🔭 astrophysics

How precisely can we measure the ages of subgiant and giant stars?

By using wide binaries as a model-independent benchmark, this study demonstrates that spectroscopic metallicity-based age estimates for subgiant stars can achieve realistic 5–10% precision, whereas photometric methods underestimate uncertainties and giant star ages remain less precise at 25–30%.

Original authors: Cheyanne Shariat, Kareem El-Badry, Soumyadeep Bhattacharjee

Published 2026-05-26

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Cheyanne Shariat, Kareem El-Badry, Soumyadeep Bhattacharjee

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Question: How Old is That Star?

Imagine you are an astronomer trying to figure out the history of our galaxy, the Milky Way. To do this, you need to know the ages of millions of stars. It's like trying to reconstruct the history of a city by knowing exactly how old every single building is.

The problem is, stars don't have birth certificates. Astronomers have to guess their ages based on how bright they are, what color they are, and what chemicals they contain. Recently, several new "star catalogs" (databases) were created that claim to have calculated these ages with high precision, saying they are usually within 10% of the truth.

But how do we know if those guesses are actually good?

The Solution: The "Twin" Test

This paper introduces a clever way to check the accuracy of these star age guesses. The authors used wide binary stars.

Think of a wide binary as a cosmic pair of twins. These are two stars that were born at the exact same time, from the same cloud of gas, and are floating near each other in space. Because they are twins:

They are the same age.
They have the same chemical makeup.

If two different astronomers (or two different computer programs) look at this pair of twins and say, "Star A is 5 billion years old" and "Star B is 10 billion years old," something is wrong. If the programs are accurate, they should give very similar answers, and the "uncertainty" (the margin of error) they report should cover the difference.

The authors used this "Twin Test" to check three different star catalogs.

The Three Contenders

The paper tested three major groups of stars:

Subgiants: Stars that are just starting to run out of fuel and expand (like a star "waking up" from its main life).
Red Giants: Stars that have swollen up and are in their old age.
Red Clump Stars: A specific, stable phase of old stars.

Here is how the three catalogs performed in the "Twin Test":

1. The "Gold Standard" (Xiang & Rix, 2022)

The Method: This team used detailed spectroscopy (splitting starlight into a rainbow) to measure the exact chemical ingredients of the stars, specifically looking at iron and alpha-elements (like oxygen and magnesium).
The Result: When they looked at the twin stars, the age estimates matched almost perfectly. If the catalog said the error was 10%, the actual difference between the twins was usually within that 10%.
The Takeaway: This method works. They achieved an accuracy of about 5–10%. It's like a clock that is so precise it only loses a few seconds a year.

2. The "Overconfident" Team (Nataf et al., 2024)

The Method: This team tried to guess the ages using only photos (photometry) and light from ultraviolet telescopes, rather than detailed chemical analysis.
The Result: They claimed their errors were small (around 10%), but when the authors checked the twins, the ages were often wildly different—sometimes off by a factor of 2 or 3.
The Takeaway: This team was overconfident. Their "margin of error" was too small. They thought they were accurate, but they were actually missing the mark by a wide margin. The paper suggests that without measuring the exact chemicals (like iron), you can't get a precise age for these stars.

3. The "Honest but Rough" Team (Wang et al., 2023)

The Method: This team focused on older Red Giant stars using a mix of data and artificial intelligence to distinguish between different types of giants.
The Result: Their age estimates were honest. When they said the error was 25–30%, the twins actually differed by about that much.
The Takeaway: They are reliable, but not super precise. It's like a clock that is accurate to within a few minutes a day. It's good enough to know if it's morning or afternoon, but not good enough to catch a train.

Why Does This Matter?

The paper concludes that to get a precise age for a star (especially the "subgiants"), you must know its chemical recipe. You can't just guess based on how bright or colorful it looks.

The authors also highlight that using "twin stars" (wide binaries) is a powerful, independent way to check our work. It doesn't rely on complex theories about how stars evolve; it just relies on the simple fact that twins should be the same age.

Summary

The Goal: Check if new star age catalogs are telling the truth about how precise they are.
The Tool: Using pairs of "twin" stars that should be the same age to see if the catalogs agree.
The Winner: The catalog that uses detailed chemical measurements (Xiang & Rix) is the most precise and honest (5–10% error).
The Loser: The catalog that relies only on photos (Nataf et al.) is too optimistic; their errors are actually 2–3 times bigger than they admit.
The Runner-up: The catalog for old giant stars (Wang et al.) is honest about its lower precision (25–30% error).

In short: If you want to know a star's age with high precision, you need to taste its soup (measure its chemicals), not just look at the bowl (measure its brightness).

Technical Summary: Precision of Subgiant and Giant Star Ages

Problem Statement
Precise stellar ages are fundamental to Galactic archaeology, specifically for reconstructing the Milky Way's formation, assembly, and chemical enrichment history. While modern surveys routinely estimate ages for millions of field stars by combining astrometry, photometry, spectroscopy, and asteroseismology, obtaining reliable age estimates and, crucially, accurate uncertainty quantification remains a significant challenge. Existing internal validation tests often rely on the same stellar evolution models used to derive the ages, limiting their ability to detect systematic errors. Furthermore, empirical checks using star clusters are constrained by narrow ranges of ages and metallicities. There is a critical need for an independent benchmark that samples a broad range of stellar populations to determine if reported age uncertainties are realistic.

Methodology
This study utilizes wide binary stars as a model-independent benchmark to test the fidelity of age uncertainties in recent catalogs of evolved stars. The methodology relies on the physical principle that the components of a wide binary system formed simultaneously (are coeval) and share a common chemical composition. Therefore, any discrepancy between the independently inferred ages of the two components, when normalized by their reported uncertainties, serves as a direct measure of the accuracy of those uncertainty estimates.

The authors constructed a sample of wide binaries by extending the El-Badry et al. (2021) catalog to 5 kpc using Gaia DR3 astrometry, resulting in approximately 1.6 million pairs. They cross-matched this sample with three major catalogs of evolved stars:

Xiang & Rix (2022): Subgiants with spectroscopic metallicities derived from LAMOST DR7.
Nataf et al. (2024): Subgiants with photometric metallicities inferred from UV-IR photometry and Gaia distances.
Wang et al. (2023): Red Giant Branch (RGB) and Red Clump (RC) stars from LAMOST DR8, utilizing pseudo-asteroseismic diagnostics.

For each binary pair where both components had age estimates, the authors calculated the uncertainty-normalized age difference ( $z$ ). If reported errors are realistic and independent, the distribution of $z$ should follow a Gaussian with a standard deviation ( $\sigma_z$ ) of approximately 1.

Key Results
The analysis of wide binary pairs yielded distinct conclusions regarding the three catalogs:

Xiang & Rix (2022) Subgiants: The ages are generally consistent with their reported uncertainties. After removing a single chemically discrepant outlier, the standard deviation of the normalized age difference was $\sigma_z = 0.81$ . This implies that fractional uncertainties of 5–10% are realistically achievable for subgiants when using spectroscopic metallicities (specifically [Fe/H] and [ $\alpha$ /Fe]).
Wang et al. (2023) Giants: The age estimates for RGB and RC stars also show reliable uncertainties, with $\sigma_z \approx 1.08$ for the full sample and $\sigma_z = 0.75$ for high signal-to-noise ratio (SNR > 50) spectra. However, the absolute precision is lower than that of subgiants, with typical fractional uncertainties of 25–30%.
Nataf et al. (2024) Subgiants: In contrast, this catalog significantly underestimates true uncertainties. The full sample exhibited $\sigma_z = 2.73$ , and even the "Primary Sample" (restricted to specific metallicities) showed $\sigma_z \approx 2.2$ . This indicates that the reported errors are too small by a factor of 2–3. The authors attribute this excess scatter to residual systematics, likely stemming from the reliance on photometric metallicity estimates rather than spectroscopic abundances.

Significance and Claims
The paper establishes wide binaries as a powerful, model-independent tool for calibrating stellar age measurements in the era of large spectroscopic surveys. The primary claim is that accurate chemical abundance measurements are essential for precise subgiant ages. The success of the Xiang & Rix (2022) catalog is attributed to the use of spectroscopic [Fe/H] and [ $\alpha$ /Fe] measurements, whereas the photometric approach of Nataf et al. (2024) introduces unaccounted systematics.

The authors emphasize that their results represent a lower limit on total age uncertainties, as wide binaries cancel out systematics dependent on shared parameters (like metallicity) and do not detect global offsets in the age scale. Therefore, while wide binaries validate the relative precision and error reporting of catalogs, they do not replace the need for absolute calibration efforts involving stellar evolution models.

Ultimately, the study concludes that for Galactic archaeology applications requiring high-precision ages, the Xiang & Rix (2022) subgiant catalog currently offers the most precise and empirically validated ages among the datasets considered. The work underscores that future efforts to estimate ages for larger samples (e.g., using Gaia XP spectra) must be rigorously validated against wide binaries to ensure uncertainty estimates are realistic.