X-Ray Expansion Velocity of the RCW 86 Northeast Rim

The Refined Shock Velocity of the X-Ray Filaments in the RCW 86 Northeast Rim

Hiroya Yamaguchi11affiliation: NASA Goddard Space Flight Center, Code 662, Greenbelt, MD 20771, USA 22affiliation: Department of Astronomy, University of Maryland, College Park, MD 20742, USA , Satoru Katsuda33affiliation: Institute of Space and Astronautical Science, JAXA, 3-1-1 Yoshinodai, Sagamihara, Kanagawa 229-8510, Japan , Daniel Castro11affiliation: NASA Goddard Space Flight Center, Code 662, Greenbelt, MD 20771, USA , Brian J. Williams11affiliation: NASA Goddard Space Flight Center, Code 662, Greenbelt, MD 20771, USA ,
Laura A. Lopez44affiliation: Department of Astronomy and Center for Cosmology & Astro-Particle Physics, The Ohio State University, Columbus, OH 43210, USA , Patrick O. Slane55affiliation: Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA , Randall K. Smith55affiliation: Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138, USA , Robert Petre11affiliation: NASA Goddard Space Flight Center, Code 662, Greenbelt, MD 20771, USA

A precise measurement of shock velocities is crucial for constraining the mechanism and efficiency of cosmic-ray (CR) acceleration at supernova remnant (SNR) shock fronts. The northeastern rim of the SNR RCW 86 is thought to be a particularly efficient CR acceleration site, owing to the recent result in which an extremely high shock velocity of  6000 km s was claimed (Helder et al. 2009). Here we revisit the same SNR rim with the Chandra X-ray Observatory, 11 years after the first observation. This longer baseline than previously available allows us to determine a more accurate proper motion of the nonthermal X-ray filament, revealing a much lower velocity of  km s (and even slower at a brighter region). Although the value has dropped to a half of that from the previous X-ray measurement, it is still higher than the mean velocity of the H filaments in this region ( 1200 km s). This discrepancy implies that the filaments bright in nonthermal X-rays and H emission trace different velocity components, and thus a CR pressure constrained by combining the X-ray kinematics and the H spectroscopy can easily be overestimated. We also measure the proper motion of the thermal X-ray filament immediately to the south of the nonthermal one. The inferred velocity ( km s) is significantly lower than that of the nonthermal filament, suggesting the presence of denser ambient material, possibly a wall formed by a wind from the progenitor, which has drastically slowed down the shock.

Subject headings:
acceleration of particles — ISM: individual objects (RCW 86) — ISM: supernova remnants — proper motions — shock waves — X-rays: ISM

1. Introduction

It is widely believed that supernova remnants (SNRs) are the predominant source of Galactic cosmic-rays (CRs) with energies up to  3 PeV (e.g., Reynolds 2008). RCW 86, probably the remnant of SN 185 (e.g., Clark & Stephenson 1975; Vink et al. 2006), is thought to be an efficient CR accelerator, based on the detection of nonthermal X-rays (Bamba et al. 2000; Borkowski et al. 2001; Rho et al. 2002) and -rays (Aharonian et al. 2009; Yuan et al. 2014). The large amount ( 1 ) of Fe ejecta in a low ionization state (Yamaguchi et al. 2011, 2014) and Balmer-dominated shocks (Smith 1997) observed in RCW 86 suggest a Type Ia supernova origin. The remnant is likely to have evolved in a wind-blown cavity formed by the progenitor binary system (Williams et al. 2011), maintaining a high enough shock velocity to accelerate CR particles over the entire lifetime of the remnant. In particular, its northeast (NE) rim is still expanding within the low-density cavity or has encountered the cavity wall very recently, so its current shock speed is the highest within the entire remnant (Vink et al. 1997; Yamaguchi et al. 2008; Broersen et al. 2014). For this reason, the NE rim of RCW 86 has been frequently studied to understand the details of the CR acceleration mechanism.

A velocity of the SNR shock wave in combination with the physical condition of the postshock plasma offers a key for constraining the CR acceleration efficiency. The most direct way to determine the shock velocity is to measure the proper motion of the shock front. Using two Chandra observations of the RCW 86 NE rim taken in 2004 and 2007, Helder et al. (2009) (hereafter H09) obtained the proper motion of the nonthermal X-ray filaments over a three-year baseline to be . This proper motion corresponds to a shock velocity of  km s at the well-agreed distance of 2.5 kpc (Rosado et al. 1996; Sollerman et al. 2003). H09 also analyzed an H spectrum from the same region and derived a postshock proton temperature of  keV based on the width of the emission lines. This was unexpected because the theoretical relationship between the postshock temperature and shock velocity, (where and are the mean particle mass and the proton mass), predicts a much higher postshock temperature ( = 42–70 keV depending on the degree of thermal equilibration) for = 6000 km s. H09 attributed this discrepancy to the efficiency of the CR acceleration: a substantial fraction of the kinetic energy is being transferred into nonthermal energy. They concluded that 50% of the postshock pressure is produced by the CR protons.

In their follow-up H observation, however, the mean proper motion of the forward shock filaments was inferred to be  yr, corresponding to a velocity of only  km s (Helder et al. 2013, hereafter H13). This velocity is far lower than their previous X-ray measurement, and consistent with the measured proton temperature (2 keV) with no energy injection to CR acceleration. H13 suggested that the H-bright filaments are biased toward denser regions with decelerated shock velocities, while the X-ray-emitting filaments represent the relatively undecelerated regions of the shock. This scenario seems reasonable, given the fact that the forward shock of the SNR is interacting with an inhomogeneous medium (e.g., Williams et al. 2011). In fact, H13 found substantial variation in the velocity of the H filaments (ranging from 300 to 3000  km s). We note, however, that there is also a substantial uncertainty in the X-ray velocity measurement of H09, due to the small separation in time (only 3 yrs) between the two Chandra observations. The positional shift they obtained () corresponds to only 3 1 pixels of the Chandra/ACIS detector. An additional X-ray observation using a much longer baseline is therefore crucial for determining the precise shock velocity.

Here we present the latest Chandra observation of the RCW 86 NE rim obtained in 2015, which is 11 years after the first observation. It reveals a much slower velocity of the nonthermal X-ray filaments than measured by H09. In §2, we describe the observations and data analysis. We discuss the results in §3, and present our conclusions in §4. Since the aim of the present work is to report the accurate velocity of the forward shock, we focus exclusively on the imaging analysis in this Letter. Detailed spectroscopy and further studies will be presented in a future paper. The uncertainties quoted in the text and the error bars in the figures represent the 1 confidence level, unless otherwise stated.

2. Observations and Results

We observed the NE rim of RCW 86 on 2015 June 26 using the Chandra/ACIS-S chips. As summarized in Table 1, the aim point and roll angle were set almost identical to those in the first observation of this region in 2004 (Vink et al. 2006), so that we can measure the proper motion as accurately as possible. We used CIAO version 4.8 and the latest calibration database (CALDB) for the data analysis. Examination of the light curves revealed no significant background flares. The net exposure we obtained was 67 ks.

Obs. ID 4611 16952
Observation Date 2004 June 15 2015 June 26
Instrument ACIS-S ACIS-S
Pointing R.A. (deg) 221.27 221.28
Pointing Dec. (deg) –62.35 –62.35
Roll Angle (deg) 295.16 293.58
Exposure Time (ks) 71.7 67.2
Table 1Summary of the Observations.
Figure 1.— Three-color X-ray image of the RCW 86 NE rim obtained from the Chandra/ACIS-S observation in 2015. Red, green, and blue correspond to the energy bands of 0.5–1.0 keV, 1.0–2.0 keV, and 2.0–5.0 keV, respectively. The three rectangles indicate where we measured the proper motions. The background point sources indicated by the green circles are used for the coordinate realignment.
Figure 2.— Projection profiles of the NE (left), NE (middle) and East (right) rims. Black and red indicate the data from the 2004 and 2015 observations, respectively. The unit of the vertical axis is  photons cm s pixel. The shock front in the 2004 data is used as the origin of the horizontal axis. The bottom panels are the profiles magnifying the regions around the shock front. The data in the green regions are used to determine the proper motion of each rim. The blue lines in the left and middle panels are the profiles expected if the shock velocity is 6000 km s as claimed by H09, which is plotted by simply shifting the 2004 profile 5.5 arcsec toward the positive direction.

Figure 1 shows the exposure-corrected ACIS image of the 0.5–1.0 keV (red), 1.0–2.0 keV (green), and 2.0–5.0 keV (blue) bands, indicating that the dominant X-ray emission component changes from thermal (red) to nonthermal (blue) along the filaments from the south to the north. The previous work has suggested that the southern part has already encountered the dense wall of the wind-blown cavity, whereas the northern part is still in the low-density cavity and thus maintains a high CR acceleration efficiency (e.g., Yamaguchi et al. 2008), which we confirm below. For accurate determination of the X-ray proper motion, we compared the images from the 2004 and 2015 observations, and aligned the coordinates using the relative positions of the four background point sources indicated in Figure 1. The details of the alignment procedure are described in Katsuda et al. (2009). The RMS of the residuals for the source positions between the two epochs is only , which we take as a systematic uncertainty in our proper motions measurements described below.

We extract projection profiles of the nonthermal filaments from the yellow and magenta rectangles in Figure 1 (hereafter “NE” and “NE”, respectively). The latter (fainter filament) is exactly where H09 measured the shock velocity. Figure 2 (left and middle) shows the profiles from both the 2004 (black) and 2015 (red) observations, using data in the 0.5–5.0 keV band. The shock front has clearly shifted over the 11 years in both filaments. Using the method established by Katsuda et al. (2008), we determine the proper motions of the NE and NE filaments to be  yr and  yr, respectively. The corresponding velocity is  km s for the NE and  km s for the NE (at the distance of 2.5 kpc), significantly less than the measurement by H09. For comparison, Figure 2 (left and middle) also shows the “expected” profiles (blue solid lines) for the case of = 6000 km s (H09). An uncertainty in the distance to the SNR,  kpc (Rosado et al. 1996; Sollerman et al. 2003), corresponds to that in the shock velocity of  km s for the NE filament (% of the mean value). In addition, there is a systematic uncertainty of  260 km s inferred from the astrometric accuracy ( between the two epoch). We also estimate another systematic uncertainty in our measurements by slightly changing the angle of the profile extraction regions, and obtain differences only within  yr or  km s. To summarize, the velocity of the NE filament never exceeds 4500 km s even with the multiple uncertainty components, and that of the NE filament must be even lower.

The large time separation between the two observations also allows us to examine if the thermal-dominated filaments have indeed decelerated due to the collision with the dense material. The right panels of Figure 2 show the profiles extracted from the white rectangle in Figure 1 (hereafter “East”). Unlike the NE filaments, we see only a little shift of the shock front. The proper motion of this region is measured to be  yr, corresponding to  km s. We establish that the East rim has a lower velocity. The value is comparable to the H-measured forward shock velocity at the southwest rim of this SNR where the shock front is also interacting with dense material ( 800 km s: Rosado et al. 1996).

3. Discussion

Our new observation of RCW 86 has revealed that the actual proper motions of the nonthermal filaments (NE and NE) are significantly slower than previously measured by H09. However, the updated shock velocities are still inconsistent with the mean velocity derived from the H proper motion in this region ( 1200 km s: H13). This implies that the NE rim of RCW 86 has a range of shock velocities due to a small-scale density inhomogeneity, and the H-bright filaments are indeed biased toward the denser regions as suggested by H13. In fact, we have revealed the significant difference between the velocities of the NE and NE filaments in our analysis. In the thermal-dominated (East) region, on the other hand, the X-ray and H velocities are consistent with each other, indicating the presence of dense material with a larger spatial scale, probably a wall formed by the progenitor wind (e.g., Williams et al. 2011), that had decelerated the overall shock propagation. Assuming pressure equilibrium and thus constant along the shell (where is an ambient density), we roughly estimate the density ratio between the East and NE regions to be .

Ours is not the first result casting doubt upon the extremely high velocity ( 6000 km s) claimed by H09. Williams et al. (2011) compared some observed characteristics of RCW 86 with their hydrodynamical simulations of a Type Ia SNR evolving in a wind-blown bubble. Although most of the characteristics were successfully reproduced with reasonable initial conditions, only the high shock velocity at the NE rim required more complex conditions, i.e., a significant offset between the explosion center and the SNR geometrical center. Recent Fermi observations of RCW 86 disfavor a hadronic origin for the -ray emission, where the inferred CR pressure is much lower than the estimate by H09 (Lemoine-Goumard et al. 2012; Yuan et al. 2014). These discrepancies have likely been resolved by our result. Notably, the refined velocity of the nonthermal X-ray filaments (1800–3000 km s) is consistent with an older estimate from the typical photon energy of the synchrotron X-rays, under the assumption that the acceleration and loss timescales for the relativistic electrons are the same (Vink et al. 2006) (see also Castro et al. (2013) for other regions of this SNR). This implies that the electrons responsible for the observed synchrotron X-rays have achieved their maximum energy at the NE region of RCW 86.

From the relationship of , a postshock proton temperature at the NE region is derived to be  11 keV, assuming thermal equilibrium (or higher if the equilibrium has not been achieved). Despite the substantial drop from the H09 value (i.e., = 42–70 keV), this temperature is still higher than that directly measured from the H spectrum ( 2.3 keV: H13). One might, therefore, consider that a significant amount of the shock energy is injected into CR acceleration. However, since the H emission is likely to be enhanced at denser regions where the shock velocity is lower, it could be inappropriate to combine the H and nonthermal X-ray measurements for studies of shock waves interacting with highly inhomogeneous environment, like the NE rim of RCW 86. We should also note that a substantial fraction of the upstream kinetic energy can be transferred into downstream turbulence, and thus the CR pressure estimated by the postshock temperature could be overestimated when naively compared with the shock velocity (Shimoda et al. 2015).

One open question is if thermal X-ray emission from the swept-up ambient medium coincides with the H-bright filaments in terms of both kinematics and plasma properties. Since H and thermal X-ray spectra constrain postshock proton and electron temperatures, respectively, they have often been combined to determine the efficiency of collisionless thermal equilibration between the protons and electrons (e.g., Rakowski et al. 2003; Helder et al. 2011). Although the proper motion measurements on the RCW 86 East (this work) as well as the northwestern rim of SN 1006 (Winkler et al. 2003; Katsuda et al. 2013) suggest a good coincidence between the thermal X-ray and H filaments, this should be verified with adiditional observations of H-bright forward shocks in other SNRs, like Kepler’s SNR (Sankrit et al. 2016) and the Cygnus Loop (Medina et al. 2014). Future observations with Hitomi (ASTRO-H) will allow us to determine the ion temperature of the swept-up medium. Comparison between it and the H-measured proton temperature will also be useful for understanding the relationship between the thermal X-ray and H observables.

4. Conclusions

We have presented new, improved measurements of the X-ray proper motion in the RCW 86 northeastern rim using Chandra data taken in 2004 and 2015 (with an 11-year separation). The velocity of the nonthermal filaments is derived to be 1800–3000 km s (at the distance of 2.5 kpc), which is much lower than the previous measurement by H09, but still significantly higher than the velocity inferred from the H proper motion (H13). This implies different origins of the filaments bright in nonthermal X-rays and H emission, with the former having a higher velocity. Caution should be exercised before combining nonthermal X-ray kinematics and H spectroscopy to study shock physics, at least in a complex environment such as this. We have also shown that the thermal filament immediately to the south of the nonthermal filament has a velocity of  km s, significantly slower than the nonthermal filament. The consistency between the velocities measured with the thermal X-ray and H emission suggests a common origin, but this should be verified with more observational work.

We thank Dr. Ryo Yamazaki for helpful discussion at Harvard-Smithsonian Center for Astrophysics. This work is supported by the Chandra GO Program grant GO5-16072A.


  • Aharonian et al. (2009) Aharonian, F., Akhperjanian, A. G., de Almeida, U. B., et al. 2009, ApJ, 692, 1500
  • Bamba et al. (2000) Bamba, A., Koyama, K., & Tomida, H. 2000, PASJ, 52, 1157
  • Borkowski et al. (2001) Borkowski, K. J., Rho, J., Reynolds, S. P., & Dyer, K. K. 2001, ApJ, 550, 334
  • Broersen et al. (2014) Broersen, S., Chiotellis, A., Vink, J., & Bamba, A. 2014, MNRAS, 441, 3040
  • Castro et al. (2013) Castro, D., Lopez, L. A., Slane, P. O., et al. 2013, ApJ, 779, 49
  • Clark & Stephenson (1975) Clark, D. H., & Stephenson, F. R. 1975, The Observatory, 95, 190
  • Helder et al. (2013) Helder, E. A., Vink, J., Bamba, A., et al. 2013, MNRAS, 435, 910
  • Helder et al. (2011) Helder, E. A., Vink, J., & Bassa, C. G. 2011, ApJ, 737, 85
  • Helder et al. (2009) Helder, E. A., Vink, J., Bassa, C. G., et al. 2009, Science, 325, 719
  • Katsuda et al. (2013) Katsuda, S., Long, K. S., Petre, R., et al. 2013, ApJ, 763, 85
  • Katsuda et al. (2009) Katsuda, S., Petre, R., Long, K. S., et al. 2009, ApJ, 692, L105
  • Katsuda et al. (2008) Katsuda, S., Tsunemi, H., & Mori, K. 2008, ApJ, 678, L35
  • Lemoine-Goumard et al. (2012) Lemoine-Goumard, M., Renaud, M., Vink, J., et al. 2012, A&A, 545, A28
  • Medina et al. (2014) Medina, A. A., Raymond, J. C., Edgar, R. J., et al. 2014, ApJ, 791, 30
  • Rakowski et al. (2003) Rakowski, C. E., Ghavamian, P., & Hughes, J. P. 2003, ApJ, 590, 846
  • Reynolds (2008) Reynolds, S. P. 2008, ARA&A, 46, 89
  • Rho et al. (2002) Rho, J., Dyer, K. K., Borkowski, K. J., & Reynolds, S. P. 2002, ApJ, 581, 1116
  • Rosado et al. (1996) Rosado, M., Ambrocio-Cruz, P., Le Coarer, E., & Marcelin, M. 1996, A&A, 315, 243
  • Sankrit et al. (2016) Sankrit, R., Raymond, J. C., Blair, W. P., et al. 2016, ApJ, 817, 36
  • Shimoda et al. (2015) Shimoda, J., Inoue, T., Ohira, Y., et al. 2015, ApJ, 803, 98
  • Smith (1997) Smith, R. C. 1997, AJ, 114, 2664
  • Sollerman et al. (2003) Sollerman, J., Ghavamian, P., Lundqvist, P., & Smith, R. C. 2003, A&A, 407, 249
  • Vink et al. (2006) Vink, J., Bleeker, J., van der Heyden, K., et al. 2006, ApJ, 648, L33
  • Vink et al. (1997) Vink, J., Kaastra, J. S., & Bleeker, J. A. M. 1997, A&A, 328, 628
  • Williams et al. (2011) Williams, B. J., Blair, W. P., Blondin, J. M., et al. 2011, ApJ, 741, 96
  • Winkler et al. (2003) Winkler, P. F., Gupta, G., & Long, K. S. 2003, ApJ, 585, 324
  • Yamaguchi et al. (2008) Yamaguchi, H., Koyama, K., Nakajima, H., et al. 2008, PASJ, 60, S123
  • Yamaguchi et al. (2011) Yamaguchi, H., Koyama, K., & Uchida, H. 2011, PASJ, 63, S837
  • Yamaguchi et al. (2014) Yamaguchi, H., Badenes, C., Petre, R., et al. 2014, ApJ, 785, L27
  • Yuan et al. (2014) Yuan, Q., Huang, X., Liu, S., & Zhang, B. 2014, ApJ, 785, L22
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
Add comment
Loading ...
This is a comment super asjknd jkasnjk adsnkj
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test description