Photometric accuracy

In Section 5 of IOM3.1 several provisions in the AOT design have been presented necessary to improve the photometric accuracy of individual observations. Assessment of the in-flight data has shown that:

• effects of detector drift after a flux step cannot easily be corrected for unless either the flux step is very small or the integration time after a step is sufficiently long (of order 32 sec or more) - especially chopped measurements can be affected by drift effects;
• the signal from the fine calibration source (FCS) is contaminated by light from the sky in the PHT near field entering the instrument. The straylight contribution depends on the detector used for the observations. It amounts to several percent of the extended background level in the ISOPHOT central field of view;
• ionizing radiation has increased both the detector dark current and noise. On a low signal level spikes are detected which - in particular for the P3 and C100 detectors - can affect the average signal level for several seconds;
• for P3 and C100 there is a ``switch-on'' effect: at the beginning of a measurement the noise in P3 and some C100 pixels is higher. This effect is particularly noticeable in the first measurement of an observation in which more filters of the same detector is applied. For relatively faint sources (  Jy in most filters) a serious degradation in the S/N is observed in the first 32 sec of the first measurement.

These observations have led to additional provisions in the observation modes of ISOPHOT:

• The duration of an FCS calibration is now 32 sec in order to ensure proper stabilization of the FCS signal.
• Comparison of the sky signal with the FCS signal for absolute photometry is performed by subsequent staring measurements of the sky and the FCS. In the pre-flight design this was performed by rapidly chopping between the sky and the FCS.
• An optional measurement against the cold, i.e. switched off FCS, is provided to determine the signal of the straylight on the FCS.
• An optional measurement of the detector dark signal is provided. This measurement is performed by configuring the filter wheels such that radiation entering the instrument is completely blocked from the detector.
• For polarimetric observations, the target measurement with no polariser is performed only after the polariser measurements in order to keep an increasing flux step on the detector.

In the following we list additional provisions to be taken by the observer in order to achieve the photometric accuracy dictated by the uncertainty in the celestial calibration sources:

• when observing targets weaker than 500 mJy with the P3 or C100 detector use at least 64 sec on-source integration time (128 sec measurement time in chopped mode) for the faintest filter, i.e. the filter giving rise to the lowest total power on the detector (cf section 4.3). The faintest filter will be the one to be executed first in the measurement sequence and is observed to be noisier than the subsequent filters. For the other filters use at least 32 sec on-source integration time.
• reliable photometric results can be obtained after at least 32 sec integration time. Signal drifts are smaller when the incident flux on the detector is increasing between consecutive filters. In practice this means that one should always observe the fainter target first, e.g. in a sparse map sequence or in the concatenation of background and source position.
• the contrast between the on- and off-chopper position should be low in order to avoid significant short term signal drifts. The chopper mode should only be used for faint targets on a strong background. As a rule, the source flux should not exceed 50% of the background flux. The observer should note, however, that a chopped measurement may suffer from loss of signal caused by drifts within the chopper plateaux and vignetting. Correction for these effects can introduce uncertainties limiting the sensitivity of chopped measurements.