White dwarf discoveries enabled by K2

Colloquium, 45 min. November 2016: NASA Ames Research Center, Mountain View, CA, USA.

Transcript

1. http://jjherm.es J.J. Hermes Hubble Fellow University of North Carolina at Chapel Hill Rogue waves, disintegrating asteroids, and precision asteroseismology: White dwarf discoveries enabled by K2 NASA Ames, 3 Nov. 2016
2. U. North Carolina: Chris Clemens, Bart Dunlap, Erik Dennihy, Josh Fuchs, Stephen Fanale U. Warwick: Boris Gaensicke, Paul Chote, Roberto Raddi, Nicola Gentile Fusillo, Dave Armstrong, Pier-Emmanuel Tremblay U. Texas: Keaton J. Bell, Mike Montgomery, Don Winget + Steve Kawaler, Agnes Bischoff-Kim, Alex Gianninas, S.O. Kepler, Alejandra Romero Rogue waves, disintegrating asteroids, and precision asteroseismology: White dwarf discoveries enabled by K2 NASA Ames, 3 Nov. 2016
4. Tremblay et al. 2016 Kleinman et al. 2013 • Most have hydrogen-dominated atmospheres (DA) • Estimate masses from observed Balmer line profiles: Teff /log(g) White Dwarfs are Remarkably Homogeneous
5. Sun White Dwarf (60% Mass of Sun) Earth (0.0003% Mass of Sun)
6. Sharyn Morrow A ‘Typical’ White Dwarf Carbon/Oxygen Core (r = 8500 km) Helium Layer (260 km) Outer Hydrogen Layer (30 km) Under Strong Gravity, Heaviest Elements Sink
7. Typical DA white dwarf log(g) = 8.0 • Settling times << years • Radiative levitation inefficient <25,000 K • Expect pure hydrogen photospheres
8. DA white dwarf + metals But many show metals!
9. • Metals must be actively accreted, tidally disrupted debris – 25-50% of all WDs are metal polluted (Koester et al. 2014) – WD debris is comparable to bulk Earth (dominated by Fe, O, Si, Mg) – Some of this debris is water-rich! (Farihi et al. 2013) • Ergo, 25-50% of all A-F stars harbor planetary systems that survive to the WD stage WDs Directly Measure Exoplanetesimal Compositions
10. An Unexpected Beneficiary of a Second Broken Reaction Wheel: WHITE DWARF STARS
11. Original Kepler Mission: 20 white dwarfs observed, 6 pulsating WDs (just two >3 months) K2 through Campaign 9: >970 WDs observed 36 pulsating WDs K2 through Campaign 13: >1300 WDs, >60 pulsating WDs (~250 known today) K1 K2, today K2, by mid-2017
12. Vanderburg et al. 2015 A disintegrating asteroid transiting a white dwarf in K2 ground-based follow-up
13. Gänsicke et al. 2016; Rappaport et al. 2016 Mark Garlick More follow-up: the object really is disintegrating 4RWD model
14. WD 1145+017 is also heavily metal-polluted So far it is the only one of the ~1000 K2 WDs to show transits
15. K2: White Dwarf Variability on All Timescales
16. K2 Shows: White Dwarfs Are Good Flux Standards >97% of all spectroscopically confirmed WDs in Kepler/K2 are flux constant to <1% on 30-min to 10-day timescales 5/252 spectroscopic WDs 9/424 with colors/proper motions Kp < 19.0 mag (Cut pulsating WDs & known WD+dM) Zhichao Xue & Brad Schaefer are exploring white dwarf stability at mmag levels
17. Caveats: Binarity, Magnetism, Pulsations Reflection effect in close binary
18. Dozens of New, Well-Constrained Pre-CV Systems 100 101 102 Orbital Period (hr) 0 2 4 6 8 10 12 14 16 N New K2 PCEB Known PCEB 0.5 d 1.0 d - 80 spectroscopic WD+dM in K2 - 20 are PCEB -- Porb for free!
19. Caveats: Binarity, Magnetism, Pulsations
20. K2 Already Doubled WD Rotation Measurements 10 1 100 101 102 White Dwarf Rotation Period (hr) 0 2 4 6 8 N K2 Asteroseismic Asteroseismic K2 Magnetic Magnetic 0.5 d 1.0 d 5.0 d 3 hr
21. EC 14012-1446, r= 15.7 mag 98.2% duty cycle for 78.9 days Caveats: Binarity, Magnetism, Pulsations
22. DA (hydrogen atmosphere) WDs pulsate when H partially ionized (DAVs, aka ZZ Cetis) 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 ob locations of the instability strips, following the nonadiabatic calculations of C´ orsico, Althaus & Bertolami (2006) for the DOVs, the pure He fits to the observations of Beauchamp et al. (1999 Annu. Rev. Astro. Astrophys. 2008.46:157-199. Downl by University of Texas - Austin on 01/28 Winget & Kepler 2008, ARA&A, 46, 157
23. Mike Montgomery • Pulsations: periodic brightness changes, caused by surface temperature variations • White dwarfs only show nonradialpulsations (strong surface gravity)
24. Empirical DAV instability strip Tremblay et al. 2011 Gianninas et al. 2011 Gianninas et al. 2014 (ELMs) 3D-corrected atmospheric parameters, ML2/α = 0.8 Pulsations are driven in a narrow temperature range
25. Empirical DAV instability strip Tremblay et al. 2011 Gianninas et al. 2011 Gianninas et al. 2014 (ELMs) 3D-corrected atmospheric parameters, ML2/α = 0.8 Pulsations are driven in a narrow temperature range
26. 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
27. Mode Identification Made Easy with K2 l=1 196.84 s l=2 146.34 s (Both values consistent with ~1.2 d rotation.)
28. A White Dwarf with Truncated Red Giant Evolution Spin Faster than Isolated WDs
29. A K2 View on Close, Evolved Binaries M-dwarf RV (VLT/FORS2) WD atmospheric parameters (SOAR) Teff = 12,330 ± 260 K log(g) = 7.99 ± 0.06 (0.601 ± 0.036 M¤ ) SDSS SOAR VLT Porb = 6.8976 hr • WD+dM in K2 Campaign 1: SDSS J1136+0409 Hermes et al. 2015
30. A K2 View on Close, Evolved Binaries M-dwarf RV (VLT/FORS2) WD atmospheric parameters (SOAR) Teff = 12,330 ± 260 K log(g) = 7.99 ± 0.06 (0.601 ± 0.036 M¤ ) SDSS SOAR VLT Porb = 6.8976 hr (Model: Doppler beaming, reflection, ellipsoidal variations using spectroscopic parameters) Folded K2 light curve • WD+dM in K2 Campaign 1: SDSS J1136+0409 Hermes et al. 2015
31. A K2 View on Close, Evolved Binaries M-dwarf RV (VLT/FORS2) WD atmospheric parameters (SOAR) Teff = 12,330 ± 260 K log(g) = 7.99 ± 0.06 (0.601 ± 0.036 M¤ ) SDSS SOAR VLT Porb = 6.8976 hr (Model: Doppler beaming, reflection, ellipsoidal variations using spectroscopic parameters) Folded K2 light curve • WD+dM in K2 Campaign 1: SDSS J1136+0409 Hermes et al. 2015 5 independent pulsation modes
32. A K2 View on Close, Evolved Binaries J1136+0409 Prot : 2.49 ± 0.53 hr l = 1 modes m = +1 m = 0 m = -1 Hermes et al. 2015
33. A K2 View on Close, Evolved Binaries 10 1 100 101 102 White Dwarf Rotation Period (hr) 0 1 2 3 4 5 6 N Non-magnetic CVs Pulsating white dwarfs J1136+0409 J1136+0409 Prot : 2.49 ± 0.53 hr ~Days ~Minutes • No isolated WD rotates this fast • No accretion history in J1136+0409 • Truncated red giant evolution affects white dwarf rotation Hermes et al. 2015 l = 1 modes m = +1 m = 0 m = -1
34. Using White Dwarf Homogeneity to Exploit Simple K2 Observations
35. The White Dwarf Seismologist’s Dilemma For any one hot DAV: 1. Small number of independent modes observed 2. Best model hinges on mode identification 3. Hidden free parameters (mostly core profile), with 8 degrees of freedom WD0111+0018: Hermes et al. 2013 Just two independent modes
36. Insights from the Aggregated Period Spectrum 239 periods from 75 hot ZZ Ceti stars (mostly ground-based, no harmonics) Histogram of periods (again, no harmonics) 0 5 10 15 20 25 30 50 100 150 200 250 300 350 400 450 500 Mode Amplitude (mma) Mode Period (s) 0 5 10 15 20 25 30 50 100 150 200 250 300 350 400 450 courtesy of Chris Clemens
37. The Kepler-Cleaned Aggregated Period Spectrum 38 periods from 16 hot ZZ Ceti stars (l=1, m=0, no harmonics, mostly Kepler/K2) 0 1 2 3 4 5 6 7 8 50 100 150 200 250 300 350 400 450 courtesy of Chris Clemens
38. Comparing to a Random Distribution of Models 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 courtesy of Chris Clemens Romero et al. 2012
39. K2 Is Constraining Red Giant Evolution • The observed distribution supports thick (canonical) hydrogen layers • Implications for improving white dwarf cooling ages (Gaia) 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 courtesy of Chris Clemens Romero et al. 2012
40. K2 Is Constraining Red Giant Evolution • The observed distribution supports thick (canonical) hydrogen layers • Implications for improving white dwarf cooling ages (Gaia) 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 courtesy of Chris Clemens Romero et al. 2012 • Slight offset suggests helium layer of models are too thick: Set by thermal pulses on the AGB
41. E. L. Robinson Nather et al. 1990 Gaps in data cause cycle-count confusion (aliasing)
42. Nather et al. 1990 Coordinated global campaigns can minimize gaps
43. Used to Be, Getting Data Required Going to the Telescope
44. *PG 1159 star = hot pre-white-dwarf (aka DOV)
45. Vauclair et al. 2002 Maidanak Observatory, Uzbekistan
46. THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev
47. 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
48. 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
49. 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
50. 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
51. 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, then came to my room ... suddenly saying no words grabbed my hair with his left hand and hit my throat with a broken knife from our kitchen. I jumped from my chair, ran in horror out of the house, but he managed to hit me twice into my back until I was out. 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
52. 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, then came to my room ... suddenly saying no words grabbed my hair with his left hand and hit my throat with a broken knife from our kitchen. I jumped from my chair, ran in horror out of the house, but he managed to hit me twice into my back until I was out. 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 in previous nights on almost everything... THE MAIN LOG Observations at Maidanak observatory in Uzbekistan. Aug 1994 Observers: E. Meistas, and local assistant Alexey V. Chernyshev
53. Aug 8th All day clear sky, in the evening no clouds too, but bad seeing. I was in Kitab at russian astronomer place all throat bandaged, 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. 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, making .op files. 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
56. GD 1212, Hermes et al. 2014 The First K2 Pulsating White Dwarf
57. The First K2 Pulsating White Dwarf Eng. Run, KP = 13.3 mag, will revisit in Field 12 9 days, K2 77 days, K2 GD 1212, Hermes et al. 2014
58. The First K2 Pulsating White Dwarf Eng. Run, KP = 13.3 mag, will revisit in Field 12 9 days, K2 77 days, K2 My throat is swollen, still hurts and ugly. END OF CAMPAIGN HERE IN THE UZBEKISTAN ------------------------------------------------------------------------ GD 1212, Hermes et al. 2014
59. “I Was Promised Rogue Waves” Draupner wave, 1995
60. The First Kepler Pulsating White Dwarf was Weird 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
61. In K2, Things Got Weirder • In the first K2 campaign we saw another case of outbursts • These outbursts are essentially rogue waves (or freak waves) on a pulsating star! • Never seen in 40+ years of pulsating white dwarf studies PG 1149+057: Hermes et al. 2015 Quiescence (1151.9 s, 1160.8 s, …) In Outburst (999.9 s, 896.6 s, …) g = 14.9 mag
62. Pulsations Persist in Outburst • White dwarf Teff = 11,060 K • é 14% mean flux = é 750 K • é >25% flux = é >1500 K Black line is 30-min running mean Event 1 Event 7 Quiescence
63. A Second Case of Outbursts in a Cool DAV Pulsations are affected by outbursts PG 1149+057: Hermes et al. 2015
64. A Second Case of Outbursts in a Cool DAV (3-day sliding window) PG 1149+057: Hermes et al. 2015
65. oDAV1 10860 K 0.70 M¤ oDAV2 11060 K 0.64 M¤ oDAV3 10570 K 0.56 M¤ oDAV4 11190 K 0.62 M¤ oDAV5 10850 K 0.53 M¤ K2 keeps finding outbursting white dwarfs: Now 8 known! All aperiodic!
66. 52 WDs within 2000 K of 10900 K do not outburst from K2 30-min-cadence data First outbursting DAVs: Coolest DAVs, deepest convection zones
67. Can Rule out Some Outburst Mechanisms • Magnetic flares unlikely: τdynamical only a few s for WDs • Nuclear burning unlikely: T < 106 K at τthermal of recurrence timescale (~8 d) • Rocky accretion unlikely: No spectroscopic metal lines • Most likely related to pulsations
68. Potential Outburst Mechanisms in Cool DAVs • Possible mechanism: Nonlinear mode coupling, via parametric instability (Wu & Goldreich2001) • Mode density increases for cooler pulsating white dwarfs: more modes to couple with
69. Potential Outburst Mechanisms in Cool DAVs • Wu & Goldreich 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)
70. Potential Outburst Mechanisms in Cool DAVs • Wu & Goldreich predicted nonlinear mode coupling could transfer energy into damped modes in the cool DAVs Wu & Goldreich2001, ApJ, 546, 469 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) ωp = 897.7 µHz (l=1, m=0, n=24) ωd1 + ωd2 = ωp + δω Limit cycle if: δω < γd Frequency mismatch < Daughter mode damping rate
71. Potential Outburst Mechanisms in Cool DAVs • Wu & Goldreich predicted nonlinear mode coupling could transfer energy into damped modes in the cool DAVs ωp = 897.7 µHz (l=1, m=0, n=24) Wu & Goldreich2001, ApJ, 546, 469 ωd1 = 407.1 µHz (l=1, m=0, n=54) l=1 l=2 ωd1 + ωd2 = ωp + δω Limit cycle if: δω < γd ωd2 = 491.1 µHz (l=1, m=0, n=45) Frequency mismatch < Daughter mode damping rate 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)
72. Potential Outburst Mechanisms in Cool DAVs • Wu & Goldreich predicted nonlinear mode coupling could transfer energy into damped modes in the cool DAVs ωp = 897.7 µHz (l=1, m=0, n=24) Wu & Goldreich2001, ApJ, 546, 469 ωd1 = 407.1 µHz (l=1, m=0, n=54) l=1 l=2 ωd1 + ωd2 = ωp + δω Limit cycle if: δω < γd ωd2 = 491.1 µHz (l=1, m=0, n=45) Frequency mismatch < Daughter mode damping rate 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) 407.1 + 491.1 = 897.7 – 0.5 δω = 0.5 µHz γd > 10 µHz
73. Potential Outburst Mechanisms in Cool DAVs • Wu & Goldreich predicted nonlinear mode coupling could transfer energy into damped modes in the cool DAVs ωp = 897.7 µHz (l=1, m=0, n=24) Wu & Goldreich2001, ApJ, 546, 469 ωd1 = 407.1 µHz (l=1, m=0, n=54) l=1 l=2 ωd1 + ωd2 = ωp + δω Limit cycle if: δω < γd ωd2 = 491.1 µHz (l=1, m=0, n=45) Frequency mismatch < Daughter mode damping rate 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) 407.1 + 491.1 = 897.7 – 0.5 δω = 0.5 µHz γd > 10 µHz Limit cycle condition easily met! Best guess: Rogue waves in pulsating stars is a rapid transfer of energy to damped oscillations via parametric resonance. These modes quickly dump their energy into the surface of the star, temporarily heating it up!
74. Enough Energy in One Mode to Power an Outburst (3-day sliding window) PG 1149+057: Hermes et al. 2015 • Of order 1033-1034 erg per outburst • At least 1033 erg kinetic energy in a single mode (e.g., l=1,n=24 ωp )
75. • Outbursts occur in the coolest pulsating white dwarfs with the deepest convection zones. This is a discovery that did not happen for more than 40 years before Kepler. • Best guess: This is manifestation of parametric resonance with longer-period modes, leeching energy into the surface layers where they are rapidly damped, heating up the star. • K2 may have solved how pulsations shut down in white dwarfs. Rogue Waves: How Pulsations Cease in WDs