Slow dynamics in systems with long-range interactions STEFANO RUFFO Dipartimento di Energetica S. Stecco and INFN Center for the Study of Complex Dynamics, Firenze Newton Institute Workshop on Relaxation Dynamics of Macroscopic Systems 9-13 January 2006 Slow dynamics in systems with long-range interactions p.1/17
PLAN Introduction Long-range interactions Extensivity vs. additivity Ensemble inequivalence: negative specific heat, temperature jumps Models XY models Free electron laser Slow dynamics Quasi-stationary states Metastability Broken ergodicity Slow dynamics in systems with long-range interactions p.2/17
Long range interactions Energy of a particle at the center of a sphere of radius where matter is homogeneously distributed The contribution of the surface of the sphere can be neglected only if Long range if ( ). Physical examples Gravity, singularity at the origin Coulomb, Debye screening Dipolar, shape dependence Onsager vortices Mean-Field, any. Slow dynamics in systems with long-range interactions p.3/17
Extensive but not additive The Curie-Weiss Hamiltonian is EXTENSIVE not ADDITIVE I II but + _ Zero magnetization state, Hence Slow dynamics in systems with long-range interactions p.4/17
Ensemble inequivalence Convex intruders s a) s b) ε ε 1 2 ε ε 1 ε 2 ε Negative heat capacity (negative susceptibility, etc.) Temperature jumps Slow dynamics in systems with long-range interactions p.5/17
XY models Simple mean-field models with Hamiltonian dynamics Simplifications of Gravitational and charged sheet models Wave-particle interactions with D. Mukamel (Weizmann) and P. De Buyl (ULB, Bruxelles) Slow dynamics in systems with long-range interactions p.6/17
Phase diagram and caloric curves 1.2 1 T/J m mic = 0 m can = 0 2.25 2 T 0.8 0.6 Canonical tricritical point m can 0 m mic = 0 1.75 1.5 0.4 0.2 m mic 0 m can 0 Microcanonical tricritical point K/J 1 2 3 4 5 1.25 1 0.75 1 0.5 0 0.5 1 E At (HMF model), second order phase transition at are equivalent. For ensembles are inequivalent. Negative specific heat for ; Temperature jumps for. Right figure shows the caloric curve for molecular dynamics simulation with. Ensembles. The points are results of a Slow dynamics in systems with long-range interactions p.7/17
Free Electron Laser y Colson-Bonifacio model x z with A. Antoniazzi (Florence), J. Barré (Nice), T. Dauxois (ENS-Lyon), D. Fanelli (Florence and Stockholm), G. De Ninno (Sincrotrone Trieste) Slow dynamics in systems with long-range interactions p.8/17
Quasi-stationary states 0.8 Laser intensity (arb. units) 0.6 0.4 0.2 Laser intensity 0.6 0.5 1 2 3 0.4 0 500 1000 1500 2000 - z 0 0 20 40 60 80 100 120 140 160 180 200 - z (curve 1), (curve 2), (curve 3) On a first stage the system converges to a quasi-stationary state. Later it relaxes to Boltzmann-Gibbs equilibrium on a time. The quasi-stationary state is a Vlasov equilibrium, sufficiently well described by Lynden-Bell s Fermi-like distribution arising from violent relaxation" (constrained maximum entropy principle). Slow dynamics in systems with long-range interactions p.9/17
Vlasov equation limit, the single particle distribution function obeys a Vlasov equation. In the #" with Slow dynamics in systems with long-range interactions p.10/17
0 1 Vlasov equilibria Coarse grained entropy maximization (Lynden-Bell 1968, Chavanis, 1996) Non Gaussian momentum distribution /. " - *,+ )( ' $&% #" /. " - * + ( ' $% #" Non-equilibrium field amplitude 0 Slow dynamics in systems with long-range interactions p.11/17
Dynamics of HMF For the HMF model (XY model with at ( ). ) there is a second order phase transition In the Vlasov equation context, the phase transition is signalled by the instability of the homogeneous state with Gaussian momentum distribution. For other initial states, this Vlasov instability can appear at energy densities that lie below. For the finite system can remain trapped in quasi-stationary states whose life-time increases with a power of. For the so called water bag" states, where the particles are homogeneously distributed on the circle and momenta have a symmetric square distribution, the Vlasov instability energy density is and the life-time of metastable states increases as,. 0.7 0.6 U=0.50 0.35 0.3 (a) 0.5 U=0.54 0.25 M(t) 0.4 0.3 U=0.58 M(t) 0.2 0.15 0.2 0.1 U=0.62 0.1 0.05 0 0 5 10 15 20 25 30 35 40 45 50 55 t 0 1 0 1 2 3 4 5 6 7 8 log 10 t Slow dynamics in systems with long-range interactions p.12/17
HMF macroparticle Refinements of maximum entropy methods should include the macroparticle concept (Tennyson, 1994) single particle distribution M 0 =0.3 M 0 =0.5 0.1 0.1 f(p) 0.01-1.5-1 -0.5 0 0.5 1 1.5 0.01-1.5-1 -0.5 0 0.5 1 1.5 0.1 0.1 M 0 =0.7 0.01-1.5-1 -0.5 0 0.5 1 1.5 0.01-1.5-1 -0.5 0 0.5 1 1.5 p M 0 =0.9 Slow dynamics in systems with long-range interactions p.13/17
Metastability At a microcanonical first order transition, temperature has a bimodal distribution. Once prepared in a local entropy maximum the system relaxes to the global entropy maximum on a time that increases with barrier size., where is the entropy density 0.65 T 0.6 T c CP A B HP (a) 80 P(T) 60 40 0.61 T(t) 0.59 0.57 (d) 0.55 0 4e+05 8e+05 t 10 9 10 3 10 4 10 5 10 5 (b) τ 10 4 10 7 10 3 0.55 D 20 10 5 C U c 0.5 2.45 2.5 2.55 2.6 U 0 0.55 0.57 0.59 0.61 0.63 0.65 T 10 3 0 20000 40000 60000 N Slow dynamics in systems with long-range interactions p.14/17
Broken ergodicity Ising model with short and long-range interactions on a ring 1.5 1 m 0.5 0-0.5 1.5 1 m 0.5 0 a b -1 0 1-1 0 1 0 25 50 75 100 (MC sweeps)/1000 with D. Mukamel and N. Schreiber (Weizmann) Slow dynamics in systems with long-range interactions p.15/17
Conclusions Microcanonical and canonical ensemble disagree for long range interactions at canonical first order transitions. Negative specific heat and temperature jumps are typical signatures of ensemble inequivalence. Collective phenomena for wave-particle interactions (FEL) are the result of constrained maximum entropy principles (Vlasov equilibria). Quasi-stationary states appear whose life-time increases with system size. Metastable states have a life-time which increases exponentially with system size. Due to non-additivity, broken ergodicity is a generic feature of systems with long-range interactions. Slow dynamics in systems with long-range interactions p.16/17
References T. Dauxois, S. Ruffo, E. Arimondo and M. Wilkens (Eds.) Dynamics and statistics of systems with long-range interactions, Springer Lecture Notes in Physics 602 (2002). J. Barré, D. Mukamel and S. Ruffo:Inequivalence of ensembles in a system with long range interactions, Phys. Rev. Lett., 87(3), 030601 June (2001). J. Barré, F. Bouchet, T. Dauxois and S. Ruffo Large deviation techniques applied to systems with long-range interactions, J. Stat. Phys., 119, 677 (2005). P. de Buyl, D. Mukamel and S. Ruffo: Ensemble inequivalence in a XY model with long-range interactions, in Unsolved Problems of Noise and Fluctuations", AIP Conference Proceedings 800, 533 (2005). J. Barré, T. Dauxois, G. De Ninno, D. Fanelli, S. Ruffo:Statistical theory of high-gain free-electron laser saturation, Phys. Rev. E, Rapid Comm., 69, 045501 (R) (2004). D. Mukamel, S. Ruffo and N. Schreiber: Breaking of ergodicity and long relaxation times in systems with long-range interactions, Phys. Rev. Lett., 95, 240604 (2005). Slow dynamics in systems with long-range interactions p.17/17