Figure 4 from Forbes, Krumholz, & Speagle (2018). These are scaling relations related to galaxy structure (stellar mass vs. half-mass radius, circular velocity, stellar mass concentration, and central stellar surface density). Observational data are shown as colored bands, and the points are drawn from the calibrated semi-analytic model.

The Evolution of Star-forming galaxies

In some sense, star-forming galaxies are simple. Their properties (stellar mass, star formation rate, metallicity, optical sizes and so on) are all tightly correlated over a wide range of mass and redshift, suggesting that they are set by straightforward physical principles which vary little from galaxy to galaxy. Despite this, it is likely that the properties of galaxies are governed by galactic winds, stellar feedback, star formation, and a variety of other complex dynamical processes. Each of these is poorly understood, often governed by very small-scale physics, and in the case of galactic winds, the dominant physical mechanism responsible (supernova momentum, cosmic rays, or radiation pressure) is not even known. 

This situation calls for flexible but detailed models of galaxy evolution, like the radially-resolved semi-analytic model I've been developing. With some help from machine learning, these reasonably sophisticated models can be fit to a wide range of observational data. With the proper calibration, these models will be useful as a general tool for researches studying the evolution of star-forming galaxies.

star formation in dwarf galaxies

Dwarf galaxies form stars slowly and in a far different environment than the more massive Milky Way-mass galaxies where much of our understanding about star formation and feedback arises. Dwarf galaxies also happen to be easier in some ways to simulate, precisely because the lower metallicity eases the resolution requirements on the cooling radius of supernova blast waves, and decreases the optical depth to FUV photons.

Our high-resolution numerical experiments explore the physics of star formation regulation in low-mass galaxies, where predictions of different models predict dramatically different outcomes.

Figure 4 of Forbes et al (2016). Newly-formed star particles indicated by the arrows suppress the ability of nearby gas to form stars as the result of FUV heating (gas near the star formation threshold is shown in red contours).

Figure 1 of Forbes & Loeb (2018). Sufficiently cold brown dwarfs (further left in the diagram) may be found even at masses exceeding the hydrogen burning limit (upper left in the diagram).

Unconventional Brown Dwarfs

The traditional dividing line between brown dwarfs and low-mass stars is a particular (composition-dependent) mass. Below this mass, objects [brown dwarfs] may burn deuterium and even hydrogen, but they never reach a steady state where their surface luminosity can be supplied by the fusion of hydrogen, while just above this mass, stars may live for up to about 10 trillion years.

While this dividing line is usually a good description of which objects are stars and which objects are brown dwarfs, it is also possible to build up brown dwarfs above this critical mass. If mass is added sufficiently slowly to a sufficiently old brown dwarf, it can substantially exceed the hydrogen burning limit, yet remain a cold degenerate brown dwarf. 

In Forbes & Loeb (2018), we predict that these objects should be reasonably rare, <1% of brown dwarfs, and are most likely to arise in cases of mass transfer via Roche lobe overflow. We hope this theoretical prediction will be borne out by future observations, and we expect further study of the phenomenology of these binaries will aid in this observational search.

Hydrodynamic Shielding

 Cold clouds moving in hot gas tend to get shredded by hydrodynamic instabilities, yet they seem to persist, likely via continuous resupply and condensation via cooling, or extra stability courtesy of magnetic fields. 

Granted that cold clouds can survive, their velocities as a function of time can give us strong constraints on the background gas through which they move, and the sizes of the clouds themselves, which are often difficult to observe directly. Ultimately these dynamics may have a profound role to play in the accretion of gas onto galaxies, galactic winds, the interface between a star-forming disk and the circumgalactic medium, and many other problems in astrophysics.

Figure 1 from Forbes & Lin (2018). A series of numerical experiments showing a line of clouds at different separations subject to a low-density wind. The denser the stream of clouds, the more effectively they resist disruption and acceleration by the background gas.

Figure 6 of Forbes (2015). These are the trajectories of spinning meter-sized objects, colored by the direction of their spin vector. If they can spin quickly enough, their lifetime in the disk can be altered by about a factor of 2 - not enough to qualitatively change the meter-size barrier, but a large enough effect to quantitatively change how the streaming instability acts to produce planetesimals.

Figure 4 of Forbes & Loeb (2018a). The fraction of stars subject to different levels of mass loss as a function of the mass of the host galaxy. At z=0, most planets in galaxies like the Milky Way or more massive have lost a mass equivalent to Mars's atmosphere, and about 10% have lost a mass comparable to Earth's atmosphere. Some may even be transformed from sub-Neptunes to super-Earth's by this process (see Chen, Forbes, & Loeb (2018))

Problems for planets

Planets are everywhere, but we're not sure how they form. Dust particles amalgamating to form larger objects run into trouble when they approach about a meter in size - collisions between meter-sized objects tend to erode the rocks rather than building them up, and on top of that drag from the gas in the protoplanetary disk causes them to spiral into the star quite quickly - in a few hundred years. A promising candidate to avoid this problem is to have smaller particles act collectively in the streaming instability to directly collapse from dust into ~10 km planetesimals.

​I was curious how the Magnus force (or the lift force) that arises for spinning objects would affect this picture. It turns out that in the right circumstances, the lift force can be about as strong as the drag force (just in a perpendicular direction), and can disrupt the streaming instability as it tries to form planetesimals. 

In addition to potentially causing problems for planet formation theories, my collaborators and I have looked into how planetary atmospheres may be affected by the supermassive black holes at the centers of galaxies. It turns out that atmospheres are sufficiently fragile, and black holes can be sufficiently bright, that a large fraction of the planets in the universe have likely sustained some damage from black holes.

Bacterial Predation

This isn't astrophysics, but you might be surprised how similar the dynamics of dust grains in protoplanetary disks and bacteria in a dish are.

What happens when a bacteria hunter is introduced into a population of prey bacteria? If one set of prey can swim and the other can't, you would expect the motile bacteria to run into their predators more frequently (as is happening in the animation on the right). But it turns out that in the real world when you run this experiment for Bdellovibrio bacteriovorus (the hunters, a "natural anti-biotic") and two sets of Vibrio cholerae, one of which has been genetically modified to not be able to swim, the immotile bacteria are rapidly killed off. Running some experiments (don't worry - I was nowhere near the lab!) with different fluid viscosities points to the fact that the greater the drag force on Bdello as it attempts to kill Vibrio, the harder it is for Bdello to kill its prey.