Abstract

Four mosaics of deep, continuum-subtracted, CCD images have been obtained over the extensive Galactic radio continuum shell, W50, which surrounds the remarkable stellar system SS 433. Two of these mosaics in the forumla and [O iii] 5007 Å emission lines, respectively, cover a field of forumla which contains all of W50 but at a low angular resolution of 5 arcsec. The third and fourth mosaics cover the eastern (in [O iii] 5007 Å) and western (in forumla 6548, 6584 Å) filamentary nebulosity, respectively, but at an angular resolution of 1 arcsec. These observations are supplemented by new low-dispersion spectra and long-slit, spatially resolved echelle spectra. The [O iii] 5007 Å images show for the first time the distribution of this emission in both the eastern and western filaments while new forumla emission features are also found in both of these regions. Approaching flows of faintly emitting material from the bright eastern filaments of up 100 km s−1 in radial velocity are detected. The present observations also suggest that the heliocentric systemic radial velocity of the whole system is 56 ± 2 km s−1. Furthermore, very deep imagery and high-resolution spectroscopy of a small part of the northern radio ridge of W50 has revealed for the first time the very faint optical nebulosity associated with this edge. It is suggested that patchy foreground dust along the ≈5 kpc sightline is inhibiting the detection of all of the optical nebulosity associated with W50. The interaction of the microquasar jets of SS 433 with the W50 shell is discussed.

1 INTRODUCTION

Most of the Galactic supernova remnants (SNRs) have been identified by their radio, optical and X-ray emission (Green 2006). G39.7−2.0 (W50) was identified in radio wavelengths by Westerhout (1958) and first classified as an evolved SNR by Holden & Caswell (1969). Its radio continuum image has a complex ‘seashell’ appearance showing a main spherical shell of non-thermal emission (∼58 arcmin in diameter) with easterly and westerly extensions or lobes. The radio spectral index varies across the eastern, central and western areas between values of 0.4 and 0.7 and the total flux densities are 71 and 160 Jy at 1465 and 327 GHz, respectively (Dubner et al. 1998). The latter authors also presented the latest H i observations showing some evidence of an interaction between W50 and its surrounding interstellar medium (ISM) at a velocity ∼76 km s−1 while a distance to W50 of ∼3 kpc was also calculated. This simple SNR interpretation of W50 is complicated by the presence at its centre of the remarkable stellar binary system SS 433 (see Fabrika 2004, for a comprehensive review) from which emission is ejected in two oppositely directed relativistic jets aligned along the same axis as the elongation of the radio lobes (see also fig. 2; Dubner 2002). The two jets are also X-ray emitters (Brinkmann, Aschenbach & Kawai 1996; Safi-Harb & Ögelman 1997) showing different morphological and spectral properties. These X-ray and IRAS (Band 1987) observations suggest that the interaction of SS 433 jets with the surrounding ISM plays an important role in the shaping of the envelope. Stirling et al. (2002) measured the distance to the jet of SS 433, hence W50, from radio proper-motion measurements, as 4.61 ± 0.35 kpc and Blundell & Bowler (2004) give 5.5 ± 0.2 kpc. Here, we will adopt a distance of ≈5 kpc.

Optical emission was first discovered by both van den Bergh (1980) and Zealey, Dopita & Malin (1980) who found two groups of faint filaments, one east and one west of SS 433 at a distance of ∼30 arcmin from it. Spectroscopic observations (Kirshner & Chevalier 1980; Murdin & Clark 1980; Shuder, Hatfield & Cohen 1980) show that all the optical filaments originated from shock-heated gas since the [S ii]/Hα∼ 2 while very strong [N ii] ([N ii] 6584 Å/Hα∼ 3) emission is also present. Faint [O iii] 5007 Å emission was also detected in these spectra ([O iii] 5007 Å/Hα∼ 0.4) but only from the eastern filaments. Mazeh et al. (1983) presented high spectral but low angular (with 120 and 200 arcsec beam diameters) resolution profiles of the [N ii] 6584 Å line from both the eastern and western optical filaments of W50. These profiles have a complex structure and are broad (∼50 km s−1); see Section 3.3 for further details and comparison with new line profiles.

A large number of papers have been published (see the review of Fabrika 2004, and references therein) concerning the association of W50 with SS 433. A consensus is emerging that the SS 433 jets have punched holes in a preceding, expanding supernova shell (e.g. Velázquez & Raga 2000) though the formation of the whole structure of W50 by the jets alone has not been completely ruled out (Fabrika 2004).

In the present paper, the first continuum-subtracted, deep CCD images of the total area of W50 in the light of forumla and [O iii] 5007 Å are presented in a mosaic which covers an area of forumla. The major eastern and western regions of nebulosity are also imaged at higher angular resolution in the [O iii] 5007 Å and forumla 6548, 6584 Å lines, respectively, to reveal the fine structure of the optical filaments. Furthermore, a very deep forumla 6548, 6584 Å image of a small region of the northern ridge of radio emission has been obtained in an attempt to detect hitherto undiscovered filamentary nebulosity that would be expected if the W50 radio shell is an SNR.

The new, faint, optical emission found over W50, together with its comparison with a high-resolution radio map, permits a better appreciation of the morphology of W50 and its interaction with the jets of SS 433. In addition, new, deep, low-resolution and spatially resolved, high-resolution spectra of this optical nebulosity have been obtained to advance the understanding of this interaction.

2 OBSERVATIONS AND RESULTS

A summary and log of our imaging and spectral observations is provided in Table 1. In the sections below, we describe the details of these observations.

Table 1

Imaging and spectral log.

Imaging   
Filter  λc (Å)  Δλ (Å)  Exposure time (s)  Area  Observatory   
forumla 6548, 6584 Å  6570  75  2400  W50 total (4)a  0.3-m Skinakas   
[O iii] 5007 Å  5005  28  2400  W50 total (4)  0.3-m Skinakas   
Cont blue  5470  230  180  W50 total (4)  0.3-m Skinakas   
Cont red  6096  134  180  W50 total (4)  0.3-m Skinakas   
[O iii] 5007 Å  5005  28  2400  East (10)  1.3-m Skinakas   
Cont blue  5470  230  180  East (10)  1.3-m Skinakas   
forumla 6548, 6584 Å  6570  75  2400  West (2)  1.3-m Skinakas   
Cont red  6096  134  180  West (2)  1.3-m Skinakas   
forumla 6548, 6584 Å  6580  90  2400  North (1)  2.1-m SPM   
Spectroscopy 
Area  Slit centres  Exposure time (s)  Offsetb (arcsec)  Aperture lengthc (arcsec)  Observatory 
α  δ 
East I (EI)  19h14m20s  forumla   3900  112 S  12  1.3-m Skinakas 
East II (EII)  19h14m20s  forumla   3900  94 S  26  1.3-m Skinakas 
East III (EIII)  19h14m20s  forumla   3900  61 S  66  1.3-m Skinakas 
East IV (EIV)  19h14m24s  forumla   3900  58 S  30  1.3-m Skinakas 
East V (EV)  19h14m38s  forumla   3900  4 N  24  1.3-m Skinakas 
West I (WI)  19h09m39s  forumla   3900  89 S  36  1.3-m Skinakas 
West II (WII)  19h09m39s  forumla   3900  50 S  43  1.3-m Skinakas 
West III (WIII)  19h09m39s  forumla   3900  193 N  44  1.3-m Skinakas 
East slit 1  19h14m18s  forumla   1800  0d  300e  2.1-m SPM 
East slit 2  19h13m57s  forumla   3600  0d  300e  2.1-m SPM 
North slit  19h11m01s  forumla   4200  90d  150e  2.1-m SPM 
Imaging   
Filter  λc (Å)  Δλ (Å)  Exposure time (s)  Area  Observatory   
forumla 6548, 6584 Å  6570  75  2400  W50 total (4)a  0.3-m Skinakas   
[O iii] 5007 Å  5005  28  2400  W50 total (4)  0.3-m Skinakas   
Cont blue  5470  230  180  W50 total (4)  0.3-m Skinakas   
Cont red  6096  134  180  W50 total (4)  0.3-m Skinakas   
[O iii] 5007 Å  5005  28  2400  East (10)  1.3-m Skinakas   
Cont blue  5470  230  180  East (10)  1.3-m Skinakas   
forumla 6548, 6584 Å  6570  75  2400  West (2)  1.3-m Skinakas   
Cont red  6096  134  180  West (2)  1.3-m Skinakas   
forumla 6548, 6584 Å  6580  90  2400  North (1)  2.1-m SPM   
Spectroscopy 
Area  Slit centres  Exposure time (s)  Offsetb (arcsec)  Aperture lengthc (arcsec)  Observatory 
α  δ 
East I (EI)  19h14m20s  forumla   3900  112 S  12  1.3-m Skinakas 
East II (EII)  19h14m20s  forumla   3900  94 S  26  1.3-m Skinakas 
East III (EIII)  19h14m20s  forumla   3900  61 S  66  1.3-m Skinakas 
East IV (EIV)  19h14m24s  forumla   3900  58 S  30  1.3-m Skinakas 
East V (EV)  19h14m38s  forumla   3900  4 N  24  1.3-m Skinakas 
West I (WI)  19h09m39s  forumla   3900  89 S  36  1.3-m Skinakas 
West II (WII)  19h09m39s  forumla   3900  50 S  43  1.3-m Skinakas 
West III (WIII)  19h09m39s  forumla   3900  193 N  44  1.3-m Skinakas 
East slit 1  19h14m18s  forumla   1800  0d  300e  2.1-m SPM 
East slit 2  19h13m57s  forumla   3600  0d  300e  2.1-m SPM 
North slit  19h11m01s  forumla   4200  90d  150e  2.1-m SPM 

a Numbers in parentheses represent the total number of different fields.

b Spatial offset from the slit centre in arcsec: N(=North), S(=South).

c Aperture lengths for each area in arcsec.

d Slit position angle (PA) in degrees.

e Slit width in μm.

2.1 Imaging

2.1.1 Wide-field imagery

The wide-field imagery was taken with the 0.3-m Schmidt-Cassegrain (f/3.2) telescope at Skinakas Observatory in Crete, Greece, from 2003 June 27 to 30. A 1024×1024 SITe (Scientific Imaging Technologies, Inc.) CCD was used which has a pixel size of 24 μm resulting in an 89 × 89 arcmin2 field of view and an image scale of 5 arcsec pixel−1.

Each of four different fields was observed for 2400 s in both filters while corresponding continuum images were also observed (180 s each) and were subtracted from those containing the emission lines to eliminate the confusing star field (see Boumis et al. 2002, for details of this technique). All fields were projected on to a common origin on the sky and were subsequently combined to create the final mosaics in forumla and [O iii] 5007 Å. During the observations the ‘seeing’ varied between 0.8 and 1.2 arcsec. The image reduction was carried out using the iraf and starlink packages. The astrometric solutions were calculated for each field using reference stars from the Hubble Space Telescope (HST) Guide Star Catalogue (Lasker, Russel & Jenkner 1999). All coordinates quoted in this paper refer to epoch 2000.

The images of the nebulosity in all of the fields considered here are detected, at most, only a few times greater than the residual noise level. In these circumstances we have always chosen to display the data with a linear scale, but negatively, and at high contrast. Furthermore, we have chosen not to suppress the noise artificially, for example, by excessive smoothing or lifting the zero level, for this often leads to detection artefacts being confused in the resultant display with real nebulous features.

The image of W50 with this system is shown in Fig. 1. This is a mosaic of four images taken through the forumla filter. The same field was observed with the [O iii] 5007 Å emission line filter. Dubner et al. (1998) and Dubner (2002) using their high-resolution radio map of W50 made a comparison between the radio/X-ray/H i emission. These radio continuum contours (Dubner et al. 1998) are compared in Fig. 1 with this new mosaic of forumla images.

Figure 1

The correlation between the forumla negative continuum-subtracted mosaic of W50 in the light of forumla and the radio emission at 1465 MHz (solid lines). The 1465 MHz (Dubner et al. 1998) radio contours scale linearly from 1 × 10−2 to 0.1 Jy beam−1. The strong radio source to the north-west is LBN 109 (see the text). The image has been smoothed to suppress the residual from the imperfect continuum subtraction. The horizontal line segments seen near overexposed stars in this figure and the next figures are due to the blooming effect. The optical features are shown in detail in Figs 2(b) and 3(b).

Figure 1

The correlation between the forumla negative continuum-subtracted mosaic of W50 in the light of forumla and the radio emission at 1465 MHz (solid lines). The 1465 MHz (Dubner et al. 1998) radio contours scale linearly from 1 × 10−2 to 0.1 Jy beam−1. The strong radio source to the north-west is LBN 109 (see the text). The image has been smoothed to suppress the residual from the imperfect continuum subtraction. The horizontal line segments seen near overexposed stars in this figure and the next figures are due to the blooming effect. The optical features are shown in detail in Figs 2(b) and 3(b).

Selected areas of the eastern and western filaments seen in Fig. 1 are shown, respectively, in Figs 2(a) and (b) and 3(a) and (b) in the light of [O iii] 5007 Å and forumla. For the first time, deep, continuum-subtracted, CCD images in the light of forumla and [O iii] 5007 Å of W50 have been obtained. The forumla mosaic of images in Fig. 1 shows new filamentary and diffuse emission while, for the first time, [O iii] 5007 Å filamentary emission from W50 is revealed. The most striking features are the differences in the filamentary nebulosities in Figs 2(a) and (b). The [O iii] 5007 Å emission in Fig. 2(a) forms a 24 arcmin-long outer arc (α≃ 19h14m30s, δ≃ 4°30′) which contains the predominantly [N ii] 6584 Å (see Section 2.2.1) emitting filaments. This eastern filamentary arc is convex with respect to SS 433 whereas the [N ii] 6584 Å emitting western filamentary arc in Figs 3(a) and (b) is concave and only has very localized [O iii] 5007 Å emitting counterparts.

Figure 2

The eastern complex filamentary structure in the light of (a) [O iii] 5007 Å and (b) forumla. The image has been smoothed to suppress the residual from the imperfect continuum subtraction.

Figure 2

The eastern complex filamentary structure in the light of (a) [O iii] 5007 Å and (b) forumla. The image has been smoothed to suppress the residual from the imperfect continuum subtraction.

Figure 3

The western filaments in the light of (a) [O iii] 5007 Å and (b) forumla. The image has been smoothed to suppress the residual from the imperfect continuum subtraction.

Figure 3

The western filaments in the light of (a) [O iii] 5007 Å and (b) forumla. The image has been smoothed to suppress the residual from the imperfect continuum subtraction.

Diffuse forumla is present to the north-east forumla and north-west forumla of the main, circular radio remnant of W50 in Fig. 1. It is also present within its western radio lobe and the possibility that it belongs to the remnant cannot be ruled out. The weak diffuse emission which is present north of W50 forumla and to the north-west (the bright, extended nebula LBN 109; Lynds 1965) is outside W50's radio borders and not associated with the remnant.

2.1.2 High-resolution imagery of the eastern and western filaments

Optical images at higher angular resolution of the eastern and western filaments were also obtained with the 1.3-m (f/7.7) Ritchey-Cretien telescope at Skinakas Observatory during 2005 September 5, 9–10 and 2006 July 27–28 using an [O iii] 5007 Å and an forumla interference filters, respectively. The detector was a 1024 × 1024 SITe CCD with a field of view of 8.5 × 8.5 arcmin2. 10 exposures through the [O iii] 5007 Å filter each of 2400 s duration and 10 corresponding exposures in the continuum, each of 180 s, were taken of the eastern filaments, and similarly two forumla 6548, 6584 Å and continuum images of the western filaments were obtained. The continuum-subtracted mosaics of these images are shown in Figs 4 and 5, respectively. The fine filamentary nature of the eastern nebulosity is revealed in Figs 4 and 5. At a distance of 5 kpc the finest [O iii] 5007 Å filaments are ≈ 7 × 1016 cm (≡1 arcsec) wide.

Figure 4

The eastern complex taken with the 1.3-m telescope in the light of [O iii] 5007 Å. The positions of slits 1 and 2 are marked. The image has been smoothed to suppress the residual from the imperfect continuum subtraction.

Figure 4

The eastern complex taken with the 1.3-m telescope in the light of [O iii] 5007 Å. The positions of slits 1 and 2 are marked. The image has been smoothed to suppress the residual from the imperfect continuum subtraction.

Figure 5

The western complex taken with the 1.3-m telescope in the light of forumla 6548, 6584 Å. The image has been smoothed to suppress the residual from the imperfect continuum subtraction.

Figure 5

The western complex taken with the 1.3-m telescope in the light of forumla 6548, 6584 Å. The image has been smoothed to suppress the residual from the imperfect continuum subtraction.

2.1.3 Deep, high-resolution imagery of the northern radio ridge

The forumla 6548, 6584 Å image in Fig. 6 was obtained on the 2003 May 1 on one position of the extensive northern ridge of radio emission (Dubner et al. 1998, and shown here in Fig. 1). For this purpose the Manchester Echelle Spectrometer at the San Pedro Martir Observatory (MES-SPM; Meaburn et al. 1984, 2003) in its imaging mode was combined with the 2.1-m San Pedro Martir (Mexico) telescope (see Section 2.2.2 for plate scale). The integration time was 2400 s and a SITe CCD (2 × 2 binned) was the detector. Faint nebular filaments parallel to the radio ridge, can be seen for the first time. No star subtraction was possible in this case for a continuum image was not obtained.

Figure 6

The northern ridge in the light of forumla. The slit position for the PV array shown in Fig. 12 is marked as a vertical line. The slit centre is indicated by a horizontal line for the slit longer than this image. The image was not continuum subtracted and attempts to remove star images by the starlink routine patch left too many confusing residuals so was not pursued. The dark patch at the top of the image is caused by the halo of a bright star image just off the field.

Figure 6

The northern ridge in the light of forumla. The slit position for the PV array shown in Fig. 12 is marked as a vertical line. The slit centre is indicated by a horizontal line for the slit longer than this image. The image was not continuum subtracted and attempts to remove star images by the starlink routine patch left too many confusing residuals so was not pursued. The dark patch at the top of the image is caused by the halo of a bright star image just off the field.

2.2 Spectroscopy

2.2.1 Low dispersion – eastern and western filaments

Low-dispersion long-slit spectra were obtained with the 1.3-m telescope at Skinakas Observatory in 2004 June 14 and 2005 September 6–7. The 1300 line mm−1 grating was used in conjunction with a 2000 × 800 SITe CCD (15 × 15 μm2 pixels) resulting in a scale of 1 Å pixel−1 and covers the range of 4750–6815 Å. The spectral resolution is ∼8 and ∼11 pixels full width at half-maximum (FWHM) for the red and blue wavelengths, respectively. The slit width is 7.7 arcsec and in all cases was oriented in the south–north direction; the slit length is 7.9 arcmin. The spectrophotometric standard stars HR 5501, HR 7596, HR 9087, HR 718 and HR 7950 (Hamuy et al. 1994) were observed to calibrate the spectra.

The deep low-resolution spectra were taken on the relatively bright optical filament in the eastern and western parts of W50 (their exact positions are given in Table 1). In Table 2, we present the relative line fluxes taken from three different apertures (I, II and III) along each slit. In particular, apertures I, II and III have an offset (see Table 1) north or south of the slit centre which were selected because they are free of field stars in an otherwise crowded field and they include sufficient line emission to permit an accurate determination of the observed line fluxes. The background extraction aperture was taken towards the northern end or the southern ends of the slits depending on the slit position. The signal-to-noise ratios (S/N) presented in Table 2 do not include calibration errors, which are less than 10 per cent. Typical spectra from the eastern and western filaments are shown in Fig. 7.

Table 2

Relative line fluxes.

Line (Å) East 
Area EI Area EII Area EIII Area EIV Area EV 
Fa Ib S/Nc F I S/N F I S/N F I S/N F I S/N 
Hβ 4861 16 35 (7) 11 35 (6) 15 35 (5) 35 (1) 23 35 (4) 
[O iii] 4959 25 52 (9) 23 68 (13) 22 49 (8) – – – – – – 
[O iii] 5007 66 134 (27) 61 174 (31) 84 181 (25) 157 488 (28) 70 103 (14) 
[O i] 6300 67 74 (38) 78 90 (61) 87 97 (38) 34 40 (13) 44 46 (19) 
[O i] 6363 17 18 (15) 22 25 (22) 26 28 (16) – – – 45 47 (20) 
[N ii] 6548 106 107 (70) 116 117 (90) 105 106 (61) 75 76 (29) 101 102 (43) 
Hα 6563 100 100 (65) 100 100 (81) 100 100 (55) 100 100 (38) 100 100 (41) 
[N ii] 6584 343 340 (154) 377 373 (210) 370 367 (159) 228 225 (78) 316 315 (109) 
[S ii] 6716 128 121 (80) 139 128 (110) 172 162 (88) 50 46 (24) 140 136 (56) 
[S ii] 6731 89 84 (61) 98 90 (81) 122 114 (63) 33 30 (15) 102 99 (41) 
Absolute Hα fluxd 16.3 14.7 5.8 2.9 4.9 
[S ii]/Hα 2.17 ± 0.06 2.36 ± 0.05 2.93 ± 0.09 0.83 ± 0.06 2.42 ± 0.11 
F(6716)/F(6731) 1.43 ± 0.11 1.42 ± 0.03 1.41 ± 0.04 1.49 ± 0.18 1.38 ± 0.06 
[O iii]/Hβ 5.27 ± 0.49 6.83 ± 0.46 6.50 ± 0.73 13.74 ± 4.03 2.91 ± 0.62 
c(Hβ) 0.98 ± 0.12 1.45 ± 0.13 1.06 ± 0.15 1.57 ± 0.37 0.53 ± 0.17 
E(BV0.68 ± 0.08 1.00 ± 0.09 0.73 ± 0.10 1.09 ± 0.26 0.37 ± 0.12 
Line (Å) West  
Area WI Area WII Area WIII  
F I S/N F I S/N F I S/N  
Hβ 4861 14 35 (2) 35 (2) 14 35 (3)  
[O iii] 4959 28 66 (2) 23 104 (3) 37 88 (4)  
[O iii] 5007 89 204 (6) 87 373 (9) 159 364 (10)  
[N ii] 6548 79 80 (10) 87 88 (18) 75 76 (10)  
Hα 6563 100 100 (12) 100 100 (20) 100 100 (12)  
[N ii] 6584 257 255 (29) 275 271 (52) 301 298 (33)  
[S ii] 6716 82 77 (11) 120 107 (25) 112 105 (14)  
[S ii] 6731 74 69 (10) 92 81 (20) 107 100 (14)  
Absolute Hα flux 1.4 2.6 1.4  
[S ii]/Hα 1.56 ± 0.27 1.86 ± 0.19 2.12 ± 0.33  
F(6716)/F(6731) 1.11 ± 0.23 1.31 ± 0.14 1.21 ± 0.16  
[O iii]/Hβ 7.64 ± 2.32 13.40 ± 1.81 12.78 ± 2.40  
c(Hβ) 1.15 ± 0.67 2.01 ± 0.76 1.15 ± 0.65  
E(BV0.80 ± 0.46 1.39 ± 0.52 0.80 ± 0.45  
Line (Å) East 
Area EI Area EII Area EIII Area EIV Area EV 
Fa Ib S/Nc F I S/N F I S/N F I S/N F I S/N 
Hβ 4861 16 35 (7) 11 35 (6) 15 35 (5) 35 (1) 23 35 (4) 
[O iii] 4959 25 52 (9) 23 68 (13) 22 49 (8) – – – – – – 
[O iii] 5007 66 134 (27) 61 174 (31) 84 181 (25) 157 488 (28) 70 103 (14) 
[O i] 6300 67 74 (38) 78 90 (61) 87 97 (38) 34 40 (13) 44 46 (19) 
[O i] 6363 17 18 (15) 22 25 (22) 26 28 (16) – – – 45 47 (20) 
[N ii] 6548 106 107 (70) 116 117 (90) 105 106 (61) 75 76 (29) 101 102 (43) 
Hα 6563 100 100 (65) 100 100 (81) 100 100 (55) 100 100 (38) 100 100 (41) 
[N ii] 6584 343 340 (154) 377 373 (210) 370 367 (159) 228 225 (78) 316 315 (109) 
[S ii] 6716 128 121 (80) 139 128 (110) 172 162 (88) 50 46 (24) 140 136 (56) 
[S ii] 6731 89 84 (61) 98 90 (81) 122 114 (63) 33 30 (15) 102 99 (41) 
Absolute Hα fluxd 16.3 14.7 5.8 2.9 4.9 
[S ii]/Hα 2.17 ± 0.06 2.36 ± 0.05 2.93 ± 0.09 0.83 ± 0.06 2.42 ± 0.11 
F(6716)/F(6731) 1.43 ± 0.11 1.42 ± 0.03 1.41 ± 0.04 1.49 ± 0.18 1.38 ± 0.06 
[O iii]/Hβ 5.27 ± 0.49 6.83 ± 0.46 6.50 ± 0.73 13.74 ± 4.03 2.91 ± 0.62 
c(Hβ) 0.98 ± 0.12 1.45 ± 0.13 1.06 ± 0.15 1.57 ± 0.37 0.53 ± 0.17 
E(BV0.68 ± 0.08 1.00 ± 0.09 0.73 ± 0.10 1.09 ± 0.26 0.37 ± 0.12 
Line (Å) West  
Area WI Area WII Area WIII  
F I S/N F I S/N F I S/N  
Hβ 4861 14 35 (2) 35 (2) 14 35 (3)  
[O iii] 4959 28 66 (2) 23 104 (3) 37 88 (4)  
[O iii] 5007 89 204 (6) 87 373 (9) 159 364 (10)  
[N ii] 6548 79 80 (10) 87 88 (18) 75 76 (10)  
Hα 6563 100 100 (12) 100 100 (20) 100 100 (12)  
[N ii] 6584 257 255 (29) 275 271 (52) 301 298 (33)  
[S ii] 6716 82 77 (11) 120 107 (25) 112 105 (14)  
[S ii] 6731 74 69 (10) 92 81 (20) 107 100 (14)  
Absolute Hα flux 1.4 2.6 1.4  
[S ii]/Hα 1.56 ± 0.27 1.86 ± 0.19 2.12 ± 0.33  
F(6716)/F(6731) 1.11 ± 0.23 1.31 ± 0.14 1.21 ± 0.16  
[O iii]/Hβ 7.64 ± 2.32 13.40 ± 1.81 12.78 ± 2.40  
c(Hβ) 1.15 ± 0.67 2.01 ± 0.76 1.15 ± 0.65  
E(BV0.80 ± 0.46 1.39 ± 0.52 0.80 ± 0.45  

aObserved fluxes normalized to F(Hα) = 100 and uncorrected for interstellar extinction. bIntrinsic fluxes normalized to F(Hα) = 100 and corrected for interstellar extinction. cNumbers in parentheses represent the S/N of the quoted fluxes.

dIn units of 10−17 erg s−1 cm−2 arcsec−2. Listed fluxes are a S/N weighted average of two fluxes.

The emission line ratios [S ii]/Hα, F(6716)/F(6731) and [O iii]/Hβ are calculated using the corrected for interstellar extinction values.

The errors of the emission line ratios, c(Hβ) and E(BV) are calculated through standard error propagation.

Figure 7

Typical observed spectra from different positions of W50 (see Table 2).

Figure 7

Typical observed spectra from different positions of W50 (see Table 2).

Interstellar reddening was derived from the Hα/Hβ ratio (Osterbrock 1989), using the interstellar extinction law by Fitzpatrick (1999) and RV= 3.1. Therefore, the interstellar logarithmic extinction coefficient c(Hβ) can be derived by using the relationship  

1
formula
where 0.348 is the relative logarithmic extinction coefficient for Hβ/Hα and 2.85 the theoretical value of F(Hα)/F(Hβ). Here we have used the ratio of 2.85 though Hartigan, Raymond & Hartmann (1987) suggest 3.0 which could be more appropriate when some collisional excitation is present. The observational reddening in magnitude E(BV) was also calculated using the relationship (Seaton 1979)  
2
formula
where the extinction parameter Xβ= 3.615 (Fitzpatrick 1999).

The errors on the measurements of c(Hβ) and E(BV) were calculated through standard error propagation of equations (1) and (2). Consequently, c was found to be between 0.53 and 2.00 (to give AV of 1.1 to 4.2) and E(BV) between 0.4 and 1.4.

2.2.2 High dispersion – eastern and western filaments

Spatially resolved, long-slit spectra were obtained of the eastern filamentary nebulosity on 2005 August 2–4 with the MES-SPM (Meaburn et al. 1984, 2003) combined with the 2.1-m San Pedro Martir (Mexico) telescope. The slit, orientated east–west, was 300 μm wide (≡3.9 arcsec and 20 km s−1). The 512 increments of the 2 × 2 binned SITe CCD detector, each 0.624 arcsec long, give a total projected slit length of 5.32 arcmin on the sky. In this spectroscopic mode MES-SPM has no cross-dispersion, consequently, for the present observations, a filter of 90 Å bandwidth was used to isolate the 87th echelle order containing the Hα and [N ii] 6548, 6584 Å nebular emission lines. Integration of 1800 and 3600 s was obtained, respectively, for slit positions 1 and 2 marked in Fig. 4. The position velocity (PV) arrays of [N ii] 6584 Å line profiles from slit positions 1 and 2 are shown in Figs 8 and 9, respectively. The [N ii] 6584 Å line profiles from the incremental lengths marked in Figs 8 and 9 are shown in Figs 10 and 11, respectively.

Figure 8

Light and dark representations of the PV array of [N ii] 6584 Å line profile from slit position 1 (see Fig. 4).

Figure 8

Light and dark representations of the PV array of [N ii] 6584 Å line profile from slit position 1 (see Fig. 4).

Figure 9

As for Fig. 8 but for slit position 2. The vertical dark lines are the continuous spectra of faint field stars also intercepted by the slit.

Figure 9

As for Fig. 8 but for slit position 2. The vertical dark lines are the continuous spectra of faint field stars also intercepted by the slit.

Figure 10

The [N ii] 6584 Å line profiles from the incremental lengths marked in Fig. 8.

Figure 10

The [N ii] 6584 Å line profiles from the incremental lengths marked in Fig. 8.

Figure 11

The [N ii] 6584 Å line profiles from the incremental lengths marked in Fig. 9.

Figure 11

The [N ii] 6584 Å line profiles from the incremental lengths marked in Fig. 9.

It can be seen in Figs 7–11 that the brightest filaments (section A in Figs 8 and 10 and section C in Figs 9 and 11) emit 22 km s−1 wide profiles when corrected for instrumental broadening, both centred on Vhel= 56 km s−1. This value is significantly different from Vsys≈ 40 km s−1 which is the mean of the estimation of 28 km s−1 from the H i measurements of Dubner et al. (1998) and 52 km s−1 from optical profiles of the eastern and western filaments (Figs 2 and 3) given by Mazeh et al. (1983). In the present paper it has been assumed that for W50, Vhel=VLSR− 14.3 km s−1.

However, the fainter [N ii] 6584 Å emitting regions between the bright eastern filaments exhibit approaching radial velocities of up to 100 km s−1 with respect to Vhel= 56 km s−1. These faint but extensive high-speed regions are particularly prominent for the sections D and B in Fig. 9 whose profiles are shown in Fig. 11.

These high-dispersion spectral observations of the eastern filamentary nebulosity were obtained in non-photometric conditions and therefore were not compared photometrically with a standard star.

2.2.3 High dispersion – northern ridge

Nearly the same instrumental set-up and data analysis as described in Section 2.2.2 was employed on 2003 May 1 for the long-slit spectroscopy of the northern nebulosity shown in Fig. 6. Here the slit was now 150 μm wide (≡1.9 arcsec and 9 km s−1), orientated north–south and centred on RA forumla, Dec. forumla (J2000). The integration time was 4200 s. The PV arrays of [N ii] 6584 Å profiles obtained in this way are shown in Fig. 12 and the line profile for the section marked A in Fig. 12 is shown in Fig. 13.

Figure 12

A negative grey-scale representation of the PV array of [N ii] 6584 Å line profiles of the northern ridge. The slit was orientated north–south (see Fig. 6). The whole slit length is shown in order to emphasize that detection of line profiles has only occurred over a small part of the slit length, i.e. we are not detecting diffuse galactic emission along the same sight line. North is to the left.

Figure 12

A negative grey-scale representation of the PV array of [N ii] 6584 Å line profiles of the northern ridge. The slit was orientated north–south (see Fig. 6). The whole slit length is shown in order to emphasize that detection of line profiles has only occurred over a small part of the slit length, i.e. we are not detecting diffuse galactic emission along the same sight line. North is to the left.

Figure 13

The [N ii] 6584 Å line profile from the incremental length A marked in Fig. 12.

Figure 13

The [N ii] 6584 Å line profile from the incremental length A marked in Fig. 12.

This spectral observation was calibrated photometrically against the slitless spectrum of the standard star Feige 56 to give a value, uncorrected for interstellar extinction, for the total emission in this [N ii] 6584 Å profile of 1.9 × 10−17 erg s−1 cm−2 arcsec−2 to 20 per cent accuracy. This is ≈29 times fainter in the [N ii] 6584 Å line than the brightest filament in the eastern nebulosity as listed in Table 1. The [N ii] 6584 Å profile is centred on Vhel= 56 ± 2 and 36 km s−1 wide when simulated by a single Gaussian. When corrected for instrumental broadening this width reduces to 35 km s−1.

3 DISCUSSION

3.1 Location of optical emission

The Hα/Hβ ratios listed in Table 2 strongly indicate that heavy and patchy absorption of optical emission occurs over W50. This is to be expected for a distance of 5 kpc along the Galactic plane and for a Galactic latitude of only forumla. This possibility is further supported by the ISOCAM (Infrared Space Observatory Camera) infrared emission map of the vicinity of the western filaments obtained by Moldowan et al. (2005). One patch of infrared emission coincides closely with the visible western filaments (Figs 3a and b and 5) but a larger region of infrared emission has no optical counterpart in our images. It is therefore probable that only parts of the shock-excited optical emission from W50 is being observed in Figs 2–5 with considerable parts heavily obscured by the foreground dust.

This patchiness by foreground dust as the cause of the limited detection of the whole of the W50 optical filaments, which would be expected if this is indeed an SNR, could be confirmed by the optical detection in Figs 6, 12 and 13 for the first time of the northern radio ridge. Although the latter's optical emission is very much fainter than that observed from the eastern and western optical filaments the [N ii] 6584 Å profile from it in Fig. 13 is centred on Vhel= 56 km s−1. This matches closely the values in Figs 10 and 11 for the brightest and narrowest profiles of the eastern filamentary nebulosity. It is suggested that in all of these regions the motions of the emitting gas are nearly perpendicular to the sightline and that Vhel= 56 km s−1 could be the best value of the systemic heliocentric radial velocity, Vsys for the whole SS 433 and W50 complex. In these circumstances the optical emission from the northern radio ridge would have to be heavily obscured by the patchy dust. One complication to this patchy dust interpretation is that the relativistic jet of SS 433 has injected ≈2 × 1051 erg into the surrounding medium over its lifetime and could have affected the recombination rates of the shocked W50 gas in unpredictable ways.

3.2 Morphology of the eastern and western optical filaments

Taken as a whole, the new optical imagery in Figs 2–5 reveals that the eastern filaments trace a broad arc that is convex with respect to the central source and follows the large-scale helical morphology of the W50 boundary in this vicinity as revealed by radio continuum maps (Dubner et al. 1998, see their more detailed maps to appreciate this structure). Interestingly, the western filaments seem to curve in an identical sense despite being on the opposite side of SS 433 (at all wavelengths the morphology of the western ‘ear’ is most likely distorted by a strong interaction with a dense portion of the local ISM).

The bi-polar relativistic jet from SS 433 appears to have broken through diametrically opposite boundaries of the W50 SNR assuming that this is the origin of large circular region of radio emission in Fig. 1. The bright optical emission (Figs 2–5) then lies at these breakout regions. In an idealized case, a breakout region should take the form of a ring of emission where the expanding jet envelope shocks the dense shell of swept up and compressed material at the SNR boundary. Optical emission from this interaction is not expected elsewhere because the interior is too hot and rarefied and the ambient medium is too low density ahead of the jet. The observed arc morphology, i.e. eastern and western arcs, could be simply the result of heavy, patchy foreground interstellar absorption (see Section 3.1) combined with the jet axis being tilted so that the eastern tip is pointing towards the observer and western tip away (see references in Fabrika 2004, that confirm this jet orientation); the rings of optical emission become partial ellipses with only localized portions on their nearsides with respect to the observer being visible. However, the interaction is likely to be much more complex than described above not least because the jet close to SS 433 has a ‘corkscrew’ structure. Furthermore, Velázquez & Raga (2000) have numerically modelled the W50/SS 433 system which in their simulations results in reflected and secondary shocks as well as the filling and acceleration of the whole SNR by the jet cocoon. None the less the breakout region remains a distinct zone in the simulations.

3.3 The shocked emission

All of the spectra in Table 2 show clearly that the observed optical emission originates in shock-heated gas, since the [S ii] 6717 and 6731 Å/Hα > 1.5. The [O iii] 5007 Å emission detected in both spectra suggests a shock velocity greater than 100 km s−1 (Cox & Raymond 1985). The absolute Hα flux covers a range of values from 1.4 to 2.6 × 10−17 and 5.8 to 16.3 × 10−17 erg s−1 cm−2 arcsec−2 for the western and eastern nebulosities, respectively. The [S ii] ratio which was found to be between 1.1 and 1.4. Using ‘temden’ in the nebular package in iraf (Shaw & Dufour 1995) the electron densities of 50 and 700 cm−3 are measured for the eastern and western filaments, respectively. A comparison to shock models shows that the densities are different (east to west) by a factor of 10 or more, while the shock velocities are nearly the same. Hence, either there is a huge pressure differential between the two sides (nT∼ pressure) or there is something incorrect with this interpretation. Furthermore, Hβ emission was detected in both areas (but with low S/N) to give [O iii] 5007 Å/Hβ ratios of between 5.6 (area EI) and 15.7 (area WII). Theoretical models of Cox & Raymond (1985) and Hartigan et al. (1987) suggest that for shocks with complete recombination zones the expected [O iii] 5007 Å/Hβ ratio is ∼6, while this limit is exceeded in the case of a shock with incomplete recombination zones (Raymond et al. 1988). Our measured values suggest in the western nebulosity that shocks with incomplete recombination zones are present while in the eastern nebulosity the presence of shocks with complete recombination zones could not be ruled out. A combination therefore of the present observations with the theoretical predictions suggests that shock velocities are ≃100 and ≃120 km s−1 for the eastern and western nebulosities, respectively, though this difference may not be real when uncertainties in the data are considered.

3.4 Velocity of the optical filaments

The previous optical spectra of Mazeh et al. (1983) at very low angular resolution of the eastern and western filaments suggest that the average heliocentric radial velocities (Vhel) are, respectively, 65 and 40 km s−1[where the LSR (local standard of rest) and heliocentric radial velocities are related by Vhel=VLSR− 14.3 km s−1]. This is in apparent contradiction with the orientation of the jet, which has the eastern portion directed towards the observer. The radio morphology also indicates that the axis of the elongated lobes of the W50 shell is tilted to the plane of the sky with the eastern side nearest the observer.

The present spectral observations, with their higher spatial resolution, show that the brightest regions of the eastern filaments (Figs 8–11) emit the narrowest lines centred on Vhel= 56 km s−1 but with extensive fainter regions, flowing off the filaments (Fig. 9), composed of high-velocity gas approaching the observer with radial velocities continuously up to 100 km s−1 from the bright filament value. It remains possible that similar high spatial resolution observations of the western filaments will show similar complex motions and eliminate this discrepancy which could only be a consequence of the low angular resolution employed in the early measurements. However, the appropriate flows emitting faintly in the eastern filaments are consistent with eastern side of W50 pointing towards the observer.

Furthermore, as noted above, the expected local shock velocities that give rise to the optical filaments are of order 100 km s−1, and the expansion of the remnant as a whole is estimated by Dubner et al. (1998) to be ∼75 km s−1. Thus the measured difference in radial velocity between the bright optical filaments is small compared with the expected velocity range (∼100 km s−1). Any residual difference could be explained easily by patchy dust absorption hiding the full extent of the optical emission.

Also the contradiction only exists if it is assumed that the optical emission is tracing directly the outflow from the star. However, the optical emission much more likely traces the interaction between the jet cocoon and the shell of the W50 SNR. Expansion of this shocked optically emitting gas could result in localized flows towards and/or away from the star (cf. simulations of Velázquez & Raga 2000, their fig. 3). In the case of the eastern filaments, the curved morphology could then be a result of an expansion back towards SS 433 whereas the western filaments could be a result of an expansion away from or stationary with respect to SS 433. The optical emission then is concentrated at the ‘rims’ of the breakout region.

Broader, competing, possibilities should also be considered for it has been suggested as above that these elongated breakout features in the radio map are the microquasar jets currently penetrating through the shell of the W50 SNR (Dubner et al. 1998). Alternatively, these eastern and western radio lobes of the SNR may simply be revealing the imprint of the precessing jets of SS 433, created early in the object's evolution after the explosion of the supernova and which have inflated along with the rest of the SNR as it expanded. The faint ≥100 km s−1 outflows from the eastern filaments around this lobe's apparent breakout region would favour the first possibility.

We would like to thank the referee for constructive comments that have improved the paper considerably. We also thank G. Dubner who kindly provided us the radio image of W50 in FITS format and the staff at Skinakas and SPM Observatories for their excellent support during these observations. JA and SA acknowledge funding by the European Union and the Greek Ministry of Development in the framework of the programme ‘Promotion of Excellence in Research Institutes (2nd Part)’. JAL acknowledges financial support from UNAM grants IN 112103, 108406 and 108506. Skinakas Observatory is a collaborative project of the University of Crete, the Foundation for Research and Technology-Hellas and the Max-Planck-Institut für Extraterrestrische Physik.

REFERENCES

Band
D. L.
,
1987
,
PASP
 ,
99
,
1269
Blundell
K. M.
Bowler
M. G.
,
2004
,
ApJ
 ,
616
,
L159
Boumis
P.
Mavromatakis
F.
Paleologou
E. V.
Becker
W.
,
2002
,
A&A
 ,
396
,
225
Brinkmann
W.
Aschenbach
B.
Kawai
N.
,
1996
,
A&A
 ,
312
,
306
Cox
D. P.
Raymond
J. C.
,
1985
,
ApJ
 ,
298
,
651
Dubner
G.
,
2002
, in
Pramesh Rao
A.
Swarup
G.
Gopal-Krishna
, eds, IAU Symp. 199,
The Universe at Low Radio Frequencies
 . p.
284
Dubner
G. M.
Holdaway
M.
Goss
W. M.
Mirabel
I. F.
,
1998
,
AJ
 ,
116
,
1842
Fabrika
S.
,
2004
,
Astrophys. Space Phys. Rev
 .,
12
,
1
Fitzpatrick
E. L.
,
1999
,
PASP
 ,
111
,
63
Green
D. A.
,
2006
,
A Catalog of Galactic Supernova Remnants
 .
Mullard Radio Astronomy Observatory
, Cambridge
Hamuy
M.
Suntzeff
N. B.
Heathcote
S. R.
Walker
A. R.
Gigoux
P.
Phillips
M. M.
,
1994
,
PASP
 ,
106
,
566
Hartigan
P.
Raymond
J.
Hartmann
L.
,
1987
,
ApJ
 ,
316
,
323
Holden
D. J.
Caswell
J. L.
,
1969
,
MNRAS
 ,
143
,
407
Kirshner
R. P.
Chevalier
R. A.
,
1980
,
ApJ
 ,
242
,
L77
Lasker
B. M.
Russel
J. N.
Jenkner
H.
,
1999
,
The HST Guide Star Catalog, version 1.1-ACT
 .
The Association of Universities for Research in Astronomy
, Inc
Lynds
B. T.
,
1965
,
ApJS
 ,
12
,
163
Mazeh
T.
Aguilar
L. A.
Treffers
R. R.
Königl
A.
Sparke
L. S.
,
1983
,
ApJ
 ,
265
,
235
Meaburn
J.
Blundell
B.
Carling
R.
Gregory
D. F.
Keir
D.
Wynne
C. G.
,
1984
,
MNRAS
 ,
210
,
463
Meaburn
J.
López
J. A.
Gutiérrez
L.
Quiróz
F.
Murillo
J. M.
Valdéz
J.
Pedrayez
M.
,
2003
,
Rev. Mex. Astron. Astrofis
 .,
39
,
185
Moldowan
A.
Safi-Harb
S.
Fuchs
Y.
Dubner
G.
,
2005
,
Adv. Space Res
 .,
35
,
1062
Murdin
P.
Clark
D. H.
,
1980
,
MNRAS
 ,
190
,
65
p
Osterbrock
D. E.
,
1989
,
Astrophysics of Gaseous Nebulae
 .
Freeman & Co
., San Francisco
Raymond
J. C.
Hester
J. J.
Cox
D.
Blair
W. P.
Fesen
R. A.
Gull
T. R.
,
1988
,
ApJ
 ,
324
,
869
Safi-Harb
S.
Ögelman
H.
,
1997
,
ApJ
 ,
483
,
868
Seaton
M. J.
,
1979
,
MNRAS
 ,
187
,
73
p
Shaw
R. A.
Dufour
R. J.
,
1995
,
PASP
 ,
107
,
896
Shuder
J. M.
Hatfield
B. F.
Cohen
R. D.
,
1980
,
PASP
 ,
92
,
259
Stirling
A. M.
Jowett
F. H.
Spencer
R. E.
Paragi
Z.
Ogley
R. N.
Cawthorne
T. V.
,
2002
,
MNRAS
 ,
337
,
657
Van Den Bergh
S.
,
1980
,
ApJ
 ,
236
,
L23
Velázquez
P. F.
Raga
A. C.
,
2000
,
A&A
 ,
362
,
780
Westerhout
G.
,
1958
,
Bull. Astron. Inst. Neth
 .,
14
,
215
Zealey
W. J.
Dopita
M. A.
Malin
D. F.
,
1980
,
MNRAS
 ,
192
,
731