Calibration of Chandra Grating Spectrometers
linked off of http://space.mit.edu/HETG/xrcf.html
-
- Spectrometer Design and Operation
- System Components
- Rowland Geometry
- XRCF Rehearsal Examples
- Calibration Products Definition
- Effective Area
- Line Response Function
- Science Drivers and Calibration Accuracy
- System Models and Grating Contribution
- Effective Area - Efficiency
- Line Response Function - Period and Alignment
-
-
-
- IR measurements
- X-Ray Tests
-
- Grating Efficiency Model
- Synchrotron Tests
- Subassembly Tests and Error Estimates
-
- Description
- Models and Uncertainties
- Calibration Considerations
-
- HETGS Predictions and Errors
- LETGS Predictions and Errors
-
-
- Off-Rowland use
- Vignetting
- HRMA PSF at XRCF
-
-
-
- Simulation of Measurements
- Grating Analysis at XRCF: ASC Pipelines
- Example Simulations for Some Measurements
-
-
Grating-related people planning to attend the Nov.13-15 meeting are:
HETG/ASC: Dan Dewey, Kathy Flanagan, Herman Marshall, Mike Juda, Jeremy Drake
LETG: Bert Brinkman, Peter Predehl, Theo Gunsing, Jelle Kaastra, Heinrich Brauminger
Other calibration review pages:
Return to HETG at XRCF page.
This is a DYNAMIC page so please direct comments and questions to
dd@space.mit.edu
Grating Spectrometer Operation
General Diagram
Rowland Geometry
Example Image from Rehearsal Data
Spectrum from above with linear scale
Calibration Product(s) for Chandra Grating Spectrometers
"... the creation of a model or set of models that describes the performance
characteristics of the instrument to within a certain accuracy..."
Effective area vs. E
The effective area calibration data must allow converting an
observed counts histogram (counts/sec bin) into an incident spectral
flux (e.g., photons/cm^2 sec). For the grating spectrometers this
generally refers to the combined (plus and minus) first-order
effective area.
Line-response function (LRF) vs E
With the presence of continuum and adjacent lines it is important to
have a known or constrained LRF to do accurate model fitting in regions
with many lines.
Response Matrix
The Effective area and LRF can be combined into a single
entity or response matrix.
Science drivers and required calibration accuracies
The absolute effective area of the grating spectrometers should be
known to +/- 10 % over the relevant energy ranges.
Plasma diagnostic methods, models, and their science lead to a desired
relative effective area calibration of +/- 3 % .
The instrumental shape of an infinitely narrow X-ray line, the
line response function LRF, should have its energy resolution,
e.g., FWHM, known to 1 part in 30 to support line-width measurements
and upper limits. The extended LRF, or wings, must be
known or constrained to within 1% of the peak of the response
in energy bins that are the instrumental width for all energies
(this includes higher orders implicitly).
To support doppler shift and line-energy determination, the
absolute energy scale of the grating spectrometers must be known
to the FWHM/30, e.g., 0.03 eV at 1 keV for HEG, or 0.002 A for the LETG at 60 A.
Definition of Effective Area and related terms for Grating Spectrometers
Effective Area (E,m,offaxis)
is ratio: detection rate (counts/sec)
------------------------------
flux at HRMA aperture (photons/cm^2sec)
where source is mono-chromatic (superposition assumed to work)
and "detection" includes PSF effects (cross-dispersion) and
detector effects (e.g., grading)
Effective efficiency (E,m,offaxis)
is the Effective Area of the HETGS/LETGS divided by the
corresponding Effective Area for the non-grating configuration.
Shell N Efficiency (E,m,offaxis)
is the ratio of photons diffracted to order m by grating facets
covering shell N to the photons leaving HRMA shell N.
Membrane Efficiency (E,m)
is the diffraction efficiency of the grating membrane, e.g. for
LEGs includes support structure vignetting, for MEGs/HEGs
includes platingbase/polyimide support. Does not include facet size
vignetting nor unfaceted regions between gratings.
Grating-bar Efficiency (E,m)
Diffraction efficiency for the gold grating bars alone.
LRF error budget and terms
The LRF is dominated primarily by the HRMA PSF (especially at XRCF)
and aspect system - except at highest dispersions. The gratings are expected
to introduce only a core blurring to the LRF based on subassembly diffraction
tests. XRCF will be an important place to look for other unexpected
components of the response, e.g. scattering.
