Lignin-Derived Methoxyterephthalates for Performance-Advantaged Polymers and Plasticizers

Abstract

Lignin-derived aromatic carboxylic acids can be produced from oxidative catalytic processes and are promising building blocks for performance-advantaged bioproducts that leverage their inherent heteroatom functionalities. Here, we synthesize 2-methoxyterephthalate and 2,6-dimethoxyterephthalate derivatives by electrochemical carboxylation of guaiacyl- and syringyl-derived lignin monomers obtained from the oxidative deconstruction of lignin. These methoxylated terephthalates are evaluated as co-monomers in poly(ethylene terephthalate) (PET) and as plasticizers that could replace petrochemically-derived isophthalate and phthalate, respectively. Specifically, we co-polymerize 2-methoxy- and 2,6-dimethoxyterephthalate with dimethyl terephthalate to form several PET co-polymers, both of which enable the properties of PET to be tuned, with an incorporation beyond 25% producing amorphous polyesters. At 10 mol% loading in the co-polymers, we demonstrate that the bio-derived co-monomers exhibit comparable behavior to isophthalic acid, a commonly used co-monomer in PET, by lowering the crystallinity and melting temperature. Moreover, methoxyterephthalate esters (2-ethyl hexyl and butyl) are compared to phthalate and terephthalate ester counterparts used as poly(vinyl chloride) (PVC) plasticizers. The bio-derived plasticizers are comparable to the petroleum-derived incumbents in reducing the glass transition temperature and increasing the thermal stability of PVC. Furthermore, the dimethoxyterephthalic esters are expected to have an extended lifetime in the polymer matrix, due to their lower volatility and by lower diffusion coefficients calculated by molecular dynamic simulations. These results demonstrate that the isophthalate and phthalate components in polyesters and plasticizers, respectively, could be substituted with bio-based methoxyterephthalate derivatives.

Graphical Abstract

Introduction

Lignin is an abundant source of bio-derived aromatic compounds with the potential to substitute for fossil carbon-derived aromatic feedstocks.¹ There are multiple catalytic strategies to depolymerize lignin into aromatic monomers,^2–9 several of which offer the ability to produce aromatic compounds with oxygenated functionalities that are distinct from fossil carbon-derived aromatics. These heteroatom functionalities can be leveraged for further functionalization or contribute directly to bioproduct performance.¹⁰ Of relevance to this study, the most common aromatic monomers from oxidative lignin depolymerization of hardwood feedstocks like poplar are 4-hydroxybenzoic acid (H), vanillic acid (G), and syringic acid (S) (Scheme 1).^11–14 These lignin-derived products have been explored for a variety of polymers and small molecule products.^15–16 Vanillic acid and its derivatives from lignin¹⁷ have been extensively studied to make bio-based products,^18–19 ranging from mimics of poly(ethylene terephthalate) (PET),^20–21 plasticizers,²² and epoxy resins.²³ Conversely, studies with syringic acid and its derivatives are less common, with reports demonstrating their use in polyacrylate esters²⁴ or polystyrenes.²⁵ Given that syringic units are abundant in hardwood lignin, finding uses for these monomers is desirable.

Scheme 1:

From lignin to aromatic dicarboxylic acids via phenol activation and electrocatalysis.

It is well established that alcohols and their activated derivatives can be converted to carboxylic acids through catalytic coupling with CO₂.²⁶ Several reports have produced terephthalic acid (TPA) from lignin, as a direct replacement for the aromatic co-monomer in the commodity plastic, PET,^27–28 which requires removal of the methoxy substituents on the aromatic ring. For performance-advantaged bioproducts, retention of the methoxy groups on the aromatic ring could provide tunability for both small molecule and polymer products. To this end, Skrydstrup and co-workers demonstrated a synthetic route to 2-methoxyterephthalic acid (^2-MeOTPA) through reductive catalytic fractionation, via a fluorosulfation–catalytic cyanation sequence followed by hydrolysis and oxidation.²⁹ The ^2-MeOTPA was co-polymerized with ethylene glycol (EG) to yield a PET-analogue, although the reported characterization of the material was limited. Another example is a report from Song and co-workers, who used hydroxycinnamic derivatives from the reductive catalytic fractionation of lignin to make dicarboxylic acid species.¹⁶ These monomers were used to make a range of aliphatic-aromatic co-polyesters with controllable mechanical, optical, and thermal performance.

In previous work (Scheme 1), we electrochemically cross-coupled 4-hydroxybenzoic, vanillic, and syringic acids to make biphenyl-dicarboxylic acid derivatives.³⁰ These compounds proved to be suitable, non-toxic candidates to replace phthalate-based plasticizers for poly(vinyl chloride) (PVC). In relation to this work, we envisioned that electrochemical catalysis could also be applied to carboxylation of phenolic lignin monomers to afford terephthalic acid derivatives.^31–36 An electrochemically-generated, low-valent Ni catalyst has been successful in the coupling of a variety of challenging, unactivated aryl halides and sulfonates.³⁵ This procedure can be adapted to lignin-derived phenols, such as the ones obtained by oxidative depolymerization of lignin to access lignin-derived methoxylated terephthalates, which are underexplored.¹⁹ Isophthalic acid (IPA) is a common co-monomer added to PET to reduce crystallinity, melt temperature, and opacity,^37–38 and ^2-MeOTPA and 2,6-dimethoxy terephthalate (^2,6-MeOTPA) are potentially replacements for the isomeric phthalic acids in commercial applications, including commodity and specialty polymers (i.e., PET and alkyds, respectively), plasticizers (i.e., phthalate esters), and more.

To this end, here we describe the 2-step electrochemical carboxylation of the lignin monomers, vanillic acid and syringic acid, into TPA derivatives and their application in both chemicals and materials. Specifically, we explore the effect of introducing methoxy groups in co-polymers of PET and whether it can offer performance advantages compared to IPA. We also describe conversion of the monomers into esters using industrially relevant alcohols typically used in plasticizers for PVC and compare their performance to the corresponding phthalate and terephthalate esters.

Results

Synthesis of methoxyterephthalates.

