Gravitational-Wave Paleontology

A New Frontier to Probe Massive Stars Across Cosmic History

Floor Broekgaarden��Junior Fellow in The Simons Society of Fellows�Postdoc at Columbia University & The Simons Foundation�Co-PI AstroAI: Center for Astrophysical AI – at Harvard University�Incoming assistant professor at Johns Hopkins University (‘25)

For accessibility:�download slides from my website

Cartoon: Getty Images/iStockphoto

It is challenging to observe Massive Stars…

NASA/ESA/Hubble

c

Black Holes

Neutron Stars

Enrichment

Reionization

Supernovae

Feedback

Star Formation

Cosmology

Transients

Galaxies

Massive

Stars

They live in binaries

Sana + (2012), Moe & DiStefano + (2017)�~70% of massive stars will interact with a companion during their lifetime (Sana + 2012)

STARS IN THE PAST

Gravitational Wave

Paleontology

How did they form?

WHAT WE SEE

ICE/LIGO/T.Pyle/R.Hurt/SXS

LIGO

NASA

stars form

collision

today

STARS IN THE PAST

Gravitational Wave

Paleontology

How did they form?

WHAT WE SEE

How do stars evolve?

mass 2

mass 1

X

X

X

X

X

X

X

X

(Rapid) Population Synthesis Models

C MPAS

Monash University

Ilya Mandel (PI)

Jeff Riley

Ryosuke Hirai

Tim Riley

Teagen Clarke

Mike Lau

COMPAS Collaboration

Harvard University

Lieke van Son

Floor Broekgaarden

Adam Boesky

U. Washington

Tom Wagg

U. of Amsterdam

Serena Vinciguerra

Cambridge U.

Isobel Ramero-shaw

MPA Munich

Selma de Mink

Stephen Justham

Alejandro Vigna-Gomez

KU Leuven

Reinhold Willcox

Johns Hopkins U.

Veome Kapil

Past Developers: Lokesh Khandelwal, Jim Barrett, Poojan Agrawal, Kit Boyett, Debatri Chattopadhyay, Sebastian Gaebel, Fabian Howitt, Floris Kummer, Coen Neijssel

Swinburne U.

Simon Stevenson

Robert Song

U. of Auckland

Avi Vajpeyi

U. of Oregon

JD Merritt

Monash University

Ilya Mandel (PI)

Jeff Riley

Ryosuke Hirai

Tim Riley

Teagen Clarke

Mike Lau

COMPAS Collaboration

Harvard University

Adam Boesky

U. Washington

Tom Wagg

U. of Amsterdam

Serena Vinciguerra

Cambridge U.

Isobel Romero-Shaw

MPA Munich

Selma de Mink

Stephen Justham

Alejandro Vigna-Gomez

KU Leuven

Reinhold Willcox

Johns Hopkins U.

Veome Kapil

Past Developers: Lokesh Khandelwal, Jim Barrett, Poojan Agrawal, Kit Boyett, Debatri Chattopadhyay, Sebastian Gaebel, Fabian Howitt, Floris Kummer, Coen Neijssel

Swinburne U.

Simon Stevenson

Robert Song

U. of Auckland

Avi Vajpeyi

U. of Oregon

JD Merritt

New York Area

Floor Broekgaarden

Lieke van Son

One of the first publicly available codes �

Optimized for running large simulations�of Gravitational Wave sources

April 5 2020

TEAM COMPAS et al. (2022; co lead by FSB)�incl.: Tom Wagg,* Lokesh Khandelwal,* and Floris Kummer*

Binary Population Synthesis

Based on tracks and prescriptions from Hurley+00,02, Pols+98�

COMPAS

TEAM COMPAS et al. (2022; co lead by FSB)�incl.: Tom Wagg,* Lokesh Khandelwal,* and Floris Kummer*

Initial conditions:

• mass 1,
• mass 2,
• separation,
• …

final conditions: mass 1, mass 2, …

COMPAS is publicly available:

https://compas.science/

STARS IN THE PAST

Gravitational Wave

Paleontology

How did they form?

WHAT WE SEE

How do stars evolve?

C MPAS

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

Floor Broekgaarden

Uncertain Formation channel

Isolated Binaries

�

​

Population-III stars

�

Chemically homogeneous evolution

​

​

Mandel & de Mink+16, de Mink & Mandel+16, Marchant+16, Riley+16, du Buisson+16

Kinugawa+14, Belczynski+17, Hijikawa+21,Liu & Bromm+21, Tanikawa+21

Isolated Triples/Multiples�

​

​

e.g. Silsbee & Tremaine+17, Antonini+17, Rodriguez & Antonini+18, Martinez+20 Hamers & Thompson+19

