NIRISS Aperture Masking Interferometry

The aperture masking interferometry (AMI) mode in JWST's Near Infrared Imager and Slitless Spectrograph (NIRISS) offers high spatial resolution, moderate contrast imaging at 2.8, 3.8, 4.3, and 4.8 μm (e.g., binary point source contrasts as high as ≈10-4), at separations between ~70–400 mas. At 2.8 μm it delivers reduced performance.

On this page

See also: NIRISS Observing Modes, NIRISS Aperture Masking Interferometry APT Template, NIRISS AMI Recommended Strategies, NIRISS AMI Science Use Case

NIRISS's aperture masking interferometry (AMI) mode (Monnier 2003) turns the full aperture of JWST into an interferometric array. Light admitted by 7 holes or sub-apertures in an otherwise opaque pupil mask interferes to produce an interferogram on the detector. The mask is designed such that each baseline (i.e., the vector linking the centers of 2 holes) is unique and forms fringes with a unique spatial frequency in the image plane. Since each spatial frequency is sampled only once, the mask is called "non-redundant." (A full aperture, on the other hand, can be considered as made up of infinite sub-apertures and multiple sets of sub-apertures to generate the same baseline making the aperture extremely redundant.) The interferogram created by the aperture mask has a sharper core than that provided by normal "direct" imaging.

The advantage is significant: while the ability to separate closely spaced objects with normal imaging is given by the familiar Rayleigh criterion (separation \delta\theta = 1.22 \, \lambda / D , where $\lambda$ is the wavelength of light and D is the diameter of the telescope), interferometry can resolve objects as close as \delta\theta = 0.5 \, \lambda / D (the Michelson criterion). The AMI mode allows planetary or stellar companions that are up to ~9 magnitudes fainter than their host star and separated by ~70–400 mas to be detected and characterized. It can also be used to reconstruct high resolution maps of extended sources, such as active galactic nuclei.

The AMI mode is enabled by a non-redundant mask (NRM) in the pupil wheel (PW), which is used in conjunction with one of 3 medium-band filters (F380M, F430M, F480M) or a wideband filter (F277W) in the filter wheel (FW).

See the JWST High-Constrast Imaging (HCI) article for a discussion on the various JWST modes that enable high-contrast imaging (HCI). Additional NIRISS-related HCI information is provided in HCI NIRISS Limiting Contrast.

For a walk-through on developing a JWST observing program using NIRISS AMI, please refer to the example science program article NIRISS AMI Observations of Extrasolar Planets Around a Host Star.

Figure 1. Prototype of the NIRISS non-redundant mask

A  prototype of the NIRISS non-redundant mask, which shows the layout of the 7 hexagonal apertures (holes) in the mask with respect to the JWST primary mirror segments and secondary mirror supports. Pairs of these apertures define 21 unique ("non-redundant") vector separations ("baselines") that produce an interferogram on the detector. The holes transmit ~15% of the light incident on the mask. They are smaller than the re-imaged mirror segments to allow for small misalignments in the optical system. Photo credit: Anand Sivaramakrishnan (STScI).

Figure 2. Elements in the NIRISS pupil and filter wheel used by the AMI mode

Optical elements used by the AMI mode

Optical elements available to the NIRISS AMI mode are shown by the green dashed circles. AMI is enabled by using the non-redundant mask in the pupil wheel in combination with one of 4 filters in the filter wheel.
Figure 3. NIRISS AMI simulations

Simulated images through the F430M filter of binary (left column) and single (right column) point sources are compared as direct images (top row) and interferograms produced by the NRM (bottom row). In these simulations, the binary system consists of an 8th magnitude star with a 9th magnitude companion at a separation of 207 mas. Arrows indicate how the interferograms produced by the NRM change in the presence of a binary source. Simulations by Thatte and Sivaramakrishnan (JWST-STScI-004484).

The PSF produced by the NRM has a narrow central diffraction core surrounded by an extended skirt resulting from overlapping fringes produced by each pair of apertures. These features give AMI 2 distinct advantages when compared to coronagraphy or full aperture imaging:

  1. The sharp core of the PSF produces better signal-to-noise ratios close to a bright host star than full aperture imaging.
  2. The interferometric fringes in the outskirts of the NRM PSF are easily measured due to their relative brightness and wider angular extent, making instrumental effects easier to calibrate out of science data.

