Quantification of poisons for Ziegler Natta catalysts and effects on the production of polypropylene by gas chromatographic with simultaneous detection: Pulsed discharge helium ionization, mass spectrometry and flame ionization

Abstract

[EN] This article describes a new simultaneous method for the analysis of sulfur-type poisons, hydrocarbons and permanent gases affecting the productivity of the Ziegler Natta catalyst during the synthesis of polypropylene on an industrial scale in a fluidized-bed reactor. The identification was achieved employing a configuration of the seven-valve chromatographic system, with events at different times, allowing distribution of the sample through multiple columns, and finally reaching the helium ionization detectors of pulsed discharge, flame ionization and mass spectrometry. The results obtained show a good precision of the method used because the variability was less than 1.02% in area and 0.49% in retention time for short term precision and longer term precision. The quantification of these species was successful after performing the calibration curve with the dynamic mixer showing an r(2) higher than 0.9945 and excellent linearity. The lowest LOD value was 0.01 mg kg(-1) for carbonyl sulphide, hydrogen sulfide, ethylmercaptan and propylmercaptan and the lowest LOQ was 0.03 mg kg(-1) for hydrogen sulfide. The highest LOD and LOQ values were for oxygen and carbon dioxide with 0.40 and 0.93 mg kg(-1) respectively. With this configuration, the correlation of data between the three detectors was simplified, having almost identical retention times for the analytes studied.

The poisons detected and quantified in the samples were: hydrogen sulfide (0.1-0.5 mg kg(-1)), carbonyl sulphide (0.012-0.06 mg kg(-1)), carbon disulphide (0.04-0.22 mg kg(-1)), methylmercaptan (0.12-12.51 mg kg(-1)), ethylmercaptan (0.9-5.5 mg kg(-1)), carbon dioxide (0.10-3.0 mg kg(-1)), oxygen (0.55-6.1 mg kg(-1)), acetylene (0.15-3.5 mg kg(-1)) and methylacetylene (0.04-0.2 mg kg(-1)). The productivity losses were between 5 and 22%. (C) 2019 Elsevier B.V. All rights reserved.