Fermi Gamma-ray Space Telescope

LAT Instrument Response Functions—Overview

The LAT performance is governed primarily by three things:

  • LAT hardware design
  • Event reconstruction algorithms
  • Background selections and event quality selections

A result of the performance analysis is the production of full Instrument Response Functions (IRFs), describing the performance as a function of photon energy, incidence angle, conversion point within the instrument, and other important parameters.

The IRF is the mapping between the incoming photon flux and the detected events. 'Detection' depends not only on the LAT hardware but also on the processing that calculates the event parameters from the observables and assigns probabilities that an event is a photon. Indeed, the different event cuts are based on tradeoffs between the non-photon background, the effective area and the spatial and energy resolution; these cuts result in analysis classes (see the section on LAT Data Products).

To evaluate the LAT response, a dedicated Monte Carlo simulation is performed. A large number of gamma-ray events are simulated in order to cover all possible photon inclination angles and energies with good statistics. This is based on the best available representation of the physics interactions, the instrument, and the on-board and ground processing to produce event classes (see Atwood et al. 2009 and Ackerman et al. 2012). The comparison between the properties of the simulated events within a given event class and the input photons gives the instrument response functions.

The LAT IRF parameters

The IRF is factored into three terms: efficiency in terms of the detector's effective area, resolution as given by the point-spread function (PSF), and energy dispersion. The components of the IRFs are the measured representation of the corresponding figures of merit in terms of the photon true energy and incidence angle. The values associated with each parameterization are defined within the Fermitools calibration database (CALDB) which is included in the Fermitools distribution. The forms for these functions are presented in the subsequent sections.

Each event class and event type selection has its own IRFs which must be used when performing analysis of events passing that selection. Internally the IRFs for each event class contain a set of IRF parameterizations for individual event types. The event types are organized in partitions that split the events within a class according to different criteria. Whereas previous data releases provided only a single partition based on the conversion plane of the reconstructed track (FRONT versus BACK), the Pass 8 data release defines three event type partitions: FRONT/BACK (two types), PSF (four types), and EDISP (four types). The selections that define the event classes and event types are described in more detail in the LAT Photon Classification section.

The Fermitools support analysis with both individual event types as well as any superset of event types within a partition (e.g. PSF2+PSF3 or FRONT+BACK). In the case that the data selection is a superset of event types within a partition, the Fermitools will apply the response function appropriate for the sum of those event types. In the typical usage scenario in which all events within a class are selected (equivalent to evtype=3) the Fermitools will apply the response function for the sum of FRONT and BACK events.

The Pass 8 release introduces a new internal format for the IRF files whereby parameter tables for event types within each partition are stored in a single aggregated IRF file (designated FB, PSF, or EDISP in the file suffix). The event type of a parameter table within the IRF file is designated by the suffix in the key of the header/data unit. For instance the FRONT and BACK effective area tables for P8R3_SOURCE_V3 are contained in the "EFFECTIVE AREA_FRONT" and "EFFECTIVE AREA_BACK" HDUs of the IRF file "aeff_P8R3_SOURCE_V3_FB.fits".

IRF selection

There are multiple IRFs delivered with the Fermi Fermitools to allow the user the flexibility necessary for the different analysis types. The LAT data currently being released by the FSSC have been processed using the "Pass 8" event-level analysis. The Pass 8 analysis uses an entirely new set of event-level reconstruction algorithms that improve the instrument performance and mitigate pile-up effects. The Pass 8 data are processed with the same calibration constants used for the P7REP data release. The current Pass 8 data set is known as "P8R3".

The P8R3_V3 IRFs provide the current description of the instrument response for the P8R3 data release. Systematic uncertainties associated with these IRFs are documented in the LAT Caveats page.

The main features of the P8R3_V3 IRFs are:

  • Inclusion of second order effects released previously: azimuthal- and livetime-dependence of effective area are also included in "P7REP".
  • Subdivision of IRFs into three event type partitions (FRONT/BACK,PSF, and EDISP). Previous IRF releases were partitioned only by conversion type into FRONT and BACK events. Note that the IRFs themselves are parameterized only for individual event types and the IRFs for any superset of event types (e.g. all events in a given class) are constructed by summing the event type IRFs.
  • Compared to the previous IRFs (P8R3_V2), we used an in-flight calibration method to change the PSF event type effective areas (as a function of energy and incidence angle) in order to improve the agreement between the observed and predicted numbers of events. The calibration was performed using the Vela pulsar and the Earth limb data. The effective area change is within 20% and the resulting data/model agreement is within 5%. No in-flight calibration has been performed for the other parts of the IRFs (FRONT/BACK or EDISP event type effective areas, PSF, energy dispersion). Systematic uncertainties in the effective area and PSF are less than 5% between 100 MeV and 10 GeV (see the LAT Caveats for more details).
  • New tables that contain a correction for the bias in the LAT event direction reconstruction (fisheye effect) as a function of energy and incidence angle. The fisheye effect is a bias in the reconstructed event direction toward the LAT boresight. This effect is largest at low energies and high incidence angles. Although the Fermi Fermitools do not currently provide a mechanism for applying this correction, these tables can be used to quantify systematic uncertainties on source localization that may be induced by this effect.

Recommendations for the appropriate selection and IRF set to be used within an analysis are provided in the Data Preparation section of the Cicerone.

The tables below give the association between the P8R3 IRF sets and the photon properties as provided in the LAT photon data section. The columns "Photon File" and "Extended File" indicate whether a given class is available in the corresponding data file type.

P8R3 IRF name Event Class (evclass) Class Hierarchy Photon File Extended File
P8R3_SOURCEVETO_V3 2048 Standard X X
P8R3_ULTRACLEANVETO_V3 1024 Standard X X
P8R3_ULTRACLEAN_V3 512 Standard X X
P8R3_CLEAN_V3 256 Standard X X
P8R3_SOURCE_V3 128 Standard X X
P8R3_TRANSIENT010_V3 64 Standard X
P8R3_TRANSIENT020_V3 16 Standard X
P8R3_TRANSIENT010E_V3 32 Extended X
P8R3_TRANSIENT020E_V3 8 Extended X
P8R3_TRANSIENT015S_V3 65536 No-ACD X

The following table lists the event types defined in the P8R3_V3 IRFs. The event types are organized in three partitions: FRONT/BACK, PSF, and EDISP. IRFs are provided for every permutation of event class and type.

P8R3 Event Type Name Event Type Partition Event Type Value (evtype)
FRONT Conversion Type 1
BACK Conversion Type 2
PSF0 PSF 4
PSF1 PSF 8
PSF2 PSF 16
PSF3 PSF 32
EDISP0 EDISP 64
EDISP1 EDISP 128
EDISP2 EDISP 256
EDISP3 EDISP 512

» Forward to Point Spread Function
» Back to the beginning of the IRFs
» Back to the beginning of the Cicerone