The X-ray laser SwissFEL on the PSI campus is a 740-metre-long facility. In it, electrons are accelerated to nearly the speed of light. Magnets then bring the elementary particles onto a slalom-like course so they emit X-ray light that is then amplified into laser flashes. "Such accelerators are among the most complex systems there are in the world," says Florian Löhl, who was responsible for construction of the main accelerator and now coordinates activities at the machine: "We have here hundreds of different systems – accelerator modules, positioning monitors, high-precision clocks, and much more – all built to the limit of what is possible."
The task that Löhl and his team had to solve, together with many of the participating groups at PSI, was correspondingly difficult: It is necessary to separate two electron packets that are rushing through the facility at extremely high speed with a time interval of just 28 nanoseconds between them (one nanosecond corresponds to a millionth of a millisecond). One packet of electrons is supposed to fly straight ahead on its way to deliver X-ray flashes in the first beamline of SwissFEL, called Aramis. The second packet of electrons has to turn and thus arrives at the second beamline, which is currently being installed under the name Athos. In this way Aramis and Athos can be operated in parallel. "With a separation element developed at PSI, we kick the two electron packets apart," Löhl explains. It is crucial for this kick to always remain stable. "If there are deviations, the electrons start to wobble and the process that amplifies the light signal no longer takes place."
New concept with "chicanes"
In Aramis, the electron packets are accelerated even more before they fly through a special arrangement of magnets called undulators. This is where the X-ray light pulses are generated. In Athos the electrons make their way directly to the undulator modules without any major further acceleration, but the way these are put together is completely different from the Aramis undulators. Between every two undulators there is a "chicane", a type of serpentine curve familiar to motor sport fans; this too consists of special magnets. This allows the electron packets to be compressed or shifted. "Our Athos undulator line is unique in the world," says Löhl: "The elaborate design gives us much better control of the properties of the light that is generated and allows us to provide a whole range of different operating modes." In this way, depending on what is required, it is possible to generate extremely short light pulses, or flashes with a narrow energy distribution, or even pulse trains in which many flashes closely follow one another.
Shortly before Christmas 2019, a first module of two undulators and a chicane delivered laser light right away. "Normally something like this takes much longer, but we were able to carry over many of the things we learned during the commissioning of Aramis," explains Löhl: "And every time more undulator modules were added, we were able to increase the light output."
Fully automatic optimisation
The Athos beamline, with a total of 16 undulators and 15 chicanes, is still under construction until the end of 2020. But important successes have already been achieved, and a tricky problem has been solved: Both beamlines should actually deliver a hundred X-ray light pulses per second, but so far there have been significantly fewer. "The amounts of data generated in Aramis were so large that they could not be processed quickly enough," says Löhl: "And at Athos we had to struggle with beam losses." The physicist therefore designed a computer program that automatically tests the effect of specific changes in the beamlines. Hundreds of parameters can be optimised fully automatically. In early September 2020, with the help of the computer program, the team succeeded for the first time in optimising the beamlines in parallel so that both delivered the maximum value of one hundred light pulses per second and, at the same time, new record values were achieved in the energy of the light pulses in both beamlines. "That was an important milestone for us," says Löhl.
Amazingly, the pulse energy in the new beamline was significantly higher than in the existing one, even though not all undulator modules have yet been installed in the new line. "It's because the new design works so well that we are already able to achieve such intense pulses with such high energy," explains Löhl. But he admits: "It was extremely difficult to optimise the accelerator so that not just one, but both beamlines deliver good values at the same time."
First experiments planned for 2021
A first end station in which experiments will be conducted with the Athos light pulses is also currently being set up. The first X-ray pulses were registered there at the end of June 2020. "The installations in the beamline and the experiment station go hand in hand," says Löhl. The first experiments are to be carried out in 2021 – one of the next milestones.
To explain how the experiments at SwissFEL work, Löhl describes a trick from photography: If you want to make a large, crowded place appear deserted, you choose a long exposure time. This blurs the people moving in the picture, and they become invisible. “With SwissFEL we do the opposite," the physicist explains. The X-ray flashes are only 1 to 60 femtoseconds (millionths of a billionth of a second) long. With this ultrashort exposure time, extremely fast processes can be recorded. The two beamlines complement each other. Aramis generates so-called hard X-rays in the short wavelength range of up to 0.1 nanometres, which is roughly the size of an atom. "This means we can use this beamline to look at individual atoms and find out, for example, how a certain biomolecule is structured," says Löhl.
Following chemical reactions
Athos, on the other hand, generates light pulses with significantly longer wavelengths of 0.65 to 5 nanometres. The experts refer to these as soft X-rays. "With this line you can look at what is happening in the electron shell," explains Löhl. This makes it possible to follow chemical reactions in real time. The researchers want to find out, for example, what happens in catalysts that purify gases or synthesise fuels. For information technology, researchers hope to gain insights into new materials that could be suitable for ultrafast switches or even more high-capacity magnetic memories. Also, molecular complexes that control cell functions and cause hereditary diseases are to be investigated with Athos.
Athos is one of the protagonists in the Alexandre Dumas novel The Three Musketeers, and another is called Aramis. But where is the third one, Porthos? "We have now started a project in which we are doing the initial brainstorming on the installation of a third beamline," says Löhl. "We are considering, for example, whether we could incorporate chicanes such as those as in Athos everywhere, in order to achieve a lot more flexibility." The parallel operation of three beamlines might prove to be an even greater technical challenge. Löhl says, "I don't think there is any one person at PSI who completely understands the accelerator from top to bottom in detail. But we have extremely competent groups that cover wide areas. And thanks to excellent cooperation, we are capable of managing such a complex system."
Text: Barbara Vonarburg