This post covers work from 09/03-09/13. I worked on the titan simulations throughout this time and I also went to San Francisco from 09/06-09/11 on vacation.

Titan Sims

Batch Script and Interactive Sessions

I finished running the single pulse simulation up to 3ps which ran without any exploding which is nice. Since the files take a while to analyze using the login node in VSCode, I decided to start doing the EPOCH analysis using an interactive session on Jupyter. I first realized that the default directory OSC places us in is the $HOME directory which is not ideal to look at the project or scratch directory where all the simulations are. Johnathon’s solution was to manually allocate resources using salloc --nodes=1 --ntasks-per-node=4 --time=1:00:00 --account=PAS1066. Alternatively, we can use salloc -N 1 -n 4 -t 1:00:00 -A PAS1066 as a shorthand. These are the same SLURM commands you would use in a batch script. You would also need --xll as an argument to salloc to enable XLL forwarding for the interactive session. OSC recommmends submitting jobs for 1 hour or less because they can run on nodes reserved for debugging. Once the nodes are allocated, can check with squeue to get the JOBID and see the node that is allocated and then use ssh <node##> to log into the node that was allocated.

My solution was to simply create a symbolic link to the project and scratch directories in the home directory. To do this (for example for the project directory), I would type (in the home directory):

ln -s /fs/project/PAS1066/ronakndesai project

which creates the directory project that corresponds to my project directory.

First Success of Single/Double Pulse Simulations

I’ll recap some powerpoint slides I made about these simulations. Here is a schematic of the target:

which is the same as my previous simulations, except that I’ve increased the resolution to 75 cells per micron or 13.33nm cells. I ran the simulations for 2-3 ps with 25 ppc for electrons/ions and 125 ppc for protons. I set the simulation at a 10eV initial temperature. Overall, I’ve found that the double pulse proton energies are higher and since they get accelerated faster, we see the energies peak earlier. The fields are stronger on the rear side of the target. Here is a plot of the double pulse at 1500fs which is when there were the highest proton energies:

as compared to the single pulse at 2000fs which was when it had its highest proton energies

We can see that the double pulse max proton energies are almost double that of the single pulse energies. Additionally, the double pulse is more symmetric. We can also look at the energy conservation in double pulse

and single pulse

where we can clearly see higher proton conversion efficiency in double pulse as well as the energies matching up to the theoretical energy (computed from a corresponding sim without a target). Another thing we could look at is the ionization level in the double

and single

and by looking at the colorbar, we definitely see more overall energy from the double pulse. Interestingly, both cap out at around an ionization level of 30.

Chris talked to me on Thursday 09/12 and gave me some feedback. He said I should watch the particle probes to see if they are slowing down the simulation. I recompiled EPOCH with the (D)NO_PARTICLE_PROBES on so that its not doing the extra step of checking for probes. Additionally, I ran a single pulse simulation with OAP2 instead of OAP1 to see if I get qualitatively the same results as OAP1. If they are similar or if OAP2 is worse, I can just ignore the results of OAP2. If the time to run is significantly less, I can just keep particle probes turned off to save computation time. The particle probes are helpful for quickly looking at the proton energies, but not necessary. I think I should be able to gather the same information from the proton energies in the simulation combined with the ejected particle energies.

He also suggested picking a certain spot near the rear surface and plotting the electric field just at that spot, which may be more relevant for analyzing TNSA fields. Another interesting thing to look at would be the proton particle positions to see how some of the protons get ejected backwards and where they are. Because if protons are only on the rear surface, it should be hard for them to go back through the target, but maybe some of them are able to go around the target.

Additionally, looking at the Brunel paper might be useful to get an analytical model for optimal incidence angle of the double laser pulse.