EFOSC2 Order Sorting filters



When observing bright blue objects (e.g. blue specphot standards) 2nd order contamination is seen for a number of the grisms. Details are available in this report. The table below lists the wavelengths above which the contamination has an effect. The conclusions of the report are:

  • Grism #1 has a too large spectral range to be subdued by the usage of a special second order cutting filter, but the effect is only present at the extreme red end. Grism #2 is also only affected at the red extreme and has a large range. For these grisms the effect is small though.
  • Grisms #11 and #13 show only a small amount of second order contamination; if the red part of the spectra is needed, the usage of a filter is recommended for observations of very blue objects; for grism #11 it has to cut at 3750Å; note though that this reduces the efficiency in the blue, and reduces the spectral range, as the blue end is at 3380Å without the order sorting filter; for grism #13 the filter has to cut at 4650Å, again reducing the original spectral range
  • Grism #4, grism #6, and grism #16 show strong second order contamination which can however easily be corrected with a filter cutting at about 3900Å for #4 and #6 and at about 5000Å for #16. Thus the spectral range is hardly diminished, but the contamination sufficiently suppressed. The usage of such filters is recommended for the observation of any blue objects.
  • The contamination in grism #17 is of such a magnitude that its usage without a second order cutting filter is strongly discouraged; the filter has to cut at 4400Å and does thus not influence the the spectral range of this grism

Recommended filters by grism

It is possible to use filters to cut off bluer wavelengths and remove the 2nd order. Here we give a list of suggested order sorting filters for this purpose (based on efficiency and cut off wavelength), below we list the full set available. They will not be loaded in EFOSC by default as it is up to the user to decide whether the removal of the 2nd order is important to his/her science goals (remembering that these filters do cut the flux redwards of their cut-off wavelength by 10-20%). Therefore the filters should be requested using the set up form if required.

Grisms Wavelength Range [Å] Second order contamination Cut off required Suggested filter
Gr#01 3185-10940 >9280Å -- --
Gr#02 5100-11000 >9850Å weak -- --
Gr#04 4085-7520 >6300Å strong 3900 GG375
Gr#06 3860-8070 >6280Å strong 3900 GG375
Gr#11 3380-7520 weak 3750 GG375
Gr#13 3685-9315 weak 4650 GG475
Gr#16 6015-10320 >7170Å strong 5000 OG530
Gr#17 6895-8765 >7000Å very strong 4400 OG530
Gr#20 6047-7147 >6900Å ~4400 OG530

Available Filters

The filters are Schott coloured glass filters (more information available at www.optical-filters.com, or see a useful transmission diagram taken from that site) and cut off at the wavelength included in the name (nm). At La Silla we have a selection of circular Schott filters with diameter 60mm (ie, the size of EFOSC2 filters) and also a larger set of square 50x50mm filters which can also be mounted in the filter wheel using special adaptors.

The efficiencies below were generated by dividing spectra of the flat field screen with the filters in by similar spectra without filters, with the same lamp settings and exposure times. Note that the flat field lamps are slightly blue, so the test was also performed with the internal red flat field lamp for some grisms and filters. Those with both are marked with an * in the table below. Blue lines are for dome flats, red for internal flats. For the dome flat lamps, the 2nd order contamination can give a dip in apparent efficiency when measuring it this way (particularly apparent with Gr#16), as there is extra light from the 2nd order in the without-filter flats and dividing by these gives the dip.

Note (29/8/09) that the filter labelled as GG475 has a redder cut off (~550nm), and the one labelled GG495 appears to really be a GG475 filter. All the filters are due to be tested by the optician and will be relabelled and assigned permanent ESO numbers. The plots below are all based on the current labelling of the filters. It is advisable to double check with arc and/or flat field exposures that the desired filter has been correctly mounted in EFOSC during the afternoon before the first night of your observing run.

Filter name ESO number Efficiency relative to filter free observations
-- -- --
-- -- --
-- -- --
GG 375 -- Gr#4, Gr#6, Gr#11, Gr#13, Gr#16,
GG 400 -- Gr#1*, Gr#4, Gr#6, Gr#11*, Gr#13*, Gr#16*, Gr#17, Gr#20,
GG 420 -- Gr#1*, Gr#4, Gr#6, Gr#11*, Gr#13*, Gr#16*, Gr#17, Gr#20,
GG 455 -- Gr#4, Gr#6, Gr#11, Gr#13, Gr#16, Gr#17, Gr#20,
GG 475 -- Gr#1*, Gr#4, Gr#6, Gr#11*, Gr#13*, Gr#16*, Gr#17, Gr#20,
GG 495 -- Gr#1*, Gr#4, Gr#6, Gr#11*, Gr#13*, Gr#16*, Gr#17, Gr#20,
OG 530 -- Gr#1*, Gr#4, Gr#6, Gr#11*, Gr#13*, Gr#16*, Gr#17, Gr#20,
-- -- --
RG 695 -- Gr#4, Gr#6, Gr#11, Gr#13, Gr#16,
RG 715 -- (tested, but cut off too red to be of use)

We have also tested a number of filter and grism combinations on sky, observing blue and red spectrophotometric standards. A page with all resulting plots (spectra and ratios of with filter to without filter spectra) is available here, but note that most of these plots were batch processed and user caution is advised when using them for reference. In the spectra plots, the coloured line is for the spectrum with the filter, and the black line is without the filter for comparison. All data have been reduced (bias subtracted, flat fielded, wavelength calibrated) using the Quick look tool for long-slit spectroscopy . Note that the flat field used in all cases is the dome flat field without any filter present, so the presence of some 2nd order light from the relatively blue flat field lamps may cause unexpected effects.