A Multi-wavelength Differential Imaging Experiment for the High Contrast Imaging Testbed

We discuss the results of a multi-wavelength differential imaging lab experiment with the High Contrast Imaging Testbed (HCIT) at the Jet Propulsion Laboratory. The HCIT combines a Lyot coronagraph with a Xinetics deformable mirror in a vacuum environment to simulate a space telescope in order to test technologies and algorithms for a future exoplanet coronagraph mission. At present, ground based telescopes have achieved significant attenuation of speckle noise using the technique of spectral differential imaging (SDI). We test whether ground-based SDI can be generalized to a non-simultaneous spectral differential imaging technique (NSDI) for a space mission. In our lab experiment, a series of 5 filter images centered around the O2(A) absorption feature at 0.762 um were acquired at nominal contrast values of 10^-6, 10^-7, 10^-8, and 10^-9. Outside the dark hole, single differences of images improve contrast by a factor of ~6. Inside the dark hole, we found significant speckle chromatism as a function of wavelength offset from the nulling wavelength, leading to a contrast degradation by a factor of 7.2 across the entire ~80 nm bandwidth. This effect likely stems from the chromatic behavior of the current occulter. New, less chromatic occulters are currently in development; we expect that these new occulters will resolve the speckle chromatism issue.

💡 Research Summary

The paper presents a laboratory demonstration of non‑simultaneous spectral differential imaging (NSDI) using the High Contrast Imaging Testbed (HCIT) at the Jet Propulsion Laboratory, with the goal of assessing whether the spectral differential imaging (SDI) techniques that have proven successful on ground‑based telescopes can be transferred to a future space‑based exoplanet coronagraph mission. The HCIT reproduces a space‑telescope environment in vacuum, combining a Lyot coronagraph with a Xinetics deformable mirror (DM) that can sculpt a “dark hole” – a region of the focal plane where speckle noise is actively suppressed to reach contrasts as low as 10⁻⁹.

The experiment focused on the O₂(A) absorption band at 0.762 µm, a wavelength region of interest for potential biosignature detection. Five narrow‑band filters, spaced roughly 16 nm apart and covering an ~80 nm total bandwidth, were used to acquire sequential images at four nominal contrast levels (10⁻⁶, 10⁻⁷, 10⁻⁸, and 10⁻⁹). Because the images are taken sequentially rather than simultaneously, the study explicitly tests the NSDI concept: after acquisition, each pair of images at different wavelengths is subtracted (single‑difference) to suppress speckles that are common to both wavelengths.

Results outside the dark hole are encouraging. Single‑difference processing improves contrast by a factor of about six across all tested contrast regimes, indicating that speckle structures remain sufficiently correlated across the modest wavelength offsets when the coronagraph is not operating at its deepest suppression. This mirrors the performance of ground‑based SDI, where simultaneous dual‑band imaging can achieve similar gains.

Inside the dark hole, however, the situation is markedly different. The authors observe a strong wavelength‑dependent evolution of speckles – termed “speckle chromatism.” As the wavelength moves away from the nulling wavelength (the wavelength at which the coronagraphic mask is optimized), both the phase and amplitude of residual speckles drift, leading to incomplete cancellation when images are differenced. Quantitatively, the contrast degrades by a factor of 7.2 across the full 80 nm band, erasing most of the benefit that NSDI would otherwise provide. The authors attribute this degradation primarily to the chromatic behavior of the current occulter (the focal‑plane mask). The mask’s transmission and phase response vary with wavelength, causing the dark‑hole region to shift slightly in the focal plane for each filter. Consequently, the speckle pattern that the DM has been tuned to cancel at one wavelength is no longer perfectly matched at another, and subtraction leaves a residual speckle halo.

The paper emphasizes that this chromatic limitation is a hardware issue rather than a fundamental flaw in the NSDI concept. Ongoing development of new occulters with reduced chromaticity—such as multi‑layer metasurface designs, high‑precision nano‑patterned masks, or broadband anti‑reflective coatings—is expected to mitigate speckle chromatism. The authors anticipate that once the mask’s wavelength dependence is suppressed, NSDI could achieve dark‑hole contrast improvements comparable to those seen outside the dark hole, making it a viable strategy for space‑based missions where simultaneous multi‑band imaging is impractical due to telemetry constraints and limited detector resources.

In addition to hardware upgrades, the authors suggest algorithmic counter‑measures: pre‑calibrated wavelength‑dependent DM command maps, model‑based speckle prediction across the band, and iterative correction loops that adjust the DM after each wavelength exposure. Such techniques could further reduce residual chromatic speckles and make NSDI robust against the inevitable small drifts that remain even with improved masks.

In summary, the study demonstrates that NSDI can deliver significant speckle suppression in regions where the coronagraph is not operating at its deepest null, but that inside the high‑contrast dark hole the current occulter’s chromatic response limits performance. The findings point to a clear development path: design and test low‑chromatic occulters, integrate wavelength‑aware DM control, and refine post‑processing algorithms. Achieving these goals will enable future space coronagraphs to exploit NSDI for biosignature‑targeted observations without the need for complex simultaneous multi‑band hardware, thereby simplifying instrument design and reducing mission risk.