Spatially-Clipped Self-Coherent Camera

• Spatially-Clipped Self-Coherent Camera (SCSCC) is a single-shot focal-plane wavefront sensor that uses a two-channel, differential imaging approach to rapidly sense wavefront aberrations.
• It employs a closer reference pinhole and a knife-edge beamsplitter to generate simultaneous fringed and unfringed images, enabling a linear, matrix-based phase reconstruction.
• Simulations demonstrate that SCSCC achieves ~50× deeper contrast in evolving speckle fields, meeting high-contrast criteria for exoplanet coronagraphy despite challenges like differential aberrations.

The Spatially-Clipped Self-Coherent Camera (SCSCC) is a single-shot focal-plane wavefront sensor proposed for high-order wavefront sensing and control in coronagraphic imaging, with NASA’s Habitable Worlds Observatory as its motivating application. It modifies the classical self-coherent camera by placing the reference pinhole much closer to the Lyot stop and by using a knife-edge beamsplitter to form simultaneous fringed and unfringed channels, so that the wavefront can be sensed from a single exposure. In monochromatic simulations, the SCSCC operating in a sensing and control loop reaches a mean normalized intensity of $\sim 4 \times 10^{-10}$ in a $5\text{–}20\,\lambda/D$ dark hole and achieves $\sim 50\times$ deeper contrast than pairwise probing in a temporally evolving speckle field (Liberman et al., 4 Sep 2025).

1. Scientific setting and lineage

Directly imaging an Earth-like exoplanet around a Sun-like star requires suppressing the host star’s light by a factor of order $10^{10}$ at separations of a few $\lambda/D$. In this regime, coronagraphs provide the first-order suppression, but tiny optical aberrations create quasi-static speckles that leak through the coronagraph and appear as planet-like signals. High-order wavefront sensing and control therefore uses a deformable mirror to minimize the complex electric field in a chosen dark hole of the science focal plane (Liberman et al., 4 Sep 2025).

Within focal-plane wavefront sensing, pairwise probing (PWP) and the self-coherent camera (SCC) are the two reference approaches emphasized in the SCSCC work. PWP plus Electric Field Conjugation is the baseline for the Roman Coronagraph, but it requires at least 4 images per estimation step, which slows the loop and makes it susceptible to time-varying speckles. Classical SCC instead creates a coherent reference beam with a small Lyot-plane pinhole and encodes the residual speckle field in Fizeau fringes, enabling single-exposure estimation in monochromatic light (Liberman et al., 4 Sep 2025).

Earlier SCC literature established the main trade-space that the SCSCC inherits. Laboratory SCC experiments reached contrast better than $3\times 10^{-8}$ in the $5\text{–}12\,\lambda/D$ region in monochromatic light and $4\times 10^{-8}$ in a 10 nm band (Mazoyer et al., 2014). The multireference SCC (MRSCC) addressed chromatic limitations by using several reference holes and demonstrated laboratory operation with an 80 nm bandwidth at 640 nm, reaching a contrast of $4.5\times 10^{-8}$ between 5 and $17\,\lambda/D$ (Delorme et al., 2016). The spectrally modulated SCC (SM-SCC) pursued the same throughput problem by spectral modulation, increasing the pinhole throughput by a factor of 32 and the wavefront sensor sensitivity by a factor of 5.7 (Haffert, 2021). In that lineage, the SCSCC is a two-channel, spatially differentiated SCC variant whose stated aim is to retain single-shot operation while improving throughput, bandwidth, and implementability (Liberman et al., 4 Sep 2025).

2. Optical architecture and the meaning of “spatially-clipped”

The classical SCC places a small reference hole outside the main Lyot pupil so that, after Fourier transformation of the focal-plane image, the interference sidebands are separated from the unmodulated central term. In the comparison used for the SCSCC study, the classical SCC pinhole is at $5\text{–}20\,\lambda/D$0, whereas the SCSCC pinhole is placed at $5\text{–}20\,\lambda/D$1 from the center. The SCSCC pinhole is therefore approximately a factor of three closer than in the classical comparison architecture (Liberman et al., 4 Sep 2025).

This change defines the “spatially-clipped” concept. The reference beam is taken from nearer the edge of the Lyot stop, effectively clipping the diffracted wavefront in a region of higher intensity and smaller spatial extent, rather than far out in the halo. Because the pinhole intercepts a brighter region of the diffracted light, the reference flux is much higher, photon-noise limits are reduced, and the beam footprint is smaller. The same design choice also boosts the sensor resolution and reduces optic size requirements (Liberman et al., 4 Sep 2025).

Downstream of the Lyot stop, the SCSCC adds a 50/50 beamsplitter and a knife-edge mask that create two channels. Channel 1 is fringed: it contains both the Lyot-pupil field and the pinhole reference, so the residual speckles interfere with the reference beam. Channel 2 is unfringed: a knife edge blocks the pinhole contribution, leaving only the Lyot-pupil field. The two images are recorded simultaneously on separate detectors or detector regions, giving single-shot access to a fringed image $5\text{–}20\,\lambda/D$2 and an unfringed image $5\text{–}20\,\lambda/D$3 (Liberman et al., 4 Sep 2025).

This two-channel arrangement is the critical distinction from the classical SCC. The SCSCC does not rely on temporal modulation, wavelength cycling, or polarization switching; instead, the interference signal is isolated by differencing two simultaneous channels that differ only by the presence of the reference beam. A plausible implication is that this architecture directly addresses the type of post-Lyot footprint constraint reported for SCC implementation on SCoOB, where post-Lyot optics were too small to allow the canonical large pinhole offset and sideband non-overlap (Derby et al., 2 Sep 2025).

3. Signal formation and reconstruction formalism

The paper defines a coronagraph operator $5\text{–}20\,\lambda/D$4 and decomposes the field into a speckle component $5\text{–}20\,\lambda/D$5 and a pinhole-filtered reference component $5\text{–}20\,\lambda/D$6. After propagation through the coronagraph, these become $5\text{–}20\,\lambda/D$7 and $5\text{–}20\,\lambda/D$8. The two detector channels are then written as

$5\text{–}20\,\lambda/D$9

and

$\sim 50\times$0

Their difference is

$\sim 50\times$1

The key approximation is

$\sim 50\times$2

so that

$\sim 50\times$3

If $\sim 50\times$4 is known from calibration and varies slowly and smoothly, this becomes a linear measurement of the real and imaginary parts of $\sim 50\times$5 at each pixel (Liberman et al., 4 Sep 2025).

The SCSCC therefore abandons the classical SCC’s explicit Fourier demodulation of fringes and instead uses a linearized, matrix-based estimator similar to the Fast Atmospheric SCC and implicit EFC methods. Writing $\sim 50\times$6 and $\sim 50\times$7, the difference image satisfies

$\sim 50\times$8

which is the paper’s linear measurement relation rewritten in component form (Liberman et al., 4 Sep 2025, Gerard et al., 2018).

Rather than solving independently for $\sim 50\times$9 at each pixel, the method calibrates the effect of deformable-mirror Fourier modes. For each mode $10^{10}$0, the resulting focal-plane field change $10^{10}$1 is computed through the SCC stop, and the corresponding channel-difference image is recorded. This produces an interaction matrix $10^{10}$2 such that

$10^{10}$3

with a factorization

$10^{10}$4

where $10^{10}$5 maps modal coefficients $10^{10}$6 to a DM phase pattern $10^{10}$7, and $10^{10}$8 maps phase to intensity modulation through the SCSCC optics. A Tikhonov-regularized pseudo-inverse then yields

$10^{10}$9

and the reconstructed phase is

$\lambda/D$0

In the simulations, $\lambda/D$1 is built numerically with HCIPy, the main reconstruction uses a regularization strength of $\lambda/D$2 relative to the maximum singular value, and all images are normalized to unit power before forming the matrix (Liberman et al., 4 Sep 2025).

4. Control law, dark hole, and simulation model

Once calibrated, the SCSCC is embedded in a closed-loop wavefront control process. The loop acquires the simultaneous fringed and unfringed images, computes $\lambda/D$3, reconstructs the phase or DM modal coefficients, and updates the deformable mirror to reduce the speckle intensity in the target dark hole. The paper writes the control problem as

$\lambda/D$4

which is an Electric Field Conjugation-like control law in the SCSCC formalism (Liberman et al., 4 Sep 2025).

The simulations dig a D-shaped dark hole spanning

$\lambda/D$5

Only pixels in that region are used in the cost function. The one-sided geometry arises from frequency-folding effects on the DM, consistent with one-sided dark-hole operation elsewhere in SCC work (Liberman et al., 4 Sep 2025, Mazoyer et al., 2014).

The numerical model is monochromatic and uses first-order physical optics in High Contrast Imaging for Python. The optical train is a circular entrance pupil of diameter $\lambda/D$6, a charge-4 scalar vortex coronagraph, and a Lyot stop undersized to $\lambda/D$7. The SCC pinhole has radial distance $\lambda/D$8 and diameter $\lambda/D$9. The deformable mirror is a 52 × 52 actuator Boston Micromachines DM, and the interaction matrix is built from Fourier modes with amplitude $3\times 10^{-8}$0. Calibration wavelength is $3\times 10^{-8}$1 nm (Liberman et al., 4 Sep 2025).

Time-varying speckles are represented by spatio-temporal power spectral density cubes on a 64×64 spatial grid over 10 s, with separable form

$3\times 10^{-8}$2

The spatial component is

$3\times 10^{-8}$3

and the temporal component is

$3\times 10^{-8}$4

For the SCSCC versus PWP comparison, the parameters are $3\times 10^{-8}$5, $3\times 10^{-8}$6, $3\times 10^{-8}$7, and peak-to-valley wavefront error amplitude $3\times 10^{-8}$8 (Liberman et al., 4 Sep 2025).

5. Performance, advantages, and limitations

In a static speckle field, SCSCC plus EFC reaches a mean normalized intensity of

$3\times 10^{-8}$9

in the $5\text{–}12\,\lambda/D$0 dark hole. The paper states that this meets HWO-like contrast requirements in monochromatic light (Liberman et al., 4 Sep 2025).

Against pairwise probing in temporally evolving speckles, the decisive result is tied to simultaneity. For long speckle lifetimes and an effective integration time of 0.1 s per frame, both SCSCC and PWP achieve mean normalized intensity $5\text{–}12\,\lambda/D$1. For short lifetimes, however, the SCSCC outperforms PWP by about 1.7 orders of magnitude, corresponding to $5\text{–}12\,\lambda/D$2 deeper contrast. The reason given is direct: PWP uses multiple non-simultaneous probe images, so the field evolves between probe states, whereas the SCSCC freezes the field in a single exposure (Liberman et al., 4 Sep 2025).

Relative to the classical SCC, the SCSCC improves sensitivity in the photon-noise-limited reconstruction tests used in the paper’s pinhole-position analysis. Both classical SCC and SCSCC recover wavefronts to within $5\text{–}12\,\lambda/D$3 nm RMS in the high-flux regime, but the SCSCC shows roughly a factor $5\text{–}12\,\lambda/D$4 improvement in sensitivity. The authors note that Haffert (2022) found $5\text{–}12\,\lambda/D$5 improvement for a spectrally modulated SCC with a closer pinhole, which they interpret as evidence that the SCSCC calibration is not yet fully optimized (Liberman et al., 4 Sep 2025, Haffert, 2021).

The architecture also changes the bandwidth discussion. Classical SCC is bandwidth-limited because the fringe pattern decorrelates across wavelength. By placing the pinhole about three times closer to the Lyot pupil, the SCSCC architecture provides an estimated $5\text{–}12\,\lambda/D$6 bandwidth improvement relative to classical SCC, but the present results are monochromatic and broadband performance is left to future work, including multi-pinhole SCC masks (Liberman et al., 4 Sep 2025). This situates the SCSCC alongside earlier achromatization strategies such as the MRSCC, which used multiple references rather than a two-channel spatial split (Delorme et al., 2016).

Its main limitations are equally explicit. First, the SCSCC is a non-common-path sensor: the 50/50 beamsplitter creates separate fringed and unfringed optical paths, so differential aberrations between channels can bias the reconstruction. In the reported tests, differential aberrations $5\text{–}12\,\lambda/D$7 nm RMS still allow $5\text{–}12\,\lambda/D$8 contrast, while aberrations $5\text{–}12\,\lambda/D$9 nm RMS cause the loop to become unstable and diverge (Liberman et al., 4 Sep 2025). Second, the beamsplitter imposes a factor-of-two throughput penalty per channel, although the higher pinhole throughput partly compensates. Third, the method requires a calibrated interaction matrix that remains valid despite drifts. A common misunderstanding is to treat the SCSCC as simply a compact classical SCC; in fact, once the pinhole is brought close to the pupil, the interference pattern is no longer easily separable in Fourier space, and the two-channel differential estimator becomes essential (Liberman et al., 4 Sep 2025).

6. Relevance to observatories and prospective development

The SCSCC is framed as a wavefront sensor concept for the Habitable Worlds Observatory because it combines measurement-based high-order wavefront sensing and control with single-shot sensing. In the authors’ formulation, time-varying aberrations are effectively frozen in place, making them easy to remove. This is particularly relevant to realistic thermal and structural drift environments in which speckle lifetimes are seconds to tens of seconds (Liberman et al., 4 Sep 2025).

The development path identified in the paper is concrete. The next steps are to design and fabricate SCSCC masks for the CACTI testbed, use a Hamamatsu ORCA-Quest2 qCMOS camera for high-speed, low-noise imaging, implement multi-pinhole SCC masks for broadband capability, and ultimately install the SCSCC on MagAO-X, with first light planned for Fall 2026 (Liberman et al., 4 Sep 2025).

This roadmap places the SCSCC within a broader transition from laboratory SCC demonstrations to operational instruments. SCC-based correction has already been demonstrated on the sky at Palomar, where SCC calibration improved the on-sky contrast by a factor of 5 between 2 and $4\times 10^{-8}$0 (Galicher et al., 2019). SCC dark-hole digging has also been demonstrated on SCoOB, where a self-coherent camera reached a final mean contrast of $4\times 10^{-8}$1 in a 3–10 $4\times 10^{-8}$2 half-annulus dark hole, although that implementation was limited by slow pinhole modulation and pointing drifts (Derby et al., 2 Sep 2025). This suggests that the SCSCC is best understood not as a replacement for the SCC family’s prior variants, but as a compact, high-throughput, two-channel SCC architecture aimed at the regime where rapid correction of evolving speckles is operationally decisive.