Research

Dust Properties and Evolution

I have been leading several projects to characterize the dust properties as a function of environment in nearby galaxies using a wide range of techniques and wavelengths. Using far-IR maps of the Magellanic Clouds obtained with Spitzer, Herschel, IRAS, and Planck, along with ground-based gas tracers (HI 21 cm and CO 1-0 emission), I have been studying the gas-to-dust ratio in the LMC and SMC (Roman-Duval+2010, 2014, 2017), showing that the dust abundance increases by a factor ~5 between the most diffuse regions, which includes the outskirts of these galaxies, and dense star-forming regions. These recent results are consistent with predictions from chemical evolution models (e.g., McKinnon+2016, Zhukovska+2016), and confirm that dust grows in the ISM with density and metallicity dependent timescales. I am the PI of a NASA ADAP proposal to extend this kind of study to the local volume (galaxies within 10 Mpc).

Because emission-based tracers suffer from degeneracies (mainly due to the lack of constraints on the FIR dust opacity and the CO-to-H₂ conversion factor, X_CO), I have been obtaining independent measurements of the amount of metals in the dust phase from UV spectroscopy. I am the PI of large Hubble program (METAL: Metal Evolution and TrAnsport in the Large Magellanic Cloud) to obtain a detailed and unambiguous census of metals in the gas and dust phases of the LMC, which has half-solar metallicity. The program obtained high S/N UV spectra toward a large sample of massive stars serving as “light bulbs” in the LMC to measure UV extinction curves and elemental depletions for the main constituents of interstellar dust (Fe, Si, Mg, Ni, Cu, Cr, Zn, S). For the first time, we are directly characterizing the variations of the dust composition and abundance as a function of density in the LMC. In parallel to the spectroscopic observations, Hubble obtained WFC3 imaging in 7 filters spanning the UV-IR in fields nearby the massive stars. These images will be used to derive extinction maps, which will be compared to the corresponding far-IR emission seen in Herschel data in order to characterize the far-IR emissivity of dust grains at half-solar metallicity.

Distribution and Properties of Molecular Gas in Nearby Galaxies

I have been leading several studies to quantify the structure of the molecular ISM in the Milky Way and Magellanic Clouds in order to understand the drivers of star formation.

Using large-scale surveys of CO 1-0 emission in the Milky Way, I compiled a large sample of molecular cloud distances and properties (mass, temperature, radius, velocity dispersion, surface density), their scaling relations, and examined the variations of those properties with location in the Galaxy (Roman-Duval+2009, 2010). I characterized the properties of turbulence in simulated and observed molecular clouds, showing that their energy spectrum is consistent with shock-dominated, intermittent turbulence, and that the scaling exponent and normalization of the turbulence spectrum does not vary significantly within the Milky Way disk (Roman-Duval+2011). My results had a significant impact on the ISM/star formation community, since the birth material of young stars could then be located and characterized.

Properties of molecular gas have traditionally been measured after identifying clouds using clump-finding algorithm. This introduces a level of subjectivity that may lead to spurious scaling relations (Ballesteros-Paredes+2012). Hence, I led a study of the structure and distribution of diffuse and dense molecular gas in the Milky Way based on the entire emission seen in the position-position-velocity cubes (Roman-Duval+2016). I demonstrated that a substantial fraction of CO gas remains in diffuse, non-star forming form, and that the contribution of diffuse gas increases away from the galactic center. This could explain the large molecular gas depletion times (low star-formation efficiency) observed at large radii in galaxies (Leroy+2008).

In low-metallicity systems, the paucity of shielding dust grains results in CO being photo-dissociated, while self-shielding H₂ molecules can exist at low A_V. As a result, the structure of molecular clouds is expected to change as metallicity decreases (Bolatto+1999). I investigated the HI-H₂ transition using HI 21 cm and CO data in combination with Herschel data of the Magellanic Clouds, providing constraints on the HI column required to allow H₂ to form (Roman-Duval+2014). I have recently started a study of the structure of CO gas and its relation to star formation using ALMA observations of 6 Magellanic Cloud regions spanning a range of star-formation rates. In the next few years, the ALMA data will allow us to characterize the column density distribution, turbulent properties, and excitation conditions of the CO gas at low metallicity, and compute the star formation efficiency on cloud scales using existing YSO catalogs.