Lecture 16. Supernova Light Curves and Observations SN 1994D

Transcription

1 Lecture 16 Supernova Light Curves and Observations SN 1994D

2 Supernovae - Observed Characteristics See also

3 Spectroscopic classification of supernovae

4 Properties: Type Ia supernovae Classical SN Ia; no hydrogen; strong Si II 6347, 6371 line Maximum light spectrum dominated by P-Cygni features of Si II, S II, Ca II, O I, Fe II and Fe III Nebular spectrum at late times dominated by Fe II, III, Co II, III Found in all kinds of galaxies, elliptical to spiral, some (controversial) evidence for a mild association with spiral arms Prototypes 1972E (Kirshner and Kwan 1974) and SN 1981B (Branch et al 1981) Brighest kind of supernova, though briefer. Higher average velocities. M bol ~ Assumed due to an old stellar population. Favored theoretical model is (used to be?) an accreting CO white dwarf that ignites at the Chandrasekhar mass.

5 Spectra of SN Ia near maximum are very similar from event to event Spectra of three Type Ia supernovae near peak light courtesy Alex Filippenko

6 The Phillips Relation (post 1993) Broader = Brighter Can be used to compensate for the variation in observed SN Ia light curves to give a calibrated standard candle. Note that this makes the supernova luminosity at peak a function of a single parameter e.g., the width.

7 Possible Type Ia Supernovae in Our Galaxy SN D(kpc) m V Tycho Kepler Expected rate in the Milky Way Galaxy about 1 every 200 years, but dozens are found in other galaxies every year. About one SN Ia occurs per decade closer than 5 Mpc.

8 Properties: Type Ib/c supernovae Lack hydrogen, but also lack the Si II 6355 feature that typifies SN Ia. SN Ib have strong features due to He I at 5876, 6678, 7065 and A. SN Ic lack these helium features, at least the 5876 A line. Some people think there is a continuum of properties between SN Ib and SN Ic Found in spiral and irregular galaxies. Found in spiral arms and star forming regions. Not found in ellipticals. Often strong radio sources Fainter at peak than SN Ia by about 1.5 magnitudes. Otherwise similar light curve. Only supernovae definitely associated with gamma-ray bursts so far are Type Ic

9 Filippenko, (1996), Ann. Rev. Astron. Ap

10 Properties: Type II supernovae Have strong Balmer lines H, H, H - in peak light and late time spectra. Also show lines of Fe II, Na I, Ca II, and, if the supernova is discovered early enough, He I. Clearly come from massive stars. Found in star forming regions of spiral and irregular galaxies. Not found in ellipticals. Several presupernova stars identified: SN 1987A = B3 supergiant; SN 1993J = G8 supergiant (Aldering et al 1994) Fainter than Type I and highly variable in brightness (presumably depending on hydrogen envelope mass and radius and the explosion energy). Typically lower speed than Type Ia. Last longer. Come in at least two varieties (in addition to 87A) Type II-p or plateau and Type II-L or linear. There may also be Type II-b supernovae which have only a trace amount of hydrogen left on what would otherwise have been a Type Ib/c (e.g., SN 1993J) Strong radio sources and at least occasionally emit neutrino bursts

11 Typical Type II-p on the Plateau Filippenko (1990)

12 2 days after SN) SN 1987A Philipps (1987) CTIO

13

14 Summary

15 A variety of light curves are seen

16 As of SN 2000ek, 1791 supernovae had been discovered;

17 550 total (incomplete) KAIT - 47 CTRS = Catalina Real Time Transient Survey

18 May 25, 2013

19 Supernova Discovery Future Rough predictions and promises PanStarrs Dark Supernova Energy Factory Survey Joint Lick Dark observatory Energy Mission SN search Large CfA Synoptic SN group Survey Telescope Carnegie SN project ESSENCE Supernova Legacy Survey

20 Supernova Frequencies Van den Bergh and Tammann, ARAA, 29, 363 (1991) Based upon 75 supernovae. 1 SNu == one supernova per century per solar luminosities for the host galaxy in the blue band. h ~ 0.7. The Milky Way is an Sb or Sbc galaxy. Its blue luminosity is about 2 x solar luminosities, so multiply numbers by 2 h 2 ~ 1 see also Tammann et al (1994) ApJS, 92, 487.

21 Sharon et al. 2007, ApJ, 660, 1165 per unit stellar mass

22 MNRAS, 395, 1409 (2009) Volume limited sample, Ia is ¼ of the total

23 MNRAS, 395, 1409 (2009) 10.5 year volume limited (28 Mpc) search for both SNe and their progenitors Most likely range for SN II from RSGs is to Heavier stars presumably become WR stars and thus not SN IIp.

