Free Access
Issue
Metall. Res. Technol.
Volume 117, Number 1, 2020
Article Number 101
Number of page(s) 7
DOI https://doi.org/10.1051/metal/2019067
Published online 03 January 2020

© EDP Sciences, 2019

1 Introduction

Ti alloys play a significant role in the aerospace industry, including fan blades, ducts and the components of engines, due to their superior properties such as a high static strength, good damage tolerance, and their mechanical behavior at elevated temperatures (i.e., higher creep resistance) [1]. Among these properties, temperature capability above about 600 °C is one of the prime requirements for Ti alloys such as IMI834 (Ti–5.8Al–4Sn–3.5Zr–0.7Nb–0.5Mo–0.35Si–0.06C) and Ti-6242 (Ti–6Al–2Sn–4Zr–2Mo). Yet the presence of trace levels of impurities such as Fe, Ni, and Cr, which inflow by the Ti sponge, has a profound effect on their creep resistance. Because of this, many studies have attempted, to explain the mechanism of Fe, Ni, and Cr impurities in the creep behavior of Ti alloys [2,3]. For Ti sponge, the precursor for the Ti alloy produced is one of the main impurity inflow sources and significantly influences the alloy’s creep properties. Hence, there is a need to study the inflow mechanism of Fe, Ni, and Cr impurities in Ti sponges.

In current industrial practice, Ti sponge is produced by the Kroll process, which consists of two major steps: the reduction of TiCl4 by Mg and the separation of the MgCl2 byproduct by vacuum distillation [4,5]. In the manufacturing process, obtaining a high-purity Ti sponge and a low energy consumption are two critical points that have attracted a considerable amount of attention [6]. However, previous work has mainly focused on the impurity sources and finding effective methods for reducing the Fe, Ni, and Cr impurity contents, and the inflow process and inflow mechanism have not been studied. For example, Chervonyj et al. studied Fe, Cr, Ni, and Mn impurity sources in Ti sponges by thermodynamic analysis [7]. The mode by which impurities entered the sponge was determined on the basis of thermodynamic calculations, but no valid evidence was given in combination with production practice. Takahiro et al. reported that the reaction vessel and Mg are sources of Fe and Ni and found that the Ti sponge will be contaminated by Fe and Ni when a Ni-containing stainless-steel retort is used [8]. However, the reasons for these results were not thoroughly analyzed.

In the paper, we present the distributions of the concentrations of Fe, Ni, and Cr impurities in a Ti sponge mass and investigate the chemical and phase compositions of specimens taken from different parts of the mass. We attempt to describe the inflow process of Fe, Ni, and Cr impurities during the Kroll process and to find effective measures that can decrease the Fe, Ni, and Cr contents to meet aerospace applications.

2 Materials and methods

2.1 Materials

Purified TiCl4 and metallic Mg (Pangang Group Vanadium Titanium Resources Co., Ltd.) were used as the raw materials. TiCl4 was produced by the chlorination of Ti slag in presence of C and molten salt. Mg was produced by electrolysis of the molten salt in a 200-KA diaphragmless cell at 710 ± 10 °C.

The chemical compositions of TiCl4 and Mg are given in Tables 1 and 2, respectively. Since the boiling points of NiCl2 and CrCl3 are higher than that of TiCl4 (136 °C), they were easily removed through multiple distillations; as a result, the concentrations of Ni and Cr were lower than the detection limit. In the experiments, the sponge mass obtained for every batch was 7.5 tons; thus, in order to reduce TiCl4 completely, 30 tons of TiCl4 were consumed, and 13 tons Mg were fed.

Table 1

Chemical composition of TiCl4 (wt.%).

Table 2

Chemical composition of Mg (wt.%).