^2-MeOTPA and ^2,6-MeOTPA derivatives can be easily accessed on a large scale via esterification and methylation of commercially available 2-hydroxy- and 2,6-dihydroxyterephthalic acid (>50 g scale, Scheme 2A, see Experimental Methods section). The procedure afforded dimethyl terephthalate (DMT) ester derivatives, ^2-MeODMT and ^2,6-MeODMT, respectively, which are desirable for downstream polycondensation reactions due to the lower melting point compared to diacids. However, we sought to establish a new route to access these structures from vanillic (G) and syringic (S) acid, two abundant lignin-derived monomers. In this strategy, the methyl esters of G and S are converted into electrophiles by reaction of the phenols with a sulfonyl chloride, TsCl [Ts = 4-toluenesulfonyl], or triflic anhydride (Tf₂O) (Scheme 2B). These compounds are subjected to Ni-catalyzed carboxylation under ambient CO₂ atmosphere to install the second carboxylate unit. We started our investigation with the mono methoxylated derivative methyl 3-methoxy-4-(tosyloxy)benzoate (G–OTs). High-throughput experimentation (HTE) using chemical reductants such as Zn powder led to rapid identification of promising catalyst systems that were used as the starting point for optimization of electrochemical conditions (see Supporting Information (SI), Table S2). The resulting catalyst and reaction conditions featured NiBr₂^●dme (10 mol%)/neocuproine (20 mol%) as the pre-catalyst, 0.2 M TBABr as electrolyte, MgCl₂ (1 equiv) and ^tBuOLi as additives, and RVC (cathode)/Zn (sacrificial anode) electrodes in anhydrous DMSO. Electrolysis was performed at a constant potential of −1.8 V vs. Fc in an undivided cell under 1 atm of CO₂ until 2.2 F·mol⁻¹ of charge was passed. These conditions generated the ^2-MeOTPA methyl ester product in 67% yield. The same conditions were then used in a recirculating-flow parallel-plate electrolysis cell, equipped with carbon felt as cathode and Zn as sacrificial anode (Scheme 2C). A constant current of −2.4 mA·cm⁻² was used to conduct a gram scale reaction (3 mmol G–OTs), and the product was obtained in 72% yield (66% isolated yield). Esterification of the product with methanol and catalytic H₂SO₄ yields the dimethyl ester, ^2-MeODMT.

Scheme 2:

(A) Synthesis of methoxyterephthalates from hydroxyterephthalic acids. (B) Synthesis of methoxyterephthalates from lignin-derived 4-hydroxyacids. (C) Optimized conditions for the electrochemical Ni-catalyzed carboxylation of G and S phenols in batch and flow. Yields are based on ¹H NMR spectroscopy, unless otherwise specified, using 1,3,5-trimethoxybenzene as an external standard. Flow electrolysis was performed under recirculation with a flow rate of 30 mL min⁻¹.

Ni-catalyzed carboxylation of the S-derived monomer proved to be more challenging due to the high steric hindrance at electrophilic site. Consequently, the more reactive triflate derivative, methyl 3,5-dimethoxy-4-(trifluoromethyl)sulfonyloxy)benzoate (S–OTf), was needed to access 2,6-dimethoxyterephthalate. The reaction conditions were reoptimized for this substrate, and the best outcome was achieved with 10 mol% NiCl₂dppf as a pre-catalyst, 0.2 M TBAPF₆ as electrolyte, MgCl₂ (1 equiv) as an additive, and RVC/Zn as electrodes in anhydrous DMSO (Scheme 2C). Electrolysis was performed by applying a constant current of −1.5 mA·cm⁻² under 1 atm of CO₂ until 2.2 F·mol⁻¹ were passed. The protodeoxygenated S–H derivative was obtained as a significant side product, even under the optimized conditions. Nonetheless, gram-scale synthesis (3 mmol S–OTf) was possible using the same recirculating flow electrolysis method, affording ^2,6-MeOTPA in 44% yield (38% isolated). While this protocol would need further optimization to support large scale application, these results illustrate a strategy for use of aromatics from lignin as building block for sustainable chemical products.

Synthesis and characterization of co-polyesters.

Polymerizations were conducted through melt polycondensation, combining different molar ratios of ^2-MeODMT or ^2,6-MeODMT and DMT in excess EG with 1 mol% Ti(OⁱPr)₄ as the catalyst (Scheme 3). The first stage of the polymerization was conducted at 180 °C under N₂ to form the transesterification product of EG and methyl esters. For co-polymers with ^2-MeODMT, this reaction was conducted for 2 hours, which is typically sufficient in PET synthesis, whereas an overnight period of 16 h for ^2,6-MeODMT was used to account for its longer esterification times due to steric hinderance of the disubstituted monomer. In the second polycondensation step, the temperature was increased to 250 °C, held for 1 h under N₂, and subsequently small molecules, including MeOH, H₂O, and EG, were removed via dynamic vacuum (ca. 2 Torr). Here again, the ^2,6-MeOTPA -containing co-polymers were allowed to react for 6 h under vacuum to achieve higher molar masses compared to only 2 h for ^2-MeOTPA co-polymers. The resulting copolyesters were fully dissolved with a 1:1 mixture of trifluoroacetic acid (TFA) and chloroform and purified by precipitation from methanol to remove any remaining small molecules and oligomers. All polymers were white to off-white in color after precipitation. The polymers are referred to as ^2-MeOPET-X and ^2,6-MeOPET-X, where X denotes the mol% of the methoxylated monomer relative to the total aromatic content.

Scheme 3:

Polycondensation of varying ratios of DMT and lignin-derived methoxyterephthalates with EG, making ^2-MeOPET-X and ^2,6-MeOPET-X co-polymers.

Co-monomer incorporation and co-polymer microstructures were analyzed by NMR spectroscopy in a 9:1 mixture of CDCl₃ and TFA-d, a solvent mixture typically used for PET. Stacked plots of ^2-MeOPET-X and ^2,6-MeOPET-X in Fig. 1A–B show the peak assignments and their evolution with changing monomer ratios. The monomer feed ratios correlate well with the final co-polymer composition measured by ¹H NMR spectroscopy (Table S7), by comparing the integral of aromatic peaks 2 and 4 in both polymers. The microstructures of ^2-MeOPET-X polymers appear random, by analysis of ¹³C NMR spectroscopy, which is expected for melt polycondensations. The quantitative ¹³C NMR spectrum for ^2-MeOPET-100 shows four peaks for the EG region, with roughly 1:1:1:1 integration, arising from the three different positions of the methoxy groups (Figs. S4–S5). This suggests that the orientation of the methoxy group is random in the homopolymer. This observation is also reflected in the resonance of the methoxy protons in the ¹H NMR spectrum, peak 7 (Fig. 1A, Fig. S6), where two peaks arise (1:1.12 relative intensity) due to the two possible methoxy group spatial environments.

Figure 1:

Stacked plots of ¹H NMR spectra (9:1 CDCl₃:TFA-d, 400 MHz) of (A) ^2-MeOPET-X and (B) ^2,6-MeOPET-X, where X = mol% co-monomer in PET.

In the ^2,6-MeOPET-X NMR spectra, peak 6 at 3.9 ppm in Fig. 1B appears with more than 25 mol% of ^2,6-MeOTPA incorporation, and the integral increases (relative to the EG region between 4.6 and 4.7 ppm in the ¹H NMR spectrum) with increasing loading of ^2,6-MeOTPA. Using 2D NMR spectroscopy, HSQC and HMBC, the peak was assigned to ^2,6-MeOTPA methyl ester end-groups, arising from the low reactivity of the monomer. Quantitative ¹³C NMR spectroscopy of the ^2,6-MeOPET-100 polymer shows that the ratio between methyl ester carbonyls and in-chain methoxy-adjacent carbonyls is 2:1 (Fig. S7), suggesting that, on average, the resulting product is a trimer. This suggests a maximum monomer sequence length for ^2,6-MeOPET-X polymers under the polymerization reaction conditions used in this work. Furthermore, the ¹H NMR spectra show distinct methoxy environments between 3.7–3.9 ppm, which can be assigned to ^2,6-MeOTPA connectivity to TPA or itself (Fig. S8).

Polymer molar mass distributions, determined by gel permeation chromatography coupled with multi angle light scattering (GPC-MALS) in 1,1,1,3,3,3-hexafluoroisopropanol (HFIP), are monomodal with dispersities between 1.7–3.0, which are typical values for polycondensation polymerizations (Fig. 2A,C). For co-polymers of ^2-MeOTPA, number average molar masses (M_n) were between 8.0–19.0 kDa, with higher incorporation of monomer trending towards lower M_n; polycondensation reactions may require longer reaction times due to the slower reaction kinetics of the methoxylated monomer. This becomes more apparent for the co-polymers of ^2,6-MeOTPA. In addition to the longer reaction times, the 2,6- substitution leads to more challenging polymerizations, and accordingly the polymer molar masses decrease substantially beyond 10% monomer incorporation. Indeed, polymers with 75 and 100 mol% ^2,6-MeOTPA produced low molar mass oligomers (below 5 kDa) which could not be resolved by GPC as they elute past column resolution (Fig. 2C).

Figure 2:

(A) Overlay of GPC traces of ^2-MeOPET-X polymers using HFIP as the eluent. The y-axis represents the normalized dRI detector response. (B) Overlay of DSC traces of ^2-MeOPET-X polymers, second heating scan at 10 °C min⁻¹. (C) Overlay of GPC traces of ^2,6-MeOPET-X polymers using HFIP as the eluent. The y-axis represents the normalized dRI detector response. The shaded grey area represents the chromatograph past column resolution. (D) Overlay of DSC traces of ^2,6-MeOPET-X polymers, second heating scan at 10 °C min⁻¹.

Glass transition temperature(s) (T_g), melting temperature (T_m), and percent crystallinity (X_c) were measured by differential scanning calorimetry (DSC). For both co-monomers, polymers are semi-crystalline up to 25% co-monomer incorporation and exhibit a decrease in both X_c and T_m with increasing incorporation (Fig. 2B–C). The crystallinity of the polymers is assumed to arise from PET crystallites only, due to the amorphous nature of methoxyterephthalate homopolymers, and hence calculations for X_c are based on the ΔH_cryst of PET. ^2-MeOPET-25 displays a cold crystallization at 150 °C during the second heating run, indicating that the sample was not able to crystallize fully from the melt at 10 °C min⁻¹ (Fig. 2B). The T_g values up to 10% incorporation (72–75 °C) are comparable to synthesized PET (74 °C). Above 25%, the T_g values decrease steadily down to 45 °C. All ^2-MeOTPA-containing polymers exhibit high thermal stability, comparable to that of PET (T_d,5% = 404 °C), with a single decomposition event at T_d,5% ≈ 400 °C (Table S8, Fig. S9). Conversely, ^2,6-MeOTPA polymers show congruent T_d,5% to PET up to 25% incorporation, after which they decrease to around 370 °C (Table S8, Fig. SS9).

These results prompted us to consider using the methoxylated terephthalates as an alternative to IPA, a co-monomer commonly used in PET to modulate material properties. We thus performed thermomechanical characterization with ^2-MeOPET-10 and ^2,6-MeOPET-10, comparing these properties to a synthesized 10 mol% IPA in PET (^IPPET-10) as well as synthesized PET (Table 1). The melting temperatures were all between 15–30 °C below PET, with ^2-MeOPET-10 exhibiting the highest melting temperature at 231 °C, and percent crystallinities were all lower than PET (Table 1). The initial thermal data demonstrate that ^2-MeOTPA and ^2,6-MeOTPA have a similar influence as IPA on the microstructure of PET, i.e., disrupting crystallinity, leading to lower T_m and X_c. This feature is advantaged because it can reduce melt processing temperatures in manufacturing.

Table 1:

Summary of thermal and mechanical properties of PET and PET co-polymers.

Polymer	Mn (kDa)	Đ	Tg,Dsc (°C)a	Tm (°C)a	Xc (%)b	Td,5% (°C)c	Tg,DMA (°C)d	E’ (GPa)e
PET	21.1	3.0	74	246	35	411	96	2.9
2-MeOPET-10	13.7	2.9	70	231	25	410	92	3.1
2,6-MeOPET-10	11.4	3.0	68	229	23	402	93	2.2
IPPET-10	14.4	2.1	70	225	27	405	90	2.7

^aMeasured by DSC from the 2^nd scan at a heating rate of 10 °C min⁻¹.

