A Review of Online Real-Time Process Analyses of Melt-State Polymer Using the Near-Infrared Spectroscopy and Chemometrics

Abstract

This review describes the online process monitoring of the melt-state polymer by near-infrared (NIR) spectroscopy and chemometrics. The spectra of linear low-density polyethylene (LLDPE), random polypropylene (RPP), block polypropylene (BPP), and ethylene-vinylacetate (EVA) copolymer in melt states measured by an online monitoring system with a Fourier transform near-infrared spectrometer are discussed. The calibration models for the density of LLDPE, the content of ethylene copolymers in RPP and BPP, and vinylacetate concentration (VA) in EVA copolymers using partial least squares regression are reported. The continuous monitoring of the LLDPE density at the real plant is described as an example of online monitoring using NIR spectroscopy and chemometrics. For the precise prediction of VA in EVA, a combination method using regression and discrimination was inducted. Three compensation methods for the effect of the temperature change in the RPP and BPP samples are shown. Conventional calibration transfer methods are introduced, and a practical calibration transfer method using two samples and its performance are reported using BPP and RPP spectra. Moreover, the possibility of a calibration correction method using one sample for the realization of long-term traceability is indicated by the example of the relocation.