The ECR ionizer is operated with the 2.45 GHz microwave, typically at a power of 20-40 W. The low magnetic field (87.5 mT) which will reduce the proton polarization is not a problem in our case. We are mainly interested in producing deuteron beams.
The early design of the ionizer followed that of HIPIOS at IUCF. The sextupole magnet consists of SmCo permanent magnets encapsulated in copper cans and was electrically isolated and biased to +7.5 kV operation voltage, while the vacuum chamber remained at the ground potential. The aperture of the extraction electrode was 1 cm in diameter. Good pumping was provided radially through the gap of the sextupole.
The problem that the ions were extracted axially to the upstream direction as well as radially through the gap of the sextupole magnet was immediately noticed. As a result the ions sputtered the RF transition units and the wall of vacuum chamber, inhibiting a long-term operation. We installed a mesh and an electrode to suppress the undesired extraction and the problem seemed to be settled. However, when the beam was injected into the cyclotron, we noticed that the suppressor was causing some instability to the ECR plasma. The best performance of the early version of the ionizer was 20 uA with 60% polarization of the ideal value.
We modified the design of the ionizer close to the one at PSI. A Pyrex plasma chamber was installed inside the sextupole magnet which was kept at the ground potential. The plasma region was pumped through the 5 cm aperture of three-stage extraction electrodes. The stable and long-term operation of the new ionizer allowed us extensive searches of optimum conditions. Eventually the ion intensity of 140 uA was routinely observed. The polarization, however, was reduced to 50% of the ideal value.
Since we thought the low polarization was primarily due to insufficient pumping which caused an increase of the unpolarized background, we installed a Ti sublimation pump downstream of the extraction stage. Although the ECR plasma was further and significantly stabilized, the polarization degree was improved only up to 60% of the ideal value.
The sextupole magnet had been designed to be capable to match an ECR condition at higher frequency, expecting to get a budget for micro wave generator in the future. The field gradient of 122 Gauss/cm2, however, appeared to be too strong compared to B0=874 Gauss (the ECR condition for 2.45 GHz micro wave), and causing undesired radial components in the plasma region. This time the inner diameter of the sextupole magnet was increased from 90 mm to 120 mm, aiming for both the reduction of the transverse magnetic field and the improved pumping of the plasma region. The resulting field gradient is 53 Gauss/cm2 (see figure below). Also another Ti sublimation pump was installed, yielding the total pumping speed of 4000 litters/sec for hydrogen. These modifications finally led to the improvement of the polarization degree. For the pure-vector mode, a polarization degree as high as 85% of the ideal value was observed.
The remaining problem of the ion source operation was the short lifetime of the insulators, particularly at the beam-extraction section and at the Wien filter. All components of the extraction electrode which had been made of stainless steel were replaced by aluminum ones. The amount of metallic vapor was significantly reduced and today the ionizer can be operated without maintenance for more than half a year. The aluminum electrode, however, does not work well for the Wien filter, where a huge intensity of nitrogen ion hits the electrodes.
The above photo shows aluminum extraction electrods followed by the first Einzel lens, mounted on a flange for ease of maintenance. The unusually large aperture (5 cm dia.) and rich spaces between elements allow good pumping of the plasma region.
Monitor & control PIS Home Page RF transitions