Publications of the Korean Astronomical Society pISSN: 1225-153430: 075 ∼ 078, 2015 September eISSN: 2287-6936c©2015. The Korean Astronomical Society. All rights reserved. http://dx.doi.org/10.5303/PKAS.2015.30.2.075

QUANTIFYING DARK GAS

Di Li1,2, Duo Xu1, Carl Heiles3, Zhichen Pan1, 2, and Ningyu Tang1

1National Astronomical Observatories, Chinese Academy of Sciences, A20 Datun Road, Chaoyang District, Beijin 100012,China; dili@nao.cas.cn , xuduo117@bao.ac.cn

2Key Laboratory for Radio Astronomy, Chinese Academy of Sciences, Nanjing 210008, China3Astronomy Department, University of California, Berkeley, CA 94720-3411

E-mail: dili@nao.cas.cn , xuduo117@bao.ac.cn

(Received November 30, 2014; Reviced May 31, 2015; Aaccepted June 30, 2015)

ABSTRACT

A growing body of evidence has been supporting the existence of so-called “dark molecular gas” (DMG),which is invisible in the most common tracer of molecular gas, i.e., CO rotational emission. DMG is be-lieved to be the main gas component of the intermediate extinction region from Av∼0.05–2, roughlycorresponding to the self-shielding threshold of H2 and 13CO. To quantify DMG relative to HI and CO,we are pursuing three observational techniques; HI self-absorption, OH absorption, and THz C+ emission.In this paper, we focus on preliminary results from a CO and OH absorption survey of DMG candidates.Our analysis shows that the OH excitation temperature is close to that of the Galactic continuum back-ground and that OH is a good DMG tracer co-existing with molecular hydrogen in regions without CO.Through systematic “absorption mapping” by the Square Kilometer Array (SKA) and ALMA, we willhave unprecedented, comprehensive knowledge of the ISM components including DMG in terms of theirtemperature and density, which will impact our understanding of galaxy evolution and star formationprofoundly.

Key words: ISM: dark gas; molecular gas; atomic gas; radio line

1. INTRODUCTION

Two relatively dense phases of the interstellar medium(ISM) are the atomic Cold Neutral Medium (CNM)traced by the the HI λ21cm hyperfine structure line andthe ‘standard’ molecular clouds (H2) as traced by CO.CO is the most important tracer of molecular hydrogen,which remains largely invisible due to lack of emissionin the temperature range of molecular ISM. Empirically,CO intensities have been used as an indicator of the to-tal molecular mass in the Milky Way and in galaxiesthrough the so-called “X-factor” with numerous well-known caveats. Gases in these two phases dominate themasses of star forming clouds on a galactic scale. Themeasured ISM gas mass from HI and CO is the foun-dation of many key quantities in understanding galaxyevolution and star formation, such as the star formationefficiency.

A growing body of evidence, however, indicates theexistence of gas traced by neither HI nor CO. Com-parative studies (e.g. de Vries et al. 1987) of InfraredAstronomy Satellite (IRAS) dust images and HI andCO gas maps revealed an apparent excess of dust emis-sion. The Planck collaboration (2011) clearly show ex-cess dust opacity (Fig.1) in the intermediate extinctionrange Av∼ 0.05–2, roughly corresponding to the self-

http://pkas.kas.org

shielding threshold of H2 and 13CO. The missing gas, orrather, the undetected gas component is widely referredto as dark gas, popularized as a common term by Gre-nier et al. (2005). They found more diffuse gamma-rayemission observed by the Energetic Gamma Ray Exper-iment Telescope (EGRET) than what can be explainedby cosmic-ray H-nuclei interaction (H+X-factor*CO).Observations of the THz fine structure C+ line also helpreveal dark gas as the C+ line strength in diffuse gas isstronger than what can be produced by collisional exci-tation with only HI gas (Langer et al. 2010).

A minority of the ISM community argued that darkgas could be explained by an underestimation of HIopacities (Fukui et al., 2014), which is in contrast withother recent works (Stanimirovic et al., 2014). Due tothe limited scope of this paper, we will only discuss thedark molecular gas (DMG) hereafter, or more specifi-cally CO-dark molecular gas.

