Glassy dynamics in the asymmetrically constrained kinetic Ising chain P. Sollich 1 and M. R. Evans 2 1 Department of Mathematics, Kings College London, Strand, London WC2R 2LS, United Kingdom 2 School of Physics, University of Edinburgh, Mayfield Road, Edinburgh EH9 3JZ, United Kingdom Received 18 March 2003; published 26 September 2003 We study the dynamics of the East model, comprising a chain of uncoupled spins in a downward-pointing field. Glassy effects arise at low temperatures T from the kinetic constraint that spins can only flip if their left neighbor is up. We give details of our previous solution of the nonequilibrium coarsening dynamics after a quench to low T Phys. Rev. Lett. 83, 3238 1999, including the anomalous coarsening of down-spin domains with typical size d ¯ t T ln 2 , and the pronounced ‘‘fragile glass’’ divergence of equilibration times as t * =exp(1/T 2 ln 2). We also link the model to the paste-all coarsening model, defining a family of interpolating models that all have the same scaling distribution of domain sizes. We then proceed to the problem of equilibrium dynamics at low T. Based on a scaling hypothesis for the relation between time scales and length scales, we propose a model for the dynamics of ‘‘superdomains’’which are bounded by up-spins that are frozen on long time scales. From this we deduce that the equilibrium spin correlation and persistence functions should exhibit identical scaling behavior for low T, decaying as g ( t ˜ ). The scaling variable is t ˜ =( t / t * ) T ln 2 , giving strongly stretched behavior for low T. The scaling function g ( ) decays faster than exponential, however, and in the limit T 0 at fixed t ˜ reaches zero at a finite value of t ˜ . DOI: 10.1103/PhysRevE.68.031504 PACS numbers: 64.70.Pf, 05.70.Ln, 05.20.-y, 75.10.Hk I. INTRODUCTION The phenomenology of glassy systems—see, e.g., Refs. 1–4for excellent reviews—has inspired many theoretical descriptions and explanations. Experimentally, long relax- ation times are observed; when these become much longer than the observation time scale a glass transition is said to occur. Other signatures of glassy dynamics are correlation functions that can be fitted by a stretched exponential decay law and aging phenomena 5where, since the system is out of thermal equilibrium, it keeps evolving as time goes by and time-translation invariance is broken. From a modeling perspective the same phenomenology arises when one studies simple model systems by computer simulation. Again, relaxation times can outstrip the time available to run a simulation and one never explores the equilibrium state. The long relaxation times in glasses typically show a pro- nounced divergence as the temperature T is lowered and are often fitted experimentally by the Vogel-Tammann-Fulcher VTFlaw = 0 exp-A / T -T 0  . 1 The relaxation time may characterize, for example, the time for a density fluctuation or an externally imposed stress to relax. Although some heuristic justifications have been offered 6, for practical purposes VTF is just a fit with three parameters 0 , A , T 0 . For T 0 =0 it reduces to an Arrhenius law. A system for which T 0 is small, so that one has some- thing close to Arrhenius behavior, is referred to as a ‘‘strong glass,’’ whereas a system exhibiting large deviations from Arrhenius behavior is called a ‘‘fragile glass.’’ Generally, T 0 is much lower than the experimental temperatures so that the mathematical singularity in the fit 1is not physically rel- evant in an experiment. From the theoretical point of view, however, there has been a long debate over whether T 0 might represent a true thermodynamic transition temperature which would in principle be measurable in the limit of infinitely slow cooling. Although 1is popular, it is not the only possibility for a fit. For example, the exponential inverse temperature squared EITSform = 0 expB / T 2 2 has been proposed as an alternative. This form does not ex- hibit a singularity at any finite T. Experimentally, or in a computer simulation, it is difficult to distinguish between VTF and EITS behavior due to obvious limitations on the longest accessible time scales; both can represent the experi- mentally observed ( T ) in many materials 7. Theoretical work is thus essential for clarifying whether VTF or EITS might be more appropriate. Stretched exponential decay of a relaxation function, let us say an autocorrelation q ( t ), is expressed by the Kohlrausch-Williams-Watt law q t exp-t / b , 3 where the stretching exponent b 1. An heuristic explana- tion for this law is that there is a broad distribution ( ) of relaxation modes with decay constants , q t = d exp-t / . 4 For example, if one assumes ( ) exp(-a) then for large t the dominant modes have =( t / a ) 1/2 which leads to Eq. 3 with b =1/2. This however leaves the physical mechanisms by which such a relaxation time distribution would arise un- clear. PHYSICAL REVIEW E 68, 031504 2003 1063-651X/2003/683/03150416/$20.00 ©2003 The American Physical Society 68 031504-1