1. Heavy quarks and leptons ‘02
KLOE results on Kaon decays and F radiative decays
E. De Lucia
(INFN Roma)
for the KLOE Collaboration
Heavy Quarks and Leptons 2002, Vietri 27/5-1/6
E. De Lucia
Heavy quarks and leptons ‘02

2. Heavy quarks and leptons ‘02
DAFNE parameters
Design parameters
Beam energy : 510 MeV
Max number of bunches : 120
Bunch spacing : 2.7 ns
Bunch current : 40 mA
Single bunch luminosity : 4·1030 cm2 s1
DEAR
50 m
L = 5 ·1032 cm2 s1
E. De Lucia
Heavy quarks and leptons ‘02

3. Heavy quarks and leptons ‘02
DAFNE Performance
KLOE Integrated Luminosity
april ’99  december ‘01
pb-1
1999
(25 pb-1  7.5 x 107 f) analysis completed
2000
(190 pb-1  5.7 x 108 f) analysis in progress
2001
1.0
0.8
0.6
0.4
0.2
0.0
s(e+e-  KSKL )(mb)
F resonance
s (MeV)
2001 day performance: peak average
L(cm2 s1) · · 10 31
day L dt (pb1)
MAY ’02 new data taking started
L(cm2 s1) · · day L dt (pb1)
machine background reduction of a factor 23
E. De Lucia
Heavy quarks and leptons ‘02

4. Heavy quarks and leptons ‘02
The KLOE detector
detection of a KS (KL) guarantees the presence of a KL(KS) with known momentum and direction
TAGGING
Neutral kaons are produced in a pure quantum state (JPC=1--) :
(contamination from KSKS,KLKL < 10-9)
EMC
pK  110 MeV/c
S = 6 mm L= 3.4 m
4 m diameter × 3.3 m length
90% helium, 10% isobutane
12582/52140 sense/tot wires
All-stereo geometry
7 m
20 < Eg < 500 MeV DC
6 m
Lead/scintillating fiber
98% coverage of solid angle
88 modules (barrel + end-caps)
4880 PMTs (two side read-out)
155 < pp < 265
E. De Lucia
Heavy quarks and leptons ‘02

5. The KLOE detector performances
mpp = MeV/c2
sm = 1 MeV/c2
KS  p+p-
sr f (mm)
sp/p = 0.4 %
sr f = 150 mm
sz = 2 mm
sV = 3 mm
p0  gg
h  gg
sE /E = 5.4% /
st = 54 ps /
s(t-r/c) (ps)
50 ps(cal) 120 ps (coll.time)
E. De Lucia
Heavy quarks and leptons ‘02

6. DAFNE energy and bunch-crossing time
 s from the b of KL interacting in the EmC
s from KS pions and Bhabha electrons momenta measurement
Energy
s (MeV)
At current luminosity Dafne energy at 0.01% within 1 minute of data taking using KL b, KS energy and Bhabha scattering events
Bunch-crossing time
e+e-
large spread of the trigger time with respect to the crossing time but the signal is synchronized with machine RF/4 (10.8 ns) at 50 ps level
Off-line determination of the bunch-crossing time T0
10.8 ns
hp: the fastest cluster with time tCLUS in EMC is due to a prompt  from I.P. with time of flight tTOF = RCLUS/c :
(ECLUS > 50 MeV and CLUS > 60 cm to avoid selection of “accidental” clusters)
E. De Lucia
Heavy quarks and leptons ‘02

7. The KLOE physics program
KS physics :
BR(KS  p+p-(g))/BR(KS  p0p0)
BR(KS  pen)
f radiative decays
f  f0g, a0g
f  h’g, hg
2000 statistics:
 first publications
O( 20 pb-1 )
rare KS decays
KL form factors, KL  2p, KL  gg
K ± decays
s(e+e- p+p-) via ISR
h decays
2001 statistics:
on tape analysis in progress
O( 200 pb-1 )
e’/e via double ratio
Semileptonic asymmetry (CPT test)
KLKS Interferometry
E. De Lucia
Heavy quarks and leptons ‘02

