1: %\documentclass[a4paper,showpacs,superscriptaddress,preprint]{revtex4}
2: \documentclass[a4paper,showpacs,superscriptaddress,twocolumn]{revtex4}
3: \usepackage{epsfig}
4: \usepackage{graphics}
5: \usepackage{color}
6:
7: \def\bra#1{\langle #1 \vert}
8: \def\ket#1{\vert #1 \rangle}
9:
10:
11: \def\imagi{\mbox{\rm i}}
12: \def\beq{\begin{equation}}
13: \def\eeq{\end{equation}}
14:
15: \def\reff#1{(\ref{#1})}
16:
17: \begin{document}
18:
19: \title{Expansion of a Finite Size Plasma in Vacuum }
20:
21: \author{S. Betti}
22: \email{betti@df.unipi.it}
23: \affiliation{Dipartimento di Fisica ``Enrico Fermi'' \& INFM , Universit\`a di
24: Pisa, Largo B. Pontecorvo 2, 56127 Pisa,
25: Italy }
26: \author{F. Ceccherini}
27: \affiliation{Dipartimento di Fisica ``Enrico Fermi'' \& INFM , Universit\`a di
28: Pisa, Largo B. Pontecorvo 2, 56127 Pisa,
29: Italy }
30: \author{F. Cornolti}
31: \affiliation{Dipartimento di Fisica ``Enrico Fermi'' \& INFM , Universit\`a di
32: Pisa, Largo B. Pontecorvo 2, 56127 Pisa,
33: Italy }
34: \author{F. Pegoraro}
35: \affiliation{Dipartimento di Fisica ``Enrico Fermi'' \& INFM , Universit\`a di
36: Pisa, Largo B. Pontecorvo 2, 56127 Pisa,
37: Italy }
38:
39: \date{\today}
40:
41: \begin{abstract}
42: The expansion dynamics of a finite size plasma is examined from
43: an analytical perspective.
44: Results regarding the charge distribution as well as the
45: electrostatic potential are presented. The acceleration
46: of the ions and the associated cooling of the electrons that
47: takes place during the plasma expansion is described.
48: An extensive analysis of the transition between the semi
49: infinite and the finite size plasma behaviour is carried out.
50: Finally, a test of the analytical results, performed through numerical
51: simulations, is presented.
52: \end{abstract}
53:
54: \maketitle
55: The production of energetic particles through the interaction of
56: an intense laser
57: pulse with a solid target is a topic that has been widely investigated
58: during the last
59: three decades. For sufficiently long pulses, the emitted particles emerge
60: from a coronal plasma formed by the laser on the target foil and the expansion
61: of the plasma in vacuum plays a key role in such a phenomenon. Thus, despite in
62: Refs.\cite{mora, allen1, allen2, lontano} the expansion of a semi infinite
63: plasma has been widely investigated, present
64: experiments often involve ``thin foils'' as targets
65: \cite{borghesi, hegelich, badziak, mackinnon, maksimchuk, clark, snavely} which are
66: in some
67: cases no
68: thicker than a few tens of a Debye length, and we believe that, in
69: these cases, the experimental results must be analyzed
70: in terms of the expansion of a finite size plasma.
71: The finite-size and the semi-infinite cases are very different from each
72: other, as in the latter case
73: an infinite amount of energy is available. As a consequence,
74: even if the ions are
75: accelerated, the energy of the electron plasma
76: remains constant. On the contrary,
77: in the case of a
78: finite size plasma an exchange of kinetic energy between electrons
79: and ions takes
80: place as long as the two populations have different velocity distributions,
81: i.e.,
82: as long as a charge separation is present.
