Caloric curves have traditionally been derived within the microcanonical ensemble via dS/dE=1/T or within the canonical ensemble via E=T^2*d(ln Z)/dT. In the thermodynamical limit, i.e., for large systems, both caloric curves give the same result. For small systems like nuclei, the two caloric curves are in general different from each other and neither one is reasonable. Using dS/dE=1/T, spurious structures like negative temperatures and negative heat capacities can occur and have indeed been discussed in the literature. Using E=T^2*d(ln Z)/dT a very featureless caloric curve is obtained which generally smoothes too much over structural changes in the system. A new approach for caloric curves based on the two-dimensional probability distribution P(E,T) will be discussed.
تحميل البحث