It is natural to infer from chemical and PDR modelsthat molecular hydrogen would exist in regions whereCO is not detectable. CO can be of low abundancedue to photo-dissociation (Fig.2) in unshielded regionsand/or can be heavily sub-thermal due to lack of colli-sions in diffuse gas. We strive to provide direct measure-ments and/or constraints of the physical conditions ofDMG. Section 2 will focus on OH absorption. Section 3introduces a CO survey toward background continuum

75

76 DI LI

Figure 1. This figure is adopted from Figure 6 in Planck(2011). Correlation plots between the gas column densityas traced by [HI+XCO*CO] and dust optical depth at IRAS100 µm (upper left), HFI 857 GHz (upper right), 545 GHz(lower left) and 353 GHz (lower right). The color scale rep-resents the density of sky pixels on a log scale. The blue dotsshow a binned average representation of the correlation. Thered line shows the best linear correlation derived at low val-ues. The vertical lines show the positions corresponding toAv = 0.37 mag and Av = 2.5 mag. A single CO X-factorXCO = 2.3 ×1020 H2 cm−2/(K km s−1) was used.

sources. Section 4 presents the combined analysis of COand OH followed by a brief outlook of upcoming surveysin the final section.

2. WHERE IS THE HYDROXYL?

OH, or Hydroxyl, was the first interstellar molecule de-tected in radio bands (Weinreb et al. 1963). It can formquickly through a series of charge exchange reactions ini-tiated by cosmic rays once H2 is present (van Dishoeck& Black 1988). OH can also form on grains. One ofthe main chemical paths associated with CO after OHformation is

OH + C+ → CO+ + H , (1)

CO+ + H2 → HCO+ + H , (2)

HCO+ + e− → CO + H . (3)

We should expect wide-spread and abundant OHalong with HCO+ and C+. HCO+ is accessible in mil-limeter bands. The main transition from C+ is its finestructure transition in the 2 THz band, which is impos-sible to map from the ground. It is somewhat puzzlingwhy large scale OH surveys of ISM have not been avail-able in the half a century since its discovery. In fact,thousands of hours of Arecibo time have been spent onsearches for OH in galaxies with mostly negative results(e.g. Schmelz & Baan, 1988).

In contrast, Dickey et al. (1981) found OH in absorp-tion against high Galactic latitude continuum sources.Important and extensive confirming absorption mea-surements by Liszt & Lucas (1996) and Lucas & Liszt(1996) found that OH and HCO+ are commonly ob-served against such sources. Lucas & Liszt (1996) found

Figure 2. A schematic view of photo-dissociation regions (Tie-lens 2005), showing the locations of different transition lay-ers. We add the blue pane to indicate the possible locationof OH.

that ∼30% of continuum sources having HI in absorp-tion exhibit HCO+ in absorption. In light these results,the dearth of OH emission should be attributed to theexcitation condition of OH rather than its abundance.

The observed antenna temperature TA is

TA = (Tex − TC)(1− e−τ ) , (4)

where TC ∼ 3.5 K is the continuum background tem-perature at L-band, composed of the CMB and galacticsynchrotron emission. When Tex approaches TC , the ap-parent signal from certain line emission vanishes. Suchgas, however, is suitable for absorption studies whenthe telescope is trained toward background sources withTC � Tex as is the case when observing quasars and/orHII regions.

Heiles & Troland (2003) published the well-cited Mil-lennium survey of 21-cm line absorption toward 79 con-tinuum sources. The ON-OFF technique and Gaussiandecomposition analysis allow them to provide crediblemeasurements of the excitation temperature and den-sity of HI components spreading through the Milky Way.Unpublished OH absorption data were taken simulta-neously with the Millennium survey. Our preliminaryanalysis of these OH absorption data confirms the suspi-cion that Tex of OH aggregates is comparable to the to-tal continuum background temperature (CBR + Galac-tic), and thus renders OH undetectable in emission.

3. A MULTI-TRANSITION CO SURVEY OF MILLEN-NIUM SOURCES

We conducted a follow-up CO survey of all Millenniumsight-lines with OH absorption. Toward 79 publishedMillennium survey sources, 43 sight-lines exhibit OHabsorption. These 43 sources with OH absorption havebeen observed in 12CO J=1-0 and 2-1, 13CO J=1-0 andC18O J=1-0. The 8 sources with strong 12CO J=1-0 and2-1 lines were also observed in the 12CO J=3-2 transi-tion.

The J=1-0 transition of 12CO, 13CO and C18O wereobserved in March and May of 2013 and May of

QUANTIFYING DARK GAS 77

Figure 3. The location of sources in galactic coordinates. Tri-angles represent sources without any absorption componentin OH. Squares represent sources with absorption compo-nent in OH. Yellow dots represent sources without any detec-tion of CO transitions. Red dots represent sources with COtransition components; these CO components correspond toall the absorption components in OH. Black dots representsources with CO transition components, and these CO com-ponents correspond to some of the absorption components inOH, but there are no detectable CO transition correspond-ing to the rest of the absorption components in OH. We callthese kind of sources “partial CO detection”.

