The KLOE Experiment

1
The KLOE Experiment at DAFNE M. Moulson,
INFN/Frascati for the KLOE collaboration BNL
Particle Physics Seminar, 30 March 2000
2
The KLOE Experiment at DAFNE M. Moulson,
INFN/Frascati for the KLOE collaboration Nuclear/P
article Seminar, 29 March 2000
3
Outline
Physics at a f-factory The DAFNE facility The
KLOE experiment Groundwork for e?/e at
KLOE Preliminary results from 1999 Outlook
The KLOE experiment in the DAFNE hall
4
Physics at a f-factory
ee- collider W m? ? 1.02 GeV
L 5 1032 cm-2 s-1 ? 2 1010 f/year
s?ee- ? ?? ? 3.2 ?b
Clean environment KSKL beams in pure quantum state
BRs for selected f decays BRs for selected f decays
KK- 49.1
KSKL 34.1
rp pp-p0 15.5
hg 1.3
f0g, a0g, h?g 10-4
(JPC 1--)
Allows for tagging and interferometry
Appreciable acceptance for KL with realistic
detector dimensions
5
The KLOE physics program

• Measurements of CP and CPT violation parameters
• double ratio KL ? 2p vs. KS ? 2p
• interferometry KS,KL ? f1,f2
• KS,KL semileptonic asymmetries esp. KS
• Kaon physics
• form factors KL ?pln, K ?pln, eventually Kl4
• rare KS decays KS ?pln, also KS ?pee, pmm, pnn
• regeneration cross section at low momentum
• Non-kaon physics
• radiative decays s(f ? f0g, a0g), BR(f
  ?h?g)/BR(f ?hg)
• s(ee- ? hadrons) using ee- ?pp-g (ISR)

6
CP violation in the KSKL system
Want to relate to weak eigenstates KS, KL,
but KL ? 2p observed since 1964!
Direct CP violation
Indirect CP violation
7
Direct CP violation in KL? 2p decay
If Im AI 0, then h- h00 e Else
parameterize direct CP by e?
h- ? e e? h00 ? e 2e?
8
KS,KL semileptonic asymmetries
Assuming DS DQ
Allowing for CPT violation
eS ? eK dK eL ? eK ? dK
Re b violates CP, CPT Im b violates CPT
If CPT violated but DS DQ AL AS ? 4 Re dK
Also useful to define (and note)
9
Interferometry with KS,KL
Start with pure quantum state
Evolve in time
Obtain decay amplitude to f1, f2
Integrate decay intensity over t1, t2 at constant
Dt t1 t2
Expression valid for Dt ? 0 For Dt ? 0, 1 ? 2
10
Identical final states f1 f2 f
For f1 f2, h1 h2 h and f1 f2
Quantity s PDG98 sstat KLOE
s(GS)/GS GS 0.893410-10 s 910-4 210-4
s(GL)/GL Gl 5.1710-10 s 810-3 510-4
s(Dm)/ Dm Dm 3.48910-12 MeV 2.610-3 210-3
For complete KLOE program 41010 fs
11
Similar final states f1 pp-, f2 p0p0
h1 ? h- ? e e? h2 ? h00 ? e - 2e? f1 f2 ?
f- - f00 ? 3 Im (e?/e)
asymmetry ? Im(e?/e)
asymmetry ? Re(e?/e)
12
Similar final states f1 p-ln, f2 pl-n
_
Assuming DS DQ
h1 ? hl ? 1 - 2dK h2 ? hl- ? -1 - 2dK f1 f2
? f l- f l- ? p - 4 ImdK
Constructive interference
13
In summary
If DS DQ assumed, 13 independent parameters
describe CP and CPT violation in the neutral kaon
system. Using interferometry and appropriate
choices of f1, f2, KLOE can completely determine
them by measuring 16 different quantities.
14
Measurements of Re(e?/e)
KTeV Feb99 28.0 4.1 10-4 NA48
average 14.0 4.3 10-4 Jun99 18.5 7.3
10-4 Feb00 12.2 4.9 10-4
World average (NA48) 19.3 2.4 10-4 c2/dof
11.5/5
Source NA48
15
Theoretical predictions for Re(e?/e)
SM includes CP violation via phase in CKM
matrix Computation of e? is difficult Hadronic
matrix elements which dominate A0, A2 terms of
Re(e?/e) nearly cancel
Source NA48
Estimates of Re(e?/e) /10-4
Group Method Gaussian Scan
Rome Lattice QCD -1.211.4
Munich Phenomenological 4.213.7 1.128
Trieste Chiral quark model 731
Standard Model can accommodate a large value for
Re(e?/e), but with difficulty
16
Measurement of e?/e via the double ratio
At KLOE
NKK and rtag cancel from R
17
Systematics and the double ratio
What doesnt cancel (identically) from R

• Geometrical acceptance
• Vertex resolution different for charged, neutral
  vertices
• Fiducial volume nominally the same but
  misalignment occurs
• Trigger efficiencies
• Reconstruction efficiencies
• Backgrounds

To obtain Re(e/e) with 10-4 accuracy, ssyst(R)
must be held down to 23 10-4
18
Addressing systematics
Systematics evaluated using abundant
processes EMC and DC provide independent
measurements
KLOE is self calibrating
Parameter Method
Abs. EMC energy scale Bhabha (incl. rad), gg, f?pp-p0, f?hg
Abs. EMC time scale Bhabha, gg
DC t0, t-s relations, eff. Bhabha, cosmic, KS?pp-
Track/vertex recon. eff. copious KL, K? decays (KL?pp-p0, K??p?p0)
Photon eff. rad. Bhabha, f?pp-p0, K??p?p0, KL?pp-p0
Neutral vtx recon. eff. KL?pp-p0, KL?p0p0p0
Fiducial volumes KL?pp-p0
19
Backgrounds
Rejection needed for dRe(e?/e) lt 0.510-4
assuming dNBkg/NBkg 1015
Signal Background Physics S/B Needed Rejection Technique
KL ? 2p0 KL ? 3p0 1/240 8 ? 104 EMC hermeticity 98 (active quads) EMC s(E)/E ? 5.7/?E
KL ? pp- KL ? pp-p0 1/60 2 ? 104 EMC hermeticity 98 (active quads) EMC s(E)/E ? 5.7/?E
KL ? pp- KL ? pp-p0 1/60 2 ? 104 s(p?)/p? ? 0.4 50ltplt300 MeV
KL ? pp- KL ? pmn 1/135 6 ? 104 s(p?)/p? ? 0.4 50ltplt300 MeV
KL ? pp- KL ? pmn 1/135 6 ? 104 EMC TOF,energy profile
KL ? pp- KL ? pen 1/190 9 ? 104 EMC TOF,energy profile
20
Statistical requirements
The KLOE goal is to measure Re(e?/e) to 10-4
4 ? 106 KL ? p0p0 events 4 ? 1010 f (with
KLOE rFV and rtag) 2 years of data taking at L
5 ? 1032 cm-2 s-1
21
The DAFNE facility
Ldesign 5 1032 cm-2 s-1
120 bunch operation 5 A/beam
Single bunch L1 ? 5 1030 cm-2 s-1 ? LVEPP-2M
Beam-beam effects ? separate ee- rings

• Acceleration and e production by two-stage linac
• Accumulator for efficient injection into main
  rings
• Main-ring injection at full energy fast topping
  off

22
DAFNE specifications
Beam energy 0.51 GeV
Trajectory 98 m
Max number of bunches 120
Bunch spacing 2.7 ns
Particles/bunch 9 1010
Crossing angle 25 mrad
Bunch length (sz) 3.0 cm
Bunch size (sx/sy) 2000/20 mm
L (single bunch) 4.4 1030 cm-2 s-1
2 low-b intersection points Flat beams cross at
an angle p?f pf ? 13 MeV/c
23
DAFNE design challenges
DAFNE is a low-energy machine with large stored
currents
Luminosity limited by beam-beam effects Low power
emission by synchrotron radiation Long damping
times for synchrotron and betatron oscillations
Lifetime limited by Touschek (intrabunch)
scattering Beam lifetime proportional to g3
24
DAFNE performance in 1999
A typical December day
Near the end of 1999, stable running with Lstart
? 3.5 1030 cm-2 s-1 Lsustained ? 1.8 1030
cm-2 s-1 I/I- ? 350/250 mA tbeam ? 60 min
25
DAFNE performance in 1999
Bunches Achieved Design/bunch
11/1998, Pre-KLOE roll-in, single-bunch mode 11/1998, Pre-KLOE roll-in, single-bunch mode 11/1998, Pre-KLOE roll-in, single-bunch mode 11/1998, Pre-KLOE roll-in, single-bunch mode
L1 1 1.6 1030 cm-2 s-1 30
I 1 20 mA 45
Imax 1 110 mA 250
11/1998, Pre-KLOE roll-in, multi-bunch mode 11/1998, Pre-KLOE roll-in, multi-bunch mode 11/1998, Pre-KLOE roll-in, multi-bunch mode 11/1998, Pre-KLOE roll-in, multi-bunch mode
L 13 1 1031 cm-2 s-1 20
I 13 200 mA 40
Imax 30 550 mA 50
At end of DAFNE commissioning, performance
results obtained were quite respectable
12/1999, Best conditions delivered to KLOE in 1999 12/1999, Best conditions delivered to KLOE in 1999 12/1999, Best conditions delivered to KLOE in 1999 12/1999, Best conditions delivered to KLOE in 1999
Lstart 2040 3.5 1030 cm-2 s-1 3
Lsustained 2040 1.8 1030 cm-2 s-1 1.5
I-/I 2040 350/250 mA
Situation changed after installation of KLOE
26
DAFNE Problems and solutions
Coupling at KLOE IP Betatron motion ? beam
rotates in KLOE solenoidal field KLOE IP
bracketed with compensator solenoids with ½?(B
dl)KLOE 6 low-b PM quads have corresponding
rotation
Imprecise alignment at KLOE IP causes tilt errors
and skew coupling
Mar 2000 Vacuum broken at KLOE IP and
adjustments made
Transverse multibunch instability due to
higher-order mode in injection kickers
Jan-Feb 2000 Kickers modified to damp
higher-order mode
27
The KLOE experiment
Be beam pipe (0.5 mm thick) Instrumented
permanent magnet quadrupoles (32 PMTs)
Drift chamber (4 m ? ? 3.3 m) 90 helium 10
isobutane 12582/52140 sense/total wires
Electromagnetic calorimeter Lead/scintillating
fibers 4880 PMTs
Superconducting coil (5 m bore) B 0.56 T ( ? B
dl 2.2 Tm)
28
Electromagnetic calorimeter
Required capabilities
Specifications

• Reconstruct KS,KL?p0p0 vertices with accuracy of
  a few mm
• Discriminate KL?p0p0 from KL?p0p0p0
• Provide fast signals to level-1 trigger
• 20 KHz Bhabha (at Ldesign)
• 3 KHz cosmic rays
• Possibly provide useful information for particle
  identification
• Km3 rejection
• Integrate with experiment

• Energy resolution 5 / Ö
  E (GeV)
• Full efficiency
  20 lt Eg lt 300 MeV
• Time resolution 70 ps
  / Ö E (GeV)
• Spatial resolution
  1 cm for g conversion point
• Hermeticity
• Fast triggering response
• Operation in B-field

29
Electromagnetic calorimeter

• Fine-sampling lead/scintillating fiber
  calorimeter (good timing)
• 1 mm fibers 0.5 mm lead foils
• fiberleadglue 484210
• Energy sampling fraction 13 (good energy
  response)
• ? 5 gr/cm3 X0 1.6 cm
• 23 cm thick 15 X0
• Both sides readout to obtain z
  coordinate

30
Electromagnetic calorimeter
2 32 endcap modules 10/15/30 cells
24 barrel modules 60 cells (5 layers) 4.3m length
2440 cells total
4880 channels
31
EMC energy calibration and resolution

1. Clean source of MIPs from cosmic rays
2. PMT HV trimmed for rough equalization of
   channel-to-channel MIP response
3. Fine equalization of column-to-column Bhabha
   response
4. ee-?gg events fix absolute energy scale

ee-? ee-g Eg from DC
Non-linearity 1
d(E)/E
s(E)/E 8 at 510 MeV
E (MeV)
Bhabha
ee-?gg
s(E)/E
E (MeV)
32
EMC mass reconstruction
f ? pp-p0 M(p0 ? gg)
f ? hg M(h ? gg)
M 134.5 MeV sM 14.7 MeV
M 546.3 MeV sM 41.8 MeV
MeV
MeV
33
EMC time calibration and resolution
ee- ? gg

• Obtain DT0 and vfib directly from TATB spectrum
• Straight cosmics provide 55 measurements of T at
  known intervals of L
• Minimize residuals to get ?T0

ns
34
EMC time-of-flight measurement
T1-T5 distribution can distinguish
incoming/outgoing ms Used to reject cosmic rays
Outgoing m
Incoming m
T5
5
4
3
T1-T5 (ns)
2
1
T1
b L/DT L from DC
m mass from TOF Fit to b vs pDC gives mm 105
MeV/c2
35
Drift chamber
Required capabilities
Specifications

• Provide a large tracking volume for decay
  products of KL with uniformly high efficiency
• Provide kinematic rejection of e.g., KL?pmn
• Be as transparent as possible to
• Regeneration at inside wall
• g conversion before EMC
• Multiple scattering in active volume
• Operate at high rates
• 20 KHz Bhabha rate
• Provide input to level-1 trigger

• Radius of sensitive volume r 2 m ?
  ?FV(KL) ? 30
• Homogeneous filling of sensitive volume
• KS,KL?pp- vertex resolution 150 mm (rf) 1 mm
  (z)
• Spatial resolution (rf) 150mm
• Momentum resolution s(p)/p ? 0.5
  
• Light mechanical structure
• Drift medium with large X0

36
Drift chamber
4 m ? 3.3 m avg. length 12582/52140 sense/total
wires All stereo geometry
Stereo angle varies from 60150 mrad as
?r Constant stereo drop d 1.5 cm
Cell geometry periodic in z 12 inner layers 2
2 cm2 cells 46 outer layers 3 3 cm2 cells
Gas mixture 90 He, 10 iC4H10 Field wires
Al(Ag), 80 mm ? Sense wires W(Au), 25 mm
? X0(gaswires) 900 m
Mechanical structure entirely in carbon-fiber
composite (? 0.1 X0) Axial load 3.5 tons
37
Drift chamber
38
DC time-to-space relations
ns

• Drift velocity not saturated and
• depends on
• drift distance (s)
• cell shape (b)
• crossing angle (f)

3 3 cm2 cells Constant f Different b
cm

• Compact parameterization of
• t-s relations
• 6 reference cells (b)
• 36 bins in crossing angle (f)
• 242 t-s relations total
• each described by 5th order Chebyshev polynomial

ns
3 3 cm2 cells Constant b Different f
cm
39
DC resolution and residuals
Bhabha
KS ? pp-
40
DC efficiency

