Chemical fingerprints of binary mass transfer in massive stars

Abstract: The majority of massive stars are born in close binary systems. As stars expand when they age, mass transfer or even a merger with their companion is inevitable. However, most binary interaction products appear as single stars, such that the main evidence of their exciting past is lost. In a comprehensive grid of detailed massive binary evolution models we find systematic trends in chemical surface abundances that allow identifying the past mass gainers. We develop an analytic framework which is independent of specific evolutionary models, to constrain the amount and composition of the accreted material from their observed surface abundances. This yields tight constraints on the uncertain mass transfer physics in massive binary stars and allows us to reconstruct the past evolutionary history of the progenitor binary system. This method, which is shown to also constrain binary mergers (for example, SN 1987A), is applied to some of the best-studied OB stars so far. For γ Columbae, suggested to be an envelope-stripped star, we show that it is a mass gainer instead, whose companion star likely formed a stripped-envelope supernova. Our results highlight surface abundance measurements as a powerful tool to improve our understanding of massive binary systems evolving towards supernovae and compact object binaries.

• The paper establishes that mass gainers occupy a distinct N/C–N/O abundance region near the CN-equilibrium dilution line, separating them from stripped stars, single-star products, and donors.
• The paper combines CNO and helium–nitrogen abundances with binary-evolution constraints to infer accreted material, mass ratios, and transfer efficiencies, successfully recovering the history of a modeled post-transfer star.
• The paper identifies γ Columbae as a likely mass gainer rather than a stripped star and shows that SN 1987A’s abundances are consistent with a rapidly mixed merger product, while highlighting uncertainties in thermohaline mixing and accretion efficiency.

Motivation and problem statement

Roughly 70% of massive stars are born in binaries close enough that mass exchange is inevitable, largely independent of metallicity. Yet the main mass transfer episode occupies less than ~0.1% of a massive star's lifetime, so it is almost never caught observationally, and full hydrodynamic simulations remain computationally prohibitive. As a result, the stability and efficiency of mass transfer in massive binaries — parameters that govern whether systems evolve toward stripped-envelope supernovae, mergers, or compact-object binaries detectable by LIGO — remain weakly constrained. Compounding the difficulty, most mass gainers and merger products appear as single stars, having lost their companion, merged with it, or left behind only a faint remnant. Up to ~30% of core hydrogen burning stars and ~70% of core helium burning stars are expected to carry a binary interaction history.

The paper addresses the diagnostic degeneracy among the three known sources of CNO surface enrichment in massive stars: wind stripping, internal mixing in single stars, and binary accretion. Its central result is that binary mass gainers occupy a distinct region of the N/C–N/O plane, providing an observational fingerprint that identifies past accretors and quantitatively constrains their accretion history.

The CNO fingerprint from detailed binary models

Jin and Langer employ a comprehensive grid of MESA binary evolution models including mass and angular momentum transfer, differential rotation, tides, and an extended nuclear network tracking all stable CNO isotopes through hydrogen burning. Plotting surface N/C against N/O during core helium burning reveals that models populate only two branches within the region bounded by the analytic "CN-eq. + dilution" and "CNO-eq. + dilution" lines:

• Mass donors and self-stripped stars lie near the "CNO-eq. + dilution" line, since envelope stripping exposes convective-core material at the high temperatures where full CNO equilibrium holds.
• Mass gainers populate a branch near the "CN-eq. + dilution" line, elevated above the "CNO-eq. + dilution" line — a region avoided by donors and single stars.

The physical explanation rests on the temperature stratification above the convective core: a thin layer at ~15–17 MK reaches CN-equilibrium but not CNO-equilibrium. After accretion, helium enrichment of the gainer's envelope drives fast thermohaline mixing on a thermal timescale (~10⁴ yr), homogenizing the accreted material before significant CN-processing can occur. Subsequently, as the inverted composition profile smooths out, "slow mixing" (assisted by rotation) brings unprocessed carbon into the CN-cycling layer at the envelope base, raising N/C while leaving N/O moderate. The authors note plainly that thermohaline mixing efficiency in massive stars has never been empirically calibrated, so the elevation of gainer models above the "CNO-eq. + dilution" line carries model uncertainty — though the framework they derive is designed to be robust to it.

An analytic, model-independent framework

The authors derive closed-form expressions for limiting lines in two diagnostic planes: the CNO diagram (N/C vs. N/O) and the HeN diagram (surface helium mass fraction vs. nitrogen enhancement). The key quantities are the dilution factor $f_{\rm CNO}$ — the fraction of the gainer's envelope consisting of accreted CNO-equilibrium material from the donor's H/He gradient layer — and $Y_{\rm CNO}$, the average helium abundance of that material. Notably, the HeN diagram alone yields only upper limits on $f_{\rm CNO}$ because subsequent CN-cycling also enhances nitrogen; combining both planes lifts this ambiguity and breaks the degeneracy with rotating single-star models.