2014 with the Purple Mountain Observatory Delingha(PMODLH) 13.7 m telescope, of the Chinese Academyof Sciences. The spectra were obtained in positionswitch mode and the reference position is selected fromIRAS Sky Survey Atlas1. With 1000 MHZ bandwidthspectroscopy, the frequency resolution is 61 kHz, result-ing in an approximately 0.18 kms−1channel width.

The 12CO J = 2-1 and J = 3-2 data were taken withthe Caltech Submillimeter Observatory (CSO) 10.4m ontop of Mauna Kea in July, October and December of2013. The velocity resolution for 12CO(J = 2-1) spectrais 0.16 kms−1. The velocity resolution for 12CO(J = 3-2) spectra is 0.11 kms−1or 0.56 kms−1due to a problemin spectroscopy.

The distribution of these spectra in the galactic coor-dinates is shown in Fig. 3.

The typical RMS of the 12CO J=1-0 observation isabout 0.06K, which corresponds to a CO detection limitof 2.6×1013 cm−2.

The astronomical software package Gildas/CLASS2

was used for baseline removal, combining spectra, andGaussian fitting.

4. COMPARISON OF HI, OH, AND CO

We compare the Gaussian components seen in HI ab-sorption, OH absorption, and CO emission. A total of115 Gaussian components were detected as specified inHeiles & Troland (2003). 52 of these gas componentshave OH absorption. The majority of these 52 have COemission, except for 13 components, which are DMGcandidates. There are no components with only COemission and no OH absorption. Three representative

1http : //irsa.ipac.caltech.edu/data/ISSA/2http : //www.iram.fr/IRAMFR/GILDAS/

CNM Dark Gas Molecular Cloud

Figure 4. Representative spectra. The 3C192 sightline hasonly HI seen in absorption. One component of the 3C154sightline has OH and HI in absorption and CO and its iso-topologues in emission. The 3C409 sightline has one gascomponent with both OH and HI , but no detectable COtransitions.

sets of spectra are shown in Fig. 4. Toward 3C192, onlyHI is present, typical of CNM. Toward 3C154, there isa component with HI, OH, and several CO and CO iso-topologue transitions, which should be representative of‘normal’ molecular clouds. Toward 3C409, there existsa component with HI and OH absorption, but no COemission in any transition observed. We posit that thisis typical of DMG. The percentage of these three cate-gories are 55% CNM, 34% molecular clouds, and 11.3%DMG.

The column densities of the OH components were cal-culated as

NOH =8πkTexν

21667

A1667c3h

16

5

∫τ1667 dv , (5)

where A1667 = 7.778×10−11s−1 is the A-coefficient andTex is its excitation temperature calculated based on arecipe similar to that for the HI absorption componentsin Heiles & Troland (2003).

The CO column densities were calculated for two cat-egories. If only the J=1-0 transition of 12CO was de-tected, the optical depth is assumed to be small andthe excitation temperature is assumed to be the sameas that of OH. If both 12CO and 13CO were detected,we derive the optical depth and the excitation tempera-ture based on multiple transitions and Local Thermody-namic Equilibrium (LTE) assumptions. The recipe forderiving CO column densities can be found in Li (2002).

The statistics of the gas column densities (Fig. 5) isconsistent with the schematic picture presented in Fig. 3and section 1. There is an apparent gas column densitythreshold for OH detection at around Av∼0.05, abovewhich OH and CO have similar distributions. OH turnsout be a good tracer of diffuse gas with ‘intermediate’extinction, namely, between the self-shielding thresholdfor H2 and 13CO.

5. DISCUSSION

The expected location, abundance, and optical depth ofOH should make it an excellent tracer of DMG. Due

78 DI LI

Figure 5. The histogram of HI colunm density for the coldneutral medium (the blue filled rectangles), molecular gas(the green line and filled rectangles) and dark gas (the redline and filled rectangles).

to insufficient collisions in diffuse gas, however, OH ishard to detect in emission. This is likely the main rea-son why a galactic scale or even any large-scale OH maphas not been accomplished. To realize its potential inquantifying dark gas throughout the ISM, upcoming ra-dio telescopes will be needed to conduct comprehensiveabsorption surveys. The Five-hundred-meter ApertureSpherical radio Telescope (FAST) is expected to startoperation in late 2016. The unprecedented sensitivityof FAST and its early science instruments (Li et al.2013) should make feasible a HI+OH absorption survey,in the mode of the Millennium survey, but with 10 timesmore sources. The SKA1 will have the survey speedand sensitivity to measure gas absorption with a sourcedensity between a few to a few tens per square degree(McClure-Griffiths et al., 2014), which means that an allsky “absorption-image” is feasible and we will have ISMtemperatures and densities everywhere! Based on simi-lar excitation and sensitivity considerations, ALMA is apowerful instrument for obtaining systematic and sensi-tive absorption measurements of millimeter lines in dif-fuse gas. CO and HCO+ in diffuse gas, in particular, willbe much better constrained in terms of excitation tem-perature and column densities through ALMA absorp-tion observations than emission measurements. Com-bining both radio and millimeter absorption surveys inthe coming decade, we will quantify DMG and providedefinitive answers to questions like the global star for-mation efficiency.

This work is supported in part by National Ba-sic Research Program of China (973 program) No.2012CB821800, NSFC No. 11373038, and ChineseAcademy of Sciences Grant No. XDB09000000. Wethank Lei Qian for his great help in running 12COJ = 3 - 2 observation at tge CSO site. We alsothank Lei Zhu for his help in CSO observations. The3×3 multibeam sideband separation superconductor-insulator-superconductor (SIS) receiver with sidebandseparation (Zuo et al., 2011; Shan et al., 2012) was usedfor observation. The 12CO ,13CO and C18O spectra was

obtained with position switch mode, the reference posi-tion is selected from IRAS Sky Survey Atlas3. The au-thors appreciate all the staff members of the PMODLHfor their help during the observation. This material ispartly based upon work at the Caltech SubmillimeterObservatory, which is operated by the California Insti-tute of Technology.

REFERENCES

Dickey, J. M., Crovisier, J., & Kazes, I., 1981, Emission-Absorption Observations of OH in Diffuse InterstellarClouds, 98, 271

Fukui, Y., Torii, K., Onishi, T., et al., 2014, Optically ThickH I Dominant in the Local Interstellar Medium: An Al-ternative Interpretation to ”Dark Gas”, The AstrophysicalJournal, 798, arXiv:1403.0999

Grenier, I. A., Casandjian, J.-M., & Terrier, R., 2005, Un-veiling Extensive Clouds of Dark Gas in the Solar Neigh-borhood, Science, 307, 1292

Heiles, C., & Troland, T. H., 2003, VizieR Online Data Cat-alog: Millennium Arecibo 21-cm Survey (Heiles+, 2003),The Astrophysical Journal Supplement Series, 145, 329

Langer, W. D., Velusamy, T., & Pineda, J. L., et al., 2010,C+ Detection of Warm Dark Gas in Diffuse Clouds, A&A,521, LL17

Li, D., 2002, Massive Cores in the Orion Molecular Cloud,Thesis (PhD). CORNELL UNIVERSITY, Source DAI-B63/04, 201

Li, D., Nan, R., & Pan, Z., 2013, The Five-Hundred-MeterAperture Spherical Radio Telescope Project and Its EarlyScience Opportunities, IAU Symposium, 291, 325

Liszt, H. & Lucas, R., 1996, Galactic OH Absorption andEmission toward a Sample of Compact Extragalactic MM-Wave Continuum Sources, A&A, 314, 917

Lucas, R., & Liszt, H., 1996, The Plateau de Bure Survey ofGalactic 3mm HCO+ Absorption toward cCmpact Extra-galactic Continuum Sources, A&A, 307, 237

McClure-Griffiths, N. M., 2014, in prep. for Proceedings ofScience “Advancing Astrophysics with the Square Kilome-tre Array”

Planck Collaboration, Ade, P. A. R., & Aghanim, N., etal., 2011, Planck Early Results. XIX. All-Sky Temper-ature and Dust Optical Depth from Planck and IRAS.Constraints on the ”Dark Gas” in our Galaxy, A&A, 536,AA19

Schmelz, J. T. & Baan, W. A., 1988, A Search for ThermalHydroxyl Emission in Nearby Galaxies, AJ, 95, 672

Stanimirovic, S., Murray, C. E., Lee, M.-Y., Heiles, C., &Miller, J., 2014, Cold and Warm Atomic Gas around thePerseus Molecular Cloud. I. Basic Properties, AJ, 793, 132

Tielens, A. G. G. M., 2005, The Physics and Chemistry ofthe Interstellar Medium, Cambridge University Press.

de Vries, H. W., Heithausen, A., & Thaddeus, P., 1987,Molecular and Atomic Clouds Associated with InfraredCirrus in Ursa Major, Astrophysical Journal, 319, 723

Weinreb, S., Barrett, A. H., Meeks, M. L., & Henry,J. C., 1963, Radio Observations of OH in the Interstel-lar Medium, Nature, 200, 829

3http : //irsa.ipac.caltech.edu/data/ISSA/