Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Formation and evolution of the Solar System wikipedia , lookup
Dyson sphere wikipedia , lookup
Aquarius (constellation) wikipedia , lookup
First observation of gravitational waves wikipedia , lookup
Negative mass wikipedia , lookup
Advanced Composition Explorer wikipedia , lookup
Nucleosynthesis wikipedia , lookup
Stellar evolution wikipedia , lookup
Star formation wikipedia , lookup
The Death of Massive Stars and the Birth of the Elements Stan Woosley Cornell, June 2, 2006 Stars are gravitationally confined thermonuclear reactors. T3 25 M Presupernova Star 900 R H, He O, Mg, Ne 1R He Si, S, Ar, Ca Fe He, C 0.1 R 0.01 R The key role of enropy… Entropy S/NAk With each progressive burning stage the central entropy decreases. Red giant formation leads to an increased entropy in the outer hydrogen envelope. Si O O, Ne, Mg, Si Fe S Ne H C He Mg C C collapses to a neutron star 20 Solar Masses Mayle and Wilson (1988) bounce = 5.5 x 1014 g cm-3 1 2 3 4 5 6 0 ms 0.5 1.0 3.0 100 ms 230 ms Some of the contributions of Hans Bethe to our understanding of core collapse supernovae: • The idea of a cold, low entropy bounce at nuclear density – Bethe, Brown, Applegate and Lattimer (Nucl. Phys. A , 324, 487, (1979)) G exp(akT ) E (akT ) 2 Fowler, Englebrecht, and Woosley (1978) not withstanding a ~A / 9 • The prompt shock model and its failure - Baron, Brown, Cooperstein works if iron core mass is < 1.1 solar masses (but it isn’t) and Kahana (PRL, 59, 736, (1987)) simplified models for the nuclear equation of state • Delayed neutrino-powered explosions – Bethe and Wilson, ApJ, 295, 14, (1985) - see also Wilson (1982) Wilson discovered in 1982 that neutrino energy deposition on a time scale longer than previously suspected could re-energize the shock. Hans joined with Jim in the first refereed publication to state this.. • Analytic Models for Supernovae. “entropy” “gain radius” “net ram” “foe” – now the “Bethe” = B Bethe, RMP, 62, 801, (1990) ApJ, 412, 192, (1993) ApJ, 419, 197, (1993) ApJ, 449, 714 (1995) ApJ, 469, 192 (1996) ApJ, 473, 343 (1996) Nuc. Phys A, 606, 195 (1996) ApJ, 490, 765, (1997) Energy deposition here drives convection Bethe, (1990), RMP, 62, 801 Velocity Neutrinosphere Neutron Star (see also Burrows, Arnett, Wilson, Epstein, ...) gain radius radius 3000 km s 1 Infall Accretion Shock; Ram pressure Inside the shock, matter is in approximate hydrostatic equilibrium. Inside the gain radius there is net energy loss to neutrinos. Outside there is net energy gain from neutrino deposition. At any one time there is about 0.1 solar masses in the gain region absorbing a few percent of the neutrino luminosity. TSI/ORNL LANL Univ. Arizona MPI No one has yet done a 3-dimensional simulation of the full stellar core including neutrino transport that the community would agree is “good”. But they are getting there and the attempts are providing insights. Scheck, Janka, et al (2006) Scheck et al. (2004) Stationary Accretion Shock Instability (SASI) Supernova shock wave will become unstable. Instability will 1. help drive explosion, 2. lead to gross asphericities. New ingredient in the explosion mechanism. Confirmed by: Scheck et al. 2004 Janka et al. 2005 Ohnishi et al. 2006 Burrows et al. 2006 Buras et al. (2003) Physics Livne et al. (2004) Physics Blondin, Mezzacappa, and DeMarino (2003) Parameterized neutrino heating/cooling. see also Foglizzo (2001,2002) Blondin, Mezzacappa, and DeMarino (2003) SASI induced flow is remarkably self similar, with an aspect ratio ~2 that is consistent with supernova spectropolarimetry data. Burrows et al. (2006) find considerable energy input from neutron star vibrations – enough even to explode the star, and surely enough to influence the r-process And so – maybe – most massive stars blow up the way Hans and others talked about: Rotation and magnetic fields unimportant in the explosion (but might be important after an explosion is launched) Kicks and polarization from “spontaenous symmetry breaking” in conditions that started spherical. Just need better codes on bigger faster computers to see it all work. But …. Dana Berry (Skyworks) and SEW Need iron core rotation at death to correspond to a pulsar of < 5 ms period if rotation and B-fields are to matter at all. Need a period of ~ 1 ms to make GRBs. This is much faster than observed in common pulsars. Total rotational kinetic energy for a neutron star E rot ~ 2 1052 (1 ms/P)2 (R/10 km)2 erg j R 2 ~ 6.31015 (1ms/P) (R/10 km)2 cm2 s-1 at M 1.4 M For the last stable orbit around a black hole in the collapsar model (i.e., the minimum j to make a disk) jLSO 2 3 GM / c 4.6 1016 M BH / 3 M cm 2 s -1 non-rotating jLSO 2 / 3 GM / c 1.5 1016 M BH / 3M cm 2 s -1 Kerr a = 1 Stellar evolution including approximate magnetic torques gives slow rotation for common supernova progenitors. Heger, Woosley, & Spruit (2004) using magnetic torques as derived in Spruit (2002) Still faster rotation at death is possible for stars born with unusually fast rotation – Woosley & Heger (2006) Yoon & Langer (2005) The spin rates calculated for the lighter (more common) supernovae are consistent with what is estimated for young pulsars So, one could put together a consistent picture … GRB – Dana Berry - Skyworks Cas-A - Chandra “slow” pulsar 90%(?) fast pulsar/magnetar? 10%(?) millisecond magnetar or accreting black hole (< 1%) ROTATION Ordinary SN IIp - SN Ib/c - GRB Nucleosynthesis in Massive Stars: Work with Alex Heger (LANL) and Rob Hoffman (LLNL) Survey - Solar metallicity: • Composition – Lodders (2003); Asplund, Grevesse, & Sauval (2004) • 32 stars of mass 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33 35, 40, 45, 50, 55, 60, 70, 80, 100, 120 solar masses. More to follow. • Evolved from main sequence through explosion with two choices of mass cut (S/NAkT = 4 and Fe-core) and two explosion energies (1.2 B, 2.4 B) – 128 supernova models • Use the Kepler implicit hydrodynamics code (1D) (Weaver, Zimmerman, and Woosley 1978; see RMP, 74, 1015, (2002) for description of physics) • Use best current information on nuclear reaction rates and opacities • Include best current estimates of mass loss at all stages of the evolution • Use recently revised solar abundances Lodders, ApJ, 591, 1220 (2003) Asplund, Grevesse, & Sauval, ASP Conf Series, (2004) • The Explosion Model The explosion can be characterized by a piston whose location and speed are free parameters. Mayle and Wilson (1988) The piston location is constrained by: • Nucleosynthesis • Neutron star masses 15 M Electron mole number Ye The piston energy is constrained by: • Light curves • Fall back Except near the "mass cut", the shock temperature to which the explosive nucleosynthesis is most sensitive is given very well by 4 R3 aT 4 Explosion energy 3 1051 erg 1B Density 20 M presupernova Entropy Density (g cm-3) Entropy per baryon (S/NAkT) The edge of the iron core sets a lower bound to the mass cut. Otherwise, too many neutron-rich isotopes … The location where the entropy S/NAkT = 4, typically at the base of the oxygen shell sets an upper limit. Stars that explode in real simulations typically develop their mass cut here. A larger value gives neutron stars that are too massive. 1.2 B explosions; mass cut at Fe core (after fall back) 1.2 B explosions Above 35 M black holes form in Z=0 stars Thorsett and Chakrabarty, (1999), ApJ, 512, 288 If in the models the mass cut is taken at the edge of the iron core the average gravitational mass for for stars in the 10 – 21 solar mass range is (12 models; above this black holes start to form by fall back): 1.38 0.16 M Ransom et al., Science, 307, 892, (2005) find compelling evidence for a 1.68 solar mass neutron star in Terzian 5 If one instead uses the S = 4 criterion, the average from 10 – 21 solar masses is 1.45 0.18M From 10 to 27 solar masses the average is Vertical line is at 1.35 0.04 M 1.53 0.22 M 15 M 2.4 B 1.2 B 25 M 2.4 B 1.2 B 1.2 B of kinetic energy at infinity gives good light curves in agreement with observations. 2.4 B gives too bright a supernova making Type II almost as brilliant as Type Ia. Though not shown here 0.6 B would give quite faint supernovae, usually with very weak “tails”. Isotopic yields for 31 stars averaged over a Salpeter IMF, G = -1.35 Intermediate mass elements (23< A < 60) and s-process (A = 60 – 90) well produced. Carbon and Oxygen overproduced. p-process deficient by a factor of ~2 for A > 130 and absent for A < 130 Conclusions • Overall good agreement with solar abundances – see also WW95. • Lightest neutron star 1.16 solar masses; average 1.4 solar masses. Black holes a likely product for some current generation stars in the 30 – 50 solar mass range (more black holes at metallicities lower than the sun) • Overproduction of C and O suggests that current estimates of Wolf-Rayet mass loss rates may be too large (and/or Lodders (2003) abundances for C and O too small). Two Mysteries • The nature of the r-process site • The origin of the p-process 90 < A < 130 r-Process Site: The Neutrino-powered Wind * Nucleonic wind 1 – 10 seconds Anti-neutrinos are "hotter" than the neutrinos, thus weak equilibrium implies an appreciable neutron excess, typically 60% neutrons, 40% protons favored at late times T9 = 5 – 10 T9 = 3 - 5 T9 = 1 - 2 Results sensitive to the (radiation) entropy, T3/, and therefore to aT4/, the energy density Duncan, Shapiro, & Wasserman (1986), ApJ, 309, 141 Woosley et al. (1994), ApJ, 433, 229 The r - process is favored by high entropy , i.e., low density at a given temperature, because the reactions that assemble - particles to heavy (seed ) nuclei between 3 and 5 billion K increase rapidly with density ( n, ) 7 Be( , n)12 C 2 Keeping the density low thus keeps most of the mass in - particles and thus increases the ratio of free neutrons to heavy seed . Woosley et al (1994) Integrated abundances in the late time wind resemble the r-process abundance pattern. But, The entropy in these calculations by Wilson was not replicated in subsequent analyses which gave s/kB about 4 times smaller s/kB ~ 80 not 300. Throughout nucleosynthesis epoch L L Early on (t about 1/2 second) Late (t >1 s) Wilson (1994) With nearly equal fluxes of neutrinos, each having about the same energy, and with the lifting of degeneracy the neutron-proton mass difference favors protons in weak equilibrium. Later, the neutrino energy difference favors neutrons. QIAN AND WOOSLEY (1996) t ~ 10 s t<½s data from Wilson (1994) Janka, Buras, and Rampp (2003) 15 solar mass star – 20 angle averaged trajectories The neutrino-assisted rp-process 0.36 s Froehlich et al (2005) Pruett, Hoffman, and Woosley (2005) T9 2.05 5 0.27 s / kb 77 Ye 0.562 X p 0.124 X 0.844 X n 1013 Unmodified trajectory number 6 from Janka et al. Entropy times two Summary – Neutrino Wind • Very rich site for nucleosynthesis – about half of all the isotopes in nature are made here • Can produce both (part of) the p-process and the r-process nearly simultaneously in one site. • In both cases, the nucleosynthesis suggests a higher entropy (and outflow with more internal energy per baryon) than traditional models have provided •Nuclear physics very uncertain – need RIA • This has important implications for the explosion model Energy input from: (Qian & Woosley 1996) vibrations - Burrows et al (2006) Alfven waves or reconnection – Suzuki and Nagataki (2005) magnetic confinement – Thompson (2003)