Search for heavy particles decaying into ψ(2S)π+π− producedin e+e− collisions with Initial State Radiation.

Tesi di Laurea Specialistica

Candidato: Stefano ZambitoRelatore: Prof. Fabrizio Bianchi1

1Dipartimento di Fisica Sperimentale & INFN

19 Luglio 2011

1 of 26

Analysis Preliminaries

� In the ”Costituent Quark Model”, only qqq baryons and qq mesons described...

BUT: QCD lattice calculations and further theoretical models predict the existenceof hadrons with complex substructures:

� Recently, many new and unexpected charmonium resonances found in thedata collected at the B-Factories...

2 of 26

Charmonium Spectrum

3 of 26

A Little Bit of Recent History

The Discovery of the Y(4260)

In 2005, studying the e+e− → γISRJ/ψπ+π− process, BABAR found a newstructure around 4.26 Gev/c2.

ISR analysis ⇒ Y(4260) has same quantum numbers of γISR : JPC = 1−−

BABAR results:

� Data sample of 232.15 fb−1

� M = 4259± 8 MeV /c2

� Γtot = 88± 23 MeV /c2

� Confirmed by Belle, CLEO-c (scan)[→ I = 0] and CLEO-III (ISR) in J/ψπ+π−,J/ψπ0π0, J/ψK +K− decay channels.

� No evidence found at present for Y(4260) → π+π−φ, π+π−pp.

� Unsuccesfully searched also in B-decays: B±,0 → K±,0J/ψπ+π−.

5 of 26

The Discovery of Y(4360), Y(4660)

Two more resonance found by BABAR around 4.36 Gev/c2 and Belle around 4.66Gev/c2 in the study of ISR ψ(2S)π+π−:

Belle results:

� M1 = 4361± 9± 9 MeV /c2

� Γtot1 = 74± 15± 10 MeV /c2

� M2 = 4664± 11± 5 MeV /c2

� Γtot1 = 48± 15± 3 MeV /c2

BABAR results:

� M = 4324± 24 MeV /c2

� Γtot = 172± 33 MeV /c2

6 of 26

Y -states Phenomenology

If Y s are conventional charmonia shoulddecay predominantly to open-charm!

BABAR → no evidence (384 fb−1)

Instead, Y (4660) recently seen in Λ+c Λ

−c :

Y -states are hybrid? moleular?tetraquark? baryonium?

More experimental and theoreticalwork is needed!

7 of 26

Analysis Motivations

Help making some light in this overpopulation of JPC = 1−−

charmonium-like states above the open charm threshold.

BABAR Detector Layout

SVT: 5 double side layers. 97% efficiency, 15µm z-hit resolution (inner layers, ⊥ tracks).SVT+DCH (momentum resolution): σ(pT )/pT = 0.13%× pT + 0.45%.DIRC: 144 quartz bars, provides K − π separation → 4.2(2.5)σ @ 2.4(4.0) GeV/c.

EMC: 6580 CsI (Tl) crystals. Energy resolution: σE/E = 2.3%× E−1/4 + 1.91%.

9 of 26

Analysis Strategy

We study the processes:

e+e− → γISR ψ(2S)π+π−

- `+`− (NEW!)

- J/ψπ+π−

- `+`−

We use 473.22 fb−1 of integrated luminosity collected bythe BABAR detector at the Υ (4S) resonance, and addittional 31.16 fb−1

at Υ (3S) and 15.90 fb−1 at Υ (2S): total of 520.87 fb−1.

The analysis strategy:

� ISR photon detection not necessary

� Require mass recoiling against final statecompatible with zero → γISR

� Missing transverse momentum must be low

10 of 26

Background Evaluation

ψ(2S)π+π−, ψ(2S)→ π+π−J/ψ decay mode.

4 pions in the final state:→ high probability of an interchange

→ expected large (dominant) combinatorial background.

Background estimated with doubly-charged combinations in the same data.Consistency check: we look also at ψ(2S) sidebands.

ψ(2S)π+π−, ψ(2S)→ `+`− decay mode.

We expect an important source of background from radiative Bhabhas, likee+e− → e+e−γ where the photon converts in the material.

Background estimated using ψ(2S) sidebands.11 of 26

ψ(2S) → π+π−J/ψ Issues: Multiple Candidates

Multiple candidates in the same event are possible with 4 pions in the final state:→ the best candidate is selected based on its ψ(2S) mass value.

MC check: m(π+π−ψ(2S)) reconstructed from a sample generated with flat massdistribution (for efficiencies and resolution studies).

The fake peak around 4.28 GeV/c2 (yellow histogram): wrong candidates wheretwo pions are swapped.

It disappears:� requiring truth-matching (blue histogram);

� selecting only the best candidate for each event (green).

12 of 26

ψ(2S) → `+`− Issues: Electron Pair Contamination

Let’s consider an e+e− → γ`+`− event where:

� the mass of the `+`− pair is compatible with ψ(2S);

� the photon converts into an electron pair (e.g. byinteracting with detector material);

� this new pair is misidentified as pions.

Get a fakeψ(2S)π+π−

candidate!!!

To remove this background we compute the π+π− invariant mass assigning to thepions the electron mass: we discard events below 100 MeV/c2 .

13 of 26

ψ(2S) → π+π−J/ψ: Are There Residual Non-ISR Events?

Some relevant kinematic variables...

Good agreement between selected data and ISR Signal MC: OK!14 of 26

ψ(2S) → `+`−: Are There Residual Non-ISR Events?

Some relevant kinematic variables...

KO: non-ISR contamination still present.15 of 26

ψ(2S) → `+`− Non-ISR Contamination Removal

ψ(2S) → `+`− channel features:

� simple decay topology, only 4 charged tracks;

� limited number of kinematic variables with effective rejection power againstbackgrounds.

