Plasmonics is the study of the interaction between electromagnetic field and free electrons in a metal. Free electrons in the metal can be excited by the electric component of light to have collective oscillations. However, due to the Ohmic loss and electron-core interactions, loss are inevitable for the plasmon oscillation, which is usually detrimental to most plasmonic devices. Meanwhile, the absorption of light can be enhanced greatly in the metal by proper designing metal patterns for SP excitation.
We theoretically predicted and experimentally demonstrated solar energy conversion by non-semiconductor metal-insulator-metal devices. Under illumination, the large spatial difference of excited electrons in the top and bottom metal guarantees net photocurrent generation from the simple MIM devices. Large absorption of light resulted from SP excitation by a Kretschmann setup boosted the photocurrent output, as surface plasmons on the top surface excited much more electrons than normal incidence. One advantage of these devices is that it's not limited by the bandgaps as in the traditional semiconductor solar cells. The absorber is a metal, so we can benefit from the tremendous studies of Plasmonics that provide metal patterns to absorb light at any specific wavelength. Optimized design of this MIM device can make it work in a wide spectrum from infrared to visible.
While the short circuit current can be enhanced by increasing the absorption, the open circuit voltage is independent on the absorption at each wavelength. The voltage is a monotonic function of photon energy. As a result, we can directly determine the wavelength of light with a single device, no reference is necessary. We demonstrated deconvolution of 3 wavelengths in a single Ti-TiO2-Au device by detecting its open circuit voltage. The devices’ working spectral range is mainly determined by the electron barrier and hole barrier at the interface between top metal and insulator. Since the ballistic electron transport process is extremely fast (~order ps), that the devices could operate at high frequencies. Unlike photocurrent, the Voc is independent of device area; thus in principle these could be scaled to nanometer dimensions without affecting the measurement.