Solid State Alkaline Depolymerization of Polyester Elastane Textiles in a Laboratory Kneader

Abstract

Elastane is ubiquitous in polyester-based textiles and complicates depolymerization-based recycling because it can undergo thermal degradation and chemical bond cleavage, consuming reagents and forming low-molecular by-products that may compromise monomer quality. Here, we investigate alkaline PET depolymerization of PET/elastane blends under an intentional base-competition scenario in a laboratory kneader. Pure PET (100/0) and PET/EL blends (95/5 and 85/15, wt/wt) were processed under quasi-solid-state conditions at 140 °C for 5 min using solid NaOH dosed at 2.1 mol per mol PET repeat unit and pelletized feedstocks to ensure scale-relevant mixing and reproducible chamber filling. Torque and bulk-temperature profiles were similar across compositions, and isolated terephthalic acid yields remained in a narrow corridor (68–71%), indicating that PET depolymerization is not measurably impaired by 5–15 wt% elastane within this reaction window. Differential scanning calorimetry of water-insoluble residues revealed pronounced changes in elastane-related thermal transitions, evidencing elastane modification during treatment. Targeted ¹H NMR screening of recovered TA against a 4,4′-methylenedianiline spiked reference showed no detectable co-isolated aromatic diamines. Overall, the study demonstrates robust monomer recovery from mixed PET/EL textiles under solid-NaOH, short-residence, solvent-lean processing, while identifying residue analytics as the key bottleneck for quantifying elastane fate and closing component balances.

1. Introduction

Global fiber production reached approximately 132 million tons in 2024 and is projected to exceed 160 million tons by 2030 [1]. Polyester (polyethylene terephthalate, PET) accounts for more than half of this output, making it the dominant synthetic textile polymer [1]. In contrast, elastane (EL, also known as spandex or polyurethane fiber) represents only about 1.2% of global fiber production, yet it is incorporated in a large fraction of garments due to its elasticity, typically between 3 wt% and 15 wt%. As a result, elastane appears in most textile waste streams and poses a persistent obstacle to recycling. Even small elastane fractions can compromise both mechanical and chemical recycling. At typical PET melt processing temperatures, elastane is prone to thermal degradation [2]. Under chemical recycling conditions, elastane linkages can undergo bond cleavage and generate low-molecular products, including aromatic diamines, polyether diols, and chain-extender-related fragments [3]. In addition, elastane is frequently present in core spun yarns where PET forms the outer sheath. Near-infrared (NIR) sensors mainly probe the textile surface and may therefore misclassify PET/EL textiles as pure PET. Consequently, low elastane contents cannot be reliably detected by NIR-based sorting [4]. This motivates the central premise of this work. Elastane degradation must be expected during PET depolymerization, and recycling processes must remain robust when elastane-derived reactions can compete for reactive agents and influence product purity.

Among available strategies to close the textile material loop, chemical recycling of polyester textiles offers the potential to recover high-purity monomers for fiber-grade repolymerization. Various depolymerization routes have been investigated for PET [5]. Among them, alkaline hydrolysis is particularly robust toward additives and mixed feedstocks, yielding terephthalic acid (TA) and ethylene glycol (EG) under moderate conditions [6]. Although PET depolymerization mechanisms in alkaline media are well understood, the behavior of elastane under such conditions has received little systematic investigation.

Recent studies investigated PET elastane blended textiles across multiple chemical depolymerization routes, including solvent-assisted alkaline hydrolysis, glycolysis, and methanolysis. In solvent mediated alkaline systems, Zhang et al. reported selective PET hydrolysis with potassium hydroxide (KOH) using tetrahydrofuran (THF) as cosolvent at 60 °C for 3 h achieving 97.5% isolated TA yield and filtration based recovery of intact elastane, and a second KOH approach in dichloromethane (DCM) ethanol solvent system at 35 °C achieving PET hydrolysis within 30 min with elastane remaining intact, while degradation at higher temperature was noted [7,8]. In contrast, glycolysis of PET elastane blends at 200 °C for 3 h with Potassium carbonate (K₂CO₃) was reported to induce elastane co-depolymerization [9]. Tanaka et al. reported methanolysis using dimethyl carbonate (DMC) with NaOMe at 50 °C for 2 h, yielding complete PET conversion to dimethyl terephthalate while polyurethane segments remained identifiable and separable as a solid fraction [10]. Overall, selective PET depolymerization from PET elastane textiles is feasible, but systematic insight into elastane fate and residue quantification under solvent-lean, solid NaOH alkaline processing at elevated temperature and short residence time remains limited.

Elastane is a segmented polyurethane consisting of alternating soft segments and hard segments. The soft segment domains, typically based on polyether or polyester diols, provide elasticity. In contrast, the hard segment domains consist of a rigid framework formed by aromatic diisocyanates, which are connected to chain extenders and the soft segments through urea and urethane linkages [11]. This network is further stabilized by strong hydrogen bonding. For the commonly used polyether-based elastane, the ether soft segments are less susceptible to alkaline attack [11]. Consequently, the urethane and urea linkages within the hard segments represent the primary sites of susceptibility to hydrolysis. Under alkaline conditions, the cleavage of these bonds may lead to the formation of diamines and polyols, resulting in a loss of structural integrity and network breakdown (Figure 1) [3,12,13]. Because elastane can undergo thermal degradation before melting and bond cleavage under alkaline depolymerization conditions, its structure may change substantially under recycling-relevant conditions. Such changes are expected to alter its characteristic thermal transitions, including the soft-segment melting endotherm used in differential scanning calorimetry (DSC) [14]. Figure 1 depicts the stoichiometric alkaline depolymerization of PET and a non-exhaustive schematic of expected product classes arising from alkaline cleavage of MDI-based polyether elastane.

Understanding these transformations is crucial for assessing the robustness of chemical recycling processes toward mixed-fiber textiles. Since alkaline reagent dosing is typically calculated stoichiometrically relative to the PET fraction, variations in feedstock composition could potentially affect the reaction environment through competitive consumption of the base by the elastane hard segments. Therefore, investigating different PET/EL ratios is essential to determine whether PET depolymerization remains robust under base competition and still delivers high monomer yields and purity across varying material compositions.

Based on these considerations, the aim of this work is to assess whether alkaline PET depolymerization remains robust when elastane can compete for base and to clarify the chemical and structural fate of elastane under the applied conditions. In our previous study using shredded PET and PET/EL fibers, EG and NaOH were charged as fixed masses, and TA quality was primarily assessed via color index. Since NaOH dosing was based on total sample mass, the effective NaOH excess relative to PET increased with higher elastane contents and thus did not represent a stringent base competition scenario. Here, pelletized feedstocks enable higher bulk density and a defined chamber filling degree, which improves loading reproducibility and reduces the relative impact of handling losses on isolated TA yields. Solid NaOH is dosed at a fixed stoichiometry relative to the PET fraction (2.1 mol per mol PET repeat unit), and EG is added as a lubricant at 25 wt% relative to pellet mass to enforce comparable conditions across compositions. Product assessment is extended by ¹H NMR analysis of recovered TA to test for aromatic diamine signals and by DSC of water-insoluble residues to evidence elastane modification and to assess limits of DSC-based residue quantification after depolymerization.

2. Materials and Methods

Depolymerization experiments were conducted in a laboratory kneader to investigate the behavior of elastane during alkaline PET depolymerization under quasi-solid-state conditions. The following sections describe the preparation of PET/EL materials, the depolymerization procedure, and the analytical methods applied to characterize both the recovered TA and the WIR. Emphasis was placed on reproducible reaction conditions and consistent analytical evaluation across materials of different compositions and bulk densities.

2.1. Material Preparation and Characterization

Two post-production PET/EL textiles containing 0.05 and 0.15 ${\mathrm{g}}_{\mathrm{E}\mathrm{L}}{\mathrm{g}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}^{-1}$ (corresponding to 5 and 15 wt%) were supplied by Sitip S.p.A, Cene, Italy. Pure post-production polyester samples were supplied by Hero Textile AG, Crailsheim, Germany. The textiles were manually cut from the roll and subsequently shredded using a cutting mill (SM300, Retsch GmbH, Haan, Germany) equipped with a V-rotor operating at 1500 min⁻¹ and a 3 mm sieve. The shredded textile flock was then pelletized using a 14–175 pellet press (Amandus Kahl GmbH and Co. KG, Reinbek, Germany) to produce cylindrical pellets with a diameter of 3 mm and a length of approximately 10 mm.

The bulk density of the pellets was determined based on DIN EN ISO 60 using a measuring cylinder and a precision balance (Highland Precision Balance HCB 2202, Adam Equipment Co. Ltd., Milton Keynes, UK) [15]. Bulk densities of 556 kg m⁻³ for 95/5 PET/EL, 455 kg m⁻³ for 85/15 PET/EL and 598 kg m⁻³ for pure PET were obtained.

2.2. Depolymerization Experiments

The alkaline depolymerization experiments were carried out in a Haake Polylab Rheomix 600 laboratory kneader (Thermo Fisher Scientific Inc., Waltham, MA, USA). The kneader consists of three electrically heated zones with an internal mixing chamber. Two roller rotors operate in the chamber at a speed ratio of 3:2, with the faster rotor set to 40 rpm. The free chamber volume is 69 cm³. Mass temperature at the center bottom of the mixing chamber and the torque acting on the rotors were recorded continuously. The chamber was preheated to 140 °C prior to each experiment. The experiments for all materials were conducted in five replicates. The experimental setup is shown in Figure 2.