24 Tammann et al (1994) says total SN rate in MW is one every years, with 85% from massive stars. The very largest spirals produce about 10 SNae per century. Good rule of thumb - 2 core collapse supernovae per centery; one SN Ia every other century

25 Type II Supernovae Models and Physics of the Light Curve

26 ~10 45? Shock break out and cooling Typical Type II Plateau Supernova Light Curve log luminosity 43 plateau (recombination) 42 Radioactive Tail 0 few months time

27 For a typical red supergiant derived from a star over 8 solar masses. Break out temperatures of 100 s of thousands K. Very brief stage, not observed so far (indirectly in 87A). Shock heating followed by expansion and cooling Plateau the hydrogen envelope expands and cools to ~5500 K. Radiation left by the shock is released. Nearly constant luminosity (T is constant and radius of photosphere does not change much around cm). Lasts until the entire envelope recombines. Radioactive tail powered by the decay of radioactive 56 Co produced in the explosion as 56 Ni. The light curve will vary depending upon the mass of the envelope, radius of the presupernova star, energy of the explosion, degree of mixing, and mass of 56 Ni produced.

28 Shock Break-out 1. The electromagnetic display begins as the shock wave erupts through the surface of the star. A brief, hard ultra-violet (or even soft x-ray) transient ensues as a small amount of mass expands and cools very rapidly. The transient is brighter and longer for larger progenitors but hotter for smaller ones New results: Tominaga et al. 2011, ApJ, 193, 20

29 Nadyozhin, Blinnikov, Woosley, in preparation T-color will be about twice T-eff, close to 5 x 10 5 K.

30 For SN 1987A detailed calculations exist. It was little hotter because it was a BSG with 10 times smaller radius than a RSG. So far these transients have escaped detection in SN II, but the effects of break out in SN1987A were measured.

31 The effect of the uv-transient in SN 1987A was observed in the circumstellar ionization that it caused. The first spectroscopic observations of SN 1987A, made 35 hours after core collapse (t = 0 defined by the neutrino burst) Kirshner et al, ApJ, 320, 602, (1987), showed an emission temperature of 15,000 K that was already declining rapidly. Ultraviolet observations later (Fransson et al, ApJ, 336, 429 (1989) showed narrow emission lines of N III, N IV, N V, and C III (all in the Angstrom band). The ionization threshold for these species is 30 to 80 kev. Modeling (Fransson and Lundquist ApJL, 341, L59, (1989)) implied an irradiating flux with T e = 4 to 8 x 10 5 K and an ionizing fluence (> 100 ev) ~ 2 x erg. This is in good agreement with the models. The emission came from a circumstellar shell around the supernova that was ejected prior to the explosion hence the narrow lines.

32 Fransson et al (1989) IUE Observations of circumstellar material in SN 1987 A

33 x-ray UV 6.1 x erg s kev Soderberg et al 2008, Nature, 453, 469 Shock breakout in Type Ib supernova SN 2008D (serendipitous discovery while observing SN 2007uy)

34 SNLS light curve FUV = 1539 A NUV = 2316 A GALLEX discovery of ultraviolet emission just after shock break-out in a Type II-p supernova (Gezari et al 2008, ApJ, 683, L131). Missed the initial peak (about 2000 s, 90 Angstroms) but captured the 2 day uv-plateau

35 Modeling shows that they captured two supernovae, SNLS-04D2dc and SNLS-06D1jd within 7.1 hr and 13.7 hr after shock break out (in the SN rest frame) SNLS = supernova legacy survey Kepler (15 Msun) plus CMFGEN models

36 2. Envelope Recombination Over the next few days the temperature falls to ~5500 K, at which point, for the densities near the photosphere, hydrogen recombines. The recombination does not occur all at once for the entire envelope, but rather as a wave that propagates inwards in mass though initially outwards in radius. During this time R photo ~ cm. The internal energy deposited by the shock is converted almost entirely to expansion kinetic energy. R has expanded by 100 or more (depending on the initial radius of the star) and the envelope has cooled dramatically. T 1 r ρ 1 r 3 ρ T 3 ε T 4 ρ 1 r As a result only ~10 49 erg (RSG; BSG) is available to be radiated away. The remainder has gone into kinetic energy, now ~10 51 erg. As the hydrogen recombines, the free electron density decreases and es similarly declines. (note analogy to the early universe)..

