Gate-Tunable Ambipolar Josephson Current in a Topological Insulator

This study demonstrates gate-tunable ambipolar Josephson current in molecular beam epitaxy-grown bulk-insulating (Bi,Sb)₂Te₃ thin films, revealing distinct transport behaviors between thin and thick films and establishing a critical foundation for realizing Dirac-surface-state-mediated topological superconductivity and Majorana modes.

The Gist

Imagine a special kind of material called a Topological Insulator (TI). Think of this material like a chocolate-covered marshmallow. The inside (the bulk) is an insulator, meaning electricity cannot flow through it—it's like the fluffy, non-conductive marshmallow. However, the outside (the surface) is a conductor, like the chocolate shell, where electrons can zip around freely.

In the world of quantum physics, these surface electrons are very special. They move in a way that is "locked" to their spin, making them perfect candidates for building future quantum computers. To study them, scientists want to turn this material into a Josephson Junction. You can think of a Josephson Junction as a narrow bridge connecting two islands of superconductors (materials where electricity flows with zero resistance). The goal is to see if the "marshmallow shell" (the TI surface) can carry a supercurrent across this bridge.

The Big Challenge
For years, scientists have struggled with a "leaky roof" problem. Even though they tried to make the inside of the marshmallow insulating, it was often still slightly conductive. This meant that when they measured the current, they couldn't tell if the electricity was flowing on the cool, special surface or just leaking through the messy inside. It was like trying to hear a whisper in a noisy room; the "bulk" noise drowned out the "surface" signal.

The Breakthrough
This paper reports a success story using a very high-quality "marshmallow" made of a material called (Bi,Sb)₂Te₃, grown layer by layer in a vacuum chamber. The researchers built tiny bridges (junctions) and used a "gate" (like a volume knob) to tune the material.

Here is what they found, explained simply:

1. The "Two-Way Street" (Ambipolar Current):
   Usually, electricity in these materials flows either with "positive" charges (holes) or "negative" charges (electrons), but not both easily. The researchers discovered that in their thinnest samples (5 layers thick), they could turn the "volume knob" (the gate) to switch the current from flowing with positive charges to flowing with negative charges. This is like a road that can instantly switch traffic direction based on a signal. This is called ambipolar behavior, and it proves the current is flowing through the special surface states, not the messy bulk.

2. The "Quiet Spot" (The Dirac Point):
   There is a specific setting on the volume knob where the material is perfectly balanced between positive and negative. In physics, this is called the "Dirac point." The researchers found that when they tuned the knob to this exact spot, the supercurrent didn't disappear completely, but it did get much weaker. It's like the road gets a little bumpy right in the middle, making it harder for the cars (electrons) to drive fast, but they can still get across.

3. The "Thick vs. Thin" Problem:
   When they made the material thicker (15 layers), the "leaky roof" problem came back. The current could still switch between positive and negative, but it became very lopsided. It was easy to get a strong current on the positive side, but the negative side was weak.

   • The Analogy: Imagine a thin sheet of paper (5 layers). If you paint a line on it, the paint soaks through evenly. But if you use a thick block of wood (15 layers), the paint might soak through the top but get stuck in the middle. The researchers used computer simulations to show that in the thick samples, the "bulk" (the wood inside) started interfering with the "surface" (the paint on top), making it hard to control the current cleanly.

4. Magnetic Sensitivity:
   The researchers also tested how these bridges held up against magnets. They found that when the current was flowing through the special surface states (especially near that "quiet spot" or Dirac point), the supercurrent was much more fragile and broke down easily in a magnetic field compared to when it was flowing through the bulk. This fragility is actually a good sign; it suggests the current is indeed traveling through the unique, delicate surface states rather than the robust, boring bulk.

The Conclusion
The paper claims that by growing these materials perfectly and making them thin enough, they have finally built a Josephson Junction where the supercurrent is clearly controlled by the special surface states. They demonstrated that this current can be tuned to flow with either type of charge (ambipolar).

This is a crucial step because it proves they can isolate the "special" physics from the "messy" background. The authors state that this success paves the way for creating Majorana modes (exotic particles that are their own antiparticles) and eventually building topological quantum computers. Essentially, they have cleared the noise so they can finally hear the whisper of the quantum world they are trying to harness.

Technical Summary

1. Problem Statement

Topological superconductivity (TSC) and Majorana zero modes (MZMs) are central to the development of fault-tolerant topological quantum computing. A promising platform for realizing these states involves hybrid structures where topological insulators (TIs) with strong spin-orbit coupling are proximitized by conventional $s$-wave superconductors.

However, a significant bottleneck has hindered progress:

• Bulk Conduction: In most TI-based Josephson junction (JJ) devices, transport is dominated by bulk conduction channels rather than the desired Dirac surface states.
• Lack of Ambipolarity: While graphene has demonstrated gate-tunable ambipolar supercurrent (carried by both electrons and holes), previous TI-based JJs have failed to show this behavior. They typically operate in regimes dominated by either bulk conduction or unipolar surface states.
• Separation Challenge: The inability to cleanly separate proximity-induced superconductivity in the Dirac surface states from bulk channels compromises the reliability of TI-based Majorana qubit schemes.

The goal of this work is to demonstrate a gate-tunable ambipolar Josephson current in a single device, providing an unambiguous signature of Dirac-surface-state-dominated transport.

2. Methodology

The researchers employed a combination of high-quality material growth, precise nanofabrication, and low-temperature transport measurements, supported by numerical simulations.

• Material Growth:

  • Material: $(Bi,Sb)_2Te_3$ thin films (a bulk-insulating TI).
  • Technique: Molecular Beam Epitaxy (MBE) grown on heat-treated insulating $SrTiO_3(111)$ substrates.
  • Optimization: The Bi/Sb ratio was optimized to tune the chemical potential near the charge neutral point (Dirac point).
  • Thickness: Films of varying thicknesses were grown, specifically 5 quintuple layers (QL) and 15 QL.

• Device Fabrication:

  • Structure: Lateral Josephson Junctions (JJ) and Superconducting Quantum Interference Devices (SQUIDs).
  • Electrodes: Niobium (Nb) superconducting contacts deposited via DC sputtering.
  • Geometry: Junction area $L \times W = 20 \text{ nm} \times 2 \text{ \mu m}$.
  • Gating: A global back gate using the $SrTiO_3$ substrate allows for fine-tuning of the chemical potential.
  • Etching: Argon plasma milling was used to etch away TI films outside the junction and electrode areas to enhance gating efficiency.

• Measurements:

  • Conducted in a dilution refrigerator at 10 mK.
  • Magnetic fields up to 9 T (out-of-plane) were applied.
  • Measurements included $I-V$ characteristics, differential resistance ($dV/dI$), and critical current ($I_c$) dependence on gate voltage ($V_g$) and magnetic field ($B_z$).

• Theoretical Modeling:

  • Method: Recursive Green's function method.
  • Model: A 3D TI Hamiltonian incorporating structural inversion asymmetry (SIA) between top and bottom surfaces, coupled with $s$-wave superconducting leads via Bogoliubov-de Gennes (BdG) formalism.

3. Key Contributions

1. First Demonstration of Ambipolar Josephson Current in TIs: The paper reports the first observation of gate-tunable ambipolar Josephson current in a TI-based JJ, where the supercurrent can be carried by both electrons and holes.
2. Thickness-Dependent Behavior: The study systematically compares 5 QL and 15 QL films, revealing that ambipolarity is robust in thinner films but suppressed and asymmetric in thicker films due to bulk channel coexistence.
3. Magnetic Resilience Analysis: The authors demonstrate that supercurrents near the Dirac point (dominated by surface states) are significantly less resilient to magnetic fields compared to those in the metallic (bulk) regime.
4. Theoretical Validation: Numerical simulations successfully reproduce the experimental asymmetry in thicker films, attributing it to the overlap of Dirac surface states with bulk conduction bands induced by SIA.

4. Key Results

A. 5 QL Device (JJ-1): Robust Ambipolarity

• Gate Tunability: The critical current ($I_c$) exhibits a pronounced "V-shape" dependence on gate voltage.
  • Dirac Point: At the charge neutral point ($V_g \approx V_{g0}$), $I_c$ reaches a minimum (~40 nA) but persists, confirming superconductivity through the Dirac surface states.
  • Off-Dirac: As the gate moves into the p-type or n-type regimes, $I_c$ increases significantly (up to ~200–210 nA).
• Resistance: The normal state resistance ($R_n$) peaks sharply at the Dirac point, while the product $I_c R_n$ (a measure of junction quality) reaches a minimum there but remains substantial (12–35 $\mu$V).
• Magnetic Field: The Fraunhofer pattern (interference of supercurrent with magnetic field) shows that while the supercurrent persists at the Dirac point, the side lobes are broader and less distinct, indicating reduced magnetic resilience compared to the doped regimes.

B. 15 QL Device (JJ-2): Asymmetry and Bulk Influence

• Reduced Ambipolarity: The device shows a strong asymmetry between p-type and n-type regimes.
  • p-type: $I_c$ varies significantly with gate voltage.
  • n-type: $I_c$ remains nearly constant and high, indicating the presence of gate-insensitive bulk conduction channels.
• Interpretation: Theoretical simulations suggest that in thicker films, the large chemical potential difference between top and bottom surfaces causes the top Dirac cone to shift and overlap with bulk bands. This creates a "precursor" to ambipolar transport that is hindered by the bulk.

C. Proximity Effect and Gap Analysis

• Gap Suppression: The induced superconducting gap ($\Delta$) extracted from differential conductance decreases as the chemical potential approaches the Dirac point (from ~20 $\mu$eV in doped regimes to ~12.5 $\mu$eV at the Dirac point).
• Ballistic Transport: The ratio $eI_c R_n / \Delta \approx 1$ suggests high interface quality and ballistic transport characteristics.
• Multiple Andreev Reflections (MAR): In the 15 QL device, MAR peaks were observed in the p-type regime, confirming coherent transport.

D. SQUID Devices

• Symmetric SQUIDs fabricated on 10 QL and 15 QL films confirmed the gate-tunability of $I_c$.
• The 10 QL device showed ambipolar behavior, while the 15 QL device remained largely in the p-type regime, consistent with the JJ results.
• No skewed current-phase relations (CPR) were observed, suggesting that pure surface-state transport has not yet been fully isolated from bulk mixing in the SQUID geometry.

5. Significance

• Validation of Surface State Transport: The persistence of Josephson current at the Dirac point, combined with the ambipolar nature in thin films, provides strong evidence that the supercurrent is mediated by the TI Dirac surface states rather than bulk channels.
• Pathway to Majorana Physics: By demonstrating that the chemical potential can be tuned across the Dirac point while maintaining a supercurrent, this work establishes a critical foundation for creating topological superconductivity and searching for Majorana zero modes in TI/superconductor hybrids.
• Material Optimization: The study highlights the critical importance of film thickness and bulk insulation. It suggests that thinner films (5 QL) are superior for isolating surface states, while thicker films suffer from bulk-surface mixing due to structural inversion asymmetry.
• Future Applications: This work paves the way for electrically tunable Majorana modes and scalable topological quantum computations using MBE-grown TI films.