JWST Multi-Object Spectroscopy

JWST multi-object spectroscopy (MOS) is used for simultaneous spectroscopy of multiple sources in a single exposure using NIRSpec's micro-shutter assembly (MSA). Observations are specified using the APT MOS spectroscopy template in APT.  

On this page

See also: JWST Multi-Object Spectroscopy RoadmapComparing Multi-Object Spectroscopy with different JWST Instruments

What is MOS?

See also: NIRSpec Multi-Object Spectroscopy

MOS is an efficient way to observe the spectra of an aggregate of sources that are within the field of view of an instrument at a single pointing. Most MOS instruments are wide field instruments with fields of view covering several square arcminutes up to a square degree. Typically, the source light exiting the telescope optics is focused onto the aperture plane of an instrument where the light passes through narrow slits or fiber optics and becomes dispersed by a refractive element like a prism or grating. The spectra are then imaged onto a detector. The data undergo rectification, registration, and extraction, as well as wavelength and flux calibration, typically using software designed specifically to handle multiplexed spectra.

MOS spectrographs

Over the past 15 years, the demand for astronomical MOS capabilities has expanded significantly. Today, most large 8–10 meter class ground-based observatories offer at least one MOS instrument option. Some employ configurable masks on sources in the FOV. Often, they are fiber-fed to get the light from the sources at different positions in the mask to the dispersive elements of the instrument.

Observation planning with these instruments typically involves the preparation of slit masks which are often milled out on-site prior to the observations. Most ground-based MOS instruments operate primarily at optical wavelengths shortward of J band. Additionally, there are large area IFU spectrographs (e.g., MUSE on the VLT), discussed in JWST Integral Field Spectroscopy. For this article, discussion is limited to the general properties of more standard MOS instruments.  

Near-infrared (NIR) spectroscopy is particularly well-suited to the study of high redshift galaxies whose important rest frame optical diagnostic emission lines are redshifted to the NIR. In galactic investigations, vibrational and rotational molecular lines in the NIR are indicative of the chemo-dynamical structure of stars of the Milky Way and can be used to study stellar radial velocities and abundances. The NIR spectra of young stars provide a means to enhance our understanding of star formation. However, the NIR wavelength regime also presents some challenges. NIR MOS spectroscopy from the ground, particularly from 3 to 5 μm, must contend with high background from the Earth’s atmosphere. Atmospheric seeing from the ground can vary on the same timescales as the observation.

On JWST, the NIRSpec MSA is the primary instrument for MOS spectroscopy. The MSA planning tool (MPT) is used to plan MOS spectroscopy with NIRSpec.

Ground-based MOS compared with JWST MOS using the NIRSpec MSA

See also: NIRSpec Micro-Shutter Assembly

Operating range

See also: NIRSpec MOS Wavelength Ranges and GapsNIRSpec Dispersers and Filters

JWST NIRSpec MSA's operating range is from 0.6 to 5.3 μm. From the ground, NIR spectroscopy is usually limited by the atmospheric absorption and OH (hydroxyl) emission. The MOSFIRE on Keck is its closest ground-based equivalent; it can operate over an extended range from 0.97 to 2.45 μm (into K band) compared to other ground-based MOS instruments which mostly stop before J band, at around 1 μm.


In ground-based instruments, apertures are either slit masks or use fibers arranged into patterns on the sources in the field. For example, at Keck, the MOSFIRE instrument uses a precisely positionable slit mask that can be configured shortly before observations begin. Fiber-fed spectrographs have fibers that can be configured on source positions manually or automatically at the telescope.

The NIRSpec MSA on JWST uses a planning tool to create MSA configurations that are the equivalent of these slit masks. 

Differences between ground- and space-based MOS spectroscopy, related to aperture considerations and angular resolution, are discussed below:

  • The angular resolution of ground-based observations depends upon the seeing, which can vary during the exposures. In space, seeing is not an issue; the angular resolution remains stable and is primarily determined by the instrument optics.

  • In ground-based MOS spectroscopy, the slits are precisely positionable. Conversely, the NIRSpec MSA is a fixed grid; therefore, not all MSA sources will be well-centered. As a result, slit losses will generally be greater compared to ground-based MOS spectroscopy. Accurate astrometry and pointing will allow for precise slit-loss corrections for point sources.

  • A small percentage of MSA shutters have impediments: there are stuck shutters (open or closed), as well as shorted rows and columns. The MSA Planning Tool will help to plan around most of the problematic shutters. The tool uses an MSA shutter operability map that is expected to be monitored and updated frequently (about every 1 to 2 months). However it should be noted that there is a population of shutters that can become stuck closed or unstuck with each new movement of the magnet arm to configure the MSA. Only sources that happen to fall in a new undetected failed closed shutter would be impacted. On the other hand, a new failed open shutter can contaminate the spectrum of a source in the vicinity. However, the number of stuck open shutters is small and is expected be stable after launch. The planning tool will plan around these problematic shutters.

  • The MSA suffers from light leakage between the shutter doors and the bars that support the doors. These leaks are discussed in detail in NIRSpec micro-shutter assembly. Typically this will not greatly affect target spectra unless there is broad scale nebular emission in the MSA field of view.

  • The MSA shutters are very small—each shutter has an open area of approximately 0.20" × 0.46", surrounded by bars of width ~0.07" on all sides. Slits on ground-based instruments are typically larger, ~1", matched to the seeing conditions. The narrower slit widths in the MSA are afforded by NIRSpec's sensitivity and JWST's high angular resolution. These tiny tolerances highlight the need for greater astrometric accuracy for MSA observing, but especially relative astrometric accuracy. Target acquisition will remove the absolute astrometric uncertainty.

  • Absolute astrometric correction for space-based operations will require target acquisition, which for NIRSpec MOS is done using reference stars across the MSA field of view at the start of a visit. Mask alignment strategies are similarly performed for ground-based MOS spectroscopy, but the accuracy requirement is less stringent because the slits are typically larger and the PSF wider.


See Also: NIRSpec MOS Recommended Strategies

Multiplexing is the ability to observe a large number of targets in one exposure, at a given pointing. The multiplexing efficiency varies for MOS instruments, largely because the size of the field of view varies. The FOV of the NIRSpec MSA is smaller than for ground-based instruments, but JWST has high spatial resolution compared to seeing-limited ground-based MOS spectrographs. Since the JWST PSF is smaller, more spectra can be packed into a smaller distance in the cross-dispersion direction. In crowded fields, this provides an advantage over ground-based instruments.

Sensitivity and background

See also: NIRSpec Sensitivity, NIRSpec Bright Source Limits, Components of the JWST BackgroundBackground-limited JWST ObservationsNIRSpec Background Recommended Strategies

At 3 to 5 μm, JWST will have a much lower background than ground-based spectrographs, as the thermal background is greatly reduced in space.

Ground-based instruments are subjected to atmospheric variability. The seeing from the ground limits the spatial resolution in the cross-dispersion direction, and can also lead to reduced spectral resolution if wider slits are required to avoid slit losses. Furthermore, seeing can vary on the same timescale as the observation itself. One component of background is the OH airglow spectrum, which is composed of many atmospheric emission lines in the NIR. Other sky lines are present as well. Often, scientific measurements are only possible between these sky lines, as they add noise and are sometimes difficult to accurately subtract. Space-based spectroscopy in the NIR is not affected by these atmospheric lines. Other types of the background are predominant in space, including zodiacal light and thermal emission from the telescope itself. But these are at a much lower level. The zodiacal emission will vary, but on much larger (seasonal) time scales with the orbital position of the telescope. A description of the background signal components at NIRSpec MOS wavelengths can be found at Components of the JWST Background.

For JWST, the dark current of its state-of-the-art detectors is low. Except for the PRISM, NIRSpec MSA observations are detector noise limited in low background environments and have high sensitivity in the NIR. 

The ETC can be used to estimate the background and to obtain a good estimate of the expected signal to noise in the presence of the various background components. The background signal for JWST can reasonably be considered to be fixed for the duration of most observations.

Process differences

See also: JWST Position Angles, Ranges, and OffsetsNIRSpec MOS and MSATA Observing Process

A number of observing process differences exist between ground-based and JWST MOS observing, including position angle assignment, observation planning iterations, target acquisition, and pre-imaging:

  • For ground-based MOS observing, the telescope orientation is usually not restricted during the cycle and the observer ultimately chooses a desired aperture position angle (APA), which is fixed only when the slit mask is completed. Because of the orbit of JWST, only certain ranges of telescope orientation and APA are available on any given date. Depending on the ecliptic latitude of a target, there may either be one long period when the target can be observed at a given position angle or two shorter periods during the cycle. Orientation angles are assigned operationally for MSA observations to provide scheduling flexibility on JWST.

  • Observers are expected to use the MSA Planning Tool to design placeholder visits for proposal submission at a feasible user-selected APA. After APA assignment, detailed planning for program updates occur about 2 months before observations are executed, again using the MSA Planning Tool, but at the assigned aperture position angle. This is also when the most current knowledge of the evolving MSA shutter operability will be available to observers. The NIRSpec MOS and MSATA Observing Process article describes this process in detail.

  • As mentioned above, the most common target acquisition for this mode is done with reference stars, which must be in the correct brightness ranges (19.5 – 25.7 ABmag in the NIRSpec TA filters). Guidance on finding reference stars and predicting their magnitudes in the NIRSpec TA-band filters can be found in the article on MSATA Reference Star Selection Recommended Strategies.

  • Pre-imaging will commonly be needed to support the increased astrometric accuracy of both the reference stars and the MSA targets. If pre-imaging is desired, the NIRCam pre-imaging observations must be specified completely at proposal submission. But the complete specification of the follow-up NIRSpec MOS observations, including reference stars, must wait for angle assignment, so there will be a second planning iteration that must happen. Other options for target acquisition in the MOS science template exist, and these are outlined in detail in NIRSpec Target Acquisition.


See also: NIRSpec MOS Dither and Nod Patterns

Dithering and nodding along with the long (cross-dispersion) axis of the slit during an observation are useful in MOS spectroscopy to place the spectrum of each source onto different areas of the detector for better estimation of source and background signal and removal of detector artifacts. This strategy is particularly useful in ground-based infrared MOS observing, where it helps to obtain accurate source fluxes in the presence of high background. In instruments that have a marginal sampling, this strategy can improve the spatial sampling and can lead to more reliable estimates of the spectral line profile and spatial distribution of the emission. 

The terminology of dithers and nods for JWST observations departs a little from the meaning of those terms in ground-based observing, where they can mean the same thing. For JWST, dithering means the telescope is moved to place sources in a different location on the MSA, requiring a reconfiguration of the MSA to observe the same targets. Nodding, or movement from one shutter to the next along the slilet in the MSA, does not require a new MSA configuration. Some distinctions between dithering and nodding for NIRSpec MOS and ground-based observing:

  • Nodding is a common practice for all NIR instruments to measure and remove variable sky background.

  • Nodding is used for background subtraction, while dithering is used for improved wavelength coverage (e.g., to provide measurements across a detector gap).

  • Both nodding and dithering can be used to mitigate detector effects and improve data quality.

  • Nodding in JWST "slitlets" moves the telescope over small separations (<1") from shutter to shutter; sources remain within the small slitlets. This is shown in Figure 1.

  • Nodding distances for ground-based spectroscopy is specified by the user.

  • The ground-based seeing-limited PSF requires larger nod offsets for sky background subtraction. Larger nod separations along the slit (>~2" typically) require longer slits and can impact the multiplexing efficiency. Since the slits are often customized to the targets, to perform nodding, the nods can be no larger than the smallest slit length.

Figure 1. Nodding in the MSA

A sketch of the MSA shutters (zoomed in) to illustrate target shutters (red) and associated local background shutters (green). The nodding technique combines the 3 exposures A, B, and C to produce background-subtracted target spectra, as indicated on the right.


See also: NIRSpec MOS Operations - Catalogs and ImagesNIRSpec MOS Operations - Pre-Imaging Using NIRCam

In general, MOS spectroscopy planning begins with a catalog of sources whose positions are known with enough accuracy to be able to place sources into slits. Typically, these positions are derived from imaging of the area of interest. When suitably accurate images are not available, the observer will need to obtain imaging from one of the JWST or HST imaging instruments.

JWST requires high astrometric accuracy catalogs (5–15 mas relative astrometry) to position the very small 0.20" wide shutters accurately onto science sources. Pre-imaging with NIRCam can provide this level of accuracy. Though pre-imaging using NIRCam is not required, it is desirable in many cases. Without this level of accuracy in the catalog, slit losses, and slit loss errors will result in increased data calibration uncertainties. Users may wish to propose for NIRCam pre-images in the same program as their NIRSpec MOS observations. For more details on pre-imaging with NIRCam, please read NIRSpec MOS Operations - Pre-Imaging Using NIRCam. JWST NIRSpec Observation Visualization Tool describes a tool to help with the planning of NIRCam pre-imaging, and NIRSpec MOS and MSATA Observing Process discusses implications to the scheduling of the MOS observation when NIRCam pre-imaging observations are added to the proposal.

NIRCam parallels

NIRCam parallels can be added to NIRSpec MOS observations. There is a one-to-one mapping between exposures of NIRspec and NIRCam, and there are extra overheads associated with the parallels because the instrument setups happen in series, not in parallel. Read JWST Parallel Observations for more details.

Data calibration considerations

Calibration of space-based and ground-based MOS spectroscopy pose different challenges, particularly in the robust estimation of the background, which is considerably higher and more variable for ground-based observing, and in slit loss correction, which must be done carefully for space-based MOS.

  • 3 to 5 μm thermal IR spectroscopy is very difficult from the ground. The background is very high and photometric conditions can vary during the observation, making the slit loss correction more difficult. Conversely, assuming good positional astrometry, and since the PSF is stable (i.e., there is no atmospheric seeing), space-based photometry is more stable and more accurate.

  • The background in ground-based instruments is measured between sky lines in J and H bands, from measurements made by nodding the target in the slit.

  • Background estimation and subtraction in JWST: For point-like sources, the local background is estimated from background shutters in the slitlets, or from a combination of background shutters in nodded exposures. Alternatively, master background regions may be designated with the MSA Configuration Editor. Master background shutter spectra are co-added before subtraction from the source spectra. Background removal is described in CALWEBB_SPEC2.

  • Aperture flux loss correction (i.e., slit loss correction) is less complicated from the ground for sources that may be moderately resolved in JWST, but are unresolved from the ground. 

  • The accuracy of wavelength calibration (and velocity measurements) depends on slit size, source centering, and the errors involved in calibrating the centering which depends on the pointing accuracy and slit losses. Wavelength calibration for NIRSpec is model-based. The data reduction pipeline wavelength assignment is expected to improve with on-orbit calibration. Wavecals are possible, but not recommended for most observations since they may cause persistence in subsequent spectroscopic observations. Off-centered sources will require a zero-point correction in the data reduction pipeline.

MOS terminology

Table 1. MOS terminology



Band or bandpass

A range of wavelengths over which the instrument mode is operable. NIRSpec MOS bands are described in NIRSpec Multi-Object Spectroscopy.


A list of candidates for observation. For NIRSpec MOS, the list should include TA reference stars and all other objects in the field for spectral contamination checking.


Sources whose spectra may inadvertently contaminate planned object spectra due to their proximity to planned sources or failed open shutters.


The quantity, in units of delta wavelength per pixel, that describes how the spectrum is sampled on the detector. NIRSpec MOS dispersion is plotted in NIRSpec Multi-Object Spectroscopy


Moderate repositioning of the telescope between exposures to place the observed sources in a different location on the detector or detectors.

Filler candidates or secondaries

A subset of sources in the catalog with lower scientific priority than the primary candidates that can be used to increase the observing efficiency or multiplexing in an exposure.


Field of view of an instrument.


Line spread function, the characteristic shape of an unresolved spectral feature. The FWHM of the LSF defines the spectral resolution.


NIRSpec's micro-shutter assembly for performing MOS spectroscopy.

MSA configuration

The file which contains the planned MSA shutter status of the NIRSpec micro-shutter assembly. This file is used onboard the JWST telescope to instruct the instrument which shutters should be used to view sources at a given pointing/exposure.


The efficiency with which the instrument can observe multiple sources at one pointing/exposure.


Repositioning the telescope slightly between exposures to place the sources into different positions within the slit. Nodding is a specific type of dithering and it implies pairwise subtraction of exposures in pipeline processing.


 A pixel refers to a physical detector pixel that is read out by some series of electronics.


The practice of obtaining imaging for improving source astrometry prior to spectroscopic observation. For JWST NIRSpec MOS spectroscopy, pre-imaging is done with NIRCam.

Primaries or primary candidates

These are the most important sources for observation defined by the observer.  For many instruments including the NIRSpec MSA, weighting of the Primary candidates is also supported to help deliver these source spectra as products of the planning software.


Point Spread Function. Typically refers to the spatial intensity profile of an unresolved point source either incident upon the instrument optics (i.e., the telescope PSF), upon the aperture slit, or in the calibrated data.

Resolving power

A measure of the disperser's ability to separate spectral lines at an average wavelength. NIRSpec disperser resolving power is plotted in NIRSpec Dispersers and Filters.


An element of the micro-shutter assembly (which contains nearly 1/4 of a million shutters) that can be configured open or closed on a source or background in the FOV. Several may be joined in the cross-dispersion direction to form a slitlet.


An aperture through which the filtered source signal is passed.


A slit in the MSA created from opening one or more adjacent shutters in the cross-dispersion direction.

Slit loss

The loss of light due to the slit size compared to the observed size of the source.

Slit mask

In many ground-based MOS spectrographs, a plate consisting of tiny holes or slits milled out to pass only the light from the sources of interest in the FOV.

Source centering

How close the source is from the center of a slit or shutter.  This depends on the pointing accuracy of the catalog and the telescope with or without target acquisition. Source centering constraints are applied in the MSA Planning Tool.

Spectral resolution

The minimum separation in wavelengths that can be resolved by an instrument unambiguously. This quantity depends on both the resolving power of the disperser and the slit aperture of the instrument mode. See LSF.

Working with MOS data

See also: NIRSpec Multi-Object Spectroscopy

Tools are available on the public github repository to help with the inspection and analysis of post-pipeline MOS data. Most are written in Python and require the installation of Astroconda. More information can be found in the JWST Analysis Tools article. Primarily intended for NIRSpec MOS data analysis are SpecViz and MOSViz. SpecViz is a tool for visualization and quick-look analysis of 1-D astronomical spectra. MOSViz is a quick-look analysis and visualization tool for multi-object spectroscopy (MOS). It is designed to work with pipeline output: spectra and associated images, or just with spectra. MOSViz is created to work with data from any telescope/instrument, but is built with the micro-shutter assembly (MSA) on the JWST/NIRSpec spectrograph in mind. As such, MOSViz has some features specific to NIRSpec data. User training videos can be found in Proposal Planning Video Tutorials.

Example science programs

 Multi-object spectroscopy is particularly well suited to a number of different science goals in the NIR. Some science use case examples include:

  • measuring redshifts of distant galaxies for wide-area cosmological surveys
  • measuring emission lines from galaxies, to obtain constraints on their physical conditions
  • deep continuum spectroscopy of distant galaxies, covering feedback-sensitive outflow diagnostics in the rest-frame UV.
  • Galactic star formation chemical abundances and stellar radial velocities

The science cases referenced below present some concrete examples that may help to understand the rationale behind observing strategies with the NIRSpec MSA. They include a rationale for different parameter choices in the MSA Planning Tool (MPT) in APT for the particular observing strategies used and other important observing considerations.

NIRSpec MOS Deep Extragalactic Survey

NIRISS WFSS with NIRCam Parallel Imaging of Galaxies in Lensing Clusters

NIRCam WFSS Deep Galaxy Observations


Wikipedia article on near-infrared spectroscopy

Latest updates
Originally published