Desulfurization behavior of molten copper alloy by a soda ash

Desulfurization behavior of molten copper alloy by a soda ash

Hatano, Takaaki

As a basic study on desulfurizing copper alloy scraps in the remelting process, desulfurization experiments were carried out with a molten Cu-8 pct Sn-0.1 pct P by using a Na^sub 2^CO^sub 3^ flux in a 2-kg high-frequency and a 5000-kg low-frequency melting furnace to investigate the influences of the melting atmosphere, the melting temperature, and the flux composition on the desulfurization behavior of the flux. Desulfurization and dephosphorization began simultaneously. Desulfurization ceased earlier than dephosphorization, and then the reversion of sulfur was found to proceed. Covering the melt with charcoal or a carbon crucible enhanced the desulfurizing ability of Na^sub 2^CO^sub 3^. The maximum desulfurizing ability was observed at 1473 K. Sulfur distribution between the melt and the flux increased with increasing the basicity of the flux and with decreasing Cu^sub 2^O content in the flux. The desulfurization rate was also evaluated based on a two-film model.


IN recent years, there has been an increasing need for effective use of scrap in metal production. Though copper alloy scrap recovered from disused products is generally returned to the smelters and recycled to electrolytic copper, a process to select and remelt scraps is essentially needed.

In general, for a variety of scraps, there is great concern about the dissolution of impurities into the melt. Among the many impurity elements, most attention should be paid to sulfur, which comes from the oil adhering to scraps in melting and must be removed in the refining process. The author examined various desulfurizing methods for a molten Cu-8 pct Sn-0.1 pct P alloy under conditions similar to that in the industrial remelting process and found that desulfurizing by soda ash (Na^sub 2^CO^sub 3^) is effective.[1]

In the steel refining process, the simultaneous desulfurizing and dephosphorizing treatment of a hot metal by Na^sub 2^CO^sub 3^ has been studied for upgrading the purity, and a number of studies have been reported on the reaction mechanism,[2,3,4] treatment process,[5,6,7] reaction with refractory,[8] recovery technique of Na^sub 2^CO^sub 3^,[19] and so on. Many of these reports are applicable to the desulfurization of a molten copper alloy. However, there are significant differences between the two systems, such as melting points of the metals and physicochemical properties of the liquid phases, which really means that some basic information is required before applying the process on an industrial scale.

In the copper refining process, removal of As, Sb, Bi, Sn, Se, Te, Pb, and Fe by Na^sub 2^CO^sub 3^ has been investigated.[10-17] Nakamura et al.[10] reported on excellent desulfurization results with Na^sub 2^CO^sub 3^; however, this required additions of Na^sub 2^CO^sub 3^ to as much as one-third of the melt weight. In this study, the influence of various factors on desulfurization with Na^sub 2^CO^sub 3^ was investigated experimentally from the point of view of industrial operation for a molten Cu-8 pct Sn-0.1 pct P alloy.


In this study, the effects of the furnace atmosphere, temperature, and flux composition on desulfurization behavior were investigated through comparison between the experimental results and the preceding theory.


The experiments were carried out in a laboratory furnace and an industrial furnace. Scraps, electrolytic copper, electrolytic tin, and P-Cu alloy were used as melting materials, which were mixed and then inserted in the furnaces. Sodium carbonate of more than 99.0 pct purity supplied by Kantochemical, Inc. (Tokyo) was used as a desulfurizing flux.

A. Experiments in the Laboratory Furnace

A high-frequency induction furnace was used for the melting furnace, in which a cylindrical crucible of 60-mm i.d. and 200-mm height made of carbon and cylindrical crucibles of 60-mm i.d. and 200-mm height and of 90-mm i.d. and 200-mm height made of alumina refractory were placed. The top of the furnace was opened to air; 1.8 to 2.0 kg of raw material was prepared, and the range of 0.004 to 0.005 wt pct of sulfur content was targeted through the addition of CuS. The experimental conditions are shown in Table I. In order to determine the effect of the reduction potential of the gas atmosphere on desulfurizing efficiency, the conditions of the melt covering and the furnace material were varied. The effect of temperature was also examined in the range from 1423 to 1573 K, which is typical for remelting copper alloys.

The raw material with pieces of charcoal of about 20 mm in diameter was inserted into the crucible and then melted. The melt was kept at the temperature indicated in Table I, and more charcoal was added to make up for oxidized charcoal. Na^sub 2^CO^sub 3^ was then added into the melt. Samples of the melt were taken at fixed time intervals by drawing 10^sup -2^-kg samples into a 4-mm-diameter silica tube and water quenching. After the Na^sub 2^CO^sub 3^ was added, the melt temperature fell down by about 50 K, and it took a few minutes to recover to the initial temperature.

B. Experiments in the Operating Furnace

A low-frequency induction furnace made of alumina refractory was used for the melting furnace, in which 5000 kg of raw material was melted. The top of the furnace was opened to air. After the material melted down, the melt surface (area: 1.05 m2) was covered with charcoal to prevent oxidation. The melt temperature was then raised to 1473 K, and the melt composition was then adjusted. Twenty-five to fifty kg of Na^sub 2^CO^sub 3^ was added, and samples of the melt and the flux were picked up periodically by a graphite ladle.

C. Analysis of Samples

For the melt sample, sulfur content was analyzed by a combustion method and phosphorous and tin contents by Xray fluorometry and wet analysis. For the flux sample, sulfur, phosphorus, tin, copper, aluminum, silicon, and sodium contents were analyzed by an inductively coupled plasma emission spectroanalysis. In addition, the total carbon content was analyzed by a combustion method, and the content of the carbon in carbonate was also analyzed by a coulometric titration. Since a considerable amount of charcoal was mixed in with the flux, carbon, except for the carbonate, was assumed to be mostly from charcoal and it was thus excluded from the flux composition. The carbonate was assumed to be Na^sub 2^CO^sub 3^ and the rest of sodium to be Na^sub 2^O. Phosphorus, silicon, copper, tin, and aluminum were taken as P^sub 2^O^sub 5^, SiO^sub 2^, Cu^sub 2^O, SnO^sub 2^, and Al^sub 2^O^sub 3^, respectively.


A. Changes in Melt and Flux Compositions: Operating Furnace

The typical changes in the melt and flux compositions after adding 50 kg of Na^sub 2^CO^sub 3^ to 5000 kg of melt in the operating furnace at 1473 K are shown in Figures 1 through 3. The typical composition of a slag sampled before adding Na^sub 2^CO^sub 3^ and the composition of Na^sub 2^CO^sub 3^ used for the experiments are shown in Table II. While [S] first decreased rapidly and then increased slowly, (S) increased rapidly and then decreased correspondingly. The amount of [P] decreased rapidly and then continued to decrease slowly. The change in (P^sub 2^O^sub 5^) coincided with that in [P] during the first half of the reaction time, but both (P^sub 2^O^sub 5^) and [P] decreased simultaneously during the second half of the test, which implies the vaporization of phosphorus.[21 While [Sn] did not change significantly, both (SnO^sub 2^) and (Cu^sub 2^O) increased in the initial stage. (Al^sub 2^O^sub 3^) and (SiO^sub 2^), which were not present in the melt, increased in the initial stage due to the erosion of the refractory and the dissolution of the initial slag. After an initial rapid increase, (Na^sub 2^O) continued to increase gradually until 30 minutes and then decreased. The change in (Na^sub 2^CO^sub 3^) was opposite to that of (Na^sub 2^O).

B. Condition in Furnace

Since the melting point of Na^sub 2^CO^sub 3^ is 1131 K,[221 it melted as soon as it was added to the melt. When the melt was covered with charcoal or a carbon crucible was used, much fume and flame arose immediately and then fuming decreased with time and finally ceased. Without the charcoal covering in an alumina crucible, however, the fume was not observed. The viscosity of the melting flux increased with time and finally solidified. When the fume ceased, the flux solidified almost simultaneously and the reversion of sulfur started.

Both the fume on the hood of the furnace about 1.5 m above the melt surface and the solidified flux were sampled after the experiment. Their chemical compositions were then analyzed by the method described previously, and their phases were also identified by X-ray diffraction, as shown in Table III and Figure 4. For the solidified flux, the main contents were Na^sub 2^CO^sub 3^, Na^sub 2^O, and P^sub 2^O^sub 5^. In the diffraction pattern, while the main peaks of Na^sub 2^CO^sub 3^ disappeared, the peaks of gamma -3Na^sub 2^O P^sub 2^O^sub 5^ appeared. In the phase diagram of Na^sub 2^O-P^sub 2^O^sub 5^,[23] 3Na^sub 2^OP^sub 2^O^sub 5^, with the melting point of 1856 K, exists in the high concentration range of Na^sub 2^O. The flux therefore likely solidified due to an increase in melting point.

The main contents of the fume were also Na^sub 2^CO^sub 3^, Na^sub 2^O, and P^sub 2^O^sub 5^. The diffraction pattern consisted of the strong peaks of Na^sub 2^CO^sub 3^ and the weak peaks of gamma – 3Na^sub 2^O P^sub 2^O^sub 5^.

Since the fume is spherical in shape, as shown in Figure 5, the fume is assumed to arise due to the reaction of Na and CO gases produced by Reaction [12] with O^sub 2^, CO^sub 2^, and P^sub 2^O^sub 5^ gases in the gas phase.

C. Influence of Melting Atmosphere and Crucible Material on Desulfurization Behavior

In the industrial remelting of copper alloys in air, either the carbon crucible or melt covering materials made of carbon are used to prevent melt oxidation. To compare the desulfurization behavior in an oxidizing atmosphere with that in a reducing atmosphere Na^sub 2^CO^sub 3^ was added in the alumina crucible in the lab furnace under both conditions. In one case, charcoal was present with the Na^sub 2^CO^sub 3^ flux; in the other, no charcoal was used. In both cases, the raw material was melted in the presence of charcoal, but it was removed before the flux addition in the second case. A test was made to confirm that both desulfurization and dephosphorization did not take place in the presence of charcoal only.

The changes in [S] and [P] with time are shown in Figure 6. Under both conditions, desulfurization and dephosphorization took place rapidly. The initial rates of [S] and [P] removal were, however, higher, and the start of the sulfur reversion was also earlier in the lab furnace than in the operating furnace, as indicated in Figure 1, because the value of V^sub m^/A in Eq. [9] for the lab furnace, 0.04 to 0.08 m, is smaller than that for the operating furnace, 0.6 m. Charcoal covering decreased the minimum [S] and suppressed both the reversion of sulfur and the dephosphorization.

To examine the effect of the crucible material on desulfurization behavior, the melts were desulfurized with charcoal covering both in the carbon crucible and in the alumina crucible using the lab furnace. The change in [S] with time is shown in Figure 7. In the carbon crucible, [S] after desulfurization became lower and the degree of sulfur reversion also became smaller than those in the alumina crucible. The degree of desulfurization, eta^sub s^, is plotted against the added weight of Na^sub 2^CO^sub 3^, as shown in Figure 8. For the carbon crucible, the equivalent desulfurizing efficiency to that for the alumina crucible was obtained by adding half the weight of Na^sub 2^CO^sub 3^. Thus, it increases the desulfurizing efficiency to make the atmosphere in the furnace reducing, because a decrease in ho leads to increase in L^sub s^ in Eq. [2] and also suppresses the dephosphorization reaction, which decreases alpha^sub Na2O^ by producing P^sub 2O^sub 5^.[25

D. Influence of Treating Temperature on Desulfurization Behavior

The desulfurization behaviors at 1423, 1473, 1523, and 1573 K under the same condition using the lab furnace are shown in Figures 9 and 10. Desulfurization proceeded rapidly at all temperatures tested. The [S] after desulfurization was a minimum and /7s was a maximum at 1473 K. The degree of sulfur reversion was a minimum at 1423 K. Temperature was not found to affect dephosphorization behavior.

Although the desulfurization Reaction [1] is supposed to be more efficient at higher temperatures according to the equilibrium constant, the degree of desulfurization was a maximum at 1473 K. This contradiction is attributed to various reactions that occurred simultaneously with the desulfurization reaction. For example, the standard free energy change for decomposition of Na2CO3 to Na (Eq. [12]), which decreases desulfurizing ability, is given as AG9 = 932,599 – 660.25T J/mol by Inoue and Suito.[261 Reaction [12] is also enhanced at higher temperatures and is influenced by temperature more significantly than Reaction [1], because the coefficient of temperature is larger than that in AGo. In hot metal, it is known that the desulfurizing efficiency is low at high temperatures due to vaporization of Na2CO3.[27] Thus, it is assumed that the decrease in desulfurizing ability below 1473 K was caused by the temperature dependency of desulfurization equilibrium, and, on the other hand, the decrease at high temperatures was caused by the enhanced decomposing reaction of Na^sub 2^CO^sub 3^.

E. Influence of Flux Composition on Sulfur Distribution: Operating Furnace


The influence of various factors on desulfurization with a molten Cu-8 pct Sn -0.1 pct P by using a Na2CO3 flux was investigated. The desulfurizing efficiency was determined by the partial pressure of oxygen in the gas atmosphere, the basicity of the flux, and also the system temperature. The oxygen pressure is affected by the furnace material, covering material, and CuO content in flux. The basicity decreases with increasing the contents of acid oxides, especially P205 in the flux. Thus, the following operations are necessary to improve the desulfurization efficiency.

1. Before adding the flux, remove the initial slag, which dissolves P205 and Cu20 in the flux.

2. Use a carbon crucible or cover the melt with charcoal, which decreases the oxygen pressure and suppresses dephosphorization producing P205.

3. Add the flux at 1473 K, where the desulfurizing ability is a maximum.

The author expresses his gratitude to the engineers at Nippon Mining and Metals Co. Ltd., especially Messrs. T. Yano and N. Kimura, who carried out much of the experimental work, and Dr. J. Miyake, who read and criticized the manuscript.


