Please cite this article in press as: R. Jaramillo, F.L. Dorman, Retention time prediction in thermally modulated comprehen- sive two-dimensional gas chromatography: Correcting second dimension retention time modeling error, J. Chromatogr. A (2018), https://doi.org/10.1016/j.chroma.2018.10.054 ARTICLE IN PRESS G Model CHROMA-359778; No. of Pages 9 Journal of Chromatography A, xxx (2018) xxx–xxx Contents lists available at ScienceDirect Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma Retention time prediction in thermally modulated comprehensive two-dimensional gas chromatography: Correcting second dimension retention time modeling error Roman Jaramillo a , Frank L. Dorman b,∗ a Department of Chemistry, The Pennsylvania State University, 104 Chemistry Building, University Park, PA 16802, United States b Department of Biochemistry and Molecular Biology, The Pennsylvania State University, 107 Althouse Lab, University Park, PA 16802, United States a r t i c l e i n f o Article history: Received 20 April 2018 Received in revised form 25 October 2018 Accepted 29 October 2018 Available online xxx Keywords: GC×GC Retention time prediction Thermodynamic modeling Method development a b s t r a c t Thermodynamic retention modeling to a thermally modulated comprehensive two-dimensional gas chromatography (GC × GC) system run under constant flow is performed. Significant errors in modeled second dimension retention time (t r,2 ) were observed, in line with past work on thermally modulated GC × GC modeling. A comprehensive study of t r,2 modeling error for alkane separations across a wide range of heating ramp rates and carrier gas flow rates was performed. Modeling errors were found to be systematic and a function of analyte elution temperature and mobile phase velocity. A model to account for these systematic errors was generated, and associated coefficients were determined which reduced average t r,2 retention time error in 144 hydrocarbon separations by an order of magnitude resulting in significant improvement in prediction accuracy. The model was used to correct the separation of 139 Grob mix analyte separations, providing an average t r,2 modeling error of 0.030 ± 0.022 s. The model successfully predicted the separation of n-alkanes on a longer second dimension column configuration. © 2018 Elsevier B.V. All rights reserved. 1. Introduction GCxGC has become an indispensable tool in the analysis of highly complex mixtures, including petrochemicals, environmen- tal waters and many other applications [1]. In GCxGC, two columns with orthogonal selectivity mechanisms are coupled to comple- ment one another and separate analytes in complex mixtures that neither column could resolve on its own. The technique’s increased peak capacity and ability to cryo focus chromatographic peaks, when using cryo-modulation, provides superior resolution and sen- sitivity for the identification and quantification of thousands of compounds present at trace levels [2]. Several approaches to modeling comprehensive two dimen- sional gas chromatography (GC × GC) separations have been explored. Quantitative structure-retention relationship (QSRR) has been examined as a way to determine retention behavior for ana- lytes based off the retention data of analytes with similar molecular structure [3,4]. D’Archivio et al. predicted the elution of polychlori- nated biphenyls (PCBs) using the retention data of select congeners [3]. They noted that prediction of second dimension retention times ∗ Corresponding author. E-mail address: fld3@psu.edu (F.L. Dorman). were suboptimal and that the application of QSRR only allowed for the prediction of congener retention times under operating conditions identical to those of the reference retention data, not allowing for the modeling of additional temperature program- ming, flow rates, etc. Second dimension retention indexes have also been applied to retention modeling [5–7]. Zhao et al. gen- erated a retention map of the second dimension using n-alkane retention indices [5]. Zhao modeled the second dimension reten- tion indices of MegaMix compounds based off their experimental retention time and elution temperature. This approach determines the retention indices of analytes, and general elution space on a chromatographic plane relative to reference n-alkanes, not analyte retention times under various operating conditions. It is primarily meant to be used in tandem with spectral library matching-based compound identification. Considerable attention has been given to modeling the thermodynamics of chromatographic separation [8–13]. Gas chromatography can be fundamentally characterized as a ther- modynamic and kinetic process. Thermodynamic chromatographic modeling relies on iteratively calculating the movement of analytes through GC columns to determine their retention time. These pro- cesses have been described extensively in the literature [14–20]. A variety of thermodynamic indices have been developed to char- acterize the solvation of analytes into and out of the stationary https://doi.org/10.1016/j.chroma.2018.10.054 0021-9673/© 2018 Elsevier B.V. All rights reserved.