physics0401145/hal.tex
1: %\documentstyle[psfig]{article}
2: \documentclass[12pt]{article}
3: 
4: \usepackage{epsfig}
5: 
6: % pour un interligne de un et demi
7: %\linespread{1.3}
8: 
9: % pour un double interligne 
10: %\linespread{1.6}
11: 
12: \setcounter{secnumdepth}{4} 
13: 
14: %%%%%%%%%%%%% serie de definition pour des commandes abregees %%%%%%%%%%%%%%%%%
15: % ces commandes sont celles du texte de notes sur  les plasmas.
16: 
17: % pour abreger les commandes d'equations
18: \newcommand{\beq}{\begin{equation}}
19: \newcommand{\eeq}{\end{equation}}
20: 
21: %pour definir le style de repr\'esentation des vecteurs et des tenseurs
22: \newcommand{\ve}[1]{{\vec{\bf #1}}}
23: \newcommand{\te}[1]{{\vec{\vec{\bf #1}}}}
24: \newcommand{\ttt}[1]{{ \vec{\vec{\vec{\bf #1}}} }}
25: \newcommand{\qe}[1]{{\vec{\vec{\vec{\bf #1}}}}}
26: 
27: %pour definir des vecteurs bares et des tenseurs bares
28: \newcommand{\vba}[1]{{\bf{\bar #1}}}
29: \newcommand{\tba}[1]{{\bf{\bar{\bar #1}}}}
30: %pour definir des moyennes avec des bra-kets
31: \newcommand{\moy}[1]{\left< #1 \right>}
32: 
33: %pour definir des derivees
34: 
35: % cas general
36: \newcommand{\dpar}[2]{\frac{\partial  #2}{\partial #1}}
37: \newcommand{\dtot}[2]{\frac{d #2}{d #1}}
38: 
39: %cas particuliers frequents
40: \newcommand{\p }[1]{\dot{#1}}
41: \newcommand{\dq }[1]{\frac{\partial  #1}{\partial q}}
42: \newcommand{\dpp }[1]{\frac{\partial  #1}{\partial p}}
43: \newcommand{\dqp}[1]{\frac{\partial  #1}{\partial \dot{q}}}
44: \newcommand{\dqi }[1]{\frac{\partial  #1}{\partial q_i}}
45: \newcommand{\dqpi}[1]{\frac{\partial  #1}{\partial \dot{q}_i}}
46: 
47: \newcommand{\dt }[1]{\frac{\partial  #1}{\partial t}}
48: 
49: \newcommand{\dr }[1]{\frac{\partial #1}{\partial \ve{r}}}
50: \newcommand{\dx }[1]{\frac{\partial #1}{\partial x}}
51: \newcommand{\dy }[1]{\frac{\partial #1}{\partial y}}
52: \newcommand{\dz }[1]{\frac{\partial #1}{\partial z}}
53: 
54: \newcommand{\dvx }[1]{\frac{\partial #1}{\partial v_x}}
55: \newcommand{\dvy }[1]{\frac{\partial #1}{\partial v_y}}
56: \newcommand{\dvz }[1]{\frac{\partial #1}{\partial v_z}}
57: \newcommand{\dv }[1]{\frac{\partial #1}{\partial {\ve v}}}
58: 
59: \newcommand{\dwx }[1]{\frac{\partial #1}{\partial w_x}}
60: \newcommand{\dwy }[1]{\frac{\partial #1}{\partial w_y}}
61: \newcommand{\dwz }[1]{\frac{\partial #1}{\partial w_z}}
62: \newcommand{\dw }[1]{\frac{\partial #1}{\partial {\ve w}}}
63: 
64: \newcommand{\Dt }[1]{\frac{d #1}{d t}}
65: \newcommand{\Dx }[1]{\frac{d #1}{d x}}
66: \newcommand{\Dy }[1]{\frac{d #1}{d y}}
67: \newcommand{\Dz }[1]{\frac{d #1}{d z}}
68: \newcommand{\Dr }[1]{\frac{d #1}{d r}}
69: 
70: \newcommand{\Dsdt }[1]{\frac{d }{d t}(#1)}
71: \newcommand{\Dsdx }[1]{\frac{d }{d x}(#1)}
72: \newcommand{\Dsdy }[1]{\frac{d }{d y}(#1)}
73: \newcommand{\Dsdz }[1]{\frac{d }{d z}(#1)}
74: \newcommand{\Dsdr }[1]{\frac{d }{d r}(#1)}
75: 
76: \newcommand{\dsdt }[1]{\frac{\partial}{\partial t}(#1)}
77: \newcommand{\dsdx }[1]{\frac{\partial}{\partial x}(#1)}
78: \newcommand{\dsdy }[1]{\frac{\partial}{\partial y}(#1)}
79: \newcommand{\dsdz }[1]{\frac{\partial}{\partial z}(#1)}
80: \newcommand{\dsdr }[1]{\frac{\partial}{\partial r}(#1)}
81: 
82: \newcommand{\dsdvx }[1]{\frac{\partial}{\partial v_x}(#1)}
83: \newcommand{\dsdvy }[1]{\frac{\partial}{\partial v_y}(#1)}
84: \newcommand{\dsdvz }[1]{\frac{\partial}{\partial v_z}(#1)}
85: \newcommand{\dsdv }[1]{\frac{\partial}{\partial {\ve v}}(#1)}
86: 
87: \newcommand{\dsdwx }[1]{\frac{\partial}{\partial w_x}(#1)}
88: \newcommand{\dsdwy }[1]{\frac{\partial}{\partial w_y}(#1)}
89: \newcommand{\dsdwz }[1]{\frac{\partial}{\partial w_z}(#1)}
90: \newcommand{\dsdw }[1]{\frac{\partial}{\partial {\ve w}}(#1)}
91: 
92: \newcommand{\dive}{\mbox{$\nabla \cdot$}}
93: \newcommand{\rot}{\mbox{$\nabla \times$}}
94: \newcommand{\grad}{\mbox{$\nabla$}}
95: \newcommand{\lap}{\mbox{$\Delta$}}
96: 
97: 
98: 
99: %%%%%%%%%%%%%%%%%%%%%%%%% DOCUMENT %%%%%%%%%%%%%%%%%%%%%%%%%%%
100: \begin{document} 
101:  
102: 
103: 
104: \title{
105: Exact nonlinear analytic Vlasov-Maxwell tangential equilibria with arbitrary density and temperature profiles} 
106:  
107: 
108:  
109: \author{F. Mottez, \\
110: %F. Mottez, 
111: Centre d'\'etude des Environnements Terrestre et Plan\'etaires (CETP),  \\
112: 10-12 Av. de l'Europe,
113: 78140 V\'elizy, France. \\
114: (e-mail: fabrice.mottez@cetp.ipsl.fr)} 
115: 
116: %\small{Centre d'\'etude des Environnements Terrestre 
117: %et Plan\'etaires, V\'elizy, France} 
118: 
119: \maketitle
120: 
121: \begin{abstract} 
122: The tangential layers are characterized by a bulk plasma velocity 
123: and a magnetic field that are perpendicular to the gradient direction.
124: They have been extensively described in the frame of the Magneto-Hydro-Dynamic (MHD) theory.
125: But the MHD theory does not look inside the transition region
126: if the transition has a size of a few ion gyroradii. 
127: A series of kinetic tangential 
128: equilibria, valid for a collisionless plasma is presented. 
129: These equilibria are exact analytical solutions of the Maxwell-Vlasov equations.
130: The particle distribution functions
131: are sums of an infinite number of elementary
132: functions parametrized by a vector potential.
133: Examples of equilibria relevant to space plasmas are shown. 
134: A model for the deep and sharp density depletions observed in 
135: the auroral zone of the Earth is proposed. Tangential equilibria are also relevant for 
136: the study of planetary environments and of
137: remote astrophysical plasmas.
138: \end{abstract} 
139: 
140: {\tiny
141: \begin{copyright}
142: Copyright (2003) American Institute of Physics. This article may be downloaded 
143: for personal use only. Any other use requires prior permission of 
144: the author and the American Institute of Physics. 
145: The following article appeared in Phys. Plasmas, Vol 10, 7, pp2501,2508, and may be found at 
146: http://link.aip.org/link/?php/10/2501.
147: \end{copyright}
148: }
149: 
150: \newpage
151: 
152: \section*{I. INTRODUCTION}
153: 
154: The swedish Viking spacecraft orbiting in the auroral region 
155: of the magnetosphere discovered deep plasma density depletions 
156: \cite{Higers et al 1992a}. Their size along the 
157: magnetic field direction reach thousands of kilometers.
