Finite-temperature dynamical correlations using the microcanonical ensemble and the Lanczos algorithm

M. W. Long, P. Prelovšek, S. El Shawish, J. Karadamoglou, and X. Zotos
Phys. Rev. B 68, 235106 – Published 15 December 2003
PDFExport Citation

Abstract

We show how to generalize the zero-temperature Lanczos method for calculating dynamical correlation functions to finite temperatures. The key is the microcanonical ensemble which allows us to replace the involved canonical ensemble with a single appropriately chosen state; in the thermodynamic limit it provides the same physics as the canonical ensemble but with the evaluation of a single expectation value. We can employ the same system sizes as for zero temperature but, whereas the statistical fluctuations present in small systems are prohibitive, the spectra of the largest system sizes are surprisingly smooth. We investigate, as a test case, the spin conductivity of the spin-1/2 anisotropic Heisenberg model and in particular we present a comparison of spectra obtained by the canonical and microcanonical ensemble methods.

  • Received 10 February 2003

DOI:https://doi.org/10.1103/PhysRevB.68.235106

©2003 American Physical Society

Authors & Affiliations

M. W. Long1, P. Prelovšek2, S. El Shawish2, J. Karadamoglou3, and X. Zotos3

  • 1School of Physics, Birmingham University, Edgbaston, Birmingham B15 2TT, United Kingdom
  • 2Faculty of Mathematics and Physics, University of Ljubljana, Ljubljana, SolveniaJ. Stefan Institute, 1000 Ljubljana, Slovenia
  • 3Institut Romand de Recherche Numérique en Physique des Matériaux (IRRMA), EPFL, 1015 Lausanne, Switzerland

References (Subscription Required)

Click to Expand
Issue

Vol. 68, Iss. 23 — 15 December 2003

Reuse & Permissions
Access Options
Author publication services for translation and copyediting assistance advertisement

Authorization Required


×
×

Images

×

Sign up to receive regular email alerts from Physical Review B

Log In

Cancel
×

Search


Article Lookup

Paste a citation or DOI

Enter a citation
×