0804.1357/v.tex
1: %\documentclass[12pt]{revtex4}
2: 
3: \documentclass[12pt,showpacs,nofootinbib]{revtex4}
4: %\documentclass[12pt]{elsart}
5: 
6: 
7: 
8: 
9: 
10: \usepackage{dcolumn,amsmath}
11: \usepackage{amsfonts}
12: \usepackage{amsthm}
13: \usepackage{mathrsfs}
14: \usepackage{amssymb}
15: \usepackage{epsfig}
16: %\usepackage{showlabels}
17: \usepackage{psfrag}
18: \usepackage{array}
19: \usepackage{graphics,graphicx,float,mfpic}
20: \usepackage{color}
21: 
22: 
23: \newcommand{\be}{\begin{equation}}
24: \newcommand{\ee}{\end{equation}}
25: \newcommand{\bea}{\begin{eqnarray}}
26: \newcommand{\eea}{\end{eqnarray}}
27: \newcommand{\X}{X}
28: \newcommand{\D}{\hat{{\ccal D}}}
29: \newcommand{\ssl}{\sl}
30: \newcommand{\iit}{\it}
31: \newcommand{\bbf}{\bf}
32: \newcommand{\ccal}{\mathcal}
33: \newcommand{\bx}{{\bf x}}
34: \newcommand{\Q}{N}
35: \newcommand{\XPEH}{\it}
36: \newcommand{\mcal}{\mathcal}
37: \newcommand{\ssigma}{{\mcal N}}
38: \newcommand{\RRR}{{R}}
39: \newcommand{\ccc}{{a}}
40: %\newcommand{\n}{{\phi}}
41: %\newcommand{\nn}{\vec{\n}}
42: \newcommand{\n}{{n}}
43: \newcommand{\nn}{{\bf \n}}
44: %\renewcommand{\baselinestretch}{1.6}
45: 
46: 
47: 
48: \def\be{\begin{equation}}
49: \def\ee{\end{equation}}
50: \def\bea{\begin{eqnarray}}
51: \def\eea{\end{eqnarray}}
52: 
53: 
54: \renewcommand{\theequation}{\arabic{section}.\arabic{equation}}
55: 
56: \begin{document}
57: %\bibliographystyle{plain}
58: %\begin{frontmatter}
59: 
60: 
61: 
62: 
63: 
64: %\title{Existence of stationary, 
65: %non-radiating ring solitons  in field theory: 
66: %knots and vortons
67: %\vspace{1 cm}
68: %}
69: 
70: \title{Stationary 
71: ring solitons  in field theory -- \\
72: knots and vortons
73: \vspace{1 cm}
74: }
75: 
76: 
77: 
78: 
79: \author{Eugen Radu%
80: \footnote{{{\tt Eugen.Radu@lmpt.univ-tours.fr}}}}
81: 
82: %and 
83: 
84: \author{Mikhail S. Volkov%
85: \footnote{{\tt Michael.Volkov@lmpt.univ-tours.fr}}}
86: 
87: 
88: 
89:  \affiliation{ \vspace{0.1 cm}
90: {Laboratoire de Math\'{e}matiques et Physique Th\'{e}orique
91: CNRS-UMR 6083, \\ Universit\'{e} de Tours,
92: Parc de Grandmont, 37200 Tours, FRANCE\\
93: %\footnote{{{\tt Eugen.Radu@lmpt.univ-tours.fr}}}
94: %\footnote{{\tt Michael.Volkov@lmpt.univ-tours.fr}}
95: }
96: \vspace{1 cm}
97: }
98: 
99: 
100: \begin{abstract}
101: We review the current status of the problem of constructing 
102: classical field theory solutions describing {stationary}  
103: vortex rings in Minkowski space in $3+1$
104: dimensions. 
105: We describe the known up to date solutions of this type, such as the  
106: static knot solitons stabilized by the topological Hopf charge, 
107: the  attempts to gauge them, the anomalous solitons
108: stabilized by the Chern-Simons number, as well as  
109: the non-Abelian monopole and sphaleron rings. Passing to the 
110: rotating solutions, we first discuss the conditions insuring that they
111: do not radiate, and then 
112: describe the spinning $Q$-balls, their twisted 
113: and gauged generalizations reported here for the first time, 
114: spinning skyrmions, and rotating monopole-antimonopole pairs. 
115: We then present the first explicit construction of global vortons as solutions
116: of the elliptic boundary value problem, which demonstrates their non-radiating 
117: character. 
118: Finally, we describe  the analogs of vortons  
119: in the Bose-Einstein condensates, analogs of spinning $Q$-balls
120: in the non-linear optics, and also moving vortex rings
121: in superfluid helium and in ferromagnetics. 
122: 
123: 
124: 
125: 
126: \end{abstract}
127: %\end{frontmatter}
128: 
129: %PACS numbers:~{11.10.Lm, 11.27.+d, 98.80.Cq}
130: \pacs{11.10.Lm, 11.27.+d, 98.80.Cq}
131: 
132: 
133: %Keywords: classical field theory, gauge fields, solitons, vortices
134: 
135: \maketitle
136: %\begin{flushright}
137: \hspace{0.7 cm}
138: to appear in {\it Physics Reports}
139: %\end{flushright}
140: 
141: \vspace{0.1 cm}
142: \newpage 
143: 
144: \tableofcontents
145: 
146: %\newpage
147: 
148: \section{Introduction}
149:  
150: 
151: 
152: 
153: Stationary vortex loops are often discussed in the literature in various contexts
154: ranging from the models of condensed matter physics 
155: \cite{Babaev:2001jt},
156: \cite{Babaev:2001zy},
157: \cite{Battye:2001ec},
158: \cite{Cooper},
159: \cite{Metlitski:2003gj},
160: \cite{Protogenov:2002bt}, 
161: \cite{Ruostekoski:2001fc},
162: \cite{Ruostekoski:2003qx},
163: \cite{Savage:2003hh}, 
164:  to high energy physics and cosmology
165: \cite{Brandenberger:1996zp},
166: \cite{Davis:1988ip},
167: \cite{Davis:1988jp},
168: \cite{Davis:1988ij},
169: \cite{Vilenkin},
170: \cite{Witten:1984eb}.
171: Such objects could be quite interesting physically and might be  
172: responsible for a plenty of important physical phenomena,
173: starting from the structure of quantum superfluids \cite{Donnelly} and 
174: Bose-Einstein condensed alkali gases \cite{Leggett} 
175: to the baryon asymmetry of the Universe, dark matter,
176: and the galaxy formation \cite{Vilenkin}. These intriguing 
177: physical aspects of vortex loops occupy therefore most of the discussions 
178: in the literature, while much less attention is usually payed 
179: to their mathematical existence. 
180: In most considerations the existence of such loops is discussed 
181: only qualitatively, using plausibility arguments, and the problem of constructing
182: the corresponding field theory solutions is rarely addressed, such that almost 
183: no solutions of this type are explicitly known.
184: 
185: The physical arguments usually invoked to justify the loop existence 
186: arise within  the effective macroscopic description 
187: of vortices \cite{Abrikosov}, \cite{NO}. Viewed from large distances,
188: their internal structure can be neglected and the vortices can be effectively 
189: described as elastic thin ropes or, if they  carry currents,
190: as thin wires 
191: \cite{Carter:1989dp},
192: \cite{Carter:1989xk},
193: \cite{Carter:1990sm},
194: \cite{Carter:1994hn},
195: \cite{Martins:1998gb}. 
196: This suggests the following  engineering 
197: procedure: to `cut out' a finite piece of the vortex, then `twist' it several times
198: and finally `glue' its ends. 
199: The resulting loop will be stabilized against contraction by the potential 
200: energy of the twist deformation, that is by the positive pressure contributed by the 
201: current and stresses 
202: \cite{Copeland:1987th},
203: \cite{Copeland:1987yv},
204: \cite{Haws:1988ax},
205: \cite{Niemi:2000ny}. 
206: Equally, instead of twisting the rope,
207: one can first `make' a loop and then  
208: `spin it up' thus giving to it an angular momentum to 
209: stabilize it against shrinking 
210: \cite{Davis:1988jq},
211: \cite{Davis:1988ij}. 
212: Depending on whether they are spinning or not, 
213: such loops are often called in the literature vortons and knots (also springs), 
214: respectively. 
215: 
216: 
217: 
218: 
219: These macroscopic arguments are suggestive. However, they
220: cannot guarantee the existence of loops as stationary field theory
221: objects,  since they do not take into account all the field degrees of freedom that 
222: could be essential. Speaking more rigorously, one can prepare twisted or spinning
223: loop configurations as the initial data for the fields. 
224: However, nothing guarantees that these data will evolve to 
225: non-trivial equilibrium field theory objects, since, up to few exceptions, the 
226: typical field theory
227: models under consideration do not have a topological energy bound for static loop 
228: solitons. For spinning loops the situation can be better, 
229: since the angular momentum 
230: can provide an additional stabilization of the system. 
231: However, nothing excludes the possibility 
232: of radiative  energy leakages from
233: spinning loops. Since there is typically a current
234: circulating along them, which means an accelerated motion of charges, 
235: it seems plausible that spinning loops should radiate, in which case
236: they would not 
237: be {\it stationary} but at best only quasistationary. 
238: In fact, loop formation is generically observed in 
239: dynamical vortex network simulations (although for currentless vortices),
240: and  these loops indeed 
241: radiate rapidly away all their energy 
242: (see $e.g.$ \cite{Bevis:2006mj}).
243: 
244: 
245: The lack of explicit solutions renders the situation even more controversial.
246: In fact, although there are explicit examples of loop solitons 
247: in {\it global} field theory, which means containing only scalar fields with global 
248: internal symmetries, almost nothing is known about such solutions in 
249: {\it gauge} field theory. Vortons, for example, were initially proposed 
250: more than 20 years ago in the context of the local U(1)$\times$U(1) 
251: theory of superconducting cosmic strings of  Witten 
252:  \cite{Witten:1984eb}. However, not a single 
field theory solution of this type has been obtained up to now.
253: Search for static knot solitons in the Ginzburg-Landau type gauge field theory models
254: has already 30 years of history, the result being always negative. 
255: All this  casts some doubts on 
256: the existence of {stationary}, non-radiating vortex loops in physically interesting 
257: gauge field theory models. 
258: However, rigorous no-go arguments are not known either. 
259: As a result, there are no solid arguments neither 
260: for nor against the existence of such solutions.
261: 
262: Interestingly, {\it exact} solutions describing ring-type objects are known
263: in curved space. These are black rings in the multidimensional generalizations 
264: of  General Relativity -- spinning toroidal black holes 
265: (see \cite{Emparan} for a review). It seems therefore that Einstein's
266: field equations, notoriously known for their complexity,
267: are easier to solve than the non-linear equations describing
268: loops made of interacting gauge and scalar fields in Minkowski space. 
269: 
270: 
271: It should be emphasized that the problem here is not related to the dynamical 
272: stability of these loops. In order to be stable or unstable they should first of 
273: all exist as stationary field theory solutions. The problem is related to the very 
274: existence of such solutions. One should be able to decide whether they exist or not,
275: which is a matter of principle, and this is the main issue that we address  in this paper.
276: Therefore, when talking about loops stabilized by some forces, 
277: we shall mean the force balance that
278: makes possible their existence,
279: and not the stability of these loops with respect to all 
280: dynamical perturbations. If they exist, their dynamical stability should be 
281: analyzed separately, but we shall call {\it solitons} all  localized, globally regular, 
282: finite energy field theory solutions, irrespectively of their stability properties. 
283: It should also be stressed that we insist on the 
284: {\it stationarity} condition for the solitons, which 
285: implies the absence of radiation.  
286: Quasistationary loops which radiate slowly and 
287: live long enough could also be physically
288: interesting, but we are only interested in loops which could live infinitely
289: longtime, at least in classical theory. 
290: 
291: 
292: 
293: Trying to clarify the situation, we review in what follows 
294: the known field theory solutions  in Minkowski space
295: in $3+1$ dimensions describing stationary ring objects. 
296: We shall divide them in two groups, depending on whether they do have 
297: or do not have an angular momentum.  Solutions without angular momentum are 
298: typically static, that is they do not depend on time and also do not typically 
299: have an electric field. 
300: We start by describing  the static knot solitons stabilized by the Hopf charge
301: in a global field theory model. Then we 
302: review the attempts to generalize these solutions within 
303: gauge field theory, with the conclusion that some additional constraints on the
304: gauge field are necessary, since fixing only the Hopf charge does not seem to be
305: sufficient  to
306: stabilize the system in this case. We then consider two known examples of 
307: static ring solitons in gauge field theory: the anomalous solitons stabilized 
308: by the Chern-Simons number, and the non-Abelian monopole and sphaleron rings. 
309: 
310: 
311: Passing to the rotating solutions, we first of all discuss 
312: the issue  of how the presence 
313: of an angular momentum, associated to some internal 
314: motions in the system, can be 
315: reconciled with the absence of radiation which is normally 
316: generated by these motions. 
317: One possibility for this is to consider
318: time-independent solitons in theories with local internal symmetries.
319: They will not radiate, but it can be shown  
320: for a number of important field theory models that the angular momentum
321: vanishes in this case.  Another possibility 
322: arises in systems with global symmetries,
323: where one can consider   {\it non-manifestly} stationary 
324: and axisymmetric fields containing 
325: spinning phases. Such fields could have a non-zero 
326: angular momentum, but the absence of radiation is not automatically
327: guaranteed in this case. The general conclusion is that the 
328: existence of non-radiating spinning solitons, although not impossible,  
329: seems to be rather restricted. Such solutions seem  to exist only in some quite special
330: field theory models, while generic spinning field systems should radiate. 
331: 
332: Nevertheless, non-radiating spinning 
333: solitons in Minkowski space
334: in $3+1$ dimensions exist, and we review below all known examples: these are  the 
335: spinning $Q$-balls, their twisted and gauged generalizations, 
336: spinning Skyrmions and rotating monopole-antimonopole pairs.
337: In addition, there are also vortons, and below we present for the first time  
338: numerical solutions of elliptic equations describing vortons in the 
339: global field theory limit. 
340: Although global vortons have 
341: been studied before by different methods,
342: our construction shows that they indeed exist as stationary, 
343: non-radiating field theory objects. 
344: 
345: We also discuss  stationary ring solitons in non-relativistic physics. 
346: Surprisingly, it turns out that the relativistic vortons can be mapped
347: to the `skyrmion' solutions of the Gross-Pitaevskii equation in the theory 
348: of Bose-Einstein condensation.  In addition, it turns out that $Q$-balls 
349: can describe light pulses in media with non-linear refraction -- `light bullets'.
350: Finally, for the sake of completeness, we discuss 
351: also the moving vortex rings stabilized by the Magnus force
352: in continuous media theories,  such as in the 
353: superfluid helium and in  ferromagnetics.   
354: 
355: In our numerical calculations  we use an elliptic PDE solver 
356: with which we have managed to 
357: reproduce most of the solutions we describe, as well as obtain a number of 
358: new results presented below for the first time.  
359: The latter include the explicit vorton solutions, 
360: spinning twisted $Q$-balls, spinning 
361: gauged $Q$-balls, 
362: as well as the `Saturn', `hoop' and bi-ring solutions for the interacting $Q$-balls.   
363: We put the main emphasise on describing how the solutions are constructed
364: and not to  their physical applications, so that 
365: our approach is just the opposite and therefore  complementary
366: to the one generally adopted 
367: in the existing literature. 
368: As a result, we outline the current status of the ring soliton existence problem --
369: within the numerical approach. Giving mathematically rigorous existence proofs 
370: is an issue  that should be analyzed separately.
371: It seems that our global vorton solutions could be generalized 
372: in the context of gauge field theory. 
373: The natural problem to attack 
374: would then be to obtain vortons in the electroweak sector 
375: of Standard Model. 
376: 
377: 
378: 
379: In this text the signature of the Minkowski spacetime metric $g_{\mu\nu}$ 
380: is chosen to be $(+,-,-,-)$, 
381: the spacetime coordinates are denoted by $x^\mu=(x^0,x^k)\equiv (t,\bx)$ 
382: with $k=1,2,3$. All physical quantities discussed below, including 
383: fields, coordinates, coupling constants and conserved quantities are
384: dimensionless.  
385: 
386: \section{Knot solitons}
387: 
388: Let us first consider solutions with zero angular momentum stabilized by their 
389: intrinsic deformations. 
390: We shall start by discussing  the famous example of 
391: static knotted solitons in a
392: non-linear sigma model. Since this  is the best known and also in some 
393: sense canonical example of knot solitons, we shall describe it in some detail.  
394: We shall then review the status of  gauge field theory 
395: generalizations of these solutions.  
396: 
397: \subsection{Faddeev-Skyrme model}
398: 
399: More than 30 years ago Faddeev introduced a field theory consisting of 
400: a non-linear O(3) sigma model augmented by adding a Skyrme-type term
401: \cite{Faddeev:1975},
402: \cite{Faddeev:1976pg}. 
403: This theory can also be obtained by a consistent truncation of the O(4) Skyrme model
404: (see Sec.\ref{versus} below). 
405: Its dynamical variables are three
406: scalar fields $\nn\equiv \n^a=(\n^1,\n^2,\n^3)$
407:  constraint by the condition 
408: $$
409: \nn\cdot\nn=\sum_{a=1}^3 \n^a \n^a=1,
410: $$
411: so that they span a two-sphere $S^2$. 
412: The Lagrangian density of the theory is 
413: \be                                                              \label{action0} 
414: {\mcal L}[\nn]=\frac{1}{32\pi^2}\left(\partial_\mu \nn\cdot\partial^\mu \nn
415: -{\ccal F}_{\mu\nu}{\ccal F}^{\mu\nu}\right)
416: \ee
417: where 
418: \be                                              \label{F}
419: {\ccal F}_{\mu\nu}=\frac12\,\epsilon_{abc}\n^a\partial_\mu \n^b\partial_\nu \n^c\equiv
420: \frac12\,\nn\cdot(\partial_\mu\nn\times\partial_\nu\nn). 
421: \ee
422: The Lagrangian field equations read 
423: \be                              \label{eqs-Hopf}
424: \partial_\mu\partial^\mu\nn+\partial_\mu{\ccal F}^{\mu\nu}(\nn\times\partial_\nu\nn)
425: %+2{\ccal F}^{\mu\nu}(\partial_\mu \nn\times \partial_\mu\nn)\,
426: =
427: (\nn \cdot\partial_\mu\partial^\mu\nn %+2{\ccal F}_{\mu\nu}{\ccal F}^{\mu\nu}
428: )\,\nn\,.
429: \ee  
430: In the static limit the 
431: energy of the system is  
432: \be                                                               \label{Faddeev}
433: E[\nn]=\left.\left.\frac{1}{32\pi^2}\int_{\mathbb{R}^3}\right((\partial_k \nn)^2
434: +({\ccal F}_{ik})^2\right)
435: d^3 \bx\,\equiv E_2+E_4. 
436: \ee
437: Under scale transformations, $\bx\to\Lambda\bx$,
438: one has $E_2\to\Lambda E_2$ and $E_4\to E_4/\Lambda$.
439: The energy will therefore be stationary for $\Lambda=1$ if only the 
440: virial relation holds,
441: \be
442: E_2=E_4.
443: \ee
444: This shows that the four derivative term $E_4$ is necessary, since otherwise  
445: the virial relation would require
446: that $E_2=0$, thus ruling out all non-trivial static
447: solutions -- in agreement with the  Hobart-Derrick theorem
448: \cite{Hobart}, 
449: \cite{Derrick:1964ww}. 
450: 
451: Any static field $\nn(\bx)$ defines a map $\mathbb{R}^3\to {S}^2$.
452: Since for finite energy configurations $\nn$ should approach
453: a constant value for $|\bx|\to\infty$, all points at infinity of $\mathbb{R}^3$ 
454: map to one point on $S^2$. Using the global O(3)-symmetry of the theory one can choose
455: this point to be the north pole of the $S^2$, 
456: \be                        \label{n-inf}
457: \lim_{|\bx|\to\infty}\nn(\bx)=\nn_\infty=(0,0,1).
458: \ee
459: Notice that this condition leaves a residual O(2) symmetry of global rotations around 
460: the third axis in the internal space,
461: \be                                       \label{residual}
462: n^1+in^2\to(n^1+in^2)e^{i\alpha},~~~~~~~~~~ n^3\to n^3.
463: \ee 
464: The position of the field configuration described by $\nn(\bx)$
465:  can now be defined as the set of points where the 
466: field is as far as possible from the vacuum value, that is the preimage
467: of the point $-\nn_\infty$ antipodal to the vacuum $+\nn_\infty$. This preimage 
468: forms a closed
469: loop (or collection of loops) called position curve. 
470: Solitons in the theory can therefore be viewed as string-like objects,
471: stabilized by their topological charge to be defined below.
472: 
473: At the intuitive level, these solitons can be viewed as closed loops made of 
474: twisted vortices.  
475:  Specifically, the theory 
476: admits solutions describing straight vortices that can be 
477: parametrized in cylindrical coordinates 
478: $\{\rho,z,\varphi\}$ as $n^3=\cos\Theta(\rho)$, 
479: $\n^1+i\n^2=\sin\Theta(\rho)\,e^{ipz+in\varphi}$,  where $n\in\mathbb{Z}$
480: is the vortex winding number \cite{Kundu}. 
481: The phase $pz+n\varphi$ thus changes both along and around the vortex and 
482: the vector $\nn$  rotates   around the third internal direction as one moves
483: along the vortex, which can be interpreted as {\it twisting}
484: of the vortex. 
485: It seems plausible that a loop made of a piece of length $L$ of such a twisted 
486: vortex, where $pL=2\pi m$, 
487: could be stabilized by the potential energy of the twist deformation. 
488: For such a twisted loop the phase increases by $2\pi n$ after a revolution around
489: the vortex core and by $2\pi m$ as one travels along the loop, where $m\in\mathbb{Z}$
490: is the number of twists. If the loop is homeomorphic to a circle, then the product
491: $nm$ gives the value of its topological invariant: the Hopf charge. 
492: 
493: 
494: 
495: 
496: 
497: \subsection{Hopf charge}
498: 
499: The condition \eqref{n-inf} allows one to view the infinity of $\mathbb{R}^3$ 
500:  as one point, thus effectively replacing 
501: $\mathbb{R}^3$ by its one-point compactification
502: ${S}^3$. Any smooth field configuration 
503: can therefore be viewed as a map
504: \be
505: \nn(\bx): {S}^3\to {S}^2.
506: \ee 
507: Any such map can be characterized by the topological
508: charge $\Q[\nn]\in \pi_3(S^2)=\mathbb{Z}$ known as the Hopf invariant.
509:  This invariant has a simple
510: interpretation. The preimage of a generic point on the target  $S^2$ is a 
511: closed loop. If a field has Hopf number $\Q$ then 
512: the two loops consisting of preimages of two generic distinct points on  $S^2$ 
513: will be linked exactly $\Q$ times (see Fig.\ref{FigHopf}). 
514: 
515: \begin{figure}[h]
516: \hbox to\linewidth{\hss%
517:   \resizebox{8cm}{3cm}{\includegraphics{Hopf.eps}}%
518: %  \resizebox{8cm}{6cm}{\includegraphics{FIG8.eps}}%
519: \hss}
520: \caption{\small 
521: Preimages %$(n^a)^{-1}(A)$ and $(n^a)^{-1}(B)$
522: of any two points $\nn_1$ and $\nn_2$ on the target space $S^2$ 
523: are two loops. The number of mutual linking of these two loops
524: is the Hopf charge $\Q[\nn]$ of the map $\nn(\bx)$. Here the case of 
525: $\Q=1$ is schematically shown. 
526: }
527: \label{FigHopf}
528: \end{figure}
529: 
530: Although there is no local formula for $\Q[\nn]$ in terms of $\nn$, one can give
531: a non-local expression as follows. 
532: The 2-form ${\ccal F}=\frac12\,{\ccal F}_{ik}dx^i\wedge dx^k$ defined by Eq.\eqref{F} is closed, 
533: $d{\ccal F}=0$, 
534: and since the second cohomology group of $S^3$ is trivial, $H(S^3)=0$, 
535: there globally exists 
536:  a vector potential ${\ccal A}={\ccal A}_k dx^k$ such that 
537: ${\ccal F}=d{\ccal A}$. 
538: The Hopf index  
539: can then be expressed as
540: \be                                                              \label{Q} 
541: \Q[\nn]=\frac{1}{8\pi^2}\int \epsilon_{ijk}{\ccal A}_i{\ccal F}_{jk}\, d^3\bx.
542: %\frac{1}{8\pi^2}\int \epsilon_{ijk}{\ccal A}_i {\ccal F}_{jk}\, d^3\bx\,.
543: \ee
544: For any smooth, finite energy field configuration $\nn(\bx)$ this integral is 
545: integer-valued \cite{Lin}. 
546: 
547: It can be shown that the maximal symmetry of $\nn(\bx)$  compatible with 
548: a non-vanishing Hopf charge is O(2) \cite{Kundu}. 
549: It follows that spherically symmetric fields are topologically trivial.  
550: However, axially symmetric fields can have any value of the Hopf charge.  
551: Using cylindrical coordinates 
552: such fields can be parametrized as \cite{Kundu} 
553: \be                                              \label{Hopf:axial}
554: \n^1+i\n^2= e^{i(m\varphi-n\psi)}\sin\Theta,~~~~~n^3=\cos\Theta
555: \ee
556: where $n,m\in\mathbb{Z}$ and $\Theta,\psi$ are functions of $\rho,z$. 
557: Since  $\nn\to\nn_\infty$ asymptotically, 
558: $\Theta$ should vanish for $r=\sqrt{\rho^2+z^2}\to\infty$. 
559: The regularity at the $z$-axis requires for $m\neq 0$ that $\Theta$ 
560: should vanish also there. 
561: As a result, one has $\nn=\nn_\infty$ both at the $z$-axis and at infinity, 
562: that is at the contour  
563: $C$ shown in Fig.\ref{FigS}.
564: Next,  one assumes that $\nn=-\nn_\infty$ on a circle $S$ 
565: around the $z$-axis which is linked to $C$ as shown in Fig.\ref{FigS} 
566: (more generally, one can have $\nn=-\nn_\infty$ on several circles
567: around the $z$-axis).   
568: The phase function $\psi$ 
569: is supposed to increase by $2\pi$ after one revolution along $C$.
570: Since  $\cos\Theta$ interpolates
571: between $-1$ and $1$ 
572: on every trajectory from $S$ to $C$, it follows that  
573: surfaces of constant $\Theta$ are homeomorphic to tori. 
574:  
575: \begin{figure}[ht]
576: \hbox to\linewidth{\hss%
577: 	\resizebox{8cm}{7cm}{\includegraphics{S1.eps}}
578: \hspace{5mm}%
579: \hss}
580: 	\caption{Knot topology: the complex phase of the field in Eq.\eqref{Hopf:axial}
581: winds along two orthogonal directions: along $S$  and along the
582:  contour $C$ consisting of the $z$-axis and a semi-circle 
583: whose radius expands to infinity. A similar winding of phases 
584: is found for other systems to be
585: discussed below: for vortons, skyrmions, and twisted $Q$-balls.}
586: \label{FigS}
587: \end{figure}
588: 
589: 
590: The preimage of the point $-\nn_\infty$ consists of $m$ copies of the circle $S$.  
591: The preimage of $\nn_\infty$ consists of $n$ copies of the contour $C$.
592: These two preimages   
593:  are therefore linked $mn$ times.  
594: 
595: One can also compute 
596: ${\ccal F}=d{\ccal A}$ according to \eqref{F}, from where one finds 
597: \be                                        \label{Hopf:axial1}
598: {\ccal A}=n\cos^2\frac{\Theta}{2}\,d\psi+
599: m\sin^2\frac{\Theta}{2}\,d\varphi
600: \ee
601:  so that  
602:  ${\ccal A}\wedge{\ccal F}=nm\cos^2\frac{\Theta}{2}\sin\Theta\, 
603: d\psi\wedge d\Theta\wedge d\varphi$.
604: Inserting this to \eqref{Q} gives the Hopf charge
605: \be						\label{Hopf-axial}
606: \Q[\nn]=mn.
607: \ee
608: Following Ref.\cite{Sutcliffe:2007ui}, we shall call the fields given by the ansatz 
609: \eqref{Hopf:axial}  ${\bf A}_{mn}$. 
610: %(in \cite{Sutcliffe:2007ui} the roles of $n$ and $m$ are interchanged).   
611: 
612: 
613: %These results  will be the same if 
614: %the preimage of the point $-\nn_\infty$ (the position curve) 
615: %is not one circle covered $m$
616: %times but $m$ different circles around the $z$-axis. In this case the field would be 
617: %given by Eq.\eqref{Hopf:axial} with $m=1$ but with $\cos\Theta$ 
618: 
619: Let us consider two explicit examples of the ${\bf A}_{mn}$ field. Let us introduce 
620: toroidal coordinates  $\{u,v,\varphi\}$ such that 
621: \be                                   \label{toroidal}
622: \rho=(R/\tau)\,{\sinh u},~~~~
623: z=(R/\tau)\,{\sin v}, 
624: \ee
625: where $\tau=\cosh u-\cos v$ with 
626: $u\in[0,\infty)$, $v=[0,2\pi)$.
627: The correct boundary conditions
628: for the field 
629: will then be achieved by choosing in the ansatz \eqref{Hopf:axial} 
630: \be                                   \label{toroidal1}
631: \Theta=\Theta(u),~~~~\psi=v,
632: \ee
633: where $\Theta(0)=0$, $\Theta(\infty)=\pi$.  
634: 
635: Another useful parametrization is achieved 
636: by expressing $\n^a$ in terms of its complex stereographic projection  coordinate
637: \be                                   \label{W-Hopf}
638: W=\frac{\n^1+i\n^2}{1+\n^3}.
639: \ee 
640: The values $W=0,\infty$ correspond to the vacuum and to the position curve of the soliton, 
641: respectively.  Passing to spherical coordinates $\{r,\vartheta,\varphi\}$ one introduces 
642: \be                                 \label{W-Hopf1}
643: \phi=\cos \chi(r)+i\sin\chi(r)\cos\vartheta,~~~~\sigma=\sin\chi(r)\sin\vartheta e^{i\varphi},
644: \ee
645:  such that $|\phi|^2+|\chi|^2=1$, where $\chi(0)=\pi$, $\chi(\infty)=0$. 
646: A particular case of the field \eqref{Hopf:axial}
647: is then obtained by setting
648: \be                                    \label{WW-Hopf}
649: W=\frac{(\sigma)^m}{(\phi)^n}.
650: \ee
651: 
652: \subsection{Topological bound and the $\Q=1,2$ hopfions}
653: 
654: The following inequality for 
655: the energy \eqref{Faddeev} and Hopf charge \eqref{Q} has been 
656: established by Vakulenko and Kapitanski 
657: \cite{Vakulenko:1979uw}, 
658: \be                                         \label{Hopf} 
659: {E[\nn]\geq c|\Q[\nn]|^{3/4}},
660: \ee
661: where $c=(3/16)^{3/8}$ \cite{Kundu}. Its derivation is non-trivial and proceeds
662: via considering a sequence of Sobolev inequalities. It is worth noting that a fractional
663: power of the topological charge occurs in this topological bound, whose value is 
664: optimal \cite{Lin}. On the other hand, it seems that the value of $c$ can be improved,
665: that is increased. Ward conjectures \cite{Ward:1998pj} that the bound holds for $c=1$,
666: which has not been proven but is compatible with all the data available. 
667: The existence of this bound shows that smooth fields
668: attaining it, if exist,  describe topologically stable solitons.
669: Constructing them implies  minimizing the energy \eqref{Faddeev} with 
670: fixed Hopf charge \eqref{Q}. Such minimum energy configurations 
671: are sometimes called in the literature Hopf solitons or hopfions, 
672: and we shall call them fundamental or ground state
673: hopfions if they have the least possible energy for a given $\Q$. 
674: 
675: Hopfions have been constructed for the first time 
676: for the lowest two values of the Hopf charge, $\Q=1,2$
677: by Gladikowski and Hellmund \cite{Gladikowski:1996mb} 
678: and almost simultaneously (although somewhat more qualitatively) 
679: by Faddeev and Niemi \cite{Faddeev:1996zj}.  
680: Gladikowski and Hellmund  
681: used the axial ansatz \eqref{Hopf:axial} expressed in toroidal 
682: coordinates, with $\Theta=\Theta(u,v)$ and $\psi=v+\psi_0(u,v)$,
683: where $\Theta(0,v)=0$, $\Theta(\infty,v)=\pi$. 
684: Assuming the functions $\Theta(u,v)$ and $\psi_0(u,v)$
685: to be periodic in $v$, they discretized the variables $u,v$ 
686: and numerically minimized the discretized expression for the energy with respect 
687: to the lattice cite values of $\Theta,\psi_0$. They found a smooth minimum energy  
688: configuration  of the ${\bf A}_{11}$ type for $\Q=1$, 
689: while for ${\Q}=2$ they obtained two solutions,  ${\bf A}_{21}$
690: and ${\bf A}_{12}$, the latter being more energetic than the former. 
691: %By construction, these $\Q=1,2$ hopfions are axially symmetric. 
692: %Clearly, for axially symmetric hopfions this position curve is the 
693: %circle $S$ shown in Fig.\ref{FigS}. 
694: 
695: We have verified the results of Gladikowski and Hellmund by integrating the field
696: equations \eqref{eqs-Hopf} in the static, axially symmetric sector. 
697: Using the axial ansatz
698: \eqref{Hopf:axial}, the azimuthal variable $\varphi$ decouples, and the equations reduce to 
699: two coupled PDE's for $\Theta(\rho,z)$, $\psi(\rho,z)$. Unfortunately, these equations 
700: are rather complicated and it is not possible to reduce them to ODE's by further separating
701: variables, via passing to toroidal coordinates, say. 
702: In fact, we are unaware of any attempts to solve these differential equations. 
703: We applied therefore our numerical method described below 
704: in Sec.\ref{sec-vortons} to integrate them, and  
705: we have succeeded in constructing the first two fundamental hopfions,
706: ${\bf A}_{11}$ and ${\bf A}_{21}$.  For the $\Q=1$
707: solution the energy density is maximal at the origin and the energy
708: density isosurfaces are squashed spheres, while for the $\Q=2$ solution
709: they have toroidal structure (see Fig.\ref{FN-en}). 
710: \begin{figure}[h]
711: \hbox to\linewidth{\hss%
712: \resizebox{5cm}{4cm}{\includegraphics{FN-m=1-E=50.eps}}%
713: \hspace{5mm}%
714: \resizebox{5cm}{4cm}{\includegraphics{FN-m=2-E=50.eps}}%
715: \hss}
716: \caption{\small The energy isosurfaces for the $\Q=1$ (left) and $\Q=2$ (right)
717: fundamental hopfions. 
718: }
719: \label{FN-en}
720: \end{figure}
721: For the solutions energies ${E}_{\Q}$ we obtained the 
722: values ${E}_1=1.22$,  ${E}_2=2.00$.
723: These features agree with the results of Gladikowski and Hellmund 
724: and with those of Refs.\cite{Battye:1998pe}, 
725: \cite{Battye:1998zn}, 
726: \cite{Sutcliffe:2007ui}, 
727: \cite{Hietarinta:1998kt}, 
728: \cite{Hietarinta:2000ci}.
729: In particular, the same values for the energy are quoted by Ward \cite{Ward:2000qj}.
730: 
731: Ward also proposes a simple analytic approximation of the solutions based on the 
732: rational ansatz formula \eqref{WW-Hopf} \cite{Ward:2000qj}. For $n=m=1$ this formula gives
733: \be                                       \label{Hopf3}
734: W=\frac{x+iy}{f(r)+iz}\,e^{i\alpha}
735: \ee
736: where $f(r)=r\cot\chi(r)$. Here the constant phase factor has been introduced
737: to account for the residual O(2) symmetry \eqref{residual}. 
738: Minimizing the energy with respect $F(r)$ 
739: it turns out that choosing  
740: \be                                        \label{Hopf4}
741: f(r)=0.453\,(r-0.878)(r^2+0.705r+1.415)
742: \ee
743: gives for 
744: the energy  a value which is 
745: less than $1\%$ above the true minimum, ${E}_1=1.22$ \cite{Ward:2000qj}. 
746: Eqs.\eqref{Hopf3},\eqref{Hopf4} provide therefore a good analytic approximation
747: for the $\Q=1$ hopfion. They show, in particular, 
748: that the position curve of the hopfion
749: is a circle of radius $R=0.848$ and that for large $r$ the field shows a dipole type 
750: behaviour, 
751: \be
752: \n^1+i\n^2\approx 2W \approx  a\,\frac{x+iy}{r^3}. 
753: \ee 
754: Ward also suggests the moduli space approximation in which the $|\Q|=1$ 
755: hopfion is viewed as an oriented circle (see Fig.\ref{FigH}). 
756: There are six continuous moduli parameters: 
757: three coordinates of the circle center, two angles determining the position 
758: of the circle axis, and also the overall phase.
759: Choosing arbitrarily a direction along the axis (shown by the vertical arrow in 
760: Fig.\ref{FigH}), there are two possible
761: orientations corresponding to the sign of the 
762: Hopf charge, changing which is achieved by $W\to W^\ast$.  
763:  Although for one hopfion the phase is not important, 
764: the relative phases of several hopfions determine their interactions. 
765: 
766: Ward conjectures  that well-separated hopfions interact as dipoles with the 
767: maximal attraction/repulsion when they are parallel/antiparallel,
768: respectively, since the like charges attract in a scalar field theory. 
769: He verifies this conjecture numerically, and then numerically relaxes
770: a field configuration corresponding to two mutually attracting hopfions.  
771: He discovers two possible outcomes of this process. If the two hopfions are initially
772: located in one plane 
773: then they approach each other till the two circles merge to one
774: thereby forming the ${\bf A}_{2,1}$ hopfion with $\Q=2$.  
775: If they are initially
776: oriented along the same line then they approach each other 
777: but do not merge even in the energy minimum, 
778: where they remain separated by a finite distance. 
779: This corresponds to the   
780: ${\bf A}_{1,2}$ hopfion, which is more energetic but locally stable. 
781: 
782: The ${\bf A}_{2,1}$ hopfion can be approximated by 
783: \be                                       \label{Hopf5}
784: W=\frac{(x+iy)^2 e^{i\alpha}}{f+i1.55\,z\,r},~~~
785: f=0.23(r-1.27)(r+0.44)(r+0.16)(r^2-2.15r+5.09)
786: \ee
787: whose energy is  $1.5\%$ above the true minimum, and there is also 
788: a similar approximation for the  ${\bf A}_{1,2}$ hopfion
789: \cite{Ward:2000qj}. 
790:  
791: \begin{figure}[ht]
792: \hbox to\linewidth{\hss%
793: 	\resizebox{12cm}{2.7cm}{\includegraphics{circ1.eps}}
794: \hspace{5mm}%
795: \hss}
796: 	\caption{Schematic representation of the hopfions as oriented circles}
797: \label{FigH}
798: \end{figure}
799: 
800: \subsection{Unknots, links and knots}
801: 
802: Similarly to the $\Q=1,2$ solutions, 
803: hopfions can be constructed within the axial ansatz \eqref{Hopf:axial} 
804: also for higher values of $n,m$. However, for $\Q>2$ they will not 
805: generically correspond
806: to the {\it global} energy minima. This can be understood 
807: if we remember that the Hopf charge
808: measures the amount of the twist deformation. 
809: Twisting an 
810: elastic rod shows that if it is twisted too much then the loop made of it
811: will not be planar, since it will find it energetically favorable to bend 
812: toward  the third direction. It is therefore expected that 
813: the ground state hopfions for higher values of $\Q$ 
814: will not be axially symmetric planar
815: loops, but 3D loops, generically without  any symmetries. 
816: 
817: The interest towards this issue was largely stirred by the work of Faddeev and Niemi
818: \cite{Faddeev:1996zj}, \cite{Faddeev:1997pf},
819: who conjectured that higher $\Q$ hopfions should be not just
820: closed lines but {\it knotted} closed lines, with the degree 
821: of knotedness expressed in terms 
822: of the Hopf charge. In other words, they conjectured that there could be 
823: a field theory realization of stable knots -- the idea that had been put forward by 
824: Lord Kelvin in the 19-th century \cite{Kelvin}  
825: but never found an actual realization. 
826: This knot conjecture of Faddeev and Niemi had 
827: a large resonance and several groups had started large scale numerical simulations 
828: to look for knotted hopfions. Such solutions have indeed been found, 
829: although not with quite the same properties as had been originally predicted in 
830: Ref.\cite{Faddeev:1996zj}, 
831: \cite{Faddeev:1997pf}. 
832: 
833: The first, really astonishing set of results has been reported by Battye and Sutcliffe
834: \cite{Battye:1998pe}, 
835: \cite{Battye:1998zn},
836: who managed to construct hopfions up to $\Q=8$ and found the first non-trivial knot --
837: the trefoil knot -- for $\Q=7$. 
838: Similar analyses have been then  
839: independently carried out by 
840: Hietarinta and Salo 
841: \cite{Hietarinta:1998kt}, 
842: \cite{Hietarinta:2000ci} 
843: and also by Ward 
844: \cite{Ward:1998pj},
845: \cite{Ward:2000qj},
846: \cite{Ward}.
847: All groups performed the full 3D energy minimization starting from an initial 
848: field configuration with a given $\Q$. 
849: Various initial configurations were used,  as for example the ones 
850: given by  the rational ansatz \eqref{WW-Hopf}, supplemented by non-axially 
851: symmetric perturbations to break the exact toroidal symmetry. 
852: The value of $\Q$ being constant during 
853: the relaxation, the numerical iterations were found to converge  to non-trivial 
854: energy minima,
855: whose structure was sometimes completely different from that of 
856: the initial configuration (an online animation of the relaxation process  
857: is available in \cite{movies}). 
858: Several local energy minima typically exist  
859: for a given $\Q$,  
860: sometimes with almost the same energy,  so that it was not always easy to know 
861: whether the minima obtained were local or global. Different initial configurations
862: were therefore tried to see if the minimum energy configurations could be reproduced
863: in a different way.  
864: As a result, it appears that the global energy minima have now been identified
865: and cross-checked up to $\Q=7$ \cite{Hietarinta:2000ci}, \cite{Sutcliffe:2007ui}, 
866: after which the analysis has been extended up to $\Q=16$ \cite{Sutcliffe:2007ui}.
867: The properties of the solutions can be summarized as follows.  
