1: \documentclass[12pt,showpacs,preprint,nofootinbib]{revtex4}
2: \usepackage[intlimits]{amsmath}
3: \usepackage{amssymb}
4: \usepackage{exscale}
5: \usepackage{graphicx}
6: \usepackage{setspace}
7:
8: \newcommand{\ud}{\mbox{d}}
9: \newcommand{\CL}{\mathcal{L}}
10: \newcommand{\CD}{\mathcal{D}}
11: \newcommand{\CE}{\mathcal{E}}
12: \newcommand{\h}{\frac{1}{2}}
13: \newcommand{\matlab}{\textsc{Matlab}\textsuperscript{\textregistered}}
14:
15: \begin{document}
16:
17:
18: ~
19:
20: \vspace{2 cm}
21:
22: \title {Spinning Q-Balls}
23:
24: \preprint{FSU TPI 01/02}
25: \preprint{hep-th/0205157}
26:
27: \author{Mikhail S. Volkov\footnotemark[2]}
28: \author{Erik W\"ohnert\footnotemark[3]}
29:
30: \vspace{1 cm}
31:
32: \affiliation{\footnotemark[2]Laboratoire de Math\'ematiques et Physique
33: Th\'eorique, Universit\'e de Tours, Parc de Grandmont, 37200 Tours,
34: FRANCE \footnotetext[2]{\tt volkov@phys.uni-tours.fr} \\
35: \footnotemark[3]Theoretisch-Physikalisches Institut, Friedrich
36: Schiller Universit\"at Jena, Fr\"obelstieg 1, 07743 Jena,
37: GERMANY \footnotetext[3]{\tt pew@tpi.uni-jena.de}}
38:
39:
40: \vspace{1 cm}
41:
42: \begin{abstract}
43:
44: \noindent
45: We present numerical evidence for the existence of spinning
46: generalizations for
47: non-topological Q-ball solitons in the theory of a complex scalar field
48: with a non-renormalizable self-interaction.
49: To the best of our knowledge, this provides the first explicit example
50: of spinning solitons in $3+1$ dimensional Minkowski space.
51: In addition, we find an infinite discrete family of radial excitations of
52: non-rotating Q-balls,
53: and construct also spinning Q-balls in $2+1$ dimensions.
54: \end{abstract}
55:
56:
57:
58:
59: \pacs{11.27.+d}
60: \keywords{Q-balls, solitons}
61: \maketitle
62:
63:
64:
65: %\setlength{\parindent}{0cm}
66:
67:
68: \section{Introduction}
69:
70: Solitons are important ingredients of models in high energy physics.
71: Apart from
72: %saturating the path integral and
73: being responsible for various non-perturbative quantum phenomena,
74: solitons are interesting in themselves, since they can be viewed as
75: field theoretic realizations %models
76: of elementary particles. It is from this viewpoint that
77: solitons were originally introduced into physics in the context of the
78: Skyrme model
79: %in the early 1960's
80: -- as models of hadrons. The subsequent developments have revealed
81: soliton solutions in many other non-linear field theories in Minkowski
82: space, such as %for example
83: monopoles, vortices, sphalerons, Q-balls, etc. These solutions describe
84: localized, particle-like objects with finite energy. Their spectra of
85: energy and charge are typically discrete. In addition, these solutions
86: are regular everywhere, a property which is especially appealing. So
87: far, however, solitons have been lacking one important feature of
88: elementary particles: the intrinsic angular momentum $J$, which is zero
89: for all known classical solutions.
90:
91: It is sensible to ask whether \textit{stationary} rotating
92: generalizations for the known static soliton solutions exist. More
93: precisely,
94: %by this are meant
95: one is interested in finite energy, globally regular, non-radiating
96: solutions for which the spatial integral of the $T_{0\varphi}$ component
97: of the energy-momentum tensor,
98: \begin{eqnarray}
99: \label{J}
100: J & = & \int T_{0\varphi} \, \ud^3 x \;,
101: \end{eqnarray}
102: is non-vanishing. In this definition the condition of stationarity
103: (absence of radiation) is important. Indeed, it is always possible to
104: construct a field configuration such that $J \neq 0$ at the initial
105: moment of time. Physically this would correspond to exciting the soliton
106: to give it an angular momentum%
107: \footnote{In the literature one can often find explicit examples of
108: `solitons in the rigid rotator approximation',
109: for which the integral (\ref{J}) does
110: not vanish, as for instance `rotating' Skyrmions
111: \cite{Adkins:1983ya}, knots \cite{Gladikowski:1997mb}, etc. These
112: configurations, however, are not solutions of the equations of motion,
113: and at best they can be viewed %only
114: as the initial values for the dynamical evolution problem.}.
115: However, when the time evolution starts, the received excitation will be
116: most probably immediately radiated away, leaving behind a non-rotating
117: object.
118:
119: When talking about rotating solitons, it also seems sensible to
120: distinguish between two types of rotation: \textit{spinning} and
121: \textit{orbiting}. \textit{Spinning} is associated with the
122: intrinsic angular momentum, in analogy with the quantum-mechanical spin.
123: Classical spin excitations, if they exist, should live in the
124: one-soliton sector because they are excitations of an individual
125: object. One expects that the corresponding angular momentum will assume
126: only discrete values. Solutions describing spinning solitons in
127: Minkowski space in $3+1$ dimensions are not known. For example, it is
128: not known whether the 't Hooft-Polyakov monopoles can be given classical
129: spin\footnote{There is also the possibility to associate the spin
130: of the monopole with that of fermionic zero modes living in the
131: monopole background. In this case, however, the spin is not classical,
132: and in fact is not related to the monopole itself.}.
133: In fact, it has been shown that, at the \textit{perturbative}
134: level, monopoles do not admit stationary rotational excitations
135: in the one-monopole sector
136: \cite{Heusler:1998ec}. This means that monopoles
137: cannot rotate \textit{slowly}, with $|J| \ll 1$. Monopoles with finite
138: (discrete) values of $J$ are not yet excluded.
139: However, to decide
140: whether such solutions exist requires to go beyond perturbation theory
141: and solving the complete coupled system of the Yang-Mills-Higgs (YMH)
142: partial differential equations (PDEs), which is an exceedingly
143: difficult task%
144: \footnote{It was argued in \cite{VanderBij:2001nm}
145: that at least within the minimal axial ansatz,
146: rotating 't Hooft-Polyakov monopoles can be excluded also
147: at the non-perturbative level.}.
148:
149:
150: On the other hand, one can consider relative \textit{orbital} motions in
151: composite many-soliton systems. For example, one can imagine a rotating
152: soliton-antisoliton pair balanced against mutual attraction by the
153: centrifugal force. The spectrum of the angular momentum is then expected
154: to be continuous. Solutions describing such orbiting solitons can
155: actually be constructed. In the case of monopoles, for instance, there
156: exist \cite{Taubes:1982ie,Taubes:1982if} solitonic solutions of the YMH
157: field equations with vanishing Higgs potential in the sector with zero
158: monopole charge. Such static, purely magnetic solutions have been
159: explicitly constructed in \cite{Kleihaus:1999sx}; they describe
160: monopole-antimonopole pairs balanced by a repulsive force of
161: topological nature. Now, there is a simple way to add angular
162: momentum to these solutions \cite{Heusler:1998ec} by using the global
163: symmetry of the field equations which mixes the Higgs field $H$ and the
164: electric potential $\Phi$:
165: \begin{eqnarray}
166: \label{hyp}
167: H & \to & H \cosh \gamma + \Phi \sinh \gamma \;, \qquad
168: \Phi \; \to \; \Phi \cosh \gamma + H \sinh \gamma \;.
169: \end{eqnarray}
170: Applying this transformation with an arbitrary $\gamma \neq 0$ leads to
171: solutions with an electric field. It is important to note that this also
172: changes the angular momentum to
173: $J \sim \sinh \gamma$ \cite{Heusler:1998ec}%
174: \footnote{This trick works only for non-BPS solutions -- those obeying
175: the second order YMH equations but not the first order
176: Bogomol'nyi equations. For BPS solutions the angular momentum is
177: invariant under (\ref{hyp}). For this reason one cannot use this method
178: to produce spinning monopoles, because non-BPS solutions with
179: finite energy and unit topological charge are not known
180: \cite{Maison:1981ze}.}
181: (in fact
182: $J=Q$, where $Q$ is the electric charge \cite{VanderBij:2001nm}).
183: These new solutions can be interpreted as describing the system of a
184: monopole and antimonopole rotating around their common center of mass.
185:
186: Another example of systems which could be classified as orbiting
187: solitons are rotating vortex loops. In certain models there exist vortices
188: with stationary currents along them; `superconducting
189: vortices'. One can argue \cite{Davis:1988ip,Davis:1989ij} that taking a
190: finite piece of such a vortex, bending and closing it to form a loop,
191: finally adding a momentum along the loop, leads to an object (vorton)
192: described by a stationary solution of the equations of
193: motion\footnote{To our knowledge such solutions have not been
194: constructed explicitly.}. The angular momentum in this case is
195: associated with the macroscopic circulation of the current along the
196: loop. This type of motion could be naturally classified as % the
197: orbital rotation.
198:
199: On the other hand, one can also consider the rotation of a straight
200: vortex \textit{along} its symmetry axis. This would % perhaps
201: correspond more closely to the notion of an intrinsic spinning
202: rotation. Although one can show that for the Nielsen-Olesen vortex such
203: spinning excitations do not exist, they can actually exist in other
204: models
205: \cite{deVega:1986eu,Jackiw:1990aw,Kim:1993mm,Piette:1995mh,Gisiger:1997vb}.
206: However, this only
207: gives spinning solitons in $2+1$ dimensions, while their $3+1$
208: dimensional analogs will have infinite energy due to the infinite length
209: of the vortex.
210:
211: In summary, we are not aware of any spinning solitons in Minkowski
212: space in $3+1$ dimensions. At the same time, such solutions are known in
213: curved space. Indeed, there are many rotating solutions in General
214: Relativity. In the pure gravity case they comprise the family of
215: Kerr-Newman black holes. These are very similar to solitons, but
216: they are not globally regular and contain a curvature
217: singularity. There are also globally regular rotating solutions
218: \cite{Neugebauer}, but these require a source of a non-field theoretic
219: origin. Interestingly, there exist
220: two explicit examples of gravitating spinning solitons in pure field
221: systems.
222:
223: The first example is provided by rotating boson stars
224: \cite{Schunk96,Yoshida:1997qf}.
225: These are solutions for a gravity-coupled massive complex scalar field
226: with harmonic dependence on time and on the azimuthal angle,
227: $\Phi(t,r,\vartheta,\varphi) = \exp(i \omega t + i N \varphi)
228: f(r,\vartheta)$, with $N$ integer. The energy-momentum tensor is
229: time-independent, and the Einstein equations together with the
230: Klein-Gordon equation for $f(r,\vartheta)$ admit globally regular,
231: stationary particle-like solutions \cite{Schunk96}. The angular momentum
232: is quantized as $J \sim N$. The solutions with $N > 0$ can be regarded
233: as spinning excitations of the fundamental static, spherically symmetric
234: solutions with $N = 0$.
235:
236: This example is instructive in the sense that it is clear
237: `what rotates': this is the phase $\omega t + N \varphi$. In
238: this connection it is worth noting that the very notion of rotation in
239: pure field systems is very different from that for ordinary rigid
240: bodies. Indeed, it is meaningless to say that a given element of volume
241: of a field system actually `performs revolutions' around a given axis.
242: Of course, one can imagine a solitonic object with a field perturbation
243: running around it. However, such a moving perturbation will be most
244: probably immediately radiated away. The example of boson stars thus
245: shows that one can nevertheless have a rotating phase which is not
246: radiated away.
247:
248: In stationary rotating systems without explicit time dependence the
249: rotation will rather be associated with certain non-linear
250: superpositions of the multipole moments of the fields. For example, in
251: systems with vector fields, angular momentum may be present
252: due to a non-vanishing integral involving the Poynting vector,
253: \begin{eqnarray}
254: \label{1}
255: \vec{J} & = & \int \vec{r}\times(\vec{E}\times\vec{B}) \, \ud^3 x \;.
256: \end{eqnarray}
257: A stationary, globally regular configuration for which this
258: integral is non-zero would correspond to a rotating
259: soliton.
260:
261: It is possible that such solutions could exist for the
262: SU(2) Yang-Mills fields coupled to gravity. The Einstein-Yang-Mills
263: (EYM) field equations admit globally regular, particle like solutions
264: \cite{Bartnik:1988am}. These gravitating EYM particles are static,
265: spherically symmetric and neutral (their purely magnetic gauge field
266: strength decays asymptotically as $1/r^3$). Now,
267: for these solutions one can {\it perturbatively} construct stationary,
268: globally regular,
269: axially-symmetric, slowly rotating generalizations
270: \cite{Brodbeck:1997ek}. Surprisingly, their spectrum of $J$ is
271: \textit{continuous}, as if they were composite objects, which is
272: presumably due to the special feature of the
273: field system consisting of only \textit{massless} physical fields.
274: Unfortunately,
275: it is not clear at the moment whether these perturbative
276: solutions exist also at the
277: non-perturbative level -- the analysis of
278: \cite{VanderBij:2001nm,Kleihaus:2002ee} indicates
279: the opposite.
280: It is however still possible that
281: the solitons exist, but
282: within a more general ansatz
283: than that considered in \cite{VanderBij:2001nm,Kleihaus:2002ee}.
284:
285:
286:
287: In summary, spinning solitons have been
288: found only in curved space. It is then sensible to ask if spinning solitons
289: without gravity exist at all. In principle, it is not excluded that only
290: gravity can support the relevant rotational degrees of freedom. In order
291: to rule out this logical possibility, we have undertaken an attempt to
292: construct spinning solitons in %Minkowski
293: flat space.
294:
295: The solitons we have chosen to `rotate' are Q-balls
296: \cite{Coleman:1985ki,Lee:1992ax}. These are solutions for a complex
297: scalar field with a non-renormalizable self-interaction arising
298: in some effective field theories. These solutions circumvent the
299: standard Derrick-type argument due to having a time-dependent phase for
300: the scalar field. In this sense they are somewhat similar to the boson
301: stars. In the simplest spherically symmetric case the fundamental Q-ball
302: solution was described by Coleman \cite{Coleman:1985ki}. Dynamical
303: properties of these objects have been studied in \cite{Battye:2000qj}.
304: Q-balls also appear in supersymmetric generalizations of the standard
305: model, where one finds leptonic and baryonic Q-balls
306: \cite{Kusenko:1997zq} which may be responsible for the generation of
307: baryon number or may be regarded as candidates for dark
308: matter \cite{Kusenko:1998si}.
309:
310: In the present paper, we first recall % describe
311: the properties of the fundamental spherically symmetric Q-balls. In
312: addition, we present an infinite family of their radial excitations
313: which have not been reported in the literature before. We then turn to
314: rotating solutions and first consider them in $2+1$ spacetime
315: dimensions, where the problem reduces to an ordinary differential
316: equation. Finally we consider the full problem in $3+1$ dimensions and
317: explicitly construct solutions with angular momentum. To our knowledge,
318: our analysis gives the first explicit example of spinning solitons in
319: flat space.
320:
321:
322: \section{The Model}
323:
324: Let us consider a theory of a complex scalar field in $3+1$ dimensions
325: defined by the Lagrangian
326: \begin{eqnarray}
327: \label{lagdens}
328: {\CL} & = & \partial_{\mu} \Phi \, \partial^{\mu} \Phi^\ast - U(|\Phi|)
329: \;.
330: \end{eqnarray}
331: It is assumed that the potential $U$ has its global minimum at $\Phi =
332: 0$, where $U(0) = 0$, while $U \to \infty$ for $|\Phi| \to \infty$. In
333: addition, the potential must fulfill a particular
334: inequality (Eq.(\ref{cond3})) which will be discussed below.
335: %must fulfill the condition (\ref{cond3}), which will be discussed below.
336: The potential may also have local minima at some finite $|\Phi|$, as is
337: shown in Fig.~\ref{fig1}, but this is not necessary.
338: \begin{figure}[ht]
339: \begin{minipage}[b]{0.45\linewidth}
340: \centering
341: \includegraphics[width=\linewidth]{potential}
342: \caption{ The qualitative shape of the potential $U(|\Phi|)$.}
343: \label{fig1}
344: \end{minipage}
345: \hspace{5 mm}
346: \begin{minipage}[b]{0.45\linewidth}
347: \centering
348: \includegraphics[width=\linewidth]{effpot}
349: \caption{The effective potential $V(\phi) = \frac{1} {2} \omega^2
350: \phi^2 - U(\phi)$ in Eq.~(\ref{radeom2}).}
351: \label{fig2}
352: \end{minipage}
353: \end{figure}
354:
355: The global symmetry of the Lagrangian under $\Phi \rightarrow \Phi
356: e^{i \alpha}$ gives rise to the conserved charge
357: %current $j_{\mu} = (\Phi^{*}
358: %\partial_{\mu} \Phi - \Phi \partial_{\mu} \Phi^{*}) /(2i)$ and to the
359: %conserved charge
360: \begin{eqnarray}
361: \label{charge1}
362: Q & = & \frac{1}{i} \int \ud^3 x \, (\Phi^{*} \dot{\Phi} - \Phi
363: \dot{\Phi}^{*}) \;.
364: \end{eqnarray}
365: The fundamental Q-ball solutions of the theory are minima of
366: the energy for a given $Q$ \cite{Coleman:1985ki}. Since $\Phi$ should
367: depend on time for $Q$ to be non-vanishing, one assumes that $\Phi$
368: has a harmonic time dependence. In the spherically symmetric case,
369: \begin{eqnarray}
370: \label{radphi}
371: \Phi & = & \phi(r) \, e^{i \omega t} \;,
372: \end{eqnarray}
373: where $\phi(r)$ is real. The potential $U(|\Phi|) = U(\phi)$ and the
374: energy-momentum tensor,
375: \begin{eqnarray}
376: \label{radt}
377: T_{\mu\nu} & = & \partial_\mu \Phi \, \partial_\nu \Phi^\ast +
378: \partial_\nu \Phi \, \partial_\mu \Phi^\ast - g_{\mu\nu} \, \CL \;,
379: \end{eqnarray}
380: ($g_{\mu\nu}$ being the spacetime metric) do not depend on time. The
381: energy distribution is therefore stationary, and the total energy is
382: \begin{eqnarray}
383: \label{radenergy}
384: E & = & 4 \pi \int_0^\infty \ud r \, r^2 (\omega^2 \phi^2 +
385: \phi^{\prime 2} + U(\phi)) \;,
386: \end{eqnarray}
387: where the prime denotes differentiation with respect to $r$. The field
388: equation,
389: \begin{eqnarray}
390: \label{eq}
391: 0 & = & \frac{1} {\sqrt{-g}} \partial_\mu (\sqrt{-g} \, g^{\mu\nu}
392: \partial_\nu \Phi) + \frac{\partial U} {\partial \Phi^\ast} \;,
393: \end{eqnarray}
394: reduces to
395: \begin{eqnarray}
396: \label{radeom}
397: 0 & = & \phi'' + \frac{2} {r} \, \phi' - \frac{\ud U(\phi)} {\ud\phi} +
398: \omega^2 \phi \;,
399: \end{eqnarray}
400: which is equivalent to
401: \begin{eqnarray}
402: \label{radeom2}
403: \h \, \phi^{\prime 2} + \h \,\omega^2 \phi^2 - U & = & \CE - 2 \int_0^r
404: \frac{\ud r} {r} \, \phi^{\prime 2} \;.
405: \end{eqnarray}
406: This effectively describes a particle moving with friction in the one
407: dimensional potential
408: \begin{eqnarray}
409: V(\phi) = \h \, \omega^2 \phi^2-U(\phi) \;.
410: \end{eqnarray}
411: $\CE$ is an integration constant playing the role of the total
412: `effective energy'. It is essential that the potential $V(\phi)$ should
413: have the qualitative shape shown in Fig.~\ref{fig2}, which is possible if
414: the following conditions are fulfilled. First, since $V''(0) < 0$, it
415: follows that $\omega^2$ should not be too large:
416: \begin{eqnarray}
417: \label{cond1}
418: \omega^2 & < & \omega^{2}_{\max} \; \equiv \; U''(0) \;.
419: \end{eqnarray}
420: On the other hand, $\omega^2$ should not be too small, since otherwise
421: $V(\phi)$ will be always negative. $V(\phi)$ will become positive for
422: some non-zero $\phi$, as is shown in Fig.~\ref{fig2}, if only
423: \begin{eqnarray}
424: \label{cond2}
425: \omega^2 & > & \omega^{2}_{\min} \; \equiv \; \min_{\phi}
426: ({2U(\phi)}/{\phi^2})
427: \;,
428: \end{eqnarray}
429: where the minimum is taken over all values of $\phi$. For the potential
430: $U(\phi)$ it is necessary that
431: %The set of allowed
432: %values of $\omega^2$ will therefore be non-empty if only
433: \begin{eqnarray}
434: \label{cond3}
435: U''(0) & > & \min_{\phi} ({2U(\phi)}/{\phi^2}) \;,
436: \end{eqnarray}
437: %which is the necessary condition the potential $U$ must fulfill.
438: since only then the set of values of $\omega^2$ will be non-empty. The
439: only possible renormalizable interaction in the theory, $U = \h \mu^2
440: |\Phi|^2 + \lambda |\Phi|^4$, does not obey this condition.
441: % for which reason
442: Thus non-renormalizable potentials have to be considered
443: \cite{Coleman:1985ki}. For example, for the potential
444: \begin{eqnarray}
445: \label{polynpot}
446: U(\phi) & = & \lambda (\phi^6 - a \phi^4 + b \phi^2)
447: \end{eqnarray}
448: the condition (\ref{cond3}) is fulfilled for any positive $\lambda$,
449: $a$, $b$, and $U(\phi)$ will have a global minimum at $\phi = 0$ if $b >
450: a^2/4$. We use this model potential with $\lambda=1$, $a=2$ and $b=1.1$
451: in all our calculations below, in which case the conditions
452: (\ref{cond1}),~(\ref{cond2}) require that $0.2 \leq \omega^2 \leq 2.2$.
453:
454:
455: \section{Fundamental Q-balls and their radial excitations}
456:
457: If conditions (\ref{cond1})--(\ref{cond3}) are fulfilled, then the field
458: equation admits globally regular solutions $\phi(r)$ with finite energy
459: \cite{Coleman:1985ki}. The necessary condition for the energy
460: (\ref{radenergy}) to be finite is that the potential $U \to 0$ for
461: large $r$, and therefore $\phi \to 0$ as $r \to \infty$. Linearizing
462: Eq.~(\ref{radeom}) around $\phi = 0$, one finds that asymptotically
463: \begin{eqnarray}
464: \label{asymptrad}
465: \phi & = & \frac{A} {r} \exp \left\{-\sqrt{(U''(0) - \omega^2)} \, r
466: \right\} \left(1+O(1/r)\right) \;,
467: \end{eqnarray}
468: where $A$ is an integration constant. In view of (\ref{cond1}) the
469: argument of the exponent is real and negative, and so $\phi$ approaches
470: zero exponentially fast.
471:
472: Solutions must also be regular at the origin of the coordinate system,
473: $r=0$. Since $r=0$ is the regular singular point of Eq.~(\ref{radeom}),
474: the solution will only be regular if it belongs to the `stable manifold'
475: characterized by the local Taylor expansion in the vicinity of $r=0$,
476: \begin{eqnarray}
477: \label{expanrad}
478: \phi(r) & = & \phi_0 + (U'(\phi_0)-\omega^2\phi_0) \, r^2 + O(r^4) \;,
479: \end{eqnarray}
480: where $\phi_0$ is an integration constant.
481:
482: Extending the two local solutions (\ref{asymptrad}),~(\ref{expanrad}) to
483: finite values of $r$ and requiring that $\phi$ and $\phi'$ for both
484: solutions agree at some $r = r_0$, yields two conditions for the free
485: parameters $A$ and $\phi_0$. Resolving these conditions determines a
486: globally regular solution in the interval $r \in [0,\infty)$. In fact,
487: in this way an infinite discrete family of globally regular solutions
488: parametrized by the number $n = 0,1,2,\ldots$ of nodes of $\phi$ is
489: obtained. The existence of these solutions can be illustrated by the
490: following qualitative argument. % below.
491:
492: \begin{figure}[htb]
493: \begin{minipage}[t]{0.45\linewidth}
494: \centering
495: \includegraphics[width=\linewidth]{radballs}
496: \caption{The profile of $\phi(r)$ for the fundamental Q-ball solution
497: and its first two radial excitations.}
498: \label{fig3}
499: \end{minipage}
500: \hspace{5mm}
501: \begin{minipage}[t]{0.45\linewidth}
502: \centering
503: \includegraphics[width=\linewidth]{radendens}
504: \caption{The radial energy density
505: $r^2T_{00}$ for the fundamental Q-ball and
506: its first two excitations.}
507: \label{fig4}
508: \end{minipage}
509: \end{figure}
510:
511: The parameter $\phi_0$ in (\ref{expanrad}) is the coordinate of the
512: `particle' at the initial moment of `time', $r=0$, when the particle
513: velocity is zero, $\phi'(0) = 0$. The particle therefore starts its
514: motion from some point on the curve $V(\phi)$ in Fig.~\ref{fig2}, with
515: the total effective energy $\CE$ being equal to its potential energy.
516: Then it moves to the right, dissipating some of its energy along its
517: way.
518: %as it goes.
519: For $r \to \infty$, the particle must end up at the local maximum of the
520: potential $V$ (at $\phi = 0$) with %zero value of
521: the total effective energy being zero. This can be achieved by
522: fine-tuning the initial position of the particle, $\phi_0$. If
523: $V(\phi_0) < 0$, the effective
524: energy of the particle will always be \textit{negative}, and therefore
525: it will not be able to end up in a configuration with zero energy. On
526: the other hand, if $\phi_0$ is such that the particle starts very close
527: to the absolute maximum of $V$, then it will stay there for a long
528: `time' $r$, during which period the dissipation term in (\ref{radeom2}),
529: which is $\sim 1/r$, will become very small. As a result, when the
530: particle will eventually start moving, its energy will be too large, and
531: so it will `overshoot' the position with zero energy. By continuity,
532: there is a value $\phi_0$ for which the total effective energy (the
533: right hand side in (\ref{radeom2})) is exactly zero for $r \to \infty$,
534: and so the particle will travel from $\phi(0) = \phi_0$ to $\phi(\infty)
535: = 0$ \cite{Coleman:1985ki}.
536:
537: Next, one can fine-tune $\phi_0$ such that the initial energy
538: $V(\phi_0)$ is slightly too large, so that the particle first overshoots
539: the $\phi=0$ position, but then it hits the barrier from the other side,
540: bounces back and dissipates just enough energy to finally arrive at
541: $\phi=0$ with zero energy. This will give a solution with one node of
542: $\phi(r)$ in the interval $r \in [0,\infty)$. Similarly one can obtain
543: solutions with $n > 1$. To recapitulate, for each $\omega^2$ subject to
544: (\ref{cond1}), (\ref{cond2}) there is a solution to Eq.~(\ref{radeom})
545: for which $\phi(r)$ smoothly interpolates between some finite value at
546: the origin and zero value at infinity, crossing zero $n$ times
547: in between. We shall call solutions with $n = 0$ fundamental Q-balls.
548: Solutions with $n>0$ are the radial Q-ball excitations. The family of
549: Q-ball solutions can thus be parameterized by $(\omega,n)$.
550:
551: As the frequency $\omega$ changes in the range $\omega_{\min}^2 <
552: \omega^2 < \omega_{\max}^2$, the charge
553: \begin{eqnarray}
554: \label{charge}
555: Q(\omega) & = & 8 \pi \omega \int_0^\infty \ud r \, r^2 \phi^2
556: \end{eqnarray}
557: changes from $Q(\omega_{\min}) = \pm \infty$ (depending on the sign of
558: $\omega$) to $Q(\omega_{\max}) = 0$.
559: %Spherically symmetric Q-balls therefore comprise a two-parameter
560: %family labeled by $(\omega,n)$, where $n=0,1,2,\ldots$ and
561: %$\omega_{\min} < \omega < \omega_{\max}$.
562: This can be understood as follows. For $\omega^2 \to \omega_{\max}^2$
563: the minimum of the effective potential $V(\phi)$ becomes more and more
564: shallow and moves closer and closer to $\phi = 0$. This implies that the
565: range of values of the solution $\phi(r)$ diminishes and the interval of
566: $r$ in which $\phi(r)$ is not constant decreases. Q-balls therefore
567: `shrink' in this limit, and the integral (\ref{charge}) tends to zero%
568: \footnote{Although the exponent in (\ref{asymptrad}) decays slower and
569: slower for $\omega^2 \to U''(0)$, the prefactor $A$ vanishes faster
570: still.}. In the opposite limit, $\omega^2 \to \omega_{\min}^2$, Q-balls
571: become large, and their charge tends to infinity.
572:
573: These considerations imply that instead of $\omega$ one can choose the
574: charge $Q$ as the independent parameter. Spherically symmetric $Q$-balls
575: therefore comprise a two-parameter family labeled by $(Q,n)$, where the
576: charge $-\infty < Q < \infty$ and the `excitation number' $n =
577: 0,1,2,\ldots$. In Figs.~3--4 the profiles of the fundamental solution and
578: its first two radial excitations are shown for $Q = 1100$. As one can
579: see from the shape of the radial energy density,
580: the $n$-th solution
581: has the structure of $n+1$ concentric spherical layers of energy%
582: \footnote{To our knowledge, solutions with $n > 0$ have not been
583: reported in the literature so far.}.
584:
585:
586: \section{Spinning Q-vortices.}
587:
588: Our aim now is to show that Q-balls admit spinning generalizations. For
589: this we modify %generalize
590: the ansatz (\ref{radphi}) according to
591: \begin{eqnarray}
592: \label{phirot}
593: \Phi & = & \phi(r,\vartheta) \, e^{i \omega t + i N \varphi} \;,
594: \end{eqnarray}
595: where $N$ is an integer.
596: %The phase of the
597: %field therefore rotates with a constant angular velocity
598: %in the $t-\varphi$ plane, and
599: This produces a non-zero component of the angular momentum along the
600: $z$-axis.
601: %\begin{eqnarray}
602: %J & = & \int T_{0\varphi} \ud^3x \; = \; 4\pi \omega N \int_0^\pi
603: %\sin \vartheta \ud \vartheta \int_0^\infty \ud r \;r^2
604: %\;\phi^2(r,\vartheta) \;.
605: %\end{eqnarray}
606: Inserting this ansatz into the field equation (\ref{eq}), the problem
607: reduces to a non-linear PDE for the function $\phi(r,\vartheta)$. Before
608: we start solving this equation, however, it is instructive to consider a
609: simpler system which effectively lives in $2+1$ dimensions. Then the
610: problem reduces to an ordinary differential equation (ODE).
611:
612: Let us pass from spherical coordinates $(t,r,\vartheta,\varphi)$ to
613: polar coordinates $(t,\rho,z,\varphi)$, and then assume that the
614: field does not depend on $z$:
615: \begin{eqnarray}
616: \label{phiaxial}
617: \Phi & = & \phi(\rho) \, e^{i \omega t + i N \varphi} \;.
618: \end{eqnarray}
619: This will correspond to a vortex-type configuration, invariant under
620: translations along the $z$-axis. The energy per unit vortex length is
621: \begin{eqnarray}
622: \label{vortenergy}
623: E & = & 2 \pi \int_0^\infty \rho \, \left(\omega^2 \phi^2 + \phi^{\prime 2}
624: + \frac{N^2} {\rho^2} \, \phi^2 + U(\phi)\right) \, \ud \rho \;.
625: \end{eqnarray}
626: If $N \neq 0$, the vortex will rotate around the $z$-axis, %with the
627: its angular momentum per unit length being given by
628: \begin{eqnarray}
629: \label{vortj}
630: J & = & \int T_{0\varphi} \, \rho \, \ud \rho \, \ud \varphi \; = \; 4
631: \pi \omega N \int_0^\infty \rho \, \phi^2 \, \ud \rho \; \equiv \; N Q
632: \;,
633: \end{eqnarray}
634: where $Q$ is the charge per unit length. The field equation now reads
635: \begin{eqnarray}
636: \label{vorteom}
637: 0 & = & \phi'' + \frac{1} {\rho} \, \phi' - \frac{N^2} {\rho^2} \, \phi
638: - \frac{\ud U(\phi)} {\ud\phi} + \omega^2 \phi \;.
639: \end{eqnarray}
640: For $N=0$ this reduces to (\ref{radeom}), with the replacement $2/r \to
641: 1/\rho$ in the friction term. All arguments above still apply, hence we
642: conclude %and the conclusion is
643: that there exist globally regular `Q-vortex'
644: solutions and their radial excitations with finite energy per unit length.
645: These solutions display the behaviors qualitatively similar to those shown in
646: Fig.~\ref{fig3}.
647:
648: Let us now consider solutions with $N > 0$. As can be seen from
649: (\ref{vortenergy}), the energy for such solutions will be finite if only
650: $\phi(0) = 0$, such that the boundary condition for small $\rho$ is now
651: different. The power series solution to (\ref{vorteom}) in the vicinity of
652: $\rho=0$ reads
653: \begin{eqnarray}
654: \label{vortexpan}
655: \phi & = & B \rho^N + O(\rho^{N+1}) \;,
656: \end{eqnarray}
657: where $B$ is an integration constant. The asymptotic behavior for large
658: $\rho$ is
659: %\footnote{The linearized version of
660: %eqn.(\ref{vorteom}) can be solved exactly and leads to $ \phi(\rho) =
661: %K_{N}(\sqrt{U''(0)-\omega^2} \rho)$ at infinity. This expression has
662: %been used for the numerical solutions.}
663: \begin{eqnarray}
664: \label{vortasympt}
665: \phi & = & \frac{A} {\sqrt{\rho}} \exp \left\{-\sqrt{(U''(0)-\omega^2)}
666: \, \rho \right\} \left(1 + O(1/\rho) \right) \;.
667: \end{eqnarray}
668: One can use essentially the same qualitative considerations as in the
669: preceding section to argue that globally regular solutions to
670: (\ref{vorteom}) with such boundary conditions exist, provided that
671: $\omega$ still fulfills conditions (\ref{cond1}), (\ref{cond2}). These
672: solutions can be obtained by numerically extending the asymptotics
673: (\ref{vortexpan}), (\ref{vortasympt}) to finite values of $\rho$ and
674: adjusting the free parameters $A$, $B$ to fulfill the matching
675: conditions at some $\rho=\rho_0$.
676:
677: \begin{figure}[ht]
678: \begin{minipage}[b]{0.45\linewidth}
679: \centering
680: \includegraphics[width=\linewidth]{rotradballsb}
681: \caption{The spinning $N=1$ fundamental Q-vortex solution and its
682: first two radial excitations; $Q=150$.}
683: \label{fig5}
684: \end{minipage}
685: \hspace{5mm}
686: \begin{minipage}[b]{0.45\linewidth}
687: \centering
688: \includegraphics[width=\linewidth]{vorticesb}
689: \caption{The first three spinning excitations of the fundamental
690: Q-vortex solution; $Q=60$.}
691: \label{fig6}
692: \end{minipage}
693: \end{figure}
694:
695: The conclusion is that there exists a family of globally regular
696: spinning Q-vortex solutions. These solutions can be labeled by
697: $(Q,n,N)$, where $Q$ is the charge per unit vortex length, $n =
698: 0,1,2,\ldots$ is the radial `quantum' number (the number of nodes of
699: $\phi(\rho)$) and $N = 0,1,2,\ldots$ is the rotational `quantum'
700: number\footnote{Spinning Q-vortices with $n=0$ have also been found in
701: Ref.~\cite{Kim:1993mm}.}.
702: For fixed $Q$, the energy per unit length, $E(Q,n,N)$, depends on
703: both $n$ and $N$, while the angular momentum is determined only by the
704: value of $N$ as $J = N Q$. The profiles of $\phi(\rho)$ for several
705: low-lying excitations of the fundamental Q-vortex are shown in
706: Figs.~\ref{fig5}--\ref{fig6}.
707:
708:
709: \section{Spinning Q-balls}
710:
711: Having considered the simpler problem in $2+1$ dimensions, we now return
712: to our main task of finding rotating Q-ball solutions in $3+1$
713: dimensions. Although, these two cases are qualitatively somewhat
714: similar, the $3+1$ dimensional problem is technically more
715: involved, since it requires solving a non-linear PDE. With $\Phi =
716: \phi(r,\vartheta) \; e^{i\omega t+i N\varphi}$ the field equation
717: reduces to
718: \begin{eqnarray}
719: \label{eom3d}
720: \left( \frac{\partial^2} {\partial r^2} + \frac{2} {r} \frac{\partial}
721: {\partial r} + \frac{1} {r^2} \frac{\partial^2} {\partial \vartheta^2}
722: + \frac{\cos\vartheta} {r^2 \sin\vartheta} \frac{\partial} {\partial
723: \vartheta} - \frac{N^2} {r^2 \sin^2\vartheta} + \omega^2 \right) \phi &
724: = & \frac{\ud U(\phi)}{\ud \phi} \;.
725: \end{eqnarray}
726: %With our choice of the potential (\ref{8aa}), the term on the right
727: %contains the qubic and quintic non-linearities.
728: The energy, $E = \int T_{00} \, \ud^3 x$, reads
729: \begin{eqnarray}
730: \label{energy}
731: E & = & 2 \pi \int_0^\infty \ud r \, r^2 \, \int_0^\pi \ud
732: \vartheta \, \sin\vartheta \left(\omega^2 \phi^2 + (\partial_r \phi)^2 +
733: \frac{1} {r^2} (\partial_\vartheta \phi)^2 + \frac{N^2\phi^2} {r^2
734: \sin^2 \vartheta} + U(\phi) \right) \;,
735: \end{eqnarray}
736: the charge
737: \begin{eqnarray}
738: \label{charge3d}
739: Q & = & 2 \omega \int \phi^2 \, \ud^3 x \; = \; 4 \pi \omega
740: \int_0^\infty \ud r \, r^2 \int_0^\pi \ud \vartheta \, \sin \vartheta \,
741: \phi^2 \;,
742: \end{eqnarray}
743: and the angular momentum
744: \begin{eqnarray}
745: \label{angmoment}
746: J & = & \int T_{0\varphi} \, \ud^3 x \; = \; N Q \;.
747: \end{eqnarray}
748: Finiteness of the energy requires that
749: \begin{eqnarray}
750: \label{bound}
751: \phi & \to & 0 \qquad \text{as} \quad r \; \to \; 0 \text{ or } \infty
752: \;.
753: \end{eqnarray}
754: The asymptotic behavior of the solutions in these limits can be easily
755: determined, since for small $\phi$ one has $\ud U/\ud \phi \approx
756: U''(0) \, \phi$, such that equation (\ref{eom3d}) actually becomes
757: linear. The variables then separate, implying that the most general
758: asymptotic solution has the form
759: \begin{eqnarray}
760: \label{sep}
761: \phi(r,\vartheta) & = & \sum_{l=N}^\infty f_{l}(r) \,
762: P^N_l(\cos\vartheta) \;.
763: \end{eqnarray}
764: Here, $P^N_l(\cos\vartheta)$ are the associated Legendre functions. At
765: the origin, the radial amplitudes $f_l(r)$ are
766: \begin{eqnarray}
767: \label{expan}
768: f_l(r) & = & C_l \, r^l + O(r^{l+1}) \;,
769: \end{eqnarray}
770: while at infinity
771: %\footnote{Again the linearized equation can be solved
772: %exactly and leads to $f_{l}(r) = K_{l+1/2}(\sqrt{U''(0)-\omega^2} r) /
773: %\sqrt{r}$.}
774: \begin{eqnarray}
775: \label{asympt}
776: f_l(r) & = & \frac{A_l}{r} \, \exp \left\{-\sqrt{(U''(0)-\omega^2)} \, r
777: \right\} \left( 1 + O(1/r)\right) \;,
778: \end{eqnarray}
779: $C_l$ and $A_l$ denoting integration constants.
780:
781: Eqs.~(\ref{sep})--(\ref{asympt}) have been obtained by linearizing the
782: field equation in the vicinity of $r = 0$ and $r = \infty$, in which
783: case modes with different values of the quantum number $l$ decouple. Our
784: strategy to construct solutions in the whole space is to employ again
785: the partial wave decomposition (\ref{sep}). This is always possible,
786: since the associated Legendre functions form a complete set. However,
787: since for arbitrary values of $r$ the equation is non-linear, harmonics
788: with different values of $l$ will no longer decouple.
789:
790: Since Eq.~(\ref{eom3d}) is symmetric with respect to reflections in
791: the $xy$-plane, $\vartheta \to \pi - \vartheta$, it follows that if
792: $\phi(r,\vartheta)$ is a solution, so is $\phi(r,\pi-\vartheta)$. In
793: addition, $-\phi(r,\vartheta)$ is also a solution, since the field
794: equation contains only odd powers of $\phi$. The associated Legendre
795: functions $P^N_l(\cos\vartheta)$ are even/odd with respect to
796: $\vartheta \to \pi - \vartheta$ for even/odd values of $l+N$,
797: respectively. As a result, half of the terms in the mode expansion
798: (\ref{sep}) will change sign under the reflection, while the other half
799: will stay invariant. Since $\phi(r,\pi-\vartheta)$ must also be a
800: solution, it follows that either all odd or all even terms in the mode
801: expansion must vanish in order to have either $\phi(r,\pi-\vartheta) =
802: \phi(r,\vartheta)$ or $\phi(r,\pi-\vartheta) = -\phi(r,\vartheta)$.
803: The conclusion is
804: that for a given value of $N$ there are two solutions with
805: different parities $P$:
806: \begin{eqnarray}
807: \label{sepeven}
808: P \; = \; +1: \quad \phi(r,\vartheta) & = & \sum_{k=0}^\infty f_{k}(r)
809: \, P^N_{N+2k}(\cos\vartheta) \;, \\
810: \label{sepodd}
811: P \; = \; -1: \quad \phi(r,\vartheta) & = & \sum_{k=0}^\infty f_{k}(r)
812: \, P^N_{N+2k+1}(\cos\vartheta) \;.
813: \end{eqnarray}
814: %one of which is symmetric and the other is antisymmetric.
815:
816: With our choice of the potential (\ref{polynpot}), the field equation
817: (\ref{eom3d}) contains cubic and quintic non-linearities. In view of
818: the completeness of the associated Legendre functions, their products
819: can be expressed in terms of their
820: linear combinations, for example, $\left(\sum f_l P_l^N(x) \right)^5 =
821: \sum a_j P_j^N(x)$, with the coefficients $a_j$ determined by the
822: $f_l$'s.
823: %and similarly for the cubic term, where
824: %$a_l(r)$'s are polynomials of $f_j(r)$'s.
825: As a result, inserting (\ref{sepeven}),~(\ref{sepodd}) into
826: (\ref{eom3d}) we obtain
827: \begin{eqnarray}
828: \label{diffform}
829: \sum_{k}^\infty \CD_k[f_{s}(r)] \, P^N_{N+k}(\cos\vartheta) & = & 0 \;.
830: \end{eqnarray}
831: Here $\CD_k[f_{s}(r)]$ are second order differential operators acting on
832: the radial amplitudes $f_s(r)$, and the sum is taken over all odd/even
833: positive $k$ for odd/even solutions, respectively. This equation is
834: equivalent to the infinite set of ODEs
835: \begin{eqnarray}
836: \label{diffset}
837: \CD_k[f_{s}(r)] & = & 0 \;.
838: \end{eqnarray}
839: Now, we %forcibly
840: truncate this system by setting all amplitudes $f_s$ with $s$ larger
841: than some maximal value $l_{\max}-N$ to zero and by discarding all
842: equations with $k > l_{\max}-N$. The indices in (\ref{diffset}) then
843: vary only in the finite limits
844: \begin{eqnarray}
845: k,s & = & 0,1,\ldots, l_{\max}-N \;.
846: \end{eqnarray}
847: As a result, we end up with a finite system of ODEs. This procedure
848: is sometimes called Galerkin's projection method. It is natural to
849: expect that if $l_{\max}$ is large enough, the resulting approximate
850: solutions will be close to the exact ones. To illustrate that this is
851: indeed the case, we show in Table I the energy and charge of
852: the solution of the truncated system with $\omega^2 = N = P = 1$
853: for several values of $l_{\max}$.
854: \begin{table}[htb]
855: \caption{Parameters of the solitons versus the truncation parameter
856: $l_{\max}$.}
857: \label{tab1}
858: \centering
859: \vglue 0.4 cm
860: \begin{tabular}{|c|*{5}{c|}}
861: \hline
862: $l_{\max}$ & $1$ & $3$ & $5$ & $7$ & $9$ \\
863: \hline
864: $E$ & 230.70 & 217.16 & 207.74 & 205.58 & 205.36 \\
865: $Q$ & 211.46 & 200.68 & 189.43 & 186.99 & 186.73 \\
866: \hline
867: \end{tabular}
868: \end{table}
869: As one can see, the energy and charge indeed approach some limiting
870: values with growing $l_{\max}$. It actually seems to be sufficient
871: %in fact that it suffices
872: to take into account only the first 3--5 lowest harmonics in order to
873: get a reasonable approximation.
874:
875: In the next step we construct solutions for a fixed charge $Q$ in
876: different $N$ sectors. More precisely, the numerical solutions to
877: Eqs.~(\ref{diffset}) have been obtained with \matlab's ODE solvers
878: by utilizing the shooting method. The asymptotic solutions
879: (\ref{expan}),~(\ref{asympt}) were used to start the integration at
880: $r=0.01$ and at $r=10$ toward the matching point, whose position has
881: been varied between $r=3$ and $r=6$. The matching conditions imposed at
882: this point determine the values of the constants $C_l$ and $A_l$ in
883: (\ref{expan}) and (\ref{asympt}). The typical matching error was found
884: to be less then $10^{-16}$. The profiles of even and odd rotating
885: solutions with $N = 1$ are shown in Figs.~\ref{phi1}--\ref{t1odd}.
886: \begin{figure}[ht]
887: \begin{minipage}[t]{0.45\linewidth}
888: \centering
889: \includegraphics[width=\linewidth]{qball1}
890: \caption{$\phi(r,\vartheta)$ for $N = 1$, $P = 1$.}
891: \label{phi1}
892: \end{minipage}
893: \hspace{5mm}
894: \begin{minipage}[t]{0.45\linewidth}
895: \centering
896: \includegraphics[width=\linewidth]{energy1}
897: \caption{$T_{00}$ for $N = 1$, $P = 1$.}
898: \label{t1}
899: \end{minipage}
900: \end{figure}
901: As one can see from these plots, the distribution of the energy density
902: $T_{00}$ is strongly non-spherical. It has the structure of a deformed
903: ellipsoid for $P = 1$, and that of dumbbells oriented along the
904: rotation axis for $P = -1$; in the latter case the energy density vanishes
905: in the equatorial plane. The energies of the first three rotational
906: excitations of the fundamental Q-ball are given in Table~\ref{tsq}.
907: \begin{table}[htb]
908: \caption{Parameters of the rotating solutions with $Q = 410$.}
909: \label{tsq}
910: \centering
911: \vglue 0.4 cm
912: \begin{tabular}{|c|c|c|c|}
913: \hline
914: $N^{P}$ & $l_{\max}$ & $E$ & $\omega$ \\
915: \hline
916: $0$ & 0 & 307.29 & 0.76099 \\
917: \hline
918: $1^{+}$ & 9 & 378.36 & 0.86017 \\
919: \hline
920: $1^{-}$ & 8 & 442.24 & 0.99164 \\
921: \hline
922: $2^{+}$ & 10 & 433.94 & 0.96258 \\
923: \hline
924: $2^{-}$ & 9 & 505.66 & 1.13284 \\
925: \hline
926: $3^{+}$ & 9 & 473.59 & 1.05277 \\
927: \hline
928: $3^{-}$ & 12 & 528.30 & 1.31037 \\
929: \hline
930: \end{tabular}
931: \end{table}
932: As one can see, the energy of the first excitation exceeds the ground
933: state energy by about $20 \%$, the next excitation lying again about $20
934: \%$ above.
935: This is in agreement with the expected properties of spinning
936: excitations of a single soliton. In summary, spinning Q-balls comprise a
937: two-parameter family labeled by the values of $(Q,N)$. We have also
938: found evidence for the existence of spinning radial excitations with $n
939: > 0$. However, the construction of such solutions is somewhat more
940: involved and we refrain from presenting them in this paper.
941: \begin{figure}[ht]
942: \begin{minipage}[t]{0.45\linewidth}
943: \centering
944: \includegraphics[width=\linewidth]{qball1odd}
945: \caption{$\phi(r,\vartheta)$ for the $N = 1$, $P = -1$.}
946: \label{phi1odd}
947: \end{minipage}
948: \hspace{5mm}
949: \begin{minipage}[t]{0.45\linewidth}
950: \centering
951: \includegraphics[width=\linewidth]{energy1odd}
952: \caption{$T_{00}$ for the $N = 1$, $P = -1$. }
953: \label{t1odd}
954: \end{minipage}
955: \end{figure}
956:
957: \vspace{-\baselineskip}
958:
959: \section{Conclusions}
960:
961: The aim of this paper was to find an example of spinning solitons in
962: Minkowski space. We have considered the model of a complex scalar field
963: with a non-renormalizable self-interaction. In the spherically symmetric
964: sector this model contains non-topological solitons, the Q-balls. In
965: addition, we have found an infinite discrete family of radial Q-ball
966: excitations, parameterized by the number of nodes $n$ of the radial field
967: amplitude. Such excited solutions have not been reported in the
968: literature before.
969:
970: In a second step we have analyzed cylindrically symmetric solutions with
971: explicit harmonic dependence on the azimuthal angle, $\exp(i N
972: \varphi)$, which we call spinning Q-vortices. For such solutions there
973: is a non-zero component of the angular momentum along the $z$-axis, $J =
974: N Q$, where $Q$ is the charge per unit vortex length. In addition, these
975: solutions exhibit radial excitations parameterized by an integer $n$. As
976: a result, such spinning solutions comprise a three-parameter family
977: labeled by $(Q,N,n)$.
978:
979: Finally, we have considered the full $3+1$ dimensional problem. We have
980: used a version of the spectral method by expanding the field with
981: respect to the complete set of associated Legendre functions. We reduced
982: the PDE to an infinite system of radial ODEs, and then truncating this
983: system at finite order. %We have found
984: The parameters of the solutions for the truncated system converge
985: rapidly to some limiting values as the truncation parameter grows. By
986: keeping the charge $Q$ fixed, we have obtained the lowest rotational
987: excitations of the fundamental Q-ball solutions. The angular momentum of
988: these solutions is quantized, $J = N Q$, the energy increases (but not
989: very rapidly) with the angular momentum. As a result, these solutions
990: can be viewed as describing \textit{spinning} excitations in the
991: one-soliton sector rather than orbital motion in a many-soliton system%
992: \footnote{One could imagine a situation where in the one-soliton sector
993: there is a soliton plus an orbiting soliton-antisoliton pair. However,
994: the spectrum of $J$ would then probably be continuous, while the energy
995: would probably be considerably higher than that for the single
996: soliton.}. To our knowledge, these spinning Q-balls provide the
997: first explicit example of spinning solitons in Minkowski space in $3+1$
998: dimensions.
999:
1000:
1001: \begin{acknowledgments}
1002:
1003: M.S.V. would like to acknowledge useful discussions with Dieter Maison
1004: during the early stages of this research,
1005: to thank Peter Forgacs for some interesting remarks,
1006: and to thank Fidel Schaposnik for
1007: bringing Refs.~\cite{deVega:1986eu,Jackiw:1990aw} to our attention. We
1008: would also like to thank Andreas Wipf for his help and useful comments,
1009: and also Tom Heinzl for a careful reading of the manuscript.
1010: The work of E.W. was supported by the DFG grant Wi 777/4-3. The work of
1011: M.S.V. was supported by CNRS.
1012:
1013: \end{acknowledgments}
1014:
1015:
1016: \begin{thebibliography}{10}
1017:
1018: \bibitem{Adkins:1983ya}
1019: G.S.~Adkins, C.R.~Nappi and E.~Witten, \emph{Static properties of
1020: nucleons in the Skyrme model}, Nucl. Phys. \textbf{B~228}, 552 (1983).
1021:
1022: \bibitem{Bartnik:1988am}
1023: R.~Bartnik and J.~McKinnon, \emph{Particle-like solutions of the
1024: Einstein-Yang-Mills equations}, Phys. Rev. Lett. \textbf{61}, 141
1025: (1988).
1026:
1027: \bibitem{Battye:2000qj}
1028: R.~Battye and P.~Sutcliffe, \emph{Q-Ball dynamics}, Nucl. Phys.
1029: \textbf{B~590}, 329 (2000).
1030:
1031: \bibitem{Brodbeck:1997ek}
1032: O.~Brodbeck, M.~Heusler, N.~Straumann and M.~S. Volkov, \emph{Rotating
1033: solitons and non-rotating, non-static black holes},
1034: Phys. Rev. Lett. \textbf{79}, 4310 (1997).
1035:
1036: \bibitem{Coleman:1985ki}
1037: S.R.~Coleman, \emph{Q-Balls}, Nucl. Phys. \textbf{B~262}, 263 (1985).
1038:
1039: \bibitem{Davis:1988ip}
1040: R.L.~Davis, \emph{Semitopological solitons}, Phys. Rev. \textbf{D~38},
1041: 3722 (1988).
1042:
1043: \bibitem{Davis:1989ij}
1044: R.L.~Davis and E.P.S.~Shellard, \emph{Cosmic vortons}, Nucl. Phys.
1045: \textbf{B~323}, 209 (1989).
1046:
1047: \bibitem{deVega:1986eu}
1048: H.J.~de~Vega and F.A.~Schaposnik, \emph{Electrically charged vortices
1049: in non-Abelian gauge theories with Chern-Simons term},
1050: Phys. Rev. Lett. \textbf{56}, 2564 (1986).
1051:
1052: \bibitem{Gladikowski:1997mb}
1053: J.~Gladikowski and M.~Hellmund, \emph{Static solitons with non-zero Hopf
1054: number}, Phys. Rev. \textbf{D~56}, 5194 (1997).
1055:
1056: \bibitem{Gisiger:1997vb}
1057: T.~Gisiger and M.B.~Paranjape, \emph{Solitons in a baby-Skyrme model
1058: with invariance under volume/area preserving diffeomorphisms},
1059: Phys. Rev. \textbf{D~55}, 7731 (1997).
1060:
1061:
1062: \bibitem{Heusler:1998ec}
1063: M.~Heusler, N.~Straumann and M.S.~Volkov, \emph{On rotational
1064: excitations and axial deformations of BPS monopoles and Julia-Zee
1065: dyons}, Phys. Rev. \textbf{D~58}, 105021 (1998).
1066:
1067: \bibitem{Jackiw:1990aw}
1068: R.~Jackiw and E.J.~Weinberg, \emph{Selfdual Chern-Simons solitons},
1069: Phys. Rev. Lett. \textbf{64}, 2234 (1990).
1070:
1071: \bibitem{Kim:1993mm}
1072: C.~Kim, S.~Kim and Y.~Kim, \emph{Global nontopological vortices},
1073: Phys. Rev. \textbf{D~47}, 5434 (1993).
1074:
1075: \bibitem{Kleihaus:1999sx}
1076: B.~Kleihaus and J.~Kunz, \emph{A monopole-antimonopole solution of the
1077: SU(2) Yang-Mills-Higgs model}, Phys. Rev. \textbf{D~61}, 025003 (2000).
1078:
1079:
1080: \bibitem{Kleihaus:2002ee}
1081: B.~Kleihaus, J.~Kunz and N.~Francisco,
1082: \emph{Rotating Einstein-Yang-Mills black holes},
1083: http://arXiv.org/abs/gr-qc/0207042.
1084:
1085:
1086: \bibitem{Kusenko:1997zq}
1087: A.~Kusenko, \emph{Solitons in the supersymmetric extensions of the
1088: standard model}, Phys. Lett. \textbf{B~405}, 108 (1997).
1089:
1090: \bibitem{Kusenko:1998si}
1091: A.~Kusenko and M.~Shaposhnikov, \emph{Supersymmetric Q-Balls as dark
1092: matter}, Phys. Lett. \textbf{B~418}, 46 (1998).
1093:
1094: \bibitem{Lee:1992ax}
1095: T.D.~Lee and Y.~Pang, \emph{Nontopological solitons}, Phys. Rept.
1096: \textbf{221}, 251 (1992).
1097:
1098: \bibitem{Maison:1981ze}
1099: D.~Maison, \emph{Uniqueness of the Prasad-Sommerfield monopole
1100: solution}, Nucl. Phys. \textbf{B~182}, 144 (1981).
1101:
1102: \bibitem{Neugebauer}
1103: G.~Neugebauer and R.~Meinel, \emph{General relativistic gravitational
1104: field of a rigidly rotating disk of dust: Solution in terms of
1105: ultraelliptic functions}, Phys. Rev. Lett. \textbf{75}, 3046 (1995).
1106:
1107: \bibitem{Piette:1995mh}
1108: B.M.A.G.~Piette, B.J.~Schroers and W.~Zakrzewski, \emph{Dynamics of baby
1109: skyrmions}, Nucl. Phys. \textbf{B~439}, 205 (1995).
1110:
1111: \bibitem{Schunk96}
1112: F.E.~Schunck and E.W.~Mielke, \emph{Rotating boson stars}, in:
1113: Relativity and Scientific Computing, edited by F.W.~Hehl, R.A.~Puntigam
1114: and H.~Ruder, Springer (1996) 138--151.
1115:
1116: \bibitem{Taubes:1982ie}
1117: C.H.~Taubes, \emph{The existence of a nonminimal solution to the SU(2)
1118: Yang-Mills Higgs equations on $R^3$. Part I}, Commun. Math. Phys.
1119: \textbf{86}, 257 (1982).
1120:
1121: \bibitem{Taubes:1982if}
1122: C.H.~Taubes, \emph{The existence of a nonminimal solution to the SU(2)
1123: Yang-Mills Higgs equations on $R^3$. Part II}, Commun. Math. Phys.
1124: \textbf{86}, 299 (1982).
1125:
1126: \bibitem{VanderBij:2001nm}
1127: J.J.~ Van der Bij and E.~Radu,
1128: \emph{On rotating regular non-Abelian solutions},
1129: Int. J. Mod. Phys. \textbf{A~17}, 1477 (2002).
1130:
1131: \bibitem{Yoshida:1997qf}
1132: S.~Yoshida and Y.~Eriguchi,
1133: \emph{Rotating boson stars in general relativity},
1134: Phys. Rev. \textbf{D~56}, 762 (1997).
1135:
1136:
1137: \end{thebibliography}
1138:
1139: \end{document}
1140: