Thermoelectric semiconductors are heavily doped in order to optimise t
he figure-of-merit. In some cases the doping levels may be in excess o
f 10(26) m-3. However, it is not always possible to achieve such dopin
g levels in practice. Even if the required levels are obtained by suit
able annealing at high temperatures, the long-term performance of the
device is affected when the dopant gradually precipitates out of the s
olution. Methods have been suggested to push up the solubility limit b
y adding selected material into the host. The method has achieved some
success. One may also look into the possibility of reducing the optim
um doping levels, so that they are well below the solubility limits. L
ow dimensional systems may become degenerate at relatively lower dopin
g levels than the bulk materials. This fact coupled with the possibili
ty of increased carrier mobility may be of interest in future thermoel
ectric material research programmes.