Figure 1: The components of the ultra clean ion source.
For the ultrasensitive detection of stable isotopes and primordial
radionuclides, a new ultra clean ion source (UCS) was developed at the
accelerator laboratory in Garching. The produced ions are detected by
an AMS system. The characteristics of the ultra clean source are: 1, a
magnetically analyzed cesium sputter beam, 2, an ultra high vacuum
(10-10 hPa) in the sputter chamber and a large volume of the
sputter chamber to suppress cross contamination of the samples, and 3,
the use of supra pure metals for slits, electrodes and the target
holder. The new ion source is installed on a new negative ion injector
of the Munich MP tandem accelerator. At the high-energy side of the
tandem, the secondary ions are detected by a Bragg ionization
chamber. The first experiments were performed in order to determine
disturbing 40K background in cryogenic sapphire detectors which
are used for the detection of dark matter particles. Impurities in
semiconductor materials have also been investigated.
Since 1979, AMS (accelerator mass spectrometry) has been used successfully to determine cosmogenic and anthropogenic radionuclides with half-lives between 5000 and ten million years. The detection limit for the ratio radionuclide/element is as low as 10-15, because of the small isotopic abundances of e.g. 10-12 for 14C or 36Cl. The isotopic abundances of stable isotopes or primordial radionuclides are in the range of 10-5 to 1. Therefore the detection limit of primordial isotopes, due to contamination, should be larger by many orders of magnitude.
One method of detecting stable isotopes is SIMS (secondary ion mass spectrometry). For SIMS the detection limit is restricted by the natural background in the ion source. Therefore mass separation with high resolution is needed in order to discriminate molecules. The latter cannot be achieved in all cases. Using AMS, the molecular background is completely suppressed by the tandem terminal stripping foil combined with a magnet which analyzes the accelerated beam.
In order to check whether the method of standard AMS is suitable for stable
isotopes measurements, we measured phosphor concentrations in several Si wafers
with P/Si concentrations of 10-4, 10-6, and 10-8. A detection limit
of about 10-6 was found, which was caused by contamination in the ion
source[1]. In order to avoid these contaminations and to reach a detection limit up to 10-12 for stable isotopes too, a new ion source was designed as an
ultra clean facility and was built at the Munich tandem accelerator
laboratory. Such SIMS set-ups with AMS detection already exist at the
University of North Texas (references are cited in[2]) and the ETH
Zürich (references are cited in[3]). A similar development is
being carried out at the CSIRO facility in Sydney (references are
cited in[4]).
The main components of the Munich ultra clean ion source are a magnetically
analyzed cesium sputter beam, an ultra high vacuum in the large sputter
chamber, and the use of supra pure metals for all parts of the set up, which
could be hit by the cesium beam or by the secondary ion beam at the low-energy
side of the tandem.
The design of the ultra clean injector is shown in Fig.1[5].
Figure 1: The components of the ultra clean ion source.
The positive Cs ions get produced by surface ionization on a 1200° C hot tungsten ionizer. The Cs beam is extracted and
preaccelerated up to 30keV by an electrode with Pierce geometry. A system
consisting of steerer/electrostatic quadrupole lens and einzel lens images
the beam on a focus. By a double focussing magnet (deflection angle
120°, D m / m = 1/220)
the cesium beam is analyzed
and thus cleaned of impurities. A second set of steerer/quadrupole lens and
einzel lens transports the Cs beam on to the target at an angle of
30° to the surface. The target is fixed on a rotatory wheel with 12
positions. In order to suppress the crosstalk between samples, the sputter
chamber has a large inner diameter of 250mm and an ultra high vacuum of 10-10 hPa, obtained by a 520l/s turbo molecular pump. The secondary ions
are vertically extracted by a voltage up to 35kV. The ions are analyzed by a
double focusing 90° magnet (D m / m = 1/121).
The negative ions are preaccelerated by a voltage of 150keV, then are
injected into the tandem accelerator, stripped in the terminal foil (3µg
carbon), accelerated to the high-energy side and are analyzed by the Wien
velocity filter and the 90° magnet. The detector is a Bragg ionization
chamber. Beam intensity attenuators up to 3 · 10-8 can be used for
calibration measurements. The whole AMS set-up is shown in
Fig.2.
The transmission from the Cs ionizer to the target is 89%. Typical currents on the target are 30µA. Typical sputter rates for producing negative ions from the target are e.g. 2.5% (Si), 4.2% (S), 11% (Cl). Typical 28Si currents are 0.7µA.
In order to check on the influence of crosstalk or memory effects, a Al2O3 target was run for 3h, which yielded a counting rate of 1011 Al ions/s in the detector. In this run attenuators were used. The counting rate of pure Si measured immediately after the Al2O3 target was 103 Al ions/s.
Experimental data were taken for material analysis of silicon and for
background determination in sapphire.
Figure3 shows the sum of five measurements of 28Si, 28Si, 30Si, 31P and 32S in the ionization chamber. The isotopes are identified by atomic number and energy. During all measurements the magnetic rigidity of lenses and magnets is kept constant. Every isotope is measured in a separate measurement by adjusting the terminal voltage. Therefore the detection limits are not given by overlapping masses in the detector, but by background contributions of the impurities. In order to determine the concentration of a sample, the events of the silicon measurement and of the impurity measurement are counted, and the corresponding ratio is calculated. The impurities are determined by using calibration standards with known concentrations. The following detection limits were obtained:
Figure 3: Five spectra of sulfur, phosphorus and silicon in the detector, one on
the top of the other.
| 23Na/Si = 2 · 10-8 | 27Al/Si = 4 · 10-9 |
| 31P/Si = 4 · 10-8 | 39K/Si = 2 · 10-11 |
The low detection limits of Na and K in Si are due to the magnetic
analysis of the Cs sputter beam. Alkaline metals are the main
pollution of the used cesium. The obtained detection limit of Na, which
is three orders of magnitude larger than that for K, may be caused by
contaminations of the Si sample.
Cryogenic sapphire Al2O3 detectors are used for the detection of dark matter (see e.g.[6]). Among others, 40K is a disturbing source of background in these detectors. In case of natural abundances, 40K concentrations can be deduced from measured 39K concentrations.
In a first experiment, Al2O3 powder, which sapphire crystals are made
of, was investigated. The powder was given in a sample holder. With the method
mentiond above, in a first attempt the K concentration was measured to be <1ppm. The measurements will be continued with pieces of a crystal to avoid
the introduction of impurities during the handling of the powder.
The use of a dedicated ultra clean ion source allows to extend the standard
method of AMS to the detection of stable isotopes or primordial radionuclides.
The main uses of this new technique are on the one hand material analysis of
semiconductor materials e.g. Si or Ge, or of diamonds, and on the other hand
background identification in low level detectors e.g. cryogenic dark
matter detectors. The first experiments which were performed for both
applications are very promising.
We wish to thank F.Bittersberger (Wacker-Siltronic AG, Burghausen)
for supplying us with Si wafers with known concentrations of elements and
Dr.L.Zerle (MPI of Physics, Munich) for his help during the measurement of
the sapphire material.
