REACT: chemometrics a predictive tool for substances on fabrics

Technology and Controls

Chemometrics allows complex experimental problems to be approached using mathematical and statistical tools, providing solutions to separate 'useful information' from the rest contained in the data, minimising time and costs. In the textile sector, a skilful use of this tool can improve end-of-life clothing sorting processes, and within the REACT project, it was exploited to detect the kind of finishing of acrylic used in outdoor applications.

The REACT project, funded by the European Horizon 2020 programme, aims to mechanically recycle acrylic fabrics from the production or end-of-life of awnings and umbrellas. The main issue addressed by the project is the removal of the substances on the fabrics, in particular the finishing present to ensure the required performance, to obtain a recycled material with high purity and without potentially hazardous substances.

To achieve this goal, the identification of the substances present on the fabric is crucial for the choice of the type of removal process to be applied. This characterisation is complex and time-consuming and requires a series of detailed analyses, some of which are destructive if these substances are not known, which add costs to the recycling process.

As part of the REACT project, Centrocot has developed a predictive method to respond to this challenge, using spectroscopic techniques currently employed in automatic sorting machines used to separate waste, which quickly identify the kind of fibre contained. Spectroscopic techniques identify the chemical characteristics of the whole fabric and are influenced by the quantity of material present; this poses the problem of recognising finishing and substances that are present in very small quantities in comparison to the fibre. Through chemometrics, only useful information can be extracted from the spectroscopic data to identify the type of chemicals present, excluding data from the fibre.

In the REACT project, a significant number of acrylic fabrics with different kinds of finishing were analysed by infrared spectroscopy: the kind of finishing and the quantity applied were known. This led to being able to ‘train’ the chemometric model with accurate information, such that the model was subsequently able to identify small differences between the various fabrics analysed and create accurate and precise predictive clusters. The model designed for acrylic textiles makes it possible to recognise the finishing present. This makes it possible to separate fabrics in real time according to the finishing types of present, and to allocate end-of-life fabrics to the appropriate treatment process. In addition, the same model can also be used to verify the effectiveness of finishing removal treatment. In other words, chemometrics becomes a core element of the REACT project: with it, it is possible to combine the effectiveness of treatments, the recognisability of the most problematic textiles and the reduction of time.

The method developed on acrylics can be extended to any type of fibre and chemical substance present; in fact, as chemometrics is a statistical method, it can be used with any type of data, substance and application, with a great variability of cases, training the model to the desired objective.

For more information, please contact:

  • Roberto Vannucci

Head of Multisectoral Research and Innovation


  • Daniele Piga, PhD

Multisectoral Research and Innovation

Author: Daniele Piga

Multisectoral Research and Innovation Dept.