3 Characteristics of DENIS images and conversion factors

The first characteristics generally measured on astronomical images are the "sky emission'' level (which includes telescope and instrument background in the $K_{\rm s}$ band), and the noise on sky images and darks.

Variations in sky level are observed in all three bands. In i, they are related to the distance to, and phase of, the Moon. In J, they are due to variations in the hydroxyl radical's emission intensity (OH^- Meinel bands), connected to the passage of density and temperature perturbations through the upper atmosphere (Ramsay et al. [1992]). In $K_{\rm s}$ they come from the temperature variations. Also note that, in crowded fields, the background value is set by the confusion level (background of faint undistinguishable stars).

From these variations, the conversion factor between ADU (analog-to-digital units) and electrons can be deduced. Table 1 compares the result of these gain measurements with the expected values, calculated from the electronics characteristics of the chips, the preamplifiers gains, and the analog-to-digital conversions. Table 2 gives for each band the median value of the read-out noise in electrons, the median and faintest values of the sky level, in mag arcsec^-2, and the median and minimal values of the sky image noise in electrons, calculated from more than 2000 images taken during the last year of observations (April 1998 to April 1999), and adopting zero-points of 23.5, 21.3, and 19.2 in i, J, and $K_{\rm s}$, respectively (see Sect. 4 and Table 5). The large values of sky and sky noise in $K_{\rm s}$ come from the thermal background of the instrument, which does not include cool stop optics. The "best'' value of the sky level in J is suspect.

Note that some care must be taken in applying the conversion factors and zero-points to the whole DENIS survey: first, J and $K_{\rm s}$ conversion factors seem to be slightly variable (11.4 to 15.3 in J, 51.6 to 57.0 in $K_{\rm s}$), and second, changes in the instrument have altered these values: the pre-amplifier boards of J and $K_{\rm s}$ cameras were changed in June 1996, and a spare J camera has been in use from April 3, 1998 to May 9, 1999. Old conversion factors valid until June 1996 were 12.47 in J and 39.3 in $K_{\rm s}$ (Chevallier [1996]). Zero-point variations will be analyzed in a future paper.

 

 

Table 1: Conversion factors (expected and measured, in $\rm e^-$/ADU) in the three bands

Gain	i	J	$K_{\rm s}$
Expected	3.08	18.4	58.6
Measured	3.04	11.4	56.2
Std dev	0.01	0.1	0.1

 

 

Table 2: Read-out noise, sky level, and sky image noise in the three bands

Band	RON	SKY		SKY noise
		median	best	median	min
	$\rm e^-$	mag arcsec-2	mag arcsec-2	$\rm e^-$	$\rm e^-$
i	7.3	19.2	20.2	15	11
J	21	16.6	18.1?	37	27
$K_{\rm s}$	40	11.2	12.0	311	222

 

 

Table 3: Mean filter wavelengths (pure filter, and filter + atmosphere + detector, hereafter "fad''), effective and isophotal wavelengths for Vega

Wavelength ($\mu$m)	i	J	$K_{\rm s}$
$\lambda_{\rm fil}$	0.802	1.248	2.152
$\lambda_{\rm fad}$	0.795	1.235	2.160
$\lambda_{\rm eff}$	0.788	1.221	2.144
$\lambda_{\rm iso}$	0.791	1.228	2.145

 

 

Table 4: Flux densities of a zero magnitude star for the three DENIS bands

Band	$\lambda_{\rm iso}$	$F_{\lambda}$	$F_{\nu}$
	$\mu$m	W/m2/$\mu$m	Jy
i	0.791	$1.20 \, 10^{-8}$	2499
J	1.228	$3.17 \, 10^{-9}$	1595
$K_{\rm s}$	2.145	$4.34 \, 10^{-10}$	665