Neutrino Oscillations:
experimental triumphs,
remaining puzzles
Giorgio Gratta
Physics Department, Stanford
What we do know:
- Status of neutrino mixing
How did we learn it: -
Solar neutrino
Atmospheric neutrino
Reactor neutrinos
K2K
Addressing remaining puzzles in neutrino mixing:
- LNSD/miniBoone Æ see Sam Zeller
- θ13 experiments
- Hierarchy/CP violation: ν-fact/β-beams
- Absolute m, Majorana/Dirac ν
Æ see Steve Elliott
Impossible to be complete: I will just make an arbitrary
-and hopefully interesting- selection
SSI 2004
Experimental Neutrino Oscillations - Gratta
2
From LEP and SLC we know that there are
2.981±0.008 active, light ν flavors
We also know that 3 ν flavors were active in
Big Bang Nucleosynthesis
If mν≠0 then the 3 weak-interactions eigenstates
νj
can be represented as superposition of mass eigenstates
ν j = ∑ U jl ν l
j
Neutrinos are produced and detected by weak interaction
but propagate in vacuum as mass eigenstates…
ν j ≅ ∑ U lj e
l
−i ( ml2 / 2 E ) L
ν l ≅ ∑∑ U lj e
j′
U *j′l ν j′
−i ( ml2 / 2 E ) L
l
Æ Neutrino oscillations
SSI 2004
Experimental Neutrino Oscillations - Gratta
3
A possible representation of the
Pontecorvo-Maki-Nakagawa-Sakata matrix
⎛ U e1
⎜
U = ⎜ Uµ1
⎜U
⎝ τ1
⎛ cosθ 12
⎜
⎜ - sinθ 12
⎜ 0
⎝
U e2
Uµ 2
Uτ 2
U e3 ⎞
⎟
Uµ 3 ⎟ =
Uτ 3 ⎟⎠
sinθ 12
cosθ 12
0
Solar ν
KamLAND
Future solar
ν exp.
θ12~32°
SSI 2004
information on Neutrino
CP is here
factory ?
0 ⎞ ⎛ cosθ 13
⎟ ⎜
0⎟ • ⎜
0
⎜
1 ⎟⎠ ⎝ - e - iδ CP sinθ 13
0 e - iδ CP sinθ 13 ⎞ ⎛ 1
0
⎟ ⎜
1
0
⎟ • ⎜ 0 cosθ 23
⎟
0
cosθ 13 ⎠ ⎜⎝ 0 - sinθ 23
0 ⎞ ⎛1
0
⎟ ⎜
- iα / 2
sinθ 23 ⎟ • ⎜ 0 e
cosθ 23 ⎟⎠ ⎜⎝ 0
0
Chooz/Palo Verde
Off axis
Future reactors
Atmospheric ν
K2K
Minos et al
θ13<7° @ 90%CL
θ23~45°
Experimental Neutrino Oscillations - Gratta
⎞
⎟
⎟
e - iα/2 + iβ ⎟⎠
0
0
Neutrinoless
double beta
decay
4
Oscillations give the proof that neutrino masses
are non-zero, but we still do not know the absolute
mass scale and the hierarchy
~20 eV
solar~2.7·10-3 eV2
~8.3·10-5 eV2
~8.3·10-5 eV2
solar~2.7·10-3 eV2
From WMAP
From 0νββ if ν is Majorana
From tritium endpoint
(Maintz and Troitsk)
0.3 eV
~1 eV
Time of flight from SN1987A
(PDG 2002)
2.8 eV
?
Sign of ∆m213 (Future accelerator experiments)
SSI 2004
Experimental Neutrino Oscillations - Gratta
5
νe are abundant by-products of
nuclear fusion in the sun
p + p→ 2H + e + + ν e + 0.42 MeV
p + e − + p → 2H + ν e + 1.44 MeV
“pep” 0.25%
“pp” 99.75%
2
H + p → 3 He + γ + 5 . 49 MeV
86%
3
He+3He → α + 2 p + 12.86MeV
14%
3
He + α →7Be + γ + 1.59MeV
“7Be”
99.89%
7
Be + e − →7 Li + γ + ν e + 0.8617 MeV
7
Li + p → α + α + 17.35MeV
0.11%
7
Be + p→8B + γ + 0.14MeV
“8B”
0.11%
8
B→8Be + e + + ν e + 14.6MeV
8
SSI 2004
“hep” 2.4*10-5
3
He + p → α + e+ +ν e
Be → α + α + 3MeV
Experimental Neutrino Oscillations - Gratta
6
The detectors are located far away from the Sun
L=108 km
But neutrinos are produced
deep inside the Sun, where
most the fusion reactions
take place
Î
Matter effects
play an essential role
SSI 2004
Experimental Neutrino Oscillations - Gratta
7
Chlorine:
37Cl+ν Æ 37Ar+ee
Homestake mine (US) Taking data from 1967-2002 (!)
The “father” of all solar neutrino experiments
Gallium:
71Ga+ν Æ 71Ge+ee
SAGE (Baksan, Russia), Gallex and GNO (Gran Sasso, Italy)
Lowest threshold
Calibrated with a neutrino source
Cerenkov: e-+νeÆ e-+νe
Radiochemical,
no real-time information
Several experiments detecting solar neutrinos
Kamiokande and SuperKamiokande (Japan), SNO (Sudbury, Canada)
Largest statistics
See D.Casper’s talk
Cerenkov: Dν scattering
SNO (Sudbury, Canada)
For the first time measure νe and νX separately
See D.Waller’s talk
SSI 2004
Experimental Neutrino Oscillations - Gratta
8
Neutrino energy spectrum includes lines and continua
SSI 2004
Experimental Neutrino Oscillations - Gratta
9
Homestake Mine, Lead SD
1400 m underground
615 tons of perchloroethilene
(C2Cl4)
2.2*1030 atoms of 37Cl
36Ar or 38Ar added to the
fluid as carrier gas
SSI 2004
Experimental Neutrino Oscillations - Gratta
10
30 years of solar neutrinos with the Chlorine detector
Expected
from SSM
SSI 2004
Experimental Neutrino Oscillations - Gratta
11
1 SNU = 1 ν interaction/sec per 1036 target atoms
SSI 2004
Experimental Neutrino Oscillations - Gratta
12
:
d
l
ho
s
e ue !
r
th niq nos
r ch ri
e
t
e
low ly t neu
s n
a
h , o pp
a te to
71 G
da tive
to nsi
se
In Gallex/GNO 30.3 ton of
Ga in aqueous solution of
GaCl4. 71GeCl4 is volatile
and is purged/counted
28 - 50 ton of metallic
Ga kept liquid (>29.8 C)
Extraction by HCl etch and
GeCl4 and Ar purge
SSI 2004
Experimental Neutrino Oscillations - Gratta
SAGE, Baksan
13
SNU
Both Ga experiments have
been calibrated with a 51Cr
source of about 1 MCi
51Cr
source
SAGE data
measurements
Expected
from SSM
SSI 2004
Experimental Neutrino Oscillations - Gratta
14
In order to better understand the situation it would be
nice to also measure the
total number of neutrinos
from the Sun,
all flavors together
The SNO detector:
1 kton of heavy water
SSI 2004
Experimental Neutrino Oscillations - Gratta
15
Neutrino interactions in D2O
Charged Current (CC):
νe+d⇒p+p+eElectron neutrinos only
ed
p
p
Elastic Scattering (ES):
νx+ e- ⇒ νx +e0.154•σ(νe)=σ(νµ)=σ(ντ)
this is all SuperK sees
νx
νx
νe
ee-
Neutral Current (NC):
νx+d⇒p+n+νx
Same for all neutrino
flavors !
SSI 2004
νx
νx
Experimental Neutrino Oscillations - Gratta
d
p
n
16
Remarkably the SNO measurement gives ΦNC>ΦCC
Phys. Rev.Lett. 89 (2002) 011301
Energy Unconstrained
For up-to-date results
see Waller’s talk
So νe are missing, but the total
number of neutrinos is right !
SSI 2004
Experimental Neutrino Oscillations - Gratta
17
1 SNU = 1 ν interaction/sec per 1036 target atoms
SSI 2004
66.9 ± 5.3
69.3 ± 5.5
Experimental Neutrino Oscillations - Gratta
18
Clearly all this smells of flavor mixing…
Global fit to all solar neutrino exps
eliminates most of the solutions
except for the MSW-LMA solution
(The solar neutrino anomaly is
a sort of neutrino-birefringence
of dense matter occurring because
neutrinos have finite masses
-- see Boris Kayser’ talk)
SSI 2004
Experimental Neutrino Oscillations - Gratta
19
All of this is very interesting…
…but wouldn’t it be great if we could
reproduce it with artificial means ?
SSI 2004
Experimental Neutrino Oscillations - Gratta
20
Nuclear reactors are very intense sources of νe deriving
from beta-decay of the neutron-rich fission fragments
n
N
n
νe
νe
νe
νe
r
a
e tor
l
c
nu eac
r
νe
SSI 2004
νe
νe
N2
n N’2
n
N1
N’1
e-
νe
Look for a deficit of νe and spectral
distortions at a distance L
νe
νx ?
L
νe
Experimental Neutrino Oscillations - Gratta
νe detector
21
Example:
235
92
U + n → X 1 + X 2 + 2n
235U
fission
94
140
stable nuclei with A
most likely from fission
94
40
Zr
140
58
Ce
together these have
98 protons and
136 neutrons
so, on average 6 n have to
decay to 6 p to reach stable matter
SSI 2004
Experimental Neutrino Oscillations - Gratta
22
Power/commercial reactors are generally used
since only requirement is to have large power
200 MeV / fission
6ν e / fission
A typical large power
reactor produces
3 GWthermal and
6•1020 antineutrinos/s
SSI 2004
Experimental Neutrino Oscillations - Gratta
the Chooz plant in France
23
>99.9% of ν- are
produced by fissions in
235U, 238U, 239Pu, 241Pu
contribution <10-3
not taken into
account for neutrino
flux calculations
SSI 2004
{
Experimental Neutrino Oscillations - Gratta
24
A specific signature is provided by the inverse-β reaction
ν e + p → e+ + n
τ ≈ 200 µs
p + n → d + γ (2.2 MeV )
Event tagging by coincidence in time,
space and energy of the neutron capture
Eν measurement
Eν ≅ Te + + Tn + ( M n − M p ) + me +
{ {
Threshold: Eν>1.8MeV
10-40 keV
1.8 MeV
→ only ~1.5 antineutrinos/fission can be detected
SSI 2004
Experimental Neutrino Oscillations - Gratta
25
The νe energy spectrum
Observed spectrum (a.u.)
νe+p→n+e+ cross
-43 cm2)
section (10-43
Reactor νe spectrum (a.u.)
Eν (MeV)
SSI 2004
Experimental Neutrino Oscillations - Gratta
26
An experiment
relevant for
solar ν will have
~100 km baseline
and hence will
need a huge
detector and a
“huge source”
SSI 2004
e
l
ca
i
st
i
t
a
St
Experimental Neutrino Oscillations - Gratta
s
r
o
rr
ly
n
o
27
~1 km high
Mt Ikenoyama
SSI 2004
Experimental Neutrino Oscillations - Gratta
28
Many reactors contribute to the antineutrino flux at KamLAND
Japan
South Korea
Kashiwazaki
Ohi
Takahama
Tsuruga
Hamaoka
Mihama
Sika
Fukushima1
Fukushima2
Tokai2
Onagawa
Simane
Ikata
Genkai
Sendai
Tomari
Ulchin
Cores
(#)
Ptherm
(GW)
Flux
(cm-2 s-1)
Rate noosc*
(yr-1 kt-1)
160
179
191
138
214
146
88
349
345
295
431
401
561
755
830
783
712
7
4
4
2
4
3
1
6
4
1
3
2
3
4
2
2
4
24.3
13.7
10.2
4.5
10.6
4.9
1.6
14.2
13.2
3.3
6.5
3.8
6.0
10.1
5.3
3.3
11.5
4.1·105
1.9·105
1.2·105
1.0·105
1.0·105
1.0·105
9.0·104
5.1·104
4.8·104
1.6·104
1.5·104
1.0·104
8.3·103
7.8·103
3.4·103
2.3·103
9.9·103
254.0
114.3
74.3
62.5
62.0
62.0
55.2
31.1
29.5
10.1
9.3
6.3
5.1
4.8
2.1
1.4
6.1
4.8
Yonggwang
986
6
17.4
7.8·103
Kori
735
4
9.2
7.5·103
4.6
Wolsong
709
4
8.2
7.1·103
4.3
Total Nominal
SSI 2004
Dist
(km)
-
70 181.7 1.3·106
Experimental Neutrino Oscillations - Gratta
*E >3.4MeV
ν
(Eprompt>2.6MeV)
Detailed power and fuel
Composition calculation used
Site
From electrical
power
Japanese average
fuel used
803.8
29
The total electric power produced “as a
by-product” of the νs is:
•~60 GW or...
•~4% of the world’s manmade power or…
•~20% of the world’s nuclear power
SSI 2004
Experimental Neutrino Oscillations - Gratta
30
A limited range of baselines contribute to the flux
of reactor antineutrinos at Kamioka
Over the data period
Reported here
Korean reactors
3.4±0.3%
Rest of the world
+JP research reactors
1.1±0.5%
Japanese spent fuel
0.04±0.02%
SSI 2004
Experimental Neutrino Oscillations - Gratta
31
KamLAND:
Kamioka Liquid scintillator
AntiNeutrino Detector
•1 kton liq. Scint. Detector
in the Kamiokande cavern
•1325 17” fast PMTs
•554 20” large area PMTs
•34% photocathode coverage
•H2O Cerenkov veto counter
SSI 2004
Experimental Neutrino Oscillations - Gratta
32
The completed detector, looking up
SSI 2004
Experimental Neutrino Oscillations - Gratta
33
Scintillator is a blend of
20% pseudocumene and
80% dodecane
Different density
paraffines are used
to tune the density of
buffer to 4·10-4 of
that of the scintillator
PPO concentration is
1.52 g/l in scintillator
SSI 2004
Experimental Neutrino Oscillations - Gratta
34
Autoradiography of KamLAND
In
8U
23
re
su
ea 8 g/g
n m -1
gio
10
re
5)x
is
0.
th
5±
(3.
=
Top flange and
chimney region
Central thermometer
Balloon
Bottom flange
and plumbing
SSI 2004
Experimental Neutrino Oscillations - Gratta
35
Live time
(days)
Jan 2002
-
Run A
(data-set of 1st paper)
Mar 9 – Oct 6
2002
145.4*
Electronics upgrade & 20”
PMT commissioning
Jan/Feb 2003
-
Run B
Oct - Jan 11
2004
369.7
Data-set presented here†
Mar 9, 2002 –
Jan 11, 2004
515.1
Start data taking
†
*Was
Dates
145.1 with
old analysis
A brief history of KamLAND
T.Araki et al, arXiv:hep-ex/0406035 Jun 13, 04
submitted to Phys Rev Lett
SSI 2004
Experimental Neutrino Oscillations - Gratta
36
Vertexing is performed using timing from the 17” PMTs
Am/Be(~8MeV)
-60 (2.6MeV)
-65 (1.1MeV)
-68 (1.0MeV)
z
SSI 2004
Experimental Neutrino Oscillations - Gratta
37
Tagged cosmogenics can be used for calibration
12B
τ=29.1ms
Q=13.4MeV
12N
τ=15.9ms
Q=17.3MeV
µ
Fit to data shows that
12B:12N ~ 100:1
SSI 2004
Experimental Neutrino Oscillations - Gratta
38
Energy calibration uses discrete γ and
12B/12N
n-p
68Ge
60Co
n-12C
65Zn
Carefully include Birks law, Cherenkov
to obtain constants for γ and
and light absorption/optics
e–type depositions
σ/E ~ 6.2% at 1MeV
SSI 2004
Experimental Neutrino Oscillations - Gratta
39
Estimate of total volume and fiducial fraction
SSI 2004
Experimental Neutrino Oscillations - Gratta
40
Fraction of volume inside the fiducial radius verified
using µ-produced 12B/12N and n (assumed uniform)
12B/12N
neutrons
SSI 2004
Experimental Neutrino Oscillations - Gratta
41
Selecting antineutrinos, Eprompt>2.6MeV
- Rprompt, delayed < 5.5 m
- ∆Re-n < 2 m
.7
3
54
n
to
5.5 m
fiducial cut
- 0.5 µs < ∆Te-n < 1 ms
- 1.8 MeV < Edelayed < 2.6 MeV
- 2.6 MeV < Eprompt < 8.5 MeV
Tagging efficiency 89.8%
Balloon edge
…In addition:
- 2s veto for showering/bad µ
- 2s veto in a R = 3m tube along track
Dead-time 9.7%
SSI 2004
Experimental Neutrino Oscillations - Gratta
42
Systematic
%
Scintillator volume
2.1
Fiducial fraction
4.2
Energy threshold
2.3
Cuts efficiency
1.6
Live time
0.06
Reactor Pthermal
2.1
Fuel composition
1.0
Time lag
0.01
Antineutrino spectrum 2.5
Antineutrino x-section 0.2
Total
6.5
SSI 2004
Experimental Neutrino Oscillations - Gratta
43
Results
766.3 ton·yr
Observed events 258
No osc. expected 365±24(syst)
Background
7.5±1.3
Inconsistent with simple 1/R2 propagation
at 99.995% CL
(Observed-Background)/Expected = 0.686±0.044(stat)±0.045(syst)
Caveat: this specific number does not have an absolute meaning in KamLAND,
since, with oscillations, it depends on which reactors are on/off
SSI 2004
Experimental Neutrino Oscillations - Gratta
44
2003 saw a substantial dip in reactor antineutrino flux
SSI 2004
Experimental Neutrino Oscillations - Gratta
45
Good correlation with reactor flux
90% CL
ed
t
ec
p
Ex
r
fo
no
ns
o
ti
a
ill
c
os
Fit constrained
through known
background
χ2=2.1/4
~0.03 for
3TW
hypothetical
Earth core
reactor
(But a horizontal line still gives a decent fit with χ2=5.4/4)
SSI 2004
Experimental Neutrino Oscillations - Gratta
46
Energy spectrum now adds substantial information
Best fit to
oscillations:
∆m2=8.3·10-5 eV2
sin22θ=0.83
Straightforward
χ2 on the histo
is 19.6/11
Using equal
probability bins
A fit to a simple rescaled reactor spectrum
is excluded at 99.89% CL (χ2=43.4/19)
SSI 2004
Experimental Neutrino Oscillations - Gratta
χ2/dof=18.3/18
(goodness
of fit is 42%)
47
Un-binned likelihood fit to 2-flavor oscillations
All solar ν
experiments
LMA2 excluded
at 99.6% CL
∆m2=8.3·10-5 eV2
sin22θ=0.83
“LMA0” disfavored
at 94% CL
SSI 2004
Experimental Neutrino Oscillations - Gratta
48
A shape-only fit gives
similar results
∆m2=8.3·10-5 eV2
sin22θ=0.98
SSI 2004
Experimental Neutrino Oscillations - Gratta
49
KamLAND uses a range of L and
it cannot assign a specific L to each event
Nevertheless the ratio of detected/expected
for L0/E (or 1/E) is an interesting quantity, as it decouples
the oscillation pattern from the reactor energy spectrum
no oscillation expectation
Hypothetical
single 180km
baseline
experiment
SSI 2004
Experimental Neutrino Oscillations - Gratta
50
More exotic, non-oscillations models for the antineutrino
channel start being less favored by data
Decay*
excluded at
95% CL
Decoherence†
excluded at
94% CL
*V.Barger et al. Phys. Rev. Lett. 82 (1999) 2640
†E.Lisi
SSI 2004
et al., Phys. Rev. Lett. 85 (2000) 1166
Experimental Neutrino Oscillations - Gratta
51
Combined solar ν – KamLAND 2-flavor analysis
0.6 × 10 −5 eV 2
∆m122 = 8.2 +
− 0 .5
0.09
tan 2 θ12 = 0.40 +
− 0.07
Includes (small) matter effects
SSI 2004
Experimental Neutrino Oscillations - Gratta
52
…it took 35 years… but we have:
1)
2)
3)
4)
detected neutrinos from the sun
found that part of the neutrino flux is missing
found that only νe are missing
found that matter effects in the sun (MSW effect)
are responsible for the deficit
5) found that neutrinos have finite masses (required
for MSW)
6) identified mixing parameters ∆m122, θ12
7) found that we can reproduce this effect artificially
with vacuum oscillations with the same ∆m122, θ12
8) found that ν and anti-ν behave the same way
9) found that flavor mixing works better than other
scenarios (e.g. decay)
10) found that our model for the sun (SSM) works
very well
SSI 2004
Experimental Neutrino Oscillations - Gratta
53
The other
oscillations:
Atmospheric
neutrinos
Higher energy
phenomenon
Measurements
dominated by
SuperK
SSI 2004
Experimental Neutrino Oscillations - Gratta
54
SuperKamiokande:
the largest “artificial”
water Cerenkov detector
ever built (50 kton)
SK1 1996-2001
>11146 50cm PMTs
40% photocathode
22.5kton fiducial mass
2003-2005
5182 PMTs recovered
19% photocathode
K2K beam
SSI 2004
2006-…
Original PMT coverage
T2K beam
Experimental Neutrino Oscillations - Gratta
55
A composite data-set
E.Kearns,
Nu2004
SSI 2004
Experimental Neutrino Oscillations - Gratta
56
SSI 2004
E.Kearns,
Nu2004
Experimental Neutrino Oscillations - Gratta
57
SSI 2004
e
µ
Experimental Neutrino Oscillations - Gratta
58
E.Kearns, Nu2004
The oscillatory pattern is now also
visible with atmospheric neutrinos
And flavor oscillation
is favored over other
models for
disappearance
SSI 2004
Experimental Neutrino Oscillations - Gratta
59
E.Kearns, Nu2004
Parameter regions fit by different channels nicely overlap
A 3 flavor analysis
favors νµ-ντ
oscillation as
responsible for
the atmospheric
neutrino anomaly
(more on this
later…)
SSI 2004
E.Kearns,
Nu2004
Experimental Neutrino Oscillations - Gratta
60
Again, the quest is open to observe the oscillation
phenomenon with artificial neutrinos
• 12 GeV KEK protons
• Mini-K (1kton water + other detectors)
near detector (100m) 1011 νµ every 2.2s
• Super-K at 250km 106 νµ every 2.2s (detect ~1event/2days)
• 8.9·1019 p.o.t. collected in ~3.5 yrs
SSI 2004
Experimental Neutrino Oscillations - Gratta
61
Although rate is low
background greatly
reduced by beam spill
timing (0.83 ms delay,
synchronized offline
using GPS)
104
103
102
10
1
In the 3.5 yrs of data
the atm nu background
in the KEK beam spill
window is 10-2 events
15
10
5
0
-4
-2
0
2
4
Tdiff(µs)
T.Nakaya,
SSI 2004 Nu2004
Experimental Neutrino Oscillations - Gratta
62
11.6
Expected no oscillation rate: 150.9 +−10
.0
K2K establishes:
no oscill,
floating
normalization
T.Nakaya,
SSI 2004 Nu2004
• νµ disappearance at 2.9σ
(standard physics
has 0.33% CL)
• Eν spectral distortion at 2.5σ
(standard physics
has 1.1% CL)
Combined we have a 3.9σ effect
(standard physics
has 0.011% CL)
Experimental Neutrino Oscillations - Gratta
63
approximate
90% CL
from global
SK fit to
atmospheric
neutrinos
T.Nakaya,
SSI 2004 Nu2004
Experimental Neutrino Oscillations - Gratta
64
What we have discovered:
There is mixing involving νe with best fit at
∆m2=8.3·10-5 eV2 and θ=32° (solar/KamLAND)
There is also mixing involving νµ with best fit at
∆m2=2.7·10-3 eV2 and θ=45° (atmospheric/K2K)
If there are only 3 neutrino families, since the two
mass differences are of different order of
magnitude, we expect to find a third ∆m2 similar
to the larger of the two above
SSI 2004
Experimental Neutrino Oscillations - Gratta
65