Carlos E. Fiore, Cláudio J. DaSilva
The development of reliable methods for estimating microcanonical averages constitutes an important issue in statistical mechanics. An alternative approach, proposed by Shida {\it et al.} [Phys. Rev. E {\bf 68}, 066125 (2003)], e valuates the microcanonical temperature by means of a formula originally deduced in the grand-canonical ensemble. Taking into account the inequivalence of ensembles for finite systems, in this paper the trustworthiness of the above-mentioned scheme is detailed investigated. Our study is carried out from the comparison between the temperature directly obtained from the derivation of entropy with respect to the energy, to the ones obtained from the density of states. A systematic analysis for finite sizes systems and different models undergoing first and second-order phase transitions is undertaken. Our results show that although the methods are inequivalent for extreme small system sizes, their conformity is achieved for rather small $L$'s. The simplicity and the great precision of our results, when compared with available exact approaches and the well established Wang-Landau method, suggest a direct and very precise protocol for obtaining microcanonical quantities.
View original:
http://arxiv.org/abs/1208.1719
No comments:
Post a Comment