Nucleosynthesis across the Galaxy: AGB Stars and Neutron Stars - - PowerPoint PPT Presentation

Nucleosynthesis across the Galaxy: AGB Stars and Neutron Stars Mergers

Diego Vescovi1,2,3, Sergio Cristallo2,3, and Marica Branchesi1,4

• 1. Gran Sasso Science Instjtute (GSSI), L’Aquila, Italy
• 2. INFN – Sectjon of Perugia, Perugia, Italy
• 3. INAF – Osservatorio Astronomico d’Abruzzo, Teramo, Italy
• 4. INFN – Laboratori Nazionali del Gran Sasso, Assergi, Italy

GSSI Admission to the 3rd year GSSI - L’Aquila - Italy, 10 October 2019

The origin of heavy elements in the Solar System

Diego Vescovi - L’Aquila, 2019

Nucleosynthesis across the Galaxy: AGB Stars and NMS

2

Neutron captures processes :

• r-process
• s-process

1)Weak component (A<90) Massive Stars 2)Main component (from Sr to Bi) AGB stars

Locatjon of peaks indicates n-captures along valley of stability s-process

H- and He-burning in TP-AGB stars

Diego Vescovi - L’Aquila, 2019

3

13C-pocket

He-fmash proton penetratjon

22Ne(α,n) 13C(α,n) 12C(p,γ)13N(β+ν)13C

• What?

Low-Mass Stars

• When?

Asymptotjc Giant Branch (AGB)

• How?

Thermally Pulsing (TP) heavy elements!!

Nucleosynthesis across the Galaxy: AGB Stars and NMS

The 13C-pocket: formatjon

Diego Vescovi - L’Aquila, 2019

4

• Protons can penetrate into the He-rich region at

each TDU (Third Dredge-Up) phenomenon Which is the physical mechanism? Classic models assume the 13C-pocket formatjon Many recent physical approaches:

• Opacity induced overshoot (Cristallo+ 2009, 2011, 2015)
• Convectjve Boundary Mixing (Battjno+ 2016)
• Magnetjc fjelds (Trippella+ 2016; Palmerini+ 2018)

botuom-up mechanism through magnetjc buoyancy 1a) Rotatjonal shears promote magnetjc fjelds? 1b) Fossil magnetjc fjelds? 2) Magnetjc structures reach the envelope 3) Protons are injested into the He-rich region

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Magnetjc buoyancy

Diego Vescovi - L’Aquila, 2019

5

• MagnetoHydroDynamics (MHD) solutjons (Nucci & Busso 2014):

➔ No numerical approximatjons (exact analytjc solutjon) ➔ Simple geometry: toroidal magnetjc fjeld

where k is the exponent of the density distributjon: Equatjons: Solutjons:

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Implementatjon

Diego Vescovi - L’Aquila, 2019

6

• Exponentjal decay of the convectjve velocity

(Straniero+ 2006, Cristallo+ 2009):

• Magnetjc contributjon (this work),

actjng when the density distributjon is ρ ∝ r k : Parameters:

➔ Radius extentjon of the overshootjng region ➔ β

Parameters:

➔ Layer “p” at the deepest coordinate from which

buoyancy starts (can be identjfjed from the corresponding critjcal toroidal Bφ value)

➔ Startjng velocity vp of the buoyant material

Calibratjon is needed!

Nucleosynthesis across the Galaxy: AGB Stars and NMS

SiC Grains

Diego Vescovi - L’Aquila, 2019

7

• Stellar models with difgerent initjal mass and metallicity

➔ difgerent numbers of thermal pulses experienced ➔ difgerent extentjon of 13C-pockets ➔ Isotopic ratjos of mainstream grains are quite well reproduced

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Intrinsic C-rich AGB Stars

Diego Vescovi - L’Aquila, 2019

8

• Stellar models with close-to-solar metallicity

➔ Low [hs/ls] ➔ High [s/Fe]

• Does magnetjsm fade out for low-to-intermediate mass (3 to 6 M⊙)?

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Post- and Intrinsic C-rich AGB Stars I

Diego Vescovi - L’Aquila, 2019

9

• Stellar models with low metallicity

➔ [hs/ls] vs. [s/Fe] consistent with observatjons ➔ Models with opacity-induced overshoot only fail

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Post- and Intrinsic C-rich AGB Stars II

Diego Vescovi - L’Aquila, 2019

10

• Stellar models at difgerent

metallicitjes

➔ [hs/ls] vs. [Fe/H] consistent with

• bservatjons

➔ Models with opacity-induced

• vershoot only fail again

Strong magnetjsm Weak magnetjsm

➔ Variable effjciency of the

MHD-induced mixing?

➔ Mass-dependent effjciency?

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Summary I

Diego Vescovi - L’Aquila, 2019

11

➔ Most of what we know has been learned through a lengthy work with parameterized

models, trying to constrain the parameters gradually, from the increasing accuracy of

• bservatjons

➔ This allowed recently the development of physical models for the mixing

mechanisms required to produce the 13C neutron source.

➔ Taking into account magnetjc fjelds in radiatjve regions might be crucial in modeling

the mixing episodes (e.g. through magnetjc buoyancy).

➔ First outcomes confjrms recent results from Trippella+ (2016), Palmerini+ (2018),

and Liu+ (2018, 2019)

➔ More extended and fmatuer 13C-pocket ➔ The majority of isotopic ratjos of mainstream grains are quite well reproduced ➔ [hs/ls] vs. [s/Fe] and [hs/ls] vs. [Fe/H] consistent with observatjons of post-AGB and

intrinsic AGB stars

➔ Magnetjsm has (most problably) variable intensity

Nucleosynthesis across the Galaxy: AGB Stars and NMS

r-process: basic ideas

Diego Vescovi - L’Aquila, 2019

12

• key reactjons: (A, Z) + n ↔ (A + 1, Z) + γ
• r-process requires initjal high nn and T

➔ high nn : τ(n,γ) << τβ-decay ➔ high nn and T: (n, γ) ↔(γ, n) along isotopic chain ➔ steady abundances intra-chain with one dominant nucleus

• β-decay rates of dominant nuclei regulate inter-chain fmow
• equilibrium freeze-out: nn drops and β-decays take over

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Neutron star mergers as r-process site

Diego Vescovi - L’Aquila, 2019

13

Matteucci +14

• r-process requires free n and seed nuclei (<A>, <Z>)
• seed propertjes/abundances depend on nuclear-statjstjcal equilibrium (NSE)

freeze-out

• in adiabatjc expansion, neutron-to-seed ratjo depends on three parameters:

1) entropy s ~ T 3/ρ 2) Ye ~ np /(nn+np) 3) τdyn (T (t) ≈ T0 exp(−t/τdyn)) nn /nseed ∝ s 3 / (τdynYe

3)

Possible scenarios high entropy r-process low entropy r-process hot CCSN winds BNS and BHNS mergers MHD supernovae First evidences of r-process nucleosynthesis in kilonova from GW170817

Nucleosynthesis across the Galaxy: AGB Stars and NMS

BNS merger + kilonova

Diego Vescovi - L’Aquila, 2019

14

Basic ideas:

• radioactjve decay of freshly

sinthetjzed r-process elements in ejecta: release of nuclear energy

• thermalizatjon of high energy

decay products with ejecta

• difgusion of thermal photons

during ejecta expansion

• thermal emission of photons at

photosphere

Metzger & Berger 12 Nucleosynthesis across the Galaxy: AGB Stars and NMS

Propertjes of GW170817/AT2017gfo

Diego Vescovi - L’Aquila, 2019

15

Nucleosynthesis across the Galaxy: AGB Stars and NMS

• 17/08/17, GW+EM detectjon of an event compatjble with BNS merger (LVC PRL 2017)
• rather bright, nIR component, with a peak at

5 days (red component) ∼

• bright, UV/O component, with a peak at

1 day (blue component) ∼

Light curves; Pian, D’Avanzo+ 2017 (left); Tanvir+ 2017 (right)

➔ Kilonova models fail in explaining the early behavior of the UV and visible light curves ➔ The presence of a larger nuclear heatjng rate at t

1 day ≲ can increase the light curves by half a magnitude during the fjrst day

Heatjng rate vs. electron fractjon Ye

Diego Vescovi - L’Aquila, 2019

16

˙ Q ˙ Qfit(t)=1010t d

−1.3 erg g−1s−1

➔ is usually approximated by an analytjc fjttjng formula as ➔ Detailed nucleosynthesis calculatjons show a complex dependence ➔ Heatjng rates normalized to point out that all the normalized heatjng rates

show considerable excess at difgerent tjmes ˙ Qfit

Nucleosynthesis across the Galaxy: AGB Stars and NMS

➔ Inclusion of new detailed nuclear heatjng rates obtained by nuclear network

calculatjons in an anisotropic, multjcomponent kilonova model (Perego+ 2017)

➔ Coupled with a parallelized Monte Carlo Markov Chain (MCMC) algorithm. ➔ Goal: re-analize AT2017gfo data by computjng the posterior distributjons

associated to several difgerent models

Implementatjon and fjrst tests

Diego Vescovi - L’Aquila, 2019

17

• First outcome (simple isotropic

dynamical ejecta) :

➔ brighter lightcurve

Next steps: 1) different matter ejection mechanisms (multi-component) 2) angular dependence (anisotropy)

Nucleosynthesis across the Galaxy: AGB Stars and NMS

[days] [erg/s]

Summary II

➔ Kilonova from GW170817 originates from the radioactjve decay of heavy elements ➔ Signature of r-process nucleosynthesis in ejecta from neutron star mergers ➔ Astrophysical site of the r-process is identjfjed, but further observatjons are

necessary

➔ Having identjfjed the astrophysical site it becomes fundamental to reduce the

nuclear physics uncertaintjes

➔ Lanthanide-rich for Ye

0.25 ≲

➔ Insensitjvity of the abundance patuern to the parameters of the merging system

because of an extremely Ye environment, which guarantees the occurrence of several fjssion cycles before the r-process freezes out

➔ Nuclear heatjng rates are, at the tjmes relevant for the kilonova emission, uncertain

for a factor a few

➔ Kilonova emission seems to be strongly afgected by non-approximated heatjng rates

Diego Vescovi – L’Aquila, 2019

18

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Future work

1) Artjcle on s-process nucleosynthesis from magnetjc AGB stars, computjng low-mass stars (1.5-2 M⊙) at difgerent metallicitjes. Submission by January/February 2)Analyzing the magnetjc contributjon to the formatjon of the 13C neutron source in in low-to-intermediate mass (3-6 M⊙) AGB stars 3)Extend the nuclear network of the open-source Skynet code in order to include the latest available fjssion rates for r-process calculatjons 4)Perform new 2- and 3-component kilonova model of AT2017gfo, also considering an angular dependence due to anisoptropy of ejecta. Expected drafu in the next few months 5)?? Implement a simplifjed gray radiatjve transport scheme (instead of a revised Arnetu’s model) in order to compute the lightcurve

Diego Vescovi – L’Aquila, 2019

19

Nucleosynthesis across the Galaxy: AGB Stars and NMS

Backup Slides

SiC Grains II

• We considered isotopic data including Sr and Ba isotope ratjos in presolar SiC

grains.

• We considered magnetjc contributjon to the partjal mixing of hydrogen.
• One stellar model: 2M⊙ Z=Z⊙
• Fixed value of β (0.1) and maximum envelope penetratjon (1.7 Hp)
• Variable vp (2, 4, 6 x10−5 cm s−1) and Bφ (0.5, 1, 2 x105 G)

Minimum δ88Sr

The 13C-pocket: parametric space

Parameter Adopted value References

• r motjvatjon

vp 2x10-5 cm/s Best fjt to the grains data β 0.1 Cristallo+ 2009 Radius extentjon of the overshootjng region 1.7 Hp Same amount of H-depleted dredged-up material of FRUITY Layer from which buoyancy starts (critjcal toroidal Bφ value) 2x105 G Best fjt to the grains data

• Our current best (not yet defjnitjve) choice can be summarized as:

Magnetjc fjeld in O-rich and C-rich AGB stars

• Generally, AGB magnetjc

fjeld measurements come from maser polarizatjon

• bservatjons (SiO, H2O and

OH) (e.g. Vlemmings+ 2012)

• These have revealed a

strong magnetjc fjeld throughout the circumstellar envelope

• B-fjeld at surface

few G ∼

• Although the maser
• bservatjons trace only
• xygen-rich AGB stars,

recent CN Zeeman splittjng

• bservatjons (Duthu+ 2017)

indicate that similar strength fjelds are found around C- rich stars

Duthu+ (2017)

Critjcal toroidal B-fjeld

• Stellar model: 2M⊙ Z=Z⊙
• The critjcal Bφ necessary for

the onset of magnetjc buoyancy instabilitjes, in radiatjve zone below the convectjve envelope varies from 10 ∼

4G to

10 ∼

6G

• Difgerent values of Bφ

correspond to difgerent values

• f the free parameter rp

➔ The strength of Bφ determines

the extension of the mixed zone and, in turn, of the 13C- pocket

Generatjon of a toroidal B-fjeld in the He-intershell

• Stellar model: 2.5M⊙ Z=Z⊙
• Stretching of a preexistjng

poloidal fjeld can generate a toroidal fjeld

• Difgerentjal rotatjon in the He-

intershell?

• An additjonal artjfjcial viscosity of

around 107cm2s−1 provides a suffjcient transport of angular momentum to match the core and envelope rotatjon rates for core He-burning stars (den Hartogh +

2019a,b)

• The critjcal polidal Bp would be
• A rough (preliminary) estjmate

gives a Bp few hundreds tjmes lower than Bφ Bp ≲ 1kG Not implausible!!

Efgectjve 13C*

• One stellar model: 2M⊙ Z=Z⊙
• Same sequence TP-interpulse
• Variable vp (2, 4, 6 x10−5 cm s−1) and Bφ (0.5,

1, 2 x105 G)

➔ The amount of efgectjve 13C is strongly

afgected by the adopted parameters

➔ The greater the initjal velocity of fmux

tubes and the deeper the buoyancy starts, the greater the velocity of the downfmow material is

➔ Larger values of Bφ correspond not only to

larger 13C-pockets but also to larger amounts of 13C

*The mass fractjon of efgectjve 13C in a given mesh point is X(13Cefg) = X(13C) − 13/14*X(14N)

The 13C-pocket: shape

• Isotopic abundances during the 2nd TDU: 1H, 13C, 14N
• Stellar model: 2M⊙ Z=Z⊙

➔ An effjcient s-process occurs when 13C overcomes 14N ➔ Outcome: more extended and a fmatuer 13C-pocket.

MHD equatjons

Magnetjc contributjon to the downward velocity

The velocity of the downward material is proportjonal to vprp

(-k+2) (with k ≥ −1)

Final abundances vs. electron fractjon Ye

➔ Threshold value Ye,crit ≈ 0.25

Ye < Ye,crit Ye > Ye,cri “robust” r-process A 130 ≳ “weak” r-process A 130 ≲ insensitjve to details of trajectory sensitjve to details of trajectory

Role of fjssion for the robust r-process

• Simulatjon starts at NSE

Role of fjssion for the robust r-process

• 1st peak is populated

• 2nd peak is populated

Role of fjssion for the robust r-process

Role of fjssion for the robust r-process

• Fissile nuclei are produced

Role of fjssion for the robust r-process

• The fjssion of heavy nuclei leads to the creatjon of nuclei around the 2nd peak.
• Fission products contjnue to capture neutrons, leading to efgectjve fjssion cycling
• An extremely neutron-rich environment guarantees the occurrence of several fjssion

cycles before the r-process freezes out.

Nuclear physics quantjtjes for modelling the r-process

1) Nuclear mass model 2) β-decay rates 3) Fission fragment distributjon models

• Late neutron captures determine the

position of the third r-process peak.

• Fission fragments distribution

shapes the region around the second r-process peak.

Eichler+ 15 Eichler+ 15

Afuer the freeze-out the release

• f neutrons from fjssion dominates
• ver β-delayed neutrons

Fission barriers and density levels above barrier

Independently of the channel, neutron- induced fjssion cross sectjons provide important data (fjssion barriers; level densitjes above barriers; etc.), which are needed to optjmize (or validate) fjssion models for r-process nucleosynthesis. Moreover, if the energy of the captured neutron is high enough to re-emit neutrons (1 or more) AND actjvate the fjssion process, multjple chance fjssion may occur. In this case, the study of multjple chance fjssion on more isotope

• f the same element allows to refjne

fjssion models.

r-process nucleosynthesis

in low entropy environment (s a few tens of ∼ kb /baryon)

➔ Ye dominant parameter

• Ye < 0.15: robust r-process, due to several fjssion cycles
• Ye

0.25: 2nd and 3rd ≲ r-process peaks, but no fjrst

• Ye ≳ 0.25: up to 2nd r-process peak

Productjon of lanthanides dramatjcally changes photon opacity (κγ ), because of electrons fjlling f-shell in ionized states

• no lanthanides:

low opacity (κγ ≲ 1 cm2/g)

• presence of lanthanides: increased
• pacity (κγ

≳ 10 cm2/g)

• key physics ingredients:

1)ejecta mass, velocity, Ye astrophysics 2)opacity κγ atomic physics 3)radioactjve heatjng rate nuclear physics

Nuclear heatjng rate

• Radioactjve decays of r-process

elements release nuclear energy ˙ Qr−process= ∑

i∈reactions

Qiλi luminosity time days matter becomes more transparent matter transparent radioactive decay: L∝ ˙ Q M time 1 s heating rate ˙ Q ∝t

−1.3

with Q=Minitial−Mfinal and λ=decay rate Ye ≳ 0.25 Ye 0.25 ≲ weak r-process (A<130) robust r-process (A>130) “blue transients” peaking afuer ~ 1 day “red transients” peaking afuer ~ 1 week ˙ Q ˙ Q

Nuclear heatjng rate – Uncertantjes I

How much variatjon can we expect from nuclear physics?

➔ Nucleosynthesis occurs near “neutron dripline” ➔ no experimental informatjon ➔ rely on theoretjcal nuclear mass models

Nuclear heatjng rate – Uncertantjes II

Comparing two frequently used nuclear mass models (Rosswog+ 2017): 1)“Finite Range Droplet Model” (FRDM; Möller+ 1995 ) 2)“Dufmo Zuker Model” (DZ31; Dufmo, Zuker 1995 ) Trans-lead region is most relevant for heatjng (Barnes+ 2016):

➔ α-decays ➔ thermalizatjon effjciency ➔ at relevant tjmes difgerence of factor ~5

Rosswog+ 17 Rosswog+ 17

Abundances Heatjng rate

MCMC – Isotropic dynamical ejecta