EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 536 (December 2011) A&A, 536 (2011) A19 Full HTML
Planck early results
Free Access
Issue
A&A
Volume 536, December 2011
Planck early results
Article Number A19
Number of page(s) 16
Section Interstellar and circumstellar matter
DOI https://doi.org/10.1051/0004-6361/201116479
Published online 01 December 2011

© ESO, 2011

1. Introduction

The matter that forms stars, that is left over after star formation, or that has never experienced star formation comprises the interstellar medium (ISM). The life-cycle and the duration of the various observable phases remains largely unknown, because the nature of the diffuse ISM is difficult to discern, owing to its low temperatures and large angular scales.

The distribution of diffuse interstellar gas, by which we mean gas not in gravitationally-bound structures and not in the immediate vicinity of active star-formation regions, has primarily been assessed using the 21-cm hyperfine line of atomic hydrogen. That line is easily excited by collisions and is optically thin for gas with temperature TK > 50K and velocity dispersion δV > 10kms-1 as long as the column density is less than 9 × 1021 cm-2 (Kulkarni & Heiles 1988). Such conditions are typical of the diffuse ISM pervaded by the interstellar radiation field (ISRF), because photoelectric heating from grain surfaces keeps the gas warm (T > 50K), and observed velocity dispersions (presumably due to turbulence) are typically >10 kms-1. Based on the observed dust extinction per unit column density, N(HI)/AV = 1.9 × 1021 cm-2 mag-1 (Bohlin et al. 1978), the upper column density for optically thin 21-cm lines corresponds to visible extinctions AV < 4.7. Thus the 21-cm line is expected to trace diffuse, warm atomic gas accurately throughout the diffuse ISM, except for lines of sight that are visibly opaque or are particularly cold.

Molecular gas is typically traced by the 2.6-mm 12CO(J = 1 → 0) rotational line in emission, which, like the 21-cm Hi line, can be easily excited because it involves energy levels that can be obtained by collisions. The CO emission line, however, is commonly optically thick, due to its high radiative transition rate. In the limit where the lines are optically thick, the primary information determining the amount of molecular gas in the beam is the line width. If the material is gravitationally bound, then the virial mass is measured and CO can be used as a tracer of molecular mass. It is common astronomical practice to consider the velocity-integrated CO line intensity as measuring the molecular column density, with the implicit assumption that the material is virialized and the mass of the virialized structures is being measured (e.g. Dickman et al. 1986; Heyer et al. 2001). In the diffuse ISM, these conditions typically do not apply. On a physical scale of R (measured in parsecs), interstellar material is only virialized if its column density N > 5.2 × 1021δV2R-1 cm-2 where δV is the velocity dispersion (measured in ). Thus the diffuse ISM is typically gravitationally unbound, invalidating the usage of CO as a virial tracer of the molecular gas mass, except in very compact regions or in regions that are visibly opaque. Although CO can emit in gas with low density, the critical density required for collisional equilibrium is of order 103 cm-3, which further complicates the usage of CO as a tracer. This again is not typical of the diffuse ISM.

To measure the amount and distribution of the molecular ISM, as well as the cold atomic ISM, other tracers of the interstellar gas are required. At least three tracers have been used in the past. These are UV absorption in Werner bands of H2, infrared emission from dust, and γ-ray emission from pion production due to cosmic-rays colliding with interstellar nucleons. The UV absorption is exceptionally sensitive to even very low H2 column densities of 1017 cm-2. Using Copernicus (Savage et al. 1977) and FUSE data, atomic and molecular gas could be measured simultaneously on the sightlines to UV-bright stars and some galaxies. A survey at high Galactic latitudes with FUSE showed that the molecular fraction of the ISM, f(H2) ≡ 2N(H2)/ [2N(H2) + N(HI)]  < 10-3 for lines of sight with total column density less than 1020cm-2, but there is a tremendous dispersion from 10-4 to 10-1 for higher-column density lines of sight (Wakker 2006). Since UV-bright sources are preferentially found towards the lowest-extinction sightlines, an accurate average f(H2) is extremely difficult to determine from the stellar absorption measurements. Along lines of sight toward AGNs behind diffuse interstellar clouds, Gillmon & Shull (2006) found molecular hydrogen fractions of 1–30% indicating significant molecular content even for low-density clouds.

The dust column density has been used as a total gas column density tracer, with the assumption that gas and dust are well mixed. The possibility that dust traces the column density better than Hi and CO was recognized soon after the first all-sky infrared survey by IRAS, which for the first time revealed the distribution of dust on angular scales down to 5′. Molecular gas without CO was inferred from comparingIRAS100μm surface brightness to surveys of the 21-cm and 2.6-mm lines of Hi and CO on 9′ or degree scale by de Vries et al. (1987); Heiles et al. (1988); Blitz et al. (1990). At 3′ scale using Arecibo, the cloud G236+39 was found to have significant infrared emission unaccounted for by 21-cm or 2.6-mm lines, with a large portion of the cloud being possibly H2 with CO emission below detection threshold (Reach et al. 1994). Meyerdierks & Heithausen (1996) also detected IR emission surrounding the Polaris flare in excess of what was expected from the Hi and CO emission, which they attributed to diffuse molecular gas. The all sky far-infrared observations byCOBE – DIRBE(Hauser et al. 1998) made it possible to survey the molecular gas not traced by Hi or CO at the 1° scale (Reach et al. 1998). This revealed numerous “infrared excess” clouds, many of which were confirmed as molecular after detection of faint CO with NANTEN (Onishi et al. 2001). Finally, there are also indications of more dust emission than seen in nearby external galaxies such as the Large Magellanic Cloud (Bernard et al. 2008; Roman-Duval et al. 2010) and the Small Magellanic Cloud (Leroy et al. 2007). This suggests that large fractions of the gas masses of these galaxies are not detected using standard gas tracers.

The γ-rays from the ISM provide an independent tracer of the total nucleon density. As was the case with the dust column density, the γ-ray inferred nucleon column density appears to show an extra component of the ISM not associated with detected 21-cm or 2.6-mm emission; this extra emission was referred to as “dark gas” (e.g. Grenier et al. 2005; Abdo et al. 2010), a term we will adopt in this paper to describe interstellar material beyond what is traced by Hi and CO emission. Grenier et al. (2005) inferred dark gas column densities of order 50% of the total column density toward locations with little or beyond detection threshold CO emission, and general consistency between infrared and γ-ray methods of detection. Recent observations using Fermihave significantly advanced this method, allowing γ-ray emission to be traced even by the high-latitude diffuse ISM. In the Cepheus, Cassiopeia, and Polaris Flare clouds, the correlated excess of dust and γ rays yields dark gas masses that range from 40% to 60% of the CO-bright molecular mass (Abdo et al. 2010).

Theoretical work predicts a dark molecular gas layer in regions where the balance between photodissociation and molecular formation allows H2 to form in significant quantity while in the gas-phase C remains atomic or ionized (Wolfire et al. 2010; Glover et al. 2010). In this paper we describe new observations made with Planck1 (Planck Collaboration 2011a) that trace the distribution of submillimeter emission at 350μm and longer wavelengths. In combination with observations up to 100μm wavelength byIRASandCOBE – DIRBE , we are uniquely able to trace the distribution of interstellar dust with temperatures down to ~10 K. The surface brightness sensitivity of Planck, in particular on angular scales of 5′ to 7°, is unprecedented. Because we can measure the dust optical depth more accurately by including the Planck data, we can now reassess the relationship between dust and gas, and relate it to previous infrared and independent UV and γ-ray results, and compare it to theoretical explanations to determine just how important the dark gas is for the evolution of the ISM.

2. Observations

2.1. Planck data

The Planck first mission results are presented in Planck Collaboration (2011a) and the in-flight performances of the two focal plane instruments HFI (High Frequency Instrument) and LFI (Low Frequency Instrument) are given in Planck HFI Core Team (2011a) and Mennella et al. (2011) respectively. The data processing and calibration of theHFI andLFI data used here is described in Planck HFI Core Team (2011b) and Zacchei et al. (2011) respectively.

Here we use only theHFI (DR2 release) data, the processing and calibration of which are described in Planck HFI Core Team (2011b). In this data the CMB component was identified and subtracted through a needlet internal linear combination (NILC) (Planck HFI Core Team 2011b).

We use the internal variance on intensity () estimated during the Planck data processing and provided with the Planck- HFI data, which we assume represents the white noise on the intensity. Note that this variance is inhomogeneous over the sky, owing to the Planck scanning strategy (Planck Collaboration 2011a), with lower values in the Planck deep fields near the ecliptic poles. We have checked that, within a small factor (<2), the data variance above is consistent with “Jack-Knife” maps obtained from differencing the two halves of the Planck rings. We also use the absolute uncertainties due to calibration uncertainties given in Planck HFI Core Team (2011b) forHFI and summarized in Table1. We note that, for a large scale analysis such as carried out here, variances contribute to a small fraction of the final uncertainty resulting from combining data over large sky regions, so that most of the final uncertainty is due to absolute uncertainties.

Table 1

Characteristics of the data used in this study.

2.2. Ancillary data

2.2.1. HI data

In order to trace the atomic medium, we use the LAB (Leiden/Argentine/Bonn) survey which contains the final data release of observations of the Hi 21-cm emission line over the entire sky (Kalberla et al. 2005). This survey merged the Leiden/Dwingeloo Survey (Hartmann & Burton 1997) of the sky north at δ >  −30° with the IAR (Instituto Argentino de Radioastronomia) Survey (Arnal et al. 2000; Bajaja et al. 2005) of the Southern sky at δ <  −25°. The angular resolution and the velocity resolution of the survey are ~0.6° and ~1.3 kms-1. The LSR velocity range −450 < VLSR < 400 kms-1 is fully covered by the survey with 891 channels with a velocity separation of ΔVch = 1.03 kms-1.

The data were corrected for stray radiation at the Institute for Radioastronomy of the University of Bonn. The rms brightness-temperature noise of the merged database is slightly lower in the southern sky than in the northern sky, ranging over 0.07−0.09 K. Residual uncertainties in the profile wings, due to defects in the correction for stray radiation, are for most of the data below a level of 20 to 40 mK. We integrated the LAB data in the velocity range −400 < VLSR < 400 kms-1 to produce an all sky map of the Hi integrated intensity (WHI), which was finally projected into the HEALPix pixelisation scheme using the method described in Sect.2.3.1.

We estimate the noise level of the WHI map as where Nch(=777) is the number of channels used for the integration, and ΔTrms is the rms noise of the individual spectra measured in the emission-free velocity range mainly in −400 < VLSR < 350 kms-1. The resulting noise of the WHI map is mostly less than ~2.5 K km s-1 all over the sky with an average value of ~1.7 K km s-1, except for some limited positions showing somewhat larger noise (~10 K km s-1).

2.2.2. CO data

In order to trace the spatial distribution of the CO emission, we use a combination of 3 large scale surveys in the 12CO(J = 1 → 0) line.

In the Galactic plane, we use the Dame et al. (2001) survey obtained with the CfA telescope in the north and the CfA-Chile telescope in the south, referred to here as DHT (Dame, Hartmann and Thaddeus). These data have an angular resolution of 8.4′ ± 0.2′ and 8.8′ ± 0.2′ respectively. The velocity coverage and the velocity resolution for these data vary from region to region on the sky, depending on the individual observations composing the survey. The most heavily used spectrometer is the 500kHz filter bank providing a velocity coverage and resolution of 332kms-1and 1.3kms-1, respectively. Another 250kHz filter bank providing the 166kms-1coverage and 0.65kms-1resolution was also frequently used . The rms noises of these data are suppressed down to 0.1–0.3 K (for details, see their Table 1). The data cubes have been transformed into the velocity-integrated intensity of the line (WCO) by integrating the velocity range where the CO emission is significantly detected using the moment method proposed by Dame (2011). The noise level of the WCO map is typically ~1.2 K km s-1, but it varies by a factor of a few depending on the integration range used.

We also use the unpublished high latitude survey obtained using the CfA telescope (Dame et al. 2010, priv. comm.). This survey is still on-going and covers the northern sky up to latitudes as high as |bII| = 70° which greatly increases the overall sky coverage. The noise level of the CO spectra are suppressed to ~0.18 K for the 0.65kms-1velocity resolution, and the total CO intensity was derived by integrating typically 10–20 velocity channels, which results in a noise level of 0.4–0.6K km s-1.

Finally, we combined the above surveys with theNANTEN12CO(J = 1 → 0) survey obtained from Chile. This survey complements some of the intermediate Galactic latitudes not covered by the Dame et al. (2001) maps with an angular resolution of 2.6′. Most of the survey along the Galactic plane has a velocity coverage of ~650 kms-1with a wide band spectrometer, but a part of the survey has a coverage of ~100 kms-1with a narrow band spectrometer. The noise level achieved was 0.4–0.5 K at a velocity resolution of 0.65kms-1. The CO spectra were sampled with the 2′ grid in the Galactic centre, and with the 4′ and 8′ grid along the Galactic plane in the latitude range |b| < 5°and |b| > 5°, respectively. The integrated intensity maps were obtained by integrating over the whole velocity range, excluding regions of the spectra where no emission is observed. The resulting rms noise in the velocity-integrated intensity map varies depending on the width of the emission. This survey along the Galactic plane is still not published in full, but parts of the survey have been analyzed (e.g. Fukui et al. 1999; Matsunaga et al. 2001; Mizuno & Fukui 2004). A large amount of the sky at intermediate Galactic latitude toward the nearby clouds is also covered with a higher velocity resolution of ~0.1 kms-1 with a narrow band spectrometer with a ≲ 100kms-1 band (e.g. Onishi et al. 1999; Kawamura et al. 1999; Mizuno et al. 2001). The velocity coverage, the grid spacing, and the noise level for these data vary, depending on the characteristics of the individual clouds observed, but the quality of the data is high enough to trace the total CO intensity of the individual clouds.

The three surveys were repixelised into the HEALPix pixelisation scheme (Górski et al. 2005) with the appropriate pixel size to ensure Shannon sampling of the beam (Nside=2048 for the NANTEN2 survey and Nside=1024 for the CfA surveys, where the Nside HEALPix parameter controls the number of pixels on the celestial sphere defined as Npix = 12 × Nside2) using the procedure described in Sect.2.3.1.

thumbnail Fig.1

Map of the 12CO(J = 1 → 0) integrated intensity used in this paper combining the Dame et al. (2001) and high latitude survey and the NANTEN survey. The data shown cover 62.8% of the sky. The map is shown at a common resolution of all the sub-surveys of 8.8′. Many small clouds at high latitude are not visible in this rendering of the data.

Each survey was smoothed to a common resolution of 8.8′ through convolution with a Gaussian with kernel size adjusted to go from the original resolution of each survey to a goal resolution of 8.8′, using the smoothing capabilities of the HEALPix software.

We checked the consistency of the different surveys in the common region observed with NANTEN and CfA. We found a reasonably good correlation between the two but a slope indicating that the NANTEN survey yields 24% larger intensities than the CfA values. The origin of this discrepancy is currently unknown. We should note that the absolute intensity scale in CO observations is not highly accurate as noted often in the previous CO papers. Since the CfA survey covers most of the regions used in this paper and has been widely used for calibrating the H2 mass calibration factor XCO, in particular by several generations of gamma ray satellites, we assumed the CfA photometry when merging the data, and therefore rescaled theNANTENdata down by 24% before merging. Note that this an arbitrary choice. The implications on our results will be discussed in Sect.6.1.

The 3 surveys were then combined into a single map. In doing so, data from different surveys falling into the same pixel were averaged using σ2 as a weight. The resulting combined map was then smoothed to the resolution appropriate to this study. Note that part of theNANTENand CfA surveys are undersampled. However, this has very minimal effect on the results presented here since the small data gaps are filled while smoothing the CO data to the coarser resolution of the Hi data (0.6°). The resulting CO integrated intensity map is shown in Fig.1.

