CandidatesCYG-2026-09-TOI3500.02

Candidate dossier · working identifier

CYG-2026-09-TOI3500.02

S64/S90 dips persist, but official UPDATED_2.0 TESS PRFs do not assign the deficit robustly between the0.18-pixel-separated Gaia A/B pair. Source preference reverses with allowed registration and1/3/5-percent model floors; static PRF residuals exceed nominal interpolation accuracy. S101 remains rejected and supplies no period preference. All16 baseline aliases remain open; Gaia RV summaries provide no orbit or companion-mass bound. Unverified lead. Six public HARPS epochs span111.785days with87.507m/s corrected-RV range. A conditional circular50.0446-day fit gives signedK about39m/s but six epochs, line-width correlation and unresolved source identity prevent confirmation or alias rejection. Eighteen further HARPS and six FEROS products returnHTTP401. A relative-spectrum attempt on23 FEROS and6 HARPS spectra fails the independent HARPS precision check (168m/s RMS); derived shifts are not accepted orbital velocities. A completed31-input line-mask rerun gives HARPS CCF-alignment disagreement RMS10.9458m/s (R50G2) and10.8503m/s (K0), both failing the10m/s requirement. FEROS midpoint barycentric replacement deltas are roughly-103 to+173m/s; release products also lack simultaneous-reference drift correction and usable ERR arrays. FEROS profile shifts remain unvalidated.

Unverified lead

Working identifier, not a designation

CYG-2026-09-TOI3500.02 is a local label used inside this project. It is not an official catalog name and does not imply the object is new or confirmed.

Measured signal

QuantityValueUnitUncertaintyMethod
E1_S90BJD 2460757.3199, depth 7553 +/- 277 ppm (box fit), shape ratio 1.08 +/- 0.06, duration ratio 1.00not recordednot recordednot recorded
E2_S101_rejectedBJD 2461107.6407, difference-image centroid 6.74 arcsec off at 11.5 sigma, POS_CORR2 +28.9 sigma, depth 1614-7796 ppm across detrendings — an aperture-loss/pointing artifact; supports no periodnot recordednot recordednot recorded
HARPS descriptive RV range0.08750691315560033km/snot recordedsix pipeline RVC epochs; range is not orbitalK or mass
alias_family_16_aliasesP = 700.62, 350.31, 233.54, 175.16, 140.12, 116.77, 100.09, 87.58, 77.85, 70.06, 63.69, 58.39, 53.89, 50.04, 36.88, 35.03 d (density limit excludes 18.44 d)not recordednot recordednot recorded
delta_T_reference_to_E1700.6242dnot recordedreference S64 to accepted E1 S90; allowed aliases P = ΔT/n, including 350.31 d
no_secondary_E1with E2 excluded, phase-0.5 windows are clean at all covered aliases (e.g. P 35.03: -80 +/- 186; 50.04: +313 +/- 167 ppm)not recordednot recordednot recorded
red_noise_significance_E1local robust statistic 36.0 relative to 300 random epochs; zero exceedances, search-wide false-alarm rate not estimatednot recordednot recordednot recorded
reference_depth7020ppm+/-282PDCSAP median residual (catalogue depth for TOI-3500.02: 8095 ppm)
reference_duration7.91hnot recordedbox fit
reference_transit_S64_bjd2460056.6957BJD_TDBnot recordedbox fit
stellar_priors1.10 Rsun, 0.94 Msun host (density limit), host G 11.35 — RV-measurable; sibling TOI-3500.01 (P 7.3436 d, PC) transit is not the signalnot recordednot recordednot recorded
superseded_bottom_line_2026-09-27Two repeat dips remain in the TIC 443666343 aperture (the S90 event E1 and the S64 reference; S101 E2 is rejected as an aperture-loss/pointing artifact). Their depth, duration and shape agree, and their difference images lie near the target. A 2026-09-27 two-source Gaussian-PSF sensitivity grid usually prefers the target over the 3.71-arcsec co-moving neighbour, but the preference reverses under plausible PSF/registration perturbations because the sources are only 0.184 TESS pixel apart. Localization therefore remains inconclusive. Sixteen aliases P = 700.6242/n over 35.03–700.62 d remain; search-wide false alarms, calibrated PRF localization, RV and phase coverage are incomplete. Unverified lead.not recordednot recordedhistorical narrative; superseded by2026-09-30 follow-up; original measurements retained

Artifact audit 30 tests

TestState
2026-09-30 SAP/PDCSAP baseline persistencePassed
2026-09-30 fixed-epoch injection sensitivityPassed
2026-09-30 search-wide false-alarm calibrationNot tested
FEROS precision calibrationInconclusive
Gaia neighbours able to mimic the depth (E1)Inconclusive
HARPS six-epoch pipeline RV retrievalPassed
Line-mask CCF precision validationFailed
RV (host G 11.35)Not tested
RV orbit and activity discriminationInconclusive
Relative-spectrum precision validationFailed
ZTF light curve (3.71 arcsec neighbour)Not tested
ZTF phase coverage (host)Not tested
alternative detrending (E1)Passed
background/centroid/pointing (E1)Passed
background/centroid/pointing (E2)Failed
blend blame (3.71 arcsec neighbour)Inconclusive
blend test (43.2 arcsec neighbour)Not tested
box fit (E1; E2)Passed
calibrated two-source PRF2026-09-30Inconclusive
difference-image centroid (E1)Passed
difference-image centroid (E2)Failed
future-sector alias test (P = 350.31 d epemeris next transit ~2027-08)Not tested
independent cached TPF localization (2026-09-27)Failed
quality flags and coverage (E1)Passed
red-noise significance (E1; E2)Inconclusive
same-CCD common mode (E1; E2)Passed
secondary eclipse at phase 0.5Inconclusive
shape vs reference (E1)Passed
stellar-density duration limit on aliasesPassed
two-source Gaussian-PSF sensitivity grid (2026-09-27)Inconclusive

“Not tested” means the test was unavailable or not run. It never counts toward a higher evidence level.

Catalog and literature audit

2026-09-30 primary literature and exact-source refresh
reports/lead-resolution-2026-09-30/PRIOR_ART_AUDIT.md; archive_refresh.json. Companion/binary evidence is not an event-time matched orbit.
Gaia DR3 cone
3.71 arcsec co-moving neighbour (parallax 5.21 mas = host's, G 14.07); 43.2 arcsec neighbour (G ~14.5)
NASA Exoplanet Archive (pscomppars, 30 arcsec)
no match within 30 arcsec (as of 2026-09-26T08:20:55Z)
SIMBAD (30 arcsec)
UCAC2 19403154 (PM*) — no eclipsing-binary class
TESS_TOI (30 arcsec)
TOI-3500.01 (PC, sibling known candidate, P 7.3436 d) and TOI-3500.02 (PC) — the phase-0.5 sibling-vs-lead relation checked through the sibling ephemeris in vet
VSX (30 arcsec)
no match (as of 2026-09-26T08:21:01Z)

Prior-art gate (ledger)

GateServiceQueryResultAs of (UTC)
catalogNASA_Exoplanet_ArchiveSELECT pl_name, hostname, ra, dec FROM pscomppars WHERE CONTAINS(POINT('ICRS', ra, dec), CIRCLE('ICRS', 186.8168840, -29.8329960, 0.0083333)) = 1no match in NASA_Exoplanet_Archive within 30" as of 2026-09-26T08:20:55Z2026-09-26T08:20:55Z
catalogSIMBADSELECT main_id, otype, ra, dec FROM basic WHERE CONTAINS(POINT('ICRS', ra, dec), CIRCLE('ICRS', 186.8168840, -29.8329960, 0.0083333)) = 11 match(es) in SIMBAD within 30" as of 2026-09-26T08:21:03Z: UCAC2 19403154 (PM*)2026-09-26T08:21:03Z
catalogTESS_TOISELECT toi, tid, tfopwg_disp, ra, dec FROM toi WHERE ra BETWEEN 186.8072776 AND 186.8264904 AND dec BETWEEN -29.8413293 AND -29.82466272 match(es) in TESS_TOI within 30" as of 2026-09-26T08:20:58Z: TOI-3500.01 (TIC 443666343, disposition PC); TOI-3500.02 (TIC 443666343, disposition PC)2026-09-26T08:20:58Z
catalogVSXSELECT "Name", "Type", "Period", RAJ2000, DEJ2000 FROM "B/vsx/vsx" WHERE CONTAINS(POINT('ICRS', RAJ2000, DEJ2000), CIRCLE('ICRS', 186.8168840, -29.8329960, 0.0083333)) = 1no match in VSX within 30" as of 2026-09-26T08:21:01Z2026-09-26T08:21:01Z

A missing match means “not found in the services searched as of that date”, not “uncatalogued”.

Competing explanations

  • Transiting planet or eclipsing companion under any of the 16 allowed aliases. At 7 ppt on an assumed 1.10 Rsun host, the central undiluted sqrt(depth) radius scale is about 10 Earth radii; grazing, dilution and unresolved multiplicity make this illustrative only. RV would constrain mass.
  • Long-period grazing eclipsing binary: secondary can hide in the seasonal gaps of the long aliases; equal depth and equal shape currently constrain it — RV is decisive.
  • Contamination by the 3.71-arcsec co-moving neighbour remains unresolved. The reviewable Gaussian-PSF grid favors the target in 71.4% of S64 trials and 85.7% of S90 trials, but target-versus-neighbour preference changes sign across the tested PSF widths and registration shifts; template correlations are 0.934-0.986 because the pair is separated by only 0.184 TESS pixel. A calibrated TESS PRF or higher-resolution observation is required.
  • Aperture-loss artifact: applies to E2 only (rejected); E1 carries no POS_CORR excursion and 237/237 usable cadences.
  • Sibling-TOI contamination (TOI-3500.01's P 7.34-d ephemeris): excluded by the sibling-ephemerides check — no catalogued sibling transit at E1's epoch.

Reproduction

2026-09-30
python reports/lead-resolution-2026-09-30/resolve_leads.py --help; config.json; see REPORT.md for ordered modes and validation
code
cygnus (src/cygnus/); historical baseline gate at commit 6a3aa66: 639 passed, 37 deselected; current workspace gate 2026-09-27: 644 passed, 37 deselected
figures
vetting/figures/ and vetting/*.png in the campaign directory
follow_up_2026_09_27
reports/lead-followup-2026-09-27/independent_tpf_check.py and toi3500_two_source_test.py; results in independent_tpf_results.json, toi3500_two_source_results.json and REPORT.md
package_versions
python 3.13; astropy/lightkurve/numpy per requirements.txt; candidates rendered by cygnus.reporting.dossier
run
python -m cygnus.multi run campaigns/toi-3500-02.yaml ; python -m cygnus.campaign vet campaigns/toi-3500-02.yaml

Next test

Use validated empirical sector PRFs or resolved time-series photometry for A/B; seek another clean event or component RV series. Retain all16 aliases, exclude S101 from period evidence, and do not treat catalog RV range as an orbital semiamplitude. Improve and validate line-mask/continuum spectral RV extraction; test the50.0446-day hypothesis against activity and blended components using additional accessible spectra.