| Address: | Gergely T. Zimanyi Physics Department University of California One Shields Avenue Davis, CA 95616 |
| Phone: Fax: Email: |
530-400-3936 530-752-4717 zimanyi@physics.ucdavis.edu |
Quantum efficiency as a function of photon energy (Los Alamos group)
Quantum yield as a function of photon energy (NREL group)
In the basic process of solar energy conversion each incident solar photon excites one electron across the gap of a semiconductor structure from the valence band into the conduction band. The efficiency of this conversion mechanism suffers from multiple limitations. Photons with energy lower than the gap do not get absorbed at all. Photons with energy higher than the gap create electrons with an energy above the bottom of the conduction band. These electrons then relax to the bottom of the conduction band by releasing their excess energy via phonon emission, in effect heating the material. These two considerations were used to estimate the optimal efficiency of solar conversion to be around 31%. As of November 12, 2006, the best solar cells on the market can reach about 20% efficiency. A large variety of innovations have been proposed to improve on this limit, including tandem cells, polymer based solar cells, dye-sensitized metal oxides, or Gratzel cells, high purity single crystals, low-cost thin films, improved material composition such as CIGS and Cu/In/S, hot carriers, diffused metal ion traps and many more.
A radically new paradigm for solar energy conversion emerged very recently, based on semiconductor nano crystals. It was proposed that the excited high energy electrons may relax by generating additional charge carriers, thus capturing a higher percentage of the solar energy. The underlying mechanism of Impact Ionization, a reverse Auger process, has been known since the fifties in crystalline systems, but its efficiency was in the one percent regime in bulk materials. It was later proposed and explored that the efficiency of this process can be enhanced in semiconductor structures with reduced dimensionality, such as quantum wires, dots and superlattices. A significant breakthrough has been achieved in 2004, when Victor Klimov's group at Los Alamos demonstrated hitherto unprecedented levels of Carrier Multiplication (CM) in Nano Crystals (NCs), or Quantum Dots (QDs). In the Los Alamos experiments PbSe Nano Crystals were used as the absorbing medium. The explored energy range E of the incident photons was considerably higher than the gap Eg. However, the excess energy (E-Eg) resulted in carrier multiplication (CM), not phonon generation. It was demonstrated that a single incident photon may excite several, possibly up to seven electrons. This CM relaxation process can keep incident solar energy in the electronic sector instead of transferring it to the phononic one, promising a solar energy conversion with much-improved conversion efficiency.
As it is always true after such breakthroughs, in order to develop functioning nano-structured solar-cells based on these exciting gains and to improve on their efficiency, the underlying fundamental processes need to be explored, clarified and understood. Our research goal is to provide an explanation of the CM process observed in recent experiments, and use state-of-the-art computational techniques to propose optimalized nanostructured semiconducting materials for solar energy conversion. We perform this analysis in the context of developing carrier multiplication into a paradigm for solar-cell design. Some of our research topics include
- Using DFT based approaches to characterize the electronic states of Si, CdSe and PbSe nano crystals in the experimentally relevant size range.
- Using ab-initio Molecular Dynamics methods to investigate the reconstruction of the surfaces of the nano crystals.
- Determining the single particle surface states in different conditions, corresponding to growth conditions involving organic solvents or embedding in a semiconductor matrix.
- Developing a representation of the high energy states, needed to characterize photon absorption at multiples of the gap. We go beyond DFT, in particular by using the GW approximation. In specific cases, we also solve Bethe-Salpeter equations to capture two-particle interactions. Our goal is to identify trends in excitation properties as a function of size and surface structure.
- Estimating the transition rates proposed in the "Virtual Exciton Generation", the "Coherent Exciton States", and the "Impact Ionization" frameworks and analyze the connection between the theories. We compare the theoretical predictions to experimental data, including the linear energy dependence of the exciton number and the lack of a staircase structure. The main goal is to provide a robust interpretation of existing experiments and then to use our physical insight to possibly propose new semiconductor nanostructures with improved efficiency for solar energy conversion.
FORC of a CoPt film
Domain wall assisted reversal in exchange spring media
The story of magnetic recording is a story of continued, remarkable progress over the last several decades. Storage densities continue rising exponentially.The last reported demonstration was 420 Gb/sq.in, a most remarkable achievement.
Bits are written into ever smaller regions containing less than 100 grains. The transitions, separating these regions, are only a few grains wide. These grains are not identical, however, and variations in their relevant characteristics, most notably the switching fields, lead to a jitter of the transitions. This jitter generates a noise, lowering the performance of the recording media. Therefore, it is essential to determine the distribution of various characteristics of the grains, most notably their switching fields.
We developed a new technique, called First Order Reversal Curves, or FORC technique, which takes a fingerprint of a disordered magnetic film and captures the distribution of the switching fields and the effective local fields very efficiently. The first Figure shows the FORC diagram of a CoPt film. In the corresponding presentation a detailed analysis of the measured and calculated FORC diagrams is presented. Our most recent step forward was the introduction of a mean field technique to improve the performance of the FORC technique.
One of our eventual goals is to create a FORC library with hundreds of such FORCs. Comparing a measured FORC to the FORCs in this library will enable the quick identification of the appropriate model for any magnetic thin film.
We also study exchange coupled recording media. This recently proposed architecture allows the increasing of the energy barrier against thermal fluctuations while keeping the write fields limited. This is achieved by introducing a media containing a soft and a hard layer. The write field reverses the soft layer, thus creating a domain wall at the soft-hard boundary. This domain wall assists the reversal of the hard layer, thus enabling the write operation at a lower external write field. The second Figure shows the domain wall formation in zero field and at the write field.
A very recent concept goes beyond the bi-layer structure and proposes that the anisotropy of the grains of the perpedicular recording media should vary continuously. The optimal anisotropy profile was suggested to be quadratic by Dieter Suess. We simulated this architecture and verified that the quadratic anisotropy indeed brings considerable enhancements. In particular, the ratio of the energy barrier and the write field - in dimensionaless units, Victora's "figure of merit"- grows more than 3 fold. However, the purely quadratic form does not take into account the finite width of the domain walls themselves. In these ultrathin media the domain wall can be ~5nm thick, while the entire film is ~20nm thick: clearly comparable length scales.
Therefore, we studied various architectures which pin the finite width domain walls better. We found that a quadratic layer, topped off with a hard layer provides the highest figure of merit, increasing it to about 3.6, quite close to its theoretical maximum of 4.
Simulation vs Measured Data on NbSe2
Some basic features of the disordered vortex matter are well settled. These include the phase diagram, which consists of a weakly disordered vortex solid at low fields and temperatures, a strongly disordered frozen phase at higher fields and disorder, and a vortex liquid at high temperatures or fields. The dynamical properties include the depinning transition for a sufficiently large driving force/current and a re-ordering transition at higher drives.
However, even after fifteen years of intense studies, there are many unresolved issues. In the phase diagram the nature of frozen state is unclear. Our studies suggest that the frozen state is closer to regular window glasses with very high viscosity but finite correlation length, rather than a "vortex glass", which has a diverging correlation length upon approaching the transition.
Also, the nature of the melting in 2 dimension requires additional studies. Our results indicate that the dislocations, which enter the vortex matter, form large scale structures, such as domain walls, instead of the dilute gas, which is sometimes assumed. (see Figure)
These and many other open questions motivate us to continue the study of the static and dynamic properties of the disordered vortex matter.