Pieter de Buck - Week 3 - Duke University

Hi everyone, it's Pieter and this is my third week at Duke.

This week has been pretty great. Mr. Patt came to visit with his son and I learned some new stuff in the lab.
I finished all my assignments that Dr. Nahrgang has given me and I am now working on my actual project with Dr. Bass.

Dr. Bass instructed me on how to use the computer model that simulates particle collisions called UrQMD, or the Ultrarelativistic Quantum Molecular Dynamics model. He, along with others, has created this program, and it is really impressive. When setting up each simulation I can specify all kinds of variables. I can choose both the projectile particle and the target particle, I can set the beam energy, which is basically the speed the particles are travelling in. For our purposes, these speeds are about 99.99996% the speed of light. That's what the term "Ultrarelativistic" comes from. The simulations of the particle collisions are done with the Monte Carlo method. This is a method for obtaining usable data from things that are not able to be expressed by some kind of function. For example, we cannot know for sure what is going to happen in each collision between particles because of minute differences. Every collision is "unique", millions of collisions and scatterings go on after two nuclei collide, and each of those collisions has a certain degree of randomness. In other words, we cannot say that a certain particle will be at this position at this time, or collide with that particle at that time. The Monte Carlo method uses randomness in each calculation, so they are all different. We then simulate millions of these events, and from that we can start seeing relationships and correlations. You can see the final product, such as a graph, as a probability distribution, we can say with a degree of certainty that a proton for example will have a momentum of 1-5 GeV. A certain proton could have a momentum of 20 GeV, but our model suggests that that is highly unlikely.

In my simulation, I collided two atoms of lead at 0.9999 times the speed of light. This single collision scattered about 4000 separate particles, and through the Monte Carlo method I simulated 200 lead+lead collisions, giving me 800.000 particles to work with. This is enough sample space to really be able to see coherent graphs. I will insert a histogram from my earlier, simpler project, which shows the momentum in GeV on the x-axis, and the number of particles that fit that momentum on the y-axis

1e7 = 10.000.000 events

After running my simulation, I adapted my earlier program to be able to read the more complex output file that the computer model uses. I was able to do this quickly and also make the program more user-friendly, the user can specify which kind(s) of particles they want to look at, and from which file they want to read the results. I will get screenshots of my new histograms next time, when I am sure that everything is working correctly.

I have really enjoyed watching the world cup, it was much better than I expected. Congrats to Germany

Pieter