Ultrafast light switch for optical data components

By pairing a nanostructure of silver with an atomically thin semiconductor layer, a team of researchers from Italy, the U.K., and Germany, led by Christoph Lienau, a physics professor at the University of Oldenburg, created an ultrafast switching mirror device that can serve as an optical transistor—and it clocks in at a speed 10,000x faster than today’s electronic transistors.

The team set out to find a material with reflective properties that can be switched within a few femtoseconds (fs) via a highly focused laser beam.

Polaritons, a hybrid excitation of light and matter states, are quasiparticles that can be created when matter excitations, such as excitons, strongly interact with electromagnetic fields. It can be achieved by strongly confining the light—either between two extremely close mirrors within microcavities or via plasmonic nanostructures.

The most direct signature of these strong couplings is a periodic energy transfer between light and matter excitations. “We’re interested in probing these energy transfer processes directly in the time domain,” says Daniel Timmer, a postdoc working with Lienau. “Unfortunately, these Rabi oscillations are difficult to observe because it requires a time resolution on the order of 10 fs.”

During the past few years, the team developed a highly sensitive and ultrafast setup to perform two-dimensional electronic microscopy (2DES), one of the best-suited laser spectroscopy methods to track how energy flows within hybrid systems. “While these 2DES experiments are very powerful, the results are usually rather complex and their interpretation often requires extensive modeling to extract all of the information they contain,” says Timmer.

Recently, the team managed to probe such ultrafast Rabi oscillation dynamics using 2DES for the first time within a plasmonic nanostructure based on molecular quantum emitters.

“In our new work, we were hoping to see a similar ultrafast and periodic energy transfer using a nanostructure made of an atomically thin semiconductor. These monolayer materials are currently a hot topic for quantum materials with extensive experimental and theoretical work being done,” says Timmer.

Nanostructure design work

The team designed a nanostructure in which a monolayer of tungsten disulfide is deposited onto a periodic array of tiny nanometer-sized slits within a silver film. The slits couple far-field light to surface plasmon polariton fields at the surface of the array.

“We obtained a first glimpse into the coherent energy between plasmons and excitons within the tungsten disulfide film,” says Lienau. “And we discovered the polaritons based on these atomically thin semiconductors are very short-lived and rapidly relax into dark excitations of the material within <50 fs. Using 2DES, we could see how not only the polaritons decay but also how microscopic many-body interactions within the material impact the polaritons. Such an ultrafast polariton transition into dark states may be used for switching light by light with unprecedented speed.”

Design and fabrication of the team’s sample, a metallic nanostructure covered with a monolayer of a transition metal dichalcogenide, was done in collaboration with Andrea Ferrari, a professor of nanotechnology the director of the Cambridge Graphene Center, and his group in Cambridge (U.K.).

“Before producing our nanostructure, we designed it via numerical simulations to optimize our geometry and fabrication parameters for the desired interaction between the monolayer excitons and plasmons of the nanostructure,” says Moritz Gittinger, a postdoc working with Lienau. “We fabricated our structure—an array of nanoslits—within a silver film via focused gallium ion beam milling. The ion beam precisely removes material from the silver film to produce an array of 45-nm-wide and -deep nanoslits with a period of 495 nm. Our partners in Cambridge added a thin layer of isolating aluminum oxide and placed a monolayer of tungsten disulfide atop the 20 × 50-µm² structure.”

How does 2DES work?

For their ultrafast 2DES experiment, the team uses a home-built laser system—a noncollinear optical parametric amplifier—that produces 9-fs laser pulses at a 175-kHz repetition rate.

“2DES is essentially a pump-probe type of experiment, in which we use ultrashort laser pulses to excite and probe the system with a time resolution only limited by pulse durations,” says Timmer. “Changing the pump-probe delay allows us to follow the system dynamics. In 2DES, we split the laser beam to create three laser pulses with variable delays and focus them onto our sample. The first two pulses are to excite the system, and we can introduce a stable delay between them using a special type of passively phase-stable interferometer, a translating wedge-based identical pulses encoding system (TWINS) invented by Giulio Cerullo, a physics professor, and his team at Polytechnic University of Milan in Italy.”

By scanning this delay, the excitation pulses effectively generate a frequency comb with variable comb spacing. “It allows us to modulate, whether we excite the specific optical resonances in the system or not, depending on the photon energy needed for the optical transition,” says Timmer. “This trick gives us information about the energy of the excitation pathway via a Fourier transform. The third pulse, labeled with the probe, then interrogates the system after some finite evolution, so that we can measure how the reflection of the sample is altered by the pump pulses.”