8. Recent results in publication
Using a sample of 5 x 107 f decays (2000 data):
(KS  p+p-() )/ (KS  p0p0)
hep-ex/ , accepted by Phys.Lett.B
BR(KS  pen)
Phys.Lett. B (2002)
BR(f  h p0 g )  BR(f  a0 g  h p0 g )
Phys.Lett. B (2002)
BR (f  p0 p0 g)  BR(f  f0 g  p0 p0 g )
Phys.Lett.B (2002)
BR(f  h’g) /BR( f  hg)
to be submitted to Phys.Lett. B
E. De Lucia
Heavy quarks and leptons ‘02

9. Heavy quarks and leptons ‘02
KS analysis at KLOE
b* after correction for different pion
velocities and trigger latency
KS tagging using time of flight identification of KL interacting in the EmC (“KL-crash”) selected as a calorimeter cluster with:
Eclus > 200 MeV
| cos(qclus)| < 0.7
0.195  b* 
(b* = KL velocity in the f rest frame)
=0.004  WDANE= 1 MeV
b* ~ 0.218
KL-crash
KS  p+ p -
 KS momentum from KL cluster position
 easier determination of trigger efficiency % of times , KL-crash triggers by itself
 relies on the measurement of b* slightly dependent on KS decay
2000 data analysis = 5.4x106 KL crash
2001 data on tape ~ 6 x107 KL crash
E. De Lucia
Heavy quarks and leptons ‘02

10. (KS  p+p - ())/(KS  p0p0)
Motivations
first step towards Re(e)
extraction of Isospin 0 and 2 amplitudes and phases from consistent treatment of soft  in KS  p+p -()
KL-crash .and.  3 neutral “prompt” clusters:
|tCLUS-RCLUS/c| < 5st .and.
ECLUS > 20 MeV .and.
cos <0.9
KS  p0p0 selection
acceptance and E cuts correction from MC
KS   +  - () selection
KL-crash.and.2 tracks from IP reaching EMC(*)
120 < Ptrack < 300 MeV/c .and. 30o <  < 150o
fully inclusive measurement  no  required in EMC Eg
cosqg
(*)
acceptance depending on the photon energy E*
epp(g)(E* ) from MC folded to theoretical  spectrum
Ptrack (MeV/c)
MC- signal
DATA
E. De Lucia
Heavy quarks and leptons ‘02

11. Efficiency evaluation
KS   +  - ()
KS  p0p0
single-track reconstruction efficiency from KS  p + p - data, used to scale MC
e+ - (sel and rec) = (57.6 ± 0.2) %
photon detection efficiency from data using f  p+ p- p0 events. The overall selection efficiency (mostly acceptance):
e00 = (90.1 ± 0.2)%
single-particle t0 and trigger efficiencies using data: KS  p+p - from KL  p +p -p0 –tagged sample (t0-independent), f  p +p -p0 and KS  p+p - with 2 trigger sectors fired by KL-crash plugged into MC
e + - (t0 and trig) = (97.9 ± 0.03) %
Cluster cuts: Prompt means t-R/c < min(5sig, 3ns), E > 20 MeV, cos theta < 0.9 (theta > 26 deg)
Photon efficiency not parameterized by single cluster as in Ke3 analysis, but parameterization put in MC via EmCFakeThresh module
T0 efficiency is actually from MC
t0 and trigger efficiency using data: KS  p0 p0 from KL  p +p -p0 –tagged sample (t0-independent), KS  p0p0 with 2 trigger sectors fired by KL-crash
e00 (t0 and trig) = (99.86 ± 0.04)%
E. De Lucia
Heavy quarks and leptons ‘02