^bMeasured by DSC from the 2^nd of heating by taking the melt enthalpy, subtracting the enthalpy of cold crystallization, and dividing by 140.1 J g⁻¹.

^cTemperature at 5% weight loss measured by TGA at a heating rate of 20 °C min⁻¹.

^dPeak in the tan(δ) curve obtained during from DMA at rate of 3 °C min⁻¹.

^eStorage modulus obtained from DMA at 30 °C.

Dynamic mechanical analysis (DMA) was also performed to determine the T_g and evaluate the thermomechamical properties of the copolymers (Table 1). Specimens were prepared by compression molding above the T_m to make thin film samples suitable for testing. The T_g values, measured by the tan(δ) peak, exhibit a similar trend to those observed by DSC (Fig. 3A), and the magnitude of the tan(δ) peak also correlates well with percent crystallization from DSC, where smaller tan(δ) values correspond to higher crystallinity. The degree of crystallinity is further reflected in the storage modulus (E’) in the rubbery-state above the T_g (i.e., 140 °C), with all copolymers exhibiting over 50% reduction in modulus compared to the unmodified PET (Fig. 3B). The larger reduction in E’ above the T_g observed for the copolymers is attributed to increased amorphous fractions imparting more segmental mobility and therefore reduced stiffness in the rubbery-state.⁴⁵ Below the T_g, in the glassy state, the storage modulus is highest for ^2-MeOPET-10 (3.1 GPa), making it the stiffest material out of the polymers tested (Fig. 3B). The ^2,6-MeOPET-10 exhibits the lowest glassy state E’ (2.2 GPa), and a higher amplitude of tan(δ) through the T_g, which is consistent with the degree of crystallinity and often correlated with improved toughness.⁴⁶

Figure 3:

(A) Tan(δ) as a function of temperature, measured by DMA at a rate of 3 °C min⁻¹. (B) Storage modulus (E’) as a function of temperature, measured by DMA at a rate of 3 °C min⁻¹. (C) Complex viscosity as a function of angular frequency at 260 °C measured on a rheometer.

Rheological frequency sweep experiments were conducted at 260 °C to investigate the shear rate dependencies and melt processability of the 10 mol% co-monomer polymers and of the unmodified PET. At low oscillatory frequencies (< 0.5 rad·s⁻¹), the co-monomer samples exhibited Newtonian plateaus and viscosities comparable to the unmodified PET, ranging from 290 to 440 Pa·s. When the oscillatory frequency was increased, the unmodified PET exhibits gradual reductions in complex viscosity and shear thinning consistent with commercial PET homopolymers.⁴⁷ Conversely, the co-monomer-containing polymers, ^IPPET-10 and ^2,6-MeOPET-10, exhibit more pronounced frequency dependencies with larger reductions in complex viscosity as the oscillatory frequency was increased (Fig. 3C). For example, the ^IPPET-10 and ^2,6-MeOPET-10 samples exhibit a 39% and 44% reduction in complex viscosity, respectively, as the frequency increased from 1.0 to 10 rad·s⁻¹, whereas the unmodified PET only demonstrates a 24% reduction across the same frequency range. For ^2-MeOPET-10, a similar trend in viscosity reduction was observed at moderate frequencies; however, the effect was less prominent than the other two co-monomers. The incorporation of meta-substituted isophthalic acid units in PET imparts kinked segments along the polymer backbone, which effectively increases free volume and alters chain reptation dynamics, promoting disentanglement and increased shear thinning behavior.^48–50

Similar reductions in viscosity at moderate frequencies have been reported in PET that was modified with 6 mol% of 2,5-furandicarboxylic acid (FDCA) co-monomer, which was attributed to the formation of polar tangling points that were more sensitive to changing frequency and more susceptible to dissociation compared to chain entanglements.³⁸ Similar polar aggregates have also been reported in PET/ionomeric polyester blends, which resulted in increased frequency sensitivity and improved shear thinning behavior compared to unmodified PET.⁵¹ The shear thinning behavior of the methoxyterephthalate copolymers observed at moderate frequencies suggests that the rheological properties are influenced by both altered chain dynamics through the incorporation of pendent groups that impart free volume and the formation of additional physical crosslinks through increased concentrations of polar functionalities.⁵² The modified rheological profiles and enhanced shear thinning behavior demonstrates the potential to reduce processing temperatures and improve the melt processability of PET by including methoxyterephthalate co-monomers.

Chemical recycling kinetics of co-polymers.

We were interested in investigating how the incorporation of these co-monomers in PET would influence rates for chemical recycling to monomers. Chemical recycling of PET is being pursued through multiple routes, and several studies have shown that small amounts of co-monomers can accelerate hydrolysis.⁵³ Of interest here, catalyzed glycolysis can yield bis-hydroxyethyl terephthalate monomers that can be repolymerized into virgin-quality polymer.⁵⁴ We investigated initial glycolysis rates for ^2-MeOPET-10 and ^2,6-MeOPET-10 compared to ^IPPET-10 and PET, using dimethylethylamine (DMEA, 5 wt%) as the catalyst at 170 °C, at 3, 8 and 13 min (Scheme 4, Fig. 4A). The polymers were initially cryomilled to obtain a consistent particle size. Small molecules released in the soluble fraction (Fig. S10) were identified and quantified by LC-MS and are summarized in Tables S10–S13. The solids at each time point were separately analyzed by GPC to track the evolution of M_n over the reaction (Fig. 4B).

Scheme 4:

Amine-catalyzed chemical recycling of PET co-polymers, yielding BHET and bis-hydroxyethyl derivatives of co-monomers as the primary products.

Figure 4:

Glycolysis of PET and co-polymers at 170 °C with 16:1 EG:polymer molar ratio and 5 wt% DMEA as a catalyst. Reactions were run in 12 mL Swagelok tube reactors using 8.5 mmol of polymer. Each data point is the average of duplicate experiments. (A) Plot of initial rates of glycolysis. BHET formation was quantified by HPLC. Quantification data for other small molecules can be found in Tables S10–13. Error between duplicate runs is smaller than the data point label. R² values for the linear fits are >0.99. (B) M_n of solids collected post-reaction, determined by GPC in HFIP. Error bars represent the value range between duplicate runs. Full GPC traces can be found in Fig. S11.

