Modeling RP-1 fuel advanced distillation data using comprehensive two-dimensional gas chromatography coupled with time-of-flight mass spectrometry and partial least squares analysis

• Published in: Analytical and bioanalytical chemistry 2015, Vol.407 (1), p.321-330
• Main Authors: Kehimkar, Benjamin, Parsons, Brendon A., Hoggard, Jamin C., Billingsley, Matthew C., Bruno, Thomas J., Synovec, Robert E.
• Format: Article
• Language: English
• Subjects: Analysis; Analytical Chemistry; Aromatic compounds; Biochemistry; Biodiesel fuels; Calibration; Characterization and Evaluation of Materials; Chemical speciation; Chemistry; Chemistry and Materials Science; Chromatography; Crude oil; Data analysis; Distillation; Food Science; Fuels; Gas chromatography; Gasoline; Kerosene; Laboratory Medicine; Mass spectrometry; Mathematical models; Monitoring/Environmental Analysis; Multidimensional Chromatography; Physical properties; Research Paper; Scientific imaging; Separation; Two dimensional
• ISSN: 1618-2642; 1618-2650
• DOI: 10.1007/s00216-014-8233-6

Recent efforts in predicting rocket propulsion (RP-1) fuel performance through modeling put greater emphasis on obtaining detailed and accurate fuel properties, as well as elucidating the relationships between fuel compositions and their properties. Herein, we study multidimensional chromatographic data obtained by comprehensive two-dimensional gas chromatography combined with time-of-flight mass spectrometry (GC × GC–TOFMS) to analyze RP-1 fuels. For GC × GC separations, RTX-Wax (polar stationary phase) and RTX-1 (non-polar stationary phase) columns were implemented for the primary and secondary dimensions, respectively, to separate the chemical compound classes (alkanes, cycloalkanes, aromatics, etc.), providing a significant level of chemical compositional information. The GC × GC–TOFMS data were analyzed using partial least squares regression (PLS) chemometric analysis to model and predict advanced distillation curve (ADC) data for ten RP-1 fuels that were previously analyzed using the ADC method. The PLS modeling provides insight into the chemical species that impact the ADC data. The PLS modeling correlates compositional information found in the GC × GC–TOFMS chromatograms of each RP-1 fuel, and their respective ADC, and allows prediction of the ADC for each RP-1 fuel with good precision and accuracy. The root-mean-square error of calibration (RMSEC) ranged from 0.1 to 0.5 °C, and was typically below ∼0.2 °C, for the PLS calibration of the ADC modeling with GC × GC–TOFMS data, indicating a good fit of the model to the calibration data. Likewise, the predictive power of the overall method via PLS modeling was assessed using leave-one-out cross-validation (LOOCV) yielding root-mean-square error of cross-validation (RMSECV) ranging from 1.4 to 2.6 °C, and was typically below ∼2.0 °C, at each % distilled measurement point during the ADC analysis. Graphical Abstract The application of chemometrics to relate GC × GC–TOFMS chromatograms to advanced distillation curve (ADC) data allows for accurate prediction of distillation curves for kerosene rocket propellant (RP-1) fuels and contributes to the chemical characterization of fuel components and their influence on fuel performance.