FY04 Si Dead Layer and Energy Calibration Studies

To first order the Si detectors were calibrated using an Am 5.486 MeV α source. Corrections were also made for energy lost in the Si dead layer. The thickness of the Si dead layer was not well known. There was also some ambiguity in the timing of signals from recoil carbons, and this varied strip by strip due to differences in cable length. A two parameter fit was performed on several runs to determine the energy loss in the dead layer, Edead, and relative time offset, t0.
    For the "on-line" polarization values, a linear approximation to the incident energy vs. energy deposited in the active Si region is used. A method using higher order polynomials to fit the correlation has also been studied. A brief description of these methods, as presented to the RHIC Spin Collaboration on April 1, '04, can be found here (MS Power Point).


  • α source calibration constant

  • "On-line" fit parameters vs. run number

    • Fitting Edead from α:
    Edead from the α's were fit for, ~350 keV < Ekin < ~550 keV. Only this low energy region was used because the α distribution began to run into the lower edge of the time window at higher energies. The t0 values from the fit to the carbons were used for this fit, making Edead the lone free parameter. Edead from carbon are also shown on the plot for reference
  • Edead from fit to α's vs. run


    • Sensitivity to Edead:
    Some studies have been done to determine the sensitivity of P and <AN> to the two different methods described in the Power Point presentation mentioned above (click here to download). I also studied the sensitivity to different various deadlayer thicknesses. I used:
  • a fixed thickness of 35 μg/cm2 (= 150 nm) for all strips, all detectors
  • add/subtract 10 μg/cm2 to the thickness given from the fit using the non-linear method
  • add/subtract 20 μg/cm2 to the thickness given from the fit using the non-linear method

    • Plots of the relative shift in P and <AN> for the different methods and deadlayer thicknesses can be found at the link below. relative shift of P and <AN>
    The relative shift is defined as:   (Pn-l - PX)/Pn-l
       where Pn-l is the polarization using the fit parameters from the non-linear method.
       and PX is the polarization using another method or deadlayer thickness.

    A comparison of how well the standard method and the non-linear method fit the data for one of the central strips is found below.
    comparison of fits   x axis = raw amplitude (i.e. ADC value),   y axis = raw time × (raw amp)½
    Lines are also drawn on the plot for tdead +/- 10 and +/- 20 μg/cm2. There was no quantitative measure of how good a fit was, but just judging by eye you can probably determine that a fit is good within +/- 10 μg/cm2.


    maintained by: Jeff Wood, wood@physics.ucla.edu
    last updated: 8/16/04