FLIX - Sensitivity estimator

Purpose

The main purpose of FLIX is to produce estimates of the detection threshold of the X-ray flux at a given point in the sky, for points where no source was detected in the 4XMM survey. This detection threshold is found empirically using the algorithm described by Carrera et al (2007) The XMM-Newton Serendipitous Survey, IV. The AXIS X-ray source counts and angular clustering. A copy of the preprint may be found in astro-ph at https://arxiv.org/abs/astro-ph/0703451 - the algorithm appears in Appendix A. These detection thresholds use the DETMSK to find which pixels in the image are valid, EXPMAP data to find the exposure time, and BKGMAP data to determine the fitted background level.

In addition FLIX provides other information:

It computes a crude estimate of actual flux near the specified position by summing counts in a circle centred there, using the background level from the BKGMAP data. The equivalent flux is converted using an approximation to the point-spread function in the calibration files. Energy conversion factors can be found in Mateos et al. 2009 (astro-ph:0901.4026).
It scans the 4XMM catalogue and lists the 6 nearest sources together with their offsets.
It provides a thumbnail image of the field using Aladin Lite with the nearest XMM catalogue source overlayed. Further informations on these sources can be obtained by clicking on the SRCID numbers which link to the source summary pages in the IRAP catalogue server.

Results are returned as a web page with an HTML table per celestial position, and a FITS file is also produced with a binary table containing one row per observation which contains useful information. This may be downloaded and examined with utilities such as FV or TOPCAT.

Input Parameters

The following parameters may be entered on the web form:

Position of interest: for a single position enter it in the text box, for multiple positions use any text editor to create a source file and upload it using the flix upload button. In either case the coordinates format is the same, and explained below.
Detection likelihood threshold for detection threshold estimation: enter it using the "Settings" link next to the input box. The default value is 10.0 which is the value used in the 4XMM catalogue, but you may choose a higher value.
Radius of the circle, called radpix below, used for actual flux estimation (arc-seconds): enter it using the "Settings" link next to the text box. The default value is 30 arcseconds (7.5 pixels) but other values may be chosen within reason.

Note that the FITS file that is produced contains a complete set of columns including encircled flux in all XMM energy bands

Likelihood Thresholds

For those who prefer to think in terms of N-sigma (assuming a 1-dimensional Gaussian distribution) then the table below shows approximate conversions from N-sigma to the likelihood values, which are actually -log_e(likelihood).

N-sigma	-log_e(likelihood)
3	6.6
4	10.4
5	15.1
6	20.7
7	27.4
8	35.0

Note that the likelihood threshold for the 4XMM surveys (and the default value for FLIX) is 10, corresponding to around 4 sigma.

Input file format

The input file is a plain text file with one position per line. This should contain an RA, a DEC and optionally a source name or position identifier. The RA and DEC may be separated by spaces or a comma. Thus all of the following formats are allowed:

8.51332233 -5.40595761 object1
09:18:48.3 -12:21:11.4984 object2
8.51332233,-5.40595761, object3
09:18:48.3,-12:21:11.4984, object4

Results

Note that for each position the program has to read around compressed 40 data product files before performing the flux calculations, which takes from 5 to 10 seconds per position typically.

For single coordinates requests, if the position in the sky was not covered by XMM, a "no match" page is returned with the coordinates of the nearest XMM pointing. Note that for some positions there may be multiple observations covering the field, up to 50 in a few cases. The processing time in such case can take a bit more than a minute. Please be patient. The FITS file contains one row per observation matching each position, so it may have more rows than there are positions in the input text file.

The FITS file will be retained on the server for at least 24 hours before being deleted.

The columns in the FITS table are described below; most of these columns are also in the HTML table.

RA	Right ascension of point of interest (decimal degrees)
DEC	Declination of point of interest (decimal degrees)
NAME	Name from input list, if any
OBS_ID	Observation identifier
REVOLUT	Revolution i.e. orbit number
RA_PNT	Pointing axis R.A. (decimal degrees)
DEC_PNT	Pointing axis declination (decimal degrees)
OFFAXIS	Approximate distance of point of interest from pointing axis (arcmins)
OBJECT	Target name of the original observation
OBSERVER	Name of principal investigator of original observation
DATE_OBS	Date and time at which observation was started
DATE_END	Date and time at which observation ended
inst_ACTIVE	Duration for which instrument was active (seconds)
inst_FILTER	Filter name for each instrument
inst_SUBMODE	Data submode - see XMM handbook for more details
inst_MASKFRAC	Fraction of pixels providing data around this position

inst_band_BGMEAN	Background counts from BKGMAP summed over the circle of radius cutrad. This is adjusted to allow for any missing pixels in the DETMSK array.
inst_band_TEXPOS	Exposure time (seconds) at the nearest pixel to the position, derived from EXPMAP product.
inst_band_ECF	Energy Conversion Factor (ECF) used to derive flux from count-rate (count sq cm/erg)
inst_band_TPSF	PSF-weighted mean exposure time (seconds)
inst_band_FUPL	Flux detection threshold for each instrument and band (erg/sq cm/sec)
inst_band_IMCIRC	Sum of the image pixels over a circle of radius radpix, not adjusted for missing pixels
inst_band_BGCIRC	Background summed over a circle of radius radpix, not adjusted for missing pixels.
inst_band_EEF	Enclosed Energy Fraction (EEF) used to derive flux from count-rate
inst_band_FEST	Flux estimate for each instrument and band using counts in circle and fitted background level (erg/sq cm/sec)
inst_band_FER	Standard error on above flux estimate (erg/sq cm/sec)

Notes:

inst: a two-character abbreviation of the instrument, M1 = MOS-1, M2 = MOS2, PN = PN.
band: a single digit band designator:

bands 1 to 5 are the standard EPIC energy bands,
band 8 is the overall band corresponding to bands 1-5.
band 6 is approx bands 1+2+3,
band 7 is approx bands 4+5,
band 9 is bands 3+4+5.

RADPIX: circle radius chosen by the user, the default is 30 arcseconds for flux measurements.
CUTRAD: radius in pixels (4 arcsecs/pixel) which optimises the flux-limit estimated for each energy band

Band	1	2	3	4	5	6	7	8	9
cutrad	5.12	5.08	5.15	5.49	5.86	5.1	5.7	5.2	5.24

In the result page the SRCID and distance for the six nearest sources are shown, but only the nearest is listed in the FITS file to avoid format complications.

Flux estimate and error

Flux estimate, FEST = (IMCIRC - BGCIRC) / (TEXPOS * ECF * EEF)

Error on flux estimate, FERR = sqrt(IMCIRC + BGCIRC) / (TEXPOS * ECF * EEF)

Note that the IMCIRC and BGCIRC values result from summing flux and background only for valid pixels, i.e. pixels in chip gaps or other places where DETMSK is zero are simply ignored. This is because there really isn’t any satisfactory way of estimating the effect of these gaps, and it could make a large difference depending on whether the source happens to be centred in a chip-gap or not.