Introduction to IFU Spectroscopy
Integral field spectroscopy (IFS) is the process of dissecting an astronomical scene into multiple spatial components and dispersing each component with a spectrograph in order to provide spatially resolved spectroscopic information. This capability is available in JWST MIRI and NIRSpec instruments.
Integral field units (IFUs) combine imaging and spectroscopic capabilities to acquire spatially resolved 2-dimensional spectroscopy in a single astronomical exposure. By dispersing the light from discrete spatial elements of the field of view, an image of the source is acquired at each wavelength, and (equivalently) a spectrum is captured at each spatial position (Figure 1). This provides a powerful dataset to study the characteristics of a wide variety of extended astronomical sources, including galaxies, nebulae and crowded stellar fields.
Type of integral field units
Each integral field spectrograph consists of two primary parts; the integral-field unit (IFU) that samples individual spatial elements of the astronomical scene, and a spectrograph+detector system that disperses and collects the light from each spatial element. IFUs are typically divided according to the method by which they sample the astronomical scene. As illustrated in Figure 2, slicers, fibers, and lenslets are all common methods.
Image Slicer IFUs
The Mid-Infrared Imager (MIRI) and Near Infrared Spectrograph (NIRSpec) instruments on JWST have IFUs with an image slicing optical design. These IFUs have purely reflective optics, and use specialized image slicer mirrors (Figure 3) at the IFU focal plane to reformat an astronomical image into aligned slices that form a slit image. The slicer has stacked rows of mirrors that reflect light from different segments of the field of view onto pupil and slit mirror arrays in the IFU optics path. The IFU slice images are then directed through a regular spectrograph slit and diffracted by a grating, thereby resulting in a spectrum for each position. When imaged (not dispersed), the IFU data is comprised of the individual slice images aligned on the detector. In the across-slice direction, an IFU image is sampled by the slicing mirror width, and in the along-slice direction is sampled by the detector pixels. In the spectral dimension, the width of the IFU image (in pixels) of the slice at the detector defines the width of the spectral sample. As a result, dithers on the sky that cross the IFU slices can move the observed spectrum to different spatial locations on the detector, which is useful for data quality to remove effects from the detector pixels or optical artifacts.
Slicer-type IFUs have the advantage of nearly 100% filling factor and a simple (albeit challenging to manufacture) reflective optical design, but the disadvantage that they do not scramble light in the across-slice direction, resulting in a spatial-spectral degeneracy (small offsets within a slice in the across-slice direction are degenerate with wavelength offsets in the dispersion direction). Additionally, they cannot feed long optical trains and must be located close to the dispersing elements of the spectrograph.
Fiber-fed integral field units use bundles of optical fibers to sample an astronomical image and disperse the light of each fiber in a manner akin to traditional multi-object fiber spectrographs. Fiber IFUs have the benefit of scrambling spatial frequencies incident on each fiber, removing concerns of spatial-spectral degeneracy for the spectrum of each fiber, are more easily deployable over large areas of a telescope focal plane, and can feed long optical trains (permitting, for instance, a telescope-mounted IFU to feed a bench-mounted spectrograph). Fiber IFUs can also be made extremely large, with the largest containing hundreds to thousands of individual fibers. However, fiber IFUs typically have smaller filling factors than the other IFU types, are subject to focal ratio degradation (FRD), and are limited in wavelength range by the transmissive properties of the optical fibers.
Lenslet-based IFUs work on the principle of dividing the astronomical scene using an array of transmissive micro-lenses. The beam coming out of each lens is then directed onto a grating (using either collimating optics or a fiber feed system) and dispersed onto a detector. Simple lenslet systems have the advantage of high filling factor, but the disadvantage of a complex data format (often with marginally overlapping spectra on the detector) and a short wavelength range.