2.2 Experimetal process

All experiments were performed in a cylindrical stainless-steel (06Cr–18Ni–11Ti) reaction retort with an outside diameter of 1836 mm and an inner diameter of 1800 mm. Before use, the retort was coated with a Ti film by a titanizing process [9]. Two different setups were used for the reduction and vacuum distillation processes, as shown in Figure 1(a) and (b). In the reduction stage, liquid Mg was poured into the retort and heated to 800 °C. Thereafter, TiCl4 was fed into the reaction retort at a speed of 400 kg/h, and the temperature of the reaction zone was controlled in the range of 800–850 °C by air cooling. In the vacuum distillation stage, a vessel of the same size was assembled as a condenser by spraying water. The distillation retort was maintained at 1000 ± 10 °C in vacuum at a pressure of 0.1–100 Pa for 90 h.

thumbnail Fig. 1

Schematic of the experimental equipment: (a) reduction and (b) vacuum distillation furnaces: (1) lower drain apparatus, (2) retort, (3) electric heater, (4) cooling air, (5) TiCl4 feed unit, (6) liquid Mg, (7) electric heater, (8) mass required after reduction, (9) thermometers, (10) vacuum pipe, and (11) condenser.

2.3 Characterization methods

The chemical compositions of the purified TiCl4 and Mg and the Fe, Ni, and Cr impurity concentrations in the Ti sponges were determined via inductively coupled plasma mass spectrometry (ICP-MS). The locations of the specimens taken from different parts of the mass were shown in Figure 2. The morphology of the specimen at location (7-1) was tested by scanning electron microscopy (SEM). Microstructural and elemental analyses of the samples taken from locations (5-5), (6-1), and (1-2) were carried out using SEM and energy-dispersive X-ray spectroscopy (EDS). The elemental distributions for the specimens at locations (6-1) was acquired by mineral liberation analysis (MLA).

thumbnail Fig. 2

Locations of the specimens taken from the Ti sponge mass.

3 Results and discussion

3.1 Concentration distribution of Fe, Ni, and Cr impurities

The concentrations of Fe, Ni, and Cr impurities in the mass are shown in Figure 3(a)–(c), respectively. It is found that the impurity contents vary according to the specimen locations, while the trends in the distributions of concentrations are similar. The side and bottom parts have relatively higher contents of Fe, Ni, and Cr impurities compared with the central and top parts. This is in agreement with the results reported by Lee et al. [10], who found that the sides and bottom of the sponge mass should be separated in order to improve the quality of the product.

thumbnail Fig. 3

Distributions of the concentrations of Fe, Ni, and Cr impurities in the sponge mass.

3.2 Ti sponge morphology and impurities enrichment characteristices

The morphology of the specimen at location (7-1) is shown in Figure 4(a) and (b), which show the part near the center and that adhered to the retort wall, respectively. As shown in Figure 4(a) and (b), the two parts have a spongy porous structure and are composed of fine scale crystallites. However, the size of grains in Figure 4(b) is larger than that in Figure 4(a) because the part near the retort wall is at a relatively higher temperature.

SEM-EDS analyses of specimens at locations (5-5), (6-1), and (1-2) are shown in Figure 5, and the EDS results are listed in Table 3. It shows that the analysis spot (6-1) contains Fe, Mn, Ti, and Cr, and the concentration of Fe (4.07 wt%) is the highest. The particles observed in Figure 5(c) shows that testing spot contains Fe, Ni, Cr, O, Mg, and Si, and it protrudes from the surface. Two different types of impurities containing Fe, Ni, and Cr are observed; one has a smooth surface, such as the specimen at location (6-1), and the other exists in protruding particles and contains other elements such as Mg and O. In addition, specimens enriched with Fe, Ni, and Cr impurities are easily found at the bottom and side parts of the sponge mass, which is in good agreement with the results in Figure 3.

To investigate the enrichment characteristics of Fe, Ni, and Cr in the specimen which taken from location (6-1), MLA was performed. The results are shown in Figure 6. There are two types of particles: one containing metallic elements such as Fe, Ni, Cr, and Ti, as clearly shown in Figure 6, and another containing nonmetallic elemental O.

thumbnail Fig. 4

Morphologies of Ti sponge: (a) the part near the center and, (b) the part adhering to the retort wall.

thumbnail Fig. 5

EDS analysis results of specimens obtained from different parts of the sponge mass: (a) location (5-5) from the center, (b) location (6-1) from the side, and (c) location (1-2) from the bottom.

