As part of an effort to develop standard test methods for the performance of these sensors in real-world environments, we employed the Fire Emulator/Detector Evaluator, an instrumented flow system originally designed to study the response of fire detectors (smoke, heat, gas), in a preliminary study to evaluate the performance of a representative selection of commercially-available hydrogen sensors.  The use of a flow system, as opposed to an exposure chamber, allows the investigation of dynamic changes in the sensor environment, as well as the ability to easily return the environment to its initial state.

The sensors under consideration here depend on a variety of sensing technologies including metal-oxide semiconductors, electrochemical cells, catalytic bead pellistors, thermal conductivity sensors, and sensors employing a combination of technologies.  They were exposed to hydrogen concentrations up to half the lower flammability limit, as well to nuisance gases (CO, CO₂, NO_x, hydrocarbon gas and vapor—all potentially present in hydrogen dispensing and storage areas).  They were also exposed to dynamic changes in environmental conditions by varying temperature, humidity, and flow velocity.  As expected, all sensors are sensitive to hydrogen within the ranges listed by their manufacturers.  However, we also find that many sensors are also affected by other gases and environmental factors, responding for example with “false positives” to propene gas and condensing water vapor, and in some cases with reduced sensitivity to hydrogen as temperature rises.  These results provide guidance for the development of a test method designed to assess real-world performance of hydrogen gas sensors.  The ultimate goal is to develop standard test methods to be employed by product certification agencies.