An induction cold crucible melting is one of the most promising technology for the reuse and decontamination of the radioactively contaminated metallic materials yielded through the dismantling of nuclear facilities, because the crucible ensures a long life operation without generating secondary wastes.

Based on the knowledge obtained through the feasibility study carried out in 1995, the MERC (Melting and Recycling of Metals by Cold Crucible) process was designed, manufactured and scaled up from 45mm to 100mm in crucible diameter. The maximum power of the high frequency generator is 150kW and the frequency is 25kHz.

In the MERC, continuously supplied fragments of stainless steels, which were the surrogates of contaminated metallic materials, were successfully melted, followed by solidified and pulled down by the withdrawal velocity of 3mm/min. The withdrawn length and the diameter of the ingot were 200mm and 100mm, respectively. The surface of the ingot was crack free. Neither macro segregation nor non-melted metallic materials inside the billet were observed. The melting characteristics, such as heat efficiency, dome height control method and removal of slag, were obtained.

This work was done under the sponsorship of Science and Technology Agency of Japan.

Melting, decontamination, formation to the suitable shape for the reuse and immobilization of radioactive nuclides are the important key technologies for the reuse of dismantled metals contaminated by low level radioactivity which are yielded from the nuclear facilities. There are several processes offering melting and immobilization, such as plasma melting and induction hot crucible melting(1,2,3). However, there are few processes in which melting, decontamination, formation and immobilization proceed simultaneously.

An induction cold crucible technology has peculiar characteristics that the dismantled metals experience melting and solidification in the same crucible consecutively and are withdrawn continuously. This enables the melting, immobilization and formation of the metals to be reused by selecting the suitable sectional shape of the cold crucible. Continuous melting and withdrawal ensure sound enough quality for the reuse of the resultant ingot compared to the batch treatment.

Another characteristic of the cold crucible is that the crucible wall is harm free against both high temperature metal and the slag in contact. Oxidation of the radioactive nuclides by the slag, followed by the transfer of the nuclides from the metal to the slag is promoted by the vigorous electromagnetic agitation of the molten metal. This is also another characteristic of the cold crucible. These features result in the maintenance free of the crucible wall and the progression of decontamination during melting.

The induction cold crucible has long been used for the melting and casting of chemically active and high purity materials, such as titanium alloy and polycrystalline silicon ingot for photovoltaic solar cell (4,5,6,7). In 1995, fundamental technologies concerning the MERC(Melting and Recycling of metals by Cold crucible)process have been established based on the conventional induction cold crucible technology by using the cylindrical crucible of 45mm in diameter(8). As a result, it was confirmed that the MERC could be applied to the continuous process of melting and solidifying the small sectional materials, such as piping and tubing pieces. Decontamination was clarified by adding surrogates of the radioactive nuclides and the necessary data for the melting and solidifying in the successive scale-up of the MERC were obtained by the experiment as well as theoretical study.

In succession to the fundamental study, the testing MERC process was designed, manufactured and scaled up from 45mm to 100mm in crucible diameter. The maximum power of the high frequency generator is 150kW and the frequency is 25kHz. Small sized fragments of stainless steel were treated by the newly developed testing MERC. Investigated items of the present study are the followings.

• Test of melting and solidifying the surrogates of dismantled metals by the scaled up MERC
• Quality of the ingot on the surface as well as inside
• Removal of flux stuck on the ingot surface
• Pollution of the ingot by the crucible

A schematic view of the testing MERC apparatus is shown in Fig. 1. A water-cooled and segmented crucible made of copper is surrounded with a multi-turn induction coil. After the primary, which is a base metal for the melting, was inserted from the bottom opening to the suitable level for the melting in the crucible where the magnetic field is most effective, the atmosphere in the chamber is replaced with Ar gas. With an increase in power input from the coil through the cold wall permeable of electromagnetic energy to the primary, it melts by Joule heating to rise in a dome shape by the electromagnetic force. The molten metal experiences electromagnetic stirring and the top of the melt is out of contact with the cold wall, resulting in the increased heat efficiency.

After the primary was melted, metallic materials are supplied from the crucible top opening onto the melt dome by the vibratory feeder. Some amount of the decontaminating flux is also supplied in the same manner. Precise velocity control concerning the supply of the metallic materials and the withdrawal of the ingot preserves the dome height at a certain fixed level during continuous melting and successive solidification. As a result, an ingot, which is the melted and successively solidified material, is obtained.

Specification of the testing MERC apparatus is shown in TableI, together with the previous fundamental testing apparatus. The electric power source, in which the high frequency electric current is generated by the power transistor, is 25kHz in frequency and 150kW in the maximum output. The internal diameter of the crucible is 100mm, corresponding to two times as large as that of the previous experimental apparatus. The maximum withdrawal length of the ingot is restricted within 300mm. This does not necessarily mean that the continuous treatment by the MERC is restricted within the amount of materials corresponding to the 300mm long ingot. In the future, the ingot is scheduled to be cut every 300mm length during withdrawal. Accordingly, the amount of treatment has no relation to the ingot length.

Typical experimental condition is shown in Table II. Fragments of stainless steel, dimension of which are 5mm in diameter and 10mm in length, are supplied from the top opening of the crucible onto the melt dome. The fragments are the representatives of piping or tubing generated during the dismantling of nuclear facilities. Withdrawal velocity of 3mm/min corresponds to the melting rate of 12kg/hr of the fragments.

Major test parameters are the dome height and the electric power. Here, the dome height is defined as the top of the cylindrical primary before melting measured from the coil top. It is equal to that measured from the crucible top, since both top levels of the coil and the crucible are arranged geometrically to coincide each other during all the tests. Sometimes, some amount of flux and tungsten powder, which are the oxidation agent for the decontamination and the tracer of identifying of solidification front, respectively, were added onto the melt dome. The flux is an admixture of SiO₂, Al₂O₃, CaO and MgO. In the present test, neither radioactive nuclides nor surrogates of them were added.

Generally, after the electric power reached some fixed value of the test, 130kW, the primary melted in a few minutes, regardless of the dome height from –10 to –70mm and the presence of the flux. Continuously supplied metallic materials floats on the dome surface on account of the surface tension and falls down into the narrow gap between the periphery of the dome and the cold wall, followed by melting by Joule heating. Non of the non-melted residuals of the supplied materials were observed on the dome surface.

Molten metal dome elevated by the electromagnetic pinch force was not necessarily stable. The surface of the melt experienced bending motion toward the cold wall or rotated around the symmetry axis. The turbulent flow of the molten metal coupled with the alternating electromagnetic filed is responsible for the appearance of unstable shape of the dome. However, Infiltration of the molten metal into the narrow slit gap did not take place because of the precise manufacturing technology of the slit gap less than 0.2mm. The instability seems to be decreased with the decrease of power input and the addition of the flux.

Consequently, supplied materials immediately and successfully melt in a crucible surrounded with the water-cold wall and are withdrawn continuously.

The whole of the primary top has not melted until the electric power increased beyond 110-120kW. The figure is almost in agreement with that predicted previously by the theoretical model, taking account of the coupling of the shape of the molten metal, electromagnetic stirring and heat transfer (8). Figure 2 shows the comparison of the solid-liquid interface obtained by the etching after the addition of tungsten powder just before power down with that of calculation. Withdrawal velocity is zero for both cases. The experimental and the calculation condition are almost the same except that the coil height and input power in the experiment are 100mm and 130kW, while those in the calculation are 65mm and 118kW, respectively. Dark part in Fig. 2a shows the melted part by Joule heating. Meanwhile, left and right half figures in Fig2b show the heat and flow patterns, respectively. Those area temperature higher than 1616K and velocity field larger than zero correspond to the molten metal. In both experimental and calculated cases, melted part proved to correspond to that area higher than the coil bottom at the input energy of about 130kW and the ingot diameter of 100mm. The validity of the conceptional design carried out previously concerning the input energy for the melting was confirmed from the experiment.

With the decrease in power, solidification did not take place even at about 60kw lower than the electric power when the melting begins. This means the presence of large hysteresis on the electric power corresponding to the beginning of melting and solidification. The difference in the shape of the primary between the increase and the decrease of the electric power seems to be responsible for the hysteresis. If the optimal shape of the primary was found, smaller electric power would be able to melt larger primary.

The energy balance of the MERC was evaluated from the increase in the water temperature and the water flow rate. Supply rate of the fragments of stainless steel was used for the evaluation of the energy transferred to the ingot. The energy balance among the crucible, ingot, coil including the coaxial bus tube, water-cooled electric cable and chamber is shown in Table III. Here, the energy consumed in the crucible does not include that of transferred from the ingot.

The energy ratio consumed in the ingot, 20%, would increase by the enlargement of the crucible size, since the specific surface area responsible for the heat loss per unit volume of the ingot decreases with the increase of the ingot size.

Quality of the ingot plays an important role in the reuse of dismantled metals. This is not only due to the requirement from the manufacturing of materials but also from the decontamination of the cast ingot. It would be difficult to remove the slag from the ingot if the slag infiltrates into the crack gaps that are made during withdrawal. Selection of the optimal casting condition concerning dome height, withdrawal velocity and power input resulted in the sound ingot of 12kg order. Surface appearance and longitudinal section along the withdrawal direction are shown in Fig. 3, respectively.

Apparently, the surface is smooth and crack free, promising that the ingot surface will be free of trapped slag. Actually some of the slag was stripped off automatically by thermal shrinkage after cooled down and some was removed by simply thrashing. A few black and dotted streaks on Fig. 3a are the remained slag on the ingot surface after withdrawn. Soundness of surface quality facilitates decontamination by the removal of slag containing nuclides.

Figure 3b shows the etched macro structure pattern on the longitudinal section of the ingot. Supplied metallic materials are completely melted and none of the slag is included in the metal, except for the final solidified part at the top of the ingot. This part is black color and filled with the slag. The inclusion of the slag was brought about by thermal shrinkage due to the unsteady state of solidification during power down.

Direction of the structure lines represents the trajectories of heat flow. They are horizontal near the surface, bend to upward in the bulk and become vertical on the axis, except for the initial and final zone of solidification. In the final zone, the structure lines are horizontal. This is responsible for the rapid power down, followed by the removal of the thermal energy mainly by the cold wall. In the initial zone of solidification, the structure lines are not symmetric on the centerline. This seems to be responsible for the eccentric arrangement of the primary and appears only at the initial stage of solidification.

To evaluate the inner quality of the ingot, concentrations of Cu, Mn, Ni, Cr and Fe elements at ingot center and 5mm deep from the surface were analyzed by ICP spectrometer along the withdrawal direction. The concentrations were proved to be homogeneous along the withdrawal direction as well as radial direction for every element, showing that there is no macro segregation. The soundness of the ingot facilitates the reuse of ingot obtained from the MERC.

In some test, the flux composed of the admixture of SiO₂, Al₂O₃, CaO and MgO was melted together with the fragments of stainless steel. Supplied flux melts by the heat transferred from the melt and preferentially gathers into the gap between the molten metal and the cold wall, just like a ring. Some of the flux was not melted and adhered to the cold wall above the molten slag. Addition of the surplus amount of slag or poor control of the molten metal height gradually results in the narrower cross section of the top opening, on account of the accumulation of the molten slag on the cold part of the crucible, followed by solidification. Precise control concerning the supply rate of the flux and dome height of the metal is important for the stable operation of the system.

Thermodynamically, alpha nuclides show the tendency that they can be easily oxidized and transferred to the slag rather than metal. Our conception on decontamination is such that the radioactive nuclides in the metal transfer to the slag during melting. The slag surrounding the ingot sticks on the surface after solidification. The solidified slag is to be collected and disposed. It is found from the test that most of the slag adhered on the ingot surface easily strips off owing to the difference in thermal expansion coefficient between the metal and slag and also to the shear stress between the ingot and crucible during withdrawal, as was shown in Fig. 3. Easiness of the slag removal facilitates the decontamination in the MERC process.

Variation of the ingot diameter along the withdrawal direction is shown in Fig. 4. When the flux was added, the ingot diameter has decreased by 0.2-0.3mm. This means that the slag infiltrates into the small gap between the ingot and the crucible and it is withdrawn together with the ingot.

It is believed that the cold wall of the crucible is corrosion free against both molten metal and molten slag. To estimate this point qualitatively, the ingot surface was analyzed by GD-Mass (Glow-Discharge Mass Spectrometer) deep to 3-micron meter. Objects of the analyzed elements are Ni, W and B that are ones of the heatproof-gilding elements covering the surface of copper-made crucible over 30 micron meter thickness. One of the results is shown in Fig. 5. Ni element is larger in the bulk, while W and B are lower in the bulk. This is responsible for the difference in the concentration of the elements between gliding layer and the ingot. It is found that the every element is leveled off within 3-micron meter from the ingot surface. This means that the element transferred from the crucible is very few and the crucible is corrosion free.

Based on the knowledge obtained through the feasibility study in 1995, the MERC (Melting and Recycling of Metals by Cold Crucible) process was designed, manufactured and scaled up from 45mm to 100mm in crucible diameter. The maximum power of the high frequency generator is 150kW and the frequency is 25kHz.In the MERC, continuously supplied fragments of stainless steels, which were the surrogates of contaminated metallic materials, were successfully melted, followed by solidified and pulled down. The following results were obtained.

1. A cylindrical ingot of 12kg order and 100mm in diameter was successfully cast by the MERC process through melting and solidifying the fragments of stainless steel together with the flux.
2. Necessary output of the high frequency generator for the operation was about 130kW, which was in agreement with that predicted in designing of the MERC process by the theoretical study carried out previously.
3. The energy consumed for the melting was more than 20% of the output of power sauce.
4. The ingot surface was smooth and crack free and almost all the slag was stripped off. There was no macro segregation inside.
5. The element transferred from the crucible to the ingot was very few, promising that the crucible is corrosion free.

1. Y. Nakayama, K. Obara, Y. Tsuji, R.Haun and R. Eschenbach, "Cobalt and Cesium Volatility Test in Plasma Arc Centrifugal Treatment " :Incineration Conference Houston (1995).
2. H. Nakamura, K. Kanazawa and K. Fujiki,"R&D on Melting of Radioactive Metal Wastes at JAERI" :The 3rd JSME/ASME Joint International Conference on Nuclear Engineering,4(1995),1785.
3. S. Abe,M. Yamamoto,S. Morishita,K. Kimura,K. Takahashi, A. Umemura, D. Sakurai and S. Iiboshi, "Demonstration Test on Manufacturing 200L-Drum Inner Shielding Material for Recycling of Reactor Operating Scrap Metal":1994 Fall Meeting of the Atomic Energy Society of Japan, (1994), M27.
4. P.G. CLITES, "The Inductoslag Melting Process" :Bulletin of the United States Bureau of Mines, 673(1982),1.
5. T. Tanaka, K. Kurita and A. Kuroda, "Liquid Metal Flow Coupled with Heat Transfer in a Cold Crucible Confined by Free Surface and Solidification Front" :Iron and Steel Institute of Japan Internatinal,31(1991),1416.
6. K. KANEKO,T. MISAWA and K. TABATA, "Cold Crucible Induction Casting of Semi-Conductor Silicon" :Proceedings of The Sixth International Iron and Steel Cngress,Nagoya,ISIJ,4(1990),254.
7. M.GARNIER, I.LECLERCQ, P.PAILLERE and J.F.WADIER, "The Cold Crucible Continuous Casting Process Applied to Titanium Scrap Melting" :Proceedings of The Sixth International Iron and Steel Cngress,Nagoya,ISIJ,4(1990),260.
8. M.SUZUKI, K.TSURUMAKI, J.KOMATSU, T.TANAKA,Y.IKENAGA and H.NOSE, "Feasibility Study on Recycling of Dismantled Metal by an Induction Cold Crucible Melting Method" :Proceeding of WM’97,50-03(1997).

BACK