Sweat Sensing through Optical Measurement of Hydrogels

Developed as a prototype sensing technology for Nano Micro Macro at the Harvard GSD & SEAS, 2018. Team Members: Nicki Adler, Berto Ceballos, Anesta Iwan, Kenneth So.

Sweat measurement and analysis is an accessible, non-invasive method which can provide physiological information directly relevant to the health and performance of an individual.

However, small flow rates and lack of easily worn sweat sensors pose challenges and have prevented sweat sensing from becoming a broadly adopted bodily measurement. This is compounded by the variability of sweating between persons, and the inability to build personalized baselines.

Current sweat measurements use a variety of methods. The simplest is absorbent patches that are then analyzed in the lab. Other passive methods use microfluidic sampling on the skin, or hydrogel based patches that uses hydrogel expansion as it hydrates to indicate sweat levels. Electronic mechanisms use printed circuits. Most of these technologies are not continuous or real time. And in many cases there is no clear pathway to integration with mainstream wearables.

Wearables often use an optical sensor to measure heart rate, using the blood's ability to absorb green light to measure pulses that correspond to heart rate.

We attemped to use existing optical sensors with an amplification layer of hydrogel. The hydrogel provides a substrate to react to minerals in the sweat. The presence of a substrate also makes it easier for a low-fidelity optical sensor to measure hydration via reflectivity using only small amounts of liquid.

Using different compositions and shapes of hydrogel we tested our ability to measure hydration via reflectivity in a low-resolution photograph.

These photographs are then pushed through image analysis software that measures luminosity in various ways. The best combinations of hydrogel substrate and software techniques results in a proof of concept that suggests optical measurement via hydrogels presents a usable mechanism for sweat sensing. In our current configuration a simple linear model of luminosity predicts hydration levels fairly well.

Of course there are significant challenges with how to actually integrate with wearables. Most hydrogels will need airation of some sort to take repeated measurements. The housing of the hydrogel will hence need to be portable and removable, adding extra heft. And reactivity to minerals present in sweat requires further scientific analysis. In most cases any hydrogel may only be able to measure reactivity to one mineral, not many at the same time.

This is hence a preliminary test of how we can get around the difficulty of measuring hydration of hydrogels in situations with extremely low fluid levels, such as sweat sensing.