1998
DOI: 10.1103/physrevlett.80.3563
|View full text |Cite
|
Sign up to set email alerts
|

Kosterlitz-Thouless-Type Metal-Insulator Transition of a 2D Electron Gas in a Random Magnetic Field

Abstract: We study the localization property of a two-dimensional noninteracting electron gas in the presence of a random magnetic field. The localization length is directly calculated using a transfer matrix technique and finite size scaling analysis. We show strong numerical evidence that the system undergoes a disorder-driven Kosterlitz-Thouless-type metal-insulator transition. We develop a mean field theory which maps the random field system into a two-dimensional XY model. The vortex and antivortex excitations in t… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

10
65
0

Year Published

2001
2001
2023
2023

Publication Types

Select...
8
1

Relationship

3
6

Authors

Journals

citations
Cited by 59 publications
(75 citation statements)
references
References 34 publications
10
65
0
Order By: Relevance
“…Following the method specified in Ref. [26], the 2M vector of wave function amplitudes (ψ n+1 ,ψ n ) T is related to the vector (ψ n ,ψ n−1 ) T according to the relation…”
Section: Model and Methodsmentioning
confidence: 99%
“…Following the method specified in Ref. [26], the 2M vector of wave function amplitudes (ψ n+1 ,ψ n ) T is related to the vector (ψ n ,ψ n−1 ) T according to the relation…”
Section: Model and Methodsmentioning
confidence: 99%
“…The electron localization length is often obtained from the transfer matrix method. For a two-dimensional system, however, it is well known that, from this quantity, it is difficult to provide a conclusive answer to questions related to the metal-insulator transition (MIT) [10]. On the other hand, level-statistics analysis [11] has been used in studying MIT.…”
Section: -1mentioning
confidence: 99%
“…Using the transfer matrix method we calculate λ M (E) by a standard iteration algorithm [4,16]. In our calculations L > 10 6 ≫ λ M (E) and self-averaging requires relatively small data ensembles to achieve good statistics.…”
mentioning
confidence: 99%