Top Quark Mass - IN2P3moriond.in2p3.fr/QCD/2016/SundayAfternoon/McCarthy.pdfT.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 1 Rencontres
Post on 08-May-2018
219 Views
Preview:
Transcript
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 1
Rencontres de Moriond QCD 2016 La Thuile, Aosta Valley, Italy
Measurements @ ATLAS & CMS
Sunday, March 20th, 2016
Top Quark Mass
Thomas McCarthy1on behalf of the ATLAS & CMS Collaborations
1Max-Planck-Institut für Physik, München
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 2
General Overview
1 Overview & Highlights of LHC Run 1 mtop Measurements (√s = 7 & 8 TeV)
2 LHC Combinations from Run 1
3 Summary & Outlook
https://twiki.cern.ch/twiki/bin/view/AtlasPublic/TopPublicResultsLink to ATLAS Top Quark Public Results
Link to CMS Top Quark Public Resultshttps://twiki.cern.ch/twiki/bin/view/CMSPublic/PhysicsResultsTOP
A nice summary of LHC mtop
measurements @ Run 1…G. Cortiana (MPI Munich) Submitted to Reviews in Physics
“Top-quark mass measurements: review and perspectives”
http://arxiv.org/abs/1510.04483
pp
Direct measurements from kinematics & event reconstructionAlternative measurements (e.g. from σtt)-
Focus today on a select number of analyses (many more results available, see backup or below)
mtop a fundamental parameter of the Standard Model Short lifetime O(10-25) ➙ no bound hadronic states formed Direct access to top quark properties via its decay products Better knowledge of mtop ➙ better performance in analyses with top backgrounds
Top Quark{
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016
tW +
bW
proton
proton
b
+
q q
3
Measurements of mtop @ the LHC: An Overview
t
W +
bb_
W
t_
_
proton
q
q
q
qproton
_
_
gg
g
7 TeV8 TeV
7 TeV8 TeV
Eur.Phys.J.C. (2015) 75:158
Combined Channels (7&8 TeV)
All-Hadronic (All-Jets) tt Channel
Single-Top Enriched Events
8 TeV
t
W +
b
b_
W
t_
_
proton
proton_
gg
g
+
g
tt + 1-Jet Channel
7 TeV
-
t
W +
bb_
W
t_
_
proton
proton
gg
g_
_
+
7 TeV8 TeV
7 TeV8 TeV
Eur.Phys.J.C. (2015) 75:330
Combined Channels (7&8 TeV)
Dileptonic tt Channel--
Top Mass from tt Cross-Section
7/8 TeV
7/8 TeV
-
t
W +
bb_
W
t_
_
proton
q
qproton
_
gg
g
+
7 TeV8 TeV
7 TeV8 TeV
Eur.Phys.J.C. (2015) 75:330
Combined Channels (7&8 TeV)
Semileptonic tt Channel-
ATLAS-CONF-2014-055 ATLAS-CONF-2014-053Eur.Phys.J. C74 (2014) 3109
CMS-PAS-TOP-13-004 *Only most precise measurements in each channel summarized on this page. See links to additional analyses in overviews on later slides.
*
CERN-PH-EP-2015-234 CERN-PH-EP-2015-234 CERN-PH-EP-2015-234
arXiv:1603.02303
arXiv:1509.04044 arXiv:1509.04044arXiv:1509.04044
CMS-PAS-TOP-15-0018 TeV
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 4
Eur.Phys.J.C. (2015) 75:158
arXiv:1509.04044
Updated & Combined Channels:
t
W +
bb_
W
t_
_
proton
q
q
q
qproton
_
_
gg
g
All-Hadronic (All-Jets) tt Channel- Direct mtop MeasurementsThe All-Hadronic tt Channel
Advantages:Largest branching ratio Full event reconstruction possible (no neutrinos in W decays)
Combinatorics: large number of jet-parton associations Tight cuts required due to very large multi-jet background
Disadvantages:
-
7 TeV
7 TeV
1D template method
2D template method
= Covered in this Talk= See Backup
2D template method8 TeV
Eur.Phys. J.C. 74 (2014) 2758
(1) W Branching Ratios from: K.A. Olive et al. (Particle Data Group), Chin. Phys. C, 38, 090001 (2014) and 2015 update
tt BR ~ 45%-(1)
CERN-PH-EP-2015-234
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 5
MC Mass
Pole MassCERN-PH-EP-2015-234Updated/Combined
Channels (7/8 TeV)√s = 7 TeV
√s = 8 TeVAll-Hadronic Channel @ 8 TeV
(∆mtop = 0.29 GeV) [0.2%] (∆mtop = 0.25 GeV) [0.1%] (∆mtop = 0.20 GeV) [0.1%] (∆mtop = 0.19 GeV) [0.1%]
Dominant Sources of Uncertainty (hybrid)
Measured (Improved!) Result:
7 TeV Result:
2D Template Method mtop extracted together with global JSF (compared w/ 1D)
Updated since Moriond 2015!
b-Dependent JEC Data Statistics Backgrounds In-Situ JEC
Jet energies modified by constraints from kinematic fits using known W boson mass Multi-jet background is modelled using an event-mixing technique (able to get fsig ~ 78%!) Likelihood built from S & B shapes, splitting signal into correct or wrong (from simulation)
arXiv:1509.04044
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 6
Direct mtop Measurements
Advantages:Good balance: large BR and modest cuts required Leptonic top identifies event (via lepton) Hadronic top can then be fully reconstructed
Combinatorics remains an issue (typically ≥ 4 jets)
Disadvantages:
t
W +
bb_
W
t_
_
proton
q
qproton
_
gg
g
+
Eur.Phys.J.C. (2015) 75:330
Semileptonic tt Channel-
The Semileptonic tt Channel-
7 TeV
7 TeV
3D template method
2D template method
= Covered in this Talk= See Backup
1D µ+jets with BEST8 TeV CMS-PAS-TOP-14-011
JHEP 12 (2012) 105
Updated & Combined Channels:2D template method8 TeV
tt BR ~ 34%-(1)
(1) W Branching Ratios from: K.A. Olive et al. (Particle Data Group), Chin. Phys. C, 38, 090001 (2014) and 2015 update (Value quoted includes all lepton types: e/µ/𝜏 except hadronic 𝜏 decays)
J/𝜓 (l+jets / dileptonic)8 TeV CMS-PAS-TOP-15-014
(Would be ~44% including hadronic tau decays)
e + jets µ + jets
e + jets µ + jets
e + jets µ + jets
e µ + jets
e/µ + J/𝜓 + jets
CERN-PH-EP-2015-234 BR ~ 0.032%arXiv:1509.04044
e/µ + hadrons + jets8 TeV lepton / secondary verticesJust Released}CMS-PAS-TOP-12-030
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 7
MC Mass
Pole Mass
√s = 7 TeV
√s = 8 TeV+ Jets Channel @ 8 TeV Updated/Combined Channels
(∆mtop = 0.35 GeV) [0.2%] (∆mtop = 0.16 GeV) [0.1%] (∆mtop = 0.16 GeV) [0.1%]
Dominant Sources of Uncertainty (hybrid)
b-Dependent JEC Semileptonic B-decay modelling Data Statistics
Left: large reduction in (mostly combinatorial) background by applying goodness-of-fit (Pgof) weighting Above: Also evaluate mt separately in different bins of kinematic variables sensitive to top quark production & decay
Final measured result (1D vs hybrid compared):
CERN-PH-EP-2015-234
Precision of ~0.3%!
Best-precision measurement in a single channel at CMS
Analysis strategy similar to 7 TeV and 7/8 TeV all-hadronic
Fitted jet energies used to reconstruct final observable
Updated since Moriond 2015!
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 8
MC Mass
Pole Mass
Physics Analysis Summary (PAS)
√s = 7 TeV
√s = 8 TeVLepton(s) + J/𝜓 Events @ 8 TeV
Measured Result:
Very small BR, but sensitivity to mtop w/o using jets to build observable (avoid JES/bJES)
Variation on semileptonic / dileptonic channels, requiring at least one of the top quarks have a final leptonic state For the main leptonic top quark decay, aim to identify cases where b-quark decays via (b ➙J/𝜓 + X ➙ µ+µ- + X)
Remains statistically limited for now
Top Quark pT ME-PS Matching Threshold Renormalization Scale
(∆mtop = -0.64 GeV) [0.4%] (∆mtop = +0.58 GeV)[0.3%] (∆mtop = +0.46 GeV)[0.3%]
Dominant Sources of Uncertainty
BR ~ 0.032%
CMS-PAS-TOP-15-014
New Result!
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 9
MC Mass
Pole Mass
Will supersede previous result
√s = 7 TeV
√s = 8 TeVSecondary Vertices + Lepton(s)
Sensitivity to mtop primarily from leptons (e/µ) and via decay lengths of charged hadrons (from b-quark decay) Target signal events in both semileptonic and dileptonic channels:
e + jets µ + jets
ee µµ eµDileptonic tt Channel:
Semileptonic tt Channel:
--
Reconstruct charged hadrons (left: J/𝜓, but also D0, D*) resulting from b-quark decays Allows one to retain a stronger sensitivity to mtop without inclusion of jets (where related systematics would otherwise dominate) Ultimately select msvl (invariant mass of lepton / secondary vertex) as mtop-sensitive observable Inclusion of b-quark fragmentation studies (leading systematic)
More generalized version of J/𝜓 analysis
Measured Result:
Dominant systematics: top quark pT & b-quark fragmentation
CMS-PAS-TOP-12-030
New Result!
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 10
Direct mtop Measurements
Advantages:Extremely low backgrounds (typically a few percent) Two lepton/b-jet permutations (combinatorics simplified)
Two neutrinos ➙ full event reconstruction not possible Loss of sensitivity to mtop ➙ energy taken by neutrinos
Disadvantages:
The Dileptonic tt Channel
t
W +
bb_
W
t_
_
proton
proton
gg
g_
_
+
Dileptonic tt Channel-
-
Eur.Phys.J.C. (2015) 75:3307 TeV
7 TeV
1D template with mlb
kinematic endpoints = Covered in this Talk= See Backuptemplate AMWT7 TeV Eur.Phys. J.C. 72 (2012) 2202
Eur.Phys. J.C. 73 (2013) 2494
template AMWT8 TeV
eµ b-jet energy peak8 TeV CMS-PAS-TOP-15-002
Updated & Combined Channels:
tt BR ~ 6%-(1)
(Would be ~11% including hadronic tau decays)
(1) W Branching Ratios from: K.A. Olive et al. (Particle Data Group), Chin. Phys. C, 38, 090001 (2014) and 2015 update (Value quoted includes all lepton types: e/µ/𝜏 except hadronic 𝜏 decays)
ee µµ eµ
ee µµ eµ
ee µµ eµ
ee µµ eµ
ee µµ eµ
CERN-PH-EP-2015-234
CMS-PAS-TOP-14-0148 TeV mlb with forward folding ee µµ eµ
arXiv:1509.04044
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 11
√s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass Eur.Phys.J.C. (2015) 75:330
One/two b-tagged jets required in event selection
Small amount of background in Nb-tag=1 events (~3%) (predominantly cases of fake leptons)
Invariant mass mlb formed by considering both lepton/b-jet permutations (two such pairings / event)
Select pairing with minimum average mlb (motivated by performance in simulation)
Dilepton Channel @ 7 TeV
t b
W +
t b
W -
-_
_
+
-
}}
1D template method employing simulated MC datasets with varying values of mtop
Measured Result:
Dominant Sources of Uncertainty
(∆mtop = 0.75 GeV) [0.4%] (∆mtop = 0.68 GeV) [0.4%] (∆mtop = 0.54 GeV) [0.3%] (∆mtop = 0.53 GeV) [0.3%]
Jet Energy Scale (JES) bJet Energy Scale (bJES) Data Statistics Hadronization Modelling
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 12
MC Mass
Pole Mass
√s = 7 TeV
√s = 8 TeVPhysics Analysis Summary (PAS) CMS-PAS-TOP-15-002Dilepton Channel @ 8 TeV
New since Moriond 2015!
