Diffuse ISM Research Articles

Planckintermediate results

We present all-sky modelling of the high resolution Planck, IRAS, and WISE infrared (IR) observations using the physical dust model presented by Draine and Li in 2007 (DL). We study the performance and results of this model, and discuss implications for future dust modelling. The present work extends the DL dust modelling carried out on nearby galaxies using Herschel and Spitzer data to Galactic dust emission. We employ the DL dust model to generate maps of the dust mass surface density, the optical extinction Av, and the starlight intensity parametrized by Umin. The DL model reproduces the observed spectral energy distribution (SED) satisfactorily over most of the sky, with small deviations in the inner Galactic disk and in low ecliptic latitude areas. We compare the DL optical extinction Av for the diffuse interstellar medium with optical estimates for 2 10^5 quasi-stellar objects (QSOs) observed in the Sloan digital sky survey. The DL Av estimates are larger than those determined towards QSOs by a factor of about 2, which depends on Umin. The DL fitting parameter Umin, effectively determined by the wavelength where the SED peaks, appears to trace variations in the far-IR opacity of the dust grains per unit Av, and not only in the starlight intensity. To circumvent the model deficiency, we propose an empirical renormalization of the DL Av estimate, dependent of Umin, which compensates for the systematic differences found with QSO observations. This renormalization also brings into agreement the DL Av estimates with those derived for molecular clouds from the near-IR colours of stars in the 2 micron all sky survey. The DL model and the QSOs data are used to compress the spectral information in the Planck and IRAS observations for the diffuse ISM to a family of 20 SEDs normalized per Av, parameterized by Umin, which may be used to test and empirically calibrate dust models.

Extinction and dust properties in a clumpy medium

(abridged) The dust content of the universe is primarily explored via its interaction with stellar photons, producing interstellar extinction. However, owing to the physical extension of the observing beam, observations may detect scattered photons, resulting in a change in the observed (or effective) extinction, depending on the spatial distribution of the dust and the resolution of the instrument. We investigate the influence of clumpy dust distributions on effective extinction toward embedded sources and those in the diffuse ISM. We use Monte Carlo radiative transfer to examine effective extinction for various geometries. By varying the number, optical depth and volume-filling factor of clumps in models of spherical shells and the diffuse ISM, we explore the evolution of extinction. Depending on the number of scattering events in the beam, the extinction curve steepens in homogeneous media and flattens in clumpy media. As a result, clumpy dust distributions can to reproduce extinction curves with arbitrary R_v, the effective ratio of total-to-selective extinction. The flattening `washes out' the 2175 \AA bump and shifts the peak to shorter wavelengths. The mean R_v of a shell correlates with the optical depth of an individual clump and the wavelength at which a clump becomes optically thick. Similar behaviour is seen for edge-on discs or tori. However, at grazing inclinations the combination of extinction and strong forward scattering results in chaotic behaviour. Caution is therefore advised when measuring extinction in, for example, AGN tori or toward SNIa or GRB afterglows. In face-on discs, the shape of the scattered continuum changes with clumpiness, however, unlike absorption features, individual features in the scattering cross-sections are preserved. Finally, we show that diffuse interstellar extinction is not modified by scattering on scales of a few kpc.

Large dust grains in the wind of VY Canis Majoris

Massive stars live short lives, losing large amounts of mass through their stellar wind. Their mass is a key factor determining how and when they explode as supernovae, enriching the interstellar medium with heavy elements and dust. During the red supergiant phase, mass-loss rates increase prodigiously, but the driving mechanism has proven elusive. Here we present high-contrast optical polarimetric-imaging observations of the extreme red supergiant VY Canis Majoris and its clumpy, dusty, mass-loss envelope, using the new extreme-adaptive-optics instrument SPHERE at the VLT. These observations allow us to make the first direct and unambiguous detection of submicron dust grains in the ejecta; we derive an average grain radius $\sim$ 0.5 $\mu$m, 50 times larger than in the diffuse ISM, large enough to receive significant radiation pressure by photon scattering. We find evidence for varying grain sizes throughout the ejecta, highlighting the dynamical nature of the envelope. Grains with 0.5 $\mu$m sizes are likely to reach a safe distance from the eventual explosion of VY Canis Majoris; hence it may inject upwards of 10$^{-2}$ M$_\odot$ of dust into the ISM.

