Bias induced drift and trapping on random combs and the Bethe lattice: Fluctuation regime and first order phase transitions

TL;DR

This paper investigates how bias-induced trapping affects fluctuations in random walks on disordered structures, revealing regimes of anomalous fluctuations and phase transitions influenced by disorder and branch length distributions.

Contribution

It provides exact calculations of variance in random combs, bounds for thresholds on Bethe lattices, and shows how fluctuation transitions can be first order depending on branch length distribution.

Findings

Variance diverges at an earlier bias threshold than velocity vanishing.

Disorder-averaged variance exhibits anomalous fluctuations.

Transition to anomalous fluctuations can be first order depending on branch length distribution.

Abstract

We study the competition between field-induced transport and trapping in a disordered medium by studying biased random walks on random combs and the bond-diluted Bethe lattice above the percolation threshold. While it is known that the drift velocity vanishes above a critical threshold, here our focus is on fluctuations, characterized by the variance of the transit times. On the random comb, the variance is calculated exactly for a given realization of disorder using a 'forward transport' limit which prohibits backward movement along the backbone but allows an arbitrary number of excursions into random-length branches. The disorder-averaged variance diverges at an earlier threshold of the bias, implying a regime of anomalous fluctuations, although the velocity is nonzero. Our results are verified numerically using a Monte Carlo procedure that is adapted to account for ultra-slow returns from long branches. On the Bethe lattice, we derive an upper bound for the critical threshold bias for anomalous fluctuations of the mean transit time averaged over disorder realizations. Finally, as for the passage to the vanishing velocity regime, it is shown that the transition to the anomalous fluctuation regime can change from continuous to first order depending on the distribution of branch lengths.