Is that so? I thought that for asymmetric airfoil, zero angle of attack is by definition the angle where it creates no lift. So, tautologically, if it's creating lift a (positive) angle of attack is required.