5.4 Dark Current Determination

In the pipeline processing, the dark current is taken as the average of the two dark current measurements performed respectively at the begining and at the end of each observation. However, it is known that this estimate of the dark current can sometimes give erroneous results when subtracted from the data, due to an intrinsic uncertainty in the measurement of the dark current. This sometimes leads to negative flux values. In such cases, OLP Version 10 choses either the dark current measurement attached to the observation, or a `fixed dark current' that was determined in dedicated calibration observations, whichever gives the best result, i.e. the less negative values after the dark subtraction. There is also the possibility to redo the dark current subtraction in LIA with the dark current chosen by the user.

The fixed dark current has been determined by three independent methods:

• Two long measurements of the dark background were taken with one of the Fabry-Pérots in the beam with its etalons set non-parallel. The first measurement was taken straight after the mid-revolution detector curing procedure and the second in the last hour before the close of the LWS science window.
• Routine checks of the dark current were made near apogee in every revolution to monitor long term trends in the detector behaviour. As no long term drifts were found in the value of the dark currents, these measurements were combined to obtain an independent determination of the dark signal.
• When the LWS is not the prime instrument, useful data can still be obtained by recording the output of the detectors with low time resolution. Much of the time during these serendipity mode operations the LWS is viewing blank portions of the sky and the data can be used to determine the detector dark currents.

The results of the three independent means of measuring the dark currents are given in Table 5.4. They agree with each other to within one or two standard deviations. The low values for some detectors seen in the serendipity mode derivation are probably due to the lack of sampling of the pre-amplifier output and the different method used for deriving the photocurrent. The higher values derived from the hand-over illuminator operation reflect the fact that these are measured with a short integration time (8 s) and are therefore prone to problems of contamination by radiation hits. It is noteworthy that there appears to be no significant change in the level of the dark current between the middle and end of a revolution. This is in disagreement with predictions from tests made during pre-launch calibration (Price et al. 1992, [36]).

Table 5.4: Detector dark currents for the ten LWS detectors determined from four different observations: the special long observations in revolution 650, the illuminator operations at apogee and the serendipity mode data. The dark currents are given in units of  A. The quoted uncertainties are one standard deviation. The last two columns give the adopted `fixed dark currents' in A and their uncertainty.

	Revolution 650		Apogee	Serendipity	Adopted	dark current
Det.	measurements		meas.	mode	fixed dark	uncertainty
	Mid Rev.	End Rev.			current
SW1	4.890.42	4.960.53	5.681.82	4.980.58	4.960E-16	5.447E-17
SW2	2.150.38	2.110.40	2.421.20	1.980.32	2.080E-16	4.255E-17
SW3	2.230.19	2.310.20	2.580.86	2.000.23	2.200E-16	2.085E-17
SW4	1.210.30	1.250.30	1.320.40	0.890.24	1.180E-16	3.404E-17
SW5	1.630.21	1.670.22	1.720.27	1.350.22	1.560E-16	2.383E-17
LW1	2.390.27	2.630.28	2.770.53	2.260.30	2.500E-16	2.936E-17
LW2	0.100.22	0.170.23	0.420.36	0.170.18	7.300E-18	2.723E-17
LW3	0.490.32	0.490.34	1.201.03	0.390.25	5.310E-17	3.915E-17
LW4	2.230.38	1.940.38	2.521.44	1.840.33	1.760E-16	4.213E-17
LW5	1.400.22	1.180.22	1.280.35	0.980.25	1.210E-16	2.511E-17