The p-type (Bi0.25Sb0.75)(2)Te-3 doped with 8 wt% excess Te alone and n-type Bi-2(Te0.94Se0.06)(3) codoped with 0.017 wt% Te and 0.068 wt% I were grown by the Bridgman method and were cut 3 into a parallelepiped of 5 x 5 x 15 mm, where the length of 15 mm is parallel to the freezing direction. The specimen is mounted on an X-Ystage and the temperature difference between two probes set at an interval of 1 mm was approximately 2.6 K. The local Seebeck coefficient alpha(l) and local electrical resistivity rho(l) along the freezing direction were measured at a scan step of 1 mm before and after annealing at 673 K for 2 h in vacuum. The local thermal conductivity kappa(l) was calculated from rho(l) using the relation between K and p obtained previously for a series of bismuth-telluride compounds. As a result, alpha(l) rho(l) and kappa(l) before and after annealing changed significantly from place to place, so that the effect of annealing on the local thermoelectric figure of merit (ZT)(l) was never uniform throughout the specimen surface. The maximum (ZT)(l) of the as-grown p- and n-type specimens reached surprisingly great values of 1.88 and 1.59 at 298 K, respectively, which correspond to about twice as large as those observed macroscopically by the conventional Seebeck coefficient apparatus with an interval (between two probes) of 8 mm. Probably, these maximum values of (ZT)(l) would be an upper limit of ZT for the p- and n-type bismuth- telluride bulk compounds, at least in the present fabrication method. (c) 2006 Springer Science + Business Media, Inc.