Pileup dependence of HLT rate

Chapter VI: Topological HLT development at √

6.3 Pileup dependence of HLT rate

The HLT rate, normalized by the number of colliding bunches, as a function of the number of pileup interactions for each razor trigger and for different data runs collected in 2015 is shown in Figures6.5and 6.6. Nominally, the dependence of the normalized HLT rate on pileup is expected to be linear,

as is the case for single-object triggers. In contrast, triggers based on sum quantities (such asHT orE_T^miss) and multi-object triggers often demonstrate a nonlinear dependence on pileup, not due to a physical increase in the cross section of the selected physics processes, but rather due to the effects of pileup contamination [165]. To illustrate this, consider the case of a QCD dijet event with no trueE^miss_T . Normally such an event would be rejected by E_T^misstriggers that requireE_T^missabove some threshold, but if some jets from pileup interactions are misinterpreted as part of the event-of-interest then the HLT-reconstructed~p_T^misswill be−∑j∈pileup~p_T^j, which may not perfectly balance to zero as illustrated in Fig.6.4.

As the razor triggers are both based on sum quantities and multiple objects, they also exhibit some nonlinear dependence on pileup. This implies that as the pileup increases at the LHC in 2016 and beyond, either trigger thresh- olds will need to rise dramatically or more sophisticated methods to deal with pileup contamination will need to implemented. One such method is delineated in Ch.9.

0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55

efficiency

0 0.2 0.4 0.6 0.8 1 1.2

CMS preliminary 13 TeV (2.3 fb^-1)

> 500 MR

signal: HLT_RsqMR240_Rsq0p09_MR200 || HLT_RsqMR240_Rsq0p09_MR200_4jet || HLT_Rsq0p25 reference: HLT_Ele27_eta2p1_WPLoose_Gsf

[GeV]

400 600 800 1000 1200

efficiency

0 0.2 0.4 0.6 0.8 1 1.2

/ ndf

χ2 1.033 / 5

p0 0.957 ± 0.01021 p1 217.9 ± 11.04 p2 99.83 ± 10.69

/ ndf

χ2 1.033 / 5

p0 0.957 ± 0.01021 p1 217.9 ± 11.04 p2 99.83 ± 10.69

CMS preliminary 13 TeV (2.3 fb^-1)

> 0.25 R2

signal: HLT_RsqMR240_Rsq0p09_MR200 || HLT_RsqMR240_Rsq0p09_MR200_4jet || HLT_Rsq0p25 reference: HLT_Ele27_eta2p1_WPLoose_Gsf

0.3 0.44 0.6 0.75 0.8 0.83 0.85 0.92

0.55 0.67 0.78 0.87 0.92 0.93 0.95 0.96

0.7 0.77 0.83 0.88 0.95 0.96 0.94 0.89

0.7 0.73 0.81 0.89 0.91 0.95 0.93 1

0.65 0.67 0.68 0.79 0.87 0.95 1 1

[GeV]

400 500 600 700 800 900 1000

0.2 0.3 0.4 0.5

efficiency

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

CMS preliminary 13 TeV (2.3 fb^-1)

signal: HLT_RsqMR240_Rsq0p09_MR200 || HLT_RsqMR240_Rsq0p09_MR200_4jet || HLT_Rsq0p25 reference: HLT_Ele27_eta2p1_WPLoose_Gsf

Figure 6.2: Trigger efficiency of the boolean “or” of the dijet, quadjet and high-R² triggers as used in the search of Ch.8, measured in a data sample of single-electron events as a function of R² (top), MR (middle), and as a function of(M_R, R²)(bottom) [44].

