Double-beam near-infrared spectroscopy to correct light source drift in aqueous glucose solution experiments

Abstract

Currently, a key problem in near-infrared (NIR) non-invasive blood glucose sensing is that the influence of the background changes during measurements, which restricts the effective extraction of unique glucose information. As the influence of the background can only be reduced by various correction techniques, standard sample correction and reference beam designs are investigated in this paper. First, the basic principles and preconditions for the background correction of single-beam measurements with reference correction and double-beam measurements with double detectors are analyzed. Second, the signal-to-noise ratio (SNR) in different measurement modes is evaluated. Third, single-day and multi-day experiments are conducted on aqueous glucose solution. The results are as follows: (1) short time interval between measurements of the sample and reference is the prerequisite for reference correction measurement with single-beam. The double-beam measurement has no special requirement for the SNR and the sampling time interval. (2) The double-beam design is more effective at eliminating the time-dependent variations. The SNR of the single-beam measurement can be improved by correcting the reference sample, which is also verified by the 2D correlation spectroscopy analysis under the perturbation of time. (3) The predictive ability of the partial least squares (PLS) model based on double-beam measurement is the best for both single-day experiments and multi-day experiments, which is consistent with results of the SNR test. Both the reference sample correction and the reference beam design can effectively reduce the light source drift and the double-beam design is more appropriate for multi-day or long-term experiments.