NEWS

Parameter optimization using Gaussian processes

One often encounters the problem of finding 'optimal domains' of parameters when a single calculation (or experiment) is time-consuming/expensive. We demonstrate for the case of an assembly of interacting Rydberg atoms which are coupled to a driven, dissipative environment, that Gaussian process optimization can be used to obtain at the same time optimal parameters and a global overview of the complete parameter space, while using only a few calculations. The specific problem considered emerges in the context of using such assemblies of Rydberg atoms as simulators for open quantum system dynamics.

The luminescent solar concentrator (LSC) is a simple device which it is hoped will enhance significantly the use and efficiency of solar devices, in particular photovoltaic (PV) cells. The basic construction of all LSC is a flat plate of polymer or glass (the host material) in which is embedded luminophores (dye molecules, quantum dots, etc.). Sunlight enters the plate, is absorbed by the luminophores and then emitted at a longer wavelength. The emitted light must have the property that in the particular host material it largely undergoes total internal reflection at the faces of the plate. In this way, most of the light is concentrated toward the thin edges of the plate where the light is emitted. The plate thereby acts essentially as a waveguide. Indeed one simple application is their use as advertising displays where letters are cut out of the plate leaving the edges of the letters illuminated. Another potential application is to direct the collected light for internal use, e.g., to illuminate a dark corridor in a building. However, the major use of LSC in solar energy is to direct the light onto PV cells at the edge of the plate.
We propose the design of a luminescent solar concentrator (LSC) in which dye aggregates with cylindrical geometry are embedded in the LSC plate with the long cylinder axis orthogonal to the faces of the plate. The most suitable aggregates are those exhibiting a blue-shifted (with respect to monomer absorption) H band and a narrow red-shifted J band. Such aggregates have high primary absorption into the H band which is polarized in the plane of the LSC and emission from the J band which is polarized perpendicular to the plane. Hence the emitted light is directed predominantly to the edges of the LSC and loss into the escape zone is much reduced.

The fabrication and utilization of nanoscale machines and devices is one of the great promises of the beginning 21st century. In particular, so-called nanoelectromechanical systems provide intriguing possibilities for applications beyond common paradigms. One archetype in this regard is the nanomechanical single-electron transistor (NEMSET), which can exhibit mechanically assisted charge transport. Here we discuss a nano-motor based on ideas from the NEMSET. We show that a time-independent field gradient can lead to a self-excitation of a continuos rotatory motion and how this dynamics is related to the measured current though the device. Hereby we identify different dynamical regimes, one of which is similar to a common NEMSET, but shows a negative differential conductance. The current-voltage characteristics can be used to infer details of the surrounding which is responsible for damping.

Energy transfer in the photosynthetic Fenna-Matthews-Olson (FMO) complex of green sulfur bacteria is studied numerically taking all three subunits (monomers) of the FMO trimer and the recently found eighth bacteriochlorophyll (BChl) molecule into account. The coupling to the non-Markovian environment is treated with a master equation derived from non-Markovian quantum state diffusion. When the excited-state dynamics is initialized at site eight, which is believed to play an important role in receiving excitation from the main light harvesting antenna, we see a slow exponential-like decay of the excitation. This is in contrast with the oscillations and a relatively fast transfer that usually occurs when initialization at sites 1 or 6 is considered. We show that different sets of electronic transition energies can lead to large differences in the transfer dynamics and may cause additional suppression or enhancement of oscillations [URL].

To investigate the effect of quantum coherence on electronic energy transfer, which is the subject of current interest in photosynthesis, we solve the problem of transport for the simplest model of an aggregate of monomers interacting through dipole-dipole forces using both quantum and classical dynamics. We conclude that for realistic coupling strengths quantum and classical coherent transport are identical. This is demonstrated by numerical calculations for a linear chain and for the photosynthetic Fenna-Matthews-Olson complex. [URL]

The direction of excitonic energy transfer along a chain of monomers can be controlled by the phase of the initial state. In the picture initially the excitation is delocalised on two monomers with a phase difference of π/2 which leads to propagation mainly to the right. In addition the interaction between the monomers is designed such that the excitation does not disperse, but is focused at the end of the chain. The coupling strengths between nearest neighbors are shown in the upper panel. [URL]

Highly excited atoms (Rydberg atoms) have almost macroscopic dimension. We have seen that their extraordinary properties allow the construction of a microscopic version of Newton's cradle. Where the metal balls of a classical cradle only impart energy and momentum on one another, the Rydberg atoms in our micro version additionally hand over the delicate quantum property of "entanglement". Entanglement is a crucial ingredient of quantum computing and may even be involved in of the most important processes of life: photosynthesis. Thus, systems to control or study its migration may be of great relevance. [PDF] [URL]