• Estimate efficiency using
• Bhabha events
• cosmic ray events

rHW ? 99 rSW ? 97
41
DC momentum and mass resolution
KS ? pp-
Bhabha
sM 1 MeV/c2
sp/p lt 0.4 45 lt q lt 135
KL ? pp-
sM 1 MeV/c2
42
DC luminous point reconstruction
Bhabha tracks extrapolated to the z axis measure
position (m) and size (L) of the luminous region
sm(x) ? sm(y) ? 0.5 mm sm(z) ? 1 mm
10 minutes of data taking at L 2 1030 cm-2
s-1 are needed to provide DAFNE with m(X,Y,Z) and
L(X,Z)

• with errors on
• m(X), m(Y) ? 60 mm
• L(X) ? 100 mm
• m(Z), L(Z) ? 1 mm

43
Beam pipe and instrumented quadrupoles
Spherical vacuum decay chamber for KS 0.5 mm Be
walls minimize regeneration, multiple scattering,
and energy loss Permanent magnet quads maximize
free solid angle, avoid use of iron in tracking
volume Quadrupole calorimeters increase rejection
for KL ? 3p0 by a factor of 5
44
Trigger requirements

• 1) Trigger efficiency for KL,S?pp- and KL,S?p0p0
  must be high
• high efficiency makes precise determination of
  inefficiency easier
• desirable to hold down inefficiency to 10-3
• 2) Trigger efficiency for KL,S?pp- and KL,S?p0p0
  should be nearly equal
• 3) Trigger efficiency should be easy to study
• Redundancy between EMC and DC triggers for many
  channels helps

At Ldesign, f rate is 1.6 KHz Trigger must
reject/downscale 3.5 KHz of Bhabhas (q gt21º, at
Ldesign) 2.6 KHz of cosmic rays ??? KHz of
machine background
Total rate must be held to 10 KHz
45
Trigger design
Trigger operates asynchronously with respect to
bunch crossing (every 2.7 ns)
T1 (200 ns after f) fast trigger synchronized
with DAFNE clock, provides start to EMC FEE
T2 (2 ms after T1) validation trigger provides
stop to DC TDCs allowing for drift time
46
Trigger
47
Calorimeter trigger
Sectors Cosmic sectors from exterior plane
Barrel 48/48/48 Endcaps 20/16/12
T1 trigger Barrel-Barrel, Endcap-Barrel, or
Endcap1-Endcap2 T2 trigger 1 Barrel sector or 3
sectors on same Endcap
Bhabha veto (T1) 2 sectors above Bhabha
threshold Barrel-Barrel or Endcap-Endcap
Thresholds
typical values Barrel Endcaps
f trigger 50 MeV 90 MeV
Bhabha veto 300 MeV 300 MeV
Cosmic veto 30 MeV 30 MeV
Cosmic veto (T2) 2 cosmic sectors above
threshold Endcap-Barrel or Barrel-Barrel (no
activity in inner part of DC)
48
Drift chamber trigger
T1 trigger N1 (15) hits in 150 ns (20 of time
spectrum) T2 trigger N2 (40) hits in 850 ns
(70 of time spectrum)
Superlayers
N1, N2, NS tuned to provide maximum rejection for
machine background, degraded Bhabhas while
retaining efficiency for (e.g.) KS ? pp-
Layers 558 ? 9 superlayers Up to NS (5) hits
counted on each Spiralizing particles do not
trigger (degraded Bhabhas, conversion e-)
Cosmic veto override (T2) TCR signal from layers
18 Preserves efficiency for ee- ? mm-, pp-g
49
Trigger performance
Trigger operation in 1999 DC trigger under test
Final HW configuration now ready Bhabha veto not
enforced L 1030 cm-2 s-1 ? Bhabha rate Hz
Cosmic veto Disabled 2.6 KHz cosmic rate
Enabled 0.7 KHz Bhabha veto 96 efficient from
offline studies
KS ? pp- rEMC () rDC () rtrig ()
Data 91.7 0.2 94.2 0.2 99.5 0.1
Blind MC 87.9 0.7 95.8 0.8 99.4 0.1
MC 87.7 0.7 95.5 0.8 99.2 0.1
Trigger efficiencies For events containing KL tag
(KS ? pp-) rEMC(rDC) estimated from data
exploiting independence of DC and EMC triggers
50
Absolute time scale

• 95 of f events contain at least 1 cluster
• Assume first cluster is from g
• t0 Rfirst/c - tfirst
• Rephase with RF t0 ? nbunchtclock
• Perform tracking and PID
• Correct t0 for the proper track length if needed
• E.g., KS ? pp-
• t0 (L/bc)first - tfirst
• Correct cluster times
• In most cases not necessary to track again

• Event t0 requires event-by-event knowledge of
  time between f decay and T1
• EMC TDCs started by T1 rephased with machine RF
• Cluster times relative to unknown bunch crossing
• Event t0 some multiple of bunch crossing period

tclust L/(b)c after first correction step
Bhabha, multi-bunch mode
KS ? pp-, single bunch mode
51
Data acquisition

• DAQ handles 23000 FEE channels
• Designed for 10 KHz event rate and sustained
  bandwidth of 50 MB/sec
• 5 KB average event size
• Low dead time (2 at 10 KHz) independent of
  event configuration
• 2.2 ms dead time incurred at T1
• All FEE channels digitized and
• buffered during this time
• Fully tested at design rate

52
Computing resources
Installed online farm handles full acquisition
rate Currently use 2 farm nodes for event
building/recording 75 of CPU is available for
monitoring, calibration tasks Installed offline
processing power 10 of final More than
sufficient to keep up with acquisition so
far Accepting bids for 2000 SpecInt95, 5 TB disk
storage
FDDI
GIGASWITCH
FDDI
Fast Eth, Gbit Switch
Tape library 6 Magstar drives 40 GB/tape 5500
slots 220 TB
ONLINE FARM 7 IBM H50 (4 PPC 604e, 330 MHz) 420
SpecInt95 0.5 TB local disk space
Fast Eth
SCSI
Tape server
OFFLINE FARM 10 Sun Enterprise 4500 (4
UltraSPARCII, 400 MHz) 700 SpecInt95
Fast Eth
SCSI
Tape server
SCSI
Gbit Eth
Offline farm disk server 2 Sun Enterprise 3500
0.5 TB RAID
53
Online monitoring
root hist. server
root browser illumination
SWITCH
L3 spy Bhabha ee- ? gg Cosmic
Remote spy
BUILDER
EMC monitor t(gg) E(Bhabha) MIP
Trigger monitor trigger performance background
rate luminosity estimate
L3
Remote spy
raw
DC monitor cell effic. residuals IP, pf monitor
RECORDER
DAFNE
Event display
DAFNE
OFFLINE
Offline monitoring W, sf, pf
Recorder
Calibration
54
Event reconstruction and classification
All raw data written to tape Reconstruction/stream
ing follows acquisition 2.6 KHz offline farm
bandwidth Recent optimization may improve by
factor 5! Fast turn-around (lt 3 hrs) for DAFNE
feedback
RAW
Translation
Cluster reconstruction
Absolute event t0
Calibration cosmics
Cosmic filter
Background filter
Calibration Bhabhas
DC hit reconstruction
DC hit reconstruction
DC track/vertex recon.
DC track/vertex recon.
Track-to-cluster assoc.
Track-to-cluster assoc.
Event classification
Dedicated
Dedicated
KLKS
rp
Rad
mm-
Bha
Cos
KK-
UFO
55
Data taking in 1999

• First collisions at DAFNE in mid-April
  Single bunch mode
  f line-shape scan (3 nb-1) to
  find peak
• Short data-taking periods in
  August, October, and November
  (200 nb-1 each)
• Sustained data-taking during November and December

1999 total ?L dt 2.42 pb-1 7.7 106 fs
1.1 106 reconstructed KSs
56
Groundwork for ??/? at KLOE
First measurement of Re(e?/e) at KLOE will use
the double ratio

• Use the first 0.4 pb-1 to hone
• Tagging strategies
• Kl3 background rejection
• Neutral vertex reconstruction
• Definition of fiducial volumes

First CP violating event observed
57
Tagging strategies
Presence of KS(KL) established by detection of
KL(KS)

• Includes
• BR for KL,S into the tagging channel
• Fiducial acceptance and reconstruction
• efficiency for tagging decay

?tagS(L) does not (ideally) depend on KS(KL)
decay mode
KL tags KL tags KS tags KS tags
KS ? pp- rtag ? 50 KL ? pp-p0 KL ? pln rtag ? 30
KS ? p0p0 rtag ? 20 KL interaction in EMC rtag ? 40
total rtag ? 70 total rtag ? 70
58
Tagging of KL by KS?pp-
sM 1.0 MeV/c2
sP 2.3 MeV/c

• KS identified by vertex
• with 2 tracks of opposite
• sign and
• r lt 5 cm from I.P.
• 50 lt P lt 170 MeV/c
• 400 lt M lt 600 MeV/c2

rtag/BR(KS?pp-) ? 72 Background lt 1 (f ? KK-)
lS 6.10.6 mm
59
Tagging of KL by KS?p0p0

• KS ? p0p0 identified by
• 4 prompt clusters (0.9 ? b ? 1.2)
• no associated tracks
• q ? 21 (rejects machine background)
• 390 lt M lt 600 MeV/c2

rtag/BR(KS?p0p0) ? 65 Background lt 1 (machine)
60
KL interaction in EMC
Clean signature (late neutral cluster) may be
used for tagging KS
Reconstruction of b is sensitive to boost and
machine energy spread dEf 1 MeV ?db 0.004
61
Tagged KL charged vertex selection

• KL ? charged selected by
• KL tag (from KS?pp- here)
• 1 charged vertex in 5 cone about pL pf - pS

20464 events in FV (0.4 pb-1)
lL 33066 cm
62
KL?pp- isolation from KL?pln, pp-p0
Apply a loose cut Pmiss lt 10 MeV/c M2miss lt (70
MeV/c2)2 55 events found 52 events expected No
efficiency correction
M(KL?pp-)
63
Kl3 background rejection
In mp hypothesis for both tracks, Kl3 has
M2missgt0 for zero En
Before (Pmiss,Mmiss) cut
3 additional events rejected
m22
m2
After (Pmiss,Mmiss) cut
m1
m12
64
Kl3 rejection using TOF from EMC
KL?pen

• Independent Kl3 backround
• rejection provided by b
• TOF from EMC
• p, L from DC
• b alone provides useful
• information for Ke3 rejection

p (MeV/c)
p
e
b may provide additional rejection for Km3 when
combined with energy deposition profile in EMC
b
65
KL neutral vertex reconstruction

• EMC measures
• impact point of g L
• total time-of-flight t tL tg
• DC measures PS and LL Pf - PS
• Solve for LL

• Each g gives an estimate of LL
• Starting point for kinematic fit

With Eg ? 100 MeV (gs from KL?p0p0)
st ? 70 ps/?E (GeV) Obtain sL ? 0.7 cm for 4 gs
66
Reconstruction of KL ? p0p0

• KL ? p0p0 selected by
• KL tag (from KS?pp- here)
• 4 neutral clusters
• Loose collinearity cut about
• pL pf pS
• Cut on M(gg) about mp0 for
• two best pairings

M(4g) 483 MeV/c2 sM 22 MeV/c2
Curve double Gaussian fit to data (30 counts in
signal peak) Hatched MC background from KL?3p0
with 2 gs lost

• Additional background rejection from
• QCAL
• Kinematic fit

67
Fiducial volumes for KS and KL
KS (spherical) KL (cylindrical)
r 10 cm ? 16lS 30 lt r lt 150 cm, z lt 125 cm rFV ? 0.3
lS 0.6 cm lL 340 cm
rz distribution of KL decay vertices

• FVs for KL ? pp- and KL ? p0p0
• are nominally the same, however
• Vertex resolutions differ
• Convolution of decay distribution in double ratio
• Residual differences in vertex positions
• drFV 0.5 mm ? dR ? 310-4

68
Charged/neutral fiducial volumes
KL ? p p-p0
Check alignment of FVs using events with
reconstructible charged and neutral vertices from
same decay K ? p?p0 KL ? p p-p0
For KL ? p p-p0 (shown), residuals not expected
to differ from KL ?p0p0 but resolution is worse
KL?p p-p0 KL ? p0p0 K ? ????
Ng 2 4 2
Eg 6080 MeV 100120 MeV 100120 MeV
69
f ? p0p0g
Investigation of intermediate state in f ?Sg?
p0p0g (S scalar meson)
If S f0, BR(f ? f0g) and shape of dN/dMpp
sensitive to f0 structure If S f0s mixing ?
study dN/dMpp

• Current situation
• BR(f ?p0p0g) 1 ? 10-4 (SND and CMD-2)
• Various models can describe dN/dMpp

• Analysis scheme
• f ? hg ? ggg Obtain sf for normalization, check
  systematics
• f ? p0g ? ggg Check systematics (10 lt Eg lt 500
  MeV as in f ? p0p0g)
• ee- ? wp0 ? p0p0g Principal background to f ?
  p0p0g
• f ? p0p0g

70
f ? hg ? ggg PRELIMINARY
Event selection 3 neutral clusters with t-R/c
lt 5st and q gt 21 No prompt hits on QCAL Clusters
assigned by kinematic fit (E,p) conservation, vg
c, M12 mh Main background f ? p0g Erad(p0g)
501 MeV, fit assigns to h Constrains energy of
remaining g Nhg 18504 Efficiency from MC BR(f ?
hg ? ggg) (0.49 0.02)
Erad 362.5 MeV sE 34.9 MeV
Mh 546.3 MeV sM 41.8 MeV
sf (3.19 0.02(stat) 0.26(syst)) mb 1.85
pb-1 KLOE preliminary sf (3.114 0.034
0.048) mb (CMD-2) Complete eval. of systematics
in progress
71
f ? p0g ? ggg PRELIMINARY
Event selection and fit similar to hg Signal Ep
501 MeV, q12 small Main backgrounds ee- ? gg
with 1 split cluster ee- ? gg with ISR dN/d(cos
q12) almost uniform Background estimated assuming
flat 2 clusters with max E back-to-back Nobs
5931 Nbkg 2144 30 Efficiency from MC
Erad 501.6 MeV sE 40.9 MeV
Mp0 133.8 MeV sM 16.0 MeV
1.85 pb-1 KLOE preliminary R3g 3.75
0.08(stat) 0.24(syst) R3g 3.77 0.40
(PDG) Complete eval. of syst. in progress
?L dt cancels from R3g
72
f ? p0p0g
First kinematic fit with minimal constraints to
improve parameter estimates Global (E,p)
conservation, vg c Select best g-pairing (of 15
possible) M(gg)1,2 closest to mp0 Second
kinematic fit in p0p0g and wp0 hypotheses with
addl constraints M(gg) mp0, M(p0g)mw
Background to f ? f0g ? p0p0g S/B
ee- ? wp0 ? p0p0g 0.52
f ? rp0 ? p0p0g 3.0
f ? a0g ? hp0g ? 5g 3.1
f ? hg ? p0p0p0g (2g lost) 0.02

• Angle between g and p0s in p0p0 CM
• flat distribution for f0g (Jf0 0)
• characteristic shape for wp (Jw 1)
• In wp fit, M(p0g) mw
• Resulting distortion of angular
• distribution allows background
• discrimination