Smarr & Blandford+76, Dominik+15, Kruckow+18, Artale+19,Neijssel+19, Spera+19,

Mapelli+20, Shao & Li+21

​

Globular Clusters

​

​

​

​

Young/Open Star Clusters

​

​

​

​

Nuclear star clusters

​

​

​

​

e.g. Clausen+13, Rodriguez+15

Antonini & Rasio+16, Askar+17, Hong+18, Kremer+20 Ye+20

e.g. Ziosi+14,Mapelli+16+20, Di Carlo+20, Kumamoto+20, Rastello+20, Santoliquido+20 Banerjee+21

e.g. Miller & Lauburg+09, Antonini & Perets+12, Petrovich & Antonini+17, Stephan+19, Arca-Sedda+20, McKernan+20

In the “field”

stars born in isolated binary/triple systems

“Dynamical”

stars born in dense stellar environments

Primordial

�

“Other”

e.g. Bird+16, Ali-Haimoud+18, Raidal+19

See also reviews by Mandel & Farmer (2020), Mapelli (2021), Gerosa & Fishbach (2021), �Mandel & Broekgaarden (2022), Arca-Sedda et al. (2023), de Mink (in prep.)

Fly-bys

�

e.g. Raveh+ 2022

1

Classic channel BH-NS merger:�

Floor Broekgaarden

e.g. Paczynski+76, Smarr & Blandford+76 | Figure based on Tauris+17

Isolated Binary Evolution Pathway

Floor Broekgaarden

Uncertain Formation channel

Isolated Binaries

�

​

Population-III stars

�

Chemically homogeneous evolution

​

​

Mandel & de Mink+16, de Mink & Mandel+16, Marchant+16, Riley+16, du Buisson+16

Kinugawa+14, Belczynski+17, Hijikawa+21,Liu & Bromm+21, Tanikawa+21

Isolated Triples/Multiples�

​

​

e.g. Silsbee & Tremaine+17, Antonini+17, Rodriguez & Antonini+18, Martinez+20 Hamers & Thompson+19

Smarr & Blandford+76, Dominik+15, Kruckow+18, Artale+19,Neijssel+19, Spera+19,

Mapelli+20, Shao & Li+21

​

Globular Clusters

​

​

​

​

Young/Open Star Clusters

​

​

​

​

Nuclear star clusters

​

​

​

​

e.g. Clausen+13, Rodriguez+15

Antonini & Rasio+16, Askar+17, Hong+18, Kremer+20 Ye+20

e.g. Ziosi+14,Mapelli+16+20, Di Carlo+20, Kumamoto+20, Rastello+20, Santoliquido+20 Banerjee+21

e.g. Miller & Lauburg+09, Antonini & Perets+12, Petrovich & Antonini+17, Stephan+19, Arca-Sedda+20, McKernan+20

In the “field”

stars born in isolated binary/triple systems

“Dynamical”

stars born in dense stellar environments

Primordial

�

“Other”

e.g. Bird+16, Ali-Haimoud+18, Raidal+19

See also reviews by Mandel & Farmer (2020), Mapelli (2021), Gerosa & Fishbach (2021), �Mandel & Broekgaarden (2022), Arca-Sedda et al. (2023), de Mink (in prep.)

Fly-bys

�

e.g. Raveh+ 2022

1

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

!! Rerun simulations for many different formation channels !!

2

Uncertain Initial conditions (where/when/how do stars form?)

Binary fraction

IMF

Initial mass ratios

Initial Periods

Initial Rotation?

Properties Environment

Birth Metallicities

Birth Metallicities can drastically alter the star’s evolution

e.g. Hurley +2000, Maeder + 1992, Heger+2003, Langer+2012

See also e.g. Chruslinska+19, Neijssel+19, Tang+2020,

Broekgaarden+21+22, Santoloquido+21, Briel+2022

Initial eccentricity?

Giacobbo & Mapelli (2019) 🡪

mass 2

mass 1

Z

Separation

Eccentricity

Metallicity

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

!! Rerun simulations for many different formation channels !!

!! Run simulations for many different initial conditions!!

Gravitational-wave sources are a Rare outcome among stars

3

only ~1 merger!

E.g. Belczynski+02

1 000 000 binaries

Neijssel (2019)

Leads to large Poisson (sampling) noise

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

!! Rerun simulations for many different formation channels !!

!! Run simulations for many different initial conditions!!

!! Run more simulations to find enough data points !!

Classic channel BH-NS merger:�

Floor Broekgaarden

4

(Binary) Stellar Evolution is Uncertain

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

!! Rerun simulations for many different formation channels !!

!! Run simulations for many different initial conditions!!

!! Run more simulations to find enough data points !!

!! Rerun simulations to test uncertain stellar evolution !!

Gravitational-Wave Paleontology today:

Floor Broekgaarden

With so many uncertainties, can we learn anything from Gravitational Waves?

Cartoon: xkcd

Floor Broekgaarden

Identify and quantify these challenges in

Gravitational-Wave Paleontology�

Address and tackle the different barriers

Adam Boesky

Kaylie Hausknecht

Floris Kummer

Lokesh

Khandelwal

Simone Abeni

Tom Wagg

Miranda Harkess

Sasha Levina

Caua Rodrigues

Ana Lam

current group

former group members

Meera Desawale

stellar evolution

Astro-statistics

Gravitational waves

Cosmic star formation history

Enrichment

AI/ML

transients

visualizations/data science/software

Floor Broekgaarden

Amedeo Romagnolo

The Gravitational Wave Paleontology Lab:

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

New sampling algorithm:

Floor Broekgaarden

Adaptive Importance Sampling (AIS)

Marin+06, Douc+07, Owen+09, Martino+15

AIS: Torrie & Valleau 1977, Hesterberg 1995,

Cappe+2004, Pennanen & Koivu 2006,

Cornuet+2012, Ortiz & Pack Kaelbling (2013), …

Simulating The Rare Outcomes Of Populations

With AIS For Efficient Learning

3

Gravitational-wave sources are Rare

Random

Explore & Refine

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

< 65 tries >

< 97 tries >

Random

Explore & Refine

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

< 65 tries >

< 97 tries >

mass 2

mass 1

Z

Separation

Eccentricity

Metallicity

Now in Higher dimensions!

Curse of dimensionality�improved using GenAIS (generational Adaptive Importance Sampling)

Khandelwal et al. (in prep.)

Based on Wraith+2009

Floris Kummer

Lokesh

Khandelwal

Stephen Justham

Floor Broekgaarden

Reduce computational costs: emulate (“interpolate”) output

Uncertainty Quantification

Lin, Bingham, Broekgaarden, Mandel (2022), Annals of applied Statistics

traditional

Uncertainty Quantification

Derek Bingham

Luyao Lin

faster

100x

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

2. Simulate binary

evolution:

How do stars

live & die?

When & Where

do stars form? (metallicity)

1. Simulate �Cosmic star formation:

​

cf Chruslinska+19, Belczynski+19 Neijssel+19, Tang+19…

Assume initial binary properties:

e.g. Sana+2012, Moe & DiStefano+ 17, Banyard+22, Offner+22, Shenar+22,…

When & Where

do stars form? (metallicity)

2

28 Cosmic models�based on Neijssel+19 �

1. Uncertainties in Cosmic modeleling:

• star formation rate
• Galaxy mass distribution
• Metallicities of galaxies/stars
• High redshift observations

​

Uncertain Initial �Conditions

Floor Broekgaarden

Population synthesis codes: isolated binaries

Isolated Binaries

�

​

Stevenson+17,19, Barrett+18, Vigna-Gomez+18, Broekgaarden+19

Based on stellar evolution tracks from Hurley+00,02, Pols+98

�

COMPAS is publicly available https://compas.science/

COMPAS collaboration, Jeff Riley et al., 2021 (incl FSB as 1 of 4 lead authors)

20 Stellar models��

2. Uncertainties in Stellar modeling:

• Supernovae remnant masses
• Mass transfer
• Common envelope phase
• Stellar winds

​

How do stars

live & die?

When & Where

do stars form? (metallicity)

How do stars

live & die?

Carl Knox

Floor Broekgaarden

Investigate & Quantify impact of uncertainties:�

cf. Chruslinska+19, Boco+19, Neijssel+19, Santoliquido+21

Stellar Evolution models

Broekgaarden et al. (2021)

Stellar Evolution models

Floor Broekgaarden

Investigate & Quantify impact of uncertainties:�

cf. Chruslinska+19, Boco+19, Neijssel+19, Santoliquido+21

Stellar Evolution models

Cosmic models

Broekgaarden et al. (2021)

Conclusion:

The expected properties of Black Hole-Neutron Star Mergers are drastically impacted by uncertainties in both Stellar evolution and the cosmic evolution history

HERE

Ilya Mandel

Andrew Levin

Michela Mapelli

2 BH-NS

Observations

My Simulations

Broekgaarden & Berger (2021)

Broekgaarden (in prep.)

When & Where

do stars form? (metallicity)

How do stars

live & die?

Carl Knox

Floor Broekgaarden

cf. Chruslinska+19, Boco+19, Neijssel+19, Santoliquido+21

Stellar Evolution models

Cosmic models

Investigate impact of uncertainties:�

Broekgaarden et al. (2021, 2022)

Floor Broekgaarden

BHBH rates dominated by “cosmic history”

cf. Chruslinska+19, Boco+19, Neijssel+19, Santoliquido+21

Stellar Evolution models

Cosmic models

Broekgaarden et al. (2022)

Floor Broekgaarden

NSNS rates dominated by “stellar evolution”

cf. Chruslinska+19, Boco+19, Neijssel+19, Santoliquido+21

Broekgaarden et al. (2022)

Stellar Evolution models

Cosmic models

Floor Broekgaarden

BHBH formation rates have a steep function of metallicity

Metallicity

Stellar Evolution models

Broekgaarden et al. (2022)

NSNS formation rates are flatter function of metallicity

Metallicity

Stellar Evolution models

Broekgaarden et al. (2022)

