Synthesis, Characterization, and Mechanical Properties of the Vinylogous Urethane Vitrimers

The here-investigated polymers were all based on bis(acetoacetate)-terminated poly(ethyleneglycol) building blocks (aPEG_x; Figure 1a). The subscript x indicates the number-average molecular weight (M_n) of the parent PEG_x. Different aPEG_x grades were prepared from commercially available, hydroxy-terminated polyethylene glycols (PEG_x) with an M_n of 1000 (PEG_1k), 2000 (PEG_2k), 3000 (PEG_3k), 10000 (PEG_10k), or 35000 (PEG_35k) g ⋅ mol⁻¹ in transesterification reactions with tert-butyl acetoacetate (see page S4 of the Supporting Information for details). Adapting a recently reported procedure,¹⁸ the reactions were carried out in the presence of an excess of tert-butyl acetoacetate (20 eq. for PEG_1k, PEG_2k, and PEG_3k; 50 eq. for PEG_10k and PEG_35k). The yield of these reactions ranged from 90 % to 94 %, while the extent of end-group functionalization varied from 92 % to 100 %, as determined by the integration of the signals in the ¹H NMR spectra of the purified aPEG_x (Figures S1–5; calculations on page S4–5). FT-IR spectra of the aPEG_x derivatives reveal the disappearance of the broad band at 3450 cm⁻¹, which is assigned to the −OH end groups of the PEG_x, and the appearance of two peaks at 1716 and 1741 cm⁻¹, which are ascribed to the C=O stretching vibrations of the acetoacetates introduced as chain-ends of the aPEG_x (Figure S9).

Polymer networks comprising VU linkages (aPEG_x-yTREN) were synthesized by reacting the various aPEG_x with tris(2-aminoethyl)amine (TREN). These condensation reactions involve the acetoacetate end groups of the aPEG_x and the amine functions of TREN; they proceed smoothly at 80 °C and afford vinylogous urethanes (VUs) and water as byproduct (Figure 1a). We synthesized several aPEG_x-yTREN grades by varying the M_n of aPEG_x and the molar ratio of amine to acetoacetate functional groups (y), which can be easily controlled by the feed of TREN and aPEG_x (Figure 1a; Table S1). aPEG_x-yTREN grades with y=1.0 (i.e., an equimolar concentration of amine and acetoacetate groups), 1.2 (i.e., a 20 % mol excess of amines), and 1.5 (i.e., a 50 % mol excess of amines) were explored (Figure 1a; Table S1). aPEG_x-yTREN films were prepared by mixing aPEG_x and TREN in DMF, casting the resulting solutions into Teflon molds, heating the mixtures overnight to 80 °C, and subsequently vacuum drying the resulting films for 12 h at 80 °C (details on page S9). The materials thus prepared were analyzed by FT-IR spectroscopy, which reveals the disappearance of the acetoacetates peaks at 1716 and 1741 cm⁻¹, and the appearance of two new vibrational bands at 1604 and 1648 cm⁻¹ (Figures S9, S10b), which indicate the formation of the vinylogous urethane bonds.^{10b, 18} The intensity of these new signals decreases (at constant y) as the M_n of aPEG_x is increased, reflecting that the cross-link density is reduced (Figure S10b). We also investigated the swelling ratio and gel fraction of the aPEG_x-yTREN networks by immersing the films in DMF (Figure S11), in which both aPEG_x and TREN are readily soluble. Within the aPEG_x-1.2TREN series, the swelling ratio increases from 500 % to 1520 %, and the gel fraction decreases from 94 % to 66 %, as the M_n of aPEG_x is increased from 1 to 10 kg ⋅ mol⁻¹ (Figure S11a). Increasing the feed of TREN (i.e., a higher y value) and keeping aPEG_x the same causes a slight decrease of the swelling ratio, while the gel fraction remains practically the same (Figure S11b). These results reflect that the parameters x and y allow one to control the cross-link density and thereby the swelling characteristics of the aPEG_x-yTREN networks. aPEG_2k-1.0TREN and aPEG_35k-1.2TREN fully dissolved in DMF, suggesting that a very low TREN feed or a very high M_n of aPEG_x prevents the formation of cross-linked networks.

The thermal properties of the various PEG_x, aPEG_x, and aPEG_x-yTREN networks were studied by differential scanning calorimetry (DSC) (Figure S12). The DSC heating traces of all polymers, except aPEG_1k-1.2TREN, show endothermic peaks at temperatures between 38 and 61 °C, which are associated with the melting of crystalline domains formed by the PEG segments.²² The melting temperature (T_m) increases with the M_n of aPEG_x (Table S2). For any given value of x, the T_m, and degree of crystallization (χ_c) decrease in the order of PEG>aPEG_x>aPEG_x-yTREN (Figures S12a–e, Table S2). For instance, the T_m, and χ_c values of PEG_2k, aPEG_2k, and aPEG_2k-1.2TREN are 52 °C and 83 %, 41 °C and 70 %, and 38 °C and 38 %, while aPEG_1k-1.2TREN is fully amorphous. The enthalpy of crystallization (ΔH_c) shows the same trend (Table S2). Similar behavior was observed for the other aPEG_x-yTREN series investigated (Figures S12a–e). Increasing the y value resulted in a small decrease of T_m, χ_c, and ΔH_c, as exemplified by the data for the aPEG_2k-yTREN series in Table S2, in which the T_m, χ_c, and ΔH_c dropped from 39 °C, 41 %, and 81 ⋅ J ⋅ g⁻¹ (y=1.0), to 38 °C, 38 %, and 74 J ⋅ g⁻¹ (y=1.2), and 35 °C, 34 %, and 69 J ⋅ g⁻¹ (y=1.5).

We next studied the thermomechanical properties of aPEG_x-yTREN films by means of dynamic mechanical analysis (DMA) (Figures 1b, 1d). All DMA traces of aPEG_x-1.2TREN (x≥2k) feature a slanted plateau regime in which the storage modulus E′ adopts values between 1110 and 570 MPa. A gradual reduction of E′ with temperature can be observed, and the E′ trace shifts to higher values as the M_n and χ_c are increased. The DMA traces all show a sharp drop of E′ at a temperature that marks the melting of the crystalline PEG domains. The corresponding maxima of the tan delta (tan(δ)) plots are in good agreement with the melting temperatures observed by DSC (see above). In the case of aPEG₂-1.2TREN, aPEG₃-1.2TREN, and aPEG₁₀-1.2TREN, the DMA traces show a rubbery plateau, confirming the formation of cross-linked architectures. In this regime, aPEG_2k-1.2TREN, which features the highest cross-link density, displays the highest E′, whereas no differences can be observed for aPEG_3k-1.2TREN, and aPEG_10k-1.2TREN. By contrast, aPEG_35k-1.2TREN fails upon melting, reflecting that the cross-link density in this material is very low. The DMA trace of aPEG_1k-1.2TREN shows no obvious thermal transitions in the temperature range explored (−20 °C to 100 °C; Figure S13a), consistent with the absence of crystalline PEG domains, as evidenced by DSC data (Figure S12a). Figure 1d, which shows the DMA traces of the aPEG_2k-yTREN series, confirms that changing the TREN content has only a minor effect on T_m, as reflected by the DSC data. However, a pronounced effect is observed above T_m, where the increased cross-link density bestows aPEG_2k-1.5TREN with a higher E′ (1.5 MPa) than aPEG_2k-1.2TREN (0.9 MPa), and the absence of cross-links causes aPEG_2k-1.0TREN to fail upon reaching T_m.

The mechanical properties of films of the aPEG_x-1.2TREN series were further investigated by tensile testing (Figure 1c; Figure S14, Table S3). The stress-strain curves of aPEG_2k-1.2TREN, aPEG_3k-1.2TREN, and aPEG_10k-1.2TREN all display the typical behavior of semicrystalline polymers, with an initial elastic regime, yield points at 30–90 % strain, and yielding deformation with modest or no strain hardening (Figure S14b). The yield stress and stress at break increase with the M_n of aPEG_x from 6.1 MPa and 7.0 MPa for aPEG_2k-1.2TREN to 15.5 MPa and 19.3 MPa for aPEG_10k-1.2TREN, and all three polymers show plastic deformation with a strain at break of more than 1800 % (Figure 1c; Table S3; Supplementary Video S1). In stark contrast, aPEG_1k-1.2TREN, in which the PEG segments do not crystallize, is an elastic material and displays much lower stress (0.9 MPa) and strain at break (98 %) values (Figure S13b). On the other hand, aPEG_35k-1.2TREN is highly crystalline and shows brittle failure at a maximum stress of 7.5 MPa and low strain at break (7.5 %). The influence of the TREN content on the tensile properties is evident from the data shown in Figure 1e. aPEG_2k-1.5TREN displays tensile properties that are very similar to those of aPEG_2k-1.2TREN, but the increased cross-link density leads to a higher stress at break (Figure 1e; Figure S15, Table S4). Somewhat surprisingly, aPEG_2k-1.0TREN shows brittle failure at a strain of 30 % and a stress of 9.3 MPa, which is higher than the yield stress observed for aPEG_2k-1.2TREN and aPEG_2k-1.5TREN (Figure S15b). We speculate that the low strain and high stress at break for aPEG_2k-1.0TREN can be attributed to the low cross-link density induced by the low TREN feed in the synthesis and the high χ_c, respectively (Table S2).

Next, we tested the stability of the aPEG_x-yTREN materials to oxidative conditions (for amino groups, i.e., high T and presence of air). Films of aPEG_10k-1.2TREN were placed in an oven at 80 °C with constant air flow for 2 days. The aPEG_10k-1.2TREN films thus treated were then subjected to tensile tests. The stress-strain curves recorded reveal a slight reduction in both stress and strain at break (21.1 MPa, 1180 %) compared to the original aPEG_10k-1.2TREN (24.0 MPa, 1380 %) (Figure S16), which we attribute to the oxidation of (some) amino groups present in the material.

Because the aPEG_x-yTREN VUs are based on hydrophilic PEG building blocks, their mechanical properties can be expected to depend on the relative humidity (RH) of the environment. To explore this, films of aPEG_1k-1.2TREN and aPEG_10k-1.2TREN were conditioned at 85 % RH for either 1 or 3 days. Subsequent tensile tests (Figure S17) show that the properties of aPEG_10k-1.2TREN are hardly affected, whereas the stress and strain at break of aPEG_1k-1.2TREN films drop from 0.9 MPa and 98 % to 0.6 MPa and 43 % after 1 day at 85 % RH and to 0.4 MPa and 29 % after 3 days at 85 % RH (Figure S17a). The different susceptibility to humidity is attributed to the crystalline nature of aPEG_10k-1.2TREN (Figure S12d), which limits the uptake of moisture.²³

Since responsiveness to moisture is normally not a desirable feature in polymeric materials, we also investigated aPTHF_2k-1.2TREN (Figure 2), which is based on a hydrophobic building block made by end-capping poly(tetrahydrofuran) with an M_n of 2000 g ⋅ mol⁻¹ with bisacetoacetate (aPTHF_2k) and TREN (synthetic details on pages S5 and S10) to understand whether the moisture sensitivity could be suppressed. Gratifyingly, aPTHF_2k-1.2TREN exhibits excellent stability in the presence of various organic solvents (such as THF, CHCl₃, ethanol, DMF, and hexane), water, THF/H₂O mixtures, and aqueous NaOH after 3 days at room temperature (Figure S18). More importantly, the mechanical properties of aPTHF_2k-1.2TREN show no appreciable changes after 1 or 3 days of exposure to 85 % RH (Figure S19). Thus, the data show clearly that moisture sensitivity can be controlled via the crystallizability and polarity of the VU-terminated building block.

Depolymerization and Recovery of the Monomers

As reflected by Figure 1a, the formation of vinylogous urethanes is a reversible process in which the position of the equilibrium can be shifted by the addition or removal of water. To test if the addition of water would lead to depolymerization of the VU-based polymers, we first investigated the hydrolysis of vinylogous urethane bonds using bis-(butyl vinylogous urethane)-terminated PEG (aPEG_2k-Btl; Figure 3a), a model compound made by reacting aPEG_2k and butylamine (Btl; page S19). The conversion of aPEG_2k-Btl into aPEG_2k and Btl was carried out by immersing 20 mg of aPEG_2k-Btl in 700 μL of deuterated water (D₂O) at 25 °C, 40 °C, and 60 °C (Figure 3a), and the progress of the reactions was monitored by following the evolution of the ¹H NMR signals assigned in Figure 3a over time (Figure 3b; Figures S21b–d). Integration of the ¹H NMR signals allowed us to quantify the degree of hydrolysis (Figure S21e). At 25 °C, the extent of hydrolysis of aPEG_2k-Btl was ≈13 % after 30 min, increased to ≈36 % after 120 min, and reached a value of ≈88 % after 9 h (Figure 3b). When the reaction was carried out at 60 °C, ca. 93 % of aPEG_2k-Btl were hydrolyzed within 1 h. Monitoring the consumption of aPEG_2k-Btl in the presence of a large excess of D₂O as a function of time and temperature (Figure S21e) allowed us to determine the pseudo-first-order kinetic rate constants (k) of the reaction (Figure S23, Table S5). An Arrhenius plot of ln(k) against 1/T affords an activation energy (E_a) of 61 kJ mol⁻¹ for the process (Figure 3c).

We also investigated the influence of the water content on the hydrolysis of aPEG_2k-Btl by mixing 20 mg of aPEG_2k-Btl and different volumes of D₂O (200, 300, 500, 700 μL) at 25 °C (Figure S22a). The extent of hydrolysis was monitored by analyzing the ¹H NMR signals (Figures S22b–f). As the amount of D₂O was increased from 200 to 700 μL, the extent of hydrolysis for aPEG_2k-Btl increased from 17 % to 36 % after 120 min, and from 76 % to 88 % after 9 h (Figure S22f). The results suggest that a large excess of D₂O is required besides high temperature to drive the hydrolysis process in VU-based vitrimers containing hydrophilic components.

Next, we studied the hydrolysis of aPEG_2k-1.2TREN in H₂O by immersing aPEG_2k-1.2TREN films in a ten-fold amount (in weight) of neutral H₂O at 25 °C (Figure 3d). The polymer initially swells and then gradually dissolves over time (Figure S24a). The pH value of the water phase increases from ca. 8.4 to 9.5 over the course of 2.5 hours, which reflects an increasing concentration of free amines (Figure S24b). The increase in pH is accompanied by a progressive dissolution of the polymer, which is consistent with the hydrolysis of the VU linkages in the polymer network. We carried out kinetic measurements for the hydrolysis of aPEG_2k-1.2TREN at temperatures between 20 and 80 °C in D₂O (further details in the Supporting Information on page S25). The progress of these reactions was again monitored by ¹H NMR spectroscopy, and the integration of the signals was used to establish kinetic profiles for these processes (Figure 3e; ¹H NMR spectra are shown in Figure S25). Note that the NMR technique only allows the detection of species produced by the hydrolysis that are soluble in D₂O, while the initial aPEG_2k-1.2TREN is NMR-silent due to its cross-linked nature. Thus, the extent of depolymerization was determined by the relative depolymerization degree (RDD), in which m_t is the ratio of the integrals of signal m (ascribed to the PEG backbone of aPEG_2k, see Figure 3d) and n (assigned to water in the NMR spectra of Figure S25) at time t, and m₀ is the initial ratio of the integrals of signals m and n in the ¹H NMR spectra of the mixtures (Eq. 1.

(1)

A comparison of the RDD determined at different temperatures (Figure 3e) shows that, also in this case, higher temperatures result in faster hydrolysis, mirroring the trend observed for aPEG_2k-Btl. For instance, the RDD is 24 % after 42 minutes at 20 °C, while it increases to 57 % after 15 min at 80 °C (Figure 3e).

We further investigated how the composition of aPEG_x-yTREN influences the water-promoted depolymerization. The first set of experiments was conducted with the aPEG_x-1.2TREN series at room temperature (page S27). All of the aPEG_x-1.2TREN/H₂O mixtures (1 : 15 wt : wt) were initially heterogeneous, and eventually turned into clear solutions (Figure S28). We also investigated the influence of the TREN feed on the depolymerization of aPEG_2k-yTREN, and the comparison shows that the time to full dissolution increases with the TREN content. The data suggest that the kinetics of the water-assisted depolymerization of the aPEG_x-yTREN can be engineered to occur within a broad time range (Figure S29), although we emphasize that the experiments reported here ultimately probe the dissolution and not the complete depolymerization of the VU networks.

We next pursued the recovery of aPEG_x (x<35 k) or aPTHF_2k (Figures S26–28). Our strategy to separate aPEG_x (x≤35k) and TREN was inspired by previous work from the Helms lab^10a and leveraged the difference in the chemical properties of these two components (page S30). In order to remove TREN, the clear solutions produced by depolymerization for 24 h at room temperature in water (conditions discussed above) were treated with a strongly acidic ion exchange resin, which was subsequently filtered off. Evaporation of the solvent (H₂O) from the filtrate afforded aPEG_x in a yield of 97 %–99 % (page S27). The low amounts of TREN used in the preparation of the VU vitrimers (<150 mg; Table S1) made the quantitative recovery of this component particularly challenging. Thus, to show the feasibility of the applied procedure, we carried out a model experiment in which the acidic ion exchange resin was impregnated with 0.5 g of TREN and further treated with an excess of trimethylamine. Filtration of the thus treated resins afforded the recovery of TREN in a yield of 88 % (Figure S27). The purity of the recovered aPEG_x and TREN was assessed by ¹H NMR spectroscopy, which confirmed the chemical structures and suggested the absence of (detectable) impurities (Figure 3f; S30–35).

As discussed above, aPTHF_2k-1.2TREN remains mechanically stable in neutral or basic water, even after being immersed for 3 days, and therefore this hydrophobic VU vitrimer requires different depolymerization conditions than aPEG_x-yTREN. Gratifyingly, aPTHF_2k-1.2TREN could be degraded by treatment with a 1 M HCl aqueous solution at room temperature. The degradation process occurred within 3 days, after which the initially clear liquid phase had turned into an off-white slurry. Extraction with CHCl₃, neutralization of the organic phase through washing with water, and removal of the solvent allowed the recovery of aPTHF_2k in high yield (95 %, page S28) and in high purity (Figure S26b).

We explored the recyclability of a (nano)filler-reinforced composite material that was prepared by combining multi-walled carbon nanotubes (MCNs) with aPEG_10k-1.2TREN (aPEG_10k-1.2TREN/MCNs) (synthesis on page S38). Immersing aPEG_10k-1.2TREN/MCNs in H₂O for 24 h afforded a suspension, from which the MCNs could be filtered off, leaving a clear solution containing aPEG_10k and TREN (Figure S36). Further treatment with the acidic ion exchange resin permitted the almost quantitative recovery (96 % of the initial mass) of aPEG_10k (Figure S36). Scanning electron microscopy (SEM) imaging and Raman spectroscopy provided qualitative indications of the unperturbed morphology of the recovered MCNs (Figure S37).

Finally, we tested the possibility to apply the depolymerization/separation sequence to VU networks in complex mixtures that simulated a mixed waste stream (Figure S38). The experiment was conducted by mixing aPEG_10k-1.2TREN with household waste (polyethylene, polypropylene, aluminum foil, tree bark, and leaves) or commodity polymer waste (polyvinyl chloride, polyethylene terephthalate, polycarbonate, and polystyrene). The mixtures were treated with water (1500 weight %) for 24 h at room temperature, and the solid components were removed via filtration. After treatment of the obtained solutions with the acidic ion exchange resin, filtration, and water evaporation, aPEG_10k was recovered with a negligible loss (ca. 97 % of material recovered) and in high purity, as confirmed by its ¹H NMR spectrum (Figures S38–39).

Thermal Reprocessing and Water-Assisted Recycling of the Vinylogous Urethane Vitrimers

As previously demonstrated by Du Prez et al., VU linkages can undergo transamination processes in the presence of auxiliary amines at sufficiently high temperature (60–150 °C) (Figure 4a, notice the exchange between the blue and green amines).^{19b, 20a, 24} Such transamination offers two paths to recycle VU vitrimers. The first one involves the thermal reprocessing at 120–150 °C,^{10b, 10c} while the second one relies on the depolymerization by treatment with monofunctional amines at 60–120 °C, followed by repolymerization with a fresh amine cross-linker.^{24a, 24c-24f} In addition to these established methods, we hypothesized that the water-mediated reversible (de)polymerization of VU-based polymer materials reported here could serve as a complementary reprocessing strategy. Thus, we attempted the reprocessing of the aPEG_2k-yTREN series by either activating the dynamic transamination process at 150 °C (Figure 4a, left), or by exploring the water-dependent depolymerization-polymerization sequence (Figure 4a, right).

The dynamic character of the transamination of the aPEG_x-yTREN systems was initially investigated in solution with model compound aPEG_2k-Btl and benzylamine (page S44). The two compounds were mixed in deuterated DMSO (DMSO-d₆) in a 1 : 6 aPEG_2k-Btl:benzylamine molar ratio, and the resulting mixtures were heated at 80 °C, 100 °C, and 120 °C (Figure S40). The progress of the exchange reaction was monitored by integrating the ¹H NMR signals of the VU linkages comprising the two amine substituents (Figure S40), which also allowed us to establish the kinetic profiles at each reaction temperature (Figure S41; Table S6), and determine the E_a of the transamination process, which is very low (35 kJ mol⁻¹) (Figure S42).

We then probed if the transamination process would allow the aPEG_2k-yTREN films to flow by performing stress relaxation experiments at different temperatures above the T_m of the PEG domains (Figure S43). No such experiments could be carried out for aPEG_2k-1.0TREN, which features a low cross-link density and flows upon melting (Figures 1d, S15). By contrast, the stress-relaxation curves of aPEG_2k-1.2TREN and aPEG_2k-1.5TREN reveal relaxations processes that we analyzed using the Maxwell model for stress-relaxation (Figures S43b, S43d, page S46). A comparison of the stress relaxation curves shows that, under the same conditions, stresses are dissipated more rapidly in aPEG_2k-1.5TREN than in aPEG_2k-1.2TREN. For instance, at 130 °C, the relaxation time (τ), at which the stress is dissipated to 1/e of its initial value, is 4.5 min for aPEG_2k-1.5TREN and 50 min for aPEG_2k-1.2TREN (Figures S43a, S43c). We attribute such difference in stress relaxation efficiency to a larger amount of free amines available in the aPEG_2k-1.5TREN network, which is tantamount to stating that a higher loading of free amines provides more efficient transamination processes.^{10c, 25} Linear fits of ln(τ) versus 1/T afforded E_a values of 81.4 kJ mol⁻¹ for aPEG_2k-1.2TREN and 127.1 kJ mol⁻¹ for aPEG_2k-1.5TREN for the topological rearrangement, which comprises the transamination and other processes, of the polymer networks (Table S7). Moreover, by fitting the data to a Maxwell equation and assuming a Poisson's ratio of 0.5,²⁶ we could identify the theoretical topology-freezing temperature (T_v) (calculations on page S47, Table S7), above which the exchange reaction in the polymer material becomes active. The T_v of aPEG_2k-1.2TREN is 66 °C, while a value of 40 °C was determined for aPEG_2k-1.5TREN.

We next investigated the possibility to thermally reprocess aPEG_2k-1.2TREN. A film of aPEG_2k-1.2TREN was cut into small pieces, which were then compression-molded at 150 °C and a pressure of 10 MPa for 30 min. This treatment afforded a homogeneous film (Figure S44). The process was repeated and the mechanical properties of the original aPEG_2k-1.2TREN film and the samples made in the first and second recycling step were compared by DMA (Figure 4b) and tensile testing (Figure 4c). Gratifyingly, the mechanical properties of the three samples are almost indistinguishable. The DMA traces of the thermally recycled aPEG_2k-1.2TREN show a slight increase of E′ in the rubbery regime (0.9 MPa for the recycled vs 0.5 MPa for the original aPEG_2k-1.2TREN), whereas the tensile tests reveal an increase in the stress at break (9.7 MPa for the recycled vs 7.0 MPa for the original aPEG_2k-1.2TREN) and a decrease in the strain at break (2020 % for the thermally recycled vs 2630 % for the original aPEG_2k-1.2TREN). These differences might be related to some unreacted amines and acetoacetate moieties in the as-prepared material, which was produced by solution casting, which were converted into VU bonds via thermal reprocessing. This seems to be supported by the lack of significant differences observed in the DMA traces and stress-strain curves of the 1^st and 2^nd thermally recycled films (Figures 4b–c), which suggests similar network architectures and density of cross-links. The hydrolysis of the VU linkages upon reprocessing could be ruled out, as this would have caused a decrease in E′ due to lowering of the density of cross-links.

Finally, we conducted similar study experiments and compared the mechanical properties of the original aPEG_2k-1.2TREN to those of an aPEG_2k-1.2TREN that was produced from depolymerized matter. Briefly, we treated aPEG_2k-1.2TREN with a 5-fold excess of deionized H₂O at 25 °C, which caused hydrolysis of the VU linkages and the solubilization of aPEG_2k and TREN (Figure S28). The solution thus produced was subsequently cast into a Teflon mold, heated to 80 °C overnight, and finally dried in vacuum at 80 °C for 12 h. The DMA and stress-strain profiles obtained by measuring the original and re-synthesized (H₂O-recycled) polymer show that the two materials display practically identical mechanical properties (Figures 4d–e), demonstrating the effectiveness of water-assisted recycling of aPEG_2k-1.2TREN.