Transcript The Mirror Crack’d*: History and Status of CP Violation Studies Eric Prebys

The Mirror Crack’d*: History and
Status of CP Violation Studies
Eric Prebys (UR ‘90*), Fermi National Accelerator Laboratory
Representing the
BELLE Collaboration
*apologies to Agatha Christie
September 26, 2001
University of Rochester
1
The BELLE Collaboration
300 people from 49
Institutions in 11 Countries:
Australia, China, India,
Korea, Japan, Philippines,
Poland, Russia, Taiwan,
Ukraine, and USA
Academia Sinica
Budker Inst. of Nuclear Physics
Chuo University
Fukui University
University of Hawaii
Institute of Single Crystal
Kanagawa University
Korea University
Kyoto University
Mindanao State University
Nagoya University
National Lien Ho Colledge of T&C
Nihon Dental College
Osaka University
Princeton University
Sankyun Kwan University
Seoul National University
University of Sydeny
Tohoku University
University of Tokyo
Tokyo Institute of Technology
Toyama N.C. of Martime
technology
Utkal University
Yonsei University
September 26, 2001
University of Rochester
Aomori University
Chiba University
University of Cincinatti
GyeongSang National University
Institute of High Energy Physics
Joint Crystal Collab. Group
KEK
Krakow Inst. of Nuclear Physics
Melbourne University
Nagasaki Inst. of App. Science
Nara Women's University
National Taiwan University
Niigata University
Osaka City University
Saga University
Univ. of Science & Technology of
China
Sugiyama Jyogakuin University
Toho University
Tohoku-Gakuin University
Tokyo Metropolitan University
Tokyo Univ. of Agricult. & Tech.
University of Tsukuba
Virginia Polytechnic Institute
2
Just to set the tone….
Dear Eric,
I just returned to Rochester and I am happy to know that Tom has invited you for a colloquium on Sep 26. Can you
send me a title of your talk at the earliest. I would like to tell you a few things that Tom may not have mentioned. First, you
will be the first speaker of the semester and, therefore, you carry a great responsibility for presenting a very good
colloquium. Second, since our colloquium attendance has thinned over the years (because of bad talks, specialized talks), I
have assured the students that I will only invite extraordinary speakers who can give a very general talk
to graduate students across all disciplines. So, I would like you to prepare your talk keeping this in mind. In particular, what
this means is that please do not make it a talk on experimental physics, rather on physics. Remember the time when
you were a student and the kinds of things you hated in colloquia, please avoid them. Not all the
students will be from high energy physics. In fact, many are from optics, astronomy and so a talk with less display of
detectors etc and with a greater balance of theoretical motivation and the explanation of results would be highly appreciated.
Why am I telling you all this? Well, first of all, you were our former student and as such I have a
right to ask you for things. Second, you will be the first speaker and if the students are not thrilled with your talk, the
attendance may shrink in the subsequent talks. On the other hand, if your talk is superb, which I hope it will be, more people
will show up for the later talks (people have a tendency to extrapolate). In any case, please keep in mind that you will be
talking to a general audience and not to a group of experimentalists.
Let me know when your itinerary is complete, but please send me a title in a couple of days.
With very best regards,
Ashok.
September 26, 2001
University of Rochester
3
Outline
• Why do we care?
• History
–
–
–
–
–
Parity Violation
V-A Currents and CP (almost) Conservation
CP Violation in the Neutral K System
The Cabbibo-Kobayashi-Maskowa Mechanism
“The” Unitarity Triangle
• The Present
– Direct CP Violation in the Neutral K System (’/)
– Indirect CP Violation in the B meson System (B-Factories)
• The Future?
September 26, 2001
University of Rochester
4
Why do We Care?
• Dirac first predicted antimatter in 1930 as a consequence of the “extra” solutions
to his relativistic formulation of quantum mechanics - and was widely ridiculed.
• The positron (anti-electron) was discovered by Anderson in 1932 and the antiproton was discovered by Segre and Chamberlain in 1955.
• Now we are all quite comfortable with the idea of antimatter as “equal and
opposite” to matter, e.g.
“Of course, there is only one correct mixing ratio of matter and
antimatter: one to one!” – Star Trek, The Next Generation
• …but why does the universe seem to be made entirely of matter?
• Why do there seem to be tiny differences in the physics of matter and antimatter?
• These legitimately qualify as “big questions”.
September 26, 2001
University of Rochester
5
Parity Violation
y
z
x
z
y
x
• The “parity” operation transforms the universe into its mirror
image (goes from right-handed to left-handed).
• Maxwell’s equations are totally parity invariant.
• BUT, in the 50’s huge parity violation was observed in weak
decays…
Example: b decay of polarized Co...
e
60
Co
J5
September 26, 2001
Ni*
J4
electron preferentially
emitted opposite spin
direction
60
University of Rochester
e
6
Weak Currents and Parity Violation
Review: QED
e

