I want to optimize unit cell of carbon nanotube adsorbed with metal atoms. I know that for device calculation k point sampling is 1*1*many. However, for perfectly periodic system transmission spectrum can be calculated from bulk. I want to know what kind of k point sampling and constraints need to be given to optimize such a structure. Also is it correct to optimize unit cell, repeat it along transport direction, convert to device and calculate transmission spectrum without any further device relaxation?
