Infrared telescopes: Spitzer, Herschel & SOFIA

Get Complete Project Material File(s) Now! »

Phases of the Interstellar Medium

This work is focused on the investigation of the gas components of the ISM, which represent around 99% of the ISM. Most of the gas is hydro-gen (≥71.8%), followed by helium (≥27%), the rest consists of gaseous met-als. The ISM gas structure is not homogeneous; on the contrary it shows a multi-component structure of diﬀerent compositions, densities, tempera-tures, abundances, etc. A rather simplified way to view the ISM, since the dominant element of the gas is hydrogen, is to consider the ISM based on the ionization state of hydrogen: ionized atomic phase, where the hydrogen is in the H+, neutral atomic phase, where the hydrogen is in the neutral atomic form, H, and molecular phase, where hydrogen is in the molecular form, H2. Typical physical properties of the diﬀerent gas phases are summa-rized in Table 1.1 and shown in schematic form in Figure 1.2. More details on the chemistry and physics of the ISM can be found in Tielens (2005) and Osterbrock & Ferland (2005).

Ionized phase

Photons with energies greater than 13.6 eV, produced by hot stars, ionize the surrounding gas creating HII regions. These are normally young stars (Æ15 Myr), typically O and B stars, which link the HII regions to the recent star formation. Therefore, HII regions are often studied to determine the present day chemical abundances (Yin et al. 2010; López-Sánchez et al. 2011; Garnett 1990). The compact HII regions around the ionizing sources are typically characterized by densities of ≥ 105 cm≠3 while the more diﬀuse ionized regions are on the order of ≥ 1 cm≠3. Temperatures of HII regions can be ≥ 102 to 104 K. Their linear scales can be a few parsecs, such as the Orion nebula (D≥8 pc) to hundreds of parsecs, such as NGC5471 (D≥1 kpc; García-Benito et al. 2011).
Ionizing photons escaping from HII regions, ionize the diﬀuse ISM, creat-ing the Warm Ionized Medium (WIM). This component is characterized by very low densities (10≠1 cm3) and can have one of the largest filling factors in galaxies.
The primary heating source of the ionized phases is photoionization. Ion-ized species collide with electrons in the gas and fine structure lines are emit-ted via radiative decay from UV to IR wavelengths. Highly ionized species, such as S3+ and Ne2+, probe the dense HII regions in close proximity to the exciting stars since they require the most energetic photons, while species with lower ionization states, such as N+ and Ne2+ originate in the more extended, diﬀuse ionized gas.
The Hot Ionized Medium (HIM) is characterized by extreme temperatures (≥105 – 106 K) and very diﬀuse gas (≥ 3◊10≠3 cm≠3 ), carrying a small mass fraction of the ISM. This phase is ionized normally by shocks generated by supernovae explosions and stellar winds and emits thermal X-ray continuum and highly-excited emission lines, such as N4+, O6+ and O7+.

Neutral atomic phase

Much of the ISM of galaxies is composed of neutral atomic hydrogen, par-ticularly in disk galaxies. The atomic phase is identified by the dominance of atomic H in the interface phase where the UV photons have energy less than 13.6 eV, but where H2 molecules do not exist. This phase, consisting of neutral atomic species such as H0, C0 and O0, is most easily traced by the 21 cm HI line.
The neutral phase can be divided in two very diﬀerent components: the Cold Neutral Medium (CNM) and the Warm Neutral Medium (WNM). The CNM is a cool, diﬀuse HI cloud with typical temperatures of ≥100 K and densities of ≥50 cm≠3. The WNM is an intercloud gas at temperatures typically of ≥8000 K and densities of ≥0.5 cm≠3.
The neutral interface between the HII region and the molecular core is the PhotoDissociation Region (PDR; Figure 1.2). The physics and chemistry of this transition phase is controlled by FUV photons which have escaped the HII region with 6 < h‹ <13.6 eV. These photons ionize species with ioniza-tion potentials less than 13.6 eV, such as C, Si and S and dissociate molecular hydrogen. Deeper into the cloud the ionizing photons are suﬃciently attenu-ated allowing H2 to exist, forming the dense, molecular core. In dense PDRs this transition between the PDR and the molecular phase occurs near visual extinction, AV (see section 1.2.5), ≥2 magnitudes.
The heating of this phase is primarily via two processes: the photoelectric eﬀect and FUV pumping of H 2. The photoelectric eﬀect occurs when a FUV photon is absorbed by a dust grain and an energetic electron is ejected into the gas, transferring energy to heating the gas. This process is illustrated in Figure 1.3. The eﬃciency of the photoelectric heating depends on the size and the charge of the grains. The eﬃciency increases inversely with grain size. Hence the most eﬃcient dust components are the smallest grains, the polycyclic aromatic hydrocarbons (PAHs) and other small grains. The charge of the grains is given by the balance between the photoionization and the electron recombination, which is controlled by “ = G0T 1/2/ne, where G0 is the radiation field (Tielens & Hollenbach 1985) and ne is the electron density. The second source of heating in the PDRs is the FUV pumping of H2. The molecule H 2 absorbs photons via the Lyman-Werner electronic transitions (ΔE > 11.2 and 12.3 eV). Most of the time the H2 cascades to an excited vibrational level in the ground electronic state via fluorescence while 10 to 15% of the FUV pumping events, the molecule will dissociate via fluorescence to the ground electronic state.
The cooling of the PDR gas is due mainly to IR fine-structure line emis-sion from species such as [CII] 157.7 µm , [OI] 63.2 µm and [OI] 145.5 µm .

