More than ring-strain: revisiting the definition of enthalpy in ring-opening polymerization

Abstract

The thermodynamics of ring-opening polymerization (ROP) are central when predicting the chemical recyclability of aliphatic polyesters and polycarbonates. Conceptually, the enthalpy of polymerization, , is widely understood as a measure of ring-strain for a given monomer. However, the ring-strain is commonly larger than , generating the question of how the release of ring-strain energy during ring-opening transforms. In this work, we propose that is the sum of the energy released by the ring-strain and the energy absorbed by the polymer conformations . Owing to the similar ring-strain, but vastly different values, δ-valerolactone, δ-caprolactone, δ-octalactone, and δ-decalactone were used as model compounds to evaluate the energy cost of polymer conformational freedom. Polymer conformation, measured by ¹³C NMR, DSC, and molecular dynamics, results are in good agreement with the hypothesis and can explain previous literature observations i.e. positive for systems with ring-strain, substituent effects, and solvent effects, that are difficult to understand when only using the analogy of ring-strain and . We believe that our results provide a deeper understanding of the underlying thermodynamics and their interpretation in ROP.

---

Introduction

Chemical recycling of plastics by converting polymers into monomers is widely investigated as an option for a circular materials economy.¹ Within this field, aliphatic polyesters are often highlighted as promising, due to the inherent chemical reversibility of ring-opening polymerization (ROP). ROP allows for the synthesis of high molecular weight polymers with, for example, mechanical properties comparable to polyolefins whilst also possessing facile chemical recyclability.^2,3 To assess the chemical recyclability of polyesters, the thermodynamics of ROP are commonly used as a benchmark.^1,4–10 The thermodynamics describe the equilibrium monomer concentration, often represented as the enthalpy and entropy of polymerization at standard state conditions or and , respectively. As such, it is important that our understanding and thereby interpretation of the physical meaning of the definitions of and are clear. To support this for ROP and facilitate new monomer design through an extended perception of structure and recyclability relationships, this work revisits the conceptual understanding of using δ-valerolactone (δVL), δ-caprolactone (δCL), δ-octalactone (δOL), and δ-decalactone (δDL) as model compounds.

IUPAC defines ring-strain as the enhanced internal energy of a cyclic molecule, due to relatively unfavourable bond lengths, bond angles, or dihedral angles compared to a linear standard.¹¹ In ROP, ring-strain is generally considered an analogue of the enthalpy of polymerization among text books,^12–14 review articles,^5,6,15 and thermodynamic modelling work.^16–20 However, there are two main issues with the assumption that enthalpy equals ring-strain. (i) First, substitutions on a ring generally lead to a lower , despite having minimal effect on ring-strain energy (Table 1, with δVL, δCL, δOL, and δDL as examples). The decrease in with ring substituents has been reported, discussed, and attributed to ring-strain for many monomers,^5,8–10 without experimentally addressing changes in ring-strain and it is possible that any potential ring-strain alteration does not support the changes in . It has for example been shown that substituents on the δVL ring only marginally affect ring-strain.²¹ (ii) Secondly, the measured ring-strain is far larger than , meaning that there must be other processes that absorb energy during ring-opening polymerization.

Table 1 The experimental and calculated ring-strain, , and values for δ-valerolactone (δVL), δ-caprolactone (δCL), δ-octalactone (δOL), and δ-decalactone (δDL). Please note that ring-strain is released during polymerization and will thus be negative in sign

Monomer	Ring-strain (kJ mol^−1)
a 30% (v/v) toluene in dichloromethane.^22 b Recalculated by review.^5 c Empirically determined in toluene (1 M).^23 d Calorimetry.^24 e Empirically determined in bulk.^7 f Empirically determined in bulk.^25 g Ring-strain determined by bomb calorimetry.^26 h Calculated theoretically.^27 i Combustion calorimetry.^28 Rings strains were here calculated using molecular dynamics simulations of the ring structure compared to the strainless linear structure using the *MM3 or **CHARMM forcefield.
	39.7^g	−8.4^a	−14.7^a
	44.3^h	−10.5^d	−15.0^d
	46.0^i	−11.8^c	−29.6^c
	32.1*	−12.2^b	−28.6^b
	36.2**
	28.9*	−13.1^c	−37.0^c
	30.3**	−13.8^a	−41.2^a
		−19.3^e	−62.0^e
	34.9*	−18.5^e	−60.0^e
	32.6**
	32.5*	−17.1^f	−68.3^f
	33.6**	−18.0^e	−57.0^e
		−20.0^b	−62.5^b

---

When trying to understand what this energy absorbing process could be, we notice a trend where systems with lower values also have lower , and when plotting values for versus from literature, Table 1, a linear trend (R² = 0.92) was evident (Fig. 1a). This indicates that the process that we are looking for is inherently tied to a simultaneous entropic process. In ROP, consists of a loss of translational entropy during polymerization, as multiple monomers are consumed into a single polymer chain, and a gain in conformational entropy, as linear structures have more possible conformations than a ring structure.¹² The more negative ΔS the more ordered the system, the more positive the less ordered.

Fig. 1 (a) Plot of experimentally obtained versus values from literature, which can be found in Table 1 with data for δVL,^5,22–24 δCL,^7,22,23 δOL,⁷ and δDL.^5,7,25 (b) Free energy of polymerization calculated at 373 K for the values in (a).

With this in mind, it is reasonable to assume that the loss in translational entropy is similar for the ring-opening polymerization of δVL, δCL, δOL, and δDL, as they follow the same mechanism. This means that a change in conformational entropy will be accompanied by an enthalpic term relating to the energy states of the new conformations that absorb part of the energy released by the release of ring-strain. The system’s driving force will of course always be the minimization of , and decreases linearly with (Fig. 1b).

We hypothesize that the deviation in ring-strain energy compared to the enthalpy of polymerization relates to a second process, when the ring-strain energy is absorbed by the newly formed polymer to increase the entropy of the system in a process we denote as conformation enthalpy . Polymer conformation is defined as the arrangement of bond orientations by means of trans, −gauche, and +gauche states, which determine the 3D spatial structure of the polymer. Entropically, the polymer will target to assume equal amounts of each of the three states, generating a random coil. However, gauche states are more energy expensive than trans states, meaning that depending on the temperature at which ROP is performed, there might not be sufficient energy to overcome the energy cost of gauche states, trans states will therefore not conform into gauche states, and the resulting polymer will be extended with long end-to-end distances. Thereby, at low temperatures, where little or no energy is absorbed by the polymer, a low will be the result. is negative when , the energy released (an inherently negative value as strain is released), is greater than , the energy absorbed (eqn (1)).

Conceptually, our hypothesis can be described from the basis of two enthalpic processes, each of which is tied to an entropic process. First, the energy stored in a ring as ring-strain, drives the ring-opening, releasing ring-strain energy and this is accompanied by an entropic penalty due to the loss of translation (Fig. 2a, step 1). The monomer now attached at the chain end of the polymer attempts to obtain equal amounts of trans, −gauche, and +gauche states, to maximize entropy. However, this process is limited by the amount of energy that is available by means of temperature (Fig. 2a, step 2). We believe that ring-strain release and energy translation to polymer conformation, is a single process in practice, exemplified by . The scenario in which a ring is opened to an all trans structure, releasing all heat, is improbable at any non-0 K temperature, yet for simplicity we will here consider and as two separate processes. The energy difference for converting a trans state into a gauche state is lower for un- or less substituted polymers than polymers with larger substituents due to the additional steric effects that the substituents provide. Therefore, at low temperatures, the un- or less substituted polymers here exemplified as δVL and δCL will have more conformational freedom and a high (less negative) value, however, more gauche states will be formed and thus a high (less negative) will be a consequence. Polymers with larger substituents, exemplified here by lactones δOL and δDL, will not have sufficient energy to convert many trans states into gauche states and hence low and values are obtained. Consequently, the substituted polymers will show greater changes in and as temperature rises and gauche states become more attainable (Fig. 2b). It must be stressed that the coil formation for polymers, although driven by an entropic process, lowers the free energy of the system compared to an extended conformation.

Fig. 2 (a) Schematic representation of the hypothesis consisting of an enthalpy driven ring-strain release step and an entropically driven polymer conformation changing step. (b) Schematic of the conversion of an extended polymer into a polymer coil due to the effect of temperature on and for PδOL and PδDL with larger substituents whilst un- and less-substituted PδVL and PδCL show little change in structure with temperature.

Results and discussion

In ROP, the enthalpy of polymerization, , has been considered equal to the ring-strain of the monomer for at least the past 40 years.^5,6,12–15 Here we challenge this assumption since: (i) the ring-strain magnitude always exceeds the and (ii) monomers, such as δVL, δCL, δOL, and δDL, that have similar ring-strain values still display drastically different values. In this work, we hypothesize that part of the energy released by ring-strain, , is transferred to attain energy expensive polymer conformations, a quantity we label .

If this holds true, then opening a single monomer should present a value similar to its ring-strain. In turn, as oligomers start forming, the measured should decrease until it levels off becoming equal to the determined enthalpy of polymerization and this occurs when all oligomers are the length of a polymer, which is considered to be around 20 repeating units.¹² To test this, ring-opening reactions of δVL, δCL, δOL, and δDL catalysed by 1,5,7-triazabicyclo[4.4.0]dec-5-ene (TBD) with three equivalents of Et₃N co-catalyst in toluene containing 1, 0.2, 0.1, 0.07, 0.05, or 0.02 equivalents (respective theoretical DP = 1, 5, 10, 15, 20, and 50) of methanol as initiator were left until equilibrium was achieved and equilibrium concentrations were measured. To account for the competing polymerization and alcoholysis reactions, the following equation was derived to measure ΔH_p (eqn (2), see ESI† for full derivation of the equation).

where R is the gas constant (8.314 J mol⁻¹ K⁻¹), M is lactone, MeOH is methanol, and P are the polymer/oligomer/monomer chains. Under these conditions the polymerization reaction conditions do not qualify as standard state due to the competitive reaction with methanol, which is why the thermodynamic terms are noted without the degree symbol. All equilibrium concentrations of M, P, and MeOH can be measured directly using ¹H NMR spectroscopy (Fig. 3) at different reaction temperatures to obtain a van’t Hoff plot where the slope equals ΔH_p and the intercept equals ΔS_p (Fig. S1–S4†).

Fig. 3 ¹H NMR spectra (CDCl₃, 400 MHz, 298 K) of the equilibrium states at different temperatures obtained during the polymerization of δVL with an equimolar amount of methanol at 2 M in toluene. Samples were quenched with acetic acid prior to analysis.

In all cases, the higher the ratio of monomer to methanol, the longer the polymer chains are, and the higher (less negative) the ΔH_p and ΔS_p (Fig. 4a and b respectively). This is a strong indication that opening a monomer releases more energy than incorporating a monomer into a polymer. The measured ΔH_p at equimolar methanol concentration for the substituted lactones is so low that it starts to approximate the calculated ring-strains and demonstrates that opening a single lactone releases all of the ring-strain energy (Fig. 4a). For δVL, even at an equimolar amount of methanol, polymers form and the measured ΔH_p is therefore substantially higher than its ring-strain (Fig. S5†).²⁶ It is possible that because δCL, δOL, and δDL open to a secondary alcohol, as opposed to the primary alcohol generated when δVL is opened, their nucleophilicity is less competitive than methanol and methanolysis occurs more easily than polymerization. Without polymer, there are no energy absorbing polymer conformations. However, as the energy released by means of ΔH_p decreases, so does the ΔS_p in a linear fashion for all of the monomers, which strengthens the notion that polymers use part of the heat to increase conformational entropy (Fig. 4c).

Fig. 4 Obtained (a) ΔH_p and (b) ΔS_p values for the polymerization of δVL, δCL, δOL, and δDL in the presence of varying methanol contents in toluene catalysed by TBD–Et₃N. (c) Plot of the obtained ΔS_pversus ΔH_p values.

To study polymer conformation, PδVL, PδCL, PδOL, and PδDL were synthesized to comparable molecular weights (∼20 kg mol⁻¹) (Fig. S6–S9†) and changes in conformational structure were analysed through ¹³C NMR spectroscopy at different temperatures in toluene (Fig. S10–S13†). The terminal carbon of the repeating unit was chosen as a proxy to represent conformational changes, as this carbon should experience the greatest chemical shift when going from a trans to a gauche state. With increasing temperature, all polymers display a downfield shift towards higher ppm values, where compared to the signal at 25 °C, PδDL and PδOL show the greatest shift and PδVL the smallest shift (a total shift of 0.116, 0.159, 0.273, and 0.270 ppm for PδVL, PδCL, PδOL, and PδDL respectively) (Fig. 5a). The magnitude of the difference in the shift between the model polymers is significant and not solely the effect of temperature on the spin relaxation,²⁹ as can be seen from the little difference in the magnitude of the shift with temperature for the methyl esters of δVL, δCL, δOL, and δDL (a total shift of 0.235, 0.201, 0.278, and 0.258 respectively) (Fig. S14–S18†). Due to their much smaller structure, the methyl esters of the lactones are less conformationally obstructed than their polymer variants and their low conformational energy thus serves as reference with regards to the conformational energy. These ¹³C NMR spectroscopy results are interpreted as changes in the polymer conformation dictated by entropy, enabled by the energy provided by increasing the temperature. The larger changes for PδOL and PδDL are a result of the higher energy difference between the trans and gauche states, meaning that at 25 °C only a few gauche states have yet been obtained. Differential scanning calorimetry (DSC) strengthens the ¹³C NMR results by showing that with increasing substitution length, more energy is absorbed per mol of repeating unit in the polymer as temperature increases (Fig. 5b), a trend that is irrespective of the heating rate (Fig. S19–S22†). The slope of the DSC curves represents the amount of additional energy that is absorbed by the polymer as the temperature increases, which can be seen as the energy penalty to satisfy entropically driven conformational changes. It should be noted that in polymers such changes are often not attributable to simple vibrational changes, but indeed linked to conformational motion.³⁰ This is further exemplified by the lower inclination of the more similar slopes obtained for the DSC traces of the methyl esters of δVL, δCL, δOL, and δDL (Fig. S23†). A steeper slope therefore represents more energy absorbed and greater polymeric conformational changes due to trans states being converted to gauche states. In terms of heat capacity, the difference between 0–120 °C was found to be 2.69–5.94, 4.44–7.38, 5.69–10.13, and 6.94–12.38 J mol⁻¹ K⁻¹ for PδVL, PδCL, PδOL, and PδDL respectively at a heating rate of 10 K min⁻¹. Calculations of the average potential energy of a polymer chain consisting of 20 repeating units at different temperatures using the MM2 forcefield, which has been reported to be among the best for accurate energy calculations of conformers,^31,32 further strengthen the ¹³C NMR and DSC observations by showing that with increasing temperature the polymers with longer substituents absorb more energy (determined from slope in Fig. 5c). The presented results all indicate that PδVL, and PδCL to a lesser extent, are entropically more satisfied and have access to a comparatively large number of conformational states even at low temperatures and therefore do not experience great conformational changes as temperature increases. PδOL and PδDL, on the other hand, show much greater conformational changes; a result that indicates that these polymers are entropically unsatisfied, but lack the energy at low temperatures to increase the number of possible conformations.

Fig. 5 Analysis of PδVL, PδCL, PδOL, and PδDL, evaluating (a) changes in chemical shift for the terminal methylene of the repeating unit with temperature derived from ¹³C NMR spectroscopy performed in toluene. (b) DSC (10 K min⁻¹, N₂) thermograms in bulk. (c) MD derived average potential energy over 10 ns simulation time using the MM2 forcefield in vacuum in the 60–360 K temperature range.

To further study polymer conformation on an atomistic scale, we decided to study the conformation of oligomers with 5 repeating units of PδVL, PδCL, PδOL, and PδDL in toluene using MD. The statistical desire to obtain equal amounts of trans and gauche states, entropy, drives extended polymers to form coil-like conformations. However, the coil structure contains more gauche states, compared to the mostly trans states in an extended polymer, and hence increases the energy of the polymer (Fig. 6a). The amount of energy in the conformation will be heavily reliant on what is allowed by the monomer structure, where the unsubstituted PδVL will coil up more easily than the sterically encumbered PδDL. To test this and observe how polymer conformations change with temperature, the distance between the alcohol chain end and the second to last repeating unit in the polymer chain were extracted from the trajectory files of the MD simulation to serve as proxy for overall polymer conformation (Fig. 6b and S24–S27†). In agreement with ¹³C NMR results, PδVL and PδCL showed little change in distance as temperature increased (Fig. 6c and S28†), whilst PδOL and PδDL presented more considerable changes in their distance distributions and increases in the alcohol to ester distance with an increase in temperature (Fig. 6d, 6e and S29†). This confirms that NMR indeed can be used as a proxy for conformational changes. Thereby, the DSC results and the observed MD changes, with regards to conformation, support that is temperature dependent and that is higher for δVL and δCL than for δOL and δDL as the prior pair show fewer conformational changes with rising temperature.

Fig. 6 Schematic representation of (a) the origins of and (b) the intramolecular distance used as proxy for polymer conformation. Conformational probability density distribution for r at different temperatures for (c) PδVL and (d) PδDL. (e) Average distance r (〈r〉) at different temperatures for polymers dissolved in toluene.

To better understand the atomistic origins of and why changing conformation affects polymerizability, we employed a recently developed molecular dynamics (MD) method used to calculate the thermodynamic polymerization parameters from collision frequencies.²³ By counting the intra- and intermolecular collisions, with a collision cut-off distance of 0.3 nm between the polymer chain-end either with itself or monomer units, and values for δVL, δCL, δOL, and δDL were obtained (Fig. 7a and b). In all cases, the obtained and are slightly lower and higher, respectively, than experimental values, however, the trend in , where δVL is the lowest and δDL the highest, is in good agreement with experimental results.

Fig. 7 (a) Schematic representation of the employed MD method for the calculation of thermodynamic parameters. (b) Modified van’t Hoff plot used to calculate the thermodynamic parameters from MD results. (c) Intramolecular and (d) intermolecular collisions counted during the 50 ns simulation in the 295–375 K temperature range in toluene.

An interesting part of calculating the thermodynamic parameters through MD, is the insight into all atomistic process that proceed, here studied by comparing the intra- and intermolecular collision frequencies recorded during the experiment. For δVL and δCL when the temperature is increased from 295 K to 375 K, they gain most of their from a sharp increase in intramolecular collisions (Fig. 7c), doubling in collision rate, with only a small decrease in intermolecular collisions (losing ∼5%) (Fig. 7d). On the other hand, δOL and δDL show a drastic drop ∼70% in intermolecular collisions with increase in temperature from 295 K to 375 K (Fig. 7d), still with a significant increase in intramolecular collisions (∼70% more at 295 K than 375 K). In summary, the equilibrium monomer concentrations of δVL and δCL are higher at higher temperatures due to accelerated intramolecular reactions, whilst the equilibrium monomer concentrations of δOL and δDL are higher at higher temperatures due to the simultaneous acceleration of the intramolecular reaction and retardation of the monomer consuming intermolecular reaction. Hence, despite having a similar structure, the origin of the polymerizability of δVL and δCL is vastly different than that of δOL and δDL.

To attempt a complete calculation of using the hypothesis described in eqn (1), MD derived average potential energies of polymers consisting of 19 repeating units or 20 repeating units, the lactones, the methyl ester of each lactone, and methanol were used to estimate , , and . It must be stressed that the following simulations and calculations are not optimized for this system and the resulting values should not be taken as literal values, but as an indication of the trend. Accurate parameterization is of high importance to obtain discreet values.^31,33 The was calculated at different temperatures by taking the difference in energy between a polymer of 20 and one of 19 repeating units (Fig. S30–S33†), and subtracting from this the potential energy of the respective methyl ester minus methanol (Fig. S34†). These equations should provide the energy of a single repeating unit in the polymer form and a linear unphysical reference state containing no end groups, respectively, meaning that the difference is due to conformation. In good agreement with experimental results, the calculated shows a greater temperature dependence for δOL and δDL compared to δVL and δCL (Fig. 8a). As a result of the linear temperature dependence, at low temperatures the calculated is higher for δVL and δCL than for δOL and δDL, but at higher temperatures (>1000 K), albeit irrelevant from a practical point of view, δOL and δDL surpass the of δVL and δCL. With the calculation of as the final goal, the at different temperatures was also calculated, which was performed by subtracting from the average potential energy of the lactone (Fig. S35†) that of the methyl ester minus methanol (Fig. S34†). These equations should provide the amount of energy that is stored in the ring versus the linear structure with the same number of atoms. From this, the could be estimated and the temperature dependence was found to be similar for all lactones at each given temperature (Fig. 8b). As a result of the equivalent ring-strain values at all temperature for all lactones, it is the that determines the , which was calculated by the addition of and (Fig. 8c). The trend in the estimated at room temperature is in good agreement with experimental results where δOL and δDL show lower values in the temperature range most commonly used for experiments (298–373 K) than δVL and δCL. It should be pointed out that and are known to be temperature dependent even when no phase changes occur.^24,34 However, their variation within the experimental temperature window (usually ∼50 °C) is small enough that they appear practically constant in equilibrium monomer concentration studies and deviations are more likely derived from experimental error.

Fig. 8 (a) and (b) calculated from the average potential energy derived from 10 ns MD simulations using the MM2 forcefield in vacuum in the 0–750 K temperature range. (c) calculated from the addition of the and . Calculations were done using non-optimized forcefields and the values should only be taken as an indication of trend.

To study if the hypothesis concerning provided in this work supports previous work, a small review of examples of monomers with positive , other variants of substituted of δVL, and solvent effects on thermodynamics was performed. Most monomers used in ROP are driven by a release of ring-strain, leading to an overall negative and thus increasing the reaction temperature promotes the depolymerization reaction. However, there are cases where is positive, for example macrolactones such as pentadecanolide (PDL),³⁵ the recently published ketal-ester,³⁶ and ethylene carbonate.³⁷ Such examples simply defy the analogy of ring-strain and . What is more intriguing, despite being widely cited as a strainless lactone within the ROP scientific community, PDL actually contains non-neglectable ring-strain experimentally measured to be ∼11 kJ mol⁻¹,³⁸ which is supported by the considerable ring-strain for other unsubstituted macrolactones of similar size,^28,38,39 and cyclohexadecane (also 16 atoms in ring).⁴⁰ Additionally, albeit having a positive , ethylene carbonate (entry 1, Table S1†) also has ring-strain. Following the hypothesis of this work, such systems can be described as having as the dominant term. Hence, information about demystifies the ROP process and explains thermodynamics by including the energy state of the polymer as an additional driving force of polymerization. Examples of this can already be found within literature, for example, di-substituting the α-position of δVL has shown to change and remarkably.¹⁰ The diethyl substituted variant of δVL, δVLEt₂, displays and , more than doubling both the enthalpy and entropy values of δVL despite having very similar ring-strain (entries 2–5, Table S1†). Changes in ring-strain for substituted δVL are small,²¹ which is supported by the lower than cyclohexane ring-strain energy for 1,1-dimethylcyclohexane.⁴¹ Therefore, changes in cannot be attributed to ring-strain for substituted δVL. Instead, going from the extended state of PδVLEt₂ to its coil state is very energy expensive, a cost which is simply too high to be supplemented by available thermal energy at the temperatures at which polymerization is performed, and as a result this polymer will absorb little energy released by the ring-strain, yielding low , and by extension low and low values as the polymer can only access a few conformational states and is entropically unsatisfied. Another clear example that demonstrates how one may change polymerizability through , are through the effect solvent has on the thermodynamics of polymerization in ROP.^42–45 The polymerization of L-lactide, for instance, in DMSO provides and , whilst in chlorobenzene as solvent and , a result which has been attributed to the difference in energy of solvation of the polymer and monomer.⁴² The differences in between different solvents are thought to be mostly the result of changes in as the internal non-equilibrium bond lengths and angles that make up are not altered in different solvents. Altogether, the hypothesis of this work provides both conceptual and mathematical support for observations made in literature, making it a powerful concept to understand the factors that govern the thermodynamics of ROP.

Conclusions

Our study explores and attempts to demystify why enthalpy of polymerization does not equal its ring-strain . We hypothesise that part of the energy released by the is absorbed to attain polymeric conformations that are entropically favourable, but enthalpically expensive. This hypothesis was tested through the combined use of NMR spectroscopy, DSC and MD simulations. ¹H NMR spectroscopy equilibrium studies demonstrated that ring-opening with methanol releases more energy than adding a monomer unit to an existing polymer chain, contributing to the concept of energy absorption by the polymer. MD simulations revealed that δVL and δCL perform faster depolymerization reactions with no change to the intermolecular reaction rate as temperature rises, whilst δOL and δDL demonstrated a steep decrease in intermolecular reaction rate as temperature rises as a result of significant polymer conformational changes for PδOL and PδDL compared to PδVL. This was further supported by ¹³C NMR, DSC, and MD potential energy calculations, which showed that polymers with larger substituents, PδOL and PδDL, absorb more energy enabling alteration of their polymer conformation as temperature rises than unsubstituted PδVL. More importantly, the hypothesis fits observations in literature that are unexplainable when considering as purely ring-strain driven and thus provides new conceptual insights to polymer chemists focusing on the thermodynamics behind ROP.

Data availability

All raw unprocessed data and NMR spectra are available on KTH Zenodo (https://doi.org/10.71775/kth.9qy5e-nea08). The supplementary information file contains experimental methods and additional data for presented experiments.^{10,23,26–28,37,46–59}

Conflicts of interest

The authors have no conflict of interest to declare.

Acknowledgements

The authors acknowledge funding from the Swedish Research Council (Grant Number 2020-03455).

References

1. G. W. Coates and Y. D. Y. L. Getzler, Chemical Recycling to Monomer for an Ideal, Circular Polymer Economy, Nat. Rev. Mater., 2020, 5(7), 501–516,  DOI:10.1038/s41578-020-0190-4.
2. V. Nieboer, K. Odelius and P. Olsén, Improving Circularity via Chemical Recycling to All Rings, Angew. Chem., Int. Ed., 2025, 64(19), e202502436,  DOI:10.1002/anie.202502436.
3. X.-L. Li, R. W. Clarke, H.-Y. An, R. R. Gowda, J.-Y. Jiang, T.-Q. Xu and E. Y.-X. Chen, Dual Recycling of Depolymerization Catalyst and Biodegradable Polyester That Markedly Outperforms Polyolefins, Angew. Chem., Int. Ed., 2023, 62(26), e202303791,  DOI:10.1002/anie.202303791.
4. P. Hodge, Recycling of Condensation Polymers via Ring–Chain Equilibria, Polym. Adv. Technol., 2015, 26(7), 797–803,  DOI:10.1002/pat.3469.
5. P. Olsén, K. Odelius and A.-C. Albertsson, Thermodynamic Presynthetic Considerations for Ring-Opening Polymerization, Biomacromolecules, 2016, 17(3), 699–709,  DOI:10.1021/acs.biomac.5b01698.
6. C. Shi, L. T. Reilly, V. S. Phani Kumar, M. W. Coile, S. R. Nicholson, L. J. Broadbelt, G. T. Beckham and E. Y.-X. Chen, Design Principles for Intrinsically Circular Polymers with Tunable Properties, Chem, 2021, 7(11), 2896–2912,  DOI:10.1016/j.chempr.2021.10.004.
7. D. K. Schneiderman and M. A. Hillmyer, Aliphatic Polyester Block Polymer Design, Macromolecules, 2016, 49(7), 2419–2428,  DOI:10.1021/acs.macromol.6b00211.
8. Y.-M. Tu, F.-L. Gong, Y.-C. Wu, Z. Cai and J.-B. Zhu, Insights into Substitution Strategy towards Thermodynamic and Property Regulation of Chemically Recyclable Polymers, Nat. Commun., 2023, 14(1), 3198,  DOI:10.1038/s41467-023-38916-5.
9. P. McMichael, X. Schultze, H. Cramail and F. Peruch, Sourcing, Thermodynamics, and Ring-Opening (Co)Polymerization of Substituted δ-Lactones: A Review, Polym. Chem., 2023, 14(33), 3783–3812,  10.1039/D3PY00657C.
10. X.-L. Li, R. W. Clarke, J.-Y. Jiang, T.-Q. Xu and E. Y.-X. Chen, A Circular Polyester Platform Based on Simple Gem-Disubstituted Valerolactones, Nat. Chem., 2023, 15(2), 278–285,  DOI:10.1038/s41557-022-01077-x.
11. A. D. Jenkins, P. Kratochvíl, R. F. T. Stepto and U. W. Suter, Glossary of basic terms in polymer science (IUPAC Recommendations 1996), Pure Appl. Chem., 1996, 68(12), 2287–2311,  DOI:10.1351/pac199668122287.
12. A. Duda and A. Kowalski, Thermodynamics and Kinetics of Ring-Opening Polymerization, in Handbook of Ring-Opening Polymerization, John Wiley & Sons, Ltd, 2009, ch. 1, pp. 1–51,  DOI:10.1002/9783527628407.
13. Thermodynamics of Cationic Ring-Opening Polymerization, in Cationic Ring-Opening Polymerization, ed. S. Penczek, P. Kubisa and K. Matyjaszewski, Springer, Berlin, Heidelberg, 1985, pp. 2–34,  DOI:10.1007/3-540-13781-5_2.
14. J. E. Mcgrath, Ring-Opening Polymerization: Introduction, in Ring-Opening Polymerization, ACS Symposium Series, American Chemical Society, 1985, vol. 286, pp. 1–22,  DOI:10.1021/bk-1985-0286.ch001.
15. O. Nuyken and S. D. Pask, Ring-Opening Polymerization—An Introductory Review, Polymers, 2013, 5(2), 361–403,  DOI:10.3390/polym5020361.
16. T. Dudev and C. Lim, Ring Strain Energies from Ab Initio Calculations, J. Am. Chem. Soc., 1998, 120(18), 4450–4458,  DOI:10.1021/ja973895x.
17. W. Saiyasombat, R. Molloy, T. M. Nicholson, A. F. Johnson, I. M. Ward and S. Poshyachinda, Ring Strain and Polymerizability of Cyclic Esters, Polymer, 1998, 39(23), 5581–5585,  DOI:10.1016/S0032-3861(97)10370-6.
18. H. Tran, A. Toland, K. Stellmach, M. K. Paul, W. Gutekunst and R. Ramprasad, Toward Recyclable Polymers: Ring-Opening Polymerization Enthalpy from First-Principles, J. Phys. Chem. Lett., 2022, 13(21), 4778–4785,  DOI:10.1021/acs.jpclett.2c00995.
19. Z.-R. Chen, J. P. Claverie, R. H. Grubbs and J. A. Kornfield, Modeling Ring-Chain Equilibria in Ring-Opening Polymerization of Cycloolefins, Macromolecules, 1995, 28(7), 2147–2154,  DOI:10.1021/ma00111a007.
20. D. Sathe, J. Zhou, H. Chen, H.-W. Su, W. Xie, T.-G. Hsu, B. R. Schrage, T. Smith, C. J. Ziegler and J. Wang, Olefin Metathesis-Based Chemically Recyclable Polymers Enabled by Fused-Ring Monomers, Nat. Chem., 2021, 13(8), 743–750,  DOI:10.1038/s41557-021-00748-5.
21. F. Ronchetti and L. Toma, Cis and Trans β,γ-Disubstituted δ-Valerolactones: Conformational Studies and Configurational Assignments by Molecular Mechanics, Tetrahedron, 1986, 42(23), 6535–6540,  DOI:10.1016/S0040-4020(01)88115-1.
