History of data scouting in CMS

Chapter III: Supersymmetry and searches at the LHC

4.3 History of data scouting in CMS

The idea of data scouting was conceived in Run I of CMS, and the technique was used to perform searches for exotic resonances decaying to dijets. Dijet resonance searches tend to be severely constrained by trigger requirements: events with two back-to-back jets are ubiquitous at hadron colliders and a trigger that recorded all of them would have prohibitively high rate.

In CMS, dijet searches traditionally use H_T triggers, which were discussed in Sec- tion 4.1. TheH_T trigger threshold rises with increasing LHC energy and luminos- ity, and this implies that the lowest resonance mass the search can probe is pushed higher and higher over time.

This challenge is illustrated by the green and light blue curves in Figure 4.2, which indicate exclusion limits from the LHC experiments on production of a hypothetical Z’ resonance decaying to quarks. The LHC sets state-of-the-art limits at high Z’

mass, but at lower masses the limits from standard offline dijet resonance searches are nonexistent (for CMS) or worse than those from the Tevatron (for ATLAS, who use prescaled trigger paths to extend their dijet search to lower masses). The ATLAS limit (light blue curve) provides an explicit demonstration of the worsening of the limits due to trigger restrictions. Their limits at lower masses rely on data from triggers with increasingly high prescales, which are needed in order to keep the rate low. The trigger with the lowest threshold (selecting jets with p_T between 59 and 99 GeV) has a prescale factor of 460000 [99].

Data scouting provides dijet searches with relief from the rising H_T trigger rates.

This was first demonstrated in 2011, when a data scouting trigger path was deployed in CMS for a few hours of data taking at 7 TeV. The trigger had a L1 requirement of H_T > 100 GeV, and it selected all events having eitherH_T >350 GeV or dijet mass greater than 400 GeV at the HLT level. The only data recorded for each event was the collection of jets reconstructed with the HLT PF algorithm. These few hours’

worth of data, collected at the very end of the 2011 run period and corresponding to 0.13 fb⁻¹of integrated luminosity, were sufficient to improve existing limits on

56

299 coupling g

_B

of a hypothetical leptophobic resonance Z

⁰_B

→

300 q q ¯ as a function of its mass. The Z

⁰_B

production cross

301 section scales with the square of the coupling g

_B

. Figure 4

302 shows the upper limits obtained with the data scouting

303 technique in the mass region from 500 to 1200 GeV,

304 extending the coverage of previous CMS searches to below

305

1200 GeV. Previous exclusions obtained with similar

306 searches at various collider energies are also shown. As

307 a result of the large data set collected by the data scouting

308 stream, the bound on g

_B

is improved by up to a factor of 3

309 for resonance masses between 500 and 800 GeV, compared

310 to previous searches. This corresponds to an order-of-

311 magnitude improvement in the cross section limit.

312 In summary, a search for narrow resonances decaying

313 into two jets was performed using data from proton-proton

314 collisions recorded by the CMS experiment at

315

p ffiffiffi s

¼ 8 TeV, corresponding to an integrated luminosity

316 of 18.8 fb

⁻¹

. The novel technique of data scouting was

317 used; by reducing the information stored per event, multijet

318 events could be collected in sufficiently large samples that a

319 sensitive search for dijet resonances down to masses as low

320 as 500 GeV was possible. No evidence for a narrow

321 resonance is found. Model-independent upper limits on

322 production cross sections are derived for quark-quark,

323 quark-gluon, and gluon-gluon resonances. Based on these

324 results, new limits are set on an extensive selection of

325

narrow s-channel resonances over mass ranges not

326 excluded by previous searches at hadron colliders.

327 Bounds on the coupling of a hypothetical leptophobic

328 resonance decaying to quark-antiquark are also provided,

329 as a function of the resonance mass. The limits obtained

330 are the most stringent to date in the dijet final state for

331 narrow resonance masses between about 500 and 800 GeV.

332 We congratulate our colleagues in the CERN accelerator

333 departments for the excellent performance of the LHC and

thank the technical and administrative staffs at CERN and 334

at other CMS institutes for their contributions to the success

335

of the CMS effort. In addition, we gratefully acknowledge 336

the computing centers and personnel of the Worldwide 337

LHC Computing Grid for delivering so effectively the 338

computing infrastructure essential to our analyses. Finally, 339

we acknowledge the enduring support for the construction 340

and operation of the LHC and the CMS detector provided 341

by the following funding agencies: BMWFW and FWF 342

(Austria); FNRS and FWO (Belgium); CNPq, CAPES, 343

FAPERJ, and FAPESP (Brazil); MES (Bulgaria); CERN; 344

CAS, MoST, and NSFC (China); COLCIENCIAS

345

(Colombia); MSES and CSF (Croatia); RPF (Cyprus); 346

MoER, ERC IUT and ERDF (Estonia); Academy of 347

Finland, MEC, and HIP (Finland); CEA and CNRS/ 348

IN2P3 (France); BMBF, DFG, and HGF (Germany); 349

GSRT (Greece); OTKA and NIH (Hungary); DAE and

350

DST (India); IPM (Iran); SFI (Ireland); INFN (Italy); MSIP

351

and NRF (Republic of Korea); LAS (Lithuania); MOE and

352

UM (Malaysia); BUAP, CINVESTAV, CONACYT, LNS,

353

SEP, and UASLP-FAI (Mexico); MBIE (New Zealand);

354

PAEC (Pakistan); MSHE and NSC (Poland); FCT

355

(Portugal); JINR (Dubna); MON, RosAtom, RAS and

356

RFBR (Russia); MESTD (Serbia); SEIDI and CPAN

357

(Spain); Swiss Funding Agencies (Switzerland); MST

358

(Taipei); ThEPCenter, IPST, STAR and NSTDA

359

(Thailand); TUBITAK and TAEK (Turkey); NASU and 360

SFFR (Ukraine); STFC (United Kingdom); DOE and 361

NSF (USA). 362

363 364 [1] C. Albajar et al. (UA1 Collaboration), Two-jet mass

365

distributions at the CERN proton-antiproton collider, Phys. 366 Lett. B 209, 127 (1988). 367

[2] J. Alitti et al. (UA2 Collaboration), A measurement of two 368 jet decays of the W and Z bosons at the CERN pp ¯ collider,

369

Z. Phys. C 49, 17 (1991). 370

[3] J. Alitti et al. (UA2 Collaboration), A search for new 371 intermediate vector bosons and excited quarks decaying 372 to two-jets at the CERN p p ¯ collider, Nucl. Phys. B400, 3 373 (1993). 374

[4] F. Abe et al. (CDF Collaboration), Two-jet invariant mass

375

distribution at p ffiffiffi s 376

¼ 1.8 TeV, Phys. Rev. D 41, 1722 (1990). 377

[5] F. Abe et al. (CDF Collaboration), Search for Quark 378 Compositeness, Axigluons and Heavy Particles Using the 379 Dijet Invariant Mass Spectrum Observed in p p ¯ Collisions, 380 Phys. Rev. Lett. 71, 2542 (1993). 381

[6] F. Abe et al. (CDF Collaboration), Search for New Particles 382 Decaying to Dijets in p p ¯ Collisions at p ffiffiffi s 383

¼ 1.8 TeV, Phys. Rev. Lett. 74, 3538 (1995). 384

[7] F. Abe et al. (CDF Collaboration), Search for new particles

385

decaying to dijets at CDF, Phys. Rev. D 55, R5263 (1997). 386 [8] T. Aaltonen et al. (CDF Collaboration), Search for new 387 particles decaying into dijets in proton-antiproton collisions 388 at p ffiffiffi s 389

¼ 1.96 TeV, Phys. Rev. D 79, 112002 (2009).

