1: % LaTeX source for MREvans Paper
2: % Preamble
3:
4: \documentclass[a4paper,12pt]{article}
5: \usepackage[dvips]{graphicx}
6:
7: % Larger page
8:
9: \setlength{\evensidemargin}{0.0in}
10: \setlength{\oddsidemargin}{0.0in}
11: \setlength{\textwidth}{16.4cm}
12: \setlength{\textheight}{23.0cm}
13: \setlength{\topmargin}{-1.5cm}
14: \setlength{\headsep}{.4in}
15: \setlength{\parskip}{0.3cm}
16:
17:
18: % Mathematics
19:
20: % For better features
21: \usepackage{amsmath}
22:
23: % A roman `d' for differentiation and integration
24: \newcommand{\D}{\ensuremath{\mathrm{d}}}
25:
26: % A roman 'd' with some space around it for use after a \int with
27: % limits
28: \newcommand{\Di}[1]{\ensuremath{\!\!\mathrm{d}#1\,}}
29:
30: % An 'identity' symbol
31: \newcommand{\thickone}{\mbox{$1\!\!1$}}
32:
33:
34: % Document body
35: \begin{document}
36:
37: % Title area
38:
39: \title{
40: Condensation Transitions
41: in Nonequilibrium systems
42: }
43: \author{ M.\ R.\ Evans \\[2ex]
44: School of Physics,\\
45: The University of Edinburgh,\\
46: Mayfield Road,\\
47: Edinburgh EH9 3JZ,\\
48: U.K.}
49: \maketitle
50: \begin{abstract}
51: Systems driven out of equilibrium can often exhibit behaviour not seen
52: in systems in thermal equilibrium- for example phase transitions in
53: one-dimensional systems. In this talk I will review several
54: `condensation' transitions that occur when a conserved quantity is
55: driven through the system. Although the condensation is spatial,
56: i.e. a finite fraction of the conserved quantity condenses into a
57: small spatial region, useful comparison can be made with usual
58: Bose-Einstein condensation. Amongst some one-dimensional examples I
59: will discuss the `Bus Route Model' where the condensation corresponds
60: to the clustering together of buses moving along a bus-route.
61: \end{abstract}
62:
63: % Main text
64:
65: % -----------------------------------------------------------------------------
66: \section{Introduction}
67:
68: Broadly speaking, one can consider two types of nonequilibrium
69: systems: those relaxing towards thermal equilibrium and those held far
70: from thermal equilibrium {\it e.g.} by the system being driven by some
71: external field. In the latter case the steady state of the system will
72: not be described by usual Gibbs-Boltzmann statistical weights rather it
73: will be a nonequilibrium steady state.
74: A natural way to construct a nonequilibrium steady state is to drive
75: the system by forcing a current of some conserved quantity, for
76: example energy or mass, through the system. Such systems are known as
77: driven diffusive systems (DDS) \cite{SZ}.
78:
79:
80: In recent years the
81: possibility
82: of phase transitions and phase separation in
83: {\em one-dimensional} nonequilibrium systems
84: has been explored and some examples are by now well
85: studied. To appreciate the significance one should recall the general
86: dictum that in one-dimensional equilibrium systems phase ordering and
87: phase transitions do not occur (except in the limit of
88: zero-temperature, or with long range interaction)
89: \cite{LL}.
90:
91: Let us briefly review work on one-dimensional
92: phase transitions in driven systems.
93: A very simple one-dimensional driven diffusive system is the
94: asymmetric simple exclusion process (ASEP). Here particles hop in a
95: preferred direction on a one-dimensional lattice with hard-core
96: exclusion (at most one particle can be at any given site). Indicating
97: the presence of a particle by a 1 and an empty site (hole) by 0 the
98: dynamics comprises the following exchanges at nearest neighbour sites
99: \begin{eqnarray}
100: 1\ 0 &\to& 0\ 1 \quad\mbox{with rate}\quad 1 \nonumber \\
101: 0\ 1 &\to& 1\ 0 \quad\mbox{with rate}\quad q
102: \end{eqnarray}
103:
104: The open system was studied by Krug\cite{Krug91} and boundary
105: induced phase transitions shown to be possible. Specifically one
106: considers a lattice of $N$ sites where at the left boundary site (site
107: 1) a particle is introduced with rate $\alpha$ if that site is
108: empty, and at the right boundary site (site $N$)
109: any particle present is removed with rate $\beta$. Thus the dynamical
110: processes at the boundaries are
111: \begin{eqnarray}
112: \mbox{ at site $1$} \quad 0 &\to& 1 \quad\mbox{with rate}\quad \alpha \nonumber \\
113: \mbox{ at site $N$} \quad 1 &\to& 0 \quad\mbox{with rate}\quad \beta\;.
114: \end{eqnarray}
115: These boundary conditions force a steady state current of particles
116: $J$ through the system. Phase transitions occur when
117: $\lim_{N{\to}\infty} J$ exhibits non-analyticities. The steady state
118: of this system was solved exactly for the totally asymmetric case
119: \cite{DEHP,SD} and more recently for the general $q$ case
120: \cite{Sasamoto,BECE}. When $q<1$ the phase diagram comprises three
121: phases: a high-density phase where the current is controlled by a low
122: exit rate $\beta$---one can think of this is queue of cars
123: at a traffic light that doesn't let many cars through;
124: a low-density phase where the current controlled by a low injection
125: rate $\alpha$---think of this as a traffic light that does not let many
126: cars onto an open road; a
127: maximal-current phase where both $\alpha,\beta$ are high
128: ($\alpha,\beta > (1-q)/2$) and the
129: current is $J=(1-q)/4$. Note that since increasing $\alpha$
130: and $\beta$ doesn't increase the current, the current is saturated.
131: In the maximal current phase generic
132: long-range correlations exist, an example being the decay of particle
133: density from the left boundary to the bulk value $1/2$ which is a
134: power law $\sim 1/x^{1/2}$ where $x$ is distance from the left
135: boundary.
136:
137:
138: On the line $\alpha= \beta < (1-q)/2$ which separates the high and low density
139: phases one finds coexistence between a region of low density in the
140: left part of lattice and a region of high density on the right separated by
141: a `shock' where the density changes sharply over a microscopic distance.
142:
143:
144:
145: Perodic systems (i.e. a ring of sites)
146: can also exhibit phase separation when inhomogeneities
147: or defects are introduced. A very simple example is to introduce
148: into the asymmetric exclusion process a
149: `slow bond' through which particles hop with a
150: reduced rate. Then in the steady state one can obtain phase separation
151: between a region of high density behind the slow bond and a region of
152: low density in front of the slow bond. Moving defects ({\it i.e.}
153: particles with dynamics different from that of the others) have also
154: been considered and exact solutions obtained \cite{MRE96,KF,Mallick}.
155: One can think of a slow agricultural vehicle on a country road with a
156: large queue of cars behind it and open road in front of it.
157:
158: A further question is whether systems related to the hopping particle
159: models described so far, but without inhomogeneities, can exhibit
160: phase ordering. A very simple model was introduced in \cite{EKKM}
161: comprising three species of conserved particles, amongst which all
162: possible exchanges are allowed. However a key feature is that the
163: dynamics has a cyclic symmetry i.e. $A$ particles move preferentially
164: to the left of $B$ particles which move preferentially to the left of
165: $C$ particles which in turn move preferentially to the left of $A$
166: particles. The model exhibits strong phase separation into pure
167: domains of $A$ $B$ $C$. Similar strong phase separation occurs
168: in other related models \cite{LBR}.
169:
170:
171: A final class of transitions in one-dimensional hopping particle
172: models is that involving spatial condensation, whereby a finite
173: fraction of the particles condenses onto the same site. Examples
174: include the appearance of a large aggregate
175: in models of aggregation and fragmentation\cite{MKB} and
176: the emergence of a single flock in dynamical models
177: of flocking \cite{OE}. We will analyse
178: a simple example of a condensation transition which occurs
179: in the zero-range process which we now define.
180:
181:
182:
183:
184:
185:
186:
187:
188:
189:
190:
191: \section{The zero-range process}
192: \label{Sec:ZRP}
193: The zero-range process was introduced by Spitzer \cite{Spitzer}
194: and recent applications and developments have been reviewed in \cite{MRE00}.
195: We consider a one-dimensional lattice of $M$ sites with sites
196: labelled $\mu =1 \ldots M$ and periodic boundary conditions
197: (more generally
198: one can consider the zero-range process on a lattice of
199: arbitrary dimension). Each site
200: can hold an integer number of indistinguishable particles. The
201: configuration of the system is specified by the occupation numbers
202: $n_{\mu}$ of each site $\mu$. The total number of
203: particles is denoted by $L$ and is conserved under the dynamics.
204: The dynamics of the system is given by
205: the rates at which a particle leaves a site $\mu$ (one can think of it
206: as the topmost particle---see Figure 1a)
207: and moves to the left nearest neighbour
208: site $\mu{-}1$. The hopping rates $u(n)$ are a function of $n$ the
209: number of particles at the site of departure. Some particular cases
210: are: if $u(n) =n$ then the dynamics of each particle is independent of
211: the others; if $u(n) = {\rm const}\quad$ for $n>0$ then the rate at
212: which a particle leaves a site is unaffected by the number of particles
213: at the site (as long as it is greater than zero).
214:
215:
216:
217: The important attribute of the zero-range process is that it has a
218: steady state described by a product measure. By this it is meant that
219: the steady state probability $P( \{ n_\mu \})$ of finding the system
220: in configuration $\{n_1, n_2 \ldots n_M\}$ is given by a product of
221: factors $f(n_\mu)$
222: \begin{equation}
223: P( \{ n_\mu \}) = \frac{1}{Z(M,L)} \prod_{\mu=1}^{M} f ( n_{\mu} )\;.
224: \label{Prob}
225: \end{equation}
226: Here the normalisation $Z(M,L)$ is introduced so that the sum of the
227: probabilities for all configurations,
228: with the correct number of particles $L$, is one.
229:
230: In the basic model described above, $f(n)$ is given by
231: \begin{eqnarray}
232: f(n) &=& \prod_{m=1}^{n} \frac{1}{u(m)}\quad\mbox{for}\quad n\ge 1
233: \nonumber \\
234: &=& 1 \quad\mbox{for}\quad n=0
235: \label{f1}
236: \end{eqnarray}
237:
238: To prove (\ref{Prob},\ref{f1}) one
239: simply considers the stationarity condition on the probability of a
240: configuration (probability current out of the configuration due to
241: hops is equal to probability current into the configuration
242: due to hops):
243: \begin{equation}
244: \sum_{\mu}
245: \theta (n_\mu) u(n_\mu) P(n_1 \ldots n_\mu \ldots n_L)
246: =
247: \sum_{\mu} \theta (n_\mu)
248: u(n_{\mu{+}1}{+}1) P(n_1 \ldots n_{\mu}{-}1, n_{\mu{+}1}{+}1 \ldots n_L) \;.
249: \label{station}
250: \end{equation}
251: The Heaviside function $\theta(n_\mu)$ highlights that it is the sites
252: with $n>1$ that allow exit from the configuration (lhs of
253: (\ref{station})) but also allow entry to the configuration (rhs of
254: (\ref{station})). Equating the terms $\mu$ on both sides of
255: (\ref{station}) and cancelling common factors assuming (\ref{Prob}),
256: results in
257: \begin{equation}
258: u(n_\mu)f(n_{\mu-1}) f(n_{\mu}) = u(n_{\mu+1}+1) f(n_{\mu}-1) f(n_{\mu+1}+1)
259: \end{equation}
260: This equality can be recast as
261: \begin{equation}
262: u(n_\mu) \frac{f(n_{\mu})}{f(n_{\mu}-1)}
263: = u(n_{\mu{+}1}+1)\frac{ f(n_{\mu{+}1}+1)}{ f(n_{\mu{+}1})}
264: =\mbox{constant}
265: \end{equation}
266: Setting the constant equal to unity implies
267: \begin{equation}
268: f(n_\mu) = \frac{f(n_{\mu}-1)}{ u(n_\mu)}
269: \label{recurr}
270: \end{equation}
271: and iterating (\ref{recurr}) leads to (\ref{f1}).
272: \begin{figure}[htb]
273: \begin{center}
274: \includegraphics[scale=0.60]{fig1.eps}
275: \end{center}
276: \caption{\label{fig:ZRP} Equivalence of
277: zero range process and asymmetric exclusion process.}
278: \end{figure}
279:
280: We can easily generalise to consider an inhomogeneous
281: system by which we mean the hopping rates are site dependent:
282: the hopping rate out of site $\mu$ when it contains
283: $n_\mu$ particles is $u_\mu(n_\mu)$.
284: It is easy to check that the steady state is simply
285: modified to
286: \begin{equation}
287: P( \{ n_\mu \}) = \frac{1}{Z(M,L)} \prod_{\mu=1}^{L} f_{\mu} ( n_{\mu} )
288: \label{Prob2}
289: \end{equation}
290: where $f_\mu$ are given by
291: \begin{eqnarray}
292: f_\mu(n) &=& \prod_{m=1}^{n} \frac{1}{u_\mu(m)}
293: \quad\mbox{for}\quad n\ge 1
294: \nonumber \\
295: &=& 1 \quad\quad \quad\mbox{for}\quad n=0
296: \label{f2}
297: \end{eqnarray}
298: The proof is identical to that for the homogeneous
299: case, with the replacement of $u(n_\mu)$ by $u_\mu(n_\mu)$
300:
301:
302:
303: There exists an exact mapping from a zero-range process to an
304: asymmetric exclusion process. This is illustrated in Figure~1. The
305: idea is to consider the particles of the zero-range process as the
306: holes (empty sites) of the exclusion process. Then the sites of the
307: zero-range process become the moving particles of the exclusion
308: process. Note that in the exclusion
309: process we have $M$ particles hopping on a
310: lattice of $M+L$ sites.
311: A hopping
312: rate in the zero range process
313: $u(m)$ which is dependent on $m$ corresponds to a hopping rate in
314: the exclusion process which depends on the gap to the particle in
315: front. So the particles can feel each other's presence
316: and one can have a long-range interaction.
317:
318:
319:
320: \section{Condensation Transitions}
321: \label{Sec:cond}
322: We now proceed to analyse the steady states of form (\ref{Prob2}) and
323: the condensation transition that may occur. The important quantity to
324: consider is the normalisation $Z(M,L)$ as it plays the role of the
325: partition sum. The normalisation is defined through the condition
326: \begin{equation}
327: Z(M,L) = \sum_{n_1,n_2 \ldots n_M}
328: \delta(\sum_{\mu} n_\mu{-}L)
329: \prod_{\mu=1}^{M} f_\mu( n_\mu )
330: \label{norm}
331: \end{equation}
332: where the $\delta$ function enforces the constraint of $L$ particles. The
333: normalisation may be considered as the analogue of a
334: canonical partition
335: function of a thermodynamic system.
336:
337:
338: We define the `speed' $v$ as the average
339: hopping rate out of a site
340: \begin{eqnarray}
341: v&=& \frac{1}{Z(M,L)}
342: \sum_{n_1,n_2 \ldots n_M}
343: \delta( \sum_{\mu} n_\mu{-}L)
344: u(n_1)
345: \prod_{\mu=1}^{M} f_\mu( n_\mu )
346: \nonumber \\
347: &=&\frac{Z(M,L{-}1)}{Z(M,L)}
348: \label{speed}
349: \end{eqnarray}
350: where we have used (\ref{Prob2},\ref{f2}). Note that (\ref{speed})
351: tells us that the speed is independent of site and thus may be
352: considered a conserved quantity in the steady state of the system. In
353: the totally asymmetric system considered in Section~\ref{Sec:ZRP} the
354: speed is equal to the current of particles flowing between
355: neighbouring sites. The
356: speed is a ratio of partition functions of different system sizes
357: (\ref{speed}) and corresponds to a fugacity.
358:
359: We now use the integral representation of the delta function to write
360: the partition function as
361: \begin{equation}
362: Z(M,L) = \oint \, \frac{dz}{2\pi i} \ z^{-(L+1)}\
363: \prod_{\mu=1}^{M} F_\mu(z)\; ,
364: \label{Zint}
365: \end{equation}
366: where
367: \begin{equation}
368: F_\mu(z) =
369: \sum_{m=0}^{\infty} z^{m} \ f_\mu(m)\; .
370: \label{Fdef}
371: \end{equation}
372: For large $M,L$ (\ref{Zint}) is dominated by the saddle point of the
373: integral and the value of $z$ at the saddle point is the
374: fugacity. The equation for the saddle point reduces to
375: \begin{equation}
376: \frac{L}{M} = \frac{z}{M} \sum_{\mu=1}^{M} \frac{\partial}{\partial z}
377: \ln F_\mu(z) \label{sad}
378: \end{equation}
379: which, defining $\phi= L/M$, can be written as
380: \begin{equation}
381: \phi = \frac{z}{M} \sum_{\mu=1}^{M} \frac{F'_\mu(z)}{F_\mu(z)}\;.
382: \label{sad2}
383: \end{equation}
384: In the thermodynamic limit,
385: \begin{equation}
386: M \rightarrow \infty \;\;\;\mbox{with}\;\;\; L =\phi M\; ,
387: \label{thermlim}
388: \end{equation}
389: where the density $\phi$ is held fixed,
390: the question is whether a valid saddle point value
391: of z can be found from
392: (\ref{sad2}). We expect that for low $\phi$ the saddle point
393: is valid but,
394: as we shall discuss, there exists a maximum value of $z$
395: and if at this maximum value the rhs of (\ref{sad2}) is finite,
396: then for large $\phi$ (\ref{sad2}) cannot be satisfied.
397: We now consider how condensation may occur in
398: the inhomogeneous
399: and the homogeneous case.
400:
401: \subsection{Inhomogeneous case}
402: \label{Sec:Condinhom}
403: To give an idea of how a condensation
404: transition may occur we consider the case
405: $u_\mu(m) = u_\mu$ for $m>0$ {\it i.e.} the hopping rate does not
406: depend on the number of particles at a site.
407: $f_\mu$ is given
408: by
409: \begin{equation}
410: f_\mu(n) = \left( \frac{1}{u_\mu}\right)^{\! n_\mu}
411: \end{equation}
412: and the probability of occupancies $\{n_1,n_2,\ldots,n_M\}$ is
413: \begin{equation}
414: P(\{ n_1,n_2,\ldots,n_M\}) = \frac{1}{Z(M,L)}
415: \prod_{\mu =1}^{M}
416: \left(\frac{1}{u_\mu}\right)^{\! n_\mu}\;.
417: \end{equation}
418: The mapping to an ideal Bose gas is evident: the $L$ particles of the
419: zero-range process are viewed as Bosons which may reside in $M$ states
420: with energies $E_{\mu}$ determined by the site hopping rates:
421: $\exp(-\beta E_{\mu}) = 1/u_{\mu}$. Thus the ground state corresponds
422: to the site with the lowest hopping rate. The normalisation $Z(M,L)$
423: is equivalent to the canonical partition function of the Bose gas. We
424: can sum the geometric series (\ref{Fdef}) to obtain $F_\mu$ and
425: $F'_\mu$ then taking the large $M$ limit allows the sum over $\mu$ to
426: be written as an integral
427: \begin{equation}
428: \phi = \int_{u_{\rm min}}^{\infty} du {\cal P}(u) \ \frac{z}{u-z}
429: \label{gce2}
430: \end{equation}
431: where ${\cal P}(u)$ is the probability distribution of site hopping rates
432: with $u_{\rm min}$ the lowest possible site hopping rate. Interpreting
433: ${\cal P}(u)$ as a density of states, equation (\ref{gce2})
434: corresponds to the condition that in the grand canonical ensemble of
435: an ideal Bose gas the number of Bosons per state is $\phi$. The theory
436: of Bose condensation tells us that when certain
437: conditions on the density of low energy states pertain we can have a
438: condensation transition. Then (\ref{sad2}) can no longer be satisfied
439: and we have a condensation of particles into the ground state, which
440: is here the site with the slowest hopping rate.
441:
442: A very simple example is to have just one `slow site'
443: i.e.
444: $u_1=p$ while the other $M-1$ sites have hopping rates
445: $u_{\mu}=1$ when $ \mu >1$.
446: Using the mapping to an exclusion process,
447: this corresponds one slow particle
448: i.e. agricultural vehicle example described earlier.
449: One can show \cite{MRE96} that
450: for a high density of particles in the zero range process
451: (low density of particles in the corresponding asymmetric exclusion process)
452: we have a condensate since
453: site 1 contains a finite fraction
454: of the particles.
455: In the low density phase the particles are evenly spread
456: between all sites.
457:
458: \subsection{Homogeneous case}
459: \label{Sec:condhom}
460: We now consider the homogeneous zero-range process where
461: the hopping rates $u(n)$ are site independent.
462: Then (\ref{Fdef}) is independent of $\mu$ and reads
463: \begin{equation}
464: F(z) = \sum_{n=0}^\infty \prod_{m=1}^n \left[ \frac{z}{u(m)} \right]
465: \label{Fhom}
466: \end{equation}
467: The fugacity $z$ must be chosen so that $F$ converges or else we could
468: not have performed (\ref{Fdef}). Therefore $z$ is restricted
469: to $z \leq \beta$ where we
470: define $\beta$ to be the radius of convergence of $F(z)$.
471: From (\ref{Fhom}) we see that
472: $\beta$ is the limiting value of the
473: $u(m)$ {\it i.e.} the limiting value
474: of the hopping rate out of a site
475: for a large number of particles at a site. We interpret
476: (\ref{sad2}) as giving a relation between the density of holes (number
477: of holes per site) and the fugacity $z$.
478: The saddle point condition (\ref{sad2}) becomes
479: \begin{equation}
480: \label{sad3}
481: \phi = \frac{z F'(z)}{F(z)}
482: \end{equation}
483: Given that the rhs of (\ref{sad3}) is a monotonically increasing function
484: of $z$ we deduce that density of
485: particle increases with fugacity. However if at $z=\beta$, the
486: maximum allowed value of $z$, the rhs of (\ref{sad3}) is still finite
487: then one can no longer solve for the density and one must have a
488: condensation transition. Physically, the condensation would correspond
489: to a spontaneous symmetry breaking where one of the sites is
490: spontaneously selected to hold a finite fraction of the particles.
491:
492:
493: Thus, for condensation to occur ({\it i.e.} when $\phi$ is large enough
494: for (\ref{sad3}) not to have a solution for the allowed values of $z$) we
495: require
496: \begin{equation}
497: \lim_{z\to \beta} \frac{F'(z)}{F(z)} < \infty\; .
498: \end{equation}
499: We now assume that $u(n)$ decreases uniformly to $\beta$ in the large $n$
500: limit as
501: \begin{equation}
502: u(n) = \beta( 1 + \zeta(n) )
503: \end{equation}
504: where $\zeta(n)$ is a monotonically decreasing function.
505: Analysis of the series
506: \begin{eqnarray}
507: F(\beta) &=& \sum_{n=0}^{\infty}
508: \exp \left\{ - \sum_{m=1}^n \ln\left[1+\zeta(m)\right] \right\}
509: \nonumber \\
510: F'(\beta) &=& \sum_{n=0}^{\infty} n
511: \exp \left\{ - \sum_{m=1}^n \ln\left[1+\zeta(m)\right] \right\}
512: \label{series}
513: \end{eqnarray}
514: reveals that the condition for condensation is simply that
515: $F'(\beta)$ is finite and this occurs if $u(n)$ decays to $\beta$ more
516: slowly than $\beta(1+2/n)$.
517: (This is easiest to see by expanding $\ln
518: \left[1+\zeta\right]$ and approximating the sum over $m$ by an
519: integral in (\ref{series}).)
520:
521: It is interesting to translate this result
522: into the language of the
523: exclusion process. In this context we can have condensation if
524: the hop rate of a particle into a gap of size $n$ decays as
525: $\beta(1+2/n)$ therefore there is an effective long range interaction.
526:
527:
528: \subsection{Bus route model}
529: As an example of this let us consider the
530: `bus route model' \cite{OEC}.
531: The model
532: is defined on a $1d$ lattice. Each site (bus-stop) is
533: either empty, contains a bus (a conserved particle) or contains a
534: passenger (non-conserved quantity). The dynamical processes are that
535: passengers arrives at an empty site with rate $\lambda$; a bus moves
536: forward to the next stop with rate 1 if that stop is empty; if the
537: next stop contains passengers the bus moves forward with rate $\beta$
538: and removes the passengers.
539:
540:
541:
542: The model thus defined has not been solved but simulations reveal two
543: regimes. At high bus density the gaps between buses are evenly
544: distributed. However at low bus density there is a condensed regime
545: where the lead bus has a large gap to the next bus in front of it with
546: bus-stops full of passengers in between. The other buses have small
547: gaps between them. Thus the buses form a jam of buses and after a
548: long delay all arrive at a bus-stop at once.
549:
550: The bus route model can be related to the zero-range process by a
551: mean-field approximation in which we integrate out the non-conserved
552: quantity (passengers). The idea is that a bus-stop, next to
553: bus 1 say, will last have been visited by a bus (bus 2) a mean time
554: ago of $n/v$ where $n$ is the distance between bus 2 and bus
555: 1 and $v$ is the steady state speed.
556: Therefore the mean-field probability that the site next to bus 1
557: is not occupied by a passenger is $\exp(-\lambda n/v)$. From this
558: probability an effective hopping rate for a bus into a gap of size $n$
559: is obtained by averaging the two possible hop rates $1,\beta$:
560: \begin{equation}
561: u(n)=\beta+(1-\beta) \exp(-\lambda n/v)\;.
562: \end{equation}
563: We can now see that this mean-field approximation to the
564: bus-route model is equivalent to a homogeneous zero-range process discussed
565: earlier.
566:
567: Since $u(n)$ decays exponentially
568: the condition for a strict phase
569: transition in the thermodynamic limit is not met.
570: However on any {\em finite} system
571: for $\lambda$ sufficiently small,
572: an apparent condensation will be seen. In the bus route
573: problem this corresponds to the universally irritating
574: situation of all the buses on the route arriving at once.
575:
576:
577: \section{Conclusion}
578: We have shown how the zero range process exhibits two kinds of condensation transition. One is due to having an inhomogeneous system
579: i.e. we get condensation of particles onto the site with the slowest
580: hopping rate. Although the condensation is spatial
581: the mechanism is equivalent to Bose condensation in an
582: Ideal Bose Gas.
583:
584: The other type of condensation occurs on a homogeneous systems and
585: involves the spontaneous selection of a site onto which a finite
586: fraction of the particles condense. Recently this condensation
587: mechanism has been used to understand the existence or non-existence
588: of phase separation in a general class of one dimensional driven
589: systems \cite{KLMST,KLMT}.
590:
591:
592:
593:
594: % -----------------------------------------------------------------------------
595: % REFERENCES
596:
597: %\bibliography
598:
599: %\newpage
600: \begin{thebibliography}{99}
601: \bibitem{KLS}
602: S. Katz, J L Lebowitz and H Spohn (1983)
603: {\it Phys. Rev. B} {\bf 28} 1655; (1984) {\it J. Stat. Phys}{\bf 34}
604: 497
605: \bibitem{SZ}
606: B. Schmittmann and R K P Zia (1995)
607: {\it Statistical Mechanics of Driven Diffusive Systems} vol. 17
608: of Domb and Lebowitz series, Academic Press, U.K.
609: \bibitem{LL} L D Landau and E M Lifshitz, 1980, {\it Statistical Physics I}
610: (Pergamon Press, New York)
611: \bibitem{Krug91}
612: J. Krug (1991) {\it Phys. Rev. Lett.} {\bf 67} 1882
613: \bibitem{DEHP}
614: B. Derrida, M. R. Evans, V. Hakim and V. Pasquier (1993)
615: {\it J. Phys. A} {\bf 26} 1493
616: \bibitem{SD}
617: G. Sch\"utz and E. Domany (1993) {\it J. Stat. Phys.} {\bf 72} 277
618: \bibitem{Sasamoto}
619: T. Sasamoto (1999) {\it J. Phys. A} {\bf 32} 7109
620: \bibitem{BECE}
621: R. A. Blythe, M. R. Evans, F. Colaiori and F. H. L. Essler (2000)
622: {\it J. Phys. A} {\bf 33} 2313
623: \bibitem{Spitzer} F. Spitzer (1970) {\it Advances in Math.} {\bf 5} 246
624: \bibitem{MRE00}
625: M. R. Evans (2000)
626: {\it Brazilian Journal of Physics} {\bf 30} 42
627: \bibitem{MRE96}
628: M. R. Evans (1996) {\it Europhys. Lett.} {\bf 36} 13
629: \bibitem{KF}
630: J. Krug and P. A. Ferrari (1996) {\it J. Phys. A} {\bf 29} L465\
631: \bibitem{Mallick}
632: K. Mallick (1996) {\it J. Phys. A} {\bf 29} 5375
633: \bibitem{EKKM}
634: M. R. Evans, Y. Kafri, H. M. Koduvely and D. Mukamel (1998)
635: {\it Phys. Rev. Lett.} {\bf 80} 425; (1998)
636: {\it Phys. Rev. E} {\bf 58} 2764
637: \bibitem{LBR}
638: R. Lahiri, M. Barma and S. Ramaswamy (1997)
639: {\it Phys. Rev. Lett.} {\bf 79} 1150; (2000) {\it Phys. Rev. E}
640: {\bf 61} 1648
641: \bibitem{MKB}
642: S. N. Majumdar, S. Krishnamurthy, M. Barma (1998)
643: {\it Phys. Rev. Lett.} {\bf 81} 3691;
644: (2000) {\it Phys. Rev. E} {\bf 61} 6337
645: \bibitem{OE}
646: O. J. O'Loan and M. R. Evans (1999)
647: {\it J. Phys. A} {\bf 32} {L99}
648: \bibitem{OEC}
649: O. J. O'Loan, M.R.Evans and M.E.Cates (1998)
650: {\it Phys. Rev. E.} {\bf 58} 1404
651: \bibitem{KLMST}
652: Y. Kafri, E. Levine, D. Mukamel, G. M. Sch\"utz, and J. Torok (2002)
653: {\it Phys. Rev. Lett.} {\bf 89}, 035702
654: \bibitem{KLMT}
655: Y Kafri, E Levine, D Mukamel and J Torok (2002)
656: {\it J. Phys. A.} {\bf 35} L459
657:
658: \end{thebibliography}
659: \end{document}
660:
661:
662: \end{document}
663:
664:
665:
666:
667:
668:
669:
670:
671:
672:
673:
674:
675: