This work stands off to the side: there is almost nobody nearby, with the nearest neighbor at a distance of 0.542. Around it lie studies on Cooper pairs, semiconductors, superfluidity, and thermal conductivity. In this direction, researchers have already explored laser cooling, atomic light switches, wireless superconductivity, anomalous waves in superatoms, sound that charges batteries, and chips where light and sound work together without interference. But the ultrafast dynamics of hot carriers at interfaces in transition metal dichalcogenides remained a blank spot.
The researchers mechanically exfoliated thin flakes of MoS₂ and MoSe₂ onto a 285 nm SiO₂/Si substrate. Ni/Au contacts for MoS₂ and Ag for MoSe₂ were deposited by vacuum thermal evaporation to obtain ohmic interfaces without Schottky barriers. For electrical readout, they used on-chip terahertz optoelectronic sampling with a photoconductive switch based on low-temperature-grown GaAs (LT-GaAs) with a thickness of 2 µm. A femtosecond laser at a wavelength of 680 nm excited carriers at the metal-semiconductor interface, while a synchronized probe beam triggered the switch to measure the transient photocurrent with subpicosecond resolution. In addition, they performed pump-probe spectroscopy with differential reflection to link the electrical response to phonon dynamics.
At zero bias, the photocurrent relaxation times were 48.5 ps for MoS₂/Au and 14.2 ps for MoSe₂/Ag. Fourier transform gave intrinsic bandwidths of 4.4 GHz and 7.5 GHz, respectively. Spatial scanning showed that the relaxation time remains nearly constant (~44.6 ps) when the pump spot is moved ±4 µm from the interface, although the photocurrent amplitude falls exponentially. Measurements with applied bias revealed two components: a slow component that depends linearly on voltage and reverses polarity, and a fast component that is completely insensitive to bias. These data rule out drift transport and point to superdiffusion of hot carriers driven by an electron temperature gradient. Comparison with optical spectroscopy showed that the electrical extraction time quantitatively matches the scattering time of optical phonons into acoustic ones for all MoS₂ thicknesses (5, 13, and 30 nm). In MoSe₂, where the difference in atomic masses is smaller, the phonon bottleneck is weaker and the response time is shorter.
Unlike neighboring studies that examine steady-state or low-frequency phenomena, this work directly resolves, for the first time, the picosecond dynamics of hot-carrier extraction at a metal-semiconductor interface in a TMDC. It shows that device speed is determined not by carrier mobility but by the lifetime of hot phonons, which opens a new channel for controlling optoelectronic properties through phonon engineering. This is the first experimental verification that a phonon bottleneck can sustain superdiffusion for tens of picoseconds.
The topic could advance toward the targeted design of phonon spectra: choosing materials with the right difference in atomic masses, creating heterostructures with a tunable acoustic-optical gap, and using strain or doping to modify anharmonic phonon decay. It may also be possible to study other polar layered materials and interfaces with different metals to optimize the response time. Studying the effects of temperature and encapsulation on phonon dynamics is another promising direction.
The results will influence the development of ultrafast self-powered photodetectors, optical interconnects, and integrated optoelectronic chips based on two-dimensional materials. They are also important for the fundamental physics of nonequilibrium processes in low-dimensional systems.
Next steps include a systematic study of other dichalcogenides and metal contacts, as well as measuring phonon spectra using inelastic scattering to directly test the predictions. There are also plans to study the effects of encapsulation and temperature on the phonon bottleneck.
The work connects to the unsolved problem of speed limitations in optoelectronic devices due to low carrier mobility at room temperature. It shows that this limit can be circumvented by using nonequilibrium hot-carrier transport governed by phonon dynamics. This opens a new path for overcoming traditional limitations in semiconductor electronics.
This section is addressed to the author of the paper and their colleagues. Neighbouring works were found by meaning-based search across our archive, not by word overlap. We do not review: we show the field around the work.
What the work does
The work directly measures the ultrafast electrical response at a metal-semiconductor interface in MoS₂ and MoSe₂ using on-chip terahertz optoelectronic sampling. Relaxation times of 48.5 ps and 14.2 ps were obtained, corresponding to bandwidths of 4.4 GHz and 7.5 GHz, and the response was shown to be independent of position and bias, pointing to superdiffusion of hot carriers. The main conclusion: the macroscopic response time is synchronized with the microscopic scattering time of optical phonons into acoustic ones (the phonon bottleneck).
Where its strength is
Directly linking the electrical response to phonon dynamics on timescales of tens of picoseconds is a rare result, providing a quantitative basis for phonon engineering in optoelectronics.
Where it could go next
Worth trying
Why this may matter
If the phonon bottleneck really sets the speed limit, this offers a universal lever for designing ultrafast self-powered optoelectronic devices based on layered materials.
Nearby works in our archive
The archive is sparse around this work: the closest review is only 0.53, against a usual 0.64. It looks like few have drilled nearby.
In MoS₂, molybdenum atoms are twice as heavy as sulfur atoms, which creates a band gap in the phonon spectrum near 1 THz. Because of this, optical phonons cannot decay quickly into acoustic ones, and hot carriers remain energetic for longer.
The idea of controlling a material's properties through lattice vibrations echoes the concept of 'phonon computers' from science fiction, where information is carried by thermal waves rather than electrons.