1. 0nbb decay to the excited state 0+ of 130Xe
Comparison of the GE and SC analyses
S. Di Domizio, December 2010

2. Part 1: comparison of the methods
In the following slides I will evaluate the efficiencies using the SC cuts with the GE and SC algorithms

3. Efficiency – scenario1 – SC – 536

4. Efficiency – scenario1 – SC – 734

5. Efficiency – scenario1 – SC – 1257

6. Efficiency – scenario2 – SC - 1257

7. Efficiency – scenario2 – SC – 1270

8. Efficiency – scenario3 – SC – 536

9. Efficiency – scenario3 – SC – 1991

10. GE – scenario
Using GE algorithms and SC cuts

11. GE – scenario
Using GE algorithms and SC cuts

12. GE – scenario
Using GE algorithms and SC cuts

13. Part 2: comparison of the results
In the following slides I will summarize the differences in the two approaches and will extract the half life limits

14. Comparison statistics efficiency Geometric onlySC
N·t = 9.11 x 1025 y
values reported in the note
N·t = 8.74 x 1025 y
N·t = 9.50 x 1025 y
N·t = 8.96 x 1025 y
Forgot to include the three “dead” channels 2, 3 and 50
“My” evaluation with “SC” method
efficiency
Geometric only
total (with psa, noise, etc.)‏ GE SC GE SC
scenario1
0.60%
0.80%
scenario1
0.48%
0.64%
scenario2
2.29%
2.58%
scenario2
1.93%
2.18%
scenario3
1.41%
1.75%
scenario3
1.19%
1.48%

15. Result (GE)‏ Posterior pdf for G G < 6.74 x 10-25 y-1 90%CL
T1/2 > 1.03 x 1024 y 90%CL
Posterior pdf for G

16. Result (SC)‏ Posterior pdf for G G < 5.98 x 10-25 y-1 90%CL
T1/2 > 1.16 x 1024 y 90%CL
Posterior pdf for G

17. Part 3: the approach proposed by Frank
In the following slides I will show the method and the results I obtained by treating the difference between GE and SC analysis as a systematic error

18. Treating the differences as syst errors
Use the approach discussed in Adam's internal note
scenario
efficiency 1
(0.56+/-0.08)%
Statistics: N·t = (9.23 +/- 0.27) x 1025 y 2
(2.06+/-0.13)% 3
(1.34+/-0.15)%

19. Result (combined)‏ Posterior pdf for G G < 6.39 x 10-25 y-1 90%CL
T1/2 > 1.09 x 1024 y 90%CL
Posterior pdf for G

20. GE: T1/2 > 1.0 x 1024 y @90%CL SC: T1/2 > 1.2 x 1024 y @90%CL
Summary
GE: T1/2 > 1.0 x 1024
SC: T1/2 > 1.2 x 1024
GE+SC: T1/2 > 1.1 x 1024

21. Method comparison
Consider the limit case of an experiment with two crystals where one has 100% dead time and the other has 0 dead time.
Since no coincidences can be recorded in these conditions, the number of signal and background counts will be zero.
The SC approach would give a finite value for both the efficiency and the accumulated statistics, thus resulting in a non trivial limit for the half life of the process.
The GE approach would give a finite value for the statistics and a null value for the efficiency, therefore nothing can be said about the half life of the process.

22. Treating the differences as syst errors