Before the experiments, the textile pellets and solid NaOH pearls (≥99%, Carl Roth GmbH and Co. KG, Karlsruhe, Germany) were weighed in a beaker, mixed, and rapidly transferred into the mixing chamber immediately after the addition of EG. The pellet mass was adjusted to achieve a volumetric chamber filling degree of 70% at the start of the reaction and therefore differed between the investigated materials. Pellet mass was calculated according to Equation (1). A filling degree of 70% was selected based on previous studies on mixing behavior in a laboratory kneader, although those experiments were performed with molten materials [16]. The NaOH mass corresponded to a molar ratio of 2.1 mol NaOH per mol PET in the pellets. EG was added prior to the reaction as a lubricant at 0.25 ${\mathrm{g}}_{\mathrm{E}\mathrm{G}}$ ${\mathrm{g}}_{\mathrm{pellets}}^{-1}$.

$${\mathrm{m}}_{\mathrm{P}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{t}}={\displaystyle \frac{\mathrm{V}·0.7}{\frac{1}{{\mathrm{\rho }}_{\mathrm{s},\mathrm{p}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{t}}}+\left[\frac{{\mathrm{\omega }}_{\mathrm{P}\mathrm{E}\mathrm{T}}}{{\mathrm{M}}_{\mathrm{P}\mathrm{E}\mathrm{T}}}·\mathrm{\upsilon }·\frac{{\mathrm{M}}_{\mathrm{N}\mathrm{a}\mathrm{O}\mathrm{H}}}{{\mathrm{\rho }}_{\mathrm{s},\mathrm{N}\mathrm{a}\mathrm{O}\mathrm{H}}}\right]}}$$

(1)

where ${\mathrm{m}}_{\mathrm{P}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{t}}$ is the required pellet mass in kg to reach 70% chamber filling degree, V is the free chamber volume in m³, ${\mathrm{\rho }}_{\mathrm{s},\mathrm{pellet}}$ and ${\mathrm{\rho }}_{\mathrm{s},\mathrm{NaOH}}$ are the bulk densities of the textile pellets and NaOH pearls in kg m⁻³, ${\mathrm{\omega }}_{\mathrm{PET}}$ is the PET mass fraction in the pellets, ${\mathrm{M}}_{\mathrm{PET}}$ and ${\mathrm{M}}_{\mathrm{NaOH}}$ are the molar masses in kg kmol⁻¹ and $\mathrm{\upsilon }$ is the molar ratio of NaOH to PET.

After filling, the mixing chamber was sealed with a pneumatic ram, and the reaction medium was mixed for 5 min inside the chamber. Data acquisition of mass temperature and torque was automatically initiated when the torque exceeded 0.2 Nm.

2.3. Work-Up and Product Separation

After the reaction, the mixing chamber of the laboratory kneader was opened, and the mixture was transferred into a beaker with deionized water added to a total volume of 400 mL. Water-soluble components fully dissolved after 1 h of stirring. The resulting suspension was filtered using a paper filter with a pore size of 4–12 µm (MN 615, MACHEREY-NAGEL GmbH and Co. KG, Düren, Germany), and the resulting filter cake, containing WIR, was washed with approximately 50 mL deionized water to remove water-soluble by-products and residual monomers. The filter cake was dried at 80 °C for 48 h.

An aliquot of the filtrate was taken, and TA was precipitated by adding sulfuric acid (≥95% p.a., Thermo Fisher Scientific Inc., Waltham, MA, USA). The precipitated TA was filtered using a Büchner funnel with a pore size of 10–16 µm, washed with approximately 50 mL deionized water, and dried at 80 °C for 48 h. After drying, the mass of the recovered TA was measured, and the yield was calculated according to Equation (2).

$${\mathrm{Y}}_{\mathrm{T}\mathrm{A}}={\displaystyle \frac{{\mathrm{m}}_{\mathrm{T}\mathrm{A}}·{\mathrm{M}}_{\mathrm{P}\mathrm{E}\mathrm{T}}}{{\mathrm{m}}_{\mathrm{p}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{t}}·{\mathrm{M}}_{\mathrm{T}\mathrm{A}}·{\mathrm{\omega }}_{\mathrm{P}\mathrm{E}\mathrm{T}}}}·{\displaystyle \frac{{\mathrm{m}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}}{{\mathrm{m}}_{\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{q}\mathrm{u}\mathrm{o}\mathrm{t}}}}$$

(2)

where ${\mathrm{Y}}_{\mathrm{TA}}$ is the TA-yield in $\mathrm{m}\mathrm{o}{\mathrm{l}}_{\mathrm{T}\mathrm{A}}$ ${\mathrm{mol}}_{\mathrm{PET}}^{-1}$, ${\mathrm{m}}_{\mathrm{TA}}$ is the mass of recovered TA in kg, ${\mathrm{m}}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ is the weight of the total product solution and ${\mathrm{m}}_{\mathrm{a}\mathrm{l}\mathrm{i}\mathrm{q}\mathrm{u}\mathrm{o}\mathrm{t}}$ is the weight of the aliquot part in kg and the remaining variables are defined as above.

2.4. Cryogenic Grinding of WIR

To ensure homogeneity for subsequent DSC analyses, the dried WIR from depolymerization experiments and reference materials were ground using a CryoMill (Retsch GmbH, Haan, Germany) with a 50 mL grinding jar and a 20 mm steel ball, under continuous cooling with liquid nitrogen (−197 °C). For the preparation of DSC calibration standards, 100% EL-fibers (The LYCRA Company LLC, Wilmington, NC, USA) and 100% PET-fabric (Carl Weiske GmbH and Co. KG, Hof, Germany) were used. The grinding process started with a pre-cooling step and comprised four grinding intervals, each separated by an intermediate step. The machine parameters for each step are summarized in

Table 1. Cryogenic grinding parameters for residues and DSC calibration materials.

Step	Interval	Frequency
Pre-cooling	2 min (100% PET reference) 2.5 min (100% EL reference) 2.5 min (WIR)	5 Hz
Grinding	4 × 2.5 min	25 Hz
Intermediate	1 min	2 Hz

.

2.5. DSC Measurements

For optimal comparability, all samples were weighed to a target mass of 5.00 mg ± 0.01 mg. Weighing was performed using a microbalance (XPR2, Mettler Toledo, Giessen, Germany; readability 1 μg) in 40 μL aluminum pans without a pin, which were then hermetically sealed using a pan press. DSC measurements of each material were performed in at least three replicates. The samples were analyzed under a nitrogen flow of 50 mL min⁻¹ using a DSC3+ (Mettler Toledo, Giessen, Germany) equipped with an FRS6+ sensor and a TC100 cooling system (Peter Huber Kältemaschinenbau SE, Offenburg, Germany). The instrument was calibrated for temperature and enthalpy using indium as a calibration standard. The temperature program, adapted from Boschmeier et al., covered a range from −80 °C to 280 °C following the sequence in Equation (3) [14].

$$25 °\mathrm{C}\to -80 °\mathrm{C}\to 280 °\mathrm{C}\to -80 °\mathrm{C}\to 280 °\mathrm{C} \mathrm{with} 10 \mathrm{K} {\min }^{-1}$$

(3)

In contrast to Boschmeier et al., the program started with a cooling step to −80 °C to capture the glass transition of the elastane soft segments, indicative of their chemical structure [17]. For undegraded PET/EL reference mixtures, the elastane mass fraction was estimated according to Boschmeier et al. from the mass-normalized melting enthalpy of the soft-segment melting endotherm (J g⁻¹) in the second heating run. Prior to analyzing residues from the depolymerization experiments, a calibration curve for elastane was established using cryogenically ground pure components. The powdered reference materials were weighed directly into the DSC pans in the desired mass ratios. The calibration curve is discussed in Section 3.2.1 together with the calibration reported by Boschmeier et al. In addition, the 95/5 and 85/15 material was measured without prior cryogenic grinding.

For WIR after depolymerization, this approach was not assumed to be quantitatively transferable. Residue modification was assessed from qualitative thermogram changes. In addition, the virgin-material calibration was applied to WIR to evaluate the transferability of the peak-area-based quantification approach. Back-calculated values are reported as an apparent elastane fraction and are not interpreted as quantitative composition.

2.6. NMR-Spectroscopy

NMR measurements of the recovered TA were performed on a Bruker AVIII 400 spectrometer (Bruker Corp., Billerica, MA, USA). Approximately 17 mg of the precipitated TA were placed in an NMR tube (SP Industries Inc., Vineland, NJ, USA) and dissolved in 0.66 mL of d₆-DMSO using ultrasonication. Chemical shifts (δ) are reported in ppm relative to tetramethylsilane (TMS, Me₄Si) as the internal standard, and coupling constants (J) are given in Hz. The spectra were recorded to verify the chemical purity of the recovered TA.

2.7. Triangulation

To evaluate the overall mass balance and correlate the solid and soluble product fractions, a triangulation approach was applied, combining independent analytical results from gravimetric analysis, DSC, and TA quantification. The total mass of the WIR was determined gravimetrically after filtration and drying (see Section 2.3). The TA yield serves as a conservative indicator of the depolymerized PET fraction (Equation (2)), as this gravimetric value does not account for potential losses in downstream processing. DSC-derived values for WIR were obtained by applying the virgin-material calibration as described in Section 2.5 and are used for internal consistency assessment.

From these data, the mass of residual PET (Equation (4)) and EL in the WIR (Equation (5)) was calculated as follows:

$${\mathrm{m}}_{\mathrm{P}\mathrm{E}\mathrm{T},\mathrm{W}\mathrm{I}\mathrm{R}}={{\mathrm{\omega }}_{\mathrm{P}\mathrm{E}\mathrm{T}}·\mathrm{m}}_{\mathrm{p}\mathrm{e}\mathrm{l}\mathrm{l}\mathrm{e}\mathrm{t}}-{\mathrm{m}}_{\mathrm{T}\mathrm{A}}·{\displaystyle \frac{{\mathrm{M}}_{\mathrm{P}\mathrm{E}\mathrm{T}}}{{\mathrm{M}}_{\mathrm{T}\mathrm{A}}}},$$

(4)

$${\mathrm{m}}_{\mathrm{E}\mathrm{L},\mathrm{W}\mathrm{I}\mathrm{R}}={\mathrm{m}}_{\mathrm{W}\mathrm{I}\mathrm{R},\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}·{\mathrm{\omega }}_{\mathrm{E}\mathrm{L},\mathrm{D}\mathrm{S}\mathrm{C}}$$

(5)

where ${\mathrm{\omega }}_{\mathrm{EL},\mathrm{DSC}}$ denotes the apparent elastane mass fraction in the WIR obtained from DSC (Section 2.5), ${\mathrm{m}}_{\mathrm{W}\mathrm{I}\mathrm{R},\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ is the total mass of the WIR in g, and the remaining variables are defined as above. The TA-based estimate of residual PET mass was combined with ${\mathrm{\omega }}_{\mathrm{E}\mathrm{L},\mathrm{D}\mathrm{S}\mathrm{C}}$ to derive an apparent component split of the gravimetrically measured WIR. Deviations from ${\mathrm{m}}_{\mathrm{W}\mathrm{I}\mathrm{R},\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}$ quantify the internal consistency of this combined estimate. While absolute transferability is not assumed, the resulting apparent values are considered suitable for relative comparison between 95/5 and 85/15 WIR, because both residues experienced comparable thermal and chemical histories and are expected to be affected by similar calibration-transfer bias. This diagnostic comparison forms the basis of the triangulation presented in Section 3.2.3.

3. Results and Discussion

The following section presents the experimental results obtained from the alkaline depolymerization of PET/EL blends and the subsequent analytical evaluation. The discussion is structured according to the sequence of analytical approaches described in Section 2. First, the DSC calibration and its applicability to degraded elastane residues are examined. This is followed by an analysis of the thermal behavior of the WIR and the quantification of TA yields from the filtrate. Finally, the analytical consistency of these datasets is assessed through a triangulation of DSC, gravimetric, and TA-based results. Together, these evaluations provide insight into the degradation behavior of elastane and its influence on PET depolymerization and product composition.

3.1. Evaluation of Depolymerization Experiments

Depolymerization experiments on pure PET textile (100/0) and polyester textile materials with 5 wt% (95/5) and 15 wt% (85/15) EL content were performed in a laboratory kneader. The experiments were conducted as described in Section 2.2 and shown in Figure 2. Figure 3 shows the characteristic mass temperature and torque profiles during the experiments.

The mass temperature profile started at approximately 120 °C, as the material was introduced into the preheated mixing chamber at room temperature. After a short heating phase of approximately 15 s, a pronounced temperature increase from 120 °C to 160 °C was observed within 30 s. Crucially, the temperature profiles for the 95/5 and 85/15 materials are virtually identical regarding peak temperature and heating rate. This similarity suggests that the dominant heat generation and heat transfer processes are governed by PET depolymerization and mixing, whereas additional thermal contributions from elastane side reactions are not resolved at the level of the mass temperature signal. If significant exothermic or endothermic degradation of the elastane were occurring at comparable rates, a deviation in the thermal profile would be expected for the material with the higher elastane content. The absence of such a deviation indicates that PET depolymerization performance is largely insensitive to increased elastane content under the studied reaction window, but it does not constitute a quantitative proof of intrinsic kinetic selectivity. The larger standard deviation observed for the 85/15 material (Figure 3b) is likely attributed to the altered rheology and heat transfer of the elastane-rich mixture rather than chemical instabilities. After the rapid reaction phase, the mass temperature gradually approached the chamber temperature of 140 °C asymptotically.

For both materials, a steep increase in torque was observed immediately after filling, coinciding with the closure of the pneumatic ram of the kneader. This initial spike resulted from the resistance the solid textile pellets exerted against the rotors. Subsequently, the torque rapidly decreased as the depolymerization proceeded, driven by the reduction in reaction volume and the lubricating effect of the formed ethylene glycol. Notably, a slight increase in torque occurred from 50 s to approximately 170 s. This minor increment is likely attributed to the delayed incorporation of residual pellets located at the lower section of the circular feed opening above the central kneading chamber. As the reaction volume decreased, the pneumatic ram forced these remaining pellets into the mixing chamber. The contact of these solid pellets with the rotating rotors caused a renewed mechanical resistance, resulting in the observed temporary rise in torque.

Compared with previous studies, the torque observed for the pelletized textiles, with a maximum of 3 Nm, was significantly lower than that reported for experiments using loose textile fibers (12 Nm) [18]. This difference is consistent with reduced fiber entanglement and more uniform packing for pellets, which might lower mechanical resistance during mixing. Lower torque indicates reduced mechanical resistance in the kneader and therefore lower frictional losses and viscous dissipation. In principle, lower dissipation could reduce frictional heating and thereby negatively affect reaction progress. However, under the present conditions, the dominant heat input is imposed by the heated chamber walls, so the torque signal alone does not imply thermal influences on PET depolymerization.

Instead, the observations lead to the working hypothesis that high shear forces are not the primary driver of PET depolymerization rate under these conditions. PET depolymerization is expected to be rapid upon NaOH contact with accessible ester sites, and overall progress is therefore likely governed mainly by macroscopic mixing and continuous renewal of reactive PET surface. This interpretation is conceptually consistent with shrinking-core type descriptions of heterogeneous alkaline PET hydrolysis [19].

In the experiments, the temperature peaks were close to 170 °C, at which thermal onset of elastane degradation has been reported in the literature [2]. Furthermore, local hotspots cannot be excluded, and therefore, thermal degradation of the elastane can be expected to occur alongside potential hydrolysis of its inherent chemical bonds. For the 85/15 material, a larger variation in mass temperature was observed, accompanied by a lower mean value compared to the 95/5 material.

3.2. Water-Insoluble Residues

After completion of the depolymerization experiments, the solid residues obtained after filtration and drying were analyzed to characterize the WIR. This fraction primarily consists of unreacted PET and elastane-derived residues that were not solubilized in water after depolymerization. In addition, DSC based elastane quantification methods described in the literature were applied to the present residue system and systematically evaluated for their suitability in this chemically modified multi-component matrix. The following sections describe the thermal behavior of the WIR determined by DSC, the derivation of elastane-specific calibration curves, and the evaluation of the elastane degradation behavior in comparison with the reference materials.

3.2.1. DSC Calibration

Figure 4 shows the DSC calibration established for cryogenically ground PET/EL reference materials. The calibration standards were prepared from pure PET and EL components that were weighed into the DSC pans in defined mass ratios, as described in Section 2.4 and Section 2.5. The data points represent the mass-normalized melting enthalpy of the elastane soft segment melting endotherm obtained in the second heating run of the temperature program given in Equation (3). The procedure follows the approach reported by Boschmeier et al. for the determination of elastane in textile blends [14].

A linear correlation between mass-normalized melting enthalpy and elastane content was obtained, described by the regression equation y = 0.51976x with a coefficient of determination R² = 0.99342. The corresponding fit reported by Boschmeier et al. is y = 0.4158x + 0.0922. The slope obtained in the present calibration is therefore 25% higher than the value reported by Boschmeier et al. Potential reasons include differences in DSC instruments, pan types, sample pretreatment, and reference standards. In addition, the present study employed a cooling step to −80 °C prior to heating (Section 2.5), which may influence soft segment crystallization behavior and thus the measured peak area. Variations in the degree of pan coverage could further contribute to discrepancies between calibrations.

The PET/EL materials and the 100 wt% EL reference are included in Figure 4 for comparison. Both materials follow the general trend of the calibration curve, and the 100 wt% EL sample lies within the calibration range. The measured peak areas for pelletized and non-pelletized fabrics are close, with a slight tendency toward higher peak areas for the pelletized samples. This small deviation may reflect residual thermal effects introduced during pelletizing or minor differences in sample packing, but it does not alter the overall linear relationship. Both types of materials were analyzed without prior cryogenic grinding; therefore, effects associated with cryogenic treatment are limited to the calibration standards.

Within the investigated concentration range, the DSC method thus provides a robust linear response for EL in PET/EL mixtures and can be used to quantify EL content in undegraded reference materials and feedstocks. This quantification approach assumes that the soft segment melting enthalpy per unit elastane mass is comparable between calibration standards and the analyzed sample. In WIR, elastane was exposed to alkaline conditions and elevated temperatures and may exhibit structural modification. The WIR may also contain additional insoluble reaction-derived material that affects thermal transitions and peak integration. Therefore, calibration-based values for WIR are reported as apparent elastane fractions and used diagnostically to assess residue modification rather than as absolute composition.

3.2.2. DSC of Water-Insoluble Residue

The WIR obtained after alkaline depolymerization was analyzed using the procedure described in Section 2.4 and Section 2.5 to probe thermal transitions associated with elastane and residual PET and to assess whether the elastane phase retains its reference-like thermal signature after treatment. The WIR represents the solid fraction recovered by filtration and drying after separation of the aqueous phase (Section 2.3). Figure 5 shows representative thermograms of the PET/EL reference materials before alkaline treatment (left column) and the corresponding WIR samples after depolymerization (right column). The upper row displays the first heating runs from −80 °C to 280 °C, and the lower row the second heating run. In each plot, the upper trace corresponds to the 85/15 material and the lower trace to the 95/5 material.

In the reference materials, only a very small soft segment melting peak was observed near 0 °C in the first heating run, while the melting peak of PET at approximately 250 °C dominated in both cases. PET melting exhibited a single peak without pronounced shoulders, indicative of a relatively homogeneous crystal morphology. The weak soft segment signal suggests that the elastane soft segments in the untreated textiles are strongly constrained and contribute only weakly to the enthalpy signal under the applied DSC conditions.

For the WIR samples, a distinct soft segment melting peak appeared already in the first heating run at about 32 °C, slightly higher than the value reported by Boschmeier et al. This change is consistent with enrichment of elastane-derived material in the solid fraction and with reduced constraint or altered crystallization behavior of the soft segments after depolymerization. In the second heating run, the soft segment peak shifted to 23 °C, close to the temperature reported by Boschmeier (≈20 °C), and the peak area increased with the elastane content of the material. The appearance and shift in the soft segment endotherm indicate that the WIR does not behave as an undegraded PET/EL blend, which supports the interpretation that virgin-based DSC calibration is not directly transferable to WIR.

