1: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
2: %\documentclass[12pt,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
3: \usepackage{dcolumn}
4: \usepackage{epsfig}
5:
6:
7: \newcommand{\be}{\begin{equation}}
8: \newcommand{\ee}{\end{equation}}
9: \newcommand{\bea}{\begin{eqnarray}}
10: \newcommand{\eea}{\end{eqnarray}}
11: \newcommand{\ww}{{w}}
12:
13:
14:
15: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
16: \begin{document}
17:
18:
19:
20:
21:
22:
23: \title{Resonant excitations of the 't~Hooft-Polyakov monopole}
24: \author{P\'eter Forg\'acs and Mikhail S. Volkov}
25: \affiliation{Laboratoire de Math\'ematiques et Physique Th\'eorique, CNRS-UMR 6083,
26: Universit\'e de Tours, Parc de Grandmont,
27: 37200 Tours, FRANCE}
28:
29:
30: \begin{abstract}
31: {The spherically symmetric magnetic monopole in an SU(2) gauge theory coupled to
32: a massless Higgs field is shown to possess an infinite number of resonances or
33: quasinormal modes.
34: These modes are eigenfunctions of the isospin 1 perturbation equations
35: with complex eigenvalues, $E_n=\omega_n-i\gamma_n$,
36: satisfying the outgoing radiation condition.
37: For $n\to\infty$, their frequencies $\omega_n$
38: approach the mass of the vector boson, $M_W$,
39: while their lifetimes $1/\gamma_n$ tend to infinity.
40: The response of the monopole to an arbitrary initial perturbation is largely
41: determined by these resonant modes, whose collective effect leads to the formation of
42: a long living breather-like excitation characterized by
43: pulsations with a frequency approaching $M_W$ and with an amplitude decaying
44: at late times as $t^{-5/6}$. }
45: \end{abstract}
46: \pacs{03.65.Ge, 11.27.+d, 14.80.Hv}
47: \maketitle
48:
49:
50: Magnetic monopoles -- magnetically charged finite energy solutions in field theories with
51: spontaneously broken gauge symmetries \cite{mon} -- play an important role in a number
52: of field theoretic considerations.
53: They can account for the quantization of the electric charge, catalyze the decay of proton,
54: might be among the relevant topological defects in the universe,
55: determine supersymmetric vacua, etc. --
56: see \cite{monrev} for reviews.
57:
58:
59: In the present Letter we point out yet another interesting aspect of the monopole
60: which seems to have gone unnoticed so far -- the existence of its quasinormal modes (QNM)
61: or resonant excitations. We show that the
62: Bogomol'nyi-Prasad-Sommerfield (BPS) monopole admits an infinite number of QNM.
63: They manifest themselves as resonance peaks in the low energy
64: scattering cross section of
65: isospin $1$ scalar particles.
66: The QNM can also be described as {\sl complex energy} solutions
67: of the isospin $1$ small fluctuation equations that are
68: regular at the origin and satisfy the {\sl outgoing radiation} condition
69: at spatial infinity. For previous studies of
70: the {\sl real energy} small fluctuations around the
71: monopole we refer to \cite{bais}.
72:
73:
74: We shall demonstrate, in particular, that the QNM of the monopole lead to a {\sl universal
75: late time behavior} of the perturbed monopole by giving rise collectively to
76: a quasiperiodic, long living excitation whose amplitude decays as $t^{-5/6}$
77: at late times ($t$ being the standard Minkowski time).
78: In a recent numerical study of Fodor and R\'acz a breather-like excitation
79: of the monopole
80: retaining a considerable fraction of the energy of the external perturbation
81: has actually been observed \cite{Racz}. It has been one of our aims
82: to explain their observations.
83:
84: It is worth noting the striking analogy of
85: the $t^{-5/6}$ asymptotic behavior of the monopole
86: with the late time evolution of massive fields in
87: black hole spacetimes \cite{5/3}. Black holes also possess
88: the QNM, which have been
89: much studied recently (see e.g. \cite{BH}).
