<Return to CalDB Release Notes>
CalDB PUBLIC Release Notes
Effective Date: (UTC) 2009-01-21T19:00:00
(Not installed in SDP.)
Public Release Date: 21 January 2009
Version: 4.1.1
Release Type: PUBLIC

I. Introduction

CaldB 4.1.1 is a patch to CalDB 4.1.0, which was released on 15 December 2008. It includes the following upgrades:
as well as the CalDB installation of the latest Cycle 11 Chandra PIMMS effective area files. These files are not effectual for users from the CalDB 4.1.1 upgrade; they may only be realized by downloading the CIAO 4.1.1 patch, available as of the same date as this CalDB release.

User Documentation:
"How CalDB 4.1.1 Affects Your Analysis"

II. Summary of Changes

A. HRMA AXEFFA (axial effective area) version N0008

LOCATION: $CALDB/data/chandra/default/axeffa
FILENAMES: hrmaD1996-12-20axeffaN0008.fits

The latest HRMA effective areas have been generated as a result of revisting the Chandra X-ray Calibration Facility (XRCF) ground calibration data and raytrace model, specifically because of known cross-calibration issues that have arisen within Chandra and between Chandra and XMM, when analyzing observations of high-temperature  galaxy clusters. The new model predicts plasma temperatures that are more in agreement with XMM for simultaneous cluster observations. In addition it brings analysis of ACIS-I and ACIS-S observations of similar clusters more into agreement as well. The new model takes into account the small surface contamination of the mirror shells at the XRCF (pre-launch), and also includes a new method of weighting the two XRCF calibration configurations (those with the flow proportional counter (FPC) in the focal plane, and those that employed the solid-state detector (SSD).)

Pipes/Tools affected:
CIAO "ARDLIB", specifically in the following parameters:
AXAF_EFFAREA_FILE_0001 = CALDB            Enter AXAF eff-area file 0001
AXAF_EFFAREA_FILE_0010 = CALDB            Enter AXAF eff-area file 0010
AXAF_EFFAREA_FILE_0100 = CALDB            Enter AXAF eff-area file 0100
AXAF_EFFAREA_FILE_1000 = CALDB            Enter AXAF eff-area file 1000
AXAF_EFFAREA_FILE_1111 = CALDB            Enter AXAF eff-area file 1111

ARDLIB is employed by the tools that generate ARFs, GARFs, weighted ARFs, and Exposure Maps.
Specifically mkexpmap, mkarf, mkgarf, and mkwarf.

Threads affected:
All threads which involve making exposure maps or effective areas.

Why Topic: "HRMA Effective Area version N0008"

B. HETG Grating Efficiencies (GREFF) N0006

LOCATION: $CALDB/data/chandra/default/greff
FILENAME: hetgD1996-11-01greffpr001N0006.fits

An adjustment of the relative calibration of the HEG and MEG 1st orders is required when a non-uniform with energy upgrade of the HRMA effective area occurs. The derivation of the changes is described in a report circulated in 2005 by the HETG gratings group, specifically by Dr. Herman Marshall. The document is available at http://space.mit.edu/ASC/calib/heg_meg/meg_heg_report.pdf , and is entitled "Improving the Relative Accuracy of the HETGS Effective Area".

Pipes/Tools affected:
CIAO 4.1 mkgarf, via the ARDLIB.

Threads affected:

C. CYCLE 11 PIMMS Effective Areas (final release)

ACIS Configurations:
LOCATION: $CALDB/data/chandra/pimms/acis/
FILENAME: acisiD2008-12-29pimmsN0011.fits

HRC Configurations:
LOCATION: $CALDB/data/chandra/pimms/hrc/

The new HRMA AXEFFA and HETG GREFF have been applied when deriving these effective areas, wherever relevant. The HRMA EA N0008 described in Section I.A above affects all configurations, producing a ~9% relative decrease in the configurations' effective areas over the 0.06-2.0 keV range, with the difference decreasing slowly to zero in the 5-6 keV range. It also produces some deltas from the previous PIMMS release in the higher energies, but at a much less significant level. The HETG GREFF affects only the first-order HEG, MEG, and HEG+MEG configurations, at the less-than-10% level at any energy, with the net change no greater than 18% when combined with the HRMA EA upgrade. The plots and input information for PIMMS CY11 may be viewed at the following sites:

"PIMMS CY11 Effective Areas"

"Cycle 11 PIMMS Effective Area Public Information"

See also the PIMMS effective area viewer.

NOTE: Users can only realize the PIMMS data upgrade by downloading and installing the CIAO 4.1.1 patch file. It does no good to install the CalDB upgrade discussed here, because the PIMMS software cannot read the CalDB. The installation of the new PIMMS data from CalDB is done at the CXC only, in the building process for the CIAO proposal planning tools. See the CIAO 4.1 web pages for details on the CIAO 4.1.1 patch, released 15 January 2008.

III. Technical Details


From the HRMA Calibration team, 18 December 2008:

Discrepancies in the temperatures of high temperature galaxy clusters derived from Chandra and XMM-Newton observationshave led to a reinvestigation of the amount of contamination on the Chandra mirrors. We have derived a more consistent model (orbit-200809-01f) which now includes contamination of the mirrors at XRCF with each shell having a different amount of contamination. A new method of weighting the two XRCF calibration configurations (those with the flow proportional counter (FPC) in the focal plane, and those that employed the solid-state detector (SSD)) has also been introduced.

A.1 Evidence for Miscalibration of the HRMA Effective Area

Comparisons between XMM-Newton and Chandra observations of clusters of galaxies as part of IACHEC 2007 indicated significant differences for high temperature clusters (see the results presented by Larry David at the 2007 CCW). While there was agreement between the derived temperatures for cool clusters (kT < 4 keV), the temperatures derived from the Chandra ACIS observations in the 2.0-7.0 keV band were systematically higher for hotter clusters when compared to those derived from

The most serious discrepency is that within the Chandra measurements (we should at least be in agreement with ourselves!). Larry showed that the ratio of the Fe XXV to FE XXVI lines is a sensitive diagnostic of temperature for hot clusters, and could be used as a good indicator of the intrinsic temperature of the cluster gas, assuming that the line and continuum emission arise from the same component.

A.2 Comparison of the new and current HRMA AXIAL effective areas
I compare the derivation of the N0007 (current) HRMA AXEFFA file to that of the N0008 (new) one below:

The basic derivation method is the same for both; the included components are different:

  1. The XRCF/HRMA configuration is raytraced at each energy and the effective area is calculated for each shell.
  2. XRCF correction functions are determined for each shell based upon the XRCF measurements.
  3. The on-orbit HRMA is raytraced at each energy, with each shell's throughput modified by the appropriate XRCF correction function. The Aeff of each shell is determined.
  4. The individual shell effective areas are summed to form the full HRMA Aeff.
The input components for each version are given in Table 1 below. Both the treatment and derivation of the included CH_2 (hydrocarbon, 1g/cm^2 density), and the weighting of ground calibration measurements at the XRCF, have changed from the previous derivation.

Table 1: The old versus new AXEFFA derivational inputs, indicating why the Shell AXEFFA curves have changed.
Input Element
Carbon layer contaminant applied to XRCF/HRMA raytrace model NONE Shell 1: 28Å
Shell 2: 18Å
Shell 3: 20Å
Shell 4: 27Å
XRCF Raytrace Model
Same model as N0007, but with the above contamination layers included.
XRCF Correction Algorithm
  • 0th order polyfit for FPC under 2keV
  • 4th order polyfit for SSD above 2keV and then
  • 50% weighted down by FPC above 2keV
  • Average the SSD data
  • Average the FPC data
  • Average the two averages.
XRCF Correction Function   (per shell)
Fitted functions given on the HRMA Twiki page at

Scalar values taken from the figures on the HRMA Twiki page at
Carbon layer contaminant applied to the on-orbit HRMA raytrace model
22 Å all mirrors Shell 1: 28Å
Shell 2: 18Å
Shell 3: 20Å
Shell 4: 27Å
On-orbit Raytrace Model

The variable contamination layers applied to each shell were derived from fitting XRCF SSD broadband traces across the Iridium M-edges with differing layer depths of CH_2 contaminant. During the process, it was discovered that an appropriate amount of contaminant could be applied to produce an essentially "grey" (energy independent ratio) scalar factor as a multiplicative correction between the
XRCF model and the SSD data.

Thus, the application of the contaminant layers to the XRCF raytrace model allows a simplification of the correction functions for the N0008 model, in which they are scalar factors applied for each shell. That is, the ad hoc polynomial correction algorythm used in the previous derivation is replaced by the effect of the contamination layers. Although eight different algorithms for generating the XRCF correction function were considered, the scalar factor method (derived from averaging all FPC effective area results and all SSD results separately, and then averaging these two averages) was selected because it provides both of the following:

A.3 The HRMA AXEFFA results

Comparison of the N0007 and N0008 derivations produces the following comparative plot and difference plot for the full axial effective area, old vs new
(all four mirrors summed):


Fig. 1: The N0008 (red) AXEFFA plotted against the current N0007 values in the CalDB. The most serious change relevant to users is the ~9% reduction in the 0 - 2 keV range; Source with spectra limited to this range will need some increase in the calculated observing time to give the anticipated statistics, when the new EA's are implemented in the proposal planning tools.

The most significant effect for Chandra data analysis is in the 0.6 - 2.0 keV range, where an essentially flat 9% reduction in the effective area for all configurations will result from using the new HRMA AXEFFA. (The decrement in the EA above 10 keV is less significant because of the strongly reduced total effective area in that range.


The update to the HRMA effective area in Section III.A. above requires an incumbent change in the HETG GREFF (grating efficiencies) that preserves the cross calibration of the HEG and MEG configurations. This is explained Herman Marshall's 14 October 2005 report "Improving the Relative Accuracy of the HETGS Effective Area".


From the above report:
"The High Energy Transmission Grating Spectrometer (HETGS) has two different grating types that disperse into two independent spectra (Canizares et al. 2005). The medium energy gratings (MEGs) have an energy range of about 0.4-7 keV, depending on the observation parameters, and the high energy gratings (HEGs) have an energy range of about 0.8 to 10 keV. Because they are built into the same structure, the MEG and HEG spectra are obtained simultaneously, facilitating cross calibration even for variable sources. This is an update of the HETG flight calibration paper (Marshall, Dewey, and Ishibashi 2004), which contained some results comparing ACIS-S quantum efficiencies (QEs)."

That report specifically refers to the development of the N0005 HETG GREFF upgrade from the previous version N0004. An exactly similar process has been employed to correct the HETG GREFF from the new HRMA N0008 release.

See sections 4.5 and 4.6.  In both cases (then and now), pre-flight grating efficiencies (from CalDB 2.0 GREFF file hetgD1996-11-01greffpr001N0004.fits) were used as a starting point and the effiency adjustments were precipitated by changes in the telescope effective area, which affects the ratio of the expected MEG and HEG count rates.  In both cases, data from blazars were accumulated in many wavelength bins from 1.7 A to 17 A in order to determine how the  MEG/HEG ratio should be corrected in order to match observations.  The ratios were computed from the data and efficiencies in such a way as to be independent of the sources' spectral properties.  In both cases, the observed errors in the MEG/HEG ratio were apportioned to the MEG and HEG efficiencies by assigning to the HEG predominantly below 1 keV and to the MEG predominantly above 1 keV.  Reasons for choosing this approach are given in the memo and have not changed.

"The differences between what was done then and what was done this time are few:

1- Five more blazar spectra were used in the analysis than used in 2005 and two were eliminated due to minor but potentially problematic spectral complexities.  In 2005, 18 observations of 5 blazars were used.  This time (January 2009), 21 observations of 3 blazars were used.

2- Details of the error in the MEG/HEG ratio differ slightly. Reference Fig. 12 in the 2005 memo referenced earlier.  In 2005, the maximum correction to the HEG efficiency was about 13%, near 17.5 A and for the MEG, the maximum correction was about 7%, near 10 A.  In the 2009 analysis, these maxima are about 21% at 17.5 A and 8% at 2.0 A, for the HEG and MEG, respectively.  Attached is the updated version of Fig. 12 from the 2005 memo."

Herman Marshall produced the following plot of the new correction factor as modeled by a polynomial (solid curve), as well as the apportioned corrections for the HEG +/- first order (dashes) and MEG +/- first order (dots), to be applied to the N0004 GREFF to produce the N0006 GREFF. See FIG 2 below.


FIG 2: Modeled correction factor (as a polynomial) for HEG vs MEG 1st orders (solid curve), and the apportioned correction factors for the MEG only (dotted curve) and the HEG only (dashes).

C. PIMMS CY11 Effective Area files (final release)

The technical details for the derivation of the PIMMS CY11 effective areas are given in the PIMMS CY11 Public Information page, and plots of the new areas in comparison with CY10 are given on the PIMMS CY11 Effective Areas page.

<Return to CalDB Release Notes>