158: They can be considered as channels parallel to the magnetic field.
159: These observation were confirmed by measurements onboard the 
160: Fast Auroral SnapshoT explorer (FAST) spacecraft
161: \cite{Strangeway et al 1998}.
162: The data showed a lot of plasma activity 
163: in and around these regions, particularly plasma acceleration 
164: (at the origin of the polar auroras) \cite{McFadden et al 1998}, 
165: electromagnetic radiation \cite{Higers 1992b}
166: and a high level of electrostatic turbulence.
167: Most of the studies focused on the acceleration 
168: \cite{Genot et al 1999}, \cite{Genot et al 2000}
169: and the nature of the turbulence \cite{Roux et al 1993}, \cite{Pottelette et al 2001}
170: \cite{Genot et al 2001}.
171: But surprisingly, little work has been devoted to the structure 
172: of these plasma cavities. Some theoretical works have been devoted to their origin,
173: but none has tackled the question : are auroral plasma cavities in equilibrium,
174: or do they vanish a soon as the cause of their formation disapears ?
175: 
176: The aim of the work presented in this paper is to construct
177: a kinetic model of the auroral plasma cavities. 
178: But as a very general solution to this problem is found, the
179: field of applications appears much wider.
180: I  present here a family of equilibrium solutions that can be used in many domains
181: of the physics of plasmas. They will be named as tangential equilibria because,
182: as in the MHD tangential discontinuities, the magnetic field and the bulk
183: velocities are tangential to the plane $(y,z)$ of invariance.
184: 
185: Previous works on tangential equilibria \cite{Harris 1962}, \cite{Channell 1976}
186: have shown the existence of simple isothermal
187: solutions that can be computed analytically. Other works, mainly focused on the 
188: study of the Earth magnetopause \cite{Roth et al 1996}, have shown
189: solutions that satisfy a larger class of constraints, but they cannot be constructed
190: analytically.
191: %%%%%%%%%%
192: 
193: %%%%%%
194: Most of these equilibria a based, for each species, on distribution functions that 
195: are products of an exponential function of the total energy and of $p_z^2$, by
196: an arbitrary $g$ function of $p_y$,
197: \begin{equation} \label{densite-distri-g-general0}
198: f= (\frac{\alpha_{z} \alpha_{\perp }^2}{\pi^3})^{1/2} 
199: [\exp{(-\alpha_{z} v_{z}^2  -\alpha_{\perp } v_{\perp }^2)} ] \,
200: g(mv_y +q A_y(x)),
201: \end{equation}
202: where $\alpha_{z}=m/2T_{z}=1/v_{tz}^2$, 
203: $\alpha_{\perp }=m/2T_{\perp }=1/v_{t \perp }^2$,
204: are the reciprocals of the squared thermal velocities ; 
205: $v_{z}=p_{z}/m$
206: and $v_{\perp}^2=v_{x}^2+v_{y}^2$. 
207: If $g(\infty) \sim 1$,
208: then $f$ is a bi-maxellian distribution for $x \sim \infty$.
209: 
210: Krall and Rosenbluth, \cite{Krall Rosenbluth 1963} have derived a solution of
211: the linearized kinetic equations, by setting $g(p_y)$ as a linear function. 
212: They used this simple solution for the study of gradient instabilities.
213: 
214: Many authors (reviewed by Roth {\em et al.} \cite{Roth et al 1996}), 
215: who studied tangential discontinuities,  considered 
216: solutions  with $g$ functions that involve 
217: error functions.
218: Most of these studies where developed for the study of 
219: the equilibrium of the Earth magnetopause.
220: They include an electric field $E_x$ and no charge neutrality.
221: There is no analytical solution. The differential equations on $E_x$ and $A$ are solved 
222: through a numerical integration.
223: 
224: Channell \cite{Channell 1976} considered solutions 
225: with no electric field. He showed that 
226: \beq \label{densite-g-Jy-dn-dAy}
227:  J_y= T_{\perp} \frac{dn}{dA_y}.  
228: \eeq
229: A more general formula will be shown in Sec. II.
230: 
231: In the same paper, Channell studied a few examples
232: of solutions with simple but nonlinear $g$ functions. In particular, he 
233: showed that there exists a solution with $g(p_y)=\exp{(-\eta p_y^2)}$, 
234: where $\eta$ is a constant number.
235: 
236: Harris \cite{Harris 1962} built an equilibrium model of a current sheet. 
237: His model was widely used later for the study of the tearing mode instability.
238: The Harris current sheet equilibrium corresponds to the case 
239: $g(p_y)=\exp{(\nu p_y)}$.
240: 
241: All the solutions mentionned above are based on elementary $g$ functions. 
242: They do not let the possibility of setting {\em a priori} a density profile, as
243: in the more simple MHD and bi-fluid models.
244: We can build a distribution function that is the sum of such elementary 
245: distribution functions and assert that each elementary part of the distribution
246: corresponds to a family of trapped particles. 
247: 
248: Because of the nonlinearity of the equations, the density and temperature profiles
249: associated to a sum of families of trapped particles are not simple combinations
250: of the profiles associated to single elementary particle distributions.
251: Nethertheless, using a numerical equation solver,
252: Roth {\em et al}. \cite{Roth et al 1996} added up to four elementary distributions functions
253: (all of them were built with error functions). Differential equations were solved numerically.
254: %%%%%%%%%%%%
255: 
256: 
257: I show in this paper how to combine an infinite number of elementary 
258: solutions and build a plasma that is made of a continuum  
259: of famillies of trapped particles.
260: The plasma equilibria
261: presented in this paper are computed analytically. They do not have the high degree 
262: of generality of those shown in Roth {\em et al.} \cite{Roth et al 1996}, 
263: (because I impose a null electric field), but they
264: are more general than those of Harris \cite{Harris 1962}, and Channell \cite{Channell 1976}. In particular, the plasma
265: does not have to be isothermal.
266: 
267: The equations of the kinetic model are developed 
268: and solved in Sec. II.
269: The properties of the solutions and some particular cases
270: are briefly analysed in Sec. III. 
271: An equilibrium model for the deep non isothermal density depletions of 
272: the Earth auroral zone is presented.
273: The last section is a discussion about 
274: other applications and further developments.
275: 
276: 
277: \section*{II. FORMULATION OF THE MAXWELL-VLASOV MODEL} %\label{section-equations}
278: We consider a monodimensional equilibrium : $\partial_t =0$ and $
279: \nabla =(d_x,0,0)$. We choose the $z$ direction along the constant 
280: direction of the magnetic field
281: $\ve{B}=(0,0,B_z(x))$, ($\nabla \cdot B =0$ is therefore ensured).
282: The magnetic field derives from a vector potential  
283: $\ve{A}(x)= (0,A_y(x),0)$ such that $B_z=d_x A_y$. This vector potential
284: satisfies the Lorentz gauge (that is $\dive \ve{A}=0$ in the stationary case).
285: The $x$ direction is called the normal direction, 
286: the $y,z$ plane is the tangential plane. 
287: 
288: In a monodimensional equilibrium, the
289: particles are confined between two points where $v_x=0$. In the case of a null electric field, 
290: the size $\Delta x$ of this area is 
291: \beq
292: \Delta x = 2 v_{\perp} /<\omega_c>
293: \eeq
294: where
295: $<\omega_c>=(q/m \Delta x) \int_{x_1}^{x_2} B_z(x)$.
296: Absolutely no restriction on the variations of $B_x$ is made. 
297: 
298: As we look for a monodimensional  equilibrium, the
299: total energy $E=\frac{mv^2}{2}+q\Phi$ of any particle and the generalized
300: momentum $p_y=mv_y + q A_y$ and $p_z=mv_z$ are the invariants of the motion.
301: Any distribution function of the form $f=f(E,p_y,p_z)$
302: is a solution of the Valsov equation. All the kinetic solutions studied in this
303: paper will be of that form.
304: The other character imposed on the solutions is a null electric field, 
305: and therefore no charge separation.
306: 
307: 
308: The basic idea is to decompose linearly the $g$ function 
309: (already mentionned in Eq. \ref{densite-distri-g-general0} ) over a set of elementary 
310: distribution functions that correspond to an analytical equilibrium solution. 
