Stellar Autopsies from White Dwarf Pulsations

Colloquium, 45 min. May 2018: University of Hawaii, Manoa, HI.

Transcript

1. http://jjherm.es J.J. Hermes Hubble Fellow University of North Carolina at Chapel Hill Stellar Autopsies from White Dwarf Pulsations
2. All low-mass stars eventually run out of fuel, lose their envelope, and become a white dwarf More than 97% of all stars in our Galaxy are or will become white dwarfs
3. D. Berry, GSFC White dwarf stars mark the endpoints of stellar, binary and planetary evolution
4. Outline: Kepler/K2 Insights into White Dwarfs • Rotation § White dwarfs relatively slow (0.5-2.2 d), solid-body rotation • Ages § Seismology constrains white dwarf H and He layers § Cooling ages in our models could be wrong by up to 10% • A New Phase of Low-Mass Stellar Evolution § Rogue waves on the coolest pulsating white dwarfs
5. Kepler 12 May 2009 – 11 May 2013
7. 
   Steve Howell The solar pressure on the Kepler spacecraft is ~50 µN m-2
8. Ecliptic K2 Campaign 18 began on Sunday!
9. Original 4-year Kepler Mission: Just 20 white dwarfs observed K2 through Campaign 8: >900 white dwarfs K2 through Campaign 15: >1750 white dwarfs K1 K2, as of 2016 K2, today Through Campaign 17: >2250 white dwarfs
10. In K2 Campaign 6 we found an interacting WD+WD binary (AM CVn-type) orbiting one another every 15.7 minutes! Green, Hermes et al. 2018
11. Vanderburg et al. 2015 The first transits of a white dwarf were discovered in K2 Campaign 1 (rare: so far a one-off in >1750 WDs with K2 data) The object is disintegrating 4RWD model Gänsicke et al. 2016
12. GD 1212, Hermes et al. 2014a Data from the 9-day K2 engineering test run, 2014 January V=13.3 mag
14. (Astero)Seismology White Dwarf Earth We match the observed periods to models to discern stellar interior!
15. 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] White dwarfs are cosmic timepieces: They are excellent age indicators But we must tune the clocks!
16. Gaia will eventually uncover more than 300,000 new white dwarfs (just ~35,000 known today) GBP - GRP MG >31,000 WDs w/in 200pc!
17. SLoWPoKES: Dhital et al. 2015 Lots of science using white dwarfs in wide binaries as age calibrators 5” 5” 5”
18. g-modes—remarkably similar to the large-amplitude DAV pulsators (Winget et al. 19 The observed pulsating white dwarf stars lie in three strips in the H-R diagram, in Figure 3. The pulsating pre-white dwarf PG 1159 stars, the DOVs, around 7 170,000 K have the highest number of detected modes. The first class of pulsating 5.5 5.0 4.5 Planetary Nebula Main sequence DOV DBV DAV 4.0 3.5 3.0 log [T eff (K)] 4 2 0 –2 –4 log (L/L ) Figure 3 A 13-Gyr isochrone with z = 0.019 from Marigo et al. (2007), on which we have drawn the obs Annu. Rev. Astro. Astrophys. 2008.46:157-199. Downloaded fr by University of Texas - Austin on 01/28/09. For Winget & Kepler 2008 H He C/O Not all white dwarfs pulsate: We must select them!
