Adam Coogan
CV

Probing dark matter with strong gravitational lensing

Source

Source galaxy with no lensing

Source parameters

Orientation of source relative to x-axis
Controls whether source is circular (q=1) or elliptical (q=0)
Higher values cause source brightness to decrease sharply with radius
Sets the size of the source

Telescope

Scale of telescope noise level
Extremely Large Telescope, 0.004-0.012'' pixel size
James Webb Space Telescope, 0.031'' pixel size
0.05'' pixel size
0.1'' pixel size
0.7'' seeing

Observation

Observation of lensed galaxy seen by telescope

Subhalo parameters

Lens parameters

Orientation of lens relative to x-axis
Controls whether lens is circular (q=1) or elliptical (q=0)
Sets the size of the lens

External shear parameters

Horizontal component of distortions from large-scale structure
Vertical component of distortions from large-scale structure

What's this all about?

A range of observations on galactic through cosmological scales demand the existence of dark matter, which outweighs normal matter by 4:1. But beyond its abundance and distribution on large scales, the identity of the fundamental constituents of dark matter remains unknown.

The fundamental properties of dark matter are closely connected with how it is distributed on small scales. For example, the images below from this simulation visualize the dark matter in a galaxy assuming it is cold(i.e. heavy and slow-moving, left plot) or warm (i.e. light and fast-moving, right plot). The difference is striking: if dark matter is warm, galaxies should contain far fewer small structures (more properly calledsubhalos) than if it is cold. The general lesson is that learning how dark matter is distributed on subgalactic scales tells us something about its fundamental properties.

Since small subhalos are made purely of dark matter, they don't emit light and are hard to search for. Instead, I use gravitational lensing to search for them.

To right is a Hubble Space Telescope image of the famous Horseshoe lens. The orange light is from the system's lens galaxy. The blue light comes from the source galaxy. The source is not really ring-shaped. Instead, it lies a good distance directly behind the lens, whose gravitational field dramatically distorts the source's light.

The lens galaxy consists of a large amount of dark matter, stars, dust and gas, and the ring-shaped distortion it produces is immediately apparent. However, the gravitational distortions caused by dark matter subhalos located in the lens are far more subtle. Measuring their distortions requires precision statistical analysis.

I use machine learning and statistics to detect and measure the distortions from subhalos in lenses. The visualization below will give you a sense of why this is a difficult problem. The left image shows a simple model for what the light could look like from an undistorted source galaxy. The image on the right shows what a telescope would see: the distorted ring of light from the source, plus the light from the lens. Some things to explore:

  • Click the "Resample subhalos" button to see how small the distortions from subhalos are. The variations between images with different subhalo populations can be quite small! Where in the observation are the differences most apparent? (Click "Show subhalos" if you want to see wherre the subhalos are located.)
  • Change the sliders controlling the source, lens and shear parameters to see how they impact the observation. Can you find configurations where the subhalos' distortions are more apparent?
  • Upcoming telescopes like the Extremely Large Telescope (ELT) and James Webb Space Telescope (JWST) will have much higher resolution than the Hubble Space Telescope. What impact does that have on how easy it is to see distortions from subhalos?
  • Longer telescope observations reduce the noise in the observation. How does this help make the effects of subhalos more apparent?
  • If you want to hide the light from the lens galaxy to make the distorted source galaxy easier to see, click "Turn off lens light". Subtracting this light from observations is typically the first stage in the data analysis.