Tight cuts already used in the selection: what still can we do?

We addittionally require that:

� OR the ISR photon is detected

� OR the ”reconstructed” ~pγ ISRpoints outside the calorimeteracceptance region.

Useful improvement, now thedistribution looks better!.

16 of 26

Efficiency and Resolution Studies

Efficiency Studies Results:

Two indipendent procedures, consistent results:

� We fit samples of signal MC events with a Voigtian PDF (gaussian→resolution,BW→resonance), to count the fraction of generated events surviving selection.

� Toy MC study for the mass-dependent efficiency.

ψ(2S)→ π+π−J/ψY(4260) Y(4360) Y(4660)∼ 5.5% ∼5.9% ∼7.3%

ψ(2S)→ `+`−

Y(4260) Y(4360) Y(4660)∼ 8.0% ∼9.2% ∼11.5%

Resolution Studies Results:

� the mass resolution remains between 3-7 MeV/c2 in [4,6] GeV/c2.

� the mass shift introduced by the reconstruction procedure is very small,∼ 0.5 MeV/c2, and will be accounted as a systematic error.

17 of 26

Results of Unblinding

Invariant Mass Spectrum, ψ(2S) → π+π−J/ψ

� Surviving background level inthe signal region is negligible.

� ψ(2S)π+π− invariant massdistribution shows clearenhancements at around 4360MeV/c2 and 4660 MeV/c2.

19 of 26

Invariant Mass Spectrum, ψ(2S) → `+`−

� Surviving background level inthe signal region is high.

� ψ(2S)→ e+e− events areexcluded from the final result.

� In ψ(2S)→ µ+µ−, there is aclear enhancement at around4360 MeV/c2.

20 of 26

π+π− Invariant Mass Distribution

Observed distributions differ from the phase space expectation and tend to peakat high masses, underlying the presence of a dynamical effect.

21 of 26

Cross-section Measurement

e+e− → π+π−ψ(2S) exclusive process in ISR events, from threshold to 6 GeV

For each π+π−ψ(2S) mass bin, the cross section is calculated according to:

σi =nobsi − nbkg

i

εiLiB(ψ(2S)→ final state), Li → i-th bin integrated luminosity

εi → i-th bin total efficiency

22 of 26

Fit to ψ(2S)π+π− Mass Spectrum

We perform an Unbinned Maximum Likelihood fit to the π+π−ψ(2S) invariantmass spectrum, only for the ψ(2S)→ π+π−J/ψ channel. We hypotize two

distinct resonances; an interference term is taken into account.

The parameters to determine are: masses (mj), widths (Γj), relative phase (φ),normalization factors (Nj , i.e. number of observed events under each resonance).

23 of 26

Conclusions

Summary of Results

� We measured the energy-dependent cross-section for the exclusive processe+e− → π+π−ψ(2S) using ISR events, for ψ(2S)→ π+π−J/ψ and forψ(2S)→ µ+µ−.

� We confirm the structures observed by BABAR at around 4.33 GeV/c2 andBelle at around 4.66 GeV/c2.

� For ψ(2S)→ π+π−J/ψ, we fit to the π+π−ψ(2S) mass spectrum under thehypostesis of two distinct resonances, and we measured their parameters.

Plans for nearby future:

Systematic uncertainties studies are close to be finalized: we will have a totalsystematic error of ∼ 10− 14%, consistent with previous published results.

The result will be presented -abstract already submitted- for the first time at:XCVII Congresso Nazionale, Societa Italiana di Fisica

L’Aquila, Sep 26-30 2011 (invited).

The final publication target is Physical Review D.

25 of 26

Thanks for Your Attention!

Backup Slides

ISR Events Selection Requirements

� Small difference between C.M. momentum of the hadronic π+π−ψ(2S) systemand the value expected for ISR events: (s −m2)/(2

√s) [∆p∗, left plot].

� Small transverse component of the missing momentum of entire event in C.M.frame [p∗

t , center plot].

� No additional charged tracks in each event [nTRK , right plot].

Selection criteria are optimized to maximize the significancy: NS/√

NS + NB .28 of 26

ψ(2S) → `+`− Issues: Electron Pair Contamination (II)

The mass of the pair is displaced asa consequence of the vertexrequirement during the fit on thetracks: in this case we obtainm(e+e−) ' 22 MeV/c2.

Zoom on the SVT:

The 2 tracks of the electron pair areclosest near the beam pipe; thisevent is probably due to γ conversionin the beam pipe material.

29 of 26

Fit to ψ(2S)π+π− Mass Spectrum (II)

We perform an Unbinned Maximum Likelihood fit to the π+π−ψ(2S) invariantmass spectrum, using the following PDF :

PDF (m) = ε · fISRm2·

∣∣∣∣∣BW1

√N1 · PS(m)

Γ1 · PS(m1)+ BW2

√N2 · PS(m)

Γ2 · PS(m2)× e iφ

∣∣∣∣∣2

,

where:

� BWj = mj Γj/(m2 −m2j + i mj Γj) is the Breit-Wigner shape;

� energy dependence of the e+e− annihilation rate: ∝ 1/s = 1/m2;

� W (s, 1−m2/s) is the ISR photon emission probability density;

� fISR = W (s, 1−m2/s)× 2m/s accounts for mass dependence of ISR emission;

� PS(mj) is the phase space factor;

� ε is the efficiency, which increases with the mass.

The parameters to determine are: masses (mj), widths (Γj), relative phase (φ),normalization factors (Nj , i.e. number of observed events under each resonance).

30 of 26