MIRI Features and Caveats

The JWST MIRI instrument has a number of features and characteristics that observers should be aware of when planning, reducing, analyzing and interpreting observations. This article provides an overview of several features and caveats that should be noted and that may affect the data and resulting science.

On this page

Detector general features and caveats

Cruciform artifact

There is a feature caused by internal scattering within the MIRI detectors that manifests as a cruciform shape in imaging and spectroscopy at wavelengths ≤ 10 μm.

The Si:As detectors used by MIRI are tasked with observing a large wavelength range, from 5.6 to 30 μm. Inherently, these detectors are lower in quantum efficiency at the shorter wavelength range of this window, absorbing less than 27% of the photons on their first pass through the absorptive layer at 5.6 μm. The remaining photons are diffracted by the square pixel lattice, resulting in a classic cruciform-shaped diffraction pattern that follows the row and column directions of the detector.

Since the MIRI imager (Field of View) FOV is rotated by 4.835 degrees from the V3 axis, the cruciform artifact is also rotated by this amount from the OTE/hexagon diffraction spikes. The cruciform artifact is strongest at the shorter wavelengths and disappears mostly by 12 μm (see Figure 1 below. The inner region of the PSF core does not show the cruciform artifact, due to photons being able to escape the detector wafer within the total internal reflection limit. For more information, please see Gaspar et al. (2021).

Figure 1. An example of the cruciform artifact observed in the MIRI Imager during commissioning

The cruciform artifact is easily identified in the flight images for wavelengths shorter than ~12 µm. 

Hot or dead pixels

The MIRI imager detector has several pixels that are flagged based on ground testing and in-orbit observations. These pixels are flagged as DO_NOT_USE in the Flight Bad Pixel MASK (Figure 2), and the pipeline will not use those pixels. 

Figure 2. MIRI Imager bad/hot/dead pixel mask

Imager flat field bad pixel mask (black). The vertical line is the mask for the shorted columns. Note that bad pixels are flagged DO_NOT_USE, while the shorted columns are flagged as UNRELIABLE_SLOPE.

Shorted columns

Two of the columns (385 & 386) on the MIRI imager array are shorted. These are flagged in the flat field bad pixel mask as UNRELIABLE_SLOPE (see Figure 2 above).

Residual dark features 

Data are still being evaluated, and updates will be provided soon.

Brighter-fatter correction

Saturated pixels affect the flux of neighboring pixels and 'fatten' the PSF. Correction for this effect is still under development.

Column/row pull up/pull down

Data are still being evaluated, and updates will be provided soon.

Residual cosmic ray artifacts (large & weird)

Data are still being evaluated, and updates will be provided soon.


See also: MIRI Imaging

Imager knife edge

The MIRI imager has a protrusion from a knife edge structure that blocks a small portion of the field of view (FOV) on the left side, and is used for optical testing. The main effect will be on mosaics (see Knife Edge Gap in MIRI Imaging Mosaics). The default MIRI imaging mosaic overlap of 10% is used to address the gap.

Bonus Lyot FOV in imaging

Because the Lyot coronagraph has no additional optics, the Lyot FOV provides valid, calibrated data except where the Lyot occulting spot and support structure prevent light from reaching the detector.

Edge brightening

The inflight imaging FOV exhibits some stray light features on the left and bottom edges. These should not have a significant impact on combined dithered or mosaicked imaging. Additional analysis is underway.

Figure 3. Edge brightening in the MIRI Imager field of view, and also showing the scattered light features in the MIRI coronagraphs (see below)

In orbit engineering F1500W image of a sparse astronomical field. Brightening around the lower edge of the imaging FOV and the knife edge are visible. Scattered light features in the coronagraphs are also evident.


See also: MIRI Coronagraphic Imaging

Why are the coronagraph masks at an angle in the focal plane?

The support structure for the Lyot coronagraph, and the quadrant boundaries for the 4QPM coronagraphs, are tilted at 4.835 degrees relative to the rows and columns of the imaging detector (see MIRI Coronagraphic Imaging). Since MIRI is tilted by this same amount in the telescope focal plane, the tilt in the coronagraphs is necessary to align the coronagraph masks with the pupil structures in the OTE.

TA PSF asymmetry in 4QPM coronagraphs

The MIRI coronagraphic imaging target acquisition (TA) process uses one of 4 broadband filters: F560W, F1000W, F1500W, and the neutral density filter (FND). Inflight data shows that the point spread function (PSF) exhibits an asymmetry caused by internal reflection in the 4QPM germanium optics. This asymmetry is a function of the difference between the central wavelengths of the TA filter and the operating wavelength of the specific 4QPM. The most extreme asymmetry is for the F560W TA filter and the F1550C coronagraph. Tests of the TA process have shown that this asymmetry is uniform across the coronagraph FOVs—it only imparts a small offset to the TA which has been compensated for in the TA process. Because there are no optics at the Lyot coronagraph occulting spot and support structure, the TA PSFs for the Lyot coronagraph are unaffected.

Glow sticks in the MIRI 4QPM coronagraphs

Fight data has shown that there is light being scattered into the coronagraphs. This is particularly apparent for the three 4QPMs, where it manifests as a glow along the horizontal boundaries of the phase masks. These are referred to as glow sticks; they can be removed by (1) a dedicated background observation, (2) angular differential Imaging (ADI) processing where a telescope roll has been used for 2 or more observations of the science target, or (3) by reference star PSF subtraction if the reference star has the same brightness and exposure time as the science target. At present, these glow sticks appear to mostly increase the shot noise.

Figure 4. Scattered light in the 4QPM coronagraphs

Comparison between the ground and flight images of the F1550C coronagraph when uniformly illuminated. The glow sticks feature is clearly visible, and "hides" the attenuation of the 4QPM phase mask along the horizontal axis. This feature is seen in all three 4QPMs.

Edge brightening around the Lyot spot and lower edge of the Lyot FOV

Inflight data show that, as for the imager, there is light scattered into the Lyot FOV that slightly illuminates the rich-hand edge of the occulting spot and the support structure. This scattered light can be removed by either (1) a dedicated background observation, (2) ADI processing where a telescope roll has been used for 2 or more observations of the science target, or (3) by reference star PSF subtraction if the reference star has the same brightness and exposure time as the science target.

Figure 5. Scattered light in the Lyot coronagraph

Comparison between the ground and flight images for the Lyot coronagraph when uniformly illuminated. The scattered light is evident along the upper edge of the Lyot spot, and the lower edge of the corornagraph FOV.

Low-resolution spectrometer (LRS)

Scattered light in the LRS

See also: MIRI Low Resolution Spectroscopy

If a very bright target is present in the imager field, its light may scatter into the region where the LRS slit is dispersed. We see scattered light artifacts in the detector substrate around bright sources (see the cruciform artifact), so these features can persist despite the presence of a focal plane mask. As the pipeline extracts a cutout region from the array in the later stages of the stage 2 pipeline (the photom and resample_spec steps), we recommend that users inspect Level 2a data (the "rate.fits" file) to check for the presence of such contaminating sources.

Figure 6. Light scattered from the imager field of view into the LRS spectral location


Image of a bright source in the Imager FOV during an LRS observation. Light from the bright source is dispersed and scatters into the LRS slit spectrum. (a) shows the entire Imager FOV, with the slit marked in green, the TA ROI in yellow, and the zoomed region in dashed white. (b) a zoomed region from the full FOV. (c) the level s2d pipeline product, which shows both LRS along-slit nods subtracted (resulting in the negative and positive spectral images). As mentioned above, the pipeline does not automatically remove this scattered light contamination.

LRS slitless spectral foldover and leak

As both slit and slitless modes use the same dispersing element, the dispersion profile is, in principle, the same for both. The nominal spectral range of 5–12 µm is dispersed over ~370 pixels. The dispersion profile however folds over below 4.5 µm (where the prism throughput is very low), superimposing 2 parts of the spectrum onto each other. A dedicated filter is mounted over the slit to block radiation shortward of 4.5 µm to avoid this contamination in the slit. This effect is not mitigated for LRS in slitless mode, causing some spectral contamination at the shortest wavelengths. 