Fig. 4 shows the conversion of polymer to BHET over time for the different polymers. It is evident that initial glycolysis rates of PET, ^IPPET-10, and ^2-MeOPET-10 are relatively similar, and that ^2,6-MeOPET-10 is roughly 3 times faster. It is worth noting that a lower molar mass batch of ^2,6-MeOPET-10 was used (Table S9), but this is not expected to affect glycolysis rates.⁵⁵ Crystallinity, which has been shown to affect depolymerization rates,⁵⁶ is similar for all polymers (X_c = 27–35%) and hence does not account for differences in glycolysis rates. The increased rate for ^2,6-MeOPET-10 is most likely due to the more labile linkages formed by ^2,6-MeOTPA, as the carbonyl between the two methoxy groups is more electrophilic and susceptible to glycolysis. The facile scission of these linkages leads to faster formation of oligomers, which in turn leads to higher BHET yield.

Methoxyterephthalate esters as plasticizers.

Phthalate-based plasticizers are typically used to tune the properties of PVC for specific applications and are currently being phased out due to their apparent reproductive toxicity.⁵⁷ Two common plasticizers found in packaging are diethyl hexyl phthalate (DEHP) and diethyl butyl phthalate (DBP).^58–59 Terephthalate-based plasticizers are gaining importance as phthalate alternatives as they have been found to be more benign, making them the highest volume, general purpose plasticizers today.⁶⁰ With this in mind, we considered the potential application for these lignin-derived aromatic compounds as plasticizers. For each methyl ester, the 2-ethyl hexyl and butyl derivatives were formed, using Ti(OⁱPr)₄ as the transesterification catalyst. The resulting esters are denoted as ^2-MeODEHT, ^2-MeODBT, ^2,6-MeODEHT, and ^2,6-MeODBT (Scheme 5 and SI for details). The analogous terephthalate esters, DEHT and DBT, were also prepared, such that their performance could be directly compared. Plasticized PVC films were prepared by dissolving unplasticized PVC (UPVC) with 10 wt% plasticizer in THF, and then slowly evaporating the solvent in a Teflon mold. The resulting materials were thermally characterized by DSC to measure the plasticization (T_g), and by TGA to measure the polymer thermostability (T_d,10%). The results are summarized in Fig. 5.

Scheme 5:

Synthesis of methoxyterephthalate ester plasticizers using excess 2-ethylhexanol or n-butanol.

Figure 5:

(A) T_d,10% of neat plasticizers, measured by TGA at 20 °C min⁻¹ (Table S14, Fig. S12). (B) T_g and T_d,10% data for PVC containing 10 wt% of different plasticizers. The T_g was measured by DSC at 10 °C min⁻¹, second heating scan. Thermographs can be found in Fig. S14. The T_d,10% was measured by TGA at 20 °C min⁻¹. Thermographs found in Fig. S13. (C) Fitting diffusion coefficients to Arrhenius equation at temperatures 600 K, 550 K, and 500 K and Williams-Landel-Ferry (WLF) equation at temperatures 500 K, 450 K, and 400 K to extrapolate the high temperature data to 300 K. Error bars are smaller than 2% of the calculated diffusion values. (D) Predicted diffusion coefficients at T = 300 K after Arrhenius and WLF fittings.

The additional methoxy groups in the lignin-derived terephthalates add mass to the plasticizer, in turn decreasing the volatility compared to their controls, as measured by TGA (Fig. 5A). It is apparent that all the aromatic esters investigated are effective plasticizers, with a significant decrease in T_g compared to UPVC (Fig. 5B). In particular, the PVC sample with ^2,6-MeODEHT as a plasticizer is most effective, with a comparable T_g (47 °C) to DEHP (45 °C). The stability of PVC is affected by the addition of an additive, which is evident by the reduction in T_d,10% for most samples. The methoxylated terephthalate esters all exhibit a lower decomposition temperature compared to UPVC, although critically are an improvement to the most common plasticizer DEHP (Fig. 5B).

For a plasticizer to volatilize or leach into the environment, it must first diffuse through the polymer matrix to reach the surface. Therefore, higher diffusion rates often correlate with higher volatility and leaching. We estimated the diffusion coefficient of the eight plasticizers using molecular dynamics (MD) simulations. The systems included 10 wt% plasticizer in PVC, with a degree of polymerization of 150. Higher temperature simulations (400–600 K) were used to estimate diffusion coefficients at 300 K as the diffusive regime was not reached due to low molecular mobility at room temperature. The Arrhenius equation was applied to diffusion data above T_g + 100 K (i.e. 500 K) and the Williams-Landel-Ferry (WLF) model below T = 500 K.^61–62 The Arrhenius approach effectively describes the exponential relationship between temperature and diffusion at higher temperatures, capturing the activation energy required for molecular movement. As molecular mobility decreases with decreasing temperature, free volume also decreases. Below the glass transition, the mobility at any temperature primarily depends on the available free volume as described by WLF equation.^61–64

In agreement with the experimental volatility study, the MD results indicate that most incumbent molecules showed higher diffusion rates compared to the lignin-derived plasticizers with methoxy groups, as shown in Fig. 5C–D. At 400 K and above, log-log plots mean square displacement (MSD) as a function time exhibit a slope of 1, which is an indicator of the systems reaching the diffusive regime (Fig. S15). The plasticizers with additional methoxy groups and alkyl chain branching showed lower mobility due to the known relationship between molar mass and diffusion that can be observed in the Arrhenius fitting of the diffusion data (Fig. 5C). The WLF model effectively captures the effect of temperature on polymer segmental mobility near T_g, using T_g values from the experimental data in Fig. 5B. The potential discrepancies between experimental and simulated T_g values were adjusted through fitting constants.⁶⁵ The three plasticizers with the highest diffusion coefficient are DBP, DEHT, and DBT (consistent with prior studies),^66–70 and the lowest three are the methoxylated plasticizers, ^2,6-MeODBT, ^2,6-MeODEHT, and ^2-MeODEHT.

