A technique for producing radiation pulses that endure for less than one attosecond (10–18 s) has been proposed by researchers in Spain and the US. If the technique can be realized in the lab, then it could produce X-ray flashes brief enough to capture the movement of an atom's inner electrons or perhaps even look directly at the movement of protons and neutrons during nuclear fission or fusion.

Ultrashort radiation pulses are the mainstay of pump–probe spectroscopy, which is used to study fast processes such as the motion of electrons taking part in chemical reactions. In this technique, a short and highly energetic laser pulse (the pump) provides the activation energy needed to kick-start a chemical reaction. A moment later, a second pulse (the probe) hits the reacting particles and is then detected. Through careful study of the detected pulse, the instantaneous state of the particles at the time of collision can be deduced. By varying the time delay between the pump and probe pulses, scientists can reconstruct the behaviour of electrons in atoms and molecules during a chemical reaction. In the future, the technique could be used to optimize the conditions for desired reactions or even to manipulate the reactions directly.

The shorter the pulses used, the faster the chemical processes that can be studied and manipulated. Pulses of a few hundred attoseconds can be produced by irradiating an atomic gas cloud with a powerful infrared laser. As the laser radiation is coherent, the electric field experienced by all the atoms in the cloud oscillates near-simultaneously. When the electric field of the radiation points in one direction, some of the negatively charged electrons are pulled a few nanometres away from the nuclei of their atoms. When the phase of the radiation changes and the electric field reverses, they are then pushed back towards the parent atoms. Some of these electrons collide with the nuclei and give up their kinetic energy in a sudden burst of broadband, coherent X-ray radiation.

Using wasted electrons

Many electrons, however, shoot straight past the nuclei without colliding and emitting radiation. Now, a new way to use these wasted electrons has been proposed by Carlos Hernández-García and colleagues at the University of Salamanca and the Centre for Pulsed Lasers, both in Spain, and the University of Colorado at Boulder in the US. What is more, the team has shown theoretically that these electrons could be used to produce sub-attosecond pulses.

The idea is that electrons that shoot straight past a nucleus the first time can be pulled back for a second time in the subsequent oscillation of the field. This gives the electrons a second opportunity to collide. The catch, however, is that these electrons produce X-rays at a slightly different frequency from those that collide the first time round.

When two adjacent notes on a musical instrument sound simultaneously, the resulting harmony produces a distressing sensation because the volume oscillates rapidly as the frequencies move in and out of phase. Similarly, the researchers' model shows that the intensity of a pulse should fluctuate as the two X-ray frequencies move in and out of phase. The attosecond pulse could be split into a pulse train with each pulse lasting just a few hundred zeptoseconds (a zeptosecond is 10–21 s). "In the paper we show how to get 800 zs, but this is scalable," says team member Tenio Popmintchev of the University of Colorado, "we don't know yet how short we can go."

Finding an appropriate laser

The team believes that the biggest obstacle to realizing this scheme in practice is the development of a suitable infrared laser – paradoxically, shorter pulses require a longer-wavelength driving laser. The scientists have received two research grants and are currently working to develop such a device.

Jon Marangos, an expert on laser–matter interactions at Imperial College London, believes that it should be possible to develop a suitable infrared laser and that an experiment to test the principle could be conducted using currently available equipment. He says that the work is one of several interesting, recent proposals to produce sub-attosecond pulses, notably one in March at the University of Strathclyde in Glasgow to do so using a free-electron laser, but that they all need to address a key issue: a pulse train would be problematic in experiments because there would be no way to attribute a signal to a specific pulse.

"You really need to work out a way of generating isolated attosecond pulses," he says, "because then you can use them in a pump–probe experiment with no ambiguity about when the pump and the probe events occurred."

The research described in Phys. Rev. Lett. 111 033002.

There is more about why ultrashort radiation pulses are used to study electron motion in the Physics World video "Can we see the motion of electrons on the atomic scale?".