What Is IFS?

Integral field spectrographs are instruments that allow you to gather spectra of the sky over a two-dimensional field-of-view. Regardless of the technique used to obtain the data, the final product is (usually) a data-cube, with axes of x, y (or RA, Dec; the two spatial axes) and wavelength (velocity).

IFS attempts to solve the main disadvantages of traditional long-slit spectroscopy. These include making poor use of the incident light when the object is extended, either intrinsically or due to poor seeing, and suffering from wavelength-dependent slit-losses due to differential atmospheric refraction (DAR). In these cases, what is really required is the ability to record a spectrum from each part of an extended object. This cannot be done with a long-slit (except by stepping the slit across the target by moving the telescope and recording separate exposures for each position — this is obviously very time-inefficient), hence the need for IFSs.

On the most part, the latest IFS instruments are all optimised for use in the optical-red and near-IR. This is due to a number of technical factors and limitations. Firstly, it is currently difficult to manufacture an optical fibre (these being used in the most common IFS design) that transmits blue light efficiently, and secondly, because the slicer technique (see below) only uses mirrors, which are inherently achromatic and can be easily cooled, they are well suited to the IR. Furthermore, current AO technology is most efficient at longer wavelengths, meaning that today's large optical telescopes are all becoming red/near-IR optimised.

What might you use IFS for?

IFS is ideal for exploring spatially extended sources in a spatially extended way. Any astronomical object where you require spatially resolved spectroscopic information is ideally suited for study with IFS. For some examples of what people have done before see the references page.

  • Pre-main sequence objects
      • Proto-stellar discs, jets, Herbig Haro objects, collimation, mass loss, mass accretion, pre-solar nebula
  • Resolved stellar populations
      • Local group and nearby galaxies, crowded fields, nuclear and bulge regions, most luminous stars, B and A supergiants, supernovae, Luminous Blue Variables, planetary nebulae, novae and cataclysmic variables, HII regions
  • Normal Galaxies
      • Stellar and gas dynamics, supermassive black holes, galaxy nuclei, elliptical galaxies, spirals/bars
  • Active Galaxies
      • Gas dynamics, stellar populations, nuclear activity/starburst, fuelling of the active nucleus, (extended) narrow emission-line regions
  • Groups and cluster of galaxies
      • Galaxy formation, galaxy evolution, tidal interaction and merging, star formation history
  • High redshift galaxies
      • Galaxy evolution, galaxy dynamics, galaxy formation, star formation history, cosmology
  • Gravitational lensing
      • Gravitational lens models, QSO unresolved structure, dark matter in lens galaxies, extinction laws in galaxies

Instrument techniques used to achieve IFS


An integral field spectrograph consists of two components: the spectrograph (bien sûr) and an integral field unit (IFU). The job of the IFU is to divide the 2D spatial plane into a continuous array. This division can be achieved in three ways as described below and illustrated in the figure following:

  • Lenslet array: The input image is split up by a microlens array (MLA). Light from each element of your observed object is then concentrated into a small dot and dispersed by the spectrograph. Because the dots are small it is possible to tilt the MLA about the optical axis of the system so that the spectra do not fall on top of each other, thus allowing the input image to be sampled contiguously (differentiating this technique from slit-less spectroscopy). The disadvantage is that the length of spectrum that can be produced without overlapping is very small and the packing of the CCD is not that efficient. For a better diagram see here.
  • Fibres (with or without lenslets): this is currently the most common technique in use. The input image is formed at the entrance to a 2D bundle of optical fibres which transfer the light to the slit of the spectrograph. The flexibility of the fibres allows the round/rectangular field-of-view to be reformatted into one (or more) "slits", from where the light is directed to the spectrograph, and the spectra are obtained without wavelength shifts between them. The disadvantages of this techniques are: (a) the sampling of the sky is not contiguous since there are gaps between the fibre cores (fibres are cylindrical) and (b) the fibres do not work efficiently at the slow focal ratios at which most telescopes work resulting in focal ratio degradation (FRD). Disadvantage (a) can be overcome by placing an array of contiguous lenslets in front of the fibre bundle in order to focus all the light collected by that lenslet into the fibre (lenslet shapes are usually square or hexagonal and thus can be packed contiguously). An additional benefit of this variation is that the microlenses slow the telescope focal beam so that FRD can be minimised. For a diagram illustrating the fibre+lenslet design see here.
  • Image-slicer: The input image is formed on a mirror that is segmented in thin horizontal sections, sending each 'slice' in slightly different directions. A second segmented mirror is arranged to reformat the slices so that, instead of being above each other they are now laid out end to end to form the slit of the spectrograph. The advantage of this technique is that FRD is avoided and the slicing arrangement gives contiguous coverage of the field at potentially high spatial resolution. Because this system uses only mirrors, it is especially suitable for the infrared since it is inherently achromatic and can be cooled to cryogenic temperatures. Potential disadvantages are: (a) that the sampling along the slices is the same as that provided naturally by the telescope meaning there is reduced scope to include a slicer within a spectrograph that must also have a normal long-slit mode and (b) the optical system might be bulky and difficult to fabricate.
      • Concept references: Content, 1997, SPIE, ADS, see also Vivès & Prieto, 2005, astro-ph
      • Examples: VLT SINFONI, Gemini NIFS

The main techniques for achieving integral field spectroscopy.
Credit: M. Westmoquette, adapted from Allington-Smith et al. 1998


The image slicer technique. See also here Credit: ESO VLT SINFONI webpage

Fabry-Pérots (tunable filters)

Fabry-Pérot imaging spectrographs are in many ways different from what is usually meant by integral field spectroscopy, but the result is a data-cube nevertheless. Fabry-Pérots allow a large FoV (e.g. 5 arcmin) to be surveyed at high spatial resolution (e.g. 0.3 arcmin/pixel) in a single exposure but only at a single wavelength — the required data volume (x, y, $\lambda$) is built up by scanning through the desired wavelength range (in a similar fashion to radio receivers). Advantages of this technique are that it can achieve a very high spectral resolution (by producing a transmission fringe which is extremely narrow in spectral bandwidth), as high a spatial resolution and coverage over the 2D field as is achievable with the telescope system, and high throughput. The disadvantage is that it is as time-inefficient to build up the cube as stepped long-slit spectroscopy. Low resolution Fabry-Pérots can also be used in place of narrow-band filters for flexible imaging purposes. [Examples: WHAM, SALT RSS]

Historically, Fabry-Pérots have been less popular than the other IFS techniques because it has proved extremely difficult to perform accurate flat-fielding and flux calibration of the data.

In a little more detail, the principles of Fabry-Pérots are as follows:
The interferometer itself, also called etalon, consists of two semi-reflecting mirrors that only allow light with wavelengths to pass which are integer times the distance between the plates d. This integer is the interference order p. An observation taken through such a system is in effect a vary-narrow-band image, however with changing wavelength over the field of view, since $p \times \lambda = 2 \, n \, d \, cos(\theta)$ where $\theta$ is the angle from the central ray, which is the only one that passes the FP perpendicular to the plates (i.e. $\theta = 0$).

After each short exposure, d is changed, resulting in the next image, also called channel. One scans through the whole (albeit usually very short) free spectral range (FSR) to build up a data cube. The FSR is given by $2 \, n \, d \, cos(\theta) / p^2 = \lambda / p$. During data reduction, the wavelength of each channel is calculated, and the flux is reorganised such that a data-cube is created with the z-axis corresponding to wavelength.

This wavelength-scanning approach is only efficient for small FSRs (usually a few hundred km/s) which are sampled in a number of channels, n, that match the resolution of the FP as given by its finesse: $R = p \times F$. The finesse depends mainly on the reflectivity, R, of the FP plates: $F = \frac{\pi \, \sqrt{R}}{(1-R)}$. Example values are p=1000, F=20, FSR=200 km/s, n=32.


Multi-object spectroscopy (MOS) is similar to IFS. However, MOS is a technique designed to obtain spectra of multiple, separated (usually small) targets over a large field-of-view — e.g. individual galaxies with a cluster or globular clusters within a galaxy. With current designs this is achieved using slit masks or positionable fibres/lenslets.

IFS Data

The raw data from an IFS observation usually consists of multiple (anything from 10s — 1000s) of spectra laid out on the detector, each originating from an individual element of the IFU. These elements are often called 'spatial pixels' (commonly shortened to 'spaxel') — the term being used to differentiate between a spatial element on the IFU and a pixel on the detector. In order to construct the data cube, it is necessary to extract the individual spectra from the detector and then rearrange them back into the geometric pattern that they held at the focal end of the telescope.


The concept of a data cube. Credit: Stephen Todd (ROE) and Douglas Pierce-Price (JAC)


The principle of IFS. Credit: M. Roth 2002

Strengths of IFS

Here is a list of a number of common but difficult problems/issues that IFS gives you the capability to address.

  • Correcting for differential atmospheric refraction (DAR) — this becomes an issue if your data cover a wide wavelength range.
  • Dealing with a wavelength-dependent point spread function (PSF) — this is an issue if your data cover a very wide wavelength range.
  • Creating sensible maps of your field of view in interesting wavelengths — single, or covering particular spectral features — and taking this further, to make line ratio maps or maps of physical parameters derived from your spectra.
  • Separating spatially blended sources (e.g. a crowded star field, nebular emission superimposed on strong background emission).
  • Extracting out interesting regions of the field (and the effect that the PSF has on how well you can separate these).
  • Spatial re-gridding/binning, and choosing whether to optimise the signal-to-noise ratio, spatial resolution, or even maintain distinctiveness of your targets in the field-of-view.

Properly addressing these points may mean a lot of work, but do not despair! The final result will be worth it — think of all the science you can get out of your IFS cubes! We hope that what is offered in this wiki can help you on your way.


Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License