311: As will be shown in this section, if the sum is parametrized by a parameter $a$
312: that correspond to shifts of the elementary distribution functions in the
313: space of the potential vectors (and not in the space of the configurations),
314: there is an analytical solution.
315: 
316: For instance, we can show that a continuous linear combination of Channell and 
317: Harris-like $g$ distribution functions given by 
318: \beq \label{distribution-gy0}
319: g({p_y})= n_0 + \int_{a_{1}}^{a_{2}} da \; n_g(a) 
320: \exp{[-\eta(a) (\frac{p_y}{m}-\frac{q}{m}a)^2 +\nu(a)(\frac{p_y}{m}-\frac{q}{m}a)]},
321: \eeq
322: leads to an equilibrium that can be computed analytically.
323: 
324: The functions $n_g(a)$, $\eta(a)>0$, and $\nu(a)$ are defined almost arbitrarily 
325: and $n_0$ is a constant scalar.
326: Each of these three functions provide a degree of freedom and it is possible, 
327: as in  MHD and bi-fluid models, to control the plasma density profile. 
328: But the distribution of the particles depends on the energy 
329: through an exponential factor that does not give any freedom to set the temperature profile.
330: Therefore, instead of using \ref{densite-distri-g-general0} and \ref{distribution-gy0},
331: let us set : 
332: \beq \label{distribution-general}
333: f= \int_{a_1}^{a_2} da \; (\frac{\alpha_{z}(a) \alpha_{\perp }^2(a)}{\pi})^{1/2} 
334: e^{(-\alpha_{z}(a) (v_{z}-u_{z})^2  -\alpha_{\perp }(a) v_{\perp }^2)} 
335: (n_0 + g_{a}({p_y}))
336: \eeq
337: and 
338: \beq \label{distribution-gy}
339: g_{a}({p_y})= n_g(a) 
340: e^{-\eta(a) (\frac{p_y}{m}-\frac{q}{m}a)^2 +\nu(a)(\frac{p_y}{m}-\frac{q}{m}a)}.
341: \eeq
342: With these distribution functions, a control of the plasma density profile (through 
343: $n_g(a)$) and 
344: of the temperature profile of each species 
345: (through $\alpha_{z}(a)$ and $\alpha_{\perp }(a)$) is possible.
346: The parameter $a$ is homogeneous to a vector potential. 
347: 
348: In order not to overload the equations, we do not express systematically 
349: the dependance of the parameters on $a$ and $x$ hereafter.
350: 
351: 
352: To solve the charge neutrality equation ($n_i=n_e$) 
353: and the Amp\`ere equation
354: \beq \label{densite-ampere}
355: - \frac{d^2}{d x^2} A_y=\mu_0 J_y, 
356: \eeq
357: we need to compute the contribution of each species $s$ to the charge 
358: density,
359: \begin{equation} \label{densite-densite-cinetique-g}
360: n_s =  \int_{a_1}^{a_2} da \;  (\frac{\alpha_{\perp}}{\pi})^{1/2} 
361: \int  dv_y \; \exp{( -\alpha_{\perp} v_y^2)} 
362: g(mv_y +q A_y(x)).
363: \end{equation}
364: and to the current density,
365: \begin{eqnarray} \label{densite-courant-cinetique-g}
366: J_{ys} &=& \int_{a_1}^{a_2} da \; q  (\frac{\alpha_{\perp}}{\pi})^{1/2}
367: \int dv_y  \; v_y \exp{( -\alpha_{\perp} v_y^2)} 
368: g(mv_y +q A_y(x)) \\
369: &=& \int_{a_1}^{a_2} da \; q  (\frac{1}{ 4 \pi \alpha_{\perp}})^{1/2}
370: \int dv_y  \;  \exp{( -\alpha_{\perp} v_y^2)} 
371: g'(mv_y +q A_y(x)) 
372: \end{eqnarray}
373: where $g'$ is the derivative of $g$. 
374: These charge and current densities can be computed explicitely. 
375: We reorder Eq. (\ref{distribution-general}) and (\ref{distribution-gy})
376: to separate the terms 
377: that depend on $v_y$ and $v_y^2$,
378: \begin{eqnarray}
379: \label{densite-s3-f-ordered}
380: f= & \int_{a_1}^{a_2}da \;  \sqrt{\frac{\alpha_{\perp}^2 \alpha_z}{\pi^3}} 
381:  e^{-\alpha_{\perp} v_x^2 -\alpha_z (v_z-u_z)^2} \times \\ 
382: \nonumber
383: &   
384:  \{n_0 e^{-\alpha_{\perp} v_y^2}  +    
385: n_g(a) 
386: e^{-\eta(a) [\frac{q}{m}(A_y(x) -a)]^2 + \nu(a) \frac{q}{m} (A_y(x)-a)}
387: e^{-P(a) v_y^2 +2 Q(a) v_y}
388:  \},
389: \end{eqnarray}
390: with 
391: \beq
392: P(a) = \alpha_{\perp} + \eta(a) \;\;\; \mbox{et}\;\; 
393: Q(a)=\frac{1}{2} \nu(a) -\eta(a) \frac{q}{m} (A_y(x)-a)
394: \eeq
395: The contribution of each species $s$ to the charge and to the current densities
396: are 
397: \beq
398: n_s=n_0 +\int_{a_1}^{a_2} da \; n_g \sqrt{\frac{\alpha_{\perp}}{\pi}} e^{-\eta (q/m)^2 A_y^2} S_{0,P,Q}
399: \eeq
400: where $S_{0,P,Q}$ is defined by 
401: \beq
402: S_{0,P,Q}= \int_{-\infty}^{+\infty} dv_y e^{(-P v_y^2 +2 Q v_y)} 
403: =
404: \sqrt{\frac{\pi}{P}} e^{(\frac{Q^2}{P})} \\
405: \eeq
406: and 
407: \beq
408: J_{ys}=\int_{a_1}^{a_2} da \; q  n_g \sqrt{\frac{\alpha_{\perp}}{\pi}} e^{-\eta (q/m)^2 A_y^2} 
409: S_{1,P,Q},
410: \eeq
411: where 
412: \beq
413: S_{1,P,Q} =\int_{-\infty}^{+\infty} dv_y \, v_y e^{(-P v_y^2 +2 Q v_y)}=
414: \frac{Q}{P} \sqrt{\frac{\pi}{P}} \exp{(\frac{Q^2}{P})}.
415: \eeq
416: Hence,
417: \begin{eqnarray} \label{densite-s3-densite-charge}
418: n_s&=&n_0 + \int_{a_1}^{a_2} da \,
419: n_{g} \sqrt{\frac{\alpha_{\perp}}{\alpha_{\perp}+\eta}} 
420: e^{\frac{\nu^2}{4(\alpha_{\perp}+ \eta)}}
421: e^{- \frac{\alpha_{\perp}}{\alpha_{\perp}+\eta} 
422: [\frac{q}{m} (A_y-a)][\eta \frac{q}{m} (A_y-a) -\nu]} \\ \nonumber
423: J_{ys}&=&-q  \int_{a_1}^{a_2} da \,
424: \frac{n_g}{2} \sqrt{\frac{\alpha_{\perp}}{\alpha_{\perp}+\eta}} 
425: e^{\frac{\nu^2}{4(\alpha_{\perp}+ \eta)}}
426: \frac{ 1 }{\alpha_{\perp}+\eta} \times \\
427: \label{densite-s3-densite-courant}
428:  &&[-2 \eta \frac{q}{m} (A_y-a)+\nu] 
429:  e^{- \frac{\alpha_{\perp}}{\alpha_{\perp}+\eta} 
430: [\frac{q}{m} (A_y-a)][\eta \frac{q}{m} (A_y-a) -\nu]}.
431: \end{eqnarray}
432: Let  
433: \begin{eqnarray} \label{densite-s3-xi}
434: \xi(a) &=& \frac{\alpha_{\perp}(a) \eta(a)}{\alpha_{\perp}(a)+\eta(a)}(\frac{q}{m})^2   \\
435: \label{densite-s3-delta}
436: \delta(a)&=&\frac{\alpha_{\perp}(a) \nu(a)}{\alpha_{\perp}(a)+\eta(a)}(\frac{q}{m})        \\
437: \label{densite-s3-N0}
438: N_0(a)&=&  n_g(a) \sqrt{\frac{\alpha_{\perp}(a)}{\alpha_{\perp}(a)+\eta(a)}}
439: e^{\frac{\nu^2}{4(\alpha_{\perp}+ \eta)}}.
