Low Temperature Phase Transition of Ilmenite during Oxidation by Chlorine

Transcription

1 Materials Transactions, Vol. 50, No. 8 (2009) pp to 2078 #2009 The Japan Institute of Metals EXPRESS REGULAR ARTICLE Low Temperature Phase Transition of Ilmenite during Oxidation by Chlorine Xiao Fu*, Yao Wang and Fei Wei Beijing Key Laboratory of Green Chemical Reaction Engineering and Technology, Department of Chemical Engineering, Tsinghua University, Beijing , P.R. China The phase transition of natural and synthetic ilmenite during oxidation by Cl 2 at C was characterized. The ilmenite phase was converted to crystallographic shear (CS) phases during the oxidation. One third of the Fe 2þ in the ilmenite phase was converted to FeCl 3. The remaining Fe 2þ was oxidized to Fe 3þ in the CS phases. The CS phase obtained from synthetic ilmenite was the Fe 2 Ti 3 O 9 (M 5 O 9 ) phase. A small amount of Mg 2þ impurity in natural ilmenite caused the resulting structure to be the M 13 O 23 structure when it was oxidized to the CS phase at temperatures up to 490 C, with the M 5 O 9 structure also formed at temperatures above 540 C. The presence of Mg 2þ was crucial in determining the particular structure of the CS phase and its thermal stability. These observations provide a useful way to get an oxidation product with a particular structure of the CS phase. They showed that there can be significant effects on the phase transition and thermal stability of the products from the presence of an impurity element. [doi: /matertrans.m ] (Received March 27, 2009; Accepted May 18, 2009; Published July 25, 2009) Keywords: phase transitions, ilmenite, low temperature oxidation, crystallography shear phase 1. Introduction Table 1 Chemical composition of natural ilmenite ore. (mass%) Natural ilmenite is commonly used as a raw material for producing titania pigment and Ti metal. An oxidation treatment is commonly used as a mineral processing method to modify the phase of ilmenite material for further utilization. 1 6) The phase transition when the Fe 2þ in ilmenite is oxidized to Fe 3þ has been investigated by a number of researchers. For the high temperature reaction of ilmenite with oxygen above 900 C, pseudobrookite and rutile are the equilibrium products. 1,7 17) At moderate and low temperatures, the phase transition product is determined by the oxidation medium and synthesis method. During electrochemical corrosion of ferrous ilmenite, pseudorutile is the oxidation product ) This can also be obtained by hydrothermal synthesis. 20) Rutile, haematite, and a series of ferric titanates with crystallographic shear (CS) structures can be formed together during the oxidation of ilmenite in dry oxygen or air. 3,9 17,22) The CS phases can be preferentially synthesized by the high energy ball milling method or fast annealing. 15,16) Chlorine gas (Cl 2 ) is another potential oxidation medium and is widely used in chlorination metallurgy. However, the oxidation of ilmenite with Cl 2 has not been sufficiently investigated. In this study, experiments and characterization were carried out to investigate the phase transition of ilmenite during oxidation by Cl Material and Experiment Both pure synthetic ilmenite and natural ferrous ilmenite were used as reactants. The synthetic ilmenite was provided by Alfa Aesar. This material was a fine and porous powder with a FeTiO 3 purity higher than 99.8%. The natural ilmenite used was supplied by Panzhihua Iron & Steel Co. (China). It has the chemical composition shown in Table 1. The x-ray powder diffraction characterization of this ilmenite showed the standard diffraction data of synthetic ilmenite (Fe 2þ TiO 3, *Graduate Student, Tsinghua University TiO 