Investigation of the interactions between analytes and stationary phases in gas chromatographic systems using simulation

Abstract

Volatile compounds are responsible for the odor of food, characterize their authenticity or potential health risks. One technique for investigating volatile compounds is gas chromatography. The method development is often resource-, time-, and cost-intensive but can be supported by computer simulation. Computational models are necessary, describing both the interaction of volatile compounds and the representation of the gas chromatographic system. As the models require corresponding data to describe and determine retention, the three presented investigations are concerned with determining and estimating this data. In the first study, a database of thermodynamic retention parameters was established for a variety of volatile compounds, including FAMEs, triglycerides, PAHs, and PCBs. Retention factors from isothermal measurements were determined for 900 substance stationary-phases-combinations, and parameters for common retention models (ABC model, K-centric model, thermodynamic model) were determined. In addition, available data from the literature was also included. A standardized approach for determining parameters was presented, and quality criteria for suitable retention parameters were established. The simulation of gas chromatographic separations using the retention parameters from the database was compared to real temperature-programmed measurements. In the second study, the relationship between measurable elution temperature and characteristic temperature was investigated. The characteristic temperature is the most important retention parameter in the “distribution-centric retention model” (K-centric model) according to Blumberg. Influences of the temperature program due to the starting temperature and the heating rate were examined. A computational model was established using the dataset, allowing an estimation of the characteristic temperature from simple temperature-programmed measurements. This extends the prediction range, especially for volatile compounds such as benzene derivatives, aldehydes and ketones, compared to previous estimation models. The prediction of retention times based on the regression model was demonstrated using the example of alcohols and phenones. In the third study, the ‘Linear Solvation Energy Relationship’ (LSER) model was used to estimate retention parameters usable for the simulation by LSER substance data. Two stationary phases were characterized. K-centric retention parameters were estimated for ca. 300 compounds, and the data were compared with parameters from isothermal measurements. Simulations of temperature-programmed GC separations using the retention parameters determined by LSER were compared with isothermal retention parameters and real measurements. The work is an important contribution for the simulation of complex GC systems like multidimensional GC (MDGC), comprehensive GC (GC×GC) or novel techniques such as spatial thermal gradient GC and furthermore for the development of auto-optimisation GC.