- Email: [email protected]
Production of titanium ion beams in an ECR ion source
Nuclear Instruments and Methods in Physics Research B 187 (2002) 111–116 www.elsevier.com/locate/nimb
Production of titanium ion beams in an ECR ion source € rje, R. Sepp€ H. Koivisto *, J. A al€ a, M. Nurmia Accelerator Laboratory, Department of Physics, University of Jyv€askyl€a, P.O. Box 35 (Y5), FIN-40351 Jyv€askyl€a, Finland Received 21 June 2001; received in revised form 20 August 2001
Abstract Intensive highly charged Ti ion beams were successfully produced in the 14 GHz ECR ion source at the Accelerator Laboratory, University of Jyv€askyl€a (JYFL). The Ti beams were produced using the MIVOC technique, i.e. by allowing the vapor of an organic compound containing titanium to diffuse into the ion source at room temperature. After optimizing the source parameters the intensity of the 48 Ti11þ ion beam reached a value of 45 lA. Ó 2002 Elsevier Science B.V. All rights reserved.
1. Introduction During the history of the ECR ion sources several methods have been developed for the production of ion beams from materials that are solids at room temperature. These include wall recycling [1], external [2] and internal ovens [3,4], insertion techniques [5,6], use of gaseous compounds, a sputtering technique [7] and the MIVOC method [8]. The selection of an appropriate method depends on properties of the desired element and partly on the structure of an ECR ion source too. The MIVOC method (metal ions from volatile compounds) was developed at JYFL for easy production of some metal ion beams. The first MIVOC beam, highly charged iron from ferrocene, was produced in 1992 [8]. The method is
*
Corresponding author. Tel.: +358-14-2602421; fax: +35814-2602401. E-mail address: [email protected].fi (H. Koivisto).
based on the fact that the consumption rate of a gaseous feed material in an ECR ion source is quite low, of the order of 0:1 cm3 =h (NTP). This applies to the ion beam gas when a plasma support gas ( ¼ gas mixing [9]) is used. It is thus possible to feed the ECRIS plasma with a low pressure vapor of a chemical compound kept at room temperature if the conductance of the feed line is sufficient. In the plasma energetic electrons break the compound molecules into individual atoms which are ionized and then extracted out of the source. If the vapor pressure of the compound used is high enough, the extra atoms can serve as plasma support gas so that no gas mixing is necessary. The production of an ion beam from pure titanium metal using the oven technique is difficult because a sufficient partial pressure of Ti in the source plasma chamber calls for a very high temperature of the solid material (melting point 1668 °C). Earlier experiments at JYFL aimed at the production of Ti beams using both MIVOC and sputtering techniques had only limited success [10].
0168-583X/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 1 ) 0 0 8 4 0 - 0
112
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
Several compounds were tried using MIVOC and a current of 200 nA of Ti10þ extracted from the K130-cyclotron was achieved using titaniumtetrachloride, a corrosive liquid. The beam currents obtained using sputtering were of similar magnitude. In this paper we report the development of the first successful highly charged Ti ion beam.
2. Experimental 2.1. Measurements with residual gas analyzer All promising MIVOC compounds are first studied with a residual gas analyzer. In this way most of the candidates can be excluded before running them in the ECR ion source. Practice has shown that a compound does not work in the ECR source if no signal of the atom of interest is seen in the gas analyzer. The compound successfully used in this experiment was (trimethyl)pentamethyl-cyclopentadienyltitanium, ðCH3 Þ5 C5 TiðCH3 Þ3 . No information concerning the vapor pressure of the compound at any temperature was found. The compound is fairly difficult to handle because of its sensitivity to air, moisture and light. As a consequence, the compound powder was loaded into the MIVOC chamber inside a nitrogen-filled glove box at a reduced light level. The MIVOC chamber was made of stainless steel with metal gaskets. As a next step the loaded chamber was connected to a residual gas analyzer, Vacscan Model 100F. Care was taken during the pump-down period to avoid drawing the light powder of the titanium compound into the pump. The operating vacuum range, below 105 mbar, was reached in a few minutes. The mass spectrum obtained is shown in Fig. 1. The partial pressures at the masses of the isotopes 46;47;48;49;50 Ti were found to be 1.6, 3.6, 13, 5.2 and 1.2, respectively, all in units of 108 mbar. The isotopic abundances calculated from these values are 6.2% (8.0%), 15% (7.3%), 54% (73.8%), 21% (5.5%) and 4.7% (5.4%), respectively, with the known abundances shown in parentheses. The deviation of the ratios from the known values indicates that the peaks are to some extent con-
Fig. 1. The measured mass spectrum from a Ti-compound in a residual gas analyzer. The original spectrum was redrawn for clarity.
