A&A 612, A10 (2018) DOI: 10.1051/0004-6361/201527773 Astronomy
&
 cESO 2018 
Astrophysics 

H.E.S.S. phase-I observations of the plane of the Milky Way Special issue 
Asearchfor veryhigh-energy fares from the microquasars GRS 1915+105, Circinus X-1, and V4641 Sgr using contemporaneous H.E.S.S. and RXTE observations 
H.E.S.S. 
Collaboration:H. Abdalla1,A. Abramowski2,F. Aharonian3, 4, 5,F. Ait Benkhali3,A.G. Akhperjanian6, 5, E. O. Anger7, M. Arrieta15, 

P. 
Aubert24,M. Backes8,A. Balzer9,M. Barnard1,Y. Becherini10,J. Becker Tjus11,D. Berge12,S. Bernhard13,K. Bernlr3,E. Birsin7, 

R. 
Blackwell14,M. Btcher1,C. Boisson15,J. Bolmont16,P. Bordas?, 3,J. Bregeon17,F. Brun18,P. Brun18,M. Bryan9,T. Bulik19,M. Capasso29, 

J. 
Carr20, S. Casanova21, 3,P.M. Chadwick?, 43, N. Chakraborty3, R. Chalme-Calvet16, R. C.G. Chaves17, 22, A. Chen23, J. Chevalier24, 

M. 
Chrétien16, S. Colafrancesco23, G. Cologna25, B. Condon26, J. Conrad27, 28,C. Couturier16,Y. Cui29,I.D.Davids1, 8, B. Degrange30, C. Deil3, 

P. 
deWilt14, H. J. Dickinson?, 44,A. Djannati-Ataï31,W. Domainko3,A. Donath3, L.O’C.Drury4,G.Dubus32,K. Dutson33,J.Dyks34,M.Dyrda21, 

T. 
Edwards3,K. Egberts35,P. Eger3, J.-P. Ernenwein?, 20,S. Eschbach36,C.Farnier27, 10,S.Fegan30,M.V. Fernandes2,A. Fiasson24, 


G.Fontaine30,A. Fster3,S. Funk36,M. Fling37,S. Gabici31,M. Gajdus7,Y.A. Gallant17,T. Garrigoux1,G. Giavitto37,B. Giebels30, 
J.F. Glicenstein18,D. Gottschall29,A.Goyal38, M.-H. Grondin26,M. Grudzi´, 
nska19, D. Hadasch13, J. Hahn3, J. Hawkes14, G. Heinzelmann2 

G. Henri32,G. Hermann3,O. Hervet15,A. Hillert3,J.A. Hinton3,W. Hofmann3,C. Hoischen35,M. Holler30,D. Horns2,A.Ivascenko1, 
A. Jacholkowska16,M. Jamrozy38,M. Janiak34,D. Jankowsky36,F. Jankowsky25,M. Jingo23,T. Jogler36,L. Jouvin31,I. Jung-Richardt36, 
M.A. 
Kastendieck2,K. Katarzynski´39,U. Katz36,D.Kerszberg16,B. Khélif31,M. Kieffer16, J. King3, S. Klepser37, D. Klochkov29, 

W. 
Klu´zniak34,D.Kolitzus13, Nu.Komin23,K.Kosack18,S. Krakau11,M. Kraus36,F. Krayzel24,P.P. Krer1,H.Laffon26, G. Lamanna24, 

J. 
Lau14, J.-P. Lees24,J. Lefaucheur31,V. Lefranc18,A. Lemière31,M. Lemoine-Goumard26, J.-P. Lenain16,E. Leser35,T. Lohse7,M. Lorentz18, 

R. 
Liu3,I.Lypova37,V. Marandon3,A. Marcowith17,C. Mariaud30,R. Marx3,G. Maurin24,N. Maxted17,M. Mayer7,P.J. Meintjes40, 

U. 
Menzler11,M.Meyer27,A. M.W. Mitchell3,R. Moderski34,M. Mohamed25,K. Morå27,E. Moulin18,T. Murach7,M.de Naurois30, 

F. 
Niederwanger13,J. Niemiec21,L. Oakes7,H. Odaka3,S. Öttl13,S. Ohm37,M. Ostrowski38,I.Oya37,M.Padovani17,M.Panter3,R.D.Parsons3, 


M.Paz 
Arribas7,N.W. Pekeur1,G. Pelletier32,P.-O. Petrucci32,B.Peyaud18,S. Pita31,H. Poon3,D. Prokhorov10,H. Prokoph10,G. Plhofer29, 

M. 
Punch31, 10,A. Quirrenbach25,S. Raab36,A. Reimer13,O. Reimer13,M. Renaud17,R.de losReyes3,F. Rieger3, 41, C. Romoli4, 

S. 
Rosier-Lees24,G.Rowell14,B. Rudak34,C.B. Rulten15,V. Sahakian6, 5, D. Salek42, D. A. Sanchez24, A. Santangelo29, M. Sasaki29, 

R. 
Schlickeiser11,F. Schsler?,18, A. Schulz37, U. Schwanke7, S. Schwemmer25, A. S. Seyffert1, N. Shaf23, I. Shilon36, R. Simoni9, H. Sol15, 

F. 
Spanier1,G. Spengler27,F. Spies2,Ł. Stawarz38,R. Steenkamp8,C. Stegmann35, 37,F. Stinzing†,36,K. Stycz37,I. Sushch1, J.-P.Tavernet16, 


T.Tavernier31,A.M.Taylor4,R.Terrier31,M. Tluczykont2,C.Trichard24,R.Tuffs3,J.van derWalt1,C.van Eldik36,B.van Soelen40, 
G.Vasileiadis17,J.Veh36,C.Venter1,A.Viana3,P.Vincent16,J.Vink9,F.Voisin14,H.J. Vk3,T.Vuillaume24,Z.Wadiasingh1,S.J.Wagner25, 
P.Wagner7,R.M.Wagner27,R. White3,A.Wierzcholska21,P.Willmann36,A. Wnlein36,D.Wouters18,R.Yang3,V. Zabalza33,D. Zaborov30, 
M. Zacharias25,A.A. Zdziarski34,A. Zech15,F. Zef30,A. Ziegler36, andN. ˙Zywucka38 
(Affiliations can be found after the references) 
Received 17 November 2015 / Accepted 28 February 2016 
ABSTRACT 
Context. Microquasars are potential γ-ray emitters. Indications of transient episodes of γ-ray emission were recently reported in at least two systems: Cyg X-1 and Cyg X-3. The identifcation of additional γ-ray-emitting microquasars is required to better understand how γ-ray emission can be produced in these systems. Aims. Theoretical models have predicted very high-energy (VHE) γ-ray emission from microquasars during periods of transient outburst. Observations reported herein were undertaken with the objective of observing a broadband faring event in the γ-ray and X-ray bands. Methods. Contemporaneous observations of three microquasars, GRS 1915+105, Circinus X-1, and V4641 Sgr, were obtained using the High Energy Spectroscopic System (H.E.S.S.) telescope array and the Rossi X-rayTiming Explorer (RXTE) satellite. X-ray analyses for each microquasar were performed and VHE γ-ray upper limits from contemporaneous H.E.S.S. observations were derived. Results. No signifcant γ-ray signal has been detected in any of the three systems. The integral γ-ray photon fux at the observational epochs is constrained to be I(>560 GeV) < 7.3× 10−13 cm−2s−1, I(>560 GeV) < 1.2× 10−12 cm−2s−1, and I(>240 GeV) < 4.5× 10−12 cm−2s−1 for GRS 1915+105, Circinus X-1, and V4641 Sgr, respectively. Conclusions. The γ-ray upper limits obtained using H.E.S.S. are examined in the context of previous Cherenkov telescope observations of microquasars. Theeffect of intrinsic absorption is modelled for each target and found to havenegligible impact on the fux of escaping γ-rays. When combined with the X-ray behaviour observed using RXTE, the derived results indicate that if detectable VHE γ-ray emission from microquasars is commonplace, then it is likely to be highly transient. 
Keywords. gamma rays: general– X-rays: binaries – X-rays: individuals: GRS 1915+105 – X-rays: individuals: Circinus X-1 – X-rays: individuals: V4641 Sgr 
? Corresponding authors: H.E.S.S. Collaboration, e-mail: contact.hess@hess-experiment.eu 
† Deceased. 

Article publishedby EDP Sciences A10, page1of 22 
1. Introduction 

Microquasars are X-ray binaries that exhibit spatially re-solved, extended radio emission. The nomenclature is moti<BR>vatedbya structural similarity with the quasarfamilyof active galactic nuclei(AGN). Both object classes are believed to com-prise a compact central object embedded in a fow of accreting material, and both exhibit relativistic, collimated jets. In the cur-rent paradigm, both microquasars andAGN derive their power from the gravitational potential energy that is liberated as ambi<BR>ent matterfalls onto the compact object. Notwithstanding their morphological resemblance, microquasars and radio-loudAGN represent complementary examples of astrophysical jet produc<BR>tion on dramatically disparate spatial and temporal scales. In-deed, conditions of accretion and mass provision that pertain to the supermassive (106 M∼ <BR><MBH ∼ <BR><109 M)black holes that powerAGNandofthe stellar-mass compact primariesof micro-quasars are markedly different. In the latter, a companion star (or donor) provides the reservoir of matter for accretion onto a compact stellar remnant (or primary), which can be either a neutron star ora black hole.Partial dissipationof the resultant power output occurs in a disk of material surrounding the pri-mary, producing the thermal and non-thermal X-ray emission, which is characteristic of all X-ray binary systems. Microquasars are segregated on the basis of associated non-thermal radio emis<BR>sion, indicative of synchrotron radiation in a collimated outfow, which carries away a sizeable fraction of the accretion luminos-ity(Fenderetal. 2004b).InAGN, superfcially similarjet struc<BR>tures are known to be regions of particle acceleration and non-thermal photon emission. The resulting radiation spectrum can extend from radio wavelengths into the very high-energy (VHE; Eγ> <BR>100 GeV) γ-ray regime. Very high-energy γ-ray emis<BR>sionhasbeenobservedfrommanyAGNintheblazar sub-class1, where the jet axis is aligned close to the observer line-of-sight, aswellasfromafew radiogalaxies(e.g.M87, Aharonianetal. 2003; Cen A, Aharonian et al. 2009; NGC 1275, Aleksi´c et al. 2012) and starburst galaxies (e.g. M82, Acciari et al. 2009; NGC 253, Abramowski et al. 2012). If similar jet production and efficient particle acceleration mechanisms operate in microquasars and AGNs, this might imply that the former object class are plausible sources of de-tectable VHE γ-ray emission as well, assuming that appropriate environmental conditions prevail. The primarily relevant environmental conditions include the density of nearby hadronic ma-terial, which provides scattering targets for inelastic proton scat-tering interactions; these interactions produce pions that produce γ-rays when they subsequently decay. The ambient magnetic feld strength is also important and infuences the rate at which electrons lose energy via synchrotron radiation. Synchrotron photons contribute to the reservoir of soft photons that are available for inverse Compton (IC) up-scattering into the VHE γ-ray regime. The argument for phenomenological parity between AGN and microquasars, possibly related to their structural re-semblance, has been strengthened in recent years as the spectral properties of both radio and X-ray emission are remarkably sim-ilar for both stellar mass and supermassive black holes. In recent years these similarities led to the postulation of a so-called fundamental plane, which describes a three-dimensional, phenomenological correlation between the radio(5 GHz) and X-ray (2−10keV) luminosities and the black hole mass(Merloni et al. 2003; Falcke et al. 2004). However, the fundamental plane doesnot appeartoextendintotheTeVband.Todate,onlyone well-established microquasar has been observed to emit in the 
http://tevcat.uchicago.edu/ <BR>
VHE γ-ray regime. This is the Galactic black hole Cygnus X-1, whichwas marginally detected(atthe ∼4σ level)bytheMAGIC telescope immediately priortoa2−50keV X-ray fare observed by the INTEGRAL satellite, the Swift Burst Alert Telescope (BAT), and the RXTE All-Sky Monitor (ASM; Albert et al. 2007;Malzac et al. 2008). Laurent et al. (2011)recently identi<BR>fed linear polarized soft γ-ray emission from Cygnus X-1 (see also Jourdain et al. 2012),thereby locating the emitter within the jets and identifying their capacity to accelerate particles to high energies (see however Romero et al. 2014). Further motivation for observing microquasars in the VHE band arises from the recent identifcation of the high-mass microquasar Cygnus X-3 as a transient high-energy (HE; 100 MeV < Eγ< 100 GeV) γ-ray source by the Fermi (Abdo et al. 2009) and AGILE (Tavani et al. 2009)satellites. The identifcation of a periodic modulation of the HE signal is consistent with the orbital frequency of Cygnus X-3 and provides compelling evidence for effective acceleration of charged particles to GeV energies withinthe binary system(Abdoetal. 2009). Basedonevidence from subsequent reobservations, the HE γ-ray fux from Cygnus X-3 appears to be correlated with transitions observed in X-rays in and out of the so-called ultra-soft state, which exhibits bright soft X-ray emission and low fuxes in hard X-rays and is typically associated with contemporaneous radio faring activity 
(e.g. Corbel et al. 2012). Unfortunately, repeated observations of Cygnus X-3 using the MAGIC telescope did not yield a signifcant detection(Aleksi´c et al. 2010), despite the inclusion of data that were obtained simultaneously with the periods of enhanced HE emission detected using Fermi. However, the intense optical and ultraviolet radiation felds produced by the Wolf-Rayet companion star in Cygnus X-3 imply a 
− 
large optical depth for VHE γ-rays due to absorption via e+epair production (e.g. Bednarek 2010; Zdziarski et al. 2012). Accordingly, particle acceleration mechanisms akin to those operating in Cygnus X-3 may yield detectable VHE γ-ray fuxes in systems withfainter or cooler donors. 
Mechanisms for γ-ray production in microquasars have been widely investigated, resulting in numerous hadronic (see e.g. Romero et al. 2003)and leptonic (see e.g.Atoyan&Aharonian 1999; Georganopoulos et al. 2002; Bosch-Ramon et al. 2006; Dermer&Btcher 2006;Dubus et al. 2010)models, describing theexpectedfuxesand spectraof microquasarsintheGeV−TeV band. In both scenarios, a highly energetic population of the relevant particles is required and, consequently, emission scenarios generally localize the radiating region within the jet structures of the microquasar. Leptonic models rely upon IC scattering of photons from the primary star in the binary system or photons produced through synchrotron emission along the jet to produce VHE γ-ray emission. In this latter scenario, they closely resemble models of extragalactic jets (Kigl 1981; Ghisellini&Maraschi 1989),but typicallyinvoke internal mag-netic felds that are stronger by factors ∼1000. Consideration of hadronic models is motivated by the detection of Doppler-shifted emission lines associated with the jets of the microquasar SS 433 (e.g. Margon 1984), indicating that at least some mi-croquasar jets comprisea signifcant hadronic component. Mod-els of VHE γ-ray production by hadronic particles generally invoke electromagnetic cascades initiated by both neutral and charged pion decays(Romero et al. 2003;Aharonian&Atoyan 1996;Romero et al. 2005). 
Electron-positron pair production, γγ → e+e− can absorb VHE γ-rays. In the case of 1 TeV γ-rays, the cross section for this process is maximised for ultraviolet target photons (Eph ∼ 10 eV), where its value may be approximated in 
A10, page2of 22 

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars 
terms of the Thomson cross section as σγγ <BR>≈ <BR>σT <BR>/5 (e.g. Gould&Schréder 1967). In high-mass systems, the companion star is expected to produce a dense feld of these target pho<BR>tons to interact with the γ-rays (e.g. Protheroe&Stanev 1987; Moskalenko 1995;Btcher&Dermer 2005;Dubus 2006). This process can be very signifcant and probably contributes to the observed orbital modulation in the VHE γ-ray fux from LS 5039 (Aharonian et al. 2006c). In contrast, the ultravio<BR>let spectrum of low-mass microquasars is likely dominated by the reprocessing of X-ray emission in the cool outer accretion fow(vanParadijs&McClintock 1994; Gierli´nski et al. 2009), although jet emission might also be signifcant(Russell et al. 2006). Regardless of their origin, the observed optical and ultraviolet luminosities of low-mass X-ray binaries (LMXBs) are generally orders of magnitude lower than those of high-mass systems(Russell et al. 2006), and the likelihood of strong γ-ray absorption is correspondingly reduced. However, microquasars may only become visiblein theTeV band during powerful faring events. These transient outbursts, characterised by the ejection of discrete superluminal plasma clouds, are usually observed at the transition between low-and high-luminosity X-ray states(Fender et al. 2004b). Monitoring black-hole X-ray binaries with radio telescopes and X-ray satellites operating in the last decade enabled a classifcation scheme of such events to be established (Homan&Belloni 2005). Hardness-intensity diagrams (HIDs) plot the source X-ray inten-sity against X-ray colour (or hardness) and have subsequently been extensively used to study the spectral evolution of black-hole outbursts. At the transition from the so-called low-hard state to the high-soft states through the hard-to-soft intermediate states, the steady jet associated with the low-hard state is disrupted. These transient ejections, produced once the accretion disk collapses inwards, are more relativistic than the steady low-hard jets(Fender et al. 2004b). Internal shocks can develop in the outfow, possibly acceleratingparticles that subsequently give rise to radio optically thin fares observed from black-hole systems; this phenomenological description is also extensible to neutron stars, although in that case jet radio power is lower by a factor5−30(Migliari&Fender 2006). 
Outburst episodes have also been observed in cases in which the source remained in the hard state without transition to the soft state(Homan&Belloni 2005). The detection (at the ∼4σ level) by the MAGIC telescope of the high-mass, black-hole binary Cygnus X-1 took place during an enhanced2−50keV fux low-hard state as observed with the INTEGRAL satellite, the Swift BAT, and the RXTE ASM(Malzac et al. 2008). However, al-though the source X-ray spectrum remained unchanged through-out theTeV fare, sucha bright hard statewas unusually long when compared with previous observations of the source. 
Here we report on contemporaneous observations with 
H.E.S.S. and RXTE of the three microquasars V4641 Sgr, GRS 1915+105, and Circinus X-1. Information on the targets, the H.E.S.S. and RXTE observations, and the corresponding trigger conditions are detailed in Sect. 2. Analysis results are reported in Sect. 3 and discussed in Sect. 4. In the appendix, detailed information on the X-ray analysis is reported, which in particular includes HIDs corresponding to the time of observations for the three studied sources. 
2.Targets and observations 
2.1. Observations 
The H.E.S.S. Imaging Atmospheric Cherenkov Telescope (IACT) array is situated on the Khomas Highland plateau of 
Table 1. Observationally established parameters of the target microquasars. 
GRS 1915+105 Circinus X-1 V4641 Sgr 
Porb [d] 33.85 ± 0.16 (1) 16.6 (2) 2.82 (3) M? [M&#12;]0.47 ± 0.27(1) 3−10(5) 2.9± 0.4(10) 0.28 ± 0.02 (4) MCO[M&#12;] 12.4+2.0 (6) ∼<1.4(8) 6.4± 0.6(10) 
−1.8 θJet[◦] 60 ± 5(6) ∼<3(9) ∼<12 (3) 
d [kpc] 8.6+2.0 (6,7) 9.4+0.8 (9) 6.2± 0.7(10) 
−1.6 −1.0 
Notes. Porb is the binary orbital period, M? is the mass of the com-panion star, MCO is the compact object mass, θJet is the inclination of the observed jet with respect to the line of sight, and d is the estimated distance to the microquasar. 
References. (1) Steeghs et al. (2013); (2) Nicolson (2007); 
(3) Orosz et al. (2001); (4) Ziolkowski (2015); (5) Johnston et al. (1999); Jonker et al. (2007); (6) Reid et al. (2014); (7) Zdziarski (2014);(8) Tennantetal.(1986b); Linaresetal. (2010);(9) Heinzetal. (2015); (10) MacDonald et al. (2014). 
Namibia (23◦1601800 south, 16◦3000000 east), at an elevation of 1800 m above sea level, and is capable of detecting a Crablike source close to the zenith at the 5σ level within <5 min under good observational conditions(Aharonian et al. 2006a). The point source sensitivity of H.E.S.S. enables it to detect 
−2
a2.0 × 10−13 cms−1 γ-ray fux above 1 TeV, at the 5σ level within 25 h, which, together with a low-energy threshold(∼100 GeV), makes H.E.S.S. an invaluable instrument for studying the VHE γ-ray emission from microquasars.Affth and larger telescope (commissioned in 2013) will allow the energy threshold to be lowered and will further increase the sensitivity of the instrument.For the analysis presented here, H.E.S.S. ob-servations were carried out using the full, original four-telescope array. Owing to the diverse morphologies of the three binary systems, unique observational trigger criteria were established for each target employing various combinations of the observed X-ray state and radio faring activity. Details are provided in sub-sequent paragraphs. 
The Rossi X-ray Timing Explorer (RXTE) was a space-based X-ray observatory launched on 30 December 30 1995 and decommissioned on 5 January 2012. The primary mission of RXTE was to provide astrophysical X-ray data with high timing resolution. This observatory occupied a circular low-earth orbit with an orbital period of ∼90 min and carried three sep-arate X-ray telescopes. The Proportional Counter Array (PCA) on board RXTE comprised fve copointing xenon and propane Proportional Counter Units (PCUs), which were nominally sen-sitive in the energy range ∼2−60keV with an energy resolution of <18%at6keV(Zhangetal. 1993).For studiesof rapidly varying sources like X-ray binaries, the PCA timing resolution of ∼1 µs can prove invaluable. However, rapid timing measurements also require a bright source to provide sufficient counting statistics within such short time bins. The High Energy X-ray-Timing Experiment (HEXTE) comprised two independent clusters of four phoswich scintillation detectors, which were sensitive to photons in the ∼12−250 keV energy range and had an energy resolution of ∼9 keV at 60 keV. The maximum timing resolution of HEXTE was ∼8 µs. The All-SkyMonitor (ASM) wasa wide feld-of-view instrument that monitored ∼80% of the sky over the course of each ∼90 min orbit. This instrument con-sisted of three identical scanning shadow cameras and was de-signed to provide near-continuous monitoring of bright X-ray sources. Nominally, the ASM was sensitive in the energy range 
A10, page3of 22 

from 2−10keV and had a rectangular feld of view spanning 110◦× 12◦ . 
Contemporaneous X-ray (RXTE) and VHE γ-ray (H.E.S.S.) observations were performed at the epochs listed inTable 2. In the following, we briefy review the observational characteristics of the target microquasars, GRS 1915+105, Circinus X-1, and V4641 Sgr. Established system parameters that characterise the three target microquasars are collatedinTable 1. 
2.2. GRS 1915+105 
GRS 1915+105 is a dynamically established black-hole binary frst identifed by the WATCH all-sky monitor on board the GRANAT satellite (Castro-Tirado et al. 1994). Observations in the optical and near-infrared using the Very Large Telescope succeeded in identifying the stellar companion as a low-mass KM III giant(Greiner et al. 2001). GRS 1915+105 gained a measure of celebrity as the prototype Galactic superluminal source (Mirabel&Rodriguez 1994). 
In a detailed study of the X-ray light curves of GRS 1915+105, Belloni et al. (2000) succeeded in identifying 12 distinct variability classes, internally characterised by the duration and juxtaposition of three separate spectral states. Episodes of class χ behaviour, belonging to state C and last-ing several days, are known as plateaux and are invariably ter-minated by faring activity in the radio, infrared, and X-ray bands(Fender&Belloni 2004). In contrast with the evidence for self-absorbed synchrotron radiation seen in the spectrally hard, low-luminosity state C, and often associated with continuous relativistic jets(Klein-Wolt et al. 2002), radio spectra ob-tained during the end-plateau faring episodes indicate optically thin synchrotron emission(Fender et al. 1997;Eikenberry et al. 1998). Occasionally, these faring episodes are linked to power-ful discrete plasma ejections with instantaneous power output reaching >∼1039 ergs−1(Mirabel&Rodriguez 1994; Zdziarski 2014). Modelling the emission from these discrete relativistic ejecta, Atoyan&Aharonian (1999)showed that inverse Comp<BR>tonisation of emitted synchrotron photons into the GeV-TeV regime could produce signifcant and persistent γ-ray fuxes that remain detectable for several days. 
Acero et al. (2009) and Saito et al. (2009) reported VHE γ-ray observations of GRS 1915+105; these authors derived in
× 10−13 −2 
tegral fux upper limitsof6.1 cms−1 above 410 GeV −2
and1.17 × 10−12cms−1 above 250 GeV, respectively. 
For the analysis presented here, GRS 1915+105 was ob-served by H.E.S.S. between 28 April and 3 May 2004 in response to an apparent decrease in the 15 GHz radio fux, which was monitoredby the RyleTelescope duringa ∼50 day plateau state(Pooley 2006), as shown in Fig. A.2, in which coloured markers indicate the H.E.S.S. observation epochs. On the basis of previously observed behaviour, it was thought likely that the observed radio evolution signalled the end of the plateau state and, therefore, that faring activity would begin within the sub-sequent 24 h. The RXTE observations of GRS 1915+105 com-prised six individual pointings, contributing to accumulated PCA and HEXTElivetimesof7.6ks and 5176 s, respectively. Fifteen contemporaneous H.E.S.S. observations were obtained, constituting an overall livetime of 6.9 h. 


2.3. Circinus X-1 
Circinus X-1 (hereafter Cir X-1) has been extensively studied since its initial identifcation(Margon et al. 1971), reveal
A10, page4of 22 
ing a somewhat confusing collection of complex observational characteristics. 
Repeated observation of typeI X-raybursts(Tennant et al. 1986a,b;Linares et al. 2010)defnitively identifes the compact primaryin Cir X-1 asalow magnetic feld(B <∼ 1011 G) neutron star. Further sub-classifcation as a Z or atoll source (see, for example, Done et al. 2007, for an explanation of the distinc<BR>tion between these two classes) is not possible since Cir X-1 exhibits several confusing spectral and timing properties, sub-sets of which are characteristic of both source types (see e.g. Shireyet al. 1999a; Oosterbroek et al. 1995). Accordingly, es-tablished paradigms for disk-jet coupling in X-ray binaries with neutron star primaries(e.g. Migliari&Fender 2006)cannotbe reliably employed. 
At radio wavelengths, the jets of Cir X-1 display notable structure on arcsecond scales, appearing as a bright core with signifcant extension along the axial direction of the arcminute jets(Fenderetal.1998).Infact,the observedextensionis rather asymmetric with a ratio of at least two between the observed fuxes of the two opposing jets. Interpreted as pure relativistic aberration, this asymmetry implies a jet velocity >∼0.1c. Cir X-1 has also been observed to eject condensations of matter with apparently superluminal velocities >∼15c (Fender et al. 2004a). These observations imply a physical velocity for the ejecta v> 0.998c with a maximum angle between the velocity vector and the line of sight θ< 5◦. These results identify Cir X-1 as a mi-croblazar, a Galactic, small-scale analog of the blazar class of AGN, several of which are known sources of VHEγ-rays. 
Defnitive classifcation of the donor star in Cir X-1 is somewhat problematic. The low apparent magnitude of the detected optical counterpart implies a dereddened luminosity consistent with a low-mass or sub-giant companion, implying that Cir X-1 is a LMXB with a high orbital eccentricity e ∼ 0.7−0.9 (e.g. Johnston et al. 1999). Nonetheless, recent near-infrared(Clark et al. 2003)andI-band optical(Jonker et al. 2007)observations reveal emission features that are consistent with a mid-B supergiant, suggesting a more moderate eccentricity e ∼ 0.45. 
ObservationsofCirX-1intheX-raybandreveala long-term evolution of the average source brightness. Fluxes rose mono-tonically from near-undetectable in the early 1970s to a peak value of ∼1.5−2 Crab (1.5−10keV) at the turn of the millennium, before returning over a period of ∼4 yr to their pre-rise levels(Parkinsonetal. 2003).Various X-ray spectra, obtained during epochs of both high and low fux, display evidence of complex and variable emission and absorption processes. 
A previous analysis of H.E.S.S. observations of Cir X-1 was presentedby Nicholas&Rowell (2008), who deriveda pre<BR>liminary upper limit to the γ-ray fux above 1 TeV of 1.9 × 
10−13 −2
cms−1 corresponding to a detector livetime of 28 h. The H.E.S.S. observations of Cir X-1 reported here began on 18 June 2004 and were scheduled to coincide with the periastron passage of the binary components. The previous observation of regular radio fares during this orbital interval were thought to provide a good chance of observing during a period of outburst with the associated possibility that superluminal ejections might occur. The RXTE observations of Cir X-1 comprised three individual pointings, corresponding to orbital phase intervals0.0486 ≤ φ ≤ 0.0498,0.1104 ≤ φ ≤ 0.1112, and 0.1718 ≤ φ ≤ 0.1725 (using the radio fare ephemeris of Nicolson 2007), and contributing to an accumulated PCA livetimeof 2576s.A data set comprising12 contemporaneous 
H.E.S.S. observations yielded a combined livetime of 5.4 h. 

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars Table 2. Observational epochs for each target microquasar. 
Target  RXTE ObsId  RXTE Observations (MJD)  H.E.S.S. Observations (MJD)  
GRS 1915+105  90108-01-01-00 90108-01-02-00 90108-01-03-00 90108-01-04-00 90108-01-05-00 90108-01-06-00  53 123.091 → 53 123.109 53 124.074 → 53 124.094 53 125.130 → 53 125.149 53 126.114 → 53 126.129 53 127.097 → 53 127.114 53 128.150 → 53 128.165  53 123.067 → 53 123.150 53 124.079 → 53 124.162 53 125.083 → 53 125.148 53 126.109 → 53 126.132 53 127.106 → 53 127.165 53 128.149 → 53 128.165  
Cir X-1  90124-02-01-00 90124-02-02-00 90124-02-03-00  53 174.749 → 53 174.761 53 175.768 → 53 175.780 53 176.781 → 53 176.793  53 174.748 → 53 174.832 53 175.735 → 53 175.822 53 176.772 → 53 176.858  
V4641 Sgr  90108-03-01-00 90108-03-02-00 90108-03-03-00  53 193.904 → 53 193.924 53 194.887 → 53 194.908 53 195.871 → 53 195.892  Not Observed 53 194.883 → 53 194.926 53 195.890 → 53 195.931  


2.4. V4641 Sgr 
V4641 Sgr is the optical designation of the habitually weak X-ray source SAX J1819.3-2525 (XTE J1819-254), which was independently identifed using the BeppoSAX(in ’tZand et al. 1999) and RXTE(Markwardt et al. 1999a) satellites. Optical spectroscopic measurements(Orosz et al. 2001;Lindstr et al. 2005)strongly suggestalateB-orearlyA-type companionwith an effective temperature Teff ≈ 10 500 K. The mass of the compact primary,6.4± 0.6M&#12; (MacDonald et al. 2014), categorises V4641 Sgr as a frm black hole candidate. 
V4641 Sgr is probably best known for its exhibition of rapid and violent outbursts. Perhaps the most spectacular of these events was the super-Eddington fare detected by the RXTE ASM in September 1999. The observed X-ray fuxes (2−12keV) increased sharply, reaching ≈12.2Crab within eight hours before fading again to below 0.1 Crab in under two hours(Revnivtsev et al. 2002).Powerful contemporaneous fares were also observed at hard X-ray (McCollough et al. 1999), optical(Stubbings&Pearce 1999), and radio(Hjellming et al. 2000) wavelengths. In fact, Very Large Array (VLA) radio observations obtained within a day of the X-ray fare re-solved a bright jet-like radio structure ≈0.25 arcsec in length (Hjellming et al. 2000). Assuming the most likely hypothesis, i.e., that the ejection is coincident with some phase of the X-ray fare, proper motions in the range 0.22 <∼ µjet ∼ <1.1 arc-sec day−1 are derived. At the minimum distance d = 5.5 kpc, the implied lower limit to the apparent velocity of the ejecta is 7c <∼ vmin ∼ <35c, which is comparable with the extragalactic jets seen in blazars. Indeed, the remarkably high apparent velocities imply that V4641 Sgr may be a microblazar with a relativistic jet moving close to the line of sight(θjet . 12◦; from Orosz et al. 2001). Subsequent, weaker broadband out-bursts have also been observed, suggesting recurrent activity on a timescale ∼1−2 yr (e.g. Hjellming 2000; Rupen et al. 2002, 2003;Swank 2004). 
Observations of V4641 Sgr with H.E.S.S. were initiated on 7 July 7 2004 (MJD 53 193) in response to the source brightening rapidly in the radio (Rupen et al. 2004b), optical(Revnivtsev et al. 2004), and X-ray(Swank 2004) bands. The resultant RXTE exposure comprised three observations, each contributing to an accumulated PCA livetime of 5ks. Two pairs of ∼30 min H.E.S.S. observations were obtained contemporaneously with the fnal two RXTE pointings. In total, the four separate exposures constitute an overall livetime of 
1.76 h. 
3. Analysis and results 
X-ray data reduction with the FTOOLS 5.3.1 software suite employed the data selection criteria regarding elevation, offset, electron contamination, and proximity to the South Atlantic Anomaly recommendedby the RXTE Guest ObserverFacility website2. For each observation, the PCA STANDARD2 data were extracted from all available PCUs. For all observations, HEXTE Archive mode data for both clusters were extracted following the recommended procedures for time fltering and background estimation. Spectral analysis was carried out using the XSPEC 12.6.0 package(Arnaud 1996). Spectral fts for GRS 1915+105 use both PCA and HEXTE data, including an energy range of3−200keV.For bright X-ray sources, such as GRS 1915+105, statistical errors on the number of counts per spectral bin become insignifcant relative to dominant uncertainties in the instrument response. Accordingly, a 1% systematic error was added to all PCA channels. The remaining sources, Cir X-1 and V4641 Sgr, were not signifcantly detected by HEXTE and therefore only PCA data in the3−20keV range were considered to ensure good data quality. These targets were sufficientlyfaint that the spectral bin uncertainties were statistically dominated and the addition of a systematic error com-ponent was not required. In the case of GRS 1915+105, power density spectra (PDS) were derived using the ftool powspec.For each RXTE pointing of GRS 1915+105, individual PDS were extracted from 128s intervals comprising214 bins. The resulting spectra were then averaged to produce a PDS for the total light curve with errors estimated using the standard deviation of the average of the power in each frequency bin. The overall PDS were logarithmically rebinned and normalised to represent the squared fractional RMS in each frequencybin (see e.g. Lewin et al. 1988). Corrections for instrument deadtime (see, for example, Revnivtsev et al. 2000)were applied (although this was found to have a negligible effect in the frequencyrange un-der consideration) and the expected white noise level was sub-tracted(Leahyet al. 1983). Similar temporal analyses for the re-maining targets proved unfeasible because of insufficient count statisticsatallbutthelowest frequencies. 
The γ-ray analysis followed the standard point-source proce-dure described in Aharonian et al. (2006b). The refected back<BR>ground model (see, for example, Berge et al. 2007)was used to deriveoverall resultsin conjunctionwithboththe hard and standard event selection cuts describedby Aharonianetal. (2006b). 
2 http://heasarc.nasa.gov/docs/xte/xhp_proc_analysis. <BR>html <BR>
A10, page5of 22 


Fig. <BR>1. <BR>RXTE ASM, and PCA light curves for GRS 1915+105 together with H.E.S.S. upper limits derived from individual ∼28 min runs using standardevent selection cuts.Theblue shadedbandsontheASMlight curveindicatetheextentofthe H.E.S.S. observations,whileonthe H.E.S.S. upper limit plots similar bands illustrate the duration of the contemporaneous PCA observations. The plotted H.E.S.S. upper limits correspond to different threshold energies and the vertical scale of each light curve has been optimised for the plotted data. 
A10, page6of 22 H.E.S.S. Collaboration: H.E.S.S. observations of microquasars Table 3. H.E.S.S. VHE γ-ray signifcances corresponding to hard and standard event selection regimes. 

Target Image Cuts NON [events] NOFF [events] α Excess [events] Signifcance[σ] 
Standard 471 7127 0.073 –51.6 –2.2 
GRS 1915+105 

Hard 36 783 0.060 –10.9 –1.6 
Standard 385 5959 0.068 –20.1 –1.0 
Cir X-1 

Hard 45 648 0.056 9.1 1.4 
Standard 161 2373 0.067 1.2 0.1 
V4641 Sgr 

Hard 11 275 0.055 –4.2 –1.11 
Table 4. H.E.S.S. VHE γ-ray integral fux upper limits above the telescope energy threshold corresponding to both event selection regimes. 
Target Cuts TLive [s] Z¯max[◦] Ethresh [GeV] I(>Ethresh)[ph cm−2s−1] 
Standard 24681 40.6 562 <7.338 × 10−13 
GRS 1915+105 

Hard 24681 40.6 1101 <1.059 × 10−13 
Standard 19433 43.6 562 <1.172 × 10−12 
Cir X-1 

Hard 19433 43.6 1101 <4.155 × 10−13 
Standard 6335 8.4 237 <4.477 × 10−12 
V4641 Sgr 

Hard 6335 8.4 422 <4.795 × 10−13 
Notes. Theupper limitsarederivedatthe99% confdencelevel, assumingapower-law spectrum(dN/dE ∝ E−Γ)with the photon indexΓstd = 2.6 for standard cuts and Γhard = 2.0for hard cuts. The rather high threshold energies derived for GRS 1915+105 and Cir X-1 are the result of large maximum observational zenith angles. 
Hard cuts (image size ≥200 photoelectrons) tend to enhance the signal of sources with power-law spectral slopes that are harder in comparison to the dominant cosmic ray background. Standard cuts (image size ≥80 photoelectrons) provide less sensitivity in such cases but allow a lower energy threshold. No signifcant detection was obtained for anyof the three targets. Upper limits to the VHE γ-ray fux above the instrumental threshold energy were therefore derived at the 99% confdence level using the profle likelihood method(Rolke et al. 2005). 


3.1. GRS 1915+105 
As illustratedbythePCAandASMlight curvesshowninFig. 1, the X-ray count rate was stable to within ∼10% during each observation and varied by no more than ∼20% between observations. Indeed, the long-term RXTE ASM light curvein Fig. 1 (top panel) clearly indicates that the H.E.S.S. observation epochs occurred during anextended and relativelyfaint plateauin the 2−10keV fux. 
The3−200keV X-ray spectra shown in Fig. A.4 also exhibit remarkable stability between observations. The individual spectra are dominatedbyahard non-thermal component and strongly suggest class χ (in stateC) behaviour (e.g. Zdziarski et al. 2001; Trudolyubov 2001), which is confrmed by the location of the observations in the HID of Fig. A.1, according to the classifcation of Belloni et al. (2000). Figure A.2 shows the contextual X-ray and 15 GHz radio light curves of GRS 1915+105 during a two-month period that brackets the H.E.S.S. observation epochs. It is evident from the fgure that H.E.S.S. observed the target during andextended radio-loud plateau(∼80 mJy; for historical fux comparison, a three-year monitoring campaign is presented in Pooley2006). The plateau ended approximately ten days later with a combined radio and X-ray faring episode. The assertion of radio-loud behaviour at the H.E.S.S. observation epochsis supportedby the quasi-periodic oscillation (QPO) analysis presented in Fig. A.3. For a detailed discussion see Appendix A.1. 
In summary, the combined spectral and temporal analyses indicate a robust association of the contemporaneous H.E.S.S. 
observation with the radio-loud χ state, and the presence of steady, mildly relativistic jets at the time of observation may be confdently inferred. 
The contemporaneous H.E.S.S. observations did not yield a signifcant VHE γ-ray detection. The signifcances corresponding to the total H.E.S.S. exposure are computed using Eq. (17) fromLi&Ma (1983)and are listedinTable 3. Figure 1 plots runwise 99% confdence level upper limits to the integral VHE γ-ray fux abovethe instrumental threshold energy and illustrates the overlap between the RXTE and H.E.S.S. observations. Inte-gral fux upper limits, which correspond to the overall H.E.S.S. exposure, are listedinTable 4. 


3.2. Cir X-1 
The ASM light curve shown in Fig. 2reveals that the H.E.S.S. observation epochs occurred during an extended ∼4 day dip in the2−10keV X-ray fux. Additionally, it should be noted that the observations reported here were obtained during an extremely faint episode in the secular X-ray fux evolution of CirX-1(Parkinsonetal.2003), whichisalsoevidentfromthe HID presented in Fig. A.5. As a consequence, the measured X-ray fuxes are signifcantly lower than most others reported for this source.As illustratedin Fig. 2, the individual PCA light curves obtained during the frst two pointings are characterised by a relatively low count rate, which remains approximately constant throughout each observation. In marked contrast, the third observation exhibits clear variability with count rates doubling on timescales of ∼50 s. 
A detailed analysis of the obtained spectra (see Appendix A.2)reveals that the observed fux variability is accompanied by marked variations in spectral shape. These can be in-terpreted as hinting towards a strong mass transfer during the periastron passage and subsequent dramatic evolution of the lo-cal radiative environment. 
H.E.S.S. observations obtained contemporaneously with the RXTE pointings yield a non-detection that is evident from the signifcances listed in Table 3. Figure 2 plots runwise 99% confdence level upper limits to the integral VHE γ-ray fux 
A10, page7of 22 


Fig. <BR>2. <BR>RXTE ASM and PCA light curves for Cir X-1 together with H.E.S.S. upper limits derived from individual ∼28 min runs using standard event selection cuts. The blue shaded bands on the ASM light curve indicate the extent of the H.E.S.S. observations, while on the H.E.S.S. upper limitplots similarbands illustratethe durationofthe contemporaneousPCA observations(OBS1 −3).The partitioningofOBS3into sub-intervals A−Dbased on2−20keV X-ray fux is illustrated in the bottom right panel. The plotted H.E.S.S. upper limits correspond to different threshold energies, and the vertical scale of each light curve has been optimised for the plotted data. 
above the instrumental threshold energy and illustrates complete overlap between the RXTE and H.E.S.S. observations. Integral fux upper limits, which correspond to the overall H.E.S.S. ex-posure, are listedinTable 4. 
3.3. V4641 Sgr 
Figure 3 shows RXTE PCA light curves derived from three pointed observations. The individual light curves indicate var-ious degrees of X-ray variability with the clearest evidence for faring visible as a sharp ∼5-fold count rate fuctuation during the frst observation. In marked contrast, the second observation is uniformly faint with the χ2 probability of constant count rate Pconst = 0.97 and, hence, this observation is consistent with a period of steady, low-level emission. Subsequently, the third observation reveals a reemergence of mild variability (Pconst = 0.07) with ∼2-fold count rate fuctuations occurring on timescales of ∼500 s. 
Radio data shown in Fig. A.7 right were obtained using the VLA and AustraliaTelescope Compact Array(ATCA) between MJD 53 190 and MJD 53 208. They indicate rapid variability with peak fux densities of ∼30 mJy observed on MJD53193(Rupenetal. 2004b; Senkbeil&Sault 2004; Rupen et al. 2004a). An optically thin radio spectrum(S ν ∝ ν−0.7)observed on MJD 53191 was interpreted byRupen et al. (2004b)as the signature of a decaying radio fare. Radio ob-servations were triggered by an optical alert from VSNET (MJD 53 190) in combination with a RXTE PCA measurement during a Galactic bulge scan (MJD 53 189) that revealed a 2−10 keV X-ray fux equivalent to 8.2 mCrab. For compari-son, the August 2003 fare of V4641 Sgr reached 66 mCrab, while quiescent fuxes are typically <0.5 mCrab(Swank 2004). As shown in Fig. A.7 right, the dedicated RXTE PCA observation and H.E.S.S. observations took place between two radio fares, which is consistent with the X-ray variability evolution illustratedin Fig. 3. 
While V4641 Sgr is evidently the most X-ray-faint binary in the studied sample, it simultaneously exhibits the hardest spectrum,asshownbythe hardnessvaluesinFig. A.7 (left-hand panel). Furthermore, the evolution of the hardness is consistent with contemporaneous observations of rapid fux evolution in the radio band(Senkbeil&Sault 2004)3 (Fig. A.7).To help place the H.E.S.S. and RXTE observations in a historical context, the HID for V4641 Sgr in Fig. A.8 displays the entire 
3 http://www.ph.unimelb.edu.au/~rsault/astro/v4641/ <BR>
A10, page8of 22 

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars 

Fig. <BR>3. <BR>RXTE ASM and PCA light curves for V4641 Sgr together with H.E.S.S. upper limits derived from individual ∼28 min runs using standard event selection cuts. The blue shaded bands on the ASM light curve indicate the extent of the H.E.S.S. observations, while on the H.E.S.S. upper limit plots theyillustrate the duration of the contemporaneous PCA observations. The plotted H.E.S.S. upper limits correspond to different threshold energies and the vertical scale of each light curve has been optimised for the plotted data. 
archival RXTE PCA data set for this target, and compares the X-ray characteristics corresponding to the H.E.S.S. observation periods (different symbols are used to indicate observations obtained on each day in the range MJD 53 193-5) with three faring episodes observed with RXTE. On 15 September 1999 (orange markers in Fig. A.8), a 1500 s RXTE observation revealed a source fux evolution that is characterised by rapid, large-amplitude variability before reverting to a soft, low intensity state after ∼1000 s. An optical fare that preceded the RXTE observations likely corresponds with the onset of the short 10-hour outburst, which Wijnands&van der Klis (2000) associated with a low M ˙accretion event. Historically, faring episodes exhibited by V4641 Sgr are often short, unpredictable, and relativelyfaint, which implies that manymaygo unnoticed. 
Data corresponding to twolonger outbursts, spanning the pe-riods 24−26May 2002and5−7August 2003, are also illustrated in Fig. A.8. In the coordinates of the HID, both episodes are topologically similar to the 1999 outburst, but shifted towards fainter harder regions. 
Evidently,the X-ray fuxes that correspond with H.E.S.S. ob-servation epochs indicated in Fig. A.8 are substantiallyfainter than anyof these historically remarkable outbursts. 
In summary, in view of the various multiwavelength data, it seems likely that V4641 Sgr underwent a period of mild activity that spanned the H.E.S.S. observation epochs. 
The contemporaneous H.E.S.S. data are consistent with non-detection with the corresponding γ-ray signifcances listed in Table 3. Technical issues prevented γ-ray data corresponding to the frst RXTE observation from being obtained. Simultaneous γ-ray observations were obtained corresponding to the second RXTE exposure, which showed no indications of X-ray variability. Although the source began to show increased X-ray activity during the third RXTE observation, the degree of overlap with the corresponding H.E.S.S. observations was minimal. At radio, optical, and X-ray energies, V4641 Sgr exhibits rapid variability on timescales ∼10 min or less 
(e.g. Uemura et al. 2005;Maitra&Bailyn 2006).Optimistically, the compelling evidence for mild broadband faring admits the possibility that the H.E.S.S. observations monitoratransient out-burst event. 
Integral fux upper limits above the instrumental threshold energy, which correspond to theoverall H.E.S.S.exposure at the positionof V4641Sgr, are listedinTable 4. 
A10, page9of 22 Table 5. Estimated maximum VHE γ-ray luminosities of the target microquasars, which would still be consistent with a non-detection given the fux upper limits presentedinTable 4. 

Target Maximum distance estimate Ethresh Luminosity above Ethresh [kpc] [GeV] [erg s−1] 
GRS 1915+105 10.6 562 <2.3× 1034 Cir X-1 10.2 562 <3.4× 1034 V4641 Sgr 6.9 237 <2.5× 1034 
Notes. Source distances correspond to the largest estimate that was found in the literature (see Sect. 2). The energy threshold of Cherenkov telescope arrays increases with observational zenith angle. 
4. Discussion 
The principalaimofthisinvestigationwasto obtain contemporaneous X-ray and VHE γ-ray observations of three known super-luminal microquasars during major faring events. However, the results presentedin Sect. 3indicate that the interpretationof the VHE γ-ray non-detections cannot proceed under the assumption of energetic faring or bulk superluminal ejections at the time of observation. Nonetheless, upper limits to the VHE γ-ray fux were derived and an analysis of the contemporaneous RXTE ob-servations has helped to reveal the X-ray behaviour corresponding to the H.E.S.S. observation epochs. These datafacilitate the straightforwards derivation of constraints on the γ-ray luminosityof the target binary systems.InTable 5 the calculated fux upper limits were used to infer the maximum γ-ray luminosities above the target-specifc, instrumental threshold energy for each target binary system by assuming the maximum source distance estimate presentedinTable 1. 

Analysis of the contemporaneous X-ray and radio observa<BR>tions conclusively places GRS 1915+105ina radio-loud plateau state at the time of observation. In contrast with the superlumi<BR>nal faring episodes, this state is characterised by the production of continuous, mildly relativistic radio jets with an estimated power of ∼3 × <BR>1038 ergs−1(Klein-Wolt et al. (2002), assum<BR>ing a distance of 11 kpc). Theoretically, it seems unlikely that bright VHE γ-ray emission would be expected from the com<BR>pact self-absorbed jets, which are typical of the plateau state of GRS 1915+105. For example, a leptonic emission model de-veloped by Bosch-Ramon et al. (2006)to simulate the broad<BR>band emission of microquasar jets in the low-hard state predicts VHE γ-ray luminosities <∼1033 ergs−1 that are consistent with the H.E.S.S. non-detection. Notwithstanding the plausibility of VHE γ-ray emission in the plateau state, a comparison of the estimated jet power with the maximum γ-ray luminosity listed inTable 5reveals thatthejetpower conversionefficiencyis con-strained to be <∼0.008% for γ-ray production above562 GeV.For comparison, corresponding efficiency estimates for γ-ray pro-duction were derived for the steady, compact jets of other micro-quasars that were observed in appropriate states. The published MAGIC upper limit on the VHE γ-ray luminosity of Cygnus X-3 during its hard state implies a somewhat larger maximum con-version efficiency of 0.07%(Aleksi´c et al. 2010)and a similar value is obtained from MAGIC upper limits on the steady VHE emissionfromCygnusX-1(Albertetal.2007). Theseefficien<BR>cies are inferred from the directly observed jet power,and should be distinguished from the higher jet powers that were indirectly derived from the observation of radio-emittingbubbles infated by microquasar jets (see e.g. Gallo et al. 2005 for Cyg X-1; and Pakull et al. 2010;Soria et al. 2010 for S26 in NGC 7793). We presented an analysis of the entire H.E.S.S. data set for GRS 1915+105(Acero et al. 2009)and we derived an upper 
−2−1 
limit to the γ-ray fux above 0.41TeV of6.1× 10−13cms, corresponding to a detector livetime of 24.1 h. The somewhat higher upper limits presented in Sect. 3.1 utilise a more limited data set and are therefore consistent with the previously published value. None of the H.E.S.S. observations of GRS 1915+105 coincide with bright faring episodes at longer wavelengths. 
Observations of Cir X-1 were obtained during an extended dip in the X-ray fux, at phase intervals close to the periastron passage of the binary components. Spectral analysis of the RXTE data showed some evidence for a recent increase in mass transfer, producing strong signatures of X-ray absorption. It was hoped that H.E.S.S. observations would coincide with one of the quasi-regular radio fares, which often accompanyperiastron passage in Cir X-1. 
The ephemeris of Nicolson(2007)predicts the onset of a radio fare ∼19−20 h before the frst RXTE observation. Unfortunately, despite the undoubted occurrence of quasi-periodic radio fares from Cir X-1 near periastron, a robust correlation between the observed X-ray and radio behaviour is yet to be identifed. Although rapid brightening of the X-ray continuum might indicate accompanying radio fares, evidence for a defnitive associationisfarfromclear(Solerietal.2007;Tudoseetal. 2008). Recent radio observations of Cir X-1 (e.g. Fender et al. 2004a;Tudose et al. 2008)focus primarily on the ultrarelativis<BR>tic ejection events that manifest as >∼3 day episodes of faring on timescales of a few hours. In principle, the lack of contemporaneous radio data admits the possibility of such persistent outbursts at the time of observation. By analogy with canon-ical black hole binaries, it is possible that the inferred varia-tion in the mass accretion rate between the frst and second RXTE observations also implies an evolution of the jet prop-erties(Migliari&Fender 2006),but thisisfar from clearin such an unusual system. Moreover, Tudose et al.(2008)report com-pelling evidence that prior to 2006, Cir X-1 underwent a ∼6yr episode of unusual radio quiescence, suggesting that jet forma-tionwas somewhat suppressedduringtheepochsof H.E.S.S.observation. Accordingly, without strictly simultaneous radio data indicating otherwise, the most likely scenario is that no outfows were present. In this context the absence of a detectable γ-ray signal is not surprising. 
As a confrmed high-mass black hole binary, V4641 Sgr is the studied target that most closely resembles the Cygnus X-1 and Cygnus X-3 systems. Moreover, the H.E.S.S. observations were obtained during a period of sporadic broadband faring, and comparing these observations with the results of Albert et al. (2007), VHE γ-ray emission might have been expected. The detection of Cyg X-1 using the MAGIC telescopes appeared to coincide with the rising part of a strong X-ray fare. In contrast, radio spectra obtained close to the H.E.S.S. observational epochs are indicative of the decay following a faring episode(Senkbeil&Sault 2004). Assuming that the γ-ray emission mechanisms operating in Cyg X-1 also occur in V4641 Sgr, 
A10, page 10 of 22 

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars 
the absence of a signifcant H.E.S.S. detection might be viewed asevidence that productionof GeV andTeV photonsisa highly transient process. Thiswould further suggest that γ-ray emission originatesina spatially compactregionthatisatmostafewlight hours in size. 
Absorption of γ-rays by pair production is expected to be negligible in GRS 1915+105, since the donor star is too cool andfaintto producea strong ultraviolet photon feld.The same is true of Cir X-1 if the conventional assumption of a low-mass companionis adopted.For completeness,Fig. 4plotsthelevelof γ-ray absorption predictedbya numerical implementationofthe model presentedby Dubus (2006), assuming that the compan<BR>ion starinCirX-1isinfacta mid-B supergiantas proposedby Jonker et al. (2007).The separate curves are representativeof the three orbital phase intervals corresponding to the H.E.S.S. observation epochs, and were derived using the system parameters de-rivedby Jonkeretal. (2007)in conjunction with typicalvalues for the temperature(Teff ≈ 20000K) and radius(R ≈ 9R&#12;)of a mid-B supergiant. It is evident that some non-negligible absorptionisexpected, particularly duringthe frst observation interval. Nonetheless,it seems unlikely thattheexpectedlevelsof attenu-ation(<∼20%) would suppress an otherwise detectable γ-ray fux sufficientlytoyieldthelow signifcances listedinTable 3. 
The situation with regard to γ-ray absorption is clearer in the case of V4641 Sgr, since the companion has been spectroscopically identifedasalateB-orearlyA-typestar.Usingthesystem parametersderivedby Oroszetal. (2001)and assuminga circu<BR>larised orbit,the model presentedby Dubus (2006)was usedto predict the expected levels of γ-ray absorption as a function of orbital phase.As illustratedin Fig. 4(bottom panel), absorption mighthave an importanteffect during the frst H.E.S.S. observation interval, although as with Cir X-1 the predicted levels of ab-sorption(<∼25%) would not attenuate a bright γ-ray signal sofar below the detection threshold. During the second H.E.S.S. ob-servation interval, when X-ray data show marginal indications of source activity, the predicted absorption due to pair production on the stellar radiation feld is negligible. We note however that, as in the case of Cir X-1, the relative inclination of the jets from V4641 Sgr with respect to the accretion disk may be low(Schulz et al. 2006)and, therefore, further absorption of ∼100 GeV−TeVγ-ray photons could occur via interaction with the disk thermal photon feld (see e.g. Carraminana 1992). 
It should also be noted that all the confrmed VHE γ-ray bi-narieslieat distancesof2−4kpc. In contrast, the targets reported here have maximum distances in the range7−11 kpc, resulting in fux dilutionfactors that are greaterby one orderof mangnitude. Obviously,this has strong implications for the detectability of anyemitted γ-ray signal. 
5. Conclusions 
Contemporaneous VHE γ-ray and X-ray observations of GRS 1915+105, Cir X-1, and V4641 Sgr were obtained using H.E.S.S. and RXTE. Analysis of the resultant H.E.S.S. data did not yield a signifcant detection for any of the target mi-croquasars. However, X-ray binaries are dynamic systems and as such are likely to exhibit evolution of their radiative properties, both as a function of orbital phase and also in response to non-deterministic properties. It follows that the non-detections presented in this work do not indicate that the target binary systems do not emit detectable VHE γ-ray emission at phases other than those corresponding to the H.E.S.S. observations. 
GRS 1915+105 appears to have been observed during an extended plateau state, the archival multiwavelength data sug-

Fig. <BR>4. <BR>Levels of γ-ray absorption due to pair production with stel-lar photons as predicted by the numerical model outlined by Dubus (2006). Top panel: expected γ-ray transmission as a function of pho-ton energy for Cir X-1 assuming an inclination i = 66◦ and using the best-ft ephemeris derived by Jonker et al. (2007), which is appropri<BR>ate for a mid-B supergiant companion. The individual curves correspond to the orbital phases of the frst (blue dashed), second (red dot-dashed), and third (black solid) H.E.S.S. observation intervals. Bottom panel: expected γ-ray transmission for V4641 Sgr as a function of orbital phase derived using the orbital solution of Orosz et al. (2001)and assuming a circularised orbit. The individual curves represent photon energies of 10 GeV (black solid), 1 TeV (blue dashed), and 10 TeV (red dot-dashed).Vertical lines indicate the frst (dot-dashed) and second (dashed) H.E.S.S. observation epochs. 
gesting the presence of continuous, mildly relativistic radio jets at the time of observation. The RXTE observations of Cir X1 yield data that are consistent with strongly varying obscura-tion of the X-ray source shortly after periastron passage, but these data are not indicative of bright faring during the H.E.S.S. observation epochs. Conversely,V4641 Sgr appears to havebeen observed during an episode of mild, transient faring, although rapid source variability, combined with the limited duration of the strictly simultaneous H.E.S.S. and RXTE exposure, complicates interpretation. 
Microquasars continue to be classifed as targets of opportunity for IACTs, requiring a rapid response to any external trigger to maximise the likelihood of obtaining a signifcant detection. These conditions are realised with the commissioning of the H.E.S.S. 28 m telescope, which aims to lower the en-ergy threshold from 100 GeV to about 30 GeV(Parsons et al. 2015;Holler et al. 2015a,b)while simultaneously enabling very rapid follow-up observations(Hofverbergetal. 2013).Toexploit these new opportunities and an increasing understanding of the behaviour of microquasars, the triggering strategies 
A10, page 11 of 22 

for TeV follow-up observations have evolved signifcantly in recent years. In the future, alternative observational strategies, including continuous monitoring of candidate microquasars in the VHE γ-ray band, may become possible using dedicated sub-arrays of the forthcoming Cherenkov Telescope Array (CTAConsortium 2011). Irrespectiveof the non-detections presented herein, the tantalising observations of Cygnus X-3 at GeV energies and Cygnus X-1 by the MAGIC telescope ensures that the motivations for observing microquasars using IACTs remain compelling. In-deed, by further constraining the γ-ray emission properties of microquasars, subsequent observations will inevitablyyield an enhanced understanding of astrophysical jet production on all physical scales. More optimistically, the detection of additional γ-ray-bright microquasarswould greatlyfacilitatea comprehensive characterisation of the particle acceleration and radiative emission mechanisms that operate in such systems. Acknowledgements. The support of the Namibian authorities and of the Universityof Namibiainfacilitatingthe constructionand operationof H.E.S.S.is gratefully acknowledged, asis the supportby the German Ministry for Education and Research (BMBF), the Max Planck Society, the German ResearchFoundation (DFG), the French Ministry for Research, the CNRS-IN2P3 and the Astroparticle Interdisciplinary Programme of the CNRS, the UK Science andTechnologyFacilities Council (STFC), the IPNP of the Charles University, the Czech ScienceFoundation, the Polish Ministry of Science and Higher Education, the South African Department of Science and Technology and National Research Foundation,andbytheUniversityof Namibia.We appreciatetheexcellentwork of the technical support staff in Berlin, Durham, Hamburg, Heidelberg,Palaiseau, Paris, Saclay,and in Namibia in the construction and operation of the equipment. 
References 
Abdo, A. A., et al. (Fermi LATCollaboration) 2009, Science, 326, 1512 Abramowski, A., Acero,F., Aharonian,F., et al. 2012, ApJ, 757, 158 Acciari,V. A., Aliu, E., Arlen,T., et al. 2009, Nature, 462, 770 Acero,F., Aharonian,F., Akhperjanian, A. G.,&H.E.S.S. Collaboration. 2009, 
A&A, 508, 1135 Aharonian,F.A.,&Atoyan,A.M. 1996, A&A,309,917 Aharonian,F., Akhperjanian,A., Beilicke,M.,etal.2003, A&A,403,L1 Aharonian,F., Akhperjanian,A.G., Bazer-Bachi,A.R.,etal. 2006a, A&A,457, 899 Aharonian,F., Akhperjanian,A.G., Bazer-Bachi,A.R.,etal. 2006b, A&A,449, 223 Aharonian,F., Akhperjanian,A.G., Bazer-Bachi,A.R.,etal. 2006c, A&A,460, 743 Aharonian,F., Akhperjanian,A.G.,Anton,G.,etal.2009, ApJ,695,L40 Albert, J., Aliu, E., Anderhub, H., et al. 2007, ApJ, 665, L51 Aleksi´c,J., Antonelli,L.A., Antoranz,P.,etal.2010, ApJ,721,843 Aleksi´
c, J., Alvarez, E. A., Antonelli, L. A., et al. 2012, A&A, 539, L2 Arnaud, K. A. 1996, in Astronomical Data Analysis Software and SystemsV, eds.G.H. Jacoby,&J. Barnes, ASP Conf. Ser., 101,17 Atoyan,A.M.,&Aharonian,F.A.1999, MNRAS,302,253 Bednarek,W. 2010, MNRAS, 406, 689 Belloni,T., Mendez,M.,King,A.R.,vanderKlis,M.,& vanParadijs,J.1997, 
ApJ, 488, L109 
Belloni,T., Klein-Wolt,M., Méndez,M.,van der Klis,M.,& vanParadijs,J. 2000, A&A, 355, 271 Belloni,T., Psaltis,D.,&vanderKlis,M.2002, ApJ,572,392 Berge,D.,Funk,S.,&Hinton,J.2007, A&A,466,1219 Bosch-Ramon,V., Romero,G.E.,&Paredes,J.M.2006, A&A,447,263 Btcher,M.,&Dermer,C.D.2005, ApJ,634,L81 Carraminana, A. 1992, A&A, 264, 127 Castro-Tirado, A. J., Brandt, S., Lund, N., et al. 1994, ApJS, 92, 469 Clark,J.S., Charles,P.A., Clarkson,W.I.,&Coe,M.J.2003, A&A,400,655 Coppi,P.S. 2000, BAAS,32, 1217 Corbel,S.,Dubus,G.,Tomsick,J.A.,etal.2012, MNRAS,421,2947 CTAConsortium. 2011, Exp. Astron., 32, 193 Dermer,C.D.,&Btcher,M.2006, ApJ,643,1081 Done,C., Gierli´nski,M.,&Kubota,A. 2007, A&ARv,15,1 Dubus,G. 2006, A&A, 451,9 Dubus,G., Cerutti,B.,&Henri,G.2010, MNRAS,404,L55 Eikenberry, S. S., Matthews, K., Morgan, E. H., Remillard, R. A., & Nelson, R.W. 1998, ApJ, 494, L61 Falcke, H., Kding, E.,&Markoff, S. 2004, A&A, 414, 895 
Fender, R.,&Belloni,T. 2004, ARA&A, 42, 317 
Fender,R.P.,Pooley,G.G., Brocksopp,C.,&Newell,S.J.1997, MNRAS,290, L65 
Fender, R., Spencer, R., Tzioumis,T., et al. 1998, ApJ, 506, L121 
Fender, R.,Wu, K., Johnston, H., et al. 2004a, Nature, 427, 222 
Fender,R.P., Belloni,T.M.,&Gallo,E.2004b, MNRAS,355,1105 
Gallo, E., Fender, R., Kaiser, C., et al. 2005, Nature, 436, 819 
Georganopoulos,M., Aharonian,F.A.,&Kirk,J.G.2002, A&A,388,L25 
Ghisellini,G.,&Maraschi,L.1989, ApJ,340,181 
Gierli´nski,M.,Done,C.,&Page,K.2009, MNRAS,392,1106 
Gould,R.J.,&Schréder,G.P.1967, Phys.Rev.,155,1404 
Greiner, J., Cuby, J. G., McCaughrean, M. J., Castro-Tirado, A. J., & Mennickent, R. E. 2001, A&A, 373, L37 
Heinz, S., Burton, M., Braiding, C., et al. 2015, ApJ, 806, 265 
Hjellming,R.M.2000, The Astronomer’sTelegram,61,1 
Hjellming,R.M.,Rupen,M.P., Hunstead,R.W.,etal.2000, ApJ,544,977 
Hofverberg, P., Kankanyan, R., Panter, M., et al. 2013, in Proc. 33rd ICRC [arXiv:1307.4550] 
Holler,M., Balzer,A., Chalmé-Calvet,R.,de Naurois,M.,&Zaborov,D. 2015a, PoS(ICRC2015)980 
Holler, M., Berge, D., van Eldik, C., et al. 2015b, PoS(ICRC2015)847 
Homan, J.,&Belloni,T. 2005, Ap&SS, 300, 107 
Iaria,R.,DiSalvo,T., Burderi,L.,&Robba,N.R.2001, ApJ,561,321 
in’tZand,J.,Heise,J., Bazzano,A.,etal.1999, IAUCirc.,7119,1 
Johnston,H.M., Fender,R.,&Wu,K.1999, MNRAS,308,415 
Jonker,P.G., Nelemans,G.,&Bassa,C.G.2007, MNRAS,374,999 
Jourdain,E.,Roques,J.P., Chauvin,M.,&Clark,D.J.2012, ApJ,761,27 
Kalberla,P.M.W., Burton,W.B., Hartmann,D.,etal.2005, A&A,440,775 
Klein-Wolt,M., Fender,R.P.,Pooley,G.G.,etal.2002, MNRAS,331,745 
Kigl, A. 1981, ApJ, 243, 700 
Laurent,P., Rodriguez, J.,Wilms, J., et al. 2011, Science, 332, 438 
Leahy,D.A., Darbro,W., Elsner,R.F.,etal.1983, ApJ,266,160 
Lewin,W.H.G.,vanParadijs,J.,& vander Klis,M. 1988, Space Sci.Rev.,46, 273 
Li,T.-P.,&Ma,Y.-Q. 1983, ApJ, 272, 317 
Linares, M.,Watts, A., Altamirano, D., et al. 2010, ApJ, 719, L84 
Lindstr, C., Griffin, J., Kiss, L. L., et al. 2005, MNRAS, 363, 882 
MacDonald,R.K.D., Bailyn,C.D., Buxton,M.,etal.2014, ApJ,784,2 
Maitra,D.,&Bailyn,C.D.2006, ApJ,637,992 
Malzac, J., Lubi ´
nski,P., Zdziarski,A.A.,etal.2008, A&A,492,527 
Margon, B. 1984, ARA&A, 22, 507 
Margon,B., Lampton,M.,Bowyer,S.,&Cruddace,R.1971, ApJ,169,L23 
Markoff,S.,Nowak,M.A.,&Wilms,J. 2005, ApJ,635, 1203 
Markwardt,C.B.,Swank,J.H.,&Morgan,E.H. 1999a, IAUCirc.,7257,2 
Markwardt,C.B.,Swank,J.H.,&Taam,R.E. 1999b, ApJ,513,L37 
McCollough,M.L., Finger,M.H.,&Woods,P.M. 1999, IAUCirc., 7257,1 
Merloni,A.,Heinz,S.,&di Matteo,T.2003, MNRAS,345,1057 
Migliari,S.,&Fender,R.P.2006, MNRAS,366,79 
Mirabel,I.F.,&Rodriguez,L.F. 1994, Nature, 371,46 
Moskalenko,I.V. 1995, Space Sci.Rev.,72,593 
Muno,M.P.,Morgan,E.H.,&Remillard,R.A.1999, ApJ,527,321 
Murdin,P., Jauncey,D.L., Lerche,I.,etal.1980, A&A,87,292 
Nicholas, B.,&Rowell, G. 2008, in AIP Conf. Ser. 1085, eds.F. A. Aharonian, 
W. Hofmann,&F. Rieger, 245 Nicolson,G.D. 2007, The Astronomer’sTelegram, 985,1 Oosterbroek, T., van der Klis, M., Kuulkers, E., van Paradijs, J., & Lewin, W. H. G. 1995,A&A, 297, 141 Orosz,J.A.,Kuulkers,E.,vander Klis,M.,etal. 2001, ApJ,555,489 Pakull,M.,Soria,R.,&Motch,C.2010,Nature,466,209 Parkinson,P.M.S.,Tournear,D.M., Bloom,E.D.,etal.2003,ApJ,595,333 Parsons,R.D., Gajdus,M.,&Murach,T. 2015, PoS(ICRC2015)826 Pooley,G. 2006,inVI MicroquasarWorkshop: Microquasars andBeyond,9 Protheroe,R.J.,&Stanev,T.1987, ApJ,322,838 Reid,M.J., McClintock,J.E., Steiner,J.F.,etal.2014, ApJ,796,2 Revnivtsev,M.,Gilfanov,M.,&Churazov,E.2000, A&A,363,1013 Revnivtsev, M., Gilfanov, M., Churazov, E.,& Sunyaev, R. 2002, A&A, 391, 1013 Revnivtsev, M., Khamitov, I., Burenin, R., et al. 2004, The Astronomer’s Telegram, 297,1 Rolke,W.A., Lez,A.M.,&Conrad,J. 2005, Nucl. Instrum. Meth.Phys. Res. A, 551, 493 
Romero, G. E.,Torres, D.F., Kaufman Bernad M. M.,&Mirabel, I.F. 2003, A&A, 410, L1 
Romero,G.E., Christiansen,H.R.,&Orellana,M.2005, ApJ,632,1093 
Romero,G.E.,Vieyro,F.L.,&Chaty,S. 2014, A&A,562,L7 
Rupen,M.P., Dhawan,V.,&Mioduszewski,A.J. 2002, IAUCirc., 7928,2 
Rupen, M. P., Mioduszewski, A. J., & Dhawan, V. 2003, The Astronomer’s Telegram, 172,1 
A10, page 12 of 22 

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars 
Rupen, M. P., Dhawan, V., & Mioduszewski, A. J. 2004a, The Astronomer’s Telegram, 303,1 
Rupen, M. P., Mioduszewski, A. J., & Dhawan, V. 2004b, The Astronomer’s Telegram, 296,1 
Russell,D.M., Fender,R.P.,Hynes,R.I.,etal.2006, MNRAS,371,1334 
Saito, T. Y., Zanin, R., Bordas, P., et al. 2009, in 31st ICRC (Ł´z) [arXiv:0907.1017] 
Schulz, N. S., Brandt, W. N., Galloway, D. K., Chakrabarty, D., & Heinz, S. 2006,in The X-ray Universe 2005, ed.A.Wilson, ESASP, 604, 201 
Schulz,N.S., Kallman,T.E.,Galloway,D.K.,&Brandt,W.N.2008, ApJ,672, 1091 
Senkbeil,C.,&Sault,B.2004, The Astronomer’sTelegram,302,1 
Shirey,R.E., Bradt,H.V.,&Levine,A.M. 1999a, ApJ,517,472 
Shirey,R.E.,Levine,A.M.,&Bradt,H.V. 1999b, ApJ,524,1048 
Soleri, P., Belloni, T., & Casella, P. 2006, in VI Microquasar Workshop: Microquasars and Beyond, 43 
Soleri,P.,Tudose,V., Fender,R.P.,& van der Klis,M. 2007,in Proc. Bursts, Pulses and Flickering: wide-feld monitoring of the dynamic radio sky; 12−15 June,Kerastari,Tripolis, Greece,37 
Soria,R.,Pakull,M., Broderick,J.,&Corbel,S.2010, MNRAS,409,541 
Steeghs,D., McClintock,J.E.,Parsons,S.G.,etal.2013, ApJ,768,185 
Stubbings,R.,&Pearce,A. 1999, IAUCirc., 7253,1 
Swank,J. 2004, The Astronomer’sTelegram, 295,1 
Tavani, M., Bulgarelli, A., Piano, G., et al. 2009,Nature, 462, 620 
Tennant,A.F.,Fabian,A.C.,&Shafer,R.A. 1986a,MNRAS,221,27P 
Tennant,A.F.,Fabian,A.C.,&Shafer,R.A. 1986b,MNRAS,219,871 
Trudolyubov,S.P. 2001,ApJ, 558, 276 
Tudose,V., Fender,R.P., Kaiser,C.R.,etal.2006,MNRAS,372,417 
Tudose,V., Fender,R.P., Tzioumis,A.K., Spencer,R.E.,&vander Klis,M. 2008, MNRAS, 390, 447 
Uemura, M., Mennickent, R., Stubbings, R., et al. 2005, Information Bulletin on Variable Stars, 5626,1 
Vadawale,S.V.,Rao,A.R.,&Chakrabarti,S.K.2001,A&A,372,793 
van Oers,P.,&Markoff, S. 2010, in Jets at all Scales, Proc. IAU, 6, 294 
vanParadijs,J.,&McClintock,J.E.1994, A&A,290,133 
Wijnands,R.,&vander Klis,M. 2000, ApJ,528,L93 
Zdziarski, A. A. 2014, MNRAS, 444, 1113 
Zdziarski,A.A.,Grove,J.E., Poutanen,J.,Rao,A.R.,&Vadawale,S.V.2001, ApJ, 554, L45 
Zdziarski, A. A., Lubinski,P., Gilfanov, M.,&Revnivtsev, M. 2003, MNRAS, 342, 355 
Zdziarski, A. A., Sikora, M., Dubus, G., et al. 2012, MNRAS, 421, 2956 
Zhang,W., Giles, A. B., Jahoda, K., et al. 1993, in SPIE Conf. Ser. 2006, eds. 
O. H. Siegmund, 324 Ziolkowski,J. 2015, ArXiv e-prints[arXiv:1509.02819] 
1 Centre for Space Research, North-West University, Potchefstroom 2520, South Africa 2 Universität Hamburg, Institut f Experimentalphysik, Luruper Chaussee 149, 22761 Hamburg, Germany 3 Max-Planck-Institut fKernphysik,PO Box 103980, 69029 Heidelberg, Germany 4 Dublin Institute for Advanced Studies, 31 Fitzwilliam Place, Dublin 2, Ireland 5 National Academy of Sciences of the Republic of Armenia, Marshall BaghramianAvenue,24, 0019Yerevan, Republicof Armenia 6YerevanPhysics Institute,2Alikhanian Brothers St., 375036 Yerevan, Armenia 7 Institut f Physik, Humboldt-Universität zu Berlin, Newtonstr. 15, 12489 Berlin, Germany 8 University of Namibia, Department of Physics, Private Bag 13301, Windhoek, Namibia 
9 GRAPPA, AntonPannekoek Institute for Astronomy, University of Amsterdam, SciencePark 904, 1098XH Amsterdam, The Netherlands 
10 Department of Physics and Electrical Engineering, Linnaeus University, 351 95 Växj Sweden 
11 Institutf TheoretischePhysik, LehrstuhlIV:Weltraumund Astro-physik, Ruhr-Universität Bochum, 44780 Bochum, Germany 
12 GRAPPA,AntonPannekoek Institutefor Astronomyand Instituteof High-EnergyPhysics,Universityof Amsterdam, SciencePark904, 1098 XH Amsterdam, The Netherlands 
13 Institut f Astro-und Teilchenphysik, Leopold-Franzens-Universität Innsbruck, 6020 Innsbruck, Austria 
14 Schoolof Chemistry&Physics, Universityof Adelaide, 5005 Adelaide, Australia 
15 LUTH, Observatoire de Paris, PSL Research University, CNRS, Université Paris Diderot, 5 Place Jules Janssen, 92190 Meudon, France 
16 Sorbonne Universités, UPMC UniversitéParis 06, UniversitéParis Diderot, SorbonneParis Cité, CNRS, Laboratoire de Physique Nu-cléaireetde Hautes Energies (LPNHE),4place Jussieu, 75252Paris Cedex 5, France 
17 LaboratoireUniversetParticulesde Montpellier,Université Montpellier, CNRS/IN2P3, CC 72, Place Eugène Bataillon, 34095 Montpellier Cedex 5, France 
18 DSM/Irfu, CEA Saclay, 91191 Gif-Sur-Yvette Cedex, France 
19 Astronomical Observatory, The UniversityofWarsaw, Al. Ujazdowskie4, 00-478Warsaw, Poland 
20 Aix Marseille Université, CNRS/IN2P3, CPPM UMR 7346, 13288 Marseille, France 
21 Instytut FizykiJa¸drowejPAN, ul. Radzikowskiego 152, 31-342 Krak, Poland 
22 Funded by EU FP7 Marie Curie, grant agreement No. PIEF-GA2012-332350, 
23 School of Physics, University of the Witwatersrand, 1 Jan Smuts Avenue, Braamfontein, 2050 Johannesburg, South Africa 
24 Laboratoire d’Annecy-le-VieuxdePhysique desParticules, Université Savoie Mont-Blanc, CNRS/IN2P3, 74941 Annecy-le-Vieux, France 
25 Landessternwarte, Universität Heidelberg, Kigstuhl, 69117 Heidelberg, Germany 
26 Université Bordeaux, CNRS/IN2P3, Centre d’Études Nucléaires de Bordeaux Gradignan, 33175 Gradignan, France 
27 Oskar Klein Centre, Department of Physics, Stockholm University, Albanova University Center, 10691 Stockholm, Sweden 
28 WallenbergAcademy Fellow 
29 Institut f Astronomie und Astrophysik, Universität Tingen, Sand 1, 72076 Tingen, Germany 
30 Laboratoire Leprince-Ringuet, École Polytechnique, CNRS/IN2P3, 91128Palaiseau, France 
31 APC, AstroParticule et Cosmologie, Université Paris Diderot, CNRS/IN2P3, CEA/Irfu, Observatoire de Paris, Sorbonne Paris Cité,10, rue Alice Domonet Léonie Duquet, 75205Paris Cedex13, France 
32 Univ. Grenoble Alpes, IPAG; CNRS, IPAG, 38000 Grenoble, France 
33 Department of Physics and Astronomy, The University of Leicester, University Road, Leicester, LE1 7RH, UK 
34 Nicolaus Copernicus Astronomical Center, ul. Bartycka 18, 00-716Warsaw, Poland 
35 Institut f Physik und Astronomie, Universität Potsdam, Karl-Liebknecht-Strasse 24/25, 14476 Potsdam, Germany 
36 Universität Erlangen-Nnberg, Physikalisches Institut, Erwin-Rommel-Str. 1, 91058 Erlangen, Germany 
37 DESY, 15738 Zeuthen, Germany 
38 Obserwatorium Astronomiczne, Uniwersytet Jagiello´nski, ul. Orla 171, 30-244 Krak, Poland 
39 Centrefor Astronomy,FacultyofPhysics, Astronomyand Informatics, Nicolaus Copernicus University, Grudziadzka5, 87-100Torun, Poland 
40 Department of Physics, University of the Free State, PO Box 339, 9300 Bloemfontein, South Africa 
41 HeisenbergFellow (DFG), ITAUniversität Heidelberg, 069120 Heidelberg, Germany 
42 GRAPPA, Institute of High-Energy Physics, University of Amsterdam, SciencePark 904, 1098XH Amsterdam, The Netherlands 
43 University of Durham, Department of Physics, South Road, Durham DH1 3LE, UK 
44 Iowa State University, Ames, IA 50011, USA 
A10, page 13 of 22 

Appendix A: Modelling and determination of the system X-raystates 


A.1. GRS 1915+105 
Figure A.1 shows the HID derived from the entire archival RXTE PCA data set for GRS1915+105. Hardness is defned as theratiooffuxes measuredinthe2−9keV and9−20keV bands, while intensity is defned as the sum of the band-limited fuxes in units of counts per second. The background has been subtracted and the light curves are sampled in 16 s intervals. These defnitions are used consistently for the three HIDs presented in this paper. In the case of GRS 1915+105, the H.E.S.S. observation took place in a low-hard state (LHS; symbols for MJD 53 1238 in top-left sector of Fig. A.1), in which compact jets are ex-pected to be present and characterised by a potentially radio-loud χ X-rayvariability class (belongingto stateCfollowingthe classifcationof Bellonietal.2000).For comparison,the orange points in Fig. A.1 correspond with data obtained on 17 Decem-ber 1997. These data were studiedby Solerietal. (2006)who associated them with a hard-intermediate state (HIMS) to soft-intermediate state (SIMS) transition. 
The power density spectra shown in Fig. A.3 show the presence of low frequency QPOs: following the approach of Belloni et al. (2002), a Lorentzian decomposition of the ob-served power spectra was performed using two broad continuum components and several narrower QPO peaks. Cru-
qcially, the characteristic frequency (νmax = ν02 +Δ2, see Belloni et al. 2002) of the higher frequency continuum com-ponent never exceeds ∼4 Hz during our observations. This is far below the characteristic cut-off frequencies associated with previous observations of the radio-quiet χ state (e.g. Trudolyubov 2001). Radio-quiet observations (Belloni et al. 1997;Trudolyubov 2001)exhibit signifcant band-limited white noise extending to high frequencies f ∼ 60−80 Hz, while in radio-loud case such noise is either absent or exhibits an exponential cut-off at ∼15 Hz(Trudolyubov 2001).Consequently,the absence of band limited noise at high frequencies is consistent with a radio-loud state C. 
Figure A.4 illustrates the spectral analysis performed for each GRS 1915+105 RXTE observation. The simple model de-scribedby Vadawale et al.(2001)has been adopted:acontinuum model comprising a disk black-body (DiskBB4)component,a hybrid thermal and a non-thermal Comptonisation component (CompST), and a separate power law (Powerlaw)to model the high-energy emission. Interstellar absorption was modelled using the Wabs model in XSPEC with the equivalent hydrogen column fxedtoavalueof6 × 1022 cm−2(Belloni et al. 1997; Markwardt et al. 1999b; Muno et al. 1999). A constant multiplicativefactorwas introducedto accountforthe normalisation of HEXTE relativetothe PCA.The additionofaGaussian com-ponent with centroid energy ELinefxedat6.4keVwas foundto signifcantly improve the resultant model ft. 
As demonstrated by the reduced χ2 values listed in TableA.1, the ftted model provides an adequate description of the RXTE data, and the derived model parameters correspond closelyto those obtainedby Vadawaleetal. (2001)during ear<BR>lier episodes of radio-loud χ-state behaviour, supporting the at-tribution of this state to the epochs of H.E.S.S. observation. However, insufficient event statistics prevent the inference of robust conclusions regarding the origin of the X-rays that were 
http://heasarc.gsfc.nasa.gov/xanadu/xspec/manual/ <BR>XspecModels.html <BR>
A10, page 14 of 22 
observedinthisstudy.Amore sophisticatedandphysicallywellmotivated model, describing an X-ray corona (EQPAIR from Coppi 2000)gives a similar goodness-of-ft,after accounting for a larger parameter count. Physically, the power-law com-ponent could result from Comptonisation by energetic electrons in a corona or could be generated by synchrotron radiation at the base of a jet. The former scenario is discussed in a number of papers (e.g. Zdziarski et al. 2003), while the latter was studied by van Oers&Markoff <BR>2010 in the context of a simi-lar plateau state of GRS1915+105. They applied a leptonic jet model(Markoff <BR>et al. 2005)to X-ray, IR, and radio data. Al-though their model provided statistically convincing broadband fts, this was only achievable when adopting extreme parameter values. The power lawwithΓ ' 2.7thatwas derivedin this study (Table A.1)cannot be extrapolated down to UV band without generating an inconsistency in subsequently inferred bolometric luminosity. It should be interpreted as a phenomenological approximation of a high-energy tail, which itself might only be partially accounted for by SSC radiation. 
A similar plateau state of GRS 1915+105 (October 1997) was studiedby Klein-Woltetal. (2002),who attributedtheob<BR>served radio emission to quasi-continuous ejecta forming the compact jet. 
In Fig. A.2, the 15 GHz radio surface brightness and the X-ray hardness and intensity corresponding with the epochs of 
H.E.S.S. observation are illustrated in a broader historical context(Pooley2006)5. The H.E.S.S. observations started approxi<BR>matelyoneweek beforetheendofalong radio-loudplateauand were triggered by a transient dip in radio fux whilst the plateau end was not yet reached. The plateau ended about two days after the last H.E.S.S. observation, followedbyaradio and X-ray fare ten days later. 
In summary,GRS 1915+105 remainedinaradio-loud χ state with steady, mildly relativistic jets at the time of H.E.S.S. observations without clear signs of a state transition. 


A.2. Cir X-1 
The HID for Cir X-1 is shown in Fig. A.5. An extensive study, examining ten days out of the 16.55 orbital period was performed by Shireyet al. (1999a) in 1997; the corresponding RXTE data are indicated with orange symbols in Fig. A.5. The study focused on the toroidally distributed data plotted in the lower right part of the HID. Shireyet al. (1999a), stud<BR>ied the spectral and temporal X-ray evolution of Cir X-1 along three distinct branches (horizontal, normal, faring) in the HID. This evolution occurred during a half-day period, approximately one day after periastron and was repeated few days later. Such behaviour is typical of a “Z source”. 
In Shireyet al. (1999a), the periastron passage corresponds to the data at low fux and hardness (dipping episode, lowerleft part of the cycle). Data contemporaneous to the H.E.S.S. observations (MJD 53 174-6) are indicated by symbols and ex-hibit low X-ray intensity and hardness values. They span two days starting19hafter the periastron, which coincides with the orbital rangeexploredby Shireyetal. (1999a),butina much fainter X-ray luminosity context. 
Inspection of the3−20keV PCA spectra shown in Fig. A.6 reveals that the observed fux variability is accompanied by marked variations in spectral shape. For the third observation, individual spectra were extracted from the four regions (A to D) shown in Fig. 2 (bottom-right panel), segregated on 
5 http://www.mrao.cam.ac.uk/~guy/ <BR>

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars 

Table A.1. XSPEC model components and best-ft parameters for GRS 1915+105. 
Component Parameter OBS1 OBS2 OBS3 
DiskBB Tin [keV] 1.695+0.42 1.137+0.61.698+0.63 
−0.59 −0.3 −0.77 
31.1+120 81.3+360 20.9+200 
DiskBB Norm 

−17 −78 −16 CompST kTe [keV] 4.195+25.057+1.14.460+0.54 
−0.54 −0.81 −0.7 CompST τ 13.145−6.97.801+3.7 10.959−4.7 
−1.8 Powerlaw 2.714+0.058 2.668+0.12 2.644+0.077 
Γphot −0.86 −0.15 −0.82 Gaussian ELine [keV] 6.4* Gaussian σ [keV] 0.776+0.24 0.891+0.15 0.730+0.22 
−0.29 −0.17 −0.24 
Gaussian W[eV] 66.44 87.76 70.44 
χ2 (NDF) 0.82 (77) 0.96 (78) 0.99 (80) 
ν 

Component Parameter OBS4 OBS5 OBS6 
DiskBB Tin [keV] 1.350+0.63 1.415+0.63 1.713+0.51 
−0.42 −0.51 −0.55 
30.8+140 23.3+110 
DiskBB Norm 15.4+28 
−28 −23 −13 CompST kTe [keV] 5.130+0.77 4.888+0.73 5.118+0.83 
−0.69 −0.69 −0.82 CompST τ 8.516+4.48.870+6.19.028+14 
−1.6 −1.7 −1.9 Powerlaw 2.503+0.12 2.521+0.026 2.441+0.13 
Γphot −0.38 −0.27 −0.37 Gaussian ELine [keV] 6.4* Gaussian σ [keV] 0.894+0.16 0.902+0.16 0.928+0.19 
−0.19 −0.18 −0.24 
Gaussian W[eV] 91.15 91.78 86.93 
χ2 (NDF) 1.12 (80) 1.19 (83) 0.82 (80) 
ν 

Notes. As discussedinthetext,an additionalWabs component (with equivalenthydrogen column density fxed to NH = 6× 1022 cm−2)was used to model the effects of interstellar absorption. The parameter errors correspond to a Δχ2of 2.71. Frozen parameters are indicatedby*. 
the basis of average 2−20keV count rates. Fitting of the Previous observations of Cir X-1 during periastron dips (e.g. spectral data from the third observation employed a simi-Shireyet al. 1999b; Schulz et al. 2008)reveal the evidence of lar approach to that of Shireyet al. (1999b) with the unab-strong, complex, and variable intrinsic X-ray absorption. Con-sorbed continuum modelled using a disk black-body compo-sequently, diagnosis of the system behaviour during the third nent (DiskBB in XSPEC)at low energies in combination with RXTE observation is critically dependent upon whether the oba single temperature black body (Bbody)that dominates above served variability represents a genuine change in the underly-∼15keV. ingcontinuum emission or is simply an artefact ofvarying ab
A10, page 15 of 22 


Table A.2. Spectral parameters for Cir X-1 during OBS3corresponding to the orbital phase interval0.1718 ≤ φ ≤ 0.1725 (according to the ephemeris of Nicolson 2007). 
Component  Parameter  OBS 3A  OBS 3B  OBS 3C  OBS 3D  
DiskBB  Tin [keV] Norm  1.059+0.03 −0.04 (1.914+0.63 )× 102 −0.44 Joint ft  
Bbody Pcfabs  kT [keV] Norm NH [×1022] CvrFract  2.954+0.67 −0.47 (1.126+0.24 )× 10−3 −0.18(2.303+0.33 )× 101 −0.34(8.630+0.22 )× 10−1 −0.20 (9.373+2.09 )× 10−4 −1.70(2.880+0.32 )× 101 −0.32(8.411+0.19 )× 10−1 −0.20 Joint ft (7.646+2.01 )× 10−4 −1.62(4.606+0.29 )× 101 −0.29(8.726+0.15 )× 10−1 −0.16 (4.992+2.61 )× 10−4 −2.44(7.998+0.51 )× 101 −0.48(8.852+0.14 )× 10−1 −0.15 
χ2 (NDF) ν  1.27 (145)  Joint ft  
Model fux  −2[erg cms−1] 6.121 × 10−10  5.512 × 10−10  3.822 × 10−10  2.515 × 10−10  

Notes. XSPEC model components, best-ft parameters, and3−20keVmodelfuxesareshownforthefour separate sub-intervals illustratedinFig. 2 in order of decreasing model fux. As discussed in the text, an additionalWabs component (with equivalenthydrogen column density fxed to NH = 1.59 × 1022 cm−2)was used to model the effects of interstellar absorption. Jointly ftted parameters assume the values quoted for OBS 3A. The parameter errors correspond to a Δχ2 of 2.71. 
sorption. Accordingly, two components are used to separately estimate of the surrounding interstellar medium density used by simulate intrinsic and extrinsic X-ray absorption characteris-Tudose et al. (2006)to model the evolution of the radio nebula tics. The bipartite intrinsic absorption is treated using a partial of Cir X-1. covering model (Pcfabs), while a simple photoelectric model In order to constrain the origin of the observed spectral (Wabs)simulates the absorbingeffect of the interstellar medium. variability, a joint ft was performed using the complete best-
Adopting a weighted average of the neutral hydrogen data of ftting model. Initially, the continuum and extrinsic absorp-
Kalberla et al. (2005)calculated using thenH ftool,a fxedeffec-tion components (DiskBB, Bbody,Wabs) were constrained tivehydrogen column with NH = 1.59× 1022 cm−2 was assumed to be equal for all individual spectra, while the component for theWabs component. This assumption is consistent with an related to intrinsic absorption (Pcfabs) was allowed to vary 
A10, page 16 of 22 

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars 
Table A.3. XSPEC model components, best-ft parameters, and3−20keV model fuxes for Cir X-1 during OBS1and OBS2corresponding to the orbital phase intervals0.0486 ≤ φ ≤ 0.0498 and0.1104 ≤ φ ≤ 0.1112, respectively (according to the ephemeris of Nicolson 2007). 
Component  Parameter  OBS 1  OBS 2  
(0.0486 ≤ φ ≤ 0.0498)  (0.1104 ≤ φ ≤ 0.1112)  
DiskBB  Tin [keV]  1.355+0.18 −0.08  1.059*  
Norm  (3.798+3.42 )× 101 −3.64 (1.914) × 1012*  

Bbody kT [keV] – 2.465+0.50 
−0.39 

Norm – (7.577+1.90)× 10−4 
−1.22

Pcfabs  NH [×1022] CvrFract  (1.353+0.44 )× 102 −0.22(8.292+0.87 )× 10−1 −2.01 (9.545+0.25 )× 101 −0.22(9.191+0.01 )× 10−1 −0.01 
Gaussian  ELine [keV] σ [keV] Norm  6.696+0.09 −0.08 0.1† (1.435+0.36 )× 10−3 −0.33 – – –  
χ2 (NDF) ν  1.07 (34)  1.22 (36)  
Model fux  −2 [erg cms−1] 2.722 × 10−10  1.912 × 10−10  

Notes. As discussedinthetext,an additionalWabs component (with equivalenthydrogen column density fxed to NH = 1.59 × 1022 cm−2)was used to model the effectsof interstellar absorption.Parameters markedby* arefxedtothe best-ftting values fromthe third observation(see TableA.2). Thevalueof theGaussian σ parameter (markedbya † symbol) was also fxed. The parameter errors correspond to a Δχ2 of 2.71. 
independently. Although this model provides a reasonable ft to the observational data(χ2 = 1.38), allowing the normalisa-
ν 
tion of theBbody component to vary between observations im-proves the ft quality somewhat, yielding χ2 = 1.27. The pa
ν 
rameters that result from ftting this more relaxed model are listed in Table A.2. A statistical comparison of the alternative model fts using theFtest yieldsa ∼1% probability that the ob-served improvement in ft quality would be obtained even if the more restrictive modelwas correct.This marginalevidence for variation of theBbody component normalisation might indicate rapid fuctuations of the X-ray continuum above ∼10keV. However, the available data cannot exclude an alternative scenario in which apparent changes in the fttedBbody parameters arise purely from imperfect modelling of substantial variations in the intrinsic X-ray absorption with no requirement for genuine evolution of the underlying continuum. 
Table A.3 lists the parameters of the spectral fts obtained from the frst and second observations. A similar continuum modeltothat obtainedfromthe third observationalsoprovidesa good ft(χ2 = 1.22) to the spectrum obtained during the second 
ν 
observation. In contrast, the spectrum obtained during the frst observation is more appropriately described by a single, heavily absorbed disk black-body component with large correlated residuals around ∼6.5 keV statistically favouring the addition of a Gaussian line component. This continuum variability is consis-tent with the results of Shireyet al. (1999a)who found that sig<BR>nifcant variation of the continuum parameters could occur on timescales of a few hours. 
Overall, the RXTE data reinforce the accepted paradigm of enhanced mass transfer during the periastron passage of the compact primary with the strong and variable intrinsic absorption attributed to obscuration by a turbulent accretion fow (see e.g. Oosterbroek et al. 1995;Murdin et al. 1980;Iaria et al. 2001).A marked disparity between best-ftting model components and pa-rameters of the frst and second observations implies a dramatic evolution of the local radiative environment.A ∼30% decrease in continuum luminosity accompanied by a similar reduction of the intrinsic absorption column suggests a signifcant decrease in the mass transfer rate. Subsequent fuctuation in the inferred magnitude of the absorption column during the third observation is indicative of dispersion or reorganisation of the recently accreted material. 
A10, page 17 of 22 


A10, page 18 of 22 

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars 

A10, page 19 of 22 


A10, page 20 of 22 

H.E.S.S. Collaboration: H.E.S.S. observations of microquasars 

A10, page 21 of 22 



A10, page 22 of 22