Transcript n,n - Osaka University

SUPERNOVA NEUTRINOS
AND THE
r- AND n-PROCESSES OF
NUCLEOSYNTHESIS
Richard N. Boyd
Ohio State University
NDM03, June, 2003
1. Understanding the stellar core collapse
process with SN neutrinos; black holes.
2. Understanding neutrinos from their
supernova signatures.
3. r-process and n-process nucleosynthesis.
4. Detecting nucleon decay.
5. OMNIS, the Observatory of Multiflavor
Neutrinos from Supernovae
STAGES OF STELLAR EVOLUTION
During each stage, each shell of a star establishes
hydrostatic equilibrium between gravity and the
energy it produces. When that stage’s fuel is gone,
the star contracts, converts gravitational potential
energy to heat, and burns the next fuel. When it’s
gone, the star contracts, ...
Stage
H
T9 r(g/cm3) t(y)
0.02 102
107
He
0.2
104
105
C
0.8
105
103
Ne
O
1.4
2.0
107
107
3
.8
Si
3.5
108
1 w.
1014
1 d.
Collapse ~40
Reactions
pp-Chains
CNO
34He12C
12C(a,g)16O
12C+12C
20Ne+a
20Ne(g,a)16O
16O+16O
28Si+a, 31P+p, ...
28Si(g,a)24Mg, ...
28Si(a,g)32S, ...
g+xn,p,a
Ashes
He, ...
C, O, ...
Ne, O, ...
O, Mg, ...
Si, S, ...
Fe, Ni
NSE
n’s,p’s,a’s
NSE: Nuclear Statistical Equilibrium; it makes everything up
to mass 100 u in < 1s out of n’s and a’s.
BASICS OF COLLAPSE PROCESS
Gravitational binding energy:
EGrav  (3/5) GMNS2/RNS
 3x1053 ergs (MNS/1.4Mo)2 (10 km/RNS)
Neutrino Diffusion time: tn ~ 3 s
Typical Luminosity per Neutrino Species:
Ln ~ (GMNS2/RNS)/1/tn ~ 1x1052 ergs/s
Epochs of Supernova Neutrino Emission:
I. Infall:
Principal n emission, high energy ne from
e- + p  n + ne
II. Shock ReHeating/Explosion:
Thermal emission from neutron star surface of ne, nm, and
nt in roughly equal fluxes.
~ Fermi-Dirac black body spectra. (?)
Ln  1052 ergs/s per flavor.
<E(ne)>  11 MeV, <E(ne)>  16 MeV, <E(nx)>  25 MeV.
III. Post-Explosion, r-Process Epoch:
Time scale ~ 10 s.
Neutron star contracts from 40 km to 10 km.
Low Ln’s, but <E>’s slowly rising. (?)
NEUTRINOS FROM
STELLAR COLLAPSE
The energy in the core is (a few)x1053 ergs; most of
it is emitted in a few seconds, ultimately to produce
a stable neutron star.
The reactions:
p + e-  n + ne (neutronization spike)
n + e+  p + ne (URCA process)
g  e+ + e e+ + e-  g + g
e+ + e-  ni + ni (once in 1019 times)
A*  A + ni + ni
Nearly all of the neutrinos emitted are from the
last process; ~ 1053 ergs of them in the order of a
second (as seen in SN 1987a).
After the neutronization spike, E(ne)E(nm)E(nt),
but the mean energies are NOT the same; <E(ne)>
< <E(ne)> < <E(nm,t)>.
DETECTING SUPERNOVA
NEUTRINOS--WHY?
1. Checking the Standard Model of core collapse and cooling of the
protoneutron star.
Are cooling times consistent with prediction? A low-entropy core
collapse? Neutrino opacities correct?
Neutrino energies correct? <Em,t>  25 MeV, <Ee>  16 MeV,
<Ee>  11 MeV? Or might other mechanisms affect <Em,t>? Are the
distributions Fermi-Dirac? Might oscillations affect these? Are there
signatures of rotation (mixing)?
2. Neutrino physics.
Measure neutrino masses from their time of flight? Or from
timing signals from collapse to a black hole. But need to detect all
the flavors! And mixing could confuse this.
Measure/check some types of oscillations: might measure 13
with incredible sensitivity.
3. Nucleosynthesis.
Measure the neutrino spectra—they’re crucial for understanding
the r-process and the n-process.
4. Black hole astrophysics.
L
Observe the collapse to a black
hole via the abrupt termination of
the neutrino signal.
Do diagnostics on the black hole
t
collapse process via differences in
termination times of signals from different neutrino flavors?
MODIFICATIONS ON THE
“STANDARD MODEL”
1. Atmospheric n observations from Super-K, Soudan 2, and
MACRO:
nm  nt vacuum oscillations. (They’re nx either way.)
2. Solar n observations from Super-K and SNO:
ne  nm MSW transitions.
As the n’s emerge from their n-spheres near the star’s core:
The nm’s, nm’s, nt’s, and nt’s have high mean energies.
As the n’s emerge from the periphery of the star:
The ne’s, nm’s, nt’s, and nt‘s have high energies.
nm
En
sin2212 = 0.8
ne
MSW region
Log R
These transformations will produce mostly high energy ne’s.
Furthermore, a non-zero 13 (which is very difficult to
measure) could produce an additional enhancement in the
high energy ne’s.
RESULT: the ne’s detected by OMNIS will be high-energy, so
reflect the energy distribution of the nm’s and nt’s emitted
from the core of the star. This results from OMNIS’s lead’s
selectivity, via its thresholds, to only high-energy neutrinos.
A SPECTRAL MODIFICATION:
”PINCHING” THE DISTRIBUTION
The neutrino distributions are generally assumed to be
Fermi-Dirac:
fn = [Tn3 F2()]-1 En2 [exp(En/Tn - ) + 1]-1
where  = chemical potential/Tn, and F2() = a normalization
factor. Tn = the temperature at the n-sphere.
But the n-sphere is determined
by the cross sections, which are
larger for higher E ns.
Thus (Raffelt) the n-sphere for
higher-E n’s is at larger r, hence
T
lower Tn, than it is for lower E
ns. Also, the low-E n’s will have
T>
their n-sphere at a smaller r,
T<
hence higher T. So both ends of
the distribution will be “pinched”
toward the center. And you can’t f’
n
infer the spectrum by sampling
2-3 points! Furthermore, since
n-induced cross sections ~ E2, the
high energy tail can be very
important to some processes!
Higher E
n-sphere
Lower E
n-sphere
r
En (MeV)
For ne + p  e+ + n
(CC Interaction)
in Super-Kamiokande
DETECTING NEUTRINOS FROM
SUPERNOVAE--HOW?
nes and nes interact via both charged-current and
neutral-current interactions. s for the former is
larger, so use
ne + p  e+ + n, and
n e + O  e+ + N
Can use Cerenkov detectors in water for this-SuperK, SNO.
m- and t-neutrinos (at SN energies) interact only
via the NC interaction, but OMNIS can detect
them by observing neutrons from
nm,t + 208Pb  208Pb* + nm,t’

207Pb + n (Q = -7.4 MeV)
206Pb + 2n (Q = -14.1 MeV)
The relative yields of 1n to 2n events test the
energy distribution, the NC interaction doesn’t
provide a direct way to measure neutrino energies.
The NC interactions also have a zero-neutron
mode that emits a distinguishable g-ray.
Detecting Neutrinos from SNe--More on How?
Electron neutrinos can also undergo CC interactions:
ne + 208Pb  e- + 208Bi*

207Bi + n (Q=-9.77 MeV)
206Bi + 2n (Q=-17.86 MeV)
With lead perchlorate (a clear liquid that dissolves
easily in water) one can detect the e-, providing a direct
measurement the energy of the ne’s from E(e-) and the
number of neutrons emitted. In most cases (except the
no-neutron NC case), detect neutrons.
Need several thousand events to do a decent statistical
analysis; 2 kT Pb slabs (2000 events) + 1.0 kT Pb[ClO4]2
(1000 events) will do that.
The CC events will measure the distribution of the high-energy n’s as they were
emitted; the (two-threshold) NC events
will provide a consistency check of the
oscillation modes and a sensitive measure
of the high-energy tails of the distributions.
Energy
Super-K will observe some NC ns from the O in the
H O, a very useful additional (high threshold) datum.
MEASURING NEUTRINO ENERGIES
IN CC INTERACTIONS
First
forbidden
states
E,
MeV
25
20
G-T states
2-n thd.
IAS
15
2-n emitting
transitions
1-n thd.
1-n emitting
transitions
10
5
208Bi
0
208Pb
NUCLEAR PHYSICS OF
NEUTRINO DETECTION: Pb
//////////
2n+206Pb, Q=-14.1 MeV
////////////
//////////////
n+207Pb, Q=-7.4 MeV
//////////////
____
____
n+207Pb threshold
n+Pbn’+Pb*Pb’+n’
(NC Interaction)
(n,n’)
But also, for 208Pb
208Pb
//////////
//////////////
2n+206Bi, Q=-17.9 MeV
n+207Bi, Q=-9.8 MeV
//////////////
n+207Bi threshold
208Pb(n,e-)208Bi*
ne+Pbe-+Bi*
e-+Bi’+n
(CC Interaction)
208Bi
208Pb
CALIBRATING OMNIS?
Using the neutrino beams from a
stopped pion facility isn’t perfect;
the neutrinos aren’t the right energy
(which we don’t know!).
So, use 208Pb(3He,t)208Bi reaction
(Fujiwara et al.). Identify the
transitions to the states of interest by
their angular distribution, and
measure the neutrons that they emit.
This is also very important
information; it determines the
detection efficiency!
How to Detect ne’s? And nx’s?
Two types of detection schemes.
1. Use vertical lead slabs alternated
with planes of neutron detectors.
n + Pb interactions produce n’s via
NC and CC interactions:
nx + APb  nx’ + A-1Pb + n,
nx
n
ne + APb  e- + A-1Bi + n.
n
The neutrons escape the lead slabs
and are detected when they thermalize and are captured in the
neutron detectors.
These detectors produce lots of events and some E information.
2. Use lead perchlorate (a clear
liquid). NC interactions again
ne
produce neutrons, which are
captured on the Cl. The e- from
the CC interactions produce
Cerenkov radiation, which
eidentifies the CC event, and
gives the energy of the incident
ne. Also neutrons. Only neutrons
means it’s a NC event.
These detectors produce the NC
to CC event ratio and measure the high E ne, hence nx, spectrum.
A Site for OMNIS?
Site I: Waste Isolation Pilot Plant, Carlsbad, NM
Nuclear waste repository. This is in a salt deposit,
2000 feet underground. Drifts and much infrastructure exist,
and waste is distant from where OMNIS would be, so is not
an issue. WIPP will be there for a very long time!
And the WIPP has been very supportive, providing
much infrastructure support. Furthermore, everything we
would need is already in place!
We plan to begin building OMNIS in the WIPP.
Site II: Deep Underground Science & Engineering Lab
DUSEL would be NSF supported. It is strongly
supported by the physics community, but isn’t a reality yet.
It’s support structure is unknown, but would be expected to
be similar, at least for our purposes, to that of the WIPP.
It would be as deep as 8000 feet. The nucleon decay
studies would require a deep site.
Site III: Boulby Mine, UK.
This is also a salt deposit, depth comparable to that
of WIPP. This is currently being used for dark matter
searches, but OMNIS group in UK wants to put an OMNIS
component there too.
Why multiple sites?
Coincidences required for REAL supernova events!
DUSEL
Deep Underground Science and Engineering Lab,
Homestake Mine, Lead, South Dakota
DIAGNOSING STELLAR COLLAPSE
Stellar collapse depends on hydrodynamics, the EOS, and the
interactions between the n’s and the nuclei in the collapsing
region of the star. SN1987A confirmed that the collapse is
low-entropy; the n’s took seconds to get out rather than the
tens of 10 ms in which they’re produced.
<E(ne)>= 11 MeV, <E(ne)>=16 MeV, and <E(nm,t)>=<E(nm,t)>
=25 MeV?? But the neutrino energy and time distributions
could be affected by:
Neutrino transport
Equation of state
1-D vs. 2-D vs. 3D
Hydrodynamics, e.g., turbulence
Pinching of energy distributions by scattering (inevitable?)
Neutrino Bremsstrahlung (n + n  n + n + ni + ni)
Neutrino inelastic scattering (n + n + n  n’ + n + n)
Convection (!)
Neutrino oscillations (!)
In addition, the time distributions (especially late time) could
exhibit interesting features such as:
Schirato-Fuller anomalies from neutrino oscillations.
Reddy late-time spikes from phase transition from neutron
matter to quark matter (?).
Cutoff from collapse to a black hole? Would all flavors
terminate at the same instant?
Supernova Neutrinos
and the r-Process
The r-process makes ~half the nuclides heavier than iron, and
all nuclides heavier than 209Bi.
It is thought to occur in ~second in the bubble just outside the
nascent neutron star, in a hot n wind.
The r-process requires a neutron density ~1020 cm-3 in order
to have it go fast enough to circumvent some short-lived
nuclides (it must get to Uranium).
PROBLEMS:
1. The nes will tend to equilibrate the neutrons and protons;
that will kill the r-process.
2. They also make 3H and 3He via n+4He, which then capture
4He’s to make 7Li and 7Be, which then can make too many
light nuclei to seed the r-process.
SOLUTIONS:
The r-process would work if one could have neutrino
oscillations involving a sterile neutrino (e.g., Caldwell, Fuller,
Qian mass scheme).
Or perhaps the energy spectral differences might solve the
problem?
REQUIREMENT:
MUST know the energy spectra!
Supernova Neutrinos and the n-Process
The n-process is thought to make some of the rarest
nuclides in the periodic table, e.g., 138La and 180Ta. It
must occur in the n-wind from a collapsing core in a
supernova.
138La
from 139La(n,n)138La and 138Xe(ne,e-)138La.
180Ta
from 181Ta(n,n)180Ta and 180Hf(ne,e-)180Ta. And half
come from 181Ta(g,n)180Ta.
But, also, 19F from, e.g., 20Ne(n,n)19Ne 19F.
And 7Li from, e.g., 4He(n,n)3He + 4He(3He,g)7Be7Li.
“Satellite yields” just below
the r-process abundance
peaks suggest n-processing
at the end of the r-process,
supporting this model of
the n-process (Haxton et
al.; Qian et al.).
.6
.5
.04
.4
.02
.3
0
.2
.1
0
170
180
190
200 A
The actual n-spectrum is crucial to the predictions of nprocess models. It’s uncertain at present, but OMNIS
will provide this.
Detecting Nucleon Decay in OMNIS
Nucleon decay:
Tests most fundamental theories of particle physics.
Has been looked for in large underground detectors.
Has good experimental limits for decay modes that emit
charged particles, especially Cerenkov light.
Has much poorer limits for processes that don’t emit charged
or strongly interacting particles.
Lead perchlorate: Pb[Cl O4]2:
VERY soluble in water--will give Cerenkov light. But what
would Pb and Cl do for nucleon decay observations?
nn+n+n
Decay of a neutron in 35Cl  34Cl*  34Cl  34S + b+.
x x
x x
xxxxxx
x
x x
xxxxxx
x x
x x x x
x x
x x x x
x
x
N
x
x
Z
3n
g
b+
x x
x x
xxxxxx
x
x x
xxxxxx
x o
x x
xxxxxx
x o
x x x x
x x
x x x x
x x
x x x x
x
x
N
x
x
Z
x
x
N
x
x x
xxxxxx
x x
x x x x
x
x
Z
Signature: Slow coincidence between g-rays from
deexcitation, then b+ from 34Cl  34S (T1/2 = 1.5 s). Then
determination of correct 34Cl lifetime! T1/2 = 1029y/year.
34Cl
Most troublesome background: High energy atmospheric
neutrino causing 35Cl  34Cl* + n. Use n for a veto!
Nucleon Decay in Lead of Lead Perchlorate
Signature of nucleon decay in 208Pb:
208Pb  207Pb* + 3n  206Pb* + n
 205Pb* + n (+ n)
 205Pb + g + 2n
(and other branches are possible). In general,
208Pb  208-jPb + 3n + (j-1)n + g.
In the LPC, multiple neutrons would be observed, and gs in
excess of 3 MeV would be observed.
Obvious background;
n + 208Pb*  207Pb* + n
…
 205Pb + g + 3n.
This is identical to neutron decay except that there’s an extra
neutron. This would be VERY difficult to identify. But could
measure the same process with nes, convert NC to CC cross
sections, and infer the background to subtract.
Should be able to achieve a half-life limit of 1030 y/y
(Boyd, Rauscher, Reitzner, Vogel)
DO SUPERNOVAE AND/OR
BLACK HOLES EVER HAPPEN?
Rate of Core-Collapse SN:
van den Bergh: 3  1 per century
Strom, Hatano et al., The historical SNe
were in a few % of the galaxy; it’s 5-10
per century.
Bahcall and Piran; Arnett, Schramm, and
Truran: 10 per century.
Large UG detectors over past 20 years: 10
per century is excluded at 85%
confidence limit.
How many Core-Collapse SNe produce black
holes?
Bahcall & Piran; Ratnatunga and van den
Bergh: 1 b.h./4 n-stars
Bethe and Brown: 1 b.h./1 n-star
Qian, Vogel, &Wasserburg: 9 b.h./1 n-star
Both areas clearly need more work!
COMPARING OMNIS TO OTHER
SN-NEUTRINO DETECTORS
FOR AN 8 kpc DISTANT SN
10000
1000
100
OMNIS-LP
OMNIS-PB
SNO-D
SNO-H
10
Super-K
nm+ nm+nt+nt
ne
ne (=nm+nt)