End-of-life upcycling of polyurethanes using a room temperature, mechanism-based degradation

A major challenge in developing recyclable polymeric materials is the inherent conflict between the properties required during and after their life span. In particular, materials must be strong and durable when in use, but undergo complete and rapid degradation, ideally under mild conditions, as they approach the end of their life span. We report a mechanism for degrading polymers called cyclization-triggered chain cleavage (CATCH cleavage) that achieves this duality. CATCH cleavage features a simple glycerol-based acyclic acetal unit as a kinetic and thermodynamic trap for gated chain shattering. Thus, an organic acid induces transient chain breaks with oxocarbenium ion formation and subsequent intramolecular cyclization to fully depolymerize the polymer backbone at room temperature. With minimal chemical modification, the resulting degradation products from a polyurethane elastomer can be repurposed into strong adhesives and photochromic coatings, demonstrating the potential for upcycling. The CATCH cleavage strategy for low-energy input breakdown and subsequent upcycling may be generalizable to a broader range of synthetic polymers and their end-of-life waste streams. Extensive crosslinking in thermosetting polymers provides their desirable durability but makes them difficult to recycle. Now acetal-based monomers containing nucleophilic pendant groups have been incorporated into polyurethanes, which are stable in aqueous acid yet degradable at room temperature under organic acidic conditions. The degradation products were upcycled into higher-value, long-lasting materials.

A major challenge in developing recyclable polymeric materials is the inherent conflict between the properties required during and after their life span. In particular, materials must be strong and durable when in use, but undergo complete and rapid degradation, ideally under mild conditions, as they approach the end of their life span. We report a mechanism for degrading polymers called cyclization-triggered chain cleavage (CATCH cleavage) that achieves this duality. CATCH cleavage features a simple glycerol-based acyclic acetal unit as a kinetic and thermodynamic trap for gated chain shattering. Thus, an organic acid induces transient chain breaks with oxocarbenium ion formation and subsequent intramolecular cyclization to fully depolymerize the polymer backbone at room temperature. With minimal chemical modification, the resulting degradation products from a polyurethane elastomer can be repurposed into strong adhesives and photochromic coatings, demonstrating the potential for upcycling. The CATCH cleavage strategy for low-energy input breakdown and subsequent upcycling may be generalizable to a broader range of synthetic polymers and their end-of-life waste streams.
The global effort to reduce polymeric waste entering oceans and landfills has become more urgent as the scale of the problem has come into focus 1,2 . Of the many approaches for making polymeric materials more sustainable, closed-loop recycling, where depolymerization produces the original monomer, has been especially promising 3,4 . The process can be quite efficient for some materials; however, others, especially thermosets, require harsh conditions and most often the breakdown is insufficiently clean and too energy intensive to be practical 5 . Dynamic covalent chemistry 6 offers one promising strategy to make entirely new polymeric materials that contain chemical linkages that are easily broken 7,8 . A complementary approach uses existing polymer chemistry but new degradation processes to generate either the original monomers for recycling or alternative breakdown products for upcycling applications [3][4][5] . In this regard, a particularly challenging class of polymers is polyurethanes.
The widespread commercial success of polyurethanes is attributable to both their ease of synthesis via polyaddition polymerization of multifunctional alcohol and isocyanate monomers, as well as their superior material properties 9 . Monomer structures and polymerization conditions can be adapted for the generation of bulk materials with a wide range of potential applications. This tunability allows polyurethanes to be used in a multitude of consumer products from durable foams, rubbers and adhesives to hard plastics and coatings. The high stability of the urethane linkages provides outstanding Article https://doi.org/10.1038/s41557-023-01151-y structure can produce high-molecular-weight polymers with useful properties that can be repeatedly cycled between monomer and polymer 15 . Our design was also influenced by stimuli-responsive degradable polymers that undergo chain cleavage through iterative cyclization processes [16][17][18][19] . Thus, we envisioned the need for water being obviated by pendant alcohol groups proximal to an acetal (see 1) that might undergo cyclization upon induction of transient oxocarbenium ion chain breaks (2a to 2b; see Fig. 1c,d). The resulting cyclic acetal (3) would serve as a kinetic and thermodynamic trap for chain cleavage in a process we refer to as cyclization-triggered chain cleavage (CATCH cleavage). As outlined in Fig. 1b, we further envisioned a gated system that would avoid premature breakdown. Thus, aqueous acid would be unable to permeate the hydrophobic polymeric network but a combination of organic solvent and organic acid (AND gate) might permeabilize the polyurethane, allowing oxocarbenium ion formation and CATCH cleavage. We describe here a simple acetal-containing polyol (5; Fig. 2) that reacts with common industrial diisocyanates to form robust polyurethanes (1) that are capable of undergoing this type of gated CATCH cleavage degradation to 3 with the derived polyol 4 capable of upcycling into higher-value polyurethane materials. Thus, polyurethanes derived from 5 are robust but can be degraded with minimal energy input and given a useful second life through repurposing.

Polyol synthesis and mechanistic verification in a model
The synthesis of tetrol 5 was accomplished in two steps using a preparation that was conveniently performed on a 50 g scale. Thus, butyraldehyde and allyl alcohol reacted to form the corresponding acetal, which underwent Upjohn dihydroxylation (Fig. 2a). To test the potential for the CATCH cleavage mechanism in a model system, 5 was reacted with butyl isocyanate to give 6 ( Fig. 2b). A small amount of product from reaction at a secondary alcohol group was observed, and this durability during the material's lifetime, but also limits opportunities for energy-efficient polyurethane recycling 3 . As a result, polyurethane waste frequently ends up in landfills, leading to myriad environmental hazards 10 . Methods have been developed recently wherein polyurethanes can be chemically recycled with carbamate exchange catalysts and elevated reprocessing temperatures. This approach is a very promising direction, particularly for existing post-consumer polyurethane materials 11 , but these methods can require a moderately high energy input and there is a need for alternative methods, especially in the area of upcycling. Given the low reactivity of the urethane linkage under mild conditions, one degradation strategy is to integrate a more readily cleavable bond into either the isocyanate or alcohol monomer. However, a key property of polyurethanes is their hydrophobicity, making hydrolysis substantially more challenging. For example, a polyurethane containing ketal groups was reported to be resistant to hydrolysis even at pH 1 and elevated temperature for 2 weeks (Fig. 1a) 12 . These studies underscore the central paradox of the plastic recycling problem 13 . Having stable physical, mechanical and chemical properties for the life span of the material is at odds with the desire for a rapid and complete degradation under the mild conditions sought for end-of-life disposal and possible recycling or upcycling. Herein we report one strategy to achieve these conflicting properties with a gated polyurethane that responds to two signals with transient polymer chain breaks that are kinetically trapped, leading to full breakdown. Furthermore, we show the upcycling of the degradation products into strong adhesives and coatings.

Design of gated, CATCH-degradable polyurethanes
Ring-chain equilibria have proven a useful strategy for depolymerizing suitable condensation polymers 14 20 . This evidence for a significant positive charge build up on the acetal carbon supports the mechanism shown in Fig. 1d.