Razor Dijet Trigger Design

2 calo jets,   pT > (70,60)

2 PF jets,  pT > 80 Calo jet

sequence L1 HTT* or SingleJet*

or DoubleJetC*  

or QuadJetC*

Calo MET sequence

Calo R² > 0.02, MR > 100,  (MR +300)(R²+0.25) > 200

PF R² > 0.09, MR > 200,  (MR +300)(R²+0.25) > 240

PF jet sequence

PF MET producer

(a) HLT RsqMR240 Rsq0p09 MR200

Calo R² > 0.02, MR > 100,  (MR +300)(R²+0.25) > 200

PF jet sequence

Razor Quadjet Trigger Design

4 calo jets,   pT>(40,40,30,30)

4 PF jets,  pT>(50,50,40,40) Calo jet

sequence L1 HTT* or SingleJet*

or DoubleJetC*  

or QuadJetC*

Calo MET sequence

PF MET producer

PF R² > 0.09, MR > 200,  (MR +300)(R²+0.25) > 240

(b) HLT RsqMR240 Rsq0p09 MR200 4jet

Calo MET sequence

High-R ² Trigger Design

2 PF jets,  pT > 80

Calo   R² > 0.16 Calo jet

sequence L1 HTT* or SingleJet*

or DoubleJetC*  

or QuadJetC*

PF   R² > 0.25 2 calo jets,   pT > (70,60) PF jet

sequence

PF MET producer

PF R² > 0.02,   MR > 300

Calo MET sequence

PF MET producer

Razor H→bb Trigger Design

3 calo jets, pT>(70,50,30)

2 L3 b-tag   CSV > (0.4, 0.2)

Calo   MR > 200

3 PF jets   pT>(80,60,40) PF jet

sequence Calo jet

sequence

2 PF b-tag   CSV > (0.7,0.4)

PF   60 < mbb < 200, 

pTb > (50,30) L1 HTT* or SingleJet*

or DoubleJetC*  

or QuadJetC*

PF b-tag sequence L3 b-tag

sequence

(d) HLT Rsq0p02 MR300 TriPFJet80 60 40 DoublePFBTagCSV0p7 0p4 Mbb60 200

Figure 6.3: Flow of the producer steps (in purple) and filter steps (in blue) in the razor triggers [44].

Many Challenges with Pileup

Si Xie 9

Jets from pileup interactions may be mis-interpreted as part of the main interaction event

Jet from Primary interaction Jet from pileup

Figure 6.4: Pileup jet misinterpreted as part of the main interaction event [44].

< PU >

0 2 4 6 8 10 12 14 16 18 20 22 24

unprescaled rate / num colliding bx [Hz]

0 0.002 0.004 0.006 0.008 0.01 0.012 0.014

f(x) = -0.00280 + x*0.00047

CMS

Rate Monitoring

36 runs:

259637 (1596 b) 259683 (1813 b) 259685 (1813 b) 259686 (1813 b) 259721 (517 b) 259809 (1813 b) 259810 (1813 b) 259811 (1813 b) 259813 (1813 b) 259817 (1813 b) 259818 (1813 b) 259820 (1813 b) 259821 (1813 b) 259822 (1813 b) 259861 (1813 b) 259862 (1813 b) 259884 (1813 b) 259890 (1813 b) 259891 (1813 b) 260373 (589 b) 260424 (2232 b) 260425 (2232 b) 260426 (2232 b) 260427 (2232 b) 260431 (2232 b) 260532 (2232 b) 260533 (2232 b) 260534 (2232 b) 260536 (2232 b) 260538 (2232 b) 260541 (2232 b) 260575 (2232 b) 260576 (2232 b) 260577 (2232 b) 260593 (2232 b) 260627 (2232 b)

(a) HLT RsqMR240 Rsq0p09 MR200

< PU >

0 2 4 6 8 10 12 14 16 18 20 22 24

unprescaled rate / num colliding bx [Hz]

0 0.0005 0.001 0.0015 0.002 0.0025 0.003

f(x) = -0.00041 + x*0.00007

CMS

Rate Monitoring

36 runs:

(b) HLT RsqMR240 Rsq0p09 MR200 4jet

Figure 6.5: Pileup dependence of the dijet (a) and quadjet (b) razor triggers throughout 2015 [44]. Each data point corresponds to a different luminosity section (23.3 seconds of data-taking). The legend denotes the run number and number of colliding bunches in each run.

