Rapid Fluorescence EEM spectra acquisition using super-cycle Hadamard-Transform Multiplexing

Hadamard-transform multiplexing has recently been applied to increasingly complex spectroscopic techniques. It had been shown that the data acquisition time for fluorescence Excitation-Emission-Matrix spectroscopy can be reduced by one or two orders of magnitude using Hadamard-Transform multiplexing of the excitation light using a programmable lightsource. In these previous studies, the data acquisition rate had been limited by the time it took to record an EEM, i.e. to complete one cycle of multiplexed excitation spectra. The extraction of chemical information, such as concentration and chemical identity, is then obtained from parallel factor (PARAFAC) analysis of the sequence of EEM spectra. In this contribution we increase the data acquisition rate by another order of magnitude, i.e. to the time it takes to record a single excitation spectrum. Our algorithm is entirely based on improved data processing, i.e. it can be applied to previously recorded Hadamard-Transform multiplexed data sets. The algorithm is based on three previously unexplored approaches: (1) we perform a PARAFAC multivariate analysis on the raw (multiplexed) data set (2) the PARAFAC loading vectors are obtained prior to obtaining the score vectors, (3) when loading vectors are difficult to obtain from the stack of EEM spectra, we instead use a rolling-average approach to considerably increase the number of spectra and the stability of the fit. Analysis of experimental data shows that fluorescence EEM spectra with 7 excitation wavelengths and over 1000 emission wavelengths can be obtained in less than 20 ms.