1. KLOE Results and Prospects for KS Physics
T. Spadaro* for the KLOE Collaboration
*Università di Roma “La Sapienza”

2. The KLOE Detector sE/E = 5.7 % / E(GeV) st = 54 ps / E(GeV)  50 ps
Lead/Scintillating-Fiber calorimeter
sE/E = 5.7 % / E(GeV)
st = 54 ps / E(GeV)  50 ps
(bunch fluctuations subtracted)
sp/p = 0.4 %
(for 90° tracks)
He-iC4H10 drift chamber
KLOE detector immersed in a 6 kG magnetic field.
Lead/Sci calorimeter: Good energy resolution, excellent timing performances
Low Z material, He based mixture, large drift chamber, very good momentum resolution
(better 150 microns point resolution)
Trigger is performed using both DC and Calorimeter information

3. KLOE data taking 21031 cm-2s-1 exceeded 1 pb-1 per day reached
35 pb-1 collected
Average trigger rate 2.1 kHz
Year 2000 (23 pb-1)
Stream
Events (M)
Volume (GB)
K+ K- 19
687
KS KL 64
2270
r p 6
166
radiative 23
520
Bhabha
127
2623
10.9 M KL tagged by KS  p+p-
7.2 M KS tagged by KL interactions in EmC
We took data during 3 different periods
Events are triggered, reconstructed and classified according to the phi decay mode.
[KsKl stream is largely redundant, not to introduce any bias in cp violation measuerement]
Useful control samples for calibration

4. KS Physics at KLOE
Integrated Lum. (pb-1)
CPT limits (KS  p+e -n /p-e+n vs KL  p+e -n / p-e+n)
rare and not so rare KS decays (gg, pee, pmm, pnn)
BR(KS  p+ p-) / BR(KS  p0 p0)
BR of semileptonic KS decays (KS  pe n)
Upper limit on BR(KS  p0 p0 p0 ) 10-5
dN/dEg (KS  p+ p-g)
2000
1000
100
+-/00 is part of the CP double ratio. 20

5. Measuring ratios of KS BR’s
KS beam: tagged by KL interactions in the calorimeter
G (KS channel a )
G (KS channel b)
observed
estimated
desired
eS selection efficiency
a,b
a,b
a,b
NS - BkgS = NKK ´ eS(tag) ´BRS ´ eS
NKK=òL dt ×sf ×BR(f®K0 K0)
drops out identically
eS (tag) almost cancel out
With tagged beams at KLOE one can also measure absolute BR’s

6. KS tagging
Clean KS tagging by time-of-flight identification of KL interacting in the EmC
(b* = KL velocity in the f rest frame)
Selection cuts:
Eclus  100 MeV
|cos(qclus)|  0.7
b*  [0.195, ]
b*distribution: nominal KL velocity 0.218
Since kaons are almost monochromatic, the velocity should be around .2
We can select the so-called Klong crash events

7. T0 fixing
Absolute time-zero of the event is fixed a posteriori using the fastest cluster in the event
Global time is synchronized to the bunch crossing frequency
g hypothesis is used at the first stage of the reconstruction
Tag efficiency is slightly dependent on the KS decay due to the different global t0 estimations, given by:
prompt g’s in KSp0p0
pion clusters in KS p+p-
pion or electron clusters in KS pen
+- clusters are slower than photons from 00, but we can estimate the difference in acceptance from data
and correct for that difference of about 5%
e+- / e00 = ( ± 0.005) %
Ratio of tag efficiencies is estimated directly from data

8. KS  p0p0 selection Selection: K crash | t – R/c|  min(5 st ; 3 ns)
cos q < 0.9
E > 20 MeV
prompt clusters distribution:
DATA vs. Monte Carlo
# prompt 3 4
 5
DATA
37.5 %
61.8 %
0.7 %
MC thr
36.7 %
63.0 %
0.3 %
The selection is performed counting the number of prompt photon clusters.
The difference in counting between data vs Montecarlo is at the level of %
All numbers normalized to 3 cluster

