LETGS Calibration Overview and Status
LETGS Calibration Overview and Status
Updated Jul 24 2012
Links
Calibration Data page:
CALDB paths, filenames, dates, versions, and product files
ACIS Calibration
HRC Calibration
LETG Calibration Overview
There are two aspects to the LETGS calibration: the throughput of the
instrument, or "effective area" (EA), and the point spread function (PSF).
Since the instrument is primarily designed to obtain spectra of single
point-like sources, standard CIAO processing only considers the PSF
in detail along the dispersion axis--the "line spread function".
Cross-dispersion profiles for energies of interest may be estimated
separately using
MARX raytrace simulations.
The LETGS EA comprises several elements corresponding to
X-ray interaction with various components in the optical path, and
additional factors related to data extraction.
For mth order of a spectrum, these are:
- HRMA effective area
- LETGS diffraction efficiency for order m (GRating EFFiciency, GREFF)
- Detector QE (including filter transmittance and QE uniformity map)
- Order Sorting Integrated Probability (OSIP; ACIS only)
- Exposure corrections (FRACEXPO) for dither across detector gaps and
bad pixels
- Extraction efficiency of the spectral region
(Enclosed Energy FRACtion, EEFRAC).
There is no calibration product for LETG/detector EA
per se. Instead, Grating Auxiliary Response Files (GARFs) are
generated for each observation with a Spectral Response (SPECRESP)
column that includes all but the last of the factors listed above.
When fitting spectra to
derive source fluxes, the model spectrum factors in the incident
source spectrum, SPECRESP, and the Enclosed Energy Fraction (EEFRAC)
of the spectral extraction region that was used. EEFRACs are
interpolated from tabulated values in the LSFPARMs files and
incorporated in the grating Response Matrix Files (gRMFs).
Theoretical grating efficiencies for all orders were based on a
rhomboidal grating bar diffraction model calibrated pre-flight from
laboratory near infrared characterization of each grating facet and
on ground measurements (mostly of 1st and 0th orders) at XRCF and
Panter. Higher order efficiencies were modified later based on LETG/ACIS-S
flight data.
Extraction efficiencies were originally estimated using
MARX raytrace simulations.
The recent (2010) recalibration of the LETG/ACIS EEFRACs used higher fidelity
SAOSAC+MARX simulations with adjustments to match
flight observations; work on the LETG/HRC EEFRACS is in progress.
LETG/HRC-S
The HRC UV-ion shield (UVIS) filter transmission was measured in the lab
and at synchrotron facilities. Only the details of oxygen and
carbon K-shell edge structure have been modified since launch.
The HRC QE Uniformity (QEU) map is derived primarily from flat field lab data,
from which an energy-dependent spatial model was built. The spatial model
is constrained to be smooth over scales ∼20x line width.
HRC-S QE was originally based on lab and XRCF measurements, but
has been modified based on in-flight data
as follows:
- Fine tuning the edge structures at C-K, O-K, and Cs-M.
- Correcting the low-energy QE (at wavelengths > 50 Å, where
ground calibration is very difficult) based on
observations of the hot white dwarfs Sirius B and HZ 43. Current efforts are
underway to further modify the low-E QE
using updated white dwarf emission models.
- Correcting the QE from 6 to ∼40 Å by comparing the predicted
spectrum of PKS 2155-304 based on contemporaneous observations
with HETG/ACIS and LETG/ACIS.
Additional small scale corrections have been
made using observations of the bright blazar Mkn 421.
LETG/ACIS-S
The LETG/ACIS-S effective area is based on the LETG transmission
efficiencies for each diffraction order and on the
ACIS-S QE model.
The ACIS-S QE is based largely on laboratory and XRCF data obtained
for the bare ACIS chips, the optical blocking filter (OBF), and the
combination of both. An important additional component is the
contamination layer on the OBF that has been calibrated using a
combination of LETG/ACIS-S observations and bare ACIS observations of
supernova remnants and clusters of galaxies.
Calibration Status
| Product |
Last Update:
Date, CALDB |
Accuracy |
LETG/HRC-S
effective area
| QE: Dec 2009,
4.2.0
QEU: Jun 2011,
4.4.5
| 10-15% over whole range.
As noted above, the LETGS EA comprises several elements, including
detector QE and QE Uniformity.
HRC-S QE was recalibrated in 2009.
Time-dependent adjustments
(via QEU) were added in 2011. The effects of the
Mar 2012 HRC-S high voltage increase
are not yet reflected in the CALDB; QE is ~5% higher than before.
|
LETG/ACIS-S
effective area
| 2010
| 10% over most of the effective range; OSIP errors may be larger on
the S1 chip. Incorporates the
OBF contamination
model last revised in 2010.
|
LETG efficiency
(m=0,1)
| Dec 1999
| 10% for 1st order. GREFF version 3.
|
LETG efficiency
(m=2-10)
| Apr 2011,
4.4.3
| 5-10% relative to 1st order for mλ<80 Å.
GREFF version 7. (The GREFF file contains data out to m=25.)
|
LETG/ACIS spectral
extraction efficiency
| Dec 2010,
4.4.1
| In combination with the Nov 2010 CALDB 4.4.0
correction for the LETG/ACIS tilt,
the newly recalibrated Encircled Energy Fractions (EEFRACS)
in the LSFPARM files are accurate to 1% for observations made
with the standard setting of Yoffset=+1.5'. The default
spectral region width was changed from |tg_d|<0.0020°
to 0.0008° at the same time.
|
LETG/HRC spectral
extraction efficiency
| Jun 2001
| The LETG/ACIS EEFRACS recalibration revealed some small errors in
the LETG/HRC
EFFRACS; LETG/HRC Effective Area is not affected but the HRC-S QE has
corresponding errors.
An updated calibration of LETG/HRC-S EA, HRC-QE, and
LETG/HRC-S EEFRACs (for the standard Yoffset=0)
is in progress (2011).
HRC-S Degap Map
& Dispersion Relation
| Nov 2005,
3.2.0
| Typical errors are
0.004 Å
for λ<20 Å, a range with many strong lines
that can be compared to HETG/ACIS spectra.
At longer wavelengths, where wavelengths are less well known and
lines are relatively weak, making calibration of dispersion
non-linearities difficult,
errors are around 0.015 Å with some localized deviations
of 5-10 pixels (0.037-0.074 Å).
An analysis method that incorporates continuum spectra in
addition to line data is
under development (ps).
|
| Line Spread Function (gRMFs)
| March 2004
| ACIS/LETG: Grating RMFs based on MARX raytraces
match observed lines well.
HRC-S/LETG:
Imaging
non-linearities in the HRC-S distort and shift the line profiles
from their expected wavelengths in some wavelength regions,
so care is needed in applying gRMFs directly, e.g.,
through XSPEC or Sherpa model fitting.
|
| HRC-S gain map
| Nov 2008,
4.2.0
Jul 2012,
4.5.1
| Relative accuracy of 2% with 1/3-tap grid (0.63 Å without dither),
but somewhat worse near the optical axis and plate ends.
This new gain map is time dependent and was implemented in
CIAO/CALDB 4.2.0 at the end of 2009, along with a new
HRC-S/LETG background filter.
The HRC-S voltage was raised
in Mar 2012 and a new gain map for the higher voltage was
released in Jul 2012. An improved gain map will be released
once temporal trends are better established.
|
HRC-S/LETG
background filter
| Nov 2008,
4.2.0
| This filter is used in conjunction with the new time-dependent gain map
and removes more than half of the L2 background at most wavelengths
with a 1.25% loss of 1st order X-ray events. The filter is still
safe to use with the new HRC-S high voltage and gain map, but
will be slightly less effective at removing background because of
the preliminary gain calibration.
| |
Notes on Current Work
The following information is somewhat dated but still relevant.
The HRC-S voltage was raised in March 2012 to restore QE and gain;
the new calibrations are preliminary and/or ongoing.
The gain of the HRC (both S and I) is very slowly decreasing.
Recently (2011), an above-trend decrease in the LETG/HRC-S
effective area was noted. It has been determined that this
effect (a few percent in the worst cases) is caused by events
falling below the HRC-S signal processing threshold in regions
of the detector where gain is especially low. We therefore plan
to raise the HRC-S voltage and restore the detector gain and QE,
probably sometime in early 2012.
Meanwhile, spectra from several soft sources are being studied to improve the
accuracy of the LETG/HRC-S effective area in the
difficult-to-calibrate intermediate wavelength range (roughly 60-100 Å).
Last modified:
10/12/12