2 FeO Fe 2 O 3 MgO CaO SiO 2 Al 2 O PDF No ). The ilmenite particles were sieved, washed and dried at 100 C before the experiments. Three sieved samples of natural ilmenite were used. The mesh sizes were 100 to 120 mesh, 150 to 180 mesh, and 200 to 300 mesh. The average diameters were mm, 92.3 mm, and 59.9 mm, respectively. The experiments were carried out in a tubular quanz tube reactor with an inner diameter of 8 mm. Ilmenite samples loaded in the reactor were heated to the required oxidation temperature in an Ar atmosphere. Then Cl 2 (9.1 kpa, diluted with Ar) was flowed in to react with the ilmenite for a specific time. The oxidized samples were digested in a HCl/HF solution in N 2 atmosphere and the content of Fe 2þ was measured by redox titration to calculate the oxidation conversion of Fe 2þ. During oxidation, gaseous FeCl 3 was produced. The FeCl 3 was collected in a condenser outside the reactor and its amount was measured by spectrophotometry. The weight lost of the sample was also recorded. The yield of FeCl 3 was the ratio of the molar quantity of FeCl 3 to that of FeO in the ilmenite reactant. X-ray diffraction (XRD) analyses were carried out on a Rigaku 2500 diffractometer with Cu K radiation. High resolution TEM characterization was carried out on a JEM 2010 microscope with an EDS microanalysis system supplied by Oxford Instruments. The EDS system was calibrated before analysis. SEM photos were obtained on a JSM-7401 microscope. The amount of chlorinated Mg was determined by plasma atomic emission spectroscopy (ICP-AES, ICP Profile spectrometer, Leeman). 3. Results and Discussion 3.1 Oxidation of synthetic ilmenite As a reference study, the oxidation of synthetic ilmenite with Cl 2 was conducted to investigate the phase transition

2 2074 X. Fu, Y. Wang and F. Wei Table 2 Oxidation of synthetic ilmenite for 60 min at different temperatures. Temperature ( C) Weight loss (%) Fe 2þ conversion (%) FeCl 3 yield (%) Fig. 1 XRD patterns of synthetic ilmenite samples oxidized with Cl 2 at 440 to 560 C for 60 min, and a sample roasted at 700 C in air for 60 min after oxidization at 560 C. behavior when there is no impurity element in the ilmenite lattice. The data for oxidation occurring in the low temperature range from 440 to 560 C are shown in Table 2. Redox titration analysis showed that most of the Fe 2þ was oxidized after 60 min reaction. Along with the oxidation of Fe 2þ, FeCl 3 was generated as a gaseous product, and the sample lost weight. It was determined that the yield of FeCl 3 was always one third of the Fe 2þ conversion. This was because O 2 ions cannot be replaced by Cl below 600 C. 23,24) Two third of the Fe 3þ remained in combination with O 2 to meet the charge balance in the oxide lattice. The weight loss determined was in good agreement with the measured amount of Fe 3þ chlorinated. Therefore, the oxidation conversion of Fe 2þ can be obtained from the FeCl 3 yield. Since this indirect measurement method was more accurate than the redox titration, it was used to analyze the oxidation conversion in the following. TiCl 4 was not detected in the gaseous product, which is in agreement with the chlorination of titanium oxide being thermodynamically infeasible. The XRD patterns of the samples after oxidation are shown in Fig. 1. Disregarding the peaks of unreacted ilmenite and weak diffraction peaks of the rutile phase, the diffraction peaks of the oxidation product was in accordance with the reference data of CS structure phases. 22,25) The broad diffraction peaks for the sample treated at 440 C indicated its poorly ordered structure. The homologous series of the CS structure phases can be considered as the intergrowth of M 3 O 5 (V 3 O 5 type) and M 2 O 4 (-PbO 2 type) slabs, where M represents a metallic ion. All the members of this series are based on a hexagonal close-packed (hcp) oxygen lattice with different metal atom distributions in the octahedral sites. For the Fe-Ti-O system, the formation of Fe 2 Ti 3 O 9 (PQ) 15,16,22,26) and Fe 2 Ti 2 O 7 (2PQ) 3,12,13) has been reported during the oxidation of ilmenite with O 2. For identifying particular homologues, the position of the XRD peak between 18.9 and 22.8 was used, which was suggested by Pownceby et al. 27) This peak is due to diffraction by the crystal plane (0; 0; nðp þ qþ), where n ¼ 1 for p even and n ¼ 2 for p odd. Its position showed the biggest difference in the XRD pattern of the series. By this method, the structure of the synthetic ilmenite oxidized with Cl 2 was found to be close to Fe 2 Ti 3 O 9. This had the characteristic peak at This result can be understood by considering that the ilmenite lattice had lost one third of the Fe ions. The reaction between synthetic ilmenite and Cl 2 is shown in eq. (1). 6FeTiO 3 (s) þ 3Cl 2 (g)! 2Fe 2 Ti 3 O 9 (s) þ 2FeCl 3 (g) ð1þ In order to examine the thermal stability of the Fe 2 Ti 3 O 9 phase obtained using Cl 2, the sample after oxidization at 560 C was subsequently roasted in air. As shown in Fig. 1, the Fe 2 Ti 3 O 9 phase decomposed to rutile and haematite after roasting at 700 C for 60 min. By contrast, the Fe 2 Ti 3 O 9 phase obtained by oxidizing synthetic ilmenite with O 2 was stable at temperatures up to 770 C. This agrees with the report by Grey et al. 22) The low thermal stability of the Fe 2 Ti 3 O 9 phase formed with Cl 2 indicated its poorer crystallinity. 3.2 Oxidation of natural ilmenite Low temperature oxidation reaction The natural ilmenite was oxidized with Cl 2 in the temperature range from 390 to 560 C. The data are shown in Fig. 2. The conversion of Fe 2þ was calculated from the FeCl 3 amount generated in the reaction. For some of the oxidized samples, titration analysis was also carried out, which confirmed that the yield of FeCl 3 was also one third the Fe 2þ conversion. The Cl 2 flow rate used was much in excess of the consumed amount so that the experiment results were not affected by the Cl 2 pressure change in the packed bed. At 560 C, 96% of the Fe 2þ ions were oxidized after 60 min treatment. By applying the method developed by Flynn, 28) the activation energy E a was calculated with a kinetics model. From the slopes of the fitted lines, the average activation energy was 71 kjmol 1. In contrast, the activation energy obtained by Fouga et al. for the reaction between ilmenite and Cl 2 at temperatures from 650 to 850 C was as large as 186 kjmol 1. The big difference in the activation energy indicated that the reaction was different in these two temperature range. At lower temperatures, Cl 2 only oxidized Fe 2þ ions, while the chlorination of ferric oxides occurred at higher temperatures. Figure 3 shows the effect of particle size on the oxidation rate of natural ilmenite at 490 C. The oxidation rate decreased with increased particle size, esp. in the initial stage of reaction. This indicated that the oxidation reaction

3 Low Temperature Phase Transition of Ilmenite during Oxidation by Chlorine 2075 Fig. 2 Effect of oxidation temperature on the conversion of Fe 2þ in natural ilmenite ( mesh) treated with 9.1 kpa Cl 2. Fig. 4 XRD patterns of natural ilmenite ( mesh samples) oxidized at temperatures from 390 to 560 C for 60 min. The inset graph shows the patterns from 20 to Fig. 3 Effect of particle size on the conversion of Fe 2þ in natural ilmenite treated with 9.1 kpa Cl 2 at 490 C. was diffusion controlled for gas transport through the ilmenite particle. Although the oxidation rate at 490 C was relatively lower, a high conversion of 94% could also be obtained with prolonged reaction time using a smaller sized sample ( mesh). The natural ilmenite material used in this study contained a high content of Mg, which was also partially converted to MgCl 2 during the reaction. By ICP analysis, the conversion of Mg was determined to be around 15% when the complete conversion of Fe 2þ was reached. This meant that a majority of Mg remained in the oxides to form the product phase Phase transition The XRD patterns of the pristine materials and samples ( mesh) treated at temperatures from 390 to 560 C for 60 min are shown in Fig. 4. The phase transition to form CS phases also occurred with the natural ilmenite sample. Using the method described in paragraph 3.1, a different CS phase was found and assigned the structure of M 13 O 23 (3P2Q), which has the characteristic peak at ,27) This was found in the oxidized samples at all the different temperatures used. In contrast, the M 5 O 9 (PQ) phase with the characteristic peak at only appeared when the temperature was 540 C and higher. The rutile phase was not detected in the oxidized natural ilmenite samples. The weak peak at 18.9 indicated the presence of the geikelite phase. The other diffraction peaks of geikelite were overlapped by the ilmenite phase. Ilmenite samples with different oxidation conversions were further analyzed to study the phase transition process. The mesh samples treated at 490 C and mesh samples treated at 560 C were used. The results are shown in Fig. 5(a) and 5(b). With increased oxidation conversion, the diffraction peaks of the ilmenite phase decreased in intensity while the peaks of the CS phases became more prominent. At 490 C, only M 13 O 23 was formed. At 560 C, there was no M 5 O 9 in the product at the beginning of the oxidation, but its characteristic peak at appeared when the conversion reached 56%. The diffraction peaks of rutile were not seen in any of the samples, even at complete oxidation. This observation indicated the CS phases were stable in a Cl 2 atmosphere at the oxidation temperatures Elemental composition In this study, the elemental compositions of the pristine ilmenite materials and the oxidized samples were investigated. The most important difference between synthetic and natural ilmenite was the presence of impurity elements in the crystalline lattice. This was the key factor for the appearance of the M 13 O 23 structure obtained from natural ilmenite. The pristine ilmenite samples were carefully ground to fine grains smaller than 1 mm and then characterized by TEM. The grains were first examined by electron diffraction to confirm that there was no impurity phase but only the ilmenite phase. Then, the ilmenite grains were analyzed by selected area

4 2076 X. Fu, Y. Wang and F. Wei Table 3 Metal element composition of the ilmenite materials. (atom%) Synthetic ilmenite Natural ilmenite Ti Fe Mg Fig. 5 XRD patterns at different conversions of (a) mesh and (b) mesh natural ilmenite samples oxidized at (a) 490 C and (b) 560 C. The insets show the patterns from 20 to EDS. The metal element compositions of both the synthetic and natural materials are shown in Table 3. The composition of synthetic ilmenite was in good accordance of the FeTiO 3 formula. For the natural ilmenite, Mg element was observed in amounts that vary from 1.2 to 19.2% in different grains. The amounts of other impurity elements, such as Ca, Si and Al, were quite low. This indicated that Mg was the major impurity element in the ilmenite lattice, while Ca, Si, and Al mainly existed in other phases. In addition, it was noted that the sum of the Mg and Fe amounts was always significantly higher than that of Ti. Therefore, the ilmenite phase in the natural material was actually a solution of FeTiO 3, MgTiO 3 and Fe 2 O 3. The natural ilmenite samples treated at 440, 490 and 560 C for 240 min were used to analyze the elemental compositions of the oxidation product phases. Ca, Si, and Al were found to remain in the impurity phases, and Mg always appeared together with Fe and Ti. Since the valence state of Fe in the CS phase was +3, we have plotted the data on the phase diagram of TiO 2,Fe 2 O 3 and MgO, as shown in Fig. 6. The element compositions of several grains fall on the line of the solid solution of Fe 2 O 3 -MgTiO 3, which matched the composition of geikelite. A majority of the points were concentrated in a small area close to Fe 2 Ti 3 O 9 and Fe 6 Ti 7 O 23. This area is shown enlarged to show the details. As reported above, the CS phases in oxidized natural ilmenite were M 5 O 9 and M 13 O 23. By considering the charge balance in the Mg 2þ -Ti 4þ -Fe 3þ -O 2 system, the M 5 O 9 phase would be a solid solution of Fe 2 Ti 3 O 9 -MgTi 4 O 9. The M 13 O 23 phase would be a solid solution of Fe 6 Ti 7 O 23 -Mg 3 Ti 10 O 23.In the oxidation products of natural ilmenite, the elemental compositions were distributed only along the line for Fe 6 Ti 7 O 23 -Mg 3 Ti 10 O 23 for samples oxidized at 440 and 490 C. This result is in agreement with the M 13 O 23 structure deduced from the XRD results. It indicated that Mg element existed in the lattice of M 13 O 23 and formed a solid solution. After oxidation at 560 C, the elemental composition deviated from that of the Fe 6 Ti 7 O 23 -Mg 3 Ti 10 O 23 line. It corresponded to that of Fe 2 Ti 3 O 9. This is in agreement with the appearance of the M 5 O 9 phase in the XRD analysis. In addition, in all the grains, Mg 2þ was seen. This indicated the M 5 O 9 phase also contained Mg 2þ and cannot be simply Fe 2 Ti 3 O 9. For contrast, the results of synthetic ilmenite treated at 560 C for 60 min are also shown. These plots matched the theoretical formula Fe 2 Ti 3 O 9. The above observations help explain the formation of the M 13 O 23 structure. Due to the presence of Mg ions in the initial ilmenite lattice, which remained during the phase transition, the structure of the CS phase was modified. In the Fe 2 O 3 -Cr 2 O 3 -TiO 2 and Cr 2 O 3 -Fe 2 O 3 -TiO 2 -ZrO 2 systems, the key effect of the elemental composition on the resulting crystalline structure of the CS phases has been shown by previous researchers. 22,25,27) In this study, Mg 2þ ions were shown to give a similar effect during the oxidation of ilmenite with Cl 2. It can also be noted that the Mg 2þ content was much lower in the CS phase in the oxidized ilmenite sample. It is likely that the geikelite phase was formed to accommodate

5 Low Temperature Phase Transition of Ilmenite during Oxidation by Chlorine 2077 Fig. 6 Element composition plots of the oxidized product in the phase diagram of the oxides. Solid lines represent the solid solution of end-point components. The bottom figure is the area in the dash rectangle enlarged. Fig. 7 XRD patterns of natural ilmenite samples after treatment with Cl 2 at (a) 490 and (b) 560 C and subsequently roasting at 700, 800 and 900 C. The insets show the patterns from 20 to excess Mg 2þ ions. The geikelite phase was also observed in the XRD patterns, as reported in paragraph Thermal stability The natural ilmenite samples oxidized at 490 C and 560 C until Fe 2þ conversions of 94% and 96% were subsequently roasted at 700, 800 and 900 C to study their thermal stability. The XRD patterns of the samples after roasting are shown in Fig. 7. It can be seen that the CS phase did not decompose up to 800 C. This was the case for both the M 5 O 9 or M 13 O 23 structures. Furthermore, the diffraction peaks, especially the peak at 26.4, were much more prominent after roasting, indicating a higher crystallinity. At 900 C, both M 5 O 9 and M 13 O 23 were converted to the pseudobrookite phase and a small amount of rutile. Thus, the thermal stability of the CS phase containing Mg 2þ was very much better than that of the Fe 2 Ti 3 O 9 phase obtained when synthetic ilmenite was oxidized by Cl Conclusion Ilmenite can be oxidized by Cl 2 at low temperatures. For both synthetic and natural ilmenite, one third of the oxidized Fe 2þ was converted to FeCl 3 during oxidation. The remaining Fe 2þ was oxidized to Fe 3þ in a solid phase in which the ilmenite phase was transformed into a CS phase. A Fe 2 Ti 3 O 9 phase with poor thermal stability was the product of synthetic ilmenite oxidized by Cl 2. The presence of Mg 2þ impurity in natural ilmenite significantly affected the phase transition behavior. For such natural ilmenite, the small amounts of Mg 2þ in the lattice of the CS phase caused the resulting structure to be M 13 O 23 (Fe 6 Ti 7 O 23 -Mg 3 Ti 10 O 23 solid solution) at temperatures up to 490 C, with M 5 O 9 also appearing at temperatures above 540 C. Both the M 13 O 23 and M 5 O 9 phases from natural ilmenite were thermally stable up to 800 C, while the M 5 O 9 phase from synthetic ilmenite decomposed to rutile and haematite at 700 C. These observations provide a useful method to obtain an oxidation product with a particular structure of the CS phase. Acknowledgement The authors thank project No supported by

6 2078 X. Fu, Y. Wang and F. Wei National Natural Science Foundation of China (NSFC), and National Basic Research Program of China (No CB219901). REFERENCES 1) M. G. M. U. Ismail, J. Amarasekera and J. S. N. Kumarasinghe: Int. J. Miner. Process. 10 (1983) ) K. Sun, R. Takahashi and J. Yagi: ISIJ Int. 33 (1993) ) G. Q. Zhang and O. Ostrovski: Int. J. Miner. Process. 64 (2002) ) E. A. Walpole and J. D. Winter: 32nd Annual Hydrometallurgy Meeting, ed. by E. M. L. Peek and G. Van Weert, (Metallurgical Society, Montreal, 2002) pp ) M. K. Sarker, A. K. M. B. Rashid and A. S. W. Kurny: Int. J. Miner. Process. 80 (2006) ) S. Itoh, T. Suga, H. Takizawa and T. Nagasaka: ISIJ Int. 47 (2007) ) J. L. Overholt, G. Vaux and J. L. Rodda: Am. Mineral. 35 (1950) ) S. Akimoto, T. Nagata and T. Katsura: Nature 179 (1957) ) D. B. Rao and M. Rigaud: High Temp. Sci. 6 (1974) ) D. B. Rao and M. Rigaud: Oxid. Met. 9 (1975) ) K. Borowiec and T. Rosenqvist: Scand. J. Metall. 10 (1981) ) S. K. Gupta, V. Rajakumar and P. Grieveson: Can. Metall. Quart. 28 (1989) ) S. K. Gupta, V. Rajakumar and P. Grieveson: Metall. Trans. B 22B (1991) ) R. A. Briggs and A. Sacco Jr.: Metall. Trans. A 24A (1993) ) Y. Chen: J. Alloy. Compd. 257 (1997) ) Y. Chen: J. Alloy. Compd. 266 (1998) ) M. Jablonski and A. Przepiera: J. Therm. Anal. Calorim. 66 (2001) ) G. Teufer and A. K. Temple: Nature 211 (1966) ) I. E. Grey and A. F. Reid: Am. Mineral. 60 (1975) ) I. E. Grey, C. Li and J. A. Watts: Am. Mineral. 68 (1983) ) M. T. Frost, I. E. Grey, I. R. Harrowfield and C. Li: Am. Mineral. 71 (1986) ) I. E. Grey and A. F. Reid: J. Solid State Chem. 4 (1972) ) G. G. Fouga, D. M. Pasquevich and A. E. Bohé: Trans. Inst. Min. Metall. C 116 (2007) ) N. Kanari, E. Allain and I. Gaballah: Thermochim. Acta 335 (1999) ) I. E. Grey, A. F. Reid and J. G. Allpress: J. Solid State Chem. 8 (1973) ) T. A. Sosedko and B. K. Kasatov: Doklady Akad. Nauk SSSR 250 (1980) ) M. I. Pownceby, M. J. Fisher-White and V. Swamy: J. Solid State Chem. 161 (2001) ) J. H. Flynn: J. Therm. Anal. Calorim. 34 (1988)