![Arxiv:1210.7240V1 [Astro-Ph.HE] 26 Oct 2012 H Nlgtcre H Nryiptms Cu Trel- Efficient](http://data.docslib.org/img/b0b54d1bea2d79010eb11703fd13f09c-1.webp)
arXiv:1210.7240v1 [astro-ph.HE] 26 Oct 2012 h Nlgtcre h nryiptms cu trel- efficient. dif- is at radiative ejecta when occur the months) must through to input fusion (weeks modify energy times significantly the late curve, to atively For light mechanisms SN two KB10). the these hereafter core the of 2010, power in Bildsten either spindown formed & which the (Kasen magnetar by the collapse rotating Alternatively, re-energized rapidly a be of 2012). may Gal-Yam 2011; ejecta Irwin 2012; SN & shock al. Chevalier inter- ma- et 2007; supernova Moriya McCray the circumstellar & the by with (Smith re-thermalizing part, energy ejecta Overluminous in effectively supernova least terial, the at of powered, categories. are action lines explana- be absorption main to hydrogen Alternative narrow thought two with SNe into which IIn 2009). (e.g., Type to for curve al. fall difficult et light models Gezari tions are optical standard 2008es, the events of SN powers these context decay of the radioactive Many in Gal-Yam 2011; explain 2010) (e.g., al. al. events. et et sub-luminous Quimby Kasliwal 2012) 2009; (e.g., al. including super-luminous et Foley and 2009; al. (SNe), et Perets supernovae of [email protected] crto noacnrlcmatrmatrepre- remnant compact central a onto Accretion classes new discovered have surveys optical Ongoing rpittpstuigL using 2012 typeset 30, Preprint October version Draft uainTp Ieet.Ti cnromyb eeatfrexplainin for supernovae. headings: relevant rare Subject be and may peculiar scenario decrease of sign This sudden classes events. experienced a have II causing may Type material, infalling duration that remaining stars the compact unbind may from events I Type fpoeio assadrdiadepoineege.Teaccret The as energies. decreases explosion hole and black radii or and star accre masses the othe estimate progenitor or to of super-luminous models producing hydrodynamical cases symmetric some spherically in curve, light optical otecmatrmat n vnulytr rudadfl back fall and around turn eventually and remnant, ( compact the to hnpeiu alaksproamdl hc goe h nu fa of ar ac input fallback curves when the light outcomes ( ignored possible resulting super-luminous The which The transients. models dim supernova efficiently. short-lived, de fallback out supernova previous the diffuse than str ejecta, can s a outgoing resulting photons if the the when or in If thermalizes stars, progenitor. which massive the outflow very in in interface energies, helium/hydrogen the explosion low at enhanced & oefato ftemtra jce nacr olpesupernova collapse core a in ejected material the of fraction Some as oe soitdwt h crto fti flbc”mater “fallback” this of accretion the with associated power days) ula cec iiin arneBree ainlLabo National Berkeley Lawrence Division, Science Nuclear 1. eatet fPyisadAtooy nvriyo Califo of University Astronomy, and Physics of Departments ACCRETION FALLBACK BY POWERED CURVES LIGHT SUPERNOVA A eatet fPyisadAtooy nvriyo Califo of University Astronomy, and Physics of Departments INTRODUCTION T E tl mltajv 8/13/10 v. emulateapj style X & uenve eea uenve niiul(N20e,S 19 SN 2008es, (SN individual supernovae: — general supernovae: tr:msie—sas eto crto,aceindss—b — disks accretion accretion, — neutron stars: — massive physics stars: — 10 44 rss ergs − 1 yeI vnso ohsotadln uain,a ela luminou as well as durations, long and short both of events II Type ) M ˙ ∝ rf eso coe 0 2012 30, October version Draft t − 5 / 3 ao Dexter Jason ailKasen Daniel ABSTRACT tlt ie,btisnraiaincnb significantly be can normalization its but times, late at eaiitcjt h iecl soitdwt h energy the with associated timescale here- a The drives 2012, envelope jet. and Kasen relativistic supported hydrogen man- rotationally & the the becomes Quataert QK12) either or after 2011; gamma if Heger 2001) & possible longer al. (Woosley et be Much (MacFadyen also tle may 1993; transients 1999). ray Woosley ray mechanism, gamma Woosley ac- collapsar & long-duration MacFadyen the even (the a where supported, (GRB) power case burst rotationally may opposite becomes energy the cretion core disappearing In iron stars hence the i.e., view. and 2008), from “un- disk are al. suddenly events et suf- a these (Kochanek lacks energy, form novae” progenitor accretion available to the the If momentum tap hole. angular black ficient a remnant, tral 1994; sup- particu- Yi rotationally & 1999). is (Narayan is Begelman This inefficient & flow Blandford radiatively accretion and material. the ported infalling when the true larly of energy ing wya uhas much e.g., associated carry as (microquasars, can is outflows holes away these al. and black Rodr´ıguez 1998), et accretion & mass (Fender ex- Mirabel stellar stars supernova object neutron and failed in 2004) Compact outflows or large-scale successful with either in plosions. amounts large energy injecting of of means potential another sents n“ald N,teetr tracee notecen- the onto accretes star entire the SNe, “failed” In aoy ylto od ekly A USA CA, Berkeley, Road, Cyclotron 1 ratory, na ekly A970 USA 94720, CA Berkeley, rnia, na ekly A S and USA CA, Berkeley, rnia, prEdntnaceindie an drives accretion uper-Eddington fiatms os crto power Accretion loss. mass ificant ws euirsproa.W use We supernovae. peculiar rwise rto oe ssgicn include significant is power cretion rswl er-nrie tatime a at re-energized be will bris inrt tlt ie o range a for times late at rate tion oeo h eetydiscovered recently the of some g o aeot h proto-neutron the onto rate ion n ees hc aefrsat forms wave shock reverse ong eso httelt time late the that show We . a a infiatyaetthe affect significantly may ial xlso a eanbound remain may explosion ceinpwradproduced and power ccretion ntebihns fsm long some of brightness the in ieetadmr diverse more and different e ∼ 0 ftegaiainlbind- gravitational the of 10% 8w N2010X) SN 98bw, akhole lack s 2 Dexter & Kasen injection corresponds roughly to the free-fall time of a At both low and high accretion rates compared to 2 stellar layer, about 0.1 s for the ironcore, but aslong as Ledd/c , where Ledd is the Eddington luminosity, ac- 1 yr for the hydrogen∼ envelope of a red giant. Powerful cretion flows onto compact objects become hot and winds∼ from the accretion disk may eventually provide suf- geometrically thick due to their inability to cool effi- ficient energy to turn the failed SN into a successful one, ciently (tcool > tinfall). Such radiatively inefficient ac- exploding the remainder of the star (Milosavljevic et al. cretion flows are expected to produce large-scale out- 2010; Lindner et al. 2011). flows (Narayan & Yi 1994; Blandford & Begelman 1999; In successful SNe, accretion from the “fallback” of the Igumenshchev & Abramowicz 2000; Pen et al. 2003; fraction of material remaing bound can be significant Begelman 2012; McKinney et al. 2012) and/or Poynting as well. The fallback may influence the resulting nu- flux dominated jets (De Villiers et al. 2005; McKinney cleosynthesis (Colgate 1971) or delay the pulsar mecha- 2006). This behavior is observed in the accretion flow nism in a young proto-neutron star such as in SN 1987A onto the Galactic center black hole, Sagittarius A*, (Michel 1988). Early time fallback may also provide where the accretion rate at the Bondi radius (e.g., a link between the explosion mechanism and the rem- Baganoff et al. 2003; Quataert 2004) is several orders of nant mass distribution (Fryer et al. 2012) or alter the magnitude larger than that onto the black hole (e.g., radiated neutrino spectrum (Fryer 2009). For red su- Marrone et al. 2007). The fallback accretion rate fol- pergiant (RSG) progenitors with typical explosion ener- lowing a successful supernova explosion is highly super- gies ( 1051erg), the fallback mass is relatively small Eddington and extremely optically thick to photons. For ( 0.