🔭 astrophysics

Precise determination of circumstellar disk lifetimes: Disk evolution in a single star-forming region

By analyzing 33 clusters within the single Scorpius-Centaurus OB association to minimize environmental biases, this study determines a circumstellar disk lifetime of $5.8 \pm 0.3$ Myr, suggesting that planet formation has approximately twice as much time as previously estimated.

Original authors: Fabian A. Polnitzky, Sebastian Ratzenböck, Josefa E. Großschedl, João Alves

Published 2026-03-18

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Fabian A. Polnitzky, Sebastian Ratzenböck, Josefa E. Großschedl, João Alves

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

Imagine that stars are like newborn babies, and the swirling disks of gas and dust around them are their nursery blankets. These blankets are crucial because they are the raw material from which planets (like Earth) are built. But just like a baby outgrows its swaddle, these disks eventually disappear.

The big question astronomers have been asking for decades is: How long does a baby keep its blanket before it's gone?

This paper is like a very precise stopwatch that finally answers that question with much greater accuracy than before. Here is the story of how they did it, explained simply:

1. The Old Way: Comparing Apples to Oranges

Previously, scientists tried to figure out how long these disks last by looking at many different star clusters scattered across the galaxy. It was like trying to figure out how long a child stays in a crib by visiting a nursery in New York, a daycare in London, and a school in Tokyo.

The problem? Every place is different. Some are far away (making it hard to see the small babies), some have different lighting, and some have different rules. This made it hard to get a clear answer. The estimates varied wildly, suggesting disks lasted anywhere from 2 to 8 million years.

2. The New Approach: One Big Neighborhood

The authors of this paper decided to stop comparing different neighborhoods and instead focus on one single, massive neighborhood: the Scorpius-Centaurus OB association.

Think of this as a giant, extended family of stars that were all born at roughly the same time and are all living in the same "city." Because they are all at the same distance from Earth, the scientists could look at them with the same "telescope glasses" and see them all clearly. This removed the confusion of comparing distant, blurry stars with nearby, sharp ones.

They found 33 distinct groups (clusters) of stars within this neighborhood, ranging from very young (3 million years old) to "teenagers" (21 million years old).

3. The Detective Work: Finding the Blankets

How do you know if a star still has its "blanket" (disk)?

The Clue: The blanket glows in infrared light (heat), which is invisible to our eyes but visible to special space telescopes like WISE and 2MASS.
The Method: The team acted like detectives. They looked at the "color" of the stars. Stars with a disk look redder and brighter in infrared than stars without one. They used two different sets of clues (called CCD and SED methods) to make sure they didn't miss any blankets or mistake a random heat source for a disk.

4. The Big Discovery: The Blankets Last Longer!

When they plotted the data—showing the percentage of stars with blankets at every age—they found a clear pattern. It looked like a slide going down: as the stars got older, fewer of them had blankets.

By fitting a mathematical curve to this slide, they calculated the "half-life" of these disks.

The Result: The disks last about 5.8 million years.
Why it matters: This is about twice as long as many previous estimates suggested.

The Analogy: Imagine you thought a child only needed a crib for 3 years. If you suddenly realized they actually need it for 6 years, you would realize they have twice as much time to learn how to walk and talk before they have to leave the crib.

Similarly, if these disks last longer, planets have more time to form. It means the "construction zone" for solar systems stays open longer than we thought, giving gravity and dust more time to build worlds like Earth.

5. Why Was This So Hard Before?

The paper explains that previous studies were biased. They often looked at distant clusters where it was hard to see the small, faint, low-mass stars (the most common type of star). Because they missed the small stars, they mostly saw the big, heavy stars.

The Metaphor: Big stars are like loud, energetic teenagers who throw their blankets away quickly. Small stars are like quiet babies who keep their blankets much longer. Previous studies were mostly listening to the loud teenagers, so they thought everyone threw their blankets away early. This study listened to the whole family, including the quiet babies, and realized the blankets actually last much longer.

The Takeaway

This paper is a major step forward because it stopped guessing and started measuring with a "homogeneous" sample (one neighborhood, one set of rules).

In a nutshell:

Old belief: Disks disappear quickly (2–4 million years).
New finding: Disks hang around longer (about 6 million years).
Implication: We have more time to build planets than we thought. The universe is a bit more patient in its construction of worlds than we previously believed.

1. Problem Statement

Determining the lifetime of circumstellar disks is fundamental to understanding the timescales of planet formation. Previous studies have estimated disk lifetimes by analyzing the fraction of young stars with infrared excess across multiple stellar clusters of varying ages. However, these studies suffer from significant systematic uncertainties and biases due to:

Distance and Completeness: Combining clusters at different distances leads to varying sample completeness, particularly affecting the census of faint, low-mass stars.
Environmental Variability: Different star-forming regions experience distinct physical conditions (e.g., UV radiation fields from massive stars) that influence disk dispersal rates.
Age Uncertainties: Estimating ages for populations at different distances using theoretical models introduces systematic discrepancies.
Sample Heterogeneity: Mixing populations with different initial mass functions (IMFs) and formation histories obscures the true evolutionary trends.

Consequently, previous estimates of disk lifetimes (often cited as 2–4 Myr or up to 8 Myr) vary widely, and the dominant mechanisms driving disk evolution remain debated.

2. Methodology

The authors address these challenges by analyzing a homogeneous sample within a single, nearby star-forming region: the Scorpius-Centaurus (Sco-Cen) OB association.

Sample Selection:
- Utilized the SigMA (Significance Mode Analysis) clustering algorithm applied to Gaia DR3 data to identify 34 co-spatial, co-moving, and coeval stellar clusters within Sco-Cen.
- Excluded one cluster (Centaurus-Far) due to contamination, leaving 33 clusters with ages ranging from 3 to 21 Myr (based on PARSEC isochrones).
- The sample covers the stellar Initial Mass Function (IMF) from the hydrogen-burning limit up to $\sim 8 M_\odot$ , comprising 12,873 stellar sources.
Disk Identification:
- Cross-matched Gaia sources with AllWISE (mid-infrared) and 2MASS (near-infrared) catalogs.
- Identified 10,087 sources with valid infrared counterparts.
- Employed three distinct selection methods to identify disk-bearing (DB) sources via infrared excess:
  1. CCD Selection (Primary): Used four Color-Color Diagrams (CCDs) combining WISE (W1–W4) and 2MASS (H, $K_s$ ) bands (e.g., $W1-W2$ vs. $W2-W3$ ). This method is robust against extinction and photometric errors.
  2. SED Selection: Used the spectral energy distribution (SED) slope ( $\alpha$ ) from $K_s$ to W3/W4, selecting sources with $\alpha \ge -2$ .
  3. Luhman Selection: Compared against an external catalog by Luhman (2022) covering a subset of the clusters.
Statistical Modeling:
- Calculated the disk fraction ( $f_i$ ) for each cluster $i$ as the ratio of disk-bearing to total sources.
- Modeled the disk fraction evolution using an exponential decay function with a shift parameter:
  $f(t) = f_0 \cdot e^{-\frac{t-t_0}{\tau}}$
  Where $t$ is age, $\tau$ is the characteristic decay time, $f_0$ is the initial disk fraction, and $t_0$ accounts for a delay in the onset of dispersal.
- Used Bayesian inference with Markov Chain Monte Carlo (MCMC) sampling (via pymc) to estimate the posterior distribution of $\tau$ .
- Defined a Fiducial Model with physically motivated priors (Half-Cauchy for $\tau$ , Beta for $f_0$ , Half-Normal for $t_0$ ) and tested robustness against "data-driven" (uninformative priors) and "no-shift" models.
- Accounted for observational uncertainties using a Beta-Binomial distribution to model the counting process of disk presence/absence.

3. Key Contributions

Homogeneous Single-Region Analysis: This is the first study to derive a disk fraction vs. age sequence for a single, coherent star-forming region (Sco-Cen) using a statistically rigorous clustering method (SigMA) that minimizes distance-related biases and age mixing.
Comprehensive Mass Sampling: The study samples the full stellar mass range (down to the hydrogen-burning limit) within a single distance, avoiding the high-mass bias common in multi-region studies.
Robust Statistical Framework: The implementation of a probabilistic fitting framework that explicitly models observational uncertainties, selection effects, and prior physical constraints.
Systematic Error Quantification: The paper rigorously tests the impact of different disk selection methods (CCD vs. SED vs. Luhman), binary contamination (via Gaia RUWE), and age determination models (PARSEC vs. Baraffe) on the derived lifetimes.

4. Key Results

Disk Lifetime: The study derives a characteristic disk decay time of $\tau = 5.8 \pm 0.3$ Myr (using the Fiducial model with CCD selection).
- This result is approximately a factor of two longer than many previous estimates (which often cited 2–4 Myr).
- The range of tested models and selection methods suggests a robust lifetime between 5 and 7 Myr for Sco-Cen.
Exponential Decay: The data strongly supports an exponential decay model for disk dispersal. The inclusion of a shift parameter ( $t_0$ ) allows for a delay in the onset of dispersal, though the results are robust even without it.
Initial Disk Fraction ( $f_0$ ): The inferred initial disk fraction is consistently below unity (typically $0.6 - 0.85$), suggesting that not all stars form with detectable disks or that disks dissipate very rapidly before the first observational epoch ( $\sim 3$ Myr).
Selection Method Impact:
- The CCD selection (primary) yields $\tau \approx 5.8$ Myr.
- SED and Luhman selections yield slightly longer decay times ( $\sim 6.5 - 6.9$ Myr) but with lower intercepts, likely due to missing young, high-fraction clusters in the Luhman sample.
Binary Influence: Subsamples separated by Gaia RUWE (a proxy for binarity) show that candidate binaries ( $RUWE \ge 1.4$ ) have a slightly shorter decay time ( $\sim 5.4$ Myr) compared to single stars ( $\sim 6.5$ Myr), consistent with the hypothesis that binarity accelerates disk dispersal.
Age Model Sensitivity: The derived lifetime is sensitive to the choice of stellar evolutionary models. While PARSEC-based ages yield $\sim 5.8$ Myr, Baraffe-based ages (specifically Baraffe-BP-RP) yield shorter timescales ( $\sim 3.3$ Myr), highlighting that age determination remains a primary source of systematic uncertainty.

5. Significance

Planet Formation Timescales: The finding of a $\sim 6$ Myr disk lifetime implies that planet formation has a significantly longer window than previously thought. This challenges models requiring rapid planet formation (e.g., within 2–3 Myr) and supports scenarios where gas giants can form over longer periods or where disk material persists well after the cessation of accretion.
Methodological Benchmark: The study demonstrates that analyzing a single, well-characterized region reduces systematic uncertainties inherent in comparing disparate clusters. It sets a new standard for future disk evolution studies, emphasizing the need for homogeneous samples.
Environmental Context: By isolating Sco-Cen, the results provide a baseline for disk evolution in a low-extinction, moderate-UV environment, distinct from the harsh environments of massive star-forming regions or the deeply embedded cores of younger clusters.
Future Directions: The paper highlights that while the exponential model fits well, the behavior of disks at ages $< 3$ Myr remains unconstrained. Future observations of younger populations are required to fully characterize the initial disk fraction and the very early stages of dispersal.

In conclusion, Polnitzky et al. provide a precise, statistically robust determination of circumstellar disk lifetimes, suggesting that the "clock" for planet formation ticks slower than previously assumed, with a characteristic timescale of nearly 6 million years in the Sco-Cen association.