Textile Waste Fibre Regeneration via a Green Chemistry Approach: A Molecular Strategy for Sustainable Fashion


 Fast fashion has been widely criticised for its excessive resource use and high generation of textile. To reduce its environmental impacts, numerous efforts have focused on finding sustainable and eco-friendly approaches to textile recycling. However, waste textiles and fibres are still mainly disposed of in landfills or by incineration and thereby pollute the natural environment, as there is still no effective strategy to separate natural fibres from chemical fibres. Herein, we developed a green chemistry strategy for the separation and regeneration of waste textiles at the molecular level. Cellulose/wool keratin composite fibres and multicomponent fibres were regenerated from waste textiles via ionic liquids. Our strategy attempts to reduce the large amount of waste textiles generated by the fast-developing fashion industry and provide a new source of fibres, which can also address the fossil fuel reserve shortages caused by chemical fibre industries and global food shortages caused by natural fibre production.

Despite the high production of textile wastes, their recycling rates remain low: only 15% of textile waste was collected and sorted for recycling in 2015, and 1.1 million tonnes were lost during the process 23 .
Most recycled textiles cascade to other industries and are downcycled into lower-value applications 3 .
Given the current low recycling rates, solving the problem of fashion waste and improving environmental and resource sustainability are left to material scientists. From a material perspective, the current major barrier hindering textile recycling is the lack of technologies for sorting and separation 12 . Most synthetic bres, such as polyesters, can be reused easily by melt spinning when pure fractions of bres can be isolated 24,25 . However, most fabrics are made of different kinds of bres 12 , such as cotton/polyester and wool/polyester blends, which could be di cult or nearly impossible to isolate once blended 26 . Such blends cannot be simply recycled as a whole by melt spinning since natural bres are mostly thermoset; therefore, the β-sheet nanocrystallites in the structure would be degraded prior to melting when subjected to thermal processing 27 . Thus, the critical question is how to separate natural bres from synthetic bres.
Traditional methods for the regeneration of natural bres usually employ concentrated acidic or alkaline solutions, which destroy the polymer structures of bres, rendering them unusable as bres in textiles and fashion. Moreover, these chemical processes unavoidably introduce hazardous substances and have been strictly prohibited in some countries 28, 29 . As we cannot cause one environmental problem while solving another, nding a green and sustainable process for waste blended bre separation will bene t the sustainable development of the fashion industry and reduce pollution, such as that by ubiquitous plastic particles. Ionic liquids (ILs) are closely associated with the green chemistry movement because of their low volatility, nonderivatisation and renewability. No hazardous byproducts or gas discharge are produced in reactions involving ILs. Furthermore, ILs are capable of dissolving recalcitrant natural biopolymers with massive inter-and intramolecular hydrogen bonds, such as cellulose 30,31 and wool 32 [DMP], to separate textile waste blends and extract natural polymers from blended fabrics, leaving polyester unchanged, and successfully spun the natural polymers into pure cellulose and cellulose/wool keratin (w/w = 1/1) composite bres. The obtained cellulose/wool keratin composite bres have a unique morphology, with wool keratin microspheres adhered on the surface, that is different from the morphologies reported in previous studies [34][35][36] , and outstanding moisture regain, which might be due to their morphology. Acrylic was also successfully dissolved in [BMIM][Cl], indicating the capability of this IL to dissolve all hydrogen-bonded polymers, both natural and synthetic. The ILs can be simply recovered and reused via distillation, thus achieving green and sustainable recycling of textile waste without the use of hazardous substances.

Results
Dissolution and separation of blended polymer fabrics Blended cotton/polyester (50%/50%) and wool/polyester (50%/50%) fabrics were separated by dissolution with ILs, leaving undissolved polyester bres to treat via melt spinning. [BMIM][Cl] and [MMIM][DMP] were applied separately to evaluate the dissolution process, and three different ratios (cotton/polyester:wool/polyester = 1:0, 1:1, 0:1) of the blended fabrics were applied in the experiment. As illustrated in Fig. 2b, the surface of the blended weaving fabric had a hair-like structure arising from the cotton bres aligned in the weft direction; this structure was disrupted by dissolution, whereas the smooth polyester bres aligned in the warp direction remained unchanged. During dissolution, the colour of [BMIM][Cl] turned from clear yellow to dark brown.

Microstructural Changes Of The Regenerated Polymers
To analyse the chemical changes occurring during the dissolution and regeneration, ATR-FTIR spectroscopy of native polymer fabrics, regenerated cellulose, wool keratin, cellulose/keratin composite and acrylic materials was carried out, as illustrated in Fig. 3a-d. To understand the crystallinity structures and polymorphy of the native and regenerated polymer samples, X-ray diffraction measurements were employed, and the spectra are compared in Fig. 3e-h.
The spectrum of the regenerated cellulose is similar to that of the cotton fabric: no new peaks were detected. This indicates that cellulose was recovered from the IL without distinct chemical structure changes. The broad band at 3326 cm − 1 is assigned to the stretching of -OH groups, which indicates the presence of hydrogen bonds. The signal at approximately 2890 cm − 1 could correspond to C-H stretching, and the band at 1645 cm − 1 is attributed to amorphous water. The peak at 1316 cm − 1 originated from O-H bending vibrations, and the small sharp band at 899 cm − 1 could be due to O-H stretching. The strong band at approximately 1011 cm − 1 is assigned to the characteristic C-O-C stretching, the intensity of which is weaker than that in the spectrum of the raw cotton fabric. The crystalline structures of cellulose can be divided into six polymorphic forms (cellulose I, II, III I , III II , IV I , and IV II ), of which cellulose I is the form in nature and can be converted into the more stable cellulose II, with a three-dimensional network structure, via regeneration or mercerization 37  (110), (020) and (004), corresponding to cellulose 39 . These results indicate the transformation from cellulose to cellulose , with rapid breaking and reforming of inter-and intramolecular hydrogen bonds among cellulose molecules during the process 40 . Moreover, the lower-and borderline-crystallinity peaks of regenerated cellulose re ect a decrease in crystallinity, probably due to a coagulation process that is unfavourable to cellulose crystallization 41 .
The spectra of the wool fabric and regenerated wool keratin show almost identical characteristic absorption bands that could be traced back to peptide bonds (-CONH), suggesting the successful recovery of wool keratin. The absorption band at 3273 cm − 1 could be due to N-H stretching from amide A vibrations. The strong peak observed at 1630 cm − 1 corresponds to C = O in the amide band, and the peak at 1518 cm − 1 is from N-H bending and C-N stretching in amide . A weak band at 1235 cm − 1 is assigned to C-O and C-N stretching and O = C-N and N-H bending in amide III. Two typical crystal structures are determined from the pattern of raw wool fabric, with a peak at approximately 10.04°a ttributed to an α-helix structure and a peak at 21.88° indicating a β-sheet structure 42 . The XRD spectrum of the regenerated wool keratin generally matches that of the raw wool fabric except that the peaks are signi cantly weaker, indicating lower crystallinity and a more amorphous structure, which can improve water absorption.
In the spectrum of the regenerated cellulose/wool keratin composite, the peaks corresponding to amide and amide II are at approximately 1638 cm − 1 and 1521 cm − 1 , but their intensities are lower than those above due to the decreased concentration of keratin in the blend 43 . The peaks at 1010 cm − 1 and 899 cm − 1 are consistent with those in the cotton fabric. Similar to that of the regenerated cellulose, the XRD spectrum of the regenerated cellulose/wool keratin lm contains peaks at 20.02° and 21.83°, indicating that the cellulose II structure is present in the composite, and the peaks at 8.89° and 12.07° could be a result of the presence of wool keratin. The small and sharp peaks at 21.59° and 24.00° in the regenerated composite bre are probably a result of the β-sheet structure from wool keratin and the cellulose II structure from cellulose. No new peaks appear in the FTIR spectrum of the regenerated composite in comparison with the spectra of regenerated pure cellulose and wool keratin, indicating that the macromolecular structure remains unchanged during the dissolution and regeneration; instead, bonds in the crystal structure break, resulting in polymer chain rearrangement.
The ATR spectrum of regenerated acrylic is similar to that of the raw fabric, except that it contains peaks at 3384 cm − 1 and 1169 cm − 1 , which may correspond to the stretching of -OH groups and C-N stretching, respectively. Compared to the spectrum of the original acrylic, in the spectrum of the regenerated acrylic, the major crystallinity peak at 16.96° is much weaker, and the peak at 29.60° is absent, which indicates that the acrylic became amorphous and that the crystal structure was damaged during recycling.

Morphology of the regenerated polymer materials
The regeneration of dissolved natural polymers can be achieved by diffusional interchange between the IL and a liquid bath, which is also known as the coagulation step of the wet-spinning process. Here, water was chosen for the coagulation bath, in which cellulose bre and cellulose/wool keratin composite bre were regenerated, and their SEM images are shown in Fig. 4a-i and Supplementary Fig. 5a-i. As shown in Fig. 4a-c, the regenerated cellulose bre exhibits a smooth and even surface with a streaky appearance, of which the brillar texture was formed by the polymer chains of D-glucose units lying alongside each other 44 . Despite the consistency of the regenerated cellulose bre morphology with that reported in previous studies 34 , it is interesting that the morphology of the regenerated cellulose/wool keratin composite bre illustrated in Fig. 4d-i is signi cantly different from that reported previous studies [34][35][36] . Two types of cellulose/wool keratin composite bres with the same blend ratio of 2.5 wt%/2.5 wt% were obtained via wet spinning from [MMIM][DMP], presenting a similar bre diameter but distinct morphology as a result of the different spinning speeds. The morphology of the composite bre regenerated at a lower spinning speed is demonstrated in Fig. 4d-f and Supplementary Fig. 5a-c, showing tightly packed globular microparticles covering the surface, while the bre spun at a higher spinning speed, shown in Fig. 4g-i and Supplementary Fig. 5d-i, exhibits a bril structure along the spinning direction and 'globules' growing from the gaps. Both types of bres display a phase-separated structure. Referring to previous studies, we found a similar globular structure in regenerated wool keratin particles and silk bres 27,[45][46][47][48] , thus indicating that the microspheres on the surface of the composite bres are polymeric micelles selfassembled by wool keratin and the bril structure was formed by cellulose (Fig. 4j).
This unique morphology may be explained by the amphiphilicity of wool keratin and the rearrangement of its chain folding. During dissolution, the breaking of disul de bridges and hydrogen bonds is oriented by the IL, leading to the disruption of the 3D structure of wool keratin and the unfolding of the α-helix. The resultant amorphous polypeptide chain is amphiphilic, consisting of both hydrophobic and hydrophilic amino acids with their side chains extending outward, potentially enabling the formation of micelles in water. The main driving force for micelle formation, a thermodynamic process, is the minimisation of the interfacial free energy of the polymer-water system 49 . When the spinning dope of dissolved polymers is added to water, the polymer concentration rapidly increases and reaches the critical micelle concentration (CMC), thus inducing the self-assembly of wool keratin. As illustrated in Fig. 4k, the polypeptide chains tend to coil and aggregate once spun into the water and eventually self-assemble into a spherical structure, with the main chains bound to each other through intermolecular disul de bonds. The R-groups from hydrophobic amino acids clump into the centre to form the core, while the hydrophilic groups from polar amino acids shield the inside from contact with water, thereby reducing the interfacial free energy to attain the most stable state. Meanwhile, cellulose, as an unbranched molecule, is regenerated from randomly oriented D-glucose subunits back to aligned polymeric chains along the spinning direction, which make up the bril structure on the bre surface.
The effect of the spinning conditions on micelle formation has also been demonstrated. As shown in Supplementary Fig. 5a-c, the regenerated cellulose/wool keratin bre obtained via a lower spinning speed presents obviously more 'globules' on its surface, which suggests that the number of distributed micelles can be controlled by the duration of the contact with water. In addition, we observed that the sizes of micelles from different parts of the same bre differ, as the diameters of the 'globules' on the anterior part of the bre obtained at a higher spinning speed ( Supplementary Fig. 5d-f) vary from 140-600 nm, similar to the bre obtained at the lower speed, while the sizes of the 'globules' on the posterior part ( Supplementary Fig. 5g-i) range between 100-300 nm. This could be explained by the principle of the wetspinning process and the properties of ILs, as the lament is formed by diffusional interchange with coagulation. When ILs diffuse into water, the surface tension of the water decreases due to the addition of this surfactant, thus limiting further formation of keratin micelles.

Mechanical Properties Of Regenerated Polymer Fibres
According to previous studies 34,43,50 , the poor mechanical properties of wool keratin have limited its potential application; the addition of cellulose may enhance its mechanical strength. The practical application of the regenerated cellulose and cellulose/keratin composite bres was studied, and the stress-strain curves are shown in Fig. 5a-c. Overall, the regenerated cellulose showed better mechanical properties than the regenerated cellulose/wool keratin composite, consistent with previous reports. The rate of moisture regain by the various bres at 20°C and 25% relative humidity (RH) is given in Fig. 5d.
Both regenerated bres showed higher moisture regain than raw cotton and wool bre. This higher moisture regain re ects better comfortability for wearing and is probably the result of lower crystallinity and more amorphous structure formation after regeneration. The moisture regain rate of wool keratin composite bre is highest, reaching 18 wt%, which may have also resulted from the increased speci c surface area due to micelle structure formation.
In addition to bres, we demonstrated the application of other regenerated cellulose/keratin composite products in the form of cyclic annular hydrogels (Fig. 5f).

Conclusion
In this study, a simple method was applied to convert waste textiles into regenerated materials. We utilised two ILs, [BMIM][Cl] and [MMIM][DMP], to separate natural bre/polyester fabrics, which achieves not only blended-fabric separation, providing a feasible method of textile recycling at scale, but also complete recycling of waste textiles and cyclic utilisation of solvents. Two types of blended fabrics, cotton/polyester and wool/polyester blends, were rst dissolved by the ILs. The remaining solid polyester could be simply recycled by melt spinning, while the extracted cellulose and wool keratin could be regenerated via wet spinning.
Several types of natural polymeric products were successfully regenerated using water for coagulation, including regenerated cellulose microspheres, cellulose bre, freeze-dried wool keratin hydrogel and cellulose/wool keratin composite bre. We also used [BMIM][Cl] to dissolve and regenerate acrylic fabric, which became hard and fragile after regeneration and will need further study for application. Cellulose/wool keratin composite bres were successfully regenerated for the rst time via [MMIM] [DMP] and presented a signi cantly different morphology from that in previous reports. Microspheres were clearly observed adhered on the surface along the direction of injection, which could indicate a wool keratin component in the mixture. Such regenerated composite bres showed good moisture regain and suitable mechanical properties, supporting their wearability.

Methods
Ionic liquid preparation 1-Butyl-3-methylimidazolium chloride ([BMIM][Cl]) was prepared by heating a mixture of 1methylimidazole and 1-chlorobutane at a ratio of 1:1.1 at 80 °C for 24 hours, with a re ux condenser tted to reduce volatilization. The water was removed by rotary evaporation to yield 1-butyl-3methylimidazolium chloride as a pale yellow, highly viscous liquid. 1,3-Dimethylimidazolium dimethyl phosphate ([MMIM][DMP]) was prepared in a similar way but with trimethyl phosphate instead of 1chlorobutane at a temperature of 145 °C, and the liquid obtained was dark brown.

Blended fabric separation
The fabrics were rst soaked in distilled water at room temperature for 72 hours to improve their solubility in the ILs and dried at 100 °C in a vacuum oven for over 2 hours before being used. Prior to dissolution, the ILs were heated under high vacuum to reduce the water concentration, and the dried fabric samples were cut into 1-2 cm rectangular pieces. [DMP] in a 100 mL round-bottomed ask with a magnetic stirrer inside and heated in an oil bath at a constant temperature of 130 °C for 3-10 hours. The ask was tted with a re ux condenser, kept in a fume hood, and stirred continuously. After dissolution, the mixture was ltered to remove the solid material and cooled to room temperature, yielding the polymer-IL solution.
Subsequently, the ltrate was washed with water, dried and weighed to determine the dissolution rate of the natural bres. An acrylic-IL solution (1:20 w/v) was prepared in a similar way by adding 5 wt% acrylic fabric to [BMIM] [Cl] with stirring under re uxing until complete dissolution was con rmed.

Polymer regeneration
Water was selected as a coagulating solvent for all polymer regeneration processes.

Wet spinning
Wet spinning was carried out using a simple laboratory setup. The cellulose-IL solution, wool keratin-IL solution, cellulose/wool keratin-IL solution and acrylic-[BMIM]Cl solution were degassed prior to spinning to remove any air bubbles and then loaded into a 20 mL syringe equipped with a pipette tip (internal diameter, 0.5 mm). The mixed solution was then extruded into a water bath and dragged at a constant speed at ambient temperature. The extruded bres were collected and dried without further stretching at room temperature for 3 days.

Microsphere regeneration
The cellulose-IL solution was immersed in water in a 100 mL glass beaker at room temperature. The solution was dispersed in water for 30 minutes to remove the IL from the cellulose. After strati cation for 10 minutes, the solution was ltered under vacuum, and the ltrate was dried. The obtained ltrate consisted of cellulose microspheres. The regenerated microspheres were washed 3 times with water via centrifugation to remove any retained IL and dried under vacuum for 3 days at 45°C.

Hydrogel regeneration
The wool keratin-IL solution was cast onto a mould, degassed to remove air bubbles, cooled to room temperature and then immersed in a deionized water bath. The water was exchanged 2 times, with each bath being used for 12 hours, to remove the IL from wool keratin and thus obtain a transparent regenerated hydrogel. The regenerated hydrogel was washed with running distilled water to fully remove the IL and freeze-dried.
After the regeneration process, the ILs were recycled by distillation.

Sample characterisation
The infrared spectra and XRD patterns of the fabrics before treatment and the polymers after regeneration were measured by a Thermo Nicolet FTIR spectrometer (Nicolet 8700) and powder X-ray diffraction (PXRD) system. Images of the sample surface morphology were collected by eld emission SEM (Zeiss Ultra-55). The mechanical properties of the regenerated bres were measured with an Instron 5943. All bre samples were cut to a length of 20 mm and tested at a speed of 5 mm min −1 at ambient temperature. The moisture regain by the regenerated bres was investigated by measuring the weight of a sample oven-dried at 100°C for 2 hours (W 0 ) and the weight of a sample placed at a constant temperature of 20°C and relative humidity of 25% for 24 hours (W 1 ). The moisture regain rate can be obtained by the following equation:     Morphology of the regenerated polymer bres. a, b, c, SEM images of regenerated cellulose bre. d, e, f, SEM images of cellulose/wool keratin (w/w=1/1) composite bre regenerated via [MMIM][DMP] at a lower spinning speed during the wet-spinning process. g, h, i, SEM images of cellulose/wool keratin (w/w=1/1) composite bres regenerated via [MMIM][DMP] at a higher spinning speed during the wet-spinning process. j, Schematic of the formation of the morphology of regenerated cellulose/wool keratin composite bre. k, Self-assembly model including chain folding and micelle formation of wool keratin.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.