9. KS  p0p0 Data vs Monte Carlo
Photon detection efficiency: estimated from data f p+p-p0
4 prompt clusters:
Energy spectrum and angular distribution
Total energy distribution 2 3 5 4
The photon detection efficiency is extreacted directly from data. The Montecarlo effective threshold has been tuned to fit to the energy shape of the efficiency.
Selection efficiency: e00 = (56.7 ± 0.1) %

10. KS  p0 p0 trigger efficiency
Assignment of the clusters to the KS and KL hemievent cutting at –15 ns
Cluster  trigger sector  unbiased multiplicities
1-etrig = S(0)•L(0) + S(1)•L(0) + S(0)•L(1)
p+ p-
p0 p0
tclu - tcrash
KL crash fragments
Dt (ns)
etrig = ± 0.03 (stat)± 0.02 (syst)
In this analysis calorimeter trigger only has been used.

11. KS  p+p- selection K crash 2 tracks from IP:
closest approach: |z|  10 cm; dxy  4 cm
first hit: |z|  40 cm; dxy  40 cm
acceptance and momentum cuts:
30< q <150
120  p  300 MeV/c (to reject f K+K-)
Both tracks have to impinge to the calorimeter
Reject machine background scattered off the quadrupole
Reject k+k- events

12. KS  p+p- selection efficiency
Conditional reconstruction efficiency is estimated from data subsamples
Selecting 1 track according to previous cuts, look for the second one with
p2, estimated = pf – pL – p1
The correlation between the two particles in the efficiency is mostly due to the acceptance
A flat correction is applied to the Monte Carlo
Systematic error is estimated comparing the plateau for different run periods
Selection efficiency:
e+- = (58.5 ± 0.1) %
Acceptance from MC…dominates the selection efficiency

13. t0 and trigger efficiency corrections
1. Single particle efficiencies are estimated using KS  p+p- subsamples and KL  pen , f  p+ p- p0 control samples t0
trigger
pion efficiencies from data
(in bins of pT and q) for p+, p-, m+, m-
2. Weighted mean of the Monte Carlo kinematics, for events with 2 selected tracks
et0+trg = eL  et0 + (1-eL) e1t0·trig = 96.5%
eL = KL crash trigger efficiency = (40.5 ± 2.5)%
3. Systematic error:
- Montecarlo statistics
- Errors on the efficiencies estimated from data

14. Contribution to systematics
KS  p+p-/KS  p0p0 result
PDG 2000 Ratio =  [1 ± 1.2  10-2 (stat) ± 0.6 10-2 (syst)]
KLOE 1999 Ratio =  [1 ± 0.4  10-2 (stat) ± 0.7  10-2 (syst)]
KLOE 2000 Ratio = 2.23  [1 ± 0.35  10-2 (stat) ± 1.5  10-2 (syst)]
preliminary Systematics under study
Contribution to systematics %
Tag bias 1
KS  p0p0 selection
KS  p0p0 trigger
0.02
KS  p+p- selection
0.1
KS  p+p- trigger and t0
0.5
Total
1.5
Summary of the systematic error contributions. Perspectives on these.
Tag bias will be corrected using a different tag, e.g. KL to p+p-p0

15. About g from KS p+p- (g)
Theoretical estimates of BR(KS p+p- (g) ) / BR(KS p0p0 ) need definition of g cutoff energy
PDG: BR(KS p+p- (g) ) = (1.78 ± 0.4)  for Eg > 20 MeV, below our present statistical precision
Effective g energy cutoff due to signal selection (loose momentum cut) must be greater than 20 MeV
p+p- invariant mass distribution + presence of g in calorimeter will be used to explicitly measure g energy spectrum

16. KS  pen Events selection summary Kcrash tag Kinematics preselection
Track to cluster association
Time of flight identification
Close kinematics
Normalization to KS  p+p-
Kcrash tag bias estimation

