Removal of Lead from Copper Alloy Scraps by Compound Separation Method

Removal of Lead from Copper Alloy Scraps by Compound-Separation Method

Atsushi Nakano

_{, Nurul Taufiqu Rochman}

_{and Hidekazu Sueyoshi}

2_{Research Center for Physics, Indonesian Institute of Science, Tangerang 1530, Indonesia}

Recently, the global scale environmental problem has become a critical issue. In metallic material, not only the cost reduction and improvement of mechanical properties but also the decrease in environmental load is required. In copper alloys, several mass% Pb was added to improve the machinability. However, due to the adverse toxicity of Pb that is harmful to the health, a new regulation to limit the amount of Pb permitted in drinking water supplies has been enforced. A huge amount of copper alloy scraps containing Pb will become industrial waste because the scrap will not be available as the raw materials.

We developed a new technique for removing Pb from copper alloy scraps in order to promote recycle of copper alloy scraps containing Pb. Pb was removed from brass and bronze using compound-separation method.

Copper alloys containing 2–6 mass% Pb were molten using a high-frequency induction furnace under nitrogen atmosphere. Ca–Si compound and NaF were added into the molten copper alloys to form large particles of a Pb compound. The large particles of the Pb compound were skimmed off from the molten copper alloys. Liquid metal extractions and castings were characterized by electron probe microanalyser (EPMA) and fluorescence X-ray (XRF) analysis.

The results show that high percentage (83% in brass and 82% in bronze) of Pb removal can be achieved. Therefore Pb-free copper alloys can be produced from copper alloy scraps, resulting in the solution of environmental problems.

(Received June 2, 2005; Accepted November 10, 2005; Published December 15, 2005)

Keywords: lead removal, lead-free copper alloys, compound-separation method, recycle

1. Introduction

Copper alloys (Brass and Bronze) containing several mass% of Pb have been widely used in water faucets and pipes for freshwater supply because of its good machinabil-ity, corrosion resistance and mechanical properties. How-ever, Pb is harmful to human health. In order to limit the amount of permitted Pb in drinking water supplies the regulations were enforced. The leaching standard value of Pb was severely revised to 0.01 mg/L in Japan in April, 2003.1)

As a countermeasure, the development of the Pb-free copper alloys has been advanced. Most of the developed Pb-free copper alloys contain Bi as an alternative of Pb. However, these Pb-free copper alloys are manufactured from the virgin materials. This is caused by low technology for removing Pb from the copper alloy scraps. If we continue the manufactur-ing method only usmanufactur-ing virgin materials, the resource consumption of not only Bi, which is the rare metal, but also Cu, Zn and Sn increases. Also a huge amount of copper alloy scraps containing Pb will be accumulated without being recycled. These are not desirable from the viewpoints of efficient use of resource and recycling.

In order to solve these issues, the development of the new technology for removing Pb from the copper alloy scraps containing Pb is needed. As the methods for Pb removal, evaporation method using Cl and oxidation method have been applied so far.2,3)_{However, these methods may not be}

applied today because of their large environmental impacts and long processing time. Authors have examined the possibility of the Pb removal by compound-separation method in brass.4–6)

In the present study, the compound-separation method in

which the floating large particles of a Pb compound were skimmed off from the molten copper alloys was investigated in both brass and bronze.

2. Experimental Procedure

Brass (JIS CAC203) containing 2.15 mass% Pb and bronze (JIS CAC406) containing 5.5 mass% Pb were used as test specimens. Table 1 shows the chemical composition of brass and bronze. These specimens were observed by electron probe microanalyzer (EPMA). Figure 1 shows SE images of the brass and the bronze. The white spots in the SE images indicate Pb, which remains undissolved within the matrix and is dispersed as particles of severalmmin size.

Figure 2 shows a schematic illustration of the experimen-tal procedure. Brass (3 kg) and bronze (5 kg) were melted using high-frequency induction furnace (20 kW) under nitro-gen gas, respectively. Carbon crucible was used for melting pot (Inner diameter: 110 mm, high: 250 mm). In order to form Pb–Ca compounds, marketed Ca–Si compound was used. Table 2 shows the chemical composition of the Ca–Si compound. According to X-ray diffraction (XRD), this compound consists of CaSi2 and Si. In the case of brass,

the Ca–Si compound was added to the molten brass at 1273 K. After agitating and holding, the formed large particles of a Pb compound rose through the molten brass. Such Pb compounds were then skimmed off from the molten brass. We designate this process as ‘‘Stage I’’. After the Stage I, an aggregation agent, NaF, was added to remove the

Table 1 Chemical composition of copper alloys used (mass%).

Cu Pb Sn P Fe Mn Al Si Zn

Brass 58.51 2.15 0.07 — 0.05 0.005 0.6 <0.005 38.58 Bronze 83.11 5.50 4.56 0.011 0.28 <0.005 <0.005 <0.005 6.33

*_{Graduate Student, Kagoshima University. Present address: Graduate}

School of Science and Engineering, Kagoshima University, Kagoshima 890-0065, Japan

remained small Pb compounds from the molten brass. We designate this process as ‘‘Stage II’’.

When the Ca–Si compound was added to the molten bronze, the percentage of Pb removal was about 20%. Bronze contains Sn as shown in Table 1. Figure 3 shows standard Gibbs energies (G_{) for synthesis of Ca}

2Pb, Ca2Sn, SnF2

and SnF4 as functions of temperature (Obtained by HSC

chemistry 5 in Outokumpu Research). It is suggested that Sn reacts with F in preference to Ca though Ca reacts with Sn in preference to Pb. Therefore, NaF was added to the molten bronze to form the Sn–F compound, followed by adding the Ca–Si compound at 1323 K. The large particles of the Pb compound were skimmed off from the molten bronze. In the above-mentioned processes, the amount of the Ca–Si

com-pound, the holding time of molten copper alloys and the Pb compound removal temperature were changed. After the experiments, quantitative analysis of Pb in the specimens was carried out by X-ray fluorescence (XRF). The microstruc-tures of brass and bronze specimens were characterized by EPMA.

3. Result and Discussion

Figure 4 shows relationships between percentage of Pb removal from brass at Stage I and bronze and the amount of the Ca–Si compound. As shown in Fig. 4, the percentage of Pb removal increases with increasing the amount of Ca–Si compound. However, it tends to saturate with further 4 mass% additions in brass and 10 mass% additions in bronze.

Figure 5 shows relationship between percentage of Pb removal and holding time of molten brass. The percentage of Pb removal increases abruptly with holding time until 10 min and it saturates after 15 min. This suggests that Ca–Si

(a)

10

µ

m

(b)

Fig. 1 SE images of (a) as-received brass and (b) as-received bronze.

Bronze

Melting

Temperature setting

Agitation

Holding

Agitation

Holding

Casting

Characterization ( XRF , EPMA ) Pb compound removal

Sampling

Sampling NaF

Brass

Melting

Temperature setting

Agitation

Holding

Pb compound removal

Agitation

Holding

Casting

Characterization ( XRF , EPMA ) Pb compound removal Stage

Stage

Ca-Si compound

Fig. 2 Schematic illustration of experimental procedure.

Table 2 Chemical composition of Ca–Si compound (mass%).

Ca Si Al Fe C

compound reacts quickly with Pb, and then most of the formed Pb compounds of solid phase move up to the surface of molten brass in about 15 min.

Figure 6 shows relationships between percentage of Pb removal and Pb compound removal temperature. In both brass (Stage I) and bronze, the percentage of Pb removal increases with decreasing Pb compound removal ature. But, in bronze, when Pb compound removal temper-ature is lower than 1183 K, the percentage of Pb removal decreases.

In 6 mass% Ca–Si compound addition to molten brass, the state of retained Pb after Stage I was characterized by EPMA. Figure 7 shows EPMA analysis of brass after Stage I. The distribution of Pb is similar to that of Ca, and Si exists around the Pb compounds. The particle size of Pb compounds is very small. On the other hand, Ca–Si compound particles which are unreactive to Pb are observed in Fig. 7. This suggests that the added Ca–Si compound was too much to Pb content in brass. In Fig. 4, when the amount of Ca–Si compound exceeds 4 mass%, the percentage of Pb removal saturates, which may be caused by the addition of too much Ca–Si compound.

-1000 -800 -600 -400 -200 0

773 973 1173 1373 1573

SnF4

SnF2

Ca2Sn

Ca2Pb

G

∆

0 / kJ · mol

-1

T / K

Fig. 3 Comparison of Standard Gibbs energies for synthesis of Ca2Pb,

Ca2Sn, SnF2and SnF4.

0 20 40 60 80 100

0 2 4 6 8 10

C cs / mass%

Percentage of Pb removal (%)

Bronze Brass ( Stage I )

Fig. 4 Relationships between percentage of Pb removal and the amount of Ca–Si compound, Ccs. (In brass, holding time: 15 min, Pb compound removal temperature: 1220 K. In bronze, holding time: 13 min, Pb compound removal temperature: 1183 K.)

0 20 40 60 80 100

0 10 20 30

Holding time, t /min

Percentage of Pb removal (%)

Brass ( Stage I )

Fig. 5 Relationship between percentage of Pb removal and holding time. (The amount of Ca–Si compound: 4 mass%, Pb compound removal temperature: 1158 K.)

0 20 40 60 80 100

1100 1150 1200 1250 1300

Percentage of Pb removal (%)

T / K

Bronze Brass ( Stage I )

Cu Zn

Pb Ca

SE

10µm

Si

Fig. 7 EPMA analysis of brass after Stage I. (The amount of Ca–Si compound: 4 mass%, holding time: 15 min, Pb compound removal temperature: 1158 K.)

10µm

Ca

Pb SE

Si

Fig. 8 EPMA analysis of skimmed off compound from molten bronze. (The amount of Ca–Si compound: 8 mass%, holding time: 13 min, Pb compound removal temperature: 1183 K.)

Cu Sn

Ca

Pb SE

10µm

Si

Fig. 9 EPMA analysis of bronze after Pb compound removal. (The amount of Ca–Si compound: 8 mass%, holding time: 13 min, Pb compound removal temperature: 1183 K.)

Pb

SE

Ca

10µm

Figure 8 shows EPMA analysis of skimmed off com-pound from molten bronze. The distribution of Pb is similar to that of Ca. This indicates that the Pb compound is formed. Si exists around the Pb compound. The Pb compound is very large (several ten mm). Such large particles of the Pb compound may be moved up to the surface of molten bronze.

In 8 mass% Ca–Si compound addition to molten bronze, Fig. 9 shows EPMA analysis of bronze after Pb compounds removal. The distribution of Pb is similar to that of Ca. This indicates that Pb compounds are formed. The particle size of Pb compounds is very small (severalmm). Besides, the Ca– Si compounds particles which are unreactive to Pb are not seen in Fig. 9. The Pb concentration in as-received bronze is 2.5 times to that in as-received brass. Therefore, it is considered that most of Ca–Si compound particle reacted with Pb.

The above-mentioned results are discussed based on equilibrium phase diagram of Ca–Pb system.7) _{Figure 10}

shows equilibrium phase diagram of Ca–Pb system. We assumed that every Ca reacts with Pb. When the 4 mass% Ca–Si compound is added to molten brass, Pb concentration in Ca–Pb componud becomes 65.1 mass%. When the 10 mass% Ca–Si compound is added to molten bronze, Pb concentration in Ca–Pb compound becomes 65.6 mass%. Ca–Si compound addition temperature is 1273 K in brass, and 1323 K in bronze. According to the equiliblium phase diagram, in both brass and bronze, both liquid phase and solid phase (Ca2Pb) coexist. The ratio of liquid phase in

bronze is larger than that in brass. After holding for a while under such a condition, it is considered that adjacent liquid phases coalesce to each other, and grow up to a large liquid phase. When the temparature is lowered, the ratio of solid phase increases. In other words, more solid phase is crystallized in a liquid phase. Consequently, in both brass and bronze, the amount of solid phase increases with the decrease in temperature. Because large particles of the Pb compound of solid phase becomes easier to move up to the surface of molten copper alloys, the percentage of Pb removal increases with decreasing Pb compound removal temperature (Fig. 6). However, in bronze, when Pb com-pound removal temperature is lower than 1183 K, percentage of Pb removal decreases (Fig. 6). This is because the Pb compounds of solid phase can not move up readily due to the high viscosity of molten bronze at near solidification temperature. Moreover, even when the Pb compound removal temperature is the same in brass and bronze, the percentage of Pb removal in brass is lower than that in bronze. This suggests that it is for small Pb compounds of solid phase difficult to move up to the surface of molten brass. As shown in Fig. 4, in bronze, percentage of Pb removal is 82% at 1183 K of Pb compound removal temperature. According to the equilibrium phase diagram (Fig. 10), not only solid phase but also liquid phase exists at this temperature. This liquid phase remains in molten bronze, because it is not possible to skim off from molten bronze. As mentioned above, percentage of Pb removal in brass was low because of small Pb compounds of solid phase (Fig. 6). In order to aggregate these small Pb compounds, NaF was added to molten brass after Stage I. As a result, the

percentage of Pb removal was improved up to 83%. Figure 11 shows EPMA analysis of brass after NaF addition. Because small Pb compounds aggregate in molten brass, the particle size of Pb compounds is very large (several tenmm). These large Pb compounds may be moved up to the surface of molten brass, resulting in the high percentage of Pb removal. However, the upper limit of the percentage of Pb removal is 83%. This is because liquid phase still remains as shown in equilibrium Phase diagram.

The result of EPMA analysis showed that Na was absent in brass and bronze ingots after Pb compound removal. A small amount of Ca and Si remained in the ingots after Pb compound removal. However, it is considered that the residual Ca and Si may be removed by oxidation refining.

4. Conclusion

The method for removing Pb from copper alloys contain-ing Pb was examined by the compound-separation method. In the case of removing Pb from brass, Ca–Si compound is added to molten brass, followed by skimming off the large Pb compounds which float on the surface of molten brass. Then, in order to aggregate the remaining small Pb compounds, NaF is added to the molten brass. As a result, the large Pb compounds are formed, resulting in the high percentage of Pb removal. Using this method, 83% of Pb can, consequently, be removed from brass containing 2.15 mass% Pb. In the case of removing Pb from bronze, NaF is added to molten bronze, followed by adding the Ca–Si compound. Therefore, the reaction of Ca with Pb is promoted without the reaction of Ca with Sn, resulting in the high percentage of Pb removal. Using this method, 82% of Pb can be removed from bronze containing 5.5 mass% Pb.

Acknowledgments

This work was supported by a Grant-in-Aid for Scientific Research from Japan Society for the Promotion of Science, No. 16510063.

Pb content, CPb / at %

Pb content, CPb / mass %

80 60 50

90 40

30 70 100

80 60 70 50 100 1073 20 1473 1273 1673 1323K 1183K Ca 2 Pb Ca 5 Pb 3 CaPb T / K 1158K 1273K Brass Bronze

REFERENCES

1) Japan Foundry Engineering Society:Cyuzo-kogaku Binran, (Maruzen, Tokyo, 2002) p. 437.

2) K. Okuma, R. Ikeda, E. Yoshida and T. Nakamura: J. the Japan Copper and Brass Research Association34(1995) 201–205.

3) K. Kunii, K. Okuma, E. Yoshida, J. Masuda, H. Okada and T. Nakamura: J. the Japan Copper and Brass Research Association 36

(1997) 132–136.

4) K. Yamada, N. T. Rochman, R. Fujimoto, S. Suehiro and H. Sueyoshi: J. Adv. Sci.13(2001) 273–276.

5) N. T. Rochman, A. Nakano, S. Suehiro, K. Higashiiriki, K. Yamada, H. Sueyoshi, K. Hamaishi, Y. Sechi and T. Matsuda:Proc. EcoMaterials and Ecoprocesses, COM2003(Vancouver Canada, 2003) 245–253. 6) A. Nakano, K. Higashiiriki, N. T. Rochaman, K. Yamada, K. Hamaishi

and H. Sueyoshi: J. Jpn. Inst. Met.69(2005) 198–201.