Optical Voltage Profiling of 2D Semiconductors via Proximal Exciton Sensing

arXiv:2608.18382 · cond-mat.mes-hall, cond-mat.str-el, physics.optics · Submitted 2026-08-18 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: I'm Kai.

Mira: And I'm Mira, here with Lev.

Kai: We're looking at the paper: "Optical Voltage Profiling of 2D Semiconductors via Proximal Exciton Sensing".

Mira: Here's the gist of the paper. Optical voltage profiling introduces a noninvasive wide-field technique that uses a proximal monolayer exciton sensor to map internal potentials in 2D devices using only source–drain contacts The optically derived resistance exhibits a metal–insulator crossover near the resistance quantum h/e2 despite MΩ-scale two-terminal resistances in the metallic region Core Methodology The technique replaces local voltage probes with a proximal monolayer MoSe2 exciton sensor, which converts the target's local electrostatic potential into spatially resolved modulations of exciton reflectance This is achieved by isolating the sensor by thin hexagonal boron nitride, allowing it to convert the target's local electrostatic potential into spatially resolved modulations of exciton reflectance.

Kai: Through pixel-wise in situ calibration, these signals yield quantitative two-dimensional voltage maps of an actively biased semiconductor device.

Mira: Sensing Mechanism and Calibration The sensing mechanism relies on the target acting as a local gate for the sensor, where a source–drain voltage applied to the target induces a position-dependent potential that shifts the interlayer bias and thereby the local Fermi level in the sensor.

Kai: This shift changes the balance between neutral exciton XX0 and trion XX−, providing an optical transduction of the target’s voltage.

Title and authors: Kai: So we're looking at this paper, "Optical Voltage Profiling of 2D Semiconductors via Proximal Exciton Sensing," and the authors are Ha-Leem Kim, Hyungbin Lim, Yuanyi Yang, Ruishi Qi, Ruichen Xia, Can Uzundal, Takashi Taniguchi, Kenji Watanabe and Feng Wang <ref:2608.18382#pg1,Optical Voltage Profiling of 2D Semiconductors via Proximal Exciton Sensing>. Mira what's the basic idea behind this title?

Mira: Well it’s trying to figure out how to map internal potentials in these two-dimensional materials without actually touching them with physical probes. It’s about using light to see the voltage distribution inside a device.

Lev: So, you're not putting probes on the chip? That sounds hard for getting good spatial resolution.

Kai: Exactly. They use a proximal monolayer exciton sensor to map those internal potentials using only source and drain contacts. It’s noninvasive wide-field mapping of the internal potential landscape in these devices.

Mira: The implication is that we can get a 2D map of the voltage profile across a semiconductor just by looking at the light reflected off it, and they find they can do this even with huge metal to insulator resistances <ref:2608.18382#pg1>.

Lev: Huge resistances are a killer for any measurement trying to see things happening inside the channel.

Kai: That’s what they’re tackling; those M-ohm scale two-terminal resistances usually just hide whatever interesting physics is actually happening in the transport channel itself.

The paper's summary: Mira: So, what they actually did in this paper is introduce this technique. They use a monolayer molybdenum disulfide sensor placed very close to the target to convert the target’s local electrostatic potential into spatial modulations of exciton reflectance.

Lev: How do you get that potential information out of just light? That sounds like a massive signal processing challenge.

Kai: The sensing mechanism relies on the target acting like a local gate for that sensor. When you apply a source drain voltage to the target, it creates a position-dependent potential that shifts the interlayer bias and changes the local Fermi level in the sensor.

Mira: That shift then changes how much of neutral exciton XXzero versus trion XX minus is present optically. So they are doing an optical transduction of that voltage via those excitons.

Lev: That’s clever, using the balance between different exciton states to encode the potential change you're trying to measure. What about imaging?

Kai: For wide-field imaging, they use incoherent Bessel illumination made by an axicon and a rotating diffuser to stop speckle from blurring the picture. They also focus that light onto a small region of interest, limiting optical power to less than sixty nano watts.

Mira: And then for calibration, they use an in situ calibration phase where they find the zero crossing of a specific signal called delta I Vcal to extract the local voltage of the target layer pixel by pixel.

The paper's improvements: Kai: They’ve actually pointed out some ways this technique can be improved for practical use. One improvement is automating the extraction of those two-dimensional maps from a single optical measurement setup.

Mira: That sounds like they’re proposing a calibration phase where you run a uniform potential measurement first, and then follow it up with a small differential bias measurement to get the map automatically.

Lev: That would reduce the manual effort needed for pixel-wise reconstruction, which is usually where these complex optical methods get bogged down.

Kai: Right, because they’re using that digital lock-in signal subtraction method where the zero crossing of delta I Vcal directly yields the local voltage of the target layer. It streamlines that process significantly.

Mira: And Lev, they are also suggesting a way to use AI to help with reconstruction. They suggest training an AI to solve an inverse problem constrained by charge conservation and Ohm's law so it can reconstruct local conductivity maps from those measured potential maps.

Lev: So the AI isn't just guessing; it’s solving a physical constraint problem, which is much more robust than just looking at the image.

Kai: And they mentioned using a forward-adjoint method within a finite element framework to solve that conductivity reconstruction problem, and they added regularization terms like Tikhonov or Total Variation to keep those sigma fields from getting too messy.

Conclusion: Mira: So to wrap up the Optical Voltage Profiling of 2D Semiconductors via Proximal Exciton Sensing paper, the main implication is that we can finally get quantitative two-dimensional voltage maps of actively biased semiconductor devices using only source and drain contacts <ref:2608.18382#pg1,Optical Voltage Profiling of 2D Semiconductors via Proximal Exciton Sensing>.

Kai: They showed that they can resolve channel resistances below one kiloohm even when the two-terminal resistance is in the megaohm scale. Plus, they found a metal to insulator crossover near the quantum of resistance h over e squared in the bilayer molybdenum disulfide device.

Lev: That crossover point, nMIT around seven times ten to eleven centimeters negative squared, that’s a concrete number we can use for real hardware design considerations for error correction overhead.

Mira: They also showed that conductivity inhomogeneity actually persists from the insulating regime right into the metallic regime, with current flowing preferentially through the higher-conductivity domains across the whole system.

Kai: It shows that even when things look metallic optically, there’s still a real structure of different conductivity regions underneath that you can map.

Lev: For me, what this means for hardware is knowing where those insulating regions are is crucial because that's where charge traps and noise usually live in qubits.

Mira: So the optical voltage profiling technique, as described in Optical Voltage Profiling of 2D Semiconductors via Proximal Exciton Sensing, gives us a powerful way to characterize these materials without destroying the sample with probes <ref:2608.18382#pg1,Optical Voltage Profiling of 2D Semiconductors via Proximal Exciton Sensing>.

University of California, Berkeley Department of Physics University of California, Berkeley Materials Sciences Division Lawrence Berkeley National Laboratory Kavli Energy NanoScience Institute University of California, Berkeley Lawrence Berkeley National Laboratory Research Center for Materials Nanoarchitectonics National Institute for Materials Science Research Center for Electronic and Optical Materials National Institute for Materials Science

cond-mat.mes-hall, cond-mat.str-el, physics.optics

Submitted: 2026-08-18

Updated: 2026-11-14

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 92/100

The gist: Optical voltage profiling introduces a noninvasive wide-field technique that uses a proximal monolayer exciton sensor to map internal potentials in 2D devices using only source–drain contacts The…

Key concepts

Proximal Monolayer Exciton Sensor
This sensor is placed close to the target material and converts its local electrostatic potential into changes in exciton reflectance. It works by isolating the sensor with hexagonal boron nitride, allowing it to sense the target's voltage directly through optical signals.
Metal-Insulator Crossover (MIT)
The paper observes a transition where resistance behavior shifts from metallic to insulating near the quantum of resistance h/e². This is driven by carrier density, indicating that conductivity in the material changes fundamentally based on how many charge carriers are present.
Inverse Problem Solving with AI
Artificial intelligence is used to reconstruct local conductivity maps from measured voltage maps. This reconstruction is constrained by physical laws like charge conservation and Ohm's law, allowing the AI to determine how current flows through the material based on the optical data.