Measured Result:
(∆mtop = 1.50 GeV) [0.9%] (∆mtop = 1.50 GeV) [0.9%] (∆mtop = 1.23 GeV) [0.7%] (∆mtop = 1.17 GeV) [0.7%]
Dominant Sources of Uncertainty
t b
W + +
Specifically target the sensitivity of the peak of the b-jet energy spectrum to mtop
pT ≥ 17 GeV (1st lepton) [Offline: 20 GeV]
pT ≥ 8 TeV (2nd lepton) [Offline: 20 GeV]
Measured peak position re-calibrated to correct for observed bias (from simulation) [lower left]
eµ channel, trigger two oppositely charged leptons:
Generator Modelling Top pT Reweighting Jet Energy Scale (JES) Data Statistics
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016
tW +
bW
proton
proton
b
+
q q
13
Direct mtop Measurements
Advantages:Access to mtop through EW production of top quarks Orthogonal datasets w.r.t. other channels ➙ combinations
Lower cross-sections and higher backgrounds
Disadvantages:
Single-Top Enriched Channel
Single-Top Enriched Events
ATLAS-CONF-2014-0558 TeV 1D template m(lb) & NN
tt BR ~ 25%-(1)
(1) W Branching Ratios from: K.A. Olive et al. (Particle Data Group), Chin. Phys. C, 38, 090001 (2014) and 2015 update (Value quoted is for single top quarks and includes all lepton types: e/µ/𝜏 except hadronic 𝜏 decays)
(Would be ~33% including hadronic tau decays)
e + jets µ + jets
e µ + jetst-channel with µ + jets8 TeV CMS-PAS-TOP-15-001
= Covered in this Talk= See Backup
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 14
MC Mass
Pole Mass
√s = 7 TeV
√s = 8 TeVATLAS-CONF-2014-055Single Top Enhanced (t-Channel) @ 8 TeV
Jet Energy Scale (JES) t-Channel Hadronization Data Statistics W+jets Normalization
Dominant Sources of Uncertainty
(∆mtop = 1.5 GeV) [0.9%] (∆mtop = 0.7 GeV) [0.4%] (∆mtop = 0.7 GeV) [0.4%] (∆mtop = 0.4 GeV) [0.2%]
Measured Result:
build control region for W+jets background by loosening b-tagging requirement
Left: Following neural net, m(lb) observable selected for mtop sensitivity (1D template method)
tW +
bW
proton
proton
b
+
qq}}e/µ
2 jets (1 b-tag)
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 15
MC Mass
Pole Mass
√s = 7 TeV
√s = 8 TeVSingle-Top Enriched with µ + Jets Physics Analysis Summary (PAS)
CMS-PAS-TOP-15-001
LO (5-flavour scheme) NLO (4-flavour scheme)
Single-Top t-Channel Productiont
W +
bW
proton
proton
b
+
qq }}µ
2 jets (1 b-tag)
Require 1 high-pT, isolated reconstructed muon Require 2 reconstructed jets w/ 1 b-tag (signal region)
Zero b-tag events characterize the W+jets background
Invariant mass of reconstructed top (mlνb) is selected as the mtop-sensitive observable
In final extended unbinned likelihood fit, the signal (single-top) normalization + shape parameters left to float
Extraction of mtop comes from mean parameter of Crystal Ball (used to parameterize signal shape)
JES systematic uncertainty dominates (∆mtop ~ 0.68 GeV)
Result shown at QCD EW Session last week (link to agenda)
Measured Result:
New Result!
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 16
Indirect Measurements
Advantages:Access to top quark pole mass directly
Compounded combinatorics
Disadvantages:
Pole Mass in tt + 1-jet Channelt
W +
b
b_
W
t_
_
proton
proton_
gg
g
+
g
tt + 1-Jet Channel-
-
ATLAS-CONF-2014-0537 TeV pole mass in tt+1-jet = Covered in this Talk
tt BR ~ 34%-(1)
(Would be ~44% including hadronic tau decays)
(1) W Branching Ratios from: K.A. Olive et al. (Particle Data Group), Chin. Phys. C, 38, 090001 (2014) and 2015 update (Value quoted includes all lepton types: e/µ/𝜏 except hadronic 𝜏 decays)
e + jets µ + jetsJHEP 10 (2015) 121
One leptonically decaying top quark ➙ keeps background low
where difference between MC & pole mass O(1 GeV)(A. Hoang: arXiv:1412.3649v1)
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 17
Pole Mass
MC Mass
JHEP 10 (2015) 121Pole Mass from tt+1-jet events @ 7 TeV-
Updated since Moriond 2015!
