CH-102 (INORGANIC)

(c:\msoffice\winword\ch102\102inorg.doc)
General methods for extraction (isolation) of metals (elements)

The concentrated ore often requires conversion into a compound suitable for
conversion into the element. For production of crude metals, pyrometallurgical
processes are employed. In pyrometallurgy thermodynamic factors are of chief
importance. Since chemical equilibrium is rapidly reached at the high temperature
employed in these methods, kinetic factors need not be considered.

Oxide ores are directly reduced (smelted) to the metal. A variety of reducing
agents are used, carbon being the one most widely employed. Al, Si and hydrogen are
other reducing agents used. These are preferred to carbon, wherever the metal to be
isolated forms a carbide. The bulk of the impurities is removed by use of suitable fluxes
to form a slag.
in every case
Sulphide ores are invariably roasted to convert them to oxide and then reduced to
the metal. The roasting process also removes volatile impurities such as arsenic. For
thermodynamic reasons, an oxide rather than a sulphide is used for reduction.
Sometimes, the sulphides are oxidized to sulphate or some other soluble salt and the
metal obtained from it by chemical or electrolytic reduction.

The highly electropositive metals, such as the alkali and alkaline earth metals are
isolated by electrolytic reduction of their fused halides, since no suitable chemical
reducing agent is available and aqueous solutions could not be used due to reactive
nature of these metals. Less electropositive metals like Cr, Cu and Zn can be prepared
by electrolysis of concentrated aqueous solutions of their salts. An advantage of the
electrolytic over oxide reduction method for extraction is high purity of the
electrolytically isolated metal.

When neither oxide reduction nor electrolysis is suitable for chemical reasons,
reduction of a metal halide by reactive metals such as Mg, Na and Ca is resorted to e.g.
reduction of TiCl4 by Mg in an inert atmosphere.

Physico-chemical principles underlying the extraction of metals. Free

energy of formation of oxides :
Free energy considerations play an important role in the understanding of many
pyrometallurgical processes. The free energy of formation Gf of an oxide may be
considered as the standard free energy of the reaction :

Y
XM + O2 → MxOy .... (1)
2

In a similar manner, one can talk of free energy of formation of an oxide per
mole of O2 as free energy of the reaction.

2X 2
M + O2 → MxOy .... (2)
Y Y

1
If ∆Gf (standard free energy of formation) has a minus sign at a given
temperature then the oxide can be expected to form spontaneously from the metal and
oxygen. If ∆Gf has a positive sign, the oxide will be expected to decompose
spontaneously into its elements.

The heats of formation of a number of oxides, expressed per mole of O2 are

listed in Table-1.

Table-1 : Heats of formation of some metal oxides

Oxide ∆Hf (per mole of O2 (g) EO (Maq / M)

MgO -288 Kcal -2.37 volts
Al2O3 -266 Kcal -1.66 volts
ZnO -166 Kcal -0.76 volts
NiO -177 Kcal -0.25 volts
CuO -74 Kcal +0.34 volts
Ag2O -15 Kcal +0.78 volts

In the formation of a metallic oxide according to eqn. (3), the entropy change is
essentially that which occurs when one mole of gaseous phase, O2(g) is removed from
the system, provided both the metal and the oxide are in solid/liquid state and are not
vaporized.

2-X
M (Solid or molten metal ) + O2(g) → MxOy (Solid or molten oxide) ..... (3)
Y

There may be small changes in disorder, hence, entropy associated with the
change from the metal phase to the oxide phase but the main factor influencing ∆Sf is
the using up of the highly disordered gaseous phase. Hence, ∆Sf will be negative and
entropy change is roughly same for all metal systems, provided the boiling point of
neither metal nor oxide is exceeded. This change in entropy is approximately 50 cal.
mole-1 deg.-1.

Fig. 1 is a plot of ∆Gf in kJ/mole O2 against temperature for various metallic

oxides. Below the boiling point of various metals, slopes of all the graphs are same,
since in ∆Gf = ∆Hf - T ∆Sf, T ∆Sf factor is the same whatever the metal may be, if slight
changes of ∆Hf with temperature are disregarded. When boiling point of the metal is
exceeded, slope will increase since the reaction now involves a larger entropy change.
For instance, above 1100 oC, three moles of gas phase are converted into solid phase in
the reaction.

2Mg(g) + O2(g) = 2MgO(s) ......(4)

Above a certain temperature, ∆Gf becomes positive for some of the oxides. This
explains why mercury(II) oxide, for instance, decomposes spontaneously into its
elements when heated. The diagram predicts decomposition of tin (IV) oxide and zinc
oxide on strong heating, but it does not give much hope for obtaining, say, pure

2
magnesium by straight forward heating of the oxide to a high temperature. Thus, heat
alone is not sufficient to decompose oxides of most metals.

