Two delays in white dwarf evolution revealed by Gaia
WWhite Dwarfs as probes of fundamental physics and tracers of plan-etary, stellar & galactic evolutionProceedings IAU Symposium No. 357, 2019A.C. Editor, B.D. Editor & C.E. Editor, eds. c (cid:13) Two delays in white dwarf evolutionrevealed by
Gaia
Sihao Cheng Department of Physics and Astronomy, The Johns Hopkins University,3400 N Charles Street, Baltimore, MD 21218, USAemail: [email protected]
Abstract.
By comparing two age indicators of high-mass white dwarfs derived from
Gaia data, twodiscoveries have been made recently: one is the existence of a cooling anomaly that produces theQ branch structure on the Hertzsprung–Russell diagram, the other is the existence of double-white-dwarf merger products. The former poses a challenge for white dwarf cooling models, andthe latter has implications on binary evolution and type-Ia supernovae.
Keywords. white dwarfs, stars: evolution, stars: kinematics, Hertzsprung-Russell diagram,methods: statistical
1. Introduction
The unprecedented astrometric and photometric power of
Gaia provides a uniqueopportunity to investigate white dwarf evolution. On the one hand, the photometricisochrone age of a large sample of white dwarfs can be accurately derived from theHertzsprung-Russell (H–R) diagram. On the other hand, their kinematic informationfrom
Gaia makes it possible to infer the true age of these white dwarfs, according tothe age–velocity-dispersion relation of the Milky Way disc. Because the photometricisochrone age is derived with a single-star evolution model and a standard white dwarfcooling model, any discrepancy between these two age indicators, or equivalently, thediscrepancy between the observed velocity distribution and the velocity distribution pre-dicted from the photoemtric ages and the age–velocity-dispersion relation, indicates de-viations from the single-star evolution or the standard cooling model. Also, by carefullydesigning the comparison, one can distinguish more than one source of the deviationbased on their different properties. In this article, I present two evolutionary delays ofwhite dwarfs revealed by this method, which include an unexpected cooling delay and amerger delay from binary evolution, and discuss related astrophysical questions.
2. Do we understand the cooling of white dwarfs?
Most white dwarfs shine at the cost of losing their thermal energy and cooling downover cosmic time. Since the seminal work of Mestel (1952), a series of detailed cool-ing models have been developed, making white dwarfs accurate cosmic clocks (see e.g.Fontaine et al. 2001 for a review).Unexpectedly, the
Gaia data (Gaia Collaboration et al. 2018) revealed an over-densityof high-mass white dwarfs on the H–R diagram, called the Q branch (Figure 1). Standardwhite dwarf models do not anticipate this feature. An explanation was proposed byTremblay et al. (2019) that the energy release from crystallisation produces this pile-upon the H–R diagram. However, Cheng et al. (2019a) found that on the Q branch, an1 a r X i v : . [ a s t r o - ph . S R ] J a n S. Cheng
Figure 1.
The H–R diagram of white dwarfs within 150 pc, colour-coded by transverse velocity.The data come from
Gaia
Data Release 2. Fast-moving high-mass white dwarfs are concentratedon the Q branch, which can only be explained by a significant slowing-down of cooling rate ina small fraction of white dwarfs. anomalously high fraction of white dwarfs have a high transverse velocity (Figure 1).According to the age–velocity-dispersion relation of the Milky Way disc (e.g., Holmberget al. 2009), these white dwarfs must be old. As argued by Cheng et al. (2019a), differentcooling behaviours are required to explain both the pile-up and the velocity distributionof high-mass white dwarfs (1.08–1.23 M (cid:12) ): while most of them cool normally as thestandard cooling model (which already includes crystallisation effect) predicts, about6% almost stop their cooling on the Q branch for as long as 8 billion years. Neithercrystallisation delay or merger delay alone cannot explain all observations. So, we calledthis delay an ‘extra cooling delay’ and the population the ‘extra delayed population’. Thisanomaly is both a challenge and an opportunity for understanding white dwarf cooling.To stop white dwarf cooling, an extra energy source is needed to provide the energyloss from shining. With a semi-analytical calculation, Cheng et al. (2019a) showed thatthe sink of Ne in C/O-core white dwarfs is a promising candidate for the extra energysource. In this explanation, the extra delayed white dwarfs are bizarre ‘sedimentars’, aspredicted and dubbed by Bildsten & Hall (2001); these white dwarfs shine out of thegravitational sedimentation of material instead of cooling.There are several aspects of this cooling anomaly and the Q branch worthy of furtherexploration.
More evidence for the extra cooling delay : Collecting more evidence for this coolinganomaly is still necessary. LP 93-21, a high-mass white dwarf just below the Q branchbut with halo kinematics (Kawka et al. 2020) may count as additional evidence. Otherevidence may come from using asteroseismology to directly measure the cooling rate (asproposed by Bart Dunlap at this symposium) in the future.
Cooling models with Ne settling : The Ne settling effect depends on the white dwarfmass, Ne abundance, core composition, crystallisation temperature, and the treatment elays in WD evolution Ne settling effect and the updated phase diagramof crystallisation for high-mass ( > . M (cid:12) ) C/O-core white dwarfs. Such models wouldtest the Ne settling hypothesis and may provide new insight into white dwarf physics.
DQ white dwarfs on the Q branch : All DQs on the Q branch belong to the extra de-layed population, whereas among the extra delayed population, about half are DQs andthe other half are DAs. Both the number density and carbon abundance suggest thatthe Q-branch DQs are evolutionary descendants of the hot-DQs discovered by Dufour etal. (2007). Further questions include the origin of the distinction between DQs and DAsand whether the Q-branch DQs also have magnetic field as the hot-DQs.
Merger origin : Three arguments support a double-white-dwarf (double-WD) mergerorigin of the extra delayed white dwarfs: the Ne settling hypothesis for the delay (be-cause it requires C/O-core), the possible relation to hot-DQs, and the lack of wide binarycompanions. Nevertheless, further definitive evidence is still needed.
Core composition : We expect that the extra delayed white dwarfs have C/O coresinstead of O/Ne cores that is expected for normal high-mass white dwarfs. Future obser-vations that can determine the core composition will be crucial to testing both the Nesettling hypothesis and their merger origin.
3. What are the origins of high-mass white dwarfs?
High-mass white dwarfs can form from single-star evolution or from merger of stars. Inparticular, a considerable fraction of high-mass white dwarfs are expected to be double-WD merger products (e.g., Temmink et al. 2019). However, despite discussions in theliterature, there is still a lack of clear evidence for the white dwarfs that are merger prod-ucts. The difficulty is to reliably distinguish the merger population from singly evolvedwhite dwarfs.The astrometric information from
Gaia has improved the situation significantly. Be-cause double-WD merger products experience additional binary evolution, their trueages are younger than their photometric isochrone ages. To get rid of the influence ofthe cooling anomaly described in the previous section, Cheng et al. (2019b) selected onlyhigh-mass white dwarfs above the Q branch and used their transverse velocity distribu-tion to infer the fraction of merger products. Thanks to the much larger sample size thanwhat Wegg & Phinney (2012) used in a previous study with similar method, Cheng etal. (2019b) found clear velocity excess that corresponds to about 20 ±
6% of high-masswhite dwarfs (0.8–1.3 M (cid:12) ) being double-WD merger products. This is strong evidencefor the merger origin of high-mass white dwarfs.
4. How frequently do two white dwarfs merge?
The fraction of double-WD merger products can also be translated into a double-WD merger rate in some mass range. Then, it can be compared with binary populationsynthesis results and other observations to improve our understanding of binary evolution,and it can also be compared with the type-Ia supernovae rate as double-WD merger isone of the promising scenarios of type-Ia supernovae.Here, I only discuss the double-WD mergers occurring in a close binary system. Ingeneral, there are three ways to measure the double-WD merger rate: to extrapolate frompre-merger binary systems, to count merger events, and to search for merger products.Previous works were mainly focused on extrapolating the orbital distribution of pre-merger systems (e.g., Brown et al. 2016, Maoz et al. 2018), which currently provides S. Cheng
Figure 2.
Estimates of the double-WD merger rate as a function of total mass, from bothobservational and theoretical sides. This figure is reproduced from the Figure 4 in Cheng et al.(2019b). no or low mass resolution and large uncertainties. The merger rate translated from thefraction of merger products in Cheng et al. (2019b) adds significant mass resolution andprecision to previous measurements, and it is shown to be close to binary simulationresults (Figure 2).If it is allowed to extrapolate the observed merger rate to a higher mass range using thesimulated mass distribution, one will conclude that the rate of mergers with total mass > . M (cid:12) is a third to a half of the type-Ia supernova rate measured for Milky-Way-likegalaxies (Li et al. 2011). The fact that the two rates are close to each other supportsthe idea that massive double-WD mergers may contribute to a significant fraction oftype-Ia supernovae, whereas the fact that the merger rate is lower than the supernovarate indicates other channels of type-Ia supernovae, including double-WD mergers withlower total mass and/or single degenerate progenitors. References
Bildsten, L., & Hall, D. M. 2001,
Ap. Lett. , 549, L219Brown, W. R., Kilic, M., Kenyon, S. J., & Gianninas, A. 2016,
ApJ , 824, 46Cheng, S., Cummings, J., & M´enard, B. 2019a,
ApJ , 886, 100Cheng, S., Cummings, J., M´enard, B., & Toonen, S. 2019b, arXiv:1910.09558Dufour, P., Liebert, J., Fontaine, G., & Behara, N. 2007,
Nature , 450, 522Fontaine, G., Brassard, P., & Bergeron, P. 2001,
PASP , 113, 409Gaia Collaboration, Babusiaux, C., van Leeuwen, F., et al. 2018,
A&A , 616, A10Holmberg, J., Nordstr¨om, B., & Andersen, J. 2009,
A&A , 501, 941Kawka, A., Vennes, S., & Ferrario, L. 2019,
MNRAS
Letters, 491, L40Li, W., Chornock, R., Leaman, J., et al. 2011,
MNRAS , 412, 1473Maoz, D., Hallakoun, N., & Badenes, C. 2018,
MNRAS , 476, 2584Mestel, L. 1952,
MNRAS , 112, 583Temmink, K. D., Toonen, S., Zapartas, E., Justham, S., & G¨ansicke, B. T. 2019,arXiv:1910.05335Tremblay, P.-E., Fontaine, G., Fusillo, N. P. G., et al. 2019,
Nature , 565, 202Wegg, C., & Phinney, E. S. 2012,