Next Contents Previous


6.1. Summary of standard molecular abundances in various regions of the Milky Way, and models of interstellar chemistry

The broad features of the relative abundances of molecular species in various standard extragalactic environments share common properties which have many similarities with those observed with much more details in the Milky Way. The latter have been comprehensively studied for decades both by millimetre observations and theoretical chemistry modelling. As discussed in Section 2.4, a few archetypes have emerged for the general patterns of molecular abundances, depending on the physical conditions which directly affect the processes of molecular formation and destruction. Indeed, despite the interest of fine studies of the peculiarities of individual Galactic sources, allowed by their relatively close distances, waiting for ALMA (Turner 2007), only very broad classes of interstellar chemistry are really relevant for comparison with the coarse and global determinations of extragalactic abundances. For this purpose, two main patterns emerge for the molecular abundances of Milky Way interstellar sources. They correspond to cold, quiescent molecular clouds, and to hotter, massive-star forming regions, respectively.

In both cases, CO is by far the most abundant, easily observed molecule, with an abundance ~ 10-4 with respect to H2, corresponding to a significant fraction of total carbon and oxygen. All other molecules are less abundant by at least a factor 100 (OH, H2O), and rather ~ 104-106 for most of them.

In very cold, dense, dark molecular condensations, with high density but no massive star formation, such as TMC-1 or L134N, a large set of complex molecules, especially carbon-rich, are observed (e.g. Ohishi et al. 1992). They are clearly built up by low temperature gas phase reactions. However, to account for the observed abundances one needs to include a time dependent chemistry, and/or the long-term effects of grain accretion producing very large depletions and increasing the C/O abundance ratio (see e.g. the comprehensive review of chemical modelling by van Dishoeck 1998, and Wakelam et al. 2006 for more recent references). Such detailed studies of small dark clouds are only possible at small distances in the Milky Way and are not directly relevant for extragalactic comparisons. However, it has been shown (see e.g. Terzieva & Herbst 1998, Turner 2000) that a similar chemistry is at work even with the lower extinction (Av ~ 1-5) of translucent clouds, where several tens of such species are broadly observed. The 38 species reported in Table 1 of Turner (2000), and partially displayed in Table 2, may be classified in four groups of similar size: basic simple interstellar molecules (OH, NH3, H2CO, HCO+, N2H+, CN, HCN, HNC); hydrocarbon radicals, from C2H through C6H which are the backbone of cold gas phase chemistry and form a highly homogeneous group (to which one may add c-C3H2 and HCnN); heterogeneous complex species such as CH3OH, CH3COH, CH3CN, which have the complexity of hydrocarbons but contain O and N as well as C; a number of sulfur-bearing species more depending on the degree of grain desorption (plus SiO whose abundance remains very weak). Similar models may even be extended to diffuse clouds (Av ~ 1), where, in addition to optically detected species (CH+, CH, CN ...), the most detailed information comes from millimetre line absorption studies (Liszt & Lucas 1999, Liszt et al. 2005 and references therein) in which basic species of the first group above, plus CS, SO, C2H, c-C3H2, are commonly detected. A number of the observed abundances (Table 2) are reasonably accounted for by models with standard lower depletion of heavy elements for the diffuse medium (Morton 1975). The addition of the effect of a few turbulence driven, weakly endothermic reactions (e.g. Spaans 1995, 1996, Turner 2000) may improve the agreement; but several serious problems remain (see e.g. Lequeux 2005, Sec. 9.4.1). Anyway, one may expect to find similar conditions in observations of cold extragalactic medium, either in beam averaged emission of cold giant molecular clouds, or rather in millimetre line absorption (Section 7.2).

Table 2. Comparison of molecular abundances (with respect to H2) in selected Galactic sources [dark cloud: TMC-1, photodissociation region: Orion Bar, hot core: Sgr B2(N), typical nuclear bulge cloud: Sgr B2(OH)], the starburst galaxies NGC 253 and M 82 (all from Tables 7 & 9 of Martín et al. 2006a); LMC/N159 (Johansson et al. 1994); and Galactic translucent clouds (Turner 2000). In addition to the molecules detected in NGC 253 displayed here, molecules detected in external galaxies include (see Table 3 of Martín et al. 2006a and Muller et al. 2006): H2, HD, CO, OH, CH, CH+, CN, CO+, H2O, C2H, HCO, HOC+, N2H+, C2H5OH, DCN, DCO+, 13CO, C18O, 13CS, H13CN, HC15N, HC18O+, HC17O+, H13CO+, HN13C, H15NC, C34S, H234S; H3+, C2H2, CO2.

Molecule NGC253 M82 LMC Sgr Sgr TMC-1 Orion Translucent
      N159 B2(N) B2(OH)   Bar cloud

HN13C -10.6 -9.5 ... -11.0 ... ... ... ...
H13CO+ -10.4 -9.9 ... -11.4 ... ... -10.3 ...
SiO -9.9 <-9.9 ... -10.7 ... < -11.6 -10.3 -10.0
NH2CN -9.7 ... ... -10.1 -10.0 ... ... ...
C2S -9.7 ... ... ... -9.6 -8.1 ... ...
CH3CN -9.5 -9.7 ... -6.7 -9.4 -9.0 < -10.3 <9.0
c-C3H -9.5 ... ... -10.5 < -10.9 -9.3 ... -8.0
HOCO+ -9.4 ... ... -10.5 -9.7 ... ... -9.0
C34S -9.4 -9.3 ... -10.2 ... ... -9.0 ...
c-C3H2 -9.3 -8.1 ... -10.5 -9.8 -8.0 -9.7 -7.4
HC3N -9.2 -8.7 ... -7.5 -9.0 -8.2 ... -9.3
NS -9.2 ... ... -7.0 ... -9.1 ... ...
H2CS -9.2 ... ... -6.8 -8.7 -8.5 ... -7.6
SO2 -9.1 ... ... -6.6 -8.7 < -9.0 -9.9 -8.2
CH2NH -9.1 ... ... -7.0 -9.2 ... ... -7.8
H2S -9.1 ... ... -9.9 ... < -9.3 -8.2 -7.6
HNC -9.0 -8.8 -10.2 ... ... -7.7 -9.0 -8.6
SO -8.9 <-8.5 -8.6 -6.9 -8.7 -8.3 -8.0 -7.5
HCO+ -8.8 -8.4 -9.7 ... ... -8.1 -8.5 -8.7
HNCO -8.8 <-8.8 ... -9.2 -8.4 -9.7 < -10.8 ...
H2CO -8.6 -8.2 -9.3 -9.3 -8.6 -7.7 -8.2 -8.2
OCS -8.4 -7.9 ... -8.6 -8.3 -8.7 ... -9.2
HCN -8.3 -8.4 -9.7 ... ... -7.7 -8.3 -7.4
CS -8.2 -8.2 -9.4 ... ... -8.0 -7.6 -8.0
CH3CCH -8.3 -7.7 ... -8.4 -8.8 -8.2 ... ...
CH3OH -7.9 <-8.3 ... -5.8 -7.3 -8.7 -9.0 ...
C2H -7.7 -7.6 ... -9.7 ... -7.1 -8.7 ...
NO -7.2 ... ... -6.0 ... < -7.5 -8.6 ...
NH3 -7.2 ... ... ... ... ... ... -7.7

Warmer molecular regions are associated with massive-star formation. In addition to warmer kinetic temperature favouring weakly endothermic reactions, one may then find enhanced photochemistry, and strong shocks whose main effect is probably efficient grain desorption not only of weakly bound species such as CO, CO2, NH3, ices and organic material, but also eventually refractory elements such as silicon. Archetypes of such regions are the so called `hot molecular cores', whose best examples are Orion KL and SgrB2(N). Their chemistry has been studied in detail, both from observations and modelling (see e.g. van Dishoeck & Blake 1998, Walmsley 1996, Ikeda et al. 2001, Wakelam et al. 2005 and papers in the proceedings of IAU Symposium 231 on astrochemistry, Lis, Blake & Herbst 2005). Temperatures are above 100 K, densities above 107cm-3 and column densities in the range 1024 - 1025cm-2. Hydrogenated species such as H2O, NH3, CH3OH, CH3CN, C2H5CN, CH3CO and a variety of other complex organic molecules, have abundances at least an order of magnitude larger than in cold dark clouds (Table 2). Enormous enhancements of SO, SO2, H2S and SiO, as well as deuterated species, are also found. Detailed models reasonably well account for the observed abundances (see van Dishoeck & Blake 1998 and references therein). The chemistry is driven by the evaporation of icy grain mantles containing a mixture of H2O, CO, CH3OH, NH3, HCN, etc. Among the observed species, one may distinguish: very stable precollapse molecules, such as CO and C2H2, which were frozen on the grains and released into the gas-phase largely unaltered; molecules made in grain-surface reactions and released in the gas (H2O, NH3, H2S, CH3OH, etc.); and those which are produced by rapid gas-phase reactions between evaporated molecules. In addition to mm-line surveys of many tens of species and isotope varieties, mid-IR infrared observations may probe species such as C2H2, HCN, OCS, NH3, etc. Because of its proximity, the Orion-IRC2/KL region is by far the best studied. Within this region one may find local differences in chemistry, e.g. between complex O- and N-bearing organics. Important lifetime effects are found in models, e.g. for SO and SO2, which may be used as clocks to date the age of various sources.

