
My research centers on theoretical study on astrophysical fluid dynamics, MHD and radiative process
in the context of accretion flows and other objects. A common thread in my research is to
utilize physical theories and numerical experiments to understand how those astronomical objects form, sustain and evolve.
A few of my research interests are:
Turbulent flows driven by Gravitational Instability
Top
Disks that are cold and/or massive may become unstable to gravitational instability. A few examples
are protostellar system which have extended massive disk at their early stage, and the outskirts of
very thin AGN disks. However, the unstable disks may still get away from the fatal fragmentation if
the shock heating due to collapsing could not be radiated away rapidly. A selfregulated state exists
when heating is balanced by cooling and is termed as "Gravitoturbulence" (Gammie 2001). To date,
numerous work have been done with either 2D local/global or 3D global only simulations. And those 3D
studies are mostly based on SPH simulations which are usually less accurate with shock waves and
often have large numerical viscosities. On the other hand, 3D local simulation could be very useful to
investigate the details of the turbulence given its higher resolution and better algorthim for handling
shock heating.

Density clip of a 3D disk to show both the midplane and the top of the box (cooling time ~ 10 dynamical time). Click vimeo link for the MOVIE

Gap Depth
Top
Gap clearing by giant planets has been proposed to explain the optically thin cavities observed in many
protoplanetary disks. How much material remains in the gap determines not only how detectable young
planets are in their birth environments, but also how strong corotation torques are, which impacts
how planets can survive fast orbital migration. We determine numerically how the average surface density
inside the gap depends on planettostar mass ratio q, ShakuraSunyaev viscosity parameter α, and disk
aspect ratio h/r. We find the surface density contrast could easily go beyond three orders of magnitude
as observed in transitional disk cavity.
Collaborators: Jeffrey Fung (University of Toronto) and Eugene Chiang (UCB).
Submitted to ApJ, preprint is available at astroph.

For Neptune mass planets(Duffell and MacFadyen 2013):
For Jupitor mass planets(this work):
For Brown Dwarf(this work):

Planetesimal Formation and Gravitational Instability
Top
Selfgravity is an attractive means of forming the building blocks of
planets, a.k.a. the firstgeneration planetesimals. For ensembles of
dust particles to aggregate into selfgravitating, bound structures,
they must first collect into regions of extraordinarily high density
in circumstellar gas disks. One question to ask therefore is how dense
is necessary to trigger the instability and form bound solid objects?
To answer this, I modified the ATHENA code to
simulate dusty, compressible, selfgravitating flows in a 3D shearing box
configuration, working in the limit that dust particles are small
enough to be perfectly entrained in gas. We find the density requirement is extremely
high to form planetesimals from small particles via direct gravitational collapse
due to the stabilized effects of gas pressure.
More detailed description of this work is provided here:
ADS, astroph.
Some movies are available here.

Image shown two dense dust clumps form via gravitational collapse after ten dynamic timescales (CLICK the image or vimeo link for the MOVIE).

3D MHD Simulations of Circumbinary Accretion Disks
Top
Circumbinary disks are commonly observed surrounding young binary stars, postAGB stars, and protoplanetary systems.
They are also believed to be present around massive binary black holes (MBBHs, a pair of massive BHs) during the galaxy
merging process. They are dynamically rich system to study on, and more importantly, the EM signals emmitted from
the disk might help identify the premergers. We performed the first 3D MHD simulations of such a circumbinary disk
surrounding an equal mass binary. We then did a thorough analysis on the disk structures (such as gap width, stream properties,
disk eccentricity, etc.) and angular momentum transport (ADS astroph. Some movies are available here).
The big accretion rate we found in this work let us consider the accretion efficiency problem of a
binary. Recently, we run parallell simulations for disk around single mass object and disks around binary objects to measure
to what extent the accretion rates are affected by the binary torque with respect to single mass object. Surprisingly,
we find the circumbinary disks have comparable or even higher rate of accretion than normal single mass disks.
Currently, we
are writing up a paper to report our findings. STAY TUNED!

Images shown surface density at initial (top panel) and the quasisteady stage (bottom). CLICK the bottom
panel for a MOVIE. This vimeo link contains a shorter but higher
time resolution version.

MHD Turbulence Driven by Magnetorotational Instability (MRI)
Top
Recently, a couple of numerical simulations of MHD turbulence in unstratified shearing box
find the turbulence strength decays toward zero as the experimental resolution increases.
One of the hypotheses trying to reconcil this problem is that the unstratified box is lack of
a real physical length scale, and the saturation level diminishes as the only
length scale, the size of the grid cell, decreases. In order to testify
it, we study the properties of the MHD turbulence in a stratified shearing box. By introducing
the disk scaleheight as the physical length scale, we find good convergence with respect to
the box resolution (see the image on the top). In addition, we find the averaged toroidal field
flips signs periodically near the midplane before being expelled to higher altitude (see the bottom
image, the "butterfly" diagram which manifests some dynamo mechanism). ADS astroph

The ratios of stress to pressure as a function of time under different resoutions:32x64x256, 64x64x256, 32x128x256, 32x64x512, and 64x128x512 from left to right and top to bottom.Red: the Maxwell stress; Green: Reynolds stress; Black: The sum of those two.

Spacetime diagram of the azimuth component of magnetic field ("butterfly" diagram).


AGN Torus
Top
The dynamics and structure of
toroidal obscuration around active galactic nuclei remain uncertain and
controversial. In this project, we extend earlier work on the dynamical
role of infrared radiation pressure by adding the effects of two kinds of
distributed heating: Compton heating due to hard Xrays from the nucleus
and local starlight heating. We find numerical solutions to the axisymmetric
hydrostatic equilibrium, energy balance, and photon diffusion equations including
these effects. Within the regime of typical parameters, the two different
sources of additional heating have very similar effects: the density profile
within the torus becomes shallower both radially and vertically, but for plausible
heating rates, there is only minor change (relative to the sourcefree case) in
the distribution of column density with solid angle. The most interesting consequence
of distributed heating is that it selects out a relatively narrow range of
parameters permitting an equilibrium, particularly (L/L_E)/\tau_T.
We discuss the implications of both the narrowness of the permitted range and
its approximate coincidence with the range inferred from observations.
(ADS
astroph)

Radiation energy density(top panel) and matter density(bottom) for AGN torus
with hard Xray luminosity 15% of the total

Last updated Oct, 2013
