Solving 'cosmological puzzles' (Credit: Bingqing Sun)
Hi! I'm Shihong Liao, a postdoc from the National Astronomical Observatories, Chinese Academy of Sciences (NAOC). Before joining NAOC, I finished my PhD at the Chinese University of Hong Kong (CUHK).
My primary research interest is the formation and evolution of galaxies. In my current research, I mainly use numerical simulations to explore the following mysteries in our Universe:
What is the nature of ultra-diffuse galaxies?How does environment affect the formation and evolution of galaxies?Where do galaxy spins come from?What is the nature of dark matter?Does dark matter annihilate?
In the meantime, I have broad interests in many other research topics. I am actively collaborating with scientists from different institutes/universities on the topics of gravitational waves, neutrino cosmology, interacting dark energy models, data analysis in astronomical observations, simulation visualizations, etc.
As a young scientist, I am enthusiastic about teaching and supervising students. When I was pursuing my PhD at CUHK, I got four years experience of being a teaching assistant, and at last I was awarded a teaching assistance prize for my performance and my help to students. During my current postdoctoral career at NAOC, I have successfully co-supervised one PhD student to finish a paper published in MNRAS and two undergraduates to finish their graduation theses.
On this website, you can find more specific information about my research and supervising & teaching experiences. Welcome!
My primary research interest is the formation and evolution of galaxies. Galaxy formation and evolution is one of the most complicated physical processes in the nature world: at the first stage, the dominating matter component, dark matter, collapses under gravitational instability and forms virialized dark matter haloes; at the second stage, gas (or baryons) falls into the potential wells created by dark matter haloes, gets shock-heated, cools radiatively, forms luminous stars, and experiences energetic feedback processes. These physical processes are highly non-linear, and currently it is impossible for us to solve the equations describing them by paper and pencil. As far as we know, the only powerful method is to solve these equations numerically with supercomputers.
Over the past years, I have devoted my efforts to study the formation of galaxies and dark matter structures in our universe using numerical simulations, namely cosmological N-body simulations and hydrodynamic simulations.
In the following, you can find the summary and paper links for some of my research projects.
Ultra-diffuse galaxies (UDGs) are a large sample of recently observed galaxies whose sizes are as large as our Milky Way Galaxy while whose luminosities are as faint as dwarf galaxies. The nature of these galaxies is still being debated.
I lead a project of using the Auriga simulations, which are a set of state-of-the-art zoom-in magetohydrodynamic simulations, to study the formation of such puzzling galaxies. In Auriga, field UDGs form in dark matter haloes with larger spins compared to normal dwarfs in the field. Satellite UDGs, on the other hand, have two different origins: about half of them formed as field UDGs before they were accreted; the remaining half were normal field dwarfs that subsequently turned into UDGs as a result of tidal interactions.
Paper LinkIn the classical picture, to form galaxies, primordial gas was first trapped into the potential wells created by dark matter haloes. Comparing with haloes, massive filaments have the same function to provide potential wells to trap gas. Understanding these processes in detail will be helpful in understanding the observed filamentary dependences of galaxy properties.
We use zoom-in hydrodynamic simulations to show that cold and dense gas preprocessed by dark matter filaments can be further accreted into residing individual low-mass haloes in directions along the filaments, resulting a very anisotropic gas accretion for filament haloes. Filament haloes have higher baryon and stellar fractions when comparing with their field counterparts. The results suggest that filaments assist gas cooling and enhance star formation in their residing dark matter haloes.
Paper LinkThe standard galaxy formation theory assumes that baryons and dark matter are initially well-mixed before becoming segregated by radiative cooling. This implies that before cooling, baryons and dark matter in a halo should have identical specific angular momentum as they have experienced identical tidal torquing.
With non-radiative hydrodynamical simulations, we show that on the contrary to this assumption, baryons and dark matter can also be segregated because of different physics obeyed by gas and dark matter during halo assembly. As a result, baryons and dark matter in many haloes do not originate from the same Lagrangian region, experience different tidal torques, and thus their angular momentum vectors are often misaligned. The result challenges the precision of some semi-analytical approaches which utilize dark matter halo merger trees to infer gas properties.
Paper LinkThe environmental effects on dark matter halo assembly history and structure are a key ingredient in understanding the environmental dependence of galaxy properties which has been observed in many galaxy observations.
With high-resolution cosmological N-body simulations, we trace the accretion history of each particle in each z = 0 halo and record its accretion time and accretion modes, to quantify the environmental effects on halo growth. On average, haloes in high-density environments accrete their z = 0 particles earlier at all radii than their low-density counterparts. High-density haloes tend to accrete their particles slightly more from major mergers and less from smooth accretion comparing to low-density haloes, and this effect is more evident for inner particles. The environmental effects are usually more pronounced for haloes with lower masses.
Paper in preparationHalo angular momentum is a key ingredient in understanding the formation of disk galaxies. Once haloes reach virial equilibrium, their properties are found to follow some universal forms which are valid for haloes with different masses, reflecting the equilibirum properties of collisionless systems. One example is the famous NFW density profile. It will be interesting to see if there is any universal spatial distribution for halo spins.
From cosmological simulations, we find that the spatial distribution of specific angular momentum in dark matter haloes is universal, and it can be fitted with a simple functional form. The profile tells us that the outer parts close to the equatorial plane of a halo tend to have larger specific angular momentum, while the inner parts near the polar direction usually have smaller specific angular momentum.
Paper LinkThere is a well-known simple power-law relation between an object's angular momentum, J, and its mass, M, i.e., J ~ M^{5/3}, the so-called J-M relation.
To understand the origin of this J-M relation, I and my collaborators use simulations to study its time-evolution, J(t) = β(t) M(t)^α(t), trying to find out when it established. In the ΛCDM model, α starts with a value of ~1.5 at high redshift z, increases monotonically, and finally reaches 5/3 near z = 0, whereas β evolves linearly with time in the beginning, reaches a maximum and decreases, and, finally, stabilizes. A three-regime scheme is proposed to understand this newly observed picture. The tidal torque theory accounts for this time-evolution behavior in the linear regime, whereas α = 5/3 comes from the virial equilibrium of halos. An updated and more complete understanding of the J-M relation is given.
(Image: M101. Credit: Subaru, HST, Robert Gendler)
Paper LinkGrid and glass configurations are two widely used pre-initial conditions when generating initial conditions for cosmological N-body simulations.
Recently, I introduce an alternative method to prepare pre-initial conditions, the Capacity Constrained Voronoi Tessellation (CCVT), which originates from computer graphics. As a geometrical equilibrium state, the CCVT configuration is uniform and isotropic, follows perfectly the minimal power spectrum, P(k) ~ k^4, and is quite stable under gravitational interactions. This new method is shown to play as good as grid and glass in cosmological simulations, and it will be helpful in studying the numerical convergence of pre-initial conditions in cosmological simulations.
Paper Link(Credit: Abstruse Goose)
Starting from my PhD studies, I have had more than 7 years experience in cosmological simulations. I am familiar with almost all commonly used techniques in performing and analyzing cosmological simulations:
(i) From preparing initial conditions (e.g. grid & glass pre-initial conditions, 1st- and 2nd-order Lagrangian Perturbation Theory, CAMB, N-GenIC/2LPTic, MUSIC) to running simulations (e.g. Gadget);
(ii) From identifying haloes (e.g. AHF, FOF, SUBFIND), to constructing merger trees (e.g. cross-matching particle IDs), to computing large-scale statistics (e.g. halo mass functions, matter power spectrum, two-point correlation function, pairwise velocity, cosmic web classfication, etc.), and to computing halo inner properties (e.g. density profiles, spins, shapes, etc.).
During my research, I have written a series of codes and shared with my collaborators, to name a few,
P3Mcode, a Particle-Particle-Particle-Mesh N-body codeMERGER, a merger tree constrution code based on halo particle IDsPOWSPEC, a code to estimate matter power spectrum from simulation particlesTPCF, an OpenMP-optimized & tree-based code to compute particle/halo two-point correlation functionsPWV, an OpenMP-optimized & tree-based code to compute particle/halo pairwise velocityWebCla, a code to perform cosmic web classification based on the Hessian matrix methodFILAMENT, a code to identify filaments from zoom-in simulations based on the Hoshen-Kopelman algorithm
I have made some codes publicly available in my GitHub, for example,
ccvt-preic, an OpenMP-optimized code to generate simulation pre-intial conditions with the Capacity Constrained Voronoi Tessellation method
As a postdoc at NAOC, I am enthusiastic about supervising/co-supervising students. Following are some student projects that I have involved in.
1. Yun Liu: Properties of dark matter haloes in the interacting dark energy models (Aug 2019 - Present; Co-supervising)Yun is a master student from SWIFAR, Yunnan University. We aim to study the halo properties in the interacting dark energy models using cosmological N-body simulations.
2. Tianchi Zhang: Impacts of pre-initial conditions on dark matter halo properties (Jan 2019 - Present; Supervising)Tianchi is a PhD student from the University of Chinese Academy of Sciences. We aim to investigate the impacts of pre-initial conditions on simulated dark matter halo properties.
3. Jia Hu: Studying the impact of filaments on galaxy formation with the Auriga simulations (Mar 2019 - May 2019; Co-supervising)Jia was an undergraduate student from Heilongjiang University. This is a short-term project for Jia's undergraduate thesis. In this project, Jia learned the basic concept of modern galaxy formation simulations, learned to analyze simulation data, and tried to investigate the impacts of filaments on the formation of their residing dwarf galaxies. By classifying dwarf galaxies into filament and field environments, Jia comfirmed that at high redshifts (z >= 2.5), galaxies in filaments tend to have higher baryonic fractions and star fractions, indicating that filaments play a role in assisting gas cooling and star formation.
4. Tianchi Zhang: Optimal gravitational softening length for cosmological N-body simulations (May 2016 - May 2018; Co-supervising)This is the first project for Tianchi's PhD studies. In this project, Tianchi mastered how to perform modern cosmological simulations with the GADGET code, and proposed an improved optimal scheme in setting simulation softening length based on the work of Power et al. (2003).
Tianchi's first science paper has been published in MNRAS 487, 1227 (2019). Congratulations to Tianchi!
5. Haonan Zheng: Halo growth in filaments (Feb 2017 - Jun 2017; Co-supervising)Haonan was an undergraduate sutdent from Beijing Institute of Technology. This is a short-term project for Haonan's undergraduate thesis. In this project, Haonan got familiar with the Pheonix zoom-in simulations, learned to analyze cosmological simulation data, and studied how filaments affect the growth and properties of residing haloes. Haonan showed that comparing to haloes in fields, high-redshift haloes in massive filaments tend to have higher concentrations and earlier formation time.
As a teaching assistant at CUHK, I was responsible to lead exercise classes, mark students' homework, answer students' questions in consultation hours, and offer help to those students who have difficulties in their study. During my four years' PhD studies, I was a teaching assistant for the following courses.
2015 Spring: PHYS3202 Quantum Physics II
2014 Autumn: PHYS2041 University Physics III
2014 Spring: PHYS3202 Quantum Physics II
2013 Autumn: PHYS2005 Quantitative Methods for Basic Physics II
2013 Spring: PHYS2005 Quantitative Methods for Basic Physics II
2012 Autumn: PHYS3201 Quantum Physics I
2012 Spring: PHYS2005 Quantitative Methods for Basic Physics II
2011 Autumn: PHYS3201 Quantum Physics I
In 2014 Summer, I was awarded a teaching assistance prize for my performance in teaching duties and my help to students.
Shihong Liao National Astronomical Observatories, CAS 20A Datun Rd, Chaoyang District Beijing 100101, China
Email: shliaoATbao.ac.cn
Github: https://github.com/liaoshong