Benchmarking pre-main sequence stellar evolutionary tracks using disk-based dynamical stellar masses

This paper benchmarks pre-main sequence stellar evolutionary tracks by comparing HR diagram-derived masses with dynamical masses from ALMA disk observations of Upper Scorpius stars, finding that models with moderate spot coverage best match the data and that incorporating dynamical mass priors significantly reduces age scatter and improves model consistency.

The Gist

Imagine you are trying to guess the age and weight of a group of teenagers. You have two ways to do it:

1. The "Growth Chart" Method: You look at how tall they are and how much they weigh, then compare them to a standard growth chart for teenagers. This tells you how old they should be based on theory.
2. The "Scale" Method: You put them on a real, physical scale. This gives you their actual, undeniable weight.

For decades, astronomers have been trying to figure out the age and weight of baby stars (called Pre-Main Sequence stars). They've mostly relied on the "Growth Chart" method (using something called a Hertzsprung-Russell diagram). But here's the problem: the growth charts they use are just theories. They depend on how the modeler thinks stars work. If the theory is slightly off, the estimated age and weight are wrong.

This paper is like a reality check. The authors decided to weigh the baby stars using a "real scale" and see if the "growth chart" theories match up.

The "Real Scale": Listening to the Spin

How do you weigh a star that is millions of miles away? You can't put it on a scale. Instead, the authors looked at the protoplanetary disks (rings of gas and dust) spinning around these stars.

Think of a figure skater spinning with their arms out. If they pull their arms in, they spin faster. The speed at which the gas in the disk spins is determined by the gravity of the star in the center. By using powerful radio telescopes (ALMA) to listen to the "song" of the gas (specifically Carbon Monoxide), the team could calculate the star's true, dynamical mass. This is their "real scale" measurement.

The "Growth Charts": Testing the Theories

Next, they took the known temperature and brightness of these 20 stars and ran them through 10 different theoretical models (the "growth charts"). These models try to predict how a star evolves, but they differ in how they handle complex physics like:

• Magnetic fields: Do stars have strong magnetic shields?
• Star spots: Do they have giant, dark sunspots (like freckles) that make them look cooler?

The Big Reveal: Which Model Got it Right?

The authors compared the "real scale" weights against the "growth chart" predictions. Here is what they found, using some fun analogies:

• The "Spotless" Models (The Optimists): Some models assumed the stars were perfectly smooth. These models tended to underestimate the weight. It's like guessing a teenager is 100 lbs when they are actually 120 lbs because you forgot to account for their muscle mass.
• The "Magnetic" Models (The Pessimists): Models that assumed strong magnetic fields tended to overestimate the weight. It's like guessing the teenager is 150 lbs because you think they are carrying a heavy backpack they aren't actually wearing.
• The "Sweet Spot" (The Winners): The model that worked best was the SPOTS model with 17% coverage. This model assumed the stars had a moderate amount of dark spots (like a teenager with a few freckles).
  • The Result: 100% of the stars matched the "real scale" weight perfectly when using this specific model! It was the only theory that didn't need to be adjusted to fit the data.

Why Does This Matter?

You might ask, "So what if we get the weight slightly wrong? Why does it matter?"

Because weight determines age.
In the world of baby stars, mass and age are locked in a dance. If you get the weight wrong, your estimate of how old the star is will also be wrong.

• Before this study: If you used different theories, you might think a group of stars was 3 years old, while another group thought they were 6 years old. The answers were all over the place.
• After this study: By using the "real scale" weight as a starting point (a "prior"), the different theories suddenly agreed with each other. The confusion vanished. The estimated ages became consistent, and the "scatter" (the disagreement between theories) dropped by 77%.

The Takeaway

Think of this paper as a quality control check for the universe's instruction manual.

For a long time, astronomers were trying to build a house using blueprints that might have had a few errors. They built the house, but the rooms didn't quite fit. This paper brought in a tape measure (the dynamical mass) to check the blueprints.

They found that the blueprint which included "freckles" (star spots) on the stars was the most accurate. Now, when astronomers look at baby stars in the future, they can use this corrected blueprint to tell us exactly how old they are and how heavy they are, helping us understand how planets (and eventually, life) form around them.

In short: They weighed the stars using the spin of their gas rings, found that the "freckled star" theory was the most accurate, and used that to fix the age estimates for the entire neighborhood of young stars.

Technical Summary

Here is a detailed technical summary of the paper "Benchmarking pre-main sequence stellar evolutionary tracks using disk-based dynamical stellar masses."

1. Problem Statement

Stellar mass is a fundamental parameter for understanding star and planet formation, yet determining the masses of Young Stellar Objects (YSOs) is challenging.

• Current Limitation: The standard method for deriving YSO masses involves placing them on the Hertzsprung–Russell (HR) diagram and comparing their position to theoretical pre-main sequence (PMS) evolutionary tracks. However, these derived masses are strongly model-dependent. Different models (incorporating varying physics like convection, magnetic fields, and stellar spots) yield significantly different mass and age estimates for the same object.
• The Gap: There is a lack of independent, empirical benchmarks to validate which evolutionary models accurately predict stellar masses, particularly for low-mass stars ($0.1 - 1.3 M_\odot$) in the 4–14 Myr age range.

2. Methodology

The authors employed a comparative approach, contrasting "dynamical" masses (independent of evolutionary models) with "evolutionary" masses (derived from HR diagram fitting).

• Sample Selection:

  • Region: Upper Scorpius (an old star-forming region, age 4–14 Myr).
  • Targets: 20 sources with well-constrained dynamical masses, reliable stellar parameters ($T_{eff}$, $L_\star$), and precise Gaia parallaxes.
  • Exclusions: Sources with poorly constrained rotation profiles, multiple systems, or inconsistent spectral types were removed from an initial parent sample of 37.

• Dynamical Masses ($M_{\star, dyn}$):

  • Derived by fitting a parametric model to ALMA Band 7 ($^{12}\text{CO } J=3-2$) visibility data.
  • The stellar mass is a key free parameter that sets the rotational velocity field of the protoplanetary disk (Keplerian rotation).
  • Uncertainty Handling: The authors noted that statistical uncertainties from the original fitting were underestimated. They performed numerical simulations (using mcfost and csalt) to assign realistic uncertainties based on disk size and inclination, creating a conservative error budget.

• Evolutionary Masses ($M_{\star, HRD}$):

  • Derived by fitting the observed effective temperatures and luminosities to ten different PMS evolutionary models using the ysoisochrone Bayesian inference package.
  • Models Tested:
    • Baraffe et al. (2015)
    • Feiden (2016) (Non-magnetic and Magnetic tracks)
    • PARSEC v2.0 (Nguyen et al. 2022)
    • Siess et al. (2000)
    • SPOTS (Somers et al. 2020) with five different cold stellar spot coverages ($f = 0\%, 17\%, 34\%, 51\%, 85\%$).

• Comparison Metric:

  • Calculated the median mass ratio $\langle R \rangle = \langle M_{\star, dyn} / M_{\star, HRD} \rangle$.
  • Determined the percentage of sources agreeing within $\pm 1\sigma$ (quadratic sum of uncertainties).

3. Key Contributions

• Empirical Benchmarking: This study provides the first large-scale, direct benchmark of modern PMS tracks against dynamical masses derived from disk kinematics in the 4–14 Myr regime.
• Quantification of Systematic Biases: The authors quantify the systematic over- or underestimation of masses by specific physical assumptions (e.g., magnetic fields vs. non-magnetic, spot coverage).
• Impact of Priors: The paper demonstrates the critical impact of using dynamical mass priors in HR diagram fitting, showing how it resolves age degeneracies and reduces scatter across different models.

4. Results

A. Model Performance (Mass Range $0.1 - 1.3 M_\odot$)

• Best Performer: The SPOTS model with a 17% cold spot coverage ($f=17\%$) provided the most accurate predictions.
  • 100% of the targets agreed with dynamical masses within $\pm 1\sigma$.
  • This model predicts masses intermediate between non-magnetic and magnetic Feiden tracks.
• Moderate/High Spot Coverage:
  • Models with higher spot fractions ($f=34\%$) performed similarly to magnetic equipartition models.
  • Models with high fractions ($f \ge 51\%$) significantly overestimated masses, with agreement dropping to 33% ($f=51\%$) and 13% ($f=85\%$).
• Magnetic vs. Non-Magnetic (Feiden 2016):
  • Non-magnetic tracks: Underestimated dynamical masses by $\sim 12\%$.
  • Magnetic tracks: Overestimated dynamical masses by $\sim 20\%$.
  • Note: While previous studies suggested magnetic tracks were superior for specific mass ranges, this study finds that in the $0.1-1.3 M_\odot$ range, they systematically overestimate mass compared to the dynamical benchmark.
• Other Models:
  • Baraffe et al. (2015): Underestimated masses by $\sim 15\%$.
  • Siess et al. (2000): Underestimated masses by $\sim 26\%$.
  • PARSEC v2.0: Showed the poorest agreement (53%), likely because these tracks were originally designed for evolved/massive stars and extended to lower masses.
• Mass Dependence: Most models showed no correlation between mass discrepancy and stellar mass. However, PARSEC v2.0 and high-spot SPOTS models ($f \ge 51\%$) showed a significant negative correlation, indicating a mass-dependent bias.

B. Impact on Age Determination

• Without Priors: Using different evolutionary tracks resulted in a median age scatter of 3.4 Myr across the sample.
• With Dynamical Mass Priors: When dynamical masses were used as priors in the HR fitting:
  • The median age scatter reduced by $\ge 77\%$ (from 3.4 Myr to 0.8 Myr).
  • The median age across all tracks converged to 6.8 Myr.
  • For individual sources, the inferred median age could shift by up to 25%.

5. Significance

• Validation of Physics: The results strongly suggest that moderate stellar spot coverage ($f \approx 17\%$) is a crucial physical component for accurately modeling PMS stars in the Upper Scorpius region. Models ignoring spots or assuming extreme coverage fail to reproduce dynamical masses.
• Refining Evolutionary Timescales: By reducing the scatter in age estimates by nearly 80% through the use of dynamical mass priors, this work offers a pathway to more precise chronologies of star and planet formation.
• Future Directions: The study establishes a robust framework for testing evolutionary models. The authors note that while the agreement holds for Upper Scorpius (4–14 Myr), testing these models at younger ages (e.g., using the DEC/O Large Program) is necessary to determine if these physics hold across the entire PMS phase.

In conclusion, this paper provides a critical empirical constraint that favors PMS models incorporating moderate stellar spot coverage and highlights the necessity of using independent dynamical mass measurements to calibrate evolutionary tracks and derive reliable stellar ages.