Subsections

• 5.3.1 ISO Point Spread Function
• 5.3.2 Straylight

5.3 Optical Performance

5.3.1 ISO Point Spread Function

Early during the Commissioning Phase of the ISO mission, the ISO Point Spread Function (PSF) was measured in order to validate the performance of the telescope optics. The best way of measuring the ISO telescope PSF was to use ISOCAM in one of its configurations which minimised any instrumental effects. For this, calibration data were obtained with a bright point source located at the centre of the detector array, to avoid possible field distortion effects. The micro-scanning mode was used with the minimum allowed raster step size of 2 arcsec in order to obtain the best sampling rate with respect to the limited pixel size of the detector arrays. To avoid the limited resolution of the ISOCAM pixels, a large PSF compared to the pixel size was needed. Thus, the configuration chosen was that of 1.5 arcsec pixel field of view with the LW9 filter at 15 m. The LW9 filter has the longest reference wavelength of the ISOCAM filters and its bandwidth is smaller than that of LW3 (also centered at 15 m). Thus, this filter was expected to provide the largest and less blurred PSF.

The obtained data were reduced using the CAM Interactive Analysis (CIA) software (see Salama et al. 2001, [140] for details on the data reduction) and compared with two independent models, the first one based on the Fourier transform of the telescope aperture, and the second one on ray-tracing with Gaussian decomposition of the beam to simulate diffraction (Okumura 2000, [131]).

Figure 5.1 shows the reconstructed PSF measured with a sampling rate of 0.5 arcsec, determined by the pixel size and the micro-scan step size. A concentric hexagonal pattern can be seen, instead of circular Airy rings, due to the diffraction pattern of the tripod. The tripod introduced a slight difference between y-axis (horizontal in Figure 5.1) and z-axis (vertical) PSF profiles. Figure 5.2 shows the data in graphical form, overlaid with the model based on the Fourier transform of the telescope aperture.

Figure 5.1: Image of the ISOCAM PSF in the configuration of the minimum instrumental effect.

The data show an asymmetry in the diagonal direction, easily visible in the brightness distribution of the first ring of the PSF. This asymmetry is observed at all wavelengths and its amplitude is larger at short wavelengths but it is not clear whether it comes from the telescope or from the camera. Laboratory ISOCAM calibration data obtained in the telescope simulator before the satellite launch already showed a small amount of detectable asymmetry above the noise. This suggests that at least a part of the asymmetry may arise from the ISOCAM instrument itself and not from the telescope but it cannot be reproduced with any model parameters setting for the camera. Introducing a slight offset of the secondary mirror of the telescope (M2). also produces an asymmetry but it cannot reproduce the wavelength dependency of its amplitude over the whole range of ISOCAM. This result should be taken with some caution, as all the model parameter space may not have been explored.

Another deviation was found between the model and the observed ISOCAM PSF width, especially at short wavelengths. The modelled PSFs are narrower than the observed ones, which result in a lower value of the flux measured using the PSF model fit compared to that obtained from aperture photometry. However, the SWS instrument did not observe such a large deviation, which suggests that it arises from the ISOCAM instrument.

Figure 5.2: Plot of the observed ISOCAM PSF in the configuration of the minimum instrumental effect. The dashed line corresponds to the model based on the Fourier transform of the telescope aperture.

Similarly to ISOCAM, the other ISO instruments were also used in-flight to determine their effective apertures, required for the calibration of extended sources, and their beam profiles, required for the calibration of point sources observed off-axis. They all reported some asymmetries in their beam profiles (see Salama et al. 2001, [140] and references therein). However, it was not possible to identify whether these are consequences of the telescope PSF or due to other instrument specific effects.

The LWS mean beam profiles and effective apertures were determined for each LWS detector from a series of observations of Mars taken at different positions in the LWS field (Lloyd 2001a, [116]; Lloyd 2001b, [117]). Off-axis observations of Mars revealed narrower than expected profiles as well as the presence of both strong fringes and spectrum fracturing, which produced an asymmetric beam profile. The fringing is known to be produced by the interference betwen the direct beam from the complex mirror M2 and the reflection from its supporting substrate, but the origin of the spectrum fracturing observed is unknown.

For SWS, raster observations of several bright point sources (HR 5340, Dra, NML Cyg, Car or HR 1457) were used to determine the instrumental beam profiles, which were also found to be narrower than expected and far from flat-topped (Beintema & Salama 2001, [10]). The asymmetries observed are likely produced by internal misalignments between the SWS detectors, especially in the cross-dispersion direction.

Similar measurements were also made with ISOPHOT to derive the instrumental beam profiles and effective solid angles. For PHT-P , dedicated raster observations of bright point sources were confronted with theoretical models of the expected beam profiles revealing strong asymmetries towards the central axis (Müller 2000a, [122]; Müller 2000b, [123]) while for PHT-C, the fraction of the PSF that actually entered a given detector was determined as a function of the source position with respect to the centre of the detector arrays C100 and C200 (Laureijs 1999, [103]). Finally, the beam profiles of the 128 pixels of PHT-S were measured with cross-scans along the spacecraft y- and z-axis of the bright star Dra. In the dispersion direction (z-axis) the beam profile was found to be fairly rectangular as expected, while in the cross-dispersion direction (y-axis) the profile was very sharp, with the peak generally deviating from the geometrical centre of the aperture (Herbstmeier et al. 2001, [80]).

Other particular effects on individual instruments related to their optical performance, like the presence of ghosts or the field distorsion in ISOCAM, or their response to observations made off-axis or on extended sources, are described in detail in the corresponding instrument specific volumes of the ISO Handbook and will not be addressed here.

5.3.2 Straylight

In general, the optical performance of the telescope and baffle system was excellent. In order to fulfill the scientific objectives of ISO, stringent straylight requirements were imposed to the optical system. First, the parasitic light level in the focal plane should not exceed 10% of the minimum diffuse astronomical background for the wavelength range from 2.5 to 240 m. Second, the thermal self-emission from the optical system should also be less than 10% of the minimum diffuse background. Main straylight sources were expected to be the Sun, Earth, Moon, and extremely bright sources like Jupiter.

Several experiments were performed in-orbit using ISOPHOT to verify both the near-field straylight (within 1 radius) and the far-field straylight suppresion.

The near-field straylight measurements were taken with PHT-S at 2.5-12 m, with PHT-P at 25 m, where the straylight contribution from Earth and Moon are largest, and with PHT-C at 170 m, the most sensitive band to detect straylight due to thermal self-emission (Lemke et al. 2001, [113]). With PHT-S, measurements performed at a distance of 60 from the bright source Dra (1600-100 Jy) resulted only in a signal consistent with dark current. In the P_25 filter, double cross-scans over Mira (1500 Jy at 25 m) were performed out to radial distances of 30. At these distances no additional contribution from the beam profile of the bright source on top of the relatively bright zodiacal light background could be detected. With the C_160 filter, the double cross-scans were performed around Saturn (32000 Jy at 160 m). Crossing the source was avoided to prevent memory effects and saturation of the detector. The maximum radial distance was 45. The radial fall-off of the signals was close to the pre-launch predictions. According to the theoretical models, most of the straylight comes from the tripod support of the secondary mirror, the secondary edge and the secondary mirror surface. An example of the straylight effects induced by very bright sources in PHT22 raster maps is shown in Figure 5.3.

Figure 5.3: Straylight observed in a PHT22 raster map taken in the vicinity of Saturn. The planet is located at a distance of 4.5 outside the displayed field close to the lower corner of the image and exactly in the direction indicated by the stripe.

