Fluid Characterisation Using Microfluidic Impedance Spectroscopy for Remote Health Monitoring

Abstract

Remote health monitoring (RHM) will play a crucial role in the development of preventative and personalised healthcare, which must be adopted in the near future to adapt to the globally growing and ageing population. The ability to continuously monitor patients will enable more accurate diagnosis, real-time disease tracking and better patient outcomes. Continuous monitoring could also provide performance advantages to athletes and wellness benefits to all. Sweat is an important and useful biofluid that contains a wealth of information about a person’s physical and mental state. Sweat can be sampled non-invasively, making it an excellent candidate for RHM. The availability of a low-cost, portable device capable of continually determining the concentration of multiple key biomarkers in sweat, including electrolytes and metabolites, is a substantial hurdle for RHM. This is an active area of research, and most sensors in the literature use either colorimetric or electrochemical detection. Colorimetric sensors use the colour change of a sensing element. The user must, for example, take a photograph on their mobile phone, making the sensing non-continuous and adding a potential source of error. Electrochemical sensors require specific chemicals that undergo a redox reaction with each target biomarker. Development of such chemicals can be complex, and their cost and limited shelf-life and working lifetime can be prohibitive. Therefore, despite many sweat sensors being under development in the literature, currently no sensors capable of continuously monitoring the concentration of multiple sweat analytes are available on the market. In contrast, impedance spectroscopy is capable of directly probing the electrical properties of the fluid, removing the intermediate steps inherent in colorimetric and electrochemical methods. In this work, a microfluidic device is proposed with integrated interdigitated electrodes produced using aerosol-jet printing, a novel additive manufacturing technique. It is shown that such devices can be used to characterise aqueous ionic solutions by using impedance spectroscopy to directly measure their electrical properties. Impedance spectroscopy is a very well-established technique for the characterisation of solids but is less commonly used for liquids. Here the development of a novel analysis technique enables impedance spectroscopy to be utilised to accurately determine the concentration of an aqueous ionic solution. An effective capacitance is calculated from the imaginary component of the impedance, and is found to contain an inflection point. The frequency of this turning point (the turning point frequency) is found to be a key and highly useful parameter in the characterisation of ionic solutions. It is strongly linearly correlated with the ionic species concentration, and is also dependent on the species present. Additionally, unlike the raw impedance values, it is consistent across samples and independent of many aspects of the electrode geometry. Therefore it is used for accurate fluid characterisation based purely on the electrical properties of the test solution, removing the intermediate steps required by optical and electrochemical techniques. Solutions of NaCl, KCl, MgCl2 and CaCl2 (important analytes in sweat) with concentrations from 0.5 mM to 2.5 mM are measured and their concentrations are determined using the turning point frequency method with a success rate of 90% and an average relative uncertainty of 5.7%. The frequency-dependent electrical behaviour of ionic solutions are explained by the formation of electrical double layers at the electrode surfaces, driven by the electric field applied by the impedance analyser. This requires the movement of ions in the solution, so the formation time is dependent on the ionic mobility (a function of temperature and the hydrated ion radius) and ionic concentration (which determines the amount of ions within a certain distance of the electrode surface). At high frequencies, insufficient time is available for Stern layer formation, leaving only the ‘normal’ dielectric component, and at low frequencies there is a high degree of Stern layer formation, giving a high total capacitance due to the small gap between the electrode surface and Stern layer. The turning point occurs when there is partial Stern layer formation. Simulations based on this phenomenon, the parallel plate capacitance equation and ionic mobility are produced, and the calculated capacitance as a function of frequency is shown to approximately match the experimental data. Solutions of NaCl, KCl, MgCl2 and CaCl2 with concentrations spanning four orders of magnitude are tested, and the lower detection limit for each is found to be ∼0.02 mM, which is below their typical concentrations in sweat. The effect of parameters such as electrode geometry and temperature are analysed using experimental data and simulations. Impedance spectroscopy is found to have poor selectivity – it is shown using machine learning techniques that ionic species and concentration have identical effects on the measured spectra, and therefore determining both the species and concentration of an unknown solution is not possible using impedance spectroscopy alone. To introduce selectivity, the use of ion-selective membranes is proposed, and proof-of-concept is shown for NaCl and KCl. Therefore the vital components of a device capable of monitoring the concentration of electrolytes in sweat using impedance spectroscopy for RHM applications are demonstrated in this work.