37 Propagation of the Recombination Front: Since T e remains constant at about 5500 K, the recombination front moves inwards (in the co-moving frame) at an ever increasing rate. τ = τ the time scale for energy in the small cylinder to be radiated away ε ρ l p = = 4 σ Te ε = σ = at a c 4 2 Energy content in cylinder (A = 1 cm ) Rate at which energy is radiated away 4 interior ρ Since T interior decreases with time, the speed of the recombination wave accelerates (in the comoving frame) v l p τ 4 σ Te ε ρ 1 T e v 4 Tinterior 4 c

38 How much mass recombines as a function of time? Woosley, ApJ, 330, 218 (1988) dm L( t) const. = ( M, t) dt ε Where dm/dt is the rate at which matter flows through the photosphere and is its internal energy. ε ( M,t) is determined by the presupernova structure and by expansion: ε ( M,t)=ε o ( M ) t o t dm dt = L ε = L ε o t t o since r t and ε 1 r ε t o o dm = t dt L ε o t o L M(t)= t 2 2 M(t)= Lt 2 2ε o t o t 2 if L = const

39 Note : ε 1 t ε

40 H ionization - 15 solar masses. As the recombination front moves in the photospheric v declines. When the recombination reaches the He core the plateau ends

41 Photospheric speed Kasen and Woosley (2009)

42

43 The luminosity of the supernova on the plateau is approximately proportional to the initial radius of the presupernova star. The shock deposits about half of its energy in the envelope of a blue or red supergiant, but the former must expand by an additional factor of 10 before it begins to recombine at cm. Since the internal energy is proportional to 1/r, the star loses an additional factor of 10 in both luminosity and total radiate output. This is why SN 1987A was a comparatively faint supernova. Adiabatic expansion with radiation entropy dominant 3 T implies = constant ρ but ρ( m) r T and ε = at r = 4 r ρ r r

44 Popov (1993, ApJ, 414, 712) gives the following scaling relations which he derives analytically: t plateau 109days κ 1/6 1/ 0.4 M 2 1/6 10 R 500 ( ) 2/3 E 1/6 51 T ion / 4500 L bol erg s R 2/3 500 E 5/6 51 ( T ion / 4500) 4/3 1/ M 2 1/3 10 κ 0.4 not quite linear in R because a more compact progenitor recombines at a slightly smaller radius. where R 500 is the radius of the presn star in units of 500 R ( cm); κ 0.4 is the opacity in units of 0.4 cm 2 gm -1 M 10 is the mass of the hydrogen envelope (not the star) in units of 10 M E 51 is the kinetic energy of the explosion in units of erg T ion is the recombination temperature (5500 is better than 4500) In fact the correct duration of the plateau cannot be determined in a calculation that ignores radioactive energy input.

45 Cosmology on the Plateau Baade-Wesselink Method (akq The Expanding Photosphere Method (EPM) φ = L 4π D 2 = 4π R 2 4 phσt e 4π D 2 2 σ D = R ph T e φ 0 R ph = R ph 0 can ignore R ph 1/2 + v ph (t t 0 ) D = distance Measure on two or more occasions, t i Solve for D and t 0. v i, T i, and φ i In reality, the color temperature and effective temperature are not the same and that requires the solution of a model. But the physics of the plateau is well understood. This is a first principles method of getting distances. No Cepheid calibration is necessary - but it is helpful.

46 Eastman, Schmidt, and Kirshner (1996) give the dilution factor for a set of models H 0 = 73± 7 The dilution factor corrects for the difference between color temperature and effective temperature and also for the finite thickness of the photosphere which leads to a spread in T

47 Modeling the spectrum of a Type II supernova in the Hubble flow (5400 km/s) by Baron et al. (2003). SN 1993W at 28 days. The spectrum suggests low metallicity. Sedonna code.

48 3. Light Curve Tail Powered by Radioactive Decay

49 1238 kev 847 kev

50 1.2 B explosion of 15 solar mass RSG with different productions of 56 Ni indicated by the labels. Kasen et al. (2009)

51

52

53

54 Predictions from Monte Carlo

55

56

57 60 years * * Norman et al, Nuc Phys A, 621, 92 (1997) Woosley & Diehl, Phys World, 11, 22, (1998) Alberger, Harbotle, Phys Rev C, 41, 2320 (1990)

58

59 2.4 foe 1.2 foe 15 M Recent models: SN II-p 1.2 foe of kinetic energy at infinity gives good light curves in agreement with observations. 2.4 foe gives too bright a supernova making Type II almost as brilliant as Type Ia. 2.4 foe 1.2 foe 25 M Though not shown here 0.6 foe would give quite faint supernovae, usually with very weak tails.

60 33M Type II-L 15 M in a binary Type Ib 60 M Type Ic Other models that have lost most or all of their hydrogen envelope, give light curves like Type II-L or Ib/c supernovae.

61 Probable explanation Low mass envelope Large radius Small radioactivity

62 Kasen and Woosley (2009). Kinetic energy is diverse but typically ~ 1 B. Colors correspond to KE. Size of smbols to mass. Squares are low Z. Circles are solar.

63 Exceptional Cases (models)

64 Faint supernovae Stars in the ~ solar mass range are virtually impossible not to explode, but the explosion energy will be very low (Kitaura, Janka, and Hillebrandt 2006 [8.8 M sun ] Dessart et al 2006 [AIC]; Janka et al [8.8 M sun ]) Density structure, 9.5 Msun (Woosley unpublished) Energy of explosion given by (unavoidable) neutrino wind ~10 50 erg. 8 M ejected so v ~ 1000 km s 1

65 Shock breakout Typical SN IIp Light curve from 9.2 M model KE = erg (T eff 6000 K) erg s -1 Could be shortentened by mass loss.

66 Magnetar illuminated supernovae: (Woosley 2010, ApJ, 719, 204) Approximately 10% of neutron stars are born as magnetars, neutron stars with exceptionally high magnetic field strength (B ~ gauss) and possibly large rotation rates. A rotational period of 6 ms corresponds to 5 x erg. A high rotational rate may be required to make a magnetar. Where did this energy go? Moderate field strengths give a large late time energy input. In fact weaker fields can make brighter supernovae (for a given rotational energy). This reflects the competition between adiabatic energy loss and diffusion.

67 The initial explosion might not be rotationally powered but assume a highly magnetized pulsar is nevertheless created. E 51 (t = 0) = 0.2 E 51 (t) = 10 ms P 0 B 2 15 t E 51 (t = 0) L = E B 51 (t) B 2 15 t 2 at late times

68 B 15 = 0.1; P o =4.5 ms E 51 = 1 15 M erg B 15 = 0.4; P o =8.9 ms E 51 = 0.25 Typical SN IIp no magnetar power B 15 = 0.2; P o =6.3 ms E 51 = 0.5 Maeda et al 2007; Woosley 2010; Kasen and Bildsten 2010

69 By the way For radioactive energy input L = L o e -t/τ ln L = ln L o - t / τ t dln L dt = dln L dlnt = - t / τ =- 2 when t = 2τ when observed at a time 2τ = 220 days (for 56 Co) the slope of the tails for the radioactive model and the pulsar powered model are identical.

70 Magnetar Illuminated Supernovae - More massive 30 M M SN =13.9M ; M He = 10.3 M KE = erg P 0 =6 ms, erg B 15 = 0.2 B 15 = 0.5 t 2 = 0 56 Ni B 15 = 2 B 15 = 1

71 and Type Ibc The dashed line is for the decay of 0.3 solar masses of 56 Ni. Other lines are the calculated light curves for B 15 = 0.5 and 0.7 and a neutron star rotational energy of 2 and 5 x erg. 7.2 M helium core erg explosion Woosley (2010)

72

73 M α 5 less M More recent studies suggest binary evolution the most likely path and M! ~ 3-4 the most likely progenitor Broad light curve. May be faint. Few examples This is for non-rotating models.for rapid rotation and high final mass gamma-ray bursts may occur

74 f WR is a reduction factor applied to older overly large WR mass loss rates

75

76 Without mixing the outer layers of the supernova lack any strong heat source and recombine quickly. This results in the abrupt shrinkage of the photosphere to a small region of hot Ni-rich material. This makes the spectrum too hot. Mixing is also essential to produce the helium line. 56 Ni mixed into the helium emits gamma-rays that non-thermally excite the helium. It may be that the chief distinction between Type Ib and Type Ic is the degree to which He and 56 Ni are mixed.

77 A typical Type Ib supernova with strong He line:

78 A typical Type Ic supernova: Here the 56 Ni and helium were barely mixed because of the thick layer of oxygen separating them in Model 7. No helium line.

79 Recent Type Ib and Ic Simulations Lots of mixing Type Ib Dessart, Hillier, & Woosley, MNRAS, 424, 2139 (2012) It is popular (mis)conception that the production of Type Ic supernovae requires the removal of the hydrogen envelope and helium shell of a very massive star in order to have a weak helium line at 5876 A. Type Ic Some mixing The two models on the left are both derived from a 5.1 helium star that originated from a binary pair in which each star was lighter than 20 solar masses (Yoon, Langer and Woosley 2010)

80 blue is continuum red is total model