1. Light Dark Matter Search in B-factory O. Tajima (KEK) for Belle collaboration

2. Contents Search for Invisible Decay of the Y(1S) Search for Invisible Decay of the Y(1S) Comments for the MeV Dark Matter search Comments for the MeV Dark Matter search New Results from Belle !! hep-ex/0611041

3. Contents Search for Invisible Decay of the Y(1S) Search for Invisible Decay of the Y(1S) Comments for the MeV Dark Matter search Comments for the MeV Dark Matter search New Results from Belle !! hep-ex/0611041

4.  : relic density h : Hubble constant v : 1/20 ~ 1/25  h 2 = 0.113  WMAP  h 2      SM  v  0.1 pb ・ c Relic density is denoted as follows Br( Y(1S)   ) ~ 6x10  3 (m  <4.73GeV/c 2 ~ m b )    SM  ~ 18 pb   SM       SM ,  S    f      bb     q q _ PRD 72, 103508 (2005) “Invisible quarkonium decays as a sensitive probe of dark matter” Presented by B. McElrath in BNM-I in KEK Sep, 2006 Motivation Extracted from PDG   q q _ Previous Upper Limit :  23x10 -3 (90% CL) by ARGUS (1986)

5. Y(3S) is the best for Y(1S) invisible Energy scan by CLEO No signal in detector  ISR ~0.02nb S/N <1/1000  ~4nb S/N~1/8  ~7nb No trigger issue Y(3S) runs : 2.9 fb -1 (Feb, 2006 : 4days) We show results, today

6. Signal finding method    S  recoil mass of     Y(3S)      Y(1S) invisible Y(1S) goes to invisible DM… We know momentum of     Energy of initial state (e + e - ) We can calculate mass of missing particle m Y(1S) (9.46 GeV/c 2 )

7. Trigger logic is important For trigger issue, we need two charged tracks Reach to outer most layer of CDC (pt ~250 MeV/c) Reach to middle layer of CDC (pt ~120 MeV/c) “Opening angle cut” is necessary to distinguish tracks   x y

8. Special Trigger for Y(3S)      Y(1S) invisible data MC Control sample  (3S)       (1S)  (1S)      Usual trigger condition (>135 o ) is too tight.  Loose cut condition is implemented ( ~850 Hz : twice rate of usual runs) To understand the trigger efficiency, “1-track” trigger was implemented with 1/500 pre-scaled rate. 2-track trigger & 1-track trigger 1-track trigger We can monitor the trigger efficiency as a function of opening angle. ?

9. Special Trigger for Y(3S)      Y(1S) invisible data MC Control sample  (3S)       (1S)  (1S)      Usual trigger condition (>135 o ) is too tight.  Loose cut condition is implemented ( ~850 Hz : twice rate of usual runs) To understand the trigger efficiency, “1-track” trigger was implemented with 1/500 pre-scaled rate. 2-track trigger & 1-track trigger 1-track trigger We can monitor the trigger efficiency as a function of opening angle. Trigger eff. = 89.8%  r  > 30 o pt full > 0.30 GeV/c pt short > 0.17 GeV/c other cuts for recon. 244 events predicted Br(Y(1S)  invisible)=6x10 -3

10. Y(3S)      Y(1S) invisible Background Two-photon 2 prong , ee,  …   p t is balanced   Boosted (  distribution)      ...     veto,  energy cut      S  recoil mass of     signal BG S/N: 1/30  1/8

11. Y(3S)      Y(1S) invisible Background     recoil mass of     Two-photon BG Y(1S)     , e  e  … (outside of acceptance) 244 events predicted Br(Y(1S)  invisible)=6x10 -3

12. Let’s see results !

13. Results data Fit BG Prediction Br(Y(1S)  invisble)=0.6% N signal = 38 ± 39  0 consistent Br( Y(1S)  invisible )  2.5x10  3 (90%C.L.) Prediction is disfavored

14. Prospects 90 %C.L Super-forward  -detector and Super-forward cal. To be reach 2x10 -4 ~500 fb -1 SM : Y(1S)  bar

15. Remaining issue … 101

16. Contents Search for Invisible Decay of the Y(1S) Search for Invisible Decay of the Y(1S) Comments for the MeV Dark Matter search Comments for the MeV Dark Matter search New Results from Belle !! hep-ex/0611041

17. Feasibility of MeV Dark Matter Search PRL 96, 141802 (2006) Presented in BNM-I WS (KEK Sep, 2006) Comments for

18. Signal Final states are Final states are  e  e   e  e   U invisible  U invisible In any case, m U,ee ~MeV << sqrt(s) In any case, m U,ee ~MeV << sqrt(s) E  cms = (s – M U 2 ) / ( 2 sqrt(s) ) E  cms = (s – M U 2 ) / ( 2 sqrt(s) ) ~ ½ sqrt(s) ~ 5 GeV (B-factory) ~ ½ sqrt(s) ~ 5 GeV (B-factory) For  e  e  mode For  e  e  mode p e cms ~ ¼ sqrt(s) ~ 2.5 GeV/c (B-factory) p e cms ~ ¼ sqrt(s) ~ 2.5 GeV/c (B-factory) We don’t need to change Beam Energy as Y invisible

19.  ee ee ee ee Suppressed by trigger “Bhabha veto” We cannot use this mode ~5 GeV,  Almost overlapped 2.5(e + )+2.5(e - ) ~ 5 GeV

20.  U invisible ~5 GeV,  No signal Due to Beam-BG suppressed “Bhabha veto” Veto with E sum of two-  rings Two-threshold logic is still under testing

21. Sensitivity Cannot reach due to Bhabha-veto with better Bhabha-veto (500 fb -1 ) with better Bhabha-veto (50 ab -1 ) PRL 96, 141802 (2006) Presented in BNM-I WS (KEK Sep, 2006)

22. Summary Y(1S) invisible search (GeV Dark Matter)  Y(3S) is the best energy for Belle  Br < 2.5 x 10 -3 ( 90% C.L. ) with 2.9 fb -1  Peaking BG suppression is necessary for further improvement (super forward detector etc.)  Sensitivity to be saturated Br ~0.2 x 10 -3 at 500 fb -1 MeV Dark Matter  We can search on usual Y(4S) running  Intelligent Bhabha veto is necessary