An X-ray study of the supernova remnant G20.0-0.2 and its surroundings
Key Words.:ISM: supernova remnants – pulsars: general – ISM: clouds – Stars: formation
Aims:We study the supernova remnant G20.0-0.2 and its surroundings in order to look for the high energy counterpart of the radio nebula and to find evidence of interaction between the shock front and the interstellar medium.
Methods:We used Chandra archival observations to analyze the X-ray emission from the supernova remnant. The surrounding gas was investigated using data extracted from the Galactic Ring Survey, the VLA Galactic Plane Survey, the Galactic Legacy Infrared Midplane Survey Extraordinaire, and the Bolocam Galactic Plane Survey.
Results:G20.0-0.2 shows diffuse X-ray emission from the central region of the radio remnant. Although the current data do not allow us to distinguish between a thermal or non-thermal origin for the X-ray diffuse emission, based on the radio properties we suggest a synchrotron origin as the most favorable. The hard X-ray point source CXO J182807.4-113516 appears located at the geometrical center of the remnant and is a potential candidate to be the pulsar powering the nebula. We found a molecular cloud adjacent to the flattest border of G20.0-0.2, indicating a probable interaction between the shock front of the remnant and the molecular gas. Several young stellar object candidates are found located in the brightest region of the molecular emission, and over a millimeter continuum source and a dark cloud. This distribution is an indication of an active star forming region around the supernova remnant.
The SNR G20.0-0.2 (hereafter G20) was classified as a plerion111Plerions are also referred as pulsar wind nebulae (PWNe). by becker85 based on the filled-center morphology, the presence of significant polarization at 6 cm, and a flat radio spectral index (). In Fig. 1, we show the best image of the radio continuum emission of the SNR at 20 cm, extracted from the MAGPIS (helfand06). The radio synchrotron emission from G20 has a complex morphology, with multiple features like bright knots, arcs and filaments. The overall emission is dominated by an elliptical central core of about 38 22, with the major axis oriented in the direction of the Galactic plane. This bright feature is surrounded by faint emission. Two arc-like filaments are located near the outer edges of G20 (green arrows in the figure). Interestingly, one border of the SNR is clearly flat and is delineated by a bright radio filament (yellow arrow), suggesting that it may have encountered a higher ambient density in this direction.
PWNe with complex morphologies have been observed in several cases, such as G328.4+0.2 (gelfand07) and G0.9+0.1 (dubner08), but the presence of a flat radio filament is a common characteristic of a surrounding radio shell, with an expected steeper radio spectrum (). However, the other available observations at 330 MHz do not have an adequate angular resolution to look for variations in the radio spectral index through the remnant and confirm the presence of a radio shell.
To the north of G20, lies the complex of ultra-compact HII regions GAL 20.08-0.14 (wood89), which appears as a compact and bright radio source adjacent to the SNR. Different masers and molecular lines have been detected toward the HII regions (avedisova02Cat; galvan09), with velocities between 40 and 45 km s (all velocities are referred to the local standard of rest, hereafter LSR). anderson09a reported the presence of a molecular cloud (MC) that spatially coincides with GAL 20.08-0.14 at a velocity of 42 km s. From an analysis of the HI absorption spectrum, they concluded that the complex of HII regions lies at its far distance, namely 12.6 kpc.
In the X-ray domain, the remnant remains almost unexplored. hands02 performed a X-ray survey of selected regions of the Galactic plane with the XMM-Newton satellite and reported the presence of some diffuse X-ray emission in direction to G20. Additionally, the SNR is the only source that lies partially inside the error box of the Fermi-LAT -ray source J1828.3-1124c (nolan12). For this reason, the authors suggested a connection between the remnant and the -ray emission.
In this work, we present the first X-ray study of the SNR G20.0-0.2 using Chandra observations with the aim to establish if this remnant is purely plerionic or a composite one. We also search for sources candidate to be the pulsar that is powering the emission. In addition, we investigate the interstellar medium (ISM) around G20 looking for signs of interaction with the expanding shock of the SNR, which could explain the peculiar morphology that this remnant shows in the radio band.
2 Results and discussion
2.1 X-ray emission from G20.0-0.2
We reprocessed an archival Chandra observation of G20 obtained on 2005 November 5 (ObsID: 5563) using ACIS-I in vfaint mode. The data were calibrated with CIAO (version 4.4) and CALDB (version 4.4.8), provided by the Chandra X-ray Center (CXC). After filtering the periods of high count-rate, we produced a cleaned event file with 34.45 ks of observation, which was used to perform an imaging and spectral analysis.
We analyzed the event file and found an enhancement of the X-ray emission toward the central radio core of G20. We constructed exposure corrected images in different energy bands and found diffuse emission above 2 keV and a single point source embedded on it. Additionally, we noticed several X-ray clumps outside this extended emission. Most of them are likely point sources that appear broadened by the smoothing. To emphasize the diffuse emission, we produced a point sources subtracted image following the standard procedure described in the Ciao Science Threads. We used the tool wavdetect to detect the point source candidates and we visually inspected the result to exclude those sources that appear to be actual diffuse emission. For each source, we created the source and background regions with the tool roi and replaced the source region with the mean background region using dmfilth. To increase the signal-to-noise ratio of the image, we convolved it with a Gaussian kernel. The final image has a spatial scale of 0.05 arcmin/pixel. Fig. 2 shows the point sources subtracted X-ray image of the SNR between 2.0 and 7.0 keV, overlayed with some contours of the radio continuum emission. We noticed that the diffuse emission is confined within an ellipse of 18 11, which is enclosed by the central bright radio core and is elongated in the same direction as in the radio band.
Fig. 3 shows an X-ray image in the 2.0-7.0 keV energy band toward the diffuse emission. Interestingly, we noted the presence of only one X-ray point source embedded in the diffuse X-ray emission (indicated in the figure with a yellow plus sign). This source has been identified in the Chandra Source Catalog (CSC, evans10) as CXO J182807.4-113516 and is centered at R.A. = 18287.43, Dec. = -11351632, almost at the geometrical center of the radio remnant. In Sect. 2.1.2, we will discuss about its nature and the connection with the remnant.
2.1.2 X-ray spectral analysis
To perform a spectral analysis of the diffuse emission from the SNR, we extracted the spectrum with the CIAO task specextract from the elliptical region indicated in Fig. 3. This region excludes some of the diffuse emission but reduces the contamination from the background. The point source CXO J182807.4-113516 was removed from the extraction region. For the background, we selected a circular region free of diffuse emission and point sources. The spectrum was binned with a minimum of 15 counts/bin. We extracted a total of 1059 counts between 1.0 and 7.0 keV. We fitted the spectrum using XSPEC (version 12.7.1) and statistics.
The Chandra ACIS-I spectrum of G20 presents a continuum without evidence of emission lines (Fig. 4). We fitted it with both non-thermal and thermal models. To probe a possible synchrotron origin, we fitted the spectrum with an absorbed power-law. For the absorption model (wabs), we fixed the hydrogen column density to the value N cm, obtained through the analysis of the neutral and molecular gas (Sect. 2.2). The best fit yields a photon index of 1.87, a value similar to that observed in other PWNe (kargal10). Letting the absorption to vary freely does not produce any significant improvement in the fitting and results in N cm, a value that is still comparable to the one measured using the neutral and molecular data.
|Model||N||T||F (1.0-7.0 keV)||L (1.0-7.0 keV)|
|[ cm]||[keV]||[ erg cm s ]||[ erg s]|
Regarding the thermal models, a thermal plasma in collisional equilibrium (apec) requires a considerably high electron temperature for a SNR (T19 keV) when the abundance parameter is kept equal to the solar value of anders89, and a temperature of 7 keV when the abundance is a free parameter. In this last case, although the temperature is not extremely high, the best fit abundance yields an unrealistic low value (10). For a non-equilibrium plasma (nei), freezing the abundance to the solar value results in T5 keV, which approaches the expected temperature of a young SNR (vink12). The obtained ionization parameter ( cms) also points to a plasma out of equilibrium. In Table 1, we list the results of the spectral fitting for the power-law and the nei models.
In conclusion, the spectral fits themselves do not preclude a thermal origin for the X-ray emission from G20. However, based on the radio properties, we propose that a synchrotron origin for the high energy emission is the most probable. The mere presence of a PWN component indicates that there is a compact source powering it. The best candidate to be a neutron star is CXO J182807.4-113516, the only X-ray point source embedded in the diffuse emission and located at the geometrical center of the SNR. There is no radio counterpart in the VLA image, nor there is any optical or infrared source within the position error box in the USNOb1.0, 2MASS222The Two Micron All Sky Survey (2MASS) is a joint project of the University of Massachusetts and the Infrared Processing and Analysis Center/California Institute of Technology, funded by the National Aeronautics and Space Administration and the National Science Foundation., and Spitzer catalogs. The closest point sources are USNOb1.0 0784-0429139, 2MASS 18280728-1135224, and Spitzer G019.9794-00.1736, located at 9, 6, and 6 from CXO J182807.4-113516, respectively. These distances are larger than three times the combined X-ray and optical/infrared positional error.
Although the low number of counts (20) from CXO J182807.4-113516 does not allow to perform a spectral fitting or timing analysis to investigate its nature, we can still use the poor spectral information to get a rough picture of this source. We found that all the counts from CXO J182807.4-113516 originate between 3.0 and 8.0 keV, so we defined the hardness ratio as the ratio of the counts in the 5.0-8.0 keV band to the counts in the 3.0-5.0 keV band. We extracted the net (i.e., background subtracted) number of counts and obtained 11.823.75 counts between 3.0 and 5.0 keV and 6.332.73 counts between 5.0 and 8.0 keV, which yields to H=0.540.29. This value can be compared with the predicted value of estimated by assuming different emission models. The X-ray emission from pulsars (PSRs) may correspond either to blackbody emission from the neutron star surface or to non-thermal emission from the magnetosphere. Using the tool W3PIMMS333http://heasarc.gsfc.nasa.gov/Tools/w3pimms.html and the net count rate between 3.0 and 8.0 keV (0.00053 cts/s), we modeled the X-ray emission from CXO J182807.4-113516. We fixed the hydrogen column density to cm, to match the absorption of the diffuse emission. For a blackbody with a temperature of 0.1 keV, the predicted is 8, in complete disagreement with the measured value. For a hot blackbody (1 keV), we obtained , but such high temperature is not expected in a young PSR. According to kargal10, the non-thermal spectrum of a PSR powering a PWN is well described by a power-law model with photon index in the range . The predicted values of are 0.59 (for ) and 0.39 (for ), in good agreement with the measured value of H=0.540.29. Thus, we conclude that the emission from CXO J182807.4-113516 is more likely non-thermal in origin, although a deeper X-ray exposure of this source is needed to confirm its nature. Using W3PIMM, we estimated its luminosity in the 0.5-8.0 keV band for a distance of 4.5 kpc, obtaining 0.7 erg/s (for ) and 1.2 erg/s (for ). From the best fit power-law model, the X-ray luminosity of G20 between 0.5 and 8.0 keV is 1.8 erg/s. The luminosities of G20 and CXO J182807.4-113516 are in good agreement with those reported by kargal10 for other PWNe and their powering PSRs. Thus, based on its position and spectral behavior, CXO J182807.4-113516 appears as a good candidate to be the pulsar created after the explosion of the supernova that originated G20 and that is powering the synchrotron nebula.
2.2 The interstellar medium around G20.0-0.2
We studied the ISM around G20 to investigate the presence of molecular clouds (MCs) that may have affected the expansion of the SNR shock front. The molecular data were extracted from the Galactic Ring Survey (GRS, jackson06). The survey maps the Galactic ring in the CO J=1–0 line with angular and spectral resolutions of 46″ and 0.2 km s, respectively. The observations were performed in both position-switching and on-the-fly mapping modes, achieving an angular sampling of 22″.
By inspecting the whole CO cube, we found a molecular feature extending from 62 to 71 km s as the most probable structure to be in contact with the SNR. In Fig. 5, we present the CO emission integrated every 1.05 km along this velocity range, displayed in Galactic coordinates. We noted the presence of a cloud that shows an arc-shape morphology and delineates the southern border of the SNR. This cloud is composed of two molecular clumps, labeled A and B. Molecular clump A lies right upon the flattest border of G20 (refer to Fig. 1), while clump B is located adjacent to the SE border of the SNR. A third molecular structure (labeled clump C) appears in the direction of the flattest border, but is not in contact with the radio emission.
The molecular clump A spatially coincides with the Bolocam444The Bolocam Galactic Plane Survey (BGPS) is a 1.1 mm continuum survey of the Galactic Plane made using Bolocam on the Caltech Submillimeter Observatory, with a 33 FWHM effective resolution (aguirre11). millimeter continuum source BGPS G19.926-0.257. Besides, this source lies within the ellipse of the Spitzer dark cloud SDC G19.928-0.257 (peretto09). In Fig. 6, we show the excellent spatial correspondence among the molecular emission, the millimeter continuum source, and the dark cloud.
Analyzing the CO spectra toward the borders of the SNR, we found some interesting spectral features. Fig. 7 displays an averaged spectrum obtained from a region toward the SNR’s flattest border, centered at l=19.944, b=-0.245. The spectrum is not symmetric and presents a slight spectral shoulder toward higher velocities. The red curve in the figure is the result of the fitting with two Gaussian functions with central velocities and FWHM v of 65.3 and 68.5 km s, and 3.8 and 5.0 km s, respectively. Such asymmetry could be evidence of turbulent motion in the gas, maybe produced by the SNR shock (see e.g. falgarone94), although we can not discard the presence of multiple molecular components.
Assuming that 66 km s is the central velocity of the cloud, according to the Galactic model of fich89, this cloud could be located at the distances of either 4.5 and 11.5 kpc. To resolve the ambiguity, we followed the procedure described in roman09. Molecular clouds have HI embedded within them, which is cooler than the Galactic inter-cloud HI. Therefore, a MC located at the near distance absorbs the radiation from the warm HI at the far distance with the same velocity. As a consequence, the HI spectrum toward a MC at the near distance shows an absorption feature whose velocity coincides with the velocity of the molecular emission line.
We constructed an HI absorption spectrum toward the clump A, which presents the most convincing signature of interaction with the SNR. The HI data were extracted from the VLA Galactic Plane Survey (VGPS, stil06), which maps the HI 21 cm line emission with angular and spectral resolutions of 1 and 1.3 km s, respectively. We obtained the CO and HI on-position spectra from a rectangular region within the 19 K km s contour level (see Fig. 6, left) and outside the SNR to avoid the effect of the radio continuum source in the absorption spectrum. The off-position HI spectrum was made by averaging three HI spectrum from rectangular regions adjacent to clump A and free of CO emission. The HI absorption spectrum results from the subtraction of the HI on- and off-position spectra. In Fig. 8, we show the HI and CO spectra toward the molecular clump A. We noted the presence of an HI absorption line at a velocity of 65 km s, which coincides with a CO emission line from the cloud. From this analysis, we favor the near distance of 4.5 kpc for molecular clump A and thus for the SNR G20.0-0.2.
An additional evidence that the near kinematic distance could be the most likely one is the presence of the millimeter source BGPS 019.926-00.257 lying inside the Spitzer dark cloud SDC G19.928-0.257, both of them associated with the molecular clump A (see Fig. 6, right). In fact, dark clouds are clouds of dust that appear dark as they absorb the mid-infrared radiation of the Galaxy. For this reason, they are usually placed at the near kinematic distance (schlin11). Thus, we consider that the SNR (located at 4.5 kpc) and the UC HII region complex GAL 20.08-0.14 (at 12.6 kpc) are not related.
To estimate the mass and density of the molecular clumps, we assumed local thermodynamic equilibrium (LTE). For the CO column density we used:
where is the excitation temperature of the CO transition and is the optical depth of the line. Assuming that the CO J=1–0 line is optically thin, we used the approximation
where , K is the background temperature, and is the brightness temperature of the line. To obtain the H column density, we took the relative abundance from simon01: N(H)/N(CO). The mass of the molecular clumps was calculated from
where is the solid angle subtended by the CO J=1–0 beam size, is the hydrogen mass, is the mean molecular weight assumed to be 2.8 by taking into account a relative helium abundance of 25 %. Our summation was performed over the area of each molecular clump. The column density was calculated assuming K and the mass and number density considering a distance of 4.5 kpc. In Table 2, we report the physical parameters of the molecular clumps A and B. The obtained masses and densities are similar to those measured toward other molecular clouds interacting with SNRs (see paron12; dubner04).
|[km s]||[km s]||[arcmin (pc)]||[ cm]||[M]||[cm]|
Additionally, we calculated the total hydrogen column density N(H), which is an important parameter in the modelling of X-ray emission (see Sect. 2.1.2). For this purpose, we summed the contributions from both neutral and molecular hydrogen column densities: N(H) = N(HI)+2N(H). The integration was performed in the velocity range between 0 and 66 km s and over the central core, in a region delimited by the 3.0 mJy/beam contour level of the radio continuum emission (see Fig. 2), where the diffuse X-rays originate. The molecular column density N(H) was obtained from Eq. 1 and the neutral column density N(HI) from the following equation:
We obtained N(H) cm.
2.3 Young stellar objects embedded in the molecular gas
The expansion of the shock front of a SNR into the ISM gives rise to a broad range of phenomena. SNRs can compress molecular clouds and induce the formation of inhomogeneities in the gas, which may collapse into dense molecular clumps where, eventually, new stars may form. Up to the present, there is no concluding observational evidences indicating that the SNRs can trigger star formation. Only a few studies have revealed the existence of young stellar objects (YSOs) in the periphery of SNRs, such as the filled-center remnants G24.7+0.6 (petriella10) and G54.1+0.3 (koo08), and the shell-type G59.5+0.1 (xu12), G0.1-0.1, G6.41-0.1 (W28), and G355.9-2.5 (marquez-lugo10). The main controversy regarding star formation triggered by a SNR is the difference between two timescales, namely the SNR’s age (10 yr) and the YSO’s characteristic age (10 yr), which would make it impossible to directly observe a SNR inducing the collapse of molecular cores to form new stars. However, the progenitors of core-collapse SNs are massive stars, whose strong stellar wind during the duration of the main sequence phase (10 yr) can provide additional injection of energy into the molecular gas prior to the SNR.
In the case of G20, the radio and X-ray properties indicate that it is a core-collapse SNR. In addition, we showed that the remnant is very likely interacting with a MC. So, it appears as a proper candidate to study star forming activity in its vicinity. To look for YSOs candidates around G20, we constructed a color-color (cc) diagram in the Spitzer-IRAC bands [5.8]-[8.0] vs [3.6]-[4.5] (not shown here) with sources that report flux measurements in the four bands. We applied the criterion of allen04 to identify class I YSO candidates, i.e. young stars at the earlier stages of evolution. In Fig. 9, we show the distribution of the class I YSO candidates, overimposed to the molecular cloud probably interacting with G20. From this figure, we noted the presence of several clusters of YSO candidates, but we call special attention to a group of 6 YSO candidates that appear projected on the clump A, which is associated with the millimeter source BGPS G19.926-0.257 and with the dark cloud SDC G19.928-0.257. The presence of dense molecular material, a dark cloud, abundant dust (traced by the millimeter source), and YSO candidates suggests that this region could be an active star forming site (rath07; rath06), representing another potential site of star formation near a SNR.
3 Concluding remarks
We have presented the first X-ray study of the SNR G20.0-0.2 using Chandra observations. We detected diffuse emission, which has a very good correlation with the bright central radio emission. The X-ray spectrum can be equally well fitted by a thermal plasma in a non equilibrium or a power-law model. Taking into account the obtained photon index and the radio properties of the SNR, we conclude that a non-thermal origin of the X-ray emission is the most probable. In addition, we have reported the presence of the hard X-ray point source CXO J182807.4-113516 located at the geometrical center of the remnant, which is characterized by a non-thermal spectrum and a X-ray luminosity between erg/s in the 0.5-8.0 keV band (similar to young pulsars), suggesting that it is a good candidate to be the central compact source powering the PWN.
The study of the surrounding interstellar gas around G20 allowed us to constrain its distance to a value of 4.5 kpc. We found a molecular cloud with some indications of interaction with the flattest border of the radio emission, and in excellent spatial correspondence with the millimeter continuum source BGPS G19.926-0.257 and the dark cloud SDC G19.928-0.257. We identified a group of class I YSO candidates located in the brightest region of the molecular cloud. The presence of these young sources, a dark cloud and abundant dust suggest that the region around G20 could be a potential star forming site.