An environmental sensor for detection of subsurface contamination

地下常见的污染物，如溶解重金属及其络合物、硝酸盐-亚硝酸盐等，由于其对人类、植物和动物的潜在影响，地质环境工程师非常关注。 通常用于检测地下污染物的方法包括从钻孔中采集土壤样品和从监测威尔斯中采集地下水样品，然后进行化学分析。这种常规取样和分析可能是昂贵的、耗时的，并且不能提供污染物迁移的真实的时间监测。本研究将结合物理化学、微波工程、人工智能等科学与工程领域的最新发展，以及PI近十年来的研究成果，开发一种用于地下污染监测与检测的环境传感器。 实验观测和理论分析表明，土壤-水系统的复介电常数随其物理化学性质的变化而变化。 因此，测量土壤和地下水的原位复介电常数的传感器指示地下的特性。 传感器测得的复介电常数数据将由人工神经网络处理，以定量和半定量描绘地下污染。 该传感器采用短路同轴线技术，具有快速、可靠、鲁棒性好、易于现场实施等优点。将开发基于人工神经网络的数据处理软件，将复介电常数与地下污染的类型和浓度联系起来。 在NSERC发现基金的支持下，环境传感器技术可能会被开发用于广泛的环境工程应用。

Contaminants commonly encountered in subsurface, such as dissolved heavy metals and their complexes, nitrate-nitrite, etc., are of great concern to geo-environmental engineers because of their potential impacts to human, plants and animals.  The methods commonly used to detect these contaminants in subsurface involve taking soil samples from boreholes and groundwater samples from monitoring wells and then performing chemical analyses. Such routine sampling and analysis can be costly, time consuming, and cannot provide real time monitoring of contaminant migration. In this study, an environmental sensor for subsurface contamination monitoring and detection will be developed based on the state-of-the-art development in science and engineering, including dielectrics, physical chemistry, microwave engineering and artificial intelligence, as well as on the PI's research in the past decade.  It has been evidenced from experimental observations and theoretical analyses that the frequency-dependent complex permittivity of soil-water systems varies with changes in their physical and chemical properties.  Therefore,  sensors that measure the in-situ complex permittivity of soil and groundwater indicate characteristics of the subsurface.  The complex permittivity data measured from the sensors will be processed by artificial neural networks for quantitative and semi-quantitative delineation of subsurface contamination.  The sensor adapts the short-circuit coaxial line technology, which is fast, reliable, robust and can be readily employed for in-situ implementation. Artificial neural networks based data processing software will be developed to link the complex permittivity with the type and concentrations of subsurface contamination.  With the support of NSERC Discovery Grant, an environmental sensor technology may be developed for broad environmental engineering applications.