AN OLD SUPERNOVA REMNANT WITHIN AN H ii COMPLEX AT ℓ ≈ 173°: FVW 172.8+1.5

The region of Auriga within which FVW172.8+1.5 lies is an H ii complex composed of Sharpless H ii regions and OB associations in the Perseus arm (Figure 5). The H ii regions are organized along a large (∼7° × 4°) filamentary structure resembling a bow tie (Kerton et al. 2007). The H ii regions have been studied previously (e.g., Israel & Felli 1978), and they are largely at two different distances (see Table 1). To the northeast of the structure, S231, S232, S233, and S235 are known to be associated with a giant molecular cloud at a distance of 1.8 kpc (Evans & Blair 1981), with CO radial velocities of −18.1 to −23.0 km s⁻¹ (Blitz et al. 1982), in which active star formation is ongoing (see Section 4.3). The distances determined by the spectrophotometry of the exciting stars of individual H ii regions range from 1.0 to 2.3 kpc, but the distance to the associated molecular cloud (1.8 kpc) is often adopted as the distance to the whole H ii complex. To the south of the complex, two 1° sized H ii regions, S229 (IC 405) and S236 (IC 410), and the large (∼5°) diffuse H ii region S230 are present. S236 is associated with a molecular cloud at −7.2 km s⁻¹ and might be at a considerably greater distance because the estimated distance to its central star cluster (NGC 1893) ranges from 3.2 to 6 kpc (Sharma et al. 2007 and references therein). For the other two H ii regions (S229 and S230), not much has been known, although Fich et al. (1990) measured Hα velocities of +4.4 and 0.0 km s⁻¹ toward S229 and S230, respectively. Toward S229, CO J = 1–0 emission at +6.7 km s⁻¹ has been detected, but it may not be related to the H ii region (Blitz et al. 1982). The small H ii regions S234 (IC 417) and S237 in the middle of the "bow tie" have velocities of −13 and −4 km s⁻¹, similar to the first and second groups, respectively. A recent photometric determination of the distance to the central cluster (Stock 8) of S234 also agrees with the distance to the northwestern group of H ii regions, i.e., 2.05 ± 0.10 kpc (Jose et al. 2008). We therefore consider that the Sharpless H ii regions S231–S235 form the active star formation complex G173+1.5 at a distance of 1.8 kpc at v_LSR ≈ − 20 km s⁻¹. The other H ii regions have considerably different velocities and distances, and they are not considered to be associated with this complex.

Zoom In Zoom Out Reset image size

The morphology of the region suggests an association between the large radio continuum filamentary structures and the H ii regions. Filaments A through D (see Figure 5) have similar concave shapes suggesting that they have resulted from a common source within the H ii complex. For the radio Filament A, we have found an H i filamentary structure along the radio feature at v_LSR = −25 to −28 km s⁻¹ (Figure 6). The correlation is not perfect in the sense that the H i feature looks less curved than the continuum filament and seems to extend outside of the field shown in Figure 6. However, the velocity is close to the CO velocity of S232, and the two could be associated. We could not find any H i structures associated with the other continuum filaments. Filament E has an enhanced brightness and its curvature is opposite to those of Filaments A–D. It is likely that the enhanced brightness and the convex shape is due to interaction with a dense ambient medium there, although no responsible molecular cloud is seen in the available CO survey data (Dame et al. 2001; see also Figure 9). Hence, Filaments A–E are likely to be associated with the H ii complex G173+1.5, although we have only circumstantial evidence, except perhaps for Filament A. Kinematic evidence is needed to confirm the association.

Zoom In Zoom Out Reset image size

The nature of the radio emission from the large continuum filamentary structure has been discussed previously, although not in detail. Kerton et al. (2007) noted this structure in the 1420 MHz Canadian Galactic Plane Survey (CGPS) map, and argued that Filaments D and E are thermal emission based on either the presence of corresponding infrared emission and/or the rising spectrum between 408 and 1420 MHz. Gao et al. (2010) claim that Filament A shows a thermal spectrum between the higher frequencies of 1.4 and 5 GHz.

Using radio continuum data from 325 MHz to 2.7 GHz available online (Table 2), we have attempted a study of the continuum spectrum of the filaments associated with the structure. The faint filamentary features are visible at most of these frequencies. Here we will focus on Filament A at (ℓ, b) ∼ (1725, +35), which stands out at 325 MHz in WENSS. To estimate values of spectral index, we (1) converted the WENSS data to brightness temperature, thus bringing all surveys to the same intensity units; (2) convolved all images to have a circular beam of FWHM = 47, the declination beam size of the CGPS 408 MHz image at δ = 37°; (3) gridded the images onto the 2' spaced pixels of the Effelsberg data; and (4) subtracted out all point sources. The final images are shown in Figure 7, where the dotted and solid boxes mark Filament A and the H ii region, S232, which provides a reference with a thermal spectrum. Finally, we removed a smooth, linear, large-scale background emission gradient using two neighboring areas, marked by pairs of boxes flanking the H ii region and the filament in Figure 7.

Zoom In Zoom Out Reset image size

Table 2. Parameters of the Continuum Surveys Used for Spectral Analysis

Name	Frequency	Resolution	rms	References
	(MHz)		(mK)
WENSS	325	54'' × 54'' csc δ	2300	Rengelink et al. (1997)
CGPS	408	28 × 28 csc δ	950	Taylor et al. (2003)
CGPS	1420	49'' × 49'' csc δ	68	Taylor et al. (2003)
Effelsberg 11 cm	2695	43	20	Fürst et al. (1990)

Download table as:  ASCIITypeset image

The continuum spectra of Filament A and S232 are shown in Figure 8 (left) by filled and open circles, respectively. The total flux densities are given in Table 3. The error bars in Figure 8 represent standard deviations in the neighboring areas after the removal of smooth backgrounds. Using all flux density values from 325 to 2695 MHz, the estimated spectral index (S_ν∝ν^−α) of S232 is α = −0.09 ± 0.02, while that of Filament A is α = +0.23 ± 0.06. We also derived spectral indices for Filament A and S232 using T–T plots (Turtle et al. 1962). The regions used for making the T–T plots are those over which the total flux densities were derived. Figure 8 (right) shows a T–T plot between 325 and 2695 MHz of Filament A. The points are well described by a linear fit. The slopes of the plots, and the values of α derived between different pairs of frequencies, are summarized in Table 4. For S232, the derived spectral indices are consistent, ranging from α = −0.04 to +0.09 which, together with the total flux density spectral index of α = −0.09, is reasonably close to the value of α = +0.1 expected for an optically thin thermal source. However, the spectral indices derived for Filament A are not straightforward to interpret. The frequency pairs that include 325 MHz result in α = +0.50 ± 0.11 and +0.32 ± 0.06, consistent with non-thermal emission, while the pairs that include 408 MHz yield α = +0.07 ± 0.10 and −0.04 ± 0.08, compatible with thermal emission. These trends are consistent with the plot of the total flux densities (Figure 8, left). In that plot, if either the 325 MHz or the 408 MHz data point is excluded, then the spectral index of Filament A would, respectively, be flatter or steeper than the α = +0.23 ± 0.06 derived using data at all frequencies.

Zoom In Zoom Out Reset image size

As mentioned at the beginning of this section, previous studies claimed that the radio emission from some of these filaments is of thermal origin. However, our analysis in this section indicates the presence of a non-thermal component in Filament A. The fact that the spectral index of S232, as derived from the same data, is consistent with optically thin thermal emission supports our result. Therefore, it is more likely that the filament has both thermal and non-thermal components (see also Section 4.3). If thermal and non-thermal components coexist in the filament, a steep spectrum at lower frequencies, and a flat spectrum at higher frequencies (as found by Gao et al.), might be expected. Also, if thermal and non-thermal components coexist in the continuum emission of Filament A, this could be the case for Filaments B–E. We have examined the polarization maps of this area. The G173 complex region was covered by the polarization surveys at 1.4 and 5 GHz (Landecker et al. 2010; Gao et al. 2010). While the 1.4 GHz polarization data of Landecker et al. do not show any apparent counterparts to the G173 structure, the 5 GHz polarization data of Gao et al. do show some filamentary features associated with the structure. In the polarized intensity map of Gao et al. (their Figure 17), two depolarized filaments, indicating a thermal nature, are recognizable; one at (ℓ, b) ∼ (174.3, +2.0), which corresponds to Filament B, and the other at ∼(171.0, +2.5), located between Filaments A and D. Other than B, the continuum filaments have no apparent counterparts in this image. Hence, it is difficult to discuss the nature of Filaments A–E using the current polarization data, with the exception that a thermal component appears to dominate in Filament B. The nature of the continuum filaments needs to be investigated with more sensitive full-Stokes continuum data.

The total extent of the putative shell based on the radio continuum filaments and our H i observations is 44 × 34, which converts to 138 pc × 107 pc at 1.8 kpc. If we adopt the velocity centers of the highest-velocity clumps given in Section 4 (+35 km s⁻¹) as an endcap velocity of the shell, its expansion velocity is 55 km s⁻¹. To derive the total H i mass and kinetic energy of the FVW172.8+1.5, we need to estimate the total mass of the expanding shell. This can be done by extrapolating the mass distribution we derive for high velocities. Figure 11 shows the distribution of H i mass in each 2.2 km s⁻¹ velocity interval between v_LSR = +14 and +46 km s⁻¹. We assume the receding and approaching hemispheres of the expanding shell to be symmetric and centered at v_LSR = −20 km s⁻¹. Note that the approaching portion of the shell is not visible because of confusion by the Galactic background/foreground H i emission. While this may not be true, the resulting kinetic energy would be roughly correct should the energy injection be spherically symmetric. A Gaussian shape is often used for the extrapolation when fitting the mass distribution of the observed H i shells (e.g., Giovanelli & Haynes 1979; Koo et al. 1990, 2006). If we adopt this Gaussian extrapolation (the solid line in Figure 11), we obtain a value of 1.3 × 10⁴d²_1.8 kpc M_☉. However, this may overestimate the total mass because for an ideal expanding thin shell of uniform density, the mass per unit radial velocity interval is constant. The mass profile at the highest velocities could be decreasing due to the velocity dispersion, or the velocity structure, within the shell. However, the mass profile at the central velocities would remain constant unless the turbulent velocity in the shell were large enough to be comparable to its expansion velocity.

Zoom In Zoom Out Reset image size

Here, we develop a simple, but physical, shell model that yields a flat mass profile for the central velocities. We consider a spherical shell with a radius R = 2°, a thickness ΔR = 10', an expansion velocity v_exp = 55 km s⁻¹, and a dispersion velocity v_σ = 5.5 km s⁻¹. These parameters are derived from the observed H i shell. We assume that the gas density within the shell is constant. For the radial profile of expansion velocity within the shell, we consider two different cases. The first is that the expansion velocity is constant with radius. The second is that the expansion velocity varies linearly with radius. Linearly decreasing or increasing velocity structures within the shell are derived by theoretical studies for different ages of radiative shells (Slavin & Cox 1992; Blondin et al. 1998). Here, we choose a velocity structure for which the expansion velocity at the outer radius drops to 50% of that at the inner radius.

To fit the derived mass distribution, we adopt a three-dimensional cube composed of 451³ pixels. The z-axis is along the LOS. The number density of H i atoms for a pixel centered at r = (x² + y² + z²)^1/2 is n(r), which is assumed to be constant in our model. This pixel has an LOS velocity v_LOS(x, y, z) = v_exp(r)z/r, where v_exp(r) is the radial expansion velocity at r. The H i 21 cm emission line from this pixel has a Gaussian distribution with the LOS velocity dispersion of v_σ (=5.5 km s⁻¹). Then the column density, N(x, y, v), at (x, y) over the LSR velocity interval from v to v + Δv and the H i mass, M(v), between v and v + Δv, are given as follows:

$$\begin{eqnarray*} N(x, y, v) &=& {n\over \sqrt{2 \pi v_\sigma ^2}}\int\!\!\! \int _v^{v+\Delta v}\nonumber\\ &&\times\, \exp \left[{-{(v^\prime - v_{{\rm LOS}}(x, y, z))^2}\over {2 v_\sigma ^2}}\right] dv^\prime dz, \end{eqnarray*}$$

$$\begin{eqnarray*} &&M(v) = m_{\rm H} \int\!\!\!\int N(x, y, v) dxdy. \end{eqnarray*}$$

The dotted line in Figure 11 is the best fit for the mass distribution of a shell with a constant expansion velocity profile. Here, the only fitting parameter is the density of the shell. The velocity range, where the mass of the observed H i shell smoothly decreases, is wider than the velocity range due to velocity dispersion, so the observed H i mass distribution cannot be explained by velocity dispersion alone. The dashed line shows the best fit for the mass distribution of a shell with an expansion velocity profile linearly decreasing toward larger radii. This dashed line fits the observed mass better than the dotted line. Variation of the expansion velocity with radius dilutes the rectangular mass distribution at high velocities, but still yields a constant mass distribution at lower velocities, from v_LSR ∼ − 50 to +10 km s⁻¹. Note that the basic shape of the mass profile is determined by the difference between the minimum and maximum expansion velocities. The detailed velocity and density structures only affect the shape of the decreasing portion of the mass profile. We will study the mass profiles resulting from various velocity and density structures in a forthcoming paper. The estimated total H i mass of the shell adopting a linearly decreasing velocity profile is 5900 d²_1.8 kpc M_☉. We will use this value as the characteristic mass of the shell. Note that the total mass derived from the Gaussian fit is about twice this. The kinetic energy of the shell, taking into account the He abundance of 10% by number, is 0.25 × 10⁵¹ d²_1.8 kpc erg. The derived physical parameters of the model H i/continuum shell are presented in Table 5.

Table 5. Parameters for An Expanding Shell

Parameter	Estimated Value
Center (α, δ)	05h33m29s, +35°50'52''
(ℓ, b)	17278, +151
Mean radius	69 × 54d1.8 kpc pc
Expansion velocity	55 ± 5 km s−1
Total H i mass	5900 ± 770 d21.8 kpc M☉
Kinetic energy	0.25 ± 0.05 × 1051 d21.8 kpc erg
Initial ISM density	0.25 ± 0.03 d−11.8 kpc cm−3
Age	0.33 ± 0.03 Myr

Notes. The errors are 1σ errors. For the estimation of the initial ISM density and age, we use the geometrical average of the mean radius (61 pc).

Download table as:  ASCIITypeset image

We can think of two possibilities for the origin of FVW172.8+1.5; stellar winds from OB stars or SN explosions. H ii regions can also produce expanding H i shells, but their expansion velocities and energies are much smaller than those of FVW 172.8+1.5. Furthermore, the H ii regions are mostly located at the boundary of the structure, so that they do not appear to be major sources of the 3°–4° sized shell. We first consider whether the stellar winds from OB stars can produce FVW172.8+1.5. There are more than a dozen O stars in this area (Garmany et al. 1982; Maiz-Apellaniz et al. 2004), with eight of them projected against the shell (Figures 5 and 9). We summarize the parameters of these stars in Table 6. Their spectral types range from O9.5 to O7, and four are associated with the Sharpless H ii regions. The other four could be possible candidates for the origin of the H i/continuum shell. If we use the relation of Howarth & Prinja (1989) and a wind speed of ∼2000 km s⁻¹, then the typical wind luminosities of O7–O9 main-sequence stars are (5–0.8)× 10³⁵ erg s⁻¹. For giant O stars, this would be larger by a factor of two. However, the required wind luminosity to explain the parameters of the shell is L_w = 77E_K/(9R/v_exp) = 6 × 10³⁷ erg s⁻¹ (Weaver et al. 1977), which is 2 orders of magnitude greater than the available wind luminosities. Even if there are groups of stars in this area (see the next section), their contribution to the wind luminosity will not be significant because they are no brighter than O9.

We thus consider that FVW172.8+1.5 is likely to be the result of an SN explosion. Certainly, the derived parameters are fully consistent with this interpretation. The required SN explosion energy is E_SN = 6.8 × 10⁴³n^1.16₀R_s^3.16v_exp^1.35χ^0.161 erg, where n₀ is the ambient H density in cm⁻³, R_s is the radius of the shell in pc, v_exp is the expansion velocity in km s⁻¹, and χ is the metalicity in units of the solar metalicity (Cioffi et al. 1988). For the above parameters, E_SN = 1.3 × 10⁵¹ erg, where the solar metalicity has been adopted. This value of E_SN is close to the canonical outburst energy of a single SN explosion of 1 × 10⁵¹ erg. The age of the shell, assuming this to be in the radiative stage of SNR evolution, is t = 0.3 R_s/v_exp = 0.33 Myr. The SNR interpretation is also consistent with the presence of X-rays in this region. The thermal energy of the X-ray-emitting gas is 3 × 10⁵⁰ erg, which is 25% of the initial explosion energy and close to that expected for a 0.33 Myr old radiative SNR (e.g., Cioffi et al. 1988). The apparent segregation between the fast-moving H i gas and the X-ray-emitting gas could be due to a non-uniform and/or inhomogeneous density distribution in the ambient medium. It is also possible that more than one SN exploded within the past few 10⁵ yr. We have checked for known pulsars in this area, but none are known within the boundary of the putative shell. The nearest pulsar, PSR J0540+32 at (ℓ, b) = (17680, +072), is ∼4° distant from the center of the H i/continuum feature, and lies at a distance of ∼6.2 kpc. PSR J0540+32 is thus unlikely to be associated with the progenitor of FVW172.8+1.5.