Chemometric Evaluation of Repeatability Using the Autocorrelation Method in High-Performance Liquid Chromatography...

Area of Science:

• Analytical Chemistry
• Chemometrics
• Chromatography

Background:

• Mixed random processes, including first-order autoregressive (AR(1)) and white noise, model analytical instrument baseline noise effectively.
• These noise models are crucial for estimating precision profiles in analytical measurements.
• Autocorrelation methods offer a potential approach for estimating noise parameters.

Purpose of the Study:

• To evaluate a novel autocorrelation method for estimating three noise parameters (two AR(1), one white noise) in High-Performance Liquid Chromatography (HPLC).
• To utilize these estimated parameters for calculating the precision profile of HPLC analyses.
• To assess the method's applicability in estriol determination using HPLC with UV detection (HPLC-UV).

Main Methods:

• Application of a chemometric autocorrelation method to estimate noise parameters from chromatographic baselines.
• Evaluation of the method's performance in repeatability assessments of estriol determination via HPLC-UV.
• Comparison of relative standard deviations (RSDs) from the autocorrelation method with actual repeated measurements.

Main Results:

• The autocorrelation method yielded a relative standard deviation (RSD) of 1.42% for 5.0 mg/L estriol, closely matching the 1.63% from actual repeated measurements (n=6).
• Theoretical RSDs calculated using the autocorrelation method fell within the 95% confidence intervals of empirical measurements.
• Noise parameters were successfully extracted from real chromatographic baselines.
• The instrumental detection limit for estriol was determined using the derived precision profile.

Conclusions:

• The autocorrelation method is a practical and effective technique for evaluating HPLC-UV analysis precision profiles.
• This method eliminates the need for repeated measurements of real samples, streamlining the precision assessment process.
• The study demonstrates the utility of noise parameter estimation for calculating precision profiles and detection limits.