A Characterization of JWST MIRI Detector Persistence and Implications for High-Contrast Imaging

The JWST MIRI detector exhibits a flux deficit persistence, but its timescales and impacts remain largely uncharacterized, particularly at the longest imaging wavelengths. In this study, we analyze full-field MIRI imager observations at 21 $μ$m (F2100W) to quantify detector persistence following a saturation event by a bright (K = 5.65 mag) nearby (8.12 $\pm$ 0.04 pc) mid M-dwarf star, IRAS 21500+5903. Unlike typical persistence that appears as excess flux, this effect presents as a flux deficit in pixels previously illuminated by the saturating or near saturating source. We measure persistence at two post-saturation epochs: shortly after saturation (11.6 minutes) and an hour later (1.39 hours). Immediately after the saturation event, we detect a persistence level of $1.69 \pm 0.10$%. By fitting a Bayesian exponential decay model to the two epochs, we estimate that persistence decreases to one-tenth of its initial value after $5.16^{+1.49}_{-0.94}$ hours. We examine the implications of persistence for MIRI high-contrast imaging using the imager (not coronagraphy). Specifically, we discuss how MIRI detector persistence can produce false-positive exoplanet signals in direct imaging surveys, as well as degrade PSF subtraction, particularly at small inner working angles. We also outline mitigation strategies to avoid these impacts in future observations.

💡 Research Summary

The Mid‑Infrared Instrument (MIRI) on the James Webb Space Telescope (JWST) is a powerful tool for imaging at wavelengths up to 21 µm, but like all infrared detectors it suffers from a persistence artifact. In most infrared arrays persistence appears as an excess signal left behind by bright sources; MIRI, however, exhibits a less common “flux‑deficit” persistence where saturated pixels become temporarily darker. The physical origin of this negative offset is not yet fully understood, but it is linked to hard saturation events.

This paper quantifies that effect in the longest MIRI imaging band (F2100W, λ ≈ 21 µm) by exploiting a serendipitous saturation of a nearby M4 V dwarf (IRAS 21500+5903, K = 5.65 mag, distance ≈ 8.12 pc). The star saturated the detector during an initial 7.7 µm (F770W) exposure and left a residual imprint that was captured in subsequent F2100W images taken 11.6 minutes and 1.39 hours later. The authors measured the flux deficit by defining a 5‑pixel‑diameter region (R1) centred on the saturated spot and two adjacent background regions (R2, R3). The fractional deficit p = 1 − F_R1/F_(R2+R3) was 1.69 % ± 0.10 % at 11.6 min and 0.99 % ± 0.10 % at 1.39 h.

Using these two points, a Bayesian Markov‑Chain Monte Carlo (MCMC) fit of an exponential decay model was performed:
R(t) = (1.85 ± 0.13) · exp