Medium resolution spectrometer (MRS)

See also: MIRI Medium Resolution SpectroscopyJWST MIRI MRS Pipeline Caveats

Cosmic ray showers and column striping

Cosmic ray showers visible in the MRS affect large regions of the detector as seen in Figure 7. The cosmic ray showers are found in both the SHORT and LONG detectors of MRS. These showers are not removed by the pipeline since the jump in the ramps introduced by these showers is too small to be detected by the cosmic ray id routines. The brightest showers can produce latents visible in the next integration. Cosmic ray showers can be mitigated by dithering and taking at least 3 integrations, while  5 or more integrations will provide better results.  

Figure 7. Cosmic ray showers and column striping

MIRI MRS observations of a dedicated sky field. Large cosmic ray showers are indicated with red arrows on two different integrations and are seen in both the SHORT and LONG detectors. Bright column striping is also visible in the SHORT detector integrations.
Column stripping from dark subtraction is most noticeable in the SHORT detector. It changes over time due to drift in the detector reset level. The column striping is the dominate noise feature left after dark subtraction at short wavelengths. There is currently no correction in the pipeline for column stripping, however it can be well corrected in science data on a per-column basis using dedicated background observations. 

Scattered light in MRS

The scattered light is an integral part of the MRS PSF, which is a telescope diffraction-limited PSF, convolved by the MRS detector response function to incoming illumination. The origin of the measured scattered light in the MRS is caused by photon scattering inside of the MRS detector substrate. Specifically, the scattering occurs between the detector pixel metalization and the anti-reflection coating on the back surface of the detector (the MRS detectors are backside illuminated).
Photon scattering in the MRS detectors manifests in the data in two ways:

  1. A broader PSF FWHM compared to diffraction limited PSF predictions
  2. A secondary diffraction at narrow gaps between the pixels, which act as narrow slits in the mid-infrared. This superimposes a traditional Airy diffraction pattern on the detector that is centered on the PSF.

Because neighboring slices on the detector do not correlate to neighboring pixels on the sky, the wings of the Airy diffraction pattern will appear in the MRS 3D-reconstructed spectral cubes as faint emission away from the observed point source. Additionally, neighboring slices on the detector are offset in wavelength. This means that the wings of the diffraction pattern produced at one wavelength will result in a faint emission line at a different wavelength (and location on the sky). This spatial scattered light is corrected for in the pipeline (see MIRI MRS Pipeline Caveats)

Figure 8. Spatial Scattered Light from the MRS PSF

Left: 2-D detector view of MRS scattered light. Right: In 3-D cubes the scattered light manifests as horizontal bars across the spatial cube.
A faint scattered component has been observed in the spectral direction of the same slice (detector vertical direction). This component is difficult to disentangle from the spectral continuum of a source, dispersed in the same direction. The scattered component in the vertical direction could be incoherent and manifest as a slightly higher background. At present, this component is still under investigation and no correction is currently available.
Figure 9. Spectral Scattered Light from the MRS PSF

The MRS PSF also extends in the vertical (spectral) direction around a bright source. This scattered light is marked as a red boxes around a bright line in the spectrum.

MRS fringes - description and correction

Like most IR spectrometers, MRS data contain fringes: large-amplitude sensitivity modulations with wavelength, caused by standing waves in the detector and, possibly, in the dichroics. Fringes change across the MRS wavelength range.  Amplitudes are highest at low wavelengths (up to ~40% peak to peak).  From channel 2C onward, fringe amplitude varies across the band due to beating. On the MIRIFULONG detector, we see an additional low-amplitude
high-frequency modulation.  It's tentatively attributed to fringing within the dichroics.

Two pipeline steps plus a post-pipeline notebook are available to correct for fringes.  Taken together, they reduce fringe contrast to safely below 6%.  More sophisticated calibration tools are under development. The current pipeline only runs the first fringe-correction step. Most high-level MRS data products on MAST will therefore still show appreciable fringes.  Users are encouraged to run ResidualFringeStep from the pipeline locally (see MIRI MRS Pipeline Caveats). 


Gaspar, A. et al. 2021, PASP, 133, 1019, 15
The Quantum Efficiency and Diffractive Image Artifacts of Si:As IBC mid-IR Detector Arrays at 5-10 μm: Implications for the JWST/MIRI Detectors

Latest updates
Originally published