Strictly speaking, hot cores are very small (leq 0.1 pc), very dense and hot. On larger scales, more relevant for starbursts in galaxies, the chemistry of active star formation regions is still dominated by the influence of massive stars through photodissociation and various kinds of shocks. It partakes some properties of the hot core chemistry such as relatively high temperature and density, and short time constants. However, it may significantly differ depending of the relative contribution of UV photodissociation and of grain desorption by shocks.

Detailed physical and chemical properties of photodisssociation regions (PDRs) have been discussed by Tielens and Hollenbach (1985), Hollenbach & Tielens (1997, 1999), Jansen et al. (1995), Tielens (2005). One fundamental feature is the rapid spatial variations of these properties and in particular of molecular abundances. Some species are specific of PDRs, such as ions (CO+, SO+, HOC+) and in a less degree products of photo-chemistry such as c-C3H2 (Fuente et al. 2005). On the other hand, complex molecules, such as CH3OH, CH3CN, etc., are underabundant in PDRs (such as Orion Bar, Table 2) compared to hot cores or Galactic Center clouds, because of rapid destruction by photodissociation.

The special concentration of molecular clouds within 200 pc of the Galactic Center (e.g. Morris & Serabyn 1996) is the medium in the Milky Way that is closest to extragalactic molecular sources. Molecules typical of grain desorption chemistry, such as C2H5OH and SiO (Minh et al. 1992), are remarkably widespread, and indeed most of the various clouds exhibit abundances of complex molecules reminiscent of hot cores (Requena-Torres et al. 2006 and references therein). The abundance ratios of various complex organic molecules relative to CH3OH are roughly constant, suggesting that all complex molecules are produced by a similar chemistry initiated by shock desorption of grains. However, the overall abundances may vary by orders of magnitude. They may be very high, even higher than in hot cores, and the abundance of CH3OH may reach 10-6. However, the prototype source, SgrB2(OH) in the envelope of the prominent star forming complex SgrB2, displays abundances smaller by one or two orders of magnitude than in the hot core Sgr(N) for molecules such as CH3OH, CH3CN, HC3N, SO, SO2 and SiO (Requena-Torres et al. 2006, as shown in Table 2). Note that the observed abundances are dramatically smaller in the few regions very close to the Galactic Center submitted to intense photodissociation from the UV radiation of starburst clusters of massive stars.

As a conclusion, in view of interpreting observed extragalactic abundances, Galactic translucent clouds seem a good reference for abundances derived from molecular absorption lines through the standard cold ISM (Section 7.2). On the other hand, typical observable extragalactic emission lines come from starburst regions and must obey some kind of warm chemistry similar to hot cores, shocks or PDRs. In the average over the telescope beam in external galaxies, one may indeed expect rather some kind of combination of these related, but distinct, types of chemistry, since they are known to coexist at very short distances in Galactic sources such as Orion or Sgr(B).