19. GBP - GRP MG >31,000 WDs w/in 200pc MG GBP - GRP 4% “most variable” WDs w/in 200pc DAV (H-atm)
20. Mike Montgomery Pulsations are periodic brightness changes, caused by surface temperature variations hotter cooler
21. m = -1 m = +1 m = 0 1000 s 200 s 500 s 125 s 316.8 s 345.3 s n = Number of radial nodes l = Number of vertical nodes m = Number of horizontal + vertical nodes n l = 1 n = 5 l = 1 n = 6 Prot = 0.9 ± 0.2 day Typical K2 data from a pulsating white dwarf
22. Model-Independent Rotation Falls Readily from K2 Data Hermes et al. 2017d k2wd.org
23. 1 10 100 White Dwarf Rotation Period (hr) 0 2 4 6 8 10 N Kepler & K2 Kawaler (2015) Most isolated white dwarfs rotate between 0.5-2.2 days Hermes et al. 2017d: k2wd.org None of the stars are currently in binaries: Representative of single-star evolution of mostly 1-3 M¤ stars Model-Independent Rotation Falls Readily from K2 Data 0.5 d 1 d 2 d 4 d
24. SDSS SOAR spectroscopy yields WD mass We have obtained SOAR spectra of all pulsating white dwarfs observed so far by K2 Hermes et al. 2017d: k2wd.org
25. 1 10 100 WD Rotation Period (hr) 0.4 0.5 0.6 0.7 0.8 0.9 WD Mass (M⊙ ) 1.0 1.5 2.0 2.5 3.0 3.5 4.0 ZAMS Progenitor Mass (M⊙ ) 1 10 100 White Dwarf Rotation Period (hr) 0 2 4 6 8 10 N Kepler & K2 Kawaler (2015) 1 d 2 d 4 d We Can Finally Probe WD Rotation as a Function of Mass The fastest-rotating pulsating white dwarf (1.13 hr) is also the most massive (0.87 M¤ ) – descended from a single 4.0 M¤ ZAMS progenitor Hermes et al. 2017c Hermes et al. 2017d: k2wd.org
26. Most white dwarfs evolve from 0.9-3.0 M¤ ZAMS stars, and rotate at 0.5-2.2 days (Possible link emerging between higher white dwarf mass and faster rotation) 1 10 100 0 1 2 3 4 N 1.7 2.0 M ZAMS WD Prot = 1.48 ± 0.94 d 1 10 100 0 1 2 3 4 N 2.0 2.5 M ZAMS WD Prot = 1.35 ± 0.74 d 1 10 100 0 1 2 3 4 N 2.5 3.0 M ZAMS WD Prot = 1.32 ± 1.04 d 1 10 100 White Dwarf Rotation Period (hr) 0 1 2 3 4 N 3.5 4.0 M ZAMS WD Prot = 0.17 ± 0.15 d We Can Finally Probe WD Rotation as a Function of Mass Hermes et al. 2017d: k2wd.org
27. in Figure 3. The pulsating pre-white dwarf PG 1159 stars, the DOVs, around 75,000 K to 170,000 K have the highest number of detected modes. The first class of pulsating stars to be 5.5 5.0 4.5 Planetary Nebula Main sequence DOV DBV DAV 4.0 3.5 3.0 log [T eff (K)] 4 2 0 –2 –4 log (L/L ) Figure 3 A 13-Gyr isochrone with z = 0.019 from Marigo et al. (2007), on which we have drawn the observed locations of the instability strips, following the nonadiabatic calculations of C´ orsico, Althaus & Miller Bertolami (2006) for the DOVs, the pure He fits to the observations of Beauchamp et al. (1999) for the DBVs, and the observations of Gianninas, Bergeron & Fontaine (2006) and Castanheira et al. (2007, and references therein) for the DAVs. 172 Winget ·Kepler 2.5 M¤ A star: Prot,ZAMS ~ 10 hr Core-He RGB: modes ~0.02-0.10 R¤ Prot : 30-180 d White dwarf: ~0.005-0.013 R¤ Prot : 0.5-2.2 d 10 100 Secondary Clump Rotation Period (d) 0 1 2 3 4 5 6 7 8 N Deheuvels et al. 2015 Tayar et al., in prep. 1 10 100 WD Rotation Period (hr) 0.4 0.5 0.6 0.7 0.8 0.9 WD Mass (M⊙ ) 1.0 1.5 2.0 2.5 3.0 3.5 4.0 ZAMS Progenitor Mass (M⊙ ) 1 10 100 White Dwarf Rotation Period (hr) 0 2 4 6 8 10 N Kepler & K2 Kawaler (2015) Kepler has mapped internal rotation evolution all the way from MS to WD
28. White Dwarfs Do Not Rotate Differentially (Solid Body) Based on detailed asteroseismic model: PG0112+104 rotates rigidlyover its outer 70% in radius with a period of Prot = 10.18 ± 0.27 hr White dwarfs appear to lack radial differential rotation Giammichele et al. 2018, in prep.
