Carbothermal reduction of Titanium monoxide (TiO)

Paper Carbothermal reduction of Titanium monoxide (TiO) Hanjung KWON and Shinhoo KANG Department of Materials Science and Engineering, Seoul National University, San 56-1 Sillim-dong Gwanak-ku Seoul 151-744, Korea The carbothermal reduction of TiO was investigated and compared with that of TiO 2 using a thermodynamic approach and experiments. Cubic-TiO powder was synthesized using a high-energy mill to understand the reduction and carburization process. During the reduction and carburization process, TiO is found to be changed into TiC and Ti 2O 3 simultaneously at low temperatures. This change is ascribed to the low thermodynamic stability of TiO compared with that of Ti 2O 3. The instability of TiO is found to result in a higher driving force for the carburization of TiO to TiC than that for relatively stable Ti 2O 3. 2008 The Ceramic Society of Japan. All rights reserved. Key-words : Carbide, Cermet, Carbothermal reduction, Titanium monoxide, Mechanical alloying [Received April 17, 2008; Accepted June 19, 2008] 1154 1. Introduction There are many methods available to prepare TiC material, such as Self-High temperature Synthesis (SHS), 1),2) mechanical alloying, 3),4) thermal plasma, 5) Mg-thermal reduction, 6) and carbothermal reduction. 7),8) Of various processes the carbothermal reduction of TiO 2 is most widely employed in the production of TiC powder. This method uses TiO 2 with carbon as starting materials to reduce and carburize TiO 2, producing CO and CO 2 gases. When this method is combined with mechanical alloying process, it could induce significant structural modifications, producing various non-equilibrium materials. The sequence of the reduction of TiO 2 has been discussed in a nitrogen-free atmosphere by Berger 9) as follows: TiO 2 Ti no 2n 1 (n > 10) Ti no 2n 1 (4 < n < 10) Ti 3O 5 (Ti 2O 3) Ti(C xo y). The Ti(C xo y) phase is reduced further and changed to TiC if the reduction is performed under vacuum. Lyubimov et al. 10) found that Ti 2O 3 would not be reduced by carbon to TiO. Rather it would be reduced and carburized directly to the solidsolution Ti(C xo y) phase. Thus, in the present study, we synthesized cubic-tio powder to understand the reduction and carburization behavior of TiO in comparison with that of TiO 2. In addition, we tried to find any benefit to use TiO in the production of TiC. The results were explained in terms of equilibrium thermodynamic approach. 2. Experimental procedure Cubic-TiO powder was produced using anatase-tio 2 (99+% purity, 43 μm ave. size) and Ti (99.5% purity, 74 μm ave. size). The mixture of TiO 2 and Ti in 1:1 molar ratio was milled for 1, 3, 5, and 20 h in ambient atmosphere using a planetary ball mill (Fritsch Pulverisette 5, Germany). A 250 ml WC-coated stainless steel jar and WC-Co ball of 4 mm in diameter were used in the milling process. The weight of the powder mixture was 33 g, resulting in a ball-to-powder ratio of 30:1. All milling was conducted at a rate of 250 rpm. The carbothermal reduction of TiO was investigated using carbon powder (1.65 μm ave. size). The TiO and C powder in 1:2 Corresponding author: S. Kang; E-mail: shinkang@snu.ac.kr molar ratio was mixed for 24 h in acetone using a horizontal ball mill with a WC-Co ball. The mixture in solution was then dried in an oven for 24 h. The dried mixture was heat-treated at 1100, 1150, 1300, and 1500 C for 2 h in a graphite vacuum furnace. TiO 2 and C were also heat-treated at 1300 C for 2 h to enable a comparison with TiO and C. The reduction study was also performed using anatase-tio 2 powder. To minimize the size effect of the starting powders on the reduction, anatase-tio 2 powder was milled under the same conditions as those used for TiO powder. The milled TiO 2 powder was then mixed with C at a 1:3 ratio for 24 h. After drying, the mixture of TiO 2 and C was heated to 1100 C and cooled immediately. TiO and C were also heat-treated under the same conditions for comparison. The powder synthesized in this study was analyzed using an X- ray diffractometer (M18XHF-SRA, Macscience, Japan), using monochromatized Cu Kα radiation (λ = 1.54 10 10 m). The powder morphology was studied using a scanning electron microscope (JSM 6360, JEOL, Japan). Gas evolution during the carbothermal reduction was measured using a mass spectrometer (QMS 403, NETZSCH, Germany) under a helium flow. Heat flow during carbothermal reduction was also studied using differential scanning calorimeter (STA409C, NETZSCH, Germany) under Ar flow. 3. Results and discussion 3.1 TiO formation X-ray diffraction (XRD) profiles of the milled mixture of TiO 2 and C are shown in Fig. 1. The cubic-tio phase formed and the (101) peak of anatase-tio 2 disappeared after 3 h of milling. The main peak (010) of Ti remained even after the mixture of TiO 2 and Ti had been milled for 5 h. The brittle nature of TiO 2 made the milling process much easier than that for Ti. Fig. 1(d) shows that the TiO 2 peaks are absent after 5 h of milling. It seems that WC peak of XRD data (Fig. 1(e)) is from the contamination of WC-Co balls during milling. TiO 2 phase is milled first and the milling of Ti is done slowly. Therefore, it is expected that the TiO precursor, which possibly forms at the Ti TiO 2 interface with a cubic crystal structure, is initially relatively rich in O. From a crystallographic viewpoint, the cubic precursor starts growing on (0001) of milled Ti as a 2008 The Ceramic Society of Japan

JCS-Japan habit plane. The milled TiO 2, which tends to transform to rutile from anatase during high-energy milling, 11) facilitates O diffusion to form stable cubic TiO. Figure 2 shows the powder morphology of the milled mixture of TiO 2 and Ti, revealing large platelets with small particles. It is likely that ductile Ti forms platelets with embedded oxide particles. The size of the platelets becomes smaller with milling time; the platelets are no longer observed after 20 h of milling. The embrittlement of milling particulates can be attributed not only to strain hardening of the metallic phase but also to the enhanced diffusion of O throughout the system due to oxidation from atmosphere. 3.2 Carbothermal reduction behavior of TiO We investigated the reduction process of TiO with carbon. The Fig. 1. XRD spectra of the milled powders of TiO 2 and Ti prepared by planetary milling for (a) 0, (b) 1, (c) 3, (d) 5, and (e) 20 h. XRD results presented in Fig. 3 show the phase formation in the TiO C mixture. When the mixture was heat-treated at 1100 or 1150 C for 2 h, we observed TiC and a Ti 2O 3 phase with corundum structure, along with the starting materials. This indicates that a portion of the TiO changes to TiC and Ti 2O 3 even at low temperatures. When the temperature was raised to 1300 C, the TiO phase disappeared completely and the amount of TiC increased. At this temperature, both TiO and Ti 2O 3 appear to change into TiC. To understand the reduction process of TiO, we examined the possible reactions involving C and O 2. When carbon is combined with oxygen, the following two reactions occur: 1 C (s) + (1) 2 O 2(g) CO (g) C () s + O 2(g) CO 2(g) The changes in the standard Gibbs free energy, ΔG, of the reactions in Eqs. (1) and (2) are expressed as a function of temperature; i.e., 111,700 87.65T (K) and 394,100 0.84T (K) (J/mole), respectively. 12) Therefore, carbon reacts with oxygen to produce CO 2 gas below 705 C. Beyond this temperature, the reaction in Eq. (1) dominates. Figure 4 summarizes the Gibbs free energy changes of various reactions at a standard state. For the transformation of TiO to Ti 2O 3, Eq. (3) below is possible. The change in the standard Gibbs free energy, ΔG, is expressed as 191,100 + 18.37T (K) (J/mole). 12) (2) 2TiO (s) + CO 2(g) Ti 2O 3(s) + CO (g) (3) The obtained change in free energy demonstrates that the reaction that produces Ti 2O 3 from TiO is favorable at any temperature because of the low formation energy of Ti 2O 3 compared with that of TiO ( 1130.9 compared with 401.2 kj at 1200 C, Fig. 2. SEM micrographs of the milled powders of TiO 2 and Ti prepared by planetary milling for (a) 1, (b) 5, and (c) 20 h. 1155