2.2.3. IR data

We use the IRIS (Improved Reprocessing of theIRASSurvey)IRAS100μm data (see Miville-Deschênes & Lagache 2005) in order to constrain the dust temperature. The data, provided in the original format of individual tiles spread over the entire sky were combined into the HEALPix pixelisation using the method described in Sect.2.3.1 at a HEALPix resolution (Nside = 2048 corresponding to a pixel size of 1.7′). TheIRAScoverage maps were also processed in the same way. We assume the noise properties given in Miville-Deschênes & Lagache (2005) and given in Table1. The noise level of 0.06MJysr-1 at 100μm was assumed to represent the average data noise level and was appropriately multiplied by the coverage map to lead to the pixel variance of the data.

Table 2

Thermal dust emissivity derived from the correlation with HI emission in the reference region with |bII| > 20° and ((Iν/NH)ref).

2.3. Additional data processing

2.3.1. Common angular resolution and pixelisation

The individual maps are then combined into HEALPix using the intersection surface as a weight. This procedure was shown to preserve photometry accuracy.

The ancillary data described in Sect.2.2 were brought to the HEALPix pixelisation, using a method where the surface of the intersection between each HEALPix pixel with each FITS pixel of the survey data is computed and used as a weight to regrid the data. The HEALPix resolution was chosen so as to match the Shannon sampling of the original data at resolution θ, with a HEALPix resolution set so that the pixel size is <θ/2.4. The ancillary data and the description of their processing will be presented in Paradis et al. (in prep.).

All ancillary data were then smoothed to an appropriate resolution by convolution with a Gaussian smoothing function with appropriate FWHM using the smoothing HEALPix function, and were brought to a pixel size matching the Shannon sampling of the final resolution.

2.3.2. Background levels

Computing the apparent temperature and optical depth of thermal dust over the whole sky requires accurate subtraction of any offset () in the intensity data, either of instrumental or astrophysical origin. Although both theIRISand the Planck- HFI data used in this study have been carefully treated with respect to residual offsets during calibration against theFIRASdata, the data still contains extended sources of emission unrelated to the Galactic emission, such as the Cosmic InfraRed Background (CIB) signal (Miville-Deschênes et al. 2002; Planck Collaboration 2011n) or zodiacal light which could affect the determination of the dust temperature and optical depth at low surface brightness.

In order to estimate the above data offsets, we first compute the correlation between IR and Hi emission in a reference region such that |bII| > 20° and . This was done using the IDL regress routine and iterative removal of outliers. The derived dust emissivities ((Iν/NH)ref) are given in Table 2. The uncertainties given are those derived from the correlation using the data variance as the data uncertainty. The derived emissivities are in agreement with the ensemble average of the values found for the local Hi velocities in Planck Collaboration (2011t) (see their Table 2) for individual smaller regions at high Galactic latitude, within the uncertainties quoted in Table2. Note that these emissivities are used only to derive the offsets in this study.

We then select sky pixels with HI column densities and compute the average Hi column density in this region to be . The offsets are then computed assuming that the dust emissivity in this region is the same as in the reference region, ie, (1)where is the average brightness at frequency ν in the sky region with .

The offset values derived from the above procedure are given in Table2 and were subtracted from the maps used in the rest of this analysis. The offset uncertainties also listed in Table2 were derived from the emissivity uncertainties propagated to the offset values through Eq. (1). When subtracting the above offsets from theIRASand Planck intensity maps, the data variances were combined with the offset uncertainties in order to reflect uncertainty on the offset determination. Note that, for consistency and future use, Table2 also lists emissivities and offset values for FIR-mm datasets not used in this study. Note also that these offsets for Planck data are not meant to replace the official values provided with the data, since they suppress any large scale emission not correlated with Hi, whatever their origin.

thumbnail Fig.2

Upper panel: thermal dust emissivity (Iν/NH)ref from Table2. The dot curve shows a modified black-body at TD = 17.5K and β = 1.8 normalized at 857 GHz, for comparison. The various colours are for different instruments:IRAS(Yellow),DIRBE(light blue), Planck- HFI (red),WMAP(dark blue) and Planck- LFI(green). Lower panel: Offsets from Table2. The error bars are plotted at ± 3σ.

3. Dust temperature and emissivity

3.1. Temperature determination

As shown in previous studies (e.g. Reach et al. 1995; Finkbeiner et al. 1999; Paradis et al. 2009; Planck Collaboration 2011t,u), the dust emissivity spectrum in our Galaxy cannot be represented by a single dust emissivity index β over the full FIR-submm domain. The data available indicate that β is usually steeper in the FIR and flatter in the submm band, with a transition around 500μm. As dust temperature is best derived around the emission peak, we limit the range of frequencies used in the determination to the FIR, which limits the impact of a potential change of β with frequency.

In addition, the dust temperature derived will depend on the assumption made about β, since these two parameters are somewhat degenerate in χ2 space. In order to minimize the above effect, we derived dust temperature maps using a fixed value of the dust emissivity index β. The selected β value was derived by fitting each pixel of the maps with a modified black body of the form Iν ∝ νβBν(TD) in the above spectral range (method referred to as “free β”). This leads to a median value of TD = 17.7K and β = 1.8 in the region at |bII| > 10°. Note that the β value is consistent with that derived from the combination of theFIRAS and Planck- HFI data at low column density in Planck Collaboration (2011t). Inspection of the corresponding TD and β maps indeed showed spurious values of both parameters, caused by their correlation and the presence of noise in the data, in particular in low brightness regions of the maps.

We then performed fits to the FIR emission using the fixed β = 1.8 value derived above (method referred to as “fixed β”). In the determination of TD, we used theIRIS100μm map and the two highestHFI frequencies at 857 and 545 GHz. Although the median reduced χ2 is slightly higher than for the “free β” method, the temperature maps show fewer spurious values, in particular in low brightness regions. This results in a sharper distribution of the temperature histogram. Since we later use the temperature maps to investigate the spectral distribution of the dust optical depth and the dust temperature is a source of uncertainty, we adopt the “fixed β” method maps in the following. The corresponding temperature and uncertainty maps are shown in Fig.3.

Temperature maps were derived at the common resolution of those three channels as well as at the resolution of lower intensity data. The model was used to compute emission in each photometric channels of the instruments used here (IRAS, Planck-HFI), taking into account the colour corrections using the actual transmission profiles for each instrument and following the adopted flux convention. In the interest of computing efficiency, the predictions of a given model were tabulated for a large set of parameters (TD, β). For each map pixel, the χ2 was computed for each entry of the table and the shape of the χ2 distribution around the minimum value was used to derive the uncertainty on the free parameters. This included the effect of the data variance and the absolute uncertainties.

thumbnail Fig.3

Upper panel: all sky map of the dust temperature in K. The temperature is derived from modeling theIRIS100μm and the Planck-HFI emission at 857 and 545 GHz. Lower panel: all sky map of the dust temperature uncertainty in %. The maps are shown in Galactic coordinates with the Galactic centre at the centre of the image. Grey regions correspond to missingIRASdata.

3.2. Angular distribution of dust temperature

thumbnail Fig.4

Details of the dust temperature (left column), dust optical depth at 857 GHz (central column) and dark gas column density (right column) for the Chamaeleon (first line), Aquila-Ophiuchus flare (second line), Polaris flare (third line) and Taurus (fourth line). The temperature and optical depth maps are shown in log scale with a colour scale ranging from 15K (black) to 20K (red) and 1 × 10-5 (black) to 3 × 10-3 (red) respectively. The dark gas column density derived from the optical depth at 857 GHz (see Sect.4) and is shown in linear scale with a colour scale ranging from –3 (black) to 7 × 1021Hcm-2 (red). The contours show the 12CO(J = 1 → 0) integrated intensity at 2, 10 and 20K km s-1. The double line shows the limit of the CO surveys.

The all-sky map of the thermal dust temperature computed as described in Sect.3.1 for β = 1.8 is shown in Fig.3. The elongated regions with missing values in the map correspond to theIRASgaps, where the temperature cannot be determined from the Planck-HFI data alone. The distribution of the temperature clearly reflects the large scale distribution of the VIS-UV starlight radiation field intensity.

Along the Galactic plane, a large gradient can be seen from the outer Galactic regions, with TD ≃ 14 − 15K to the inner Galactic regions around the Galactic center regions with TD ≃ 19K. This asymmetry was already seen at lower angular resolution in theDIRBE (Sodroski et al. 1994) and the FIRAS (Reach et al. 1995) data. The asymmetry is probably due to the presence of more massive stars in the inner Milky Way regions, in particular in the molecular ring. The presence of warmer dust in the inner Galaxy is actually clearly highlighted by the radial distribution of the dust temperature derived from Galactic inversion of IR data (e.g. Sodroski et al. 1994; Paladini et al. 2007; Planck Collaboration 2011q). The origin of the large scale region near (lII, bII)=(340°, − 10°) with TD ≃ 20K is currently unclear, but we note that it corresponds to a region of enhanced X-ray emission in the Rosat All-Sky Survey (RASS). It may therefore correspond to warm dust associated with hot gas pervading the local bubble around the Sun, or a pocket of hot gas in Loop I. Similar large regions with enhanced dust temperature, such as near (lII, bII)=(340°, − 30°) or (lII, bII)=(315°, + 30°) may have a similar origin. Loop I (lII, bII)=(30°, + 45°) is seen as a slightly warmer than average structure at TD ≃ 19K. Running parallel to it is the Aquila-Ophiuchus flare (lII,bII)=(30°, + 20°) with apparent TD ≃ 14K extending to latitudes as high as 60°. The Cepheus and Polaris Flare (lII, bII)=(100–120°, +10–+20°) (see Planck Collaboration 2011t, for a detailed study) is also clearly visible as a lower temperature arch extending up to bII=30° into the North Celestial Pole loop and containing a collection of even colder condensations (TD ≃ 12 − 13   K).

On small angular scales, which are accessible over the whole sky only with the combination of theIRAS and Planck- HFI data at 5′, the map shows a variety of structures that can all be identified with local heating by known single stars or Hii regions for warmer spots and with molecular clouds for colder regions. Figure4 illustrates the high resolution spatial distribution of dust temperature and dust optical depth around some of these regions. Warmer regions include the tangent directions to the spiral Galactic arms in Cygnus (lII, bII)=(80°, 0°) and Carina (lII, bII)=(280°, 0°), hosts to many OB associations, and many Hii regions along the plane. At higher Galactic latitude, dust heated by individual hot stars such in the Ophiuchi region (lII, bII)=(340°, + 20°) with individual stars σ − Sco, ν − Sco, ρ − Oph, ζ − Oph, in Orion (lII, bII)=(210°, − 20°) with the Trapezium stars or in Perseus-Taurus (lII, bII)=(160°, − 20°) with the California Nebula (NGC1499) can clearly be identified. Note the Spica HII region at (lII, bII)=(300°, + 50°) where dust temperatures are TD ≃ 20K due to heating by UV photons from the nearby (80 pc) early-type, giant (B1III) star α Vir.

At intermediate and high latitudes, nearby molecular clouds generally stand out as cold dust environments with TD ≃ 13K. The most noticeable ones are Taurus (lII,bII)=(160°, − 20°) (see Planck Collaboration 2011u, for a detailed study), RCrA (lII, bII)=(0°, − 25°), Chamaeleon (lII, bII)=(300°, − 20°) and Orion (lII, bII)=(200°, − 20°). Numerous cold small scale condensations can readily be found when inspecting the temperature map, which mostly correspond to cold cores similar to those discovered at higher resolution in the Herschel data (e.g. André et al. 2010; Könyves et al. 2010; Molinari et al. 2010; Juvela et al. 2010) and in the Planck Cold-Core catalog (see Planck Collaboration 2011r,s).

Individual nearby Galaxies are also readily identified, in particular the Large (lII, bII)=(279°, − 34°) and the Small Magellanic Cloud (lII, bII)=(301°, − 44°) (see Planck Collaboration 2011m, for a detailed study), as well as M31 and M33.

Near the Galactic poles, the temperature determination becomes noisy at the 5′resolution due to the low signal levels.

3.3. Optical depth determination

thumbnail Fig.5

Maps of the dust optical depths on a log scale, in theIRAS100μm (first row left) and Planck- HFI bands at (first row right), (second row left), (second row right), (third row left), (third row right) and GHz (fourth row). All maps are shown in Galactic coordinates with the Galactic centre at the centre of the image. The missing data in all images correspond to theIRASgaps. The upper and lower bounds of the colour scale are set to τmin = 5 × 10-5 × (λ/100μm)-1.8 and τmax = 10-2 × (λ/100μm)-1.8 respectively.

Table 3

Derived parameters for the dark gas, computed in the region with available CO data and |bII| > 10°.

Maps of the thermal dust optical depth (τD(λ)) are derived using: (2)where Bν is the Planck function and Iν(λ) is the intensity map at frequency ν. The maps are shown in Fig.5. We used resolution-matched maps of TD and Iν(λ) and derived τD(λ) maps at the various resolutions of the data used here. The maps of the uncertainty on τD(λ) (ΔτD) are computed as: (3)

4. Dust/gas correlation

thumbnail Fig. 6

Correlation plots between the dust optical depth atIRAS100μm (upper left),HFI857 GHz (upper right), 545 GHz (lower left) and 353 GHz (lower right) and the total gas column density in the solar neighbourhood (|bII| > 10°). The color scale represents the density of sky pixels on a log scale. The blue dots show a -binned average representation of the correlation. The red line shows the best linear correlation derived at low values (). The vertical lines show the positions corresponding to AV = 0.37mag and AV = 2.5mag. These figures are shown for a single XCO = 2.3 × 1020H2 cm-2/(K km s-1).

We model the dust opacity (τM) as (4)where is the reference dust emissivity measured in low NH regions and XCO = NH2/WCO is the traditional H2/CO conversion factor. It is implicitly assumed that the dust opacity per unit gas column density is the same in the atomic and molecular gas. If this is not the case, this will directly impact our derived XCO since only the product of XCO by the dust emissivity in the CO phase can be derived here. The fit to derive the free parameters of the model is performed only in the portion of the sky covered by all surveys (infrared, Hi, and CO) and where either (1) the extinction is less than a threshold , or (2) the CO is detected with WCO > 1K km s-1. Criterion (1) selects the low-column density regions that are entirely atomic and suffer very small Hi optical depth effects, so that the dust in this region will be associated with the Hi emission at 21-cm. Criterion (2) selects regions where the CO is significantly detected and the dust is associated with both the Hi and the 12CO emission lines. We fit for the following three free parameters: , XCO and . Note that, although does not enter Eq.(4) it is nonetheless derived from the fit since it enters the selection of the data points we use in the fit. The threshold measures the extinction (or equivalently the column density) where the correlation between the dust optical depth and the Hi column density becomes non-linear.

The correlation between the optical depth for various photometric channels and the total gas column density () is shown in Fig.6. The correlations were computed in the region of the sky where the CO data is available (about 63% of the sky) and at Galactic latitudes larger than |bII| > 10°. The τD and WCO maps used were smoothed to the common resolution of the Hi map (0.6°). For these plots, we used a fixed value of XCO = 2.3 × 1020H2 cm-2/(K km s-1). The colours show the density of points in and τD bins. The dots show the binned average correlation. The larger scatter of these points at high comes from the limited number of points in the corresponding bins. The red line shows the τM model values derived from the fit (slope=) to the low part of the data.

It can be seen that the correlation is linear at low values and then departs from linear at (). Above (), where becomes dominated by the CO contribution, the dust optical depth again is consistent with the observed correlation at low for this given choice of the XCO value. Between these two limits, the dust optical depth is in excess of the linear correlation. The same trend is observed in all photometric channels shown, with a similar value for the threshold. It is also observed in theHFI bands at lower frequencies, but the increasing noise at low prevents an accurate determination of the fit parameters.

The best fit parameters for , XCO and are given in Table 3. They were derived separately for each frequency. The uncertainty was derived from the analysis of the fitted χ2 around the best value. The values decrease with increasing wavelength, as expected for dust emission. The resulting dust optical depth SED is shown in Fig.7. The dust optical depth in low column density regions is compatible with β = 1.8 at high frequencies. The best fit β value between theIRAS100μm and theHFI857 GHz_is actually found to be β = 1.75. The SED then flattens slightly at intermediate frequencies with a slope of β = 1.57 around λ = 500μm then steepens again to β = 1.75 above 1mm. The XCO values derived from the fit are constant within the error bars, which increase with wavelength. The average value, computed using a weight proportional to the inverse variance is given in Table3 and is found to be XCO = 2.54 ± 0.13 × 1020H2 cm-2/(K km s-1). Similarly, the parameter does not significantly change over the whole frequency range and the weighted average value is found to be  mag.

The excess column density is defined using the difference between the best fit and the observed dust opacity per unit column density using, (5)The map is used to derive the total excess mass () assuming a fiducial distance of 200 pc to the gas responsible for the excess.

thumbnail Fig. 7

Dust optical depth derived from this study using theIRASand Planck- HFI frequencies. The square symbol shows the emissivity at 250μm derived by Boulanger et al. (1996). The dash and dash-dot lines show a power law emissivity with λ-1.8 and λ-1.75 respectively, normalized to the data at 100μm. The error bars shown are ± 1σ.

We also computed the atomic and molecular total gas masses over the same region of the sky, assuming the atomic and molecular gas emissions arise from a gas volume at the same distance. In the region covered by the CO survey, the Hi to CO mass ratio derived for XCO = 2.54 × 1020H2 cm-2/(K km s-1) is MHI/MCO = 4. Using the average and XCO values above, the ratio of the dark gas mass to the atomic gas mass () and to the molecular gas mass () are given in Table3. On average, at high Galactic latitudes, the dark gas masses are of the order of 28% ± 3% of the atomic gas mass and ≃ 118% ± 12% of the molecular mass. We note that, since the scale-height of the HI layer is larger than that of the molecular layer, and that the dark gas component is likely to have an intermediate scale-height, the true and values are likely to be respectively lower and larger than the values quoted above, which assume the same distance for all three components.

thumbnail Fig.8

Map of the excess column density derived from the 857 GHz data. The map is shown in Galactic coordinates with the Galactic centre at the centre of the image. The grey regions correspond to those where noIRASor CO data are available, regions with intense CO emission (WCO > 1K km s-1) and the Galactic plane (|bII| < 5°).

5. Dark-gas spatial distribution

The spatial distribution of the dark gas as derived from τD computed from theHFI857 GHz channel is shown in Fig.8. It is shown in the region where the CO data are available and above Galactic latitudes of |bII| > 5°. Regions where WCO > 1K km s-1 have also been excluded. The map clearly shows that the dark gas is distributed mainly around the best known molecular clouds such as Taurus, the Cepheus and Polaris flares, Chamaeleon and Orion. The strongest excess region is in the Aquila-Ophiuchus flare, which was already evident in Grenier et al. (2005).

Significant dark gas is also apparent at high latitudes, south of the Galactic plane in the anticenter and around known translucent molecular clouds, such as MBM53 (lII=90°, bII= −30°). As with all the molecular clouds, the spatial distribution of the dark gas closely follows that of the Gould-Belt (Perrot & Grenier 2003) and indicates that most of the dark gas in the solar neighbourhood belongs to this dynamical structure.

6. Discussion

6.1. Dust emissivity in the atomic neutral gas

In the solar neighbourhood, Boulanger et al. (1996) measured an emissivity value in the diffuse medium of 10-25cm2/H at 250 μm assuming a spectral index β =  2 which seemed consistent with their data. The optical depth of dust derived in our study in the low regions at |bII| > 10° is shown in Fig.7. The figure also shows the reference value by Boulanger et al. (1996) which is in good agreement with the values derived here, interpolated at 250μm (in fact 10% above when using β = 1.8 and 6% above when using β = 1.75). Our study does not allow us to measure the emissivity in the molecular gas, since we are only sensitive to the product of this emissivity with the XCO factor. However, we note that our derived average XCO = 2.54 × 1020H2 cm-2/(K km s-1) is significantly higher than previously derived values. Even if we account for the possible uncertainty in the calibration of the 12CO(J = 1 → 0) emission (24%) discussed in Sect.2.2.2, increasing the CO emission by the corresponding factor would only lower our XCO estimate to XCO = 2.2H2 cm-2/(K km s-1). In comparison, a value of (1.8 ± 0.3) × 1020H2 cm-2/(K km s-1) was found at |bII| > 5° from the comparison of the Hi, CO, and IRAS 100μm maps (Dame et al. 2001). Similarly, values derived from γ-ray Fermi data can be as low as XCO = 0.87 × 1020H2 cm-2/(K km s-1) in Cepheus, Cassiopea and Polaris (Abdo et al. 2010). This could be evidence that the dust emissivity in the high-latitude molecular material could be larger than in the atomic phase by a factor  ≃3. Such an increase in the dust emissivity in molecular regions has been inferred in previous studies (e.g. Bernard et al. 1999; Stepnik et al. 2003) and was attributed to dust aggregation.

6.2. Dark molecular gas

The nature of “dark molecular gas” has recently been investigated theoretically by Wolfire et al. (2010), who specifically address the HI/H2 and C/C+ transition at the edges of molecular clouds. The nominal cloud modeled in their study is relatively large, with total column density 1.5 × 1022cm-2, so the applicability of the results to the more translucent conditions of high-Galactic-latitude clouds is not guaranteed. The envelope of the cloud has an Hi column density of 1.9 × 1021cm-2 which is more typical of the entire column density measured at high latitudes. Wolfire et al. (2010) define fDG as the fraction of molecular gas that is dark, i.e. not detected by CO. In the nominal model, the chemical and photodissociation balance yields a total H2 column density of 7.0 × 1021cm-2, while the “dark” H2 in the transition region where CO is dissociated has a column density of 1.9 × 1021cm-2. The fraction of the total gas column density that is molecular, (6)is 93% in the nominal model, which suggests that the line of sight through such a cloud passes through material which is almost entirely molecular. To compare the theoretical model to our observational results, we must put them into the same units. We define the dark gas fraction as the fraction of the total gas column density that is dark, (7)For the nominal Wolfire et al. (2010) model, fDG = 0.29 so we can infer fDARK = 0.27. The smaller clouds in Figure 11 of their paper have larger fDG, but f(H2) is also probably smaller (not given in the paper) so we cannot yet definitively match the model and observations. These model calculations are in general agreement with our observational results, in that a significant fraction of the molecular gas can be in CO-dissociated “dark” layers.

If we assume that all dark molecular gas in the solar neighbourhood is evenly distributed to the observed CO clouds, the average fDG measured is in the range fDG = 1.06 − 1.22. This is more than three times larger than predicted by the Wolfire et al. (2010) mass fraction. This may indicate that molecular clouds less massive than the ones assumed in the model actually have a dark gas mass fraction higher by a factor of about three. This would contradict their conclusion that the dark mass fraction does not depend on the total cloud mass. We also note that Wolfire et al. (2010) used the Solomon et al. (1987) value for the mean column density of GMCs, but that a more recent study by Heyer et al. (2009) shows column density values 2–5 times lower. It can be expected that the reduced column density would allow deeper UV penetration and perhaps, a larger fractional composition of dark gas, as implied by our results.

The location of the Hi-to-H2 transition measured here () is comparable, although slightly higher than that predicted in the Wolfire et al. (2010) model (). Again, this difference may indicate variations with the cloud size used, since UV shadowing by the cloud itself is expected to be less efficient for smaller clouds, leading to a transition deeper into the cloud.

Our results can also be compared with the amount of dark molecular gas inferred from recent observations of the CII transition with Herschel – HIFIin the framework of the GOT C+ survey (Langer et al. 2010). In transition clouds showing emission but no emission, Velusamy et al. (2010) found an average , which is significantly smaller than our value of 1.18. This difference could be due to the different characteristics of the clouds included in the GOT C+ survey, which are more distant and probably of larger physical size than those studied here in the solar neighborhood. However, it is most likely that the differences are due to the CII emission sampling preferentially warm and dense regions of the ISM, owing to the density and temperature sensitivity of the excitation of the CII transition, while the dust based detection of the dark-gas described here is in principle sensitive to all densities. This difference would explain why we see larger dark gas fractions with respect to CO, because the CII observations are likely to miss part of the diffuse dark gas.

6.3. Other possible origins

The observed departure from linearity between τD and the observable gas column density could also in principle be caused by variations of the dust/gas ratio (D/G). However, such variations with amplitude of 30% in the solar neighbourhood and a systematic trend for a higher D/G ratio in denser regions would be difficult to explain over such a small volume and in the presence of widespread enrichment by star formation. However, the fact that the dark gas is also seen in the γ-ray with comparable amplitudes is a strong indication that it originates from the gas phase. The dark gas column-densities inferred from the γ-ray observations are also consistent with a standard D/G ratio (Grenier et al. 2005).

The observed excess optical depth could also in principle be due to variations of the dust emissivity in the FIR-Submm. We expect such variations to occur if dust is in the form of aggregates with higher emissivity (e.g. Stepnik et al. 2003) in the dark gas region. We note however that such modifications of the optical properties mainly affect the FIR-submm emissivity and are not expected to modify significantly the absorption properties in the Visible. Therefore, detecting a similar departure from linearity between large-scale extinction maps and the observable gas would allow us to exclude this possibility.

Sky directions where no CO is detected at the sensitivity of the CO survey used (0.3–1.2K km s-1) may actually host significant CO emission, which could be responsible for the excess dust optical depth observed. Evidences for such diffuse and weakly emitting CO gas have been reported. For instance, in their study of the large-scale molecular emission of the Taurus complex, Goldsmith et al. (2008) have found that half the mass of the complex is in regions of low column density NH < 2 × 1021cm-2, seen below WCO ≃ 1K km s-1. However, Barriault et al. (2010) reported a poor spatial correlation between emission by diffuse CO and regions of FIR excess in two high Galactic latitude regions in the Polaris Flare and Ursa Major. The difficulty at finding the CO emission associated to dark gas is that the edges of molecular clouds tend to be highly structured spatially, which could explain why many attempts have been unsuccessful (see for instance Falgarone et al. 1991). In our case, it is possible to obtain an upper limit to the contribution of weak CO emission below the survey detection threshold, by assuming that pixels with undetected CO emission actually emit with WCO = 0.5K km s-1. This is the detection limit of the survey we use at |b| > 10° so this should be considered an upper limit to the contribution of undetected diffuse CO emission. In that case, the dark gas mass is reduced by a factor lower than 20%. This indicates that, although diffuse weak CO emission could contribute a fraction of the observed excess emission, it cannot produce the bulk of it.

Finally, we recognize that the optically thin approximation used here for the Hi emission may not fully account for the whole atomic gas present, even at high latitude. Hi emission is subject to self absorption and NH can be underestimated from applying too high a spin temperature (Ts) while deriving column densities. Ts is likely to vary from place to place depending on the relative abundance of CNM clumps (with thermodynamical temperatures of 20–100K) and WNM clouds (at several thousand K) in the telescope beam. The effective spin temperature of 250–400K to be applied to correct for this blending and to retrieve the total column density from the Hi spectra does not vary much in the Galaxy (Dickey et al. 2003, 2009). It indicates that most of the Hi mass is in the warm phase and that the relative abundance of cold and warm Hi is nearly constant across the Galaxy (outside of the inner molecular ring). The correlation between the Fermi γ-ray maps and the Hi column densities derived for different spin temperatures also support an average (uniform) effective spin temperature > 250K on and off the plane (Ackermann et al. 2011). In order to test these effects, we performed the analysis described in this paper using a very low choice for the Hi spin temperature. We adopted a value of Ts = 80K when the observed Hi peak temperature is below 80K and Ts = 1.1 × Tpeak when above. Under this hypothesis, we obtained dark gas fractions which are about half of those given in Table3 under the optically thin approximation. We consider this to indicate that significantly less than half of the detected dark gas could be dense, cold atomic gas. We further note that, under the optically thin Hi hypothesis, the dark gas fraction appears very constant with Galactic latitude down to |bII| ≃ 3° (see Sect.6.4), while it varies more strongly using Ts = 80K. This does not support the interpretation that the bulk of the dust excess results from underestimated Hi column densities.

6.4. Dark-gas variations with latitude

thumbnail Fig. 9

Fractional mass of the dark gas with respect to the neutral gas mass as a function of the lower bIIvalue used in the analysis. The solid curve is computed under the assumption of optically thin Hi, the dashed curve is for computed using Ts = 80K. Error bars are 1σ.

We investigate the distribution of the dark gas as a function of Galactic latitude. This is important, since the dark gas template produced here for the solar neighbourhood is also used in directions toward the plane for Galactic inversion purpose in Planck Collaboration (2011q). We performed the calculations described in Sect.4 for various values of the Galactic latitude lower cutoff (bmin) in the range bmin < |bII| < 90° with bmin varying from 0° to 10°. For each value, we used the best fit parameters derived from bmin = 10° and given in Table3.

Figure9 shows the evolution of the dark gas mass fraction with respect to the atomic gas mass as a function of bmin. It can be seen that the ratio changes only mildly (increases by a factor 1.12 from bmin = 10° to bmin = 2°) as we approach the Galactic plane. This indicates that a fairly constant fraction of the dark gas derived from the solar neighbourhood can be applied to the rest of the Galaxy.

Figure9 also shows the same quantity computed using the Hi column density derived using Ts = 80K. It can be seen that, in that case, the dark gas fraction is predicted to decrease by a factor 2.12 from bmin = 10° to bmin = 2°. This is caused by the much larger inferred Hi masses toward the plane under this hypothesis. We consider it unlikely that the dark gas fraction varies by such a large factor from the solar neighbourhood to the Galactic plane, and consider it more likely that the correction applied to NH by using a spin temperature as low as Ts = 80K actually strongly overestimates the Hi opacity, and therefore the fraction of the dark gas belonging to atomic gas.

7. Conclusions

We used the Planck-HFI andIRASdata to determine all sky maps of the thermal dust temperature and optical depth. The temperature map traces the spatial variations of the radiation field intensity associated with star formation in the Galaxy. This type of map is very important for the detailed analysis of the dust properties and their spatial variations.

We examined the correlation between the dust optical depth and gas column density as derived from Hi and CO observations. These two quantities are linearly correlated below a threshold column density of corresponding to AV < 0.4mag. Below this threshold, we observed dust emissivities following a power-law with β ≃ 1.8 below λ ≃ 500μm and flattening at longer wavelengths. Absolute emissivity values derived in the FIR are consistent with previous estimates.

This linear correlation also holds at high column densities () corresponding to AV = 2.5mag where the total column density is dominated by the molecular phase for a given choice of the XCO factor. Under the assumption that the dust emissivity is the same in both phases, this leads to an estimate of the average local CO to H2 factor of XCO = 2.54 × 1020H2 cm-2/(K km s-1). The optical depth in the intermediate column density range shows an excess in all photometric channels considered in this study. We interpret the excess as dust emission associated with dark gas, probably in the molecular phase where H2 survives photodissociation, while the CO molecule does not.

In the solar neighbourhood, the derived mass of the dark gas, assuming the same dust emissivity as in the Hi phase is found to correspond to ≃ 28% of the atomic mass and ≃ 118% of the molecular gas mass. The comparison of this value with the recent calculations for dark molecular gas around clouds more massive than the ones present in the solar neighbourhood indicates a dark gas fraction about three times larger in the solar neighbourhood. The threshold for the onset of the dark gas transition is found to be ≃ 0.4mag and appears compatible to, although slightly larger than, the thresholds predicted by this model. Finally, we stress that the Hi 21 cm line is unlikely to be fully optically thin and to measure all the atomic gas. Therefore, the dark gas detected here could well represent a mixture of dark molecular and dark (optically thick) atomic gas seen through its dust emission. For an average Hi spin temperature of 80K, the mixture is predicted to be 50% atomic and 50% molecular.

1

Planck (http://www.esa.int/Planck) is a project of the European Space Agency (ESA) with instruments provided by two scientific consortia funded by ESA member states (in particular the lead countries: France and Italy) with contributions from NASA (USA), and telescope reflectors provided in a collaboration between ESA and a scientific consortium led and funded by Denmark.

Acknowledgments

A description of the Planck Collaboration and a list of its members can be found at http://www.rssd.esa.int/index.php?project=PLANCK&page=Planck_Collaboration

References

  1. Abdo, A. A., Ackermann, M., Ajello, M., et al. 2010, ApJ, 710, 133 [NASA ADS] [CrossRef] [Google Scholar]
  2. Ackermann, M., Ajello, M., Baldini, L., et al. 2011, ApJ, 726, 81 [NASA ADS] [CrossRef] [Google Scholar]
  3. André, P., Men’shchikov, A., Bontemps, S., et al. 2010, A&A, 518, L102 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  4. Arnal, E. M., Bajaja, E., Larrarte, J. J., Morras, R., & Pöppel, W. G. L. 2000, A&AS, 142, 35 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  5. Bajaja, E., Arnal, E. M., Larrarte, J. J., et al. 2005, A&A, 440, 767 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  6. Barriault, L., Joncas, G., Falgarone, E., et al. 2010, MNRAS, 406, 2713 [NASA ADS] [CrossRef] [Google Scholar]
  7. Bernard, J., Reach, W. T., Paradis, D., et al. 2008, AJ, 136, 919 [NASA ADS] [CrossRef] [Google Scholar]
  8. Bernard, J. P., Abergel, A., Ristorcelli, I., et al. 1999, A&A, 347, 640 [NASA ADS] [Google Scholar]
  9. Blitz, L., Bazell, D., & Desert, F. X. 1990, ApJ, 352, L13 [NASA ADS] [CrossRef] [Google Scholar]
  10. Bohlin, R. C., Savage, B. D., & Drake, J. F. 1978, ApJ, 224, 132 [NASA ADS] [CrossRef] [Google Scholar]
  11. Boulanger, F., Abergel, A., Bernard, J., et al. 1996, A&A, 312, 256 [NASA ADS] [Google Scholar]
  12. Dame, T. M. 2011, [arXiv:1101.1499] [Google Scholar]
  13. Dame, T. M., Hartmann, D., & Thaddeus, P. 2001, ApJ, 547, 792 [NASA ADS] [CrossRef] [Google Scholar]
  14. de Vries, H. W., Thaddeus, P., & Heithausen, A. 1987, ApJ, 319, 723 [NASA ADS] [CrossRef] [Google Scholar]
  15. Dickey, J. M., McClure-Griffiths, N. M., Gaensler, B. M., & Green, A. J. 2003, ApJ, 585, 801 [NASA ADS] [CrossRef] [Google Scholar]
  16. Dickey, J. M., Strasser, S., Gaensler, B. M., et al. 2009, ApJ, 693, 1250 [NASA ADS] [CrossRef] [Google Scholar]
  17. Dickman, R. L., Snell, R. L., & Schloerb, F. P. 1986, ApJ, 309, 326 [NASA ADS] [CrossRef] [Google Scholar]
  18. Falgarone, E., Phillips, T. G., & Walker, C. K. 1991, ApJ, 378, 186 [NASA ADS] [CrossRef] [Google Scholar]
  19. Finkbeiner, D. P., Davis, M., & Schlegel, D. J. 1999, ApJ, 524, 867 [NASA ADS] [CrossRef] [Google Scholar]
  20. Fukui, Y., Onishi, T., Abe, R., et al. 1999, PASJ, 51, 751 [NASA ADS] [Google Scholar]
  21. Gillmon, K., & Shull, J. M. 2006, ApJ, 636, 908 [NASA ADS] [CrossRef] [Google Scholar]
  22. Glover, S. C. O., Federrath, C., Mac Low, M., & Klessen, R. S. 2010, MNRAS, 404, 2 [NASA ADS] [Google Scholar]
  23. Goldsmith, P. F., Heyer, M., Narayanan, G., et al. 2008, ApJ, 680, 428 [NASA ADS] [CrossRef] [Google Scholar]
  24. Górski, K. M., Hivon, E., Banday, A. J., et al. 2005, ApJ, 622, 759 [NASA ADS] [CrossRef] [Google Scholar]
  25. Grenier, I. A., Casandjian, J., & Terrier, R. 2005, Science, 307, 1292 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  26. Hartmann, D., & Burton, W. B. 1997, Atlas of Galactic Neutral Hydrogen, ed. D. Hartmann, & W. B. Burton [Google Scholar]
  27. Hauser, M. G., Arendt, R. G., Kelsall, T., et al. 1998, ApJ, 508, 25 [NASA ADS] [CrossRef] [Google Scholar]
  28. Heiles, C., Reach, W. T., & Koo, B. 1988, ApJ, 332, 313 [NASA ADS] [CrossRef] [Google Scholar]
  29. Heyer, M., Krawczyk, C., Duval, J., & Jackson, J. M. 2009, ApJ, 699, 1092 [NASA ADS] [CrossRef] [Google Scholar]
  30. Heyer, M. H., Carpenter, J. M., & Snell, R. L. 2001, ApJ, 551, 852 [NASA ADS] [CrossRef] [Google Scholar]
  31. Juvela, M., Ristorcelli, I., Montier, L. A., et al. 2010, A&A, 518, L93 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  32. Kalberla, P. M. W., Burton, W. B., Hartmann, D., et al. 2005, A&A, 440, 775 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  33. Kawamura, A., Onishi, T., Mizuno, A., Ogawa, H., & Fukui, Y. 1999, PASJ, 51, 851 [NASA ADS] [Google Scholar]
  34. Könyves, V., André, P., Men’shchikov, A., et al. 2010, A&A, 518, L106 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  35. Kulkarni, S. R., & Heiles, C. 1988, in Galactic and extragalactic radio astronomy (Berlin, New York: Springer), 2nd edn, 95 [Google Scholar]
  36. Langer, W. D., Velusamy, T., Pineda, J. L., et al. 2010, A&A, 521, L17 [Google Scholar]
  37. Leroy, A., Bolatto, A., Stanimirovic, S., et al. 2007, ApJ, 658, 1027 [NASA ADS] [CrossRef] [Google Scholar]
  38. Matsunaga, K., Mizuno, N., Moriguchi, Y., et al. 2001, PASJ, 53, 1003 [NASA ADS] [Google Scholar]
  39. Mennella, A., Butler, R. C., Curto, A., et al. 2011, A&A, 536, A3 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  40. Meyerdierks, H., & Heithausen, A. 1996, A&A, 313, 929 [NASA ADS] [Google Scholar]
  41. Miville-Deschênes, M., & Lagache, G. 2005, ApJS, 157, 302 [NASA ADS] [CrossRef] [Google Scholar]
  42. Miville-Deschênes, M., Lagache, G., & Puget, J. 2002, A&A, 393, 749 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  43. Mizuno, A., & Fukui, Y. 2004, in Milky Way Surveys: The Structure and Evolution of our Galaxy, ed. D. Clemens, R. Shah, & T. Brainerd, ASP Conf. Ser., 317, 59 [Google Scholar]
  44. Mizuno, A., Yamaguchi, R., Tachihara, K., et al. 2001, PASJ, 53, 1071 [NASA ADS] [CrossRef] [Google Scholar]
  45. Molinari, S., Swinyard, B., Bally, J., et al. 2010, A&A, 518, L100 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  46. Onishi, T., Kawamura, A., Abe, R., et al. 1999, PASJ, 51, 871 [Google Scholar]
  47. Onishi, T., Yoshikawa, N., Yamamoto, H., et al. 2001, PASJ, 53, 1017 [NASA ADS] [Google Scholar]
  48. Paladini, R., Montier, L., Giard, M., et al. 2007, A&A, 465, 839 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  49. Paradis, D., Bernard, J., & Mény, C. 2009, A&A, 506, 745 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  50. Perrot, C. A., & Grenier, I. A. 2003, A&A, 404, 519 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  51. Planck Collaboration 2011a, A&A, 536, A1 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  52. Planck Collaboration 2011b, A&A, 536, A2 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  53. Planck Collaboration 2011c, A&A, 536, A7 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  54. Planck Collaboration 2011d, A&A, 536, A8 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  55. Planck Collaboration 2011e, A&A, Vol., A9 [Google Scholar]
  56. Planck Collaboration 2011f, A&A, 536, A10 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  57. Planck Collaboration 2011g, A&A, 536, A11 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  58. Planck Collaboration 2011h, A&A, 536, A12 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  59. Planck Collaboration 2011i, A&A, 536, A13 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  60. Planck Collaboration 2011j, A&A, 536, A14 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  61. Planck Collaboration 2011k, A&A, 536, A15 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  62. Planck Collaboration 2011l, A&A, 536, A16 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  63. Planck Collaboration 2011m, A&A, 536, A17 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  64. Planck Collaboration 2011n, A&A, 536, A18 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  65. Planck Collaboration 2011o, A&A, 536, A19 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  66. Planck Collaboration 2011p, A&A, 536, A20 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  67. Planck Collaboration 2011q, A&A, 536, A21 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  68. Planck Collaboration 2011r, A&A, 536, A22 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  69. Planck Collaboration 2011s, A&A, 536, A23 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  70. Planck Collaboration 2011t, A&A, 536, A24 [Google Scholar]
  71. Planck Collaboration 2011u, A&A, 536, A25 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  72. Planck Collaboration 2011v, The Explanatory Supplement to the Planck Early Release Compact Source Catalogue (ESA) [Google Scholar]
  73. Planck Collaboration 2011w, A&A, 536, A26 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  74. Planck HFI Core Team 2011a, A&A, 536, A4 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  75. Planck HFI Core Team 2011b, A&A, 536, A6 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  76. Reach, W. T., Koo, B., & Heiles, C. 1994, ApJ, 429, 672 [NASA ADS] [CrossRef] [Google Scholar]
  77. Reach, W. T., Dwek, E., Fixsen, D. J., et al. 1995, ApJ, 451, 188 [NASA ADS] [CrossRef] [Google Scholar]
  78. Reach, W. T., Wall, W. F., & Odegard, N. 1998, ApJ, 507, 507 [NASA ADS] [CrossRef] [Google Scholar]
  79. Roman-Duval, J., Israel, F. P., Bolatto, A., et al. 2010, A&A, 518, L74 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  80. Savage, B. D., Bohlin, R. C., Drake, J. F., & Budich, W. 1977, ApJ, 216, 291 [NASA ADS] [CrossRef] [Google Scholar]
  81. Sodroski, T. J., Bennett, C., Boggess, N., et al. 1994, ApJ, 428, 638 [NASA ADS] [CrossRef] [Google Scholar]
  82. Solomon, P. M., Rivolo, A. R., Barrett, J., & Yahil, A. 1987, ApJ, 319, 730 [NASA ADS] [CrossRef] [Google Scholar]
  83. Stepnik, B., Abergel, A., Bernard, J., et al. 2003, A&A, 398, 551 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  84. Velusamy, T., Langer, W. D., Pineda, J. L., et al. 2010, A&A, 521, L18 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  85. Wakker, B. P. 2006, ApJS, 163, 282 [NASA ADS] [CrossRef] [Google Scholar]
  86. Wolfire, M. G., Hollenbach, D., & McKee, C. F. 2010, ApJ, 716, 1191 [NASA ADS] [CrossRef] [Google Scholar]
  87. Zacchei, A., Maino, D., Baccigalupi, C., et al. 2011, A&A, 536, A5 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]

All Tables

Table 1

Characteristics of the data used in this study.

In the text
Table 2

Thermal dust emissivity derived from the correlation with HI emission in the reference region with |bII| > 20° and ((Iν/NH)ref).

In the text
Table 3

Derived parameters for the dark gas, computed in the region with available CO data and |bII| > 10°.

In the text

All Figures

thumbnail Fig.1

Map of the 12CO(J = 1 → 0) integrated intensity used in this paper combining the Dame et al. (2001) and high latitude survey and the NANTEN survey. The data shown cover 62.8% of the sky. The map is shown at a common resolution of all the sub-surveys of 8.8′. Many small clouds at high latitude are not visible in this rendering of the data.
In the text
thumbnail Fig.2

Upper panel: thermal dust emissivity (Iν/NH)ref from Table2. The dot curve shows a modified black-body at TD = 17.5K and β = 1.8 normalized at 857 GHz, for comparison. The various colours are for different instruments:IRAS(Yellow),DIRBE(light blue), Planck- HFI (red),WMAP(dark blue) and Planck- LFI(green). Lower panel: Offsets from Table2. The error bars are plotted at ± 3σ.

In the text
thumbnail Fig.3

Upper panel: all sky map of the dust temperature in K. The temperature is derived from modeling theIRIS100μm and the Planck-HFI emission at 857 and 545 GHz. Lower panel: all sky map of the dust temperature uncertainty in %. The maps are shown in Galactic coordinates with the Galactic centre at the centre of the image. Grey regions correspond to missingIRASdata.
In the text
thumbnail Fig.4

Details of the dust temperature (left column), dust optical depth at 857 GHz (central column) and dark gas column density (right column) for the Chamaeleon (first line), Aquila-Ophiuchus flare (second line), Polaris flare (third line) and Taurus (fourth line). The temperature and optical depth maps are shown in log scale with a colour scale ranging from 15K (black) to 20K (red) and 1 × 10-5 (black) to 3 × 10-3 (red) respectively. The dark gas column density derived from the optical depth at 857 GHz (see Sect.4) and is shown in linear scale with a colour scale ranging from –3 (black) to 7 × 1021Hcm-2 (red). The contours show the 12CO(J = 1 → 0) integrated intensity at 2, 10 and 20K km s-1. The double line shows the limit of the CO surveys.
In the text
thumbnail Fig.5

Maps of the dust optical depths on a log scale, in theIRAS100μm (first row left) and Planck- HFI bands at (first row right), (second row left), (second row right), (third row left), (third row right) and GHz (fourth row). All maps are shown in Galactic coordinates with the Galactic centre at the centre of the image. The missing data in all images correspond to theIRASgaps. The upper and lower bounds of the colour scale are set to τmin = 5 × 10-5 × (λ/100μm)-1.8 and τmax = 10-2 × (λ/100μm)-1.8 respectively.
In the text
thumbnail Fig. 6

Correlation plots between the dust optical depth atIRAS100μm (upper left),HFI857 GHz (upper right), 545 GHz (lower left) and 353 GHz (lower right) and the total gas column density in the solar neighbourhood (|bII| > 10°). The color scale represents the density of sky pixels on a log scale. The blue dots show a -binned average representation of the correlation. The red line shows the best linear correlation derived at low values (). The vertical lines show the positions corresponding to AV = 0.37mag and AV = 2.5mag. These figures are shown for a single XCO = 2.3 × 1020H2 cm-2/(K km s-1).

In the text
thumbnail Fig. 7

Dust optical depth derived from this study using theIRASand Planck- HFI frequencies. The square symbol shows the emissivity at 250μm derived by Boulanger et al. (1996). The dash and dash-dot lines show a power law emissivity with λ-1.8 and λ-1.75 respectively, normalized to the data at 100μm. The error bars shown are ± 1σ.
In the text
thumbnail Fig.8

Map of the excess column density derived from the 857 GHz data. The map is shown in Galactic coordinates with the Galactic centre at the centre of the image. The grey regions correspond to those where noIRASor CO data are available, regions with intense CO emission (WCO > 1K km s-1) and the Galactic plane (|bII| < 5°).
In the text
thumbnail Fig. 9

Fractional mass of the dark gas with respect to the neutral gas mass as a function of the lower bIIvalue used in the analysis. The solid curve is computed under the assumption of optically thin Hi, the dashed curve is for computed using Ts = 80K. Error bars are 1σ.
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.