Measured Result:
Dominant Sources of Uncertainty
(∆mtop = 1.5 GeV) [0.9%] (∆mtop = 0.94 GeV) [0.5%] (∆mtop= +0.93/-0.44 GeV)[+0.5/-0.3 %] (∆mtop = 0.72 GeV) [0.4%]
arXiv:1303.6415Motivated by theory:
where:
Clean channel w/ ~6% background !
Very promising pole mass result
Statistical uncertainty quite large (but √s = 7 TeV dataset)
(m0 = 170 GeV)
(largest: single top)
Extract top quark pole mass directly using differential tt+1-jet cross-section-
√s = 7 TeV
√s = 8 TeV
Data Statistics JES & bJES Scale Variations ISR/FSR
ATLAS-CONF-2014-053
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 18
Indirect Measurements
Advantages:Access to top quark pole mass directly Can take advantage of low backgrounds of eµ channel
Sensitivity not as strong as in direct measurements Systematic uncertainties typically larger
Disadvantages:
mtop from Production Cross-Section
Top Mass from tt Cross-Section-
Phys.Lett.B. 738 (2014) 526
7/8 TeV
7 TeV
pole mass eµ channel
first mass from XS
= Covered in this Talk= See Backup
pole mass eµ channel7 TeV
JHEP 07 (2011) 049
Eur.Phys.J. C74 (2014) 3109 ee µµ eµ
ee µµ eµ
ee µµ eµ
ee µµ eµ7/8 TeV pole mass eµ channel CMS-PAS-TOP-13-004
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 19
Pole Mass √s = 8 TeV
MC Mass √s = 7 TeVEur.Phys.J. C74 (2014) 3109Pole Mass from Cross-Section at √s = 7 & 8 TeV
Measured Result:
MC Mass
Pole Mass √s = 8 TeV
√s = 7 TeV
Great improvement over previous measurement:
Pole Mass from Cross-Section at √s = 7 & 8 TeV
Measurement relies on a given choice of PDF sets and 𝛼s (quoted result uses NNPDF3.0, 𝛼s = 0.118 ± 0.001)
Consistent results using CT14 and MMHT2014, and also comparing √s = 7 & 8 TeV separately Left: √s = 7 & 8 TeV measured values & prediction vs. mtop
Measurement of σtt together with NNLO theoretical prediction allows for extraction of the pole mass (mt)
Luminosity
Dominant Uncertainties
PDFs & 𝛼s
Considered PDF sets: MSTW2008, CT10, and NNPDF2.3Ultimately quoted value comes from maximizing a product of likelihoods based on √s = 7 & 8 TeV
Final result reflects combination of 7 & 8 TeV datasets (as above)
Submitted to JHEP
arXiv:1603.02303
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 20
Combinations of ATLAS & CMS Results
ATLAS/CMS mtop Combinations
Updated since Moriond 2015!
https://twiki.cern.ch/twiki/bin/view/AtlasPublic/TopPublicResults
Link to ATLAS Top Quark Public Results
Link to CMS Top Quark Public Resultshttps://twiki.cern.ch/twiki/bin/view/
CMSPublic/PhysicsResultsTOP
ATLAS-CONF-2013-102
LHC / Tevatron (World) Combination
LHC Combination (√s = 7 TeV)
CMS-PAS-TOP-13-014
ATLAS-CONF-2014-008
CDF Note 11071
D0 Note 6416
CMS-PAS-TOP-13-005
Precision of ~0.3%!
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 21
Summary & Outlook
https://twiki.cern.ch/twiki/bin/view/LHCPhysics/SingleTopRefXsechttps://twiki.cern.ch/twiki/bin/view/LHCPhysics/TtbarNNLO
were produced during 2012 data-taking!
