We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cond-mat.stat-mech

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Condensed Matter > Statistical Mechanics

Title: High-precision Estimate of the Critical Exponents for the Directed Ising Universality Class

Authors: Su-Chan Park
Abstract: With extensive Monte Carlo simulations, we present high-precision estimates of the critical exponents of branching annihilating random walks with two offspring, a prototypical model of the directed Ising universality class in one dimension. To estimate the exponents accurately, we propose a systematic method to find corrections to scaling whose leading behavior is supposed to take the form $t^{-\chi}$ in the long-time limit at the critical point. Our study shows that $\chi\approx 0.75$ for the number of particles in defect simulations and $\chi \approx 0.5$ for other measured quantities, which should be compared with the widely used value of $\chi = 1$. Using $\chi$ so obtained, we analyze the effective exponents to find that $\beta/\nu_\| = 0.2872(2)$, $z = 1.7415(5)$, $\eta = 0.0000(2)$, and accordingly, $\beta /\nu_\perp = 0.5000(6)$. Our numerical results for $\beta/\nu_\|$ and $z$ are clearly different from the conjectured rational numbers $\beta/\nu_\| = \frac{2}{7} \approx 0.2857$, $z = \frac{7}{4}= 1.75$ by Jensen [Phys. Rev. E, {\bf 50}, 3623 (1994)]. Our result for $\beta/\nu_\perp$, however, is consistent with $\frac{1}{2}$, which is believed to be exact.
Subjects: Statistical Mechanics (cond-mat.stat-mech)
Journal reference: J. Korean Phys. Soc. 62, 469 (2013)
DOI: 10.3938/jkps.62.469
Cite as: arXiv:1406.7070 [cond-mat.stat-mech]
  (or arXiv:1406.7070v1 [cond-mat.stat-mech] for this version)

Submission history

From: Su-Chan Park [view email]
[v1] Fri, 27 Jun 2014 05:47:33 GMT (402kb)

Link back to: arXiv, form interface, contact.