122
J. Ru et al. / Journal of Alloys and Compounds 583 (2014) 121–127
3. Experimental
2. Thermodynamics analysis
3.1. Materials
The synthesis of TiN from FeTiO3 includes the preparation of
TiN–Fe composite from FeTiO3 by carbothermic reduction–nitrida-
tion and the removal of Fe form the prepared TiN–Fe composite by
acid leaching. The carbothermic reduction–nitridation process for
synthesizing TiN–Fe composite from FeTiO3 is a very complex pro-
cess which involves a series of intermediate reactions. The possible
reactions in the process and the corresponding changes in Gibbs
free energy are shown in Table 1. The proceeding sequence of the
reduction reactions depends on the CO partial pressure, while the
formation of TiN depends on both N2 and CO partial pressures. In
order to describe the effect of N2 and CO partial pressures on the
formation of TiN–Fe composite, the predominance area diagrams
for the system Ti–Fe–N–C–O at 800 °C, 900 °C and 1000 °C are
depicted by Outokumpu HSC Chemistry as shown in Fig. 1, where
PCO and PN2 are the equilibrium partial pressures of CO and N2 in
gas phase, and P° is standard atmospheric pressure.
Fig. 1 demonstrates that the TiN–Fe composite can be prepared
from the oxides by controlling CO and N2 partial pressures at
elevated temperatures, as shown in the shaded areas in the figures.
By comparing the shaded areas in Fig. 1, it can be seen that curve
abdefg shifts upward gradually as the temperature rises. This im-
plies that the rising of temperature is favorable for the formation
of TiN–Fe composite. In addition, the TiN formed is in equilibrium
with carbon and oxygen via reaction C + 1/2O2 = CO [28], hence the
residual oxygen and carbon as interstitial atoms may exist in the
TiN lattice, which will result in the formation of N, C and O contain-
ing phases, such as TiC, TiNxC1ꢂx (0 6 x 6 1), TiCyO1ꢂy (0 6 y 6 1)
and TiNxCyO1ꢂxꢂy (x + y = 1; x, y P 0). Due to lack of the thermody-
namic data for TiNxC1ꢂx, TiCyO1ꢂy and TiNxCyO1ꢂxꢂy, their possible
predominance areas are speculated in accordance with literatures
[28,29] and the principles of phase equilibrium, as shown in
Fig. 1. Lines ck, ij, bi and hi are the phase boundary between
The reduction experiments were conducted with mixtures of FeTiO3 and carbon
powders (99 wt% C). FeTiO3 powder was synthesized via co-precipitation combined
with microwave calcination method and the XRD revealed no evidence of other
phases being present. The particle size of FeTiO3 and carbon powders are both
ꢀ75
lm. High purity nitrogen (99.99 wt% N2) which was dried using silica gel
and anhydrous potassium hydroxide was used to maintain a reducing atmosphere
with a flow rate of 100 mL/min in all experiments.
3.2. Procedures
Before mixing, all the reagents were dried in an oven at 100 °C for 24 h to re-
move the moisture existed in the raw materials. The samples with a 1:3 M ratio
of FeTiO3 to carbon powder were thoroughly mixed and ground with 6 wt% polyvi-
nyl alcohol (PVA). Then the mixtures were pressed with a pressure of 10 MPa in a
closed die to obtain a cylindrical pellet (2 mm in height and 13 mm in diameter).
The reduction experiments were carried out in a tubular high-temp microwave
reactor (2.45 GHz, 3 kW) under nitrogen atmosphere. The pellets were placed in a
crucible which was enclosed in a quartz tube (100 cm in length and 8 cm in diam-
eter) and then subjected to microwave irradiation. The sample temperatures with
microwave heating were determined by inserting a metallic-sheathed Ni(Cr) ther-
mocouple into the center of sample. The pellets reduced were ground to ꢀ75
lm
and then leached by HCl solution in a mechanical-agitated vessel for 24 h at room
temperature. The agitation speed was held at 500 rpm and the ratio of solid to
liquid was maintained constant at 1:4 for all the leaching experiments. The initial
concentrations of HCl in leaching solution were 30 wt%. After filtering, the remain-
ing powder was washed several times with distilled water until the pH is about 7,
and then dried in a vacuum oven at 100 °C for 12 h.
3.3. Characterization
The products were analyzed by XRD (D/Max-2200 model) with Cu Ka radiation
at a scan rate of 10o/min in the range of 2h = 10–90°. The product X-ray diffraction
results were used for determining the lattice parameters of the TiN phase for a
given set of processing conditions. SEM (XL 30 ESEM TMP model) was used to
estimate the particle size of samples and characterize the morphology of products.
