Analysis of brominated flame retardants in plastics by infrared reflectance spectroscopy

Abstract

Flame retardants are added to plastics to make the use of plastic products safer. However, brominated flame retardants (BFR) are a source of pollution to the environment, and bioaccumulation of BFRs affects the health of the human beings. For this reason, in the recycling process, waste plastics containing brominated flame retardants (BFR) have to be recycled separately. This thesis focuses on extracting the relation between BFR concentration in plastics and the infrared spectral information. Distinction of waste plastics containing BFR with concentration higher than a specified threshold is also considered.

To achieve its goal, this thesis made use of infrared reflectance spectroscopy (IRS). Samples consisting both of pure polymers doped with different type of BFRs and waste plastics were used in the thesis. Near-infrared (NIR) and mid-infrared (MIR) spectra are acquired from the samples. For the plastics made of pure polymers, partial least squares regression (PLS) and variable selection methods were applied to the spectra to build prediction models. The model with the minimum prediction error was obtained using the MIR range. Classification models using the k-nearest neighbors classifier (k-NN) was built using the MIR spectra of waste plastics to distinguish plastics containing higher BFR concentration than a specific threshold, reaching an accuracy of 81%.

The results indicate that it is possible to quantify BFRs in pure polymers doped with one type of BFR using IR spectroscopy. The classification model shows that it can identify the plastics containing a high level of BFR concentration.