Constraints on Dark Matter Mass from Lyman-α forest

This is a julia code repository for Computational Cosmology final project, Winter 2016.

Summary

We use Eulerian N-body hydrodynamics simulations to study the evolution of warm dark matter models. We compare the artificial Lyman-alpha spectra depending on the mass of the dark matter particles and check whether there really appears difference in small scale features of the spectra. We argue that careful comparison of the actual observation and simulation can give us a lower bound on the dark matter mass.

Summary of Numerical Simulation

All codes are written in Julia.

We use Particle Mesh(PM) method for realizing a system with interacting particles (dark matter) and fluid (ordinary matter). The basic principle of PM method is to convert the system of particles into a grid of density values. For density assignment, we choose Cloud-in-Cell; i.e. particles are cubes of uniform density and of one grid cell size.

There are five differential equations in time and spatial coordinates x,y,z which govern the system, in terms of density, energy, 3D momentums, and pressure. We discretize those differential equations and evolve the system in time discretely.

We use periodic 2D volume of 100x100 grid size.

More details of the work can be found in final report.

Sample Plots

The sample numerical simulation of dark matter and ordinary matter(baryon) distribution.

In the figure above, from top to bottom, we show how dark matter and ordinary matter distribution evolves in time by gravity and hydrodynamical pressure. The left two columns are for mass = 100 GeV (cold dark matter), and the right two columns are for mass = 100 keV (warm dark matter). For each mass, images in the left column show the evolution of dark matter distributions, and those in the right column show the evolution of ordinary matter. From the top to bottom, the corresponding redshifts are z = 11.42, 6.83, 4.97, 3.92 (high z means more in the past). Note that although ordinary matter started with uniform distribution, they quickly fall into the potential well of dark matter.

The simulated Lyman-alpha forest spectra

Note

It is built upon the original NoName repository by Tom Abel which provides an environment for exploring various algorithms in computational cosmology. Please refer to original ReadMe or NoName repository for further details of the original code.