Molecular phase

The molecular phase is the densest component of the ISM and exists where H2 dominates the gas phase. While H2 is the most abundant molecule, other molecules are forming within this phase, shielded from dissociating photons, such as CO, HCN, O2. The molecular phase is characterized by densities > 102 cm≠3 and temperatures on the order of ≥10 K.
Molecules can form on the surface of dust grains and can be destroyed by UV photons if their self-shielding eﬃciency or the shielding from the dust is insuﬃcient. For example, for H2 the Lyman-Werner bands become optically thick at high density; hence the molecule can be shielded from starlight. On the other hand, CO, which is the second most abundant molecule, is not as eﬃciently self-shielded, but relies on the dust to shield it from the dissociating photons.
FUV photons and cosmic rays are the main heating sources of molec-ular clouds. Molecules, excited by collisions and shocks, cool the gas via vibrational and rotational transitions emitted in the FIR to submillimeter (submm) wavelength range. A prominent coolant in the molecular gas phase is CO in its various rotational transitions.

Observational tracers

The state of the ionized, PDR, and molecular phases can be derived by studying the emission lines, which emit at various wavelengths. Baldwin et al. (1981), for example, proposed the [OIII]⁄5007/H— and [NII]⁄6584/H– diagnostic diagram (the BPT diagram) to separate the line-emission objects based on the excitation mechanism: HII regions, planetary nebulae, and shock heated media, such as active galactic nuclei (AGN). Moreover, even simple line ratios can sometimes characterize the physical properties of the emitted gas. For example, the ratio of lines of the same element but diﬀerent ionization stage, such as [NeIII] 15.5 µm /[NeII] 12.8 µm , traces the tem-perature of the ionizing source, while the ratio of lines of the same ion but diﬀerent transition, can estimate the electron density (Osterbrock & Ferland 2005; Abel et al. 2005). Consider two levels with the same excitation energy but with diﬀerent Einstein A coeﬃcients (rate of spontaneous emission and absorption) or collisional de-excitation rates, the excitation of these levels will be related only by diﬀerent densities. Examples of ratios that can be used to measure the gas electron density are [SII]6716/[SII]6731 Å in the optical and [OIII]88/[OIII]52 µm in the IR. Figure 1.4 shows examples of in-frared line ratios that trace diﬀerent density regimes. Each line ratio, in fact, can trace the density of the gas when the value of the density lies between the critical densities of the two lines. The critical density is the density for which the collisional rate coeﬃcient is equal to the radiative de-excitation co-eﬃcient. For densities higher that this threshold density, collisions dominate the de-excitation process.
Unlike the optical lines, infrared observations are slightly eﬀected by ex-tinction (see Section 1.2.5) and they cover a wide range of excitation poten-tials and critical densities (Figure 1.5; Kennicutt et al. 2011). The wave-lengths, excitation potentials and critical densities of the MIR and FIR lines observed with Spitzer and Herschel space telescopes (see Chapter 2), are pre-sented in Table 1.2.
The MIR fine-structure lines, observed by Spitzer, are diagnostics mostly of the compact ionized gas (Spinoglio et al. 2015; Abel et al. 2005), due to their high critical density and excitation potential. In the following I present the MIR tracers that I have used in my PhD study.
• [ArII] 6.9 µm and [ArIII] 8.9 µm
Ar0 and Ar+ have ionization potentials of 15.7 and 27.6 eV, respec-tively, and critical densities for collisions with electrons of 4 ◊ 105 cm≠3 for [ArII] 6.9 µm and 3 ◊ 105 cm≠3 for [ArIII] 8.9 µm . Since these two lines are emitted by two diﬀerent ionization stages of the same element, the ratio [ArIII] 8.9 µm /[ArII] 6.9 µm traces the hard-ness of the radiation field. [ArIII] 8.9 µm , requiring the higher energy to ionize, would originate closer to the radiation source in the dense HII regions than [ArII] 6.9 µm , which would be emitted more from the outer shell of the HII regions. Therefore, the ratio [ArIII] 8.9 µm /[ArII] 6.9 µm can trace the leakage of ionizing photons from the HII regions (Section 4.3.2).
• [NeII] 12.8 µm and [NeIII] 15.5 µm
Of the two MIR neon lines, the [NeIII] 15.5 µm has a much higher excitation potential, requiring 41 eV photons to ionize the Ne+ while Ne0, requires 21.6 eV to create Ne+ (Table 1.2). The [NeII] 12.8 µm and [NeIII] 15.5 µm are exclusively found in HII regions. The critical density of [NeII] 12.8 µm for collisions with electrons is 7 ◊ 105 cm≠3, while for [NeIII] 15.5 µm the critical density with electrons is lower, 3 ◊ 105 cm≠3. However, while [NeIII] 15.5 µm is associated only with dense, hot HII regions, [NeII] 12.8 µm can arise from HII regions as well as relatively diﬀuse low-ionization gas (Cormier et al. 2012; Dimaratos et al. 2015). Since these two lines are emitted by diﬀerent ionization stages of the same element, the ratio [NeIII] 15.5 µm/[NeII] 12.8 µm traces the hardness of the radiation field.
• [SIII] 18.7 µm , [SIII] 33.5 µm and [SIV] 10.5 µm
S+ has an ionization potential of 23.3 eV and traces the ionized gas. The two [SIII] transitions at 18.7 µm and 33.5 µm have critical den-sities with electrons of 2 ◊ 104 cm≠3 and 7 ◊ 103 cm≠3, respectively. Hence the ratio [SIII] 33.5 µm/[SIII] 18.7 µm is a useful density tracer. S2+, has an ionization potential of 34.7 eV and [SIV] has a critical density for collisions with electrons of 5 ◊ 104 cm≠3, tracing the dense hot HII regions. The combination of [SIV] 10.5 µm with [SIII] 18.7 µm , as well as [SIII] 33.5 µm , provides useful ratios to measure the hardness of the radiation field.
• [FeII] 25.9 µm and [SiII] 34.8 µm
Fe+ and Si+ ions require ionizing photons lower than that of hydro-gen (7.9 eV to ionize Fe0 and 8.2 eV to ionize Si0; see Figure 1.5). Therefore, Fe+ and Si+ may exist in the neutral gas, excited mostly by H0, H2 and free electrons, as well as in the ionized gas, excited by elec-trons. In order to understand which phase these lines trace, modeling of the full suite of lines is required.

READ Environmental fiscal incentives: Effectiveness or free-riding effect? An econometric evaluation of the French energy tax credit

FIR fine-structure lines

The FIR fine-structure lines are important diagnostics of the properties of the ionized phase as well as of the PDR phase (Wolfire et al. 1990; Kaufman et al. 2006). They have been observed in the past with ISO and the KAO, in more recent past with Herschel and now with SOFIA.
• [OIII] 88.4 µm
[OIII] 88.4 µm has an excitation potential of 35.1 eV and a critical density 500 cm≠3, thus arises only from ionized gas. The combination of [OIII] 88.4 µm with [OIII] 52 µm, is a useful constraint of electron density the HII region (Lebouteiller et al. 2012). [OIII] 52 µm was not accessible with Herschel, but can now be observed with SOFIA (Sec-tion 2.4.3).
• [CII] 157.7 µm
[CII] 157.7 µm is one of the most important cooling lines of the ISM. It has been studied in many diﬀerent environments, including Galactic PDRs (Bennett et al. 1994), dwarf galaxies (e.g. Madden et al. 1997; Cormier et al. 2015; Fahrion et al. 2016), spiral galaxies (e.g. Kapala et al. 2017), high-redshift galaxies (Neeleman et al. 2017). Usually it is the brightest FIR emission line in galaxies, which can sometimes be used to determine the redshift of galaxies (e.g. Bradačet al. 2017) and its luminosity can be used as a measure of the star formation rate in galaxies (e.g. De Looze et al. 2011).
The ionization potential of C0 is 11.3 eV, less than that of hydrogen (13.6 eV). Thus C+, in principle, may exist in the ionized phase, excited by collisions with electrons with a critical density of 50 cm ≠3, as well as in the neutral phase, excited by H or H2 with a critical density of 2.8◊103cm≠3. The potential for [CII] 157.7 µm to exist in diﬀerent phases complicates the interpretation of this popular diagnostic. In our Galaxy, [CII] 157.7 µm has been found to trace the surface layers of PDRs. For example, Bernard-Salas et al. (2012) have found that in the Orion Bar > 82% of the [CII] emission is coming from the PDR region. Similar conclusions have been determined for [CII] 157.7 µm observed around the 30Doradus region in the Large Magellanic Cloud (Chevance et al. 2016). However [CII] 157.7 µm has also been observed to originate in the ionized gas phase (Madden et al. 1993; Abel et al. 2005; Cormier et al. 2012).
There are several methods to disentangle the origin of [CII] 157.7 µm. It can be done through detailed modeling (e.g. Bernard- Salas et al. 2012; Cormier et al. 2012) or by inspecting key diagnostic line ratios, such as [CII] 157.7 µm/[NII] 121.9 µm as explained below (e.g. Oberst et al. 2011).
• [NII] 121.9 µm and [NII] 205.2 µm
N+ with an excitation potential of 14.5 eV originates only in the ionized phase. The two FIR transitions, [NII] 121.9 µm and [NII] 205.2 µm, with relatively low critical densities of 300 cm≠3 and 45 cm≠3, respec-tively, when used as a ratio, become a useful electron density tracer of the diﬀuse ionized gas, as is shown in Figure 1.6. Since these lines can only arise from the ionized phase while the origin of [CII] 157.7 µm is ambiguous, the ratios [NII] 121.9 µm/[CII] 157.7 µm and [NII] 205.2 µm/[CII] 157.7 µm can be used to disentangle the origin of [CII] 157.7 µm . In particular, since [CII] 157.7 µm and [NII] 205.2 µm have very similar critical densities and excitation temperatures, the ratio of [CII]/[NII] depends mostly on the abundances of N+ and C+ (Oberst et al. 2006; Oberst et al. 2011; Bernard-Salas et al. 2012).
Figure 1.6: Theoretical ratios of [NII]122/205µm, [CII]/[NII]205µm and [CII]/[NII]122µm as a function of the density for a temperature of 9000 K. (Figure from Bernard-Salas et al. 2012).
• [OI] 63.2 µm and [OI] 145.5 µm
The two [OI] lines at 63.2 and 145.5 µm, together with [CII] 157.7 µm , are the dominant coolants in PDRs. These two lines have critical densities of 5◊105 and 105 cm≠3, respectively and excitation energies of 228 and 325 K. The ratio [OI] 145.5 µm /[OI] 63.2 µm is a popular diagnostic used to determine the physical conditions in PDRs (Kauf-man et al. 1999). In particular, this ratio can be used to estimate the density and temperature of the neutral gas: the ratio increases with increasing temperature and decreases with decreasing density (Tielens & Hollenbach 1985). The [OI] 63.2 µm is one of the brightest cool-ing lines (Oberst et al. 2011; Bernard-Salas et al. 2012; Cormier et al. 2015) but is known to be aﬀected by optical depth eﬀects (Tielens & Hollenbach 1985; Abel et al. 2007; Chevance et al. 2016). The [OI] 145.5 µm line, however, is fainter than the 63.2 µm line, often by at least an order of magnitude, making it a challenge, even for Herschel, to map large regions in galaxies.
• H2
H2 is the most abundance molecule in galaxies. However, it is a sym-metric molecule and does not have a permanent dipole moment, making this molecule diﬃcult to observe. Nevertheless, its rotational quadrupole transitions (ΔJ = ±2; defining J as the rotational quantum number) have been observed in the MIR with Spitzer. These lines (3.4 to 28µm) are pure rotational transitions (Δ‹= 0), while the NIR lines (1 to 4µm) are vibrational transitions (Δ‹= ±1). However, these transitions need excitation temperatures > 500 K, hence they trace a warm neutral phase instead of the bulk of the cold molecular gas found in the star-forming disks of galaxies, for example.
H2 can also be observed in absorption in the UV in the Lyman-Werner band, from warm as well as cold diﬀuse gas heated by UV sources. It can also be a prominent emission line diagnostic in shocked regions tracing the turbulent cascade throughout the ISM (e.g. Guillard et al. 2009; Guillard et al. 2010)
This molecule has been observed in a variety of objects, including star-forming regions and PDRs in our galaxy (e.g. Tielens & Hollenbach 1993; Lebouteiller et al. 2006), and in other galaxies (Roussel et al. 2007; Naslim et al. 2015).
• CO
The H2 reservoir is important to quantify since it is the fuel for star formation. Since the cold H2 is diﬃcult to observe, the usual tracers of the molecular phase are the CO emission lines. The CO rotational transitions emit in the FIR to submm wavelengths, excited by collisions with H2. The CO rotational transitions can sometimes, arbitrarily, be divided into low-J transitions, namely J = 1-0, 2-1 and J=3-2; and high-J transitions where J = 3-2, 4-3, 5-4 and higher. Since the criti-cal densities and the excitation temperatures increase with increasing transition level, the low-J transitions normally trace the bulk of the cold molecular gas, while the high-J transitions trace warmer molec-ular gas. The most commonly detected lines are 12CO (1-0) and the isotopic version, 13CO (1-0) (e.g. Wilson et al. 2009; Chevance et al. 2016). However, since CO forms deep inside the cloud, its lower rota-tional transitions can be aﬀected by optical depth eﬀects, in particular 12CO(J =1 -0) transition is often optically thick.

Dust phase

Dust grains are formed in the envelopes of evolved stars as well as in novae and supernovae and are then ejected into the ISM by stellar winds and su-pernovae, which can also destroy them in the associated shock waves. Even though the dust represents only ≥1% of the total ISM mass, it is a fundamen-tal component of the ISM. Dust grains provide the surfaces for the accretion and reaction of species that lead to the formation of molecules. They are also responsible for the stellar light extinction at wavelengths longer than 912Å, through the absorption and scattering of radiation, and its re-emission in the infrared.

Table of contents :

1 The Interstellar Medium of Dwarf Galaxies
1.1 Galaxy
1.2 Phases of the Interstellar Medium
1.2.1 Ionized phase
1.2.2 Neutral atomic phase
1.2.3 Molecular phase
1.2.4 Observational tracers
1.2.5 Dust phase
1.3 Metallicity
1.4 Dwarf galaxies
1.4.1 Stellar population
1.4.2 Star formation
1.4.3 Cloud structure
1.4.4 Dust properties
1.5 Dwarf Galaxy Survey
1.5.1 DGS analysis
1.5.2 Dust properties
1.5.3 Gas properties
1.6 The dwarf irregular galaxy IC10
2 Infrared telescopes: Spitzer, Herschel & SOFIA
2.1 Infrared observations
2.2 The Spitzer Space Telescope
2.2.1 Spitzer Mission
2.2.2 IRAC
2.2.3 MIPS
2.2.4 IRS
2.2.5 Data reduction – Spitzer/IRS
2.3 The Herschel Space Observatory
2.3.1 Herschel Mission
2.3.2 SPIRE
2.3.3 PACS
2.3.4 Data reduction – Herschel/PACS
2.4 SOFIA Telescope
2.4.1 SOFIA overview
2.4.2 FIFI-LS
2.4.3 SOFIA IC 10 observations
3 Spitzer and Herschel Observations of IC 10
3.1 Dataset
3.1.1 Spitzer/IRS maps
3.1.2 Herschel maps
3.2 Spatial distribution of ISM tracers
3.2.1 Available tracers
3.2.2 MIR and FIR ratios
3.2.3 Conclusion
4 Modeling the ISM physical properties with Cloudy
4.1 State-of-the-art modelling
4.1.1 Radiative transfer theory and energy balance
4.1.2 Photoionization models
4.2 Photoionization and photodissociation code: CLOUDY
4.2.1 Overview
4.2.2 Input parameters
4.3 Cloudy models applied to IC10
4.3.1 Setting the input parameters
4.3.2 Stopping criteria
5 Modeling the ionized gas at different spatial scales
5.1 Tracers used
5.2 Spatial decomposition
5.2.1 Description of the spatial scales used
5.2.2 Disentangling the “clumps”
5.3 Building a set of observational constraints for the models
5.3.1 Foreword on the available methods
5.3.2 Line Ratio Method
5.3.3 Absolute Flux Method
5.4 Conclusion