Next: 4.3 Glitches Up: 4. Calibration and Performance Previous: 4.1 Responsivity

4.2 Dark Current

With the term `dark current' we indicate the level of the signal measured when the detector is in darkness, i.e. when no external flux reaches the detector. Strictly speaking, this is not a current, but an electronic reference level including both real dark current and electrical offsets.

4.2.1 Dark level (SW)

With the camera closed, the charge generation in the SW detectors was not due to thermal effects, but to quantum tunneling effects. It increased slowly with the logarithm of on-chip integration time. Thus very long integration times were possible without running into serious dark current accumulation problems. The dark current level and the detector noise just after launch were the same as during the ground tests, and stayed very stable for the whole duration of the mission, with the exception of a small increase in the noise (about 2.5%) after revolution 700 (Boulade & Gallais 2000, [14]). Darks obtained from calibration measurements are included in the CGSWDARK calibration file described in Section 6.1.2. The dark currents provided in the calibration file are estimated to be accurate to about 0.5 ADU/G/s (Biviano et al. 1998a, [5]).

4.2.2 Dark level (LW)

The dark level of the LW array arose as the sum of 2 effects, a leakage of charges generated during the commutation of the reset transistor, and thermal charge generation in the photoconductors. The first effect was dominant for on-chip integration times up to 10 seconds. For longer integration times, charge generation was dominant. The dark frame pattern exhibits a strong line effect with a separation between odd and even lines as can be seen in Figure 4.3.

**Figure 4.3:** *An example of the LW dark frame, measured with 2.1 seconds on-chip integration time. Note the very evident line pattern.*
$\resizebox {11cm}{!}{\includegraphics{dark_calg2s.ps}}$

During in-orbit operations an extensive campaign was performed to study the dark current behaviour. A description of the data and the analysis performed can be found in Biviano et al. 1998a, [5]; 2000, [9] and Román & Ott 1999, [50]. Three type of significant drifts of the LW dark current were found:

a short term drift, i.e. a modification of the dark current within a single revolution, from the beginning of the revolution to the end; this drift was in general non-linear, and it was different for different pixels and integration times;
a long term drift, i.e. a linear decrease of dark current from a revolution to another, with a typical time scale of days; this drift was also found to be different for different pixels and integraton times;
a correlation of the dark current with the temperature of the ISOCAM focal plane which follows a linear relation and is the same for all pixels and for all integration times.

The impact of each of these effects can be summarized as follows:

short term: the drift within each revolution was $\approx$ $\pm$ 0.5 ADU from the activation to the de-activation of the instrument;
long term: the drift was $\approx$ ADU from the beginning to the end of the ISO mission;
temperature dependence: $\approx$ $\pm$ 0.25 ADU over the full range of temperature variation (few hundreths of a Kelvin).

Consequently, a model was developed that takes the different dependencies found into account and corrects the observed LW dark current per pixel as a function of the orbital position of the spacecraft and the temperature of the ISOCAM detector. The parameters of this model are provided in the CCGLWDMOD-file (Section 6.1.3). Once the corrections are applied, the median LW dark current residual amounts to $\approx$ 0.25 ADU (Biviano et al. 2000, [9]). No similar trend was found for the SW channel.

**Figure 4.4:** *The mean signal (top) and noise (bottom) level of the 2.1 seconds on-chip integration time handover LW dark current measurements during the ISO lifetime.*
$\resizebox {13cm}{!}{\includegraphics*[54,537][415,775]{hand_dark_lw2.ps}}$

**Figure 4.5:** The (median-filtered) difference of the mean dark currents of the odd and even line pixels of the LW detector. We show measurements, as a function of time since instrument activation, for 5 seconds on-chip integration time in ISOCAM parallel mode, during three ISO revolutions. No correction for the long term trend has been applied to the data so that a shift is visible from revolution to revolution.
$\rotatebox {90}{ \resizebox{!}{11cm}{\includegraphics{parall5sdark3rev.ps}}}$

To demonstrate the long term trend we show in Figure 4.4 the behaviour of the mean signal (top) and noise (bottom) level of the 2.1 seconds on-chip integration time handover LW dark current measurements (Section 3.7.4) through the ISO lifetime (Gallais & Boulade 1998, [35]). One immediately notes a decrease of the dark current throughout the mission. A more careful inspection demonstrates that the dark pattern on the array changed, the behaviour of odd and even lines being different. The short term trend in the dark level, i.e. the dark current as a function of orbital position or, in effect, as a function of the time since activation of ISOCAM, is shown in Figure 4.5. Note the very similar trend in the difference between odd and even lines of the LW detector. The example is given for three revolutions with the 5 seconds on-chip integration time. For other examples see Biviano et al. 1998a, [5].

Next: 4.3 Glitches Up: 4. Calibration and Performance Previous: 4.1 Responsivity

ISO Handbook Volume II (CAM), Version 2.0, SAI/1999-057/Dc