SPECTROSCOPIC INFRARED EXTINCTION MAPPING AS A PROBE OF GRAIN GROWTH IN IRDCs

We present spectroscopic tests of MIR to FIR extinction laws in IRDC G028.36+00.07, a potential site of massive star and star cluster formation. Lim & Tan (2014) developed methods of FIR extinction mapping of this source using ${\it Spitzer}$-MIPS ${\rm 24\mu m}$ and ${\it Herschel}$-PACS ${\rm 70\mu m}$ images, and by comparing to MIR ${\it Spitzer}$-IRAC $3$--${\rm 8\mu m}$ extinction maps, found tentative evidence for grain growth in the highest mass surface density regions. Here we present results of spectroscopic infrared extinction (SIREX) mapping using ${\it Spitzer}$-IRS (14 to ${\rm 38\mu m}$) data of the same IRDC. These methods allow us to first measure the SED of the diffuse Galactic ISM that is in the foreground of the IRDC. We then carry out our primary investigation of measuring the MIR to FIR opacity law and searching for potential variations as a function of mass surface density within the IRDC. We find relatively flat, featureless MIR-FIR opacity laws that lack the $\sim{\rm 12\mu m}$ and $\sim{\rm 35\mu m}$ features associated with the thick water ice mantle models of Ossenkopf & Henning (1994). Their thin ice mantle models and the coagulating aggregate dust models of Ormel et al. (2011) are a generally better match to the observed opacity laws. We also find evidence for generally flatter MIR to FIR extinction laws as mass surface density increases, strengthening the evidence for grain and ice mantle growth in higher density regions.

Self-consistent two-phase AGN torus models

We assume that dust near active galactic nuclei (AGN) is distributed in a torus-like geometry, which may be described by a clumpy medium or a homogeneous disk or as a combination of the two (i.e. a 2-phase medium). The dust particles considered are fluffy and have higher submillimeter emissivities than grains in the diffuse ISM. The dust-photon interaction is treated in a fully self-consistent three dimensional radiative transfer code. We provide an AGN library of spectral energy distributions (SEDs). Its purpose is to quickly obtain estimates of the basic parameters of the AGN, such as the intrinsic luminosity of the central source, the viewing angle, the inner radius, the volume filling factor and optical depth of the clouds, and the optical depth of the disk midplane, and to predict the flux at yet unobserved wavelengths. The procedure is simple and consists of finding an element in the library that matches the observations. We discuss the general properties of the models and in particular the 10mic. silicate band. The AGN library accounts well for the observed scatter of the feature strengths and wavelengths of the peak emission. AGN extinction curves are discussed and we find that there is no direct one-to-one link between the observed extinction and the wavelength dependence of the dust cross sections. We show that objects of the library cover the observed range of mid IR colors of known AGN. The validity of the approach is demonstrated by matching the SEDs of a number of representative objects: Four Seyferts and two quasars for which we present new Herschel photometry, two radio galaxies, and one hyperluminous infrared galaxy. Strikingly, for the five luminous objects we find pure AGN models fit the SED without a need to postulate starburst activity.

The CH+abundance in turbulent, diffuse molecular clouds

The intermittent dissipation of interstellar turbulence is an important energy source in the diffuse ISM. Though on average smaller than the heating rates due to cosmic rays and the photoelectric effect on dust grains, the turbulent cascade can channel large amounts of energy into a relatively small fraction of the gas that consequently undergoes significant heating and chemical enrichment. In particular, this mechanism has been proposed as a solution to the long-standing problem of the high abundance of CH+ along diffuse molecular sight lines, which steady-state, low temperature models under-produce by over an order of magnitude. While much work has been done on the structure and chemistry of these small-scale dissipation zones, comparatively little attention has been paid to relating these zones to the properties of the large-scale turbulence. In this paper, we attempt to bridge this gap by estimating the temperature and CH+ column density along diffuse molecular sight-lines by post-processing 3-dimensional MHD turbulence simulations. Assuming reasonable values for the cloud density (30 / cm^3), size (20 pc), and velocity dispersion (2.3 km / s), we find that our computed abundances compare well with CH+ column density observations, as well as with observations of emission lines from rotationally excited H2 molecules.