Synthesis and degradation of polyurethane elastomers
Following the successful verification of the proposed cleavage mechanism, we sought to demonstrate an analogous acetal degradation in a crosslinked polyurethane. Thus, elastomeric polyurethane film 10a was prepared from tetrol 5 and toluene 2,4-diisocyanate-terminated poly(propylene glycol) (PPG-TDI; number-averaged molecular mass (M n ) = ∼2,300) by mixing 5 and PPG-TDI in a 1:2 ratio (Fig. 2c,d). Different ratios of 5 with PPG-TDI were tested, with a 1:2 ratio providing the lowest amount of 5 while maintaining the desired CATCH cleavage degradation. An analogous elastomeric film (11), physically indistinguishable from that shown in Fig. 2d, was prepared from tetrol 9 (Fig. 2c). Polyurethane film 11 served as a control because 9 lacks the acetal group and therefore the ability to degrade by CATCH cleavage. Beyond the two films having identical physical appearances, their thermal and viscoelastic properties, as measured by differential scanning calorimetry (DSC) and dynamic mechanical analysis (DMA) are similar (Fig. 2e). More importantly, thermal gravimetric analysis (TGA) of both films showed similar onset degradation temperatures (≳200 °C), indicating their comparable thermal stability (Extended Data Fig. 3). The TGA results provided evidence that thermal stability is not impaired by the degradable bonds of the acetal moiety. Additionally, these data were consistent with industrially relevant polyurethane elastomers, which begin to decompose at or below 250 °C (ref. 21 ). Film 10a with a 2:1 PPG-TDI-to-tetrol ratio was further characterized by Fourier-transform infrared spectroscopy, which showed a clear alcohol stretch at ~3,400 cm −1 but no detectable residual isocyanate peaks. In   Article https://doi.org/10.1038/s41557-023-01151-y contrast, film 10b with a 5:1 PPG-TDI-to-tetrol ratio showed a residual isocyanate band at ~2,270 cm −1 but a negligible alcohol absorption, consistent with nearly full reaction of all four alcohol groups. Based on the characterization and degradation data, the polyurethane elastomer prepared with a 2:1 PPG-TDI-to-tetrol ratio (10a) was used for further studies because each acetal unit was likely to have at least one residual unreacted alcohol group to participate in CATCH cleavage.
To test the potential degradation of the bulk polyurethane elastomeric materials, 40 mm × 10 mm × 1 mm films of 10a and 11 containing rhodamine B were immersed in a range of organic solvents containing methanesulfonic acid (MSA) as well as 1 M aqueous HCl (see Fig. 3 and Supplementary Figs. 1 and 2). Films of 10a immersed in tetrahydrofuran (THF) containing 1 M MSA completely dissolved within 20 min at room temperature (Supplementary Video 1) but showed no visible change with immersion in 1 M HCl (aq) or 1 M MSA (aq) for at least 2 d. This striking difference supports the notion of rapid degradation of 10a in low-moisture environments via the CATCH cleavage mechanism and further highlights the stability of hydrophobic acetal-containing polyurethanes in strongly acidic aqueous environments (see Fig. 3d for the logic gate table). In contrast, the acetal-free film 11 experienced swelling and some dye loss but otherwise remained fully intact upon immersion in THF with 1 M MSA.
Furthermore, the effect of acid concentration on degradation rate was investigated with both elastomer degradation studies and small-molecule kinetics experiments. Lower concentrations of MSA resulted in slower degradation rates of elastomer 10a (Extended Data We performed tensile DMA studies to test the thermomechanical responses of 10a and 11 to gain more insight into their relative stabilities in aqueous acid. Films 10a and 11 were soaked in 1 M HCl for 24 h and compared with their pristine counterparts. DMA studies on 11 indicated direct overlap of the tan δ and E′ curves for the pristine and soaked films (Fig. 3e). Although 10a had a slight shift in E′ and tan δ after soaking in 1 M HCl, the rubbery plateau for the soaked films indicated that the bulk material remained crosslinked, providing evidence that bulk polymer degradation is limited under aqueous acidic conditions (Fig. 3f). Additional studies must be performed to determine the cause of these shifts in tan δ and storage modulus, but from a practical point of view, such durability in aqueous acid suggests that polyurethane materials derived from tetrol 5 perform well under real-world conditions. To further quantify the degradation and relative stability of these crosslinked polyurethane materials, we performed in situ storage modulus measurements while the films were immersed in 1 M MSA in THF solution. We observed a rapid decrease in E′ of 10a after the   Article https://doi.org/10.1038/s41557-023-01151-y addition of 1 M MSA in THF at 100 s, indicating rapid degradation of the film, whereas the E′ of 11 remained unchanged, providing further evidence that films without acetal groups were stable in MSA/THF solutions (Fig. 3g).
The DMA results indicate degradation for acetal-containing polyurethane 10a, whereas 11 maintains its integrity. Indirect support for the importance of the hydroxyl groups in the breakdown of polyurethane 10a comes from studies of 10b prepared with a 5:1 PPG-TDI-to-tetrol ratio. Analogous to 11, these elastomeric films minimally changed with the MSA/THF treatment, at least in part because nearly all of the hydroxyl groups were converted to urethane linkages and thus were not able to participate in CATCH cleavage. More direct support for the Fig. 1c mechanism was provided by NMR. The heteronuclear single quantum coherence spectrum of the degradation mixture of 10a showed a correlation between large and small 1 H resonances at δ = 4.8 ppm and 13 C signals at 104 ppm, consistent with the acetal protons and carbons of 1,3-dioxolane (major) and 1,3-dioxane (minor) units of 12 (Fig. 4a and Extended Data Fig. 5). Importantly, this signal disappeared upon hydrolysis with aqueous HCl to give 13. The extent and nature of the CATCH cleavage-based degradation was examined by matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS) and gel permeation chromatography (GPC) on 12. To better grasp how the degraded material compared with the starting material, we reacted PPG-TDI with two equivalents of hexylamine to give 14 (Fig. 4b) Fig. 6b). One higher-mass series was observed for 13, corresponding to two PPG-TDI segments tethered together by a glycerol unit. The origin of this degradation product can be traced to the reaction of two PPG-TDI units with a single 1,2-diol unit of 5 during the polymerization. This higher-mass series is derived from the crosslinking that occurs during curing and is not seen in hexylamine-capped PPG-TDI control 14. The GPC provided Ð = 2.31 and M n = 7.7 kDa for 13 and Ð = 1.93 and M n = 2.85 kDa for 14 (Extended Data Fig. 6c). The GPC data further support complete degradation with generation of the original PPG-TDI length and a smaller amount of higher-molecular-weight material derived from crosslinking. We predict that this product distribution should not affect upcycling into new materials.

Repurposing elastomers into strong adhesives and coatings
In addition to their utility as elastomers, polyurethanes are frequently applied as coatings, adhesives and sealants. Indeed, they are strong and desirable adhesives with an ability to adhere to various substrates and be water repellant, and as thermosets they have a wide range of operating temperatures 22,23 . Whereas facile degradation of crosslinked polyurethane materials into reprocessable components is an important first step towards a sustainable polyurethane lifecycle, the utility of the post-degradation materials for reuse or recycling is equally crucial.
To demonstrate the reusability of the degraded materials generated from 10a via CATCH cleavage, we chose to reformulate the post-degradation polyol urethane prepolymer 13 as an adhesive. Polyol 13 was polymerized with 15 wt% of industrially relevant poly[(phenyl isocyanate)-co-formaldehyde] (M n = ~340) (PAPI 2027) (15; Fig. 4c). Adhesives fabricated with PAPI 2027 exhibited strong adhesion to both aluminium and glass substrates. The adhesives withstood a 20 lb weight for 24 and 3 h for glass and aluminium, respectively (Fig. 4c). The significantly stronger adhesion to glass probably arises from the hydrogen bonding between the polyurethane adhesive and the hydroxylated silicate surface 24 . Additionally, lap shear tests were carried out on the PAPI-based adhesive for both substrates. Our repurposed polyurethane adhesive exhibited a magnitude of shear strength on the aluminium substrate that was comparable to cyanoacrylate glue (super glue) (Fig. 4d). For the glass substrate, adhesion of super glue was measured to have a shear strength of 0.8 ± 0.3 MPa, but no measurement was obtainable with the available instrument for the repurposed polyurethane adhesive because the glass substrates fractured before any failure in the adhesive could be observed (Fig. 4e). That is, the adhesion was sufficiently strong that the glass substrate would break during the lap shear test. Thus, we were able to degrade and repurpose our polyurethane material into a practical product that demonstrates competitive performance and potential commercial value.
In addition to forming adhesives, polyol prepolymer 13 can also be repurposed for use in functional coatings 22,25 . Polyurethane coatings are ubiquitous and their properties can vary depending on the desired application, but they are often hard and chemically resistant and provide appealing high-gloss finishes. As a proof of concept, we focused on demonstrating the utility and versatility of the polyol prepolymer 13 to formulate a photochromic coating as a light-responsive coating, loosely analogous to smart window coatings. Thus, we blended a neat mixture of 13, 2,4,6-trioxotriazine-1,3,5(2H,4H,6H)tris(hexamethylene) isocyanate (16) and the well-studied spiropyran dye 26 17 to generate an ultraviolet-responsive coating material. This mixture was applied onto a glass slide using a paintbrush and cured overnight (Fig. 5). The photochromic activity of the repurposed polyurethane coating was tested by incorporating a Z-shaped mask on top of the coating and irradiating the substrate with 365 nm light for 3 min, which generated a purple Z on the coated section of the glass slide corresponding to Article https://doi.org/10.1038/s41557-023-01151-y formation of the merocyanine dye 18. The Z entirely disappeared in 30 min when exposed to white light. The process was repeated for four cycles (Fig. 5 and Supplementary Video 2).