Cross-dispersion Range for Calibration
The Effective Area / LRF calibrations apply to a cross-dispersion
width sufficiently large that variations in this width will not
produce large (few percent) changes in detected counts.
For the HEG and MEG a cross-dispersion width of 200um (d = 0.100 mm in
Dave H's plots below) encloses more than 99% of the flux.
For the LEG sprectra, 200 um is used for dipsersions below 60 mm, and
500 um for dispersions greater than 60 mm.
Further away from the dispersion axis are the support structure
orders. The first of these has ~1/100 of the main peak intensity,
and each subsequent is ~0.8 times the previous. The first two
"blips" from both sides would then have ~ 3.6 % of the main counts.
Grating Calibration Products
Component pieces:
HRMA area (E,shell,offaxis)
LETG/HETG Effic (E,m,shell)
Detector Effic (E,other)
Drake reflectivity(E,shell)
HRMA PSF (E,offaxis)
HEG,MEG,LEG LRF (E,m)
Drake PSF (E, shell)
Detector PSF
Aspect PSF Effects
Top level:
Response matrices for:
HETGS(HEG,ACIS-S)
HETGS(MEG,ACIS-S)
HETGS(HEG,HRC-I)
HETGS(MEG,HRC-I)
LETGS(HRC-S)
LETGS(Drake)
LETGS(HRC-I)
LETGS(ACIS-S)
LETG
Subassembly IR/Diffractometer data
Panter, Zeta, BESSY X-ray measurements
Likely errors from IR or X-ray tests
Range of LEG parameters
Subassembly prediction, modeling it
Significant edges/features
LETG Membrane efficiencies derrived from IR Data
Assumes a factor of 0.809 for the support-structure vignetting.
ASCII tables of LETG Effective Efficiency predictions
Effective efficiency of combined shells
(mirror weighted) for direct comparison with XRCF measurements.
For LETG using an extra vignetting factor of 0.75 in addition to the
"membrane efficiencies".
Plots of LETG Effective Efficiency predictions
Basis for model is scalar diffraction theory for trapazoidal
or many vertex modeled bar shape.
Synchrotron and X-ray facility testing.
HETG
HETG Sub-assembly tests
Synchrotron results, compare with model and X-GEF
Likely error on X-GEF measurements
Range of grating parameters for shells 1,3,4,6
Subassembly prediction, modeling it
Significant edges/features
Efficiency Model composed of many gratings
(and for HETG different mirror shells)
Can be fit with a several-grating model?
Adjusting model to match data
HETG Membrane Efficiencies derrived from X-GEF data
ASCII tables of HETG Effective Efficiency predictions
Effective efficiency of combined shells
(mirror weighted) for direct comparison with XRCF measurements.
For HETG using an extra vignetting factor of 0.93 in addition to the
"membrane efficiencies".
Plots of HETG Effective Efficiency predictions
- Jeremys Presentation (130Kb .ps)
- Model Details for Drake Flat(360kb .ps.gz)
- Physical models of key sub components
HRMA (P1H1 P3H3 P4H4 P6H6)
Reflectivity off of Ir coated surface, Ir optical constants
Uncertainty for all HRMA shells of order 10%, more at edges.
Created and tested by MST and TS.
Significant edges/features: Ir N and M edge regions, high energy fall off
Expect it will be well constrained after XRCF.
ACIS-S, ACIS-I
CCDs
UV/Optical-S, UV/Optical-I
Significant edges/features: O edge, Al edge, Si edge
HRC-S, UV/ion
HRC-I, UV/ion
MCPs
UV/ion
Significant edges/features:
For each physical model above summarize:
Basis for model
Model Creation and Testing
Significant edges/features
Uncertainties (pre XRCF)
Unconstrained (even) after XRCF
- Pre-XRCF Calibration Status
Plots of "products" and their errors if only subassembly data are used
HETG and LETG effective efficiency predictions
Resolving Power plots and data files based on "worst-case" error budget
Off-Rowland Operation
The gratings are built for operation at a Rowland diameter
(grating to focal plane distance) of 8635 mm. At XRCF they
are located at a distance of 8772.8 mm (TBR) from the focal plane
driven by the requirement that the gratings intercept the HRMA rays.
This Rowland diameter "error" produces insignificant (1 micron)
additional blur terms.
Effective Area: Vignetting
XRCF-Flight differences
As outlined in a detailed memo the
spacing of 8772.8 mm (TBR) is not optimum even from a vignetting point of
view and some additional LETG vignetting (about 5% for shell 1) will occur.
Somewhat different regions of the gratings will be illuminated at XRCF
as compared to flight. It is expected that the size of these effects and
the uniformity of the gratings are such that these diferences will not
compromise the XRCF to flight calibration transfer.
XRCF Point Spread Function
Because of a variety of effects the HRMA PSF at XRCF will differ
from the flight PSF. In addition, because the XRCF shutters allow
selection of mirror shells/quadrants it is useful to consider the
predicted performance of each shell/quadrant when planning XRCF tests.
The images and plots below provide just such detailed looks for the Al
1.487 keV line.
Images and 1D projection plots for each shell/quadrant. The
projection along the dispersion axis (horizontal, XRCF-Y axis) is
shown solid and the cross-dispersion axis (vertical, XRCF-Z)
projection is dashed.
Summary of identified calibration priorities
Alignment
Measure grating dispersion directions and (roll) angles in facility
Measure XRCF vigneting by shell
Measure axial repeatability of grating insertion mechanism at XRCF
Focus
Use a "Shutter Focus" iterative test to focus detector and define X=0
Make "Focus Check" measurements - images in each quadrant - to
determine focus of dispersed orders.
LRF Inner Core
Verify grating dp/p values
Measure LRF near core for a few energies
Measure cross-dispersion response
LRF Outer Core
Search LRF for unexpected features (in wings)
LRF Modeling
Verify defocus and offaxis effects
Effective area related:
Efficiency/Effective area vs E to high accuracy for 1st order
Measurement of LETG high-order efficiency at several energies
Scans over known edges/features
Search for unexpected edges/features
Verify/measure detector effects on effective area
(e.g., grading, mode, gain, rate)
Wavelength Scale
Set energy scale relative to edges/features
Generally parasitic on other measurements
Molecular Contamination
Use "featureless" continuum to search for unexpected edges and
features in the response
Measurement priorities for Grating Calibration Measurements
Test list - CMDB
[Test lists in /proj/asc/cal/Calibration/{LETG|HETG}/{phase1|phase2}]
Subphases for Phase 1:
A - find any major surprises in HRMA/Gratings
B - minimum ground calibration of HRMA/Gratings
C - proper calibration of HRMA/Gratings, model over determination, etc.
.exp file written to clearly indicate purpose of measurements
Phase 1 HXDS
CMDB Summary Table Phase 1A
CMDB Summary Table Phase 1X
Phase 2 HRC-I, HRC-S, ACIS-2C
Summary Material
Model constraints that these measurements will provide
Assummed calibrations (XSS,HXDS,HRMA,FPSIs,...)
- have HSI map identifying bad areas, gain variations
- HSI effic at 130A calibrated post test
- SSD-500 and BND-500 cross-calibrated
- FPC efficiency @ 108 eV > 1%
- Effic(HSI_for_real) > Effic(HSI_rehearsal)/20
- comparable no-grating measurements are made
Simulation of XRCF Measurements
Both SAOSAC and MARX can be used to simulate XRCF measurements.
The SDS/ASC-MIT developed MARX ("csim") is being used with an IDL
XRCF spectrum generator, is having its parameter files
generated from a cmdb.trw database file, and is being modified to include
the HXDS instruments. Details of csim/MARX were recently presented at
the '96 ADASS meeting, manuscript in preparation.
Pipeline Processing for XRCF Grating Tests
PURPOSE: The purpose of pipeline processing is to automatically process as much
data as possible with minimal human interaction in order to prepare
data products for inspection, modeling, and analysis by calibration
scientists. Pipelines have been defined for each type of test, with
the detector/grating combination primarily specifying variant
parameters and slightly alternative processing (e.g., if HSI or HRC,
then skip PHA-based order sorting). Processing specifications were
derived from input from Calibration scientists.
CONTENT: For Phase 1, there is strong reliance on MST processing of the
focal-plane detector data. Those results are then subjected to
grating-specific processing (for example, scan-positions are converted
to diffraction coordinates before profiles are extracted and
measured).
The tables below diagram the types of tests and the detectors used,
based on a recent Calibration Measurement Data Base (CMDB). Numeric
designations correspond to individual pipelines. Sub-designations
(1a, 1b, etc) mean that the processing is identical at Level 2 (ASC
level 2 - individual test measurements). (priorities are not
indicated)
After Level 2 pipelines, data for each test and type are collected
into tables for detailed analysis (e.g., collect all effective areas
vs energy).
HETG XRCF tests:
| HXDA | SI |
Test | HSI SSD FPC | 2C AS HS AI HI |
______________|________________|_____________________|
1a Shutter focus | x | x x x x x |
1b Focus check | x | x x x x x |
2 Alignment | x | |
3 Mol. Contam | x | x x x |
4a PSF/1D | x x | |
4b PSF/2D | | |
4c PSF/Inner | | x x x x x |
4d PSF/Outer | | |
4e dFocus | | |
4f Dispersion | x | x x x |
5a EffArea/EE | x x x | |
5b EffArea/QE | | x x x x x |
6 Scattering | x | x |
LETG XRCF tests:
Test | HSI SSD FPC | 2C AS HS AI HI |
______________|________________|______________________|
1a Shutter Focus | x | x x x x |
1b Focus check | x | x x x x x |
2 Alignment | x | |
3 Mol. Contam | x | |
4a PSF/1D | x | |
4b PSF/2D | x | |
4c PSF/Inner | x x | x x x x |
4d PSF/Outer | | x x |
4e dFocus | x | |
4f Dispersion | x | x x x |
5a EffArea/EE | x | |
5b EffArea/QE | x | x x x x |
6 Scattering | | |
HSI = High Speed Imager SSD = Solid State Detector
FPC = Flow Proportional Counter
2C = ACIS-2C AS = ACIS-S AI = ACIS-I
HS = HRC-S HI = HRC-I
RELEASED GRATING PIPES:
Released pipelines can be found on the
ASCDS web pages.
They can be executed under the ASSIST environment, or via the UNIX
command-line. The current XRCF release of grating pipes includes:
Level-1 (formatting, "linearization") / Rehearsal:
ACIS/HRC event processing
ACIS/HRC SAO-simulation (SAOSAC) connector
ACIS/HRC MIT-simulation (csim/MARX) connector
ACIS-2C quicklook PSF
ACIS-2C quicklook spectra
HSI quicklook PSF
Level-1.5 (diffraction coordinates)
ACIS
ACIS-2C
HSI
Level-2 (test-specific measurements):
ACIS/HSI Molecular contamination
ACIS/HSI PSF (inner, outer, defocus, dispersion)
ACIS/HSI Shutter focus, focus check
PIPELINES under development:
HSI Alignment: 50% done. Borrows from HRC PSF
(radial/azimuthal profiles). Completion requires profile
ratios and statistical test of constancy with azimuth.
SSD/FPC effective area and PSF: These borrow heavily from MST
processing for L1 products, after which they in turn
borrow from existing grating pipes - e.g., for L1.5
diffraction coordinate transformation and then profile
extraction and measurement.
HSI effective area: borrows heavily from ACIS and HRC
effective area (which are released pipelines).
ACIS/HRC effective area: will be similar to HSI.
HRC L1.5: release depends upon coordinate transformation
library updates which include HRC-S and grating routines.
Otherwise, this is very much like ACIS-S, with the
exception of PHA-based order sorting.
HRC PSF: follows easily after L1.5 - similar to ACIS, HSI.
POST-PIPELINE PROCESSING:
After pipeline processing, a large variety of software is available
on the 3rd floor for detailed, interactive analysis. Many
ASC-developed tools have been released in the data-system. IDL is
available, as well as IRAF/PROS, XSPEC, and FTOOLS.
ASC-DS released tools can be found on the ASCDS web pages.
(http://???). They include:
- image manipulation
- data conversion
- event manipulation and filtering
- table manipulation
- source detection
- image processing (radial profiles)
- curve fitting
- graphics display tools
Some examples of post-pipeline processing are:
- Fitting effective area vs energy
- determination of dispersion constants and clocking angles from
PSF-pipe output line-centroids vs energy.
- spectral resolution vs energy.
- modeling shutter focus results to determine whether centroid
variance is due to astigmatic focus, or exceeds it due to
defocus.
XRCF Data Analysis plan (Tasks, who, dates: prelim, final)
o
MSFC-RQMT-2229 "Scientific Requirements for the Calibration of AXAF"
o HETG Homepage
o LETG Homepage
o HETG SOP01 Operations Handbooks
o LETG SOP01 Operations Handbooks
o ASC SIN
o Various SPIE papers
