1: %\usepackage{amsmath}
2:
3:
4: \documentclass[aps]{revtex4}
5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
6: \usepackage{amsmath}
7: \usepackage{graphicx}
8: \usepackage{epsfig,amsmath}
9:
10: %TCIDATA{OutputFilter=Latex.dll}
11: %TCIDATA{LastRevised=Tue May 17 12:22:49 2005}
12: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
13: %TCIDATA{Language=American English}
14:
15:
16: \begin{document}
17:
18: \title{Reconnections of Vortex Loops in the Superfluid Turbulent HeII. Rates of the
19: Breakdown and Fusion processes.}
20: \author{Sergey K. Nemirovskii}
21: \affiliation{Institute for Thermophysics, Lavrentyeva, 1, 630090
22: Novosibirsk, Russia}
23: \date{today}
24:
25: \begin{abstract}
26: Kinetics of merging and breaking down vortex loops is the
27: important part of the whole vortex tangle dynamics. Another part
28: is the motion of individual lines, which obeys the Biot-Savart
29: law in presence of friction force and of applied external velocity
30: fields if any. In the present work we evaluate the
31: coefficients of the reconnection rates $A(l_{1},l_{2},l)$ and $%
32: B(l,l_{1},l_{2})$. Quantity $A$ is a number (per unit of time and
33: per unit of volume) of events, when two loops with lengths
34: $l_{1}$and $l_{2}$ collide and form the single loop of length $\ l=l_{1}+l_{2}$. Quantity $%
35: B(l,l_{1},l_{2}) $ describes the rate of events, when the single
36: loop of the length $l$ breaks down into two the daughter loops of
37: lengths $\ l_{1}$ and $l_{2}$. These quantities
38: ave evaluated as the averaged numbers of zeroes of vector $\mathbf{S}%
39: _{s}(\xi _{2},\xi _{1},t)$ connecting two points on the loops of
40: $\xi _{2}$ and $\ \xi _{1}$ at moment of time $t$. Statistics of
41: the individual loops is taken from the Gaussian model of vortex
42: tangle.\\
43: PACS-number 67.40
44: \end{abstract}
45:
46: \maketitle
47:
48:
49:
50: \section{Introduction}
51:
52: Quantized vortices appeared in quantum fluids and other systems play a
53: fundamental role in the properties of the latter. For that reason they have
54: been an object of intensive study for many years (for review and
55: bibliography see e. g. \cite{Don_book}). The greatest success in
56: investigations of dynamics of quantized vortices has been achieved in
57: relatively simple cases such as a vortex array in rotating helium or vortex
58: rings. However these simple cases are rather exception than a rule. Due to
59: extremely involved dynamics initially straight lines or rings evolve to form
60: highly chaotic structure, so called vortex tangle (see e.g. papers \cite
61: {Feynman}, \cite{Vinen} and review \cite{NF}). The vortex tangle consists of
62: a set of vortex loops of different lengths. The individual vortex loops
63: evolves obeying the Biot-Savart law, undergoing the friction force and the
64: applied external velocity field. All these processes are good understood,
65: although the according equations of motion are highly nonlinear (and
66: nonlocal) therefore general analysis can be made with use of numerical
67: methods.
68:
69: Besides motion of each of the individual loops there is one more
70: element of general evolution of the vortex tangle related to
71: collision of loops, or intersection of elements of vortex lines.
72: During intersection of lines the very complicated process, related
73: to arrangement of the vortex core takes place \cite{Koplik}.
74: Nevertheless this process is relatively short, therefore it is
75: usually accepted that the filaments instantly reconnect whenever
76: they intersect each other. It is widely appreciated that the
77: reconnection processes influence both the structure and evolution
78: of vortex tangle. However the questions how it happens and what
79: mechanisms are responsible for these remain open. For instance,
80: Feynman in his pioneering paper devoted to superfluid turbulence
81: proposed scenario how the vortex tangle decays in absence of
82: applied external counterflow. According to this scenario a fusion
83: of small vortex rings into larger ones as well as a breakdown into
84: smaller ones is possible at the moments of the reconnection events
85: (see Fig.~\ref{Feynman}).
86: \begin{figure}[tbp]
87: \includegraphics[width=7cm]{Feynman_decay.eps}
88: \caption{The reconnection process schematically (Feynman, 1955, Fig.\ 10). (a)
89: initial stage, (b) and (c) are stages of collapse, (d) reconnection stage,
90: (e) stage of degeneration of vortex rings into thermal excitations.}
91: \label{Feynman}
92: \end{figure}
93: In assumption that \textit{\ on the average} the last property
94: dominates i.e. the cascade like process of formation of smaller
95: and smaller loops forms. When the scale of the small rings becomes
96: of the order of the interatomic distances, which is the final
97: stage of the cascade, the vortex motion is degenerated into
98: thermal excitations. This key idea that degeneration of the vortex
99: tangle occurs due to cascade-like transferring of the length in
100: space of scale of sizes of vortex loops was indirectly confirmed
101: only in numerical calculations, where the procedure of artificial
102: elimination of small loops had been used
103: \cite{Schwarz88}-\cite{Tsubota00}. \
104:
105: In spite of the recognized importance of the reconnecting loops kinetics,
106: the numerical results remain main source of information about this process.
107: The obvious lack of theoretical investigations interferes with deep insight
108: in the nature of this phenomena (this question had been recently discussed
109: in \cite{Barenghi2004}). For instance it is not clear how the cascade of
110: length in space of vortex loops sizes is formed, what mechanisms are
111: responsible for this, what quantities determine an intensity of cascade, and
112: why at all breakdown of the loop prevails.
113:
114: Of course the scarcity of analytic investigations related to
115: incredible complexity of the problem. Indeed we have to deal the
116: set of objects with not fixed number of elements, they can born
117: and die. Thus, some analog of the secondary quantization method is
118: required with the difference that objects (vortex loops)
119: themselves possess an infinite number of degree of freedom with
120: very involved dynamics. Clearly this problem can hardly be
121: resolved in nearest future. Recently in \cite {Copeland98} much
122: more modest approach, based on the ''rate equation'' for
123: distribution function $n(l)$ was elaborated in context of cosmic
124: strings. Following this work we introduce distribution function
125: $n(l,t)$ of density of loop in ''space'' of their lengths. It is
126: defined as a number of loops (per unit of volume) with lengths
127: between $l$ and $l+dl$. Due to reconnection processes $n(l,t)$ can
128: vary.
129:
130: \begin{figure}[tbp]
131: \includegraphics[width=8cm]{A_B_loops_1}
132: \caption{Schematic sketch of the fusion and breaking-down of vortex loops.