22. M. Save, M. Schappacher and A. Soum, Controlled Ring-Opening Polymerization of Lactones and Lactides Initiated by Lanthanum Isopropoxide, 1. General Aspects and Kinetics, Macromol. Chem. Phys., 2002, 203(5–6), 889–899,  DOI:10.1002/1521-3935(20020401)203:5/6<889::AID-MACP889>3.0.CO;2-O.
23. V. Nieboer, P. Olsén, K. Odelius and J. Wohlert, Predicting Chemical Recyclability Thermodynamics via Molecular Simulations, Macromolecules, 2024, 57, 9546–9554,  DOI:10.1021/acs.macromol.4c01469.
24. A. A. Yevstropov, B. V. Lebedev, T. G. Kulagina and N. K. Lebedev, The Calorimetric Study in the 13·8–340°K Range of δ-Valerolactone, Its Polymer, and of the δ-Valerolactone Polymerization, Polym. Sci. U.S.S.R., 1982, 24(3), 628–636,  DOI:10.1016/0032-3950(82)90053-3.
25. M. T. Martello, A. Burns and M. Hillmyer, Bulk Ring-Opening Transesterification Polymerization of the Renewable δ-Decalactone Using an Organocatalyst, ACS Macro Lett., 2012, 1(1), 131–135,  DOI:10.1021/mz200006s.
26. M. L. P. Leitão, G. Pilcher, Y. Meng-Yan, J. M. Brown and A. D. Conn, Enthalpies of Combustion of γ-Butyrolactone, γ-Valerolactone, and δ-Valerolactone, J. Chem. Thermodyn., 1990, 22(9), 885–891,  DOI:10.1016/0021-9614(90)90176-Q.
27. N. L. Allinger and S. H. M. Chang, Conformational Analysis—CXXIII: Carboxylic Acids and Esters in Force Field Calculations, Tetrahedron, 1977, 33(13), 1561–1567,  DOI:10.1016/0040-4020(77)80161-0.
28. K. B. Wiberg and R. F. Waldron, Lactones. 2. Enthalpies of Hydrolysis, Reduction, and Formation of the C4-C13 Monocyclic Lactones. Strain Energies and Conformations, J. Am. Chem. Soc., 1991, 113(20), 7697–7705,  DOI:10.1021/ja00020a036.
29. H. J. Schneider and W. Freitag, The Temperature Dependence of Carbon-13 Nuclear Magnetic Resonance Shifts. Limiting Factors and Stereochemical Applications, J. Am. Chem. Soc., 1976, 98(2), 478–481,  DOI:10.1021/ja00418a025.
30. B. Wunderlich, The Heat Capacity of Polymers, Thermochim. Acta, 1997, 300(1), 43–65,  DOI:10.1016/S0040-6031(96)03126-7.
31. T. Lewis-Atwell, P. A. Townsend and M. N. Grayson, Comparisons of Different Force Fields in Conformational Analysis and Searching of Organic Molecules: A Review, Tetrahedron, 2021, 79, 131865,  DOI:10.1016/j.tet.2020.131865.
32. K. Gundertofte, T. Liljefors, P. Norrby and I. A. Pettersson, Comparison of Conformational Energies Calculated by Several Molecular Mechanics Methods, J. Comput. Chem., 1996, 17(4), 429–449,  DOI:10.1002/(SICI)1096-987X(199603)17:4<429::AID-JCC5>3.0.CO;2-W.
33. E. M. Engler, J. D. Andose and P. V. R. Schleyer, Critical Evaluation of Molecular Mechanics, J. Am. Chem. Soc., 1973, 95(24), 8005–8025,  DOI:10.1021/ja00805a012.
34. B. V. Lebedev, Thermodynamics of Polylactones, Russ. Chem. Rev., 1996, 65(12), 1063,  DOI:10.1070/RC1996v065n12ABEH000265.
35. A. A. Yevstropov, B. V. Lebedev and Y. G. Kipatisova, Thermodynamics of Pentadecanolactone, of the Process of Its Polymerization and of the Resultant Polypentadecanolactone in the 0–400 K Region, Polym. Sci. U.S.S.R., 1983, 25(8), 1951–1960,  DOI:10.1016/0032-3950(83)90317-9.
36. X.-B. Meng, T. Zhou, C. Yang, X.-Y. Cheng, X.-T. Wu, C. Shi, F.-S. Du and Z.-C. Li, Thermally Stable and Chemically Recyclable Poly(Ketal-Ester)s Regulated by Floor Temperature, J. Am. Chem. Soc., 2024, 146, 15428–15437,  DOI:10.1021/jacs.4c03523.
37. L. Vogdanis, B. Martens, H. Uchtmann, F. Hensel and W. Heitz, Synthetic and Thermodynamic Investigations in the Polymerization of Ethylene Carbonate, Makromol. Chem., 1990, 191(3), 465–472,  DOI:10.1002/macp.1990.021910301.
38. C. Galli, G. Illuminati, L. Mandolini and P. Tamborra, Ring-Closure Reactions. 7. Kinetics and Activation Parameters of Lactone Formation in the Range of 3- to 23-Membered Rings, J. Am. Chem. Soc., 1977, 99(8), 2591–2597,  DOI:10.1021/ja00450a031.
39. C. Galli and L. Mandolini, The Role of Ring Strain on the Ease of Ring Closure of Bifunctional Chain Molecules, Eur. J. Org. Chem., 2000, 2000(18), 3117–3125,  DOI:10.1002/1099-0690(200009)2000:18<3117::AID-EJOC3117>3.0.CO;2-5.
40. E. V. Anslyn and D. A. Dougherty, Modern Physical Organic Chemistry, Chapter 2: Strain and Stability, 2006.
