Development of spreadsheet-based simulation to reduce number of experimental trials in separation with comprehensive two-dimensional gas chromatography

Abstract

Method development and optimization in comprehensive two- dimensional gas chromatography (GC´GC) is a process with the requirement to perform a large number of experiments maximizing the number of separated components in each sample. These are the skills which cannot be developed by students in a general lecture class; while, they can be achieved in a laboratory although with consumption of large amount of budget, energy and time. This thesis developed a spreadsheet in Microsoft Excel format to simulate a GC´GC retention time plot and a number of separated peaks for a given sample within a few milliseconds. Application of this approach waives the experimental expense in enhancing the optimization skills of students and can be performed within a short period of time such as in a few hours of a lecture class. The thesis demonstrated this advantage by distributing the spreadsheet simulation to the group of 25 students consisting of bachelor, master and PhD students. They were then instructed on how to use the spreadsheet simulation and practice with a given separation goal. In this program, the GC´GC parameters, which can be changed leading to different results, include column lengths, gas flow rate, initial oven temperature, temperature increase rate, and the first and the second columns. The students then performed the test to separate a model kerosene sample containing 641 compounds. They showed the average score of 12 out of 15 corresponding to the average number of separated peaks of 226. The students could achieve 96% of 235 separated peaks obtained from the ideally optimized condition for this sample. This could also be an indicative of sufficient trial-and error skills that the students could learn within the a few hours learning time compared with ~30 hours of the actual GC´GC experiments that each student needed to spent on learning such skills. In addition, the simulated results were in agreement with the experimental separation of an essential sample with the correlation coefficients (R2) between the simulated and experimented peak retention times ranging from 0.63 to 0.96.