Search for m eg with the MEG experiment at PSI:
results and prospects
A.M. Baldini March 23rd
for the MEG collaboration
Hep-ex:0908.2594v1 18 Aug 2009
Most recent m+  e+ g Experiments (Hincks and Pontecorvo 1948)
Lab.
Year
Upper limit
Experiment or Auth.
PSI
1977
< 1.0  10-9
A. Van der Schaaf et al.
TRIUMF
1977
< 3.6  10-9
P. Depommier et al.
LANL
1979
< 1.7  10-10
W.W. Kinnison et al.
LANL
1986
< 4.9  10-11
Crystal Box
LANL
1999
< 1.2  10-11
MEGA
PSI
~2012
~ 10-13
MEG
Two orders of magnitude improvement
tough experimental challenge!
2
Signal and background
signal
meg
background
accidental
menn
physical
e+ m+
g
megnn
n
qeg = 180°
Ee = Eg = 52.8 MeV
e+ m+
n
Te = Tg
g
megnn
ee  g g
eZ  eZ g
n
n
e+ m+
g
3
The sensitivity is limited by the accidental background
nsig  Rμ , nphys.b.  Rμ , nacc.b.  Rμ2
The n. of acc. backg events (nacc.b.) depends quadratically on the muon rate
and on how well we measure the experimental quantities: e-g relative
timing and angle, positron and photon energy
Effective BRback (nback/Rm T)
BRacc  Rμ  Δteγ  Δθeγ2  ΔE e  ΔE γ2
Integral on the detector
resolutions of the Michel and
radiative decay spectra
4
Required Performances
BR (meg)  10-13 reachable
BRacc.b.  2 10-14 and BRphys.b.  0.1 BRacc.b. with the following
resolutions
FWHM
Exp./Lab
Year
DEe/Ee
(%)
DEg /Eg
(%)
Dteg
(ns)
Dqeg
(mrad)
Stop rate
(s-1)
Duty
cyc.(%)
BR
(90% CL)
SIN
1977
8.7
9.3
1.4
-
5 x 105
100
3.6 x 10-9
TRIUMF
1977
10
8.7
6.7
-
2 x 105
100
1 x 10-9
LANL
1979
8.8
8
1.9
37
2.4 x 105
6.4
1.7 x 10-10
Crystal Box
1986
8
8
1.3
87
4 x 105
(6..9)
4.9 x 10-11
MEGA
1999
1.2
4.5
1.6
17
2.5 x 108
(6..7)
1.2 x 10-11
MEG
2012
0.8
4
0.15
19
2.5 x 107
100
1 x 10-13
5
Need of a DC muon beam
Experimental method
Detector outline
g
Stopping Target
Muon Beam
e+
Stopped beam of 3 107 m
/sec in a 150 mm target
2.
Solenoid spectrometer &
drift chambers for e+
momentum
3.
Scintillation counters for
e+ timing
4.
Liquid Xenon calorimeter
for g detection
(scintillation)
Liq. Xe Scintillation
Detector
Liq. Xe Scintillation
Detector
Thin Superconducting Coil
1.
g
e+
Timing Counter
Drift Chamber
Drift Chamber
1m
• Method proposed in 1998: PSI-RR-99-05: 10-14 possibility
• MEG proposal: september 2002: 10-13 goal: A. Baldini and T. Mori
spokespersons: Italy, Japan, Switzerland, Russia
6
Detectors responsibilities
Switzerland
Drift Chambers
Beam Line
DAQ
Russia
LXe Tests
Beam line
Italy
Japan
e+ counter
Trigger LXe
Calorimeter
LXe Calorimeter,
Spectrometer’s
magnet
USA(UCI)
Calibrations/Target/DC
pressure system
7
APD Cooled Support
APD F.E. Board
TC Final Design
Fibers (longitudinal position):
mainly needed for trigger
• A PLASTIC SUPPORT
APD ARRANGES THE
STRUCTURE
SCINTILLATOR BARS AS
REQUESTED
• PM
THE BARS ARE GLUED
ONTO
THE SUPPORT
Muon beam
direction
• INTERFACE ELEMENTS ARE
GLUED ONTO THE BARS AND
Divider
Board
SUPPORT THE
FIBRES
• FIBRES ARE GLUED AS
WELL
Main Support
• TEMPORARY ALUMINIUM
BEAMS ARE USED TO HANDLE
THE DETECTOR DURING
INSTALLATION
Scintillator Slab
• PTFE SLIDERS WILL
ENSURE
A SMOOTH MOTION ALONG
THE RAILS
Scintillator Housing
8
PM-Scintillator Coupler
9
The Liquid Xe calorimeter
•
•
•
•
•
800 l of Liquid Xe
848 PMT immersed in LXe
Only scintillation light
High luminosity
Unsegmented volume
Refrigerator
Experimental
check
In a Large
Prototype
H.V.
Signals
Cooling pipe
Vacuum
for thermal insulation
Liq. Xe
Al Honeycomb
window
PMT
Plasticfiller
1.5m
10
The liquid xenon calorimeter
11
m radiative decay
g
e Lower beam intensity < 10
Is necessary to reduce pilem
ups
n
n
Better s , makes it possible
7
t
to take data with higher
beam intensity
(rough) relative
timing calib.
< 2~3 nsec
A few days ~ 1 week to get
enough statistics
p0 gg
LED
Laser
PMT Gain
Higher V with
light att.
p- + p  p0 + n
Can be repeated
frequently
p0  gg (55MeV, 83MeV)
p- + p  g + n (129MeV)
Laser
10 days to scan all
volume precisely
alpha
(faster scan possible
with less points)
e+
LH2 target
Xenon
Calibration
PMT QE & Att. L
Cold GXe
LXe
g
e-
Proton Acc
Li(p,g)Be
Nickel g Generator
LiF target at
COBRA center
K
17.6MeV g
Bi
Tl
F
Li(p, g1) at 14.6 MeV
Li(p, g0) at 17.6 MeV
~daily calib.
Can be used
also for initial
setup
off
9 MeV Nickel γ-line
on
quelle
Illuminate Xe from
the back
Source (Cf)
transferred by
comp air  on/off
NaI
3 cm 20 cm
Polyethylene
0.25 cm Nickel plate
Calorimeter energy Resolution and uniformity at 55
MeV by means of
Another (movable) detector (NaI ) is
placed at 180° wrt the LXe
calorimeter
sR = 1.5%
FWHM = 4.6 %
Energy
resolution on
the calorimer
Entrance face
CW beam line
14
LiF target
LITHIUM g - spectrum +
FLUORINE g - spectrum
Automatic
insertion/Extraction from
the experiment center
(target)
15
First physics run in 2008
-First 3 months physics data taking
(september-december 2008)
-Xe LY increase
-DCHs instability on part of the chambers
after some months of operation: reduction
of efficiency to 30%
- APD: noise on DCHs turned off
CW Calibration each three
days during 2008 run
16
2008 run : 1014 muons stopped in target
We also took RMD data once/week at reduced beam intensity
RD
Programmed
beam
shutdowns
RD
RD
RD
RD
RD
Air test in
COBRA
Cooling system
repair
18
2008 data analysis: blind analysis: Eg
vs Dtge window
Sidebands are used to MEASURE
accidental background distributions
g Energy
Radiative decay + In flight positron annihilation +
resolution + pileup: in agreement with MCs
NO unwanted
backgrounds
Radiative Muon decays (low photon energy)
DCH resolutions from 2008 data
Tracks with two turns in the spectrometer
are used to detetrmine the
Angular resolutions
The edge of Michel positrons used to
determine momentum resolution
score = 374 keV (60%)
stail1 = 1.06 MeV (33%)
stail2 = 2 MeV (7%)
s(Df) =14 mrad
 s(f) =10 mrad
s(Dq) = 25 mrad
 s(q) = 18 mrad
Probability distribution functions
• Signal: from data except positron angular resolutions which is based on MC
• Background: from sidebands (D timing flat)
• Radiative decay: MC based on theoretical distributions + experimental resolutions
Analysis cuts
Likelihood analysis: accidentals + radiative
+ signal PDFs to fit data + Feldman Cousins
Best fit in the signal region
0  N Sig  14.6
Agreement of 3
different analyses
Kinematical distributions with a different analysis
Normalization: measured Michel events simultaneous with
the normal MEG trigger
Neg  BR (m   eg )  k
dove:
 fS
k  N enn  
 fM

 
 (TRG  MEG | e g )
  A(g | track )   (g )  Psc( Mtr )
 

   (TRG  Michel | track  em  TC ) 
f S  A(DC )   (track , p e  50MeV|DC )   (TC| p e  50MeV ) S
fM   M
pre-scaling 107
-Independent of instantaneous beam rate
- Nearly insensitive to positron acceptance and efficiency factors associated with
DCH and TC
90% CL limit
90 % C.L. NSig  14.6 corresponds to BR(m→eg)  2.8 x 10-11
Computed sensitivity 1.3 x 10-11
Statistical fluctuation ~5%
From side bands analysis we expected 0.9 (left) and 2.1
(right) x 10-11
• Bad luck
•
•
•
•
Xenon
purification
New (2009) custom liquid phase
purification system : Oxysorb-like +
“silent” pump (piston-type)
50 cc / cycle, 60 rpm
operation
180 liter/h liquid circulation
2 months of data
taking in 2009:
31
• Problem on DCHs  problem in HV distribution cards
• All chambers repaired before start of 2009 beam
time
Hit map 2008
2009
32
2009 run
• Smoother: LXe clean, DCHs working properly
• Shorter run: another experiment (muonic atom Lamb shift) having good results
• Transverse (fibers+APD) timing counter still missing: noise induced in DCHs
• Preliminary DCHs resolutions though improved are not yet at the proposal
level. Synchronization between different electronic channels measuring timing
not yet at good level
33
Prospects
• 2 months of stable data taking at the end of 2009
• Improvement in sensitivity due to stable conditions: 6 * 10-12 for 2009
data (analysing now): ready this summer
• Started running in stable conditions at the end of 2009: continue at
least until 2012 (no competitor)
• Data taking now paused due to accelerator maintenance will resume
next month
• Start thinking of possibile improvements/upgrades
Planning
1998
1999
R&D
2000
2001
2002
now
Data Taking
Assembly
2003
2004
2005
2006
2007
2008
2009
2010
http://meg.psi.ch
More details at
34
2011
Present: 2009 analysis
A.M. Baldini PSI February
17° 2010
Likelihood analysis
Pm
(Y. Kuno et al., MEG TN1, 1997 and references)
H.E. g in m   e nng : (1  Pm cosq g )
H.E. e  in m   e nn : (1  Pm cosq e )
e
qD
m
Pm
g
Det. 1
Det. 2
 Suppression factor  (for isotropic m  eg decay)
1

qd cosq
cos
D
D
(1  Pm cos qD )(1  Pm cos qD )
1
 d cosq
D
cosqD
• For suitable geometry big  factors can be obtained
• This is not the case for MEG (detailed calculations are necessary )
• In some theories (minimal SU(5) model) the positron has a definite helicity
 Pm is less effective
36