Wireless and untethered electrochemical oxygen sensor for respiratory gas exchange analysis

Abstract

The objective of this work is to create a novel medical sensor for monitoring the concentration of oxygen in exhaled human breath. The ultimate goal is to combine this work with a contiguous research project to deliver a sensor that measures several breath constituents, including CO2, and end-tidal volume. Current respiratory monitoring systems suffer from limitations including high cost and power consumption, large size and usually require breath to be drawn through sampling tubes. These factors lead to a lack of portability in current monitoring systems. Further motivation for a portable oxygen sensor is the importance of oxygen as a breath marker for prognostics and ambulatory monitoring. Electrochemical sensing techniques are chosen for the oxygen sensor due to their relatively low cost, small size and high selectivity. The design process combines several other disciplines including nanofabrication, electrical engineering and mechanical engineering. Nafion 117 membrane is chosen from the onset as the solid polymer electrolyte (SPE) for the sensor. Early work explored a variety of platinum electrodes including the use of carbon nanotubes (CNT) to increase surface area. Apart from the challenges of fabricating with carbon nanotubes, no significant advantages could be obtained with the CNT-based electrodes. Instead, a SPE-based oxygen sensor is developed using the membrane electrode assemblies commonly found in proton exchange membrane (PEM) fuel cells. A Teflon coating allows the sensor to function under the influence of varying humidity. The sensor implements two porous electrodes rather than the typical 3-electrode sensor cell and therefore does not require a potentiostat circuit to operate. Early embodiments are replaced with a small, open channel sensor design. The small, portable sensor is powered by two 3-volt lithium coin batteries and transmits the signal wirelessly to a USB receiver on a laptop. Unlike similarly sized commercial sensors, the developed sensor does not require a pneumatic sampling system. It is thus a compact untethered sensor very suitable for ambulatory applications. The raw signal is post-processed using special signal processing algorithms. The total material cost of the sensor (including wireless components) is roughly $200. The sensor is tested at the University of Minnesota using a MedGraphics Cardiopulmonary Exercise System (CPX) and against a mass spectrometer (MS) at St. Mary’s Hospital in Rochester MN. The final test is conducted on February 10th, 2011 during which 20 data sets are obtained for various breathing patterns. Post processing techniques include algorithms for breath-detection, slope-correction, baseline-correction, model inversion and calibration. The low frequency drift is removed and end tidal oxygen concentration is measured with an RMS error of 0.2 %O2 over steady, resting-rate respiration (max error less than ±0.75 %O2). During faster respiration (such as exercise) or largely-varying respiration, the end tidal accuracy of the sensor deteriorates to an error of ±1 %O2 (max error less than ±4 %O2). A new configuration of the sensor is developed with an on-board fan. The fan speeds up the sensor response and is also able to provide a rotational speed signal from which breath flow rate through the sensor can be computed. Even with the on-board fan, the sensor can continue to be wireless and untethered. The data shows that the improved speed of the sensor provides more accurate O2 concentration measurement and a closer match with the mass spectrometer. However, the calibration gains needed for the sensor vary with the breathing flow rate. An algorithm that calibrates sensor response using flow rate measurements is found to yield consistent performance over a variety of breathing patterns and varying flow rates. With the fan flow sensor in place, the sensor error decreased to ±0.35 %O2 (±1.2 %O2 maximum) for erratic breathing patterns, including widely-varying flow rates. The fan flow meter also enables the oxygen sensor to measure volume of flow which is useful for many clinical diagnostic applications.