Estimating the dynamical masses of dwarf galaxies in the presence of binary-star contamination

This paper demonstrates that accounting for undetected binary stars in ultra-faint dwarf galaxies reduces their estimated dynamical masses by a factor of 1.5 to 3, challenging current dark matter models and the classification of certain systems as galaxies, while showing that multi-epoch observations can effectively mitigate this contamination.

The Gist

The Big Picture: Weighing Ghosts in the Dark

Imagine you are trying to weigh a ghost. You can't see the ghost, but you can see how fast the air around it is swirling. If the air is spinning wildly, you assume the ghost is heavy. If the air is calm, the ghost must be light.

In astronomy, these "ghosts" are Dark Matter, and the "air" is the stars in tiny, faint galaxies called Ultra-Faint Dwarfs (UFDs). For years, astronomers have looked at how fast these stars move and concluded that these tiny galaxies are incredibly heavy—so heavy that they are almost entirely made of invisible dark matter. In fact, they seemed to be the most "dark-matter-heavy" things in the universe.

But there's a problem. The paper argues that we might have been tricked by a cosmic illusion.

The Culprit: The "Cosmic Dance Partner"

The problem is binary stars.

Think of a binary star system like a couple dancing in a crowded room. They are holding hands and spinning around a common center. To an observer standing far away (like us on Earth), the couple looks like a single point of light. However, as they spin, one star moves toward us and the other moves away. This makes the single point of light appear to wobble back and forth.

In the past, astronomers looked at these "wobbles" and thought, "Wow, the stars in this galaxy are moving really fast! There must be a massive amount of dark matter holding them together."

The Reality: The stars weren't moving fast because of dark matter; they were just dancing with their partners. The "wobble" of the binary stars was being mistaken for the "swirl" of the whole galaxy.

What This Paper Did: The "Truth Detector"

The authors of this paper built a new mathematical tool to separate the "dancers" from the "swirl."

1. The Single-Photo Problem: Most data we have for these tiny galaxies is like a single snapshot. You see the stars in one position, but you don't know if they are just standing still or if they are in the middle of a dance.
2. The New Method: The team used computer models based on how stars dance in our own neighborhood (the Solar System) to predict how much "wobble" binary stars create. They then mixed this "wobble" into their calculations.
3. The Result: When they subtracted the "dance wobble" from the total speed, the galaxies turned out to be much lighter than we thought.

The Shocking Conclusions

When they corrected for the binary stars, three major things happened:

1. The Galaxies Got Lighter (By a Lot)
The estimated mass of these galaxies dropped by a factor of 1.5 to 3.

• Analogy: Imagine you thought a backpack weighed 30 pounds because it was bouncing around. Once you realized the bouncing was just a cat jumping inside, you realized the backpack only weighs 10 pounds.
• Why it matters: If these galaxies are lighter, they are less "dark-matter-heavy." This makes it harder to detect signals from dark matter particles (like annihilation or decay) because there's less dark matter to look for.

2. Some "Galaxies" Might Just Be Star Clusters
Astronomers define a "galaxy" as a system held together by dark matter. If a system has no dark matter, it's just a globular cluster (a ball of stars).

• The paper suggests that three specific systems—Leo IV, Sagittarius II, and UNIONS 1—might actually be just globular clusters. Once you remove the "dance wobble," their internal speed drops so low that they might not have any dark matter at all. They might be imposters!

3. The Solution: Time-Traveling Cameras
How do we fix this in the future? We need multi-epoch data.

• Analogy: Instead of taking one photo of the dancing couple, you take a video. If you watch them for a year, you can clearly see the dance. You can identify the couples, ignore them, and measure the speed of the people who are just standing still.
• The paper shows that if we observe these galaxies over a period of one year, we can spot the binary stars, remove them from our data, and get a true measurement of the galaxy's mass.

The Bottom Line

For a long time, we thought these tiny, faint galaxies were the ultimate dark matter powerhouses. This paper says, "Hold on a minute. We were confusing the stars' dance moves with the galaxy's gravity."

By accounting for the binary stars, these galaxies turn out to be much less extreme than we thought. Some might not even be galaxies at all. It's a reminder that in the universe, sometimes the most dramatic movements are just a local dance, not a sign of a giant invisible force.

To summarize in one sentence: We thought these tiny galaxies were heavy with dark matter, but we were actually just watching stars dance with their partners; once we stop the music and count the real weight, the galaxies are much lighter, and some might not be galaxies at all.

Technical Summary

1. Problem Statement

Ultra-faint dwarf galaxies (UFDs) are critical laboratories for studying dark matter (DM) and testing cosmological models (e.g., $\Lambda$CDM). Their dynamical mass-to-light ratios ($M/L$) are estimated using the line-of-sight velocity dispersion ($\sigma_{\text{los}}$) of their stellar components. However, UFDs present unique challenges:

• Low Velocity Dispersion: Typical $\sigma_{\text{los}}$ values are very low (3–6 km s$^{-1}$).
• Small Sample Sizes: Spectroscopic samples often contain only a few dozen stars, making measurements highly sensitive to outliers.
• Binary Contamination: Undetected spectroscopic binary stars introduce orbital velocity variations that artificially inflate the observed $\sigma_{\text{los}}$. This inflation leads to overestimated dynamical masses, potentially misclassifying globular clusters as DM-dominated galaxies or distorting constraints on DM halo profiles.
• Data Limitations: Most UFD datasets are single-epoch (one observation per star), preventing the direct detection of binaries via velocity variations.

The paper addresses the need to correct $\sigma_{\text{los}}$ and dynamical mass estimates for binary contamination in both single-epoch and multi-epoch regimes.

2. Methodology

The authors developed a Bayesian mixture model framework to disentangle the intrinsic velocity dispersion of the galaxy from the kinematic signature of binary stars.

A. Binary Population Models

• Source: The authors utilize updated binary population models from the solar neighborhood (Moe & Di Stefano 2017) rather than older models (e.g., Duquennoy & Mayor 1991).
• Parameters: They assume a primary star mass ($m_1$) of 0.8 $M_{\odot}$ (typical for old, metal-poor RGB and turn-off stars). They adopt specific distributions for mass ratio ($q$), eccentricity ($e$), and orbital period ($P$), with physical cutoffs to prevent Roche-lobe overflow.
• Velocity Calculation: The line-of-sight velocity component of the binary ($v_{\text{los,bin}}$) is calculated analytically based on orbital parameters.

B. The Mixture Model (Single-Epoch)

For single-epoch data, the authors define a likelihood function ($\mathcal{L}$) where the observed velocity distribution is a mixture of:

1. Single Stars: Modeled as a Gaussian with systemic velocity ($v_{\text{sys}}$) and intrinsic dispersion ($\sigma_{\text{los}}$).
2. Binary Stars: Modeled as the convolution of the intrinsic Gaussian with the binary velocity distribution ($B_i$).

• Free Parameters: Systemic velocity ($v_{\text{sys}}$), intrinsic dispersion ($\sigma_{\text{los}}$), and the binary fraction ($f$).
• Truncation: To mimic observational selection effects, the binary distribution is truncated at $\pm 3\sigma$ from the systemic velocity, removing stars that would be rejected as non-members in real data.
• Inference: The posterior probability distribution (PPD) is sampled using the EMCEE (Affine Invariant MCMC) package.

C. Generalization to Multi-Epoch Data

The methodology is extended to handle multi-epoch data:

• Detection: Stars exhibiting velocity variations $> 3 \times \Delta v_{\text{los}}$ between epochs are identified as binaries and removed from the sample.
• Residual Correction: For the remaining "undetected" binaries (those with low amplitude or long periods), the authors compute a new binary distribution ($B_i$) based on the specific time baseline and measurement errors of the dataset. This distribution is then used in the mixture model to correct the remaining sample.

3. Key Contributions

• Systematic Correction Framework: A unified Bayesian formalism to correct $\sigma_{\text{los}}$ for binary contamination without imposing strong priors on the binary fraction ($f$).
• Application to UFDs: Application of this method to 12 UFDs with $\sigma_{\text{los}} < 4.5$ km s$^{-1}$ (including Bootes I, Carina II, Crater II, Leo IV, Segue 1, etc.).
• Mock Data Validation: Testing the multi-epoch correction strategy on a simulated UFD (20 stars, intrinsic $\sigma_{\text{los}} = 2.4$ km s$^{-1}$) to demonstrate the efficacy of time-domain spectroscopy.
• Low-Number Statistics Analysis: Investigation of how binary contamination affects systems with very small sample sizes (as few as 10 stars), a regime where previous studies (e.g., Spencer et al. 2017) focused on larger samples.

4. Key Results

A. Impact on Dynamical Masses

• Mass Reduction: Accounting for undetected binaries reduces the estimated dynamical masses ($M_{\text{dyn}}$) of UFDs by a factor of 1.5 to 3.
• M/L Ratios: The corrected mass-to-light ratios are significantly lower. For the most affected systems (e.g., Leo IV, UNIONS 1, Sagittarius II), the lower limits of the mass estimates drop so low that they challenge the classification of these objects as galaxies.
• Specific Cases:
  • Leo IV, Sag II, UNIONS 1: After correction, their velocity dispersions become unresolved (consistent with 0 km s$^{-1}$). Combined with unconstrained metallicity spreads, these systems can no longer be confirmed as DM-dominated galaxies; they may be globular clusters.
  • Reticulum II: Remains a robust galaxy candidate with a resolved $\sigma_{\text{los}}$ even after binary correction.
  • Crater II: The corrected $\sigma_{\text{los}}$ aligns closely with Modified Newtonian Dynamics (MOND) predictions but exacerbates tensions with $\Lambda$CDM NFW halo models, suggesting an even more cored DM profile.

B. Multi-Epoch Efficacy

• Mitigation: A dedicated multi-epoch campaign (spanning ~1 year) that removes stars with clear velocity variations significantly reduces binary-induced inflation.
• Residual Effect: Even after removing detected binaries, undetected binaries still cause a slight inflation in $\sigma_{\text{los}}$. Therefore, statistical correction for the remaining population remains necessary for high-precision mass estimates.

C. Low-Number Statistics Regime

• Inflation Limits: For systems with only ~10 stars and single-epoch data, a binary fraction of $f=0.6$ can inflate the observed $\sigma_{\text{los}}$ to 5.5 km s$^{-1}$ (within 1$\sigma$), regardless of the intrinsic dispersion.
• Multi-Epoch Improvement: With a 1-year multi-epoch baseline, this upper limit drops to < 3 km s$^{-1}$, highlighting the necessity of time-domain spectroscopy for faint systems.

5. Significance and Implications

• Cosmological Tensions: The reduction in inferred masses worsens existing tensions between UFD observations and $\Lambda$CDM predictions (specifically regarding the minimum halo mass and cusp-core problems). Systems like Bootes I and Crater II now appear even more difficult to reconcile with standard NFW halos.
• Dark Matter Detection: Lower dynamical masses imply lower J-factors (integrated DM density), which weakens the prospects for detecting DM annihilation or decay signals from UFDs and increases upper limits on particle-physics cross-sections.
• Galaxy vs. Cluster Classification: The study suggests that several recently discovered "UFDs" (Leo IV, Sag II, UNIONS 1) may actually be globular clusters, necessitating a re-evaluation of the low-mass end of the galaxy population.
• Observational Strategy: The paper establishes that single-epoch data is insufficient for robust mass estimation in the lowest dispersion regime. Future surveys must prioritize multi-epoch observations with time baselines of at least one year to mitigate binary contamination.

In conclusion, this work provides a critical correction to the dynamical mass estimates of the faintest galaxies, demonstrating that binary stars have been systematically inflating our view of their dark matter content, with profound implications for galaxy formation theories and dark matter searches.