≃1 M⊙). However, in more compact stars (e.g., all timescales of interest here (& 1000s after the explo- blue∼ supergiants, BSGs) the formation of a strong reverse sion), the disk is not dense enough to cool by neutrino shock at the H/He interface can decelerate the ejecta and emission (Kohri et al. 2005). We expect then that it enhance the fallback mass to 2M⊙ (Chevalier 1989; should be radiatively inefficient, geometrically thick, and Zhang et al. 2008). For weak∼ explosions, most of the should drive large-scale outflows. star may fall back, with only a small fraction of the mass The resulting mass outflow rate can be estimated fol- ejected in a dim SN (Moriya et al. 2010). lowing Kohri et al. (2005) by assuming that the accretion While the dynamics of supernova fallback have been rate increases as some power of radius studied in a various contexts, little has been said about s how fallback may impact the optical SN light curve. The r M˙ (r)= M˙ fb , (1) energy released from fallback accretion may profoundly rfb affect what we observe, if two conditions are met. First, the accretion energy must be injected at relatively late where M˙ fb and rfb are the accretion rate and radius at times (& days) otherwise it will be largely degraded by the outer disk edge, and 0
-2 1019 10 )
-3 18 10 -3 10 1017 -4 1016 / s) 10 sun 15 10 10-5 14 10 t = 100 s t = 300 s -6 1013 10
Pressure (ergs cm t = 800 s 1012 10-7 100.0
Accretion Rate (M S33 ) 3 10-8 U45 10.0 Z29 10-9 102 103 104 105 106 1.0 Time (s)
Density (g cm 0.1 Figure 2. Numerical (lines) and semi-analytic (points) fallback accretion rates for three models from Table 1. For the small pro- 4·108 genitor in S33, the asymptotic scaling M˙ ∝ t−5/3 applies after the 8 3·10 first ≃ 100s. At much lower explosion energies (U45), the entire 2·108 accretion rate curve is well described by freefall. In BSG progeni- tors (Z29), the accretion rate can be significantly enhanced at late 1·108 times by material re-captured by the reverse shock formed at the H/He interface. 0
8 Velocity (cm / s) -1·10 for de-pressurized models with analytic freefall solutions -2·108 (Eq. 5). 109 1010 1011 1012 The initial conditions are taken from a wide r (cm) range of pre-supernova progenitor star models from Woosley et al. (2002)1. Three sets of models are consid- Figure 1. Pressure (top), density (middle), and velocity (bottom) ered: zero and solar metallicity progenitors with ZAMS vs. radius at t = 100s, 300s, and 800s for an E = 1.2 × 1051 erg −4 0 masses of 11 40M⊙, and 10 Z⊙ progenitors with explosion of a 25M⊙ zero metallicity progenitor using the fallback − (lines) and piston (points) inner boundary conditions. The two ZAMS masses of 11 60M⊙. The lower mass solar metal- methods are in good agreement away from the inner boundary, licity progenitors retained− their hydrogen envelopes and and the results are similar to Zhang et al. (2008) Figure 1. The tended to be red supergiants (R 1014cm), while those points have been down-sampled by a factor of 5 for clarity. in the high mass range were bare∼ helium or C/O stars ˙ 2 ǫMfbc would be nearly identical to the case of a disk (R 1011cm). Low metallicity stars tended to be blue wind. This is the scenario discussed for failed supernova supergiants∼ of smaller radii (R 1012−13cm). Due to explosions by Woosley & Heger (2011) and QK12. The large uncertainties in prescriptions∼ for semi-convection, results presented below only depend on the energy injec- convective overshoot, and mixing, we view the very low tion rate and thus are the same for either a disk or a jet. metallicity models as alternative outcomes for possible The results are also expected to be insensitive to whether massive star progenitors rather than necessarily corre- the central object is a proto neutron star or black hole. sponding to extremely metal-poor environments. In 5.1 we discuss the dissipation of accretion energy in § Explosions are simulated using a moving inner bound- the infalling material and outgoing ejecta, and its impli- ary (“piston,” e.g. Woosley & Weaver 1995). For the cations for the viability of wind and jet scenarios. −3 first 0.45s the boundary moves inwards from the loca- We use ǫ = 10 throughout, although we discuss disk tion where the specific entropy s =4 to r = 5 107cm, formation and size in 5.2. The outflow energy, either × § after which time it moves outwards at constant velocity. from a wind or jet, is then set by the fallback accretion The inner boundary velocity is set to zero either after a rate. specified amount of time or after the internal energy has 3. FALLBACK ACCRETION changed by the desired amount. The resulting explosions are fairly insensitive to the piston velocity as long as it is 3.1. Numerical Hydrodynamics large enough to deposit the desired amount of energy in We estimate the fallback accretion rate by simulating a few seconds. We use the same number of radial zones supernova explosions using a 1D lagrangian finite differ- for the hydrodynamics calculations as were used for the ence hydrodynamics code. The code uses a staggered stellar evolution ( 500 1000), but verify that doubling mesh and artificial viscosity shock prescription (Castor the number of zones≃ and− interpolating using the near- 2004). The artificial viscosity parameter is chosen to est neighbor from the initial condition does not lead to smooth shocks over 7 zones. This is fairly diffusive, significant changes in the evolution. helping with code stability≃ but gives nearly identical re- An outflow boundary condition is employed by copy- sults to much smaller coefficients. The Courant factor ing the acceleration from the outer zone to a single ghost used is ∆x/cs∆t = 0.5, and the time step is set by zone. The inner boundary condition can be either a hard the minimum required by any zone. The hydro code (reflective) piston or inflow, depending on the time. Ini- has been verified via comparisons to the Sedov-Taylor tially, the piston is used to blow up the star, after which and 1D shock tube problems, by verifying that pre- the inner boundary velocity is set to zero. To allow in- supernova stellar models with no explosions remain in hydrostatic equilibrium, and by comparing the solutions 1 http://homepages.spa.umn.edu/ alex/stellarevolution/data.shtml 4 Dexter & Kasen
1045 ) -1 1044
1043
42 10 R < 10 Rsun
Peak Luminosity (ergs s 10 Rsun < R < 30 Rsun
30 Rsun < R < 300 Rsun
R > 300 Rsun E < 1051 ergs 1041 E > 1051 ergs
10 100 Time to Peak (days)
Figure 3. Peak luminosity vs. time to peak for all events, measured from light curves calculated with the methods in Appendix A. The points are only shown when the peak luminosity is larger than the standard thermal supernova luminosity (Eq. 9). Luminous, long duration Type II events come from high energy explosions in massive BSGs, some with strong reverse shocks. Superluminous events with durations ≃ 2 − 40 days come from weak explosions with low ejecta masses. Luminous Type I events are from weak explosions in compact initial stars.
Table 1 Sample Event Parameters
51 Name MZAMS (M⊙) MSN (M⊙) Z (Z⊙) Eexp (10 ergs) Mej (M⊙) Mrem (M⊙) vej (km / s) vf (km / s) Comparison S33 33.0 11.4 1 0.34 2.1 9.2 2800 24000 SN 1998bw S39 39.0 8.49 1 0.21 1.1 7.4 3000 17000 SN 2008D S39O 39.0 8.49 1 0.055 0.39 8.1 2600 11000 SN 2010X U45 45 44.7 10−4 1.0 × 10−3 0.45 44 340 13000 SN 2008es U60 60 59.2 10−4 1.0 43 17 1100 5500 — Z29 29 28.8 0 1.1 22 6.7 1600 4700 —
accretion rate once material begins to fall back. When 1045 the inner zone passes through the radius corresponding 8 to the assumed outer disk edge, rmin = 10 cm, its prop- )
-1 44 erties are saved and it is removed from the calculation. 10 The outside of the accreted zone then becomes the inner boundary for the subsequent evolution. 51 1043 We show in Figure 1 an example of the 1.2 10 erg explosion of a 25M⊙, zero metallicity stellar progenitor.× In this star, a significant density discontinuity at the he- 1042 10 R < 10 Rsun
Peak Luminosity (ergs s lium/hydrogen interface (r 2 10 cm) as well as the 10 R < R < 30 R sun sun ≃ × −2.5 30 Rsun < R < 300 Rsun compact hydrogen envelope (ρ r ) lead to a strong R > 300 Rsun 51 ∝ 41 E < 10 ergs reverse shock forming at t 20 s. The two-shock struc- 10 51 E > 10 ergs ture can be seen clearly in≃ the curves of ρ(r) and v(r). 108 109 Ejecta Velocity (cm / s) The filled circles show the results using a pure piston inner boundary condition, while the solid lines show the v Figure 4. Peak luminosity vs. f for the events in Figure 3. results for the “fallback” boundary condition (i.e., piston The peak luminosity scales roughly with final ejecta velocity as L ∝ v2 . switched to inflow after the shock was initiated). The two p f are in excellent agreement in the portion of the star with v> 0, but differ slightly in the inner regions, as pressure flow, once the velocity of the inner zone drops below zero, support slows the infall in the pure piston model. In the it is copied to the inner boundary. This allows us to use fallback calculation, the reverse shock turns around and the inner boundary to blow up the star and to record the Accretion Powered Supernova Light Curves 5 leads to a jump in the accretion rate at t 2000 s. This formulae in Matzner & McKee (1999), which are typi- calculation is very similar to that shown≃ in Figure 1 of cally an excellent approximation to the numerical calcu- Zhang et al. (2008), and the solutions are in qualitative lations. Then the total fallback time for each mass ele- agreement. The quantitative differences are likely due to ment can be calculated from Eq. 3.7 of Chevalier (1989), a difference in progenitor models. and its time derivative is an approximate accretion rate. We follow the explosions until late time (t = 108 s), and This assumes that pressure effects are negligible, which is calculate the accretion rate through the inner boundary incorrect. However, the true acceleration measured from from the properties of accreted zones. Sample accretion the numerical calculations described below turns out to rate curves are shown in Figure 2. In some cases, we be roughly constant at half of the gravitational acceler- −6 −1 find large accretion rates (& 10 M⊙ s ) for a week ation. or so after the explosion. This late time accretion is due This ballistic estimate reproduces the fallback accre- either to the fallback of stellar layers at large radii, or tion rate at all times in many progenitors. However, from the deceleration of inner layers by the reverse shock. in some cases (particularly blue supergiants such as The energy associated with accretion at these rates is SN1987A, Chevalier 1989) the reverse shock formed at sufficient to power luminous supernova light curves. the hydrogen-helium interface is strong enough to decel- erate portions of the ejecta below the escape speed. This 3.2. Semi-Analytic Treatment enhances the accretion rate at late times, and can sig- The general behavior of the fallback accretion rate can nificantly add to the remnant mass (Zhang et al. 2008). be easily understood in the two limits where the ma- The reverse shock formation and evolution is analagous terial is either highly bound or mildly bound. For the to that formed when the forward shock breaks out of highly bound material (i.e., those layers where the ve- the star and into the interstellar medium (e.g., McKee locity following the shock propagation is much less than 1974; Chevalier 1982). As the simplest possible reverse the escape velocity) the accretion rate can be estimated shock prescription, we solve the strong shock jump con- from the free-fall time, ditions for the reverse shock velocity and the downstream velocity at the boundary of 100% helium and hydrogen 3/2 layers: v 0.6v , where v is the shock velocity. The πr RS ≃ 0 0 tff = , (4) reverse shock velocity evolves in time as the densities 8GM(r) in both the expanding ejecta and unshocked hydrogen envelope change, and eventually it turns around. For from each radial and massp coordinate in the progenitor star: simplicity, we ignore this and take vRS to be constant at its initial value. Then the location of intersection be- dM dr dr tween ejecta and the reverse shock can be found, as well M˙ =4πρ(r)r2 . (5) as the resulting ballistic t(M) for material that is re- ≡ dr dt dtff captured after passing through the reverse shock. The For an approximately power law density profile in a reverse shock prescription is important for the Z29 curve 3−α particular shell of a star, ρ(r) = ρ0(r/r0) , the en- in Figure 2. This approximate semi-analytic description 3 α does a reasonable job reproducing the numerical calcula- closed mass is M(r)= ρ0(r0) r0 (r/r0) (for 0 <α< 3), and the fallback accretion rate is: tions in all cases. The largest disagreement is in the re- verse shock cases, where the semi-analytic accretion rate − − 8πα ρ r3 t 3(α 1)/(3 α) overestimates (underestimates) the numerical results at M˙ = 0 0 , (6) early (late) times. For the remainder of the paper, we 3 α t t − 0 0 use the results from the numerical fallback calculations. −1/2 where t0 (2Gρ0) (cf. Eq. 2 of QK12). For α< 0, the enclosed≡ mass is roughly constant, and the accretion rate is: 4. POSSIBLE OUTCOMES − We detail here the possible outcomes of supernova light 8π ρ r3 t (2α 3)/3 M˙ = 0 0 , (7) curves powered by accretion energy. We first assume that 3 t t 0 0 a supernova explodes via the traditional core collapse 3/2 mechanism, whatever that may be. For each progenitor, where now t0 πr0 / 2GM(r0). In this way, the we ran explosions with a variety of energies, in the range freefall accretion≡ rate is set by the density profile of the 1048 to 1051 ergs, in order to explore the full range of progenitor star. p possible outcomes. Only explosions with positive total For the other limit of mildy bound material with vesc energy of non-accreted material at t = 108 are consid- 2 2 −1 ≃ v, the maximum radius, r1 r0(1 v /v ) , becomes ered, and the resulting remnant vs. initial mass distri- ≡ − esc much larger than the initial one, r0. Then the asymptotic bution from these explosions is in excellent agreement fallback rate, M˙ t−5/3, applies (Michel 1988; Chevalier with Zhang et al. (2008). 1989). This asymptotic∝ scaling applies at the latest times The ejection of some stellar layers and the fallback of in all three curves in Figure 2. others is then calculated numerically as described in 3, Using the ballistics solution from Chevalier (1989), we which determines the energy input rate from fallback.§ can bridge these two asymptotic limits to analytically es- We then calculate approximate one zone light curves us- timate the fallback accretion rate at all times for compar- ing the methods described in Appendix A. For these cal- ison with our numerical calculations. For each mass shell, culations, we need the effective diffusion time through the downstream shock velocity is taken from the analytic homologously expanding ejecta Arnett (1979), 6 Dexter & Kasen
SN 2008es SN 1998bw 1044 SN 2008D
44 SN 2010X ) 10 -1 1043 ) -1 L (ergs s 42 1043 10
42 10000
Luminosity (ergs s 10
1000 (K)
41 phot 10 T S33 100 U45 0 20 40 60 80 100 Z29 Days since explosion U60 10
Figure 5. Comparison of fallback powered light curves (solid 2.5·104 lines) from models U45 (purple), S33 (green), S39 (red), and S39O (orange) with some observed supernovae. The parameters for these 2.0·104 events are given in Table 1. The orange dashed curve assumes toff = 7 days. Data points are taken from Gezari et al. (2009, 1.5·104
SN 2008es), Mazzali et al. (2008, SN 1998bw and SN 2008D), and (km / s) 4 Kasliwal et al. (2010, SN 2010X). phot 1.0·10 V
5.0·103
0 3 Mκ 3 (Mej + Mfb)κ td = = , (8) 0 100 200 300 400 500 r4π vc s4π vf c Days since explosion ˙ where Mfb = ξ Mfbdt is the total outflow mass, Figure 6. Sample fallback powered light curves (top), and photo- 2 Efb = ǫM˙ fbc is the injected accretion energy, and spheric temperatures (middle) and velocities (bottom). The dashed R curves in the top panel show the rate of energy injection from fall- vf = (Esn + Efb)/(Mej + Mfb) is the final ejecta ve- back accretion. The temperature is estimated from the one zone locity. Note that there is an ambiguity in determining model, while the velocity is taken to be the maximum of vf and Mfb, dependingp on the interpretation of the fudge factor that calculated from the expanding ejecta. The temperature re- mains fixed at TI during the plateau phase for events where hy- ξ. If ξ indicates the fraction of outflow mass that inter- drogen is present. acts with the supernova ejecta, then the above expression for Mfb applies. If on the other hand, the mass transfer R > 1013cm (RSGs), purple for 1012cm < R < 1013cm, to the ejecta is more efficient while the specific energy of blue for 1011cm < R < 1012cm (BSGs), and green the outflow is lower, Mfb could be significantly larger. 11 We assume a constant opacity κ =0.2 g cm−1, appro- for R < 10 cm (He or C/O stars). This radius also priate for electron scattering for fully ionized elements corresponds to the zero age main sequence metallicity: solar without significant mass loss for RSGs, zero for heavier than hydrogen. This is clearly a coarse approxi- −4 mation, as the actual opacity will depend on the compo- BSGs, 10 Z⊙ for stars in between, and solar with large sition and the presence of Doppler broadened lines. The amounts of mass loss for compact He and C/O stars. effects of recombination on the opacity are, however, in- Events are only plotted if Lp is larger than the thermal cluded in an approximate way (Appendix A). supernova luminosity, While our one zone light curve models account for the E0 R acceleration of the ejecta due to the input accretion en- Lsn . (9) ergy, they lack any information on the radial structure ∼ td vtd of the ejecta. The radiation hydrodynamical calculations The number of points is then set by the number of explo- of KB10 show that energy deposition at the base of the sion energies and progenitor models, as well as the frac- ejecta (in that case from a magnetar) blows a bubble in tion of cases where that condition is met. The number the inner regions, piling up material into a dense shell. of points does not represent an expected rate, since both We expect a similar effect in fallback powered SNe, which the choices of explosion energies and progenitor models will likely also induce an asymmetry if the energy depo- are arbitrary. sition is anisotropic. Figure 3 illustrates the wide range of light curves that For each of the light curves, we measure the time to may result when fallback power is included. Many of the 51 peak, tp, and the peak luminosity, Lp. The results are successful explosions with energies 10 ergs lead to −3 ∼ 41−44 −1 shown in Figure 3 for ǫ = 10 . Each point represents events with tp 50 200 days, Lp 10 ergs s . a single explosion energy and progenitor model, color- The long durations≃ are− similar to those∼ of Type II plateau coded by the radius of the pre-supernova star: red for SNe, and a result of the large ejecta masses and corre- Accretion Powered Supernova Light Curves 7
42−44 −1 spondingly long diffusion times. The final velocities of Lp 10 ergs s . The most luminous cases are ei- these events are also fairly typical of core-collapse super- ther∼ from very massive stars (& 40M⊙) at low metallicity nova explosions ( 3000 kms−1). This is because the or from zero metallicity stars with strong reverse shocks. amount of fallback∼ is much less than the ejecta mass, In both cases, the ejecta masses are 10 40M⊙ with ≃ − −1 so that fallback energy does not appreciably change the low expansion velocities, vf 2000 6000kms . total kinetic energy of the explosion. For smaller ejecta Subluminous Type I and≃ II events− are possible on a masses, the fallback energy can dominate the total ex- variety of timescales. As an example, Figure 5 shows plosion energy, significantly increasing the final velocity. a comparison of a Type I explosion with the transient The diffusion timescale therefore decreases with decreas- 2010X (Kasliwal et al. 2010). The steep decay in this ing ejecta mass both from the smaller total mass and case requires that the accretion turn off about 7 days because of the increasing final ejecta velocity. after explosion (see 5.1). § These effects lead to a strong scaling of Lp with vf , Approximate light curves from examples of each of shown in Figure 4. The roughly L v2 dependence these type of events are shown in Figure 6 along with p ∝ f can be recovered by assuming the fallback energy always photospheric temperatures and velocities. The model 2 parameters are listed in Table 1. The photospheric tem- dominates the supernova energy (Efb vf ), while the fallback mass contributes negligibly to∼ the ejecta mass. perature is taken from the one zone light curve calcu- Furthermore, the scaling assumes that the total fallback lations (see Appendix A). The photospheric velocity is 2 taken to be the maximum of vf and the photospheric energy scales with peak luminosity (Lp Efb vf ), which is true if the accretion rate at late∼ times∼ scales velocity in the expanding ejecta in the absence of in- with its integral over all times. The apparent maximum jected accretion energy. For light curves with recombi- 9 −1 nation, the photospheric properties are meaningless after in vf . 3 10 cms is from the case where the fall- back mass× and energy dominate that of the supernova the plateau phase, since then formally the ejecta are com- pletely optically thin. For relatively short events, the ex- explosion: v ǫ/ξc = 0.01c for our standard pa- 9 −1 f ≃ pansion velocities are high (& 10 cms , and the fallback rameters. In the context of the simple outflow models energy sets the velocity since the ejecta mass is small described in Sectionp 2, this maximum velocity scales as ( 1 2M⊙). Much slower velocities occur in the longer −1/2 ≃ − vf,max rout . The considerable scatter in Figure 4 is duration events with large ejecta masses ( 10 40M⊙). from the∝ breakdown of the above assumptions. The photospheric temperatures are very high≃ at− peak in 4.1. Candidate Events II-L type events ( 20000K). When recombination≃ isn’t important, the light curves Different classes of progenitor stars lead to different are in excellent agreement with the semi-analytic for- outcomes in Figure 3. First, solar metallicity RSG pro- mula in Eq. (A6) for a power-law injection of energy genitors for the most part lead to relatively low luminos- with n =5/3. This is because in nearly all cases the late 43 −1 ity events (Lp . 10 ergs s ). At high ZAMS masses, time accretion rate falls as M˙ t−5/3, while any energy these stars undergo substantial mass loss and become injected on timescales . 1 day∝ is lost to adiabatic expan- stripped He or C/O stars. These progenitors can lead to 42−43 −1 sion, so that its time-dependence does not influence the events with tp 20days, Lp 10 ergs s . These light curve. could potentially≃ explain broad∼ line Type Ibc GRB SNe: high velocities are a natural outcome of the injection of 5. CAVEATS large amounts of fallback energy. Example fits are shown Gravitational energy liberated through fallback accre- in Figure 5 for SN 1998bw (Galama et al. 1998) and SN tion at late times can power unusual supernova light 2008D (Soderberg et al. 2008). In the context of the col- curves (Figures 5 and 6). The calculations in this pa- lapsar model, this suggests that the central engine could per have made many simplifying assumptions; we dis- be responsible for all of the observed properties: early cuss here some of the uncertainties. We have treated time accretion leading to black hole formation, the GRB, the explosion of stars with crude 1D hydrodynamic cal- and the initial supernova explosion; and late time ac- culations using a piston. This method has frequently cretion powering the resulting light curve and the large been used to simulate core collapse supernova explosions expansion velocities. and the resulting fallback (e.g., Woosley & Weaver 1995; BSG progenitors lead to two classes of outcomes de- MacFadyen et al. 2001; Zhang et al. 2008), and the un- pending on the explosion energy. At low explosion ener- certainties in the numerically calculated fallback accre- 43−45 −1 gies, they can produce luminosities Lp 10 ergs s tion rates are probably less than those in the outflow ∼ and peak times of tp 2 40 days. The short du- physics and/or parameters (e.g., ǫ). The light curve cal- ∼ − rations are from the very small ejecta masses, Mej culations further assume simple one zone prescriptions −3 0 ∼ 10 10 M⊙, with nearly all of the star falling back. for the bolometric luminosity and photospheric temper- For ξ−= 0.1 used here, the wind mass is comparable to ature. More sophisticated techniques would be required the ejecta mass, and the injected fallback energy is much for spectral calculations. larger than the initial explosion energy. This leads to The density structures of the pre-supernova models, large final velocities and short diffusion times. Events which directly impact the fallback rate, depend sensi- with tp 25 days can have light curve shapes very similar tively on uncertain prescriptions for convection (semi- to observed∼ luminous Type II-L events. An example fit convection and overshoot) and compositional mixing in to the superluminous Type II-L SN 2008es (Gezari et al. stellar evolution calculations (e.g., Woosley et al. 2002). 2009) is shown in Figure 5. Probably a bigger issue is that the calculations here are At high explosion energies, BSG progenitors lead to a based on a limited set of stellar progenitors, and ignore range of long duration events with t 100 300 days, the effects of rotation and binarity, which may be very p ≃ − 8 Dexter & Kasen
S39 S33 U45 1044 ) Outflow trapped in fallback -1
1043
10 Outflow unbinds fallback 42
Luminosity (ergs s 10
Outflow escapes ejecta 1041
0 20 40 60 80 100 Days since explosion Outflow Opening Angle (degrees) 1
1 10 100 Z29 U60 Time (days) 1044
Figure 7. Regions of θj vs. t parameter space for model Z29 )
where i) the outflow cannot escape the accreting material before -1 depositing most of its energy (Eq. 12), ii) the outflow escapes the outgoing supernova ejecta before losing most of its energy (Eq. 15), and iii) the energy deposited in the accreting material by the out- flow exceeds its binding energy (Eq. 13). The remaining parameter space is where an outflow could plausibly power a supernova light 43 ◦ 10 curve without shutting off continuing accretion. For θj < 10 , the outflow is arbitrarily changed from a wind (vj = 0.1c) to an ul- trarelativistic jet (vj ≃ c). Constraints i) and ii) fix the range of Luminosity (ergs s allowed θj for any disk formation time, ton, while constraint iii) sets the time at which fallback accretion will stop (toff ). common in massive stars (e.g., Sana et al. 2012). There may be additional variety in the range of possible fallback 1042 powered transients from stellar progenitors not consid- 0 100 200 300 400 500 ered here. Days since explosion Further, we have assumed that the stellar material that falls back after the explosion has sufficient angular mo- Figure 8. Sample fallback powered light curves (lines) for models mentum to form a disk, and that this disk can efficiently (top panel) U45, S33, S39O; and (bottom panel) U60 and Z29. The drive a massive wind and/or ultrarelativistic jet. These solid curves assume continued energy injection, while the dashed curves turn off at a range of times, toff = 0.3 − 4.0td. The dashed are both important open questions. The angular momen- curves all assume constant opacity, and the strong effects of recom- tum distribution and surface rotation rates of massive bination can be seen on the light curves in the bottom panel. stars remain highly uncertain (Woosley & Heger 2011). Although previous studies have found prominent polar outflows from geometrically thick black hole accretion 10000 Z29 flows (Stone et al. 1999; Igumenshchev & Abramowicz U45 2000), more recent calculations have found large-scale 1000 S33 circulations to be more common than unbound massive winds (McKinney et al. 2012; Narayan et al. 2012). If so, ultrarelativistic jets may be a more natural explanation 100 for injecting energy into the ejecta. Finally, we have assumed that this outflow can ther- 10 malize in the outgoing supernova ejecta without expelling infalling material and halting accretion. We outline the 1 requirements below to satisfy these assumptions, and es- Initial Angular Velocity (km / s) timate in a few sample cases the required rotation rates for disk formation. 0.1 1.0 10.0 100.0 Disk Formation Time (days)
5.1. Outflow Collimation and Interaction with Ejecta Figure 9. Angular velocities required to form disks at radii at 8 7 6 In order for a fallback accretion powered outflow (ei- r = 10 cm (solid), 10 cm (dotted), and 10 cm (dashed) in a few models from material falling back at a range of turn on times. ther a ultrarelativistic jet with speed v c or massive j ≃ Accretion Powered Supernova Light Curves 9 wind with vj 0.1c) to power a supernova light curve, it must be able∼ to: i) escape the remaining infalling ma- 2 2/3 t fθ vj terial ii) without unbinding it and iii) thermalize in the l 20 j . (15) outgoing supernova ejecta. To order of magnitude, we tesc ≃ vej ! assess the plausibility for this scenario as follows. Following QK12, we assume a magnetically-dominated The requirements i) tesc,FB/tl,FB < 1, ii) Edep/Egrav < outflow, collimated with an opening angle θj . The prop- 1, and iii) tl,ej/tesc,ej < 1 amount to constraints on the agation of the outflow through the remaining bound fall- outflow opening angle, θj , and the time over which fall- back is similar to the propagation of a jet through a host back accretion can continue. Excluded regions of θj vs. t star during a long GRB (e.g., Begelman & Cioffi 1989; parameter space from enforcing these constraints for the Matzner 2003). The speed of the head of the collimated model Z29 are shown in Figure 7. outflow, vh, is determined by pressure balance between For the interaction of accretion energy with remaining the outflow and the host star (Eq. 4 of QK12): fallback material, we find the maximum radius reached at time t by material that will ultimately accrete (Rfb), 1/2 and its remaining total mass (Mfb) and binding energy 4LjR (Egrav). For the interaction of the accretion energy with vh 2 , (10) ≃ 3Mvjθj ! the ejecta, we use Mej and vf estimated at time t. The timescale constraints essentially place limits on where R is the maximum radius of bound material and θj for each type of outflow for all disk formation times, M its total mass. The outflow escape timescale is then ton: at small (large) opening angles, the outflow escapes tesc R/vh. The outflow also drives a lateral shock into (is captured). These ratios also depend on the other the∼ surrounding material, whose speed is approximately quantities, leading to differences between various mod- (QK12), els. Generally, smaller vf (t) leads to higher ejecta densi- ties and help to trap the outflow. Outflows escaping the 1/4 1/2 1/4 3/4 vl f θ v v , (11) ejecta before thermalizing could appear as long duration, ≃ j j h high energy transients (QK12, Woosley & Heger 2011). where f is the efficiency of depositing outflow energy in Outflows trapped in the material still falling back would the surrounding material. If the outflow is dominated by likely deposit energy there more effectively, either un- toroidal magnetic field (e.g., in a helical jet or outflow binding the material or prolonging its accretion to later from a rotating disk), a typical value from numerical cal- times. culations is f 0.03 (Bucciantini et al. 2007). For the At late times, the outflow will unbind any remaining ≃ outflow to escape, tesc should be shorter than the time material, shutting off further accretion. This is because for the lateral shock to envelope the star, tl R/vl. For the energy deposition into the accreting material at late ∼ −5/6 interactions with the remaining bound fallback material, times scales as Ltesc t , while its binding energy this ratio is: scales as M /r t−∼4/3. Equating these gives the turn fb fb ∼ off time (toff ) for each event. This turn off time tends to 3 6 1/8 be shorter in higher energy explosions, since the bulk of t Mfbv θ l 0.6 j j . (12) the late time accretion comes from loosely bound mate- tesc ≃ Lj Rfb ! rial. Sample light curves for events where accretion shuts off are shown in Figure 8 for a range of toff , assuming a In addition to requiring tl > tesc, continued accretion constant opacity. Once the injected energy runs out, the requires that the energy deposited, light curve decays according to Eq. (A5), but with an
ton+tesc initial luminosity L(toff ). This may be particularly rele- vant for long duration transients in models like U60 and Edep = f Lj(t)dt, (13) ton Z29, where the turn off time ( 80 days for Z29) is likely Z to be comparable to the time≃ to peak. should be less than the binding energy of the remain- Although these estimates demonstrate the plausibil- ing fallback material, where ton (toff ) is the time after ity of accretion-driven outflows powering supernova light explosion at which the outflow turns on (off). curves, detailed physical calculations will be required to Conversely, for the outflow energy to be deposited ef- assess this scenario accurately. Further, the statement ficiently in the ejecta the outflow escape time should be that the outflow cannot escape the ejecta does not pro- shorter than the energy deposition time. Since the mech- vide an efficient means of thermalization, since we have anism for thermalizing the outflow energy and its associ- assumed f = 0.03. Using a larger value of f 1 would ated timescale are unknown, we can instead use the same shift the range of allowed opening angles to≈ favor rela- comparison of tl and tesc as above. tivistic jets, and lead to the outflow unbinding the ac- In this case, vej is larger than vh, and the escape time creting the material time at proportionally earlier times. can be estimated from setting R = vejt and finding when The efficiency of thermalization depends on how ex- vh = vej (QK12): actly energy is transported from the accretion disk to the supernova ejecta. We have assumed that this mech- 2 3Mejvejvjθj anism is a highly magnetized disk wind or an ultra- tesc = . (14) 4Lj relativistic jet. If instead the wind is not highly mag- netized, a double (forward/reverse) shock structure will Similarly, we can find the ratio tl/tesc under the same form when it catches up with the slowly moving inner assumptions. The result is: layers of the supernova ejecta (KB10). The situation is 10 Dexter & Kasen analogous to the commonly case of supernovae interact- The accretion power released when material falls back ing with circumstellar material, only here the interaction onto a compact remnant at late times could power un- happens inside, rather than outside the remnant. In ei- usual supernova light curves. We have explored the con- ther case, shocks should be efficient in thermalizing the sequences for a variety of progenitors and explosion en- kinetic energy of the wind. Some recent semi-analytic ergies, using numerical calculations of the fallback ac- (Begelman 2012) and numerical (McKinney et al. 2012; cretion rate and order of magnitude estimates of the Narayan et al. 2012) calculations of non-radiative accre- resulting energy injection. While most of the fallback tion flows have found large-scale circulations or convec- typically occurs at early times, it may be significant at tive motions (Lindner et al. 2011) as well as or instead late times in very massive stars, for low explosion ener- of outflows. This may also be a relevant mechanism for gies, or when a strong reverse shock forms at the hydro- transporting accretion energy to large radius. gen/helium boundary. We have demonstrated that it is If the accretion energy cannot efficiently thermalize, it plausible that, under certain circumstances, the energy will likely still lead to high ejecta velocities vf . In the available from accretion could power an outflow which case of an event like S33, this could still explain≃ broad then thermalizes in the supernova ejecta. line Type Ibc supernovae: radioactivity would power the The events we have described are different and more light curve and accretion energy would lead to the high diverse than what have previously been studied as “fall- observed photospheric velocities. This is also a possi- back supernovae”. Fryer et al. (2009), for example, con- ble outcome of early time accretion onto a magnetar sidered the case of massive star collapse in which most (Piro & Ott 2011). of the material fell into the central black hole and only a fraction was ejected. Because they also assumed that the 5.2. Angular Momentum and Disk Formation surrounding medium was very dense and extended (due Given the viable range of disk formation times for fall- to mass loss prior to explosion) the supernova shock wave back accretion powered supernovae 5.1 and the initial did not breakout of the circumstellar gas until late times. stellar radii accreting at those times, we can calculate the The result was a dim, shock-powered transient lasting required angular velocity. Curves for models Z29, U45, from weeks to months. Moriya et al. (2012) similarly and S33 are shown in Figure 9 for forming disks at radii considered the case in which most of the star fell back 6−8 from 10 cm, or r 1 100 for a 10M⊙ black hole. and only a very small amount ( 0.1M⊙) was ejected. Naively assuming rigid∼ rotation,− in all cases disks can By assuming that this ejecta was∼ enriched with 56Ni, form at the required times without exceeding breakup at they found a brief and sub-luminous radioactively pow- the outer edge of the star. The required rotation rates ered transient similar to SN 2005E. Both of these pre- essentially scale with explosion energy: for large explo- vious scenarios neglected the possible input of accretion sion energies, the envelope is expelled, and larger rotation energy from fallback (i.e., they assumed ǫ = 0). As we rates are required for the disk to form from material that have shown, accretion may re-energize the ejecta at late was originally at smaller radius. times and hence power much brighter emission. Under this assumption, we can also calculate the max- The power from fallback accretion may be relevant imum disk size from fallback accretion, and the corre- for explaining recently discovered classes of peculiar su- 2 −1 sponding viscous time, tvisc (R/H) α tdyn, where pernovae. These may include the Type IIL supernovae H/R is the accretion flow scale∼ height and α is the stan- that are extremely luminous and of relatively short du- dard dimensionless viscosity parameter in accretion the- ration (e.g., Gezari et al. 2009; Miller et al. 2009) as ory (Shakura & Sunyaev 1973). Even with conservative well as those that are moderately bright and of very assumptions (H/R =0.1, α =0.01), this timescale only long duration (e.g., Miller et al. 2010; Rest et al. 2011; becomes larger than the fallback timescale for the high- Chatzopoulos et al. 2011). Several of the observed est rotation rates and early disk formation times . 104s. Type II events, however, also show narrow hydrogen This is because assuming rigid rotation, the total disk emission features in their spectra, indicating that inter- size never greatly exceeds its formation radius. action with a dense circumstellar medium is occurring Stellar cores spin up as they contract, and depend- and may be responsible for the luminosity. ing on the efficiency of angular momentum transfer from Many of the predicted Type II events with 1051 erg magnetic torques (Spruit 2002) can transfer much of explosion energies have very long times to peak∼ (100 200 the core angular momentum to the outer layers (e.g., days). Accretion energy is likely to unbind the remaining− Heger et al. 2005). If this mass is retained, as in our infalling material on a comparable timescale, turning off models from low metallicity progenitors (e.g., Z29 and the power source for the light curve ( 5.1). We therefore U45), it will likely form a disk upon fallback for modest predict that these events could be seen§ as very bright ZAMS rotation rates. If instead this mass is lost (e.g., Type II supernovae that disappear suddenly from view. S33), insufficient angular momentum may remain to form In general, late time turn off would be an observational a disk. If the red supergiant is in a binary system, tidal signature of fallback accretion powered supernovae. interactions may be an efficient means to spin up the star Fallback accretion could also power very bright Type I sufficiently to cause disk formation even if the envelope events. The models considered in this paper reached is lost during subsequent evolution (Woosley & Heger peak luminosities of 1043 ergs s−1, similar to the 2011). The latter scenario may be fairly common, given broad-lined SNe Ic like∼ SN 1998bw. If the accretion the frequency of massive stars in binaries (Sana et al. efficiency is assumed to be higher than our fiducial 2012). These scenarios should be considered in more de- case, it is possible for some events to reach luminosities tail in future work. & 1044 ergs s−1, in which case fallback could power the super-luminous hydrogen poor events such as SN 2005ap 6. CONCLUSIONS Accretion Powered Supernova Light Curves 11
(Quimby et al. 2011). light curve are apparently quite special, as the scenario Another effect that may produce super-luminous requires sufficient angular momentum to form a disk and events like SN 2005ap involves mass loss. Some events an evolution that permits fallback to persist long enough considered here (e.g., Z29) are brightened considerably to drive energetic outflows at late times. Such a conflu- by enhanced accretion from material decelerated by a ence of factors may be rare in the Universe. On the other reverse shock forming at the H/He interface. A similar hand, observational surveys show that the rate of pecu- outcome could occur in both Type I/II events where the liar SNe – in particular the rate of very luminous ones progenitor has experienced considerable mass loss shortly – is only a small fraction that of standard core collapse before explosion. In this case, the reverse shock would events. It is possible that fallback power plays a role in be formed when the outgoing shock wave reaches the some of these spectacular events. interface between the progenitor star and the massive wind or ejected shell. The subsequent inward propaga- tion of the reverse shock could lead to order of mag- We thank A. Heger for making a large number of pre- nitude increases in the fallback accretion rate at later supernova stellar models publicly available. JD thanks times. Interaction of supernova ejecta with circumstellar L. Bildsten, B. Metzger, C. Ott, T. Piro, E. Quataert, shells at radii 1015 cm is commonly considered to ex- E. Ramirez-Ruiz, and S. Woosley for stimulating discus- plain superluminous∼ supernovae via thermalization of the sions related to this work. This work is supported by the kinetic energy (e.g., Gal-Yam 2012). Surprisingly, inter- Director, Office of Energy Research, Office of High En- action with circumstellar material at much smaller radii ergy and Nuclear Physics, Divisions of Nuclear Physics, ( 1011 1012 cm) may also lead to an super-luminous of the U.S. Department of Energy under Contract No. event,∼ but− by very different means – by enhancing fall- DE-AC02-05CH11231, and by a Department of Energy back accretion and feeding the central compact object at Office of Nuclear Physics Early Career Award. late times. Also of potential relevance to the fallback scenario are REFERENCES dimmer supernovae that decline very rapidly after peak (e.g., SN 2002bj and SN 2010X, Poznanski et al. 2010; Arnett, W. D. 1979, ApJ, 230, L37 Kasliwal et al. 2010). The short duration of these events —. 1982, ApJ, 253, 785 Baganoff, F. K., et al. 2003, ApJ, 591, 891 makes it difficult to explain them as radioactively pow- Balberg, S., Zampieri, L., & Shapiro, S. L. 2000, ApJ, 541, 860 ered transients. In the case of SN 2002bj at least, the Begelman, M. C. 2012, MNRAS, 420, 2912 mass of 56Ni inferred from the light curve peak exceeds Begelman, M. C., & Cioffi, D. F. 1989, ApJ, 345, L21 the total ejecta mass inferred from the light curve dura- Blandford, R. D., & Begelman, M. C. 1999, MNRAS, 303, L1 tion (diffusion time), seemingly ruling out a radioactively Bucciantini, N., Quataert, E., Arons, J., Metzger, B. D., & Thompson, T. A. 2007, MNRAS, 380, 1541 powered events. In the fallback scenario, short lived tran- Castor, J. I. 2004, Radiation Hydrodynamics (Cambridge, UK: sients are possible, especially if the energy injection from Cambridge University Press) accretion cuts off the fallback abruptly (Figure 8). Chatzopoulos, E., et al. 2011, ApJ, 729, 143 The true range of possible light curves powered by Chevalier, R. A. 1982, ApJ, 258, 790 fallback accretion is likely much larger than shown in —. 1989, ApJ, 346, 847 Chevalier, R. A., & Irwin, C. M. 2011, ApJ, 729, L6 Figure 3. Those events are limited in both the variety Colgate, S. 1971, ApJ, 163, 221 of progenitor models and by our neglect of the fallback De Villiers, J.-P., Hawley, J. F., Krolik, J. H., & Hirose, S. 2005, accretion physics. The latter could conceivably lead to ApJ, 620, 878 variations in ǫ in either direction: smaller disks and/or Fender, R., Wu, K., Johnston, H., Tzioumis, T., Jonker, P., more efficient thermalization of the outflow could lead Spencer, R., & van der Klis, M. 2004, Nature, 427, 222 45 −1 Foley, R. J., et al. 2009, AJ, 138, 376 to higher peak luminosities & 10 ergs s . Conversely, Fryer, C. L. 2009, ApJ, 699, 409 lower efficiencies could help explain a wider variety of Fryer, C. L., Belczynski, K., Wiktorowicz, G., Dominik, M., sub-luminous supernovae (e.g., SN 2008ha, Foley et al. Kalogera, V., & Holz, D. E. 2012, ApJ, 749, 91 2009). Fryer, C. L., et al. 2009, ApJ, 707, 193 Further modeling is needed to identify the observa- Gal-Yam, A. 2012, Science, 337, 927 Galama, T. J., et al. 1998, Nature, 395, 670 tional signatures of fallback accretion powered super- Gezari, S., et al. 2009, ApJ, 690, 1313 novae, and determine how we might distinguish these Heger, A., Woosley, S. E., & Spruit, H. C. 2005, ApJ, 626, 350 events from other means of generating unusual light Igumenshchev, I. V., & Abramowicz, M. A. 2000, ApJS, 130, 463 curves. Possible signatures include a tail with L t−5/3 Kasen, D., & Bildsten, L. 2010, ApJ, 717, 245 ∝ Kasen, D., & Woosley, S. E. 2009, ApJ, 703, 2205 at late times compared to the diffusion time (but be- Kasliwal, M. M., et al. 2010, ApJ, 723, L98 fore the ejecta become completely optically thin), some- Kochanek, C. S., Beacom, J. F., Kistler, M. D., Prieto, J. L., what different from those from radioactivity or magne- Stanek, K. Z., Thompson, T. A., & Y¨uksel, H. 2008, ApJ, 684, tar spindown. More promisingly, at late times ( 100 1336 ∼ Kohri, K., Narayan, R., & Piran, T. 2005, ApJ, 629, 341 days) it seems likely that the accretion energy will un- Lindner, C. C., Milosavljevic, M., Shen, R., & Kumar, P. 2011, bind the infalling material ( 5.1), shutting off accre- eprint arXiv, 1108.1415 tion and leading to a sudden§ decrease in luminosity. If MacFadyen, A. I., & Woosley, S. E. 1999, ApJ, 524, 262 instead accretion continues for decades after the explo- MacFadyen, A. I., Woosley, S. E., & Heger, A. 2001, ApJ, 550, sion, the black hole could emerge as an observable X-ray 410 Marrone, D. P., Moran, J. M., Zhao, J.-H., & Rao, R. 2007, ApJ, source (Balberg et al. 2000), as has been suggested for 654, L57 SN 1979C (Patnaude et al. 2011). Matzner, C. D. 2003, MNRAS, 345, 575 The conditions for fallback to influence the supernova Matzner, C. D., & McKee, C. F. 1999, ApJ, 510, 379 Mazzali, P. A., et al. 2008, Science, 321, 1185 12 Dexter & Kasen
McKee, C. F. 1974, ApJ, 188, 335 Perets, H. B., et al. 2009, eprint arXiv, 0906, 2003 McKinney, J. C. 2006, MNRAS, 368, 1561 Piro, A. L., & Ott, C. D. 2011, ApJ, 736, 108 McKinney, J. C., Tchekhovskoy, A., & Blandford, R. D. 2012, Popov, D. V. 1993, ApJ, 414, 712 ArXiv e-prints Poznanski, D., et al. 2010, Science, 327, 58 Michel, F. 1988, Nature, 333, 644 Quataert, E. 2004, ApJ, 613, 322 Miller, A. A., et al. 2009, ApJ, 690, 1303 Quataert, E., & Kasen, D. 2012, MNRAS, 419, L1 —. 2010, MNRAS, 404, 305 Quimby, R. M., et al. 2011, Nature, 474, 487 Milosavljevic, M., Lindner, C. C., Shen, R., & Kumar, P. 2010, Rest, A., et al. 2011, ApJ, 729, 88 eprint arXiv, 1007.0763 Sana, H., et al. 2012, Science, 337, 444 Mirabel, I. F., & Rodr´ıguez, L. F. 1998, Nature, 392, 673 Shakura, N. I., & Sunyaev, R. A. 1973, A&A, 24, 337 Moriya, T., Tominaga, N., Tanaka, M., Nomoto, K., Sauer, D., Smith, N., & McCray, R. 2007, ApJ, 671, L17 Mazzali, P., Maeda, K., & Suzuki, T. 2010, ApJ, 719, 1445 Soderberg, A. M., et al. 2008, Nature, 453, 469 Moriya, T. J., Blinnikov, S. I., Tominaga, N., Yoshida, N., Spruit, H. C. 2002, A&A, 381, 923 Tanaka, M., Maeda, K., & Nomoto, K. 2012, ArXiv e-prints Stone, J., Pringle, J., & Begelman, M. 1999, MNRAS, 310, 1002 Narayan, R., Sadowski, A., Penna, R. F., & Kulkarni, A. K. 2012, Woosley, S. E. 1993, ApJ, 405, 273 ArXiv e-prints Woosley, S. E., & Heger, A. 2011, eprint arXiv, 1110.3842 Narayan, R., & Yi, I. 1994, ApJ, 428, L13 Woosley, S. E., Heger, A., & Weaver, T. A. 2002, Reviews of Patnaude, D. J., Loeb, A., & Jones, C. 2011, New Astronomy, 16, Modern Physics, 74, 1015 187 Woosley, S. E., & Weaver, T. A. 1995, ApJS, 101, 181 Pen, U.-L., Matzner, C. D., & Wong, S. 2003, ApJ, 596, L207 Zhang, W., Woosley, S. E., & Heger, A. 2008, ApJ, 679, 639
APPENDIX LIGHT CURVE MODELING KB10 described a one zone diffusion estimate for bolometric supernova light curves powered by an injection of energy with arbitrary time-dependence, H(t). The argument follows along the lines of Arnett (1979, 1982). As the ejecta expand, energy is lost both due to escaping radiation (L) and adiabatic losses from expansion: ∂(E t) ∂V int = p + H L (A1) ∂t − ∂t − Assuming optical depth τ 1, the diffusion equation gives an approximate relationship between E and L: ≫ int L c E /V F = int . (A2) 4πR2 ≈ 3κρ R With the definition of the diffusion time (Eq. 8) and assuming V t3, Eq. (A1) can be re-written as: ∝ d t t + L(t)= H(t). (A3) dt t2 t2 d d The general solution for L(0) = 0 is:
2 2 −t /2t e d 2 2 L(t)= dttet /2td H(t). (A4) t2 d Z For an initial shock energy H = E δ(t t ), this gives: 0 − 0 E0 t0 2 2 2 L(t)= e−(t −t0)/2td , (A5) td td in agreement with Eq. (9) with t0 = R/vsh modulo the exponential factor of order unity at peak. −n For accretion powered light curves, a power law form, H = L0(t/t0) with n = 5/3 provides an excellent ap- proximate description for the numerical light curves integrated with Eq. (A4). The semi-analytic solution for t>t0 is:
n n/2 2 2 t0 2 2 1 n t n t L(t)= L e−t /2td γ 1 , 0 γ 1 , , (A6) 0 t −2 − 2 −2t2 − − 2 −2t2 d d d where γ(s, x) is the lower incomplete Gamma function. The incomplete Gamma function is complex for negative arguments. Since the observed light curve and the integral in Eq. (A4) are real, the imaginary part in Eq. (A6) vanishes. In the special case of constant energy injection (n = 0), the solution is (cf. Eq. 13 of KB10):
2 2 2 L(t)= L 1 e−(t −t0)/2td , (A7) 0 − h i for t toff , where H = 0 for t>toff . This≤ light curve estimate assumes a constant opacity. A different limit occurs when the outer portion of the ejecta drops below the ionization temperature and recombines. The opacity drops suddenly in the recombined material, and Accretion Powered Supernova Light Curves 13 the effect is that of a recombination wave passing through the ejecta. This effect significantly alters the light curve evolution of Type II-P supernovae (e.g., Popov 1993; Kasen & Woosley 2009). During the passage of the recombination wave through the ejecta, the photosphere remains at the ionization tem- perature, TI . The luminosity can then be calculated from the time-dependent photospheric radius:
2 4 L =4πRp(t)σTI , (A8) where Rp(t)= xi(t)vt and xi is the dimensionless position of the photosphere in the expanding ejecta. We can write 3 the equivalent of Equation (A1) for the evolution of the internal energy of the ionized region, ǫVi =4/3πǫ(xivt) : 1 ∂ǫ 4 ∂x 4 H L + i + = − . (A9) ǫ ∂t xi ∂t t ǫVi
We again use the diffusion equation to write L in terms of the internal energy, except now using Rp instead of R. Finally, we equate the photospheric luminosity with that from diffusion in the ionized region, which gives an expression for ǫ in terms of xi. The result is a non-linear first order differential equation for xi(t): dx 2x t 1 H i = i + . (A10) dt − 5t − 5t2x 5x3t 4πv2t2σT 4 d i i d I In the absence of heating (H = 0), Eq. (A10) is similar to Eq. 14 of Popov (1993), except with slightly different numerical coefficients. In this case, the analytic solution for the luminosity starting at time ti, such that xi(ti) = 1, is: t2 t4 L(t)=4πσT 4v2 t6/5t4/5 1+ i . (A11) I i 7t2 − 7t2 d d In general, we calculate the luminosity assuming constant opacity using Eq. (A4). Then, the approximate one 4 2 2 zone photospheric temperature is given by σTp = L/4πv t . When this drops below TI , we numerically integrate Eq. 2 2 2 4 (A10) for xi(t), and then calculate L(t)=4πvf t xi σTI . The recombination wave can significantly increase the peak luminosity in hydrogen rich progenitors (see Section 4). More accurate radiative transfer calculations would likely find smoother light curves than those estimated from this one zone approach.