Discussion

As our ability to refine biomass improves, it is important to find suitable applications for new bio-based building blocks. In this work, we targeted two applications of interest because: 1) PET is becoming increasingly important in the context of circular polymers, as it is the most widely recycled commodity plastic, and 2) the phasing out of phthalate esters due to toxicity concerns means that alternatives must be found. Moreover, minor co-monomers and plasticizers are used in lower volumes than commodity plastics like PET, thus offer promising avenues for scale-up of bio-based chemicals.

We found that the methoxylated terephthalates, which can be derived from the carboxylation of lignin-derived aromatic hydroxy acids, provide performance advantages as co-monomers in PET and as PVC plasticizers. In PET, the methoxy substitutions disrupt the crystallinity of PET in a similar fashion to the 1,3-substitution of IPA. In particular, ^2-MeOPET-10 shows equivalent stiffness to PET while exhibiting lower crystallinity and improved processability. Polymerizations are also easily transferrable to existing manufacturing, given the reactivity of the monomer. Although thermal data for ^2,6-MeOPET-10 are comparable, the co-polymer has a lower modulus and generally yields lower molar masses because the dimethoxylated monomer exhibits low reactivity. This also leads to more facile cleavage of ester linkages, which was apparent in the initial rates of glycolysis of ^2,6-MeOPET-10. For similar reasons, ^2,6-MeOTPA is also more challenging to synthesize through electrochemical carboxylation, leading to overall lower yields.

The ^2-MeOTPA monomer should be explored further as a potential bio-based component in PET; particularly, the resulting materials would be interesting to test as polyester fiber, where the addition of a co-polymer could ease processing conditions without compromising performance. Additionally, small amounts of ^2,6-MeOTPA could be added to accelerate polymer glycolysis and improve polymer circulatity. Good targets for future work include investigating mixtures of ^2-MeOTPA and ^2,6-MeOTPA that balance material performance and glycolysis rates. An ideal scenario is to use a mixture from a lignin stream, which would have G and S acids in the desirable ratio. Other examples of bio-based PET co-polyesters reported in the literature have shown modulation of PET properties, using 2,5-furandicarboxylic acid,³⁸ isosorbide,⁷¹ and lactide.⁷² Most of these examples cause significant changes to PET such as crystallization rates, which can be seen by prominent cold crystallization peaks by DSC under standard 10 C min⁻¹ heating and cooling cycles. According to our results, the advantage of ^2-MeOTPA as a co-monomer is its similarity to IPA at low percentage incorporation, and can hence be readily substituted.

Plasticizers are imperative for the diverse uses of PVC in industries including construction, automotive, and food. The application space is broad, with a variety of structures dictating different applications. We targeted ortho-phthalates and terephthalates as our benchmark plasticizers due to their similarity in structure to common plasticizers today. The plasticizers produced from ^2-MeOTPA and ^2,6-MeOTPA show plasticizing behavior similar to the relevant general plasticizers, and their lower volatility and predicted diffusion is desirable for extending the performance lifetime of PVC. Typically, syringic di-methoxylated phenol units are more abundant in lignin, however their applications are scarcer. The performance advantages for ^2,6-MeOTPA-derived plasticizers makes it a good candidate for bio-based non-phthalate general plasticizers. Although endocrine disruption studies were not performed here, addition of methoxy groups to aromatics has been shown to decrease toxicity relative to bisphenol A.⁷³

Conclusion

The present study shows that vanillic and syringic esters obtained from lignin can be used as building blocks to access methoxy-substituted terephthalates, ^2-MeOTPA and ^2,6-MeOTPA, via electrochemical carboxylation. These compounds show utility in two notable applications, as co-monomers in PET and as PVC plasticizers. The addition of ^2-MeOTPA and ^2,6-MeOTPA as co-monomers modulate the properties of PET, lowering the crystallinity and melt temperatures, and producing amorphous polymers when incorporated beyond 25 mol%. At 10 mol% incorporation, similar to that of isophthalate in certain PET applications, the lignin-derived monomers showed improvements in storage modulus, lower complex viscosities, and faster initial rates of chemical recycling by glycolysis. These polymers merit further investigation as polyester fiber, which would benefit from added bio-based content and from lower manufacturing temperatures. The monomers also show promise as plasticizers for PVC, decreasing the T_g comparably to the state-of-the-art general plasticizers. Their lower volatility also guarantees longer-term plasticization. Molecular dynamics simulations predict reduced diffusion coefficients that should make these compounds less susceptible to leaching. Overall, the methoxy functionalities inherent in lignin-based aromatics contribute to performance advantages relative to products where fossil-based phthalates are typically used.

Experimental Methods

Chemical synthesis of methoxyterephthalates via hydroxyterephthalic acids

2-Hydroxyterephthalic acid dimethyl ester.

Concentrated sulfuric acid (50.0 mL, 943 mmol, 1.72 equiv) was added slowly to a solution of 2-hydroxyterephthalic acid (100 g, 549 mmol, 1 equiv) in methanol (800 mL) in a 1 L round bottom with a ~4 in stir bar at 0 °C. A reflux condenser was attached and the reaction was placed under nitrogen. The solution was then heated to 65 °C by placing in a preheated oil bath. The reaction was stirred for 20 h at 65 °C. The hot product mixture was diluted in 4.0 L of ice-cold deionized water to precipitate the product. The product was collected via vacuum filtration and washed with ice cold deionized water (3 × 300 mL). The product was air dried overnight and then dried further for 24 h at 60 °C under vacuum to provide 2-hydroxyterephthalic acid dimethyl ester as white solid (106 g, 92%). The dried product obtained in this way was used without further purification in the subsequent step. NMR spectroscopic data for 2-hydroxyterephthalic acid dimethyl ester obtained in this way were in agreement with those previously reported.³⁹

2-Methoxyhydroxyterephthalic acid dimethyl ester (^2-MeODMT).

