Weak localization as probe of spin-orbit-induced spin-split bands in bilayer graphene proximity coupled to WSe$_2$

This paper demonstrates that high-quality bilayer graphene/WSe$_2$ heterostructures exhibit gate-tunable transitions between weak anti-localization and weak localization, providing direct spectroscopic evidence of proximity-induced spin-split valence bands driven by Rashba-type spin-orbit coupling.

The Gist

Imagine you have a super-fast, ultra-thin sheet of carbon atoms called graphene. It's like a highway for electrons (the tiny particles that carry electricity), where they can zip along without hitting any bumps. However, there's a catch: in pure graphene, electrons don't really care about their "spin" (a quantum property that makes them act like tiny spinning tops). For future computers that use spin instead of just charge, we need to make these electrons spin in a specific way.

Scientists have tried sticking graphene next to a special crystal called WSe2 (a type of transition metal dichalcogenide). Think of WSe2 as a "spin magnet" that tries to force the electrons in the graphene to spin. But usually, this effect is weak and hard to control.

This paper is about a clever new trick the researchers used to turn this weak effect into a powerful, tunable tool. Here is the story in simple terms:

1. The Sandwich and the "Magic Knob"

The researchers built a sandwich:

• Bottom Bun: A layer of graphite (acting as a gate).
• Filling: Bilayer Graphene (two sheets of graphene stacked like a double-decker bus) sitting on top of the WSe2 crystal.
• Top Bun: Another layer of graphite and a gold gate.

The secret sauce is that they can push and pull on this sandwich with two different knobs (voltages).

• Knob A (Bottom Gate): Controls how many electrons are in the system (like turning on a faucet).
• Knob B (Top Gate): Controls the "pressure" or electric field squeezing the sandwich.

Because the graphene is a double-decker bus, the electric field pushes the "passengers" (electrons) to either the top floor or the bottom floor. Since the WSe2 is only touching the bottom floor, it can only influence the passengers on that specific floor. By turning the knobs, the scientists can decide exactly which floor the electrons are on and how they behave.

2. The Traffic Jam vs. The Superhighway

In the world of quantum physics, electrons act like waves. When they travel, they can interfere with themselves, creating patterns.

• Weak Anti-Localization (WAL): Usually, when electrons have strong spin-orbit coupling (the "spin magnet" effect), they act like a chaotic crowd. If you send them through a maze, they tend to avoid going back the way they came. This makes the material conduct electricity better when you apply a tiny magnetic field. The researchers saw this "chaotic crowd" behavior when the electrons were in the conduction band (the upper energy levels).
• Weak Localization (WL): But here is the surprise. When they tuned the knobs to move the electrons into the valence band (the lower energy levels), the behavior flipped completely. The electrons suddenly started acting like a disciplined marching band. They preferred to retrace their steps, creating a "traffic jam" that made the material conduct electricity worse in a magnetic field.

3. The "Spin-Split" Discovery

Why did the traffic jam happen? This is the big discovery.

The researchers realized that the "spin magnet" (WSe2) didn't just push the electrons; it actually split the road in two.
Imagine a highway that suddenly splits into two lanes:

• Lane 1: For electrons spinning "Up."
• Lane 2: For electrons spinning "Down."

In the lower energy levels (the valence band), the "spin magnet" was so strong that it pushed all the "Down" spins into a different energy zone, leaving only the "Up" spins on the road.

Because there was only one lane left for the electrons to travel in, they lost their ability to be chaotic. They were forced to march in a single file, which caused the "Weak Localization" (the traffic jam) effect.

4. Why This Matters

This is like finding a switch that can instantly change a highway from a chaotic, multi-lane highway (where cars weave and dodge) into a single-lane, one-way street (where everyone must follow the same path).

• Proof of Concept: This is the first time scientists have directly "seen" (through electrical measurements) that the spin-orbit coupling actually splits the energy bands in this way. It's like looking at a map and seeing the road physically divide.
• Future Tech: This gives us a way to build "Spin Valves" or "Spin Filters." Imagine a computer chip where you can use a simple voltage knob to decide: "Today, only allow electrons spinning 'Up' to pass." This is the holy grail for spintronics—a new type of computing that is faster and uses less energy than today's technology.

The Bottom Line

The researchers built a high-quality, ultra-clean graphene sandwich. By using two gates to squeeze the electrons, they proved that they could force the electrons to split into separate spin lanes. When only one lane was open, the electrons behaved in a predictable, "localized" way, proving that the spin-orbit coupling was working exactly as the theories predicted. This opens the door to building future quantum devices that control electron spin with the precision of a light switch.

Technical Summary

1. Problem Statement

While combining graphene with transition metal dichalcogenides (TMDs) enhances the weak intrinsic spin-orbit coupling (SOC) of graphene, a major challenge remains: the lack of gate-tunability in standard graphene/TMD heterobilayers. This limitation hinders the development of tunable spintronic devices like spin valves or transistors.

Bilayer graphene (BLG) offers a solution due to its unique electronic structure. When an out-of-plane electric displacement field ($D$) is applied, a band gap opens, and charge carriers (electrons and holes) become spatially separated into distinct layers. Since proximity-induced SOC acts primarily on the layer in direct contact with the TMD, applying a displacement field allows researchers to switch which layer hosts the carriers, thereby tuning the effective SOC strength and the nature of the spin-split bands. However, direct spectroscopic evidence of this tunable, proximity-induced spin-splitting via transport measurements has been elusive.

