Accelerators and ion storage ring TSR
During the last years the TSR has become an important tool in molecular physics. Only a storage ring offers the opportunity to get vibrational cold molecules just by storing them and waiting until they are relaxed. With the installed extraction system which uses the same electrostatic septum that is also used for injection, the TSR has become a source for cold molecular ion beams. The third order resonance of the horizontal tune Qx = 8/3 driven by one sextupole magnet is used to extract the stored ion beam.
Scheme of an extraction cycle using noise on the horizontal kicker
In an extraction experiment with 12C6+ ions (E=73.3 MeV) following extraction scheme as shown was applied: The ions were injected at a horizontal tune of Qx = 2.64. To reduce the transverse phase space of the ion beam electron cooling was used. After 1 second the cooler was switched off and the horizontal tune was shifted to Qx = 2.662 close to Qx = 8/3 by ramping one quadrupole family. To extract the ion beam the emittance of the stored beam was increased by applying noise to the horizontal kicker. The bandwidth of the noise was about 30 kHz. The center frequency f of the noise signal depends on the revolution frequency of the beam f0 and the non integer part of the tune q and has to fulfill following relation: f = f0 · (n ± q). The integer number n was chosen to 2. The extraction rate of the ions can be controlled by the noise level. In the extraction experiments the beam was extracted by typical noise powers of 10 W to 50 W. After 9.5 s the noise was switched off.
Time evolution of an extraction experiment
Applying noise to the kicker increases the beam profile. The development of the width of the horizontal beam profile (top) and the extraction rate (bottom) during the extraction process is shown in the figure beside. The standard deviation of a Gaussian fit to the beam profile is marked with blue crosses and the half beam width containing 68% of the ions is shown as red squares. The extraction rate was measured with a channeltron detector (green curve below). The first vertical dashed straight line marks the time when the cooler is switched off. Between the first and the second line the quadrupole is being ramped and between the second and the third line the noise is applied to the kicker. In the first second the reduction of the beam width due to electron cooling can be seen. In the time between 1 s and 1.3 s after injection the distance to the third order resonance is reduced by changing the quadrupole strength of one quadrupole family, corresponding to a reduction of the separatrix size. The electron pre-cooled beam size, however, is small enough that all ions are inside the separatrix, therefore no particles are extracted during the ramping time of the quadrupole. After 1.3 s the beam size is increased by using noise on the the kicker but still no ions are extracted. After 1.8 s some ions reach the separatrix and the extraction rate begins to increase. The maximum value of the extraction rate is reached after 4 seconds and the horizontal beam width of the stored ion beam remains constant. The spill of the extracted beam can be controlled by the noise level. An example for that is given in the last figure.
The efficiency ε of the slow extraction is defined by
In order to determine the efficiency the extraction rate and the number of stored ions before and after extraction were measured. It was necessary to use small stored ion currents below 1 µA in order to limit the extraction rate R due to the channeltron detector (CEM). The stored ion current was measured with the beam profile monitor detecting ionized residual gas molecules. The counting rate of the beam profile monitor, which is proportional to the stored ion current, was then calibrated with ion currents between 10 and 30 µA measured with the DC transformer. In the experiment shown in this figure the number of the stored particles decreased from N(ti ) = (270000 ± 10000) to N(tf ) = (52500 ± 5000) whereas the number of extracted particles measured with the CEM detector was 184000 resulting in a extraction efficiency of ε = 85 ± 5 %.
Controlling the extraction rate
This figure shows the count rate of the extracted particles as a function of time (blue curve). The noise applied to the kicker was periodically switched on and off.