But in the late 1980s, scientists noticed that the original kilogram was about 50 micrograms lighter than its brethren. Because mass measurements are relative, it's tough to determine whether the replicas are getting heavier or the original is getting lighter. [...] Over the last two decades, Cumpson and others have developed such a method. It's based not on manually scrubbing the metal chunks but on exposing them to ultraviolet light and ozone about once per decade. Recently, Cumpson fine-tuned the procedure and added another step: a pure water rinse to remove dust particles. The final recipe is now available online in the journal Metrologia. [...]
There's another kilo-contaminant, though, one that sticks too stubbornly to remove: Mercury atoms from trace amounts of free mercury in labs that once used mercury-based thermometers. "Mercury's a most unfortunate contaminant," Cumpson said. "It's difficult to remove, it forms a strong metallic bond with platinum."
It's scientific understatement like that that explains why you never hear anything about the "misadventure" with the platinum-iridium CONS cell in Paris back in 1987. The horror, the horror.