Measuring the probe reflectivity with a fast and sensitive spectrometer lets the team obtain information about the detection energy. The result is a set of energy-energy maps that correlates these excitations and detection energies of the system as a function of the delay between pump pulses and the probe. Compared to more-often-applied pump-probe spectroscopy, in which only a single pump pulse is applied, this extra excitation dimension allows the researchers to discover hidden interactions and pathways into quantum systems.

“One of our most surprising observations was the large enhancement in optical nonlinearity of the hybrid nanostructure,” says Timmer. “While the monolayer only shows a weak optical nonlinearity (pump-induced change in reflectivity) when placed on a bare silver film, simply milling an array of nanoslits into the silver film before adding the monolayer dramatically enhances its nonlinearity by more than 20x.”

These nanoslits allow for extremely efficient excitation of surface plasmons, and their hybridization with excitons of the monolayer boosts the optical nonlinearity substantially. In the team’s previous work with molecular emitters they didn’t observe this type of enhancement, and they also weren’t expecting it in this case.

During the past few years, the team has worked to improve the stability and sensitivity of their 2DES setup to reach a high resolution of 10 fs or less. “We did it via ytterbium-based laser system and home-built noncollinear optical parametric amplifiers. Today, we can acquire almost shot-noise limited spectra at a rate of 100 kHz with very high short- and long-term stability—even up to several days,” says Timmer. “This is a gamechanger compared to our previously used 5-kHz system based on a titanium-sapphire amplifier. The dramatic boost in sensitivity and speed opens up a new way of working and performing experiments.”

Simulations work

2DES experiments tend to provide rich experimental data, but interpretation of these types of experiments can be challenging. “This is particularly true for the 2DES spectra of molecular and solid-state nanostructures, because the electronic dephasing times of these systems are usually rather short and the spectra are often congested,” says Lienau. “There’s an intense worldwide effort toward improving the theoretical understanding of 2DES spectra—and, in particular, the signatures of polaritons within these spectra.”

For this work, the team performed their own quantum mechanical model simulations to better understand the nonlinear optical properties of the polaritons and their ultrafast dynamics via 2DES signatures. It allowed them to learn about the signals within the nanostructure—especially the difference between coherent polaritons, which can in principle represent a coherent energy transfer between light and matter excitations, and incoherent polaritons and dark states. Since the latter can have similar optical nonlinearities, it’s difficult to distinguish them from the initially excited coherent polaritons. The team could see their transition by combining 2DES with their model simulations.

Photonic computing based on polaritons

There’s a push toward ultrafast switching of light by light for photonic computing based on polaritons. Within the context of the team’s work, it means how quickly can the sample be reversibly changed and how can pump pulse energy (the number of photons) needed for this switching process be minimized?

“Our partners in Milan recently showed that polaritons based on atomically thin semiconductors can indeed be used for this type of switching—with switching times of a few-hundred femtoseconds and moderate pump pulse energies of a few picojoules,” says Lienau. “The rapid loss of coherent polariton populations within our sample during <50 fs suggests that much faster switching times within the 10-fs range are realistic. Today, such ultrafast switching dynamics are still obscured by the presence of the rapidly populated dark polariton states. Suppressing these dark state populations appears to be a crucial task to make use of this fast switching in the future.”

Ultrafast coherent polariton dynamics

The team is making use of sensitivity and stability improvements, thanks to their 2DES experiment, and continuing to explore ultrafast coherent polariton dynamics.

Currently, studies that could probe the coherent polariton dynamics, such as Rabi oscillations, are limited despite considerable efforts and no shortage of ultrafast studies of polariton systems. “The main reasons are the short periods of oscillations (10 to 100 fs) and usually ubiquitous rapid dephasing mechanisms (<100 fs) that dampen such coherent phenomena and make their observation a challenge,” says Lienau.

Designing a system with sufficiently strong coupling strengths and much longer dephasing times “will open up new possibilities for coherent all-optical switching and information processing,” Lienau adds. “The coupling of atomically thin semiconductor layers to low-loss photonic cavities may be of crucial importance for this. Work in this direction is currently underway. Also, it would be exciting to further enhance the time resolution of 2DES to the 1-fs regime or below. It could provide an entirely new glimpse into the quantum dynamics of functional nanostructures and novel quantum materials.”

FURTHER READING

D. Timmer et al., Nat. Nanotechnol. (2026); https://doi.org/10.1038/s41565-025-02054-4.