83: The aim of this work is to provide a detailed analytical
84: description of the thermal expansion of a globally neutral,
85: finite size, unidimensional plasma. Analytical predictions
86: will be compared with numerical results obtained
87: with a Particle in Cell (PIC) code.\\
88: The configuration
89: at the initial time $t_0$, which formally can be defined as
90: the hydrostatic equilibrium in the limit of infinitely
91: massive ions, is specified by the ion density
92: $
93: n_{i0}(x) \equiv n_i(x, t_0) = n_0~ \theta(a - |x|),
94: $
95: and by the Boltzmann-like electron profile
96: $
97: n_{e0}(x) \equiv n_e(x, t_0) = {\bar n}
98: ~ \exp{ \left( e \Phi(x, t_0) / T_{e0} \right) } $, where
99: $a$ is the half-thickness of the plasma, $2a n_0$ is the total number of
100: positively charged particle, $\theta(x) = 0$ for $x<0$ and $\theta(x) = 1$
101: for $x>0$, $\Phi(x, t_0)$ is the
102: electrostatic potential, ${\bar n}$ is
103: the density at the position
104: where $\Phi(x, t_0) = 0$, $T_{e0}$ the initial
105: electron plasma temperature and
106: $\int_{-\infty}^\infty n_{e0}(x)\mbox{d}x = 2a n_0$ because the system is
107: globally neutral. For the sake of notational simplicity we are considering
108: a plasma where the ion
109: charge is equal and opposite of that of the electrons.
110: Measuring space in
111: units of the initial electron Debye length $\lambda_{d,0}
112: = (T_{e0}/4 \pi e^2 n_0)^{1/2}$
113: and rescaling the potential $\Phi(x,t)$ as
114: $
115: \Phi(x,t) \to \phi(x,t) =e \Phi(x,t)/T_{e0},
116: $
117: we rewrite the Boltzmann equation in the simpler form
118: \begin{equation}\label{scaled}
119: n_{e0}(\zeta) = {\bar n} e^{\phi(\zeta)}.
120: \end{equation}
121: In order to obtain the electrostatic potential $\Phi(x, t_0)$,
122: and hence $n_{e0}(x)$,
123: we need to solve the Poisson equation
124: $
125: \nabla^2 \Phi(x, t_0) = -4 \pi e [n_{i0}(x)- n_{e0}(x)],
126: $
127: coupled to Eq.(\ref{scaled}).
128: Because of the inversion symmetry ($\zeta \to -
129: \zeta$) of the configuration, we will restrict our analysis to the half-plane
130: $\zeta \geq
131: 0$.
132: We choose the potential to be zero in $\zeta= a$, where $a$ denotes now the
133: normalized half
134: width of the plasma slab, so that ${\bar n} = n_{e0}(a)\equiv n_a$ and
135: Poisson
136: equation reads
137: \begin{equation}
138: \label{poisson}
139: \frac{d^2 \phi}{d\zeta^2} = \left( \frac{n_a}{n_0} e^{\phi (\zeta)} - 1 \right)
140: ~
141: \theta(a - \zeta) + \left( \frac{n_a}{n_0} e^{\phi (\zeta)} \right) ~
142: \theta(\zeta - a).
143: \end{equation}
144: The different functional form of
145: $n_{i0}(x)$ for $|x| \leq a$ (internal region) and for $|x| >
146: a$ (external region), implies that separate treatments are required.
147: First we will calculate the electron density at the ion front $x = a$.
148: Using this result, we will obtain an approximate analytical expression for the
149: electrostatic potential in the internal region, while in the external
150: region we will use
151: the analytical solution of Eqs.(\ref{scaled}, \ref{poisson}) found in
152: \cite{allen1}.\\
153: The value of $n_a$ can be derived in the following way. Integrating
154: Eq.(\ref{poisson}) once gives the electric field $E (\zeta)$ in the form
155: \begin{equation}
156: \label{E_int}
157: E(\zeta)
158: = \sqrt{2[(n_a/n_0) \left(e^{\phi(\zeta)} - e^{\phi(0)}
159: \right) - \left(\phi(\zeta) - \phi(0) \right) ]
160: }
161: \end{equation}
162: for $0 \leq \zeta \le a$ and, for $ \zeta > a$, in the form
163: \begin{equation}
164: \label{E_ext}
165: E(\zeta) = \sqrt{ 2(n_a/n_0) e^{\phi(\zeta)} } ,
166: \end{equation}
167: where the integration constants have been fixed such that
168: in Eq.(\ref{E_int}) $E(0) = 0$, as follows from the inversion symmetry,
169: and that in Eq.(\ref{E_ext}) $
170: \lim_{\zeta \to \infty} E(\zeta) = 0$
171: which expresses charge neutrality
172: (together with $\lim_{\zeta \to \infty} (n_a/n_0)
173: e^{\phi(\zeta)} = 0$ since there are no electrons at infinity).
