Minerals Engineering 49 (2013) 184–189

Contents lists available at ScienceDirect

Minerals Engineering
journal homepage: www.elsevier.com/locate/mineng

Recirculation of high arsenic content copper smelting dust to smelting

and converting processes
Victor Montenegro a,*, Hiroyuki Sano b, Toshiharu Fujisawa b
a
Department of Materials Science and Engineering, Graduate School of Engineering, Nagoya University, Nagoya 464-8603, Japan
b
EcoTopia Science Institute, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-8603, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Flue dusts generated in Chilean copper smelters are partly recycled to smelting units, such as flash smelt-
Received 1 June 2009 ers and Teniente Converters, and partly processed separately by hydrometallurgical methods, in order to
Accepted 23 March 2010 recover copper and stabilize arsenic in the waste. Flue dust generation is increasing due to increasing
Available online 9 April 2010
copper production and arsenic content in the copper concentrates, thus creating environmental prob-
lems. Additionally, the flue dusts contain associated metals, such as zinc and lead, which can be poten-
Keywords: tially recovered. Tests on concentrate smelting and matte converting, with various additions of flue dust,
Copper smelting dust
allowed the determination of the distribution coefficients of arsenic, zinc and lead among the products,
Converting process
Dust recirculation
and, as a result of flue dust addition, oxygen enrichment and temperature. The distribution coefficients
Minor elements distribution permit determination of impurities in blister copper, which affect the quality of cathodic copper
Pyrometallurgy produced.
Recycling Crown Copyright Ó 2010 Published by Elsevier Ltd. All rights reserved.

1. Introduction
The removal of impurities is critical in copper metallurgy for the
production of high quality copper, but it is also of great importance
from an environmental aspect. Due to the constant increase in
impurities in copper ore over the years, an even more critical situ-
ation can be reasonably expected in the future. It is essential to
keep impurities below the levels accepted by national and interna-
tional standards for the copper industry, and to decrease the total
emission of harmful components. Currently, the majority of smelt-
ers around the world face this problem.
Impurity removal in copper smelting process is carried out
through oxidation, followed by slagging and volatilization during
smelting, conversion and refining (Yazawa and Azakami, 1969). A
combination of intensive reactors, using oxygen or oxygen-
enriched air as the blowing gas, has provided a number of improve-
ments in copper smelter processes, such as that of the Teniente
Converter, Inco Flash, Outokumpu Flash and Mitsubishi processes.
However, the method of removing impurities has not changed, but
has instead become a serious limitation in the modern process of
copper smelting and converting (Utigard et al., 1997).
Copper processes generate gaseous emissions that contain SO2,
N2, O2, water vapor, heavy metals and other impurities. Metallurgi-
* Corresponding author.
E-mail address: vmontene@gmail.com (V. Montenegro).
cal dust is generated as a result of the gas cleaning process; the gas
contains condensate matter and fine particles from the semi-
melted concentrate, which are transported with the off-gas. The
mass of the dust produced in smelting and converting processes
varies mainly according to the type of reactor. The dust usually
contains copper (Cu), iron (Fe) and other minor elements, such as
lead (Pb), arsenic (As), cadmium (Cd), antimony (Sb), zinc (Zn), bis-
muth (Bi) and selenium (Se), but the form of the elements depends
on the conditions and the operating parameters of the smelting
and converting processes. Arsenic is the main impurity of concern
in the dust and during the smelting process around 15% of the total
input of As from the concentrate is transferred to the matte and
22% to the slag phase, but the main part is eliminated through
the gas phase, creating the copper smelter flue dust (Alguacil
et al., 1996).
The flue dust generated by the Chilean copper smelters is partly
recycled to the smelting units, such as the Flash smelter and the
Teniente Converter, and partly processed separately by hydromet-
allurgical methods, in order to recover copper and stabilize the
impurities (Maldini et al., 2006). The final residue is stored in a
special landfill. In recent years, increase in copper production
and degradation of the quality of copper ore, have resulted in high-
er amounts of dust from the process. In addition, environmental
regulations have become stricter. Due to environmental concerns
the amount of disposed dust must be reduced. Moreover, flue dust
contains associated metals, such as zinc and lead, which can be
potentially recovered. In this study, simulated scale tests of

0892-6875/$ - see front matter Crown Copyright Ó 2010 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.mineng.2010.03.020
 V. Montenegro et al. / Minerals Engineering 49 (2013) 184–189 185

concentrate smelting and matte conversion with addition of flue

dust, allowed the determination of the distribution coefficients of 1 CuSO4 2 Cu 2O 3 ZnSO4
1 6 4 ZnO 5 As2O3 6 PbSO4

Relative intensity
As, Zn, Sb and Pb among the products, as a result of dust recycling,
matte grade, temperature variation and oxygen enrichment. The 56 6
6
outcome also permits evaluation of the effect of dust addition on 6
6 5 66 3
the quality of the matte and blister copper. 4 2 1
3 1

2. Experimental methods
20 30 40 50 60
Industrial matte, slag, copper concentrates and flue dust from a 2θ (deg.,CuKα)
Chilean smelter were used as the test materials. Table 1 shows the
Fig. 1. XRD results of Chilean smelter copper dust.
chemical composition of the samples. X-ray diffraction analysis
(XRD) of the dust samples showed that the main components are
CuSO4, Cu2O, As2O3, ZnO and PbSO4 (see Fig. 1).
Simulated tests of dust addition were conducted in an electric Gas
H2 SO4 aq. NaOHaq.
furnace, as shown in Fig. 2. The furnace was heated to a fixed tem- Outlet
perature (1523–1573 K) under a continuous flow of nitrogen
200 cm3 min1. A Pt/Pt–Rh thermocouple in an alumina sheath Water-Cooled Condenser
was introduced into the furnace to monitor the temperature. Ar or N2 -O2
Two types of experiments were conducted, in order to simulate
the smelting and converting processes, respectively. In the smelt- Water-Cooled Brass Flange
ing test, 13 g of slag, 10 g of matte and 20 g of charge (concentrate,
dust and silica mixed in different ratios) were initially charged in Sample Injection Tube
an alumina crucible (12 mm ID, 100 mm H) and placed in an alu- Gas Injection Tube
mina reaction tube, closed by a water-cooled brass flange. After
the melting and temperature stabilization, an alumina tube was Thermocouple
introduced to start the gas blowing. The ratio of matte/slag used Alumina Crucible
in the test was the same as in the industrial Teniente Converter.
The silica addition was adjusted, so that the ratio Fe/SiO2 in the Sample
slag was constant at about 2. Mullite Pedestal
For the converting experiments, a synthetic fayalite was used in
order to assess the influence of the slag on the impurities with dust Electric Resistance Furnace
addition (slag forming stage). A total mass of 37 g of different Alumina Reaction Tube
materials was initially charged with copper and slag to reproduce
the 3-phase system. To prevent the volatilization of dust during the Nitrogen Gas Inlet
injection, a copper tube (6 mm OD and 5.4 mm ID) was filled with
2 g of dust, which was continuously charged through an alumina Fig. 2. Experimental apparatus.
tube. The number of charged tubes varied with the dust addition
ratio. The temperature of the furnace was maintained at 1473 K.
To control the oxygen partial pressure, a mixture of N2–CO2 was in- NaOH to remove harmful components and to neutralize the SO2.
jected to the furnace to control the partial pressure between 103 The condensed volatile matter, the slag and the matte phases were
and 105 atm. To determine the influence of converting process analyzed for their elemental content, by inductively coupled plas-
time, an initial test was conducted using 5% addition of dust, with ma atomic emission spectrometry (ICP-AES).
a continuous gas blowing of 0.1 l/min of CO2 for 1, 2, 4 and 8 h. The
results showed that the concentration of impurities decreased
3. Results
remarkably in the first 2 h of the process, but after the 4th hour
the concentration remained constant. From these results, the blow-
3.1. Smelting process
ing time for the experimental test was fixed at 4 h.
When the gas injection for both processes was finished, the
3.1.1. Distribution of minor elements
sample was kept at the same temperature for 30 min to promote
The dust addition ratio represents the amount of dust charged
settling and phase separation. The crucible was then taken out of
to the process, which is defined by the following equation:
the furnace and quenched immediately by jetting helium gas.
The off-gas from the reaction tube was continuously passed Addition of dust
through a water-cooled condenser, which collected the volatile Addition ratio ð%Þ   100
Charge ðconcentrate þ dust þ silicaÞ
matter, and then was directed to solutions of 1 M H2SO4 and 1 M
ð1Þ
The distributions of As, Pb and Zn were evaluated by alternating
Table 1
Chemical composition of samples.
the amount of dust added to the smelting process, which used
21 vol.% of oxygen in the blowing gas at 1523 K. Previous research
Element Cu Fe S Zn Pb Bi As Sb (Yazawa and Azakami, 1969; Itagaki, 1986; Montenegro et al.,
Concentrate 36.1 22.9 32.6 0.7 0.2 0.05 0.7 0.01 2008) showed that distributions of Sb and Bi were almost indepen-
Matte 76.4 1.4 19.8 0.6 0.08 ND 0.4 0.01 dent of the dust addition ratio, and for these reasons the impurities
Slag 8.03 38.4 1.9 7.9 0.1 0.1 0.4 0.06
of concern are As, Pb and Zn in the present study.
Dust 10.4 0.8 10.4 15.6 7.8 3.5 19.4 0.1
The results are shown in Fig. 3. The terms h%Xi, (%X) and {%X}
(Unit: mass%) represent the fractional distribution of the element X in the gas,
186 V. Montenegro et al. / Minerals Engineering 49 (2013) 184–189

Gas, <% X > Slag, (% X )

Matte, {% X } X in Matte

100 0.10 4.0

As Pb Zn

Distribution (%)

X in Matte (%)
80 0.08
3.0
60 0.06
2.0
40 0.04

20 0.02 1.0

0 0 0
0 10 20 0 10 20 0 10 20
Dust Addition Ratio (%)
Fig. 3. Distribution of As, Pb and Zn among matte, slag and gas phases in relation to dust addition ratio amount at 1523 K, matte grade 60% Cu.

slag and matte phases, respectively. The closed circle represents
the concentration of each element in the matte. The concentration
of As and Zn in the matte increased remarkably with higher dust
addition ratio, while the amount of h%Asi increased, while h%Pbi
and h%Zni decreased. Additionally, the amount of Pb in the matte
increased slightly. Previous work (Itagaki, 1986) showed that when
the concentration of As in the charge increases, its concentration in
matte also increases. The results showed that the volatilization of
As increased up to 10%, but for higher additions it is almost inde-
pendent of the concentration of As in the initial charge. Moyano
et al. (2006) reported that the elimination of As is almost stable,
even if its concentration increases in the initial charge. This behav-
ior is ascribable to the presence of the metallic oxides of Cu, Zn and
Pb, which in contact with As react and form arsenates.
The results show that the concentrations of Zn and Pb increase
strongly in the matte and the amounts of h%Pbi and h%Zni slightly
decrease with higher dust addition ratio. When the volatilization of
Pb and Zn decreases, the elimination of these impurities by slag-
ging is higher, which is attributed to the fact that Pb and Zn are
originally in the form of oxides in the dust (see Fig. 1), but also
due to the presence of sulfates which react as follows:
MSO4 Dust ) ðMOÞslag þ hSO3 igas
MSO4 Dust ) ðMOÞslag þ hSO2 igas þ 1=2hO2 igas
bution of minor elements in the gas phase slightly increases and in
the matte phase decreases. Their distribution in the slag phase
slightly increases with matte grade.
3.1.3. Effect of oxygen enrichment
Fig. 5 shows the effect of oxygen enrichment in the blowing gas
on the distribution of minor elements and their concentration in
the matte at 1523 K, using 5% dust addition ratio. The distribution
in the slag (%X) increases with oxygen enrichment, and the distri-
bution and concentration of each element in the matte {%X} was
suppressed. These characteristic behaviors were basically caused
by the increase of the oxidation of metals following slagging, when
the oxygen partial pressure increased.
3.1.4. Effect of temperature
Fig. 6 shows the effect of temperature on the distribution of
minor elements, using 10% dust addition ratio. Temperature has
an effect mainly on the vapor pressure of volatile species. The va-
por pressure of minor elements increases with temperature, so a
higher elimination by volatilization would be expected. As is
shown in the results, the concentration in the matte of As, Pb
ð2Þ
and Zn slightly decreased at higher temperature, suggesting that
ð3Þ the effect of temperature on the phase distribution is small.

3.1.2. Effect of matte grade 3.2. Converting process

The matte grade is one of the most important operation param-
eters in the copper smelting process. The distribution of As, Pb and
Zn among the phases was examined by varying the matte grades,
at 1523 K. It was observed that the matte grade had the same effect
for all the dust addition ratios. An example of the obtained results,
using 5% of dust addition, is shown in Fig. 4. In general, the distri-
3.2.1. Slag making stage
During the slag forming stage in the copper converting process,
blister, white metal and slag phases can coexist in the reactor.
White metal is a fine grade of copper sulfide (70–78% Cu) obtained
at a certain stage in copper smelting.

100 0.10 4.0

As Pb Zn
Distribution (%)

X in Matte (%)

80 0.08 3.0
60 0.06
2.0
40 0.04
1.0
20 0.02

0 0 0
55 65 75 80 55 65 75 80 55 65 75 80
% Cu in matte
Fig. 4. Distribution of As, Pb and Zn among matte, slag and gas phases against matte grade at 1523 K, 5% dust addition ratio.
 V. Montenegro et al. / Minerals Engineering 49 (2013) 184–189 187

100 0.10 4.0

As Pb Zn

Distribution (%)

X in Matte (%)
80 0.08 3.0
60 0.06
2.0
40 0.04
20 1.0
0.02
0 0 0
20 30 40 50 20 30 40 50 20 30 40 50
Oxygen Enrichment (%)
Fig. 5. Distribution of As, Pb and Zn among matte, slag and gas phases against oxygen enrichment in blowing gas at 1523 K, 5% dust addition ratio.

100 0.10 4.0

As Pb Zn
Distribution (%)

X in Matte (%)
80 0.08
3.0
60 0.06
2.0
40 0.04

20 0.02 1.0

0 0 0
1523 1573 1523 1573 1523 1573
Temperature (K)
Fig. 6. Distribution of As, Pb and Zn among matte, slag and gas phases against temperature for 10% of dust addition ratio.

Gas Slag Blister White Metal X in Blister

100 2.0 0.75 0.20

As Pb Zn
Distribution (%)

X in Blister (%)
80 0.15
1.5
60 0.50
1.0 0.10
40
0.25
0.5 0.05
20

0 0 0 0
1 5 10 1 5 10 1 5 10
Dust Addition (%)
Fig. 7. Distribution of Pb, Zn and As among blister, white metal, slag and gas phases in relation to dust addition ratio.

The distribution of As, Pb and Zn among the phases was exam-
ined by varying the dust addition ratio, at 1473 K. Fig. 7 shows the
experimental results. The distributions of As, Pb and Zn in the blis-
ter were inversely proportional to the dust addition. The results
show that the presence of multiple phases enhance the elimination
of Pb and Zn. Generally, Pb and Zn were transferred to the slag
phase, because of their initial form in the dust as oxide or sulfide
(see Eqs. (2) and (3)), and due to the oxidation from white metal.
The results show that As was mainly distributed into the blister
and gas phases, due to the high affinity of As with metallic copper.
Previous work (Chaubal and Nagamori, 1983), reported that elim-
ination of As by volatilization is important during the first stage of
converting, when As species have a high vapor pressure.
3.2.2. Copper making stage
At the beginning of the copper making stage, some of the blister,
white metal and gas phases are in contact, while at the end of the
converting stage, blister and gas phases coexist. The distributions
of As, Pb and Zn among the phases were examined using a system
with two and three phases during the copper making stage, vary-
ing the dust addition ratio. The results presented in Fig. 8 show
that the distribution of As, Pb and Zn in the blister decreased when
the dust addition ratio increased. The presence of multiple phases
enhances the elimination of Pb and Zn. During the second stage of
converting, the vapor pressure of As species is low and for that rea-
son the volatilization decreased.
In the refining stage (Fig. 9), where liquid copper is in direct
contact with the gas phase, the elimination of impurities through
the gas phase increased, together with higher dust addition, espe-
cially for As and Pb.
The decrease of Zn was limited during copper blowing. This sug-
gests that Zn reacts principally during the slag making stage.
The concentration of Pb in the blister increased slightly during
the first and second stages of converting, with a slight decrease
at the end of the copper blowing. In the case of As, the presence
of multiple phases suppressed its volatilization from the blister,
suggesting that the contact between gas and blister is absolutely
important for the elimination of As by volatilization.
188 V. Montenegro et al. / Minerals Engineering 49 (2013) 184–189

100 2.0 0.75 0.25

As Pb Zn

X in Blister (%)
Distribution (%)
80 0.20
1.5
60 0.50 0.15
1.0
40 0.10
0.25
0.5 0.05
20

0 0 0 0
1 5 10 1 5 10 1 5 10
Dust Addition (%)
Fig. 8. Distribution of Pb, Zn and As among blister, white metal and gas phases in relation to dust addition ratio.

100 2.0 0.75 0.20

As Pb Zn
Distribution (%)

X in Blister (%)
80 0.15
1.5
60 0.50
1.0 0.10
40
0.25
0.5 0.05
20

0 0 0 0
1 5 10 1 5 10 1 5 10
Dust Addition (%)
Fig. 9. Distribution of Pb, Zn and As among blister and gas phases in relation to dust addition ratio.

4. Discussions injected was converted to SO2. The difference of temperature and

The results of smelting and converting processes with dust
addition, allowed the determination of the distribution coefficients
of arsenic, zinc and lead.
Impurities such as Pb and Zn are mainly eliminated through the
slag phase by increasing the matte grade. However, As remains in
the copper rich phase, because of its high affinity with copper.
Fig. 10 shows the effect of matte grade on the distribution ratios
of As, Pb and Zn between slag and matte, reported by Yazawa
et al. (lines). The present results were compared with their results
and plotted together (points) in the same figure. The present
experiments were conducted at 1523 K without copper precipita-
tion. The partial pressure of SO2 prevailing in the system was con-
sidered to be about 0.2 atm, under the assumption that all oxygen
pSO2 compared with the previous results might not significantly
affect the distribution ratios. Since the present results agree fairly
well with the Yazawa results, it is estimated that the equilibrium
between slag and matte was maintained during the present
experiments.
Figs. 11 and 12 show the distribution coefficients as a result of
metal addition in the charge (As, Zn and Pb) of the smelting and
converting processes. As shown in Fig. 11, the results show that
dissolution of As in the slag is easier than that of Pb and Zn when
dust is added to the smelting process. The results also show that
the behavior of Pb and Zn might limit the amount of dust in the
smelting process. Oxygen enrichment may be effective in sup-
pressing the accumulation of these elements in the matte, because
it is possible to remove them by oxidation, especially Zn.

102
101

101
As
As ( ) Zn
L S/M

0
Zn ( )
S/M

10 0
10
L

Pb
Pb ( )
10-1

-2 -1
10 10
40 50 60 70 80 0 1 2 3 4 5
Matte grade (%) M in Charge(%)

Fig. 10. Distribution ratios of As, Pb and Zn between slag and matte (LS/M). (Plots: Fig. 11. Variation of the distribution ratios of As, Pb and Zn between slag and matte
present work, pSO2  0.2 atm, 1523 K lines: Yazawa et al., pSO2 = 0.1 atm, 1573 K.) (LS/M) with the dust addition ratio during the smelting process.
 V. Montenegro et al. / Minerals Engineering 49 (2013) 184–189
10
3 copper is not linear with the initial charge. This means that some
10
2 Zn
101
Pb
L S/Cu
0
10
-1
10
As
-2
10
10
-3
0 1 2 3
M in Charge(%)
Fig. 12. Variation of the distribution ratios of As, Pb and Zn between slag and
copper (LS/Cu) with the dust addition ratio during the converting process.
During the converting process (Fig. 12), the dissolution of Zn
and Pb is much easier than that of As in the silicate slag. These dif-
ferences seem to be due to the large affinity of As to copper. The
results show that As is mainly transferred to the copper phase,
but the distributions of Pb and Zn are almost stable when the pres-
ence of these elements was greater in the initial charge. At Codelco
Norte Smelter, the concentration of As in blister is around 0.43%,
while Zn and Pb are around 0.40% and 0.50%, respectively (Cabal-
lero et al., 2007). Setting this data as limiting values, the behavior
of Pb and Zn did not influence the final quality of the blister copper.
The behavior of As might limit the recirculation and the quality of
copper. The present results are in agreement with an industrial
evaluation at the Chuquicamata Smelter (Mendoza et al., 1995).
As the results demonstrate, during the smelting process most
impurities were at least partly transferred to the slag or the dust.
Base metals with a high enthalpy of oxide formation, such as Zn
and Pb, have to be transferred to the slag in several steps. During
the converting process, the elements are either volatilized or trans-
ferred to the slag by selective oxidation. After the oxidation of Zn,
the oxygen content increased. This allowed the oxidation of Pb, as
the oxidation rate of Pb is lower than that of Zn. The evidence indi-
cates that the activity coefficient of PbO (cPbO) has a lower value in
slags, where Cu2O coexists with ZnO.
The present results show that the addition of recycling materi-
als (slag, scrap, dust, reverts, etc.) to the process is possible, and the
increase of the content of impurities in the matte and the blister
189
of the impurities are transferred to the slag phase, where it stabi-
lizes. At the same time, the amount of material recycled depends
on the quality of the products in every stage, a quality that is
essential in order to reach the required levels for electro-refining.
5. Conclusions
The present results show that the addition of recycling materi-
als (slag, scrap, dust, reverts, etc.) to the process is possible, and the
increase in the content of impurities in the matte and the blister
copper is not linear with respect to the initial charge. Materials
with high arsenic content may be recycled to the smelting process,
where the concentration of Pb and Zn limit the amount of material
recycled in the smelting process. However, metallurgical streams
with zinc, lead and low concentrations of arsenic can be recycled
to the converting stage. The concentration of As in the blister cop-
per defines the level of recirculation of streams with high concen-
tration of impurities, during the converting and refining stages.
References
Alguacil, F., Magne, L., Navarro, P., Simpson, J., 1996. Tratamiento Hidrometalúrgico
de los Polvos de Fundición de Cobre. Rev. Metal Madrid 32, 400–407.
Caballero, C., Zuñiga, J., Moyano, A., Font, J., 2007. Effect of the codelco norte smelter
operation on the impurities content (As And Sb) in the anode required by the
electrolytic refinery. In: Alvarez, J., Padilla, R., Parra, R., Sánchez, M., Vergara, F.
(Eds.), Proc. of the First Meeting on Minor Element Contaminants in Copper
Metallurgy, Chile, pp. 189–200.
Chaubal, C., Nagamori, M., 1983. Volatilization of arsenic and antimony in copper
matte converting. Metall. Trans. B 14B, 303–306.
Itagaki, K., 1986. Thermodynamic evaluation of distribution behavior of various
elements and effect of the use of oxygen in copper smelting. Metall. Rev. MMIJ
3, 87–100.
Maldini, M., Osorio, J., Mella, M., Herrera, M., 2006. Diagnosis of Current Situation of
Smelter Flue Dust in the Chilean Copper Mining Industries Proc. Hydroprocess,
Iquique, Chile, pp. 402–414.
Mendoza, H., Luraschi, A., Riveros, G., Cerna, M., 1995. Use of a predictive model for
impurity behavior at the chuquicamata smelter. In: Chen, W.J., Diaz, C.,
Luraschi, A., Mackey, P.J. (Eds.), Copper 95, Pyrometallurgy of Copper, vol. 4.
Canadian Institute of Mining, Metallurgy and Petroleum, Quebec, pp. 281–299.
Montenegro, V., Sano, H., Fujisawa, T., 2008. Recirculation of Chilean copper
smelting dust with high arsenic content to the smelting process. Mater. Trans.
49 (9), 2112–2118.
Moyano, A., Caballero, C., Mackay, R., Itagaki, K., Font, J., 2006. In: Kongoli, F., Reddy,
R. (Eds.), Proc. Sohn International Symposium, TMS, vol. 7, pp. 105–113.
Utigard, T., Sanchez, G., Manriquez, J., Luraschi, A., Diaz, C., Cordero, D., Almendras,
E., 1997. Reduction kinetics of liquid iron oxide-containing slags by carbon
monoxide. Metall. Trans. B 28B, 821–826.
Yazawa, A., Azakami, T., 1969. Thermodynamics of removing impurities during
copper smelting. Can. Met. Quart. 8 (3), 257–261.