73
f ? p0p0g PRELIMINARY
s(ee- ? wp0 ? p0p0g) Nobs 529 Nbkg(MC) 93
7 1.85 pb-1 KLOE preliminary (0.65 0.04
0.07) nb s (0.64 0.08) nb SND BR(f ? M(gt700
MeV)I0g ? p0p0g) Nobs 307 Nbkg(MC) 114
11 1.85 pb-1 KLOE preliminary (0.80 0.10
0.08) 10-4 BR (1.14 0.10 0.12) 10-4
SND BR (1.06 0.09 0.06) 10-4 CMD-2
Complete analysis of systematics in progress
74
f ?hp0g
5g and 9g final states

• Expected contributions to
• f ? hp0g
• f ? Vp0, V ? hg
• V vector meson r (w)
• BR 5 10-6
• f ? a0g, a0 ? hp0
• a0 scalar meson, I 1, I3 0
• KK, qq BR 10-5
• qqqq BR 10-4

Background S/B (incl. misrecon prob.)
5g f ? hp0g ? p0ggg 5g f ? hp0g ? p0ggg
f ? f0g ? p0p0g f ? rp0 ? p0p0g ee- ? wp0 ? p0p0g 1/10
f ? hg ? ggg (3g) f ? hg ? p0p0p0g (7g) 1/2
9g f ? hp0g ? p0p0p0p0g 9g f ? hp0g ? p0p0p0p0g
f ? KSKL ? 5p0 (10g) 1/60
f ? hg ? p0p0p0g (7g) 1/4
Measurements of BR(f ? hp0g) (0.83 0.23) 10-4
CMD-2 (0.89 0.14 0.06) 10-4 SND
75
f?hp0g
5g analysis 5 prompt gs, no tracks, Etot lt 900
MeV Minimal kinematic fit to improve
params. Construct c2 for g-pairing in hp0g,
p0p0g, hg hypotheses c2(hp0g) lt 12 c2(hp0g) lt
c2(p0p0g) c2(hg) gt 2 Kinematic fit to best
pairing with c2 cut M(gg) mp0 M(gg)
mh Background estimated from MC
9g analysis 9 prompt gs, no tracks All Eg lt 340
MeV to suppress hg Erad(hg) 363 MeV, very
effective t-R/c lt 2st to suppress KSKL LKL lt 8
cm (eliminates 98 of KL) Minimal kinematic fit
to improve params Construct c2 for g-pairing in
hp0g hypothesis Kinematic fit to best pairing
with c2 cut M(gg) mp0 (4) M(p0p0p0)
mh Background from f ? KSKL estimated from MC
76
f?hp0g PRELIMINARY
5g analysis (2.35 pb-1) Nobs 153 Nbkg(MC)
84 BR(f ? hp0g) (0.78 0.15(stat)) 10-4 9g
analysis (1.85 pb-1) Nobs 41 Nbkg(MC) 14.6
3.8 BR(f ? hp0g) (0.91 0.24(stat))
10-4 First measurement in this channel
Erad from 5g analysis
Evaluation of systematic errors in progress
SND BR (0.89 0.14 0.06) 10-4 CMD-2 BR
(0.83 0.23) 10-4
77
f ? h?g, hg
Theoretical predictions for BR(f ? h?g) range
from 2 10-4 down to 10-6 in case of large
gluonic component of the h?
h and h? can be related to the SU(3) octet and
singlet states h8 and h0 through the mixing angle
qP
Precise measurement of qP can help to distinguish
between phenomenological and chiral-perturbative
approaches

• Measure h,h? ? pp-ggg and h,h? ? 7g
• independent measurements for each channel
• check selection efficiencies for h? by estimating
  well-known BR for f ? hg
• many systematic effects cancel in the ratio BR(f
  ? h?g)/BR(f ? hg)

78
f ? h?g, hg

• Charged/neutral analysis
• Selection criteria
• 1 vertex from origin, 3 prompt neutral clusters
• f? hg ? pp-ggg
• Egmax gt 300 MeV
• Ep Ep-lt 550 MeV
• f? h?g ? pp-ggg
• Eggg gt 520 MeV
• Ep Ep- lt 412 MeV
• Selection efficiency
• rhg 48.3, reject(pp-p0) 103
• rh?g 43.8, reject(pp-p0) 104
• Unambiguous assignment of radiative g
• Kinematic fit with constraints
• M(gg) mh and M(hpp-) mh? (for f ? h?g)
• M(gg) mp0 and M(pp-p0) mh (for f ? hg)

79
f ? h?g, hg

• Fully neutral analysis
• Selection criteria
• 7 prompt neutral clusters
• Mf - Etot lt 2sE
• Egmax gt 300 MeV (for f ?hg)
• Selection efficiency
• rhg 42.8
• rh?g 46.7
• Pairing
• Identification of radiative g sufficient for f ?
  hg
• Best g-pairing chosen for f ?h?g
• Kinematic fit with constraints
• M(6g) mh (for f ? hg)
• M(g1g2) mh
• M(g3g4) M(g5g6) mp0 (for f ? h?g)
• M(p0p0h) mh?

Eg from f ? hg
Eg from f ? h?g
80
f ? h?g, hg

• Background for f ? h?g
• KSKL and pp-p0 backgrounds removed as for f ? hg
• Only substantial background to f ? h?g is
  then f ? hg, suppressed by
• Cuts in Pc2(h?g) vs Pc2(hg) plane
• Egmax gt 345 MeV (7g final state)
• pp-ggg final state
• S/B gt 35 (rel. to hg), rsel 22.9
• 7g final state
• S/B gt 20 (rel. to hg), rsel 12.6

• Background for f ? hg
• f ? hg ? 7g is background free
• Backgrounds to f ? hg ? pp-ggg from
  misreconstruction
• f ? KSKL ? 1 charged and 1 neutral vertex
• f ? pp- p0 (1 split cluster)
• Event selection criteria
• cut on Pc2(hg) gt 1
• S/B gt 900

81
f ? h?g, hg PRELIMINARY
82
s(ee-? hadrons)
Anomaly in muon magnetic moment am (gm-2)/2
amQED amhadr amweak ( amelse ?)
results from hadronic vacuum polarization is not
calculable using pQCD can be determined from
s(ee- ? hadrons) for ?s lt 5 GeV by dispersion
integral
amQED ( 11 658 470.7 0.03)
10-10 amhadr ( 685.0 16.0 )
10-10 amweak ( 15.1 0.4 )
10-10
Currently amtheor ( 11 659 171 16)
10-10 amexp ( 11 659 230 85) 10-10
E821 expects to measure amexp to 4 10-10 Need
s(ee- ? hadrons) to 1 to test amweak and look
for new physics
KLOE can measure ds/dQ2(ee- ? hadrons) using
ee- ?pp-g withg radiated in initial state
(ISR) and Q2 M2(pp-) W2 2EgW Restricted
to Q2 lt W2 66 of dam comes in r mass
region Requires good separation of ISR/FSR Can
be handled by topological cuts at 99
level Requires precise understanding of
ISR Errors on ?L, W enter only once
83
ee-? pp-g PRELIMINARY
Analysis of FSR and f ?f0g
50? lt qp lt 130? 60? lt qg lt 120? Cuts ISR,
enhances FSR and f ?f0g

• 2 track vertex at IP
• 1 prompt neutral cluster
• Kinematic cuts to isolate pp-g from ee-g,
  mm-g, and pp-p0
• Likelihood cut to add rejection for ee-g using
  TOF, energy profile in EMC

MC
data

• Comparison with MC
• Simulation includes ISR and FSR
• Detailed study of efficiencies (incl. trigger)
  from data

ds/dQ2 (nb/GeV2)
Q2 GeV2
84
ee-? pp-g PRELIMINARY
1.8 pb-1 KLOE preliminary Data QED background
for Eg 20120 MeV 35 160 events Assigning
all events to f0 resonance and interference BR(f
?pp-g) (0.2 0.9) 10-4 lt 1.64 10-4
90 CL
MC
data
BR(f ? f0g ? p0p0g) (0.80 0.10) 10-4
KLOE (0.50 0.06) 10-4 SND Some suggestion of
destructive interference between f0g and FSR
contributions to pp-g
ds/dQ2 (nb/GeV2)
Q2 GeV2
85
f ? pp-p0
2.14 pb-1 ? 330K events
3 contributions to Dalitz plot f ? r?,0p0,? f ?
pp-p0 (direct) ee- ? wp0 Amplitudes of rp and
direct terms not well established Analysis of
Dalitz plot sensitive to r line shape Measure
r?,0p0,? simultaneously

• 2-track vertex from origin
• 90 lt Mmiss lt 180 MeV/c2
• 2 prompt neutral clusters
• cos qgg lt -0.98
• q- lt 175? (reject ee-?ee-gg)

E0 mp0 GeV
(E - E-)/?3 GeV
86
f ? pp-p0 PRELIMINARY
Full functional dependence of efficiency and
experimental resolution not yet included
Statistical sensitivity to r parameters already
comparable to world averages
Fit to Dalitz plot with amplitudes for f ?
r?,0p0,?, f ? pp-p0 (direct), ee- ? wp0, and
all interference terms
Fit with m? m0 mr Fit with m, m0 Compare to
c2/dof 1205/919 1115/918
m0 775.10.7 MeV/c2 779.8 0.8 MeV/c2 770.0 0.8 PDG 98 global avg 776.0 0.9 PDG 98 mix chg t and ee-
m-m0 -5.0 0.6 MeV/c2 0.1 0.9 PDG 98 avg
Gr 143.2 1.7 MeV 145.3 1.7 MeV 150.7 1.1 PDG 98 avg
adirect 0.141 0.012 0.130 0.010 -0.16 to 0.11 CMD-2 98
fdirect (120 5) (108 5)
87
Data taking in 2000

• Next milestone 100 pb-1
• (1 year at L 1031 cm-2 s-1 ? 100 pb-1 ? 108 KL)
• Measurement of Re(e?/e) to 10-3 statistical
  accuracy
• thorough workout for measurement systematics
• Measurement of Kl3 form factors
• Measurement of KS semileptonic decays
• Measurement of regeneration cross section at p
  100 MeV
• Limits on C-violation in h decays
• Complete analysis of f radiative decays
• Measurement of s(ee- ? hadrons) around mr to 1

88
Beams circulating in the machine... Collisions
scheduled for April 9... ...KLOE is ready to go!