Universal response model for a corona charged aerosol detector

• Published in: Journal of Chromatography A Vol. 1217; no. 47; pp. 7418 - 7427
• Main Authors: Hutchinson, Joseph P., Li, Jianfeng, Farrell, William, Groeber, Elizabeth, Szucs, Roman, Dicinoski, Greg, Haddad, Paul R.
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 19.11.2010; Amsterdam; New York: Elsevier; Elsevier
• Subjects: Aerosols; Aerosols - analysis; Analytical chemistry; Charged aerosol detector; Chemistry; Chromatographic methods and physical methods associated with chromatography; Chromatography, High Pressure Liquid - instrumentation; Concentration (composition); Corona; Detectors; Elution; Error analysis; Exact sciences and technology; Liquid chromatography; Mathematical models; Other chromatographic methods; Phase composition; Response factor; UHPLC; Universal detection; Model; Corona; Liquid chromatography; Response factor; Universal detection; UHPLC; Charged aerosol detector; Ultra performance liquid chromatography; Elution; Gradient; Chemical analysis; HPLC chromatography; Instrumental response; Non volatile compound; Detection; Flow injection; Corona charged aerosol detector
• ISSN: 0021-9673; 1873-3778
• DOI: 10.1016/j.chroma.2010.09.056

The universality of the response of the Corona Charged Aerosol Detector (CoronaCAD) has been investigated under flow-injection and gradient HPLC elution conditions. A three-dimensional model was developed which relates the CoronaCAD response to analyte concentration and the mobile phase composition used. The model was developed using the response of four probe analytes which displayed non-volatile behavior in the CoronaCAD and were soluble over a broad range of mobile phase compositions. The analyte concentrations ranged from 1 μg/mL to 1 mg/mL, and injection volumes corresponded to on-column amounts of 25 ng to 25 μg. Mobile phases used in the model were composed of 0–80% acetonitrile, mixed with complementary proportions of aqueous formic acid (0.1%, pH 2.6). An analyte set of 23 compounds possessing a wide range of physicochemical properties was selected for the purpose of evaluating the model. The predicted response was compared to the actual analyte response displayed by the detector and the efficacy of the model under flow-injection and gradient HPLC elution conditions was determined. The average error of the four analytes used to develop the model was 9.2% ( n = 176), while the errors under flow-injection and gradient HPLC elution conditions for the evaluation set of analytes were found to be 12.5% and 12.8%, respectively. Some analytes were excluded from the evaluation set due to considerations of volatility (boiling point <400 °C), charge and excessive retention on the column leading to elution outside the eluent range covered by the model. The two-part response model can be used to describe the relationship between response and analyte concentration and also to offer a correction for the non-linear detector response obtained with gradient HPLC for analytes which conform to the model, to provide insight into the factors affecting the CoronaCAD response for different analytes, and also as a means for accurately determining the concentration of unknown compounds when individual standards are not available for calibration.