In stop-action photography, a moving body is frozen on film by the speed of the shutter or the flash of a strobe light. In this photograph (Fig. 1) of the great American pastime baseball, the shutter speed was sufficient to freeze the motion of the crowd, and the batter's anxious expression, as the pitcher throws the ball towards him. But there is one detail that the camera failed to capture — the motion of the ball. The ball's speed was so high (around 150 kilometres per hour) that it is blurred in the photograph, masking such details as the stitching on its surface. So there are different timescales for the action on the ball field, some too fast to register easily. The same limitation applies at the atomic scale, but the paper by Drescher et al.1, on page 803 of this issue, is set to change this. The authors have introduced a new type of 'photography' that freezes an atom in motion and heralds a new field of research — attophysics.
Figure 1: Spot the ball — the limitations of action photography.
![Figure 1 : Spot the ball |[mdash]| the limitations of action photography. Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, or to obtain a text description, please contact npg@nature.com](/nature/journal/v419/n6909/images/419789a-f1.0.jpg)
High resolution image and legend (98K)
Consider a water molecule moving through a solution on a picosecond timescale (1ps = 10-12 s). On a faster timescale (down to 10-14 s), the atomic nuclei of hydrogen and oxygen inside the water molecule also vibrate and bend relative to each other. Scientists have studied this motion since the late nineteenth century by unravelling the mysteries contained in the spectrum of radiation that can be emitted by such a molecule. But the advent of ultrafast laser pulses has revolutionized time-resolved spectroscopy over the past quarter-century.
Using the technique of 'pump–probe' laser spectroscopy, the experimentalist can initiate (pump) and watch (probe) the motion of molecules in real time with a resolution of a few femtoseconds (1 fs = 10-15 s). This has not only resulted in a direct measure of how molecules move during a chemical reaction but, perhaps more importantly, it has opened up a new school of thought that views the quantum world in terms of moving 'wave packets' representing the evolution of quantum-mechanical amplitudes and phases. In 1999, Ahmed Zewail received the Nobel prize in chemistry for his seminal contributions to this field of femtochemistry.
Although femtosecond light pulses are still an important scientific tool, they cannot freeze every aspect of atomic motion that is relevant in the structure of matter, just as the photograph fails to capture the details of the baseball. The frontier in ultrafast spectroscopy is the study of the motion of electrons bound inside the atom, in orbits close to the nucleus. This new scale is defined by the time it takes the electron in the innermost orbit of the hydrogen atom to complete one turn around the proton nucleus. The period of this orbit — 24 
10-18 s, or 24 attoseconds — is at least 100 times shorter than the duration of the shortest laser pulse.
But the period of the laser-pulse cycle at or near visible wavelengths imposes a fundamental limit on the resolution that can be achieved, at approximately a few femtoseconds. Consequently, researchers trying to make light pulses of attosecond duration have looked to other wavelengths, particularly those around the border between the X-ray and ultraviolet (XUV) regimes. One successful approach, using the strong-field phenomenon known as high-harmonic generation, produced the first measurable light pulses in the attosecond regime2, 3.
But forming attosecond light pulses is only one of the scientific challenges to be faced in making attophysics a reality. Equally challenging is the detection and propagation of these fragile pulses. Drescher et al.1, however, have solved all of these problems for the specific case of measuring the decay time of an inner-shell electron excitation, which until now had been studied only indirectly through the emission spectrum.
The authors used an attosecond XUV pulse to initiate the excitation process in an atom of krypton. This pulse frequency is high enough to ionize the atom, knocking out an inner-shell electron and leaving behind an unstable 'hole'. The decay path from this excited state involves two electrons; one drops from a higher shell to fill the hole, and another — the 'Auger electron' — is shaken out of the atom. The Auger decay process occurs between 10-14 and 10-16 s after ionization, and by sending in a second, longer (femtosecond) optical pulse, Drescher et al. could pick up the emission of the Auger electron. The time of release and the final energy of the Auger electron in the intense optical field are related. So by measuring the Auger electron's energy as a function of the time delay between the XUV and optical pulses, Drescher et al. were able to measure the Auger decay time with attosecond resolution.
The emergence of new areas with broad scientific impact is usually the result of creative contributions from a large community of researchers over a long period of time. But there are a few papers that announce the beginning of a new era, and the paper by Drescher et al.1 falls into this category. We are entering a new realm of hyperfast measurement; the age of attophysics has begun.