Generality and sustainability of CATCH cleavage
From a sustainability perspective, the use of MSA in the elastomer degradation is advantageous, partially due to the relatively small amount necessary for efficient degradation of polyurethanes. Additionally, MSA is readily biodegradable and substantially less toxic than more commonly used mineral acids such as hydrochloric acid 27 . The environmental impact of this approach can be further limited by isolating the solvent via vacuum distillation and successfully reusing it for additional degradation cycles (Supplementary Fig. 3). We have demonstrated that simple polyol monomer 5 forms thermally and aqueous stable polyurethane elastomeric films that can be degraded into new polyols that can be reused to make different polyurethane materials. From a sustainability standpoint, one disadvantage of 5 is that its preparation utilizes the toxic osmium tetroxide for dihydroxylation. To test the generality of the CATCH cleavage mechanism and provide a greener alternative to 5, we prepared a new tetrol acetal from trimethylolethane as a starting material. The new acetal-containing polyol monomer 21 was conveniently prepared in three steps (Extended Data Fig. 7). Despite the alcohol groups all being primary and the CATCH cleavage mechanism exclusively forming six-membered 1,3-dioxane rings, model polyurethane elastomeric films still degraded within 20 min under anhydrous acidic conditions and the TGA exhibited a similar onset degradation temperature (251 °C) to 10a.
To further explore the generalizability of this approach, we prepared additional polymeric materials using 5, as well as analogous monomers that could degrade via the proposed CATCH cleavage mechanism (Fig. 6). First, we prepared rigid polyurethane resins (26)  Article https://doi.org/10.1038/s41557-023-01151-y diisocyanate produced a hard resin with a higher crosslinking density and reduced swelling ability. Nonetheless, this material underwent degradation at room temperature, albeit slower, probably by surface erosion (Supplementary Fig. 4). To apply this methodology to other polymerizations, we copolymerized tetrol 5 and activated diester 22 to prepare a crosslinked polyester resin (27) that rapidly degraded in 1 M MSA in dichloromethane at room temperature ( Supplementary  Fig. 5). Additionally, we synthesized diazide 24 and triyne 25, which were copolymerized in the presence of copper(i) to form a rubbery crosslinked polytriazole material (28) that similarly degraded within minutes at room temperature in 1 M MSA in dichloromethane (Supplementary Fig. 6).

Conclusions
Because bulk polyurethane waste is made up of highly durable thermosets, it is usually incinerated or simply discarded in the environment. With the demand for polyurethane growing unabated, the safe and economically viable management of polyurethane waste will remain a key challenge facing modern society. We developed an approach to this challenge by designing an acetal-containing polyol that is compatible with the current industrially relevant isocyanate-polyol polyurethane chemistry. The water-resistant, crosslinked hydroxyacetal polyurethane demonstrated rapid acid-catalysed degradation in various organic solvents at room temperature, whereas the acetal-free control polyurethane and the hydroxyl-free acetal polyurethane analogues were largely unaffected by the same treatment. Additionally, the recovered polyol from the degraded waste product was successfully repurposed in two practical applications: a polyurethane adhesive with comparable performance to super glue and a photochromic coating. Again, both processes employed commercially available, industrial polyisocyanates as co-reagents. The generality of the hydroxyacetal approach and compatibility with existing polyurethane technologies suggests its application to other polyurethane systems that include isocyanate-free polyurethane chemistries 28,29 , thereby opening a practical pathway to a new generation of sustainable polyurethanes. More broadly, the CATCH cleavage strategy has shown promise in other polymer chain functionalities, allowing for facile breakdown of a variety of polymeric materials. It is encouraging that this degradation mechanism is not limited to polyurethanes and has the potential to be implemented into other classes of polymers, thus increasing the long-term environmental benefits of this approach.

