Mechanisms of Hydrogen-Assisted Magnesiothermic Reduction of TiO 2 HYRUM LEFLER, Z. ZAK FANG, YING ZHANG, PEI SUN, and YANG XIA Direct reduction of TiO 2 powder has been attempted for decades by researchers in an effort to decrease titanium (Ti) metal production costs. The main objective has been to avoid energy-intensive steps involved in production of primary Ti by the Kroll process. The emerging hydrogen-assisted magnesiothermic reduction process, which uses Mg to directly reduce TiO 2 powder under a H 2 atmosphere, has been shown to have the potential to compete directly with the Kroll process. The present studies represent an effort to understand the reduction reaction mechanisms of this process. Phase transformations and the reaction pathways are examined by SEM/EDX analysis of partially reduced powder cross-sectional, X-ray diffraction, and other analytical techniques. The results show important morphological changes, the prominent intermediate and final phases under the H 2 atmosphere, as well as the local deposition behavior of the MgO byproduct. The effect of the specific surface areas of the initial particles is also discussed. https://doi.org/10.1007/s11663-018-1399-0 Ó The Minerals, Metals & Materials Society and ASM International 2018 I. INTRODUCTION TITANIUM (Ti) metal has long been valued as a highly useful metal for its high specific strength, high-temperature capability, corrosion resistance, and biocompatibility. [1,2] Although it is one of the more abundant elements in the earth’s crust, its application in most industries has been limited because of the high energy and high cost associated with refining it to pure metallic form. [3] This is primarily because of its high chemical affinity for oxygen. For decades, researchers have sought to find a less-expensive route for the production of Ti metal to replace the Kroll process, [4–6] where TiCl 4 (made by chlorinating concentrated titania minerals) is reduced by molten magnesium (Mg). For a scale of a few tons of titanium per batch, the reduction process typically takes several days, and the energy-in- tensive vacuum distillation step takes several more days to minimize residual Mg and Cl contents in the Ti-sponge product. [7] Although many approaches have been investigated, so far none have been able to replace the Kroll process. A succinct review on the characteristics of those different approaches was recently given by Zhang et al. [8] One method worth noting here is the FFC process, [9–11] where TiO 2 (or other titania precursors) instead of TiCl 4 is used as the feedstock material and is directly reduced via electrolysis in a molten calcium chloride electrolyte to produce Ti powder. [12–15] And another is the Armstrong pro- cess, [7,16] where TiCl 4 is continuously reduced by Na to form a sponge-like Ti powder. [17,18] Each of these technologies, among many other R&D processes, has shown advantages and disadvantages, but none of these approaches to date have been able to compete with the Kroll process commercially with respect to both quality and cost. [19] Another general approach that has long held promise for lower-cost Ti metal production is to thermochem- ically reduce TiO 2 directly using calcium (Ca) or Mg. [6,20] Direct TiO 2 reduction involves many interme- diate phases, as the Ti-O system is quite complex. [21] The use of Ca in this approach has been studied extensively and reported in literature. [22–26] And yet, to date, none of the calciothermic processes have been fully developed which is at least partially attributable to the high cost of Ca. Compared to Ca, Mg is more cost-effective, but is thermodynamically unable to produce Ti with suffi- ciently low oxygen content to meet industry stan- dards, [25,27–29] even with significantly excess Mg. [30] The thermodynamic limit in oxygen removal by Mg ranges from 2 wt pct (at approx. 800 °C) to 1 wt pct (at approx. 600 °C). [28] Fortunately, this thermodynamic limit can be lowered drastically in the HAMR process by introducing H 2 gas to the system, which assists by HYRUM LEFLER, Z. ZAK FANG, PEI SUN, and YANG XIA are with the Department of Metallurgical Engineering, University of Utah, Salt Lake City, UT 84112. Contact e-mail: zak.fang@utah.edu YING ZHANG is with the Department of Metallurgical Engineering, University of Utah and also with the Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China. Contact e-mail: zhangying@ipe.ac.cn Manuscript submitted December 27, 2017. METALLURGICAL AND MATERIALS TRANSACTIONS B