Floor Broekgaarden

Matching BHBH, BHNS and NSNS rates

Broekgaarden et al. (2022)

expected detectable mass distribution

Observed rates from O1+O2

Floor Broekgaarden

BH-BH

Change in Stellar evolution models

Cosmic models

Broekgaarden et al. (2022)

expected detectable mass distribution

Observed rates from O1+O2

Floor Broekgaarden

BH-NS

Change in Stellar evolution models

Broekgaarden et al. (2022)

Conclusion I:

The expected properties of BH-NS, BH-BH, and NS-NS Mergers are drastically impacted by uncertainties in both Stellar evolution and the cosmic evolution history

Conclusion II:

Uncertainties impact each flavor in different ways

🡪 Simultaneous constraints can be used to tackle the Uncertainty Challenge

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

Floor Broekgaarden

Uncertain Formation channel

Isolated Binaries

�

​

Population-III stars

�

Chemically homogeneous evolution

​

​

Mandel & de Mink+16, de Mink & Mandel+16, Marchant+16, Riley+16, du Buisson+16

Kinugawa+14, Belczynski+17, Hijikawa+21,Liu & Bromm+21, Tanikawa+21

Isolated Triples/Multiples�

​

​

e.g. Silsbee & Tremaine+17, Antonini+17, Rodriguez & Antonini+18, Martinez+20 Hamers & Thompson+19

Smarr & Blandford+76, Dominik+15, Kruckow+18, Artale+19,Neijssel+19, Spera+19,

Mapelli+20, Shao & Li+21

​

Globular Clusters

​

​

​

​

Young/Open Star Clusters

​

​

​

​

Nuclear star clusters

​

​

​

​

e.g. Clausen+13, Rodriguez+15

Antonini & Rasio+16, Askar+17, Hong+18, Kremer+20 Ye+20

e.g. Ziosi+14,Mapelli+16+20, Di Carlo+20, Kumamoto+20, Rastello+20, Santoliquido+20 Banerjee+21

e.g. Miller & Lauburg+09, Antonini & Perets+12, Petrovich & Antonini+17, Stephan+19, Arca-Sedda+20, McKernan+20

In the “field”

stars born in isolated binary/triple systems

“Dynamical”

stars born in dense stellar environments

Primordial

�

“Other”

e.g. Bird+16, Ali-Haimoud+18, Raidal+19

See also reviews by Mandel & Farmer (2020), Mapelli (2021), Gerosa & Fishbach (2021), �Mandel & Broekgaarden (2022), Arca-Sedda et al. (2023), de Mink (in prep.)

Fly-bys

�

e.g. Raveh+ 2022

1

Floor Broekgaarden

Ping us if your paper is missing!

Publicly available code/data:

Mandel & Broekgaarden (2022) Living Review in Relativity

1

Floor Broekgaarden

Uncertain Formation channel

Isolated Binaries

�

​

Population-III stars

�

Chemically homogeneous evolution

​

​

Mandel & de Mink+16, de Mink & Mandel+16, Marchant+16, Riley+16, du Buisson+16

Kinugawa+14, Belczynski+17, Hijikawa+21,Liu & Bromm+21, Tanikawa+21

Isolated Triples/Multiples�

​

​

e.g. Silsbee & Tremaine+17, Antonini+17, Rodriguez & Antonini+18, Martinez+20 Hamers & Thompson+19

Smarr & Blandford+76, Dominik+15, Kruckow+18, Artale+19,Neijssel+19, Spera+19,

Mapelli+20, Shao & Li+21

​

Globular Clusters

​

​

​

​

Young/Open Star Clusters

​

​

​

​

Nuclear star clusters

​

​

​

​

e.g. Clausen+13, Rodriguez+15

Antonini & Rasio+16, Askar+17, Hong+18, Kremer+20 Ye+20

e.g. Ziosi+14,Mapelli+16+20, Di Carlo+20, Kumamoto+20, Rastello+20, Santoliquido+20 Banerjee+21

e.g. Miller & Lauburg+09, Antonini & Perets+12, Petrovich & Antonini+17, Stephan+19, Arca-Sedda+20, McKernan+20

In the “field”

stars born in isolated binary/triple systems

“Dynamical”

stars born in dense stellar environments

Primordial

�

“Other”

e.g. Bird+16, Ali-Haimoud+18, Raidal+19

See also reviews by Mandel & Farmer (2020), Mapelli (2021), Gerosa & Fishbach (2021), �Mandel & Broekgaarden (2022), Arca-Sedda et al. (2023), de Mink (in prep.)