41. A. M. De Lio, B. L. Durfey and T. M. Gilbert, Estimating Ring Strain Energies of Highly Substituted Cyclohexanes with the Semi-Homodesmotic Approach: Why Substantial Ring Strain Exists for Nominally Tetrahedral Ring Carbon Atoms, J. Org. Chem., 2015, 80(20), 10234–10243,  DOI:10.1021/acs.joc.5b01861.
42. L. Cederholm, J. Wohlert, P. Olsén, M. Hakkarainen and K. Odelius, “Like Recycles Like”: Selective Ring-Closing Depolymerization of Poly(L-Lactic Acid) to L-Lactide, Angew. Chem., Int. Ed., 2022, 61(33), e202204531,  DOI:10.1002/anie.202204531.
43. P. Olsén, J. Undin, K. Odelius, H. Keul and A.-C. Albertsson, Switching from Controlled Ring-Opening Polymerization (cROP) to Controlled Ring-Closing Depolymerization (cRCDP) by Adjusting the Reaction Parameters That Determine the Ceiling Temperature, Biomacromolecules, 2016, 17(12), 3995–4002,  DOI:10.1021/acs.biomac.6b01375.
44. K. Matyjaszewski, S. Słomkowski and S. Penczek, Kinetics and Mechanism of the Cationic Polymerization of Tetrahydrofuran in Solution: THF-CH₃NO₂ System, J. Polym. Sci., Polym. Chem. Ed., 1979, 17(1), 69–80,  DOI:10.1002/pol.1979.170170108.
45. L. Cederholm, P. Olsén, M. Hakkarainen and K. Odelius, Chemical Recycling to Monomer: Thermodynamic and Kinetic Control of the Ring-Closing Depolymerization of Aliphatic Polyesters and Polycarbonates, Polym. Chem., 2023, 14(28), 3270–3276,  10.1039/D3PY00535F.
46. H. J. C. Berendsen, D. van der Spoel and R. van Drunen, GROMACS: A Message-Passing Parallel Molecular Dynamics Implementation, Comput. Phys. Commun., 1995, 91(1), 43–56,  DOI:10.1016/0010-4655(95)00042-E.
47. M. Parrinello and A. Rahman, Polymorphic Transitions in Single Crystals: A New Molecular Dynamics Method, J. Appl. Phys., 1981, 52(12), 7182–7190,  DOI:10.1063/1.328693.
48. S. Nosé, A Unified Formulation of the Constant Temperature Molecular Dynamics Methods, J. Chem. Phys., 1984, 81(1), 511–519,  DOI:10.1063/1.447334.
49. S. Nosé and M. L. Klein, Constant Pressure Molecular Dynamics for Molecular Systems, Mol. Phys., 1983, 50(5), 1055–1076,  DOI:10.1080/00268978300102851.
50. B. Hess, P-LINCS: A Parallel Linear Constraint Solver for Molecular Simulation, J. Chem. Theory Comput., 2008, 4(1), 116–122,  DOI:10.1021/ct700200b.
51. M. J. Abraham, T. Murtola, R. Schulz, S. Páll, J. C. Smith, B. Hess and E. Lindahl, GROMACS: High Performance Molecular Simulations through Multi-Level Parallelism from Laptops to Supercomputers, SoftwareX, 2015, 1–2, 19–25,  DOI:10.1016/j.softx.2015.06.001.
52. S. Jo, T. Kim, V. G. Iyer and W. Im, CHARMM-GUI: A Web-Based Graphical User Interface for CHARMM, J. Comput. Chem., 2008, 29(11), 1859–1865,  DOI:10.1002/jcc.20945.
53. S. Kim, J. Lee, S. Jo, C. L. Brooks, H. S. Lee and W. Im, CHARMM-GUI Ligand Reader and Modeler for CHARMM Force Field Generation of Small Molecules, J. Comput. Chem., 2017, 38(21), 1879–1886,  DOI:10.1002/jcc.24829.
54. J. Lee, X. Cheng, J. M. Swails, M. S. Yeom, P. K. Eastman, J. A. Lemkul, S. Wei, J. Buckner, J. C. Jeong, Y. Qi, S. Jo, V. S. Pande, D. A. Case, C. L. I. Brooks, A. D. MacKerell Jr, J. B. Klauda and W. Im, CHARMM-GUI Input Generator for NAMD, GROMACS, AMBER, OpenMM, and CHARMM/OpenMM Simulations Using the CHARMM36 Additive Force Field, J. Chem. Theory Comput., 2016, 12(1), 405–413,  DOI:10.1021/acs.jctc.5b00935.
55. J. Lee, M. Hitzenberger, M. Rieger, N. R. Kern, M. Zacharias and W. Im, CHARMM-GUI Supports the Amber Force Fields, J. Chem. Phys., 2020, 153(3), 035103,  DOI:10.1063/5.0012280.
56. CHEM3D 19.1.0.8, Cambridge Scientific Computing.
57. U. Burkert and N. L. Allinger, Molecular Mechanics, American Chemical Society, 1982.
58. M. Chen, R. Xiao, L. Zhao, J. Weng, D. Ouyang, Q. Chen, J. Yao and J. Wang, Experimental Study on the Combustion Characteristics of Carbonate Solvents under Different Thermal Radiation by Cone Calorimeter, Appl. Therm. Eng., 2022, 211, 118428,  DOI:10.1016/j.applthermaleng.2022.118428.
59. A. F. Reinertsen, K. G. Primdahl, A. E. Shay, C. N. Serhan, T. V. Hansen and M. Aursnes, Stereoselective Synthesis and Structural Confirmation of the Specialized Pro-Resolving Mediator Resolvin E4, J. Org. Chem., 2021, 86(4), 3535–3545,  DOI:10.1021/acs.joc.0c02913.

---

Footnote

† Electronic supplementary information (ESI) available: Experimental methods and additional data for presented experiments. See DOI: https://doi.org/10.1039/d5fd00060b

---