Solar Metal Oxides Reduction under Vacuum, Experimental Comparison between Alumina and Magnesia Irina Vishnevetsky * , Rami Ben-Zvi, Michael Epstein Solar Research Facilities Unit, Weizmann Institute of Science, POB 26, Rehovot 76100, Israel * Irina.Vishnevetsky@weizmann.ac.il Aluminum and magnesium are metal elements with strong oxygen bonds and their oxides reduction demand high energy input in their production processes. Therefore, solar thermal carboreduction can be considered as a promising option when at least part of the necessary energy can be provided from concentrated solar radiation. A major difficulty is the high reaction temperature at atmospheric pressure that according to Le-Chatelier principle can be reduced in vacuum by decreasing the partial pressure of the product gases. This however, requires additional energy consumption for the pumping. Thermodynamic considerations associated with this process will be discussed along with experimental results obtained in a wide range of tests parameters. The main features predicted by thermodynamics can be formulated as follow:  Relatively high reaction temperature to achieve full alumina to aluminum reduction even at vacuum below 1 mbar compared to the temperatures required for carboreduction of magnesia [1] albeit it needs also high reaction temperature at atmospheric pressure (Fig1a).  Complicated chemistry of alumina carboreduction because by-products such as oxycarbides, carbides and volatile sub-oxide can be formed during the forward and/or reverse reactions [1, 2, and 3], whereas the chemistry of magnesia carboreduction is simple and demonstrates merely the presence of Mg and MgO.  Significantly higher condensation temperature of the aluminum vapor comparing to magnesium which can promote oxidation and carbonization of the reduced aluminum in reverse reactions (Fig.1b).  Higher volume of product gas released per unit weight of reduced Al compared to Mg (Fig.1c) that requires higher electrical energy consumption in pumping. a b c Figure 1: Minimal reaction temperatures for full conversion of oxides to metal as a function of product gas (CO) pressure (a); condensation/desublimation temperatures of metal vapors as a function of pressure (b); and amount of gas released per 1 kg reduced metal for alumina and magnesia carboreduction processes (c). Tests were performed using experimental setup based on fast induction heating (IH) as described in details in [1]. The setup in addition to GC was complemented with Siemens infrared analyzer Ultramat 23. The CO composition in the outlet gases was recorded every 10 sec. Useful oxygen