At peak instantaneous luminosities @ ATLAS & CMS:
~2 top pairs/sec~1 single top/sec
~4M top quarks produced in 2011 (√s = 7 TeV)
~24M top quarks produced in 2012 (√s = 8 TeV)
~13M top quarks produced in 2015 (√s = 13 TeV)
Era of Precision Top Quark Measurements
(per experiment)
Overall from both experiments combined:
Cross-section values taken from:
pp
√s = 13 TeV
Increasing 𝛾top = Etop/mtop
qq_
b
tq
q
_
t b
W +
Planned measurements of mtop @ √s = 13 TeV Higher √s ➙ greater % of boosted top quarks As always, trade off between with stats / syst Avoiding jets ➙ avoid (b)JES systematics…but b-quark fragmentation / top pT systematics increase New methods should aim to strike optimal balance (see nice LHC talk in EW session last week) [link]
Great success by ATLAS & CMS in making precision mtop measurements during Run 1! Several novel approaches with a variety of final-state signatures (& differing backgrounds)
Thanks for your attention.
Looking Ahead to Run II…
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 22
Backup Material
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 23
Combinations of ATLAS & CMS Results
CMS-TOP-14-0027 & 8 TeV
CMS mtop Combinations
A variety of analyses in all main channels
Many analyses employ orthogonal datasets (greater impact on combination)
Dominant contribution from √s = 8 TeV lepton plus jets channel measurement (72.5% BLUE combination coefficient)
Individual per-systematic correlation values used for performing combination are itemized (by decay channel / year)
There remains some tension with results of the Tevatron combination (2014)
New since Moriond 2015!
Run 1 Comb CMS-PAS-TOP-14-015
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 24
Combinations of ATLAS & CMS Results
CMS-TOP-14-0027 & 8 TeV
CMS mtop Combinations
Run 1 Comb CMS-PAS-TOP-14-015
New Summary March 2016
CMS-PAS-TOP-15-001
CMS-PAS-TOP-15-014
arXiv:1603.02303
CMS-PAS-TOP-12-030
Eur. Phys. J. C 73 (2013) 2494
CMS-PAS-TOP-15-002
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 25
Combinations of ATLAS & CMS Results
ATLAS-CONF-2013-102
LHC / Tevatron (World) Combination
LHC Combination (√s = 7 TeV)
CMS-PAS-TOP-13-014
ATLAS-CONF-2014-008
CDF Note 11071
D0 Note 6416
CMS-PAS-TOP-13-005
Summary Plots
ATLAS mtop Combinations
Current combinations include only results using √s = 7 TeV dataset
Expect increased precision at √s = 8 TeVATLAS Sumary PlotsSemi-/di-leptonic (7 TeV) Eur.Phys.J.C. (2015) 75:330
As with CMS, greatest sensitivity in the semileptonic (lepton+jets) channel
In addition several analyses of pole mass performed
Also see:
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 26
Top Quark Production Cross-Sections (Top Quark Pairs)
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 27
Top Quark Production Cross-Sections (Single Top)
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 28
Top Quark Pole Mass Measurements
https://twiki.cern.ch/twiki/bin/view/AtlasPublic/TopPublicResultsLink to ATLAS Top Quark Public Results
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 29
Top Quark Mass World (LHC + Tevatron) CombinationLHC / Tevatron (World) Combination
CMS-PAS-TOP-13-014
ATLAS-CONF-2014-008
CDF Note 11071
D0 Note 6416
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 30
(Some!) Observables Sensitive to mtop in Template Measurements
Sensitivity to
mtop
Invariant mass of the lepton/b-jet pair
t b
W +
t b
W -
-_
_
+
-
Peak of the b-jet energy spectrum
Invariant triple-jet mass
Ratio of reconstructed top-to-W masses (‘R’ or ‘R3/2’)
Advantages:No full kinematic event reconstruction required
Avoids to some extent ambiguity from MET (2 neutrinos)
Lost sensitivity to mtop from neutrinoDisadvantages:
Advantages:Reconstructed top quark most sensical observable to probe true top quark properties
Linearity with generator mtop is likely the most intuitive
t b
W +
Very large susceptibility to fluctuations in JES
