{"id":"cd2a22cb-b08c-42c5-b261-b0c94abe3911","arxiv_id":"2507.21967","paper_version":3,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Archival TESS images show 3I/ATLAS was brighter than distance effects alone can explain, implying possible cometary activity at ~6 au from the Sun.","lead":"3I/ATLAS, the third interstellar object found passing through the Solar System, was recovered in archival TESS satellite images from May and June 2025, about two months before its official discovery. The object appears brighter than its bare nucleus would be, suggesting it was already releasing dust or gas at roughly 6 times the Earth-Sun distance.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Distant-activity claim hinges on combining two epochs with different cameras, a single adopted zeropoint, and the Farnham color term; the weaker epoch (Camera 1 CCD 2) is where the activity evidence lives.","rationale":"The paper does a careful job on the detection side: shift-stacking thousands of FFIs, validating with 896 Sphinx, testing multiple apertures and background models, and publicly releasing code/data. The strongest claim as stated by the reader is the 19σ/11σ detection with corresponding HV values implying activity at 6 au, and the paper itself hedges that claim in Section 4. My stress-test focuses on the differential step: the activity interpretation is not primarily a single-epoch photometry claim but a two-epoch comparison across two different TESS camera/CCD pairs. The weakest identified assumption in the reader's verdict is the TESS-to-V color conversion (Equation 2), and that is certainly a real uncertainty, but I see the more load-bearing structural issue as the cross-detector zeropoint and crowding-dependent photometry on the second epoch. If the Camera 1 CCD 2 photometry is contaminated by unresolved field stars or a detector-specific zeropoint offset, then the brightening is not necessarily activity. The authors acknowledge the color uncertainty explicitly, but they do not quantify cross-detector zeropoint differences for Sector 92, and the by-eye rejection of 5–9% of frames plus the half-crowded-frame masking rule is not fully algorithmic. My concrete proposal — relative zeropoint checks and injected-source recovery through the same pipeline — is exactly the kind of independent absolute/relative calibration that would settle the issue. Given the reader's verdict was already CONDITIONAL, my concern does not move the verdict, but it does sharpen the condition: the activity interpretation should be verified against a detector-relative zeropoint calibration and/or an independent calibration of the second-epoch photometry.","tokens_in":18963,"tokens_out":2003,"duration_ms":21409,"concrete_test":"Recalibrate both deep stacks against a common absolute or relative standard: (1) inject a synthetic point source of known Tmag into the actual FFIs at the 3I/ATLAS positions and run the identical shift-stack pipeline, verifying that the recovered flux matches the injected flux on both detectors; (2) cross-match a set of isolated, non-variable stars that appear on both Camera 2 CCD 3 and Camera 1 CCD 2 in Sector 92 and compute the relative zeropoint offset between the two detectors; report whether the 3I/ATLAS brightness difference survives a detector-specific zeropoint correction. If the offset is >0.1 mag, the 1.55 mag brightening and the distant-activity inference weaken.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim — that 3I/ATLAS was active at ~6 au — rests on the 1.55 mag brightening between the Camera 2 CCD 3 deep stack (Tmag = 20.83, HV = 13.72) and the Camera 1 CCD 2 deep stack (Tmag = 19.28, HV = 12.52). This comparison is intra-survey but cross-detector: the two epochs come from different cameras/CCDs with potentially different zeropoints, and the authors adopt a single global zeropoint (Equation 1, +0.05 mag error) without recalibrating it for Sector 92. The crowding filter (Section 2.2) removes over half of Camera 1 CCD 2 frames by a criterion the authors describe as tuned by-eye, and the 3x3 aperture on a median stack of a crowded-field object can mix in unresolved stellar flux; the claimed 11σ detection has a background of 0.141 ± 0.039 e-/s, roughly 24% of the source, so aperture contamination is not negligible. Additionally, the conversion V = Tmag + 0.8 (Equation 2) carries ±0.3 mag color uncertainty, which the authors themselves state permits the first-epoch HV to be ~1 mag fainter. The 1.55 mag brightening, and hence the 'factor 5 flux increase' and the inconsistency with an inactive body, is sensitive to these systematics. Section 4 explicitly acknowledges the 3σ color uncertainty, but the cross-camera zeropoint and crowding-systematics issue is not quantified. This is the load-bearing weak point: the activity claim is a differential statement, yet the differential is evaluated across two detectors, with only a single common zeropoint and partially subjective frame rejection.