17. Kinematics preselection
Two tracks of opposite charge from a vertex in the fiducial volume:
r < 4 cm and | z | < 8 cm
Cut:
invariant mass Mpp (pp hypothesis)
KS momentum PKS in the f rest frame
Mpp(MeV)
PKS (MeV)
Efficiency taken directly from Monte Carlo
e = (62.4 ± 0.3) %
Efficiency estimation from KL  pen decays before the drift chamber wall (work in progress)
Efficiency dominated by the acceptance contribution

18. Track to cluster association (TCA)
Extrapolation of both tracks to the calorimeter surface is required. Acceptance from Monte Carlo:
eGEO = (51.1 ± 0.2) %
Both tracks associated to calorimeter clusters: cut on impact parameter (dPCA )
dPCA (cm)
Among all clusters with dPCA< 30 cm, the one with maximum energy is used for time of flight identification
Acceptance is very low, due to the low pt electrons of this channel

19. TCA T0 and Trigger efficiencies
2 methods to estimate TCA, T0 and calorimeter trigger efficiencies:
Single particle efficiencies taken from data using KS  p+p- subsamples and KL  pen , f  p+p- p0 control samples. Efficiencies for p+ , p- , m+ , m- , e used to weight the Monte Carlo:
e = (82.1 ± 0.5) %
Efficiencies directly from data using KS  p+p- , KL  pen control sample, selected using DC information only.
e = (81.4 ± 0.6) %
Emiss(ep)-Pmiss (MeV)
Verifica sti numerelli e soprattutto gli errori
eTCA+T0+TRG = (81.7 ± 0.5) %
Emiss(pe)-Pmiss (MeV)

20. Time of flight identification
Main background from KS  p+p- (p  mn before the DC wall)
Mass hypothesis are used
dt (mass) = tcl – length / c b(mass)
Independence with respect to the global T0 estimation
Ddt (mass1, mass2) = dt (mass1) - dt (mass2)
Two cuts applied:
|Ddt (p, p)| > 1.5 ns
|Ddt (p, e)| < 1 ns, |Ddt (e, p)| > 3 ns
|Ddt (e, p)| (ns)
KS pen from MC
Data Summer 2000
Ddt (p, e) (ns)

21. Time of flight efficiency
Ddt distributions in KL  pen before the DC wall used to evaluate the efficiency
Efficiency for the cut |Ddt (p, p)| e1 = (91.5 ± 0.5) %
Efficiency for the cut |Ddt (p, e)| e2 = (91.8 ± 0.5) %
Total efficiency eTOF = (84.0 ± 0.6) %

22. Kinematic identification
Kinematics closed with KS momentum estimated using:
- direction of the KL cluster
- f 4-momentum
KS  pen and KS  p+ p- MC FIT
Data: 2000  11.5 pb-1
E(missing)pe = ES – Ep – Ee
Signal yield estimation:
Data fit using MC spectra for background and signal
Log-likelihood function takes into account contribution due to finite MC statistics
Signal
Emiss(pe)-Pmiss (MeV)

23. KS  pen results Data: 2000  11.5 pb-1
overall efficiency eTOT = (21.80.3)%
Yield N(KS pen)=283 19 events
N(KS pen)=248 17 events
N(KS pp) = 9.0 1.4 events
In a 5 MeV window
PDG 2000* BR(KS pen) = [7.2 ± 1.2]  (75 ± 13 events)
GS = GL BR(KS pen) = [6.70 ± 0.07]  10-4
KLOE BR(KS pen) = [7.47 ± 0.51(stat)]  10-4
preliminary Systematics under study
Qua devi mettere la referenza all’articolo di
VEPP-2M Phys. Lett. B456(1999)90-94

24. Conclusions and perspectives
The measurement of KS  p+p- / KS  p0p0 branching ratio has been presented.
A first measurement of KS  pen branching ratio has been obtained using a subsample of data. The statistical error is a factor of 2 lower than that of the only existing measurement.
Better understanding on systematics is needed:
Tag efficiency bias
Photon counting
Work is in progress on:
KS  p+p-g
KS  p0p0p0
KS  gg