GeV-TeV Counterparts of SS 433/W50 from Fermi-LAT and HAWC Observations
DDraft version January 13, 2020
Typeset using L A TEX twocolumn style in AASTeX63
GeV-TeV Counterparts of SS 433/W50 from
Fermi -LAT and HAWC Observations
Ke Fang,
1, 2
Eric Charles, Roger D. Blandford, NHFP Einstein Fellow Kavli Institute for Particle Astrophysics and Cosmology (KIPAC), Stanford University, Stanford, CA 94305, USA SLAC National Accelerator Laboratory, 2575 Sand Hill Road, Menlo Park, CA 94025, USA
ABSTRACTThe extended jets of the microquasar SS 433 have been observed in optical, radio, X-ray, and recentlyvery-high-energy (VHE) γ -rays by HAWC. The detection of HAWC γ -rays with energies as great as25 TeV motivates searches for high-energy γ -ray counterparts in the Fermi -LAT data in the 100 MeV–300 GeV band. In this paper, we report on the first-ever joint analysis of
Fermi -LAT and HAWCobservations to study the spectrum and location of γ -ray emission from SS 433. Our analysis findscommon emission sites of GeV-to-TeV γ -rays inside the eastern and western lobes of SS 433. The totalflux above 1 GeV is ∼ × − cm − s − in both lobes. The γ -ray spectrum in the eastern lobe isconsistent with inverse-Compton emission by an electron population that is accelerated by jets. Toexplain both the GeV and TeV flux, the electrons need to have a soft intrinsic energy spectrum, orundergo a quick cooling process due to synchrotron radiation in a magnetized environment. Keywords:
Gamma-ray sources, X-ray binary stars INTRODUCTIONSS 433 is a microquasar in the supernova remnantW50 (see Margon 1984; Fabrika 2004 and referencestherein). It is likely composed of a ∼
20 M (cid:12) blackhole orbiting a ∼ M (cid:12) supergiant companion with a13.1 day period. The exotic system is located at a dis-tance of 5 . ◦ below the Galacticplane. It produces two remarkable jets with kineticpower L kin ∼ erg s − . The jets are heavily loadedwith baryons and move at a speed of 0 .
26 c while pre-cessing with a period of 162 days. The angle betweenjets and the axis is ∼ ◦ . Other periods are measuredbut the dynamics is poorly understood (Eikenberry et al.2001).Extended X-ray jets are observed on the eastern andwestern sides (in Galactic coordinates) of SS 433 asshown by the white contours in Figure 1 (Safi-Harb &¨Ogelman 1997). They interact with and distort the shellof the W50 nebula (Watson et al. 1983; Gregory et al.1996) which is shown by the grey contours. A set ofemission regions, denoted as e1, e2, and e3 centered at24 (cid:48) , 35 (cid:48) and 60 (cid:48) east of SS 433, and w1 and w2 centeredat 18 (cid:48) and 31 (cid:48) west of SS 433 have been investigated indetail (Safi-Harb & ¨Ogelman 1997). A bright knot isseen in soft X-rays at e2 (Safi-Harb & ¨Ogelman 1997;Brinkmann et al. 2007), and emission from e1, e2 and w1, w2 is observed in hard X-rays (Safi-Harb & Petre1999; Moldowan et al. 2005). The X-ray emission canbe explained by synchrotron radiation of ∼ − ∼ µ G magnetic field.Very-high-energy (VHE) γ -ray emission has recentlybeen detected from the SS 433 lobes by the High Al-titude Water Cherenkov (HAWC) observatory (HAWCCollaboration et al. 2018). In a dataset based on1,017 days of measurements, photons with energies ofat least 25 TeV are observed. The TeV hotspots arelocated close to e1, e2 and w1 with spatially unresolvedemission profiles. The flux can be explained by theinverse-Compton emission of the same electron popu-lation whose synchrotron emission is observed by X-ray telescopes. On the other hand, 40–80 h obser-vations with the Major Atmospheric Gamma ImagingCherenkov telescopes (MAGIC) and High Energy Spec-troscopic System (H.E.S.S.) reports no evidence of γ -rayemission between a few hundred GeV and a few TeVfrom the jet termination regions, nor from the centralbinary (MAGIC Collaboration et al. 2018). A similarupper limit is reported by the Very Energetic RadiationImaging Telescope Array System (VERITAS) (Kar &VERITAS Collaboration 2017).Searches (Bordas et al. 2015; Xing et al. 2019; Sunet al. 2019; Rasul et al. 2019) have been made for a GeVcounterpart in the data observed by the Fermi
LargeArea Telescope (LAT) (Atwood et al. 2009). Analysis a r X i v : . [ a s t r o - ph . H E ] J a n Fang et al. of LAT data in this region faces two complications. Apoint source FL8Y J1913.3+0515 from the preliminaryLAT 8-year point source list (FL8Y) is tagged as pos-sibly associated with W50. It is no longer a source inthe 4FGL catalog due to different emission models (TheFermi-LAT collaboration 2019). Different conclusionsabout detection of SS 433 have been reached dependingon whether this source is included in the backgroundmodel. In addition, analysis of the emission profile andspectrum of the SS 433 region is heavily impacted by thenearby pulsar PSR J1907 + 0602 (4FGL J1907 . γ -rays yetclose to the bright Galactic plane, it is difficult to studysolely with the GeV or the TeV measurements. Here wejointly analyze a region-of-interest (ROI) surroundingSS 433 observed by LAT and HAWC. A simultaneousfit to the 100 MeV to 100 TeV data directly addressesthe question whether γ -ray emission over six decades canbe produced by a common cosmic-ray population insidethe SS 433 lobes. We explain the methods in Section 2,present the results of the LAT analysis in Section 3.1and that of the joint analysis in Sections 3.2 and 3.3.We discuss immediate implications of this analysis inSection 4. METHODSOur analysis uses 10.5 years of
Fermi -LAT data and1,017 days HAWC data. Details of the LAT and HAWCanalyses, as well as background sources in each bandare presented in Appendices A and B. The setup of ajoint analysis of the LAT and HAWC data based onthe 3ML framework (Vianello et al. 2015) is presentedin Appendix C. Here we describe the procedure for thejoint analysis.We first build a source model to describe the broad-band γ -ray emission of SS 433. Three types of modelsare considered.I. γ -rays follow a power-law spectrum, dN/dE γ = K γ ( E γ /E γ, piv ) − α γ .II. γ -rays follow a LogParabola spectrum, dN/dE γ = K γ ( E γ /E γ, piv ) − α γ − β γ log( E γ /E γ, piv ) .III. Electrons are injected with a rate d ˙ N /dE e = Q e γ − α e e exp ( − E e /E e, max ), where γ e = E e /m e c https://fermi.gsfc.nasa.gov/ssc/data/access/lat/fl8y/ is the Lorentz factor of an electron. They up-scatter the cosmic microwave background (CMB)and infrared photons in W50 to γ -rays throughthe inverse-Compton process, and produce syn-chrotron emission in a magnetic field B.Models I and II are simple descriptions of γ -ray spectral shapes. Model III is physically moti-vated. The cooling time of VHE electrons, t e, cool =2 . B/ µ G) − ( γ e / ) − kyr, is much less than thesource age, t age ∼
30 kyr (Fabrika 2004). In order totake into account the effects of cooling, we solve a trans-port equation for each set of parameter values. De-tails about the cooled electron spectrum can be foundin Appendix D. Once a steady-state cosmic-ray spec-trum is obtained, the γ -ray flux is calculated using theradiative functions of the naima package (Zabalza 2015;Khangulyan et al. 2014).The models are then converted into data space andcompared to observation through the joint analysisframework (see Appendix C). Since the GeV and TeVobservations are carried out independently, a total loglikelihood is evaluated by summing the log likelihoodsfrom the GeV and TeV analyses. The total likelihoodis then maximized by adjusting model parameters toobtain the best-fit source model. Finally, the likelihoodtest statistic (TS) of a target source is computed as twicethe difference of the log likelihoods of the data given themodels with and without the source. RESULTS3.1.
LAT analysis results
Here we present results from the LAT-only analysis.The method is detailed in Appendix A. The main differ-ence between our analysis and previous works is that weuse a dataset for which the PSR J1907+0602 is gated off,that is, the arrival times of photons are phase folded, andthe photons that arrive during the pulsar’s pulse peakare removed (see Li et al. 2020 for details). Throughoutthe work we use the 4FGL catalog and the correspondingdiffuse emission models to model background sources.The significance of the residual γ -ray excess fromthe SS 433/W50 region in the LAT data between100 MeV and 300 GeV is shown in the left panel ofFigure 1. The most statistically significant excess isnear the location of FL8Y J1913.3+0515. We callthis excess J1913+0515 to differentiate it from FL8YJ1913.3+0515. J1913+0515 is at the boundary of W50,well outside the extended X-ray jets. When describingthe SED with a power-law function, we obtain K γ =1 . × − MeV − cm − s − , E γ, piv = 0 . α γ = 2 .
4. The best-fit location is very close but slightly eV-to-TeV γ -rays from SS 433/W50 Figure 1.
The SS 433/W50 region in the 10.5-year
Fermi -LAT data between 100 MeV and 300 GeV (left) and from jointanalysis of the
Fermi -LAT data and the 1,017-day HAWC data (right) in Galactic coordinates. Left: The color scale indicatesthe statistical significance for a point source following an E − spectrum as a function of position. The figure is a test statisticmap after fitting γ -rays from known sources in the 4FGL catalog. Right: The background includes the 4FGL sources andJ1913+0515 in the GeV band, and MGRO J1908+06 in the TeV band. The color scale indicates the improvement of the totallikelihood of the ROI by a test point source that follows a log parabola spectrum in each 0 . ◦ × . ◦ grid inside the purplesquares. The maps are smoothed by a Gaussian interpolation. The γ -ray hotspots revealed by joint analysis are inside the lobesand close to hard X-ray emission sites. For comparison, the locations of SS 433, the jet termination regions e1, e2, e3, w1 andw2 observed in the X-ray data are indicated as orange crosses. FL8Y J1913 . ∼ . − different from the location listed in the FL8Y catalog.The test statistic of the source is TS = 32 . . <
25) excess is evident at thenortheastern side of J1913+0515. Because it is spatiallyclose to the TeV excess in the eastern lobe, we referto it as the “eastern hotspot”. It is not significant inthe LAT data and has TS = 5 . ∼ −
50 GeV, as shownby the SED in the top panel of Figure 2.In the western lobe, a sub-threshold excess is foundbetween w1 and w2 (which we refer to as the “west-ern hotspot” below). The excess region partially over-laps with the X-ray jets and touches the boundary of W50. We find a TS of 16.1 for the western hotspotwhen adding it to the baseline model. Its spectrum canbe described by a power law of index 2 . ∼ ∼ J1913+0515 and the TeV emission
To investigate whether J1913+0515 and the TeV emis-sion in the eastern lobe share a common origin, we testtwo ways of combining the GeV and TeV hotspots.First, we replace J1913+0515 and the TeV excess witha single source centered between them and assume thatit has a power-law spectrum. The joint fit has six freeparameters in total, including spectral index, flux nor-
Fang et al.
Table 1.
Fit resultsSource Position TS (Individual) Model* Significance(R.A., Dec. indegree) LAT HAWC Individual Totaleastern hotspot (288.56, 4.95) 1.9 21.6 I 4 . σ . σ western hotspot (287.58, 5.01) 8.9 12.1 I 3 . σ eastern hotspot (288.56, 4.95) 4.3 21.7 II 4 . σ . σ western hotspot (287.58, 5.01) 4.6 12.4 II 3 . σ eastern hotspot (288.56, 4.95) 3.3 19.9 III 4 . σ . σ western hotspot (287.58, 5.01) 5.2 10.8 III 3 . σ * For particular models, certain parameters are held constant. They include: Model I, E piv = 875 .
753 MeV, α γ,W = 2 . α γ,E = 2 .
1; Model II, α γ = 1 . β γ = 0 . E γ, piv = 60 GeV; Model III, α e = 1 . B = 20 µ G, and E e, max = 1 PeV. RA andDec are for epoch J2000. See text for additional details. malization, extension of MGRO J1908+06, and flux nor-malization and location (RA, Dec) of the test source.Due to the low statistics, it is difficult to fit the spec-tral index and the flux normalization of the test sourcesimultaneously. We thus fix the index as α γ = 2 . . . . . σ standard deviation.Alternatively, we assume that the sources share a spec-trum but differ in emission sites. The fit results inTS = 30 . . . σ , despitethe two extra degrees of freedom due to the additionalemission site. In general, we find that the LAT TS of thecommon source increases and the HAWC TS decreaseswhen the test source is moved toward J1913+0515, andthe trend is reversed when the test source is moved to-ward the VHE hotspot. Such a trend, along with theconsiderable difference in the statistical significances ofthe models with one and two source locations, suggestthat J1913+0515 is unlikely to be a counterpart of theTeV hotspot in the eastern lobe.3.3. Joint analysis results
Motivated by results from the last section, we per-form a joint analysis of LAT and HAWC data withJ1913+0515 added to the background. The parameters E d N / d E [ e V c m s ] east Fermi-LATHAWC MAGIC-HESS 95% UL (2018)VERITAS 99% UL (2017) E [eV] west E L E [ e r g s ] Figure 2.
The best-fit γ -ray spectra in the eastern andwestern lobes obtained by joint analysis assuming that γ -rays are produced by an electron population (Model III). Theparameters and TS of the model are listed in Table 1. Thegrey shaded area indicates the 68% statistical uncertaintyfrom a fit that varies the normalization. For comparison, weshow the SED from the LAT-only analysis (Section 3.1, redmarkers), HAWC-only analysis (Appendix B, HAWC Col-laboration et al. 2018; blue markers), upper limits on γ -raysfrom nearby regions by VERITAS (Kar & VERITAS Collab-oration 2017) and HESS (MAGIC Collaboration et al. 2018)(grey markers). For the LAT data points, 95% upper limitsare shown when TS <
4, otherwise 1 σ error bars are shown.Since IACT limits are converted from integral limits, theydo not have horizontal error bars. We find that the γ -rayemission in the eastern lobe can be explained as the inverse-Compton emission by a cooled electron population. eV-to-TeV γ -rays from SS 433/W50 l = 0 . ◦ and ∆ b = 0 . ◦ thatcover the e1, e2 and w1, w2 regions, we compute the TSof a test source at every 0 . ◦ × . ◦ grid point assuminga log parabola spectrum (Model II). The scanned re-gions are enclosed by the purple squares in Figure 1. Alog parabola spectrum is chosen because it describes theLAT SED and the HAWC flux better than a power-lawspectrum. We take Model II with α γ = 1 . β γ = 0 . E γ, piv = 60 GeV but have verified that alternative logparabola shapes (for example with α γ = 1 . β γ = 0 . E γ, piv = 5 GeV) leads to similar results. We leavethe normalization K γ as a free parameter. For back-ground sources, we free the normalization and index ofJ1913+0515, the normalization and extension of MGROJ1908+06, and fix parameters of the rest of sourcesin the ROI. The map of TS for the best source posi-tions considered is shown in the right panel of Figure 1.We find that when including the TeV data, the easternhotspot becomes significant, and can be resolved fromJ1913+0515.To directly check whether the GeV-to-TeV emissioncan be explained as inverse-Compton emission of thesame electron population, we perform a joint fit withthe electron model (Model III). We fix the parameters inboth lobes as α e = 1 . B = 20 µ G, and E e, max = 1 PeVand free the normalizations of the electron spectra.These parameters and their values are motivated by thefit to the broadband multi-wavelength data in HAWCCollaboration et al. (2018). We do not scan the pa-rameter space for these parameters, but note that theelectron energy needs to be higher than 150 TeV to pro-duce the measured 20 TeV photons. In general higher E e, max leads to better fits. The best-fit model has aTS of 40 when fitting both lobes simultaneously. With6 free parameters including the two normalizations andthe coordinates of the two hotspots, the TS correspondsto a significance of 5 σ for a two-sided Gaussian distri-bution. The fit results using all three models are listedin Table 1. They are all significant, suggesting that theGeV-to-TeV γ -rays can be explained by common sourcesinside the SS 433 lobes.The SED is shown in Figure 2. For comparison, wealso show the SED obtained from the LAT-only analy-sis, the upper limits (UL) on nearby γ -ray emission byimaging air Cherenkov telescopes (IACT), and the fluxat the pivot energy E piv = 20 TeV from the HAWC-onlyanalysis. We find that the γ -ray flux and the cosmic-rayinjection rates of the east and the west hotspots are verysimilar. To explain both the GeV and TeV flux, a softelectron spectrum dN/dE ∼ E − is needed. This can beachieved by a relatively inefficient acceleration (Bland- ford & Eichler 1987) or by cooling of electrons as sug-gested by HAWC Collaboration et al. (2018). The fluxat GeV energies is higher than that predicted by HAWCCollaboration et al. (2018). This suggests that a far-infrared background needs to be present (whose energydensity is discussed in Appendix D).The best-fit models predict a sub-TeV γ -ray flux thatis higher than the upper limits from IACTs. The upperlimits are based on observations of e2, w2 (H.E.S.S.)and w1 (VERITAS) with small angular extents definedby X-ray observations, and so they do not necessarilyapply to the actual source locations in these models.The western source is less significant in all cases, whichcould be due to confusion by Galactic diffuse emissionand with MGRO J1908+06. The location of the γ -rayemission site in the western lobe is less clear. Like theeastern side, the localized GeV emission could be a com-bination of emission inside the lobe and at the boundaryof W50, though more statistics is needed to verify thisscenario. DISCUSSIONBecause of its proximity and exotic structure, SS 433has been one of the most observed Galactic high-energysources for over 40 years. Nonetheless, no consen-sus has been reached about what happens inside theSS 433/W50 complex. The detection of multi-tens ofTeV photons from the object confirms the existence ofparticles at extreme energies, but deepens the ques-tion why lower-energy γ -rays have not been observed.By jointly analyzing an ROI measured by both the Fermi -LAT and the HAWC Observatory, we find com-mon sites of GeV and TeV γ -ray emission inside theSS 433 lobes. The spectral energy distribution is con-sistent with inverse-Compton emission of an electronpopulation accelerated by the jets but quickly cooleddue to synchrotron radiation in a magnetized environ-ment. We use a dataset that suppresses emission bya nearby pulsar that highly impacts previous analyses.Our joint analysis concludes that the GeV point sourceJ1913+0515 located at the boundary of W50 is unlikelya counterpart to the TeV emission. This addresses thedilemma encountered by Xing et al. (2019); Sun et al.(2019); Rasul et al. (2019).This is the first joint-ROI analysis across γ -ray obser-vatories to our knowledge. Using a framework built onindividual data analysis toolkits from γ -ray observato-ries, we have shown that such an approach is feasible.The joint analysis is designed to study shared proper-ties of sources of γ -rays over a very broad spectrum. Itmaximizes the usage of data including sub-threshold in-formation, and is more powerful than simply combining Fang et al. results from each experiment. Future data from HAWC,especially with refined angular resolutions (HAWC Col-laboration et al. 2019) will help improve the understand-ing of γ -ray emission from the western lobe. Future ob-servations by IACTs, as well as by X-ray and radio tele-scopes, at the revised source locations will help furtherconstrain the emission models.The GeV-TeV spectrum can be explained as inverse-Compton scattering by X-ray synchrotron-emitting ∼
100 TeV electrons. Three scenarios can be entertainedto account for the acceleration. The first is to invokedirect, diffusive shock acceleration of the electrons atthe termination shocks of the precessing jets launchedby the accretion disk. If the post-shock field strengthis ∼ µ G then acceleration to these energies is possi-ble, though only a small electron power ∼ erg s − is needed to account for the γ -rays. Secondly, ∼ ∼ ∼
100 TeV are just possible. How-ever, in order to account for the γ -ray power, a protonpower ∼ erg s − is necessary. The third possibilityis that a hitherto unobserved, ultra-relativistic, electro-magnetic jet is formed by the spinning black hole. Sucha jet can create an EMF ∼
100 ( L jet / erg s − ) / PVthat suffices to accelerate the emitting particles. Finally,if the gas density in the lobes is high ( ∼ > − ), pionproduction can also make a contribution to the γ -rayflux. High-energy neutrinos would be produced simul-taneously and could be measured by IceCube (IceCubeCollaboration et al. 2018). Future multi-messenger observations of the γ -rayemission regions have the potential to discriminate be-tween these scenarios.The Fermi
LAT Collaboration acknowledges gener-ous ongoing support from a number of agencies andinstitutes that have supported both the developmentand the operation of the LAT as well as scientific dataanalysis. These include the National Aeronautics andSpace Administration and the Department of Energy inthe United States, the Commissariat `a l’Energie Atom-ique and the Centre National de la Recherche Scien-tifique / Institut National de Physique Nucl´eaire et dePhysique des Particules in France, the Agenzia SpazialeItaliana and the Istituto Nazionale di Fisica Nucleare inItaly, the Ministry of Education, Culture, Sports, Sci-ence and Technology (MEXT), High Energy AcceleratorResearch Organization (KEK) and Japan Aerospace Ex-ploration Agency (JAXA) in Japan, and the K. A. Wal-lenberg Foundation, the Swedish Research Council andthe Swedish National Space Board in Sweden.Additional support for science analysis during the op-erations phase is gratefully acknowledged from the Is-tituto Nazionale di Astrofisica in Italy and the Cen-tre National d’´Etudes Spatiales in France. This workperformed in part under DOE Contract DE-AC02-76SF00515.We thank Henrike Fleischhack, Colas Rivi´ere, andGiacomo Vianello for their help with the usage of the and
HAWC-HAL packages. We thank Jian Li andMatthew Kerr for their help with performing the pulsargating in the LAT data analysis.REFERENCES
Atwood, W. B., Abdo, A. A., Ackermann, M., et al. 2009,ApJ, 697, 1071Blandford, R., & Eichler, D. 1987, Physics Reports, 154, 1Blundell, K. M., & Bowler, M. G. 2004, ApJ, 616, L159Bordas, P., Yang, R., Kafexhiu, E., & Aharonian, F. 2015,The Astrophysical Journal, 807, L8. https://doi.org/10.1088%2F2041-8205%2F807%2F1%2Fl8Brinkmann, W., Pratt, G. W., Rohr, S., Kawai, N., &Burwitz, V. 2007, A&A, 463, 611Chernoff, H., & Lehmann, E. L. 1954, Ann. Math. Statist.,25, 579Eikenberry, S. S., Cameron, P. B., Fierce, B. W., et al.2001, ApJ, 561, 1027Fabrika, S. 2004, Astrophysics and Space Physics Reviews,12, 1 Gregory, P. C., Scott, W. K., Douglas, K., & Condon, J. J.1996, ApJS, 103, 427HAWC Collaboration, Abeysekara, A. U., Albert, A., et al.2018, Nature, 82—. 2019, arXiv e-prints, arXiv:1905.12518IceCube Collaboration, Aartsen, M. G., Ackermann, M.,Adams, J., et al. 2018, arXiv e-prints, arXiv:1811.07979Kar, P., & VERITAS Collaboration. 2017, InternationalCosmic Ray Conference, 35, 713Khangulyan, D., Aharonian, F. A., & Kelner, S. R. 2014,ApJ, 783, 100Li, J., Torres, D. F., Liu, R.-Y., et al. 2020, submitted toNature Astronomy
MAGIC Collaboration, Ahnen, M. L., Ansoldi, S., et al.2018, A&A, 612, A14 eV-to-TeV γ -rays from SS 433/W50 Margon, B. 1984, Annual Review of Astronomy andAstrophysics, 22, 507Marshall, H. L., Canizares, C. R., Hillwig, T., et al. 2013,The Astrophysical Journal, 775, 75Moldowan, A., Safi-Harb, S., Fuchs, Y., & Dubner, G. 2005,Advances in Space Research, 35, 1062Rasul, K., Chadwick, P. M., Graham, J. A., & Brown,A. M. 2019, MNRAS, 485, 2970Safi-Harb, S., & ¨Ogelman, H. 1997, ApJ, 483, 868Safi-Harb, S., & Petre, R. 1999, ApJ, 512, 784Sun, X.-N., Yang, R.-Z., Liu, B., Xi, S.-Q., & Wang, X.-Y.2019, A&A, 626, A113 The Fermi-LAT collaboration. 2019, arXiv e-prints,arXiv:1902.10045Vernetto, S., & Lipari, P. 2016, Phys. Rev. D, 94, 063009Vianello, G., Lauer, R. J., Younk, P., et al. 2015, arXive-prints, arXiv:1507.08343Watson, M. G., Willingale, R., Grindlay, J. E., & Seward,F. D. 1983, ApJ, 273, 688Wilks, S. S. 1938, The Annals of Mathematical Statistics,9, 60Wood, M., Caputo, R., Charles, E., et al. 2017,International Cosmic Ray Conference, 35, 824Xing, Y., Wang, Z., Zhang, X., Chen, Y., & Jithesh, V.2019, ApJ, 872, 25Zabalza, V. 2015, Proc. of International Cosmic RayConference 2015, 922
Fang et al.
APPENDIX A. FERMI -LAT ANALYSIS
Table A1.
Significance of the candidate sources in the LAT dataNote Source Position (RA, Dec indegree, J2000) 1 σ uncertainty (indegree) TSFit individually J1913+0515 (288.30, 5.24) 0.06 32.8With J1913 eastern hotspot (288.56, 4.95) 0.27 5.0Fit individually western hotspot (288.53, 4.93) 0.13 16.1Fit two sourcessimutaneously J1913+0515 (288.31, 5.24) 0.06 28.4western hotspot (287.58, 5.01) 0.16 9.7Fit three sourcessimultaneously J1913+0515 (288.31, 5.24) 0.06 26.1eastern hotspot (288.56, 4.95) 0.33 5.0western hotspot (287.58, 5.01) 0.17 9.6 We analyze 10 . taken between 2008-08-04 15:43:36 UTC and 2019-01-28 00:00:00 UTC usingversion 0.17.4 of fermipy and version ScienceTools-11-04-00 of Fermitools . We define the ROI as the 15 ◦ × ◦ regionin Galactic coordinates centered at SS 433 ( l = 39 . , b = − . γ -ray events with energies between 100 MeV and300 GeV are selected. Other event selection criteria include a 90 ◦ zenith cut and a filter expression of “DATA QUAL > Fermi
Science Support Center.We use the
P8R2 SOURCE event selection, and the corresponding
P8R2 SOURCE V6
LAT instrument responsefunctions. Unlike previous works analyzing the SS 433 region (Xing et al. 2019; Sun et al. 2019; Rasul et al. 2019),here we use the latest LAT 8-year Point Source Catalog (The Fermi-LAT collaboration 2019), together with thecorresponding Galactic diffuse model gll iem v07.fits and the isotropic diffuse model. The 4FGL catalog and theupdated diffuse emission model turn out to considerably impact the analysis of 100–300 MeV photons in this region,comparing to the FL8Y catalog.As the nearby pulsar PSR J1907+0602 is very bright in the GeV band, we follow the method from Li et al. (2020)to suppress the pulsar emission. The same pulsar ephemeris is adopted in pulsar gating which amounts to 44% of theobserving time. The exposure is scaled accordingly.There are 33 4FGL sources within 5 ◦ of SS 433 and 61 4FGL sources within 8 ◦ . The baseline ROI analysis isperformed using the fermipy.job sub-package fermipy-analyze-roi . The optimized model is referred to as “baselinemodel”.A significance map of the residual γ -ray excess is shown in the left panel of Figure 1. The color scale correspondsto the square root of the TS when there is a new point source at a given location, in addition to known sources fromthe 4FGL catalog, the Galactic diffuse emission and the isotropic diffuse emission. The test point source is assumedto have an E − spectrum and the TS is evaluated for each location on a grid with 0 . ◦ × . ◦ spacing.After setting up the baseline model, we add J1913+0515, the eastern and western hotspots to the background modeland refit the new model to the data. The new model is re-fit to the data using GTAnalysis.optimize , which fits sourcesin the order of their fluxes. The fit returns TS values of 26.1, 5.0 and 9.6 for the three candidate sources, respectively.We also tested an alternative fitting method, where we fixed the parameters of background sources to their best-fitvalues, and vary only the normalization of new sources using
GTAnalysis.fit . This approach returned TS values of28.0, 4.9 and 10.4 for the three. Since the difference of the results from the two fitting methods is minor while the https://fermi.gsfc.nasa.gov/ssc/data/analysis/documentation/Cicerone/Cicerone Data/LAT DP.html https://github.com/fermiPy/fermiPy https://github.com/fermi-lat/ScienceTools eV-to-TeV γ -rays from SS 433/W50 B. HAWC ANALYSISIn the TeV band, we analyze the public data from the High Altitude Water Cherenkov (HAWC) Observatory(HAWC Collaboration et al. 2018). The dataset contains 1,017 days of γ -ray events collected between 26 November2014 and 20 December 2017. The reconstruction of the arrival direction of primary γ -rays is based on the relativearrival times of photoelectron hits detected by the photomultipliers inside the water Cherenkov detectors. (This kindof reconstruction is referred as the nhit method.) Angular resolution from the nhit analysis ranges from around 1 ◦ below 1 TeV to < . ◦ above 10 TeV.We adopt the same ROI as in HAWC Collaboration et al. (2018), which is defined to be a semicircular region with aradius of 2 . ◦ centered on the position of MGRO J1908+06 (as shown in Extended Data Fig 1 of HAWC Collaborationet al. 2018). By masking the sources close to the Galactic plane, the contamination from the Galactic diffuse emissionis significantly reduced.Three sources remain in the ROI: MGRO J1908+06, the eastern and the western hotspots in the SS 433 lobes.Following HAWC Collaboration et al. (2018), we use the electron diffusion model to describe the spatial morphologyof MGRO J1908+06. Other spatial models with Gaussian and power-law radial profiles lead to similar results.Unlike the analysis in HAWC Collaboration et al. (2018), which is based on the HAWC analysis framework AERIE,here we redo the analysis using the HAWC Accelerated Likelihood (HAL) framework. HAL provides faster convolutionwith the detector response functions which is needed by the joint analysis in our work. We have confirmed that thetwo analysis frameworks lead to results that are consistent at the 1% level. C. JOINT ANALYSISThe work flow of a joint analysis is diagrammed in Figure C1. The joint analysis is implemented in the Multi-Mission Maximum Likelihood framework (3ML) (Vianello et al. 2015) . 3ML is a data analysis architecture thatconverts emission models for an ROI into data spaces for specified instrument(s), and compares the model predictionsto the corresponding data based on the likelihood formalism. The Fermi -LAT module of the package provides awrapper of fermipy (Wood et al. 2017) and the
Fermi Science Tools . The HAWC module links to the HAWC analysistools including the Accelerated Likelihood (HAL) framework .Since a ROI analysis is not implemented in 3ML, we use fermipy to perform a “baseline” analysis externally (seeAppendix A) and use that as a starting point for 3ML analysis. Meanwhile, the source model and a model of MGROJ1908+06 are passed to the HAWC plugin (see Appendix B). In this way the contribution of background sources istaken into account properly. D. RADIATIVE COOLING OF ELECTRONSThe cooling of relativistic electrons in the lobes of SS 433 can be described by a transport equation ∂N e ∂t + ∂∂γ e [ ˙ γ e N e ( γ e , t )] = Q e ( γ e , t ) , (D1)where ˙ γ e = − / γ e c σ T ( u B + u γ ) / ( m e c ) ≡ − ν γ e is the energy loss rate due to inverse-Compton and synchrotronemission. σ T is the Thomson cross section. N e and Q e are the spectrum and injection rate of electrons, respectively. u B = B / (8 π ) and u CMB = 0 .
26 eV cm − are the energy density of magnetic field and the CMB. We also adopt a far-infrared (FIR) background at 20 K with u FIR = 0 . − motivated by the dust emission in the solar neighborhood(Vernetto & Lipari 2016). Background photons with higher energies are not important due to the Klein-Nishina https://data.hawc-observatory.org/datasets/ss433 2018/index.php https://github.com/threeML/threeML https://fermi.gsfc.nasa.gov/ssc/data/analysis/software/ https://github.com/threeML/hawc hal Fang et al.
Figure C1.
Diagram of the work flow of a joint analysis of the
Fermi -LAT and HAWC data around SS 433. The source modeldescribes the γ -ray emission of the potential sources and may depend on the parent particle types, injection spectra, sourcelocations, and magnetic field strength. The background model is composed of dozens of 4FGL sources, the diffuse emissionmodels in the GeV band, and MGRO J1908+06 in the TeV band. The models are passed to analysis pipelines for the two datasets, converted to data space and compared to data separately. The total likelihood is maximized to obtain the best-fit model. effect. Due to its much lower energy density, the synchrotron radio emission of W50 and the lobes is not expectedto contribute significantly to the cooling of electrons or production of high-energy γ -rays. In equation D1 we haveignored the diffusion of electrons as it is a slower process than cooling for TeV electrons, and also because doing sosaves computing time. Assuming that electrons are injected with a simple power-law spectrum constantly over time, Q e ( γ e , t ) = Q e, γ − αe , the solution of equation D1 can be written as N ( γ e , t ) = Q e, γ e (cid:90) tt min dt i (cid:2) γ − e − ν ( t − t i ) (cid:3) α − (D2)where t min = max[0 , t − ν − ( γ − e − γ − e, max )] is the earliest time that an electron with γ e can be injected and still notcooled after time t , and γ e, maxmax