We propose the no-recoil approximation, which is valid for heavy systems, for a double folding nucleus-nucleus potential. With this approximation, the non-local knock-on exchange contribution becomes a local form. We discuss the applicability of this approximation for the elastic scattering of 6 Li + 40 Ca system. We find that, for this system and heavier, the no-recoil approximation works as good as another widely used local approximation which employs a local plane wave for the relative motion between the colliding nuclei. We also compare the results of the no-recoil calculations with those of the zero-range approximation often used to handle the knock-on exchange effect.PACS numbers: 25.70.Bc, The double folding model has been widely used to describe the real part of optical potential for heavy-ion collisions [1,2,3]. The direct part of the double folding potential is constructed by convoluting an effective nucleonnucleon interaction with the ground state density distributions of the projectile and target nuclei. In the double folding model, the exchange contribution originating from the antisymmetrization of the total wave function of the system is customarily taken into account simply through the single nucleon knock-on exchange term. The exchange term leads to a non-local potential. Since it is cumbersome to handle the resultant integro-differential equation, a local approximation has usually been employed. In the past, many calculations have been performed along this line by introducing a pseudo zero-range nucleon-nucleon interaction to mock up the knock-on exchange effect [1,2,4]. The strength of the pseudo interaction has been tuned so as to reproduce exact results of the integro-differential equation for proton scattering from various target nuclei at several incident energies [4]. This approach, in conjunction with the (density dependent) Michigan-three-range Yukawa (M3Y) interaction [5,6], has successfully accounted for observed elastic and inelastic scattering for many colliding systems [1,2].Recently, a more consistent treatment for the exchange term has also been considered [3,7,8,9,10]. This approach obtains a local potential by employing a local approximation to the momentum operator (local momentum approximation) [11,12]. Since the local momentum depends explicitly on the potential itself, there arises the self-consistency problem, which however can be solved iteratively. Since the exchange potential is directly constructed from a given nucleon-nucleon interaction of finite range, this approach is more favorable than the zero range approximation. In fact, the finite range treatment for the exchange term has enjoyed a success in reproducing the experimental angular distributions for light heavy-ion scattering where the zero range approximation fails [7,8,13].Despite its success, however, there is a potential difficulty in this approach. That is, the iterative procedure for the self-consistent problem may not work in the classically forbidden region, where the local momentum is imaginary. Althoug...