Aluminum metal is manufactured by passing a large electrical current downward through a layer of molten salt, where aluminum ore has been dissolved, and a layer of molten aluminum beneath. The process uses tremendous amounts of energy -- 3 percent of all the world's electricity. Improving its efficiency would reduce costs, save energy, and reduce emissions. However, the efficiency is limited by a fluid dynamics disruption, the metal pad instability, which can drive an aluminum smelter out of control. In today's smelters, the instability is prevented by thickening the salt layer, but because salt is a poor electrical conductor, thickening it sacrifices efficiency and turns nearly 40% of the electrical energy into waste heat. However, simulations predict that adding an alternating component to the current or to nearby magnetic fields could prevent the instability even with a thin salt layer, thereby increasing efficiency and reducing cost. In this project, researchers will perform experiments to reproduce the instability in the laboratory, measure its growth rate, and seek to prevent it by adding an alternating current component and/or magnetic field, and subsequently determine optimal parameters for preventing it. Small, inexpensive laboratory experiments could pave the way for implementation in full-size smelters (involving about 50 tons of aluminum). If the efficiency of aluminum manufacture can be increased by even 10%, worldwide savings would be about $1 billion per year