| 研究課題/領域番号 |
26400393
|
|---|
| 研究機関 | 大阪大学 |
|---|
研究代表者 |
SLEVIN KEITH 大阪大学, 理学(系)研究科(研究院), 准教授 (90294149)
|
|---|
| 研究期間 (年度) |
2014-04-01 – 2019-03-31
|
| キーワード | semiconductor physics / Anderson localisation / Anderson transition / Coulomb interaction |
|---|
| 研究実績の概要 |
A central goal of the project is to understand how the long-range Coulomb interaction affects the Anderson transition. We have in mind a particular example, the metal insulator transition in doped semiconductors. In the previous year data was accumulated for the case of a compensated semiconductor. Analysis of this data continued this year. While the data show a metal insulator transition a clear-cut finite size scaling analysis was not possible.
One of the bottlenecks of the simulation is the diagonalization step. In the original version of the simulation this is performed using the JADAMILU library. In an attempt to speed up this step, this part of the simulation was rewritten to use the recently developed FEAST algorithm. Unfortunately, the results were disappointing.
An important feature of the metal insulator transition in semiconductors is the clear experimental evidence in spin susceptibility measurements for the existence of magnetic moments for doping concentrations near the metal insulator transition. These moments break time reversal symmetry and change the universality class of the transition. A better understanding of this symmetry breaking was obtained by simulating an Anderson-Heisenberg model, which is a model with an electron in a random potential coupled to randomly distributed classical spins.
|
| 現在までの達成度 (区分) |
現在までの達成度 (区分)
3: やや遅れている
理由
Difficulties have been encountered in understanding the reason for the very weak size dependence of the simulation data for the compensated semiconductor.
Difficulties have been encountered in speeding up the diagonalization step of the simulation.
|
| 今後の研究の推進方策 |
We will continue efforts to reduce the computational time needed for the simulations by carefully reconsidering each step of the simulation with a view to further optimization and speed up of the simulation.
|
| 次年度使用額が生じた理由 |
The project involves intensive numerical simulations. This requires up-to-date hardware and software tools particularly programming tools such as compilers, source control, calculation tools and tools for analysis of data.
|
| 次年度使用額の使用計画 |
Upgrade of necessary software licenses including source control (PureCM), fortran compilers (Intel and PGI), computational tools (MATLAB, Mathcad), graphing tools (Origin), software development tools (Dreamspark subscription, Wing IDE, Enterprise Architect etc.), reference and manuscript preparation tools (EndNote, WinEdt).
Travel expenses for participation in relevant conference and workshops.
|