Oven dried potassium carbonate (139 g, 1.01 mol, 2.00 equiv) was added to a solution of 2-hydroxyterephthalic acid dimethyl ester (106 g, 0.504 mol, 1.00 equiv) in acetone (1 L) in an oven dried 2 L round bottom with a ~4 in stir bar at 22 °C. Methyl iodide (47.1 mL, 756 mmol, 1.50 equiv) was added via syringe at 22 °C. The heterogenous solution was heated to 56 °C by placing it in a preheated oil bath. The reaction was stirred for 69 h at 56 °C. The product mixture was concentrated. The residue was suspended in dichloromethane (300 mL) and filtered through Whatman filter paper (150 mm). The collected solid was washed with dichloromethane (3 × 150 mL). The filtrates were combined and washed with deionized water (3 × 350 mL). The washed organic layer was dried over sodium sulfate. The dried solution was filtered and the filtrate concentrated to provide 2-methoxyterephthalic acid dimethyl ester as an off white solid (111 g, 98%).

NMR spectroscopic data for 2-methoxyterephthalic acid dimethyl ester obtained in this way were in agreement with those previously reported.³⁹

The product was further purified to obtain polymer-grade purity monomer. The dimethyl ester was sublimed at 90 °C under vacuum (100 mTorr) and then recrystallized in an ethyl acetate/hexane mixture to obtain white crystals (m. p. = 72.3 °C, 99.5% pure by DSC).

2,6-Dihydroxyterephthalic acid dimethyl ester.

Concentrated sulfuric acid (25.0 mL, 472 mmol, 1.86 equiv) was added slowly to a solution of 2,6-dihydroxyterephthalic acid (50.2 g, 253 mmol, 1 equiv) in methanol (500 mL) in a 1 L round bottom with a ~4 in stir bar at 0 °C. A reflux condenser was attached, and the reaction was placed under nitrogen. The solution was then heated to 65 °C for 20 h. The product mixture was concentrated, and the residue was suspended in water (500 mL) and filtered through Whatman filter paper (150 mm). The collected solid was washed with water (3 × 250 mL). The product was air dried overnight and then dried further for 24 h at 60 °C under vacuum to provide 2,6-dihydroxyterephthalic acid dimethyl ester as white solid (36.4 g, 64%). The dried product obtained in this way was used without further purification in the subsequent step.

NMR spectroscopic data for 2,6-dihydroxyterephthalic acid dimethyl ester obtained in this way were in agreement with those previously reported.⁴⁰

Dimethyl 2,6-dimethoxyterephthalate (^2,6-MeODMT).

Oven dried potassium carbonate (78.0 g, 564 mmol, 3.50 equiv) was added to a solution of 2-hydroxyterephthalic acid dimethyl ester (106 g, 504 mmol, 1 equiv) in acetone (1 L) in an oven dried 2 L round bottom with a ~4 in stir bar at 22 °C. Methyl iodide (30.1 mL, 483 mmol, 3.00 equiv) was added via syringe at 22 °C. The heterogenous solution was heated to 56 °C for 69 h. The product mixture was concentrated, and the residue was suspended in dichloromethane (150 mL) and filtered through Whatman filter paper (150 mm). The collected solid was washed with dichloromethane (3 × 150 mL). The filtrates were combined and washed with deionized water (3 × 100 mL). The washed organic layer was dried over sodium sulfate. The dried solution was filtered and the filtrate concentrated to provide 2,6-dimethoxyterephthalic acid dimethyl ester as an off white solid (35.4 g, 87%).

NMR spectroscopic data for 2,6-dimethoxyterephthalic acid dimethyl ester obtained in this way were in agreement with those previously reported.⁴¹

The product was further purified to obtain polymer-grade purity monomer. The dimethyl ester was sublimed at 130 °C under vacuum (100 mTorr), then passed through a silica plug using EtOAc as eluent, and finally recrystallized in a water/methanol mixture to obtain white crystals (m. p. = 121.5 °C, 98.4% pure by DSC).

Electrochemical synthesis of methoxyterephthalates via vanillic and syringic methyl ester

General procedure for synthesis of aryl sulfonate esters.

The following procedure was adapted from precedents reported in the literature.^42–43 Methyl vanillate or methyl syringate (50 mmol, 1.0 equiv) and anhydrous DCM (200 mL) were added to a 500 mL round-bottom flask. Et₃N (10.4 mL, 75 mmol, 1.5 equiv) was then injected into the solution. Another 200 mL round-bottom flask was charged with the sulfonylating reagent (60 mmol, 1.2 equiv) and anhydrous DCM (50 mL), mixed thoroughly, then this solution was added dropwise to the 500 mL flask. The reaction mixture was stirred at room temperature and stopped when full conversion was observed via TLC. The DCM solution was washed with 300 mL water and 300 mL brine sequentially. The aqueous layers were collected and extracted with 200 mL DCM. The organic layers were combined, dried over anhydrous Na₂SO₄, and concentrated in vacuo. The crude product was purified by column chromatography to furnish the desired product, G-OTs and S-OTf (eluent: hexane/ethyl acetate, 2/1).

General batch electrolysis procedure for the carboxylation of aryl sulfonates.

To an undivided cell (Fig. S2) was added a cross-shaped stir bar, NiBr₂^●dme (10 mol%) aryl sulfonate (1.0 equiv) and neocuproine (20 mol%). The cell was then transferred into a N₂-filled glove box where anhydrous TBABr (0.2 M), MgCl₂ (1.0 equiv) and LitBuO (50 mol%) were added together with 3 mL of anhydrous DMSO. The cell was sealed with a rubber septum containing an RVC(−) as cathode and a Zn plate as sacrificial anode (+) as well as a Teflon tube as CO₂-inlet. The RVC is cut into a piece of 3 × 1 × 0.5 cm; during electrolysis the electrode is submerged only by 2 cm in height. The mixture was fully solubilized before being subjected to CO_2. At this point, the cell was removed from the glovebox and connected through the thin Teflon tube to a CO₂-line and CO₂ was bubbled for 20 min into the cell before electrolysis. The reaction mixture was electrolyzed under an applied constant potential of −1.8 V vs Fc until 2.2 F mol⁻¹ were passed. For quantification by ¹H NMR spectroscopy, 0.15 mmol of TMB (1,3,5-trimethoxybenzene) was added to the crude reaction, and 0.1 mL of this solution was directly diluted in 0.5 mL of CDCl₃.