868: 
869: For $\Q=1,2$ these are the toroidal hopfions of  Gladikowski and Hellmund  
870: \cite{Gladikowski:1996mb}, ${\bf A}_{11}$ and ${\bf A}_{21}$. 
871: Although initially obtained within the constrained, axially symmetric 
872: relaxation scheme, they also correspond to the global minima of the full 3D energy 
873: functional. Axially symmetric hopfions ${\bf A}_{m1}$  
874: exist also for higher $\Q=m$ \cite{Battye:1998pe}, 
875: \cite{Battye:1998zn}, but they no longer correspond to global energy minima. 
876: For $\Q=3$ the ground state hopfion is not planar and is called in \cite{Sutcliffe:2007ui}
877: $\tilde{{\bf A}}_{31}$,
878: which can be viewed as deformed ${\bf A}_{31}$,
879: with a pretzel-like position curve bent in 3D to brake the axial symmetry. 
880: However, for $\Q=4$ the axial symmetry is restored again in the ground state,
881: ${\bf A}_{22}$, which seems to have a similar to the ${\bf A}_{12}$ two-ring structure 
882: \cite{Hietarinta:2000ci}, \cite{Ward}. 
883: The bent $\tilde{{\bf A}}_{41}$ also exists, 
884: but its energy it higher. 
885: 
886: 
887: Up to now all hopfions have been the simplest knots
888: topologically equivalent to a circle, called {\it unknots}
889: in the knot classification. 
890: A new phenomenon arises for $\Q=5$, 
891: since the fundamental hopfion in this case consists of two 
892: linked unknots. This has nothing to do with the linking of 
893: preimages determining  the value of $\Q$.
894: This time the position curve itself consists of two disjoint loops,
895: corresponding to  a charge 2 unknot linked to a charge 1 unknot.  
896: The Hopf charge is not simply the sum of charges of each component, 
897: but contains in addition the sum of their linking numbers 
898: due to their linking with the other components. It is worth noting that 
899: the linking number of the oriented circles can be positive or negative, 
900: depending on how they are linked \cite{Hietarinta:1998kt} (see Fig.\ref{FigL}).  
901: \begin{figure}[ht]
902: \hbox to\linewidth{\hss%
903: 	\resizebox{10cm}{2cm}{\includegraphics{circles.eps}}
904: \hspace{5mm}%
905: \hss}
906: 	\caption{Two possible ways to link $\Q=1$ hopfions. Left: the total linking
907: number is $2=1+1$ and the total charge if $\Q=1+1+2=4$.
908: Right: the total linking number is $-2=-1-1$ and the total charge is 
909: $\Q=1+1-2=0$.  
910: The numerical relaxation of these configurations gives therefore completely
911: different results \cite{movies}.}
912: \label{FigL}
913: \end{figure}  
914: Using the notation of Ref.\cite{Sutcliffe:2007ui},
915: the $\Q=5$ hopfion can be called 
916: ${\bf L}_{1,2}^{1,1}$,  
917: where the subscripts label the Hopf charges of the unknot components of the link,
918: and the superscript above each subscript counts the extra linking number of that component.
919: The total Hopf charge is the sum of the 
920: subscripts plus superscripts. Similarly, for $\Q=6$ the ground state hopfion is 
921: a link of two charge 2 unknots, so that it can be called ${\bf L}_{2,2}^{1,1}$. 
922: 
923: For $\Q=7$ the true knot appears at last: the ground state configuration 
924: corresponds in this case to the simplest non-trivial torus knot: trefoil knot. 
925: Let us remind that a $(p,q)$ torus knot is formed by wrapping a circle around 
926: a torus $p$ times in one direction and $q$ times in the other, 
927: where $p$ and $q$ are coprime integers, $p>q$. One can explicitly parametrize it as
928: $\rho = R+\cos\left({p\varphi}/{q}\right)$,
929: $z = \sin\left({p\varphi}/{q}\right)$, where $R>1$. 
930: A $(p,q)$ torus knots can also be obtained as the intersection of the unit three sphere
931: $S^2\in\mathbb{C}^2$ defined by $|\phi|^2+|\sigma|^2=1$ with the complex algebraic curve
932: $\sigma^p+\phi^q=0$. The trefoil knot is the $(3,2)$ torus knot, and it determines the 
933: profile of the position curve of the $\Q=7$ fundamental hopfion denoted 
934: $7{\bf K}_{3,2}$ in Ref.\cite{Sutcliffe:2007ui}. 
935: 
936: Sutcliffe \cite{Sutcliffe:2007ui} extends the energy minimization up to $\Q= 16$
937: specially 
938: looking for other knots. For the input configurations in his numerical procedure  
939: he uses fields parametrized
940: by the rational map ansatz,
941: \be                                \label{pq}
942: W=\frac{\sigma^a \phi^b}{\sigma^p+\phi^q},
943: \ee
944: where $0<a\in\mathbb{Z}$, $0\leq b\in\mathbb{Z}$ and $W,\phi,\sigma$ are defined by 
945: Eqs.\eqref{W-Hopf},\eqref{W-Hopf1}. The position curve for such a field configuration 
946: coincides with the $(p,q)$ torus knot, since the condition $W=\infty$ reproduces the 
947: knot equation. The parameters $a,b$ determine the value of the 
948: Hopf charge, $\Q=aq+bp$ \cite{Sutcliffe:2007ui}. If $p,q$ are not coprime, then the 
949: denominator in \eqref{pq} factorizes and the whole expression describes
950: a link. 
951: As a result, fixing a value of $\Q$ one can construct many different knot or link
952: configurations compatible with this value. Numerically relaxing these 
953: configurations, Sutcliffe finds many new energy minima,
954: discovering  new knots and links \cite{Sutcliffe:2007ui}. 
955: He also obtains configurations that he calls $\chi$  
956: whose position curve seems to self-intersect and so it is not quite clear to what
957: type they belong, unless the self-intersections are only apparent and can be resolved
958: by increasing the resolution. 
959: 
960: 
961: 
962: 
963: 
964: 
965: 
966: 
967: \begin{table} 
968: %\begin{center}
969: \caption{Known Hopf solitons 
970: according to Refs. 
971: \cite{Hietarinta:2000ci}, \cite{Sutcliffe:2007ui}, \cite{Battye:1998zn}. }
972: \label{tab1}
973: \begin{tabular}{c|c|c|c|c|c|c|c|c|c} \hline\hline
974: $\Q$ 
975: & {1} 
976: & {2} 
977: & 2 
978: & {3} 
979: & 3 
980: & {4} 
981: & 4 
982: & 4 
983: & {5} 
984: \\    
985: & $\underline{{\bf A}_{1,1}}$ 
986: & \underline{${\bf A}_{2,1}$}  
987: &  ${{\bf A}}_{1,2}$ 
988: & \underline{$\tilde{{\bf A}}_{3,1}$} 
989: & ${{\bf A}}_{3,1}$   
990: & \underline{${\bf A}_{2,2}$} 
991: & $\tilde{{\bf A}}_{4,1}$ 
992: & ${{\bf A}}_{4,1}$ 
993: & \underline{${\bf L}_{1,2}^{1,1}$ } 
994: \\ 
995: ${\mcal E}$ 
996: & {1} 
997: & {0.97 }
998: & 0.98 
999: & {1.00 }
1000: & 1.01 
1001: & {1.01 }
1002: & 1.03 
1003: & 1.06 
1004: & {1.02 }
1005: \\   \hline       
1006: $\Q$ 
1007: & 5 
1008: & 5 
1009: & {6} 
1010: & 6 
1011: & 6 
1012: & {7} 
1013: & 7 
1014: & {8} 
1015: & 8 
1016: \\    
1017: & $\tilde{{\bf A}}_{5,1}$  
1018: & ${{\bf A}}_{5,1}$  
1019: & \underline{${\bf L}_{2,2}^{1,1}$}  
1020: & ${\bf L}_{1,3}^{1,1}$ 
1021: & ${\bf A}_{6,1}$ 
1022:  & \underline{${\bf K}_{3,2}$ } 
1023: & ${\bf A}_{7,1}$      
1024: & \underline{${\bf L}_{3,3}^{1,1}$  }
1025: & ${\bf K}_{3,2}$ 
1026: \\ 
1027: ${\mcal E}$ 
1028: & 1.06 
1029: &  1.17 
1030: & {1.01} 
1031: & 1.09 
1032: & 1.22 
1033: & {1.01 } 
1034: & 1.20
1035: & { 1.02 }
1036: & 1.02 
1037: \\     \hline   
1038: $\Q$ 
1039: & 8 
1040: & {9} 
1041: & 9 
1042: & {10} 
1043: & 10 
1044: & 10 
1045: & {11} 
1046: & 11 
1047: & 11 
\\    
1048: & ${\bf A}_{8,1}$ 
1049: & \underline{ ${\bf L}_{1,1,1}^{2,2,2}$ }
1050: & ${\bf K}_{3,2}$ 
1051: & \underline{ ${\bf L}_{1,1,2}^{2,2,2}$ }
1052: & ${\bf L}_{3,3}^{2,2}$   
1053: & ${\bf K}_{3,2}$ 
1054: & \underline{ ${\bf L}_{1,2,2}^{2,2,2}$ }
1055: & ${\bf K}_{5,2}$  
1056: & ${\bf L}_{3,4}^{2,2}$
1057: \\ 
1058: ${\mcal E}$ 
1059: & 1.40 
1060: & { 1.02 }
1061: & 1.02 
1062: & { 1.02 }
1063: & 1.02 
1064: & 1.03 
1065: & { 1.02 }
1066: & 1.03 
1067: & 1.04 
1068: \\     \hline   
1069: $\Q$ 
1070: & 11 
1071: & 12 
1072: & 12 
1073: & 12 
1074: & 12 
1075: & {13} 
1076: & 13 
1077: & 13 
1078: & 13 
1079: \\    
1080: & ${\bf K}_{3,2}$                
1081: & \underline{ ${\bf L}_{2,2,2}^{2,2,2}$ }
1082: & ${\bf K}_{4,3}$ 
1083: & ${\bf K}_{5,2}$ 
1084: & ${\bf L}_{4,4}^{2,2}$
1085: & \underline{  ${\bf K}_{4,3}$ }
1086: &  ${\chi}_{13}$
1087: &  ${\bf K}_{5,2}$ 
1088: &  ${\bf L}_{3,4}^{3,3}$  
1089: \\ 
1090: ${\mcal E}$ 
1091: & 1.05           
1092: & { 1.01 }
1093: & 1.01 
1094: & 1.04 
1095: & 1.04 
1096: & { 1.00 }
1097: & 1.03 
1098: & 1.04 
1099: & 1.05 
1100: \\     \hline   
1101: $\Q$ 
1102: & {14} 
1103: & 14 
1104: & 14 
1105: & {15} 
1106: & 15 
1107: & 15   
1108: & 16
1109: \\    
1110: & \underline{  ${\bf K}_{4,3}$ }
1111: &  ${\bf K}_{5,3}$  
1112: &  ${\bf K}_{5,2}$ 
1113: &  \underline{ ${\mathbf \chi}_{15}$ }
1114: &  ${\bf L}_{1,1,1}^{4,4,4}$ 
1115: &  ${\bf K}_{5,3}$ 
1116: &  \underline{$\chi_{16}$}    
1117: \\ 
1118: ${\mcal E}$ 
1119: & { 1.00 }
1120: & 1.01 
1121: & 1.05 
1122: & { 1.01 }
1123: & 1.02 
1124: & 1.02 
1125: & { 1.01 } 
1126: \\   \hline      \hline   
1127:  \end{tabular}
1128: %\end{center}
1129: \end{table}
1130: 
1131:  
1132: The properties of all known hopfions, according to the results of
1133: Refs. \cite{Hietarinta:2000ci}, \cite{Sutcliffe:2007ui}, \cite{Battye:1998zn}, 
1134: are summarized 
1135: in Table \ref{tab1} and in Fig.\ref{Fig:knots}.  
1136: Table \ref{tab1} shows the Hopf charge, 
1137: the type of the solution, with the ground state configuration 
1138: in each topological sector underlined,
1139: and also the relative energy ${\mcal E}$ defined by the relation 
1140: \be                              \label{reden}
1141: E_\Q/E_1={\mcal E}\Q^{3/4},
1142: \ee 
1143: where $E_1$ is the energy of the $\Q=1$ hopfion. 
1144: Of the two decimal places of values of ${\mcal E}$ shown in the table the second one 
1145: is rounded.  
1146: Different groups give slightly different values for the energy, 
1147: but one can expect the relative energy to be less sensitive to this.  
1148: The values of ${\mcal E}$ for $\Q\leq 7$ shown in the table correspond to the data 
1149: of Hietarinta and Salo \cite{Hietarinta:2000ci} and of Sutcliffe 
1150: \cite{Sutcliffe:2007ui}, and it appears  that for  solutions described by both of these 
1151: groups these values are the same. The data for $8\leq \Q\leq 16$ are given 
1152: by Sutcliffe \cite{Sutcliffe:2007ui}, apart from  those for the ${\bf A}_{\Q,1}$ hopfions
1153: for $\Q=5,6,7,8$, which are found in the earlier  work of Battye and 
1154: Sutcliffe \cite{Battye:1998zn}.
1155: %(they are probably 1-2$\%$ inaccurate). 
1156: Although ${\bf A}_{\Q,1}$ solutions seem to exist
1157: also for $\Q>8$, no data are currently available for this case. 
1158: To obtain the energies from Eq.\eqref{reden} one can use the value $E_1=1.22$,
1159: which is known to be accurate to the two decimal places \cite{Ward:2000qj}. 
1160: 
1161: As one can see, the hopfion energies follow closely the topological lower 
1162: bound \eqref{Hopf}.
1163: This suggests that the ground state hopfions actually attain this 
1164: bound, so that they should be topologically stable. 
1165: According to the 
1166: data in Table \ref{tab1} one has $\inf\{{\mcal E}\}=0.97$ for $\Q\leq 16$, 
1167: and if this is true  
1168: for all values of $\Q$ then the optimal value for the constant 
1169: in the bound \eqref{Hopf} is $c=E_1\inf\{{\mcal E}\}=1.18$. 
1170: 
1171: A rigorous existence proof for the hopfions was given by Lin and Yang \cite{Lin}, 
1172: who demonstrate the existence of a smooth least energy configuration in every
1173: topological sector whose Hopf charge value belongs to an infinite (but unspecified)
1174: subset of $\mathbb{Z}$. 
1175: This shows that ground state hopfions exist, although perhaps 
1176: not for any $\Q\in\mathbb{Z}$.   
1177: As shown in Ref.\cite{Lin}, their {energy} 
1178: is bounded not only below but also {\it above} as
1179: \be
1180: E<C|\Q|^{3/4},
1181: \ee
1182: where $C$ is an absolute constant. This  
1183: implies that knotted solitons are energetically preferred 
1184: over widely separated unknotted multisoliton configurations when $\Q$
1185: is sufficiently large. Indeed, for a 
1186: decay into charge one elementary hopfions
1187: the energy should grow at least as $\Q$ for large $\Q$, but it grows slower.  
1188: 
1189: 
1190: 
1191: 
1192: \begin{figure}[h]
1193: \hbox to\linewidth{\hss%
1194:   \resizebox{14cm}{6cm}{\includegraphics{knots.eps}}%
1195: %  \resizebox{8cm}{6cm}{\includegraphics{FIG8.eps}}%
1196: \hss}
1197: \caption{
1198: Schematic profiles of the position curves for the known hopfions 
1199: (excepting the 
1200: $\chi$-solutions) 
1201: according to the results of Refs.\cite{Hietarinta:2000ci},\cite{Sutcliffe:2007ui}. 
1202: }
1203: \label{Fig:knots}
1204: \end{figure}
1205: 
1206: The position curves of the solutions, schematically shown in Fig.\ref{Fig:knots},
1207: present an amazing variety of shapes. It should be stressed though that other 
1208: characteristics of the solutions, as for example their energy density, do not 
1209: necessarily show the same knotted pattern. 
1210: Several types of links and 
1211: knots appear, and  each particular type can appear several times, for different
1212: values of the Hopf charge. Intuitively, one can view the position curves as wisps 
1213: made of two intertwined lines corresponding to preimages of two infinitely close to 
1214: $-\nn_\infty$ points on the target space \cite{Sutcliffe:2007ui}. Increasing the twist
1215: of the wisp increases the Hopf charge, without necessarily 
1216: changing the topology of the position curve. A more detailed inspection 
1217: (see pictures in \cite{Sutcliffe:2007ui}) actually shows that configurations
1218: appearing several times in Fig.\ref{Fig:knots}, as for example the trefoil knot 
1219: ${\bf K}_{3,2}$, become more and more distorted by the internal twist as $\Q$ increases. 
1220: Finally, for some critical value of $\Q$, the excess of the intrinsic 
1221: deformation makes it energetically favorable to change 
1222: the knot/link type and pass to other, more complicated knot/link configurations. 
1223: Estimating the length of the position curves shows that it grows as $\Q^{3/4}$,
1224: so that the energy per unit length is approximately the same for all hopfions
1225: \cite{Sutcliffe:2007ui}.
1226: 
1227: \subsection{Conformally invariant knots} 
1228: 
1229: Hopf solitons in the Faddeev-Skyrme model, also  
1230: sometimes called in the literature  
1231: knot solitons of Faddeev-Niemi,  
1232: of Faddeev-Skyrme, or of Faddeev-Hopf
1233: provide the best known
1234: example of knot solitons in field theory.
1235: However, there are also other field theory models admitting knotted solitons
1236: with a non-zero Hopf index.
1237: An interesting example proposed by Nicole \cite{Nicole} is obtained
1238: by taking the first term in the  Faddeev-Skyrme model \eqref{action0} and
1239: raising it to a fractional power,
1240: \be                                      \label{Nicole}
1241: {\mcal L}_{\rm Nicole}=\left(-\partial_\mu\nn\cdot\partial^\mu\nn\right)^{3/2}\,.
1242: \ee 
1243: A similar possibility, suggested by Aratyn, Ferreira and Zimerman (AFZ) 
1244: \cite{Aratyn}, uses the second term in the  Faddeev-Skyrme Lagrangian,
1245: \be                                      \label{AFZ}
1246: {\mcal L}_{\rm AFZ}=\left({\mcal F}_{\mu\nu}{\mcal F}^{\mu\nu}\right)^{3/4}\,.
1247: \ee 
1248: Both of these models are conformally invariant in three spatial dimensions
1249: and so the existence of static solitons is not excluded for them by the Derrick argument. 
1250: In fact, static knot solitons in these models exist and can even be
1251: obtained in close analytical form, which is achieved by simply using the 
1252: axial ansatz \eqref{Hopf:axial} with the function $\Theta,\psi$ expressed 
1253: in toroidal coordinates \eqref{toroidal} according to Eq.\eqref{toroidal1}
1254: \cite{Aratyn:1999cf}, \cite{Adam:2006wg}. 
1255: Curiously, this separates away the $v,\varphi$ 
1256: variables in the field equations reducing the problem
1257: to an {\it ordinary} differential equation for $\Theta(u)$ 
1258: (in the Faddeev-Skyrme theory this does not work). In the AFZ model
1259: solutions for $\Theta(u)$ can be expressed in terms of elementary functions 
1260: for any $n,m$ \cite{Aratyn:1999cf}, while in the Nicole model they are obtained 
1261: numerically \cite{Adam:2006wg}, 
1262: apart from the $|n|=|m|=1$ case, where the solution turns out to be
1263:  the same in both 
1264: models  and is given by 
1265: $\tan(\Theta/2)=\sinh u$ 
1266: \cite{Nicole}, 
1267: \cite{Aratyn}.  
1268: These results remain, however,  interesting mainly from the purely
1269: mathematical point of view at the time being, since it is difficult 
1270: to justify physically 
1271: the appearance of the fractional powers in Eqs.\eqref{Nicole},\eqref{AFZ}. 
1272: 
1273: Other examples of solitons with a Hopf charge 
1274: will be discussed below in Sec.\ref{o} and Sec.\ref{magnet}.
1275: 
1276: 
1277: 
1278: 
1279: 
1280: 
1281: 
1282: 
1283: \subsection{Can one gauge the knot solitons ?}
1284: 
1285: The Faddeev-Skyrme theory is a {\it global} field model, 
1286: so that it cannot be a fundamental physical theory like gauge field theory models, 
1287: but perhaps can be viewed as an effective theory. This suggests using the Faddeev-Skyrme 
1288: knots for an effective description of some physical objects, and so it has
1289: been conjectured by Faddeev and Niemi 
1290: that they could be used for an effective  description 
1291: of glueballs in  the strongly coupled Yang-Mills theory 
1292: \cite{Faddeev:1998eq},
1293: \cite{Faddeev:2006sw},
1294: \cite{Shabanov:1999xy},
1295: \cite{Shabanov:1999uv},
1296: \cite{Wipf},
1297: \cite{Faddeev:2003aw}. 
1298: This conjecture is very interesting, quite in the spirit of the original
1299: Lord Kelvin's idea to view atoms as knotted ether tubes 
1300: \cite{Kelvin}, and perhaps it could apply in some form. In fact,  
1301: when describing the $\eta(1440)$ meson, 
1302: the Particle Data Group says  (see p.591 in Ref.\cite{Yao}) 
1303: that ``the $\eta(1440)$ is an excellent candidate 
1304: for the $0^{-+}$ glueball in the flux tube model \cite{Faddeev:2003aw}''. 
1305: 
1306: Some other physical applications of the global field theory knot solitons 
1307: could perhaps be found. However, if they could be promoted to  {\it gauge} field theory
1308: solutions, then they would be much more interesting physically, since in this case 
1309: they would find many interesting applications, as for example in the theories of 
1310: superconductivity and  of Bose-Einstein condensation
1311: \cite{Babaev:2001jt},
1312: \cite{Babaev:2001zy},
1313: in the theory of plasma 
1314: \cite{Faddeev:2000rp}, 
1315: \cite{Faddeev:2000qw}, 
1316: in Standard Model
1317: \cite{Cho:2001gc},
1318: \cite{Fayzullaev:2004xa},
1319: \cite{Niemi:2000ny},
1320: or perhaps even in cosmology, where they could presumably 
1321: describe knotted cosmic strings \cite{Vilenkin}. 
1322: For this reason it has been repeatedly conjectured in the literature 
1323: that some analogs of the Faddeev-Skyrme knot solitons could also exist 
1324: as static solutions   
1325: of the gauge field theory equations of motion. This conjecture 
1326: is essentially inspired
1327: by the fact that the Faddeev-Skyrme theory already contains 
1328: something like a gauge field:
1329: ${\mcal F}_{\mu\nu}$. Moreover, we shall now see that 
1330: changing the variables one can rewrite the 
1331: theory in such a form that it looks almost identical to a gauge field theory 
1332: (or the other way round). 
1333: 
1334: 
1335: 
1336: 
1337: \subsection{Faddeev-Skyrme model versus semilocal Abelian Higgs model}
1338: 
1339: Let $\Phi$ be a doublet of complex scalar fields satisfying a constraint, 
1340: \be                                                    \label{norm}
1341: \Phi=
1342: \left(\begin{array}{c}
1343: \phi\\
1344: \sigma\\
1345: \end{array}
1346: \right),~~~~~~~~~
1347: \Phi^\dagger\Phi=|\phi|^2+|\sigma|^2=1,
1348: \ee
1349: such that $\Phi\in S^3$. 
1350: Let us consider a  field theory defined by the Lagrangian density 
1351: \be                                                        \label{action1}
1352: {\mcal L}[\Phi]=-\frac{1}{4}\,{\ccal F}_{\mu\nu}{\ccal F}^{\mu\nu}
1353: +({\ccal D}_\mu\Phi)^\dagger {\ccal D}^\mu\Phi
1354: \ee
1355: with 
1356: ${\ccal F}_{\mu\nu}=\partial_\mu {\ccal A}_\nu-\partial_\nu {\ccal A}_\mu$
1357: and ${\ccal D}_\mu\Phi=(\partial_\mu-i{\ccal A}_\mu)\Phi$ and with
1358: \be                                                     \label{A} 
1359: {\ccal A}_\mu=-i\Phi^\dagger\partial_\mu\Phi\,.
1360: \ee 
1361: In fact, this  theory is again the  Faddeev-Skyrme model 
1362: but rewritten in different variables, 
1363: since upon the identification  
1364: \be                                             \label{project}
1365: n^a=\Phi^\dagger\tau^a\Phi
1366: \ee 
1367: ($\tau^a$ being
1368: the Pauli matrices) the fields ${\ccal A_\mu}$ and 
1369: ${\ccal F}_{\mu\nu}$ coincide with those in 
1370: \eqref{action0} \cite{Ren} and the whole action 
1371: \eqref{action1} reduces (up to an overall factor) 
1372: to \eqref{action0}  
1373: \cite{Faddeev:1997pf}. 
1374: More precisely, \eqref{project} is the Hopf projection from $S^3$ 
1375: parametrized by $(\phi,\sigma)$  
1376: to $S^2$ parametrized by the complex projective coordinate $\phi/\sigma$.
1377: For example, the axially symmetric fields \eqref{Hopf:axial}, \eqref{Hopf:axial1}
1378: are obtained in this way by choosing the $CP^1$ variables 
1379: \be                                        \label{axial-CP}
1380: \phi=\cos\frac{\Theta}{2}\,e^{in\psi},~~~~~
1381: \sigma=\sin\frac{\Theta}{2}\,e^{im\varphi},
1382: \ee
1383: such that the phases of $\phi$ and $\sigma$ wind, respectively, along the
1384: two orthogonal direction  as shown in Fig.\ref{FigS} and the Hopf charge is 
1385: $\Q=nm$. 
1386: 
1387: 
1388: 
1389: The fields in the static limit, $\Phi=\Phi(\bx)$, can now
1390: be viewed 
1391: as maps $S^3\to S^3$, but their energy 
1392: \be                                                                    \label{en1}
1393: E[\Phi]=\int\left( |{\ccal D}_k\Phi|^2+
1394: \frac{1}{4}\,({\ccal F}_{ik})^2
1395: \right)
1396: d^3 \bx 
1397: \ee
1398: is still bounded from below as in Eq.\eqref{Hopf}.
1399: The  topological charge $\Q=\Q[\Phi]$, 
1400: still expressed by Eq.\eqref{Q}, 
1401: is now interpreted as the index 
1402: of map $S^3\to S^3$,
1403:  $\Q\in\pi_3(S^3)=\mathbb{Z}$. 
1404: The theory  therefore  
1405: admits the same knot solitons as the original Faddeev-Skyrme model. 
1406: However, in the 
1407: new parametrization the theory looks almost like a gauge field theory,
1408: in particular it exhibits a local 
1409: U(1) gauge invariance under $\Phi\to e^{i\alpha}\Phi$, 
1410: $A_\mu\to A_\mu+\partial_\mu\alpha$.
1411: 
1412: 
1413: 
1414: Let us now compare the model \eqref{action1} to a genuine gauge field theory 
1415: with the Lagrangian  density 
1416: \be                                                           \label{action2}
1417: {\mcal L}[\Phi,A_\mu]=-\frac{1}{4}\,F_{\mu\nu}F^{\mu\nu}
1418: +(D_\mu\Phi)^\dagger D^\mu\Phi-\frac{\lambda}{4}\,(\Phi^\dagger\Phi-1)^2\,.
1419: \ee
1420: Here $\Phi$ is again a doublet of complex scalar fields, but this time 
1421: without the normalization condition \eqref{norm}, the condition  \eqref{A}
1422: being also relaxed, so that $A_\mu$ is now an independent field. One has 
1423: ${F}_{\mu\nu}=\partial_\mu {A}_\nu-\partial_\nu {A}_\mu$
1424: and ${D}_\mu\Phi=(\partial_\mu-i{A}_\mu)\Phi$. 
1425: This semilocal \cite{Achucarro:1999it}
1426: Abelian Higgs model with the SU(2)$_{\rm global}\times$U(1)$_{\rm local}$
1427: internal symmetry arises in different contexts, in particular it can be viewed 
1428: as the Weinberg-Salam model in the limit where the weak mixing angle is $\pi/2$
1429: and the SU(2) gauge field decouples. 
1430: The non-relativistic limit of this theory
1431: is the two-component Ginzburg-Landau model \cite{Ginzburg}. 
1432: 
1433: Let us now consider the limit $\lambda\to\infty$. In this sigma model limit the
1434: constraint \eqref{norm} is enforced and the potential term in \eqref{action2} vanishes.  
1435: The theories \eqref{action1} and \eqref{action2} then look identically the same,
1436: the only difference being that in the first case the vector field ${\ccal A}_\mu$ 
1437: is defined by Eq.\eqref{A} and so is composite, 
1438: while in the second case $A_\mu$ in an independent field. 
1439: 
1440: \subsection{Energy bound in the Abelian Higgs model}
1441: 
1442: The question now arises: does the gauge field theory \eqref{action2},
1443: at least in the limit where $\Phi^\dagger\Phi=1$, admit 
1444: knot solitons analogues to those of the global model \eqref{action1} ? 
1445: If exist, such solutions would correspond to minima 
1446: of the energy in the theory \eqref{action2} 
1447: in static, purely magnetic sector,  
1448: \be                                                                    \label{en2}
1449: E[\Phi,A_k]=\int\left( |{D}_k\Phi|^2+
1450: \frac{1}{4}\,({F}_{ik})^2
1451: \right)
1452: d^3 \bx \,.
1453: \ee
1454: At first thought, one may think that the answer to this question should be affirmative. 
1455: Indeed, the energy functionals \eqref{en1} and \eqref{en2} look identical. 
1456: They have the same internal symmetries and the same scaling behaviour under
1457: $\bx\to \Lambda \bx$. The two theories also have the same topology 
1458: associated to the field $\Phi$,
1459: since in both cases $\Phi(\bx)$ defines a 
1460: mapping $S^3\to S^3$ with the topological 
1461: index \eqref{Q}. 
1462: 
1463: The gauged model \eqref{en2} contains in fact even more 
1464: charges than the global theory \eqref{en1}, 
1465: since it actually has two vector fields: the independent gauge field 
1466: $A_k$  and the composite field 
1467: ${\ccal A}_k=\frac{i}{2}(\partial_k\Phi^\dagger\Phi-\Phi^\dagger\partial_k\Phi)$. 
1468: It is convenient to introduce their difference
1469: $C_k=A_k-{\ccal A}_k$. 
1470: Defining the linking number between two vector fields, 
1471: \be
1472: I[A,B]=\frac{1}{4\pi^2}\int \epsilon_{ijk}{A}_i \partial_j{B}_k\, d^3\bx\,,  
1473: \ee
1474: one can construct three different charges,
1475: \be                             \label{charges}
1476: \Q[\Phi]\equiv I[{\ccal A},{\ccal A}],~~~~~
1477: {\rm L}=I[C,{A}],~~~~~
1478: N_{\rm CS}[A]\equiv I[{A},{A}],
1479: \ee
1480: which are, respectively, the topological charge \eqref{Q}, 
1481: the linking number between $A_k$ and $C_k$, 
1482: and the Chern-Simons number of the gauge field.
1483: The following 
1484: inequality, established by Protogenov and Verbus \cite{Protogenov:2002bt},
1485: holds:
1486: \be                                                           \label{Proto} 
1487: E[\Phi,A_k]\geq c_1|\Q|^{3/4}\left(1-\frac{|{\rm L}|}{|\Q|}\right)^2\,,
1488: \ee
1489: where $c_1$ is a positive constant. This can be considered 
1490: as the generalization of the 
1491: topological bound \eqref{Hopf} of Vakulenko-Kapitanski. 
1492: 
1493: Given all these, one may believe that the local  
1494:  theory \eqref{en2} admits topologically 
1495: stable knot solitons similar to those of the global theory \eqref{en1}. 
1496: 
1497: 
1498: \subsection{The issue of charge fixing}
1499: 
1500: The difficulty with implementing the Protogenov-Verbus  bound 
1501: \eqref{Proto} in practice 
1502: is that it contains two different charges, $\Q$ and L,
1503: but without invoking additional 
1504: physical assumptions  there is no reason why L should be fixed while minimizing 
1505: the energy.  
1506: 
1507: Let us consider first the charge $\Q=\Q[\Phi]$. Its variation vanishes identically,
1508: so that it 
1509: does not change under smooth deformations of $\Phi$. It is therefore a genuine  
1510: topological charge whose value is completely determined by the boundary 
1511: conditions, and so it should be fixed when minimizing the energy. 
1512: 
1513: Let us now consider the linking number L=$I[C,A]$. The analogs of this quantity have been
1514: studied in the theory of fluids, where they are known to be integrals 
1515: of motion \cite{Zakharov}. In the context of gauge field theory,
1516: this quantity is gauge invariant. However, it is not 
1517: a topological invariant,
1518: since its variation does not vanish identically and so it does change 
1519: under arbitrary smooth deformations of $C_k=A_k-{\ccal A}_k$.
1520: On cannot fix L using only continuity arguments,
1521: because there are no topological conditions imposed on  
1522: $A_k$, whose  arbitrary deformations are allowed.  
1523: The only topological quantity associated to $A_k$ 
1524: could be a magnetic charge related to a non-trivial U(1) bundle structure.
1525: However, since we are interested in globally regular solutions, 
1526: the bundle base space
1527: is $\mathbb{R}^3$ (or $S^3$) without removed points, 
1528: in which case the bundle is trivial.  
1529: 
1530: As a result,  on continuity grounds only,
1531: one can fix $\Q$  but not L. But then, 
1532: as is obvious from Eq.\eqref{Proto},  there is no 
1533: non-trivial lower bound for the energy, since one can always choose L=$\Q$
1534: in which case the expression on the right in \eqref{Proto} vanishes. 
1535: More precisely, since there are no constraints  for $A_k$, 
1536: nothing prevents one from smoothly deforming it to zero, 
1537: after which one can scale away the 
1538: rest of the configuration. Explicitly, given fields $\Phi,A_k$ 
1539: one can reduce $E[\Phi,A_k]$ to {\it zero} via a continues 
1540: sequence of smooth field deformations
1541: preserving the value of the  topological charge  $\Q[\Phi]$,  
1542: \be
1543: E[\Phi(\bx),A_k(\bx)]\to E[\Phi(\Lambda \bx),\gamma A_k(\bx)]
1544: \ee
1545: by taking first the limit $\gamma\to 0$ and then $\Lambda\to \infty$ \cite{Forgacs}. 
1546: 
1547: The conclusion is that without constraining $A_k$, with only $\Q[\Phi]$ fixed, 
1548: the absolute minimum of $E[\Phi,A_k]$ is zero, so that 
1549: there can be no absolutely stable knots. To have such solutions, 
1550: one would need to constraint
1551: somehow the vector field $A_k$, for example 
1552:  it would be enough
1553: to insure that $C_k=A_k-{\ccal A}_k$ be 
1554: zero or small.
1555: Such a condition 
1556: is often assumed in the literature 
1557: \cite{Babaev:2001zy},
1558: \cite{Babaev:2001jt},
1559: \cite{Cho:2001gc},
1560: \cite{Fayzullaev:2004xa},
1561: but usually simply {\it ad hoc}. 
1562: Unfortunately, it cannot be justified
1563: on continuity grounds only, without an additional physical input.
1564: 
1565: 
1566: 
1567: \subsection{Searching for gauged knots} 
1568: 
1569: The above arguments do not rule out
1570:  all solutions in the theory. Even though the 
1571: global minimum of the energy is zero, there could still 
1572: be non-trivial local minima or saddle points. 
1573: The corresponding static solutions would  be metastable or unstable. 
1574: One can  therefore wonder whether  such solutions exist. This question has 
1575: actually a long 
1576: history, being first addressed by de Vega \cite{deVega:1977rk} and by 
1577: Huang and Tipton \cite{Huang:1980bz} over 30 years ago,
1578: and being then repeatedly reconsidered by different authors 
1579: \cite{Koma:1999sm},
1580: \cite{Niemi:2000ny},
1581: \cite{Forgacs},
1582: \cite{Ward:2002vq},
1583: \cite{Jaykka:2006gf},
1584: \cite{Doudoulakis:2006iw},
1585: \cite{Doudoulakis:2007ti},
1586: \cite{Doudoulakis:2007xz}.
1587: However, the  answer is still unknown. 
1588: No solutions have been found up to now, neither has it been shown 
1589: that they do not exist. 
1590: 
1591: Trying to find the answer, all the authors were minimizing  the energy 
1592: given by the sum of $E[\Phi,A_k]$ 
1593: and of a potential term that can be either of the 
1594: form contained in \eqref{action2} 
1595: or  a more general one. A theory 
1596: with two vector fields with a local U(1)$\times$U(1) 
1597: invariance -- Witten's model of superconducting strings \cite{Witten:1984eb} --
1598: has also been considered in this context 
1599: \cite{Doudoulakis:2006iw},
1600: \cite{Doudoulakis:2007ti},
1601: \cite{Doudoulakis:2007xz}. 
1602: The energy was minimized within classes of fields with a 
1603: given topological charge $\Q$ and having 
1604: profiles of a  loop of radius $R$. 
1605: The resulting minimal value of energy was always found to 
1606: be a monotonously growing  
1607: function of $R$, thus always showing the tendency of the loop to shrink  
1608: thereby reducing
1609: its energy.
1610: 
1611: These results render the existence of solutions somewhat implausible. 
1612: However, they do not yet prove
1613: their absence. Indeed, local energy minima  may be difficult to 
1614: detect via energy minimization, 
1615: as this would require starting the numerical iterations 
1616: in their close vicinity, because otherwise the minimization procedure
1617: converges to the trivial global minimum. 
1618: In other words, one has to choose a good initial configuration. 
1619: However, since `there is a lot of room in function space', 
1620: chances to make the right choice are not high. 
1621: 
1622: It should be mentioned that a positive result was once reported 
1623: in Ref.\cite{Niemi:2000ny},
1624: were the energy was minimized in the $\Q=2$ sector 
1625: and an indication 
1626: of a convergence to a non-trivial minimum was observed. However, 
1627: this result was not confirmed in an independent analysis 
1628: in Ref.\cite{Ward:2002vq}, so that  
1629: it is unclear whether it should be attributed to a 
1630: lucky choice of the initial configuration or to 
1631: some numerical artifacts.  
1632: 
1633: \section{Knot solitons in gauge field theory}
1634: 
1635: 
1636: 
1637: 
1638: 
1639: As we have seen, fixing only the topological charge does not guarantee 
1640: the existence of knot solitons in gauge field theory. 
1641: In order to obtain such solutions
1642: one needs to constraint  the gauge field in order that it could not 
1643: be deformed to zero. 
1644: Below we describe two known examples of such solutions.  
1645: 
1646: \subsection{Anomalous solitons}
1647: 
1648: One possibility to constraint the gauge field is to fix 
1649: its Chern-Simons number.
1650: An example of how this can be done was suggested long ago by 
1651: Rubakov and Tavkhelidze 
1652: \cite{Rubakov:1986am}, 
1653: \cite{Rubakov:1985nk}, 
1654: who showed that the Chern-Simons number 
1655: can be fixed by including fermions into the system.  
1656: They considered the 
1657: Abelian Higgs model \eqref{action2}, but with a {\it singlet} 
1658: and not doublet Higgs field, augmented by including 
1659: chiral fermions.
1660: In the weak coupling limit at zero temperature this theory contains 
1661: states with $N_F$ non-interacting fermions and with the bosonic fields
1662: being in vacuum, $A_\mu=0$, $\Phi=1$. The
1663: energy of such states is  $E\sim N_F$. Rubakov and Tavkhelidze argued that 
1664: this energy could be decreased  via exciting the 
1665: bosonic fields in the following way. 
1666: 
1667: Owing to the axial anomaly, when the gauge field $A_\mu$ varies, 
1668: the fermion energy levels can cross zero and dive into 
1669: the Dirac see. 
1670: The fermion number can therefore change, but 
1671: the difference $N_F-N_{\rm CS}$ is conserved. As a result, starting from
1672: a purely fermionic state and increasing the gauge field, 
1673: one can smoothly deform this state  
1674: to a purely
1675: bosonic state, whose 
1676: Chern-Simon number will be fixed by the initial conditions,
1677: \be                                                     \label{CS}
1678: N_{\rm CS}=N_F.
1679: \ee
1680: Now, the energy of this state, 
1681: \be                                                                    \label{en3}
1682: E[\Phi,A_k]=\int\left( |{D}_k\Phi|^2+
1683: \frac{1}{4}\,({F}_{ik})^2+\frac{\lambda}{4}\,(|\Phi|^2-1)^2
1684: \right)
1685: d^3 \bx \,,
1686: \ee
1687: can be shown to be bounded from {\it above} by $E_0(N_{\rm CS})^{3/4}$
1688: where $E_0$ is a constant, and so for large $N_F=N_{\rm CS}$ it grows slower 
1689: than the energy of the original fermionic state,  
1690: $E\sim N_{\rm CS}$ \cite{Rubakov:1986am}, \cite{Rubakov:1985nk}.
1691: Therefore, for large enough $N_F$, it is energetically favorable for 
1692: the original purely fermionic state 
1693: to turn into a purely bosonic state. The latter is called {\it anomalous} 
1694: \cite{Rubakov:1986am}, \cite{Rubakov:1985nk}.
1695: The energy of this anomalous state can be obtained by minimizing the 
1696: functional \eqref{en3} with the Chern-Simons number fixed by the condition \eqref{CS}.  
1697: 
1698: Such an energy minimization was carried out in the recent work of Schmid and 
1699: Shaposhnikov \cite{Schmid:2007dm}. 
1700: First of all, they established the following inequality,
1701: \be                                  \label{CS:bound}
1702: E[\Phi,A_k]\geq c(N_{\rm CS})^{3/4},
1703: \ee  
1704: which reminds very much of the Vakulenko-Kapitanski bound \eqref{Hopf} 
1705: for the Faddeev-Skyrme model,
1706: but with the topological charge replaced by $N_{\rm CS}$. 
1707: This gives a very good example of how constraining the gauge field
1708: can stabilize the system: even though in the theory 
1709: with a singlet Higgs field there is no
1710: topological charge similar to the Hopf charge, 
1711:  its role can be taken over by the Chern-Simons charge.
1712: 
1713: 
1714: 
1715: \begin{figure}[h]
1716: \hbox to\linewidth{\hss%
1717:   \resizebox{7cm}{5cm}{\includegraphics{spindle.eps}}%
1718: \hss}
1719: \caption{\small 
1720: Spindle torus shape of the anomalous solitons for large $N_{\rm CS}$. 
