PRIMARY MELTING OF TITANIUM WITH PLASMA ARC
Tatsuo Fujiwara, Koshi Kato, Kiyoo Ono and Hiroyuki Yamada Daido Steel Co., Ltd. Japan
Introduction
Plasma arc has a good characteristic as a heat source for melting metals and alloys. That is, its temperature is higher than that of ordinary electric arc and its power can be controll ed very easily. When an inert gas like argon is used as. a plasma forming gas, very clean heat, or non-reactive heat, is obtained. Furthermore, the contamination by plasma torch may be substan tially none. Accordingly, plasma torch is thought to be a very useful nonconsumable electrode.
Recently, the use of plasma arc for melting metals and alloys has been well studied. In USSR and GDR two types of plasma arc furnaces have been developed and working on an in dustrial scale. One is a plasma arc furnace in which the carbon electrodes and its construction were, as a whole, similar to those of the conventional arc furnace (1,2). And the other is the plasma arc remelting (PAR) furnace (3,4).
In the meantime, in Japan, our company has been doing the research on plasma arc melting of steels and superalloys for more than 10 years, and has already developed an industrial Plasma Induction Furnace (5,6), which is the combination of plasma arc and induction heating. We also have studied the use of plasma arc for melting titanium. This may be thought to be one of the most promising fields in which the characteristics of a plasma arc -high temperature, easily controllable and clean heat source would be made the best use of.
At present, conventional double vacuum arc melting (VAR) process for titanium has such problems that, as generally known, it is necessary to make a primary consumable electrode from titanium sponge, .but it is a very troublesome process and recy cling of various shaped titanium scrap is difficult.
Concerning titanium melting with plasma arc we made a presentation of the results on our own PPC-flasma frogressive Casting-furnace at the 6th International Vacuum Metallurgy Con ference held in San Diego, April 23-27, 1979. In this we de scribed the test results obtained from an examination of avail ability of the PPC process for melting of titanium, and it was concluded that it was more economical to utilize the PPC as the primary melting process, that is, the double melting consisting of the primary melting with PPC and the secondary melting with VAR was particularly effective for recycling of scrap materials. 2136 T. Fujiwara et al.
After this, we fully made an examination of the i'PC primary melt ing conditions from the qualitative and economical point of view.
Description of Plasma Progressive Casting Furnace
Plasma arc furnace designed and constructed for this experi ment is shown in Fig.l. We call this furnace PPC-Elasma Erogres sive Casting-furnace. A transfer type plasma torch was set ver tically just above the mold. The section of the plasma torch is shown in Fig.2. The W-Th cathode is attached at the end of water cooled copper pipe and the cathode is encircled by the water cooled copper nozzle. Argon gas, which we used as a plasma form ing gas, is flowed between these, and a part of argon gas is ionized by the arc and pinched at the nozzle, so that the high energy density plasma arc is gained. The anode was the molten metal in the mold.
The water cooled copper mold llOmm in dia. shown in Fig.l installed a stirring coil and was combined with a movable bottom plate, which was withdrawn progressively and proportionally to the melt rate of the melting materials from the hopper so that the pool surface might be kept at the settled position.
Before starting the melting, the furnace chamber was evac uated to the reduced pressure of the order of 10-3 Torr and then argon gas of five-nine purity was supplied to latm. Then the plasma arc was ignited. The argon gas from the torch was flowed out through the leak valve of the furnace. So melting was done under argon atmosphere of latm.
The power source of this furnace had a maximum capacity of o.c. 150KW and the ingot obtained was maximum 25Kg in weight.
Photo.l shows the upper and lower view of the PPC furnace. It can be seen the plasma torch and its moving mechanism, the hopper, the viewing port and the melting chamber in the upper view, the mold, of which jacket for water cooling is only seen, and the ingot cooling chamber in the lower view. This PPC furnace was used for studying the plasma arc remelting of steels and superalloys, too (7).
PPC Primary Melting Primary meltings were done under the various melting con ditions -plasma power, feeding rate and type of melting materials using titanium sponge or scraps and Ti-6Al-4V alloy scrap. The chemical compositions of these melting materials are shown in Table 1. One of the requirements to the PPC primary melted ingots is that they can be safely used as the electrodes for the VAR. This requirement is satisfied by proper selection of the melting conditions. We adopted the apparent densities of the primary melted ingots as the parameter of the soundness. PRIMARY MELTING BY PLASMA ARC 2137
Table 1 Chemical compositions of melting materials used for this experiments (ppm) Melting Material 0 N c Fe H Mg Cl Ti sponge(JIS Grade1) 500 50 100 200 20 800 400 Ti sponge(JIS Grade 2) 700 70 100 300 30 800 400 Ti cut wire scrap 900 80 120 500 13 - - Ti cut plate scrap 1070 80 120 300 15 Al(°lo) V(°lo) Ti-6Al-4V cut wire scrap 1500 150 200 1000 50 6.25 4.10
Fig.3 indicates the results of the examination, in which the dependency of the apparent densities of the PPC primary melted ingots on the plasma power and the feeding rate of the melting materials is shown. At the same power level, the inverse rela tion between the apparent densities and the feeding rates was seen. And also at the high plasma power the feeding rate was made higher to obtain the same density. Photo.2 shows the surface appearances of the PPC primary ingots melted from the various melting materials. The reasons why the primary melted ingots had the lower apparent densities than the true density of 4.5gr/cm3 were that the fed materials had not been completely melted, as seen in Photo.2, and that there had been blow holes in the interior of the ingots (Photo.3)
But from Photo.3 their interior quality was proved to be fairly better than that expected from the ingot surfaces. Therefore, they were thought to be satisfactory as the electrodes for the subsequent VAR. However, for establishing the proper conditions for PPC primary melting it is necessary to determine the lower limit of the apparent density usable as the electrode .for VAR. So, we examined the density distributions on the horizontal sections of the primary ingots with various apparent densities. The results of the examination are shown in Fig.4. In this figure, A and B were melted at the high and low melt rate at 90KW, respectively, and c were at the high melt rate at llOKW. And the base planes of the figures correspond to the density of 4.0gr/cml.
comparing these, it is clearly seen that the lower the melt rate and the higher the plasma power, the higher the density of the ingot became. Also there could be seen the tendency that when the melt rate was high, the density of the outer side of the ingot became lower, even though the plasma power had been high.
From these examination, it was confirmed that the PPC pri mary melted ingots which had the apparent densities over 3.5 gr/crri1 had satisfactory qualities and could be safely used as 2138 T. Fujiwara et al. the electrodes for VAR. All of those shown in Fig.3 had the apparent densities over 3.Sgr/cm3 and so they can be used as the electrodes for VAR. In the meantime, the melting efficiency is also a matter of concern. Fig.S shows the melting efficiencies of PPC primary melting under the various melting conditions. It is seen that the melting efficiencies were dependent on the feeding rate at all plasma power level tested, and that the high efficiencies could be obtained under the higher power conditions, because the feeding rate could be increased more than that under the lower power conditions without any decrease in the apparent densities. Here the typical melting conditions and the apparent dens ities of the ingots are shown in Table 2, in which titanium sponge only, mix of titanium sponge and scrap and Ti-6Al-4V alloy scrap only were respectively used as the melting materials. And Table 3 shows the results of the chemical analysis of the PPC primary ingots obtained from titanium sponge only, as well as those of sponge.
