1: %\documentclass{aastex}
2: \documentclass[12pt,preprint]{aastex}
3:
4: %\documentstyle[emulateapj,apjfonts,epsfig]{article}
5:
6: %\shorttitle{} \shortauthors{}
7:
8: %\slugcomment{}
9:
10: \DeclareGraphicsExtensions{.eps}
11:
12: \begin{document}
13:
14: \title{Signatures of hadronic cosmic rays in starbursts?\\
15: High-energy photons and neutrinos from NGC253}
16:
17: \author{Gustavo E. Romero\footnote{Instituto Argentino de Radioastronom\'{\i}a
18: (IAR), C.C.\ 5, 1894 Villa Elisa, Argentina. E-mail:
19: romero@irma.iar.unlp.edu.ar} \& Diego F. Torres\footnote{Lawrence
20: Livermore National Laboratory, 7000 East Ave., L-413, Livermore,
21: CA 94550, USA. E-mail: dtorres@igpp.ucllnl.org}}
22:
23:
24:
25: \begin{abstract}
26: We show that it appears possible for starburst galaxies, like the
27: nearby NGC 253, recently identified as a TeV
28: source by the CANGAROO collaboration, to emit a significant amount
29: of high-energy $\gamma$-rays and neutrinos through hadronic
30: processes in their cores. We suggest that proton illumination of
31: the inner winds of massive stars can be a viable mechanism for
32: producing TeV $\gamma$-rays and neutrinos without a strong MeV-GeV
33: counterpart. The rich stellar content of the starbursts, with
34: millions of early-type stars concentrated in the central regions,
35: where collective effects of the stellar winds and supernovae can
36: produce a significant enhancement of the cosmic ray density,
37: provides an adequate scenario for TeV $\gamma$-ray generation.
38: Close starbursts are also found to be potential sources for
39: km-scale neutrino telescopes, like ICECUBE, within reasonable
40: integration times.
41: \end{abstract}
42:
43: \keywords{galaxies: starburst, gamma rays: observations,
44: gamma rays: theory, galaxies: individual (NGC 253)}
45:
46:
47: \section{Introduction}
48:
49: Very recently, the CANGAROO collaboration reported the
50: detection of the nearby starburst galaxy NGC 253 at TeV
51: $\gamma$-ray energies (Itoh et al. 2002). The purpose of this Letter is
52: to show that it is possible for starburst galaxies, like
53: NGC 253, to produce TeV photons and neutrinos through hadronic processes.
54:
55: \section{A plausible scenario}
56:
57: Collective effects of strong stellar winds and supernova
58: explosions in star-forming regions are expected to result in
59: particle acceleration up to multi-TeV energies, or even beyond
60: (e.g. Montmerle 1979; Cass\'{e} \& Paul 1980; Bykov \& Fleishman
61: 1992a,b; Anchordoqui et al. 1999; Bykov 2001). Recently, the
62: region co-spatial with the Cygnus OB2 association has been
63: detected as a TeV $\gamma$-ray source by the HEGRA telescope array
64: (Aharonian et al. 2002). TeV cosmic rays (CRs) accelerated in the
65: association might be responsible for the high-energy $\gamma$-ray
66: emission through the hadronic illumination of some suitable target
67: (Butt et al. 2003). A nearby EGRET source (3EG J2033+4118)
68: has also a likely stellar origin (Chen et al. 1996, Romero et al.
69: 1999, Benaglia et al. 2001).
70:
71: In a starburst, where the ambient density of CRs is enhanced by a
72: local high rate of supernova explosions and the collective effects
73: of strong stellar winds (e.g. Bykov 2001), $\gamma$-rays can be
74: produced by interactions of relativistic protons with the rich
75: interstellar medium. This interaction produces neutral pions that
76: quickly decay into $\gamma$-rays. The latter can also be produced
77: by leptonic processes, like inverse Compton interactions with the
78: strong FIR field, and relativistic bremsstrahlung in the ambient
79: gas. All these processes have been modelled with a set of
80: reasonable parameters for NGC 253 by Paglione et al. (1996). The
81: expected total $\gamma$-ray flux, however, is below the EGRET
82: sensitivity limit, in accordance with the non-detection of the
83: galaxy in the MeV-GeV range (Bhattacharya et al. 1994, Blom et al.
84: 1999).
85:
86: A source of $\gamma$-rays not previously considered in the
87: mentioned analysis is the production at the base of the strong
88: stellar winds of early-type stars. When the relativistic particles
89: are accelerated in the stellar wind itself, for instance through
90: multiple shocks produced by line-driven instabilities (e.g. White
91: \& Chen 1992), the $\gamma$-ray flux from pion decays is limited
92: to MeV energies. However, if the star is immersed in an external
93: bath of relativistic particles accelerated at larger scales, e.g.
94: in the core of the starburst, TeV protons can penetrate the base
95: of the wind to produce TeV $\gamma$-rays. The injection of
96: MeV---GeV particles into the wind of the stars will be strongly
97: attenuated by the modulation effects of the wind itself.
98:
99: The dense medium in which these $\gamma$-rays are produced is,
100: nonetheless, transparent to $\gamma$-ray propagation. The optical
101: depth to pair production is $\tau \sim n\,\sigma\,R$, where $n$ is
102: the photon number density, $\sigma$ is the photon-photon cross
103: section, and $R$ is the distance that the photon must travel to
104: escape. In order to compute $n$, we need the photon energy density
105: at the base of the stellar wind, $U$. We shall assume that target photons in the
106: wind have typical energies of $\sim$ 1~eV. Then,
107: $U=3\,L/(4\pi\,R^2\,c)\sim 2\times 10^{10}$~eV~cm$^{-3}$, where we
108: have taken, for order-of-magnitude estimates, the following
109: typical values: $R\sim 5\times 10^{14}$~cm and $L\sim
110: 10^{39}$~erg~s$^{-1}$. Then, $n=U/(1~\rm eV)$, and the optical depth
111: is very small, $\tau\sim 2 \times 10^{-5}$. The computation
112: of $\tau$ makes use of the $\gamma-\gamma$ (1 TeV---1eV) cross
113: section, $\sigma \sim 2 \times 10^{-30}$ cm$^2$ (Lang 1999). Once
114: TeV photons are produced and escape from the wind region, they must yet leave
115: the galaxy to reach the Earth. The starburst has a FIR field with $U
116: \sim 200$ eV cm$^{-3}$ (e.g. Paglione et al. 1996), and photons
117: have to travel distances of about 70-100 pc to leave the active
118: region of the galaxy. The optical depth is, in this case, $\tau =
119: (200$ photons cm$^{-3})\; \times (2\times 10^{-30}$ cm$^{2})\;
120: \times (100$ pc) $\approx$ 1.2 $\times 10^{-7}$, which is also small
121: enough to secure that photons can escape from the galaxy.
122:
123: %The overabundance of massive stars with strong winds in the
124: %CR-enhanced core of a nearby starburst like NGC 253 might then
125: %lead to a TeV $\gamma$-ray source without a strong counterpart at
126: %MeV energies. The identification of the hadronic origin of the TeV
127: %emission could, furthermore, be made through the detection of the
128: %associated neutrino emission.
129:
130:
131: \section{NGC 253}% as a starburst}
132:
133:
134: NGC 253 has been described as an archetypal starburst galaxy by
135: Rieke et al. (1980), and it has been extensively studied from
136: radio to $\gamma$-rays (e.g. Beck et al. 1994, Paglione et al.
137: 1996, Ptak 1997).
138: %More than 60 individual compact radio sources
139: %have been detected within the central 200 pc (Ulvestad et al.
140: %1999), most of which are supernova remnants (SNRs) of only a few
141: %hundred years old.
142: The supernova rate is estimated to be as high
143: as $0.2-0.3$ yr$^{-1}$, comparable to the massive star formation
144: rate, $\sim 0.1$M$_\odot$ yr$^{-1}$ (Ulvestad et al. 1999, Forbes
145: et al. 1993). The central region of this starburst is packed with
146: massive stars. Watson et al. (1996) have discovered four young
147: globular clusters near the center of NGC 253; they alone can
148: account for a mass well in excess of 1.5$\times 10^6 M_\odot$ (see
149: also Keto et al. 1999). Assuming that the star formation rate has
150: been continuous in the central region for the last 10$^9$ yrs, and
151: a Salpeter IMF for 0.08-100 $M_\odot$, Watson et al. (1996) find
152: that the bolometric luminosity of NGC 253 is consistent with 1.5
153: $\times 10^8 M_\odot$ of young stars. Based on this evidence,
154: it appears likely that there are at least tens of millions
155: of young stars in the central region of the starburst.
156: Many stars
157: might be obscured by the large amount of molecular material, as in
158: the case of the massive Galactic cluster Westerlund 1 (Clark and
159: Negueruela 2002).
160:
161: Physical, morphological, and kinematic evidence for the existence
162: of a galactic superwind has been found for NGC 253 (e.g. McCarthy
163: et al. 1987, Heckman et al. 1990). This superwind creates a cavity
164: of hot ($\sim10^8$ K) gas, with cooling times longer than the
165: typical expansion time scales. As the cavity expands, a strong
166: shock front is formed on the contact surface with the cool
167: interstellar medium.
168: %Shock interactions with low and high density
169: %clouds can produce X-ray continuum and optical line emission,
170: %respectively, both of which have been directly observed (McCcarthy
171: %et al. 1987).
172: The shock velocity can reach thousands of kilometers per second,
173: and iron nuclei can be efficiently accelerated up to high energies
174: $\sim 10^{20}$ eV by diffusive mechanisms operating at large
175: scales. Indeed, NGC 253 has been proposed as the origin for some
176: of the observed ultra-high energy CRs (Anchordoqui et al. 1999,
177: 2002).
178: %In the
179: %central region of the galaxy, however, where the first step of
180: %particle acceleration occurs, the highest energy of the particles
181: %is expected to be $\sim 10^{15}$ eV.
182:
183:
184: %\section{The TeV photon spectrum for NGC 253}
185:
186:
187: The CANGAROO collaboration did not give a precise photon spectrum
188: for NGC 253. The reported data were only the integrated flux above
189: 0.5 TeV (securing an 11$\sigma$ confidence detection) and the
190: possible range of power law indices $\Gamma=$(1.85, 3.75) (Itoh et al.
191: 2002). We can summarize these results expressing the differential
192: photon spectrum as:
193: $
194: F_{\gamma}(E_{\gamma}) =
195: B\,\, \left({ E_\gamma}/{1\, {\rm TeV}}\right)^{-\Gamma} \,\,\,\,
196: {\rm cm^{-2}\, s^{-1}\, TeV^{-1}},
197: $
198: where $B = (3.8\pm 1.2) \times 10^{-12} $ for $\Gamma=1.8$, or
199: $B = (3.2\pm 1.0) \times 10^{-12} $ for $\Gamma=3.75$. It is worth
200: noticing that the direct extrapolation of the spectrum at TeV
201: energies down to the GeV band would yield an integrated value
202: compatible with the EGRET upper limits for NGC 253 only for power
203: law indices $\Gamma \lesssim 2$. Spectra softer than those with
204: $\Gamma=2$ at TeV energies should present a break at a few
205: hundred GeV.
206:
207:
208: \section{Hadronic origin of the TeV emission from NGC 253?}
209:
210: The differential $\gamma$-ray photon number distribution from
211: neutral pion decays at the source is given by (e.g. Gaisser 1990)
212: $ N_{\gamma}(E_{\gamma})= N_p(E_\gamma)\, \xi_A \, 4 \pi\,
213: \sigma_{pp}(E_\gamma) \, N \, {2Z_{p\rightarrow
214: \pi^0}^{(\alpha)}}/{\alpha}. $ Here, $E_\gamma$ is the photon
215: energy, $N=\Sigma/m_p$ is the column density, $m_p$ is the proton
216: mass, and $\alpha$ is the proton spectral index, such that
217: $N_p=K_p E^{-\alpha}$. In addition, $\xi_A\sim 1.5$ is a
218: correction factor that takes into account possible effects of
219: heavier nuclei, $\sigma_{pp}(E_\gamma)\sim 35$ mb is the hadronic
220: interaction cross section, and $Z_{p\rightarrow \pi^0}^{(\alpha)}
221: \sim 0.17$ stands for the fraction of kinetic energy
222: %$E_{\rm KIN}=E-m_p c^2$
223: of the parent proton that is transferred to the neutral pion
224: during the collision. For a photon/proton spectral index $ \alpha
225: \sim 2$ we get $N_\gamma(E_{\gamma})=1.12\times 10^{-25} K_p
226: E_{\gamma}^{-2}$ cm$^{-3}$ TeV$^{-1}$ (notice that 1 erg $\sim$ 1
227: TeV). Assuming typical grammages for the stellar winds of OB stars
228: in the range $\Sigma \sim 50-150$ g cm$^{-2}$ (e.g. White 1985),
229: we get $N_\gamma(E_{\gamma})=(3.4-10.1) K_p E_\gamma^{-2}$
230: cm$^{-3}$ TeV$^{-1}$ . The luminosity in the CANGAROO observing
231: range is
232: \begin{eqnarray}
233: L &=& 4 \pi R^2 c \int_{\sim 0.5{\rm TeV}}^{\sim 20{\rm TeV}}
234: N_\gamma(E_\gamma) E_\gamma dE_\gamma \\ \nonumber &=& (4.7-13.9)
235: \times 10^{12} R^2 K_p\; {\rm erg}\; {\rm s}^{-1},
236: \end{eqnarray}
237: where $R$ is the size of the base of the wind and $c$ is the
238: speed of light (all numerical coefficients are for cgs units). The
239: CR energy density is, on the other hand,
240: \begin{equation}
241: \omega_{\rm CR} = \int_{\sim 1{\rm GeV}}^{\sim 20{\rm TeV}}
242: N_p(E_p) E_p dE_p = 9.9 K_p\; {\rm erg}\; {\rm s}^{-1} \equiv
243: \varsigma \omega_{\rm CR,\odot},
244: \end{equation}
245: where $\varsigma$ is the amplification factor of the CR energy
246: density with respect to the local value, $ \omega_{\rm CR,\odot}\sim
247: 10^{-12}$ erg cm$^{-3}$. This implies that the TeV luminosity can
248: be written as $L=(0.47-1.4) R^2 \varsigma$ erg s$^{-1}$. The
249: latter corresponds to the luminosity of a single star illuminated
250: by TeV CR protons. In the central region of the starburst,
251: millions of stars could be illuminated at the same time. The total
252: luminosity would then be $ L_{\rm
253: total}=\sum_{i=1}^{N_\star} (0.47-1.4)\, R_i^2 \, \varsigma_i \;
254: {\rm erg}\; {\rm s}^{-1},
255: $ where $N_\star$ is the number of contributing
256: stars, and we have allowed for the possibility that not all of them will
257: have the same $R$, nor will they be located within regions having
258: the same CR enhancement factor.
259:
260: \begin{figure*}[t]
261: \centering
262: \includegraphics[width=0.3\textwidth,height=7cm]{f1a.eps}\hspace{2cm}
263: \includegraphics[width=0.3\textwidth,height=7cm]{f1b.eps}
264: \caption{Left (a): Results of 1000 Monte Carlo trials to compute
265: the total TeV luminosity of a starburst lying at the distance of
266: NGC 253. Right (b): The hadronic
267: model expectation of neutrinos from NGC 253 compared with the
268: upper limit for the neutrino background in a ${1^\circ \times
269: 1^\circ}$-generation neutrino telescope.} \label{fig2}
270: \end{figure*}
271:
272: If we now assume the following ranges for the different parameters
273: involved: $1\times 10^{12}<R[{\rm cm}]<1\times 10^{15}$ (Lamers
274: 1999), $300 < \varsigma < 3000$ (Suchkov et al. 1993), and
275: $10^6<N_\star<5 \times 10^7$ (see above), there is ample room for
276: the total luminosity to reach values above 10$^{39}$ erg s$^{-1}$,
277: i.e. the observed CANGAROO luminosity above 0.5 TeV, given the
278: distance to NGC 253 (2.5 Mpc, de Vacoulers 1978). To show this, we
279: have carried out Monte Carlo numerical simulations where the
280: number of stars illuminated by the accelerated protons, the
281: grammage, the size of the base of the wind for each star, and
282: the ambient enhancement of CRs are independently allowed to take values
283: within their assumed ranges. For each of these cases, the total
284: luminosity was computed. We show the results in Fig. 1a, for
285: a sample of 1000 trials. Only for less than 40\% of
286: the trials, the total luminosity is below 10$^{39}$ erg s$^{-1}$.
287: This scenario then appears to be a plausible explanation for the
288: TeV observations of NGC 253.
289:
290: Note that multi-TeV electrons in the starburst central region will
291: experience inverse Compton losses with the FIR fields (typical
292: energy density $\sim 200$ eV cm$ ^{-3}$) in the Klein-Nishina
293: regime, which result in a very steep $\gamma$-ray spectrum (that
294: can be approximated by power-law indices $\gtrsim 3.2$ (although
295: notice that the actual spectrum is not a pure power law, see e.g.
296: Georganopoulos et al. 2001). The total leptonic contribution from
297: the central starburst region at TeV energies is then negligible in
298: comparison to the hadronic one.
299:
300: MeV---GeV particles cannot easily penetrate the
301: stellar winds, and hence these winds will glow mainly at TeV
302: energies. The modulation of CRs by stellar winds is a complex
303: subject that has been studied in detail only in the context of the
304: relatively weak solar wind.
305: %The modulation includes a number of
306: %different effects like diffusion, convection, particle drifts,
307: %energy change, and terminal shock barriers.
308: In such case,
309: %solar wind,
310: three-dimensional models including diffusion and the effects of
311: the terminal shocks have been developed (e.g. K\'ota \& Jokipii
312: 1983, Jokipii et al. 1993). These models can explain fairly well
313: the basic observed features of the nucleonic CR component in the
314: solar system, where strong gradients in the CR density are known
315: to exist from interplanetary spacecraft measurements (e.g. Quenby
316: et al. 1990).\footnote{Although it is highly dependent of the solar
317: activity and the inclination angle of the incident particles, the
318: proton flux decreases to less than $\sim 30$\% of its initial
319: value at 80 A.U. for 2.5 GeV particles, and to less than $\sim
320: 10$\% for 250 MeV particles.}
321: In OB stars, with extremely
322: supersonic winds and mass loss rates orders of magnitude higher
323: than the solar values, these effects should be far more
324: pronounced, leading to an almost complete rejection of externally
325: injected MeV---GeV particles at the base of the winds and thus to
326: an absence of a strong (above EGRET sensibility) $\gamma$-ray
327: counterpart at these energies. The modulation effect depends on
328: the parameter $\epsilon \sim u R_{\rm w}/D$,
329: where $u$ is the
330: stellar wind velocity, $R_{\rm w}$ is the radius of the wind shell, and $D$ is the diffusion
331: coefficient. $\epsilon$
332: measures the ratio between the diffusive timescale of the particle to their convective timescale.
333: The former should be shorter than the latter for a TeV particle to enter into the wind cavity.
334: For a typical O star,
335: $R_{\rm w}> 10$ pc and $u\sim 2500$ km/s. Hence,
336: the diffusion coefficient of the TeV particles should be at least $10^{28}$ cm$^{2}$ s$^{-1}$,
337: similar to what is estimated for our Galaxy (e.g. Berezinskii et al. 1990).
338:
339: \section{Neutrino signal}
340:
341:
342: If the $\gamma$-ray emission is explained by the decay of neutral
343: pions, their charged counterparts must produce a neutrino flux.
344: Following Alvarez-Mu\~niz \& Halzen (2002), the $\nu_\mu +
345: \bar\nu_\mu$ neutrino flux $F_\nu(E_\nu)$ produced by the decay of
346: charged pions in the source can be derived from the observed
347: $\gamma$-ray flux $F{_\gamma}(E_\gamma)$ by imposing energy
348: conservation (see also Stecker 1979):
349: \begin{equation}
350: \int_{E_{\gamma}^{\rm min}}^{E_{\gamma}^{\rm max}} E_\gamma
351: F{_\gamma}(E_\gamma) dE_\gamma = K \int_{E_{\nu}^{\rm
352: min}}^{E_{\nu}^{\rm max}} E_\nu F_\nu(E_\nu) dE_\nu~.
353: \end{equation}
354: Here, ${E_{\gamma\; [\nu]}^{\rm min}}$ ($E_{\gamma \; [\nu]}^{\rm
355: max}$) is the minimum (maximum) energy of the photons (neutrinos)
356: and the pre-factor $K=1$.\footnote{Note that there are two
357: muon-neutrinos (out of the pion-to-muon-to-electron decay chain)
358: for every photon produced. Each of the latter carries, on average,
359: an energy $E_\pi/4$, whereas each photon carries on average half
360: the energy of the pion. Therefore, there is as much energy in
361: photons as there is in neutrinos, and $K=1$.}
362: %In $p\gamma$ interactions,
363: %instead, neutrinos are produced via the $\Delta$-resonance, for
364: %which 2/3 of the interactions give rise to a neutral pion which
365: %later decays into two photons with energies $E_\pi/2$. The
366: %remaining 1/3 of the interactions produce $\pi^+$, which later
367: %decay into 2 neutrinos of energy $E_\pi/4$. Therefore, in this
368: %case there is four times as much energy in photons than there is
369: %in neutrinos, and $K=4$.
370: The maximum neutrino energy is fixed by the maximum energy of the
371: accelerated protons ($E_p^{\rm max}$) which can in turn be
372: obtained from the maximum observed $\gamma$-ray energy
373: ($E_{\gamma}^{\rm max}$) as: $ E_p^{\rm max}\sim 6
374: E_{\gamma}^{\rm max}\;,\;\;E_{\nu}^{\rm max}\sim {1 / 12} E_p^{\rm
375: max} $ (Alvarez-Mu\~niz \& Halzen 2002).
376: %For the $p\gamma$ case it is just about 5\% of the maximum proton
377: %energy.
378: The minimum neutrino energy is fixed by the threshold for pion
379: production,
380: $
381: E_p^{\rm min}=\delta ~ {(2m_p+m_\pi)^2-2m_p^2 \over 2m_p},
382: \label{ethpp}
383: $
384: where $\delta$ is the Lorentz factor of the accelerator relative
385: to the observer.
386: %It is important to
387: %note that it is the assumption of a hadronic origin for the TeV
388: %emission what allows the extension of the $\gamma$-ray spectrum
389: %towards the lower integration limit. It is not true that this same
390: %spectrum should be seen at lower energies, however, since other
391: %processes, if they exist, may affect its shape.
392: %If the TeV emission is hadronically produced, protons must be
393: %accelerated up to $E_p^{\rm max}\sim 100$ TeV. $E_\nu^{\min}$ is
394: %obtained from the previous relationships
395: Assuming a Lorentz factor of order 1, the $\nu$-spectrum results
396: in $F_{\nu}(E_\nu)\sim 4.2 \times 10^{-12} (E/{\rm TeV})^{-1.85}$
397: TeV$^{-1}$ cm$^{-2}$ s$^{-1}$ for the harder photon index, and
398: $F_{\nu}(E_\nu)\sim 9.5 \times 10^{-13} (E/{\rm TeV})^{-3.75}$
399: TeV$^{-1}$ cm$^{-2}$ s$^{-1}$ for the softer photon index.
400:
401:
402:
403: %The signal-to-noise (S/N) ratio for the detection of this
404: %$\nu$-signal in a km-scale detector can be estimated as follows
405: %(see Anchordoqui et al. 2002b).
406: The event rate of atmospheric
407: $\nu$-background is (see Anchordoqui et al. 2002b)
408: \begin{equation}
409: \left. \frac{dN}{dt}\right|_{\rm B} = A_{\rm eff}\, \int dE_\nu
410: \,\frac{d\Phi_{\rm B}}{dE_\nu}\, P_{\nu \to \mu}(E_\nu)\,\,\Delta
411: \Omega_{1^\circ \times 1^\circ}\,, \label{background}
412: \end{equation}
413: where $A_{\rm eff}$ is the effective area of the detector, $\Delta
414: \Omega_{1^\circ \times 1^\circ} \approx 3 \times 10^{-4}$~sr is an
415: assumed bin size for the observation (ICECUBE-generation detector,
416: Karle et al. 2002), and $d\Phi_{\rm B}/dE_\nu \lesssim 0.2
417: \,(E_\nu/{\rm GeV})^{-3.21}$~GeV$^{-1}$ cm$^{-2}$ s$^{-1}$
418: sr$^{-1}$ is the $\nu_\mu + \bar \nu_\mu$ atmospheric
419: $\nu$-flux~(Volkova 1980, Lipari 1993). Here, $P_{\nu \to \mu}
420: (E_\nu) \approx 1.3 \times 10^{-6} \,(E_\nu/{\rm TeV)}^{0.8}$
421: denotes the probability that a $\nu$ of energy $E_\nu
422: > 1~{\rm TeV}$, on a trajectory through the detector, produces a
423: muon~(Gaisser et al. 1995). On the other hand, the $\nu$-signal is
424: \begin{equation}
425: \left. \frac{dN}{dt}\right|_{\rm S} = A_{\rm eff} \,\int dE_\nu
426: \, F_{\nu}(E_\nu)\,P_{\nu \to \mu}(E_\nu)\ \,.
427: \label{yellowsubmarine}
428: \end{equation}
429: The signal-to-noise (S/N) ratio in the 1--20 TeV band is then $(
430: \left. {dN}/{dt}\right|_{\rm S})/(\sqrt{\left.
431: {dN}/{dt}\right|_{\rm B}})=0.1-2.5$ for one year of operation,
432: depending on the photon spectral index. A NEMO-generation
433: detector, with a bin size $\Delta \Omega\sim {0.3^\circ \times
434: 0.3^\circ}$ (Riccobene et al. 2002), would have a S/N ratio
435: roughly three times larger for the same integration time. The
436: effect of neutrino oscillations may reduce the flux by a factor of
437: three at Earth (Bilenky et al. 1999). However, for some possible
438: values of the photon spectral index, this signal could be detected
439: in a reasonable time span. The possible starburst neutrino spectra
440: producing the previously quoted S/N ratio are plotted, together
441: with an upper limit for the atmospheric neutrino background with
442: vertical incidence, in Fig.~1b.\footnote{Regrettably, the south
443: pole location of the ICECUBE telescope would render its
444: observation of NGC 253 nearly impossible; only neutrinos going
445: through Earth can be identified. The smaller northern telescope
446: ANTARES (0.1 km$^2$, Becherini et al. 2002) would need much longer
447: integration times to be able to distinguish NGC 253 from the
448: background, if at all possible. Nevertheless, the analysis
449: presented here can be directly applied to M82, a northern-sky
450: starburst at the same distance as NGC 253.}
451: %Starbursts are strong potential sources for high energy
452: %neutrinos; their detection can provide definite evidence of the
453: %main mechanism responsible for the high energy emission.
454:
455:
456: \section{Concluding remarks}
457:
458: Through hadronic interactions in their central regions, nearby
459: starburst galaxies can generate high-energy $\gamma$-ray emission
460: detectable at Earth. Proton illumination of the inner winds of
461: massive stars in a high density CR environment may result in TeV
462: sources without strong counterparts at MeV---GeV energies. The
463: high number of early-type stars with strong stellar winds in this
464: kind of galaxies might produce a strong collective effect,
465: providing the setting for accelerating protons up to multi-TeV
466: energies. In addition, we have found that starbursts might also be
467: potential sources for km-scale neutrino telescopes within
468: reasonable integration times, depending on the photon spectral
469: index. The detection ---or non-detection--- of the neutrino signal
470: may be an essential piece of evidence in the determination of the
471: origin of the high-energy radiation.
472:
473: Gamma-rays from neutral pion decays will be produced not only in
474: the stellar winds at the core of the starburst, but also in the
475: diffuse matter among the stars. If we model the starburst region
476: as a disk with a radius of $\sim300$ pc and a thickness of
477: $\sim70$ pc, where the average density of H atoms in the diffuse
478: ISM is $\sim 300$ cm$^{-3}$ (Paglione et al. 1996), we get that
479: the expected flux above 1 TeV is $F(E>1\;{\rm TeV})\simeq 3.9
480: \times 10^{-13}$ ph cm$^{-2}$ s$^{-1}$, for an average CR
481: enhancement of $\varsigma\sim1000$ (see Torres et al. 2002 for
482: details of calculation). This means an additional diffuse
483: contribution of $\sim3\times10^{38}$ erg s$^{-1}$ to the total TeV
484: $\gamma$-ray luminosity of the starburst. In most of our models
485: this diffuse contribution is only a small fraction of the total
486: $\pi^0$-decay $\gamma$-ray emission, which is dominated by
487: interactions occurring in the stellar winds.
488:
489: Finally, it is important to remind that the CANGAROO observations
490: are consistent with an extended source, whose size might be even
491: bigger than the galaxy itself. If this is true, it would imply
492: that there might be another component generating TeV radiation,
493: related with the galactic halo. This component might probably be
494: the result of Inverse Compton scattering of CMB photons by TeV
495: electrons accelerated at the superwind. In such case, since the
496: interactions occur in the Thomson regime, the spectrum should be
497: harder than that produced by hadrons in the center of the
498: starburst.\footnote{After the submission of this paper we were
499: informed that Itoh et al. (2003) discussed in detail this point.
500: We thank Prof. Yanagita for calling our attention to this.}
501:
502:
503:
504:
505: \acknowledgements
506:
507: Comments by L. Anchordoqui, P. Benaglia, Y.M. Butt, C. Mauche,
508: M.M. Kaufman-Bernad\'o, M. Pohl, and an anonymous referee are
509: gratefully acknowledged. G.E.R.'s research is mainly supported by
510: Fundaci\'on Antorchas. Additional support was provided by CONICET
511: (under grant PIP N$^o$ 0430/98) and ANPCT (PICT 03-04881). The
512: work of D.F.T. was performed under the auspices of the U.S. DOE
513: (NNSA), by UC's LLNL under contract No. W-7405-Eng-48. His
514: research was done while he was a Visiting Scientist at SISSA
515: (Italy) and a Visiting Professor at the University of Barcelona
516: (Spain), to which he gratefully acknowledge their hospitality.
517:
518: \begin{thebibliography}{}
519:
520:
521: \bibitem{} Aharonian, F., et al. 2002, A\&A 393, L37
522:
523: \bibitem{} Alvarez-Mu\~niz, J., \& Halzen, F. 2002, ApJ 576, L33
524:
525: \bibitem{}Anchordoqui, L.A., Romero, G.E., \& Combi, J.A. 1999,
526: Phys. Rev. D60, 103001
527:
528: \bibitem{}Anchordoqui, L.A., Goldberg, H., \& Torres, D.F. 2002,
529: astro-ph/0209546
530:
531: \bibitem{} Anchordoqui, L.A., et al. 2002b, ApJ, in press,
532: hep-ph/0211231
533:
534: \bibitem{} Bhattacharya, D., et al. 1994, ApJ 437, 173
535:
536: \bibitem{} Blom, J.J., Paglione, T.A., Carrami\~nana, A. 1999,
537: 516, 744
538:
539: \bibitem{} Benaglia, P., Romero, G.E., Stevens, I.R., \&
540: Torres, D.F. 2001 A\&A 366, ~605
541:
542: %\cite{Becherini:2002wg}
543: \bibitem{Becherini:2002wg}
544: Becherini, Y., et al. [ANTARES Collaboration] 2002,
545: %``The ANTARES neutrino telescope,''
546: arXiv:hep-ph/0211173.
547: %%CITATION = HEP-PH 0211173;%%
548:
549: \bibitem{beck} Beck, R., Carilli, C.L., Holdaway, M.L., \& Klein, U. 1994,
550: A\&A 292, 409
551:
552: \bibitem{} Berezinskii, V.S., Bulanov, S.V., Dogiel, V.A., Ginzburg, V.L.,
553: \& Ptuskin, V.S. 1990, ``Astrophysics of Cosmic-Rays'', Amsterdam,
554: North Holland, Ch. 9
555:
556:
557: %\cite{Bilenky:1998dt}
558: \bibitem{Bilenky:1998dt}
559: Bilenky S.M., Giunti C., \& Grimus W. 1999,
560: %``Phenomenology of neutrino oscillations,''
561: Prog.\ Part.\ Nucl.\ Phys.\ 43, 1.
562: %[arXiv:hep-ph/9812360].
563: %%CITATION = HEP-PH 9812360;%%
564:
565: \bibitem{}Butt, Y. M., et al. 2003, in preparation.
566:
567: \bibitem{} Bykov, A.M. \& Fleishman, G.D. 1992a, MNRAS 255, 269
568:
569: \bibitem{} Bykov, A.M. \& Fleishman, G.D. 1992b, Sov. Astron. Lett. 18, 95
570:
571: \bibitem{} Bykov, A.M. 2001, Space Sci. Rev. 99, 317
572:
573: \bibitem{} Cass\'{e}, M., \& Paul, J.A. 1980, ApJ 237, ~236
574:
575: \bibitem{} Chen, W., White, R.L., \& Bertsch, D. 1996, A\&AS 120, 423
576:
577: \bibitem{} Clark J. S., \& Negueruela I. 2002,
578: A\&A 396, L25
579:
580: \bibitem{} de Vacoulers, G. 1978, ApJ 224, 710
581:
582: \bibitem{forbes} Forbes, D.A., et al. 1993, ApJ 406, L11
583:
584: \bibitem{} Gaisser, T.K. 1990, ``Cosmic Rays and Particle Physics",
585: Cambridge University Press, Cambridge
586:
587: \bibitem{Gaisser:1994yf}
588: Gaisser, T.K., Halzen, F., \& Stanev, T. 1995,
589: %``Particle astrophysics with high-energy neutrinos,''
590: Phys.\ Rept.\ { 258}, 173 [Erratum-ibid.\ { 271}, 355 (1996)].
591: %[arXiv:hep-ph/9410384].
592: %%CITATION = HEP-PH 9410384;%%
593:
594: \bibitem{}Georganopoulos, M., Kirk, J. G., \& Mastichiadis, A. 2001,
595: ApJ 561, 111
596:
597: %\bibitem{} Hartman R.C., et al. [EGRET Collaboration] 1999, ApJS 123, 79
598:
599: \bibitem{heckman} Heckman, T.M., Armus, L., \& Miley, G.K. 1990, Ap. J. Suppl.
600: 74, 833
601:
602: %\bibitem{}
603: %Hill G.C. et al. [the AMANDA Collaboration] 2001,
604: %``Results from AMANDA,''
605: %astro-ph/0106064.
606: %%CITATION = ASTRO-PH 0106064;%%
607:
608: \bibitem{}Itoh, C., et al. [CANGAROO Collaboration] 2002, A\&A 396, L1
609:
610: \bibitem{}Itoh, C., et al. 2003, astro-ph/0301147, submitted to
611: ApJ Lett.
612:
613: \bibitem{} Jokipii, J.R., K\'ota, J., \& Mer\'enyi, E. 1993, ApJ 405,
614: 782
615:
616:
617: %\cite{Karle:2002dv}
618: \bibitem{}Karle A., et al. [the IceCube Collaboration] 2002,
619: %``IceCube: The next generation neutrino telescope at the South Pole,''
620: arXiv:astro-ph/0209556.
621: %%CITATION = ASTRO-PH 0209556;%%
622:
623:
624: \bibitem{} Keto E., Hora J.L., Fazio G.G., Hoffman W., \& Deutsch
625: L. 1999, ApJ 518, 133
626:
627: \bibitem{} K\'ota, J., \& Jokipii, J.R. 1983, ApJ 265,
628: 573
629:
630: \bibitem{} Lamers, H.J.G.L.M., \& Cassinalli, J.P. 1999,
631: ``Introduction to Stellar Winds", Cambridge University Press,
632: Cambridge
633:
634: \bibitem{} Lang, K.R. 1999, Astrophysical Fromulae Vol. I,
635: Springer, Berlin, p.434
636:
637: \bibitem{}
638: Lipari, P. 1993,
639: %``Lepton Spectra In The Earth's Atmosphere,''
640: Astropart.\ Phys.\ 1, 195
641: %%CITATION = APHYE,1,195;%%
642:
643:
644: %\bibitem{} Longair M.S. 1994, ``High Energy Astrophysics, Vol.2:
645: %Stars, the Galaxy and the Interstellar Medium", Cambridge
646: %University Press, 2nd ed.
647:
648:
649: \bibitem{mccarthy} McCarthy, P.J., Heckman, T., \& van Breugel, W. 1987,
650: AJ 92, 264
651:
652:
653: \bibitem{} Montmerle, T. 1979 ApJ 231, ~95
654:
655:
656: \bibitem{paglione} Paglione, T.A.D., Marscher, A.P., Jackson, J.M., \&
657: Bertsch D.L. 1996, ApJ 460, 295
658:
659: \bibitem{} Quenby, J.J., Lockwood, J.A., Webber, W.R. 1990, ApJ
660: 365, 365
661:
662: \bibitem{ptak} Ptak, A., et al. 1997, AJ 113, 1286
663:
664:
665: %%\cite{Riccobene:2002xk}
666: \bibitem{Riccobene:2002xk} Riccobene, G., et al. [NEMO Collaboration] 2001,
667: %``Status of the NEMO project,''
668: Hamburg DESY - DESY-PROC-2002-01. Published in ``Hamburg 2001,
669: Methodical aspects of underwater/ice neutrino telescopes'', p.61.
670:
671:
672: \bibitem{rieke} Rieke, G.H., et al. 1980, Ap. J. 238, 24
673:
674: \bibitem{} Romero G.E., Benaglia P., Torres D.F. 1999 A\&A 348, ~868
675:
676: \bibitem{} Torres D.F., et al. 2002, astro-ph/0209565
677:
678: \bibitem{} Suchkov A., Allen R.J., \& Heckman T.M. 1993, ApJ 413,
679: 542
680:
681: \bibitem{}Stecker, F.W. 1979, ApJ, 228, 919
682:
683: \bibitem{ulvestad} Ulvestad J.S., \& Antonucci R.R.J. 1997, ApJ
684: 488, 621
685:
686: %\bibitem{} V\"olk H.J., \& Forman M. 1982 ApJ 253, ~188
687:
688:
689: \bibitem{Volkova:sw}
690: Volkova L.V. 1980,
691: %``Energy Spectra And Angular Distributions Of Atmospheric Neutrinos,''
692: Sov.\ J.\ Nucl.\ Phys.\ { 31}, 784 [Yad.\ Fiz.\ { 31}, 1510
693: (1980)].
694: %%CITATION = SJNCA,31,784;%%
695:
696: \bibitem{} Watson A.M., et al. 1996, AJ 112, 534
697:
698:
699: \bibitem{} White R.L. 1985, ApJ 289, 698
700: \bibitem{} White R.L., \& Chen W. 1992, ApJ 387, ~L81
701:
702: \end{thebibliography}
703:
704: %\clearpage
705:
706: \end{document}
707: