arXiv subjects

D. Roth

Publications and source records attributed to D. Roth.

6 recordsLinked to original sources

arXiv2017

Electronic Stopping of Slow Protons in Oxides: Scaling Properties

Electronic stopping of slow protons in ZnO, VO2 (metal and semiconductor phases), HfO2, and Ta2O5 was investigated experimentally. As a comparison of the resulting stopping cross sections (SCS) to data for Al2O3 and SiO2 reveals, electronic stopping of slow protons does not correlate with electronic properties of the specific material such as band gap energies. Instead, the oxygen 2p states are decisive, as corroborated by density functional theory calculations of the electronic densities of states. Hence, at low ion velocities the SCS of an oxide primarily scales with its oxygen density.

cond-mat.mtrl-sci↗

arXiv2017

Electronic Stopping of Slow Protons in Transition and Rare Earth Metals: Breakdown of the Free Electron Gas Concept

The electronic stopping cross sections (SCS) of Ta and Gd for slow protons have been investigated experimentally. The data are compared to the results for Pt and Au to learn how electronic stopping in transition and rare earth metals correlates with features of the electronic band structures. The extraordinarily high SCS observed for protons in Ta and Gd cannot be understood in terms of a free electron gas model, but are related to the high densities of both occupied and unoccupied electronic states in these metals.

cond-mat.mtrl-sci↗

arXiv2013

Quasi-resonant neutralization of He+ ions at a germanium surface

When low-energy He ions are scattered from a Ge surface, the fraction of positive ions exhibits characteristic oscillations as a function of ion energy. These oscillations are caused by quasi-resonant neutralization (qRN), a process which can be observed for materials with a narrow band nearly resonant with the unperturbed He-1s level. In the present manuscript, we employ recently developed theoretical methods to extract quantitative information on the efficiency of qRN. Our evaluation reveals that qRN is a highly efficient process leading to ion fractions two orders of magnitude lower than in systems for which neutralization is only due to Auger processes.

cond-mat.mtrl-sci↗

arXiv2011

Efficiency versus Convergence of Boolean Kernels for On-Line Learning Algorithms

The paper studies machine learning problems where each example is described using a set of Boolean features and where hypotheses are represented by linear threshold elements. One method of increasing the expressiveness of learned hypotheses in this context is to expand the feature set to include conjunctions of basic features. This can be done explicitly or where possible by using a kernel function. Focusing on the well known Perceptron and Winnow algorithms, the paper demonstrates a tradeoff between the computational efficiency with which the algorithm can be run over the expanded feature space and the generalization ability of the corresponding learning algorithm. We first describe several kernel functions which capture either limited forms of conjunctions or all conjunctions. We show that these kernels can be used to efficiently run the Perceptron algorithm over a feature space of exponentially many conjunctions; however we also show that using such kernels, the Perceptron algorithm can provably make an exponential number of mistakes even when learning simple functions. We then consider the question of whether kernel functions can analogously be used to run the multiplicative-update Winnow algorithm over an expanded feature space of exponentially many conjunctions. Known upper bounds imply that the Winnow algorithm can learn Disjunctive Normal Form (DNF) formulae with a polynomial mistake bound in this setting. However, we prove that it is computationally hard to simulate Winnows behavior for learning DNF over such a feature set. This implies that the kernel functions which correspond to running Winnow for this problem are not efficiently computable, and that there is no general construction that can run Winnow with kernels.

cs.LG↗

arXiv1996

Influence of boundaries on pattern selection in through-flow

The problem of pattern selection in absolutely unstable open flow systems is investigated by considering the example of Rayleigh-B\'{e}nard convection. The spatiotemporal structure of convection rolls propagating downstream in an externally imposed flow is determined for six different inlet/outlet boundary conditions. Results are obtained by numerical simulations of the Navier-Stokes equations and by comparison with the corresponding Ginzburg-Landau amplitude equation. A unique selection process is observed being a function of the control parameters and the boundary conditions but independent of the history and the system length. The problem can be formulated in terms of a nonlinear eigen/boundary value problem where the frequency of the propagating pattern is the eigenvalue. PACS: 47.54.+r, 47.20.Bp, 47.27.Te, 47.20.Ky

patt-sol↗

arXiv1996

Pattern selection in the absolutely unstable regime as a nonlinear eigenvalue problem: Taylor vortices in axial flow

A unique pattern selection in the absolutely unstable regime of a driven, nonlinear, open-flow system is analyzed: The spatiotemporal structures of rotationally symmetric vortices that propagate downstream in the annulus of the rotating Taylor-Couette system due to an externally imposed axial through-flow are investigated for two different axial boundary conditions at the in- and outlet. Unlike the stationary patterns in systems without through-flow the spatiotemporal structures of propagating vortices are independent of parameter history, initial conditions, and system's length. They do, however, depend on the axial boundary conditions, the driving rate of the inner cylinder and the through-flow rate. Our analysis of the amplitude equation shows that the pattern selection can be described by a nonlinear eigenvalue problem with the frequency being the eigenvalue. Approaching the border between absolute and convective instability the eigenvalue problem becomes effectively linear and the selection mechanism approaches that one of linear front propagation. PACS:47.54.+r,47.20.Ky,47.32.-y,47.20.Ft

patt-sol↗