Astatine: Halogen or Metal?
quantumchymist.blogspot.com
quantumchymist.blogspot.com
As the post says, this is a cooling problem. I wonder if it would work to deposit a thin film of Bi-209 on a thin diamond plate, the other side cooled by liquid nitrogen. Bombard the bismuth with alpha to make astatine. The diamond's very high thermal conductivity (peaks at over 40,000W/m-°K at ~ 100°K [1]) could keep the mixed bismuth/astatine film in solid form.
Since you'd want to do the alpha irradiation in a vacuum, the setup would allow for a wide range of surface characterization techniques.
[1] https://en.wikipedia.org/wiki/Material_properties_of_diamond...
“There’s no material safety data sheet for astatine. If
there were, it would just be the word “NO” scrawled over
and over in charred blood.”
― Randall Munroe, What If?: Serious Scientific Answers to Absurd Hypothetical QuestionsThe first reference is quite entertaining ...
"“Being a high energy oxidizer, dioxygen difluoride reacted vigorously with organic compounds, even at temperatures close to its melting point. It reacted instantaneously with solid ethyl alcohol, producing a blue flame and an explosion. When a drop of liquid 02F2 was added to liquid methane, cooled at 90°K., a white flame was produced instantaneously, which turned green upon further burning. When 0.2 (mL) of liquid 02F2 was added to 0.5 (mL) of liquid CH4 at 90°K., a violent explosion occurred.” And he’s just getting warmed up, if that’s the right phrase to use for something that detonates things at -180C (that’s -300 Fahrenheit, if you only have a kitchen thermometer). The great majority of Streng’s reactions have surely never been run again. The paper goes on to react FOOF with everything else you wouldn’t react it with: ammonia (“vigorous”, this at 100K), water ice (explosion, natch), chlorine (“violent explosion”, so he added it more slowly the second time), red phosphorus (not good), bromine fluoride, chlorine trifluoride (say what?), perchloryl fluoride (!), tetrafluorohydrazine (how on Earth. . .), and on, and on. If the paper weren’t laid out in complete grammatical sentences and published in JACS, you’d swear it was the work of a violent lunatic. I ran out of vulgar expletives after the second page. "
[1] http://blogs.sciencemag.org/pipeline/archives/2010/02/23/thi...
[2] http://blogs.sciencemag.org/pipeline/archives/2014/10/10/thi...
[3] http://blogs.sciencemag.org/pipeline/archives/2013/01/09/thi...
What I remember from this time was that our calculations were ridiculously intense because we could never get a machine with enough ram to store, or computational power to compute on the fly, all the matrix elements <phi_1|Operator|phi_2> we needed, so we wound up setting up huge tables on disk, and doing interpolation.
My 64 atom supercells would take, initially a week of runtime to get 100 timesteps. After a few years, my SGI R8000 could do that calc about 90 seconds per time step. 10 years after I finished up, my home computer could do 10s wall clock per time step. I've not tried it recently, could be fun.
Dirac eqn is a harder thing to deal with in general. We used Schrodinger eqn parameterized for soft pseudopotentials. Also had a restriction on how many k-points we could handle, and ionization was sadly impossible in the code, as the repeated volume of supercells made any unbalanced charge effectively infinite. I came up with some ways to handle this, but we lacked computer power for that. Likely it would work today.
Honestly, I miss that work. It was fun.
Calculations of charged periodic systems are possible through the use of Ewald summation techniques (or particle mesh Ewald/related techniques). It basically introduces a homogeneous neutralizing background charge, which is unphysical and makes it so that a the total energy of a charged particle in a box converges to the "true" result as a function of 1/volume. This is applicable to both classical and quantum simulations.
There is no way to do a "true" charged periodic system, but honestly that's never what you want (because true infinite systems with net charge don't exist!). You can do better than homogeneous neutralizing background charge by introducing a neutralizing electrolyte through a continuum solvent model. Here's the implementation I'm most familiar with (https://arxiv.org/abs/1601.03346), but I know the fully open-source JDFTx also implements this (which has a bunch of other fancy approaches to incorporating solvation into periodic DFT calculations).