Grassmann time-evolving matrix product operators for fermionic impurities coupled to a superconducting bath

This paper introduces the Grassmann time-evolving matrix product operator (GTEMPO) method extended to the Nambu formalism, providing a scalable approach for simulating fermionic impurities coupled to superconducting baths in both imaginary and real time.

The Gist

The Quantum "DJ" and the Superconducting Remix: A Simple Guide

Imagine you are a DJ at a massive music festival. You have one specialized turntable (the "Impurity") that plays a very specific, complex beat. Surrounding you is a massive, roaring crowd of thousands of people (the "Bath").

In physics, scientists want to understand how that one turntable behaves when it’s plugged into that massive crowd. Does the crowd drown out the beat? Does the beat change how the crowd dances? This is called a Quantum Impurity Problem.

The Problem: The "Superconducting" Crowd

Usually, scientists study "normal" crowds—people who just move around randomly. But this paper deals with a much weirder, more organized crowd: a Superconducting Bath.

In a superconducting crowd, people aren't just dancing randomly; they are dancing in perfectly synchronized pairs (called Cooper Pairs). If one person moves, their partner moves in a specific, mirrored way. This makes the math incredibly messy. It’s like trying to predict the movement of a crowd where everyone is holding hands with a partner and moving in a choreographed ballet.

Until now, our mathematical "calculators" were great at handling normal crowds, but they kept "crashing" when they tried to process the synchronized, hand-holding superconducting crowd.

The Solution: The "Nambu-GTEMPO" Method

The authors of this paper have created a new mathematical tool called Nambu-GTEMPO. To understand how it works, let’s use two metaphors:

1. The "Bogoliubov" Magic Trick (The Transformation)
The researchers realized that instead of trying to fight the synchronized dancing, they could use a mathematical "magic trick" called a Bogoliubov Transformation.

Imagine if, instead of trying to track every individual person and their partner, you suddenly changed your perspective. Instead of seeing "Person A" and "Partner B," you start seeing "The Pair" as a single unit. By changing how you look at the crowd, the chaotic, hand-holding ballet suddenly looks like a much simpler, normal dance. This "simplification" allows the math to run smoothly without crashing.

2. The "GTEMPO" Memory Trick (The Matrix Product Operator)
Even with the magic trick, the history of the dance matters. To know what the turntable will do at minute 10, you need to know what happened at minute 1, 2, and 3. Usually, this requires a massive amount of memory—so much that even supercomputers run out of space.

GTEMPO is like a highly efficient way of taking notes. Instead of writing down every single detail of every single dancer at every single second (which would fill a billion notebooks), it uses a "smart summary" system (called a Matrix Product Operator). It only keeps track of the essential patterns. It’s like a DJ who doesn't memorize every single note of a song, but instead remembers the rhythm and the melody, which is all they actually need to keep the beat going.

Why Does This Matter?

By combining the "Magic Trick" (to simplify the crowd) and the "Smart Summary" (to save memory), the researchers created a tool that can:

• Predict the Future: They can simulate how the system evolves in real-time (the "Real-Time" part).
• Understand the Present: They can calculate the stable, resting state of the system (the "Imaginary-Time" part).
• Work in Complex Environments: It works even when the "turntable" is very complicated and interacting heavily with itself.

The Big Picture: This tool helps scientists design better materials for the future—like faster computers, more efficient power grids, or even the building blocks of quantum computers—by giving them a way to finally "hear" the music playing in the complex, superconducting world.

Technical Summary

Technical Summary: Grassmann Time-Evolving Matrix Product Operators for Superconducting Quantum Impurity Models

1. Problem Statement

The study of quantum impurity problems (QIPs)—where a localized electronic state is coupled to a structured environment (bath)—is fundamental to condensed matter physics, particularly for understanding phenomena like Andreev reflection, Majorana zero modes, and the interplay between local correlations and superconductivity.

A central challenge in modern computational physics is solving the Superconducting Anderson Impurity Model (SAIM) within the framework of Dynamical Mean-Field Theory (DMFT) using the Nambu–Gor’kov formalism. While existing methods like Continuous-Time Quantum Monte Carlo (CTQMC) are numerically exact, they are largely restricted to the imaginary-time axis and suffer from the "dynamical sign problem" in real-time (non-equilibrium) simulations. Conversely, real-time methods like Numerical Renormalization Group (NRG) or standard Matrix Product States (MPS) often struggle with bath discretization errors or finite-temperature complexities.

2. Methodology: Nambu-GTEMPO

The authors introduce Nambu-GTEMPO, an extension of the previously developed Grassmann Time-Evolving Matrix Product Operator (GTEMPO) method. The core innovation lies in adapting the Feynman-Vernon influence functional (IF) to accommodate a superconducting bath.

• Bogoliubov Transformation: The primary technical hurdle was that a superconducting bath mixes spin species, preventing the standard decomposition of the IF into independent spin components. The authors overcome this by applying a Bogoliubov transformation to the BCS bath, effectively mapping the superconducting bath onto a "normal" bath representation.
• Nambu Formalism Integration: By transforming the bath, the hybridization Hamiltonian is rewritten in terms of "global" impurity operators that act on both spin species simultaneously. This allows the derivation of a compact, analytic expression for the Feynman-Vernon IF in the Nambu representation.
• Grassmann MPS (GMPS): The method represents the augmented density tensor (ADT) as a Grassmann Matrix Product State. They utilize a partial IF algorithm, which regrouping terms in the exponent of the influence functional by shared Grassmann variables. This allows the IF to be constructed as a product of $2M$ partial IFs, each with a low bond dimension, making the algorithm computationally efficient and scalable.
• Contour Flexibility: The method is designed to work on multiple time contours, including the imaginary-time contour (equilibrium), the Keldysh contour (real-time/non-equilibrium), and the L-shaped Kadanoff contour.

3. Key Contributions

• Theoretical Extension: Successfully derived the analytic expression for the Feynman-Vernon influence functional for fermionic impurities coupled to superconducting baths.
• Algorithmic Development: Developed the "Nambu-GTEMPO" algorithm, which generalizes the partial IF construction to handle the spin-mixing inherent in superconductivity.
• Unified Solver: Created a versatile numerical tool capable of performing both equilibrium (imaginary-time) and non-equilibrium (real-time) calculations within a single framework.

4. Results and Validation

The authors validated the method through three tiers of testing:

• Exact Benchmarking: In a single-orbital toy model and a non-interacting case ($U=0$), Nambu-GTEMPO showed excellent agreement with Exact Diagonalization (ED). They demonstrated that errors decrease linearly with the time step ($\delta\tau$ or $\delta t$) and that the bond dimension ($\chi$) can be controlled to saturate the error.
• DMFT Convergence: For the non-integrable SAIM ($U=-1$), the method was used as an impurity solver in DMFT iterations. The results showed a step-by-step convergence and matched CTQMC results on the imaginary-time axis with high precision. Notably, Nambu-GTEMPO provides results free of the sampling noise and sign problems that plague CTQMC.
• Non-Equilibrium Dynamics: The method successfully simulated the real-time evolution of impurity state probabilities (vacuum, spin-up, spin-down, and double occupation). This demonstrated its ability to tackle problems where CTQMC fails due to the dynamical sign problem.

5. Significance

Nambu-GTEMPO represents a significant advancement in the toolkit for strongly correlated electron systems. Its ability to provide noise-free, sign-problem-free, and real-time dynamics makes it a powerful candidate for studying:

1. Non-equilibrium superconductivity: Understanding how superconducting states respond to external perturbations or quenches.
2. Advanced DMFT studies: Serving as a robust impurity solver for Nambu-DMFT in both equilibrium and non-equilibrium regimes.
3. Complex Impurity Systems: The framework is extensible to multi-orbital and multi-site quantum impurity problems, offering a scalable path toward more complex materials science simulations.