Though on increasing charge carrier concentration more of charge carriers should diffuse from hot to cold end which should result in increase of thermopower but we observe opposite. Why is it so?
You are talking about change in con entration of charge carriers hence I can assume that you are talking about thermoelectric power in semiconductors. In semiconductors, thermoelectric power is about 1000 times larger than metals. The major contribution comes from the position of Fermi level Ef from Ec in n type and from EV in case of p type semiconductor.When Ec - Ef will be more thermoelectric power will be more. The concentration of free charge carriers though will be small for large Ec - Ef, yet the contribution due to Fermi level position will be more. Similarly when Ec - Ef will be small, thermoelectric power will be small though free carrier concentration will more.
The Seebeck coefficient is related to "n", "m" by the equation mentioned by Mr. Sikanth. Like metals by w sintering a material in argon atmosphere the sintered material will behave like metal i.e. Semiconductor to metallic transition (SMT) behaviour will occur and hence the the no of carriers will be increased and will result in a lower Seebeck coefficient.
The Seebeck coefficient (S - also known as thermoelectric power) and electrical conductivity strongly depend on the Fermi level, which in turn depends on the carrier concentration, carrier effective mass, and temperature. It has been explained above. In addition, according to Fermi-Dirac statistics, we can obtain the equation for thermoelectric power then this is inversely proportional to Fermi energy and electrical conductivity. The electrical conductivity is proportional to carrier density and mobility (=n.q.u, where n is carrier density n-p), so that the S will decrease with increase of charge carrier concentration.
Since the Seebeck is inversely proportional to the carrier density. n increases, Seebeck decreases as you said. Note that sometimes it exists a bipolar conduction effect in which the Seebeck is not linear decay as usual. Please have a look at: Article Thermoelectric Penta-Silicene with a High Room-Temperature F...