440: \end{eqnarray}
441: The contribution of each particle species to the particle density is:
442: \beq \label{densite-s3-densite-charge-reduit}
443: n_s(x)=n_0 + \int_{a_1}^{a_2} da \, N_0
444: e^{- \xi (A_y-a)^2 +\delta (A_y-a)}
445: \eeq
446: Let us set 
447: \beq
448: n_a(x)=N_0(a) e^{- \xi(a) (A_y(x)-a)^2 +\delta(a) (A_y(x)-a)},
449: \eeq
450: then 
451: \beq \label{densite-s3-densite-charge-resume}
452: n_s(x)=n_0 + \int_{a_1}^{a_2} da \, n_a(x)
453: \eeq
454: If the following equations are satisfied for each value of $a$, 
455: \begin{eqnarray}
456: \xi(a)_{ions} = \xi(a)_{electrons} \\
457: \delta(a)_{ions} = \delta(a)_{electrons} \\
458: N_0(a)_{ions}= N_0(a)_{electrons},
459: \end{eqnarray}
460: then, the charge neutrality is satisfied.
461: It is easy to show that 
462: one can freely choose the functions $\alpha_{\perp}, \, \eta, \nu, \, n_g$ 
463: that correspond to a particle species, then compute the $\xi$, $\delta$ and $N_0$ functions
464: and find the $\eta, \, \nu, \, n_g$ functions of the other species
465: that are associated with the same $\xi$, $\delta$ and $N_0$. 
466: There is only one restriction: $\eta$ must be positive.
467: If we choose electron parameters, the condition 
468: for a positive ion $\eta_i$ parameter is, for each value of $a$
469: \beq \label{densite-g-nonlineaire-extreme-eta-0}
470: \frac{m_e}{\eta_e(a)} > 2 [(\frac{m_i}{m_e})T_{\perp i}(a) +T_{\perp e}(a)]
471: \eeq
472: which requires small enough values of $\eta_e(a)$. 
473: Noticing that $\eta$ homogeneous to the reciprocal of a squared velocity $\eta=v_{\eta}^2$, 
474: the above relation writes 
475: \beq \label{densite-g-nonlineaire-extreme-v-eta-0}
476: v_{\eta}^2 > (\frac{m_i}{m_e} \frac{T_i}{T_e}+1) v_{te}^2
477: \eeq
478: Let us call $\eta_c(a)= v_c(a)^{-2}$ the value that corresponds to an equality in 
479: Eq. (\ref{densite-g-nonlineaire-extreme-eta-0}) and (\ref{densite-g-nonlineaire-extreme-v-eta-0}).
480: The case of $\eta_e=\eta_c$ corresponds to an infinite value of $\eta_i$. 
481: We shall see in the applications that because of the condition $\eta < \eta_c$, $v_{\eta}$ can reach very high values, 
482: and it is necessary to remember that $v_{\eta}$ is not an actual velocity, 
483: it does not correspond to a propagation phenomenon. 
484: %In the particular case of $\eta_e=\eta_c$, $\nu=0$,  $T_{iz}(a)=$constant, and 
485: %$T_{i \perp}(a)=$constant,
486: %the $g$ function is null for the ions and the ion velocity distribution is Maxwellian.
487: 
488: Let us define  
489: \beq \label{densite-s3-k(a)}
490: k(a)=  -\mu_0 N_0(a) 
491: (\frac{m_e}{\alpha_{\perp e}(a)}+\frac{m_i}{\alpha_{\perp i}(a)})
492: = -2 \mu_0 N_0(a)(T_{\perp e}(a)+T_{\perp i}(a)).
493: \eeq
494: Negative values of $k(a)$ correspond to a positive value of $n(a)$, and therefore
495: imply an increase of the plasma density. On the contrary, positive values
496: of $k(a)$ induce a reduction of the density.
497: The total current density is the sum of the ion and the electron currents,
498: \beq \label{densite-s3-densite-courant-reduit}
499: J_y(x)=-\frac{1}{\mu_0} \int_{a_1}^{a_2} da \; \frac{k}{2} 
500: [-2 \xi (A_y -a) + \delta] e^{-\xi (A_y(x)-a)^2 +\delta (A_y(x)-a)}.
501: \eeq
502: Let us notice that Eq. (\ref{densite-s3-densite-charge-reduit}),
503: (\ref{densite-s3-k(a)}), and (\ref{densite-s3-densite-courant-reduit}) imply 
504: \beq
505: J_y= \int_{a_1}^{a_2} da \; T_{\perp}(a) \frac{d n_a}{d A_y}.
506: \eeq
507: The above relation is a generalization of Eq. (\ref{densite-g-Jy-dn-dAy}) to non isothermal equilibria.\\
508: The Amp\`ere equation becomes
509: \beq \label{densite-s3-equation-courant-ordre2}
510: \frac{d^2 A_y(x)}{dx^2}=
511: \int_{a_1}^{a_2} da \; \frac{k(a)}{2} 
512: [-2 \xi (A_y(x) -a) + \delta] e^{-\xi (A_y(x)-a)^2 +\delta (A_y(x)-a)}.
513: \eeq
514: We can deduce a first order equation 
515: \beq \label{densite-s3-equation-courant-ordre1}
516: B_z(x)^2= (\frac{d A_y}{dx})^2 =C +  
517: \int_{a_1}^{a_2} da \,
518: k(a) e^{-\xi(a) (A_y-a)^2+\delta (A_y(x)-a)}.
519: \eeq
520: The vector potential is the solution of
521: \beq \label{densite-s3-solution-x-Ay}
522: x = s \int_{A_y(0)}^{A_y(x)} 
523: \frac{dA}{\sqrt{C+  \int_{a_1}^{a_2} da \,
524: k(a) \exp{[-\xi(a) (A-a)^2+\delta(a) (A-a)]}}}.
525: \eeq
526: The sign of the magnetic field is given by $s=\pm 1$, and 
527: $C$ is an integration constant. The parameters 
528: $k$, $\xi$, and $\delta$ are arbitrary functions of $a$.
529: The solution has a physical meaning if it is defined for any value of $x$:
530: the integral in Eq. (\ref{densite-s3-solution-x-Ay}) 
531: considered as a function of $A_y$
532: must be able to vary from $-\infty$ to $+\infty$.
533: 
534: Equation (\ref{densite-s3-equation-courant-ordre1}) 
535: can be modified in order to provide
536: an integro-differential equation in $B_z(x)$. Let $y$ defined by 
537: $A_y(y)=a$ be the new integration variable.
538: Then, 
539: \beq \label{densite-s3-magnetique-ordre1}
540: B_z(x)^2 = C + \int_{y_1}^{y_2} dy B_z(y) k(y) 
541: e^{ 
542: -\xi(y) [\int_{y}^{x}B_z(u) du]^2+\delta(y) [\int_{y}^{x}B_z(u) du] 
543:  }.
544: \eeq
545: The same process applied on Eq. (\ref{densite-s3-densite-charge-reduit})
546: provides
547: \beq \label{densite-s3-densite-ordre1}
548: n(x) = n_0 + \int_{y_1}^{y_2} dy B_z(y) N_0(y) 
549: e^{ 
550: -\xi(y) [\int_{y}^{x}B_z(u) du]^2+\delta(y) [\int_{y}^{x}B_z(u) du] 
551:  }.
552: \eeq
553: These equations will be used in section III.B to analyse the asymptotic behaviour of the solutions.
554: 
555: \section*{III. EXAMPLES AND PROPERTIES OF TANGENTIAL EQUILIBRIA} \label{section-properties}
556: 
557: 
558: \subsection*{A. Examples of elementary solutions with parameters relevant to space plasmas}
559: The trivial case of a uniform plasma can be easily recovered with several sets of parameters 
560: (including $n_g \ne 0$ and $\xi \ne 0$).
561: The case of the Harris current sheet \cite{Harris 1962}, 
562: reviewed in details in \cite{Roth et al 1996}, can be recovered with
563: $\xi=0$, $\nu \ne 0$ and $n(a)=n_g \delta_0$, where $\delta_0$ is the Dirac distribution.
564: 
565: Here are a few examples of equilibria where  $\nu=0$, $n_g=n_c \delta_0$, 
566: and where the temperature functions are constant.
567: They correspond to the case $g(p_y) = n_c \exp{(- \eta (p_y/m)^2)}$ already mentionned
568: by Channell \cite{Channell 1976}. The figure given in Channell's paper
569: to illustrate this case 
570: displays the equilibrium of an evanescent plasma.
571: We briefly show in this section 
572: a few examples of equilibria of finite size current layers
573: in non evanscent plasmas,
574: in order to emphasize a few interesting properties not discussed in Channell's paper.
575: 
576: 
577: Figure 1 shows an example of a structure that is computed with parameters
578: typical of high altitude auroral plasmas : asymptotic magnetic field
579: $\sqrt{C}=3000$ nT, $n_0=10 \mbox{cm}^{-3}$, $T_i=T_e=10$ eV, 
580: and $\eta_e=1.53 \, 10^{-16} (m/s)^{-2}=0.99 \eta_c$, very close to the 
581: critical value imposed by
582: Eq. (\ref{densite-g-nonlineaire-extreme-eta-0}).
583: A high negative value $n_c=-8 \mbox{cm}^{-3}$ allows for a deep density depletion.
584: The abcsissa is normalized to the ion Larmor radius which is
585: $\rho_L = 152$ m. We can see on Fig. 1 that the size of the structure is about 
586: $4 \rho_L \sim 600$ m. Such a narrow and deep structure 
587: (the density in the cavity is only 20\% of the density of the surrounding plasma)
588: can be used in a first approximation for the density depletions encountered in the
589: high altitude auroral zone. 
590: 
591: The solution displayed in Fig. 2 corresponds to the limiting case for $\eta$, obtained for  
592: the weakly magnetized interplanetary plasma encountered in the solar wind at one 
593: astronomical unit : 
594: $\sqrt{C}=5$ nT, $n_0=10 \mbox{cm}^{-3}$, $T_i=T_e=100$ eV, 
595: and $\eta_e=1.53 \, 10^{-17} (m/s)^{-2}=0.99 \eta_c$. The ion Larmor radius 
596: is $280$ km. We can notice that for the same density 
597: depletion, the variation of the magnetic field is much higher than in the case 
598: of a higly magnetized plasma. We also notice that the total size of the structure
599: hardly exceed one ion Larmor radius.
600: 
601: 
602: As the size is of the order of the ion Larmor radius,
603: such solutions cannot be described through the fluid theories.
604: Of course, we can also build very large density structures, if we choose a small
605: value of ${\eta}$ in order to get a small value of $\xi$. Such large 
606: structures can more simply be described by fluid models.
607: Figure 3 shows the solution with $\eta =10^{-20} (m/s)^{-2}\sim 10^{-3} \eta_c $. 
608: The structure is much larger 
609: but has the same amplitude. Comparing Fig. 2 and Fig. 3, 
610: we can notice that, contrary to the case of solitons, 
611: there is no correlation between the size and the amplitude of the structures.
612: 
613: 
614: 
615: 
616: \subsection*{B. Asymptotic behaviour of the solutions} 
617: 
618: We can caracterize two kinds of non trivial elementary solutions : 
619: those with $\eta \ne 0$, and those where $\eta = 0$ (and therefore $\nu \ne 0$, otherwise the solution is trivial).
620: Those of the first kind correspond to localized solutions, they are centered around an abscissa $x$ such as
621: $A_y(x) \sim a$. Therefore, when the distribution is the 
622: infinite sum of elementary functions indexed by $a$ as in Eq. (\ref{densite-s3-densite-charge-resume}), 
623: the parameter $a$ indicates where the 
624: density $n_a(x)$ is non negligible. 
625: Let us consider an equilibria where, for $x \sim \infty$, the magnetic field has a finite value. 
626: Then, the potential vector goes to infinity. Taking $a_1 \sim -\infty$ and $a_2 \sim +\infty$ means that 
627: the plasma distribution will be influenced by some of those elementary 
628: functions even when $x \sim \infty$. On the contrary, equilibria defined with finite values 
629: of $a_1$ and $a_2$ (and $B(\infty) \ne 0$) means that 
630: for $x \sim \infty$, the equilibria will converge toward the trivial solution of a uniform plasma with the density $n_0$ 
631: that appears in Eq. (\ref{distribution-general}).
632: 
633: The case of a null magnetic field at $x \sim \infty$ (less useful for space plasmas applications) 
634: is different. Let us consider the case of  $B(+\infty)=0$. For $x \sim +\infty$, 
635: the vector potential has a constant value. 
636: Let us call $A_{\infty}$ this value. Then, as $A$ cannot exceed $A_{\infty}$, 
637: the main integral in Eq. (\ref{densite-s3-solution-x-Ay})
638: is divergent for $A_y(x)=A_{\infty}$ (otherwise the equilibria would not be defined for any value of $x$). 
639: As the elementary solutions parametrized 
640: by $a$ are non negligible only for abscissas $x$ such that $A_y(x)$ is close to $a$, it is not 
641: necessary to consider values or $a_2$ that exceed notably $A_{\infty}$ : they would have no influence on 
642: the value of the integral in Eq. (\ref{densite-s3-solution-x-Ay}). In brief, for equilibria with
643: $B(\infty)=0$, the range of values where the parameter $a$ is significant is bounded, 
644: $a_1$ and $a_2$ do not have to be 
645: infinite. 
646: 
647: 
648: In the Harris model, the magnetic field reverses and has two opposite finite values for 
649: $x \sim - \infty$ and $x \sim + \infty$. We shall see
650: that there exist non elementary solutions where the magnetic field has finite values
651: on both sides, and that these values have not necessarily the same absolute value.
652: 
653: We consider the case $a_1=-\infty$ and $a_2=+\infty$.
654: Let $u^-=\lim_{x \rightarrow -\infty} u$ be the limit of 
655: any physical value $u$ when $x$ tends to $-\infty$,
656: and $u^+=\lim_{x \rightarrow +\infty} u$.
657: The integral in Eq. (\ref{densite-s3-magnetique-ordre1}), where $y_1=-\infty$
658: and $y_2=+\infty$ can be cut in three parts : an integral $I1$ from
659: $y_1$ to $x-\Lambda$, an integral $I2$ from $x-\Lambda$ to $x+\Lambda$,
660: and an integral $I3$ from $x+\Lambda$ to $y_2$.
661: We will consider large values of $x$ (positive or negative), and large values
662: of $\Lambda$.
663: 
664: For $\Lambda$ large enough,  $y$ has large negative values, 
665: \begin{eqnarray}
666: I1 
667: &\sim& 
668: \int_{y_1}^{x-\Lambda} dy B_z^- k^- 
669: \exp{(-\xi^- (B_z^-)^2 (x-y)^2-\delta^- B_z^-(x-y)  )} \\
670: &\sim& 
671: -\frac{k^-}{\delta^-} [\exp{\delta^- B^- \Lambda}-\exp{\delta^- B^- y_1}]
672: \end{eqnarray}
673: %&\sim& 
674: %\int_{y_1}^{x-\Lambda} dy B_z^- k^- 
675: %\exp{(-\delta^- B_z^-(x-y)  )} \\
676: This function tends to zero for $y_1 \rightarrow -\infty$ if $\delta^- B^-<0$,
677: and diverges if $\delta^- B^->0$. In the case $\delta^- =0$, 
678: \beq
679: I1 \sim
680: \int_{y_1}^{x-\Lambda} dy B_z^- k^- 
681: \exp{(-\xi^- (B_z^-)^2 (x-y)^2 )} 
682: \eeq
683: tends toward zero.
684: Similarily, $\lim_{y_1 \rightarrow -\infty} I3 = 0$ if $\delta^+ B^+ \ge 0$.
685: With the same kind of technique, we can show that, for large negative values of $x$,
686: \begin{eqnarray}
687: I2 
688: &\sim& 
689: k^- \sqrt{\frac{\pi}{\xi^-}} \sigma_{B_z^-} \exp{\frac{(\delta^-)^2}{4\xi^-}},
690: \end{eqnarray}
691: where $\sigma_{B_z^-}$ is the sign of $B_z^-$.
692: Provided that $\delta^+ B^+ \ge 0$, and $\delta^- B^- \le 0$,
693: the asymptotic expansion of Eq. (\ref{densite-s3-magnetique-ordre1}) shows that
694: \beq
695: (B_z^-)^2 = 
696: C +
697: k^- \sqrt{\frac{\pi}{\xi^-}} \sigma_{(B_z^-)} \exp{\frac{(\delta^-)^2}{4\xi^-}}.
698: \eeq
699: Computing an asymptotic value of $I2$ for $x \rightarrow +\infty$
700: bring a similar  relation for $(B_z^+)$.
701: The same method used with Eq. (\ref{densite-s3-densite-ordre1})  provides
702: the asymptotic value
703: of the  contribution of each species to the particle density:
704: \beq
705: n^{+}=  n_0 + N_0^{+} \sigma_{(B_z^+)} 
706:           \sqrt{\frac{\pi}{\xi^+}} \exp{\frac{(\delta^+)^2}{4\xi^+}}.
707: \eeq
708: The equation for $n^{-}$ is analogous.
709: As $d_x n (\infty)=0$, we deduce
710: $J_y^+=J_y^-=0$.
711: %
712: There is no current density carried by the plasma for $x \sim \infty$.
713: This result can also be found with an asymptotic expansion 
714: of Eq. (\ref{densite-s3-densite-courant-reduit}).
715: Moreover, Eq. (\ref{densite-s3-densite-courant-reduit}) shows that the 
716: ion and  electron contribution to
717: the current density have the same sign, their mean velocities are therefore 
718: opposite. As there is no current for $x \sim \infty$, for each species 
719: $v^+=v^-=0$. The plasma does not flow perpendicularly to the magnetic field 
720: at $x \sim \infty$.
721: 
722: 
723: 
724: 
725: \subsection*{C. Examples of solutions with and without field reversal}
726: 
727: 
728: Figure 4 shows an example of equilibrium where the magnetic field amplitude 
729: and the density are not the same for 
730: $x \sim - \infty$ and $x \sim + \infty$. We have set 
731: constant values $\eta=1.53 \;  10^{-17}(m/s)^{-2}$ close to $\eta_c$, $\nu=0$, and a function 
732: $n_g(a)=640 \mbox{atanh} (1000 a)$. The temperature 
733: is 10 eV. With $C=(5nT)^2$, the asymptotic magnetic field amplitude 
734: is 5nT. The ion Larmor radius is $\rho_i \sim 0.2 \, 10^{6}$m.
735: The variable $a$ is homogeneous to $B_z$ times $x$, therefore $1000 a \sim x/\rho_i$ 
736: is of the order
737: of magnitude of the absissa $x$ divided by the ion Larmor radius.
738: The factor $640$ that multiplies the inverse hyberbolic tangent 
739: function was set heuristically in order to
740: get a variation of density of the order of 60\%.
741: The relatively low value of $\sqrt{C}=5$ nT 
742: allows for high relative variations of the magnetic field amplitude. 
743: The asymptotic values of the magnetic field and of the density 
744: fit the analytical formulas given in section III.B.
745: 
746: 
747: Figure 5  shows an example where $\eta = 0$. 
748: We have chosen  a function 
749: $n_g(a)=4200 e^{-(1000 a)^2}$. The temperature 
750: is 10 eV, and $\sqrt{C}=5$ nT.
751: The value $\nu=10^{-10}$ was chosen, because, in the exponential, the 
752: factor $\nu a \sim \nu B_z x$ is again of the order of $x/\rho_i$.
753: We can see on Fig. 5 that the sign of the magnetic field changes
754: for $x \sim - \infty$ and $x \sim + \infty$. The reversal occur 
755: for a value of $A_y$ that cancels the square root in Eq. (\ref{densite-s3-solution-x-Ay}).
756: Although the magnetic field is reversed on both sides, 
757: this is not a Harris equilibrium. 
758: 
759: \subsection*{D. Example of a non isothermal plasma cavity in the 
760: Earth auroral zone} 
761: 
762: Are the cavities built in the Earth auroral zone in an equilibrium state of the plasma ?
763: 
764: Hilgers {\em et al.} \cite{Higers et al 1992a} made a careful analysis of the Langmuir probe
765: current measurements onboard the swedish Viking satellite in low plasma densities.
766: They showed a case, taken at an altitude of 7000 km, of a deep  auroral cavity.
767: The density reach less than one particle
768: per cubic centimeter, that is less than 10 \% of the surrounding plasma
769: density. The electron plasma temperature in the cavity (although not precisely measured) 
770: is of the order of 1 keV, 
771: compared to 1 eV outside the cavity. The boundaries of the cavity are sharp, 
772: Hilgers {\em et al}. measured 1.4 km, of the order of a few ion Larmor radii. 
773: The magnetic field amplitude is  $B_{z0}=6800$ nT. 
774: 
775: Apart from the temperature gradient, all these features have been qualitatively reproduced 
776: in the elementary equilibria shown in Fig. 1. 
777: In order to reproduce a cavity with sharp temperature gradients, 
778: an elementary solution cannot be used. 
779: Let $\rho_i$ be the Larmor radius outside the cavity, it corresponds to the cold palsma (1 eV), 
780: its value is 21 m. Inside the cavity, the plasma temperature is 1 keV, and the ion Larmor radius is
781: $\rho_{i,hot}=660$m$=32 \rho_i$. 
782: We choose an electron temperature function 
783: $T_e(a)=1+1000 \exp{(-(a/B_{z0} 32 \rho_i)^2)}$ that varies 
784: from 1 to 1000 eV on a scale $\Delta x \sim 32 \rho_i \sim \rho_{i,hot}$.
785: The density is controled through the constant scalar $n_0=10$ cm$^{-3}$, 
786: and the $n_g$ function, 
787: $n_g(a)=-n_{g0} \exp{(-(a/B_{z0} 32 \rho_i)^2)}$.
788: The scalar $n_{g0}$ is set (heuristically) to 50000 in order to have 
789: one particle per cubic centimeter inside the cavity. The $\nu$ function is null.
790: The $\eta$ function is chosen is order to get a sharp cavity, 
791: that is with a value close to the $\eta_c$ function:  
792: $\eta(a)=0,98 \eta_c (a)$ have values of the order of $10^{-15} (m/s)^{-2}$. Actually, the 
793: equilibria is not very sensitive on the value of $\eta$. Taking a lower value like $\eta(a)=0.7 \eta_c (a)$ 
794: (and a slightly lower value of $n_{g0}$) bring similar results. 
795: 
796: The result, displayed on Fig. 6, is in quantitative agreement with the prescriptions given 
797: by the observations. The size of the gradient is $70 \rho_i =1400$m, as measured by Hilgers {\em et al}.
798: 
799: We can notice on figure 6 that the magnetic field amplitude variation $\Delta B$ in the cavity is about 0.5 nT. 
800: A similar equilibria (not displayed)
801: set with a lower value of the ambient magnetic field ($B_{z0}=300$ nT) bring a larger variation: $\Delta B=5$ nT. 
802: 
803: We conclude from this short study that deep plasma cavities can be equilibrium structures of the 
804: Earth auroral plasma. Therefore, they won't be destroyed immediately after the
805: extinction of their cause. It is however interesting to know 
806: if this equilibrium is stable. The theoretical treatement 
807: of this question goes beyhond the scope of this paper. 
808: I will only give a few hints
809: in favour of the stability of the auroral cavities. 
810: 
811: The Auroral zone of the Earth is a radio source. 
812: The power of the emissions, called Auroral Kilometric Radiation (AKR),
813: can reach 10 MW. 
814: The Viking spacecraft has gone through the sources of AKR: 
815: the waves are emitted inside the cavities, where they take their
816: free energy from the hot rarefied plasma \cite{Roux et al 1993}.
817: The waves are strongly refracted at the edges of the AKR source,
818: and it is probable that a part of the AKR is guided inside the cavities
819: \cite{deFeraudy 1987}.
820: Auroral kilometric radiations can be observed on time scales of 10 minutes.
821: The cavities that contain AKR sources are expected to last at least for the same duration.
822: Therefore, two situations are possible : 
823: (1) the cavities are the consequence of a phenomena that lasts 
824: for tens of minutes, or 
825: (2) they are generated by more transient phenomena and they are stable.
826: 
827: An another hint about the stability of hot plasma cavities 
828: comes from numerical simulations carried by Genot {\em et al}. 
829: \cite{Genot et al 2001}. 
830: The simulations start with an isothermal (a few eV) auroral cavity. The cavity is stable. 
831: An Alfv\'en wave is added. The interaction of the Alfv\'en wave with the cavity
832: triggers strong electron acceleration and turbulence. When the free energy is completely 
833: removed from the Alfv\'en wave by the accelerated electrons, the plasma inside
834: the cavity is hot, and the surrounding plasma is cold.
835: We are in the situation shown in Fig 6. The simulations show that the heated cavity,
836: which is still very deep and sharp, is stable.
837: 
838: Most of the models about the generation of auroral cavities invoke strong kinetic Alfv\'en 
839: Waves (SKAW) \cite{Wu et al 1997} \cite{Shukla et al 1999}, observed onboard the Freja 
840: and FAST satellites  \cite{Louarn et al 1994},
841: \cite{Chaston et al 1999}. 
842: The associated magnetic field perturbation $\Delta B$ 
843: is of the order of 50 nT. But the observations made onboard 
844: FAST show that the SKAW's are mainly observed in the cusp and the polar cap boundary layer,
845: while the cavities are observed in the lower latitude auroral region where 
846: the magnetic fluctuations
847: $\Delta B$ do not exceed 5 nT. It is possible that the cavities are built in the 
848: polar cap boundary layer by SKAWs, they subsist after the
849: disapearance of the Alfv\'enic perturbations and finally they 
850: are gently convected to lower latitudes
851: by the large scale convection electric field. This suppose of course the stability of the
852: auroral cavities.  An other scenario is that the cavities are created 
853: through Field Line Resonance (FLR) \cite{Lotko et al 1998} 
854: directly in the auroral region. The 
855: $\Delta B \sim $ 5 nT  (on a time scale that corresponds to the crossing
856: of a cavity by a spacecraft) associated to the FLR is compatible with 
857: the value shown in Fig. 6. The equilibria shown in Fig. 6 may be 
858: a relevant short scale Vlasov-Maxwell description
859: of the part of the FLR that is in the high altitude auroral zone 
860: (the FLR has been modeled on a global scale
861: in the multifluid approximation up to now).
862: 
863: \section*{IV. CONCLUSION AND FURTHER DEVLOPPMENTS}
864: 
865: This paper presents a large class of monodimensional analytic kinetic equilibria.
866: They are based on particle distribution functions Eq. (\ref{distribution-general}) 
867: that depend on a set of almost arbitrary functions. These distribution
868: functions are solutions of the Vlasov equation, 
869: and the vector potential can be computed through the evaluation of the integral
870: function given in Eq. (\ref{densite-s3-solution-x-Ay}).
871: In some particular cases, like the uniform plasma and the Harris current sheet,
872: this integral is a combination of elementary mathematical functions.
873: In the other cases, its numerical evaluation is straightforward.
874: 
875: Unlike most of the kinetic 
876: tangential equilibria given in the litterature,
877: the freedom in the choice of the density profile, altough more difficult to control,
878: is almost as large as with bi-fluid models. The kinetic 
879: equilibria have the great advantage of giving a complete description of the distribution
880: function (not provided with the fluid theory), 
881: and can describe equilibria where strong gradients
882: develop on the scale of a few ion Larmor radii.
883: These solutions can be used as initial conditions in particle in cell
884: and Vlasov numerical simulations.
885: 
886: 
887: Most of the authors who build models and are cited in this paper   
888: have based their elementary solutions
889: on distribution functions that depend on the invariants of the motion. 
890: Some of them combined a small number of such elementary
891: solutions and asserted that each 
892: of these elementary solutions
893: correspond to a family of trapped particles. The distribution function 
894: given in Eq. (\ref{distribution-general}) is the superposition, 
895: not of a finite number, but of
896: a continuum of families of trapped particles. 
897: This is why we have a very large degree of freedom
898: in the choice of the density and temperature profiles.
899: For the first time, we show that such a supperposition of families 
900: of trapped particles lead to an analytically integrable equation, 
901: whose solution is given in Eq. (\ref{densite-s3-solution-x-Ay})
902: 
903: 
904: The equilibria discussed in section III.B where the plasma
905: is uniform for $x \sim \infty$ can be compared to the solutions of 
906: the jump equations 
907: developed in the frame of the MHD theory.
908: The equilibria developed in the present paper are characterized 
909: by jumps of the density and of the
910: magnetic field, the normal velocity is null as well as the normal magnetic field.
911: They therefore belong to the family of tangential discontinuities.
912: But they concern only a sub-category: 
913: the magnetic field direction is uniform and
914: the solutions have no velocity shear: the velocities
915: are equal to zero at $x \sim -\infty$ and $x \sim +\infty$, 
916: even if they can take other values
917: at finite distances.
918: These restrictions are not required for general tangential discontinuities.
919: 
920: Building a kinetic model, inspired from the present results,
921: of a tangential discontinuity where the magnetic field can turn
922: is straighforward and will be presented in a forthcomming paper in order to 
923: analyse experimental data provided by the Cluster satellites.
924: 
925: A kinetic model of a tangential equilibria whith a velocity shift 
926: ($x \sim -\infty$ and $x \sim +\infty$) is not compatible with exact 
927: charge neutrality. A quasi neutral model would do. It could be computed, as 
928: a first order perturbation in $n_i-n_e$ added to 
929: the charge neutral equilibria presented
930: in this paper.
931: 
932: The examples given in the present paper come from the magnetospheric and solar wind 
933: physics because
934: these media offers the opportunity of in situ observations of 
935: non collisional astrophysical plasmas.
936: Tangential equilibria can be used to describe  the magnetopause 
937: (present around all the magnetized planets with an atmosphere),
938: the distant neutral sheet, the Earth auroral cavities, and some 
939: density fluctuations of the 
940: solar wind \cite{Safrankova et al 2000}. 
941: 
942: But tangential equilibria do not exist only in the terrestrial environnment.
943: They exist whenever plasmas with different origins and velocities meet. 
944: Such situations exist in many astrophysical plasmas.
945: Theoretical models (based on MHD and plasma multifluid theories) of the
946: boundary of the Heliosphere show that it is constituted of a termination shock, 
947: followed by the heliopause, that is a tangential discontinuity, an hydrogen wall,
948: and possibly an heliospheric shock \cite{Zank 1999}. 
949: Tangential discontinuities also exist around other stars. 
950: M\"uller {\em et al.} \cite{Muller et al 2001} studied the interaction of 
951: the very active 
952: binary star $\lambda$ Andromedae, and of the neaby star $\epsilon$ Indi 
953: with the interstellar medium.
954: They found, as in the case of the heliopause, the existence of four boundaries, 
955: one of them,
956: the asteropause being a tangential discontinuity.
957: It is clear that tangential layers exist in other astrophysical plasmas where 
958: they may play a very important role, as frontiers, or in 
959: acceleration, heating and radiative processes. 
960: 
961: The studies of the magnetospheric plasmas
962: have shown that most of the heating and acceleration phenomena 
963: can be only explained through 
964: kinetic processes.
965: So far, most of the remote astrophysical plasmas have been studied through 
966: MHD or multifluid theories.  The understanding of acceleration processes in remote 
967: astrophysical plasma might as well require kinetic models. The equilibria 
968: presented in this paper might be a good start for some of those studies.
969: 
970: 
971: % ******************** ACKNOWLEGEMENTS AND APPENDICES *******************
972: %
973: \newpage
974: 
975: 
976: 
977: \section*{ACKNOWLEDGMENTS}
978: The author gratefully acknowledges stimulating and usefull discussions with 
979: G\'erard Belmont and Alain Roux. 
980: 
981: 
982: % ******************** BIBLIOGRAPHIE *************************
983: %
984: \newpage
985: 
986: \begin{thebibliography}{}
987: 
988: 
989: \bibitem{Higers et al 1992a}
990: Hilgers, A.,B. Holback, G. Holmgren, and R. Bostr\"om,
991: %Probe measurements of low plasma densities with applications to 
992: %the auroral acceleration region and auroral kilometric radiation sources.,
993: { J. Geophys. Res. 97}, A6, 8631, (1992).
994: 
995: 
996: \bibitem{Strangeway et al 1998}
997: Strangeway R.J., L. Kepko, R.C. Elphic, C.W. Carlson, R.E. Ergun,
998: J.P. McFadden, W.J. Peria, G.T.Delory, C.C.Chaston, M. Temerin, 
999: C.A. Cattell, E. M\"obius,L.M. Kistler, D.M. Klumpar, W.K. Peterson,
1000: E.G. Shelley, and R. Pfaff,
1001: %FAST observations of VLF waves in the auroral zone : 
1002: %evidence of very low plasma densities,
1003: { Geosphys. Res. Let., 25}, 12, 2065, (1998).
1004: 
1005: 
1006: \bibitem{McFadden et al 1998}
1007: J.P. McFadden, C.W. Carson, R.E. Ergun, C.C. Chaston, F.S. Mozer, M. Temerin,
1008: D. M. Klumpar, E.G. Shelley, W.K. Peterson, E. Moebius, L. Kistler, 
1009: R. Elphic, R. Strangeway, C. Cattell, R. Pfaff
1010: %Electron modulation and ion cyclotron waves observed by FAST,
1011: { Geophys. Res. Lett., 25}, 12, 2045, (1998).
1012: 
1013: 
1014: \bibitem{Higers 1992b}
1015: Hilgers, A.,
1016: %The auroral radiating plasma cavities,
1017: { Geophys. Res. Lett., 19} 3, 237-240, (1992).
1018: 
1019: 
1020: \bibitem{Genot et al 1999}
1021: G\'enot, V., P. Louarn, and D. Le Qu\'eau, A study of the propagation of
1022: Alfv\'en
1023: %waves in the auroral density cavities, 
1024: { J. Geophys. Res. 104}  22,649, (1999).
1025: 
1026: \bibitem{Genot et al 2000} 
1027: V. G\'enot, P. Louarn, and F. Mottez, 
1028: %Electron acceleration by Alfv\'en waves in density cavities, 
1029: { JJ. Geophys. Res. 105}, A12, 27611,  (2000).
1030: 
1031: 
1032: \bibitem{Roux et al 1993}
1033: A. Roux, A. Hilgers, H de Feraudy, D. Le Queau, P. Louarn, S. Perraut, 
1034: A. Bahnsen, M. Jesperen, E. Ungstrup, M. Andre,
1035: %Auroral kilometric radiation sources: In situ and remote observations from Viking,
1036: { J. Geophys. Res., 98}, 11657, (1993).
1037: 
1038: \bibitem{Pottelette et al 2001} 
1039: R. Pottelette, R.A. Treumann, M. Berthomier,
1040: %Auroral plasma turulence and the cause of auroral kilometric 
1041: %radiation fine structure,
1042: { J. Geophys. Res., 106}, 8465, (2001).
1043: 
1044: \bibitem{Genot et al 2001} 
1045: V. G\'enot, P. Louarn, and F. Mottez, 
1046: %Fast evolving spatial structure of auroral parallel electric field, 
1047: { J. Geophys. Res. 106}, A12, 29633, (2001).
1048: 
1049: \bibitem{Harris 1962}
1050: E.G. Harris, 
1051: %On a plasma sheath separating regions of oppositely directed magnetic field,
1052: {Il Nuovo Cimento 23} 1, 115, (1962).
1053: 
1054: \bibitem{Channell 1976}
1055: P. J. Channell, 
1056: %Exact Vlasov-Maxwell equilibria with sheared magnetic field,
1057: {Phys. Fluids 19} 10, 1541, (1976).
1058: 
1059: \bibitem{Roth et al 1996}
1060: M. Roth, J. De Keyser, M.M. Kuznetova,
1061: %Vlasov theory of the equilibrium structure of tangential 
1062: %discontinuities in space plasmas,
1063: { Space Sci. Rev., 76}, 251, (1996).
1064: 
1065: \bibitem{Krall Rosenbluth 1963}
1066: N.A. Krall and M.N. Rosenbluth, 
1067: %Low-frequency stability of nonuniform plasmas,
1068: {Phys. Fluids 19} 2, 254, (1963).
1069: 
1070: \bibitem{deFeraudy 1987}
1071: de Feraudy,H., B.M. Pedersen, A. Bahnsen, and M. Jespersen,
1072: %Viking observations of auroral kilometric radiation from the plasmasphere
1073: %to the night auroral oval source regions,
1074: { Geophys. Res. Lett., 14} 511, 1987.
1075: 
1076: \bibitem{Wu et al 1997}
1077: Wu,D.J., D.Y. Wang, and G.L. Huang,
1078: %Two-dimensional solitary kinetic Alfv\'en waves
1079: %and dipole vortex structures,
1080: { Phys. Plasmas, 4} (3), 611, (1997).
1081: 
1082: \bibitem{Shukla et al 1999}
1083: Shukla, P.K., L. Stenflo, R. Bingham,
1084: %Nonlinear propagation of inertial Alfv\'en waves in auroral plasmas,
1085: { Phys. of Fluids, 6}, 5, 1677, (1999).
1086: 
1087: \bibitem{Louarn et al 1994}
1088: Louarn, P., J.E. Wahlund,T. Chust, H de Feraudy and A. Roux,
1089: B. Holback, P.O. Dovner, A.I. Eriksson and G. Holmgren,
1090: %Observation of kinetic Alfv\'en waves by the FREJA spacecraft,
1091: { Geophys. Res. Lett., 21}, 17, 1847, (1994).
1092: 
1093: \bibitem{Chaston et al 1999}
1094: Chaston, C.C., C.W. Carlson, W.J. Peria, R.E. Ergun and J.P. Mc Fadden,
1095: %FAST observations of inertial Alfv\'en waves in the dayside aurora,
1096: { Geophys. Res. Lett. 26},6, 647, (1999).
1097: 
1098: \bibitem{Lotko et al 1998}
1099: Lotko,W., A.V. Streltsov, C.W. Carlson,
1100: %Discrete auroral arc, electrostatic shock and suprathermal electrons
1101: %powered by dispersive, anomanously resistive field line resonance,
1102: { Geophys. Res. Lett., 25}, 24, 4449, (1998).
1103: 
1104: \bibitem{Safrankova et al 2000}
1105: Safrankova, J., Prech, L., Nemecek, Z., Sibeck, D.G. Mukai, T,
1106: %Magnetosheath response to the interplanetary magnetic field tangential discontinuity,
1107: { J. Geophys. Res., 105}, 11, 25113, (2000).
1108: 
1109: \bibitem{Zank 1999}
1110: Zank, G.P.,
1111: %Interaction of the solar wind with the local interstellar medium: 
1112: %A theoretical perspective,
1113: { Space Sci. Rev., 89}, 413, (1999).
1114: 
1115: 
1116: \bibitem{Muller et al 2001}
1117: M\"{u}ller, H.R., G.P. Zank, and B.E. Wood,
1118: %Modeling the ISM-stellar wind interactions of lambda Andromedae and epsilon Indi,
1119: { Astrophys. J.551}, 495, (2001).
1120: 
1121: 
1122: 
1123: \end{thebibliography}
1124: 
1125: % **************** LES FIGURES ***************
1126: \newpage
1127: 
1128: 
1129: \begin{figure}[htp]
1130: \centering
1131: \vspace{1cm}
1132: \psfig{file=figureA1.ps,width=12cm}
1133: \caption{ FIG. 1. A deep and narrow density depletion in a highly magnetized plasma.
1134: The parameters of this equilibrium are given in Sec. III.A.}
1135: \end{figure}
1136: 
1137: \begin{figure}[htp]
1138: \centering
1139: \vspace{1cm}
1140: \psfig{file=figureA2.ps,width=12cm}
1141: \caption{FIG. 2. A deep and narrow density depletion plasma with a weak magnetic field.
1142: The parameters are the same as for Fig. 1, exepct for $C$ that is much weaker.
1143: }
1144: \end{figure}
1145: 
1146: \begin{figure}[htbp]
1147: \centering
1148: \vspace{1cm}
1149: \psfig{file=figureA3.ps,width=12cm}
1150: \caption{FIG. 3. A deep and large density depletion, the only difference with Fig. 2 is
1151: a larger value of $\eta$.
1152: See Sec. III.A..}
1153: \end{figure}
1154: 
1155: \begin{figure}[htp]
1156: \centering
1157: \vspace{1cm}
1158: \psfig{file=figureA4.ps,width=12cm}
1159: \caption{FIG. 4. An example of non symetric equilibrium where $n_g$ is a function of $a$.
1160: See Sec.  III.C for details.}
1161: \end{figure}
1162: 
1163: \begin{figure}[htp]
1164: \centering
1165: \vspace{1cm}
1166: \psfig{file=figureA5.ps,width=12cm}
1167: \caption{FIG. 5. An exemple of an equilibrium with $\eta=0$ and 
1168: a reversal of the magnetic field. Details are given in Sec III.C.}
1169: \end{figure}
1170: 
1171: \begin{figure}[htp]
1172: \centering
1173: \vspace{1cm}
1174: \psfig{file=figureA6.ps,width=12cm}
1175: \caption{FIG. 6. An example of a deep plasma cavity containing hot electrons (1 keV) 
1176: surrounded by a cold (1 eV) highly magnetized plasma.  
1177: See  section III.D for details.}
1178: \end{figure}
1179: 
1180: 
1181: 
1182: \end{document}
1183: