GWinters041508
From MariachiWiki
Stony Brook course CEB558/PHY315: Hands-On Science with Cosmic Rays
Return to my course page. Return to previous week's class.
Contents |
Week 11: April 15, 2008: Start Last Project
This week we started out last project for the course. In my case, I continued with the same project, but other groups may have decided to change directions. Brad, from Bay Shore HS, has joined me in my quest to analyze some of the data that has been accumulating at the high schools. During class time I mostly read a couple of papers and worked with Brad to access data from Bay Shore HS. Since I didn't accomplish much during class (other than reading, but there isn't much to show for that), I did some analysis in the week before our next class.
Papers
Prof. Mike Marx had sent me links to two online references, so I read them.
A Cosmic Ray Muon Detector for Astronomy Teaching is from the Electronic Publications of the Astronomical Society of Australia. In it Clay (from Adelaide) et. al describe a muon detector very similar to our MARIACHI detectors, that they use for teaching purposes. They use a scintillation detector with photomultiplier tube (PMT), and, as I mentioned in an earlier posting, have found a 0.2% decrease in cosmic ray counts per millibar (hPa) of pressure. They have a nice explanation about the details of why there should be a decrease in cosmic ray count with increased pressure: basically, the atmosphere acts as an absorber for muons, so higher pressure (more atmosphere) results in more absorption of muons. They also talk about the effect of temperature. I have seen elsewhere (ref? maybe Joe Willie?) an assumption that temperature would affect the cosmic ray counts. In this paper they specify that the lab temperature (as opposed to atmospheric temperature) affects the photomultiplier/scintillator combination, but that this effect is not significant, or at lease not for the level of analysis expected at the undergraduate level. They also reference a site for real time neutron monitor data. This site has neutron data in graphical and tabular form, and says that "neutron monitor variation corresponds to 10-20 GeV primary cosmic rays." I don't know enough about either neutrons or about cosmic rays in the 10-20 GeV range to comment. They also reference NOAA's solar data site, a good place to start when looking for solar data.
The second link was to cosmicrays.org, which has short discussions on
- galactic cosmic rays vs. solar cosmic rays
- variations in cosmic ray counts with the solar cycle (long term: 11 year cycle; short term: day/night)
- Forbush decrease which follows solar flares
- cosmic rays and weather on earth
It's a good starting point and not too detailed.
Retrieving Data
Brad has joined me in looking at and analyzing data that we have been accumulating from detectors at our high schools. Brad is from Bay Shore HS, so I expect he will concentrate on Bay Shore HS data, and will hopefully be able to compare some of those results to the results I have found from Smithtown HS.
Our first task was to be able to access the Bay Shore HS data. Brad should be able to find all his data on the hard drive of the MARIACHI computer at his school. We spent some time trying to access it from the lab. I was able to access the database, so retrieved one day of data for him. Brad was not able to, presumably because 1) he doesn't have a grid certificate and 2) his (non-existent) grid certificate hasn't been added to the approved list for access to the data.
During the week that followed I exchanged e-mail with John Hover about access to the data. John has created an excellent tool, available to all, for graphing our cosmic ray data.
- &graph=1 include to get a graph; leave out to get a text file of the data
- &datatype=counts include to get just the data from the counts-yyyymmdd.txt file
- &datatype=events include to get just the data from the timestamp-yyyymmdd.txt file
At the moment the data retrieved from the counts file is compressed a little, but John says that later this week he will change the query to retrieve all of the original data. These data queries are available to everybody, so all students will be able to use them to get access to the data, regardless of whether they have a grid certificate.
Data Analysis
3-Fold, 4-Fold, and 5-Fold Coincidences
First I looked at the pressure dependence of 3-fold, 4-fold and 5-fold coincidences, and compared the relative changes in count rates. I used the same week of data: 3/3/08-3/9/08, with the same atmospheric pressure as before.
The number of 3-fold coincidences was different from the number of 2-fold coincidences in the number of permutations of the detector setups. While the Smithtown HS detectors are set up to count 2-fold coincidences only when an event occurs in detectors 1 and 2, 3-fold coincidences are counted each time events occur in detectors 1, 2 and 3; in detectors 1, 2 and 4; and in detectors 1, 2 and 5. So 3-fold coincidences are counted three times more often than 2-fold coincidences because of the 3 different permutations of the detectors making measurements. To account for the 3 different permutations, I used the number of 3-fold coincidences as the total number of 3-fold coincidences over a 30-minute period, divided by 3.
Similarly, there are three different permutations for 4-fold coincidence detection: 1-2-3-4 or 1-2-3-5 or 1-2-4-5. So I used the number of 4-fold coincidences as the total number of 4-fold coincidences over a 30-minute period, divided by 3.
Since there is only 1 possible permutation for 5-fold coincidences: 1-2-3-4-5, I used the number of 5-fold coincidences as the total number of 5-fold coincidences over a 30-minute period. Graphs are below.
The percent change for each of the 3-fold, 4-fold and 5-fold coincidences were similar, certainly similar within the 2 significant figure quoted elsewhere. Although the 3-fold coincidences had a larger absolute decrease in number of counts per unit of pressure, since there were more 3-fold counts detected, the percent change was similar to the percent change of the 4-fold and 5-fold coincidences.
Note that while the percent change for each of these multiple coincidences was similar, about -0.6%/hPa, it was a more dramatic difference than the percent decrease in the 2-fold coincidence rate: about -0.2%/hPa (both in this study and at Adelaide and by Joe Willie, as discussed in an earlier posting). I don't know why the coincidences over a larger area (3-fold, 4-fold and 5-fold) would be affected more by a change in pressure than the 2-fold coincidences (where the detectors are stacked on top of each other). Hmmmm...
 3-fold coincidences. Slope is -0.0049 cts/Pa or -0.57%/hPa |
 4-fold coincidences. Slope is -0.0012 cts/Pa or -0.64%/hPa |
 5-fold coincidences. Slope is -0.0015 cts/Pa or -0.64%/hPa |
Temperature Dependence
I started to look at the temperature dependence of the cosmic ray data. I don't have indoor temperatures (which, at school, may fluctuate by 10C or more in either the winter or summer, but less in spring and fall). I do have outside temperatures recorded at Mount Sinai, the same place where the barometric pressure was recorded. Graphing the corrected 1-2 coincidence cosmic ray data (corrected for pressure dependence, as outlined above) vs. outside temperature yielded the following graph:
The 2-fold coincidence rate does appear to decrease with increased outside temperature. At this point I am using corrected cosmic ray data, where the pressure dependence, found earlier, has been subtracted to get the "corrected" cosmic ray counts.
So now I subtracted the apparent outside temperature dependence, and got the graph that follows. If you look closely (it helps if you squint), it really looks like there is a diurnal variation, just like it said in the cosmicrays.org paper I was reading. That gives more motivation for learning R, so that I can figure out how to do a Fourier deconvolution "easily", without writing a routine for Excel. So much to do, so little time ...
In the meantime, I had talked to Mr. Rich Lefferts about rigging a temperature-controlled environment to test the dependence of the detectors on the local temperature. We talked about putting some detectors in a box, with an air conditioner or heating device and a thermometer and varying the temperature slowly. It occurs to me that it would make sense to use a total of 3 detectors, so that we could look at the efficiency of the middle detector as a function of temperature, independent of other factors such as ambient pressure. With luck we will rig something up at the next class.
Go on to next week's class.