1721: }
1722: \label{Fig1}
1723: \end{figure}
1724: 
1725: In order to numerically minimize the energy, 
1726: Schmid and Shaposhnikov considered the Euler-Lagrange equations
1727: for the functional
1728: \be
1729:  E[\Phi,A_k]+\mu\int\epsilon_{ijk}A_i\partial_j A_k d^3\bx
1730: \ee
1731: where $E[\Phi,A_k]$ is given by \eqref{en3} and 
1732: $\mu$ is the Lagrange multiplier. In the gauge where $\Phi=\Phi^\ast\equiv\phi$
1733: these equations read (with $\vec{A}=A_k$)
1734: \begin{subequations}               
1735: \begin{align}
1736: &\Delta\phi-\vec{A}^{\,2}\phi-\frac{\lambda}{2}\,(\phi^2-1)\phi=0, \label{e1} \\
1737: &\vec{\nabla}\times(\vec{\nabla}\times\vec{A})
1738: +2\mu\vec{\nabla}\times\vec{A}+2\phi^2\vec{A}=0,\label{e2}
1739: \end{align}
1740: \end{subequations} 
1741: where $\vec{\nabla}$ and $\Delta=(\vec{\nabla})^2$ 
1742: are the standard gradient and Laplace operators,
1743: respectively. 
1744: Multiplying Eq.\eqref{e2} by $\vec{A}$ and integrating gives the expression for the 
1745: Lagrange multiplier,
1746: \be
1747: \mu=\frac{1}{32\pi^2 N_{\rm CS}}\left.\left.\int\right(
1748: (\nabla\times \vec{A})^2+|\vec{A}|^2\phi^2\right)d^3x.
1749: \ee  
1750: Solutions of the elliptic equations \eqref{e1},\eqref{e2} 
1751: were studied in \cite{Schmid:2007dm}
1752: in the 
1753: axially symmetric sector, where $\phi=\phi(\rho,z)$ and 
1754: $\vec{A}=\vec{A}(\rho,z)$, with the boundary conditions 
1755: \be
1756: \phi=1,~~~~~\vec{A}=0
1757: \ee
1758: at infinity and 
1759: \be
1760: \partial_\rho\phi=\partial_\rho A_z=0,~~~~
1761: A_\rho=A_\varphi=0
1762: \ee
1763: at the symmetry axis $\rho=0$. The solutions obtained are quite interesting.
1764: They 
1765: are very strongly localized in a compact region, $\Omega$, of the $(\rho,z)$
1766: plane centered around a point $(\rho_s,0)$. Inside $\Omega$ the 
1767: field ${A}_k$ is non-zero, while $\phi$ is almost constant 
1768: and very close to zero. As one approaches the boundary of the region,
1769: $\partial\Omega$, the field ${A}_k$ tends to zero, while $\phi$ is still almost 
1770: zero.
1771: Finally, in a  small neighborhood of 
1772: $\partial\Omega$ whose thickness is of the order of the Higgs boson wavelength,
1773: $\phi$ starts varying and quickly increases up to its asymptotic value. 
1774: Outside $\Omega$ one has everywhere $\phi\approx 1$ and $A_k\approx 0$,
1775: such that the energy density is almost zero.  The energy for these solutions scales as 
1776: $(N_{\rm CS})^{3/4}$. 
1777: For large $N_{\rm CS}$ the region $\Omega$ 
1778: can be described by the simple analytic formula,
1779: \be
1780:  (\rho-\rho_s)^2+z^2<R_s^2,
1781: \ee
1782: where $R_s>\rho_s$. 
1783: In other words, the 3D domain where the soliton energy is concentrated 
1784: can be obtained by rotating around the $z$-axis a disc centered at a point 
1785: whose distance from the $z$-axis is less than its radius (see Fig.\ref{Fig1}). 
1786: Such a geometric figure is called spindle torus \cite{Schmid:2007dm}. 
1787: The solutions for large $N_{\rm CS}$ are 
1788: very well approximated by setting $\phi=1$ and $A_k=0$
1789: outside the spindle torus, while inside it one has $\phi=0$ and $A_k$ is obtained
1790: by solving the linear equation \eqref{e2}. The spindle torus approximation
1791: becomes better and better for large $N_{\rm CS}$, in which  limit 
1792: Schmid and Shaposhnikov obtain the following asymptotic formulas for the energy 
1793: and parameters of the torus, which agree very well with their
1794: numerics,
1795: \be
1796: E=118\,\lambda^{1/4}M_{W}(N_{\rm CS})^{3/4},~~~~~
1797: \rho_s=\frac{1.7}{M_{W}}\left(\frac{N_{\rm CS}}{\lambda}\right)^{1/4},~~~~~
1798: R_s=1.49\rho_s\,,
1799: \ee
1800: where $M_W$ is the vector field mass.
1801: 
1802: It is likely that these solution attain the lower energy bound \eqref{CS},
1803: which means that they should be topologically {\it stable}. 
1804: It should, however, be emphasized that the anomalous solitons require a rather exotic physical 
1805: environment, since for them to be energetically favoured  
1806: as compared to the free fermion condensate, the density of the latter should attain
1807: enormous values possible perhaps only in the core  of neutron stars. 
1808: 
1809: 
1810: Summarizing, fixing the Chern-Simons number forbids deforming 
1811: the gauge field to zero
1812: and gives rise to  stable knot solitons in gauge 
1813: field theory, even if
1814:  the Higgs field 
1815: is topologically trivial. It is unclear whether
1816: this result can be generalized within the context of the model \eqref{action2} 
1817: with two component Higgs field -- since the Protogenov-Verbus
1818: formula \eqref{Proto} contains 
1819: not the Chern-Simons number but the linking number L. 
1820: 
1821: \subsection{Non-Abelian rings \label{rings}} 
1822: 
1823: Another interesting class of objects arises in the non-Abelian gauge
1824: field theory, where one can have smooth, finite energy loops stabilized by 
1825: the magnetic energy.
1826: 
1827: 
1828: \subsubsection{Yang-Mills-Higgs theory}
1829: Let us parametrize 
1830: the non-Abelian Yang-Mills-Higgs theory for a compact and simple  gauge group 
1831: ${\mathcal G}$ as
1832: \be                                           \label{YMH}
1833: {\mcal L}[A_\mu,\Phi]=-\frac14 \langle F_{\mu\nu}F^{\mu\nu}\rangle
1834: +({D}_\mu\Phi)^\dagger {D^\mu\Phi}-
1835: U(\Phi)\,.
1836: \ee
1837: Here the gauge field strength is 
1838: $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu
1839: -ig[A_\mu, A_\nu]\equiv F^a_{\mu\nu}{\bf T}_a$
1840: where $A_\mu=A^a_\mu{\bf T}_a$ is the gauge field, 
1841: $a=1,2,\ldots ,{\rm dim} ({\mathcal G})$,
1842: and $g$ is the gauge coupling. The Hermitian gauge group
1843: generators ${\bf T}_a$ satisfy the relations
1844: \be
1845: [{\bf T}_a,{\bf T}_b]=i C_{abc}{\bf T}_c,~~~~~ 
1846: {\rm tr}({\bf T}_a{\bf T}_b)={\rm K}\delta_{ab}\,.
1847: \ee
1848: The Lie algebra inner product
1849: is defined as $\langle AB\rangle =\frac{1}{{\rm K}}\,{\rm tr}(AB)=A^a B^a$. 
1850: The Higgs field $\Phi$ is a vector in the representation space of 
1851: ${\mathcal G}$ where the generators ${\bf T}_a$ act; this space
1852: can be complex or real. The covariant derivative of the Higgs field
1853: is ${D}_\mu\Phi=(\partial_\mu-igA_\mu)\Phi$ and the Higgs
1854: field potential can be  chosen as 
1855: $U(\Phi)=\frac{\lambda}{2}\,(\Phi^\dagger\Phi-1)^2$. 
1856: The Lagrangian is invariant under the local gauge transformations,
1857: \be                              \label{UUU}
1858: \Phi\to {\rm U}\Phi,~~~~A_\mu\to {\rm U}
1859: (A_\mu+\frac{i}{g}\partial_\mu){\rm U}^{-1},
1860: \ee
1861: where ${\rm U}=\exp(i\alpha^a(x^\mu) T_a)\in{\mathcal G}$. 
1862: The field equations read
1863: \begin{align}                                         \label{YMHeqs}
1864: {\hat {\mcal D}}_\mu F^{\mu\nu}&=ig\left\{({D}^\nu\Phi)^\dagger {\bf T}_a\Phi
1865: -\Phi^\dagger{\bf T}_a{D}^\nu\Phi\right\}{\bf T}_a \,,\notag \\
1866: {D}_\mu{D}^\mu\Phi&=
1867: -\frac{\partial U}{\partial(\Phi^\dagger\Phi)}\,\Phi\,,
1868: \end{align}
1869: where $\hat{{\mcal D}}_\mu=\partial-ig[A_\mu,~~]$ is the covariant derivative in the 
1870: adjoint representation. The energy momentum tensor is 
1871: \be                              \label{Tmunu}
1872: T^\mu_\nu=-\langle F_{\nu\sigma}F^{\mu\sigma}\rangle 
1873: +({D}_\nu\Phi)^\dagger {D^\mu\Phi}
1874: +({D}^\mu\Phi)^\dagger {D_\nu\Phi}-
1875: \delta^\mu_\nu{\mcal L}\,.
1876: \ee
1877: 
1878: \subsubsection{Monopole rings}
1879: 
1880: The ring solitons in the theory \eqref{YMH} have been first constructed by 
1881: Kleihaus, Kunz and Shnir \cite{Kleihaus:2003xz},\cite{Kleihaus:2004is}
1882: in the case where ${\mathcal G}$=SU(2) and the Higgs field 
1883: is in its adjoint, $\Phi\equiv\Phi^a$, such that the gauge group
1884: generators are $3\times 3$ matrices with components 
1885: $({\bf T}_a)_{bc}=-i\epsilon_{abc}$. The fundamental solutions  in this theory 
1886: are the magnetic monopoles of 't Hooft and Polyakov 
1887: \cite{'tHooft:1974qc},\cite{Polyakov:1974ek},
1888: while the ring solitons are more general 
1889: solutions. Specifically, 
1890: in the static, axially symmetric and purely magnetic case 
1891: it is consistent to choose the following ansatz for 
1892: the fields in spherical coordinates:
1893: \begin{align}                                   \label{anz0}
1894: A_\mu dx^\mu&=(K_1 dr+(1-K_2)d\vartheta){\bf T}_\varphi
1895: +m(K_3{\bf T}_r+(1-K_4){\bf T}_\vartheta)\sin\vartheta\, d\varphi, \nonumber \\
1896: {\bf T}_a\Phi^a&=\phi_1{\bf T}_r+\phi_2{\bf T}_\vartheta\,,
1897: \end{align}
1898: where functions $K_1,K_2,K_3,K_4,\phi_1,\phi_2$ depend on $r,\vartheta$
1899: and are subject of suitable boundary conditions 
1900: at the symmetry axis and at infinity 
1901: \cite{Kleihaus:2003xz},\cite{Kleihaus:2004is}. Here 
1902: \begin{align}                                   \label{mn}
1903: {\bf T}_r&=\sin({k}\vartheta)\cos({m}\varphi) {\bf T}_1+
1904: \sin({k}\vartheta)\sin({m}\varphi){\bf T}_2
1905: +\cos({k}\vartheta){\bf T}_3, \nonumber \\
1906: {\bf T}_\vartheta&=\frac{1}{k}\,\frac{\partial}{\partial \vartheta}{\bf T}_r\,,
1907: ~~~~~~~
1908: {\bf T}_\varphi=\frac{1}{m\sin\vartheta}\,
1909: \frac{\partial}{\partial \varphi}{\bf T}_r\,,~~~
1910: \end{align}
1911: with ${k},{m}\in Z$. Using 
1912: the gauge invariant tensor
1913: \be                                            \label{hooft}
1914: {\mcal F}_{\mu\nu}=\Phi^a F^a_{\mu\nu}-\epsilon_{abc}\Phi^a D_\mu\Phi^b D_\nu\Phi^c
1915: \ee
1916: and its dual, $\tilde{{\mcal F}}_{\mu\nu}=
1917: \frac12\epsilon_{\mu\nu\alpha\beta}{\mcal F}_{\alpha\beta}$,
1918: one can define the  
1919: electric and magnetic currents, respectively, as 
1920: \be                                              \label{charge-current}
1921: j_\mu=\partial^\alpha{\mcal F}_{\alpha\mu},~~~~~~~~
1922: \tilde{j}_\mu=\partial^\alpha\tilde{{\mcal F}}_{\alpha\mu}\,.
1923: \ee   
1924: It turns out that the solutions depend crucially on values of $k,m$
1925: in \eqref{anz0}. 
1926: \begin{figure}[h]
1927: \hbox to\linewidth{\hss%
1928: \resizebox{7cm}{5cm}{\includegraphics{ring-phi.eps}}%
1929: \hspace{5mm}%
1930: \resizebox{6cm}{5cm}{\includegraphics{p1.eps}}%
1931: \hss}
1932: \caption{\small Left: Higgs field amplitude for the $k=2$, $m=3$ monopole ring solution
1933: in the  limit where the Higgs field potential is zero. $|\Phi|$ vanishes
1934: at a point in the $(\rho,z)$ plane away from the $z$-axis, which corresponds to a ring.   
1935: Right: schematic shape of charge/current distribution for this solution. 
1936: }
1937: \label{Fig2}
1938: \end{figure}
1939: In particular, their 
1940: magnetic charge is given by 
1941: \be                                 \label{Qmagn}
1942: {\mcal Q}=\frac{m}{2}\,[1-(-1)^k].
1943: \ee
1944: The following solutions are known 
1945: in the limit of vanishing Higgs potential
1946: (for a generic potential their structure is more complicated)
1947: \cite{Kleihaus:2003xz},\cite{Kleihaus:2004is}: \\
1948: ${k}=1,{m}=1$ -- the spherically symmetric 't Hooft-Polyakov monopole.  \\
1949: ${k}=1,{m}>1$ -- multimonopoles.\\
1950: ${k}>1,{m}=1,2$ -- monopole-antimonopole sequences.\\
1951: ${k}=2l\geq2,{m}\geq 3$ -- monopole {vortex rings}. \\
1952: In the first three cases the Higgs field has discrete zeros  
1953: located at the $z$-axis. 
1954: Solutions of the last type are  especially interesting in the 
1955: context of our discussion, since 
1956: zeros of the Higgs fields in this case are not discrete but continuously 
1957: distributed along a circle (for $k=2$)
1958: around the $z$-axis (see Fig.\ref{Fig2}). Solutions in this 
1959: case can be visualized as stationary rings
1960: stabilized by the magnetic energy.  The mechanism of their 
1961: stabilization is quite interesting
1962: and can be elucidated as follows \cite{Shnir:2005te,Shnir}.
1963: If one studies the profiles of the currents \eqref{charge-current} for these solutions,
1964: it turns out that  both  
1965: the magnetic charge density $\tilde{j}_0$ and the 
1966: electric current density $j_k$ have ring shape distributions,
1967: as qualitatively shown in Fig.\ref{Fig2}. 
1968: 
1969:   Although the total magnetic charge is zero, 
1970: locally the charge density is non-vanishing and 
1971: the system can be visualized  as a pair of 
1972: magnetically charged rings with opposite charge located at 
1973: $z=\pm z_0$, accompanied by a circular
1974: electric current in the $z=0$ plane.  
1975: The two magnetic rings create a magnetic field orthogonal to the $z=0$ plane. 
1976: This magnetic field forces the electric charges in the 
1977: plane to Larmore orbit, which 
1978: creates  a circular current. The Biot-Savart magnetic field produced by this current
1979: acts, in its turn, on the magnetic rings keeping them away from each other, so that 
1980: the whole system is in a self-consistent equilibrium  \cite{Shnir:2005te}
1981: (assuming the magnetic rings to be rigid). 
1982: 
1983: It is, however, unlikely that this sophisticated balance  mechanism 
1984: stabilizing the rings against contraction 
1985: could also guarantee their stability with respect to all possible 
1986: deformations. In fact, the monopole-antimonopole solution is  
1987: known to be unstable \cite{Taubes:1982ie}, while the monopole rings 
1988: can be viewed as generalizations of this solution. They are therefore likely
1989: to be saddle points of the energy functional and so they should be unstable as well.  
1990: 
1991: 
1992: 
1993: \subsubsection{Sphaleron rings}
1994: 
1995: Very recently, a similar ring construction was carried out by 
1996: Kleihaus, Kunz and Leissner \cite{Kleihaus:2008gn} 
1997: within the context of the Yang-Mills-Higgs theory \eqref{YMH}
1998: with ${\mathcal G}$=SU(2) and with the Higgs field 
1999: in its  fundamental complex doublet representation, where 
2000: ${\bf T}_a =\frac12 \tau_a$. 
2001: This theory can be viewed as the SU(2)$\times$U(1) 
2002: Weinberg-Salam theory in the limit where 
2003: the weak mixing angle vanishes and the U(1) gauge field decouples. 
2004: Kleihaus, Kunz and Leissner used exactly the same field ansatz \eqref{anz0},
2005: only modifying the Higgs field as  
2006:  \begin{align}                                   \label{anz0a}
2007: \Phi&=(\phi_1{\bf T}_r+\phi_2{\bf T}_\vartheta)\Phi_0\,,
2008: \end{align}
2009: where $\Phi_0$ is a constant 2-vector.  As in the monopole case,
2010: in this case too the solutions depend strongly on the choice of 
2011: the integers $k$ and $m$ in Eq.\eqref{mn}. The fundamental solutions in this case are 
2012: the sphalerons -- 
2013: unstable saddle point configurations that can be smoothly deformed to vacuum. 
2014: They can be characterized by the Chern-Simons number, given   
2015: by the same formula as the magnetic charge in the monopole case, 
2016: up to the factor $1/2$, so that half-integer 
2017: values are now allowed: ${\mcal Q}=m[1-(-1)^k]/4$.
2018: 
2019: Setting $k=m=1$ gives 
2020: the Klinkhamer-Manton sphaleron \cite{Klinkhamer:1984di}, in which case the 
2021: Higgs field vanishes at one point. Choosing $k=1$, $m>1$  or $k>1$, $m=1$ gives
2022: multisphalerons or sphaleron-antisphaleron
2023: solutions for which  the Higgs field has several isolated zeros
2024: located at the symmetry axis \cite{Kleihaus:2008gn}.
2025: A new type of solution arises for $k\geq 2$, $m\geq 3$, in which case the Higgs field
2026: vanishes on one or more rings centered around the symmetry axis. 
2027: In this respect these sphaleron rings are quite 
2028: analogues to the monopole rings. It is unclear at the moment whether 
2029: their existence can be qualitatively explained by a mechanism similar to that  
2030: for the monopole rings, shown in Fig.\ref{Fig2}. 
2031: 
2032: Since sphalerons are unstable objects, it is very likely that sphaleron rings
2033: are also unstable. However, similar to the sphalerons, they could perhaps
2034: be interesting physically as mediators of baryon number violating processes 
2035: \cite{Klinkhamer:1984di}. 
2036: 
2037:  
2038: 
2039: 
2040: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2041: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2042: 
2043: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2044: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2045: \section{Angular momentum and radiation in field systems}
2046: 
2047: 
2048:  
2049: 
2050: We are now passing to the spinning systems with the ultimate intention to discuss 
2051: spinning vortex loops stabilized by the 
2052: centrifugal force -- vortons. As was already said in the Introduction, 
2053: usually vortons are considered within a qualitative macroscopic description 
2054: as loops made of vortices and stabilized by rotation. 
2055: This description is suggestive, but it does not take into account
2056: the radiation damping. At the same time, 
2057: the presence of the vorton angular momentum requires 
2058: some internal motions in the system (see Fig.\ref{FigV}),  
2059: as for example circular currents, and these  are likely to  
2060: generate radiation carrying away both the energy and angular momentum. 
2061: It is therefore plausible that macroscopically constructed vortex loops will not
2062: be stationary field theory objects, but at best only quasistationary, with a 
2063: finite lifetime determined by the radiation rate. 
2064: 
2065: \begin{figure}[ht]
2066: \hbox to\linewidth{\hss%
2067: 	\resizebox{7cm}{4cm}{\includegraphics{VORTON.eps}}
2068: \hspace{5mm}%
2069:         \resizebox{7cm}{4cm}{\includegraphics{VORTON1.eps}}	
2070: \hss}
2071: \caption{One can make a loop from a vortex carrying a current $j$ and momentum $P$. 
2072: It will have an angular momentum, but it may be radiating.}
2073: \label{FigV}
2074: \end{figure}
2075: 
2076: The best way to decide whether vortons are truly stationary or only 
2077: quasistationary is to explicitly resolve the corresponding field theory equations. 
2078: The current situation  in this direction is not, however, very suggestive. 
2079: Within the original local U(1)$\times$U(1) Witten's
2080: model of superconducting cosmic strings 
2081: \cite{Witten:1984eb} vorton solutions have never been constructed.    
2082: To the best of our knowledge, the only explicit vorton solutions 
2083: have been presented by Lemperier and Shellard \cite{Lemperiere:2003yt} 
2084: within the global 
2085: version of Witten's model, and also by Battye, Cooper and Sutcliffe  
2086: \cite{Battye:2001ec} in a special limit of the same global model. 
2087: In addition, vortons in a $2+1$ dimensional field theory toy model 
2088: have been recently analyzed  \cite{Battye-kink}. 
2089: 
2090: Lemperier and Shellard \cite{Lemperiere:2003yt} considered the full 
2091: hyperbolic evolution problem for the fields, with the initial data corresponding to 
2092: a vortex loop. Evolving dynamically  this loop in time, 
2093: they saw it oscillate  around a visibly stationary equilibrium position, 
2094: and they could follow these oscillations for several dozens 
2095: characteristic periods of the system. This suggests that vortons exist and are stable
2096: against perturbations. 
2097: However, one cannot decide on these grounds  whether the equilibrium configurations 
2098: are truly stationary or only quasistationary, since the radiation damping 
2099: could be non-zero but too small to be visible in their numerics. 
2100: In addition, Lemperier and Shellard did not actually consider precisely 
2101: the global version of Witten's model (see Eq.\eqref{lag} below),
2102: but, in order to improve the numerics, 
2103: added to it a $Q$-ball type interaction 
2104: term of the form $|\phi|^6 |\sigma|^2$, where $\phi,\sigma$ are the two scalars 
2105: in the model.
2106: 
2107: 
2108: Battye, Cooper and Sutcliffe \cite{Battye:2001ec} did not study 
2109: Witten's model but minimized the energy of a non-relativistic Bose-Einstein condensate, 
2110: which  seems to be mathematically equivalent to solving equations 
2111: of Witten's model in a special limit.  
2112: They found non-trivial energy minima
2113: saturated by configurations of vorton type, which again suggests that
2114: vortons exist, at least in this limit. Moreover, this suggests   
2115: that they are indeed non-radiative --
2116: since being already in the energy minimum they cannot loose energy anymore.  
2117: It would therefore be interesting to construct these solutions in a different way,
2118: extending the analysis to the full Witten's model. 
2119: 
2120: 
2121: 
2122: 
2123: The method we shall employ below to study truly non-radiating vortons
2124:  will be to construct them as stationary solutions of the elliptic boundary 
2125: value problem obtained by separating the time variable. However, 
2126: first of all we need to  
2127: understand how in principle a non-radiating field system can have a non-zero 
2128: angular momentum. Both angular momentum and radiation are associated to some 
2129: internal 
2130: motions in the system, and it is not completely  clear how to reconcile 
2131: the presence of the former with the  absence of the latter.
2132: 
2133: \subsection{Angular momentum for stationary solitons}
2134: 
2135: In what follows we shall be considering field theory systems obeying
2136: the following four conditions: 
2137: 
2138: \noindent 
2139: (1) stationarity \\
2140: (2) finiteness of energy \\
2141: (3) axial symmetry   \\
2142: (4) non-vanishing angular momentum
2143: 
2144: 
2145: 
2146: The angular momentum is defined as the Noether charge
2147: associated to the global spacetime symmetry generated by the axial Killing
2148: vector $K=\partial/\partial\varphi$,
2149: \be                                    \label{JJ}
2150: J=\int T^0_\varphi d^3x\,.
2151: \ee
2152: Let us discuss the first three conditions. 
2153: 
2154: (1) A system is stationary if its energy momentum tensor $T^\mu_\nu$ does not depend on time.  
2155: According to the standard definition 
2156: of symmetric fields \cite{Forgacs:1979zs}, for stationary fields 
2157: the action of time translations can be compensated by 
2158: internal symmetry transformations.  
2159: 
2160: If all internal symmetries of the theory are local, then there is a gauge where
2161: the compensating symmetry transformation is trivial, so that the stationary  fields
2162: are time-independent.  
2163: We shall call them 
2164: {\it manifestly} stationary. 
2165: If the theory contain also global 
2166: internal symmetries and if the compensating symmetry transformation is global,  
2167: then its action cannot be trivialized and so the action of time translations
2168: will  be non-trivial. The fields will explicitly depend on
2169: time in this case, typically via time-dependent phases, and we shall call them 
2170: {\it non-manifestly} stationary. 
2171: 
2172: For example, in a system with two complex scalars coupled to a U(1) gauge field 
2173: one cannot gauge away simultaneously 
2174: phases of both scalars. A non-manifestly stationary field configuration will then be $\phi_1(\bx)$,
2175: $\phi_2(\bx)e^{i\omega t}$, $A_\mu(\bx)$.  However, if there is  only one
2176: scalar, then it is always possible to gauge away its time-dependent phase. 
2177: 
2178: (2) Even if $T^\mu_\nu$ is time-independent, 
2179: one can still have a constant radiation flow compensated by the 
2180: energy inflow from infinity. However, if the energy is finite, then 
2181: the fields fall-off fast enough at infinity to eliminate this 
2182: possibility.  
2183: 
2184: In principle, one can also
2185: have situations where  $T^\mu_\nu$ is time-dependent, but radiation 
2186: is nevertheless absent, as for the breathers in 1+1 dimensions \cite{Rajaraman}.    
2187: However, such cases are probably less typical and we shall not discuss them. 
2188: 
2189: 
2190: (3)  It is intuitively clear that asymmetric spinning systems will more likely 
2191: radiate than symmetric ones. 
2192: It is therefore most natural to assume spinning 
2193: non-radiating solitons to be axially symmetric. 
2194: The axially symmetry can  be manifest or non-manifest. 
2195: In fact, it appears that the axial symmetry condition can sometimes be relaxed, 
2196: but such a possibility seems to be more exotic and will be discussed below
2197: only very briefly in Sec.\ref{mon-pairs}. 
2198: 
2199: 
2200: 
2201: Let us now analyze possibilities for condition (4) 
2202: to coexist  with (1)--(3). 
2203: 
2204: 
2205: 
2206: 
2207: 
2208: 
2209: 
2210: 
2211: 
2212: 
2213: 
2214: 
2215: 
2216: 
2217: \subsubsection{The case of manifest symmetries -- no go results}
2218: In theories where {\it all} 
2219: internal symmetries are local the  
2220: stationary field are time-independent. 
2221: Can one have  an angular momentum in this case ? 
2222: 
2223: In fact, 
2224: even without an explicit time dependence 
2225: one can have 
2226: a non-vanishing field momentum expressed by the Poynting vector, 
2227: $\vec{\mcal E}\times\vec{\mcal B}$, 
2228: and this could give a contribution to the angular momentum,  
2229: \be                                  \label{Poynt}
2230: \int \vec{r}\times(\vec{\mcal E}\times\vec{\mcal B}) \, d^3 x \;.
2231: \ee 
2232: If the theory contains gauged scalars, they will give an additional 
2233: contribution. If there exists a stationary, globally regular, finite energy
2234: on-shell configuration for which the
2235: integral \eqref{JJ} is non-zero, this would correspond to a rotating and non-radiating 
2236: soliton.  The existence of such solutions is  not {\it a priori} excluded.  
2237: However, it turns out that for a number of physically interesting cases
2238: they can be ruled out.  
2239: 
2240: Specifically, since the time translations and spatial rotations commute, 
2241: the system is not only manifestly stationary but also manifestly axially symmetric
2242: \cite{Forgacs:1979zs}. 
2243: It turns out that the latter condition allows one to transform the volume integral in  
2244: Eqs.\eqref{JJ}  to a {surface integral},
2245: which is often enough to conclude that $J=0$. 
2246: 
2247: To me more precise,   let  us consider the Yang-Mills-Higgs theory \eqref{YMH}. 
2248: If the fields $(A_\mu,\Phi)$ 
2249: are manifestly invariant under the action of a Killing symmetry generator 
2250: $K=\partial/\partial s$, then there is a gauge where they do not depend
2251: on the corresponding spacetime coordinate $s$ and 
2252: their Lie 
2253: derivatives along $K$ vanish, ${\mcal L}_K A_\mu={\mcal L}_K\Phi=0$.
2254: In some other gauge the fields could depend on $s$, but only in such a way that 
2255: \cite{Forgacs:1979zs}
2256: \be                                         \label{sym}
2257: {\mcal L}_K A_\mu=\hat{{\mcal D}}_\mu W(K)\,,~~~~~~
2258: {\mcal L}_K\Phi= ig W(K)\Phi\,.
2259: \ee
2260: Here 
2261: $W(K)$ takes its values in the Lie algebra of the gauge group and
2262: transforms as connection under gauge transformations.
2263: If $K=\partial/\partial\varphi$ is the axial Killing vector and 
2264: $W_\varphi\equiv W(K)$, then in spherical coordinates these conditions reduce to 
2265: \be                                         \label{sym1}
2266: \partial_\varphi A_\mu=\hat{{\mcal D}}_\mu W_\varphi\,,~~~~~~
2267: \partial_\varphi \Phi= ig W_\varphi\Phi\,.
2268: \ee
2269: If these conditions are fulfilled, then it is 
2270: straightforward to check with the field equations  \eqref{YMHeqs}
2271: that the $T^0_\varphi$ component of the energy-momentum tensor \eqref{Tmunu}
2272: has a total derivative structure 
2273: \cite{VanderBij:2001nm}, \cite{Volkov:2003ew}, 
2274:  \be                              \label{Tmu}
2275: T^0_\varphi=\partial_k\langle (W_\varphi-A_\varphi)F^{0k}\rangle\,.
2276: \ee
2277: Since both $A_\varphi$ and $W_\varphi$ transform as connections 
2278: under gauge transformations, their difference is gauge covariant, so that 
2279: this formula is gauge invariant. 
2280: For globally regular solutions the 
2281: electric field ${\mathcal E}^k=F^{0k}$ 
2282: is everywhere bounded, which allows one to transform  
2283: the volume integral
2284: of $T^0_\varphi$ to a surface integral
2285: over the boundary of $\mathbb{R}^3$ at infinity. This gives the surface 
2286: integral representation for the angular momentum
2287: \be                                             \label{JJJ}
2288: J=\oint \langle
2289: (W_\varphi-A_\varphi)\vec{{\mcal E}}\rangle\, \vec{dS}\,.
2290: \ee
2291: This formula imposes rather strong restrictions on the existence of spinning
2292: solitons. 
2293:  For example, it shows 
2294: that if the electric field 
2295: ${\mcal E}$ decays at infinity faster than $1/r^2$, such that the electric charge 
2296: is zero, then $J=0$. A non-zero electric charge is therefore necessary 
2297: to have a non-zero angular momentum \cite{VanderBij:2001nm}.   
2298: Similarly, 
2299: using the asymptotic conditions in the far field zone, 
2300: one can show that for a number of important cases 
2301: the asymptotic behavior of the fields does not allow for 
2302: the surface integral to be non-zero. 
2303: 
2304: \begin{figure}[h]
2305: \hbox to\linewidth{\hss%
2306:   \resizebox{10cm}{5cm}{\includegraphics{126-new.eps}}%
2307: \hss}
2308: \caption{\small 
2309: Do static solitons admit stationary, spinning generalizations ?
2310: }
2311: \label{FigDo}
2312: \end{figure}
2313: 
2314: 
2315: 
2316: Specifically, one can consider any known static, spherically symmetric soliton,
2317: as for example the 't Hooft-Polyakov monopole, and ask if it admits stationary,
2318: spinning generalizations (see Fig.\ref{FigDo}) ? For example, 
2319: the static Schwarzschild black hole solution of Einstein's equation in General Relativity
2320: can be generalized to the manifestly stationary, spinning Kerr black hole. One can wonder
2321: if similar generalizations are possible for solitons 
2322: in non-linear field theories in Minkowski space.
2323: If they exist, then in the far field zone their 
2324: fields should approach those of the original static soliton, so that they should be
2325:  spherically symmetric, up to small corrections which could 
2326: contribute to the surface integral \eqref{JJJ}.  Since the 
2327: corrections are small, their most general form can be determined
2328: from the linearized field equations \cite{Volkov:2003ew}. 
2329: 
2330: Surprisingly, 
2331: one discovers in this way 
2332: that none of the well known solitons in the gauge field theory \eqref{YMH}
2333: with gauge group SU(2), as
2334: for example 
2335: the magnetic monopoles of 't~Hooft-Polyakov 
2336: \cite{'tHooft:1974qc}, \cite{Polyakov:1974ek}, 
2337: dyons of Julia-Zee \cite{Julia:1975ff}, and sphalerons
2338: of Klinkhamer-Manton \cite{Klinkhamer:1984di},
2339: admit spinning generalizations
2340:  within the manifestly stationary and manifestly axisymmetric sector 
2341: \cite{VanderBij:2001nm},
2342: \cite{vanderBij:2002sq},
2343: \cite{Volkov:2003ew}.
2344: The picture in Fig.\ref{FigDo} thus does not apply for these solitons. 
2345: If the gauge symmetry is completely broken in the Higgs vacuum,
2346: as for the sphalerons, then the fields approach their vacuum values
2347: exponentially fast and the surface integral vanishes. For 
2348: the monopoles and dyons there is a long-range field associated with the unbroken
2349: U(1), so that some additional analysis is required to show  
2350: that in the asymptotic region there are actually no field modes  giving 
2351: a non-zero contribution to the surface integral \cite{Volkov:2003ew}. 
2352: 
2353: As a result, the best known SU(2) solitons cannot spin in the 
2354: manifestly stationary and manifestly axisymmetric sector. 
2355: This restricts rather strongly 
2356: the existence of spinning solitons with manifest symmetries,
2357: although does not rule them out completely. 
2358: Such solutions might exist in theories with other gauge groups.  
2359: Their explicit 
2360: examples in U(1) gauge field theories will be presented below. 
2361: Non-manifestly symmetric rotational excitations  could perhaps
2362: exist for the sphalerons, since only an SU(2) part of their  U(2) 
2363: internal  symmetry is gauged. A more exotic possibility to have stationary
2364: rotation without axial symmetry will be briefly discussed in Sec.\ref{mon-pairs}. 
2365: 
2366: 
2367: \subsubsection{The case of non-manifest symmetries}
2368: More general possibilities to have spinning solitons arise in theories where 
2369: {\it not all} internal symmetries are 
2370: local, since in this case the conditions \eqref{sym} can be generalized as 
2371: %\be                                         \label{sym1}
2372: %{\mcal L}_K\Psi=W(K)\Psi+T(K)\Psi\,,
2373: %\ee  
2374: \be                                         \label{sym1a}
2375: {\mcal L}_K A_\mu=\hat{{\mcal D}}_\mu W(K)\,,~~~~~~
2376: {\mcal L}_K\Phi= igW(K)\Phi+iT(K)\Phi\,,
2377: \ee
2378: where $T(K)$ is a function of the {\it global} 
2379: symmetry generators. One can gauge away $W(K)$ 
2380: but not $T(K)$, so that the action of the Killing symmetry 
2381: will always be non-trivial, since $\Phi$ will 
2382: depend explicitly on the corresponding spacetime 
2383: coordinate $s$ as
2384: \be
2385: \Phi(s)=e^{is T(K)}\Phi_0\,.
2386: \ee
2387: This means that the $s$-dependence is equivalent
2388: to a sequence  of internal symmetry transformations, in which case
2389: the invariant objects like $T^\mu_\nu$ will not 
2390: depend on $s$ at all. Such a symmetry can be called 
2391: {\it non-manifest}. Non-manifestly stationary and non-manifestly 
2392: axisymmetric fields would typically depend on 
2393: $t,\varphi$ via the complex phase factor 
2394: \be                                                               \label{om}
2395: \exp\{i(\omega t + m\varphi)\},
2396: \ee
2397: in which case one can say that it is the phase that spins. 
2398: The angular momentum 
2399: in this case cannot be totally 
2400: expressed by a surface integral 
2401: and will contain a volume integral contribution, in which case
2402: typically  $J\sim \omega m$. 
2403: 
2404: 
2405: As a result, one can have $J\neq 0$. However, 
2406: the absence of radiation is not yet guaranteed. 
2407: Separating the time variable makes 
2408: the equations elliptic, but with the mass term(s) modified as 
2409: \be
2410: M^2\to M^2-\omega^2,
2411: \ee
2412: where $M^2$ collectively denotes masses of the 
2413: field excitations in the asymptotic zone.  
2414: Now, if the equations admit globally regular solutions with 
2415: \be                                       \label{rad}
2416: \omega^2<M^2
2417: \ee
2418:  then these
2419: solutions will behave asymptotically as $\exp\{-\sqrt{M^2-\omega^2}~ r\}$ 
2420: and there will be no radiation.
2421: On the other hand, if $\omega^2>M^2$ then solutions in the 
2422: asymptotic region will oscillate as 
2423: $\exp\{\pm i\sqrt{\omega^2-M^2}~ r\}$ thus showing the presence
2424: of the ingoing and outgoing radiation, even though $T^\mu_\nu$ is time-independent,
2425: so that 
2426: the total energy will be infinite.  
2427: As a result, the no-radiation condition \eqref{rad} becomes crucial in this case.
2428: 
2429: \subsubsection{Spinning solitons as solutions of elliptic equations}
2430: 
2431: Summarizing the above discussion, one can conclude that the 
2432: existence of non-radiating spinning solitons, although not forbidden,  
2433: is not guaranteed either.  
2434: Even for known static solitons their spinning 
2435: generalizations may or may not exist,
2436: and if they do exist, this should be considered 
2437: as something exceptional rather than the general rule.
2438: At first view, such a conclusion may seem to contradict our experience, since normally 
2439: one knows that compact objects can spin. 
2440: However, we are not saying
2441: that generic field theory solitons cannot spin. They always can, but
2442: it seems that  they should generically radiate at the same time. 
2443:  One can always `give a kick' 
2444: to any static soliton, as for example to the magnetic monopole
2445: or to any of the knotted solitons of Faddeev-Skyrme shown in Fig.\ref{Fig:knots},
2446: so that they will start spinning. However, at the same time 
2447: they will start radiating away all the received energy and angular momentum,
2448: till they relax back to the original static, non-spinning configuration.
2449: It seems that 
2450: `spinning and radiating'  represents the generic way the  
2451: field systems behave, while `spinning without radiating' should rather be
2452: considered as something exceptional, possible only  
2453: in some special field theory models.   
2454: 
2455: Perhaps  the best way to really establish the
2456: existence of spinning and non-radiating solitons 
2457: would be to construct them as solutions of the
2458: elliptic boundary value problem obtained by separating the time variable. 
2459: In gauge field theory, as for example in the local 
2460: U(1)$\times$U(1) Witten's model 
2461: \cite{Witten:1984eb}, one could look for manifestly stationary 
2462: and axisymmetric vortons, 
2463: in which case there is no radiation.
2464: However, since the angular momentum in this case can be 
2465: expressed by the surface integral \eqref{JJJ} and so is
2466: determined only by the asymptotic behavior of the fields, 
2467: there are higher chances that it could  vanish.  
2468: For global field theories 
2469: one could consider systems with non-manifest symmetries, 
2470: in which cases
2471: there are better chances to have a non-zero angular momentum, 
2472: but also higher chances to have
2473: radiation, unless the no-radiation condition \eqref{rad} 
2474: is fulfilled. 
2475: 
2476: \section{Explicit examples of stationary spinning solitons}
2477: 
2478: Before coming to vortons, one can wonder, in view of the above 
2479: discussion, if there exist at all any 
2480: known examples of non-radiating spinning 
2481: solitons. As we said, for the SU(2) magnetic monopoles,
2482: dyons and sphalerons there are no spinning generalizations,
2483: at least in the manifestly stationary and 
2484: manifestly axisymmetric case. Nevertheless, explicit examples of 
2485: spinning solitons in Minkowski space in $3+1$ dimensions exist
2486: and below we shall review all known solutions of this type. These are
2487: the spinning $Q$-balls, spinning Skyrmions and also rotating 
2488: monopole-antimonopole pairs, apart from the vortons. 
2489: Interestingly, it seems that spinning 
2490: solitons are in some sense more easily constructed in curved space,
2491: because the spinning degrees of freedom can be naturally associated
2492: to the non-radiative  dipole moment of 
2493: gravitational field. For this reason 
2494: a number of articles cited below actually describe the  Minkowski space 
2495: spinning solitons 
2496: only as a special limit of the more general, self-gravitating 
2497: configurations. 
2498: 
2499: 
2500: 
2501: \subsection{$Q$-balls \label{Qballs}}
2502: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2503: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2504: 
2505: This is an important for our discussion example  which 
2506: shares many features with the vortons. At the same time, 
2507: spinning $Q$-balls are simpler than vortons,
2508: and so 
2509: they can be used to introduce a number of notions to be applied later. 
2510: We shall therefore discuss them  in some detail. 
2511: 
2512: $Q$-balls have been introduced by Coleman \cite{Coleman:1985ki}.
2513: These 
2514: are non-topological solitons \cite{Lee:1991ax} 
2515: found in a theory  with 
2516:  a single complex scalar field with the Lagrangian 
2517: density 
2518: \be                   \label{lQ}
2519: {\mathcal L}_{Q}[\Phi]=  
2520: \partial_\mu \Phi^\ast \partial^\mu \Phi  -U(|\Phi|),
2521: \ee 
2522: the corresponding field equation being 
2523: \be                                   \label{Qe00}
2524: \partial_\mu\partial^\mu\Phi+
2525: \frac{\partial U}{\partial |\Phi|^2}\,\Phi=0,
2526: \ee
2527: while the energy-momentum tensor 
2528: \be
2529: T_{\mu\nu}=\partial_\mu\Phi^\ast\partial_\nu\Phi+
2530: \partial_\nu\Phi^\ast\partial_\mu\Phi-g_{\mu\nu}{\mcal L}_{Q}.
2531: \ee
2532: The  potential $U$ should have the absolute minimum, $U(0)=0$, and 
2533: should also satisfy the condition 
2534: \be                                \label{omega+}
2535: \omega_{\rm -}^2= \min_{f}\frac{U(f)}{f^2}<
2536: \omega_{\rm +}^2= \frac12\left.\frac{d^2U}{df^2}\right|_{f=0}, 
2537: \ee
2538: whose meaning will be explained below. 
2539: The value of $\omega_{+}$ determines the mass of the 
2540: field quanta, $M=\omega_{+}$.  
2541: For the condition \eqref{omega+} to be fulfilled
2542: $U$ should (if it is even in $|\Phi|$)  
2543: contain powers of $|\Phi|$ higher than four, 
2544: which means that the theory cannot be renormalizable \cite{Coleman:1985ki}. 
2545: A convenient choice is
2546: \be                           \label{pQ}
2547:  U=\lambda |\Phi|^2(|\Phi|^4-a |\Phi|^2+b),
2548: \ee 
2549: where $\lambda$, $a$, $b$ are positive 
2550: constants, so that $\omega_{+}^2=\lambda b$
2551: and $\omega_{-}^2=\omega_{+}^2(1-a^2/4b)$. 
2552: In our numerics below we shall always 
2553: choose $\lambda=1$, $a=2$, $b=1.1$ \cite{Volkov:2002aj}, which is not a restriction 
2554: (for $\omega_{-}\neq 0$) 
2555: since $\lambda,a$ can be changed by rescaling the coordinates and field, while the mass 
2556: $M=\sqrt{\lambda b}$ enters 
2557: the field equation \eqref{Qe} only in the combination $M^2-\omega^2$ where
2558: $\omega$ is another free parameter. 
2559: 
2560: 
2561: The global invariance of the theory under 
2562: $\Phi\to \Phi e^{i \alpha }$ implies the 
2563: conservation of the Noether charge 
2564: \be                                                      \label{Noether}
2565: Q=i\int (\partial_t\Phi^\ast\Phi-\Phi^\ast\partial_t\Phi) d^3x\,.
2566: \ee
2567: The scaling argument of Derrick  applies for the theory \eqref{lQ} and 
2568: rules out all static solutions with finite energy. Therefore, in order to  
2569: circumvent this argument, solutions should depend  on time,
2570: \be                    \label{QQQ}
2571: \Phi=\phi(\bx)e^{i\omega t},
2572: \ee 
2573: in which case 
2574: the Noether charge is 
2575: \be                               \label{QQ}
2576: Q=2\omega \int \phi^2 d^3 {\bf x} \equiv 2\omega \ssigma.
2577: \ee
2578: In what follows we shall assume $\omega$ to be positive.  
2579: More explicitly, with 
2580: \be
2581: E_2=\int(\nabla\phi)^2d^3 {\bf x},~~~~~ 
2582: E_0=\int U d^3 {\bf x},
2583: \ee 
2584: the Lagrangian reads
2585: \be
2586: L= \int {\mcal L}_{Q}\, d^3 {\bf x} =\omega^2\ssigma-E_0-E_2.
2587: \ee
2588: Under scale transformations $\bx\to \Lambda\bx$ one has  
2589: $
2590: L\to \Lambda^3(\omega^2\ssigma-E_0)-\Lambda E_2
2591: $
2592: so that $L$ will be stationary for $\Lambda=1$ if the virial relation 
2593: is fulfilled,
2594: \be                                        \label{Qvirial}
2595: 3\omega^2\ssigma=3E_0+E_2,
2596: \ee
2597: which is only possible (for $E_0\neq 0$, $E_2\neq 0$) if $\omega\neq 0$. 
2598: 
2599: $Q$-balls are finite energy solutions of the Lagrangian 
2600: field equations for the ansatz \eqref{QQQ},
2601: \be                                   \label{Qe}
2602: \left(\Delta+\omega^2\right)\phi=
2603: \frac{\partial U}{\partial |\phi|^2}\,\phi.
2604: \ee
2605: Even though $\Phi$ depends on time, 
2606: $T_{\mu\nu}$ is time independent --
2607: the system is non-manifestly stationary. 
2608: Equivalently, $Q$-balls can be obtained 
2609: by minimizing the total energy
2610: \be                       \label{EneQ}
2611: E=\int T^0_0 d^3\bx=\int(\omega^2|\phi|^2+|\nabla\phi|^2+U) 
2612: d^3 {\bf x}\equiv \omega^2\ssigma+E_0+E_2
2613: \ee
2614: by keeping fixed the charge $Q$. 
2615: Indeed, rewriting the energy as 
2616: \be                       \label{EneQ1}
2617: E=\frac{Q^2}{4\ssigma}+E_0+E_2
2618: \ee
2619: shows that minimizing   $E$ with $Q$ fixed is equivalent to 
2620: extremizing $L$,
2621: since 
2622: \be                         \label{deltaE}
2623: \delta E=-\frac{Q^2}{4\ssigma^2}\,\delta\ssigma+\delta (E_0+E_2)=
2624: -\omega^2\delta\ssigma+\delta (E_0+E_2)=
2625: -\delta L\,,
2626: \ee
2627: so that the on-shell condition $\delta L=0$ follows from $\delta E=0$. 
2628: It is also instructive to see how the same thing comes about within the 
2629: Lagrange multiplier method. Introducing 
2630: \begin{align}                      \label{EQ}
2631: E_Q&=\int(\omega^2|\phi|^2+|\nabla\phi|^2+U) 
2632: d^3 {\bf x}
2633: +\mu\left(2\omega\int|\phi|^2 d^3\bx-Q\right), 
2634: \end{align}
2635: the condition $\partial E_Q/\partial\mu=0$ fixes the charge,   
2636: the condition $\partial E_Q/\partial\omega=0$ 
2637: gives $\mu=-\omega$ insuring  that $E_Q=L+const.$, so that the condition 
2638: $\delta E_Q/\delta\phi=0$ reproduces the field equation. 
2639: 
2640: Equivalently, one can minimize the potential energy of the field, 
2641: $E_0+E_2$, by keeping fixed 
2642: \be                           \label{NN}
2643: \ssigma=\int |\phi|^2 d^3\bx
2644: \ee
2645: via extremizing the functional 
2646: \begin{align}                    \label{NE}
2647: E_\ssigma&=\int(|\nabla\phi|^2+U) 
2648: d^3 {\bf x}
2649: +\mu_1\left(\int|\phi|^2 d^3\bx-\ssigma\right).
2650: \end{align}
2651: The condition $\partial E_\ssigma/\partial\mu_1=0$ imposes the constraint \eqref{NN},    
2652: while the condition 
2653: $\delta E_\ssigma/\delta\phi=0$ reproduces the field equation \eqref{Qe} with 
2654: $\omega^2=-\mu_1$. The time dependence of the solutions is very implicit 
2655: in this approach, since $\omega$ appears only as a Lagrange multiplier. 
2656: It follows from Eq.\eqref{deltaE} that 
2657: \be                             \label{omega2}
2658: \omega^2=\frac{\partial(E_0+E_2)}{\partial \ssigma}, 
2659: \ee
2660: so that 
2661: fixing $\ssigma$ fixes also the charge 
2662: $Q=2\omega(\ssigma)\ssigma$.  
2663: In the non-relativistic theory  $\ssigma$ 
2664: can be viewed  as the particle number (see Sec.\ref{VS} below). 
2665: 
2666: 
2667: Let us also note that 
2668: using the charge definition 
2669: \eqref{QQ} and the virial relation \eqref{Qvirial}
2670: the energy \eqref{EneQ} can be rewritten   as
2671: \be                             \label{Eomega}
2672: E=\omega Q+\frac23\,E_2\,.
2673: \ee
2674: 
2675: 
2676: $Q$-balls could exist in the supersymmetric extensions of Standard Model 
2677: \cite{Kusenko:1997zq} and could perhaps contribute to    
2678: the dark matter \cite{Kusenko:1997si}.
2679: 
2680:  
2681: 
2682: 
2683: \subsubsection{Non-spinning $Q$-balls}
2684: 
2685: Let us briefly  consider the simplest  spherically symmetric 
2686: $Q$-balls \cite{Coleman:1985ki}, since this will help to understand 
2687: more complex, spinning solutions.  
2688: Setting 
2689: $\Phi=e^{i\omega }f(r)$, the 
2690: real amplitude $f(r)$ satisfies the equation 
2691: \be                                   \label{Qeq0}
2692: f^{\prime\prime}+\frac2r\,f^\prime+\omega^2f=
2693: \frac12\frac{\partial U}{\partial f}\,.
2694: \ee
2695: For the energy to be finite
2696: $f$ should vanish at infinity, so that  
2697: for large $r$ one has 
2698: \be                                    \label{larger}
2699: f\sim\frac{1}{r}\,\exp\{-\sqrt{M^2-\omega^2}\,r\}.
2700: \ee
2701: 
2702: \begin{figure}[h]
2703: \hbox to\linewidth{\hss%
2704:   \psfrag{x}{$r$}
2705:   \psfrag{n=0}{$n=0$}
2706:   \psfrag{n=1}{$n=1$}
2707:   \psfrag{n=2}{$n=2$}
2708:   \resizebox{7cm}{6cm}{\includegraphics{Qn.eps}}%
2709: \hspace{1mm}
2710:   \psfrag{x}{$f$}
2711:   \psfrag{f0}{$f_0$}
2712:   \psfrag{fi}{$f_\infty=0$}
2713:   \psfrag{y}{$U_{\rm eff}$}
2714:   \psfrag{o0}{$A$}
2715:  \psfrag{B}{$B$}
2716:  \psfrag{C}{$C$}
2717:  \psfrag{D}{$D$}
2718:   \psfrag{o1}{$\omega_{+}$}
2719:   \psfrag{o2}{$\omega_{-}$}
2720:   \resizebox{7cm}{6cm}{\includegraphics{Qpot.eps}}%
2721: \hss}
2722: \caption{Left: the amplitude $f(r)$ for the $n=0,1,2$ spherically symmetric $Q$-balls 
2723: with $\omega=0.2$. Right: the effective potential $U_{\rm eff}$ in Eq.\eqref{fric}
2724: for three values of $\omega$.}
2725: \label{FIG:Q}
2726: \end{figure}  
2727: 
2728: 
2729: 
2730: 
2731: Solutions of Eq.\eqref{Qeq0} comprise an infinite family 
2732: labeled by an integer 
2733: $n=0,1,\ldots $ counting the 
2734: nodes of $f(r)$ \cite{Volkov:2002aj} (see Fig.\ref{FIG:Q}). 
2735:  The energy increases with $n$. 
2736: These solutions admit a simple qualitative interpretation 
2737: \cite{Coleman:1985ki}, 
2738: since Eq.\eqref{Qeq0} is then equivalent to 
2739: \be                                   \label{fric}
2740: f^{\prime 2}+U_{\rm eff}(f)=\tilde{E}-4\int_0^r f^{\prime 2}\,\frac{dr}{r}\,.
2741: \ee
2742: This describes a particle moving with friction in one-dimensional potential
2743: $U_{\rm eff}(f)=\omega^2 f^2-U(f)$ shown in Fig.\ref{FIG:Q}, 
2744: the integration constant $\tilde{E}$ playing a role of the total energy.
2745: At the `moment' $r=0$ the `particle' rests 
2746: at a point with coordinate $f_0$ 
2747: close to the potential maximum $A$ (see Fig.\ref{FIG:Q}), so that its energy is 
2748: $\tilde{E}=U_{\rm eff}(f_0)$.  Then it starts moving to the left,  
2749: dissipating  its energy as it goes. One can  adjust the value of $f_0$ so that 
2750: for $r\to\infty$ it dissipates all its energy and arrives at the local maximum $B$ 
2751: of the potential with zero total energy to rest there -- either directly or after 
2752: $n$ oscillations between the 
2753: two potential hills. 
2754: 
2755: It follows that $\omega$ should belong to the interval 
2756: $\omega_{-}<\omega < \omega_{+}$, the condition 
2757: Eq.\eqref{omega+} making  sure that this interval is non-empty. 
2758:  $Q$-balls become large as $\omega\to\omega_{\pm}$, their charge and energy
2759: growing without bounds. 
2760: 
2761: As $\omega\to\omega_{-}$ the maximum $A$ of the potential descends
2762: towards the position $D$. The `particle' then stays for a `long time' at $D$,
2763: till the friction term becomes suppressed by the $1/r$ factor. 
2764: Then it starts moving, crosses the potential well in a finite `time' $\Delta r$ 
2765: and asymptotically approaches the potential maximum $B$. In the simplest $n=0$ case 
2766: solutions in this limit  
2767: can be described by a smoothed step-function, 
2768: \be                       \label{thin}
2769: f(r)\approx f_0\Theta(R-r),
2770: \ee 
2771: where $R\to\infty$ as $\omega\to\omega_{-}$ and  
2772: $f^\prime\approx f_0/\Delta r$ 
2773: in a region of a fixed size $\Delta r\ll R$ around $r=R$.  
2774: This is sometimes called 
2775: $Q$-balls in the thin wall approximation \cite{Coleman:1985ki}. 
2776: Their charge $Q\sim R^3$
2777: while $E_2\sim R^2$ so that for large $R$ 
2778: one can neglect the second term in Eq.\eqref{Eomega},
2779: which gives for their energy 
2780: %\be                                       \label{Easympt}
2781: $E=\omega_{-} Q$.
2782: %\ee
2783: The same value is obtained inserting \eqref{thin} to Eq.\eqref{EneQ1},
2784: neglecting the $E_2$ term and minimizing with respect to $R,f_0$. 
2785: 
2786: 
2787: When $\omega\to\omega_{+}$ the local minimum $C$
2788: of the potential approaches the local maximum $B$. The potential 
2789: in their vicinity becomes approximately 
2790: $U_{\rm eff}\approx f^2(f^2-M_\omega^2)$ with $M_\omega^2=\omega_{+}^2-\omega^2$. 
2791: Inserting this to Eq.\eqref{fric} gives $Q$-balls in the thick wall
2792: approximation \cite{Kusenko}, 
2793: \be                          \label{thick}
2794: f(r)\approx \sqrt{2}M_\omega\, y(M_\omega r),
2795: \ee
2796: where $y(x)$ fulfills 
2797: $y''+(2/x)y'+(y^2-1)y=0$. Solutions of this 
2798: equation satisfy 
2799: $y(0)=y_n$ and 
2800: $y\sim e^{-x}$ for large $x$ (one has $y_0=4.33$).
2801: Although $f\sim M_\omega$ is small in this case, 
2802: the typical configuration radius $R\sim 1/M_\omega$ is large. The charge 
2803: is also large, $Q\sim   1/M_\omega$, while $E_2\sim  M_\omega$,
2804: so that  
2805: one can again neglect the second term in Eq.\eqref{Eomega},
2806: which gives for the energy 
2807: %\be                                       \label{Easympt}
2808: $E=\omega_{+} Q$.
2809: %\ee
2810: 
2811: \subsubsection{Spinning $Q$-balls}
2812: 
2813: The spinning, axially symmetric generalizations for the 
2814: spherically symmetric $Q$-balls 
2815: have been constructed 
2816:  by Volkov and Wohnert \cite{Volkov:2002aj} 
2817: and have been analysed also 
2818: in Refs.\cite{Axenides:2001pi},
2819: \cite{Kleihaus:2005me},
2820: \cite{Kleihaus:2007vk},
2821: \cite{Brihaye:2007tn}. 
2822: %Axenides $et ~al$ \cite{Axenides:2001pi}
2823: %studied them 
2824: %within the energy minimization method. 
2825: These {non-manifestly} stationary and non-manifestly axisymmetric
2826: solutions provide the first explicitly constructed
2827: example of stationary, spinning solitons in Minkowski 
2828: space in a relativistic field theory in $3+1$ spacetime dimensions. 
2829: They have the typical ring structure, very similar to that  for vortons,
2830: and they can also have the same topology 
2831: as vortons. They admit simple generalizations to the gauged case. 
2832: Spinning $Q$-balls thus provide a simple prototype example of vortons. 
2833: Their analogs also exist  in the context of non-linear optics
2834: \cite{mihalache}, where they describe spinning light pulses
2835: (see Sec.\ref{light} below).   
2836: 
2837: 
2838: Spinning $Q$-balls have a spinning phase:  
2839: the field being non-manifestly stationary and non-manifestly
2840: axisymmetric, 
2841: \be
2842: \Phi=e^{i(\omega t +m\varphi)}f(r,\vartheta).
2843: \ee
2844: The real field amplitude $f(r,\vartheta)$ satisfies the equation 
2845: \be                                   \label{Qeq}
2846: \left(\Delta-\frac{m^2}{r^2\sin^2\vartheta}+\omega^2\right)f=
2847: \frac12\frac{\partial U}{\partial f}\,.
2848: \ee
2849: For $m\neq 0$ the energy will be finite if only  $f$ vanishes at the symmetry axis,
2850: while at infinity the asymptotic behavior \eqref{larger} still applies. 
2851: The energy momentum $T^\mu_\nu$ depends only on $r,\vartheta$
2852: and the angular momentum is 
2853: \be
2854: J=\int T^0_\varphi\, d^3 {\bf x} =2m\omega\int f^2 d^3 {\bf x}=2m\omega\ssigma,
2855: \ee
2856: so that it is classically quantized as 
2857: \be                               \label{JQ}
2858: J=mQ\,.
2859: \ee
2860: In view of this relation, 
2861: spinning $Q$-balls correspond to minima of 
2862: energy with fixed angular momentum. They could therefore be obtained
2863: by extremizing 
2864: \be                                                   \label{EJ}
2865: E=\frac{J^2}{4m^2\int f^2 \,d^3\bx}+\int\left((\nabla f)^2
2866: +\frac{m^2 f^2}{r^2\sin^2\vartheta}+U\right)d^3\bx
2867: \ee
2868: with constant $J,m$, the corresponding extremum condition being given by
2869:  Eq.\eqref{Qeq}. A direct minimization of this functional 
2870: carried out in Ref.\cite{Axenides:2001pi} 
2871: (although for a different choice of the potential $U$)
2872: suggests  that its extrema exist but seem to have negative directions. 
2873: One should therefore integrate the field equation to construct these solutions. 
2874: 
2875: 
2876: 
2877: \begin{figure}[ht]
2878: \hbox to\linewidth{\hss%
2879: 	\resizebox{6cm}{6cm}{\includegraphics{X-Q-balls-even.eps}}
2880: \hspace{1mm}%
2881:         \resizebox{6cm}{6cm}{\includegraphics{E-Qballs-even.eps}}
2882: \hspace{1mm}%
2883:         \resizebox{4cm}{4cm}{\includegraphics{Qball-even-constant-E=0,65-new.eps}}	
2884: \hss}
2885: 	\caption{The amplitude $f(r,\vartheta)$, energy density $T^0_0$
2886: and the energy isosurface with $T^0_0=0.65$  
2887: for the $1^{+}$ spinning $Q$-ball with $\omega=0.9$.}
2888: \label{FigQ1}
2889: \end{figure}
2890: 
2891: 
2892: 
2893: Taking into account the boundary conditions at the $z$-axis and at infinity, 
2894: Eq.\eqref{Qeq} with $m\neq 0$ admits 
2895: two different types of solutions determined by the 
2896: behaviour of $f$ under $z\to-z$ \cite{Volkov:2002aj}. 
2897: 
2898: For even-parity solutions, called $m^{+}$, 
2899: the amplitude $f(r,\vartheta)$, energy-momentum and charge densities are 
2900: maximal in the equatorial plane and the energy is  concentrated
2901: in a toroidal region encircling the $z$-axis (see Fig.\ref{FigQ1}). 
2902: 
2903: For odd-parity solutions, called $m^{-}$, 
2904: the amplitude $f(r,\vartheta)$
2905: vanishes in the equatorial plane, while the 
2906: energy-momentum and charge densities  
2907: show two maxima located symmetrically with 
2908: respect to the plane, so that the solutions 
2909: exhibit in this case a double torus or dumbbell-like 
2910: structure (see Fig.\ref{FigQ2}). 
2911: 
2912: 
2913: \begin{figure}[ht]
2914: \hbox to\linewidth{\hss%
2915: 	\resizebox{6cm}{6cm}{\includegraphics{X-Q-balls-odd.eps}}
2916: \hspace{1mm}%
2917:         \resizebox{6cm}{6cm}{\includegraphics{E-Qballs-Pm.eps}}
2918: \hspace{1mm}%
2919:         \resizebox{4cm}{5cm}{\includegraphics{Qball-odd-constant-E=0,65-new.eps}}	
2920: \hss}
2921: 	\caption{The amplitude $f(r,\vartheta)$, energy density $T^0_0$
2922: and the energy isosurface with $T^0_0=0.65$  
2923: for the $1^{-}$ spinning $Q$-ball with $\omega=0.9$.}
2924: \label{FigQ2}
2925: \end{figure}
2926: 
2927: 
2928: Apart from choosing the parity and the value of $m$, 
2929: solving the differential equation \eqref{Qeq} also 
2930: requires choosing $\omega^2$ as an input parameter,
2931: the energy and charge being computed from the numerical output.
2932: As in the spherically symmetric case,  
2933: spinning solutions are also found to 
2934: exist only in a finite frequency range, 
2935: $\omega_{-} < \omega< \omega_{+}$. 
2936: Here $\omega_{+}=M$ as in Eq.\eqref{omega+}, while $\omega_{-}$ seems to be
2937: $m$-dependent. 
2938: As $\omega$ approaches the limiting values, 
2939: the energy and charge seem to grow
2940: without bounds (see Fig.\ref{FigQ3}), while somewhere in between there is 
2941: a critical value of frequency, $\omega_{\rm crit}$, 
2942: for which  both $E(\omega)$ and $Q(\omega)$ attain 
2943: their minimal values. This behaviour is found for all $m$ \cite{Kleihaus:2005me} 
2944: and it is  qualitatively the same as in the $m=0$ case. 
2945: 
2946: The existence of a minimal value of the $Q$-ball charge implies that 
2947: the angular momentum \eqref{JQ} cannot be arbitrarily small. $Q$-balls
2948: cannot therefore rotate slowly. They  show a {\it discrete} spectrum of spinning 
2949: excitations. 
2950: 
2951: 
2952: 
2953: Although it is difficult  to  qualitatively analyse  the behaviour  
2954: of the $m\neq 0$ solutions, there are some analogies with the $m=0$ case. 
2955: Specifically, $Q$-balls become large as $\omega\to\omega_{\pm}$.
2956: For $\omega\to\omega_{-}$ they can be viewed as squashed spheroids,
2957: homogeneously filled inside, which reminds of the thin wall approximation.  
2958: Unfortunately, an analog of the step function solution \eqref{thin} does not 
2959: directly apply for $m\neq 0$, 
2960: since the function $f$ cannot be constant 
2961: inside the ball because it must vanish at the symmetry axis.  Instead,
2962: $f$ increases as one moves away from the  axis and reaches maximal values at the 
2963: surface of the spheroids, after which it rapidly goes to zero. 
2964: The energy density is approximately constant inside the spheroid, with a slight
2965: increase at its surface, and rapidly vanishes outside it.   
2966:  For $\omega\to\omega_{+}$ solutions also become large spheroids,
2967: but this time they are hollow, with the maximal energy density concentrated
2968: at the surface and being close to zero everywhere else. It seems that the thick wall
2969: approximation \eqref{thick} can be directly generalized to the axially 
2970: symmetric case to give   
2971: $
2972: f(r,\theta)\approx \sqrt{2}M_\omega\, y(M_\omega r,\theta)
2973: $
2974: where $M_\omega\to 0$ and $y(r,\theta)$ fulfills 
2975: $(\Delta-\frac{m^2}{r^2\sin^2\theta} +y^2-1)y=0$.  
2976:  
2977: 
2978: If one considers $Q$ and not $\omega$ as the solution parameter 
2979: then, using $E(\omega)$ and $Q(\omega)$  to express $E$ in terms of $Q$,
2980: one discovers that the function $E(Q)$ is double-valued with a cusp,
2981: as shown in  Fig.\ref{FigQ3} 
2982: \cite{Kleihaus:2005me}.
2983: For a given $Q$ there are thus two different 
2984: spinning $Q$-ball solutions with different energies.   
2985: Solutions from the less energetic branch correspond to the 
2986: $\omega<\omega_{\rm crit}$ parts of the $E(\omega)$, $Q(\omega)$ curves
2987: in Fig.\ref{FigQ3}. As in the $m=0$ case, the plots shown in Fig.\ref{FigQ3} 
2988: demonstrate for large $Q$ the 
2989: linear dependence, $E=\omega_{\pm}Q$. This can again be explained
2990: by the general relation \eqref{Eomega}, implying  that for large $Q$
2991: one can neglect
2992: the gradient energy $E_2$ as compared to $Q$. 
2993: 
2994: \begin{figure}[ht]
2995: \hbox to\linewidth{\hss%
2996: 	\resizebox{7cm}{6cm}{\includegraphics{Qballs-varw.eps}}
2997: \hspace{5mm}%
2998:         \resizebox{7cm}{6cm}{\includegraphics{Qballs-QE2.eps}}	
2999: \hss}
3000: 	\caption{Energy and charge 
3001: $E(\omega),Q(\omega)$ (left) and $E(Q)$ (right)
3002: for the $1^{+}$ spinning $Q$-balls. For large $Q$ one has 
3003: $E(Q)=\omega_{-} Q$ and $E(Q)=\omega_{+} Q$ for the lower
3004: an upper branch, respectively. The upper branch solutions are unstable,
3005: since $E/(MQ)>1$, while the lower branch solutions are stable for large enough $Q$.}
3006: \label{FigQ3}
3007: \end{figure}
3008: 
3009: The existence of two different solutions with the same $Q$ suggests that the 
3010: more energetic of them is unstable. In fact, it follows from Eq.\eqref{Eomega}
3011: that the energy-to-charge 
3012: ratio for $Q$-balls is $E/Q=\omega+2E_2/3Q$.
3013: For the upper branch solutions  one has $\omega\to\omega_{+}=M$ for large $Q$, 
3014: implying that $E/Q=M+O(Q^{-2})$ where the subleading term can in principle 
3015: be positive or negative, depending on the details of the $Q$-ball potential. 
3016: In the present case it is positive, as can be seen in Fig.\ref{FigQ3}, so that 
3017: the upper branch $Q$-balls are unstable with respect 
3018: to decay into free particles.     
3019: The same argument for the lower branch solutions gives 
3020: $E/Q=\omega_{-}<M$ for large $Q$, so that they cannot decay into free particles. 
3021: However, there can be other decay modes, and so the stability analysis is needed.
3022: In fact, the $m=0$ lower branch $Q$-balls are known to be stable 
3023: \cite{Correia}.       
3024: The stability analysis of the even parity solutions with $m\neq 0$ will be described below 
3025: in Sec.\ref{light}. 
3026: It seems that for  $|m|=1$ the lower branch solutions are stable only for large
3027: enough $Q$, while already for $|m|=2$ all of them are unstable -- 
3028:  they seem to decay by splitting
3029: into several non-spinning $Q$-balls.  
3030: 
3031: 
3032: It is plausible that spinning generalizations could also be constructed 
3033: for the excited spherically symmetric $Q$-balls with $n>0$. 
3034: The complete family of spinning $Q$-balls should therefore contain 
3035: not only $m^\pm$ solutions, but also the excited $(n,m^{\pm})$ solutions with  
3036: $n=1,2,\ldots $ for which the amplitude $f(r,\vartheta)$ exhibits nodes. 
3037: More precisely, parametrizing the $(\rho,z)$ plane by a complex variable
3038: $w=\rho+i z$, one can expect the 
3039: amplitude $f(w)$ of the $(n,m^{+})$ and $(n,m^{-})$ solutions to have the same zeros as,
3040: respectively,  $\Re(F^{+}_n(w))$ and $\Im(F^{-}_n(w))$,
3041: where 
3042: \be                                     \label{complex} 
3043: F^\pm_n(w)=\prod_{j=1}^n \alpha^\pm_j(w-\rho^\pm_j),
3044: \ee
3045: with some suitably chosen 
3046: $\alpha^\pm_j\in\mathbb{C}$ and $\rho^\pm_j>0$. 
3047: However, very little is known about such excited solutions at present.  
3048: 
3049: 
3050: 
3051: 
3052: \subsubsection{Twisted $Q$-balls}
3053: 
3054: Here and in the next subsection we briefly describe our 
3055: new results on further generalizations of the spinning $Q$-balls,
3056: not yet discussed in the literature. Let us consider again 
3057: the theory \eqref{lQ}, but 
3058: generalize the field ansatz to include an independent phase,
3059: \be
3060: \Phi=e^{i(\omega t +m\varphi -n\psi(r,\theta))}f(r,\theta)\equiv 
3061: (X(r,\vartheta)+iY(r,\vartheta))e^{i(\omega t +m\varphi)}.
3062: \ee
3063: We require the phase function $\psi(r,\vartheta)$ to increase by $2\pi $
3064: after one revolution around the contour $C$  shown
3065: in Fig.\ref{FigS}. The overall phase 
3066: $\omega t +m\varphi-n\psi(r,\theta)$ then winds 
3067: %in the two orthogonal  directions, 
3068: around the circle $S$ and along the contour $C$,
3069: exactly as for the Faddeev-Hopf field Eq.\eqref{Hopf:axial}. 
3070: For regular fields $f(r,\vartheta)$ 
3071: vanishes  at $C$ and has  $n$ zeros inside $C$. 
3072: The solutions are thus characterized by two integers $(n,m)$ giving rise 
3073: to the `topological charge' $\Q=nm$, 
3074: although now this does not represent a genuine  topological invariant.  
3075: We shall call $Q$-balls with $\Q\neq 0$ twisted, by analogy with twisted loops
3076: in the Faddeev-Hopf theory,
3077: while those with $\Q=n=0$ described above will be called simply spinning
3078: or `non-twisted'. 
3079: 
3080: \begin{figure}[ht]
3081: \hbox to\linewidth{\hss%
3082: 	\resizebox{7cm}{6cm}{\includegraphics{X-top-Q-balls.eps}}
3083: \hspace{5mm}%
3084:         \resizebox{7cm}{6cm}{\includegraphics{Y-top-Q-balls.eps}}	
3085: \hss}
3086: 	\caption{Profiles of the $(1,1)$ twisted $Q$-ball with $\omega=0.9$.
3087: %Both $X$ and $Y$ vanish 
3088: %on a circle in the $z=0$ plane.
3089: }
3090: \label{FigQ0}
3091: \end{figure}
3092: 
3093: The field equations read  
3094: \begin{align}                                   \label{Qeq1}
3095: \left(\Delta-\frac{m^2}{r^2\sin^2\vartheta}+\omega^2\right)X&=
3096: \frac12\frac{\partial U(\sqrt{X^2+Y^2})}{\partial X}\,, \notag \\
3097: \left(\Delta-\frac{m^2}{r^2\sin^2\vartheta}+\omega^2\right)Y&=
3098: \frac12\frac{\partial U(\sqrt{X^2+Y^2})}{\partial Y}\,,
3099: \end{align}
3100: where one can require $X,Y$ to be symmetric and antisymmetric, respectively, 
3101: with respect to reflections in the equatorial plane. 
3102: %Setting $Y=0$ gives again the non-twisted $Q$-balls. 
3103: The simplest twisted solutions are obtained 
3104: for $n=m=1$ and are shown in Fig.\ref{FigQ0}.  
3105: 
3106: 
3107: \begin{figure}[ht]
3108: \hbox to\linewidth{\hss%
3109: 	\resizebox{7cm}{6cm}{\includegraphics{top-Q-balls-E3D.eps}}
3110: \hspace{5mm}%
3111:         \resizebox{7cm}{6cm}{\includegraphics{top-Q-ball-E=0,3-new.eps}}	
3112: \hss}
3113: 	\caption{Energy density for the $(1,1)$ twisted $Q$-ball 
3114: with $\omega=0.9$ (left) and the isosurfaces of constant
3115: energy density with 
3116: $T^0_0=0.3$ (right).}
3117: \label{FigQQ}
3118: \end{figure}
3119: The profiles of the energy density look spectacular; see Fig.\ref{FigQQ}.
3120: It is unclear  at present 
3121: whether these solutions are dynamically
3122: stable, but they certainly exist
3123: as solutions of the elliptic system  \eqref{Qeq1}.
3124: The twisted $Q$-balls are much more heavy 
3125: than the non-twisted ones. For example, for $\omega=0.9$   
3126: we found for the twisted $(1,1)$ solution the energy
3127: $E=929.34$,
3128: while the non-twisted  $1^{+}$ and $1^{-}$ spinning
3129: $Q$-balls have, respectively, 
3130: $E=169.27$ and $E=338.66$.
3131: 
3132: %\input{Qgauged}
3133: 
3134: 
3135: 
3136: 
3137: \subsubsection{Spinning gauged $Q$-balls}
3138: 
3139: Another possibility to generalize the spinning $Q$-balls 
3140: is to couple them to a gauge field 
3141: %This exercise is important for our discussion, since 
3142: %spinning $Q$-balls can be considered
3143: %as simple prototypes of global vortons, so that gauging 
3144: %them gives an idea of how a similar procedure would work for vortons. 
3145: %The resulting solutions are 
3146: %{manifestly} stationary and {manifestly} axially symmetric,
3147: %and yet their angular momentum is not zero. 
3148: within the theory 
3149: \be                                                     \label{lQg}
3150: {\mcal L}_{\rm gQ}[A_\mu,\Phi ]=
3151: -\frac14\,F_{\mu\nu}F^{\mu\nu}+
3152: {\mcal L}_{\rm Q}[\Phi ].
3153: \ee
3154: Here $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$ while 
3155: ${\mcal L}_{\rm Q}[\Phi ]$ is the same $Q$-ball Lagrangian as in \eqref{lQ} but with 
3156: the derivatives of $\Phi$  
3157: replaced by the covariant derivatives, 
3158: $\partial_\mu\Phi\to D_\mu\Phi=(\partial_\mu-i g A_\mu)\Phi$ 
3159: where $g$ is the gauge coupling constant. 
3160: This theory is in fact the U(1) version  of the general  gauge field model \eqref{YMH}.
3161: The local gauge transformations \eqref{UUU} now read
3162: %\be                          %\label{UUU1}
3163: $\Phi \to \Phi e^{i g \alpha }$,
3164: $A_\mu\to A_\mu +\partial_\mu \alpha$ 
3165: while the field equations \eqref{YMHeqs} assume the form
3166: \begin{align}
3167: \label{gauged-Q-ball-eqs}  
3168:  \partial^\mu   F_{\mu\nu}&= 
3169: ig \left \{({D}_\mu\Phi)^\ast\Phi- \Phi^\ast ({D}_\mu\Phi) \right\}
3170: \equiv g j_\nu,  \notag \\
3171: {D}_\mu {D}^\mu \Phi &=-\frac{\partial U}{\partial |\Phi|^2}\,\Phi.
3172:  \end{align} 
3173: The conserved Noether charge  analogues to the $Q$-ball charge 
3174: Eq.\eqref{Noether} is 
3175: \be
3176: Q=\int j_0\, d^3\bx=\frac{1}{g}\oint 
3177: \vec{\mcal E}d\vec{S}\equiv \frac{4\pi Q_{\rm el}}{g},
3178: \ee 
3179: where $Q_{\rm el}$ is the electric charge. 
3180: Spherically symmetric solutions of this model were discussed in 
3181:  Ref.\cite{Lee:1988ag}. 
3182: 
3183: 
3184: 
3185: We make the ansatz
3186: \be                          \label{gQfields}        
3187: A_\mu dx^\mu=A_0(r,\vartheta)dt+A_\varphi(r,\vartheta)\sin \theta d\varphi, ~~~~  
3188: \Phi=f(r,\vartheta)e^{i (m\varphi+\omega t) },
3189: \ee
3190: with real $f(r,\vartheta)$, and require  
3191: the gauge field to vanish at infinity, while at the $z$-axis 
3192: $A_\varphi=\partial_\theta A_0=0$. 
3193: Although $\Phi$ depends on $t,\varphi$, this dependence can be gauged away,
3194: so that the system is 
3195: {manifestly} stationary and {manifestly} axially symmetric. 
3196: Numerically solving the field equations gives spinning $Q$-balls with 
3197: a long range gauge field which behaves for large $r$ as   
3198: (${\bf m}$ being  the magnetic dipole moment)
3199: \be                           \label{Ainf}
3200: A_0=\frac{Q_{\rm el}}{r}+\dots,~~~~~~~
3201: A_\varphi=\frac{{\bf m}\sin \theta }{r^2}+\dots \,.
3202: \ee 
3203: In the limit $g\to 0$ they reduce to the non-gauged spinning $Q$-balls. 
3204: It seems that solutions exist if only $g$ does not exceed 
3205: a certain maximal value $g_{\rm max}(\omega)$ (see Fig.\ref{FigQg}). 
3206: This feature can be understood qualitatively \cite{Lee:1988ag}: since 
3207: $Q$-balls can be viewed as condensate states of mutually attracting scalar particles,
3208: gauging them creates  an electric repulsion that destroys  
3209:  the condensate for large enough $g$.  
3210: 
3211: 
3212:   
3213: 
3214: \begin{figure}[ht]
3215: \hbox to\linewidth{\hss%
3216: 	\resizebox{7cm}{6cm}{\includegraphics{gauged-Q-Balls-E-Q2.eps}}
3217: \hspace{5mm}%
3218:         \resizebox{7cm}{6cm}{\includegraphics{var-g1.eps}}	
3219: \hss}
3220: 	\caption{Left: the energy  $E(\omega)$ and $E(Q)/(QM)$ 
3221: for several values of $g$ for the $1^{+}$ gauged $Q$-balls.
3222: Right: the energy $E(g)$ for several values of $\omega$
3223: for the $1^{+}$ solutions. The plots of $Q(\omega)$ and $Q(g)$
3224: look qualitatively similar.
3225: }
3226: \label{FigQg}
3227: \end{figure}
3228: 
3229: Since the symmetries of the solutions are manifest, one can use the 
3230: surface integral formula \eqref{JJJ} for the angular momentum. 
3231: Applying the symmetry equations 
3232: \eqref{sym1} to the ansatz \eqref{gQfields} gives $W_\varphi=m/g$, 
3233: inserting which to \eqref{JJJ} and using \eqref{Ainf} yields 
3234: \be                                             \label{JJJ1}
3235: J=\oint 
3236: \left(\frac{m}{g}-A_\varphi\right)\vec{{\mcal E}}\, \vec{dS}
3237: =\frac{4\pi Q_{\rm el} m}{g}=Qm.
3238: \ee
3239: This shows that spinning is possible in manifestly stationary and {manifestly} axially symmetric systems where all the spinning phases can be gauged away.
3240: 
3241: 
3242: 
3243: 
3244: 
3245: 
3246: As seen in Fig.\ref{FigQg},  the dependence of solutions on $\omega$ 
3247: is similar to that in the ungauged case: they exist  
3248: for a limited range of $\omega$. 
3249: Expressing $E$ in terms of $Q$ gives again a two-branch 
3250: function $E(Q)$, the solutions from the lower branch being stable
3251: for large enough $Q$ with respect to decay into free particles. 
3252: $E,Q$ cannot be arbitrarily large for $g\neq 0$, since 
3253: the electric and scalar field contributions to the energy grow as $Q_{\rm el}^2\sim Q^2$
3254: and $Q$, respectively, and so for large $Q$ the electric term dominates
3255: and destroys the soliton. 
3256: 
3257: 
3258: For small $g$  solutions can be represented as 
3259: $\Phi=\Phi^{(0)}+g^2\Phi^{(2)}+\ldots$ and 
3260: $A_\mu=gA^{(1)}_\mu+\ldots$ where 
3261: $\Phi^{(0)}$ is the non-gauged $Q$-ball. 
3262: The energy   is  
3263: $
3264: E(g)=E^{(0)}+g^2E^{(2)}+\ldots 
3265: $
3266: and calculating $E^{(2)}$ reveals  that it is not sign definite -- 
3267: due to the electric field contribution. 
3268: The plots in Fig.\ref{FigQg} show that it can be both 
3269: positive and negative, depending on the solution. 
3270: 
3271: 
3272: 
3273: 
3274: 
3275: 
3276: 
3277: 
3278: 
3279: 
3280: \subsubsection{Spinning interacting $Q$-balls}
3281: \begin{figure}[ht]
3282: \hbox to\linewidth{\hss%
3283: 	\resizebox{6cm}{4cm}{\includegraphics{Saturn-Q-balls-E=0,1-new.eps}}
3284: \hspace{5mm}%
3285:         \resizebox{6cm}{4cm}{\includegraphics{2-rings-E=0,3-new.eps}}	
3286: \hss}
3287: 	\caption{Energy isosurfaces with $T_0^0=0.1$ for 
the $m_1=0,~m_2=3$ {`Saturn'}  (left) and    
3288: with  $T_0^0=0.3$ for
3289: the $m_1=1,~m_2=3$ {bi-ring}  (right) solutions in the theory \eqref{2Q}. }
3290: \label{Fig:saturn}
3291: \end{figure}
3292: 
3293: Yet another way to generalize $Q$-balls is to couple to each other
3294: several copies of the theory. This allows one to consider `non-linear superpositions'
3295: of individual $Q$-balls. 
3296: Although this idea has already been considered
3297: in the literature \cite{Brihaye:2007tn}, we have reconsidered it 
3298: and found some curious solutions which could be interesting in the context of
3299: our discussion. In the simplest case one can choose 
3300: \be
3301: \label{2Q}
3302:  L[\Phi_1,\Phi_2]=L_Q(\Phi_1)+L_Q(\Phi_2) -\gamma |\Phi_1|^2|\Phi_2|^2,
3303: \ee
3304: where $L_Q(\Phi_1)$  and $L_Q(\Phi_2)$ are two copies of the $Q$-ball
3305: Lagrangian (\ref{lQ}). Setting 
3306: \be
3307: \Phi_1=e^{i(\omega_1 t +m_1\varphi)}f_1(r,\theta),~~~~~~
3308: \Phi_2=e^{i(\omega_2 t +m_2\varphi)}f_2(r,\theta),
3309: \ee
3310: it is interesting to consider solutions 
3311: with $m_1\neq m_2$. 
