Mar 302014

By Molly Peeples, Postdoctoral Fellow at STScI

As all heavy elements are produced in stars and stellar deaths, the eventual fates of metals are unique tracers of the large scale gas flows driving galaxy evolution. Some metals will remain in the ISM, some will get trapped in stars during subsequent episodes of star formation, and some will be blown out of the galaxy via large-scale galactic winds. I have recently conducted an inventory of metals in and around star forming galaxies at z~0 in order to place constraints on the histories of gas flows into, within, and out of galaxies (Peeples et al. 2014).

Figure 1 shows our main result, the relative distribution of metals in galactic and circumgalactic components relative to the total amount of metals galaxies have produced, as a function of galaxy stellar mass.  “100%” on this diagram denotes the total mass of metals a galaxy has produced by Type II supernovae, Type Ia supernovae, and AGB stars throughout its lifetime.

Figure 1: Cumulative fraction of metals in interstellar gas (blue), stars (red), interstellar dust (orange), the highly ionized circumgalactic medium (CGM; green), the low-ionization CGM (purple), circumgalactic dust (brown), and the hot X-ray traced CGM (yellow) of star forming galaxies. The points correspond to the median stellar mass of the COS-Halos galaxies. 100% corresponds to the total mass of metals a typical star forming galaxy of a given stellar mass has produced in its lifetime. Adapted from Peeples et al. (2014).

With regards to the metals that remain in galaxies, there are a few surprising results this analysis has uncovered.  The relative distribution of metals within galaxies depends on the galaxy stellar mass: massive galaxies have most of their retained metals in stars, while low mass galaxies have the bulk of their retained metals in the interstellar medium. Remarkably, star forming dwarf galaxies have more metals in interstellar dust than in stars; models of the chemical evolution of these galaxies cannot ignore this relatively massive component.  Most strikingly, by combining the metals in stars and the ISM gas and dust, we find that star forming galaxies have retained a nearly constant 20-25% of the metals they have produced. That this fraction is so constant is extremely counter-intuitive because our current understanding of galaxy evolution has as a central facet the idea that low mass galaxies are more efficient at driving galactic winds, owing to their relatively shallow potential wells.

An important implication of the fact that galaxies have retained only ~20% of their metals is that, because this fraction is so low, most of the metals galaxies have produced must no longer be in the galaxies, but instead have been expelled into their surroundings. With the installation of the Cosmic Origins Spectrograph (COS) on HST, we can finally systematically characterize galaxies’ gaseous halos’ baryons and metals. With 129 orbits, the COS-Halos program (Tumlinson et al., 2013) has measured the mass of metals and baryons in the circumgalactic medium (CGM) of galaxies with stellar mass ~1010 Msun out to impact parameters of 150 kpc (Peeples et al., 2014; Werk et al., 2014). We find that the CGM is multi-phase: there is a warm, diffuse, highly-ionized phase (traced by the highly ionized oxygen ion, OVI) and a colder, denser low-ionization phase (traced by lower ionization species such as Si II, Mg II, and C II). The relative masses we measure for these phases are shown as the purple and green wedges in the Figure; the points denote the median stellar mass of the COS-Halos galaxies. (The wedge-like shape owes to the fact that we do not detect a difference in the CGM mass within 150kpc that depends on the galaxy stellar mass.) We find that the CGM out to 150kpc has a mass of metals comparable to the mass remaining in the interstellar medium.

The COS-Halos team is now working on observationally extending this metal inventory by characterizing the CGM beyond 150kpc, and, with the COS-Dwarfs survey, of lower mass galaxies within 150kpc. From a theoretical standpoint, I am using analytic models and hydrodynamic simulations to try to reconcile how galaxies can have outflow efficiencies that strongly depend on the depth of their potential wells while simultaneously retaining a constant fraction of the metals they produce and maintaining a CGM structure that is consistent with our observations.

Tumlinson et al. 2013 (ApJ, 777, 59)
Peeples et al. 2014 (ApJ, in press)
Werk et al. 2014 (ApJ, submitted)

Mar 172014

Using Hubble to probe the dynamic interaction between the Magellanic Clouds

By Andrew Fox, Astronomer at STScI

Interactions between spiral galaxies and their dwarf satellites are often spectacular, producing extended streams of stripped gas and triggering new generations of star formation. The most striking local example of such an interaction lies in the outer halo of the Milky Way in the form of the Magellanic Stream. Extending for over 140 degrees across the Southern Sky, the Stream is a giant ribbon of gas trailing the orbit of the Large and Small Magellanic Clouds as they journey around the Galaxy. Since its discovery over 40 years ago, the Stream has puzzled observers and theorists alike and raised many questions. How was it physically removed from the Magellanic Clouds? Did it originate in the LMC or SMC? And what will its ultimate fate be? New spectroscopic observations with the Hubble Space Telescope and the Very Large Telescope are addressing these questions and finding the origin of the Stream to be surprisingly complex.


Figure 1 :  Top: In this combined all-sky radio and visible-light image, the Magellanic Stream is shown in pink. The radio observations, taken from the Leiden-Argentine-Bonn (LAB) Survey, have been combined with a visible-light panorama. The Milky Way is the light blue band in the center of the image. The brown clumps are interstellar dust clouds in our galaxy. The Magellanic Clouds are seen in white at bottom right. Bottom: close-up of the Stream with our HST/COS sightlines marked with crosses. Credit: NASA, ESA, D. Nidever et al., NRAO/AUI/NSF, A. Mellinger, LAB Survey.

Measuring the chemical abundance (metallicity) of interstellar gas clouds requires finding UV-bright background sources, such as quasars. By splitting the quasar light into its constituent colors, the absorption lines imprinted by foreground gas clouds can be measured. These lines encode detailed information on the chemical composition and motion of the foreground clouds. Using observations from the Cosmic Origins Spectrograph (COS) installed on Hubble in 2009, we observed eight active galactic nuclei (AGN) lying behind or near the Stream. By comparing the strength of the neutral oxygen (O I) and ionized sulfur (S II) UV absorption lines to the strength of the atomic hydrogen (H I) 21 cm emission measured by radio telescopes, we derived the Stream’s metallicity in each direction. O I and S II were chosen for these measurements since they are largely unaffected by ionization and dust-depletion effects, so their ratios with H I provide robust metallicity indicators.

We found the Stream’s metallicity to be only ≈10% of the solar value in three separate directions sampling most of its length, considerably lower than the current-day average metallicity of the SMC (≈20% solar) and the LMC (≈50% solar). However, the age of the Stream is estimated from tidal models to be around 2 billion years, and information on the metallicity evolution of the Magellanic Clouds indicates that 2 billion years ago, the SMC abundance was ~10% solar, matching the value we measure in the Stream, whereas the LMC abundance was much higher, at ~30-40% solar. Our results thus support a scenario in which most of the Stream was stripped from the SMC (not the LMC). It has not self-enriched since its formation, because there is no evidence for ongoing star formation in the gas. In a sense, we have measured a fossil record of the Stream at the time of its birth in the SMC about 2 billion years ago.

However, a fourth sightline we studied (toward the AGN Fairall 9) tells a very different story. In this direction, which lies close to the Magellanic Clouds on the sky, the sulfur abundance in the Stream is found to be 50% solar, five times higher than the value measured in the other directions, and much higher than expected for gas that has been stripped from the SMC. Furthermore, the Fairall 9 direction intercepts a filament of the Stream that appears to connect kinematically to the LMC. Our measurement of a higher metal abundance supports this claim, and points toward a dual origin for the Stream, with two interwoven strands of material, one pulled out of the SMC about 2 billion years ago, and another pulled out of the LMC more recently.

Ongoing work by our team is investigating the total mass and inflow rate of the Magellanic gas onto the Milky Way, where it will potentially be able to fuel future generations of star formation. However, the gas must first survive the perilous journey through the hot Galactic corona, which can evaporate passing gas clouds. This survivability is being tested with computer simulations.

For more details, see Fox et al. 2013 (ApJ, 772, 110) and Richter et al. 2013 (ApJ, 772, 111), and the NASA press release and Huffington post article

Mar 032014

By Andrea Banzatti, Postdoctoral Fellow at STScI

As the number of confirmed (exo)planets grows rapidly, infrared spectroscopy is providing an exceptional opportunity to study the molecular environments where planets are being formed. The Spitzer Space Telescope recently revealed a dense forest of emission lines from water, OH, and some organic molecules (Figure 1) tracing warm/hot gas (200 < T < 1000 K) in young protoplanetary disks inward of the water snow line (the condensation/evaporation boundary between gas and ice in the disk)1. This “steam” emission offers a unique observational link to a variety of processes ongoing inside planet factories.


Figure 1 : Infrared emission from water and other gas molecules in the planet formation region of a young protoplanetary disk (around RNO90, a G5 star), as observed with the Spitzer Space Telescope. All the prominent emission features, apart from those labelled differently, are due to water vapor (a model2 of water emission is superimposed in blue).

Analyzing the rich infrared molecular emission, we recently had the opportunity to investigate its connection to disk evolution and planet formation processes. By comparing infrared spectra of young solar-mass stars, taken at different phases of their accretion of circumstellar material, I and my collaborators found that strong accretion outbursts are able to dramatically affect the molecular content at planet-forming radii in the disk3. An increased heating causes a recession of the snow line to larger disk radii, probably triggering evaporation of water ice, while a harsher UV radiation photodissociates water vapor producing OH (as seen in the infrared spectrum of EX Lupi during a strong recent outburst, see Figure 2). The fate of organic molecules, which disappear during outburst, remains unclear. If accretion outbursts are ubiquitous in star formation, the evolution of material in the planet formation region may be strongly linked to the accretion histories, probably affecting also the chemical/physical properties of forming solid bodies.


Figure 2 : Changes in molecular emission from the disk of EX Lupi during a strong accretion outburst (from Banzatti et al. 2012). Water and OH emission increases, in connection to a recession of the snow line and a stronger UV radiation. The emission from organic molecules instead disappears.






Infrared spectroscopy of planet formation regions also allows us to study the migration of icy solids in the midplane of disks. As icy dust grains stick together in the outer disk, they form larger particles that at some point decouple from the gas and are dragged toward the star. When they reach and cross the snow line, the ice is evaporated providing large abundances of water vapor in the inner disk, unless forming planets accrete them at outer radii4. Evidence for ongoing ice migration is provided by inner disks where the water vapor abundance exceeds the oxygen budget available to form it in situ, while “drier” disks may be advanced in depleting the outer disk from migrators, or have already formed large accreting planetesimals outward of the snow line. In a recent work5, we have shown how a rotation diagram analysis of infrared water vapor emission offers a useful tool to distinguish between these two scenarios, from the spread of the rotational scatter (Figure 3). Such studies of water vapor emission are lifting the veil on processes taking place in disk midplanes that have been until now elusive to our observations.


Figure 3 : Rotation diagram analysis of infrared water vapor emission in protoplanetary disks (from Banzatti et al. 2013). Line opacities are color-coded in blue, while dot sizes are proportional to line intensities. The larger the amount of water vapor in the inner disk, the larger the spread of the rotational scatter in the diagram. The plot to the right shows the case of an inner disk water abundance larger than the oxygen budget available in situ, providing evidence for enrichment from inward icy migrators that evaporate at the snow line.

The number of protoplanetary disks and of molecular emission lines from planet formation regions observed with Spitzer is by far the largest provided by any other telescope to date. This unique dataset has already given the opportunity for pioneering studies of the properties and evolution of the molecular environments during planet formation. Yet, it is likely that we are just scratching the surface. These studies offer us a fertile ground for planning observations with the upcoming James Webb Space Telescope, which promises to narrow the gaps in our understanding of how planets form and to bring us closer to deciphering planet diversity, so to understand better even our own Earth.


  1. Carr & Najita 2008 (Science, 319, 1504); Salyk et al. 2008 (ApJ, 676, L49); Pascucci et al. 2009 (ApJ, 696, 143); Pontoppidan et al. 2010 (ApJ, 720, 887), 2014 (PPVI chapter)
  2. Pontoppidan et al. 2009 (ApJ, 704, 1482)
  3. Banzatti et al. 2012 (ApJ, 745, 90), 2014 (ApJ, 780, 26)
  4. Ciesla & Cuzzi 2006 (Icarus, 181, 178)
  5. Banzatti 2013 (PhD thesis)