taminated by impurity molecules such as TiH. The peaks disappeared after closing the MIVOC chamber valve and a background check with an empty MIVOC chamber confirmed that they originated from the titanium compound. Because of the encouraging results the MIVOC chamber with the compound was next connected to the JYFL 14 GHz ECR ion source [11] for a trial production of highly charged Ti ions. 2.2. Measurements at the JYFL 14 GHz ECRIS The MIVOC chamber loaded by approximately 100 mg of the ðCH3 Þ5 C5 TiðCH3 Þ3 powder was prepumped using the pumping system of the residual gas analyzer. The pre-pumped chamber was next connected axially to the injection side of the 14 GHz ECR ion source, as close to the source as possible. The connecting plastic tube (L 20 cm, i.d. ¼ 4 mm) was the conductance-limiting part of the vapor feeding system. A rough 2-turns on/offvalve was used between the MIVOC chamber and the source to control the flow rate of the compound vapor. This made tuning of the vapor feed somewhat cumbersome. Inside the source the vapor diffused through the stainless steel pipe into the plasma chamber. The pipe was made of two pieces (L1 23 cm, i.d. ¼ 12 mm and L2 12 cm, i.d. ¼ 5 mm) and was used to prevent the vapor from flowing into the injection side pump of the ion source. Additionally, in this test run the MIVOC chamber connection was to the flange of the
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
113
Fig. 2. The MIVOC chamber connection to the injection side of the JYFL 14 GHz ECRIS.
main gas feeding line so that no gas mixing could be used. Fig. 2 shows the arrangement used. In the experiment the on/off-valve was fully opened and the vapor of Ti-compound was allowed to flow into the plasma chamber. In the beginning the pressure exceeded the normal operation value. The MIVOC chamber was pumped for some time in order to evacuate all the residual gas from the chamber and pipe walls. Within 1 h the pressure in the plasma chamber had decreased to a suitable value and the microwave power was turned on to ignite the plasma. The normal ECRIS settings for production of highly charged ions were used. The ion source was tuned for the charge state of 11+. After an optimization procedure 45 lA of 48 Ti11þ ion beam was achieved. The microwave input power was about 350 W and an extraction voltage of 10 kV was used. A bias voltage was applied to the Faraday cup in order to suppress the emission of secondary electrons. An accurate regulation system for the vapor feed was not yet available in this test experiment. The pressures in the plasma chamber measured in plasma-on and -off conditions were 4:1 107 and 1:1 106 mbar, respectively. These values are slightly higher than normally needed for the production of highly charged ion beams. The gas feeding system was constructed after this experiment and was used
with the long-term stability tests described later in this paper. Fig. 3 shows the spectrum measured. As can be seen, the 48 Ti11þ and 48 Ti10þ peaks are clear without any interference from other elements. Some minor isotopes of titanium are also visible in the spectrum. Table 1 shows the best intensities obtained. The asterisk denotes intensity values estimated from the intensity of the same charge state of a different isotope using the known isotopic abundances in natural titanium. For example, the intensity of 48 Ti11þ ion beam was measured to be 45 lA. The abundance of the isotope of 46 Ti is 7.9%. Using these values its intensity was calculated to be 4.8 lA. Typical Hþ and C5þ;4þ;3þ intensities during titanium runs are approximately 120, 45, 130 and 100 lA, respectively. During the titanium run the vapor pressure in the plasma chamber was approximately the same as in the earlier ferrocene test run [11]. In the later case 115 lA of Fe11þ ion beam was achieved. This is more than twice the intensity of Ti11þ ion beam measured in the present experiment. The pressures in the plasma chamber measured in plasma-on conditions were 4:1 107 and 4:4 107 mbar in the titanium and iron runs, respectively. In both cases the pressure inside the plasma chamber was higher than needed. A possible explanation for this intensity ratio is in the structure of the molecules.
114
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
Fig. 3. The Ti ion spectrum obtained using the MIVOC method at the new JYFL 14 GHz ECRIS. The ion source and the optics were tuned for the charge state of 11+. The typical current of C4þ is 120–140 lA.
Table 1 Intensities of measured Ti ion beams in lA Isotope/charge 46
Ti Ti 48 Ti 49 Ti 50 Ti 47
7.9% 7.3% 73.9% 5.5% 5.3%
8+
9+
3.6 3.3 33 2.5 2.5
10+
4.7 4.3 43 3.2 3.2
11+
4.8 4.4 44 3.3 3.3
4.8 4.4 45 3.3 3.3
12+
13+
3.9 3.6 36 2.7 2.7
1.6 1.4 14.3 1.1 1.1
The abundance of the titanium isotopes in natural titanium is also shown. The asterisk denotes intensities estimated from the intensity of the same charge state of a different isotope. The extraction voltage was 10 kV.
The Ti compound, ðCH3 Þ5 C5 TiðCH3 Þ3 , contains 38 atoms and the Fe compound, FeðC5 H5 Þ2 , 21 atoms while both compounds contain only one atom of interest. For the same total number of atoms in the plasma, the amount of titanium in the plasma would be about one half of that of iron. As a next step, the long-term stability of the beam and the consumption rate of the material were studied. The gas feeding system of the new source was completed in order to make the accurate tuning of the vapor flow rate possible. Fig. 4 shows the arrangement. In the test the JYFL 14 GHz ECRIS was tuned for Ti10þ ion beam. After a short tuning procedure 30 lA of beam was obtained. The use of the gas mixing did not improve the intensity of Ti10þ ion beam. The tuning of
the source was kept constant during the test of 43 h. The intensity of the ion beam was very stable during the experiment. At the end of the test the intensity of Ti10þ beam was 29 lA. The carbon– titanium ratio was constant during the experiment indicating that no remarkable contamination was caused. The consumption of the compound was measured to be 47 mg giving the value of 0.22 mg/ h for the consumption of titanium. This gives the value of 3.0% for the production efficiency of Ti10þ ion beam. Here the production efficiency is mðTiqþ Þ=mðTiÞ;
ð1Þ
where mðTiqþ Þ is the mass flow rate of charge state q+ measured as an ion beam current in the Fara-
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
115
Fig. 4. TiF4 was tested with the miniature oven of the JYFL 14 GHz ECRIS. The gas feeding system has been completed. The MIVOC chamber is connected to the right-side gas feeding valve.
day cup and mðTiÞ is the mass flow rate of titanium fed into the ECR ion source. In another experiment the titanium ion beam was delivered for an experiment in nuclear physics. The duration of the experiment was 282 h. The requirement concerning the ion beam intensity on the target was 1 lA. The feeding rate of titanium compound was kept as low as possible to fulfil that requirement. In this way the possible carbon contamination of the ion source was minimized. Around 18–19 lA of Ti10þ ion beam was extracted from the source. The ion beam was very stable during the period of 282 h. After the experiment the consumption rate of titanium was measured to be 0.06 mg/h. As a consequence, the production efficiency for Ti10þ ions was as high as 7.3%. As a comparison, TiF4 was tested with the JYFL 14 GHz ECRIS. The compound has been used for example at LBNL where up to 10 lA of Ti10þ ion beam has been produced using their 6.4 GHz ECR ion source [2]. The compound requires about 50–80 °C before the ionization is possible. In order to heat the sample it was placed inside the
crucible of the miniature oven [12] of the new source. The oven was inserted axially (Fig. 4) to the position where the plasma chamber begins. However, this arrangement did not give a very encouraging result. The oven was overheated by the plasma and the neutral pressure went too high in order to get intensive, highly charged titanium ion beams. Next the oven was moved further from the plasma chamber in order to decrease the plasma heating of the oven. As a result, no positive gain in the intensity of the ion beam was achieved. During the test the maximum intensity of around 3 lA of Ti11þ was obtained.
3. Summary The MIVOC method has been used at the JYFL ECR ion source since 1992. During the past years the method has been adopted at several laboratories around the world. At the present time the number of different elemental ion beams produced worldwide using this technique is around 20 [13].
116
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
The MIVOC method offers a ‘‘gas-like’’ way to produce some ion beams of solids. Ion beams of non-volatile elements like Fe, Ni, W and Os, have been successfully produced using the MIVOC technique [14]. In this work the MIVOC method was applied to the titanium at the new JYFL 14 GHz ECR ion source. A beam of 45 lA of 48 Ti11þ was produced from ðCH3 Þ5 C5 TiðCH3 Þ3 powder at room temperature. Acknowledgements This work has been supported by the Academy of Finland under the Finnish Centre of Excellence Programme 2000–2005 (Project No. 44875, Nuclear and Condensed Matter Programme at JYFL). H.K. would like to acknowledge financial support from the Academy of Finland (Project No. 46323).
References [1] F. Bourg, R. Geller, B. Jacquot, M. Pontonnier, in: M.C. Boraso (Ed.), Proceedings of the 4th International Workshop on ECRIS, Grenoble, CENG, 1982, p. 5.1. [2] D.J. Clark, C.M. Lyneis, in: Proceedings of the International Conference on the Physics of Multiply Charged Ions and E.C.R. Ion Sources, 1998, Grenoble, Journal de Physique, Tome 50, Col. C1, Suppl. No. 1, 1989, p. C1-759.
[3] D. Hitz, G. Melin, M. Pontonnier, T.K.N. Guyen, in: Proceedings of the 11th International Workshop on Electron Cyclotron Resonance Ion Sources, Groningen, 6–7 May 1993, p. 91. [4] R. Harkewicz, Rev. Sci. Instrum. 67 (6) (1996) 2176. [5] R. Geller, B. Jacquot, in: Contributed Papers of the 7th Workshop on ECR Ions Sources, ISSN 0344-5798, 1986, p. 31. [6] T. Nakagawa, T. Kageyama, E. Ikezawa, M. Hemmi, Y. Miyazawa, in: Proceedings of the 10th Workshop on ECR Ion Sources, Oak Ridge National Laboratories, Knoxville, 1990, p. 163. [7] R. Harkewicz, P.J. Billquist, J.P. Greene, J.A. Nolen Jr., R.C. Pardo, Rev. Sci. Instrum. 66 (4) (1995) 2883. € rje, M. Nurmia, Nucl. Instr. and Meth. B [8] H. Koivisto, J. A 94 (1994) 291. [9] A.G. Drentje, in: Proceedings of the 11th International Workshop on ECRIS, Groningen, 1993, p. 155. € rje, M. Nurmia, in: [10] H. Koivisto, R. Harkewicz, J. A Proceedings of the 13th International Workshop on Electron Cyclotron Resonance Ion Sources, TAMU, College Station, 26–28 February 1997, p. 167. [11] H. Koivisto, P. Heikkinen, V. H€anninen, A. Lassila, H. € rje, Leinonen, V. Nieminen, J. Pakarinen, K. Ranttila, J. A E. Liukkonen, Nucl. Instr. and Meth. B 174 (2001) 379. € rje, E. [12] H. Koivisto, P. Heikkinen, K. Ranttila, J. A Liukkonen, in: Proceedings of the 16th International Conference on Cyclotrons and their Applications 2001, East Lansing, USA, 2001, p. P3-02. [13] http://www.phys.jyu.fi/research/accelerator/ionsources/mivoc.html. € rje, Y. Miyazawa, M. Hemmi, T. Chiba, [14] T. Nakagawa, J. A N. Inabe, M. Kase, T. Kageyama, O. Kamigaito, A. Goto, Y. Yano, Rev. Sci. Instrum. 69 (2, Part II) (1998) 637.
Production of titanium ion beams in an ECR ion source € rje, R. Sepp€ H. Koivisto *, J. A al€ a, M. Nurmia Accelerator Laboratory, Department of Physics, University of Jyv€askyl€a, P.O. Box 35 (Y5), FIN-40351 Jyv€askyl€a, Finland Received 21 June 2001; received in revised form 20 August 2001
Abstract Intensive highly charged Ti ion beams were successfully produced in the 14 GHz ECR ion source at the Accelerator Laboratory, University of Jyv€askyl€a (JYFL). The Ti beams were produced using the MIVOC technique, i.e. by allowing the vapor of an organic compound containing titanium to diffuse into the ion source at room temperature. After optimizing the source parameters the intensity of the 48 Ti11þ ion beam reached a value of 45 lA. Ó 2002 Elsevier Science B.V. All rights reserved.
1. Introduction During the history of the ECR ion sources several methods have been developed for the production of ion beams from materials that are solids at room temperature. These include wall recycling [1], external [2] and internal ovens [3,4], insertion techniques [5,6], use of gaseous compounds, a sputtering technique [7] and the MIVOC method [8]. The selection of an appropriate method depends on properties of the desired element and partly on the structure of an ECR ion source too. The MIVOC method (metal ions from volatile compounds) was developed at JYFL for easy production of some metal ion beams. The first MIVOC beam, highly charged iron from ferrocene, was produced in 1992 [8]. The method is
*
Corresponding author. Tel.: +358-14-2602421; fax: +35814-2602401. E-mail address: [email protected].fi (H. Koivisto).
based on the fact that the consumption rate of a gaseous feed material in an ECR ion source is quite low, of the order of 0:1 cm3 =h (NTP). This applies to the ion beam gas when a plasma support gas ( ¼ gas mixing [9]) is used. It is thus possible to feed the ECRIS plasma with a low pressure vapor of a chemical compound kept at room temperature if the conductance of the feed line is sufficient. In the plasma energetic electrons break the compound molecules into individual atoms which are ionized and then extracted out of the source. If the vapor pressure of the compound used is high enough, the extra atoms can serve as plasma support gas so that no gas mixing is necessary. The production of an ion beam from pure titanium metal using the oven technique is difficult because a sufficient partial pressure of Ti in the source plasma chamber calls for a very high temperature of the solid material (melting point 1668 °C). Earlier experiments at JYFL aimed at the production of Ti beams using both MIVOC and sputtering techniques had only limited success [10].
0168-583X/02/$ - see front matter Ó 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 0 1 ) 0 0 8 4 0 - 0
112
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
Several compounds were tried using MIVOC and a current of 200 nA of Ti10þ extracted from the K130-cyclotron was achieved using titaniumtetrachloride, a corrosive liquid. The beam currents obtained using sputtering were of similar magnitude. In this paper we report the development of the first successful highly charged Ti ion beam.
2. Experimental 2.1. Measurements with residual gas analyzer All promising MIVOC compounds are first studied with a residual gas analyzer. In this way most of the candidates can be excluded before running them in the ECR ion source. Practice has shown that a compound does not work in the ECR source if no signal of the atom of interest is seen in the gas analyzer. The compound successfully used in this experiment was (trimethyl)pentamethyl-cyclopentadienyltitanium, ðCH3 Þ5 C5 TiðCH3 Þ3 . No information concerning the vapor pressure of the compound at any temperature was found. The compound is fairly difficult to handle because of its sensitivity to air, moisture and light. As a consequence, the compound powder was loaded into the MIVOC chamber inside a nitrogen-filled glove box at a reduced light level. The MIVOC chamber was made of stainless steel with metal gaskets. As a next step the loaded chamber was connected to a residual gas analyzer, Vacscan Model 100F. Care was taken during the pump-down period to avoid drawing the light powder of the titanium compound into the pump. The operating vacuum range, below 105 mbar, was reached in a few minutes. The mass spectrum obtained is shown in Fig. 1. The partial pressures at the masses of the isotopes 46;47;48;49;50 Ti were found to be 1.6, 3.6, 13, 5.2 and 1.2, respectively, all in units of 108 mbar. The isotopic abundances calculated from these values are 6.2% (8.0%), 15% (7.3%), 54% (73.8%), 21% (5.5%) and 4.7% (5.4%), respectively, with the known abundances shown in parentheses. The deviation of the ratios from the known values indicates that the peaks are to some extent con-
Fig. 1. The measured mass spectrum from a Ti-compound in a residual gas analyzer. The original spectrum was redrawn for clarity.
taminated by impurity molecules such as TiH. The peaks disappeared after closing the MIVOC chamber valve and a background check with an empty MIVOC chamber confirmed that they originated from the titanium compound. Because of the encouraging results the MIVOC chamber with the compound was next connected to the JYFL 14 GHz ECR ion source [11] for a trial production of highly charged Ti ions. 2.2. Measurements at the JYFL 14 GHz ECRIS The MIVOC chamber loaded by approximately 100 mg of the ðCH3 Þ5 C5 TiðCH3 Þ3 powder was prepumped using the pumping system of the residual gas analyzer. The pre-pumped chamber was next connected axially to the injection side of the 14 GHz ECR ion source, as close to the source as possible. The connecting plastic tube (L 20 cm, i.d. ¼ 4 mm) was the conductance-limiting part of the vapor feeding system. A rough 2-turns on/offvalve was used between the MIVOC chamber and the source to control the flow rate of the compound vapor. This made tuning of the vapor feed somewhat cumbersome. Inside the source the vapor diffused through the stainless steel pipe into the plasma chamber. The pipe was made of two pieces (L1 23 cm, i.d. ¼ 12 mm and L2 12 cm, i.d. ¼ 5 mm) and was used to prevent the vapor from flowing into the injection side pump of the ion source. Additionally, in this test run the MIVOC chamber connection was to the flange of the
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
113
Fig. 2. The MIVOC chamber connection to the injection side of the JYFL 14 GHz ECRIS.
main gas feeding line so that no gas mixing could be used. Fig. 2 shows the arrangement used. In the experiment the on/off-valve was fully opened and the vapor of Ti-compound was allowed to flow into the plasma chamber. In the beginning the pressure exceeded the normal operation value. The MIVOC chamber was pumped for some time in order to evacuate all the residual gas from the chamber and pipe walls. Within 1 h the pressure in the plasma chamber had decreased to a suitable value and the microwave power was turned on to ignite the plasma. The normal ECRIS settings for production of highly charged ions were used. The ion source was tuned for the charge state of 11+. After an optimization procedure 45 lA of 48 Ti11þ ion beam was achieved. The microwave input power was about 350 W and an extraction voltage of 10 kV was used. A bias voltage was applied to the Faraday cup in order to suppress the emission of secondary electrons. An accurate regulation system for the vapor feed was not yet available in this test experiment. The pressures in the plasma chamber measured in plasma-on and -off conditions were 4:1 107 and 1:1 106 mbar, respectively. These values are slightly higher than normally needed for the production of highly charged ion beams. The gas feeding system was constructed after this experiment and was used
with the long-term stability tests described later in this paper. Fig. 3 shows the spectrum measured. As can be seen, the 48 Ti11þ and 48 Ti10þ peaks are clear without any interference from other elements. Some minor isotopes of titanium are also visible in the spectrum. Table 1 shows the best intensities obtained. The asterisk denotes intensity values estimated from the intensity of the same charge state of a different isotope using the known isotopic abundances in natural titanium. For example, the intensity of 48 Ti11þ ion beam was measured to be 45 lA. The abundance of the isotope of 46 Ti is 7.9%. Using these values its intensity was calculated to be 4.8 lA. Typical Hþ and C5þ;4þ;3þ intensities during titanium runs are approximately 120, 45, 130 and 100 lA, respectively. During the titanium run the vapor pressure in the plasma chamber was approximately the same as in the earlier ferrocene test run [11]. In the later case 115 lA of Fe11þ ion beam was achieved. This is more than twice the intensity of Ti11þ ion beam measured in the present experiment. The pressures in the plasma chamber measured in plasma-on conditions were 4:1 107 and 4:4 107 mbar in the titanium and iron runs, respectively. In both cases the pressure inside the plasma chamber was higher than needed. A possible explanation for this intensity ratio is in the structure of the molecules.
114
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
Fig. 3. The Ti ion spectrum obtained using the MIVOC method at the new JYFL 14 GHz ECRIS. The ion source and the optics were tuned for the charge state of 11+. The typical current of C4þ is 120–140 lA.
Table 1 Intensities of measured Ti ion beams in lA Isotope/charge 46
Ti Ti 48 Ti 49 Ti 50 Ti 47
7.9% 7.3% 73.9% 5.5% 5.3%
8+
9+
3.6 3.3 33 2.5 2.5
10+
4.7 4.3 43 3.2 3.2
11+
4.8 4.4 44 3.3 3.3
4.8 4.4 45 3.3 3.3
12+
13+
3.9 3.6 36 2.7 2.7
1.6 1.4 14.3 1.1 1.1
The abundance of the titanium isotopes in natural titanium is also shown. The asterisk denotes intensities estimated from the intensity of the same charge state of a different isotope. The extraction voltage was 10 kV.
The Ti compound, ðCH3 Þ5 C5 TiðCH3 Þ3 , contains 38 atoms and the Fe compound, FeðC5 H5 Þ2 , 21 atoms while both compounds contain only one atom of interest. For the same total number of atoms in the plasma, the amount of titanium in the plasma would be about one half of that of iron. As a next step, the long-term stability of the beam and the consumption rate of the material were studied. The gas feeding system of the new source was completed in order to make the accurate tuning of the vapor flow rate possible. Fig. 4 shows the arrangement. In the test the JYFL 14 GHz ECRIS was tuned for Ti10þ ion beam. After a short tuning procedure 30 lA of beam was obtained. The use of the gas mixing did not improve the intensity of Ti10þ ion beam. The tuning of
the source was kept constant during the test of 43 h. The intensity of the ion beam was very stable during the experiment. At the end of the test the intensity of Ti10þ beam was 29 lA. The carbon– titanium ratio was constant during the experiment indicating that no remarkable contamination was caused. The consumption of the compound was measured to be 47 mg giving the value of 0.22 mg/ h for the consumption of titanium. This gives the value of 3.0% for the production efficiency of Ti10þ ion beam. Here the production efficiency is mðTiqþ Þ=mðTiÞ;
ð1Þ
where mðTiqþ Þ is the mass flow rate of charge state q+ measured as an ion beam current in the Fara-
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
115
Fig. 4. TiF4 was tested with the miniature oven of the JYFL 14 GHz ECRIS. The gas feeding system has been completed. The MIVOC chamber is connected to the right-side gas feeding valve.
day cup and mðTiÞ is the mass flow rate of titanium fed into the ECR ion source. In another experiment the titanium ion beam was delivered for an experiment in nuclear physics. The duration of the experiment was 282 h. The requirement concerning the ion beam intensity on the target was 1 lA. The feeding rate of titanium compound was kept as low as possible to fulfil that requirement. In this way the possible carbon contamination of the ion source was minimized. Around 18–19 lA of Ti10þ ion beam was extracted from the source. The ion beam was very stable during the period of 282 h. After the experiment the consumption rate of titanium was measured to be 0.06 mg/h. As a consequence, the production efficiency for Ti10þ ions was as high as 7.3%. As a comparison, TiF4 was tested with the JYFL 14 GHz ECRIS. The compound has been used for example at LBNL where up to 10 lA of Ti10þ ion beam has been produced using their 6.4 GHz ECR ion source [2]. The compound requires about 50–80 °C before the ionization is possible. In order to heat the sample it was placed inside the
crucible of the miniature oven [12] of the new source. The oven was inserted axially (Fig. 4) to the position where the plasma chamber begins. However, this arrangement did not give a very encouraging result. The oven was overheated by the plasma and the neutral pressure went too high in order to get intensive, highly charged titanium ion beams. Next the oven was moved further from the plasma chamber in order to decrease the plasma heating of the oven. As a result, no positive gain in the intensity of the ion beam was achieved. During the test the maximum intensity of around 3 lA of Ti11þ was obtained.