6.2. Observed abundance variations in local galaxies: I. the Magellanic Clouds

Because of the distances, extragalactic studies of molecular emission are severely limited compared to the details which are achieved in studies of the molecular gas in the Milky Way, such as those dealing with fine structure or the abundances of rare species. Even in our closest neighbours, the typical sensitivity for small sources is reduced in the Magellanic Clouds by almost two orders of magnitude with respect to the Galactic Center, and by four orders of magnitude in the other galaxies of the Local Group such as Andromeda (M 31). Outside of widespread CO (and to a less extend 13CO), practically all extragalactic molecular studies are limited to the strongest emitting sources, generally associated with active well localized massive-star formation regions, except in starburst galaxies where they are more widespread.

Therefore, there is practically no detailed studies of a large set of molecular abundances except in a few sources of the nearby Magellanic Clouds, and in local starburst infrared galaxies or nuclear starbursts, where star formation activity fills most of the telescope beam.

The Magellanic Clouds have the special interest to provide tests of interstellar chemistry with a much lower metallicity and enhanced UV radiation compared to the Milky Way, a situation which could be similar in many high redshift galaxies. One thus expects a large decrease in the overall abundances of molecules, and first of H2 whose abundance is further reduced by the lower rate of formation on rarer dust grains. A large programme with the space telescope FUSE has detected H2 UV absoption lines along ~ 50 interstellar lines of sight in LMC and SMC (Tumlinson et al. 2002). The amount of H2 is on average an order of magnitude smaller than along lines of sight of the Galactic disk over a similar range of reddening. These results imply that the diffuse H2 mass is only about 0.5% and 2% of the HI mass derived from 21 cm emission measurements in SMC and LMC, respectively. The high UV radiation enhances the excitation of upper rotational levels. Far-UV lines of CO and HD have also been measured in a few lines of sight (Bluhm and de Boer 2001, André et al. 2004). Recent VLT/UVES optical observations in a dozen of lines of sight have compared the abundances of CH, CH+ and CN with Galactic ones (Welty et al. 2006). The CH/H2 ratio is comparable or smaller than the values found for Galactic diffuse clouds. The observed relationships between the column density of CH and those of CH+ and CN show the same trends as in the local Galaxy. The authors discuss in detail the extension of chemical models for diffuse clouds to the smaller metallicities and higher UV radiation field of these galaxies. A significant fraction of the CH and CH+ observed may arise in photon-dominated regions which should be more extensive than in our Galaxy.

Millimetre studies of CO in the Magellanic Clouds are discussed in Section 4. Millimetre detections of a number of other molecular species (13CO, CS, SO, CCH, HCO+, HCN, HNC, C2H, CN, H2CO, and C3H2) achieved near two peaks in the CO emission of the LMC and SMC, have provided some information on molecular abundances, isotopic ratios, and cloud structure (Johansson et al. 1994; Chin et al. 1997, 1999). The molecular abundances are about an order of magnitude lower (and even more for CN) than the corresponding values found for Orion KL and TMC-1 (Table 2). However, molecular studies of the Magellanic Clouds have suffered from the low number of large millimetre facilities in the Southern Hemisphere. This situation is now changing with new facilities. ATCA has already produced first 3 mm results at high angular resolution (Wong et al. 2006). APEX (12 m) has replaced SEST, allowing the extension of molecular studies to the submm range with a better receiver equipment (see also AST/RO, Bolatto et al. 2005); the ASTE 10 m and Mopra-22m telescopes have undertaken extensive survey works of the LMC, SMC and Bridge (Hughes et al. in prep., Muller et al. in prep.; see also the extensive contributions from the 4 m NANTEN telescope for CO observations of the LMC and SMC). ALMA will allow a full development of interstellar chemistry in the Magellanic Clouds.

6.3. Observed abundance variations in local galaxies: II. Nearby starbursts and other galaxies

In contrast to the Magellanic Clouds, the diversity of the molecular content is better documented in a few galaxies outside of the Local Group despite distances almost 100 times larger. This is due to the fact that these galaxies are powerful starbursts and are observable with the most sensitive millimetre facilities such as the IRAM-30m telescope. However, the number of different molecules presently detected, ~ 40 (plus isotopic varieties, Table 3), remains far behind the total number, ~ 150, known in local sources of the Milky Way (see updated list in, see also As for these archetype Galactic sources, the best success has been achieved by a systematic spectroscopic survey of a broad frequency range (most of the 2 mm atmospheric window) that Martín et al. (2006a) have recently achieved on the nuclear region of the starburst galaxy NGC 253. This work extends and synthesizes the results of 20 years of continuous similar studies by members of this team (see e.g. Henkel et al. 1991, García-Burillo et al. 2006a). More than 100 spectral features are identified in the 2 mm band alone with the IRAM 30m telescope, corresponding to transitions from 25 different molecular species by Martín et al. (2006a, see Table 2). The derived abundances show striking similarities with those observed in the molecular clouds of the Galactic Center, such as SgrB(OH), which are believed to be dominated by low-velocity shocks. A comparison of the chemical composition of the nuclear environment of NGC 253 with other well observed nuclear starbursts of nearby galaxies (Martín et al. 2006a, Fig. 7) demonstrates the chemical similarity of galaxies such as IC 342 and NGC 4945 to NGC 253. On the other hand, the chemistry of NGC 253 appears clearly different from that of M 82 which is another archetype of a nearby nuclear starburst galaxy. M 82 abundances of molecules like SiO, CH3OH, HNCO, CH3CN and NH3 are systematically low in comparison to NGC 253 (Table 2), while species like HCO (García-Burillo et al. 2002) and C3H2 (Mauersberger et al. 1991) are overabundant. This suggests that photo-dissociation dominates the heating and the chemistry in most of the nuclear region of M 82, which has much more many HII regions than NGC 253. The detection of widespread emission of HCO, HOC+ and CO+ in this galaxy disk reveals that the nucleus of this prototypical starburst is a giant (~ 650 pc) extragalactic PDR (García-Burillo et al. 2002, Fuente et al. 2005, 2006). The detection of a  500 pc molecular gas chimney in SiO indicates the occurrence of large-scale shocks in the disk-halo interface of this galaxy (García-Burillo et al. 2002). However, the detection of abundant and high excitation CH3OH (Martín et al. 2006b) suggests its injection from dust grains and the existence of dense warm cores, shielded from the UV radiation and similar to the molecular clouds in other starbursts. Such examples show that we have already clues to understand the most active extragalactic chemistry in various starbursts. Similar less powerful Galactic sources, such as Galactic Center clouds, PDR and shock regions, may help to disentangle the complexity of extragalactic sources in various stages of evolution, including merger and AGN effects.

A few molecules which have strong lines and are widely distributed, have been studied in greater detail than the bulk of the molecules reported in Table 2. Besides CO and OH (Sections 3 & 4), HCN, CS and HCO+ have a particular interest because of their abundance, their simple linear structure and their large dipole moments (~ 20-50 times larger than CO). As discussed in Section 5 for HCN, they are thus ideal tracers of high-density star-forming clouds. HCN is by far the most used for this purpose (see the systematic work of Gao and Solomon 2004a and Section 5.3). HCN has now been detected in the center of about 60 CO-bright galaxies, including mostly luminous and ultra-luminous infrared galaxies, and also a number of the nearest normal spiral galaxies, and mapped in the disk of some of them (see Gao & Solomon 2004a for a complete list of references). CS has also been detected in some of those galaxies (see references in Gao & Solomon 2004a), as well as HCO+ (Nguyen-Q-Rieu et al. 1992, Brouillet et al. 2005, Muller et al. 2005).

However, the number of detected molecules remains very limited in `normal', non starburst nearby galaxies. Even in M 31, where large efforts have been devoted to intensive studies of the distribution of CO (Section 4 and Fig. 1), only a few molecules such as HCN and HCO+ (Brouillet et al. 2005), could be detected. Even absorption line studies which have the advantage of a sensitivity independent of distance, are practically limited to the special cases of OH and H2CO lines, or to the exceptional nearby radio source Cen A (Wiklind & Combes 1997a) and a few exceptional high-z systems (Section 7.2).

A few symmetrical molecules lacking permanent electric dipole cannot be detected through millimetre rotation lines, but only in the infrared. Besides H2 (Section 10), the extragalactic detections are still limited to infrared absorption lines of a few prominent molecules in front of the strong continuum of a very few starburst galaxies: H3+ in IRAS 08572+3915 NW with UKIRT (Geballe et al. 2006), CO2 and C2H2 (together with CO, HCN, ice and silicates) in deeply obscured ULIRG nuclei with Spitzer (Spoon et al. 2004, 2005, Armus et al. 2006, Lahuis et al. 2007).

6.4. Abundance ratios of isotopic varieties, and inferences for chemical evolution of galaxies

The study of interstellar CNO isotope ratios is important for tracing the chemical evolution of galaxies and their nucleosynthesis. The widely separated molecular lines of different isotopic varieties (isotopologues) provide a unique way of measuring isotope ratios even in regions heavily obscured. However, great care must be taken in dealing with radiative transfer, line formation and isotopic fractionation (see Section 2 and e.g. Wilson & Matteucci 1992). In the Milky Way significant variations of the CNO isotopic ratios are known in different environments, reflecting their various nucleosynthesis history (see e.g. the review by Wilson & Rood 1994 and references therein, and examples of isotopic ratios in Table 3, reproduced from Muller et al. 2006, for the local ISM, the Solar System, the Galactic Center and a C-rich AGB star).

In external galaxies, since the first detection of 13CO by Encrenaz et al 1979, only a few isotopic abundance ratios have been reported. These studies were reviewed in particular by Henkel & Mauersberger (1993) and Muller et al. (2006). The best studied cases are the LMC and a few prominent local starbursts such as NGC 253, NGC 4945 and M 82 (Table 3). The trends observed in the isotopic ratios are explained by the various levels of nuclear processing. The ratios in NGC 253 and NGC 4945 appear characteristic of a starburst environment in which massive stars dominate the isotopic composition of the surrounding interstellar medium with isotopes such as 16O and 18O. In contrast, the ISM in the LMC has been much enriched by the products of low-mass AGB stars such as 17O. The various Galactic components are intermediate. The isotopic ratios measured from absorption lines in the z = 0.89 spiral galaxy PKS 1830-211 (see Section 7.2) show the same trends as in starbursts, with very little contribution from low-mass stars, but less enrichment from massive stars than in starburst galaxies.

The interstellar D/H ratio is an indicator of the degree of destruction in stars of the primordial deuterium produced in the Big Bang nucleosynthesis. However, the determination of the D/H ratio from observation of deuterated molecules is difficult because the abundance of deuterated molecules is greatly enhanced by fractionation in the cold and dense gas (Section 2.4). Modelling of deuterium chemistry (see e.g. Roberts & Millar 2000) predicts D/H values as large as 0.01-0.1 in molecules in very cold clouds. Such high ratios which have been observed in various Galactic sources, have been found in the LMC where DCN and DCO+ were detected by Chin et al. (1996) (see also Heikkilä et al., 1997). This result is consistent with a D/H ratio of about 1.5 × 10-5, as observed in the Galaxy. However, this limit is not very constraining because of the uncertainty on the large fractionation enhancement. Upper limits for the DCN abundance were found for other galaxies (see e.g. Mauersberger et al. 1995, Muller et al. 2006).

Table 3. Comparison of the C, N, O and S isotopic ratios in various Galactic and extragalactic environments (reproduced from Table 7 of Muller et al. 2006)

  12C / 13C 14N / 15N 16O / 18O 18O / 17O 32S / 34S

Solar System (a) 89 270 490 5.5 22
Local ISM (a) 59 ± 2 237 -21+27 672 ± 110 3.65 ± 0.15 19 ± 8
Galactic Center (a) 25 ± 5 900 ± 200 250 ± 30 3.5 ± 0.2 18 ± 5
IRC+10216(C-rich AGB)(a) 45 ± 3 > 4400 1260 -240+315 0.7 ± 0.2 21.8 ± 2.6

LMC (b) 62 ± 5 114 ± 14 > 2000 1.8 ± 0.4 18 ± 6
NGC 253 (c) 40 ± 10 - 200 ± 50 6.5 ± 1 8 ± 2
NGC 4945 (d) 50 ± 10 105 ± 25 195 ± 45 6.4 ± 0.3 13.5 ± 2.5
PKS 1830-211(z=0.89)(e) 27 ± 2 130 -15+20 52 ± 4 12 -2+3 10 ± 1

References: a) See references in Table 7 of Muller et al. (2006) for Galactic sources; b) Chin (1999); c) Henkel & Mauersberger (1993), Harrison et al. (1999) and Martín et al. (2005); d) Wang et al. (2004); e) Muller et al. (2006).

Next Contents Previous