Chemical Recycling of Thermoset Composites: Why Solvolysis Chemistry Must Match the Material

Conventional thermoset polymers are irreversibly cured. The reaction of resin and hardener creates a hard, 3D network of covalent bonds that can’t be broken with heat or regular solvents. The reason for this impressive structural performance of thermoset composites and the reason that conventional recycling doesn’t work is that thermosets contain permanent covalent networks and therefore they do not melt and flow when heated. Although solvents may swell into the network, depolymerization requires cleavage or exchange of network bonds.

Solvolysis involves a reactive solvent, such as water, an alcohol or an organic acid, sometimes together with an acid, base or metal-salt catalyst which breaks specific covalent bonds in the network under controlled temperature and time. During solvolysis, part or all the matrixes may be converted into soluble monomers or oligomers, while insoluble residues may remain depending on the resin and reaction conditions. (Lahive et al. 2025.) Under optimal selection of conditions, the fibers can be obtained with minimal damage, and properties close to those of virgin fibers. (Kuang, Zhou, Shi, Wang & Qi 2018; Lahive et al. 2025).

However, solvolysis cannot regenerate well-defined molecular components if those components did not exist as discrete chemical species in the original material. The effectiveness of each method is primarily dependent upon the type of bonds present in the thermoset to be treated, and the ability of the chemical system to react under the applied conditions. (Lahive et al. 2025; Ahrens, Bonde, Sun, Wittig, Hammershøj, Batista, Sommerfeldt, Frølich, Birkedal & Skrydstrup 2023.)

However, alkaline systems can also react with amine cured epoxy matrices, as shown below. Zhao and co-workers showed in 2020 that a matrix can be 98.82% degraded in 90 minutes at 160°C using a solution of monoethanolamine (MEA) and potassium hydroxide (KOH), and that the recovered carbon fibers had a loss of tensile strength of 6.5% compared to virgin fiber. (Zhao et al. 2020.) The mechanism here is not C-N crosslink bond but C-O (β-hydroxy ether) in the DGEBA resin. This is significant for the recovered resin fraction: In the case of a C-O cleavage, hydroxyl-functional (-OH) oligomers are obtained instead of amine-functional (-NH₂) oligomers. This affects how the recovered resin can be reused. Hydroxyl-functional oligomers are suitable for transesterification-based reformulation. Alkaline C-O cleavage does not liberate primary amine groups that are needed for formulation of imine bonds with aldehyde crosslinkers for reprocessable vitrimer network synthesis.

In 2025, Lim et al. at USC discovered a more potential alkaline approach: the molten sodium and potassium hydroxide eutectic at 220°C breaks both the aryl ether backbone bonds as well as the amine linkages in aerospace grade epoxy composites, recovering bisphenol-A up to quantitative yields. (Lim, Yu, Cherepakhin, Williams, & Nutt 2025.) The recovered carbon fiber fabric was successfully remanufactured into second-generation composites, demonstrating the viability of the process for circular material cycling. Both bond types are cleaved at 220°C; at lower temperatures, only the C-O backbone is targeted.

A team from Aarhus University has published a new approach in 2023 that used ruthenium catalysis to cleave C-O backbone bonds of bisphenol A based epoxy resins back to near-quantitative recovery of the bisphenol A monomer and intact fibers. (Ahrens et al.  2023.) The cost, recovery, and repeated use of the ruthenium catalyst should be considered when evaluating industrial scale-ups.

The choice among these routes for amine-cured epoxy depends on what is the end goal. For fiber recovery alone, MEA/KOH at 160°C (Zhao et al. 2020) and acetic acid acetolysis (Lahive et al. 2025) both recover high-quality carbon fibers; however, the acetolysis process requires high-temperature pressurized conditions (approximately 280-300°C in sealed reactors). For selective C-N bond cleavage, resulting in nitrogen-containing oligomers with recovery of fibers, Lewis acid-mediated processes based on AlCl₃ or ZnCl₂ can be chosen. (Wang et al.; Deng et al. 2015.)

