Reimaging Lens Polarization

1
Initial Data Reduction for the FPP
Spectro-Polarimeter October, 2004
Bruce W. Lites 303 497 1517 lites_at_ucar.edu
2
FPP Spectro-Polarimeter Data

• OBJECTIVE
• The objective of the initial processing is to
  prepare the FPP-SP data in a form suitable for
  scientific data analysis.
• PROPOSED METHOD
• I propose to adapt extensive data reduction tools
  developed for the Diffraction-Limited
  Spectro-Polarimeter (DLSP) at the National Solar
  Observatory/Dunn Solar Telescope to the FPP-SP.
• DLSP is an instrument that evolved from the
  Solar-B Concept Model Spectro-Polarimeter
• DLSP codes written in general way to be easily
  adapted to other spectro-polarimeters
• Software both in IDL and FORTRAN (for speed of
  reduction)

3
Steps for FPP-SP Initial Data Reduction

• Prior to analysis of science data
• Prepare the dark and flat field correction images
• Prepare the polarization calibration matrix
• For science data, the steps in order are
• Dark and flat correct
• Apply polarization calibration
• Remove spectral skew
• Merge the two polarization beams
• Fringe removal
• Correct for spectral curvature
• Compensate for residual I?Q,U,V crosstalk

4
Illustration of the Reduction Process
I illustrate this data reduction scheme with the
procedure as adapted to data from the new
spectro-polarimeter at the Swedish Solar
Telescope (SST). These data from 1 April 2006
demonstrate the correction end-to-end resulting
in fully calibrated data.
5
Dark and Flat Field generation

• Dark measurements from ground-based instruments
  are simple just block the beam to the
  spectrograph.
• FPP-SP darks are problematic, no shutter.
• Flat field images for FPP-SP must average many
  independent images of quiet granulation near disk
  center.
• rms contrast of granulation expected to be 15
• To achieve flats accurate to 0.5 rms, need 900
  independent measurements of granulation
• Many coarse maps of quiet Sun required!
• Flat procedure for spectra obtain an average
  spectral profile from the average of flat images,
  then divide the spectrum by this profile to
  obtain the flat image. The multiplicative
  flat-field correction is the inverse of the flat
  image.

6
Multiplicative Flat Field Corrections
Dark 0 Flat 0 Dark 1 Flat 1

• Sample dark, multiplicative flat field images for
  the FPP-SP obtained in sun tests on 13 June 2005.
  
• Fewer flat images were needed than on orbit
  because the seeing was bad
• Dark images are very uniform
• Multiplicative flat images show little trace of
  spectral lines
• Flat images scaled 10
• Corrected image (Raw-Dark)xMflat

7
SST Flat Field Correction
Dark-corrected Flat Field Data
Multiplicative Flat Field Correction
8
SST Dark Corrected Data
Q
I
U
V
9
SST Dark/Flat Field Corrected Data
Q
I

• Opposite Q,U,V signatures in two orthogonal
  polarization image pairs

U
V
10
FPP-SP Calibration Matrices X-1
Smoothed
Residual
Original
Spectral ROI 112-224
CCDSIDE0
CCDSIDE1
11
FPP-SP Variation of X over Slit Scan Range

• Error Bars polarization matrix requirement
• Slit scan position -225 taken at low light
  level, so discarded

12
SST Polarization Calibrated Data
Q
I

• Same Q,U,V signatures in two orthogonal
  polarization pairs
• Symmetric Q,U
• Antisymmetric V
• Opposite seeing crosstalk in Q,U,V pairs

U
V
13
SST Skew Corrected Data
Q
I
U
V
14
SST Merged Data
Q
I
V
U
Seeing crosstalk eliminated
15
SST Spectral Curvature Removed
Q
I
V
U
16
SST High Sensitivity Q,U,V
Q
I
V
U
Q,U,V Grey Scale 0.5 Ic
17
SST Residual I?Q,U,V Crosstalk Removed
Q
I
V
U
18
Other FPP-SP Reduction Issues

• Slit Scan Vignetting
• Variation of SP throughput exists as a function
  of slit scan position. Also a 2-D variation vs.
  slit scan position (x) and distance along the
  slit length (y)???
• Polarized Spectral Fringes
• Known to exist in the polarization calibration
  matrices
• Smoothed over in the representation of the
  polarization calibration matrix as a function of
  (?,y).
• Refinement of the Calibration Matrix
• It is possible to use solar observations of a
  sunspot umbra to refine the polarization
  calibration matrix. This will be difficult in
  view of variations of the matrix in (?,x,y).

19
Slit Scan Vignetting
Scan Mirror Step Number
Scan Mirror Step Number
19 August 2004 NAOJ SP intensity vs. scan mirror
position before pre-slit repair. FPP on OBU with
solar feed.
26 May 2005 NAOJ SP intensity vs. scan mirror
position after pre-slit repair. FPP on optical
bench. Solar feed with telescope simulator.
20
Slit Scan Vignetting

• Additional measurements Careful observational
  study on-orbit of flat field observations taken
  over full range of slit scan positions
• Analysis Derive variation of intensity of these
  flat field observations as corrected by a flat
  field derived at the center of the scan range.
  Derive the normalization factor as a function of
  (x,y) ASP(x,y)
• Corrections Apply the normalization function
  ASP(x,y) to all FPP-SP map data. Applies equally
  to Stokes I,Q,U,V

21
FPP-SP Polarized Spectral Fringes

• Spectral fringes are apparent at the few x 10-3
  level (or less) in the calibration matrices
• These fringes are not represented in the
  interpolated, smoothed representations of the
  calibration matrices
• Will they show up in the final data on orbit?

CCDSIDE1, Spectral ROI 0-112
22
Polarized Spectral Fringes
An example of fringe removal from DLSP spectral
data. One must examine final calibrated data
from space to look for residual fringes of
concern.
Before Fringe Correction
After Fringe Correction
23
Reduction Code Strategy

• Preliminary analyses done in IDL
• Calculation of flat field corrections
• Calculation of vignetting corrections
• Preparation of polarization response matrix
  corrections

These corrections are determined only occasionally
24
Reduction Code Strategy

• Routine Map corrections done with FORTRAN code
  spawned from IDL control routine
• Application of dark/flat corrections
• Polarization calibration
• Skew removal
• Merging orthogonal polarization images
• Spectral curvature removal
• Residual I? Q,U,V crosstalk correction
• Vignetting correction
• Fringe removal?