Strongly correlated Josephson junction: proximity… — Plain-Language Explanation

✨

This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you have a very stubborn, grumpy neighbor (the Hubbard model) who hates sharing anything. In their neighborhood, electrons (the people) are so crowded and aggressive that they refuse to move around. This creates a Mott Insulator: a material that should conduct electricity but acts like a perfect wall because the electrons are stuck in place, fighting each other.

Now, imagine you build a super-highway right next to this grumpy neighborhood. This highway is a Superconductor, a place where electrons flow perfectly without any friction. Usually, when you put a superconductor next to a normal metal, the "super" magic leaks over, and the metal starts conducting too. This is called the proximity effect.

But what happens when you put a superconductor next to a grumpy, stubborn Mott insulator? That is the question this paper answers.

Here is the story of their interaction, told in simple terms:

1. The Two Personalities (The M-Phase and S-Phase)

The researchers found that this grumpy neighbor has two distinct moods, depending on how hard you try to push them to cooperate.

The "M-Phase" (The Stubborn Wall):
If you try to connect the superconductor but the grumpiness (electron repulsion) is too strong, the neighbor stays stubborn. The superconducting "magic" tries to leak in, but the electrons inside are so busy fighting each other that they ignore the highway.
- The Result: The material remains an insulator. It acts like a wall that doesn't even care about the phase of the supercurrent. It's "Josephson-inactive," meaning it won't let the special supercurrent flow through it. It's like trying to whisper a secret through a brick wall; the wall just doesn't hear you.
The "S-Phase" (The Cooperating Metal):
If you increase the connection strength (make the tunnel between the neighbor and the highway wider), you eventually break through their stubbornness. The electrons are forced to cooperate.
- The Result: The material suddenly becomes a metal that conducts electricity perfectly. It acts like a normal superconductor, letting the current flow and responding to the "phase" (the timing) of the wave.

2. The Switch and the Hysteresis (The Light Switch with a Stuck Button)

The most fascinating part is how the material switches between these two moods. It's not a smooth slide; it's a sudden, violent jump.

Imagine a light switch that is stuck.

If you push it gently, it stays off (M-Phase).
If you push it really hard, it suddenly snaps to "On" (S-Phase).
But if you try to turn it off again by pulling back, it stays "On" until you pull much harder than you pushed.

This is called hysteresis. The material remembers its history. You can use the "phase bias" (the timing difference between the two superconductors) or the "transparency" (how wide the tunnel is) as a knob to flip this switch. This means you could theoretically build a switch that toggles between a perfect insulator and a perfect conductor just by tweaking the settings.

3. The "Ghost" in the Machine (The Self-Energy)

Why does the stubborn wall stay stubborn? The paper explains this using a concept called self-energy.

Think of the electrons in the Mott insulator as having a "ghost" inside them. In a normal insulator, this ghost is a single point right in the middle of the energy gap. When the superconductor tries to talk to it, this ghost splits into two smaller ghosts that dance around the center.

In the M-Phase: These two ghosts are so strong and so close to the center that they create a barrier. They block the superconducting information from passing through. It's like having two bouncers at a club door who are so aggressive that no one gets in.
In the S-Phase: The connection is so strong that the bouncers are pushed aside, and the "ghost" disappears, allowing the supercurrent to flow freely.

4. The Magic Trick at the End (The $\pi$ -Junction)

There is one final trick. In the "S-Phase" (the conducting state), if you twist the phase bias to a specific angle (called $\pi$ ), the superconducting gap closes up completely. The material turns into a "correlated metal"—a weird, messy metal that is still conducting but has no gap.

It's as if you twist the knob so hard that the wall dissolves entirely, leaving just a chaotic crowd of people running around. This happens because the superconducting "push" cancels out the internal "fight" of the electrons at that specific angle.

Why Does This Matter?

This isn't just about theory. Scientists are currently building devices using van der Waals materials (like stacking thin sheets of graphene and other crystals). They are trying to make tiny superconducting circuits.

This paper tells them: "Be careful!"
If you stack a superconductor on top of a strongly correlated material, you might accidentally create a switch that turns the current off completely, or you might get a device that behaves unpredictably. But, if you understand this "M-Phase" vs. "S-Phase" switch, you could use it to build new types of quantum switches or sensors that are controlled by pressure or spacing.

In a nutshell:
The paper describes a tug-of-war between a material that wants to be an insulator (due to internal fighting) and a superconductor that wants to make it conduct. Depending on how hard they pull, the material snaps between being a perfect wall or a perfect wire, and this snap can be used as a powerful switch for future electronics.

1. Problem Statement

The paper investigates the proximity effect in a strongly correlated electron system described by the single-layer Hubbard model when coupled to BCS superconductors. Specifically, it addresses the behavior of a Mott insulator embedded within a Josephson junction (two superconductors with a phase bias $\phi$ ).

The central question is how strong electron-electron interactions (characterized by the on-site Coulomb repulsion $U$ ) compete with induced superconducting pairing (characterized by the hybridization strength $\Gamma$ ) to determine the ground state. The authors aim to understand whether a Mott insulator can sustain a supercurrent or if it becomes "Josephson-inactive," and how the system transitions between insulating and conducting regimes under phase bias.

2. Methodology

The study employs Dynamical Mean-Field Theory (DMFT) combined with the Numerical Renormalization Group (NRG) as the impurity solver.

Model: The system is modeled as a single active layer of the Hubbard model (on a Bethe lattice with hopping $t$ and interaction $U$ ) coupled to semi-infinite superconducting leads (L and R) via tunneling. The leads are described by BCS Hamiltonians with gap $\Delta$ and phase difference $\phi$ .
Mapping: The lattice problem is mapped onto a self-consistent Superconducting Anderson Impurity Model (SAIM). The effective hybridization function $\Delta_{\text{eff}}$ includes both the lattice self-consistency ( $t^2 G_{\text{loc}}$ ) and the self-energy contribution from the superconducting baths ( $\Sigma_{\text{BCS}}$ ).
Solver: The NRG Ljubljana implementation is used to solve the impurity problem at zero temperature ( $T \approx 10^{-8}$ ). This allows for high-resolution spectral functions and the calculation of thermodynamic quantities.
Parameters: Calculations are performed at half-filling. The critical interaction strengths for the Mott transition in the isolated system are $U_{c1} \approx 2.39$ and $U_{c2} \approx 2.92$ . The superconducting gap is fixed at $\Delta = 0.05$ .

3. Key Contributions

The paper establishes a direct correspondence between the lattice phase transitions in the proximitized Hubbard model and the singlet-doublet quantum phase transitions (QPT) found in the single-impurity SAIM.

Two Distinct Phases: The authors identify two stable, gapped solutions separated by a first-order phase transition with hysteresis:
1. M-phase (Mott-like): A solution with free local moments, a large charge gap, and negligible supercurrent.
2. S-phase (Superconducting): A solution with induced pairing correlations, coherence peaks, and finite supercurrent.
Phase Diagram: They map out the coexistence region in the $(U, \Gamma)$ plane, showing that the transition lines are linear, consistent with an atomic limit competition between effective interaction and induced pairing.
Mechanism of Suppression: The paper provides a microscopic explanation for why the M-phase suppresses the Josephson current, distinguishing it from conventional band insulators.

4. Key Results

A. Spectral Properties and Phase Transition

M-phase ( $\Gamma < \Gamma_c$ ):
- The spectral function retains the Hubbard bands at $\pm U/2$ .
- The Mott gap is reduced from $\approx U - 2D$ to a scale set by $\Delta$ .
- Self-energy Structure: The imaginary part of the self-energy ( $\text{Im}\Sigma$ ) exhibits sub-gap resonances symmetrically located around the Fermi level. These arise from the splitting of the "mid-gap pole" characteristic of Mott insulators.
- Susceptibility: The spin susceptibility shows a peak at $\omega=0$ , indicating well-defined local moments.
S-phase ( $\Gamma > \Gamma_c$ ):
- The system develops a proximitized superconducting gap $\Delta^*$ with coherence peaks in both normal and anomalous spectral functions.
- The self-energy becomes gapless (Fermi-liquid like, $\text{Im}\Sigma \propto \omega^2$ ) within the induced gap.
- Local moments are screened (singlet ground state).
Transition: The transition at $\Gamma_c$ is discontinuous. The sub-gap resonances in the M-phase vanish abruptly, and the system jumps to the S-phase. This is accompanied by a sign change in the derivative of the pairing expectation value (though not the value itself).

B. Phase Bias Dependence ( $\phi$ )

M-phase (Josephson Inactive):
- The spectral function and local observables are almost completely insensitive to the phase bias $\phi$ .
- The critical current $I_c$ is suppressed by at least an order of magnitude compared to conventional SIS junctions.
- Physical Origin: The Mott charge gap is dynamically generated by self-energy poles, not a single-particle band gap. Consequently, there are no low-energy quasiparticle states to support coherent pair transfer across the barrier. The local moments remain decoupled from the superconducting leads.
S-phase (Conventional 0-junction):
- The ground state energy follows $E(\phi) \propto \cos(\phi)$ , characteristic of a standard 0-junction.
- Gap Closing at $\phi = \pi$ : As $\phi \to \pi$ , the proximitized gap $\Delta^*$ closes ( $\Delta^* \to 0$ ), and the system becomes a correlated metal with a quasiparticle peak at the Fermi level. This is a remnant of the "doublet chimney" physics in the impurity model, where the spin cannot be screened at $\phi=\pi$ .

C. Comparison with Quantum Dots

The authors highlight a crucial distinction between the lattice problem and single-impurity quantum dots:

In a quantum dot SAIM, the doublet phase (M-like) corresponds to a $\pi$ -junction with a sign-reversed but finite Josephson current.
In the lattice problem, the self-consistency condition leads to the complete suppression of the critical current in the M-phase, rather than a sign reversal.

5. Significance and Implications

Fundamental Physics: The work clarifies the nature of the Mott metal-insulator transition in the presence of superconductivity, demonstrating that strong correlations can fundamentally alter the proximity effect, rendering the system "Josephson-inactive."
Experimental Relevance: The results are directly applicable to van der Waals heterostructures (e.g., Nb $_3$ $_{3}$ Br $_8$ $_{8}$ or heavy-fermion lattices on superconductors).
- The predicted hysteretic switching between high-current (S-phase) and low-current (M-phase) states suggests a mechanism for tunable Josephson junctions.
- The suppression of critical current in the M-phase could be observed as an anomalously small $I_c$ in junctions with correlated barriers.
Tunability: Phase bias and junction transparency ( $\Gamma$ ) act as "tuning knobs" to switch between conducting and insulating regimes, offering a pathway for designing novel quantum devices based on correlated materials.

In summary, the paper demonstrates that a proximitized Mott insulator can exist as a distinct phase with vanishing supercurrent, driven by the specific structure of the self-energy in strongly correlated systems, and that this state can be switched to a superconducting state via coupling strength or phase bias.

Strongly correlated Josephson junction: proximity effect in the single-layer Hubbard model