A comparison of Fig. 1 with the data given in Table-1 clearly indicates that those
metal oxides with least negative heat of formation have lowest stability towards heat.
Also stability order of the oxides of metals parallels roughly, but not exactly the order in
the redox series.

Reduction of metallic oxides :

Since many metallic oxides cannot be reduced by the action of heat alone, it is
necessary to employ a reducing agent for their reduction. Aluminium, carbon and
hydrogen are quite useful for such purposes. The ensuing discussion throws light on
how to find whether a given substance will reduce the oxide of a given metal or not.

Thermit process :
This method employs aluminium metal as reducing agent. Let us consider the
reaction between chromium(III) oxide and aluminium.

4 2
Al + O2 → Al2O3 ∆Ho = −266.0 Kcal at 25oC .... (5)
3 3

4 2
Cr + O2 → Cr2O3 ∆Ho = −179.8 Kcal at 25oC .... (6)
3 3

4 2 2 4
Al + Cr2O3 → Al2O3 + Cr , ∆Ho = −86.2 Kcal .... (7)
3 3 3 3

It may be seen that the reaction will be accompanied by evolution of heat.

Similarly, from a knowledge of ∆Gf values for the two oxides, ∆Gf for the reaction can
be shown to be negative. Fig. 1 shows the two curves obtained by plotting ∆Gf for the
two oxide against temperature.

At all accessible temperatures, ∆Gf is markedly negative and we would expect

the reaction to proceed. In fact, the entropy change, ∆So, of this reaction will be quite
small, since no gaseous products or reactants are involved and the effect due to different
structures of various phases will be very small. Hence, ∆Go will be approximately same
at room temperature, as well as at higher temperatures, but the reactions need to be
raised to a higher temperature to ‘trigger it off’. This can be done by priming it with
magnesium ribbon and barium peroxide. Once started, the reaction is highly exothermic
and very high temperatures will be reached. The reduction of oxides with aluminum is
called “thermit” process or “Goldschmidt” process. Even though this process could
be applied in principle, to all but most electropositive metals, it actually finds very little
application on an industrial scale as other reducing agents (e.g. carbon) cheaper than
aluminum are available. However, some manganese and chromium are produced in this
way.

3
Carbon as Reducing agent :

Since carbon is the only reducing element, which occurs native in large
quantities, it has been used on a large scale for reduction of metal oxides in the
extraction of metals. Carbon burns in oxygen at temperatures below 710oC to form
carbon dioxide but at higher temperatures it gives carbon monoxide as the product.

C(s) + O2(g) = CO2(g) ... (8)

2C(s) + O2(g) = 2CO(g) ... (9)

The formation of CO2 (by equation 8) is accompanied with little entropy change
as one mole of gaseous O2 forms one mole of gaseous CO2. The disappearance of well
ordered solid phase lattice of carbon has little effect on the entropy change, ∆So. For
this reaction, we have ∆Go = -94.1 Kcal at 25o C, and ∆So = +0.7 cal. deg.-1.

Carbon can also react with CO2 to form CO, which is favoured at high
temperatures since it involves an increase in the disorder, ∆S is positive because two
moles of CO are formed from only one mole of CO2(g).

C(s) + CO2 = 2CO(g) ... (10)

For this reaction (eqn. 10) we have, ∆Ho=+41.8 Kcal at 25oC, ∆So = +42.1 Kcal deg.-1

Adding equations (8) and (10), we see that carbon reacts with oxygen to form carbon
monoxide (by equation 9). For this reaction we have ∆Ho = -52.8 Kcal, ∆So = +42.8
Kcal. deg-1.

Equation (9) also has a favourable entropy change, which will make the entropy term
T∆S more significant at higher temperatures :

Fig. 1 shows that of the two possible reactions between carbon and oxygen, the one
which actually occurs at a given temperature is the one which has more negative ∆Go.
The bold line in the free energy graph indicates this. It is the +∆So for equation (9)
which causes free energy plot to slope downwards above about 710o C. It may also be
seen from Fig. 1 that carbon monoxide could be an effective reducing agent only at
temperature below 710o C according to the reaction :

CO(g) + O2(g) = 2CO2(g) ... (11)

In the region around 710oC, where the two reactions (8) and (9) have
approximately equal ∆Go values (the two lines actually cross), the products of
combustion will be a mixture of 1C0(g)+ CO2(g). The downward slope of carbon-oxygen
graph shows that carbon can reduce oxides of most metals provided the temperature is
high enough.

Consider the possible reaction:

4
MgO(s) + C(s) = Mg(s) + CO(g) .... (12)

for which ∆Ho = +117.4 Kcal at 25oC. Free energy change of the reaction can be
derived by considering separate ∆Go values for the two reactions.

