Just outside Campinas – the third-largest city in the state of São Paulo – the diggers are excavating huge piles of dark, red soil in preparation for one of Brazil’s biggest scientific endeavours. The large hole is destined to become Sirius, a next-generation synchrotron source estimated to cost R$650m ($280m) that promises to be one of the world’s first “ultimate storage rings”. When complete in 2017, it will boast the lowest ever emittance – confining its photons to a beam narrower than any other in the world – to let scientists probe the structure and properties of materials with unprecedented detail.
Antonio José Roque da Silva, director of Brazil’s National Synchrotron Light Laboratory (LNLS), which houses the country’s existing second-generation light source UVX, says that the aim is to achieve an emittance of 0.28 nm rad. “The best synchrotrons in the world go down to 1 nm rad and only MAX IV [being built in Sweden] will be comparable in terms of emittance,” he says. But more than just providing a powerful new tool, Roque believes that Sirius confirms a shift in ambition among the Brazilian scientific community. “Sirius shows that Brazil can be a leader, not just a follower,” he adds. “We can share knowledge on the same level as other countries, rather than just learning from what others have done.”
Roque’s view is backed by Yves Petroff – a former head of the European Synchrotron Research Facility (ESRF) and long-time advisor to the LNLS – who says that Sirius is part of a “small revolution in synchrotron technology”. It will be one of the first light sources to exploit groups of magnets known as multi-bend achromats to focus a coherent, high-brightness beam more tightly than ever. Researchers will for the first time be able to resolve complex biological structures, even when the sample size is small, and to track spatial and temporal changes in materials with nanoscale resolution.
Time is money
With work starting by mid-2014, Roque admits that the construction schedule will be tight. If all goes to plan, commissioning will get under way in mid-2016, with the first beamlines open to users a year later. But that plan assumes that the LNLS can raise all the funds it needs for the project. Although the R$650m price tag has been approved in principle, the Ministry for Science, Technology and Innovation has committed less than half of the overall budget. Negotiations are still under way with other contributors, including the National Bank for Development, the São Paulo Research Foundation (FAPESP) and major industrial users such as national oil giant Petrobras.
But even the funding for this year is not yet secure. “What is certain is that the science ministry has committed R$137m for 2014, while R$230m is needed to maintain the target schedule,” says Carlos Alberto Aragão de Carvalho, director general of the National Centre for Research in Energy and Materials, which comprises the LNLS plus three other national laboratories for nanotechnology, biosciences and ethanol production. Warning that “there are no guaranteed commitments”, Aragão is nevertheless confident that the money will be found. “Money flows more quickly from the ministry because the agreement is already in place, but for other contributors it takes time to complete their evaluation processes and establish the necessary infrastructure,” he says.
Petroff, however, strikes a note of caution, adding that secure financing is crucial for building a machine like Sirius on time and on budget. “You need to know your budget for a new project in advance, since it influences your design choices and makes it possible to place bulk orders,” he says. “If the budget needs to be negotiated and supplied each year, purchasing must be piecemeal and will increase costs.”
But even if the money does not come through right away, all three researchers agree that Sirius will still happen, although it might take longer to complete. Indeed, when the existing UVX synchrotron was built at Campinas in the 1980s, it was handcrafted by scientists and engineers at the LNLS, partly to reduce costs and partly because there were no Brazilian firms that could supply parts and components. “Noone in Brazil had any experience of using or building synchrotrons,” recalls Petroff, who advised on that project too. “Everything was built in-house and so it took almost eight years to construct. The upside is that the staff at the lab have a deep understanding of the technology.”
Today, that knowledge is being used to help Brazilian companies develop a commercial capability in building the core components for Sirius. Roque and his LNLS colleagues have, for example, been working with WEG, a large electrical-motor business based in Santa Catarina in southern Brazil, to develop high-performance magnets, with production having started in January. Meanwhile, Brazilian metals manufacturer Termomecanica is supplying the alloy for the vacuum chamber and discussions are ongoing with several other potential partners too.
Creating a domestic supply chain is in fact a key objective for the project. “Sirius is a big investment in basic science, but that investment should also yield positive outcomes for the entrepreneurial community,” says Aragão. “We are inviting companies to take part in the challenge, to develop new technologies, to train their engineers in specialist technology and ultimately to boost their international competitiveness.”
Try something special
Designing and building a synchrotron with an emittance of just 0.28 nm rad is quite a challenge. In fact, Sirius was originally meant to have an emittance of 1.7 nm rad, similar to the PETRA machine in Germany, but in July 2012 an international advisory committee recommended a more ambitious target. “The committee questioned whether the previous target would stand the test of time,” says Liu Lin, head of accelerator physics at the LNLS who has worked at the lab since the initial feasibility studies for the current UVX. “We all thought it was worth taking the opportunity to try for something special.”
Once the decision was taken, it took Liu and her team about a month to come up with an optical design that reached the target emittance while working within other key constraints, such as the size of the ring. “We chose to use multi-bend achromats, which exploit multiple dipole magnets to bend the beam through successive angles. Most new projects such as MAX IV and the planned ESRF upgrade have been designed with this approach, while existing machines and projects that got under way some time ago are based on double-bend or triple-bend achromats,” she says.
Strong focusing is needed between the dipoles because multiple deflections of the electron beam create dispersion that must be controlled. Sirius will use five groups of bending and focusing magnets, and Liu says it should deliver a slightly lower emittance than the seven-bend lattice at MAX IV because it has tighter focusing between the dipoles. Although Sirius and MAX IV incorporate 20 multi-bend achromats and have rings with a similar circumference, Sirius has fewer dipoles and longer straight sections to make it easier to insert devices, such as undulators and wigglers, to modify and enhance the beam.
The need for such strong focusing has led to some creative engineering solutions. For example, vibrations under the focusing quadrupole magnets must be kept within a few nanometres because any external movement is effectively magnified in the beam by a factor of about 50. And as such narrow row apertures are also difficult to pump, it in turn forced an entire rethink of Sirius’s vacuum system. In particular, stainless steel had to be replaced by copper coated with non-evaporable getters (NEGs) – a technology originally developed for the Large Hadron Collider at CERN. Staff visited CERN for guidance and training, and the LNLS is now one of the few places in the world where the technology has been installed. Moreover, Rafael Seraphim – head of vacuum at the LNLS – says that CERN is now interested in working with his team to study the behaviour of vacuum systems when exposed to synchrotron radiation.
Putting Brazil on the map
There is no doubt that Sirius will be breaking new technology ground, but future users will be more concerned with its scientific potential. Harry Westfahl Jr, scientific director at the LNLS, says that 13 experimental stations will be up and running when Sirius opens, with a capacity to expand to 40 beamlines over time. “We want users to have access to the newest technology,” he says. “Our initial focus will be to develop five undulator-based soft and hard X-ray beamlines, since these will offer improvement to the beam of several orders of magnitude.”
One potential challenge, says Westfahl, will be producing X-ray mirrors to the exacting standards needed to conserve low emittance through the beamline. “We have opted to exploit mirrors rather than specialist lenses for focusing, partly because mirrors are more flexible and partly because we have industry expertise in Brazil,” he says. “We are now working with industrial suppliers to produce mirrors with very low slope errors, but they could become a bottleneck because of the time needed to fabricate them.”
Westfahl’s team has also been busy refurbishing the beamlines of the existing UVX synchrotron, with the aim of transferring some of the new experiments to Sirius. One recent success has been the installation of an infrared beamline with a diffractionbeating resolution of 20 nm, achieved by making the light interact with the tip of an atomic force microscope. “Infrared light is ideal for probing delicate biological samples, but its resolution is normally limited to the micrometre range,” says Westfahl. “This technique has already been shown to achieve an imaging resolution of 100 nm and we are hoping to push it one step further.”
In fact, Westfahl’s ambition reflects the commitment of the whole LNLS team to make something special happen here in Campinas. “Sirius will put Brazil on the scientific map,” says Aragão. “Every detail requires creativity and opens up new areas of research. We want to enlarge our scientific community around our new star.”