MIRI Coronagraphic Recommended Strategies

Pre-launch recommendations are available for astronomers planning coronagraphic observations with JWST's Mid-Infrared Instrument

Note that what follows are pre-launch recommendations that will be updated as the knowledge about the instrument's optimal usage and calibration progresses.


Parent page: JWST Observing Guidance → JWST Recommended Strategies  → MIRI Observing Strategies 
See also: MIRI Observing StrategiesJWST Coronagraphic Observation Planning,

as well as relevant science use case examples MIRI Coronagraphy of GJ 758 b and MIRI and NIRCam Coronagraphy of the Beta Pictoris Debris Disk.

The Mid-Infrared Instrument (MIRI) on board JWST is equipped with four coronagraphs that will enable cutting-edge science at small inner working angles (IWAs). In order to prepare for such high-contrast imaging (HCI), observers will have to decide on the choice of filter and focal plane mask, target acquisition strategy, and order of observations, among others. 

Planning a coronagraphic observation with MIRI

Which coronagraph (or no coronagraph)?

See also MIRI Coronagraphic Imaging, MIRI Coronagraphic Performance, and MIRI Coronagraph Masks

MIRI incorporates two coronagraphic imaging architectures, each with their own advantages and drawbacks. The coronagraphic focal plane mask assembly comprises of three 4-quadrant phase masks (4QPM) and a traditional coronagraphic occulting Lyot spotEach occulting mask has a corresponding wavelength range and associated IWA.  

The intended science will ultimately determine the choice of coronagraph, depending on the wavelength(s) of interest and, to some extent, on contrast and separation of the science target. When assessing feasibility, you may need to use models to extrapolate shorter wavelength measurements from the 3 to 23 µm range to determine companion contrasts at MIRI wavelengths. 

Lyot coronagraph

The MIRI Lyot spot has a radius of 2.16", which is 3 λ/D in projected radius at 23 µm, and is suspended in the focal plane by two supporting struts in the mounting bracket, which themselves block light in the field of view (FOV). The Lyot coronagraph works only in a broad band centered at 23 μm, provided to maximize the sensitivity on colder and extended objects (cold silicates in circumstellar disks, for instance). It provides an inner working angle (IWA) of ~3 λ/D, making it useful for detecting objects, structures, or diffuse emission at an apparent separation of ≥2.16" from their host. This could include the outer regions of protoplanetary and debris disks, extended structures around post-AGB stars or galaxies, and scattering/ionization "cones" of AGN. For companions very close to their host (within 3 λ/D), observers should consider using the 4QPMs.

Four-quadrant phase masks (4QPMs)

MIRI's 4QPMs are able to reduce the IWA to near 1 λ/D, but at the price of strong sensitivity to optical aberrations and source misalignments. Another defect is the partial attenuation of a point source when it lies in-between two adjacent quadrants. The MIRI 4QPMs are monochromatic, working only in narrow bands (R = 20) centered at 10.65 μm, 11.4 μm, and 15.5 μm. These masks are optimized for detecting mid-IR features in gas giant and terrestrial atmospheres: searching for absorption in the Ammonia (NH3) band at 10.65 µm; measuring the adjacent continuum at 11.4 µm and the longer wavelength continuum at 15.5 µm. They are also useful for studying a variety of other objects, structures, and emissions very close to bright point sources such as the inner regions of debris disks, very tight binary systems, or the near-nuclear environments of AGN.

Utilizing the Coronagraphic Visibility Tool (CVT)

See also JWST Coronagraphic Visibility Tool Help and Target Visibility Tools.Which_coronagraph

The JWST Coronagraphic Visibility Tool (CVT) was created as a resource to assist in the pre-planning and observational strategizing of MIRI (and NIRCam) coronagraphic observations. The GUI-based tool provides the visibility and allowed position angles (PAs) for a given target across the year or at a specific date. It is also particularly useful for visualizing the coronagraphic field of view and its physical limitations (as mentioned above). Observers should use the CVT to evaluate the feasibility of their observations and address the following questions: will scheduling be an issue for my target? What are the restrictions on the allowed PAs and time for my observations? What about the roll flexibility for this observing period? Are there any restrictions that I need to place on my observations? 

Utilizing the JWST Backgrounds Tool (JBT)

See also Backgrounds Tool

The Backgrounds Tool (JBT) is a simple command line tool that accesses the JWST background models to return a plot of the total background intensity, and its components, as a function of time. It can be useful for estimating and visualizing the impact of the background on the scheduability of JWST observations as it returns the number of day per year that a given target is observable at low background for a given wavelength and selectable threshold. 

Selecting a PSF subtraction strategy

When using the JWST coronagraphs it is critical to obtain at least one good reference point spread function (PSF) that matches the science PSF. There are three complementary strategies for obtaining an adequate reference PSF, that are supported by MIRI:

  •  Referenced differential imaging (RDI)

    The RDI strategy uses an observation of a reference star to calibrate the coronagraphic point spread function (PSF) and subtracting it from the image of the target star that contains the astrophysical signal. The quality of RDI, however, is sensitive to several factors. This includes thermally-induced wavefront drifts of the observatory that cause the PSF to change from when the science target was acquired, and imperfect target acquisitions of the science and PSF reference targets. 

  • Angular differential imaging (ADI)

    Here, two coronagraphic observations of the target are taken that differ by a telescope roll (i.e., taken at different spacecraft orients). Differencing the PSF on the same star as the astrophysical signal rotates introduces diversity between instrument artifacts from the astrophysical signal. This technique is highly effective as it allows PSF subtraction at nearly the same attitude (for wavefront stability), mitigates detector artifacts and also eliminates stellar color mismatch terms. However, ADI is sensitive to self-subtraction biases, especially given the limited available roll (~10º max) of JWST.

  • Small grid dithers (SGDs)

    This technique utilizes a defined set of sub-pixel dithered exposures to optimize coronagraphic PSF subtraction–it's recommended only to be applied to the PSF reference target in cases when the highest suppression of the target star light is needed. SGDs essentially provide a small library of reference images that effectively samples the PSF diversity close to the center of the coronagraph mask. Post-processing optimization algorithms (such as LOCI or KLIP) are then used to construct an optimal synthetic reference to be subtracted from the target star image.

    In general, SGDs improve contrast at the coronagraph's IWA by a factor of more than 10 for MIRI. At larger separations, the gain subsides to between 5 and 10. However, while the small grid dither technique enhances performance, it comes at a non-negligible cost; the exposure time is increased by a factor approximately equal to the number of dithered observations. Thus we recommend that the feature is only requested when it is scientifically justified. The 4QPMs, which are very sensitive to alignment differences, are the most likely to benefit most from this option.

Note that because of the limitations and biases associated with both ADI and RDI, for robustness, we advocate for PSF subtraction calibration via both techniques for standard coronagraphic observations, especially now at pre-launch, when predictions are necessarily imperfect. Furthermore, because the JWST PSF will vary in time from wavefront thermal evolution, all science and PSF calibration observations are required to be taken back-to-back (i.e., in a non-interruptible sequence). The inference from this is that the science target and PSF reference target (for RDI and SGDs) must be schedulable in the same visibility windows (you will need to use the CVT).    

Choosing a reference PSFs target

See also Selecting Suitable PSF Reference Stars for JWST High-Contrast Imaging.

The JWST PSF is expected to be time variable, which has important consequences on the choice of PSF reference targets. Their observation is crucial in MIRI coronagraphic observations and will be used to calibrate wavefront errors and other uncertainties in the TA process. By STScI policy, observers are required to define a reference PSF within every program. To ensure effective PSF subtraction, whilst also minimizing overheads and potential contrast losses, the observer should follow these guidelines when selecting their PSF reference(s):

  • Choose a known "good" PSF reference – "good references" are usually stars that are not astrophysically contaminated (i.e., without additional astrophysical signal from a debris disk or companion.)
  • To mitigate thermal changes, choose a reference star in as close temporal and physical proximity to the science target as is feasible. A general guideline is to find a reference star within about 20° of your science target.
  • Choose a reference star that is spectro-photometrically similar to the science target. Choosing a star that is of similar brightness or greater than the science target will ensure an equal or greater signal-to-noise for the reference star, in a similar or shorter amount of total integration time. 
  • The PSF reference star must observable within the same visibility window as the science target, at the desired time of the observation. Observers can confirm this by utilizing the CVT.

A useful tool to aid in the selection of a good PSF reference target is the Jean-Marie Mariotti Center (JMMC)'s SearchCal, a GUI that allows to select suitable, non-resolved calibrators using a number of search criteria.

Telescope rolls

The roll capacity of JWST is limited at a given time; typically 10° total (±5° off nominal), but larger rolls can be obtained between different epochs depending on the target position on the sky. In coronagraphic observations, telescope rolls are used in two applications:

  1. Angular differential imaging

    This type of roll was mentioned previously to obtain a self-reference for PSF subtraction; it also provides better sampling and increased robustness against detector defects or artifacts by rotation of the PSF. In this case, the roll angle must be sufficient to prevent self-subtraction of astrophysical targets of interest (e.g., for point source, the rotation must move the point source by at least one PSF FWHM). This technique is wavelength dependent, thus longer wavelengths are associated with larger IWAs. For MIRI at 10.65 μm, the minimum separation to avoid self-subtraction of a point source is obtained for an IWA of 1.9" with the typical ±5° roll. At 15.5 μm, this IWA becomes 2.8". For all science observations at separations below this IWA, the rolled science target cannot be used as a self-reference in PSF subtraction unless the second observation is taken at a roll difference greater than 10° (and, therefore, at a significantly later epoch).

  2. Moving astrophysical targets of interest off of PSF structures

    These include structures such as diffraction spikes, or axes of the 4QPM masks, obtained at two orientations. However, this typically requires rolls larger than 10° and therefore an observation at a later date, albeit with more significant changes in the wavefront error compared to back-to-back observations. As such, we recommend the second observation require its own PSF reference star to calibrate these wavefront changes. The CVT can be useful in planning for larger roll offsets to assess the availability of multiple position angles and to estimate the time separation between observations.

Target acquisition

Target acquisition will be required for all coronagraphic observations with MIRI to achieve the pointing requirements for optimal contrast and performance. The acquisition filters available to users for all coronagraphic imaging modes are F560W, F1000W, F1550W, and the neutral density filter (FND). The FND provides the strongest flux attenuation and is recommended (especially for the 4QPMs) to avoid saturation and persistence when observing bright targets.

In general, users should consider using the FAST readout mode; however, for TA with the Lyot coronagraph concerning fainter stars (where longer than the shortest integration times are needed), FASTGRPAVG should be used. Users must determine the exposure time required to obtain a sufficient signal-to-noise for the TA procedure to achieve the desired centroid accuracy. In order to achieve the centroid accuracy requirements for MIRI target acquisition, the minimum required integrated (within the extraction aperture) signal-to-noise ratio (SNR) must be ≥30. Saturation can also affect the accuracy of the centroiding procedure, and should be avoided.

The TA can be achieved in any of four locations concentrically distributed about the center of each coronagraphic subarray. It is generally recommended to select the quadrant in which the companion source of interest does not reside. If this is not an option, or if performing TA on bright targets, performing a "SECOND EXPOSURE" may be considered (see below). If selected, the observation will be repeated with the target acquisition performed in the diagonally-opposite quadrant (i.e. 13, 24, 31 and 42). 

There are two effects that make target acquisition with MIRI’s coronagraphs complex: (1) the phase masks can distort the image of a star close to its center, undermining the centroid determination; and (2) the detector's arrays have latent images that could mimic astronomical phenomena. 

Avoiding latent images

Slowly decaying latents can produce residual signals that will interfere with the centroid measurement algorithm and further limit the accuracy of any TA procedure around bright stars. If the user decides to utilize a single position for TA, we recommend performing TA as far out as possible and keeping in mind the corresponding slew accuracy (see below), in order to avoid strong latent images close to science targets. Estimates indicate that this position can be as far as 4" away and suffer minimal performance degradation. Performing TA at smaller angular separations may imprint long-duration latents and undermine the entire rationale for the MIRI coronagraphs of observing expolanets down to λ/D from their stars. To mitigate this, observers can consider using the SECOND EXPOSURE strategy; here two observations are taken: one with TA using the 1st ROI and second with TA using a 2nd ROI that is diagonally opposed to the first one. This will allow for discrimination of persistence images because they are time variable, but at the cost of precision in the position repeatability. 

Avoiding centroid error

In order to minimize the effect of centroid error on TA induced by the 4QPM, it is recommended for observers to avoid the 4QPM axes (which can introduce errors as large as 100 mas on the centroid measurement), and aim for regions with the smallest errors, such as on the diagonal.

The error on the observatory slews is also an important factor during TA. Observers must therefore try to find a compromise between large slews, which have inherently larger errors, and close-in TA, with inherently larger error on the centroid induced by the 4QPM and possible contamination from latent images at the location of scientifically interesting objects.  In general, we recommend performing target acquisition no closer than 500 mas to the 4QPM center and no further than ~750 mas.

If using SECOND EXPOSURE TA, the centroid measurements of the two symmetric acquisitions are averaged and used to make the final slew to the center of the coronagraph. This results in smaller mean offsets when compared to scenarios involving a single TA, especially for the 4QPMs since the two acquisitions help average out the symmetrical effect of the 4QPM on the centroid error. However, this technique requires more slews, which translate in a larger final dispersion of the pointings. Performing a single acquisition yields the best repeatability, albeit less accurate centering, making it the favorable TA procedure for satisfying the pointing requirements for the MIRI 4QPMs. We recommend restricting the use of the SECOND EXPOSURE strategy to cases where possible latent images will compromise coronagraph science (e.g. searching for planet companions at small angular separations).

Figure 1. Centroid error as a function of position on the 4QPM at 11.4 μm (for positions ≤500 mas)

The vectors point from the true position to the actual centroid (measured) position. We see that the 4QPM can introduce errors as large as 100 mas on the centroid measurement depending on the position of the star relative to the 4QPM center and that the position of the center of the 4QPM is only known to within 1-2 mas. The cross-like pattern, in which the centroid measurement error is largest, is attributed to the axes of the 4QPM.

Utilizing the Exposure Time Calculator (ETC)

See also JWST Coronagraphy in ETC.

The estimation of exposure times is a science-critical aspect of observation planning. It is crucial to estimate exposure times for both science observations and target acquisition. in fact, ETC workbook and specific calculation ID #s are now a required field for coronagraphic observations specified within the Astronomer's Proposal Tool (APT, version 25.4.2).

The purpose of the JWST Exposure Time Calculator tool is to provide users the information necessary to design a proposal that will yield successful observations (i.e., that the astronomical observables sought by a given program will be accessible at the level of SNR required to carry out the scientific interpretation of the data). For MIRI coronagraphic observations, the current implantation of the ETC is useful for the following tasks: (1) investigating detector readout patterns and associated saturation and (2) computing the SNR of an off-axis source, and (3) obtaining an idea of which mask and coronagraph faint companion source under the ideal contrast assumption, with background and PSF subtraction. Doing so will provide the observer with the information necessary to design and determine the best observation strategy for their science case.

PSF subtraction in the ETC

See also ETC Coronagraphy Strategy

Currently for PSF subtraction, the ETC quantifies the shot noise in the wings of either the host or reference sources, as appropriate, at the position of the faint companion source. It estimates the PSF using one or multiple reference images, calling for the choices of reference target and PSF calibration method in the "strategy" tab. It is also here that the user can specify the "Scene rotation", which allows for rotation of the scene with respect to the axes of the MIRI FQPMs. When used in conjunction with the CVT, this feature allows the user to determine the optimal Aperture Position Angle Special Requirements for the scene of interest. 

 Because the PSF stability is unknown until launch, the ETC does not support ADI or SGD—only RDI.


While the JWST web-based ETC is perfectly adapted and trustworthy for the SNR calculations occurring in the background limited regime, typically at 1" separation and beyond, below 1" (in the speckle-limited regime) and down to the IWA of the coronagraph, the Pandeia-Coronagraphy extension of the ETC's Pandeia engine is recommended for more realistic contrast and detection limit calculations. The Python-based package also provides capability to introduce mis-registration of the PSF reference star and to experiment with small-grid dithers.

*Note that the IWA (in arcseconds) of the 10.65, 11.4, and 15.5 4QPMs are 0.33", 0.36", 0.49", respectively.


Implementation into APT programs

See also JWST High-Contrast Imaging in APTMIRI Coronagraphic Imaging APT Template and APT Coronagraphic Sequence Examples.

In the Astronomer's Proposal Tool a standard coronagraphic sequence involves a set of linked observations in a non-interruptible sequence. Coronagraphic programs with several observations bundled in a non-interruptible sequence can benefit largely from the APT Smart Accounting tool, which will aid in reducing the total slew times charged for such sequences. 

The APT MIRI Coronagraphic Imaging Template provided for coronagraphic imaging observations with MIRI provides sections for entering information on a variety of parameters including target acquisition, exposure times, special requirements for linking observations, observations of PSF-reference stars, and full-frame astrometric images, if needed.  Best practice is to always decide ahead on the observation strategy and observation sequence that will be used for a given pair or set of targets to be observed. The JWST High-Contrast Imaging in APT article provides helpful tips on how to organize observation folders and the observations to which they pertain. 

Setting the appropriate Special requirements (SRs) is a crucial step for MIRI coronagraphic observations. At the very least, users will need to place the "SEQUENCE NON-INTERRUPTIBLEtiming SR on their coronagraphic sequence. For observations with the 4QPMs, observers will likely require an enforced APERTURE PA RANGE on their observations to place the companion source of interest optimally with respect to the mask, especially to ensure they are not coincident with any of the 4QPM axes. The ideal orientation of a single companion on the mask is usually ~45° from the 4QPM axes (if available), however this can differ for targets with multiple companions. The CVT can be used to assess the available position angles.

In order to define a roll dither between science target observations, the appropriate APERTURE PA OFFSET SR must be placed on them. If a second sequence at a larger PA offset is needed, the APERTURE PA OFFSET SR must be set between the two sequences; and so forth. For programs requiring this level of attention to detail, observers should use the observing constraints determined from advanced work with the CVTThis will avoid the disappointment of later discovering that the needed angles are impossible.

The implementation of a multi-roll coronagraphic sequence is a non-trivial process and requires an understanding of the roll angles accessible by JWST for a given target. In light of this, we offer the following recommendation: assuming the user requires roll dither of angle α (where α < 14°) and a larger roll offset of angle θ (where θ > 15 °)

  1. Create a three observation, NON-INTERRUPTIBLE coronagraph sequence such that: Obs. 1 is of the science target, Obs. 2 is of the science target and Obs. 3 is of the reference star.
  2. Add a Special Requirement on the sequence such that Obs. 2 has an APERTURE PA OFFSET of ° with respect to Obs. 1
  3. Create a second three-observation, NON-INTERRUPTIBLE sequence such that: Obs. 4 is of the science target, Obs. 5 is of the science target and Obs. 6 is of the reference star.

  4. Add a Special Requirement on the second sequence such that Obs. 5 has a APERTURE PA OFFSET of +α° with respect to Obs. 4.

  5. Add an APERTURE PA OFFSET Special Requirement to Obs. 4, with respect to Obs. 1, with an allowed APERTURE PA RANGE of [θ°, 180−θ°]
  6. If the above is not schedulable, switch the APERTURE PA RANGE on Obs. 4 w.r.t. Obs.1 to [θ−180°,−θ°].

Figure 2. Illustration of allowed ranges of Obs 4 and 5 given the above algorithm

Obs 1 and 2 are shown as solid and dashed red lines, respectively, separated by +α°. Obs 4 (blue shaded region) is allowed to be scheduled from [θ°, 180−θ°] or [θ−180°,−θ°]. Obs 5 (black dashed region) is rotated by +α° with respect to Obs 4. If the user does not constrain the PA of Obs 1, the entire diagram would rotate correspondingly.

Last updated

Not published