Slashdot Mirror
Astronomers Explode Virtual Supernova
DynaSoar writes "Scientists at the University of Chicago's Center for Astrophysical Thermonuclear Flashes have created a simulation of a white dwarf exploding into a type 1a supernova. Using 700 processors and 58,000 hours, they produced a three second movie showing the initial burst that is thought to be the source of much of the iron in the universe. Understanding these supernovas is also important to testing current cosmological theories regarding dark matter and dark energy, as their brightness is used as a measurement of distance, and discrepancies found in the brightness of very distant supernovas consistently seem to indicate a change in the speed of expansion of the universe over time."
http://flash.uchicago.edu/website/research/gallery /home.py
for all alternative OS users out there.
It's probably 58,000 processor hours, which on 700 processors is closer to 83 hours in real time.
"It is possible to commit no errors and still lose. That is not a weakness. That is life." -Peak Performance
Maybe they can simulate a sense of humor for you.
THere are a lot of interesting things about this. Supernovas are believed to be a major (though not the only) source of all elements heavier than iron in the universe.
For a brief overview (based on Fowler's Nobel Prize lecture) on element formation... This is all from memory (and I am not a physicist) so do your own verification. Basically small stars burn Protium (1H). These fuse to product 2He which immediately decays into Deuterium (2H), emiting a positron. This P-P process eventually allows Deuterium to fuse forming the stable 4He.
As the amount of Helium in a star increases, it eventually becomes possible for Helium to fuse. The only problem is that 8Be is unstable and alpha decays almost immediately back into He. However, you get a small amount of 8Be sitting around for a while, and it can fuse with 4He to produce 12C (Carbon-12). From here things get interesting...
For stars with more than about 1.1 times the mass of our sun, The carbon becomes the basis for Helium production, replacing the P-P process. The basic process (called the CNO cycle) involves single captures of protons (2 of which decay into neutrons and positrons) and then the alpha decay back into 12C. In short this allows Carbon to act as a sort of catalyst for Hydrogen fusion. All elements heavier than Carbon are produced using one of a number of processes. These include fast proton capture, slow proton capture, and alpha capture. The problem is that these become endothermic at the point of Iron. So while smaller stars can produce some of the heavier elements, they are limited in the quantities they can produce. Supernovas, however, can rapidly create much larger quantities of heavier elements.
Also note that at a point in the distant past, stars were more massive than they generally are today. This means that at different points in the history of the universe, we saw large amounts of heavier elements generated.
So this is all quite interesting. I am sure at that many hours we are probably talking about a pretty detailed atomic model. The movie probably shows noting near what the simulation shows.
LedgerSMB: Open source Accounting/ERP
Gah that article is awful. They link to pretty pictures and blurbs mostly and never really explain what these things are, why they are important or give you any real sense of scale. So since I like to beat on the drum of better communication of science, here is a little more detail to add to the good einhverfr's post.
The progenitors of SNIa are most likely white dwarfs composed of carbon nitrogen and oxygen, probably with a companion star from which they are stripping matter. They are very compact on the order of a few thousand kilometers at most, and really dense - more than the mass of the sun. They aren't hot enough to support fusion - they are supported by Pauli pressure; quantum mechanics doesn't allow two electrons in the same state at the same time so though gravity tries to compact these objects there is a Pauli pressure outward to balance it.
This can't go on forever in these progenitor systems however, and if the white dwarf strips enough matter of its companion to get to ~1.4 solar masses (the Chandrashekar limit) then Pauli pressure isn't strong enough to balance gravity and the star begins to collapse and when that happens pressure and temperature rises and somewhere a nuclear fusion flame ignites. Details about what happens near collapse, and where and how the flame ignites, and how many there are and how they progress are still debated. In this particular model they are considering only a single flame (so far) and its a "gravitationally confined detonation" (GCD - the name of this particular model).
Its a little difficult to get a sense of scale from those videos, though there are numbers in the bottom corner. The flame starts of near or just of center and becomes bubble/mushroom shaped through a Rayleigh-Taylor instability and breaks the stellar surface in under a second. Its less than another second before the ash and flame from the bubble collides at the opposite end of the star. This flame crashing into itself (see video 1) causes compression and a detonation.
Theres been a lot of debate as to whether its a deflagration or a detonation or whether it transitions from one to the other and how and when that happens and us poor graduate students just hope they don't go crazy over details of the progenitors during our qualifying examinations. This is notable because there appears to be a growing number of voices who are saying that a detonation is necessary. These events are so standard because they all become SNIa if they get near 1.4 solar masses. There is a fair bit of diversity (and some just crazy objects) and most of that probably arises from details during the explosion which is why modeling them is partly why the models are so important.
There is still a lot of modeling left to do. This flame is producing a lot of heavy elements (there is O, S, Ca, Mg and Si in the early spectra - the silicon feature is around 6150 angstrom in the rest frame and is the marker of a Ia at low to moderate redshifts). As the outer layers expand and become more transparent you see more of the material produced during the explosion and a lot of this is Nickel (Ni-56) which decays to cobalt and powers the light curve so you get this typically 2 week rise and then a slow fall off. Later times most of the Ni has become cobalt which is decaying to iron and you see these elements in the spectrum. The energies we are talking about here are about 10^45 Joules. A H bomb by contrast is 10^15 Joules so 30 order of magnitude. Unless you can picture 10^30 H bombs going off its hard to get a feeling for this number but thats generally the case with numbers in cosmology.
There are a lot of empirical relations you see from the lightcurve, which are exploited to standardize them (for instance the brighter the supernova, the slower its rate of decline, and there are relations for the colour...) and if a model can replicate them and match the observed lightcurves and spectra then this is a very impressive accomplishment. I skim
Reality must take precedence over public relations, for nature cannot be fooled.