Compressive hyperspectral microscopy of nanomaterials

In the important and exciting field of plasmonics, dark field characterization of metal nanoparticles is particularly insightful. In this case, the spectrum of the scattered light directly reveals the plasmon absorption in the various nanoparticles. However, the common strategy first involves acquiring a wide field image and then centering the particular particle of interest before acquiring the spectra of that particle. When there are only a few particles, this is not so difficult but it becomes increasingly onerous as the number of particles on the surface increases or the nanoparticles become spatially extended. There are now academic [1, 2] and commercial systems [3] that use pushbroom spectroscopy to speed up this process but need a scanning slit coupled to a very expensive CCD camera. More recently, there exist other strategies for hyperspectral microscopy, both compressive such as the computational CASSI system developed at Duke University [4] and the hardware solution of the Image Mapping Spectrometer developed by Tomas Tkaczyk at Rice University [5]. However these too at their heart are still essentially pushbroom hyperspectral imaging systems. There has been an explosion over the last decade in applying computational imaging techniques such as compressive sensing (CS) to radically alter the way various technologies acquire data.