[Paper Review] Spectroscopic time series performance of the Mid-Infrared Instrument on the JWST

J. Bouwman, Sarah Kendrew+22 more|arXiv (Cornell University)|Nov 29, 2022

Stellar, planetary, and galactic studies5 citations

TL;DR

This paper presents the first mid-infrared spectroscopic time series observation of the transiting exoplanet L 168-9 b using the James Webb Space Telescope's Mid-Infrared Instrument (MIRI) in Low Resolution Spectroscopy (LRS) mode. Achieving ~50 ppm spectro-photometric precision at R=50 over 7–8 µm, the study demonstrates MIRI’s capability for high-precision exoplanet atmosphere characterization, with measured transit depths consistent with known values within 1σ despite excess noise at shorter wavelengths due to detector effects like RSCD and scattering.

ABSTRACT

We present here the first ever mid-infrared spectroscopic time series observation of the transiting exoplanet \object{L 168-9 b} with the Mid-Infrared Instrument (MIRI) on the James Webb Space Telescope. The data were obtained as part of the MIRI commissioning activities, to characterize the performance of the Low Resolution Spectroscopy (LRS) mode for these challenging observations. To assess the MIRI LRS performance, we performed two independent analyses of the data. We find that with a single transit observation we reached a spectro-photometric precision of $\sim$50 ppm in the 7-8 \micron range at R=50, consistent with $\sim$25 ppm systematic noise. The derived band averaged transit depth is 524 $\pm$ 15 ppm and 547 $\pm$ 13 ppm for the two applied analysis methods, respectively, recovering the known transit depth to within 1 $σ$. The measured noise in the planet's transmission spectrum is approximately 15-20 \% higher than random noise simulations over wavelengths $6.8 \lesssim λ\lesssim 11$ $μ$m. \added{We observed an larger excess noise at the shortest wavelengths of up to a factor of two, for which possible causes are discussed.} This performance was achieved with limited in-flight calibration data, demonstrating the future potential of MIRI for the characterization of exoplanet atmospheres.

Motivation & Objective

• To assess the performance of the MIRI Low Resolution Spectroscopy (LRS) mode for spectroscopic time series observations on JWST.
• To evaluate the spectro-photometric precision achievable for transiting exoplanet atmospheres in the mid-infrared.
• To identify and characterize systematic noise sources affecting high-precision time series data, particularly at short wavelengths.
• To validate the reliability of independent analysis pipelines (CASCADe and Eureka!) for exoplanet transmission spectroscopy using real JWST data.

Proposed method

• Conducted a single transit observation of L 168-9 b using MIRI LRS mode during JWST commissioning.
• Performed independent data analysis using two pipelines: CASCADe and Eureka! for spectral extraction and light curve modeling.
• Applied time series processing with the STScI JWST calibration pipeline, treating each integration as a separate unit to preserve temporal resolution.
• Used white light curves and transmission spectra to derive transit depths and assess noise levels across 6.8–11 µm.
• Compared measured noise levels with random-noise simulations to quantify excess systematics.
• Investigated potential causes of excess noise, including Reset Switch Charge Decay (RSCD), detector scattering, and spectral extraction method differences.

Experimental results

Research questions

• RQ1 What is the spectro-photometric precision achievable with MIRI LRS mode for a single transit observation of an exoplanet in the mid-infrared?
• RQ2 How do systematic noise sources such as RSCD and detector scattering affect high-precision time series data, especially at short wavelengths?
• RQ3 To what extent do different spectral extraction methods (fixed aperture vs. PSF-weighted) influence noise estimates in low-signal regimes?
• RQ4 Can independent analysis pipelines (CASCADe and Eureka!) produce consistent results for exoplanet transmission spectra from JWST MIRI data?
• RQ5 How does the measured noise in the transmission spectrum compare to predictions from random-noise simulations?

Key findings

• The spectro-photometric precision reached ~50 ppm in the 7–8 µm range at R=50, consistent with ~25 ppm systematic noise, demonstrating high stability for mid-infrared time series observations.
• The derived band-averaged transit depth was 524 ± 15 ppm (CASCADe) and 547 ± 13 ppm (Eureka!), both consistent with the known value within 1σ.
• Measured noise in the transmission spectrum was approximately 15–20% higher than random-noise simulations across 6.8 ≲λ ≲11 µm.
• Excess noise at wavelengths shorter than ~7 µm was up to a factor of two higher than predicted, likely due to RSCD and multiple scattering paths in the MIRI detector.
• Discrepancies at longer wavelengths (λ ≳11 µm) between the two analysis methods may stem from differences in spectral extraction—fixed aperture vs. PSF-weighted methods.
• The initial 30 minutes of the time series showed a 0.25% signal drift that decayed with a time constant of ~20 minutes, which was mitigated via exponential fitting to improve light curve accuracy.