We use a single-ion electric-field noise sensor in combination with in situ surface treatment and analysis tools, to investigate the relationship between electric-field noise from metal surfaces in vacuum and the composition of the surface. These experiments are performed in a setup that integrates ion trapping capabilities with surface analysis tools. We find that treatment of an aluminum-copper surface with energetic argon ions significantly reduces the level of room-temperature electric-field noise, but the surface does not need to be atomically clean to show noise levels comparable to those of the best cryogenic traps. The noise levels after treatment are low enough to allow fault-tolerant trapped-ion quantum information processing on a microfabricated surface trap at room temperature. © 2014 American Physical Society.