Research at Fermilab Part 1

I am already halfway done with my time at Fermilab.  The goal of this trip was to advance my research on the Pion Charge Exchange Cross Section to prepare it for a conference in late January.  I certainly have been making progress, but with particle physics the progress is often slow at first as you learn the tools.  Once you can understand the computer, progress is nearly instant as the computational power at my disposal is fairly immense.

So far I have spent my time attempting to understand the behavior of photons in our detector using simulations.  The difficult part of this is that the simulations are extremely sophisticated and there isn’t a friendly user manual.  Finding where to change one line of code to change an output is like finding a needle in a haystack if you don’t know what you’re doing.  However, this is where being on site at Fermilab has proved advantageous.  There are many experts whom I can query about the simulations I am using, and this has produced great results.  For instance, one of my problems was that I couldn’t figure out what happens in the simulation when a photon interacts.  I expected the photon to create “daughter” particles, but couldn’t find these anywhere in the records of a photon interaction when I looked at them.  It turns out that photons produce particles that in turn produce particles and all in all many hundreds of particles are created when a photon interacts.  To keep the logs from being too verbose tracking all of these particles is turned off by default.  There is a single parameter hidden in thousands of lines of configuration code that one can switch to see a photon interaction reflected in the output logs.  I was able to diagnose the problem of missing particles on my own, launch an inquiry to experts, and get an answer in a fairly short time frame.  I have now fixed this problem moving forward.

An intermediate goal of this research is to accurately be able to identify photons in particle interactions.  To be sure that I have the right answer I can compare my results for some parameter to the simulation’s logs.  This is why it is imperative that I have all the information in these logs at my disposal.  Much of my time so far had been in exploring different methods to identify photons, but until this morning I did not know that the simulation logs, which we call “truth information” were garbage.  Now later today I can continue exploring with the accurate truth information at my disposal.

Another reason that the truth information is useful is that I can double check the “reconstruction” process. Essentially, a particle is simulated and an algorithm tries to use physics first-principles to figure out what is happening.  Then, I can reveal the truth and compare it to the reconstruction.  If they match, it is a good algorithm; and if not, it needs some work.  I have been able to use other parts of the truth information that were not being hidden to verify parts of our reconstruction.  For example, there should be only a constant (to account for differences in units) difference between the energy known to the simulation and the energy reconstructed by the algorithm.  I was able to verify this for our reconstruction, so I am confident that the energy reconstruction is working, but I now need to verify that the energy is attributed to the correct particle when many of them are moving within the apparatus.  This and much more will hopefully be in the works this week.