Terminology

Summary

Optical voltage profiling introduces a noninvasive wide-field technique that uses a proximal monolayer exciton sensor to map internal potentials in 2D devices using only source–drain contacts The optically derived resistance exhibits a metal–insulator crossover near the resistance quantum h/e2 despite MΩ-scale two-terminal resistances in the metallic region

Core Methodology

The technique replaces local voltage probes with a proximal monolayer MoSe2 exciton sensor, which converts the target's local electrostatic potential into spatially resolved modulations of exciton reflectance This is achieved by isolating the sensor by thin hexagonal boron nitride, allowing it to convert the target's local electrostatic potential into spatially resolved modulations of exciton reflectance. Through pixel-wise in situ calibration, these signals yield quantitative two-dimensional voltage maps of an actively biased semiconductor device. p. 0

Sensing Mechanism and Calibration

The sensing mechanism relies on the target acting as a local gate for the sensor, where a source–drain voltage applied to the target induces a position-dependent potential that shifts the interlayer bias and thereby the local Fermi level in the sensor. This shift changes the balance between neutral exciton XX0 and trion XX−, providing an optical transduction of the target’s voltage. p. 1 For wide-field imaging, incoherent Bessel illumination generated by an axicon and a rotating diffuser is employed to suppress speckle. p. 2 To account for variations, an in situ, spatially resolved calibration is implemented where the digital lock-in signal ΔI(Vcal) = Ical(Vcal) − Ibias is used to extract the local voltage of the target layer by finding the zero-crossing of ΔI(Vcal). p. 3

Transport Characterization

The method is validated on bilayer MoSe2 where channel resistances below 1 kΩ are resolved despite MΩ-scale two-terminal resistances. The study investigates the carrier-density-driven metal–insulator transition (MIT), observing a crossover near the quantum of resistance h/e2. p. 0 Reconstruction of local conductivity and current flow shows that conductivity inhomogeneity persists from the insulating regime well into the metallic regime, with current preferentially routed through higher-conductivity domains throughout. p. 5

Quantitative Analysis of MIT

For a quantitative analysis, voltage maps are used to reconstruct local conductivity by solving an inverse problem constrained by charge conservation and Ohm's law. The extracted average channel resistance R□(n) exhibits a metal–insulator crossover near the quantum of resistance h/e2. p. 5 The critical density of the metal–insulator transition is identified as nMIT ∼ 7 × 1011 cm−2 based on temperature-independent resistivity behavior. p. 4

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Automated extraction of two-dimensional voltage maps from single optical measurements can be achieved by implementing an in situ calibration process that involves a calibration phase (uniform potential) followed by a biasing phase (small differential bias).

  2. The improved system can perform pixel-wise voltage reconstruction using the digital lock-in signal subtraction method, where the zero-crossing of delta I(Vcal) yields the local voltage of the target layer.

  3. The AI can be trained to solve an inverse problem constrained by charge conservation and Ohm's law to reconstruct local conductivity maps from measured potential maps, as described in Supplementary Note 2.

  4. The system can utilize a forward-adjoint method within a finite-element framework to solve the conductivity reconstruction problem, including regularization terms like Tikhonov or Total Variation to suppress unphysical variations in sigma fields.

  5. The AI can perform density-dependent resistance analysis by extracting average channel resistance R = Delta V/ISD optically and plotting it against carrier density, identifying the metal-insulator transition via temperature-dependent traces.

Related papers