Two Attosecond Flashes, No Infrared in Between: A Cleaner Look at an Electron Hole

A femtosecond is a millionth of a billionth of a second, and for chemistry it is the natural unit: atoms in a molecule take tens or hundreds of femtoseconds to rearrange. Electrons are quicker. They redistribute in attoseconds, a thousand times faster still, and the only way to catch them is with a flash of light shorter than the motion you want to see.

Attosecond science has had such flashes for two decades, and the 2023 Nobel Prize in physics went to the people who made them. What it has not had is a clean way to use two of them in a row. The standard experiment, attosecond transient absorption spectroscopy, pairs one extreme-ultraviolet attosecond pulse with an infrared laser pulse hundreds of times longer, and the time resolution comes out of the infrared field's own sub-cycle dynamics rather than from the pulse duration.

That works, and it has produced a great deal of physics. It also means the strong field you are using to read the system is part of the measurement. Writing in Nature Communications on July 28, Mikhail Volkov, Bernd Schütte and five colleagues at the Max Born Institute for Nonlinear Optics and Short Pulse Spectroscopy in Berlin report the first version of the experiment with the infrared removed: both the pump and the probe are extreme-ultraviolet attosecond pulses. They call it all-attosecond transient absorption spectroscopy, and the qualifier "all" is the whole claim.

"By using attosecond pulses in both steps, we can excite and observe electronic motion without applying an additional strong infrared field," Schütte said in the institute's announcement of the work on July 29.

What they watched

The system is xenon, and the choice is deliberate. The first attosecond pulse strips an electron out of a xenon atom, leaving a vacancy, or hole, in the outer 5p shell. It does not leave the ion in one definite state; it leaves it in a superposition of two, and a superposition of two states with slightly different energies beats, like two piano strings tuned a fraction apart. The hole sloshes back and forth. The second pulse arrives after an adjustable delay and reads the ion's absorption, and as the delay is scanned the absorption rises and falls in step with the sloshing.

The measured period is three femtoseconds. The abstract puts it plainly: the experiment revealed "previously unresolved electronic coherences in Xe, revealing oscillatory valence hole motion with a 3-femtosecond period." The Max Born Institute adds two numbers the abstract does not carry, and they should be read as the institute's own account of its instrument rather than as figures checked against the paper: pulses of about 270 attoseconds, and an oscillating feature in the absorption spectrum at about 16 electronvolts.

The number was already known

Here is where the result becomes a methods story rather than a discovery. Three femtoseconds is not a surprise. Planck's constant, written in units convenient for this kind of arithmetic, is about 4.14 femtosecond-electronvolts; divide it by a three-femtosecond period and you get an energy splitting of roughly 1.4 electronvolts. The splitting between the two spin-orbit states of the xenon ion's 5p hole has been in the spectroscopy tables for generations, and it is about 1.3 electronvolts. The new method recovered a number that older, slower techniques had already pinned down.

For a measurement technique, that is the best possible outcome. Xenon is a calibration target, not a mystery, and agreement to within a few percent with an independently known quantity is how you show that a new instrument is telling the truth. Nothing here revises what physics thought about xenon. What is new is that the answer came out of two attosecond pulses and nothing else.

The coherences themselves were not previously resolved, which is a narrower statement than saying the old approach got them wrong. It did not. Infrared-assisted attosecond spectroscopy has produced solid results for two decades; it simply could not separate this particular oscillation, because the field that made the measurement fast was also acting on the states being measured.

Why a table top matters

The other half of the claim is where the experiment sits. Attosecond pulses bright enough to serve as both pump and probe have generally meant a large facility with a queue. This one runs on high-harmonic generation from a table-top laser, in which an intense infrared pulse is fired into a gas and the gas radiates a comb of much higher frequencies. The authors close their abstract on exactly that point, calling high-harmonic generation "an ideal source" for the technique thanks to its broad bandwidth, high stability and easy accessibility, and "offering the potential for replication in numerous laboratories."

Xenon was not the only gas. The paper also reports systematic measurements of electron dynamics in krypton, argon and neon, which matters for the generality argument. Four noble gases make a better basis for claiming a technique than one does.

What the abstract does not claim is universal reach. It points to ultrafast electron dynamics in atoms, molecules and solids as the potential application area, with possible impact across physics, chemistry and biology. That is three classes of material and a stated hope, and it stops well short of any light-driven process whatsoever. Charge migration along a molecule after it absorbs a photon is the obvious next target, and molecules are messier than a noble gas in every way that matters to a spectroscopist.

This account rests on the abstract, the journal's registry record and the Max Born Institute's release; the 270-attosecond pulse duration and the 16-electronvolt feature come from the release alone. And nothing in the experiment produced an image. No electron was filmed. What exists is a set of absorption spectra recorded at a series of delays, from which the oscillation of the electron hole was reconstructed.