On the possibility of using model potentials for collision integral calculations of interest for planetary atmospheres