Ti3O5 and TiCyO1ꢂy, TiNxC1ꢂx and TiN, TiNxCyO1ꢂxꢂy and TiNxC1ꢂx
,
TiNxCyO1ꢂxꢂy and TiN, respectively. The values of x, y, and z are
determined by composition of N, C, and O. The locations of these
lines in the diagram depend on the reaction equilibrium constant
K and the values of x, y, and z. Fig. 1 also shows that the reduction
of iron oxides is dependent on CO partial pressure and independent
off N2 partial pressure, while the formation of TiN from TiO2 is
determined by CO and N2 partial pressures at the same time. For
example, TiO2 is firstly reduced to sub-oxides and then reacted
with carbon and nitrogen to form TiNxCyO1ꢂxꢂy, which is further
nitridized to TiN at 1000 °C when lg(PN2/P°) is lower than 2.88.
According to Fig. 1, the phase transition sequence of TiO2 as a
function of N2 partial pressure can be obtained, as shown in Table 2.
It is observed that the intermediate steps from TiO2 to TiN are
increased with decreasing in N2 partial pressure.
4. Results and discussion
As mentioned above, the carbothermic reduction–nitridation of
FeTiO3 is a complex process. In order to study the effect of micro-
wave heating on the phase changes occurring during the reduction
process, the XRD patterns of reaction products for FeTiO3:C = 1:3
after microwave-assisted carbothermic reduction–nitridation at
800 °C, 900 °C and 1000 °C for different reaction time (10, 20, 40,
and 60 min) are shown in Fig. 2. Depending on atmosphere, tem-
perature and reaction time, FeTiO3 can be reduced and nitridized
to different products such as Fe, TiO2, Ti2O3, Ti4O7, TiN. From
Fig. 2, it is found that FeTiO3 can be transformed into TiN–Fe com-
posite, which is in a good agreement with the results predicted by
thermodynamics.
Table 1
Equations for the changes in Gibbs free energy and lg(PCO/P°)–lg(PN2/P°) relationshipsa.
GoT (J)
Chemical reaction
lg(PCO/P°)–lg(PN2/P°) relationships
D
800 °C
900 °C
1000 °C
FeO + C = Fe + CO(g)
111505–154.87T
195676–188.94T
284019–251.51T
245699–208.90T
1101300–836.84T
622144–526.34T
lg(PCO/P°) = 0.61
lg(PN2/P°) = ꢂ2.04
lg(PN2/P°) = ꢂ3.35
lg(PN2/P°) = ꢂ3.77
lg(PN2/P°) = ꢂ4.18
lg(PCO/P°) = 1.24
lg(PN2/P°) = ꢂ1.06
lg(PN2/P°) = ꢂ2.13
lg(PN2/P°) = ꢂ2.54
lg(PN2/P°) = ꢂ3.09
lg(PCO/P°) = 1.78
lg(PN2/P°) = ꢂ0.27
lg(PN2/P°) = ꢂ1.06
lg(PN2/P°) = ꢂ1.51
lg(PN2/P°) = ꢂ2.18
9TiO2 + C = Ti9O17 + CO(g)
Ti9O17 + 5/4C = 9/4Ti4O7 + 5/4CO(g)
3Ti4O7 + C = 4Ti3O5 + CO(g)
Ti3O5 + 8C = 3TiC + 5CO(g)
2TiO2 + 4C + N2(g) = 2TiN + 4CO(g)
lg(PN2/P°) = 0.25ꢃlg(PN2/P°)–2.43 lg(PN2/P°) = 0.25ꢃlg(PN2/P°)–1.65 lg(PN2/P°) = 0.25ꢃlg(PN2/P°)–0.99
Ti9O17 + 17C + 9/2N2(g) = 9TiN + 17CO(g) 2603970–2179.58T lg(PN2/P°) = 0.27ꢃlg(PN2/P°)–2.46 lg(PN2/P°) = 0.27ꢃlg(PN2/P°)–1.68 lg(PN2/P°) = 0.27ꢃlg(PN2/P°)–1.03
Ti4O7 + 7C + 2N2(g) = 4TiN + 7CO(g)
Ti3O5 + 5C + 3/2N2(g) = 3TiN + 5CO(g)
2TiN + 2C = 2TiC + N2(g)
1031090–856.92T
711893–590.46T
259602–164.25T
lg(PN2/P°) = 0.29ꢃlg(PN2/P°)–2.38 lg(PN2/P°) = 0.29ꢃlg(PN2/P°)–1.65 lg(PN2/P°) = 0.29ꢃlg(PN2/P°)–1.03
lg(PN2/P°) = 0.3ꢃlg(PN2/P°)–2.31
lg(PN2/P°) = ꢂ6.21
lg(PN2/P°) = 0.3ꢃlg(PN2/P°)–1.60
lg(PN2/P°) = ꢂ4.97
lg(PN2/P°) = 0.3ꢃlg(PN2/P°)–1.01
lg(PN2/P°) = ꢂ3.92
a
For pure solid phases, the all activities of them are assumed to be 1. All thermodynamic data used above comes from Outokumpu HSC Chemistry.