By comparing the distribution of flux in an interferogram obtained for a target of scientific interest (e.g., the binary shown in the left panel of Figure 3) with the distribution obtained for a reference star that is known to be (or strongly suspected of being) single (e.g., the right panel of Figure 3), AMI observations allow secure detections of faint companions at separations that are not accessible to the coronagraphic modes of JWST. AMI observations acquired through multiple filters allow the spectral properties of the secondary to be measured.

Advantages and disadvantages of space-based aperture masking interferometry

Space-based near-infrared interferometry has 3 advantages over similar ground-based interferometry. The first is that fringe visibilities (or relative photometry) from space are many orders of magnitude more stable than ground-based visibilities. Atmospheric scintillation corrupts ground-based visibility amplitudes, which renders symmetric structures hard to recover reliably. Good fringe visibilities provide more reliable disk brightness and structure.

Secondly, JWST's low thermal background enables fainter targets to be observed, since ground-based 3–5 µm interferometry can be thermal background-limited.

The third advantage is that fringe phases, closure or kernel phases derived from space-based data, are an order of magnitude more accurate than ground data can provide. Interferometric phase data are only sensitive to radially anti-symmetric structure in the target, so fringe phases and amplitudes from space-based data provide leverage for high-contrast binary point sources where the symmetric component is essentially the bright central star and the anti-symmetric component is dominated by the faint companion.

On the other hand, JWST's largest baseline is 6.5 m which is significantly smaller than that of single-dish ground imaging (8–10 m) or multiple telescope interferometers (up to hundreds of meters). Combining AMI data with ground-based near-simultaneous interferometric data can improving the resulting science.

AMI exposure sequence

See also: NIRISS AMI Template APT Guide, NIRISS Non-Redundant MaskNIRISS Target Acquisition, NIRISS Detector Subarrays, NIRISS Filters

An AMI observation sequence usually involves a target acquisition (TA) followed by images using the NRM in combination with the desired filters. A TA is required to ensure accurate and reproducible placement of targets within the small subarray that is typically used for AMI. The TA procedure is optional for applications of AMI that require full frame exposures.

Target acquisitions are accomplished by taking short integrations in a predefined subarray through the F480M filter in the FW and either the NRM (for brighter targets) or CLEARP element (for fainter targets) in the PW. The TA procedure autonomously determines the centroid of the brightest object in the TA subarray. Accurate knowledge of the position of the source at this location is used to command a small slew that places it accurately in the subarray used for AMI science.

The TA is followed by the science exposures, which use the NRM in the PW and one or more of the 4 filters in the FW available to AMI:  F277W, F380M, F430M, and F480M. The science exposure may be taken with a subarray or a full frame aperture, depending on the science requirements or the brightness of the source. Optionally,  one or more direct images using the CLEARP aperture and the same suite of FW filters as those used for the NRM images may be obtained for PSF characterization or related analyses.

This entire sequence is typically repeated for a nearby "reference star," which is single and ideally of similar magnitude and color. When contrast limits are not very demanding, a reference star from an unrelated observation, or possibly a synthetic reference PSF can be used. For more demanding cases, the science target(s) and reference star(s) observations should not be separated by any adjustment of JWST's primary and secondary mirrors.

Filters used with NRM to enable AMI mode 

See also: NIRISS FiltersNIRISS Non-Redundant Mask, NIRISS SensitivityNIRISS Bright LimitsHCI NIRISS Limiting Contrast 

NRM will be used in conjunction with F277W, F380M, F430M, or F480M; these filters were chosen to capture spectral regions of scientific interest. The bandpasses are relatively narrow to preserve the non-redundancy of the u − v (i.e., spatial frequency) coverage. The properties, including estimated saturation limits in the NIRISS filter bandpasses and in the WISE W1 (3.4 μm) and W2 (4.2 μm) bands, are listed in Table 1.

Figure 4 shows the transmission curve of the AMI filters.

Figure 4. Filters for use with AMI mode

A figure showing transmission curves of AMI filters, based on measurements at cryogenic temperatures by the manufacturer.

Table 1. NIRISS AMI filter properties

Filterλavg (μm)aΔλ/λbIWA (mas)cCPFMagnitude saturation (20,000 e-) limit
for Ngroups = 2

























a \lambda_{\rm avg} corresponds to the average wavelength of the filter (\frac{\int P_\lambda\, \lambda\, d\lambda}{\int P_\lambda\, d\lambda}), where P_{\lambda} is the filter throughput.

b Δλ/λ is the fractional bandpass, defined as \Delta \lambda/\lambda = RW/\lambda_{\rm avg}, where RW (the Rectangular Width) is the Equivalent Width/max(P_\lambda), and  EW = \int P_\lambda\, d\lambda.

Inner working angle (IWA) for deepest contrast. Beyond 400–500 mas NIRCam coronagraphs provide higher contrasts.

d CPF is the central pixel fraction, corresponding to the fraction of the total PSF flux in the brightest pixel. It is reported here assuming a 79x79 pixel field-of-view which is relevant to the SUB80 subarray used for most AMI observations.

e When 2 neighboring pixels accumulate charge at very different rates, the brighter pixel “spills” photoelectrons on to its neighbor, but the reverse does not occur. This effect becomes pronounced above about 20,000 e- in the bright pixel. We mitigate this effect in AMI data by setting a signal limit lower than the true non-linearity-based saturation limit for the NIRISS detector. The magnitude system uses the CALSPEC Sirius model from Bohlin 2022 as a standard star with a magnitude of -1.395 in all filters (Rieke et al. 2022).

For Ngroups = 1 the bright limit will be approximately 0.60 mag brighter.

The NIRISS magnitudes in filters F277W and F380M roughly correspond to WISE W1 magnitudes. The NIRISS magnitudes in filters F430M and F480M roughly correspond to WISE W2 magnitudes. There is a ±0.05 magnitude uncertainty due to the conversion from NIRISS magnitude to WISE magnitudes, which is a function of the spectral shape of the source. The magnitudes of the WISE and NIRISS filters should match for an average A0V star and WISE magnitudes are predicted to be slightly smaller than the NIRISS magnitudes for later spectral types. There is an additional uncertainty of order ±0.1 in the simulated NIRISS magnitudes for a given spectrum due to uncertainties in the NIRISS throughputs and quantum efficiency.

AMI detector array and subarrays

See also: NIRISS Detector Subarrays

Science targets for the AMI mode are typically bright point sources. The AMI mode usually uses an 80 × 80 subarray (which includes 4 rows of reference pixels that are not sensitive to light). The subarray can be read out quickly enough to ensure that sources as bright as M ≈ 3 (Vega magnitude system) will not saturate in the F480M filter. For faint targets, the full frame mode can be used. Target acquisition for the AMI mode uses a 64 × 64 subarray. More details are available on the NIRISS subarrays article.

AMI Flat field errors

Analysis of on-sky flat field measurements indicate that current flat fielding errors do not affect AMI detection limits substantially. Point source simulations of a target and calibrator star separated by 2 mas in each of the detector X and Y axes were modeled along with two sources of noise: flat field errors (from on-sky flat measurements) and photon noise. These simulations were analyzed to derive detection limits, which are shown in Figure 5. The simulations considered a shallow exposure to detect 108 photons and a deeper exposure to detect 1010 photons. Detection limits for these two exposures differ by a factor of 10, or 2.5 magnitudes, which is consistent with photon noise-limited results.

Figure 5. Analysis of AMI detection limits incorporating flat field error and photon noise for a "shallow" and "deep" exposure

Click on the figure for a larger view. 

The effect of flat field errors on AMI contrast detection limits. Exposures with 108 and 1010 photons were simulated with appropriate photon noise using 10 flat fields that were created with Gaussian errors that match the actual per-pixel F480M filter flat field standard deviations. The target was placed at the pixel center, the center of a pixel edge, and at one of the pixel corners. The calibrator was offset by (2,2) mas from the target. The detection limits of the target-calibrator pars were calculated and are shown in this plot.

AMI dither patterns

See also: NIRISS AMI Dithers, NIRSS Dithers

The five-point SUBPIXEL dither pattern is recommended. Dithered AMI data may help to mitigate:

  • residual errors in the flat field,
  • effects of hot pixels or bad pixels,
  • effects of cosmic ray hits that are not identified in standard processing, and
  • inter-pixel capacitance variations.

Since these effects occur on different spatial scales on the detector, dither patterns involving spatial offsets of many pixels as well as small fractions of a pixel may be required to treat them.

The dither patterns for the AMI mode are implemented as "primary" dithers that perform ~30 pixel offsets (with up to 4 positions within the 80 × 80 pixels science subarray) in conjunction with "secondary" dithers that are subpixel offsets (0.20–0.33 pixels) designed to obtain adequate pixel phase coverage. More details on the available AMI dither patterns may be found on the NIRISS dithers page


Artigau E., Sivaramakrishnan A., Greenbaum A. Z. et al., 2014, SPIE, 9143E, 40A

NIRISS aperture masking interferometry: an overview of science opportunities

Bernat D., Bouchez A. H., Ireland M. et al. 2010, ApJ, 715, 724B
A Close Companion Search Around L Dwarfs Using Aperture Masking Interferometry and Palomar Laser Guide Star Adaptive Optics 

Bracewell, R., 2003, Fourier Analysis and Imaging, Springer

Ford K. E. S., McKernan B., Sivaramakrishnan A. et al. 2014, ApJ, 783, 73
Active Galactic Nucleus and Quasar Science with Aperture Masking Interferometry on the James Webb Space Telescope

Greenbaum A.Z., Pueyo L., Sivaramakrishnan A., Lacour S. 2015, ApJ, 798, 68
An Image-Plane Algorithm for JWST's Non-Redundant Aperture Mask Data

Hinkley S., Carpenter J. M., Ireland M. J., & Kraus A. L.,2011, ApJL, 730, L21
Observational Constraints on Companions Inside of 10 AU in the HR 8799 Planetary System 

Ireland M. 2013, MNRAS, 433, 1718
Phase errors in diffraction-limited imaging: contrast limits for sparse aperture masking 

Jennison R. C. 1958, MNRAS, 118, 276
A phase sensitive interferometer technique for the measurement of the Fourier transforms of spatial brightness distributions of small angular extent 

Jennison, R. C and Lanham, V. 1959, MNRAS, 119, 174
The brightness distribution within the radio sources Cygnus A (19N4A) and Cassiopeia A (23N5A) 

Keszthelyi L., Grundy W., Stansberry J. et al., 2016, PASP, 128 8006
Observing Outer Planet Satellites (Except Titan) with the James Webb Space Telescope: Science Justification and Observational Requirements 

Kraus A. L. & Ireland M. J. 2012, ApJ, 745, 5
LkCa 15: A Young Exoplanet Caught at Formation? 

Lacour S., Tuthill  P., Ireland M. et al. 2011, Msgnr, 146, 18
Sparse Aperture Maskin on Paranal 

Lacour, S. et al. 2011, A&A 532, A72
Sparse aperture masking at the VLAT. I. Faint companion detection limits for the two debris disk stars HD 92945 and HD 141569

Lawson P. R. 2000
Principles of Long Baseline Stellar Interferometry
Course notes from the 1999 Michelson Summer School, held August 15-19, 1999. Edited by Peter R. Lawson. Published by NASA, Jet Propulsion Laboratory, California

Lloyd J.P., Martinache F., Ireland, M. J. et al. 2006, ApJ, 650, L131
Direct Detection of the Brown Dwarf GJ 802B with Adaptive Optics Masking Interferometry 

Monnier J. D. 2003, Reports on Progress in Physics, 66, 789 
Optical interferometry in astronomy

Rieke, G., H., Su, K., Sloan, G., C., Schlawin, E., 2022, AJ, 163, 45
Infrared Absolute Calibration. I. Comparison of Sirius with Fainter Calibration Stars

Sivaramakrishnan A., Lafreniere D., Tuthill P. et al. 2010, SPIE, 7731, 3
Planetary system and star formation science with non-redundant masking on JWST 

Sivaramakrishnan A. & Artigau E. 2014, STScI Newsletter, Volume 31, issue 01
Aperture-Masking Interferometry with Webb's NIRISS

Sivaramakrishnan, A., Tuthill, P., Lloyd, J. P., et al., 2023, PASP, 135, 015003
The Near Infrared Imager and Slitless Spectrograph for the James Webb Space Telescope – IV. Aperture Masking Interferometry.

Thatte, D. & Sivaramakrishnan A. 2016 JWST-STScI-004484
NIRISS AMI Target Scene Simulations

Thatte, D., Sivaramakrishnan, A., & Stansberry, J., STScI Newsletter, Volume 32, issue 02
Vulcanism on Io with Aperture Masking Interferometry on Webb’s NIRISS

Tuthill P., Lloyd J., Ireland M. et al. 2006 SPIE, 6272, 103
Sparse-Aperture Adaptive Optics 

Thompson A. R., Moran J. M., and Swenson G. W. Jr, 2001
Interferometry and synthesis in Radio Astronomy, Wiley 

Notable updates
    Updated dithering section and added discussion about flat field errors.

    Updated to include a section to discuss the advantages and disadvantages of space-based aperture masking interferometry.

    Updated to reflect in-flight performance as measured during commissioning.

  • Updated Figure 3, made minor changes to the first sentence, updated bright limit in the section AMI detector arrays and subarrays.

    Added 2.8 µm filter option in summary text block

  • Mostly additional text to introduction
Originally published