{"id":"2154ef5e-47a1-4f69-bdc8-2d45c1036e1e","arxiv_id":"2608.02962","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"low","formal_verification":"none","parameter_count":4,"one_line_summary":"Simulations and preliminary bench data show LDFC and EKF dark zone maintenance can stabilize vortex coronagraph contrast on the SCoOB testbed, with LDFC failing on low-order testbed drift.","lead":"This paper tests two algorithms that keep a coronagraph's dark hole stable during long exoplanet observations. Simulations on the SCoOB testbed model and early bench results show both can hold contrast, though the bench demo reveals a known weakness with low-order drifts.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Simulations omit detector noise, so the quantitative claim that LDFC and EKF maintain contrast near the iEFC level is not yet established for realistic SCoOB photon/read-noise levels.","rationale":"The reader weakened the result over POPPY model fidelity and the absence of code/data; my concern is more specific and directly load-bearing for the quantitative claim: the simulations are noiseless, and both DZM algorithms operate on intensity measurements whose useful signal is small exactly in the maintained-dark-hole regime. Because the authors transparently defer detector noise to future work, the paper is not unsound, but the central quantitative conclusion should be read as conditional on noiseless simulations. This matches the reader's CONDITIONAL verdict, so I recommend no change to the verdict.","tokens_in":8169,"tokens_out":13013,"duration_ms":125542,"concrete_test":"Add measured SCoOB science-camera read noise and photon noise, at the exposure and throughput used for the 3e-9 contrast result, to the Section 3 LDFC and EKF simulations; rerun the cases in Figures 2, 3, 5, and 6 with the same Q, R, and dither settings, and check whether the maintained contrast curves remain within a factor of two of the noiseless iEFC curve. Also require the authors to state the dither amplitude and R matrix used for EKF.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3 explicitly says 'We have not considered detector noise for these simulations.' The central quantitative claim — that LDFC and EKF maintain dark-hole contrast near the final iEFC level — is supported only by these noiseless simulations. For EKF the measurement is the dark-hole intensity image, and in closed-loop maintenance the residual field is deliberately driven to a very small value, so the signal available to the filter is tiny; without simulated photon/read noise, the estimator can exploit intensity changes that a real camera cannot detect. For LDFC the signal is a small intensity change in the bright field; photon noise on that bright background sets a detection floor that the noiseless calibration curves (Figures 1 and 7) do not address. The paper also does not state the measurement-noise covariance R or the EKF dither amplitude, both of which determine how much of the performance is attributable to the filter rather than to the noiseless idealization. The testbed LDFC results are preliminary, lack error bars, and already show an uncorrected low-order drift, so they do not close this gap. The concern is not that the algorithms are invalid; it is that the quantitative comparison to the iEFC contrast is underdetermined until a noise-inclusive simulation or an error-barred testbed run is reported.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript (SCoOB X) reports simulations and preliminary testbed demonstrations of two dark-zone maintenance (DZM) algorithms for the SCoOB vortex coronagraph: model-free linear dark field control (LDFC) and an extended Kalman filter (EKF) estimator/controller. In POPPY simulations, both algorithms are claimed to recover and hold dark-hole contrast near the final iEFC contrast level under static and dynamic wavefront errors, including a single DM eigenmode, random-walk actuator drift, and accumulating low-order Zernike modes. On the testbed, LDFC is shown to correct an injected DM eigenmode and a sum of low-order DM eigenmodes, while it fails to correct the bench's own low-order drift. The paper honestly states that detector noise is omitted from simulations and that the testbed LDFC results are preliminary.","tokens_in":8429,"tokens_out":3819,"duration_ms":35228,"significance":"The work addresses an important operational problem for space coronagraphs: maintaining high contrast during long exposures without re-digging a dark hole. Its strengths include explicit step-by-step algorithm descriptions with matrix dimensions, a public code reference, and an honest report of LDFC's failure on low-order drift, which is a useful falsifiable observation. The comparison of LDFC and EKF on the same simulated SCoOB model is also valuable. However, the quantitative headline claim—that both algorithms maintain contrast at the final iEFC level—is supported only by noiseless simulations, so the significance of that specific quantitative statement is currently limited until realistic noise is included or the claim is deliberately softened.","major_comments":[{"comment":"The sentence 'We have not considered detector noise for these simulations' is a load-bearing limitation. The quantitative claim in Sections 3.1 and 3.2 that LDFC and EKF maintain contrast at the level of the final iEFC contrast is established only in a noiseless idealization. For EKF, the measurement is the dark-hole intensity and the residual field is deliberately driven to a small value, so without simulated photon and read noise the filter can exploit intensity changes far below what a real camera can detect. For LDFC, the signal is a small intensity change in the bright field, and photon noise on that bright background sets a detection floor. Please either include noise-inclusive simulations with stated Q, R, and dither values, or explicitly soften the quantitative claims to 'noiseless-simulation performance' in the abstract, Section 3, and conclusions.","section":"Section 3, first paragraph"},{"comment":"The testbed LDFC demonstration is preliminary: the plotted curves have no error bars, the runs are single realizations, and the bench's own low-order drift remains uncorrected. The text states that the contrast curves with only the DH command and with the WFE added were recorded only every 30 and 15 iterations, respectively, so the baselines are sparsely sampled. This is acceptable for a qualitative demonstration, but it cannot substitute for the noise-inclusive simulation needed to support the quantitative maintenance claim or to establish repeatability.","section":"Section 4, Figures 8 and 9"},{"comment":"The EKF results depend on quantities that are not reported: the process noise Q, the measurement noise R, the dither amplitude, and the control gain schedule. The text mentions that the gain is set to zero for the first 10-20 iterations but does not state the subsequent gain or any parameter values. Without these, the reader cannot judge whether the demonstrated maintenance is robust to reasonable parameter choices or is specific to the simulated WFE cases. Please state the values used and, ideally, include a short sensitivity study over Q, R, and dither amplitude.","section":"Section 3.2, EKF equations"},{"comment":"The simulation conclusions are extrapolated to the bench through the POPPY Fresnel model that 'accounts for the surface error measurements.' The only direct model-testbed comparison is qualitative (Figures 1 versus 7). Given that the testbed shows a low-order drift that LDFC cannot correct and the simulations already predict insensitivity to low-order Zernikes, the transfer of the EKF simulation results to the real SCoOB bench remains unvalidated. Please add a quantitative model-versus-testbed comparison of the LDFC response curves or state a clear validation plan for EKF, or alternatively limit the conclusions to the simulation model.","section":"Section 3 and Section 4, model validation"}],"minor_comments":[{"comment":"The phrase 'scoob stategithub respository' appears to be a malformed reference; please provide a proper URL or a complete citation for the repository.","section":"Section 4"},{"comment":"The sentence 'we assume that the electric field has not drifted and that the state/open loop electric field is the field at the end of DH digging' is unclear; please define the state vector explicitly, stating that it contains the real and imaginary parts of the field at each dark-hole pixel, and clarify the relationship between the state and the open-loop field.","section":"Section 2.2, Step 1"},{"comment":"In Equation (4), the notation N(0, sigma_drift^2 I) is consistent with a multivariate normal distribution, but the text describes sigma_drift as 'the standard deviation' of the normal distribution; please state explicitly that sigma_drift is the per-actuator standard deviation and that the increments are independent across actuators.","section":"Equation (4)"},{"comment":"The dashed, dotted, and solid curves are described in the captions but are not labeled inside the panels; adding a legend or explicit curve labels would make the figures substantially easier to read.","section":"Figures 2 through 6"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The new content here is not the algorithms—LDFC and EKF-DZM are both prior published work—but their application to the SCoOB vortex coronagraph testbed, plus a simulation finding that LDFC is insensitive to low-order Zernikes when used with a VVC. That negative result is genuinely useful, and the paper does not oversell it. The testbed data are preliminary and honestly presented: LDFC corrects injected DM eigenmodes but does not fix the bench's own low-order drift, which is exactly the kind of finding people need before trusting LDFC on a vortex coronagraph.\n\nThe paper does several things well. The algorithm descriptions are clear, the claims are modest, and the simulated cases (single eigenmode, random walk, accumulating Zernikes) are reasonable probes. The testbed contrast curves, while lacking error bars, show the expected behavior for the injected disturbance. The reference list covers the prior work properly, and the self-citations are appropriate for a series paper.\n\nThe main soft spot is the noiseless simulation. The paper says explicitly that detector noise is not considered, and the central quantitative claim—that both algorithms maintain contrast near the final iEFC level—rests entirely on those simulations. That matters because EKF's measurement is dark-hole intensity, and in closed-loop maintenance the residual field is deliberately small; without photon or read noise the estimator can exploit changes that a real camera would not detect. LDFC has the same issue in the bright field, where photon noise on a bright background sets a detection floor. The paper also does not report the EKF's Q, R, or dither amplitude, so the EKF result is hard to reproduce or assess for sensitivity to those choices. These are real gaps, but they are not fatal to the paper's main qualitative points.\n\nWho gets value from this? Someone working on dark-zone maintenance for a vortex coronagraph, especially anyone considering LDFC, will want to know about the low-order blindness result. The paper is appropriate for a SPIE-style proceedings. If it were submitted to a journal, it would deserve peer review, but reviewers should ask for a noise-inclusive simulation and for the missing EKF parameters before acceptance.\n\nI would not desk-reject this. It is a solid, honest engineering data point that advances a practical question, and it does not overclaim. Send it to a referee who knows DZM literature and ask for a focused revision.","headline":"A modest, honest testbed paper: known DZM algorithms applied to a vortex coronagraph, with a useful negative finding about LDFC and low-order aberrations, but the quantitative contrast-maintenance claim rests on noiseless simulations.","tokens_in":9026,"tokens_out":2186,"would_cite":true,"duration_ms":22316,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper claims that dark zone maintenance can hold a vortex coronagraph's dark hole contrast near its post-digging level while wavefronts drift, with simulations of two algorithms and bench tests of one.","keywords":["high contrast imaging","coronagraphy","wavefront control","dark zone maintenance","linear dark field control","extended Kalman filter","vector vortex coronagraph"],"falsifier":"On the SCoOB bench, inject a slowly growing low-order Zernike mode into the deformable mirror while LDFC is closed-loop and record the dark hole contrast: if contrast returns to the post-iEFC floor, the claim that LDFC is insensitive to such modes with a vortex coronagraph is refuted; if contrast degrades while simulation predicts it should hold, the Fresnel model is the part to doubt.","tokens_in":8002,"feed_emoji":"🔭","tokens_out":9370,"duration_ms":72270,"temperature":0.7,"pith_summary":"This paper tries to establish that dark zone maintenance (DZM) algorithms can hold a coronagraph's dark hole—the region of suppressed starlight where exoplanets would appear—near the contrast level reached right after digging, despite wavefront drifts. It reports simulations of two DZM schemes on the SCoOB testbed: linear dark field control (LDFC), which uses the linear response of bright speckles, and an extended-Kalman-filter (EKF) approach that estimates the drifting electric field in the dark hole. Both simulations show contrast maintained near the post-digging floor for static and slowly accumulating wavefront errors. Preliminary bench tests confirm LDFC corrects deliberately injected deformable-mirror eigenmode errors, but not the bench's own low-order drift, which matches the simulated insensitivity of LDFC to low-order Zernike modes with a vortex coronagraph.","feed_headline":"Two algorithms hold dark hole contrast through wavefront drift","feed_subtitle":"LDFC and EKF hold the dark hole near its dug-in floor in simulations; on the bench, LDFC misses low-order drift.","key_machinery":"The central mechanism is the asymmetric response of the focal plane: in the bright field, intensity changes linearly with wavefront perturbations, while inside the dark hole the response is quadratic. LDFC exploits the linearity with a two-stage calibration that builds a control matrix from Hadamard probe patterns decomposed into DM eigenmodes, requiring one image per loop and no extra DM probes. The EKF alternative maintains an estimate of the complex electric field at each dark-hole pixel from single images, using a small DM dither for phase diversity, and converts the field estimate into DM commands through the EFC Jacobian. The vector vortex coronagraph, a focal-plane mask that suppresses on-axis starlight, is the optical element whose properties make LDFC's bright-field response weak for low-order modes.","core_discovery":"On its own terms, the paper's central claim is that dark zone maintenance can be decoupled from dark hole digging: once implicit electric field conjugation (iEFC) has dug the hole, LDFC and EKF-based DZM can each keep the monochromatic contrast at the final iEFC level while the wavefront drifts, using one image per iteration and no probing of the dark hole. The simulation evidence covers single DM eigenmodes, random walks of DM actuators, and slowly accumulating low-order Zernike errors. The testbed evidence is for LDFC only: it restores the iEFC floor after injected DM eigenmode errors, but the bench's ambient low-order drift leaks through because LDFC with a vortex coronagraph is insensitive to low-order modes. The paper identifies that limitation explicitly and names low-order wavefront sensing as the complementary correction.","pith_inferences":["A direct continuation is to run the EKF algorithm on the bench under the same injected eigenmode tests; success would cover the low-order drift that LDFC misses.","Marrying LDFC with low-order wavefront sensing on the same bench should close the gap, since the two sensors respond to complementary mode ranges.","For mission design, the DZM choice may reduce to a trade-off: LDFC is model-free and simple but blind to low-order modes, while EKF needs a state-space model and covariance tuning but handles those modes in simulation.","The single-image-per-iteration property suggests DZM could be interleaved with science frames rather than occupying dedicated calibration time, an operational pattern the paper does not develop."],"forward_implications":["A coronagraph could observe a target for hours while DZM keeps the dark hole at its dug-in contrast, eliminating the need to alternate to a reference star.","LDFC adds no perturbation to the dark hole during maintenance, so it can run while science data are being taken.","The simulated insensitivity of LDFC to low-order modes means a practical LDFC system for a vortex coronagraph should include a low-order wavefront sensor.","EKF-based DZM can stabilize contrast against slowly accumulating low-order Zernike drift and random-walk DM drift, covering regimes LDFC misses.","The convergence behavior of EKF, with a brief contrast dip in early iterations, means operational loops should zero the controller gain for the first iterations."],"supporting_citations":[{"why":"Defines spatial LDFC and the linear bright-field response that the algorithm exploits.","marker":"[14]"},{"why":"Demonstrates spatial LDFC maintaining high contrast on a high-contrast instrument with a vector-apodizing phase plate coronagraph.","marker":"[15]"},{"why":"Provides an early on-sky demonstration of spatial LDFC, supporting the method's practicality.","marker":"[16]"},{"why":"Introduces dark hole maintenance with an extended Kalman filter for speckle drift.","marker":"[18]"},{"why":"Describes an implementation of a dark zone maintenance algorithm for speckle drift correction in a space coronagraph.","marker":"[19]"},{"why":"Demonstrates a modal dark zone maintenance algorithm that the EKF variant builds on.","marker":"[20]"},{"why":"Defines implicit electric field conjugation, whose final contrast is the baseline the DZM algorithms must hold.","marker":"[6]"},{"why":"Supplies the physical optics propagation code used for the SCoOB simulations.","marker":"[22]"},{"why":"Reports the vacuum contrasts achieved by SCoOB that establish the bench performance being stabilized.","marker":"[5]"},{"why":"Provides the Fresnel model of SCoOB with measured surface errors used in the simulations.","marker":"[13]"}],"fun_headline_variants":["Dark zone maintenance: hold contrast without re-digging","LDFC and EKF stabilize dark hole under drift","Sims: DZM holds contrast; bench LDFC misses low-order","Decoupling digging: LDFC and EKF maintain dark hole"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the Fresnel model of the bench, including its measured surface errors, predicts real optical behavior closely enough that simulation conclusions about LDFC and EKF transfer to the testbed.","fun_headline_variants_meta":{"raw":{"variants":["Dark zone maintenance: hold contrast without re-digging","LDFC and EKF stabilize dark hole under drift","Sims: DZM holds contrast; bench LDFC misses low-order","Decoupling digging: LDFC and EKF maintain dark hole"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000198,"raw_usage":{"total_tokens":1327,"prompt_tokens":864,"completion_tokens":463,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":480,"completion_tokens_details":{"reasoning_tokens":391}},"tokens_in":480,"tokens_out":463,"duration_ms":4754,"temperature":1.0,"reasoning_tokens":391,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T14:53:09.097066+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"On the SCoOB bench, inject a slowly growing low-order Zernike mode into the deformable mirror while LDFC is closed-loop and record the dark hole contrast: if contrast returns to the post-iEFC floor, the claim that LDFC is insensitive to such modes with a vortex coronagraph is refuted; if contrast degrades while simulation predicts it should hold, the Fresnel model is the part to doubt.","supporting_citations":[{"cited_title":"Spatial linear dark field control: stabilizing deep contrast for exoplanet imaging using bright speckles,","cited_arxiv_id":null,"evidence_quote":"Defines spatial LDFC and the linear bright-field response that the algorithm exploits."},{"cited_title":"Spatial linear dark field control on Subaru/SCExAO. Maintaining high contrast with a vAPP coronagraph,","cited_arxiv_id":null,"evidence_quote":"Demonstrates spatial LDFC maintaining high contrast on a high-contrast instrument with a vector-apodizing phase plate coronagraph."},{"cited_title":"First on-sky demonstration of spatial Linear Dark Field Control with the vector-Apodizing Phase Plate at Subaru/SCExAO,","cited_arxiv_id":null,"evidence_quote":"Provides an early on-sky demonstration of spatial LDFC, supporting the method's practicality."},{"cited_title":"Dark Hole Maintenance and A Posteriori Intensity Estimation in the Presence of Speckle Drift in a High-contrast Space Coronagraph,","cited_arxiv_id":null,"evidence_quote":"Introduces dark hole maintenance with an extended Kalman filter for speckle drift."},{"cited_title":"Implementation of a dark zone mainte- nance algorithm for speckle drift correction in a high contrast space coronagraph,","cited_arxiv_id":null,"evidence_quote":"Describes an implementation of a dark zone maintenance algorithm for speckle drift correction in a space coronagraph."},{"cited_title":"Demonstration of a modal dark zone maintenance algorithm for high-contrast direct imaging,","cited_arxiv_id":null,"evidence_quote":"Demonstrates a modal dark zone maintenance algorithm that the EKF variant builds on."},{"cited_title":"The space coronagraph optical bench (scoob): 11. modeling and correction of chromatic aberrations,","cited_arxiv_id":null,"evidence_quote":"Provides the Fresnel model of SCoOB with measured surface errors used in the simulations."}],"review_version":1}