Ryan McGreal IRES at CERN: LHCb Summer Internship Note

Transcription

1 Ryan McGreal IRES at CERN: LHCb Summer Internship Note I. Introduction The Large Hadron Collider beauty experiment is designed to study the decay of B mesons. B mesons are particles which contain the b, sometimes referred to as bottom or beauty, quark. The LHCb experiment measure Charge Parity violation in hopes of better understanding the matter antimatter asymmetry of the universe. II. Detector Function The LHCb detector is a narrow angle particle detector which only used downstream data. Particles travel in opposite directions and interact in the vertex locator, or Velo. The products of the particle interactions will travel both to the left, upstream, and right, downstream, of the Velo as shown in the schematic below.

2 Image is in the public domain, created by CERN. Only the downstream, or right hand side, data is taken as the detector is oriented so that the particle collisions occur at the far left end of the detector. Protons travel through the center of the detector through the tube shown in the diagram. The LHCb detector is called a narrow angle detector as the detector components downstream form a small angle as compared to other particle detectors with the particle collision locations in the Velo, as can be seen with the schematic. The reason for this is that B mesons which travel downstream after a proton proton collision will likely deviate very little from the proton beam. Each layer of the detector is designed to measure different properties of the products of the proton proton collisions. This allows for the energy, momentum, path and charge of the produced particles to be detected which allows for the identity of the products to be known. III. Vertex Locator For my work I compared the location of particles interactions within the Velo to those generated through Monte Carlo simulations. The purpose of this type of comparison is to determine whether there is a beam misalignment as well as checking that the series of

3 silicon modules which make up the Velo have a uniform material distribution. A beam misalignment would systematically shift the particle interaction points, while a material distribution issue would result in hot spots and dead zones in the Velo data. IV. Methods Data was generated by files kept in the LHCb book keeping system at CERN. The files I used can be seen in the program GangaVertexDown, labeled 1 in the appendix. GangaVertexDown is run with ganga in root. It is to note that the macro shown in appendix 1 was provided by my supervisor Walter Bonivento as well as Victor Coco. GangaVertexDown can pull n-tuples from actual data files as well as from Monte Carlo simulation files. The n-tuples used were the z position of events as well as the r position. The r position is defined as r=sqrt(x^2 +y^2), where the y position is shown on the diagram above and the x position is the distance from the beam pipe into and out of the page. A graph of r vs vz shows the distribution of events. If this is different from Monte Carlo simulations then effort must be made to try to explain these differences. V. Results What I found is a lack of data points near the 1000mm mark along the z axis and between 30 and 35 mm along the radial, r, direction. This can be most clearly seen in the 2011 data with the beam run at 3.5 TeV. For the following graphs the y axis represents r and the x axis depicts z. The unit used is millimeters. Plot 1: r vs z

4 This can be seen more clearly in the next plot, which is zoomed in on the area of question. A box is used to indicate the area of the graph which is lacking points. Plot 2: A zoomed in portion of plot 1

5 VI. Discussion Initially I thought such a gap was the result of not having enough data. After creating plots from different runs I found that this gap remained in every plot. I chose the 2011 data show above as it was the plot which I had the most entries for. The presentation Tomography using downstream tracks by Victor Coco, NIKHEF, from October 11, 2012 suggests that this gap may be caused by the beam pipe exit, however further

6 consistency checks were needed. VII. Future Research In order to better understand any inconsistencies with the data generated by the Velo I have developed macros for each set of n-tuples created per file found in the LHCb book keeping. These macros will allow for large data sets to be managed, as the methods used to create plot 1 and 2 would not be suitable for very large samples. Use of such macros could compare data from different runs. Appendix 1 GangaVertexDown Provided by Victor Coco params = { 'OutputName' : 'NTUPLES.root', 'Mag': 'mu', 'Simulation': True,'Year':'2012' } myfiles = {'MC':{'2011':["/MC/2011/Beam3500GeV-2011-MagDown-Nu2- Pythia8/Sim08a/Digi13/Trig0x /Reco14a/Stripping20r1NoPrescalingFlagged/ /ALLSTREAMS.DST", "/MC/2011/Beam3500GeV-2011-MagUp-Nu2- Pythia8/Sim08c/Digi13/Trig0x /Reco14a/Stripping20r1NoPrescalingFlagged/ /ALLSTREAMS.DST"], '2012':["/MC/2012/Beam4000GeV-2012-MagDown-Nu2.5- Pythia8/Sim08c/Digi13/Trig0x409f0045/Reco14a/Stripping20NoPrescalingFlagged/ /ALLSTREAMS.DST", "/MC/2012/Beam4000GeV-2012-MagUp-Nu2.5- Pythia8/Sim08a/Digi13/Trig0x409f0045/Reco14a/Stripping20NoPrescalingFlagged/ /ALLSTREAMS.DST"]},

7 'Data':{'2011':["/LHCb/Collision12/Beam4000GeV-VeloClosed- MagUp/Reco14/Stripping20/ /EW.DST"], "2012":["/LHCb/Collision11/Beam3500GeV-VeloClosed-MagUp/Real Data/Reco14/Stripping20r1/ /EW.DST"] } } for sim in ['MC']: for year in ['2011','2012']: for mag in ['MagUp','MagDown']: for k,sample in enumerate(myfiles[sim][year]): magnet = 'mu' if mag == 'MagDown': magnet = 'md' j = Job( application = Bender( version = 'v23r3' ) ) j.name = sim+'_'+mag+'_'+year+'_sample'+str(k) j.application.module = '/afs/cern.ch/user/r/rmcgreal/downvertex.py' j.backend = Dirac() ## Setup some parameters for the Bender Job params['year'] = year params['simulation'] = False params['mag'] = magnet params['outputname'] = 'VertexDown'+j.name+'.root' if sim == 'MC': params['simulation'] = True params['outputname'] = 'VertexDown'+j.name+'.root' j.outputfiles = [SandboxFile(params['OutputName'])] dataset = LHCbDataset() mypath = myfiles[sim][year][k].replace('magup',mag) bkq = None if params['simulation'] : bkq = BKQuery( dqflag = "ALL", path = mypath, type = "Path" ) else : bkq = BKQuery( dqflag = "OK", path = mypath, type = "Path" ) dataset=bkq.getdataset() print "I am here" j.inputdata = dataset if params['simulation']: j.splitter = SplitByFiles( filesperjob = 20 ) else: j.splitter = SplitByFiles( filesperjob = 20 ) j.application.params=params j.submit()