29. Used to Be, Getting Data Required Going to the Telescope
31. *PG 1159 star = hot pre-white-dwarf (aka DOV)
32. Vauclair et al. 2002 Maidanak Observatory, Uzbekistan 2-week coordinated, global (14-site) Whole Earth Telescope run in 1994
33. THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev
34. Jul 27th Uzbeks introduced new rules for the visas ... spent long 8 night hours in the old stinking Russian bus, which, using longest possible route and stopping more than ten times for the repairs, after which passengers were supposed to push the bus to start the engine, brought us to Shakhrisabz. Jul 28th Old military jeep, which exhaust went more inside than via its pipes, after 5 hours brought us to Maidanak [Observatory]. ... Some windows of our living house were broken, no clean sheets ... no butter, meat, sugar. Running water system was not working anymore, not to mention hot water. THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev
35. Jul 29th I checked telescope; tracking and positioning were working, but telescope mirrors needed cleaning... Jul 30th Managed to repair distiller and to get 3 L of water late in the evening only. Decided to wash mirrors next day. Still lots of yellow Afghanistan dust in the sky. Jul 31st Washed mirrors, cleaned telescope inner surfaces from thick dust layer. Started the full scale system test. THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev
36. Aug 1st All day no clouds, but wind increasing to the evening. Worked all night. Aug 3rd All day clear sky with some clouds. Quite strong wind in day time but diminished before the night. THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev
37. Aug 5th It was first night there on the mountain without me. I was at that time in Kitab Hospital severely injured by the Tashkent Astrophysical Institute Director son Iskander Yuldashbaev, apparently mentally ill young man of about 21. THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev
38. Aug 5th It was first night there on the mountain without me. I was at that time in Kitab Hospital severely injured by the Tashkent Astrophysical Institute Director son Iskander Yuldashbaev, apparently mentally ill young man of about 21. He did some cleaning ... suddenly saying no words grabbed my hair with his left hand and hit my throat with a broken knife from our kitchen. I ran in horror, but he managed to hit me twice into my back. I ran to the Russian house for the help all in the blood. It was no phone connection with outside world and two of them had to run all the way to Maidanak to soldiers, and in three hours at last I was delivered to Kitab hospital in rather weak condition. THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev
39. THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev Aug 5th It was first night there on the mountain without me. I was at that time in Kitab Hospital severely injured by the Tashkent Astrophysical Institute Director son Iskander Yuldashbaev, apparently mentally ill young man of about 21. He did some cleaning ... suddenly saying no words grabbed my hair with his left hand and hit my throat with a broken knife from our kitchen. I ran in horror, but he managed to hit me twice into my back. I ran to the Russian house for the help all in the blood. It was no phone connection with outside world and two of them had to run all the way to Maidanak to soldiers, and in three hours at last I was delivered to Kitab hospital in rather weak condition. ... He is in a custody now and cannot say the reason either, says he did not like the way I looked at him. But he was smart enough to steal before that event good sum of my money ... Until helicopter arrived I explained the basics of the work with the quilt program to Alexey -- my assistant. Luckily I trained him on almost everything...
40. Aug 10th Alexey arrived from the Maidanak in the afternoon. Everything seems OK. Aug 11th Aug 12th I lived in the Russian hotel in Kitab ... working with data: writing logs, marking bad points. Tomorrow night Uzbeks promised to bring me to the Samarkand airport. My throat is swollen, still hurts and ugly. END OF CAMPAIGN HERE IN THE UZBEKISTAN ------------------------------------------------------------------------ THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev Aug 8th I ... practically defected from Kitab hospital, where black bugs were running on the walls at night even in the patient's beds, over the face too. Throat is badly swollen and hurts.
43. Today We Are Spoiled with Telescopes in Space
45. 1000 s 200 s 500 s 125 s White Dwarfs: g-modes, not all modes are observed excited BiSON; Thompson et al. 2003 5 min 4 min 6 min Solar p-modes
46. n = Number of radial nodes l = Number of vertical nodes n Each white dwarf has a spectrum of g-modes: standing waves that naturally resonate Adiabatic Model: 11,245 K, 0.632 M¤ , 10-4.12 MH /MWD (Romero et al. 2012) 1000 s 200 s 500 s 125 s l=1 l=2
47. n = Number of radial nodes l = Number of vertical nodes n 1000 s 200 s 500 s 125 s l=1 l=2 l=1 l=2
48. n = Number of radial nodes l = Number of vertical nodes n 1000 s 200 s 500 s 125 s l=1 l=2 l=1 l=2
49. If we only plot identified l=1 modes: 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 l = 1 n = 1 l = 1 n = 2 l = 1 n = 3 Kepler makes mode identification relatively trivial Mode Period (s) N SDSSJ0051+0339, g=17.6, K2 Campaign 8 n = 1 n = 2 n = 3 n = 4 Clemens, Dunlap, Hermes et al. 2018, in prep.
50. 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 Mode Period (s) N l = 1 n = 1 l = 1 n = 2 l = 1 n = 3 n = 1 n = 2 n = 3 n = 4 Clemens, Dunlap, Hermes et al. 2018, in prep. If we only plot identified l=1 (m=0) modes: Kepler makes mode identification relatively trivial
51. 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 Mode Period (s) N n = 1 n = 2 n = 3 n = 4 l = 1 n = 1 l = 1 n = 2 l = 1 n = 3 Clemens, Dunlap, Hermes et al. 2018, in prep. If we only plot identified l=1 (m=0) modes: Kepler makes mode identification relatively trivial
52. 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 l=1 DAV periods, observed Full evolutionary models computed by Romero et al. 2012 Clemens, Dunlap, Hermes et al. 2018, in prep.
53. Drawing from a random distribution of hydrogen layer masses Full evolutionary models computed by Romero et al. 2012 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 l=1 hDAV periods, observed 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 l=1 random MH simulation Clemens, Dunlap, Hermes et al. 2018, in prep.
54. Most (>80%) DA White Dwarfs Have Thick H Layers 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 l=1 hDAV periods, observed 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 l=1 random MH simulation l=1 canonical MH simulation Full evolutionary models computed by Romero et al. 2012 Only drawing from the models with canonically thick hydrogen layers Clemens, Dunlap, Hermes et al. 2018, in prep.
55. Asteroseismology:Model He Layers Too Thick 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 l=1 hDAV periods, observed 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 l=1 canonical MH simulation Full evolutionary models computed by Romero et al. 2012 10-15 s offset: Suggests He-layer masses too thick in canonical models à Would lead to systematically younger WD cooling ages (~10%) Only drawing from the models with canonically thick hydrogen layers Clemens, Dunlap, Hermes et al. 2018, in prep.
56. We Have Only Scratched the Surface of Analyzing the ~100 Pulsating White Dwarfs Observed by Kepler WDs Evolve (Cool) à Blue: Observed by Kepler Open: Ground-based
57. Coming Soon: NASA/TESS will observe all bright WDs every 2 min All systems nominal. Sector 1 planned to start end of June! At least 28 days per sector ~70 I < 16 mag white dwarfs per sector
58. 2 Bell et al. Fig. 1.— Representative sections of the Kepler light curve of KIC 4552982 in units of days since the start of observations. The top p shows the full Q11 light curve. The one-month shaded region in the top panel is expanded in the middle panel. The one-week sh region in the middle panel is expanded in the bottom panel. The solid line is the light curve smoothed with a 30-minute window. point-to-point scatter dominates the pulsation amplitudes in the light curve, so pulsations are not apparent to the eye. The dram increases in brightness are discussed in detail in Section 3. to medium-resolution spectra for the white dwarf and fit the Balmer line profiles to models to determine its val- tion rate. We summarize our findings and conclud Section 5. KIC 4552982: Bell et al. 2015 3 months: 1 month: 1 week: Brightenings every ~2.7 d, lasting for 4.0-25.0 hr A surprising discovery with Kepler: Aperiodic Outbursts
59. This outburst phenomenon never seen before in 40+ years of pulsating white dwarf studies A surprising discovery with Kepler: Aperiodic Outbursts Quiescent pulsations (1151.9 s, 1160.8 s, …) In Outburst (999.9 s, 896.6 s, …) PG 1149+057: Hermes et al. 2015b
60. • White dwarf Teff = 11,060 K • é 15% mean flux = é 750 K • ~1034 erg energy Black line is 30-min running mean Event 1 Event 7 Quiescence Pulsations Persist in Outburst, But Surface >700 K Hotter
61. A surprising discovery with Kepler: Aperiodic Outbursts Keaton Bell 2017 (PhD thesis) Outbursts seen in at least 13 white dwarfs (this is not rare!) None of the outbursts in pulsating WDs are periodic; appear chaotic Outburst recurrence times can be as short as a few hours and as long as 45+ days
62. The vast majority of cool pulsating white dwarfs outburst! All white dwarfs pulsate at the appropriate temperature, and it appears all outburst at some point, too. This is likely a new phase of stellar evolution! Outbursting DAVs Blue: Observed by Kepler Open: Ground-based
63. GD 1212, Hermes et al. 2014a Data from the 9-day K2 engineering test run V=13.3 mag
64. A surprising discovery with Kepler: Aperiodic Outbursts Quiescent pulsations (1135.2 s, 856.9 s, …) In Outburst (864.1 s, 846.4 s, …) GD 1212: Hermes et al. 2018, in prep. >60 days between outbursts! K2 Campaign 12
65. The vast majority of cool pulsating white dwarfs outburst! All white dwarfs pulsate at the appropriate temperature, and it appears all outburst at some point, too. This is likely a new phase of stellar evolution! Outbursting DAVs Blue: Observed by Kepler Open: Ground-based
66. 150 s 1000 s The coolest pulsating white dwarfs have longer-period pulsations à Higher mode density for coupling
67. • Wu & Goldreich (2001) predicted nonlinear mode coupling could transfer energy into damped modes in the cool DAVs l=1 l=2 Adiabatic Model: 11,245 K, 0.632 M¤ , 10-4.12 MH /MWD Observed: 11,060(170) K, 0.64(0.03) M¤ (Romero et al. 2012) (Gianninas et al. 2011) Likely Cause: Mode Coupling via Parametric Resonance ωp = 897.7 µHz (l=1, m=0, n=24) à Standing waves Evanesce ß
68. • Wu & Goldreich (2001) predicted nonlinear mode coupling could transfer energy into damped modes in the cool DAVs ωp = 897.7 µHz (l=1, m=0, n=24) ωd1 = 435.9 µHz (l=2, m=0, n=88) l=1 l=2 ωd2 = 461.9 µHz (l=1, m=0, n=48) Adiabatic Model: 11,245 K, 0.632 M¤ , 10-4.12 MH /MWD Observed: 11,060(170) K, 0.64(0.03) M¤ (Romero et al. 2012) (Gianninas et al. 2011) Likely Cause: Mode Coupling via Parametric Resonance ωd1 + ωd2 = ωp + δω Limit cycle if: δω < γd
69. ωp = 897.7 µHz (l=1, m=0, n=24) ωd1 = 435.9 µHz (l=2, m=0, n=88) l=1 l=2 ωd2 = 461.9 µHz (l=1, m=0, n=48) Adiabatic Model: 11,245 K, 0.632 M¤ , 10-4.12 MH /MWD Observed: 11,060(170) K, 0.64(0.03) M¤ (Romero et al. 2012) (Gianninas et al. 2011) Likely Cause: Mode Coupling via Parametric Resonance ωd1 + ωd2 = ωp + δω Limit cycle if: δω < γd
71. Key Takeaways • Kepler/K2 has us in a new regime of looking at white dwarfs • Rotation: Stars lose most core angular momentum before WD stage § The endpoints of 1-3 solar-mass stars rotate at 0.5-2.2 days § They are not rotating with any detectable radial differential rotation § Evidence for trend of faster rotation with higher mass • Ages: It is rare for white dwarfs to have thin H layers § Seismology suggests >80% of DAs have H layers ~10-4 MH /Mstar § We see evidence that the He layers from models are 10x too thick § Cooling ages may be overestimated by up to 10% • Outbursts: Nonlinear mode coupling rampant in pulsating WDs § Nearly all cool pulsating WDs undergo stochastic outbursts ~1034 erg § Parametric resonance of excited mode(s) with damped daughter modes
72. KIC08626021: Giammichele et al. 2018 Core Surface ß 99% of mass X(O) = 78.03% ± 4.2% X(C) = 21.96% ± 4.2% X(He) = 0.0113% ± 0.006% With enough pulsation modes we can model entire white dwarf C/O ratio constrains 12C(α,γ)16O reaction rate