M31 (the Andromeda galaxy), with a star formation rate of 0.35–0.4 solar masses per year, is predicted to host 320–700 supernova remnants (SNRs) younger than roughly 20,000 years, yet only about 120–130 are known. Recent optical spectroscopy confirms the existing catalog but shows that the searches are biased against specific supernova types: Type Ia remnants expanding into low-density interstellar or halo gas produce faint Balmer-dominated filaments that fail the standard [S II]/Hα classification test, and most core-collapse supernovae explode inside superbubbles where their remnants are dim and easily confused with bubble walls. The open problem is to resolve this census deficit: either recover the missing populations through deeper radio, X-ray, ultraviolet and optical spectroscopic observations, or establish quantitatively that they radiate below any feasible detection threshold — in which case extragalactic SNR counts must be reinterpreted as probes of galactic gas environments rather than of supernova rates.

Background

Supernova remnants (SNRs) are the expanding nebulae of shocked gas created when supernova ejecta collide with the surrounding interstellar medium (ISM). Because a remnant radiates detectably for only some tens of thousands of years before blending into the ISM, the number of SNRs in a galaxy is tied to its recent supernova rate, making SNR censuses a potential proxy for star formation and supernova activity.

Identifying SNRs in external galaxies traditionally rests on optical diagnostics. Shock-heated gas cools through collisionally excited forbidden lines, so the ratio of the doubly ionized sulfur lines [S II] λλ6716, 6731 to the hydrogen recombination line Hα exceeds roughly 0.4 in radiative shocks, whereas photoionized H II regions show values near 0.1. Kinematics provide a second test: shocks move at 100–300 km s−1^{-1}, so the Hα line profile of an SNR, measured as the full width at 20% of peak flux (FW20), exceeds about 50 km s−1^{-1}, while H II regions expand at only 10–20 km s−1^{-1}. Independent confirmation comes from non-thermal radio synchrotron shells, soft thermal X-rays, and ultraviolet emission. Integral-field-unit (IFU) spectrographs — instruments that record a spectrum at every spatial pixel, such as MUSE on the VLT and SITELLE at CFHT — now combine all of these diagnostics in a single observation.

M31, the nearest large spiral at a distance of about 780 kpc, has a star formation rate of 0.35–0.4 M⊙M_\odot yr−1^{-1}. Lee & Lee (2014) identified 156 SNR candidates there through narrowband imaging for elevated [S II]/Hα. Caldwell, Raymond, Long & Lee (2025) obtained fiber-fed spectroscopy of 152 of these with the Hectospec (resolution R≈1000R \approx 1000) and Hectochelle (R≈32000R \approx 32000) instruments on the MMT: 104 are confirmed SNRs, 33 are probably not SNRs (mostly H II regions), and only about 50 have supporting X-ray, ultraviolet or radio detections. The 1.5″ fibers subtend about 5.5 pc and sample only portions of most remnants, and many high-resolution line profiles are non-Gaussian — narrow photoionized cores on broad shock-excited wings — so mapping observed profiles to shock speeds and three-dimensional geometry remains approximate. The classification thresholds themselves ([S II]/Hα = 0.4, FW20 = 50 km s−1^{-1}) are operational conventions with no sharp physical boundary.

The central quantitative fact is the census deficit. Combining M31's star formation rate with supernova rates measured per unit stellar mass (Li et al. 2011) predicts 320–700 SNRs younger than about 20,000 years; the detected sample of roughly 120–130 therefore represents only 20–40%. The missing objects are precisely the classes that fade in low-density gas. Type Ia supernovae — thermonuclear explosions of white dwarfs, which occur far from star-forming regions — leave remnants that expand into tenuous ISM or halo gas and shine mainly through Balmer-dominated emission (hydrogen lines excited by a collisionless shock with little metal-line cooling, as in the Galactic remnants SN 1006 and Tycho), with X-ray luminosities of order 103410^{34} erg s−1^{-1}; they fail the [S II] criterion by construction. About 80% of core-collapse supernovae explode inside superbubbles — cavities blown in the ISM by stellar winds and earlier supernovae (Lingenfelter 2019) — where their remnants are faint and easily confused with bubble walls. Deep narrowband imaging keeps revealing very large, extremely faint SNRs in the Milky Way's halo (Fesen et al. 2024) that would be undetectable in M31, and the Galactic catalog itself (Green 2025: 310 known versus roughly 1000 expected remnants) shows the deficit is generic.

The observational landscape is changing. IFU surveys have resolved SNR censuses in more distant but angularly compact galaxies — about 2200 new MUSE candidates across 19 PHANGS galaxies (Li et al. 2024), kinematic classification in M33 (Duarte Puertas et al. 2024), and the recovery of rare broad-line oxygen- and nitrogen-rich young remnants (Kravtsov et al. 2025) — but M31 spans more than 3 degrees on the sky, impractical to mosaic with existing IFUs. On the radio side, LOFAR 150 MHz imaging of M31's center (Venkattu et al. 2024) and the new VLA Local Group L-band Survey, LGLBS, with continuum resolution under 2″ (Koch et al. 2025), provide the sharpest radio views yet, with SNR-specific catalogs still pending; X-ray depth still rests on XMM-Newton surveys (Sasaki et al. 2012; Huang et al. 2023). Caldwell et al. (2025) close with the interpretive crux: given well-measured star formation and supernova rates, SNR counts "tells us more about the galactic environments in which we can detect SNRs" than about supernova rates themselves.

Problem Statement

Resolve the supernova-remnant census deficit of M31 (the Andromeda galaxy). About 120–130 remnants and candidates are known, against 320–700 expected from the galaxy's measured star-formation rate (0.35–0.4 M⊙M_\odot yr−1^{-1}) and supernova rates for remnants younger than roughly 20,000 years. The problem is to determine where the missing 60–80% of the expected population are, by either: (1) recovering them observationally — detecting and classifying the predicted but undetected remnants, above all Type Ia remnants (remnants of thermonuclear explosions of white dwarfs) expanding into low-density interstellar or halo gas and core-collapse remnants inside superbubbles (cavities blown in the interstellar gas by stellar winds and earlier supernovae), through deeper radio (e.g., LOFAR, the Low-Frequency Array, and VLA Local Group L-band survey imaging), X-ray, ultraviolet and optical spectroscopic observations, using criteria appropriate to Balmer-dominated remnants (whose emission is almost purely hydrogen lines, with little metal-line cooling) and other metal-line-poor remnants that the standard optical shock tests — [S II]/Hα > 0.4 (the ratio of the doubly ionized sulfur forbidden doublet to the hydrogen Balmer line Hα) and FW20 > 50 km s−1^{-1} (the Hα full width at 20% of peak flux) — exclude by construction; or (2) establishing quantitatively, anchored to M31's independently measured gas distribution, star-formation rate and supernova rates, that the missing remnants radiate below the surface-brightness limits of the deepest surveys of M31 that are feasible with existing or already-committed facilities — concretely, the Hα narrowband detection depth reached by Lee & Lee (2014) and the sensitivities of the LOFAR 150 MHz (Venkattu et al. 2024), VLA Local Group L-band (Koch et al. 2025) and XMM-Newton (Sasaki et al. 2012; Huang et al. 2023) observations, with any extrapolation to deeper future facilities stated explicitly — so that the deficit is intrinsic to the low-density environments in which those supernovae explode rather than a selection effect. Either branch alone resolves the task. A complete answer on that branch states, under branch (1), the number of remnants recovered and classified within the roughly 20,000-year visibility window — the observationally detectable count, distinct from the galaxy's intrinsic remnant population — or, under branch (2), the total population implied by that same supernova-rate modeling together with the bound on the fraction of it lying below the adopted sensitivity floor; it identifies which missing classes are recoverable in principle and with what instrumentation under branch (1), or which of the remaining classes are provably undetectable below the sensitivity floor set out under branch (2), and in either branch fixes what fraction of the deficit is a survey selection effect versus a physical consequence of remnants fading in low-density gas — thereby determining whether extragalactic SNR counts can be used to infer supernova rates, or instead measure the distribution of detectable environments.

References

  1. Caldwell, N., Raymond, J. C., Long, K. S., & Lee, M. G. 2025, 'Spectroscopy of Supernova Remnants and Candidates in M31', The Astrophysical Journal, 983, 150, doi:10.3847/1538-4357/adbf98, arXiv:2503.15423, https://arxiv.org/abs/2503.15423
  2. Lee, J. H., & Lee, M. G. 2014, 'A New Optical Survey of Supernova Remnant Candidates in M31', The Astrophysical Journal, 786, 130, doi:10.1088/0004-637X/786/2/130, https://doi.org/10.1088/0004-637X/786/2/130
  3. Li, W., Chornock, R., Leaman, J., et al. 2011, 'Nearby supernova rates from the Lick Observatory Supernova Search - III. The rate-size relation, and the rates as a function of galaxy Hubble type and colour', Monthly Notices of the Royal Astronomical Society, 412, 1473, doi:10.1111/j.1365-2966.2011.18162.x, https://doi.org/10.1111/j.1365-2966.2011.18162.x
  4. Rahmani, S., Lianou, S., & Barmby, P. 2016, 'Star formation laws in the Andromeda galaxy: gas, stars, metals and the surface density of star formation', Monthly Notices of the Royal Astronomical Society, 456, 4128, doi:10.1093/mnras/stv2951, https://doi.org/10.1093/mnras/stv2951
  5. Lingenfelter, R. E. 2019, 'The Origin of Cosmic Rays: How Their Composition Defines Their Sources and Sites and the Processes of Their Mixing, Injection, and Acceleration', The Astrophysical Journal Supplement Series, 245, 30, doi:10.3847/1538-4365/ab4b58, https://doi.org/10.3847/1538-4365/ab4b58
  6. Fesen, R. A., Drechsler, M., Strottner, X., et al. 2024, 'Deep Optical Emission-line Images of Nine Known and Three New Galactic Supernova Remnants', The Astrophysical Journal Supplement Series, 272, 36, doi:10.3847/1538-4365/ad410a, https://doi.org/10.3847/1538-4365/ad410a
  7. Green, D. A. 2025, 'An updated catalogue of 310 Galactic supernova remnants and their statistical properties', Journal of Astrophysics and Astronomy, 46, 14, doi:10.1007/s12036-024-10038-4, https://doi.org/10.1007/s12036-024-10038-4
  8. Sasaki, M., Pietsch, W., Haberl, F., et al. 2012, 'Supernova remnants and candidates detected in the XMM-Newton M 31 large survey', Astronomy & Astrophysics, 544, A144, doi:10.1051/0004-6361/201219025, https://doi.org/10.1051/0004-6361/201219025
  9. Huang, R., Li, J.-T., Cui, W., et al. 2023, 'An XMM-Newton View of the Andromeda Galaxy as Explored in a Legacy Survey (New-ANGELS). I. The X-Ray Source Catalog', The Astrophysical Journal Supplement Series, 268, 36, doi:10.3847/1538-4365/ace4b8, https://doi.org/10.3847/1538-4365/ace4b8
  10. Venkattu, D., Lundqvist, P., Pérez Torres, M., et al. 2024, 'SN 1885A and Supernova Remnants in the Centre of M31 with LOFAR', The Astrophysical Journal, 976, 213, doi:10.3847/1538-4357/ad890f, arXiv:2410.22073, https://arxiv.org/abs/2410.22073
  11. Koch, E. W., Leroy, A. K., Rosolowsky, E., et al. 2025, 'The Karl G. Jansky Very Large Array Local Group L-band Survey (LGLBS)', The Astrophysical Journal Supplement Series (in press), arXiv:2506.11792, https://arxiv.org/abs/2506.11792
  12. Li, J., Kreckel, K., Sarbadhicary, S., et al. 2024, 'Discovery of ~2200 new supernova remnants in 19 nearby star-forming galaxies with MUSE spectroscopy', Astronomy & Astrophysics, 690, A161, doi:10.1051/0004-6361/202450730, arXiv:2405.08974, https://arxiv.org/abs/2405.08974
  13. Duarte Puertas, S., Drissen, L., Robert, C., et al. 2024, 'Properties of supernova remnants in SIGNALS galaxies – I. NGC 6822 and M33', Monthly Notices of the Royal Astronomical Society, 533, 2677, doi:10.1093/mnras/stae1641, arXiv:2407.01153, https://arxiv.org/abs/2407.01153
  14. Kravtsov, T., Anderson, J. P., Kuncarayakti, H., Maeda, K., & Mattila, S. 2025, 'Discovery of young, oxygen-rich supernova remnants in PHANGS-MUSE galaxies', Astronomy & Astrophysics, 700, A223, doi:10.1051/0004-6361/202349083, arXiv:2409.06504, https://arxiv.org/abs/2409.06504
  15. Raymond, J. C., Caldwell, N., & Long, K. S. 2026, 'Shock Wave Precursors in M33 Supernova Remnants', The Astrophysical Journal, 998, 13, doi:10.3847/1538-4357/ae3752, https://doi.org/10.3847/1538-4357/ae3752