Purification of Cobalt, Nickel, and Titanium by Cold-Crucible Induction Melting in Ultrahigh Vacuum

Transcription

1 Materials Transactions, Vol. 47, No. 1 (2006) pp. 156 to 161 #2006 The Japan Institute of Metals Purification of Cobalt, Nickel, and Titanium by Cold-Crucible Induction Melting in Ultrahigh Vacuum Seiichi Takaki and Kenji Abiko Institute for Materials Research, Tohoku University, Sendai , Japan Three separate charges of cobalt 9.2 kg in weight, 3.0 kg nickel and 1.6 kg titanium have been melted in ultrahigh vacuum (UHV) using cold-crucible induction melting (CCIM) furnace, in which UHV of better than Pa is attainable. The effects of CCIM in UHV on the purification of these metals due to the removal of gaseous impurities, and the characteristics of CCIM in UHV have been investigated by the analysis of the melting conditions such as the total pressure and the mass spectra of gases during CCIM, and analytical results of these ingots. CCIM in the present UHV degree and quality is found to be very effective for the purification due to the removal of gaseous impurities in cobalt and nickel, in particular, for the decrease in oxygen, and very important for melting of ultrahigh-purity titanium without any contamination by gaseous impurities, in particular, oxygen. (Received August 29, 2005; Accepted November 24, 2005; Published January 15, 2006) Keywords: cobalt, nickel, titanium, gaseous impurities, high-purity, ultrahigh-purity, purification, ultrahigh vacuum, cold-crucible induction melting, trace analysis 1. Introduction The nominal purity of high-purity metals described as, for example, 4-nine (4N) or 5N is known to exclude the total concentration of gaseous impurities, which is equal to or higher than that of metallic impurities in many cases (for example, see Sect. 2.1). The gaseous impurities have a strong influence on the properties not only as solute atoms but also as precipitates such as carbides, oxides, and sulfides. Purification due to decrease in gaseous impurities is thus very important for the fundamental research on the intrinsic properties of a metal and also to determine the inherent effects of each impurity element on the properties. There have been many studies of the purification of cobalt, nickel and titanium reported in the literature, 1 12) but very little is known about the effect of melting in ultrahigh vacuum (UHV) on it. The main reason is that it is not easy even in a small furnace to keep the vacuum at the pressure of less than Pa before melting and less than Pa during melting of these metals, still less in a large-scale melting furnace. We reported in UHPM-97 on the construction of a new cold-crucible induction melting furnace 13) in which ultrahigh vacuum of less than Pa is attainable and an ingot up to 10 kg in weight for iron can be made. Since then we have published many papers 14 16) on preparation, trace analysis, and characteristic properties of ultrahigh-purity metals and alloys. We prepared, for example, ultrahighpurity iron ingot 10 kg in weight of higher than % purity after the analysis of 33 elements and containing the concentration of C þ N þ O þ S of less than 3.4 mass ppm using the CCIM furnace, 17) high-purity Fe Cr system alloys, 18,19) high-purity Ti Al alloys, 20) and so on. The purpose of the present paper is to investigate the effect of CCIM under ultrahigh vacuum of 10 7 Pa on the purification of high-purity nickel, cobalt and titanium due to the removal of gaseous impurities and thus to establish a new purification procedure to obtain high-purity these metals even on the scale of one to ten kg in weight. 2. Experimental Procedure 2.1 Starting materials High-purity electrolytic cobalt of 5N-purity, pure electrolytic cobalt (3N), high-purity electrolytic nickel (5N), and high-purity titanium produced by Kroll method of 5N purity were used as the starting materials. The concentration of impurities in two kinds of starting cobalt is shown in Table 1. The concentration of gaseous impurities and the other impurities in starting nickel is shown in Tables 4 and 5, respectively. That in starting titanium is shown in Tables 6 and 7. It is found from these Tables that the total concentration of gaseous impurities in each metal is excluded from the above indications of nominal-purity by the suppliers. Table 1 Concentration of impurities except gaseous impurities in starting high-purity cobalt (5N) and pure cobalt (3N) (mass ppm). Elements Starting Co (5N) Starting Co (3N) C N O S Ag <0:1 Al <0:1 <10 B <0:1 Ca <0:1 Cl <0:1 Cr <0:1 <10 Cu <0:1 <10 Fe K <0:1 <1 Mg <0:1 Mn <0:1 Na <0:1 <1 Ni Si <0:1 Sn <0:1 V <0:1 Zn 1.5

2 Purification of Cobalt, Nickel, and Titanium by Cold-Crucible Induction Melting in Ultrahigh Vacuum Cold-crucible induction melting Two induction melting furnaces with a water-cooled copper-crucible have been used for the present melting. The larger furnace, which is described in detail in Ref. 13), can make an ingot up to 10 kg in weight for iron. The main chamber which has two flanges of 1.4 m in diameter and a volume of about 2.3 m 3 can be evacuated to a base pressure of 6: Pa by a system of an oil diffusion pump and a cold trap. The other small furnace can make an ingot up to 3 kg in weight for iron. The chamber can be evacuated to a base pressure of 6: Pa by a system of two oil diffusion pumps and a cold trap. The common power supply provides maximum 260 kw and a frequency of 10 khz. The total pressure and the mass spectra of mass numbers between 1 and 50 in both main chambers were measured before, during and after melting as seen below. 2.3 Analysis The samples for analysis were cut from each starting material and each ingot melted in the present experiment. They were mechanically and then electrolytically polished. The gaseous impurities in cobalt, nickel and titanium were analyzed as follows: carbon and sulfur by the combustion infrared absorption method on the LECO CS444LS, nitrogen by the fusion thermal conductivity method on the LECO TC436, and oxygen by the fusion infrared absorption method on the LECO TC436. It is difficult to analyze low concentration of oxygen in titanium. The oxygen analysis was, thus, carefully performed in the same way as Ref. 21); that is, it was improved by the addition of nickel as an accelerator for complete combustion of titanium, because the conventional analytical method without the addition of nickel gives a lower analytical value for oxygen. The impurities except gaseous impurities in cobalt, nickel and titanium were analyzed by glow-discharge mass-spectrometry (GDMS). Pressure, p / Pa power on Total pressure melt-down melt-beginning Mass spectrum Time, t / sec power off Mass No. Fig. 1 Changes in the total pressure and the mass spectra of gases of mass numbers 2, 18, 28, 32 and 44 during induction heating and melting of highpurity cobalt (5N) Ion Current, i / A 3. Results and Discussion 3.1 Purification of cobalt High-purity cobalt (5N) 9.2 kg in weight was melted using the large CCIM furnace, and also pure cobalt (3N) 2.4 kg in weight was melted using the small CCIM furnace for comparison. Figure 1 shows the changes in the total pressure and the mass spectra of gases of mass numbers 2, 18, 28, 32 and 44, which are most likely to be H 2,H 2 O, CO, O 2 and CO 2, during induction heating and melting of pure cobalt (5N). The vacuum before melting was 1: Pa and the pressure during melting after melt-down to surface solidification was kept from 1: to 4: Pa for especially long time 32 min. The change in weight before and after melting was so small as to be 0.7%. This means that the temperature of the melt is not so higher than the melting point, because the melt stirred strongly by eddy current is cooled by water-cooled cupper-wall of the crucible. The picture of the high-purity cobalt ingot made of 5N cobalt by cold-crucible induction melting in UHV is shown in Fig. 2. The sampling locations for chemical analysis in a diametrical vertical section of the high-purity cobalt ingot (5N) and pure cobalt ingot (3N) are shown in Figs. 3 and 4, respectively. The analytical results of the high-purity cobalt ingot (5N) and Fig. 2 The 9.1 kg ultrahigh-purity ingot of cobalt (5N) melted in the coldcrucible furnace at the pressure of 1: to 4: Pa during melting. pure cobalt ingot (3N) are shown in Tables 2 and 3, respectively. In both cobalt ingots oxygen decreased appreciably by CCIM in UHV, although the intensity of the oxygen spectrum during heating and melting of 5N cobalt is as low as the order of A before melt-down and the order of A after melt-down. In high-purity cobalt ingot (5N) carbon and nitrogen decreased from 1.2 and 0.8 mass ppm to nearly 0.2 mass ppm, almost the level of their detection limit, respectively. In pure cobalt ingot (3N) carbon and nitrogen decreased from 6.7 and 1.4 mass ppm to nearly 1 mass ppm, respectively. The difference in the value attained is probably caused by the difference in melting time between 5N-cobalt (32 min) and 3N-cobalt (12 min) and/or also in concentration

3 158 S. Takaki and K. Abiko Table 2 Concentration of gaseous impurities in starting high-purity cobalt (5N) and the purified cobalt ingot melted in UHV in cold-crucible furnace (mass ppm) C N O S Starting Co (5N) <0: Ingot <0:1 <0: Fig. 3 Sampling locations for chemical analysis in a diametrical vertical section of the high-purity cobalt ingot (5N). L R Table 3 Concentration of gaseous impurities in starting pure cobalt (3N) and the pure cobalt ingot melted in UHV in cold-crucible furnace (mass ppm). C N O S Starting Co (3N) L Ingot R B Fig. 4 Sampling locations for chemical analysis in a diametrical vertical section of the pure cobalt ingot (3N). of impurities such as Al, Cr and V, which are nitrideformation elements. Each sulfur content in starting materials and ingots of both high-purity and pure cobalt is considered to be almost the same and on the detection limit of sulfur analysis in cobalt of 1 ppm. The removal of gaseous impurities has been performed often by heating and melting the metals in hydrogen. Isshiki et al. 5) prepared the purest cobalt in the previous literature by anion exchange separation, electrolytic extraction, floatingzone refining, and dry hydrogen treatment for :3 mm 3 thin sheet, and analyzed carbon, nitrogen and oxygen by the activation analysis. 6) The analytical values of carbon, nitrogen and oxygen are 1.7, 1.5, and 10.5 mass ppm, respectively, which are one order of magnitude higher than each content in our high-purity 9 kg cobalt ingot. We conclude that CCIM in the present UHV degree and quality is very effective for the purification of cobalt due to the removal of gaseous impurities. B 3.2 Purification of nickel High-purity nickel (5N) 3.0 kg in weight was melted using the small CCIM furnace. Figure 5 shows the changes in the total pressure and the mass spectra of gases of mass numbers 2, 18, 28, 32 and 44 during induction heating and melting of high-purity nickel. The change in weight before and after melting was so small as to be 0.7% as well as cobalt. Figure 6 shows the sampling locations for chemical analysis in a diametrical vertical section of the nickel ingot, and the analytical results of gaseous impurities and the other impurities are shown in Tables 4 and 5, respectively. The vacuum before melting was 1: Pa and the pressure during melting after melt-down to surface solidification was kept from 1: to 4: Pa for 14.5 min. The intensity of the oxygen spectrum increased from the order of A to that of 10 9 A during heating and melting. It should be emphasized that such a marked increase in the intensity of oxygen spectrum has seldom seen during melting of the other metals such as iron 17) and titanium, as shown below. This result consists with that of chemical analysis shown in Table 4; that is, the oxygen concentration was reduced from 49 mass ppm in starting nickel to 3 8 mass ppm in the ingot. The oxygen content of the sample B is higher than those of sample L and R, because location B is in the scar layer. It is thus out of the discussion. Judging from the standard free energy of formation of NiO (Ellingham diagram), nickel has not so strong affinity for oxygen as titanium in the present vacuum degree of Pa during heating and melting. The above results is thus reasonable. Carbon content decreased a little bit. Nitrogen content was not reduced, but it is enough low to be below 1 mass ppm. The main conclusion is that CCIM in the present UHV degree and quality is very effective for the purification of nickel due to the removal of oxygen. It is found from Table 5 that impurities such as Al, As, Ca, Mn, Pb and Zn decrease in concentration by CCIM in UHV. It is reasonable because melting points of these elements are lower than that of nickel and their vapour pressure is higher than that of nickel. Also Table 5 shows that most of impurities analyzed are lower than each detection limit of analysis. 3.3 Purification of titanium Titanium is known to have a very strong affinity for oxygen and to be not deoxidized in UHV of Pa during heating and melting on the basis of standard free energy of formation of TiO 2. The present main subject is thus

