Parallel-processing concepts for solar energy conversion are being explored at NASA and through university grants as part of a high-risk effort to design a 50% efficient solar cell for the 1990's. The aim is to overcome the 56% loss associated with mismatch between the broad solar spectrum and the monoenergetic conduction electrons used to transport energy in conventional silicon solar cells (Fig. 1). The parallel-processing strategy would use surface plasmons or other guided electromagnetic waves for broadband energy transport, and an array of tunable diodes for energy extraction (Fig. 2). Tunable diode research has focused on inelastic tunneling in metal-oxide-semiconductor-metal films, with the idea of designing energy conversion devices which can be tuned to different frequencies by varying the load. This paper describes the technical barriers which must be overcome in order to harness surface plasmons for solar energy conversion, and describes current research on four key problems: practical techniques to phase-match sunlight to surface plasmons, structures to minimize ohmic loss and reradiation, mode conversion techniques to couple energy into the tunnel diodes, and junction designs to maximize the probability of plasmon capture by tunneling electrons.
|