On the possibility of finding a suitable potential model for liquid CO2