Full event reconstruction required (hadronic side)
Combinatorics (jet-parton assignment difficult)
Disadvantages:
q
q
_
t b
W + +
t b
W + q
q
_
}}}
}}Full event reconstruction required (hadronic side)
Susceptibility to bJES uncertainty persists
Disadvantages:
Advantages:Similar sensitivity as mjjj
Less susceptibility to JES ➙ mitigated by similar sensitivity to JES in denominator
Also avoids loss of stats in 2D (/3D) template method
No longer have leptons or neutrinos as probes to top
Increased JES and top quark pT modelling systematics
Disadvantages:
Advantages:No full kinematic event reconstruction required
Presence of leptons + b-tagged jets enough to trigger event (very low backgrounds in dilepton channel)
Technically possible in any channel
Combinatorics not an issue
Eb-jet
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 31
√s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass
Kinematic likelihood employed for event reconstruction
1D template method with R3/2 = mjjj/mjj as mtop-sensitive observable
R3/2 observable: suppress the contribution from JES systematic uncertainty Data-driven ‘ABCD’ method employed for multi-jet background estimation
In final fit two parameters are left to float: top mass (mt)& the fraction of bkgd events (fbkg)
Eur.Phys.J.C. (2015) 75:158All-Hadronic Channel @ 7 TeV
Data Statistics bJet Energy Scale (bJES) Hadronization Modelling Jet Energy Scale (JES)
Dominant Sources of Uncertainty
(∆mtop =1.4 GeV) [0.8%] (∆mtop = 0.62 GeV) [0.4%] (∆mtop = 0.50 GeV) [0.3%] (∆mtop = 0.51 GeV) [0.3%]
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 32
√s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass
Employs ideogram method Investigated precision of 1D or 2D method (2D method simultaneously extracts global JES) Likelihood variable based on S & B parameterizations:
Updated/Combined Channels (7&8 TeV)
Particular Analysis Eur.Phys. J.C. 74 (2014) 2758All-Hadronic Channel @ 7 TeV
1D Method
2D Method together with
global JES
Signal split into three permutation cases: (correct, incorrect, unmatched) from simulation Event weights lessen impact of incorrect jet-parton associations from kinematic fit (keeping total number of events unchanged)
Jet energies modified by constraints from kinematic fits Jet Energy Scale (JES)
Data Statistics Data Statistics
(∆mtop = 0.97 GeV) [0.6%] (∆mtop = 0.69 GeV) [0.4%] (∆mtop = 0.49 GeV) [0.3%]
Dominant Sources of Uncertainty (1D)
CERN-PH-EP-2015-234
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 33
MC Mass
Pole Mass
Updated/Combined Channels (7&8 TeV)
√s = 7 TeV
√s = 8 TeVAll-Hadronic Channel @ 8 TeV CERN-PH-EP-2015-234
arXiv:1509.04044
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 34
√s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass Eur.Phys.J.C. (2015) 75:330+ Jets Channel @ 7 TeV
Targets three different observables, each sensitive to a particular key parameter Background also sensitive to JSF, bJSF
Build template parameterizations and perform a global 3D fit to extract all parameters simultaneously
Observable
Target sensitivity
Observable
Target sensitivity
Observable
Target sensitivity
Where
is defined differently for
events with 1 or ≥ 2 b-tagged jets{
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 35
√s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass Eur.Phys.J.C. (2015) 75:330
Background fraction from ~ 25% (1 b-tag events) to 3% (≥2 b-tag events)
+ Jets Channel @ 7 TeV
Dominant Sources of Uncertainty (3D)
(∆mtop = 0.75 GeV) [0.4%] (∆mtop = 0.58 GeV) [0.3%] (∆mtop = 0.50 GeV) [0.3%] (∆mtop = 0.32 GeV) [0.2%]
Measured Results:
(mostly W+jets) (mostly single top)
1D Template Method
Use of 3D template method (over 1D) improves JES / bJES syst uncertainties
2D Template Method 3D Template Method
Also correlations between this and dilepton analysis reduced (useful for combinations)
Statistics (incl. JSF/bJSF cont.)
