First analysis of the fully simulated channel

First analysis of the fully simulated channel

Marco ZanettiCMS Padova

Muon Week – Bologna 10/4/2003

WWH 0

Past analyses (e.g. M.Dittmar) have demostrated that this could be a very fast channel (with l , 5fb-1 ~ 1 month @ LL) for the discovery of the SM Higgs in the intemediate mass region 130< MH <2 MZ, filling the gap of

Because of the 2 impossible to distinguish a peak in the invariant mass distribution => not easy counting channel: => the precise knowledge of all possible backgrounds processes is fundamental;

Marco Zanetti CMS Padova

llllWWH 0

b

s

n

nS

)4( 0 lZZHBR


Signal

x-sec (pb)events

in the sample

MH=120 0.036 9999

MH=140 0.114 9997

MH=160 0.182 9998

MH=200 0.104 9999

WWH 0Reference sample

WWnr is the main bkgr, with the same final states but reducible (next slides)


sBackground

Process Ptmin x-sec (pb) N events L

5,5 77 17599 0.2 fb-1

(Pthard>20) 7,7 (1.4)·107 48471 3.5 nb-1

3,3 106 49489 49 nb-1

5,5 621 17991 29 pb-1

5,5 29, 12 36990 O( fb-1 )

WWqq

XZ */

Xtt

XZZZW ,

Xbb

Not yet considered:

• Single top ( )

• WWgg

XWtbpp


Signal vs WWnr 1:

small |muons

Higgs produced mainly via g-g fusion (in 130< MH <2MZ 80% of the times)

WWnr produced by

z-boost

qq

21

21

4 xx

xxz

z~1

z~10

WWnr evs more boosted along the beam

More isotropic signal muons spatial distribution


Signal vs WWnr 1:

Could depend on PDF used, but no significant differences between

GRV94 (old) and CTEQ5L (DC04)

Di- system polar angle

8000 evsbarrel endcap


Signal vs WWnr 2:

Signal:

Higgs scalar, W vectors => to have JTot=0 spin wave function must be:

nearby

LL

WW

TT

WW

TT

WWWW zzzzzz

z

)0()0()()()()(

3

1)0,0(

TL coupling not allowed

Case TT: each muon has the same SZ of the mother and must be produced with definite elicity => max prob to have emitted in the same direction

Case LL: not easy to demostrate that => again max prob to have emitted in the same direction

))((

WWnr:

Unpolarized initial state => TL coupling is allowed and @ MWW~160 half of the production rate

Case TL: no correlation in emission


Signal vs WWnr 2:

Big difference in shape

Indipendece from the PDF

Implications also for MEt

opening angle

8000 evs


Production chain

Pythia 6.158 (GRV94LO)

CMSIM 125

ORCA 6_2_3 (digi and reco)

Analysis

based on L3 reconstruction

Pile up


HLT SelectionAlgorithmic Efficiency (Single+Di- standard sel. in |)

MH=120 MH=160 MH=200

L1 98.2% 99.2% 99.3%

L1+L2 95.0% 97.6% 98.1%

L1+L2+L3 91.1% 93.7% 96.3%

Analisys starting point: HLT Di- standard sel.

Eff (%) Rate (Hz) Events

Signal 80.6 2.3 ·10-4 8055

WWnr 72.8 1.1 ·10-3 12812

ZZ+WZ 77.6 2.0 ·10-3 28704

49.4 4.4 ·10-2 8888

Z/ 5.2 2.29 2574

28.1 1.81 13620

tt

bb

Signal overwhelmed (4 ord. mag.)


