Critique of tight binding (orbital removal) method: Ideal vacancy and surface states