User:MB Lab PP
Log Book Outline of the P335 Project By Henry J. Frisch, frisch@hep.uchicago.edu HEP 320, EFI, 5640 S. Ellis Ave., Chicago, Il. 60637 Phone:773-702-7479 | Fax:773-834-5959 Goals: Take a set of shifts and get some hands-on experience taking data at a hadron collider. Enumerate possible causes of the discrepancy in the UC CDF Charged Higgs analysis. Investigate and understand detector effects that could cause the discrepancy: e.g. tracking errors, cosmic rays, decays-in-flight, multiple-events, detector readout errors, ... This involves understanding energy loss by muons, hard bremstrahlung, the kinematics of decay-in-flight, timing measurements in calorimeters and tracking, including velocity measurement of the track, etc. Try to confirm or eliminate each explanation. Calculate a limit on the contribution from each. Write it up as a CDF note and 335 report. Reading: CDF TDR Steve Levy CDF notes and PRD draft; Rossi Perkins Some original papers on fundamentals (e.g. energy loss) Feng and Strassler Papers Project Requirements Sent to me by Jim Pilcher, j-pilcher@uchicago.edu HEP 203, EFI, 5640 S. Ellis Ave., Chicago, Il. 60637 Phone:773-702-7443 Guidelines Topic should involve "hands on" work with apparatus and data Work should be separate from any RA activity Work should not be on PhD thesis topic Fall Quarter Identify topic and advisor Prepare 1 page summary of project signed by student and advisor Submit by November 20 I will review and approve projects in time for modification or registering for 334 Begin work on projects Winter Quarter Meeting early in quarter with 10 minute oral presentations on topics Complete bulk of work on projects Oral presentations of 15 minutes late in quarter Spring Quarter Complete work on projects Submit 10 page final report by May 18 Poster session Friday, June 1 Gathering Information Using various sources, llisted below, I began my research by learning and revising essential facts about particle physics, as well as detector physics. Reading Griffiths, David J. Introduction to elementary particles. New York : Wiley, c1987. National Research Council (U.S.). Committee on Elementary Particle Physics in the 21st Century. Revealing the hidden nature of space and time : charting the course for elementary particle physics / Committee on Elementary Particle Physics in the 21st Century, Board on Physics and Astronomy, Division on Engineering and Physical Sciences. Washington, D.C. : National Academies Press, 2006. Perkins, Donald H. Introduction to high energy physics. Cambridge ; New York : Cambridge University Press, 2000. Rossi, Bruno Benedetto. High-energy particles. New York, Prentice-Hall, 1952. Papers Han, Tao. Collider Phenomenology, Basic Knowledge and Techniques. arXiv:hep-ph/0508097 v1, 9 Aug 2005 Willenbrock, Scott. Symmetries of the Standard Model. arXiv:hep-ph/0410370 v2, 24 Feb 2005 Martemyanov, B.V. Nine lectures on quark structure of light hadrons. arXiv:hep-ph/9903256 v1, 4 Mar 1999 Internet Resources ftp://root.cern.ch/root/doc/Users_Guide_5_14.pdf from the CERN Root Homepage http://en.wikipedia.org providing various definitions http://www-cdf.fnal.gov/upgrades/tdr/tdr.html has the Technical Design Report of the CDF II Detector by the CDF II Collaboration http://hep-www.colorado.edu/~will/fmfman.ps for the manual for creating Feynman diagrams Programming with ROOT In order to analyse the data we first need to get to grips with using ROOT. Browse .root files in ROOT The command line is as follows: >source ~cdfsoft/cdf2.cshrc >setup cdfsoft2 6.1.4 >root -l x.root ... root[1] TBrowser b Inside the browser is the complete information of the ntupled data View and Safe Histograms HMet_met: Inside the subdirectory met, we choose met once again and are presented with a histogram, which we can save and is shown below: [insert image: /home/cdf/bock/Projects/muon_1jet_tautype2/Otherchicago_Events_1/HMet_met.eps] Note that .eps is the most convenient file format. Similarly we can view the time of flight in each event, HTof_ntm: [insert image: /home/cdf/bock/Projects/muon_1jet_tautype2/Otherchicago_Events_1/HTof_ntm.eps] Ntupling data files For convenience individual event files of one or more runs are grouped in so called ntuples. There are different types of ntuples and in this example we shall ucntuple our data files. First, we look for a machine which has SLF305 running and is available for our purposes. Under http://hep.uchicago.edu/help/hardware/wksta/ we find a list of machines and whether they are in queue or not. In our case cdf49 happens to be available. We logon via the terminal window: >ssh cdf49 cdf49> Once loged on we go to our folder the individual data files and setup the run time environment (see 3.1). We can view more detailed information about the files before ntupling via the following command: cdf49>Edm_EventLister bhm* Having taken one last look at our chosen events, we are now ready to ntuple: cdf49>cd /cdf/s3/cwolfe/6.1.4_ntuple_emtiming_caf/ cdf49>setenv FLINPUT /cdf/home/cdf/bock/Projects/muon_1jet_tautype2/bhm* cdf49>setenv FLOUTPUT /cdf/home/cdf/bock/Projects/muon_1jet_tautype2/ucntuple_tester.root cdf49>ucntuple.exe FlatNtuple/tcl/53X_multi.tcl If we return to our directory we find the .root file and can view it accordingly. Extracting information of .root files outside root With a development environment setup, we can code C scripts that will extract and manipulate any of the information contained in the ntuples. The disadvantage with this procedure is that any changes to the .C file only become effective after compiling it, where as in ROOT there is no need to compile. The template file is a good starting point. We shall create a copy and edit it for our specific needs: >cd /cdf/s4/bock/devel/ >cp template.C showRunsAndEvents.C >emacs showRunsAndEvents.C & The last command opens the editor. To simply view all the run and event numbers we can use the simple script, which we insert into the test class:
// display run and event numbers cout << "Run: "
<< uber.header.run << " Event: " << uber.header.event << endl; Notice that our variables correspond to the ntuple structure, i.e. [uber.subdirectory.histogram]. To see if everything works we need to compile the file. To do this we first change one line of Makefile to: PROGRAM = showRunsAndEvents entering the name of our .C file. Now we can close the editor and compile and run our little program: >make >showRunsAndEvents -i ../../../home/cdf/bock/Projects/muon_1jet_tautype2/ucntuple_tester.root You will see a list displaying for each event the run number and its event number. Merging root files containing nTuple and Histograms Use the following script after setting up the run-time environment for root: hadd result.root file1.root file2.root filen.root Run a C script in ROOT Instead of entering line by line into ROOT, we can also create a .C file with the root commands as r_graph.C does: void r_graph() {
TF3 *f = new TF3("f","sin(x*x+y*y+z*z-36)",-2,2,-2,2,-2,2); f->SetNpx(10); f->SetNpy(10); f->SetNpz(10); f->Draw("A");
return; } After setting up the runtime environment (see 3.1) we add the following to the root command: >root r_graph.C And thus without compiling we can make changes to the script without having to enter the whole script over and over again. Creating the Release Directory A good way to find the ht values for the ucntuples is by creating the release directory after setting up the runtime environment: >newrel -t 6.1.4 release >cd release/ >cvs co -r HEAD FlatNtuple >cd FlatNtuple/anal/ In this directory we find all the important files to compute the ht. Create .log Files of ntuples As we are looking at multiple entries of the nuple we shall run a .C file to avoid having to reenter the same commands for every event we would like to make the log file of: //running inside ~/Projects/muon_1jet_tautype2/work/champNtuple/
void dumpEventsOfNtuple() { gROOT->ProcessLine(".X ChampsNtAna/scripts/LoadLibs.C"); TChain* t = new TChain("ChampNtModule/Champ"); t->Add("/home/cdf/bock/Projects/muon_1jet_tautype2/Otherchicago_Events_0/otherchicago_events_0.root"); ChampNtCommon cnt(t); for(i=0;i<60;i++) { cnt.Dump(i); } return; } In order to run the .C file, we enter the following command into the terminal >source makeDumpEvent1.tcsh Which is a batch with the following content:
- !/bin/tcsh
source ~cdfsoft/cdf2.cshrc setup cdfsoft2 6.1.4
- goto champNtuple/
cd ~/Projects/muon_1jet_tautype2/work/champNtuple/
- root command
root -l -q -b dumpEventsOfNtuple.C >&! dumpEvent1.log Note: The ChampNt scripts have been provided by Dr. Thomas Phillips (Thomas.Phillips@duke.edu). Chi-Distribution Histogram See chi.C for details. Displaying the Data For data analysis it is important to be able to display the information. The Detector Display In order to view our events we setup our run-time environment and open the individual data file: >evd bhm012345.0123456 A panel will open up and under the window tab we can choose various types of displays. Creating Feynman Diagrams In order to make sophisticated Feynman diagrams we use the LaTeX / TeX package feynMF. Once having written the .tex file, we compile it into a .ps using the following steps: >cd ~/Latex/work >latex ../papers/feynex Where /work is the location of your latex libraries and feynex is the name of your latex file. To create the diagrams in the .tex file, we need to address the diagrams individually: >mf diagram_one.mf Now we create the final .dvi file with the diagrams and convert the .dvi to a .ps file: >latex ../papers/feynex >dvips feynex Alternatively to dvips we can enter pdflatex to get the output file in a .pdf format. Analysis of the SS Muons, 1 Jet, TauType1 We have three possible theories for the anomalous tau production in b-tagged Top quark events using 955pb-1 of data: There has been an error in the readings of the detector The expectation model is not applicable/erroneous We are dealing with new physics In our analysis, we shall look at each possibility seperately, starting by taking a careful look at our data. Collecting the Individual Data Files After logging into the fcdflnx6.fnal.gov shell, we extract the data found under /cdf/scratch/carla/steve/ via: fcdflnx6:/cdf/scratch/carla/steve>scp strip_bhmu0* bock@cdf49.uchicago.edu:/cdf/s4/bock/ Now we classify the data in four categories depending on the number of jets and the tau type. Extract Data Files from the Database Make a scratch book and then copy individual data files into the directory from the list of events provided by Collin. fcdflnx6:~>cd /cdf/scratch/frisch/ fcdflnx6:/cdf/scratch/frisch>mkdir bock fcdflnx6:/cdf/scratch/frisch>cd bock fcdflnx6:/cdf/scratch/frisch/bock>scp bock@cdf17.uchicago.edu:~cwolfe/topana/bock.txt fcdflnx6:/cdf/scratch/frisch/bock>less bock.txt fcdflnx6:/cdf/scratch/frisch/bock>head -1 bock.txt 195678 12026850 fcdflnx6:/cdf/scratch/frisch/bock>source ~cdfsoft/cdf2.cshrc fcdflnx6:/cdf/scratch/frisch/bock>setup cdfsoft2 6.1.4 fcdflnx6:/cdf/scratch/frisch/bock>setenv USE_SAM_METADATA 1 fcdflnx6:/cdf/scratch/frisch/bock>pickEvent -r 195678 -e 12026850 -d bhmu0h SS Muons, 1 Jet, Tau Type 1 We make the log entries as described in section 3.7 of the 10 anomalies with 1 jet and tau type 1. From table 7 of Updated Measurement of Anomalous Tau Production in b-tagged Top Quark Events Using 955pb-1 of Data by Henry Frisch and Stephen Levy, we have one of the highest discrepancies between the data and the expected results. We expect to see 1.81±0.26 and find 10. The ten events are listed below. We shall now take a look at each event and see why they have been selected:
Table 1: List of ss muons, 1jet, tautype1 with their Ht and MET Selection Process The criteria for events being selected in the decay of pair-produced top quarks are as follows: There is either a tau, an electron or a muon We have large missing transverse energy Two or more jets with at least one bottom quark*
- The bottom quark is identified using a secondary vertex tagging algorithm.
Note: These criteria have been taken from Henry Frisch and Stephen Levy's Paper (see 4.2.0). Analysis of the 10 Events Our first step is check the data for any irregularities. Event bhmu0d.165873.1726584 Plot t0 by SL For each of our event we analysis the values of t0 by SL and notice by mere inspection that they are no clear ascending order, except if we consider them pairwise; i.e. 1 and 2, 3 and 4 ... up to 7 and 8. We shall use the globCotT0 value as our global t0. Plot of the t0's For the same sign muons with one jet and of tau type 1, we get the following plot:
Note: We only have 9 entries as we do not have the champ analysis for the event 11913629 of the run 198695. Data sets of ttbar events root://fcdfdata047.fnal.gov//cdf/scratch/top/scratch/results/topnt/ttbar/topnt_Inclusive_from_toronto_531_latest_ttbar_*.root dsfg