Table 2 Melting conditions at PPC primary melting and the apparent densities of the ingots obtained
Heat PPC Condition Melting Ingot Appar. Mel ting Material Efficiency Density No. Power(KW) F~~i~pte (Kg/KWH) (gr/cm3) 125 Ti, sponge (JIS Grade 1) 110 1.16 0.54 3.8 713 Ti, sponge (JIS Grade 2) 110 0.96 0.43 3.9 128 Ti, cut wire scrap(5)+sponge(1) 110 1. 17 0.53 3.8 118 Ti, cut plate scrap(S)+sponge(l) 110 1. 20 057 4.0 120 Ti, cut plate scrap(3)+sponge(1) 110 1.36 0.60 3.8 129 Ti, cut wire scrap(3)+sponge(2) 110 1.47 0.67 3. 7 119 Ti, cut plate scrap(l)+sponge(l) 110 1. 14 0.55 3.9 150 Ti-6Al-4V. cut wire scrap 100 0.95 0.48 4.0
Table 3 Change of chemical composition by PPC primary melting (ppm) 0 N c Fe H Mg Cl Ti sponge (JIS Gradel) 500 50 100 200 20 800 400 After PPC 550 50 80 210 30 <30 <50 Ti sponge(JIS Grade 2) 700 70 100 300 30 800 400 After PPC 740 60 100 320 25 <30 <50 PRIMARY MELTING BY PLASMA ARC 2139
From this Table it was seen that by PPC melting magnesium and clorine contained in titanium sponge were fully removed, and that the content of hydrogen was unchanged. Therefore, it is seen that the PPC primary melted ingots can be safely remelted with an usual VAF used for the remelting of steels or super alloys, because no release of magnesium and clorine occur during VAR.
VAR.of PPC Primary Melted Ingots VAR of the PPC primary melted ingot was done with the 160mm dia. mold. During the VAR, it was not perceived remark able release of gases from the electrodes, and the vacuum was kept nearly constant. The PPC + VAR double melted ingots seen in Photo.4 had no detrimental surface defects and almost the same surface appearances as those of the conventional double VAR ingots.
Quality Of PPC + VAR Ti, Ti-6Al-4V Alloy
The mechanical properties and the chemical compositions of the PPC + VAR double melted Ti and Ti-6Al-4V alloy are shown in Table 4. All tension tests were done with the specimens ma chined from the 20mm dia. annealed bars directly forged from the ingots. The chemical compositions of the melting materials were already shown in Table 1.
Table 4 Chemical compositions and mechanical properties of PPC + VAR Ti and Ti-6Al-4V bars forged directly from ingots
Chemical Composition (ppm) Mechanical Properties Heat Melting Material No. 0 N Fe H ~·v~ 1.::. HB c ·~ c·T.> ~/~ P2406 Ti, sponge(JIS Grade 1) 500 60 30 200 11 20.7 31.5 64.0 82.2 107 P9102 Ti,sponge(JIS Grade2) 910 120 90 400 9 32.8 41.6 33.9 69.3 119
PD108 Ti, wire scrap(5)+sponge(1) 600 90 60 400 9 24.4 34.3 54.5 78.7 109 P2401 Ti, plate scrap(5).sponge(1) 1080 50 90 700 12 409 50.0 29.7 56.5 141 P2403 Ti, plate scrap(3)osponge(2) 1050 60 130 1000 11 41.8 50.9 33.8 58.8 142 PD109 Ti, wire scrap(3).sponge(1) 1040 70 30 400 13 32.3 42.1 36.4 68.1 130 P2402 Ti, plate scrap(1)osponge(1) 1030 80 110 600 16 38.7 49.4 33.7 580 140 P3401 Ti-6Al-4V, cut wiresaap 1900 157 250 1100 10 105.4 111.4 15.7 37.6 311 2 C, P. Ti Grade 1 ~1500 "S500 - :::2CXXJ '.::150 - !'42 ?:27 - ?:100 J IS 35 C, P, Ti Grade 2 4£2000 -=500 - '.::20CI '.::150 - - 52 ?:23 - '.:-110 * In the annealed state ** Number in parenthesis represents mixing ratio *** Al=6.20%, V=4.15% 2140 T. Fujiwara et al.
Heats melted from such titanium sponge that would be gener ally used for the production of grade 1 and grade 2 titanium ingots in JIS showed that they satisfied the requirements of each aimed standard. Heats melted from C.P.Ti scraps showed somewhat high oxygen and iron contents. This is because, as shown in Table 1, their contents in the scraps used were rela tively high. Hydrogen was fully removed after VAR. Conse quently, all heats melted from C.P.Ti scraps showed the satis factory properties required for grade 2 titanium in JIS.
The changes of alminum and vanadium in Ti-6Al-4V alloy were scarcely recognized. And the heat melted from Ti-6Al-4V alloy scrap, had the satisfactory mechanical properties as the alloy with such compositions.
In the meantime, the distribution of the impurity elements in the PPC + VAR double melted ingot from titanium sponge seen in Table 5 was uniform and no remarkable segregation of each element could be seen. Only the amount of oxygen was somewhat high in the bottom position.
Table 5 Distribution of impurity elements in a PPC + VAR c.P.Ti ingot .
lmwrity •lements (%) Position Fe c N Ma 0 H surfoc• 0.02, o.om 0.006 <0.003 o.°'' 0.0005 1,-radius 0.032 0.005 0.004 <0.003 0. 041 00005 Top C•nfH 0.032 0.003 0.005 <0.003 0.00 0.0005 fradius 0.032 0.005 0.005 <0.003 0.0~ 0.0005 surtac. 0.0~ 0.003 0.005 <0.003 0.04~ 0.0001 surface O.OJS 0.005 0.008 <0.003 0.074 0.fllW1 irodius o.~ 0.003 0.006 <0.003 O.Oli3 0.0008 Mid CMIH 0.02'J 0.00, 0.005 <0.003 0. 06.§ 0.0001 yrodius 0.025 0.004 0.006 <0.003 0.078 0.0008 surface 0.038 0.004 0.007 <0.003 0.0:111 0.0011 surfac• 0.046 0.004 0.009 '<0.003 o on 0.0008 frodius 0.037 0005 0.009 <0.003 O.nAl 0.0008 Bot cent"' o.°'o 0.004 0009 <0.003 005' 0.0007 tradius 0.035 0.005 0.009 Fig.6 also shows the distribution of hardness. It can be seen that at the each position of the same height the Brinell hardness were almost the same, but that at the bottom position the hardness were a bit higher, which might correspond to the distribution of oxygen. Finally, it was examined whether the contamination by copper or tungsten from the plasma torch would occur or not. The results by the mass spectrum analysis is shown in Table 6. It is seen from Table 6 that the amount of tungsten in the PPC or PPC + VAR ingots were both below 2ppm as well as that of in the conventional double VAR ingot and that the amount of copper was nearly same. It was proved that there would occur no contamination by the plasma torch during the melting. PRIMARY MELTING BY PLASMA ARC 2141 Table 6 Results of mass spectrum analysis of PPC only, PPC + VAR double and double VAR melted titanium ingots (ppm) w Cu Ca Mn Zn Fe As Sn Sb Pb 63 78 -'3 90 590 2.0 160 5.1 3. 7 PPC <2 I I I I I I I I I ingots 110 28 83 98 950 330 -'.2 7. 8 PPC+VAJ; 82 12 38 85 600 "·'2.9 190 5.8"' <2 I I I I I I I I I ingots 85 33 76 90 900 8.7 -'90 16 "· 7 53 10 29 90 -'70 -'.6 360 16 1.0 VAR+ VAR <2 I I I I I I I I I ingots 1-'0 17 "7 110 700 7.2 -'30 22 -'.2 Summary Primary melting of titanium with plasma arc was investi gated using an experimentally designed and constructed Plasma Progressive Casting (PPC) furnace using a llOmm dia. water cooled mold and a power source of maximum n.c. 150KW. The apparent densities and the soundness of the PPC ingots melted under various melting conditions -plasma power, feeding rate and type of melting materials- were examined. These PPC ingots were vacuum arc remelted. It was found that the PPC ingots having the apparent densities over 3.5gr/cm3 were safely used for VAR electrodes. The PPC + VAR double melted ingots had almost the same surface appearances and qualities as those of the ingots obtained by the conventional VAR double melting. PPC primary melted ingots with the appar ent densities over 3.5gr/cm3 could be obtained at the melting efficiency of 0.5 ~o. 7Kg/KWH. This figure may be lowered by the scale up of the melting capacity. Consequently, the primary melting with PPC was thought to be very flexible and promising process for recycling of scrap materials. I 4. E. F. Marley Jr. and J. R. Wamsley: Metal Progress, 19(1976)4, 46. s. c. Asada, I. Eguchi and T. Adachi: 7th U.I.E., Warsaw ( 1972) I 123. 6. H. Tezuka, s. Sugiura and T. Sugiyama: Proceed., 4th ICVM, Tokyo (1974)2, 142. 7. T. Fujiwara, K. Kato, K. Ono and H. Yamada: Trans. ISIJ, 19(1979)1, 48. · "Plasma TorchB TV.Camera ' ' -,_~ Viewing Port Ingot Cooling Chamber . Fig.l Scheme of PFC (Plasma Progressive Casting) furnace Photo.l View of Plasma Progressive Casting (PFC) furnace PRIMARY MELTING BY PLASMA ARC 2143 M«ting Matwials ~Ti only -···--·-___..., Path of argon gas Cooling water r Cut Wire Ti only \ Fig.2 Shceme of Plasma Arc Photo.2 Surface appea Torch rances of PPC primary melted Ti ingots from various melting materials . ... ·. {:. . . '·_. . . ' .. . / ' ... : ... f" • .••'- ~ ••• (.. c (A} 4.14 (8) 3. 65 (CJ 3. 92 Photo.3 Sections of PPC primary melted Ti ingots having various apparent densities (gr/cmJ) 2144 T. Fujiwara et al. Melt. Mat...... ~4.5 i} Sponge Ti ~L.; O'I 0 Sponge Ti 90KW ' 0 +ScrapTi -~ 75KW \ ' ·~ "in c ~ ~ . ··~ ~4.0... A 'tA ~~ ~ 0 ~ \·~A All ·, 61. . 8: \ o 'o <{ A . ' , 3.5 ,.... 110KW Fig.3 Dependency of apparent densities of PPC primary melted ingots on the feeding rate of melting material and plasma power 08 ,.. Power(KW) Melt.Mat. x 75 } -J: • 90 SpongeTi ~07 "'" A 110 0 11 0 Sponge 1:i +Scrap Ti ~ ,.... oy -0.6 00 > u A~ c .SI! 0.5 ..... u ,y..A ..... :;:: A -wo.4 .... g' ~0.3 ~ /'x 02. "'" I I I - 05 1.0 1.5 Feeding Rate (Kg/min) Fig.5 Dependency of melting efficiency of PPC primary melting on the feeding rate of melting material and plasma power PRIMARY MEL TING BY PLASMA ARC 2145 (A) (8) (CJ x (AJ (8) (CJ Density i (gr/cm') ,.25 3.98 ,.05 Plasma Power(kWJ 90 90 110 Melt Rote (gr/min) 586 930 880 Fig.4 Density distribution in the PPC primary melted Ti ingots from sponge Ti 120 110 100 90...._~-~~-~~-~-~~ 1201 ::~--0...... -~"-----o~----~o--~o-_..,.,..._o~--~o--~o -.. 120 o..-o__ Mid 0 __ __ 0_0 ___ o-o ~ 110 0 ...... 100 ~ 90·'--~-~-'--~-~-'--~-~ 0 ::C ...... _ Ill 120~ o-o....._ _....o_ 0 ,,,_o :~-~-~~-""~o_.,-_.__~-...... -~o-"'_~ 120 0 ...... Bot 110 o__ o,,,..,,--0...... _o __ o_o_o 100 9·v~-ou"'-r-~-~-~ce-nTter~~-~-~out Photo.4 Surface appearances Fig.6 Distribution of hard of PPC + VAR double ness in a PPC + VAR melted Ti ingots c.P.Ti ingot