Because three measured abundances must constrain five unknowns, additional constraints come from single-star core-mass relations (linking accreted CNO-equilibrium mass to donor initial mass), the current gainer mass, supernova timing, and the stability criterion for mass transfer. The framework is validated against a "Mock star": a post-mass-transfer model from an initially 22.4 + 7.8 $M_\odot$, 10-day binary, whose accreted masses, $f_{\rm CNO}$, and progenitor configuration are successfully recovered. Importantly, the inferred accretion efficiencies generally differ from those assumed in the underlying MESA grid, making them empirical benchmarks for future numerical models rather than circular restatements of model input.

Application to observed OB stars

Applying the method to four well-studied enriched OB stars yields concrete reconstructions:

Star	$f_{\rm CNO}$	$Y_{\rm CNO}$	Initial mass ratio $q_i$	Overall $\beta$
HD 48279	22%	0.8	0.38–0.79	0.25–0.82
$\gamma$ Columbae	17%	~0.9	0.26–0.35	0.08–0.16
HD 93840 (F)	32%	0.4–0.7	0.31–0.48	0.30–0.53
$Y_{\rm CNO}$0 Ophiuchi	~0%	unconstrained	0.45–0.95	~0.00

The most consequential case is $Y_{\rm CNO}$1 Columbae, previously proposed as a recently stripped pulsating core. The authors show instead that its surface composition reflects nearly undiluted CN-equilibrium abundances — the signature of a mass gainer, not a donor. About 0.8 $Y_{\rm CNO}$2 of matter with average helium mass fraction ~90% was accreted; the donor must have had an initial mass of at least ~14–15 $Y_{\rm CNO}$3, implying an initial mass ratio below 0.35 and highly non-conservative transfer. The stripped-star interpretation faces three independent problems the gainer scenario avoids: thermal disequilibrium at the observed HRD position, the absence of an expected bright companion, and normal surface gravity inconsistent with stripping. Under the gainer scenario, the companion likely exploded as a stripped-envelope supernova.

For HD 93840, one of two published parameter sets admits no solution; the other requires a strongly non-conservative Case B event. For $Y_{\rm CNO}$4 Ophiuchi, $Y_{\rm CNO}$5 implies zero late-phase accretion efficiency, which the authors suggest may reflect rapid rotation hindering late accretion. Across all cases, the observed stars span a larger parameter space than the model grid covers, which the authors attribute to the fixed (and uncertain) assumptions about accretion and thermohaline efficiencies — arguing that larger samples will tighten these physics inputs.

Distinguishing mergers: SN 1987A

Post-main-sequence merger products, whose envelopes mix rapidly on dynamical timescales, should show strong helium enhancement and lie near the "CNO-eq. + dilution" line without a CN-cycling component — distinguishing them from stable mass gainers. Applied to SN 1987A using circumstellar ring abundances (after correcting for LMC initial composition), the method confirms diluted CNO-equilibrium composition with no discernible CN contribution, consistent with rapid post-merger mixing. With $Y_{\rm CNO}$6, the analysis requires 6–8 $Y_{\rm CNO}$7 of pristine material lost during the merger and a companion of ~2.5 $Y_{\rm CNO}$8 merging with a primary above ~17 $Y_{\rm CNO}$9. Incorporating helium-core material would widen the solution space but is not required by the data. The caveat that slow post-merger envelope mixing cannot be fundamentally excluded applies here.

Limitations and open questions

The framework's scope is deliberately restricted: it targets the dominant channel (~63% of stable mass transfer events) of single-episode Case B transfer onto main-sequence secondaries, though the authors argue extension to Cases A and C is straightforward where transferred material is pristine or hydrogen-burning product. Several assumptions bear directly on the quantitative results. Complete rejuvenation is assumed, so inefficient core growth would inflate inferred envelope masses and shift derived accreted masses. Solar initial abundances are adopted throughout; defensible for the nearby $f_{\rm CNO}$0 Columbae and $f_{\rm CNO}$1 Ophiuchi, but untested for the more distant objects, and the authors do not attempt per-object abundance constraints given measurement uncertainties — indeed, adopting a higher N/O value for $f_{\rm CNO}$2 Columbae materially changes its allowed progenitor parameter space. Mass accretion efficiency in the grid is rotationally limited, so higher-efficiency channels may populate regions of the diagnostic planes not covered here. Donor models sit slightly above the "CNO-eq. + dilution" line due to shell-burning convective mixing, an effect neglected in the main analysis. Finally, the interplay of thermohaline and rotational mixing remains incompletely understood, and the Geneva single-star grids overlap considerably with the gainer models in the HeN plane — degeneracy broken only when CNO ratios are included.

Conclusion

This work establishes that the joint use of N/C–N/O and helium–nitrogen surface abundances provides a quantitative, largely model-independent diagnostic of past binary mass accretion in apparently single massive stars. It overturns the stripped-star interpretation of $f_{\rm CNO}$3 Columbae, constrains accretion efficiencies for several benchmark OB stars, and extends naturally to merger products as demonstrated for SN 1987A. The immediate open question the paper poses is empirical: systematic, high-precision abundance measurements across larger samples are needed to calibrate thermohaline mixing and accretion efficiency, and thereby to sharpen predictions for the supernovae and compact-object binaries that massive binaries produce.