Charged aerosol detection – use of the power function and robust calibration practices to achieve the best quantitative results

Significance of the topic

Charged aerosol detection (CAD) has become an indispensable tool in liquid chromatography for quantifying non-chromophoric and low-UV-absorbing analytes. Its near-universal response, broad dynamic range, and compatibility with gradient elution make CAD attractive in pharmaceutical, biochemical, and industrial QC laboratories. However, the inherent nonlinearity of CAD response over wide concentration ranges poses challenges for accurate quantitation and chromatographic performance assessment. The application of a power function (PF) to CAD signal processing can extend the quasi-linear range and simplify calibration and data analysis tasks.

Objectives and overview of the study

This technical note aims to demonstrate how to optimize CAD linearity by:

• Applying a power function exponent (PFV) to the raw CAD signal
• Identifying the optimal PFV for a given analyte class and quantitation range
• Establishing best practices for calibration curve fitting and quality assessment

Model compounds (sulfonamide drugs and their impurities) were used to illustrate the effect of PFV settings on detector response, peak shape, signal-to-noise ratio (S/N), and calibration accuracy. A comparison between experimental PFV settings and simulated values using Thermo Scientific™ Chromeleon™ CDS Power Law was conducted to streamline PF optimization.

Methodology and instrumentation

The study employed a Thermo Scientific™ charged aerosol detector interfaced to an UHPLC system, controlled by Chromeleon CDS software. Key elements included:

• Data acquisition at PFVexp values (1.0 and 1.4)
• Post-processing of recorded data with the Power Law feature in Chromeleon CDS to generate PFVcalc data (1.2, 1.4, and 1.6)
• Calibration over more than two orders of magnitude (1.56–200 ng on column) for sulfamerazine, sulfamethizole, and sulfadimethoxine
• Evaluation of linear least squares regression with and without 1/amount weighting, using residuals plots as the primary metric
• Use of Thermo Scientific guidance on PF theory and prior technical notes on CAD response characteristics and ELSD comparison

Main results and discussion

Response curves of non-volatile analytes exhibited sub-linear behavior (exponent b < 1) when PFV = 1.0, leading to up to 70 % deviation at low analyte levels. Incremental application of PFVcalc values demonstrated that:

• PFVcalc 1.4 substantially flattens the response curve, bringing exponent b closer to unity
• Linear calibration with 1/amount weighting at PFVcalc 1.4 yielded residuals within ±15 % across all levels above the LOQ
• Experimental verification at PFVexp 1.4 confirmed simulated results, with improved peak sharpness, reduced baseline drift, and higher, more reliable S/N values

Methods for distinguishing non-volatiles (optimal PFV 1.0–1.6) from semi-volatiles (avoid PFV ≠ 1.0) were outlined. The inability of a single PFV to perfectly linearize the full native dynamic range was noted, but a “best-fit” PFV can greatly simplify routine analysis.

Benefits and practical applications of the method

Optimizing PFV for CAD delivers multiple advantages:

• Enables accurate linear calibration curves, reducing reliance on higher-order polynomial or log–log models
• Improves quantitation precision at both low and high analyte concentrations
• Provides truer measurements of chromatographic metrics such as resolution and efficiency
• Yields more reliable limits of detection by aligning S/N measurements with linear response conditions
• Facilitates polymer mass distribution analysis and other applications relying on relative peak areas

Future trends and possibilities of application

Potential developments include:

• Automated PFV optimization algorithms integrated into instrument control software
• Extension of PF-based linearization to multi-detector configurations (e.g., CAD + ELSD)
• Machine-learning models to predict optimal PFV from compound physicochemical properties
• Standardized PF-based workflows for regulatory methods requiring robust calibration

Conclusion

Applying a power function exponent to CAD output is a highly effective strategy to expand the quasi-linear response range and simplify quantitative workflows. Through robust calibration practices—especially evaluation via residuals plots and appropriate weighting—analysts can identify an optimal PFV (typically 1.0–1.6 for non-volatiles) that supports linear calibration and enhances chromatographic performance metrics. The Chromeleon CDS Power Law feature streamlines PFV determination, enabling rapid translation of simulated settings to experimental operation.

References

• Gamache, P. H., ed. (2017). Charged aerosol detection for liquid chromatography and related separation techniques. John Wiley & Sons.
• Thermo Fisher Scientific. (2018). Technical Note 72806: Charged Aerosol Detection: Factors Affecting Uniform Analyte Response.
• Kiser, M. M., & Dolan, J. W. (2004). Selecting the best curve fit. LC GC North America, 22(2), 112–117.
• Almeida, A. M. d., et al. (2002). Linear regression for calibration lines revisited: weighting schemes for bioanalytical methods. Journal of Chromatography B, 774(2), 215–222.
• Dolan, J. W. (2009). Calibration curves, Part III: a different view. LC GC North America, 27(5), 392–400.
• Raposo, F. (2016). Evaluation of analytical calibration based on least-squares linear regression for instrumental techniques: a tutorial review. TrAC Trends in Analytical Chemistry, 77, 167–185.