12. (KS  p+p - ())/(KS  p0p0)
Result
Statistical error (%)
0.14
0.20
photon counting
0.23
trigger and t0
0.26
tracking
0.68
Overall systematic error
0.55
KS  p0p0/KS  p+p- tag %
Contribution to systematic error
Nev (KS  p+p- ) = x 106
Nev (KS  p0p0 ) = x KLOE 2000 data
2.239  0.003stat  0.015syst
PDG 2000 average  (without clear indication of E* )
With 2001 data (180 pb-1) improvement s on:
absolute scale  tagging efficiency bias
statistics of control sub-samples
Eg* spectrum
Cluster efficiency biggest error because EmCFakeThresh used to evaluate. Accidental clusters and splitting also discrepant
Tag bias may go down when data with wider cuts analyzed
Syst error on inclusive measurement due to contribution to KL -> p+p-g must be insignificant:
Eff vs. Eg is 15% lower at 20 MeV than at 0 MeV, and falls off slowly. Meanwhile, Eg spectrum falls off sharply (1/Eg).
So vast majority of p+p-g decays are detected.
E. De Lucia
Heavy quarks and leptons ‘02

13. Heavy quarks and leptons ‘02
BR( KS  p e  n)
Motivations
if (CPT).and.(DS.eq. DQ) then
BR( KS  p e  n) = BR( KL  p e  n)x GL/ GS
from PDG values = (6.704  0.071)x10-4
only one measurement (CMD ): (7.2  1.4)x10-4
KL-crash.and.charged vertex at IP (r<8cm , |z|<10cm) .and.2 tracks with associated EmC clusters
invariant mass of the tracks in p hp Mpp < 490 MeV/c2 (against background from KS  p+p-)
p/e identification using time-of-flight
Emiss - |pmiss|
Selection
b  0.8
t  9 ns
b = 1
t  7 ns e p
Ddt (p,e) =[t1CLU - t2CLU] – [L1 /c (p) –L2 /c (e)]
|Ddt (p, p)| > 1.5 ns to reject KS  p+p-
Cuts on Ddt (p, e) and Ddt(e, p)
p/e identification using time-of-flight
Vertex cuts: dxy < 8 cm, |z| < 10 cm. Preselection in P* vs Mpp: Neither is peaked since nu carries away momentum, anticorrelation since small Mpp = smaller mass of track system, larger momentum for same energy ?
TCA cut: dtTCA < 30 cm
Purity of Klpen sample: contmination < 0.3%
E. De Lucia
Heavy quarks and leptons ‘02

14. Efficiency evaluation
2000
Fit to Emiss-Pmiss spectrum using MC
spectra for signal and p+p- background
Normalization to KS  p+p- decays
Vertex reconstruction, fiducial cuts and Mpp efficiency from MC but also from data KL  pen near I.P. (high-purity sample (> 99.7 %), by kinematic cuts) and KS  popo to scale MC
Tracking efficiency for MC and data from KS  p+p-
Single-particle t0 , track-cluster, and trigger efficiencies from data using KL  pe n near origin and KS  p+p- but also fp+p-p MC efficiency scaled accordingly
Time of flight ID efficiency from KL  pen decays near origin and KS  p+p
Overall selection efficiency:
(20.8 0.4)% 
Single particle efficiencies: KL used for all e and for low-p pions, pi-pi used for most of p range, pi+pi-pi0 for high-p
Separate overall t0-TCA-trig measurement possible from KL decays alone
Fit takes into account finite statistics in MC background
Normalization sample same as in BR(+-)/BR(00) analysis
Standard t0/trig unbiasing and isolation cuts when obtaining single-particle efficiencies from pi+pi- and pi+pi-pi0
E. De Lucia
Heavy quarks and leptons ‘02

