PEFTEC 2017 - Abstract

Abstract Title: Exhaustive TD-GC-MS/FID as first line selection tool for low VOC PP development
Abstract Type: Oral
Presenter Name: Dr Ward DAutry
Co-authors:Mr Gerard Kwakkenbos
Mr Peter Tummers
Mr Peter Degenhart
Company/Organisation: SABIC
Session Choice: Analytical Techniques: Chromatography and Separations

Abstract Information :

In automotive, construction or consumer good markets, the release of volatile organic compounds (VOCs) from thermoplastic materials is of crucial importance. Understanding, controlling and testing polymer materials on VOCs is driven by public health concerns and regulatory activities. Therefore, various analytical methods are available for the analysis of VOC emissions from polymer materials. One of these methods is direct TD-GC-MS/FID, which is a widely used technique to measure VOC emissions from materials intended for the abovementioned markets. However, when applying TD-GC-MS/FID in R&D programs in search for low VOC PP materials, its relative standard deviation (RSD) values exceeded 20 %. This large RSD hampers the selection of best performing, low VOC PP grades.

A project was set up to identify the major sources of variability upon measuring PP emissions using TD-GC-MS/FID. Five key factors potentially influencing precision were defined and investigated: (i) exhaustiveness of emission, (ii) random blank responses, (iii) user dependent data processing, (iv) sample amount and (v) desorption flows through the TD tube and trap.

The desorption temperature was increased well above the melting temperature of PP to enhance the exhaustiveness of emission. Next to that, cryomilling the samples prior to analysis was investigated. A more stringent procedure with respect to blank analysis was implemented for neutralizing random blank response. A data processing algorithm was programmed in MATLAB that offers fully automated baseline correction, blank subtraction, integration, calculation of emission values per carbon cluster and report table output. This tool provides a very reproducible way of user independent data processing. Additionally, four TD methods were investigated, in which the sample amount and desorption flows were varied. The F-test statistic was performed on replicate results to draw conclusions about statistical significance on observed variabilities in emission response. This work demonstrates that more exhaustive emission conditions, more rigorous blank corrections and using our smart data processing algorithm demonstrated the largest positive effect on the variability. In tangible figures, the intra-lab reproducibility decreased from 20% to 5% RSD on the total emission of a PP reference sample.

Examination of the additional method parameters showed an adverse effect on variability (up to 50% RSD) when reducing the sample amount, whereas cryomilling or varying desorption flow parameters showed no effect. In conclusion, more exhaustive conditions and improved automated data processing provide an improved first line emission test in R&D phase of PP development, enabling a more reliable pre-selection of PP grades to be submitted for accredited emission testing.