Gas phase chemistry
In the gas phase, chemistry is driven by two-body reactions. Due to the very low densities in the ISM, three-body reactions have very low probabilities, and can be thus neglected.
The most important gas-phase process in PDRs involve photons, due to the high far-UV flux. Photodissociation fragments molecules, such as CO, into smaller species. This is the dominant destruction mechanism of the molecules. Far-UV photons can also ionize species, such as C. This way, C+, C and O can remain abundant in the PDR. Deeper inside the cloud, where the radiation field has been attenuated considerably, cosmic rays can still ionize atoms and molecules, producing species such as HCO+, H3O+ and H+ 3 . Cosmic rays can also excite H2, which later de-excites leading to a constant source of secondary far-UV photons, that can then photo-dissociate other molecules. In radiative association reactions, the collision between two species leads to a product that is stabilized through the emission of a photon that takes away the excess energy. One example is the radiative association of C+ with H2, leading to CH+ 2 .
Neutral-neutral reactions are usually exothermic, but they often present activation barriers because the bond between atoms has to be broken in order to rearrange the molecule. Some neutral-neutral reactions involving atoms or radicals, such as C and OH, have no activation energy and are thus important in cold environments. Ion-neutral reactions are faster than the neutral-neutral ones because the dipole moment induced by the ion on the neutral species creates an attraction force, that increases the cross section of the reaction and overcomes any activation barrier. This type of reactions are important in PDRs, because ions are abundant.
Recombination with electrons and charge transfer reactions between species are important for the ionization balance of the gas. One important charge transfer is the one between H+ and O, where H+ is produced by cosmic-ray ionization. The charge transfer of C+ to S is an important source of neutral C in the C+/C/CO transition layers of PDRs. In highly illuminated PDRs, CO can be formed from the charge-transfer between CO+ and H. Chargetransfer reactions between cations, such as C+, S+ and Si+, and PAHs are also important in PDRs, as they can appreciably increase the column density of neutral atoms through the PDR. In the dissociative recombination reactions, an ion captures a free electron and forms a neutral species in an excited electronic state, that can then dissociate into two neutral species. The dissociative electron recombination of CH+ to C can be an important source of C in warm, high-density PDRs. PAH cations can also capture free electrons to recombine. This last process is balanced by the photoelectric effect, where electrons are ejected from the surface of PAHs.
Grain surface chemistry
Dust grains are chemically active in the ISM. Atoms and molecules can be absorbed onto their surfaces, they can move and encounter other atoms or molecules, react with them to form more complex species, and they can be desorbed back into the gas phase.
There are two types of interactions between the atoms and the surface of dust grains. The first one, called physisorption, is a weak force through Van der Waals interactions, with binding energies of ∼ 0.01 eV. The other one, called chemisorption, is a strong force through chemical bonds, with binding energies of ∼ 1 eV. Fig. 1.5 shows the interaction energy profile between an atom and the surface, where the two types of interaction are summarized. Gasphase atoms can easily enter physisorbed sites and become physisorbed atoms. In contrast, in order for a gas-phase atom to enter a chemisorbed site, it first has to overcome a barrier that depends on the gas temperature. Adsorbed atoms can migrate from site to site by tunneling effects and thermal hopping. At low temperatures, physisorbed atoms can easily migrate on the surface of grains to other physisorbed sites. Physisorbed atoms can also migrate to chemisorbed sites, depending on the barrier against chemisorption. At higher temperatures, the physisorbed species evaporate, while chemisorbed species can remain on the surface.
Molecules can be formed on the surface of dust grains through two mechanisms: the Langmuir-Hinshelwood mechanism, that involves physisorbed species, and the Eley-Rideal mechanism, that involves chemisorbed species (see Fig. 1.6). In the Langmuir-Hinshelwood binding through physisorption (Ephys) and through chemisorption (Echem). The dashed line displays a surface with a high barrier against chemisorption, while the continuous line displays a surface with no barrier. Figure from Cazaux & Spaans (2009). mechanism, physisorbed atoms diffuse on the surface, encounter another physisorbed atom and form a molecule. They can also leave the surface through thermal desorption or another desorption process such as photodesorption. In the Eley-Rideal mechanism, an impinging gasphase atom reaches an occupied chemisorbed site and react with it to form a molecule. The formation rate by the Langmuir-Hinshelwood mechanism depends on the dust temperature (allowing diffusion of species on the surface), while the formation rate by the Eley-Rideal mechanism depends on the gas temperature (allowing gas-phase atoms to reach a chemisorbed atom).
Two different environments less than 40” away
Gerin et al. (2009) observed interferometric maps (5′′) of the HCO and H13CO ground state lines towards the Horsehead edge. They detected bright HCO emission delineating the illuminated edge of the nebula, and faint emission toward the shielded molecular cloud. The HCO emission almost coincides with the PAH and CCH emission. At the emission peak, HCO reaches a similar abundance to HCO+ (1 − 2 × 10−9 with respect to H2). At this position, the gas is warm (Tkin ∼ 60 K) and relatively dense (nH ∼ 6 × 104 cm−3). Gerin et al. (2009) proposed that HCO is a good tracer of dense far-UV illuminated gas. On the other side, Pety et al. (2007) detected very bright DCO+ lines in the Horsehead, arising from a dense (nH ≥ 2 × 105 cm−3) and cold (Tkin ≤ 20 K) condensation less than 40′′ away from the UV-illuminated edge of the nebula. This dense core is confined by the pressure from the external far-UV radiation field (Ward-Thompson et al. 2006). A large DCO+/HCO+ of H2 is also given.
Chemistry in the Horsehead
• Carbon chemistry: The carbon chemistry will be described in Chapter 4.
• Sulfur chemistry: The main sulfur bearing species in molecular clouds are S+, S, SO, CS and H2S. Because the observed abundances of these species in dark clouds are too low compared to standard chemical models predictions, it was assumed that sulfurbearing molecules are depleted onto grains in these regions. Goicoechea et al. (2006) studied the emission of CS and HCS+ in the PDR of the Horsehead and found that the sulfur abundance required to reproduce the observations is very close to the solar sulfur elemental abundance. Therefore, they showed that the gas sulfur depletion in the Horsehead PDR is orders of magnitude lower than in previous studies of the sulfur chemistry. This implies that there is something important lacking in the chemical models or that an important sulfur-bearing carrier is missing.
• Electron abundance: The electron abundance ([e−] = ne/nH) plays a fundamental role in the chemistry and dynamics of interstellar gas. Electron can excite molecules with large dipole moments, such as CH3CN. Goicoechea et al. (2009b) investigated the electron abundance gradient across the edge of the Horsehead nebula, using observations of DCO+, H13CO+ and HOC+. They showed that the ionization fraction follows a steep gradient in the edge of the Horsehead, with a scale length of ∼ 0.05 pc (or ∼ 25′′), from [e−] ≃ 10−4 (or ne ∼ 1 − 5 cm−3) in the PDR to a few times ∼ 10−9 in the core. They also confirmed that PAH− anions play a role in the charge balance of the cold and neutral gas if their abundance is significant ([PAH] > 10−8).
The Horsehead WHISPER line survey
The Horsehead WHISPER (Wideband High-resolution Iram-30m Surveys at two Positions with Emir Receivers, PI: J. Pety) survey is a complete and unbiased line survey of the Horsehead. Two positions were observed, namely the PDR and the dense core. The observation of these two different positions enables a detailed comparison of the chemistry of the UVilluminated and UV-shielded gas. The survey covers the 3, 2 and 1 mm bands with an unprecedented combination of bandwidth (36GHz at 3mm, 25GHz at 2mm and 76GHz at 1mm), high spectral resolution (49 kHz at 3 and 2 mm; and 195 kHz at 1 mm), and high sensitivity (median noise 8.1 mK, 18.5 mK and 8.3mK, respectively). The total telescope time used to complete this project is ∼145 hours. detected per band (with S/N > 5) at each positions are shown in the left panel of Fig. 2.6.
Most of the detected lines are in the 3 mm band at both positions. The 2 mm spectrum has the lowest sensitivity of the three spectra and covers a smaller frequency range, which could explain the low number of detected lines compared to the other bands. This is because the 2 mm EMIR reciever at the 30 m was a SSB (Single Side Band, 4 GHz bandwidth) receiver, while the 1 and 3 mm EMIR receivers are 2SB (side band separated mixers, 8 GHz bandwidth) receivers. The 2 mm EMIR receiver has been upgraded to a 2SB mixer at the end of September 2013. From now on, it will thus be possible to do 2 mm line surveys with similar sensitivity to those at 3 and 1 mm, in a similar observing time. In the right panel of density is, on average, 5 lines/GHz (3 mm), 1 lines/GHz (2 mm) and 1 lines/GHz (1 mm) at the PDR, and 4 lines/GHz (3 mm), 1 lines/GHz (2 mm) and 1 lines/GHz (1 mm) at the dense core. It is interesting to note that a similar number of lines are detected at the PDR and dense core positions.
A cumulative distribution of the number of lines as a function of their integrated intensities is shown in Fig. 2.7. The number of lines with integrated intensities larger than 0.1 and 10 Kkms−1 can be roughly fitted by a power law of slope −0.9. The power law breaks down at the lower and higher end of the distribution. The same slope has been found in other millimeter surveys (e.g., Schilke et al. 2001b; Comito et al. 2005; Caux et al. 2011).
At the PDR, the total line flux is 77, 11 and 68 Kkms−1 at 3, 2 and 1 mm, respectively. At the dense core, the total line flux is 88, 20 and 69 Kkms−1 at 3, 2 and 1 mm, respectively. Around 75% of the total line flux is due to the emission of CO and its isotopologues at both positions. The contribution of molecular lines to the total flux at 1.2 mm is estimated to be 14% at the PDR and 16% at the dense core.
The species detected (with S/N > 5) and identified in the survey are listed in Table 2.3. Approximately 30 species (plus their isotopologues) are detected with up to 7 atoms in the PDR and the dense core. In Table 2.3, the species in red are only detected at the PDR, while the species in blue are only detected at the dense core. Species in black are detected at both positions. It is interesting to note that most of the species are detected at both positions. However, for a few species, the emission detected at one of the positions is actually beam pick-up from the other position (e.g., CF+). In addition, for a few species detected at the dense core, the emission arises from the illuminated skin of the cloud and not from the dense core itself (e.g., HCO). A few species are detected only at one of the two positions, and show the different chemistries present in the PDR and dense core. The dense core is characterized by the presence of deuterated species (e.g., CCD, D2CO, DCN and N2D+), while the PDR is characterized by the presence of radicals and reactive species (e.g., HOC+and 13CCH).
CF+ and HF in the interstellar medium
Following the theoretical study of Neufeld et al. (2005), which predicted that CF+ could be abundant enough to be detected in UV-irradiated cloud surfaces, where C+ is the dominant carbon reservoir, the authors searched for the low-energy rotational lines of CF+ toward the Orion Bar. Neufeld et al. (2006) reported the first detection of the J = 1 − 0, J = 2 − 1 and J = 3 − 2 rotational transitions of CF+, and infer a CF+ abundance of ∼ 10−10. This provided support to the theoretical model of fluorine chemistry, and opened the possibility of detecting the J = 1 − 0 transition of HF with Herschel in the following years. HF was detected for the first time in its J = 2 − 1 transition (2.5 THz) by Neufeld et al. (1997). The line was observed in absorption toward the far-infrared continuum source Sgr B2, with the Infrared Space Observatory (ISO) at low spectral resolution. More recently, the HF J = 1 − 0 transition at 1.2 THz has been routinely observed with Herschel/HIFI at high spectral resolution, towards the envelope of the carbon star IRC +10216 (Agúndez et al. 2011), and also in absorption against strong continuum sources (e.g., Neufeld et al. 2010; Sonnentrucker et al. 2010; Monje et al. 2011a; Emprechtinger et al. 2012). These studies revealed that HF is ubiquitous in Galactic molecular clouds and has the potential to become an excellent tracer of molecular gas in diffuse clouds. From observations of 13CO and CH, they estimated the H2 column density, and derived a relative abundance of N(HF)/N(H2) = (1−2)×10−8 in diffuse molecular clouds, with an average value of 1.4×10−8. This value is about two times lower than the abundance predicted by theory of N(HF)/N(H2) = 3.6×10−8, if one assumes the interstellar gas-phase abundance of fluorine to be N(F)/NH = 1.8×10−8.
This last value is the one found in diffuse cloud sight lines, where fluorine is expected to be mainly in its atomic form (Federman et al. 2005; Snow et al. 2007). More recently, Indriolo et al. (2013) derived the abundance ratio N(HF)/N(H2) in diffuse molecular clouds towards three background sources, where the H2 column density was previously determined directly from observations of H2. This allowed them to remove the uncertainties associated to the relation between the column density of H2 and the tracer molecule (CH and CO), and compute a more accurate value for the HF abundance. They derived relative abundances of N(HF)/N(H2) = (0.3 − 1.2) × 10−8, which are well below the predicted value, but in agreement with the previous studies. They also detected HF absorption from warm gas associated with a massive protostar. They derived N(HF)/N(H2) = (1.7 − 2.9) × 10−8 for this component. The observed underabundance of gas-phase HF compared to the predicted value if nearly all gas-phase fluorine is in the form of HF may indicate that 1) the gas-phase abundance of fluorine is lower than expected, 2) not all fluorine is in its molecular form, which could be due to a larger destruction rate of HF, and/or 3) it could also be the result of the freeze-out of HF onto dust grains.
Besides diffuse foreground clouds, HF has been detected in dense molecular clouds, in absorption toward the high-mass star-forming region Orion KL (Phillips et al. 2010) and in NGC 6334 I (Emprechtinger et al. 2012). A lower limit on the HF abundance of N(HF)/N(H2) = (1.6 − 5.0) × 10−10 was found for these sources, which is ∼100 times lower than the HF abundance found in diffuse clouds, suggesting that HF is efficiently depleted onto grains in dense regions.
Table of contents :
1.1 The interstellar medium
1.1.1 Interstellar dust
1.1.2 Molecules in the ISM
1.1.3 The structure of the ISM
1.2 Photon-dominated Regions
1.3 Physical and chemical processes in PDRs
1.3.1 Penetration of far-UV radiation
1.3.2 Thermal balance
1.4 Aims and structure of this thesis
2 The Horsehead nebula as a template PDR
2.1 The Horsehead nebula
2.1.1 A dense PDR seen edge-on
2.1.2 Star formation in the Horsehead
2.1.3 Formation and evolution of the Horsehead
2.1.4 Two different environments less than 40” away
2.1.5 Chemistry in the Horsehead
2.2 Spectral line surveys
2.3 The Horsehead WHISPER line survey
I Simple molecules
3 CF+ Fluoromethylidynium
3.1.1 Fluorine chemistry
3.1.2 CF+ and HF in the interstellar medium
3.1.3 The importance of C+
3.1.4 CF+ hyperfine structure
3.2 Detection of CF+ in the Horsehead
4 C3H+ Propynylidynium
4.1.1 Chemistry of small hydrocarbons
4.1.2 The abundance of hydrocarbons in PDRs
4.2 A new molecule in space: Tentative detection of l-C3H+ in the Horsehead
II Complex molecules
5 H2CO and CH3OH
5.2 Formation of H2CO and CH3OH
5.2.1 Gas-phase chemistry
5.2.2 Grain surface chemistry
5.2.3 Photo-desorption into the gas-phase
5.3 H2CO and CH3OH as tracers of physical properties
5.4 Observations of H2CO and CH3OH in the Horsehead
5.4.1 Determination of column densities
6 CH3CN, CH3NC and HC3N
6.1 Formation of CH3CN
6.1.1 Gas-phase chemistry
6.1.2 Grain surface chemistry
6.2 Nitriles in the Horsehead nebula
7 HCOOH, CH2CO, CH3CO and CH3CCH
7.3 Rotational temperatures and column densities
7.3.4 Formic acid
7.4 Discussion and conclusions
8 Comparison with other environments
8.2 Prototypical PDRs
8.2.2 NGC 7023
8.2.3 Orion Bar
8.2.4 Mon R2
8.3 Other environments
8.3.1 Diffuse medium
8.3.2 Dense core: L1498
8.3.3 Hot corino: IRAS 16293-2422
8.3.4 Hot core: Orion KL
8.3.5 Galactic center GMC: Sgr B2(N)
8.3.6 Starburst galaxy: M82
8.3.7 Active galactic nucleus (AGN): NGC 1068
8.3.8 High-z galaxy: FG0.89
8.4 Abundance ratios
8.4.1 HCO/H13CO+ and CO+/H13CO+
8.4.4 HNC/HCN and HCN/HCO+
8.4.5 C2H/HCO+ and C3H2/HCO+
8.4.7 H2CO, HCOOH, CH2CO, CH3CHO and CH3CCH relative to CH3OH
8.5 Tracers of /n and density in PDRs
8.6 Abundance estimator
9 Conclusions and Perspectives
9.1 The Horsehead: a benchmark to chemical models
9.1.1 CF+: a tracer of C+ and a measure of the fluorine abundance
9.1.2 Detection of a new molecule in space, tentatively attributed to l-C3H+
9.1.3 Photo-desorption of dust grain ice mantles: H2CO and CH3OH
9.1.4 Nitrile molecules: CH3CN, CH3NC and HC3N
9.1.5 Complex molecules in PDRs
9.1.6 Chemical diagnostics
9.2 On the impact of instrumental progresses in radioastronomy on astrochemistry
A Rotational diagrams