Building 3D superconductor-based Josephson junctions using a via transfer approach

This paper presents a gentle, lithography-free via transfer method to construct high-quality 3D superconductor-graphene Josephson junctions with low contact resistance, enabling the observation of gate-tunable supercurrents and providing a pathway for engineering novel superconducting heterostructures on sensitive materials.

The Gist

The Big Idea: Building a "Super-Bridge" Without Breaking the Road

Imagine you want to build a bridge between two islands. One island is made of a special, fragile material (like a delicate sheet of glass) that represents Graphene. The other island is a heavy, powerful engine (a Superconductor) that can carry electricity with zero resistance.

The goal is to connect them so the "super-power" of the engine flows smoothly onto the glass island. This is called the Superconducting Proximity Effect.

The Problem:
Usually, to build this bridge, engineers use a "construction crew" that involves harsh chemicals, high heat, and heavy machinery (lithography and deposition). Think of this as trying to lay a heavy steel bridge onto a sheet of ice using a sledgehammer. The result? The ice cracks, the surface gets dirty, and the connection is weak. The electricity gets stuck at the border.

The Solution (The "Via Transfer" Method):
The researchers in this paper came up with a gentler way. Instead of building the bridge on top of the fragile island, they built the bridge inside a mold first, then gently lifted it and placed it down like a stamp.

They call this the "Via Transfer" approach.

1. The Mold: They took a piece of hard, protective rock (Hexagonal Boron Nitride, or h-BN) and dug a small, smooth pit into it.
2. The Casting: They poured their super-conductive metal (Niobium Nitride, or NbN) into that pit. Because the pit was smooth, the metal settled perfectly flat.
3. The Transfer: They picked up the whole rock with the metal inside and gently lowered it onto the fragile graphene island.
4. The Result: The metal fits perfectly into the graphene's surface, like a key sliding into a lock, without scratching or damaging the graphene.

What Did They Discover?

Once they built this gentle bridge, they tested how well electricity flowed across it. Here is what they found, translated into everyday terms:

1. A Smooth Highway (Low Resistance)
Usually, when two different materials touch, there is a "bumpy border" where electricity gets stuck. In this experiment, the border was so smooth that the resistance was incredibly low. It's like going from a dirt road to a perfectly paved highway with no speed bumps.

2. The "Super-Current" (Josephson Effect)
They found that they could send a "super-current" (electricity with zero resistance) across the graphene.

• The Analogy: Imagine a crowd of people (electrons) trying to walk through a narrow hallway. Usually, they bump into each other and slow down. But because the bridge was so smooth, they started walking in perfect unison, like a synchronized dance troupe, moving without any friction.
• The Control: They could turn this super-current on and off, or change its strength, just by applying a tiny voltage (like turning a dimmer switch). This is crucial for building future quantum computers.

3. The "Magic Pattern" (Fraunhofer Pattern)
When they applied a magnetic field, the super-current created a specific wave pattern (like ripples in a pond). This proved that the current was flowing evenly across the entire width of the graphene, not just in a few lucky spots. It confirmed the bridge was uniform and high-quality.

4. The "Ghost Gap" (Induced Gap)
The researchers noticed something interesting. The super-conductor (NbN) has a huge "energy gap" (a big jump in energy levels). However, when it touched the graphene, the graphene only got a tiny "ghost" of that gap.

• The Analogy: Imagine a loudspeaker (the superconductor) playing music next to a quiet room (the graphene). You expect the room to get loud, but it only gets a faint whisper.
• Why? They realized the metal they used (NbN) was a bit "messy" or disordered inside, like a speaker with a broken wire. This messiness prevented the full power from transferring. They suggest that if they use a "cleaner" metal in the future, the "whisper" could become a "shout," making the device even better.

Why Does This Matter?

This paper is a breakthrough for two main reasons:

1. It's Gentle: This method works for materials that are too sensitive for traditional construction. If you tried to build a bridge on a topological insulator (a very sensitive material used in quantum physics) using old methods, you'd destroy it. This "stamp" method keeps it safe.
2. It's Scalable: Because the method is clean and works well, it opens the door to building complex quantum circuits. We might soon see quantum computers built using these "gentle bridges" that can perform calculations impossible for today's computers.

In a Nutshell

The team figured out how to connect a powerful super-conductor to a fragile quantum material without breaking it. They used a "mold-and-stamp" technique to create a perfect, smooth connection. While the connection wasn't quite as strong as they hoped (due to some internal "messiness" in the metal), the method itself is a game-changer for building the next generation of quantum devices.

Technical Summary

1. Problem Statement

The integration of superconductors with two-dimensional (2D) materials and sensitive surfaces (e.g., topological insulators, transition metal dichalcogenides) is crucial for exploring emergent quantum phenomena like topological superconductivity and 0-$\pi$ Josephson junctions. However, conventional fabrication methods face significant challenges:

• Interface Damage: Direct sputtering or evaporation of superconductors (SCs) onto 2D materials often introduces surface damage, amorphous layers, and chemical contamination.
• Air Sensitivity: Many promising 2D materials (e.g., topological insulators like Bi$_2$Se$_3$) degrade rapidly in air, making standard lithography and deposition difficult without in-situ cleaning that can still deform the interface.
• Limited Material Scope: While van der Waals (vdW) superconductors (like NbSe$_2$) allow for gentle transfer, they are limited to specific materials. There is a lack of reliable methods to create high-quality, low-resistance interfaces between 3D superconductors (like NbN) and sensitive 2D materials without polymer residues or lattice mismatch issues.
• Previous Failures: Prior attempts to create superconducting via contacts to materials like ZrTe$_5$ and MoS$_2$ failed to observe Josephson currents, likely due to poor interface transparency.

2. Methodology

The authors developed a via transfer approach to construct 3D superconductor (NbN/Pd) to 2D material (graphene) Josephson junctions (JJs) with a "van der Waals-like" interface.

• Fabrication Process:
  1. Via Etching: Pits (approx. 500 nm wide) were etched into hexagonal boron nitride (h-BN) flakes using electron-beam lithography (EBL) and reactive ion etching (RIE).
  2. Deposition: A 45 nm NbN / 4 nm Pd film was sputtered directly into the etched pits. The Pd layer acts as a buffer to prevent NbN from sticking to the SiO$_2$ substrate and ensures a transparent interface with graphene.
  3. Dry Transfer: The NbN/Pd-embedded h-BN flake was picked up using a polymer (PC) stamp inside an Argon-filled glovebox. It was then sequentially stacked onto monolayer or bilayer graphene (MLG/BLG), followed by an h-BN dielectric and a graphite backgate.
  4. Device Completion: Standard lithography was used to pattern electrodes (Au/Cr) contacting the NbN strips and the exposed graphene edges.
• Characterization Techniques:
  • Microscopy: Scanning/Transmission Electron Microscopy (STEM), Energy-Dispersive X-ray Spectroscopy (EDX), and Electron Energy-Loss Spectroscopy (EELS) were used to verify the structural integrity and chemical purity of the interface.
  • Transport Measurements: Measurements were conducted in dilution refrigerators (base temp ~10 mK) to characterize normal state resistance, critical current ($I_c$), Fraunhofer patterns, and Andreev reflections.

3. Key Contributions

• Novel Contact Geometry: Successfully demonstrated a lithography-free, polymer-free method to create smooth, conformal interfaces between a 3D superconductor (NbN) and 2D graphene.
• Low Contact Resistance: Achieved a unit-length contact resistance of ~130 $\Omega \cdot \mu$m, comparable to conventional top/edge deposition techniques, proving that the via method does not compromise electrical coupling.
• Observation of Josephson Effect: For the first time in this specific geometry, the authors observed robust gate-tunable supercurrents, Fraunhofer patterns, and Andreev reflections in NbN/Pd-graphene junctions, overcoming previous failures with similar via contacts.
• Generalizability: Demonstrated the method's applicability to air-sensitive materials by successfully transferring NbN/Pd contacts to Bi$_2$Se$_3$ (a topological insulator) without bubble formation or degradation.

4. Key Results

• Interface Quality: STEM and AFM confirmed a void-free, atomically smooth interface with a root-mean-square (RMS) roughness of 1.8 Å. EDX/EELS maps showed no contamination or intermixing at the Pd/graphene interface.
• Superconducting Properties:
  • Critical Current: Gate-tunable critical currents ($I_c$) up to 110 nA were observed in bilayer graphene devices.
  • Fraunhofer Pattern: The magnetic field dependence of $I_c$ followed a standard Fraunhofer pattern, indicating a uniform supercurrent distribution across the junction width.
  • Proximity Gap: The devices operated in a diffusive regime. Analysis revealed a proximity-induced superconducting gap ($\Delta'$) in the graphene beneath the contact of 10–30 $\mu$eV.
• Mechanism Analysis:
  • Induced Gap Limitation: While the parent NbN gap ($\Delta_{NbN}$) is ~0.9 meV, the induced gap $\Delta'$ is only a few percent of this value.
  • Cause of Suppression: The authors attribute the small $\Delta'$ not to interface disorder (which was minimal) but to disorder within the NbN film itself. The sputtered NbN had a relatively low $T_c$ (~7 K) and high resistivity, suggesting strong scattering (short quasiparticle relaxation time $\tau_S$) in the superconductor reduces the induced gap according to the relation $\Delta_N \approx \Delta_S \tau_S / (\tau_S + \tau_N)$.
  • Transparency: Despite the small induced gap, the interface transparency was high ($\mathcal{T} \approx 0.8 - 0.92$), as confirmed by Andreev reflection measurements and BTK analysis.

5. Significance

• Enabling New Heterostructures: This work provides a robust, gentle fabrication route for constructing superconducting heterostructures on air-sensitive and damage-prone 2D materials (e.g., topological insulators, TMDs) where conventional deposition fails.
• Path to Topological Quantum Computing: By enabling high-quality SC-normal metal interfaces on topological insulators, this method paves the way for realizing Majorana zero modes and topological qubits.
• Material Optimization Insight: The study highlights that for via contacts to be effective, the quality of the 3D superconductor film (low disorder, high $T_c$) is as critical as the interface smoothness. Future improvements should focus on using higher-quality SC films (e.g., epitaxial Nb or Nb/Au) to maximize the induced gap.
• Scalability: The dry transfer, glovebox-compatible process is scalable and compatible with the fabrication of complex van der Waals heterostructures for next-generation quantum devices.