Extension of Trilinear Decomposition Method with an Application to the Flow Probe Sensor

The trilinear decomposition algorithm (TLD) is a method for calibration of second-order instrumentation (e.g., LC-UV). This method, like the generalized rank annihilation method (GRAM), estimates the intrinsic profiles (e.g., spectra and chromatograms) of each component present in each sample by solving an eigenvector/eigenvalue problem. The relative concentration of each component between the samples is found by the least squares fitting of the intrinsic profiles to the instrument response of the samples. The advantage the TLD algorithm has over GRAM is the ability to analyze data from multiple samples simultaneously. The previously published algorithm provided unreliable calibration estimates when imaginary eigenvectors were included in the solution of the eigenproblem