HESS J1858+020: A GeV-TeV source possibly powered by CRs from SNR G35.6-0.4
aa r X i v : . [ a s t r o - ph . H E ] J a n Astronomy & Astrophysicsmanuscript no. 39041_final © ESO 2021January 15, 2021
HESS J1858+020: A GeV-TeV source possibly powered by CRsfrom SNR G35.6-0.4
Y. Cui , Y. Xin , S. Liu , P.H.T. Tam , G. Pühlhofer , and H. Zhu School of Physics and Astronomy, Sun Yat-Sen University, Guangzhou, 510275, Chinae-mail: [email protected], [email protected] School of Physical Science and Technology, Southwest Jiaotong University, Chengdu 610031, People’s Republic of China e-mail: [email protected] Key Laboratory of Dark Matter and Space Astronomy, Purple Mountain Observatory, Chinese Academy of Sciences, Nanjing210033 Institut für Astronomie und Astrophysik, Eberhard Karls Universität Tübingen, Sand 1, D 72076 Tübingen, Germany Key Laboratory of Optical Astronomy, National Astronomical Observatories, Chinese Academy of Sciences, Beijing 100101,ChinaReceived 28 / / / / ABSTRACT
Context.
The supernova remnant (SNR) G35.6 − γ -ray or X-ray counterparts havebeen found for it thus far. One TeV source, HESS J1858 + Aims.
To attain a better understanding of the origin of HESS J1858 + Methods.
We performed the Fermi-LAT analysis to explore the GeV emission in and around the SNR. We explored the SNR physicswith previously observed multi-wavelength data. We built a hadronic model using runaway CRs of the SNR to explain the GeV-TeVobservation.
Results.
We found a hard GeV source (SrcX2) that is spatially coincident with both HESS J1858 +
020 and a molecular cloud complexat 3.6 kpc. In addition, a soft GeV source (SrcX1) was found at the northern edge of the SNR. The GeV spectrum of SrcX2 connectswell with the TeV spectrum of HESS J1858 + γ -ray spectrum ranges from several GeV up to tens of TeV and itfollows a power-law with an index of ∼ + ff usion coe ffi cient that is much lower than the Galactic value. Key words. acceleration of particles − (ISM:) cosmic rays − gamma rays: ISM − ISM: supernova remnants
1. Introduction
HESS J1858 +
020 is one of the first unidentified TeV sourcereported by Aharonian et al. (2008). In the HESS Galacticplane survey (HGPS) (H. E. S. S. Collaboration et al. 2018),HESS J1858 +
020 is also considered as one of eleven HGPSsources that do not yet have any associations with knownphysical objects. In the context of the HAWC telescope ( >
100 GeV), which is mostly sensitive at ∼
10 TeV, there is noknown 2HWC catalog source associated with HESS J1858 + − +
020 was identified as a SNR by Green (2009). Theradio boundary of this SNR is clearly shown in the 1.4GHz im-age with VGPS (Stil et al. 2006; Zhu et al. 2013) and the 610MHz image with GMRT (Paredes et al. 2014).A molecular cloud (MC) complex composed of two clumpshas been found to be spatially coincident with HESS J1858 + +
020 is likely due to the interaction between theSNR CRs and this MC complex. This argument is further cor-roborated in Paron et al. (2011) by excluding the possibility of ayoung stellar object as the power source. The velocity of this MC complex is ∼ − , corresponding to a near-side distance of ∼ . ∼ . i study byZhu et al. (2013) suggested a distance to this SNR-cloud systemof 3.6 ± . i study by Ranasinghe, & Leahy (2018). It was suggested thattwo nearby pulsars – PSR J1857 + + . ± . + + ff use X-ray emission has been found in or aroundSNR G35.6 − − ff usion Article number, page 1 of 7 & Aproofs: manuscript no. 39041_final coe ffi cient in order to explain the lack of GeV emission. In the3FHL catalog, an extended source, 3FHL J1857.7 + + + +
026 region, with HESS J1858 +
020 is not straightfor-ward. Recently, two point 4FGL sources at SNR G35.6 − − − + γ -ray spectrum extends from sev-eral GeV up to tens of TeV. Assuming the GeV-TeV emis-sion of HESS J1858 +
020 is powered by escaped CRs fromSNR G35.6 − .
100 GeV). In Section 3, we look for evidence of a middle-agedSNR from previous multi-wavelength observations. In Section 4,we build a hadronic model to explain the GeV-TeV spectrum ofHESS J1858 + Fermi -LAT data analysis
In the following analysis, we selected the latest Fermi-LATPass 8 data with a “Source” event class (evclass =
128 & ev-type = ◦ are excluded to reducethe contamination from the Earth Limb. The region of interest(ROI) is a 14 ◦ × ◦ square region centered at the position ofHESS J1858 +
020 (Aharonian et al. 2008) and the standard LATanalysis software,
FermiTools, was adopted. The Galactic andisotropic di ff use background emissions are modeled according to gll_iem_v07.fits and iso_P8R3_SOURCE_V2_v1.txt . Allthe sources listed in the 4FGL catalog (Abdollahi et al. 2020)within a radius of 20 ◦ from the ROI center, together with the twodi ff use backgrounds, are included in the background model.In the vicinity of HESS J1858 + + = ◦ ,Dec. = ◦ ) and 4FGL J1857.6 + = ◦ ,Dec. = ◦ ), and yet these two 4FGL sources have no as-sociations. Further to the north, also there is an extended sourceknown as 4FGL J1857.7 + +
026 (Aharonian et al. 2008).Recent MAGIC observations revealed that the > +
026 can be spatially separated into twosources: MAGIC J1857.2 + + + +
026 recorded by MAGIC Collaboration et al. (2014).Close to the SNR, we can find the TeV sourceHESS J1858 + http: // fermi.gsfc.nasa.gov / ssc / data / analysis / software / http: // fermi.gsfc.nasa.gov / ssc / data / access / lat / BackgroundModels.html extension of 5 ′ (Aharonian et al. 2008), which is shown asa white circle in Fig. 1. The observations with MAGIC(MAGIC Collaboration et al. 2014) also show a point-likesource at HESS J1858 +
020 (named as MAGIC-south). How-ever, the TeV emission around the SNR is not investigated bythe MAGIC Collaboration et al. (2014) and no flux or spectralinformation of MAGIC-south is given due to its relatively lowexposure at its angular distance from the MAGIC pointing po-sitions. In our Fermi-LAT analysis below, we also performed ananalysis with the HESS / MAGIC image of HESS J1858 +
020 asan extended template.
For the purposes of carrying out a more detailed study ofthe emission around SNR G35.6-0.4 / HESS J1858 + + + + > >
10 GeVTS maps are shown in Fig 1. As clearly seen in the SNR region,the 5 GeV photons are mostly concentrated at the northeasternedge of the SNR, while the 10 GeV photons are mostly aroundHESS J1858 + ff erence between 5 GeV and10 GeV TS maps seems to indicate a two source scenario. As-suming these two sources are point sources, we hereafter fit thecoordinates of the two sources with the command gtfindsrc and rename them as SrcX1 and SrcX2.In the course of the following analysis, which focus on find-ing the spatial and spectra information of SrcX1 and SrcX2, theextended source 4FGL J1857.7 + + +
026 (the cyan contours north of the SNR in Fig. 1)were used as the spatial template of 4FGL J1857.7 + > = . ◦ , Dec. = . ◦ , with 1 σ er-ror radius of 0.040 ◦ . And for SrcX2, the best-fit position and its1 σ error radius are R.A. = . ◦ , Dec. = . ◦ and 0.029 ◦ ,respectively. The TS (Mattox et al. 1996) values of SrcX1 andSrcX2 were fitted to be 49.1 and 39.3, corresponding to the sig-nificant level of 6.2 σ and 5.4 σ , respectively. More comparisonsbetween SrcX1 and SrcX2 in di ff erent energy bands are list inTable 1.Additionaly, extended spatial templates of SrcX1 & SrcX2were also explored to test their spatial extensions. These tem-plates include uniform disks with a di ff erent radius, the MAGIC-south image of HESS J1858 +
020 and the HESS image of HESSJ1858 + / SrcX2and point sources are enough to describe their γ -ray emissions.In Fig 1, the radio contours of SNR G35.6 − CO J = − (Paron & Giacani 2010) is shown inpurple contours and named "CloudX2" for the purposes of thiswork. As can be seen, the position of SrcX2 is in good correspon-dence with the TeV image of HESS J1858 +
020 and CloudX2,while SrcX1 appears to be corresponding to the radio shell ofSNR G35.6 − Article number, page 2 of 7ui et al.: Fermi-LAT analysis & hadronic modeling of HESS J1858 + Table 1.
Best-fit position, spectral parameters, and TS values of SrcX1 / SrcX2 for di ff erent energy bands with point source assumptions σ error Spectral Photon Flux TSradius Index (10 − ph cm − s − ) ValueSrcX1 284 . ◦ & 2 . ◦ . ◦ . ± .
49 2 . ± .
49 49.1SrcX2 284 . ◦ & 2 . ◦ . ◦ . ± .
27 1 . ± .
45 39.35-10 GeVSrcX1 284 . ◦ & 2 . ◦ . ◦ . ± .
94 2 . ± .
47 48.6SrcX2 284 . ◦ & 2 . ◦ (fixed) 2.31(fixed) 1 . ± .
39 13.310-500 GeVSrcX1 284 . ◦ & 2 . ◦ (fixed) 3.73(fixed) < . ◦ & 2 . ◦ . ◦ . ± .
41 0 . ± .
22 28.2
With the best positions of SrcX1 and SrcX2 obtained from the 5-500 GeV data (see Table 1), we used gtlike to fit the power-lawspectra of them in the energy range of 1-500 GeV. The spectralindex and total photon flux of SrcX1 are 3 . ± .
09 and (6 . ± . × − photon cm − s − . While the spectral index and totalphoton flux of SrcX2 are fitted as 2 . ± .
14 and (1 . ± . × − photon cm − s − .To derive the spectral energy distribution (SED) of SrcX2at di ff erent energies, we binned the data with six logarithmi-cally even energy bins between 1 GeV and 500 GeV and weperformed the same likelihood fitting analysis to the data. Con-sidering the much softer spectrum of SrcX1, we derived the en-ergy band of 1-50 GeV into eight bins to obtain its SED. For theenergy bin with the TS value of SrcX1 / SrcX2 smaller than 5.0,an upper limit with 95% confidence level was calculated. Theresults of the spectral analysis are shown in Fig 2.In summary, we performed the Fermi-LAT analysis aroundSNR G35.6 − +
020 and CloudX2; the other one displaying asoft GeV spectrum (SrcX1) is located at the northern edge of theSNR. The GeV spectrum of SrcX2 connects well with the TeVspectrum of HESS J1858 +
020 and together, they show a hardGeV-TeV spectrum with a power-law index of ∼
3. Multi-wavelength observations aroundSNR G35.6 − In adopting the 1.4GHz image with VGPS (see Fig. 3), we ob-tained a close-up of the complex with an angular size of ∼ ′ − ′ and a center at RA = . ◦ , Dec = . ◦ (Zhu et al.2013).The most up-to-date H i & CO studies by Zhu et al. (2013)and Ranasinghe, & Leahy (2018) suggest that the distance tothe SNR-cloud system (SNR G35.6 − ± . ± . ∼ As described in Section 2.3, the γ -ray spectrum at CloudX2 ex-tending down to several GeV seems to indicate that this is thecase of a middle-aged SNR. So far, there is no direct observa-tional evidence for the shock velocity. In following: we discuss the evidence based on multi-wavelength studies in favor of amiddle-aged SNR. γ -ray emission extending down to several GeV: The flat γ -ray spectrum at CloudX2 ranges from several GeV to tensof TeV, which is similar to those of the middle-aged SNRsassociated with MCs, such as SNR W28, W44, and IC443. Itsflat spectral shape, in particular, resembles those of the clouds(240A,B,C) next to but not in direct proximity to SNR W28(Abdo et al. 2010; Hanabata et al. 2014; Cui et al. 2018). If ahadronic origin is assumed, then the SNR has already releasedCRs with energies down to tens of GeV into CloudX2. A slowshock ( v SNR ≪ − ) at present can achieve an escapeenergy of ∼
10 GeV with the help of the damping of magneticwaves by neutrals (O’C Drury et al. 1996). It could be arguedthat a young SNR can also release the GeV CRs to MCs througha shock-cloud collision scenario (Cui et al. 2019; Tang et al.2015). Nonetheless, there is no evidence from the linewidthof the molecular clouds that are interacting with the SNR; forexample, no line asymmetry was found at CloudX2 (Paron et al.2011).
An intrinsic weak X-ray emission : The non-detection ofdi ff use X-ray emission (Paredes et al. 2014) can be either due tothe intrinsic nature of the SNR or the heavy absorption. The totalH column density along the line of sight (LOS) of G35.6 − N H , total ≈ . × H cm − (Willingale et al. 2013). Theextinction curve along the LOS of the SNR becomes flat behind ∼ . N H , total . Obviously, this1 . × H cm − is not thick enough to absorb most keVphotons; see, for example, the SNRs with higher N H displayingX-ray emissions (Zhu et al. 2017). Hence, the non-detection ofdi ff use X-ray emission is likely due to an intrinsic weak source. A hard radio index:
The non-thermal radio spectrum ofSNR G35.6 − α = . ± .
07 (Green2009). This value is much harder than the ones ( α ≈ . − . − − , a SNRdistance of 3 . <
100 GeV CRs – hence,a relatively higher ambient density is required.
Article number, page 3 of 7 & Aproofs: manuscript no. 39041_final
20 40 60 80 100 120 : . : . : : . PSF (5GeV)
MAGIC-south (MAGIC)HESS J1858+020 (HESS)HESS J1857+026 (MAGIC) : . : : . PSF (10GeV)
HESS J1858+020 (HESS)
SrcX2 SrcX1
Fig. 1.
TS maps for photons above 5 GeV(top) and 10 GeV(bottom)are shown. Both of the maps are smoothed with a σ = . ◦ Gaussiankernel. The PSF of a single 5 GeV /
10 GeV photon is also shown indashed yellow circles. The red cross and circle show a point source –4FGL J1855.8 + + + + σ error radius of SrcX1 and SrcX2 are also shown in dashed circles. Thecyan dashed circle in the top panel shows the TeV extension of HESSJ1857 + CO J = Following the CO study by Paron & Giacani (2010) and thedistance study by Zhu et al. (2013), CloudX2 is shown to have amass of ∼ . × M ⊙ ; see also Fig. 3. The projected distancebetween CloudX2 and the SNR center is ∼ −1 E [GeV]10 −13 −12 −11 E d N / d E [ e r g c m − s − ] SrcX1SrcX2HESS J1858+020
Fig. 2.
Fermi-LAT SED of SrcX1 (black dots) and SrcX2 (blue dots),with arrows indicating the 95% upper limits. Red dots represent theHESS observation of HESS J1858 +
020 (Aharonian et al. 2008). Graybutterfly indicates the best-fit power-law of SrcX1 in the energy rangeof 1-500 GeV. Solid green line is the joint fit for the Fermi-LAT dataof SrcX2 and HESS data of HESS J1858 + ff erential sensitivities of LHAASO (1 year)with di ff erent sizes of photomultiplier tube (PMT; Bai et al. 2019).Black dotted line represents the di ff erential sensitivity of CTA-North(50 hrs; CTA Consortium 2019). : . : . : : . : . : . HII PN
PSR J1858+0215PSR J1857+0210PSR J1858+0215
SrcX2 SrcX1
Fig. 3. CO map around SNR G35.6 − γ -ray counterpartsare neglected in this work. The 1.4GHz image from VGPS is markedin green contours. The best fitted position of SrcX1 and SrcX2 aremarked in red diamonds, red circles represent their position uncertain-ties (2 σ ). The X-ray point sources and the nearby pulsars are marked inwhite crosses and circles, respectively. The H ii region G35.6 − − + Fig. 4.
SNR evolution profiles. The radial profiles of the circumstellardensity, shock velocity, and the escape energy are shown in black, red,and blue lines, respectively. model, shown below, they are assumed to be located far fromthe SNR.Noticeably, a H i shell sometimes could be associated witha pre-SN wind bubble or a SNR; see, for example, the slow-expanding H i shell around the Wolf-Rayet star HD 156385(Cappa de Nicolau et al. 1988) and the fast-expanding H i shellsaround SNR CTB 80 (Park et al. 2013) & SNR Cygnus Loop(Leahy 2003). Finding such a shell could help us to confine theprogenitor type and the ambient density, which eventually leadsto a more accurate SNR age. Unfortunately, H i shells are verydi ffi cult to find and only several are known among all ∼ i data used in Zhu et al. (2013) and we do not find any evidencefor either a slow or a fast H i shell around SNR G35.6 − ii region is located within the 2 σ uncertainty circle.Hence, SrcX1 will be considered as a background source in thiswork.
4. A hadronic explanation of GeV-TeV emission ofHESS J1858+020
As described in the previous section, one of the most challengingcharacteristics of the γ -ray spectrum of SrcX2 is the broad en-ergy range. A successful acceleration model for such a spectrumshould be able to release the &
100 TeV CRs during its earlySNR stage, as well as the 10 −
100 GeV CRs during its late SNRstage.A simple SNR evolution history with a homogeneouscircumstellar medium is adopted in our model. To calcu-late the SNR evolution history, the analytical solution of v SNR ∝ t − / (Chevalier 1982; Nadezhin 1985) was adoptedfor the ejecta-dominated stage, while the thin-shell approxima-tion (Ptuskin & Zirakashvili 2005) was adopted for the Sedov-Taylor stage; the result is essentially consistent with the an-alytical solution of v SNR ∝ t − / (Ostriker & McKee 1988;Bisnovatyi-Kogan & Silich 1995) and the analytical solution of v SNR ∝ t − / by Cio ffi et al. (1988) is adopted for the pressure-driven snowplow (PDS) stage. The SNR evolution profiles areshown in Fig. 4.In explaining the hard TeV tail of SrcX2, we adopt theacceleration theory of nonresonant streaming instability devel- oped by Bell (2004) and Zirakashvili & Ptuskin (2008). Thistheory can boost the escape energy up to hundreds of TeV inyoung SNRs. Given a SNR’s evolution history and an accel-eration e ffi ciency, an analytical approximation of this theory(Zirakashvili & Ptuskin 2008) can provide us with the runawayCR flux J and the escape energy E max . In a strong shock, onlyCRs with energies above E max can escape from the shock up-stream and become runaway CRs. A magnetic field of B = µ Gand an initial magnetic fluctuation of B b = B in the ICMare assumed in the calculation, following Zirakashvili & Ptuskin(2008).In explaining the GeV spectrum of SrcX2, the dampingof the magnetic waves by the neutrals is adopted in the lateSNR stage. This damping e ff ect can significantly lower E max and it is considered important in mid-aged and old SNRs; bothShull, & McKee (1979) and Sutherland, & Dopita (2017) notedthat a shock slower than ∼
100 km s − will lead to a significantdrop of the UV ionization at shock precursor. The relationshipof E max = v n . n − TeV (O’C Drury et al. 1996) is adopted toestimate the escape energy in partially ionized medium, where v is shock velocity in unit of 10 km s − , n H is the circumstel-lar density, and n n is the neutral density. A homogeneous di ff u-sion coe ffi cient was adopted in the entire space, which follows apower-law rule of D = D E δ . By integrating the SNR surfaceas well as the entire SNR evolution history, it is possible to ob-tain the present CR density at CloudX2; see also the equationsin Section 2.4.1 of Cui et al. (2016). The escape energy E max is highly sensitive to the shock veloc-ity, v SNR . To generate ∼
10 GeV CRs, one requires a shockvelocity of v SNR ≪ − during the late stage of theSNR evolution. Hence, we adopt a relative high circumstellardensity of n H =
20 H cm − . Such a density at a Galactocen-tric distance of 5 kpc indicate that the circumstellar medium iscold neutral medium (Wolfire et al. 2003; Cox 2005). When theSNR reaches 8 pc, our model gives a SNR age of t SNR =
18 kyr,a shock velocity of v SNR =
142 km s − , and an escape energy of E max =
26 GeV, see Fig. 4.More details of the best fitted parameters are shown in Ta-ble 2, where we also show the dependencies of our fitting resultson those parameters. A higher E ej gives a overall higher v SNR .Both a higher v SNR and a higher η can eventually leads to a highertotal CR production and a higher escape energy E max . The valueof E ej is suggested to be around 1 E . The value of η is limitedby that the energy of total accelerated CRs should not be too farfrom 10% E ej . A lower di ff usion coe ffi cient, D , and longer SNR-cloud distance, L , will suppress the CRs (mostly GeV CRs) fromreaching CloudX2. However, once the CRs can easily reach thecloud (mean di ff usion distance after certain time is beyond theSNR-cloud distance L ), a lower di ff usion coe ffi cient helps con-fining the CRs in the SNR-cloud region from spreading too thin.The neutral density n n in the shock precursor lacks observa-tional constraints. Following the recent simulation work (Fig. 5in Sutherland, & Dopita (2017)), we adopt an estimation of X = . · ( v SNR /
100 km s − ) − in a range of v SNR = −
500 km s − ,where X = n n / n H and 1 − X is the ionization ratio. In a moredetailed model, for a future study, X is dependent on manyother factors, such as metallicity or magnetic field; see alsoSutherland, & Dopita (2017). The damping of magnetic wavestakes e ff ect when the shock velocity is below ∼
500 km s − , andthis 500 km s − is chosen to get a smooth transition from the the- Article number, page 5 of 7 & Aproofs: manuscript no. 39041_final
Table 2.
Parameters of the SNR model E ej a η b D c δ c L d parameter value 1 . E γ -ray flux e + + − spectral index f + + − − + a Explosion energy of the SN. E = erg. b The acceleration e ffi ciency η represents the ratio betweenthe energy flux of runaway CRs and the kinetic energyflux of incoming gas onto the shock, and it remains con-stant through the entire SNR evolution. The total energyof all the released CRs is 23% E ej . c The di ff usion coe ffi cient follows a rule of D = D E δ ,where D is in unit of 10 cm s − . D = δ = . − . ff usion coe ffi cient. d The three dimensional distance between the SNR andCloudX2 (The projected distance is ∼ e “ + " / “-" means that with a increasing value of the pa-rameter, the γ -ray flux (total energy of GeV-TeV band)of SrcX2 increases / decreases. For parameter D , both “ + "and “-" could happen, that is why we leave them blank,see also the explanations in text. f “ + " / “-" means that with a increasing value of the pa-rameter, the γ -ray spectral index becomes harder / softer(higher / lower TeV to GeV ratio). Fig. 5.
Hadronic model results using SNR CRs in explaining the GeV-TeV emission of SrcX2. The Fermi-LAT data and HESS data of SrcX2are marked in red and blue, respectively. Our model results are shownin solid lines, the contribution of runaway CRs and CR sea are markedin dashed lines and dotted lines, respectively. ory of Zirakashvili & Ptuskin (2008) to that of O’C Drury et al.(1996).Using the nonresonate acceleration model in the early stageand the damping model in the late stage, over the entire SNRevolution, the runaway CRs cover a large energy range, from ∼
10 GeV up to more than 100 TeV, see Fig. 4. To explain theflat GeV-TeV spectrum of SrcX2, the early released TeV CRsshould not di ff use too far at an age of 18 kyr, meanwhile thelate-released GeV CRs should be able to reach CloudX2. Hence,a relative hard index of di ff usion coe ffi cient (0.25) is adopted;see also Table 2. Ultimately, these SNR CRs explain the GeV-TeV emission of SrcX2 with a di ff usion coe ffi cient that is muchlower than the Galactic value; see the spectrum fitting in Fig. 5.The SNR has a Galactocentric distance of 5 kpc, while the cor-responding CR sea contribution (Yang et al. 2016; Acero et al.2016) is very little and can be ignored.
5. Discussions and observational expectations
One of the arguments in support of SNR G35.6 − ff use X-ray emission withChandra. More sensitive observations with XMM-Newton maysolve this question by detecting either the thermal or the non-thermal emission. For instance, the thermal emission foundin SNR W28 with XMM-Newton (Zhou et al. 2014) shows atemperature of ∼ . ∼ × H cm − . The column density at SNR G35.6 − ∼ . × H cm − . Future X-ray observations may also discover thealternative power sources if they are leptonic dominated – forexample, a PWN origin for HESS J1640-465 (Xin et al. 2018) –as well as the potential shock-cloud collisions – for example, thethermal X-ray emission at the northeastern shell of SNR W28(Zhou et al. 2014). Noticeably, millimeter observations of ion-ization lines may also shed some light on the shock-cloud colli-sions.One of the most interesting feature of SrcX2 is the hard TeVtail and we expect future observations with LHAASO / CTA mayfurther characterize the hard tail (see Fig. 2). In addition, theymay also reveal more detailed TeV features at or around theSNR.
6. Summary
We carried out an analysis the Fermi-LAT data atSNR G35.6 − > +
020 and the molecularcloud complex at east – CloudX2. The spectral index of SrcX1and SrcX2 are 3.09 ± ± + − ff use X-rayemission, especially for the keV band, is likely due to an intrin-sic weak source rather than the heavy absorption. Secondly, if theSrcX2 is indeed powered by the SNR CRs, then the GeV emis-sion found at CloudX2 indicates that CRs with energies down to ∼
10 GeV have been released from the SNR. Thirdly, the radioindex of SNR G35.6 − − &
100 TeV during the early SNR stage. Thedamping of magnetic waves by the neutrals was adopted for thelate SNR stage and it leads to the release of CRs with energiesdown to ∼
10 GeV. Our model requires a di ff usion coe ffi cientthat is much lower than the Galactic value and, in particular, ahard index of di ff usion coe ffi cients is needed to suppress the dif-fusion of early-released TeV CRs. Acknowledgements.
We like to thank Guangxing Li, Wenwu Tian for discus-sions on radio studies of SNR G35.6 − Article number, page 6 of 7ui et al.: Fermi-LAT analysis & hadronic modeling of HESS J1858 + and the International Partnership Program of Chinese Academy of Sciences(No.114332KYSB20170008). Hui Zhu is supported by National Key R&D Pro-gram of China (2018YFA0404203) and NSFC (11603039). References
Abeysekara, A. U., Albert, A., Alfaro, R., et al. 2017, ApJ, 843, 40Abdo, A. A., Ackermann, M., Ajello, M., et al. 2010, ApJ, 718, 348Acero, F., Ackermann, M., Ajello, M., et al. 2016, ApJS, 223, 26Abdollahi, S., Acero, F., Ackermann, M., et al. 2020, ApJS, 247, 33Aharonian, F., Akhperjanian, A. G., Barres de Almeida, U., et al. 2008, A&A,477, 353Bai, X., Bi, B. Y., Bi, X. J., et al. 2019, arXiv:1905.02773Bell, A. R. 2004, MNRAS, 353, 550Bisnovatyi-Kogan, G. S., & Silich, S. A. 1995, Reviews of Modern Physics, 67,661Cappa de Nicolau, C. E., Niemela, V. S., Dubner, G. M., et al. 1988, AJ, 96, 1671Cio ffi , D. F., McKee, C. F., & Bertschinger, E. 1988, ApJ, 334, 252Chevalier, R. A. 1982, ApJ, 259, 302Cox, D. P. 2005, ARA&A, 43, 337CTA Consortium 2019, Science with the Cherenkov Telescope Array,9789813270091Cui, Y., Pühlhofer, G., & Santangelo, A. 2016, A&A, 591, A68Cui, Y., Yeung, P. K. H., Tam, P. H. T., & Pühlhofer, G. 2018, ApJ, 860, 69Cui, Y., Yang, R., He, X., et al. 2019, ApJ, 887, 47Dubner, G., & Giacani, E. 2015, A&A Rev., 23, 3O’C Drury, L., Du ff y, P., & Kirk, J. G. 1996, A&A, 309, 1002Green, D. A. 2009, MNRAS, 399, 177Green, G. M., Schlafly, E., Zucker, C., et al. 2019, ApJ, 887, 93Hanabata, Y., Katagiri, H., Hewitt, J. W., et al. 2014, ApJ, 786, 145Han, J. L., Manchester, R. N., Lyne, A. G., et al. 2006, ApJ, 642, 868H. E. S. S. Collaboration, Abdalla, H., Abramowski, A., et al. 2018, A&A, 612,A1Leahy, D. A. 2003, ApJ, 586, 224Leahy, D. A., & Ranasinghe, S. 2012, MNRAS, 423, 718MAGIC Collaboration, Aleksi´c, J., Ansoldi, S., et al. 2014, A&A, 571, A96Mattox, J. R., Bertsch, D. L., Chiang, J., et al. 1996, ApJ, 461, 396Morris, D. J., Hobbs, G., Lyne, A. G., et al. 2002, MNRAS, 335, 275Nadezhin, D. K. 1985, Ap&SS, 112, 225Ostriker, J. P., & McKee, C. F. 1988, Reviews of Modern Physics, 60, 1Paredes, J. M., Ishwara-Chandra, C. H., Bosch-Ramon, V., et al. 2014, A&A,561, A56Park, G., Koo, B.-C., Gibson, S. J., et al. 2013, ApJ, 777, 14Paron, S., & Giacani, E. 2010, A&A, 509, L4Paron, S., Giacani, E., Rubio, M., & Dubner, G. 2011, A&A, 530, A25Phillips, J. A., & Onello, J. S. 1993, Massive Stars: Their Lives in the InterstellarMedium, 35, 419Ptuskin, V. S., & Zirakashvili, V. N. 2005, A&A, 429, 755Ranasinghe, S., & Leahy, D. A. 2018, AJ, 155, 204Sutherland, R. S., & Dopita, M. A. 2017, ApJS, 229, 34Stil, J. M., Taylor, A. R., Dickey, J. M., et al. 2006, AJ, 132, 1158Shull, J. M., & McKee, C. F. 1979, ApJ, 227, 131Tang, Y., Yang, C., Zhang, L., et al. 2015, ApJ, 812, 32Torres, D. F., Li, H., Chen, Y., et al. 2011, MNRAS, 417, 3072Willingale, R., Starling, R. L. C., Beardmore, A. P., Tanvir, N. R., & O’Brien,P. T. 2013, MNRAS, 431, 394Wolfire, M. G., McKee, C. F., Hollenbach, D., & Tielens, A. G. G. M. 2003,ApJ, 587, 278Xin, Y.-L., Liao, N.-H., Guo, X.-L., et al. 2018, ApJ, 867, 55Yang, R., Aharonian, F., & Evoli, C. 2016, Phys. Rev. D, 93, 123007Zhou, P., Safi-Harb, S., Chen, Y., et al. 2014, ApJ, 791, 87Zhu, H., Tian, W. W., Torres, D. F., et al. 2013, ApJ, 775, 95Zhu, H., Tian, W., Li, A., et al. 2017, MNRAS, 471, 3494Zirakashvili, V. N., & Ptuskin, V. S. 2008, ApJ, 678, 939y, P., & Kirk, J. G. 1996, A&A, 309, 1002Green, D. A. 2009, MNRAS, 399, 177Green, G. M., Schlafly, E., Zucker, C., et al. 2019, ApJ, 887, 93Hanabata, Y., Katagiri, H., Hewitt, J. W., et al. 2014, ApJ, 786, 145Han, J. L., Manchester, R. N., Lyne, A. G., et al. 2006, ApJ, 642, 868H. E. S. S. Collaboration, Abdalla, H., Abramowski, A., et al. 2018, A&A, 612,A1Leahy, D. A. 2003, ApJ, 586, 224Leahy, D. A., & Ranasinghe, S. 2012, MNRAS, 423, 718MAGIC Collaboration, Aleksi´c, J., Ansoldi, S., et al. 2014, A&A, 571, A96Mattox, J. R., Bertsch, D. L., Chiang, J., et al. 1996, ApJ, 461, 396Morris, D. J., Hobbs, G., Lyne, A. G., et al. 2002, MNRAS, 335, 275Nadezhin, D. K. 1985, Ap&SS, 112, 225Ostriker, J. P., & McKee, C. F. 1988, Reviews of Modern Physics, 60, 1Paredes, J. M., Ishwara-Chandra, C. H., Bosch-Ramon, V., et al. 2014, A&A,561, A56Park, G., Koo, B.-C., Gibson, S. J., et al. 2013, ApJ, 777, 14Paron, S., & Giacani, E. 2010, A&A, 509, L4Paron, S., Giacani, E., Rubio, M., & Dubner, G. 2011, A&A, 530, A25Phillips, J. A., & Onello, J. S. 1993, Massive Stars: Their Lives in the InterstellarMedium, 35, 419Ptuskin, V. S., & Zirakashvili, V. N. 2005, A&A, 429, 755Ranasinghe, S., & Leahy, D. A. 2018, AJ, 155, 204Sutherland, R. S., & Dopita, M. A. 2017, ApJS, 229, 34Stil, J. M., Taylor, A. R., Dickey, J. M., et al. 2006, AJ, 132, 1158Shull, J. M., & McKee, C. F. 1979, ApJ, 227, 131Tang, Y., Yang, C., Zhang, L., et al. 2015, ApJ, 812, 32Torres, D. F., Li, H., Chen, Y., et al. 2011, MNRAS, 417, 3072Willingale, R., Starling, R. L. C., Beardmore, A. P., Tanvir, N. R., & O’Brien,P. T. 2013, MNRAS, 431, 394Wolfire, M. G., McKee, C. F., Hollenbach, D., & Tielens, A. G. G. M. 2003,ApJ, 587, 278Xin, Y.-L., Liao, N.-H., Guo, X.-L., et al. 2018, ApJ, 867, 55Yang, R., Aharonian, F., & Evoli, C. 2016, Phys. Rev. D, 93, 123007Zhou, P., Safi-Harb, S., Chen, Y., et al. 2014, ApJ, 791, 87Zhu, H., Tian, W. W., Torres, D. F., et al. 2013, ApJ, 775, 95Zhu, H., Tian, W., Li, A., et al. 2017, MNRAS, 471, 3494Zirakashvili, V. N., & Ptuskin, V. S. 2008, ApJ, 678, 939