On the scaling approach to electron-electron interactions in a chaotic quantum dot

Shaffique Adam,Piet W. Brouwer,Prashant Sharma

DOI: https://doi.org/10.1103/PhysRevB.68.241311

2003-09-03

Abstract:A scaling theory is used to study the low energy physics of electron-electron interactions in a double quantum dot. We show that the fact that electrons are delocalized over two quantum dots does not affect the instability criterion for the description of electron-electron interactions in terms of a ``universal interaction Hamiltonian''.

Mesoscale and Nanoscale Physics

What problem does this paper attempt to address?

The problem that this paper attempts to solve is about the low - energy physical behavior of electron - electron interactions in a double - quantum - dot system. Specifically, the author explores whether the distribution of electrons between two quantum dots will affect the instability conditions of the "universal interacting Hamiltonian" that describes electron - electron interactions.

### Main problems and background

1. **Application of random matrix theory**:
- In chaotic quantum dots or disordered metal particles, the statistical distribution of single - particle energy levels and wave functions can be described by random matrix theory. This is because there is a large parameter - the dimensionless conductance \( g \), which is the ratio of the Thouless energy \( E_T \) to the average level spacing \( \Delta \).

2. **Universal interacting Hamiltonian**:
- The "universal interacting Hamiltonian" proposed by Kurland, Aleiner and Altshuler is used to describe electron - electron interactions in quantum dots and metal particles. In the case of large \( g \), the main contributions come from the capacitive charging energy, the long - range exchange interaction and the "Cooper - channel" interaction (responsible for superconducting instability).

3. **Role of off - diagonal interaction matrix elements**:
- Although the off - diagonal interaction matrix elements are smaller than the diagonal elements by a factor of \( 1/g \), their number is large, so it is necessary to study their role. Murthy et al. studied these problems using the renormalization group method and found that for strongly attractive Fermi - liquid interactions, the system will become unstable.

4. **Special features of the double - quantum - dot system**:
- This paper studies a special double - quantum - dot system, which is formed by coupling two quantum dots of approximately the same size through a point contact. The wave functions and interactions inside each quantum dot can be described by random matrix theory and the "universal interacting Hamiltonian". The author's goal is to calculate the wave - function correlations in detail in order to avoid using approximation methods in the renormalization group analysis.

### Main conclusions of the paper

The author's main finding is that, after considering the correct non - universal wave - function correlations, the instability of the universal Hamiltonian occurs at the same interaction strength as in the double - quantum - dot system without point contacts. This means that in the results of Murthy et al., the phenomenon that the finite - size system is unstable while the bulk system is stable may be due to the use of random matrix theory to describe the wave - function statistics in the region up to a distance of \( E_T/2 \) from the Fermi level.

### Summary of mathematical formulas

- **Interaction matrix elements**:
\[
V_{\alpha\beta\gamma\delta} = \int d\mathbf{r}_1 d\mathbf{r}_2 \phi_\alpha^*(\mathbf{r}_1) \phi_\beta^*(\mathbf{r}_2) V(\mathbf{r}_1, \mathbf{r}_2) \phi_\gamma(\mathbf{r}_2) \phi_\delta(\mathbf{r}_1)
\]

- **Average of wave - function correlations**:
\[
\langle \phi_\nu^*(k) \phi_\mu(k) \phi_\mu^*(l) \phi_\nu(l) \rangle = \frac{g \sigma_k \sigma_l \Delta^2}{4N^2 [g^2 \Delta^2 + \pi^2 (\varepsilon_\mu - \varepsilon_\nu)^2]} + \frac{\delta_{\mu\nu} + \delta_{kl}}{4N^2} - \frac{1 + \sigma_k \sigma_l}{8N^3}
\]

- **Renormalization group flow equations**:
\[
\frac{d\bar{u}_0}{dM} = 0
\]
\[
\frac{d\bar{u}_1}{dM} = \frac{\bar{u}_1^2}{g}