Online content
Any methods, additional references, Nature Portfolio reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/s41557-023-01151-y.

Instrumentation
GPC was performed on a Tosoh EcoSEC 8320 GPC system with 14.5 mM LiBr in N,N-dimethylformamide as the mobile phase. Samples were prepared by dissolving 2 mg polymer in 1 ml mobile phase and filtering the solution with a 0.2 μm nylon syringe filter before injection into the GPC. A refractive index detector with a polymethyl methacrylate standard was utilized to determine the molecular weight (M n and weight-averaged molecular mass (M w )) and dispersity (Ð).
TGA was performed on a Q50 Thermogravimetric Analyzer equipped with a vertical beam balance (sample capacity = 1,000 mg) and a purge gas system. Samples were prepared by cutting small pieces of polyurethane films (5-10 mg). Samples were heated at a rate of 5 °C min −1 from 25-600 °C under a nitrogen atmosphere.
DSC was performed on a TA Instruments Discovery DSC 250 with an RCS90 refrigeration system. 5-10 mg of each sample was placed in a Tzero pan and sealed with a Tzero lid. The samples were cooled to −85 °C then heated at a rate of 10 °C min −1 to 90 °C before being cooled again to −85 °C at the same rate. The temperature was then increased again at a rate of 10 °C min −1 until it reached 90 °C. The glass transition temperature (T g ) of the sample was determined by the maximum value of the derivative of heat flow with respect to temperature. Data were obtained for a minimum of three specimens per material.
Tensile DMA was performed on a TA Instruments RSA III equipped with thin-film grips and a liquid nitrogen cooling system. Samples were cured into rectangle geometries of 40 mm × 10 mm × 1 mm. The gauge length was maintained at 10 mm and dynamic loading was performed at 1 Hz and 0.1% strain amplitude while increasing the temperature linearly at 5 °C min −1 from −60 °C to 25 °C. The glass transition temperature of the samples was determined by the temperature at which the peak of tan δ occurred. Data were obtained for a minimum of three specimens per material. 1 H and 13 C NMR spectra were recorded on a 400 or 500 MHz Varian Unity Inova spectrometer. MestReNova software was utilized to process NMR spectra and the chemical shifts were recorded in ppm. All 1 H and 13 C spectra are reported with residual solvent peaks as a reference. The integration is provided, along with coupling constants (J), which are reported in Hz. Electrospray ionization mass spectra were obtained using electrospray ionization on a Waters Micromass Q-TOF spectrometer and field desorption on a Waters 70-VSE spectrometer. Attenuated total reflection Fourier-transform infrared spectroscopy measurements were performed in a Thermo Nicolet Nexus 670 at room temperature. The infrared spectra were obtained by averaging 16 scans over the 4,000-600 cm −1 range.
For small-molecule degradation studies, 6 ml glass vials were submerged in 1 M acetic acid overnight, then rinsed thoroughly with distilled water and acetone. The vials were dried in the oven for 6 h before use. In a 6 ml vial, a 0.5 mM solution of small-molecule model 3 in CD 3 CN was prepared, along with 0.3 mol equivalents of hexamethylbenzene as an internal standard. 0.6 ml of this solution was transferred to an NMR tube, along with 0.5 mol% TsOH. The reaction was monitored by 1 H NMR until complete degradation of 6 was observed. The percentage of 6 remaining was calculated from integration of the 6 acetal peak relative to integration of the hexamethylbenzene peak.
Lap shear tests were performed on a Parker 081-6079 load frame with a Transducer Techniques 300 lb MDB Series load cell at an extension rate of 1 mm s −1 . Strain within the samples was measured using a virtual extensometer. Samples were prepared according to ASTM D1002 standards. Briefly, 10 mg adhesive mixture was evenly applied on a 12.7 mm × 25.4 mm (L × W) overlap and lap shear tests were reported as the average and standard deviation of three measurements. All equipment was controlled via LabVIEW and a custom program.

Data availability
The data supporting the findings in this study are available within the Supplementary Information. Source data are provided with this paper.