133: Rates of these processes characterized by the coefficients $A(l_{1},l_{2},l)$
134: and $B(l,l_{1},l_{2})$.}
135: \label{A_B_loops}
136: \end{figure}
137:
138: We discriminate two types of processes, namely the fusion of two
139: loops into the larger single loop and the breakdown of single
140: loop into two daughter loops. The kinetic of vortex tangle is
141: affected by the intensity of the introduced processes.
142: The intensity of the first process is characterized by the rate of collision $%
143: A(l_{1},l_{2},l)$ of two loops with lengths $l_{1}$and $l_{2}$ and forming
144: the loop of length $\ l=l_{1}+l_{2}$. The intensity of the second process is
145: characterized by the rate of self-intersection $B(l,l_{1},l_{2})$ of loop of
146: the length $l$ into two daughter loops with lengths $\ l_{1}$ and $l_{2}$
147: (see Fig.~\ref{A_B_loops}). In view of exposed above we can directly write
148: out the master ''kinetic'' equation for rate of change distribution function
149: $n(l,t)$.
150: \begin{align}
151: \frac{\partial n(l,t)}{\partial t}& = \label{kinetic equation} \\
152: \int \int A(l_{1},l_{2},l)n(l_{1})n(l_{2})\delta
153: (l-l_{1}-l_{2})dl_{1}dl_{2}\;\;\;\;\;\;\;\;\;\;\;\;l_{1}+l_{2}& \rightarrow l
154: \notag \\
155: -\int \int A(l_{1},l,l_{2},)\delta
156: (l_{2}-l_{1}-l)n(l)n(l_{1})dl_{1}dl_{2}\;\;\;\;\;\;\;\;\;\;\;\;\;l_{1}+l&
157: \rightarrow l_{2} \notag \\
158: -\int \int A(l_{2},l,l_{1},)\delta
159: (l_{1}-l_{2}-l)n(l)n(l_{1})dl_{1}dl_{2}\;\;\;\;\;\;\;\;\;\;\;\;l_{2}+l&
160: \rightarrow l_{1} \notag \\
161: -\int \int B(l_{1},l_{2},l)n(l)\delta
162: (l-l_{1}-l_{2})dl_{1}dl_{2}\;\;\;\;\;\;\;\;\;\;\;\;\ \;\;\;\;\;\;\;\;\;l&
163: \rightarrow l_{1}+l_{2}\;\;\;\;\; \notag \\
164: +\int \int B(l,l_{2},l_{1})\delta
165: (l_{1}-l-l_{2})n(l_{1})dl_{1}dl_{2}\;\;\;\;\;\;\;\
166: \;\;\;\;\;\;\;\;\;\;\;l_{1}& \rightarrow l+l_{2} \notag \\
167: +\int \int B(l,l_{1},l_{2})\delta
168: (l_{2}-l-l_{1})n(l_{1})dl_{1}dl_{2}\;\;\;\;\
169: \;\;\;\;\;\;\;\;\;\;\;\;\;\;l_{2}& \rightarrow l+l_{1} \notag
170: \end{align}
171: All of the processes are depicted at the left of each line. Clear that the
172: ''kinetic'' equation has ''bookkeeping'' character. Physics of this approach
173: lies in the ''correct'' determinations of coefficient of that equations $%
174: A(l_{1},l_{2},l)$ and $B(l,l_{1},l_{2})$ on the base of some more
175: or less plausible model. \newline In the present work we derive
176: some general (independent on the model) relations for reconnection
177: rates $A(l_{1},l_{2},l)$ and $B(l,l_{1},l_{2}))$ ( Section II).
178: Then in Sec. III we apply these relations to evaluate the
179: reconnection rates on the base of Gaussian model of the vortex
180: tangle elaborated earlier \cite{Nemirovskii_97_1}. Section IV is
181: devoted to conclusion and plans for future investigations.
182: \section{Mathematical identities for $A(l_{1},l_{2},l)$ and $%
183: B(l,l_{1},l_{2}).$} In this section we will formulate mathematical
184: definition for quantities $\ A(l_{1},l_{2},l)$ and$\,$\
185: $B(l,l_{1},l_{2})$. We will start with the latter quantity
186: $B(l,l_{1},l_{2}).$ By
187: definition its physical meaning is just frequency of events when part of line with total length $%
188: l $ intersects the same line to create two two daughter loops with lengths $%
189: l_{1}$ and $l_{2}$ (see Fig.~\ref{A_B_loops}), that is
190: self-crossing event. As it is already stated we assume that each
191: crossing event leads to reconnection of lines. The elements of
192: vortex line are described as a function $\mathbf{s}(\xi,t)$ \ (See
193: Fig.~\ref{intersection}) that is time dependant radius-vector of
194: the points resting on loop. Variable $\xi $ labels the points of
195: the loop. It is convenient to choose variable $\xi $ to be equal
196: to the arc length, $(0\leq \xi \leq l)$. Let us consider function
197: \begin{equation}
198: \mathbf{S}_{s}(\xi _{2},\xi _{1},t)=\mathbf{s}(\xi _{2},t)-\mathbf{s}(\xi
199: _{1},t), \label{distance_single}
200: \end{equation}
201: that is just $3D$ vector connecting points $\mathbf{s}(\xi
202: _{2},t_{2})$ and $\mathbf{s}(\xi _{1},t_{1})$. Clearly that
203: condition $\mathbf{S}_{b}(\xi _{2},\xi _{1},t)=0$ implies that the
204: self-crossing event of parts of line with label-coordinates $\xi
205: _{2},\xi _{1}$ occurs at moment of time $t$. \ The quantity
206: $\mathbf{S}_{b}(\xi _{2},\xi _{1},t)$ is fluctuating 3-component
207: function of three arguments $\xi _{2},\xi _{1},t$. We are
208: interested in how often $\mathbf{S}_{b}(\xi _{2},\xi _{1},t)$
209: vanishes
210: in cube of space $\mathbf{\zeta =\{}\xi _{2},\xi _{1},t\}$. Points in space $%
211: \mathbf{\zeta =\{}\xi _{2},\xi _{1},t\}$ where function
212: $\mathbf{S}_{b}(\xi _{2},\xi _{1},t)$ vanishes, are points of
213: cross of $3$ surfaces ${S}_{i}(\xi _{2},\xi _{1},t)=0,\;\;i=x,y,z$
214: in space $\mathbf{\zeta =\{}\xi _{2},\xi _{1},t\}$ as it is shown
215: in Fig.~\ref{3_sufaces}.
216:
217: \begin{figure}[tbp]
218: \includegraphics[width=6cm]{intersection.eps}
219: \caption{Schematic sketch of vortex loop. Elements of line are described as
220: vectors $\mathbf{s}(\protect\xi )$, where the label variable $\protect\xi $
221: is taken here as the arc length. We associate the moment of intersection with the
222: vanishing of vector $\mathbf{S}(\protect\xi _{1},\protect\xi _{1},t)$
223: connecting points $\mathbf{s}(\protect\xi _{2},t)$ and $\mathbf{s}(\protect%
224: \xi _{1},t)$. \ }
225: \label{intersection}
226: \end{figure}
227:
228: \begin{figure}[tbp]
229: \includegraphics[width=7cm]{3_surfaces.eps}
230: \caption{The reconnection events can be described as zeroes of function $%
231: \mathbf{S}_{b}(\protect\xi _{2},\protect\xi _{1},t)$. In space of
232: its variables $\mathbf{\protect\zeta =\{}\protect\xi
233: _{2},\protect\xi _{1},t\}$ these points are cross of $3$ surfaces
234: ${S}_{i}(\protect\xi _{2},\protect\xi _{1},t)=0,\;\;i=x,y,z$.}
235: \label{3_sufaces}
236: \end{figure}
237:
238: From theory of generalized function it follows that number of these points
239: (we denote them below as $\mathbf{\zeta }_{a}$) can be expressed via $\delta
240: $-function of quantity $\mathbf{S}_{b}(\xi _{2},\xi _{1},t\dot{)}$ with the
241: help of following formula .
242: \begin{equation}
243: \sum_{a} \delta (\mathbf{\zeta }-\mathbf{\zeta }_{a})=\left| \frac{\partial
244: (X,Y,Z)}{\partial (\xi _{2},\xi _{1},t)}\right| _{\mathbf{\zeta }=\mathbf{%
245: \zeta }_{a}}\delta (\mathbf{S}_{b}(\xi _{2},\xi _{1},t))
246: \label{s_zeroes_number}
247: \end{equation}
248: Here $X,Y,Z$ are the components of vector $\mathbf{S}_{b}(\xi _{2},t_{2},\xi
249: _{1},0) $ . By integration of both parts of (\ref{s_zeroes_number}) over $%
250: d\xi _{1}d\xi _{2}$ we would obtain the full number of
251: intersections (per unit time). If, further to introduces
252: additional constraint $\delta (\xi _{2}-\xi _{1}-l_{1})\,$\
253: implying that the distance (along line) between chosen points is
254: equal to $l$ and to integrate over $d\xi _{1}d\xi _{2}$, we obtain
255: the rate of self-intersection of line with length $l$ and
256: breakdown it into pieces $l_{1}$ and $l-l_{1}$.
257:
258: In addition we have to do averaging \ over all possible fluctuating
259: configurations. Thus the coefficient $B(l_{1},l-l_{1},l)$ with dimension $%
260: \left[ s\right] =s^{-1}cm^{-1}$ is equal to
261: \begin{equation}
262: B(l_{1},l-l_{1},l)=\int \int d\xi _{1}d\xi _{2}\delta (\xi _{2}-\xi
263: _{1}-l_{1})\,\left\langle \left| \frac{\partial (X,Y,Z)}{\partial (\xi
264: _{2},\xi _{1},t)}\right| _{\mathbf{\zeta }=\mathbf{\zeta }_{a}}\delta (%
265: \mathbf{S}_{b}(\xi _{2},\xi _{1},t))\right\rangle . \label{B_def}
266: \end{equation}
267:
268: To obtain coefficient $A(l_{1},l_{2},l)$ we use the similar procedure. Let
269: us consider two loops with length $l_{1}$ and $l_{2}$. Our purpose now to
270: find the rate $A(l_{1},l_{2},l)$ of fusion of these two loops into one loop
271: of length $l=l_{1}+l_{2}$. Dimension of $A(l_{1},l_{2},l)$ is $%
272: [A]=cm^{3}s^{-1}$. As previously we describe vortex filaments by positions
273: of radius vectors their elements $\mathbf{s}(\xi _{1},t)$ and $\mathbf{s}%
274: (\xi _{2},t)$. Here we have two label variables $\xi _{1},\xi _{2}$
275: belonging to different loops and running in limits $(0\leq \xi \leq l_{1})$
276: and $(0\leq \xi \leq l_{2})$ respectively. One more important difference
277: with the previous case is that both functions $\mathbf{s}(\xi _{1},t)$ and $%
278: \mathbf{s}(\xi _{2},t)$ should depend on ''initial'' positions $\mathbf{s}%
279: (\xi _{2}=0,t)=\mathbf{R}_{1}(t)$ and $\mathbf{s}(\xi _{2}=0,t)=\mathbf{R}%
280: _{2}(t)$, chosen arbitrary. Of course in previous case of
281: self-intersection of single loop, quantity $\mathbf{s}(\xi ,t)$
282: also depended on ''initial'' positions $\mathbf{R(t)}$, but it did
283: not influence the rate of self-intersection. Now for case of the
284: fusion this dependance is important, since very distant loops have
285: small probability to collide. Let us introduce the ''fusion''
286: functions
287: \begin{equation}
288: \mathbf{S}_{f}(\xi _{2},\xi _{1},t_{1})=\mathbf{s}(\xi _{2},t)-\mathbf{s}%
289: (\xi _{1},t). \label{distance_double}
290: \end{equation}
291: Repeating the considerations for case of the single loop we find that number
292: of reconnection (per unit of time) of points $\xi _{2},\xi _{1}$ formally
293: coincides with (\ref{s_zeroes_number})\bigskip
294: \begin{equation}
295: \sum \delta (\mathbf{\zeta }-\mathbf{\zeta }_{a})=\left| \frac{\partial
296: (X,Y,Z)}{\partial (\xi _{2},\xi _{1},t)}\right| _{\mathbf{\zeta }=\mathbf{%
297: \zeta }_{a}}\delta (\mathbf{S}_{f}(\xi _{2},\xi _{1},t))
298: \label{m_zeroes_number}
299: \end{equation}
300: with the difference that $\xi _{2},\xi _{1}$ belong to different curves.
301: Since intersections of any elements of lines lead to fusion of the loops we
302: have to integrate (\ref{m_zeroes_number}) over $d\xi _{1}d\xi _{2}$. The
303: result obtained is valid for chosen pair of loops. To obtain the
304: total number of events we have to multiply the result obtained by quantity $ n(l_{1})n(l_{2})d
305: \mathbf{R}_{1}d\mathbf{R}_{2}$, which is just a full number of loops of chosen sizes in the whole volume.
306: Comparing with the master kinetic equation (%
307: \ref{kinetic equation}) we find the final expression for fusion coefficient $%
308: A(l_{1},l_{2},l)$
309:
310: \begin{equation}
311: A(l_{1},l_{2},l)=\frac{1}{\text{VOLUME}}\int \int d\mathbf{R}_{1}d\mathbf{R}%
312: _{2}\int \int d\xi _{1}d\xi _{2}\,\left\langle \left| \frac{\partial (X,Y,Z)%
313: }{\partial (\xi _{2},\xi _{1},t)}\right| _{\mathbf{\zeta }=\mathbf{\zeta }%
314: _{a}}\delta (\mathbf{S}(\xi _{2},\xi _{1},t))\right\rangle . \label{A_def}
315: \end{equation}
316:
317: Thus we obtained expressions (\ref{B_def})and (\ref{A_def}), which
318: allow to calculate the rates of reconnections of the fusion and
319: breakdown vortex loops. They are however just formal mathematical
320: identities. Concrete results depend on statistics of individual
321: lines. Therefore to move further we have to ascertain the
322: procedure for averaging. It will be done in the next section.
323:
324: \section{Gaussian model case}
325:
326: \subsection{\protect\bigskip The generalized Wiener distribution \ }
327:
328: To evaluate quantities $B(l_{1},l_{2},l)$ and $A(l_{1},l_{2},l)$ written in
329: form (\ref{B_def}) and (\ref{A_def}) one needs to know statistics of
330: individual loops. In general case this statistics should be extracted from
331: investigation of the full dynamical problem. The according statement of such
332: problem includes equation of motion (Biot-Savart law for quantum vortices),
333: dissipative effects (interaction with normal component) and additional
334: Langevin force responsible for chaotic behavior. The problem becomes more
335: involved by circumstance that other loops also influence dynamics of the
336: chosen loop. At this stage we choose another way, namely we use the Gaussian
337: model of the vortex tangle elaborated by author \cite{Nemirovskii_97_1}. To
338: develop this approach the trial distribution functional of Gaussian form had
339: been consructed. This functional absorbed all properties of the superfluid
340: turbulence known from both experimental studies and numerical simulations.
341: According to this model the ''average'' vortex loop has a
342: typical structure shown in Fig.~\ref{RW}.
343: \begin{figure}[tbp]
344: \includegraphics[width=7cm]{RW_1.eps}
345: \caption{Snapshot of the ''average'' vortex loop obtained from analysis of
346: the statistical properties. Close $(\Delta \protect\xi \ll R)$ parts of the
347: line are separated in $3D$ space by distance $\Delta \protect\xi .$ The
348: distant parts $(R\ll \Delta \protect\xi )$ are separated in $3D$ space by
349: the distance $\protect\sqrt{\protect\xi _{0}\Delta \protect\xi }$, i.e. the
350: vortex loop has the typical random walking structure. }
351: \label{RW}
352: \end{figure}
353: The close parts of the loop separated (along line) by distance
354: $\xi _{2}-\xi _{1}$ smaller then the mean radius of curvature $\xi
355: _{0}$ are strongly correlated, $\left\langle \mathbf{s}^{\prime
356: }(\xi _{1},t)\mathbf{s}^{\prime }(\xi _{2},t)\right\rangle
357: \rightarrow 1,$ ($\mathbf{s}^{\prime }$ is the tangent vector) and
358: line is smooth. \ Remote parts of the line $\ \ (\xi _{2}-\xi
359: _{1}\gg \xi _{0})$ are not correlated at all, $\left\langle \mathbf{s}%
360: ^{\prime }(\xi _{1},t)\mathbf{s}^{\prime }(\xi _{2},t)\right\rangle
361: \rightarrow 0$. Thus for large separations the vortex loop has a typical
362: ''random walking structure'' with the Wiener distribution. This
363: ''semifractal'' behavior satisfies to the generalized \ Wiener distribution.
364: Namely, the probability $\mathcal{P}(\left\{ \mathbf{s}(\xi ,t)\right\} )$\
365: to find some particular configuration $\left\{ \mathbf{s}(\xi ,t)\right\} $
366: is expressed by the probability distribution functional (see for details in
367: \cite{Nemirovskii_97_1})
368: \begin{equation}
369: \mathcal{P}(\left\{ \mathbf{s}(\xi ,t)\right\} )=\mathcal{N}\exp
370: \left( -\int\limits_{0}^{l}\int\limits_{0}^{l}\mathbf{s}^{\prime
371: \alpha }(\xi _{1},t)\Lambda ^{\alpha \beta }(\xi _{1}-\xi
372: _{2})\mathbf{s}^{\prime \beta }(\xi _{2},t)d\xi _{1}d\xi
373: _{2}\right) . \label{Gauss_model}
374: \end{equation}
375: Here $\mathcal{N}$ is normalizing factor, $l$ is the length of
376: curve, it is supposed to be much larger then mean radius of
377: curvature $\xi _{0}$. The probability distribution functional \
378: (\ref{Gauss_model}) introduced in \cite {Nemirovskii_97_1}
379: distribution considered anisotropy and \ polarization of vortex
380: tangle in counterflowing turbulent helium. This fact is reflected
381: by the circumstance that $\Lambda ^{\alpha \beta }(\xi _{1}-\xi
382: _{2})$ is the full matrix. In practice it is more convenient to
383: deal with characteristic functional $W(\{\mathbf{P}(\xi,t)\})$
384: defined as
385: \begin{equation}
386: W(\{\mathbf{P}(\xi,t)\})\;=\left\langle \exp \left( i\int\limits_{0}^{l}%
387: \mathbf{P}(\xi ,t)\mathbf{s}^{\prime }(\xi ,t)d\xi \ \right) \right\rangle
388: \;. \label{CF_def}
389: \end{equation}
390: The characteristic functional enables us calculate any averages depending on
391: vortex lines configuration $\left\{ \mathbf{s}(\xi ,t)\right\} $ by simple
392: functional differentiation. For instance the average tangent vector $%
393: \left\langle \mathbf{s}_{\alpha }^{\prime }(\xi _{1})\right\rangle $ or the
394: correlation function between orientation of the different elements of the
395: vortex filaments $\left\langle \mathbf{s}_{\alpha }^{\prime }(\xi _{1})%
396: \mathbf{s}_{\beta }^{\prime }(\xi _{2})\right\rangle $ are readily expressed
397: via characteristic functional accordingly to the following rules:
398:
399: \begin{equation}
400: \left\langle \mathbf{s}_{\alpha }^{\prime }(\xi _{1})\right\rangle =\;\left.
401: \frac{\delta W}{i\delta \mathbf{P}^{\alpha }(\xi _{1})}\right| _{\;\mathbf{P}%
402: \;=\;0},\;\;\;\;\;\left\langle \mathbf{s}_{\alpha }^{\prime }(\xi _{1})%
403: \mathbf{s}_{\beta }^{\prime }(\xi _{2})\right\rangle =\;\left. \frac{\delta
404: ^{2}W}{i\delta \mathbf{P}^{\alpha }(\xi _{1})\;\;i\delta \mathbf{P}^{\beta
405: }(\xi _{2})}\right| _{\mathbf{P}\;=\;0} \label{r'(W)}
406: \end{equation}
407: Calculation of the characteristic functional $W(\{\mathbf{P}(\xi ,t)\})$ (%
408: \ref{CF_def}) on the base of the probability functional \
409: (\ref{Gauss_model}) is reduced to functional integration, which,
410: in turn, reduces to the ''full square procedure''. \ The result is
411:
412: \begin{equation}
413: W(\{\mathbf{P}(\xi,t)\})\;=\exp \left\{
414: -\int\limits_{0}^{l}\int\limits_{0}^{l}\mathbf{P}^{\alpha }(\xi
415: _{1})N^{\alpha \beta }(\xi _{1}-\xi _{2})\mathbf{P}^{\beta }(\xi _{2})d\xi
416: _{1}d\xi _{2}\right\}. \label{CF_N}
417: \end{equation}
418:
419: To avoid unnecessary lengthy calculations we simplify the model
420: expressed by the probability distribution functional \
421: (\ref{Gauss_model}), namely we omit both the anisotropy and \
422: polarization. In this case the matrix $N^{\alpha \beta }(\xi
423: _{1}-\xi _{2})$ used in (\cite{Nemirovskii_97_1}) can be taken as
424: \begin{equation}
425: N^{\alpha \beta }(\xi _{1}-\xi _{2})=\delta _{\alpha \beta }\frac{1}{6\left(
426: 1-2\xi _{0}\sqrt{\pi }/l\right) }\left( \exp \left[ -\frac{(\xi _{1}-\xi
427: _{2})^{2}}{4\xi _{0}^{2}}\right] -\frac{2\xi _{0}\sqrt{\pi }}{l}\right) .
428: \label{N_isotrop}
429: \end{equation}
430: For small separation $\ (\xi _{2}-\xi _{1}\ll \xi _{0})$ , $%
431: \sum\nolimits_{\alpha }N^{\alpha }(\xi _{1}-\xi _{2})\rightarrow
432: 1/2$ \ that guarantees that $\left\langle \mathbf{s}^{\prime }(\xi
433: ,t)\mathbf{s}^{\prime }(\xi,t)\right\rangle =1,$ as it should be
434: for smooth lines (we recall that for pure random walking \ line
435: the quantity $\left\langle \mathbf{s}^{\prime }(\xi
436: ,t)\mathbf{s}^{\prime }(\xi ,t)\right\rangle $ does not exist at
437: all). For large separation $(\xi _{2}-\xi _{1}\gg \xi _{0})$
438: exponents tends to $\delta _{\alpha \beta }(2\sqrt{\pi } \xi
439: _{0}/6)\delta (\xi _{1}-\xi _{2}),$ and correlation between
440: tangent
441: vectors weakens,\ $\left\langle \mathbf{s}^{\prime }(\xi ,t)\mathbf{s}%
442: ^{\prime }(\xi ,t)\right\rangle \rightarrow 0.$ \ The second term
443: in parenthesis of expression (\ref{N_isotrop}) appears due to
444: closeness of the loops. It ensures that
445: \begin{equation*}
446: \int\limits_{0}^{l}\int\limits_{0}^{l}d\xi _{1}d\xi _{2}\left\langle \mathbf{%
447: s}^{\prime }(\xi _{1},t)\mathbf{s}^{\prime }(\xi _{2},t)\right\rangle
448: =\left\langle (\mathbf{s}(l,t)-\mathbf{s}(0,t))^{2}\right\rangle =0
449: \end{equation*}
450: as it should be for the closed lines (see explanations in (\cite
451: {Nemirovskii_97_1})). Thus the characteristic functional with
452: function $\ N$ satisfies to necessary ''semifractal'' behavior of
453: line \ and will be used further for evaluation of the rate
454: reconnection coefficients. Here it is necessary to do two
455: important comments. First one concerns the ''semifractal''
456: structure of the vortex loop. In principle, reconnections events
457: occur due random walking behavior with non-zero probability of the
458: meeting of any two elements. However, to evaluate the according
459: rates on base of relations (\ref{B_def}), (\ref{A_def}), we need
460: to know not only an average of $\delta (\mathbf{S}(\xi _{2},\xi
461: _{1},t))$ but also an average of the Jacobian. But the latter
462: includes averaged derivatives (with respect two both label
463: variable$\xi _{2},\xi _{1}$ and time). It is known fact however
464: that quantities like $\left\langle \left| \mathbf{s}_{\alpha
465: }^{\prime }(\xi _{1})\right| \right\rangle ,\;\left\langle
466: \mathbf{s}_{\alpha }^{\prime }(\xi _{1})\mathbf{s}_{\beta
467: }^{\prime }(\xi _{2})\right\rangle $ etc. are ill-defined for pure
468: random walking structure, For that reason we need to smooth away
469: the curve on small scales. The second remark concerns the mean
470: radius of curvature $\xi _{0}$. This quantity\ appears here as
471: important parameter of theory of random walk. It is frequently \
472: called as ''an elementary step'' or persistence length. In fact
473: the theory described by the Wiener distribution is failed for
474: small scales less then $\xi _{0}$ , therefore usually this value
475: appears as a low cut-off of the whole approach.