Changes were also observed in the PET melting region of the WIR samples. The altered peak shape can reflect changes in PET crystallinity induced by alkaline exposure and thermal history, and it may also include overlapping contributions from elastane hard segment-related transitions in a similar temperature range [14]. Consequently, both soft segment and possible hard segment transitions can be identified qualitatively in the WIR, but their relative contributions cannot be quantified from DSC alone.

The presence of hard segment-related features in the WIR may arise from incomplete solubilization of elastane-derived fragments, including aromatic diamines formed via cleavage of urethane or urea linkages and remaining oligomeric hard segment blocks. Aromatic diamines and oligomeric fragments can exhibit limited solubility under the alkaline conditions used and are therefore expected to partition at least partially into the solid phase. The glass-transition step at approximately −70 °C is consistent with a polyether-type soft segment as discussed by Król [17]. A slight downward drift of the DSC baseline was also observed, which may reflect ongoing thermal processes such as the release of residual moisture or continued degradation during heating.

In summary, the DSC results reveal pronounced differences between the untreated reference materials and the WIR obtained after depolymerization. Both soft segments and additional high-temperature features consistent with elastane hard segment-derived material are observed qualitatively in the residues. These changes indicate substantial modification of the elastane phase under the applied conditions and provide a mechanistic basis for treating DSC based WIR quantification as apparent rather than absolute.

3.2.3. Component Balance and Triangulation

To relate soluble and insoluble product fractions and to critically assess the transferability of the applied quantification methods to the present residue system, an apparent component balance was derived by combining (i) gravimetric TA yield, (ii) DSC derived apparent elastane fraction in WIR, and (iii) gravimetrically measured WIR mass. The triangulation is therefore used to assess internal consistency between complementary measurements rather than to claim strict mass balance closure.

The gravimetrically determined TA yields were 68% ± 3.4% for pure PET (100/0), 71% ± 6.4% for 95/5, and 68% ± 7.0% for 85/15 (Figure 6). All three materials fall within a narrow yield corridor despite the intentionally imposed base competition scenario. The scatter is consistent with the heterogeneous nature of the quasi-solid-state process and may additionally reflect batch-to-batch differences in pellet incorporation during the first minutes of mixing, as indicated by the transient torque increase discussed in Section 3.1. Importantly, the similar TA yields across 95/5 and 85/15 indicate that PET depolymerization performance is not measurably impaired at higher elastane content within the studied reaction window, despite two effects that could have reduced PET conversion. These effects are partial NaOH consumption by elastane side reactions and a potentially reduced probability of PET NaOH contact due to the elastane fraction. This outcome supports the working hypothesis that PET depolymerization is rapid once NaOH contacts accessible ester sites and that the overall progress is governed mainly by macroscopic mixing and surface renewal under the applied conditions.

A comparison with previous investigations in the same reactor setup shows that pure PET fibers yielded approximately 54% TA under similar nominal conditions (140 °C, 5 min, 40 rpm) [18]. In the present study, isolated TA yields for PET/EL blends remained in a comparable range, whereas pure PET reached a higher yield while the charged mass was increased by densified, pelletized feedstocks at a defined chamber filling degree (approximately +11 g versus loose fibers) [18]. This supports the working hypothesis that PET depolymerization is rapid upon NaOH contact and that overall performance is governed largely by macroscopic mixing and surface renewal, while isolated yield still reflects combined reaction and recovery efficiency rather than intrinsic conversion.

Because NaOH dosing was fixed relative to the PET fraction, the effective NaOH availability per unit elastane mass differs by a factor of 3.3 between 95/5 and 85/15. The absence of a systematic decrease in TA yield at higher elastane loading suggests that any NaOH consumption by elastane reactions does not become significantly limiting for PET depolymerization under the applied stoichiometry and mixing regime. This conclusion is consistent with literature reports that ester hydrolysis is generally more facile than cleavage of urethane and urea linkages under alkaline conditions, although the present study does not derive kinetic constants [20,21]. In addition, the macromolecular architecture of elastane may limit accessibility of reactive hard segment sites, because hard segments represent only a minor fraction of the elastane mass and can be shielded by soft segment domains [11].

Mean apparent EL recoveries were about 90% for 95/5 and close to 100% for 85/15, with individual measurements exceeding 100% (Figure 7). Values above 100% indicate that applying a virgin material DSC calibration to WIR is not quantitatively transferable. A plausible reason is that partial elastane modification can change the soft segment melting enthalpy per unit elastane mass and can increase soft segment mobility, which directly affects the integrated signal used for calibration. Consequently, the DSC derived values are treated as apparent elastane fractions and are used diagnostically. The slightly higher apparent recovery for 85/15 is not interpreted as higher true retention, but as a method-sensitive response to differences in residue composition and elastane modification. Importantly, the stable TA yields (Figure 6) indicate that elastane-related effects are secondary with respect to PET depolymerization performance under the investigated conditions.

To further assess the consistency of the mass balance, the gravimetrically measured total WIR mass was compared with the back-calculated component masses of residual PET and EL, calculated from the TA yield and DSC data according to Equations (4) and (5) (Figure 8). For both materials, the sum of the calculated component masses ${\mathrm{m}}_{\mathrm{P}\mathrm{E}\mathrm{T},\mathrm{W}\mathrm{I}\mathrm{R}}$ + ${\mathrm{m}}_{\mathrm{E}\mathrm{L},\mathrm{W}\mathrm{I}\mathrm{R}}$ exceeded the experimentally measured WIR mass. This indicates a systematic positive bias in the combined estimation rather than evidence of process instability.

Two methodological characteristics likely contribute to this behavior. First, DSC based back calculation can overestimate the elastane-related fraction in WIR when soft segment mobility is enhanced by partial modification. Second, the TA-based PET back calculation assumes that unrecovered TA corresponds to unreacted PET. However, isolated TA yield integrates precipitation efficiency and may also be affected by product entrapment within WIR, which can lead to an overestimation of residual PET when used for back calculation. Accordingly, the triangulation result is interpreted as a diagnostic indicator that the present quantification toolbox is not sufficient for a closed component balance on WIR.

To resolve the elastane fate and NaOH consumption more directly, future work should combine solvent-based fractionation of WIR, for example, dissolution in DMF, followed by analysis of the soluble fraction, with complementary spectroscopic and chromatographic methods such as solid-state NMR and HPLC-based quantification. In addition, alkalinity titration of the aqueous phase and carbon inorganic analysis could help constrain base consumption pathways.

3.2.4. TA Quality Analysis

Based on the mechanism of alkaline hydrolysis, the cleavage of urethane and urea containing hard segments in MDI-based elastane can generate aromatic diamines, most notably MDA, alongside polyols and other products (Figure 1) [3]. Consequently, the recovered TA was analyzed by ¹H NMR spectroscopy and compared with a reference sample intentionally spiked with MDA (Figure 9). The objective was to determine whether MDA co-isolated with TA could be detected in the recovered product.

The spectrum of the recovered TA shows only the characteristic aromatic proton signal of terephthalic acid at δ ≈ 8.0 ppm and the carboxylic proton signal at δ ≈ 13.1 ppm, both matching the reference spectrum of pure TA [18]. In contrast, the MDA-spiked sample exhibits additional resonances between δ = 6.47–6.90 ppm, which correspond to the aromatic protons of MDA [19]. No corresponding resonances were observed for the recovered TA above the baseline noise level. This indicates that MDA, if present in the isolated TA, is below the detection capability of the present ¹H NMR measurement under the applied acquisition conditions. A slight downfield shift in the carboxylic proton resonance in the spiked sample indicates weak hydrogen-bonding interactions between TA and MDA, an effect that is absent in the recovered product.

Importantly, this targeted NMR screen addresses aromatic diamine contamination in the isolated TA fraction. It does not exclude formation of MDA during reaction if it remained in the aqueous phase, was removed during washing, or did not co-precipitate with TA. It also does not rule out other low-level impurities that are not resolved in this spectral window. Overall, the NMR analysis supports that the recovered TA is free of detectable co-isolated MDA within the scope of the applied method.

4. Conclusions and Outlook

Alkaline depolymerization of PET and PET/EL textiles was demonstrated in a laboratory kneader under quasi-solid-state conditions (140 °C, 5 min, 40 rpm) using pelletized feedstocks and PET-based stoichiometric NaOH dosing. Across pure PET (100/0) and PET/EL blends (95/5 and 85/15), isolated TA yields remained within a narrow corridor, indicating that PET depolymerization performance is largely robust toward increased elastane content within the investigated reaction window. This robustness is consistent with the working hypothesis that PET depolymerization is rapid upon NaOH contact with accessible ester sites and that overall progress depends on macroscopic mixing and surface renewal, conceptually in line with shrinking-core type models for heterogeneous alkaline PET hydrolysis [19].

DSC thermograms of water-insoluble residues showed clear changes in elastane-related thermal transitions, consistent with structural modification under alkaline and thermal stress. Applying a virgin-material DSC calibration to these residues yielded apparent elastane fractions that could exceed 100% recovery, demonstrating that the DSC quantification approach described in the Literature is not directly transferable to chemically modified residues [14]. The component-balance triangulation, therefore, served primarily as a consistency assessment and highlighted methodological limitations in closing a component-resolved mass balance with gravimetry, isolated TA yield, and DSC alone.

¹H NMR screening of recovered TA against an MDA-spiked reference showed no detectable co-isolated aromatic diamine resonances under the applied measurement conditions, supporting high TA purity in the isolated monomer fraction. However, this does not exclude the formation of soluble elastane degradation products that remain in the aqueous phase or are removed during washing.

Overall, the results support that alkaline PET depolymerization can tolerate elastane-containing textile feedstocks without requiring quantitative pre-separation, while elastane side reactions and residue modification must be expected and considered in analytical workflows. Future work should focus on (i) identifying and quantifying soluble elastane-derived products by chromatographic methods, (ii) applying solvent-based fractionation to better resolve the chemical fate of elastane residues, and (iii) systematically investigating the influence of reaction parameters such as temperature, rotor speed, and residence time. These steps are required to close the mass balance with higher confidence and to inform transfer to continuous processing concepts such as reactive extrusion.