3. Summary The MIVOC method has been used at the JYFL ECR ion source since 1992. During the past years the method has been adopted at several laboratories around the world. At the present time the number of different elemental ion beams produced worldwide using this technique is around 20 [13].
116
H. Koivisto et al. / Nucl. Instr. and Meth. in Phys. Res. B 187 (2002) 111–116
The MIVOC method offers a ‘‘gas-like’’ way to produce some ion beams of solids. Ion beams of non-volatile elements like Fe, Ni, W and Os, have been successfully produced using the MIVOC technique [14]. In this work the MIVOC method was applied to the titanium at the new JYFL 14 GHz ECR ion source. A beam of 45 lA of 48 Ti11þ was produced from ðCH3 Þ5 C5 TiðCH3 Þ3 powder at room temperature. Acknowledgements This work has been supported by the Academy of Finland under the Finnish Centre of Excellence Programme 2000–2005 (Project No. 44875, Nuclear and Condensed Matter Programme at JYFL). H.K. would like to acknowledge financial support from the Academy of Finland (Project No. 46323).
References [1] F. Bourg, R. Geller, B. Jacquot, M. Pontonnier, in: M.C. Boraso (Ed.), Proceedings of the 4th International Workshop on ECRIS, Grenoble, CENG, 1982, p. 5.1. [2] D.J. Clark, C.M. Lyneis, in: Proceedings of the International Conference on the Physics of Multiply Charged Ions and E.C.R. Ion Sources, 1998, Grenoble, Journal de Physique, Tome 50, Col. C1, Suppl. No. 1, 1989, p. C1-759.
[3] D. Hitz, G. Melin, M. Pontonnier, T.K.N. Guyen, in: Proceedings of the 11th International Workshop on Electron Cyclotron Resonance Ion Sources, Groningen, 6–7 May 1993, p. 91. [4] R. Harkewicz, Rev. Sci. Instrum. 67 (6) (1996) 2176. [5] R. Geller, B. Jacquot, in: Contributed Papers of the 7th Workshop on ECR Ions Sources, ISSN 0344-5798, 1986, p. 31. [6] T. Nakagawa, T. Kageyama, E. Ikezawa, M. Hemmi, Y. Miyazawa, in: Proceedings of the 10th Workshop on ECR Ion Sources, Oak Ridge National Laboratories, Knoxville, 1990, p. 163. [7] R. Harkewicz, P.J. Billquist, J.P. Greene, J.A. Nolen Jr., R.C. Pardo, Rev. Sci. Instrum. 66 (4) (1995) 2883. € rje, M. Nurmia, Nucl. Instr. and Meth. B [8] H. Koivisto, J. A 94 (1994) 291. [9] A.G. Drentje, in: Proceedings of the 11th International Workshop on ECRIS, Groningen, 1993, p. 155. € rje, M. Nurmia, in: [10] H. Koivisto, R. Harkewicz, J. A Proceedings of the 13th International Workshop on Electron Cyclotron Resonance Ion Sources, TAMU, College Station, 26–28 February 1997, p. 167. [11] H. Koivisto, P. Heikkinen, V. H€anninen, A. Lassila, H. € rje, Leinonen, V. Nieminen, J. Pakarinen, K. Ranttila, J. A E. Liukkonen, Nucl. Instr. and Meth. B 174 (2001) 379. € rje, E. [12] H. Koivisto, P. Heikkinen, K. Ranttila, J. A Liukkonen, in: Proceedings of the 16th International Conference on Cyclotrons and their Applications 2001, East Lansing, USA, 2001, p. P3-02. [13] http://www.phys.jyu.fi/research/accelerator/ionsources/mivoc.html. € rje, Y. Miyazawa, M. Hemmi, T. Chiba, [14] T. Nakagawa, J. A N. Inabe, M. Kase, T. Kageyama, O. Kamigaito, A. Goto, Y. Yano, Rev. Sci. Instrum. 69 (2, Part II) (1998) 637.