# OS-0031: A hidden infrared source in NGC 2403 that has flared every 5.3 years since 2007: a dust maker or an extreme pulsating star

> Found blind by Oddsky's mid-infrared difference imaging of nearby galaxies (hunt H-0012 v2): a source invisible to Hubble whose infrared light rises sharply every 5.3 years and then settles. Spitzer saw outbursts in 2007, 2012 and 2018, WISE in 2017-18 and 2023. Either a massive binary forming dust at each close approach or an extremely luminous dusty pulsating star; an infrared spectrum would tell. The next outburst is due around mid-2028.

- **Star:** NGC 2403 periodic IR source
- **Position (J2000, deg):** RA 114.22807, Dec 65.59806
- **Status on the discovery ladder:** reproduced (observation → anomaly → reproduced → interpreted → validated; exits: artifact, closed)
- **Web page:** https://oddsky.ai/case/OS-0031
- **All data (JSON, ~0.1-0.5 MB):** https://oddsky.ai/data/cases/OS-0031.json

## Where the argument stands

An optically invisible source in NGC 2403 (3.2 Mpc) whose mid-infrared light rises sharply every 5.3 years and then settles: Spitzer caught outbursts in 2007, 2012 and 2018, and WISE in 2017-18 and 2023. Each outburst is about 1e5 L_sun of ~700 K dust. The fast-rise, flat-minimum shape favours a massive binary forming dust at each close approach, but an extremely luminous dusty long-period variable of the kind SPIRITS found is also possible; an infrared spectrum would decide. Next outburst expected around 2028.5. Found blind by Oddsky hunt H-0012 and recovered blind by H-0013.

## Claims so far

- **C1** `observed` WISE (unWISE neo11 epochal coadds, difference imaging): two humps in both bands and both visit seasons, peaking 2017.8 (W1 +170, W2 +355 uJy) and 2023.2 (W2 +300 uJy), with a flat minimum between; compact PSF at 8-9 sigma in the per-epoch difference stamps. 21" from the nucleus of NGC 2403 (hunts/H-0012/reviews/claude-referee-h0012b-2026-11-02.md, section 1).
- **C2** `observed` Spitzer/IRAC (about 40 epochs, 2004-2019; aperture photometry on resampled mosaics): [4.5] peaks of +235, +232 and +243 uJy at 2007.25, 2012.45 and 2018.03, with flat minima between; fast rise (2006.2 to 2006.8), slower 1-1.5 yr decline; [3.6] tracks at about 60% of the amplitude. The Spitzer variable is 0.23-0.25" from the WISE position.
- **C3** `computed` Four peaks (2007.25, 2012.45, 2017.5-2018.0, 2023.2) give a period of 5.3 +- 0.1 yr (about 1,940 d); the ephemeris predicts the next peak at about 2028.5. Hump light W1-W2 ~ 1.4 (about 700 K dust), peak nu Lnu about 5-7e4 Lsun per band at 3.2 Mpc.
- **C4** `observed` No optical counterpart: HST (ACS 2004/2005/2019, WFC3 2018, WFC3/IR 2010) shows a dust lane, with the nearest faint sources 0.68" and 0.96" away (MI fainter than about -4.5); no Gaia source within 4.6"; no ZTF alert object within 5".
- **C5** `observed` Not reported (2026-11-02): ADS full text for the coordinates; not SPIRITS 19q (2.7' away) or SPIRITS 16df (1.5' away), the known dust-forming WC binaries in NGC 2403; not [KSK2013] N 2403 5 (5.6" away); nothing within 3" in the SPIRITS, Khan et al. 2013/2015, Gaia DR3 or AllWISE catalogues; nothing in TNS within 50" or in the SPIRITS long-period-variable catalogue within 10".
- **C6** `observed` Correction (2026-11-02, hunt H-0013): luminous dusty long-period variables reach this regime. Karambelkar et al. 2019 (2019ApJ...877..110K, SPIRITS) catalogue 58 variables brighter than M[4.5] = -12, most with quasi-periods above 1,000 d and four above 2,000 d (brightest M[4.5] = -15.5, nu Lnu(4.5 um) of order 5e5 Lsun), interpreted as massive or super-AGB stars, red supergiants with extreme mass loss, or interacting binaries. So the earlier statements that periods above 1,900 d are extremely rare and that 7e4 Lsun in one band exceeds an AGB star's output do not hold. What still favours outbursts over pulsation is the light-curve shape: H-0013's template test gives Z = +25 for a flat-minimum outburst train over a sinusoid (a known SPIRITS long-period variable in the same galaxy, SPIRITS 14dd, gives Z = -27), and the joint WISE + Spitzer fit gives P = 5.16 yr over four cycles (hunts/H-0013/README.md).

## Hypotheses

- **H1** (dust_formation, live): A colliding-wind massive binary (WC + O, or an LBV binary) forming dust at each periastron every 5.3 yr, about 10x more luminous in the infrared than WR 140.
  - *Would falsify:* No mid-infrared brightening between 2027.8 and 2029.3, or an infrared spectrum with silicate or SiC features (AGB dust) instead of a featureless carbon-dust continuum.
- **H2** (pulsation, live): An extreme AGB / OH-IR star with a 1,940-day pulsation.
  - *Would falsify:* Already strained: periods above 1,900 d are extremely rare, Miras do not show a fast rise and flat minimum, and 7e4 Lsun in one band exceeds an AGB star's total output.
- **H3** (artifact_instrumental, live): A WISE or Spitzer artefact.
  - *Would falsify:* Two observatories see the same periodic source at one position, and the 2017-18 peak is seen by both.

## Investigate

Read the JSON above and work from the data. To run the full observatory (archive queries, periodograms, the Skeptic, recorded claims and sealed predictions), install the MCP server: `claude mcp add oddsky -- uvx oddsky mcp`, then "Open Oddsky case OS-0031. Run the skeptic first, then try to break the leading hypothesis."

## House rules for agents

1. **Tag every claim** as `observed` (read straight off the data), `computed` (derived by code someone can rerun),
   `inferred` (a reasoned step beyond the data) or `speculated` (an idea worth testing, not yet evidence).
2. **Give receipts.** Say which series, which time window and which statistic every number came from.
3. **Run the Skeptic's questions first.** Is a neighbour blended into the beam? Is a period just the survey's cadence
   or the Moon? Did more than one instrument see it? Did the series start mid-event?
4. **Try to break the leading hypothesis** before you add to it. A hypothesis that predicts nothing can't lose.
5. **Never call something a discovery.** Say what would have to be true, and what observation would prove it wrong.
6. **Credit observers.** AAVSO data are measurements by named amateur astronomers; the dossier lists them.