Hyperspectral Luminescence Imaging in Combination with Signal Deconvolution Enables Reliable Multi-Indicator-Based Chemical Sensing

Abstract

Although real-time monitoring of individual analytes using reversible optical chemical sensors (optodes) is well established, it remains a challenge in optical sensing to monitor multiple analyte concentrations simultaneously. Here, we present a novel sensing approach using hyperspectral imaging in combination with signal deconvolution of overlapping emission spectra of multiple luminescent indicator dyes, which facilitates multi-indicator-based chemical imaging. The deconvolution algorithm uses a linear combination model to describe the superimposed sensor signals and employs a sequential least-squares fit to determine the percent contribution of the individual indicator dyes to the total measured signal. As a proof of concept, we used the algorithm to analyze the measured response of an O₂ sensor composed of red-emitting Pd(II)/Pt(II) porphyrins and NIR-emitting Pd(II)/Pt(II) benzoporphyrins with different sensitivities. This facilitated chemical imaging of O₂ over a wide dynamic range (0-950 hPa) with a hyperspectral camera system (470-900 nm). The applicability of the novel method was demonstrated by imaging the O₂ distribution in the heterogeneous microenvironment around the roots of the aquatic plant Littorella uniflora. The presented approach of combining hyperspectral sensing with signal deconvolution is flexible and can easily be adapted for use of various multi-indicator- or even multianalyte-based optical sensors with different spectral characteristics, enabling high-resolution simultaneous imaging of multiple analytes.

Keywords: chemical imaging; chemometrics; hyperspectral imaging; image analysis; linear unmixing; planar optodes.