A

jEM
e

C



A  jCA
jDB,  uC  u A u D  uB

*
eB
jEM ,
eD
Transform like vectors
For weak interactions, try (“four fermion interaction”)
e

A
B
axial vector

jweak
jweak , 
September 26, 2001
C
eD


j   uC cv   c A 5  u A
vector
Manifestly Violates Parity!!
University of Rochester
7
“V-A” Current
Experimentally, it was found that data were best described by



jweak
 uC     5  u A
Maximum Parity Violation!!!!
Recall that for Direct Spinors, the left handed projection operator is
1  5 

u  jweak
u L  PLu  
 uL  u L
 2 
“Left-handed” current
For massless particles, spinor state = helicity state
Only Left-handed Neutrinos
September 26, 2001
University of Rochester
8
CP Conservation (sort of)
When we apply the usual Dirac gymnastics, we find that for anti-particles



jweak
 vC     5  vA  vR  vR
Right-handed current
Only Right-handed anti-Neutrinos
Overall symmetry restored under the combined
operations of C(harge conjugation) and P(arity).
CP Conservation!!!
well, maybe not….
September 26, 2001
University of Rochester
9
The Neutral Kaon System
In experiments in the 1950s, it was found that there were two
types of neutral strange particles, of indistinguishable mass (498
MeV), but with different decay properties.
CP = -1
K L ( ong)  3
K S ( hort)  2
CP = +1
Because 3*m  mK , the KL lives about 600 times longer than the KS, hence the
names.
Strangeness eigenstates
Possible explanation:
KS
KL


1

K0  K0
2
1

K0  K0
2


close, but not quite correct…
September 26, 2001
University of Rochester
10
CP Violation in the Neutral K System
In 1964, Fitch, Cronin, etal, showed that in fact KL2 with a
branching ratio on the order of 10-3.
Interpretation:
CP Eigenstates
1
K1 
K0  K0
2
1
K2 
K0  K0
2
Mass Eigenstates


KS
KL


K S  K1   K 2
5
10
15
 (1010 s)
K L  K 2   K1
  2.3 10 3
September 26, 2001
University of Rochester
11
The Significance
In other words…
KL,S  aL,S K  bL,S K
0
0
where
a L ,S  bL ,S
This generated great interest (not to mention a Nobel Prize), and has
been studied in great detail ever since, but until recently had only
been conclusively observed in the kaon system.
Unlike parity violation, it is not trivial to
incorporate CP violation into the standard
model. To understand how it is done, we
must now digress a bit into some details of
fundamental particle interactions….
September 26, 2001
University of Rochester
12
Weak Interactions in the Standard Model
• In the Standard Model, the fundamental particles are leptons
and quarks
quarks combine as
qqq, q q q, or q q
to form hadrons
leptons exist
independently
• In this model, weak interactions are analogous to QED.
e
e

September 26, 2001
e
e
W
University of Rochester
OR
u
d
W
13
Quark Mixing
  e        
      
e


    
In the Standard Model, leptons
can only transition within a
generation (NOTE: probably
not true!)
Although the rate is suppressed,
u
c
t
    quarks can transition between
    generations.
 d  s   b 
September 26, 2001
University of Rochester
14
The CKM Matrix (1973)
• The weak quark eigenstates are related to the strong (or mass) eigenstates
through a unitary transformation.
 u  c  t 
   
 d '  s'  b' 
d ' Vud Vus Vub  d 
 s'   V
 s 
V
V
cs
cb   
   cd
 b'  Vtd Vts Vtb   b 
Cabibbo-Kobayashi-Maskawa (CKM) Matrix
d
Vud
W
u
d
Vud*
u
W
• The only straightforward way to accommodate CP violation in the SM is by
means of an irreducible phase in this matrix
• This requires at least three generations and led to prediction of t and b
quarks … a year before the discovery of the c quark!
September 26, 2001
University of Rochester
15
Wolfenstein Parameterization
The CKM matrix is an SU(3) transformation, which has four
free parameters. Because of the scale of the elements, this is
often represented with the “Wolfenstein Parameterization”
 1  2 2


2


1  2
 A3 (1    i)  A2

First two generations almost
unitary. = sine of “Cabbibo
Angle”
September 26, 2001
University of Rochester
A3 (  i)

2
A


1

CP Violating
phase
16
“The” Unitarity Triangle
• Unitarity imposes several constraints on the matrix, but one
(product first and third columns)...
VtdVtb*  VcdVcb*  VudVub*  0
results in a triangle in the complex plane with sides of similar
length  A3 , and appears the most interesting for study


2
VudVub*
3
VtdVtb*
1
VcdVcb*
(Note! in US : 1  b, 2  , 3   )
September 26, 2001
University of Rochester
17
The  Plane
• Remembering the Wolfenstein Parameterization
 1  2 2




1  2 2
 A3 (1    i)  A2

A3 (  i)

A2


1

we can divide through by the magnitude of the base (A3)….
ρ, η
VudVub*
VcdVcb*
2
3
VtdVtb*
VcdVcb*
1
1,0
0,0
CP violation is generally discussed in terms of this plane
September 26, 2001
University of Rochester
18
Direct CP Violation
• CP Violation is manifests itself as a difference between the
physics of matter and anti-matter
 (i  f )   (i  f )
• Direct CP Violation is the observation of a difference between
two such decay rates; however, the amplitude for one process
can in general be written
iw is
A Ae e
 A Ae
Weak phase changes sign
 iw is
e
Strong phase does not
• Since the observed rate is only proportional to the amplitude, a
difference would only be observed if there were an interference
between two diagrams with different weak and strong phase.
 Rare and hard to interpret
September 26, 2001
University of Rochester
19
Direct CP Violation in the Neutral Kaon System
(’/ Measurement)
Recall…
K S  K1   K 2
K L  K 2   K1
If there is only indirect CP violation, then ALL 2 decays really come from K1 , and
we expect (among other things)
Br ( K L     ) Br ( K1     ) Br ( K S     )


0 0
0 0
Br ( K L    ) Br ( K1    ) Br ( K S   0 0 )
But the Standard Model allows
Br ( K 0  2 )  Br ( K 0  2 )
 K 2  2
September 26, 2001
Direct CP Violation
University of Rochester
20
Direct CP Violation in the Neutral Kaon System
(cont’d)
Formalism:
CP=+1
CP=-1
K L  K 2   K1
 
’

CP=+1
A( K L     )

  '
 
A( K S    )
A( K L   0 0 )
00 
   2 '
0 0
A( K S    )
Br ( K L    ) / Br ( K S    )   

0 0
0 0
Br ( K L    ) / Br ( K S    ) 00




2
 1  6 Re( ' /  )
Theoretical estimates for ’/ range from 4-30 x 10-4
September 26, 2001
University of Rochester
21
Easy to Measure….NOT!

KL


 0

KS

  0
0
0









Must take great steps to understand acceptances and
systematic errors!!
September 26, 2001

University of Rochester
Detector
22
KTeV Experiment (Fermilab)
(Images from Jim Graham’s Fermilab “Wine and Cheese” Talk)
September 26, 2001
University of Rochester
23
Current Status of ’/
This bothered people
At this point, the
accuracy of this
measurement is better
than that of the
theoretical prediction:
(4-30 x 10-4)
(ibid.)
September 26, 2001
University of Rochester
24
Indirect CP Violation in the B Meson System
• Let’s Look at B-mixing…
V
d
B
0
b
td
t
V
W
V
*
tb
W
t
*
tb
V
b
B
0
d
td

B 0 (t )  e  i ( mi /  cos 2mt  B 0  i sin  2mt e 2im B 0

*

arg(
V
V
Mixing phase
td tb )  1
September 26, 2001
University of Rochester
25
Indirect CP Violation (cont’d)
• If both B and B can decay to the same CP eigenstate f,
there will be an interference
B
0
f
B
0
And the time-dependent decay probability will be
Difference between B mass eigenstates
P(t)  e
 |t |
1  CP sin( M   D ) sin( m * t )
Decay phase
CP state of f
September 26, 2001
Mixing phase
University of Rochester
26
The  Resonances
At the right energies, electrons and positrons can produce a
spectrum of bound resonant states of b and anti-b quarks
e
The 1- states are called the
“ (‘Upsilon’)resonances”
b

*
b
e
b

b
Starting with the (4S), they can decay
strongly to pairs of B-mesons.
b
The lighter states must
decay through quarkantiquark annihilation
September 26, 2001

b

u
b
University of Rochester
B0
d
u
b
b
B
B
b
b
d
b
B0
27
The Basic Idea
• We can create B0 B0 pairs at the ( 4S) resonance.
• Even though both B’s are mixing, if we tag the decay of one
of them, the other must be the CP conjugate at that time. We
therefore measure the time dependent decay of one B relative
to the time that the first one was tagged (EPR “paradox”).
• PROBLEM: At the ( 4S) resonance, B’s only go about 30
m in the center of mass, making it difficult to measure timedependent mixing.
e
-
 30 m
B0
e
B0
September 26, 2001
University of Rochester
28
The Clever Trick (courtesy P. Oddone)
• If the collider is asymmetric, then the entire system is Lorentz
boosted.
• In the Belle Experiment, 8 GeV e-’s are collided with 3.5
GeV e+’s so
e
 30 m
-
B0
B0
e

e
 200 m
-
B0
e
B0
• So now the time measurement becomes a z position
measurement.
September 26, 2001
University of Rochester
29
“Gold-Plated” Decay
B
b
V
0
W
*
J
/

c
cb
V
d
(  e  e  ,    , etc)
c
Total state CP
cs
s
d
K S (CP  1), K L (CP  1)


  , 
0
0
D  arg(VcsVcb* )  0
probes M  1 ( b)
September 26, 2001
University of Rochester
30
Predicted Signature
t = Time of tagged decays
September 26, 2001
University of Rochester
31
“Tin-Plated” Decay
V
ud
W
B
0
b
V
*
ub
d

d
u
u

d


D  arg(VudVub* )  ( 1  2 )
probes M  D  1  (2  1 )  2 ( )
Complicated by “penguin pollution”, but still promising
September 26, 2001
University of Rochester
32
Review - What B-Factories Do...
Make LOTS of b b pairs at the (4S) resonance in an asymmetric collider.
Detect the decay of one B to a CP eigenstate.
Tag the flavor of the other B.
Reconstruct the position of the two vertices.
Measure the z separation between them and calculate proper time
separation as
t  z form
/(bCM  CM c)
• Fit to the functional
•
•
•
•
•
e
 |t |
1  CP sin 21 sin mt
• Write papers.
• Over the last ~8 years, there have been two dedicated experiments under
way to do this – BaBar (SLAC) and Belle (KEK)
September 26, 2001
University of Rochester
33
Motivations for Accelerator Parameters
• Must be asymmetric to take advantage of Lorentz boost.
• The decays of interest all have branching ratios on the order
of 10-5 or lower.
– Need lots and lots of data!
• Physics projections assume 100 fb-1 = 1yr @ 1034 cm-2s-1
• Would have been pointless if less than 1033 cm-2s-1
September 26, 2001
University of Rochester
34
The KEKB Collider (KEK)
• Asymmetric Rings
– 8.0GeV(HER)
– 3.5GeV(LER)
• Ecm=10.58GeV=
M((4S))
• Target Luminosity:
1034s-1cm-2
• Circumference: 3016m
• Crossing angle: 11mr
• RF Buckets: 5120
•  2ns crossing time
September 26, 2001
University of Rochester
35
The PEP-II Collider (SLAC)
• Asymmetric Rings
– 9.0GeV(HER)
– 3.1GeV(LER)
• Ecm=10.58GeV=
M((4S))
• Target Luminosity:
3x1033s-1cm-2
• Crossing angle: 0 mr
• 4ns crossing time
September 26, 2001
University of Rochester
36
Motivation for Detector Parameters
•
•
•
•
•
Vertex Measurement
– Need to measure decay vertices to <100m to get proper time distribution.
Tracking…
– Would like p/p.5-1% to help distinguish B decays from BK and
BKK decays.
– Provide dE/dx for particle ID.
EM calorimetry
– Detect ’s from slow, asymmetric 0’s  need efficiency down to 20 MeV.
Hadronic Calorimetry
– Tag muons.
– Tag direction of KL’s from decay BKL .
Particle ID
– Tag strangeness to distinguish B decays from Bbar decays (low p).
– Tag ’s to distinguish B decays from BK and BKK decays (high p).
Rely on mature, robust technologies whenever possible!!!
September 26, 2001
University of Rochester
37
The Belle Detector
September 26, 2001
University of Rochester
38
BaBar Detector (SLAC)
September 26, 2001
University of Rochester
39
The Accelerator is Key!!!
STOP Run
+HV Down
+Fill HER
+Fill LER
+HV Up
+START Run
= 8 Minutes!
September 26, 2001
University of Rochester
40
Luminosity
Our Records:
Daily integrated luminosity
World Records!!
•Instantaneous: 4.49 1033 cm -2s -1
•Per (0-24h) day: 229.1 pb -1
•Per (24 hr) day: 241.3 pb -1
Total integrated luminosity
•Per week:
1478 pb-1
•To date:
 29.9 fb -1
(on peak)
Note: integrated numbers
are accumulated!
Total for these Results:
29.1 fb -1
Total for first CP Results
(Osaka): 6.2 fb -1
September 26, 2001
University of Rochester
41
The Pieces of the Analysis
•
•
•
•
September 26, 2001
Event reconstruction and selection
Flavor Tagging
Vertex reconstruction
CP fitting
University of Rochester
42
J/ and KS Reconstruction
    
K S   
s4 Mev
Require mass
within 4s of PDG
  ee
September 26, 2001
University of Rochester
43
BKS Reconstruction
• In the CM, both energy
and momentum of a real
B0 are constrained.
• Use “Beam-constrained
Mass”:
M
2
BC
E
2
beam

Signal
 p
2
123 Events
3.7 Background
September 26, 2001
University of Rochester
44
All Fully Reconstructed Modes (i.e. all but KL)
Mode
September 26, 2001
Events Background
BKS
457
12
All Others
290
46
Total
747
58
University of Rochester
45
BKL Reconstruction
KLM Cluster
KL
J/ daughter
particles
• Measure direction (only) of
KL in lab frame
• Scale momentum so that
M(KL+)=M(B0)
• Transform to CM frame
and look at p(B0).
September 26, 2001
University of Rochester
46
BKL Signal
0<pB*<2 GeV/c
Biases spectrum!
346 Events
223 Background
September 26, 2001
University of Rochester
47
Flavor Tagging
ud , u s, c s, e   e , or   
W
b
B
0
c
s
q
K  , K 0 , , , etc.
d
X
Statistica lly, B 0 ' s will tend to produce high momentum
e  ,   , and/or K  , while B 0 ' s will produce the opposites.
September 26, 2001
University of Rochester
48
Flavor Tagging (Slow Pion)
ud , u s, c s, e   e , or   
W
b
B
0
c
D
c
u
*
u
d
d
D

0

Very slow pion
B 0 ' s will tend to produce slow  .
Combined effective efficiency eff = t(1-2w)2 = 27.0.2%
September 26, 2001
University of Rochester
49
Vertex Reconstruction (SVD)
Overall efficiency = ~85%.
September 26, 2001
In total 1137 events for the CP fit.
University of Rochester
50
CP Fit (Probability Density Function)
f (t ;sin 21 )  e

t
B

t 
1  sin 21 sin xd

B 

PDF   (1  f BG ) f (t ) R (t   t ) dt   f BG PDFBG (t )
•fBG = background fraction. Determined from a 2D fit of E vs M.
•R( t) = resolution function. Determined from D*’s and MC.
•PDFBG( t) = probability density function of background.
Determined from K sideband.
September 26, 2001
University of Rochester
51
Resolution Function
Fit with a double-Gaussian…
 main  0.09 ps
s main 1.54 ps
 tail  0.78 ps
s tail 3.78 ps
f tail
September 26, 2001
University of Rochester
0.018
52
Test of Vertexing – B Lifetime
 B0  1.55  .02 ps (PDG :1.55  .03 ps)
 B   1.64  .03 ps (PDG :1.65  .03 ps)
September 26, 2001
University of Rochester
53
The Combined Fit (All Charmonium States)
September 26, 2001
University of Rochester
54
Sources of Systematic Error
Source
Vertex Algorithm
Flavor Tagging
Resolution Function
KL Background Fraction
Background Shapes
m d and B Errors
Total
s
.04
.03
.02
.02
.01
.01
.06
• Bottom Line
sin 21  .99  .14(stat )  .06(syst.)
Published in Phys.Rev.Lett. 87, 091802 (2001)
September 26, 2001
University of Rochester
55
The BaBar Measurement
Based on 32 million B-Bbar pairs
sin 2 b  .59  .14  .05
Phys.Rev.Lett. 87 (2001)
September 26, 2001
University of Rochester
56
Summary of 21 Measurements
September 26, 2001
University of Rochester
57
How About That  Plane?
World Average Sin21 (1s)
Constraints of Everything but Sin21
Looks good for the Standard Model, but a little dull for experimenters !
September 26, 2001
University of Rochester
58
Current Status
• The study of CP Violation has been going on for almost 40 years!
• A number of experiments are currently taking data which seem to
be confirming the Standard Model (CKM) explanation of CP
Violation, and thereby constraining that model
– Direct CP violation is observed in the neutral K system!
– CP is violated in the B-Meson system!
• Over the next several years, the existing B-Factories will continue
to take data, providing tighter and tighter constraints.
• New players are also coming on the scene:
–
–
–
–
Fermilab Run II (CDF and D0) - now
BTeV (dedicated B Experiment at Fermilab) - ~2005
LHC (Atlas and CMS) - 2006
LHC-B (dedicated B Experiment at LHC) - ?
September 26, 2001
University of Rochester
59
More “Out There”
• CP Violation in the  sector? (probably there, hard to study)
• CPT Violation?
– CPT Conservation is a direct consequence of the Lorentz invariance of the
Lagrangian.
– Evidence of its violation would be observation (direct or indirect) of
m( p )  m( p ) or ( p)  ( p)
and would be big news.
• We still can’t answer why the unverse is all matter. Maybe it isn’t!
– The AMS experiment, set to fly on the ISS, will look for massive anti-nuclei
to test the hypothesis that distant parts of the universe might be antimatter
(!!)
September 26, 2001
University of Rochester
60
Are Two B-Factories Too Many?
• These are not discovery machines!
• Any interesting physics would manifest itself as small
deviations from SM predictions.
• People would be very skeptical about such claims without
independent confirmation.
• Therefore, the answer is NO (two is not one too many,
anyway).
September 26, 2001
University of Rochester
61
Differences Between PEP-II (BaBar) and KEKB
(Belle)
•PEP-II has complex IR optics to force beams to collide head-on.
Pros:
Interaction of head-on beams well understood.
Cons: Complicates IR design.
More synchrotron radiation.
Can’t populate every RF bucket.
• In KEK-B, the beams cross at ±11 mr.
Pros:
Simple IR design.
Can populate every RF bucket.
Lower (but not zero!!!) synchrotron radiation.
Cons: Crossing can potentially couple longitudinal
and transverse instabilities.
September 26, 2001
At present, bothUniversity
designs
seem to be working.
of Rochester
62
Differences (cont’d)
Readout:
• BaBar uses an SLD-inspired system, based on a continuous digitization. The
entire detector is pipelined into a software-based trigger.
Pros:
Extremely versatile trigger.
Less worry about hardware-based trigger systematics.
Can go to very high luminosities.
Cons: Required development of lots of custom hardware.
• Belle’s readout is based on converting signals to time-pulses. The trigger is an
“old-fashioned” hardware-based level one. Events satisfying level one are read out
after a 2 µs latency.
Pros:
Simple.
Readout relies largely on “off-the-shelf” electronics.
Cons: Potential for hardware-based trigger systematics.
Possible problems with high luminosity.
September 26, 2001
University of Rochester
63
Particle ID needs
Technology
Pros
Cons
Comment
TOF
Simple.
Only for low
momentum.
Included in
Belle
dE/dx
Proven.
Comes for
free.
Only for low
momentum
Included in
Belle.
TMAE based
RICH
Proven in
SLD and
DELPHI
Universally
despised.
Rejected.
CSI RICH
September 26, 2001
Once seemed No one could
promising.
build a
working
prototype.
Rejected.
DIRC
Rugged.
Excellent
separation.
New.
Contstrants
on detector
geometry
Babar choice
Aerogel
threshold
Cerenkov
Simple.
Barely
adequate
Belle choice
University of Rochester
64