The Menagerie of Hydrogen-Deficient White Dwarfs

Conference presentation, 30 min. September 2018: Hydrogen Deficient Stars 2018, Armagh, Northern Ireland, UK.

Transcript

1. The menagerie of hydrogen-deficient white dwarfs http://jjherm.es J.J. Hermes Hubble Fellow University of North Carolina at Chapel Hill
2. Outline: The Menagerie of Hydrogen-Deficient WDs via single-star progenitors via binary coalescence DA DB hot DQ LP 40-365
3. Outline: The Menagerie of Hydrogen-Deficient WDs • Spectral evolution of descendants of single stars • Asteroseismic constraints of H layers in DA (H-dominant) WDs • Kinematics and masses of DAs and non-DAs • Evidence for a population of WD+WD mergers: the hot DQs • New class partially burnt supernova remnants slungshot from the Galaxy (e.g., LP 40-365)
4. What Do We Mean by ‘White Dwarf’? • a stellar remnant that is no longer fusing in its core • the endpoints of everything < 8 M¤ • electron degeneracy limits WD mass to < 1.4 M¤ A ‘typical’ white dwarf electron degenerate C/O core (r = 8500 km) non-degenerate He layer (260 km) non-degenerate H layer (30 km) [thermal reservoir] [insulating blanket] DA
5. 4000 4500 5000 5500 6500 DA DA: H DB: He DQ: C2 (‘Swan bands’) DC: [continuum] DZ: [metals] DB DZ DQ DC The Most Common White Dwarf Flavours in Nature adapted from Wesemael et al. 1993
6. Appeal to the Modern: the Gaia Colour-Magnitude Diagram Gaia Collaboration, Babusiaux et al. 2018
7. 100 pc sample: 18,702 high-probability WDs 100 pc sample: Gentile Fusillo et al. 2018
8. 100 pc sample: 2145 WDs with spectroscopy DA: log(g) = 8.0 DA: log(g) = 9.0 100 pc sample: Gentile Fusillo et al. 2018 crossed with the Montreal White Dwarf Database: Dufour et al. 2015
9. 100 pc sample w/ spectra: 1578 (~75%) are DAs DA: log(g) = 8.0 DA: log(g) = 9.0 100 pc sample: Gentile Fusillo et al. 2018 crossed with the Montreal White Dwarf Database: Dufour et al. 2015
10. 100 pc sample w/ spectra: 88 (~4%) are DBs DA: log(g) = 8.0 DA: log(g) = 9.0 100 pc sample: Gentile Fusillo et al. 2018 crossed with the Montreal White Dwarf Database: Dufour et al. 2015
11. Gaia Shows Spectral Types are Temperature-Dependent
12. The Balmer Jump Strongly Affects WD Atmospheres DA: log(g) = 8.0 200 pc sample: Gentile Fusillo et al. 2018 crossed with the Montreal White Dwarf Database: Dufour et al. 2015
13. Gaia CMD Shows Spectral Types are Evolution-Dependent DA: log(g) = 8.0 200 pc sample: Gentile Fusillo et al. 2018 crossed with the Montreal White Dwarf Database: Dufour et al. 2015
14. 100 pc sample w/ spectra: ~25% non-DA
15. DA vs. non-DA Ratio is a Function of Sample Selection 100 pc sample (volume-limited): DA: ~65% non-DA: ~35% SDSS sample (magnitude-limited): DA: ~80% non-DA: ~20% e.g., Kleinman et al. 2013 Kilic et al. 2018
16. Gaia CMD Shows Spectral Types are Evolution-Dependent 200 pc sample: Gentile Fusillo et al. 2018 crossed with the Montreal White Dwarf Database: Dufour et al. 2015
17. Dredge-Up from Convection Leads to Spectral Evolution • DB à DQ when He convection reaches c2 =nC /(nHe +nC ) > 10-6 • This naturally explains the cool DQs • Eventually all line opacities fade the WD into a DC Fontaine & Wesemael 1991 45 kK 18 kK 8 kK He convection zone c2 = nC /(nHe +nC ) c2 = 10-10 c2 = 0.99 core photosphere log M/M★
18. When Cool Enough, a DA can Transform to a DB • With a thin enough H layer (figure shows 10-11 MH /M ★ ), a DA convection zone dredges up He • Again, eventually all line opacities fade the WD into a DC Fontaine & Wesemael 1991 H convection zone towards core log M/M★
19. First Clue to Spectral Evolution: The ‘DB gap’ • First big clue of spectral evolution: the ‘DB gap’, which is a dearth of DBs between 30-45 kK • Requires H to be very thin (<10-14 MH /M ★ ) 45 kK 10 kK 6 kK Greenstein et al. 1986
20. How Thick Is the Hydrogen Layer in Typical DA WDs? • We can explore chemical layers via asteroseismology as well as eclipsing binaries! See wonderful reviews by: Winget & Kepler 2008 Fontaine & Brassard 2008 Althaus, Córsico, Isern & García-Berro 2010 A ‘typical’ white dwarf electron degenerate C/O core (r = 8500 km), 99% M ★ non-degenerate He layer (260 km) 1% MHe /M★ non-degenerate H layer (30 km) <0.01% MH /M★ [thermal reservoir] [insulating blanket] DA
21. WD+dM Eclipsing Binaries: <2% WD Masses, Radii • No evidence for very thin H layers in 13 WD in close WD+dM binaries • All have <10-8 MH /M ★ Parsons et al. 2017 He-core models C/O-core models Thick H (10-4) Thin H (10-10)
22. Asteroseismology: Pulsations Constrain Envelope Masses Detailed study of two superficially similar pulsating WDs: GD 165 and Ross 548 Giammichele et al. 2015 Time (s) Rel. Flux Rel. Flux Both white dwarfs have Teff ~ 12,100 K and are ~0.64 Msun but quite different pulsation properties
23. Asteroseismology: Pulsations Constrain Envelope Masses Thick H Layer: 10-4.23±0.15 MH /M ★ He Layer: 10-1.70±0.13 MHe /M ★ Giammichele et al. 2016 Thin H Layer: 10-7.45±0.12 MH /M ★ He Layer: 10-2.92±0.10 MHe /M ★
24. Asteroseismology: Insights from the Aggregated Periods Clemens et al. 2018, in prep. Ross 548 GD 165 l = 1, k = 2 l = 1, k = 1 Thick H Layer: ~10-4 MH /M ★ He Layer: ~10-1.7 MHe /M ★ “Canonical” nuclear burning sets envelope masses Thin H Layer: <10-7 MH /M ★ ~He Layer: 10-2.9 MHe /M ★ Very late thermal pulses? Giammichele et al. 2016 size = amplitude of mode
25. Asteroseismology: Insights from the Aggregated Periods Clemens et al. 2018, in prep. Ross 548 GD 165 l = 1, k = 2 l = 1, k = 1 Thick H Layer: ~10-4 MH /M ★ He Layer: ~10-1.7 MHe /M ★ “Canonical” nuclear burning sets envelope masses Thin H Layer: <10-7 MH /M ★ ~He Layer: 10-2.9 MHe /M ★ Very late thermal pulses? Interpulse interaction? Giammichele et al. 2016 size = amplitude of mode ~80% of DAs have canonically thick (~10-4 MH /M ★ ) envelopes ~20% of DAs have thinner (~10-7-9 MH /M ★ ) envelopes N = 14 N = 4
26. (Photometric) Mass Distribution of DA vs. DB 200 pc sample: Gentile Fusillo et al. 2018 DA: <M> = 0.65 M¤ DB: <M> = 0.61 M¤ Mass (M¤ )
27. Kinematic Distribution of DA vs. DB 200 pc sample: Gentile Fusillo et al. 2018 DB: <vtan > = 40.7 km/s DA: <vtan > = 39.7 km/s • Historically it has been found there is no difference in kinematics between DA vs. DB • This is consistent with 200 pc sample from Gaia Sion et al. 1988
28. BUT: DQ and maybe DC show bimodality at ~50 km/s DB DA 200 pc sample DQ DC • Kinematics can give us insights into possible merger history
29. Stars Get Stirred Up Over Time in the Galaxy • Hot (>15 kK) massive WDs descending from single stars were born <1 Gyr ago • They should thus have low velocity dispersions Dunlap & Clemens 2015 1.2 M¤ WD cooling kinematics
30. Most Massive DAs Have Low Kinematics DA: M < 0.75 M¤ Teff > 15,000 K DA: M > 0.90 M¤ Teff > 15,000 K Wegg & Phinney 2012 • PG & SDSS samples suggest most massive DAs are evolved single stars
31. Specifically, Hot DQs Appear to be Merger Byproducts • Hot DQs: 18,000-26,000 K Dunlap et al. 2018, submitted DA < 0.75 M¤ DA > 0.75 M¤ hot DQs (all >0.90 M¤ ) Dufour et al. 2008 see especially Dunlap & Clemens 2015 • Hot DQs: No H: Mostly C Williams et al. 2013
32. Specifically, Hot DQs Appear to be Merger Byproducts Dufour et al. 2013 • Hot DQs: 18,000-26,000 K • 0.9-1.2 M¤ WDs (massive) • ~70% strongly magnetic (>2 MG) • Most 5-20 min monoperiodic variables (very fast rotation; most WD rotate 0.5-2 d) Dunlap et al. 2018, submitted DA < 0.75 M¤ DA > 0.75 M¤ hot DQs (all >0.90 M¤ ) Dufour et al. 2008 Dunlap et al. 2018 see especially Dunlap & Clemens 2015 Williams et al. 2016 • Hot DQs: No H: Mostly C Williams et al. 2013
33. Can We Find Further Evidence of Mergers in the CMD? 100 pc sample: Gentile Fusillo et al. 2018
34. Tangential Velocity Cuts in the Gaia CMD 100 pc sample: Gentile Fusillo et al. 2018
35. Tangential Velocity Cuts in the Gaia CMD 100 pc sample: Gentile Fusillo et al. 2018
36. Tangential Velocity Cuts in the CMD 100 pc sample: Gentile Fusillo et al. 2018
37. Using Gaia to Revisit the GD Sample Hermes et al. 2018, in prep. 0.50 0.25 0.00 0.25 0.50 0.75 1.00 1.25 GBP GRP [mag] 0 2 4 6 8 10 12 14 MG = G + 5 ⇥ log $ 10 [mag] DA, log(g) = 8.0 Z=0.019 Z=10 2 Z=10 3 0 100 200 300 400 500 v? [km s 1] • Gaia CMD: <400 WDs among >1700 WD suspects in Giclas Dwarf catalog (Giclas, Burnham & Thomas 1980)
38. Using Gaia to Revisit the GD Sample Hermes et al. 2018, in prep. 0.50 0.25 0.00 0.25 0.50 0.75 1.00 1.25 GBP GRP [mag] 0 2 4 6 8 10 12 14 MG = G + 5 ⇥ log $ 10 [mag] DA, log(g) = 8.0 Z=0.019 Z=10 2 Z=10 3 0 100 200 300 400 500 v? [km s 1] • Gaia CMD: <400 WDs among >1700 WD suspects in Giclas Dwarf catalog (Giclas, Burnham & Thomas 1980) GD 492
39. GD 492 = LP 40-365: A Hyper-Runaway WD discovery: Vennes et al. 2017 follow-up: Raddi et al. 2018a, 2018b • LP 40-365 has vrad = +499(6) km/s and vrf = 852(10) km/s • It is unbound, a hyper-runaway not from Galactic center • Gaia: 0.18(1) R¤ , crossed Z = 0 <5.3 Myr ago
40. GD 492 = LP 40-365: A Hyper-Runaway, Ne-rich WD discovery: Vennes et al. 2017 follow-up: Raddi et al. 2018a, 2018b Iax unburnt remnants Iax yields C/O or C/O/Ne Iax models from Fink et al. 2014 and Kromer et al. 2015 • LP 40-365 is >30% Ne and ~2% O by mass • H/He < 10-5 • (He invisible at 8900 K) • Alpha elements indicate C, Si processing • [Mn/Fe] > 7x solar • Hypothesis: Remnant of SN Iax, near-MCh , ejected from <40-min binary! (Flip side of coin from D6 stars Shen et al. 2018)
41. A New Class of Hyper-Runaway, Mg/Ne-rich WD Raddi et al. 2018c, in prep. • With Gaia we found 2 more! (a) (b) a) 3 mag brighter than GD 492; much hotter and larger; on retrograde but bound orbit b) Complete twin to GD 492; vRV = -480 km/s
42. Spectroscopic Twin and Kinematic Doppelgänger to GD 492 Raddi et al. 2018c, in prep. • Nearly identical radius and mass; vrf = 800 km/s (also unbound) • GD 492 has a slightly higher Mg abundance • Otherwise, they are startlingly similar • Formation mechanism for these slung-shot remnants must be similar
43. LP 40-365 and D6 Friends LP 40-365 LP 40-365 D6 LP 40-365: Raddi et al. 2018b D6: Shen et al. 2018 DOx
44. Are ’DOx’ Cooled-Down, Non-Ejected Versions of LP 40-365? first O-rich WDs: Gänsicke et al. 2008 Most O-rich: Kepler, Koester & Ourique 2016 • SDSSJ1240+6710 is 21 kK • Composed of 99.9% O • log(g) suggests 0.56 M¤ • Vrf = 260 km/s, but on a retrograde orbit Kepler, Koester & Ourique 2016 • SDSS has found spectra of a few log(g) ~ 8.0 WDs with no H and very high O content: so-called DOx LP 40-365 D6 DOx
45. Summary: The Menagerie of Hydrogen-Deficient WDs • Gaia shows WD spectral types are strongly dependent on cooling • Spectral evolution (DAà DB à DQ, etc.) involves both convective dredge- up and requires a range of H-layer masses (>10-4 to <10-15 MH /M ★ ) • Asteroseismology: ~80% of DAs have canonically thick H layers (>10-4 MH /M ★ ) • Gaia: Possible evidence of kinematic difference between DA and DQ? • Gaia: No difference in mean mass between DA and DB • The hot DQs (>18 kK) are massive, magnetic, rapid rotators, kinematicallyhot • LP 40-365 is the first in a class of Mg- and Ne-rich, hyper-runaway remnants work led by Roberto Raddi work led by Bart Dunlap work led by Chris Clemens
46. Thank you!
47. 10-4-6 MH /M ★ 10-6-14 <10-15 DAO DA DC DAO DA DC PG1159 DO DA DB DQ DC ~60 kK ~6 kK ~60 kK ~10 kK ~100 kK ~45 kK ~30 kK ~12 kK ~6 kK WD+WD hot DQ DC SN Iax LP 40-365 DOx? Mostly from single-star evolution Mostly from binary coalescence ~65% DA ~35% non-DA <1% DQ ~80% DAs (~50% total) ~20% DAs (~15% total) bound remnant unbound; Ne-rich O/C > 1 kinematics; mass; magnetic; fast Prot DA CZ <13 kK DB CZ <13 kK gravitation settling dominates <80 kK radiative levitation impactful >25 kK winds possible > 35 kK?