The far-field straylight of the Sun and the Moon were searched for during solar eclipses (Klaas et al. 1998, [99]; Lemke et al. 2001, [113]). For the Sun straylight the telescope pointed at a fixed sky position 60 away, i.e. at the minimum avoidance angle. For the expected drop and rise at the entry and exit of the eclipse phase, an upper limit of 9 mJy at 25 m and 19 mJy at 170 m was derived for a 180 aperture or the sum of the four C200 pixels, respectively. A sequence of 3 successive eclipses were used to assess the Moon straylight. ISO was pointed at three different fixed positions of the sky located at distances of 62, 49 and 28 with the intention of detecting relative differences in the background levels. Again, the variation was below the detection limit and only upper limits of 100 mJy and 86 mJy were derived at 25 m and 170 m, respectively.

Finally, the measurement of straylight contributions from the Earth was performed with the Earth positioned at 3 different directions relative to the oblique conical sunshade, keeping the limb at the minimum angle allowed from the ISO sky pointing. The observations were designed to detect differences in the measured background level, depending on the Earth illumination into the sunshade. The upper limits derived in this case were of 101 mJy at 25 m and 526 mJy at 170 .

Surprisingly, the 170 m serendipity survey contributed to our knowledge of instrumental straylight. A composition of all slews relative to Saturn produced after the mission revealed the presence of a very faint straylight ring emission at a distance of 0.7 to 1.0 with Saturn at its centre. The brightness of the ring is 1% of the central source and was tentatively assigned to grazing reflections at the oversized secondary mirror cover (Lemke et al. 2001, [113]).

In the case of SWS, off-axis responses were measured early in the PV phase on the bright source W Hya, out to a distance of 2. Later in the mission, specific straylight checks were performed, prompted by scientific needs for a better characterisation of the influence of straylight from Saturn on planetary satellites such as Titan and Europa.

One measurement of straylight rejection was performed during observations of Titan on revolution 384. The spectral line due to CH at 13.7 m was measured on Saturn and at a position corresponding to a distance of 3. This gave a rejection factor of 10 (Salama 1998, [139]).

For LWS, spot checks of the flux entering the instrument at distances of 5 and 9 from a very strong source like Jupiter were also performed when the LWS was taking the off-source spectra corresponding to Ganymede and Callisto. The results obtained show that there is a good correlation in both cases between the flux measured as a function of the distance to Jupiter and the model of the ISO PSF by Okumura 2000, [131]. However, the measured fluxes are systematically higher than what was expected from the optical model, which could be due either to the fact that Jupiter is not a point source or to the existence of significant wings in the beam profile.

Straylight effects from extended sources in ISOCAM were investigated through observations of zodiacal background regions with three CVF step positions corresponding to the wavelengths of 7.7, 11.4 and 15 m, and all possible pixel field of views. The images obtained show some contamination by straylight which is found to be an important limitation in CVF images, where real physical structures with an average flux per pixel below 10% of the background are hard to detect.

In addition, straylight measurements were also performed in the surroundings of the bright source NML Cyg in revolution 46. Dedicated raster scans around the source with 6 different orientations taken at a distance of 4.6 from the ISOCAM field centre showed the presence of a faint straight line pattern and a diffuse excess brightness in the direction of the source. However, the flux of NML Cyg is roughly 3000 to 5000 Jy at the ISOCAM wavelengths while the brightest pixel of the straylight is about 10 to 10 times fainter. This corresponds very well with the predicted values in the pre-flight optical modelling (Okumura et al. 1998, [130]).

In CAM parallel observations, images taken in the neighbourhood of very bright sources were also found to show bands of illuminated pixels pointing towards the direction where the bright source was located. The effect is again very small but it can be seen in some cases at distances of several arcmin (see Figure 5.4).

Figure 5.4: Straylight observed in a CAM parallel image (left panel) in the neighbourhood of the bright Mira variable HD 117287 (the bright star in the right panel). The distance to this source from the centre of the image is about 4.