1: %\documentstyle[aasms4,amstex,amsfonts,epsfig,rotating,float,12pt]{article}
2: \documentclass[12pt,preprint]{aastex}
3: %\documentclass{aastex}
4:
5: \newcommand{\ds}{\displaystyle}
6: \newcommand{\be}{\begin{equation}}
7: \newcommand{\ee}{\end{equation}}
8: \newcommand{\rmin}{ {R_{\rm min} }}
9: \newcommand{\cross}{ {\langle \sigma \rangle}}
10: \newcommand{\sigwig}{ {\widetilde \sigma}}
11: \newcommand{\dwig}{ {\widetilde D}}
12:
13: \begin{document}
14:
15: \title{High-Energy Activity in the Unusually Soft TeV Source \\
16: HESS J1804-216 toward the Galactic Center}
17:
18: \medskip
19: \author{Marco Fatuzzo$^1$, Fulvio Melia$^2$, and Roland M. Crocker$^{3}$}
20: \bigskip
21: \affil{$^1$Physics Department, Xavier University, Cincinnati, OH 45207}
22: \affil{$^2$Physics Department and Steward Observatory,
23: The University of Arizona, AZ 85721}
24: \affil{$^3$ School of Chemistry and Physics, The University of Adelaide,
25: South Australia 5005}
26:
27: \begin{abstract}
28: In recent years, apparent anisotropies in the ~EeV cosmic ray (CR) flux
29: arriving at Earth from the general direction of the galactic center have
30: been reported from the analysis of AGASA and SUGAR data. The more recently
31: commissioned Auger Observatory has not confirmed these results. HESS has now
32: detected an unusually soft TeV source roughly coincident with the location of the
33: previously claimed CR anisotropy. In this paper, we develop a model for the
34: TeV emission from this object, consistent with observations at other wavelengths,
35: and examine the circumstances under which it might have contributed to the
36: $\sim$ EeV cosmic ray spectrum. We find that the supernova remnant
37: G8.7-0.1 can plausibly account for all the known radiative characteristics
38: of HESS J1804-216, but that it can accelerate cosmic rays only up to an
39: energy $\sim 10^5$ GeV. On the other hand, the pulsar (PSR J1803-2137)
40: embedded within this remnant can in principle inject EeV protons into
41: the surrounding medium, but it cannot account for the broadband
42: spectrum of HESS J1804-216. We therefore conclude that although G8.7-0.1
43: is probably the source of TeV photons originating from this direction, there
44: is no compelling theoretical motivation for expecting a cosmic ray anisotropy
45: at this location. However, if G8.7-0.1 is indeed correctly identified
46: with HESS J1804-216, it should also produce a $\sim$ GeV flux detectable in
47: a one-year all sky survey by GLAST.
48: \end{abstract}
49:
50: \keywords{acceleration of particles---cosmic rays---Galaxy: center---galaxies:
51: nuclei---radiation mechanisms: nonthermal---supernova remnants}
52:
53: \section{Introduction}
54: The High Energy Stereoscopic System (HESS) array provides sensitivity to gamma
55: rays with energy $>100$ GeV at a level
56: below $1\%$ of the flux from the Crab Nebula, with an angular resolution for
57: individual photons better than $0.1^\circ$. Thus, the position of even relatively
58: faint sources may be determined with an error of only $30$ sec. A scan of the
59: inner $60^\circ$ of the galactic plane has identified fourteen discrete TeV-emitting
60: sources, half of which have plausible identifications at other wavelengths (Aharonian
61: et al. 2005). HESS J1804-216 is the source at galactic coordinates $l=8.40^\circ$ and
62: $b=-0.033^\circ$, with a TeV-size of $\approx 22$ arcmin. Its estimated flux above
63: 200 GeV is $5.3\times 10^{-11}$ cm$^{-2}$ s$^{-1}$, with a statistical error of
64: $10$ to $35\%$. Its spectral (power-law) index is $-2.72\pm0.06$, the steepest of
65: all the TeV sources in this survey (Aharonian et al. 2006).
66:
67: This source coincides with the southwestern rim of the shell-type SNR G8.7-0.1
68: (W30), whose radio-emitting radius has been set at $\approx 26$ arcmin (see,
69: e.g., Handa et al. 1988). Its radio flux at $1$ GHz may be derived from the Green
70: (2004) catalog, assuming a radio spectral index $\alpha\sim-0.65$, and integrating
71: from $10^7$ to $10^{11}$ Hz (Helfand et al. 2005). Such an estimate yields a
72: radio flux for G8.7-0.1 of $1.1\times 10^{-11}$ ergs cm$^{-2}$ s$^{-1}$.
73:
74: From CO observations, it is known that the W30 complex is surrounded by molecular
75: gas (Blitz, Fich, \& Stark 1982), in which new stars are forming. By associating
76: G8.7-0.1 with coincident H II regions whose distances are known, Kassim \& Weiler
77: (1990) have estimated a distance to the SNR of $6\pm 1$ kpc which, combined with
78: its angular size of $\sim 50$ arcmin, implies a physical size of $\sim 80$ pc.
79: Thus, if HESS J1804-216 is indeed associated with this SNR, its location is not
80: quite at the galactic center (i.e., at a distance of $\approx 8.5$ kpc).
81:
82: G8.7-0.1 may also be linked with the (relatively) young pulsar PSR J1803-2137,
83: an association suggested by their coincidence on the sky, and by the observed
84: dispersion measure, which points to a distance of $\sim 5.3$ kpc (Clifton \&
85: Lyne 1986). Both sources were observed with the Position Sensitive Proportional
86: Counter at the focus of ROSAT (Finley \& \"Ogelman 1994), and both were detected
87: in the soft X-ray energy band $0.1-2.4$ keV. For G8.7-0.1, the unabsorbed flux
88: in this range of wavelengths is estimated to be $\sim (1-3)\times 10^{-10}$
89: ergs cm$^{-2}$ s$^{-1}$, corresponding to a luminosity $(0.4-1.3)\times
90: 10^{36}$ ergs s$^{-1}$ (for a distance of $6$ kpc). On the other hand, there is
91: no known detection of either source at $\sim$ GeV energies, implying an upper
92: limit to their EGRET (i.e., $100$ MeV to $\sim 30$ GeV) flux of $\sim 4\times
93: 10^{-8}$ cm$^{-2}$ s$^{-1}$ (Hartman et al. 1999).
94:
95: Several of these new TeV sources have raised important questions concerning their
96: origin. HESS J1804-216, in particular, is characterized by an unusually steep
97: gamma ray spectrum, yet it was not detected by EGRET in the GeV range, as one
98: might naively expect based on an extrapolation of the HESS data to lower energies.
99: Clearly, the spectrum cannot be a simple power law, and must display some interesting
100: physics between GeV and TeV energies, possibly shedding some light on the connection
101: between particle acceleration, diffusion, and emissivity. Addressing these questions
102: is one of the principal goals of this paper.
103:
104: But interest in the high-energy activity in this region of the Galaxy extends
105: beyond just these issues (see, e.g., Melia and Falcke 2001). The analysis,
106: in recent years, of data from two different
107: cosmic ray detectors has suggested the presence of an anisotropic overabundance of
108: cosmic rays coming from the general direction of the galactic center at energies
109: around an EeV ($10^{18}$ eV). Statistically, the most robust determination for an
110: anisotropy has been made by the Akeno Giant Shower Array (AGASA) Group
111: \citep{Hayashida1999}, which found a strong---4\% amplitude---anisotropy in the
112: energy range $10^{17.9} - 10^{18.3}$ eV. Two-dimensional analysis of the data showed
113: that this anisotropy could be interpreted as an excess of air showers from two regions
114: each of $\sim 20^\circ$ extent, one of 4$\sigma$ significance near the galactic center and
115: another of 3$\sigma$ in Cygnus. Interestingly, AGASA also saw a CR deficit towards
116: the Galactic anti-center.
117:
118: Prompted by this result, Bellido et al. (2001) re-analyzed the data collected by
119: the SUGAR cosmic ray detector, which operated near Sydney from 1968 to 1979.
120: They confirmed the existence of an anisotropy, consistent with a point source
121: located 7.5$^\circ$ from the galactic center---and 6$^\circ$ degrees from the
122: AGASA maximum over an energy range of $10^{17.9} - 10^{18.5}$ eV.
123:
124: Not surprisingly, this tentative result has generated some intense theoretical
125: interest because although it has long been speculated that diffusive shock
126: acceleration of protons and ions at shock fronts associated with supernova
127: remnants (SNRs) is the mechanism likely responsible for energizing the bulk of
128: high energy cosmic rays, the observational evidence for this has been elusive.
129: In addition, the conditions at almost all known SNRs seem not to promote the
130: acceleration of CRs beyond the `knee' feature in the spectrum, at $\simeq 5
131: \times 10^{15}$ eV. Thus, the origin of CRs between the knee and the `ankle'
132: at few $\times 10^{19}$ eV has been a mystery.
133:
134: In a series of earlier papers (Crocker et al. 2005a, 2005b; see also Bossa et al.
135: 2003, Aharonian and Neronov 2005, and the original discussion in Markoff et al. 1997,
136: 1999), we explored in detail the physics of particle acceleration and radiative emission
137: in the central few parsecs of the Galaxy to examine the extent to which detectors on Earth
138: would be able to sense neutrinos originating at the galactic center, and to understand how a cosmic
139: ray anisotropy might fit into the overall astroparticle scheme associated with this
140: dynamic region of the Galaxy.
141:
142: But there are several serious difficulties that a comprehensive and self-consistent
143: model of the energetic particle activity at the galactic center must overcome (Melia \& Falcke
144: 2001; Melia 2006). For example, not all of the data can be accommodated within a solitary
145: framework, at least not one focusing on the nucleus itself; indeed, the data are inconsistent
146: amongst themselves in two important instances: (i) the SUGAR results indicate a point
147: source offset by $7.5^\circ$ (toward positive galactic latitude) from the galactic center, in
148: (at least partial) disagreement with AGASA, and (ii) the $\sim$ TeV and (EGRET)
149: $\gamma$-ray emissivities are not simple extensions of a single spectrum.
150:
151: Even at $\sim$ EeV energies, charged particles would not be able to reach Earth
152: directly without being significantly deflected by the intervening magnetic field. A
153: consensus has developed that, if real, this CR anisotropy would be caused by
154: neutron emission at the galactic center, an idea first mooted by Jones (1990).
155: The anisotropy `turn on' at a definite energy of $\sim$ EeV finds a natural
156: explanation in the fact that this energy corresponds to a Lorentz factor for
157: neutrons large enough that they can reach us from the center of the Milky Way.
158: Neutrons below this energy decay in transit and are then diverted by galactic
159: magnetic fields. Above $\sim 10^{18.4}$ eV, the anisotropy ceases due to either
160: a very steep galactic center source spectrum or an actual cut-off in the source
161: so that the background takes over again at this energy.
162:
163: Note, moreover, that the galactic center, with declination $\delta = -28.9^\circ$,
164: is outside the
165: field of view of AGASA (which is limited to $\delta > -24^\circ$; Bossa et al. 2003),
166: but not so for SUGAR. Thus, the discrepancy between the two source positionings
167: may simply be due to the fact that AGASA is seeing protons produced during in-flight
168: neutron decay, whereas SUGAR sees the neutron source directly. That the SUGAR
169: anisotropy is not coincident with the galactic center, however, still presents a challenge
170: to all scenarios that would posit an EeV source right at the nucleus. Either, all
171: such models are incorrect or the SUGAR directional determination is somewhat in error.
172: If not the galactic center, then the SUGAR anisotropy could be due to another source
173: displaced from the galactic center by approximately $8^\circ$ toward positive galactic latitude.
174:
175: Another serious problem with the data sets is that whereas the galactic center is outside the
176: field of view of AGASA, the position of the SUGAR maximum is inside the AGASA
177: field of view so that the putative SUGAR source should be seen by AGASA. Of course,
178: one way out of this dilemma is that the source may have varied between the SUGAR
179: and AGASA observation times; this might occur, for example, if the source of
180: energetic particles were a pulsar.
181:
182: One cannot ignore the fact that the new TeV source, HESS J1804-216, coincides
183: with the previously cataloged SNR, G8.7-0.1, at about the position where SUGAR
184: detected the source of CR anisotropy. Could this be the smoking gun that finally
185: provides the observational evidence of a link between EeV particle acceleration and
186: a known source?
187:
188: Unfortunately, such an evaluation is not quite straightforward. The Pierre Auger Observatory
189: has now acquired enough data to search for an excess of events near the direction of the
190: galactic center in several energy bands around an EeV. With the accumulated statistics---1155
191: events, compared with an expected number of $1160.7$ for the energy range $(1.0-2.5)$
192: EeV---already larger than that of any earlier experiment, including AGASA and SUGAR (the
193: event number of AGASA in this region is only $1/3$ of this), the Auger
194: Observatory does not confirm the previously reported CR anisotropy (Letessier-Selvon
195: et~al. 2005). This raises several pertinent questions that we wish to address in
196: this paper. (1) Though unlikely, could it be that Auger is wrong, and that the
197: TeV characteristics of HESS J1804-216 would support the view that SUGAR (and possibly
198: AGASA) are in fact correct? Finding a model of HESS J1804-216 as the source of TeV gamma
199: rays, not violating observed flux limits at other wavelengths, and producing $\sim$ EeV
200: particles, could lend important support to this viewpoint. (2) On the other hand, if Auger
201: is correct, and no CR anisotropy is evident from the galactic center, then what is the
202: nature of HESS J1804-216, and can one account for it using more `conventional' means,
203: i.e., by identifying it as a member of a known class of object? (3) Finally, how likely
204: is it that HESS J1804-216 may have produced variable EeV emission over the past
205: $\sim 30$ years? In the next section, we shall develop viable models of its high-energy
206: activity, and then discuss its role in a broader context in \S\ 3.
207:
208: \newpage
209: \section{Two Possible Sources for the TeV Photons}
210:
211: \centerline{\sl 2.1 The Pulsar PSR J1803-2137}
212: Young, rapidly spinning pulsars have long been viewed as viable sources of cosmic rays,
213: possibly capable of accelerating hadrons to energies greater than $\sim 10^{20}$ eV
214: (Blasi et al. 2000; Arons 2003). We consider here the possibility that relativistic
215: particles produced by PSR J1803-2137 lead to a pionic-decay photon emissivity,
216: accounting for the HESS J1804-216 source. We also ascertain the likelihood that
217: PSR J1803-2137 is the source of excess cosmic rays reported by SUGAR and AGASA.
218:
219: Following Blasi et al. (2000) and Arons (2003), we consider the acceleration of
220: charged particles across voltage drops in the relativistic winds near the light cylinder
221: of young, rapidly rotating neutron stars. These particles do not suffer from
222: the radiation losses that limit their polar cap or outer gap counterparts to energies
223: $\sim 10^{16}$ eV (see, e.g., Arons 2003); as such, they can reach a maximum energy of
224: \be
225: E_{\rm max} (\Omega_0) \approx \eta\, Ze\, \Phi_{\rm wind} = \eta\, Ze {\Omega_0^2\, \mu\over
226: c^2} = 3\times 10^{18}\, Z \left({\eta\over 0.1}\right) \left({\Omega_0\over
227: 10^4\; \hbox{s}^{-1}}\right)^2 \left({B_*\over 10^{12}\;\hbox{G}}\right) \hbox{eV},
228: \ee
229: where $\eta$ is the fraction of the voltage drop, $\Phi_{\rm wind}$, experienced
230: by the charges, $B_*$ is the surface magnetic field strength (giving rise to
231: a magnetic moment $\mu\equiv B_*R_*^3$, in terms of the stellar radius $R_*$) and
232: $\Omega_0$ is the initial pulsar rotation frequency. In the case of PSR J1803-2137,
233: its measured spin-down age (Kassim and Weiler 1990) implies a magnetic field strength
234: $B_*\approx 8.6\times 10^{12}$ G, and therefore $E_{max} \sim 24$ EeV.
235:
236: As rotational energy is transferred to the particles, the pulsar spins down, and the
237: injected particle energy drops. As a result, the particles accelerated over the pulsar's
238: spin-down lifetime produce a power-law distribution
239: \be
240: N(E) = {9\over 4} {c^2 I \over Ze\mu} E^{-1}\;,
241: \ee
242: where $I$ is the neutron star's moment of inertia. The fact that PSR J1803-2137 has a
243: present day period of 133 ms, means that the distribution in this case extends down in
244: energy only as far as $E_{min}\approx 6\times 10^{5}$ GeV, estimated from Equation (1) by
245: setting $\Omega_0$ to its current value.
246:
247: The interaction of the accelerated, power-law (i.e., $N(E) = N_0 E^{-1}$) particles,
248: assumed here to be protons, with an ambient medium of density $n_p$, leads to
249: the production of pions and, subsequently, photons from the decay of neutral pions
250: and secondary lepton emission. Since Equation (2) directly yields the number of
251: relativistic protons injected into the surrounding medium once the magnetic field
252: strength is specified, the cascade-induced emissivity depends solely on the properties
253: of the ambient medium, e.g., its number density. (A full description of the pp induced
254: particle cascade and resulting broadband radiative emissivity is provided in Fatuzzo \& Melia
255: [2003; hereafter FM03] and Crocker et al. [2005a], and will therefore not be reproduced
256: in detail here.\footnote{It is worth pointing out, though, that FM03 use the empirical
257: fits to the accelerator data based on a hybrid isobar/scaling model (Dermer 1986a,b).
258: These fits are similar, but slightly different, from the cross sections used in
259: other SNR treatments (see, e.g., Drury et al. 1994; Gaisser et al. 1998; Baring et al.
260: 1999).}) However, the rate at which these cosmic rays diffuse out of the scattering region
261: depends on their rigidity and therefore on their energy. As a result, the energy distribution
262: of the particles remaining in the interaction zone will differ from the simple scaling
263: implied by Equation (2) and, moreover, this distribution evolves in time as the diffusion
264: process differentiates between particles of different energy.
265:
266: But one does not need to analyze the diffusion process in great detail to see that
267: PSR J1803-2137 could not be the source of TeV photons in HESS J1804-216.
268: The reason is that $E_{min}>> 1$ TeV, and regardless of how
269: the energy-dependent diffusion modifies the cosmic ray distribution at the source,
270: the photon spectrum below $E_{min}$ would have an index $\approx -1$, at odds with the
271: observed value $-2.7$. One can see this in Figure~1, where we show the calculated photon
272: spectrum in comparison with the TeV data, under the most favorable assumption that diffusion
273: could somehow produce a proton distribution with index $-2.7$. The photon emissivity below
274: $E_{min}$ is dominated by the $\pi^0$ cascade initiated by cosmic rays in the energy
275: range $\sim 6\times 10^5-6\times 10^6$ GeV, and this is true for all proton power-law
276: indices $<0$. It therefore appears that under all circumstances, the contribution of
277: PSR J1803-2137 to the TeV photon spectrum is too flat to account for the HESS data.
278:
279: Even so, PSR J1803-2137 could still account for the putative cosmic ray anisotropy
280: at $\sim$ EeV energies, without contributing measurably to the observed TeV flux, if
281: the proton spectral index is $\sim -1$ (see Crocker et al. 2005a). It should be noted,
282: however, that the production of neutrons within this scheme results from charge exchange
283: in $pp$ scattering. A present day neutron production associated with particles
284: accelerated by PSR J1803-2137
285: would then require that the diffusion time for EeV protons (which would have been injected
286: very early on by the pulsar) out of the surrounding region could not be much less than
287: the pulsar age, and that the surrounding region have a density below $10$ cm$^{-3}$
288: so as to not violate the observed HESS data, as shown in Figure 1 (note that the
289: photon emissivity scales directly with the ambient density).
290:
291: \bigskip
292: \centerline{\sl 2.2 The SNR G8.7-0.1}
293: The expansion of an SNR into a dense molecular cloud
294: can produce shock accelerated protons whose interactions with that medium
295: (via pp scattering) may also lead to an observable pionic decay signal above $\sim 100$
296: MeV (Drury et al. 1994; Sturner et al. 1997). Indeed, this mechanism has been invoked
297: to account for the association of several EGRET sources with SNRs interacting with their
298: dense surroundings (see, e.g., Gaisser et al. 1998; Baring et al. 1999; Berezhko \& V\"olk
299: 2000; Fatuzzo \& Melia 2005).
300:
301: In this section, we consider whether the broadband emission powered by SNR shocks
302: in G8.7-0.1 can account for the HESS J1804-216 source without violating the
303: flux limits observed at other wavelengths. We assume that shock acceleration
304: within the SNR environment injects a power-law distribution of relativistic
305: protons with a spectral index $\alpha = 2.0-2.4$ and maximum energy $E_{max}$ into
306: a dense ($500$ cm$^{-3}$) shell at the SNR-cloud boundary (e.g., Chevalier 1999;
307: Fatuzzo \& Melia 2005). The value of $E_{max}$ can be approximated by taking the
308: product of the remnant's age with the energy-gain rate for particles in a shock,
309: \begin{equation}
310: \dot E(t)=10^8\,{B\,v_8^2(t)\over fR_J}\;\hbox{eV}\;{\hbox{s}}^{-1}\;,
311: \end{equation}
312: where $B$ ($\approx 10^{-5}$ G) is the magnetic field strength, $v_8$ is the
313: shock velocity in units of $10^8$ cm s$^{-1}$, $f\sim 10$ is the particle
314: mean free path along the magnetic field in units of its gyroradius, and
315: $R_J\sim 1$ is a factor that accounts for the orientation of the shock relative
316: to the magnetic field (Sturner et al. 1997). For the SNR age $T_{SNR}\sim 15,000$ years
317: (Kassim and Weiler 1990), and canonical values of the parameters, we estimate
318: $E_{max} \sim 10^5$ GeV. Since this value exceeds the HESS date range for
319: HESS J1804-216, we ignore any high-energy truncation in the proton distribution
320: when calculating the pion-decay spectrum.
321:
322: In order to calculate the pp-initiated cascade leading to the gamma ray emissivity,
323: one must know the current relativistic proton distribution in the interaction region.
324: Once these particles leave the shock acceleration region, they diffuse into the
325: dense ($n\sim 500$ cm$^{-3}$) molecular environment where they scatter with the
326: low-energy ambient medium, though some eventually leave the system without scattering.
327: Given the $\sim 40$ mbarn pp scattering cross section, we estimate the cooling time
328: scale for these cosmic rays to be $\sim 70,000$ years, much longer than the age
329: of the remnant. (In reality, the pp scattering cross section increases slowly with
330: energy, such that this time scale shrinks to $\approx 15,000$ years for cosmic
331: rays with energy $\approx 1.7$ EeV. At this energy, however, the protons escape
332: directly from the source.) As such, the injected proton distribution evolves
333: primarily under the influence of energy-dependent diffusion, according to the
334: simplified equation
335: \begin{equation}
336: {dN(E)\over dt}=Q(E)-{N(E)\over\tau(E)} \;,
337: \end{equation}
338: where $Q(E)$ is the injection rate and
339: \begin{equation}
340: \tau\equiv {R^2\over D(E)}
341: \end{equation}
342: is the diffusion time scale, in terms of the system size $R$ and the diffusion
343: coefficient $D(E)$. To simplify the procedure, yet retain the essential physics,
344: we solve this equation in two limits: (i) for energies such that $\tau(E)>>T_{SNR}$,
345: we put $N(E)=Q(E)\,T_{SNR}$; (ii) for energies such that $\tau(E)<<T_{SNR}$,
346: we take $N(E)=Q(E)\,\tau(E)$. The resulting power laws are connected smoothly at the
347: energy $E_{roll}$ where $T_{SNR}=\tau(E)$.
348:
349: Unfortunately, the process of diffusion (here characterized by the diffusion
350: coefficient $D[E]$) is not well understood, and several viable prescriptions
351: exist. To quantify the dependence of our results on the range of possibilities,
352: we will use the formulation
353: \begin{equation}
354: D(E)\equiv \left({\lambda_{max}\,c\over 3\mu}\right)\left({E\over ZeB\,
355: \lambda_{max}}\right)^{2-\beta}\;,
356: \end{equation}
357: where $\lambda_{max}$ is the scale size of the largest magnetic fluctuations in the
358: system (often similar to the scale size of the dynamical disturbance), $\mu$ ($\sim 1$)
359: represents the magnitude of the magnetic fluctuations relative to the underlying global
360: field, and $\beta$ is the index characterizing the fluctuation spectrum for the various
361: prescriptions of turbulence, i.e., $\beta=1$ for Bohm diffusion, $3/2$ for Kraichnan
362: diffusion, and $5/3$ in the case of Kolmogorov diffusion. In what follows, we shall
363: examine the impact of all three forms of diffusion, and ascertain whether any may be
364: ruled out for this system given the currently available data for HESS J1804-216.
365:
366: The actual value of $E_{roll}$ is rather sensitive to the choice of $R$ and
367: $\lambda_{max}$ (and to a lesser extent, $B$). For reasonable choices ($R\sim
368: 1-10$ pc and $\lambda_{max}\sim 1-10 R$) of these variables, $E_{roll}$ ranges
369: over several decades and spans the whole HESS energy range. As such,
370: $E_{roll}$ effectively functions as a free parameter, since neither $R$ nor
371: $\lambda_{max}$ are known precisely.
372:
373: We show in Figures 2, 3, and 4 the gamma ray spectra of particle distributions
374: injected into the medium by shocks in SNR G8.7-0.1, and subsequently modified in energy
375: by diffusion. For all three cases of diffusion, the particle distribution index above
376: $E_{roll}$ is $-2.7$, but in order to attain this value with the different $\beta$'s,
377: the assumed proton injection power-law must be modified accordingly. A quick inspection
378: of these three figures indicates that the principal impact of this feature is to alter
379: the spectrum below $E_{roll}$, where the particles have not yet had time to diffuse
380: through the medium. As such, the photon spectrum in this region is shaped by $Q(E)$.
381: Thus, the spectrum below $E_{roll}$ is hardest for Bohm diffusion (Figure~2), and it
382: softens progressively from Bohm to Kraichnan (Figure~3), and then to Kolmogorov (Figure~4).
383:
384: Under the assumption of a uniform and steady injection rate $Q(E)$, as we have adopted
385: here, only the Bohm diffusion scenario results in a photon spectrum that does
386: not violate the EGRET upper limit. Because the ensuing emissivity scales as the product
387: $N\cdot n_p$, the energy content of the relativistic particles is given by $U_E
388: \approx 10^{49}\,(n_p/500\,{\hbox{cm}}^{-3})^{-1}$ ergs, which represents approximately 1\%
389: of the supernova energy for the assumed density. In this case (of Bohm diffusion),
390: the integrated flux between 100 MeV and 30 GeV is found to be $6\times 10^{-8}$ cm$^{-2}$
391: s$^{-1}$, and thus falls right at the EGRET threshold. In the near future, GLAST
392: will greatly improve the sensitivity of measurements in this energy range, and will
393: therefore provide an important probe into the possible association of HESS J1804-216 with
394: SNR G8.7-0.1. In fact, the $\sim 1$ GeV emission should be detectable with GLAST after a
395: one-year all-sky survey (indicated by the thick solid lines in Figures~2, 3, and 4)
396: regardless of which prescription of diffusion is active.
397:
398: For completeness we calculate the broadband emissivity for the case of Bohm diffusion
399: by including the flux contributions from secondary leptons
400: (as outlined in detail in Fatuzzo \& Melia 2003). The decay of charged pions
401: leads to the production of electrons and positrons in the scattering environment that
402: in turn radiate via bremsstrahlung, synchrotron and inverse Compton scattering with the
403: 5.7 eV cm$^{-3}$ stellar photon field that permeates this region (Helfand et al. 2005;
404: strictly speaking, this is the photon density at the galactic center, but given that we
405: need it primarily to provide an upper limit, it will suffice as an estimate of the photon
406: field at the location of G8.7-0.1 as well). The resulting population of these secondary
407: leptons depends on the injection rate (which is tied directly to the rate of pion decay,
408: and subsequently, to the $\pi^0$ decay emissivity fixed by the HESS data) and the energy
409: loss rates from the aforementioned emission processes and from Coulomb processes. The
410: cooling time $E/\dot E$ for each of these four processes is shown in Figure~5 for an
411: assumed ambient density $n_p = 500$ cm$^{-3}$ and magnetic field strength $B=10^{-5}$ G.
412: It is clear that leptons with energy between $\sim 10^2$ and $10^7$ MeV will not have
413: time to cool during the remnant's estimated lifetime. The lepton distribution is
414: therefore approximated by using the lower part of the steady-state lepton distribution
415: curve and the secondary lepton injection rate curve multiplied by the age of the remnant.
416:
417: The broadband spectrum (from radio to TeV energies) for the case of Bohm diffusion is
418: shown in Figure~6. This now includes the various contributions from the secondary leptons
419: as well as photons emitted during pion decays. Clearly, the radiative flux produced by
420: secondary particles falls well below all broadband limits. It is worth noting, however,
421: that a secondary lepton population created by the proton distribution with spectral index
422: of $\sim -2.3$ (also mirrored by the injected secondaries) can reproduce the observed
423: radio emission if the magnetic field strength is $\sim 0.3$ mG. Such field strengths
424: have been associated with SNR - molecular cloud interactions (Claussen et al. 1997;
425: Koralesky et al. 1998; Brogan et al. 2000). However, this choice of spectral index
426: for the protons is clearly at odds with the EGRET observations.
427:
428: \section{Conclusion}
429: We are left with a rather intriguing situation. On the one hand, we have shown that the TeV
430: emission associated with the new source HESS J1804-216 may be a signature of particle
431: acceleration and injection into the ambient medium by the shell of SNR G8.7-0.1, but
432: not by the pulsar PSR J1803-2137 embedded within it. Reasonable values of the physical
433: variables are sufficient to account for its radiative characteristics, and associated
434: flux limits observed from this region at other wavelengths.
435:
436: On the other hand, the maximum energy of cosmic rays energized by the presumed shock
437: in G8.7-0.1 is severely limited by the acceleration efficiency and time, and would
438: seem to be restricted to values below about $10^5$ GeV. This scenario would not support
439: the viability of a CR anisotropy from this region, and would be consistent with the
440: findings of the Auger Observatory, which does not confirm earlier claims based on
441: the SUGAR and AGASA data.
442:
443: The pulsar model for HESS J1804-216 does not work at TeV energies, but unlike the
444: SNR source, the energy of particles accelerated near the pulsar's light cylinder
445: reach values as high as $\sim 24$ EeV. The capability of PSR J1803-2137 to produce
446: such energetic cosmic rays means that a CR anisotropy in the direction of HESS J1804-216
447: could be accounted for with this model. Such a scenario would require that EeV protons
448: produced nearly 15,000 years ago have not all diffused out of the surrounding,
449: lower density ($< 10$ cm$^{-3}$) region. In contrast, particles associated with
450: the HESS source would result from the SNR-cloud interaction, and populate a
451: denser (and smaller) shell. Such an anisotropy, though, would require the
452: unlikely circumstance of Auger's findings being incorrect.
453:
454: To see how these considerations impact the broader context of high-energy
455: activity at the galactic center, let us firstly take it that, as the AUGER data
456: suggest, there really is (currently) no EeV CR anisotropy in this direction.
457: Then we are led to one of the two following conclusions:
458: \begin{enumerate}
459: \item if the SUGAR and AGASA data and the analyses thereof (suggesting the existence
460: of galactic center CR anisotropies) are broadly correct, then a source both variable
461: on decadal timescales and capable of accelerating particles to beyond $10^{18}$ eV
462: exists. Only a source associated with a compact object, most likely a pulsar, could
463: satisfy these requirements but, as may be seen above, it is then difficult to conceive
464: of a source that might produce both the required variability and a power-law spectrum
465: of accelerated particles; it would seem, then, that the HESS J1804-216 source and the
466: SUGAR anisotropy cannot be directly attributed to the same object. Likewise, the
467: high-energy galactic center source scenarios described in Crocker et al. (2005a) could
468: not account for the required variability plus power-law spectrum of accelerated particles.
469: \item if, on the other hand, the SUGAR and AGASA data and/or the analyses thereof are in error,
470: then the scenarios outlined in Crocker et al. (2005a)---which are predicated on the
471: SUGAR and AGASA anisotropies being real---cannot hold. There also necessarily exists
472: an implicit upper limit on the strength of any $10^{18}$ eV neutron source located
473: within the field of view of AUGER and, in particular, towards the galactic center.
474: Furthermore, the in-situ high-energy hadronic population inferred on the basis of the
475: TeV radiation detected by the HESS instrument from the galactic center (and also from
476: the J1804-216 source) must cut off below $\sim$ EeV. Alternatively, it may be that,
477: even if this population does continue up to the EeV energy scale as an undistorted
478: power law, the simple scaling behavior for the $pp \to n X$ type interaction assumed by
479: Crocker et al. (2005a) and also implicitly employed above (given the dearth of direct
480: experimental data on neutron production at these very high center-of-mass energies)
481: fails and, in fact, significantly over-estimates neutron production (Grasso and Maccione
482: 2005).
483:
484: \end{enumerate}
485:
486: Should the analysis of the currently available AUGER data be incomplete, so that, in
487: particular, the galactic center (angular) region be CR over-abundant at $\sim$ EeV
488: energies with an amplitude suggested by the AGASA and SUGAR data and consequent
489: analyses, a proton acceleration model with PSR J1803-2137 as the source is a viable
490: explanation for the SUGAR CR anisotropy. We note in this regard that the minimum energy
491: associated with this pulsar's particle injection is well beyond the HESS data range, so
492: this eventuality would not be impacted by the low energy data. Alternatively,
493: it may be that, as discussed by
494: Crocker et al. (2005a), the SUGAR point source be real but actually located at the
495: galactic center and, therefore, associated not with HESS J1804-216, but rather the
496: galactic center source detected by HESS (this would, of course, require that SUGAR's
497: directional determination be in error by $\sim 8^\circ$, but then the problem with
498: AGASA's non-observation of the SUGAR point source would be resolved).
499:
500: Pulsars such as PSR J1803-2137, embedded within the environment surrounding the
501: SNR G8.7-0.1, can be viable sources of EeV cosmic rays within the Galaxy. This
502: warrants further analysis. In future work, we will examine whether the known
503: population of sources such as this, distributed throughout the Milky Way, can
504: conceivably account for the isotropic component of the CR spectrum observed at
505: $\sim$ EeV energies.
506:
507: \bigskip
508: \centerline{\bf Acknowledgments}
509:
510: This research was partially supported by NSF grant AST-0402502 (at Arizona).
511: FM is very grateful to the University of Melbourne for its support (through a
512: Miegunyah Fellowship). MF is supported by the Hauck Foundation through Xavier
513: University.
514:
515: \begin{thebibliography}{}
516:
517: \bibitem[{Aharonian} et~al. 2006]{Aha06}
518: Aharonian, F., et~al. 2006, ApJ, 636, 777
519: \bibitem[{Aharonian} et~al. 2005]{AharonianHESS2005}
520: {Aharonian}, F. et~al. 2005, Science, 307, 1938
521:
522: \bibitem[{Aharonian} \& {Neronov} 2005]{Aharonian2005}
523: {Aharonian}, F., \& {Neronov}, A. 2005, ApJ, 619, 306
524:
525: \bibitem[{Arons} 2003]{Arons2003}
526: {Arons}, J. 2003, ApJ, 589, 871
527:
528: \bibitem[]{}
529: Baring, M. G., et al. 1999, ApJ, 513, 311
530:
531: \bibitem[{Bellido} et~al. 2001]{Bellido2001}
532: {Bellido}, J.~A., {Clay}, R.~W., {Dawson}, B.~R., \& {Johnston-Hollitt}, M.
533: 2001, Astroparticle Physics, 15, 167
534:
535: \bibitem[{Benbow} et al. 2005]{Benbow2005}
536: {Benbow}, W., \& Hess Collaboration 2005, AIP Conf. Proc. 745: High
537: Energy Gamma-Ray Astronomy, 745, 611
538:
539: \bibitem[Berezhko \& V\"olk 2000]{Ber2000}
540: Berezhko, E. G., \& V\"olk, H. J. 2000, ApJ, 540, 923
541:
542: \bibitem[{Blasi} et al. 2000]{Blasi2003}
543: {Blasi}, P., {Epstein}, R. I., \& {Olinto}, A. V. 2000, ApJ Letters, 533, L123
544:
545: \bibitem[{Blitz} et al. 1982]{Blitz82}
546: {Blitz}, L., Fich, M., \& Stark, A. A. 1982, ApJ Supplements, 49, 183
547:
548: \bibitem[{Bossa} et~al. 2003]{Bossa2003}
549: {Bossa}, M., Mollerach, S., \& Roulet, E. 2003, J.Phys.G., 29, 1409
550:
551: \bibitem[{Brogan} et al 2000]{Brogan2000}
552: Brogan, C. L., Frail, D. A., Goss, W. M., \& Troland, T. H. 2000, ApJ, 537, 875
553:
554: \bibitem[{Chevalier} 1999]{Chevalier1999}
555: {Chevalier}, R. A. 1999, ApJ, 511, 798
556:
557: \bibitem[{Claussen} et al. 1997]{Claussen1997}
558: Claussen, M. J., Frail, D. A., Goss, W. M., \& Gaume, R. A.
559: 1997, ApJ, 489, 143
560:
561: \bibitem[{Clifton} et al. 1986]{Clifton86}
562: {Clifton}, T. R. \& Lyne, A. G. 1986, MNRAS, 174, 267
563:
564: \bibitem[{Crocker} et~al. 2005a]{Crocker2005a}
565: {Crocker}, R., {Fatuzzo}, M., {Jokipii}, R., {Melia}, F., \& {Volkas}, R.
566: 2005, ApJ, 622, 892
567:
568: \bibitem[{Crocker} et~al. 2005b]{Crocker2005b}
569: {Crocker}, R., {Melia}, F., \& {Volkas}, R.
570: 2005, ApJ Letters, 622, L37
571:
572: \bibitem[{Drury} et al. 1994]{Drury1994}
573: Drury, L. O'C., Aharonian, F. A., \& V\"olk, H. J. 1994, A\&A, 287, 959
574:
575:
576: \bibitem[{Fatuzzo} \& {Melia} 2003]{FM03}
577: {Fatuzzo}, M., \& Melia, F. 2003, ApJ, 596, 1035
578:
579: \bibitem[Fatuzzo \& Melia 2005]{Fat2005}
580: Fatuzzo, M., \& Melia, F. 2005, ApJ, 630, 321
581:
582: \bibitem[{Finley} \& {\"Ogelman} 1994]{Finley94}
583: {Finley}, J. P. \& {\"Ogelman}, H. 1994, ApJ Letters, 434, L25
584:
585: \bibitem[Gaisser et al. 1998]{Gaisser98}
586: Gaisser, T. K., Protheroe, R. J., \& Stanev, T. 1998, ApJ,
587: 492, 227
588:
589: \bibitem[Grasso \& Maccione(2005)]{2005APh....24..273G} Grasso, D., \&
590: Maccione, L.\ 2005, Astroparticle Physics, 24, 273
591:
592: \bibitem[{Green} 2004]{Green04}
593: {Green}, D. A. 2004, Bulletin of the Astronomical Society of India, 32, 335
594:
595: \bibitem[{Handa} et al. 1988]{Handa88}
596: {Handa}, T. et al. 1988, PASJ, 39, 709
597:
598: \bibitem[{Hartman} et al. 1999]{Hartman99}
599: {Hartman}, R. C. et al. 1999, ApJ Supplements, 123, 79
600:
601: \bibitem[{Hayashida} et~al. 1999]{Hayashida1999}
602: {Hayashida}, N. et~al. 1999, Astroparticle Physics, 10, 303
603:
604: \bibitem[{Helfand} et al. 2005]{Helfand05}
605: {Helfand} D. J., {Becker}, R. H., \& {White}, R. L. 2005,
606: ApJ Letters, submitted (astro-ph/0505392)
607:
608: \bibitem[{Jones} 1990]{Jones1990}
609: {Jones}, L. 1990, in Proceedings of the 21st ICRC (Adelaide), Ed. R. Protheroe,
610: 2, 75
611:
612: \bibitem[{Kassim} \& {Weiler} 1990]{Kassim90}
613: {Kassim}, N. E. \& {Weiler}, K. W. 1990, Nature, 343, 146
614:
615: \bibitem[{Koralesky} et al 1998]{Koralesky98}
616: Koralesky, B., Frail, D. A., Goss, W. M., Claussen, M. J., \&
617: Green, A. J. 1998, AJ, 116, 1323
618:
619: \bibitem[{Letessier-Selvon} et~al. 2005]{Letessier2005}
620: {Letessier-Selvon}, A. \& Auger Collaboration 2005, 29th International
621: Cosmic Ray Conference, Pune, 101
622:
623: \bibitem[{Markoff} et~al. 1997]{Markoff1997}
624: {Markoff}, S., {Melia}, F., \& {Sarcevic}, I.
625: 1997, ApJ Letters, 489, L47
626:
627: \bibitem[{Markoff} et~al. 1999]{Markoff1999}
628: {Markoff}, S., {Melia}, F., \& {Sarcevic}, I.
629: 1999, ApJ, 522, 870
630:
631: \bibitem[{Melia} 2006]{Melia06}
632: Melia, F. 2006, The Galactic Supermassive Black Hole (Princeton University Press:
633: New York), in press
634:
635: \bibitem[{Melia} \& {Falcke} 2001]{Melia01}
636: {Melia}, F. \& Falcke, H. 2001, ARAA, 39, 309
637:
638: \bibitem[{Sturner} et al 1997]{Sturner97}
639: Sturner, S. J., Skibo, J. G., Dermer, C. D., \& Mattox, J. R.
640: 1997, ApJ, 490, 619
641:
642: \end{thebibliography}
643:
644: \newpage
645: \begin{figure}
646: \figurenum{1}
647: {\epsscale{0.9} \plotone{fig1.ps} }
648: \figcaption{Illustrative $\gamma$-ray emissivity from the pulsar-powered pion-decay model
649: with $E_{max}\sim 24$ EeV and $E_{min}\sim 6\times 10^5$ GeV (both calculated from
650: its inferred spin-down age) and an assumed ambient density of $10^3$ cm$^{-3}$.
651: The dotted curve shows the resulting emissivity if the injected particles do not
652: diffuse out of the region. The solid curve shows the resulting emissivity
653: for an idealized distribution that, as a result of diffusion, has a spectral
654: index of -2.7. In both cases, the $\pi^0$ cascade induced by protons with energy $E$
655: produces a photon spectrum with index $\approx -1$ below $E_{min}$. In the case where
656: diffusion acts to remove high-energy particles from the emission region,
657: the spectrum below $E_{min}$ is almost entirely due to protons with
658: energy $\sim E_{min}$ (whose contributions are shown by the short-dashed curve).
659: Similarly, the long-dashed curve shows the photon spectrum
660: resulting from the $\pi^0$ cascade initiated by protons at $E\approx 6\times 10^{8}$ GeV.
661: Regardless of how much the energy-dependent diffusion modifies the injected cosmic ray
662: spectrum above $E_{min}$, the pulsar model therefore cannot account for the steep
663: spectrum (index $-2.7$) measured by HESS.}
664: \end{figure}
665:
666: \newpage
667: \begin{figure}
668: \figurenum{2}
669: {\epsscale{0.9} \plotone{fig2.ps} }
670: \figcaption{The $\gamma$-ray emissivity (solid line) for a particle distribution
671: injected with index $-2.0$ and modified by Bohm diffusion (and $E_{roll}=7\times 10^3$
672: GeV), and the corresponding spectrum (dashed curve) produced without diffusion.
673: The EGRET bar is an upper limit, and the GLAST curve is the simulated one-year
674: all sky survey limit.}
675: \end{figure}
676:
677: \newpage
678: \begin{figure}
679: \figurenum{3}
680: {\epsscale{0.9} \plotone{fig3.ps} }
681: \figcaption{Same as Figure~2, except now for Kraichnan diffusion with $E_{roll}=
682: 10^3$ GeV, and an injected particle distribution index $-2.2$.}
683: \end{figure}
684:
685: \newpage
686: \begin{figure}
687: \figurenum{4}
688: {\epsscale{0.9} \plotone{fig4.ps} }
689: \figcaption{Same as Figures~2 and 3, except now for Kolmogorov diffusion
690: with $E_{roll}=100$ GeV, and an injected particle distribution index $-2.4$.}
691: \end{figure}
692:
693: \newpage
694: \begin{figure}
695: \figurenum{5}
696: {\epsscale{0.9} \plotone{fig5.ps} }
697: \figcaption{The cooling time $\tau = E / \dot E$ as a function of energy
698: for leptons interacting with a medium of density $n_H = 500$ cm$^{-3}$
699: and magnetic field strength $B = 10^{-5}$ G. Short dashed curve: bremsstrahlung
700: cooling; solid curve: synchrotron cooling; long dashed curve: Coulomb losses;
701: dotted curve: inverse Compton scattering with the ambient stellar photon field.}
702: The dot-dashed curve represents the value of the SNR age.
703: \end{figure}
704:
705: \newpage
706: \begin{figure}
707: \figurenum{6}
708: {\epsscale{0.9} \plotone{fig6.ps} }
709: \figcaption{The $\gamma$-ray emissivity from the SNR-powered pion-decay
710: model with $n_H = 500$ cm$^{-3}$, and $B = 10^{-5}$ G.
711: The remnant's age is assumed to be $15,000$ years in order to determine the secondary
712: leptons' (non steady-state) distribution. Solid line: photons produced via the decay
713: of neutral pions; long dashed curve: bremsstrahlung emission from the secondary
714: leptons; dot-dashed line: inverse Compton scattering emission from the secondary
715: leptons interacting with the background stellar field; dotted line: synchrotron
716: emission from the secondary leptons. The HESS data are represented as dark squares.
717: Also shown are the EGRET upper limit, and the measurements made with ROSAT and the
718: VLA, both of which must be considered as upper limits as well, given the differences
719: in field-of-view.}
720: \end{figure}
721:
722: \end{document}
723: