Transactions

of theJapanInstituteof Metals,Vol. 23, No. 6 (1982),pp. 328to 333

Equilibrium Relations between Liquid Copper

and Calcium Ferrite Slag

By Akira Yazawa* and Yoichi Takeda*

Experimentswerecarriedout on equilibriumbetweencalciumferriteslag and metalliccopper

undera controlledpartial pressureof oxygenby using a mixtureof CO2and CO. Lime,magnesia
or iron crucibleswere used in the temperaturerange from 1423to 1623K. The resultswere sum-
marizedas follows:
(1) The solubilityof copper in the stagsis affectedby the oxygenpotentialand temperature,
and the solubilityamountsto 70% of that in the caseof iron silicateslags.
(2) The value of (Fe+++/Fe++) in slags fits the equation previouslyobtained in the plain
FeO-FeO1.5-CaOternarysystem,and thisratio in the basiccalciumferriteslagsis ten timesas large
as that in acidiciron silicateslags.
(3) The iron contentsin metalliccopperare relatedto the oxygenpotential,temperatureand
iron contentin the slag.
(Received
February 19,1982)

brought into equilibrium in an iron, lime or

Ⅰ. Introduction magnesia crucible at a fixed temperature be-
tween 1423 to 1623K, usually under a current
Iron silicate slags are widely used in extrac- of CO-CO2 gas mixture. Experimental con-
tive metallurgy, but there exist various inherent ditions are summarized in Table 1, and most
inconveniences(1) such as valuable metal losses procedures were similar to those described in
to the slag and low holding capacity for iron the previous report(2). Because the system has
oxide, especially for magnetite. On the other four or five components (Fe-Cu-O-CaO,
hand, interesting features of a new slag system +MgO) and four phases (metal, slag, gas and
composed of calcium ferrite have recently crucible), the freedoms are derived from the
been recognized(1), and phase equilibria in the phase rule, and at a fixed temperature the
FeO-Fe2O3-CaO system under the controlled equilibrium should be established by designat-
oxygen potential were reported by the au- ing the oxygen potential and/or the CaO
thors(2). Such calcium ferrite slag is used suc- content in slag, as shown in Table 1. The oxygen
cessfully in Mitsubishi converting furnace potential was derived from the composition
under coexistence with liquid copper(3)(4). of gas mixture by use of the following data(5):
Thus, the distribution of various elements be-
2CO(g)+O2(g)=2CO2(g),
tween liquid copper and calcium ferrite slag
ΔG°/J=-564840+173.30T, (1)
is very interesting from both basic and practical
standpoints. The equilibrium experiments were and

carried out for the plain copper and calcium logPO2=2log(PCO2/PCO)-29510/T+9.05. (2)
ferrite slag system under the controlled oxygen
After confirming the time necessary to attain
potential, and results obtained are summarized
in this paper. equilibrium as illustrated in Fig. 1, the heating
time required was determined for each case as
Ⅱ. Experimental Procedure shown in Table 1. The lime crucible was
available only under higher oxygen potential,
Calcium ferrite slag and copper metal were where the formation of dicalcium ferrite
(2CaO・Fe2O3) phase was observed as an

* The Research Institute of Mineral Dressing and intermediate layer between liquid slag and the
Metallurgy (SENKEN), Tohoku University, crucible. The intermediate phase observed
Sendai 980, Japan. between liquid slag and magnesia crucible was
 Equilibrium Relations between Liquid Copper and Calcium Ferrite Slag 329

Table 1 Experimental conditions.

Ⅲ. Experimental Results and Discussion

1. Copper contents in slag

In Fig. 2, copper contents in slag equilibrat-

ing with liquid copper are plotted against the
oxygen potential. Approximately, a linear
relation is observed for a respective tempera-
ture, but a steeper slope may be expected
above -5 of log PO2, because liquid copper
will be oxidized completely at around -4 of
log PO2 at these temperatures. The dash-dot line
in Fig. 2 illustrates the contents of copper in
Fig. 1 Effect of melting time on copper content in
slag. liquid iron silicate slag at 1537K(8). It is rec-
ognized that the copper content in calcium
ferrite slag corresponds to around 70 per cent
magnesio-wustite (FeO-MgO solid solution)
of that in iron silicate slag, that is, a lower
or magnesio-ferrite (Fe3O4-MgO・Fe2O3 solid
copper loss in slag is expected in ferrite slag.
solution) depending on the melting conditions. However, any appreciable difference could not
Methods of chemical analysis for calcium, be observed in the temperature dependence
ferrous iron and total iron in the slag sample of copper solubility between ferrite and silicate
were described in the previous report(2). slags.
Methods of atomic absorption were adopted
to determine the contents of magnesium in
slag and iron in metal. Copper contents in
slag were usually determined by EDTA-
spectrophotometric titration, but when the
content was small the method of atomic absorp-
tion was also employed. When the slag is
dissolved by acid for chemical analysis, it is
recognized that the following substitution
reaction takes place(6)(7):
Cu+(melt)+Fe3+(melt)→

Cu2+(aq)+Fe2+(aq). (3)

Thus, the analytical results of ferrous iron in

slag were corrected by the copper contents.
Fig. 2 Effect of oxygen potential on copper content
in slag.
330 Akira Yazawa and Yoichi Takeda

where, a, γ and MCu are the activity, the activity

coefficient, and the atomic weight of copper,

respectively, and the parentheses denotes the

value in slag. (nT) is the total mole number of

constituents in 100g of the slag phase when all

constituents are expressed with mono-nuclear

metal atom base such as FeO-FeO1 .5-CaO-

MgO-CuO0 .5. This way of expression of toal
mole number is very convenient, because the
value of nT can be taken as a fixed one at a

first approximation.
nrst In the present experiment

the values of (nT) vary from 1.41 to 1.49 even

under considerable variation in slag composi-

tions. Thus, assuming aCu=1 and (nT)=1.45,

the following equation can be derived from eq.

Fig. 3 Effect of CaO content on copper content in
(6).
slag.
(%Cu)=92K5P1/4O2/(γCuO0.5). (7)
Copper contents in slag at a fixed oxygen
If (γCuO0 .5) is assumed to be constant, the slope
potential are plotted against CaO contents as
of the lines in Fig. 2 may be expected to be 1/4,
shown in Fig. 3, and any distinct dependence
but a slightly lower value of the slope is ob-
on CaO contents are hardly observed. The
served experimentally, as shown in Fig. 2
bottom line represents data obtained in iron
and eq. (4). Figure 4 shows the results when
crucibles, where minute decrease in the oxygen
the difference in the slopes is ascribed to the
potential was confirmed(2) with increasing CaO
variation of (γCuO0 .5), and is approximately
content. The copper contents are approxi-
expressed with the following equation:
mately 0.3 per cent, but tend to increase slightly
with increasing temperature and decreasing log(γCuO0 .5)=0.029logPO2+581/T+0.488. (8)
CaO content. It should be noticed that copper
The values of (γCuO0 .5) for calcium ferrite slag
solubility can decrease usually with decreasing
lie between 3 and 6, while those for iron silicate
oxygen potential, but hardly decrease to the
slag are reported to be 2 to 3(8).
level less than 0.3 per cent.
Because the effect of CaO contents is neg- 2. Ratio of Fe3+/Fe2+ in slag
ligible, the copper contents in calcium ferrite
Detailed description was made in the pre-
slag are determined mainly by the oxygen
vious report(2) for the effects of oxygen poten-
potential and temperature. From the linear
tial, slag compositions and temperature on the
relations in Fig. 2, the following equation can
ferric to ferrous ratio of FeO-FeO1.5-CaO
be derived:
slag. Similar treatments are possible to the
log(%Cu)=0.221logPO2+2490/T+0.455. (4) present slag containing cuprous oxide, and the
The mono-valent dissolved copper species Fe3+/Fe2+ ratios are plotted against the oxygen
are accepted to exist in slag which is in equi- potential or CaO content in Fig. 5 or 6. The
librium with the phase of liquid copper, and
the following equations have been used for the
thermodynamic discussion(7)-(9):

Cu(1)+1/4O2(g)=CuO0
Cu(1)+1/4O2(g)=CuO0.5(1), .5(1),

ΔG°/J=-58800+19.55T(5), (5)

and

Fig. 4 Activity coefficient of CuO0.5 in slag plotted

(6) against oxygen potential.
 Equilibrium Relations between Liquid Copper and Calcium Ferrite Slag 331

Fig. 5 Ferric to ferrous iron ratio in slag plotted

against oxygen potential.

solid lines in both figures depict the results

obtained for the plain FeO-FeO1.5-CaO
system(2), and are agreeable well with the
Fig. 6 Ferric to ferrous iron ratio in slag plotted
plotted points obtained for the present slag
against CaO content.
containing CuO0.5 and MgO. This suggests
that the effects of CuO0.5 and MgO on the the existence of cuprous oxide in slag.
Fe3+/Fe2+ ratio are relatively small, and the The solubilities of MgO in ferrite slag from
following equation for Fe3+/Fe2+ ratio derived magnesia crucible are illustrated in the top
in the previous report(2) is applicable also to of Figs. 5 and 6. The contents are 1 to 2 per
the slag equilibrating with liquid copper: cent and can be expressed as follows:
log(Fe3+/Fe2+)=0.170logPO2+0.018(%CaO)
log(%MgO)=-5650/T+0.035logPO2+4.23.
+5500/T-2.52. (9)
(10)
The Fe3+/Fe2+ ratios for ferrite slag are around
In Fig. 5, MgO contents tend to increase with
10 times higher than those for iron silicate
increasing oxygen potential, but this may be
slag.
caused by increasing CuO0 .5 in the slag.
In Fig. 6, the data obtained for the slag sam-

ple equilibrating with copper in lime crucible, 3. Iron content in liquid copper
which are plotted with mark ●, represent the
The behaviour of iron in the liquid copper
CaO content higher by 1 to 3 per cent than the
phase may be discussed based upon the fol-
line RR' which is the CaO content in the plain lowing equations:
FeO-FeO1.5 CaO melt saturated with dical-
cium ferrite. Such an increase in the CaO
(11)
solubility in the present slag may be ascribed to
 332 Akira Yazawa and Yoichi Takeda

and
K11=αFeO v/(aFe・Pv/2O2). (12)

Using a method similar to eq. (6),

(13)

where [nT] and (nT) are total mole numbers in

100g of metal phase and slag phase and may

be taken as 1.57 and 1.45, respectively, under

the present experimental conditions. If the

dependence of the ratio of (γFeO

v)/[γFe] on the
oxygen potential is negligible, linear relations

with a slope of -v/2 are expected between

logPO2 and log[%Fe]. Figure 7 shows such

linear relations having the slope between 1/2

to 3/4, where the solid lines are correspon iding

to the case of 55 per cent of Fe in slag. The

slopes of the lines become steeper with lower

oxygen potential. In the top of the figure, the

solubilities of iron in liquid copper from iron

crucible are illustrated. The values are 3.2,

4.3, 5.5 and 7.7 per cent at 1423, 1473, 1523

and 1573K, respectively, not influenced by

Fig. 8 Solubility of iron in copper metal plotted
slag compositions, and are somewhat lower
againstiron content in slag.

than those suggested by Hansen(10).

In Fig. 8, the iron contents in copper metal
are plotted against those in the slag under each
fixed oxygen potential. Combining Figs. 7
and 8 with eq. (13), the following equation
could be derived:
log[%Fe]=-0.553logPO2+1.92log(%Fe)
-17200/T+2 .09. (14)
This equation shows a good agreement with
the present data except the case of the ap-
proximate iron-saturation or of the iron
content in copper metal lower than 0.01 per
cent. Any serious difference in the behaviour
of iron in copper metal can not be recognized
between the data obtained for the ferrite slag
system and those for the silicate slag one(11)(11).

Ⅳ. Summary

Liquid copper and calcium ferrite slag were

equilibrated in lime or magnesia crucible at
Fig. 7 Solubility of iron in liquid copper plotted 1423 to 1623K under a current of CO-CO2
against oxygen potential. gas. The results are as follows:

(1) Dissolved copper contents in slag are
affected by the oxygen potential and tempera-
ture, and little influence is observed by the
CaO content in slag. The contents are expressed
by the following equation:
log(%Cu)=0.221logPO2+2490/T+0.455,
and they correspond to around 70 per cent of
that in silicate slag at a fixed oxygen potential
and temperature. Dissolved oxide will be in
the form of CuO0 .5, and its activity coefficient,
(γCuO0.5), is 3 to 6 in ferrite
3 in silicate slag.
(2) (Fe3+/Fe2+)ratioin slagcan be de-
scribedby the oxygen potential,temperature
and CaO contentas follows:
log(Fe3+/Fe2+)
=0.170logPO 2+0.018(%CaO)+5500/T-2.52.
This equation is just similar to that derived for
plain FeO-FeO1.5-CaO slag described in the
previous report, and gives a good agreement
with the data at PO2from one order above that
at the iron-saturation to 10-4. This ratio in
ferrite slag is around ten times larger than
that in iron silicate slag under the same con-
ditions.
(3) Iron content in liquid copper is ex-
pressed as follows:
log[%Fe]=-0.553logPO2+1.92log(%Fe)
-17200/T+2 .09,
Akira Yazawa and Yoichi Takeda 333
when the oxygen potential is one order higher
than the iron-saturation and [%Fe] is more
than 0.01.
Acknowledgement
The present investigation was supported
partly by a Grant-in-Aid for Scientific Re-
search by the Ministry of Education.
REFERENCES
(1) A. Yazawa, Y. Takeda and Y. Waseda: Can.
Met. Quart. 20 (1981), 129.
slag, whereas 2 to
(2) Y. Takeda, S. Nakazawa and A. Yazawa: Can.
Met. Quart. 19 (1980), 297.
(3) T. Nagano and T. Suzuki: Extractive Metallurgy
of Copper, Vol. 1, Ed. by J.C. Yannopoulos and
J.C. Agarwal, RIME, (1976), p. 439.
(4) A. Yazawa: Erzmetall, 30 (1977), 511.
(5) A. Yazawa (ed.): Nonferrous Extractive Metal-
lurgy, Jap. Inst. Metals, (1980), p. 315 (in Japa-
nese).
(6) R.L. Altman: Anal. Chim. Acta, 63 (1973), 129.
(7) M. Eguchi and A. Yazawa: Trans. Jap. Inst.
Metals, 18 (1977), 353.
(8) R.L. Altman and H.H. Kellogg: Trans. Inst.
Min. Met., 81 (1972), C163.
(9) C. Diaz: The Thermodynamic Properties of
Copper-Slag Systems, Monograph III on the
Metallurgy of Copper, INCRA, (1974), p. 150.
(10) M. Hansen: Constitution of Binary Alloys,
McGraw-Hill, (1958), p. 580.
(11) R.W. Ruddle, B. Taylor and A.P. Bates: Trans.
Inst. Min. Met., 75 (1966), Cl.