Spectral and transmission properties of multiple correlated quantum dots made simple

Technical Summary: Spectral and Transmission Properties of Multiple Correlated Quantum Dots Made Simple

Problem Statement
The theoretical description of strongly correlated electronic systems, particularly multiple quantum dots (MQDs) coupled to metallic leads, presents a significant computational challenge. While the Single Impurity Anderson Model (SIAM) is well-understood using various many-body techniques (e.g., NRG, DMRG, QMC), extending these methods to double or multiple quantum dots is difficult. The physics of MQDs is qualitatively richer, involving interdot tunneling, capacitive coupling, and complex quantum phase transitions (QPTs) such as SU(4) Kondo correlations and charge frustration. Existing many-body approaches suffer from a computational cost that grows exponentially with system size, limiting their application to small clusters. Furthermore, capturing the full transmission properties of MQDs requires access not only to local spectral weights but also to the off-diagonal elements of the many-body spectral function matrix, which encode interference effects. There is a need for a framework that remains computationally manageable as the number of dots increases while retaining access to full many-body transmission and spectral functions in the strongly correlated regime.

Methodology: Steady-State i-DFT in the Ideal STM Limit
The authors employ a specific formulation of Density Functional Theory (DFT) known as steady-state DFT or i-DFT. Unlike standard DFT which describes equilibrium states, i-DFT describes a system in a steady state driven by a bias, utilizing both the steady-state density and the steady current as basic variables.

The core of the methodology relies on the "ideal Scanning Tunneling Microscope (STM) limit." In this setup, a second, infinitesimally weakly coupled probe lead (with broadening matrix $\Gamma_P$) is attached to the interacting system. By biasing this probe, a steady current $I$ is driven through the system. The authors utilize the Meir-Wingreen formula to establish that, in the limit where the probe coupling vanishes ($\Gamma_P \to 0$), the differential conductance is directly proportional to the equilibrium spectral function matrix $A(\omega)$ of the interacting system.

Key technical components include:

1. Decoupling of Equations: In the STM limit, the i-DFT equations for the equilibrium densities decouple from the equation for the current. This allows the densities to be solved self-consistently as an equilibrium DFT problem first. Subsequently, the current (and thus spectral properties) is determined by solving an independent non-linear equation for each frequency.
2. Reconstruction Formula: The interacting spectral quantity $\tau(\omega) = \text{Tr}\{\Gamma_P A(\omega)\}$ is reconstructed from the non-interacting Kohn-Sham (KS) spectral quantity $\tau_s(\omega)$ and the exchange-correlation (xc) bias $V_{xc}$ via the exact relation:
   $$\tau(\omega) = \frac{\tau_s(\omega + V_{xc})}{1 - \frac{1}{\pi} \frac{\partial V_{xc}}{\partial I} \tau_s(\omega + V_{xc})}$$
   where $V_{xc}$ and its derivative are evaluated at the current corresponding to the external bias $V = \omega$.
3. Bi-orthogonal Representation: To handle the non-Hermitian effective KS Hamiltonian in the presence of coupling to leads, the authors utilize a bi-orthogonal spectral representation of the resolvent, allowing for closed-form expressions of integrals and efficient numerical implementation.
4. Exchange-Correlation Functionals: The practical implementation requires approximate functionals for the Hartree-exchange-correlation (Hxc) potential and the xc bias. The authors construct these for density-density interactions (on-site Hubbard $U_l$ and intersite $U_{lm}$) in "Regime I" (where on-site interactions dominate). The functionals incorporate step structures characteristic of the Coulomb blockade and are supplemented with a Kondo prefactor to capture the Kondo regime at low temperatures.

Key Contributions and Results
The paper demonstrates the application of this framework to various MQD geometries, achieving results that agree well with many-body benchmarks at a fraction of the computational cost.

• Coulomb Blockade Regime: The authors computed local spectral functions for systems with $M=4$ quantum dots.
  • In the absence of interdot hopping, the i-DFT results matched exact Grand Canonical Ensemble (GCE) benchmarks across different on-site interactions.
  • With nearest-neighbor hopping introduced (Constant Interaction Model), the method successfully reproduced the splitting of degenerate peaks and the redistribution of spectral weight, showing excellent agreement with GCE results without modifying the xc functionals.
• Kondo Regime: By lowering the temperature and incorporating the Kondo prefactor into the xc bias, the method was applied to MQDs with $M=2, 3, 4$ dots.
  • The results show the emergence of narrow Kondo resonances at the Fermi level for dots near half-filling, alongside Coulomb blockade side peaks.
  • As no exact benchmarks exist for multi-orbital Kondo regimes in the literature for these system sizes, these results serve as genuine predictions of the i-DFT framework.
• Transmission Spectral Functions and Quantum Phase Transitions: A significant application was the calculation of transmission spectral functions for a double quantum dot (DQD) with off-diagonal coupling to the reservoir.
  • The study investigated a system where the ratio of level detuning ($\Delta\varepsilon$) to intersite interaction ($\Delta U$) acts as a control parameter for a quantum phase transition.
  • The i-DFT results reproduced the transition identified in previous Numerical Renormalization Group (NRG) studies: a broad Kondo resonance for $\Delta\varepsilon/\Delta U < 1$ (ferromagnetic-like) and a suppression of zero-frequency transmission for $\Delta\varepsilon/\Delta U > 1$ (antiferromagnetic-like).
  • The authors provided an analytical explanation for this transition within the KS framework: it arises from the self-consistent splitting of KS levels, which opens an interference antiresonance. The width of this dip serves as a spectral measure of the emergent low-energy scale, closely related to the second-stage Kondo scale found in NRG.

Significance and Claims
The paper claims that the ideal STM limit of i-DFT provides a "compact, flexible, and computationally efficient route" to the spectral and transmission properties of interacting nanoscale systems.

• Computational Efficiency: The method reduces the problem to solving $M$ coupled equations for densities (standard DFT cost) followed by independent scalar non-linear equations for spectral properties at each frequency. This avoids the exponential scaling of Fock space methods, allowing the treatment of larger systems.
• Access to Non-Local Information: By generalizing the probe coupling matrix, the framework accesses off-diagonal elements of the spectral function, enabling the calculation of transmission properties and interference effects that are often inaccessible to purely local spectral methods.
• Accuracy: The results are in "excellent agreement" with many-body approaches (GCE, NRG) for the tested regimes (Coulomb blockade and Kondo), validating the constructed functionals.
• Analytical Insight: The formulation allows for analytical investigations, such as the derivation of the QPT mechanism in the DQD, linking the suppression of transmission to KS level splitting and interference.

The authors maintain a modest tone regarding the limitations, acknowledging that the quantitative accuracy depends on the quality of the approximate functionals. They note that the current functionals are designed for specific interaction regimes (Regime I) and that future work is needed to systematically construct functionals for more general interaction regimes and improved temperature dependence. The paper does not propose new experimental setups but rather offers a theoretical tool to interpret and predict properties of existing quantum dot architectures.