Offline selections 1) Pt cut: Pt1,2 > 25,15; isolation @ 97% for both; opposite charge;

High eff. for the signal ~70%

Huge Rej. factor ( =Rbefore/Rafter) ~500 for sample; only 28 events survive

most of the * contribution suppressed

good Rej. (5) for thanks to isolation

bb

tt

2) Missing Et cut: MEt > 40 GeV;

Only 1 event survives (low pt s) with a integral rate of the same order of signal events (1.3·10-4 Hz)

Big Rej. (~21) for Z/

WWnr more rejected (2.1) than signal (1.3) as foreseen (spin correlation)

bb


Offline selections

3) Invariant Mass cut: M < 70 GeV;

To reject events from Z peak;

I will select events with nearby muons:

M = {2|p1||p2|(1-cos)}1/2 => No lower bound

Rej. = 22.5 for Z/ sample; 3 events surving but with a integral rate greater than the signal (2.7·10-3 vs 1.3·1 0-4 Hz)

Big rejection also for WZ+ZZ

no more eventsbb


Offline selections 4) Jet veto: rejected events with one jet with Et>20 in | ;

VETO REGION

Specific for top events (tt & Wtb)

Huge Rej. (~85) for ; only 5 events survive giving a smaller rate contribution than the signal (2.5·10-5 vs 4.7·10-5 Hz)

Much less effective for sample (if applied)

no more Z/ events

not so small rejection also for the signal (Rejec=2.7)

tt

bb


Offline selections 5) Di- polar angle cut: => Di- System in the barrel8.0cos Di

Signal Rejec = 1.9 WWnr Reject = 2.4;

no more eventstt

6) Nearby s: < 0.8 &&< 1.5

Final plot

Eff.(%) Rate Rej

Signal 4.7 1.4·10-5 1.8 4.6

WWnr 0.5 7.9·10-6 3.3 10.5

WW+ZZ 0.003 0.7·10-6 3.1 30

N/1

Final numbers


Statistical Results

MH=160 Nev @ 5fb-1 Nev @ 20fb-1

Signal 35 138

Background 22 88

7.4 14.7

Results for other masses:

Eff (%) @ 5fb-1 @ 20fb-1 L5(fb-1)

MH=120 1.9 0.5 1.1 390

MH=140 3.2 3.0 6.0 14

MH=200 2.7 2.5 4.9 21

MH=120: not viable; other H production processes like Vector Bosons Fusion can be used (CMS Note 2002/26: pp->jjH, H->WW->ll)

MH=200: different kinematic (e.g. no so effective spin correlation) different selections should be used

sample depleted at the very beginning of the offline selection

before beeing cut away, events give a integral rate contribution smaller than the signal and following cuts more effective to them than to signal ones.

Huge rejection for Z events by cuts 5 and 6 ( inizial state and muons back-to-back in transvers plane)

in the present Z/ sample not included qg->q(Z/) and => foreseen good rejection also by Jet veto


Systematics

Statistics of the background samples used insufficient, anyway:

tt

bb

qq

)/*( Zgqq

No effective contribution after offline selections

foreseen for these background processes


DC04 Production

Xbb

XZ */

Sample Pt1 Pt2 M N events Rfin (Hz)

20 10 10 106 3·10-6

20 10 10 3·106 10-6

7 7 0 600K 2·10-7Xtt

95 % CL upper limit of Rate contribution after offline sel.if no events survive

WWnr sample: 200K events, Pt>7,7 => statistical error after offline selection %3/1 FinN

WZ ZZ samples: 200K +200Kevents, Pt>7,7 => statistical error after offline selection %9/1 FinN

(for tt forseen contribution O(10-6 Hz))


Missing bkgr processes

WWgg

NLO process, x-sec 1 ord. smaller than WWnr; because of the gg inizial state should be taken into account.

Not yet implemented in Pythia. Not yet inculded in any analysis

XWtbpp x-sec ~ 0.5 and known to give a comparable contribution.

particularly dangerous, to be generated with CompHEP

tt

Wtbgg

(Stefano Marcellini)


Conclusions

Although with not sufficient bkgr statistics, it has been shown the present channal is very promising for a fast ( 5fb-1 ) discovery of SM H in the intermediate mass region

waiting for DC04 high statistics samples to confirm these results

First analysis of the fully simulated channel

Documents