The XFEL particle accelerator will be like no other: its unique action, zigzagging electrons at high speed along a tunnel, will enable researchers to film atoms in motion. Jon Excell and Christopher Sell report.
It’s just as well for those living in the leafy environs of Hamburg’s DESY Research Centre that sub-atomic particles colliding at close to the speed of light are pretty harmless. Because if they weren’t, the residents of this peaceful suburb would be living in the physics equivalent of a war zone. DESY, along with CERN in Switzerland and the UK’s Rutherford Appleton Laboratory, is a leading centre for particle accelerator research. The organisation operates four accelerators of varying sizes, including the HERA super-electron microscope which accelerates photons and electrons around a 6.3km underground ring. This collection of powerful scientific instruments looks set to be joined by a new neighbour: XFEL, an immensely powerful and highly unusual particle accelerator that will send electrons hurtling down a 3.4km underground tunnel to generate pulses of brilliant X-ray light. Expected to provide bursts of light that are orders of magnitude brighter than any existing light source, XFEL will enable researchers to film at atomic detail the precise sequence of events in chemical reactions, material deformation and biological processes. If all goes well, construction of the €908m facility should begin later this year, with completion expected in 2012. But the project is at a decisive stage. Last year science ministers from the UK, Denmark, France, Greece, Hungary, Italy, Poland, Russia, Spain, Sweden and Switzerland signed a memorandum of understanding to participate in the project and Prof Massimo Altarelli, who heads the European XFEL team, is confident that full European support will be forthcoming. He told The Engineer that when this is confirmed, construction can begin immediately. When up and running, the system will use a new type of superconducting linear accelerator to bring electrons to high energies of up to 20GeV (giga-electron volts). This will be achieved using a series of resonators made from the rare metal niobium. During operation these resonators are cooled to -271 Deg C, at which point they will conduct electricity without any electrical resistance, making them superconductors. This enables the system to operate for longer periods and generate a higher-quality electron beam than other accelerators that use conducting copper resonators. Once the electrons reach this point of high excitement, they will race at nearly the speed of light through arrays of permanent magnets known as undulators. These are arranged in such a way that they will force the electrons to zigzag down the tunnel. Because of the high speed of the electrons, every time these slaloming particles change direction they will produce radiation, and because of their high speed this will be X-ray radiation with the characteristics of a laser. The light pulses will then be distributed to a series of 10 experimental cells at the far end of the tunnel — where researchers hope to shed light on some of science’s mysteries. But before this can happen the project’s leaders are faced with a number of fundamental engineering challenges. The first is digging the tunnel, a 5m-diameter shaft that will start 27m beneath the DESY site and terminate 3.4km away under the town of Schleswig-Holstein. ‘As a tunnelling project it is pretty demanding. The tolerances and parameters enforced by an accelerator are extreme,’ said Altarelli. One of the biggest problems is that the spirit levels used to keep the tunnel aligned may create a perfectly horizontal tunnel, but not a straight one — and there’s a crucial difference. Because these indispensable tools obey the laws of gravity they will, he said, help to produce a 3.4km tunnel along a curve that is effectively an arc of a circumference. The problem with this is that the highly charged electrons travelling through the device are unaffected by the Earth’s gravitational field and will therefore travel in a straight line. But rather than employ techniques to bend the trajectory of these electrons, the solution has been to develop adjustable support systems within the tunnel. ‘The supports of the accelerating modules and the vacuum chamber where the electrons are running are engineered in such a way that we can make some adjustments to ensure that the thing is kept perfectly straight,’ said Altarelli. This situation is further complicated by the fact that while the electrons may be unaffected by the Earth’s gravitational field, the liquid helium used in the cooling system for the superconducting accelerating cavities is, so a sophisticated pumping system is required to keep the temperature precisely controlled. Another challenge peculiar to this project is the design and development of the undulators used to induce the slaloming effect. While undulators have been developed before, those in XFEL will be unique. ‘People have never made undulators to this length or accuracy,’ said Altarelli. ‘All the other synchrotrons use undulators, but they are typically 5m long. These are 200m.’ Plus, to produce coherent radiation these have more stringent specifications. For instance, the distance between the magnets has to be extremely well adjusted throughout the length of the undulator, otherwise the fields will be affected and the instrument will not work properly. Each 200m undulator is made from 5m segments that have to be installed one after the other and aligned from one segment to the next within +/-100 micrometres. Within each segment the accuracy and distance of 15mm between upper and lower set of magnets must be precise to within 1 micron. Because of the huge forces involved, these precisely arranged magnets also require tough mechanical engineering solutions. The system effectively consists of two immensely strong magnets with opposing poles just 15mm apart. Altarelli explained that the resulting forces between these magnets — in the order of tonnes per cm2 — mean that without some exceptionally hardy support mechanisms the whole system will collapse in on itself. The magnets are therefore held in place by 50cm x 10cm stainless steel girders which are attached to a support structure also made of steel. This material was chosen to avoid bending effects resulting from the temperature in the tunnel and also for its very low magnetic permeability. Even though it is holding magnets in place it will not develop a magnetic field itself. But as well as being sturdy, the system must also be adjustable to within 1 micron. So the whole thing is motorised, and a system of very high-precision optical encoders continuously monitors the relative positions of the magnets. Should adjustment be required, a number of motors that are electronically synchronised with a central control system will adjust the screws that connect the girders to the main support structure. Beyond the core features of the accelerator there are thousands more elements to the system. And all of them, from the mirrors used to deflect radiation, to the detectors that will capture radiation scattered by the samples used in the experiments, have been carefully designed by the DESY team. The further development and manufacture of many of these systems will be carried out by subcontractors. But although a number of organisations are in the frame, contracts have not yet been awarded and Altarelli declined to reveal which companies might be involved. He is, however, extremely vocal on the nature of the experiments, and the scientific significance of the resource once it is completed. ‘The X-rays that are being produced compared to what you would get in a lab are distinctive in three ways: they are more intense, they come in very short 100 femtosecond pulses (10-15 seconds), and they are spacially coherent like a laser. This means that when you send them to a target, the X-ray is diffused by every single atom in the structure, they are in phase so they will be able to interfere and you produce a sort of hologram.’ He said that it will be possible to record these holograms which encode the positions of the atoms in materials and then to observe how these positions evolve on a timescale of 10-13 seconds. Expanding on the applications of the instrument, Altarelli said that researchers will even be able to take pictures of atoms moving during their vibrations — which could prove to be particularly useful for materials development. ‘During a phase transformation of a material, as well as making a crystallographic examination of the material before and afterwards, you are now able to make a movie of how you get from one thing to another,’ he said. It could also be used to follow the rearrangement of atoms in a chemical reaction, he claimed: ‘It will be a tremendous advancement to our understanding of things that we don’t really understand — there are catalytic processes used in industry that people know but don’t fully understand why they work. With this you’ll be able to see what happens.’ Clearly, there are a number of other light sources around the world either in use or under development. So where does XFEL fit in? The facility’s closest rival is on the other side of the Atlantic, where engineers are working on the US Department of Energy-funded Linac Coherent Light Source (LCLS), a machine that will produce X-rays in a manner similar to XFEL, emitting pulses 1,000 times shorter than existing instruments. Due to be switched on in 2009 , the US project is a few years ahead of the European effort. But Reinhard Brinkmann, XFEL project leader at DESY, suggested that the technology in the US linear accelerator could hinder its future development. ‘They’re using existing conventional copper accelerator technology which has been around for 30 years. This makes it cheaper and quicker to develop, but they are limited with the properties of this conventional accelerator to have short radio frequency pulses. They cannot provide as many electron bunches as we can, so we can provide much larger average intensity of laser light.’ He added that use of a copper accelerator will also mean the facility cannot upgrade to continuous mode as the XFEL has the option to do, thus limiting the scope of possible research. Conventional linear accelerators using copper resonators can only accelerate electrons for around 0.01 per cent of the time they are operating. Meanwhile, in the UK, a group at the CCLRC’s Rutherford Appleton Laboratory is laying much of the groundwork for a so-called fourth- generation light source that will also use FEL technology to generate bright flashes of light. However, Wendy Flavell, chair of strategy for 4GLS, explained that such a system would be underpinned by slightly different technology and used for slightly different research than XFEL. ‘4GLS is operating on a lower energy regime, so it will produce light from THz range — from far-infrared right through infrared to UVV and soft X-rays. It produces everything XFEL doesn’t,’ she said. In the more immediate future Rutherford Appleton is also getting ready to switch on the diamond synchrotron third-generation light source next year. But again the applications are different — and in terms of power and brightness, there is, said Altarelli, no competition between the two. ‘XFEL is a completely new ballpark, orders and orders of magnitude above. It’s like saying how does a jet compare with a ship when you go from Europe to the US. The intensity of difference is nine orders of magnitude — this is revolutionary.’