1. T. Hatano, T. Yano, M. Tsuji, and T. Ogura: J Jpn. Inst. Met., 1995, vol. 59, pp. 44-49.

2. H. Suito, A. Ishizaka, R. Inoue, and Y. Takahashi: Tetsu-to-Hagane,

1979, vol. 65, pp. 1848-57.

3. R. Inoue and H. Suito: Tetsu-to-Hagane, 1982, vol. 68, pp. 417-25. 4. S. Yamamoto, Y. Fujikake, and S. Sakaguti: Tetsu-to-Hagane, 1982,

vol. 68, pp. 1896-1904.

5. M. Hanmyo, E. Ogura, S. Kuriyama, O. Yamase, K. Yamada, and K. Iwasaki: Tetsu-to-Hagane, 1983, vol. 69, pp. 1849-55. 6. Y. Nakajima, M. Mukai, and T. Moriya: Tetsu-to-Hagane, 1983, vol.

69, pp. 1863-70.

7. S. Yamamoto, H. Ishikawa, Y. Fujikake, C. Satto, and H. Kajioka: Tetsu-to-Hagane, 1983, vol. 69, pp. 1871-77. 8. K. Hiroo: Tetsu-to-Hagane, 1983, vol. 69, pp. 1924-30. 9. T. Kato, K. Tajima, K. Yamashita, E. Ogura, and M. Hammyo: Tetsu

to-Hagane, 1982, vol. 69, pp. 1878-84.

[0. T. Nakamura, F. Noguchi, Y. Ueda, and T. Kusakabe: J Min. Met. Inst. Jpn., 1984, vol. 100, pp. 675-80.

ll. I.V. Kojo, P. Taskinen, and K. Lilius: Erzmetallurgy, 1984, vol. 37, pp. 21-26.

12. I.V. Kojo: Acta Polytech. Scand. Chem. Technol. Metall. Ser., 1985, vol. 161, pp. 148.

13. T.T. Stapurewicz and N.J. Themelis: Metall. Trans. B, 1990, vol. 21B,

pp. 967-75.

14. H. Fukuyama, T. Fujisawa, and C. Yamauchi: J. Jpn. Inst. Met., 1993, vol. 57, pp. 521-26.

15. H. Fukuyama, M. Tonsho, T. Fujisawa, and C. Yamauchi: J Min.

Met. Inst. Jpn., 1993, vol. 109, pp. 601-06. 16. H. Fukuyama, J. Shimotori, T. Fujisawa, and C. Yamauchi: J. Jpn. Inst. Met., 1993, vol. 57, pp. 905-13.

17. H. Fukuyama, E. Wada, T. Fujisawa, and C. Yamauchi: J Jpn. Inst.

Met., 1993, vol. 57, pp. 1149-57.

18. I. Barin, O. Knacke, and O. Kubaschewski: Thermochemical Properties of Inorganic Substances, Springer-Verlag, Berlin, 1973. 19. K. Sudo: Sci. Rep. Res. Inst. Tohoku Univ., Ser. A, 1950, vol. 2, p. 519.

20. M.M.A. El-Nagaar and N.A.D. Parlee: Metall. Trans., 1970, vol. 1, pp. 2975-77.

21. T. Masanori: Tetsu-to-Hagane, 1983, vol. 65, pp. 1699-1713. 22. O. Kubaschewski, E.LI. Evans, and C.B. Alcock: Metallurgical

Thermochemistry, 4th ed., Pergamon, Oxford, 1968. 23. E.T. Turkdogan and W.R. Maddocks: J. Iron Steel Inst., 1952, vol. 152, p. 1.

24. Y. Nakamura, K. Harashima, Y. Hukuda, N. Tokumitsu, and S. Yamamoto: Tetsu-to-Hagane, 1980, vol. 66, pp. 2023-31. 25. S. Yamaguchi and K.S. Goto: J. Jpn. Inst. Met., 1983, vol. 48, pp. 43-49.

26. R. Inoue and H. Suito: Tetsu-to-Hagank, 1979, vol. 65, pp. 1838-47. 27. K. Narita: Tetsu-to-Hagane, 1971, vol. 57, pp. 411-29. 28. T. Moriya and M. Tawara: Tetsu-to-Hagane, 1977, vol. 63, pp. 207076.

29. K. Narita: The 54 and 55th Nishiyama Memorial Sem., Iron and Steel Institute of Japan, Tokyo, 1978, p. 101.

30. K. Bornemann and F. Sauerwald: Z. Metallkd., 1922, vol. 14, p. 10.

TAKAAKI HATANO, Metallurgical Engineer, is with the R&D Department, Kurami Works, Nippon Mining and Metalls Co., Ltd.. Kanagawa 253-01, Japan.

Manuscript submitted March 26, 1996.

Copyright Minerals, Metals & Materials Society and ASM International Feb 1998

Provided by ProQuest Information and Learning Company. All rights Reserved