The Pair-Instability Mass-Gap Question

Pair creation, pulsational pair instability (PPISN), and complete pair-instability supernovae (PISN) robustly reshape the first-generation remnant spectrum of very massive stars. Current gravitational-wave populations also show a transition or depletion beginning near $40$–$45\,M_\odot$, most clearly in secondary or low-spin component masses. At the evidence cutoff of 2026-08-20, the mechanism and a qualified population feature are supported. A universal empty interval in observed primary black-hole masses is not: hierarchical mergers and other channels can populate masses above the stellar first-generation cutoff, and flexible GWTC-5 inference does not resolve a sharp primary gap.

Helpful background. Pair Instability and Proposed Mass Gaps explains the thermodynamic instability, PPISN/PISN regimes, and mapping from stellar cores to remnants. This record assesses the current edge measurements and why “the gap” is not one distribution.

Which mass distribution has a gap

A stellar calculation may report helium-core mass, pre-collapse mass, or black-hole gravitational mass. A population analysis may report ordered primary $m_1$, ordered secondary $m_2$, an exchangeable component distribution, or a spin-conditioned subpopulation. Those are not interchangeable. Pairing and hierarchical mergers can make $m_1$ cross a first-generation cutoff even if most $m_2$ values do not.

Five axes delimit the proposition: core-to-remnant mapping; PPISN mass loss versus full PISN disruption versus collapse above the disruption range; isolated first-generation births versus all channels; component labeling; and nuclear, stellar, detector, waveform, pairing, redshift, spin, and population-model uncertainties. The useful claim is a depletion or transition in a specified population—not a forbidden universal mass interval.

Assessment at 2026-08-20	Meaning
Question status	Partially resolved
Evidence status	Supported
Confidence	Moderate because the edge-scale agreement is notable but data and model independence are incomplete
Domain	First-generation stellar remnants and selection-corrected GWTC-4/GWTC-5 binary component distributions, with explicit component and spin conditioning
What could change the conclusion	A revised stellar-physics calculation or selection-corrected population release

Entry project: distinguish a depletion from an empty interval

This is an inference project for someone comfortable with posterior samples, ordered component masses, and selection effects. Begin with population summaries rather than a new stellar-evolution grid. A figure-level reproduction is computationally modest, but the official posterior archive is large; regenerating population fits or MESA models requires separate storage, computing, and specialist preparation.

Read the LVK GWTC-5 population manuscript, P2600045-v7, §§6–7 and Appendix H, extracting exactly which distribution is plotted: primary mass, secondary mass, or a conditional population. Use LVK Zenodo 20292639 v1, README and figure notebooks to regenerate one relevant summary. Inspect the small scripts and event-list archives before downloading popsummary_files.tar.gz, which is about 25 GB. Keep the released environment and sample selection together. Then read Farmer et al. 2019, §§3–5, separating a stellar remnant edge from a detected binary mass distribution.

Deliver the regenerated plot and an explicit comparison of an empty interval, a smooth depletion, and a component-conditioned edge using the published definitions. Do not invent posterior samples for a model absent from the archive: a new model comparison needs a new likelihood calculation. The training outcome is an explanation of which feature the released inference supports and which proposed statistic it cannot answer.

A candidate extension, with novelty unchecked, is a controlled injection study measuring how mass ordering and selection turn an imposed first-generation cutoff into different primary and secondary distributions. Stop if the study omits detection efficiency, has too few effective selection samples, or assumes a hierarchical-merger fraction that determines the answer in advance. Pivot to an analytic or simulated demonstration labeled as such. Linking the inferred edge to nuclear reaction rates additionally requires an independently specified stellar-model uncertainty calculation.

What stellar evolution and populations establish

The underlying mechanism is well established. Heger and Woosley compute pulsations and complete disruption for very massive zero-metallicity helium cores, producing a remnant deficit Heger and Woosley 2002, §§ 2–4 and outcome tables. The historical grid does not fix modern boundaries, but it establishes the physical origin of the feature.

