1. Measurement of DE and INT in K±→±0g with NA48/2
Mauro Raggi
FlaviaNet Workshop on K decays
Frascati 18/05/2007

2. NA48/2 beam and detector LKr EM calorimeter Spectrometer
- 4 Drift Chambers (DCH) - Magnet
Mauro Raggi
FlaviaNet Workshop

3. Gamma production mechanismDE IB IB DE
INT
Inner Bremsstrahlung(IB) : (2.75±0.15)·10-4 PDG (55<T*p<90 MeV) Direct Emission (DE) : (4.4±0.8)· PDG (55<T*p<90 MeV) Interference (INT) : not yet measured
Mauro Raggi
FlaviaNet Workshop

4. K±→p±p0g amplitudes Two type of contributions:
Electric (J=l±1) dipole (E1)
Magnetic (J=l) dipole (M1)
Electric contributions are dominated by Inner Bremsstrahlung (EIB)
Thanks to Low’s theorem IB contribution can be related to the non radiative decay p±p0 using QED corrections.
DE shows up only at order O(p4) in CHPT
Is generated by both E and M contributions:
Magnetic contributions are dominated by chiral anomaly
Electric contributions come from L4 CHPT lagrangian and loops L2
Present experimental results suggest a M dominated DE
Mauro Raggi
FlaviaNet Workshop

5. General expression for decay rateIB DE
INT
P*K=4-momentum of the K± P*p=4-momentum of the p± P*g=4-momentum of the radiated g
Mauro Raggi
FlaviaNet Workshop

6. Experimental status for DE and INT
DE measurements
All the measurements have been performed in the T*p region MeV to avoid p±p0p0 BG
All of them are assuming the Interference term to be = 0
INT estimates [20][21]:
BNL E787
KEK E470
Mauro Raggi
FlaviaNet Workshop

7. E787 2000 measurement 20K K+ events Fit in 8 bins 0.1-0.9
Fit function:
with b set to 0 (INT=0).
Mauro Raggi
FlaviaNet Workshop

8. What’s new in NA48/2 measurement
In flight Kaon decays
Both K+ and K- in the beam (possibility to check CP violation)
Very high statistics (220K p±p0g candidates of which 124K used in the fit )
Enlarged T*p region in the low energy part (0<T*p<80 MeV)
Negligible background contribution <1% of the DE component
Good W resolution mainly in the high statistic region
14 bins in the fit to enhance sensitivity to INT
Order ‰ g mistagging probability for IB, DE and INT
Mauro Raggi
FlaviaNet Workshop

9. Enlarging T*p region
T*p(IB)
T*p(DE)
T*p(INT)
Standard region
Standard region
Standard region
The standard region 55<T*p<90 MeV is safe WRT K3pn BG rejection
Unfortunately the region below 55 MeV is the most interesting for DE and INT measurement
This measurement has been performed in the region 0<T*p<80 MeV
Mauro Raggi
FlaviaNet Workshop

10. Reconstruction strategy
We can get two independent determination of the K decay vertex: - The charged ZV(CHA) using the K and p flight directions (spectrom.) - The neutral ZV(NEU) choosing g pair with the best p0 mass (LKr)
ZV(CHA)
Once the neutral vertex has been chosen we also know which is the radiated g. g1 g3 g2
ZV(2-3)
ZV(1-3)
ZV(NEU)
LKr
Mauro Raggi
FlaviaNet Workshop

11. Main BG sources Decay BR Background mechanism K±→p±p0
(21.13±0.14)%
+1 accidental or hadronic extra cluster
K±→p±p0p0
(1.76±0.04)%
-1 missing or 2 overlapped gammas
K±→p0 e±n
(4.87±0.06)%
+1 accidental g and e misidentified as a p
K±→p0 m±n
(3.27±0.06)%
+1 accidental g and m misidentified as a p
K±→p0 e±n(g)
(2.66±0.2)·10−4
e misidentified as a p
K±→p0 m±n(g)
(2.4±0.85)·10−5
m misidentified as a p
Physical BG rejection:
For p±p0 we can relay on the cut in T*p< 80 MeV, MK and COG, cuts
For p±p0p0 we have released the T*p cut, but we can anyway reach the rejection needed (MK COG (missing g) and overlapping g cut (fused g))
Accidental BG rejection (p±p0, Ke3(p0 e±n), Km3(p0 m±n))
Clean beam, very good time, space, and mass resolutions.
Mauro Raggi
FlaviaNet Workshop

12. BG rejection performance
Thanks to the overlapping gamma cut, good MK and COG resolutions, the BG coming from p±p0p0 is under control.
K±→p±p0p0
K±→p±p0g
Selected region 220K events
Source
%IB
%DE
p±p0p0
~1·10-4
~0.61·10-2
Accidental
<0.5·10-4
~0.3·10-2
All physical BG can be explained in terms of the p±p0p0 events only. Very small contribution from accidentals neglected
Total BG is less than 1% with respect to the expected DE contribution even in the range 0<T*p<80 MeV
Mauro Raggi
FlaviaNet Workshop

13. The miss tagged g events: a self BG
■ DE ▼IB
The miss tagged gamma events behave like BG because they can induce fake shapes in the W distribution. In fact due to the slope of IB W distribution they tend to populate the region of high W simulating DE events.
The identification of radiative gamma has 2 steps: 1. Compatibility of charged and neutral vertices (2.5% mistagging)
2. Distance between best and second best neutral vertices>xx cm
The mistagging probability has been evaluated in MC as a function of the mistagging cut to be 1.2‰ at 400 cm
Mauro Raggi
FlaviaNet Workshop

14. Data MC comparison
The IB dominated part of W spectrum is well reproduced by MC
The radiated g energy (IB part of the spectrum) is well reproduced
W(Data)/W (IBMC)
Fitting region
IB dominated region
Data MC for Eg (W<0.5)
Egmin cut
Mauro Raggi
FlaviaNet Workshop

15. Fitting algorithm
To get the fractions of IB(a), DE(b), INT(g), from data we use an extended maximum likelihood approach:
The fit has been performed in 14 bins, between , with a minimum g energy of 5 GeV, using a data sample of 124K events.
To get the fractions of DE and INT the raw parameter are corrected for different acceptances
Mauro Raggi
FlaviaNet Workshop

16. Preliminary results Preliminary First measurement:
in the region 0<T*p<80 MeV and of the DE term with free INT term.
First evidence of INT term ≠0
The error is dominated by statistics and could be reduced using SS0 and 2004 data sets
Unfortunately parameters are highly correlated r=0.93%
Mauro Raggi
FlaviaNet Workshop

17. INT=0 fit: just for comparison
For comparison with previous experiments the fraction of DE, with INT=0 has been also measured. A likelihood fit using IB+DE MC only has been performed in the region 0<T*p<80 MeV and the result extrapolated to MeV using MC
The analysis of fit’s residuals shows a bad c2
Description in term of IB+DE only is unable to reproduce W data spectrum.
Mauro Raggi
FlaviaNet Workshop

18. Effect of the DE form factor
Could the negative INT be faked by DE ff? No
DE(no ff)-DE(ff)
Only INT>0 could be faked
DE could be underestimated
Could IB+DE(ff) be enough to improve MC agreement with data spectrum? No
Could the DE ff change the results for DE and INT? Yes (results moving on the correlation line)
Mauro Raggi
FlaviaNet Workshop

19. Conclusions
A new measurement of the of fractions of DE and INT in K±→±0g has been performed with 124K events
New kinematic region (T*p<55 MeV) explored
First evidence of a non zero INT term found
Preliminary
Mauro Raggi
FlaviaNet Workshop

20. Future prospects
The total number of events in the final result ( ) will be >500K
A fit with form factor in DE is foreseen
The CP violation will be investigated
A statistical error on CP violation ≈ 10-3 could be reached using the full data sample
The systematic error on CP violation will be probably <10-3
Mauro Raggi
FlaviaNet Workshop