476:
477: \section{\protect\bigskip Evaluation of $A(l_{1},l_{2},l)$ and $%
478: B(l,l_{1},l_{2}).$}
479:
480: \subsection{evaluation of $\ B(l,l_{1},l_{2})$}
481:
482: Let us come back to the rate coefficients. We start with the
483: self-intersection processes. Positions of line elements $\mathbf{s}(\xi
484: _{2},t),\mathbf{s}(\xi _{1},t)$ and relative vector $\mathbf{S}_{b}(\xi
485: _{2},\xi _{1},t)$ are strongly fluctuating quantities having the Gaussian
486: statistics. Due to the Wick theorem \ the average in integrand of (\ref
487: {B_def}) \ can be taken as a sum of all possible pairs of quantity $\mathbf{S%
488: }_{b}(\xi _{2},t,\xi _{1},t)\,\ \ $and its derivatives. Because of
489: uniformity in $\xi $ space, quantity $\mathbf{S}_{b}(\xi _{2},\xi _{1},t)\,$%
490: \ depends on $\left| \xi _{2}-\xi _{1}\right| ,$ for that reason all
491: averages of structure $\left\langle (\partial X/\partial \xi _{1})\delta
492: (X(\xi _{2},\xi _{1},t))\right\rangle $ vanish, therefore only pairs
493: separately from $\mathbf{S}_{b}(\xi _{2},\xi _{1},t)\,$ and from its
494: derivatives survive. As a result the average of production is equal to
495: production of \ averages and each of the factors can be evaluated separately
496: \begin{equation}
497: \left\langle \left| \frac{\partial (X,Y,Z)}{\partial (\xi _{2},\xi _{1},t)}%
498: \right| _{\mathbf{\zeta }=\mathbf{\zeta }_{a}}\delta (\mathbf{S}_{b}(\xi
499: _{2},\xi _{1},t))\right\rangle =\,\left\langle \left| \frac{\partial (X,Y,Z)%
500: }{\partial (\xi _{2},\xi _{1},t)}\right| _{\mathbf{\zeta }=\mathbf{\zeta }%
501: _{a}}\right\rangle \left\langle \delta (\mathbf{S}_{b}(\xi _{2},\xi
502: _{1},t))\right\rangle. \label{self_separate}
503: \end{equation}
504: As mentioned, the use of the characteristics functional (\ref{CF_def}), (\ref
505: {CF_N}) allows to calculate any averaged functional of configurations $%
506: \left\{ \mathbf{s}(\xi ,t)\right\}.$ Let us show how to evaluate $%
507: \left\langle \delta \mathbf{S}_{s}(\xi _{2},\xi _{1},t)\right\rangle $. With
508: use of the standard integral representation for $\delta $-function
509: \begin{equation*}
510: \delta (x)=\frac{1}{(2\pi )}\int_{-\infty }^{\infty }e^{ixy}dy,
511: \end{equation*}
512: \bigskip we rewrite $\left\langle \delta \mathbf{S}_{b}(\xi _{2},t,\xi
513: _{1},t)\right\rangle $ as
514: \begin{eqnarray}
515: \left\langle \delta \mathbf{S}_{b}(\xi _{2},\xi _{1},t)\right\rangle &=&%
516: \frac{1}{(2\pi )^{3}}\int \left\langle \exp \left[ i\mathbf{y(s(}\xi
517: _{2},t_{2}\mathbf{)-s(}\xi _{1},t_{1}\mathbf{))}\right] \right\rangle d^{3}%
518: \mathbf{y=} \notag \\
519: &&\frac{1}{(2\pi )^{3}}\int \left\langle \exp \left( i\int\limits_{\xi
520: _{1}}^{\xi _{2}}\mathbf{ys}^{\prime }(\xi ,t)d\xi dt\ \right) \right\rangle
521: d^{3}\mathbf{y.} \label{Delta_integral}
522: \end{eqnarray}
523: Comparing (\ref{CF_def}) and (\ref{Delta_integral}) we conclude
524: that the integrand in last term of (\ref{Delta_integral}) is just
525: the characteristic functional $W(\{\mathbf{P}(\xi ,t)\}),$ taken
526: at value of $\mathbf{P}(\xi ,t) $
527: \begin{equation}
528: \mathbf{P}(\xi )\;=\;-\mathbf{y}\theta (\xi -\xi _{1})\theta (\xi _{2}-\xi ).
529: \label{TETA}
530: \end{equation}
531: Here $\theta (\xi )$ is the unit step-wise function. Relation (\ref{TETA})
532: implies that we choose in integrand of the characteristic functional only
533: points lying in interval from $\xi _{1}$ to $\xi _{2}$ on \ the curve.
534: Substituting this value $\mathbf{P}(\xi ,t)$ expressed by (\ref{TETA}) in CF
535: (\ref{CF_N}) with function \ $N^{\alpha \beta }(\xi _{1}-\xi _{2})$ in form (%
536: \ref{N_isotrop}) we obtain
537: \begin{equation}
538: \left\langle \delta \mathbf{S}_{b}(\xi _{2},t,\xi _{1},t)\right\rangle =%
539: \frac{1}{(2\pi )^{3}}\int \exp \left[ -\mathbf{y}^{2}\frac{2\sqrt{\pi }\xi
540: _{0}}{6}\left( (\xi _{2}-\xi _{1})-(\xi _{2}-\xi _{1})^{2}/l\right) \right]
541: d^{3}\mathbf{y=}\left( \frac{3}{2\pi ^{3/4}\xi _{0}\left( (\xi _{2}-\xi
542: _{1})-(\xi _{2}-\xi _{1})^{2}/l\right) }\right) ^{3/2} \label{delta_self}
543: \end{equation}
544: Evaluation of absolute value of Jacobian in (\ref{self_separate}) we perform
545: by use of $\left| J\right| =\sqrt{J^{2}}.$ Furthermore it would be
546: convenient for the sake of generalization to use vector velocity $\mathbf{V}%
547: _{l}=d\mathbf{s}/dt,$ instead of calculation of it in explicit form,
548: expressing velocity via the vortex filament configuration $\{\mathbf{s}(\xi )\}.$
549: Calculation of $J^{2}$ can be fulfilled writing \ Jacobian in explicit form
550: and subsequent applying the Wick theorem. Simple but tedious calculations
551: leads to result that
552: \begin{equation}
553: J^{2}=2\left\langle \mathbf{V}_{lx}^{2}\right\rangle \left\langle (\partial
554: \mathbf{s}_{y}\mathbf{/}\partial \xi _{1})^{2}\right\rangle \left\langle
555: (\partial \mathbf{s}_{z}\mathbf{/}\partial \xi _{2})^{2}\right\rangle
556: +2\left\langle \mathbf{V}_{lx}^{2}\right\rangle \left\langle (\partial
557: \mathbf{s}_{y}\mathbf{/}\partial \xi _{2})^{2}\right\rangle \left\langle
558: (\partial \mathbf{s}_{z}\mathbf{/}\partial \xi _{1})^{2}\right\rangle +p.p.
559: \label{J_squared}
560: \end{equation}
561: where $p.p.$ all permutations with respect to $x,y,z$. \ Estimating $%
562: \left\langle (\partial \mathbf{s}_{y}\mathbf{/}\partial \xi
563: _{1})^{2}\right\rangle $ and similar terms as $1/3$ we obtain that $\left|
564: J\right| =\frac{2}{3}\left| \mathbf{V}_{l}\right| .$\bigskip\ After use of
565: the integration $\int \int d\xi _{1}d\xi _{2}\delta (\xi _{2}-\xi
566: _{1}-l_{1})\,$\ (see \ref{B_def}) we finally obtain
567: \begin{equation}
568: B(l_{1},l-l_{1},l)=b_{s}\ast V_{l}\frac{l}{(\xi
569: _{0}(l_{1}-l_{1}^{2}/l))^{3/2}} \label{B_final}
570: \end{equation}
571: where constant $b_{s}=\left( \sqrt{3}/8\pi ^{9/4}\right) \approx
572: 0.0164772.$ We introduced in coefficient $B$ the additional factor
573: $1/2$ in order to avoid the over-counting of the reconnection
574: events, since decays $l\rightarrow l_{1}+l_{2}$ and $l\rightarrow
575: l_{2}+l_{1}$ describe the same process, though the both enter into
576: equations. Let us recall that the quantity $\xi _{0}$ is the mean
577: radius of curvature. The numerical factor $b_{s}\approx 0.0164772$
578: is the result of particular approximation used in paper (\cite
579: {Nemirovskii_97_1}). If for instance we used pure Wiener
580: distribution \ with the persistency length $\ \xi _{0}$ and
581: disregard the closeness of line, we would have for coefficient
582: $B(l_{1},l-l_{1},l)$ the expression
583: \begin{equation}
584: B_{C}(l_{1},l-l_{1},l)=b_{s}\ast V_{l}\frac{l}{(\xi _{0}l_{1})^{3/2}}
585: \end{equation}
586: where $b_{s}\approx 0.11.$ This result is very similar to result
587: obtained earlier in paper \cite{Copeland98} from the qualitative
588: consideration. Yet it is important also that in this way it is not
589: clear how to relate the persistency length $\ \xi _{0}$ with the
590: mean radius of curvature.
591:
592: \subsection{\protect\bigskip evaluation of $A(l_{1},l_{2},l)$}
593:
594: Let us now evaluate quantity $A(l_{1},l_{2},l)$ defined by
595: relation (\ref {A_def}). We again (as for the previous case)
596: evaluate average from Jacobian and $\delta $-function separately.
597: Contribution from Jacobian coincides with the previous result\
598: $\left| J\right| =\frac{2}{3}\left| \mathbf{V}_{l}\right| $. The
599: rest $\delta $-function part can be evaluated with the help of the
600: CF obtained above. Unlike previous case we have to know two-loop
601: distribution function. Since we omit interaction of loops (until
602: the reconnection event occurs) the CF \ for two loops with lengths $l_{1}$%
603: and $\ l_{2}$ is just production of the expressions of type (\ref{CF_N})
604:
605: \begin{eqnarray}
606: W(\{\mathbf{P}_{1}(\xi )\},\{\mathbf{P}_{2}(\xi )\})\; &=&\exp \left\{
607: -\int\limits_{0}^{l_{1}}\int\limits_{0}^{l_{1}}\mathbf{P}^{\alpha }(\xi
608: _{1})N_{1}^{\alpha \beta }(\xi _{1}-\xi _{2})\mathbf{P}^{\beta }(\xi
609: _{2})d\xi _{1}d\xi _{2}\right\} \times \label{W_double} \\
610: &&\exp \left\{ -\int\limits_{0}^{l_{2}}\int\limits_{0}^{l_{2}}\mathbf{P}%
611: ^{\alpha }(\xi _{1})N_{2}^{\alpha \beta }(\xi _{1}-\xi _{2})\mathbf{P}%
612: ^{\beta }(\xi _{2})d\xi _{1}d\xi _{2}\right\} . \notag
613: \end{eqnarray}
614: Quantities $N_{1}^{\alpha \beta }(\xi _{1}-\xi _{2})$ and $N_{2}^{\alpha
615: \beta }(\xi _{1}-\xi _{2})$ differ from each other only by lengths of loops $%
616: l_{1}$and $\ l_{2}$, entering expressions for $N_{1}^{\alpha \beta
617: }$.
618: Further, by use of the standard integral representation for $\delta $%
619: -function we have \bigskip
620: \begin{equation}
621: \left\langle \delta \mathbf{S}_{f}(\xi _{2},\xi _{1},t)\right\rangle =\frac{1%
622: }{(2\pi )^{3}}\int \left\langle \exp \left[ i\mathbf{y(s}_{2}\mathbf{(}\xi
623: _{2},t_{2}\mathbf{)-s}_{1}\mathbf{(}\xi _{1},t_{1}\mathbf{))}\right]
624: \right\rangle d^{3}\mathbf{y.} \label{delta_fusion}
625: \end{equation}
626: We stress again that the label variables $\xi _{2}$ and $\xi _{1}$ belongs
627: two different loops. Let us introduce initial points $\mathbf{s}_{1}\mathbf{(%
628: }0\mathbf{)}$ and $\mathbf{s}_{2}\mathbf{(}0\mathbf{)}$ and rewrite (\ref
629: {delta_fusion}) in the following form:
630: \begin{equation*}
631: \frac{1}{(2\pi )^{3}}\int \exp \left[ -i\mathbf{y(s}_{2}\mathbf{(}0\mathbf{%
632: )-s}_{1}\mathbf{(}0\mathbf{))}\right] \left\langle \exp \left[ i\mathbf{y(s}%
633: _{2}\mathbf{(}\xi _{2}\mathbf{)-s}_{2}\mathbf{(}0\mathbf{))}\right] \exp %
634: \left[ -i\mathbf{y(s}_{1}\mathbf{(}\xi _{2}\mathbf{)-s}_{1}\mathbf{(}0%
635: \mathbf{))}\right] \right\rangle d^{3}\mathbf{y.}
636: \end{equation*}
637: Identifying further \ the ''initial'' positions $\mathbf{s}_{1}\mathbf{(}0%
638: \mathbf{),s}_{2}\mathbf{(}0\mathbf{)}$ with quantities $\mathbf{R}_{1},%
639: \mathbf{R}_{2}$ in formula (\ref{A_def}) we rewrite it as
640: \begin{eqnarray}
641: A(l_{1},l_{2},l) &=&\frac{1}{\text{VOLUME}}\frac{2}{3}\left| \mathbf{V}%
642: _{l}\right| \int \int d\mathbf{R}_{1}d\mathbf{R}_{2}\int \int d\xi _{1}d\xi
643: _{2} \label{A_inter} \\
644: &&\frac{1}{(2\pi )^{3}}\int \exp \left[ -i\mathbf{y(\mathbf{R}_{2}-\mathbf{R}%
645: _{1})}\right] \left\langle \exp \left[ i\mathbf{y(s}_{2}\mathbf{(}\xi _{2}%
646: \mathbf{)-s}_{2}\mathbf{(}0\mathbf{))}\right] \exp \left[ -i\mathbf{y(s}_{1}%
647: \mathbf{(}\xi _{2}\mathbf{)-s}_{1}\mathbf{(}0\mathbf{))}\right]
648: \right\rangle d^{3}\mathbf{y}. \notag
649: \end{eqnarray}
650: Let us introduce variables $\mathbf{R}_{1}-\mathbf{R}_{2},(\mathbf{R}_{1}+%
651: \mathbf{R}_{2})/2.$ \ Integration over $\mathbf{\mathbf{R}}_{2}\mathbf{-%
652: \mathbf{R}}_{1}$ gives $\mathbf{\delta }\mathbf{(y)}$, integration over $(%
653: \mathbf{R}_{1}+\mathbf{R}_{2})/2$ gives the total volume of
654: system. Further, integration over $\mathbf{y}$ gives unity, and
655: integration over $\xi _{1},\xi _{2}$ gives the production
656: $l_{1}l_{2}$. \ Thus we obtain the remarkable result, that for
657: noninteracting loops quantity $A(l_{1},l_{2},l)$ \ responsible for
658: merging of loops does not depend on statistics of individual loop
659: at all and is equal
660: \begin{equation}
661: A(l_{1},l_{2},l)=b_{m}V_{l}l_{1}l_{2} \label{A_final}
662: \end{equation}
663: Here $b_{m}=1/3.$ As earlier we introduced additional factor $1/2$ to avoid the over-counting of the reconnection events.
664:
665: Results (\ref{A_final}) and (\ref{B_final}) \ (with not well determined
666: factors $b_{s}$ and $b_{m}$ ) were also obtained in paper \cite{Copeland98}.
667: Authors used some qualitative picture of moving and colliding elements of
668: lines. This fact confirms the validity of approach made in our work,
669: which allows to use it for more complicated (in comparison with random walk)
670: cases.
671:
672: \section{Conclusion}
673:
674: \bigskip In the paper we discussed a role of kinetics of the merging and
675: breaking down vortex loops in the whole vortex tangle dynamic. Following the
676: work \cite{Copeland98} we introduced the reconnection rates $%
677: A(l_{1},l_{2},l) $ and $B(l,l_{1},l_{2})$. Quantity $A$ is a number (per
678: unit of time and per \ unit of volume) of the fusion of two loops of lengths
679: $l_{1}$and $l_{2}$ with formation of the single loop of length $\
680: l=l_{1}+l_{2}$. Quantity $B(l,l_{1},l_{2})$ describes rate of the breakdown
681: of \ the loop with the length $l$ into two daughter loops with lengths $\
682: l_{1}$ and $l_{2}$. We developed mathematical formalism, which enables us to
683: calculate $A(l_{1},l_{2},l)$ and $B(l,l_{1},l_{2}).$ Briefly, these
684: quantities can be found as an average number of zeroes of vector $\mathbf{S}%
685: _{b}(\xi _{2},\xi _{1},t)$ connecting points of $\xi _{2}$ and $\ \xi _{1}$
686: at moment of time $t$. The averaging procedure should be fulfilled
687: by using statistics of individual loops, expressed e.g. by the probability
688: distribution functional $\mathcal{P}(\left\{ \mathbf{s}(\xi ,t)\right\} ).%
689: \mathcal{\ }$\ We applied this formalism to evaluate the reconnection rates $%
690: A(l_{1},l_{2},l)$ and $B(l,l_{1},l_{2})$ in case when vortex loop
691: is semifractal object, which is smooth for small separations along
692: the line and has a random work structure for distant (along line)
693: points.\\
694: Relations (\ref{A_def}) and (\ref{B_def})are the key
695: results of the paper. They are formal, and do not depend on
696: statistic of loop. They can be applied for any physical situation
697: provided that we are able to get statistics from
698: the dynamical problem. In that sense the application of the results (%
699: \ref{A_def}) and (\ref{B_def}) for Gaussian case is just
700: illustration how to use it in concrete sense. One of serious
701: problem in this illustration is that in Gaussian approach the low
702: scale was restricted by mean radius of curvature. In real vortex
703: tangle $\xi _{0}$ in superfluid turbulent HeII this scale exceeds
704: greatly the scales at \ which the real dissipative mechanisms
705: occur. Thus, the study of what would happen for scales smaller of
706: mean radius of curvature is on of actual problems. Nevertheless
707: results expressed by relations (\ref{A_final}) and (\ref{B_final})
708: are significant, they allows to analyze the kinetic equation
709: (\ref{kinetic equation}), the attempts were made in papers
710: \cite{Copeland98}, \cite{Sreer99}. We also will apply these result
711: in the forthcoming paper for study of the kinetic
712: equation.\newline
713:
714: I am grateful to participants of the workshop ''Superfluidity
715: under Rotation'' (Manchester, 2005) for useful discussion of the
716: results exposed above. This work was partially supported by grant
717: N 03-02-16179 of the Russian Foundation of Fundamental Research.
718:
719: \begin{thebibliography}{99}
720: \bibitem{Don_book} R.J.Donnelly, {\it Quantized Vortices in HeliumII},
721: (Cambridge University Press, 1991).
722:
723: \bibitem{Feynman} R.P.Feynman, in {\it Progress in Low Temperature
724: Physics}, edited by C.J.Gorter (North-Holland, Ameterdam, 1955),
725: Vol.I, p.17.
726:
727: \bibitem{Vinen} W.F.Vinen, Proc. R. Soc. London A {\bf 242}, 493(1957).
728:
729: \bibitem{NF} S.K.Nemirovskii and W.Fiszdon, Rev. Mod. Phys. {\bf 67},
730: 37 (1995).
731:
732: \bibitem{Koplik} J.Koplik and H.Levine, Phys. Rev. Lett. {\bf 71},
733: 1375 (1993).
734:
735: \bibitem{Schwarz88} K.W.Schwarz, Phys. Rev. B {\bf 38}, 2398 (1988).
736:
737: \bibitem{Barenghi} C.F.Barenghi, D.C.Samuels, G.H.Bauer and
738: R.J.Donnelly, Phys. Fluids {\bf 9}, 2631(1997).
739:
740: \bibitem{Aarts} R.G.K.M.Aarts and A.T.A.M.de Waele, Phys. Rev. B {\bf
741: 50}, 10069 (1994).
742:
743: \bibitem{Tsubota00} M. Tsubota, T. Araki and S. K. Nemirovskii, Phys.
744: Rev. B {\bf 62}, 11751 (2000).
745:
746: \bibitem{Barenghi2004} C.F.Barenghi and D.C.Samuels, Journal of Low
747: Temperature Physics, Vol. 136, Nos. 5/6, (2004).
748:
749: \bibitem{Copeland98} E.J.Copeland, T.W.B.Kibble and D.A.Steer, Phys.Rev. D,
750: textbf\{58, \} 043508, (1998).
751:
752: \bibitem{Nemirovskii_97_1} S.K. Nemirovskii, Phys. Rev. B {\bf 57},
753: 5792 (1997).
754:
755: \bibitem{Sreer99}Joao Magueijoa, Havard Sandvik, and Dani`ele A.
756: Steer, http://lanl.arxiv.org/abs/astro-ph/9905363.
757:
758: \end{thebibliography}
759:
760: \end{document}
761: