1: \documentclass[12pt]{revtex4}
2: \usepackage{amsmath,amssymb,graphicx}
3:
4:
5:
6: \def\al{\alpha}
7: \def\be{\beta}
8: \def\ga{\gamma}
9: \def\de{\delta}
10: \def\ep{\epsilon}
11: \def\ve{\varepsilon}
12: \def\ze{\zeta}
13: \def\et{\eta}
14: \def\th{\theta}
15: \def\vt{\vartheta}
16: \def\io{\iota}
17: \def\ka{\kappa}
18: \def\la{\lambda}
19: \def\vpi{\varpi}
20: \def\rh{\rho}
21: \def\vr{\varrho}
22: \def\si{\sigma}
23: \def\vs{\varsigma}
24: \def\ta{\tau}
25: \def\up{\upsilon}
26: \def\ph{\phi}
27: \def\vp{\varphi}
28: \def\ch{\chi}
29: \def\ps{\psi}
30: \def\om{\omega}
31: \def\Ga{\Gamma}
32: \def\De{\Delta}
33: \def\Th{\Theta}
34: \def\La{\Lambda}
35: \def\Si{\Sigma}
36: \def\Up{\Upsilon}
37: \def\Ph{\Phi}
38: \def\Ps{\Psi}
39: \def\Om{\Omega}
40: \def\mn{{\mu\nu}}
41: \def\pt#1{\phantom{#1}}
42: \def\prt{\partial}
43: \def\cl{{\mathcal L}}
44: \def\vev#1{\langle {#1}\rangle}
45: \def\expect#1{\langle{#1}\rangle}
46: \def\bra#1{\langle{#1}|}
47: \def\ket#1{|{#1}\rangle}
48: \def\fr#1#2{{{#1} \over {#2}}}
49: \def\Frac#1#2{{\textstyle{{#1}\over {#2}}}}
50: \def\half{{\textstyle{1\over 2}}}
51: \def\lsim{\mathrel{\rlap{\lower4pt\hbox{\hskip1pt$\sim$}}
52: \raise1pt\hbox{$<$}}}
53: \def\gsim{\mathrel{\rlap{\lower4pt\hbox{\hskip1pt$\sim$}}
54: \raise1pt\hbox{$>$}}}
55: \def\sqr#1#2{{\vcenter{\vbox{\hrule height.#2pt
56: \hbox{\vrule width.#2pt height#1pt \kern#1pt
57: \vrule width.#2pt}
58: \hrule height.#2pt}}}}
59: %\def\square{\mathchoice\sqr66\sqr66\sqr{2.1}3\sqr{1.5}3}
60: \newcommand{\beq}{\begin{equation}}
61: \newcommand{\eeq}{\end{equation}}
62: \newcommand{\bea}{\begin{eqnarray}}
63: \newcommand{\eea}{\end{eqnarray}}
64: \newcommand{\rf}[1]{(\ref{#1})}
65:
66: \def\kf{\hat k_F}
67: \def\kaf{\hat k_{AF}}
68: \def\kfd#1{k_{F}^{(#1)}}
69: \def\kafd#1{k_{AF}^{(#1)}}
70: \def\klm#1#2#3{k^{(#1)}_{(#2)#3}}
71: \def\sylm#1#2{\phantom{}_{#1}Y_{#2}}
72: \def\mbf#1{\mbox{\boldmath$#1$}}
73:
74: \def\kI{\klm{d}{I}{lm}}
75: \def\kE{\klm{d}{E}{lm}}
76: \def\kB{\klm{d}{B}{lm}}
77: \def\kV{\klm{d}{V}{lm}}
78: \def\kVdlm#1#2{\klm{#1}{V}{#2}}
79:
80: \begin{document}
81:
82: \title{Lorentz Violation, Electrodynamics,
83: and the Cosmic Microwave Background\footnote{
84: Presented at the Fourth Meeting on CPT and Lorentz
85: Symmetry, Bloomington, Indiana, August, 2007.}}
86:
87: \author{Matthew Mewes}
88: \address{Physics Department, Marquette University, Milwaukee, WI 53201}
89:
90: \begin{abstract}
91: Vacuum birefringence is a signature of
92: Lorentz-symmetry violation.
93: Here we report on a recent search for
94: birefringence in the cosmic
95: microwave background.
96: Polarization data is used
97: to place constraints on
98: certain forms of Lorentz violation.
99: \end{abstract}
100:
101: \maketitle
102:
103: \section{Introduction}
104:
105: The properties of light have proved
106: to be a valuable testing ground for
107: special relativity for more than a century.
108: Contemporary experiments are motivated in part
109: by a possible breakdown of special relativity
110: with origins in Planck-scale physics
111: \cite{kost,cpt04,ck}.
112: These experiments include modern versions of the
113: classic Michelson-Morley and
114: Kennedy-Thorndike experiments
115: that use highly stable resonant cavities
116: to search for violations of
117: rotation and boost symmetries
118: \cite{cavities}.
119: However, the highest sensitivities
120: to relativity violations
121: in electrodynamics are found in
122: searches for vacuum birefringence in
123: light from very distant sources
124: \cite{cfj,km1,km2,km3}.
125: Birefringence studies take advantage of the
126: extremely long baselines that allow the miniscule
127: effects of a Lorentz violation
128: to accumulate to (potentially)
129: detectable levels over the billions of years
130: it takes for the light to reach Earth.
131: The cosmic microwave background (CMB)
132: is the oldest light available to observation
133: and therefore provides an excellent
134: source for birefringence searches.
135: Here we summarize a recent search for
136: signals of Lorentz violation
137: using CMB polarimetry
138: \cite{km3}.
139:
140: General Lorentz violation is
141: described by a framework known
142: as the Standard Model Extension (SME)
143: \cite{ck}.
144: The SME provides the theoretical
145: backbone for studies in a number of areas
146: \cite{cpt04},
147: including photons
148: \cite{cavities,km1,km2,km3,photons}.
149: Most tests of Lorentz violation
150: focus on the minimal SME,
151: which assumes usual gauge invariance
152: and energy-momentum conservation
153: and restricts attention to superficially
154: renormalizable operators.
155: Operators of dimension $d\leq4$
156: are of renormalizable dimension
157: and are included the minimal SME.
158: Two types of Lorentz-violating operators
159: appear in the minimal SME,
160: $CPT$-odd operators with coefficients
161: $(k_{AF})_\ka$ and $CPT$-even operators
162: with coefficients $(k_F)^{\ka\la\mu\nu}$.
163:
164: In this work,
165: we also consider non-minimal
166: higher-dimensional operators
167: in the photon sector with $d>4$.
168: In general there are an infinite
169: number of possible operators
170: that emerge when we relax the
171: renormalizable condition.
172: These operators are
173: phenomenologically and theoretically
174: relevant in that they help
175: provide a connection to the
176: underlying Planck-scale physics.
177: They also add a number of new and
178: interesting signals for Lorentz
179: violation that may be tested experimentally.
180:
181: \section{Theory}
182:
183: General Lorentz-violating
184: electrodynamics is given
185: by a lagrangian that takes the
186: same basic form as the minimal-SME
187: photon sector
188: \cite{km3}:
189: \beq
190: \cl = -\fr14 F_\mn F^\mn
191: + \fr12\ep^{\ka\la\mu\nu} A_\la (\kaf)_\ka F^{\mu\nu}
192: -\fr14 F_{\ka\la} (\kf)^{\ka\la\mu\nu} F_{\mu\nu} \ .
193: \eeq
194: We assume a linear theory
195: and impose the usual U(1) gauge invariance.
196: The key difference between this theory
197: and the minimal-SME photon sector
198: is that here the $\kaf$ and $\kf$ coefficients
199: are differential operators.
200: The effects of these operators
201: mimic the effects of a permeable medium
202: whose activity depends on the
203: photon energy and momentum.
204: This introduces a plethora of new effects not
205: found in either the conventional
206: Lorentz-conserving case or the minimal SME.
207: These include drastically different frequency
208: dependences and direction-dependent propagation
209: of light.
210:
211: Expanding the $\kaf$ and $\kf$ operators
212: in the 4-momentum $p_\mu = i\prt_\mu$
213: leads to the expressions
214: \begin{align}
215: (\kaf)_\ka&=\sum {(\kafd{d})_\ka}^{\al_1\ldots\al_{(d-3)}}
216: \prt_{\al_1}\ldots\prt_{\al_{(d-3)}} \ , \\
217: \allowbreak
218: (\kf)^{\ka\la\mu\nu}&=\sum (\kfd{d})^{\ka\la\mu\nu\al_1\ldots\al_{(d-4)}}
219: \prt_{\al_1}\ldots\prt_{\al_{(d-4)}} \ .
220: \end{align}
221: The coefficients for Lorentz violation
222: associated with the dimension $d$ operators
223: are now given by
224: ${(\kafd{d})_\ka}^{\al_1\ldots\al_{(d-3)}}$ and
225: $(\kfd{d})^{\ka\la\mu\nu\al_1\ldots\al_{(d-4)}}$.
226: The $\kaf$ expression contains all
227: $CPT$-breaking effects, and the sum
228: is restricted to odd-dimensional operators,
229: $d=$ odd.
230: The $\kf$ coefficients control all $CPT$-even
231: violations and have $d=$ even.
232: Imposing gauge invariance places
233: various constraints on these coefficients.
234: For $\kafd{d}$ coefficients, any trace of the
235: that involves the first index vanishes identically.
236: For $\kfd{d}$, the antisymmetrization on
237: any three indices vanishes.
238: Standard group theory techniques allow a
239: counting of the independent coefficients
240: for Lorentz violations
241: \cite{km4}.
242: For a given dimension $d$,
243: we find $\half(d+1)(d-1)(d-2)$
244: independent $\kafd{d}$ coefficients
245: in the $CPT$-odd case and $(d+1)d(d-3)$
246: independent $\kfd{d}$ coefficients
247: in the $CPT$-even case.
248:
249: For studies of Lorentz-violation
250: induced birefringence,
251: certain linear combinations of these
252: general coefficients are important.
253: They result from a spherical-harmonic
254: expansion of plane-waves propagating
255: in the vacuum.
256: This plane-wave expansion is
257: best characterized using the language
258: of Stokes parameters.
259: We begin by defining
260: a Stokes vector $\mbf s = (s^1,s^2,s^3)^T$.
261: The direction in which this vector
262: points in the abstract 3-dimensional
263: Stokes space uniquely characterizes
264: the polarization of the radiation.
265: Stokes vectors lying in the $s^1$-$s^2$
266: plane correspond to all possible linear
267: polarizations,
268: while Stokes vectors parallel and
269: antiparallel to the $s^3$ axis
270: give the two circular polarizations.
271: General right-handed elliptical
272: polarizations point in the upper-half Stokes space, $s^3>0$,
273: while left-handed are given by the lower half, $s^3<0$.
274:
275: This formalism provides an intuitive
276: picture of birefringence.
277: It can be shown that birefringence causes a
278: rotation of the Stokes vector $\mbf s$
279: about some axis $\mbf \vs$.
280: This occurs whenever the usual
281: degeneracy between the various
282: polarizations in broken.
283: Formally,
284: we solve the modified equations of motion.
285: We find that some types of violations
286: lead to two propagating eigenmodes
287: that have slightly different velocities.
288: They also differ in polarization,
289: and light of an arbitrary polarization
290: is a superposition of the two eigenmodes.
291: This superposition is altered
292: as the eigenmodes propagate
293: at different velocities,
294: causing an oscillatory effect
295: that reveals itself as a rotation
296: of the Stokes vector.
297: The rotation takes the form
298: \beq
299: d\mbf s/dt = 2\om \mbf \vs\times\mbf s \ ,
300: \label{rot}
301: \eeq
302: where $\om$ is the wave frequency,
303: and the rotation axis $\mbf\vs$
304: corresponds to the Stokes vector
305: of the faster eigenmode.
306: In general, $\mbf\vs$
307: may depend on both the direction
308: of propagation and the frequency.
309:
310: The basic idea of a birefringence
311: test is to examine light from a
312: distant polarized source
313: for the above rotation.
314: To do this we need to express the
315: rotation axis $\mbf \vs$ in terms
316: of the coefficients for Lorentz violation.
317: The general result is rather complicated,
318: but can be written in a relatively simple
319: form in terms of a set of ``vacuum'' coefficients,
320: which are linear combinations of the general coefficients.
321: The calculation involves decomposing $\mbf\vs$
322: into spin-weighted spherical harmonics.
323: The result takes the form
324: \begin{align}
325: \vs^1\mp i\vs^2 &= \sum_{dlm} \om^{d-4}
326: (\kE \pm i \kB)\, \sylm{\pm2}{lm}(\mbf{\hat n})\ , \\
327: \vs^3 &= \sum_{dlm} \om^{d-4} \kV\, \sylm{0}{lm}(\mbf{\hat n})\ ,
328: \end{align}
329: where $\sylm{s}{lm}$ is a spin-weighted
330: spherical harmonic with spin-weight $s$,
331: and $\mbf{\hat n}$ is the radial unit
332: vector pointing toward the source on the sky.
333: The vacuum coefficients
334: $\kV$, $\kE$, and $\kB$ represent
335: the minimal combinations
336: of coefficients for Lorentz violation
337: that cause birefringence and affect polarization.
338: The designations $E$ and $B$ refer to
339: the parity of the coefficient and
340: is borrowed from radiation theory.
341: In the next sections,
342: we describe a search for these effects
343: in existing CMB polarization data.
344:
345:
346: \section{CMB}
347:
348: The CMB is conventionally
349: parameterized by a spin-weighted
350: spherical-harmonic expansion
351: similar to the expansion of $\mbf\vs$ given above
352: \cite{cmbrev,polrev}.
353: The complete characterization
354: of radiation from a given point
355: on the sky includes the temperature $T$,
356: the linear polarization, given by
357: Stokes parameters $s^1$ and $s^2$,
358: and the circular polarization, given by $s^3$.
359: The global description is
360: given by the expansion
361: \begin{gather}
362: T=\sum a_{(T)lm}\, \sylm{0}{lm}(\mbf{\hat n})\ ,
363: \quad\quad
364: s^3=\sum a_{(V)lm}\, \sylm{0}{lm}(\mbf{\hat n})\ , \notag \\
365: s^1\mp is^2 = \sum (a_{(E)lm} \pm i a_{(B)lm})\, \sylm{\pm2}{lm}(\mbf{\hat n})\ .
366: \end{gather}
367: One then constructs various power spectra,
368: \beq
369: C^{X_1X_2}_l = \Frac1{2l+1} \sum_m \vev{a^*_{(X_1)lm}a_{(X_2)lm}}\ ,
370: \eeq
371: where $X_1,X_2=T,E,B,V$.
372: These spectra quantify the
373: angular size variations in
374: each mode and any correlation
375: between different modes.
376: Smaller $l$ correlates to larger
377: angular size on the sky.
378:
379: Within conventional physics,
380: we expect a nearly isotropic ($l=0$)
381: temperature distribution.
382: However, tiny fluctuations in
383: temperature during recombination
384: not only introduce higher-order
385: multipole moments ($l>0$)
386: but also provide
387: the necessary anisotropies to
388: produce a net polarization.
389: Only linear polarizations are
390: expected since no circular polarization
391: is produced in Thomson scattering.
392: Furthermore, $E$-type polarization
393: is expected to dominate and be correlated
394: with the temperature.
395: No correlation is expected between the
396: much smaller $B$ polarization
397: and temperature.
398: This general picture agrees with observation
399: to the extent to which CMB radiation
400: has been measured
401: \cite{cmbexp,boompol}.
402:
403: A breakdown of Lorentz symmetry
404: may alter these basic features.
405: Some of the new effects
406: can be readily understood
407: as consequences of the Stokes rotations.
408: For example,
409: the $CPT$-odd coefficients $\kV$
410: lead to a Stokes rotation axis that
411: points along the $s^3$ direction.
412: The resulting local rotations
413: in polarization leave the circularly
414: polarized component unchanged.
415: However,
416: it does lead to a rotation
417: in the linear components,
418: causing a simple change in the
419: polarization angle at each point on
420: the sky.
421: Globally this causes a mixing between
422: the $E$ and $B$ polarization.
423: This could introduce an unusually
424: large $B$ component,
425: which gives a potential signal
426: of $CPT$ and Lorentz violation.
427: Similar mixing can arise from
428: the $\kE$ and $\kB$ coefficients.
429: However,
430: since these give a rotation
431: axis that lies in the $s^1$-$s^2$ plane,
432: the rotations in this case
433: also introduce circular polarization.
434: So a large circularly polarized
435: component in the CMB might indicate a $CPT$-even
436: violation of Lorentz invariance.
437:
438: All except the $d=3$ coefficients
439: result in frequency-dependent rotations.
440: Also, only the $l=0$ coefficients cause
441: isotropic rotations that are uniform
442: across the sky.
443: As a result,
444: the coefficient $\kVdlm{3}{00}$
445: provides a simple isotropic
446: frequency-independent special case.
447: A calculation shows that this
448: case causes a straightforward
449: rotation between
450: $C_l^{EE}$, $C_l^{BB}$, and $C_l^{EB}$,
451: as well as between
452: $C_l^{TE}$ and $C_l^{TB}$
453: \cite{feng}.
454: In contrast,
455: more general anisotropic and frequency-dependent
456: cases cause very complicated mixing
457: between the various $C_l^{X_1X_2}$
458: and require numerical integration of
459: the rotation \rf{rot} over the sky and
460: frequency range.
461:
462:
463: \section{Results}
464:
465: To illustrate the kinds of sensitivities
466: that are possible in CMB searches for birefringence,
467: we next examine the results of the
468: BOOMERANG experiment
469: \cite{boompol}.
470: This balloon-based experiment
471: made polarization measurements
472: in a narrow band of frequencies
473: at approximately 145 GHz.
474: This relatively high frequency
475: implies that BOOMERANG
476: is well suited to birefringence tests
477: since for all violations,
478: except those with $d=3$,
479: higher photon energy implies a larger
480: rotation in polarization.
481: The small frequency range is also
482: helpful since we can approximate
483: all frequencies as $\sim 145$ GHz.
484:
485: \begin{figure}
486: \centerline{\includegraphics[width=\textwidth]{fig.eps}}
487: \caption{\label{fig}
488: Relative likelihood
489: versus coefficients for Lorentz violation.
490: Boxes indicate numerically
491: calculated values,
492: and the curve is the smooth
493: extrapolation of these points.
494: The dark-gray shaded region
495: indicates the 68\% confidence level,
496: and the light-gray shows the 95\% level.}
497: \end{figure}
498:
499: In our calculation,
500: we assume conventional polarization
501: is produced during recombination and
502: numerically determine the rotated
503: polarization for points across the sky.
504: The resulting $C_l^{X_1X_2}$ for various
505: values of coefficients for Lorentz violation
506: are determined and compared to published
507: BOOMERANG results.
508: Figure \ref{fig} shows the calculated
509: relative likelihood
510: for a sample of 12 coefficients for
511: Lorentz violation.
512: In each case,
513: we vary the value of one coefficient,
514: setting all other coefficients to zero.
515: The $1\si$ and $2\si$ regions are shown.
516:
517: Some generic features are seen in our survey.
518: In each case, the coefficient is
519: nonzero at the $1\si$ level,
520: hinting at possible Lorentz violation.
521: However,
522: since this occurs in every case,
523: it is likely that this indicates
524: some systemic feature of the
525: BOOMERANG data or our analysis.
526: We also see that each case
527: is consistent with no Lorentz violation
528: at the $2\si$ level,
529: giving conservative upper
530: bounds on the 12 coefficients
531: in Figure \ref{fig}.
532:
533: These results demonstrate
534: the potential of the CMB for
535: testing Lorentz invariance.
536: Due to the long propagation times,
537: the sensitivities to $d=3$ coefficients
538: afforded by the CMB are near the
539: limit of what can be expected
540: in birefringence tests.
541: However, for $d\geq 4$,
542: better sensitivities might be
543: obtained using high-frequency
544: sources like gamma-ray bursts
545: \cite{km2}.
546: Regardless, because of its all-sky nature,
547: the CMB provides a useful
548: probe that can simultaneously probe
549: large portions of coefficient space,
550: which is difficult in searches involving
551: a handful of point sources that
552: access a limited number of propagation directions.
553:
554: \section*{Acknowledgments}
555: This work was supported in part
556: by the Wisconsin Space Grant Consortium.
557:
558: \begin{thebibliography}{xx}
559: \def\etal {{\it et al.}}
560:
561: \bibitem{kost}
562: V.A.\ Kosteleck\'y and S.\ Samuel, Phys.\ Rev.\ D {\bf 39}, 683 (1989);
563: V.A.\ Kosteleck\'y and R.\ Potting, Nucl.\ Phys.\ B {\bf 359}, 545 (1991).
564: \bibitem{cpt04}
565: For a recent review of various
566: experimental and theoretical
567: approaches to Lorentz violation,
568: see, for example,
569: V.A.\ Kosteleck\'y, ed.,
570: {\it CPT and Lorentz Symmetry III, } World Scientific, Singapore, 2005.
571: \bibitem{ck}
572: D.\ Colladay and V.A.\ Kosteleck\'y,
573: Phys.\ Rev.\ D {\bf 55}, 6760 (1997); {\bf 58}, 116002 (1998);
574: V.A.\ Kosteleck\'y, Phys.\ Rev.\ D {\bf 69}, 105009 (2004).
575: \bibitem{cavities}
576: J.\ Lipa \etal, Phys.\ Rev.\ Lett.\ {\bf 90}, 060403 (2003);
577: H.\ M\"uller \etal, Phys.\ Rev.\ Lett.\ {\bf 91}, 020401 (2003);
578: P.\ Antonini \etal, Phys.\ Rev.\ A {\bf 71}, 050101 (2005);
579: S.\ Herrmann \etal, Phys.\ Rev.\ Lett.\ {\bf 95}, 150401 (2005);
580: P.L.\ Stanwix \etal, Phys.\ Rev.\ D {\bf 74}, 081101 (2006).
581: \bibitem{cfj}
582: S.M.\ Carroll, G.B.\ Field, and R.\ Jackiw, Phys.\ Rev.\ D {\bf 41}, 1231 (1990).
583: \bibitem{km1}
584: V.A.\ Kosteleck\'y and M.\ Mewes,
585: Phys.\ Rev.\ Lett.\ {\bf 87}, 251304 (2001);
586: Phys.\ Rev.\ D {\bf 66}, 056005 (2002).
587: \bibitem{km2}
588: V.A.\ Kosteleck\'y and M.\ Mewes, Phys.\ Rev.\ Lett.\ {\bf 97}, 140401 (2006).
589: \bibitem{km3}
590: V.A.\ Kosteleck\'y and M.\ Mewes, Phys.\ Rev.\ Lett.\ {\bf 99}, 011601 (2007).
591: \bibitem{photons}
592: H.\ M\"uller \etal, Phys. Rev. D {\bf 67}, 056006 (2003);
593: R.\ Lehnert and R.\ Potting, Phys.\ Rev.\ Lett.\ {\bf 93}, 110402 (2004);
594: Q.G.\ Bailey and V.A.\ Kosteleck\'y, Phys.\ Rev.\ D {\bf 70}, 076006 (2004);
595: M.\ Mewes and A.\ Petroff, Phys.\ Rev.\ D {\bf 75}, 056002 (2007);
596: A.J.\ Hariton and R.\ Lehnert, Phys.\ Lett.\ A {\bf 367}, 11 (2007);
597: B.\ Altschul, Phys.\ Rev.\ D {\bf 75}, 105003 (2007);
598: A.\ Kobakhidze and B.H.J.\ McKellar, Phys.\ Rev.\ D {\bf 76}, 093004 (2007).
599: \bibitem{km4}
600: V.A.\ Kosteleck\'y and M.\ Mewes, in preparation.
601: \bibitem{cmbrev}
602: For review of current CMB theory and experiment,
603: see, for example,
604: Particle Data Group (http://pdg.lbl.gov),
605: S.\ Eidelman et al., Phys. Lett. B 592, 1 (2004).
606: \bibitem{polrev}
607: For a review of CMB polarimetry, see, for example,
608: W.\ Hu and M.\ White, New Astron.\ {\bf 2}, 323 (1997).
609: \bibitem{cmbexp}
610: G.F.\ Smoot {\it et al.}, Ap.\ J.\ {\bf 396}, L1 (1992);
611: J.\ Kovac {\it et al.}, Nature {\bf 420} 772 (2002);
612: A.C.S.\ Readnead {\it et al.}, Science {\bf 306}, 836 (2004);
613: D.\ Barkats {\it et al.}, Ap.\ J.\ {\bf 619}, L127 (2005);
614: E.M.\ Leitch {\it et al.}, Ap.\ J.\ {\bf 624}, 10 (2005);
615: W.C.\ Jones \etal, Ap.\ J.\ {\bf 647}, 823 (2006);
616: G.\ Hinshaw \etal, Ap.\ J.\ Supp.\ {\bf 170}, 288 (2007);
617: L.\ Page \etal, Ap.\ J.\ Suppl.\ {\bf 170}, 335 (2007).
618: \bibitem{boompol}
619: T.E.\ Montroy \etal, Ap.\ J.\ {\bf 647}, 813 (2006);
620: F.\ Piacentini \etal, Ap.\ J.\ {\bf 647}, 833 (2006).
621: \bibitem{feng}
622: B.\ Feng {\it et al.}, Phys.\ Rev.\ Lett.\ {\bf 96}, 221302 (2006);
623: P.\ Cabella \etal, Phys.\ Rev.\ D {\bf 76}, 123014 (2007).
624: \end{thebibliography}
625:
626: \end{document}
627: