V652 Her: evidence that white dwarf mergers really do happen

1
V652 Her evidence that white dwarf mergers
really do happen

• Simon Jeffery1, Hideyuki Saio2, Vincent Woolf1
  and Pilar Montañés Rodríguez1.
• 1Armagh Observatory, 2Tohoku University

The British Council, the Japan Society for the
Promotion of Science and the NI Dept of Culture
Arts and Leisure D.Pollacco, QUB, R.Starling,
MSSL, P.W.Hill, St Andrews Universitybased on
work published in MNRAS 313, 671, MNRAS 321,111
and accepted by MNRAS and AA
2
Outline

• Extreme Helium Stars
• V652 Her the unique pulsating helium star
• Previous evolution models for V652 Her
• The HeHe merged binary white dwarf model
• V652 Her as an HeHe merger
• V652 Her pulsation models
• V652 Her as a future subdwarf B star
• Previous evolution models for other EHes
• The COHe merged binary white dwarf model
• EHes (and RCrBs) as COHe mergers
• Conclusions

3
Extreme Helium Stars
A and B supergiants low mass dimensions
post-AGB stars no planetary nebulae no
binaries weak or absent hydrogen lines strong
carbon lines (most) rare (3 in HD catalogue)

• what are they?
• where did they come from?
• where are they going?

4
V652 Her the pulsating helium star

• Extreme helium star (Berger Greenstein 1953)
• comparable to other helium stars HD124448,
  HD168476, BD10 2179
• Light variations (Landolt 1973)
• P0.108 days similar to ? Cepheids, but obviously
  not Pop I MS
• Radial velocity variations (Hill et al. 1981, )
• amplitude 70 km/s, rapid acceleration
  free-fall
• Radius measurement from Baades method
  (Lynas-Gray et al. 1984)
• M g.R2 0.7 / 0.3 Msun, L 103 Lsun

• Period change (Kilkenny Lynas-Gray 1982 - 1996)
  
• corresponds to R/R0 -2 .10-4 yr-1 but
  not-linear (R, R)
• Nonadiabatic linear pulsation models (Saio 1983 -
  1995)
• 1993 OPAL and OP opacities gt Z-bump opacity
  driving
• Nonlinear models (Fadeyev Lynas-Gray 1996)
• Best agreement for M0.72 MSun, Teff23 500K, L
  1062 LSun, Z0.0156
• Stellar atmosphere analysis (Jeffery et al. 1999)
• 1 H, Fe/H0, N-rich, C and O poor, log g 3.7
  / 0.1, Teff 24 500 / 500 K

5
spectrum of V652 Her
MgII
H?
HeI
HeI
WHT 1998 spectrum (histogram) plus model
(curve) Jeffery, Woolf Pollacco 2001
6
light and radial velocity curves
P 0.108 days
SAAO 1.0m StAPCCD V-band data from Kilkenny
(priv. comm.)
WHTISIS 1998 Jeffery, Woolf Pollacco 2001
?
7
Radial velocity curve

• cross-correlation
• template chosen close to maximum radius
• random errors lt 1 km/s
• calibration errors lt0.2 km/s
• zero-point fixed by synthetic spectrum
• ccf width reflects mean line width
• velocity smearing
• gravity broadening
• turbulence/shocks.

8
Spectral fitting

• Automatic fitting FFIT
• interpolation in grid of fluxes
• ?2 minimization AMOEBA
• single and binary stars
• Teff, ?, EB-V
• Automatic fitting SFIT
• interpolation in grid of spectra
• ?2 minimization AMOEBA
• single and binary stars
• Teff, log g, vrot sin I, ni
• Automatic fitting SFIT_SYNTH
• uses grid of models
• ?2 minimization AMOEBA
• real-time radiative transfer calculation
• single stars
• vturb, ni,i3,.

• Model atmospheres STERNE
• LTE, HSE
• line-blanketed (odf dfbhe90)
• produces models and fluxes
• Linelists LTE_LINES
• Jeffery 1991
• Spectrum synthesis SPECTRUM
• LTE
• H Vidal et al. 1973, Lemke 1997
• HeIBarnard et al, 1969,74,75, Gieske Griem
  1969, Dimitrijevic et al. 1984,90
• HeII Schöning Butler 1989
• Z Voigt profiles
• produces HIRES normalized spectra, total fluxes
  (line and continuum) and specific intensities
  (line and continuum)

http//www.arm.ac.uk/csj/atmospheres.html http//
www.arm.ac.uk/csj/software_store.html
9
Automated analysis
composition
High-resolution spectrograms
UV and visual spectrophotometry
STERNE
model structure grid
model flux grid
vt
FFIT
SPECTRUM
Teff, E(B-V), q
Teff, EB-V, ?
high-resolution model grid
SFIT
Atomic Data
Teff, g, ni, v sin i
SFIT_SYNTH
LTE_LINES
vt, composition
10
Fitted spectrum
11
Results of fits
Teff effective temperature. geff gravity felt
by atmosphere, increased by outward
acceleration geff g d2r/dt2 v sini ?
allowance for additional line broadening formal
(and hence differential) errors are very
small systematic errors will be much larger
12
Radius measurement 1
Measure geff and dr/dt from spectrum Integrate
dr/dt to obtain ?r Differentiate to obtain
d2r/dt2 Hence evaluate g geff - d2r/dt2 By
identity, if g0 gravity at some radius
r0 ?(g/g0) 1 ? (rr0)/r0 ? ?r/r0 Plot against
?r, then gradient is r0 ? ltrgt0.990.02R? Mass M
g0r02/G 0.11M ? Problems caused by poor
measure of geff at minimum radius -
non-equilibrium effects and overall lack of
sensitivity to geff. The method could be useful
in some circumstances.
13
Radius measurement 2
Measure Teff and dr/dt from spectrum V light
curve from Kilkenny Lynas-Gray (1982). Use
Teff, V magnitude and model atmospheres to
establish angular diameter, ?. Integrate dr/dt
to obtain ?r By identity, ??/?0 ? ?r/r0 Plot
??/?0 against ?r, then gradient is r0 ?
ltrgt1.260.00R? Mass M g0r02/G 0.170.05M ?
Problems may be caused by using V as a proxy
for the angular diameter and, possibly, the
accuracy of Teff from spectral analysis. An IR
light curve would be helpful!
14
Extinction 1
IUE SWPLWR image pairs with Strömgren (uvbV)
photometry (Kilkenny Lynas-Gray 1982). Use ?2
minimization (FFIT) to solve for EB-V0.060.01
15
Extinction 2
IUE SWP HIRES spectra (merged) Synthetic spectrum
around Ly?, including contribution from
interstellar hydrogen (Groenewegen Lamers 1989,
Bohlin et al. 1978), yields EB-V0.070.02
16
Radius measurement 3
Measure Teff and ? from IUE SWPLWR image pairs
with Strömgren (uvbV) photometry (cf. Lynas-Gray
et al. 1984) Integrate dr/dt to obtain ?r By
identity, ??/?0 ? ?r/r0 Plot ??/?0 against ?r,
then gradient is r0 ? ltrgt2.310.02R? Mass M
g0r02/G 0.590.18M ? Problems 1)
reconciling photometric and spectroscopic
measurements of Teff. 2) lack of IUE data close
to minimum radius (one datum) A good HST light
curve would be helpful!
17
Infrared spectrum

• First IR spectrum for an extreme helium star
• LTE models
• HeI Dimitrevic Sahal Brechot (1990)
• H Lemke (1997)

• Weakness of 23S-23P line (10830Å)
• theoretically NOT the strongest HeI line in the
  spectrum of a helium star - atomic data?
• observationally much weaker than -predicted.
  NLTE?

Strength of Paschen lines
18
Linebroadening

• minimum radius occurs at ?0.140
• increased blending in weak lines
• asymmetries in strong lines
• ? doubling in HeI, HeII

19
V652 Her evolutionary models

• Constraints
• CNO-processed surface with some H
• M 0.7 MSun, L 103 Lsun,dP/dt ? rapid
  contraction

• merged white dwarf models
• HeHe (Iben 1990)
• dM/dt gt Eddington ? sdB star
• HeCO (Iben 1990 previous authors)
• dM/dt gt Eddington ? RCrB star
• Evolution critically sensitive to WD temperature
  at merger
• COCO (Saio Nomoto 1998)
• HeHe (Saio Jeffery 2000)
• dM/dt half Eddington
• HeCO (Saio Jeffery 2002)

• ad hoc models
• helium horizontal branch (Jeffery 1984)
• He core with luminous shell contracts onto Helium
  Main Sequence, reproduces M, L and dP/dt
• BUT no plausible progenitor
• mixed red giant branch (Sweigart 1996)
• Physics implausible, and see above
• binary mass transfer (case BB Iben Tutukov
  1984)
• V652 not a binary
• Final-flash white dwarf
• Luminosity too high for V652 Her
• Carbon abundance too high for V652 Her

20
The HeHe white dwarf merger model

• hypothesis
• HeHe white dwarf formed from binary star
  evolution (observed)
• orbit decays through gravitational, tidal and
  magnetic interaction
• less massive WD disrupted when Porb 4 minutes
  and forms thick disk
• more massive WD accretes material from disk
• ?model

21
V652 Her
accretion turned off at selected final mass
shell burns inwards in series of mild flashes
lifts degeneracy
helium-burning shell forces star to expand to
yellow giant, 103 yr
Helium core-burning star (sdB?) formed as shell
reaches centre
helium ignites in shell at core-envelope boundary
22
internal details
inward migration of helium-burning shell and
response of surface to shell flashes
extent of shell and surface convection zones
during first five shell flashes
23
pulsation properties linear analysis of
evolutionary models gives fundamental pulsation
period dP/dt, derivative of period wrt time (or
dP/dn) also obtained evolution track through
P-dP/dn diagram looks good !
V652 Her
24
V652 Her

• HeHe WD merger
• mass ?
• radius ?
• luminosity ?
• pulsation period ?
• dP/dt ?
• composition ?

25
V652 Her the pulsating helium star

• Period change (Kilkenny Lynas-Gray 1982 - 1996)
  
• corresponds to R/R0 2 .10-4 yr-1 (xdx/dt)
  but not-linear (R, R)
• Nonadiabatic linear pulsation models (Saio 1983 -
  1995)
• 1993 OPAL and OP opacities gt Z-bump opacity
  driving
• Nonlinear models (Fadeyev Lynas-Gray 1996)
• Best agreement for M0.72 MSun, Teff23 500K, L
  1062 LSun, Z0.0156
• Stellar atmosphere analysis (Jeffery et al. 1999)
• 1 H, Fe/H0, N-rich, C and O poor, log g 3.7
  / 0.1, Teff 24 500 / 500 K

• Extreme helium star (Berger Greenstein 1953)
• comparable to other helium stars HD124448,
  HD168476, BD10 2179
• Light variations (Landolt 1973)
• P0.108 days similar to ? Cepheids, but obviously
  not Pop I MS
• Radial velocity variations (Hill et al. 1981, )
• amplitude 70 km/s, rapid acceleration
  free-fall
• Radius measurement from Baades method
  (Lynas-Gray et al. 1984)
• M g.R2 0.7 / 0.3 Msun, L 103 Lsun

26
V652 Her pulsation models
log Te M log L
2.84
2.82
2.94
3.01
Best Obs log Te M log L4.32 0.59 2.96
27
V652 Her pulsation modelsMontañés Rodríguez
2002 PhD Thesis
28
V652 Her as an exotic star

• Unique
• Surface composition CNO-processed He
• Surface gravity and effective temperature
  low-mass evolved star
• Pulsations radius, contraction rate, internal
  structure
• Challenge to evolution
• Not a post-AGB star
• Not a HB star
• Looks like a WD merger
• A key for understanding less exotic stars ?

29
V652 Her as a future subdwarf B star

• four observationally distinguishable groups of
  sdB stars
• 1. short-period single spectrum nHelt0.01
• a. WD companion
• b. M-dwarf companion
• 2. long-period(?) composite spectrum nHe lt0.01
• BS companion
• 3. single spectrum, non-variable RV
• only group with nHegt0.03
• range 0.03ltnHelt1
• a. HeHe mergers
• b. high mass loss near RGB tip
• Given timescales 5000 He-rich sdB stars for
  every V652 Her in galaxy.

30
(No Transcript)
31
Extreme Helium Stars
A and B supergiants low mass dimensions
post-AGB stars no planetary nebulae no
binaries weak or absent hydrogen lines strong
carbon lines (most) rare (3 in HD catalogue)

• what are they?
• where did they come from?
• where are they going?

V652 Her may be the exception. What about more
luminous C-rich helium stars?
32
proposals
early ideas mass-loss on AGB hot
bottom burning final-flash a white dwarf
evolves Iben et al.1984 back to the
AGB white dwarf binary merger one white dwarf
breaks up, Webbink 1984, Iben Tutukov
1984 falls onto its companion and forms a
giant quantitative models? observational tests?
33
0.6 M? , X0.001
accretion turned off at selected final mass
helium-burning shell forces star to expand to
yellow giant, 103 yr
0.5 M? CO-WD
helium ignites in shell at core-envelope boundary
34
Temporal evolution of accreting WD
He ignition
H ignition
Mi0.6X0.001
convection zone
hydrogen-burning shell
helium-burning shell
35
Observational tests for COHe merged binary
white dwarf models

• 1. Binarity
• 2. R, M and L measurements for pulsating stars
• 3. Gravity measurements
• 4. Contraction rates
• 5. Surface abundances
• 6. Numbers

36
Binarity
COHe merger model Test 1

• No extreme helium star has been found to be a
  binary
• (radial velocity searches)
• (IR excess searches)
• (UV excess searches)

37
Radius measurement (Baades method)
IUE SWPLWP LORES fluxes model atmospheres
?Teff and ?? integrating radial velocities
??R R?. ?R/ ? ? L R2Teff4 M gR2/G PV Tel
2 others
38
COHe merger model Test 2 EHe masses
3 methods for estimating masses of EHes Ms
spectroscopic mass Mc-Ls g Mp pulsation mass ?
g Md direct mass ?R, ?, ??, g
39
COHe merger model Test 2
COHe mergers solid 0.6M?COHe dashed
0.5M?COHe light accretion heavy
contraction EHes Baade radii from pulsating
EHes Masses from log g
(0.94 M?)
EHe stars
(0.79 M?)
40
COHe merger model Test 3
HD168476
HD160641
41
Contraction measurement
COHe merger modelTest 4
150 IUE LORES spectra over 17-year baseline Teff
and ? measured from SWPLWP image pairs
42
vectors represent predicted temperature evolution
over 10 years for 0.7 and 0.9 (solid) Msun helium
stars respectively dT/dt?expected
43
COHe merger modelTest 4
HD160641
BD-9 4395
BD-1 3438
HD168476
44
COHe merger model Test 4 EHe contraction
contraction rates with masses will discriminate
between evolution models
45
COHe merger model Test 5surface abundances
1. Model assumed homogeneous accretion 2. A
simple recipe a. assume realistic composition
of different layers of progenitor white dwarfsb.
mix together in proportion to expected layer
masses 3. Observations of EHe, and RCrB stars
46
COHe merger model Test 6number densities

• 20 of all WD pairs include COHe WD (Nelemans
  et al 2001)
• COHe WD merger rate ? ? 4.4 10-3 yr-1
  (Nelemans et al. 2001) (Iben et al. give 2.3
  10-3 yr-1)
• Heating rates between 10 000 and 40 000 K are 10
  - 100 K yr-1, or evolution timescales ? ? 300 -
  3000 yr.
• Merger rate ? timescales gives number of EHes
  (N) in Galaxy between 1.3 and 13.
• There are 17 known EHes in this temperature
  range
• Stars cooler than 10000 K have ? ? 105 yr, ? N
  ? ? ? 30 - 300 cool COHe merger products.
• There are an estimated 200-1000 RCrBs in galaxy
  (Lawson et al. 1990), although only 33 are known
  (Alcock et al. estimate 3000 RCrBs).
• Model builders reckon anything within a factor
  three is excellent!

47
Observational tests for COHe merged binary
white dwarf models

• 1. Absence of binaries ?
• 2. Radii and masses for pulsating stars ?
• 3. Gravity measurements ?
• 4. Contraction rates ?
• 5. Surface abundances ?
• 6. Number densities ?

not bad!
48
Conclusions
The excellent correspondence between observation
and theory for the N-rich extreme helium star
V652 Her suggests that HeHe WD mergers really do
happen. A significant fraction of single
helium-rich subdwarf B stars should also be
formed through this channel. The COHe WD merger
model provides an excellent fit for all of the
observed properties of other (C-rich) luminous
Extreme Helium stars and, by association, the
RCrB and luminous He-sdO stars. It is the only
model to do this. Work is still required to match
the detailed surface abundances in both V652 Her
and EHes. The hydrodynamics of the merger event
itself must also be explored properly.
49
About the movie...