Isis keeps beaming

When physicists at ISIS, the world's leading pulsed neutron and muon source, at the

in Oxfordshire, began planning a second target station for neutron production in April 2003, they knew they faced a big challenge.

They also knew that building the new target station would give them an opportunity to improve construction of some of the major components of neutron production — the most fundamental being the hardware for the proton beam.

To make the improvements, ISIS called on

(MAC) to provide highly specialist metallised ceramic components for new monitoring equipment to be installed inside the proton beam apparatus. The ceramic vacuum tubes in the first target station were sealed with indium wire, but experience showed these could become unreliable if disturbed.

'Wires are a perfectly good vacuum seal, but the difficulty is that they are not very robust,' said Eamonn Quinn, senior project manager at Rutherford Appleton. 'If you move the [proton beam] instrument or if you put loading on its flanges, there is a risk that the indium seal will fail and you have to remake it or at least retighten that seal.'

The challenge for MAC engineers was to design a seal that would require less maintenance. Quinn explained that to reduce the risk of exposing workers to radiation 'you want to remove as much maintenance as you can from this type of machine'. Also, he added, if the machine leaks it will not function and that means downtime for ISIS.

The engineers were given the task of creating a design-and-manufacturing process that would produce a strong, high-integrity vacuum seal with a leak rate of 10-8 milibars litres/second.

Another problem they faced was that the 158mm-long proton beam apparatus contains a series of components — two nickel-plated mild steel flanges of 240mm diameter separated by a pre-formed diamond-ground alumina ceramic insulator — which expand at different rates when heated. To mitigate this effect, the engineers decided to use Nilo K, a nickel iron cobalt alloy that is a low coefficient expansion material similar to alumina ceramic. In their design, a Nilo K flange is brazed to the tube and then welded to a stainless steel intermediate plate, which is welded to the mild steel flange.

'We used Nilo K because you have to match the coefficient of the expansion of the alumina ceramic to the metal that you have to braze to it,' said Quinn.

'This braze operation happens at high temperatures, 800-8500C. If you use something like stainless steel, there is a large thermal mis-match strain when you cool that down to room temperature. Nilo K is much closer in thermal expansion coefficient to alumina, so the mismatch strain is much less and you have more joint integrity and less strain in the joints.'

Quinn said the newly-designed proton beam is not only more efficient but also retrofitable to the existing extracted proton beam system. MAC engineers tried to remain as consistent as possible with the old design, he said, 'so that over a programme period of maintenance we can actually build up a spare intensity monitor and have it as a direct replacement.'

Martin Davidson, a senior project engineer at Rutherford Appleton, said he and the other engineers are confident that their new design will outlast the old one.

'This system should outperform [the existing proton beam apparatus] and once it is brazed and leak-tight and assembled it is not going to change its state,' he said. 'It will be secure for as long as they need it.'