Change page style: 

Scattered Light

Both GMOS-N and GMOS-S suffer fairly significant scattered light in spectroscopic mode. The six gratings exhibit different levels of scattered light, with some gratings being more affected by red scattered light than blue. The scattered light affects both science observations and GCAL lamp (flat and CuAr) calibration exposures. Various examples of the level of scattered light in GCALflat spectra can be found on the Flat/Arc Exposure Times webpage. Note that the GCALflat lamp is very red, and scattered light from astronomical sources will have different relative strengths of blue and red scattered light depending on their intrinsic spectral properties.

The two dimensional shape of the scattered light is rather smooth and spatially fills the entire detector, as can be seen from the example images below. The result of this is that a faint science target can suffer significant contamination from a much brighter object within the GMOS field of view (eg. bright objects also falling on the longslit or bright alignment targets in MOS masks). Because the scattered light character is spatially smooth it is removed fairly adequately as part of the sky subtraction procedure, but the increased Poisson noise can negatively impact the final signal-to-noise of the science target. This is particularly a concern for very faint targets observed in the blue during new moon, when the nominal sky background is very low.

In order to minimize the effects of scattered light from other bright targets one may wish to consider one of the following strategies:


  • Avoid placing much brighter objects on the slit, IFU, or MOS slitlet with your science target if there is no scientific reason to include them.
  • For MOS, it may be beneficial to include an offset iterator in your science observation to offset the field a small amount (eg. q=5 arcsec) after the mask is aligned in order to move the alignment stars out of their 2 arcsec wide boxes. If adopting this strategy it is essential to ensure that either the MOS science slitlets have sufficient length to accommodate the offset or employ the corresponding y-offset to the science targets when designing the MOS mask.

Unfortunately, there is little that can be done to avoid scattered light from the science target itself, particularly when observing with the lower resolution gratings. For the higher resolution gratings, it may be advantageous to observe through a broadband filter in order to avoid scattered light contamination from wavelengths beyond the region of scientific interest. Any of these strategies should be adopted with care, we recommend you consult with your contact scientist or the GMOS-N/S instrument scientist if you are concerned with minimizing the effects of spectroscopic scattered light on your science data.

The following images illustrate the spatial distribution of scattered light in longslit spectral images from a single bright star. Both a red star and a blue star were observed on two separate occasions, these data all used the B600 grating.

EG131 (blue star)
Central Wavelength 400nm (262 - 538 nm)
no filterg-filterr-filteri-filterz-filter
[EG131 B600 400nm no filter] [EG131 B600 400nm r filter] [EG131 B600 400nm i filter] [EG131 B600 400nm z filter]
Central Wavelength 900nm (762 - 1038 nm)
[EG131 B600 900nm no filter] [EG131 B600 900nm g filter] [EG131 B600 900nm r filter]
HD195275 (red star)
Central Wavelength 400nm (262 - 538 nm)
no filterg-filterr-filteri-filterz-filter
[HD195275 B600 400nm no filter] [HD195275 B600 400nm r filter] [HD195275 B600 400nm i filter] [HD195275 B600 400nm z filter]
Central Wavelength 900nm (762 - 1038 nm)
[HD195275 B600 900nm no filter] [HD195275 B600 900nm g filter] [HD195275 B600 900nm r filter]

Scattered light longslit B600 spectral images arising from a single bright star. Click to view full images.

The following two sets of images illustrate the relative contribution of scattered light from GCALflat data. All data were taken using identical GCAL configurations and exposure times. The B600, R400 and gratings are compared at both red and blue central wavelengths. Because the GCAL flat lamp is very red the effect of red scattered light in blue data is much more noticeable than the effect of blue scattered light in red data for both gratings. Because identical configurations and display stretches are used one can see how the B600 grating is more affected by red scattered light than the R831 grating.

B600 grating
Central Wavelength 400nm (262 - 538 nm)
no filterg-filterr-filteri-filterz-filter
[GCALflat B600 400nm no filter] [GCALflat B600 400nm g filter] [GCALflat B600 400nm r filter] [GCALflat B600 400nm i filter] [GCALflat B600 400nm z filter]
Central Wavelength 850nm (712 - 988 nm)
[GCALflat B600 850nm no filter] [GCALflat B600 850nm g filter] [GCALflat B600 850nm r filter] [GCALflat B600 850nm i filter] [GCALflat B600 850nm z filter]
R400 grating
Central Wavelength 400nm (192 - 608 nm)
no filterg-filterr-filteri-filterz-filter
[GCALflat R400 400nm no filter] [GCALflat R400 400nm g filter] [GCALflat R400 400nm r filter] [GCALflat R400 400nm i filter] [GCALflat R400 400nm z filter]
Central Wavelength 850nm (642 - 1058 nm)
[GCALflat R400 850nm no filter] [GCALflat R400 850nm g filter] [GCALflat R400 850nm r filter] [GCALflat R400 850nm i filter] [GCALflat R400 850nm z filter]
R831 grating
Central Wavelength 400nm (296 - 504 nm)
no filterg-filterr-filteri-filterz-filter
[GCALflat R831 400nm no filter] [GCALflat R831 400nm g filter] [GCALflat R831 400nm r filter] [GCALflat R831 400nm i filter] [GCALflat R831 400nm z filter]
Central Wavelength 850nm (746 - 954 nm)
[GCALflat R831 850nm no filter] [GCALflat R831 850nm g filter] [GCALflat R831 850nm r filter] [GCALflat R831 850nm i filter] [GCALflat R831 850nm z filter]

Scattered light MOS B600 and R831 spectral images corresponding to GCALflats centered on 400 and 850nm. The MOS mask consists of two widely separated (in y) slitlets (centered in x). The scattered light images in blue light are very faint, mainly due to the lack of blue light in the GCAL flat lamp. The scattered red light images are very smooth for blue central wavelengths, though for red central wavelengths they show some structure both spatially and in wavelength implying that some fraction of the scattered light contribution is dispersed light. All exposures were 120s long using the ND4-5 filter in GCAL. All images are displayed with the same stretch to reveal the difference in scattering between the two gratings. Click to view full images.