Analytical methods capable of in situ monitoring of water quality have been in high demand for environmental safety,
the identification of minute impurities and fundamental understanding of potential risks of these molecular species.
Raman spectroscopy, which provides 'fingerprint' information about molecular species in the excitation volume, is a
powerful tool for in vivo diagnostics. However, due to a relatively weak Raman signal (~ 1 out of 1014 incident photons
produces the useful signal) there is a need to significantly (by many orders of magnitude) enhance this signal, to raise the
detection sensitivity of this technique. Traditionally, surface enhanced Raman spectroscopy is employed to dramatically
increase the local field intensity and substantially improve the efficiency of Raman scattering. However, the above
enhancement occurs only in "hot spots", which represent only a small percent of the total surface are of the substrate.
Plasmonic nanostructures are also found to be hard to manufacture in large quantities with the desired degree of
reproducibility and to be unable to handle high laser power. We propose and experimentally demonstrate a new type of
approach for ultrasensitive Raman sensing. It is based on manufacturing a random porous structure of high-index
material, such as GaP, and use the effect of light localization to help improving the detection sensitivity of such sensor.
The desired structure was manufactured using electrochemical etching of GaP wafers. The observed Raman signal
amplitudes are favorably compared to the best known plasmonic substrates.