Page 94 -
P. 94

Hard disks and spheres                                                      2






     In the first chapter ofthis book, weconsidered simpleproblems in statis-
     tics: pebbles on the beach, needles falling butnever rolling, and people  2.1 Newtonian deterministic
                                                                        mechanics               83
     strolling onheliports by night.Wenow move onto study model systems
     in physics—particles with positions, velocities, and interactions—that  2.2 Boltzmann’s statistical
                                                                        mechanics               92
     obey classical equations ofmotion.To understand how physical systems
                                                                     2.3 Pressure and the
     can be treated with the tools of statistics and simulated with Monte  Boltzmann distribution  108
     Carlo methods, weshall consider the hard-sphere model, which lies at  2.4 Large hard-sphere
     the heart of statistical mechanics.Hard spheres, which are idealizations  systems         119
     ofbilliard balls, in free space orinabox,behaveasfree particles when-  2.5 Cluster algorithms  122
     ever they are notin contact with other particles or with walls, and obey  Exercises       128
     simplereflectionrules oncontact.                                 References                130
       The hard-sphere model played a crucial role in the genesis ofsta-
     tistical mechanics.Since the early days of machine computing, in the
     1950s, and upto the present day, the hard-sphere model has spurred
     the development ofcomputer algorithms, and both the explicit numeri-
     cal integration of Newton’sequations and the Markov-chain Monte Carlo
     algorithm were first tried out onthis model. Weshall use such algorithms
     to illustrate mechanics and statistical mechanics, and to introduce the
     fundamental concepts of statistical mechanics: the equiprobability prin-
     ciple, the Boltzmann distribution, the thermodynamic temperature, and
     the pressure.Weshall also be concerned with the practical aspects of
     computations and witness the problems of Markov-chain algorithms at
     high densities.Weshall conclude the chapter with a first discussion of
     sophisticated cluster algorithms which are commonto many fields of
     computational physics.
       In the hard-sphere model,all configurations havethe same potential
     energy and there is no energetic reasonto prefer any configuration over
     anyother. Only entropic effects come into play. Inspite ofthis restric-
     tion, hard spheres and disks show a rich phenomenology and exhibit
     phase transitions fromthe liquid to the solid state.These “entropic
     transitions”were once quite unsuspected, and then hotly debated, be-
     fore they ended uppoorly understood, especially in two dimensions.
     The physics ofentropywill appear in several places in this chapter, to
     be taken up again in earnest in Chapter 6.
   89   90   91   92   93   94   95   96   97   98   99