JCS-Japan Kwon et al.: Carbothermal reduction of Titanium monoxide (TiO) Fig. 3. XRD spectra of the powder of (a) TiO and C mixture, and carbothermally reduced powders at (b) 1100 C, (c) 1150 C, and (d) 1500 C for 2 h. Fig. 4. Standard Gibbs free energy change of the reactions related to the carbothermal reduction of TiO. 12),13) respectively). 12) The remaining TiO forms TiC as follows since Ti 2O 3 is much more stable than TiC: TiO (s) + 2C (s) TiC (s) + CO (g) (4) Figure 3(b) and (c) shows that TiC forms immediately from TiO at low temperatures (< 1100 C). The early-formed Ti 2O 3 phase also transforms to TiC at higher temperatures (1100 1500 C; see Eq. 5). Based on the results shown in Fig. 3(c) and (d), the following reaction is also feasible: Ti 2O 3(s) + 5C (s) 2TiC (s) + 3CO (g) (5) The changes in free energy indicate that the reduction of Ti 2O 3 to TiC is more difficult to accomplish than that of TiO to TiC. The metastable TiO phase changes to TiC more readily than does stable Ti 2O 3. Figure 4 shows the Gibbs free energy changes of Eqs. (4) and (5) with respect to the temperature. It can be found that the reaction temperatures of Eqs. (4) and (5) are ~1100, 1300 C, respectively in the case of p CO = 1atm (1.01 10 5 Pa). However, TiO was reduced in a vacuum furnace as stated above. The partial pressure of CO in the presence of C in the considered system is represented by the equation 222,200 151.86T (K) = RTlnP CO, which is the condition necessary for the reaction in Eqs. (4), (5). 12),13) Therefore, if the CO gas evolved from the reaction in Eqs. (4), (5) is continuously evacuated, the reaction temperatures of the reactions in Eqs. (4), (5) would decrease significantly (~910, 1060 C at p CO = 13.33 Pa). The unstable TiO nanocrystallites synthesized by high-energy milling also act to lower the reaction temperature. We also consider the possible reduction of Ti 2O 3 to TiO as follows: Ti 2O 3(s) + C (s) 2TiO (s) + CO (g) (6) According to the data (ΔG = 361800 192.83T (K), J/mole) of Eq. (6), the reaction in Eq. (6) does not occur below 1600C; consequently, the TiO phase cannot be formed during the reduction of TiO 2 at the standard state. 12) 3.3 Comparison of the reduction behaviors of TiO and TiO 2 A comparative study was undertaken with a mixture of anatase-tio 2 and C. The oxide was milled in the same way as that for the Ti TiO 2 mixture, and mixed with carbon. In our previous study, anatase-tio 2 transformed to rutile-tio 2 during high-energy milling, which is stable only at high pressure and temperature. 11),14) Figures 5(a) and (b) show the results of CO evolution measured using a mass spectrometer. The CO gas is detected from the mixture of the milled TiO 2 and C at a lower temperature (~100 C lower) than that from the mixture of TiO and C. CO degassing of TiO and TiO 2 are completed in a different manner as shown in Figs. 5(a) and (b). That is, in the TiO case, the evolution of CO gas stopped abruptly at around 1470 C, whereas in the case of TiO 2 a small amount of residual CO gas continuously evolved beyond 1500 C. The Differential Scanning Calorimeter (DSC) result in Fig. 5(c) reveals that the temperature range of the reaction in the TiO and C case is narrower than that in the TiO 2 and C case. The DSC results show a similar evolution of CO gas to that in Fig. 5(a and b). It can therefore be concluded that the reaction of TiO and C begins at a higher temperature and ends at a lower temperature than that of TiO 2 and C. Figure 6(a) and (b) shows the oxide phases detected by XRD after the mixtures of milled TiO and C, and TiO 2 and C were exposed to relatively low temperatures for a short time; they were heated to 1100 C and cooled immediately from the corresponding temperatures. The Ti 3O 5 and Ti 2O 3 phases formed at 1100 C in the TiO 2 and C case (Fig. 6(b)), confirming the results obtained by Berger. 9) The Ti 3O 5 phase was not observed in the TiO case (Fig. 6(a)). Furthermore, the absence of the TiC phase from the powder mixture in the TiO 2 case (Fig. 6(b)) might reflect the high temperatures required for the carburization of TiO 2 (higher than those required for TiO). In the case of TiO 2 and C heat-treated at 1300 C for 2 h (Fig. 6(d)), phases with higher oxygen contents than Ti 2O 3 are not observed, as with the TiO case; however, the amount of TiC in the powder is less than that in the TiO case. The above findings indicate that the reaction finishes at a higher temperature in the TiO 2 case compared with the TiO case. Considering both the evolution of CO gas shown in Fig. 5 and the XRD results in Fig. 6, the omission of the reduction process in TiO 2 Ti 3O 5 appears to explain the higher starting temperature in the case of TiO and C. Because the carburization progresses rapidly in the case of TiO and C, the reaction for carbide synthesis is completed at a lower temperature compared with the TiO 2 and C case. It appears that the two peaks in CO gas evolution in Fig. 5(a) reflect the carburization of TiO and Ti 2O 3, respectively, even 1156

JCS-Japan (a) (b) (c) Fig. 5. CO gas detected by a mass-spectrometer from (a) TiO and C and (b) TiO 2 and C, and (c) heat flow estimated by DSC during carbothermal reduction of TiO 2 and TiO. TiC phases were observed (Fig. 6(c)). At 1500 C, the TiC phase remained in the powder, without Ti 2O 3 (Fig. 3(d)). This indicates that the carburization of Ti 2O 3 occurs at the second peak of CO gas evolution in Fig. 5(a). It seems that the separation of reduction (TiO 2 Ti 4O 7 Ti 3O 5 TiO x, x 1) and carburization in TiO 2 and C are more definite than that in TiO and C. The CO gas evolution during the carbothermal reduction of TiO 2 (Fig. 5(b)) reveals a single large peak and a smaller broad peak at high temperatures. A comparison of XRD data (Fig. 6(b), (d)) with CO gas evolution indicates that the reduction and partial carburization occurs at the first peak, while carbide synthesis is achieved at the second. 4. Summary and conclusions Fig. 6. XRD spectra of the milled mixtures of (a) TiO and C and (b) TiO 2 and C exposed at 1100 C and cooled immediately from the corresponding temperatures, and (c) TiO and C and (d) TiO 2 and C exposed at 1300 C for 2 h. though there is a degree of overlap. The carburization temperature of TiO is lower than that of Ti 2O 3 (see Fig. 4). When the holding time at 1100 C is increased from 0 (Fig. 6(a)) to 2 h (Fig. 3(b)), the amount of TiC increases as a result of TiO carburization. It is therefore considered that TiC is mainly synthesized from TiO and C, although Ti 2O 3 is partly carburized. When TiO and C were heat-treated at 1300 C, only Ti 2O 3 and Cubic-TiO powder was synthesized from a mixture of TiO 2 and Ti using a high-energy mill. The carbothermal reduction of TiO was investigated and compared with that of TiO 2 using a thermodynamic approach and experiments. The following points summarize the conclusions of this study. (1) In the process of carburization, a portion of the TiO phase changes to Ti 2O 3. This outcome reflects the low thermodynamic stability of TiO compared with that of Ti 2O 3. (2) TiO shows a greater tendency to form TiC than does Ti 2O 3, possibly first forming intermediate Ti(C xo y). The instability of TiO resulted in a higher driving force for the carburization of TiO than that of relatively stable Ti 2O 3. (3) Carbothermal reduction in TiO and C begins at a lower temperature and ends at a higher temperature than that in the 1157

JCS-Japan Kwon et al.: Carbothermal reduction of Titanium monoxide (TiO) TiO 2 case due to the omission of the reduction process and early formation of carbide. Acknowledgements This work was supported in part by the MOCIE research fund through the National R&D Project for Nano Science and Tech. under contract #10022970-2007-22 and in part by grants-in-aid for the National Core Research Center Program from MOST/KOSEF (No. R15-2006-022-03001-0). References 1) C. Deidda, F. Delogu, F. Maglia, U. Anselmi-Tamburini and G, Cocco, Mater. Sci. Eng. A, 375 377, 800 803 (2004). 2) C. Benoit, H. Ellen, K. Nikhil, V. Dominique and D. Sylvain, Powder Technol., 157, 92 99 (2005). 3) Z. Xinkun, Z. Kunyu, C. Baochang, L. Qiushi, Z. Xiuqin, C. Tieli and S. Yunsheng, Mater. Sci. Eng. C, 16, 103 105 (2001). 4) M. Razavi, M. R. Rahimipour, A. Hossein and R. Zamani, J. of Alloys Compd., 436, 142 145 (2007). 5) L. Tong and R. G. Reddy, Scripta Mater., 52, 1253 1258 (2005). 6) D. W. Lee and B. K. Kim, Scripta Mater., 48, 1513 1518 (2003). 7) L.-M. Berger, W. Gruner, E. Langholf and S. Stolle, Int. J Refract Met. Hard Mater., 17, 235 243 (1999). 8) L.-M. Berger, E. Langholf, K. Jaenicke-Robler and G. Leitner, J. Mater. Sci. let., 18, 1409 1412 (1999). 9) L.-M. Berger, J. Hard. Mater., 3, 3 15 (1992). 10) V. D. Lyubimov, S. I. Alyamovskii and G. P. Shveikin, Zh. Neorg. Khim., 26, 2314 2322 (1981). 11) H. Kwon and S. Kang, Met Trans., in print (2008). 12) D. R. Stull and H. Prophet, JANAF thermochemical tables, Ed. by U. S. Government Printing Office, Washington D. C. (1971). 13) D. R. Gaskell, Introduction to the Thermodynamics of Materials, 3rd ed., Taylor & Francis, Washington D. C. (1995) pp. 369 376. 14) J. L. Murray and H. A. Wriedt Bull., Alloy Phase Diag., 7[2], 148 165 (1987). 1158