Long-time tails in the velocity autocorrelation function of hard-rod binary mixtures

The temporal evolution of binary mixtures of hard rods in a ring is simulated in a computer with random initial velocities ±{}v. The time the system takes to reach a Maxwellian distribution dramatically diverges as the mass ratio {ε}{→}1 and it also increases, although rather slowly, when {ε}{→}{∞}. A negative ``long-time tail,'' i.e., a slow, power-law decay in the velocity autocorrelation function at large values of the time t, is observed whose behavior changes from t^-3 to t^-δ, {δ}{}1, as {ε} is increased from {ε}=1. .AE