ionsource-test

Ion Sources and Related Techniques

After release from the catcher the reaction products have to be re-ionized to charge-state 1⁺ with high efficiency. Therefore various ion sources are available comprising three gaseous discharge ion sources of the FEBIAD-type operating in the temperature range of 1700 - 2300K, and hot-cavity thermoionizers operating up to 2800K. The latter may also be used for multi-step resonant laser ionization, or for ionization of metal-fluorides formed on-line by addition of CF₄- vapour, both methods enabling highly selective chemical separations. As a general rule, elements with an enthalpy of adsorption of H_a < 6 eV with respect to the walls of the ion source can be handled with the discharge sources, elements with H_a < 7 eV and an ionization potential W_i < 6.5 eV by the thermoionizers. For ions above argon FEBIAD-sources have ionization efficiencies between 20 and 70%. The ionization efficiency of a thermoionizer is close to 100% for W_i < 5 eV and decreases to a few percent for W_i = 6.5 eV. The feasibility of an experiment depends thus in general only on the halflife of the considered nucleus and on the delay between production and detection. Delays arise both in the catcher (mean solid-matter diffusion time) and in the enclosure of the ion source (mean effusion time due to molecular flow plus the manifold ad/desorption processes at the walls). These delays do not only set efficiency limits as outlined in the table below, but are also the base of "chemical" selectivity, using differences in diffusion coefficients, H_a, and W_i for neighbouring elements. Also the formation of molecular ions may enable selectivity or the on-line separation of species otherwise impossible. In some cases selectivity (even for several elements simultaneously) can be reached by the technique of bunched beam-release which may also be of further advantage, e.g. for background reduction in collinear laser experiments.

Classification of the elements for their ISOL-performance

Only elements above argon are considered, since lighter elements are in general not considered typical for GSI-ISOL experiments. The numbers below the element symbols indicate that separation efficiencies of the order of 1% or more are reached for half-lives above 30 ms (1), above 1 second (2), above 1 minute (3), respectively, or are excluded from ISOL unless tried as molecular ion (4). A "z" indicates possibilities to reach Z-selectivity at least to some extent, often for half-lifes not too short, and eventually at the expense of efficiency. A "?" indicates that classification is plausible but not based on strict experimental evidence.

1?z

Latest News: On-line Chemisty for Short-lived Tin Isotopes.

The detailed spectroscopic investigation of short-lived isotopes of tin, e.g. of the isotopes more neutron-deficient than ¹⁰⁴Sn formed in the reaction ⁵⁰Cr(⁵⁸Ni, a xn), was so far always hindered by the dominance of the simultaneously produced isobaric indium, cadmium, and silver isotopes. Initiated by the recent www-release [1] from Oak Ridge that the formation of SnS molecules permits a powerful separation of tin from antimony and tellurium, we looked into the possibility whether this on-line chemistry would also suppress indium, silver and cadmium.

The results of these off-line tests using vapours of stable Sn, In, Cd, and Ag can be summarized as follows:

The formation of SnS molecules, which was found in Oak Ridge as a consequence of a sulphur-contaminated target, can be performed in a defined way by using vapours of either gaseous H₂S, CS₂ (a liquid), or from solid sulphur powder. As shown in the figure, depending on the operation mode of the FEBIAD-B2-C [2] ion source, about 40 – 60% of the tin ions coming out of the ion source appear in the sulfide sideband, at sulphur partial pressures that do not obstruct the ion-source operation. The sulfide-forming agents and the test vapours of tin, plus either indium, cadmium, or silver, were fed into the ion source via two separate gas lines to avoid an unrealistic simulation of the on-line situation by forming sulfides already outside the ion source. The data displayed in the figure stem from six ion-source set-ups differing in test vapour composition and/or sulfide-forming agent, thus showing the reproducibiltity of the method. The relatively strong dependence of the SnS⁺/Sn⁺ ratio on the mean temperature of the ion-source enclosure indicates that the SnS⁺ molecular ion is at the limits of its stability. The fact that sulphur is not mono-isotopic (^32,33,34,36S with 95%, 0.75%, 4.2%, and 0.02% abundance) is of minor relevance for neutron-deficient tin isotopes: a potential contamination by lighter tin isotopes is negligible due to both the much lower production cross-section and the low abundance of the competing sulphur isotopes.

Figure: Measured probability for the formation of SnS⁺ versus the current of ³²S⁺ out of the ion source. This current is a direct measure for the partial pressure of sulphur in the ion source, since the ionization efficiency (monitored via krypton from a test-leak) is hardly affected by the addition of sulphur within the tested range. Apart from the strong toxicity of H₂S, the stated sulfide-forming agents are about equally unproblematic, only the additionally tested SF₆obstructs the ion source operation long before a useful sulphur partial pressure is reached. The variation of the mean temperature of the ion-source enclosure was performed by small changes of the cathode temperature, thus changing the discharge current without changing the discharge voltage.

At ³²S⁺ currents of 700-900 nA, the suppression of contaminations in the sulfide sideband was measured to be

SbS⁺/Sb⁺ £ 5*10^-3

InS⁺/In⁺ £ 3*10^-3

CdS⁺/Cd⁺ < 3*10^-4

AgS⁺/Ag⁺ < 10^-4.

Even though the suppression of SbS⁺ is not relevant in our case, it was measured, since it was as ever-present impurity available anyway, and confirms the result by Oak Ridge. Since the ion currents in the sidebands were, due to the strong suppression, only of the order of pA, i.e. of the order of the background-at-every-mass of a high-temperature ion source, the actual presence of contaminating sulfides is not easily traced. The ratios given above may well be zero for cadmium and silver, while the measured currents for SbS⁺ and InS⁺ reconcile with the respective isotopic signatures well enough to indicate that the ratios are at or close to the upper limits stated above.

Oak Ridge concluded from their data that the transport of SnS may be fast. Our measurements of the mean wall sticking times (which are critical for short-lived tin isotopes) by the „bunching technique" [3] with a heatable cold-trap were not perfectly conclusive. The most likely interpretation of the characteristic variation of the ion currents versus the cooling and heating of the cold-trap is, however, that the wall sticking times for the tin species become negligible at a sufficiently high sulphur vapour pressure. Actually under these conditions no difference in the behaviour of Sn⁺ and SnS⁺ isnoticed: this indicates strongly that tin migrates in the ion source to practically 100% as (neutral) SnS, and the splitting into Sn⁺ and SnS⁺ is due to the ionization process (electron impact) or due to thermal dissociation of the molecular ions. Also the mobility of antimony seems to be very much increased by the presence of sulphur vapour. This finding is not necessarily in contradiction to the fact that the SbS⁺ ion hardly shows up: the faster migration of antimony may well be via SbS molecules which, however, in contrast to SnS, dissociate moreless completely when ionized. This explanation is analogue to the situation for alkali-fluorides, where the respective molecular ions are practically suppressed to zero, although the neutral molecules are known to have a sufficient thermal stability [4].

[1] D.C. Stracener, http://www.phy.ornl.gov/hribf/usersgroup/news/spring2_01.html

[2] R. Kirchner et al., Nucl. Instr. Meth. A234 (1985) 224

[3] R. Kirchner et al., Nucl. Instr. Meth. A247 (1986) 265 and B26 (1987) 204

[4] R. Kirchner, Nucl. Instr. Meth. B126 (1997) 135

R. Kirchner, Aug. 31, 2001, revised Sept. 11, 2001