1 of 62

(What) Can We Learn about the Lives & Deaths of Massive Stars from Gravitational Wave Mergers?

Floor Broekgaarden

Center for Astrophysics | Harvard & Smithsonian

Aspen 2022

@FloorAstro

Cover images credit: pngegg (diver) – LIGO/T. Pyle (GW merger) – adjusted by FB

With contributions from: Edo Berger, Simon Stevenson, Lieke van Son, Tom Wagg, Alejandro Vigna-Gomez, Martyna Chruslinska, Debatri Chattopadhyay, Ilya Mandel, Stephen Justham, Selma de Mink, Coen Neijssel, Will Farr, TEAM COMPAS, and many others

A “flipped” talk

Full pptx & PDF for accessibility available on my website

2 of 62

“Flipped” Talk

Flipped:

~15-20 min: talk (me + you!!)

~2 min: Quick (v. urgent) Questions (you)

15 min: ”Break out Discussion rooms” (we)

20 min: reconvene + discussion with everyone (we)

~2 min: closing

Use your Post-it notes!

3 of 62

Let’s practice: How are you feeling today?

Dynamic!

(a bit) tired

Confused

Excited

Other

4 of 62

We are at the tipping point of entering

“The Big Data Era for NS/BH binaries”

 

 

Illustration by Jon Krause 

5 of 62

BH-BH

NS-NS

BH-NS

We are here! ~100 detections (GWTC-3)

Figure: Baibhav+19

 

Cosmic Explorer Einstein Telescope

3G

2G

We are at the tipping point of entering

“The Big Data Era for NS/BH mergers:”

6 of 62

Floor Broekgaarden

Higher redshifts

BH/NS systems spiral in over Myr-Gyr timescales

Images credit: pngegg (diver) – NASA (GW merger) – adjusted by FB

-> GW detections probe Massive Stars throughout “Cosmic History”!

7 of 62

Floor Broekgaarden

Higher redshifts

BH/NS systems spiral in over Myr-Gyr timescales

Images credit: pngegg (diver) – NASA (GW merger) – adjusted by FB

-> GW detections probe Massive Stars throughout “Cosmic History”!

8 of 62

Floor Broekgaarden

Higher redshifts

BH/NS systems spiral in over Myr-Gyr timescales

Mix of Myr-Gyr gravitational-wave inspiral times makes GW detections unique probes to study the lives & deaths of massive stars throughout Cosmic History

LIGO,

Virgo,

KAGRA

-> GW detections probe Massive Stars throughout “Cosmic History”!

Images credit: pngegg (diver) – NASA (GW merger) – adjusted by FB

9 of 62

Floor Broekgaarden

Higher redshifts

BH/NS systems spiral in over Myr-Gyr timescales

Mix of Myr-Gyr gravitational-wave inspiral times makes GW detections unique probes to study the lives & deaths of massive stars throughout Cosmic History

3G GW networks (CE & ET)

3G GW networks will enable to probe mergers at higher redshift where we might observe different properties, species and/or formation channels

Images credit: pngegg (diver) – NASA (GW merger) – adjusted by FB

-> GW detections probe Massive Stars throughout “Cosmic History”!

We are at the tipping point of the “Big Data” era of GWs.

3G detectors will uniquely enable us to study the “stellar graveyard” of BHs/NSs as a function of redshift

10 of 62

How do these BH/NS systems form? What can we learn from their detections?

Images credit: pngegg (diver) – LIGO/T. Pyle (GW merger) – adjusted by FB

11 of 62

Floor Broekgaarden

How do BH/NS mergers form?

See also review by Mandel & Broekgaarden (subm. to LRR), Mapelli (2021), Gerosa & Fishbach (2021)

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

12 of 62

Floor Broekgaarden

Which channel below contributes > 20% of current GW detections?

See also review by Mandel & Broekgaarden (subm. to LRR), Mapelli (2021), Gerosa & Fishbach (2021)

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

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

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

13 of 62

Floor Broekgaarden

How do BH/NS mergers form?

See also review by Mandel & Broekgaarden (subm. to LRR), Mapelli (2021), Gerosa & Fishbach (2021)

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

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

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

Compare population synthesis simulations to observed rates and properties (masses, mass ratio) of BH/NS mergers to learn about massive stars!

14 of 62

Floor Broekgaarden

Does this sound familiar?

Yes!

No!

15 of 62

Floor Broekgaarden

What I work on lately / my message of tonight:

Gravitational waves will teach us everything!!!!

(the hope of) Population Synthesis modellers

16 of 62

Floor Broekgaarden

What I work on lately / my message of tonight:

Gravitational waves will teach us everything!!!!

model uncertainties

(the hope of) Population Synthesis modellers

Not so fast..

See also:

Chruslinska et al. (2019)

Belczynski et al. (2021),

the discussion in Zevin et al. (2021),

Broekgaarden et al. (2021a, 2021b)

17 of 62

What Can We Learn from GW Merger Rates…

Images credit: pngegg (diver) – LIGO/T. Pyle (GW merger) – adjusted by FB

18 of 62

Floor Broekgaarden

1. Comparing to Observations: Merger Rates

Ping us if your paper is missing!

Publicly available code/data:

Mandel & Broekgaarden (2021) Living Review in Relativity

1. Only very few formation channels can be “ruled out”

2. Population Synthesis predicted BH/NS merger rates span enormous ranges

19 of 62

Floor Broekgaarden

Common envelope physics

e.g. Dominik+12, Ivanova+20, Marchant+21

Muhammad & Noam+21

Stellar Winds

e.g. Renzo+17, Vink+18a,18b

Initial conditions (e.g. IMF)

e.g. de Mink & Belczynski+15, Klencki+17

Fraction of Binaries

e.g. Sana+12, Moe & di Stefano+17,

Stellar tracks of Massive Stars

e.g. Laplace+20, Agrawal+20

Supernova kicks

e.g. Belczynski & Bulik+99, Shao & Li+18

Pair instability Supernovae

e.g. Farmer+19, Marchant+19, Stevenson+19

Electron-capture supernovae

e.g. Giacobbo & Mapelli+18

AGN disk density and lifetime

e.g. Mckernan+19

Sampling (Poisson) uncertainties

e.g. Andrews+19, Broekgaarden+19

Initial cluster properties

e.g. Fragione & Kocsis+18

Ultra-stripped supernovae?

e.g. Tauris+15, Mandel & Muller+20

Mass transfer stability criteria

e.g. Soberman+97 , Ge+05, Claeys+14

Depth of convective envelope

e.g. Klencki+20

Over Contact Binaries?

e.g. Marchant+16, du Buisson+20, Riley+20

Expansion of Pop III stars

e.g. Marigo+01, Kinugawa+21

Escape velocities cluster

e.g. Rodriguez+18,19, Gerosa&Fishbach+21

What are key uncertainties in population synthesis?

Floor Broekgaarden

20 of 62

Floor Broekgaarden

Common envelope physics

e.g. Dominik+12, Ivanova+20, Marchant+21

Muhammad & Noam+21

Stellar Winds

e.g. Renzo+17, Vink+18a,18b

Initial conditions (e.g. IMF)

e.g. de Mink & Belczynski+15, Klencki+17

Fraction of Binaries

e.g. Sana+12, Moe & di Stefano+17,

Stellar tracks of Massive Stars

e.g. Laplace+20, Agrawal+20

Supernova kicks

e.g. Belczynski & Bulik+99, Shao & Li+18

Pair instability Supernovae

e.g. Farmer+19, Marchant+19, Stevenson+19

Electron-capture supernovae

e.g. Giacobbo & Mapelli+18

AGN disk density and lifetime

e.g. Mckernan+19

Sampling (Poisson) uncertainties

e.g. Andrews+19, Broekgaarden+19

Initial cluster properties

e.g. Fragione & Kocsis+18

Ultra-stripped supernovae?

e.g. Tauris+15, Mandel & Muller+20

Mass transfer stability criteria

e.g. Soberman+97 , Ge+05, Claeys+14

Depth of convective envelope

e.g. Klencki+20

Over Contact Binaries?

e.g. Marchant+16, du Buisson+20, Riley+20

Expansion of Pop III stars

e.g. Marigo+01, Kinugawa+21

Escape velocities cluster

e.g. Rodriguez+18,19, Gerosa&Fishbach+21

What are key uncertainties in population synthesis?

