Supernova Grand Challenges on ATLAS - Lawrence Livermore ...

monkeyresultΜηχανική

22 Φεβ 2014 (πριν από 3 χρόνια και 1 μήνα)

191 εμφανίσεις

Supernova Grand
Challenges on
ATLAS

R. D. Hoffman

Nuclear Theory & Modeling Group

N
-
DIV
-

LLNL

This work performed under the auspices of the U.S. Department of Energy by

Lawrence Livermore National Laboratory under Contract DE
-
AC52
-
07NA27344

UCRL
-

PRES
-

401463

CAC
-

Collaborators



UCSC
-

S. E. Woosley & D. Kasen


LBNL (CCSE)
-

J. Bell, A. Almgren,


M. Day, A. Aspden, & P. Nugent


SUNY Stony Brook
-

M. Zingale &


C. Malone


LLNL (CASC)
-

L. Howell & M. Singer

SUPRNOVA

4M CPU hrs



Feb 08: 3.4M

Big Questions:


How did the Universe
begin?


How did it evolve to its
present state (extent,
composition, dynamics)?


Where is it headed (a big
crunch, long coast, a
bounce)?


These and other pressing
questions are the purview
of COSMOLOGY


Current best theory: The
Big Bang

What’s new? DARK ENERGY


Could be 2/3 of all matter
and energy in the Universe.


Causing the observed
expansion to accelerate.


Need to determine EOS.


Many theories, many
conflicts, little guidance.




All agree: observations at

high red
-
shift are necessary.



SNe Ia
-

“standard candle”

SN 1994D

Entering a “Precision Era”


Use of SNe Ia as standard
candles has caused a
revolution in cosmology.


In fact most theories are
based on nearby SNe Ia’s.



Evolutionary effects like
metallicity, rotation, or even
asymmetric explosions could
influence our interpretation
of cosmological parameters
at high
-
Z
.

Observations
of higher
-
Z

Ia’s suggest
they have a
larger intrinsic
scatter in their
brightness.

Supernova Discovery History

Asiago Catalog (all supernova types)

Rvd. Evans: 41 SN (81
-
05)

KAIT: 490 SN (88
-
06)

Supernova Factory

Lick observatory SN
search

CfA SN group

Carnegie SN project

ESSENCE

Supernova Legacy Survey

Supernova Discovery Future

Rough predictions and promises…

PanStarrs

Dark Energy Survey

JDEM

Large Synoptic Survey




Telescope (LSST)

Can we use Type Ia SNe as
reliable standard candles at the
few % level?


Systematic error
, not statistical
error,

is the issue (e.g., luminosity
evolution)

SN Ia Progenitors

Accreting white dwarf near the Chandrasekhar limit

Accretion rate:

10
-
7
M
sun

/ year

Issues with the single degenerate scenario


Where is the hydrogen?

How do you make them in old (~10 Gyr)
systems?

What about observed “Super
-
Chandra”

events?


Could double white dwarf systems be the
answer?

C/O

boom

Fe

56
Ni

Si/S/Ca

C/O

M
WD
=1.38 M
sun

r
c
=3x10
9

g/cc

Type Ia Supernova Light Curves

powered by the beta decay:
56
Ni
56
Co
56
Fe

Type Ia Width
-
Luminosity
Relation

brighter supernovae have broader light curves

L
p

= f(w)

Type Ia Supernova Spectrum

Most Sne Ia’s
look similar:


line features of
doubly ionized
Mg, Si, S, Ca
(intermediate Z)


as well as Fe, Co

Day 35 after explosion

Time Evolution of Spectrum

Recession of photosphere reveals deeper layers

Fe

56
Ni

Si/S/Ca

C/O

Day 15 after explosion

Model

SN1994D

free expansion

Light Curves / Spectra

(~1
-
100 days
)

radioactive decay / radiative transfer


ignition

Presupernova Evolution

(
~1000
-
10
9

years
)

accreting, convective white dwarf


Explosion

(~1
-
100 secs
)

turbulent nuclear combustion / hydrodynamics


t = 0.0 sec

t = 0.5 sec

t = 1.0 sec

t = 1.5 sec

w ~ 10
-
4

cm

R
WD

~ 1800 km

Ma 2007

Roepke 2007

Kasen 2007

What are the progenitors?


How and where does ignition happen?


How might the deflagration transition into a

detonation?


How do the light curves and spectra
depend upon the progenitor, its
environment and the nature of the
explosion?



Pressing Questions

The Theoretical Understanding
of Type Ia Supernovae

Able to take large time steps based on the fluid velocity
rather than the speed of sound in the star.



SNe
:
designed to study the

microphysics of nuclear
flames and how the flame interacts with turbulence.
Forms the basis of the sub
-
grid model needed for the full star
calc’s.



MAESTRO
:
incorporates background density
stratification of the star and compressibility effects due
to heat release and buoyancy.


CASTRO
:
our compressible rad
-
hydro code used
for late time simulations when the low Mach number
assumption is no longer valid. Also for SNII & GRB’s.

SNe & MAESTRO

low Mach number hydro codes

3
-
dimensional Time
-
Dependent


Monte Carlo Radiative Transfer

SEDONA Code

Expanding atmosphere

Realistic opacities

Three
-
dimensional

Time
-
dependent

Multi
-
wavelength

Includes spectropolarization

Treats radioactive decay


and gamma
-
ray transfer

Iterative solution for


thermal equilibrium

Non
-
LTE capability



Kasen et al 2006 ApJ

2D Deflagration Model

M
Ni

= 0.2 M
sun

E
K

= 0.3 x 10
51

ergs

Roepke, Kasen, Woosley

2D Delayed Detonation

M
Ni

= 0.5 M
sun

E
K

= 1.2 x 10
51

ergs

Roepke, Kasen, Woosley

The stronger the deflagration phase



the more pre
-
expansion



the lower the densities at
detonation



the less
56
Ni produced

Off
-
center Detonation

M
Ni

= 1.0 M
sun

E
K

= 1.3 x 10
51

ergs

Roepke, Kasen, Woosley

An alternative to

super
-
chandra SNe?

Howell et al, 2006

Hillebrandt, Sim, Roepke 2007

Spectrum of Off
-
center
Detonation

expansion velocities depend on orientation

Kasen (2006) ApJ

I
-
Band

Asymmetry and

Polarization

Model polarization
spectrum at maximum light

as seen from different

viewing angles


Transition to Detonation

Hot ash plumes surrounded by
the flame are buoyant. As
they rise, encountering lower
densities, shear gives rise to
turbulence, which cascades
to smaller length scales
where it affects the motion of
the flame, it thickens.





A critical length
-
scale in
turbulent combustion is the
Gibson scale
l
G

the scale at
which the flame can just
burn away a turbulent eddy
before it turns over

where
s
L

is the laminar flame speed, L is the integral
scale

and v'(L) is the turbulent intensity on that scale
(with assumed Kolmogorov scaling).

Simulating turbulence


At around 10
7

g cm
-
3
, the flame
becomes thick enough that
turbulent eddies can disrupt its
structure before they burn
away, that is, the flame
thickness is larger than the
Gibson scale.


At this point, the burning
fundamentally changes
character and the flame is said
to be in the distributed burning
regime.


3
-
D simulations showing the
distribution of nuclear energy
generation in turbulent carbon
fusion flames spanning:


The flamelet regime (0.3 m)
2


r

= 8x10
7
g/cc , u = 0.1 s
L


Transitional stage (0.3m)
2


r

= 3x10
7
g/cc , u = 1.8 s
L


The distributed regime (1.0 m)
2


r

= 1x10
7
g/cc, u = 70.0 s
L


where u is an imposed
turbulence level.


Q: In the distributed burning
regime, can a mixed region of
partially burned fuel and ash
grow large enough such that it
can ignite a detonation?








































Here the turbulence is
dominated by the
flame, which remains
fairly coherent and
burns in a similar way
to a flat laminar flame.
The
red line

is the locus
of a laminar flame at
the same density.

Turbulent disruption
of the flame leads to
thermodiffusively
stable behavior
expected of a high
Lewis number flame,
where regions of
negative and positive
curvature experience
greatly enhanced and
reduced burning
rates, respectively.

















Intense burning
regions and local
extinction are both
observed. The width
of the flame is slightly
increased, but the
overall burning rate
remains close to the
laminar value.

Turbulent mixing dominates over diffusive
processes shredding the flame. Its thickness
is greatly increased accompanied by a 5
-
fold
increase in burning rate.

We are currently
generating statistics that will further refine
the subgid model for our full star studies.

SNe Ia Highlights on ATLAS


Code development is nearly complete on
MAESTRO
, the low Mach
-
number code, and
CASTRO
, the compressible radiation
-
hydro code.
SEDONA
now has non
-
LTE capability
-

distributed MC in progress.
Full star 3D studies to begin in summer 08.



The light curves and spectra of a set of 1D and 2D models for Type
Ia supernovae were calculated. The physical origin of the WLR has
been determined. Significant variations in spectra and brightness as
a function of viewing angle for asymmetric explosions were
observed, which could explain the so called ``super
-
Chandrasekhar
mass Type Ia supernovae'’ for a single degenerate progenitor.



Turbulent nuclear combustion in the distributed regime has been
studied analytically and simulated. We see the broadening of the
flame by turbulence and have derived the necessary criteria for a
transition to detonation.

SNe & MAESTRO

low Mach number hydro codes

References


“Type Ia Supernovae”, Woosley et al. Journal of Physics: Conference
Series {
\
bf 78}, (2007) 012081


“The Light Curves and Spectra of Supernova Explosions: Multi
-
Dimensional
Time
-
Dependent Monte Carlo Radiative Transfer Calculations”, Kasen et al.
Journal of Physics: Conference Series 78, (2007) 012037


"Adaptive low Mach number simulations of nuclear flame microphysics", J.
B. Bell, M. S. Day, C. A. Rendleman, S. E. Woosley, and M. A. Zingale,
LBNL Report 52395, J. Comp. Phys, 195, 677
-
694, 2004.


"MAESTRO: A Low Mach Number Stellar Hydrodynamics Code", Almgren,
A.S., Bell, J.B., & Zingale, M., Journal of Physics: Conference Series 78,
(2007) 012085


SEDONA: "Time Dependent Monte Carlo Radiative Transfer Calculations
for 3
-
Dimensional Supernova Spectra, Lightcurves, and Polarization", D.
Kasen, R.C. Thomas, & P. Nugent, astro
-
ph/0606111 (2006) URL:
http://arxiv.org/abs/astro
-
ph/0606111