Anhydride-cured epoxy composites are mostly employed in electrical encapsulation and printed circuit boards and are also used in some wind-turbine parts, During the anhydride curing reaction, ester bonds are formed in the main chain of the network. Alcohol, amines and water under mild base or acid catalysis can break ester bonds, a chemistry that is well known and doesn’t need special Lewis acids or precious metals.

The researchers found that they could dissolve the anhydride-cured epoxy composites to 95 percent mass loss in 70 minutes at 170°C and ambient pressure, leaving the carbon fibers with no loss in mechanical properties by using solvent combined with a reactive alcohol. (Kuang et al. 2018.) Zhao et al. (2022) demonstrated a closed-loop recycling strategy in which potassium phosphate (K₃PO₄) catalyzed alcoholysis cleaved the ester bonds selectively, followed by hydrolysis to yield carboxyl-rich degradation products. The catalyst was readily recovered by filtration owing to its low solubility in ethanol at room temperature, and the degradation products were reused as curing agents for new anhydride-cured epoxy thermosets without compromising their strength or stability (Zhao, Liu, Feng, An, Tian, Du, Xu, Chen, Wu & Wang 2022). Liu and co-workers were able to re-use in the closed loop using phosphotungstic acid as a catalyst, with up to 40 wt.% of the degraded products incorporated into a new formulation without a loss in mechanical performance (Liu, Guo, Liu, Hao, Wang, Hiscox, Liu, Jin, Xin & Zhang 2017).

One key aspect of anhydride-cured epoxy systems is that the ester bonds are accessible to both alcoholysis and aminolysis. Alcoholysis at 170°C depolymerizes the network and enables recovery of carbon fibers (Kuang et al. 2018). Subsequent studies exhibited that the resulting degradation products can be reincorporated into new epoxy formulations (Liu et al. 2017). While Zhao et al. (2022) further demonstrated that carboxyl-rich degradation products could be directly reused as curing agents for new anhydride-cured epoxy thermosets.

Unsaturated polyester resin (UPR) is the thermoset material widely used in wind turbine blade applications. Moreover, UPR is also employed in boat hulls, body panels of automobiles, and building components. First, it is important to recognize that UPR composites contain two distinct types of bonds. These bond types must be considered separately to understand and evaluate potential recycling approaches. (Arturi, Sokoli, Søgaard, Vogel & Bjelič 2018, Van de Moosdijk, van de Runstraat, van Someren, Roelands, Krauklis, Nazemi, Muzyka & Shuaib 2025.)

The polyester backbone is made from condensation of unsaturated diacids or anhydrides and glycol during the manufacture. These ester bonds in the backbone are in theory susceptible to the same hydrolytic and aminolytic cleavage pathways exploited in chemical recycling of anhydride-cured epoxies. The cross-links between polyester chains, however, are made by free-radical copolymerization with styrene, and involve the formation of carbon-carbon (C-C) bonds. Under hydrolytic and aminolytic conditions, degradation proceeds primarily through cleavage of the polyester ester bonds, whereas the styrene-derived C-C crosslinks are considerably more resistant to these reactions. (Arturi et al. 2018, Van de Moosdijk et al. 2025.)

Subcritical hydrolysis using acetone/water in the range 250-300 °C can be used to extract phthalic acid and dipropylene glycol from the polyester backbone, leaving the styrenic fraction as a carbonaceous solid residue (Arturi et al. 2018). A study at TNO in the Netherlands identified MEA/KOH at 170°C as the most promising combination for recovery of glass fiber from automotive SMC for practical use. No significant reduction in fiber tensile strength was observed under the selected conditions, although the treatment affected fiber-surface sizing. (Van de Moosdijk et al. 2025.) In the polyester, ester bonds are aminolyzed by MEA, and the alkalinity is provided by KOH to speed up the reaction. The styrene-crosslinked portion is not soluble and becomes a residue which needs to be managed separately. Any UPR recycling process which fails to consider the non-recoverable styrene fraction overestimates the process’s effectiveness. (Van de Moosdijk et al. 2025.)

In addition to these factors, particle size is also an important factor defining the success of reaction.  A particular chemical system may be suitable or unsuitable for a given material. MEA/KOH at 160^oC to 170^oC can effectively dissolve amine-cured epoxy and recover glass and carbon fibers (Zhao et al. 2020) but only results in hydroxyl-functional resin recyclate and not amine-functional. Lewis acid-mediated recycling selectively cleaves the C-N bond, preserving nitrogen-containing oligomeric products (Wang et al. 2015; Deng et al. 2015). In contrast, Lahive et al. (2025) demonstrated catalyst-free acetolysis of amine-cured epoxy thermosets using glacial acetic acid at high temperature, recovering carbon fibers and depolymerization products without reporting selective preservation of amine-functional oligomers.

Selective alcoholysis at 170°C depolymerizes anhydride-cured epoxy networks through ester bond cleavage and generates epoxy-derived oligomers that can be reused as reactive ingredients in new epoxy materials. (Kuang et al. 2018.) For UPR, ester-targeting solvolysis alone cannot achieve complete matrix degradation because the styrene-derived C–C crosslinks are more resistant; additional oxidative chemistry may be required for their cleavage.

Life cycle assessment of supercritical water solvolysis demonstrated that the environmental benefit of CFRP recycling arises primarily from avoiding the production of virgin carbon fiber, rather than from the recycling process itself (Prinçaud, Aymonier, Loppinet-Serani, Perry & Sonnemann 2014).

Overall, the life cycle studies indicate that the environmental benefit of solvolysis can be substantial when recovered fibers replace energy and resource-intensive virgin fibers. However, the real outcome depends on the functional unit, allocation method, fiber quality, substitution ratio, process energy, solvent and catalyst recovery, and treatment of liquid and solid residues.

The bond type in the matrix determines the method. C-N and C-O ether bonds are the main constituents of amine-cured epoxy networks. Ester bonds are present in anhydride-cured epoxy network. UPR combines a polyester backbone containing hydrolysable ester bonds with a cross-linked network formed through styrene monomers. A different chemical approach is required for each of them.

Multiple methods work for amine-cured epoxy, for different outcomes. Lewis acid routes (AlCl3/AcOH; ZnCl2/water) break C-N bonds and recover functional oligomers with amine rich moieties.  C-O backbone bond is cleaved by MEA/KOH at 160°C aiming at recovery with hydroxyl-functional oligomers. Acetolysis using glacial acetic acid enabled the depolymerization of epoxy thermosets cured with both aliphatic and aromatic amines, allowing the recovery of carbon fibers without the need for a metal salt catalyst. Molten NaOH-KOH at 220°C breaks both bond types and recovers bisphenol-A. In a nutshell, the choice of chemical reaction is dependent on the recovery goal.

Anhydride-cured epoxy contains ester linkages that can enable comparatively mild depolymerization and reuse of the resulting products in new formulations.

UPR recycling has a structural limit. MEA/KOH-assisted ester bond cleavage depolymerizes the polyester matrix and enables recovery of clean glass fibers with minimal loss of mechanical properties because esterolytic processes primarily target the polyester backbone, the styrene-derived C-C crosslinked fraction remains more resistant and may persist as an insoluble residue.  Literature review has shown that an additional Fenton oxidation step is required to cleave these stable C-C bonds and achieve more complete degradation of the UPR matrix.

The functional group of recovered oligomers matters for reuse. Alkaline C-O cleavage routes yield Hydroxyl-rich functional oligomers that suit transesterification-based applications. C-N cleavage routes yield Amine-functional oligomers that suit amine-reactive downstream chemistry. These are not interchangeable.

This work was carried out as part of the CompositeCircle project at Centria University of Applied Sciences. This project is supported by funding from Interreg BSR Programme, whose contribution is greatly acknowledged. The aim of the Composite Circle project is to promote sustainable development and circular solutions for composite materials, with peculiar emphasis on enhancing resource efficiency, material recovery, and the reuse of composite constituents.