General flow electrolysis procedure for the carboxylation of aryl sulfonates.

For the electrochemical flow reactions, a commercial Micro Flow Cell (purchased from ElectroCell) with an electrode area of 10 cm² was used, the active reactor volume is 5 mL, and a Pine WaveNow PGstat was used as power supply. The undivided flow cell consists of PTFE end frames, zinc plate as the anode, stainless steel plate and graphite plate overlay together as the cathodic electron collector. The flow cell also contains the flow frames and gaskets. Carbon-felt with approx. dimensions = 3.5 × 3.0 × 0.5 cm was used to increase the surface area of the cathode and as a turbulence material for diffusion. All electrolysis reactions were performed in DMSO solutions. A magnetic stir bar was used in the reservoirs and the reaction mixture was stirred (1000 rpm) during flow electrolysis reactions. The reaction mixture is pumped through the system via peristaltic pump with a flow rate of 30 mL/min. The components of the electrochemical cell are shown in Fig. S3. Experimental details for the flow electrolysis of G-OTs and S-OTf can be found in the Supporting Information.

General procedure for preparation of polyesters

Dimethyl ester (1 equiv.) and EG (2.5 equiv.) were combined in a three-neck flask equipped with a short path distillation piece and an overhead stirrer. The flask was evacuated and re-filled with N₂ three times, after which Ti(OⁱPr)₄ (1 mol%) was added under a positive flow of N₂. The flask was lowered into a preheated molten metal bath at 180 °C. The reaction was stirred for 2 h, or 16 h for ^2,6-MeODMT-containing polymers. The temperature of the bath was then raised to 250 °C and kept under N₂ for 1.5 h. The reaction was subsequently subjected to ~2 Torr of vacuum for 2 h, or 8 h ^2,6-MeODMT-containing polymers. In the case of PET synthesis, the temperature was raised to 280 °C and kept under vacuum for 2 h. At the end of the polymerization, the flask was allowed to cool completely under N₂. The crude polymer was dissolved in a 1:1 mix of CHCl₃ and trifluoroacetic acid and precipitated in MeOH. The isolated purified polymer was then dried in a vacuum oven at 90 °C for 18h prior to further analysis.

Analysis of glycolysis kinetics of PET and PET co-polymers

Glycolysis reactions were run in 12 mL stainless steel Swagelok tube reactors in a fluidized sand bath purchased from Techne. Reactions were run at 170 °C with a 16:1 EG:PET molar ratio based on the PET repeat unit and 5 wt% DMEA catalyst relative to PET. Reactions were performed in duplicate and each time point was a discrete reaction in a separate reactor. For a given substrate and timepoint, PET or a derivative (8.5 mmol), 8.44 g EG (7.60 mL, 136 mmol), and 82 mg DMEA (117 μL, 1.12 mmol) were added to a reactor. The reactor threads were covered with anti-seize paste and the reactor lid was tightened using a wrench and bench vise. The mass of the filled reactor was recorded. This process was repeated for each timepoint. Each reactor was thoroughly mixed before being heated to avoid diffusion limitations. The reactors were then dropped into the preheated sand bath reactor in reverse time order (longest timepoints first). After the desired reaction time, all reactors were removed from the sand bath at once and quenched in ice water for 15 minutes.

After cooling, the mass of each reactor was measured to check for leaks (any leaked reaction was rerun). The reactors were reheated to 120 °C for 15 minutes to solubilize monomer products. The hot reaction mixture was quickly filtered using a pre-weighed 60 mL capacity 10 μm polyethylene filter, and the soluble fraction was collected directly into a pre-weighed 3-dram vial. The solids remaining in the filter were washed with at least 30 mL deionized water with thorough mixing in a separate vacuum flask to remove excess EG. Solids in the filter were dried at 50 °C in a vacuum oven overnight. This resulted in the monomeric products being collected in the liquid fraction and the remaining polymeric material being isolated in the solid fraction. The soluble fraction was diluted in THF, and HPLC was used to analyze monomer yields. The solid fraction was analyzed by HFIP GPC to analyze residual polymer molar mass and molecular mass distributions.

General procedure for preparation of plasticizers

Titanium iso-propoxide (3.70 mol%) was added to a solution of dimethyl ester (1.00 g, 1 equiv) in the desired plasticizer alcohol (10.0 mL) in a 40 mL microwave vial (Biotage^®) at 22 °C. The reaction mixture was sealed with a crimp cap and was heated at 170 °C in a heating block for 20 h. The product mixture was cooled to room temperature and the excess alcohol was removed via vacuum distillation. The residue obtained was purified via flash-column chromatography.

Procedure for preparation of PVC films

UPVC powder (100 mg) and plasticizer (10 wt%) were dissolved in THF (2 mL) overnight. The solution was poured into a PTFE mold, and the solvent was left to evaporate over 24 h at ambient conditions. The resulting films were dried in a vacuum oven at 70 °C for 18 h.

Molecular Dynamics (MD) simulations.

Diffusive behavior of plasticizers in PVC systems were studied using MD simulations, which infers the leaching potential of plasticizers from the matrix. Each system contains 10 wt% plasticizer, approximately 80,000 atoms, and with a degree of polymerization of 150 for PVC (ensures molar mass is in the polymer regime according to the Fox-Flor relationship for T_g).⁴⁴ Lignin-derived methoxylated compounds were compared to four industrially-relevant incumbent plasticizers, namely diethylhexyl phthalate (DEHP), dibutyl phthalate (DBP), diethylhexyl terephthalate (DEHT), and dibutyl terephthalate (DBT). Table S1 provides details on the number of PVC chains and corresponding plasticizer content. See the Supporting Information for details on the MD simulations.

Supplementary Material

The Supporting Information is available free of charge at (link).

Description of experimental methods, small molecule and polymer characterization (NMR spectroscopy, DSC, TGA), and computational methods.

Synopsis.

Monomers derived from lignin were used as co-monomers in poly(ethylene terephthalate) and as plasticizers for PVC, and their performance for each application was assessed.