Ionic liquids were evaluated as potential extraction solvents for liquid-liquid extraction to remove vanadium (VO2+) and nickel (Ni2+) from petroleum. It was reasoned that ionic liquids had the potential to selectively extract metal ions by changing the dissociation equilibrium of the ionic metal-containing species and because it was anticipated that ionic liquids and petroleum would be immiscible. Eight different ionic liquids were evaluated for the extraction of diluted bitumen (similar to 170 mu g of V/g and similar to 70 mu g of Ni/g) and partially deasphalted bitumen. It was found that phase separation between the bitumen-containing phase and the ionic liquid phase was incomplete and that some of the ionic liquids measurably co-extracted bitumen. In one specific instance, a 1:1 wt/wt mixture of diluted bitumen and 1-ethyl-3-methylimidazolium acetate was found to be completely miscible. Single-stage metal extraction from the bitumen-containing phase was generally <20 mu g of V/g and <5 mu g of Ni/g, with slightly better extraction achieved in only a few cases. Multistage liquid-liquid extraction increased metal removal, but the rate of phase separation and sizable rag layer may make such operation difficult in practice. It was found that the ionic liquids tested did not measurably affect the dissociation equilibrium of the metal-containing species in the bitumen. Evidence was presented that vanadium and nickel extraction was likely governed by phase equilibrium only, which meant that it was the relative solubility of the metal-containing species in the two liquid phases that mattered. This implied that the ionic liquids investigated provided no specific benefits related to their ionic nature and that it was only the relative solubility of the metal-containing species that determined partitioning between the two liquid phases. Liquid-liquid extraction of petroleum with ionic liquids was no different from liquid-liquid extraction of petroleum with other solvents that are not ionic liquids.