A wave of sound passing beneath a semiconductor only three atoms thick can alter the light it emits. The effect repeats in billionths of a second, too quickly for an ordinary time-averaged spectrum to reveal which part of the wave caused which optical change. Researchers in Japan solved that timing problem by making the acoustic wave and their laser share a clock.
The experiment used monolayer tungsten diselenide, WSe₂, a two-dimensional semiconductor about 0.7 nanometers thick. Light creates excitons in the material: bound states of negatively charged electrons and the positively charged vacancies, or holes, they leave behind. When an exciton recombines and emits a photon, the photon’s energy, intensity, spectral width and decay can disclose how the material’s electronic states are changing.
The collaboration included doctoral student Yuta Takahashi, former master’s student Takumi Yamamoto and master’s student Hidetoshi Kanzawa of Keio University’s Graduate School of Science and Technology; Keio physics assistant professor Shun Fujii, who is also a visiting researcher at RIKEN; and RIKEN Center for Advanced Photonics team director Yuichiro K. Kato. Kato is concurrently chief scientist of RIKEN’s Nanoscale Quantum Photonics Laboratory.
The eight-author paper appeared online in Science Advances on August 21, or August 22 in Japan. It is titled “Multidimensional spectro-temporal imaging of exciton dynamics under propagating acoustic strain in two-dimensional semiconductors.” The authors reported no competing interests.
Interlocking electrodes launch sound across a surface
A surface acoustic wave is a mechanical disturbance concentrated near the surface of a solid. The researchers patterned interdigitated electrodes on a piezoelectric substrate. A radio-frequency voltage applied to the comb-like electrodes launched a 377-megahertz wave. Monolayer WSe₂ placed on the device was alternately stretched and compressed as the crest and trough passed.
Even a small strain changes atomic spacing and shifts the edges of the valence and conduction bands. Exciton emission consequently changes energy. Yet acoustic strain in this type of device is typically below 0.1%, while the propagating wave also brings a piezoelectric electric field and radio-frequency heating. A spectrum that averages over many cycles can mix the three effects.
The team used a frequency-converted 532-nanometer femtosecond pulse laser. Its pulse repetition rate was phase-locked to the surface acoustic wave. By stepping the radio-frequency phase, the researchers sampled successive moments of the same repeatable 2.65-nanosecond cycle. An actively stabilized Michelson interferometer measured the minute vertical displacement of the surface so that each optical measurement could be compared with the wave itself.
Tension shifted the emission lower and brighter
During the tensile part of the acoustic cycle, exciton emission moved to lower photon energy and became brighter. During compression, it shifted to higher energy and dimmed. The spectral linewidth also oscillated with acoustic phase.
The peak-energy change was about plus or minus 2.2 millielectronvolts. Combining the optical response with interferometric surface-displacement measurements placed the maximum dynamic strain at about 0.045%. The energy movement is a measurable spectral shift; it does not mean the sample visibly changes between dramatically different colors to the naked eye.
Time-resolved photoluminescence then showed that the decay after laser excitation was also modulated at the 2.65-nanosecond acoustic period. The paper’s coupled rate-equation model attributes that behavior to strain-dependent intervalley scattering. The wave changes the energy separation between electronic states associated with the K and Q valleys, altering how excitons redistribute before radiative recombination.
The exciton emission continued to respond over multiple acoustic cycles while it decayed. A combined time-and-energy measurement found the emission peak oscillating at the same period during the decay, direct evidence that the instantaneous bandgap was being modulated rather than merely displaced in a time-averaged spectrum.
| Quantity | What it represents | Observed cycle dependence | What it does not mean |
|---|---|---|---|
| Emission energy | Photon energy, corresponding to color | Lower under tension and higher under compression; about ±2.2 meV. | Not a large naked-eye color switch. |
| PL intensity | Detected brightness of photoluminescence | Brighter under tension and dimmer under compression. | Time-averaged intensity can also be reduced by the wave’s electric field. |
| Linewidth | Spread of the emission spectrum | Oscillated with acoustic phase. | The experiment did not track only one fixed wavelength. |
| Decay dynamics | How emission weakens after laser excitation | Varied through strain-dependent intervalley scattering and persisted across cycles. | “Lifetime” does not mean device durability. |
Phase resolution separates strain from field and heat
At high radio-frequency drive, the time-averaged emission intensity and energy both decreased. The paper attributes the quenching primarily to exciton dissociation by the in-plane piezoelectric field carried with the wave. It attributes the dominant time-averaged red shift to radio-frequency heating. Those effects are why an on-versus-off comparison alone cannot isolate dynamic strain.
The advance lies in combining phase-resolved microscopy, time-resolved photoluminescence and interferometric surface-motion measurements under common synchronization. The researchers could place surface displacement, emission energy, intensity, linewidth and decay on the same acoustic phase axis and identify which components followed tension and compression.
The measurements are multidimensional in a practical sense: position, acoustic phase, photon energy and time after optical excitation can be connected. Each view answers a different question—where the wave is, what color is emitted, how much light is produced and how quickly the emission evolves.
What “lifetime of light” means in this report
The Japanese institutional announcement describes ultrafast changes in the “color, brightness and lifetime” of emission. The underlying measurement is a time-resolved photoluminescence decay: after a short laser pulse creates excitons, the detector follows how the emitted signal changes and fades.
This is not the operational lifetime of a semiconductor device. It is the timescale over which an emitting exciton population evolves through radiative recombination, nonradiative processes and scattering between electronic valleys. The experiment goes beyond reporting one fixed lifetime constant by mapping how the full decay curve depends on acoustic phase.
The distinction matters because optical function and material reliability are separate engineering questions. The timing and energy of emitted photons could be useful for an optical source or sensor. Whether a device can survive years of operation, tolerate defects or be manufactured reproducibly requires different tests.
- The team correlated monolayer WSe₂ emission with acoustic phase, position, photon energy and decay time.
- It connected about 0.045% maximum strain to a roughly ±2.2 meV emission-energy modulation.
- A model incorporating strain-dependent intervalley scattering reproduced key decay changes.
- The work did not establish room-temperature device performance, long-term reliability or scalable fabrication.
- The phase-stepped method reconstructs periodic behavior; it does not capture arbitrary one-time transients in a single shot.
From reading the acoustic clock to designing photon flow
Surface acoustic waves are electrically programmable through their frequency, amplitude, phase and direction. Coupling them to atomically thin materials could eventually produce moving strain landscapes that route excitons or determine when and at what energy photons are emitted—functions that static bending cannot provide in the same way.
The researchers identify exciton-transport devices, sensitive strain sensors and single-photon sources as possible destinations, especially if the method is combined with longer-lived excitons or quantum emitters. Each application will impose its own requirements for temperature, loss, coherence, photon statistics, power and material uniformity.
For now, the result is a measurement and control platform. It converts a cycle that was blurred into an average into a phase-resolved map of energy, brightness, spectral width and decay. Before sound can engineer the photon stream of a two-dimensional device, researchers need to know precisely how the material keeps time. This experiment supplies that clock.