4 Purification of Cobalt, Nickel, and Titanium by Cold-Crucible Induction Melting in Ultrahigh Vacuum 159 Pressure, p / Pa power on Total pressure melt-down melt-beginning Mass spectrum power off power down Time, t / sec Mass No Fig. 5 Changes in the total pressure and the mass spectra of gases of mass numbers 2, 18, 28, 32 and 44 during induction heating and melting of highpurity nickel (5N). L T B Fig. 6 Sampling locations for chemical analysis in a diametrical vertical section of the nickel ingot (5N) prepared by cold-crucible melting. Table 4 Concentration of gaseous impurities in starting nickel (5N) and the purified nickel ingot melted in UHV in cold-crucible furnace (mass ppm). not purification of titanium but to make the point clear whether by the CCIM even in the present UHV degree and quality the concentration of oxygen in high-purity titanium increases or not. High-purity titanium 1.6 kg in weight was melted using the small CCIM furnace. Figure 7 shows the changes in the total pressure and the mass spectra of gases of mass numbers 2, 18, 28, 32 and 44 during induction heating and melting of highpurity titanium. The vacuum before melting was 1: Pa and the pressure during melting after melt-beginning to surface solidification was kept from 2: to 5: Pa for 13.3 min. The mass spectra show that hydrogen is the main spectrum during heating and melting which determines the total pressure. The other spectra are two orders of magnitude lower than hydrogen spectrum, and 32 R C N O S Starting Ni (5N) T Ingot L R B Ion Current, i / A Table 5 Concentration of impurities except gaseous impurities in starting nickel (5N) and the purified nickel ingot melted in UHV in cold-crucible furnace (mass ppm). Ingot Starting Ni (5N) L T R Ag <0:01 <0:01 <0:01 Al <0: As 0.25 <0:01 <0:01 <0:01 Au <0:01 <0:01 <0:01 Bi <0:005 <0:005 <0:005 Ca <0:1 <0:01 <0:01 <0:01 Cd <0:01 <0:01 <0:01 Cl Co <0:1 <0:001 <0:001 <0:001 Cr <0: Cu Fe Ga <0:01 <0:01 <0:01 Ge <0:05 <0:05 <0:05 Hf <0:01 <0:01 <0:01 In <0:05 <0:05 <0:05 Ir <0:01 <0:01 <0:01 K <0:01 <0:01 <0:01 Li <0:001 <0:001 <0:001 Mg <0:005 <0:005 <0:005 Mn <0:1 <0:001 <0:001 <0:001 Mo <0:01 <0:01 <0:01 Na Nb <0:01 <0:01 <0:01 P <0:01 <0:01 <0:01 Pb 0.06 <0:005 <0:005 <0:005 Pd <0:01 <0:01 <0:01 Pt <0:05 <0:05 <0:05 Re <0:01 <0:01 <0:01 Rh <0:01 <0:01 <0:01 Ru <0:01 <0:01 <0:01 Sb <0:01 <0:01 <0:01 Si <0:01 <0:005 <0:005 Sn <0:01 <0:01 <0:01 Ta <1 <1 <1 Ti V <0:005 <0:005 <0:005 W <0:01 <0:01 <0:01 Zn 1.5 <0:01 <0:01 <0:01 Zr <0:01 <0:01 <0:01 oxygen spectrum is further two orders of magnitude lower than them. The change in weight before and after melting was also so small as to be 0.6%. Figure 8 shows the sampling locations for chemical analysis in a diametrical vertical section of the titanium ingot. The analytical results of gaseous impurities and the other impurities in the high-purity titanium ingot made by cold-crucible induction melting in UHV are shown in Tables 6 and 7, respectively. Titanium is so active that the analysis of gaseous impurities is very difficult and the detection limits are high. Hydrogen, carbon and nitrogen are lower than each detection limit, 10 mass ppm. The oxygen content in the ingot is almost the

5 160 S. Takaki and K. Abiko power on Total pressure melt-down power down melt-beginning power off Table 6 Concentration of gaseous impurities in starting titanium (5N) and the purified titanium ingot melted in UHV in cold-crucible furnace (mass ppm). H C N O O Pressure, p / Pa Mass spectrum Mass No Ion Current, i / A Starting Ti (5N) <10 <10 < T <10 <10 <10 30 Ingot M B 40 Melting condition CCIM EBM Pa Pa 0 same as that of the starting titanium, taking consideration of the accuracy of oxygen analysis in titanium. On the other hand, when the same starting titanium was melted by electron-beam melting (EBM) in the lower vacuum degree of Pa, the oxygen concentration of the titanium ingot obtained increased from 30 to 80 mass ppm by 50 mass ppm. It is concluded from these results that highpurity titanium can be melted without any increase in oxygen in the present conditions of CCIM in UHV, although titanium has a very strong affinity for oxygen. Also it is found from Table 7 that impurities such as Al, Cl, Na, S and so on decrease in concentration by CCIM in UHV. It is reasonable because melting points of these elements are lower than that of titanium and their vapour pressure is higher than that of titanium. 4. Conclusion Time, t / sec Fig. 7 Changes in the total pressure and the mass spectra of gases of mass numbers 2, 18, 28, 32 and 44 during induction heating and melting of highpurity titanium (5N). GDMS T M B Fig. 8 Sampling locations for chemical analysis and GDMS in a diametrical vertical section of the titanium ingot (5N) prepared by coldcrucible melting. Cold-crucible induction melting in UHV of Pa is very effective for the purification of 9.2 kg cobalt and 3.0 kg nickel, in particular for decreasing oxygen concentration; for example, for the starting material and the ingot melted by CCIM of cobalt (3N) from 19 to 6 mass ppm, of Table 7 Concentration of impurities except gaseous impurities in starting titanium (5N) and the purified titanium ingot melted in UHV in coldcrucible furnace (mass ppm). Elements Starting Ti (5N) Ingot Al B <0:005 <0:005 Be <0:01 <0:01 Ca <0:02 <0:02 Cl 0.06 <0:01 Co <0:005 <0:005 Cr Cu F <0:05 <0:05 Fe Ga <0:05 <0:05 K <0:01 <0:01 Li <0:005 <0:005 Mg <0:01 <0:01 Mn <0:005 <0:005 Na 0.06 <0:01 Ni <0:01 <0:01 P <0:01 <0:01 S Sc <0:005 <0:005 Si V Zn <0:05 <0:05 Zr cobalt (5N) from 6 to 2 mass ppm, and of nickel (5N) from 49 to 6 mass ppm, respectively. High-purity titanium can be melted without any increase in oxygen in the present conditions of CCIM in UHV, although titanium has a very strong affinity for oxygen. For further purification of cobalt, nickel, and titanium due to the removal of gaseous impurities it is absolutely necessary to develop and establish the analysis technique of accurately measuring the concentration of gaseous impurity element at levels below 0.1, 1 and 10 mass ppm, respectively. Acknowledgements The authors wish to thank Nikko Materials Co. for supplying high-purity cobalt, nickel, and titanium and also for GDMS analysis. We are grateful to Dr. K. Takada and the members of the analytical group in our institute for analysis

6 Purification of Cobalt, Nickel, and Titanium by Cold-Crucible Induction Melting in Ultrahigh Vacuum 161 of trace amounts of gaseous elements in high-purity cobalt, nickel, and titanium, and also to Messrs. N. Harima, G. Kanou, Ch. Kawarada, N. Suda, and Y. Sasada for help in the present experiments. REFERENCES 1) Ultra High Purity Base Metals, Proc. 1st. Int. Conf. on Ultra High Purity Base Metals (UHPM-94), ed. by K. Abiko, K. Hirokawa and S. Takaki, (The Japan Institute of Metals, Sendai, 1995). 2) Purification Process and Characterization of Ultra High Purity Metals, ed. by Y. Waseda and M. Isshiki, (Springer-Verlag, Berlin Heidelberg, 2002). 3) B. Dubois, F. Rocquet, M. Nardin and V. Q. Kinh: Mem. Sci. Rev. Met. 66 (1969) ) M. Isshiki, T. Kikuchi and K. Igaki: Trans. JIM 25 (1984) ) M. Isshiki, Y. Fukuda and K. Igaki: J. Less-Common Met. 105 (1985) ) M. Isshiki, Y. Fukuda and K. Igaki: Trans. JIM 27 (1986) ) V. J. Albano and R. R. Soden: J. Electrochem. Soc. 113 (1966) ) R. R. Soden and V. J. Albano: J. Electrochem. Soc. 113 (1966) ) L. Kuchar, N. Deschamps, A. M. Wache, B. Dubois and O. Dimitrov: Mem. Sci. Rev. Met. 64 (1967) ) M. Nardin, A. M. Wache and V. Q. Kinh: Mem. Sci. Rev. Met. 66 (1969) ) Ti 2003 Science and Technology, Proc. 10th World Conf. on Titanium, ed. by G. Lutjering and J. Albrecht, (Wiley-VCH Verlag, 2003). 12) C. Ouchi, H. Iizumi and S. Mitao: Mater. Sci. Eng. A 243 (1998) ) K. Abiko, T. Nakajima, N. Harima and S. Takaki: Phys. Status Solidi A 167 (1998) ) Special Issue on Ultra-High Purity Metals, Mater. Trans. 41 (2000) No ) Proc. 7th Int. Conf. on Ultra High Purity Base Metals (UHPM-2000), ed. by A. Seeger and A. Korhonen, Phys. Status Solidi A 189 (2002) No ) Special Issue on Ultra-High Purity Metals (II), Mater. Trans. 43 (2002) No ) S. Takaki and K. Abiko: Meter. Trans., JIM 41 (2000) ) K. Abiko and Y. Kato: Phys. Status Solidi A 167 (1998) ) G. Kanou, N. Harima, S. Takaki and K. Abiko: Mater. Trans. 41 (2000) ) T. Nakajima, Y. Morimoto, S. Takaki and K. Abiko: Mater. Trans. 41 (2000) ) T. Nakajima, Y. Morimoto, S. Takaki and K. Abiko: Phys. Status Solidi A 167 (1998)