Textile Recycling With NIR Spectroscopy

INTRODUCTION

In today's environment, environmental issues are of particular concern worldwide. And one of the biggest international issues is textile waste. The US Environmental Protection Agency estimates that more than 15 million tons of textile waste are sent to landfills each year.

At some point in its life cycle, a textile product will become worn or soiled and no longer suitable for reuse. In this case, recycling can give the material a new lease of life. To add value to recycled material and ensure that it is of good enough quality as an input material for further recycling processes, it is essential to be able to identify and sort items according to their material content.

Many current methods for sorting textiles for recycling involve cumbersome manual handling, but despite this, the proportion of textiles recycled is increasing year by year. Traditional manual sorting in textile recycling is not a completely reliable method of identifying existing materials and is ultimately ineffective. According to a study by Circle Economy, up to 41% of labels on blended materials contain incorrect information. There are methods available for identifying textiles, such as ISO standardized quantitative methods based on different dissolution behaviors (ISO 1833-1, etc.), morphological differences detected by microscopy, DNA identification, and differences in thermal behavior detected by differential thermometry, thermogravimetric analysis, and gas chromatography. These methods are accurate, but require sample preparation and are therefore too slow for automated identification and sorting of textiles required for recycling. Therefore, to increase the use of recycled materials, a feasible and cost-effective way to identify and sort textiles is needed.

Near infrared (NIR) spectroscopy has emerged as a powerful tool with potential to improve the efficiency and accuracy of textile recycling processes. This application note explores the use of NIR spectroscopy in the textile recycling industry, highlighting its advantages, methodology, and practical applications.

METHODS AND RESULTS

Near-infrared spectroscopy

NIR light, typically with wavelengths between 700 and 2500 nm, interacts with molecular vibrations, especially the overtones and combinations of fundamental molecular vibrations. This interaction provides valuable information about the chemical composition and structure of the material. For textile recycling applications, the absorption spectrum of the material will be used. Based on the fingerprints contained in the NIR spectrum, we will be able to identify different textile fibers such as: cotton, polyester (PET), polycotton...

In addition, it will be necessary to build a data library that gathers information about different types of fabrics to use as reference information. This means that it is possible to perform automatic analysis of the infrared spectrum using a suitable algorithm to determine the likelihood of matching with the reference spectrum and then calculate the expected composition. Therefore, NIR spectroscopy is considered as a qualitative and quantitative technique, which can be used to determine the relative percentage composition of finished textiles.

Benefits of NIR Spectroscopy in Textile Recycling

• Fast and non-destructive analysis: NIR spectroscopy allows for fast and non-destructive analysis of textiles. This is important for recycling operations where materials need to be sorted without damaging the fabric.
• High throughput: NIR spectroscopy can analyze multiple samples quickly, making it suitable for high-throughput recycling processes where large volumes of textile waste need to be processed efficiently.
• Versatility: NIR spectroscopy can identify a wide range of textile materials, including natural fibers (cotton, wool, silk), synthetic fibers (polyester, nylon, acrylic) and blends. This versatility is essential for efficient sorting and recycling.
• Cost-effective: Compared to other analytical techniques, NIR spectroscopy offers a cost-effective solution with minimal sample preparation and low operating costs.

In fact, we will present the outstanding results of NIR spectroscopy with the actual results obtained below.

Figure 1: NIR Spectral from 0/100 % PET/CO to 0/100 % CO/PET

Figure 1 shows the spectral overview of 30 samples with different compositions from 100 % PET and 0 % cotton to 0 % PET and 100 % cotton. The characteristic peaks of PET and CO are observed here as follows. 

Polyester is polymerized from the monomers ethylene glycol and terephthalic acid. The characteristic peaks of the NIR spectrum of polyester can be seen between 1100 nm and 1150 nm and around 1630 nm - 1680 nm. C-O bonds from secondary alcohols, such as ethyleneglycol, can cause second overtone vibrations, which lead to peaks between 1087 nm - 1124 nm. In addition, absorptions at the wavelength of 1130 nm can result from the third overtone vibration of CH2 bonds. The peak around 1648 nm can be traced back to stretching of the C = C double bonds in the aromatic ring of the terephthalic acid of the polyester and/or C = O bonds in the polymer. Both bonds can also cause absorption around 1648 nm to 1680 nm. Furthermore, in the range of wavelengths between 1650 nm and 1800 nm, first overtone vibrations of the C–H bond may be responsible for absorption. Cotton mainly consists of cellulose, which is made of glycose units. The most characteristic peaks of the cotton spectrum are at 1355 nm, 1416 nm, 1470 nm, and 1605 nm. In the wavelength range of 1400 nm to 1500 nm, first overtone oscillations of O–H bonds of the glucose may be responsible for the absorption. A peak in the raw spectrum is at about 1550 nm, as this is where the transition between the low point at 1416 nm and the high point at 1605 nm is located. The first overtone vibration from the O–H bonds contained in the cotton is responsible for this increased absorption at 1550 nm.

From this it can be clearly seen that there is a correlation in absorption intensity between the preparations. The results obtained with the mean absolute error (MAE) represents the difference of the setpoint from the measurement is about 4.0 %, there is a deviation of ±4.0 % of the predicted and true cotton share. The present value R2 from 0.975 to 0.99 indicates a very high suitability of the linear regression model to describe the data as well as a very high correlation of the data with each other. Based on the low MAE (4 %) and high R2 -score, it can be stated that a differentiation of textiles according to different polycotton blends based on the NIR spectra is possible with high accuracy.

Figure 2: (a) Natural fibers. (b) Manmade fibers

The survey in Figure 2 focuses on the deeper NIR wavelength region, Figure 2a shows the NIR spectra of natural fibers (i.e. cotton, linen, wool and silk), while synthetic fibers are shown in Figure 2b. Similarities and differences between those samples can be observed in different NIR spectral ranges (i.e. the similarity between cotton and linen is high, both are similar to viscose, while synthetic fibers are very different from them). The spectral region around 1942 nm and 2100 nm is characteristic for cotton, the same was observed for linen and viscose, as both are cellulose-based fibers. In the case of polyamide, a special band was observed at 1713 nm, corresponding to the combined overtones of the NH bond from the amide functional group. This band was also observed in wool and silk (due to the amide functional groups from proteins). This demonstrates that the classification capability of NIR is extremely effective.

Figure 3: Average raw reflectance spectra of the analyzed textile samples

Figure 3 shows the spectral information for textiles of animal origin, interesting bands are visible between 1540 and 1580 nm, related to NH groups. Absorption in the region around 1900 nm, the combination of O–H stretching and H–OH bending vibrations of the hydroxyl group from water is shown. The overtones of the CH stretching of protein and lipid side chains are at around 1730 nm, while the range from 2000 to 2500 nm will provide information on amino acid composition and animal keratin fibers - a characteristic range. In terms of chemical composition, the combined bands of man-made textiles (i.e. polyester, acetate, viscose, nylon, etc.) can be associated with C–H groups, methyl (CH3) and ester (CO–O) bonds. Animal-derived textiles will concentrate the spectral information with larger wavelength regions.

The above surveys all yielded very high R2 values ranging from 0.97 to 0.99, which is sufficient to demonstrate that the NIR spectroscopy method is an extremely reliable choice for this application.

APPLICATION

• Sorting and Classification: NIR spectroscopy is used to sort textiles into different categories based on fiber type and composition. This improves the efficiency of the recycling process by ensuring that materials are correctly identified and processed.
• Quality Control: During the recycling process, NIR spectroscopy can monitor the quality of the recycled material, ensuring that it meets the desired specifications and standards.
• Material Recovery: By identifying and quantifying different fiber types, NIR spectroscopy helps in recovering valuable materials from textile waste, contributing to a circular economy and reducing the environmental impact of textile disposal.

RECOMMENDED EQUIPMENT

For applications using Near-infrared spectroscopy (NIRS), INTINS offers the Ocean Optics NIRQuest+ spectrometer product line with high sensitivity and good resolution from 2.8 nm to 6.48 nm. NIRQuest+ is available in three versions covering different wavelengths from 900-2500 nm.

NIRQuest+ can measure very small changes in absorbed signal, which are typical of harmonic overtones in the NIR. Its higher sensitivity enables better measurement accuracy, particularly in low-light conditions. Also, at longer wavelengths, NIRQuest+ measures diffuse reflection at low noise levels, resulting in cleaner spectra at wavelengths where distinct spectral features can be difficult to capture. With its compact, handheld design, NIRQuest+ can be easily and conveniently integrated into industrial systems in manufacturing plants to directly collect spectra right on the production line. Therefore, this is also a plus point for our system.

At the same time, we can provide other accompanying optical devices such as fiber, light sources... suitable depending on the customer's application needs.

CONCLUTION

NIR spectroscopy offers a robust and efficient solution for textile recycling, enabling rapid, non-destructive analysis and accurate sorting of materials. Its ability to handle diverse textile compositions and high-throughput analysis makes it a valuable tool in advancing sustainable textile recycling practices. By integrating NIR spectroscopy into recycling processes, the industry can enhance material recovery, improve recycling efficiency, and contribute to environmental sustainability.