Free Access
Issue
A&A
Volume 578, June 2015
Article Number A135
Number of page(s) 12
Section Galactic structure, stellar clusters and populations
DOI https://doi.org/10.1051/0004-6361/201525902
Published online 17 June 2015

© ESO, 2015

1. Introduction

The large-scale structure of spiral arms in the Milky Way has been a subject of great interest for understanding the dynamics of the Galaxy and for interpreting its properties. However, there have been long-standing disagreements about the number of arms and their physical parameters. While a majority of published papers favor a four-arm structure, others prefer a two-arm structure with a small pitch angle, allowing nearly two turns of the arms within the solar circle. In a series of papers, Vallée (2013, 2014a,b,c) attempted a statistical modeling analysis of all assembled recent positional data on the Milky Way’s spiral arms and observed tangencies in a number of tracers such as CO, H i, methanol masers, hot and cold dust, and far-infrared (FIR) cooling lines, such as [C ii]. The most recent version of his idealized synthesized Galactic map can be found in Vallée (2014a). Modeling the Galactic spiral structure is based mostly on the data at the spiral arm tangents as observed by different gas tracers and stars. However, each of these spiral arm tracers can occupy a separate lane, or layer, across an arm (e.g., Vallée 2014a,c), resulting in an inconsistency among the various models extracted from observational data.

In this paper we present new [C ii] spectral line l-V maps of the Galactic plane covering to 341.̊4 and to 305.̊9 as obtained by Herschel1 HIFI on-the-fly (OTF) mapping. These maps illuminate the structure of different gas components in the spiral arms. Using the [C ii], H i and 12CO maps of the Crux, Norma, and start of the Perseus tangencies, we derive the intensity profiles of their emissions across these spiral arms and quantify the relative displacement of the compressed warm ionized medium (WIM) and the atomic and molecular gas lanes with respect to the inner and outer edges of the arms. Our results reveal an evolutionary transition from the lowest to the highest density states induced by the spiral arm potential. This compressed WIM component traced by [C ii] is distinct from the ionized gas in H ii regions, which traces the spiral arms with characteristics similar to those of molecular gas traced by CO (cf. Vallée 2014a; Downes et al. 1980).

The spiral tangent regions (cf. Vallée 2008; Benjamin 2009) are ideal laboratories in which to study the interaction of the interstellar gas and spiral density waves in the Milky Way. The tangents provide a unique viewing geometry with sufficiently long path lengths in relatively narrow velocity ranges to detect the diffuse WIM component traced by [C ii] emission and to study its relationship to the neutral H i and molecular 12CO gas components within spiral arms and the influence of spiral arm density waves on the interstellar medium (ISM).

COBE FIRAS observed strong [C ii] and [N ii] emission along the Galactic spiral arm tangencies, and Steiman-Cameron et al. (2010) fit the COBE results with four well-defined logarithmic spiral arms in the gaseous component of the ISM. However, COBE’s 7° beam and spectrally unresolved lines preclude obtaining detailed information on the scale and properties of the gas within the spiral tangencies, nor can one use the COBE data to separate the emission that arises from the photon -dominated regions (PDRs) from that in the WIM. The HIFI Galactic Observations of Terahertz C+ (GOT C+) survey (Langer et al. 2010) of the Milky Way also detects the strongest [C ii] emission near the spiral arm tangential directions (Pineda et al. 2013; Velusamy & Langer 2014). In the velocity-resolved HIFI spectra, Velusamy et al. (2012) separated the WIM component of the [C ii] emission in velocity space from that in the molecular and neutral gas. They suggested that excitation by electrons in the WIM, with a density enhanced by the spiral arm potential, accounts for the low surface brightness [C ii] excess observed at the tangent velocities along the Scutum-Crux spiral tangency. To determine whether a similar spatial and density distribution is a general property of Galactic spiral arms, it is important to observe the velocity structure of the [C ii] emission in other spiral arm tangencies and compare it with the corresponding H i and 12CO emissions.

Here we present a large-scale (~15°) position-velocity map of the Galactic plane in [C ii] and derive the following characteristics of the spiral arm features: the relative locations of the peak emissions of the WIM, H i, and molecular gas lanes, including the PDRs, and the width of each layer. In addition, we use the [C ii] emission to derive the mean electron density in the WIM. These results confirm our earlier conclusion (Velusamy et al. 2012) that in the velocity profile of [C ii] emission at the Scutum tangency, the WIM and molecular gas components of [C ii] are distinguished kinematically (appearing at well-separated velocities around the tangent velocity).

In the analysis presented here we use the fact that [C ii] emission can arise in the three major constituents of the interstellar gas: fully or partially ionized gas (as in the WIM), neutral atomic gas (as in HI clouds), and H2 molecular gas (such as CO clouds or PDRs) excited by electrons, H atoms, and H2 molecules, respectively. These components are identified as demonstrated in the GOT C+ results in the velocity-resolved HIFI spectra (cf. Pineda et al. 2013; Langer et al. 2014; Velusamy & Langer 2014). Furthermore, Velusamy & Langer (2014) showed that in the inner Galaxy ~60% of the [C ii] emission traces the H2 molecular gas, while at least 30% of the [C ii] traces the WIM. They also showed that emission in H i excited by H atoms is not a major contributor. To study spiral arm structure, we here identify two major [C ii] components, one component that stems from the WIM, the other from PDRs. We find that in the spiral arm tangencies the [C ii] spectral line data alone can be used to study the relative locations of the WIM and molecular gas PDR layers.

The outline of our paper is as follows. The observations are discussed in Sect. 2. In Sect. 3 we construct the spatial-velocity maps and compare the distributions of [C ii] with H i and 12CO and their relation to the spiral arms. In Sect. 4 we analyze the velocity structure of these gas components at the spiral arm tangencies and use it to infer the relative locations of the different gas lanes within the spiral arm. Note that in discussing the internal structure of the spiral arms we refer to the emission layers of the gas tracers as lanes, analogous to the terminology used in external spiral galaxies. We also derive the average electron density in the WIM using the [C ii] emission and a radiative transfer model. We summarize our results in Sect. 5.

thumbnail Fig. 1

Spiral arms in the fourth quadrant are represented in a VLSR-longitude (V-l) plot, adapted from Vallée (2008): red: Norma-Cygnus; blue: Scutum-Crux; green: Sagittarius-Carina; magenta: Perseus. The rectangular boxes indicate the longitude extent in the Galactic plane at a latitude b = 0° of the HIFI [C ii] spectral line map data presented here. Note that the maps cover the tangencies of the Norma, Crux, and start of the Perseus arms. The vertical dashed line represents a line of sight at a given longitude that intercepts multiple spiral arms, thus demonstrating the need for velocity-resolved spectral line data to separate them.

Open with DEXTER

2. Observations

The longitudinal and velocity coverage of the [C ii] observations at a latitude b = 0°presented here are summarized in Fig. 1. We also show the velocity-longitude relationship of the spiral arm in a schematic view. All [C ii] spectral line map observations were made with the high spectral resolution HIFI (de Graauw et al. 2010) instrument onboard Herschel (Pilbratt et al. 2010). These observations were taken between October 2011 and February 2012. We used 37 OTF scans for the large-scale [C ii] map of the Galactic plane (at b = 0°), covering a 15° range in longitude between 326.̊6 and 341.̊4, which include the Norma and Perseus tangencies. The observations of the fine-structure transition of C+ (2P3/22P1/2) at 1900.5369 GHz were made with the HIFI band 7b using the wideband spectrometer (WBS). Each OTF scan was taken along the Galactic longitude at latitude b = 0° and was 24 arcmin long. For the Crux tangency two OTF longitude scan data were used: (i) one 24 arcmin long centered at and b = 0° (note an earlier version of this map was presented in Velusamy & Langer (2014) and it is included here for completeness in the analysis of the tangencies); and (ii) a shorter 6 arcmin long scan centered at and b = 0.̊15. All HIFI OTF scans were made in the LOAD-CHOP mode using a reference off-source position about 2 degrees away in latitude (at b = 1.̊9). However, we did not use off-source data (see below). All the 24 arcmin long OTF scans were sampled every 40 arcsec and the shorter scan every 20 arcsec. The total duration of each OTF scan was typically ~2000 s, which only provides a short integration time on each spectrum (pixel). Thus the rms (0.22 K in Tmb) in the final maps with an 80′′ beam and 2 km s-1 wide channels in the OTF spectra is much larger than the rms in the HIFI spectra observed in the HPOINT mode, for example, in the GOT C+ data. The observations for the Crux tangency used longer integrations.

We processed the OTF scan map data following the procedure discussed in Velusamy & Langer (2014). The [C ii] spectral line data were taken with HIFI Band 7, which used hot electron bolometer (HEB) detectors. These HEBs produced strong electrical standing waves with characteristic periods of ~320 MHz that depend on the signal power. The HIPE Level 2 [C ii] spectra show these residual waves. We found that applying the task fitHifiFringe2 to the Level 2 data produced satisfactory baselines. However, removal of the HEB standing waves has remained a challenge up until the recent release of HIPE-12, which includes a new tool, hebCorrection3, to remove the standing waves in the raw spectral data by matching the standing-wave patterns (appropriate to the power level) in each integration using a database of spectra at different power levels (see Herschel Science Center (HSC) HIPE-12 release document for details). We used this HSC script to apply hebCorrection to create the final pipeline-mapping products presented here. Following one of the procedures suggested by David Teyssier at the HSC, the OTF map data presented here were processed by re-doing the pipeline without off source subtraction to produce Level 1 data. The hebCorrection was then applied to these new Level 1 data. The fact that hebCorrection subtracts the matching standing-wave patterns from a large database of spectra eliminates the need for off-source subtraction. In our analysis the processed spectral data are therefore free from any off-source contamination. While fitting the HEB waves, we also used the feature in the hebCorrection script to exclude the IF frequencies with strong [C ii] emissions. Finally, the Level 2 data were produced from the HEB-corrected Level 1 data.

thumbnail Fig. 2

Examples of a [C ii] OTF longitudinal scan l-V map centered at and b = 0.̊0. The intensities are in main-beam antenna temperature (Tmb) with values indicated by the color wedge at the top. A square-root color stretch is used to bring out the low-brightness emission features. The velocity resolution in all maps is 2 km s-1, and the restored beam size in longitude is 80′′.

Open with DEXTER

From Level 2 data the [C ii] maps were made as spectral line cubes using the standard mapping scripts in HIPE. Any residual HEB and optical standing waves in the reprocessed Level 2 data were minimized further by applying fitHifiFringe to the gridded spectral data (we took the additional precaution in fitHifiFringe of disabling DoAverage so as not to bias the spectral line window). The H- and V-polarization data were processed separately and were only combined after applying fitHifiFringe to the gridded data. This approach minimizes the standing-wave residues in the scan maps by taking into account the standing-wave differences between H- and V-polarization.

All OTF scan map data were reprocessed and analyzed in HIPE 12.1, as described above, to create spectral line data cubes. We then used the processed spectral line data cubes to make longitude-velocity (l-V) maps of the [C ii] emission as a function of the longitude range in each of the 39 OTF observations. For HIFI observations we used the Wide Band Spectrometer (WBS) with a spectral resolution of 1.1 MHz for all the scan maps. The final l-V maps presented here were restored with a velocity resolution of 2 km s-1. At 1.9 THz the angular resolution of the Herschel telescope is 12′′, but the [C ii] OTF observations used 40′′ sampling. Such fast scanning results in heavily undersampled scans broaden the effective beam size along the scan direction (Mangum et al. 2007). Therefore all [C ii] maps were restored with effective beam sizes corresponding to twice the sampling interval along the scan direction (~80′′). Figure 2 shows an example of a l-V map reconstructed using the map data processed in HIPE 12.1 for a single OTF scan map observed at longitude .

To compare the distribution of atomic and molecular gas with the ionized gas, we used the 12CO and H i data in the southern Galactic plane surveys available in the public archives. The 12CO(10) data were taken from the Three-mm Ultimate Mopra Milky Way Survey4 (ThrUMMS) observed with the 22m Mopra telescope (Barnes et al. 2011). The H i data were taken from the Southern Galactic Plane Survey (SGPS) observed with the Australia Telescope Compact Array (McClure-Griffiths et al. 2005). Although the Park HIPASS Galactic plane map of the H 166,167, and 168α radio recombination lines (RRL) is now available (Alves et al. 2015), its low-velocity resolution (~20 km s-1) precludes using it for the analysis presented here. Furthermore, as the H ii regions are associated with star formation and dense molecular gas (e.g., Anderson et al. 2009), the RRL emission from H ii regions probably does not add more information to the spiral arm structures than is already traced by 12CO.

thumbnail Fig. 3

[C ii] longitude-velocity (l-V) map covering ~15° in longitude over the range to 341.̊4 at b = 0.̊0. The intensities are in main-beam antenna temperature (Tmb) with values indicated by the color wedge. A square-root color stretch is used to bring out the low-brightness emission features. The velocity resolution in all maps is 2 km s-1 and the beam size along the longitudinal direction is 80′′. A sketch of the spiral arms in the fourth quadrant, adopted from Vallée (2008), is overlaid and indicated in the following colors: red-solid: Norma-Cygnus; white-solid: Scutum-Crux; white-dashed: Sagittarius-Carina; red-dashed: Perseus. The rectangular boxes indicate the extent of the spiral tangencies as labeled.

Open with DEXTER

3. Results

In this section we present the l-V emission maps and analyze the structure of the different gas components. We use a schematic of the expected relationship between velocity (VLSR) and location with respect to Galactic center to guide the analysis of the gas lane profile across the arm. We show that the emissions reveal an orderly change in gas components across the arms, leading from the least dense WIM to the densest molecular clouds.

thumbnail Fig. 4

12CO(10) longitude-velocity (l-V) map covering ~15° in longitude over the range to 341.̊4 at b = 0.̊0. The 12CO(10) data are taken from Mopra ThrUMMS survey (Barnes et al. 2011). See the caption of Fig. 3 for the color labels.

Open with DEXTER

thumbnail Fig. 5

H i longitude-velocity (l-V) map covering the ~15° range in longitude l = 326.̊6 to 341.̊4 at b = 0.̊0. The H i data are taken from the SGPS survey (McClure-Griffiths et al. 2005). See the caption of Fig. 3 for the color labels.

Open with DEXTER

thumbnail Fig. 6

a) Schematic view of the Norma spiral arm tangency. The emissions (distinguished by color) tracing the spiral arm are shown as a cross cut of the layers from the inner to outer edges (adapted from Figs. 2 and 3 in Vallée 2014a). b) A sketch indicating the velocity (VLSR) structure of the corresponding spectral line intensities near the tangency for each layer. Note that this cartoon is intended to be a schematic and is not to scale.

Open with DEXTER

3.1. Longitude-velocity maps

We stitched all the individual OTF maps (an example is shown in Fig. 2) within the longitude range 326.̊6 to 341.̊4 to create a single longitude-velocity map that includes the Perseus and Norma tangencies. This l-V map is shown in Fig. 3. An l-V representation of the spiral arms is overlaid on the [C ii] map. Note that strong [C ii] emissions are seen in the tangencies (denoted by the boxes in Fig. 3) as seen in the GOT C+ survey data (Pineda et al. 2013; Velusamy & Langer 2014) and COBE data (Steiman-Cameron et al. 2010). However, the rest of the spiral arm trajectories show poor correspondence with the brightness of [C ii] emission, probably because of the uncertainties in the model parameters (e.g., pitch angle) used for the spiral arms. We assembled the l-V maps that match the [C ii] map for the 12CO(10) (Fig. 4) and H i (Fig. 5) maps using the Mopra ThrUMMS survey and SGPS data. Note that the intensities for ThrUMMS 12CO(10) data are uncorrected for main-beam efficiency (Ladd et al. 2005; Barnes et al. 2011).

The [C ii] maps in Fig. 3 contain in addition to information about the tangency a rich data set on [C ii] emission in the diffuse gas, the molecular gas, and PDRs; these properties have been analyzed in detail for a sparse galactic sample using the GOT C+ data base (Pineda et al. 2013; Langer et al. 2014; Velusamy & Langer 2014). However, the continuous longitude coverage in the data presented here offers a better opportunity for studying the Galactic spiral structure and the internal structure of the arms. In principle, it is possible to derive a 2D spatial-intensity map of this portion of the Galaxy using the kinematic distances for each velocity feature in the maps. However, such a study is subject to the near- far-distance ambiguities inherent in lines of sight inside the solar circle (see discussion in Velusamy & Langer 2014), except along the tangencies. The tangencies offer a unique geometry in which to study kinematics of the interstellar gas without the distance ambiguity. Furthermore, the tangential longitudes provide the longest path length along the line of sight through a cross section of the spiral arm, thus making it easier to detect weak [C ii] emission from the WIM and the molecular gas. In this paper we limit our analysis to the [C ii] emission in tangencies and compare it with those of H i and 12CO to understand the structure of the spiral arms in different ISM components.

3.2. Tangent emission-velocity profiles

In Fig. 6 we show a schematic model of the internal structure of a spiral arm based on the results of Vallée (2014a) and Velusamy et al. (2012), using the Norma tangency as an example. Figure 6a illustrates how the emission for the tangency probes the lanes seen in different tracers. In Vallée’s schematic (see his Fig. 5), both H i and [C ii] emissions occur displaced from 12CO toward the inner edge (on the near side of the Galactic center). In Vallée’s sketch the displacement of [C ii] with respect to 12CO is based on the COBE data (Steiman-Cameron et al. 2010). However, in the velocity-resolved HIFI data for the Scutum tangency (Velusamy et al. 2012), the WIM component of [C ii] emission occurs near the inner edge, while that associated with molecular gas and PDRs is coincident with 12CO emission. The boxed region in Fig. 6a represents the area in the tangency over which the emission spectra are computed. The expected velocity profiles at the tangency, relative to the other emission layers, are shown schematically in Fig. 6b and can be compared with the actual observed velocity profiles for each tracer in the maps in Figs. 3 to 5. Thus we can examine the location of the respective emission layers relative to each other within each spiral arm by analyzing the velocity profiles of different gas tracers.

The trajectory of each spiral arm radial velocity (VLSR) as a function of Galactic longitude is shown as a line in the l-V maps. In reality, however, it is expected to be much broader and complex. Obviously, there is no single unique value of the longitude that can be assigned as a tangency. Furthermore, the longitudes listed in the literature (e.g., Vallée 2014a) often correspond to data averaged over a wider range of Galactic latitudes. But the l-V maps presented here are all in the Galactic plane (b = 0.̊0) observed with narrow beam sizes in latitude (12′′, 33′′, and 150′′ for [C ii], 12CO, and H i, respectively). In the maps in Figs. 3 to 5 we refer to a range of longitudes for each spiral tangency as indicated by the box sizes.

Using the l-V maps in Figs. 3 to 5, we averaged the emissions within a 2° and 2.̊5 longitude range for the Perseus and Norma tangencies, respectively, and in all observed longitudes for the Crux tangency. The resulting averaged spectra are plotted in Figs. 7b and 8b. Note that the 12CO spectra shown have new baselines fitted to the map data in Fig. 4. For clarity we limited the velocity range to only cover the tangencies. Furthermore, the intensity scale for each spectrum was adjusted such that the highest value corresponds to its peak brightness listed in Table 1. The tangent velocity is indicated in the spectra in each panel in Figs. 7, 8, and 10 to provide a reference to the VLSR velocities. The tangent velocities were estimated using the mean longitude and assuming Galactic rotational velocity (220 km s-1) at the tangent points (cf. Levine et al. 2008). We note that the tangent velocity varies from panel to panel, corresponding to its longitude range. The [C ii] spectra of the tangencies show remarkably distinct differences compared to spectra of H i or 12CO. [C ii], unlike H i or 12CO, shows a clear emission peak near the tangent velocity, well separated from the H i and 12CO peaks. To bring out the uniqueness of these differences in the tangencies, we compare in Figs. 7 and 8 the spectra at tangencies (labeled “On-tangent”) with those at neighboring longitudes (labeled “Off-tangent”). As discussed below, only the spectra at the tangencies show the excess [C ii] emission peak at more negative velocities. As illustrated in the schematic in Fig. 6, it is possible to individually delineate the emission layers or the lanes, highlighting the internal structure of the spiral arms, for [C ii] in both the WIM and the PDRs, as well as that of molecular gas in 12CO and atomic gas in H i. The characteristics of the observed velocity profiles in each tangency are summarized Table 1.

thumbnail Fig. 7

Norma tangency spectra. The [C ii], H i, and 12CO emission spectra are plotted against velocity (VLSR). Each panel shows the spectra for the indicated longitude ranges. Note that the intensity scale is normalized to the peak emission within the velocity range. The corresponding 1σ error bars are indicated on the [C ii] and 12CO spectra. The tangent velocity is marked on each panel by a vertical arrow. Panel a): Off-tangent. Panel b): On-tangent.The dashed lines indicate the VLSR shift between the [C ii] and 12CO peaks.

Open with DEXTER

thumbnail Fig. 8

Spectra for the location of the start of the Perseus tangency. The caption is same as for Fig. 7. Panels a) and c): Off-tangent. Panel b): On-tangent. The dashed lines indicate the shifted VLSR of peak emission of the [C ii] and 12CO.

Open with DEXTER

Table 1

Observed spiral arm parameters at the Perseus, Norma, and Crux tangencies.

3.2.1. Norma tangency

The Norma tangency has been determined as 328° for 12CO (Bronfman et al. 2000) and H i (Englmaier & Gerhard 1999), 329° for 60 μm dust (Bloemen et al. 1990), and 332° for 870 μm dust (Beuther et al. 2012). García et al. (2014) assigned tangent directions to the Crux (Centaurus), Norma, and 3 kpc expanding arms of 310.̊0, 330.̊0, and 338.̊0, respectively, by fitting a logarithmic spiral arm model to the distribution of giant molecular clouds (GMCs). As discussed above, the detection of [C ii] from the WIM is strongest along the tangencies and thus is a good discriminator of the tangent direction of a spiral arm. The emission profiles shown in Fig. 7 therefore provide strong evidence that the longitude of the tangent direction is well constrained to . The 12CO emission shows a small peak near the tangent velocity in the on-tangent spectrum. However, this feature is relatively weak when compared to the prominent emission seen in [C ii].

3.2.2. Start of Perseus tangency

The tangency at the start of the Perseus arm has been determined as 336° for 12CO (Bronfman et al. 2000), 338° for the [C ii] and [N ii] FIR lines (Steiman-Cameron et al. 2010), and 338° for 870 μm dust (Beuther et al. 2012). Green et al. (2011) suggested that part of the Perseus arm might harbor some of the methanol masers found toward the tangent direction of the 3 kpc expanding arm. According to the spiral arm model of Russeil (2003), the starting point of the Perseus arm would be found in the region between the Norma and the 3 kpc expanding arms. Although there are some doubts about this longitude being the start of the Perseus arm (cf. Green et al. 2011), we adopted the longitude range l = 336°–338° as the start of the Perseus arm following the work of Vallée (2014a). The emission profiles in Fig. 8 clearly show the detection of a tangent direction in this longitude range, as seen by the strong [C ii]–WIM emission compared to the weaker [C ii] emission in the neighboring longitudes on either side.

3.2.3. Crux tangency

Vallée (2014a) listed the Crux5 tangency as between l = 309°and 311° in different tracers with CO at l = 309° (Bronfman et al. 2000) and dust at l = 311°(Drimmel 2000; Bloemen et al. 1990; Beuther et al. 2012). However, using the GLIMPSE source counts, Benjamin et al. (2005) and Churchwell et al. (2009) placed the Crux tangency within a broader longitude range 306°<l< 313°. Furthermore, it has been suggested that the Crux tangency provides an ideal testing ground for models of spiral density wave theory (Benjamin 2008), as the l = 302°to 313° direction is known to have several distinct anomalies, including large deviations in the H i velocity field (McClure-Griffiths & Dickey 2007) and a clear magnetic field reversal (Brown et al. 2007). Although we do not have a complete map of this region, it is of sufficient interest that we present a partial map covering l = 304.̊9 and 305.̊9, which is close to the Crux tangency. This partial l-V map, which comes from another of our [C ii] Herschel projects, is shown in Fig. 9 along with the corresponding 12CO and H i maps. Note that the velocity range for the emission in the 12CO map is much narrower than in the [C ii] map; this difference is indicative of a broader diffuse emission component in [C ii].

thumbnail Fig. 9

Longitude-velocity (l-V) map covering part of the Crux tangency between longitudes to 305.̊9. a) HIFI [C ii] maps covered in two OTF longitude scans centered at l = 305.1° and 305.76° at b = 0.̊0 and 0.̊15, respectively. b) The corresponding 12CO map from the Mopra ThrUMMS survey. c) The corresponding H i map from SGPS survey. This map region is off to the right of the l-V trajectory of the Crux tangency, as shown in Fig. 1 (see text). See also the caption of Fig. 3 for details on the display.

Open with DEXTER

The emission profiles for the Crux tangent region shown in Fig. 10 are very similar to those observed for the Scutum tangency (Velusamy et al. 2012). The [C ii] emission shows a clear excess beyond the tangent velocity. However, unlike the Norma or Perseus tangencies, we do not detect a resolved [C ii] emission peak, and the WIM component appears as an enhanced emission shoulder under the H i emission profile. This difference may be partly due to the fact that the longitude range of these maps is well outside the nominal tangent direction and may also be due to much stronger emissions seen in both [C ii] and 12CO close to the tangent velocity. Nevertheless, the detection of the [C ii] excess associated with low H i and little or no 12CO and its similarity to the results for the Scutum tangency strongly favor its interpretation as the WIM. This detection of WIM in the longitudes to 305.̊9 indicates that the Crux tangency probably is much broader (l = 302°to 313°) in longitude than the other arms, as was indicated by the analysis of star counts in the Spitzer data (Benjamin et al. 2005).

thumbnail Fig. 10

Crux tangency emission spectra showing [C ii], H i, and 12CO emission plotted against velocity (VLSR). The spectra are averaged over the longitude range indicated. Note that the intensity scale is normalized to the peak emission within the velocity range. The corresponding 1–σ error bars are indicated in the [C ii] and 12CO spectra at VLSR = 50 km s-1. The tangent velocity is also marked. The dashed lines indicate the [C ii] excess beyond the tangent velocity and 12CO peak in the spectra.

Open with DEXTER

4. Discussion

The spectra in Figs. 7, 8, and 10 clearly highlight the exceptional characteristics of the [C ii] emission at the tangencies. These are as follows:

  • 1.

    At the highest velocities near the tangent velocity, only[C ii] shows a peak in emission, while there islittle or no 12CO and H i is weak, but increasingslowly with velocity. (The low-intensity 12CO peak near thetangent velocity in Fig. 7 is less prominent whencompared to the dominant [C ii] emission).

  • 2.

    The velocity of the [C ii] peak near the tangent velocity corresponds to the radial distance closest to the Galactic center. Therefore it is near the inner edge of the spiral arm, representing the onset of the spiral arm feature.

  • 3.

    The peak emissions for H i and 12CO appear at still higher velocities than for the first [C ii] peak corresponding to distances away from the inner edge.

  • 4.

    [C ii] emission shows two peaks: one near or beyond the tangent velocity representing the WIM component traced by [C ii], and the second corresponding to the molecular gas traced by [C ii] observed in association with 12CO arising from the PDRs of the CO clouds.

  • 5.

    The observed velocity profiles are consistent with the schematic shown in Fig. 6 for the internal structure of the spiral arm.

The anomalous excess [C ii] emission in the velocity profiles for all the spiral arm tangencies represents the direct unambiguous detection of the large-scale Galactic diffuse ionized gas (WIM) through its 158 μm [CII] line emission. Our Herschel HIFI detection of the diffuse ionized gas provides detailed spatial and kinematic information on the nature of this gas component in the spiral arms that has not been possible with previous [C ii] surveys. For example, in contrast to the direct detection of the WIM in [C ii] in our HIFI data, the deduction of the WIM component in the COBE [C ii] data (Steiman-Cameron et al. 2010) is indirect because it depends on using the [N ii] intensities to separate the fraction of [C ii] intensity from the WIM from other gas components, and is model dependent.

We can be certain that the [C ii] emissions near the tangent velocity come from the highly ionized WIM and are the result of electron excitation of C+ and not from H atom excitation from diffuse H i clouds (warm neutral medium or cold neutral medium) for the following reasons, as first noted in our earlier study of the WIM in the Scutum-Crux arm (Velusamy et al. 2012; Velusamy & Langer 2014). In the spectra in Figs. 7 and 8, the H i emission at the lowest velocities (near the left around the tangent velocities) is only seen along with [C ii] for the longitudes of tangent directions that are identified to have [C ii] in the WIM (Figs. 7b and 8b). However, in all other longitude directions (Figs. 7a, 8a, and 8c), strong H i emission is seen with little or no associated [C ii] emission. If H atom collisional excitation contributed to any of the [C ii] emission we identify as coming from the WIM, then we should have seen [C ii] associated with the H i emission for all longitudes in Figs. 7 and 8. However, we see none. It is even more unlikely that [C ii] excess in the tangents is associated with CO-dark H2 gas. In the velocity profiles of the On-tangent regions the [C ii] emission starts appearing at lower velocities than 12CO. In contrast, in the velocity profiles of the Off-tangent regions both [C ii] and 12CO begin to appear simultaneously. Considering the relatively weak 12CO emission, we only expect that a small fraction of the [C ii] excess at the tangent is excited by H2 molecules, if at all. Therefore we assume that [C ii] excess is dominated by contribution from excitation by electrons.

The detection of WIM emission from our HIFI OTF survey is surprising because the average electron density in the WIM throughout the disk, ~ a few × 10-2 cm-3, is too low to result in detectable [C ii] emission at the sensitivity of our HIFI OTF maps. Our explanation for the [C ii] emission detected in our survey is that it originates from denser ionized gas along the inner edge of the spiral arms that has been compressed by the spiral density wave shocks (Fig 6), as previously discussed by Velusamy et al. (2012) for the Scutum-Crux arm. Indeed, as shown below, the electron density is significantly higher in the spiral arm WIM than between the arms. Below we derive the physical characteristics of the spiral arm gas lanes and the electron density in the WIM lane.

4.1. Internal structure of spiral arms traced by [C II], H I, and 12CO

We can resolve the anatomical structure of the spiral arms, namely how each gas component is arranged spatially as a function of distance from its edges, using the velocity profiles in the tangencies. We located the peak emissions in the layers traced by [C ii]–WIM component, H i (diffuse atomic clouds), 12CO (the GMCs), and the [C ii] in molecular gas (PDRs), as indicated in the cartoon in Fig. 6, as a function of radial distance from the Galactic center. The radial distances were derived from the observed radial velocities (VLSR) assuming a constant Galactic rotation speed of 220 km s-1 for radius >3 kpc at the distance of these spiral arm tangencies. The Galactocentric distances derived from the VLSR are listed in Table 1. The uncertainty in the Galactocentric distances, which depends on the assumed Galactic rotational speed at the tangency, does not significantly affect our results because we are only interested in the relative displacement between them. It is easy to characterize the emission profiles on the near side of the Galactic center near the tangent velocities where the emissions increase from zero. On the far side of the Galactic center, the emissions become too complex to extract the profile parameters as a result of confusion by emission from adjacent arms. This confusion is especially prominent on either side of the peak of the [C ii] -WIM component. To derive the width of the emission lanes of each tracer, we therefore only used the cleaner profile on the rising portion of spiral feature on the near side to the Galactic center. We computed the Galactocentric radial distances to the VLSR of peaks and the half-intensity points of each emission profile. With these radial distances, we derived the distance between the peak and half-intensity points in each emission lane. The total width was obtained by doubling this value, and the results are summarized in Tables 1 and 2.

We derive an approximate lower bound on the width for the spiral arm by assuming [C ii]–WIM traces the inner edge and 12CO the outer edge, and we list them in Table 2. Note that in the anatomy suggested by Vallée (2014a) the hot dust emission traces the inner edge on the near side of the Galactic center with 12CO tracing the midplane while [C ii] is in between. Using the parameters in Table 1 we sketch the cross-cut view of the spiral arm in Fig. 11. We plot each lane tracing the [C ii]–WIM, H i, and molecular gas (12CO) by an approximate Gaussian profile. Note that we did not plot the lanes for the Crux arm because the available data do not include the main part of this tangency (see Sect. 3.3). We list the results in Table 2 as a rough estimate for this tangency because of the poor longitude coverage of the tangency. Nevertheless, the largest arm width inferred for Crux seems to be consistent with narrower widths for Perseus and Norma. The Crux spiral tangency is at the largest radial distance (~7 kpc), in contrast to ~3.54.5 kpc for the other arms studied here.

Table 2

Derived spiral arm structure at the Perseus, Norma, and Crux tangencies.

The cross-cut emission profiles of the internal structures (Fig. 11) for Norma and the start of the Perseus spiral arms are quite consistent with each other. The overall sizes are ~480 pc and ~500 pc for the Perseus and Norma arms, respectively. To be consistent with Vallée (2014a), we present the location of the emission lanes of [C ii]–WIM and H i with respect to the location of the 12CO peak emission. We identify the inner edge of the arm as the location of the half power point in the [C ii] emission profile on the near side of the Galactic center and on the far side from 12CO. Similarly, we identify the outer edge as the location of the half power point in 12CO emission profile on the far side from the Galactic center. We calculated the spiral arm sizes as the distance between inner and outer edges as marked in Fig. 11. With the exception of the Crux arm (which also has the poor longitude coverage of the tangency in our analysis), the overall sizes are ~480 pc and ~500 pc for the Perseus and Norma arms, respectively. The similar width for these arms suggests that a width of ~500 pc is probably characteristic of all spiral arms in the Galaxy. Our result is close to that obtained by Vallée (2014a), who estimated a mean width ~600 pc for the spiral arms using the range of tangent directions observed using hot dust at the inner edge and 12CO at the outer edge. Our analysis and the values for the widths presented here are likely to be less ambiguous than those of Vallée (2014a) because our data are based on fully sampled longitudinal maps in the Galactic plane observed with similar spatial and velocity resolutions. Another difference between our data and those used by Vallée (2014a) is that in his data the tangent directions were observed using maps averaged over a range of latitudes, while ours are only observed in the plane at b = 0.̊0.

thumbnail Fig. 11

Fitted spiral-arm structure of the emission lanes. Cross-cut view of the structure of the spiral arms Norma and start of Perseus are shown, including the relative locations and widths of the emission lanes. The intensity scale is arbitrary.

Open with DEXTER

As noted above, Vallée (2014a) constructed a cross-section view of the spiral arms that shows where each spiral arm tracer occurs. The view was based on a single value assigned for the longitude of each tangency for the tracers. This approach provided a useful insight into the different gas lanes in a cross-cut of the profile of the spiral arms. However in reality, as seen in the maps in Figs. 3 to 5, the emissions occur over a range of longitudes, and it is too complex to assign a single longitude as the tangency. In other words, all tracers have emission within a range of longitudes representing their tangency. Therefore we took a different approach to resolve and delineate the emission layers within each tangency kinematically by studying the velocity structure in each tracer spectrum. As illustrated in the cartoon of the tangency (Fig. 6), the emissions from different layers will show separate velocity VLSR as a result of the Galactic rotation.

4.2. [C II] emission in the compressed WIM along the spiral arm

The spectra in Figs. 7 to 10 clearly highlight the exceptional characteristics of the [C ii] emission at the tangencies. Near the extreme low velocities (near the inner edge of the spiral arm), only [C ii] shows an emission peak (representing the onset of the spiral feature) while there is little or no 12CO and the H i intensity still increases with velocity. In contrast, both H i and 12CO emission peaks appear at still higher velocities away from the inner edge. This anomalous excess [C ii] emission in the velocity profiles is only observed for the spiral arm tangencies, where the path lengths are longest, and its detection represents the direct unambiguous identification of the large-scale Galactic diffuse ionized gas (WIM) by the 158 μm [C ii] line. The results presented here corroborate the previous detection of the WIM in the velocity-resolved HIFI data for the Scutum tangency (Velusamy et al. 2012). Our Herschel HIFI detection of the diffuse ionized gas provides detailed spatial and kinematic information on the nature of this gas component in the spiral arms.

It has been suggested that any contribution to the [C ii] emission from excitation by H atoms is small and negligible in the WIM (Velusamy et al. 2012; Velusamy & Langer 2014). For the warm neutral medium and WIM conditions (Tk = 8000 K; Wolfire et al. 2003), the critical density for excitation of [C ii] by H atoms is ~1300 cm-3 and for electrons it is ~45 cm-3 (see Goldsmith et al. 2012). Using the H i intensities integrated over the line width of the [C ii]–WIM component and the path lengths listed in Table 2, we estimate a mean H density n(H) ⟩ ~ 0.59 cm-3 and 0.31 cm-3 in the Perseus and Norma tangencies, respectively. As shown for the Scutum tangency (Velusamy et al. 2012), such low H i densities cannot account for the [C ii] emission detected at the tangencies. Our interpretation that the [C ii] emissions near the tangent velocity are the result of C+ excitation by the collisions with electrons in the WIM and not with H atoms is further corroborated by strong evidence seen in the data presented here, as discussed above in Sect. 4.

Although the [C ii] emission from the compressed WIM is observed only at or near the tangent longitudes, it is ubiquitous along the spiral arms in all directions. But, unlike the tangent direction, it is not easy to detect this enhancement in the spiral arms layers because (i) of the insufficient path length through the WIM: it is a factor of 4 to 5 shorter when viewed in any other direction than along the tangency, therefore much higher sensitivities are required to detect the weaker emission; (ii) at other longitudes, the velocities are blended with other components and it is difficult to distinguish the diffuse WIM emission from [C ii] emissions from the molecular gas and PDRs due to the viewing geometry. However, it is possible to separate the emissions with spectral line data obtained with higher sensitivity and by making a spaxel-by-spaxel comparison with CO emission (e.g., Velusamy & Langer 2014). Thus, the WIM component may significantly contribute to the total [C ii] luminosity in galaxies, while being difficult to detect along the average line of sight.

4.2.1. Electron densities in the spiral arm WIM

The electron density of the WIM is an important parameter for understanding the conditions in the ISM, such as the pressure and ionization rate. To estimate the electron densities required to produce the observed [C ii] emission in the WIM, we followed the approach in Velusamy et al. (2012). At the low densities of the diffuse medium, the excitation is subthermal and the emission is optically thin, therefore the intensity in an ionized gas is given by (see Velusamy et al. 2012) (1)which assumes a fully ionized gas, a fractional abundance of C+ with respect to the total hydrogen density, nt, X(C+) = 1.4 × 10-4, and where Lkpc is the path length in kpc, T3 is the kinetic temperature in 103 Kelvin, and (2)is the intensity in K km s-1. The [C ii] intensities given in Table 2 are integrated over the velocity widths for the [C ii]–WIM profiles. The path lengths listed in Table 2 are derived using the Galactocentric radial distance and the thickness of the [C ii] emission layer, as discussed above, assuming approximately circular geometry at the tangencies.

Using this approach and assuming a fully ionized gas and a fractional abundance x(e) = n(e) /nt = 1, we calculated n(e) ⟩ for all three tangencies and found n(e) ⟩ in the range 0.42 to 0.74 cm-3 (see Table 2) for Tk = 8000 K. For a fully ionized gas this implies a total density n(H+) = n(e). These values are strictly a lower limit if the gas is partially ionized, x(e) < 1, but only weakly so. The densities in the WIM at the leading edge of the spiral arms are an order of magnitude higher than the average density in the disk, which is dominated by the interarm gas. Our determination of the WIM density from the [C ii] emission is several times higher than the LOS-averaged densities inferred from pulsar dispersion and Hα measurements, n(e) ~ few × 10-2 cm-3 (Haffner et al. 2009), and we argue that our higher mean value is a result of compression by the WIM-spiral arm interaction.

5. Summary

We presented large-scale [C ii] spectral line maps of the Galactic plane from to 341.̊4 and to 305.̊9 observed with Herschel HIFI using OTF scans. All maps are shown as longitude-velocity (l-V) maps at a latitude b = 0.̊0, except for l = 305.̊7 to 305.̊9, for which b = +0.̊15. The [C ii] l-V maps along with those for H i and 12CO, available from southern Galactic plane surveys (Barnes et al. 2011; McClure-Griffiths et al. 2005) were used to analyze the internal structure of the spiral arms as traced by these gas layers in the Crux (–305.̊9), Norma (l = 327°–329.̊5), and start of Perseus (l = 336°–338°) tangencies. Our key results are as follows:

  • 1.

    We derived the internal structure of the spiral arm features usingthe velocity-resolved emission profiles of[C ii], H i, and 12CO averagedover each tangency. These yielded the relative locations of thepeak emissions of the compressed WIM, H i,and molecular gas lanes, including the PDRs, and we derived thewidth of each gas lane with these values.

  • 2.

    We found that the [C ii] emission has two components. At the extreme velocities near the tangent velocity, only [C ii] shows a peak in emission, while there is little 12CO and H i is weak. This [C ii] component traces the compressed WIM and is displaced by about 9 km s-1 in VLSR corresponding to ~200 pc toward the inner edge of the spiral arm with respect to the 12CO emission. The second [C ii] component is roughly coincident with 12CO and traces the PDRs of the molecular gas. The WIM and molecular gas components of [C ii] are distinguished kinematically (appearing at well-separated velocities around the tangent velocity). Thus, we found that in the spiral arm tangencies the [C ii] spectral line data alone can be used to study the relative locations of the WIM and molecular gas PDR layers.

  • 3.

    The peak velocity of the [C ii]–WIM component is near the tangent velocity and corresponds to the radial distance closest to the Galactic center. Thus it is near the inner edge of the spiral arm, representing the onset of the spiral arm feature. Both H i and 12CO peak emissions appear at still higher velocities, corresponding to distances away from this inner edge. The 12CO profile thus defines the outer edge of the spiral arm. We derived the width of the spiral arm as the distance between the two extremes of the half-power points in the [C ii]–WIM and 12CO emission profiles. We estimated the spiral arm widths as ~500 pc near the start of the Perseus arm and for the Norma arm.

  • 4.

    We interpreted the excess [C ii] near the tangent velocities as compression of the WIM induced by the spiral density waves and as the innermost edge of spiral arms. We used the [C ii] intensities and a radiative transfer model to determine the electron densities of the WIM traced by [C ii]. The electron densities in the compressed WIM are ~0.5 cm-3, about an order of magnitude higher than the average for the disk. The enhanced electron density in the WIM is a result of compression of the WIM by the spiral density wave potential.

  • 5.

    Finally, we suggested that the WIM component traced by [C ii] at the spiral arm tangencies exists all along the spiral arms in all directions, but unlike in the tangent direction, it is difficult to detect because of the insufficient path length of C+ across the arms and because of confusion due to velocities blended with other components. Thus, the WIM component along the spiral arms may significantly contribute to the total [C ii] luminosity in galaxies, while being difficult to detect along the average line of sight.

We demonstrated the usefulness of spectrally resolved Herschel HIFI OTF scan maps of [C ii] emission in unraveling the internal structure of spiral arms using the velocity-resolved spectral line profiles at the spiral arm tangencies. Our results provide direct observational evidence of the cross-section view of the spiral arms, in contrast to the synthetic model by Vallée (2014a), who used the longitude tangents as traced by different tracers. Combining [N ii] with [C ii] yields additional constraints (e.g., Langer et al. 2015; Yildiz et al. 2015), and future [N ii] spectral line maps of the spiral arms are needed to fully characterize the compressed WIM detected here and the Galactic arm-interarm interactions.


1

Herschel is an ESA space observatory with science instruments provided by European-led Principal Investigator consortia and with important participation from NASA.

5

Also referred to as Centaurus as it appears in this constellation.

Acknowledgments

We thank the staffs of the ESA Herschel Science Centre and NASA Herschel Science Center, and the HIFI, Instrument Control Centre (ICC) for their help with the data reduction routines. In addition, we owe special thanks to David Teyssier for clarifications regarding the hebCorrection tool. This work was performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.

References

All Tables

Table 1

Observed spiral arm parameters at the Perseus, Norma, and Crux tangencies.

Table 2

Derived spiral arm structure at the Perseus, Norma, and Crux tangencies.

All Figures

thumbnail Fig. 1

Spiral arms in the fourth quadrant are represented in a VLSR-longitude (V-l) plot, adapted from Vallée (2008): red: Norma-Cygnus; blue: Scutum-Crux; green: Sagittarius-Carina; magenta: Perseus. The rectangular boxes indicate the longitude extent in the Galactic plane at a latitude b = 0° of the HIFI [C ii] spectral line map data presented here. Note that the maps cover the tangencies of the Norma, Crux, and start of the Perseus arms. The vertical dashed line represents a line of sight at a given longitude that intercepts multiple spiral arms, thus demonstrating the need for velocity-resolved spectral line data to separate them.

Open with DEXTER
In the text
thumbnail Fig. 2

Examples of a [C ii] OTF longitudinal scan l-V map centered at and b = 0.̊0. The intensities are in main-beam antenna temperature (Tmb) with values indicated by the color wedge at the top. A square-root color stretch is used to bring out the low-brightness emission features. The velocity resolution in all maps is 2 km s-1, and the restored beam size in longitude is 80′′.

Open with DEXTER
In the text
thumbnail Fig. 3

[C ii] longitude-velocity (l-V) map covering ~15° in longitude over the range to 341.̊4 at b = 0.̊0. The intensities are in main-beam antenna temperature (Tmb) with values indicated by the color wedge. A square-root color stretch is used to bring out the low-brightness emission features. The velocity resolution in all maps is 2 km s-1 and the beam size along the longitudinal direction is 80′′. A sketch of the spiral arms in the fourth quadrant, adopted from Vallée (2008), is overlaid and indicated in the following colors: red-solid: Norma-Cygnus; white-solid: Scutum-Crux; white-dashed: Sagittarius-Carina; red-dashed: Perseus. The rectangular boxes indicate the extent of the spiral tangencies as labeled.

Open with DEXTER
In the text
thumbnail Fig. 4

12CO(10) longitude-velocity (l-V) map covering ~15° in longitude over the range to 341.̊4 at b = 0.̊0. The 12CO(10) data are taken from Mopra ThrUMMS survey (Barnes et al. 2011). See the caption of Fig. 3 for the color labels.

Open with DEXTER
In the text
thumbnail Fig. 5

H i longitude-velocity (l-V) map covering the ~15° range in longitude l = 326.̊6 to 341.̊4 at b = 0.̊0. The H i data are taken from the SGPS survey (McClure-Griffiths et al. 2005). See the caption of Fig. 3 for the color labels.

Open with DEXTER
In the text
thumbnail Fig. 6

a) Schematic view of the Norma spiral arm tangency. The emissions (distinguished by color) tracing the spiral arm are shown as a cross cut of the layers from the inner to outer edges (adapted from Figs. 2 and 3 in Vallée 2014a). b) A sketch indicating the velocity (VLSR) structure of the corresponding spectral line intensities near the tangency for each layer. Note that this cartoon is intended to be a schematic and is not to scale.

Open with DEXTER
In the text
thumbnail Fig. 7

Norma tangency spectra. The [C ii], H i, and 12CO emission spectra are plotted against velocity (VLSR). Each panel shows the spectra for the indicated longitude ranges. Note that the intensity scale is normalized to the peak emission within the velocity range. The corresponding 1σ error bars are indicated on the [C ii] and 12CO spectra. The tangent velocity is marked on each panel by a vertical arrow. Panel a): Off-tangent. Panel b): On-tangent.The dashed lines indicate the VLSR shift between the [C ii] and 12CO peaks.

Open with DEXTER
In the text
thumbnail Fig. 8

Spectra for the location of the start of the Perseus tangency. The caption is same as for Fig. 7. Panels a) and c): Off-tangent. Panel b): On-tangent. The dashed lines indicate the shifted VLSR of peak emission of the [C ii] and 12CO.

Open with DEXTER
In the text
thumbnail Fig. 9

Longitude-velocity (l-V) map covering part of the Crux tangency between longitudes to 305.̊9. a) HIFI [C ii] maps covered in two OTF longitude scans centered at l = 305.1° and 305.76° at b = 0.̊0 and 0.̊15, respectively. b) The corresponding 12CO map from the Mopra ThrUMMS survey. c) The corresponding H i map from SGPS survey. This map region is off to the right of the l-V trajectory of the Crux tangency, as shown in Fig. 1 (see text). See also the caption of Fig. 3 for details on the display.

Open with DEXTER
In the text
thumbnail Fig. 10

Crux tangency emission spectra showing [C ii], H i, and 12CO emission plotted against velocity (VLSR). The spectra are averaged over the longitude range indicated. Note that the intensity scale is normalized to the peak emission within the velocity range. The corresponding 1–σ error bars are indicated in the [C ii] and 12CO spectra at VLSR = 50 km s-1. The tangent velocity is also marked. The dashed lines indicate the [C ii] excess beyond the tangent velocity and 12CO peak in the spectra.

Open with DEXTER
In the text
thumbnail Fig. 11

Fitted spiral-arm structure of the emission lanes. Cross-cut view of the structure of the spiral arms Norma and start of Perseus are shown, including the relative locations and widths of the emission lanes. The intensity scale is arbitrary.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.