On top of this many uncertainties in the metallicity-dependent star formation history

Floor Broekgaarden

21 of 62

Floor Broekgaarden

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

Cartoon: xkcd

Floor Broekgaarden

22 of 62

Floor Broekgaarden

Method

Isolated Binaries

Figure from Broekgaarden et al. (2021a)

Broekgaarden et al. (2021a, 2021b)

Floor Broekgaarden

23 of 62

Floor Broekgaarden

1. Simulate Cosmic star formation:

based on models from Neijssel et al., incl. FSB (2019)

28x variations

Method

Isolated Binaries

“Cosmic History”

When & Where

do stars form? (metallicity)

Figure from Broekgaarden et al. (2021a)

Floor Broekgaarden

24 of 62

Floor Broekgaarden

1. Simulate Cosmic star formation:

based on models from Neijssel et al., incl. FSB (2019)

28x variations

Method

Isolated Binaries

“Cosmic History”

When & Where

do stars form? (metallicity)

2. Simulate binaries:

FSB+19 (“STROOPWAFEL”), FSB+21, Team COMPAS; Riley et al., incl. FSB, (2021).

How do stars

live & die?

20x variations

Figure from Broekgaarden et al. (2021a)

Floor Broekgaarden

25 of 62

Floor Broekgaarden

1. Simulate Cosmic star formation:

based on models from Neijssel et al., incl. FSB (2019)

Method

Isolated Binaries

“Cosmic History”

When & Where

do stars form? (metallicity)

2. Simulate binaries:

FSB+19 (“STROOPWAFEL”), FSB+21, Team COMPAS; Riley et al., incl. FSB, (2021).

How do stars

live & die?

20x variations

560 publicly available model realizations

Broekgaarden (2021a, 2021b)

Figure from Broekgaarden et al. (2021a)

28x variations

Floor Broekgaarden

26 of 62

Floor Broekgaarden

Both uncertainties in stellar evolution & cosmic history can impact the BHNS merger rates with factors > 10-100x

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

Broekgaarden & Berger (2021); Broekgaarden et al. (2021b)

Stellar Evolution models

27 of 62

Floor Broekgaarden

Both uncertainties in stellar evolution & cosmic history can impact the BHNS merger rates with factors > 10-100x

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

Broekgaarden & Berger (2021); Broekgaarden et al. (2021b)

Stellar Evolution models

Cosmic models

28 of 62

Floor Broekgaarden

Both uncertainties in stellar evolution & cosmic history can impact the BHNS merger rates with factors > 10-100x

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

Broekgaarden & Berger (2021); Broekgaarden et al. (2021b)

Stellar Evolution models

Cosmic models

29 of 62

Floor Broekgaarden

Both uncertainties in stellar evolution & cosmic history can impact the BHNS merger rates with factors > 10-100x

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

Broekgaarden & Berger (2021); Broekgaarden et al. (2021b)

Stellar Evolution models

Cosmic models

Both stellar evolution & cosmic history uncertainties can impact the BHNS rate >10x

30 of 62

Floor Broekgaarden

BHBH rates dominated by “cosmic history”

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

Broekgaarden & Berger (2021);Broekgaarden et al. (2021b)

Stellar Evolution models

Cosmic models

31 of 62

Floor Broekgaarden

BHBH rates dominated by “cosmic history”

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

Broekgaarden & Berger (2021);Broekgaarden et al. (2021b)

Stellar Evolution models

Cosmic models

Particularly cosmic history uncertainties can impact the BHBH rate >10x

32 of 62

Floor Broekgaarden

NSNS rates dominated by “stellar evolution”

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

Broekgaarden & Berger (2021);Broekgaarden et al. (2021b)

Stellar Evolution models

Cosmic models

33 of 62

Floor Broekgaarden

NSNS rates dominated by “stellar evolution”

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

Broekgaarden & Berger (2021);Broekgaarden et al. (2021b)

Stellar Evolution models

Cosmic models

Particularly stellar evolution uncertainties can impact the NSNS rate >10x

34 of 62

Floor Broekgaarden

What can we learn from the distribution shapes?

Stellar

Evolution

models

  • Stellar evolution changes dominate BHNS distribution shape (compared to cosmic variations)�
  • Observing the BH-NS distributions aids distinguishing between models�
  • BHBH and NSNS distribution shapes are dominated by both

Broekgaarden et al. (2021b)

BH-NS:

Cosmic

models

35 of 62

Floor Broekgaarden

What can we learn from the distribution shapes?

Stellar

Evolution

models

Cosmic

models

  • Stellar evolution changes dominate BHNS distribution shape (compared to cosmic variations)�
  • Observing the BH-NS distributions aids distinguishing between models�
  • BHBH and NSNS distribution shapes are dominated by both

Broekgaarden et al. (2021b)

BH-NS:

stellar evolution uncertainties dominantly impacts the BHNS distribution shape

36 of 62

Floor Broekgaarden

What can we learn from the distribution shapes?

Stellar

Evolution

models

Cosmic

models

  • Stellar evolution changes dominate BHNS distribution shape (compared to cosmic variations)�
  • Observing the BH-NS distributions aids distinguishing between models�
  • BHBH and NSNS distribution shapes are dominated by both

Broekgaarden et al. (2021b)

BH-NS:

Detections

stellar evolution uncertainties dominantly impacts the BHNS distribution shape

37 of 62

Floor Broekgaarden

What can we learn from the distribution shapes?

Stellar

Evolution

models

Cosmic

models

  • Stellar evolution changes dominate BHNS distribution shape (compared to cosmic variations)�
  • Observing the BH-NS distributions aids distinguishing between models�
  • BHBH and NSNS distribution shapes are dominated by both

Broekgaarden et al. (2021b)

BH-NS:

BH-BH:

Both cosmic history & stellar evolution uncertainties significantly impact the BHBH distribution shape

38 of 62

Floor Broekgaarden

Particularly the SNe assumptions affect the BBH mass distributions

Broekgaarden et al. (2021b)

39 of 62

Floor Broekgaarden

Particularly the SNe assumptions affect the BBH mass distributions

Broekgaarden et al. (2021b)

40 of 62

Floor Broekgaarden

Particularly the SNe assumptions affect the BBH mass distributions

Broekgaarden et al. (2021b)

41 of 62

Floor Broekgaarden

Particularly the SNe assumptions affect the BBH mass distributions

Broekgaarden et al. (2021b)

42 of 62

Floor Broekgaarden

Particularly the SNe assumptions affect the BBH mass distributions

Population synthesis urgently requires better supernova remnant models

Broekgaarden et al. (2021b)

43 of 62

Can current population synthesis models “resolve” the remnant mass function with resolution < 2 Msun?

Yes!

No!

Other…

As also discussed in Jessica Lu’s talk (and questions by Chris Belczynski, Saavik Ford)

44 of 62

Can current population synthesis models “resolve” the remnant mass function with resolution < 5 Msun?

Yes!

No!

Other…

As also discussed in Jessica Lu’s talk (and questions by Chris Belczynski, Saavik Ford)

45 of 62

So what now?

Some thoughts

46 of 62

Floor Broekgaarden

Mandel & Broekgaarden (2021) Living Review in Relativity

Ping us if your paper is missing!

Publicly available code/data:

47 of 62

Floor Broekgaarden

Mandel & Broekgaarden (2021) invited Living Review in Relativity

Ping us if your paper is missing!

Publicly available code/data:

Most dynamical channels seem not efficient in making NS-BH

48 of 62

Floor Broekgaarden

Mandel & Broekgaarden (2021) invited Living Review in Relativity

Ping us if your paper is missing!

Publicly available code/data:

Most dynamical channels might not seem efficient in making NS-NS?

49 of 62

Floor Broekgaarden

Matching BHBH, BHNS and NSNS rates

Broekgaarden et al. (2021b)

Floor Broekgaarden

50 of 62

Do all three flavors of GW mergers have the same dominant formation channel?

Yes!

No!

51 of 62

Floor Broekgaarden

What will we learn from looking at redshift dependent rates?

And redshift dependent distribution functions?

Redshift dependent BHBH merger rate for different formation channels. From Ng et al. (2021)

See also van Son et al. (2021, incl FSB)

But….

52 of 62

Floor Broekgaarden

What will we learn from looking at redshift dependent rates?

And redshift dependent distribution functions?

The 3 models used in Ng et al. (2021)

But….

53 of 62

Floor Broekgaarden

What will we learn from looking at redshift dependent rates?

And redshift dependent distribution functions?

The 3 models used in Ng et al. (2021)

But….

+ more isolated models

54 of 62

Floor Broekgaarden

What will we learn from looking at redshift dependent rates?

And redshift dependent distribution functions?

The 3 models used in Ng et al. (2021)

But….

+ more isolated models�+ other models

55 of 62

Floor Broekgaarden

Other observational constraints?

See talks Jessica Lu, Katie Breivik

Detecting BH-NS in our Milky Way

1. Pulsars binaries

e.g. BH-PSR: Debatri Chattopadhyay, Simon Stevenson, Jarrod Hurley, Matthew Bailes & FSB (2021), ArXiv:2011.13503

2. BH/NS binaries in LISA

Thomas Wagg, FSB, Selma de Mink et al. (2021)

56 of 62

Floor Broekgaarden

We need to improve many

things in our models…

Floor Broekgaarden

Population Synthesis

BSE code that

Hurley’s SSE/BSE code from 2000/2002

e.g. Laplace+20, Agrawal+20,21, Klencki+20

Original: xkcd

See, for example, the Aspen talk about POSYDON by Jeff Andrews

57 of 62

BH-BH

NS-NS

BH-NS

We are here! (GWTC-3)

Figure: Baibhav+19

 

Cosmic Explorer Einstein Telescope

3G

2G

We are at the tipping point of entering

“The Big Data Era for NS/BH mergers:”

58 of 62

Floor Broekgaarden

What I work on lately / my message of tonight:

Gravitational waves will teach us everything!!!!

model uncertainties

(the hope of) Population Synthesis modellers

Not so fast..

See also:

Chruslinska et al. (2019)

Belczynski et al. (2021),

the discussion in Zevin et al. (2021),

)

Going forward I believe it will be vital to understand and quantify the model uncertainties in population synthesis (& observations)�e.g. Broekgaarden et al. (2021a, 2021b)

59 of 62

How are you feeling today?

We are

doomed

More

pessimistic

Confused

More

optimistic

other

60 of 62

Questions

Discussion Questions

  • What have we learned from GW rates?
  • Can we constrain formation channels?
  • What key uncertainties should we spend (computational) time on fixing first?
  • What can we learn from mass/spin population distributions?
  • How do we improve our SN remnant mass function?
  • What can we learn from EM counterparts/observations?
  • Do all GWs have the same dominant formation channel?
  • How do we go forward?

61 of 62

What’s next?:

Flipped: 15 min: ”Break out Discussion rooms” (we)

Find the number on your

yellow post-it note

1

Go to the place for your number and find your fellow group members

In the coming 15 min:� - Introduce yourself quickly

- Discuss the questions (or other related questions)� - Assign 1 person* who report back from the group / pose a question on behalf of the group.

*aim for the person who so far has spoken the fewest

62 of 62

Final Poll: write down your answers & I’ll collect them

1. The dominant formation channel(s) of LIGOs (z<1) detections are:

BH-BH: ………………………………………………………………..…………………..…….…………………….

BH-NS: …………………………………………………………………………………………………………………..

NS-NS: ……………………………………………………………………………………………………………………

2. Once we have 1000 GW detections we will

…………………….…………………..…….…………………….…………………….…………………..…….…………………….

3. Once we have 1000 GW detections we will not

�…………………….…………………..…….…………………….…………………….…………………..…….…………………….

Name (optional): …………………………………………………………..