HETEROSTRUCTURE DEVICE SIMULATION USING THE

WIGNER FUNCTION

by

KIRAN KUMAR GULLAPALLI, Ph.D, MS.

DISSERTATION

Presented to the Faculty of the Graduate School of

The University of Texas at Austin

in Partial Fulfillment

of the Requirements

for the Degree of

DOCTOR OF PHILOSOPHY

 

THE UNIVERSITY OF TEXAS AT AUSTIN

May 1994

 

ABSTRACT

Kiran Kumar Gullapalli, Ph. D.

The University of Texas at Austin 1994

Supervisor Dean Neikirk

In the last decade advanced heteroepitaxial technology has allowed the exploration of a wide variety of semiconductor heterostructures in which the electronic properties can be varied significantly over atomic length scales. Semiclassical models of electron transport are not useful for the analysis of such structures. Perhaps the most dramatic phenomenon illustrating the need for advanced quantum transport models is resonant-tunneling across a double barrier quantum well structure. Since the early work on resonant tunneling, many devices have been proposed that incorporate double barrier quantum wells, the motivation being the increased functionality per device. Because significant regions in these devices can still be described by semiclassical equations, it is highly desirable that the quantum transport equation have the structure of the semiclassical Boltzmann equation, and reduce to it when appropriate. Such a model is afforded by the Weyl transform and the associated Wigner function. In this work, a quantum transport model based on the Wigner function is developed for the analysis of heterostructure devices. Past work on the use of the Wigner function has assumed that the effective-mass is spatially uniform, clearly not the case in heterostructure devices. In this work, the spatially varying effective-mass has been correctly incorporated in the Wigner transport equation. While past work using the Wigner function has been restricted to relatively unimportant AlxGa1-xAs / GaAs heterostructure devices, this work presents improvements in the numerical treatment of the Wigner transport equation that are necessary to extend its application to the study of the more important devices based on InyAl1-yAs / InxGa1-xAs heterojunctions.

With the aid of the quantum transport models developed during this work, an intriguing memory switching phenomena was discovered in double barrier resonant tunneling diodes that contain N- - N+ - N- spacers. Experimentally, the devices can be reversibly switched between two conduction curves. They retain memory of the curve last switched to, even after removal of the external bias. Within the scope of the quantum transport models, the phenomenon is explained by the existence of two self-consistent charge distributions in the device, even when no external bias is applied.

 

Table of Contents

 

List or Figures ............... x

Chapter 1. Introduction .................. 1

Chapter 2. Schrödinger-Poisson Model for Quantum Transport .................. 6

2.1 Effective-mass Hamiltonians .................. 7

2.2 The Schrödinger equation on a discrete lattice .................. 14

2.2.1 Boundary conditions .................. 16

2.2.2 Resonances .................. 18

2.2.3 Electron density .................. 19

2.2.4 Current density .................. 21

2.2.5 Self-consistent potential .................. 23

2.3 G - X bandstructure .................. 24

2.4 Resonant-tunneling diodes .................. 26

2.5 Summary .................. 31

Chapter 3. Wigner-Poisson Model of Quantum Transport .................. 33

3.1 The Weyl transformation and the Wigner function .................. 31

3.2 Equation of motion of the Wigner function .................. 37

3.2.1 Parabolic energy band .................. 39

3.2.2 General energy band .................. 42

3.3 Summary .................. 48

Chapter 4. Numerical Aspects of Quantum Transport Simulation using the Wigner-Poisson Model .................. 50

4.1 Discrete Wigner-Poisson model: general considerations .................. 51

4.2 Transport in a parabolic energy band .................. 55

4.3 Transport in a tight-binding energy band .................. 56

4.4 Discretization of the drift term .................. 57

4.4.1 First order upwind/downwind discretization .................. 55

4.4.2 Second order upwind/downwind discretization .................. 67

4.4.3 Improved discretiziaion schemes .................. 69

4.5 Current density in the discrete model .................. 74

4.6 Solution method .................. 75

4.7 Summary .................. 77

Chapter 5. Memory twitching in AlAs/GaAs tunneling diodes .................. 79

5.1 Observation of memory switching in double barrier diodes .................. 81

5.2 Simulation of memory switching in double barrier diodes .................. 86

5.3 Summary .................. 103

Chapter 6. Summary and Recommendations .................. 104

Appendix A. Comments on the model of Tsuchiya, Ogawa and Miyoshi .................. 108

Bibliography .................. 113

Vita .................. 122