Colorimetric nutrient sensing in aquatic ecosystems using perovskite and metal oxide photodetectors

Abstract

Efficient sensors have become a crucial and indispensable part of today's technological landscape. However, nutrient monitoring in the aquatic environment is primarily dependent on the infrequent and discrete collection of water samples followed by analytical analysis in the laboratory. Real-time monitoring of nutrient dynamics in complex aquatic ecosystems has the potential to provide transformative insight into the effects of these nutrients on complex natural environments, particularly in addressing the global issue of harmful algal blooms. Harmful algal blooms can contaminate water supply, create dead zones, and a contributing factor for their occurrence is the presence of excess nutrients, specifically dissolved nitrate, nitrite, and phosphates in water. The existing advanced technologies for nutrient quantification in aquatic ecosystems are constrained by their single-use, unimodal feature, or the requirement of frequent replacements of reagents. In addition, the few field-deployable nutrient sensors with real-time data reporting, such as some phosphorus sensors, can cost up to $30k USD, strongly restricting the number of nutrient sensors that can be deployed. To address this challenge, aquatic researchers need real-time, high-frequency, low-cost nutrient sensors with low detection limits that can be embedded in existing smart sensor networks that monitor suites of other water variables. Driven by the pressing need to predict harmful algal blooms like those covering Lough Neagh, the largest freshwater lake in Northern Ireland, an international collaboration with researchers in the United States, the Republic of Ireland, and Northern Ireland aims to prototype a microbe-assisted colorimetric nutrient detection system to detect multiple forms of nitrogen. Dr.Panagiotis Manesiotis (Chemistry and Chemical Engineering, Queen’s University Belfast, Northern Ireland) and Dr.Margaret McCaul (School of Chemical Science, Dublin City University, Republic of Ireland) contribute selective molecular imprinted polymers to capture nitrogen products and existing field-tested microfluidic colorimetric systems, respectively. In addition to the detection module, research conducted within the Sawyer group through benchtop experiments using bacteria to change nitrogen products to nitrites is translated into a bacterial conversion module for integration into the system.The primary work presented in this thesis is the fabrication of perovskite photodetectors and the study of zinc oxide device structures to optimize the performance of the detector module. Variation in the concentration of dissolved nutrients in water samples can be monitored by a corresponding change in detector current, which reflects changes in absorbance within the colorimetric detection system. The material approach to detection improvement includes fabrication of an organic-inorganic perovskite active layer chosen for its tunable bandgap, high absorption, and cost-efficient fabrication methods. The bandgap of the mixed halide perovskite thin film is tailored to match the peak absorption spectra of the colorimetric-based nitrite detection system. The active layer thickness and composition are tuned to optimize the absorption and photocurrent-to-dark-current ratio. The dark current and photocurrent of the fabricated MAPbBr3 photodetector with a MABr top layer is 17.4 nA and 12.4 µA, measured over a 12 V bias, and the responsivity is 4.7 mA/W under a 528 nm illumination. Additionally, the perovskite photodetector is also suitable for operation at 370 nm, making it capable of detecting phosphate using molybdovanadophosphoric acid ‘yellow’ colorimetric reagents. The device structure approach to detection improvement is to explore a capacitive model to optimize the recovery time of the zinc oxide photodetector while maintaining its high UV photoresponsivity. Even under a very small bias of 3.3 V across the device, the recovery time is observed to improve from ~ 16.846 s to ~ 0.146 s with a 10 V backgate bias. An alternative approach to improve the recovery time of the ZnO detector is to optimize the anneal recipe and environment, with a longer and controlled anneal, the fall time significantly increases from ~ 127.6 s to ~ 20.3 s for a 3.5 V bias measurement. The perovskite photodetectors are developed for the colorimetric detection of nitrite and phosphate, either directly present in the sample or generated through the bacterial conversion module. The microbe-assisted conversion module expands the capability of our sensing system to detect various forms of nitrogen compounds, eliminating the need for multiple sensors for each nitrogen species and reducing cost and complexity. To evaluate the responsiveness of the MABr-coated MAPbBr2.92Cl0.08 photodetector, nitrite standards of up to 15.625 µM were reacted with the Griess assay to generate varying colors, and the detector could successfully process the signals, producing 1.79 µA to 59.97 nA across the blank to 250 µM nitrite concentration. Additional insights into the detector's behavior at lower nitrite concentrations can be explained from the performance of the MAPbBr2.89Cl0.11 detector with the simulated nitrite standards. The lighter colors emulate concentrations of nitrite much less than 15.625 µM, and the output trend of the photodetector current indicates that the photodetector is capable of responding to concentrations much weaker than 15.625 µM. The system successfully achieved a detection limit sufficiently below the EPA's maximum contaminant level goal of 71.4 µM of nitrite for drinking water. Similarly, the current output of the simulated phosphate assays confirms the detector's phosphate-sensing capabilities, producing 420 nA to 6 nA across the blank to the highest simulated concentration. Furthermore, the zinc oxide detector was also integrated into the simulated phosphate detection system, producing 121 µA to 730 nA across the blank to the highest simulated concentration. Compared to the MABr-coated MAPbBr2.92Cl0.08 photodetector, the alternative ZnO system produced a ~ 288 times increase in photocurrent for DI water and ~ 121 times rise for the darkest (simulated highest concentration) sample. The current bench-top system can be further improved to make a portable real-time nutrient monitoring and early warning system for nutrient analysis in water.