The achievement rests on a technique that has become one of the most powerful control knobs in attosecond science: two-colour waveform synthesis. Instead of hitting the tunnel junction with a single laser beam, the researchers combined two coherent light fields of different colours, a fundamental frequency and its second harmonic. When these two fields overlap, their sum is no longer a simple sinusoidal oscillation. Within each optical cycle, the combined electric field becomes an asymmetric waveform, with a strong half-cycle in one direction and a weaker, differently shaped excursion in the other. For an electron contemplating a quantum jump across the tunnel barrier, that asymmetry matters enormously. The instantaneous field tilts the barrier, lowers it on one side, and determines the precise instant, within the optical cycle, when tunnelling is most probable. By adjusting the relative phase between the two colours, the team could shift that favoured instant back and forth, and with it, flip the direction of the resulting current pulse.

What makes the result remarkable is the combination of timescale and spatial precision. The junction in a scanning tunnelling microscope is a few nanometres or less across, and the tunnelling process itself is confined to a region far smaller than that. The researchers demonstrated sub-ångström topographic sensitivity, meaning the laser-driven current responded to height changes on the surface at the level of individual atoms, while achieving a lateral spatial resolution of two nanometres. In other words, the attosecond current they generated and steered remained a genuinely microscopic probe, not a diffuse photoresponse smeared across the sample. And they did it under ambient conditions, at room temperature and in air, rather than in the cryogenic vacuum that most ultrafast microscopy experiments demand.

A persistent obstacle in laser-driven STM has been heat. When laser light strikes the tip or the sample, some of its energy is inevitably absorbed, raising the temperature of the junction and producing thermal expansion of the tip, thermally assisted currents and other artefacts that can masquerade as genuine ultrafast dynamics. Earlier femtosecond and picosecond experiments in laser-assisted STM were often haunted by exactly this problem, and distinguishing a true sub-cycle electronic response from a slow thermal signal became a central methodological challenge for the field. The Technion team addressed this directly. Because their current control depends on the sub-cycle waveform of the two-colour field, flipping the carrier-envelope phase flips the current direction, whereas a thermal effect would be completely indifferent to the sign of the electric field. The observed phase-dependent, direction-reversing current is therefore an unambiguous signature of coherent, lightwave-driven electron transport, and the experiment was explicitly designed to be free of thermal artefacts.

To understand the mechanism, the researchers turned to theory at two levels of description. On the one-electron side, they employed a strong-field model of tunnelling, in the spirit of the Keldysh framework that has underpinned strong-field physics since 1965, treating the electron as a wavepacket launched through a field-tilted barrier. On the many-body side, collaborators performed time-dependent density functional theory calculations of the full metal-insulator-metal junction, capturing the collective response of the electrons in the electrodes. Projecting the experimental data onto both theoretical descriptions revealed a consistent physical picture: a three-step transport process operating in the non-adiabatic tunnelling regime. In this regime, the optical field changes so quickly that the electron cannot follow it adiabatically; instead, the field launches the electron into the barrier, the wavepacket propagates across the gap while the field continues to evolve, and the electron is finally collected on the far side, with the whole sequence compressed into a fraction of an optical cycle.

From their theoretical analysis, the team extracted a concrete number for the duration of the current burst: 860 attoseconds. That figure, less than a single femtosecond, quantifies how briefly the electrons stream through the junction during each pulse. It also places the experiment firmly in the lightwave electronics regime, where the oscillating carrier wave of light, rather than a sequence of electronic switches, dictates the timing of current flow. Lightwave electronics has been advancing rapidly in recent years, with petahertz-scale field control demonstrated in nanotips, nanoantennas and solid-state devices, but the STM junction offers something none of those platforms can: the ability to position the driven current with atomic-scale precision on a surface and, in principle, to image ultrafast charge dynamics as they unfold at specific sites, such as a single molecule or a defect.

The lineage of this work runs through several landmark experiments. Attosecond science itself was born from the recognition that strong laser fields can ionize atoms within a fraction of an optical cycle, and that the emitted electrons carry timing information about that process. In 2011, researchers at the Max Planck Institute of Quantum Optics showed attosecond control of electrons emitted from a nanoscale metal tip, extending strong-field physics to nanostructures. Later experiments traced attosecond electron emission from tungsten nanotips and demonstrated attosecond field emission from nanotips driven by terahertz-scale waveforms. In parallel, ultrafast STM developed along its own track, first with pulsed lasers that suffered from thermal complications, then with single-cycle terahertz pulses that enabled femtosecond imaging of molecular motion, and more recently with all-optical sub-cycle microscopy on atomic length scales. The new work synthesizes these threads, bringing waveform-level attosecond control, the hallmark of gas-phase strong-field physics, into the spatially resolved world of the STM.

The two-colour approach itself has a distinguished history. Phase-controlled ionization with two colours was explored in atoms in the 1990s, and coherent control of above-threshold photoemission from tungsten nanotips followed in 2016. What the Technion team added was the junction geometry and the diagnostic power of the STM platform. In a tunnel junction, the current direction is a measurable, meaningful quantity: electrons can flow from tip to sample or from sample to tip, and the net laser-induced current reflects the asymmetry of the driving waveform. By scanning the two-colour phase and recording how the current reverses, the researchers effectively performed attosecond streaking inside the junction, using the optical waveform itself as the clock that timestamps the tunnelling event.

The practical implications reach toward a long-sought goal: triggering and imaging ultrafast charge dynamics at specific locations on a surface. Many of the most important processes in chemistry, materials science and nanotechnology, from charge transfer in molecular junctions to carrier dynamics in photovoltaic materials, begin with electron motion on attosecond to femtosecond timescales localized to nanometre-scale regions. A microscope that can launch a directional attosecond current pulse at a chosen point and then read out the response could watch those processes unfold in real space and real time simultaneously. The two-nanometre lateral resolution and sub-ångström vertical sensitivity reported here, achieved under ambient conditions, suggest that such spatiotemporal microscopy need not be confined to specialized cryogenic laboratories.

There remain challenges to overcome before attosecond STM becomes a routine tool. The current bursts are generated by intense optical fields, and extending the technique to more delicate samples, to insulating substrates and to single-molecule junctions will require careful control of the near-field enhancement at the tip. The theoretical framework, spanning one-electron strong-field models and full many-body simulations, will need to keep pace as experiments move to more complex materials. But the conceptual barrier has now been crossed. Electrons tunnelling through the gap of a scanning tunnelling microscope have been caught in the act, timed to within a few hundred attoseconds, and steered at will by the waveform of light. The fastest event in electronics and the sharpest eye in microscopy are, at last, looking at the same thing at the same time.

Subject of Research: Attosecond control of laser-driven electron tunnelling currents in a scanning tunnelling microscope junction

Article Title: Attosecond current control and timing in a scanning tunnelling microscope

Article References: Davidovich, D., Ma, B., Goldner, A., Cohen, S., Chen, Z., Borisov, A. G., & Krüger, M. (2026). Attosecond current control and timing in a scanning tunnelling microscope. Nature Photonics. https://doi.org/10.1038/s41566-026-02021-y

Image Credits: AI Generated

DOI: 10.1038/s41566-026-02021-y

Keywords: attosecond science, scanning tunnelling microscopy, lightwave electronics, two-colour laser pulses, quantum tunnelling, ultrafast microscopy, strong-field physics, TDDFT, nanoscale junctions, carrier-envelope phase, electron transport, Nature Photonics