A sample-effective calibration design for multiple components.

Area of Science:

• Analytical Chemistry
• Chemometrics

Background:

• Multivariate calibration requires efficient experimental designs for accurate chemical mixture analysis.
• Existing uniform designs can be computationally intensive and inflexible regarding the number of experimental points or components.

Purpose of the Study:

• To introduce a novel, computationally simple experimental design for mixture analysis in multivariate calibration.
• To assess the performance of this design in a UV-Vis spectroscopic experiment for quantifying multiple inorganic components.

Main Methods:

• The proposed design distributes experimental points uniformly within a concentration hypercube.
• It avoids fixed 'levels' and accommodates varying numbers of mixtures and components.
• Partial Least Squares (PLS) regression models were used to evaluate the design's performance.

Main Results:

• The new design achieved comparable or superior prediction accuracy compared to cyclic permutation and Kennard-Stone designs.
• It required a lower number of experimental points to achieve high accuracy.
• The design demonstrated effectiveness in a UV-Vis spectroscopic experiment for simultaneous quantification of four inorganic components.

Conclusions:

• The introduced experimental design is a computationally efficient and flexible method for mixture analysis.
• It offers a practical alternative for multivariate calibration, improving accuracy and reducing experimental effort.
• This approach is valuable for applications requiring simultaneous quantification in complex mixtures.