Page tree

Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.

Content Block

The JWST NIRSpec integral field unit (IFU) provides spatially resolved imaging spectroscopy over a 3" × 3" square region. Each spatial element in the resulting IFU data cube is 0.1" × 0.1". 

Content Block


Parent articleNIRSpec Observing Modes
See also: JWST Integral Field Spectroscopy , NIRSpec IFU Spectroscopy APT Template

The JWST integral field spectroscopy mode enables spectral imaging over small fields of view. The NIRSpec integral field unit (IFU) is designed for studies of astronomical targets that are extended over a few arcseconds, or for sampling small regions of more extended objects. The NIRSpec IFU use cases range from the study of spatially resolved kinematics and emission lines in distant galaxies to atmospheric or topological features in extended moons or planets within the solar system.

The primary benefit of an IFU is the ability to obtain the spectrum of a contiguous, extended area on the sky (Figure 1). Optical realignment of the spectra is accomplished by an “image slicing” technique using mirrors in the IFU. The resulting slit images are dispersed without confusion from neighboring spatial elements. The 3" × 3" square field of view is dissected into 30 slices (Figure 2) and mapped onto optics to create 900 spatial elements within the field of view. NIRSpec spatial IFU elements are 0.1" × 0.1" on the sky, and can be dispersed using any of the NIRSpec gratings or the prism to acquire spectral imaging data cubes over the 0.6–5.3 μm wavelength range where NIRSpec is sensitive.

For further details on the IFU optics, please see the NIRSpec IFU instrument page.

Figure container
Figure title

Figure 1. NIRSpec IFU 3" × 3" field of view on the sky

Figure caption

Left: WFC3 color image of the field around Supernova 1987A showing a box representing the cutout area displayed on the right. Right: The 3" × 3" NIRSpec IFU field of view shown over-plotted in white on a Hubble Space Telescope WFC3/UVIS F625W image of Supernova 1987A within the Large Magellanic Cloud. A detailed example program for observing SN1987A with both the MIRI and NIRSpec IFU is presented in the article NIRSpec and MIRI Observations of SN1987A.

Properties of the IFU mode

See also: NIRSpec Integral Field Unit

Figure 2 shows the 3" × 3" IFU field of view and how each slice maps to a different position on the detector. This is how the slices would appear in imaging mode on the detector. Each of the 30 slices are 0.1" wide and 3" in length on the sky, which corresponds to a height of 30 pixels on the detector. The spectral dispersion direction is also indicated.

Figure container
Figure title

Figure 2. The 3" × 3" NIRSpec IFU image slices and sky view

Figure caption

The JWST NIRSpec IFU aperture as seen on the Sky (left). Each slice maps to a different virtual slit on the detector (right). The numbers on both panels indicate the slice on the sky (left) and its corresponding location on the detector (right). This can best be seen by clicking on the figure to expand it. As shown in this figure, there are two NIRSpec detector arrays: NRS1 on the left, and NRS2 on the right, separated by a physical gap. The spectra from each slice can be extracted and combined to form a three-dimensional "data cube" that contains the source image for each spectral resolution element. The gap is responsible for some missing wavelengths in the extracted spectra. See Figure 4 for an example.

The NIRSpec IFU entrance aperture is in the same focal plane as the micro-shutter assembly (MSA) shutters and the NIRSpec fixed slits (FSs). This arrangement is shown in a diagram of the NIRSpec entrance apertures in Figure 3. The IFU entrance aperture is a square cutout in the metal mounting of the MSA. The aperture is opened during IFU operations, and closed when the IFU is not in use. In imaging mode, however, the IFU remains open (e.g for verification images and for TA).  

Figure container
Figure title

Figure 3. IFU entrance aperture location and characteristics

Figure caption

The location of the IFU entrance aperture in the MSA focal plane is shown as the small red square. A photograph of the zoomed view is presented to the left.

In IFU mode, the NIRSpec MSA shutters are configured "all closed," and the IFU aperture is opened. Light from the NIRSpec IFU is dispersed onto the detector across the same pixel region as spectra obtained in the NIRSpec MSA mode. There are, however, two MSA instrument characteristics that affect IFU data and sensitivity:

  • Failed open MSA shutters: the MSA has a population of shutters that are permanently or intermittently stuck in an open configuration (for an example of how this affects the data, see Figure 4). These failed open shutters always produce spectra of the sky, no matter the state of the MSA. For an IFU mode observation in a low density region of the sky, these shutters might only contain sky background flux from a blank field. However, in very crowded fields, the IFU spectra will likely be contaminated by the spectra of field objects through these failed open MSA shutters. If this characteristic is considered problematic for science data quality, the IFU aperture can be closed and a MSA leakage correction "leakcal” exposure or set of exposures can be acquired. Such observations can be used in post-processing to subtract the spectral contamination from failed open MSA shutters.  

  • MSA contrast: MSA shutters have finite contrast and are not completely opaque. Flux contamination in IFU observations from finite contrast MSA shutters is referred to as "print-through." This problem is described more completely in the MSA instrument article. MSA shutter contrast is defined as the flux through an open MSA shutter divided by the flux through a closed MSA shutter; higher contrast numbers characterize more opaque shutters. MSA contrasts range from patchy regions of 2,000–3,000 in MSA quadrant Q3, to contrasts of 80,000+ (very opaque) seen in many regions across MSA quadrant 4. While the print-through is generally small for an individual shutter, it can be significant across all 365 MSA shutters in a row in one quadrant. All the shutters in the row contribute in a cumulative way to the flux at a given pixel on the detector. This can result in increased background and a loss of sensitivity in IFU mode observations. The effect can be problematic for very faint IFU targets and IFU targets in crowded regions or with strong nebulosity. If this is likely to affect the sensitivity of science exposures, the MSA leakage correction "leakcal” exposures (as described above) can also be used to correct this contamination from the finite contrast MSA shutters. 

Note:  Dithering can help mitigate the effects of individual failed open MSA shutters on IFU science exposures, but it will not improve the effect of the cumulative MSA print-through from the finite contrast shutters. Beside dithering, there are strategies that can help to mitigate or remove MSA leakage originating from failed open or poor contrast MSA shutters. More information can be found in NIRSpec MSA Leakage Subtraction Recommended Strategies and NIRSpec Bright Spoilers and the IFU Recommended Strategies.

Spectral configurations

See also: NIRSpec Dispersers and Filters

All grating-filter combinations and the prism can be used in the NIRSpec IFU mode. The instrument configurations, spectral resolutions, and wavelength ranges that can be used are outlined in Table 1. For the F070LP filter + G140 grating settings, the minimum measured wavelength is shifted and the blue-ward spectra lie off the detector. This is because the IFU aperture is located on the short-wavelength side of the NIRSpec focal plane (Figure 3). 

Table 1. Instrument configurations, spectral resolutions, and wavelength ranges for the NIRSpec IFU

Disperser-filter combinationNominal resolving powerWavelength range (μm)
G140M/F070LP 1~1,000



Multiexcerpt include
PageWithExcerptMR:APT GUI footnote

The NIRSpec IFU 0.1" spatial elements are projected onto the detector with a factor of two magnification in the dispersion direction. As a result, the undispersed IFU slice images are two pixels wide, even though they project to an area of  0.1" x 3.0" on the sky. The magnification helps to match the spectral sampling of the IFU to that of the MSA shutters and the 0.2" wide fixed slits.

Detector wavelength gaps

See also: NIRSpec IFU Wavelength Ranges and Gaps

There is a physical gap between the two NIRSpec detectors in the focal plane array (Figure 2). This affects NIRSpec IFU observations with the high resolution (R = 2,700) gratings because the spectra are long enough to span both NIRSpec detectors. Unfortunately, the wavelengths of the spectra which fall in the detector gaps are not recoverable in IFU mode. The NIRSpec IFU Wavelength Ranges and Gaps article discusses in detail these gaps and provides figures that show the the gaps for dispersers G140H, G235H, and G395H. The G140H/F070LP configuration in the NIRSpec IFU mode does not have a gap in wavelength because its spectra fall onto one of the two detectors (NRS1). Each IFU slice has a very slightly different wavelength gap range.  