Modern MESA calculations place the lower remnant edge in the mid-$40\,M_\odot$ range for representative assumptions and demonstrate sensitivity to the $^{12}\mathrm C(\alpha,\gamma)^{16}\mathrm O$ rate, convection, winds, rotation, and stripping Farmer et al. 2019, §§ 3–5 and Figs. 3–8. “Nuclear-rate measurement from black holes” is therefore possible only after those mappings are propagated, not by equating a population break with one reaction rate.

Two GWTC-4 analyses recover remarkably similar scales using different conditioning. Tong and collaborators infer a lower boundary $44^{+5}_{-4}\,M_\odot$, visible in $m_2$ but not in $m_1$, and interpret gap-filling primaries as hierarchical Tong et al. 2026, Figs. 1–4 and Methods. Antonini and collaborators obtain $44.3^{+5.9}_{-3.5}\,M_\odot$ for a low-spin population while a high-spin component extends through the nominal gap Antonini et al. 2026, Figs. 1–4 and supplementary tests. The agreement supports a transition scale, but shared catalog data limit independence.

Results and open questions at 2026-08-20

Claim or evidence	What it supports	What it does not establish
Foundational PISN stellar models	A physical remnant-deficit mechanism	Current edge calibration
Modern MESA uncertainty grid	Mid-$40\,M_\odot$ lower-edge plausibility and nuclear sensitivity	Universal edge across stellar codes and environments
GWTC-4 $m_2$ boundary	A component-conditioned depletion near the predicted scale	Empty observed $m_1$ interval
GWTC-4 low-spin boundary	A first-generation-like cutoff plus gap-crossing high-spin population	Unique hierarchical channel or direct nuclear-rate measurement
GWTC-5 flexible population model	Steep $m_2$ decline above about $40\,M_\odot$	A resolved sharp $m_2$ gap or clear $m_1$ gap
Official population archive	Version-pinned figure and posterior reproduction	Independent population inference or stellar calculation

The theory and catalog results reinforce one another at the level of a broad transition. The stronger claim of a sharp, empty observed interval fails because it is neither predicted for a mixed population nor recovered flexibly.

Strongest competing interpretations

The pair-instability interpretation identifies the low-spin or secondary transition with the first-generation stellar cutoff. Hierarchical mergers then produce heavier, often higher-spin primaries. The concordant $44\,M_\odot$ scales make this physically appealing and potentially connect black-hole demographics to the carbon–oxygen nuclear rate.

The competing interpretation treats the feature as an emergent population transition. Pairing preferences, multiple formation channels, redshift evolution, selection, and flexible mass functions can generate a decline without a literal empty interval. Even if pair instability contributes, the observed boundary need not map one-to-one to a stellar edge. A statistical component labeled “hierarchical” is not itself a measured merger history; spin and mass overlap among channels.

Counterexamples, null results, and failed routes

The official GWTC-5 population analysis is the strongest adverse test of sharpness. Its flexible PixelPop reconstruction finds a steep decline in $m_2$ above roughly $40\,M_\odot$ but does not resolve a sharp secondary gap, while the primary distribution extends smoothly beyond $100\,M_\odot$ LVK 2026, §§ 6–7, Figs. 7–9, and Appendix H Fig. 23. This rejects the unqualified claim that observed black holes cannot occupy “the gap.” It corroborates the more careful depletion claim.

The numerical agreement between the two GWTC-4 studies is not two independent experiments: they share the catalog and some expertise. Their structured models can sharpen an edge that a nonparametric reconstruction smooths. Conversely, PixelPop regularization can blur a real discontinuity. Simulation-based coverage is needed to determine which model recovers which truth.

Translating a population edge into stellar physics

The translation should be hierarchical. A stellar code maps nuclear and mixing parameters to a first-generation remnant function. A formation model adds binaries, metallicity and redshift. A mixture model adds hierarchical and other channels. A detector model applies selection, and event likelihoods propagate mass and spin uncertainty. Only the full chain can turn an observed transition into a nuclear-rate posterior.

Component ordering matters at every stage. If hierarchical mergers preferentially occupy $m_1$, an $m_2$ cutoff may better trace first-generation births. If exchange or pairing changes labels, the interpretation changes. Reporting $m_1$, $m_2$, exchangeable, and spin-conditioned results side by side is therefore a physical test, not a presentation choice.

What would change the assessment

A blind calibration should generate no-gap, smooth-cutoff, true first-generation-gap, and hierarchically filled populations, pass them through detection and parameter estimation, and compare several physical and nonparametric models. Recovery and false-gap rates should be reported separately for $m_1$, $m_2$, exchangeable masses, and spin-conditioned components before applying the procedure to GWTC-5.

Confidence in a universal gap would require the same edges across labels and models and consistency across independently evolved stellar grids. More plausibly, a stable secondary/low-spin depletion plus a filled primary distribution would strengthen the bounded current verdict. A new stellar rate or mixing calculation that moves the edge outside the population feature would weaken the simple identification.

Research checklist

• State whether masses are core, pre-collapse, remnant, source-frame, detector-frame, primary, secondary, exchangeable, or spin conditioned.
• Pin event selection, parameter-estimation release, selection injections, waveform model, redshift law, pairing function, and priors.
• Compare a smooth depletion, a sharp gap, and physical mixture models; use posterior-predictive and simulation-based coverage checks.
• Test $m_1$, $m_2$, exchangeable, and spin-conditioned distributions separately.
• Track shared events and authors when judging independence among population analyses.
• Pin stellar code, reaction-rate distribution, convection, winds, rotation, metallicity, and binary stripping.
• Propagate the inferred population feature through several stellar codes rather than one fit.
• No independent computational reproduction was performed for this assessment; the official notebooks reproduce released analyses, not an external model.

Source and data map

Heger and Woosley establish the mechanism; Farmer et al. map modern stellar uncertainties. Tong et al. and Antonini et al. provide the two structured GWTC-4 edge measurements. The current official adverse and confirmatory result is the GWTC-5 population manuscript DCC P2600045-v7, which remained an arXiv/DCC record at the cutoff.

The exact analysis object is Zenodo record 20292639 v1, including popsummary_files.tar.gz, scripts.tar.gz, Event_list.tar.gz, README, and hashes. Its environment pins Python 3.11, NumPy 2.3.2, SciPy 1.16.1, Matplotlib 3.10.5, Bilby 2.6.0, popsummary 0.1.0, and zenodo-get 2.0.0. The roughly 25.4 GB package supports published figures and posterior comparisons; a separate MESA or other stellar calculation is required for the edge mapping. For a new event-level rerun, GWOSC GWTC-5.0, DOI 10.7935/bk00-6a89 points to stable-release-9 parameter-estimation records 20348005 and 20348006 and candidate record 20348004. Those newer event products cannot be swapped into the v1 population package without regenerating or validating the corresponding likelihood and selection calculation.

References

• Antonini, F., et al. “Gravitational-wave constraints on the pair-instability mass gap and nuclear burning in massive stars.” Nature Astronomy 10, 1049–1056 (2026). DOI.
• Farmer, R., Renzo, M., de Mink, S. E., Marchant, P., and Justham, S. “Mind the gap: the location of the lower edge of the pair-instability supernova black hole mass gap.” Astrophysical Journal 887, 53 (2019). DOI.
• Heger, A. and Woosley, S. E. “The nucleosynthetic signature of population III.” Astrophysical Journal 567, 532–543 (2002). DOI.
• LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration. “GWTC-5.0: Population Properties of Merging Compact Binaries.” Data release v1, Zenodo (2026). DOI.
• LIGO Scientific Collaboration, Virgo Collaboration, and KAGRA Collaboration. “GWTC-5.0: Population Properties of Merging Compact Binaries.” DCC P2600045-v7; arXiv:2605.27226 (2026). Official manuscript.
• Tong, H., et al. “Evidence of the pair-instability gap from black-hole masses.” Nature 652, 874–877 (2026). DOI.