15. Heavy quarks and leptons ‘02
BR( KS  p e  n )
CPT and DS=DQ predicts: G(KS  p e n ) = G(KL  p e n ) and then: BR(KS  p e n ) = BR(KL  p e n ) x (GL/GS)
Using PDG:
BR(KS  p e n ) = ( ± ) x 10-4
KLOE 2000 data, (6.79  0.33stat  0.16syst)x  30 evts
CMD , (7.2  1.4)x  13 evts
Result
Main contributions to the total error %
Statistics
4.9
Tracking + vertex efficiency
2.0
Cluster, t0 , trigger
0.9
TOF selection eff
0.8
Tag eff
0.6
Total
5.9
lower with the 2001 data !!
600 lb systematic gorilla is track and vertex efficiency
E. De Lucia
Heavy quarks and leptons ‘02

16. F Radiative decays: f  (f0g) p0p0g; f  (a0g) hp0g
Motivations
various models for f0 and a0 mesons : qqqq state, KK molecule, ordinary qq meson.
f  f0g , a0g BR and mass spectra sensitive to f0 ,a0 nature [Achasov, Ivanchenko 1989]: phenomenological framework (kaon loop model) coupling constants
g(fKK) from G(fK+K-)
g(f0KK) g(a0KK) f0 , a0 model
g(f0pp) g(a0hp) M(p0p0) M(hp) spectra
f0, a0
Kaon loop final state f
radiative g
a0 analysis : uses f  a0g  (hp0 ) g gg
2tracks + 5g final state
p+p-p0
5g final state 
first observation
0.2 nb taken for indicative CS into f0g and a0g for background est.
f0 analysis : uses f  f0g  (p0p0) g
5g final state
Overlap = structure dependent function k = f0 momentum
E. De Lucia
Heavy quarks and leptons ‘02

17. Heavy quarks and leptons ‘02
F Radiative decays: f  f0g , a0 g  5 g final state
Selection
5 “prompt” g with Eg > 7 MeV
| cos| < 0.93 (to avoid background from I.P.)
5Ei > 700 MeV to reject KLKS neutrals
kinematic fit (4-mom. + |t-r/c| conservation ) 1
0
000
00
|M(1) - M (2) | (MeV)
M (MeV/c2)
e+e  wp0 w  p0g f  00 f  hp0g h  gg
f  hg h  gg g final state
Background sources
5g final states
h  p0p0p g final state
in hp 1.
0.2 nb taken for indicative CS into f0g and a0g for background est.
Photon pairing in the hypothesis
kinematic fit with mass constraint
00
0
000 (M(0)=M())
3 (rejecting events  and  with E  =363 MeV)
for (2)  e 0 mass
for (1) e (3) two 0
2 cut
E. De Lucia
Heavy quarks and leptons ‘02

18. F Radiative decays: f  p0p0g (f0g)
After 2 cut in 00 hp + DMgg cut + background sub.
Results
Nev = 2438  61
BR(f  p0p0g ) = (1.09  0.03stat  0.05syst)x10-4
CMD (0.92 0.08 0.06)x10-4
SND (1.14 0.10 0.12)x10-4
(from the fit to the spectrum)
The fit to the M0 0 spectrum (kaon loop)
contributions from f  f0g and f  sg with “strong” destructive interference (Mpp < 700 MeV) Negligible contribution from f  r0p0 p0p0g
0.2 nb taken for indicative CS into f0g and a0g for background est.
M(f0) =  MeV
g2(f0KK)/4p =  GeV2
g(f0pp) /g(f0KK) =  0.01
g(fsg) =  0.008
BR(f  f0g  p0p0g ) = (1.49  0.07)x10-4
E. De Lucia
Heavy quarks and leptons ‘02

19. F Radiative decays: f  hp0g (a0g)
Sample (a) from h  gg final state
(a)
(b)
Comb. fit
M0 (MeV)
Parabolic cut to reject 0 .and. M < 760 MeV to reject f0 .and. 3 cut on M to reject  000
BR(f  hp0g) = (8.5  0.5stat  0.6syst)x10-5
Main background :p0p0g.
(Nev = Nbck = 309  20 )
Neutral selection as for (a) .and. 1 vtx in I.P. + 2 tracks .and. Minv(p+p-) < 425 MeV to reject KS  p+p .and.kinematic fit with mass constraint on  and 0
BR(f  hp0g) = (8.0  0.6stat  0.5syst)x10-5
CMD-2 (9.0 2.4 1.0) x 10-5
SND (8.8 1.4 0.9) x 10-5
Sample (b) from h  p+p-p0 final state
Combined fit to the Mhp0 spectra
using PDG for a0 mass ( MeV )
g2(a0KK)/4p =  0.04 GeV2
g(a0hp) /g(a0KK) =  0.09
0.2 nb taken for indicative CS into f0g and a0g for background est.
Negligible background with the same topology .
(Nev = Nbck = 4  4)
BR(f  a0g  hp0g) = (7.4  0.7)x10-5
E. De Lucia
Heavy quarks and leptons ‘02

20. Summary of the results on f0 and a0
Within the context of the kaon-loop model with point-like coupling of scalars to kaon pairs, we have obtained:
KLOE q model
g2f0KK/(4)  ”super-allowed” (few GeV2)
(GeV2)
gf0 /gf0KK  f0
g2a0KK/(4) 0.40  ”super-allowed” (few GeV2)
ga0/ga0KK  a0
f0 parameters compatible with 4q model
a0 parameters not well described by the 4q model (2001 data  more accurate study of a0)
E. De Lucia
Heavy quarks and leptons ‘02

21. F Radiative decays: f  h‘ g ,hg
Motivations
Precise measurements of BR(f  h’g) and BR(f  hg) and the determination of h-h’ mixing angle probe hidden strangeness and gluonium content of h’
3 “prompt” g with Eg > 7 MeV and q > 21o .and tracks vertex in IP
Preliminary kinematic fit: conservation of total E, p and b = 1 for each g
Simple kinematic cuts to eliminate background:
f  p+p-p0 with extra g
f  KSKL  p+p- p0 p0 with g lost
negligible background but N(a)N(b)/100
Selection
Used decay chains: a) f  h’g  p+p-hg  p+p-3g
 hg  p+p-p0g  p+p-3g
the topology is the same b)
E. De Lucia
Heavy quarks and leptons ‘02

22. F Radiative decays: f  h‘ g ,hg
Result
Invariant mass spectrum of h’g
N(b) =  220
N(a) = 120  12stat ±5bck R = BR(f  h’g)/BR(f  hg )
KLOE 2000 data
R = (4.70  0.47stat  0.31syst) x 10-3
From R we obtain the pseudoscalar mixing angle:
(flavor basis)
(octet-singlet basis)
p= ( 41.8  1.7)o [ qP = (-12.9  1.7)o ]
BR (f  h’g ) x 10-5 10 8 6 4 2
2000 data
Using PDG value for BR(f  hg) we obtain:
BR(f  h’g ) = (6.10  0.61stat  0.43syst) x most precise determination to date
CMD SND KLOE
consistent with the gluonium contents of the h‘ < 15%
E. De Lucia
Heavy quarks and leptons ‘02

23. Heavy quarks and leptons ‘02
Conclusions
DAFNE improved continuously during the first two years of KLOE data taking and improvements on the background are already visible in data taking
The KLOE detector is performing well
With 2000 data we measured:
(KS  p+p- (g))/ (KS  p0p0), BR(KS  pe n)
f  f0g  p0p0g, f  a0g  hp0g
With 2001 data we will improve the results and expect results from:
KL  2p, K ± decays, KL  gg decays
Rare KS decays ( BR ~ 10-5)
s(e+e-  p+p-) by ISR
 decays (6106  tag from f  hg Eg = 363 MeV photon)
E. De Lucia
Heavy quarks and leptons ‘02