Recent Constraints on Jet physics and Properties Obtained from High Energy Observations of Microquasars

TL;DR

This paper reviews recent high-energy observations of microquasars, exploring their emission mechanisms, media composition, and jet physics, highlighting new insights from X-ray and gamma-ray data.

Contribution

It provides a synthesis of recent observational results from XMM-Newton, INTEGRAL, and Fermi, offering new interpretations of microquasar jet emissions at high energies.

Findings

Detection of microquasars at energies up to TeV.

Insights into emission processes and media composition.

Correlation between accretion and high-energy emissions.

Abstract

The recent detections of microquasars at energies above a few hundred keV up to the TeV in one case has stimulated a strong interest and raised several questions. How are the MeV, GeV, and even TeV emissions, produced? What are the emission processes and in-fine the physics and media at the origin of the broad band spectra? What is the content of these media, and how are they powered? These sources are machines accelerating particles and matter to very high speed in collimated jets. While these are now known for more than twenty years (mainly with radio observations) and their behavior known to be tied to the accretion processes onto the compact object (studied via X-ray observations), their potential influence at high energy is just being recognized. In this review I will present a selection of recent results obtained with XMM-Newton, INTEGRAL, Fermi on 4U 1630$-$40, Cygnus X-1, and V404 Cygni and will discuss the possible interpretations of these results.

Full text

accretion, accretion disks — jets sources — black hole physics — X-rays: binaries — Gamma-rays: observations — stars: individual: Cyg X-1, V404 Cyg, 4U 1630-47

1 Introduction

Microquasars (MQs) are Galactic X-ray binaries showing episodes of ejections in certain phases of their activity. Most MQs host a black hole as compact object, and the majority are transient sources that spend most of their lives in a dormant, so-called quiescent state. They are detected (usually in X-rays) when they enter into month to year long periods of activity called outbursts. All MQs are multi-wavelength emitters. The radio to infrared (IR) emission is attributed to relativistic jets, either a persistent compact jet or in form of large scale discrete ejections. The X-rays probe the inner accretion flow(s): the accretion disc ($\sim$0.1–10 keV) and the so-called “corona” (10–200 keV) (e.g. Fender, 2006; Remillard & McClintock, 2006). The relative contributions of these two media seen in the X-ray spectra have led to the classical classification into spectral states. The two canonical ones are the “(high) soft state” (HSS) and the “(low) hard state” (LHS). In the former the X-ray spectrum is dominated by the thermal emission from the disc, and in the latter it is dominated by a (cut-off) power law with a hard (\Gamma$$<2) photon index attributed to inverse Comptonization (IC) of the soft disc photons by the coronal electrons. The past 25 years have allowed us to enrich this initial vision, and, in particular, to find strong connections between X-rays (accretion) and radio-IR (ejection) (e.g. Mirabel et al., 1998; Fender & Pooley, 1998; Corbel et al., 2000; Fender, 2001b; Corbel et al., 2003; Rodriguez et al., 2003, 2008; Corbel et al., 2013b, to cite but a few): a powerful compact jet is always present in the LHS, while transitions from the LHS to the HSS are accompanied by discrete ejections.

The picture may have recently been completed by the detection of a few sources at energies above 1 MeV with OSSE and INTEGRAL, and/or in the GeV domain with Fermi and Agile, and possibly in the TeV range with the Magic telescope. These detections pose questions regarding the media and physical processes at work, with answers that will certainly impact our understanding of particles acceleration, particle-matter and particle-particle interaction, and eventually will help us constrain feedback processes in the interstellar medium.

As mentioned above these sources are either transients, and they are anyway highly variable. One thus needs alerts from surveying instruments in order to be able to conduct multi-wavelength, multi-instrumental campaigns during the most relevant periods of their activity. Before presenting recent high energy observations and analysis of a few MQs, I will first start this review by presenting the obvious but necessary needs of all sky surveying/monitoring instruments without which most of the results of the past 30 years will not have been obtained.

2 Transient and variable sources need to be surveyed: the central role of all-sky monitors

MQs have unpredictable behavior. They enter into outburst and/or show (spectral) transitions at “random” moments. These are ideal times to probe the interactions and causal relations between the different media (jet-disc-corona-companion). It is thus necessary to keep an eye on the entire sky to detect new or know object as early as possible during their outbursts, or simply check for any spectral transition that may ’act’ as a smoking gun for major change of configuration of the inner flows and possible ejection events. Wide field monitoring instruments (hereafter ASM for all sky monitors of any types) are the only instruments to do so.

Fig. 1 shows the MAXI light curves covering a multi-year survey of GX 339$-$4 that, in many ways, can be considered as the prototype MQ (e.g. Corbel et al., 2013a). Over the 7 years of X-ray coverage, one can clearly distinguish three outbursts separated by long intervals of quiescence. The outbursts, moreover, show very different profiles, with two long ($>$6 months), bright ($>$100 mCrab at all energies) episodes, and a shorter and much fainter one (dubbed “failed” in Fig 1). The spectral capabilities of ASMs permit to have an overview of the spectral dependence of the sources light curves, build hardness ratios and hardness intensity diagrams (HID). This allows to survey the sources spectral evolution/transitions and classify their states (Fig 2) to eventually trigger/organise observing campaigns during specific periods of activity.

It is thus possible to use the entire database of any ASM to classify a source’s behavior from multi-color light-curves and HID without a full model dependent approach to the data (e.g Punsly & Rodriguez, 2013; Grinberg et al., 2014, in the respective cases of GRS 1915+105 and Cyg X-1; see Fig. 2). This also allows to fill the temporal gaps and/or uneven temporal sampling of dedicated pointed observing campaign made with more efficient, smaller field instruments (XMM-Newton, Chandra, RXTE, Nustar, for instance). This is an approach that we followed and that allowed us to refine the study of the high energy properties of Cyg X-1 (Rodriguez et al., 2015b).

3 Cyg X-1 and the possible origin hard spectral tails seen in (some) microquasars

While emission above $100$ keV from microquasars is known for more than 20 years (e.g. Laurent et al., 1993; Churazov et al., 1994; Grove et al., 1998; McConnell et al., 2000, to cite just a few), no consensus on its origin has been reached yet. It can be interpreted as due to IC, either purely thermal (spectrum with an exponential cut-off at about 100 keV, Titarchuk, 1994; Rodriguez et al., 2003; Cadolle Bel et al., 2007), or thermal/non-thermal in an hybrid plasma (e.g Zdziarski et al., 2001; Del Santo et al., 2013), or from thermal IC from two different populations of electrons (e.g. Bouchet et al., 2009, to fit the 1E 1740.7$-$2942 INTEGRAL spectrum). In the LHS, jet emission can also be invoked (e.g Markoff et al., 2005) through various radiation processes (Comptonisation, Synchrotron and Synchrotron-Self Compton - SSC). Apart from pure thermal IC, all these models predict similar 20–1000 keV spectral shape, and a single spectral analysis, even with the high quality data obtained today, does not permit to break the model degeneracy (see also Nowak et al., 2011).

Using the large Cyg X-1 INTEGRAL archive Laurent et al. (2011) with IBIS, and Jourdain et al. (2012) with SPI , reported the detection of a high energy tail extending to about 1 MeV in addition to a standard IC component below typically 200–300 keV, and a strong polarized signal above $\sim$400 keV. The large polarisation fraction reported ($\gtrsim 60\%$) led both teams to conclude that the origin of this $>400$ keV tail was Synchrotron emission coming from a compact jet (Laurent et al., 2011; Jourdain et al., 2012).

In both studies, however, the entire INTEGRAL data acquired before 2009 was used, mixing thus all spectral states. As no radio data was considered by any of the two teams the presence of a jet could not be proved either although Cyg X-1 is a known jet source. To go further, and discriminate the potential state dependence of the results, we classified the whole INTEGRAL and AMI-Ryle databases into three basic states based on the ASM-based model independent classification (Grinberg et al., 2014, in Fig. 2, Sec. 2). Stacked high energy spectra, polarigrams and radio properties were then obtained and studied for all three states.

A high energy tail, dominating the spectrum above $\sim$400 keV, is visible only in the LHS (Fig. 3 Rodriguez et al., 2015b). A high degree ($\sim 75\%$) of polarisation is also measured in this state at energies where the tail dominates, while only an upper limit is obtained at lower energies, where the spectrum is dominated by IC, and in the other states. The radio observations independently indicate a high level of emission (a mean of about 13 mJy) in the LHS while in the HSS we report a mean level of 3 mJy, probably due to residual emission from discrete ejections occurring at LHS to HSS transitions. Our refined analysis therefore clearly confirms the presence of both a highly polarized $>$400 keV tail and a compact radio jet during the LHS. We interpret this tail as Synchrotron emission from the jet (Laurent et al., 2011; Jourdain et al., 2012; Rodriguez et al., 2015b).

Alternative possibilities, such as hybrid thermal-non thermal Comptonization in the corona (Zdziarski et al., 2014), or Synchrotron from the corona (see Romero et al., 2014, for all details), have, however, also been suggested. The clear detection of Cyg X-1 with Fermi/LAT up to $\sim$10 GeV (Bodaghee et al., 2013; Zanin et al., 2016; Zdziarski et al., 2016) in the LHS, in addition to the possible TeV detection with MAGIC during a flare (Albert et al., 2007) renders the non-jet interpretation less likely (Zdziarski et al., 2014; Pepe et al., 2015). In particular, and as I discuss in Sec. 5, the coronal model implies a configuration that is not entirely clear, and the MeV-GeV emission is better modeled by emission from a jet (Zdziarski et al., 2014; Pepe et al., 2015; Zanin et al., 2016; Zdziarski et al., 2016; Zhang et al., 2017, Sec. 5), than by a model invoking IC up to the MeV.

4 The extreme case of V404 Cygni

The low mass microquasar V404 Cygni entered into outburst on June 15, 2015 after more than 25 years of quiescence. It was detected with all X-ray wide field instruments (Barthelmy et al., 2015; Negoro et al., 2015; Kuulkers et al., 2015). It was soon recognized that the source had a dramatic behavior, showing intense flares (up to 50 Crab in the 20–40 keV range) on timescales of about an hour (e.g. Rodriguez et al., 2015a; Roques et al., 2015a; Jourdain et al., 2017, Fig.  4). It was also clearly detected up to high energy early in the outburst (Roques et al., 2015b), and a multi-instrumental follow-up and analysis showed probable detection of the source with Fermi in the GeV range (Loh et al., 2016, Fig.  4).

Until then all GeV binaries were high mass systems with either a Be or a Supergiant companion. In these systems, the GeV emission is explained by IC scattering of the stellar (UV) photons on relativistic electrons near the compact object (this model is known as the leptonic model, e.g Bednarek, 2013; Dubus, 2013, and references therein, Sec. 5). In V404 Cyg, however, the secondary is a $0.7_{-0.2}^{+0.3}$ M⊙ K3 III companion (Khargharia et al., 2010, e.g.). In the context of the leptonic model, one cannot expect GeV emission from this source. The discovery of a variable transient feature in the 511 keV region, attributed to $e^{-}/e^{+}$ annihilation further brings the question about the production of pair plasma (Siegert et al., 2016), and the potential role of a jet (that is detected at other wavelengths). The annihilation flux seems correlated with the X-ray flaring activity, which itself somehow correlates with the flares seen at longer wavelengths (Rodriguez et al., 2015a; Siegert et al., 2016; Loh et al., 2016; Muñoz-Darias et al., 2016). Loh et al. (2016) report a $\sim$4$\sigma$ detection in the 0.1–100 GeV Fermi observations at a position coincident with V 404 Cyg. The detected emission at GeV energies occurs in conjunction with the brightest hard X-ray flare, itself associated with a spectral change in the source. Both events appear soon after ($\lesssim$6 hr) a giant radio flare was detected (Loh et al., 2016). The temporal and spatial coincidence of all events strongly support a $\gamma$-ray flare from V404 Cyg during a major flare. The simultaneous detection of the 511 keV annihilation line and $\gamma$-rays of higher energies imply that the latter emission occurs far from the corona, pointing in the direction of jet emission (Loh et al., 2016). The multi-wavelengths properties near the Fermi detection are actually compatible with a Blandford-Znajek (BZ) powered jets interpretation during states when the inner accretion disk is magnetically arrested (Loh et al., 2016).

5 Models for high energy emission and jet contents

The origin of the high energy ($>$MeV) emission is debated. Two main families of models co-exist and predict strong differences at very high energies only (TeV), a domain were there is, to date, only one marginally significant detection of Cyg X-1 (Albert et al., 2007). In the first family (leptonic models) the high energy emission is due to IC scattering of the strong stellar photon field (until 2015 all GeV detected binaries were high mass systems) on relativistic electrons. The second family (hadronic models) invoke the presence of a jet, and the MeV-GeV is instead due to post break Synchrotron and SSC radiations (see e.g. Pepe et al., 2015, and references therein). Leptonic models are popular in the context of $\gamma$-ray binaries where the companion is a high mass star and the primary a pulsar.

In the LHS of MQs the radio-IR emission (Fig. 5) is well explained by self-absorbed Synchrotron radiation from a population of electrons in a compact jet (e.g. Blandford & Konigl, 1979; Fender, 2001b, 2006; Corbel et al., 2013b). Here the presence of the relativistically accelerated particles in a strongly collimated outflow might change our interpretation of the origin the high energy emission(s). The strong polarization fraction detected in the $\gamma$-ray tail of Cyg X-1 excludes IC (Rodriguez et al., 2015b), but not necessarily coronal emission (Romero et al., 2014). However the latter explanation contains intrinsic difficulties that led us to favor the jet emission. The high level of polarization measured with INTEGRAL needs an highly ordered magnetic field which is likely to be found close to the base of the jet, where high energy emission could also occur111It is interesting to note that Zhang et al. (2017) could also reproduce the polarized hard tail with a jet model involving both large and small scale turbulent magnetic fields. . Instead, Synchrotron photons produced by coronal electrons would either have to undergo many Compton scattering if produced deep into the corona, and the polarisation would be lost, or they would be produced too far from the highly ordered field to explain the high degree of polarization we measured.

Pepe et al. (2015) recently applied a lepto-hadronic model to represent the broad band spectrum of Cyg X-1, the hadronic component being required by the TeV flare detected by MAGIC. While globally, the two sets of parameters they use fit the data rather well, the one based on coronal IC and IC of the companion star fails to represent the MeV tail, while this component is well fitted when considering direct jet Synchrotron and SSC instead (Fig. 5). The recent detection of MeV and GeV emission during the V404 Cyg outburst (Siegert et al., 2016; Loh et al., 2016; Jourdain et al., 2017) implies similar conclusions, especially in a system where the secondary has no wind and a much weaker and cooler photon field than high mass systems.

Knowing the composition of jets is of prime importance as it permits to have a proper estimate of their total kinetic power and in-fine their feedback (matter and energy) into the ISM. Baryonic jets should indeed cary significantly more energy than pure leptonic ones, and this of course puts even stronger constraints on the mechanisms responsible for their production and acceleration. It seems widely admitted that jet might cary a very significant part of the accretion energy (Fender, 2001a; Gallo et al., 2005), in the LHS. Jet acceleration is thought to either be based on the extraction of the BH spin (BZ) or magnetic energy from the disc (Blandord & Payne mechanism). Until recently, SS 433 was the only MQ were baryonic jets had been detected (e.g. Kotani et al., 1994), but this source is a peculiar MQ in many other ways, and this could be just another of its peculiarities. In this respect the recent XMM-Newton detection of Doppler shifted emission lines in 4U 1630$-$47  (Díaz Trigo et al., 2013), a more typical MQ, may shed new light on this aspect. Díaz Trigo et al. (2013) could use a simultaneous coverage at radio wavelengths that showed that the Doppler shifted lines are coincident with the detection of a jet. The spectral analysis of the X-ray lines are compatible with being emitted from a $\sim$0.7 $c$ outflow. The flux ratio between the red and the blue shifted lines is, furthermore, consistent with Doppler boosting (Díaz Trigo et al., 2013). This exciting result may imply that typical MQ jets do carry baryons, and that, therefore, their energetic budget is even more important than previously thought. Additionally, and as discussed by Díaz Trigo et al. (2013), the detection of a baryonic jet favours models of accretion disk powered jet rather than BZ jets. A baryonic jet implies that 4U 1630$-$47 should be a strong source of $\gamma$-rays, and this source is not a known as an high energy emitter. These lines have, furthermore never been confirmed since their first report with new Chandra or re-analysis of the XMM-Newton observation (Neilsen et al., 2014; Wang & Méndez, 2016). It is therefore not possible to fully conclude on this issue, which leaves the mystery complete.

6 Conclusions

Jets are clearly important channels for energetic, material redistribution and feedback into the ISM. Their detection at high energy will certainly modify our understanding of MQs in general. The chosen results described in this short review are just the tip of the iceberg, and certainly do not allow to have firm conclusions on most of the questions I raised in the introduction. Soon, the advent the CTA at TeV energy, and/or the detection of neutrinos from a MQ, in conjunction with observations at all wavelengths may help to make a step forward into many of these aspects. One should still keep in mind that studying the accretion-ejection physics relies on multi-instrumental campaigns triggered at appropriate times. This cannot be done without constant survey of the transient sky, and our ability to understand better these objects definitely rely on all sky monitorings such as those made with RXTE/ASM in the past, or currently with MAXI, and Swift.

Acknowledgements

I Warmly thank C. Gouiffès, P. Laurent, & A. Loh for useful discussions and comments on this manuscript. Partial funding support is provided through the CHAOS project (ANR-12-BS05-0009), and UnivEarthS Labex (ANR-10-LABX-0023 and ANR-11-IDEX-0005-02), and by the French Space Agency (CNES) in support to the analysis of the INTEGRAL data.