The TA Energy Scale

1
The TA Energy Scale

• Douglas Bergman
• Rutgers University
• Aspen UHECR Workshop 2007
• 17 April 2007

2
Introduction

• TA is just now being deployed
• Cant say honestly what we will do
• Can only say what I think we should do
• TA is a Hybrid Detector
• Use fluorescence detector (FD) to calibrate
  surface detector (SD)
• This limits model dependence in SD reconstruction

3
Outline

• Working from what we know to what we wish to know
• Fluorescence Detector Energy Scale
• Surface Detector Energy Calibration using the
  Fluorescence Detector
• Using the Surface Detector at the High Energies

4
Outline

• Working from what we know to what we wish to know
• Fluorescence Detector Energy Scale
• Surface Detector Energy Calibration using the
  Fluorescence Detector
• Using the Surface Detector at the High Energies

5
The FD Energy Scale

• From the shower to the data (and back again!)
• Missing energy and the EM portion of the shower
• Depositing shower energy in the atmosphere
• Fluorescence yield
• Atmospheric losses
• Light collection
• Photometric scale
• Shower parameterization and fitting
• Verification with MC (Data/MC comparisons)
• Finding the aperture (more Data/MC comparisons)

6
Missing Energy

• Only the EM portion of the EAS is accessible to
  FD
• Calculate using Corsika and QGSJet
• Depends on composition
• Composition from Xmax distribution in Data/MC

Song et al, APP 14 (2000) 7
7
Missing Energy

• Only the EM portion of the EAS is accessible to
  FD
• Calculate using Corsika and QGSJet
• Depends on composition
• Composition from Xmax distribution in Data/MC
• There are measurements
• Do we believe Yakutsk?

Knurenko et al, NPB(ps) 151 (2006) 92
8
Energy Deposited in Atmosphere

• The energy deposited by an electron depends on
  its energy
• The average electron energy depends on the age,
  but also on the model
• At any age have given average energy

Song et al, APP 14 (2000) 7
9
Energy Deposited in Atmosphere

• The energy deposited by an electron depends on
  its energy
• The average electron energy depends on the age,
  but also on the model
• At any age have given average energy
• Average over all ages
• New version of QGSJet gives 10 different ltdE/dxgt

Song et al, APP 14 (2000) 7
10
Fluorescence Yield

• Each MeV deposited in the atmosphere gives some
  number of fluorescence photons
• Use fit to FLASH, Nagano and Kakimoto results
• This is
• Kakimoto (1.000.06)

11
Atmospheric Losses

• Not all fluorescence photons reach the detector
• Rayleigh scattering
• Slowly varying over small range
• Aerosol scattering
• Varies dramatically
• Have to measure the VAOD
• Will use a measurement scheme similar to the in
  HiRes

Abbasi et al, APP 23 (2005) 157
12
Atmospheric Losses

• Not all fluorescence photons reach the detector
• Rayleigh scattering
• Slowly varying over small range
• Aerosol scattering
• Varies dramatically
• Have to measure the VAOD
• Will use a measurement scheme similar to the in
  HiRes

Abbasi et al, APP inpress
13
Light Collection

• We have to measure the amount of light reflected
  by the mirrors, and the amount lost at the filters

14
Photometric Calibration

• The response of each PMT must be adjusted and
  calibrated.
• Can calibrate many mirrors with one source at
  Middle Drum!
• Will include mirror and filter losses

15
Photometric Calibration

• The response of each PMT must be adjusted and
  calibrated.
• Can calibrate many mirrors with one source at
  Middle Drum!
• Will include mirror and filter losses
• Will also have an electron beam!
• End-to-end calibration with a known Ne

16
Fitting Showers

• Have to fit profile in data to a parametric form
• Gaussian in age
• Gaisser-Hillas

17
Fitting Showers

• Have to fit profile in data to a parametric form
• Gaussian in age
• Gaisser-Hillas
• Gaisser-Hillas works well for HiRes, well use it
  in TA too

18
Data/MC Comparisons

• Have to put all of the above into the detector
  simulation
• How does one know that the simulation is right?
• Any and all distributions observed in the data
  should be able to be reproduced by MC
• Any distribution that does not agree between data
  and MC is a systematic error
• To the level one can tell that distributions
  agree lets one assign systematic uncertainties

19
The Aperture Calculation

• For FD, MC is needed to calculate the aperture
• How can we be confident in the calculation?
• Data/MC comparisons
• Which distributions are most important?
• How far away (RP)
• What directions (? or ?)
• How much light (shower brightness)
• Xmax (if cutting on shower shape parameters)

Abbasi et al, APP inpress
20
Outline

• Working from what we know to what we wish to know
• Fluorescence Detector Energy Scale
• Surface Detector Energy Calibration using the
  Fluorescence Detector
• Using the Surface Detector at the High Energies

21
From the FD to the SD Energy

• Geometric Reconstruction
• Hybrid reconstruction gives both FD and SD
  reconstruction resolution
• SD Energy Reconstruction
• MIP normalization of counter response
• S(1000) vs EFD
• Attenuation correction
• LDF slope compared to Xmax
• Rise time compared to Xmax (no muons!)

22
Finding S(1000)

• Have to fit Lateral Distribution Function to find
  S(1000) or S(600)
• Use NKG parameterization
• Can use FD energy to check linearity of S(x) vs E
  relation

Yoshida et al APP 3 (1995) 105
23
Attenuation Corrections

• Can get attenuation correction from data using
  hybrid without making constant intensity cuts
• But only at energies with sufficient hybrid data

Sakaki et al 27th ICRC (2001) 333
24
Rise time compared to Xmax

• TA can measure the rise time of the shower front
  (and how it changes with distance from core)
• Can be used to measure composition

Walker Watson JPG 7 (1981) 1297
25
Outline

• Working from what we know to what we wish to know
• Fluorescence Detector Energy Scale
• Surface Detector Energy Calibration using the
  Fluorescence Detector
• Using the Surface Detector at the High Energies

26
SD at the Highest Energies

• Highest half-decade in energy is just SD
• 10 FD duty cycle
• Extrapolations
• Extrapolate S(1000) vs E and attenuation
• Base the extrapolation on MC
• Check MC by Data/MC comparisons
• LDF Slopes
• If linearity changes differently between data and
  MC, will LDF slope distribution show it?
• Rise times
• Showers only accurately modeled if rise times
  agree in data and MC
• Zenith Angle Distribution