s'g'ryanherts'ac'uk - PowerPoint PPT Presentation

Title: s'g'ryanherts'ac'uk


1
s.g.ryan_at_herts.ac.uk
Lithium abundances and isotope ratios in halo
dwarfsSean G. Ryan School of Physics,
Astronomy and MathematicsUniversity of
Hertfordshire
Principal collaborators Ana García Pérez
(UH), Adam Hosford (UH), John Norris
(ANU) Structure of this talk Lithium problem
(7Li) Lithium isotope ratio (6Li/7Li) Mg/Fe
ratio good news and bad news for chemical tagging
NGC 4414Hubble Heritage Team (AURA/STScI/NASA)
Hack/Ryan (OU)
2
Lithium problem

• Measurements of CMBR by WMAP give baryon density
  fraction ?Bh2 0.02240.0009 (Spergel et al.
  2003).
• BBN depends on ?B.WMAP in excellent
  agreementwith ?B derived from 2H/1H.
• Uncomfortable discrepancy for 7Lithe Lithium
  problem

Coc Vangioni (2005)
3
Lithium problem

• Large spread in Pop.I 7Li reflects fragility in
  stars
• 7Li p ? 4He 4He at T 2.6106 K (at stellar
  densities)
• Hence 7Li survives only in outermost region of
  stars with shallow convective zones.
• Destroyed in stars with deep convective zones
• Disk-metallicity stars high
  metallicity deep SCZ
• Cool main-sequence stars low
  temperature deep SCZ
• First ascent giants Hayashi boundary
  convective
• Survives in warm (6000 K) dwarfs only
• Also removed due to diffusion (sinking out of
  atmosphere)

1.0 ppb
0.1 ppb
Ryan et al. 2000, ApJ, 549, 55 (RKBSRMR)
4
Lithium problem

• 7Li also produced in Galactic sources late in
  evolution
• AGB stars during hot bottom burning, and RGB
  stars if deep mixing, by Cameron Fowler
  mechanism
• 4He 3He ? 7Be ? ? mixed to surface 7Be
  e ? 7Li ?e (t1/253d)
• Novae ?? SN ??
• GCE models dont reproduce steep synthesis of 7Li
  at Fe/H gt -0.5

1.0 ppb
0.1 ppb
Ryan et al. 2000, ApJ, 549, 55 (RKBSRMR)
5
Lithium problem

• Several explanations offered to explain
  discrepancy
• Failure of SBBN modelparticle physics
  possibilities
• survival of metastable particlesfor a few 103
  s, i.e. during BBNBird, Koopmans Pospelov
  2007,hep-ph/0703096 X- 7Be ? 7BeX- 7BeX-
  (p,?) 8BX- ? 8BeX- ß ?e Pospelov, M.
  2007, hep-ph/0712.0647 X- 4He ? 4HeX-
  4He ? 8BeX- 8BeX- n ? 9BeX- ? 9Be X-
• decay or annihilation of massive supersymmetric
  particle,modifying 7Li and 6Li production
  Jedamzik 2004

6
Lithium problem

• Several explanations offered to explain
  discrepancy
• Stars have destroyed some 7Li
• Inhibited diffusion?
• More mundane explanationsdid we get the
  abundances wrong?
• Largest uncertainty is effective-temperature
  scale for metal-poor stars
• E.g. comparison betweenRyan et al (2001) and
  hot Melendez Ramirez (2005) Teff scales
  shows difference of up to 400K for Fe/H lt -3
• ?Teff 400 K ? ?A(Li) 0.3 dex close to
  discrepancy

7
Lithium problem

• PhD Adam Hosfordeffective temperaturescale for
  metal-poor stars
• Fe I lines
• Use T-dependence of level populationsBoltzmann
  factor exp-(?/kT)
• 1D,LTE analysisHosford, Ryan, Garcia Perez,
  Norris Olive 2009, AA, 493, 601
• Attention to error propagation

8
Lithium problem

• Attention to error propagation
• Uncertainty in ? vs A(Fe) slope being nulled
• 60-80 K
• Assumptions/constraints regarding evolutionary
  state (weakly constrained for distant weak-lined
  stars)
• affect adopted isochrones and hence model
  atmosphere
• 12-24 K
• Uncertainty in inferred ?
• affects slope being nulled
• wrong physics anyway need 3D
• 30-90 K

9
Lithium problem

• Hosford T-dependence of level populations
• Asplund05Ha Balmer profile fits
• Melendez RamirezIRFM
• Teff similar to R01, Asplund05

T(Ryan)
T(Asplund)
T(Hosford)
T(Hosford)
200 K
T(MR)
T(Hosford)
10
Lithium problem

• cf. Asplund et al. (2005) analysis
• Asplund et al. (2005) temperatures in good
  agreement with b-y and V-K IR flux method Teff of
  Nissen et al. (2002,2004) ?Teff -34 95 K
• not consistent with hot Melendez Ramirez
  (2004) Teff scale (?Teff 182 72 K _at_ Fe/H lt
  -2.6).
• Abundance lower than Melendez Ramirez also due
  to steeper (T,t) relation in K93 models used by
  MR
• Log e(7Li) (2.409 0.020) (0.103
  0.010)Fe/H
• Primordial value 0.5 dex below predicted value
  assuming WMAP OB.

11
Lithium problem

• Current T? analysis assumes LTE ? level
  populations
• LTE holds at tcontinuum gt 1, but lines form at
  tcontinuum lt 1
• NLTE difficult to calculate reliably
• Collisional excitation very uncertain
• Collisions with hydrogen parametrized via SH (
  0.001? 1?)
• Information not available for some levels of the
  Fe atom
• Ideally calculate
• populations
• radiative and collisional transition rates (need
  all oscillator strengths)
• for all levels (population of a level affected
  by populations of others through radiative and
  collisional transitions)
• NIST lists 493 levels for Fe I, 578 levels for Fe
  IIcf. our model atom contains just 524 levels in
  total (Fe I, II and III)
• Need photoionisation rates from Fe I levels to Fe
  II levels
• Usually just to ground state of Fe II

12
Lithium problem

• Many previous calculationse.g. Asplund et al.
  (1999, AA, 346, L17) point to biggest effect at
  low Z being overionisation relative to LTE, which
  underpopulates all levels
• transparent layers at tcontinuum lt 1 see photons
  from deep/hot atmosphere, so photon intensity J?
  gt local B?
• In UV, excess flux sufficiently energetic to
  photoionise excited Fe I states (e.g. Asplund
  2005 ARAA, 43, 481, 3.7) underpopulates all
  energy levels

13
Lithium problem

• Additional factor lack of collisions at
  tcontinuum lt 1
• reduces collisional excitation of excited levels
  compared to what local T suggests via
  Boltzmann(populations not in thermal equilibrium
  with local temperature)
• hence excited level populations lower than in LTE
• Faint hope that overionisation dominates over
  differential excitation ???
• Assess using MULTI calculations ...

14
Lithium problem

• Net effect of NLTE on level populations
• b nNLTE/nLTE
• (SH 1)
• NLTE effects clearly depend on ?
• Roles of collisional ionisation and
  photoionisationdepend on how close level is to
  continuum
• Demonstrates that slope of ? vs A(Fe) plot
  affected by NLTE
• Hence T? affected by NLTE

15
Lithium problem

• Transparency of atmosphere (esp. in metal-poor
  stars)
• photon field characteristic of deeper
  layers/higher temperatures than local
  temperature(source function not characterised by
  local temperature)
• S/B ranges from lt1 to gt1

16
Lithium problem

• Work in progress
• Our preliminary calculations suggest T 90-110 K
  hotter than in LTE
• To be revised shortly, following exploration of
  sensitivities in the NLTE modelling
• ?A(Li) lt 0.1 dex still wont solve Lithium
  problem

17
Lithium problem

• Efforts (uncompelling) to reconcile 7Li with
  WMAP
• Systematic errors in abundance analysis not big
  enough?
• Dilution (diffusion) or destructionclassical
  diffusing or destroying this much 7Li implies
  huge initial abundances of 6Li creates a bigger
  problemsolution may be turbulent diffusion
  (Korn et al. 2006)though model parameters not
  known a priori.
• Errors in nuclear reaction ratesnot big enough
• Non-standard BBN, e.g. supersymmetric particles
  speculative and interesting!
• Non-standard stellar evolutiontoo many
  unconstrained free parameters

18
Lithium-6 problem

• 6Li isotope shift 0.15 Å same as fine
  structure splitting
• Adds a little asymmetry to an already asymmetric
  line
• Turbulence/convection also adds asymmetry hard
  to model?
• 6Li not produced significantly in standard bbn
• lt 0.00001 ppb Serpico et al. 2004
• 6Li not produced in stars
• H-burning p-p chain produces 4He
• He-burning 3a produces 12C
• No stable nuclei of A 5 or 8, so light-element
  nucleosynthesis of A 6 and 7 frustrated
• 4He p does not proceed
• 4He 4He ? 8Be either a-decays or a-captures to
  12C.

19
Lithium-6 problem

• 6Li produced alongside 9Be and 10,11B in ISMvia
  galactic cosmic ray (GCR) spallation
• p,4HeISM 12C,14N,16OGCR ? X Y primary
• 12C,14N,16OISM p,aGCR ? X Y secondary
• 4HeISM aGCR ? 6,7Li Y primary Li
  at low Z Steigman Walker 1992, ApJ, 385,
  L13 Yoshii et al. 1997, ApJ, 485, 605 (YKR)
• Boesgaard et al. 1999, AJ, 117, 1549 (BDKRVB)
  Duncan et al. 1997, ApJ, 488, 338
  (DPRBDHKR)

boron
beryllium
20
Lithium-6 problem

• 6Li and 7Li destroyed in stars
• Reaction rate governed by S-factor
• S(0) 3140 keV barns for 6Li(p,3He)4He Elwyn et
  al. 79, PhysRevC, 20, 1984 (EHDMMR)
• S(0) 55 keV barns for 7Li(p,4He)4He Pizzone
  etal. 03, AA, 398, 423 (PSLCMPRTCDI)
• 6Li p ? 4He 3He at 2.0106 Kcf. 7Li
  p ? 4He 4He at 2.6106 K for 7Li
• 6Li survives only in hottest metal-poor starsdue
  to extreme sensitivity of 6Li survival to
  deepening convection in stars of lower Teff
  Brown Schramm 88, ApJ, 329,
  L103

21
Lithium-6 problem

• Two major results from Asplund et al. 2006
• Abundance high compared to models that are
  consistent with spallative 9Be, 10,11B,
  especially if depletion allowed for.
• Trend with Fe/H looks like plateau, unlike
  strong Fe/H dependence of models.

22
Lithium-6 problem

• Aoki et al. 2004, AA, 428, 579 (AIKRSST)
• S/N 1000 R 90000 6Li/7Li
  0.00, 0.04, 0.08

23
Lithium-6 problem AEGP

• Subaru/HRS data on 6 stars analysed
  by Ana García Pérez
• Isotope ratio VERY sensitive to
  systematic uncertainties e.g.
  macroturbulent width, wavelength
  shifts, continuum errors, flat field
  errors, 7Li abundance
• fair choices ? ?(6Li/7Li) 3-4
  

24
Lithium problem AEGP
Subaru data very similar
to Asplund et al. VLT/UVES data
... But we are not
confident of our detections
Working at margins of significance due
to systematic limitations
VLT observations
4 3 2 1
Asplund et al. 2006
25
Lithium problem

• Optimists view
• in metal-poor stars, 7Li direct (?) from the big
  bang
• uncrowded spectrum, high S/N achievable
• Pessimists view
• destroyed in most parts of most stars
• where it survives, most is ionised and
  unobservable
• cant model its recent Galactic formation
• hard to derive abundance?
• strongly dependent on uncertain Teff scale and
  (T,t) scales because of high degree of
  ionisation
• stellar depletion (including diffusion) hard to
  quantify

26
Inhomogeneous GCE
Theoretical basis for origin of yields Diversity
in early ISM depends on how sensitive SN yields
are to progenitor mass. Expect Mg/Fe yield to
depend strongly on progenitor mass Argast et al.
2001, AA, 388, 842 --- using yields of
Thielemann et al.
but observations show almost no spread?obs
0.06 dex but ?err 0.06 so ?intrinsic lt 0.04
dex Arnone et al. 2005, AA, 430, 507 (ARANB)
27
Inhomogeneous GCE
Theoretical basis for origin of yields Diversity
in early ISM depends on how sensitive SN yields
are to progenitor mass. Expect Mg/Fe yield to
depend strongly on progenitor mass Argast et al.
2001, AA, 388, 842 --- using yields of
Thielemann et al.
but new observations show almost no spread?obs
0.06 dex but ?err 0.06 so ?intrinsic lt
0.04 dex Arnone et al. 2005, AA, 430, 507
(ARANB)
Argast et al. model
Cayrel et al. obs
Arnone et al. obs
28
Inhomogeneous GCE
The low Mg/Fe spread is not the full
picture see large Mg/Fe values for CS
22949-037 Mg/Fe 1.2 Fe/H
-3.8McWilliam et al. 1995, AJ, 109, 2757 (MPSS)
Norris, Ryan Beers 2001, ApJ, 561, 1034
Depagne et al. 2002, AA, 390, 187 CS
29498-043 Mg/Fe 1.8 Fe/H -3.5Aoki et
al. 2002, ApJ, 576, L141 (ANRBA) Aoki et al.
2004, ApJ, 608, 971 (ANRBCTA)
Explain with mixing and fallback in
hypernovae?e.g. 30 MSun, 50x1051 erg, Mr 2.44
- 12.6 MSun, f 0.004 Aoki et al. 2004,
ApJ, 608, 971 (ANRBCTA) cf Umeda Nomoto
2004, Nature, 422, 871 (30 MSun, 20x1051
erg, Mr 2.33-8.56 MSun, f 0.002) Also
Umeda Nomoto 2005, ApJ, 619, 427 Many
free parameters?
Does diversity in intermediate-mass elements
appear mostly at Fe/H lt -3.5?
29
Concluding remarks
Lithium problem for 7Li too little seen compared
to SBBN Do stars destroy it? 6Li Some analyses
point to high 6Li abundances compared to models,
and apparent plateau ... ... but our analysis
emphasises extreme sensitivity of inferred
isotope ratio to reasonable choices in analysis,
at level 3-5. Mg/Fe is remarkably uniform in
most (not all) metal-poor stars... Suggests
chemical tagging CAN achieve the accuracy
required, 0.06 dex ...... but is there a
signal to detect in many stars? More data under
analysis ...
30
NGC 4414Hubble Heritage Team (AURA/STScI/NASA)
Hack/Ryan (OU)
31
Inhomogeneous GCE

• Low spread for Mg/Fe requires
• Mg/Fe not dependent on progenitor mass?
• Only very narrow mass range of progenitors _at_
  18 MSun effective?
• Always see Mg/Fe averaged over IMF, despite
  these being amongst the first stars?
• Are cooling timescales longer than mixing
  timescales?

?
?
??
?