A semiclassical study of bubble formation (nucleation) in nuclei at finite temperature is presented. The effects of the Coulomb interaction and of the density profile parametrization are investigated. Significant nucleation probabilities are found.Since the equation of state of hot nuclear matter is of van der Waals type, one might expect that hot nuclei exhibit phase transitions. Their existence has been conjectured in experiments with high-energy protons on nuclei and in heavy-ion collisions [ 1,2 ]. The role of statistical and quantum fluctuations which lead to phase transitions has been first analyzed in a microscopic approach in ref. [ 3 ], based on relativistic mean field theory of nuclei [4 ]. In a recent work [ 5 ], a similar approach, also using the path integral technique to calculate phase transitions, has been applied to bubble formation in hot and dense nuclear matter. In this work we improve upon ref.[5 ] by including Coulomb forces which are of great importance in realistic calculations. In addition, we use a more appropriate parametrization of the bubble density profiles and take gradient corrections to the free kinetic energy functional into account.The picture we have in mind is that hot compressed nuclear matter produced in a heavy-ion reaction, may by subsequent (isentropic) expansion reach the region of metastability, where bubbles of lower than liquid density may be formed. The formation of droplets in a gas background can be excluded by the argument that the gas density is below freeze out. The phase transitions are due to statistical fluctuations. Quantum fluctuations have been shown to be relevant only for temperatures T< 1 MeV [ 5 ] n. The transition probability W from a metastable phase to a stable one has a WKB-like formwhere ~max is the maximum of the thermodynamical potential [3,5], see eq. (3) and fig. 1. We normalize W to 1 at vanshing potential barrier [ 5 ]. The relevant potential for isothermal processes in a grand canonical ensemble is the thermodynamical potential I2, defined aswhere F=E-TS is the free energy calculated from the internal energy E and the entropy S at given temperature T. The chemical potential/t is the Lagrange multiplier of the particle number N. We assume a homogeneous liquid background density PL, on which bubble formation can take place. Therefore we introduce a thermodynamical potential ~, in which the constant potential associated with the homogeneous background S'2[pL] is subtracted. The thermodynamical treatment of charged systems, however, is a difficult issue, since the Coulomb interaction is a long-range force and hence has no clear place in thermodynamics. We adopt the subtraction prescription of refs. [6][7][8] where the thermodynamical potential for a charged bubble takes the * Work supported by Deutsche Forschungsgemeinschaft, grant no. Br 733/4-1.:l This holds true also for the present calculation, as the mass parameters resulting from eq. (5) are almost the same as in ref.[5].0370-2693/86/$03.50