Minimize SCIAMACHY Product Handbook

Channels 6-8 (SWIR)

Table of Contents


where fcoadd and tPET are the co-adding factor of the cluster and the pixel exposure time, respectively. Note that the analogue offset is only multiplied with the co-adding factor since it is not dependent on the integration time and is added to the signal for every detector readout. Linear fitting to dark measurements with different integration times yields the in-flight dark signal correction. The dark signal in the UV-VIS-NIR channels is dominated by the analogue offset while the leakage current amounts to only 0.04-0.5 BU/sec and has roughly doubled since launch.

Channels 6-8 (SWIR)

The SWIR channels do not suffer from the Memory effect. However, these channels display a significant non-linearity, i.e. a deviation in the detector response from a (chosen) linear curve. The non-linearity has been measured during the on-ground calibration campaign and a correction algorithm was defined. The maximum value of the non-linearity is around 250 BU which can be significant for weak absorbers such as CO. A separate non-linearity correction for the channels 6, 6+, 7 and 8 has been derived. Within these channels the non-linearity differs for odd and even pixels (starting pixel numbering with 0) because of the different multiplexers used for odd and even pixels. Additionally, there is a clear difference in the non-linearity between pixel numbers higher and lower than pixel number 511. This leads to 14 correction curves, four per channel with the exception of channel 6+, which covers only pixels 794 to 1024. Fig. 5-3 shows the non-linearity curves derived for channel 8. The accuracy of the non-linearity correction corresponds to 5-21 BU for detector fillings from 10000-40000 BU, depending on the channel. As for the Memory effect correction, the non-linearity has to be corrected before any other correction is applied. More details about the non-linearity can be found in (Kleipool 2003).

In addition to the non-linearity, Channels 6+, 7 and 8 contain a significant number of unusable pixels due to the lattice mismatch between the light detecting InGaAs layer and the InP substrate. These channels are doted with a higher amount of Indium (see chapter 3). It changes the lattice constant of the light detecting layer so that it no longer matches the lattice constant of the substrate on which the detecting layer is grown. The resulting degraded pixels are called ‘bad’ or ‘dead’ pixels. There are various effects making these pixels unusable:


image

disconnected pixels preventing any signal readout

image

so-called Random Telegraph (RT) pixels which spontaneously and unpredictably jump between two levels of dark current leading to different detected signals for the same intensity

image

other effects including excessive noise or too high leakage current that saturates the detector




All these effects were measured on-ground and a Bad and Dead Pixel Mask (BDM) was created. Pixels of the BDM have to be ignored in any retrieval. As a result of radiation damage to the detectors in orbit, previously sound pixels may become flagged as bad or dead. A dynamic BDM is determined in-flight based on monitoring and calibration measurements, and updated as pixels change their status as defined by the quality criteria listed above. In order to be less affected by noise on the measurements, the dynamic BDM is smoothed in time.

After the application of the non-linearity and the BDM, the dark signal has to be corrected. The dark signal correction in channels 7 and 8 is complicated by the presence of a large thermal background BGth and the unforeseen growth of an ice layer on the detector (see chapter 6.3). The ice layer slowly changes the detector temperature and attenuates the signal on the detector, including the thermal background. The dark signal in these channels becomes (equ. 4-3)


image



where Γice is the transmission coefficient that changes due to the ice layer and QE is the quantum efficiency for the detector. (fig. 4-3)


Table of Contents
0 Attachments
3600 Views
Average (0 Votes)
The average rating is 0.0 stars out of 5.
Comments