Assessing the star formation history of all-sky and part-sky 100pc white dwarf samples

Basic Information

• Paper ID: 2510.20624
• Title: Assessing the star formation history of all-sky and part-sky 100pc white dwarf samples
• Authors: Emily K. Roberts, Pier-Emmanuel Tremblay, Antoine Bédard (University of Warwick)
• Classification: astro-ph.SR (Solar and Stellar Astrophysics)
• Publication Date: November 21, 2025 (Preprint)
• Journal: MNRAS (Monthly Notices of the Royal Astronomical Society)
• Paper Link: https://arxiv.org/abs/2510.20624

Abstract

This paper constructs multiple volume-limited white dwarf samples within 100 parsecs (pc) using Gaia satellite data and large-scale spectroscopic follow-up surveys (4MOST, DESI, WEAVE, SDSS-V), systematically assessing their completeness and residual biases. Through white dwarf population simulations compared with observational samples, the study quantitatively characterizes the star formation history (SFH) and Galactic disk age. The research evaluates the value of Gaia XP spectra, the applicability of partial-sky versus all-sky surveys, and predicts the contribution of the 4MOST 100PC sub-survey to understanding local stellar populations.

Research Background and Motivation

Core Questions

White dwarfs, as end products of stellar evolution, have well-defined relationships between cooling time and temperature, luminosity, and mass, serving as "cosmic clocks" to infer star formation history. However, realizing this potential depends on:

1. Complete and unbiased white dwarf samples: Gaia provides unprecedented large samples, but different selection criteria and survey coverage lead to sample variations
2. Accurate physical models: Requiring white dwarf cooling models, main-sequence lifetimes, initial mass functions (IMF), initial-final mass relations (IFMR), etc.
3. Quantitative assessment of observational biases: Effects of spectroscopic completeness, apparent magnitude limits, sky coverage, and other factors

Research Significance

• Galactic archaeology: Reconstructing the star formation history of the solar neighborhood through white dwarfs, with cross-validation against main-sequence star results
• Sample selection strategy: Providing guidance for sample design in future large surveys (e.g., Vera Rubin)
• Methodological benchmark: The 40pc sample has established a methodological framework (Roberts et al. 2025); the 100pc sample can test its extrapolation

Limitations of Existing Methods

• 40pc sample: High completeness (~97%) but small statistical sample (1,073 objects) with large Poisson noise
• 100pc sample: Large sample size (1-1.5 million objects) but completeness and biases insufficiently quantified
• Partial-sky surveys: Spectroscopic coverage only ~30-50% of the sky, with questionable representativeness
• Different selection criteria: Multiple white dwarf catalogs derived from Gaia DR3 (Gentile Fusillo, GCNS, GSPC-WD, etc.) show significant differences

Core Contributions

1. Systematic comparison of multiple 100pc white dwarf samples: Quantitatively assessing completeness and biases of 6 major samples (3 all-sky + 3 partial-sky)
2. Validating the robustness of star formation history reconstruction methods: Demonstrating consistent results across different samples using luminosity function and absolute magnitude distribution methods
3. Quantifying the value of XP spectra: The GSPC-WD sample, despite having spectroscopic classification, exhibits red-end incompleteness (Teff<6000K)
4. Assessing the applicability of partial-sky surveys: DESI DR1 shows unbiased characteristics comparable to all-sky samples, while SDSS exhibits significant biases
5. Predicting 4MOST 100PC sub-survey performance: Simulations show that its apparent magnitude limit (G<19.5) causes faint-end incompleteness, but is sufficient to distinguish different SFH models
6. Confirming the optimal star formation history form: Both late-peaked SFH (Roberts et al. 2025) and constant star formation rate fit the data well

Methodology Details

Task Definition

Input: Observational data of white dwarfs within 100pc (Gaia parallax, magnitudes, spectroscopic parameters, etc.)
Output: Local Galactic disk star formation history SFH(t) and age
Constraints: Volume-limited (100pc spherical or conical region), mass range (0.54-1.34 M⊙, excluding binary products)

Model Architecture: Single-Star Population Synthesis

Core Workflow

1. Initial stellar generation:
   • Generate initial masses according to Salpeter IMF (α=2.35±0.075)
   • Assign formation times according to assumed SFH(t)
   • Assign initial metallicity (Z=0.0134±0.0104)
2. Stellar evolution:
   • Main-sequence + giant branch lifetime: tMSGP (Byrne et al. 2025 model, ±3.5% uncertainty)
   • Convert to white dwarf mass via IFMR
   • White dwarf cooling: Bédard et al. (2020) model (±4.5% uncertainty)
   • Crystallization delay: on/off scenarios
   • Merger delay: tΔ (43% probability of activation)
3. Kinematic evolution:
   • Galactic disk scale height: h(t) = 10.71 × t^0.65 pc
   • Age-velocity dispersion relation: probability of leaving sample p_left = 1 - 65/h
   • Correction for missing main-sequence stars (not yet evolved to white dwarfs)
4. Observational simulation:
   • Atmospheric composition: 25% helium-dominated (±6.5%), DC type <5000K assumed hydrogen atmosphere
   • Magnitude calculation: accounting for extinction (average AV=0.02±0.015 mag)
   • Partial sky: applying actual survey footprints (Figure 3)
   • Apparent magnitude limits: DESI G<20, 4MOST G<19.5

Uncertainty Handling

Running 50 Monte Carlo simulations, with random sampling in parameter space listed in Table 2 for each run:

• Systematic uncertainties: Population age (10.6±0.5 Gyr), IMF slope, main-sequence lifetime, cooling model, etc.
• Individual uncertainties: Metallicity, Gaia measurement errors

Three Star Formation History Reconstruction Methods

1. Luminosity Function Method

• Observable: White dwarf absolute bolometric luminosity Mbol distribution
• Calculation: Requires Teff and radius R → depends on mass-radius relation and atmospheric models
• Challenge: Requires known atmospheric composition (H or He dominated)
• Treatment: Assume pure hydrogen atmosphere for samples without spectra; use custom models for spectroscopic samples

2. Absolute G-band Magnitude Distribution Method

• Observable: Gaia absolute G-band magnitude MG distribution
• Advantage: Only requires parallax and apparent magnitude, no spectroscopic parameters needed
• Simulation: Encodes 25% helium atmosphere fraction in population synthesis
• Extinction correction: Uses AV values from Gentile Fusillo et al. (2021)

3. Direct Age Method (Appendix A)

• Calculation: ttotal = tWD + tMSGP + tΔ
• Application: Requires complete Teff, mass, and spectral type
• Correction: Missing main-sequence stars, age-velocity dispersion effects
• Result: Normalized star formation history dN/dt

Technical Innovations

1. Low-mass problem correction: For Teff<6000K, Lyman-α red-wing opacity errors cause luminosity mass underestimation; applying O'Brien et al. (2024) fifth-order polynomial correction to median 0.6 M⊙
2. Probabilistic volume membership: Considering parallax uncertainty for 100pc membership probability (Figure 1), but finding <2,000 boundary objects with P100pc<0.5, negligible impact on results
3. Binary exclusion strategy: Mass limits 0.54-1.34 M⊙ exclude double white dwarfs and non-single-star evolution products, reducing completeness but decreasing systematic biases
4. Partial-sky simulation: Analytically modeling survey footprints (Figure 3) to simulate directional dependence of Galactic scale height and age-velocity relations

Experimental Setup

Datasets

All-sky 100pc samples (3)

• GF21: Reference sample, ~93% completeness (compared with SDSS footprint)
• GCNS: Independent selection but highly overlapping with GF21 (14,687/14,699)
• GSPC-WD: Requires XP spectra, G<20.5, red-end incomplete (Teff<5000K)

Partial-sky 100pc samples (3)

Evaluation Metrics

1. χ² goodness-of-fit: Bin-by-bin comparison of simulated and observed distributions
2. Best-fit population age: Age that matches the faint-end decline of the luminosity function
3. Star formation history form: Relative performance of four assumed SFH models
   • Constant star formation rate (Cukanovaite et al. 2023)
   • Late-peaked (Roberts et al. 2025)
   • Early-peaked (Fantin et al. 2019)
   • Bimodal (Mor et al. 2019)

Comparison Baselines

• 40pc sample (O'Brien et al. 2024): 960 white dwarfs, 100% spectroscopic completeness
• Different SFH models: Figure 7 from Roberts et al. (2025)

Implementation Details

• Simulation runs: 50 Monte Carlo runs per configuration
• Error calculation: Square root of sum of standard deviation (systematic uncertainty) and Poisson error (observational)
• Extinction handling: Sample from N(0.02, 0.015²) when AV is missing
• Parallelization strategy: Independent simulations for different SFH and samples

Experimental Results

Main Results: All-sky Samples

GF21 Sample (Figure 4)

Luminosity function:

• Best fit: Roberts et al. (2025) late-peaked SFH, followed by Mor et al. (2019)
• Peak problem: All simulations predict slight upturn at Mbol≈15, observations show downward curvature
• Age constraint: Best fit 11.0 Gyr (within 1σ of default 10.6 Gyr)

Absolute G-band magnitude distribution:

• Consistent trend: Late-peaked and bimodal superior to constant and early-peaked
• Fantin model: Clearly predicts too many faint stars, too few bright stars

GCNS Sample (Figure 5)

• High similarity: Results nearly identical to GF21
• Best-fit age: Exactly 10.6 Gyr
• Conclusion: Cross-validation of two independent Gaia selection methods

GSPC-WD Sample (Figure 6)

• Clear incompleteness: Luminosity function faint-end declines prematurely
• Age bias: Requires reduction to 8.5-9.3 Gyr to match (3σ deviation)
• Cause: Missing white dwarfs with Teff<6000K (red), G≈20 apparent magnitude limit

Key finding: XP spectra provide spectroscopic classification value, but sample incompleteness introduces significant age bias

Main Results: Partial-sky Samples

SDSS Sample (Figure 7)

• Incompleteness: Requires age reduction to 7.5-9.3 Gyr
• Gradient mismatch: Decline slope matches none of the simulations
• Cause: Intentional avoidance of featureless DC-type white dwarfs (Teff<5000K)

DESI DR1 Sample (Figure 8)

• Excellent performance: 10.6 Gyr perfectly fits all SFH models
• Random sampling advantage: Both tiles and targets randomly selected, no color/brightness bias
• Apparent magnitude limit: Empirical G<20 excludes ~5% of simulated sample (Figure 9)
• All-sky extrapolation: Partial-sky results can represent all-sky properties

Breakthrough conclusion: DESI's random sampling strategy achieves unbiased representation of partial-sky regions

4MOST Prediction (Figures 10-11)

• Sky coverage effect: 50% southern coverage has no significant impact on luminosity function (top panel)
• Magnitude limit: G<19.5 causes faint-end incompleteness (bottom panel)
• SFH discrimination ability: Can distinguish four SFH forms within error bars (Figure 11)
• Expected sample size: ~8,000 white dwarfs

Ablation Study: Direct Age Method (Appendix A)

Normalized star formation history dN/dt (Figures A1-A2):

• GF21 and GCNS: Highly consistent with 40pc sample, low early star formation → high star formation rate in recent 5 Gyr
• GSPC-WD: Peak time advanced, steep increase in 5-10 Gyr (due to red-end bias)
• DESI: Between all-sky and SDSS
• Age upper limit: 12-13 Gyr bin, but errors >2 Gyr

Key Experimental Findings

1. Inter-sample consistency:
   • GF21, GCNS, and 40pc show completely consistent star formation history
   • Different Gaia selection criteria (HR diagram position, probability threshold) effects negligible
2. Completeness hierarchy:

   DESI (unbiased) > GF21 ≈ GCNS (mild) > GSPC-WD (moderate) > SDSS (severe)
   

3. Spectroscopy value trade-off:
   • Spectroscopic classification improves atmospheric models (GSPC-WD, SDSS)
   • But selection bias offsets or even exceeds model improvement gains
   • Luminosity and absolute magnitude methods have low spectroscopic dependence
4. Peak mystery:
   • All simulations predict upturn at Mbol≈15
   • All observations show smooth or downward curvature
   • Possible causes: Cooling physics (crystallization?) or selection effects
5. 40pc→100pc extrapolation:
   • Poisson noise significantly reduced (12.5 mag bump in 40pc sample disappears in 100pc)
   • Systematic uncertainties still dominate (Table 2)
   • Methodological robustness verified

Related Work

White Dwarf Star Formation History Studies

• Early work: Winget et al. (1987) first proposed using white dwarf luminosity function for dating
• SDSS era: Rowell (2013), Fantin et al. (2019) early-peaked SFH from SDSS samples
• Gaia revolution:
  • Mor et al. (2019): 20pc sample, bimodal SFH
  • Cukanovaite et al. (2023): 40pc sample, constant SFH, age 10.6±0.7 Gyr
  • Roberts et al. (2025): Three-method comparison, late-peaked SFH

Main-sequence Star Comparison Studies

• Gallart et al. (2024), Fernández-Alvar et al. (2025): Gaia main-sequence SFH
• Objective: Cross-validation of white dwarf and main-sequence results

Gaia White Dwarf Catalogs

• DR2: Gentile Fusillo et al. (2019) - 260,000 candidates
• EDR3: Gentile Fusillo et al. (2021) - 359,000, PWD probability
• DR3 additions:
  • GCNS (Gaia Collaboration 2021): Random forest classification
  • GSPC-WD (Vincent et al. 2024): XP spectra + SDSS synthetic photometry
  • Jiménez-Esteban et al. (2023): Independent XP analysis

Ground-based Spectroscopic Surveys

• SDSS I-IV: Kleinman et al. (2013), Kilic et al. (2025) - ~30% northern sky
• LAMOST: Cui et al. (2012) - Chinese survey
• DESI: Cooper et al. (2023), Manser et al. (2024) - Random sampling strategy
• Future: 4MOST (de Jong et al. 2019), WEAVE (Jin et al. 2024), SDSS-V (Kollmeier et al. 2025)

Cooling Physics

• Bédard et al. (2020): Mass-radius relation and cooling models
• Caron et al. (2023), O'Brien et al. (2024): Lyman-α red-wing corrections
• Crystallization delay: Still-controversial physical process

Conclusions and Discussion

Main Conclusions

1. Star formation history robustness:
   • Late-peaked SFH (Roberts et al. 2025) or constant star formation rate best fits all samples
   • Results independent of specific Gaia selection method or sky coverage
   • 40pc and 100pc sample consistency validates methodological extrapolation
2. Sample completeness ranking:

   DESI DR1 (optimal) > GF21 ≈ GCNS > GSPC-WD > SDSS
   

   • DESI random sampling achieves unbiased partial-sky representation
   • XP spectroscopy value offset by incompleteness
3. Age constraints:
   • Galactic disk age 10.6±0.5 Gyr (consistent with 40pc)
   • Apparent magnitude limits hinder precise faint-end measurement
   • Systematic uncertainties >2 Gyr dominate error budget
4. Partial-sky applicability:
   • Random sampling strategies (DESI, 4MOST) can represent all-sky
   • Target selection bias (SDSS) introduces severe systematic errors
   • 50% sky coverage sufficient (Galactic scale height effects small)
5. Future survey recommendations:
   • 4MOST 100PC will provide ~8,000 southern white dwarfs
   • G<19.5 limit acceptable (can distinguish SFH models)
   • Deep surveys (Vera Rubin) more critical for faint-end

Limitations

1. Physical model uncertainties:
   • Cooling model ±4.5% (crystallization physics incompletely understood)
   • Main-sequence lifetime ±3.5% (metallicity, rotation effects)
   • IFMR scatter (~0.1 M⊙, insufficiently modeled)
2. Single-star assumption:
   • Mass limits exclude binaries but merger history simplified
   • 43% merger delay ratio based on 40pc sample empirical values
3. Atmospheric composition:
   • 25% helium atmosphere fraction crude
   • DC-type <5000K composition uncertainty not propagated
4. Small-scale structure:
   • Mbol≈15 peak discrepancy unexplained
   • Cannot constrain short-timescale SFH variations (<1 Gyr)
5. Volume limit:
   • Halo white dwarf contamination within 100pc negligible
   • Larger volumes require multi-population decomposition

Future Directions

1. Binary populations:
   • 4MOST complete spectra can measure binary parameters
   • Mass ratio, orbital period distributions
   • Mass transfer history constraints
2. Chemical abundances:
   • Spectroscopic metal line measurements
   • Joint population-chemical evolution modeling
   • Comparison with main-sequence Fe/H
3. Deeper samples:
   • Vera Rubin LSST: r<24.5, probing 200pc
   • Constraining Galactic disk age to ±0.5 Gyr
   • Halo-disk separation
4. Physical improvements:
   • 22Ne sedimentation effects on crystallization
   • Magnetic field effects on cooling (10% white dwarfs)
   • 3D atmospheric convection models
5. Machine learning:
   • End-to-end SFH inference (bypassing intermediate parameters)
   • Simulation-observation domain adaptation
   • Uncertainty quantification

In-depth Evaluation

Strengths

1. Systematic and comprehensive:
   • First quantitative comparison of 6 major 100pc samples
   • Cross-validation via three independent methods (luminosity, magnitude, age)
   • Coverage of all-sky and partial-sky, photometric and spectroscopic samples
2. Methodological rigor:
   • 50 Monte Carlo simulations quantify systematic uncertainties
   • Detailed error budget (Table 2)
   • Low-mass problem correction (critical technical detail)
3. High practical value:
   • Provides design guidance for future surveys (4MOST, etc.)
   • Validates partial-sky strategy feasibility
   • Clarifies XP spectroscopy value trade-offs
4. Robust results:
   • Strong inter-sample consistency
   • Agreement with 40pc baseline
   • Reasonable physical explanations (DESI random sampling advantage)
5. Clear presentation:
   • Table 1 comprehensively summarizes sample characteristics
   • Figure 3 intuitively displays survey footprints
   • Appendix A provides complete third method

Shortcomings

1. Peak discrepancy unresolved:
   • Mbol≈15 systematic bias lacks physical explanation
   • May indicate cooling model defects or selection effects
   • Requires independent verification (e.g., spectroscopic Teff)
2. 4MOST prediction simplified:
   • Fiber allocation algorithm (neighbor exclusion) not simulated
   • Time-pressure incompleteness not quantified
   • Actual completeness may be lower than expected
3. SFH form limited:
   • Only 4 parameterized forms tested
   • Short-timescale variations (<1 Gyr) indistinguishable
   • Systematic uncertainties mask fine structure
4. Binary treatment crude:
   • Mass limits simplistic
   • Merger delay binary (on/off)
   • Mass ratio distribution effects unexplored
5. Insufficient statistical testing:
   • Lacks quantitative χ² values or p-values
   • "Best fit" based on visual judgment
   • Bayesian model comparison more rigorous

Impact

Academic contributions:

• Establishes benchmarks for Gaia-era white dwarf population studies
• Demonstrates partial-sky survey representativeness (important methodological conclusion)
• Quantifies marginal utility of XP spectroscopy in large samples

Practical value:

• Directly guides 4MOST, WEAVE observing strategies
• Provides analysis framework for DESI subsequent data releases
• Reference for Vera Rubin and other deep survey design

Reproducibility:

• Code based on publicly available Roberts et al. (2025) framework
• Data from public catalogs (Gaia DR3, DESI DR1)
• Complete simulation parameters listed in Table 2

Potential impact:

• Extension building on Roberts et al. (2025) foundation
• Likely becomes standard reference for 100pc sample selection
• Promotes random sampling strategy adoption in MOS surveys

Applicable Scenarios

Direct applications:

1. Solar neighborhood star formation history research (<100pc)
2. White dwarf sample design and quality control
3. Gaia DR4/DR5 white dwarf catalog validation

Extended applications:

1. Galactic disk-halo decomposition (>100pc samples)
2. Star cluster white dwarf age determination
3. Initial-final mass relation calibration

Inapplicable scenarios:

1. Galactic center/spiral arm non-disk environments
2. Short-timescale starburst events (insufficient resolution)
3. Individual white dwarf precise dating (>10% systematic error)

Key References

Methodological foundations:

• Roberts et al. (2025, MNRAS 538, 2548): 40pc sample three-method comparison
• Cukanovaite et al. (2023, MNRAS 522, 1643): Constant SFH baseline

Gaia white dwarf catalogs:

• Gentile Fusillo et al. (2021, MNRAS 508, 3877): GF21 sample
• Gaia Collaboration et al. (2021, A&A 649, A6): GCNS catalog
• Vincent et al. (2024, A&A 682, A5): GSPC-WD XP spectroscopic sample

Ground-based spectroscopic surveys:

• Kilic et al. (2025, ApJ 979, 157): SDSS 100pc sample
• DESI Collaboration et al. (2025, arXiv:2503.14745): DESI DR1
• de Jong et al. (2019, Messenger 175, 3): 4MOST survey design

Physical models:

• Bédard et al. (2020, ApJ 901, 93): White dwarf cooling and mass-radius relation
• O'Brien et al. (2024, MNRAS 527, 8687): Low-mass problem correction
• Byrne et al. (2025, MNRAS 537, 2433): Main-sequence lifetime model

Overall Assessment: This is a high-quality astrophysical methodology paper that provides important benchmarks for star formation history research in the Gaia era through systematic comparison of multiple 100pc white dwarf samples. Its core contribution lies in quantifying biases from different sample selection strategies, particularly demonstrating the superiority of DESI random sampling and the limited value of XP spectroscopy in large samples. The research provides direct guidance for future surveys (4MOST, Vera Rubin), though improvements remain possible in physical model details and statistical testing. The paper's robust conclusions (late-peaked or constant SFH) provide reliable constraints for Galactic archaeology.