Minimize SCIAMACHY Product Handbook

Stray Light

Table of Contents

4.4 Stray Light

There are two types of stray light (Sstray in equ. 5-1), the spectral stray light and the spatial stray light. Stray light is characterised as a fraction of the total measured intensity for a given pixel. Spectral stray light is light of a certain wavelength which is scattered to a detector pixel ‘belonging’ to a different wavelength. It can lead to distortions in the shape of the spectrum. This type of stray light may be caused by a reflection in the instrument after the dispersion of the light beam, or by periodic errors in the spacing of the ruled grooves in a diffraction grating. The source of spectral stray light can be within the same channel, referred to as intra-channel stray light, or it can scale with the intensity in a different channel, referred to as inter-channel stray light.
Spatial stray light is light entering the telescope from outside the instantaneous field of view. It is dispersed just like light from the observation target. Depending on the source of the stray light, the spatial stray light component can add an additional offset to the spectrum and/or distort the spectrum, if the primary source of the stray light has spectral characteristics that differ significantly from the observed target.

Spectral Stray Light

In a full matrix approach, the spectral stray light determination would measure the stray light contribution from each individual pixel to all other pixels separately. In practice, however, this is not always possible. In the case of SCIAMACHY an 8192 x 8192 matrix would be needed making the calculation of stray light too slow. Initially the spectral stray light for SCIAMACHY was separated into three types: uniform stray light, ghost stray light and channel 1 stray light. As it turned out, the split into uniform and ghost stray light did not provide a sufficient correction of the stray light, so the uniform stray light correction was expanded to include a reduced matrix correction.
Stray light was characterised on-ground using measurements employing a monochromator. A monochromator produces light in a narrow, pre-defined spectral band. The centre wavelength of the spectral band can be adjusted. In the derivation of the stray light fractions from monochromator measurements it is assumed that any signal in detector pixels outside this spectral band is caused by stray light. During the on-ground calibration the spectral stray light was measured by changing the central wavelength of the monochromator spectral band, thus covering the whole wavelength range of SCIAMACHY. Dividing the integrated light of the monochromator peak(s) by the light detected outside the peak yielded the stray light fraction. The resulting data is part of the calibration data set and is used to correct the spectral stray light in-flight.

Ghost stray light is caused by a more or less focused reflection of one part of a spectrum to another part of the spectrum. It can distort the shape of the ‘true’ spectrum, because it does not add signal to all pixels. During the on-ground measurements many tens of ghost signals were detected in channels 3-8, of which the 20 strongest ones were characterised. The total sum of ghost stray light in a channel is at maximum 1% of the incoming intensity.

The reduced stray light matrix describes any of the remaining stray light as a matrix multiplication of an input spectrum and a stray light matrix. A measured spectrum is resampled to lower resolution (1022 pixels instead of 8192), and yields a stray light spectrum of 2048 pixels after matrix multiplication. This stray light spectrum is interpolated to the full 8192 pixel grid and subtracted from the input spectrum. As of 0-1 processor version 7 that became operational in February 2010 the reduced matrix multiplication is implemented for channel 2. The complete reduced matrix multiplication for all channels will allow for inter-channel stray light correction and will be implemented in the version 8 processor.

For channel 1 the situation is less favourable with respect to stray light levels. The on-ground measurements revealed that the spectral stray light in channel 1 can reach levels of up to 10% of the incoming signal for a typical input spectrum. It is also highly wavelength dependent. The main reason for the larger stray light fraction in channel 1 is the high dynamic range of the spectra in this channel, with the lowest signal 3 orders of magnitude smaller than the highest signal. The initial coarse, artificial separation in uniform and ghost stray light turned out to be insufficient for a correction in channel 1 and an alternative method was formulated already before launch, i.e. before the stray light matrix was implemented for channel 2. The chosen approach combines the correction of uniform and ghost stray light in a modified matrix approach. In addition to the matrix approach discussed above, the approach in channel 1 also considers the polarisation of the incoming light.

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