2C(s) + O2(g) = 2CO(g)

(Subtract) 2 Mg(s) + O2(g) = 2MgO(s) ...... (13)

Fig. 1 shows that this reaction will have a value of ∆Go = O at a point where the two
lines representing (9) and (12) cross each other. The equilibrium constant equals unity
at ~1900oC (from the graphs). Above this temperature, ∆Gf for the reaction becomes
negative and the reaction is favored. It is to be noted that the formation of magnesium
vapor steepens the MgO curve and makes the temperature for reduction lower than it
would otherwise be.

The same conclusion can be arrived at in a qualitative manner in other ways also.
Reaction (11) is endothermic. Hence equilibrium will be according to the Le Chatelier
principle, shifted towards the r.h.s.. Formation of metal is thus favored at high
temperatures. Furthermore, when magnesium vapor is formed as a result of the
reduction of MgO by carbon, ∆S for the reaction is strongly positive and if T becomes
high enough for -T ∆S term to off-set the adverse positive value of ∆H the forward
reaction will become predominant.

Fig. 1 shows a complete diagram of the free energies of some metal oxide
systems, together with carbon and also hydrogen. The temperature at which the
‘carbon’ curve intersects the ‘metal’ curve is the temperature at which the equilibrium
constant for the reduction becomes unity. The diagram shows for instance, that TiO2
requires about 1800o C for reduction to titanium and ZnO only 1000 oC. In the case of
reduction of copper(II) oxide, the diagram predicts that carbon dioxide rather than
monoxide will be the main product of the reaction. as low temp will be needed

2CuO(s) + C(s) = 2Cu(s) + CO2(g) .... (14)

Hydrogen as a reducing agent :

Hydrogen is not a very effective reducing agent for obtaining metals from their
oxides, as shown by Fig. 1. The reason is that So for the reaction

2H2(g) + O2(g) → 2H2O(g) .... (15)

3 moles of gas , 2 moles of gas

is negative as r.h.s. is less disordered. Hence, the plot of ∆Go against T rises with
temperature, indicating that not many metal oxide plots are intercepted by H2 curve.
Thus, hydrogen reduces oxides such as those of copper(I) and copper(II), but not the
oxide of aluminium magnesium and calcium. Oxides of iron are reduced only with
difficulty. In case of magnetic oxide of iron, Fe3O4 an equilibrium is readily
established.

5
Reduction of sulfides and chlorides of metals :
Although carbon reduces oxides quite effectively, it is not so effective when
used directly on sulfides or chlorides of metals. In the case of chlorides, reaction is
prevented by the fact that entropy of the reaction is not favorable. (Fig. 2). This carbon
cannot be used as reducing agent in the case of chlorides and hydrogen has only a few
applications e.g. in reduction of SiCl4. Hence, recourse is made to reactive
electropositive metals, such as Mg or Na as in production of zirconium and titanium,
from their tetrachlorides.

Similar case occurs with sulfides, carbon forms only one stable sulfide, CS2 (b.p.
o
46 ). The situation is analogous to the case which would occur if carbon is incapable of
forming CO. The downward sloping part of the curve in Fig. 1 does not exist in Fig. 3.
Hence carbon will reduce sulfides directly, but also with considerable difficulty. As
may be seen from Fig. 3, free energies of formation of most of metal sulfides are greater
than those for CS2 and H2S, the former is in fact an endothermic compound. Thus,
neither H2 nor C is a suitable reagent for reduction of metal sulfides. This difficulty can
be overcome by using a more powerful reducing agent in a few exceptional cases e.g.
reduction of stibnite (Sb2S3) by scrap iron.
Sb2S3(s) + 3Fe(s) = 2Sb(s) + 3FeS(s) .... (16)

in reverboratory furnace when a fused ferrous sulfide is formed. Similar method is

applicable to bismuth.

In actual practice, sulfide ores are roasted by heating them in a stream of air.
This reaction converts the ore into oxide, with the formation of SO2(g), e.g. galena (PbS).
2 PbS(s) + 3/2 O2(g) = PbO(s) + PbS(s) + SO2(g) ...... (17)

Metallic lead may then be obtained by incomplete roasting to produce a mixture

of PbS and PbO, followed by heating in absence of air. This kind of reduction involving
partial oxidation followed by heating in absence of air is called autoreduction (or air
reduction).

PbS(s) + 2PbO(s) → 3Pb(s) + SO2(g) .... (18)

Such a method is possible for lead since the formation of PbO is not very
exothermic. It would not be possible for metals where ∆H of the oxide is strongly
negative. In such cases, a reducing agent such as carbon must be used.

Mercury(II) oxide, which occurs as “Cinnabar” is converted directly to the metal

when roasted,

HgS(s) + O2(g) → Hg(s) + SO2(g) .... (19)

This happens unlike the reaction observed for lead above because at the roasting
temperature, the free energy of formation of HgO is positive. Thus, HgO decomposes
into Hg and O2.
ssg/1.1.1998 (c:\msoffice\winword\ch102\102inorg.doc)