Fly-bys

�

e.g. Raveh+ 2022

1

Classic channel BH-NS merger:�

Floor Broekgaarden

e.g. Paczynski+76, Smarr & Blandford+76 | Figure based on Tauris+17

Isolated Binary Evolution Pathway

Classic channel BH-NS merger:�

Floor Broekgaarden

e.g. Paczynski+76, Smarr & Blandford+76 | Figure based on Tauris+17

Isolated Binary Evolution Pathway

Classic channel with a Common Envelope

Classic channel BH-NS merger:�

FB

Floor Broekgaarden

e.g. Paczynski+76, Smarr & Blandford+76 | Figure based on Tauris+17

Isolated Binary Evolution Pathway

Only stable mass transfer channel without a Common Envelope

different

codes and

stellar evolution

uncertainties

Do Majority of BH+BH mergers experience an unstable, “common envelope”, phase??

Yes!

No!

Do Majority of BH+BH mergers experience an unstable, “common envelope”, phase??

Conclusion

The expected formation pathway within the isolated formation channel is drastically impacted by uncertainties in Stellar evolution

Ana Lam

Analyzing formation pathways of massive stars

Lam et al. (in prep)

Floor Broekgaarden

Extra dimension: Redshift dependent rates

Adam Boesky*, Broekgaarden ,

Berger (2024a, 2024b, submitted)

See also van Son et al. (2022), Santoloquido (2022) and many others

Stellar Evolution �models

Floor Broekgaarden

Redshift dependent rates

Isolated binary evolution

Globular clusters

Pop III stars

Broekgaarden (in prep)

Floor Broekgaarden

What will we learn from looking at redshift dependent rates?

Isolated binary evolution

Globular clusters

Pop III stars

Redshift dependent rates

Broekgaarden (in prep)

Floor Broekgaarden

What will we learn from looking at redshift dependent rates?

And redshift dependent distribution functions?

Isolated binary evolution

Globular clusters

Pop III stars

Chemically

homogeneous

Redshift dependent rates

Broekgaarden (in prep)

Sasha Levina

Lieke van Son

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

Uncertainties are also common in the lives of scientists

“Underrepresented minorities^1-3 and especially multiple marginalized minorities such as women of color face from early childhood through university most barriers including harassment, anxiety, and discrimination,^4-7 leading to their structural exclusion from the science community.”^8-10

[1] The AIP National Task Force to Elevate African American Representation in Undergraduate Physics & Astronomy (TEAM-UP), (2019), [2] 2022 Decadal Survey Astronomy, [3] Cooper & Berry (2019), [4] Norman et al. (2013), [5] Clancy et al. (2017), [6] Richey et al. (2019), [7] Steve Fund and Jed Foundation 2017, [8] Pauls, S. (2014, July 14 , [9] National Research Council. 2013 , [10] Anne Marie Porter and Rachel Ivie (2019)

CuWIP survey

Experiences

of Workplace Harassment

25.7%

74.3%

No experiences

Aycock et al. 2024/APS

Uncertainties are also common in the lives of scientists

10 ways to improve support for Workplace Civility in Astronomy

1. Implement a public code of ethics �
2. Designate point persons for harassment concerns �
3. Provide adequate training for the persons entrusted to deal with code of ethics concerns�
4. When concerns are brought forward “instigators should be swiftly, justly, and consistently sanctioned”�
5. Organize diversity and cultural awareness training �
6. Organize bystander training

​

• Move beyond legal compliance to address culture and climate �
• Improve transparency and accountability�
• Connect members to local resources�
• Avoid saying “They/I had the best intentions”�
• [bonus] Learn more about the topic

Cortina et al. (2013), Clancy et al. (2014)

See also “Astronomy Allies”

Cortina et al. (2013)

Cortina et al. (2013), National Research Council (2013),

NASEM (2018)

Ackerman et al. (2018), NASEM (2018)

See, e.g., “Particles for Justice” and the “ASP Resource Guides”

USA equal employment opportunity commission 2016 report

Upcoming�Events

Early Career Astronomers Resources (including online workshops on “how to apply to Postdoc/faculty)”

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

BH-BH

NS-NS

BH-NS

The Big Data Era for Double Compact Object mergers:

today

Number of observations

Animation from �Broekgaarden (2024; accepted ApJS yesterday!)

Gravitational Wave Paleontology “a unique frontier to study massive stars for $50/star” if we can connect GW sources to their formation pathway histories

Floor Broekgaarden

From Hall & Evans (2019)

LIGO

Floor Broekgaarden

From Hall & Evans (2019)

We might see *every* merging BH-BH in the Universe!

Cosmic Explorer

Einstein Telescope

LIGO

“Gravitational-Wave Astrophysics is one of the most exciting frontiers in science”

“The US-based Cosmic Explorer is central to achieving the science vision laid out in the survey’s roadmap”

- 2020 Decadal Survey on Astronomy & Astrophysics

2020

2040

2030

2025

2035

JWST

early star formation

LSST

HWO

Gaia

DR4

DR5

ngVLA

Hubble

DESI/Euclid

VLA/Chime/LOFAR/ALMA/Meerkat

SKA/DSA-2000

Roman

Microlensing BHs

Stellar Streams, Disrupted halo dwarfs�dormant black holes

Nustar/Chandra/NICER

AXIS/Athena?

Luminous Red Novae, �kilonovae, PISN

Pulsars,

FRB, galaxy

COSI

Fermi/Swift

Floor Broekgaarden

Other observational constraints?

Detecting BH-NS in our Milky Way

Pulsars binaries

e.g. BH-PSR: Debatri Chattopadhyay, Simon Stevenson, Jarrod Hurley, Matthew Bailes & FSB (2021), MNRAS

BH/NS binaries in LISA

Thomas Wagg*, FSB, Selma de Mink et al. (2022)

The Uncertainty Challenge

How do stars evolve?

1

2

3

4

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Uncertain

Stellar

Evolution

Gravitational Wave Paleontology

Today

Future

The Uncertainty Challenge

Presents hurdles in simulations of GW progenitors

STARS IN THE PAST

Gravitational Wave

Paleontology

WHAT WE SEE

How do stars evolve?

STARS IN THE PAST

Gravitational Wave

Paleontology

WHAT WE SEE

How do stars evolve?

How did they form?

STARS IN THE PAST

Gravitational Wave

Paleontology

How did they form?

WHAT WE SEE

How do stars evolve?

STARS IN THE PAST

Gravitational Wave

Paleontology

How did they form?

WHAT WE SEE

How do stars evolve?

Massive

Stars

Massive

Stars

Black Holes

Neutron Stars

Enrichment

Reionization

Supernovae

Feedback

Star Formation

Cosmology

Transients

Galaxies

Adam Boesky

Kaylie Hausknecht

Floris Kummer

Lokesh

Khandelwal

Simone Abeni

Tom Wagg

Miranda Harkess

Sasha Levina

Caua Rodrigues

Ana Lam

current group

former group members

Meera Desawale

stellar evolution

Astro-statistics

Gravitational waves

Cosmic star formation history

Enrichment

AI/ML

transients

visualizations/data science/software

Floor Broekgaarden

Amedeo Romagnolo

The Gravitational Wave Paleontology Lab:

You?

Or join our upcoming events/efforts:

Adam Boesky

Kaylie Hausknecht

Floris Kummer

Lokesh

Khandelwal

Simone Abeni

Tom Wagg

Miranda Harkess

Sasha Levina

Caua Rodrigues

Ana Lam

current group

former group members

Meera Desawale

stellar evolution

Astro-statistics

Gravitational waves

Cosmic star formation history

Enrichment

AI/ML

transients

visualizations/data science/software

Floor Broekgaarden

Amedeo Romagnolo

The Gravitational Wave Paleontology Lab:

You?

​

Thank you!

​

Gravitational-wave paleontology is a promising new frontier to study Massive stars across cosmic time.�

We need stellar evolution simulations to make most of the rapidly growing gravitational-wave data!��Simulations are limited by the Uncertainty Challenge: �uncertainties in both the initial conditions �(cosmic history), stellar evolution, and formation �pathways can drastically impact the expected �properties of BH-BH, BH-NS, and NS-NS mergers.

�We need to identify and quantify these uncertainties �& identify areas where we can start constraining �and addressing uncertainties. �

Thoughts on how to help us?

Cartoon: xkcd

Extra slides

Slide: Ethan Siegel

“We are made of star-stuff”

Carl Sagan

“we are mostly made of massive star-stuff”

Floor Broekgaarden

Where do Black Holes Collide?

Floor Broekgaarden

Hints from Masses…

Abbott et al (2022)

Upper Black Hole mass gap

Floor Broekgaarden

Hints from observed spins…

Callister+(2022)

Floor Broekgaarden

Finding Formation channel sub-populations

Godfrey+(2023)

Population of low mass, low (aligned) spin of BBHs that contributes to 82% of the population

See also Callister+2021,2022, Li+2022, Edelman+2022, Tong+2022, Fishbach+2023

But see Cheng+2023 for challenges for doing this based on synthetic universes

Floor Broekgaarden

1) Exploring phase

prior

Floor Broekgaarden

1) Exploring phase

2) Create adapted distribution

Gaussian

around hits

prior

Gaussian width

Floor Broekgaarden

1) Exploring phase

2) Create adapted distribution

3) Refinement phase

Gaussian

around hits

prior

Gaussian width

When to switch from exploring to refinement?

Floor Broekgaarden

Missing an “island”

Getting more “hits”

uncertainty

View From Outside the Viewing Sphere - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/Photograph-of-the-constellation-Orion-in-the-night-sky-thus-evidently-a-view-from-the_fig8_325315980 [accessed 21 Mar, 2024]

​

Orion

How many massive stars among these?

Answer:

At least 12!!

View From Outside the Viewing Sphere - Scientific Figure on ResearchGate. Available from: https://www.researchgate.net/figure/Photograph-of-the-constellation-Orion-in-the-night-sky-thus-evidently-a-view-from-the_fig8_325315980 [accessed 21 Mar, 2024]

​

Orion

How many massive stars among these?

Answer:

At least 12!!

cf Chruslinska+19, Belczynski+19 Neijssel+19, Tang+19…

Assume initial binary properties:

e.g. Sana+2012, Moe & DiStefano+ 17, Banyard+22, Offner+22, Shenar+22,…

When & Where

do stars form? (metallicity)

2

Uncertain Initial �Conditions

cf Chruslinska+19, Belczynski+19 Neijssel+19, Tang+19…

Assume initial binary properties:

e.g. Sana+2012, Moe & DiStefano+ 17, Banyard+22, Offner+22, Shenar+22,…

When & Where

do stars form? (metallicity)

2

Uncertain Initial �Conditions

cf Chruslinska+19, Belczynski+19 Neijssel+19, Tang+19…

Assume initial binary properties:

e.g. Sana+2012, Moe & DiStefano+ 17, Banyard+22, Offner+22, Shenar+22,…

When & Where

do stars form? (metallicity)

2

28 Cosmic models�based on Neijssel+19 �

1. Uncertainties in Cosmic modeleling:

• star formation rate
• Galaxy mass distribution
• Metallicities of galaxies/stars
• High redshift observations

​

Uncertain Initial �Conditions

Floor Broekgaarden

Population synthesis codes: isolated binaries

Isolated Binaries

�

​

Stevenson+17,19, Barrett+18, Vigna-Gomez+18, Broekgaarden+19

Based on stellar evolution tracks from Hurley+00,02, Pols+98

�

COMPAS is publicly available https://compas.science/

COMPAS collaboration, Jeff Riley et al., 2021 (incl FSB as 1 of 4 lead authors)

20 Stellar models��

2. Uncertainties in Stellar modeling:

• Supernovae remnant masses
• Mass transfer
• Common envelope phase
• Stellar winds

​

Floor Broekgaarden

Matching BHBH, BHNS and NSNS rates

Broekgaarden et al. (2022)

expected detectable mass distribution

Observed rates from O1+O2

Floor Broekgaarden

BH-NS

Change in Stellar evolution models

BHNS mass distributions are dominantly impacted by stellar evolution models

Broekgaarden et al. (2022)

Conclusion I:

The expected properties of BH-NS, BH-BH, and NS-NS Mergers are drastically impacted by uncertainties in both Stellar evolution and the cosmic evolution history

Conclusion II:

Uncertainties impact each flavor in different ways

🡪 Simultaneous constraints can be used to tackle the Uncertainty Challenge

Zoom in on simulations

Many different formation channels

Different flavors �BNS always CE

Stellar Evolution models

Only Stable mass transfer

Classic (Common Envelope)

Zoom in on simulations

Many different formation channels

Different flavors �BNS always CE

Only Stable mass transfer

Common Envelope at 1^st mass transfer

Classic (Common Envelope)

Double Core Common Envelope

BHNS, BNS, BBH channel changes per

1. Merger flavor
2. Stellar evolution model

5

Improving Stellar Tracks:�POSYDON

Population Synthesis

Hurley’s SSE/BSE code from 2000/2002

Original: xkcd

Hurley+2000

e.g. StarTrack, binary_c, COMPAS,

COSMIC

5

Improving Stellar Tracks:�

Population Synthesis

Original: xkcd

See also, METISSE, COMBINE

POSYDON Fragos +22

SEVN: Iorio+23

5

Choice of stellar Tracks impact Stellar Evolution �METISSE

Star’s property at onset of mass transfer can drastically differ for different stellar tracks

Agrawal+23

Intuitively:

Initially most massive star

M_A

M_B

_ZAMS

M_B

M_A

M_A -> SN

M_B

M_A

Chapter 5

forms more massive BH

Other dimensions: Which black hole spins?

Broekgaarden, Stevenson & Thrane (2022)

less massive BH spins

Intuitively:

Initially most massive star

M_A

M_B

_ZAMS

M_B

M_A

M_A -> SN

M_B

M_A

Chapter 5

Mass transfer can alter this:

Initially less massive star

M_A

M_B

_ZAMS

M_B

M_A

M_B

M_A

forms more massive BH

forms more massive BH

“mass ratio reversal”

Other dimensions: Which black hole spins?

Broekgaarden, Stevenson & Thrane (2022)

less massive BH spins

more massive BH spins

Broekgaarden, Stevenson & Thrane (2022)

Mass ratio reversal for BH-BH is common: typically >30%

Other dimensions: Which black hole spins?

14

8

Chapter 5

Conclusion:

Which Black Hole is spinning, can inform scientists about mass transfer and Stellar evolution uncertainties

Broekgaarden, Stevenson & Thrane (2022)

Test in observations:

• Hints that ~80% of binary black holes have both Black Holes spinning

Floor Broekgaarden

Arca-Sedda (2023)

B-POP

1

Modeling multiple Formation channels

Floor Broekgaarden

Understanding initial conditions

Srinivasan+(2023)

See also Chruslinka+2021, Broekgaarden+2021, Giacobbo&Mapelli+2020; Iorio+2023

Higher birth metallicity 🡪

Iorio+(2023)

2

2

Creating fast globular cluster codes to probe Cluster conditions

Kritos + (2023)

Rapster

3

Reducing Simulations using active learning algorithm

3

Psy-cris

Rocha+22

Reduction to 25%

computational costs

Reducing Simulation costs using MCMC

3

Mandel+22

Analytical fit to Peters+64 formula

Update on DartBoard (Andrews+19) by including sampling in hyper-parameters

Wong+22

GW150914

4

5

Constrain Stellar Evolution Physics: Stripped Stars

Floor Broekgaarden

A few found, e.g., as Black Hole Imposters through spectroscopy:

• LB-1 (Liu +2019, 2020, Shenar+2020, El-Badry+2020)
• HR 6819 (Rivinius+2020, Bodesteiner+2020, El-Badry+2021, Frost+2022, Romagnolo+2022)

See also e.g., Gies et al. 1998; Groh et al. 2008; Peters et al. 2015; Peters et al. 2013; Wang et al. 2017).�

4

5

Constrain Stellar Evolution Physics: Stripped Stars & Stellar Winds

Done using three UV filters spanning 1928–2600 A with a resolution of 2.5 ′′.

Discovery of the missing intermediate-mass helium stars stripped in binaries�Drout & Gotberg et al. 2023a�Gotberg & Drout et al. 2023b

4

5

Constrain Stellar Evolution Physics: Supernovae/Kilonovae

(Early) supernova spectra provides key information on the temperature, density, and chemical abundances of the ejecta

e.g. Vasylyev+2022,+2023, Bostroem+2023

Bostroem+2023

5

Using properties of galaxies:

Rauf+ (2023)

Volumetric GW rate for galaxies in observed survey

Completeness

Stevance+ (2023)

Different behavior to the same parameter:

Broekgaarden (2021, 2022),

Briel (2021;2022),

Santoloquido (2021), Dorozsmai &Toonen (2022),

Stevenson & Teagan (2022)

Broekgaarden+ (in prep.)

Boesky, Broekgaarden (in prep.)

1. Only very few formation channels can be “ruled out” based on rates alone

2. Population Synthesis predicted merger rates depend on many uncertainties

Who Ordered That? Unequal-mass binary black hole mergers have larger effective spins

Callister+21

LIGO–Virgo correlations between mass ratio and effective inspiral spin: testing the active galactic nuclei channel

McKernan+21

Floor Broekgaarden

What can we learn from combining GW information with other observational constraints?

Predictions for GW merger rates and other cosmological transients from the same population model. By Eldridge+18

Multi waveband observations with LISA + LIGO

https://media.giphy.com/media/xTiTnooneW4SYfch8Y/giphy.gif

1

2

3

4

5

Form, Initial Conditions Stellar Evolution Stellar evolution, Rare (FIReE!)

Where WIREE

Uncertain

Formation

Channels

Uncertain

Initial �Conditions

They are Rare (sampling

Uncertainties)

Stellar Evolution (I)

Stellar Evolution (II)

How do stars evolve?

The GW Paleontology Uncertainty Challenge

!! Rerun simulations for many different formation channels !!

!! Run simulations for many different initial conditions!!

!! Run more simulations to find enough data points !!

!! Rerun simulations to test uncertain stellar evolution !!

Move to more expensive (detailed?) simulations!

5

(Binary) Stellar Evolution is really Uncertain !!!

Common-envelope phase modeling

In most population synthesis simulations:

See e.g., Ivanova + 2013

Credit: Sebastian Ohlmann / HITS

Common-envelope phase modeling

Hydro simulations

Yes!

No!

Do Majority of BH+BH mergers experience an unstable, “common envelope”, phase??

Broekgaarden+ (in prep.)

No!

Do Majority of BH+NS mergers experience an unstable, “common envelope”, phase??

Yes!

Broekgaarden+ (in prep.)

Boesky, Broekgaarden (in prep.)

Different codes

Yes!

Yes! 1^st mas transfer & “Double Core”

Broekgaarden+ (in prep.)

Boesky, Broekgaarden (in prep.)

Different codes

Different behavior to the same parameter across models:

Broekgaarden (2021, 2022),

Briel (2021;2022),

Santoloquido (2021), Dorozsmai &Toonen (2022),

Stevenson & Teagan (2022)

Broekgaarden+ (in prep.)

Boesky, Broekgaarden (in prep.)

See also, e.g., van Son (2022), Briel (2022), Belczynski (2022)

BH-BH mass distribution per channel

Only Stable mass transfer

Common Envelope at 1^st mass transfer

Broekgaarden et al. (in prep)

Classic (with Common Envelope)

2

Inferring uncertain star formation histories (globular clusters)

Fishbach & Fragione (2023)

~61% of BBHs are

dynamically assembled

2

Creating fast globular cluster codes to probe Cluster conditions

Kritos + (2023)

Rapster

3