An arXiv preprint reports that applying surface acoustic waves to two thin-film materials produced acoustoelectric currents that changed near charge-density-wave, or CDW, transitions. In NbSe3, the current polarity changed with temperature, and reversing the wave's propagation direction reversed the signal. A metallic tantalum control was nearly temperature-independent and did not show the same polarity reversal.
The document is an arXiv preprint, arXiv:2608.27959v1, dated 28 August 2026. Its central question was whether SAW-induced acoustoelectric currents are modulated at CDW transitions in NbSe3 and 2H-TaSe2, whether polarity depends on propagation direction, and whether strain-modified conductivity can account for the observed changes.
The direction mattered
One representative NbSe3 measurement made the directional effect especially clear. At 50 K, reversing SAW propagation reversed the acoustoelectric signal's polarity, while increasing SAW power increased its magnitude linearly. The authors describe this as a linear-response regime.
The devices used exfoliated NbSe3 and 2H-TaSe2 thin films on a 128° Y-cut black LiNbO3 substrate, with metallic Ta thin films as controls. The main measurements used SAW propagation along X + 90° or X. The crystallographic orientation of the 2H-TaSe2 flake was not determined.
Researchers measured acoustoelectric voltage across radio-frequency sweeps, fitted the spectra, and converted the result to current with each film's four-terminal resistance. They used lock-in detection synchronized to 100% amplitude modulation at 173 Hz. That setup allowed comparisons of current as temperature, SAW power and propagation direction changed.
Two temperature patterns
In NbSe3 with X + 90° propagation, resistivity anomalies appeared near approximately 140 K and 60 K, the higher and lower CDW transitions reported for the device. The acoustoelectric current had a distinct peak near 140 K and increased more strongly below 60 K. Its polarity was positive around the higher transition and negative elsewhere.
2H-TaSe2 showed a broader temperature pattern. Its acoustoelectric current divided by SAW power began increasing below approximately 120 K, reached a broad peak around approximately 95 K and then declined. The signal was positive from 115 K to 65 K and negative outside that interval. The analysis treats the link between the broad feature and a specific transition as possible, rather than certain.
The contrast with the Ta controls was notable: their current was nearly temperature-independent and did not reverse polarity regardless of film size. In the NbSe3 device using X propagation, the current still showed enhancement below the CDW transitions and a sign change near the first transition, but its polarity was opposite to the X + 90° device over almost the entire temperature range.
Additional NbSe3 devices reproduced the CDW-state acoustoelectric modulation. Their comparison suggested that flake thickness did not determine polarity, while X propagation again produced the opposite sign from X + 90° propagation. Some reproducibility devices were damaged before all temperatures could be measured.
A model points to strain
The paper's explanation uses a phenomenological rectification model, designed to capture the observed signal rather than serve as a definitive mechanistic test. It treats the SAW-induced electric field together with strain, while longitudinal strain perturbs the film's conductivity. Numerical calculations showed only weak temperature dependence in the calculated electric-field/strain prefactor. Within the model, the authors therefore treated strain-induced conductivity modulation as the main source of the temperature variation.
The model-derived normalized conductivity response to strain was much larger for the CDW materials than for Ta. The report lists the values as approximately 102 for 2H-TaSe2 and approximately 103 to 104 for NbSe3, compared with approximately negative 2 for Ta Device C. It describes the CDW-material estimates as orders of magnitude larger than Ta. These are model-derived estimates, not direct strain-response measurements.
An illustrative conventional-model calculation produced an acoustoelectric current of approximately 1.3 pA/mW, about four orders of magnitude below the experimental value. Under the tested conditions, the authors judged that contribution negligibly small. The comparison does not establish that the proposed strain mechanism is the sole explanation.
What the findings do not settle
The study's limits are important. The device comparisons were descriptive, the 2H-TaSe2 flake orientation was unknown, and the conductivity model was phenomenological and supported qualitatively by numerical calculations rather than direct strain-response data. No formal inferential tests or confidence intervals were reported.
The experiment also remained in a linear regime. The induced field was at most 0.01 V/cm, compared with a reported 10 V/cm threshold for NbSe3 thin films. Resistivity changes with and without SAW were negligible, and no CDW depinning or sliding was observed. The results therefore do not address behavior outside that linear regime.
Across the reported devices, the acoustoelectric signal was associated with CDW temperature patterns and with the direction of the applied wave. The proposed strain-modified-conductivity model offers a qualitative account of those changes, but the preprint does not establish a universal strain-sensitivity value or show that strain is the sole explanation.
Paper data and sources
Original title: Significant modulation of acoustoelectric current associated with charge density wave transitions
Authors: Natsumi Nikaido, Takuya Kawada, Koji Fujiwara et al.
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-28
DOI: Not available
Original paper · Full text