Weekly Reports
Week 1 (Pre-selection):
Went to many lectures and learned about the various projects that are available here. I came here for HEP (High Energy Physics), however, so I hope for that work.
Many people wanted HEP. I and three others were chosen. My partners are mentioned on the previous page.
Week 1 (Post-selection):
Learned a lot from the mentors about neutrinos and other physics.
Tasked to read the information on: Double Chooz proposal, KamLAND proposal and safety/shift manuals, and Geant4.
Week 2:
Tasked with learning the tools necessary to study neutrinos: C++, Root, and, eventually, Geant4.
Tasked to make a program for a specific topic. Different topics were given to each of the four of us.
Started making programs for whatever topic we chose individually. I chose to study the antineutrino interaction that is mentioned on the previous page. My goal was to simulate an antineutrino entering the simulated detector at some angle and hitting a proton that's relatively at rest. I was to also simulate the neutron and positron leaving the proton location with the appropriate energy and random, isotropic angles.
Week 3:
Continued working on our programs until Tuesday.
Presented the work done on the programs. Two of the four of us had programs that ran. My own didn't have all the features that I wanted, but it did run and I accomplished my goals.
Tasked to experiment with Root and Geant4 next.
Tornado hit Wednesday. My car’s rear window was damaged. No one here was killed.
Nothing Thursday due to the extensive damage to the campus. Volunteered for some clean-up.
Arrived to work on Friday. However, power went out in the morning and didn't come back on until much later afternoon. Didn't get much done.
Week 4:
Placed my data into Root. I created a pretty histogram or two.
Started up Geant4 to learn it. Had to have a bit of instruction on the software.
Discovered that all four of us will go to KamLAND in two groups in July.
Discussed topics for our "official" research. I chose to study free neutron spallation and their interactions within the Double Chooz detector. Basically, a cosmic ray will generate a muon in the atmosphere which travels down through the rock surrounding the detector. It imparts a significant amount of energy to a random nucleus. A neutron is ejected with significant momentum. Assuming it does not get absorbed where it's supposed to, then it will stand a chance of impacting a proton in the fluid, which generates a photon burst similar to the positron/electron annihilations resulting from the neutrino interactions. The free neutron background (and other backgrounds) must be determined to accurately remove those photon signatures from the neutrino measurements.
Week 5:
Started looking at the Geant4 macros and source code in depth to understand them enough to manipulate them to give the data required.
Mentors left on a trip and will probably be back on Monday.
The tornado caused some damage to the air filtration system and we have been evacuated from the building for asbestos testing.
Help is required to understand everything we’re doing with Geant4. Our mentors being gone and the asbestos problem are keeping us from working at peak efficiency. Dr. Larry Weaver is giving us some physics information related to our projects while our mentors are away.
Week 6:
Continued working on project. Dr. Horton-Smith returned and has been assisting all of us. Also, as of today, I’m able to work completely from my computer in the dorms. Installing and configuring Linux was a pain in the butt.
Came up with an idea to improve the Geant4 simulations. Basically, the simulation simulates rocks
surrounding the detector but the elements it uses are not the entire list of
elements that I would use. I will do
some research on this and probably have to look at the Geology of northern
Week 7:
Emailed a list of elements that I think should be included in the simulations. Dr. Horton-Smith emailed back an old report detailing sampling of the Chooz site rocks. The information in it not only confirmed all the elements that I suggested, but even added an additional element that I had not considered. Hopefully this information will be put to good use.
I’ve been discovering a number of annoying bugs in Geant4 that slow down my simulations. This is my last full week here and slowing down is the last thing I need.
Things have finally started working for me! I eliminated as much of the processes that I could to speed up simulations. Basically, any photonic processes are not important to my studying of the fast neutrons so they are removed via the program macro. Working from the inside out, I discovered that out of 10,000 events, only 2 neutrons reach the target in the detector, 6 the gamma catcher, 17 the buffer, and 74 the inner veto. Well, based on these numbers, I would need 1,000,000 events to get >100 in the target, which is what I need for better than 10% relative statistical error. Sounds like there’s a lot of run time needed.
Set Geant4 to run two separate 1,000,000 event simulations with a single parameter difference between them (one has rock thickness of 0.4 and the other 4 meters). It’s the 0.4 meter thickness that produced the 2 neutron interactions in the target with the 10,000 events. Hopefully these will be done by Monday.
Using the rest of my time this week to prepare for the trips and
presentation on Friday. All four of us
might go to FermiLab on Wednesday and Alex and I, our flight for
Week 8:
Okay, so we’re not going to be able to go to FermiLab. Just too many things going on. Oh well. At least I live in Omaha and can drive it myself later.
Continued working on my project. Analyzed the data from the million event runs and discovered the inner veto of the detector blocks almost all of the neutrons. Also, none of the neutron events seem to correlate to neutrino events. This is good news. More analysis to come.
Spent some time today beautifying my program (organizing and commenting appropriately). I also added into the program a few lines of code so that Root would spit out initial and final positions of the neutrons. Dr. Horton-Smith later told me the portion of code I used for final positions isn’t really giving final positions; it’s giving something more like final weighted positions based on deposited energies. This is good enough because actually getting the final positions is excessively complicated, according to my mentor.
So, all my data to this point has a flaw in it, Dr. Horton-Smith and I
discovered. I was attempting to
correlate neutron and neutrino events using a 100 microsecond window in which
the prompt and delayed energies seem to match up. I consistently found no correlated
events. The reason is not because they
don’t happen at all, but because the macro running the simulation, by default,
merged events that were shorter than 1 millisecond. What this means is that I had to change my
macro and re-run my simulations. Now
they merge when events are shorter than 1 microsecond. This, combined with my program’s instruction
to cut-off analysis of events at 100 microseconds, gives me the required
window. I ran 10,000 events for a quick
test and found one correlation event, confirming that the previous macro was
faulty. Because I’m leaving for
Analysis of the runs yielded excellent results. We discovered two correlation events with the 0.4m rock thickness run and zero with the 4.0m run. This is promising as there is a bit more than 4m of rock on the side of the detector and above the detector is an extension of the inner veto shield, which was quite effective in blocking the neutrons. I’ll attempt to complete this webpage while I’m gone. I’m sure some actual numbers and pictures will serve well in spicing up this page. Time for the trip! Later!
Weeks 9 – 10:
In
The training took place in the mine and we did a daily walk-around check as well as hourly checks of the data via computer.
The work took place on the surface, in the office. Only the hourly check is required for our particular job.
On the last day, during changeover, the online instructions directed us to perform a task that was not supposed to be performed any more. Consequently, this wreaked havoc with the data collection electronic units and the entire system started freezing up. A full reboot was required. Everything was working by the end of the night.
All done with shift and we are going home soon!
Here are some pictures of Japan and KSU!