3 MIR flux measurements and uncertainties

   
3.1 PHT-S flux measurements and reproducibility

For each epoch of observation the continuum flux was measured over two different intervals, at short wavelengths (SW: 2.5-4.7$\mu$m) and long wavelengths (LW: 5.8-9.9$\mu$m). Table 1 lists the mean intensities over these intervals and their uncertainties, while the light curves are shown in Fig. 2.

Figure 2: IR and optical light curves of Mrk 279. The top two panels show the IR light curves from the SW and LW detectors (2.52-4.70 $\mu$m and 5.8-9.9$\mu$m, respectively). The third panel shows V-band photometric measurements from Wise Observatory. The forth panel shows the 5100Å continuum flux measured from the spectra, in units of 10$^{-15}\,$ergs s-1cm-2Å-1. The H$\beta$ flux is shown in the bottom panel, in units of 10$^{-13}\,$ergs s-1cm-2. The optical photometry, optical continuum and H$\beta$ fluxes are available in electronic form at the CDS

An accurate determination of the flux uncertainties is essential when discussing source variability. We have therefore investigated the different source of errors which could potentially affect our PHT-S measurement.

 

In staring mode, the overall responsivity of PHT-S is known to remain stable within $\pm\,10$% (Schulz 1999). We have assessed the stability of the PHT-S responsivity more specifically at the time of each of the Mrk 279 observations and verified that no other systematic effects were present. For this purpose, two different types of calibration measurements were used as diagnostic:

1.

The detector dark current measurements, which are obtained immediately prior to each PHT-S observation: throughout the campaign, the dark current signal retained its nominal value of $\sim$zero V/s. We can therefore be confident that none of the Mrk 279 observations suffered from detector remanence induced by a prior exposure to a bright source;

2.

Measurements with an internal calibration source which are carried out systematically at the beginning of each ISO revolution: averaged over all pixels, the variations of detector responsivity from epoch to epoch are $\simeq\,\pm\,2$%, with upper limits of 3% and 5% in the SW and LW range, respectively.

A conservative upper limit of $\pm 10$% was thus adopted for the systematic uncertainty on the PHT-S fluxes of Mrk 279. Internal measurement errors were added in quadrature to this systematic uncertainty. The internal errors were computed as the dispersion about the mean flux in the SW and LW integration intervals, after normalization of the spectra. The purpose of the normalization is to remove the spectral curvature. Each spectrum was first divided by its best-fit power-law continuum ( $F_\nu = 8.44\, 10^{-14}~ \nu ^{-0.8}$ ergs^-1 cm^-2 Hz^-1; see Sect. 5) and the rms deviations were computed. The results show that the internal errors associated with each flux measurement are of 2.8% and 1.7% for the 2.52-4.70$\mu$m and 5.76-9.89$\mu$m bands, respectively.

As a consistency check, errors were also computed by comparing PHT-S fluxes obtained within 30 days from each others. This gives a conservative error estimate since it assumes that there are no flux variations on time scales shorter than 30 days. Taking every pair of fluxes within 30 days and measuring the error on their means, we get mean relative errors of 3.5% and 1.6% for the SW and LW bands, respectively.

   

Table 1: PHT-S observation log and fluxes

UT	MJD	F(2.5-4.7$\mu$)	F(5.8-9.9$\mu$)
	(-2450000)	(mJy)	(mJy)
(1)	(2)	(3)	(4)
1996 Feb. 5	119	73.5 $\pm$ 7.6	132 $\pm$ 13
1996 Mar. 3	146	75.3 $\pm$ 7.8	123 $\pm$ 12
1996 Mar. 12	155	65.2 $\pm$ 6.8	125 $\pm$ 13
1996 Apr. 2	176	69.5 $\pm$ 7.2	129 $\pm$ 13
1996 Apr. 27	201	72.7 $\pm$ 7.6	127 $\pm$ 13
1996 May 11	215	72.8 $\pm$ 7.6	128 $\pm$ 13
1996 May 29	233	73.2 $\pm$ 7.6	131 $\pm$ 13
1996 Jul. 29	294	63.6 $\pm$ 6.6	127 $\pm$ 13
1996 Aug. 12	308	61.8 $\pm$ 6.4	122 $\pm$ 12
1996 Aug. 27	323	62.0 $\pm$ 6.4	123 $\pm$ 12
1996 Sep. 15	342	60.0 $\pm$ 6.2	124 $\pm$ 13
1996 Oct. 17	374	73.8 $\pm$ 7.7	128 $\pm$ 13
1996 Nov. 1	389	62.6 $\pm$ 6.5	120 $\pm$ 12
1996 Nov. 18	406	73.7 $\pm$ 7.7	128 $\pm$ 13
1996 Dec. 5	423	70.1 $\pm$ 7.3	124 $\pm$ 13
1997 Feb. 13	493	81.8 $\pm$ 8.5	128 $\pm$ 13

   

Table 2: ISOCAM narrow band filter intensities

Filter	$\lambda_{\rm c}$	Range	Flux	FWHM
	($\mu$m)	($\mu$m)	(mJy)	('')
(1)	(2)	(3)	(4)	(5)
SW1	3.57	3.05-4.10	68	3.9
SW5	4.25	3.00-5.5	60	3.3
LW4	6.00	5.50-6.50	106	4.1
LW2	6.75	5.00-8.50	115	4.8
LW5	6.75	6.50-7.00	106	4.2
LW6	7.75	7.00-8.50	120	5.0
LW7	9.62	8.50-10.7	159	3.8
LW8	11.4	10.7-12.0	212	4.5
LW3	15.0	12.0-18	209	5.0

   
3.2 Comparison with ground based measurements and estimation of the host galaxy contribution

Spinoglio et al. (1985) measured L-band ($\sim$3.5 $\,\mbox{$\mu$ m}$) fluxes of $100 \pm 21$mJy, $112 \pm 27$mJy, and $68 \pm 15$mJy through apertures of 12'', 12'', and 17'', respectively, consistent with our results to within the measurement uncertainties.

Given the spectrograph aperture ( $24''\times 24''$), the host galaxy of Mrk 279 could, in principle, contribute to the PHT-S flux. Indeed, a faint extended nebulosity is apparent in the K-band ($\sim$2.2 $\,\mbox{$\mu$ m}$) image of McLeod & Rieke (1995). This extended flux arises from the integrated emission of giants and supergiants in the galactic disk whose energy distribution is maximum at $\sim$2 $\,\mbox{$\mu$ m}$ and falls-off abruptly at longer wavelengths. In practice, stellar emission will therefore make a negligible contribution to the PHT-S flux. Nevertheless, this was positively verified by comparison with ground-based data as follows:

1.

In the K-band, McLeod & Rieke (1995) estimated that the AGN contributes 90% of the flux within the central $1\hbox{$.\!\!^{\prime\prime}$ }5$ (FWHM) and 55% (35 mJy) of the total K-band flux (68 mJy) integrated over the whole galaxy (i.e., out to a radius of 35''). Assuming ``normal'' near-IR colors ( $K-L=0.22\pm0.02$mag; Clavel et al. 1989) for the stellar population, we derive a total host-galaxy flux of 18mJy at $3.5\,\mbox{$\mu$ m}$;

2.

Using the B-and R-band nucleus-galaxy decomposition of Granato et al. (1993), and assuming normal optical-to-IR (V-K=3.22mag; Clavel et al. 1989) and K-L colors, we estimate a total galaxy flux of 20mJy and 17mJy respectively at $3.5\,\mbox{$\mu$ m}$.

These consistent estimates can be used to infer the amount of stellar light which enters the $24''\times 24''$ PHT-S aperture, $5\pm2$ mJy at $3.5\,\mbox{$\mu$ m}$. Such a small contamination is within the measurement errors and can be neglected.

The galaxy contributes $11\pm 4$mJy, $14\pm 5$mJy and $11\pm 4$mJy to the $12\,\hbox{$^{\prime\prime}$ }$photometric measurements in the J, H, and K bands respectively (Granato et al. 1993). These values are used in Sect. 6 to infer the intrinsic spectral energy distribution of the active nucleus in Mrk 279.