You are here: Archives > July 2010 > Hydrogen sensors

The Importance of Fast Responding Hydrogen Sensors in the Detection of Hydrogen Leaks
Lois Brett, European Commission, DG JRC - Institute for Energy


The Institute for Energy (IE) is one of 7 scientific research institutes of the Joint Research Centre (JRC) of the European Commission. Its mission is to provide independent support to European Community policies related to energy in order to ensure sustainable, safe, secure and efficient energy production, distribution and use.

At the JRC-IE we have been involved in impartial hydrogen sensor performance testing since 2001. In the execution of our activities we collaborate with sensor developers, manufacturers, end users, renowned research institutions (e.g. NREL) and international standards organisations to support the safe and effective use of hydrogen. During the past decade, we have witnessed a large increase in the number of hydrogen detection devices available on the commercial market, as well as significant improvements in their performance. Sensors are now capable of measuring hydrogen concentration with greater accuracy and under more challenging ambient operating conditions. Furthermore, both sensor size and price have decreased over this period.
In recent times there has been a change in the focus of external interest in our work. Enquiries now relate more to the practical application and correct deployment of hydrogen sensors rather than to the details of the various types of sensing technology as in the past.

Traditionally performance evaluation of chemical sensors, including hydrogen sensors, has focussed on their selectivity and sensitivity to the species of interest, as well as their operating life stability. However, the recognised importance of hydrogen sensor response time in safety applications means that this property is also receiving much attention, and increasingly stringent targets are being set [1]. Rapid detection of hydrogen gas is essential to provide an early warning of an unwanted hydrogen release. In this respect the shorter the sensor's response time (RT) the better (here we define response time, t(90), as the interval between the time when an instantaneous variation from clean air to the standard test gas, with a given hydrogen concentration, is produced at the inlet of the hydrogen sensor and the time when the response reaches 90% of the maximum indication).

While a short RT is a critically important performance parameter of any hydrogen safety sensor, the total time of response to a leak depends not only on the RT of the sensor itself, but also on the location of the sensor relative to the point of release. Consequently correct sensor placement is another equally if not more important consideration in facilitating rapid leak detection and promoting safe hydrogen use.

At present there are several commercially available hydrogen sensors with a claimed response time of >5 s. Developments in micro- and nano-structured sensor technologies promise smaller, cheaper and even faster hydrogen sensors in the future. The emergence of progressively faster hydrogen sensors demands improved methods to accurately measure, validate and compare the response time of these devices.

However accurate measurement of this property is not trivial, as it can be difficult to separate the response time of the sensor from the time taken for the hydrogen to reach it. As defined, a sensor's RT is measured relative to an instantaneous gas change and so any experimental method requires an instantaneous, step-wise gas change from clean air to the desired hydrogen concentration at the sensor inlet. In reality however such a stepwise gas exchange is not possible and the hydrogen concentration at the sensor inlet will always increase in a transient manner. Consequently any experimental measurement of a sensor's RT includes not only the time taken for the sensor to react with and respond to hydrogen but also a time period required to complete gas exchange at the sensing element.

Various sensor RT measurement methods exist and the JRC-IE has evaluated the two methods which are described in ISO 26142:2010 - Hydrogen detection apparatus. The RTs of two different commercial hydrogen sensors were measured using these methods. The purpose of the work was to assess the suitability of these different methods for the measurement of sensor response time. The methods were evaluated in terms of the measurement repeatability and accuracy when compared with the sensor manufacturer's specifications. We have shown that the measured RTs depend highly on the method used to evaluate them. Based on the observations and results new RT measurement methods were developed and evaluated yielding significant improvements in measurement accuracy and repeatability.

The two methods suggested in ISO 26142 employ different principles for measuring hydrogen sensor RT. One method uses diffusion and the other dynamic flow to transport gas to the sensor's sensing element. In our realisation of the flow-based method the hydrogen sensor is fixed to the side of a copper pipe, down-stream of a 3-way valve. Switching of this valve allows either air or the hydrogen test gas mixture to be selectively flowed through the pipe to the sensor. Following switching of the valve from air to test gas the response of the sensor is recorded to yield the RT. This flow-based method proved to have a reasonable repeatability however the accuracy of the measurements was not very good. The main advantages of this method include its ease of execution, simple set-up and the ability to also measure sensor recovery time, which was not always possible using the other methods.

In the diffusion based method or "Membrane method", the sensor is placed inside a small holder cell which is sealed at the top with a latex membrane. The sensor holder is then placed inside a 30L diffusion chamber which is subsequently filled with hydrogen in air. At the start of the measurement the membrane is ruptured using a scalpel, allowing the hydrogen gas mixture to diffuse to the sensor. The sensor response is recorded to yield the RT. Extensive measurements of RT were made using this method however the measured RTs were much longer than those reported by the sensor manufacturer and showed a very large variation. In addition execution was cumbersome and time consuming, large volumes of gas were required and sensor recovery time measurements were not possible.

The low accuracy and repeatability of measurements made using the Membrane method was largely attributable to inconsistent rupturing of the membrane, diffusion of hydrogen through the membrane and uncertainty regarding the exact start time. For these reasons this method was modified by replacing the latex membrane with a latched aluminium lid. All other aspects of the set-up and method remained the same. At the start of the measurement the lid was removed by releasing the latch holding it in place thereby exposing the sensor rapidly to the test gas mixture. Results obtained using this “Lid method” showed very significant improvements in both the accuracy and repeatability of the RT measurements. This was achieved with minor, but nonetheless effective, modifications to the original method suggested in ISO 26142. Similar to the Membrane method, execution of this method was time consuming, large volumes of gas were required and sensor recovery time measurements were not possible.

The final method tested and evaluated used a fast acting gate valve to separate the sensor under test from the hydrogen gas mixture. The "Gate valve method" was designed to allow measurement of not only the sensor response time but also the recovery time. This was made possible by locating the sensor holder outside the 30L diffusion chamber. The gate valve separated the two volumes, each of which could then be independently flushed with either air or hydrogen test gas mixture. In this way, both sensor response and recovery time measurements were theoretically possible. Following opening of the gate valve the sensor response was recorded and the RT measured. The measurements were quite accurate and repeatability was good however this method was technically complex, costly and recovery time measurement was not always possible under certain conditions.

The full results and conclusions from this work have recently been published [2] and they were also presented during the NHA Hydrogen Conference and Expo 2010 [3]. While these results were not available for consideration during the preparation of the current version of ISO 26142:2010 the JRC-IE will continue to provide independent scientific and technical support for any future revisions of this standard.


Sources

US DoE Multi-Year Research, Development and Demonstration Plan: Planned Program Activities for 2005-2015
http://www1.eere.energy.gov/hydrogenandfuelcells/mypp/pdfs/safety.pdf

L. Boon-Brett, G. Black, P. Moretto, J. Bousek, A comparison of test methods for the measurement of hydrogen sensor response and recovery times, International Journal of Hydrogen Energy, Volume 35, Issue 14, July 2010, Pages 7652-7663

Conference details and proceedings available from http://www.hydrogenconference.org