Spectroscopic Science Modes


This instrument was built primarily for spectroscopy. EFOSC2 is one of the most efficient instruments available when low resolution (dispersion < 1000) is the need of the hour. The different science spectroscopic modes modes are described below. Note that there is a separate page for the Multi-Object Spectroscopy mode. The observer is also referred to the pages on Acquisition images and Miscellaneous information.

The median seeing at La Silla is 0.8", so the unbinned pixel size of 0.12" spreads the light over a large number of pixels in the spatial direction. The binned modes adequately sample the resolution afforded by all slits. Thus unless spectral resolution is of first importance and one has bright targets using the CCD in the binned mode is strongly recommended. At the NTT, binning modes of 1x1 and 2x2 are currently ofered, while we plan to offer 1x1, 1x2, 1x3, 1x4, 2x1, 2x2, 2x3 and 2x4 modes in the near future.

Long Slit Spectroscopy

In this mode one can obtain one or more spectra of 1-2 objects through 1 or more grisms. However, it is essential that all templates in an OB have the same slits. There is no provision for moving the object along the slit from one exposure to another.

The following acquisition templates may be used in this mode

  • Move to slit : to acquire a single target in the slit, usually at the parallactic angle (-9999).
  • Rotate To Slit : to acquire 2 targets in the slit.

A sample jP2PP observing block

The above example shows an OB comprising a sequence of:

  • A RotateToSlit acquisition template to move one of the targets into Slit#1.2  at x-pixel 1100 with the slit oriented along the line joining 2 targets (to be selected at run-time).
  • Grism #7  spectra of 1800 seconds  each with Slit#1.2normal readout  and 2x2 binning  followed by
  • Grism #12  spectrum of 1200 seconds with normal readout  and  2x2 binning  followed by
  • 5 spectroscopic flats with internal (quartz) lamp with Grism #12 and Slit#1.2  with normal readout and 2x2 binning.

The on-source internal flats may be needed for Grism #12 (and other red grisms) to correct for the strong fringing
seen in their spectra. While CCD flexure with pointing is in general much smaller than a pixel, it may still be sufficient to result in fringe residuals after flat-fielding. An on-source Helium-Argon arc lamp spectrum may not be needed unless the science goal is to get accurate radial velocities since the flexure is much less than a pixel.

The FITS file is called  EFOSC_Spectrum.#.fits

Parameter Values   One does not have to (in fact, should not) change the value of any parameter other than:

  • Starplate/Slit:  See the Slit set for a list of the available choices (all the templates should have the same slit)
  • Grism:  See the Grism Set for a list of the available choices
  • Exposure Time:  whatever appropriate in seconds - less than 2700-3600sec per exposure.
  • CCD readout speed: normalfast  or  slow .
  • CCD binning: 2x42x32x22x11x41x31x2  or  1x1 (only for Slit#0.5)
  • Number of exposures:  whatever appropriate.  Try to have at least 2 exposures to eliminate cosmics

Long Slit Spectroscopy - Scanning

This is very similar to Long Slit Spectroscopy except that the telescope is moved across the extended target while integrating in order to sample the spectrum across a larger part of the target. There are 2 ways in which this can be achieved:

1. Unguided scanning across the target
One can apply a velocity to the telescope in the RA and/or DEC directions after the shutter has opened and with the AutoGuider turned off. This is to be used when the telescope scans many tens or hundreds of acrseconds during an exposure or if the required velocity is large (say, less than 0.1 arcsec/sec). Since the telescope is not being guided the motion is not precise and one may have a 1" wander around the desired track.The problem with large translations is that the guide star may fall off the edge of the guide probe field. In the case of large velocities the star may move out of the target square between one guide correction and the next (done every 5 seconds). Follow the recipe given below :

  • Identify the starting and ending points and calculate the speed.
  • Acquire an object near the starting point into the slit using the MoveToSlit template (with guiding on)
  • Pause the templates after the acquisition
  • Turn he guiding off
  • Type the required velocity value on the Telescope Control Software (TCS) panel.
  • Wait for shutter to open (the status is shown on the EFOSC Status Display panel)
  • Start the telescope movement
  • Stop the telescope movement after the exposure has ended.

One may have any number of templates in the OB but each template should ask for just 1 exposure because the OB has to be paused and the telescope brought back to the starting position before starting the next exposure. The grisms used may (or may not) change from one template to another. If different exposure times are used the telescope velocity has to be modified accordingly.

You will see the guide star starting to move away from the guide target area once the guiding is switched off. If one wants to repeat the exposure,  with the same grism or another one, one can ask the telescope operator to offset the telescope so as to bring the guide star back into the guide target area - this will ensure that the telescope is again pointing to the starting position. This should be much faster than requiring the image with an acquistion template. An alternative is to repeat the telescope movement backwards for the next exposure.

Some other Issues

Try and reduce the scan speed so as to be able to use the differential tracking method; aligning the slit along the major axis of the extended source is one way of soing so.

Always confirm the direction of motion of the telescope - there are several sources of confusion : whether the velocity refers to the movement of the telescope or the movement of the source across the field. One should ask the support astronomer to confirm the sense by trailing a star on the image with a velocity of known sign.

These observations require constant communication between the observer and the telescope operator and manual intervention by the latter over and above the regular actions - there is scope for a lot of confusion. Try to keep the intervention as simple as possible by, for example, confining the scan to just RA or DEC and providing clear instructions to the telescope operator.


The SpectroPolarimetric mode is very similar to the Long Slit Spectroscopy  mode  except for the following differences:

  • MoveToPixel  is the acquisition template (as for long slit spectroscopy) but
    • the parallel Wollaston Mask  (WollMask=) should be chosen against  Slit for reference
    • the choice of x pixels is limited by the locations of the transparent sections of the SpectroPolarimetric Wollaston mask/slit  (refer to  Identifying the Acquisition Reference Pixel). For the same reason the RotateToSlit acquisition template may not be of much use but it can be used (for aligning the slit along the major axis of a small extended galaxy, for example)
  • A different science template is used: EFOSC_spec_obs_Polarimetry:
    • The 20" Wollaston prism has to be chosen for the filter wheel (the template helpfully labels the field Wollaston Prism instead of Filter). The 10" prism cannot be used since the parallel Wollaston Mask  (WollMask=) should be chosen against  Starplate
    • Typically, the half wave plate should be moved in and set to fixed values (continuous rotation false  and a list of HWP rotator positions).

Refer to the page on Polarimetric optical elements as well.

Any one of the 2 available SpectroPolarimetric Masks/Slits  may be mounted on the aperture wheel but the name in jP2PP will be the same. The Wollaston Prism in the grism wheel must be aligned in a manner appropriate for a parallel mask.

Important  When the half wave plate (HWP) is set to continuous rotation in a template the counter ends up with 360xN degrees. When a signal is sent to set the HWP to some small fixed value (say, 45 degrees)  it tries to undo all the N rotations and in doing so triggers a time-out error. So we have included a command at the end of every Polarimetry template to initialise the HWP if it was set into continuous rotation mode - this just takes a few seconds. However when such a template is aborted, for whatever reason, this initialisation does not take place. The only solution is to cold-start the EFOSC2 instrument which takes about 1-2 minutes. So avoid aborting a Polarimetry template - use the STOP option instead.

A sample jP2PP observing block

The above example shows an OB comprising a sequence of

  • A MoveToSlit acquisition template to locate a spectropolarimetric target on x-pixel 1100 (assuming that this is on a transparent section of the mask/slit with the rotator set to parallactic angle (calculated by the script).
  • 1 exposure of 1200 seconds (2 simultaneous spectra with orthogonal polarisations)  at each of 4 HWP positions (0, 22.5, 45, 67.5 degrees)  with  Grism #6 ,  normal readout and 2x2 binning (i.e. 4 frames in all).

The FITS file is called EFOSC_SpecPol.#.fits

Parameter Values   One does not have to (in fact, should not) change the value of any parameter other than:

  • Grism :  See the Grism Set for a list of the available choices
  • Exposure Time  :  whatever appropriate in seconds - less than 2700-3600sec per exposure.
  • CCD readout speed : normal , fast  or  slow .
  • CCD binning :  1x1 is never necessary since the narrowest available slit is 1" wide.
  • Number of exposures  :  whatever appropriate.  Try to have at least 2 exposures to eliminate cosmics


The use of filters in combination with a grism can be useful to isolate the spectral region of interest and to reduce crowding and sky background intensity. Note that the spectral coverage depends on the position of the objects. Measures are needed to define the wavelength ranges as a function of the object position and the overall efficiency in this mode.
back to top