Table 3

Elemental compositions of the testing spots (wt.%).

thumbnail Fig. 6

Elemental enrichment characteristics of Fe, Ni, and Cr in the specimen taken from location (6-1).

3.3 Inflow process analysis

3.3.1 Ti sponge generation characterisces

As pointed out by Nagesh et al., TiCl4 is first reduced by Mg at the inner surface of the wall of the reaction retort above the liquid Mg level, and this process is an exothermic reaction [11]. Thus, the size of grains in Figure 4(b) is larger than that in Figure 4(a), indicating that the Ti sponge grew toward the center of the retort from the edge when TiCl4 is fed. When MgCl2 is exhausted, the Ti sponge adhering to the retort wall will sink owing to the loss of buoyancy. We speculate that the Ti sponge formed at the beginning lies at the bottom part of the mass, and the part adhering to retort wall after MgCl2 is exhausted most likely settles at a location near the crucible wall.

The distributions of the concentrations of Fe, Ni, and Cr impurities in Figure 3(a)–(c) indicate that these impurities are heterogeneously enriched. Although the bottom and sides have relatively higher impurity contents, the Fe, Ni, and Cr impurity concentrations in the central part are approximately 0.013, 0.015 and 0.008 wt%, respectively. SEM-EDS indicates that a lower amount of the oxide phase containing Ti, Fe, Ni, Cr, Mg, and O is formed, and a large amount of the alloy phase comprising Ti, Fe, Ni, and Cr may have formed.

3.3.2 Formation enthalpies calculation

To predict the affinity between the Fe, Ni, and Cr impurities and Ti, and confirm whether an alloy phase has formed, the standard enthalpies of formation of the alloys formed between Ti, Fe, Ni, Cr, and Mg have been calculated using the Miedema and extended Miedema models [12]. Since the Troop model considers the effect of an element added to a binary alloy, this model is chosen to calculate the enthalpies of the Ti–Mg–Fe, Ti–Mg–Ni, and Ti–Mg–Cr ternary systems [13]. Figure 7 shows the formation enthalpies of these systems; it is observed that the formation enthalpies are negative for the Ti–Fe, Ti–Ni, Mg–Ni, and Ti–Cr binary systems, which is in agreement with the results proposed by Boer [14], whereas they are positive for the Mg–Ti, Mg–Fe, Mg–Cr systems. The standard formation enthalpies of the Ti–Ni binary system are more exothermic than those for the Ti–Fe and Ti–Cr systems at the same molecular ratio, and the formation enthalpy of the Mg–Ni system is more exothermic than those of Mg–Fe and Mg–Cr systems. Figure 7 also shows that the formation enthalpies of the Mg–Ti–Fe, Mg–Ti–Ni, and Mg–Ti–Cr ternary systems are negative when atomic fractions of Mg, Ti, and Fe (Ni and Cr) are 10, 50, and 50% respectively. The results indicate that an alloy phase consisting of Ti and Fe, Ni, and Cr impurities will form during the reduction process, even though liquid Mg exists, because the formation enthalpies are negative in the zone that has a lower Mg content. This is consistent with the phase diagrams of the Ti–Fe, Ti–Ni, and Ti–Cr binary systems.

thumbnail Fig. 7

Formation enthalpies of the Ti–Mg–Fe, Ti–Mg–Ni, and Ni–Ti–Mg ternary systems.