< PU >

0 2 4 6 8 10 12 14 16 18 20 22 24

unprescaled rate / num colliding bx [Hz]

0 0.0005 0.001 0.0015 0.002 0.0025

f(x) = -0.00034 + x*0.00005

CMS

Rate Monitoring

36 runs:

< PU >

0 2 4 6 8 10 12 14 16 18 20 22 24

unprescaled rate / num colliding bx [Hz]

0 0.005 0.01 0.015 0.02 0.025

f(x) = -0.00296 + x*0.00080

CMS

Rate Monitoring

36 runs:

(d) HLT Rsq0p02 MR300 TriPFJet80 60 40 DoublePFBTagCSV0p7 0p4 Mbb60 200

Figure 6.6: Pileup dependence of the high-R² (c) and H(bb)(d) razor triggers [44]. A detailed description of the graphs is given in Fig.6.5

Part III

Searches for new physics at the LHC

Chapter 7

SEARCHES FOR SUPERSYMMETRY AT √

S = 8 TEV

As discussed in Sec.3.8, models of SUSY predict additional, undiscovered fundamental particles which correspond to the heavy superpartners of SM particles. Of particular interest is the production of top squarks, bottom squarks, and gluinos due to their role in taming the quadric divergence of the Higgs mass in the SM (see Sec. 3.5). The residual fine-tuning inherent in these models is dependent on the masses of these superpartners, with a preference for smaller masses to avoid large fine-tuning. These consid- erations have motivated searches for the lightest allowed top and bottom squarks, as well as gluinos that may couple to top/bottom squarks, whose decays would produce final states enriched in b-jets. Moreover, due to the possible presence of top quarks that produce leptons∼ 30% of the time in the decay chain, the presence of electrons or muons may be used as part of the event selection to enhance the signal-to-background ratio. We exploit both of these features (b-jets and leptons) in the event classification.

We classify events into different “boxes,” or data categories, based on the jet multiplicity, b-jet multiplicity, and lepton (e orµ) multiplicity (see Fig.7.1).

The advantages of this classification are (i) by isolating different SM background processes, like tt, we can better model them individually, and (ii) in the event of a discovery, we may be able to infer the values of certain SUSY branching fractions based on the boxes where the signal is present. The like- lihood functions of the different boxes are statistically combined to exclude or discover particular SUSY models. In the following, this classification and statistical combination is referred to as the razor box approach.

In this chapter, we present an inclusive search for gluinos and top squarks¹ using pp collision data collected by CMS at√

s=8 TeV in the context of the minimal natural SUSY spectrum outlined in Sec.3.8[94]. Previous searches for natural SUSY by CMS [102, 103, 104, 105, 106] and ATLAS [97, 98, 99, 100, 101] at √

s = 7 and 8 TeV have probed gluino masses up to 1.3 TeV

1Though we don’t explicitly interpret our results in the context of bottom squark production, many of the conclusions regarding top squark production carry over.

and top squark masses up to 700 GeV under the assumptions of specific decay modes for the SUSY particles. One important feature that sets this search apart is that the SUSY parameter space of gluino and top squark branching ratios is explored for the first time at the LHC [94]. Notably, the razor box approach ensures good sensitivity to a wide range of branching ratios. In addition, we combine the results from the hadronic razor search with those from a previous search [102] for top-squark production in the single-lepton (eor µ) channel to obtain an improved bound on top-squark pair production.

The remainder of this chapter is organized as follows. The event selection and box definitions are detailed in Sections 7.1 and 7.2, respectively. The modeling of the SM backgrounds through a fit using an empirical function is explained in Sec.7.3. In particular, the motivation for this empirical function and the properties that make it a suitable description are examined in Sec. 7.3. The results of the fits to data are presented and compared to the corresponding results in a signal injection scenario in Sec.7.4. Finally, lim- its are derived in the context of the natural SUSY scenario of Sec.3.8 and a summary is given in Sections7.5and7.8, respectively.

Dalam dokumen Naturalness Confronts Nature: Searches for Supersymmetry with the CMS Detector in pp Collisions at √s = 8 and 13 TeV (Halaman 111-120)

Chapter VI: Topological HLT development at √

6.3 Pileup dependence of HLT rate

Razor Dijet Trigger Design

Razor Quadjet Trigger Design

High-R 2 Trigger Design

Razor H→bb Trigger Design

Many Challenges with Pileup

Part III

Searches for new physics at the LHC

SEARCHES FOR SUPERSYMMETRY AT √

S = 8 TEV

High-R ² Trigger Design