Filename
stringlengths 22
64
| Paragraph
stringlengths 8
5.57k
|
|---|---|
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | plex interwoven streams, elongated arcs, bubbles and narrow filaments in Hα. The newly identified candidate is shown in Fig. 1 where Hα is coloured red and the short red continuum image appears white due to both blue and green colours be- ing assigned to the same image. Although these structures can be closely correlated with radio emission, they usually become quite fragmented in old remnants and may extend beyond, within or along the boundaries of the radio remnant (see Stupar & Parker 2009). In addition to the Hα/SR com- bined images, we used the Magellanic Cloud Emission-Line Survey (MCELS; Smith et al. 1998) to directly compare the strength of the three narrow bands of [O iii] 5007 (blue), Hα (red) and [S ii] 6716+6731 (green). When the colour maps are evenly combined the SNR candidate becomes evident by the strong green ([S ii] line) colour. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | RP1577 has a central position of (J2000) RA 05h 12m 28s DEC -67◦ 07(cid:48) 20(cid:48)(cid:48) determined using SuperCOSMOS as- trometry (Hambly et al 2001) and a measured angular extent from the Hα imagery of 56 × 64 arcsec. This equates to a physical extent of 13.5 × 15.5 pc at the LMC distance. With physical diameters ranging from 0 to 120 pc (Badenes et al. 2010), the cumulative distribution of SNR diameters in the LMC have been found to be roughly linear (eg. Chu & Ken- nicutt 1988; Bandiera & Petruk 2010) up to a cutoff at a physical radius of rcut ∼30 pc (Badenes et al. 2010). This means that the 15.5 pc size of RP1577 places it among the smaller 15 per cent of LMC SNRs, joining six other SNRs in the 10-20 pc range. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure 2. The (Hα/SR) quotient image of RP1577 showing cir- cular loops to the north of parallel, stratified filaments which outline the exterior of the expanding shell. The position of the 2dF spectroscopic observation is shown by the yellow circle close to the center of the object. Horizontal lines mark the positions of long-slit observations. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure 3. The 2dF spectroscopic observation of RP1577 indi- cated by the yellow circle in Figure 2. The Hα/[S ii] ratio is 0.46±0.02. The extinction of c(Hβ) = 0.45 is based on the Balmer decrement. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Table 2. Observed line intensities and flux ratios for four different spectroscopically observed positions across three areas of RP1577. Line measurements were made using the SPLOT script in the IRAF package. Errors for individual flux line measurements are a constant ±1.9×10−17 for 2dF 300B, ±1.2×10−17 for 2dF 1200R, ±1.5×10−17 for AAOmega, ±2×10−17 for 1.9-m measurements. Column details are provided in the text. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure 4. The 2dF 1200R medium resolution spectroscopic ob- servation of RP1577 indicated by the yellow circle in Figure 2. The higher resolution of this spectrum shows the weak [O i]6300 & 6363 lines. The [S ii]/Hα ratio is 0.66±0.02. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure 5. The long-slit, spectroscopic observation of RP1577 at position (J2000) RA 05h 12m 31s and DEC –67◦ 06(cid:48) 59(cid:48)(cid:48). The [S ii]/Hα observed ratio at this position is 0.78±0.01. The extinc- tion of c(Hβ) = 0.28 is based on the Balmer decrement. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | In Table 1 we summarise details regarding the spectro- scopic follow-up observations. A five night observing run on the AAT using 2dF (Lewis et al. 2002) was undertaken in December 2004 to spectroscopically confirm LMC emission candidates. The large corrector lens incorporates an atmo- spheric dispersion compensator, which is essential for wide wavelength coverage using small diameter fibres. Specifically targeted long-slit spectra were obtained using the 1.9-m tele- scope at the SAAO in March, 2014. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure 6. The long-slit, spectroscopic observation of RP1577 at position (J2000) RA 05h 12m 26s and DEC –67◦ 07(cid:48) 05(cid:48)(cid:48). The [S ii]/Hα observed ratio of 0.93±0.01 at this position is the high- est ratio of all our spectroscopic observations of RP1577. The ex- tinction of c(Hβ) = 0.33 at this position is based on the Balmer decrement. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | umn (9) provides the [S ii]/Hα ratio and line measurement errors for estimating the excitation of the shock-front. Col- umn (10) shows the ratio of [S ii] 6716/6731 used for esti- mating the electron density. The electron density shown in column (11) is derived from the [S ii] 6716/6731 ratio derived using the TEMDEN script found in the STSDAS package from IRAF. Once the [S ii] ratio reaches 1.4, representing a den- sity of 26 cm3 the low density limit using this method has been reached. Column (13) gives the electron temperature, derived from the [O iii] 5007+4959/[O ii] 4363 emission lines. The faint [O ii] 4363 line was only detectable for the purpose of measurement at one position. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | where [f (Hα)-f (Hβ)] = –0.34 from the standard interstellar extinction curve. The resulting extinction values c(Hβ) are given in the caption below each spectrum in Figs. 3, 5 and 6. The [S ii] 6717/6731 ratio of 1.35 ±0.07 from the 2dF 300B exposures results in an electron density value near 70 ±10 cm3. In Fig. 2 we show the (Hα/SR) quotient image which effectively removes stars from the area. The yellow circle in the image marks the position observed with 2dF. This is approximately 10 arcsec south of the optical centre as indicated by the elliptical shape of the diffuse Hα emission. Spectroscopic results from this position are shown in Fig. 3 for the 300B exposure and Fig. 4 for the 1200R exposure. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | near the continuum therefore their low level in the reduced spectra (Figs. 5 and 6) may also in part be due to sky sub- traction where the combined nebulae and telluric [O i] 6300 lines span the width of the slit. Morphologically, although the overall shape is somewhat elliptical, the interior detail of the SNR shows a series of elongated filaments, which be- come more dense and finally curved in the form of loops to the north. With such strong spectroscopic confirmation, we now assign RP1577 the name MCSNR J0512−6707 which includes the basic coordinates as recommended by the IAU. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Table 3. Heliocentric radial velocities and systematic errors for key emission lines at the observed positions as shown in Fig. 2. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | density of ∼1.2 cm−3. For the lower error estimate we adopt a value of 0.2 cm−3 since a value of 0.3 cm−3 was found by plotting the SNR on the correlation provided by Berezhko & V¨olk (2004). |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | The upper error for our adopted energy is estimated from the radio observations for this remnant (see section 3) and based on a surface brightness of 1.57 × 10−20 W m2 Hz−1 sr−1 at a frequency of 1 GHz, placing the object at a position on the Σ - D diagram by Berezhko & V¨olk (2004) equivalent to an energy (E0) of ∼ 0.25 × 1051 erg. We adopt this estimated value as it represents an upper estimate for such an SNR in the LMC. For our lower error we adopt a value of 9.5×1051 erg, which is just below the canonical 1051 erg value for SNRs. The radius of 7.75 pc is found from the maximum diameter given above. This method pro- vides an estimated dynamical age of the remnant between ∼2,000 yrs and ∼5,000 yrs. The method assumes that the swept-up mass will become proportionally larger than the ejected mass and the velocity will decelerate accordingly. Using the same equations, the expansion velocity may lie between ∼ 600 and 1400 km s−1. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | The heliocentric radial velocities at each observed position of the SNR are shown in Table 3. Emission lines were measured using the IRAF splot and EMSAO tasks against standard rest velocities. Column 1 gives the name of the observa- tion, which corresponds to the positions marked in Fig. 2. Columns 2 to 5 then present the heliocentric radial velocities in km s−1 for the [O iii] 5007, Hα, [N ii] 6583 and [S ii] 6716 emission lines. The estimated velocity error, given in column 6 is a standard systematic error for all velocities measured on a particular observation. It was determined according to the size of the pixels, related to the resolution of the in- dividual spectrum. The rest velocities for each element are provided in the column headers. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | imagery in section 2, above, the [O iii] emission is less affected by temperatures greater than 104K and so, can be seen to move at a slower velocity than Hα which leads the shock front. This effect can be measured, even though expansion at positions 1 & 2 are directionally near to the plane of the sky. Our velocity mea- surements show that this is true of the [O iii] line at both positions observed by the SAAO 1.9-m telescope. Velocities on the forbidden lines of [N ii] and [S ii] are in every case higher than those found for Hα. This is to be expected for [S ii] since the bulk abundance of this element is produced in the SN explosion. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure 7. Radio contours from PMN at 4850 MHz super- imposed on the 843 MHz SUMSS greyscale map of SNR MCSNR J0512−6707. Contours are spaced at 0.01 mJy beam−1. The flux density varies from –0.66 mJy to 3.16 mJy across the area shown. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Radio detection has traditionally been the main method for identification and confirmation of SNRs because of their non-thermal spectra, radio shell structures and polarization (e.g., Bozzetto et al. 2014b). Radio observations also allevi- ate problems associated with optical detection where dust often obscures optical detail. With the development of new instrumentation, the sensitivity of existing radio telescopes is constantly being enhanced, allowing us to improve our knowledge of SNRs within the LMC. For this project we have used a combination of dedicated radio observations complimented by freely available radio survey data. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | The radio counterpart to this optically detected SNR was originally found using data from the Parkes-MIT-NRAO (PMN; beam size = 300(cid:48)(cid:48)) radio survey at 4 850 MHz (Con- don, Griffith, Wright, 1993) and the Sydney University Mo- longlo Sky Survey (SUMSS; beam size = 45(cid:48)(cid:48)) at 843 MHz (Bock, Large, Sadler, 1999). Although a radio source is de- tected at the position of MCSNR J0512−6707in both of it is more extensive in the PMN due to these surveys, the lower resolution of the survey (see Fig. 7). The con- tours therefore do not represent the true size. A detection of MCSNR J0512−6707 was made in Filipovi´c et al. (1995; listed as LMC B0512-6710) where a radio continuum anal- ysis of the LMC was performed at various frequencies using the Parkes radio telescope. The SUMMS 843 MHz measure- ment was omitted from spectral index calculations as it was superseded by the 843 MHz Molonglo Sythesis Telescope (MOST) measurement, due to the higher sensitivity of the MOST image and its greater UV–coverage. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure 8. MCSNR J0512−6707 with contours from the Dickel et al. (2010) 8640 MHz mosaic image (beam size = 22 × 22 arcsec). The local rms is σ = 0.5 mJy beam−1 and contours are 3, 4, 5, 6, 7 and 8 σ. The radio contours are superimposed on the continuum- subtracted MCELS map of the immediate area where Hα is red, [O iii] is blue and [S ii] is green. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | structure and so the latest radio-continuum data and our own pointed observations were used to provide details and additional frequencies to demonstrate the non-thermal na- ture of MCSNR J0512−6707. Radio-continuum data used in this project includes a 36 cm (843 MHz) MOST mo- saic image (as described in Mills et al. 1984), a 20 cm (1 377 MHz) mosaic image from Hughes et al. (2007), and 6 cm (4 800 MHz) and 3 cm (8 640 MHz) mosaic images from Dickel et al. (2005, 2010) which are shown as contours against the continuum-subtracted MCELS three-colour map (see section 2) in Fig. 8. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | These mosaic data were complemented by pointed observations of MCSNR J0512−6707 from our Australian Telescope Compact Array (ATCA) observations taken on 2011 November 15 and 16, using the Compact Array Broad- band Backend (CABB) receiver which has a bandwidth of 2 GHz. The compact array configuration EW367 was chosen in an effort to obtain as much emission from the shorter spac- ings at 6 and 3 cm (ν = 5 500 and 9 000 MHz, respectively) as possible. The resulting 6 cm contours are overlaid on the MCELS Hα map of MCSNR J0512−6707 in Fig. 9. Total integration time over the two days for this source amounted to ∼64 min. The miriad2 (Sault, Teuben & Wright 1995) and karma3 (Gooch 1995) software packages were used for reduction and analysis. Images were formed using miriad multi-frequency synthesis (Sault & Wieringa 1994) and nat- ural weighting. They were deconvolved with primary beam correction applied. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure 9. Contours using the entire 2 GHz bandwidth centered at 5 500 MHz are overlaid on the Hα image of MCSNR J0512−6707. The contours are spaced at ∼ 0.01 mJy beam−1 commencing at 0.3 mJy beam−1. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | integrated flux density measure- MCSNR J0512−6707, ments were taken from the four mosaics at frequencies of 843, 1 377, 4 800, and 8 640 MHz, in addition to our observations at 5 500 and 9 000 MHz. However, as in- terferometers such as the ATCA are an arrangement of individual antennae, they lack the short spacings that are responsible for the large-scale emission (e.g., of objects such as SNRs), resulting in a loss in flux density which becomes more pronounced at higher frequencies. For this reason, the 9 000 MHz integrated flux density measurements were discarded. The 5 500 MHz observations were split into four 512 MHz bands (4 732, 5 244, 5 756, and 6 268 MHz), and integrated flux density measurements were taken per band with fitted gaussians using the miriad task IMFIT. The results of these measurements and those taken from the mosaics are listed in Table 4 and shown in Fig. 10. Errors in these measurements predominately arose from uncertainties in defining the ‘edge’ of the remnant. However, we estimate errors to be less than 10 per cent. The rms noise estimates were taken using the miriad task sigest. The resulting spectral index between 843 and 8 640 MHz of α = − − 0.52 ± 0.04 is consistent with the typical SNR spectral index value of − − 0.5. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Figure of MCSNR J0512−6707 between 843 and 8 640 MHz. The blue markers represent flux density error margins of 10 per cent. |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | consequently, the filling factor was unable to be estimated. Therefore, Table 5 shows the average equipartition field over the whole shell of MCSNR J0512−6707 in addition to the estimated minimum energies based on various filling factors ranging from 1 (a completely filled SNR) to 0.25 (a mostly empty shell-like remnant). The range of values for the mag- netic field, from 124 − 184 µG, are relatively high for an SNR, and may be indicative of a younger remnant. This can be compared with other LMC SNRs using the same method, e.g., those listed in Table 6. Such values may sug- gest a remnant of a few thousand years, which is equal to or less than the age estimate of ∼2,280 to ∼4,690 yrs derived in section 2.1 but still in-keeping with the remnant’s rela- tively small size which ranks it as the 6th smallest currently known in the LMC (Badenes et al. 2010; Bozzetto et al. in prep). |
Processed_Optical_discovery_and_multiwavelength_investigatio.txt | Table 6. Magnetic field strength for a sample of LMC SNRs with their associated ages. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | Abstract. About 25% of the optical extragalactic sky is obscured by the dust and stars of our Milky Way. Dynamically important structures might still lie hidden in this zone. Various approaches are presently being employed to uncover the galaxy distribution in this Zone of Avoidance (ZOA). Results as well as the different limitations and selection effects from these multi-wavelengths explorations are being discussed. Galaxies within the innermost part of the Milky Way — typically at a foreground obscuration in the blue of AB > — remain particularly difficult to uncover except for H i-surveys: the Galaxy is fully transparent at the 21cm line and H i-rich galaxies are easy to trace. We will report here on the first results from the systematic blind −1) in the southern Zone of Avoidance which is currently H i-search (v ≤ 12700 km s being conducted with the Parkes Multibeam (MB) Receiver. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | Is the dipole in the Cosmic Microwave Background Radiation (direction and amplitude) entirely explained by the gravity on the LG from the irregu- lar mass/galaxy distribution? As the nearest galaxies (v < 300 km s−1) gener- ate 20% of the total dipole moment (Kraan-Korteweg 1989) nearby individual galaxies are equally important as massive groups, clusters and voids. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | Various approaches are presently being employed to uncover the galaxy distribu- tion in the ZOA: deep optical searches, far-infrared (FIR), near-infrared (NIR) surveys and blind H I searches. All methods produce new results, but all suffer from (different) limitations and selection effects. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | OPTICAL: Nearly the whole southern ZOA has been systematically surveyed for highly obscured but still visible galaxies using existing sky surveys (cf. Woudt 1998, for a detailed overview). These surveys achieve a considerable reduction of the ZOA and have uncovered distinct LSS unrelated with the foreground extinction. Follow-up redshift observations have revealed a number of dynami- cally important structures such as e.g., the nearby overdensity in Puppis (La- hav et al. 1993) and the massive cluster A3627 at the core of the GA (Kraan- Korteweg et al. 1996, cf. Fig. 1). Deep optical surveys are not biased with respect to any particular morphological type. However, for foreground extinctions above AB >∼ 5m (H I-column-densities NHI >∼ 6 · 1021cm−2), the ZOA remains fully opaque (cf. inner contour in Fig. 1). For the southern Milky Way this corre- sponds roughly to |b| <∼ ±5◦. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | FIR: The IRAS Point-Source Catalog (PSC) has been exploited in the last decade to identify galaxy candidates behind the ZOA. Using different colour se- lection criteria, galaxy candidates were followed up by HI radio surveys (e.g., Lu et al. 1990) or by inspection of plates (e.g., Takata et al. 1996). To avoid confu- sion with Galactic sources, K-band snapshots have proved very efficient (Saun- ders et al. 1994). Confirmed IRAS galaxies can be merged with IRAS galaxy samples outside the ZOA to produce uniform whole-sky samples for LSS stud- ies. But bright spiral and starburst galaxies dominate these samples. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | NIR: The recent near infrared (NIR) surveys, 2MASS (Skrutskie et al. 1997) and DENIS (Epchtein 1997), provide complementary data. NIR surveys are sensitive to early-type galaxies, are tracers of massive groups and clusters missed in IRAS and H I surveys, have little confusion with Galactic objects and are less affected by absorption than optical surveys. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | NIR data, and therewith the possibility to extend the peculiar velocity field into the ZOA via the NIR Tully – Fisher relation. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | H I: In the regions of highest obscuration and infrared confusion the Galaxy is fully transparent to the 21-cm line radiation of neutral hydrogen. H I-rich galaxies can readily be found at lowest latitudes through detection of their redshifted 21-cm emission. Only low-velocity extragalactic sources (blue- and redshifted) within the strong Galactic HI emission will be missed, and – because of baseline ripple – galaxies close to radio continuum sources. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | Until recently, radio receivers were not sensitive and efficient enough to at- tempt large systematic surveys of the ZOA. In a pilot survey with the late 300-ft telescope of Green Bank, Kerr & Henning (1987) surveyed 1.5% of the ZOA and detected 16 new spiral galaxies. Since then a systematic shallow search for nearby, massive galaxies has been completed in the north (Henning et al. 1998), yielding five objects including Dwingeloo 1 (Kraan-Korteweg et al. 1994). |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | In March 1997, a systematic blind H I survey began with the Multibeam Receiver (13 beams in the focal plane array) at the 64 m Parkes telescope in the most opaque region of the southern Milky Way (213◦ ≤ ℓ ≤ 33◦; |b| ≤ 5◦). The ZOA will be surveyed along constant Galactic latitudes in 23 contiguous fields of length ∆ℓ = 8◦. The ultimate goal is 25 scans per field where adjacent strips will be offset in latitude by ∆b = 1.′5 for homogeneous sampling. With a total observing time of 1500h, we will obtain an effective integration time of 25 min/beam with a 3 σ detection limit of 15 mJy. Roughly 3000 detections are predicted for the covered velocity range of −1200 <∼ v <∼ 12700 km s−1 (Staveley-Smith 1997). This allows the detection of dwarfs with H I-masses as low as 106 M⊙ in the local neighbourhood, and will be sensitive to normal Sc galaxies well beyond the Great Attractor region. As a byproduct, the survey will produce a high resolution integrated column density map of the southern Milky Way and a detailed catalog of high velocity clouds (cf. Putnam et al. 1998). |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | At the time of this meeting, the whole southern ZOA survey had been surveyed twice (∆b = 17′, rms ∼ 20 mJy). The cubes of the Hanning smoothed data (26 km s−1 resolution) were inspected visually for 21 of the fields (220◦ ≤ ℓ ≤ 4◦) and all galaxy candidates with H I fluxes >∼ 100 mJy were catalogued. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | Fig. 2. Redshift slices from data in Fig. 1: 500<v<3500 (top), 3500<v<6500 (middle), 6500<v<9500 km s−1 (bottom). The open circles mark the nearest ∆v=1000 km s−1 slice in a panel, then triangles, then the filled dots the 2 more distant ones. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | on the southern Milky Way. Note the near full lack of galaxy data for extinction levels AB ∼ 1.m5 (outer contour). The middle panel results from the follow-up observations of the optical galaxy search by Kraan-Korteweg and collaborators. Various new overdensities become apparent at low latitudes. But the innermost part of our Galaxy remains obscured (AB >∼ 5m, |b| <∼ 5◦). Here, the blind H I data (cf. lower panel) finally can provide the missing link for LSS studies. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | continuity of the thin filamentary sine-wave-like structure that dominates the whole southern sky, and the prominence of the Local Void. This feature is very different from the thick, foamy Great Wall-like structure, the GA, in the middle panel. With the full sensitivity aimed at with the MB-survey, we will be able to fill in the LSS in the more distant panels of Fig. 2 as well. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | ATCA follow-up observations of three very extended (20′ to >∼ 1◦), nearby (v < 1500 km s−1) sources revealed them to be interesting galaxies/complexes, with unprecedented low H I column densities (cf. Staveley-Smith et al. 1998). |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | The combination of the complementary multiwavelength surveys allow a new probing of LSS in the ’former’ ZOA. The H I surveys are particularly powerful at the lowest latitudes. But future merging of ZOA data with catalogs outside the ZOA will have to be done with care to obtain ’unbiased’ whole-sky surveys. From the sensitivity attained with the first 2 scans of the ZOA MB-survey it can be maintained that no Andromeda or other H I-rich Circinus galaxy is lurking undetected behind the extinction layer of the southern Milky Way. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | Acknowledgements — The help of the HIPASS ZOA team members R.D. Ekers, A.J. Green, R.F. Haynes, P.A. Henning, R.M. Price, E. Sadler, and L. Staveley-Smith is gratefully acknowledged. |
Processed_Galaxies_behind_the_deepest_extinction_layer_of_th.txt | Takata, T., Yamada, T., Saito, M. (1996): ApJ 457, 693 Woudt, P.A. (1998): Ph.D. thesis, Univ. of Cape Town. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | In this paper we study 21-cm absorption spectra and the corresponding emis- sion spectra toward bright continuum sources in the test region (326◦< l < 333◦) of the Southern Galactic Plane Survey. This survey combines the high resolution of the Australia Telescope Compact Array with the full brightness temperature information of the Parkes single dish telescope. In particular, we focus on the abundance and temperature of the cool atomic clouds in the inner galaxy. The resulting mean opacity of the HI, < κ >, is measured as a function of Galactic radius; it increases going in from the solar circle, to a peak in the molecular ring of about four times its local value. This suggests that the cool phase is more abundant there, and colder, than it is locally. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The distribution of cool phase temperatures is derived in three different ways. The naive, “spin temperature” technique overestimates the cloud temperatures, as expected. Using two alternative approaches we get good agreement on a histogram of the cloud temperatures, Tcool, corrected for blending with warm phase gas. The median temperature is about 65 K, but there is a long tail reaching down to temperatures below 20 K. Clouds with temperatures below 40 K are common, though not as common as warmer clouds (40 to 100 K). |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Using these results we discuss two related quantities, the peak brightness temperature seen in emission surveys, and the incidence of clouds seen in HI self-absorption. Both phenomena match what would be expected based on our measurements of < κ > and Tcool. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Absorption by the 21 cm line of HI was first detected 48 years ago (Hagen, Lilley, and McClain, 1955), just four years after the line was detected in emission from interstellar hydrogen in the Galactic plane. As spectral line interferometers were developed in the 1960’s (Clark, Radhakrishnan, and Wilson 1963, Radhakrishnan et al. 1972a) it became possible to measure absorption spectra toward many more continuum sources at high and low latitudes. The striking result from these pioneering studies was that the HI absorption spectra look very different from emission spectra in the same directions (Clark 1965, Hughes, Thompson, and Colvin, 1971, Radhakrishnan et al. 1972b). This can be understood as a result of the huge range of temperatures in the interstellar hydrogen (Field 1958, Field, Goldsmith, and Habing 1969). Interstellar atomic hydrogen can exist in stable thermodynamic equilibrium either at 150 K; these two warm temperatures, phases are reflected in two regimes of excitation temperature for the 21-cm line. Generally this “spin temperature” is close to the kinetic temperature of the gas, but see Liszt (2001) for a detailed discussion of the exceptions to this rule. There is good evidence that interstellar hydrogen is present at temperatures intermediate between those of the classic warm neutral medium (WNM) and cool neutral medium (CNM) (Heiles 2001), but generally the different morphology of the emission and absorption spectra can be explained by a mixture of gas along the line of sight at widely different excitation temperatures (see Kulkarni and Heiles 1988 for a deeper discussion of the phases of the interstellar medium). |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | of whether the differences between emission and absorption could be due simply to small scale structure in the interstellar medium (Faison and Goss 2001, Deshpande 2000). To get around this problem requires an emission-absorption study using the smallest possible beam size for the emission and using background sources which are relatively large, if possible the same size as the beam. The new mosaic surveys of the Galactic plane, combining single dish and interferometer data, allow this for the first time. The Southern Galactic Plane Survey (SGPS, McClure-Griffiths et al. 2000) and the Canadian Galactic Plane Survey (CGPS, Taylor et al. 2002, Strasser et al. 2002) provide good quality emission-absorption spec- trum pairs using extended continuum sources at low latitudes, with a beam small enough to mitigate the effects of variations in the emission and absorption over small angles. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | In this paper we concentrate on a small, test region of the SGPS (McClure-Griffiths et al. 2001) which has been mapped at relatively high resolution (FWHM 90′′). Over the next year maps of the rest of the 210 square degrees of the SGPS with similar resolution will become available, which will provide some 30 times the number of background sources as those considered here. The purpose of this paper is to test methods of analysing the emission- absorption spectrum pairs, and their interpretation. The next section (2) discusses the best way to obtain the spectra from the data. We then consider what the absorption spectra alone tell us about the opacity of the ISM in the inner galaxy (section 3). Then comes the tricky question of how best to combine the information from the emission and absorption spectra to estimate the spin temperatures of the cool clouds (section 4). We consider several fitting techniques that parallel approaches used in past studies at intermediate latitudes. Finally we discuss the implications of the spin temperature distribution for observable quantities like the peak brightness temperature of the HI emission and HI self-absorption (section 5). |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The HI 21-cm line is one of the only transitions in the entire electromagnetic spectrum for which emission and absorption are both relatively easy to detect from the same region. The two fundamental spectra are the brightness temperature of the emission, TB(v), and the optical depth, τ (v). Where there is no background continuum we observe TB directly, and toward an extremely strong background source we observe τ with negligable contribution from the emission, but toward most continuum sources we see a mixture of emission and absorption. This section deals with the optimum method for obtaining TB and τ from the survey data. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | the absorption spectra from the data, since the fainter continuum sources give absorption spectra of marginal quality. For this experiment, as for most studies of Galactic 21-cm emission and absorption, the limiting factor which sets the noise level in the absorption spectrum is not the radiometer noise but the precision with which we can subtract the emission in the directions of the background continuum sources, leaving the absorption only. This is a severe problem at low latitudes; the difficulty of doing accurate emission subtraction has made single dish low latitude emission-absorption studies very difficult, even for Arecibo (Kuchar and Bania, 1991). |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Interferometer studies offer such high resolution that the emission can always be elim- inated, at least toward the most compact background sources, but lacking the zero-spacing (total power) information, interferometer data alone cannot be used to find the emission spectrum. So for emission-absorption studies we need both single dish and interferometer data. The standard method for many years has been to measure the absorption spectrum with an interferometer and compare with an emission spectrum from a single dish telescope (Radhakrishnan et al. 1972c, Dickey et al. 1983, Mebold et al. 1982). In the SGPS we have carefully combined single dish (Parkes) and interferometer (ATCA1) data to achieve com- plete uv coverage for all angular scales larger than our synthesized beam size (90′′). Thus for absorption studies we have the best of both worlds, with sufficient resolution to remove the emission in the direction of the continuum sources to get accurate absorption spectra at low latitudes, while still preserving the total power information which gives us the corresponding emission spectra. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 1.— The rms of the emission after high-pass spatial filtering. The abscissa is the high-pass filter cutoff radius on the uv plane, and the ordinate is the rms of the remaining emission. The 21-cm line emission is shown by the circles, the continuum is indicated by the stars. The continuum is made up of a combination of diffuse emission and discrete sources, while the line emission is mostly diffuse. Thus filtering has a stronger effect on the line channels than on the continuum. The pixel size in this case is 1.4 (deg)−1; the corresponding spatial wavelengths are shown on the top scale. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | three by the spatial filtering step described above. The accuracy of both the interpolated emission spectrum and the absorption spectrum is limited by this filtering and/or interpo- lation step, as neither can be observed directly without some pollution by the other. Thus the matched-filter based on figure 1 is of central importance. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The brightest continuum sources in the SGPS test region are listed on table 1, which gives the source name, rms noise in optical depth (στ ), integrated flux density in Jy, peak brightness temperature in K, and recombination line velocity, in columns 1 - 5. Columns 6 - 10 give the values of various spectral integrals defined below. All of these continuum sources are HII regions except for G328.42+0.22 which is a supernova remnant. The HII regions all have recombination lines detected in the survey of Caswell and Haynes (1987); the measured recombination line center velocity is an indication of the kinematic distance. In a few cases the HI absorption continues to velocities beyond (i.e. more negative than) the recombination line velocity. This may be an indication that the HII region is at the far distance corresponding to its velocity (G328.82-0.08 is an example). In some cases however the HI absorption spectrum shows a single deep component extending 5 to 10 km s−1 beyond the recombination line velocity, and then no absorption at higher velocities (G326.45+0.90 and G326.65+0.59 are examples). These suggest local departures from the smooth circular rotation described by the rotation curve, perhaps due to the effect of the Sgr-Car arm at about -50 km s−1. The recombination line velocities themselves may be offset from the HII region velocity due to asymetric expansion. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 2.— Emission absorption spectrum pairs toward G331.35+1.07 (top) and G326.45+0.90 (bottom). For both, the emission spectrum is at the top and the absorption spectrum is below. Both spectra have error envelopes, which are computed from the errors of the interpolation of the emission in the direction of the continuum source. The numbered boxes in between show the velocity ranges of the different blended lines. The rms noise in the absorption, coming primarily from the emission fluctuations, is indicated as στ , the peak continuum brightness temperature is indicated as Tc,max (in K), and the equivalent width, EW is indicated (in km s−1). Two values of the emission integral over negative velocities are shown, converted to units of 1020 cm−2, as discussed in the text. Also shown are Gaussian fits to the absorption (equation 9) and their residuals (data minus fit) for spectra toward continuum sources with Tc,max > 30 K. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 3.— Emission-absorption spectrum pairs for G326.65+0.59 and G327.99-0.09, presented as in figure 2. The absorption spectrum toward G326.65+0.59 was observed by Goss et al. (1972) in their pioneering study of Galactic 21-cm absorption. Our spectra are consistent with theirs within the uncertainties caused by emission fluctuations in their data. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 4.— Emission-absorption spectrum pairs for G328.31+0.45 and G328.42+0.22, pre- sented as in figure 2. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | where Ω is the angular velocity of rotation around the Galactic center, Rgal is the Galacto- centric radius, and Ω⊙ and R⊙ are the corresponding values for the LSR at the solar circle (220 km s−1 and 8.5 kpc). |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | where n is the HI density and Tsp the excitation temperature of the λ21-cm line, and the , is taken over ∆s which corresponds to the line of sight average denoted by the brackets, h i range of integration in velocity, ∆V , used to compute EW . CH is the familiar conversion K km s−1 , giving κ units of km s−1 constant 1.823 kpc−1. In these units κ typically has values of a few to a few tens in the Milky Way disk. Note that < κ > is the frequency integral of the more familiar κν, (e.g. Spitzer 1977 eq. 3-2 and 3-14), which has units (length)−1, although the range of integration may be only one spectral line channel. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 8.— Emission-absorption spectrum pair for G331.52-0.07, presented as in figure 2. As for G326.65+0.59 on figure 3, this absorption spectrum was first observed by Goss et al. (1972), with similar results to within their errors. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The curves on figure 9 indicate observational selection based on our noise level in τ for the case of G328.42+0.22, one of our brighter sources. The lower curve shows the opacity which would result from optical depth equal to one sigma, as defined by the emission fluctuations discussed above. Weaker absorption than this is not detectable, so multiplying by the velocity gradient gives the corresponding lower limit for detectable opacity. Spectra toward fainter sources that give higher στ will have higher lower limits. Points near or below this curve on figure 9 are upper limits on κ. On the high side, when the absorption lines are very deep, the ln (2στ ). noise prevents us from distinguishing between optical depths greater than τmax = The line saturates at this point (typically τ 3), and we set the optical depth to this value ∼ if e−τ < 2στ . Multiplied by the velocity gradient, this maximum detectable τ gives an upper limit on the measurable κ, shown for the case of G328.42+0.22 by the upper curve. Again, for spectra with higher noise level the upper limit is lower, as evidenced by chains of points at various levels corresponding to saturated lines. Points near the upper curve on figure 9 are thus lower limits on κ. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 9.— Measurements of the 21-cm line opacity as a function of Galactic radius. The circles represent individual spectral channels, translated into optical depth per unit path length using the velocity gradient, as described in the text. The histogram shows the averages of many channels on many lines of sight for radial bins, weighting by line of sight length. The upper and lower curves indicate the observational limits due to saturation (top) and noise (bottom). |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 10.— A face-on view of the Galaxy, with the longitude range of the SGPS test region indicated by two lines at longitudes 326◦and 333◦. The right panel has the radial velocities predicted from the rotation curve drawn as contours; the left panel shows some annular bins which are crossed by the lines of sight at the edges of the region. The annuli used for the averages shown on figure 9 are much narrower than those shown here, and there are many more of them. Both figures have the spiral arm model of Taylor and Cordes (1993) superposed. Common names for the spiral features are Norma (labelled “a”), Scutum-Crux (“b”), Sagitarius-Carina (“c”), and Perseus (“d”). |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | features from the model of Taylor and Cordes (1993) are reproduced. We see that the Norma spiral arm dominates the radial range 0.5 to 0.6 R⊙ near the tangent point, the Scutum-Crux arm at the near distance dominates radii 0.6 to 0.7 R⊙, and Sagitarius-Carina dominates at about 0.85 R⊙. There is a gap in the range 0.7 to 0.85 which is between arms. This appears to be reflected in a dip in κ at about 0.8 R⊙ on figure 9. Even in this dip, and at all other radii in the inner Galaxy, the opacity is larger than its solar circle value of about 5 km s−1 kpc−1. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | This narrow longitude range may not be representative of the entire inner galaxy, so we must wait for the full SGPS survey area to be studied before deciding conclusively that the opacity increases as we go to smaller Galactic radii. In the first quadrant, Garwood and Dickey (1989) found about the same result for the solar circle opacity, but smaller values for the inner galaxy. However, recently Kolpak et al. (2002) have made more extensive observations with the VLA in the first quadrant, and they find a trend similar to what we see on figure 9. The difference may be due to the small numbers of lines of sight sampled in the Garwood and Dickey study, which may have preferentially sampled the regions between the spiral arms, which appear here to show opacities of 5 km s−1 kpc−1 or less, in contrast to the arm regions which can have < κ >> 20 km s−1 kpc−1. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | If the increase in < κ > by a factor of three to five with decreasing Galactic radius from the solar circle to the molecular ring region at about 0.5 R⊙ is typical of the disk overall, we may ask whether it is due simply to an increase in the surface density of gas. The HI surface density as traced by 21-cm emission surveys does not show an increase with decreasing Rgal in this longitude range (Burton 1988, fig. 7.15), but the molecular gas does. The strong variation of < κ > tells us that the cool atomic gas, in contrast to the warm neutral medium that dominates the λ21-cm emissivity, is much more abundant in the inner galaxy than at the solar circle. This underlines the role of the CNM clouds as an intermediate population between structures in the WNM and the much denser and colder molecular clouds. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The effect of this correction is to increase the column density by a factor that is typically for the same range of integration, between one and two. Values of the ratio fc ≡ ∆V are given on table 1, column 10. These correction factors show typical values of 1.4 to 1.6, which is in agreement with similar correction factors found by Dickey and Benson (1982). Note that the Nt values on Table 1, column 8 are typically larger than Nc on column 9 because the range of integration is broader for the former than for the latter, the full inner Galaxy range for the former, and only to the recombination line velocity for the latter. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | which is the nominal value of the spin temperature needed to give the emission and absorption integrals if the gas were at a single temperature. This is given on table 1, column 11 for the velocity range corresponding to columns 7 and 9. Comparison of the emission and absorption spectra channel-by-channel shows that this value is far from realistic, even under the “one-phase” assumption. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Methods for combining the emission and absorption spectra to find the distribution of spin temperatures in the interstellar HI can get complicated. The complications arise from blending of gas at different temperatures in the same spectral channel. Any line of sight at low Galactic latitudes will contain a mixture of several cool clouds which appear as Gaussian line components in the absorption spectrum, plus warm gas which is hardly visible in the absorption spectrum, but which contributes the bulk of the emission. The challenge is to combine the absorption and emission spectra in a way which separates these two thermal phases, giving an estimate of the temperature of the cool gas without the bias introduced In this section we discuss and compare several mathematical by the warm gas emission. techniques for this. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | where TB is the brightness temperature of the emission in velocity channel v, and τ (v) is the optical depth of the absorption in the corresponding velocity channel. This gives the harmonic mean of the temperatures of the various neutral atomic regions which contribute to that velocity channel, i.e. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | where i is an index for the different regions on the line of sight contributing, each with column density Ni and excitation temperature Ti. The distribution of Tsp vs. number of channels for the nine sources with Tc >30K is given on figure 11. We must assume that the temperatures resulting from equation 7 shown on figure 11 are strongly biased to higher values than the true cool HI temperature due to blending with the warm gas. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Improved estimates of the cool phase temperature are based on the shapes of the features in the emission and absorption spectra. A simple approach is to fit the absorption spectrum with a number of Gaussian components. Low latitude λ21-cm emission spectra do not generally look like the sum of distinct Gaussian components, but absorption spectra do, so this is a reasonable approach. When the Gaussian components blend together the best fit parameters become ambiguous, and for a complicated, low latitude spectrum even the number of Gaussian components needed is not clear. Thus there is no unique solution to the Gaussian fitting problem. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 11.— Histogram of the temperatures derived from the naive, one phase model (Tsp, eq. 7). The weighting is simply by number of channels, which roughly corresponds to weighting by volume rather than by mass. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | below the stronger ones in these velocity ranges. Our detection limit for absorption features is thus a function of velocity, and is generally higher than that set by the noise. Note that these boxes do not represent velocity ranges over which the fitting is performed. In cases where several line components are blended we perform the fitting of all Gaussian parameters simultaneously over the full velocity range covered by all components, plus some baseline on either side. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | where the parameters τ0,i, v0,i, and σi for each line component i are varied to minimize χ2 in the usual way (see table 2, columns 3-5). In spite of the double exponential this is no more difficult to fit to the data than a sum of ordinary Gaussians, and it provides a better match to the flat tops seen on the deeper lines. More importantly, the values of τ0,i are significantly larger than the depths of the corresponding Gaussian fits, because of saturation. Thus the derived column densities for the cool gas are larger, often by a factor of two or more, than what would be obtained from simple Gaussian fits. The fitted parameters are obtained using an interactive program based on the Levenberg-Marquardt method of non-linear chi-squared minization (Press et al., 1992). The numbered boxes above the spectra on figures 2-8 indicate the velocity ranges dominated by the different line components (given on Table 2, columns 6 and 7). Weaker components may be hidden below the stronger ones in these velocity ranges. Our detection limit for absorption features is thus a function of velocity, and is generally higher than that set by the noise. Note that these boxes do not represent velocity ranges over which the fitting is performed. In cases where several line components are blended we perform the fitting of all Gaussian parameters simultaneously over the full velocity range covered by all components, plus some baseline on either side. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | distribution of atomic velocities. So we need another approach to combine the emission and absorption spectra to get estimates for the temperatures of the cool phase gas and undo the blending of equation 8. The next simplest thing to do after the one-phase assumption of equation 7 is to assume the HI comes in two phases. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | where c2 is the cool cloud temperature reduced by the background continuum, c2 = (Tcool − Tcont) and ǫ 1). Here we will assume ǫ is constant across the velocity width of each absorption line component, though in general it may be a function of velocity. Note that the continuum brightness, Tcont, is not the brightness temperature Tc of the background source towards which we measure the absorption, but the diffuse continuum in the directions of the nearby pointings which give the interpolated emission spectrum. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | where the sum is taken over all spectral channels, i = 1 to n, covered by a distinct absorption feature (which presumably corresponds to a single cool phase temperature). The parameter ǫ is carried through all the calculations explicitly, so that results can be found for any selected value of ǫ. The advantage of this approach, compared with the Gaussian fits discussed above, is that there is no ambiguity in the results; the unique solution is found directly with no iteration and no need for a “first guess” input by hand. The equations needed to find this best fit are given in the appendix. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 12.— Detailed analysis of the variation of brightness temperature vs. optical depth. The top left panel shows at the top the emission-absorption spectrum pair toward G326.65+0.59 as in figure 3. The upper portions of the other panels show the same two spectra, but blowing up the velocity ranges indicated by the numbered boxes. The emission spectrum has superposed the least squares fit, and below the residuals (data minus fit). The linear function, Tw, is also shown with the emission spectrum. The graphs just below the absorption e−τ ). The x’s on these plots show the individual channel spectra are plots of TB vs. x values. The solid curve on the TB vs. x plots shows the least squares fitted function of the form of equation 14. The coefficients of this function are indicated for each absorption feature sub-plot by the equation at the bottom. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The approach taken here is very similar to that of Mebold et al. (1997) and Dickey et al. (2000) in their studies of the cool HI in the Magellanic Clouds. In those papers the fitting was done by hand by determining the “ridge-line” of each feature measured on the Tem vs. τ plane like those in figures 12 and 13. That approach works best for simple spectra without blended absorption lines, as seen in the Magellanic Clouds and some high and intermediate latitude Galactic directions. At low latitudes the many overlapping absorption features make it difficult to draw the ridge lines by hand, so the fitting technique described in this section is needed. The results for unblended lines are very similar to those derived using the ridgeline method. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The fitted values of c0, c1, and c2 tell us both the temperature of the cool cloud (Tcool = c2 + Tcont) and the linear fit to the warm gas emission across the velocity range of the absorption line component (equation 13). These fit results are shown on table 2, columns 8 through 10. For this analysis we use only the five spectra toward continuum sources with Tc,max >75 K, for which the rms errors in the absorption are less than 0.025. From these fits we can derive new histograms of cool gas temperatures. These are shown on figure 14, for three values of ǫ=0.25, 0.50, and 0.75, which should bracket the true foreground - background distribution. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Figure 14 shows simply the number of spectral channels covered by clouds with a given value of Tcool. This is not a mass weighted average, nor a volume average, but just the statistics of clouds sampled on the small number of lines of sight toward the five strongest continuum sources. Since the sample is small, and groups of channels come out with the same cloud temperature, the histograms have gaps at the lower temperatures, which are certainly unrealistic. These histograms must be considered merely representative of the distribution of HI cloud temperatures, with the general conclusion that most clouds have cool phase temperatures in the range 30 to 100 K, with occasional cooler values. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 14.— Histogram of Tcool. The histograms show the number of spectral channels at each temperature, derived using method 2 (linear least squares). Results for three values of ǫ are shown. The median values of Tcool are 79 K, 65 K, and 51 K for ǫ = 0.75, 0.5, and 0.25, respectively. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 15.— Detail of the linear least squares fitting results for G328.31+0.45. The upper spectrum (emission) shows both the data (x’s and the fitted emission profile, which are nearly indistinguishable, as well as the linear fits to the warm phase emission (c0 + c1v). These piecewise linear fits are not continuous, either with each other or with the emission spectra on either side of the absorption features. This motivates the third method (non-linear least squares) for finding Tcool, which forces the warm phase models to be continuous. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The linear least-squares fit technique described by equations 10-15 and in the appendix does a good job of matching the observed values of TB using the measured absorption values as the independent variable, but this technique ignores what we know about the warm gas emission at velocities on either side of the absorption line. Ideally the linear approximation to Tw given by c0 +c1v should be continuous with the emission spectrum above and below the velocities covered by the absorption line, and in the case of multiple overlapping absorption lines, piecewise linear fits should be continuous between sequential velocity ranges. But this constraint is not incorporated into the fitting algebra, and typically it is not satisfied. An example is shown on figure 15, where the separate least squares fits toward G328.31+0.45 are plotted together. The discontinuities between successive linear approximations to Tw show that we are still missing some information that could improve the model of the HI temperatures. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | boxes on figure 15. This forces the piecewise linear model Tw to be continuous from one component to the next. If we further require that Tw be continuous with TB on the first and last velocity edge, i.e. where the absorption drops below its detection threshold, then for a set of n overlapping line components we have to fit just n values of c2 and (n 1) values of Tw(vi), where i = 2, ..., n and Tw(v1) and Tw(vn+1) are fixed. Figure 16 shows the same spectral pair as in figure 15 (G328.32+0.45), but now fitted using method 3. The estimate of the warm phase contribution is much better behaved. The residuals of the fit to Tem are shown at the bottom. Given that there are fewer than two free parameters to fit for each absorption component, the agreement between the data and the fit is impressive. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Figure 17 (lower panel) illustrates the distribution of cool phase temperatures resulting from this fitting for ǫ=0.5, compared with a copy of the histogram for ǫ=0.5 from figure 14. The third method of fitting changes the results slightly, mainly by reducing the higher values of Tcool and so narrowing the peak near 70K. There is still a tail reaching to low temperatures, and the median value of Tcool is hardly changed (67 K vs. 65 K from method 2). |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | For a single, isolated absorption component, the warm gas level is determined by a linear interpolation between the values of TB on the edges of the line, so there is only one parameter to fit, i.e. c2, that determines the cool gas temperature. This effectively reduces to the method of interpolation used by Kalberla, Schwarz, and Goss (1985) in their study of the emission and absorption in the vicinity of 3C147, and by Mebold et al. (1982) in their study of several high and intermediate latitude regions. If we were to assume that the absorption and warm phase emission components could all be fitted by blended Gaussians then we could derive the cool phase temperature directly. This is the approach taken by Heiles and Troland (2002a,b) in their Arecibo survey of absorption and emission at intermediate latitudes. In appendix 2 we make a detailed comparison of their results with the results of our Method 2, and find that the differences are minor. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The mean opacity, < κ >, derived in section 3 above plays a fundamental role in deter- mining the radiative transfer of the 21-cm line across the Galactic plane. At low latitudes, cool clouds with optical depths of one or more are common enough that some, but not all, of the velocity range in a typical direction will be covered by optically thick gas. How likely this is depends on a contest between < κ > and the velocity gradient, dv ds , along the line of sight. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 16.— Detail of the non-linear least squares fitting results for G328.31+0.45, as in figure 15. The piecewise linear fits for the warm phase emission are now much more reasonable. The data and the fitted values for the total emission are also shown, but almost indistinguishable. The residuals (data minus fit) are shown at the bottom. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Fig. 17.— Comparison of the Tcool distributions obtained with method 2 and method 3. Both histograms assume ǫ=0.5. The upper histogram gives median Tcool = 65 K, the lower histogram gives 67 K. Both are unweighted counts of spectral channels at each temperature, without regard to the total mass of gas in each channel. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | that depends on the local properties of the medium (the abundance of cool clouds) as well as on our particular vantage point (the longitude) and the shape of the rotation curve. In the extremely low velocity gradient case, as near longitudes 0◦and 180◦where dv 0, this ratio ρ is much greater than one, and the optical depth is high at all allowed velocities. In contrast, 90 deg for in the outer galaxy at a longitude where the velocity gradient is moderate, e.g. 10 km s−1 along most of the line of sight, high optical depths are rare, since which ρ is less than one (< κ > in the outer Galaxy is mostly below 5 km s−1 kpc−1). Since the absorbing gas is not smoothly distributed in space but is collected in discrete clouds with small filling factor, the spectra show occasional narrow features of significant optical depth even in regions where the mean optical depth is low. The ratio ρ is useful for predicting optical depths averaged over velocity width ∆V broader than the linewidth of a single cloud, δv . 5 km s−1. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | Using the cool cloud temperature distribution, derived in section 4, together with < κ >, we can derive an approximate value for the peak brightness temperature of the Galactic 21- It was noticed as early as the 1950’s in surveys of the Galactic plane that cm emission. the 21-cm emission typically peaks at TB 100 to 125 K, regardless of the direction or ∼ the velocity gradient. This is a curious fact, clearly seen in the spectra of the Weaver and Williams (1972) survey, for example. Since we know that the neutral hydrogen is spread fairly smoothly from the inner Galaxy to well outside the solar circle, we might expect the peak brightness temperature to correspond to low values of the velocity gradient, as it naturally would if most lines of sight were optically thin. On the other hand, we know that most of the HI is in the warm medium (roughly 75 %, Kulkarni 1983), so the peak brightness temperature does not simply correspond to the cloud temperature, Tcool, since the warm phase gas contributes most of the column density at any velocity. It is clear from emission-absorption spectrum pairs at low latitudes, like those in figures 2-8, that the peak brightness in emission does not always correspond to the velocities of high optical depths; on the contrary, deep absorption lines often correspond to dips in the emission brightness (i.e. HI self-absorption). It is a combination of the opacity, < κ >, the cloud temperature, Tcool, and the warm phase density, nw, that determines the peak brightness temperature. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | For ǫ=0.5 and x=0.63 (τ =1) this gives 102 K. In both cases the contributions from the warm and cool phases are about equal at about 50 K each. Going to higher optical depths does (x not change these numbers much. For τ 115 K and in equation 24 the peak brightness actually decreases to approach Tcool 65K. Of course the actual value of the peak brightness temperature will fluctuate around these values due to the relatively wide distribution of values of Tcool. The highest values we see presumably correspond to the warmer clouds (Tcool & 100 K) and regions of relatively low < κ >, i.e. long path lengths between clouds. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | warm gas with it. Payne, Salpeter and Terzian (1983), estimate that only about 30% of the total HI emission is from the widespread, “intercloud” medium, the rest is associated with the clouds, of which about half must be cool and the rest warm. Braun (1997) sees many examples of brightness temperatures above 150 K; the SMC shows a similarly high peak brightness temperature of 179 K (Stanimirovic et al., 1999). The implication of the high ends of the histograms in figures 14 and 17 is that there must be some clouds in the Milky Way with Tcool as high as 150 K or more, although they are rare compared with the 50 to 100 K clouds. If we could get outside the Milky Way and look back, we might see those clouds unobscured by the cooler, more common clouds, so that seen from outside the Milky Way also may show brightness temperatures of 150 K or more. Braun (1997) also sees a radial variation in the brightness temperatures of the “high brightness network”, with brightness temperatures increasing with radius. This may be due to a radial variation of Tcool, but it could also be explained by a radial variation in < κ > similar to what we find for the Milky Way in section 3; if < κ > decreases with Rgal then Tw,u increases with Rgal in equation 22, which makes the peak brightness temperatures increase. |
Processed_Fitting_Together_the_HI_Absorption_and_Emission_in.txt | The recognition that the peak brightness temperature of the 21-cm line at low latitudes results from a trade-off between absorption in the cool phase and emission from both warm and cool was first described by Baker and Burton (1975) in the context of radiative transfer models to fit the low latitude survey data then available. The role of the velocity gradient in the peak brightness temperature problem is discussed further by Burton (1993). The only innovation in the discussion above is to use the measured values of < κ > together with the mean HI density to get a rough quantitative estimate for the value of the peak brightness temperature, rather than making a full cloudy medium model of the radiative transfer like that of Liszt (1983). The opacity function, < κ(r) >, incorporates the optical depth predictions of these models without making specific assumptions about clouds sizes or geometries. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.