Quantitative analysis of time-resolved FTIR specta : steps towards full automation

Abstract

Inline, time-resolved FTIR spectra are commonly recorded after completion of the experiments. The abilities and versatility of FTIR spectroscopy can, however, also be utilised in the in situ quantification of absorbing mixtures. Recent developments, in the laboratory where this investigation was conducted, demands the inline quantification of PTFE pyrolysis products for process control purposes. This investigation is primarily focused on the development of a procedure and software capable of processing, fitting and quantifying real-time, time-resolved spectra. Processing methods were evaluated with respect or improvement in SNR, smoothing and baseline tracking of infrared spectra. Execution speed was also considered due the need for real-time analysis. The asymmetric least squares method proved to be the optimal choice with respect to the mentioned criteria. An asymmetric lineshape fitting function together with a Levenberg-Marquardt nonlinear solving function was introduced to represent pure component spectra mathematically. A method for quantitative analysis by means of solving a linear set of equations was developed. The software was implemented on the batch pyrolysis of PTFE pyrolysis as test case. Experiments were conducted to obtain sufficient samples of the components such that FTIR spectra could be captured. Infrared spectra of the perfluorobutenes were experimentally determined. These spectra could not be found in the available literature and are deemed to be novel. The ability of the software to perform real-time quantification of the PTFE pyrolysis stream was demonstrated over a range of experimental conditions spanning the temperature range 650 ºC to 850 ºC, and pressures from <1kPa to 70 kPa.