Iron oxides are among the most common materials in Earth's crust, and exhibit properties relevant to a wide range of applications including catalysis, biomedicine, and magnetism. The most stable iron oxide under atmospheric conditions is hematite (alpha-Fe2O3), which has received much attention as a promising material for photoelectrochemical (PEC) water splitting due to its stability in water, ecological agreeability and 1.9-2.2 eV bandgap. In theory, hematite can achieve a maximum solar-to-hydrogen efficiency of 15%, sufficient for practical application. However, its actual performance is hindered by a low absorption coefficient, short minority carrier lifetime, low conductivity and sluggish reaction kinetics. Furthermore, the most stable alpha-Fe2O3 surface, the (012) facet, is poorly understood at an atomic scale. In this thesis, the two terminations of the alpha-Fe2O3(012) surface and their reactivity were investigated by scanning tunnelling microscopy (STM), low energy electron diffraction (LEED), x-ray photoelectron spectroscopy (XPS), ultraviolet photoelectron spectroscopy (UPS), Auger electron spectroscopy (AES) and temperature programmed desorption (TPD). The data support the previously assumed bulk termination model for the oxidized surface, but none of the models proposed for the (2x1) surface reconstruction under reducing conditions appear to be correct. A new model for the (2x1) termination is therefore proposed. Reactivity of the surfaces to water fit previously published results in that H2O is found to partly dissociate on both surface terminations. Two distinct phases of hydroxyl ordering, depending on coverage, are found on the (2x1) reconstruction with STM and subsequently analysed with LEED, TPD and XPS. Preliminary data on oxygen adsorption, the effects of atomic hydrogen, as well as surface interaction with titanium and platinum adatoms is presented. The previous understanding of O2 adsorption on the surface is expanded in the temperature regime below 100 K. Low temperature physisorption of O2 is found to be a precondition for a transformation of the (2x1) surface.