174: Imposing the continuity of $E(\zeta)$ across the boundary $\zeta = a$
175: and using Eq.(\ref{scaled}), we obtain the following
176: relationship for the potential and the
177: electron density at $\zeta=0$
178: $
179: \phi (0) = (n_a/n_0) e^{ \phi(0) } = n_{e0}(0)/ n_0
180: $. The integration in
181: Eq.(\ref{E_int}) for $\phi(x)$ can not be performed explicitly. Numerical
182: integration
183: gives the electron density $n_{e0}(0)$ at $\zeta = 0$ and the electron density
184: $n_{a}$ at $\zeta = a$ as a function of the plasma length $a$.
185: %%%%%%%%%%%%%%%
186: \begin{figure}
187: \resizebox{0.4\textwidth}{!}{%
188: \includegraphics{figure1.eps}
189: }
190: \caption{ Behaviour of $n_a / n_0$ and of $n_{e0}(0) /
191: n_0$ as a function of $a$.}
192: \label{density.eps}
193: \end{figure}
194: %%%%%%%%%%%%%%%
195: Fig. \ref{density.eps} shows that, for $a \gg 1$, (i.e., in dimensional units
196: for
197: $a\gg \lambda_{d,0}$), we find
198: $\phi (0) = n_{e0}(0)/ n_0 \simeq 1$ and $ n_a/ n_0 \simeq 1/{\rm{e}}$, so that
199: for most
200: cases of interest, the electron density at the ion front is
201: equal to that obtained in the limit of a semi-infinite plasma.
202: The solution of Eq.(\ref{E_ext}), with the condition $\phi(a) = 0$,
203: is given by
204: \begin{equation}
205: \label{field scaling}
206: E = \frac{2 \sqrt{n_a/ n_0} }{\sqrt{n_a/ n_0} \left( \zeta - a \right) + \sqrt{2}},
207: \end{equation}
208: with electrostatic potential given by
209: \begin{equation}\label{tt}
210: \phi (\zeta) = \mathrm{ln} \frac{2}{ \left( \sqrt{n_a/ n_0}~(\zeta - a) + \sqrt{2} \right)^2 }.
211: \end{equation}
212: In order to derive an approximate solution of Eq.(\ref{E_int}),
213: we define a layer of thickness $\delta$ inside the plasma,
214: specified by the assumptions that
215: ${n_{e0}(\zeta)} \simeq {n_{0}}$ for $ 0 < \zeta < (a-\delta)$, where
216: $E \simeq 0$ and
217: $\phi \simeq \mathrm{const} = \phi(0)$,
218: and that the electron density distribution differs significantly from the
219: fixed ion distribution only for
220: $\zeta > (a - \delta)$. Inside this layer
221: we adopt an approximate
222: parabolic fit for the
223: electrostatic potential of the form
224: $
225: \phi(\zeta) = \mathcal{C}_1 (\zeta - a) + \mathcal{C}_2 (\zeta - a)^2/2
226: $,
227: where the constant term has been set equal to zero since, by continuity
228: with the external solution,
229: $
230: \phi(a) =0$. This fit of the potential
231: corresponds to a rough approximation in Poisson's equation (\ref{poisson})
232: where the electron density inside the
233: layer is taken to be constant and equal to an intermediate value between its
234: two values, $n_a$ and $n_0$, at the borders
235: of the layer. This value must be determined,
236: together with
237: $C_1$ and $\delta$, self-consistently by requiring
238: that the potential and the electric field be
239: continuous at $\zeta =a, a-\delta$, which
240: ensures in particular charge conservation.
241: From
242: the continuity of the electric field at $\zeta = a$ we obtain
243: from Eq.(\ref{field scaling})
244: $\mathcal{C}_1 = - \sqrt{2 / {\rm e} }$.
245: At $\zeta = a-\delta$, the continuity of the electric field gives
246: $\mathcal{C}_2 = - \mathcal{C}_1 / \delta$,
247: while the continuity of $\phi$ defines $\delta$ as $\delta =
248: \sqrt{2 {\rm e} } \phi_0 \approx \sqrt{2 {\rm e} }$.\\
249: We now derive a physical model capable of describing
250: the electron cooling that occurs as the plasma expansion takes place and,
251: consequently, ion acceleration.
252: Measuring temperatures in units of $T_0$, where $T_0 = T_{e0}$,
253: mass in units of the ion mass $m_i$, and
254: introducing the dimensionless time unit $\tau = \omega_{\rm pi} t$,
255: with $\omega_{\rm pi} = \sqrt{4 \pi n_0 e^2}$,
256: we denote with $T_e$ the electron time
257: dependent temperature,
258: with $m_e$ the electron mass and
259: assume that both the initial fluid and thermal ion energy
260: is zero. The quantity $T_e$ can in principle depend both on $x$ and $\tau$,
261: however in
262: what follows it will be assumed spatially constant.
263: In order to derive an analytical expression for
264: the cooling of the electron population we need
265: to introduce a simplified description of the
266: ion expansion and of the time evolution of the
267: electric field.
268: %%%%%%%%%%%%%%%
269: \begin{figure}
270: \resizebox{0.4\textwidth}{!}{%
271: \includegraphics{figure2.eps}
272: }
273: \caption{Model profile of the ion density as plasma
274: expands.}
275: \label{ion.eps}
276: \end{figure}
277: %%%%%%%%%%%%%%%
278: For this purpose, we introduce the approximate
279: ion profile $n_i(x,\tau)$ sketched in Fig.(\ref{ion.eps})
280: which is specified once the quantities $x_s$ and $x_f$ are known.
281: The point $x_s$ models the position of the rarefaction
282: front, which obeys the equation of motion
283: \begin{equation}
284: \label{x_s}
285: x_s = a - \delta - \int_0^{\tau}{c_s d \tau}.
286: \end{equation}
287: with $c_s$ the time dependent ion acoustic speed which,
288: in dimensionless units, is given by $c_s = \sqrt{T_e}$.
289: The quantity $x_f$ represents the ion front position, defined by
290: the condition $n_i(x, \tau) = 0$ for $x > x_f$, which obeys
291: the equation of motion
292: \begin{equation}
293: \label{x_f}
294: \ddot{x}_f = E_{f},
295: \end{equation}
296: with initial conditions $\dot{x}_f |_{\tau = 0} = 0$, $x_f |_{\tau = 0} = a$
297: and $E_{f}$ the self consistent electric field at point $x = x_f$.
298: Profile (\ref{ion.eps}) then becomes
299: $
300: n_i(x, \tau) = n_0 ~ \theta(x_s - |x|) + n_0 D ~ \theta(|x| - x_s)
301: - n_0 D ~ \theta(|x| - x_f)
302: $,
303: with $D = (a - x_s)/(x_f-x_s)$, which corresponds to
304: redistributing the charge $n_0 \left( a-x_s \right)$
305: on the interval $\left( x_f-x_s \right) ~ $\footnote{Note that this
306: profile can be
307: used only if $x_s \geq 0$.}.
308: Assuming hydrostatic equilibrium at every time $\tau$
309: between the electron and
310: ion plasma we find $~ D ~ n_0 / n_f = \rm{e} ~ $ for
311: the electron density $n_f$ at point $x = x_f$,
312: thus $E_f = \sqrt{ \left( 2 / {\rm e} \right) D T_e}$,
313: which completes the description of both the ion profile and dynamics.
314: Regarding the electrons, we assume quasineutrality in the whole region
315: $[0, x_f - \delta]$ except for a small interval $[x_s - \Delta x, x_s]$, with
316: $\Delta x$ to be determined, in
317: which charge separation occurs such that Boltzmann relation
318: (\ref{scaled}) can be satisfied.
319: The electric field $E_{cs}$
320: in $[x_s - \Delta x, x_s]$ is approximated by
321: $E_{cs} \simeq \left( 1 - D \right)$, which is the field of a plane
322: capacitor with surface charge density
323: $\sigma = \left( 1 - D \right) / 4 \pi$ in dimensionless units.
324: Therefore $- \int_{x_s - \Delta x}^{x_s} E_{cs} dx = T_e \left[ \phi \left( x_s \right)
325: - \phi \left( x_s - \Delta x \right) \right]$ which defines
326: $\Delta x$ as $\Delta x = - T_e (\ln{D})/(1-D)$.
327: The cooling phenomenon is associated with the reflections that the electrons
328: undergo at the potential barriers in the region $[x_s - \Delta
329: x, x_s]$ ({\it internal} barrier) and $[x_f - \delta,+\infty]$
330: ({\it external} barrier), which are moving with velocities
331: $-c_s$ and $v_f$, respectively. Assuming the reflections
332: to be elastic, $m_e \ll 1$ implies that the power loss due to a
333: reflection of an electron with velocity $v$ is given by
334: $
335: \Delta U (v) = 2m_e \left( v - \bar{v} \right) \bar{v}
336: $,
337: with $\bar{v} = \{ - c_s, ~ v_f \}$.
338: The corresponding power is given by $\Delta U(v)$
339: times the number of electrons with velocity $v >
340: \bar{v}$ that hit the barrier
341: per unit time, integrated over the electron distribution function.
342: Assuming the latter to be Maxwellian,
343: the number of electrons with velocity in the interval $[v, v+dv]$
344: reflected per unit time by the internal and external barrier
345: is given by $(1-D) n_0 v f(v)$ and $D n_0 v f(v) ~ \theta(v-v_f)$
346: respectively, where $f(v) = \left( 1 / \sqrt{2 \pi} ~ v_{th} \right)
347: e^{- v^2 / 2 v_{th} } $ and $v_{th} = \sqrt{T_{e}/m_e}$.
348: The total energy of the electron plasma is $U = n_0 a T_e/2$, hence, taking the
349: temperature $T_e$ to be uniform, $ d U /d \tau
350: = (n_0 a/2) ~{d T_e}/{d \tau}$ and,
351: taking $v_f, ~ c_s \ll v_{th}$, we find
352: \begin{equation}\label{cooling law}
353: \frac{1}{T_e} \frac{ d T_e }{d \tau}
354: = - \left( 2 / a \right) \left[D v_f-\left(1-D\right) c_s\right],
355: \end{equation}
356: which is the electron cooling equation, to be coupled both to Eq.(\ref{x_s})
357: and Eq.(\ref{x_f}) and then integrated numerically.
358: The time dependent total kinetic energy $U_i$ of the ions
359: satisfies the conservation equation $U_i = n_0 a / 2 - n_0 a T_e/2 - U_{el}$, where $U_{el}$
360: is the contribution from the release of the electrostatic energy present in the
361: initial ``equilibrium'' configuration described previously.
362: Note that $2 U_{el} / (T_e) \approx \left(2d U_{el} / dt \right) / \left(
363: d T_e/ dt \right) \approx
364: \delta / a$ and thus $U_i \approx 1 - T_e/2$ except for the very beginning
365: of the expansion process, when the electrostatic energy
366: contribution is significant and must be taken
367: into account.\\
368: We may summarize the results of the above modeling of the plasma expansion in
369: vacuum by referring to $a$, $\tau / (a \sqrt{m_e})$ and $\tau / \sqrt{a}$
370: as the
371: relevant physical quantities characterizing the plasma dynamics.
372: The condition $a \gg 1$ implies that, at every time $\tau$,
373: the electron density at the expanding front
374: is $n_f = D n_0 / {\rm e}$, as in the hydrodynamic equilibrium of a
375: semi-infinite slab. The quantity
376: $\tau / (a \sqrt{m_e})$, where $a \sqrt{m_e}$ is the
377: typical time it takes an electron to cross the target,
378: roughly gives the ratio of particles which, at time
379: $\tau$, have undergone at least one reflection at the ion front.
380: Thus the initial expansion phase when $\tau / (a \sqrt{m_e}) \ll 1$
381: corresponds to the semi-infinite, constant temperature
382: plasma limit discussed in \cite{mora, allen1, allen2}.
383: The quantity $\tau / \sqrt{a}$ measures, at the early stage of the
384: process, the expansion of the ion component in the plasma. In fact, taking
385: $v_f \propto \tau$, which corresponds to a uniformly accelerated motion of
386: the front and neglecting coefficients of order unity, we obtain $(x_f -a)/a
387: = dV/V \approx \tau^2/a$ where $V$ is the volume occupied by the ions. In
388: the initial expansion phase, from Eq.(\ref{cooling law}) we obtain
389: ${d T_e}/{T_e} \approx - 2 ~ (v_f \tau ) / a \approx - 2 ~{dV}/{V}$,
390: which implies that in this phase the plasma behaves like a perfect
391: monoatomic gas obeying the law of unidimensional
392: adiabatic expansion.
393: %%%%%%%%%%%%%%%
394: \begin{figure}
395: \resizebox{0.4\textwidth}{!}{%
396: \includegraphics{figure3.eps}
397: }
398: \caption{\label{parameter.eps} Plasma expansion regimes in
399: the $(a,\tau)$-plane.
400: Electron cooling is important in region C. The uniform electron
401: temperature assumption used
402: in our analytical model applies for $a< 1/m_e$.}
403: \end{figure}
404: %%%%%%%%%%%%%%%
405: %%%%%%%%%%%%%%
406: \begin{figure}
407: \resizebox{0.4\textwidth}{!}{%
408: \includegraphics{figure4.eps}
409: }
410: \caption{\label{T.eps} Total normalized ion kinetic energy versus time $\tau$
411: as obtained from the analytical model (dashed line) and from a PIC simulation
412: (solid line) in a plasma with $a = 50 ~ \lambda_d$.}
413: \end{figure}
414: %%%%%%%%%%%%%%%
415: %%%%%%%%%%%%%%
416: \begin{figure}
417: \resizebox{0.4\textwidth}{!}{%
418: \includegraphics{figure5.eps}
419: }
420: \caption{\label{confronto.eps} Normalized electric field $E_f$
421: at the ion front (a) and ion front position $x_f$ (b) versus
422: time $\tau$ as obtained
423: from the analytical
424: model (dashed line) and
425: from the PIC simulation (solid line),
426: in a plasma with $a = 50 ~ \lambda_d$.
427: The dotted line corresponds to semi-infinite case \cite{mora}.}
428: \end{figure}
429: %%%%%%%%%%%%%%%
430: Thus, the plane $(a, \tau)$ can be divided into four regions by the
431: two curves $\tau = \sqrt{a}$ and $\tau = a \sqrt{m_e}$,
432: as shown in Fig.(\ref{parameter.eps}). The two curves cross at $a =1/m_e$.
433: In the three regions $A, B, D$ defined by
434: $\tau < max (a \sqrt{m_e} ,\sqrt{a})
435: $, electron cooling is not important and the slab expansion can be
436: approximated by the semi infinite case.
437: In region $C$ the effects of the
438: the finite size of the plasma becomes apparent and ion acceleration
439: takes place together with electron cooling.\\
440: Now we reexamine the approximations
441: used in the derivation of Eq.(\ref{cooling law}).
442: The assumption of uniform
443: temperature rests on the comparison between
444: the energy redistribution time, approximated by $a \sqrt{m_e}$,
445: and the typical ion plasma expansion time
446: $\sqrt{a}$, which roughly represents the energy
447: confinement time. Therefore the temperature
448: inside the plasma can be taken to be uniform
449: as long as $a m_e \ll 1$ (the area to the left of the vertical dotted line in
450: Fig.(\ref{parameter.eps})) and the predictions of our analytical model
451: are thus restricted to this domain which, however,
452: covers the typical experimental configurations
453: \cite{mackinnon, snavely}. The approximation that the electron
454: reflections are elastic in the frame co-moving with the reflecting barrier,
455: can be written as
456: $\sqrt{m_e} {\partial}[ \ln{(
457: T_e \phi )}] /{\partial \tau} \ll 1$, where we
458: take the electron reflection time to be of order $\sqrt{m_e}$.
459: The validity of this condition has been verified by integrating
460: Eqs.(\ref{x_s}, \ref{x_f}, \ref{cooling law}) numerically. Finally,
461: because of the significant spread of the ion front, the effective
462: value of $E_f$ is smaller than the one obtained within this model.
463: This may be corrected by introducing the reduction factor
464: $ 1 - \left( 1 - D \right)^\alpha$, with $\alpha$ a constant to be fixed
465: numerically.
466: In order to determinate the value of $\alpha$ and to verify
467: our analytical results, a series of numerical simulations has
468: been performed using a unidimensional PIC code.
469: The numerical results that we present here describe the expansion of
470: a plasma slab $50 ~ \lambda_d$ thick with
471: initial conditions given by Eqs.(\ref{scaled},\ref{tt}).
472: An overall good agreement between the analytical and
473: the numerical results is achieved with $\alpha =
474: 1.4$. The ion total kinetic
475: energy and the
476: value of $E_f$ and the position of the ion front are plotted versus time in
477: Figs.(\ref{T.eps}) and (\ref{confronto.eps}),
478: respectively. Different simulations for slabs up to $150 \lambda_d$
479: thick have been performed and have
480: confirmed the validity of the
481: analytical model and in particular of the choice of the value
482: of the parameter $\alpha$. The comparison between the finite size
483: results and those described in \cite{mora} is shown in
484: Fig.(\ref{confronto.eps}) and indicates that the semi-infinite and
485: the finite size plasma behaviours are
486: very similar only at the early stage of the expansion process,
487: while they become, as expected, substantially different when
488: the electron kinetic energy is
489: significantly reduced.\\
490: In conclusion our analytical model shows that, in the case of a finite size
491: plasma, the process of electron cooling leads to a nearly complete
492: energy transfer from electrons to ions on time intervals of the order of a few
493: tens of the transit time $a \sqrt{me}$. This sets
494: an upper limit to
495: the energy ions can acquire in the thermal expansion of a finite size plasma.
496: Furthermore, different expansion regimes as function of time and plasma size have
497: been identified and the numerical verification of the proposed model
498: has been discussed.\\
499: This work was supported by the INFM Parallel Computing Initiative.
500:
501:
502: \begin{thebibliography}{300}
503:
504: \bibitem{mora} P. Mora, Phys. Rev. Lett. {\bf 90}, 185002 (2003).
505:
506: \bibitem{allen1} J. E. Allen, J. G. Andrews, J. Plasma Physics,
507: \textbf{4}, 187 (1970)
508:
509: \bibitem{allen2} J. E. Crow, {\it et al.}, P. L. Auer, J. E. Allen,
510: J. Plasma Physics,\textbf{14}, 65 (1975)
511:
512: \bibitem{lontano} M. Passoni, {\it et al.}, Phys. Rev. E {\bf 69}, 026411
513: (2004).
514:
515: \bibitem{borghesi} M. Borghesi {\it et al.}, Phys. Rev. Lett. {\bf 92}, 055003
516: (2004).
517:
518: \bibitem{hegelich} M. Hegelich {\it et al.}, Phys. Rev. Lett. {\bf 89}, 085002
519: (2002).
520:
521: \bibitem{badziak} J. Badziak {\it et al.}, Phys. Rev. Lett. {\bf 87}, 215001
522: (2001).
523:
524: \bibitem{mackinnon} A. J. Mackinnon {\it et al.}, Phys. Rev. Lett. {\bf 86},
525: 1769
526: (2001).
527:
528: \bibitem{maksimchuk} A. Maksimchuk {\it et al.}, Phys. Rev. Lett. {\bf 84},
529: 4108
530: (2000).
531:
532: \bibitem{clark} E. L. Clark {\it et al.}, Phys. Rev. Lett. {\bf 85}, 1654
533: (2000).
534:
535: \bibitem{snavely} R. A. Snavely {\it et al.}, Phys. Rev. Lett. {\bf 85}, 2945
536: (2000).
537:
538: \end{thebibliography}
539:
540: \end{document}
541:
542:
543:
544:
545: