1. The Death of Massive Stars and the Birth of the Elements Stan Woosley Cornell, June 2, 2006

3. Stars are gravitationally confined thermonuclear reactors.

6. Fe Si, S, Ar, Ca O, Mg, Ne He, C H, He He

7. With each progressive burning stage the central entropy decreases. Red giant formation leads to an increased entropy in the outer hydrogen envelope. Entropy S/N A k The key role of enropy…

8. Fe O, Ne, Mg, Si He H Si O S Ne Mg C C C collapses to a neutron star

9.  bounce = 5.5 x 10 14 g cm -3 20 Solar Masses Mayle and Wilson (1988) 1 0 ms 2 0.5 3 1.0 4 3.0 5 100 ms 6 230 ms

10. 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)) Fowler, Englebrecht, and Woosley (1978) not withstanding The prompt shock model and its failure - Baron, Brown, Cooperstein and Kahana (PRL, 59, 736, (1987)) works if iron core mass is < 1.1 solar masses (but it isn’t) simplified models for the nuclear equation of state

11. Delayed neutrino-powered explosions – 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) 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..

12. Infall Accretion Shock; Ram pressure radius Velocity gain radius Neutrinosphere 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. Energy deposition here drives convection Bethe, (1990), RMP, 62, 801 (see also Burrows, Arnett, Wilson, Epstein,...) Neutron Star

14. TSI/ORNL LANL MPI Univ. Arizona

15. 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.

16. Scheck, Janka, et al (2006)

17. Scheck et al. (2004)

18. 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 Blondin, Mezzacappa, and DeMarino (2003) Buras et al. (2003) Physics Parameterized neutrino heating/cooling. Livne et al. (2004) Physics see also Foglizzo (2001,2002)

19. SASI induced flow is remarkably self similar, with an aspect ratio ~2 that is consistent with supernova spectropolarimetry data. Blondin, Mezzacappa, and DeMarino (2003)

21. 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

22. 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 ….

23. Dana Berry (Skyworks) and SEW

24. 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. For the last stable orbit around a black hole in the collapsar model (i.e., the minimum j to make a disk)

25. Heger, Woosley, & Spruit (2004) using magnetic torques as derived in Spruit (2002) Stellar evolution including approximate magnetic torques gives slow rotation for common supernova progenitors. Still faster rotation at death is possible for stars born with unusually fast rotation – Woosley & Heger (2006) Yoon & Langer (2005)

26. 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 …

27. ROTATION Ordinary SN IIp - SN Ib/c - GRB “slow” pulsar 90%(?) fast pulsar/magnetar? 10%(?) millisecond magnetar or accreting black hole (< 1%) Cas-A - Chandra GRB – Dana Berry - Skyworks

28. Nucleosynthesis in Massive Stars: Work with Alex Heger (LANL) and Rob Hoffman (LLNL)

29. 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/N A kT = 4 and Fe-core) and two explosion energies (1.2 B, 2.4 B) – 128 supernova models

30. Use the Kepler implicit hydrodynamics code (1D) Use best current information on nuclear reaction rates and opacities Include best current estimates of mass loss at all stages of the evolution (Weaver, Zimmerman, and Woosley 1978; see RMP, 74, 1015, (2002) for description of physics)

31. Use recently revised solar abundances Lodders, ApJ, 591, 1220 (2003) Asplund, Grevesse, & Sauval, ASP Conf Series, (2004)

32. Mayle and Wilson (1988) The explosion can be characterized by a piston whose location and speed are free parameters. The piston location is constrained by: Nucleosynthesis Neutron star masses The piston energy is constrained by: Light curves Fall back The Explosion Model

34. 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/N A kT = 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. Density (g cm -3 ) Entropy per baryon (S/N A kT) Density Entropy

35. (after fall back) 1.2 B explosions; mass cut at Fe core

36. 1.2 B explosions

37. Thorsett and Chakrabarty, (1999), ApJ, 512, 288 Ransom et al., Science, 307, 892, (2005) find compelling evidence for a 1.68 solar mass neutron star in Terzian 5 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): If one instead uses the S = 4 criterion, the average from 10 – 21 solar masses is From 10 to 27 solar masses the average is

38. 2.4 B 1.2 B 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”.

40. 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 over- produced. p-process deficient by a factor of ~2 for A > 130 and absent for A < 130

41. 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).

42. Two Mysteries The nature of the r-process site The origin of the p-process 90 < A < 130

43. Anti-neutrinos are "hotter" than the neutrinos, thus weak equilibrium implies an appreciable neutron excess, typically 60% neutrons, 40% protons favored at late times r-Process Site: The Neutrino-powered Wind * Duncan, Shapiro, & Wasserman (1986), ApJ, 309, 141 Woosley et al. (1994), ApJ, 433, 229 T 9 = 5 – 10 T 9 = 3 - 5 T 9 = 1 - 2 Results sensitive to the (radiation) entropy, T 3 / r, and therefore to aT 4 / r, the energy density Nucleonic wind 1 – 10 seconds

45. 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/k B about 4 times smaller s/k B ~ 80 not 300. Woosley et al (1994)

46. 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.

47. QIAN AND WOOSLEY (1996) t ~ 10 s t < ½ s data from Wilson (1994)

48. Janka, Buras, and Rampp (2003) 15 solar mass star – 20 angle averaged trajectories

49. Froehlich et al (2005) Pruett, Hoffman, and Woosley (2005) The neutrino-assisted rp-process

50. Unmodified trajectory number 6 from Janka et al.

51. Entropy times two

52. 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)