Strange tagging at future colliders

Matthias Schlaffer University of Chicago

PRD 101 (2020) 11, 115005 (1811.09636) with: Duarte-Campderros, Perez, and Soffer Higgs properties

35.9-137 fb-1 (13 TeV) V

> learned a lot about the Higgs ν t m 1 V CMS Z κ W

mH = 125.38 GeV

or −1

> Higgs is main source of EWSB F 10

ν p-value = 44% m F κ b > couples to third generation as 10−2 τ

Vector expected −3 10 rd µ 3 generation − SM Why bother? 10 4

1.5

> No understanding of flavor 1

Ratio to SM 0.5 puzzle 10−1 1 10 102 mass (GeV) > Yukawa modifications could [CMS: 2009.04363] affect only first and second generation [E.g. Giudice & Lebedev ’08, Bauer et.al 15, Ghosh et.al ’15, Altmannshofer et.al ’15, Egana-Ugrinovic et.al ’19]

Matthias Schlaffer 1 Exclusive decay h φγ [Bodwin et.al ’13, Kagan et.al ’14] → s¯ o φ s¯ o φ s s + h h

γ γ

> Clean decay: BR(φ(ss¯) K+(us¯) + K (¯us)) 50% → − ≈ > BUT: BR(h φγ) 2 10 6 [König et.al ’15] → ≈ × − > compare BR(h ss¯) 2 10 4 → ≈ × − > only weak limit at future () colliders [Kagan et.al ’14] estimate: µ (107) @HL-LHC ss . O > current limit: BR(h φγ) < 4.8 10 4 [ATLAS ’17] → × − Ideas to use differential distributions [see e.g. Bishara et.al ’16, Soreq et.al ’16, Yu ’16, Carpenter et.al ’16]

Matthias Schlaffer 2 Our brute force method Alternative ansatz: > FCC-ee will produce 106 Higgses via

e− Z Z∗

e+ h > (200) of which decay into strange O > tag strange jets > Done before in Z ss¯ – Measurement of→ the strange forward backward around the Z0 peak [DELPHI Collaboration, Eur.Phys.J. C14 (2000)] – Light quark fragmentation in polarized Z0 decays [SLD Collaboration, Nucl.Phys.Proc.Suppl. 96 (2001)]

Matthias Schlaffer 3 Part I: > Clean sample with hadronic Higgses > We know which jets originate from the Higgs decay > Generate and shower with PYTHIA and Herwig > No detector simulation

Setup and assumptions

kinematic separation h jj data → s-tagger limit ⇒ cut&count, BDT,... other bkg ⇒

Matthias Schlaffer 4 Setup and assumptions

kinematic separation h jj data → s-tagger limit ⇒ cut&count, BDT,... other bkg ⇒

Part I: > Clean sample with hadronic Higgses > We know which jets originate from the Higgs decay > Generate and shower with PYTHIA and Herwig > No detector simulation

Matthias Schlaffer 4 Neutral : > Decay length 80 cm ∼ > Needs to decay to π± within 5 mm < R < 1 m > reco efficiency 80%

Kaon reconstruction Charged kaons: > stable on detector scales > tracking efficiency 95% > Particle ID .

π± K±

2 σ bench marks e.g.: > no ID

> K = 95% π = 12%

some observable

Matthias Schlaffer 5 Neutral kaons: > Decay length 80 cm ∼ > Needs to decay to π± within 5 mm < R < 1 m > reco efficiency 80%

EXPERIMENT ENVIRONMENTAND DETECTOR DESIGNS the DCH design is the drift chamber of the KLOE experiment [490], which was more recently developed as the MEG2 [491] drift chamber. The DCH is a unique-volume, high-granularity, all-stereo, low-mass, cylindrical, short-drift, wire chamber, co-axial with the 2 T solenoid field. It extends from an inner radius Rin = 0.35 m to an outer radius Rout = 2 m, for a length L = 4 m and consists of 112 co-axial layers, at alternating-sign stereo angles, arranged in 24 identical azimuthal sectors. The approximately-square cell size varies between 12.0 and 14.5 mm for a total of 56 448 drift cells. The challenges potentially arising from a large number of wires are addressed by the peculiar design of the wiring, which was successfully employed for the recent construction of the MEG2 drift chamber [492]. The chamber is operated with a very light gas mixture, 90% He – 10% iC4H10 (isobutane), corresponding to a maximum drift of 400 ns. The number ∼ 1 of ionisation clusters generated by a minimum ionising particle (m.i.p.) is about 12.5 cm− , allowing cluster counting/timing techniques to be employed to improve both spatial resolution (σx < 100 µm) and particle identification (σ(dN /dx)/(dN /dx) 2%). The angular coverage extends down to 13◦, cl cl ≈ ∼ and could be further extended with additional silicon disks between the DCH and the calorimeter end caps. A drift distance resolution of 100 µm hasKaon been reconstruction obtained in a MEG2 drift chamber prototype [493] (7 mm cell size), with very similar electrostatic configuration and gas mixture. A better resolution is expected for the DCH, as a result of the longerCharged drift distances kaons: and the employment of cluster timing > stable on detector scales techniques. Analytical calculations for the expected momentum, transverse momentum and angular > tracking efficiency 95% resolutions, conservatively assuming a 100 µm point> Particle resolution, ID are plotted in the left panel of Fig. 7.10. The expected particle identification performance. is presented in the right panel of Fig. 7.10. Results are IDEA Drift chamber

Par7cle"Separa7on"(dE/dx"vs"dN/dx)"π± K± Momentum(and(Angular(Resolu9ons((theta(=(90)( 10" 1.E$02' µ-π π-Κ Κ-p 9" 2 σ bench marks e.g.: 8" Δp /p t t 7" > no ID #"of"sigma" 1.E$03' 6" > K = 95% 5" π = 12% 4" 1.E$04' 3" Δϑ [rad] 2" some observable 1" Δϕ [rad] 1.E$05' 0" 0.1" 1" 10" 100" 0.1' 1' 10' 100' [FCC-ee CDR] Transverse(Momentum([GeV/c]( Momentum"[GeV/c]" Matthias Schlaffer 5

Figure 7.10: IDEA drift chamber performance. Left: momentum and angular resolutions for θ = 90◦ as a function of momentum. Right: particle type separation in units of standard deviations as a function of momentum, with cluster counting (solid curves) and with dE/dx (dashed curves). based on cluster counting, where it is assumed that the relative resolution on the measurement of the number of primary ionisation clusters (Ncl) equals 1/√Ncl. For the whole range of momenta, particle separation with cluster counting outperforms the dE/dx technique by more than a factor of two. The expected / separation is better than three standard deviations for all momenta except in a narrow range from 850 MeV to slightly above 1 GeV. A layer of silicon micro-strip detectors surrounds the outside of the drift chamber providing an additional accurate space point as well as precisely defining the tracker acceptance.

7.4.3 IDEA Tracking System Performance Simulations were performed to obtain a first estimate of the performance of the IDEA tracking system. In this study, a seven-layer cylindrical vertex detector, and a two-layer silicon wrapper, both with a rφ pitch of 20 µm, were placed inside and around the cylindrical drift chamber, respectively. Details

205 PREPRINT submitted to Eur. Phys. J. ST Kaon reconstruction Charged kaons: > stable on detector scales > tracking efficiency 95% > Particle ID

Neutral kaons: > Decay length 80 cm ∼ > Needs to decay to π± within 5 mm < R < 1 m > reco efficiency 80%

Matthias Schlaffer 5 Jet-Flavor

> define flavor of light jet > strange quarks fragment more likely into hard kaons

h ss¯ h gg → → 101 101 2 2 2 2 Q = mh Q = mh ± ± K ± K ± 0 0 10 Herwig 10 Herwig Pythia 8 Pythia 8 ) ) z z

1 ( 1 (

s 10 10 g F F

10 2 10 2

10 3 10 3 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 z z

Matthias Schlaffer 6 K− K+

Jet-Flavor

> define flavor of light jet > strange quarks fragment more likely into hard kaons

π− π+ K+

 P ˆ  1 H = K∓ H j ~pH jRH ± J = ∈ · R = 1 H = K , min. J F P ˆ H S F H j ~pH j ± ∈ ·  0 else

Matthias Schlaffer 7 π− π+

Jet-Flavor

> define flavor of light jet > strange quarks fragment more likely into hard kaons

K− K+ K+

 P ˆ  1 H = K∓ H j ~pH jRH ± J = ∈ · R = 1 H = K , min. J F P ˆ H S F H j ~pH j ± ∈ ·  0 else

Matthias Schlaffer 7 Jet-Flavor > define flavor of light jet > strange quarks fragment more likely into hard kaons

> J : R ± = 1, R = 1 minimizing J , else 0 s K ∓ Ks ± s > counts collinear hard strange content > not safe against collinear emission

h uu¯ → h dd¯ → 101 h ss¯ → h gg → Herwig 0.02 Pythia 8 /

100 fraction of events

1 10−

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 Js Matthias Schlaffer 8 Reject heavy flavor

> Minimalistic approach: Just cut on largest impact parameter

> Require plab > 5 GeV ∆d 10 µm ⇒ 0 . > Smear truth values > Include 5 µm uncertainty on IP

h uu¯ → h dd¯ 0.08 → h cc¯ → h ss¯ m → µ h b¯b 5

. →

0 0.06 h gg

/ → W had. → Herwig Pythia 8 0.04

fraction of0 events .02

0.00 0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 d0 [µm]

Matthias Schlaffer 9 Signal vs. background efficiencies

1.0 Pythia Herwig non-jj bkg 0.8 h jj bkg → combined bkg h cc¯ → 0.6 h b¯b → h gg

bkg →  W had. → 0.4

0.2

0.0 0.0 0.2 0.4 0.6 0.8 1.0 ss

Matthias Schlaffer 10 Part II: Realistic Collider

Existing studies for h bb, cc, gg: → > Cut&Count: mh = 120 GeV [Ono et.al ’12] > BDT [Talk by Yu Bai @ CEPC meeting]

Assumptions: > hνν final state (don’t consider h`` or hqq) > Non-h jj flavor composition as in C&C study: → 60% ν`qq 20% ννqq 10% qq flavor W bb uu dd cc ss gg relative abundance [%] 66 6 7 6 7 6 0 >  from h qq qq → >  from ee WW W → Matthias Schlaffer 11 Results

kinematic separation h jj data → s-tagger limit ⇒ cut&count, BDT,... other bkg y ⇒

x-Axis: N = σ BR  jj L h jj jj P y-Axis: Nnon-jj = i non-jj i L ∈

For each point (x,y) find best cut values to minimize upper limit

Matthias Schlaffer 12 Results

Upper limit on µ PYTHIA

108 200 Cut & Count BDT 7 1 100 10 = 250 fb− L 1 = 5 ab− L 1 = 20 ab− 50 106 L µ

jj 20 − 105 non

N 10 104 95% CL on 5

3 10 2

102 1 102 103 104 105 106 107 Njj

Matthias Schlaffer 13 Conclusion

> s-tagger in the context of h ss¯ → > proof-of-concept, can be improved > validation possible with large data sets of WW and Z 1 > with 10 ab− (FCC-ee): µs . 20 7 > compare with HL-LHC: µs . 10 > applicable to other searches with s-jets (up to some modifications)

Thank You

Matthias Schlaffer 14 BACKUP

Matthias Schlaffer 15 1/6 vis.

K± inv. 1/3

Strange hadronization In which kaons can a s quark hadronize?

0 KS

K± 0 KL

Matthias Schlaffer 16 0 KS

K± 0 KL

Strange hadronization In which kaons can a s quark hadronize?

1/6 vis.

K± inv. 1/3

Matthias Schlaffer 16 Impact parameter resolution r  2 ∆d = ∆2 + (5 µm)2 + 10 0 IP p sin3/2 θ

20

15 ]

m p=2GeV

[ μ 10

0 p=5GeV d

Δ p=10GeV 5 p=20GeV

0 0.0 0.5 1.0 1.5 θ

Matthias Schlaffer 17