Isolated electrons/photons separation based on VD-VFT space points

A package is available which performs an electron/photon separation in topologies where they are isolated and highly energetic (E>10 GeV), in the polar angular region covered by EMF and HPC calorimeters, looking at VD hits or VFT space points along the trajectory between the beam spot and the electromagnetic shower position, taking into account the B field bending.

Last update July 20, 2000

Version v07:

In this version, which superseed the old one, the functions ACOS and ATAN are used instead of ACOSD and ATAND, to assure Linux compatibility of the code.

How to use the package:

  1. The code is located on /afs/cern.ch/user/v/verlato/public/vft/siveto_v07.car

  2. The VFT pam's, alignment and databases files needed by the code are located on /afs/cern.ch/delphi/tasks/mvtx/vfpams
    Also the pam file vdkeep.car located on /afs/cern.ch/delphi/tasks/mvtx/vdpams has to be read.

  3. Remember to put, at the initialization stage of your program, the calls: 
         CALL VDIDST
     CALL VFIDST
    and inside the event loop the calls: 
         CALL VDALUN(LUN)
     CALL VDRDST
     LUNM=31
     LUNP=32
     CALL VFALIN(LUNM,LUNP)
     CALL VFRDST(LUNHOT)
    LUN and LUNHOT are dummy integer arguments.

  4. The key routine is

    CALL SIVETO(THETA,PHI,ENE,IBS,VDGAM,NHI,IPAT,IRVF,IZVF,IRVD).

    THETA and PHI are the polar and azimuthal angles given in radians and according to DELPHI conventions (0<THETA<PI and 0<PHI<2PI).
    In principle they can be any direction, but obviously the routine is meant to have as input the direction of a HPC/EMF shower or cluster of showers of total energy ENE given in GeV.
    The helix trajectories (one for each charge sign) passing through the beam spot and shower position at calorimeter surface are used, as well as a straight line extrapolation, to look for VD-VFT hits confirmation, as is shown in the figure here. The input directions given to the routine should be corrected for the beam spot position. For example, the directions taken from QHPC or QEMF vectors in the skelana common refers to DELPHI origin, so the real direction of the primary particle must be calculated from the beam spot position and the HPC radius or EMF surface position. If you don't apply this correction already in your code, the routine can do it for you setting then input variable IBS=1

    The output logical variable VDGAM is set .FALSE. if there are:

    1. at least 2 VD R-phi hits in 2 different layers within 0.30 cm (XY plane distance) from the straight line or the helix curves

      OR

    2. at least 1 VFT hit within 0.30-0.30-0.40-0.40 cm (XY plane distance) AND 0.30-0.30-0.30-0.30 cm (Z distance) respectively according to the VFT layer, or 1 VD hit in the first plaquette of the inner layer within 0.30 cm (XY plane distance) from the straight line or the helix curves

    The output variable NHI gives the total number of hits satisfying the cuts written above.

    The output variables IPAT and IRVF,IZVF and IRVD give the hits pattern and Rphi and Z residuals bitted in a way described in the code.

  5. In the skelana cradle the following lines must be included: 
         +use,p=vdkeep.
     +use,p=vdkeep,d=vdblck,t=i.
     +pam,11,t=a,c,r=vdkeep.           vdkeep.car
     %
     +use,p=vfkeep.
     +pam,11,t=a,c,r=vfkeep.           vfkeep.car
     %
     +use,p=sitcom,t=e.
     +use,p=siveto,d=siveto,t=e.
     +pam,11,t=a,c,if=unix.            siveto_v06.car
    and you must load the VFCLAP library. In addition, since several routines rely on VDCLAP variables, the user has to load VDCLAP library too.

  6. Concerning the Rphi VD hits, a cut on the signal to noise ratio S/N>8 is applied, based on the observation that a big amount of noise/background hits populate the region below 8, as can be seen from this plot, where the normalized distributions of S/N for associated (open area) and unassociated (shaded area) hits for DATA and MC 97 samples are shown.

  7. The following files compares the Rphi residuals, for Bhabhas DATA and MC samples for processing 97_C1 at Ecm=184 GeV, calculated in the following way:
    in the first case (left side of the plots) a straight line from calorimeters to beam spot is traced, and the Rphi residual is the distance from the VD-VFT space point and this line. Since Bhabhas at high energies are almost t-channel only, it means that electrons go mostly forward and positrons mostly backward, so you can clearly see the B field effect separing the two hemispheres;
    in the second case (right side of the plots), the B field effect is accounted for tracing two helix (one for each charge ipothesis) from calorimeters to beam spot where the curvature is determined from the transversal shower energy. The Rphi residual is now the minimum distance between the VD-VFT space point and the three trajectories;

  8. The following files show some results for Bhabhas DATA and MC samples for processing 97_C1 at Ecm=184 GeV, and an e+e- --> gamma gamma (gamma) MC sample produced at Ecm=183 GeV. Both linear and helix extrapolations (two helix extrapolations, one for each charge ipothesis) from calorimeters to VD-VFT layers are performed, and the Rphi residuals show the minimum distance among the VD-VFT point and the three trajectories.

    all the results refers to events passing cuts similar to the selection A of Method 2 extended down to 10° and without the VD-VFT hit requirement. This selection is essentially based on the following cuts:

    1. E1>75%Eb
    2. E2>55%Eb
    3. Acollinearity < 20°

    For the polar angle region above 40° the events were required to pass also the selection B of Method 2, which is based on charged tracks (with TPC) only (of course this is not applied to the photon efficiency plots). This suppress the background from QED events with two photons in the final state which were selected from the "modified" selection A based on calorimeters only. The global selected sample includes ~1% of QED events with two photons in the final state for theta < 25° and ~4% for 25° < theta < 40°.

  9. A few comments about the VFT and VD efficiency plots:
    what i called global efficiency include the effect of the first plaquette of VD Inner Layer, extending down to 20°;
    the efficiencies for each VFT layer vs. the azimuthal angle are given for the corresponding polar angle geometrical acceptance;
    the efficiencies for each VFT layer vs. the polar angle are given for the efficient sectors as seen from the azimuthal angle plots;
    the apparent efficiency loss of the MC VFT as compared to the REAL DATA is difficult to explain as a noise effect on the data, as can be seen from the residuals plots, and up to now is not yet understood.

  10. The program works running on LONG and XS DST data type, and the VFT information on the DATA is only included since the cassette LONG96_E1/C269-.
Questions and comments to Marco Verlato (verlato@padova.infn.it)

Last update July 20, 2000