Star Forming Dense Cloud Cores in the TeV -ray SNR RX J1713.7-3946
RX J1713.7-3946 is one of the TeV -ray supernova remnants (SNRs) emitting synchrotron X rays. The SNR is associated with molecular gas located at 1 kpc. We made new molecular observations toward the dense cloud cores, peaks A, C and D, in the SNR in the CO(=2-1) and CO(=2-1) transitions at angular resolution of 90”. The most intense core in CO, peak C, was also mapped in the CO(=4-3) transition at angular resolution of 38”. Peak C shows strong signs of active star formation including bipolar outflow and a far-infrared protostellar source and has a steep gradient with a variation in the average density within radius . Peak C and the other dense cloud cores are rim-brightened in synchrotron X rays, suggesting that the dense cloud cores are embedded within or on the outer boundary of the SNR shell. This confirms the earlier suggestion that the X rays are physically associated with the molecular gas (Fukui et al. 2003). We present a scenario where the densest molecular core, peak C, survived against the blast wave and is now embedded within the SNR. Numerical simulations of the shock-cloud interaction indicate that a dense clump can indeed survive shock erosion, since shock propagation speed is stalled in the dense clump. Additionally, the shock-cloud interaction induces turbulence and magnetic field amplification around the dense clump that may facilitate particle acceleration in the lower-density inter-clump space leading to the enhanced synchrotron X rays around dense cores.
Supernova remnants (SNRs) have a profound influence on the interstellar medium (ISM) via shock interaction and injection of heavy elements. The dynamical interaction also affect the evolution of SNRs through the distortion of the shell morphology, if the ISM is dense enough. It is therefore important to study the detailed physical properties in the interaction between SNRs and ISM, in oder to understand what can occur.
RX J1713.7-3946 (G347.3-0.5) is one of the high energy ray SNRs associated with dense molecular gas and is a good target to study the interaction. The satellite detected the SNR at X rays for the first time (Pfeffermann Aschenbach 1996). Koyama et al. (1997) showed that the X ray emission is synchrotron emission without thermal features using the satellite and derived a distance of 1kpc based on a relatively small X ray absorption, corresponding to an HI column density of (6.21)10 cm for the northwestern rim. Slane et al. (1999), on the other hand, claimed a distance of 6 kpc based on three CO clouds of around of -95 km s observed at 8.8 arcmin resolution, which were suggested to be interacting with the SNR by the authors. They argued that the small X ray absorption is ascribed to an HI hole that happens to be located toward the SNR.
Subsequently, Fukui et al. (2003) revealed that molecular gas at a much lower of -11 km s to -3 km s is interacting with the SNR and derived a kinematic distance of 1 kpc by using the NANTEN CO dataset at a higher angular resolution of 2.6 arcmin (Mizuno Fukui 2004). Figure 1 shows the CO distribution in this velocity range overlayed with the X ray distribution, and demonstrates that the molecular gas shows a clear anti-correlation with the X ray distribution and that intense molecular peaks are spatially correlated with the X ray peaks. Observations with the satellite derived a distance around 1.3 kpc from a study of X ray absorption at a higher angular resolution than (Cassam-Chenai et al. 2004). Moriguchi et al. (2005) showed further details of the CO(=1-0, 3-2) distributions and confirmed the association between the SNR and the low velocity molecular gas. Aharonian et al. (2004, 2006 and 2007) revealed a shell like TeV distribution with the H.E.S.S. Cerenkov ray telescope and showed that the SNR is a very energetic emitter of TeV rays. They compared the rays with the NANTEN CO distribution and showed that the CO and rays exhibit a fairly good correlation particularly toward the molecular peaks in the northwestern rim, supporting the physical association between the SNR and the molecular gas. Tanaka et al. (2008) presented a detailed comparison of X ray data with TeV rays and discussed the origin of the rays, whereas the connection between high energy photons and the molecular gas remained unexplored. To summarize, RX J1713.7-3946 shows a close connection among X rays, rays and molecular gas, a unique case where high energy particles, either protons or electrons, and the low energy ISM are co-existent and physically interacting (Fukui et al. 2008).
In the course of the above studies Fukui et al. (2003) found that CO peak C shows broad CO wings and suggested that the wings may result from dynamical acceleration by the SNR blast wave. Such broad molecular wings are found in several SNRs including IC443, W44, W28, etc. (e.g., Denoyer et al. 1979, Wootten 1977, 1981). Moriguchi et al. (2005) showed that the CO(=3-2) distribution in peak C shows a hint of a bipolar nature being associated with an infrared compact source with the spectrum of a protostar. This may indicate an alternative possibility that the broad wings are driven by the outflow from a protostar but are not due to the shock interaction. It is an open question whether the SNR is accelerating molecular gas to high velocities in peak C. It is therefore important to clarify whether the CO broad wings are due to the blast-wave acceleration or due to protostellar activity in our efforts to better understand the interaction.
We have carried out new observations of the molecular peaks associated with the X rays in the CO(=2-1, 4-3) and CO(=2-1) transitions at mm and sub-mm wavelengths in order to derive the molecular properties of the core and broad wings in peak C and the other cloud cores nearby in RX J1713.7-3946. We present these molecular results and discuss star formation in the SNR in this paper. We will publish separate papers that deal with a detailed X ray analysis of results, a comparison with TeV rays and an analysis of the results toward the molecular peaks. In the following we shall adopt a distance 1 kpc and assume that the SNR has an age of 1600 yrs, a radius of 7 pc and an expansion speed of 3000 km s (Uchiyama et al. 2007). This is consistent with the SNR corresponding to a historical SNR recorded in the Chinese literature in AD393/4 (Wang et al. 1997).
The present paper is organized as follows. Section 2 describes the observations and Section 3 their results. Section 4 gives an analysis of density and temperature. The paper is summarized in Section 5.
We carried out CO(=2-1, 4-3) and CO(=2-1) observations with the NANTEN2 4 m sub-mm telescope of Nagoya University installed at Pampa La Bola (4865 m above the sea level) in the northern Chile.
Observations in the CO(=2-1) and CO(=2-1) lines were conducted from August to November and in December 2008, respectively. The backend was a 4 K cooled Nb SIS mixer receiver and the single-side-band (SSB) system temperature was 250 K for both transitions, including the atmosphere toward the zenith. The telescope had a beam size of 90” at 230 GHz. The pointing was checked by observing the Jupiter every two hours and was found to be as accurate as 15”. We used acoustic optical spectrometers (AOS) with 2048 channels having a bandwidth of 390 km s and resolution of 0.38 km s. Observations in CO were carried out in the on-the-fly (OTF) mode, scanning with an integration time of 1.0 to 2.0 sec. The chopper wheel method was employed for the intensity calibration and the derived Trms was better than 0.66 and 0.51 K ch with 1.0 and 2.0 sec integrations, respectively. The observed area was about 2.25 square degrees and is shown in Figure 1. Observations in CO were also carried out in the on-the-fly (OTF) mode for the area of 22 square arcminutes including the peaks A, B, C, and D (Moriguchi et al. 2005) with an integration time of 2.0 sec with as Trms better than 0.68 K ch. The absolute intensity was calibrated by observing Oph EW4 [ (J2000)] (Kulesa et al. 2005).
Observations in CO(=4-3) were carried out from November to December 2007 covering a 9 square arcminutes region including peak C and toward a point in peak A. The telescope had a beam size of at 460 GHz as measured by observing Jupiter. The front end was a SIS receiver having SSB temperature of 300 K including the atmosphere toward the zenith. The typical rms noise fluctuations were 0.28 K ch. The absolute intensity calibration is as described by Pineda et al. (2008).
Together with the data above, we made use of the data by Moriguchi et al. (2005). They observed the peaks A, C, and D with the ASTE sub-mm telescope in CO(=3-2) in November, 2004. The data were taken by a position switching mode with a 30” grid spacing with 23” beam. The spectrometer was an AOS with 450 km s band width and 0.43 km s resolution. The system temperature was 300 - 400 K (DSB) and the typical rms achieved is 0.4 - 0.9 K with 30 sec integration.
3.1 Co(=2-1) and Co(=2-1) distributions
Figures 2a and 2b show the CO(=2-1) and CO(=2-1) distributions, respectively, of the western rim of the SNR in the box shown by dashed lines in Figure 1. There are nearly ten CO peaks and the observational parameters of nine of them are given in Table 1 as taken from Fukui et al. (2003) and Moriguchi et al. (2005). Three of the cores marked by crosses in Figure 2a, peaks A, C and D, are associated with point sources with a protostellar spectral energy distribution (SED) (Moriguchi et al. 2005). Bright cores in CO are all associated with sources and we mapped them in the CO(=2-1) transition for the two areas shown by solid lines in Figure 2. All the cores with sources are detected in the CO transition and peak C is the most intense among them.
3.2 Co(=4-3) distribution
We observed the CO(=4-3) transition in an area of 3 arcmin 3 arcmin around peak C in equatorial coordinates and toward the peak position of peak A. Figure 3 shows four images of peak C in the CO(=2-1, 3-2, and 4-3) and CO(=2-1) transitions, where the CO(=3-2) distribution is taken from Moriguchi et al. (2005). The CO(=4-3) core is most compact and the size of the core increases toward the lower transitions, suggesting a sharp density decrease with radius, since the higher transitions have higher critical densities for collisional excitation.
3.3 Co(=4-3) broad wings
Figure 4 shows that the broad CO(=4-3) wings first detected by Fukui et al. (2003) reveal a clear bipolar signature centered on 17089-3951 (Table 2) and on the peak position of the dense cloud core in peak C. The bipolarity verifies that the wings are driven by a protostar and are not driven by the SNR blast wave. The source position also shows a good correlation with an extended MSX sources at 8.28 m (from IPAC Infrared Science Archive) (Figure 4a).The wings toward peak C are also recognized in the present CO(=2-1) data in addition to the CO(=1-0) and CO(=3-2) transitions (Moriguchi et al. 2005), whereas the wing intensities of these lower- transitions are more than a few times weaker than the CO(=4-3) wings. It is not clear if the other peaks A and D show sings of bipolar outflow either in the CO(=1-0) and CO(=3-2) data (Moriguchi et al. 2005) or in the present CO(=2-1) data.
3.4 Comparison with X rays
Figure 5(a) shows an overlay of the CO(=2-1) image with the X ray image in the 1-5keV range over the western rim of the SNR (Tanaka et al. 2008). As noted earlier in Moriguchi et al. (2005), in the lower resolution CO(=1-0) image at 2.6 arcmin resolution, and the CO(=3-2) distributions of more limited coverage at 30 arcsec resolution for individual peaks A, C and D, the X ray distribution shows a clear correlation with the molecular distribution. First, peak C is surrounded by bright X ray emission both on its north and south with a local minimum toward the center of the core. Figure 5(b) shows a remarkable coincidence between the X ray depression and the cloud core at a 0.1 pc scale. This depression is not due to interstellar absorption because we find similar distribution of X ray in the higher energy band (5-10 keV) that is hardly absorbed (Figure 5(c) and (d)). Therefore, this morphology suggests that the X ray emission is enhanced on the surface of the cloud core. Peak D is also associated with bright X ray emission, whereas only the eastern part, the inner side of the shell, is emitting the X rays. This is ascribed to that peak D is located on the outer surface of the SN shell and only the inner side is being illuminated by the shock. Another X ray feature is toward peaks A and B where the X rays are bright between the two peaks, showing a similar anti-correlation. Such a trend is also seen in the middle of the SNR rim between peaks E and G and between peaks E and I. All these features suggest that the X ray emission is enhanced on the surface of the molecular gas.
4 Analysis of molecular properties
4.1 LVG analysis
We shall use the large velocity gradient model of line radiation transfer to estimate density and temperature from the muti- transitions of CO; i.e., CO(=3-2, 4-3) and CO(=2-1). We applied the large velocity gradient (LVG) analysis (Goldreich Kwan 1974; Scoville Solomon 1974) to estimate the physical parameters of the molecular gas toward peaks A and C by adopting a spherically symmetric uniform model having a radial velocity gradient dv/dr. The CO(=2-1) transition was not included in the analysis because the transition may be subject to self absorption due to low excitation foreground gas (e.g. Mizuno et al. 2010). We calculate level populations of CO and CO molecular rotational states and line intensities. The LVG model requires three independent parameters to calculate emission line intensities; i.e., kinetic temperature, density of molecular hydrogen and X/(dv/dr). X/(dv/dr) is the abundance ratio from CO to H divided by the velocity gradient in the cloud. We use the abundance ratio [CO]/[CO] 75 (Gsten et al. 2004) and [CO]/[H] 510 (Blake et al. 1987), and estimate the mean velocity gradient between the peaks as 12.5 km s pc. Accordingly, we adopt that X/(dv/dr) is 4.010 (km s pc) for CO.
In order to solve temperatures and densities which reproduce the observed line intensity ratio, we calculate chi-square defined as below;
where is the observed line intensity ratio between different excitation lines or different isotopes, is the line ratio of the LVG calculations, is the standard deviation for in the analysis. The error in the observed intensity is estimated by considering the noise level of the observations and the calibration error. We assume that the error of calibration from to is 10 for all the line intensities. The data used are derived from the line profiles in Figure 6 upper and three ratios are estimated for a 1.5 km s velocity interval in the three peaks as listed in Table 3.
The lower panels of Figure 6 shows the results of fitting to the data obtained with a chi-square minimization approach to find the solution of temperature and density. Each locus of a black solid line surrounding the cross indicates the chi-square , which corresponds to the 95 confidence level of a chi-square distribution. The crosses denote the lowest point of chi-square. Additionally, we are able to reject each region outside the black solid line at the 95 confidence level. Table 3 summarizes the results of the LVG analysis. Density and temperature are relatively well constrained in peaks A and C. The temperature of peaks A and C is in a range of 10 - 12 K and density is somewhat higher in peak C, 10 cm, than in peak A, 610 cm.
4.2 The density distribution of peak C
Peak C is associated with a dense cloud core with a strong intensity gradient (Figure 3). The total molecular mass of the core is estimated to be 400 M from the CO(=1-0) integrated intensity (Moriguchi et al. 2005) for an X factor of 2.0 10 [(CO)/(K km s)](Bertsch et al. 1993). We shall derive the density distribution by employing a simple power-law analysis assuming a spherical symmetry. First, we de-convolve the intensity distribution in order to correct for the beam size. We assume the relation to derive a de-convolved core radius in each transition; = + , where , and stand for the observed radius, the beam radius and the de-convolved radius, respectively, at a half power level of the peak intensity (Table 4). Then, we estimate the averaged density within , , so as to match the observed averaged integrated intensity to the LVG estimate within the radius by assuming kinetic temperature of 12 K and the same model parameters in Section 4.1. Considering the low luminosity of the source (300 ), we infer that the local temperature variation in the core is not significant and that a uniform temperature is a good approximation. The result is shown in Figure 7. This presents that the average density distribution is well approximated by a power law, . Such a steep density gradient is consistent with a star forming cloud core (Larson 1969, Penston 1969, Lizano and Shu 1989, Onishi et al. 1999). The line width does not vary much among the different density regimes as shown in Figure 7, suggesting the infall motion is not very large. An alternative interpretation for the density distribution will be discussed later in Section 5.
4.3 Physical parameters of the outflow
The bipolarity of the broad CO wings in Figure 4 verifies that the wings are caused by bipolar outflow driven by a protostar and is consistent with the compact dense core with a steep density gradient (Section 4.2). The physical parameters of the outflow are estimated by using the method described by Moriguchi et al. (2005) (Table 5). The average intensities of the red component are 8.9 K and 2.0 K for the CO(=2-1) and CO(=4-3) transitions in a velocity range from -8 km s to -3 km s. The density in the line emitting wings are then estimated to be cm from the line intensity ratio, 0.23, between the CO(=2-1) and CO(=4-3) transitions. The beam filling factor is estimated to be as small as 5 for an assumed excitation temperature of 12 K for the rotational levels of CO. This result is similarly obtained from the blue component, that yields beam filling factor of . Therefore, the CO wings represent highly clumped gas rather typical for outflows. The present CO(=4-3) data have verified that the wings represent outflow and showed that the CO(=4-3) wings are much more compact than the CO(=1-0) wings, indicating that the CO(=4-3) wings are denser outflow gas whose spatial extent is smaller than the lower density wings (Table 5, for the CO(=1-0) wings see Table 4 of Moriguchi et al. 2005). The most likely driving source of the outflow is the far infrared source 17089-3951. General properties for bipolar outflows are found elsewhere (e.g., Lada 1986, Fukui 1989, Fukui et al. 1993) and the present outflow is seen to be typical of an outflow associated with a low mass protostar in terms of its size and velocity.
4.4 An evolutionary scheme
We present a general scenario following discussions given by Fukui et al. (2003) and Moriguchi et al. (2005). The progenitor of the SN was a high mass star formed some Myrs ago in the region, and which created a cavity in the ISM by its stellar wind. This is a star-forming region with a loose spatial association extended over a few 10 pc. The molecular peaks, including A, B, C, D, and the others are in part of the cavity wall. The total mass of the molecular complex is at least a few 1000 (Moriguchi et al. 2005). The star formation in the peaks A, C, and D may have been triggered by the progenitor of the SNR (Koo et al. 2008, Desai et al. 2010). The progenitor star caused the supernova explosion 1600 yrs ago, as recorded in an ancient Chinese document (Wang et al. 1996) and the blast wave expanded into the cavity. The SNR is now interacting with the molecular gas and this interacting layer is traced by the X rays, as supported by the good spatial anti-correlation with the molecular gas at a 0.1 pc scale.
The SNR with an age of 1600 yrs is still in the free expansion phase for ambient density less than 1 cm (e.g., Truelove and McKee 1999), while the expansion may experience a sudden deceleration when it hits the dense ISM wall. The interactions of such a young SNR with the interstellar medium are numerically simulated by Inoue, Yamazaki Inutsuka (2009). These authors postulated an inhomogeneous initial density distribution in a range of 1 - 1000 cm with a speed of the blast wave of 1000 km s and an age of 1000 yrs. The density distribution is basically consistent with the observed clumped nature of the CO gas. Based on the numerical results, we are able to estimate the disturbing depth of the blast wave into the interacting medium as a function of the initial density. It is found that the penetrating velocity of the shock front into a clump is inversely proportional to , where is the density of the clump. From the results of the numerical simulations, we find that the velocity is given as 3000 km s/, where the ambient density =1 cm in the present SNR. The penetrating depth in a typical timescale of the interaction 1000 yrs is then estimated to be 0.1 pc for =10 cm and 3 pc for =1 cm, respectively. Peak C has a present diameter of at least 0.6 pc at density of 10 cm (Figure 7) and should have not been affected significantly by the shock penetration according to the argument above, whereas the ambient lower density gas is significantly disturbed and accelerated in a scale length of the SNR radius. We, therefore, infer that peak C is able to survive against the shock wave as originally suggested by Fukui et al. (2003), while the lower density ambient gas is significantly disturbed and accelerated .
It seems essential in the interaction that the initial density distribution in and around the stellar wind cavity was highly inhomogeneous. This inhomogeneity is indeed verified by the clumpy molecular distribution in the present region. According to the theoretical results (Inoue et al. 2009), dense clumps tend to retard shock fronts. The global morphology of blast wave is ,however, not deformed significantly because the shock wave propagates through lower density inter-clump regions between dense clumps. This explains the nearly circular shape of the SNR with little observable deformation, even if the dense clumps are located inside the shell. Another implication of the numerical simulations is that the magnetic field may be amplified considerably to mG by turbulence in the interaction where the initial density is higher. Such amplified fields are able to offer an explanation of the enhanced synchrotron X ray emission, because the synchrotron emission is proportional to and the number density of high energy electrons. The rim-brightened X rays toward the molecular peaks are explained by such field amplification that leads also to efficient acceleration of high energy electrons in the inter-clump space.
We present millimeter and sub-millimeter spectroscopic observations of the molecular cloud cores in the TeV ray SNR RX J1713.7-3946. Our main conclusions are summarized as follows;
Three of the dense cores, peaks A, C and D, in RX J1713.7-3946 are associated with point sources with protostellar spectra as noted by Moriguchi et al. (2005). These cores show CO(=2-1) emission and have densities around 10 cm. Peak C is the most outstanding among the three and shows strong CO(=4-3) emission.
The spatial distributions of the four transitions, CO(=2-1, 3-2, 4-3) and CO(=2-1), indicate that the core of peak C has a strong density gradient consistent with an average density distribution of , where is the radius from the center of the core. The density and temperature, averaged over 90” (=0.2 pc) are 0.8 - 1.7 10 cm and 11 - 16 K, as derived from an LVG analysis. Peak C is also associated with a bipolar outflow as evidenced by the CO(=4-3) broad wings of at least 30 km s velocity extent. Along with the far infrared spectrum, these identify the region as a site of recent low-mass star formation within Myr. This verifies the broad wings are not produced by the shock acceleration driven by the SNR as suggested before by Fukui et al. (2003).
Peak A has density and temperature of 5 - 8 10 cm and 9 - 11 K, somewhat lower than peak C. The sources in peaks A and D are also likely protostars, because of the high density of these peaks.
A morphological comparison with X rays indicates that the dense cloud cores show a good anti-correlation with the X rays, suggesting that the X rays are enhanced on the surface or surroundings of the dense molecular gas where magnetic field strength is enhanced. We show that these features are consistent with theoretical results of shock propagation in the highly inhomogeneous ISM. Numerical simulations support that peak C is a dense clump which survived against shock erosion, since shock propagation speed is stalled in the dense clump.
|(deg)||(deg)||(K)||(km )||(km )||(pc)||
( K km )
Note. – Col. (1): Cloud name. Cols. (2)–(3): Position of the observed point with the maximum CO(=1-0) intensity. Cols. (4)–(6): Observed properties of the CO(=1-0) spectra obtained at the peak positions of the CO clouds. Col. (4): Peak radiation temperature . Col. (5): derived from a single Gaussian fitting. Col. (6): FWHM line width . Col. (7): Size defined as (/), where is the total cloud surface area defined as the region surrounded by the contour of 8.5 or 6.5 K (see text). If the contour is unclosed, the boundary is defined as the intensity minimum between the nearby peaks. Col. (8): The CO luminosity of the cloud . Col. (9): Mass of the cloud derived by using the relation between the molecular hydrogen column density () and the CO(=1-0) intensity (CO), () = 2.0 [(CO)/(K km )] () ( Bertsch et al. 1993). Col. (10): point source name nearby CO(=3-2) peaks.(see Cols. (1)–(9) are Moriguchi et al. 2005, Table 1)
|A||17082-3955||346.94||-0.31||17 11 41.04||-39 59 11.21||23||5||98||5.4||3.8||17.5||138||137|
|C||17089-3951||347.08||-0.39||17 12 26.46||-39 55 17.98||23||6||98||4.4||13.0||98.5||234||311|
|D||17079-3926||347.31||0.01||17 11 25.59||-39 29 53.29||39||6||98||2.0||20.0||88.6||739||562|
Note. – Cols. (1)–(2): Cloud name (Moriguchi et al. 2005) and point source near the CO(=3-2) peaks. Cols. (3)–(6): Position of the sources. Cols. (7)–(9): Semimajour axis, semiminor axis, and position angle of the position error of the sources. Cols (10)–(13): Fluxes of 12, 25, 60, and 100 , respectively. Col. (14): luminosity estimated using formula of Emerson (1988). Col. (18).(see Cols. (2)–(4) and (10)–(15) are Moriguchi et al. 2005, Table 3)
Note. – Col. (1): Cloud name. Col. (2)–(3): Position of the observed point with the maximum (=3-2) intensity peak. Col. (4)-(6): Radiation temperature averaged to the line of sight over a velocity integral of 1.5 km s. Col. (7): Density of molecular hydrogen. Col. (8): Kinetic temperature. The parameter, () (km s pc), is used.
|Beam size (arcsec)||90||22||38||90|
|Core radius (pc)||0.62||0.44||0.23||0.32|
|Average brightness (K)||0.79||1.87||1.28||0.29|
|Number density ()||0.330.05||1.510.23||3.550.53||1.780.27|
Note. – Number density is derived by assuming K and () (km s pc). We assume that the error of density caused by calibration from to is 15 for all intensities.
|Integrated intensity (K km )||17.5||14.7|
Note. – Integrated intensity is derived by summing the integrated intensities of the observed points in the area enclosed by a contour of 13.4 K within the velocity range of km s km s for blueshifted component and 11.0 K within km s km s for redshifted component, respectively. Size is defined as an effective diameter=, where is the region enclosed by a contour of 13.4 K km s and 11.0 K km s for the blueshifted and redshifted component, respectively. is the velocity range of the wing component. The dynamical age, , is defined as 2.
- slugcomment: version 2010.09.04 accepted
- Aharonian, F. A., et al. 2004, Nature, 432, 75
- Aharonian, F. A., et al. 2006, A&A, 449, 223
- Aharonian, F. A., et al. 2007, A&A, 464, 235
- Bertsch, D. L., Dame, T. M., Fichtel, C. E., Hunter, S. D., Sreekumar, P., Stacy, J. G., & Thaddeus, P. 1993, ApJ, 416, 587
- Blake, G. A., Sutton, E. C., Masson, C. R., & Phillips, T. G. 1987, ApJ, 315, 621
- Cassam-Chenai, G., Decourchelle, A., Ballet, J., Sauvageot, J, L., Dubner, G., & Giacani, E. 2004, A&A, 427, 199
- Denoyer, L. K. 1979, ApJL, 232, 165
- Desai, K. M., et al. 2010, AJ, 140, 584
- Inoue, T., Yamazaki, R., & Inutsuka, S. 2009, ApJ, 695, 825
- Fukui, Y. 1989, in Proc. ESO Workshop on Low Mass Star Formation and Pre-Main Sequence Objects, ed. B. Reipurth (ESO: Garching), 95
- Fukui, Y., Iwata, T., Mizuno, A., Bally, J., & Lane, A. P. 1993, in Protostars and Planets III, ed. E. H. Levy & J. I. Lunine (Tucson: Univ. Arizona Press), 603
- Fukui, Y., et al. 2003, PASJ, 55, 61
- Fukui, Y. 2008, in AIP Conf. Proc., Vol. 1085, Proc. of 4th International Meeting on High-Energy Gamma-Ray Astronomy, ed. F. A. Aharonian, W. Hofmann, & F. Rieger (Melville, NY: AIP), 104
- Gsten, R., & Philipp, S. D. 2004, in Proc. the Fourth Cologne-Bonn-Zermatt Symposium ed. S. Pfalzner, C. Kramer, C. Staubmeier, & A. Heithausen (Heidelberg: Springer), 253
- Goldreich, P., & Kwan, J. 1974, ApJ, 189, 441
- Koo, B.-C., et al. 2008, ApJL, 673, 147
- Koyama, K., Kinugasa, K., Matsuzaki, K., Nishiuchi, M., Sugizaki, M., Torii, K., Yamauchi, S., & Aschenbach, B. 1997, PASJ, 49, 7
- Kulesa, C. A., Hungerford, A. L., Walker, C. K., Zhang, X., & Lane, A. P. 2005, ApJ, 625, 194
- Lada, C. J. 1986, S&T, 72, 334
- Larson, R. B. 1969, MNRAS, 145, 271
- Lizano, S., & Shu, F. H. 1989, ApJ, 342, 834
- Mizuno, A., & Fukui, Y. 2004, in ASP Conf. Ser. 317, Milky Way Surveys: The Structure and Evolution of our Galaxy, ed. D. Clemens, R. Shah, & T. Brainerd (San Francisco, CA: ASP), 59
- Mizuno, Y., et al. 2010, PASJ, 62, 51
- Moriguchi, Y., Tamura, K., Tawara, Y., Sasago, H., Yamaoka, K., Onishi, T., & Fukui, Y. 2005, ApJ, 631, 947
- Onishi, T., et al. 1999, PASJ, 51, 871
- Penston, M. V. 1969, MNRAS, 145, 457
- Pfeffermann, E., & Aschenbach, B. 1996, in Proc. Rntgenstrahlung from the Universe, ed. H. U. Zimmermann, J. Trmper, & H. Yorke (MPE Rep. 263; Garching: MPE), 267
- Pineda, J. L., et al. 2008, A&A, 482, 197
- Scoville, N. Z., & Solomon, P. M. 1974, ApJL, 187, 67
- Slane, P., Gaensler, B. M., Dame, T. M., Hughes, J. P., Plucinsky, P. P., & Green, A. 1999, ApJ, 525, 357
- Tanaka, T., et al. 2008, ApJ, 685, 988
- Truelove, J. K., & McKee, C. F. 1999, ApJS, 120, 299
- Uchiyama, Y., Aharonian, F. A., Tanaka, T., Takahashi, T., & Maeda, Y. 2007, Nature, 449, 576
- Wang, Z. R., Qu, Q.-Y., & Chen, Y. 1997, A&A, 318, 59
- Wootten, H. A. 1977, ApJ, 216, 440
- Wootten, A. 1981, ApJ, 245, 105