Wearable capacitors in leaves and machine learning for real-time analysis of multiple plant physiological parameters

Abstract

Physiological parameters such as water loss and photosynthesis rate are biomarkers of plant health and, therefore, show a correlation with microclimate and harvest productivity. As a consequence, the analysis of these parameters covers important implications for the environment and Agriculture. As a result of the newest revolution in the area of chemical sensors, the use of wearable electrodes on leafs is a promising alternative for the remote and real-time monitoring of physiological parameters of plants with high sensibility. The biggest challenges facing this area are adherence, biocompatibility, and sensitivity of the sensor for long-term assays. It is also worth noting simultaneous variations on two or more parameters in response to a single stress factor are observed. As limitations, the platforms described in the literature for this application have not encompassed reproducibility or accuracy tests, as well as they have considered just the variation of single parameters. Thus, this project is intended to develop a new flexible electrochemical sensor for the simultaneous determination of multiple parameters (multidetermination), physiological of plants and of microclimate, based on biocapacitance measurements using a portable instrumentation and smartphone. The parameters, associated with the water stress of the soybean, will be individually determined through machine learning supervised models. Due to its rapid growth, ease of cultivation, high water loss rate (roughly 50.0% w/w for 100 min), and economic relevance for the Brazil's trade balance, soy will be used as study model. Monitoring of soybean leaves will be performed by analyzing the photosynthesis rate, water loss, temperature, and humidity from a single impedance spectrum. Nickel (Ni) thin films coated with gold nanoparticles will be fixed onto the leaves using double-sided tape and will act as polarizable electrodes for bioimpedance measurements. Such films will be obtained by photolithography and electrodeposition, reproducible methods compatible with mass production. This new sensor will exhibit anti-fouling properties due to the presence of holes in the Ni film (mesh) with specific dimensions and density. Nanoparticles will be used to increase the conductivity of the electrodes and, thus the sensitivity of the capacitive sensing method. As it has already been done in our group, a smartphone will be used to control the potentiostat, treat the data, and display the final result on your screen. This so-called sample-to-answer system enables the wireless transmission of results and eliminates the data processing step by the operator, which is commonly laborious, particularly in the cases of multivariate data.The platform is expected to be a promising alternative toward point-of-use analysis with regard to real-time and high-sensitivity monitoring of different parameters that affect soy health, with innovative implications for scientific research (e.g., nanotoxicology) and Agriculture 4.0. The candidate will have the necessary technical support and infrastructure from the Brazilian Nanotechnology National Laboratory (LNNano) of the Brazilian Center for Research in Energy and Materials (CNPEM), which covers all the steps proposed herein. Prior to the section of objectives and expected results, this project shows preliminary results that indicate that our sensor may be promising for the analysis of leaf water loss. The main goals proposed herein are: sensor prototyping, soybean cultivation, studies of analytical performance, adhesion and biocompatibility of the electrodes, multi-determinations for monitoring soybean leaves and in-situ applications. (AU)