3.3.3 Inflow process of Fe, Ni, and Cr impurities

The SEM-EDS and MLA results indicate that the stainless-steel (06Cr–18Ni–11Ti) reaction retort is one of the sources of the Fe, Ni, and Cr impurities, particularly for the sides of the sponge mass. This result is strongly supported by the EDS results of the retort shown in Figure 8 and Table 4. It demonstrate that content of Fe, Ni, and Cr impurities decreases gradually from the inner surface to the outside of the retort, and the decrease range of Ni impurity is lowest. For Mg–Fe and Mg–Cr binary systems, the formation enthalpies are positive, while for Mg–Ni binary system it is negative, hence there are strong affinity between Mg and Ni atoms, and they are miscible in liquid state. However, the force between Mg–Fe and Mg–Cr atoms are dominated by repulsive forces, they are almost immiscible in liquid state. Thus, the dissolution rate of Ni atoms in liquid Mg is higher than that for Fe and Cr. When the Ti sponge adheres to the retort wall, and Fe, Ni, and Cr impurity atoms have dissolved in liquid Mg, Fe, Ni and Cr atoms will form alloys with Ti, owing to the strong affinity between Ti–Fe, Ti–Ni, and Ti–Cr binary systems. Eventually, Fe, Ni, and Cr impurities are enriched at bottom and side parts of the sponge mass.

As discussed above, the inflow process of Fe, Ni, and Cr impurities consists of two steps: the dissolution of impurities in liquid Mg and the formation of alloys with the Ti sponge, as depicted in Figure 9. The dissolution rate of the impurities is significantly influenced by the temperature at the reaction sites and whether the inner surface of the retort is coated with a titanizing film as a barrier layer. Therefore, the temperature of the reaction zone and the uniformity and thickness of the coating are key to lowering the Fe, Ni, and Cr impurity contents. Additionally, to obtain higher quality Ti sponge with Fe, Ni, and Cr impurity contents that are nearly zero, Mg that has lower concentrations of Fe, Ni, and Cr impurities should be used, and stainless steel should not selected as the material for the reaction retort. For controlling the Fe content, the temperature of the reaction zone and the quality of coating film should be the primary focus, and the side and bottom parts of the sponge mass need be separated during the cracking process.

thumbnail Fig. 8

EDS analysis results of the retort from the inner surface to the outside.

Table 4

Elemental compositions of the testing spots (wt.%).

thumbnail Fig. 9

Schematic of the inflow process of Fe, Ni, and Cr impurities.

4 Conclusions

Fe, Ni, and Cr impurities are heterogeneously enriched. The bottom and side parts of the sponge mass have relatively higher impurity contents, and the central part of the mass has lower concentrations of Fe, Ni, and Cr impurities. The stainless-steel reaction retort is one of the sources of the impurities, and the impurity inflow process consists of two steps: the dissolution of impurities in liquid Mg and the formation of alloys with the Ti sponge. The reaction retort material, the temperature of the reaction zone, and the uniformity and thickness of the coating directly influence the concentrations of the Fe, Ni, and Cr impurities. For high-quality Ti sponge production, the inner surface of the reaction retort should not be made of stainless steel and must be coated with a Ti film as a barrier layer. The reaction temperature should meet the technical parameters, and the side and bottom parts of the sponge mass need to be separated during the cracking process.

Funding

This study was funded by the Science & Technology Department of Sichuan Province (Project: 2017GZYZF0039) and Pangang Group Vanadium Titanium Resources Co., Ltd.

Conflict of interest

The authors declare that they have no conflict of interest.

References

  1. R.R. Boyer, Aerospace applications of beta titanium alloys, JOM, 20–23 (1994) [CrossRef] [Google Scholar]
  2. H. Mishra, P. Ghosal, T.K. Nandy, et al., Influence of Fe and Ni on creep of near α–Ti alloy IMI834, Mater. Sci. Eng. A, 222–231 (2005) [CrossRef] [Google Scholar]
  3. R.W. Hayes, G.B. Viswanathan, M.J. Mills, Creep behavior of Ti-6Al-2Sn-4Zr-2Mo: I. The effect of nickel on creep deformation and microstructure, Acta Mater 50, 4953–4963 (2002) [Google Scholar]
  4. W.J. Kroll, The production of ductile titanium [J], Trans Electrochem Soc 78, 35–47 (1940) [CrossRef] [Google Scholar]
  5. L. Liang, D. Liu, H. Wan, et al., Removal of chloride impurities from titanium sponge by vacuum distillation[J], Vacuum 152, 166–172 (2018) [Google Scholar]
  6. C.R.V.S. Nagesh, C.S. Ramachandran, R.B. Subramanyam, Methods of titanium sponge production, Trans. Indian Inst. Met 61, 341–348 (2008) [CrossRef] [Google Scholar]
  7. I.F. Chervonyj, D.O. Listopad, Thermodynamic laws of impuries in the titanium sponge inflow during its production, Acta Mech Slovaca 13, 40–47 (2009) [CrossRef] [Google Scholar]
  8. L. Takahiro, N. Nobuo, A. Tadao, Establishment of the high purity titanium billet production method using titanium sponge produced by the Kroll process, Proceedings of the 13th World Conference on Titanium, TMS, San Diego, America, 2016, pp. 103–105 [Google Scholar]
  9. D.W. Lee, Method of reforming inner surface of reactor for manufacturing sponge titanium having high purity, Korea, 20100119668A[P]. 2010-11-10 [Google Scholar]
  10. J.C. Lee, H.S. Sohn, J.Y. Jung, Effect of TiCl4 feeding rate on the formation of titanium sponge in the Kroll process, Korean J. Met Mater 13, 40–47 (2012) [Google Scholar]
  11. C.R.V.S. Nagesh, C.S. Rao, N.B. Ballal, et al., Mechanism of titanium sponge formation in the Kroll reduction reactor, Metall. Mater. Trans B 35B, 65–74 (2004) [CrossRef] [Google Scholar]
  12. A.R. Miedema, P.F. Chatel, F.R. Boer, Cohesion in alloys-fundamentals of a semi-empirical model, Physica 100B, 1–28 (1980) [Google Scholar]
  13. G.W. Toop, Extented Miedema’s model for solid solution formation of ternary alloys, Tran. ALME 233, 850–855 (1965) [Google Scholar]
  14. F.R. Boer, R. Boom, A.R. Miedema, Enthalpies of formation of liquid and solid binary alloys based on 3d metals, Physica 101B, 294–319 (1980) [Google Scholar]

Cite this article as: Sheng Zhuo, Li Kaihua, Li Liang, Cheng Xiaozhe, Inflow process of Fe, Ni, and Cr impurities in Ti sponge during kroll process, Metall. Res. Technol. 117, 101 (2020)

All Tables

Table 1

Chemical composition of TiCl4 (wt.%).

Table 2

Chemical composition of Mg (wt.%).

Table 3

Elemental compositions of the testing spots (wt.%).

Table 4

Elemental compositions of the testing spots (wt.%).

All Figures

thumbnail Fig. 1

Schematic of the experimental equipment: (a) reduction and (b) vacuum distillation furnaces: (1) lower drain apparatus, (2) retort, (3) electric heater, (4) cooling air, (5) TiCl4 feed unit, (6) liquid Mg, (7) electric heater, (8) mass required after reduction, (9) thermometers, (10) vacuum pipe, and (11) condenser.

In the text
thumbnail Fig. 2

Locations of the specimens taken from the Ti sponge mass.

In the text
thumbnail Fig. 3

Distributions of the concentrations of Fe, Ni, and Cr impurities in the sponge mass.

In the text
thumbnail Fig. 4

Morphologies of Ti sponge: (a) the part near the center and, (b) the part adhering to the retort wall.

In the text
thumbnail Fig. 5

EDS analysis results of specimens obtained from different parts of the sponge mass: (a) location (5-5) from the center, (b) location (6-1) from the side, and (c) location (1-2) from the bottom.

In the text
thumbnail Fig. 6

Elemental enrichment characteristics of Fe, Ni, and Cr in the specimen taken from location (6-1).

In the text
thumbnail Fig. 7

Formation enthalpies of the Ti–Mg–Fe, Ti–Mg–Ni, and Ni–Ti–Mg ternary systems.

In the text
thumbnail Fig. 8

EDS analysis results of the retort from the inner surface to the outside.

In the text
thumbnail Fig. 9

Schematic of the inflow process of Fe, Ni, and Cr impurities.

In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.