Accelerator challenges move the frontiers

Tom Shelley reports on the large-scale precision requirements being set by plans to build the world’s next big particle accelerator

The International Linear Collider (ILC) – the next generation of particle accelerator – will set new benchmarks in machine design. It will need to focus 10MW into micron-sized beams, accelerate and steer them over kilometres, then re-focus them to nanometre sizes and make them collide. Many of the technologies being developed to make it work are immediately applicable to solving problems in leading-edge mechanical engineering. In particular, the required engineering to microns over kilometres has led to the development of a novel automatic surveying system. This is expected to find use in the monitoring and adjustment of large structures made to high degrees of precision – such as radio telescopes and large scale machine tools – as well as improving the maintenance and safe functioning of future high speed rail lines. According to Professor Phil Burrows of the John Adams Institute, a leading research centre into accelerators: “We plan to use the ILC like a precision microscope to home in on the features of the new physics landscape. It is vital for the UK to secure a piece of the scientific action. If the UK provides funding we'll be able to make major contributions to the project – wherever it is located”. Conventional methods of alignment and adjustment – even those based on laser beams – are nowhere near good enough to satisfy the alignment needs of the ILC, which are measured in tenths of millimetres over kilometre-long baselines. Laser beams in air are not sufficiently straight because of variations in refractive index resulting from changes in temperature. Researchers from Oxford and Deutsches Elektronen-Synchrotron (Desy) in Hamburg, Germany have formed the Linear Collider Alignment and Survey Group (Licas). They have developed a system called a Rapid Tunnel Reference Surveyor (RTRS) – otherwise known as the Licas train. According to Dr Armin Reichold, who leads the Licas team at Oxford: "If you have a very large reference system, that needs to be maintained to a high degree of precision over a long period, a conventional system to survey it becomes pretty difficult to handle. Take a high speed rail network for example or, as in our case the approximately 100km of beam line in the ILC. The ground moves and there are frequent occasions when you have to resurvey it - the ILC will need this once or twice a year. For this an automatic surveying system is quite wanted as it is much faster and thus reduces downtime of a very precious facility. “The next generation of very large telescopes will reach dimensions in the order of 100m,” he says. “These need to be constantly monitored and adjusted because of the wind and other effects.” He says that the Licas train supplies a way of transporting a geometrically straight reference system over long distances – but the breakthrough developments are inside it. It has two types of sensing system. One is a distance-measuring system using Frequency Scanning Interferometry (FSI). This is unlike conventional interferometry because it measures absolute distance rather then distance changes. The other is the Laser Straightness Monitors (LSMs), which use laser beams in a vacuum tube to represent a straightness reference. This has advantages over hydrostatic levelling systems, which measure height with respect to an equipotential surface (Geoid) as realised by a fluids surface, or highly stretched wires, which sag. The exact shape of the geoid is very hard to determine and changes for example with tides. The cost of such fixed installations also scales with the number of elements you wish to survey whereas an RTRS only needs some extra time to complete a longer survey. “We also needed something that represented a definition of ‘geometrically straight’ without incurring horrendous expense,” says Dr Reichold. “To do this, we chose what would normally be considered a bad laser, which has an extremely short coherence length. Because of the short coherence length, it does not produce speckle patterns, which would make it difficult for the CCDs to locate the beam centre accurately.” In practice, the scheme involves having corner cube retro reflectors opposite the machine every 4.5m in the tunnel wall, and six measurement cars in a 25m long train suspended from a track running along that wall or in the floor. Each car will be stopped closely in front of one of the reflectors and measures its coordinates by making six simultaneous distance measurements to it. The cars use their LSMs to measure their positions and orientations relative to a central laser beam that runs up and back down the vacuum tube connecting all cars together. The same vacuum tube houses five further sets of six FSIs, which measure the distance between cars. The trio of instruments in each car is complemented by a high precision tilt sensor that levels the car in front of each wall marker. Once the system has worked out precisely where the wall markers are, a conventional computer controlled survey instrument can measure the positions of accelerator components relative to these wall markers. Open air instruments can be used, as measurements are only made across the width of the tunnel – which is short enough to overcome refraction effects. This conventional instrument is ideally located directly on one of the cars of the RTRS so it can ride along with it. With these measurements done is possible to make the necessary positional adjustments of the accelerator components. In addition to the six measurement cars that hold the actual sensing units, the readout electronics, propulsion system, power supplies, interlocks and control computers for each measurement unit are supported in six separate service cars. This helps to isolate the measurements from vibrations and thermal effects produced by the electronics and drive mechanics. The FSIs can measure absolute distances to sub micron accuracies over lengths of up to 10m though a longer range is readily achievable. The DAQ control software for the laboratory development phase was originally written in LabView and has subsequently been ported to C++. The prototype currently being deployed in a 70m tunnel at Desy has three measurement cars. The ILC is expected to need at least three, six-car RTRS units – one for each arm of the main linacs and one for the damping rings. The final parts of the beams will be steered into each other using electromagnetics and real time control, using a system that will have to have a latency of no more than about 100ns. The beams will be located in space using electromagnetic pickup devices. The shape of the beam profile will be measured using lasers. Despite the UK being keen to support the project, the ILC is likely to be built in the US – the Fermilab in Illinois is the leading candidate to host it. Pointers * Beams of particles will have to be focussed and steered into each other to an accuracy of nanometres – after each has traversed a distance of some 20km * Precise alignment of components will be achieved with an automatic measurement train, which establishes the positions of its components relative to corner cube reflectors, to each other, and the exact positions of the reflectors relative to each other using novel optical techniques