Fluorescent microscopy, time-resolved spectroscopy, optical frequency metrology, laser micromachining, laser surgery and femtochemistry are just a few of the technologies that have evolved fostered by our ability to generate, control and measure ultrashort laser pulses. Moreover, ultrafast lasers and the nonlinear optical processes they generate have successfully made their way into industrial settings with the development of compact, turnkey and cost-effective systems. From the point of view of materials science, ultrafast lasers are a great tool to excite, control and probe in real time ultrafast dynamics such as the motion of atoms within molecules, providing unprecedented insights into the molecular and atomic structures of materials.

Credit: Tatyana Khanko / Alamy Stock Photo

But what is an ultrashort laser pulse? The term ultrashort is intrinsically subjective, but, in the late 1990s, it encompassed pulse durations in the range between hundreds of picoseconds and a few femtoseconds. Solid-state mode-locked lasers typically emit laser pulses with pulse widths of tens to hundreds of femtoseconds. By means of spectral broadening, ultrabroadband amplification processes and pulse compression, pulses consisting of a few optical cycles can reach pulse durations as low as a few femtoseconds1. However, laser technologies that emit pulses in the infrared and visible are inherently unable to break the femtosecond barrier, because the light oscillation periods at those emission wavelengths are longer than a femtosecond. For instance, the single-cycle period of an 800-nm pulse — the central emission wavelength of the widely used titanium–sapphire laser — has a light oscillation period of 2.7 fs. Indeed, by the late 1990s, achieving sub-10 fs pulse durations was becoming routine2. Yet, breaking the femtosecond barrier and making the leap from the femtosecond to the attosecond realm required even more exceptional science and engineering.

The first bit of ‘magic’ came with the serendipitous observation of high-order harmonics that were generated when intense ultrashort, but multi-cycle, infrared pulses were focused into a noble gas. Owing to their high frequency (in the extreme ultraviolet) and extremely broad bandwidth, such high harmonics could, potentially, enable the realization of attosecond pulses3. However, there was a caveat. Analogous to mode-locked laser pulses, these harmonics would need to be in phase for them to effectively generate sub-femtosecond pulses. It was at the turn of the millennium, when it was finally revealed that indeed these harmonics were in phase and the first attosecond pulse trains were ultimately measured4.

In a separate study published at the same time, the generation of the first isolated attosecond pulses was reported, employing strong infrared pulses that worked in this case in the few-cycle regime; this discovery led to the development of the attosecond streak camera5 and marked the beginning of the attosecond era.

It has now been two decades since the field of attosecond physics was launched, and although it is still in its infancy, the field has progressed substantially, namely in the generation of more intense and shorter attosecond pulses, and the generation of high harmonics has moved beyond operating solely with noble gases to more complex condensed-matter systems. Furthermore, attosecond X-ray pulses with gigawatt peak powers have been generated in large-scale free-electron laser facilities.

Attosecond lasers are now being used in distinct fields of science to, for instance, deepen understanding of the temporal electron dynamics of photoionization, and proposals for advanced attosecond signal processing and even for controlling chemical reactions are being put forward. It is still early days and to boost their impact in diverse research fields attosecond technologies will need to transition towards more compact, cost-effective and user-friendly systems. Nonetheless, the attosecond era is surely coming. After all, it provides glimpses to uncharted physical phenomena, a basic mandate of scientific discovery. A question that now remains is, when will talk of the zeptosecond era start?