Start Date: September 2007
H.-S. Philip Wong, Peter Peumans, Yoshio Nishi, Electrical Engineering; Mark Brongersma, Materials Science and Engineering, Stanford University
This research aims to develop a novel concept for multijunction photovoltaic devices. The classical multijunction structure consisting of multiple series-connected, vertically-stacked cells is revisited to overcome intrinsic limitations such as photocurrent matching in varying illumination conditions and expensive epitaxial growth. The proposed device uses nanowire-based subcells connected in parallel and a plasmonic electrode serving both as a lateral spectral filter and as a light concentrator.
Multijunction solar cells, first developed for space-based applications in the early 80’s, are the most mature photovoltaic concept for high efficiency (>30%) solar energy conversion. Multijunctions are designed to optimize the absorption of solar radiation by using multiple series-connected semiconductor layers that absorb different portions of the solar spectrum (see Figure 1a). This approach reduces energy losses caused by the thermalization of electron-hole pairs produced by photons with energy larger than the bandgap. Higher conversion efficiencies can therefore be achieved compared to wafer-based or thin-film single-junction devices: the thermodynamic efficiency limit for triple junctions is close to 50% under global illumination and to date, III-V-based triple-junction laboratory cells have reached record efficiencies exceeding 40%. The flip side of this technology is its high production cost. Although concentrator systems are being proposed to decrease the amount of active material in a panel, the intrinsic cost of the absorber remains high due to the low-rate, high-vacuum deposition methods used to grow the high-purity crystals necessary to achieve sufficient carrier lifetime.
The present research investigates novel multijunction architectures that may overcome cost barriers and other limitations such as current matching between series-connected subcells. The proposed structure – illustrated in Figure 1b – uses nanowire-based subcells deposited in a monolayer and connected in parallel. Incoming radiation is filtered and distributed laterally by a plasmonic electrode that also serves as a concentrator to enhance light absorption in the nanowire monolayer.
The lateral multijunction device consists of an array of vertically aligned nanowire-shaped p-n junctions with different bandgaps grown on top of a nanostructured electrode. The metal structure of this electrode is the keystone of the proposed concept. Its plasmonic capabilities allow concentration of incident light up to 105-fold on a small layer (~100nm) close to its surface, which should dramatically enhance light absorption in the nanowire monolayer. Additionally, the plasmonic film is engineered to be capable of filtering light in a space and frequency selective manner. The incident broadband solar spectrum gets split and photons within distinct spectral ranges are localized in different spatial locations coinciding with the location of nanowires with the appropriate bandgap, as illustrated in Figure 2. Electron-hole pairs are produced upon photon absorption at the p-n junction within the nanowires and current is extracted separately from each nanowire. Nanowires are connected in parallel such that photocurrent matching is not required. In order to maximize power output, however, sections with low bandgap will be connected in series to match the voltage produced by the highest bandgap section.
The two-dimensional nanostructured electrode is designed to tailor plasmonic resonances for efficient spectral splitting (efficient sunlight separation into lateral hotspots of different wavelengths) and light concentration. Optimal nanostructure design is investigated using both automated design procedures based on genetic algorithms and electronic filter design approaches. The best structures are fabricated using electron beam lithography and focused ion beam milling, and optically characterized both in the near- and far-field.
The optical response of the metal nanostructured film is then used to direct the growth of the nanowires to match their position to the location of the appropriate frequency hotspots produced by the plasmonic concentrator. Multiple strategies are investigated to do this, including local laser-driven heating of catalyst nanoparticles to direct the vapor-liquid-solid growth of the semiconductor nanowires.
The project includes the integration of all above elements into a working prototype that will be tested using traditional opto-electrical diagnostic methods and deep-level transient spectroscopy to measure trap density profiles.