90:
91:
92: We consider a Yang-Mills-Higgs (YMH) theory
93: with gauge group SU(2) defined by the Lagrangian
94: \be \label{1}
95: {\cal L}=-\frac14\,F^a_{\mu\nu}F^{a\mu\nu}+\frac12
96: D_\mu\Phi^a D^{\mu}\Phi^a-
97: \frac{\lambda}{4}(\Phi^a\Phi^a-1)^2.
98: \ee
99: Here
100: $ F^a_{\mu\nu}=\partial_{\mu} A^a_{\nu} - \partial_{\nu} A^a_{\mu}
101: + \varepsilon_{abc}A^b_{\mu} A^c_{\nu}$ is the non-abelian field strength tensor,
102: $D_\mu\Phi^a=\partial_\mu\Phi^a+\varepsilon_{abc}A^b_\mu\Phi^c$
103: denotes the covariant derivative, and in our units
104: the mass of the gauge bosons is equal to one, $M_W=1$, while the mass of the Higgs
105: particle is $\sqrt{2\lambda}\,$.
106: The energy-momentum tensor of the theory defined by Eq.(\ref{1}) is
107: $T^\mu_\nu=-F^a_{\nu\rho}F^{a\mu\rho}
108: +D^\mu\Phi^a D_\nu\Phi^a
109: -\delta^\mu_\nu {\cal L}$.
110:
111: We restrict our analysis to the `minimal' spherically symmetric
112: sector, where the ansatz for the YMH fields is given by $A^a_0=0$,
113: \be \label{3}
114: A^a_i=\varepsilon_{aik}\frac{x^k}{r^2}(1-W(t,r))\,,
115: ~~\Phi^a=\frac{x^a}{r^2}\,H(t,r)\,,
116: \ee
117: where $a,i,k=1,2,3$ and $r^2=x^kx^k$\,.
118: With $\Box\equiv
119: \frac{\partial^2}{\partial t^2 }-\frac{\partial^2}{\partial r^2 }$
120: the YMH equations reduce to
121: \bea \label{5}
122: (r^2\Box
123: +{W}^2+{H}^2-1){W}=0\,, \nonumber \\
124: (r^2\Box+
125: 2{W}^2+\lambda ({H}^2-r^2)){H}=0\,.
126: \eea
127: The static, finite energy solution of these equations
128: is the 't~Hooft-Polyakov monopole \cite{mon}.
129: For the special case, $\lambda=0$, the solution is analytically known -- the BPS monopole,
130: \be \label{6}
131: W(r)=\frac{r}{\sinh r}\,,\quad
132: H(r)=r\coth r-1\,.
133: \ee
134:
135: We shall consider small fluctuations around the static monopole background:
136: $W\to W(r)+w(t,r)$ and $H\to H(r)+\sqrt{2}\,h(t,r)$.
137: Linearizing Eqs.(\ref{5}) with respect to $w$ and $h$, we obtain
138: \bea
139: (r^2\Box+3W^2+H^2-1)w=-q\,{2\sqrt{2}WH}\,h,&& \label{2:1} \\
140: (r^2\Box+{2W^2}+\lambda(3H^2-r^2))h=-
141: q\,2\sqrt{2}WH\,w.&& \label{2:2}
142: \eea
143: Here an auxiliary parameter, $q$, has been introduced,
144: presently $q=1$.
145: In this Letter we concentrate on the case $\lambda=0$,
146: when $W,H$ are given by Eq.(\ref{6}).
147:
148: {\bf Resonant scattering.--}
149: First we shall demonstrate that Eqs.(\ref{2:1}),(\ref{2:2}) describe
150: resonance phenomena indeed.
151: Separating the variables as
152: $w=\Re(e^{-i \omega t}w_\omega(r))$ and
153: $h=\Re(e^{-i \omega t}h_\omega(r))$ with real $\omega$,
154: Eqs.(\ref{2:1}),(\ref{2:2}) become a standard
155: two-channel Schr\"odinger system. As it will be clear from what follows, the
156: frequency spectrum is continuous, $\omega^2\geq~0$.
157: Regular solutions of Eqs.(\ref{2:1}),(\ref{2:2}) have to satisfy the conditions
158: $w_\omega\sim h_\omega\sim r^2$ for $r\to 0$,
159: and they can be normalized for $r\to\infty$ such that
160: \be \label{stat}
161: h_\omega(r)\to \sin(\omega r+\delta(\omega))\,,\quad
162: w_\omega(r)\to C(\omega) r^{\frac{1}{\nu}} e^{-\nu r}\,,
163: \ee
164: where $\nu=\sqrt{1-\omega^2}$. The $h$-field is massless,
165: so it oscillates as $r\to\infty$ for any value of $\omega$.
166: The $w$-field is massive,
167: and for $\omega^2<1$ it shows a bound state type behavior,
168: with exponential decay as $r\to\infty$.
169: The fact that the $w$-field is {\it non-radiative} for $\omega^2<1$ plays the
170: crucial role in our analysis, and below we shall concentrate
171: to this frequency range.
172: Eqs.(\ref{2:1}),(\ref{2:2}) describe in this case the scattering of a massless
173: $h$-radiation on the monopole surrounded by a confined
174: massive $w$-field. We note that for $\omega^2<1$ this is effectively a {\it one-channel}
175: scattering problem, so that the scattering cross section is given by the standard formula
176: $\sigma(\omega)=({4\pi}/{\omega^2})\sin^2(\delta(\omega))$.
177:
178:
179: It is worth noting that the interaction of the $h$-field with the
180: monopole is in fact {\it short range}, so that the cross
181: section is finite. We integrate Eqs.(\ref{2:1}),(\ref{2:2})
182: numerically to obtain $w_\omega(r)$ and $h_\omega(r)$ subject to
183: the boundary conditions (\ref{stat}).
184: The resulting cross section $\sigma(\omega)$
185: shown in Fig.\ref{fig2} exhibits a sequence of resonant peaks
186: accumulating near the value $\omega=1$. This can be so interpreted that for
187: certain energies of the incident $h$-radiation the monopole core
188: gets strongly excited.
189:
190:
191:
192: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
193: \begin{figure}
194: \epsfxsize=9cm
195: \epsfysize=4cm
196: \centerline{\epsffile{SECTION.eps}}
197: \caption{The scattering cross section $\sigma(\omega)$.
198: }
199: \label{fig2}
200: \end{figure}
201: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
202:
203: {\bf QNM -- numerical results.--} The scattering resonances can
204: be usually related to the quasinormal modes --
205: complex energy solutions satisfying the purely outgoing wave condition at $r=\infty$.
206: To construct the QNM, we integrate Eqs.(\ref{2:1}),(\ref{2:2})
207: with $w=\Re(e^{-i E t}w_E(r))$ and $h=\Re(e^{-i Et}h_E(r))$, where
208: $w_E$ and $h_E$ are complex, the energy $E=\omega-i\gamma$, and
209: \be \label{compl}
210: Ar^2\leftarrow h_E(r)\to e^{iEr},~~~ Br^2\leftarrow w_E(r)\to C r^{\frac{1}{\nu}}
211: e^{-\nu r}
212: \ee
213: for $0\leftarrow r\to\infty$.
214: With the `shooting to a fitting point' numerical method we find a discrete
215: family of global solutions $w_E(r)$, $h_E(r)$ subject to the boundary conditions (\ref{compl})
216: labeled by $n=1,2,\ldots $, the
217: number of nodes of $\Im(w_E(r))$ (see Fig.\ref{fig3}).
218: Notice that $h_E\sim e^{i\omega r+\gamma r}$
219: {\sl grows} at infinity -- the QNM are not physical solutions themselves,
220: but only approximate such solutions for a fixed $r$ and for $t\to\infty$.
221: The first
222: $10$ eigenvalues $E_n=\omega_n-i\gamma_n$ are listed in Table~I.
223:
224:
225: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
226: \begin{figure}
227: \epsfxsize=9cm
228: \epsfysize=4cm
229: \centerline{\epsffile{RES.eps}}
230: %{\epsffile{fYM5.eps}}
231: \caption{The complex solutions of Eqs.(\ref{2:1}),(\ref{2:2}) for $n=3$.
232: }
233: \label{fig3}
234: \end{figure}
235: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
236:
237:
238: Table I clearly indicates that $\omega_n\to 1$ and $\gamma_n\to 0$ for growing $n$.
239: It seems that the QNM can be obtained for any $n$, thus
240: comprising an infinite family. The values of
241: $\omega_n$ coincide well with the positions of the
242: resonance peaks shown in Fig.\ref{fig2}.
243:
244:
245:
246: {\bf QNM -- qualitative analysis.--}\
247: The existence of the QNM
248: of the BPS monopole can be qualitatively understood as follows.
249: Let us consider
250: $q$ introduced in Eqs.(\ref{2:1}),(\ref{2:2})
251: as a free parameter, $q\in[0,1]$, and
252: denote the corresponding solutions by
253: $w_{(q)}$ and $h_{(q)}$.
254: For $q=0$ the equations decouple.
255: Setting $h_{(0)}(t,r)= e^{- i\omega t}(C_{+}h_{+}+C_{-}h_{-})$
256: with constant $C_{\pm}$, Eq.(\ref{2:2})
257: is then solved by
258: \be\label{h}
259: h_{\pm}(r)=(\coth r\mp i\omega)e^{\pm i\omega r}.
260: \ee
261: Eq.(\ref{2:1}) with $w_{(0)}(t,r)= e^{-i\omega t}w(r)$
262: reduces to the eigenvalue problem
263: \be \label{bound}
264: \left(-
265: \frac{d^2}{d r^2}+\frac{3W^2+H^2-1}{r^2}\right)w(r)=\omega^2 w(r).
266: \ee
267: The potential in this equation has an attractive Coulombian tail,
268: since it behaves as
269: $1-2/r+O(e^{-r})$ for $r\to\infty$, thus there are
270: infinitely many bound states, $w(r)=w_n(r)$
271: for $\omega^2=\omega^2_n$, $n=1,2,\ldots $.
272: Several low lying $\omega_n$'s are given in Table II.
273: The $n$-th eigenfunction
274: $w_n(r)$ has $n-1$ nodes in the interval $r\in[0,\infty)$ and
275: can be normalized by the condition $\int_0^\infty w_n^2dr=1$.
276: For large $n$ the $w_n(r)$'s extend to the
277: asymptotic region where the potential is Coulombian.
278: As a result, they can be well approximated by solutions of the
279: hydrogen atom problem. This implies that for $n\to\infty$ one has
280: $\omega_n^2=1-1/n^2+O(n^{-3})$ and also that
281: $w_n\sim n^{-3/2}$. More precisely,
282: for $\omega=1$
283: Eq.(\ref{bound}) admits a limiting solution
284: with infinitely many nodes, $w_\infty(r)$, which itself is not normalizable,
285: but $w_\infty(r)$ corresponds to the pointwise limit of $w_n(r)$; i.e.
286: for a {\sl fixed} $\bar{r}$
287: \be \label{large}
288: \lim_{n\to\infty}w_n(r)=n^{-3/2}w_\infty(r)~~~~~~~~~\forall r<\bar{r}\,.
289: \ee
290:
291:
292:
293: Let us consider a solution of Eqs.(\ref{2:1}),(\ref{2:2})
294: given by $h_{(0)}=0$
295: and
296: $w_{(0)}= \Re(A_n e^{-i\omega_n t}w_{n}(r))$,
297: where $A_n$ is a constant.
298: ``Switching on'' a small value of the coupling between the channels,
299: $q\ll 1$, the $w$-bound state
300: will start loosing its energy to the $h$-channel, where
301: this energy will be radiated to infinity.
302: To find approximatively the corresponding solutions $h_{(q)}$
303: and $w_{(q)}$ by successive iterations,
304: we solve first Eq.(\ref{2:2}) for $h_{(q)}$ by replacing $w$ by $w_{(0)}$
305: in its right hand side.
306: The solution is regular at the origin and reduces to an outgoing wave at infinity:
307: $h_{(q)}\sim qA_ne^{i\omega_n (r-t)}$ as $r\to\infty$.
308:
309:
310:
311:
312:
313:
314: Since there is now an outgoing flux of the $h$-radiation,
315: the energy of the $w$-bound state will be slowly decreasing.
316: In the adiabatic approximation this process is described by
317: a decrease of the amplitude
318: of the $\ww$-field by replacing $A_n\to A_n(t)$.
319: To determine $A_n(t)$
320: we use the law of energy
321: conservation, $\partial_\mu T^\mu_0=0$,
322: whose integral form is
323: \be\label{flux}
324: \frac{d}{dt}\left(\int_0^\infty r^2T^0_0 dr \right)=
325: -\lim_{r\to\infty}r^2 T^r_0\, .
326: \ee
327: Expanding $T^\mu_\nu$ up to terms
328: quadratic in $w$ and $h$,
329: the expression on the left in Eq.(\ref{flux}) is proportional to $\frac{d}{dt}A_n^2$ and
330: determines the decrease of the bound state energy.
331: The expression on the right is proportional to $A_n^2$ and gives the
332: energy flux at infinity.
333: Thus
334: $\dot{A_n}=-\gamma_n A_n$, from which $A_n(t)=c_ne^{-\gamma_n t}$ with a
335: constant $c_n$. The coefficient $\gamma_n$ is given by
336: \be\label{gamma}
337: \gamma_n=q^2\frac{4}{\omega_n^2(1+\omega_n^2)}
338: {\left(\int_0^\infty \frac{h_{+}-h_{-}}{2ir^2}\,WH w_n dr \right)^2 }.
339: \ee
340: Summarizing, upon switching on a small value of $q$,
341: the stationary bound state of the $w$-field, $w_n(r)$, becomes quasistationary
342: and is approximately given by
343: \be\label{approx}
344: {w}_{(q)}= \Re(e^{-i E_n t}c_n w_n)\,,\quad
345: {h}_{(q)}= q\,\Re(e^{-i E_n t}c_n h_n)\,,
346: \ee
347: where $E_n=\omega_n-i\gamma_n$, and
348: $h_n(r)$ can be expressed by quadratures in terms of $w_n(r)$.
349: Note the appearance of a complex energy, $E_n$, in
350: Eq.(\ref{approx}).
351: Evaluating the integral in Eq.(\ref{gamma}) (see Table II)
352: shows that $\gamma_n$ {decreases} as $n$ grows.
353: This feature
354: can be understood qualitatively by viewing the QNM as quasibound states of the
355: massive $w$-field slowly decaying via an energy transfer to the
356: $h$ channel. For large $n$ the coupling between the $w$ and $h$
357: channels becomes weaker and weaker, since it is proportional to
358: $WH w_n\sim n^{-3/2}$, and so the decay becomes more
359: and more unlikely. Using the asymptotic relation (\ref{large})
360: in Eq.(\ref{gamma}) yields $\gamma_n\sim n^{-3}$ for $n\to\infty$.
361: \begin{table}
362: \caption{The eigenvalues of the first ten QNM}
363: \begin{ruledtabular}
364: \begin{tabular}{cccc}
365: n & $\omega_n-i\gamma_n$ & $n$ & $\omega_n-i\gamma_n$ \\ \hline
366: 1 & $0.8473-i\, 0.5077\times 10^{-1}$ & 6 & $0.9888-i\, 0.8396\times 10^{-3}$ \\
367: 2 & $0.9332-i\, 0.1384\times 10^{-1}$ & 7 & $0.9915-i\,0.5499\times 10^{-3}$ \\
368: 3 & $0.9637-i\,0.5218\times 10^{-2}$ & 8 & $0.9933-i\,0.3794\times 10^{-3} $ \\
369: 4 & $0.9774-i\,0.2488\times 10^{-2} $ & 9 & $0.9946-i\,0.2731\times 10^{-3}$ \\
370: 5 & $0.9846-i\, 0.1375\times 10^{-2}$ &10 & $0.9956-i\,0.2030\times 10^{-3} $ \\
371: \end{tabular}
372: \end{ruledtabular}
373: \end{table}
374:
375: The approximative formulas derived above
376: under the assumption $q\ll 1$ give in fact, when straightforwardly extrapolated to $q=1$,
377: a good approximation for $\omega_n-i\gamma_n$ (see Tables I, II).
378: This approximation is actually getting better with growing $n$.
379: We can therefore use the asymptotic relations derived above
380: to obtain for $n\gg 1$
381: \be \label{llarge}
382: 1-\omega_n^2=n^{-2}+O(n^{-3})\,,\quad \gamma_n=bn^{-3}.
383: \ee
384: Using the results of Table I, $b\approx 0.2$.
385:
386: \begin{table}
387: \label{tab1}
388: \caption{Values of $\omega_n$ and $\gamma_n$ defined by
389: Eqs.(\ref{bound}),(\ref{gamma})}
390: \begin{ruledtabular}
391: \begin{tabular}{cccccc}
392: n & 1 & 2 & 3 & 5 & 10 \\ \hline
393: $\omega_n$ & $0.798$ & $0.926$ & $0.961$ & $0.984$ & $0.995$ \\
394: $\gamma_n/q^2$ & $0.057$ & $0.010$ & $0.0035$ & $0.0009$ & $0.0001$ \\
395: \end{tabular}
396: \end{ruledtabular}
397: \end{table}
398: {\bf Collective effect of the QNM.--}
399: The numerical simulations in Ref.\cite{Racz} of the non-linear temporal dynamics of the
400: BPS monopole hit by a strong spherically symmetric pulse
401: have shown that a considerable fraction of the energy received from the pulse is
402: not radiated away immediately. It gets `trapped' by the
403: monopole and forms a long living quasiperiodic excitation that radiates
404: very slowly, and whose late time behavior is to a large extent
405: independent of the structure
406: of the initial perturbation pulse, i.e.,\ it shows certain type of {\sl universality}.
407: We can now offer an explanation to these observations.
408:
409: It is intuitively clear that, according to the decomposition of the eigenfunctions of the
410: linearized problem, there is a `radiative' sector containing the massless $h$-modes
411: with $\omega^2>0$ and massive $w$-modes with $\omega^2>1$, and also
412: a `non-radiative' sector consisting of the $w$-modes with $\omega^2<1$.
413: One expects that a part of the energy of the pulse received by the monopole
414: will be distributed among the radiative modes and will be
415: radiated away. However,
416: as a generic perturbation will have a nonzero overlap also with the non-radiative
417: modes, the remaining energy will get trapped in the non-radiative sector.
418: Since it cannot be radiated away directly, a
419: long living excitation on the monopole forms that
420: will decay only due to a slow energy leakage to the radiative
421: channels. In the terminology of black hole physics,
422: the perturbed monopole keeps some of its `hair' for a long time.
423:
424: Although initially the dynamic of the system will be nonlinear,
425: one expects the nonlinear effects to become negligible
426: when a sufficient amount of the received energy is radiated away.
427: The linear approximation will therefore give a good description of the late time behavior,
428: providing also an explanation of the observed {\sl universality}.
429:
430: Let $t=0$ be the starting point of linear regime, when
431: the perturbed monopole
432: is described by $w(0,r)=\delta W(r)$ and $h(0,r)=\delta H(r)$.
433: The subsequent temporal evolution is determined to a large extent
434: by the QNM, since they
435: hold their energies for a long time. Using (\ref{approx}),
436: we therefore approximate the general solution for $t>0$ by
437: \be \label{late}
438: w(t,r)=\Re\left(\sum_{n=1}^{\infty}c_n e^{-i\omega_n t-\gamma_n t}w_n(r)\right),
439: \ee
440: and similarly for $h(t,r)$. Here
441: $c_n=\int_0^\infty w_n(r)\,\delta W(r) dr$.
442: This sum of damped oscillations
443: corresponds to a long living excitation with the following properties.
444:
445:
446: For a localized $\delta W$ the overlap coefficients $c_n$ will be maximal
447: for small $n$. Therefore terms with small $n$ will dominate at first the sum in Eq.(\ref{late}).
448: They will soon be damped, however, since their damping rates, $\gamma_n$,
449: are the largest.
450: Terms with higher $n$ will then become more important at later times.
451: This `dying out' of modes has indeed been observed in Ref.\cite{Racz} (Fig.6),
452: and the values of frequencies $\omega_n$ measured there are
453: in good agreement with those given in Table~I.
454:
455: For any $t$
456: there is a number, $k(t)$, determined by the condition
457: $\gamma_k t\sim 1$, such that all terms with $n\ll k$ in (\ref{late}) are
458: already damped, while those with
459: $n\gg k$ are not yet important, since their $c_n$'s are small compared to $c_k$.
460: The sum will
461: therefore be dominated by terms with $n\sim k(t)$.
462: Considering for simplicity the sum of only two of these terms,
463: with frequencies $\omega_k$, $\omega_{k+1}$, gives beats with the base frequency
464: $\omega(t)=\frac12(\omega_{k+1}+\omega_{k})$ whose amplitude is
465: modulated with the frequency $\Omega(t)=\frac12(\omega_{k+1}-\omega_{k})$.
466: This explains qualitatively the behavior shown in Fig.\ref{fig4}.
467: Since $\gamma_k$ decreases with $k$, the value of $k(t)$
468: will grow with $t$, and so will do the base frequency $\omega(t)$.
469: Since $\gamma_k\sim k^{-3}$
470: for large $k$, it follows that $k(t)\sim t^{1/3}$ for large $t$.
471: Using (\ref{llarge}), we conclude that for large $t$
472: one has $1-\omega^2(t)\sim t^{-2/3}$, which explains the feature observed
473: in Ref.\cite{Racz} (Fig.5).
474: In a similar way we obtain for the modulation frequency
475: $\Omega(t)\sim t^{-1}$.
476:
477: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
478: \begin{figure}
479: \epsfxsize=9cm
480: \epsfysize=4cm
481: \centerline{\epsffile{BREATH.eps}}
482: \caption{The profile of $w(t,1)$ in Eq.(\ref{late}) corresponding to
483: the initial step-like perturbation, $\delta W(r)=\theta(1-r)$.
484: }
485: \label{fig4}
486: \end{figure}
487: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
488: Even though each term in the sum (\ref{late}) tends to zero exponentially
489: fast, the sum as a whole decreases considerably slower.
490: Since for large $t$ only terms with large $n$ are relevant in (\ref{late}),
491: we can use Eqs.(\ref{large}),(\ref{llarge}) to obtain
492: $\omega_n\approx 1-{1}/{2n^2}$ and $w_n(r)\approx n^{-3/2}w_\infty(r)$
493: for $r\leq \bar{r}=\int_0^\infty rw_n^2dr\sim n^2$. For a localized $\delta W$
494: this implies that
495: $c_n\approx {\cal N}n^{-3/2}$ with ${\cal N}=\int_0^\infty w_\infty\delta Wdr$.
496: As a result, for $t\to\infty$ and for $r<\bar{r}(t)\sim n(t)^2\sim t^{2/3}$
497: the solution (\ref{late}) reduces to $w(t,r)={\cal N}w_\infty(r)\Re(G(t))$, with
498: \be \label{final}
499: G(t)=e^{-it}\sum_{n> t^{1/3}}^{\infty}\frac{1}{n^3}
500: \exp\left(\frac{it}{2n^2}-\frac{bt}{n^3} \right).
501: \ee
502: This shows that the late time dynamics is indeed {\sl universal}, since
503: changing the initial conditions affects only the
504: normalization ${\cal N}$.
505: The final task is to determine the asymptotic behavior
506: of $G(t)$. Transforming the sum (\ref{final}) to a contour integral
507: and using the saddle point method, we find that $|G(t)|\sim t^{-5/6}$ for $t\to\infty$.
508: This $t^{-5/6}$ exponent
509: explains the feature observed in Ref.\cite{Racz} (Fig.5).
510:
511: The perturbed monopole thus ends up
512: in a long living breathing state
513: dominated by the confined, slowly radiating massive modes of the gauge field.
514: In the region $r\leq t^{2/3}$
515: this state is characterized by modulated pulsations
516: whose base frequency approaches the vector boson mass, while the
517: amplitude decreases as $t^{-5/6}$.
518:
519:
520: The total energy of the breather can be obtained by summing over the QNM,
521: $E=\sum_n c_n^2\omega_n^2 e^{-2\gamma_n t}
522: \sim
523: \sum_n n^{-3}\exp(-2bn^{-3}t)\sim
524: t^{-2/3}$.
525: This includes the energy of the confined massive $w$-modes and also that of
526: the massless $h$-radiation emitted by these modes.
527: $E$ decreases, since there is a flux of the $h$-radiation at infinity,
528: $S=\dot{E}\sim t^{-5/3}$. This value agrees with the result of \cite{Racz}.
529: Specifically, Fig.2 in \cite{Racz} shows
530: the flux $S\sim T^{-5/3}$ at future null infinity, with
531: $T\sim t-r$.
532: By continuity, the flux
533: through a 2-sphere of a very large but finite radius $r$ is still $S\sim T^{-5/3}$.
534: But since $t$ is finite then, for $t\gg r$ one has $T\sim t$, which agrees with out result
535: $S\sim t^{-5/3}$.
536:
537:
538:
539:
540:
541: In conclusion, we have found resonant excitations of the monopole
542: and studied some of their applications. Although we have considered above only
543: the BPS monopole, the existence of similar resonances can also be shown
544: in the case
545: of a non-zero Higgs self-coupling $\lambda$.
546: %, as well as in the {\sl odd parity} perturbation sector.
547:
548:
549:
550:
551: {\bf Acknowledgments.--}
552: We would like to thank Gyula Fodor and Istv\'an R\'acz
553: for communicating to us their results prior to publication and
554: stimulating discussions.
555:
556:
557:
558: \begin{thebibliography}{10}
559:
560:
561:
562:
563:
564:
565: \bibitem{mon}
566: G.~'t Hooft, {\it Nucl.Phys.}, {\bf B 79} (1974) 276;
567: A.M.~Polyakov, {\it JETP.Lett.}, {\bf 20} (1974) 430;
568: M.K.~Prasad and C.M.~Sommerfield, {\it Phys.Rev.Lett.}, {\bf 35} (1975) 760;
569: E.B.~Bogomol'nyi, {\it Sov.J.Nucl.Phys.}, {\bf 24} (1976) 861.
570:
571:
572:
573: \bibitem{monrev}
574: P.~Goddard and D.I.~Olive, {\it Rep.Progr.Phys.}, {\bf 41} (1978) 1357;
575: J.~Preskill, {\it Ann.Rev.Nucl.Part.Sci.}, {\bf 34} (1984) 461;
576: M.~Duff, R.~Khuri, and J.~Lu, {\it Phys.Rep.}, {\bf 259} (1995) 213.
577:
578: \bibitem{bais}
579: F.A.~Bais and W.~Troost, {\it Nucl.Phys.}, {\bf B 178} (1981) 125; \\
580: H.~Sonoda, {\it Phys.Lett.}, {\bf B 143} (1984) 142;
581: J.~Baake, {\it Z.Phys.}, {\bf C 53} (1992) 399.
582:
583:
584: \bibitem{Racz}
585: G.~Fodor and I.~R\'acz, {\tt hep-th/0311061}.
586:
587:
588:
589: \bibitem{5/3}
590: H.~Koyama and A.~Tomimatsu, {\it Phys.Rev.}, {\bf D 65} (2002) 2084031;
591: F.~Finster, N.~Kamran, J.~Smoller, and S.T.~Yau, {\it Commun.Math.Phys.},
592: {\bf 230} (2002) 201.
593:
594: \bibitem{BH}
595: K.~Kokkotas, B.~Schmidt, {\it Living.Rev.Rel.}, {\bf 2} (1999)~2;
596: L.~Motl, {\it Adv.Theor.Math.Phys.}, {\bf 6} (2003) 1135.
597:
598:
599:
600: \end{thebibliography}
601:
602:
603:
604: \end{document}
605: