I. Gosachinskiy " HI Distribution in the Region of the Supernova Remnant G78.2+2.1 ".


Based on RATAN-600 21cm line observations with an angular resolution of 2.4' over a wide range of radial velocities, we analyze the neutral-hydrogen distribution in the region of the supernova remnant ( SNR ) G78.2+2.1.
In addition to an HI shell at low radial velocities immediately surrounding the radio remnant, we detected an extended expanding HI shell, 3 dg. in diameter, at a radial velocity of -25 km/s, which closely coincides in coordinates and angular sizes with the outer X-ray shell discovered by Lozinskaya et al. ( 2000 ).
The Halpha emission studied by these authors in the SNR region also has a secondary peak at radial velocities from -45 to -20 km/s. Since the radial velocities of these two objects differ markedly, their distances can be assumed to differ as well; i.e., a chance projection of two distinct objects is observed.


The well-known shell-type supernova remnant ( SNR ) G78.2+2.1 see Green ( 1996 ), Gosachinskiy et al. ( 1999 ), and Lozinskaya et al. ( 2000 ) for references is projected onto the region of the Cygnus X radio source.
More than half a hundred various objects, mostly compact H II regions ( Piepenbrink and Wendker 1988 ), are located there in a small area of the sky ( a mere 6-7 dg. in diameter ). Only three of these objects, judging from their H110alpha radial velocities measured by the above authors, are projected by chance onto this region from more distant spiral arms.
At the same time, the interpretation of observations in this region is complicated by uncertainties in distances, especially kinematic ones, because of the small dv/dr, and by the fact that the line of sight here passes along the Orion-Cygnus spiral arm.
The interstellar medium in this region has been studied by dozens of authors over various wavelength ranges and in various spectral lines. Gosachinskiy et al. ( 2000 ) discovered two annular HI structures, which morphologically resemble expanding shells.
One of them, with a larger angular size, is located around the Cygnus X radio source, and the other, at a mean radial velocity of +3 km/s, immediately surrounds SNR G78.2+2.1. A quantitative analysis shows that it could be produced both by the SNR shock wave and by the progenitor star's stellar wind. Here, we also give references to previous studies of neutral hydrogen in the SNR region and critically analyze the results of some of them.
The archival ROSAT and ASCA X-ray data were analyzed by Lozinskaya et al. ( 2000 ). Their analysis revealed two clumpy nested X-ray shells, which may be associated with this SNR. The surprising thing is that, whereas one of them, composed of three clouds, is only slightly larger than the SNR, the angular diameter of the second, weaker shell is three times that of the SNR.
Bearing in mind that this region is very rich in various objects, a chance projection of X-ray features onto one another and onto the SNR cannot be ruled out either.
On the other hand, high-angular-resolution observations are known to commonly reveal shells or cavities in the HI distribution where active objects with high energy release are located: SNRs, H II regions, or stellar-wind-associated shells,let alone X-ray emitting regions.
For this reason, when investigating the relationship of the HI distribution to SNR G78.2+2.1 and other objects, it makes sense not to restrict the analysis to radial velocities near zero, which corresponds to the SNR distance, but consider the entire HI-line range. Below, we present our results.


We used 18 cuts in right ascension at 0.6 dg. intervals in declination obtained with the RATAN-600 radio telescope to investigate the interstellar-gas distribution in the region of the Cygnus X radio source and SNR G78.2+2.1.
The instrumentation and techniques are detailed in Gosachinskiy et al. ( 1999 ), but, for convenience, we give their brief description. In this elevation range, the RATAN-600 antenna at 21 cm has an angular resolution of 2' x 12', an effective area of 875 m2, and substantial ( up to 40% ) losses in brightness temperature.
The latter is attributable to the peculiar antenna design and to the fact that the observations were carried out near zenith. The system noise temperature is 60 K; the 39-channel filter spectrum analyzer has a channel bandwidth of 30 kHz ( 6.3 km/s ).
The cut at each declination consists of two series of three observations each obtained by shifting the receiver tuning frequency by half the channel bandwidth. As a result, each cut has 78 channels that follow at intervals of 3.15 km/s.
This observing technique also allows an effective noise cleaning. The mean square of fluctuations in spectral channels on an averaged scan is 0.2 K. The antenna and equipment parameters were checked in each series of observations by measuring a set of reference sources ( Venger et al. 1979 ).
Subsequently, we subtracted an extended background obtained by spline interpolation at lower brightness-distribution level from the transit curves in each spectral channel and then reduced the transit curves containing only features of small angular sizes.
The subtracted background part of the transit curves includes:
  • (1) large-scale features of the interstellar-gas distribution, such as spiral arms or giant complexes;
  • (2) emission from the intercloud medium, if present;
  • (3) features of small angular sizes unresolved by the RATAN-600 beam;
  • (4) and spurious large-scale background produced by distant side lobes and the RATAN-600 stray field.
It should be noted that subtracting the background by the above method could result in an underestimation of the brightness and angular sizes of the remaining small-scale features, whose parameters were determined by using a code of Gauss analysis.
The measured parameters have the following errors. The radial velocity of an isolated medium-brightness HI feature is measured with an accuracy of 1 km/s. In several cases noted below, the accuracy is reduced because of the difficulties in separating the object from the background or from neighboring features.
Given the antenna calibration errors, the measurement error of the HI line brightness temperature is about 0.5 K, while the error of the estimated angular sizes in right ascension is 0.1 dg.
The same qualifications as those for the radial velocities hold in the latter case. The antenna resolution in declination is much lower, and, accordingly, the accuracy of measuring angular sizes is lower.
The accuracy of estimating distances depends on the method of their determination, and it should be considered separately in each case. As a result, the accuracy of estimating the HI mass turns out to be no higher than 0.5-1 order of magnitude.


The brightness distribution of HI features at a radial velocity of +10 km/s after the extended radio-line emission background in the region of the Cygnus X radio source and SNR G78.2+2.1 has been subtracted is shown in Fig. 1.
The declinations of the cuts are indicated on the right; the vertical and horizontal scales slightly differ. The annular HI structures that can be correlated with the above objects, as was done by Gosachinskiy et al. ( 1999 ), are represented by thin lines.
A schematic 21cm radio image of the SNR from Higgs et al. ( 1977 ) is shown in Fig. 2 ( heavy circle ) together with X-ray ( 0.5-2.0 keV ) isophotes from Lozinskaya et al. ( 2000 ). In this figure, the vertical and horizontal scales are approximately the same.
Gosachinskiy et al. ( 1999 ) found that the smaller annular HI structure immediately surrounds the radio remnant and determined the following observed parameters of the HI shell around the SNR:
  • Coordinates of the center R.A. alpha ( 2000.0 ) = 20h21.3m, delta ( 2000.0 ) = +40.8 dg.;
  • Angular sizes 2.8 dg. x 3.5 dg. ( inner ) and 2.0 dg. x 2.5 dg. ( outer );
  • Mean brightness temperature of the line 8 +- 0.5 K;
  • Mean radial velocity +3 km/s;
  • The radial-velocity range in which the HI shell was observed > 20 km/s.
When identifying the HI shell, we carefully allowed for the effect of the HI absorption line, because the SNR continuum emission was intense. We also compared the detected structure with the data of other authors who analyzed the H I distribution in this region.
We assumed the SNR distance to be 1.5 kpc ( Landecker et al. 1980 ) when calculating the following physical parameters of the HI shell around G78.2+2.1:
  • Large-scale velocity of radial motions > 10 km/s;
  • Outer diameter 75 pc;
  • Inner diameter 55 pc;
  • Gas density 2.5 cm-3;
  • Shell mass 8.1 x 103 M of the Sun.
If the HI shell around SNR G78.2+2.1 results from the impact of the shock wave generated by a supernova explosion on the interstellar gas, then the initial explosion energy and the SNR age can be estimated from observed HI-shell parameters.
The required density of the ambient interstellar medium, n0 = 1.6 cm-3, was estimated by Gosachinskii et al. ( 1999 ) by assuming that the HI-shell gas was initially "spread" over its entire volume.
The presence of an expanding HI shell around the SNR was also assumed to suggest that the remnant is in the radiative phase. In this case, using relations from Wheeler et al. ( 1980 ), we obtained the initial explosion energy and the SNR age:
E0 = 3 x 1050 erg and t = 1.1 x 106 years.
Although the derived initial explosion energy is in good agreement with the universally accepted value, the age proves to be too large for an X-ray emitting remnant.
As an alternative explanation, Gosachinskiy et al. ( 1999 ) assumed that an extended slow HI shell could be produced by the supernova progenitor's stellar wind. Using the calculations by Weaver et al. ( 1977 ), we then estimated the stellar-wind intensity and duration required for the formation of an HI shell from its observed parameters to be:
Lw = 1036 erg and t = 2.3 x 106 years.
Such an intensity is characteristic of the winds from OB stars with Minit > 8 M of the Sun, which produce supernova explosions, while the wind duration is an order of magnitude shorter than the main-sequence lifetimes of these stars. This confirms that the outer HI shell around SNR G78.2+2.1 could be produced by the supernova progenitor's stellar wind.
Data on the soft X-ray emission from the G78.2+2.1 region ( Lozinskaya et al. 2000 ) ( see Fig. 2 ) confuse the picture further still. The presence of the so-called inner X-ray shell, which consists of three extended clouds and coincides almost exactly with the radio remnant in coordinates and angular sizes, can still be explained somehow if the remnant is still in the adiabatic phase (see Lozinskaya et al. ( 2000 ) for a detailed discussion.
The weak outer clumpy X-ray shell ( of course, if it actually exists ) is offset by almost 1 dg. northwest of the radio-remnant center. In addition, its angular size is almost triple that of the remnant, which also completely disagree with the sizes and location of the HI shell detected by Gosachinskiy et al. ( 1999 ).
Note, however, that, in accordance with the assumed distance to the with the SNR only at nearly zero radial velocities. Clearly, the pattern of HI distribution should be studied over a much wider range of radial velocities.
The possible association of HI features with the above objects in the radio, optical, and X-ray ranges was analyzed in the radial-velocity range -125 to +35 km/s. It turned out that, apart from the features at positive radial velocities noted by Gosachinskiy et al. ( 1999 ), there is an interesting gas structure in the range -15 to -35 km/s.
The distribution of HI features at radial velocities in the range -20 to -30 km/s is shown in Fig. 3. A comparison of this map with Fig. 2 reveals a distinct cavity or even an HI shell, which closely coincides in coordinates and angular sizes with the outer X-ray shell from Lozinskaya et al. ( 2000 ).
This shell has the following observed parameters:
  • Coordinates of the center alpha (2000.0) = 20h20.0m, delta (2000.0) = +40.8 dg.;
  • Angular sizes 3.7 dg. x 4.7 dg. (inner) and 2.5 dg. x 3.5 dg. (outer);
  • Mean brightness temperature of the line 12 +- 0.5 K;
  • Mean radial velocity -25 km/s;
  • The radial-velocity range in which the HI shell was observed +- 10 km/s.
  • Note also that the angular sizes of the shell are at a maximum at a radial velocity of -25 km/s; this velocity can thus be considered to
  • This velocity can thus be considered to be the mean for the object.
At lower and higher radial velocities, the angular sizes of the annular structure decrease, while the center slightly displaces in right ascension. Such an isophotal structure is characteristic of a shell with a large-scale radial velocity component (expansion) and rotation.
Nevertheless, it should be noted that the cavity is not quite circular and is elongated from north to south. This, it cannot be reliably interpreted as an expanding shell. However, in the radial-velocity range studied, we failed to detect any other HI features whose morphology would correspond so closely to the outer X-ray shell found by Lozinskaya et al. (2000).
Whether the annular structure that we detected in the HI distribution is real can be verified, for example, by using data from a survey made with the 90-m NRAO radio telescope with an angular resolution of 12' and a radial-velocity resolution of 2 km/s (Westerhout and Wendlandt 1982).
Data from this survey are now accessible via the Internet in the ADS archive (Strasburg). Figure 4 shows the distributions of HI-line brightness temperature in Galactic longitude taken from this survey at the source latitude b = +2.1 dg. in the radial-velocity range -20.0 to -40.0 km/s.
The heavy and thin vertical lines mark, respectively, the position of the SNR center and its sizes at "zero" continuum radio brightness as inferred by Reich et al. (1990). There is a clear reduction in HI-line intensity southwest of the SNR at longitudes 76-77 dg.
This effect is at a maximum at radial velocities of about -30 km/s. Of course, the H I distribution from Westerhout and Wendlandt (1982) differs in details from our distribution, particularly the mean radial velocity of the cavity on the profiles in Fig. 4.
This can be explained, first, by a marked difference between the antenna beam shapes and a different radial-velocity resolution and, second, by the fact that Westerhout and Wendlandt (1982 provided data only for the southeastern part of the cavity up to b = +2.4 dg.
Lozinskaya et al. (2000) measured the Halpha radial velocities in the G78.2+2.1 region. It turned out that there is a secondary peak of the line profile at radial velocities in the range -20 to -45 km/s outside the bright remnant, but inside the outer weak X-ray shell, which closely corresponds to the velocity range of the H I shell we detected and confirms that it is real. However, Lozinskaya et al. (2000) point out that the above velocity range is also observed in the Halpha profiles in the broader region of the Cygnus X source.
The question arises as to whether the two objects, which can be characterized as a supernova remnant, on the one hand, and as an outer X-ray shell of a large diameter, on the other, are interrelated. In the range of Galactic longitudes under study, the objects in the Orion-Cygnus arm have radial velocities of Galactic rotation within 5 km/s for any Galactic-rotation models (see Burton and Gordon 1978; Kerr and Linden Bell 1986).
Given the peculiar velocities and systematic noncircular motions, this range extends to +- 15 km/s. The radial velocity -25 km/s is well outside this range and formally corresponds to a distance of 6 kpc from the Sun in the most compact Galactic-rotation model by Gluskova et al. (1998).
Of course, we can assume that these objects spatially coincide and that the radial-velocity difference is explained by their internal kinematics, for example, by radial motions. However, there are no clear traces of interaction between these objects in the HI distribution at radial velocities between +3 and -25 km/s.
Therefore, we believe a chance projection of these objects to be most likely. This comes as no surprise in a region that is so rich in objects of different nature: the SNR G78.2+2.1 itself is projected onto the Cygnus X thermal complex, while the well-known gaseous nebula around the star gamma Cyg is projected onto the SNR shell.


A thorough analysis of the HI distribution in the G78.2+2.1 region over a wide radial-velocity range has revealed an extended expanding elliptical HI shell, 3 dg. in diameter, at radial velocities from -10 to -25 km/s, which closely coincides in coordinates and angular sizes with the outer X-ray shell from Lozinskaya et al. (2000).
The Halpha emission investigated by the above authors in the SNR region also has a secondary peak inside the outer shell at radial velocities from -45 to -20 km/s.
Since the radial velocities of the SNR with the inner X-ray shell and of the outer X-ray shell with the HI shell and Halpha-emitting regions differ markedly, their distances can be assumed to differ as well; i.e., we observe a chance projection of two distinct objects, which is not surprising for the Cygnus region under study.
Of course, if subsequent observations in all ranges will reveal unquestionable evidence for a physical association of these remarkable objects, then this simplest interpretation will have to be rejected.