv), monomer 1 (50.0 equiv) and catalyst 4 (0.25 equiv) were added to a sealed vial and heated for 6 h under an N₂ atmosphere. Conversion determined via ¹H NMR analysis. M_n^exp and Đ determined on purified products via SEC-MALS. The olefins generated by the polymerization were formed exclusively as the E-isomers with an anti relative configuration in all cases.

The resulting chain-end provides an opportunity for functionalization and chain extension after polymerization. Subjecting the isolated polymer 2 (M_n^exp = 15 kg/mol) to the standard conditions and 150 equivalents of additional monomer generated a 41 kg/mol polymer with a Đ = 1.11 (Figure 3, 10). Importantly, the polymer underwent a unimodal shift to higher molecular weights with excellent matching between M_n^exp and M_n^theo and a low Đ, indicating efficient chain extension and high chain-end fidelity.

Figure 3.

Demonstration of polymer chain extension.

Previous studies from our group have shown that the ring-opening carbonyl-olefin metathesis of TDDs^14g or other cyclic olefins^14a,d is highly stereoselective, resulting in only trans olefins. This outcome is a result of the preference for cycloaddition of the (E)-hydrazonium isomer via an exo-transition state, leading to cycloadduct 12¹⁴ and thence (E)-olefins via stereospecific cycloreversion to hydrazonium 13 (Figure 4). In addition, despite the necessarily syn configuration of the fused-ring cyclobutene stereocenters of the TDD monomer, the relative orientation of these sites upon ring-opening is exclusively anti, resulting from epimerization of the intermediate hydrazonium ion (or the free aldehyde) to the more stable configuration 14. Consistent with our previous study,^14g NMR analysis of polymer 2 reveals that the relative orientation of the allylic protons is indeed anti.¹⁹ To investigate the impact this high stereoselectivity imparts on the thermal properties of the resulting polymers, thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) were performed (see ESI for experimental details). Poly-7 was found to have a decomposition temperature (T_d) of 294 °C and a glass transition temperature (T_g) of 178 °C. In contrast, previously reported TGA and DSC analyses of poly-7 synthesized using the Grubbs catalyst indicate polymer decomposition begins below 300 °C, with no T_g observed below 250 °C.^16a The significant difference in T_g is likely due to orthogonal stereselectivity of these two methods.

Figure 4.

Stereochemical rationale.

In summary, we have developed an efficient hydrazine-catalyzed ROMP of TDD monomers. The reaction occurs via a chain-growth mechanism with living characteristics. The resulting polymers show good agreement between M_n^exp and M_n^theo, low Đs, and high chain-end fidelity. Most importantly, the polymers are obtained inherently free of metal contaminants, obviating the need for the purification procedures that are typically needed to remove these impurities. As such, this chemistry should facilitate the use of ROMP in applications that require metal-free materials. Furthermore, while this reaction does not currently allow for the polymerization of common monomers such as norbornenes, we expect that the design of more reactive hydrazine catalysts should enable a broader application of this process.