S. Tan et al. / Journal of Alloys and Compounds 475 (2009) 368–372
371
reduction of Ti is quite sluggish. After 6 h of electrolysis, the greater
portion of Ti is still tied up in CaTiO3.
Using the thermodynamic data, the decomposition voltages of
various oxides have been calculated and are listed below with
descending order. The values here were determined for the reac-
◦
tions taking place at 900 C, the anode product being CO(g).
Fe TiO5 + 3C = 2Fe + TiO + 3CO, ꢀE = 0.212
(2)
(3)
(4)
(5)
(6)
2
2
Fe TiO + 2C = 2Fe + TiO + 2CO, ꢁE = 0.051
2
4
2
FeTiO + C = Fe + TiO + CO, ꢁE = 0.036
3
2
TiO + 2C = Ti + 2CO, ꢁE = −0.784
2
CaTiO + 2C = Ti + CaO + 2CO, ꢁE = −1.015
3
It should be noted that with the reactions (2)–(4) which give
rise to Fe have reduction potentials more positive than those which
give rise to Ti, i.e. the reactions (5) and (6). The early formation of
Fe and the persistence of Ti bearing oxides until the very end of
electrolysis are consistent with the above values. It is also worth
noting that CaTiO has a higher decomposition potential than pure
3
TiO2 which implies that the reduction process is made difficult by
CaTiO3 formation.
Electrolysis, with the sequence of changes described above,
may be contrasted with the sintering process. While during sin-
tering Fe O3 and TiO2 are combined into more complex oxides, i.e.
2
Fe TiO5, with electrolysis, the structure first attacked probably by
2
CaO which is normally present in CaCl2 as it forms during the dry-
ing process, is then gradually reduced into simpler oxides yielding
first Fe and then Ti. Even though the structure formed by sintering
is later disintegrated, it appears that the state, which had evolved
during sintering, exercises a considerable effect on the nature of the
reduction process. This effect of sintering is most probably related
to the length scale of the chemical species in the oxide preforms. At
◦
Fig. 8. X-ray diffractograms of the oxide pellets sintered at 1300 C and deoxidized
◦
◦
900 C, the sintered pellet is grossly heterogeneous in that it con-
in CaCl2 at 900 C (3.2 V). The diffractograms refer to pellets; (a) in the sintered
condition, (b) sintered and immersed into the salt bath for 30 min, (c), (d) and (e)
refer to sintered pellets electrolyzed for 30 min, 6 h and 24 h respectively.
tains unreacted Fe2O3 and TiO2 phases as well as Fe2TiO5. With the
porosity of 47%, elements that are obtained as a result of deoxida-
tion, i.e. Fe and Ti, especially from the unreacted phases, are not
close enough to react with one another. Accordingly deoxidation
yields a range of phases, namely Fe, Ti (i.e. TiC see above) as pure
The sequence of changes that occurred in the samples dur-
ing electrolysis can be seen in Fig. 8. The X-ray diffractogram
shows a systematic change from the sintered condition to that after
elements, and Fe Ti and FeTi as intermetallics.
2
the 24 h electrolysis. In the immersed sample, Fe TiO5 and TiO ,
At higher sintering temperatures, the reaction goes to comple-
2
2
i.e. the phases in the sintered sample, continue to be the major
constituents, but there are also other phases. These are ilmenite
tion. Thus all Fe O3 is consumed and the preform is made up of
2
two phases Fe TiO5 and TiO . The structure, therefore, is more
2
2
(
FeTiO ), a spinel phase (Fe TiO ), and a perovskite phase (CaTiO ).
homogenous than that given above. Even though this structure is
modified upon immersion into the salt bath, this modification prob-
ably occurs with a length scale that is comparable to the original
structure. Thus upon reduction, elements are close to each other
and with diffusion over small distances they react with each other
yielding the intermetallics. The result, at the end of electrolysis, is
that FeTi makes up the considerable portion of the product.
As a final remark, it should be pointed out that, with the current
3
2
4
3
After 30 min of electrolysis, the phases did not change much, though
CaTiO3 is much stronger and a metallic phase Fe is already present
in the sample. After 6 h, the Fe TiO5 and TiO2 are all consumed.
Fe is now the dominant phase accompanied by the intermetallic
Fe Ti. The target composition FeTi can be identified in this sample,
2
2
but it is far from being a major phase. It is interesting to note that
CaTiO3 continues to be the major phase in this sample. It appears
that beyond this, the electrolysis mainly consumes CaTiO and with
processing conditions, the phases Fe Ti and a small amount of Ti
3
2
the production of Ti as a result, the greater portion of the sample is
converted into FeTi, see Fig. 8(e).
The observations reported above imply that, from the oxide state
to the final composition FeTi, electroreduction follows quite a sys-
tematic route. It appears that the oxides are already modified upon
immersion into the salt bath even before the electrolysis. Thus a
(i.e. TiC) do form in the final product. Since the target composition
is FeTi, the formation of these other phases need to be minimized
or should be prevented altogether. Following the approach given
above, it may be suggested that one method of achieving this,
would be to make use of mixed oxides with refined structures, i.e.
to employ a processing route that would yield the mixed oxide pel-
let with as fine a structure as possible. The other route would be to
two-phase structure, i.e. Fe TiO5 and TiO , is in part converted into
2
2
a mixture of phases comprising CaTiO , FeTiO3 and Fe TiO
produce a single-phase pellet, namely ilmenite (FeTiO ). It appears
3
2
4.
3
With the progress of electrolysis, as the oxygen is discharged,
these complex oxides are consumed and instead the remaining oxy-
gen is tied up in simpler oxides such as FeO and CaO. It appears that
Fe is reduced quickly from the iron bearing oxides. In contrast, the
that the ilmenite approach, with Fe:Ti in 1:1 proportion, would be
particularly worthwhile since this would make the electrolysis pos-
sible in a variety of sintering conditions. For instance, pellets of
increased porosity could be electrolyzed which might accelerate