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper reports precovery photometry of interstellar object 3I/ATLAS in TESS Sector 92 full-frame images. Using a shift-and-stack algorithm on roughly 9,800 calibrated FFIs, the authors detect the object on two camera/CCD combinations at 19 sigma and 11 sigma, with TESS magnitudes Tmag = 20.83 ± 0.05 and 19.28 ± 0.05. Applying the TESS-to-V color relation of Farnham et al. (2021) and an n = 2 distance normalization, they derive absolute visual magnitudes HV = 13.72 ± 0.35 and 12.52 ± 0.35, which are brighter than the HST nucleus limit H > 15.4. The 1.55 mag brightening between the two epochs exceeds the geometric factor of about 1.5 expected from the changing Sun and observer distances, leading the authors to suggest possible activity at heliocentric distances near 6 au. The paper also extracts a 20-day light curve and finds no statistically significant rotation period. The pipeline is validated by recovering the known 21.04-hour rotation period of minor planet 896 Sphinx.","tokens_in":19319,"tokens_out":8924,"duration_ms":99611,"significance":"If the activity inference holds, these are the earliest precovery observations of 3I/ATLAS and would indicate volatile activity at about 6 au, which is important for understanding hypervolatile-driven mass loss in interstellar objects. The paper's strengths are its public data and code, the validation against 896 Sphinx, and the explicit null result for rotation. However, the activity claim is a differential measurement across two different detectors, and it depends on external color and zeropoint calibrations whose uncertainties the authors themselves acknowledge in Section 4. The result is therefore a tentative suggestion rather than a secure detection of distant activity, and the central claim needs to be hardened or softened accordingly.","major_comments":[{"comment":"The central activity claim rests on the 1.55 mag brightening between the Camera 2 CCD 3 and Camera 1 CCD 2 deep stacks, but this is a cross-detector differential measured with a single global TESS zeropoint (Eq. 1) and an adopted error of only 0.05 mag. The manuscript states that it is standard practice not to recalculate the zeropoint, but it does not bound the expected zeropoint variation across cameras and CCDs in Sector 92. Because the activity evidence is concentrated in the Camera 1 CCD 2 epoch, an inter-detector zeropoint offset of only a few tenths of a magnitude would erase the anomaly. Please calibrate the relative zeropoint using field stars common to both detectors, or demonstrate from similar TESS data that the cross-detector zeropoint dispersion is smaller than about 0.1 mag, and propagate this into the reported magnitudes.","section":"§3.1, Eq. (1), Table 1"},{"comment":"The Camera 1 CCD 2 deep stack has a central-pixel flux of 0.589 ± 0.016 e−1 s−1 and a background scatter of 0.141 ± 0.039 e−1 s−1, while the 3×3 aperture flux is 2.904 ± 0.016 e−1 s−1. The crowding filter in Section 2.2 only removes frames where more than half of the pixels in an 11×11 box are flagged as >2σ outliers; it does not ensure that the specific 3×3 aperture is free of unresolved stellar flux. In a field this crowded, undetected neighbours could contribute a substantial fraction of the 3×3 flux, biasing Tmag bright and mimicking activity. Please quantify this by comparing apertures of different sizes, fitting a PSF, or performing the same stacking on nearby off-source positions with the same filter, and report the resulting contamination correction.","section":"§2.2, §2.5, Fig. 5"},{"comment":"The statement in Section 3.1 that the observations are 'statistically inconsistent with an asteroid-like reflectance model' is based only on the 0.05 mag photometric errors. It does not include the ±0.3 mag uncertainty in V = Tmag + 0.8 (Eq. 2), nor the cross-detector and crowding systematics discussed above. The authors themselves note in Section 4 that the 3σ color uncertainty permits the object to be about 1 mag fainter, which would substantially reduce the inferred brightness anomaly. The inconsistency claim should be rephrased as tentative, and the HV values should be reported with a full systematic budget (color, zeropoint, crowding) rather than the current 0.35 mag uncertainty, which appears to be dominated by the color term alone.","section":"§3.1, §3.3 (Eq. 2), §4"}],"minor_comments":[{"comment":"The arithmetic of the claimed flux excess is inconsistent: the expected Tmag of about 20.5 versus the observed 19.28 differs by 1.22 mag, which is a flux ratio of about 3, not the 'factor of 5' stated in the text.","section":"§3.1"},{"comment":"The abstract says 'average TESS magnitude' but reports two separate epoch values; this phrasing should be changed to avoid implying a single averaged measurement.","section":"Abstract and Table 1"},{"comment":"The removal of contaminated FFIs is described as by-eye identification, and the crowding threshold and Savitzky-Golay window length are chosen by visual inspection; a reproducibility statement or machine-readable list of excluded frames would strengthen the analysis.","section":"§2.1, §2.2"},{"comment":"The conclusion that there is no statistically significant rotation period would be more informative with an upper limit on the allowed light-curve amplitude, rather than only a null periodogram comparison.","section":"§3.2"},{"comment":"The caption for Figure 9 should clarify that the lower time axis applies to 3I/ATLAS and not to 2I/Borisov; the current wording is confusing.","section":"Fig. 9"},{"comment":"There are several minor typographical and grammatical issues, including 'The top axes represents' in the Fig. 9 caption and an orphaned '/gtb' markup artifact in Section 1.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"This is a timely and useful precovery report, and the detection itself appears well supported. The main scientific claim, however, is more weakly supported than the detection: the activity inference is a cross-detector differential with an uncalibrated single zeropoint, and the Camera 1 CCD 2 epoch that carries the activity signal is also the one most affected by crowding. I recommend major revision to either quantify these systematics or soften the title and conclusions to a tentative brightness excess. The paper is otherwise within the journal's scope, and the data/code availability is a positive feature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThe thing you should know up front: the detection is the result, and it is solid. They recovered 3I/ATLAS in TESS Sector 92 with a shift-stack on deep images, and they validated the whole pipeline by recovering the known rotation period of 896 Sphinx from the same field. That gives two clean precovery epochs, 19σ and 11σ, with magnitudes and a ~20 day light curve. This is genuinely new, predating the ZTF and Rubin precovery papers.\n\nThe distant-activity interpretation is carefully hedged and mostly holds up. The first epoch alone gives HV = 13.72 ± 0.35, about 1.7 mag brighter than the HST nucleus limit of H > 15.4. Even if you take the Farnham TESS-to-V color term at its full 3σ, the first epoch still lands ~0.6 mag brighter than an inactive nucleus. So the activity claim is not as fragile as the color uncertainty alone would suggest.\n\nThe softer spot is the 1.55 mag brightening between epochs. The two epochs come from different cameras/CCDs with a single global zeropoint, and the second epoch sits in a much denser field where their by-eye crowding filter removes over half the frames. A 3x3 aperture in that stack can mix in unresolved stellar flux, and they don't quantify that risk. If the second epoch is blended, the brightening is overstated. The activity claim for the first epoch still stands, but the differential is less secure.\n\nA couple of small things: the text says 'factor of 5 flux increase' while 1.55 mag is about a factor of 4.2; fix the wording. And the frame-rejection criterion deserves a more algorithmic description than 'tuned by-eye.'\n\nThe rotation-period null result is handled well, with background-pixel controls and no overclaiming.\n\nWho is this for? Anyone working on interstellar objects or TESS precovery. It deserves a serious referee. The main requests should be a transparent frame-rejection criterion and an independent photometric check of the second-epoch stack, not rejection of the detection.\n\nI'd cite this for the precovery magnitudes.","headline":"Solid TESS precovery detection of 3I/ATLAS; the distant-activity claim is plausible and honestly hedged, but the second-epoch differential in a crowded field is the weak link.","tokens_in":19905,"tokens_out":5522,"would_cite":true,"duration_ms":60564,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"TESS precovery data suggest 3I/ATLAS was active at 6 au, months before it was discovered, with brightening that cannot be explained by distance geometry alone.","keywords":["interstellar objects","3I/ATLAS","TESS","precovery photometry","shift-stacking","cometary activity","hypervolatiles","absolute magnitude"],"falsifier":"A direct test is to obtain archival V-band (or g- and r-band) photometry of 3I/ATLAS from May-June 2025: if those measurements place it about one magnitude fainter than the TESS-derived H_V, the apparent excess over the HST nucleus limit disappears and the distant-activity argument fails. A confirming result would be a resolved coma or dust production signature in deep precovery images from that same window.","tokens_in":18741,"feed_emoji":"☄️","tokens_out":8599,"duration_ms":91395,"temperature":0.7,"pith_summary":"3I/ATLAS, the third interstellar object found crossing the Solar System, was caught on camera two months before it was discovered: TESS full-frame images from May to June 2025 show the object in two deep-stacked epochs. The paper argues that the object brightened between those epochs by about 1.55 magnitudes, roughly three times more than the changing Sun and observer distances can explain, and that its absolute brightness was higher than the limit set for a bare nucleus by HST. Taken together, this suggests 3I/ATLAS was already shedding material at heliocentric distances near 6 au, months before perihelion. If true, the activity cannot be powered by ordinary water-ice sublimation and points to more volatile ices such as carbon monoxide or carbon dioxide. The same images yield a 20-day light curve, but no statistically convincing rotation period.","feed_headline":"3I/ATLAS was shedding material by 6 au, TESS data hint","feed_subtitle":"The interstellar visitor brightened faster than geometry allows, hinting at outgassing far from the Sun.","key_machinery":"The machinery is the shift-stack precovery pipeline: predict 3I/ATLAS's pixel position in each TESS full-frame image from its known orbit, cut out a small postcard around that position, and sum thousands of cutouts into one deep image per detector. A data-driven smoothing filter removes slowly varying background structure before stacking, and contaminated frames are rejected. A 3x3-pixel aperture on the median stack supplies the flux, which is converted to a TESS magnitude with a fixed zeropoint and then to a visual absolute magnitude through the relation V = Tmag + 0.8 and the standard distance normalization. The pipeline is validated by recovering the main-belt asteroid 896 Sphinx and its known rotation period.","core_discovery":"Using a shift-stack technique on nearly ten thousand TESS frames, the authors recover 3I/ATLAS at 19 sigma on one detector and 11 sigma on another, measuring TESS magnitudes of 20.83 +/- 0.05 and 19.28 +/- 0.05. Converted to absolute visual magnitudes, these give H_V = 13.72 +/- 0.35 and 12.52 +/- 0.35, both of which are brighter than the H > 15.4 nucleus limit from HST. The observed brightening of 1.55 mag between the two epochs is about 1.1 mag larger than the roughly 0.4 mag expected from geometry alone, and the excess is interpreted as cometary activity rather than a bare, rotating nucleus. The paper explicitly notes the activity inference is tentative: the color-based magnitude conversion carries a +/- 0.3 mag uncertainty, and the 3-sigma errors allow the object to have been about a magnitude fainter in the first epoch. Its conclusion is that the data are consistent with weak pre-discovery activity, likely driven by hypervolatiles.","pith_inferences":["(Editorial) The same shift-stack treatment of other TESS ecliptic-sector data could yield precovery detections of future interstellar objects, since TESS is now observing the ecliptic plane.","(Editorial) A multi-filter precovery campaign, even a single simultaneous V-band point, would directly test the color assumption that anchors the H_V values and turn a tentative signal into a firm one.","(Editorial) If distant activity is confirmed, the interstellar-object population may be routinely volatile-rich at large heliocentric distances, which would affect models of planetesimal formation in other systems.","(Editorial) A natural follow-up is to search for nongravitational acceleration in 3I/ATLAS's orbit; activity at 6 au, if real, should leave a measurable dynamical signature."],"forward_implications":["The light curve of 3I/ATLAS now extends back to May 2025, roughly two months before discovery, at heliocentric distances of 5.5 to 6.4 au.","If the activity is real, water-ice sublimation is ruled out as the driver at these distances, and CO, CO2, or another hypervolatile must be responsible.","The brightening between the two TESS epochs indicates activity increasing as the object approached the Sun and warmed.","The 16-hour rotation-period candidate reported elsewhere is not confirmed; the TESS light curves are dominated by systematics that also appear in background pixels.","These precovery magnitudes are consistent with the brighter end of the July 2025 photometry, supporting a gradual activity evolution rather than a sudden outburst."],"supporting_citations":[{"why":"Supplies the V = Tmag + 0.8 color conversion used to derive H_V; its +/- 0.3 mag uncertainty is the dominant systematic in the activity claim.","marker":"T. L. Farnham et al. 2021"},{"why":"HST imaging gives the nucleus size limit R < 2.8 km and H > 15.4, the inactive baseline against which the bright H_V values are compared.","marker":"D. Jewitt et al. 2025"},{"why":"Provides the ZTF discovery and precovery light curve to which the new TESS points are appended in the secular light curve.","marker":"D. Z. Seligman et al. 2025"},{"why":"Reports precovery photometry and the secular light curve of 2I/Borisov used as the comparison object in Figure 9.","marker":"Q. Ye et al. 2020"},{"why":"Announces the discovery of 3I/ATLAS on 2025 July 1, defining the epoch that the TESS precovery data precede.","marker":"L. Denneau et al. 2025"},{"why":"The TESS Instrument Handbook provides the counts-to-magnitude zeropoint used in Equation 1 and its 0.05 mag error.","marker":"R. Vanderspek et al. 2018"},{"why":"Supplies the known 21.038-hour rotation period of 896 Sphinx used to validate the shift-stack and periodogram pipeline.","marker":"T. Polakis 2018"},{"why":"Proposes a rotation period near 16 hours that the TESS periodograms are compared with.","marker":"R. de la Fuente Marcos et al. 2025"},{"why":"Independently proposes a similar roughly 16-17 hour period; the TESS data cannot confirm it.","marker":"T. Santana-Ros et al. 2025"}],"fun_headline_variants":["TESS precovery reveals 3I/ATLAS active at 6 au","Interstellar object 3I/ATLAS was outgassing by 6 au","3I/ATLAS shone too bright at 6 au, hinting at cometary activity","Shift-stacked TESS images show early activity on 3I/ATLAS","TESS data suggest 3I/ATLAS was active far from the Sun"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Because the paper itself concedes the color conversion allows the first-epoch brightness to be about 1 mag fainter, the claim that 3I/ATLAS was active at 6 au stands on the assumption that the object's color matches the cometary average used to convert TESS magnitudes to V.","fun_headline_variants_meta":{"raw":{"variants":["TESS precovery reveals 3I/ATLAS active at 6 au","Interstellar object 3I/ATLAS was outgassing by 6 au","3I/ATLAS shone too bright at 6 au, hinting at cometary activity","Shift-stacked TESS images show early activity on 3I/ATLAS","TESS data suggest 3I/ATLAS was active far from the Sun"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000529,"raw_usage":{"total_tokens":2644,"prompt_tokens":1133,"completion_tokens":1511,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":749,"completion_tokens_details":{"reasoning_tokens":1401}},"tokens_in":749,"tokens_out":1511,"duration_ms":12548,"temperature":1.0,"reasoning_tokens":1401,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T12:11:28.179787+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct test is to obtain archival V-band (or g- and r-band) photometry of 3I/ATLAS from May-June 2025: if those measurements place it about one magnitude fainter than the TESS-derived H_V, the apparent excess over the HST nucleus limit disappears and the distant-activity argument fails. A confirming result would be a resolved coma or dust production signature in deep precovery images from that same window.","supporting_citations":[],"review_version":1}