1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2:
3: %% Trim Size: 9.75in x 6.5in
4: %% Text Area: 8in (include Runningheads) x 5in
5: %% ws-ijmpd.tex : 25-10-04
6: %% Tex file to use with ws-ijmpd.cls written in Latex2E.
7: %% The content, structure, format and layout of this style file is the
8: %% property of World Scientific Publishing Co. Pte. Ltd.
9: %% Copyright 1995, 2002 by World Scientific Publishing Co.
10: %% All rights are reserved.
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12: %
13:
14: %\documentclass[draft]{ws-ijmpd}
15: \documentclass{ws-ijmpd}
16:
17: \begin{document}
18:
19: \markboth{V. Bosch-Ramon \& D. Khangulyan}
20: {Understanding the Very High-Energy Emission from Microquasars}
21:
22: %%%%%%%%%%%%%%%%%%%%% Publisher's Area please ignore %%%%%%%%%%%%%%%
23: %
24: \catchline{}{}{}{}{}
25: %
26: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
27:
28: \title{Understanding the Very High-Energy Emission from Microquasars}
29:
30: \author{Valent\'i Bosch-Ramon}
31:
32: %\address{\footnote{State completely without abbreviations, the
33: %affiliation and mailing address, including country. Typeset in
34: %8~pt Times italic.}\\
35: %first\_author@university.edu}
36:
37: \author{Dmitry Khangulyan}
38:
39: \address{Max Planck Institut f\"ur Kernphysik, postfach 10 39 80
40: 69029 Heidelberg, Germany\\
41: Valenti.Bosch-Ramon@mpi-hd.mpg.de/Dmitry.Khangulyan@mpi-hd.mpg.de}
42:
43: \maketitle
44:
45: \begin{history}
46: \received{Day Month Year}
47: \revised{Day Month Year}
48: \comby{Managing Editor}
49: \end{history}
50:
51: \begin{abstract}
52: Microquasars are X-ray binaries with relativistic jets. These jets are powerful energy carriers, thought
53: to be fed by accretion, which produce non-thermal emission at different energy bands. The processes behind the bulk of the
54: non-thermal emission in microquasars may be of leptonic (synchrotron and inverse Compton) and hadronic (proton-proton
55: interactions, photo-meson production, and photo-disintegration) nature. When leptonic, the fast particle cooling would allow
56: the obtention of relevant information about the properties close to the accelerator, like the radiation and the magnetic
57: field energy densities, and the acceleration efficiency. When hadronic, the extreme conditions required in the emitter would
58: have strong implications on the physics of jets and their surroundings. The very high-energy part of the spectrum, i.e.
59: $>100$~GeV, is a good energy range to explore the physics behind the non-thermal radiation in these compact variable sources.
60: In addition, this energy range, when taken altogether with lower energy bands, is a key piece to construct a comprehensive
61: picture of the processes occurring in the emitter. Until recently, the very high-energy range was hard to probe due to the
62: lack of sensitivity and spatial and spectral resolution of previous instrumentation. Nowadays, however, powerful gamma-ray
63: instruments are operating and the quality of their observations is allowing, for the first time, to start to understand the
64: production of high-energy emission in microquasars.
65:
66: To date, several Galactic sources showing extended radio emission, among them at least one confirmed microquasar, Cygnus~X-1,
67: have shown a TeV signal. All of them show complex patterns of spectral and temporal behavior. In this work, we
68: discuss the physics behind the very high-energy emission in the microquasar Cygnus~X-1, and also in the other two TeV
69: binaries with detected extended outflows, LS~5039 and LS~I~+61~303, pointing out relevant aspects of the complex phenomena
70: occurring in them. We conclude that the TeV emission is likely of leptonic origin, although hadrons cannot be discarded. In
71: addition, efficient electromagnetic cascades can hardly develop since even relatively low magnetic fields suppress them.
72: Also, the modeling of the radiation from some of the detected sources points to them as either extremely efficient
73: accelerators, and/or having the TeV emitter at a distance from the compact object of about $\sim 10^{12}$~cm. Finally, we
74: point out that the role of a massive and hot stellar companion, due to its strong photon field and wind, cannot be neglected
75: when trying to understand the behavior of microquasars at high and very high energies. The complexity of microquasars
76: precludes straightforward generalizations to a whole population, and are better studied presently in a source by source base.
77: The new and future gamma-ray instrumentation
78: will imply a big step further in our understanding of the processes in microquasars and gamma-ray emitting binaries.
79: \end{abstract}
80: \keywords{Microquasars; radiative processes; outflows; gamma-ray emission}
81:
82: \section{Introduction}
83:
84: Microquasars are an X-ray binary (XRB) subclass formed by those sources that present extended radio jets (e.g. Mirabel \&
85: Rodr\'iguez \cite{mirabel99}). These systems are formed by a non-degenerated star, which can be in different stages of its
86: evolution, and a compact object, which can be a black-hole or a neutron star. Depending on the mass of the non-degenerated
87: stellar companion, the system is considered a low or a high-mass microquasar. Typically, systems harboring an OB star are
88: considered high-mass XRBs (HMXB), and XRBs with later type stellar companions are classified as intermediate or low-mass
89: XRBs. It is thought that the compact object powers the relativistic jets via accretion of matter expelled from the companion.
90: This material, when reaching the surroundings of the compact object, forms an accretion disk that is usually detected in the
91: X-rays. Simplifying very much the case, depending on the accretion state, the X-ray spectrum varies strongly, from a
92: multi-color black body peaking around 1~keV plus a minor soft power-law spectrum at higher energies, for high accretion rates
93: (high-soft state), to one dominated by a hard power-law spectrum plus an exponential cutoff around 100~keV, for low accretion rates (low-hard
94: state; for an extensive
95: description of the X-ray phenomenology, see McClintock \& Remillard \cite{mcclintock06}). It is expected that a persistent jet will be present during the low-hard state, and a transient ejection will form
96: when switching from the low-hard to the high-soft state (Fender et al. \cite{fender04a}). Correlations between the radio and
97: the X-ray luminosity, and the accretion/ejection activity, have been proposed (e.g. Gallo et al. \cite{gallo03}, Corbel et
98: al. \cite{corbel03}, Fender et al. \cite{fender03}).
99:
100: The jet formation and the production of non-thermal radiation in the jet are major ingredients that distinguish a microquasar
101: among other types of XRBs. The non-thermal radiation produced in microquasar jets has been resolved in radio at very
102: different spatial scales (e.g. Rib\'o \cite{ribo05}) and also in X-rays at large scales (e.g. Corbel et al. \cite{corbel02}).
103: This emission is a clear evidence that particle acceleration takes place in different locations of microquasar jets under
104: very different conditions (for a discussion on this, see, e.g., Bosch-Ramon \cite{bosch07}). There are also radio emitting
105: XRBs in which extended emission has not been detected. It has been proposed that these sources could be microquasars as well
106: (e.g. Fender \cite{fender04}), like the low-mass system XTE~J1118$+$480 (e.g., Chaty et al. \cite{chaty03}), the jet of which
107: has not been resolved yet.
108:
109: Microquasars show up themselves as compact and rapidly variable sources from radio to very high energies. In such a type of
110: emitters, when radiating particles are leptons, the highest energy part of the spectrum is a good range to explore
111: non-thermal processes. It is due to the short timescales associated to the particles that produce the emission, which implies
112: that the accelerator and the emitter are likely the same or similar regions. In case of hadrons, it is possible to derive
113: important information concerning the jet hadronic content, at least of its relativistic part, whereas at the same time it is
114: giving information of the conditions of the emitter, like very dense matter and/or target photon fields. Moreover, photons
115: generated by very high-energy (VHE) electrons and/or protons give a better insight on the mechanism of acceleration and the
116: conditions under which it takes place, helping to understand better the processes that accelerate particles up to such high
117: energies. Finally, the presence of a hot and massive star renders a scenario in which photon-photon absorption, and the
118: occurrence of electromagnetic cascades, can be studied. This can give important information on the conditions of the massive
119: star surroundings.
120:
121: Historically, the poor spatial resolution and sensitivity of the available instruments working at gamma-ray energies were not
122: enough for accurate theoretical modeling, although these sources were proposed to be gamma-ray emitters more than a
123: decade ago (e.g. Levinson \& Blandford \cite{levinson96}; Levinson \& Mattox \cite{levinson96b}; Paredes et al.
124: \cite{paredes00}). Nowadays, however, powerful gamma-ray instruments are operating or will start soon, and the quality of
125: their observations is allowing us, for the first time, to probe the physical processes that take place in microquasars and
126: their jets. These observations at very high energies give the necessary input to constrain the theoretical models that lacked
127: in the past.
128:
129: The recent evidence of detection of a transient event, at the (post-trial) significance level of 4.1~$\sigma$, from Cygnus
130: X-1 (Albert et al. \cite{albert07}) have shown that microquasars can indeed produce VHE emission, and VHE observations can
131: bring to us important information on these sources. Another two interesting cases are LS~5039 and LS~I~+61~303. The former
132: has been detected in the TeV range by the Cherenkov telescope HESS (Aharonian et al. \cite{aharonian05}), showing a periodic
133: behavior in the VHE radiation (Aharonian et al. \cite{aharonian06}) with the same period as the orbit (Casares et al.
134: \cite{casares05}). LS~I~+61~303 has been also detected at TeV energies by the Cherenkov telescope MAGIC, being its emission
135: variable (Albert et al. \cite{albert06}). Unlike Cygnus~X-1, which is a firmly established microquasar, LS~5039 and
136: LS~I~+61~303 are no considered at present microquasars due to some peculiarities in their X-ray and radio characteristics.
137: Nevertheless, both sources present, like Cygnus~X-1, extended radio emission, and harbor OB stars.
138:
139: In this review, we want to take profit of the new microquasar phenomenology at very high energies, to draw a theoretical
140: picture of the non-thermal processes that could take place in these sources. This is to be put in the context of the
141: historical evolution of the field, which will be summarized. To explore which are the relevant processes in the microquasar
142: scenario, we will carry out a detailed review of different mechanisms: particle transport, radiation and photon-photon
143: absorption, in the context of microquasars. Using this sound and basic theoretical background, plus the observational
144: knowledge at very high energies from the sources presented above, the microquasar Cygnus~X-1, and also the TeV binaries
145: LS~5039 and LS~I~+61~303, constraints on their physical conditions will be inferred. We note that the type of approach
146: applied to these sources is applicable to large extent to any close binary system emitting in the TeV regime. It is
147: worth mentioning here a recent review of a broader topic by Levinson (\cite{levinson06}) on VHE radiation from jets including
148: active galactic nuclei, gamma-ray bursts, and microquasars. It is also interesting to note that
149: massive young stellar objects, also presenting jets,
150: have been proposed to be gamma-ray emitters by Araudo et al. (\cite{araudo07}).
151:
152: This work is organized in the following manner. In Sections~\ref{genst} and \ref{difsc}, we try to look comprehensively at
153: the microquasar phenomenological picture, setting up a general scenario for these sources. In addition, there we summarize in
154: a non-exhaustive way previous studies of different topics related to microquasars at high energies. In Sect.~\ref{theor}, a
155: short review on different particle transport, acceleration and radiation mechanisms is done, treating also briefly the issue
156: of gamma-ray absorption and electromagnetic cascading in microquasars. All this can help to understand how the new VHE data
157: fits in previous ideas and frameworks. Also, we put forward a plain but physically sound leptonic model to explore the
158: processes that are involved in the generation and absorption of VHE emission in jet galactic sources, trying to understand
159: which kind of physics is relevant there. For this, the recent observational findings at very high energies concerning
160: microquasars are used. In Sect.~\ref{disc}, some hot topics of the field are discussed with some extension. Finally, in
161: Sect.~\ref{summ}, the main ideas of this work are summarized.
162:
163: \section{The study of microquasars}\label{genst}
164:
165: At the discovery of microquasars, their physics was object of speculation due to the morphological similarities of these
166: sources with their larger scale analogs, the quasars (Mirabel et al. \cite{mirabel92}). The question was to find out to which
167: extent the processes taking place in extragalactic jet sources could be extended to galactic ones. In addition to morphology,
168: kinematical similarities became apparent when superluminal moving ejecta were found at radio wavelengths in GRS~1915$+$150
169: (Mirabel et al. \cite{mirabel94}). Nowadays it is widely spread the opinion that galactic and extragalactic relativistic jet
170: sources share much more than just morphological and kinematical resemblance. In fact, since both types of source harbor a
171: compact object surrounded by an accretion disk and a relativistic outflow that is an efficient non-thermal emitter, the
172: original analogy argument has been extended not only to the mechanisms that are producing such non-thermal radiation but even
173: beyond. The aim would be to embrace also the physical link between the relativistic particle generation, the accretion
174: phenomena, the jet formation and to some extent the interaction between the jet and its environment. A well known example
175: would be the phenomenological scaling laws relating the physics of the radio and the X-ray emission in microquasars with the
176: accretion rate (e.g. Corbel et al. \cite{corbel03}; Gallo et al. \cite{gallo03}), and its more general version that includes
177: the mass of the black-hole, being extended to extragalactic objects (e.g. Merloni et al. \cite{merloni03}; Falcke et al.
178: \cite{falcke04}; K\"ording et al. \cite{koerding06}). As already mentioned, these empirical scaling laws include the
179: accretion rate and the central object mass as the key parameters to account for (at some stage Doppler boosting is to be
180: added), as well as phenomenological links, sometimes accretion and jet formation theory motivated, between the radio, the
181: X-ray, and the jet kinetic luminosities and the accretion rate itself.
182:
183: It is usual and seems natural to associate the properties of the accretion disk, the jet and the compact object mass. The
184: latter will influence strongly the properties of the infalling matter, like the accretion rate, the accreted matter
185: temperature, the magnetic field strength or the accretion disk radiative output. In many XRBs, standard accretion theory can
186: explain the X-ray spectrum quite accurately via thermal emission from the inner parts of the accretion disk plus a
187: corona-like emitter\footnote{A region surrounding the compact object that would be filled with a hot population of electrons
188: and ions.}, although there is an on-going debate concerning the origin of the hard X-rays, whether they come from the jet
189: base, or from a corona-like region close to the compact object, both of similar properties (e.g. Markoff et al.
190: \cite{markoff05}; Maccarone \cite{maccarone05}). In any case, some sources do not fit in any of these schemas, since their
191: X-ray radiation does not seem to come from the regions close to the compact object but further out, like the case of SS~433
192: and, perhaps as well, LS~5039 and LS~I~+61~303. This could be explained by recent theoretical studies that show that the role
193: of the magnetic field can be more crucial than the compact object mass itself and can lead to radiatively inefficient
194: accretion (Bogovalov \& Kelner \cite{bogovalov05}), in which case the jet could dominate in the X-ray band (see also Fender
195: et al. \cite{fender03}).
196:
197: Theoretically, the relationship between the compact object mass and the jet physics (e.g. ejected mass rate, matter content,
198: internal energy, carried magnetic energy, bulk speed) is unclear since the accretion/ejection physics is not well known.
199: Furthermore, the link between the main jet properties and the production of non-thermal particles and their radiation cannot
200: be based on first principles, since it relies on the particular conditions of the jet plasma (sound speed, diffusion
201: coefficients, local magnetic field, degree of turbulence, presence of shocks and their velocity, radiation field, plasma
202: density and temperature, etc.), which are not properly constrained. Actually, the situation is even more complicated due to
203: the influence on the jet processes and the non-thermal radiation itself of external factors like, for instance, an external
204: radiation field, the wind produced by the star, the properties of the interstellar medium (ISM), etc. Therefore, although
205: empirical laws can help to classify objects via significant features of their emission and be powerful heuristic tools, it is
206: sometimes difficult to motivate them from first principles, and fundamental approaches become necessary. Such fundamental
207: approaches have to be applied source by source and grounded on detailed observational data. It requires keeping a
208: comprehensive approach to study the source (i.e. phenomenological studies), but developing more fundamental theoretical
209: models.
210:
211: \section{High energy processes in microquasars: a historical perspective}\label{difsc}
212:
213: In this section, a schematic and descriptive picture for the relevant processes taking place in microquasars, like particle acceleration, and generation and absorption of VHE
214: radiation, is presented together with a historical perspective of the field. This will set up the context within which a more detailed physical treatment will be carried out in
215: Sect.~\ref{theor}.
216:
217: \subsection{A descriptive picture}
218:
219: For illustrative purposes, we show in Figure~\ref{mq} a very schematic picture of a microquasar together with its main
220: elements, and some of the processes taking place in these objects. Some regions of the jet are labeled depending on the main
221: radiative products. This simplified picture would consist on several important elements: the companion star, the stellar
222: radiation field, the stellar wind, the compact object, the accretion disk, and the jet itself. The relativistic particles
223: inside the jet will interact with the stellar radiation field as well as the jet magnetic field producing inverse Compton
224: (IC) and synchrotron radiation, respectively. Under the strong stellar radiation field, the creation of pairs will be
225: unavoidable if VHE gamma-rays are produced. Finally, the stellar wind may have a significant impact on the jet. All this,
226: plus some additional physical processes (not included in Fig.~\ref{mq} for clarity), like hadronic emission, will be
227: discussed in more detail in Sect.~\ref{theor}.
228:
229: \begin{figure}[pb]
230: \centerline{\psfig{file=mq1.eps,width=9cm}}
231: \vspace*{8pt}
232: \caption{A sketch of a microquasar. We explicitly show the star, the stellar wind, the accretion
233: disk, the jet, the jet relativistic electrons, some leptonic radiation processes, and pair creation
234: under the stellar photon field.\label{mq}}
235: \end{figure}
236:
237: \subsection{Jet formation, evolution and termination}
238:
239: We present here a brief description of a plausible picture for the jet development, from the launching to the termination
240: point.
241:
242: Although a complete theory for jet generation is still lacking, many studies on jet powering, acceleration and collimation
243: have been carried out during the last decades (e.g. Blandford \cite{blandford76}; Blandford \& Znajek \cite{blandford77};
244: Blandford \& Payne \cite{blandford82}; Meier \cite{meier96}; Koide et al. \cite{koide02}; Chattopadhyay \& Chakrabarti
245: \cite{chatto02}; Meier \cite{meier03}; Hujeirat \cite{hujeirat04}; Meier \cite{meier05}; Bogovalov \& Kelner
246: \cite{bogovalov05}; De Villiers et al. \cite{deVilliers05}; Ferreira et al. \cite{ferreira06};
247: McKinney \cite{mcKinney06};
248: Hawley \& Krolik \cite{hawley06};
249: Komissarov et al. \cite{komissarov07}, Barkov \& Komissarov \cite{barkov08}). At present, due to the apparent correlation between accretion and jet activity (e.g.
250: Fender et al. \cite{fender04a}), a widely accepted scenario is one in which jets are powered by accretion. The accreted
251: matter starts to move following ordered magnetic field lines that thread the inner regions of the accretion disk. By
252: magneto-centrifugal forces, the plasma is ejected by these magnetic field lines from the accretion disk in the direction
253: perpendicular to it. The differential rotation of the accretion disk would create a spiral-like shape of the magnetic lines.
254: This magnetic field configuration would accelerate and collimate the plasma. Although it is unclear at which scales the jet
255: is already formed, VLBI observations of extragalactic jets show evidences that the collimation region could be located at
256: $\sim 100$--1000~$R_{\rm Sch}$ (e.g. Junor et al. \cite{junor99}; Horiuchi et al. \cite{horiuchi06})\footnote{The jets of the
257: microquasar SS~433 present collimation distances, inferred from X-ray observations, apparently larger, of about $\sim
258: 10^6$~$R_{\rm Sch}$ (Namiki et al. \cite{namiki03}). We note however that the jet of SS~433 is quite different from a typical
259: compact microquasar jet because of its huge kinetic power, messy environment, and thermal emitting nature.}.
260:
261: Once the microquasar jet is formed, it may interact with dense material ejected by the accretion disk or the stellar
262: companion, the latter being almost unavoidable in high-mass systems. Although the role of an accretion disk wind could be to
263: give further collimation and stability to the jet (e.g. Hardee \& Hughes \cite{hardee03}; Tsinganos et al.
264: \cite{tsinganos04}), the role of a strong lateral stellar wind may lead to jet bending and even disruption (Perucho \&
265: Bosch-Ramon \cite{perucho07}). In any case, jet/environment interaction can lead to shock formation and the radiative
266: counterpart may be observable either as a transient phenomena, when the jet (or a discrete ejection) penetrates for the first
267: time the surrounding medium, or as a (quasi) steady one, when the jet formation is continuous at the relevant timescales and
268: recollimation shocks occur (Perucho \& Bosch-Ramon \cite{perucho07}).
269:
270: Despite the environment of microquasars at large spatial scales may be quite different depending on the Galaxy region or the
271: strength of the companion star wind, the jets should stop somewhere, terminating either via disruption, in a similar way to
272: extragalactic Fanaroff-Riley I sources (e.g. Fanaroff \& Riley \cite{faranoff74}; Perucho \& Mart\'i \cite{pm07}), or via
273: strong shocks in the ISM, as seems to be the case for Fanaroff-Riley II sources (e.g Fanaroff \& Riley \cite{faranoff74};
274: Kaiser \& Alexander \cite{kaiser97}; Scheck et al. \cite{sch02}). In either case, the radiative outcomes would be different.
275: A classical example of the interaction between a microquasar jet and its environment at large scales is the case of the
276: W~50-SS~433 system (e.g. Zealey et al. \cite{zealey80}; Vel\'azquez \& Raga \cite{velazquez00}).
277:
278: In the following sections, we list several non-thermal processes, i.e. particle acceleration and radiation, that may occur in
279: the jet giving a (non exhaustive) overview on the literature. For the sake of clarity, we have divided the jet in several
280: regions. The jet base, close to the compact object ($\sim 100-1000$~R$_{\rm Sch}$), a region farther out, at the binary
281: system scales (typically $\sim 10^{11}-10^{13}$~cm), well outside the binary system -jet middle scales- (around
282: $10^{15}-10^{16}$~cm), and the termination regions of the jet, where it ends interacting somehow with the ISM ($\ge
283: 10^{17}$~cm). We show in Fig.~\ref{mqq} a sketch of the different considered regions. Later on, with the most recent
284: observational findings at hand, we will go deeper in the modeling to find out which of the different considered scenarios is
285: the most suitable to explain the data.
286:
287: \begin{figure}[pb]
288: \centerline{\psfig{file=mqq.eps,width=9cm}}
289: \vspace*{8pt}
290: \caption{Sketch of a microquasar with the different relevant scales resumed in this section. \label{mqq}}
291: \end{figure}
292:
293: \subsection{Particle acceleration}
294:
295: Particle acceleration processes in microquasar jets can be of different types. At the base, the jet could be magnetically
296: dominated (e.g., in extragalactic jets: Sikora et al. \cite{sikora05}), and a mechanism of particle acceleration may be
297: magnetic energy dissipation via MHD instabilities in the jet base (e.g. Zenitani et al. \cite{zenitani01}). Also, if jet
298: velocities were high enough in the base, the dense available photon (from an accretion disk/corona) and matter fields could
299: allow the converter mechanism to take place (e.g. Derishev et al. \cite{derishev03}; Stern \& Poutanen \cite{stern06}).
300: Magnetic field reconnection in the surrounding corona could inject a non-thermal population of particles in the jet as well
301: (e.g. Gierlinski \& Done \cite{gierlinski03} and references therein). A magneto-centrifugal mechanism could also operate very
302: close to a rotating black-hole (e.g. Neronov \& Aharonian \cite{neronov07}; Rieger \& Aharonian \cite{rieger07}).
303:
304: At binary system scales, plausible mechanisms to generate relativistic particles in the jet are the different versions of the
305: Fermi process: shock diffusive (Fermi~I); random scattering (Fermi~II); and shear acceleration (e.g. Drury \cite{drury83};
306: Fermi \cite{fermi49}; and Rieger \& Duffy \cite{rieger04}, respectively; see also Rieger et al. \cite{rieger06}). Fermi~I
307: mechanism could take place due to internal shocks in the jet; Fermi~II acceleration could take place if magnetic turbulence
308: is strong enough, with high Alfven velocity; shear layer would be a natural outcome of an expanding jet or different
309: jet/medium velocities. Interactions with the stellar wind may also trigger particle acceleration via, e.g., a recollimation
310: shock formed in the jet that expands against the dense material expelled by the companion star (e.g. Perucho \& Bosch-Ramon
311: \cite{perucho07}). The velocities of the shocks mentioned here could be either mildly or strongly relativistic. In the latter
312: case, the converter mechanism may be effective in very bright star systems.
313:
314: At microquasar jet middle scales, some sort of shock acceleration might still take place. For instance,
315: intermittent ejections
316: at timescales of $\sim$hours--days and different velocities could create shocks at distances of about $\sim 10^{15}-10^{16}$~cm. Also
317: Fermi~II type and shear acceleration appear plausible for a continuous outflow at these scales (something similar could
318: happen in the intra-knot regions of extragalactic jets, see, e.g., Rieger et al. \cite{rieger06}). Regarding the environment
319: dynamical influence on the jet, it is not expected to be significant given the high jet ram pressure compared with the medium one
320: at these scales.
321:
322: At the microquasar jet termination point, as in AGN hot spots and radio lobes (e.g. Kaiser \& Alexander \cite{kaiser97}), the
323: external medium plays an important dynamical role. When the swept ISM inertia starts to affect the jet advance, two shocks may
324: be formed, one moving backwards in the jet, the so-called reverse shock, and another one moving forward, the so-called
325: forward or bow shock. Under these conditions, Fermi~I type acceleration mechanism seems the most reasonable option, although
326: high diffusive and convective rates in the downstream regions of the forward/reverse shocks could prevent efficient
327: acceleration to occur. It might be the case as well that hydrodynamical instabilities distorted the jet and mixed jet matter
328: with the ISM without forming strong shocks (e.g. Heinz \& Sunyaev \cite{heinz02}).
329:
330: \subsection{Radiative processes}
331:
332: In the jet base, depending on the dominant conditions, the relevant leptonic radiative mechanisms could be synchrotron
333: emission (e.g. Markoff et al. \cite{markoff01}), relativistic Bremsstrahlung from electrons interacting with jet ions
334: (Bosch-Ramon et al. \cite{bosch06b}), SSC (e.g. Bosch-Ramon \& Paredes \cite{bosch04}) and IC with corona and/or disk photons
335: (e.g. Romero et al. \cite{romero02}; Georganopoulos et al. \cite{georganopoulos02}). Regarding hadronic processes, there are
336: several radiative mechanisms that could produce gamma-rays, neutrinos and, as a by-product, low energy emission from
337: secondary pairs. Two of these mechanisms are the collisions of relativistic protons with ions (pp) in the jet, and
338: interactions between jet relativistic protons and X-ray photons (photo-meson) from the disk, the corona or the jet itself
339: (e.g. Levinson et al. \cite{levinson01}; Aharonian et al. \cite{aharonian06b}; Romero \& Vila \cite{romero08} -who also
340: account for proton synchrotron emission-). These relativistic proton collisions with ions and photons would produce neutral
341: pions ($\pi^0$) that decay to gamma-rays, and charged pions ($\pi^\pm$) that decay to muons and neutrinos, the former
342: decaying then to electron-positron pairs and neutrinos. Another possible hadronic mechanism is photo-disintegration, which
343: requires the presence of UHE heavy nuclei and a dense field of target photons of large enough energy. This process produces
344: lower mass hadrons and gamma-rays.
345:
346: At binary system scales, possible radiative leptonic processes taking place in microquasars are synchrotron emission (e.g.
347: Yuan et al. \cite{yuan05}; Paredes et al. \cite{paredes06}), relativistic Bremsstrahlung (e.g. Bosch-Ramon et al.
348: \cite{bosch06b}), SSC (e.g. Atoyan \& Aharonian \cite{atoyan99}; Dermer et al. \cite{dermer06}), and external IC (e.g.
349: Paredes et al. \cite{paredes00,paredes02}; Kaufman Bernad\'o et al. \cite{kaufman02}; Georganopoulos et al.
350: \cite{georganopoulos02}; Dermer et al. \cite{dermer06}; Khangulyan et al. \cite{khangulyan07}). At these spatial scales, jet
351: proton collisions with target nuclei of the stellar wind (e.g. Romero et al. \cite{romero03}; Aharonian et al.
352: \cite{aharonian06b}) seem to be the most efficient hadronic process. As noted above, this mechanism would lead as well to
353: neutrino production (e.g. Romero et al. \cite{romero05}; Aharonian et al. \cite{aharonian06b}). Other hadronic processes,
354: which were also discussed in the literature, are photo-meson production (e.g. Aharonian \cite{aharonian06c}) and
355: photo-disintegration (e.g. Bednarek \cite{bednarek05}).
356:
357: The emission at larger scales are commonly characterized by synchrotron radiation. At higher energies, stellar IC scattering
358: is quite inefficient because the large distances to the companion star and the subsequent dilution of the stellar photon
359: field. Nevertheless, for powerful ejections, SSC could still be significant (e.g. Atoyan \& Aharonian \cite{atoyan99}).
360: Regarding the particle energy distribution, its evolution is likely dominated by convective and adiabatic energy losses (van
361: der Laan \cite{vanderlaan66}). At the termination of the jet, in case particle acceleration and confinement were efficient,
362: synchrotron, relativistic Bremsstrahlung and IC radiation could be produced and even detected for galactic sources (e.g.
363: Aharonian \& Atoyan \cite{aharonian98}; Bordas et al. \cite{bordas08}). Hadronic acceleration could take place as well, which
364: could lead to gamma-ray production (e.g. Heinz \& Sunyaev \cite{heinz02}) and secondary leptonic emission (e.g. Bosch-Ramon
365: et al. \cite{bosch05}).
366:
367: Regarding the variability of the emission discussed above, several factors are relevant: injection could change via
368: variations in the accelerator (e.g. injection power, injection spectrum); the target densities could vary via changes in the
369: magnetic, photon and matter fields; geometry changes, due to e.g. orbital motion or jet precession, could affect anisotropic
370: gamma-gamma absorption and IC scattering or the level of radiation Doppler boosting. Thus, the timescales of the variability
371: could be linked to the injection mechanism, the radiative cooling, particle escape, and the orbital motion, depending on
372: which mechanism plays a major role. Depending on the emitter location and size, some kinds of variability will not play a
373: role because they will be smeared out. We come back to this issue for a more detailed discussion in Sect.~\ref{theor}.
374:
375: \subsection{Photon-photon absorption and electromagnetic cascades}
376:
377: Close to the compact object, either in the inner accretion disk, in the corona-like region, or in the jet base, quite extreme
378: conditions should be present. Magnetic fields are likely high, and the same applies to the present photon fields, which may
379: be dominated by thermal radiation from accretion peaking at UV/X-rays. Dense UV/X-ray photon fields imply that $\sim$~GeV
380: gamma-ray absorption close to the compact object will be very high. The strong magnetic field may suppress efficient
381: electromagnetic cascading, although the occurrence of cascades (e.g. Akharonian et al. \cite{akharonian85}) cannot be
382: discarded. At binary system scales, the photon-photon opacities of the stellar photon field for gamma-rays are high in
383: massive systems. Although high-mass stars can have quite large magnetic fields in their surfaces, up to $10^3$~G for a young
384: O star (e.g. Donati et al. \cite{donati02}), electromagnetic cascades may still develop in the system since the magnetic
385: field strength at several stellar radii is not well known. Several authors have studied the photon-photon absorption effects
386: in the VHE lightcurve (e.g. Protheroe \& Stanev \cite{protheroe87}; Moskalenko \& Karakula \cite{moskalenko94}; Bednarek
387: \cite{bednarek97}; Boettcher \& Dermer \cite{boettcher05}; Dubus \cite{dubus06}; Khangulyan et~al. \cite{khangulyan07};
388: Reynoso et~al. \cite{reynoso08}). The effects of absorption on the radiation variability are important not only because the
389: photon column density changes along the orbit, but also due to the angular dependence of the cross section and the low-energy
390: threshold of the pair creation process. The interaction angle between gamma-rays and stellar photons changes with the orbital
391: phase. Several studies have been done in the recent years studying the impact of cascading in the gamma-ray spectrum of
392: microquasars. Some authors have assumed that particles get deflected after creation, using a three-dimensional code to
393: compute cascading, (e.g. Bednarek \cite{bednarek06}), whereas others have computed one-dimensional cascades in the direction
394: to the observer (e.g. Aharonian et al. \cite{aharonian06b}; Orellana et al. \cite{orellana07}; Khangulyan et al.
395: \cite{khangulyan07}). In case the magnetic field is high enough, synchrotron secondary radiation produced in the system will
396: be significant in the radio and X-ray domains (e.g. Khangulyan et al. \cite{khangulyan07}; Bosch-Ramon et al.
397: \cite{bosch08}). In the next section, we discuss further this issue.
398:
399: \section{Theoretical interpretation of observations}\label{theor}
400:
401: To interpret observations, one can use some analogies comparing one specific object with the general properties of a
402: population of sources. Unfortunately, neither our knowledge on the physics of microquasars emitting at TeV, nor the number of
403: these sources, are sufficient to follow this strategy. Another approach, which seems to us more suitable to study the
404: complexity of the behavior of these objects at very high energies, is to start from basic physical processes to understand
405: the observations in a rough but sound way. It implies to check, with the help of data and the available knowledge on each
406: particular source, which are the most reasonable physical conditions, and processes, that lead to the observed VHE emission.
407: To perform such an analysis, it is required to constrain the conditions in which particle acceleration is possible up to the
408: observed energies, to know which mechanism is behind gamma-ray production (either leptonic -IC- or hadronic -pp, p$\gamma$
409: and nuclei disintegration-), and the impact of photon-photon absorption (and the subsequent energy release channel
410: -synchrotron or IC-). In addition, it is important to explore the impact of particle transport (advection, diffusion or
411: particle escape) on the final spectrum. In the following, each one of these elements is considered individually. It is worth
412: to remark that in reality the mentioned processes will be likely coupled, and a very complex outcome can be expected, as
413: shown in Sects.~\ref{mQc}, \ref{ls} and \ref{lsi}, when focusing in the leptonic scenario.
414:
415: %Radiating particles could also be losing energy via work transferring to surrounding matter, i.e. adiabatical losses, although the lack of
416: %knowledge on the outflow hydrodynamics prevent a sound modeling of this and is neglected here. In addition, to characterize the microquasar
417: %scenario, it is probably unavoidable to take into account the impact of the stellar wind material on the jet evolution and radiative processes, at
418: %least in high-mass systems. In a scenario in which the acceleration is triggered by wind/jet interaction, the accelerator here may be associated to
419: %the places where this interaction is stronger. Therefore, this can be absorbed via the conditions of the accelerator at $Z_0$, although neglecting
420: %at the moment any further effect as long as particles propagate through the jet.
421:
422: \subsection{Non-thermal processes}\label{nonther}
423: %\subsubsection{Particle acceleration and transport}\\
424:
425: In this section, we explore those processes that are relevant to understand the non-thermal emission produced at very high
426: energies for the expected conditions given in microquasars. We do not focus particularly in the jet but also in its
427: environment. In particular, we consider particle acceleration, particle energy losses and transport, and photon-photon
428: absorption and electromagnetic cascading.
429:
430: \subsubsection{Particle acceleration}
431:
432: Our aim is not to model accurately what is observed, but to derive basic and solid constraints from the available data. The
433: first step is to set a necessary condition for acceleration of a charged particle of a certain energy to occur, which is
434: given by the Hillas criterium (Hillas \cite{hillas84}). This consists on the fact that particles can only be accelerated if
435: their Larmor radius ($r_{\rm L}=E/qB_{\rm a}$; where $B_{\rm a}$ is the accelerator magnetic field, and $q$ and $E$ are the
436: charge and energy of the particle, respectively) is smaller than the accelerator size ($l_{\rm a}$), since otherwise they
437: will escape the accelerator. This limits the highest achievable energy to:
438: \begin{equation}
439: E<qB_{\rm a}l_{\rm a}\,,
440: \label{larm}
441: \end{equation}
442: which can be rewritten in the following form:
443: \begin{equation}
444: E<30\,\frac{q}{e}\,B_{\rm a,G}\,l_{11}\, {\rm TeV}\,,
445: \label{maxelarm}
446: \end{equation}
447: where $e$ is the electron charge, $B_{\rm a,G}$ is the accelerator magnetic field ($B_{\rm a}$) in Gauss units,
448: and $l_{11}=l_{\rm a}/10^{11}{\rm cm}$.
449: Nevertheless, to determine whether particles can be accelerated up to a certain energy, the specific acceleration and
450: energy loss (or particle escape)
451: mechanisms are to be known: $t_{\rm acc}=t_{\rm cool/esc}$\,. In general, the acceleration timescale can be expressed as:
452: \begin{equation}
453: t_{\rm acc}=\eta{r_{\rm L} \over c}\simeq 0.1\eta \frac{e}{q}E_{\rm TeV}B_{\rm a,G}^{-1}\,{\rm s}\,,
454: \label{accrate}
455: \end{equation}
456: where $\eta$ is a dimensionless phenomenological parameter (or function) representing the acceleration efficiency, and $E_{\rm
457: TeV}$ is the particle energy in TeV units. The particular case of $\eta=1$ corresponds to the shortest possible acceleration time
458: independently of the acceleration mechanism. Another instance for $\eta$ can be given for the case of
459: non-relativistic diffusive shock acceleration (plane shock with weak magnetic field, in the test particle
460: approximation) (e.g. Protheroe \cite{protheroe99}):
461: \begin{equation}
462: \eta=2\pi{D\over D_{\rm Bohm}}\left(c\over v_{\rm sh}\right)^2\,,
463: \label{accef}
464: \end{equation}
465: where $ v_{\rm sh}$ is the shock velocity, and $D$ is the diffusion coefficient ($D_{\rm Bohm}$ in the Bohm limit). For $v_{\rm
466: sh}=3\times 10^9$~cm~s$^{-1}$ and $D=D_{\rm Bohm}$, $\eta\sim 10^3$.
467:
468: As an example of the importance of using acceleration constraints, we show in Fig.~\ref{acc} a 2-dimensional $d_{\rm
469: a}-B_{\rm a}$ map for a compact high-mass microquasar with stellar luminosity $L_*=10^{39}$~erg~s$^{-1}$, $d_*=3\times
470: 10^{12}$~cm and $kT_*\approx 3$~eV. $d_{\rm a}$ is the distance from the accelerator to the star. As done by Khangulyan et
471: al. \cite{khangulyan07}, it is possible to restrict $d_{\rm a}$ and $\eta$. In the case of LS~5039, for instance, for
472: reasonable $\eta\ge 10$, and the detected 30~TeV photon energies, the accelerator should be outside the system, i.e.
473: $d_{\rm a}> 2\times 10^{12}$~cm.
474:
475: \begin{figure}[pb]
476: \centerline{\psfig{file=2d.eps,width=9cm, angle=0}}
477: \vspace*{8pt}
478: \caption{
479: 2-dimensional $d-B$ map that shows the maximum achievable energy for $\eta=10$ for
480: different $d_{\rm a}$ and $B_{\rm a}$ values. The adopted parameter values are $L_*=10^{39}$~erg~s$^{-1}$,
481: $d_*=3\times 10^{12}$~cm and $kT_*\approx 3$~eV. \label{acc}}
482: \end{figure}
483:
484: {\bf Particle transport}
485:
486: Without focusing on any particular acceleration mechanism, we note that, either the maximum energy is limited by cooling,
487: which implies that particles
488: radiate most of their energy before escaping the accelerator, or their escape stops acceleration.
489: In the
490: latter case, the emitter itself can be considered
491: larger than the acceleration region. The escape time ($t_{\rm esc}$) can be characterized, in
492: the accelerator as well as in the whole emitter, by the minimum among
493: the diffusion ($t_{\rm diff}$) and the advection ($t_{\rm adv}$) times:
494: \begin{equation}
495: t_{\rm esc}=\min(t_{\rm diff}, t_{\rm adv})\,,
496: \end{equation}
497: which can be expressed either as:
498: \begin{equation}
499: t_{\rm diff}={l^2\over2D(E)}\sim 2\times 10^4 l_{\rm 12}^2
500: \left(\frac{D_{\rm Bohm}}{D(E)}\right)B_{\rm G} E_{\rm TeV}^{-1}\,{\rm s}\,,
501: \end{equation}
502: for 1-dimensional diffusion in the accelerator/emitter,
503: where $l$ ($l_{12}=l/10^{12}\, {\rm cm}$) is the (accelerator/emitter) size covered by diffusion,
504: and $B_{\rm G}$ is the emitter magnetic field; or as:
505: \begin{equation}
506: t_{\rm adv}={l\over V_{\rm adv}}\sim 10^4 l_{\rm 12}V_{8}^{-1}\,{\rm s}\,,
507: \end{equation}
508: for advection, where $V_{\rm adv}$ is the advection speed ($V_{8}=V_{\rm adv}/10^{8}$~cm~s$^{-1}$).
509:
510: Under the impact of cooling, the typical distance up to which particles can propagate is:
511: \begin{equation}
512: l_{\rm diff}=10^{10}\,E_{\rm TeV}^{1/2} B_{\rm G}^{-1/2}t_{\rm cool}^{1/2}
513: \left(\frac{D}{D_{\rm Bohm}}\right)^{1/2}\,\,{\rm cm}\,,
514: \label{diffs}
515: \end{equation}
516: under diffusive transport, and:
517: \begin{equation}
518: l_{\rm adv}=10^{10}\left(\frac{V_{\rm adv}}{10^{10}\,{\rm cm/s}}\right)t_{\rm cool}\,\,{\rm cm}\,,
519: \label{advs}
520: \end{equation}
521: under advective transport. $t_{\rm cool}$ is the cooling timescale (in $s$) of the dominant loss mechanism.
522: We discuss in the following different cooling processes that may be
523: relevant in the microquasar context.
524:
525: \subsubsection{Radiative processes}\label{rad}
526:
527: {\bf Synchrotron}
528:
529: In the presence of a disordered magnetic field, electrons radiate via synchrotron emission.
530: The synchrotron cooling timescale
531: is approximately:
532: \begin{equation}
533: t_{\rm sy}\approx 4\times 10^2 \left(\frac{e}{q}\right)^4(m/m_{\rm e})^4B_{\rm a,G}^{-2} E_{\rm TeV}^{-1}\, {\rm s}\,,
534: \label{sy_loss}
535: \end{equation}
536: where $m$ and $m_{\rm e}$ are the mass of the particle and the electron, respectively.
537: This gives the maximum particle energy:
538: \begin{equation}
539: E_{\rm max} \approx 60 \, \left(\frac{e}{q}\right)^{3/2}(m/m_{\rm e})^2\,(\eta B_{\rm a,G})^{-1/2} \, {\rm TeV}\,.
540: \label{syn_max_en}
541: \end{equation}
542: This process is only relevant for leptons unless the magnetic field is very strong,
543: in which case hadronic synchrotron may become efficient.
544:
545: To see what may happen in a microquasar regarding $E_{\rm max}$ under synchrotron cooling, we take $B_{\rm 0,G}=10^6$~G as a reasonable value for the jet base magnetic field (at
546: $\sim 50\,R_{\rm Sch}\sim 10^8$~cm from the compact object), obtaining: $E_{\rm max}\approx 0.06\,\eta^{-1/2}$~TeV. If we assume a distance dependence of $B$ of the form $1/Z$ and
547: locate the accelerator at $10^{12}$~cm, we obtain: $E_{\rm max}\approx 6\,\eta^{-1/2}$~TeV, i.e. in such a situation, it seems unlikely to produce VHE leptons in the jet even at
548: the system scales.
549:
550: At spatial scales similar or smaller than the binary system, it seems reasonable to expect large magnetic fields related either to the jet, to the accretion disk, or to the
551: companion star. It implies that synchrotron emission can be an efficient radiation process, and electrons may release most of their energy via synchrotron emission. At larger
552: scales, if no significant magnetic field enhancement takes place, efficiencies will decrease strongly.
553:
554: It is worthy using Eqs.~(\ref{diffs}) and (\ref{advs}) to calculate the impact of synchrotron losses on the propagation of
555: electrons via diffusion and advection, since they tell how far relativistic electrons can go without assuming additional acceleration and
556: neglecting other sources of losses. Under the next parameter choice, e.g. $B\sim 1$~G, $V_{\rm
557: adv}=10^{10}$~cm~s$^{-1}$, $D=D_{\rm Bohm}$, advection is the most effective transport mechanism, and TeV particles
558: may reach distances of $\sim 3\times 10^{12}$~cm from their injection point.
559:
560: {\bf Inverse Compton}
561:
562: Under the radiation field of the primary star in a high-mass microquasar, or in the case of strong accretion disk/corona
563: emission, IC scattering is to be considered and may limit particle acceleration. Synchrotron self-Compton could become a
564: dominant process, although it is not treated here, since it would require significantly large magnetic fields, making
565: acceleration up to very high energies unlikely due to strong synchrotron/Thomson IC energy losses. In addition, once the
566: electron enters in the KN IC regime, the interaction efficiency reduces strongly and synchrotron cooling becomes dominant,
567: making SSC not very efficient in the TeV range under UV photon fields. We note that, due to the same reason as in the
568: synchrotron case, IC losses are only significant for leptons\footnote{For hadrons, other cooling processes, mainly
569: photo-meson production, discussed further in the text, will overcome hadronic IC scattering.}.
570:
571: For a Planckian distribution of target photons with temperature $T$, the IC energy loss rate can be approximated,
572: with an accuracy of less than a 3\%, by:
573: \begin{equation}
574: \dot{\gamma}_{IC}=5.5\times 10^{17}T_{\rm mcc}^3\gamma {ln(1+0.55\gamma T_{\rm mcc})\over 1+25T_{\rm mcc}\gamma}
575: \left(1+{1.4\gamma T_{\rm mcc}\over1+12\gamma^2T_{\rm mcc}^2}\right)\,{\rm s}^{-1},
576: \label{gammabb}
577: \end{equation}
578: where $T_{\rm mcc}=kT/m_{\rm e}c^2$.
579:
580: Regarding the particle maximum achievable energy, at the energies in which we are interested in this work,
581: IC scattering proceeds in the Klein-Nishina (KN) regime ($\gamma\gg 1/kT$ in $m_{\rm e}c^2$ units).
582: In such a case,
583: a simple power-law fit for the cooling time can be used for a black-body type of target photon distribution:
584: \begin{equation}
585: t_{\rm IC}\approx 10^2\,
586: \left(R\over R_*\right)^2
587: T_{4}^{-2.3}\,
588: E_{\rm TeV}^{0.7}\,{\rm s}.
589: \label{tkn1}
590: \end{equation}
591: where $R$ and $R_*$ are the distance to the origin and the radius of the source of target photons,
592: and $T_4=T_*/10^4\, {\rm K}$.
593: Eq.~(\ref{tkn1}) can be expressed in a more convenient form in case of a hot and massive primary star as the
594: dominant source of target photons:
595: \begin{equation}
596: t_{\rm IC}=10^2
597: (w_{100})^{-1}
598: \left(T_*\over 3\times 10^4\,{\rm K}\right)^{1.7}\,
599: E_{\rm TeV}^{0.7}\,{\rm s}\,,
600: \label{tkn2}
601: \end{equation}
602: where $w=100\,w_{100}$~erg~cm$^{-3}$ is the target photon field energy density, and $T_*$ the stellar temperature.
603: This expression gives a reasonable agreement for the IC cooling time, within a factor $\le 2$, in the relevant energy range:
604: \begin{equation}
605: 0.1\ {\rm TeV}
606: \left(T_{\rm star}\over 3\times 10^4{\rm K}\right)^{-1}<
607: E<
608: 3\times 10^3\ {\rm TeV}\left(T_{\rm star}\over 3\times 10^4{\rm K}\right)^{-1}\,.
609: \end{equation}
610: The corresponding value of the maximum energy is:
611: \begin{equation}
612: E_{\rm max}\approx
613: 4\times 10^{10} \,
614: [B_{\rm a~G}\eta^{-1}w_{100}^{-1}]^{3.3}~{\rm TeV}\,,
615: \label{emax}
616: \end{equation}
617: which shows that KN IC limits particle acceleration much less than synchrotron radiation, and can be hardly dominant
618: at the maximum particle energy for reasonable magnetic fields in any region of the microquasar.
619:
620: To estimate the impact of IC losses on the propagation of electrons via diffusion and advection, Eqs.~(\ref{diffs}) and
621: (\ref{advs}) can be used as well. Like in the case of synchrotron losses, for the conditions $B\sim 1$~G, $V_{\rm
622: adv}=10^{10}$~cm~s$^{-1}$ and $D=D_{\rm Bohm}$, and IC loss dominance, advection is again the most efficient transport
623: mechanism, under which TeV particles could reach distances up to $\sim 10^{12}$~cm.
624:
625: {\bf Proton-proton interactions}
626:
627: As mentioned in the previous section, purely hadronic processes like pp interactions
628: have been discussed in the past in the context of microquasars. We consider them here as well,
629: since they may be relevant in some cases. The energy loss timescale for pp collisions (Kelner et al. \cite{kelner06}) is:
630: \begin{equation}
631: t_{\rm pp}\approx {10^6\over n_9}\, {\rm s}\,,
632: \label{tpp}
633: \end{equation}
634: where $n_9=n_{\rm t}/10^9~{\rm cm}^{-3}$ is the target density.
635: The energy threshold of this process in the reference frame of the interaction center of masses
636: is the pion rest mass, $\approx 140$~MeV.
637: From Eqs.~(\ref{accrate}) and (\ref{tpp}), the maximum particle energy can be derived:
638: \begin{equation}
639: E_{\rm max}\approx 10^7 {B_{\rm a,G}\over \eta n_9}\, {\rm TeV}\,.
640: \label{accrate}
641: \end{equation}
642: Given the long cooling timescale, the maximum energy will be likely
643: limited by the accelerator size.
644:
645: Defining $L_{\rm p}$ as the luminosity injected in the form of relativistic protons, the
646: luminosities in gamma-rays, neutrinos and secondary electron-positron pairs are,
647: with differences of about a factor of 2 (Kelner et al. \cite{kelner06}), the next ones:
648: \begin{equation}
649: L_{\gamma}\approx \min(1,t_{\rm esc}/t_{\rm pp})\,c_{\rm pp}\,L_{\rm p}\,,
650: \label{lumpp}
651: \end{equation}
652: where $c_{\rm pp}$ is the energy transfer efficiency from relativistic protons to secondary particles ($\sim 10$\%).
653: In the context of high-mass microquasars, a reasonable lower limit for $t_{\rm esc}$ is the wind advection time, i.e. the
654: time required for the stellar wind to cross the orbital radius ($R_{\rm orb}/V_{\rm w}$; where $R_{\rm orb}\sim 10^{12}-10^{13}$~cm
655: and $V_{\rm w}\sim 1-3\times 10^8$~cm~s$^{-1}$), $t_{\rm esc}\le 10^4$~s; and a very lower limit is set by
656: the speed of light, i.e. $t_{\rm esc}\ge 10^2$~s. All this, plus adopting a $\dot{M}=10^{-6}$~M$_{\odot}$~yr$^{-1}$
657: (typical for O stars), yields:
658: $$
659: L_{\gamma}\sim 10^{-5}-10^{-3}L_{\rm p}\,.
660: $$
661: Larger efficiencies cannot be discarded in some specific cases, like density enhancements in the wind or even in the
662: jet itself via, e.g., shocks. Once relativistic protons leave the binary system, the expected density decreases in the
663: stellar wind as well as in the jet, making pp collisions negligible.
664:
665: {\bf Photo-meson production}
666:
667: Among different hadronic processes, photo-meson production (Kelner et al. \cite{kelner08}) is worthy also to be considered.
668: The energy threshold for this process is:
669: \begin{equation}
670: E_{\rm th~p\gamma}={m_{\rm p}c^2\epsilon_{\rm th~p\gamma}}/2\epsilon=(5\times 10^4\,{\rm TeV})\,(T_4)^{-1}\,,
671: \label{thres1}
672: \end{equation}
673: where $m_{\rm p}$ is the proton mass and $\epsilon$ and $\epsilon_{\rm th~p\gamma}\approx 140$~MeV
674: are the energy of the target photon in the laboratory and the hadron rest frames, respectively. The loss rate is given by:
675: \begin{equation}
676: t_{\rm p\gamma}\sim{10^{18}\over N_{\rm X}}\, {\rm s}\,,
677: \label{pgam1}
678: \end{equation}
679: where:
680: \begin{equation}
681: N_{\rm X}\approx {L\over4\pi\epsilon R^2c}\approx 2\times 10^{14}L_{38}T_4^{-1}R_{12}^{-2}\,{\rm cm}^{-3}\,.
682: \end{equation}
683: $L_{38}$ is the star luminosity $L_*/10^{38}$~erg~s$^{-1}$, $R_{12}$ is the distance to the star,
684: and $R/10^{12}\, {\rm cm}$. Eq.~(\ref{pgam1}) can be rewritten as:
685: \begin{equation}
686: t_{\rm p\gamma}\sim 10^4L_{38}^{-1}T_4R_{12}^2\, {\rm s}\,.
687: \end{equation}
688: With this cooling timescale, the corresponding maximum energy is:
689: \begin{equation}
690: E_{\rm max}\sim 10^5 \eta^{-1} B_{\rm a, G} L_{38}^{-1} T_4R_{12}^2\,{\rm TeV}\,.
691: \label{emaxpm}
692: \end{equation}
693:
694: As for pp interactions, the long cooling timescales of photo-meson production imply that the maximum energy is in fact
695: limited by the accelerator size. We note that only in case a substantial part of the energy in accelerated protons were
696: $>E_{\rm th~pm}\sim 10^4$~TeV, under an UV photon field, and the fastest protons escaped at the speed of light, photo-meson
697: production could reach efficiencies of $L_{\gamma}\sim 10^{-3}L_{\rm p}$\footnote{At the involved proton energies, the
698: stellar wind can hardly confine the particles, thereby we adopt the speed of light as escape velocity.} for $E_{\rm p}>E_{\rm
699: >th~p\gamma}$. The energy transfer
700: efficiency for this process is $c_{\rm p\gamma}=0.1$. In the inner regions of the jet, close to the accretion disk and corona
701: photon fields, the radiation energy density may be high enough to get even higher efficiencies. The larger target photon
702: energy would imply a reduced threshold energy. Nevertheless, the constraint imposed by
703: Eq.~(\ref{maxelarm}) is quite restrictive and could prevent hadrons from reaching energies $>E_{\rm th~p\gamma}$ for
704: reasonable $B_{\rm a}$ and $l_{\rm a}$ values. Namely, comparing Eqs.~(\ref{emaxpm}), (\ref{thres1}) and (\ref{maxelarm}), two
705: conditions are obtained:
706: \begin{equation}
707: B_{\rm G}T_4l_{11}>2\times 10^3\,,
708: \label{cond1}
709: \end{equation}
710: for basic physical conditions; and
711: \begin{equation}
712: \eta<B_{\rm a,G}T_4^2R_{12}^2/2L_{38}\,.
713: \label{cond2}
714: \end{equation}
715: for the acceleration efficiency.
716:
717: As in the case of pp collisions, the strong dilution of the photon field far
718: from its source prevents photo-meson production from being efficient outside the binary system.
719:
720: {\bf Photo-disintegration}
721:
722: If ultra relativistic
723: heavy nuclei are present, they can suffer photon disintegration under the ambient photon field. The expression for the
724: threshold energy is similar to that presented in Eq.~(\ref{thres1}):
725: \begin{equation}
726: E_{\rm th~pd}={m_{\rm N}c^2\epsilon_{\rm th~pd}}/2\epsilon\,,
727: \label{thres}
728: \end{equation}
729: where $m_{\rm N}$ is the mass of the nucleus and $\epsilon_{\rm th~pd}=8$~MeV. Effectively,
730: since $m_{\rm N}$ can be up to $\sim 100\,m_{\rm p}$, $E_{\rm th~pd}$ could be $>E_{\rm th~p\gamma}$.
731:
732: The disintegration of the nuclei has as typical timescale:
733: \begin{equation}
734: t_{\rm pd}\sim 10^3L_{38}^{-1}T_4R_{12}^2\,{\rm s}\,.
735: \label{dis}
736: \end{equation}
737: For simplicity, we have assumed that the mass of the nucleus is its charge times the mass of the proton,
738: which slightly overestimates the efficiency within
739: a factor of $\sim 2$. In addition, a slow dependence on the mass of the nuclei in the cross section has
740: been neglected. Photo-disintegration would stop the acceleration of heavy nuclei at energies:
741: \begin{equation}
742: E_{\rm max}\sim 10^4\frac{q}{e} B_{\rm a, G} L_{38}^{-1}T_4R_{12}^2\,{\rm TeV}\,.
743: \label{dis}
744: \end{equation}
745: As in the case of pp collisions and photo-meson production, the accelerator size instead of photo-disintegration cooling
746: will probably stop acceleration of heavy nuclei.
747: Taking into account that the energy transfer efficiency is $c_{\rm p\gamma}\sim 0.01$, and
748: adopting the speed of light to derive a lower-limit for $t_{\rm esc}$,
749: photo-disintegration can yield in gamma-rays $L_{\gamma}\sim 10^{-3}L_{\rm N}$,
750: where $L_{\rm N}$ is the energy stored in heavy nuclei with $E>E_{\rm th~pd}$.
751: Hillas constraint allows energies $m_{\rm N}/m_{\rm p}$ times larger for nuclei than for protons, although
752: as noted above
753: $E_{\rm th~pd}$ can be $>E_{\rm th~p\gamma}$\footnote{Actually, similar constraints to those derived for photo-meson production
754: regarding the interaction physical conditions and the acceleration efficiency apply here as well.}. Like in the case of photo-meson production,
755: it is unclear whether enough energy can be in the form of ultra relativistic
756: heavy nuclei above the threshold to produce a significant signal via photo-disintegration.
757:
758: {\bf Photon-photon absorption and secondary radiation}
759:
760: The presence of the star, a powerful source of UV photons, is very relevant to understand TeV emission from compact sources.
761: A massive and hot star is an excellent provider of target photons for photon-photon absorption. The anisotropy of the stellar
762: photon field and the relative position of the system with respect to the observer makes the exploration of the observational
763: impact of photon-photon absorption a non-trivial subject. In Fig.~\ref{absor}, we show a 2-dimensional map of the
764: opacity coefficient ($\tau$) for different energies of the incoming photon ($E_{\gamma}$) emitted
765: near the compact object, and different directions starting
766: with the one pointing away from the star. Here, $\theta$ is the angle with the line joining the TeV emitter and the star. The
767: point-like approximation for the target photon field has been used, which implies that $\tau\propto 1/d$. The used parameter
768: values are $L_*=10^{39}$~erg~s$^{-1}$, $d_*=3\times 10^{12}$~cm and $kT_*=3$~eV. It is worthy noting that the point-like
769: approximation for the target photon emitter works in a wide range of situations, but there are cases relevant for compact TeV
770: sources in which is necessary to account for the finite size of the star (see also Dubus \cite{dubus06}). In Fig.~\ref{fin},
771: we show a 2-dimensional map of the ratio $\tau_{\rm p}/\tau_{\rm f}$ (i.e. point-like versus finite size $\tau$) for
772: different distances to the star ($d$) and values of $\theta$. The radius of the star has been taken $R_*=7\times 10^{11}$~cm.
773: The energy of the incoming photon has been taken 1~TeV. The deep blue/black regions in this map represent the cases when the
774: point-like approximation gives almost no absorption, i.e. when the emitter points roughly away from the star, or when the
775: photons in the finite size approximation collide with the star surface giving infinite opacity.
776:
777: The importance of cascading has been already stated above. It is worth now to study the possibility of cascade development in
778: the surroundings of massive and hot stars. There are two extreme situations. Either the magnetic field energy density is much
779: lower compared with the radiation energy density, and in such a case pure cascade will develop, or the magnetic field energy
780: density is a significant fraction, or above, the radiation one, and then the electron energy will be mostly released via
781: synchrotron radiation and single-scattering IC. If the magnetic field is low enough (and therefore suitable for cascading to
782: occur), very high-energy particles will be only slightly deflected before radiating their energy via IC. This allows us for a
783: 1-dimensional approach to compute cascading. For the same parameter values as those presented above, we have performed a
784: 1-dimensional electromagnetic cascade simulation, the result of which is shown in Fig.~\ref{casc}. The injected particle
785: spectrum was a power-law of photon index 2.
786:
787: To give an idea of the magnetic field importance,
788: next formula shows the critical value of the magnetic field that allows cascading to occur for
789: TeV electrons:
790: \begin{equation}
791: B_{\rm c}=10\,\frac{R_*}{R}\left(L_*\over 10^{39}\,{\rm erg~s}^{-1}\right)^{1/2}\,{\rm G}\,,
792: \label{eq:b_crit}
793: \end{equation}
794: In fact, $B_{\rm c}$ is valid for the 1-dimensional cascade case. If electrons suffer strong deflection in the ambient
795: magnetic field, the longer time required by particles to escape the region of dense photon field will increase the
796: synchrotron outcome with respect to the IC one. This happens since photons convert to electrons more times before escaping,
797: releasing at the end more energy in form of synchrotron and low energy IC radiation. From all this, and the fact that
798: magnetic fields of hundreds of G are typical in the surface of OB stars, we conclude that effective electromagnetic cascading
799: is unlikely in the environment of high-mass systems. For the same conditions as those taken for Fig. 6, plus a stellar surface
800: magnetic field of 100~G and a primary gamma-ray injection luminosity of $3\times 10^{35}$~erg~s$^{-1}$, we show the secondary
801: pair spectral energy distribution (SED) in Fig.~\ref{sedsec}. We note the moderately high radio and X-ray luminosities (see
802: also Bosch-Ramon et al. \cite{bosch08}).
803:
804: \begin{figure}[pb]
805: \centerline{\psfig{file=big2.eps,width=9cm, angle=270}}
806: \vspace*{8pt}
807: \caption{2-dimensional $\theta-E_{\rm gamma}$ map of the opacity coefficient.
808: The adopted parameter values are $L_*=10^{39}$~erg~s$^{-1}$, $d_*=3\times 10^{12}$~cm and $kT_*=3$~eV. \label{absor}}
809: \end{figure}
810:
811: \begin{figure}[pb]
812: \centerline{\psfig{file=ratio2.eps,width=9cm, angle=270}}
813: \vspace*{8pt}
814: \caption{2-dimensional $\theta-d$ map of the ratio $\tau_{\rm p-l}/\tau_{\rm fin}$ for an incoming photon of 1~TeV.
815: The adopted parameters are the same as those of Fig.~\ref{absor}, plus the stellar radius $R_*=7\times 10^{11}$~cm.
816: The shaded area to the left of the plot corresponds to a region of opacities $\sim 0$. \label{fin}}
817: \end{figure}
818:
819: \begin{figure}[pb]
820: \centerline{\psfig{file=cascade2.eps,width=9cm, angle=270}}
821: \vspace*{8pt}
822: \caption{1-dimensional electromagnetic cascade. Note the difference
823: of about a factor of 10 between the 1~TeV and the 10~GeV fluxes.
824: The adopted parameters are those of Fig.~\ref{absor}.
825: \label{casc}}
826: \end{figure}
827:
828: \begin{figure}[pb]
829: \centerline{\psfig{file=sedsec.eps,width=9cm, angle=0}}
830: \vspace*{8pt}
831: \caption{Computed SED of the synchrotron and IC emission from the pairs created in a compact
832: binary. The TeV emitter location is at $d=3\times 10^{12}$~cm. The adopted parameters are those of
833: Fig.~\ref{absor}, plus a stellar surface magnetic field of 100~G and a primary gamma-ray injection luminosity
834: of $3\times 10^{35}$~erg~s$^{-1}$.\label{sedsec}}
835: \end{figure}
836:
837: \subsection{Modeling non-thermal emission in microquasars: the case of Cygnus~X-1}\label{ls}
838:
839: In Sect.~\ref{nonther}, we have reviewed the main radiative processes that could take place under typical conditions in
840: microquasars. After roughly estimating the efficiency of different leptonic and hadronic processes, it is our aim now to
841: focus on the mechanism that, to our understanding, is the most likely to produce the VHE radiation observed in some of these
842: sources. From the point of view of efficiency, IC scattering is clearly a good candidate in this sense. In addition, the fact
843: that a hot and massive star is present in all these systems detected at TeV energies makes IC scattering very
844: attractive\footnote{ It does not discard low-mass microquasars as VHE emitters. Actually, from the observational point of
845: view, hints of TeV emission from GRS~1915$-$105 were found using imaging Cherenkov techniques by HEGRA (Aharonian
846: \& Heinzelmann \cite{aharonian98b}) and Whipple (Rovero et al. \cite{rovero02}). In addition, a number of theoretical works
847: have proposed these sources as VHE emitters (e.g. Atoyan \& Aharonian \cite{atoyan99}; Bosch-Ramon et al. \cite{bosch06b};
848: Romero et al. \cite{romero08}).}. Finally, the pattern of variability of the spectrum and flux along the orbit found in
849: LS~5039 can be, together with photon-photon absorption, explained due to the angular dependence of the cross section of IC
850: scattering and the anisotropic nature of the stellar photon field (e.g. Khangulyan \& Aharonian \cite{khangulyan05};
851: Khangulyan et al. \cite{khangulyan07}).
852:
853: To carry out a simple but thorough treatment of the processes relevant to VHE emission in microquasars, we have simplified
854: the scenario as much as possible. We show a schema of this scenario in Fig.~\ref{sq}. We consider the jet as a weakly
855: relativistic convective flow without going into the details of its physical nature, i.e. whether it is a magneto- or a
856: hydrodynamical plasma, or even a pointing flux dominated flow. This outflow is ejected perpendicular to the orbital plane
857: bearing a disordered magnetic field attached to it. Efficient acceleration of non-thermal particles can occur at a certain
858: location in the jet ($Z_0$ from the compact object, $d_0$ from the star); such acceleration regions are treated as point-like,
859: and their position in the jet, $Z_0$, could change. For simplicity, we adopt a model in which acceleration takes place only
860: in one point.
861:
862: Because of the strong uncertainties affecting the physics of microquasar jets, we do not treat but very generally the
863: acceleration processes that may take place in there. The injection electron spectrum adopted here is phenomenological: a
864: power law of exponent -2 plus an exponential high-energy cutoff. The accelerator sets up the initial conditions of the
865: emitter, which is treated as a 1-dimensional region where particles are transported by jet convection. Adiabatic losses have
866: been neglected here. The magnetic field depends on $Z$ like $B=B_{\rm a}(Z_0/Z)$. To model the non-thermal processes included
867: in our scenario, it is necessary to consider the presence of different ingredients: the magnetic field, possibly present
868: radiation fields, and the material in which the emitting particles may be embedded. As discussed above, we will deal here
869: with synchrotron and stellar IC emission, since we restrict ourselves to high mass systems and to the most relevant processes
870: to produce or affect VHE radiation. Nevertheless, other mechanisms cannot be discarded. We take into account the impact of
871: photon photon absorption on the VHE spectrum. We recall the importance of the geometry of IC scattering and photon photon
872: absorption for the spectra and lightcurves when the star is the main source of target photons. For further mathematical
873: details of the model, we refer to Khangulyan et al. \cite{khangulyan07}.
874:
875: In the following, we apply this model to the microquasar Cygnus~X-1 (Sect.~\ref{mQc}), and to the other two X-ray binaries
876: with extended radio emission, LS~5039 and LS~I~+61~303 (Sect.~\ref{lslsi}).
877:
878: \begin{figure}[pb]
879: \centerline{\psfig{file=systemls.eps,width=9cm}}
880: \vspace*{8pt}
881: \caption{Schema of the system. \label{sq}}
882: \end{figure}
883:
884: \subsubsection{Application to Cygnus~X-1}\label{mQc}
885:
886: Cygnus~X-1 is a HMXB of uncontroverted accreting nature with relativistic radio jets (Stirling et al. \cite{stirling01}). The
887: compact object is a black-hole of $\sim 20$~M$_{\odot}$, the primary is an O9.7Iab star of $\sim 40$~M$_{\odot}$, and the
888: system is located at $\approx 2.1$~kpc (Ziolkowski \cite{ziolkowski05}). The orbit is thought to be circular, with an
889: inclination $i\sim 30^{\circ}$, orbital radius of $\approx 3\times 10^{12}$~cm and period of 5.6~days (Gies \& Bolton
890: \cite{gies86}). The primary star presents a luminosity of $\approx 10^{39}$~erg~s$^{-1}$ and a temperature of $\approx
891: 30000$~K. An evidence of detection from this source at the 4.1~$\sigma$ level has been found by MAGIC in the TeV range during
892: a transient event that may be correlated with the hard X-ray behavior (Albert et al. \cite{albert07}). At present, the origin
893: of the VHE emission is unclear. In the context of the hadronic scenario, Romero et al. \cite{romero03} and Orellana et al.
894: \cite{orellana07} computed the emission for a microquasar with the characteristics of Cygnus~X-1. Regardless the hadronic or
895: leptonic origin of the gamma-ray radiation, electromagnetic cascading and/or secondary synchrotron emission should occur in
896: the system (e.g. Bednarek \cite{bednarek07}, Orellana et al. \cite{orellana07}, Bosch-Ramon et al. \cite{bosch08}).
897:
898: {\bf On the spectral energy distribution and orbital variability}
899:
900: It is worthy to see the aspect of the multiwavelength non-thermal emission inferred from the VHE radiation detected from the
901: source. This is shown in Fig.~\ref{sedcyg}. We have located the accelerator/emitter at $Z_0=3\times 10^{12}$~cm. The
902: magnetic field is $\sim$~G. The synchrotron X-ray flux is well below the observed level, of likely thermal (comptonized)
903: origin (Sunyaev \& Truemper \cite{sunyaev79}). This SED has been calculated for an emitter located in the borders of the
904: system, in which opacities are moderate. However, deeper, closer to the compact object, the attenuation factor grows orders
905: of magnitude.
906:
907: For Cygnus X-1, the spectrum of the emission from $\sim 0.1-1$~TeV electrons is available for a narrow orbital phase range
908: around $\phi=0.9$, i.e. this is not averaged for wide ranges of the orbit. In addition, these electrons have the shortest
909: lifetimes under IC cooling, implying that they radiate most of their energy before propagating significantly. These two facts
910: mean that the observed radiation was produced under similar conditions. In addition, the stellar photon field is very dense
911: in the compact object surroundings, implying very large opacities in almost all the directions. All this, plus the known
912: orbital observer/system geometry at the observation phase, allows us, and makes interesting, to estimate the absorbed energy
913: via photon-photon absorption depending on the distance between the emitter and the compact object. The calculation of the
914: energy processed in this system via photon-photon absorption can give a model-independent constraint on the the location of
915: the TeV emitter.
916:
917: In Fig.~\ref{abscyg}, we show the luminosity divided by $4\pi d_{\odot}^2$ ($d_{\odot}$: distance to the Earth) of the
918: secondary pairs injected in the system as a function of the distance to the compact object. This is computed calculating
919: first the deabsorbed VHE spectrum and flux given the distance to the star and the observed spectrum and flux (Albert et al.
920: \cite{albert07}). In our calculations we have not taken into account the effect of cascading, although the likely
921: moderate-to-high magnetic fields present in massive binary systems would imply a dominant synchrotron channel suppressing
922: cascading. Since in our
923: context a significant fraction of this energy rate is released via synchrotron emission, and the X-ray/soft gamma-ray
924: luminosities are typically of $\sim 10^{37}$~erg~s$^{-1}$ (e.g. McConnell et al. \cite{mcconnell02}), the emitter location
925: can be set, conservatively, to at least few times $10^{12}$~cm (see also Bosch-Ramon et al. \cite{bosch08b}). This is a
926: strong indication that the TeV emitter is located in a jet far away from the compact object. It is relevant also to note that
927: the secondary synchrotron radiation may also explain the observed soft gamma-ray fluxes (e.g. McConnell et al.
928: \cite{mcconnell02}). Another interesting result is the absorbed luminosity at distances $\sim 10^{13}$~cm, indicating that
929: there can be significant injection of relativistic pairs at distances where their radio emission may be resolved using VLBI
930: techniques (see also Bosch-Ramon et al. \cite{bosch08}).
931:
932: Regarding the time evolution of the observed radiation, the flaring nature of the VHE emission points to intrinsic changes of
933: the emitter properties, instead of geometric effects or target density variations, as the origin of the variability.
934:
935: \begin{figure}[pb]
936: \centerline{\psfig{file=bbcyg.eps,width=9cm, angle=0}}
937: \vspace*{8pt}
938: \caption{Computed synchrotron and IC SED for Cygnus~X-1, together with the observed VHE SED, is shown. The accelerator/emitter
939: location has been set to $Z_0=3\times 10^{12}$~cm. The total contribution from the jet and the counter-jet is
940: shown.
941: \label{sedcyg}}
942: \end{figure}
943:
944:
945: \begin{figure}[pb]
946: \centerline{\psfig{file=energy.ps,width=9cm, angle=270}}
947: \vspace*{8pt}
948: \caption{Luminosity divided by $4\pi d_{\odot}^2$ of the secondary pairs injected in the system with the emitter at different
949: distances from the compact object. This is computed calculating first the deabsorbed VHE spectrum and flux given the distance
950: to the star and the observed spectrum and flux. The case in which the radiation process is isotropic, like proton-proton
951: collisions may be (solid line), and the case in which the radiation process is IC (dot-dashed line) -strongly anisotropic-,
952: are shown. The production curves for both cases are also shown.
953: \label{abscyg}}
954: \end{figure}
955:
956:
957:
958: \subsection{TeV binaries with extended radio emission: LS~5039 and LS~I~+61~303}\label{lslsi}
959:
960: \subsubsection{LS~5039}\label{ls}
961:
962: LS~5039 is an HMXB (Motch et al. \cite{motch97}) located at 2.5$\pm 0.5$~kpc (Casares et al. \cite{casares05}). The source
963: presents radio jets (Paredes et al. \cite{paredes00,paredes02}), shows X-ray variability possibly associated to the orbital
964: motion (Bosch-Ramon et al. \cite{bosch05}), and has been detected at very high-energy (VHE) gamma-rays (Aharonian et al.
965: \cite{aharonian05}), which virtually confirms its association with a CGRO/EGRET source (Paredes et al. \cite{paredes00}).
966: Interestingly, the TeV emission varies with the orbital period (Aharonian et al. \cite{aharonian06}). The most recent orbital
967: ephemeris of the system were obtained by Casares et al. \cite{casares05}. The orbital period is $3.9060$~days, the
968: eccentricity is moderate $e=0.35\pm 0.04$, and the inclination angle is not well constrained, $i\approx
969: 15^{\circ}$--$60^{\circ}$. The orbital semi-major axis of the system is $\approx 2\times 10^{12}$~cm. The nature of the
970: compact object is still uncertain. Casares et. al. \cite{casares05} suggested that it may be a black hole\footnote{Assuming
971: pseudo-synchronization of the orbit.}, although there is an on-going debate on this issue, and some authors have proposed
972: that LS~5039 is in fact a young non-accreting pulsar (see, e.g., Martocchia et al. \cite{martocchia05}; see also Dubus
973: \cite{dubus06b}). In this regard, the strongest argument would the lack of accretion features in the X-ray spectrum of
974: LS~5039.
975:
976: \begin{figure}[pb]
977: \centerline{\psfig{file=lslchess.eps,width=9cm, angle=0}}
978: \vspace*{8pt}
979: \caption{At the top, the lightcurve of the photon flux above 1~TeV of LS~5039 presented by Aharonian
980: et al. (2006). Two full phase periods are shown for clarity. The blue solid arrows depict periastron
981: and apastron. The thin red dashed lines represent the superior and inferior conjunctions of the
982: compact object, and the thick red dashed line depicts the Lomb-Scargle Sine coefficients for the
983: period giving the highest Lomb-Scargle power (see Aharonian et al. 2006 for details). At the
984: middle, the fitted pure power-law photon index (for energies 0.2 to 5~TeV) versus phase interval of
985: width $\phi=0.1$ is presented. At the bottom, the power-law normalization (at 1~TeV) versus. phase
986: interval of width $\Delta\phi=0.1$ is shown. \label{lslc}}
987: \end{figure}
988:
989: \begin{figure}[pb]
990: \centerline{\psfig{file=lssphess.eps,width=9cm, angle=0}}
991: \vspace*{8pt}
992: \caption{Spectra of LS~5039 for the two broad orbital phase intervals $0.45<\phi\le 0.9$ (red circles),
993: and $\phi\le 0.45$ and $\phi>0.9$ (blue triangles), from Aharonian et al. (2006). The shaded regions
994: represent the 1~$\sigma$
995: confidence bands on the fitted functions (see Aharonian et al. 2006 for details). \label{lssed}}
996: \end{figure}
997:
998: The radio emission in LS~5039 is of non-thermal origin, slightly variable at month--year timescales ($\sim 20$--$30$\%) and
999: extended -$\sim$1~--~100~milliarcseconds (mas)- (Mart\'i et al. \cite{marti98}; Rib\'o et al. \cite{ribo99}; Paredes et al.
1000: \cite{paredes00,paredes02}, Rib\'o et~al. \cite{ribo08}; for an exhaustive study of this source, see Rib\'o \cite{ribo02}),
1001: the $\sim 60$--$80$\% being produced within a core of $\sim$~mas. The radio emission, when observed at VLA angular
1002: resolution, appears unresolved, and optically thin (Mart\'i et al. \cite{marti98}).
1003:
1004: In the X-rays, the source shows flux variations by a factor of $\sim 2$ peaking smoothly around phase 0.8 and more sharply at
1005: other phases (Bosch-Ramon et al. \cite{bosch05}). These peaks were apparently accompanied by spectral hardening. At phases
1006: $\sim 0.8$, higher and harder emission than in the rest of the orbit has been also observed at TeV energies (Aharonian et al.
1007: \cite{aharonian06}), and simultaneous Chandra observations have apparently shown a similar behavior in X-rays (Horns et al.
1008: \cite{horns06}; Bosch-Ramon et al. \cite{bosch07b}).
1009:
1010: At VHE, the flux and the photon index change periodically, by a factor of $\sim 4$ the former and between $\sim 1.9$--$3.1$
1011: the latter, with the spectrum hardening when flux increases (Aharonian et al. \cite{aharonian06}). The observed lightcurve
1012: and VHE SED are shown in Figs.~\ref{lslc} and \ref{lssed}, respectively. The maximum of the emission takes place around phase
1013: 0.8, similarly as it seems to happen at X-rays. Also, sudden increases/hardening in the flux and spectrum on hour timescales
1014: could have been detected at phase $\sim$0.85 (de Naurois et al. \cite{denaurois06}), being similar to what has been mentioned
1015: above concerning sharp peaks in X-rays. The TeV emission from LS~5039 has been studied by several authors in the microquasar
1016: (e.g. Paredes et al. \cite{paredes06}) and the pulsar context (e.g. Dubus et al. \cite{dubus07}; Sierpowska-Bartosik \&
1017: Torres \cite{sierpowska07,sierpowska08}). A more general approach to study the VHE emission from LS~5039 has been recently carried out by
1018: Khangulyan et al. \cite{khangulyan07}.
1019:
1020: In the following, we present the multiwavelength SED and the orbital VHE lightcurve computed with the model sketched above
1021: for LS~5039. The values given to the free parameters of the model are chosen with the intention to show that a simple
1022: leptonic model can roughly explain the observed features of the emission taking into account the TeV data and the
1023: phenomenology of the source. At this stage, no electromagnetic cascading should be introduced.
1024:
1025: {\bf On the spectral energy distribution and orbital variability}
1026:
1027: In Fig.~\ref{sedls}, for illustrative purposes, we show the synchrotron and IC SED for LS~5039 averaged over the orbital
1028: phase interval $0.9>\phi>0.45$, corresponding to the inferior conjunction of the compact object ($\phi=0.72$). We use this
1029: format for a better comparison with the spectral results from observations (Aharonian et al. \cite{aharonian06}), shown also
1030: in the figure. For the very high energy spectrum in superior conjunction, see Khangulyan et al. \cite{khangulyan07}, fig.~23;
1031: at that phase interval, the synchrotron emission will be similar to that at inferior conjunction. In Fig.~\ref{lcls}, the VHE
1032: lightcurve along the orbit is shown. The adopted parameter values in both figures are the following: $V_{\rm adv}=0.1\,c$,
1033: $Z_0=10^{12}$~cm, $i=25^{\circ}$, $\eta=10$ and $B=0.05$~G. The exact values of $B$ and $V_{\rm adv}$ are actually no
1034: crucial. In the case of $B$, its value just needs to fulfill acceleration constraints (i.e. to allow electrons to reach high
1035: enough energies to explain the TeV data), and be low enough such that VHE electrons will lose most of their energy via IC in
1036: the KN regime. Regarding $V_{\rm adv}$, moderately different values will render similar results. To reach the observed
1037: emission levels, the total injected luminosity in relativistic electrons is $L_{\rm e}\approx 1.5\times
1038: 10^{35}$~erg~s$^{-1}$. We note that we have computed the contribution from both the jet and the counter-jet and summed them
1039: up. Photon-photon absorption is taken into account. Remarkably, the synchrotron part of the spectrum is well below the
1040: observed fluxes, indicating that the dominant X-rays in this source likely come from a different emitting region. This region
1041: should be nevertheless physically connected to the TeV emitter, given the similar behavior in time of both the X-ray and the
1042: TeV radiation.
1043:
1044: Regarding variability, we show in Fig.~\ref{lcls} the orbital VHE lightcurve obtained for the same parameter choice adopted
1045: to compute the SED. Although the matching with observations is not perfect, it is necessary to state that the lightcurve
1046: provided in Aharonian et al. \cite{aharonian06} is constructed in bins with relatively small statistical significance, the
1047: spectra are simplified to pure power laws (and they could be more complicated than just a power law, as shown in
1048: Fig.~\ref{sedls}), and includes data points from observations of very different epochs. In any case, just playing with
1049: propagation effects, IC scattering and photon photon absorption, one can already obtain very different lightcurves.
1050:
1051: For an extensive discussion on the role of the different parameters present in the used model, we refer to Khangulyan et al.
1052: \cite{khangulyan07}. Here our purpose is just to note that the leptonic scenario, accounting for changes in the interaction
1053: geometry due to the orbital motion of the system, can reproduce quite nicely the observed spectra for different phases. It is
1054: worthy mentioning that, to reach the observed maximum photon energies, $Z_0$ is to be $\ge 10^{12}$~cm unless the
1055: acceleration efficiency is close to the maximum one, i.e. $\eta<10$\footnote{In the case of a deep emitter, the magnetic
1056: field in the system should be well below several G, as noted by Bosch-Ramon et al. \cite{bosch08b}.}. In addition, the
1057: magnetic field in the accelerator/emitter is restricted to a relatively narrow band around 0.1~G.
1058:
1059: \begin{figure}[pb]
1060: \centerline{\psfig{file=bb_av2.eps,width=9cm, angle=270}}
1061: \vspace*{8pt}
1062: \caption{
1063: Computed synchrotron and IC SED of LS~5039 averaged over the orbital phase interval $0.9>\phi>0.45$. The HESS data points are
1064: also shown (Aharonian et al. 2006). The adopted parameter values are $V_{\rm adv}=0.1\,c$, $Z_0=10^{12}$~cm, $i=25^{\circ}$,
1065: $\eta=10$, $B=0.05$~G, and $L_{\rm e}=1.5\times 10^{35}$~erg~s$^{-1}$. The total contribution from the jet and the counter-jet is
1066: shown.
1067: \label{sedls}}
1068: \end{figure}
1069:
1070: \begin{figure}[pb]
1071: \centerline{\psfig{file=print2.eps,width=9cm, angle=270}}
1072: \vspace*{8pt}
1073: \caption{Computed IC orbital lightcurve of the differential photon flux at 1~TeV for LS~5039.
1074: The different components, jet, counter-jet and summation of both, are labelled. The adopted
1075: parameters are the same as those adopted in Fig.~\ref{sedls}. \label{lcls}}
1076: \end{figure}
1077:
1078: \subsubsection{LS~I~+61~303}\label{lsi}
1079:
1080: LS~I~+61~303 is a quite eccentric ($e=0.72$) HMXB, located at $\approx 2$~kpc, with an orbital semi-major axis of $\approx
1081: 6\times 10^{12}$~cm, a Be primary star of luminosity $\approx 10^{38}$~erg~s$^{-1}$ and temperature $\approx 28000$~K, and an
1082: orbital period of 26.4960~days (Hutchings \& Crampton \cite{hutchings81}; Gregory \cite{gregory02}; Casares et~al.
1083: \cite{casares05b}). The inclination of the system is not well constraint, being in the range $i=15-60^{\circ}$. Massi et al.
1084: (\cite{massi01,massi04}) detected radio jets, with apparently relativistic motion, at $\sim 100$~milliarcsecond scales and
1085: classified LS~I~+61~303 as a microquasar. Otherwise, the source does not show signatures of accretion in the X-rays (e.g.
1086: Sidoli et al. \cite{sidoli06}; Paredes et al. \cite{paredes07}; and references therein), which are of likely non-thermal
1087: origin. This fact, altogether with other arguments -mainly related to the extended radio emission morphology (see Dhawan et
1088: al. \cite{dhawan06})- have let several authors to put forward a non-accreting pulsar as the compact object in the system
1089: (e.g. Dubus \cite{dubus06b}; Chernyakova et al. \cite{chernyakova06}; Dhawan et al. \cite{dhawan06}), which had been proposed
1090: for the first time by Maraschi \& Treves \cite{maraschi81}. Nevertheless, the microquasar scenario cannot be discarded, and
1091: it may indeed explain some of the inferred properties of LS~I~+61~303 better than the pulsar scenario (e.g. Romero et al.
1092: \cite{romero07}). In fact, the apparently slow and collimated radio structures detected in this source may be rather
1093: difficult to explain in the colliding wind context, in which very fast motions of the shocked pulsar wind could be expected
1094: (see Bogovalov et al. \cite{bogovalov08}). We recall that LS~I~+61~303 has been detected in the TeV range by MAGIC (Albert et
1095: al. \cite{albert06}) with the maximum around the phase 0.6, being not detected during periastron passage, at phase 0.23
1096: (Casares et al. \cite{casares05b}). This source had been also detected by EGRET (Kniffen et al. \cite{kniffen97}). It is
1097: worthy noting that the lightcurves in the radio, X-rays, and high-energy and very high-energy gamma-rays look somewhat
1098: similar (see fig.~3 in Chernyakova et al. \cite{chernyakova06} and references therein). Several authors have adopted
1099: different frameworks and mechanisms to explain the high energy radiation from this source (in the microquasar framework: e.g.
1100: Romero et~al. \cite{romero05b} -hadronic-, Bosch-Ramon et al. \cite{bosch06c} -leptonic-, Bednarek \cite{bednarek06}
1101: -cascading-; in the pulsar framework: e.g. Leahy \cite{leahy04}, Dubus \cite{dubus06b} -leptonic-, Chernyakova et al.
1102: \cite{chernyakova06} -hadronic-; regarding neutrino detectability, see Christiansen et al. \cite{christiansen06} and Torres
1103: \& Halzen \cite{torres07}).
1104:
1105: {\bf On the spectral energy distribution and orbital variability}
1106:
1107: In Fig.~\ref{sedlsi}, we show the computed synchrotron and IC SED for LS~I~+61~303 at $\phi=0.6$. The adopted parameter
1108: values are the following: $V_{\rm adv}=0.1\,c$, $Z_0=10^{12}$~cm, $i=30^{\circ}$, and $\eta=10$. In this case, the magnetic
1109: field energy density could be larger than the radiation one compared to LS~5039, since the VHE spectrum is less hard, leaving
1110: more room for synchrotron losses (which yield a softer spectrum). The value of $B$ obtained here to reproduce the observed
1111: radiation is 0.5~G. The total adopted $L_{\rm e}$ is $1.5\times 10^{35}$~erg~s$^{-1}$. In LS~I~+61~303, the X-ray levels
1112: would not be far from the observed ones. Still, for the adopted magnetic field, adjusted to explain the TeV spectrum, the
1113: fluxes are still too low. Unlike LS~5039, neither the observer/binary system geometry nor the impact of photon-photon
1114: absorption are enough when understanding the TeV variability, given the distance to, and the luminosity of, the primary star.
1115: At the phase at which this emission is detected, the gamma-ray attenuation is low, and the IC scattering angle changes
1116: relatively little. In Figs.~\ref{lclsi1}, \ref{lclsi2}, and \ref{lclsi3}, the differential photon flux lightcurves at 0.1, 1
1117: and 10~TeV, respectively, are shown. In the observed lightcurve (Albert et al. \cite{albert06}), the maximum of the emission
1118: is at phases (using the ephemeris of Casares et al. \cite{casares05b}; see also Grundstrom et al. \cite{grundstrom07})
1119: different from those predicted by our IC/photon-photon absorption model. This fact, plus the similarity of the VHE lightcurve
1120: with those at other wavelengths,v suggest intrinsic changes of the emitting region. Interestingly, the 10~TeV emission in our
1121: scenario peaks roughly around periastron passage, due to the characteristics of the pair creation and IC cross sections.
1122:
1123: \begin{figure}[pb]
1124: \centerline{\psfig{file=lsi_bb2.eps,width=9cm, angle=0}}
1125: \vspace*{8pt}
1126: \caption{The Computed Synchrotron and IC SED for LS~I~+61~303 at phase $\phi=0.6$,
1127: together with the observed VHE SED, is shown. The adopted parameter values are
1128: $V_{\rm adv}=0.1\,c$, $Z_0=10^{12}$~cm, $i=30^{\circ}$, $\eta=10$, $B=0.5$~G
1129: and $L_{\rm e}=1.5\times 10^{35}$~erg~s$^{-1}$. The total contribution from the jet and the counter-jet is
1130: shown.
1131: \label{sedlsi}}
1132: \end{figure}
1133:
1134: \begin{figure}[pb]
1135: \centerline{\psfig{file=light_curves1.eps,width=9cm, angle=270}}
1136: \vspace*{8pt}
1137: \caption{
1138: The light-curve (differential photon flux) at 100~GeV for LS~I~+61~303. The jet and the counter-jet
1139: components are labelled. The adopted parameter values are the
1140: same as those in Fig.~\ref{sedlsi}.
1141: \label{lclsi1}}
1142: \end{figure}
1143:
1144: \begin{figure}[pb]
1145: \centerline{\psfig{file=light_curves2.eps,width=9cm, angle=270}}
1146: \vspace*{8pt}
1147: \caption{
1148: The same as in Fig.~\ref{lclsi1} but at 1~TeV.
1149: \label{lclsi2}}
1150: \end{figure}
1151:
1152: \begin{figure}[pb]
1153: \centerline{\psfig{file=light_curves3.eps,width=9cm, angle=270}}
1154: \vspace*{8pt}
1155: \caption{
1156: The same as in Fig.~\ref{lclsi1} but at 10~TeV.
1157: \label{lclsi3}}
1158: \end{figure}
1159:
1160: \section{Further comments}\label{disc}
1161:
1162: We have discussed in this work the different processes that may take place in microquasars, and explored the possibility that
1163: stellar IC scattering is behind the VHE radiation from three sources recently detected in the TeV range. We have shown also
1164: the importance of photon-photon absorption, whose occurrence can be used to constrain the emitter properties. In this
1165: section, we critically discuss three hot topics that are, in our view, very relevant for the kind of source discussed here.
1166: The first point is related to the possibility, already mentioned and commented above, that a non-accreting pulsar may power
1167: the non-thermal broadband emission in LS~I~+61~303 and LS~5039. The second point, affecting Cygnus~X-1 as well, is the role
1168: played by the companion star concerning high-energy processes and the dynamics of the jet. Finally, some comments are made
1169: regarding the prospects of the study of microquasars with the new/future VHE intrumentation.
1170:
1171: \subsection{Evidences for a pulsar nature}
1172:
1173: Although we do not consider here that LS~I~+61~303 and LS~5039 may harbor young non-accreting pulsars instead of an accreting
1174: black-hole, this possibility cannot be discarded. At present, the strongest argument in favor of this is the lack of
1175: accretion features in the X-ray spectrum. An {\it extension} of this argument is the fact that the timing properties of the
1176: X-ray emission in these two sources, and even the radio/X-ray behavior, do not correspond to what is believed a microquasar
1177: should be. Needless to say, this argument is phenomenological, and relies on a supposed regularity of the microquasar
1178: behavior at X-rays. The same could be said regarding the radio behavior and the radio/X-ray connection.
1179:
1180: Indeed, there is a whole set of phenomenological studies, as already mentioned in Sect.~\ref{genst}, majorly grounded on
1181: observations but also to some extent on theoretical models, which try to give an unified microquasar picture. LS~5039 and
1182: LS~I~+61~303 do not fit in such a picture regarding the points mentioned above. Nevertheless, most of the high-mass
1183: microquasars, and to a lesser extent several low-mass microquasars, present some level of discrepancy from the standard
1184: picture, and not all the sources may have occurring in them the same mechanisms as those comprehended in the standard
1185: scenario. All this shows that such a picture or framework is a useful working hypothesis, but it cannot still be a
1186: discrimination tool when trying to find out whether or not a source pertains to some class of objects. In the same line,
1187: sometimes the morphology of the radio emission from these systems is used as an argument, again phenomenological, against
1188: their microquasar nature. In this regard, it is claimed that the extended radio emission does not fit the standard picture of
1189: microquasar jets. Nevertheless, it is worthy pointing out here that galactic compact jets are sometimes detected at the
1190: resolution limits of radio VLBI interferometers. It makes therefore difficult to argue, based on solid observational grounds,
1191: about how a microquasar jet should look like. From the theoretical point of view, to define how a canonical microquasar jet
1192: should be is presently not possible, since we lack a complete theory for jet formation and collimation, and there might be
1193: more than one mechanism operating.
1194:
1195: There are otherwise two good indicators of the presence of a pulsar in the system, i.e. the detection of radio and X-ray
1196: pulsations, and the lack of strong X-ray radiation (as is the case) due to any form of accretion onto a possible neutron star
1197: surface (in case the compact object were known to be a neutron star). To date, none of both questions has been answered.
1198: Neither pulsations have been detected, nor the problem of the compact object nature has been solved. The lack of pulsations
1199: may be explained by dense stellar wind ionized material smearing out the pulsed signal via free-free absorption, although the
1200: powerful pulsar wind would allow the observer to see the pulsar without the interference of the stellar wind around inferior
1201: conjunction. The radiation beam may also point away from us preventing us from seeing it. Regarding accretion X-ray bursts,
1202: since the masses of the components in LS~5039 and LS~I~+61~303 are not yet properly constraint, we cannot tell whether the
1203: compact object is a black hole or a neutron star.
1204:
1205: Therefore, the question whether there is an accreting or a non-accreting compact object in LS~5039 and LS~I~+61~303 remains
1206: open. Fortunately, because of this fact, their complex multiwavelength spectral and temporal behavior, and their TeV
1207: detections, these two sources are extensively studied nowadays. This will bring for sure fruitful and surprising results in
1208: the near future.
1209:
1210: \subsection{The relevance of the primary star}
1211:
1212: LS~5039, LS~I~+61~303, and Cygnus~X-1 contain a massive and hot OB type star which embeds the jet/accelerator/emitter with
1213: dense matter and photon fields. Unlike low-mass microquasars, where the accretion/ejection phenomena could be naturally
1214: reduced to the disk/jet system, in high-mass microquasars (SS~433 and Cygnus~X-3 are two additional instances of high-mass
1215: microquasars, both being certainly peculiar), the strong photon field should play an important role in the radiation
1216: processes via, e.g., stellar IC, photo-meson production and photo-disintegration, photon-photon absorption, secondary pair
1217: radiation and electromagnetic cascading. In addition, the strong stellar wind is likely to have an impact in the radiation
1218: processes via dynamical interactions with the jet/accelerator/emitter, providing targets for pp radiation, confining
1219: relativistic particles, either secondary pairs or protons, absorbing part of the radio and X-ray photons produced in the
1220: system, and determining the medium at large scales (and thereby influencing the processes that take place in the termination
1221: regions of the microquasar jets). As shown along this review, numerous studies have been or are being carried out to clarify
1222: the importance of the primary star in the mentioned processes.
1223:
1224: Beside the fact that the physics to extract from theoretical studies plus present and future observations can teach us a lot
1225: about jets, acceleration and radiation processes, and stellar winds, it is also important to remark that the presence of the
1226: star cannot be neglected when modeling phenomenologically the observations of high-mass microquasars. It seems very likely
1227: that these systems cannot be understood as just accretion/ejection systems to first, or even zero, order of approximation,
1228: but require a much more complex approach including hydrodynamical and magnetohydrodynamical simulations of jets and their
1229: environment, a consistent treatment of acceleration and emission of particles, and the physics of OB stars and their winds.
1230:
1231: \subsection{Prospects}
1232:
1233: The future instruments like MAGIC~II, HESS~II, CTA, and the already on-flight GLAST will allow to make a big step in our
1234: understanding of microquasars and gamma-ray emitting binaries. The better performance at low energies of all these
1235: instruments will allow to study for instance the possible development of electromagnetic cascades (in case of weak magnetic
1236: fields) or the secondary single-scattering IC component, giving information on the physical conditions of the VHE emitter
1237: environment. Furthermore, radiation components below 100~GeV
1238: coming from regions invisible in TeV (e.g. due to severe absorption or maximum particle energy <~100~GeV),
1239: could also be investigated. The good sensitivity in the whole
1240: energy range would allow the study of fast variability and accurate modeling, which is of primary importance to understand the
1241: structure of the source, and the nature of the particles and processes behind the gamma-ray emission. Finally, a extension of
1242: the operation energy range up to $\sim 100$~TeV (e.g. CTA) would bring the opportunity to study in detail the acceleration
1243: processes with extreme efficiencies, like in the case of LS~5039.
1244:
1245: \section{Summary}\label{summ}
1246:
1247: Microquasars are sources in which non-thermal processes occur. Particle acceleration is taking place in these systems,
1248: although the mechanisms involved are still unclear (e.g. non-relativistic and relativistic shocks, magnetic turbulence,
1249: magnetic reconnection, etc.). The complexity of such a multiwavelength emitters, in which different radiative processes may
1250: be relevant in the same energy range, and emission reprocessing via photon-photon absorption is common, yields the study in
1251: detail of the underlying physics quite difficult. Nevertheless, the detection of TeV photons can help to constrain the
1252: fundamental properties of the emitting region, since the relevant timescales are short. Because of this, Cherenkov astronomy
1253: is allowing for the first time to really probe the accelerator/emitter in microquasars. Cygnus~X-1, observed by MAGIC, has
1254: been found flaring during phases when photon-photon absorption is expected to be very severe, giving information on the
1255: emitter location and the stellar magnetic field. LS~5039, detected by the Cherenkov telescope HESS as a periodical TeV
1256: emitter, is at the moment the best laboratory to understand particle acceleration and radiative processes in galactic compact
1257: sources. LS~I~61~+303, detected as well by MAGIC, shows TeV emission that varies along the orbit. Since geometric effects
1258: would not play in this source a role for the orbital variability as important as in the case of LS~5039, intrinsic properties
1259: of the emitter should also change. The same would apply to Cygnus~X-1.
1260:
1261: We conclude that a leptonic scenario can explain the radiation from microquasars at very high energies, although hadronic
1262: emission, energetically very demanding, cannot be discarded. In case of LS~5039 and Cygnus~X-1, acceleration of particles and
1263: emission seem to take place in the borders of the system. Due to their relatively hard VHE spectra, the emitter magnetic
1264: field in LS~5039 and LS~I~+61~303 should be low. Due to the magnetic field produced by an OB star, gamma-rays, although
1265: likely photon-photon absorbed, can hardly trigger efficient electromagnetic cascades and secondary energy may be radiated
1266: mainly via the synchrotron process.
1267:
1268: TeV microquasars are showing us that they behave in a different way from their extragalactic analogs, the blazars, being very
1269: much affected by the presence of the primary star and the the orbital motion.
1270:
1271: \section*{Acknowledgments}
1272: The authors thank the anonymous referee for constructive comments and useful
1273: suggestions. The authors are grateful to Felix A. Aharonian for fruitful
1274: discussion and advice. The authors thank Anabella T. Araudo for helping to
1275: improve the manuscript. V.B-R. gratefully acknowledges support from the
1276: Alexander von Humboldt Foundation. V.B-R. acknowledges support by DGI of MEC
1277: under grant AYA2007-68034-C03-01, as well as partial support by the European
1278: Regional Development Fund (ERDF/FEDER).
1279:
1280: \section{References}
1281: \begin{thebibliography}{0} %for 1 digit
1282: \bibitem{mirabel99} Mirabel, I.~F. \& Rodr\'iguez, L.~F. {\it ARA\&A} {\bf 37}, 409 (1999)
1283: \bibitem{mcclintock06} McClintock, J.~E., Remillard, R.~A. Black Hole Binaries
1284: in {\it Compact stellar X-ray sources}, eds. W. Lewin \& M. van der Klis
1285: (Cambridge University Press) (2006)
1286: \bibitem{fender04a} Fender, R.~P., Belloni, T.~M., Gallo, E. {\it MNRAS}, {\bf 355}, 1105 (2004)
1287: \bibitem{gallo03} Gallo E., Fender R.~P., Pooley G.~G. {\it MNRAS}, {\bf 344}, 60 (2003)
1288: \bibitem{corbel03}
1289: Corbel, S., Nowak, M.~A., Fender, R.~P., Tzioumis, A.~K., Markoff, S. {\it A\&A}, {\bf 400}, 1007
1290: (2003)
1291: \bibitem{fender03} Fender, R.~P., Gallo, E., Jonker, P.~G. {\it MNRAS} {\bf 343}, 99 (2003)
1292: \bibitem{ribo05} Rib\'o, M. in {\it Future Directions in High Resolution Astronomy: The 10th Anniversary of the VLBA},
1293: (ASPC, 2005) 340, 421 [astro-ph/0402134]
1294: \bibitem{corbel02} Corbel, S., Fender, R. P., Tzioumis, A. K., {\it et al.}, {\it Science} {\bf 298}, 196 (2002)
1295: \bibitem{bosch07} Bosch-Ramon, V., {\it Ap\&SS} {\bf 309}, 321 (2007)
1296: \bibitem{fender04} Fender, R. {\it NewAR} {\bf 48}, 1399 (2004)
1297: \bibitem{chaty03} Chaty, S., Haswell, C. A., Malzac, J. {\it MNRAS} {\bf 346}, 689 (2003)
1298: \bibitem{levinson96} Levinson, A. \& Blandford, R.~D. {\it ApJ} {\bf 456}, L29 (1996)
1299: \bibitem{levinson96b}Levinson, A. \& Mattox, J.~R. {\it ApJ}, {\bf 462}, 67 (1996)
1300: \bibitem{paredes00} Paredes, J.M., Mart\'{\i}, J., Rib\'o, M., Massi, M., {\it Science} {\bf 288}, 2340 (2000)
1301: \bibitem{albert07} Albert, J. et al. {\it ApJ} {\bf 665}, L51 (2007)
1302: \bibitem{aharonian05} Aharonian, et al. {\it Science} {\bf 309}, 746 (2005)
1303: \bibitem{aharonian06} Aharonian, F.~A. et al. {\it A\&A} {\bf 460}, 743 (2006)
1304: \bibitem{casares05} Casares, J., Rib\'o, M., Ribas, I., et al., {\it MNRAS} {\bf 364}, 899 (2005)
1305: \bibitem{albert06} Albert, J. et al. {\it Science} {\bf 312}, 1771 (2006)
1306: \bibitem{mirabel92} Mirabel I. F., Rodr\'iguez L. F., Cordier B., Paul J., Lebrun F., {\it Nature} {\bf 358}, 215 (1992)
1307: \bibitem{mirabel94} Mirabel, I. F. \& Rodriguez, L. F. {\it Nature} {\bf 371}, 46 (1994)
1308: \bibitem{merloni03} Merloni, A., Heinz, S., di Matteo, T. {\it MNRAS} {\bf 345}, 1057 (2003)
1309: \bibitem{falcke04} Falcke, H., Körding, E., Markoff, S., {\it A\&A} {\bf 414}, 895 (2004)
1310: \bibitem{koerding06} K\"ording, E., Falcke, H., Corbel, S. {\it A\&A} {\bf 456}, 439 (2006)
1311: \bibitem{markoff05} Markoff, S., Nowak, M.~A., Wilms, J., {\it ApJ} {\bf 635}, 1203 (2005)
1312: \bibitem{maccarone05} Maccarone, T.~J., {\it MNRAS} {\bf 360}, L68 (2005)
1313: \bibitem{bogovalov05} Bogovalov, S.~V. \& Kelner, S.~R. {\it Astron. Rep.} {\bf 49}, 57 (2005)
1314: \bibitem{levinson06} Levinson, A. {\it Int. J. Mod. Phys.~A} {\bf 21}, 30 (2006)
1315: \bibitem{araudo07} Araudo, A.~T., Romero, G.~E., Bosch-Ramon, V., Paredes, J.~M. {\it A\&A} {\bf 476}, 1289 (2007)
1316: \bibitem{blandford76} Blandford, R. D. {\it MNRAS} {\bf 176} 465 (1976)
1317: \bibitem{blandford77} Blandford, R. D. \& Znajek, R. L. {\it MNRAS} {\bf 179}, 433 (1977)
1318: \bibitem{blandford82} Blandford, R. D. \& Payne, D. G. {\it MNRAS} {\bf 199}, 883 (1982)
1319: \bibitem{meier96} Meier, D. {\it ApJ}, {\bf 459}, 185 (1996)
1320: \bibitem{koide02} Koide, S., Shibata, K., Kudoh, T., Meier, D.~L. {\it Science} {\bf 295}, 1688 (2002)
1321: \bibitem{chatto02} Chattopadhyay, I. \& Chakrabarti, S.~K. {\it MNRAS} {\bf 333}, 454 (2002)
1322: \bibitem{meier03} Meier, D. {\it NewA Rev.}, {\bf 47}, 667 (2003)
1323: \bibitem{hujeirat04} Hujeirat, A. {\it A\&A} {\bf 416}, 423--435 (2004)
1324: \bibitem{meier05} Meier, D.~L. {\it Ap\&SS} {\bf 300}, 55 (2005)
1325: \bibitem{deVilliers05} De Villiers, J.-P., Hawley, J.~F., Krolik, J.~H., Hirose, S. {\it ApJ} {\bf 620}, 878 (2005)
1326: \bibitem{ferreira06} Ferreira, J., Petrucci, P.~O., Henri, G., Saugé, L., Pelletier, G. {\it A\&A} {\bf 447}, 813 (2006)
1327: \bibitem{mcKinney06} McKinney, J.~C. {\it MNRAS}, {\bf 368}, 1561 (2006)
1328: \bibitem{hawley06} Hawley, J.~F. \& Krolik, J.~H. {\it ApJ} {\bf 641}, 103 (2006)
1329: \bibitem{komissarov07} Komissarov, S.~S., Barkov, M.~V., Vlahakis, N., K\"onigl, A. {\it MNRAS} {\bf 380}, 51 (2007)
1330: \bibitem{barkov08} Barkov, M.~V. \& Komissarov, Serguei S. {\it MNRAS} {\bf 385}, 28 (2008)
1331: \bibitem{junor99} Junor, W., Biretta, J.~A., Livio, M. {\it Nature} {\bf 401}, 891 (1999)
1332: \bibitem{horiuchi06} Horiuchi, S., Meier, D.~L., Preston, R.~A., Tingay, S.~J. {\it PASJ} {\bf 58}, 221 (2006)
1333: \bibitem{namiki03} Namiki, M., Kawai, N., Kotani, T., Makishima, K., {\it PASJ} {\bf 55}, 281 (2003)
1334: \bibitem{hardee03} Hardee, P.~E., Hughes, P.~A. {\it ApJ} {\bf 583}, 116 (2003)
1335: \bibitem{tsinganos04} Tsinganos, K., Vlahakis, N., Bogovalov, S.~V., Sauty, C., Trussoni, E. {\it Ap\&SS} {\bf 293}, 55 (2004)
1336: \bibitem{perucho07} Perucho, M. \& Bosch-Ramon, V. {\it A\&A} {\bf 482}, 917 (2008)
1337: \bibitem{faranoff74} Fanaroff, B.~L. \& Riley, J.~M. {\it MNRAS} {\bf 167}, 31 (1974)
1338: \bibitem{pm07} Perucho, M. \& Mart\'{\i}, J.~M. {\it MNRAS} {\bf 382}, 526 (2007)
1339: \bibitem{kaiser97} Kaiser, C.~R. \& Alexander, P. {\it MNRAS} {\bf 286}, 215 (1997)
1340: \bibitem{sch02} Scheck, L., Aloy, M.A., Mart\'{\i}, J.M$^{\underline{\rm a}}$, G\'omez, J.L., M\"uller, E. {\it MNRAS} {\bf 331}, 615 (2002)
1341: \bibitem{zealey80} Zealey, W.~J., Dopita, M.~A., Malin, D.~F. {\it MNRAS} {\bf 192}, 731 (1980)
1342: \bibitem{velazquez00} Vel\'azquez, P. F. \& Raga, A. C. {\it A\&A}, {\bf 362}, 780 (2000)
1343: \bibitem{sikora05} Sikora, M., Begelman, M.~C., M., Greg M., Lasota, J.~P., {\it ApJ} {\bf 625}, 72 (2005)
1344: \bibitem{zenitani01} Zenitani, S., Hoshino, M., {\it ApJ} {\bf 562}, 63 (2001)
1345: \bibitem{derishev03} Derishev, E.~V., Aharonian, F.~A., Kocharovsky, V.~V., Kocharovsky, Vl. V.,
1346: {\it PhRvD} {\bf 68}, 3003 (2003)
1347: \bibitem{stern06} Stern, B., Poutanen, J. {\it MNRAS} {\bf 372}, 1217 (2006)
1348: \bibitem{gierlinski03} Gierli\'nski, M., Done, C., {\it MNRAS} {\bf 342}, 1083 (2003)
1349: \bibitem{neronov07} Neronov, A. \& Aharonian, F.~A. {\it ApJ} {\bf 671}, 85 (2007)
1350: \bibitem{rieger07} Rieger, F. \& Aharonian, F.~A. {\it A\&A} {\bf 479}, L5 (2008)
1351: \bibitem{drury83} Drury, L.~O., {\it Rep. Prog. Phys.} {\bf 46}, 973 (1983)
1352: \bibitem{fermi49} Fermi, E., {\it Phys. Rev.} {\bf 75}, 1169 (1949)
1353: \bibitem{rieger04} Rieger, F.~M.; Duffy, P., {\it ApJ} {\bf 617}, 155 (2004)
1354: \bibitem{rieger06} Rieger, F.~M., Bosch-Ramon, V., Duffy, P. {\it Ap\&SS} {\bf 309}, 119 (2007)
1355: \bibitem{heinz02} Heinz, S., Sunyaev, R., {\it A\&A} {\bf 390}, 751 (2002)
1356: \bibitem{markoff01} Markoff S., Falcke H., Fender R., {\it A\&A} {\bf 372}, 25 (2001)
1357: \bibitem{bosch06b} Bosch-Ramon, V., Romero, G~E., Paredes, J.~M. {\it A\&A} {\bf 447}, 263 (2006)
1358: \bibitem{bosch04} Bosch-Ramon, V., Paredes, J. M., {\it A\&A} {\bf 417}, 1075 (2004)
1359: \bibitem{romero02} Romero, G.~E., Kaufman Bernad\'o, M.~M., Mirabel, I.~F., {\it A\&A} {\bf 393}, 61 (2002)
1360: \bibitem{georganopoulos02} Georganopoulos, M., Aharonian, F.~A., Kirk, J.~G., {\it A\&A} {\bf 388}, 25 (2002)
1361: \bibitem{levinson01} Levinson, A., Waxman, E. {\it PhRvL} {\bf 87}, 171101 (2001)
1362: \bibitem{aharonian06b} Aharonian, F.~A., Anchordoqui, L.~A., Khangulyan, D., Montaruli, T. {\it J.Phys.Conf.Ser.} {\bf 39} 408 (2006)
1363: \bibitem{romero08} Romero, G.~E. \& Vila G.~S. {\it A\&A} {\bf 485}, 623 (2008)
1364: \bibitem{yuan05} Yuan, F.; Cui, W., Narayan, R., {\it ApJ} {\bf 620}, 905 (2005)
1365: \bibitem{paredes06} Paredes, J.~M., Bosch-Ramon, V., Romero, G.~E., {\it A\&A} {\bf 451}, 259 (2006)
1366: \bibitem{atoyan99} Atoyan, A.~M., Aharonian, F.~A., {\it MNRAS} {\bf 302}, 253 (1999)
1367: \bibitem{dermer06} Dermer, C., B\"ottcher, M., {\it ApJ} {\bf 643}, 1081 (2006)
1368: \bibitem{paredes02} Paredes, J. M., Rib\'o, M., Mart\'i, J., {\it A\&A} {\bf 393}, 99 (2002)
1369: \bibitem{kaufman02} Kaufman Bernad\'o M.M., Romero G.E., Mirabel I.F., {\it A\&A} {\bf 385}, 10 (2002)
1370: \bibitem{khangulyan07} Khangulyan, D., Aharonian, F., Bosch-Ramon, V., {\it MNRAS}, {\bf 383}, 467 (2008)
1371: \bibitem{romero03} Romero, G. E., Torres, D. F., Kaufman Bernad\'o, M. M., Mirabel, I. F., {\it A\&A} {\bf 410}, 1 (2003)
1372: \bibitem{romero05} Romero, G.~E., Orellana, M., {\it A\&A} {\bf 439}, 237 (2005)
1373: \bibitem{aharonian06c} Aharonian, F. Invited talk at {\it The International Conference on Neutrino Physics and Astrophysics} (2006)
1374: [astro-ph/0702680]
1375: \bibitem{bednarek05} Bednarek, W. {\it ApJ} {\bf 631}, 466 (2005)
1376: \bibitem{vanderlaan66} van der Laan, H., {\it Nature} {\bf 211}, 1131 (1966)
1377: \bibitem{aharonian98} Aharonian, F. A.; Atoyan, A. M. {\it NewAR} {\bf 42}, 579 (1998)
1378: \bibitem{bordas08} Bordas, P., Bosch-Ramon, V., Paredes, J. M. {\it Int.~J.~Mod Phys.~D} {\bf 17}, 1895 (2008)
1379: \bibitem{bosch05} Bosch-Ramon, V., Aharonian, F. A., Paredes, J. M., {\it A\&A} {\bf 432}, 609 (2005c)
1380: \bibitem{akharonian85} Akharonian, F. A., Vardanian, V. V., {\it Ap\&SS} {\bf 115}, 31 (1985)
1381: \bibitem{donati02} Donati, J.~F., Babel, J., Harries, T.~J. et al. {\it MNRAS} {\bf 33}, 55 (2002)
1382: \bibitem{protheroe87} Protheroe, R.~J. \& Stanev, T. , {\it ApJ} {\bf 322}, 838 (1987)
1383: \bibitem{moskalenko94} Moskalenko I.V., Karakula S., {\it ApJS} {\bf 92}, 567 (1994)
1384: \bibitem{bednarek97} Bednarek, W. {\it A\&A}, {\bf 322}, 523 (1997)
1385: \bibitem{boettcher05} Böttcher, M., Dermer, C.~D. {\it ApJ}, {\bf 634}, L81 (2005)
1386: \bibitem{bottcher05} B\"ottcher, M., Dermer, C., {\it ApJ} {\bf 634}, 81 (2005)
1387: \bibitem{dubus06} Dubus, G., {\it A\&A} {\bf 451}, 9 (2006)
1388: \bibitem{reynoso08}
1389: Reynoso, M.~M., Christiansen, H.~R., Romero, G. E. {\it Astrop. Phys.} {\bf 28}, 565 (2008)
1390: \bibitem{bednarek06} Bednarek, W., {\it MNRAS} {\bf 368}, 579 (2006)
1391: \bibitem{orellana07} Orellana, M., Bordas, P., Bosch-Ramon, V., Romero, G.~E., Paredes, J.~M. {\it A\&A} {\bf 476}, 9 (2007)
1392: \bibitem{bosch08} Bosch-Ramon, V., Khangulyan, D., Aharonian, F.~A. {\it A\&A} {\bf 482}, 397 (2008)
1393: \bibitem{hillas84} Hillas, A.~M. {\it ARA\&A} {\bf 22} 425 (1984)
1394: \bibitem{protheroe99} Protheroe, R.~J. Acceleration and Interaction of Ultra High Energy Cosmic Rays in
1395: {\it Topics in cosmic-ray astrophysics}, eds. M. A. DuVernois (Nova Science Publishing) (1999) [astro-ph/9812055]
1396: \bibitem{kelner06} Kelner, S.~R., Aharonian, F.~A., Bugayov, V.~V. {\it Phys. Rev. D} {\bf 74}, 4018 (2006)
1397: \bibitem{kelner08} Kelner, S. R. \& Aharonian, F. A. {\it Phys. Rev. D} {\bf 78}, 4013 (2008)
1398: \bibitem{aharonian98b} Aharonian F.~A. \& Heinzelmann, G. {\it Nucl. Phys. Proc. Suppl.} {\bf 60}, 193 (1998)
1399: \bibitem{rovero02} Rovero A.~C. et al., {\it BAAA} {\bf 45}, 66 (2002)
1400: \bibitem{khangulyan05} Khangulyan, D. \& Aharonian, F.~A. in {\it High Energy Gamma-Ray Astronomy},
1401: (AIP Conference Proceedings 2005), 745, 359 [astro-ph/0503499]
1402: \bibitem{stirling01} Stirling, A. M., Spencer, R. E., de la Force, C. J. {\it MNRAS}, {\bf 327}, 1273 (2001)
1403: \bibitem{ziolkowski05} Ziolkowski, J. {\it MNRAS}, {\bf 358}, 851 (2005)
1404: \bibitem{gies86} Gies, D.~R., \& Bolton, C. T. {\it ApJ}, {\bf 304}, 371 (1986)
1405: \bibitem{bednarek07} Bednarek, W. \& Giovannelli, F. {\it A\&A} {\bf 464}, 437 (2007)
1406: \bibitem{sunyaev79} Sunyaev, R. A., \& Truemper, J. {\it Nature} {\bf 279}, 506 (1979)
1407: \bibitem{mcconnell02} McConnell, M. L., Zdziarski, A. A., Bennett, K. {\it ApJ} {\bf 572}, 984 (2002)
1408: \bibitem{bosch08b} Bosch-Ramon, V., Khangulyan, D., Aharonian, F. A. {\it A\&A} {\bf 489}, L21 (2008)
1409: \bibitem{motch97} Motch, C., Haberl, F., Dennerl, K., Pakull, M., Janot-Pacheco, E., {\it A\&A} {\bf 323}, 853 (1997)
1410: \bibitem{bosch05} Bosch-Ramon, V., Paredes, J. M., Rib\'o, M. et al. {\it ApJ} {\bf 628}, 388 (2005)
1411: \bibitem{martocchia05} Martocchia, A., Motch, C., Negueruela, I., {\it A\&A} {\bf 430}, 245 (2005)
1412: \bibitem{ribo99} Rib\'o, M., Reig, P., Mart\'i, J., Paredes, J. M., {\it A\&A}, {\bf 347}, 518 (1999)
1413: \bibitem{ribo08} Rib\'o, M., Paredes, J. M., Moldon, J., Mart\'i, J., Massi, M. {\it A\&A} {\bf 481}, 17 (2008)
1414: \bibitem{ribo02} Rib\'o, M. Ph. D. Thesis, Universitat de Barcelona (2002)
1415: \bibitem{marti98} Mart\'i, J., Paredes, J. M., Rib\'o, M., {\it A\&A} {\bf 338}, 71 (1998)
1416: \bibitem{horns06} Horns, D., for the HESS collaboration, talk presented in 2nd Workshop On TeV Particle Astrophysics (2006)
1417: \bibitem{bosch07b} Bosch-Ramon, V., Motch, C., Rib\'o, M., et al. {\it A\&A} {\bf 473}, 545 (2007)
1418: \bibitem{denaurois06} de Naurois, M. for the HESS collaboration, talk presented in The keV to TeV connection (2006)
1419: \bibitem{dubus07} Dubus, G., Cerutti, B., Henri, G. {\it A\&A} {\bf 477}, 691 (2008)
1420: \bibitem{sierpowska07} Sierpowska-Bartosik, A., Torres, D.~F. {\it ApJL}, {\bf 671}, 145 (2007)
1421: \bibitem{sierpowska08} Sierpowska-Bartosik, A., Torres, D.~F. {\it ApJL}, {\bf 674}, 89 (2008)
1422: \bibitem{hutchings81} Hutchings, J. B. \& Crampton, D. {\it PASP} {\bf 93}, 486 (1981)
1423: \bibitem{gregory02} Gregory, P. C. {\it ApJ}, {\bf 575}, 427 (2002)
1424: \bibitem{casares05b} Casares, J., Ribas, I., Paredes, J. M., Mart\'i, J., Allende Prieto, C. {\it MNRAS} {\bf 360}, 1105 (2005)
1425: \bibitem{massi01} Massi, M., Rib\'o, M., Paredes, J. M., Peracaula, M., Estalella, R. {\it A\&A} {\bf 376}, 217 (2001)
1426: \bibitem{massi04} Massi, M., Rib\'o, M., Paredes, J. M., et al. {\it A\&A} {\bf 414}, L1 (2004)
1427: \bibitem{sidoli06} Sidoli, L., Pellizzoni, A., Vercellone, S., et al. {\it A\&A} {\bf 459}, 901 (2006)
1428: \bibitem{paredes07} Paredes, J. M., Rib\'o, M., Bosch-Ramon, V., et al. {\it ApJ} {\bf 664}, 39 (2007)
1429: \bibitem{dhawan06} Dhawan, V., Mioduszewski, A., Rupen, M. The VI Microquasar Workshop: Microquasars and Beyond (Proceedings of Science), 52, 1 (2006)
1430: \bibitem{dubus06b} Dubus, G. {\it A\&A} {\bf 456}, 801 (2006)
1431: \bibitem{chernyakova06} Chernyakova, M., Neronov, A., Walter, R. {\it MNRAS} {\bf 372}, 1585 (2006)
1432: \bibitem{maraschi81} Maraschi, L. \& Treves, A. {\it MNRAS} {\bf 194}, 1 (1981)
1433: \bibitem{romero07} Romero, G.~E., Okazaki, A.~T., Orellana, M., Owocki, S.~P. {\it A\&A} {\bf 474}, 15 (2007)
1434: \bibitem{bogovalov08} Bogovalov, S.~V., Khangulyan, D., Koldoba, A.~V., Ustyugova, G.~V., Aharonian, F.~A. {\it MNRAS}
1435: {\bf 387}, 63 (2008)
1436: \bibitem{kniffen97} Kniffen, D.~A., Alberts, W.~C.~K., Bertsch, D.~L. et al. {\it ApJ} {\bf 486} 126 (1997)
1437: \bibitem{romero05b} Romero, G. E., Christiansen, H. R., Orellana, M. {\it ApJ} {\bf 632}, 1093 (2005)
1438: \bibitem{bosch06c} Bosch-Ramon, V., Paredes, J.~M., Romero, G. E., Rib\'o, M. {\it A\&A} {\bf 459}, L25 (2006)
1439: \bibitem{leahy04} Leahy, D. A., {\it A\&A}, {\bf 413}, L1019 (2004)
1440: \bibitem{christiansen06} Christiansen, H.~R., Orellana, M., Romero, G.~E. {\it Phys. Rev. D.} {\bf 73}, 3012 (2006)
1441: \bibitem{torres07} Torres, D.~F. \& Halzen F. {\it Astropart.Phys.} {\bf 27}, 500 (2007)
1442: \bibitem{grundstrom07} Grundstrom, E.~D., Caballero-Nieves, S.~M., Gies, D.~R. et al. {\it ApJ} {\bf 656}, 437 (2007)
1443: \end{thebibliography}
1444: \end{document}
1445: