next previous
Up: The neutral gas environment


Subsections

3 Results


  \begin{figure}
\par\includegraphics[width=9cm,clip]{MS2223f2a.ps} \end{figure} Figure 2: From a) to d): images of the H  I data (interferometric only) between LSR velocities $\sim $-52 and $\sim $+34 km s-1, each one integrated over 5 consecutive channels (spanning 4 km s-1). The velocity shown in the upper right corner of each image corresponds to the first channel of integration. The plotted H  I contours are 1.5, 2, 3, 4, 5, 6, 7, 8, 9, 12 and 15 in units of 1.4 K. The data are Gaussian smoothed to an angular resolution of 5$^\prime $ $\times $ 5$^\prime $. The rms noise is 0.3 K. The MOST continuum image of SN 1006 at 843 MHz (from Bock et al. 1999) is included in greyscale for comparison.


 \begin{figure}
\par\includegraphics[width=9cm,clip]{MS2223f2b.ps}
\end{figure} Figure 2: continued.


 \begin{figure}
\par\includegraphics[width=9cm,clip]{MS2223f2c.ps}
\end{figure} Figure 2: continued.


 \begin{figure}
\par\includegraphics[width=9cm,clip]{MS2223f2d.ps}
\end{figure} Figure 2: continued.

Figure 2 (from (a) to (d)) displays in contour lines the H  I distribution within the velocity interval (-52, +30) km s-1, the range over which significative H  I emission is detected. To produce this image we have used the interferometric data (smoothed with a 5$^\prime $ FWHM Gaussian) in order to easily recognize the occurrence of small scale structure. All calculations are, however, carried out using the combined (interferometric plus single dish) data. Each H  I line image was obtained by averaging 5 consecutive channel images, thus spanning 4 km s-1. The velocity corresponding to the first channel is shown in the top right corner of each image. These contour plots, with isophotes ranging from 2.1 K to 21 K, provide both a qualitative and quantitative image of the H  I distribution in an extended field around SN 1006. The MOST image of the radio continuum emission of SN 1006 is included in greyscale in all the images for comparison.

At high negative velocities ( $ -52 \leq V_{\rm LSR} \leq -40$ km s-1) the most conspicuous feature is the extended H  I concentration located near the S-SE border of SN 1006. From $V_{\rm LSR} \sim {{-}31}$ km s-1 to $V_{\rm LSR} \sim {{-}23}$ km s-1, the presence of an H  I concentration projected onto the center of the SNR is quite striking. Between $V_{\rm LSR} \sim {{-}19}$ km s-1 and $V_{\rm LSR} \sim {{-}11}$ km s-1, SN 1006 appears surrounded by smooth, tenuous gas, with a few H  I concentrations scattered over the field. At $V_{\rm LSR} \sim {{-}6.6}$ km s-1 an extended H  I structure encompass SN 1006 along its E, N and NW sides. Within this structure, a bright H  I concentration is observed towards the NW, in the direction where the brightest optical filaments are detected. This feature has two maxima centered near (15$^{\rm h}$01$^{\rm m}$00$^{\rm s}$, -41$^\circ$45$^\prime $) and (15$^{\rm h}$02$^{\rm m}$30$^{\rm s}$, -41$^\circ$35$^\prime $). Part of this feature can still be detected at $V_{\rm LSR} \sim {{-}2.5}$ km s-1. At this last velocity, an H  I feature delineates the E-NE border of SN 1006.

At $V_{\rm LSR} \sim {{+}1.7}$ km s-1, a bright H  I concentration is detected to the SW of the SNR, with a maximum centered near 15$^{\rm h}$01$^{\rm m}$20$^{\rm s}$, -42$^\circ$10$^\prime $. The outer contour of this feature mimics the shape of the SW limb of the radio remnant. At $V_{\rm LSR} \sim {{+}5.8}$ km s-1, this feature acquires an elongated NS morphology. At +5.8 km s-1 and at +9.9 km s-1, the maximum of this feature appears adjoining the SNR outer border in coincidence with a small indentation of the radio continuum. At these velocities the presence of an H  I concentration adjacent to the flat NW border of SN 1006, peaking near $\sim $15$^{\rm h}$02$^{\rm m}$, -41$^\circ$40$^\prime $, is also apparent. Because this is the direction where the bright Balmer filaments are detected, we will analyze this feature despite the anomalous positive velocity.

At $V_{\rm LSR}= {{+}9.9}$ km s-1 and +14 km s-1, a small H  I emission feature is observed projected onto the center of SN 1006. From $V_{\rm LSR} \sim {{+}18}$ km s-1 to $V_{\rm LSR} \sim {{+}30}$ km s-1, the only noticeable feature is the concentration detected to the SE, which appears to overlap a small portion of the SE corner of SN 1006.

3.1 Analysis of the H I features

In order to understand the peculiar characteristics of SN 1006 (e.g. the bilateral appearance, the position of the brightest optical filaments, and the reason why the TeV $\gamma-$ray emission originates only on the NE lobe) it is important to estimate the preshock conditions towards different directions around SN 1006, looking for inhomogeneities and/or anisotropies in the H  I distribution.

A fundamental issue in establishing a physical association between galactic H  I and SN 1006 is the determination of the systemic velocity for this SNR. Several estimates for the distance to SN 1006 have been carried out using different methods: d=1.4-2.1 kpc (Kirshner et al. 1987); $d=0.7 \pm 0.1$ kpc (Willingale et al. 1996); $d=1.6 \pm 0.1$ kpc (Schaefer 1996); d=2 kpc (Winkler & Long 1997); $d \leq
2.1$ kpc (Burleigh et al. 2000) and d=1.4 kpc (Allen et al. 2001). Collectively, these determinations suggest that the distance to SN 1006 lies in the range 1.4-2 kpc. The Galactic circular rotation model (Fich et al. 1989) suggests a $V_{\rm LSR}$ between $\sim $-16 and $\sim $-25 km s-1 for this distances range, with possible uncertainties of the order of 7 km s-1 because of peculiar or non-circular motions (Burton 1992). However, one must be very cautious when applying circular rotation models for sources located far from the Galactic plane, since these models are only strictly valid for low Galactic latitudes. Moreover, the warping of the Galactic plane towards negative latitudes in the fourth Galactic quadrant (Burton 1976) increases the height above the plane of an object located at a positive latitude. Here, we analyze the H  I information based on morphological comparisons among the H  I radiocontinuum, X-rays and optical emission from SN 1006, irrespective of the H  I velocity.

As mentioned before, between $V_{\rm LSR} \sim {{-}31}$ km s-1 and $V_{\rm LSR} \sim {{-}23}$ km s-1, an H  I concentration appears projected onto the center of SN 1006. X-ray observations indicate that the interior of SN 1006 is almost free of neutral hydrogen, thus this H  I feature is either in front or behind SN 1006. Based on H  I observations, there is no way to accurately locate this feature along the line of sight, because the hollow center of this SNR precludes the use of absorption tests. If this H  I cloud lies behind the SNR, an upper limit of $\sim $-20 km s-1 can be derived for the systemic velocity of SN 1006, thus suggesting an upper limit of 1.7 kpc for the distance to the remnant. Within the large uncertainties involved in the models, this limit for the distance is compatible with the average of the various distance estimates for SN 1006, as summarized above. If, on the other hand, the central H  I feature would be localized in front of SN 1006, then the systemic velocity of the SNR would be any value more negative than $\sim $-31 km s-1, more difficult to reconcile with the distance estimated for SN 1006 based on independent methods. We thus conclude that adopting $V_{\rm sys}\sim {{-}20}$ km s-1 and $d \sim 1.7$ kpc for the possible systemic velocity and distance to SN 1006, respectively, is a plausible hypothesis.

To investigate the average H  I density in the environs of SN 1006, we have calculated the atomic density in different directions around the remnant and in each velocity plane within the interval $-20 \leq V_{\rm LSR} \leq {{-}16}$ km s-1, assuming that the systemic velocity of SN 1006 lies somewhere within this range, as suggested by the present and previous results. An average density of n0 = 0.5 d-1 cm-3 is obtained, where d is the distance in kpc. If $d \sim 1.7$ kpc, then $n_0 \simeq 0.3$ cm-3. To carry out these calculations we have assumed that the gas is optically thin, a good approximation for gas at $b \sim {{+}15}^{\circ}$, $z \sim 450$ pc. The determined value is an upper limit since contribution from unrelated distant gas with the same kinematical velocities may be included due to the distance ambiguity in this direction of the Galaxy.

The features to the north-west:

As discussed above, the flat NW border of SN 1006 is faint in radio continuum and in X-rays, but is delimited by sharp, bright optical filaments with strong Balmer lines (Long et al. 1988; Smith et al. 1991; Winkler & Long 1997). For these H$_\alpha$ filaments proper motions of the order of 0.30§ yr-1 (Long et al. 1988) and shock velocities within the range 2200-3500 km s-1 (Smith et al. 1991), have been estimated. Winkler & Long (1997) also detected fainter emission 2$^\prime $ beyond the bright filaments to the NW.


  \begin{figure}
\par\includegraphics[width=9cm,clip]{MS2223f3.ps}\end{figure} Figure 3: a) Comparison of the H  I distribution (contours) at $V_{\rm LSR}= {{-}4.9}$ km s-1 as obtained from the interferometric data, with the radio continuum emission of SN 1006 (from Bock et al. 1999) (left) and with the optical emission towards the NW corner of SN 1006 (from Winkler & Long 1997) (right); b) the same for the H  I distribution at $V_{\rm LSR}= {{+}9.9}$ km s-1.

Balmer-dominated filaments are characterized by a spectrum of Balmer lines with little or none of the characteristic forbidden-line emission of radiative filaments (such as [SII], [NII] or [OIII]). Chevalier & Raymond (1978) explained the nature of this spectrum as due to a collisionless shock moving into partially neutral interstellar material. Winkler & Long (1997) conclude that the bright optical filaments observed in SN 1006 are, in fact, thin sheets observed nearly edge-on, which originate where the shock encounters a "wall'' with density of about 1 cm-3 (half of which would be ionized), oriented nearly parallel to the line of sight. The fainter filaments are formed where the SN shell encounters a lower density region. The fact that the northwest portion of SN 1006 appears flattened and expands less rapidly than elsewhere (Moffett et al. 1993; Winkler & Long 1997) is consistent with the existence of denser preshock material in this direction.

We can model a neutral "wall'' with the geometry and density suggested by Winkler & Long (1997): a narrow sheet of gas with a width of 2$^\prime $, comparable to the separation between bright and faint filaments (about 1 pc at the adopted distance of 1.7 kpc), a length of 15$^\prime $, which is the angular size of the bright optical filaments ($\sim $7.5 pc at $d \sim 1.7$ kpc), and a depth of 30$^\prime $ along the line of sight, similar to the diameter of SN 1006 ($\sim $15 pc at $d \sim 1.7$ kpc), with a neutral gas density of $\sim $0.5 cm-3. This neutral hydrogen feature has a column density of $\sim $ $ 2\times 10^{19}$ cm-2, and the correspondent brightness temperature would be $\sim $15 K. Therefore, the present H  I images should reveal the presence of the predicted feature, even taking into account the possibility that the signal may suffer from beam dilution and/or is spread over several channels.

From an inspection of the entire observed H  I cube (from -250 to +400 km s-1), we conclude that H  I emission enhancements towards the NW occur only from $V_{\rm LSR} \sim {{-}4.5}$ km s-1 to $V_{\rm LSR} \sim {{-}6.6}$ km s-1 and from $V_{\rm LSR} \sim {{+}9}$ km s-1 to $V_{\rm LSR}\sim {{+}11}$ km s-1 (Figs. 2b and 2c). One of these two concentrations must be responsible for the particular formation of bright optical filaments exclusively on this side of the SNR. In Fig. 3 we display the H  I distribution overlaid on the radio continuum and the optical emission (towards the NW) at these velocity ranges (Figs. 3a and 3b respectively). From the morphological point of view, the best agreement is found near +9.9 km s-1, where there is an excellent correspondence between the H  I  and the optical emission. The major problem with this association is the discrepancy between this velocity of $\sim $+10 km s-1 and the probable systemic velocity of SN 1006 of $\sim $-20 km s-1. Colomb & Dubner (1982) have shown the existence of two concentric expanding H  I shells of swept up interstellar matter around the SNR Lupus Loop (G330.0+15.0). Therefore, in spite of the good morphological agreement, it is possible that the observed features near +9.9 km s-1 are part of the slowly expanding external shell around Lupus Loop and are located at $\sim $300-500 pc.

On the other hand, for the NW feature observed near $\sim $-5 km s-1, the existence of a second coincidence between the shape of the H  I features and the radio continuum emission reinforces the hypothesis of association between the H  I and the SNR. In effect, near 15$^{\rm h}$03$^{\rm m}$30$^{\rm s}$, -41$^\circ$45$^\prime $, an H  I concentration is seen apparently adjacent to SN 1006 at the position where the radio shell is broken and appears to branch off to the interior of the SNR. Of course, the LSR velocity of this H  I concentration is still anomalous if the systemic velocity of SN 1006 is close to -20 km s-1, and some kinematical perturbation has to be assumed. If there is a physical link between the NW cloud and the SNR, then an atomic density n = 0.8 d-1 cm-3 is derived (n = 0.5 cm-3 at a distance of $\sim $1.7 kpc). In order to calculate this density we have assumed for the elongated feature a dimension along the line of sight equal to the diameter of the SNR, and integrated the brightness temperature between -7 and -4 km s-1. This value is in good agreement with the density proposed by Winkler & Long (1997) for the neutral component of the "wall'' based on shock model fitting to the optical and X-ray data. The origin of the anomalous kinematical velocity remains an unsolved problem.

Unfortunately, there is little kinematic evidence to support the physical association of the H  I with the SNR; projection effects and chance alignments cannot be ruled out. Thus, a quantitative analysis of the kinematical effects of this young SNR on the surrounding ISM, is ruled out.

3.2 The neutral hydrogen column density

In Fig. 4 we show the spatial distribution of the $N_{\rm H}$,

  \begin{figure}
\par\includegraphics[width=6.4cm,clip]{MS2223f4.ps}\end{figure} Figure 4:I column density integrated between 0 and -20 km s-1. The greyscale varies between $6.6 \times 10^{20}$ and $7.5 \times 10^{20}$ cm-2. The contours (in units of 1020 cm-2) are 6.6, 7.0, 7.4 and 7.8, i.e. every three times the uncertainty in the column density estimate. The outer contour of the radio-continuum image of SN 1006 is included for reference.

obtained from an integration of the brightness temperature (interferometric plus single dish data) between $V_{\rm LSR}\sim
{{-}20}$ km s-1 and $V_{\rm LSR}\sim 0$ km s-1, i.e. all the H  I included between the Sun and a likely systemic velocity of SN 1006. Due to the distance ambiguity in this direction of the Galaxy, gas located between $\sim $12 kpc and $\sim $14 kpc, would, in theory, contribute to the column density distribution. It is highly unlikely, however, that there is much gas located so far. Thus, the major uncertainty in considering this distribution as representative of the intervening column to the SNR, is the selection of the integration limit of $V_{\rm LSR}\sim
{{-}20}$ km s-1. From Fig. 4 we notice that the absorbing column density decreases from the interior of SN 1006 (N  H $\sim $ $7 \times 10^{20}$ cm-2) to the radio lobes (N  H $\sim $ $6.6 \times 10^{20}$ cm-2). The maximum N  H is observed towards the NW ( $8 \times 10^{20}$ cm-2), while in the opposite direction to the SE (closer to the Galactic plane) the column density is a minimum ( $6 \times 10^{20}$ cm-2). This latter position coincides with the portion of SN 1006 with the lowest emission in all wavelengths.


next previous
Up: The neutral gas environment

Copyright ESO 2002