Likely Detection of GeV Gamma-Ray Emission from the Composite Supernova Remnant COMP G327.1+1.1 with Fermi-LAT

COMP G327.1+1.1, as a nonthermal radio source, was reported by Clark et al. (1973, 1975). Its mean angular diameter was 163, and its flux density was 10.6 Jy at 408 MHz, and that at 5000 MHz was 4.3 Jy (Clark et al. 1973). The emission at 408 MHz was described as a shell source with a strong peak, and that of 5000 MHz also showed a similar peak, but did not exhibit a shell structure. At the 843 MHz band observed by the Molonglo Observatory Synthesis Telescope (MOST), its flux was 7.6 Jy and was identified as a composite supernova remnant (SNR) with an unusual off-center plerionic component and a faint shell, with the flux density of the unusual component being 2 Jy (Whiteoak & Green 1996).

Initially, X-ray detection was reported by the Einstein Observatory (Lamb & Markert 1981). Its energy flux of 1–3 keV was (0.81 ± 0.12) × 10⁻¹² erg cm⁻² s⁻¹, and its luminosity was (0.73 ± 0.11) × 10⁻³⁴ erg s⁻¹. For 0.7–2.2 keV from the ROSAT Position Sensitive Proportional Counter observations, 40% of the X-ray counts came from the brighter central region, and the spectrum of the central region was harder than that of the shell of the south (Seward et al. 1996; Sun et al. 1999). Based on the observation data from ROSAT and the Advanced Satellite for Cosmology and Astrophysics, Sun et al. (1999) considered that COMP G327.1+1.1 contained a nonthermal emission component and a diffuse thermal emission component. The nonthermal emission component was from the synchrotron nebula powered by an undiscovered central pulsar, and the thermal component was from the interaction between the shock and interstellar medium. Using the X-ray observation from BeppoSAX, Bocchino & Bandiera (2003) studied the origin of its X-ray emission; they considered COMP G327.1–1.1 to have gone through a longer evolutionary time than previously estimated (1.1 × 10⁴ yr) according to the theoretical model of Sedov or radiative expansion. The X-ray observatories of Chandra and XMM-Newton show a pulsar wind nebula (PWN) inside COMP G327.1-1.1 according to an unusual morphology (Temim et al. 2009).

In the TeV energy band, Acero et al. (2011) reported the hard TeV γ-ray emission inside the source with an integrated energy flux of 1–10 TeV 1.12 × 10⁻¹² erg cm⁻² s⁻¹. Abdalla et al. (2018a) verified that the TeV γ-ray emission was mainly from the PWN G327.15-1.04 within COMP G327.1-1.1 based on both the coincidence of a spatial position and the approximate size of the multiwavelength emission region. In fact, the X-ray and radio observations have also revealed that the morphology of the central PWN is shaped by reverse shock interactions inside the SNR (Temim et al. 2009, 2015; Ma et al. 2016), as reverse shocks inside SNRs can accelerate electrons to high enough energies to produce gamma-ray emission. However, the pulsed emission of the central putative pulsar has not been detected from the radio to high-energy γ-ray bands so far.

Compared with other composite SNRs, COMP G327.1-1.1 is an excellent source to explore the interaction between its host and inside PWN in combination with the multiwavelength observations from radio to γ-ray (Acero et al. 2013). In previous works (Acero et al. 2013, 2016), the GeV γ-ray emission from the source had not been detected by the Fermi-LAT. In this paper, we search for the potential GeV emission from COMP G327.1-1.1 with Fermi-LAT to better understand its nature of emission in the high-energy band.

The paper is organized as follows. The routines of data analysis are given in Section 2, and the related results are presented in Section 3. The likely origin of the GeV radiation is discussed in Section 4, and we conclude in Section 5.

The analysis was performed using the Fermi Science Tools version v11r5p3. ³ We follow the data analysis method as documented in Fermi Science Support Center. ⁴ The Pass 8 data with "Source" event class (evtype = 3 and evclass = 128) and the instrumental response function "P8R3_SOURCE_V2" were adopted. To minimize the contamination from the Earth limb, we excluded events with the zenith angle > 90°. The time range of the photon events is from 2008 August 4 to 2020 October 24, and the energy range is from 800 MeV to 500 GeV considering a larger point-spread function (PSF) in the lower-energy band. The center of the region of interest (ROI) of 20° × 20° square is at (R.A. = 23863, decl. = −5506). ⁵ We used the Fermi Large Telescope Fourth Source Catalog (4FGL; Abdollahi et al. 2020) and the script make4FGLxml.py ⁶ to generate the source model file, which in total included 737 objects within 30° centered at the center of ROI, and then we added a point source with a power-law spectral model in the position of COMP G327.1+1.1. The fit was performed by using the tool gtlike in the Fermi Science Tools package. ⁷ Here, we select to free the normalizations and spectral parameters from all sources within 5°, and the normalizations of the two diffuse backgrounds including iso_P8R3_SOURCE_V2_v1.txt and gll_iem_v07.fits. ⁸

The command gttsmap is used to compute test statistic (TS) maps of 1° × 1° centered at the position of COMP G327.1+1.1 in 0.8–500 GeV. We find significant γ-ray radiation with TS = 22.94 from the direction of COMP G327.1+1.1 as shown in the left panel of Figure 1. Considering the smaller PSF in ≥0.8 GeV (Abdollahi et al. 2020), we do not consider deducting other unknown γ-ray residual emission more than 10 in our analysis. ⁹ After subtracting the point source as well, there is no other significant residual γ-ray radiation from this location as shown in the middle panel of Figure 1. Therefore, the γ-ray emission is likely to come from the region. Next, the best-fit position (R.A., decl. = 238.65, −55.13, with 1σ error radius of 004) from COMP G327.1+1.1 was calculated by running gtfindsrc, and it is adopted to replace the original position of COMP G327.1+1.1; the new γ-ray source is marked as SrcX for all subsequent analyses. We found that the location of COMP G327.1+1.1 is within 2σ error radius, but the TeV counterpart HESS J1554-550 associated with it is within 1σ error radius (Abdalla et al. 2018a), and most of its contours of the brighter center of the 843 MHz radio band are within 2σ error radius as shown in the right panel of Figure 1, which implies that the SrcX is likely to be a counterpart of COMP G327.1+1.1.

Zoom In Zoom Out Reset image size

To check whether SrcX is an extended source, we fit the spatial distribution of the γ-ray emission from the source using uniform disk models. The radius for the disk model was set in a range of 001–005 with a step of 001. The fitting results are shown in Table 1, and we did not find any significant improvement from the uniform disk model by judging TS_ext values ≈0 calculated from 2log(L_ext/L_ps) (e.g., Lande et al. 2012). Here L_ext and L_ps represent the likelihood values for extended uniform disk with radius of 001 and point source, respectively. We used the point-source template to do all subsequent analyses.

Table 1. Spatial Distribution Analysis for SrcX with Two Types of Spatial Models in 0.8–500 GeV

Spatial Model	Radius (σ)	Spectral Index	Photon Flux	TS Value	Degrees of Freedom
	(deg)		(10−10 ph cm−2s−1)
Point source	⋯	2.35 ± 0.24	5.21 ± 1.52	22.94	4
uniform disk	001	2.61 ± 0.07	9.19 ± 0.74	22.06	5
	002	2.61 ± 0.07	9.18 ± 0.74	22.02	5
	003	2.60 ± 0.07	9.17 ± 0.74	21.96	5
	004	2.59 ± 0.07	9.16 ± 0.74	21.87	5
	005	2.58 ± 0.07	9.14 ± 0.74	21.76	5

Download table as:  ASCIITypeset image

3.1. Variability Analysis

To check whether SrcX is a variable source, we generate the light curve of ∼12.2 yr with 20 time bins in the 0.3–500 GeV band. Here, we assess the variability by calculating TS_var defined by Nolan et al. (2012). For the light curve of 20 time bins, TS_var ≥ 36.19 is used to identify variable sources at a 99% confidence level. However, with a TS_var = 19.62 for SrcX, we do not find any significant variability from Figure 2.

Zoom In Zoom Out Reset image size

3.2. Pulsation Search

3.3. Spectral Analysis

For the emission of 0.1–0.3 GeV in the position of SrcX, we find there is no significant emission with TS = 0 by gtlike. The photon flux of the global fit calculated by the binned likelihood analysis method is (2.51 ± 0.31) × 10⁻⁹ ph cm⁻² s⁻¹ with a spectral index of 2.44 ± 0.07 in 0.3–500 GeV. We selected the energy band of 0.3–500 GeV to generate its spectral energy distribution (SED). Here, we choose to divide the energy range of 0.3–500 GeV into three logarithmic bins. Afterward, each energy bin is separately fitted by using the binned likelihood analysis method, as was done in the global fit. For the third bin, an upper limit with a 95% confidence level is given for the TS value ≈0. The SED of SrcX is shown in Figure 3. Meanwhile, we calculate the systematic uncertainty of the first and the second energy bins from the effective area estimated by the bracketing Aeff method. ¹⁰ Considering that SrcX is located in the observation range of the Cherenkov Telescope Array in the south (CTA Consortium et al. 2019), it will be able to be probed in the absence or presence of a cutoff of the γ-ray spectrum as shown in Figure 3. The data from three bins are shown in Table 2.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Table 2. The Energy Flux Measurements from SrcX with Fermi-LAT

E	Band	${E}^{2}{dN}(E)/{dE}$	TS
(GeV)	(GeV)	(10−13erg cm2s−1 )
1.03	0.3–35.57	11.84 ± 3.86${}_{-0.71}^{+0.45}$	9.58
12.25	35.57–42.17	4.04 ± 2.03${}_{-0.37}^{+0.21}$	4.79
145.21	42.17–500	8.04	0.0

Note. Fluxes with uncertainties from the first and the second energy bins are given with TS > 4, and the first and the second uncertainties are statistic and systematic ones, respectively. The flux of the third energy bin is the 95% upper limit.

Download table as:  ASCIITypeset image

4.1. Source Detection

4.2. Model Fitting

Here, we also combine the following three simple stationary models to explain its SED:

• Model 1: One-zone leptonic model.
• Model 2: Two-zone leptonic model.
• Model 3: One-zone leptohadronic model.

Ma et al. (2016) suggested that the emission in the KeV and TeV energy bands could respectively originate from synchrotron and inverse-Compton processes for the same photon indexes from the two energy bands, so we assume that these models all have the same particle distribution of a power law with an exponential cutoff model (PLEC) for radio and γ-ray bands. Here, the form of PLEC is as follows (e.g., Xing et al. 2016; Xin et al. 2019):

$$\begin{eqnarray}&&N(E)={N}_{0}{\left(\displaystyle \frac{E}{{E}_{0}}\right)}^{-\alpha }\exp \left(-\displaystyle \frac{E}{{E}_{\mathrm{cutoff}}}\right),\end{eqnarray} \tag{ 1 }$$

where N₀ is the amplitude, α is the spectral index, E is the particle energy, E_cutoff is the break energy, and E₀ is fixed to 10 TeV. Here, we use NAIMA (Zabalza 2015) to fit a multiwavelength spectrum using the Markov Chain Monte Carlo method in the emcee package (Foreman-Mackey et al. 2013). The CMB is considered in the inverse-Compton scattering process for the above three models. For Model 3, the pion production and cross-section of the proton–proton energy losses are selected from PYTHIA 8 (Kafexhiu et al. 2014).

First, we select Model 1 to fit the SED, and we find that this model cannot well explain the high-energy γ-ray radiation in the GeV band with a reduced χ² ∼ 2.28. The discrepancy may imply a more complicated model for the broadband emission (e.g., multi-emission zones; Xing et al. 2016; Lu et al. 2020). For the case of multi-emission zones, in fact, Lu et al. (2020) considered a two-zone model with different diffusion processes from the SNR extended region and the PWN region to explain the SED of plerionic SNR G21.5-0.9. Here, for simplicity, we consider Model 2 containing two unknown emission regions to explain the SED and find Model 2 can well explain its GeV high-energy γ-ray radiation with a reduced χ² ∼ 1.55. However, the hadronic origin cannot be excluded even if the ambient mediums regarded as a natural target from the COMP SNR are rich, and they can provide subsequent γ-ray production by the interaction of proton–proton collisions (e.g., Temim et al. 2013; Lu et al. 2020). Therefore, we also considered the case of hadrons by using Model 3. Here, we assume the gas density is 10 cm⁻³ (e.g., Xin et al. 2019). With a reduced χ² ∼ 2.06, we found that the model can also explain the broadband observations well. The fitting results of the three models are shown in Figure 4. The fitting parameters for the above three models are given in Table 3.

Table 3. The Best-fit Parameters of Three Models

Model Parameter	B	α	Ecutoff	We (or Wp)	χ2/Ndof
	(μG)		(TeV)	(1048 erg)
Model 1					$\tfrac{29.6}{18-5}=2.28$
Leptonic component	${10.1}_{-1.08}^{+1.55}$	${2.31}_{-0.03}^{+0.03}$	${45.3}_{-11.8}^{+19.4}$	${0.67}_{-0.23}^{+0.17}$
Model 2					$\tfrac{12.4}{18-10}=1.55$
Leptonic component1	${4.14}_{-1.28}^{+1.86}$	${1.57}_{-0.11}^{+0.13}$	${0.30}_{-0.09}^{+0.08}$	${3.13}_{-0.73}^{+0.65}$
Leptonic component2	${9.93}_{-1.4}^{+1.43}$	${2.08}_{-0.25}^{+0.17}$	${38.7}_{-9.09}^{+12}$	${0.21}_{-0.08}^{+0.06}$
Model 3					$\tfrac{20.6}{18-10}=2.06$
Leptonic component	${10.2}_{-1.03}^{+1.62}$	${2.31}_{-0.03}^{+0.03}$	${45}_{-9.51}^{+9.8}$	${0.73}_{-0.07}^{+0.05}$
Hadronic component	⋯	${2.65}_{-0.19}^{+0.10}$	${1}_{-0.5}^{+0.5}$	${9.90}_{-0.75}^{+0.42}$

Download table as:  ASCIITypeset image

However, with fewer data points with a lower confidence level in the GeV band and larger errors in the TeV band, more observations are necessary to infer the likely origin of the high-energy emission in the future.

In this work, we found significant γ-ray emission of 0.8–500 GeV from the region of COMP G327.1+1.1 with ∼ 4σ significance level by using Fermi-LAT Pass 8 data. Meanwhile, we analyzed the variability from its light curve of ∼12 yr, and no significant variability is found with TS_var = 19.62. Its spatial position well coincides with those at the 843 MHz radio and the 0.2–100 TeV energy band. Moreover, the ranges of the spectral index and the GeV luminosity for the new γ-ray source are well consistent with the observed COMP SNRs in the Milky Way. We believe that the emission is probably from COMP G327.1+1.1. On this basis, in combination with the central brighter structure, we suggest the GeV emission may be powered by the PWN. By comparing three radiation models, we found that the two-zone leptonic model can better explain its multiwavelength data. More data in the high-energy band are necessary to firmly confirm the association between the γ-ray source and COMP G327.1+1.1 in the future, which will help us better understand its emission origin and acceleration mechanism.

We sincerely appreciate the referee for invaluable comments, and we gratefully acknowledge Xian Hou from Yunnan Observatories of Chinese Academy of Sciences for her generous help. Meanwhile, we also appreciate the support for this work from National Key R&D Program of China under grant No. 2018YFA0404204, the National Natural Science Foundation of China (NSFC U1931113, U1738211), the Foundations of Yunnan Province (2018IC059, 2018FY001(-003)), and the Scientific research fund of Yunnan Education Department (2020Y0039).