Gaia astrometry disfavors a binary origin for long secondary periods

Cheyanne Shariat; Emily Leiner; Kareem El-Badry; Morgan MacLeod

arxiv: 2604.09767 · v1 · submitted 2026-04-10 · 🌌 astro-ph.SR

Gaia astrometry disfavors a binary origin for long secondary periods

Cheyanne Shariat , Kareem El-Badry , Morgan MacLeod , Emily Leiner This is my paper

Pith reviewed 2026-05-10 17:27 UTC · model grok-4.3

classification 🌌 astro-ph.SR

keywords long secondary periodsred giant starsGaia astrometrybinary companionsRUWEstellar variabilitybrown dwarf desert

0 comments

The pith

Gaia astrometry shows most long secondary periods in red giants are not caused by binary companions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tests the leading binary hypothesis for long secondary periods, which are stable photometric cycles of months to years seen in about one-third of luminous red giant stars. It combines Gaia radial velocities with forward modeling of the expected astrometric wobble to predict that 0.1 solar-mass companions at 1-3 AU would produce systematically high RUWE values. The observed RUWE for nearby LSP stars is instead low and consistent with single stars. This discrepancy removes low-mass companions as the main explanation and points toward intrinsic stellar processes.

Core claim

Interpreting the radial-velocity variability of LSP stars as orbital motion requires companions narrowly clustered near 0.1 solar masses at separations of 1-3 AU. Forward modeling of the astrometric signature these companions would produce in Gaia DR3 predicts elevated RUWE values. The actual RUWE distribution for nearby LSP stars lies systematically below those predictions and matches expectations for single stars, disfavoring low-mass stellar or substellar companions as the dominant origin of LSPs.

What carries the argument

Forward modeling of Gaia DR3 astrometric signatures to predict RUWE for binary systems whose masses and separations are inferred from observed radial-velocity amplitudes.

If this is right

The brown-dwarf desert around solar-type stars extends to the red-giant phase for the relevant separations.
Most LSP stars should show no detectable astrometric or spectroscopic signature of a companion.
Alternative single-star mechanisms must account for both the photometric LSP and the RV modulation.
The fraction of true binaries among LSP stars is much lower than previously inferred from RV data alone.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Similar RUWE tests can be applied to other classes of variables whose RV signals have been attributed to companions.
Pulsation models for red giants will need to incorporate a mechanism that produces stable multi-year photometric cycles without requiring a companion.
Higher-precision astrometry from future missions could tighten the upper limit on any residual binary fraction.

Load-bearing premise

That the observed radial-velocity variability is produced by orbital motion of a companion.

What would settle it

A sample of LSP stars in which the measured RUWE values match the elevated predictions calculated for 0.1-solar-mass companions at 1-3 AU, or direct imaging or radial-velocity confirmation of such companions.

Figures

Figures reproduced from arXiv: 2604.09767 by Cheyanne Shariat, Emily Leiner, Kareem El-Badry, Morgan MacLeod.

**Figure 1.** Figure 1: Color-magnitude diagram of the LSP samples used in this work. We show 1.5 kpc LSPs from the Gaia Focused Product Release with RV time-series (red), which constitutes the primary sample used in this work, as well as the ASAS–SN LSP sample (blue). The cleaned Gaia 100 pc sample is shown for comparison (black). All Gaia photometry is dust-corrected. Gaia RV-selected sample, especially at lower luminosities … view at source ↗

**Figure 3.** Figure 3: RV semi-amplitude (left) and companion mass (right) probability distribution of LSPs in the Gaia 1.5 kpc sample (N = 224). On the left, we show the K = 0.75 km s−1 line, above which our sample is roughly complete. On the right, the companion mass, M2, is derived from K assuming a solar-mass primary and circular orbit. We show the result for both an isotropic inclination distribution (solid black) and one r… view at source ↗

**Figure 4.** Figure 4: Observed vs predicted RUWE for LSP stars assuming the binary hypothesis. The left panel shows the observed Gaia RUWE while the right panel shows the predicted RUWE assuming the RV variability is due to a binary companion. We display results for three different samples, restricted to 1.5 kpc: Gaia LSPs (top), ASAS–SN LSPs (middle), and Gaia ellipsoidal variables (ELL; bottom). Note that the ASAS–SN sample … view at source ↗

**Figure 5.** Figure 5: Comparison of LSP and non-LSP pulsators in the Gaia FPR sample. Left: dust-corrected Gaia CMD for nearby LSPs (red) and non-LSP pulsators (blue), shown together with the cleaned Gaia 100 pc sample (black) for reference. Right: observed RUWE versus distance for the same LSP and non-LSP pulsator samples. Both LSP and non-LSP pulsators show similar RUWE values concentrated near RUWE ≈ 1, disfavoring a model w… view at source ↗

**Figure 6.** Figure 6 [PITH_FULL_IMAGE:figures/full_fig_p013_6.png] view at source ↗

**Figure 7.** Figure 7: RUWE predictions for alternative binary-model assumptions. The top row shows DR3 forward models for eccentric binaries with isotropic inclinations, assuming either fixed e = 0.3 (left) or e ∼ U(0.1, 0.6) (right). The bottom row shows the corresponding DR3 predictions when the same eccentricity prescriptions are combined with an edge-on inclination prior, 45◦ < i < 135◦ . In each panel, the left-hand column… view at source ↗

**Figure 8.** Figure 8: RUWE for LSP stars assuming the binary hypothesis in Gaia DR3 (left) and DR4 (right). The binary assumptions are the same as in [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗

**Figure 9.** Figure 9: Dust-photocenter time models and their predicted astrometric signatures. The top row shows example photocenter motion in ∆α for one representative LSP star, using the same dust-inferred photocenter semi-amplitude but different time dependences: sinusoidal motion on the LSP timescale (left), smooth stochastic wander with coherence time τ = 2 yr (middle), and random epoch-to-epoch jitter (right). The bottom … view at source ↗

**Figure 10.** Figure 10: Phase-folded RV and light curves for all Gaia LSPs within 750 pc [PITH_FULL_IMAGE:figures/full_fig_p018_10.png] view at source ↗

**Figure 11.** Figure 11: Continuation of Figure [PITH_FULL_IMAGE:figures/full_fig_p019_11.png] view at source ↗

read the original abstract

Approximately one-third of luminous pulsating red giant stars exhibit long secondary periods (LSPs): stable photometric variability with periods of several months to years in addition to their much shorter primary pulsation cycles. Now nearly a century after their discovery, the physical origin of LSPs remains unresolved. A leading explanation invokes binarity, in which the LSP corresponds to the orbital period of a low-mass companion responsible for both the photometric variability and the radial-velocity (RV) modulation. We test this hypothesis using a nearby sample of LSP stars from the {\it Gaia} Focused Product Release, which provides multi-epoch RVs and contemporaneous optical photometry. We find that interpreting the observed RV variability as orbital motion implies companion masses narrowly distributed around $M_2 \approx 0.1~{\rm M_\odot}$ with separations of 1--3 au, placing them squarely in the brown dwarf desert observed around their solar-type progenitors. Assuming such companions exist, we then forward-model the astrometric signature expected in {\it Gaia} DR3 and predict systematically elevated {\tt RUWE} values for nearby LSPs. In contrast, the observed {\tt RUWE} of nearby LSP stars is systematically lower than these predictions and consistent with most systems exhibiting LSPs being single. This discrepancy disfavors low-mass stellar or substellar companions as the dominant origin of LSPs in evolved stars, motivating a further exploration of alternative stellar mechanisms.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Gaia RUWE measurements for LSP stars fall below what the RV-inferred companions would produce, but the forward model may not fully capture how Gaia absorbs long-period signals into its single-star fits.

read the letter

The paper's core result is that nearby LSP stars show RUWE values consistent with single stars, while the companion masses and separations pulled from their Gaia RVs should produce noticeably higher RUWE if the variability is orbital. That discrepancy is the new observational handle they add to a long-standing question about what drives LSPs in red giants. They pull the RV amplitudes directly from the Focused Product Release, convert them to companion parameters around 0.1 solar masses at 1-3 au, then run a forward model to predict the astrometric excess noise. The observed distribution sits lower, which undercuts the binary channel as the main explanation. That step is useful because it turns an existing RV sample into a testable astrometric prediction without needing new observations. The modeling itself is the soft spot. Periods of several hundred days sit close to the DR3 time baseline, so the orbital motion can be partly absorbed into the fitted parallax and proper motion; the scanning law adds further modulation. If the simulator does not inject the actual Gaia epoch times, attitude history, and noise realization for each target before refitting the single-star solution, the predicted RUWE excess is likely inflated. The abstract gives no numbers on how they handled that, so the size of the claimed discrepancy is hard to judge from the summary alone. This work is aimed at people who follow red-giant variability and binary statistics. A reader who already knows the LSP problem and wants a quick observational constraint will find the comparison worth seeing. It is worth sending to referees so they can check the details of the RUWE forward model and the sample cuts. The central claim is plausible but needs that verification before it can be treated as settled.

Referee Report

2 major / 2 minor

Summary. The paper claims that RV variations in a sample of nearby LSP red giants, if interpreted as orbital motion, imply low-mass companions (M2 ≈ 0.1 M⊙ at 1–3 au). Forward-modeling the expected Gaia DR3 astrometric perturbations for these parameters predicts systematically elevated RUWE values, yet the observed RUWE distribution for the same stars is low and consistent with single-star solutions. This discrepancy is used to disfavor a binary origin for LSPs and to motivate alternative stellar mechanisms.

Significance. If the forward-modeling of RUWE is accurate and the RV-to-orbit conversion holds, the result provides a direct, falsifiable test that weakens the long-standing binary-companion hypothesis for LSPs. The approach usefully combines Gaia RVs, photometry, and astrometry on a well-defined nearby sample and yields a clear observational prediction that can be checked with future data releases.

major comments (2)

[RUWE forward-modeling section] The RUWE forward-modeling section does not appear to inject the actual Gaia DR3 epoch times, attitude history, and scanning-law geometry for each target before refitting the single-star astrometric solution. For orbital periods of several hundred days (comparable to the 34-month DR3 time span), the astrometric signal can be partially absorbed into the fitted parallax and proper-motion terms; without the real scanning law the predicted excess noise is likely overestimated, directly affecting the claimed discrepancy with the observed low RUWE values.
[RV analysis and companion-parameter derivation] The conversion from observed RV semi-amplitude to companion mass and separation (leading to the narrow M2 ≈ 0.1 M⊙, 1–3 au distribution) assumes the RV variability is purely Keplerian and that the stars are viewed at random inclinations. No quantitative assessment is given of how non-orbital contributions to the RV signal or selection effects in the LSP sample would alter the predicted RUWE distribution.

minor comments (2)

[Sample selection] Clarify the exact definition and error treatment of the 'nearby' sample cut and whether RUWE values are taken directly from the Gaia archive or recomputed.
[Results] The abstract states that the observed RUWE is 'consistent with most systems exhibiting LSPs being single'; a quantitative statement of the fraction of stars whose RUWE lies within the single-star locus would strengthen this claim.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We are grateful to the referee for their insightful comments, which have helped us improve the clarity and robustness of our analysis. Below, we provide a point-by-point response to the major comments. We maintain that our conclusions are supported by the data, but we will make revisions to address the concerns raised.

read point-by-point responses

Referee: The RUWE forward-modeling section does not appear to inject the actual Gaia DR3 epoch times, attitude history, and scanning-law geometry for each target before refitting the single-star astrometric solution. For orbital periods of several hundred days (comparable to the 34-month DR3 time span), the astrometric signal can be partially absorbed into the fitted parallax and proper-motion terms; without the real scanning law the predicted excess noise is likely overestimated, directly affecting the claimed discrepancy with the observed low RUWE values.

Authors: We thank the referee for highlighting this limitation in our forward-modeling of the RUWE. Our approach uses a simplified estimation of the astrometric wobble without the full Gaia DR3 scanning law and epoch data. We agree that this can lead to an overestimation of the predicted RUWE for orbital periods comparable to the DR3 baseline, as some of the signal may be absorbed into the fitted parameters. This is a valid point, and we will revise the manuscript to explicitly discuss this approximation and its implications. Specifically, we will note that while the predicted RUWE may be somewhat overestimated, the observed values are still systematically lower and consistent with single-star solutions, thereby maintaining the discrepancy with the binary model. A full simulation is beyond the scope of the current work but will be pursued in follow-up research. We believe this addition will strengthen the paper by providing necessary context. revision: partial
Referee: The conversion from observed RV semi-amplitude to companion mass and separation (leading to the narrow M2 ≈ 0.1 M⊙, 1–3 au distribution) assumes the RV variability is purely Keplerian and that the stars are viewed at random inclinations. No quantitative assessment is given of how non-orbital contributions to the RV signal or selection effects in the LSP sample would alter the predicted RUWE distribution.

Authors: Our conversion from RV semi-amplitudes to companion parameters follows the standard procedure for testing the binary hypothesis, assuming Keplerian motion and random inclinations. We did not include a quantitative assessment of non-orbital RV contributions or selection effects because the primary aim is to evaluate the binary scenario under the assumptions commonly adopted in the literature. If non-orbital effects are present, it would undermine the binary interpretation, supporting our overall conclusion. We will revise the manuscript to include a discussion of these assumptions and their potential effects on the RUWE predictions. This will involve adding a qualitative analysis of how such factors might influence the results, without altering the main findings. The revision will be made in the methods and discussion sections. revision: partial

Circularity Check

0 steps flagged

No significant circularity; derivation uses independent RV-to-orbit mapping and external Gaia forward model

full rationale

The paper's chain starts from observed RV variability interpreted as orbital motion to derive companion masses and separations (M2 ≈ 0.1 M⊙, 1-3 au). It then applies an external forward-modeling procedure based on Gaia DR3 scanning law, attitude, and single-star solution fitting to predict RUWE values for those parameters. Observed RUWE is compared directly to this prediction. This is a test against external astrometric data and instrument properties rather than any self-referential fit, definition, or self-citation chain. No step reduces by construction to the inputs; the discrepancy is falsifiable outside the paper's equations. Minor self-citations, if present, are not load-bearing for the central claim.

Axiom & Free-Parameter Ledger

1 free parameters · 2 axioms · 0 invented entities

The analysis rests on interpreting RV variations as orbital motion and on the fidelity of Gaia RUWE as an astrometric indicator; no new physical entities are introduced.

free parameters (1)

companion mass and separation distribution
Narrowly peaked around 0.1 solar masses at 1-3 au, derived from RV data under the orbital assumption.

axioms (2)

domain assumption Observed RV variability is produced by Keplerian orbital motion of a companion
Invoked to convert RV amplitudes into companion masses and separations.
standard math Gaia RUWE reliably traces excess astrometric noise from unresolved companions
Used as the observable to compare model predictions against data.

pith-pipeline@v0.9.0 · 5566 in / 1397 out tokens · 66251 ms · 2026-05-10T17:27:14.953414+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

2 extracted references · 2 canonical work pages

[1]

2020, MNRAS, 496, 1922, doi: 10.1093/mnras/staa1522

Belokurov, V., Penoyre, Z., Oh, S., et al. 2020, MNRAS, 496, 1922, doi: 10.1093/mnras/staa1522 Chiavassa, A., Freytag, B., & Schultheis, M. 2018, A&A, 617, L1, doi: 10.1051/0004-6361/201833844 Chiavassa, A., Kudritzki, R., Davies, B., Freytag, B., & de Mink, S. E. 2022, A&A, 661, L1, doi: 10.1051/0004-6361/202243568 Courtney-Barrer, B., Haubois, X., Wood,...

work page doi:10.1093/mnras/staa1522 2020
[2]

J., et al

https://dms.cosmos.esa.int/COSMOS/doc fetch. php?id=1566327 MacLeod, M., Blunt, S., De Rosa, R. J., et al. 2025, ApJ, 978, 50, doi: 10.3847/1538-4357/ad93c8 Marcy, G. W., & Butler, R. P. 2000, PASP, 112, 137, doi: 10.1086/316516 Nicholls, C. P., Wood, P. R., Cioni, M.-R. L., & Soszy´ nski, I. 2009, MNRAS, 399, 2063, doi: 10.1111/j.1365-2966.2009.15401.x O...

work page doi:10.3847/1538-4357/ad93c8 2025

[1] [1]

2020, MNRAS, 496, 1922, doi: 10.1093/mnras/staa1522

Belokurov, V., Penoyre, Z., Oh, S., et al. 2020, MNRAS, 496, 1922, doi: 10.1093/mnras/staa1522 Chiavassa, A., Freytag, B., & Schultheis, M. 2018, A&A, 617, L1, doi: 10.1051/0004-6361/201833844 Chiavassa, A., Kudritzki, R., Davies, B., Freytag, B., & de Mink, S. E. 2022, A&A, 661, L1, doi: 10.1051/0004-6361/202243568 Courtney-Barrer, B., Haubois, X., Wood,...

work page doi:10.1093/mnras/staa1522 2020

[2] [2]

J., et al

https://dms.cosmos.esa.int/COSMOS/doc fetch. php?id=1566327 MacLeod, M., Blunt, S., De Rosa, R. J., et al. 2025, ApJ, 978, 50, doi: 10.3847/1538-4357/ad93c8 Marcy, G. W., & Butler, R. P. 2000, PASP, 112, 137, doi: 10.1086/316516 Nicholls, C. P., Wood, P. R., Cioni, M.-R. L., & Soszy´ nski, I. 2009, MNRAS, 399, 2063, doi: 10.1111/j.1365-2966.2009.15401.x O...

work page doi:10.3847/1538-4357/ad93c8 2025