pith. machine review for the scientific record. sign in

arxiv: 2604.09435 · v1 · submitted 2026-04-10 · 🌌 astro-ph.EP · astro-ph.IM· astro-ph.SR

Recognition: unknown

Finding Circumbinary Planets: A Semi-Automated Transit Search of TESS Eclipsing Binaries

Benjamin D. R. Davies , David J. A. Brown , Samuel Gill , Jenni R. French
Authors on Pith no claims yet

Pith reviewed 2026-05-10 16:59 UTC · model grok-4.3

classification 🌌 astro-ph.EP astro-ph.IMastro-ph.SR
keywords circumbinary planetsTESSeclipsing binariestransit searchexoplanet detectionlight curve analysisplanet formation
0
0 comments X

The pith

A semi-automated pipeline removes stellar eclipses from TESS binary light curves to search for single planetary transits, recovering at least half the known circumbinary planet signals and identifying one new candidate.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops and tests a method to find planets that orbit both stars in a binary system by processing TESS observations of eclipsing binaries. It first strips away the regular eclipses between the two stars, applies custom detrending to the remaining light curve, and then looks for isolated dips that could be planetary transits before applying automated checks to discard false alarms. When run on the known population of transiting circumbinary planets, the method recovers at least 50 percent of the transits for every planet and more than 75 percent for nine of the fourteen planets, provided the signals exceed the data noise level. Applied to a catalog of TESS eclipsing binaries, it produces one candidate transit event. The framework is built to run on large samples with low computational cost and is intended for use with future photometric surveys.

Core claim

The mono-cbp framework identifies planetary transit events in TESS eclipsing binary light curves by removing stellar eclipses, applying a custom detrending procedure, searching for individual transit events, and using automated vetting to filter false positive signals, achieving a recovery rate of at least 50 percent for each known transiting circumbinary planet in Kepler and TESS data and yielding one candidate in the searched TESS Eclipsing Binary Catalogue sample.

What carries the argument

mono-cbp, the semi-automated pipeline that removes stellar eclipses from binary light curves, applies custom detrending, detects individual transit events, and applies automated vetting procedures to filter false positives.

If this is right

  • The method can be applied to large samples of TESS eclipsing binaries with little computational burden.
  • Recovery performance is a strong function of transit duration and the specific metrics used to reject false positives.
  • The same framework can be extended to photometry from future space-based photometric surveys.
  • One candidate transit event was identified in the TESS Eclipsing Binary Catalogue.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Wider use of the pipeline could enlarge the known population of circumbinary planets and tighten constraints on their formation and orbital evolution.
  • The dependence of recovery on transit duration suggests the method may systematically miss short events unless the false-positive filters are tuned.
  • Combining the search with radial-velocity or multi-band follow-up could provide independent confirmation of any candidates.

Load-bearing premise

The custom detrending and automated vetting procedures preserve genuine planetary transits while removing false positive signals, and the recovery rates measured on known planets and injections apply to the broader TESS eclipsing binary sample.

What would settle it

Follow-up observations that confirm or rule out the single candidate transit event as a genuine circumbinary planet would directly test whether the pipeline's detection and vetting steps work as claimed on real TESS data.

Figures

Figures reproduced from arXiv: 2604.09435 by Benjamin D. R. Davies, David J. A. Brown, Jenni R. French, Samuel Gill.

Figure 1
Figure 1. Figure 1: The detection of a transit of TOI-1338 b by mono-cbp. Top: A zoom-in of the identified event. The grey points depict the detrended flux. The legend shows the estimated parameters, including the depth (𝛿), duration (𝑡dur), binary phase at the time of the event (𝜙𝑏), and the event SNR (see Eq. 1). Middle: The full detrended light curve. Bottom: The full raw light curve. The solid red curve shows the detrendi… view at source ↗
Figure 2
Figure 2. Figure 2: Examples of events identified by mono-cbp and vetted by the model comparison (Section 2.4). Top left: Sector 6 transit of TOI-1338 b (same as [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Top: TESS magnitude distribution of our sample of EBs. The median TESS magnitude is ≈ 10.2. Bottom: Period distribution of our sample of EBs. The median period in our sample is ≈ 11.4 d. 6 EBs with periods > 50 d are not show in the period distribution for visualisation purposes. however, exist small CBPs with stable, misaligned orbits in these systems which may sparsely transit their host stars (Martin et… view at source ↗
Figure 4
Figure 4. Figure 4: presents an overview of the number of TCEs that were identified in our search, and how these were processed through mono-cbp. In summary, we applied mono-cbp to 3808 TESS-SPOC FFI light curves (up to Sector 79) of the 512 EBs in our sample, iden￾tifying 7176 TCEs. After the default vetting metrics were applied (Skye metric flag, detrending-dependence flag, SNR and duration cuts; see Section 2.4 for details… view at source ↗
Figure 5
Figure 5. Figure 5: The candidate event identified in the Sector 7 TESS-SPOC light curve of TIC 319011894. Bottom: The Sector 7 TESS-SPOC light curve of TIC 319011894, with the mono-cbp detrending depicted by a red curve. There are three eclipses with depth ≈ 2.2 per cent and period of 8.3 d. The time of the candidate event (∼ 1495.8 BTJD) is marked with a blue vertical line. If the candidate event was a secondary eclipse, th… view at source ↗
Figure 6
Figure 6. Figure 6: Candidate CBP transit in the Sector 7 light curve of TIC 319011894 extracted from five different TESS data reduction pipelines (top to bottom): TESS-SPOC (blue; Caldwell et al. 2020), SPOC (green; Jenkins et al. 2016), QLP (orange; Huang et al. 2020a,b), eleanor (pink; Feinstein et al. 2019), and TGLC (yellow; Han & Brandt 2023). For each light curve, the raw/SAP flux is shown. The candidate was identified… view at source ↗
Figure 7
Figure 7. Figure 7: The detection efficiency (i.e., the proportion of individual transits identified by the framework) of mono-cbp for the known transiting CBPs. The bars are colour-coded based on the mean SNR of an individual transit. For SNR ≳ 10 (the majority of the known planets), the framework identifies nearly all transits. Below SNR ≲ 8, the framework struggles to identify all individual transits consistently, but in a… view at source ↗
Figure 8
Figure 8. Figure 8: Results of our injection-recovery tests. For each square in the duration-depth grid, 1000 transit profiles with the corresponding depth and duration were injected into random light curves in our sample. The proportion of injections that were recovered by mono-cbp are visualised by the colour bar and by the text in each grid square. Left: Recovery rates before we apply our vetting metrics (i.e., injected ev… view at source ↗
Figure 9
Figure 9. Figure 9: Recovered Signal-to-Noise Ratio (SNR) of injected transit signals as a function of transit depth and duration, before our vetting metrics are applied. Most of the signal is retained for transits with durations of ≤ 0.7 d, while our detrending approach removes a significant proportion of the signal for longer duration transits. corresponds to only 4 or 5 cadences. Therefore, whether an event passes the dete… view at source ↗
Figure 10
Figure 10. Figure 10: Model comparison classifications of the recovered transits from the injection-recovery tests, after applying the default vetting cuts. Most are classified as transits/ambiguous transits. Many of the longer duration transits are classified as steps due to the detrending altering the transit morphology. of the injected transits are misclassified, with many of these events being best fit by the step model. A… view at source ↗
read the original abstract

The discovery of circumbinary planets (CBPs) has advanced our understanding of planet formation and dynamical evolution in complex environments. However, the population of such planets remains small, leading their underlying physical properties to be loosely constrained. In this work, we have developed a semi-automated framework to identify planetary transit events in light curves of eclipsing binaries observed by the Transiting Exoplanet Survey Satellite (TESS). Our search method, ${\tt mono-cbp}$, removes stellar eclipses and applies a custom detrending procedure, searching for individual transit events and applying automated vetting procedures to filter false positive signals. We searched a sample of binaries from the TESS Eclipsing Binary Catalogue, yielding one candidate transit event. ${\tt mono-cbp}$ was also tested on the known population of transiting CBPs, using the Kepler long-cadence photometry for the Kepler transiting CBPs and the TESS Full Frame Image photometry for the TESS CBPs. Excluding transits that are shallower than the intrinsic noise of the Kepler/TESS data, ${\tt mono-cbp}$ achieved a recovery rate of $\geq50$ per cent for each planet, reaching >75 per cent for 9 of the 14 planets. To test the limits of our framework, we injected simulated transit profiles with varying depth and duration into our sample of TESS light curves, finding that our recovery rate is a strong function of transit duration and the metrics used to filter false positive signals. This framework may be applied to large samples of TESS eclipsing binaries with little computational burden and to photometry from future space-based photometric surveys.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 2 minor

Summary. The paper presents a semi-automated framework called mono-cbp for detecting individual planetary transits in TESS light curves of eclipsing binaries. Eclipses are removed, followed by custom detrending, a search for transit-like events, and automated vetting to suppress false positives. The method is applied to a sample from the TESS Eclipsing Binary Catalogue, yielding one candidate transit event. Validation on 14 known transiting CBPs (using Kepler long-cadence and TESS FFI photometry) reports recovery rates of ≥50% per planet (>75% for 9 of 14), after excluding transits shallower than the data noise. Injection tests into TESS light curves show recovery depends strongly on transit duration and the vetting metrics employed. The framework is positioned as computationally lightweight for large EB samples and future surveys.

Significance. If the detrending, search, and vetting steps generalize reliably beyond the development cases, the approach could efficiently expand the small known population of circumbinary planets by mining existing and future TESS EB photometry with low computational cost. The reported recovery statistics on known systems and the injection tests provide concrete supporting evidence of sensitivity to transit duration and vetting choices; however, the single candidate and broader applicability rest on untested extrapolation from a small, heterogeneous test set.

major comments (3)
  1. [Validation on known CBPs] Validation on known CBPs (as described in the testing section): recovery rates of ≥50% (and >75% for 9/14) are reported after excluding transits shallower than intrinsic noise, but the quantitative definition of this noise threshold, the per-light-curve noise estimation procedure, and the exact vetting metrics/thresholds are not specified. These details are load-bearing for interpreting the sensitivity and for assessing whether the rates generalize to the searched TESS EB sample.
  2. [Application to TESS Eclipsing Binary Catalogue] Application to TESS Eclipsing Binary Catalogue (results section): the search yields one candidate transit event, yet no information is given on the total number of binaries processed, the precise sample selection criteria from the catalogue, or any error analysis/statistical significance assessment for the candidate. This leaves the central claim of a new detection only partially substantiated.
  3. [Injection tests] Injection tests (limits of framework section): recovery is stated to be a strong function of transit duration and the metrics used to filter false positives, but the specific numerical thresholds, how they were chosen, and any cross-validation or blind tests on held-out real TESS EBs are not described. Given that the test set comprises only 14 known CBPs plus injections (drawn from Kepler/TESS FFI data with potentially different noise properties), this raises a correctness risk for generalization to the full catalogue.
minor comments (2)
  1. [Abstract] The abstract would be strengthened by including the total sample size searched and a brief quantitative summary of the vetting metrics.
  2. [Methods] Notation for the detrending parameters and vetting thresholds should be defined explicitly when first introduced to improve reproducibility.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive comments on our manuscript describing the mono-cbp framework for detecting circumbinary planets in TESS eclipsing binary light curves. We address each of the major comments below and will update the manuscript to incorporate additional details where necessary to improve clarity and substantiation of our results.

read point-by-point responses

  1. Referee: [Validation on known CBPs] Validation on known CBPs (as described in the testing section): recovery rates of ≥50% (and >75% for 9/14) are reported after excluding transits shallower than intrinsic noise, but the quantitative definition of this noise threshold, the per-light-curve noise estimation procedure, and the exact vetting metrics/thresholds are not specified. These details are load-bearing for interpreting the sensitivity and for assessing whether the rates generalize to the searched TESS EB sample.

    Authors: We agree that these specifics are crucial for a full understanding of the method's performance. In the revised manuscript, we will explicitly define the quantitative noise threshold used to exclude shallow transits, detail the procedure for estimating the intrinsic noise on a per-light-curve basis, and list the exact vetting metrics along with their applied thresholds. This will enable better assessment of the reported recovery rates and their potential generalization. revision: yes

  2. Referee: [Application to TESS Eclipsing Binary Catalogue] Application to TESS Eclipsing Binary Catalogue (results section): the search yields one candidate transit event, yet no information is given on the total number of binaries processed, the precise sample selection criteria from the catalogue, or any error analysis/statistical significance assessment for the candidate. This leaves the central claim of a new detection only partially substantiated.

    Authors: We acknowledge the need for more complete information on the search application. The revised manuscript will report the total number of binaries processed from the TESS Eclipsing Binary Catalogue, specify the precise sample selection criteria employed, and include an error analysis or statistical significance assessment for the identified candidate transit event, such as through comparison to the distribution of vetting scores from the injection tests. revision: yes

  3. Referee: [Injection tests] Injection tests (limits of framework section): recovery is stated to be a strong function of transit duration and the metrics used to filter false positives, but the specific numerical thresholds, how they were chosen, and any cross-validation or blind tests on held-out real TESS EBs are not described. Given that the test set comprises only 14 known CBPs plus injections (drawn from Kepler/TESS FFI data with potentially different noise properties), this raises a correctness risk for generalization to the full catalogue.

    Authors: We will expand the description of the injection tests to include the specific numerical thresholds for the false positive filtering metrics and explain the process by which they were selected, for instance by tuning to achieve high recovery on the known CBP sample. While we did not perform explicit cross-validation or blind tests on held-out real TESS EBs owing to the small number of known systems available for validation, the recovery statistics on all 14 known CBPs and the injections into real TESS photometry provide relevant constraints on the framework's performance. We will discuss the limitations of the test set more thoroughly in the revision. revision: partial

Circularity Check

0 steps flagged

No circularity: validation uses independent known planets and injections

full rationale

The paper describes a search pipeline (eclipse removal, custom detrending, transit search, automated vetting) applied to the TESS Eclipsing Binary Catalogue. Its performance claims rest on recovery statistics measured against 14 known transiting CBPs (Kepler long-cadence and TESS FFI photometry) and on separate injection tests of simulated transits into TESS light curves. These benchmarks are external to the searched sample and to any parameter tuning performed on the development cases; no equation, fit, or self-citation reduces the reported recovery rates or candidate detection to the inputs by construction. The framework is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 0 invented entities

The framework rests on several tunable parameters in the custom detrending and vetting steps whose values are not specified, plus standard assumptions about transit signals in photometric data.

free parameters (2)
  • detrending parameters
    Custom detrending procedure requires choices of smoothing scale or filter parameters that are not detailed.
  • vetting thresholds
    Automated false-positive filters use metrics with decision thresholds that must be set by hand or tuning.
axioms (1)
  • domain assumption Stellar eclipses can be modeled and subtracted without removing or distorting superimposed planetary transits.
    The method begins by removing stellar eclipses as a prerequisite step.

pith-pipeline@v0.9.0 · 5621 in / 1317 out tokens · 59464 ms · 2026-05-10T16:59:56.267388+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

1 extracted references · 1 canonical work pages

  1. [1]

    Information

    Agol E., Luger R., Foreman-Mackey D., 2020, AJ, 159, 123 Akaike H., 1998, Information Theory and an Extension of the Maxi- mum Likelihood Principle. Springer New York, New York, NY, pp 199–213,doi:10.1007/978-1-4612-1694-0_15,https://doi.org/10. 1007/978-1-4612-1694-0_15 Armstrong D., et al., 2013, MNRAS, 434, 3047 Armstrong D. J., Osborn H. P., Brown D. ...

    work page doi:10.1007/978-1-4612-1694-0_15 2020