Research
My research centres on understanding the interstellar medium, or "ISM", of galaxies (the material floating in the space between stars) — what its properties are, and what it can tell us about how galaxies have changed over the history of the Universe. I especially focus on cosmic dust, and the unique window it provides onto how galaxies have developed from clouds of basic primordial gas into evolved systems — enriched with the heavy elements that go on to form new stars and planets (and occasionally astronomers). My scientific home at STScI is the supportive & excellent ISM*@ST research group.
The James Clerk Maxwell Telescope, at the summit of Mauna Kea on Hawai'i, observing for the JINGLE survey during my New Year 2017 observing run (with Venus & the Moon shamelessly photobombing in the backround).
I am currently a European Space Agency / AURA Astronomer at the Space Telescope Science Institute, investigating how the properties if the ISM change within and between nearby galaxies, and how this traces galaxy evolution. I was previously a postdoctoral fellow at STScI, and before that a postdoctoral research associate at Cardiff University, where I did my PhD. I'm an architect of the JINGLE large programme at the James Clerk Maxwell Telescope (JCMT), which is observing the cold ISM in a statistical sample of low-redshift galaxies. I'm also actively involved with the Scylla large programme on the Hubble Space Telescope (investigating the ISM and star-formation histories of dwarf galaxies); with the HASHTAG (observing gas & dust in the Andromeda galaxy), NESS (observing evolved stars in the Milky Way), and DOWSING (Dust Observations With Scuba-2 In Nearby Galaxies) large programmes at the JCMT; and with the the IMEGIN large programme at the IRAM 30m telescope (which is using the advanced new NIKA2 camera to observe nearby galaxies at under-exploited millimetre wavelengths). I'm also a member of the EU-funded DustPedia project.
Preliminary JWST NIRCam data from our first field in M101 to be observed. RGB channels assigned to F090W, F335M, and F444W, respectively. Blue shows stellar & ionised hydrogen emission; green show emission from carbonaceous nanoparticles (with some stellar emission); red shows stellar & hot dust emission.
In 2024, we have been receiving the first data the our 25 hour JWST program that I am leading. In this program, we're aiming to pin down the factors driving the variation in emission from carbonaceous nanoparticles (often called PAHs), especially with respect to metallicity, via a comprehensive multi-parameter benchmarking M101. We will use MIRI and NIRCam imaging to map the strengths of the 3.3, 7.7, and 11.3 um carbonaceous emission features across 6 fields in M101. M101 provides an unparalleled 'controlled laboratory' for this investigation, and has exquisite ancillary data. Our fields sample 0.93 dex in metallicity, 1.6 dex in neutral gas surface density, 2.6 dex in ionized gas surface density, and 3.0 dex in UV radiation field strength. Our data will provide >3500 measurements of each feature with 2" resolution, giving us the statistical power to disentangle the phenomena driving carbonaceous evolution, by seeing how one parameter varies whilst others are held constant. With this data, will be able to tackle open questions about the formation, evolution, and destruction of carbonaceous nanoparticles; and test existing models in a uniquely wide range of environments.
Maps of how the dust-to-gas ratio varies within the four largest galaxies in our Local Group: the Large Magellanic Cloud, the Small Magellanic Cloud, M31, and M33. Large variations in dust-to-gas ratio can be seen both within and between these galaixes (driven mainly by changes in density and in metallicity, respectively).
In Clark et al. (2023), I explore how the dust-to-gas ratio (one of the fundamental measures of a galaxy's chemical evolution) varies within the four largest galaxies in the Local Group. I show that the dust-to-gas ratio can vary with density by over a factor of 20 within a single galaxy, far more than previously thought. This is strong evidence for high levels of dust grain growth in the denser regons of the interstellar medium of galaxies. I also demonstrate that dust-to-gas properties can differ markedly between similar-appearing galaxies. This work was the subject of a NASA-JPL press release, and covered by publications worldwide.
A comparison of an original map of the Small Magellanic Cloud galaxy (left), compared to my new corrected map (right), with previousy-missed dust emission restored via "feathering".
In Clark et al. (2021), I produced new Herschel Space Observatory maps of nearest, most-extended galaxies in the sky: the Local Group galaxies Andromeda (M31), Triangulum (M33), the Large Magellanic Cloud (LMC), and the Small Magellandic Cloud (SMC). Previous Herschel maps of the these galaxies missed a large fraction (over 20% in some cases) of the total flux present, becuase Herschel is unable to detect very extended emission. However, using a technique called "feathering", I was able to combine the Herschel data with observavtions from other telescopes (Planck, COBE, and IRAS). My new maps capture the diffuse extended dust emission that was previously missed, whilst still preserving the excellent resolution provided by Herschel. These new maps enabled me to detect dust in exceptionally low-density environments, below ΣH = 1 M☉ pc-2.
NASA's SOFIA observatory in flight, with the telescope doors open. (Image credit NASA / Jim Ross; public domain.)
I have also been awarded time to use NASA's flying telescope, the Stratospheric Observatory For Infrared Astronomy (SOFIA), to look for carbon absorption in the ISM. With this data, we will be able to pin down what fraction of the carbon in space is locked up in dust grains. Despite carbon being one of dust's main constituents, we still don't have a good idea of what fraction of all carbon is actually locked up in dust.
A Hubble Space Telescope image of extremely metal-poor galaxy Leo P, showing the stars we will use to measure extincton due to dust (crosses), and the regions within which we will combine extinction meaurements to map Leo P's dust mass (lines). Also marked are potential dust sources and sinks: purple hexagons are dusty evolved stars; the red ring is a candidate supernova remnant; the double orange ring outlines Leo P's HII region; and the pink circles indicate areas of higher density atomic gas.
In 2020, I succesfully proposed for time on the Hubble Space Telescope to observe the extremely metal-poor dwarf galaxy Leo P. The prperties of the ISM in such primitive galaxies are poorly understood, especially how & where dust forms, the role of dust role in creation of molecular gas, and the charactaristics of dust at low metallicity. By using Hubble to observe Leo P in multiple UV–optical–IR wavelengths, we will be able to measure the dust extinction along the line-of-sight to many hundreds of its stars. By combining extinction measurements for stars located close together, we will be able to create a map of dust in Leo P. This will tell us where the dust is, how much of it there is, its properties, how it relates to Leo P's various potential sites of dust creation and destruction. Leo P is the only extremely metal-poor galaxy close enough for us to perform this sort of analysis, and can provide a window on how the ISM evolves in primirtive galaxies in the early Universe.
The first map of the dust mass absorption coefficient, κd, in a galaxy, for nearby spiral M83. The panel on the left shows the map of κd, whilst the panel on the right shows a multiwavelength reference image of M83 for comparison (where blue is ultraviolet, green is near-infrared, and red is far-infrared).
In 2019, I created the first ever maps of the dust mass absoroption coefficient in galaxies, showing how it varies within nearby spirals M74 and M83. The dust mass absorption coeffecient, κd, is what astronomers depend upon to calculate the actual mass of dust in a system, from observations of the dust's thermal emission. However, its value is notoriously poorly constrained, and it was previously totally unknown how it varied within galaxies. In Clark et al. (2019), we show that κd appears to be lower in regions of denser ISM - this is the opposite of what had been predicted by models, but the result cannot be avoided without making unphysical or contrived alterations to our method.
Multiwavelength imagery of DustPedia galaxy NGC 3686, showing the apertures used by the CAAPR photometry pipeline to measure the galaxy's flux in each of the 26 ultraviolet—submillimetre wavelengths shown. CAAPR performed this process for all 875 DustPedia galaxies; see Clark et al. (2017).
In Clark et al. (2018), I describe the extensive database of imagery and photometry that I created for the DustPedia project, covering 875 nearby galaxies and spanning 42 ultraviolet—microwave bands. For this, I created the publicly-available CAAPR photometry pipeline, tailored to handle the difficult task of automatically conducuting robust aperture-matched photometry for nearby galaxies.
Microscope image of a liver tissue section, as processed by my AstroCell software. The original image is shown in the top-left panel, with the other panels showing the results of AstroCell automatically identifying and classifying the cells.
In 2017, I was awarded funding by the Data Innovation Research Institute to take image-analysis techniques used in astronomy, and apply them to microscope imagery of cancerous tissues, in collaboration with the School of Biosciences at Cardiff University. When studying microscope images of tissues, cancer researchers still regularly count and classify the many thousands of cells by hand — whereas in astronomy we use software to detect and analyse the stars and galaxies in our data. I worked on developing AstroCell, a user-friendly and freely-available piece of software to count and classify cells (in contrast to the complex, specialised, and typically commercial software generally avaialble).
Compilation of values for κd (the dust mass absorption coefficient) published 1984—2019. This shows show poorly-constrained κd is, with values spanning over 3 orders of magnitude! The empirical value reported in Clark et al. (2016) is shown in blue. The range of values I find in M74 and M83 in Clark et al. (2019) are shown by the shaded purple bars.
In Clark et al. (2016), I used data from the Herschel Reference Survey to empircally determine a value for κd (the dust mass absorption coefficient), the notoriously-poorly-constrainted conversion factor that astronomers need to work out dust masses. By taking advantage of the fact that the dust-to-metals ratio in the ISM only shows minimal variation, I find a value of κd at 500 μm of 0.051 (+0.070, −0.026) m2 kg−1.
Six of the enigmatic blue and dusty gas rich galaxies revealed in Clark et al., 2015, as they appear in four different parts of the spectrum. Note their abundant star formation (in the ultraviolet, 1st row), irregular and flocculent morphologies (in the optical, 2nd row), modest populations of evolved stars (in the near-infrared, 3rd row), and dust-richness (in the submillimetre, 4th row).
For my PhD thesis (and published in Clark et al., 2015), I used Hershcel-ATLAS to create the first dust-selected volume-limited survey of nearby galaxies, providing our first unbiased view of dust in the local Universe. Unexpectedly, I found that most dusty galaxies belong to a severely under-studed population dubbed 'BADGRS' — Blue And Dusty Gas Rich Sources (examples shown above). BADGRS contain only 5% of the stars in the local Universe, but 35% of the dust, and over 50% of the atomic gas. Their dust is very cold (12—16 K), and absorbs an exceptionally small fraction of their stellar emission. BADGRS represent a fascinating intermediate stage in galaxy evolution. They’ve built up a lot of dust very quickly, and are relatively metal-rich (metallicities of 0.5—1.1 solar); however, they are still early in the process of converting their gas into stars, unlike most commonly-studied galaxies.
Resolved component separation of the Crab Nebula as it appears at 160 μm (published in Gomez et al., 2012), showing the synchrotron (left), warm dust (centre), and cool dust (right) emission. This was the first ever map of the distribution of cold dust manufactured within a supernova remnant.
I also have an active interest in evolved stars, following on from work I carried out for my PhD thesis (also published in Gomez et al., 2012). I used Herschel data to perform a resolved component separation of the far-infrared and submillimetre emission of three supernova remnants, including the Crab Nebula. My resolved component separation for the Crab (shown above) generated the first ever map of the distribution of cold supernova dust within a supernova remnant. This map showed the dust lies protected within the Crab's dense filaments, proving that it was manufacted by the Crab, and showing that the dust will survive to be injected into the Galactic environment — the first remnant for which this can be said.