INTERSTELLAR DUST PROPERTIES FROM A SURVEY OF X-RAY HALOS

Interstellar dust grains produce X-ray halos around bright sources due to small-angle X-ray scattering. Numerous studies have examined these halos, but no systematic study has yet tested the available halo data against the large number of well-defined dust models in circulation. We have therefore obtained the largest sample to date of X-ray dust halos from XMM-Newton and Chandra, and fitted them with 14 commonly used dust grain models, including comparisons with the optical extinction, AV, where available in the literature. Our main conclusions are summarized as follows. (1) Comparing AV with NH values measured via X-ray spectral fits, we find a ratio of AV/NH (1021 cm?2) = 0.48 ? 0.06, in agreement with previous work. (2) Out of 35 halos, 27 could be fit by one or more grain models, with the most successful models having maximum grain radius ?m and fewer large grains than the less successful models. This suggests that the diffuse ISM does not contain a signicant presence of grains with ?m. (3) Most halos were best fit assuming a single dust cloud dominated the scattering, rather than smoothly distributed dust along the sightline. (4) Eight sources could not be fit with the models considered here, most of which were along distant ( kpc) sight lines through the Galactic thin disk. (5) Some sight lines had halos with observed X-ray scattering optical depth ?sca/AV that were signicantly different than expected. This may result from an inhomogeneous dust distribution across the halo extraction area.

Quantifying the Magnetic Alignment of Hi and Dust in the Diffuse ISM

AbstractSensitive, high resolution observations of Galactic neutral hydrogen (Hi) reveal an intricate network of slender linear features, much as sensitive surveys of dust in Galactic molecular clouds reveal ubiquitous filamentary structure. Across the high Galactic latitude sky, diffuse Histructures are aligned with the interstellar magnetic field, as revealed by background starlight polarization (Clark, Peek, & Putman 2014) and by Planck 353 GHz polarized dust emission (Clark et al. 2015). These discoveries were enabled by the Rolling Hough Transform, a recently developed, open source machine vision algorithm.

Polarized thermal dust emission as seen by Planck : A comparison with MHD simulations and lessons from a toy model

AbstractThe Planck satellite has mapped the polarized microwave sky (from 30 GHz to 353 GHz) with unprecedented sensitivity and angular resolution. This wealth of data yields the first complete map of polarized thermal emission from dust in our own Galaxy, shedding new light on the formation of dense cold structures within which new stars and planetary systems are born, under the combined effects of gravity, turbulence and magnetic fields. We present a statistical analysis of this polarized emission from nearby molecular clouds, with an emphasis on the evolution of the maximum polarization fraction observed as a function of column density, and on the anti-correlation between the polarization fraction and the local dispersion of polarization angles. To interpret this data, numerical simulations of anisotropic MHD turbulence underline the essential role played by the topology of the interstellar magnetic field, in particular its large-scale component. Indeed, the polarization of dust thermal emission at the scales observed by Planck is essentially related to the geometry of the magnetic field. Polarization fractions anti-correlate with column densities, which may be due to a succession of variously polarized structures on the line of sight. They also anti-correlate with the local dispersion of polarization angles. These features are well reproduced by MHD simulations of the diffuse ISM, with comparable correlation coefficients. As an extension to this work published in Planck Intermediate Results XX (A&A, 576, 105, 2015), the statistical properties of the random component of the interstellar magnetic field are explored using a toy model of the turbulent magnetized ISM based on fractional Brownian motion (fBm) fields. A least-squares analysis to retrieve the statistical properties of the interstellar magnetic field from Planck observations is pursued. Application of this method on the toy model shows good promise, and we are currently working towards its application on Planck data.

Polarized thermal dust emission as seen by Planck : A comparison with MHD simulations and lessons from a toy model

AbstractThe Planck satellite (Planck 2015 Results I) has mapped the polarized microwave sky (from 30 GHz to 353 GHz) with unprecedented sensitivity and angular resolution. This wealth of data yields the first complete map of polarized thermal emission from dust in our own Galaxy (Planck Intermediate Results XIX),, shedding new light on the formation of dense cold structures within which new stars and planetary systems are born, under the combined effects of gravity, turbulence and magnetic fields. We present a statistical analysis of this polarized emission from nearby molecular clouds, with an emphasis on the evolution of the maximum polarization fraction observed as a function of column density, and on the anti-correlation between the polarization fraction and the local dispersion of polarization angles. To interpret this data, numerical simulations of anisotropic MHD turbulence (Fromang, Hennebelle, & Teyssier 2006, Hennebelle et al. 2008) underline the essential role played by the topology of the interstellar magnetic field, in particular its large-scale component (Planck Intermediate Results XX). Indeed, the polarization of dust thermal emission at the scales observed by Planck is essentially related to the geometry of the magnetic field. Polarization fractions anti-correlate with column densities, which may be due to a succession of variously polarized structures on the line of sight. They also anti-correlate with the local dispersion of polarization angles. These features are well reproduced by MHD simulations of the diffuse ISM, with comparable correlation coefficients. As an extension to this work published in Planck Intermediate Results XX, the statistical properties of the random component of the interstellar magnetic field are explored using a toy model of the turbulent magnetized ISM based on fractional Brownian motion (fBm) fields. A least-squares analysis to retrieve the statistical properties of the interstellar magnetic field from Planck observations is pursued. Application of this method on the toy model shows good promise, and we are currently working towards its application on Planck data.

Colours and luminosities ofz = 0.1 galaxies in the eagle simulation

We calculate the colours and luminosities of redshift z = 0.1 galaxies from the EAGLE simulation suite using the GALAXEV population synthesis models. We take into account obscuration by dust in birth clouds and diffuse ISM using a two-component screen model, following the prescription of Charlot and Fall. We compare models in which the dust optical depth is constant to models where it depends on gas metallicity, gas fraction and orientation. The colours of EAGLE galaxies for the more sophisticated models are in broad agreement with those of observed galaxies. In particular, EAGLE produces a red sequence of passive galaxies and a blue cloud of star forming galaxies, with approximately the correct fraction of galaxies in each population and with g-r colours within 0.1 magnitudes of those observed. Luminosity functions from UV to NIR wavelengths differ from observations at a level comparable to systematic shifts resulting from a choice between Petrosian and Kron photometric apertures. Despite the generally good agreement there are clear discrepancies with observations. The blue cloud of EAGLE galaxies extends to somewhat higher luminosities than in the data, consistent with the modest underestimate of the passive fraction in massive EAGLE galaxies. There is also a moderate excess of bright blue galaxies compared to observations. The overall level of agreement with the observed colour distribution suggests that EAGLE galaxies at z = 0.1 have ages, metallicities and levels of obscuration that are comparable to those of observed galaxies.

ULTRAFAINT DWARF GALAXIES—THE LOWEST-MASS RELICS FROM BEFORE REIONIZATION

New observations indicate that ultrafaint dwarf galaxies (UFD) -- the least luminous systems bound by dark matter halos (<10^5 Lsun) -- may have formed before reionization. The extrapolated virial masses today are uncertain with estimates ranging from 10^8 Msun to 10^9 Msun. We show that the progenitor halo masses of UFDs can be as low as Mvir = 10^7 Msun. Under the right conditions, such a halo can survive the energy input of a supernova and its radiative progenitor. A clumpy medium is much less susceptible to both internal and external injections of energy. It is less prone to SN sweeping because the coupling efficiency of the explosive energy is much lower than for a diffuse ISM. With the aid of the 3D hydro/ionization code Fyris, we show that sufficient baryons are retained to form stars following a single supernova event in dark matter halos down to Mvir ~ 10^7 Msun with radiative cooling. The gas survives the SN explosion, is enriched with the abundance yields of the discrete events, and reaches surface densities where low mass stars can form. Our highest resolution simulations reveal why cooling is so effective in retaining gas compared to any other factor. In the early stages, the super-hot metal-enriched SN ejecta exhibit strong cooling, leading to much of the explosive energy being lost. Consistent with earlier work, the baryons do *not* survive in smooth or adiabatic models in the event of a supernova. The smallest galaxies carry signatures of the earliest epochs of star formation, which may distinguish a small primordial galaxy from one that was stripped down to its present size through tidal interaction. We discuss these results in the context of local UFDs and damped Ly-alpha systems (z~2) at very low metallicity ([Fe/H] ~ -3). We show that both classes of objects are consistent with primordial low-mass systems that have experienced only a few enrichment events.

Understanding the two-dimensional ionization structure in luminous infrared galaxies

We investigate the 2D excitation structure of the ISM in a sample of LIRGs and Seyferts using near-IR IFS. This study extends to the near-IR the well-known optical and mid-IR emission line diagnostics used to classify activity in galaxies. Based on the spatially resolved spectroscopy of prototypes, we identify in the [FeII]1.64/Br$\gamma$ - H_2 1-0S(1)/Br$\gamma$ plane regions dominated by the different heating sources, i.e. AGNs, young MS massive stars, and evolved stars i.e. supernovae. The ISM in LIRGs occupy a wide region in the near-IR diagnostic plane from -0.6 to +1.5 and from -1.2 to +0.8 (in log units) for the [FeII]/Br$\gamma$ and H_2/Br$\gamma$ line ratios, respectively. The corresponding median(mode) ratios are +0.18(0.16) and +0.02(-0.04). Seyferts show on average larger values by factors ~2.5 and ~1.4 for the [FeII]/Br$\gamma$ and H_2/Br$\gamma$ ratios, respectively. New areas and relations in the near-IR diagnostic plane are defined for the compact, high surface brightness regions dominated by AGN, young ionizing stars, and SNe explosions, respectively. In addition, the diffuse regions affected by the AGN radiation field cover an area similar to that of Seyferts, but with high values in [FeII]/Br$\gamma$ that are not as extreme. The extended, non-AGN diffuse regions cover a wide area in the diagnostic diagram that overlaps that of individual excitation mechanisms (i.e. AGN, young stars, and SNe), but with its mode value to that of the young SF clumps. This indicates that the excitation conditions of the diffuse ISM are likely due to a mixture of the different ionization sources. The integrated line ratios in LIRGs show higher excitation conditions i.e. towards AGNs, than those measured by the spatially resolved spectroscopy. If this behaviour is representative, it would have clear consequences when classifying high-z, SF galaxies based on their near-IR integrated spectra.

Dust variations in the diffuse interstellar medium: constraints on Milky Way dust fromPlanck-HFI observations

The Planck-HFI all-sky survey from 353 to 857GHz combined with the 100 microns IRAS show that the dust properties vary in the diffuse ISM at high Galactic latitude (1e19<NH<2.5e20 H/cm2). Our aim is to explain these variations with changes in the ISM properties and grain evolution. Our starting point is the latest core-mantle dust model. It consists of small aromatic-rich carbon grains, larger amorphous carbon grains with aliphatic-rich cores and aromatic-rich mantles, and amorphous silicates with Fe/FeS nano-inclusions covered by aromatic-rich carbon mantles. We explore whether variations in the radiation field or in the gas density distribution in the diffuse ISM could explain the observations. The dust properties are also varied in terms of mantle thickness, Fe/FeS inclusions, carbon abundance, and size distribution. Variations in the radiation field intensity and gas density distribution cannot explain the observed variations but radiation fields harder than the standard ISRF may participate in creating part of them. We further show that variations in the grain mantle thickness coupled with changes in the grain size distribution can reproduce most of the observations. We put a limit on the mantle thickness of the silicates (~10-15nm), and find that aromatic-rich mantles are needed for the carbon grains (at least 5-7.5nm thick). We also find that changes in the carbon abundance in the grains could explain part of the observed variations. Finally, we show that varying the composition of Fe/FeS inclusions in the silicates cannot account for the variations. With small variations in the dust properties, we are able to explain most of the variations in the dust emission observed by Planck-HFI in the diffuse ISM. We also find that the small realistic changes in the dust properties that we consider almost perfectly match the anti-correlation and scatter in the observed beta-T relation.

Sulphur-bearing molecules in diffuse molecular clouds: new results from SOFIA/GREAT and the IRAM 30 m telescope

We have observed five sulphur-bearing molecules in foreground diffuse molecular clouds lying along the sight-lines to five bright continuum sources. We have used the GREAT instrument on SOFIA to observe the 1383 GHz $^2\Pi_{3/2} J=5/2-3/2$ transitions of SH towards the star-forming regions W31C, G29.96-0.02, G34.3+0.1, W49N and W51, detecting foreground absorption towards all five sources; and the EMIR receivers on the IRAM 30m telescope at Pico Veleta to detect the H$_2$S 1(10)-1(01), CS J=2-1 and SO 3(2)-2(1) transitions. In nine foreground absorption components detected towards these sources, the inferred column densities of the four detected molecules showed relatively constant ratios, with N(SH)/N(H$_2$S) in the range 1.1 - 3.0, N(CS)/N(H$_2$S) in the range 0.32 - 0.61, and N(SO)/N(H$_2$S) in the range 0.08 - 0.30. The observed SH/H$_2$ ratios - in the range (0.5-2.6) $\times 10^{-8}$ - indicate that SH (and other sulphur-bearing molecules) account for << 1% of the gas-phase sulphur nuclei. The observed abundances of sulphur-bearing molecules, however, greatly exceed those predicted by standard models of cold diffuse molecular clouds, providing further evidence for the enhancement of endothermic reaction rates by elevated temperatures or ion-neutral drift. We have considered the observed abundance ratios in the context of shock and turbulent dissipation region (TDR) models. Using the TDR model, we find that the turbulent energy available at large scale in the diffuse ISM is sufficient to explain the observed column densities of SH and CS. Standard shock and TDR models, however, fail to reproduce the column densities of H$_2$S and SO by a factor of about 10; more elaborate shock models - in which account is taken of the velocity drift, relative to H$_2$, of SH molecules produced by the dissociative recombination of H$_3$S$^+$ - reduce this discrepancy to a factor ~ 3.

Variation of the ultraviolet extinction law across the Taurus-Auriga star-forming complex. A GALEX based study

The Taurus-Auriga molecular complex (TMC) is the main laboratory for the study of low mass star formation. The density and properties of interstellar dust are expected to vary across the TMC. These variations trace important processes such as dust nucleation or the magnetic field coupling with the cloud. In this article, we show how the combination of near ultraviolet (NUV) and infrared (IR) photometry can be used to derive the strength of the 2175 \AA\ bump and thus any enhancement in the abundance of small dust grains and PAHs in the dust grains size distribution. This technique is applied to the envelope of the TMC, mapped by the GALEX All Sky Survey (AIS). UV and IR photometric data have been retrieved from the GALEX-AIS and the 2MASS catalogues. NUV and K-band star counts have been used to identify the areas in the cloud envelope where the 2175 \AA\ bump is weaker than in the diffuse ISM namely, the low column density extensions of L1495, L1498 and L1524 in Taurus, L1545, L1548, L1519, L1513 in Auriga and L1482-83 in the California region. This finding agrees with previous results on dust evolution derived from Spitzer data and suggests that dust grains begin to decouple from the environmental galactic magnetic field already in the envelope.

The silicate absorption profile in the interstellar medium towards the heavily obscured nucleus of NGC 4418

The 9.7-micron silicate absorption profile in the interstellar medium provides important information on the physical and chemical composition of interstellar dust grains. Measurements in the Milky Way have shown that the profile in the diffuse interstellar medium is very similar to the amorphous silicate profiles found in circumstellar dust shells around late M stars, and narrower than the silicate profile in denser star-forming regions. Here, we investigate the silicate absorption profile towards the very heavily obscured nucleus of NGC 4418, the galaxy with the deepest known silicate absorption feature, and compare it to the profiles seen in the Milky Way. Comparison between the 8-13 micron spectrum obtained with TReCS on Gemini and the larger aperture spectrum obtained from the Spitzer archive indicates that the former isolates the nuclear emission, while Spitzer detects low surface brightness circumnuclear diffuse emission in addition. The silicate absorption profile towards the nucleus is very similar to that in the diffuse ISM in the Milky Way with no evidence of spectral structure from crystalline silicates or silicon carbide grains.

DUST HEATING BY LOW-MASS STARS IN MASSIVE GALAXIES AT $z\lt 1$

Using the Hubble Space Telescope/Wide Field Camera 3 imaging data and multi-wavelength photometric catalog, we investigated the dust temperature of passively evolving and star-forming galaxies at 0.2<z<1.0 in the CANDELS fields. We estimated the stellar radiation field by low-mass stars from the stellar mass and surface brightness profile of these galaxies and then calculated their steady-state dust temperature. At first, we tested our method using nearby early-type galaxies with the deep FIR data by the Herschel Virgo cluster survey and confirmed that the estimated dust temperatures are consistent with the observed temperatures within the uncertainty. We then applied the method to galaxies at 0.2<z<1.0, and found that most of passively evolving galaxies with Mstar > 10^{10} Msun have a relatively high dust temperature of Tdust > 20 K, for which the formation efficiency of molecular hydrogen on the surface of dust grains in the diffuse ISM is expected to be very low from the laboratory experiments. The fraction of passively evolving galaxies strongly depends on the expected dust temperature at all redshifts and increases rapidly with increasing the temperature around Tdust ~ 20 K. These results suggest that the dust heating by low-mass stars in massive galaxies plays an important role for the continuation of their passive evolution, because the lack of the shielding effect of the molecular hydrogen on the UV radiation can prevent the gas cooling and formation of new stars.

DUST AND GAS IN THE MAGELLANIC CLOUDS FROM THE HERITAGE HERSCHEL KEY PROJECT. II. GAS-TO-DUST RATIO VARIATIONS ACROSS INTERSTELLAR MEDIUM PHASES

The spatial variations of the gas-to-dust ratio (GDR) provide constraints on the chemical evolution and lifecycle of dust in galaxies. We examine the relation between dust and gas at 10-50 pc resolution in the Large and Small Magellanic Clouds (LMC and SMC) based on Herschel far-infrared (FIR), H I 21 cm, CO, and Halpha observations. In the diffuse atomic ISM, we derive the gas-to-dust ratio as the slope of the dust-gas relation and find gas-to-dust ratios of 380+250-130 in the LMC, and 1200+1600-420 in the SMC, not including helium. The atomic-to-molecular transition is located at dust surface densities of 0.05 Mo pc-2 in the LMC and 0.03 Mo pc-2 in the SMC, corresponding to AV ~ 0.4 and 0.2, respectively. We investigate the range of CO-to-H2 conversion factor to best account for all the molecular gas in the beam of the observations, and find upper limits on XCO to be 6x1020 cm-2 K-1 km-1 s in the LMC (Z=0.5Zo) at 15 pc resolution, and 4x 1021 cm-2 K-1 km-1 s in the SMC (Z=0.2Zo) at 45 pc resolution. In the LMC, the slope of the dust-gas relation in the dense ISM is lower than in the diffuse ISM by a factor ~2, even after accounting for the effects of CO-dark H2 in the translucent envelopes of molecular clouds. Coagulation of dust grains and the subsequent dust emissivity increase in molecular clouds, and/or accretion of gas-phase metals onto dust grains, and the subsequent dust abundance (dust-to-gas ratio) increase in molecular clouds could explain the observations. In the SMC, variations in the dust-gas slope caused by coagulation or accretion are degenerate with the effects of CO-dark H2. Within the expected 5--20 times Galactic XCO range, the dust-gas slope can be either constant or decrease by a factor of several across ISM phases. Further modeling and observations are required to break the degeneracy between dust grain coagulation, accretion, and CO-dark H2.

COSMOLOGICAL SIMULATIONS OF GALAXY FORMATION WITH COSMIC RAYS

We investigate the dynamical impact of cosmic rays in cosmological simulations of galaxy formation using adaptive-mesh refinement simulations of a $10^{12}$ solar mass halo. In agreement with previous work, a run with only our standard thermal energy feedback model results in a massive spheroid and unrealistically peaked rotation curves. However, the addition of a simple two-fluid model for cosmic rays drastically changes the morphology of the forming disk. We include an isotropic diffusive term and a source term tied to star formation due to (unresolved) supernova-driven shocks. Over a wide range of diffusion coefficients, the CRs generate thin, extended disks with a significantly more realistic (although still not flat) rotation curve. We find that the diffusion of CRs is key to this process, as they escape dense star forming clumps and drive outflows within the more diffuse ISM.