Lost wavelength ranges for the R = 2,700 instrument configurations for the NIRSpec IFU mode are provided in Table 2.

Table 2. Detector wavelength gaps in NIRSpec IFU Mode

Spectral configuration

Detector wavelength gap range (μm)

G140H/F070LP None
G140H/F100LP 1.442–1.479
G235H/F170LP 2.417–2.479



NIRSpec IFU exposures are only acquired in FULL frame 2048 × 2048 detector pixel readout; no subarrays are used.  

Exposure specification

See also: NIRSpec DetectorsNIRSpec Detector Recommended Strategies

NIRSpec IFU exposure times are tied to the timing of the detector readout patterns. There are four readout patterns available for NIRSpec IFU observations:

  • NRS 

The readout patterns are split over two readout modes: (1) traditional and (2) improved reference sampling and subtraction (IRS2). The traditional mode, which is used for the NRSRAPID and NRS readout patterns, is similar to the detector readout for NIRCam and NIRISS. In FULL detector readout, NRSRAPID has a single frame (10.7 s), and NRS has four frames averaged into a single group (42.9 s).

The IRS2 mode, used by NRSIRS2RAPID and NRSIRS2 readout patterns, intersperses reference pixels within the science pixel reads to improve the noise characteristics during data processing,  This results in longer frame times and higher data volumes. In addition, NRSIRS2, has five frames averaged into a single group (72.9 s), compared to NRS with four (42.9 s). These IRS2 readout patterns improve performance and sensitivity in long exposure IFU observations of faint objects.

Additional information on NIRSpec IFU exposure specification and how this translates to exposure time and sensitivity can be found using the JWST Exposure Time Calculator (ETC). The NIRSpec Detector Recommended Strategies article provides guidance about the selection of appropriate readout patterns to use in different science cases.

Options for target acquisition

See also: NIRSpec Target AcquisitionNIRSpec Target Acquisition Recommended Strategies

Target acquisition to place the science target into the small (3" x 3") IFU aperture is likely necessary for most science use cases.  The default option, WATA, will suffice for most cases, but MSATA is offered when WATA is not feasible.  The MSATA process requires an aperture position angle assignment and the selection of reference stars which in turn may require pre-imaging, or the availability of HST imaging to obtain accurate coordinates of nearby stars in the right brightness range. Considerations for the different methodologies for IFU target acquisition are discussed in NIRSpec Target Acquisition Recommended Strategies.

Options for dithering

See also: NIRSpec IFU Dither and Nod PatternsNIRSpec Dithering Recommended StrategiesNIRSpec Bright Spoilers and the IFU Recommended Strategies

Most observations with JWST will require dithering to mitigate detector effects and improve sensitivity. Since the NIRSpec PSF is under-sampled at most wavelengths, dithering is required to achieve nominal spectral and spatial resolution.  For observations with the NIRSpec IFU, there are several options available, including dithers and nods.  Both dither and nod options move science sources on the detector, but nods differ because the offsets are also pairwise subtracted in data processing to remove astrophysical background flux.  IFU dithers are optimized to provide a factor of two sub-sampling of the PSF in both directions. 

Table 2. Summary of NIRSpec IFU dither and nod options

No dither or nodNo dithering or nodding is performed. The target is positioned at the IFU aperture center.
Two-point nod

Two points separated by ~1.6" in both X and Y directions.

Four-point ditherFour points forming a box that's ~0.4" on a side. These four points are separated by a small fraction of the size of the IFU field of view, so they share large areas in common, and the data will be co-added in the calibration pipeline.
Four-point nod

Four points forming a box that's ~1.6" on a side. The data from the 4 nod positions will be paired and subtracted by the calibration pipeline.

Cycling dither pattern

Up to 60-point cycling dither pattern (with either ~1", 0.5" and 0.25" extents) 

Sparse cycling dither patternUp to 60-point cycling dither pattern (with a variety of separations).

The NIRSpec Dithers and Nods page provides an in-depth view of the available options briefly described in Table 2.  The NIRSpec Dithering Recommended Strategies article provides dithering advice tailored to different science goals.  When choosing dithers, one should take into account target acquisition uncertainties, as described in the NIRSpec Target Acquisition Recommended Strategies article. 

Options for background subtraction

See also: NIRSpec IFU Dither and Nod PatternsNIRSpec Background Recommended StrategiesNIRSpec Bright Spoilers and the IFU Recommended StrategiesJWST Background Model

The JWST Background Model describes the different components of the background. The JWST ETC can be used to model and estimate the background in IFU mode, to determine whether background subtraction is needed and what observing parameters will yield the desired S/N on the science source. Bright science sources may not need this correction since the background is expected to be low at NIRSpec operational wavelengths.

Users have the option to do in-frame nodding or off-source backgrounds for background removal. All available options are covered in NIRSpec IFU Dither and Nod Patterns The article NIRSpec Background Recommended Strategies can help the user decide which options to use for their science case.

What do NIRSpec IFU data look like? 

Figure 4 shows NIRSpec IFU mode data acquired with a ground calibration test lamp using the high resolution (R = 2,700) disperser G140H and filter F100LP. Spectra of the 30 IFU slices are the horizontal bands in the figure with labels 1 to 30 (on the left). The fixed slit apertures are always open, so data is acquired through them even in IFU mode. They are shown and labeled "Fixed Slits" in the figure. 

Spectra produced by the IFU are dispersed onto the same region of the detectors used by the NIRSpec MOS mode, so the IFU and MOS modes are not used simultaneously. The red arrows in Figure 4 point to the spectra of failed open MSA shutters that contaminate the IFU spectra.

Pipeline algorithms transform the data read out from the detectors into 3-D data cubes. One wavelength slice of a processed cube of simulated data is shown in Figure 4 in the lower left corner. Example wavelength and flux calibrated spectra of individual pixels are presented in the lower right corner of the figure. 


Figure container
Figure title

Figure 4. NIRSpec IFU example data

Figure caption

An example of a NIRSpec IFU mode calibration showing the detector view (top), the cube view (bottom left), and the spectrum view (bottom right). In the detector view, 30 IFU data slices are shown as horizontal bands that span the two NIRSpec detectors. Spectra in the middle of the detector view are from the NIRSpec fixed slits. The fixed slits are always open and spectra are always acquired on the sky. The red arrows point to spectra from example failed open MSA shutters. The cube view represents one wavelength slice of the pipeline processed cube from a simulated dataset. The spectrum view shows the wavelength and flux calibrated spectra obtained for three of the cube pixels.

Content Block


Lutzgendorf, N. 2016 ESAC JWST "On Your Mark" Workshop
The NIRSpec IFU 

Closs, M., Ferruit, P., Lobb, D. R. et al. 2008 SPIE, 7010, 701011-1
The Integral Field Unit on the James Webb Space Telescope's Near-Infrared Spectrograph

Content Block

Multiexcerpt include
MultiExcerptNameNIRSpec icon and Name
PageWithExcerptMR:NIRSpec icon and name

Content Block



Last updated

Updated March 26, 2018

  • Minor change to indicate frame times for IRS2 patterns

Updated March 23, 2018

  • Minor change to clarify Figure 1 caption

Updated June 30, 2017

  • Table 1 updated

Published December 29, 2016



The format for 'Last updated' is shown below. Enter updates above this internal"Comment" box. Only enter major updates (not typos or formatting changes).

Updated April 05, 2017

  • Lorem ipsum dolor sit amet, consectetur adipiscing elit. Aliquam fermentum vestibulum est. Cras rhoncus. 
  • Pellentesque habitant morbi tristique senectus et netus et malesuada fames ac turpis egestas. Sed quis tortor. 

Published March 02, 2017