Bumpy Light Curves of Superluminous Supernovae · Hydrogen-poor superluminous supernovae M.Nicholl...

Bumpy Light Curves of SuperluminousSupernovae

Elena SorokinaSternberg Astronomical InstituteMoscow University

in collaboration withS.BlinnikovA.TolstovK.Nomoto

Ringberg Workshop 26 July 2017

• SLSNe (Type I (no hydrogen), Type II) are brigher than -21 magnitude in any optical band at the maximum brightness

• Subclasses: SLSN I normal, SLSN I-R, SLSN I - fast, SLSN II-n, SLSN IIn-peculiar, SLSN II-L, SN Ia-CSM • Rise ~ 20-60 d, decline ~ 20-500 d, Erad ~ (1-10) .1051 erg,

rate/CC ~ 0.1%, ~ 100 SLSNe.

Superluminous supernovae (SLSNe)

Superluminous supernovae (SLSNe)

• SLSNe-II: Occur in all galaxies

SLSNe-I: Exclusively in M* < 109.2 M ⊙ , Z < 0.5 Z⊙ galaxies

Gal-Yam 2012

(Perley 2016)

2

Hydrogen-poor superluminous supernovae

M.Nicholl et al. 2015 A.Papadopoulos et al. 2015griz pseudobolometric light curves

−40 −20 0 20 40 60 80 100 120Rest-frame days from maximum light

−23

−22

−21

−20

−19

−18

−17

−16

−15

Mgr

iz

2005ap2007bi2008es2010gx2011ke2011kf2012il2013dg

2013hxCSS121015iPTF13ajgLSQ12dlfLSQ14bdqLSQ14moPS1-10bzjPS1-10ky

PS1-11apPTF09cndPTF09cwlPTF10hgiPTF11rksPTF12damSCP06F6SSS120810

3

Observational properties of SLSN-I

• Mmax <= − 21m; Lbol,max > 7 · 1043 erg/s.• Typically, they are very blue and emit qiute much inUV range.

• The range of Mmax is not so large (' 1m); the rangeof the slopes after max is quite large.

• V ' 10, 000km/s, starting from maximum.• Many, if not all, of SLSNe has small bump on therising part of the light curve.

• For some of SLSNe-I: H lines in nebular spectrawith v ' 4, 000km/s, R ' 4× 1016cm (iPTF13ehe;Lin+ 2015).

An ideal model must explain all the properties at once.We are still far from the ideal.

4

Three scenarios proposed for SLSNe-I

• Pair instability Supernovae, PISN• “Magnetar” pumping (BUT observed magnetars areslowly rotating, and here millisecond periods areneeded)

• Shock interaction with CSM, e.g. as a result ofPulsational pair instability, PPISN

5

3 outcomes of pair-instability

Here are only He-core models,labeled by “He” and the mass ofthe core. They all reach pairinstability, subsequentlyexperiencing 1) pulsations

(He48),2) complete disruption (He80), or3) direct collapse (He160).

8.5

8.6

8.7

8.8

8.9

9

9.1

9.2

9.3

9.4

9.5

9.6

9.7

9.8

9.9

10

3 4 5 6 7 8 9 10

log(T

c)

log(ρc)

Fe disintegration

Pair instability

He 48

He 80

He 160

6

Bolometric light curve and “magnetar” fit for PTF 12dam,Nicholl+, 2013, simple analytical model by S.Jerkstrand,colored curves – PISN models

7

Compact PISN models with Hydrogen lost; Kozyreva+ 2016

0 50 100 150 200 250 30041

42

43

44

45

Time [days]

log

Lb

ol

[erg

s−

1]

P250

PTF12dam

correctedP200

0 100 200 300

−22

−21

−20

−19

−18

−17

mag

nit

ud

e

u

P200

P250

0 100 200 300

−22

−21

−20

−19

−18

−17

gP200

P250

0 100 200 300

−22

−21

−20

−19

−18

−17

Time [days]

r

P200

P250

0 100 200 300

−22

−21

−20

−19

−18

−17

Time [days]

mag

nit

ud

e

i

P200

P250

0 100 200 300

−22

−21

−20

−19

−18

−17

Time [days]

z

P200

P250

8

PISN: Kozyreva+ 2014

0 100 200 300 400 500 60041

42

43

44

45

Time, days

250M

SN2010gxslsn Ic

PS1−10bzjslsn Ic

SN2007bislsn Ic

PS1−10kyslsn Ic

PS1−10awhslsn Ic

log

Bo

lom

etr

ic L

um

ino

sit

y

It is clear that at least some SLSNe are not PISN.

9

Ejecta–CSM interaction models with modest energy

PISN and magnetar models requires very high explosion energy andextremly high radioactive nickel production.

In many cases, CSM interaction scenario doesn’t require so extremeparameters.

Our Lagrangean 1D code STELLA with multigroup radiative transferallows us to get more economical models.

Sorokina+ 2016, Tolstov+ 2016,2017

10

Radiative shock waves: a powerful source of light in SLSNe.Cold Dense Shell, Smith et al. 2008, a cartoon

11

STELLA reproduces the range of SLSNe in shock model:2 extreme cases

Explosion energy is just 2 - 4 foe12

Light curve model for SN2010gx

N0

Synthetic light curves for the model N0, one of the best for SN 2010gx,in r, g, B, and u filters compared with Pan-STARRS and PTF observations.Pan-STARRS points are designated with open squares (u, g, and R bands),

PTF points, with filled circles (B and r bands).

13

14

15

Many SLSNe-I have a pre-maximum bumpNicholl & Smartt 2015

0 10 20 30 40 50 60 70 80 90Rest-frame days

−26

−24

−22

−20

−18

−16

−14

Abs

olut

e m

agni

tude

+ c

onst

ant

LSQ14bdq (g - 4.5 mag)PTF09cnd (g - 3)SN1000+0216 (1600 Å)PS1-10ahf (2900Å - 1.7)

iPTF13ajg (u + 0.5)SNLS 06D4eu (2420Å + 1)PS1-10pm (2840Å + 1.8)SN 2006oz (g + 3.5)

16

Some others have many bumps on the declining stage

SLSN-I PTF15esb: bumpy light curveSLSN-I PTF15esb: bumpy light curve

17

SN 2015bn – another example of very bumpy SLSN(Nicholl+ 2015). Pre- and post-maximum bumps might havesimilar or differnt origin.

−100 −50 0 50 100 150 200 250Rest-frame days from maximum light

−23

−22

−21

−20

−19

−18

−17

Abs

olut

e m

agni

tude

z-1i-0.5rg+0.5u+1

15bn: 12dam:-1.4-1.1-0.4+0.3+1

PTF12dam

0 50 100 150 200 250 300 350Rest-frame days from first detection

43

44

log 1

0Lbo

l(e

rgs−1

)

2015bnPTF12dam

18

Radioactive origin of pre-maximum bump?

Doubled peak of SLSN-I (by R. Quimby)Doubled peak of SLSN-I (by R. Quimby)

19

Some more realistic explanaitions of bumps

• 1st bump - shock break-out through an extended pre-SN envelope,main max - another (whichever) main source of energy.

• Two (or more) subsequent explosions/ejections (quite natural forPPISN scenario – see recent paper by Woosley 2017, with lots ofmodels)

• Stratification of ejected elements along the radius:Innermost layes of CO-rich gas become opaque at lower T, thenphotosphere stalls waiting for helium layers to become opaque athigher T.

20

Shock breakout – analytical formula

21

Shock breakout looks similar for any energy sourceMagnetar inside:

0 10 20 30 40 50days since explosion

1043

1044

lum

inosi

ty (

erg

s/s)

Msn = 5 M¯

Msn = 10 M¯

Msn = 15 M¯

Msn = 20 M¯

0 10 20 30 40 50days since explosion

1043

1044

Esn =1×1051 ergs

Esn =2×1051 ergs

Esn =3×1051 ergs

0 5 10 15 20 25 30 35 40

days since explosion

1043

1044

lum

inosi

ty (

erg

s/s)

FullThermalization

shocksupernovatotal

0 5 10 15 20 25 30 35 40

days since explosion

1043

1044

InefficentThermalization

Thermal bomb:

22

Shock breakout

0 50 100 150 200 250 300time since explosion (days)

43.0

43.2

43.4

43.6

43.8

44.0

44.2

44.4

log

10 b

olo

metr

ic lum

inosi

ty

LSQ14bdqmodel 1model 2

Kasen+ 2016 (Magnetar model) 23

24

Two explosions details

25

26

1st piston with E = 4B, then thermal bomb with E = 20B

27

Problems with 2-explosion models

It is the easiest model for light curve calculation,BUT it is unclear if it is physically possible to produce 2 energeticexplosion or mass ejection so close in timeand WHY this time delay and brightness ratio of two maxima are sosimilar for many SLSNe.

28

Post-shock-breakout cooling, then interaction(or subsequent interactions with several ejections)

Piston expands 8 Msun of He-envelope with E = 1B;density distribution in 1e7 s and 2e7 s

29

Then 2nd explosion produces the whole light curve

30

Opacities of CO mixture (red) and He (black)

T=7000K T=11000K

31

Different composition

Helium C:O=1:1 C:O=9:1

Sorokina+ (2016)

32

Pure helium vs. CO/He models

33

Conclusions

• Interaction models are able to reproduce both narrow and wide LCsof SLSNe. They require quite large mass of CO-rich material ejectedwithin few months to years before the final explosion. The problemof high velocities will be discussed in the next talk.

• The origin of pre-maximum bumps still remain questionable.• Most natural explanation of the post-maximum bump is theinteraction of SCM layers or bullets.

• The combination of the SLSN scenarios is promising.• The ideal scenario have to explain ALL observational features atonce: LCs including bumps, high velocities, etc.

The work is in progress

34

Thank you!

35