2. Methodology

The authors fabricated and characterized a high-quality double-gated BLG/WSe2 heterostructure:

• Device Architecture: A bilayer graphene flake was encapsulated between hexagonal boron nitride (hBN) crystals. A graphite bottom gate and a narrow gold top gate were used to create a dual-gated region.
• Geometry: The top gate is smaller than the bottom gate, creating "leads" (single-gated regions) and a central "cavity" (double-gated region). This geometry allows for the electrostatic formation of lateral p-n-p or n-p-n junctions.
• Materials: High-quality WSe2 (bilayer) was selected via optical contrast and Raman spectroscopy (confirming the $B_{2g}$ interlayer mode).
• Measurements:
  • Transport: Low-temperature ($T = 60$ mK) two-terminal resistance measurements as a function of top gate ($V_{tg}$) and bottom gate ($V_{bg}$) voltages.
  • Finite-Bias Spectroscopy: Differential conductance ($dI/dV$) measurements to map the band gap ($E_g$).
  • Magneto-transport: Measurements of conductance changes ($\Delta \sigma$) under small perpendicular magnetic fields ($B$) to detect quantum interference effects (Weak Localization - WL, and Weak Anti-Localization - WAL).
  • Theory: Tight-binding Hamiltonian modeling of the BLG/WSe2 system to calculate band structures, Berry phases, and expected WL/WAL signatures based on Rashba ($\lambda_R$) and Valley-Zeeman ($\lambda_{VZ}$) SOC strengths.

3. Key Contributions

• Demonstration of Quasi-Ballistic Transport: The device exhibits exceptionally high mobility ($\mu_e \approx 260,000$ cm$^2$V$^{-1}$s$^{-1}$) and a mean free path ($l_m \approx 1.0-2.0$ $\mu$m) comparable to the top gate length ($2$ $\mu$m). This places the device in a quasi-ballistic regime, essential for observing clear quantum interference effects.
• Gate-Tunable Band Gap: The device successfully opens a clean, tunable band gap in BLG, with the gap width scaling linearly with the displacement field $|D|$. The absence of significant screening by the WSe2 layer indicates a high-quality interface.
• Direct Observation of Spin-Split Bands via WL: The paper provides the first direct transport spectroscopic evidence of proximity-induced spin-split valence bands in BLG. It demonstrates a distinct transition from Weak Anti-Localization (WAL) to Weak Localization (WL) as the Fermi level is tuned into the upper spin-split valence band.

4. Results

A. Electrostatic Control and Band Gap

• Resistance maps show a clear "diamond" shape in finite-bias spectroscopy, confirming the opening of a band gap in the double-gated region.
• The band gap energy ($E_g$) is tunable from 0 to $\sim$40 meV by varying the displacement field.
• The system allows independent control of the Fermi level ($\mu_F$) and the band gap, enabling the formation of p-n-p cavities where quantum interference is enhanced by internal reflections at the p-n junctions.

B. Magneto-Transport Signatures (WAL vs. WL)

• Weak Anti-Localization (WAL): When the Fermi level is in the conduction band (or deep in the valence band where both spin-split bands are occupied), the system exhibits a WAL peak (a conductance dip at zero field). This is consistent with Rashba-type SOC inducing a Berry phase of $\pm \pi$.
• Weak Localization (WL): A remarkable finding is the emergence of a sharp, pronounced WL peak (conductance dip) when the Fermi level is tuned to the very top of the valence band (specifically for negative displacement fields, $D < 0$).
  • This WL signal is observed only when the transport is dominated by the upper spin-split valence band.
  • The amplitude of the WL signal is up to an order of magnitude larger than the WAL signal.
  • As hole density increases and the lower spin-split band becomes populated, the WL signal suppresses, reverting to WAL.

C. Theoretical Interpretation

• Berry Phase Analysis: Calculations show that near the band edge (K-point) of the upper spin-split valence band, the accumulated Berry phase approaches zero. A Berry phase of 0 leads to constructive interference and WL, whereas a phase of $\pi$ leads to destructive interference and WAL.
• SOC Splitting Estimation: By analyzing the gate voltage range where WL is observed, the authors estimate the proximity-induced spin splitting ($\Delta_{SOC}$) to be approximately 2 meV at $D = -0.3$ V/nm and 3 meV at $D = -0.4$ V/nm. These values align with theoretical predictions.
• Scattering Model: A scattering model incorporating the cavity dimensions and Fermi velocity successfully reproduces the experimental trend: WL amplitude increases with hole density until the second band is populated, after which it drops.

5. Significance

• Spectroscopic Proof: This work provides unambiguous spectroscopic evidence that proximity coupling to TMDs induces spin-split bands in BLG, and that this splitting can be tuned and probed via transport.
• Berry Phase Control: It demonstrates that the Berry phase in graphene can be continuously tuned from $\pi$ (WAL) to $0$ (WL) by controlling the band occupancy relative to the spin-splitting, a fundamental requirement for topological spintronics.
• Platform for Spintronics: The combination of high mobility, low disorder, and precise gate control establishes BLG/WSe2 as a robust platform for future spin-based quantum technologies, including spin filters, spin valves, and quantum dots with tunable spin-orbit coupling.
• Methodological Advance: The use of quasi-ballistic p-n-p cavities to enhance quantum interference signals offers a new strategy for probing SOC effects in 2D materials where diffusive transport might obscure these features.