1: \documentclass[twocolumn,showpacs,amsmath,amssymb,prl]{revtex4}
2: \usepackage{graphicx}% Include figure files
3: \usepackage{dcolumn}% Align table columns on decimal point
4: \usepackage{bm}% bold math
5:
6: \begin{document}
7: \input{epsf}
8:
9: \title{The Cumulative Bakground of High-Energy Neutrinos from Starburst
10: Galaxies}
11: \author{Abraham Loeb$^{1,2}$ \& Eli Waxman$^3$}
12: \affiliation{$^1$ Astronomy Department, Harvard University, 60 Garden Street, Cambridge, MA 02138, USA}
13: \affiliation{$^2$ Einstein Minerva center, Weizmann Institute of Science, Rehovot 76100, Israel,}
14: \affiliation{$^3$Physics Faculty, Weizmann Institute of Science, Rehovot 76100, Israel}
15:
16: \begin{abstract}
17:
18: We show that starburst galaxies convert efficiently cosmic-rays into pions,
19: which in turn decay into high-energy neutrinos and photons. The cumulative
20: background of GeV neutrinos is $E_\nu^2\Phi_\nu\approx 10^{-7}~{\rm
21: GeV~cm^{-2}~s^{-1}~sr^{-1}}$, and its extrapolation to higher neutrino
22: energies depends on the energy spectrum of the injected cosmic-rays, with
23: $E_\nu^2\Phi_\nu\propto E_\nu^{-0.15\pm0.1}$ up to $E_\nu\sim0.3$~PeV and
24: possibly higher neutrino energies. This flux, which constitutes a lower
25: limit to the high energy extra-Galactic neutrino flux, is potentially
26: detectable by forthcoming km-scale neutrino telescopes.
27:
28: \end{abstract}
29:
30: \pacs{95.85.Ry, 98.54.Ep, 98.70.Sa, 14.60.Pq}
31:
32: \date{\today}
33: \maketitle
34:
35: Large area, high-energy neutrino telescopes are being constructed to detect
36: extra-Galactic neutrino sources \cite{HalzenHooper02,Gaisser03}. It is
37: widely believed that high-energy neutrinos are produced at cosmological
38: distances based on the fact that the cosmic-ray energy spectrum extends to
39: $>10^{20}$~eV and is most likely dominated above $\sim3\times10^{18}$~eV by
40: extra-Galactic protons \cite{Gaisser03}. High-energy neutrinos are then
41: likely to be emitted by the same sources that produce the high-energy
42: protons, through the decay of charged pions generated by interactions of
43: these protons with the radiation field (or nucleons) within these sources
44: \cite{Man}. Gamma-ray bursts (GRBs,\cite{Waxman05}) and jets of active
45: galactic nuclei (AGN, \cite{Stecker91}) have been suggested as possible
46: sources of high-energy neutrinos that are associated with high-energy
47: cosmic-rays.
48:
49: If the source size is not much larger than the mean-free-path of ultra
50: high-energy cosmic-rays (UHECRs) for pion production,
51: as is the case for both AGN jets and GRBs, then UHECR observations set
52: an upper bound of $E_\nu^2\Phi_\nu<E_\nu^2\Phi_\nu^{\rm
53: WB}=5\times10^{-8}{\rm GeV~cm^{-2}~s^{-1}~ sr^{-1}}$ on the flux of
54: high-energy neutrinos produced by pion decay \cite{WBBound}. This
55: so-called Waxman-Bahcall (WB) upper bound, implies that km (Gton)
56: scale neutrino detectors are required for detecting the high energy
57: neutrino flux produced by sources of UHECRs in the energy range
58: of $\sim1$~TeV to $\sim1$~PeV, and larger detectors are required at
59: yet higher energies \cite{WBBound}.
60:
61: The extra-Galactic neutrino number flux per unit energy is expected to be
62: close to $\Phi_\nu^{\rm WB}$ in the energy range of $10^{19}$--$10^{20}$~eV
63: \cite{gzk_nu}, since protons originating at cosmological distances with
64: $>5\times10^{19}$~eV lose all their energy to pion production through their
65: interaction with cosmic microwave background (CMB) photons \cite{gzk}.
66: However, the existence of a neutrino flux at the level of $\Phi_\nu^{\rm
67: WB}$ for lower energies is not guaranteed. Protons in this energy range may
68: lose only a small fraction of their energy to pion production, making a
69: neutrino flux much lower than $\Phi_\nu^{\rm WB}$.
70:
71: In this {\it Letter} we show that radio observations of starburst galaxies
72: imply a lower limit on their cumulative extra-Galactic neutrino background
73: flux $\Phi_\nu^{\rm SB}$ which is of order $\Phi_\nu^{\rm WB}$ at the
74: energy range of $\sim1$~GeV to $\sim300$~TeV (0.3~PeV), and possibly
75: extending also to higher energies. Previous discussions \cite{Torres04}
76: have used a different model-dependent approach to estimate the neutrino
77: luminosities of individual galaxies and did not integrate the background
78: neutrino flux over cosmic history. The main uncertainty in our estimate
79: involves the spectral index of the injected cosmic-rays. As we show next,
80: starburst galaxies are "hidden" cosmic-ray sources in the sense that they
81: dissipate the cosmic-ray energy produced within them, and so their neutrino
82: background could in principle exceed the WB bound. Nevertheless, the
83: neutrino flux we derive is comparable to the WB bound, because the energy
84: production rate of cosmic rays in starbursts is close to the cosmic
85: production rate of UHECRs. As discussed below, this is not necessarily a
86: coincidence.
87:
88: The detection of multi-TeV neutrinos by upcoming km-scale detectors would
89: provide unique constraints on models for cosmic-ray generation, as well as
90: a unique handle on the physical properties of the interstellar medium in
91: starburst galaxies. The expected detection rate may allow km scale
92: detectors to study the flavor ratio of astrophysical neutrinos, thus
93: probing neutrino oscillations and physics across unprecedented scales of
94: energies and lengths \cite{nu_flavors}.
95:
96:
97: \paragraph*{Starbursts as neutrino factories.}
98: A substantial fraction of the cosmic star formation activity at redshift
99: $z\sim 2$ \cite{Reddy,Jun} occurs in transient starburst episodes. These
100: episodes are often triggered by galaxy merger events, which channel fresh
101: gas towards the center of the merger remnant \cite{Mihos,Springel}. The
102: characteristic scale over which the gas is observed to be concentrated,
103: ($\ell\sim$ hundreds of pc) and the typical gas velocities, ($v\sim$few
104: hundred ${\rm km~s^{-1}}$) imply a dynamical time of a few million years,
105: comparable to the lifetime of massive stars. Core-collapse supernovae (SNe)
106: are therefore expected to enrich the gas with relativistic protons and
107: electrons, i.e. cosmic-rays, which are accelerated in the collisionless
108: shocks produced by these explosions. Additional relativistic particles may
109: be injected by an outflow from a central supermassive black hole, in cases
110: where quasar activity accompanies the starburst phase
111: \cite{Genzel,Springel}. Synchrotron radio emission is routinely observed
112: in starburst galaxies, confirming the presence of relativistic electrons
113: within them \cite{Condon01}.
114:
115: Relativistic protons, injected along with the electrons into the starburst
116: interstellar medium, would lose energy primarily through pion production by
117: inelastic collisions with interstellar nucleons. The decay of charged
118: pions, $\pi^+\rightarrow\mu^++\nu_\mu\rightarrow
119: e^++\nu_e+\bar{\nu}_\mu+\nu_\mu$ and
120: $\pi^-\rightarrow\mu^-+\bar{\nu}_\mu\rightarrow
121: e^-+\bar{\nu}_e+\bar{\nu}_\mu+\nu_\mu$, would then convert part of the
122: proton energy to neutrinos. In what follows we estimate the neutrino flux
123: by showing that: {\it (i)} protons lose essentially all their energy to
124: pion production, and {\it (ii)} a lower limit to the energy loss rate of
125: protons may be directly derived from the synchrotron radio flux of the
126: secondary $e^\pm$.
127:
128: Protons would lose all their energy to pion production provided that the
129: energy loss time is shorter than both the starburst lifetime and the
130: magnetic confinement time within the starburst gas. In the energy range of
131: interest, the inelastic nuclear collision cross section is $\approx50$~mb,
132: with inelasticity of $\approx0.5$ \cite{GaisserBook}. The energy loss time,
133: $\tau_{\rm loss}\approx (0.5 n\sigma_{\rm pp}c)^{-1}$ where $n$ is the
134: interstellar nucleon density, would be shorter than the starburst lifetime
135: which is at least a dynamical time, $\sim(2\ell/v)$, as long as
136: \begin{equation}
137: \Sigma_{\rm gas}\gtrsim \Sigma_{\rm crit}\equiv {m_p \beta \over \sigma_{\rm
138: pp}}= 3\times 10^{-2} \beta_{-3}~{\rm g~cm^{-2}}.
139: \label{sig}
140: \end{equation}
141: Here $\Sigma_{\rm gas}\sim m_pn\ell$ is the surface mass density of the
142: gas and $\beta=v/c=\beta_{-3}(v/300~{\rm km~s^{-1}})$. As it turns out, the
143: critical surface density, $\Sigma_{\rm crit}$, is comparable to the minimum
144: $\Sigma_{\rm gas}$ in known starburst galaxies \cite{Thompson05}.
145:
146: The ratio of confinement time, $\tau_{\rm conf}$, and loss time, $\tau_{\rm
147: loss}$, is less straightforward to estimate, since magnetic confinement of
148: cosmic-rays is not well understood. For starburst galaxies $\Sigma_{\rm
149: gas}\approx10^9M_\odot/{\rm kpc^2}=0.2{\rm g~cm^{-2}}$, which for a disk
150: height of $\sim 100$~pc \cite{Condon91} implies $\tau_{\rm
151: loss}\approx10^5$~yr. The confinement time of 10~GeV protons in our Galaxy
152: is $\tau_{\rm conf}\simeq10^7$~yr \cite{Yanasak01}. The total gas column
153: density traversed by protons of energy $\le1$~TeV before they escape the
154: Galaxy is $\Sigma_{\rm conf}\approx9(E/10{\rm GeV})^{-s}{\rm g~cm^{-2}}$
155: with $0.5\lesssim s\lesssim0.6$ \cite{Sigma_conf}, suggesting $\tau_{\rm
156: conf}\simeq10^7(E/10{\rm GeV})^{-0.6}$~yr for the most commonly-used value
157: of $s=0.6$. If the confinement time in starburst galaxies was similar to
158: that of the Galaxy, then the comparison of $\tau_{\rm loss}$ with
159: $\tau_{\rm conf}$ would have implied that protons with $E\lesssim30$~TeV
160: lose all their energy prior to their escape from starbursts. However, the
161: magnetic field in starbursts is much larger than in the Galaxy. Thompson et
162: al. \cite{Thompson05} have found that the magnetic field strength within
163: starburst galaxies scales linearly with $\Sigma_{\rm gas}$, and has an
164: amplitude that is $\sim 100$ times larger than in the Galaxy for
165: $\Sigma_{\rm gas}=0.2{\rm g~cm^{-2}}$. Assuming that the confinement time
166: depends on energy only through the proton's Larmor radius ($\propto E/B$),
167: this suggests that protons with $E\lesssim3\times10^3$~TeV lose all their
168: energy prior to their escape from typical starburst galaxies. Moreover,
169: the neutrino flux is expected to be dominated by starbursts at $z\gtrsim
170: 1$, for which the typical surface density should be even higher than in
171: local starbursts. It is therefore reasonable to assume that most of the
172: energy injected into starburst galaxies in $E\lesssim3\times10^3$~TeV
173: protons is converted to pions.
174:
175: We now proceed to calibrate the luminosities of starburst galaxies in
176: high-energy neutrinos based on their observed synchrotron luminosities. As
177: recently shown by Thompson et al. \cite{Thompson05}, the synchrotron
178: cooling time of relativistic electrons radiating at a frequency of few GHz
179: is typically much shorter than the starburst lifetime. The synchrotron
180: luminosity reflects, therefore, the energy production rate in these
181: electrons. The production rate of electrons with energy $E_e$,
182: $E_e^2d\dot{N}_e/dE_e$, is related to the synchrotron radio luminosity
183: $L_\nu$ per unit frequency $\nu$ produced by these electrons, as
184: $E_e^2d\dot{N}/dE_e\approx2\nu L_\nu$. The factor of 2 originates from the
185: fact that the synchrotron frequency is proportional to the square of the
186: electron Lorentz factor, $\nu\propto E_e^2$, so that the energy of
187: electrons in a given decade of $E_e$ is spread over two decades of photon
188: frequency $\nu$. Denoting the ratio of injected power in protons and
189: electron at a fixed particle energy as $\eta_{p/e}$, the luminosity in
190: $\nu_\mu$ and $\bar{\nu}_\mu$ per logarithmic energy bin $E_\nu$ is given
191: by $E_\nu dL/dE_\nu=(1/3)\eta_{p/e}E_e^2d\dot{N}/dE_e= (2/3)\eta_{p/e}\nu
192: L_\nu$. The factor of $1/3$ comes from the fact that $\approx2/3$ of the
193: proton energy is carried by charged pions ($\sim1/3$ by neutral pions), and
194: $\approx1/2$ of the pion energy is carried, after it decays, by muon
195: neutrinos. Cosmic-ray data in the Milky-Way galaxy indicates
196: $\eta_{p/e}\sim50$ (e.g. \cite{Eichler}). However, we expect a lower ratio
197: in starburst galaxies. The decay of the charged pions produces secondary
198: $e^\pm$ which carry $\approx2/3\times1/4=1/6$ of the proton energy (since
199: the $e^\pm$ carry $\approx 1/4$ of the decaying $\pi^\pm$ energy). This
200: implies $\eta_{p/e}\approx6$ and a neutrino luminosity $E_\nu
201: dL/dE_\nu=4\nu L_\nu$. We conservatively ignore the ionization loss by
202: the electrons \cite{Thompson05}, whose inclusion would only increase our
203: predicted neutrino flux.
204:
205: In the Milky way, electrons radiating at few GHz have energies
206: $E_e\sim10$~GeV. In starburst galaxies the magnetic field is typically
207: $\sim100$ times larger, implying that the radio luminosity observed is
208: produced by $E_e\sim1$~GeV electrons. Since the electrons and neutrinos
209: produced in pion decays carry similar energy per particle, the radio
210: luminosity provides a direct estimate of the GeV neutrino luminosity,
211: $E_\nu dL/dE_\nu(E_\nu=1{\rm GeV})= 4\nu L_\nu|_{\nu\sim1{\rm GHz}}$. The
212: local 1.4~GHz energy production rate per unit volume is $\nu
213: (dL_\nu/dV)|_{\nu=1.4\rm GHz} \approx 10^{43}~ {\rm erg~Mpc^{-3}~yr^{-1}}$
214: \cite{Condon01}. Most of the stars in the Universe formed at redshifts
215: $z\sim 2$-4 \cite{Ostriker} inside starburst galaxies \cite{Reddy,Jun}.
216: The fact that the starburst mode of star formation dominated at $z> 2$ is
217: also supported by the prominence of quasar activity around the same
218: redshift, as both processes are believed to result from mergers of gas-rich
219: galaxies. We extrapolate the local 1.4~GHz energy production rate per unit
220: volume (of which a dominant fraction is produced in quiescent spiral
221: galaxies) to the redshifts where most of the stars had formed through the
222: starburst mode, based on the observed redshift evolution of the cosmic star
223: formation rate \cite{Ostriker}, and calculate the resulting neutrino
224: background. The cumulative GeV neutrino background from starburst galaxies
225: is then
226: \begin{eqnarray}
227: E_\nu^2\Phi_{\nu}(E_\nu&=&1{\rm GeV})\approx \frac{c}{4\pi}\zeta t_H
228: \left[4\nu (dL_\nu/dV)\right]_{\nu=1.4\rm GHz}
229: \nonumber\\
230: &=& 10^{-7}\zeta_{0.5}~ {\rm GeV~cm^{-2}~s^{-1}~sr^{-1}}.
231: \label{back}
232: \end{eqnarray}
233: Here, $t_H$ is the age of the Universe, and the factor
234: $\zeta=10^{0.5}\zeta_{0.5}$ incorporates a correction due to redshift
235: evolution of the star formation rate relative to its present-day value.
236: The value of $\zeta_{0.5}\sim 1$ applies to activity that traces the cosmic
237: star formation history \cite{WBBound}. Note that flavor oscillations would
238: convert the pion decay flavor ratio, $\nu_e:\nu_\mu:\nu_\tau=1:2:0$ to
239: $1:1:1$ \cite{nu_flavors}, so that $\Phi_{\nu_e}=\Phi_{\nu_\mu}
240: =\Phi_{\nu_\tau}=\Phi_{\nu}/2$.
241:
242: \begin{figure} [ht]
243: \centerline{\epsfxsize=3.4in \epsfbox{fig1.eps}}
244: \caption{The shaded region brackets the range of plausible choices for the
245: spectrum of the neutrino background. Its upper boundary is obtained for a
246: power-law index $p=2$ of the injected cosmic-rays, and its lower boundary
247: corresponds to $p=2.25$ for $E_\nu<10^{14.5}$~eV.
248: The solid green line corresponds to the likely value $p=2.15$
249: (see text). Other lines: the WB upper bound on the
250: high energy muon neutrino intensity from optically-thin sources;
251: the neutrino intensity expected from interaction with CMB photons (GZK);
252: the atmospheric neutrino background; experimental upper bounds of optical
253: Cerenkov experiments (BAIKAL \cite{Baikal} and AMANDA \cite{amanda_bound});
254: and the expected sensitivity
255: of 0.1~km$^2$ and 1~km$^2$ optical Cerenkov detectors \cite{HalzenHooper02}.}
256: \label{fig:flux}
257: \end{figure}
258:
259: Equation~(\ref{back}) provides an estimate of the GeV neutrino
260: background. The extrapolation of this background to higher neutrino
261: energies depends on the energy spectrum of the high energy protons. If the
262: proton energy distribution follows a power-law, $dN/dE\propto E^{-p}$, then
263: the neutrino spectrum would be, $E_\nu^2\Phi_{\nu_\mu}\propto E_\nu^{2-p}$.
264: The energy distribution of cosmic-ray protons measured on Earth follows a
265: power-law $dN/dE\propto E^{-2.75}$ up to the "knee" in the cosmic-ray
266: spectrum at a few times $10^{15}$~eV \cite{Eichler,BalloonProton}. (The
267: proton spectrum becomes steeper, i.e. softer, at higher energies
268: \cite{Gaisser03}.) Given the energy dependence of the confinement time,
269: $\propto E^{-s}$ \cite{Sigma_conf}, this implies a production spectrum
270: $dN/dE\propto E^{-p}$ with $p=2.75-s\approx2.15$. This power-law index is
271: close to, but somewhat higher than, the theoretical value $p=2$, which
272: implies equal energy per logarithmic particle energy bin, obtained for
273: Fermi acceleration in strong shocks under the test particle approximation
274: \cite{Krymskii77}. We note that the cosmic-ray spectrum observed on Earth
275: may not be representative of the cosmic-ray distribution in the Galaxy in
276: general. The inferred excess relative to model predictions of the $>1$~GeV
277: photon flux from the inner Galaxy, implies that the cosmic-rays are
278: generated with a spectral index $p$ smaller than the value $p=2.15$
279: inferred from the local cosmic-ray distribution, and possibly that the
280: spectral index of cosmic-rays in the inner Galaxy is smaller than the local
281: one \cite{innerCR}. The spectrum of electrons accelerated in SNe is
282: inferred to be a power law with spectral index $p=2.1\pm0.1$ over a wide
283: range energies, $\sim1$~GeV to $\sim10$~TeV, based on radio, X-ray and TeV
284: observations (e.g. \cite{SN}).
285:
286: For a steeply falling proton spectrum such as $dN/dE\sim E^{-2}$, the
287: production of neutrinos of energy $E_\nu$ is dominated by protons of energy
288: $E\approx20E_\nu$ \cite{GaisserBook}, so that the cosmic-ray "knee"
289: corresponds to $E_\nu\sim0.1$~PeV. In analogy with the Galactic injection
290: parameters of cosmic-rays, we expect the neutrino background to scale as
291: \begin{eqnarray}
292: E_\nu^2\Phi_{\nu}^{\rm SB}\approx10^{-7}(E_\nu/1{\rm GeV})^{-0.15\pm0.1}
293: {\rm GeV~cm^{-2}~s^{-1}~sr^{-1}}
294: \label{back_h}
295: \end{eqnarray}
296: up to $\sim0.1$~PeV. In fact, the "knee" in the proton spectrum for
297: starburst galaxies may occur at an energy higher than in the Galaxy. The
298: steepening (softening) of the proton spectrum at the knee may be either due
299: to a steeper proton production spectrum at higher energies, or a faster
300: decline with energy for the proton confinement time. Since both the
301: acceleration of protons and their confinement depend on the magnetic field,
302: we expect the "knee" to shift to a higher energy in starbursts, where the
303: magnetic field is much stronger than the Galactic value. The predicted
304: neutrino intensity is shown as a solid line in Fig.~\ref{fig:flux}. The
305: shaded region illustrating the range of uncertainty in the predicted
306: neutrino background. This range is bounded from above by the intensity
307: obtained for $p=2$, corresponding to equal proton energy per logarithmic
308: bin, and from below by the intensity obtained for $p=2.25$, corresponding
309: to the lower value of the confinement time spectral index, $s=0.5$.
310:
311: The extension of the neutrino spectrum to energies $E_\nu>1$~PeV is highly
312: uncertain. If the steepening of the proton spectrum at the knee is due to
313: a rapid decrease in the proton confinement time within the Galaxy rather
314: than a change in the production spectrum, then the neutrino background may
315: follow Eq.~(\ref{back_h}) above $1$~PeV, provided that the magnetic
316: confinement time in starbursts does not decline rapidly at these
317: energies. This extension is shown as the dashed line in
318: Fig.~\ref{fig:flux}. If, on the other hand, the steepening at the knee is
319: due to a steepening in the production spectrum, then the neutrino
320: background may decrease rapidly following the lower boundary of the shaded
321: region. Note that although transient sources such as GRBs or AGN may not
322: contribute at the present-time to the observed Galactic cosmic-rays, they
323: could dominate the cosmic integral of the neutrino emission by
324: starburst galaxies \cite{Galactic}. Such a contribution would be
325: particularly important at $E_\nu\gtrsim0.1$~PeV if the same sources
326: produce the UHECRs.
327:
328: \paragraph*{Gamma-ray Production.}
329: The neutrino emission from $\pi^\pm$ decay will be accompanied by a
330: comparable flux of $\gamma$-rays from $\pi^0$ decay with energies
331: $E_\gamma\gtrsim1$GeV. Since the protons carry more energy than the
332: electrons and the electrons lose only a minor fraction of their energy to
333: $\gamma$-rays (through inverse-Compton scattering of soft photons
334: \cite{Thompson05}), the $\gamma$-rays provide a robust measure of the pion
335: production rate by protons. This allows a sanity check on our predicted
336: neutrino fluxes. Our expected $\gamma$-ray flux from the nearest starburst
337: galaxies must be smaller than existing upper-limits. For example, analysis
338: of HESS and EGRET data for NGC253 provided an upper limit of $\sim
339: 1.9\times 10^{-12}~{\rm erg~cm^{-2}~s^{-1}}$ on its $\gamma$-ray flux as a
340: point source at $E_\gamma\gtrsim 1$~TeV \cite{NGC253,Cillis}. Based on its
341: 1.4GHz flux of $10^{-13}~{\rm erg~cm^{-2}~s^{-1}}$, we predict a
342: $\gamma$-ray flux of $\lesssim 10^{-12}~{\rm erg~cm^{-2}~s^{-1}}$, well
343: below the existing upper limit. We have repeated this calculation for other
344: starburst galaxies with known upper limits \cite{Cillis}, and found similar
345: results. Our expected $\gamma$-ray fluxes are consistent with previous
346: models for individual sources \cite{Torres04}, which were based on other
347: considerations. The forthcoming GLAST
348: \footnote{http://www-glast.stanford.edu} mission will have the sensitivity
349: to detect the predicted $\gamma$-ray fluxes from nearby starbursts and
350: revalidate our expectation for the neutrino background.
351:
352:
353: \paragraph*{Discussion.} The neutrino intensity predicted by Eq.~(\ref{back_h})
354: at $100$~TeV is $\approx2\times10^{-8\pm0.5}{\rm
355: Gev~cm^{-2}~s^{-1}\,sr^{-1}}$, implying a detection rate at $E_\nu>100$~TeV
356: of $\approx10^{1.5\pm0.5}$ events per yr in a 1~km$^2$ detector. The
357: neutrino background is therefore potentially detectable by forthcoming
358: km-scale telescopes. If detected, the signal would serve as a
359: highly-effective probe of cosmic-ray production in star forming
360: environments.
361:
362: As already mentioned, our inferred background of high-energy neutrinos is
363: similar to $\Phi^{\rm WB}$ since the present-day energy production rate of
364: cosmic-ray protons by starbursts,
365: $E^2d\dot{n}/dE=12\nu(dL_\nu/dV)|_{\nu=1.4\rm GHz} \approx 10^{44}{\rm
366: erg~Mpc^{-3}~yr^{-1}}$, is similar to the energy production rate of UHECRs
367: \cite{UHECRrate}. While the neutrino background originates primarily in
368: starburst galaxies with $\Sigma_{\rm gas}\gtrsim \Sigma_{\rm crit}$,
369: extra-Galactic cosmic-rays would originate in galaxies with $\Sigma_{\rm
370: gas}< \Sigma_{\rm crit}$, where they do not encounter substantial loses on
371: their way out. Since the two classes of galaxies are responsible for making
372: comparable fractions (up to a factor of $\sim 2$) of the stellar mass
373: reservoir in the local universe \cite{Jun}, the coincidence between
374: $\Phi_\nu^{\rm SB}$ and $\Phi_\nu^{\rm WB}$ may simply reflect the matching
375: condition at $\Sigma_{\rm gas}\sim \Sigma_{\rm crit}$ that known galactic
376: environments account for the extragalactic flux of UHECRs. Thus, as long as
377: the sources of UHECRs (such as GRBs or AGN) exist in starburst galaxies,
378: they would produce a measurable background of neutrinos with energies
379: $E_\nu\gtrsim 10^{14}$ eV (near the upper envelope of the shaded region in
380: Fig. 1), that extends well beyond the atomspheric neutrino background.
381:
382: Our results do not depend on whether the cosmic-rays originate from
383: supernovae or the supermassive black holes in galactic nuclei. While
384: supernova-produced $\gamma$-rays will be spread throughout the star-forming
385: volume of these nuclei, a black hole origin would bias the $\gamma$--rays
386: to be centrally concentrated. GLAST will be able to distinguish between
387: these possibilities by imaging nearby starburst galaxies in $\gamma$-rays.
388:
389: \paragraph*{Acknowledgments.}
390: This work was supported in part by ISF and Minerva grants (E. W.) and by
391: NASA grants 5-7768 and NNG05GH54G (A. L.).
392:
393: \begin{references}
394:
395: \bibitem{HalzenHooper02} F. Halzen \& D. Hooper, Rep. Prog. Phys. {\bf 65},
396: 1025 (2002).
397:
398: \bibitem{Gaisser03} T. K. Gaisser, Nucl. Phys. B (Proc. Supp.) {\bf 117},
399: 318 (2003).
400:
401: \bibitem{Man} Learned, J. G., Mannheim, K., Ann. Rev. of Nuc. \&
402: Part. Sci. {\bf 50}, 679 (2000).
403:
404: \bibitem{Stecker91} F. W. Stecker. C. Done, M. H. Salamon, \& P. Summers,
405: Phys. Rev. Lett., {\bf 66}, 2697 (1991); K. Mannheim, Phys. Rev. {\bf D48},
406: 2408 (1994); A. P. Szabo \& R. J. Protheroe, Astroparticle Phys., {\bf 2},
407: 375 (1994).
408:
409: \bibitem{Waxman05} E.~Waxman, Proc. Nobel Symp. 129: Neutrino Physics
410: (Sweden 2004); [arXiv:astro-ph/0502159]; S. Razzaque, P. Meszaros, \&
411: E. Waxman, Int. J. Mod. Phys., {\bf 20}, 3099 (2005).
412:
413: \bibitem{WBBound} E. Waxman \& J. N. Bahcall, Phys. Rev. {\bf D59}, 023002
414: (1999); J. N. Bahcall \& E. Waxman, Phys. Rev. {\bf D64}, 023002 (2001).
415:
416: \bibitem{gzk_nu} V. S. Berezinsky \& G. T. Zatsepin, Phys. Lett. {\bf 28B},
417: 423 (1969); R. Engel, D. Seckel \& T. Stanev, Phys. Rev. {\bf D64}, 093010
418: (2001).
419:
420: \bibitem{Cillis} A. N. Cillis, D. F. Torres, \& O. Reimer,
421: Astrophys. J. {\bf 621}, 139 (2005).
422:
423: \bibitem{Torres04} D. F. Torres, Astrophys. J. {\bf 617}, 966 (2004);
424: E. Domingo-Santamar{\'{\i}}a, \& D. F. Torres, Astron. \& Astrophys. {\bf
425: 444}, 403 (2005); G. E. Romero, \& D. F. Torres, Astrophys. J., {\bf 586},
426: L33 (2003).
427:
428: \bibitem{gzk}
429: K. Greisen, Phys. Rev. Lett. {\bf 16}, 748 (1966);
430: G.T. Zatsepin, V.A. Kuzmin, JETP {\bf 4}, 78 (1966).
431:
432: \bibitem{nu_flavors}
433: J.~G.~Learned and S.~Pakvasa, Astropart.\ Phys.\ {\bf 3}, 267 (1995);
434: H.~Athar, M.~Jezabek and O.~Yasuda, Phys.\ Rev.\ D {\bf 62}, 103007 (2000);
435: J.~F.~Beacom, N.~F.~Bell, D.~Hooper, S.~Pakvasa and T.~J.~Weiler,
436: Phys.\ Rev. {\bf D68}, 093005 (2003);
437: L.~A.~Anchordoqui et al., Phys. Rev. {\bf D72}, 065019 (2005);
438: T. Kashti \& E. Waxman, Phys. Rev. Lett. {\bf 95}, 181101 (2005).
439:
440: \bibitem{Reddy} N. A. Reddy,D. K. Erb, C. C. Steidel,
441: A. E. Shapley, K. L. Adelberger, \& M. Pettini, Astrophys. J., {\bf 633},
442: 748 (2005).
443:
444: \bibitem{Jun} S. Juneau, et al. Astrophys. J. Lett. {\bf 619}, L135 (2005).
445:
446: %\bibitem{Smail} I. Smail, R. J. Ivison, \& A. W. Blain,
447: %Astrophys. J. Lett. {\bf 490}, L5 (1997).
448:
449: %\bibitem{Barger} A. J. Barger, L. L. Cowie, \& D. B. Sanders,
450: %Astrophys. J. Lett. {\bf 518}, L5 (1999).
451:
452: \bibitem{Mihos} J. C. Mihos, \& L. Hernquist, Astrophys. J. {\bf
453: 464}, 641 (1996).
454:
455: \bibitem{Genzel} H. Dannerbauer, D. Rigopoulou, D. Lutz, R. Genzel,
456: E. Sturm, \& A.~F.~M.\ Moorwood, Astron. \& Astrophys. {\bf 441}, 999
457: (2005).
458:
459: \bibitem{Springel} V. Springel, T. Di Matteo, \& L. Hernquist,
460: Mon. Not. R. Astron. Soc. {\bf 361}, 776 (2005).
461:
462: \bibitem{Condon01} M.~S. Yun, N.~A. Reddy, \& J.~J. Condon,
463: Astrophys. J. {\bf 554}, 803 (2001).
464:
465: \bibitem{GaisserBook}
466: T. K. Gaisser, "Cosmic-Rays and Particle Physics," Cambridge Univ. Press 1990.
467:
468: \bibitem{Thompson05} T. A. Thompson, E. Quataert, E. Waxman, N. Murray, \&
469: C. L. Martin, Astrophys. J., submitted (2006) [arXiv:astro-ph/0601626]; see
470: also J. H. Voelk, Astron \& Astrophys., {\bf 218}, 67 (1989).
471:
472: %\bibitem{Brinchmann04} J. Brinchmann, S. Charlot, S. D. M. White, C.
473: %Tremonti, G. Kauffmann, G., T. Heckman, \& J. Brinkmann,
474: %Mon. Not. R. Astron. Soc. {\bf 351}, 1151 (2004).
475:
476: \bibitem{Condon91} J. J. Condon, Z.-P. Huang, Q. F.Yin,
477: \& T. X. Thuan, Astorphys. J. 378, 65 (1991);
478: E. R. Seaquist \& N. Odegard, Astorphys. J. {\bf 369}, 320 (1991).
479:
480: \bibitem{Yanasak01} W. R. Webber, \& A. Soutoul, Astroph. J. {\bf 506}, 335
481: (1998); N. E. Yanasak et al., Astorphys. J. {\bf 563}, 768 (2001).
482:
483: \bibitem{Sigma_conf} J. J. Engelmann et al., Astron. Astrophys. {\bf 233},
484: 96 (1990); S. A. Stephens \& R. E. Streitmatter, Astrophys. J. {\bf 505},
485: 266 (1998); W. R. Webber, F. B. McDonald \& A. Lukasiak, Astrophys. J.
486: {\bf 599}, 582 (2003).
487:
488: \bibitem{Eichler} R. Blandford, \& D. Eichler, Phys. Rep. {\bf 154}, 1
489: (1987).
490:
491: \bibitem{Ostriker} K. Nagamine, R. Cen, S. R. Furlanetto,
492: L. Hernquist, C. Night, \& J. P. Ostriker,
493: New Astron. Rev. {\bf 50}, 29 (2006).
494:
495: \bibitem{BalloonProton} K. Asakimori et al., Astrophys. J. {\bf 502}, 278
496: (1998); V. A. Derbina et al., Astrophys. J. {\bf 628}, L41 (2005).
497:
498: \bibitem{Krymskii77} G. F. Krymskii, Sov. Phys. Dokl {\bf 22}, 6 (1977); W. I. Axford, E. Leer \& G. Skadron, Proc. 15th Int. Cosmic-Ray Conf., Plovdiv (Budapest: Central Research Institute for Physics) {\bf 11}, 132 (1977); A. R. Bell, Mon. Not. R. Astron. Soc. {\bf 182}, 147 (1978); R. D. Blandford \& J. Ostriker, Astrophys. J. {\bf 221}, L29 (1978).
499:
500: \bibitem{innerCR} S. D. Hunter et al., Astrophys. J. {\bf 481}, 205 (1997);
501: E. G. Berezhko \& H. J. V\"olk, Astrophys. J. {\bf 540}, 923 (2000).
502:
503: \bibitem{SN} S. P. Reynolds, Astrophys. J. {\bf 459}, L13 (1996);
504: G. E. Allen, R. Petre, E. V. Gotthelf, Astrophys. J. {\bf 558}, 739 (2001);
505: F. Aharonian et al.\, Astron. \& Astrophys. {\bf 437}, L7 (2005);
506: Astron. \& Astrophys., {\bf 442}, 177 (2005); Nature {\bf 432}, 75 (2004).
507:
508: \bibitem{Baikal} V. Balkanov et al., Nucl. Phys. B (Proc. Suppl.) {\bf
509: 110}, 504 (2002).
510:
511: \bibitem{amanda_bound} J. Ahrens et al., Phys. Rev. Lett. {\bf 90}, 251101
512: (2003).
513:
514: \bibitem{Galactic} A. Loeb, \& E. Waxman, preprint (2002);
515: [arXiv:astro-ph/0205272].
516:
517: \bibitem{NGC253} F. Aharonian et al., preprint (2005);
518: [arXiv:astro-ph/0507370].
519:
520: \bibitem{UHECRrate} E. Waxman, Astrophys. J. {\bf 452}, L1 (1995);
521: J. N. Bahcall, E. Waxman, Phys. Lett. B {\bf 556}, 1 (2003).
522:
523: %\bibitem{Juneau05} S. Juneau, et al.\ Astrophys. J. Lett. {\bf 619}, L135
524: %(2005).
525:
526:
527: \end{references}
528:
529:
530: \end{document}
531:
532:
533:
534:
535:
536: