Contributions to a reliable hydrogen sensor based on surface plasmon surface resonance spectroscopy

Abstract

Hydrogen is being seen as a potentially inexhaustible, clean power supply. Direct hydrogen production and storage techniques that would eliminate carbon by-products and compete in cost are accelerated in R&D due to the recent sharp price increase of crude oil. But hydrogen is also linked with certain risks of use, namely the danger of explosions if mixed with air due to the very low energy needed for ignition and the possibility to diminish the ozone layer by undetected leaks. To reduce those risks efficient, sensitive and very early warning systems are needed. This paper will contribute to this challenge in adopting the optical method of Surface-Plasmon-Resonance (SPR) Spectroscopy for a sensitive detection of hydrogen concentrations well below the lower explosion limit. The technique of SPR performed with fiberoptics would in principle allow a remote control without any electrical contacts in the potential explosion zone. A thin palladium metal layer has been studied as sensing element. A simulation programme to find an optimum sensor design lead to the conclusion that an Otto-configuration is more advantageous under intended “real world” measurement conditions than a Kretschmann configuration. This could be experimentally verified. The very small air gap in the Otto-configuration could be successfully replaced by a several hundred nm thick intermediate layer of MgF 2 or SiO 2 to ease the fabrication of hydrogen sensor-chips based on glass slide substrates. It could be demonstrated that by a separate detection of the TM- and TE-polarized light fractions the TE-polarized beam could be used as a reference signal, since the TE-part does not excite surface plasmons and thus is not influenced by the presence of hydrogen. Choosing the measured TM/TE intensity ratio as the analytical signal a sensor-chip made from a BK7 glass slide with a 425 nm thick intermediate layer of SiO 2 and a sensing layer of 50 nm Pd on top allowed a drift-free, reliable and reversible determination of hydrogen concentrations up to about 10 vol.% in dry or humid air with a detection limit of 0.04 vol.% with response times of around 2 min.