Surveys for High-Latitude Molecular Clouds

Abstract

Searches for high latitude molecular gas have followed different strategies and been subjects of different systematics and biases. To show how these affect the emergent physical picture, we describe in some detail a few surveys and overview the development of the searches in the millimeter wave tracers.

Surveys aren’t the most important thing in astronomy - they’re the only thing. Jerry Ostriker, with apologies to Vince Lombardi and Red Sanders.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

9.1 Introduction

The vast majority of molecular clouds are located below | b |  = 15^∘. Given the thinness of the disk, when one sees higher clouds, they are likely nearby (e.g., the Taurus/Aurigae dark clouds ( ∼ 140 pc), the ρ Ophiuchi clouds ( ∼ 130 pc), and even the Orion GMC at ∼ 0.5 kpc is fairly close). Throughout the 1970s and early 1980s, searches for molecular gas at high latitudes were deemed a waste of time because little molecular gas was expected there, but by the mid 1980s, a population of small molecular clouds numbering less than 200 was identified at | b | ≥ 20^∘. Of these high-latitude molecular clouds, 57 were cataloged by Magnani et al. (1985) and became known as the MBM clouds (numbered MBM 01–MBM 57).¹ It became clear after a few years that these objects are for the most part small, diffuse and translucent, molecular clouds. There are a few dark clouds in the mix that had been previously identified (e.g., Lynds 1642 = MBM 20), but most of the objects in the MBM catalog were new and constituted the largest population of diffuse and translucent clouds identified to that date by CO(1-0). Of course, there are plenty of diffuse and translucent molecular clouds at | b |  < 20^∘, but the long sightlines through the Galactic plane at low latitudes usually intersect more distant GMCs and dark cloud complexes so that any smaller foreground or background diffuse and translucent clouds are hard to pick out. The initial interest in these objects wasn’t so much their peculiar latitudes but, instead, they represented a different population than the usual dark clouds and GMCs.

Because the high-latitude clouds are close, anywhere from ∼ 100 pc to a few hundred pc away, they can be studied at the highest spatial resolution. The few high-latitude molecular clouds that are forming stars (see McGehee 2008 for a review) are among the nearest birthing stars to the Sun. However, their dynamical state is more important than their proximity: the vast majority of high-latitude clouds are not in virial equilibrium. Their kinetic energies are anywhere from a few to hundreds of times greater than their gravitational potential energies, thus, unless an exterior medium is confining them, they are breaking up on a sound crossing time. They seem to be condensing out of the atomic CNM and then rapidly dissipating. It was clear from the beginning that, unlike their denser brethren, these objects are structured primarily by turbulence rather than gravity and, as such, they are likely the best laboratory for studying the effects of turbulence in the diffuse ISM.

There have been many searches for high-latitude molecular gas, but not all of them are familiar to researchers in other fields. Many claims about the distribution and amount of molecular gas in the last 10 years have ignored previous work. So, in this chapter, we will describe the major searches and surveys and their biases and limitations. We will use chronological order only to set in context the evolution of the sample selection criteria and of the methods employed for the searches.

9.2 The Initial Searches

The large scale CO surveys of the Galaxy Sect. 7.2.1 all concerned the midplane distribution of GMCs. Despite the marked concentration of molecular gas towards the Galactic midplane, the presence of CO emission in Fig. 7.1 away from the midplane raises the question of the existence of gas outside the latitude boundaries of the map (approximately ± 35^∘), the so-called “high” Galactic latitudes. The demarcation line where “high” begins depends very much on the particular observer; values ranging from | b | = 10^∘ to 30^∘ have been used at various times. In this book we will define high latitude to begin at | b |  = 25^∘. The reason for this, admittedly subjective, choice is to avoid including in the high-latitude molecular gas inventory the southern extension of the Taurus/Aurigae dark clouds below the Galactic plane and the northern extension of the ρ Ophiuchi dark clouds above the plane. With this definition, virtually all molecular radio astronomers in the 1970s were of the opinion that regions with | b | ≥ 25^∘ were so deficient in molecular gas as to make surveys not worthwhile, despite the fact that from the very first observations of the CO(1-0) line it was clear that some molecular gas was present there.

Table 9.1 lists the 19 Lynds dark clouds known before 1983 to be located at high Galactic latitudes. Dickman (1975) had detected CO emission from four of them: L169, L183, L1642, and L1778. The first detection of a high-latitude molecular cloud was reported by Penzias et al. (1972) who described CO observations of 12 dark clouds including L134 at b = 35.8^∘. But in addition to the CO line detections in a few Lynds dark clouds at | b | ≥ 25^∘, there were at least five other early indications that appreciable quantities of high-latitude molecular gas might exist: (1) the bright nebulae noted by Lynds (1965); (2) the absorption observations of the 18 cm OH lines and the 21 cm H I line in the direction of extragalactic continuum sources (Dickey et al. 1981; Kazès and Crovisier 1981); (3) the CO(1-0) emission detected by Knapp and Jura (1976) along lines of sight previously studied only via optical absorption lines; (4) the Blitz et al. (1982) catalog of CO emission from Sharpless H II regions; and (5) the two populations of quasars present in a study of color excess of quasars compared to the atomic hydrogen column density in their directions (Teerikorpi 1981). We will examine each of these indicators in turn.

Table 9.1 Lynds dark clouds at | b | ≥ 25^∘

We discussed the Lynds Bright Nebulae (LBN) in Sect. 6.4 There are more of these at high Galactic latitude than there are Lynds dark nebulae. Because these clouds are rich in dust, they likely contain some molecular gas, but this was just a suspicion.

Dickey et al. (1981) searched for OH absorption lines at 18 cm in the direction of 58 extragalactic sources and detected OH absorption and/or emission in 16 directions, including two at b = −38.2^∘ and −33.6^∘. Similarly, Kazès and Crovisier (1981) searched for CO(1-0) emission in 76 lines of sight showing strong H I 21 cm absorption. CO(1-0) was detected from 21 directions including four at | b | ≥ 25^∘ (two of the four detections were the same as those in the Dickey, Crovisier, and Kazès sample). The molecular gas along these four high-latitude lines of sight not coinciding with any Lynds dark clouds was thought to arise from diffuse, low column density clouds. To our knowledge, these clouds still have not been mapped although two of the lines of sight are almost surely part of the MBM 53–55 high-latitude cloud complex. It is important to note that, at that time, the astronomy community which studied optical and ultraviolet interstellar absorption lines included few radio astronomers, so it is not surprising that the detections by Crovisier, Dickey, and Kazès were not followed up.

A few years earlier, Knapp and Jura (1976) detected CO(1-0) emission from interstellar clouds along lines of sight to stars which showed CO absorption in the ultraviolet. Although some of the detections were questioned by Lada and Blitz (1988), none had | b | ≥ 24^∘. Similarly, a study by Dickman et al. (1983) included lines of sight to other stars known to exhibit CO absorption lines, but none of those were above | b |  = 21^∘. However, the fact that several intermediate-latitude sight lines outside of known Lynds dark clouds had sufficiently high CO column densities to be detected by radio techniques should have provoked more interest in this type of low-extinction, molecule-rich, interstellar clouds. It did not.

Three clear detections of CO(1-0) emission at high-latitudes from objects other than Lynds dark clouds were made by Blitz et al. (1982) during the compilation of their catalog of CO radial velocities toward Galactic H II regions, chosen from the Sharpless (1959) catalog. Of the more than 300 regions observed, six objects at | b | ≥ 25^∘ showed CO emission. Despite the fact that these objects had been cataloged by Sharpless because they are associated with red nebulosity on the POSS plates, none of the six detections is actually an H II region. The red emission is produced either by reflected starlight or photoluminescence of carbon grains (see Sect. 3.5.4). Fich et al. (1990) detect Hα emission from most of these objects, but the profiles are wide in velocity and are detections of the WIM, not of high-latitude H II regions.

Finally, using a completely different line of reasoning, Teerikorpi (1981) showed that a substantial number of molecular clouds has to exist at high latitudes. His novel method consisted of examining residuals in the (B-V) magnitude of quasars compared to their unreddened color derived from theoretical considerations (in other words a type of quasar color excess: [(B-V) − (B-V) _Zo ] where (B-V) _Zo is the unreddened quasar color). A plot of the residuals of B-V versus N(H I) showed a two-branch structure in the quasar population, with the quasars in the directions of the local spiral arms showing a greater residual than quasars in the direction away from spiral arms. Teerikorpi concluded that the more reddened branch of the quasar population is caused by greater absorption in the local Galactic spiral arm. If the gas to dust ratio is constant, this extra absorption can not be produced by H I associated dust and must come from dust associated with molecular hydrogen. He obtained for these 2N(H₂)/N(H _total ) ∼ 1, A _V  ≥ 0.5 magnitudes, linear size less than 2 pc if n > 50 cm⁻³, R _V  ≈ 3.3 for the dust component, and a surface filling factor of 40% of the sight lines at moderate latitudes. This intriguing argument was completely ignored by the molecular cloud observing community.

These detections and lines of reasoning pointed to some molecular gas residing at high latitudes, but the only actual published search or survey for CO emission in that territory came up empty (Dewey et al. 1983). They searched for CO(1-0) emission along 68 lines of sight around b ∼ 59^∘ and b ∼ 84^∘. This area had been chosen based on its high reddening over several square degrees determined from the Shane and WIrtanen (1967) galaxy counts. It is likely, however, that many of the early CO observers conducted high-latitude CO searches while they were at the telescope surveying the Galactic plane. Yet, with the exception of the Dewey et al. survey, the results were not published. Evidently, the repeated failure to detect high-latitude CO emission spread by word-of-mouth and added to the lore that such gas was not to be found.

A successful search for high-latitude CO emission was conducted by Michel Fich and Leo Blitz in January of 1982 yielding detections toward the North Galactic Pole (M. Fich, private communication), but the exact coordinates were forgotten and the results were never published. This success, however, inspired the Blitz, Magnani, and Mundy high-latitude survey conducted in 1983 and published in 1984.

Thus, by the summer of 1983, the inventory of molecular gas with sufficient column density to provide a CO(1-0) signal at | b | ≥ 25^∘ included 7 separate Lynds (1962) dark cloud complexes (comprising the 19 Lynds dark clouds in Table 9.1) with visual extinctions in their core regions in all cases less than one magnitude, four lines of sight to diffuse, unmapped interstellar clouds, and three lines of sight toward regions of red nebulosity on the POSS plates. Despite a few additional, indirect, indications to the contrary, there appeared to be very little dust and extinction at high Galactic latitudes. The IRAS mission (Low et al. 1984) changed this perception with the discovery of the infrared cirrus primarily at 60 and 100 μm (see Sect. 5.3.1).

9.3 The MBM Survey

The 1970s and 1980s were the watershed period in Galactic molecular astronomy and surveys were an essential part of the development. The probing of the high latitude sky was seen as a risky undertaking, one that would likely come up empty, and for that reason was not a high priority for most researchers. In this section, where some of the description will be necessarily based on personal experiences, we are neither nostalgically reminiscing nor indulging in vapid autobiography. By describing the methods used to discover high-latitude clouds we can illustrate how current data sets, especially in the optical and near infrared, can be used to find new objects, and we can dispel the common misconception that the clouds were found by surveying the IRAS cirrus. More important, the methods carry implicit biases that can skew the results if not examined more closely. Thus, concurrently with, but independent of the IRAS mission, Leo Blitz at the University of Maryland initiated a project to systematically search for high-latitude molecular gas. As outlined above, the Blitz et al. (1982) survey and unpublished observations by Blitz and Fich had hinted at the presence of undiscovered molecular clouds at high Galactic latitudes.

Blitz assigned Loris Magnani, then a graduate student at the University of Maryland, the task of poring over prints of the POSS plates and the Whiteoak extension to the POSS to find all regions at | b | ≥ 20^∘ showing any sign of obscuration by dust. This was the traditional method for finding high dust column density regions which the earliest molecular surveys had shown to often contain molecular gas. The choice | b | ≥ 20^∘ was originally made because at these latitudes there did not appear to be significant molecular gas associated with lower latitude clouds. The northern and southern extension of the Taurus-Aurigae and Ophiuchus dark cloud complex later forced the “boundary” to be shifted to be | b | ≥ 25^∘, where only the Polaris Flare shows a possible connection to lower latitudes.

Each red and blue print of the POSS plates (6^∘× 6^∘) was examined for regions of size 5^′ or or larger showing: (1) a noticeable drop in stellar density, (2) any emission nebulosity with a sharp boundary, possibly indicating the edge of a molecular cloud (the emission is produced primarily by reflecting the integrated starlight of the Galactic plane (however, some of the emission could be due to red luminescence Sect. 3.5.4), or (3) any change in the stellar density from the blue to the red print possibly indicating reddening produced by dust. All obvious regions of obscuration (i.e., those with A _V greater than a few magnitudes) had been found by Lynds (1962); thus, Magnani had to effectively identify those regions which had obscuration equivalent to ∼ 1 magnitude of visual extinction or less. At | b | ≥ 50^∘, candidate regions become especially difficult to identify because of the significant drop in stellar density and the ensuing relatively larger stellar surface density fluctuations. The lack of blue prints in the Whiteoak extension to the POSS also complicated the search in the range −44^∘ < δ < −37^∘. In many instances, candidate regions extended over a large area of a print, or several regions were in close proximity so that several search positions for a given candidate region had to be included. A total of 458 candidate regions were selected and 439 were searched for CO(1-0) emission during November and December 1983 using the University of Texas Millimeter Wave Observatory’s (MWO) 5-m radiotelescope (see Vanden Bout et al. 2012). The initial observations were made in frequency-switched mode and the presence of the telluric CO(1-0) line in every spectrum caused considerable confusion at the beginning of the run. However, by the end of the run, it was clear that the number of detections was much greater than had been anticipated. Emission had been detected from 105 of the 439 observed candidate positions—a detection rate of 24%.

The initial results of the first survey were published in a letter in The Astrophysical Journal (Blitz et al. 1984). The detections were divided into objects in the 20^∘ ≤ | b |  < 30^∘ range and those with | b | ≥ 30^∘. Of the clouds at | b | ≥ 30^∘, CO emission was detected from 37 of 287 observed positions (13%), with the detections grouped into 29 distinct clouds. Eleven clouds were mapped in a very undersampled pattern (typically 10^′–20^′ grids with a 2.3^′ beam). Some of the Lynds clouds in the sample had been partially mapped by others (e.g., L1642 by Sandell et al. 1981) and five clouds in the vicinity of and including L134 had been mapped in CO by Joe Montani and Mark Morris with the Columbia 1.2 meter mm-wave radiotelescope (now at Harvard), although the data were never published. With the Montani and Morris results, a total of 16 clouds at | b | ≥ 30^∘ had sufficient mapping data to permit Blitz et al. (1984) to characterize some of the basic CO and physical properties of these “high-latitude molecular clouds” (see Table 9.2). Blitz et al. (1984) also opined that these objects had to be associated with the IRAS cirrus clouds which had been identified concurrently by Low et al. (1984).

Table 9.2 Properties of high-latitude molecular clouds based on the initial CO survey by Blitz et al. (1984)

The distances to the clouds were obtained statistically using the velocity dispersion of the cloud ensemble and the local mass surface density (see Chap. 10). With a mean distance of 100 pc, the initial survey established that at least a portion of the infrared cirrus was local; a point made explicitly by Gautier (1986) and Weiland et al. (1986). With a distance estimate, many of the observed properties of the clouds could be converted to physical properties. It was thus clear that the high-latitude clouds are small, low-density, low-extinction clouds quite unlike GMCs or dark clouds.

Besides the discovery of substantial number of new molecular clouds, the Blitz et al. (1984) paper made a strong assertion: most of the high-latitude clouds are not gravitationally bound. Many high-latitude clouds have centroid velocity dispersions which imply virial masses larger by 1–2 orders of magnitude than the masses derived from the CO emissivity. Before Blitz et al. (1984) no one had explicitly stated in the literature that at least some molecular clouds are unbound. A few researchers (Mahoney et al. 1976; Clark et al. 1977) had noticed that the kinetic energy of some clouds appears to be substantially greater than their gravitational potential term, but this discrepancy was attributed to rotation or other kinematic effects (i.e., orbiting fragments). The large velocity dispersions in the handful of mapped clouds was the first clear indication that the majority of the high-latitude clouds could represent a population different from GMCs and dark clouds. At this point in time, the diffuse/translucent nature of the high-latitude clouds was not yet appreciated. They were assumed have molecular properties like dark clouds with somehow lower extinctions (attributed to different grain properties). Some of the more savvy astronomers referred to them as “CO-rich diffuse clouds”, but their nature was still uncertain.

Many of the detections are associated with regions of marked emission, especially on the POSS red prints. Magnani then went through the prints a second time and subsequently identified 35 additional regions of apparent obscuration, focussing on the presence of low-level blue and red emission. Many of the new candidates had previously been noted by Sharpless (1959) and Lynds (1965). The final search catalog constituted 493 candidate regions, with over 400 completely new lines of sight.

The second set of observing runs consisted of a few runs at the MWO during the first half of 1984. The observing setup was similar to that of the first run and is further described by Magnani et al. (1985). Of the 493 positions from the search catalog, 488 had been observed by June, 1984. Some of the five unobserved positions were eventually detected (they were not observed because the Texas 5 meter telescope could not point to the North Celestial Pole). CO(1-0) emission was found from a total of 133 positions. Mapping showed that some of these were from the same contiguous structure so 124 clouds were identified. At 20^∘ ≤ | b |  < 25^∘, 67 positions were detected (68% detection rate), and 57 (14%) at | b | ≥ 25^∘. The relatively high detection rate in the 20^∘ ≤ | b |  < 25^∘ region was largely due to the many candidate positions near or in the Taurus-Auriga and Ophiuchus cloud complexes and, finally, only those objects at | b | ≥ 25^∘ were called high-latitude molecular clouds. Tables 1 and 2 in the MBM paper list all the detections from the initial observing runs.

The distribution of the original detections in Galactic coordinates at | b | ≥ 20^∘ is shown in Fig. 9.1. The clustering of detections at b ∼ −20^∘ and ℓ ∼ 155^∘-180^∘ and b ∼ +20^∘ and ℓ ∼ 350^∘ to 5^∘ indicated that the sample was contaminated by portions of the Taurus/Perseus and Scorpius/Ophiuchus dark cloud systems. A north-south asymmetry (29%–71% at | b | ≥ 25^∘) was noticeable. The implications of this asymmetry for the location of the Sun with respect to the Galactic midplane will be discussed in Chap. 10.

Fig. 9.1

Distribution of CO(1-0) detections in Galactic coordinates and Aitoff projection obtained during the MBM high-latitude survey in 1983–4. The data were taken at the MWO of the University of Texas with the 5-m radiotelescope at McDonald Observatory. The beamsize at 115 GHz was 2.3^′. The region in white at ℓ ∼ 120^∘ and b ∼ +30^∘ is centered on the North Celestial Pole and could not be reached with the Texas instrument. The large white region mostly in the southern Galactic hemisphere either did not rise above the horizon or culminated at too low an altitude to be observed from McDonald Observatory. Figure from Magnani et al. (1985)

Seventy percent of the objects detected at | b | ≥ 25^∘ are located within one degree of a Lynds bright nebula (LBN; Lynds 1965). The LBN at high latitudes are prominent on both the blue and red POSS prints and are most likely dusty nebulae reflecting the starlight of the Galactic plane. Similarly, some of the high-latitude clouds are at the same locations as Sharpless H II regions (Sharpless 1959). The molecular clouds are also associated with H I features as determined from the Heiles and Habing (1974) H I survey.

The MBM survey as described by Magnani et al. (1985) established that a significant population of high-latitude molecular clouds existed, that the clouds as an ensemble were relatively nearby, and that many of them were not gravitationally bound. The last significant step in understanding their nature was to determine their connection with the IRAS infrared cirrus.

9.4 The IRAS Cirrus: High Latitude Cloud Connection

As soon as the IRAS data were released, a comparison between the infrared images at 100 μm and the dozen mapped high-latitude clouds immediately confirmed the speculation by Blitz et al. (1984) that there was a correspondence between some of the infrared cirrus and the molecular clouds. Figure 9.2, reproduced from Weiland et al. (1986) shows the striking similarity between the dust and gas tracers. The high-latitude clouds have cirrus counterparts, but many regions with similarly high 100 μm emission do not show CO(1-0). In the subsequent two decades, determining how to use the infrared data to predict the presence of molecular gas in a given region became a small cottage industry. Some of the various efforts in this direction are discussed in below.

Fig. 9.2

Relationship between IRAS 100μm emission and CO(1-0) mapping for three high-latitude molecular clouds from the MBM survey (MBM 30, 16, 20; clockwise from the bottom left). The bottom right panel shows the relationship between the IRAS emission and atomic hydrogen in MBM 20. Unlike the first three panels, the peak of the atomic hydrogen is offset from the peak of the infrared emission. This figure first appeared in Weiland et al. (1986)

The most important contribution of the groundbreaking Weiland et al. (1986) paper was establishing that at least some of the cirrus emission is interstellar. While many had suspected this, the lack of velocity information in the infrared images precluded any definitive assignment of a location. Thanks to the CO data and the distance estimates to the high-latitude clouds, the cirrus became firmly established as a feature of the local ISM.

Weiland et al. (1986) also established that some regions displaying CO emission had the same infrared properties as others that do not. The dust temperature in the CO-rich clouds is ∼ 20 K and the relatively low infrared luminosity-to-mass ratio indicates that there are no significant embedded heating sources. A linear relationship exists between W _CO and the 100 μm intensity [also noted at about the same time by Cor de Vries (1986)] but with a significantly higher slope than the value obtained by Hauser et al. (1984) in a survey of the central part of the Galaxy. The latter property was discussed as indicating that high-latitude clouds have anomalously large CO emission relative to the far-infrared cirrus clouds. Finally, they found excess infrared emission at 12 and 25 μm for the CO-rich infrared cirrus clouds, a property that had been noted for the cirrus in general (e.g., Boulanger et al. 1985). The nature of this emission was attributed either to the presence of very small dust particles (e.g., Draine and Anderson 1985) or to polycyclic aromatic hydrocarbons (PAHs; Puget et al. 1985; Allamandola et al. 1985).

The Weiland et al. (1986) paper, and similar work by the Dutch infrared/ molecular group led by Harm Habing were the first salvos of many subsequent infrared-CO comparisons for the high-latitude clouds. It is worth noting, however, that some of the questions raised by the Weiland et al. (1986) paper such as: (1) what induces the transition between predominantly atomic vs. predominantly molecular regions? (2) why is the CO/IR ratio much larger in the high-latitude clouds than is generally the case for the Galaxy or external galaxies? (3) what is the precise nature and evolutionary state of the “very small grain” component? have not been completely answered even three decades later. To explore further the infrared-molecular connection, a more complete census of the molecular content of the high-latitude sky was needed.

9.5 Post-MBM CO Surveys at High Latitudes

Shortly after the MBM survey, Magnani et al. (1986) attempted a blind survey over random regions to determine the fractional completeness of their survey, ε, sampling 10^∘× 10^∘ grids in one degree increments in ℓ and b for a total of 100 points per grid. Seventeen grids covering 1750 points (three of the grids were not completed, and one was sampled in RA and Dec and included more than 100 points) in the northern Galactic hemisphere resulted in only four detections, while 8 grids covering 750 points in the southern Galactic hemisphere produced five detections.

The results of the survey confirmed that the southern Galactic hemisphere is richer in high-latitude clouds than the north and yielded a value of ε of 0.55—implying that nearly twice as many clouds of the MBM type were still to be found. From this blind survey, estimates of the basic parameters of high latitude molecular gas could be determined: the total number of high latitude clouds is 120₋₃₀ ⁺⁴⁰, the total local molecular mass ∼ 5 × 10³ M_⊙, and the mass surface density ∼ 0.2 M_⊙ pc⁻². High-latitude molecular gas was estimated to contribute about 5–10% of the overall local molecular mass.

Other groups began to search for high latitude clouds. Keto and Myers (1986) searched for high-latitude clouds from the southern hemisphere focussing on the portion of the Galaxy that was not visible from the north. Using the Columbia University 1.2 m antenna on Cerro Tololo, Chile, they detected CO(1-0) emission from 15 high-latitude clouds ( | b | ≥ 20^∘) and from 3 small clouds associated with the Chamaeleon dark cloud complex. The objects were found in a similar manner to MBM objects: looking for regions of obscuration or reflection using the ESO sky survey plates, so the Keto and Myers survey is subject to the same selection biases as the MBM survey. Nevertheless the Keto and Myers survey showed that regions not surveyed by MBM contain high-latitude clouds, and that these objects are not gravitationally bound. However, Keto and Myers (1986) did raise the possibility that the clouds might be confined by the external pressure of the ISM.

de Vries et al. (1987) mapped a large region ( ∼ 10^∘× 8^∘) in Ursa Major and showed that six of the MBM clouds are almost certainly related. MBM 27–32 are located in this region, embedded in a large infrared loop (sometimes called the North Celestial Pole Loop—see Figs. 9.3 and 9.4, and Heithausen (1987)). Using the Columbia 1.2 m mm-wave radiotelescope,² they convincingly showed that individual high-latitude clouds from the MBM survey are parts of large scale filamentary structures, something that was not noticed by MBM because of their severely undersampled maps. The high-latitude clouds in this region have come to be known as the Ursa Major clouds and half of them were the subject, a decade later, of a more comprehensive, higher resolution study by Pound and Goodman (1997). One of the most important results by de Vries et al. (1987) was an attempt to calibrate the CO-H₂ conversion factor by analyzing the relationship between the infrared emission from the dust in the clouds, the atomic gas contribution from the 21 cm H I line, and the molecular contribution from the CO data.

Fig. 9.3

Dust map from Schlegel et al. (1998) of the infrared cirrus clouds near the North Celestial Pole. Initial detections of several molecular clouds in the region were made by Magnani et al. (1985), Heithausen (1987), and de Vries et al. (1987). These clouds are sometimes known as the North Celestial Pole Loop or the Ursa Major clouds. The prominent structure in the southwest portion of the image is the top of the Polaris Flare. Image made with the Skyview Virtual Observatory

Fig. 9.4

Image from the Planck satellite of approximately the same region as in Fig. 9.4. This image is centered on (ℓ, b = 135^∘, +40^∘) and is 30^∘× 30^∘. The colors represent dust emission with the warmer colors representing greater emission. The “waves” show the orientation of the magnetic field in the region from the 353 GHz Planck polarization data using the Linear Integral Convolution technique devised by Diego Falceta-Goncalves. Image: ESA

A different approach was taken by Heiles et al. (1988). Instead of selecting regions based on obscuration, they identified 26 isolated, degree-sized IRAS cirrus clouds and then observed their HI emission to define S₁₀₀/N _HI , the ratio of the infrared surface brightness at 100 μm to the atomic hydrogen column density. Heiles, Reach, and Koo found that regions with high values of S₁₀₀/N _HI contained sufficient molecular gas for the detection of the CO(1-0) line. They found that CO emission can be detected even in clouds with hydrogen column density as low as ∼ 2.4 × 10²⁰ cm⁻², and identified 11 high-latitude clouds, later increased to a total of 14 by Reach et al. (1994). The CO-emitting regions in these objects tend to be compact and smaller than the clouds in the MBM survey. Moreover, their latitude distribution differs markedly from that of the MBM clouds. This is partly due to the search criteria employed by Heiles, et al., but may also highlight a real difference in cloud populations. In a somewhat similar vein, Clemens and Barvainis (1988) optically selected a sample of round or elliptical, small, obscured regions. These objects are mostly at lower latitudes and are similar to Bok globules. Five of the objects are at | b | ≥ 25^∘ and exhibit CO emission so they can be considered high-latitude clouds.

Heithausen et al. (1993) surveyed 620 square degrees in the region 117^∘ ≤ ℓ ≤ 160^∘ and 16^∘ ≤ b ≤ 44^∘. They found that 13% of the region showed CO(1-0) emission. Their sampling was quite good: mostly 15^′ with an 8.7^′ beam. with 245 square degrees mapped at 0.5^∘ resolution. Although their survey dipped significantly below the Blitz et al. surveys in latitude, a 25^∘ minimum latitude still produced a surface filling fraction of 9.2%. However, the large feature known as the Polaris Flare dominated the CO detections: Of the nearly 80 square degrees with CO emission, 40.21 of them were from the Polaris Flare. Excluding the Polaris Flare, the surface filling fraction decreases to ∼ 4.1%, still significantly greater than the Magnani et al. (1986) blind survey. Along with the surface filling fraction, the mass surface density of molecular gas in this region also increases to 0.86 M_⊙ pc⁻² if one assumes that even at b > 44^∘ the CO detection rate does not decrease, and to 0.47 M_⊙ pc⁻² if there is no CO at b > 44^∘. It was clear by comparing these results to those of Magnani et al. (1986), or even MBM, that the region in the northern Galactic hemisphere near the North Celestial Pole is relatively rich in molecular gas, but whether similar high-latitude regions existed in either hemisphere was not known at the time.

A search for molecular clouds toward intermediate-to-high latitude IRAS point sources (signposts of star formation) in the southern Galactic hemisphere was made by Yonekura et al. (1999). Using the NANTEN 4-m millimeter-wave telescope to survey 29 IRAS-selected targets, they detected CO(1-0) emission from five of them, with three sources at | b | ≥ 24.7^∘. However, all 3 objects had been previously detected by Knapp et al. (1977), MBM, and Bally et al. (1991). A more extensive survey was made, using the same telescope, by Onishi et al. (2001), who searched the far-infrared excess clouds identified by Reach et al. (1998) using the DIRBE instrument on the COBE satellite. Onishi et al. observed those regions which rose at least 30 degrees above the horizon at Las Campanas, Chile, the site of the telescope, 68 out of 153 candidates. They observed all 68 detecting CO emission from 32 and often mapping them systematically. The immediate consequence of this work was to increase the number of known high-latitude molecular clouds significantly. Figure 9.5 shows all the CO detections or clouds at | b | ≥ 30^∘ that were known at the end of 2011 before the Planck satellite and the systematic high-latitude CO survey by the Thaddeus/Dame group.

Fig. 9.5

All known spectroscopic CO(1-0) detections at Galactic latitudes | b | ≥ 25^∘ as of 2011. See text for details

The studies described above all used CO(1-0) emission line as a surrogate for tracing molecular hydrogen, but CO(1-0) absorption can also be used. At 115 GHz, CO(1-0) absorption lines can be detected against the continuum of background quasars or active galaxies that have sufficient fluxes at 2.6 mm to allow detection of molecular line absorption. A search for continuum sources suitable for millimeter absorption studies was made by Liszt and Wilson (1993) and was discussed in the context of diffuse molecular clouds in Chap. 7 From the survey point of view, over 100 sources were selected for CO(1-0) emission observation, 64 of which were at | b | ≥ 20^∘. Since Liszt and Wilson were searching for lines of sight with CO emission, this result was an almost blind high-latitude CO survey in the spirit of Magnani et al. (1996). Of the 64 high-latitudes sources, only one (B2251+158) showed clear CO absorption. This line of sight to the quasar 3C454.3 had been known as a source of molecular emission as early as the OH survey of Kazès and Crovisier (1981) and is likely associated with the envelope or environs of the nearby MBM 53–55 complex. The low detection rate at | b | ≥ 20^∘ corroborates the results of Magnani et al. (1986) and the original Magnani et al. (1985) results. A smaller survey for CO(1-0) absorption at somewhat higher Galactic latitudes was conducted by Liszt and Lucas (1998). The lines of sight included had been previously studied in other molecular tracers. Although they detected CO absorption in 8 of the 9 sources they observed, only one, B2251+158, was at what we would consider high-latitude. An earlier survey for HCO⁺ absorption towards extragalactic sources by Lucas and Liszt (1996) had better high-latitude coverage (9 lines of sight), and produced only two detections (B2145+067 and B2251+158). Although, in principle, absorption line observations can detect molecular gas, a more thorough survey at high-latitudes would have to be conducted to yield results comparable to the CO emission surveys.

9.6 The Georgia-Harvard CfA High Latitude Survey

In the late 1990s a truly unbiased CO survey of the high-latitude sky was made using the Harvard Center for Astrophysics 1.2-m millimeter wave telescope. This is the instrument used for the wide-scale Galactic plane survey by Thaddeus and Dame (see Fig. 7.1). The northern and southern sky at | b | ≥ 30^∘ which culminated at least 30^∘ above the horizon from Cambridge, Massachusetts was surveyed every degree in Galactic latitude and longitude. The survey pattern for the northern and southern skies is shown in Figs. 9.6 and 9.7.

Fig. 9.6

Polar projection of the mapping grid for the northern Galactic hemisphere for the Georgia-Harvard CfA High-Latitude CO survey. The color image represents IRAS 100 μm emission while the white dots are the grid pattern observed with the 1.2 m Harvard-Smithsonian center for Astrophysics millimeter-wave telescope. The results of this survey were reported by Hartmann et al. (1998)

Fig. 9.7

Same as Fig. 9.7 but for the southern Galactic hemisphere. More information on this survey can be found in the paper by Magnani et al. (2000)

The results from the surveys are presented in Hartmann et al. (1998) and Magnani et al. (2000). In the Northern Galactic Hemisphere (NGH) 10,562 out of a maximum of 11,478 points at | b | ≥ 30^∘ were observed. The missing points did not rise above 30^∘ altitude from Cambridge, Massachusetts, the site of the 1.2 m radiotelescope. Only 26 lines of sight showed evidence of CO(1-0) emission at T _mb  > 0.3 K in antenna temperature (3σ rms). This was not surprising as the NGH does not have many high-latitude molecular clouds.

In the Southern Galactic Hemisphere (SGH) 4982 points were observed with the same sensitivity. There were 133 detections, 75 of which coincide with previously known molecular clouds. The average centroid velocity of the detections is −1.67 ± 0.58 km s⁻¹ attesting again to the local nature of the clouds. The average T _A ^∗, line width (FWHM), and W _CO are 0.83 K, 2.08 km s⁻¹, and 1.81 K km s⁻¹. Both the average antenna temperature and W _CO are lower than in the MBM survey which was biased towards larger, more opaque clouds. The number of CO detections in both hemispheres is significantly lower at | b | ≥ 50^∘ than would be expected from extrapolating the molecular in-plane results from the Dame et al. (2001) survey. This may be an indication that supernovae from the Scorpius/Ophiuchus associations have “blown through” the disk and cleared out most of the denser neutral material at the highest latitudes. This is not a new idea. The Galactic Fountain model (see Chap. 1) can produce blowouts of hot gas from the plane to the halo, and Heiles (1993) wrote an interesting paper titled “The Worm-Ionized ISM” where supernovae produced chimneys or “worms” (so-called because of their morphology) of hot gas that vent to the halo. Whether this explains the marked drop in molecular clouds at | b | ≥ 50^∘ or that something else is at work requires further investigation.

Both surveys calculated a surface filling fraction for molecular gas, although this quantity is uncertain without a model for the distribution. However, taken at face value, ε quantifies how easy it may be to detect CO emission in random directions for | b | ≥ 30^∘. For the NGH, ε = 0.0025 and 0.03 for the SGH. This significant discrepancy between NGH and SGH was explained in the original MBM survey as a displacement of the Sun above the Galactic midplane (see Magnani et al. 1996, for details). However, it is important to note that this result, i.e., that there is an order of magnitude more molecular gas in the SGH than the NGH at high latitudes, was called into question by a couple of surveys for molecular gas which used infrared emission from dust as the tracer for molecular gas. We discuss these in the next section. Finally, the mass surface density of molecular gas in the NGH is 0.025 M_⊙ pc⁻² and 0.09 M_⊙ pc⁻² in the SGH. Comparing these values to the local value of molecular gas from all sources (1-2 M_⊙ pc⁻²) indicates that the results of the Georgia-Harvard CfA high-latitude CO surveys are consistent with previous work: The high-latitude molecular clouds contribute about 10% to the local inventory of molecular gas.

There is a certain artificiality about ε and mass surface density estimates given for a 30^∘ cutoff in latitude. If the cutoff was lowered to 25^∘, a good portion of the Polaris Flare (Heithausen and Thaddeus 1990) would be included in the survey and both numbers would increase. If the survey includes all points above 20^∘, the entire Polaris Flare is covered along with much of the Ophiuchus dark cloud complex (De Geus 1988), increasing both ε and mass surface density. Nevertheless, comparing the northern vs. southern Galactic hemispheres for a given cutoff, the north has significantly less molecular gas than the south, for CO(1-0) emission surveys with 0.1 K, 1σ sensitivity.

The bulk of the detections are clustered around the MBM 53–55 region (ℓ ∼ 90^∘, b ∼ −35^∘) and south of the Taurus/Auriga dark clouds (ℓ ∼ 170^∘, b ∼ −40^∘) There may also be an enormous partial loop structure extending from southwest of the Taurus dark clouds, extending to ℓ ∼ 130^∘, b ∼ -45^∘), and curling back up east of the MBM 53–55 region. This feature was noted by Bhatt (2000) who connected it to the Ursa Major clouds in the north and attributed their formation to a single (dynamical) event. Even if the whole loop postulated by Bhatt is not a real feature, the lower portion could still be a connected feature. A velocity-coded map of the southern detections is shown in Fig. 9.8 and a gradient may be present in this feature.

Fig. 9.8

Detections of CO(1-0) emission from the SGH. Color coded for LSR velocity. The color bar is at the top right of the figure. The green diamond represents a cloud at a velocity of ∼ 12 km s⁻¹

The mass surface density value of 0.09 M_⊙ pc⁻² was derived by Magnani et al. (2000) using X _CO  = 1 × 10²⁰. A factor of two increase in X _CO brings the mass surface density estimate more in line with the result from Magnani et al. (1996) derived from an inventory of the known high-latitude clouds (at that time), encompassing a sample of objects identified using very different selection criteria. This encouraging concordance implies that the two NGH and SGH surveys are an excellent snapshot of the general properties of the high-latitude molecular clouds detected by the previous non-systematic searches.

However, there are three worrisome aspects of the results from the Georgia-Harvard CfA high-latitude CO surveys: (1) The search grid was chosen on the basis of the typical size cloud from Magnani et al. (1985). While all the “big fish” were certain to be caught in the sampling net, it is clear that some of the smaller objects would be missed. A particularly telling example is shown in Fig. 9.9. The question of the size distribution of high-latitude clouds has never been satisfactorily addressed because of the lack of complete mapping data for the majority of the objects. Even among the mapped clouds, many researchers have focused on the core regions and ignored the more extended, low CO intensity regions. Indications of that a population of smaller clouds likely exists is clear from the surveys of Keto and Myers (1986), Heiles et al. (1988), and Heithausen (2002). A systematic search for this population of objects at high-latitude (or even at low latitude) has never been carried out. (2) Similarly, a population of intermediate-velocity molecular clouds is known to exist at high-latitudes (Magnani and Smith 2010; Heiles et al. 1988). The Draco cloud is a famous, but atypical example of this population (Mebold et al. 1985; Magnani et al. 1985), but the other half-dozen or so known clouds are smaller and are likely much more distant than the typical high-latitude clouds. How many of these clouds exist, and where they are located with respect to the Galactic plane is still unknown. (3) The surveys targeting regions with excess infrared emission have always determined that there was more molecular gas than the CO(1-0) surveys detect. In particular Reach et al. (1998) claim that roughly equal amount of molecular gas are present at intermediate and high-latitude in the NGH and SGH, in disagreement with the Georgia-Harvard CfA surveys results. We now examine this last point in greater detail.

Fig. 9.9

High-latitude cloud (MBM 25) that was missed by the sampling grid pattern from the Georgia-Harvard-CfA survey shown in Fig. 9.7. This figure is from Hartmann et al. (1998)

9.7 High-Latitude Molecular Gas via Infrared Techniques

Gas and dust are well-mixed in the ISM and H I column density maps of large areas are relatively easy to make. Thus, if one removes the dust emission in a band anywhere from 60–300 μm that is due to grains associated with the H I, the remainder should trace H₂. This relatively straightforward approach was made possible by the opening up of the infrared electromagnetic window in the early 1980s. Désert et al. (1988) used this technique to identify over 500 “infrared excess” regions at | b | ≥ 5^∘. Could all these objects be detected in CO(1-0)? Blitz et al. (1990) answered “no”, detecting only 13% in a survey of 201 DBB clouds, The undetected DBB objects seem to be clouds with enough H₂ for significant column densities, but not enough molecular hydrogen to sufficiently self-shield to yield CO detections (i.e., regions composed primarily of dark gas as discussed in Sect. 8.4). Others arrived at this conclusion from similar types of observations (e.g., Reach et al. 1994; Meyerdierks and Heithausen 1996).

The most complete work of this type was done by Reach et al. (1998) who improved on the method of Desert, Bazell, and Boulanger by using COBE data (in particular, the 240 μm DIRBE data) to pick out the colder infrared clouds. In addition, the H I data that they used were of higher quality and better calibrated than what was used in the Désert, Bazell, and Boulanger catalog. The RWO catalog consists of 60 infrared excess regions associated with previously identified molecular clouds either from the MBM catalog or the Magnani et al. (1996) compilation, and 81 objects not associated with known molecular clouds at that time. All of these objects are at | b | ≥ 20^∘. The surface mass density of the infrared excess clouds is similar to that of the original MBM CO survey: 0.3 M_⊙ pc⁻², but their results differ significantly from those of the Georgia-Harvard CfA high-latitude CO surveys in one key aspect: Their SGH mass surface density is similar to Magnani et al. (2000), but an order of magnitude greater in the NGH than what was obtained by Hartmann et al. (1998). In other words, the North/South asymmetry in the CO distribution originally noted in Blitz et al. (1984) and confirmed by the Georgia-Harvard CfA blind surveys is not evident in the RWO infrared excess clouds.

As described above, a search for CO emission from a sample of RWO objects was made in 2001 by Onishi et al.. They observed 68 infrared-excess objects (all the RWO sources that were visible from the site of the telescope used for the survey, the NANTEN 4 m mm-wave radiotelescope) from the RWO catalog and detected CO emission from 32, a significantly higher percentage than Blitz et al. found from the DBB catalog. This underscores how the temperature correction in the RWO catalog made a significant difference in identifying infrared regions that might contain H₂. Similarly, the temperature corrected dust maps from Schlegel et al. (1998) (Sect. 6.7.1) are an excellent tool for identifying regions in the diffuse ISM where molecular gas might be present.