ISO observations of the dusty quasar BR 1202-0725

Abstract

We present mid- and far-infrared photometry of the high-redshift (z = 4.69) dusty quasar BR 1202−0725. The quasar was detected in the near-infrared, at a flux level (0.7 ± 0.2 mJy) consistent with an average radio-quiet quasar at its redshift. Only upper limits for the emission were obtained in the far-infrared. These upper limits, when combined with data from ground-based telescopes, are the first direct evidence for a turnover in the far-infrared emission, and hence confirm that a blackbody dominates the spectral energy distribution at far-infrared wavelengths. This blackbody is most probably cool dust, constrained to have a temperature below 80 K, for a β of 1.5.

galaxies: active, galaxies: photometry, quasars: individual: BR 1202−0725, infrared: galaxies

1 Introduction

It came as a surprise when groups working with IRAS data (e.g. Soifer et al. 1984; Lawrence et al. 1986) discovered galaxies with bolometric luminosities equivalent to those of quasars but emitting predominantly in the infrared wavelength region. Even more surprising was the discovery by Rowan-Robinson et al. (1991) of IRAS FSC 10214+4724, a galaxy at a redshift of with an apparent infrared luminosity of . While F10214+4724 has turned out to be gravitationally lensed (Eisenhardt et al. 1996), leading to a lower luminosity of , it is still of extreme luminosity.

Five high-redshift radio-quiet QSOs were selected, primarily from an APM catalogue, by McMahon et al. (1994) because their similarity to F10214+4724 in redshift and overall properties made them natural objects in which to look for infrared emission from dust. BR 1202−0725 (see Fig. 1 for a DSS image of the quasar and surrounding field), at a redshift of 4.69, was the only one to be detected at 1.25 mm with IRAM. Assuming that the detection was due to dust and that the dust had a single temperature (80 K was assumed), McMahon et al. (1994) derived a dust mass of the order of , a considerable amount of dust to exist at such an early epoch. If the far-infrared spectrum of BR 1202−0725 is similar to that of F10214+4724 then it has an infrared luminosity of the order of 10¹⁴ L_⊙, making it one of the most luminous objects in the Universe.

Figure 1.

The field around the quasar BR , taken from the DSS. The quasar is in the middle of the field and is indicated. The scale is consistent with Fig. 3.

Later infrared observations of this object by Isaak et al. (1994) at 450, 800 and 1100 μm, and by Benford et al. (1999) at 350 μm, supported slightly lower dust temperatures of 68 and respectively, although none of these observations saw the turnover of the emission to confirm that it comes from a blackbody. Indeed, its spectral index of between 1100 and 800 μm, calculated by Isaak et al., does not formally rule out synchrotron emission, although a spectral index above 2.5 is difficult to explain by anything other than cool dust.

The detection of CO emission from approximately of molecular hydrogen (comparable to that in a present-day luminous galaxy) by Ohta et al. (1996) and Omont et al. (1996b), and the detection by Omont et al. (1996b) of a close companion also containing large amounts of molecular gas, further reinforced the idea that the galaxy was undergoing a large burst of star formation, possibly interaction-triggered.

2 Observations and Data Reduction

2.1 Observing strategy

The observations of BR 1202−0725 had two goals: to try to detect the turnover in the far-infrared emission; and to detect the near-infrared emission from the continuum to make a better determination of the underlying power law. Early predictions from Isaak et al. (1994) were for flux densities in the range to reach of the order of 100 mJy, depending on dust temperature. We therefore proposed to make observations through the wide-band C90 and C160 filters (centred at 90 and 170 μm) of the ISOPHOT instrument (Lemke et al. 1996) on-board ESA's Infrared Space Observatory (ISO, Kessler et al. 1996). We hoped these two far-infrared observations would lie either side of the peak of the emission.

The ISOCAM instrument, described by Cesarsky et al. (1996), was used to obtain a single observation through the LW10 filter with the aim of detecting the near-infrared emission of BR 1202−0725. A normal radio-quiet quasar with the redshift of BR 1202−0725 would have a flux of approximately 1 mJy in this band, and the ISOCAM observation was therefore planned to reach that sensitivity by employing the microscanning mode.

2.2 Observations

Observations, listed in Table 1, were performed on 1996 July 14. The ISOCAM observation was taken in microscanning mode, with the field of view of the instrument scanned over the object of interest in a raster pattern. The detector was read out every 2.1 s, leading to redundancy in the data set. The redundancy is used to distinguish glitches and other artefacts from valid sources. The ISOPHOT observations were taken as PHT32 rasters of size . The ISOCAM (Siebenmorgen et al. 1999) and ISOPHOT (Laureijs et al. 2000) handbooks give complete descriptions of these observing modes.

Table 1.

ISO observations, taken on 1996 July 14.

2.3 Data reduction

2.3.1 ISOCAM data reduction

The data were sliced and dark-corrected in the standard way (Delaney 2000), making use of a time-dependent dark correction. A thorough first-order deglitching was performed using an iterative sigma-clipping method (Metcalfe et al., in preparation).

No correction was made at this stage for the effects of responsive transients, as the existing correction algorithms can adversely affect the signal-to-noise ratio for very faint sources. The transient correction was instead made at the end of the data reduction by applying an appropriate scaling factor. This factor was established by transient correcting reference sources identified for this purpose in representative data sets.

Dedicated faint-source processing followed essentially the method described by Altieri et al. (1998) and Metcalfe et al. (in preparation).

Long-term baseline drifts were removed by smoothing (via median filtering) the time-history of each individual detector pixel, as sampled by the hundreds of readouts made during the observation, and then subtracting the smoothed baseline from the nominal history. The images recorded at the several raster positions were organized into a cube so that all samples of a given sky position line up. For each sky position, this sky position vector was then sigma-clipped to give a second-order deglitching which has been found to be extremely effective in removing residual glitches. It should be noted that first-order deglitching acts on the time-history of a detector pixel, as it moves over the sky during a raster. Second-order deglitching acts on the history of samples of each sky position, and so takes implicit advantage of the sampling redundancy of the raster measurement technique. The two deglitching steps are therefore highly complementary. Following this deglitching stage the raster mosaic was constructed straightforwardly from the raster position images and standard aperture photometry could be performed on the resulting map.

In order to calibrate the aperture photometry, the data were re-processed after inserting a set of fake sources into the raw data cube. These consisted of theoretical model point spread functions matched to the optical configuration in use. By performing aperture photometry on these fake sources in a manner identical to that used for the actual source, any flux-altering effects of the complex data reduction algorithm could be calibrated.

The fake-source insertion operation was repeated for a range of fake-source brightnesses. In this way the source strength in ADUgs (ADU per gain per second) in the raw data, corresponding to a given source strength in the reduced map, was determined.

At this point it was only necessary to scale the source strength in the raw data for the effects of responsive transients (as described above) and to apply the standard calibration scaling factor relating ADUgs to incident mJy for each ISOCAM filter (Siebenmorgen et al. 1999), in order to arrive at an estimate of the flux of the target source.

2.3.2 ISOPHOT data reduction

The ISOPHOT PHT32 data were reduced using the PHOT Interactive Analysis (pia) package including a pre-release version of PHT32 processing routines developed by R. Tuffs (Tuffs et al., in preparation). Initial data processing (ERD to SCP) used the new PHT32 processing routines, with the later processing (SCP to AAP) using the standard processing routines. Default values for the reduction parameters were used in both the old and new routines.

Maps were generated using the Trigrid interpolation and the first quartile normalization flat-fielding methods.

3 Results

In the ISOCAM data set (see Fig. 2), BR 1202−0725 was detected with a flux of 0.7 mJy. Deriving the scatter in photometry of 10 randomly placed apertures near the centre of the image produces a 1σ value of 0.07 mJy, while the results of the simulations give a (random plus systematic) uncertainty of 0.2 mJy. We therefore take a value of for the flux from this object in the wavelength region. The background flux was determined to be per square arcsecond , consistent with the typical background predictions for CAM LW10 given in the user manual.

Figure 2.

The ISOCAM image overplotted with contours from a DSS2 red image. The quasar is in the centre of the image, with the black line above being the missing column 24. There is a slight shift between the two images of approximately 5 arcsec. This offset is not significant, as the uncertainty in the ISOCAM astrometry arising from the wheel positioning jitter is greater than this. The position of the quasar is marked.

Other surrounding sources (visible on the DSS image) are also present in the ISOCAM image. As they are not in the centre of the image, the observation was not optimized for them, leading to a higher uncertainty on their detections, and we therefore do not quote fluxes for them.

No object was detected at the position of the quasar in either the 90- or 170-μm data set. The ISOPHOT 170-μm map is shown in Fig. 3. The effective noise in both maps was derived by moving a box the size of one C100 or C200 array pixel over the two maps (avoiding the outer parts of the maps) and converting to jansky. This produced 1σ noise values of 16 and 15 mJy at 90 and 170 μm – effectively the 3σ upper limit for both maps is 50 mJy.

Figure 3.

The 170-μm ISOPHOT map. The quasar is located in the centre of the image but has not been detected by ISOPHOT. Its position is marked. This figure is available in colour in the electronic version of the article on Synergy.

Combined results of this and other studies are presented in Table 2. Fig. 4 shows the infrared data, along with a greybody spectrum at a temperature of 68 K. This was derived using the method outlined by Benford et al. (1999), i.e.

with , where , the critical frequency at which the source becomes optically thin, and taking . The current data, with other parameters as given, do not support the dust temperature of 50 K presented by Benford et al. (1999). The data are, however, consistent with the higher temperature of 68 K as given by Isaak et al. (1994). The maximum temperature allowed by the 170-μm upper limit is 80 K, although then the fit to the longer wavelength data becomes poor.

Table 2.

BR 1202-0725 fluxes.

Figure 4.

Data points from Table 2 plotted with a blackbody of temperature 68 K. The ISOCAM data point is plotted as a diamond, the ISOPHOT 3σ upper limits are plotted as squares, and observations with other instruments are plotted as asterisks. The 68-K blackbody is shown as a double-dot-dashed line.

Fig. 5, shown as normalized L against ν, extends the previous figure to include the 6-cm data. It can be compared with the average spectrum for a radio-quiet quasar given in either fig. 10 of Elvis et al. (1994) or fig. 7 of Polletta et al. (2000). If we approximately normalize the fluxes of BR 1202−0725 at 6 cm and R to those of the average radio-quiet quasar, the 15-μm data point is also in agreement, whereas the fluxes from BR 1202−0725 are higher than those of the average radio-quiet quasar, indicating the presence of cool dust in this quasar.

Figure 5.

All data points from Table 2 plotted as relative luminosity against frequency. The ISOCAM data point is plotted as a diamond, the ISOPHOT 3σ upper limits are plotted as squares, and observations with other instruments are plotted as asterisks. The 68-K blackbody is shown as a double-dot-dashed line, and the average radio-quiet quasar spectral energy distribution from Elvis et al. (1994) is shown as a dashed line. The 15-μm flux of BR 1202−0725 is consistent with the average for radio-quiet quasars, whereas it exhibits an excess of emission at the far-infrared wavelengths.

4 Conclusions

We have detected the distant, dusty quasar BR 1202−0725 in the near-infrared and obtained upper limits for the flux at two far-infrared wavelengths. The upper limits for the flux levels at 90 and 170 μm, when combined with the previous ground-based measurements, are consistent with the far-infrared and submillimetre being emitted from a greybody, most probably arising from dust. The maximum temperature allowed by the 170-μm upper limit is 80 K, while the probable temperature is 68 K.

The near-infrared flux level of at 11.5 μm is at a level consistent with it originating in a normal quasar at the distance of BR 1202−0725.

Acknowledgments

The authors thank Robert Priddey for a fruitful discussion concerning models of BR 1202−0725. This paper is based on observations with the Infrared Space Observatory (ISO). ISO is an ESA project with instruments funded by ESA member states (especially the PI countries: France, Germany, the Netherlands and the United Kingdom) and with the participation of ISAS and NASA. The ISOPHOT data presented in this paper were reduced using pia, which is a joint development by the ESA Astrophysics Division and the ISOPHOT Consortium with the collaboration of the Infrared Processing and Analysis Center (IPAC). Contributing ISOPHOT Consortium institutes are DIAS, RAL, AIP, MPIK and MPIA. The ISOCAM data presented in this paper were analysed using cia, a joint development by the ESA Astrophysics Division and the ISOCAM Consortium. The ISOCAM Consortium is led by the ISOCAM PI, C. Cesarsky.

The Digitized Sky Survey was produced at the Space Telescope Science Institute under US Government grant NAG W-2166. The images of these surveys are based on photographic data obtained using the Oschin Schmidt Telescope on Palomar Mountain and the UK Schmidt Telescope. The plates were processed into the present compressed digital form with the permission of these institutions.