[GeV]

ZB

M

100 200 300 400 1000 2000 3000

0 0.5 1 1.5

2 2.5

= 1.96 TeV, [21]

s , p p

(2009) CDF 1.13 fb-1

(8 TeV) 18.8 fb-1

CMS

(Data scouting) CMS 18.8 fb-1

(Expected) CMS 18.8 fb-1

1 std. deviation (Expected)

±

2 std. deviation (Expected)

±

= 0.63 TeV, [21]0 63 TeV [2 ss

, pp pp

(1993)( ) UA2 10.9 pbp -1

= 1.8 TeV, [21]

s , p p

(1997) CDF 106 pb-1

1000 200

(Gaussian resonance shapes) = 8 TeV, [14]

s pp,

(2015) ATLAS 20.3 fb-1

= 8 TeV, [18]

s pp,

(2015) CMS 19.7 fb-1

BgCoupling

F4:1 FIG. 4. Observed 95% CL upper limits on the coupling F4:2 g

_B

of a hypothetical leptophobic resonance Z

⁰_B

→ q q ¯ [23]

F4:3 as a function of its mass. The results from this study are F4:4 compared to results obtained with similar searches at different

F4:5

collider energies [14,23].

2

3 4

5

5

[23]

[23]

[23] [19]

Figure 4.2: Limits on the coupling of a hypothetical leptophobic Z’ resonance as a function of the resonance mass, with results from a variety of hadron collider experiments. The green and light blue curves are the results from the CMS and ATLAS 2012 dijet resonance searches, respectively. The black curve and the green and yellow bands indicate the observed and expected limits from the CMS low-mass resonance search using data scouting [100].

a number of resonance models in the range 0.6-0.9 TeV [101]. The dijet mass spectrum observed in this search is displayed in the top part of Fig. 4.3.

After this success, a second scouting trigger was designed and deployed at the HLT for most of the 2012 CMS data taking period. The trigger requiredH_T > 250 GeV and had a maximum rate of 1 kHz. This rate was too high for the PF algorithm to be run for every event, so the trigger instead reconstructed and saved calorimeter jets (‘calo jets’), which are clustered directly from energy deposits in the ECAL and HCAL. Calo jets require negligible HLT resources to reconstruct, and at high mo- mentum their mass resolution is adequate despite the lack of tracking information.

A dijet search was again conducted using the scouting dataset, and the resulting limits on massive Z’ resonances were the best to date between 500 and 800 GeV.

These limits are indicated by the black line in Fig. 4.2. The dijet mass spectrum from this search is shown in the bottom part of Fig. 4.3.

/dm (pb/GeV)σd

10-6

10-5

10-4

10-3

10-2

10-1

1 10

102 CMS Preliminary (0.13 fb-1₎

Fit QCD Pythia

Jet Energy Scale Uncertainty

= 7 TeV s

| < 1.3 η

∆

| < 2.5, | η

| Wide Jets W’ (0.7 TeV)

D (0.7 TeV)

D (1.5 TeV)

Dijet Mass (GeV)

500 1000 1500 2000 2500 3000

Residuals

-2 -1 0 1 2

400 600 800 1000 1200 1400 1600 1800

[pb/GeV] jj / dmσd

−5

10

−4

10

−3

10

−2

10

−1

10 1 10

102 ^Data

Background fit

= 0.37) (M = 700 GeV, gB

Z'B

= 0.84) (M = 1200 GeV, gB

Z'B

Wide jets

| < 1.3 ηjj

| < 2.5, |∆

|η

(8 TeV) 18.8 fb-1

CMS

Dijet mass [GeV]

400 600 800 1000 1200 1400 1600 1800

statσ(Data-Fit)

−3 2

−

−1 0

1 2 3

Figure 4.3: Dijet mass spectra obtained in CMS searches using data scouting in 2011 [101] (top) and 2012 [100] (bottom). For each spectrum, a fit to a predefined functional form is displayed, along with the significance of the data’s deviation from the fit (bottom panel of each plot).