In the Fray-Farthing-Chen (FFC) Cambridge process, a metal oxide is made at the cathode and polarised to a sufficiently negative potential in a molten salt, typically CaCl2 and LiCl, that the metal oxide is electro-deoxidised to produce metal and oxide ion, the latter then dissolving in and moving through the molten salt to discharge at the anode. Replacing an O-conducting molten salt such as CaCl2 with a mixture of MgCl2-NaCl-KCl, which has very poor solubility of the oxide ion, has been proposed to prevent severe corrosion of the graphite anode caused by the discharge of O2, for example when MoS2 is electrolysed in molten CaCl2, as an alternative to the development of expensive inert anodes or the use of solid oxide-ion conductors, ultra-high temperatures or a steep temperature gradient. Advantages of disengaging O2 from the carbon anode include a more stable anode, lower cost for materials and operation, minimum diffusion in the electrolyte, smaller anode polarisation and faster anode reaction, lower melting temperature, lower energy consumption, higher energy efficiency, no CO2 emissions, zero or low C in product, and useful anode-produced Cl2. Challenges include precipitation of MgO in cathode, O in produced metal, narrower working voltage windows, electrolyte consumption, use of acid to reduce MgO, need to convert MgO to MgCl, Cl2 emission control and so on. Preliminary demonstration and comparison between CaCl2 and eutectic MgCl2-NaCl-KCl (MNK) has shown a current efficiency of 93.0% and energy consumption of 1.4 kWh/kg for Ta2O5 with MNK, compared with 78.0% and 2.4 kWh/kg with CaCl2, amd fpr ZrO2 39.0% and 7.1 kWh/kg, compared with 35.4% and 10.5 kWh/kg with CaCl2; the reduction rate with MgCl2-based molten salts is fast, CO2 emissions low if not zero, and energy and C consumption also low, the C anode remaining sufficiently inert for Cl- discharge and precipitated MgO being prevented from blocking cathode pores by use of a metal oxide cathode with sufficie, In the Fray-Farthing-Chen (FFC) Cambridge process, a metal oxide is made at the cathode and polarised to a sufficiently negative potential in a molten salt, typically CaCl2 and LiCl, that the metal oxide is electro-deoxidised to produce metal and oxide ion, the latter then dissolving in and moving through the molten salt to discharge at the anode. Replacing an O-conducting molten salt such as CaCl2 with a mixture of MgCl2-NaCl-KCl, which has very poor solubility of the oxide ion, has been proposed to prevent severe corrosion of the graphite anode caused by the discharge of O2, for example when MoS2 is electrolysed in molten CaCl2, as an alternative to the development of expensive inert anodes or the use of solid oxide-ion conductors, ultra-high temperatures or a steep temperature gradient. Advantages of disengaging O2 from the carbon anode include a more stable anode, lower cost for materials and operation, minimum diffusion in the electrolyte, smaller anode polarisation and faster anode reaction, lower melting temperature, lower energy consumption, higher energy efficiency, no CO2 emissions, zero or low C in product, and useful anode-produced Cl2. Challenges include precipitation of MgO in cathode, O in produced metal, narrower working voltage windows, electrolyte consumption, use of acid to reduce MgO, need to convert MgO to MgCl, Cl2 emission control and so on. Preliminary demonstration and comparison between CaCl2 and eutectic MgCl2-NaCl-KCl (MNK) has shown a current efficiency of 93.0% and energy consumption of 1.4 kWh/kg for Ta2O5 with MNK, compared with 78.0% and 2.4 kWh/kg with CaCl2, amd fpr ZrO2 39.0% and 7.1 kWh/kg, compared with 35.4% and 10.5 kWh/kg with CaCl2; the reduction rate with MgCl2-based molten salts is fast, CO2 emissions low if not zero, and energy and C consumption also low, the C anode remaining sufficiently inert for Cl- discharge and precipitated MgO being prevented from blocking cathode pores by use of a metal oxide cathode with sufficie