Steven Yuvan

Synchronization of Coupled Oscillators -- Phase Transitions and Entropies

Over the last half century the liquid-gas phase transition and the magnetization phase transition have come to be well understood. After an order parameter, $r$, is defined, it can be derived how $r=0$ for $T>T_c$ and how $r \propto (T_c - T)^γ$ at lowest order for $T < T_c$. The value of $γ$ appears to not depend on physical details of the system, but very much on dimensionality. No phase transitions exist for one-dimensional systems. For systems of four or more dimensions, each unit is interacting with sufficiently many neighbors to warrant a mean-field approach. The mean-field approximation leads to $γ= 1/2$. In this article we formulate a realistic system of coupled oscillators. Each oscillator moves forward through a cyclic 1D array of $n$ states and the rate at which an oscillator proceeds from state $i$ to state $i+1$ depends on the populations in states $i+1$ and $i-1$. We study how the phase transitions occur from a homogeneous distribution over the states to a clustered distribution. A clustered distribution means that oscillators have synchronized. We define an order parameter and we find that the critical exponent takes on the mean-field value of 1/2 for any $n$. However, as the number of states increases, the phase transition occurs for ever smaller values of $T_c$. We present rigorous mathematics and simple approximations to develop an understanding of the phase transitions in this system. We explain why and how the critical exponent value of 1/2 is expected to be robust and we discuss a wet-lab experimental setup to substantiate our findings.