PhD Project

Hydrodynamic Simulations of supernova remnants:
Dust Destruction by the reverse shock

My main research focus is currently on simulations of dust destruction mechanisms in supernova remnants as part of the SNDUST project (PI: Prof. Michael Barlow, UCL).  This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (ERC advanced grant N0 694520).  Please visit our official project webpage for a project overview and work by other group members.  Visualizations of results can be found on our Vimeo group SNDUST.

Astrophysical Background

Cosmic Dust

Cosmic dust plays a significant role in the composition and evolution of the universe.  It reprocesses the radiative output of stars, transforming ultraviolet and optical light into far-infrared and submillimeter emission while also influencing the thermal and chemical balance of gas by depleting or enriching it.  Gas cooling due to dust can trigger the collapse of molecular clouds, leading to the formation of young stars (Tsuribe and Omukai, 2006) and dust grains themselves can act as the formation site for molecular hydrogen which facilitates star formation (Hirashita and Ferara, 2002).  Due to its importance, significant effort has been put into investigating cosmic dust and of particular interest here is the formation and abundance of cosmic dust.

For a long time, scientists believed that cosmic dust was mainly formed in the atmosphere of asymptotic giant branch (AGB) stars.  However, in recent years Sloan Digital Sky Survey and Hubble Deep Field observations have revealed significant amounts of dust in high redshift (z > 6) quasars (Bertoldi and Cox, 2002).  Stars such as our sun will take about 10 billion years to transition from a main sequence star (a star burning hydrogen in its core) to an AGB star (a star that has burned through all its hydrogen in the core).  But at redshift 6, the universe was only around 1 billions years old, making it very unlikely for the discovered dust to have been produced by AGB stars alone.  Instead, core-collapse (CCSNe) have been suggested as alternative dust producers, possibly dominating in the early universe

Cosmic dust in the Horsehead Nebula
(Hubble Space Telescope).
Source: Wikipedia.

SNR Cassiopeia A (Spitzer Space Telescope, Hubble Space Telescope, Chandra X-ray Observatory).
Source: Wikipedia.

Core-Collapse Supernovae as Dust Producers

Core-collapse supernovae (CCSN) are one of the most violent and energetic phenomena in our universe.  When a star of a mass larger than about 8 solar masses has burned through all its fuel sources the fusion processes in its core halt and the core temperature and the pressure, once sustained through the burning of chemical elements, start to plummet.  This causes the outer layers to fall inwards onto the inner core, compressing it further and further.  Eventually, the core becomes too dense and the energy deposited by neutrinos triggers a massive explosion, expelling the outer material into space.  The inner regions of the supernova (SN) ejecta are cool, dense, metal-enriched and ideal for dust formation (Silvia et al., 2010).  Theoretical models suggest, that SN explosions of the first stars might have produced large amounts of dust within the first 150 to 800 days after the explosion (Todini and Ferrara, 2001).  Such dust has indeed been inferred from observations of several supernova remnants (SNR) like for example SN 1987A and the Crab Nebular (0.25 to 0.8 solar masses of dust, Barlow et al. 2010, Matsuura et al., 2011, De Looze et al., 2017) as well as  SN1980K, SN 1993J, Cassiopeia A (0.08 to 1.1 solar masses of dust, Bevan et al., 2017).

Unfortunately, not all of the dust that is produced in those early days after the explosion survives long enough to be incorporated into the interstellar medium where it then becomes part of the cosmic matter lifecycle.

Core-Collapse Supernovae as Dust Destroyers

While space is often described as a vacuum, it is in fact not completely empty.  Instead, the space between star systems is filled with what we call interstellar medium (ISM).  The ISM (made out of gas, small molecules, dust, and atoms) might be of a density that we would consider negligible in everyday life, but in space, even the smallest densities can have a huge effect.  When a star explodes as a CCSN, the material it expells will eventually impact the ISM which causes a reverse shock to propagate back through the ejecta (Chevalier, 1977).  This reverse shock leads to sputtering (the dust grains shrink due to interactions with the shocked gas) and grain-grain collisions (the dust grains can be shattered or completely destroyed when they impact another dust grain) as shown in Barlow (1978), and Drain and Salpeter (1979).  In general, the expected lifetime of grains against sputtering is proportional to the dust grain radius:  Small dust
grains are more likely to be completely destroyed and they are trapped more easily in regions of hot gas where sputtering processes are particularly effective.  Larger grains, on the other hand, are thought to survive long enough to eventually transition into the ISM.

The effect of the reverse shock on the newly formed dust grains has been studied by various groups in recent years including Nozawa et al. (2006), Bianchi and Schneider (2007), Nozawa et al. (2007), and Silvia et al. (2010).  Silvia et al. (2010), used the magnetohydrodynamics code ENZO to simulate dust grains enclosed in ejecta clumps and subjected to sputtering due to interactions with the reverse shock.  They found that the initial radii of the particles, the density of the SN ejecta and the relative velocity between ejecta and reverse shock play a pivotal role in determining how much dust survives.  Likewise, Nozawa et al. (2006 & 2007), found that depending on the initial conditions (ISM density, SN explosion energy) between 20% and 100% of the newly formed dust is destroyed before it reaches interstellar space.

While all models suggest that dust grains are strongly processed by the reverse shock, the exact extent of the dust destruction is not well constrained.  In this study, we develop a new dust processing algorithm including the effects of sputtering, grain-grain collisions, and drag to determine whether CCSNe are major contributors to the dust budgets of early galaxies.

The interaction between the SN remnants and the previously ejected material (Hubble Space Telescope).
Source: Wikipedia.


(This section is kept short as publications are currently in the works)

Following the example of previous studies, we make use of a public MHD code (AstroBEAR) to simulate the SN ejecta environment. We adopt the “Cloud Crushing” model presented in Silvia et al, 2010, but refrain from using tracer particles to represent our dust grains.  Instead, we have developed a “Dusty Grid” model based on a combination of approaches summarized by Haworth et al. (2016).  Currently, we are working on two dust processing treatments: External treatment (Dr Florian Kirchschlager, UCL) and internal treatment (Franziska Schmidt, UCL).  For a visualization of a preliminary test with external tracer particles see the video at the bottom of the page.


To simulate the conditions within a SNR, we make use of the publicly available MHD code AstroBEAR developed at the University of Rochester, USA.  AstroBEAR is a parallelized, adaptive mesh refinement, 3D MHD code widely used in the astrophysics community.  Our internal treatment approach aims to include dust processing directly into the MHD code and is supported by a collaboration with AstroBEAR developer Dr Erica Fogerty (Los Alamos National Laboratory, Los Alamos, USA).

The “Cloud Crushing” Model.

The ‘Cloud Crushing’-Model

An illustration of the “Cloud Crushing” model is shown on the left.  We split the computational domain into a pre- and a post-shock region.  Density, pressure, temperature, and velocity are calculated according to the Rankine-Hugoniot conditions across the shock front.  A cloud or clump of denser gas, which is assumed to hold the majority of the new dust (Biscaro and Cherchneff, 2016), is placed in the pre-shock medium.

The ‘Dusty Grid’-Model

Unlike previous studies, we do not include dust in form of tracer particles.  Instead, we introduce overlaying grids representing a distribution of dust bins.  The dust is then transported between cells according to partial differential equations describing the advection and inter-bin transport equations describing the dust processing.

Visualization of an initial external dust post-processing test combined with the underlying MHD simulation.

Funding Information

The SNDUST project (grant agreement N0 694520) has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme.

This work is partially performed using the Cambridge Service for Data Driven Discovery (CSD3), part of which is operated by the University of Cambridge Research Computing on behalf of the STFC DiRAC HPC Facility ( The DiRAC component of CSD3 was funded by BEIS capital funding via STFC capital grants ST/P002307/1 and ST/R002452/1 and STFC operations grant ST/R00689X/1. DiRAC is part of the National e-Infrastructure. (Project Codes: ACSP190 and ACSP217)