Development of Photonic Multi-Sensing Systems Based on Molecular Gates Biorecognition and Plasmonic Sensors: The PHOTONGATE Project.

This paper presents the concept of a novel adaptable sensing solution currently being developed under the EU Commission-founded PHOTONGATE project. This concept will allow for the quantification of multiple analytes of the same or different nature (chemicals, metals, bacteria, etc.) in a single test with levels of sensitivity and selectivity at/or over those offered by current solutions. PHOTONGATE relies on two core technologies: a biochemical technology (molecular gates), which will confer the specificity and, therefore, the capability to be adaptable to the analyte of interest, and which, combined with porous substrates, will increase the sensitivity, and a photonic technology based on localized surface plasmonic resonance (LSPR) structures that serve as transducers for light interaction. Both technologies are in the micron range, facilitating the integration of multiple sensors within a small area (mm2). The concept will be developed for its application in health diagnosis and food safety sectors. It is thought of as an easy-to-use modular concept, which will consist of the sensing module, mainly of a microfluidics cartridge that will house the photonic sensor, and a platform for fluidic handling, optical interrogation, and signal processing. The platform will include a new optical concept, which is fully European Union Made, avoiding optical fibers and expensive optical components.

Using staining as reference for spectral imaging: Its application for the development of an analytical method to predict the presence of bacterial biofilms.

At present, although spectral imaging is known to have a great potential to provide a massive amount of valuable information, the lack of reference methods remains as one of the bottlenecks to access the full capacity of this technique. This work aims to present a staining-based reference method with digital image treatment for spectral imaging, in order to propose a fast, efficient, contactless and non-invasive analytical method to predict the presence of biofilms. Spectral images of Pseudomonasaeruginosa biofilms formed on high density polyethylene coupons were acquired in visible and near infrared (vis-NIR) range between 400 and 1000 nm. Crystal violet staining served as a biofilm indicator, allowing the bacterial cells and the extracellular matrix to be marked on the coupon. Treated digital images of the stained biofilms were used as a reference. The size and pixels of the hyperspectral and digital images were scaled and matched to each other. Intensity color thresholds were used to differentiate the pixels associate to areas containing biofilms from those ones placed in biofilm-free areas. The model facultative Gram-negative bacterium, P. aeruginosa, which can form highly irregularly shaped and heterogeneous biofilm structures, was used to enhance the strength of the method, due to its inherent difficulties. The results showed that the areas with high and low intensities were modeled with good performance, but the moderate intensity areas (with potentially weak or nascent biofilms) were quite challenging. Image processing and artificial neural networks (ANN) methods were performed to overcome the issues resulted from biofilm heterogeneity, as well as to train the spectral data for biofilm predictions.