Remarks on a possible physical sense of the value ZT = 1 for semiconductor thermoelectrics

The problem of the existence of an upper limit of thermoelectric figure of merit Z is of principal importance for thermoelectric energy conversion. It was discussed actively during the period of development of basic aspects of semiconductor thermoelectricity but was not solved. After a long time of silence it attracts attention again. C.B.Vining at ICT'92 considered the well known fact that the dimensionless parameter ZT/spl les/1 for all real thermoelectrics. He concluded that one doesn't know any physical reasons to consider this value as a specific one. Our paper indicates those rather general assumptions at which ZT/spl ap/1 for semiconductors at a qualitative evaluation. It is shown too that the dependence of ZT increase on decrease of lattice thermal conductivity is weaker for semiconductors with ZT>1 than for one with ZT<1.