3312: \begin{figure}[ht]
3313: \hbox to\linewidth{\hss%
3314: 	\resizebox{7cm}{4.5cm}{\includegraphics{Qball-vortex.eps}}
3315: \hspace{5mm}%
3316:         \resizebox{7cm}{4.5cm}{\includegraphics{m3E0,1-new.eps}}	
3317: \hss}
3318: 	\caption{The energy density $T_0^0$ and the $T_0^0=0.1$ energy  
3319: isosurface for the $m_1=0$, $m_2=3$ `hoop' solution   
3320: in the theory \eqref{2Q}. }
3321: \label{hoop}
3322: \end{figure} 
3323: Choosing $\omega_1=\omega_2=1$, $\gamma=1$ and 
3324: restricting to the even parity sector, we find solutions
3325: of the `Saturn type', with a central concentration of 
3326: the energy density produced by the first scalar with $m_1=0$ and 
3327: surrounded  by a ring created 
3328: by the second scalar with $m_2>0$ (Fig.\ref{Fig:saturn}), 
3329: solutions for $m_2>m_1>0$ with bi-ring profiles (Fig.\ref{Fig:saturn}),
3330: as well as solutions describing  a superposition of 
3331: a straight   `$Q$-ball vortex' with $m_1=0$
3332: `hooped'  by a $Q$-ball ring with $m_2>0$  (Fig.\ref{hoop}).
3333: 
3334:  
3335: 
3336: 
3337: 
3338: 
3339: 
3340: \subsection{Skyrmions}
3341: 
3342: The first known example of topological solitons in $3+1$ 
3343: dimensions was suggested  almost 50 years ago
3344: within the non-linear relativistic field theory model 
3345: proposed by T.H.R.~Skyrme 
3346: \cite{Skyrme:1961vq}. These
3347: solitons are now called  skyrmions. 
3348: Skyrme himself considered them as field theoretic realizations of baryons. 
3349: Nowadays the Skyrme model is regarded as an effective, low energy approximation 
3350: of  QCD 
3351: \cite{Adkins:1983ya},\cite{Adkins:1983hy}.
3352: In this approximation the static skyrmions are 
3353: promoted to spinning objects by making use of the effective rigid body 
3354: approximation that will be 
3355: considered below.  The masses of the spinning skyrmions obtained in this way
3356: are then compared to the hadron masses.  
3357: 
3358: The question of whether spinning skyrmions 
3359: really exist as stationary field theory
3360: objects was addressed only very recently in Ref.\cite{Battye:2005nx},
3361: whose authors constructed spinning skyrmions by applying 
3362: the same mechanism as for the $Q$-balls and arrived 
3363: at conclusions which differ considerably from those obtained 
3364: within the rigid 
3365: body approximation. 
3366: 
3367: Since there exist excellent descriptions of the Skyrme model and its solutions 
3368: in the literature \cite{Makhankov},\cite{Manton:2004tk}, 
3369: we shall 
3370: very briefly summarize only the features essential for our discussion. 
3371: The fundamental field variables in the theory  
3372: comprise an SU(2)-valued matrix $U(x^\mu)$
3373: which satisfies field equations for  the Lagrangian
3374: \be                                                      \label{Skyrme}
3375: {\mcal L}_{\rm S}[U]={\rm tr}
3376: \left(\frac{1}{2}\,\partial_\mu U^\dagger \partial^\mu U+
3377: \frac{1}{8}\,[\partial_\mu U^\dagger,\partial_\nu U]
3378: [\partial^\mu U^\dagger,\partial^\nu U]\right).
3379: \ee
3380: It is convenient to introduce a pure gauge SU(2) connection, 
3381: \be                                       \label{puregauge}
3382: {\bf A}_\mu=iU^\dagger\partial_\mu U\equiv \tau^a{\bf A}^a_\mu,~~~~~ 
3383: \partial_\mu{\bf A}_\nu-\partial_\nu{\bf A}_\mu=
3384: i[{\bf A}_\mu,{\bf A}_\nu],
3385: \ee 
3386: in terms of which 
3387: the field equations read
3388: \be                                       \label{eqs-skyrme}
3389: \partial^\mu\left({\bf A}_\mu-\frac{1}{4}\,[[{\bf A}_\mu,{\bf A}_\nu] ,{\bf A}^\nu]\right)=0.
3390: \ee
3391: The energy for static fields is
3392: %\be                                                      \label{Skyrme1}
3393: %E[U]=\frac{1}{12\sqrt{2}\pi^2}\,\int {\rm tr}\left(\frac12({\bf A}_k)^2
3394: %+\frac{1}{8}({\bf F}_{ik})^2\right) d^3\bx
3395: %\ee
3396: \be                                                      \label{Skyrme1}
3397: E[U]=\int {\rm tr}\left(\frac12({\bf A}_k)^2
3398: -\frac{1}{8}([{\bf A}_i,{\bf A}_k])^2\right) d^3\bx\,.
3399: \ee
3400: For the energy to be finite
3401: the field  $U(\bx)$ should approach a constant value at infinity, which can be
3402: chosen to be the unit matrix, so that $\lim_{|\bx|\to\infty}U(\bx)=1$.   
3403: This allows one to replace $\mathbb{R}^3$ by its one-point compactification 
3404: $S^3$. Since SU(2) is topologically also $S^3$ it follows that any 
3405: finite energy field configuration can be viewed as a map,
3406: $
3407: U(\bx): S^3\to S^3,
3408: $
3409: and can therefore be characterized by the integer degree of map,
3410: \be                                  \label{baryon}
3411: \Q[U]=\frac{i}{24\pi^2}\int 
3412: %U dU^{\dagger} \wedge U dU^{\dagger}\wedge  U dU^{\dagger}\,.
3413: {\rm tr}\left(\epsilon_{ijk}{\bf A}_i {\bf A}_j {\bf A}_k\right)d^3\bx \,.
3414: \ee
3415: This topological charge is called in Skyrme theory `baryon number'. 
3416: Fixing it the energy obeys the Bogomol'nyi type inequality
3417: \be
3418: E[U]\geq 12\sqrt{2}\pi^2|\Q[U]|,
3419: \ee
3420: which can be easily obtained by rearranging terms in the integrand in \eqref{Skyrme1}.  
3421: The existence of this lower bound suggests looking for  energy
3422: minimizers in each topological sector. 
3423: 
3424: It will be convenient for what follows to parametrize the matrix $U$ in terms 
3425: of two complex scalar fields $\phi,\sigma$ 
3426: subject to the normalization condition
3427: $
3428: |\phi|^2+|\sigma|^2=1$
3429: as 
3430: \be 						\label{U}
3431: U=
3432: \left(
3433: \begin{array}{cc}
3434: \phi & i\sigma^\ast \\
3435: i\sigma & \phi^\ast 
3436: \end{array}
3437: \right). 
3438: \ee
3439: Simplest skyrmions are spherically symmetric, 
3440: $U=\exp(i\chi(r)\,{\bx\cdot\mbox{\boldmath$\tau$}}/{r})$, hence 
3441: \be                               \label{SU}
3442: \phi=\cos\chi(r)+i\sin\chi(r)\cos\vartheta ,~~~~~
3443: \sigma=\sin\chi(r)\sin\vartheta e^{i\varphi}.
3444: \ee
3445: Inserting this to Eqs.\eqref{eqs-skyrme} the variables separate
3446: and the equations reduce to an  
3447:  ODE for $\chi(r)$. This equation admits globally regular solutions
3448: with the boundary conditions $\chi(0)=\Q\pi$ and $\chi(\infty)=0$
3449: for any value of the topological charge $\Q$. 
3450: However, it seems that only for $\Q=\pm 1$ these solutions correspond to the
3451: absolute  energy minimum, while those for $|\Q|>1$ are local minima or saddle
3452: points. Global energy minima for $|\Q|>1$ are not spherically symmetric and can be 
3453: obtained by directly minimizing the energy \eqref{Skyrme1}. 
3454: For $|\Q|=2$ they are
3455: axially symmetric, with the constant baryon number isosurfaces having 
3456: toroidal shape, so they are somewhat similar to knots. 
3457: However, higher charge skyrmions do not resemble knots at all and 
3458: look like polyhedral shells. Their detailed description 
3459: can be found in the monograph  \cite{Manton:2004tk}. 
3460: 
3461: \subsubsection{Skyrme versus Faddeev-Skyrme \label{versus}}
3462: 
3463: It is worth emphasizing the similarity between the Skyrme and Faddeev-Skyrme 
3464: models. Both can be represented in the form (up to normalization) 
3465: \be						\label{vers}
3466: {\mcal L}[\mbox{\boldmath$\phi$} ]=
3467: \partial_\mu \mbox{\boldmath$\phi$}\cdot\partial^\mu\mbox{\boldmath$\phi$}
3468: -\frac14\,(\partial_\mu \mbox{\boldmath$\phi$}\cdot\partial^\mu\mbox{\boldmath$\phi$})^2
3469: +\frac14\,(\partial_\mu \mbox{\boldmath$\phi$}\cdot\partial_\nu\mbox{\boldmath$\phi$})
3470: (\partial^\mu \mbox{\boldmath$\phi$}\cdot\partial^\nu\mbox{\boldmath$\phi$})
3471: \ee
3472: where \mbox{\boldmath$\phi$} is a vector normalized by the condition
3473: $
3474: \mbox{\boldmath$\phi$}\cdot \mbox{\boldmath$\phi$}=1.
3475: $ 
3476: In the Faddeev-Skyrme case one has $\mbox{\boldmath$\phi$}=\phi^a$ with $a=1,2,3$ so that 
3477: the target space is $S^2$.
3478: In the Skyrme model one has  $\mbox{\boldmath$\phi$}=(\phi^0,\phi^a)$ with $a=1,2,3$
3479: so that the target space is $S^3$, the Skyrme field being 
3480: $U=\phi^0+i\phi^a\tau^a$. 
3481: One can also  
3482: construct a generalized theory whose target space `interpolates' between $S^2$ and $S^3$
3483: and so it includes both models as special cases \cite{Ward:2004gr}.  
3484: 
3485: 
3486: 
3487: The Faddeev-Skyrme model can be regarded as a consistent
3488: truncation of the Skyrme theory.  
3489: Specifically, if $\nn=\n^a$ is a solution of the 
3490: Faddeev-Skyrme model, then $\mbox{\boldmath$\phi$}=(0,\n^a)$ solves the equations 
3491: of the Skyrme model. 
3492:  As a result, 
3493: knot solitons can be embedded into the Skyrme model \cite{Meissner}, \cite{Cho},
3494: although their stability properties will then be different. 
3495: 
3496: Another embedding of knot solitons into Skyrme theory is obtained by 
3497: expressing the fields  ${\mcal A}_\mu$ and $\nn$ 
3498: in terms of the $CP^1$ coordinates 
3499: $\phi,\sigma$ using Eqs.\eqref{norm},\eqref{A},\eqref{project}. 
3500: The Skyrme field $U$ is then given by
3501: Eq.\eqref{U}. Although this correspondence does not map solutions to 
3502: solutions, the advantage now is that the 
3503: baryon number $\Q[U]$ is exactly equal to the 
3504: Hopf charge $\Q[\nn]$. This follows from the fact that 
3505: ${\bf A}^3_\mu=-{\mcal A}_\mu$ and  
3506: ${\mcal F}_{\mu\nu}=2({\bf A}^1_\mu{\bf A}^2_\nu-{\bf A}^1_\nu{\bf A}^2_\mu)$ 
3507: due to Eq.\eqref{puregauge}. 
3508: With this, the expression  \eqref{baryon} for the baryon number exactly reduces to the
3509: expression \eqref{Q} for the Hopf charge. 
3510: For example, using the parametrization \eqref{axial-CP} for the $CP^1$ 
3511: scalars $\phi,\sigma$ with the 
3512: Hopf charge $\Q[\nn]=nm$ gives with \eqref{U}
3513: the Skyrme field $U$ with $\Q[U]=nm$.  
3514: 
3515: The Hopf charge $\Q[\nn]$ will determine in this case also the winding 
3516: number of the pure gauge field
3517: ${\bf A}_k=iU^\dagger\partial_k U$. If $\nn$ is an energy minimizer
3518: then the pure gauge ${\bf A}_k$  
3519: is {maximally abelian} 
3520: \cite{vanBaal:2001jm}.
3521: 
3522: The correspondence can also be used in the opposite direction to construct 
3523: $\nn$ with a given Hopf charge out of $U$ with a given baryon number. 
3524: If $U$ is spherically symmetric, then $\nn$ is axially symmetric,
3525: as can be seen comparing Eq.\eqref{SU} and Eq.\eqref{W-Hopf1}.  
3526: Using approximations for $U$ gives in this way approximate solutions
3527: for $\nn$ \cite{Ward2}. 
3528: 
3529: 
3530: 
3531: 
3532: 
3533: 
3534: \subsubsection{Spinning skyrmions}
3535: 
3536: Let us consider the  $\Q=1$ static, spherically symmetric skyrmion  
3537: described by  Eq.\eqref{SU}. Its spinning analog  
3538: in the
3539:  rigid body approximation is obtained by simply replacing 
3540: in Eq.\eqref{SU} 
3541: \be
3542: \varphi\to\varphi+\omega t\,,
3543: \ee 
3544: which 
3545: gives a non-zero value to the angular momentum. It is in fact 
3546: precisely this type of approximation 
3547: that  is often implicitly assumed in the literature when talking
3548: about spinning solitons. 
3549: In this approximation  spinning solitons do not change their shape 
3550: and the radiation effects are neglected, which are natural assumptions  
3551: if the rotation is slow. However, if $\omega$ is not small, then 
3552: this description does not approximate the true spinning 
3553: solutions any more but simply 
3554: gives field configurations with  $J\neq 0$.  
3555: If one uses them as initial data, then their temporal evolutions will be certainly 
3556: accompanied by radiation carrying $J$ away, and it is not clear
3557: if the system will finally relax to a state with $J\neq 0$.   
3558: 
3559: 
3560: 
3561: In order to construct truly spinning skyrmions with $\Q=1$, 
3562: Battye, Krusch and Sutcliffe (BKS) \cite{Battye:2005nx}
3563: generalize the spherically symmetric ansatz 
3564: \eqref{U},\eqref{SU} to the axially symmetric, 
3565: non-manifestly stationary one, which can be parametrized as  
3566: %\be                                                      \label{SU1}
3567: %\phi=X(r,\vartheta)+iY(r,\vartheta) ,~~~~
3568: %\sigma=Z(r,\vartheta)e^{i(\varphi+\omega t)},
3569: %\ee
3570: \be                                        \label{SU1}
3571: \phi=\cos\frac{\Theta}{2}\,e^{i\psi},~~~~~
3572: \sigma=\sin\frac{\Theta}{2}\,e^{i(\varphi+\omega t)}.
3573: \ee 
3574: It is instructive to compare this with the ansatz \eqref{axial-CP} for the 
3575: axially symmetric hopfions with $m=n=1$.  
3576: They also add to the Lagrangian \eqref{Skyrme} the mass term 
3577: \be                         \label{Skyrmmass}
3578: M^2\,{\rm tr}(U-1). 
3579: \ee   
3580: Instead of solving the field
3581: equations for $\Theta(r,\vartheta),\psi(r,\vartheta)$ they 
3582: minimize the energy with fixed $J$, whose expression is  
3583: similar to the one in Eq.\eqref{EJ}. 
3584: The frequency $\omega$ for 
3585: their solutions is a parameter in the interval 
3586: \be                          \label{limit}               
3587: 0\leq \omega^2 <M^2,
3588: \ee
3589: both the energy and angular momentum increasing with $\omega$. 
3590: The latter feature is very interesting -- for spinning skyrmions $J$ is a 
3591: continuous parameter that can be arbitrarily small, so that they can 
3592: rotate slowly. 
3593: To compare, $Q$-balls cannot rotate slowly. 
3594: Although BKS do not emphasize  this, it seems that both $E$ and $J$ should blow up 
3595: as $\omega^2\to M^2$ since the fields localized by the factor 
3596: $\exp(-\sqrt{M^2-\omega^2}\, r)$
3597: become then long range. 
3598:  
3599: 
3600: 
3601: 
3602: 
3603: Trying to adjust $M^2$ and the energy scale 
3604: to reproduce the pion and hadron masses and their spins, 
3605: BKS discover that this requires moving to the parameter region where deviations
3606: of spinning skyrmions from spherical symmetry are large. 
3607: They conclude that 
3608: the rigid body approximation does  not provide an adequate 
3609: description of spinning skyrmions. 
3610: 
3611: The results of BKS have been confirmed  
3612: by solving the field equations by Ioannidou,
3613: Kleihaus and Kunz \cite{Ioannidou:2006nn} (who studied in fact gravity-coupled
3614: skyrmions, but considered  the Minkowski space limit as well).  
3615: %This example of spinning skyrmions is very instructive 
3616: %for our discussion, since it shows once again that the presence of a non-zero 
3617: %angular momentum is not always compatible with 
3618: %the absence of radiation. 
3619: %Spinning skyrmions
3620: %radiate if they spin too fast. 
3621: 
3622: 
3623: The spinning skyrmions have been obtained by BKS by minimizing
3624: the energy within the axially symmetric ansatz \eqref{SU1}.
3625: Strictly speaking, this does not 
3626: exclude the possibility of non-axially symmetric instabilities. However,
3627: since the non-spinning skyrmions are stable, one can expect that at least for 
3628: small $J$ their spinning analogs should be stable as well. 
3629: 
3630: Spinning solitons have also been studied  
3631: in the context of the baby Skyrme model \cite{Piette},\cite{Betz}. This is simply
3632: the Faddeev-Skyrme model restricted to $2+1$ spacetime dimensions and 
3633: written in the form \eqref{vers} with an additional  mass term 
3634: $M^2(\mbox{\boldmath$\phi$}^3-1)$. Spinning fields are then chosen to be 
3635: \be                                           \label{baby}
3636: \mbox{\boldmath$\phi$}^1+i\mbox{\boldmath$\phi$}^2=\sin\Theta(\rho)e^{i(\omega t+\varphi)},~~~~~
3637: \mbox{\boldmath$\phi$}^3=\cos\Theta(\rho).
3638: \ee
3639: The field equations reduce in this case to an ODE for $\Theta(\rho)$ whose solutions 
3640: are easy to study. Solutions exist for $\omega< M$ and  both $E$ and $J$
3641: blow up as $\omega\to M$.   
3642: 
3643: Trying to further increase $\omega$, solutions become oscillatory, since  one has 
3644: $\Theta\sim \exp(-\sqrt{M^2-\omega^2}\, \rho)$ for large $\rho$, 
3645: both $E$ and $J$ being then infinite.  However, one can consider the initial data 
3646: of the form \eqref{baby}, where $\omega>M$ and $\Theta(\rho)$ 
3647: vanishes identically for large $\rho$.
3648: $E,J$ will then be finite.   Evolving these data in time, 
3649: the system radiates away a fraction of its energy and angular momentum and relaxes 
3650: to a stationary, rotating configuration with $\omega<M$ \cite{Piette}.  
3651: 
3652: \subsubsection{Spinning gauged skyrmions}
3653: 
3654: 
3655: Spinning solutions have also been constructed within the gauged version of 
3656: the Skyrme model \cite{Callan:1983nx} 
3657: by Radu and Tchrakian \cite{Radu:2005jp}. 
3658: These solutions are 
3659:  manifestly stationary and manifestly axisymmetric. 
3660: Parametrizing the Skyrme field $U$ in terms of two complex scalars $\phi,\sigma$ 
3661: according to Eq.\eqref{U}, 
3662: the gauged Skyrme model is obtained from the Skyrme model as
3663: \be                                                               \label{Skyrme2}
3664: {\mcal L}_{\rm gS}[A_\mu,\phi,\sigma]=
3665: -\frac14\,F_{\mu\nu}F^{\mu\nu}+
3666: {\mcal L}_{\rm S}[\phi,\sigma].
3667: \ee
3668: Here ${\mcal L}_{\rm S}$ is the Skyrme Lagrangian \eqref{Skyrme} with the mass term 
3669: \eqref{Skyrmmass} included, 
3670: $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$, and  the derivatives 
3671: of the field $\sigma$ in ${\mcal L}_{\rm S}[\phi,\sigma]$ 
3672: are replaced by the covariant derivatives,
3673: $\partial_\mu\sigma\to(\partial_\mu-igA_\mu)\sigma$,  
3674: while the derivatives  $\phi$ do not change, so that the theory is invariant 
3675: under local U(1) gauge transformations
3676: \be
3677: A_\mu\to A_\mu+\partial_\mu\alpha,~~~~\sigma\to e^{i\alpha}\sigma,~~~~
3678: \phi\to\phi.
3679: \ee
3680: Radu and Tchrakian \cite{Radu:2005jp} generalize the ansatz \eqref{SU1} as 
3681: \be                          \label{SU2}        
3682: A_\mu dx^\mu=A_0dt+A_\varphi\sin\vartheta\, d\varphi,~~~
3683: \phi=\cos\frac{\Theta}{2}\,e^{i\psi},~~~~~
3684: \sigma=\sin\frac{\Theta}{2}\,e^{i(m\varphi+\omega t)}, 
3685: \ee
3686: where $A_0,A_\varphi,\Theta,\psi$ depend on $r,\vartheta$. 
3687: This ansatz 
3688: is manifestly stationary and manifestly axially symmetric,
3689: since its $t,\varphi$-dependence can be removed by a gauge 
3690: transformation.  The topological charge \eqref{baryon} is $\Q=m$. 
3691: Integrating the field equations gives
3692: globally regular solutions with finite energy and a longrange gauge field, 
3693: which is at large $r$
3694: \be
3695: A_0=\frac{Q_{\rm el}}{r}+\ldots,~~~~
3696: A_\varphi=\frac{\bf{m}}{r^2}\,\sin\vartheta+\ldots ,
3697: \ee
3698: where $Q_{\rm el},{\bf m}$ are the electric charge and dipole moment. 
3699: This feature is similar to that for the gauged $Q$-balls in Eq.\eqref{Ainf}. 
3700: The angular momentum can be calculated in the same way as in Eq.\eqref{JJJ1},
3701: with a similar result, 
3702: \be                \label{Jmm}
3703: J=\frac{4\pi Q_{\rm el}m}{g}\,.
3704: \ee
3705: However, unlike for $Q$-balls, $J$ can be arbitrarily small if $\omega$ is small.
3706: The dependence of solutions on $\omega$ is 
3707:  qualitatively similar to that     
3708: for the global skyrmions: they exist only for a finite frequency range,
3709: both $E$ and $J$ growing with $\omega$.
3710: %As in the  gauged $Q$-balls case, 
3711: %it is difficult to approach  the upper limit of $\omega$ numerically,
3712: %since the accuracy decreases drastically then. 
3713: For $\omega=0$ the solutions reduce to the non-spinning gauged skyrmions 
3714: found in Ref.\cite{Piette:1997ny}. 
3715:  
3716: 
3717: The stability of spinning gauged skyrmions has not been studied. They
3718: probably have less chance to be stable than their
3719: global counterparts -- since gauging introduces additional degrees of freedom
3720: that can produce instabilities. 
3721: 
3722: 
3723: 
3724: \subsection{Rotating monopole-antimonopole pairs  \label{mon-pairs}}
3725: 
3726: There are other known  solutions with non-zero angular momentum. They  
3727: are manifestly stationary and manifestly axisymmetric, but they  
3728: describe rotations in  multisoliton systems and not 
3729: spinning of a single soliton. 
3730: Let us consider the already mentioned above monopole-antimonopole solutions in the 
3731:  Yang-Mills-Higgs theory \eqref{YMH}. Their existence was 
3732: demonstrated by Taubes \cite{Taubes:1982ie} and they were explicitly constructed by 
3733: Kleihaus and Kunz \cite{Kleihaus:1999sx} within the ansatz \eqref{anz0}
3734: with $k=2$, $m=1$. 
3735: These solutions  are static, purely magnetic and manifestly axisymmetric 
3736: (since their $\varphi$-dependence can be gauged away) and  
3737: with vanishing magnetic charge. The Higgs field $\Phi$ shows two simple zeros at
3738: two spatial points separated by a finite distance (see Fig.\ref{FigPqa}). 
3739: These two points correspond to 
3740: the positions of the monopole and antimonopole. The
3741: analysis of the charge and current distributions 
3742: similar to that described in Sec.\ref{rings} shows  that the system also 
3743: contains a circular electric current, as shown in Fig.\ref{FigPqa}. The magnetic 
3744: field created by this current acts against the Coulombian attraction between  
3745: the monopole and antimonopole, which presumably stabilizes the system. 
3746: \begin{figure}[h]
3747: \hbox to\linewidth{\hss%
3748:   \resizebox{7cm}{5cm}{\includegraphics{MAP-phi.eps}}%
3749: \hspace{5mm}  \resizebox{6cm}{5cm}{\includegraphics{P.eps}}%
3750: \hss}
3751: \caption{\small Left: the Higgs field amplitude for the monopole-antimonopole solution
3752: in the limit where the Higgs field potential vanishes.  
3753: It is shown only for a limited region around the origin, while at infinity 
3754: it approaches the unit value. Right: the charge and current distributions 
3755: for this solution exhibits essentially the same structure as for the monopole rings,
3756: shown in Fig.\ref{Fig2}. 
3757: }
3758: \label{FigPqa}
3759: \end{figure}
3760: 
3761: Since there is no electric field for these solutions, their angular momentum
3762: \eqref{JJJ} is  zero. However, in the  
3763: limit where the Higgs potential
3764: vanishes, the Lagrangian \eqref{YMH} admits the global symmetry 
3765: \cite{Bogomolnyi},\cite{Coleman:1976uk}
3766: \be                               \label{hyper} 
3767: A_0\to A_0\cosh\gamma+\Phi\sinh\gamma,~~~~~~~~
3768: \Phi\to \Phi\cosh\gamma+ A_0\sinh\gamma,~~~~~~~A_k\to A_k,
3769: \ee 
3770: and this allows one to produce solutions with an electric field starting from 
3771: purely magnetic solutions. As was noticed in \cite{Heusler:1998ec}, this 
3772: transformation may also generate a non-zero angular momentum
3773: \be
3774: J\sim\sinh\gamma,
3775: \ee
3776: provided 
3777: that the original  purely magnetic solution is not spherically symmetric and 
3778: does not satisfy the first 
3779: order Bogomol'nyi equations \cite{Bogomolnyi}. 
3780: The latter condition excludes from consideration
3781: the (multi)monopole solutions of these equations.
3782: However, applying the symmetry \eqref{hyper}
3783: to the monopole-antimonopole solution, 
3784: which does not fulfill the Bogomol'nyi equations and 
3785: whose Higgs field is long-range in the limit of vanishing Higgs potential, 
3786: gives an electrically charged system, since both monopole and antimonopole
3787: then receive an electric charge of the 
3788: same sign \cite{Hartmann:2000ja}. The total 
3789: electric charge calculated with the gauge-invariant definition \eqref{charge-current} is 
3790: $Q_{\rm el}\sim\sinh\gamma$, and applying the surface integral formula \eqref{JJJ} 
3791: gives the angular momentum \cite{vanderBij:2002sq}
3792: \be                                           \label{JQQ} 
3793: J=\frac{4\pi Q_{\rm el}}{g}
3794: \ee
3795: directed along the symmetry axis passing through the monopole 
3796: and antimonopole. %Here $m$ is the same as in Eq.\eqref{mn}. 
3797: 
3798: \begin{figure}[h]
3799: \hbox to\linewidth{\hss%
3800: \resizebox{7cm}{4cm}{\includegraphics{Pq.eps}}%
3801: \hss}
3802: \caption{\small 
3803: Schematic visualisation of the rotating monopole-antimonopole pair as superposition
3804: of a monopole-electric charge pair  $(+P,q)$ and an antimonopole-electric charge 
3805: pair $(-P,q)$, both of which making the same contribution to the angular momentum. 
3806: }
3807: \label{FigPq}
3808: \end{figure}
3809: 
3810: It is interesting to understand the origin of this angular momentum. 
3811: If it was an orbital angular momentum, then it would be orthogonal 
3812: to the monopole-antimonopole symmetry axis, but it is parallel to it instead. 
3813: Let us remember that a static system containing an electric charge 
3814: $q$ and a magnetic charge $P$ has an angular momentum $J=Pq$ independent 
3815: of the distance between the charges and 
3816: directed from $q$ to $P$ \cite{Coleman:1982cx}.
3817: This angular 
3818: momentum can be obtained by integrating the 
3819: Poynting vector as in Eq.\eqref{Poynt}. 
3820: We know that a rotating monopole-antimonopole pair contains two magnetic charges
3821: of the opposite sign and two electric charges of the same sign. It can therefore be
3822: visualized as a superposition of two pairs, $(+P,q)$ and $(-P,q)$, both of which making 
3823: {\it the same} contribution to the angular momentum, as shown in Fig.\ref{FigPq}. 
3824: 
3825: 
3826: 
3827: A similar construction of rotating solutions 
3828: can be carried  out using 
3829: other known static solutions of the non-Abelian theory \eqref{YMH},
3830: as for example the monopole rings described in Sec.\ref{rings}, 
3831: also for the generic Higgs field potential \cite{Paturyan:2004ps},\cite{Kleihaus:2005fs}. 
3832: It turns out that the following general relation holds
3833: \be
3834: J=\frac{2\pi m\, Q_{\rm el}}{g}\,[1+(-1)^k]
3835: \ee 
3836: where $m,k$ are the winding
3837: numbers in the ansatz \eqref{mn}
3838:  \cite{Paturyan:2004ps}, \cite{vanderBij:2002sq}, \cite{Kleihaus:2005fs}. 
3839: Therefore, using Eq.\eqref{Qmagn}, the 
3840: angular momentum is zero in the presence of a 
3841: nonvanishing magnetic charge, even though  
3842: $T_\varphi^0\neq 0$ \cite{Kleihaus:2005fs},
3843: while the relation 
3844: \eqref{Jmm} is still valid for solutions with zero magnetic charge. 
3845: 
3846: Since they are related to the monopole-antimonopole solution known to be 
3847: unstable  \cite{Taubes:1982ie}, rotating monopole-antimonopole pairs are very 
3848: probably unstable as well. However, the mechanism of their rotation is interesting
3849: and suggests, in particular, a 
3850: possibility to have stationary, non-radiating spinning solitons
3851: {\it without} axial symmetry.  
3852: 
3853: Specifically, in theories without radiation, as in classical mechanics, 
3854: steadily  rotating bodies can be totally asymmetric.  However, if they carry 
3855: an electric charge distribution, say, then their electric dipole momentum will depend on time, 
3856: hence  
3857: producing radiation, unless they are axially symmetric. This suggests that 
3858: non-radiating spinning field systems should be axially symmetric -- the assumption 
3859: usually made when studying spinning solitons.   
3860: 
3861: An interesting illustration  of this has been found by Hen and 
3862: Karliner \cite{Hen}, who analyze situations where rapidly spinning systems can loose 
3863: axial symmetry. A classical example of this phenomenon
3864: is provided by the rotating Jacobi ellipsoids in Newtonian gravity --
3865: a theory without radiation. Hen and Karliner study spinning baby skyrmions on a 
3866: {\it compact} (sphere or disk) 2D space, in which case there is no radiation,
3867: and find that the system looses axial symmetry for large enough $J$. 
3868: However, as was discussed above, in the limit of infinite $\mathbb{R}^2$,
3869: when radiation can exist, spinning baby skyrmions remain 
3870: axially symmetric for any value of $J$ \cite{Piette} --
3871: since otherwise they would radiate. 
3872: 
3873: 
3874:  
3875: 
3876: Now, the rotating monopole-antimonopole pairs suggest a possibility to have
3877: non-radiating, non-axially symmetric rotation even in theories with radiation. 
3878: This can be achieved by simply taking three or more dyons not aligned 
3879: along one direction.  Of course, this should be realized by constructing a 
3880: smooth, static, non-axially symmetric, finite energy solutions  
3881: in a non-Abelian gauge field theory. If they exist, they may 
3882: have a non-zero $J$ due to the same mechanism as 
3883: for the rotating monopole-antimonopole pairs. 
3884: 
3885: 
3886: 
3887: 
3888: 
3889: 
3890: 
3891: 
3892: 
3893: 
3894: 
3895: 
3896: 
3897: 
3898: 										
3899:  
3900: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
3901: \section{Vortons \label{sec-vortons}}
3902: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
3903: We are now ready to explicitly construct vortons 
3904: as localized,
3905: finite energy  
3906:  solutions of the elliptic field equations. By construction, they are  
3907: stationary and non-radiating. 
3908: Our results can be viewed as complementary 
3909: to those obtained by 
3910: Battye, Cooper and Sutcliffe 
3911: \cite{Battye:2001ec} (although in a completely different context) 
3912: and by Lemperier and Shellard \cite{Lemperiere:2003yt}.  
3913: 
3914: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
3915: \subsection{The Witten model}
3916: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%  
3917: 
3918: 
3919: Vortons were originally  suggested 
3920: \cite{Davis:1988jp},
3921: \cite{Davis:1988ij},
3922: \cite{Davis:1988ip},
3923: \cite{Davis:1988jq} 
3924: in the context of Witten's model of
3925: superconducting cosmic strings \cite{Witten:1984eb}. 
3926: This model contains two Abelian vectors $A^{(a)}_\mu$ $(a=1,2)$ interacting
3927: with two complex scalars $\phi$ and $\sigma$ with the Lagrangian density 
3928: \be  
3929: \label{W}
3930: {\mathcal L}_{W}=-\frac14\sum_{a=1,2} F^{(a)}_{\mu\nu}F^{(a)\mu\nu}  
3931: +D_\mu \phi^\ast D^\mu \phi + 
3932: D_\mu \sigma^\ast D^\mu \sigma -U.
3933: \ee 
3934: Here $F^{(a)}_{\mu\nu}=\partial_\mu A^{(a)}_\nu-\partial_\nu A^{(a)}_\mu$
3935: are the Abelian field strengths, the gauge covariant derivatives of the 
3936: scalars are
3937: $D_\mu \phi=(\partial_\mu-ig_{1}A^{(1)}_\mu)\phi$ and 
3938: $D_\mu \sigma=(\partial_\mu-ig_{2}A^{(2)}_\mu)\sigma$
3939: where $g_{1}$ and $g_{2}$ are the gauge coupling constants. 
3940: The scalar field potential is 
3941: \be                                                          \label{potential}
3942: U=
3943: \frac{1}{4}\lambda_\phi (|\phi|^2-\eta_\phi^2)^2
3944: +\frac{1}{4}\lambda_\sigma |\sigma|^2(|\sigma|^2-2\eta_\sigma^2) 
3945: +\gamma |\phi|^2 |\sigma|^2,
3946: \ee
3947: where $\lambda_\phi,\lambda_\sigma~,\eta_\phi,\eta_\sigma$ and $\gamma$ 
3948: are positive constants. This theory is invariant under local 
3949: U(1)$\times$U(1) gauge transformations, 
3950: so that there are two conserved Noether 
3951: currents.  
3952: 
3953: The theory admits stationary, cylindrically symmetric solutions of the vortex type,
3954: supporting a constant non-zero value of one of the Noether currents
3955: \cite{Witten:1984eb},
3956: \cite{Babul:1987me},
3957: \cite{Haws:1988ax},
3958: \cite{Hill:1987qx},
3959: \cite{Davis:1988jp},
3960: \cite{Amsterdamski:1988zp},
3961: \cite{Peter}.
3962: These 
3963: are the superconducting cosmic strings.
3964: Vortons are supposed to be loops
3965: made of these strings. 
3966: 
3967: 
3968: 
3969: Within the full gauged model \eqref{W} explicit vorton constructions 
3970: have never  been attempted -- due to the complexity of the problem.  
3971: However, the problem simplifies in the global 
3972: limit of this model, for $g_1=g_2=0$. The gauge fields then decouple 
3973: and the theory reduces to   
3974: \be                         \label{lag}
3975: {\mathcal L}=  
3976: \partial_\mu \phi^\ast \partial^\mu \phi + 
3977: \partial_\mu \sigma^\ast \partial^\mu \sigma -U
3978: \ee 
3979: so that the U(1)$\times$U(1) internal symmetry becomes global. 
3980: This global theory still keeps some essential 
3981: features of the original local model. 
3982: In particular, it still admits superconducting vortex solutions, even though
3983: the corresponding current is now global and not local. One can therefore
3984: study global vortons made of these vortices, which was in fact 
3985: the subject of Refs.\cite{Battye:2001ec},\cite{Lemperiere:2003yt}. 
3986: Below we shall 
3987: construct the global vortons as stationary solutions of the field equations
3988: in the model \eqref{lag}, 
3989: which has not been done before. The question of whether
3990: these solutions can be generalized within the full gauged model \eqref{W}
3991: remains open. 
3992: 
3993:  
3994:  
3995: 
3996: One can  absorb $\eta_\phi$  in the definition
3997: of $\phi,\sigma$ in \eqref{lag} to achieve $\eta_\phi=1$. 
3998: Since the overall 
3999: normalization of the potential could be changed by 
4000: rescaling the spacetime coordinates, $x^\mu\to\Lambda x^\mu$,
4001: one can impose one more condition on the remaining four parameters
4002: $\lambda_\phi,\lambda_\sigma,\eta_\sigma,\gamma$, although we 
4003: do not use this option. 
4004: 
4005: 
4006: 
4007: A minimal value of the potential is achieved for
4008: $|\phi|=\eta_\phi =1$ and $|\sigma|=0$, in which case $U=0$. 
4009: This minimum is global if $4\gamma^2>\lambda_\sigma\lambda_\phi$.
4010: The perturbative spectrum 
4011: of field excitations around this vacuum consists of two massless 
4012: Goldstone particles, corresponding to excitations of the phases of the fields,
4013: and of two Higgs bosons with the masses
4014: \be                                                        \label{MASSES}
4015: M_\phi=\sqrt{\lambda_\phi},~~~~~
4016: M_\sigma=\sqrt{\gamma-\frac12\,\lambda_\sigma\eta_\sigma^2}.
4017: \ee  
4018: 
4019: 
4020: 
4021: The global U(1)$\times$U(1) symmetry of the theory,
4022: $\phi\to\phi e^{i\alpha_1}$, $\sigma\to\sigma e^{i\alpha_2}$, 
4023: leads to the conserved currents
4024: \be
4025: \label{curr}  
4026: j^\mu_{(\phi)}=2\Re(i\phi^\ast\partial^\mu \phi),~~~~~~~~~~
4027: j^\mu_{(\sigma)}=2\Re(i\sigma^\ast\partial^\mu \sigma),
4028: \ee 
4029: with $\partial_\mu J^\mu_{(\phi)}=\partial_\mu J^\mu_{(\sigma)}=0$.
4030: The energy-momentum tensor is
4031: \be
4032: \label{Tik}  
4033: T_{\mu\nu}=
4034: \partial_\mu \phi^\ast \partial_\nu \phi+\partial_\nu \phi^\ast \partial_\mu \phi
4035: +\partial_\mu \sigma^\ast \partial_\nu \sigma
4036: +\partial_\nu \sigma^\ast \partial_\mu \sigma
4037: -g_{\mu\nu}{\mathcal L}~,
4038: \ee
4039: where $g_{\mu\nu}$ is the spacetime metric. 
4040: 
4041: The Lagrangian field equations in the theory \eqref{lag} read 
4042: \be                             \label{eW}
4043: \partial_\mu\partial^\mu\phi+
4044: \frac{\partial U}{\partial |\phi|^2}\,\phi=0,~~~~~~~~~
4045: \partial_\mu\partial^\mu\sigma+
4046: \frac{\partial U}{\partial |\sigma|^2}\,\sigma=0.~
4047: \ee
4048: Expressing the complex scalar fields in terms of their amplitudes and phases as
4049: \be
4050: \phi=f_1\,e^{i\psi_1},~~~~~
4051: \sigma=f_2\,e^{i\psi_2},
4052: \ee
4053: equations \eqref{eW} assume the form ($a=1,2$)
4054: \be                                           \label{eqs}
4055: \partial_\mu\partial^\mu f_a=(\partial_\mu\psi_a\partial^\mu\psi_2)f_a
4056: -\frac12\,\frac{\partial U}{\partial f_a} \, ,~~~~~~
4057: \partial_\mu(f_a^2\partial^\mu\psi_a)=0.
4058: \ee
4059: These equations admit cylindrically symmetric vortex solutions discussed 
4060: by several authors (see $e.g.$
4061:  \cite{Davis:1988jp},
4062: \cite{Lemperiere:2003yt},
4063: \cite{Hartmann:2008yr}).
4064:  For these solutions one has  
4065: \be
4066: \label{vortices}  
4067: \phi=f_1(\rho)e^{i n \varphi},~~~~~~~~
4068: \sigma =f_2(\rho)e^{i(p z+\omega t)}~, 
4069: \ee
4070: where $n\in\mathbb{Z}$ and $p,\omega\in\mathbb{R}$, with 
4071: $0\leftarrow f_1(\rho)\to 1$ and $f_2(0) \leftarrow f_2(\rho)\to 0$
4072: as $0\leftarrow\rho\to\infty$, respectively. 
4073: The field $\phi$ thus vanishes in the vortex core and approaches the finite vacuum 
4074: value at infinity, while $\sigma$ develops a non-zero condensate value in the core
4075: and vanishes at infinity. The fields $\phi,\sigma$ are sometimes 
4076: referred to as vortex field and condensate field, respectively. 
4077: The phase of $\phi$ changes by $2\pi n$ after one revolution around the vortex,
4078: while the phase of $\sigma$ increases along the vortex. 
4079: The $z$-dependence of the condensate field gives rise to a non-zero momentum 
4080: along the vortex, $P=\int T^0_z d^3 x\sim p$. 
4081: 
4082: The qualitative vorton construction usually discussed in the literature 
4083: suggests taking a piece of length $L$ of the vortex and identifying its 
4084: extremities 
4085: to form a loop. The momentum along the vortex will then circulate along the loop,
4086: thus becoming an angular momentum, and the centrifugal force which arises 
4087: is supposed to be able to compensate the tension of the loop, 
4088: thereby  producing an equilibrium configuration. 
4089: The coordinate $z$ along the vortex then becomes periodic and can be replaced
4090: by the azimuthal angle $\varphi$, so that 
4091: $p$ will have to assume discrete values, $p=2\pi m/L$ with $m\in\mathbb{Z}$.
4092: The central axis of the vortex where the field $\phi$ vanishes 
4093: then becomes a circle of radius $\RRR=L/2\pi$.  
4094: 
4095: \subsection{Vorton topology and boundary conditions}
4096: 
4097: The above considerations 
4098: suggest describing the vortons by the 
4099: ansatz
4100: \be                              \label{anz}
4101: \phi=f_1(\rho,z)e^{i\psi_1(\rho,z)},~~~~~~
4102: \sigma=f_2(\rho,z)e^{im\varphi+i\omega t}.~
4103: \ee
4104: The phase of $\sigma$ 
4105: increases by $2\pi m$ after one revolution around the $z$-axis, so that  
4106: $f_2(\rho,z)$ should vanish at the axis for the field to be regular there. 
4107: The phase $\psi_1$ 
4108: increases by $2\pi n$ after one revolution around 
4109: the boundary $C$ of the $(\rho,z)$ half-plane, which is the $z$-axis plus the
4110: infinite semi-circle (see Fig.\ref{Fig:top}). 
4111: The regularity condition then implies that $\phi(\rho,z)$ must have $n$ zeros
4112: somewhere inside $C$.   
4113: \begin{figure}[ht]
4114: \hbox to\linewidth{\hss%
4115: %	\resizebox{8cm}{4cm}{\includegraphics{S.eps}}
4116: %\hspace{5mm}%
4117: 	\resizebox{3.7cm}{7cm}{\includegraphics{phase.eps}}
4118: 	\hss}
4119: 	\caption{Vorton topology. 
4120: The phase of $\phi$ increases by $2\pi n$ after
4121: one revolution around the boundary $C$ of the half-plane, 
4122: $\phi$ then vanishing $n$ times somewhere inside $C$. 
4123: In the simplest $n=1$ case the zero is located at a point $(\RRR,0)$
4124: corresponding to the vortex center.
4125: The vortex has a characteristic thickness $\ccc$.}
4126: \label{Fig:top}
4127: \end{figure}
4128: Therefore any vorton configuration can be characterized by a 
4129: pair of integers ($n,m$), which we shall call, respectively,
4130: vortex winding number and azimuthal winding number. 
4131: 
4132: It is instructive to compare these boundary conditions to those 
4133:  for the Faddeev-Hopf knots in Eq.\eqref{axial-CP}.
4134: They are the same --
4135: the phases of $\phi,\sigma$ wind, respectively, along  the orthogonal directions 
4136: shown in Fig.\ref{FigS}. The vortons and knots have therefore similar topology. 
4137: This suggests introducing the `topological charge' 
4138: analogues to the Hopf charge \eqref{Hopf-axial}
4139: \be                                 \label{topol}
4140: \Q=nm.
4141: \ee
4142: Strictly speaking, this quantity will be topologically invariant, that is
4143: unchanged under arbitrary continues field deformations, only in the 
4144: sigma model limit defined by the condition \eqref{constraint} below.
4145: However, we shall call it topological charge in all cases and shall  
4146: require  the vortons to have $\Q\neq 0$.  
4147: 
4148: 
4149: 
4150: 
4151: 
4152: Let us  pass in \eqref{anz} to spherical coordinates,
4153: \be                        \label{ansatz} 
4154: \phi = f_1(r,\vartheta) e^{i\psi_1(r,\vartheta)},
4155: ~~~~~~~
4156: \sigma = f_2(r,\vartheta) \exp[ i ( m \varphi+\omega t)].
4157: \ee
4158: Inserting this to Eqs.\eqref{eqs} gives 
4159: \bea
4160: \Delta f_1-\left((\nabla\psi_1)^2+\frac{\lambda_\phi}{2}(f_1^2-1)
4161: +\gamma f_2^2\right)f_1&=&0\,, \notag \\
4162: \Delta f_2-\left(\frac{m^2}{r^2\sin^2\vartheta}-\omega^2
4163: +\frac{\lambda_\sigma}{2}(f_2^2-\eta_\sigma^2)
4164: +\gamma f_1^2\right)f_2&=&0\,, \notag \\
4165: \nabla(f_1^2\nabla\psi_1)&=&0.
4166: \eea
4167: At infinity the fields are required to approach the vacuum values, $f_1\to 1$,
4168: $f_2\to 0$. One finds for large $r$ 
4169: \bea                              \label{tail}
4170: \psi_1&=&\frac{A\cos\vartheta}{r^2}+\ldots,~~~~~~~~~~
4171: f_2=\frac{B}{r}\,\exp(-\sqrt{M_\sigma^2-\omega^2}~ r)+\ldots \,, \nonumber \\
4172: f_1&=&1+\frac{A^2}{r^6}\,(1+3\cos^2\vartheta)+\ldots
4173: +\frac{C}{r}\exp(-M_\phi~ r)+\ldots\,,
4174: \eea
4175: where $A,B,C$ are integration constants and the dots denote the subleading terms. 
4176: These asymptotic expansions  show the presence of 
4177: two massive Higgs modes and a long range massless Goldstone mode.
4178: The second Goldstone field is not excited within the ansatz \eqref{ansatz}. 
4179: Introducing the notation 
4180: \be                           \label{ansatz:1}
4181: X=f_1\cos\psi_1,~~~~~Y=f_1\sin\psi_1,~~~~~Z=f_2,
4182: \ee
4183: the equations assume the form 
4184: \begin{eqnarray}
4185: \nonumber 
4186: &&\Delta X=\left(\frac{\lambda_\phi }{2}(X^2+Y^2-1)+\gamma Z^2\right)X,
4187: \\
4188: \label{eqs1} 
4189: &&\Delta Y=\left(\frac{\lambda_\phi }{2}(X^2+Y^2-1)+\gamma Z^2\right)Y,\nonumber 
4190: \\
4191: \label{eqs2} 
4192: &&\Delta Z=\left(
4193: \frac{m^2}{r^2\sin^2\vartheta}-\omega^2+
4194: \frac{\lambda_\sigma }{2}(Z^2-\eta_\sigma^2)+\gamma (X^2+Y^2)\right)Z. \label{eqs3} 
4195: \end{eqnarray} 
4196: Using \eqref{tail}, one has at large $r$
4197: \be
4198: X=1+O(r^{-4}),~~~~Y=O(r^{-2}),~~~~
4199: Z\sim \frac{1}{r}\,\exp(-\sqrt{M_\sigma^2-\omega^2}~ r).
4200: \ee
4201: Introducing 
4202: \be
4203: \label{T} 
4204: E_2=\int(|\nabla\phi|^2+|\nabla\sigma|^2)d^3x\,,~~~
4205: E_0=\int U d^3x, ~~~
4206: \ssigma=\int Z^2 d^3x\,, 
4207: \ee 
4208: the energy is 
4209: \be
4210: \label{Energy}  
4211: E=\int T_0^0 d^3x= \omega^2\ssigma+E_2+E_0~,
4212: \ee 
4213: while the Lagrangian
4214: \be
4215: L=\int{\mathcal L}\,d^3x=\omega^2\ssigma-E_2-E_0\,.
4216: \ee
4217: One of the two Noether charges is non-vanishing,  
4218: \be
4219: \label{charge}
4220: Q=\int d^3 xJ^0_{(\sigma)}=2\omega\ssigma,
4221: \ee
4222: and this gives rise to the angular momentum  
4223: \be
4224: \label{J}  
4225: J=\int T_\varphi^0 d^3x =mQ.
4226: \ee
4227: Rescaling the spatial coordinates as $\bx\to\Lambda \bx$, the 
4228: Lagrangian  $L$ should be stationary, which implies 
4229:  the virial relation
4230: \be
4231: \label{virial} 
4232:  E_2=3(\omega^2\ssigma-E_0).
4233: \ee
4234: The presence of the term $({m^2Z^2}/{r^2\sin^2\vartheta})$ in  
4235: $|\nabla\sigma|^2$ requires that for finite energy fields one should have $Z=0$
4236: at the $z$-axis.   
4237: Let us now remember that the phase $\psi_1$ of $\phi=X+iY=f_1e^{i\psi_1}$ 
4238: should increase by 
4239: $2\pi n$ after one revolution around the boundary of the $(\rho,z)$ half-plane.  
4240: Since one has $X=1$, $Y=0$ at $r\to\infty$, it follows that 
4241: the phase is constant at infinity,  so that it 
4242: can only change along the $z$-axis (see Fig.\ref{Fig:top}). Therefore,
4243: \be                                            \label{topology}
4244: \psi_1(\rho=0,z=\infty)-\psi_1(\rho=0,z=-\infty)=-2\pi n,
4245: \ee
4246: from where it follows that the functions $X$ and $Y$ have (at least) 
4247: $2n$  and $2n+1$ zeros at the $z$-axis, respectively. Such a behavior
4248: is compatible with the assumption that $X$ is symmetric and $Y$ is 
4249: antisymmetric under $z\to -z$. Assuming that $Z$ is symmetric,
4250: one arrives  at the following parity conditions, 
4251: \be
4252: X(r,\vartheta)=X(r,\pi-\vartheta),~~~~~
4253: Y(r,\vartheta)=-Y(r,\pi-\vartheta),~~~~~
4254: Z(r,\vartheta)=Z(r,\pi-\vartheta),
4255: \ee
4256: which allow one to restrict in the analysis the range of $\vartheta$ to the 
4257: interval $[0,\pi/2]$. Summarizing everything together,
4258: solutions of Eqs.\eqref{eqs3} should satisfy the following boundary conditions. 
4259: 
4260: {At the symmetry axis, $\vartheta=0$},
4261: functions $X,Y$ should 
4262: have, respectively, $n$ and $n-1$ zeros for $0<r<\infty$,
4263: and one should also have 
4264: \be                                      \label{1}
4265: \partial_\vartheta X=\partial_\vartheta Y=Z=0.
4266: \ee 
4267: {At the origin, $r=0$},
4268: \be                                \label{2}
4269: \partial_r X=Y=Z=0.  
4270: \ee
4271: {At infinity, $r=\infty$}, 
4272: \be                                 \label{3}
4273: X=1,~~~~ Y=Z=0.  
4274: \ee
4275: {In the equatorial plane, $\vartheta=\pi/2$},
4276: \be                                   \label{4}
4277: \partial_\vartheta X=\partial_\vartheta Z=Y=0.
4278: \ee
4279: 
4280: 
4281: 
4282: 
4283: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
4284: 
4285: \subsection{The sigma model limit \label{o}}
4286: We have succeeded for the first time 
4287: to solve the equations  in the special limit 
4288: of the theory obtained by choosing  
4289: the parameters in the potential \eqref{potential} as
4290: \be                                  \label{sm-limit}
4291:  \lambda_\sigma=\lambda_\phi=\beta,~~\eta_\sigma=1,~~
4292:  \gamma=\frac{1}{2}\,\beta+\gamma_0,
4293:  %{\rm with}~~\beta\to\infty~,
4294: \ee
4295: such that the potential becomes
4296: \be                                \label{pot1} 
4297: U=\frac{\beta}{4}(|\phi|^2+|\sigma|^2-1)^2+\gamma_0|\phi|^2|\sigma|^2.
4298: \ee
4299: Taking 
4300: $\beta\to\infty$
4301: enforces the constraint 
4302: \be                                       \label{constraint}
4303: |\phi|^2+|\sigma|^2=1\,
4304: \ee
4305: and the theory \eqref{lag} becomes a sigma model,
4306: \be 
4307: \label{sig}
4308: {\mathcal L}=  
4309: \partial_\mu \phi^\ast \partial^\mu \phi + 
4310: \partial_\mu \sigma^\ast \partial^\mu \sigma -\gamma_0|\phi|^2|\sigma|^2.
4311: \ee
4312: %
4313: %It is worth noting that the quantity $\Q=mn$ becomes now a genuine topological 
4314: This model is simpler than the full theory \eqref{lag}.  
4315: It has less degrees of freedom
4316: and no free parameters, since the value of $\gamma_0$ 
4317: can be changed by rescaling the spacetime
4318: coordinates. The field mass $M_\sigma$ determined by  
4319: \eqref{MASSES} reduces now 
4320: to $\sqrt{\gamma_0}$, while the mass $M_\phi$ 
4321: becomes infinite, which means that field 
4322: degrees of freedom violating the constraint  
4323: \eqref{constraint} are excluded from the dynamics. 
4324: The theory therefore contains only one massive 
4325: particle and two Goldstone bosons in the spectrum. 
4326: It is worth noting that $\Q=mn$ now becomes genuinely
4327: topological charge, invariant under arbitrary smooth field deformations.   
4328: 
4329: 
4330:  It is convenient to use the Lagrange multiplier method by  
4331: adding to the Lagrangian \eqref{sig} 
4332: a term $\mu (1-|\phi|^2-|\sigma|^2)$ 
4333: and varying with respect to $\phi,\sigma$ and $\mu$.
4334: Using the same ansatz \eqref{ansatz},\eqref{ansatz:1}
4335: as before gives  
4336: \begin{align}                  \label{eqsa}
4337: \nonumber 
4338: &\Delta X=\left(\gamma_0 Z^2+\mu\right)X,     \\
4339: &\Delta Y=\left(\gamma_0 Z^2+\mu\right)Y,\nonumber   \\
4340: &\Delta Z=\left(
4341: \frac{m^2}{r^2\sin^2\vartheta}-\omega^2+
4342: \gamma_0 (X^2+Y^2)+\mu\right)Z. 
4343: \end{align} 
4344: and also 
4345: \be                   \label{c}
4346: X^2+Y^2+Z^2-1=0.
4347: \ee
4348: Multiplying the three equations \eqref{eqsa}, respectively, by 
4349: $X,Y,Z$ and taking their sum and using the constraint \eqref{c}
4350: one finds the expression for the Lagrange multiplier,  
4351: \bea                                    \label{lambda}
4352:  \mu &=& X\Delta X+
4353: Y\Delta Y
4354: +
4355: Z\Delta Z
4356: -2\gamma_0 (X^2+Y^2)Z^2
4357: -(\frac{m^2}{r^2\sin^2 \theta}-\omega^2)Z^2 \nonumber \\
4358: &-&\frac12\, \Delta(X^2+Y^2+Z^2),
4359: \eea
4360: where the second line has been added in order to 
4361: cancel the second derivatives appearing in the first line. 
4362: Inserting this to \eqref{eqsa}, 
4363: the constraint \eqref{c} will be imposed automatically
4364: on the solutions, so that it can be excluded from considerations 
4365: from now on. 
4366: 
4367: 
4368: The boundary conditions for Eqs.\eqref{eqsa} are given by 
4369: Eqs.\eqref{1}--\eqref{4}. 
4370: The regularity condition 
4371: $Z|_{\theta=0,\pi}=0$ imposes the 
4372: constraint $X^2+Y^2=1$ on the $z$-axis.
4373: In addition, since at the origin one has $Y=0$, 
4374: the constraint requires that 
4375: $X^2=1$, and so one can choose 
4376: \be                                 \label{n}
4377: X|_{r=0}=-1\,.
4378: \ee
4379: Since one has $X|_{r=\infty}=+1$, this will guarantee a non-zero 
4380: value of $n$, which is the principal advantage of the 
4381: sigma model theory \eqref{sig}. 
4382:  
4383: In fact, making sure that the phase of $\phi$ winds $n\neq 0$ times around 
4384: the contour $C$ is not simple. 
4385: This condition can be naturally implemented  in 
4386: toroidal coordinates \eqref{toroidal}, but these coordinates 
4387: are somewhat singular at infinity. 
4388: Spherical or cylindrical coordinates are  better suited for numerics, but 
4389: the field topology is then determined by 
4390: zeros of the $X,Y$ amplitudes 
4391: (see Eq.\eqref{topology}), whose number 
4392:  cannot be generically prescribed,
4393: since they  are able to disappear  during numerical iterations. 
4394:  Now,
4395: the condition \eqref{n} enforces
4396: a non-trivial field topology at least in the $n=1$ case. 
4397: 
4398: We can now solve the problem.
4399: Starting from a field configuration satisfying the 
4400:  boundary conditions \eqref{1}--\eqref{4}, \eqref{n}
4401: we numerically iterate it 
4402: until the convergence is achieved. 
4403:  The resulting solutions   
4404: are qualitatively very similar to the generic vortons described
4405: in detail below, and they also 
4406: agree with the `skyrmions' 
4407: in the theory of Bose-Einstein condensation previously 
4408: obtained in a completely different way  by Battye, Cooper and Sutcliffe 
4409: \cite{Battye:2001ec}. The latter issue will be discussed below in more detail. 
4410: 
4411: 
4412: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4413: \subsection{Explicit vorton solutions}
4414: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4415: 
4416: Having obtained vortons in the sigma model limit, 
4417: we can use them as the starting profiles in order to iteratively 
4418: descend from  infinite to finite values of $\beta$ in 
4419: the potential \eqref{pot1}, thus relaxing the sigma model condition.
4420:  After this, we can relax also 
4421: the conditions \eqref{sm-limit}, thereby  recovering vorton
4422: solutions within the full original theory \eqref{lag}.  
4423: 
4424: 
4425: 
4426: 
4427: 
4428: 
4429: In our numerics we use the program FIDISOL (written in Fortran),
4430: based on the iterative Newton-Raphson method.
4431:  A detailed presentation of the FIDISOL code is given in \cite{FIDISOL},\cite{FIDISOL1}.
4432: In this approach  the field equations are discretized on a ($r,~\vartheta$) 
4433: grid with $N_r\times N_{\vartheta}$ points,
4434: and the resulting system is iterated until convergence is achieved. 
4435: The grid spacing in the $r$-direction is non-uniform, while the 
4436: values of the grid points
4437: in the angular direction are given by $\vartheta_k=(k-1)\pi/(2(N_{\vartheta}-1))$.
4438: Instead of $r$, 
4439: a new radial variable $x$ is introduced which maps the 
4440: semi-infinite region $[0,\infty)$ to the finite interval $[0,1]$.
4441:  There
4442: are various possibilities for such a mapping, 
4443: a flexible enough choice being $ x=r/(c+r)$, where $c$
4444: is a properly chosen constant.
4445: Typical grids for the $n=1$ solutions have around $200 \times 30$ points. 
4446: The typical  numerical error 
4447: for the solutions is estimated to be of order $10^{-3}$.
4448: In addition, the virial relation (\ref{virial}) was also monitored.
4449: The deviation from unity of the  ratio  $E_2/3(\omega^2\ssigma-E_0)$
4450: is less than $10^{-3}$ for most of the solutions we have considered.
4451: 
4452: \begin{figure}[ht]
4453: \hbox to\linewidth{\hss%
4454: %	\resizebox{7cm}{7cm}{\includegraphics{vector-vorton0.eps}}
4455: %\hspace{5mm}%
4456:         \resizebox{7cm}{7cm}{\includegraphics{plot-phi-psi.eps}}	
4457: \hss}
4458: 	\caption{Levels of constant amplitude $|\phi|$ (closed lines)
4459: and phase $\psi_1$ (radial lines, red online) on the $\rho$ 
4460: (horizontal direction)--$z$ (vertical direction) plane 
4461: for the vortex field $\phi=|\phi|\exp(i\psi_1)$ 
4462: corresponding to the
4463: vorton solution with $n=1$, $m=2$, $\omega=0.85$, 
4464: $\lambda_\phi=41.12$,
4465: $\lambda_\sigma=40$,
4466: $\eta_\sigma=1$, 
4467: $\gamma=22.3$.}
4468: \label{Fig:phase}
4469: \end{figure}
4470: 
4471: As a result, for given values of the parameters  
4472: $\lambda_\phi$, $\lambda_\sigma$, $\eta_\sigma$ and $\gamma$ 
4473: in the potential, we can specify the azimuthal winding number $m$
4474: and the frequency $\omega$ and obtain a vorton solution. 
4475: The value of $n$, the charge $Q$ and   
4476: energy $E$ are computed from the numerical output.
4477: A complete analysis of the parameter space of solutions  
4478: is a time consuming  task that we did not aim at.
4479: Instead,  
4480: we analyzed in detail a few particular classes of solutions, 
4481: which hopefully reflects the essential properties 
4482: of the general pattern.
4483: 
4484: 
4485: %\subsubsection{$n=1$ vortons}
4486: 
4487: 
4488: 
4489: We mainly considered the case where  the function $X$ vanishes
4490: once at the positive $z$ semi-axis, which corresponds to the vortex winding
4491: number  $n=1$.
4492: The amplitude of the   scalar field $\phi$
4493: has in this case one zero 
4494: located on a circle in the $\vartheta=\pi/2$ plane.
4495: The corresponding vorton topology  
4496: can be illustrated
4497: by the diagram in Fig.\ref{Fig:phase}, where the level lines
4498: of the complex vortex function $\phi=|\phi|\exp(i\psi_1)$ are shown. 
4499: The levels of constant amplitude, $|\phi|$, are closed lines encircling
4500: the center of the vortex where $\phi$ vanishes. 
4501: Emanating from this center there are lines of constant phase. 
4502: The phase, $\psi_1$, 
4503: increases  by $2\pi$ after one revolution around the center.  
4504: 
4505: \begin{figure}[ht]
4506: \hbox to\linewidth{\hss%
4507: 	\resizebox{7cm}{6cm}{\includegraphics{vort-X-n=1.eps}}
4508: \hspace{5mm}%
4509:         \resizebox{7cm}{6cm}{\includegraphics{Y-vorton.eps}}	
4510: \hss}
4511: 	\caption{The amplitudes $X(\rho,z),Y(\rho,z)$ 
4512: for the typical vorton solution; here 
4513: $n=1$, $m=2$, $\omega=0.85$,
4514:    $\lambda_\phi=41.12$,
4515: $\lambda_\sigma=40$, $\eta_\sigma=1$, $\gamma=22.3$.}
4516: \label{Fig3}
4517: \end{figure}
4518: 
4519: \begin{figure}[ht]
4520: \hbox to\linewidth{\hss%
4521: 	\resizebox{7cm}{6cm}{\includegraphics{vort-Z-n=1.eps}}
4522: \hspace{5mm}%
4523:         \resizebox{7cm}{6cm}{\includegraphics{vort-Phi-n=1.eps}}	
4524: \hss}
4525: 	\caption{The amplitudes $Z(\rho,z)$ and $|\phi(\rho,z)|$ for the same
4526: solution as in Fig.\ref{Fig3}.}
4527: \label{Fig3a}
4528: \end{figure}
4529: 
4530: All vortons have toroidal distributions of the 
4531: energy density and charge.
4532: It seems that they  do not 
4533: exist for arbitrary values of the parameters of the model but only for 
4534: some regions in the parameter space.  
4535: In Figs. \ref{Fig3}, \ref{Fig3a} the  3D plots of $X,Y,Z$ and also 
4536: of $|\phi|=\sqrt{X^2+Y^2}$
4537: for a typical solution with  $m=2$ are presented. 
4538: One can see that these functions exhibit a strong  dependence
4539: both on $\rho$ and $z$, with the zero of $|\phi|$ 
4540: located in the $z=0$ plane at $\RRR\simeq 0.82$.
4541: \begin{figure}[ht]
4542: \hbox to\linewidth{\hss%
4543: 	\resizebox{7cm}{6cm}{\includegraphics{vorton-Xu.eps}}
4544: \hspace{5mm}%
4545:         \resizebox{7cm}{6cm}{\includegraphics{vorton-Yu.eps}}	
4546: \hss}
4547: 	\caption{Profiles of  $X(r,\vartheta), Y(r,\vartheta)$ for 
4548: fixed $\vartheta$
4549: 	for the $n=m=1$ vorton solution with   $\omega=0.8$,   $\lambda_\phi=33.26$,
4550: $\lambda_\sigma=32.4$, $\eta_\sigma=1$, $\gamma=18.4$.}
4551: \label{Fig4}
4552: \end{figure}
4553: \begin{figure}[ht]
4554: \hbox to\linewidth{\hss%
4555: 	\resizebox{7cm}{6cm}{\includegraphics{vorton-Zu.eps}}
4556: \hspace{5mm}%
4557:         \resizebox{7cm}{6cm}{\includegraphics{vorton-phiu.eps}}	
4558: \hss}
4559: 	\caption{Profiles of $Z(r,\vartheta),|\phi(r,\vartheta)|$ 
4560: for fixed values of $\vartheta$
4561: 	for the same solution as in Fig.\ref{Fig4}. 
4562: }
4563: \label{Fig4a}
4564: \end{figure}
4565: 
4566: In Fig. \ref{Fig4}, \ref{Fig4a} the 2D profiles of $X,Y$,$Z$ and
4567: $|\phi|$ for a typical  $m=1$ solution for several values of $\vartheta$ are shown; 
4568: these pictures remain 
4569: qualitatively the same for higher values of $m$.
4570: As seen in these plots,
4571: the space can be partitioned into three regions. 
4572: In the first region, located near the origin, the functions  $X,Y,Z$ present  a
4573: strong variation, with $Y$ approaching its extremum.
4574: The second region is located inside the vortex, where $X$
4575: stays very close to zero, while the field $Z$ is almost constant 
4576: and close to its maximal value.
4577: As seen in the insertion in the first plot, $X$ crosses zero value with
4578: a non-zero first derivative, as it should, since the degree of zero of $\phi$
4579: at the vortex center should be one. 
4580: In the third region, outside the vortex ring, 
4581: the fields approach the vacuum values.
4582: Both $X$ and $Z$ show strong variations in a transition domain between second 
4583: and third regions. We notice also that the value of $X$ at the origin, which was  
4584: $-1$ in the sigma model limit, is now larger, although 
4585: it stays always negative and smaller that $-0.5$ for all solutions we have found.
4586: 
4587: 
4588: \begin{figure}[ht]
4589: \hbox to\linewidth{\hss%
4590: 	\resizebox{7cm}{6cm}{\includegraphics{vorton-En1.eps}}
4591: \hspace{5mm}%
4592:         \resizebox{7cm}{6cm}{\includegraphics{vorton-En3.eps}}	
4593: \hss}
4594: \caption{The energy density $T_0^0$ for 
4595: the $n=m=1$ (left) and  $n=1$, $m=3$ (right) vortons with 
4596: $\omega=0.8$,   $\lambda_\phi=39.25$,
4597: $\lambda_\sigma=38.4$, $\eta_\sigma=1$, $\gamma=21.39$.  }
4598: \label{Fig6}
4599: \end{figure}
4600: \begin{figure}[ht]
4601: \hbox to\linewidth{\hss%
4602: 	\resizebox{7cm}{6cm}{\includegraphics{Qn1.eps}}
4603: \hspace{5mm}%
4604:         \resizebox{7cm}{6cm}{\includegraphics{Qn3.eps}}	
4605: \hss}
4606: 	\caption{The charge density for the same 
4607: solutions as in Fig.\ref{Fig6}.}
4608: \label{Fig6a}
4609: \end{figure}
4610:  For the $m=1$ solutions the 
4611: energy density is maximal at the origin.
4612: Emanating from this central maximum there is 
4613: a toroidal shell of energy containing  the ring of radius $\RRR$ 
4614: where $\phi$ vanishes 
4615: (see Fig.\ref{Fig6}). When $m$ increases
4616: $R$ grows, 
4617: whereas the height of the maximum at the origin 
4618: decreases (see Fig.\ref{Fig6}), so that  
4619: the shape of the  energy density resembles   a hollow tube 
4620: (see Fig.\ref{Fig6aa}). 
4621: There is, however, an almost  constant  energy density  
4622: inside the tube, whose value decreases with $m$. 
4623: As seen in Fig.\ref{Fig6a},
4624: the charge of the solutions, although also localized in a compact region,
4625: does not present the tube structure of the energy density.
4626: 
4627: \begin{figure}[ht]
4628: \hbox to\linewidth{\hss%
4629: 	\resizebox{7cm}{6cm}{\includegraphics{n1m3vorton-constant-E=1,05-new.eps}}
4630: \hspace{5mm}%
4631:         \resizebox{7cm}{6cm}{\includegraphics{vorton-varw.eps}}	
4632: \hss}
4633: 	\caption{Left: the surface of constant $(T^0_0=1.05)$ energy density 
4634: for the same $n=1$, $m=3$ vorton solution as in Fig.\ref{Fig6}. 
4635: It has the structure of a torus containing
4636: inside another toroidal surface. Right: 
4637: the energy $E(\omega)$ and charge $Q(\omega)$ for the 
4638: $n=1, m=2$ vorton solutions.}
4639: \label{Fig6aa}
4640: \end{figure}
4641: 
4642: % effect of the winding number  %%%%%%%%%%%%%%%%%%%%%%
4643: Nontrivial vorton solutions are likely to exist for all values of $m$.
4644: We have studied solutions up to $m=5$, but beyond this value 
4645: the numerical accuracy of our procedure decreases drastically.
4646: The energy of the solutions increases with $m$, and 
4647: for solutions with the same $\omega$ one has 
4648: $
4649: E(m)<mE(m=1).
4650: $ 
4651: All solutions we have found  have the ratio
4652: $\RRR/\ccc$ between 1 and 1.5, so that the radius of the vortex ring
4653: is almost the same as the thickness of the vortex. 
4654: The construction of vortons in the thin ring limit, for $\RRR\gg \ccc$, 
4655: remains a numerical challenge.
4656: 
4657: 
4658: 
4659: 
4660: 
4661: Fig.\ref{Fig6aa} (right) shows the energy $E$ and charge $Q$ 
4662: as functions of the frequency 
4663: $\omega$ for $m=2$ vortons.  
4664: These functions are actually quite similar to those
4665: found in the $Q$-ball case. 
4666: As one can see,  
4667: solutions exist for a limited range of frequencies, 
4668: $\omega_{-}<\omega<\omega_{+}$.
4669: Unfortunately, the numerical accuracy decreases considerably near the limits
4670: of this interval. 
4671: However, it seems that
4672:  both $Q(\omega)$ and $E(\omega)$ diverge
4673: in these limits, as in the $Q$-ball case.
4674: 
4675: In analogy with $Q$-balls, the vortons become large as $\omega\to\omega_{\pm}$. 
4676: However, it is not the vorton radius $\RRR$ that becomes large, since it appears to be
4677: not very sensitive to $\omega$ 
4678: and remains finite and nonzero when $\omega$ approaches 
4679: its limiting values. It appears that the torus thickness increases for 
4680: $\omega\to\omega_{\pm}$ so that its volume grows, thereby increasing the 
4681: energy and charge. Both $E$ and $Q$
4682: assume their minimal values at
4683: $\omega=\omega_{\rm crit}$.
4684:  The dependence $E(Q)$ 
4685: appears to be double-valued, very similar to the one shown in Fig.\ref{FigQ3} for $Q$-balls, 
4686: again exhibiting two branches with a cusp.  Solutions with $\omega<\omega_{\rm crit}$
4687: are less energetic than those for $\omega>\omega_{\rm crit}$ and with the
4688: same $Q$. 
4689: 
4690: Using again the analogy with $Q$-balls, it seems likely that only solutions 
4691: from the lower branch can be stable, probably those for large enough charge. 
4692: Stable vortons certainly exist, perhaps not for all values of the parameters
4693: of the potentials, but at least for large enough values, when the theory 
4694: approaches the sigma model limit. Vortons in the sigma model limit 
4695: are stable, since, as will be discussed below, they can be obtained  via 3D 
4696: energy minimization \cite{Battye:2001ec}. This suggests 
4697: that they should remain stable at least
4698: when they are close to this limit. 
4699: 
4700: 
4701: 
4702: 
4703: Although we did not study  in detail the dependence of the solutions 
4704: on the parameters of the potential, 
4705: one should mention that for $\eta_\sigma\simeq 1$ and $\omega\simeq 0.85$
4706: (the values we mainly considered)
4707: we could find solutions only for $\lambda_\phi\gtrsim 22$,
4708: while
4709: the ratio $\lambda_\phi/\lambda_\sigma$ was always
4710:  close to one. 
4711: In other words, our vortons are not that far from the sigma model limit,
4712: which suggests that they could be stable. 
4713:  
4714: 
4715: Vortons with $n>1$ also exist, and  we were able to construct them 
4716: for $n=2$. 
4717: For these solutions there are two concentric rings 
4718: in the $\vartheta=\pi/2$ plane where $|\phi|$ vanishes.
4719: All $n=2$ solutions we have found have very large values of
4720: $\lambda_\phi$, $\lambda_\sigma$, $\gamma$. 
4721: 
4722: Finally, an  interesting class of solutions of Eqs.(\ref{eqs1})
4723: is obtained by setting $Y=0$, which means that the phase of $\phi$ 
4724: is trivial and so $n=0$. Such solutions 
4725: have not been discussed in the literature. 
4726: Although $m\neq 0$ in this case, 
4727: one has $\Q=nm=0$, 
4728: and so according to our definition these solutions are not vortons,
4729: since they are `topologically trivial'. 
4730: However, they  also have a ring structure, resembling  somewhat 
4731: spinning $Q$-balls. One can view them 
4732: as `$\phi$-dressed' $Q$-balls made of the complex $\sigma$ field, 
4733: the non-renormalizable
4734: $|\sigma|^6$ term in the $Q$-balls potential 
4735: being replaced by the interaction with a real scalar field $\phi$,
4736: as was first suggested in Ref.\cite{Friedberg}. 
4737: 
4738: 
4739: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4740: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4741: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4742: \section{Ring solitons in non-relativistic systems}
4743: 
4744: So far we have been considering relativistic ring solitons that could be
4745: relevant in models of high energy physics. At the same time, 
4746: similar objects also exist 
4747: in non-relativistic physics, and below we shall consider some of them which 
4748: are most closely related to our previous discussion. 
4749: 
4750: A famous example of soliton-type 
4751: toroidal systems is provided by the magnetically confined plasma in 
4752: the TOKAMAK's. However, since considering this subject would lead us 
4753: far beyond the scope of our present 
4754: discussion, we shall restrict ourselves to simply giving a reference 
4755: to a recent review \cite{plasma}. In addition, since 
4756: the external field is necessary to confine the plasma in this case, 
4757: this example does not quite fit in with our discussion, because we are
4758: considering closed, self-interacting systems. Another example of plasma confinement,
4759: although somewhat controversial, but
4760: which could perhaps be explained by some solitonic structures, 
4761: is provided by the phenomenon of ball lightning 
4762: (see \cite{Ball} for a bibliography). 
4763: It has in fact been conjectured that some analogs of Faddeev-Skyrme solitons 
4764: in plasma  might be responsible for their existence 
4765: \cite{Faddeev:2000rp},\cite{Faddeev:2000qw}, although we are unaware
4766: of any confirmations of this conjecture via constructing explicit solutions. 
4767:  
4768: Field theory models of condensed matter physics are more closely related to our
4769: discussion.  As in the relativistic case, they  can be either 
4770: theories with local gauge invariance,
4771: as for example the Ginzburg-Landau theory of superconductivity \cite{Ginzburg}, 
4772: or models with 
4773: global internal symmetries,  as in the Gross-Pitaevskii  
4774: \cite{Pitaevskii},
4775: \cite{Pitaevskii1},
4776: \cite{Gross}
4777: theory of Bose-Einstein 
4778: condensation. It has been conjectured that some analogs of the 
4779: Faddeev-Skyrme knot solitons
4780: could exist in the multicomponent Ginzburg-Landau models \cite{Babaev:2001jt}.
4781: However, we are again
4782: unaware of any explicit solutions, and moreover, as was discussed above, 
4783: their existence is not very plausible, at least in the purely
4784: magnetic case. 
4785: On the other hand, global models of condensed matter physics do admit 
4786: ring type solitons. 
4787: 
4788: 
4789:  \subsection{Vortons versus `skyrmions' in Bose-Einstein condensates  \label{VS}}
4790: 
4791: Surprisingly, it turns out that the described above vortons
4792: in the global limit of Witten's model can also be considered  as solutions of the 
4793: Gross-Pitaevskii (GP) equation, 
4794: \be                                    \label{GP}
4795: i\frac{\partial\Psi_a}{\partial t}=
4796: \left(-\frac12\,\Delta + V(\bx)+\frac12\sum_{b}\kappa_{ab}|\Psi_b|^2\right)\Psi_a\,.
4797: \ee
4798: This equation, also called nonlinear Schr\"odinger equation, 
4799: provides an effective description of the Bose-Einstein condensates. 
4800: Here $\Psi_a$ are the condensate order parameters, $V(\bx)$ is the external
4801: trapping potential and $\kappa_{ab}$ describe interactions between different condensate
4802: components. In the case of a two-component condensate the indices  $a,b$ assume 
4803: values $1,2$.  
4804: 
4805:  Let 
4806: \be                              \label{an}
4807: \phi=\phi(\bx),~~~~~~
4808: \sigma=\sigma(\bx)e^{i\omega t}
4809: \ee
4810: be a solution of the equations \eqref{eW} of the global model of Witten \eqref{lag} 
4811: in the case where the parameters in the potential \eqref{potential} are related as
4812: \be                                \label{rel}
4813: \lambda_\phi=\lambda_\sigma\eta_\sigma^2.
4814: \ee
4815: Then setting 
4816: \be
4817: \Psi_1=\,e^{-i\frac{\lambda_\phi}{4}\,t}\,\phi(\bx),
4818: ~~~~\Psi_2= e^{-i\left(\frac{\lambda_\phi}{4}+\frac{\omega^2}{2}\right) t}\,\sigma(\bx)
4819: \ee
4820: gives a solution of the GP  equation \eqref{GP} in the case where 
4821: \be
4822: V(\bx)=0,~~~\kappa_{11}=\frac{\lambda_\phi}{2},~~~
4823: \kappa_{22}=\frac{\lambda_\sigma}{2},~~~
4824: \kappa_{12}=\kappa_{21}=\gamma.
4825: \ee
4826: It follows that any vorton solution satisfying the condition \eqref{rel} (this condition
4827: is typically not difficult to achieve numerically) provides a solutions to the  
4828: Gross-Pitaevskii equation and hence can be interpreted in terms of condensed
4829: matter physics. 
4830: 
4831: In particular, the condition \eqref{rel} is achieved in the 
4832: sigma model limit \eqref{sm-limit},\eqref{constraint},
4833: in which case one obtains solutions of the GP equation with 
4834: \be
4835: V(\bx)=\kappa_{11}=\kappa_{22}=0,~~~~~~~~~
4836: \kappa_{12}=\kappa_{21}=\gamma_0
4837: \ee
4838: by setting $\Psi_1=\phi(\bx)$, $\Psi_2= e^{-i\frac{\omega^2}{2} t}\,\sigma(\bx)$. 
4839: Remarkably, solutions of the GP equation in this case have been independently 
4840: studied by Battye, Cooper and Sutcliffe (BCS) 
4841: \cite{Battye:2001ec}, so that we can compare our results in Sec.\ref{o} with theirs.  
4842: They 
4843: parametrize the fields as 
4844: \be                              \label{anz:3}
4845: \phi(\bx)=\cos\frac{\Theta}{2}\,e^{in\psi}\,,~~~~~~
4846: \sigma(\bx)=\sin\frac{\Theta}{2}\,e^{im\varphi}\,,~
4847: \ee
4848: which exactly coincides with the parametrization \eqref{axial-CP} of the 
4849: Faddeev-Hopf field.  
4850: Instead of directly solving the GP equation, BCS minimize the energy
4851: \be                                    
4852: E_2+E_0=\int(|\nabla\phi|^2+|\nabla\sigma|^2+\gamma_0|\phi|^2|\sigma|^2)d^3\bx
4853: \ee
4854: in a given $\Q$ sector 
4855: by keeping fixed the particle number
4856: \be
4857: {\mcal N}=\int |\sigma|^2 d^3\bx\,.
4858: \ee
4859: They call the resulting configurations $\phi_\ssigma(\bx),\sigma_\ssigma(\bx)$ 
4860: `skyrmions' -- because Eq.\eqref{anz:3} agrees with the 
4861: skyrmion parametrization \eqref{SU1}.  
4862: At first view it is not completely obvious how these 
4863: `skyrmions' are
4864: related 
4865: to our vortons $\phi_\omega(\bx),\sigma_\omega(\bx)$ obtained by solving 
4866: Eqs.\eqref{eqsa} for a given $\omega$. 
4867: However, these are actually the same solutions.
4868: In particular,  the profiles of the vortons described in the previous sections,
4869: their energy density distributions, 
4870: are similar to those for the `skyrmions' 
4871: given in Ref.\cite{Battye:2001ec}.
4872: 
4873: 
4874: 
4875: The two different ways to obtain the solutions, 
4876: either by minimizing the energy or via
4877: solving the equations, are actually equivalent\footnote{We thank 
4878: Richard Battye for explaining this point to us.}. 
4879: This issue has in fact already been discussed above in the $Q$-ball context. 
4880: $Q$-balls  can be obtained either by solving Eq.\eqref{Qe}, which gives 
4881: $\phi_\omega(\bx)$, or via 
4882: minimizing the truncated energy functional $E_0+E_2$ \eqref{NE} 
4883: with the particle number $\ssigma$ \eqref{NN} fixed, which gives 
4884: $\phi_\ssigma(\bx)$. Once one knows 
4885: the value  $\omega=\omega(\ssigma)$, obtained either  
4886: with the formula \eqref{omega2} or from the virial relation \eqref{Qvirial}, the 
4887: two results can be related to each other: 
4888: $\phi_\omega(\bx)=\phi_{\omega(\ssigma)}(\bx)=\phi_\ssigma(\bx)$.
4889: 
4890: Similarly, the BCS skyrmions  $\phi_\ssigma(\bx),\sigma_\ssigma(\bx)$  
4891: can be related to our vortons $\phi_\omega(\bx),\sigma_\omega(\bx)$ 
4892: by simply establishing the relation $\omega(\ssigma)$. 
4893: The knowledge of $\omega$ is important to make sure that the 
4894: no radiation condition
4895: \be
4896: \omega^2\leq M_\sigma^2=\gamma_0\,
4897: \ee
4898: is fulfilled, which is the case for our solutions,
4899: while their profiles agree  with the features described by BCS 
4900: in \cite{Battye:2001ec}.
4901: This provides an independent confirmation of our numerical results 
4902: described in Sec.\ref{o}.  Since they can be obtained
4903: by the 3D energy minimization, these solutions are stable. 
4904: 
4905: Apart from the work of BCS there have been other studies of `skyrmions'
4906: in Bose-Einstein condensates 
4907: \cite{Savage:2003hh},
4908: \cite{Ruostekoski:2003qx},
4909: \cite{Ruostekoski:2004pj},
4910: \cite{Ueda},
4911: \cite{Metlitski:2003gj}. Such solutions  
4912: have been  constructed by numerically resolving the GP equation  
4913: for more general choices of 
4914: $V(\bx)$ and $\kappa_{ab}$ and also for more 
4915: than two condensate components. 
4916: Possibilities for an experimental creation
4917: and observation of such objects have also been discussed 
4918: \cite{Ruostekoski:2001fc},
4919: \cite{Ruostekoski:2005zd}.  
4920: In all studies the solutions are typically
4921: presented for the winding numbers $n=1$, with the  
4922: exception of Ref.\cite{Ruostekoski:2004pj},
4923: where results for $n=2$ are reported, although without giving many
4924: details. In our language this corresponds to the double vortons. 
4925: 
4926: \subsection{Spinning rings in non-linear optics -- 
4927: $Q$-balls as light bullets \label{light}}
4928: 
4929: Yet another application of the considered above ring solitons arises in the 
4930: theory of light propagating in media whose polarization vector $\vec{\mcal P}$
4931: depends non-linearly
4932: on the electric field $\vec{\mcal E}$. The sourceless Maxwell equations become in this case 
4933: non-linear and in a number of important cases they can be reduced to a 
4934: non-linear Schr\"odinger (NLS) equation  similar to the Gross-Pitaevskii equation \eqref{GP}.
4935: For a Kerr medium, when the non-linear part of  $\vec{\mcal P}$ 
4936: is cubic in $\vec{\mcal E}$, the NLS equation has exactly the same structure as Eq.\eqref{GP}, 
4937: but for more general media it can have different non-linearities. Soliton solutions 
4938: of this NLS equation describe non-linear light pulses, 
4939: sometimes called light bullets, which are
4940: very interesting from the purely theoretical viewpoint and which can 
4941: actually be observed,
4942: as for example in the optical fibers.  
4943: Unfortunately, discussing these issues  in more detail would lead us away
4944: from our subjects,  and so we
4945: simply refer to a monograph \cite{akhmediev} on optical solitons.  
4946: 
4947: Most of the known solitons of the NLS equation describe plane waves or 
4948: cylindrical beams of light \cite{akhmediev}. However, solutions
4949: describing solitons localized in all three dimensions are also known, 
4950: they have been studied by 
4951: Mihalache $et~ al$ \cite{mihalache}. These solutions describe spinning rings, 
4952: which is
4953: very interesting in the context of our discussion. 
4954: 
4955: Mihalache $et~ al$ consider light pulses 
4956: travelling in the $z$ direction with a group velocity $V$ 
4957: in a medium with cubic and quintic non-linearities. 
4958: After a suitable rescaling of the coordinates and fields, 
4959: the envelope of the pulse, $\Psi$, satisfies the equation \cite{mihalache}
4960: \be                                       \label{bullet}
4961: i\frac{\partial \Psi}{\partial z}=\left(
4962: -\frac{\partial^2}{\partial x^2}
4963: -\frac{\partial^2}{\partial y^2}
4964: -\frac{\partial^2}{\partial \zeta^2}
4965: +
4966: |\Psi|^4-|\Psi|^2\right)\Psi
4967: \ee
4968:  where $\zeta=z-Vt$. Solutions of this equation, numerically constructed in 
4969: Ref.\cite{mihalache}, describe `spinning light bullets'. 
4970: Interestingly, it turns out that 
4971: these solitons correspond to the spinning
4972: $Q$-balls already described above.  
4973: 
4974: Specifically, let us rename the variables as 
4975: $z\to t$, $\zeta\to z$
4976: and set
4977: $\Psi=\exp\{i(\omega^2-M^2)t\}\phi$. 
4978: Eq.\eqref{bullet} then assumes the form 
4979: \be                                       \label{bullet1}
4980: i\dot{\phi}=(-\Delta+|\phi|^4-|\phi|^2+M^2-\omega^2)\phi\,.
4981: \ee
4982: Let us now consider Eq.\eqref{Qe00} describing $Q$-balls with the 
4983: potential \eqref{pQ}. We notice that rescaling the spacetime 
4984: coordinates and the field as $x^\mu\to \Lambda_1 x^\mu$, $\Phi\to \Lambda_2\Phi$
4985: is equivalent to changing the values of the parameters $\lambda,a$ in the potential. 
4986: Therefore, as long as they do not vanish, one can choose without any loss of generality
4987: $\lambda=1/3$, $a=3/2$. 
4988: Setting then  $\Phi=e^{i\omega t}\phi$, Eq.\eqref{Qe00} assumes the form 
4989: \be                                       \label{bullet2}
4990: -\ddot{\phi}-2i\omega\dot{\phi}   
4991: =(-\Delta+|\phi|^4-|\phi|^2+M^2-\omega^2)\phi\,,
4992: \ee
4993: with $M^2=\lambda b$.  
4994: We see that if 
4995: $\dot{\phi}=0$ then Eq.\eqref{bullet1} and  Eq.\eqref{bullet2} become identical.
4996: They have therefore the same solutions. 
4997: We thus conclude that the $Q$-balls solutions discussed above in Sec.\ref{Qballs}
4998: can describe light pulses in non-linear media.  
4999: The solutions found in Ref.\cite{mihalache} correspond to the 
5000: simplest $m^{+}$, that is non-twisted, even parity $Q$-balls. Let us remind that for 
5001: a given $m$ these solutions can be parametrized by the value of $\omega$ varying 
5002: within the range 
5003: $\omega^2_{-}(m) < \omega^2< \omega^2_{+}=M^2$. 
5004: 
5005: Mihalache $et~ al$ also study the dynamical stability of 
5006: the solutions by analysing their small perturbations \cite{mihalache}.  
5007: Perturbing in 
5008: Eq.\eqref{bullet1} the field as $\phi\to\phi+\delta\phi$ and 
5009: linearizing with respect to $\delta\phi$
5010: gives
5011: \be                                       \label{bullet3}
5012: i\dot{\delta\phi}=\hat{{\mcal D}}[\phi]\delta\phi\,,
5013: \ee
5014: where $\hat{{\mcal D}}[\phi]$ is a linear differential operator. 
5015: If the spectrum of this operator
5016: is known, 
5017: \be                                       \label{bullet4}
5018: \hat{{\mcal D}}[\phi]\xi_\lambda=\lambda\xi_\lambda\,,
5019: \ee
5020: then Eq.\eqref{bullet3} is solved by setting 
5021: $\delta\phi=\exp\{i\lambda t\}\xi_\lambda$. 
5022: It follows that  if there are complex 
5023: eigenvalues $\lambda$ with a negative imaginary part, 
5024: then the perturbations grow for $t\to\infty$ and so the 
5025: background is unstable. 
5026: Mihalache $et~ al$ study the spectral  problem \eqref{bullet4} 
5027: and find complex eigenvalues for all 
5028:  spinning solitons with $m=2$. They  also  
5029: integrate Eq.\eqref{bullet} to trace 
5030: the full non-linear perturbation dynamics, and they observe that the spinning
5031: toroidal  soliton splits into {\it three} individual non-spinning spheroidal pieces
5032: \cite{mihalache}. However, they find that the $m=0,1$ solitons can be stable,
5033: both at the linear and non-linear levels, if  
5034: $\omega$ is less than a certain value $\omega_\ast(m)<M$
5035: but still within the domain of the solution existence, 
5036: $\omega_{-}(m) < \omega<\omega_\ast(m)$. 
5037: According to their results, for $m=0$ one has 
5038: $\omega_\ast=\omega_{\rm crit}$, which is the value where 
5039: $E(\omega),Q(\omega)$ pass through the minimum (see Fig.\ref{FigQ3}).  
5040: For $m=1$ the stability region shrinks, since 
5041: $\omega_\ast<\omega_{\rm crit}$, so that 
5042: not all parts of the less energetic solution branch 
5043: (see Fig.\ref{FigQ3}) correspond to stable
5044: solutions. 
5045: 
5046: These results are very interesting, 
5047: since they allow us to draw  certain
5048:  conclusions about stability of the relativistic $Q$-balls. 
5049: Linearizing  Eq.\eqref{bullet2} gives 
5050: \be                                       \label{bullet5}
5051: -\ddot{\delta\phi}-2i\omega\dot{\delta\phi}=\hat{{\mcal D}}[\phi]\delta\phi\,,
5052: \ee
5053: whose solution is $\delta\phi=\exp\{i\gamma t\}\xi_\lambda$ with 
5054: $\gamma=-\omega\pm\sqrt{\lambda+\omega^2}$. All eigenmodes with complex 
5055: $\lambda$ therefore correspond to complex $\gamma$ and so to unstable modes. 
5056: In addition, there could be unstable modes with 
5057: real $\lambda<-\omega^2$, since $\gamma$ would then be complex. 
5058: As a result, the relativistic $Q$-balls 
5059: have the same unstable modes as their non-relativistic optical counterparts
5060: (although with different instability growth rates) 
5061: and perhaps also some additional instabilities.   
5062: 
5063: It follows that 
5064: the $2^{+}$ spinning $Q$-balls are unstable, 
5065: while the $1^{+}$ ones, if they belong to the lower solution branch 
5066: and their charge is large enough, should be stable   
5067: (unless they have negative modes with  
5068: $\lambda<-\omega^2$). The latter conclusion agrees with the 
5069: results shown in Fig.\ref{FigQ3}. 
5070: 
5071: 
5072: 
5073: As was said above, 
5074: $Q$-balls corresponding to the lower solution branch in Fig.\ref{FigQ3}
5075: cannot decay into free particles. As we now see,
5076: this property does not yet guarantee their stability,
5077: since there can be other decay channels. Their non-relativistic optical counterparts 
5078:  can decay into several solitonic constituents. 
5079: In order to see if the same could be true
5080: for the relativistic $Q$-balls, one should integrate 
5081: the full time evolutions equation Eq.\eqref{bullet2}. 
5082: The existence of negative modes for spinning $Q$-balls is also suggested 
5083: by the results of Ref.\cite{Axenides:2001pi},\cite{BattyeQ}. 
5084: 
5085:  
5086: 
5087: The results of Ref.\cite{mihalache} have been generalized in 
5088: Refs.\cite{mihalache1},\cite{mihalache2} to 
5089: describe the case of light pulses with two 
5090: independent polarizations and also for 
5091: more general non-linearities of the medium.
5092: 
5093: 
5094: 
5095: 
5096: 
5097: \subsection{Moving vortex rings }
5098: 
5099: For the sake of completeness, we would also like to consider  
5100: another well known type of ring solitons: moving vortex rings in continuous media. 
5101: Although they are stabilized by the 
5102: interaction with the medium and not by intrinsic forces, 
5103: their structure can be quite similar to that for the 
5104: other solitons considered above. 
5105: 
5106: 
5107: A vortex ring moving in a medium encounters the `wind'  that 
5108: produces  a circulation around the vortex core, giving  rise to the 
5109: Magnus force -- the same force that acts on a spinning ping-pong ball in 
5110: the direction orthogonal to its velocity. 
5111: This force is directed outwards orthogonally to the ring velocity, 
5112: stabilizing the ring against shrinking (see Fig.\ref{FigVV}). As a result, 
5113: the ring travels with a constant speed (if there is no dissipation),
5114: dynamically keeping a constant radius, so that in the comoving reference 
5115: frame its configuration is stationary. 
5116: 
5117: \begin{figure}[h]
5118: \hbox to\linewidth{\hss%
5119:   \resizebox{12cm}{4cm}{\includegraphics{ring.eps}}%
5120: %  \resizebox{8cm}{6cm}{\includegraphics{FIG8.eps}}%
5121: \hss}
5122: \caption{\small 
5123: Moving vortex rings are stabilized by an analog $\vec{F}_M$ of the Magnus force
5124: produced by the phase rotation {\it around} the vortex core.
5125: Vortons are stabilized by the centrifugal force
5126: produced by the phase rotation {\it along} the vortex. 
5127: }
5128: \label{FigVV}
5129: \end{figure}
5130: 
5131: The  moving vortex rings are best known in fluid dynamics, as for example 
5132: smoke rings. The vortices themselves correspond in this case to their traditional 
5133: definition -- these are line type objects containing a non-zero 
5134: rotation of the fluid in their core, as for example swirls in water or tornadoes. 
5135: Vortices and 
5136: vortex rings can be created in the wake of rigid bodies travelling  through the fluid. 
5137: Their hydrodynamical theory was created in the 19-th 
5138: century by Helmholtz and Kelvin and has been developing ever since, 
5139: finding numerous practical applications, as for example in the aircraft engineering. 
5140: We do not intend  to discuss
5141: this vast subject here and simply refer to the standard monograph on vortices
5142: \cite{Saffman}. We also give the classical 
5143: formulas of Kelvin relating the vortex ring velocity $V$ and its energy $E$  
5144: and momentum $P$
5145: to the ring radius, $\RRR$, and the vortex core radius, $\ccc$, 
5146: \be                                           \label{Kelvin}
5147: E=\frac12\,\eta^2\Gamma^2\left(\ln\frac{8\RRR}{\ccc}-\frac{7}{4}\right),~~~~
5148: P=\pi\eta\Gamma\RRR^2,~~~~
5149: V=\frac{\Gamma}{4\pi \RRR}\left(\ln\frac{8\RRR}{\ccc}-\frac{1}{4}\right),
5150: \ee
5151: which are valid in the thin ring limit, $\RRR\gg \ccc$.
5152: Here $\eta$ is the fluid mass density
5153: and $\Gamma$ is 
5154: the circulation of the fluid velocity around the vortex.
5155: These formulas show that large rings have large energy and momentum
5156: but move slowly. 
5157:  Very large and slowly moving rings, which could 
5158: perhaps be described by some version of hydrodynamics of magnetized plasma, 
5159: are observed in Sun's atmosphere.  
5160: 
5161: \subsubsection{Moving vortex rings in the superfluid helium} 
5162: 
5163: Moving vortex rings exist also in the 
5164: superfluid helium, where they are created by moving impurities 
5165:  \cite{Rayfield} (see \cite{Volovik} for a survey of defects in helium).
5166: It turns out that large rings in helium
5167: can be  well described by the classical hydrodynamics. However, 
5168: in the generic case 
5169: their adequate description should be quantum. 
5170: The superfluid helium can be approximately modeled by a weakly interacting  
5171: Bose-Einstein condensate (although in reality the interactions 
5172: in helium are not weak, see $e.g.$ \cite{Berloff4}),
5173: in which case vortices can be described by solutions of 
5174: the Gross-Pitaevskii equation \eqref{GP}. 
5175: In the simplest case 
5176: one can consider the one-component version of this equation, 
5177: which, upon setting $V(\bx)=0$, 
5178: $\kappa_{11}=\kappa$ 
5179: and $\Psi_1\equiv e^{-i\frac{\kappa}{2} t}\,\Psi$ reduces to
5180: \be                                    \label{GP1}
5181: 2i\frac{\partial\Psi}{\partial t}=
5182: \left(-\Delta +\kappa(|\Psi|^2-1)\right)\Psi\,.
5183: \ee
5184: Rescaling coordinates and time one can set $\kappa=1$, if it is positive. 
5185: Choosing $\Psi=e^{in\varphi}f(\rho)$ gives
5186: \be                                    \label{GP2}
5187: f^{\prime\prime}+\frac{1}{\rho}\,f^\prime-\frac{n^2}{\rho^2}\,f
5188: =(f^2-1)f\,,
5189: \ee
5190: whose solution with the boundary conditions $f(0)=0$ and $f(\infty)=1$
5191: describes a global vortex with $n$ quanta of circulation
5192: \cite{Pitaevskii}. 
5193: 
5194: Solutions of the GP equation \eqref{GP1} describing moving vortex rings 
5195: and their various properties 
5196: have been comprehensively analyzed by P.H.~Roberts and collaborators in a series of articles 
5197: written over the last 25 years,
5198: of which we only mention 
5199: Refs.\cite{Grant},
5200: \cite{Jones},
5201: \cite{Berloff},
5202: \cite{Berloff1},
5203: \cite{Berloff3}.
5204: The fundamental vortex ring solutions 
5205: have been numerically obtained by Jones and Roberts 
5206: \cite{Jones} by setting 
5207: \be
5208: \Psi=\Psi(\rho,z-Vt),
5209: \ee
5210: such that $\Psi(\rho,z)$ satisfies the equation 
5211: \be                                    \label{GP3}
5212: -i2V\partial_z\Psi=
5213: \left(-\Delta +(|\Psi|^2-1)\right)\Psi\,.
5214: \ee
5215: They imposed the reflection condition $\Psi(\rho,z)=\Psi(\rho,-z)^\ast$
5216: and also the asymptotic conditions for large $r$,
5217: \be
5218: \Re(\Psi)=1+O(r^{-3}),~~~~\Im(\Psi)=O(r^{-2}).
5219: \ee
5220: Numerically integrating Eq.\eqref{GP3}, they obtained
5221: a family of solutions labeled by $V\in(0,V_{\rm max})$,
5222: where $V_{\rm max}=1/\sqrt{2}$ 
5223: is the speed of sound in the dimensionless units chosen. 
5224: For small $V$ the solutions have ring profiles: the function $\Psi$ 
5225: vanishes at a point $(\RRR,0)$ in the $(\rho,z)$ plane while its phase
5226: increases by $2\pi$ after one revolution around this point,
5227: so that $\Psi$ has the `vorton topology'  (see
5228: Fig.\ref{Fig:top}). 
5229: When $V$ is small, the radius of the ring, $\RRR$,
5230: as well as its energy and momentum, 
5231: \begin{align}
5232: E&=\frac12\int\left(|\nabla\Psi|^2+\frac12(|\Psi|^2-1)^2\right)d^3\bx\,,\notag \\
5233: P_3&=\Im\int (\Psi^\ast-1)\partial_z\Psi\, d^3\bx,
5234: \end{align}
5235: are large, and in the leading order  they 
5236: agree with the classical Kelvin  formulas 
5237: \eqref{Kelvin} \cite{Grant}.  
5238: As $V$ increases, $E$, $P_3$ and $\RRR$ all decrease, 
5239: and the following relation holds,
5240: \be
5241: V=\frac{\partial E}{\partial P_3}\,.
5242: \ee
5243: The ring radius 
5244: $\RRR$ vanishes for 
5245: $V=V_0\approx 0.62$ and  
5246: for $V_0<V<V_{\rm max}$ solutions change their character and describe 
5247: acoustic excitations in helium, called in  \cite{Jones}  rarefaction
5248: pulses, in which case $\Psi$ 
5249: no longer has zeroes . 
5250: The energy and momentum  diverge when $V$ approaches $0,V_{\rm max}$, 
5251:  and they both  assume their minimal values for some 
5252: $V=V_{\rm crit}\approx 0.657$.
5253: The functions $E(V)$ and $P_3(V)$ exhibit therefore qualitatively the same 
5254: behaviour as $E(\omega)$ and $Q(\omega)$ for the $Q$-balls (see Fig.\ref{FigQ3}). 
5255: The function $E(P_3)$ shows the two-branch structure with a cusp, 
5256: similarly to $E(Q)$ in Fig.\ref{FigQ3}. The moving rings belong to the 
5257: lower branch, which suggests that they could be stable.
5258: Their perturbative stability has been demonstrated by Berloff and Roberts
5259: \cite{Berloff}, who also studied the 
5260: ring creation by solving 
5261: the GP equation for a superfluid flow around a moving solid sphere \cite{Berloff1}
5262: and showed that rings are created in its wake,
5263: provided that its velocity exceeds some critical value.  
5264: 
5265: The moving ring solutions have been generalized for a non-zero trapping
5266: potential $V(\bx)$ in the GP equation (see \cite{Komineas} for a review)
5267:  and for the two-component
5268: condensate case \cite{Berloff2}. An interesting solution of the 
5269: GP equation describing  a non-linear superposition of a straight vortex and
5270: a vortex ring encircling it and moving along it, 
5271: somewhat similar to the `hoop' solution shown in Fig.\ref{hoop}, 
5272:  has been presented by Berloff in
5273:  \cite{Berloff3}.  
5274: 
5275: One may also wonder whether the vortex ring solutions 
5276: could be generalized within a
5277: relativistic field theory (as is the case for the vortex 
5278: solution of Eq.\eqref{GP2}). 
5279: In fact, the answer to this question is affirmative, since taking
5280:  a solution $\Psi(\rho,z)$ of Eq.\eqref{GP3} and  setting
5281: \be
5282: \phi=e^{iV(z\pm t)}\Psi(\rho,z)
5283: \ee
5284: and also $\sigma=0$ solves
5285: Eqs.\eqref{eW} of Witten's model  
5286: for $\lambda_\phi=2$.
5287:  However, since $\Psi\to 1$ at infinity,
5288: it follows that $\phi\to e^{iV(z\pm t)}$ in this limit, which 
5289: corresponds to a massless
5290: Goldstone radiation.
5291: The relativistic
5292: analogs of the moving vortex rings are therefore 
5293: radiative solutions and their total 
5294: energy is infinite. 
5295: 
5296: 
5297: 
5298: 
5299: 
5300: \subsubsection{Moving magnetic rings \label{magnet}}
5301: 
5302: An interesting example of moving ring solitons, 
5303: found by Cooper \cite{Cooper} and  
5304: studied also by 
5305: Sutcliffe  \cite{Sutcliffe:2007vm}, 
5306: arises in ferromagnetic systems
5307: whose dynamics is described in the continuum 
5308: limit by the Landau-Lifshitz equation
5309: \be                                                               \label{LL1}
5310: \frac{\partial{n^a}}{\partial t}=\epsilon_{abc}n^b\Delta n^c.
5311: \ee  
5312: Here $n^a$ is a unit three-vector, $n^a n^a=1$, giving the local orientation
5313: of the magnetization. The temporal evolution defined by the 
5314: Landau-Lifshitz equation
5315: preserves the value of the energy 
5316: \be
5317: E= \frac{1}{4}\int (\partial_k n^a)^2d^3\bx\,.
5318: \ee
5319: In addition, it does not change the 
5320: momentum,
5321: \be                       \label{P}
5322: P_i=\frac12 \int \epsilon_{ijk}(x_j {\mcal B}_k) d^3\bx,
5323: \ee
5324: where ${\mcal B}_i=\frac{1}{2}\epsilon_{ijk}{\mcal F}_{jk}$ and ${\mcal F}_{jk}$ is defined
5325: as in the Faddeev-Skyrme model  by Eq.\eqref{F}. 
5326: The number of spin flips,
5327: \be
5328: {\mcal N}=\frac12\int(1-n^3)d^3\bx,
5329: \ee
5330: is also preserved by the evolution; 
5331: for this number to be finite one should have $n^3=1$ at infinity.
5332: Since the theory contains a unit vector field,
5333: it can be characterized by the topological index of Hopf defined by the same 
5334: equation \eqref{Q} as in the Faddeev-Skyrme model. 
5335: An  axially symmetric field configuration 
5336: with a non-zero Hopf charge can be parametrized 
5337: exactly as in Eq.\eqref{Hopf:axial} (with $n=1$),
5338: \be                        \label{nn}
5339: n^1+in^2=\sin\Theta(\rho,z) \exp\{i[m\varphi-\psi(\rho,z)]\},~~~~
5340: n^3=\cos\Theta(\rho,z).
5341: \ee  
5342: The energy density then has a ring structure and the 
5343: Hopf charge is 
5344: $\Q=m.$
5345: Using this to calculate
5346: the momentum \eqref{P} shows that its $z$-component
5347: does not vanish,
5348: \be
5349: P_z=\frac14\int \rho(\partial_z \Theta\partial_\rho\psi
5350: -\partial_\rho \Theta\partial_z\psi) d^3\bx\,.
5351: \ee 
5352: This means that if one uses \eqref{nn} as the initial data for the 
5353: Landau-Lifshitz equation, then this data set will have a 
5354: non-zero initial momentum.
5355: To verify this assertion, 
5356: Sutcliffe \cite{Sutcliffe:2007vm} directly integrates the Landau-Lifshitz equation 
5357: to reconstruct the temporal dynamics of the  ring.  
5358: For largely arbitrary functions $\Theta(\rho,z)$, $\psi(\rho,z)$ in \eqref{nn}
5359: he then sees the whole ring configuration drift  with an almost constant 
5360: velocity along the $z$-axis. He also observes 
5361: that the phase of $n^1+in^2$ increases as a linear function of the 
5362: distance travelled by  the ring,
5363: so that there are internal rotations in the ring as it moves. 
5364: 
5365: These empirical features can be understood by making the following ansatz 
5366: for the dynamical fields \cite{Cooper},
5367: \be                        \label{nn1}
5368: n^1+in^2=\sin \Theta(\rho,z-Vt)\exp\{i[m\varphi+\omega t-\psi(\rho,z-Vt)]\},~~~~
5369: n^3=\cos\Theta(\rho,z-Vt),
5370: \ee  
5371: with constant $V,\omega$. Inserting this ansatz to the Landau-Lifshitz equation
5372: \eqref{LL1}, the $t,\varphi$ variables separate and the problem reduces 
5373: to the two dimensional elliptic problem for $\Theta(\rho,z),\psi(\rho,z)$,
5374: \begin{align}                    \label{eq-magn}
5375: \Delta \Theta&=(\omega+V\partial_z\psi)\sin\Theta
5376: +\left[(\partial_\rho\psi)^2+(\partial_z\psi)^2)
5377: +\frac{m^2}{\rho^2}\right] \sin\Theta\cos\Theta, \notag \\
5378: \sin\Theta\Delta\psi&=-V\partial_z \Theta
5379: -2\cos\Theta[\partial_\rho \Theta\partial_\rho\psi+\partial_z \Theta\partial_z\psi].
5380: \end{align}
5381: This problem is in fact very much similar to that in the 
5382: vorton case.  This can be seen by applying the same change
5383: of variables which was used to rewrite the Faddeev--Skyrme model \eqref{action0}
5384: in the $CP^1$ form \eqref{action1}, 
5385: $
5386: n^a=\Phi^\dagger \tau^a\Phi\,,
5387: $
5388: with $\Phi^\dagger\Phi=1$.  This parametrization of the system allows one to  
5389: write down a simple Lagrangian for the  Landau-Lifshitz equation \eqref{LL1},
5390: \be                               \label{LagLL}
5391: {\mcal L}=i\Phi^\dagger\partial_t\Phi
5392: -\frac14(\partial_k n^a)^2
5393: +\mu(\Phi^\dagger\Phi-1),
5394: \ee
5395: where $\mu$ is the Lagrange multiplier. One can parametrize
5396: $\Phi$ exactly in the same way as in the vorton case,
5397: $$
5398: \Phi=
5399: \left(\begin{array}{c}
5400: X+iY\\
5401: Z e^{i\omega t+i m\varphi}\\
5402: \end{array}
5403: \right).
5404: $$
5405: Inserting this  to the Lagrangian \eqref{LagLL} and 
5406: varying with respect to $X,Y,Z,\mu$ gives equations similar to 
5407: Eqs.\eqref{eqsa} for vortons in the sigma model limit. 
5408: These equations can be solved  with the same boundary conditions 
5409: \eqref{1}--\eqref{4}
5410: for  $X,Y,Z$ as for vortons. On the other hand, setting 
5411: $X+iY=\cos(\Theta/2)e^{i\psi}$ and $Z=\sin(\Theta/2)$,
5412: as in Eq.\eqref{anz:3}, gives 
5413: again Eqs.\eqref{eq-magn}. 
5414: 
5415: Fixing $m$  
5416: one can solve the equations for suitable  input values of $\omega,V$. 
5417: Having obtained the solutions, one can calculate
5418:  $E(\omega,V)$, ${\mcal N}(\omega,V)$ and $P_3(\omega,V)$.  
5419: Alternatively, 
5420: instead of solving the equations, one can 
5421: minimize the energy $E$ with fixed 
5422: ${\mcal N},P_3$. Having found the 
5423: minimal value $E({\mcal N},P_3)$, one can reconstruct  $\omega,V$ as 
5424: \cite{Cooper},\cite{Sutcliffe:2007vm}
5425: \be
5426: \omega=\frac{\partial E}{\partial {\mcal N}},~~~~~
5427: V=\frac{\partial E}{\partial P_3}.
5428: \ee
5429: Cooper \cite{Cooper} and
5430:  Sutcliffe \cite{Sutcliffe:2007vm} choose the latter approach:
5431: they minimize the energy with fixed $\ssigma,P_3$ 
5432: for several values of $m=\Q$.
5433: Their numerics converge to non-trivial  
5434: configurations  $\Theta(\rho,z),\psi(\rho,z)$, inserting which 
5435: to \eqref{nn1} 
5436: gives dynamical rings travelling with a constant velocity.  
5437: Fixing all other parameters, the ring velocity decreases with growing $m$. 
5438: 
5439: In fact, these rings also have an angular momentum,
5440: $J=m\ssigma $,
5441: but apparently this feature is  not so essential for their 
5442: stabilization as it is for vortons, 
5443: since solutions with $m=\Q=J=0$ also exist 
5444: and do not exhibit any particular features  
5445: as compared to the solutions with $m\neq 0$ \cite{Sutcliffe:2007vm}.  
5446: Similarly to the other moving ring solitons discussed above,
5447: magnetic rings are stabilized by a Magnus-type force produced 
5448: by the phase circulation 
5449: {\it around} the vortex core and directed outwards, 
5450: orthogonally to the velocity (see Fig.\ref{FigVV}). 
5451: The magnetic rings are  stable not only against shrinking but also
5452: with respect to all other deformations -- since they can be obtained via the 3D energy 
5453: minimization  \cite{Sutcliffe:2007vm}.  
5454: 
5455: 
5456: 
5457: 
5458:  
5459: 
5460: 
5461: \section{Concluding remarks}
5462: 
5463: We have reviewed  the known field theory solutions 
5464: describing stationary vortex loops.
5465: From the physical point of view, loops of magnetic 
5466: flux would perhaps be the most interesting. 
5467: However, as we have seen, almost all explicit solutions  
5468: describe vortex loops  in global field theories,
5469: which are effective theories with a relatively limited range of applicability. 
5470: Perspectives for generalizing these solutions in the context of gauge field theory 
5471: do not look very promising at the moment, at least as far as  knot solitons are concerned.  
5472: The best known static knot solitons, Faddeev-Skyrme knots,
5473: do not seem to admit immediate gauge field theory generalizations, unless 
5474: one makes additional physical assumptions to fix the charges in the 
5475: Protogenov-Verbus formula. Such assumptions can be made, as shows the example
5476: of Schmid and Shaposhnikov. However this requires a rather exotic physical 
5477: environment.  
5478: 
5479: The non-Abelian monopole and sphaleron 
5480: rings are very interesting 
5481: theoretically, however,  they are presumably unstable. 
5482: It is therefore unclear whether they can find important
5483: physical applications. Summarizing, the existence of knots -- magnetic vortex loops 
5484: stabilized against shrinking by internal stresses -- 
5485: does not seem to be very plausible
5486: in the context of physically interesting relativistic gauge field theories. 
5487:   
5488: 
5489: Perhaps gauged vortons have more chances 
5490: to exist. 
5491: As we have seen, their global counterparts  
5492: do exist  as stationary, non-radiating field theory objects.
5493: Already for these global vortons an additional analysis to study the structure 
5494: of the parameter space is needed.  
5495: A natural problem to address would be their analysis  
5496: in the thin ring limit, for  large values of the azimuthal winding number $m$, 
5497: when the ring radius is much larger than the vortex core thickness. 
5498: One can expect that the field theory solutions in this limit should agree 
5499: with the effective macroscopic description. This limit, however, is difficult to 
5500: explore numerically, since the vorton fields are then almost everywhere 
5501: constant, except for  a narrow ring region containing all the field gradients.
5502: It is difficult to properly adjust the grid in this case, so that some  
5503: other numerical methods are necessary, as perhaps 
5504: the multi-domain spectral method \cite{Grandclement}. 
5505: Is seems also that some other methods
5506: are necessary to study solutions with $n>1$. 
5507: 
5508: 
5509: 
5510: Yet another interesting related problem would be to find 
5511: vortons with {\it three} 
5512: winding
5513: numbers. We know that the phase of $\phi$ winds around 
5514: the vortex core, while
5515: that of $\sigma$ winds along the ring, the 
5516: corresponding winding numbers being $n$ and $m$. 
5517: However, nothing forbids the phase of $\sigma$ to wind {\it both} 
5518: around the core and
5519: along the ring, as does the phase of the twisted $Q$-balls, 
5520: in which case one
5521: would need a third integer, $k$, counting the 
5522: windings of $\sigma$ around the 
5523: vortex core. Constructing vortons with $k\neq 0$ remains a challenge.   
5524: 
5525: A very important open problem related to all vortons is their 
5526: dynamical stability analysis. This problem can be handled either by 
5527: studying the spectrum of small field  fluctuations around the vorton background,
5528: or by looking for vorton negative modes within the energy minimization method,
5529: or by reconstructing the full temporal evolution of the perturbed
5530: vorton configuration. Stable vortons definitely exist, at least in the 
5531: sigma model limit, but in the 
5532: generic case their stability should be studied. 
5533: 
5534: 
5535: 
5536: Generalizing global vortons within gauge field theory does not seem
5537: to be impossible. In theories where all the internal symmetries are gauged,
5538: as in Witten's model \eqref{W}, all spinning phases of the fields can 
5539: be  gauged away, so that if the fields are stationary and axisymmetric, these  
5540: symmetries are manifest.
5541: In this case the surface integral formula \eqref{JJJ} for the 
5542: angular momentum applies, providing  higher chances for
5543: $J$ to vanish due to 
5544: the fast asymptotic falloff of the fields. Although   
5545: the example of spinning gauged $Q$-balls shows that spinning is possible
5546: in local theories, 
5547: it seems that vortons have better chances to exist in models with both local
5548: and global internal symmetries. 
5549: Such models contain gauge fields, such that one can have magnetic vortices, 
5550: but the gauge freedom is not enough to gauge away  all the phases of the scalars. 
5551: The symmetries of the fields are therefore not manifest and the surface integral 
5552: formula does not apply, in which case the 
5553: angular momentum should normally be non-zero. 
5554: 
5555: For example, it would be natural to analyze
5556:  the existence of stationary vortons in a half gauged
5557: version of Witten's modes obtained by coupling the two scalars in Eq.\eqref{lag}
5558: to a U(1) gauge field. 
5559: Such solutions could then perhaps be generalized within the
5560: fully gauged U(1)$\times$U(1) Witten's modes to confirm  the existence of 
5561: stationary loops made of superconducting cosmic strings  -- the idea first 
5562: put forward more than 20 years ago. 
5563: 
5564: However, a real challenge would be to construct stationary superconducting loops
5565: in Standard Model. There are  indications that this might be possible.
5566: The electroweak sector of Standard Model contains stationary 
5567: current carrying vortices -- superconducting strings \cite{Volkov:2006ug}. 
5568: Their stability 
5569: has been studied so far only in the semilocal limit, where the SU(2) gauge field 
5570: decouples and the current becomes global \cite{Forgacs:2006pm},\cite{Forgacs:2005sf}, 
5571: and it has been found that sufficiently short pieces of strings 
5572: are perturbatively stable \cite{Garaud:2007ti}.
5573: This suggests that small loops 
5574: made of these strings may also be stable. If such loops could be 
5575: constructed for the physical value of the weak mixing angle, this would 
5576: give stable solitons in Standard Model. 
5577: 
5578: For the sake of completeness, having in mind possible applications of vortex loops
5579: in astrophysics and cosmology, it is also interesting to  mention the effects
5580: of gravity on these solutions. On general grounds, one expects all known ring 
5581: solitons to admit self-gravitating generalizations with essentially the same properties
5582: as in the zero gravity limit,  
5583: at least for small enough values of the gravitational coupling constant. However,
5584: the soliton structure can change considerably
5585: in the strong gravity regime \cite{VG}.
5586: The solitons can then become gravitationally closed and may also contain 
5587: a small black hole in their center. The spinning  of the latter can 
5588: endow the whole configuration with an angular momentum, 
5589: even if the soliton 
5590: itself cannot spin -- as for example the monopole \cite{monBH}. 
5591: However, if the soliton can spin on its own,
5592: as for example a vorton, it would be  interesting to put a spinning black hole in its center. 
5593: Since there are no asymptotically flat toroidal black holes in four dimensions 
5594: \cite{Friedman}, the resulting configuration is expected to be a black hole with spherical 
5595: horizon topology, surrounded by a vortex ring.  
5596:  
5597: 
5598: \vspace{1 cm}
5599: {\bf Acknowledgements.}
5600: The work of E.R. was supported by the ANR grant NT05-$1_{-}$42856 
5601:  `Knots and Vortons'. We would like to thank Richard Battye, Brandon Carter, 
5602: Maxim Chernodub, Ludwig Faddeev, Peter Forgacs, Philippe Grandclement, Betti Hartmann, 
5603: Xavier Martin, Antti Niemi, 
5604: Mikhail Shaposhnikov,
5605: Matthias Schmid, Paul Shellard, Paul Sutcliffe 
5606: and Tigran Tchrakian  
5607:  for discussions during various stages of this work. 
5608: %M.S.V. would also like to thank the organizers of the Peyresq   
5609: %workshop where an early version of this work was presented.  
5610: 
5611:  
5612: 
5613: 
5614: %\bibliography{v}
5615: \begin{thebibliography}{100}
5616: 
5617: \bibitem{Abrikosov}
5618: A.A. Abrikosov.
5619: \newblock {\ssl On the magnetic properties of superconductors of the second
5620:   group}.
5621: \newblock {\em {\XPEH Sov.Phys.JETP}}, {\bbf 5}, 1174--1182, 1957.
5622: 
5623: \bibitem{Achucarro:1999it}
5624: A.~Achucarro and T.~Vachaspati.
5625: \newblock {\ssl Semilocal and electroweak strings}.
5626: \newblock {\em {\XPEH Phys.Rept.}}, {\bbf 327}, 347--426, 2000.
5627: 
5628: \bibitem{Adam:2006wg}
5629: C.~Adam, J.~Sanchez-Guillen, R.A. Vazquez, and A.~Wereszczynski.
5630: \newblock {\ssl Investigation of the Nicole model}.
5631: \newblock {\em {\XPEH J.Math.Phys.}}, {\bbf 47}, 052302, 2006.
5632: 
5633: \bibitem{Adkins:1983hy}
5634: G.~Adkins and C.R. Nappi.
5635: \newblock {\ssl The Skyrme model with pion masses}.
5636: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B233}, 109--115, 1984.
5637: 
5638: \bibitem{Adkins:1983ya}
5639: G.~Adkins, C.R. Nappi, and W.~Witten.
5640: \newblock {\ssl Static properties of nucleons in the Skyrme model}.
5641: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B228}, 552--566, 1983.
5642: 
5643: \bibitem{akhmediev}
5644: N.N. Akhmediev and A.~Ankiewicz.
5645: \newblock {\em {\ssl Solitons, nonlinear pulses and beams}}.
5646: \newblock {\XPEH Chapman and Hall, London}, 1997.
5647: \newblock 299 p.
5648: 
5649: \bibitem{Amsterdamski:1988zp}
5650: P.~Amsterdamski and P.~Laguna-Castillo.
5651: \newblock {\ssl Internal structure and the spacetime of superconducting bosonic
5652:   strings}.
5653: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D37}, 877--884, 1988.
5654: 
5655: \bibitem{Aratyn:1999cf}
5656: H.~Aratyn, L.A. Ferreira, and A.H. Zimerman.
5657: \newblock {\ssl Exact static soliton solutions of 3+1 dimensional integrable
5658:   theory with nonzero Hopf numbers}.
5659: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 83}, 1723--1726, 1999.
5660: 
5661: \bibitem{Aratyn}
5662: H.~Aratyn, L.A. Ferreira, and A.H. Zimerman.
5663: \newblock {\ssl Toroidal solitons in 3+1 dimensional integrable theories}.
5664: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B456}, 162--170, 1999.
5665: 
5666: \bibitem{Axenides:2001pi}
5667: M.~Axenides, E.~Floratos, S.~Komineas, and L.~Perivolaropoulos.
5668: \newblock {\ssl Q-rings}.
5669: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 86}, 4459--4462, 2001.
5670: 
5671: \bibitem{Babaev:2001jt}
5672: E.~Babaev.
5673: \newblock {\ssl Knotted solitons in triplet superconductors}.
5674: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 88}, 177002, 2002.
5675: 
5676: \bibitem{Babaev:2001zy}
5677: E.~Babaev, L.D. Faddeev, and A.J. Niemi.
5678: \newblock {\ssl Hidden symmetry and duality in a charged two-condensate Bose
5679:   system}.
5680: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf B65}, 100512, 2002.
5681: 
5682: \bibitem{Babul:1987me}
5683: A.~Babul, T.~Piran, and D.M. Spergel.
5684: \newblock {\ssl Bosonic superconducting cosmic strings. I. Classical field
5685:   theory solutions}.
5686: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B202}, 307--314, 1988.
5687: 
5688: \bibitem{Battye:1998pe}
5689: R.~Battye and P.~Sutcliffe.
5690: \newblock {\ssl Knots as stable soliton solutions in a three-dimensional
5691:   classical field theory}.
5692: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 81}, 4798--4801, 1998.
5693: 
5694: \bibitem{Battye:1998zn}
5695: R.~Battye and P.~Sutcliffe.
5696: \newblock {\ssl Solitons, links and knots}.
5697: \newblock {\em {\XPEH Proc.Roy.Soc.Lond.}}, {\bbf A455}, 4305--4331, 1999.
5698: 
5699: \bibitem{BattyeQ}
5700: R.~Battye and P.~Sutcliffe.
5701: \newblock {\ssl $Q$-ball dynamics}.
5702: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B590}, 329--363, 2000.
5703: 
5704: \bibitem{Battye:2001ec}
5705: R.A. Battye, N.R. Cooper, and P.M. Sutcliffe.
5706: \newblock {\ssl Stable Skyrmions in two-component Bose-Einstein condensates}.
5707: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 88}, 080401, 2002.
5708: 
5709: \bibitem{Battye:2005nx}
5710: R.A. Battye, S.~Krusch, and P.M. Sutcliffe.
5711: \newblock {\ssl Spinning Skyrmions and the Skyrme parameters}.
5712: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B626}, 120--126, 2005.
5713: 
5714: \bibitem{Battye-kink}
5715: R.A. Battye and P.~Sutcliffe.
5716: \newblock {\ssl Kinky vortons}, 2008.
5717: \newblock {\tt e-Print: arXiv:0806.2212 [hep-th]}.
5718: 
5719: \bibitem{Berloff2}
5720: N.G. Berloff.
5721: \newblock {\ssl Nucleation of solitary wave complexes in two-component mixture
5722:   Bose-Einstein condensates}, 2005.
5723: \newblock arXiv:cond-mat/0412743.
5724: 
5725: \bibitem{Berloff3}
5726: N.G. Berloff.
5727: \newblock {\ssl Solitary waves on vortex lines in Ginzburg-Landau models for
5728:   the example of Bose-Einstein condensates}.
5729: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 94}, 010403, 2005.
5730: 
5731: \bibitem{Berloff1}
5732: N.G. Berloff and P.H. Roberts.
5733: \newblock {\ssl Motion in a Bose condensate: VII. Boundary-layer separation}.
5734: \newblock {\em {\XPEH J.Phys.}}, {\bbf A33}, 4025--4038, 2000.
5735: 
5736: \bibitem{Berloff4}
5737: N.G. Berloff and P.H. Roberts.
5738: \newblock {\ssl Nonlinear Schr\"odinger equation as a model of superfluid
5739:   helium}.
5740: \newblock In {\em {\XPEH Quantized Vortex Dynamics and Superfluid Turbulence}},
5741:   volume~{\bbf 571} of {\em Lecture Notes in Physics}. Springer-Verlag, 2001.
5742: \newblock Edited by C.F. Barenghi, R.J. Donnelly and W.F. Vinen.
5743: 
5744: \bibitem{Berloff}
5745: N.G. Berloff and P.H. Roberts.
5746: \newblock {\ssl Motion in a Bose condensate: X. New results on stability of
5747:   axisymmetric solitary waves of the Gross-Pitaevskii equation}.
5748: \newblock {\em {\XPEH J.Phys.}}, {\bbf A37}, 11333--11351, 2004.
5749: 
5750: \bibitem{Betz}
5751: M.~Betz, H.B. Rodrigues, and T.~Kodama.
5752: \newblock {\ssl Rotating skyrmion in (2+1)-dimensions}.
5753: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D54}, 1010--1019, 1996.
5754: 
5755: \bibitem{Bevis:2006mj}
5756: N.~Bevis, M.~Hindmarsh, M.~Kunz, and J.~Urrestilla.
5757: \newblock {\ssl CMB power spectrum contribution from cosmic strings using
5758:   field-evolution simulations of the Abelian Higgs model}.
5759: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D75}, 065015, 2007.
5760: 
5761: \bibitem{Bogomolnyi}
5762: E.B. Bogomol'nyi.
5763: \newblock {\ssl Stability of classical solutions}.
5764: \newblock {\em {\XPEH Sov.J.Nucl.Phys.}}, {\bbf 24}, 449--454, 1976.
5765: 
5766: \bibitem{plasma}
5767: A.H. Boozer.
5768: \newblock {\ssl Physics of magnetically confined plasmas}.
5769: \newblock {\em {\XPEH Rev.Mod.Phys.}}, {\bbf 76}, 1071--1141, 2004.
5770: 
5771: \bibitem{Brandenberger:1996zp}
5772: R.H. Brandenberger, B.~Carter, A.C. Davis, and M.~Trodden.
5773: \newblock {\ssl Cosmic vortons and particle physics constraints}.
5774: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D54}, 6059--6071, 1996.
5775: 
5776: \bibitem{Brihaye:2007tn}
5777: Y.~Brihaye and B.~Hartmann.
5778: \newblock {\ssl Interacting Q-balls}, 2007.
5779: \newblock {\tt arXiv:0711.1969 [hep-th]}.
5780: 
5781: \bibitem{Callan:1983nx}
5782: C.G. Callan and E.~Witten.
5783: \newblock {\ssl Monopole catalysis of skyrmion decay}.
5784: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B239}, 161--176, 1984.
5785: 
5786: \bibitem{Carter:1989dp}
5787: B.~Carter.
5788: \newblock {\ssl Duality relation between charged elastic strings and
5789:   superconducting cosmic strings}.
5790: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B224}, 61--66, 1989.
5791: 
5792: \bibitem{Carter:1989xk}
5793: B.~Carter.
5794: \newblock {\ssl Stability and characteristic propagation speeds in
5795:   superconducting cosmic and other string models}.
5796: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B228}, 466--470, 1989.
5797: 
5798: \bibitem{Carter:1990sm}
5799: B.~Carter.
5800: \newblock {\ssl Mechanics of cosmic rings}.
5801: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B238}, 166--171, 1990.
5802: 
5803: \bibitem{Carter:1994hn}
5804: B.~Carter and P.~Peter.
5805: \newblock {\ssl Supersonic string model for Witten vortices}.
5806: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D52}, 1744--1748, 1995.
5807: 
5808: \bibitem{Cho:2001gc}
5809: Y.M. Cho.
5810: \newblock {\ssl Knot solitons in Weinberg-Salam model}, 2001.
5811: \newblock arXiv:hep-th/0110076.
5812: 
5813: \bibitem{Cho}
5814: Y.M. Cho.
5815: \newblock {\ssl Monopoles and knots in Skyrme theory}.
5816: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 87}, 252001, 2001.
5817: 
5818: \bibitem{Coleman:1982cx}
5819: S.R. Coleman.
5820: \newblock {\ssl The magnetic monopole fifty years later}.
5821: \newblock Lectures given at Int. Sch. of Subnuclear Phys., Erice, Italy, Jul
5822:   31-Aug 11, 1981, at 6th Brazilian Symp. on Theor. Phys., Jan 7-18, 1980, at
5823:   Summer School in Theoretical Physics, Les Houches, France, and at Banff
5824:   Summer Inst. on Particles $\&$ Fields, Aug 16-28, 1981.
5825: 
5826: \bibitem{Coleman:1985ki}
5827: S.R. Coleman.
5828: \newblock {\ssl Q balls}.
5829: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B262}, 263--283, 1985.
5830: 
5831: \bibitem{Coleman:1976uk}
5832: S.R. Coleman, S.~Parke, A.~Neveu, and C.M. Sommerfield.
5833: \newblock {\ssl Can one dent a dyon?}
5834: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D15}, 544--545, 1977.
5835: 
5836: \bibitem{Cooper}
5837: N.R. Cooper.
5838: \newblock {\ssl Propagating magnetic vortex rings in ferromagnets}.
5839: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 82}, 1554--1557, 2001.
5840: 
5841: \bibitem{Copeland:1987yv}
5842: E.J. Copeland, D.~Haws, M.~Hindmarsh, and N.~Turok.
5843: \newblock {\ssl Dynamics of and radiation from superconducting strings and
5844:   springs}.
5845: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B306}, 908--930, 1988.
5846: 
5847: \bibitem{Copeland:1987th}
5848: E.J. Copeland, M.~Hindmarsh, and N.~Turok.
5849: \newblock {\ssl Dynamics of superconducting cosmic strings}.
5850: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 58}, 1910--1913, 1987.
5851: 
5852: \bibitem{Correia}
5853: F.~Paccetti Correia and M.G. Schmidt.
5854: \newblock {\ssl Q balls: some analytical results}.
5855: \newblock {\em {\XPEH Eur.Phys.Journ.}}, {\bbf C21}, 181--191, 2001.
5856: 
5857: \bibitem{Davis:1988ip}
5858: R.L. Davis.
5859: \newblock {\ssl Semitopological solitons}.
5860: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D38}, 3722--3730, 1988.
5861: 
5862: \bibitem{Davis:1988jp}
5863: R.L. Davis and E.P.S. Shellard.
5864: \newblock {\ssl The physics of vortex superconductivity}.
5865: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B207}, 404--410, 1988.
5866: 
5867: \bibitem{Davis:1988jq}
5868: R.L. Davis and E.P.S. Shellard.
5869: \newblock {\ssl The physics of vortex superconductivity. 2}.
5870: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B209}, 485--490, 1988.
5871: 
5872: \bibitem{Davis:1988ij}
5873: R.L. Davis and E.P.S. Shellard.
5874: \newblock {\ssl Cosmic vortons}.
5875: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B323}, 209--224, 1989.
5876: 
5877: \bibitem{deVega:1977rk}
5878: H.J. de~Vega.
5879: \newblock {\ssl Closed vortices and the Hopf index in classical field theory}.
5880: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D18}, 2945--2951, 1978.
5881: 
5882: \bibitem{VanderBij:2001nm}
5883: J.J.~Van der Bij and E.~Radu.
5884: \newblock {\ssl On rotating regular nonabelian solutions}.
5885: \newblock {\em {\XPEH Int.J.Mod.Phys.}}, {\bbf A17}, 1477--1490, 2002.
5886: 
5887: \bibitem{vanderBij:2002sq}
5888: J.J.~Van der Bij and E.~Radu.
5889: \newblock {\ssl Magnetic charge, angular momentum and negative cosmological
5890:   constant}.
5891: \newblock {\em {\XPEH Int.J.Mod.Phys.}}, {\bbf A18}, 2379--2393, 2003.
5892: 
5893: \bibitem{Derrick:1964ww}
5894: G.H. Derrick.
5895: \newblock {\ssl Comments on nonlinear wave equations as models for elementary
5896:   particles}.
5897: \newblock {\em {\XPEH J.Math.Phys.}}, {\bbf 5}, 1252--1254, 1964.
5898: 
5899: \bibitem{Wipf}
5900: L.~Dittmann, T.~Heinzl, and A.~Wipf.
5901: \newblock {\ssl A lattice study of the Faddeev-Niemi action}.
5902: \newblock {\em {\XPEH Nucl.Phys.Proc.Suppl.}}, {\bbf B106}, 649--651, 2002.
5903: 
5904: \bibitem{Donnelly}
5905: R.J. Donnelly.
5906: \newblock {\em {\ssl Quantized vortices in helium II}}.
5907: \newblock {\XPEH Cambridge University Press}, 1991.
5908: \newblock 346 p.
5909: 
5910: \bibitem{Doudoulakis:2007ti}
5911: C.G. Doudoulakis.
5912: \newblock {\ssl On vortices and rings in extended Abelian models}.
5913: \newblock {\em {\XPEH Physica}}, {\bbf D234}, 1--10, 2007.
5914: 
5915: \bibitem{Doudoulakis:2007xz}
5916: C.G. Doudoulakis.
5917: \newblock {\ssl On vortices and solitons in Goldstone and Abelian-Higgs
5918:   models}, 2007.
5919: \newblock {\tt arXiv:0709.3709 [hep-ph]}.
5920: 
5921: \bibitem{Doudoulakis:2006iw}
5922: C.G. Doudoulakis.
5923: \newblock {\ssl Search of axially symmetric solitons}.
5924: \newblock {\em {\XPEH Physica}}, {\bbf D228}, 159--165, 2007.
5925: 
5926: \bibitem{Emparan}
5927: R.~Emparan and H.S. Reall.
5928: \newblock {\ssl Black rings}.
5929: \newblock {\em {\XPEH Class.Quant.Grav.}}, {\bbf 23}, R169, 2006.
5930: 
5931: \bibitem{movies}
5932: J.~Hietarinta et~al.
5933: \newblock {\tt http://users.utu.fi/hietarin/knots/}.
5934: 
5935: \bibitem{Yao}
5936: W.M.~Yao {\it et al.}
5937: \newblock {\ssl The review of particle physics}.
5938: \newblock {\em {\XPEH Journ.Phys.}}, {\bbf G33}, 1--1232, 2006.
5939: 
5940: \bibitem{Faddeev:1975}
5941: L.D. Faddeev.
5942: \newblock {\ssl Quantization of solitons}, 1975.
5943: \newblock Princeton preprint IAS-75-QS70; also {\ssl Einstein and several
5944:   contemporary tendencies in the theory of elementary particles} in {
5945:   Relativity, quanta, and cosmology, Vol.1}, M. Pantaleo and F. De Finis, eds.,
5946:   pp. 247-266 (1979), reprinted in {\it L. Faddeev, 40 years in mathematical
5947:   physics}, pp. 441-461 (World Scientific, 1995).
5948: 
5949: \bibitem{Faddeev:1976pg}
5950: L.D. Faddeev.
5951: \newblock {\ssl Some comments on the many-dimensional solitons}.
5952: \newblock {\em {\XPEH Lett. Math. Phys.}}, {\bbf 1}, 289--293, 1976.
5953: 
5954: \bibitem{Faddeev:2000qw}
5955: L.D. Faddeev, L.~Freyhult, A.J. Niemi, and P.~Rajan.
5956: \newblock {\ssl Shafranov's virial theorem and magnetic plasma confinement}.
5957: \newblock {\em {\XPEH J.Phys.}}, {\bbf A35}, L133--L140, 2002.
5958: 
5959: \bibitem{Faddeev:1996zj}
5960: L.D. Faddeev and A.J. Niemi.
5961: \newblock {\ssl Knots and particles}.
5962: \newblock {\em {\XPEH Nature}}, {\bbf 387}, 58--61, 1997.
5963: 
5964: \bibitem{Faddeev:1997pf}
5965: L.D. Faddeev and A.J. Niemi.
5966: \newblock {\ssl Toroidal configurations as stable solitons}, 1997.
5967: \newblock {\tt hep-th/9705176}.
5968: 
5969: \bibitem{Faddeev:1998eq}
5970: L.D. Faddeev and A.J. Niemi.
5971: \newblock {\ssl Partially dual variables in SU(2) Yang-Mills theory}.
5972: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 82}, 1624--1627, 1999.
5973: 
5974: \bibitem{Faddeev:2000rp}
5975: L.D. Faddeev and A.J. Niemi.
5976: \newblock {\ssl Magnetic Geometry and the Confinement of Electrically
5977:   Conducting Plasmas}.
5978: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 85}, 3416--3419, 2000.
5979: 
5980: \bibitem{Faddeev:2006sw}
5981: L.D. Faddeev and A.J. Niemi.
5982: \newblock {\ssl Spin-charge separation, conformal covariance and the SU(2)
5983:   Yang-Mills theory}.
5984: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B776}, 38--65, 2007.
5985: 
5986: \bibitem{Faddeev:2003aw}
5987: L.D. Faddeev, A.J. Niemi, and U.~Wiedner.
5988: \newblock {\ssl Glueballs, closed fluxtubes and eta(1440)}.
5989: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D70}, 114033, 2004.
5990: 
5991: \bibitem{Fayzullaev:2004xa}
5992: B.A. Fayzullaev, M.M. Musakhanov, D.G. Pak, and M.~Siddikov.
5993: \newblock {\ssl Knot soliton in Weinberg-Salam model}.
5994: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B609}, 442--448, 2005.
5995: 
5996: \bibitem{Forgacs:1979zs}
5997: P.~Forgacs and N.~Manton.
5998: \newblock {\ssl Space-time symmetries in gauge theories}.
5999: \newblock {\em {\XPEH Commun.Math.Phys.}}, {\bbf 72}, 15--35, 1980.
6000: 
6001: \bibitem{Forgacs:2005sf}
6002: P.~Forgacs, S.~Reuillon, and M.S. Volkov.
6003: \newblock {\ssl Superconducting vortices in semilocal models}.
6004: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 96}, 041601, 2006.
6005: 
6006: \bibitem{Forgacs:2006pm}
6007: P.~Forgacs, S.~Reuillon, and M.S. Volkov.
6008: \newblock {\ssl Twisted superconducting semilocal strings}.
6009: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B751}, 390--418, 2006.
6010: 
6011: \bibitem{Forgacs}
6012: P.~Forgacs and M.S. Volkov.
6013: \newblock {\ssl On the existence of knot solitons in gauge field theory}, 2002.
6014: \newblock unpublished.
6015: 
6016: \bibitem{Friedberg}
6017: R.~Friedberg, T.D. Lee, and A.~Sirlin.
6018: \newblock {\ssl A class of scalar-field soliton solutions in three space
6019:   dimensions}.
6020: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D13}, 2739--2761, 1976.
6021: 
6022: \bibitem{Garaud:2007ti}
6023: J.~Garaud and M.S. Volkov.
6024: \newblock {\ssl Stability analysis of the twisted superconducting electroweak
6025:   strings}.
6026: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B799}, 430--455, 2008.
6027: 
6028: \bibitem{Ginzburg}
6029: V.L. Ginzburg and L.D. Landau.
6030: \newblock {\ssl On the theory of superconductivity}.
6031: \newblock {\em {\XPEH Zh.Eksp.Teor.Fiz.}}, {\bbf 20}, 1064, 1950.
6032: 
6033: \bibitem{Pitaevskii}
6034: V.L. Ginzburg and L.P. Pitaevskii.
6035: \newblock {\ssl On the theory of superfluidity}.
6036: \newblock {\em {\XPEH Sov.Phys.JETP}}, {\bbf 34}, 858--861, 1958.
6037: 
6038: \bibitem{Gladikowski:1996mb}
6039: J.~Gladikowski and M.~Hellmund.
6040: \newblock {\ssl Static solitons with non-zero Hopf number}.
6041: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D56}, 5194--5199, 1997.
6042: 
6043: \bibitem{Grandclement}
6044: P.~Grandclement and J.~Novak.
6045: \newblock {\ssl Spectral methods for numerical relativity}, 2007.
6046: \newblock arXiv:0706.2286 [gr-qc].
6047: 
6048: \bibitem{Grant}
6049: J.~Grant and P.H. Roberts.
6050: \newblock {\ssl Motion in a Bose condensate III. The structure and effective
6051:   masses of charged and uncharged impurities}.
6052: \newblock {\em {\XPEH J.Phys.}}, {\bbf A7}, 260--279, 1974.
6053: 
6054: \bibitem{Ball}
6055: B.~Greenwood.
6056: \newblock {\ssl Ball lightning bibliography.} 
6057: \newblock {\tt www.project1947.com/shg/${\tt bl_{-}db.html}$}.
6058: 
6059: \bibitem{Gross}
6060: E.P. Gross.
6061: \newblock {\ssl Structure of quantized vortex}.
6062: \newblock {\em {\XPEH Nuovo Cimento}}, {\bbf 20}, 454--477, 1961.
6063: 
6064: \bibitem{Hartmann:2008yr}
6065: B.~Hartmann and B.~Carter.
6066: \newblock {\ssl The logarithmic equation of state for superconducting cosmic
6067:   strings}, 2008.
6068: \newblock {\tt arXiv:0803.0266 [hep-th]}.
6069: 
6070: \bibitem{Hartmann:2000ja}
6071: B.~Hartmann, B.~Kleihaus, and J.~Kunz.
6072: \newblock {\ssl Dyons with axial symmetry}.
6073: \newblock {\em {\XPEH Mod.Phys.Lett.}}, {\bbf A15}, 1003--1012, 2000.
6074: 
6075: \bibitem{Haws:1988ax}
6076: D.~Haws, M.~Hindmarsh, and N.~Turok.
6077: \newblock {\ssl Superconducting strings or springs ?}
6078: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B209}, 255--261, 1988.
6079: 
6080: \bibitem{Hen}
6081: I.~Hen and M.~Karliner.
6082: \newblock {\ssl Spontaneous breaking of rotational symmetry in rotating
6083:   solitons: A toy model of excited nucleons with high angular momentum}.
6084: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D77}, 116002, 2008.
6085: 
6086: \bibitem{Heusler:1998ec}
6087: M.~Heusler, N.~Straumann, and M.~Volkov.
6088: \newblock {\ssl On rotational excitations and axial deformations of BPS
6089:   monopoles and Julia-Zee dyons}.
6090: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D58}, 105021, 1998.
6091: 
6092: \bibitem{Hietarinta:1998kt}
6093: J.~Hietarinta and P.~Salo.
6094: \newblock {\ssl Faddeev-Hopf knots: Dynamics of linked un-knots}.
6095: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B451}, 60--67, 1999.
6096: 
6097: \bibitem{Hietarinta:2000ci}
6098: J.~Hietarinta and P.~Salo.
6099: \newblock {\ssl Ground state in the Faddeev-Skyrme model}.
6100: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D62}, 081701, 2000.
6101: 
6102: \bibitem{Hill:1987qx}
6103: C.T Hill, H.M. Hodges, and M.S. Turner.
6104: \newblock {\ssl Bosonic superconducting cosmic strings}.
6105: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D37}, 263--282, 1988.
6106: 
6107: \bibitem{Hobart}
6108: R.H. Hobart.
6109: \newblock {\ssl On the instability of a class of unitary field model}.
6110: \newblock {\em {\XPEH Proc.Phys.Soc.}}, {\bbf 82}, 201--203, 1963.
6111: 
6112: \bibitem{Huang:1980bz}
6113: K.~Huang and R.~Tipton.
6114: \newblock {\ssl Vortex excitations in the Weinberg-Salam theory}.
6115: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D23}, 3050--3057, 1981.
6116: 
6117: \bibitem{Ioannidou:2006nn}
6118: T.~Ioannidou, B.~Kleihaus, and J.~Kunz.
6119: \newblock {\ssl Spinning gravitating skyrmions}.
6120: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B643}, 213--220, 2006.
6121: 
6122: \bibitem{Jaykka:2006gf}
6123: J.~Jaykka, J.~Hietarinta, and P.~Salo.
6124: \newblock {\ssl Investigation of the stability of Hopfions in the two-component
6125:   Ginzburg-Landau model}.
6126: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf B77}, 094509, 2008.
6127: 
6128: \bibitem{Friedman}
6129: J.L.Friedman, K.~Schliech, and D.M. Witt.
6130: \newblock {\ssl Topological censorship}.
6131: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 71}, 1486--1489, 1993.
6132: 
6133: \bibitem{Jones}
6134: C.A. Jones and P.H. Roberts.
6135: \newblock {\ssl Motion in a Bose condensate: IV. Axisymmetric solitary waves}.
6136: \newblock {\em {\XPEH J.Phys.}}, {\bbf A15}, 2599--2619, 1982.
6137: 
6138: \bibitem{Julia:1975ff}
6139: B.~Julia and A.~Zee.
6140: \newblock {\ssl Poles with both magnetic and electric charges in nonabelian
6141:   gauge theory}.
6142: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D11}, 2227--2232, 1975.
6143: 
6144: \bibitem{Ueda}
6145: Y.~Kawaguchi, M.~Nitta, and M.~Ueda.
6146: \newblock {\ssl Knots in a spinor Bose-Einstein condensate}, 2008.
6147: \newblock {\tt arXiv:0802.1968 [cond-mat.other]}.
6148: 
6149: \bibitem{Kleihaus:1999sx}
6150: B.~Kleihaus and J.~Kunz.
6151: \newblock {\ssl A monopole antimonopole solution of the SU(2) Yang-Mills- Higgs
6152:   model}.
6153: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D61}, 025003, 2000.
6154: 
6155: \bibitem{Kleihaus:2008gn}
6156: B.~Kleihaus, J.~Kunz, and M.~Leissner.
6157: \newblock {\ssl Sphalerons, antisphalerons and vortex rings}, 2008.
6158: \newblock arXiv:0802.3275 [hep-th].
6159: 
6160: \bibitem{Kleihaus:2005me}
6161: B.~Kleihaus, J.~Kunz, and M.~List.
6162: \newblock {\ssl Rotating boson stars and Q-balls}.
6163: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D72}, 064002, 2005.
6164: 
6165: \bibitem{Kleihaus:2007vk}
6166: B.~Kleihaus, J.~Kunz, M.~List, and I.~Schaffer.
6167: \newblock {\ssl Rotating boson stars and Q-balls II: negative parity and
6168:   ergoregions}, 2007.
6169: \newblock {\tt arXiv:0712.3742 [gr-qc]}.
6170: 
6171: \bibitem{monBH}
6172: B.~Kleihaus, J.~Kunz, and F.~Navarro-Lerida.
6173: \newblock {\ssl Rotating black holes with monopole hair}.
6174: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B599}, 294--300, 2004.
6175: 
6176: \bibitem{Kleihaus:2005fs}
6177: B.~Kleihaus, J.~Kunz, and U.~Neemann.
6178: \newblock {\ssl Gravitating stationary dyons and rotating vortex rings}.
6179: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B623}, 171--178, 2005.
6180: 
6181: \bibitem{Kleihaus:2003xz}
6182: B.~Kleihaus, J.~Kunz, and Y.~Shnir.
6183: \newblock {\ssl Monopoles, antimonopoles and vortex rings}.
6184: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D68}, 101701, 2003.
6185: 
6186: \bibitem{Kleihaus:2004is}
6187: B.~Kleihaus, J.~Kunz, and Y.~Shnir.
6188: \newblock {\ssl Monopole-antimonopole chains and vortex rings}.
6189: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D70}, 065010, 2004.
6190: 
6191: \bibitem{Klinkhamer:1984di}
6192: F.R. Klinkhamer and N.S. Manton.
6193: \newblock {\ssl A saddle point solution in the Weinberg-Salam theory}.
6194: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D30}, 2212--2220, 1984.
6195: 
6196: \bibitem{Koma:1999sm}
6197: Y.~Koma, H.~Suganuma, and H.~Toki.
6198: \newblock {\ssl Flux-tube ring and glueball properties in the dual
6199:   Ginzburg-Landau theory}.
6200: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D60}, 074024, 1999.
6201: 
6202: \bibitem{Komineas}
6203: S.~Komineas.
6204: \newblock {\ssl Vortex rings and solitary waves in trapped Bose-Einstein
6205:   condensates}.
6206: \newblock {\em {\XPEH Eur.Phys.J. Special Topics}}, {\bbf 147}, 133--152, 2007.
6207: 
6208: \bibitem{Kundu}
6209: A.~Kundu and Yu.P. Rybakov.
6210: \newblock {\ssl Closed-vortex-type solitons with Hopf index}.
6211: \newblock {\em {\XPEH J.Phys.}}, {\bbf A15}, 269--275, 1982.
6212: 
6213: \bibitem{Kusenko}
6214: A.~Kusenko.
6215: \newblock {\ssl Small Q-balls}.
6216: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B404}, 285--290, 1997.
6217: 
6218: \bibitem{Kusenko:1997zq}
6219: A.~Kusenko.
6220: \newblock {\ssl Solitons in the supersymmetric extensions of the standard
6221:   model}.
6222: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B405}, 108--113, 1997.
6223: 
6224: \bibitem{Kusenko:1997si}
6225: A.~Kusenko and M.~Shaposhnikov.
6226: \newblock {\ssl Supersymmetric Q-balls as dark matter}.
6227: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B418}, 46--54, 1998.
6228: 
6229: \bibitem{Lee:1988ag}
6230: K.~Lee, J.A. Stein-Schabes, R.~Watkins, and L.M. Widrow.
6231: \newblock {\ssl Gauged Q-balls}.
6232: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D39}, 1665--1673, 1989.
6233: 
6234: \bibitem{Lee:1991ax}
6235: T.~D. Lee and Y.~Pang.
6236: \newblock {\ssl Nontopological solitons}.
6237: \newblock {\em {\XPEH Phys.Rept.}}, {\bbf 221}, 251--350, 1992.
6238: 
6239: \bibitem{Leggett}
6240: A.J. Leggett.
6241: \newblock {\ssl Bose-Einstein condensation in the alkali gazes: Some
6242:   fundamental concepts}.
6243: \newblock {\em {\XPEH Rev.Mod.Phys.}}, {\bbf 73}, 307--356, 2001.
6244: 
6245: \bibitem{Lemperiere:2003yt}
6246: Y.~Lemperiere and E.~P.~S. Shellard.
6247: \newblock {\ssl Vorton existence and stability}.
6248: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 91}, 141601, 2003.
6249: 
6250: \bibitem{Lin}
6251: F.~Lin and Y.~Yang.
6252: \newblock {\ssl Existence of energy minimizers as stable knotted solitons in
6253:   the Faddeev model}.
6254: \newblock {\em {\XPEH Comm.Math.Phys.}}, {\bbf 249}, 273--303, 2004.
6255: 
6256: \bibitem{Makhankov}
6257: V.G. Makhankov, Y.P. Rybakov, and V.I. Sanyuk.
6258: \newblock {\em {\ssl The Skyrme model, Fundamentals, Methods, Applications}}.
6259: \newblock Springer-Verlag, Berlin, 1993.
6260: 
6261: \bibitem{Manton:2004tk}
6262: N.~Manton and P.~Sutcliffe.
6263: \newblock {\em {\ssl Topological Solitons}}.
6264: \newblock Cambridge University Press, 2004.
6265: \newblock 493 p.
6266: 
6267: \bibitem{Martins:1998gb}
6268: C.J.A.P. Martins and E.P.S. Shellard.
6269: \newblock {\ssl Vorton formation}.
6270: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D57}, 7155--7176, 1998.
6271: 
6272: \bibitem{Meissner}
6273: U.G. Meissner.
6274: \newblock {\ssl Toroidal solitons with unit Hopf charge}.
6275: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B154}, 190--192, 1985.
6276: 
6277: \bibitem{Metlitski:2003gj}
6278: M.~Metlitski and A.R. Zhitnitsky.
6279: \newblock {\ssl Vortex Rings in two Component Bose-Einstein Condensates}.
6280: \newblock {\em {\XPEH JHEP}}, {\bbf 06}, 017, 2004.
6281: 
6282: \bibitem{mihalache}
6283: D.~Mihalache, D.~Mazilu, L.C. Crasovan, I.~Towers, A.V. Buryak, B.A. Malomed,
6284:   L.~Torner, J.P. Torres, and F.~Lederer.
6285: \newblock {\ssl Stable spinning optical solitons in three dimensions}.
6286: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 88}, 073902, 2002.
6287: 
6288: \bibitem{mihalache1}
6289: D.~Mihalache, D.~Mazilu, L.C. Crasovan, I.~Towers, B.A. Malomed, A.V. Buryak,
6290:   L.~Torner, and F.~Lederer.
6291: \newblock {\ssl Stable three-dimensional spinning optical solitons supported by
6292:   competing quadratic and cubic nonlinearities}.
6293: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf E66}, 016613, 2002.
6294: 
6295: \bibitem{mihalache2}
6296: D.~Mihalache, D.~Mazilu, I.~Towers, B.A. Malomed, and F.~Lederer.
6297: \newblock {\ssl Stable spatiotemporal spinning solitons in a bimodal
6298:   cubic-quintic medium}.
6299: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf E67}, 056608, 2003.
6300: 
6301: \bibitem{Nicole}
6302: A.D. Nicole.
6303: \newblock {\ssl Solitons with non-vanishing Hopf index}.
6304: \newblock {\em {\XPEH J.Phys.}}, {\bbf 4}, 1363--1369, 1978.
6305: 
6306: \bibitem{NO}
6307: H.B. Nielsen and P.~Olesen.
6308: \newblock {\ssl Vortex line models for dual strings}.
6309: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B61}, 45--61, 1973.
6310: 
6311: \bibitem{Niemi:2000ny}
6312: A.J. Niemi, K.~Palo, and S.~Virtanen.
6313: \newblock {\ssl (Meta)stable closed vortices in (3+1)+dimensional gauge
6314:   theories with an extended Higgs sector}.
6315: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D61}, 085020, 2000.
6316: 
6317: \bibitem{Paturyan:2004ps}
6318: V.~Paturyan, E.~Radu, and D.H. Tchrakian.
6319: \newblock {\ssl Rotating regular solutions in Einstein-Yang-Mills-Higgs
6320:   theory}.
6321: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B609}, 360--366, 2005.
6322: 
6323: \bibitem{Peter}
6324: P.~Peter.
6325: \newblock {\ssl Superconducting cosmic string: Equation of state for spacelike
6326:   and timelike current in the neutral limit}.
6327: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D45}, 1091--1102, 1992.
6328: 
6329: \bibitem{Piette}
6330: B.M.A.G. Piette, B.J. Schroers, and W.~Zakrzewski.
6331: \newblock {\ssl Dynamics of baby skyrmions}.
6332: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B439}, 205--238, 1995.
6333: 
6334: \bibitem{Piette:1997ny}
6335: B.M.A.G. Piette and D.H. Tchrakian.
6336: \newblock {\ssl Static solutions in the U(1) gauged Skyrme model}.
6337: \newblock {\em {\XPEH Phys.Rev.}}, {\bf D62}, 025020, 2000.
6338: 
6339: \bibitem{Pitaevskii1}
6340: L.P. Pitaevskii.
6341: \newblock {\ssl Vortex lines in an imperfect Bose gas}.
6342: \newblock {\em {\XPEH Sov.Phys.JETP}}, {\bbf 13}, 451--454, 1961.
6343: 
6344: \bibitem{Polyakov:1974ek}
6345: A.M. Polyakov.
6346: \newblock {\ssl Particle spectrum in quantum field theory}.
6347: \newblock {\em {\XPEH JETP Lett.}}, {\bbf 20}, 194--195, 1974.
6348: 
6349: \bibitem{Protogenov:2002bt}
6350: A.P. Protogenov and V.A. Verbus.
6351: \newblock {\ssl Energy bounds of linked vortex states}.
6352: \newblock {\em {\XPEH JETP Lett.}}, {\bbf 76}, 53--55, 2002.
6353: 
6354: \bibitem{Radu:2005jp}
6355: E.~Radu and D.H. Tchrakian.
6356: \newblock {\ssl Spinning U(1) gauged skyrmions}.
6357: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B632}, 109--113, 2006.
6358: 
6359: \bibitem{Rajaraman}
6360: R.~Rajaraman.
6361: \newblock {\em {\ssl Solitons and Instantons}}.
6362: \newblock North Holland, Amsterdam, 1982.
6363: \newblock 418 p.
6364: 
6365: \bibitem{Rayfield}
6366: G.W. Rayfield and F.~Reif.
6367: \newblock {\ssl Quantized vortex rings in superfluid helium}.
6368: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf A136}, 1194--1208, 1964.
6369: 
6370: \bibitem{Ren}
6371: J.~Ren, R.~Li, and Y.~Duan.
6372: \newblock {\ssl Inner topological structure of Hopf invariant}.
6373: \newblock {\em {\XPEH Journ.Math.Phys.}}, {\bbf 48}, 073502, 2007.
6374: 
6375: \bibitem{Rubakov:1985nk}
6376: V.A. Rubakov.
6377: \newblock {\ssl On the electroweak theory at high fermion density}.
6378: \newblock {\em {\XPEH Prog.Theor.Phys.}}, {\bbf 75}, 366--385, 1986.
6379: 
6380: \bibitem{Rubakov:1986am}
6381: V.A. Rubakov and A.N. Tavkhelidze.
6382: \newblock {\ssl Stable anomalous states of superdense matter in gauge
6383:   theories}.
6384: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B165}, 109--112, 1985.
6385: 
6386: \bibitem{Ruostekoski:2004pj}
6387: J.~Ruostekoski.
6388: \newblock {\ssl Stable particlelike solitons with multiply-quantized vortex
6389:   lines in Bose-Einstein condensates}.
6390: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf A70}, 041601, 2004.
6391: 
6392: \bibitem{Ruostekoski:2005zd}
6393: J.~Ruostekoski and Z.~Dutton.
6394: \newblock {\ssl Engineering vortex rings and systems for controlled studies of
6395:   vortex interactions in Bose-Einstein condensates}.
6396: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf A72}, 063626, 2005.
6397: 
6398: \bibitem{Ruostekoski:2001fc}
6399: J.~Ruostekoski and J.~R. J.R.~Anglin.
6400: \newblock {\ssl Creating vortex rings and three-dimensional skyrmions in
6401:   Bose-Einstein condensates}.
6402: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 86}, 3934--3937, 2001.
6403: 
6404: \bibitem{Ruostekoski:2003qx}
6405: J.~Ruostekoski and J.~R. J.R.~Anglin.
6406: \newblock {\ssl Monopole core instability and Alice rings in spinor Bose-
6407:   Einstein condensates}.
6408: \newblock {\em {\XPEH Phys. Rev. Lett.}}, {\bbf 91}, 190402, 2003.
6409: 
6410: \bibitem{Saffman}
6411: P.G. Saffman.
6412: \newblock {\em {\ssl Vortex dynamics}}.
6413: \newblock {\XPEH Cambridge University Press}, 1992.
6414: \newblock 311 p.
6415: 
6416: \bibitem{Savage:2003hh}
6417: C.M. Savage and J.~Ruostekoski.
6418: \newblock {\ssl Energetically stable particle-like Skyrmions in a trapped
6419:   Bose-Einstein condensate}.
6420: \newblock {\em {\XPEH Phys.Rev.Lett.}}, {\bbf 91}, 010403, 2003.
6421: 
6422: \bibitem{FIDISOL1}
6423: M.~Schauder, R.~Wei\ss\, and W.~Sch\"onauer.
6424: \newblock {\ssl The CADSOL Program Package, Universit\"at Karlsruhe Interner
6425:   Bericht Nr. 46/92 }, 1992.
6426: 
6427: \bibitem{Schmid:2007dm}
6428: M.~Schmid and M.~Shaposhnikov.
6429: \newblock {\ssl Anomalous Abelian solitons}.
6430: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B775}, 365--389, 2007.
6431: 
6432: \bibitem{FIDISOL}
6433: W.~Sch\"onauer and R.~Wei\ss.
6434: \newblock {\ssl The Fidisol blackbox solver}.
6435: \newblock {\em {\XPEH J. Comput. Appl. Math.}}, {\bbf 27}, 279--297, 1989.
6436: 
6437: \bibitem{Shabanov:1999xy}
6438: S.V. Shabanov.
6439: \newblock {\ssl An effective action for monopoles and knot solitons in
6440:   Yang-Mills theory}.
6441: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B458}, 322--330, 1999.
6442: 
6443: \bibitem{Shabanov:1999uv}
6444: S.V. Shabanov.
6445: \newblock {\ssl Yang-Mills theory as an Abelian theory without gauge fixing}.
6446: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B463}, 263--272, 1999.
6447: 
6448: \bibitem{Shnir:2005te}
6449: Y.~Shnir.
6450: \newblock {\ssl Electromagnetic interaction in the system of multimonopoles and
6451:   vortex rings}.
6452: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D72}, 055016, 2005.
6453: 
6454: \bibitem{Shnir}
6455: Ya.M. Shnir.
6456: \newblock {\em {\ssl Magnetic monopoles}}.
6457: \newblock {\XPEH Springer-Verlag, Berlin-Heidelberg}, 2005.
6458: \newblock 532 p.
6459: 
6460: \bibitem{Skyrme:1961vq}
6461: T.H.R. Skyrme.
6462: \newblock {\ssl A nonlinear field theory}.
6463: \newblock {\em {\XPEH Proc.Roy.Soc.Lond.}}, {\bbf A260}, 127--138, 1961.
6464: 
6465: \bibitem{Sutcliffe:2007ui}
6466: P.~Sutcliffe.
6467: \newblock {\ssl Knots in the Skyrme-Faddeev model}, 2007.
6468: \newblock {\tt arXiv:0705.1468 [hep-th]}.
6469: 
6470: \bibitem{Sutcliffe:2007vm}
6471: P.~Sutcliffe.
6472: \newblock {\ssl Vortex rings in ferromagnets}, 2007.
6473: \newblock {\tt arXiv:0707.1383 [cond-mat.mes-hall]}.
6474: 
6475: \bibitem{'tHooft:1974qc}
6476: G.~'t~Hooft.
6477: \newblock {\ssl Magnetic monopoles in unified gauge theories}.
6478: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B79}, 276--284, 1974.
6479: 
6480: \bibitem{Taubes:1982ie}
6481: C.H. Taubes.
6482: \newblock {\ssl The existence of a non-minimal solution to the SU(2)
6483:   Yang-Mills-Higgs equations on $\mathbb{R}^3$}.
6484: \newblock {\em {\XPEH Commun.Math.Phys.}}, {\bbf 86}, 257--298, 1982.
6485: 
6486: \bibitem{Kelvin}
6487: W.H. Thomson.
6488: \newblock {\ssl On vortex motion}.
6489: \newblock {\em {\XPEH Trans.R.Soc.Edin.}}, {\bbf 25}, 217--260, 1867.
6490: 
6491: \bibitem{Vakulenko:1979uw}
6492: A.F. Vakulenko and L.V. Kapitansky.
6493: \newblock {\ssl Stability of solitons in $S^2$ in the nonlinear
6494:   $\sigma$-model.}
6495: \newblock {\em {\XPEH Sov.Phys.Dokl.}}, {\bbf 24}, 433--434, 1979.
6496: 
6497: \bibitem{vanBaal:2001jm}
6498: P.~van Baal and A.~Wipf.
6499: \newblock {\ssl Classical gauge vacua as knots}.
6500: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B515}, 181--184, 2001.
6501: 
6502: \bibitem{Vilenkin}
6503: A.~Vilenkin and E.P.S. Shellard.
6504: \newblock {\em {\ssl Cosmic Strings and Other Topological Defects}}.
6505: \newblock Cambridge University Press, 1994.
6506: \newblock 517 p.
6507: 
6508: \bibitem{Volkov:2006ug}
6509: M.S. Volkov.
6510: \newblock {\ssl Superconducting electroweak strings}.
6511: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B644}, 203--207, 2007.
6512: 
6513: \bibitem{VG}
6514: M.S. Volkov and D.V. Gal'tsov.
6515: \newblock {\ssl Gravitating non-Abelian solitons and black holes with
6516:   Yang-Mills fields}.
6517: \newblock {\em {\XPEH Phys.Rep.}}, {\bbf 319}, 1--83, 1999.
6518: 
6519: \bibitem{Volkov:2002aj}
6520: M.S. Volkov and E.~Wohnert.
6521: \newblock {\ssl Spinning Q-balls}.
6522: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D66}, 085003, 2002.
6523: 
6524: \bibitem{Volkov:2003ew}
6525: M.S. Volkov and E.~Wohnert.
6526: \newblock {\ssl On the existence of spinning solitons in gauge field theory}.
6527: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D67}, 105006, 2003.
6528: 
6529: \bibitem{Volovik}
6530: G.E. Volovik.
6531: \newblock {\em {\ssl The Universe in a helium droplet}}.
6532: \newblock {\XPEH Oxford University Press}, 2003.
6533: \newblock 507 p.
6534: 
6535: \bibitem{Ward:1998pj}
6536: R.S. Ward.
6537: \newblock {\ssl Hopf solitons on $S^3$ and $R^3$}, 1998.
6538: \newblock {\tt hep-th/9811176}.
6539: 
6540: \bibitem{Ward:2000qj}
6541: R.S. Ward.
6542: \newblock {\ssl The interaction of two Hopf solitons}.
6543: \newblock {\em {\XPEH Phys.Lett.}}, {\bbf B473}, 291--296, 2000.
6544: 
6545: \bibitem{Ward2}
6546: R.S. Ward.
6547: \newblock {\ssl Hopf solitons from instanton holonomy}, 2001.
6548: \newblock {\tt hep-th/0108082}.
6549: 
6550: \bibitem{Ward:2002vq}
6551: R.S. Ward.
6552: \newblock {\ssl Stabilizing textures with magnetic fields}.
6553: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D66}, 041701, 2002.
6554: 
6555: \bibitem{Ward:2004gr}
6556: R.S. Ward.
6557: \newblock {\ssl Skyrmions and Faddeev-Hopf solitons}.
6558: \newblock {\em {\XPEH Phys.Rev.}}, {\bbf D70}, 061701, 2004.
6559: 
6560: \bibitem{Ward}
6561: R.S. Ward.
6562: \newblock {\ssl Hopf solitons on the lattice}.
6563: \newblock {\em {\XPEH J.Phys.}}, {\bbf A39}, L105--L109, 2006.
6564: 
6565: \bibitem{Witten:1984eb}
6566: E.~Witten.
6567: \newblock {\ssl Superconducting strings}.
6568: \newblock {\em {\XPEH Nucl.Phys.}}, {\bbf B249}, 557--592, 1985.
6569: 
6570: \bibitem{Zakharov}
6571: V.A. Zakharov and E.A. Kuznetsov.
6572: \newblock {\ssl Hamiltonian formalism for nonlinear waves}.
6573: \newblock {\em {\XPEH Phys.Usp.}}, {\bbf 40}, 1087--1116, 1997.
6574: 
6575: \end{thebibliography}
6576: 
6577: \end{document}
6578: 
6579:  
6580: