Can a Hexapole magnet of an ECR Ion Source be too strong?

Drentje A. G. , Barzangy F. , Kremers H. R. , Meyer D. , Mulder J. , Sijbring J.
K.V.I. University Groningen, Zernikelaan 25, 9747AA Groningen, The Netherlands

Experience of many ECRIS designers and users during more than a decade has given a few experimental rules, or "scaling laws". Many of these have been discussed at the ECRIS workshops. After the 1993 workshop it was concluded that the properties of the magnetic trap, in particular the strength of the radial component, determine to a great deal the confinement characteristics.

For that reason it was decided at the KVI to choose a very strong magnet for the new 14 GHz ECRIS4 to be used in the Atomic Physics experiments. The hexapole magnet designed by the Giessen group is a good example. The induction, measured 2.5 mm inside the pole tips (i.e. at the wall of the plasma chamber) amounts more than 1.2 T. The measured radial field component Br obeys closely the expression Br= 0.00136 r2. (with B in T, r in mm). The quality of the magnetic trap can be given by the "radial mirror ratio", which is usually defined as R = Bmax / Breson, with Breson equal 0.5 T for a 14 GHz ECRIS. For the KVI magnet this would give R= 2.4 .

This hexapole magnet was installed in ECRIS4, which is in operation since 1996. ECRIS4 is to a large extent a copy of ECRIS3 at KVI; the main difference is that the hexapole magnet of ECRIS3 has a considerably lower mirror ratio, i.e. R = 1.86. ECRIS3 is in operation since 1995; its performance has been reported at the 1995 ECRIS workshop.

It turns out that the source with the stronger magnet produces much lower currents of highly charged ions. Various changes were made in the source set-up, including major alterations of the (axial) magnet configuration. The results of some of these changes, in terms of O7+ and Ar14+ currents, will be discussed.

One conclusion could be that, at the given RF frequency of 14 GHz, the strong hexapole magnet forms a basic limitation, because of the smaller ECR volume. At present this conclusion may seem obvious, but that was certainly not the case after the 1993 ECRIS workshop.

Presenting Author : Drentje A.G.
Presentation : Poster
Tuesday Poster Session at 16:00