Extraordinary optical transmission for surface- plasmon-resonance-based sensing

In 1998, Ebbesen and co-workers demonstrated the surprising result that arrays of small holes in a metal show optical transmission resonances [1]. This result was surprising because Bethe's aperture theory predicts negligible transmission through a single small hole in a thin metal film. As a result the phenomenon was termed extraordinary optical transmission (EOT). Unfortunately, the term EOT is a bit of a misnomer. Recently it has been shown that Bethe's theory, when applied to arrays of holes, predicts 100% transmission [2]. Therefore, when compared to Bethe's theory, EOT of imperfect conductors is actually a reduction in transmission! There is also a common misconception that surface plasmons somehow enable EOT. Once again, EOT is predicted by Bethe's theory using an infinitesimally thin, perfect electric conductor [2]. As that theory does not allow for surface plasmons, surface plasmons are not required for EOT. This is an important point because we do not wish to limit the application of EOT to surface-plasmon bearing cases. EOT may be (and has been) exploited in other parts of the electromagnetic spectrum without surface plasmons, for various applications. EOT may even be demonstrated and exploited using superconductor films. (Surface plasmons are relevant to EOT for metals that support them, but they are not required for EOT.) While it was recognized early on that EOT could be used to sense changes in bulk refractive index [3], we were the first to demonstrate monolayer sensitivity by using nanohole arrays for sensing based on surface plasmon resonance (SPR) [4]. There are considerable benefits to using nanohole arrays for SPR-based sensing, as compared with the conventional prism coupling geometry. The nanohole array approach to SPR-based sensing benefits from: (1) a collinear optical geometry for dense integration, (2) the potential for multiplexing from several nanohole arrays, (3) lower limits of detection, and (4) a reduction in the diffusionlimited adsorption time by using the nanoholes as conduits. Up until recently, the nanohole-SPR technique has faced the challenges of reduced sensoroutput sensitivity with respect to conventional SPR methods. (This is different than the analytical sensitivity, which refers to the absolute amount of analyte detected. The analytical sensitivity is large for the nanohole arrays due to the small sensing area.) Significant improvements in the sensor-output sensitivity have been achieved using crossed polarizers [5]. Thereby, the linewidth of transmission is narrowed significantly since the transmitted signal comes only from scattering between the two orthogonal polarization states by the array of holes. The interfering processes of direct transmission and co-polarization resonant transmission are suppressed. As a result, sensitivities approaching 10 refractive-index units (RIU) are possible [5]. Another limiting factor is the need for a spectrometer, an angle-scanner, or a tunable laser, which limits the portability and cost-effectiveness of the nanohole-SPR device. A recent innovation has been the removal of a spectrometer (without using a tunable laser or anglescanner) by using biaxial arrays [6]. Biaxial arrays have two different periodicities, one along each orthogonal axis. In this configuration, lasers of two different polarization states are used Journal of Nanophotonics, Vol. 2, 020305 (15 October 2008)