Hydrogen-modified oxides for sustainable energy technology

Abstract

Titanium dioxide (TiO2) is a wide-band gap semiconductor with band gaps around 3 eV for the principal polymorphs rutile and anatase. Although TiO2 has been commercialized in applications that utilise its special chemical and optical properties, its band gap should be reduced to improve its performance, especially as an active photo catalyst. Band gap engineering of TiO2 has therefore attracted many researchers looking to extend its applicability as a functional material. Reduction of TiO2 introduces oxygen vacancies, initially forming disordered TiO2–x and eventually forming the ordered Magnéli phases TinO2n–1 ( n 1 x ), which have been commercialized in battery electrodes. Reduction of TiO2 under a hydrogen atmosphere is a promising method which can increase the visible-light absorption efficiency of TiO2, but the mechanism by which hydrogen exposure enhances its electrical properties is subject to controversy. TiO2 can also be reduced by carbon-containing gases, including CO, CO2 and CH4. Using methane as a reducing agent has the advantage of consuming a greenhouse gas in favour of producing oxygen deficient TiO2 for cocatalyst free methane decomposition processes. Reduced rutile is bluish in colour, while reduced anatase is black. Slightly reduced anatase or rutile nanoparticles have been reported to be yellow. Hydrogen-modified TiO2 In the first part of the project, carried out at Griffith University, the focus was on fundamentals, particularly the production and influence of O vacancies under vacuum and hydrogen, aiming to understand the action of hydrogen as a reducing agent. Oxygen deficient TiO2‒x was produced by exposing rutile to hydrogen at temperatures up to 500 C. Magnéli phases were produced by exposing rutile to vacuum at temperatures up to 1100bC. The absorption and desorption of hydrogen were studied by thermogravimetric analysis at temperatures up to 730 C, with simultaneous mass spectrometry measurements. The structural modifications caused by hydrogen absorption and desorption were confirmed by in-situ x-ray powder diffraction measurements at temperatures up to 1100 C. It was found that the Magnéli phases produced also absorbed hydrogen and desorbed higher amounts than hydrogen-modified rutile. Recent explanations of the enhancement of the electrical properties of hydrogen-modified TiO2 propose mid-band gap states just below the conduction band and, based on the absence of obvious structural changes in x-ray diffraction measurements, relate these to surface disorder. The reasoning behind this conclusion is that the volume of material subject to structural change must be too small to contribute noticeably to the measured diffraction pattern. On the other hand, x-ray diffraction is insensitive to light elements such as oxygen. In-situ high-resolution neutron powder diffraction with deuterium in place of protium was carried out to test this hypothesis, based on the high neutron scattering length of O relative to Ti. A small contraction of the unit cell was found, accompanying the introduction of oxygen vacancies. By refining the O occupancy, it was determined that TiO2–x with x = 0.2 (equivalent in stoichiometry to a Ti5O11 Magnéli phase) was formed under 50 bar of deuterium at 500 C. This indicates that vacancies are introduced throughout the volume of the TiO2 particle, because a surface-only structural change would not be resolvable. The sample was bluish in colour as is usual for reduced rutile. It therefore appears that the explanation of enhanced electrical properties owing to surface-only processes is wrong, or at best incomplete. Reduction and Carburization of TiO2 by Methane The second part of the project, carried out at University of California Santa Barbara, focused on applications of reduced TiO2. TiO2 reduction was studied using methane-containing gas (CH4-H2-Ar). In addition, catalytic decomposition of methane to hydrogen and carbon over reduced TiO2 surface was investigated. Oxide reduction using methane-containing gas occurs through adsorption and dissociation of methane with formation of adsorbed active carbon. Methane decomposition on metal oxides and solid solutions has been limited by carbon formation and deactivation. Carbon formation in the alkane dehydrogenation process is problematic because even small amounts of carbon can deactivate catalytic surfaces by physically blocking active sites. Methane pyrolysis experiments were performed in a lab-scale fixed-bed reactor and molten salt environment by flowing CH4 through a molten halide (LiCl-KCl eutectic mixture)/TiO2(Degussa P25) mixture. The highest degree of CH4 conversion (~34% initially) occurred at 1000 °C, but owing to catalyst coking and sintering fell quickly to ~20.0%. Temperature-programmed reaction (TPR) was also performed on the molten salt/additives mixture. The H2 yield of the LiCl-KCl/TiO2 mixture was not much higher than that of plain salt. The salt mixture turned yellow and TiO2 particles precipitated in the bottom of the reactor. Since it is known that CH4 reduces TiO2 to TiO2–x, the yellow colour of the molten salt/TiO2 mixture was likely due to the presence of TiO2–x. The higher density of the TiO2 particles relative to the molten salt, and ability to be wetted by the molten salt, caused them to settle in the reactor. It is concluded that the precipitation of TiO2 particles in the bottom of the reactor caused the low yield of the salt/catalyst mixture. More experiments should be done to confirm the catalytic activity and stability of TiO2 and Magnéli phases for methane pyrolysis.