Real-time Deconvolution of Complex Chemical Matrices
Abstract
Distributed chemical sensors have the potential to transform the way in which we obtain and process chemical information. Rapid transduction of chemical signals into actionable insights is critical to the development of new knowledge. For example, on-site and timely quantification of complex chemical matrices underpin our ability to monitor the fate of emerging contaminants, to create accurate models of transport phenomena, and to formulate well-informed plans of treatments. Electronic sensing arrays (Òelectronic tonguesÓ) that detect molecular interactions at solid-liquid interfaces are, however, prone to instability in real-world conditions. Alternatively, this proposal focuses on interrogating interactions at liquid-liquid interfaces. Specifically, simultaneous monitoring of the dynamic interfacial tensions at multiple interfaces will be used to generate chemical fingerprints without degradation in repeated measurements. There are two fundamental challenges preventing this progress toward all-liquid sensors. The first challenge is that current methods of measuring interfacial tensions do not provide high dimensional data. That is, only a single interface could be measured at once. This limitation slows the speed with which hypotheses can be tested and prevents a possibility for high-throughput array-based sensors. The second challenge is that only surface-active analytes are detectable by traditional methods. This shortcoming also leads to the inability to differentiate among the many types of surfactants. This proposal addresses both of these issues. The goals of this proposal are thus (1) to understand and control behaviors of emerging contaminants at liquid-liquid interfaces and (2) to leverage this knowledge for a high-throughput deconvolution of complex chemical matrices. The core innovations are the use of all-liquid responsive complex emulsions (RCEs) to simultaneously probe multiple interfaces and machine-learning algorithms to classify unknown analytes in real-time. This proposal substantially builds on the PIÕs previous results through three interrelated tasks. In Task 1, the effects of cross-reactive recognition units on interfacial tensions will be validated. In particular, the hypothesis that adsorption and desorption kinetics govern the fidelity of analyte-selector interactions will be tested. Additionally, the role of conformational transformation of polymeric selectors on dynamic interfacial tensions will be verified. Task 2 will focus on the high-throughput method to generate chemical fingerprints using responsive complex emulsions (RCEs). Multiplexed and simultaneous measurements will be combined to classify analytes in synthetic groundwater. In Task 3, we propose an automatic classification of real-time data using machine-learning algorithms for continuous operation. Ultra-sensitive, array-based sensors would impact a range of application of interest to the Army. This proposal, while providing a novel mode of array-based sensing, will also lead to new fundamental knowledge of how analytes behave at complex interfaces and how geochemical parameters affect interfacial behaviors.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 25, 2021
- Source ID
- W911NF2110310
Entities
People
- Suchol Savagatrup
Organizations
- Army Contracting Command
- United States Army
- University of Arizona