Jet Energy Scale (JES) b-Tagging ε / mistag ISR/FSR
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 36
MC Mass
Pole MassPhysics Analysis Summary (PAS)
√s = 7 TeV
√s = 8 TeVCMS-PAS-TOP-14-011+ Jets Channel @ 8 TeV
Alternate √s = 8 TeV analysis performed in µ+jets channel
Template method employing R observable (ratio of reconstructed top/W masses)
Bi-Event Subtraction Technique (BEST) used to estimate background shape
Note good agreement in both distributions:
(Left)
(Right)
Also note shapes differ
Crucially R is more resistant to shifts in JEC
{}
Reconstructed top (from event X)
Reconstructed W
1 2
Replace 1 non-tagged jet with
one from event Y
{} R = _________m( )m( )
Yields a distribution of R representative of combinatorial background
W-mixing with BEST (b-mixing similar)
Combinatorial W
Combinatorial top
New since Moriond 2015!
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 37
MC Mass
Pole MassPhysics Analysis Summary (PAS)
√s = 7 TeV
√s = 8 TeVCMS-PAS-TOP-14-011+ Jets Channel @ 8 TeV
Measured Result:
(∆mtop = 0.57 GeV) [0.3%] (∆mtop = 0.40 GeV) [0.2%] (∆mtop = 0.34 GeV) [0.2%] (∆mtop = 0.23 GeV) [0.1%]
Dominant Sources of Uncertainty
Left: Good agreement in between signal and data-minus-background
Total systematic using R (∆mtop = 0.90 GeV)Total systematic using mbjj (∆mtop = 2.45 GeV)
Note this is an alternate analyses of the full √s = 8 TeV dataset in l+jets channel
Precision of ~0.3%!
The analysis covered in the main talk yields the following result:
Data Statistics Renorm/Factor. Scale Flavor-Dep. Hadronization b Fragment. + B had. BR
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 38
Combined Channels (7&8 TeV)
Particular Analysis √s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass
JHEP 12 (2012) 105+ Jets Channel @ 7 TeV
Build 2D likelihood function to extract simultaneously mtop and a global JES
Per-permutation weights applied to heighten contribution from correct jet-quark assignments
Left (upper): all permutations for reconstructed mtop prior to kinematic fit and Pgof weighting
Left (lower): following Pgof cut, weighting and kinematic fit
Dominant Systematics: Color Reconnection (0.54 GeV), bJES (0.61 GeV)
Build templates for all simulated signal{Here ‘correct’
permutations shown but done for all 3 categories
Strong sensitivity to generator mtop as expected
Measured Results:
CERN-PH-EP-2015-234
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 39
MC Mass
Pole Mass
√s = 7 TeV
√s = 8 TeV+ Jets Channel @ 8 TeV
Comparisons between 1D, 2D and hybrid template method Hybrid: incorporates prior knowledge about JSF using a Gaussian constraint Best precision obtained in hybrid case as in all-had channel
Updated since Moriond 2015!
Updated/Combined Channels CERN-PH-EP-2015-234
arXiv:1509.04044
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 40
√s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass Eur.Phys.J.C. (2015) 75:330+ Jets + Dilepton Channel @ 7 TeV
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 41
√s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass Eur.Phys.J.C. (2015) 75:330+ Jets + Dilepton Channel @ 7 TeV
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 42
Combined Channels (7&8 TeV)
Particular Analysis √s = 7 TeV
√s = 8 TeV
MC Mass
Pole Mass
Eur.Phys. J.C. 73 (2013) 2494Dilepton Channel @ 7 TeV
Measured Result:
(∆mtop = +1.3/-1.8 GeV) [+0.7/-1.0%] !(∆mtop ~ ±0.6 GeV) [0.3%]
Dominant Sources of Uncertainty
Analysis strategy ideally suited to scenarios when pairs of parent particles each decay to a partially invisible final state, e.g.
General approach allows for measurement of all three masses – mtop, mW and m𝜈 (or DM rather than neutrinos)
Alternatively constrain system using mW and m𝜈 for precision mtop measurement (as is done here)
t b
W + +
t
W -
-b_
_
-Key observable (MT2) based on the more familiar transverse mass variable (MT):
Sensitivity to kinematic endpoints in distributions
QCD Effects Background Shape Fit Range
Jet Energy Scale (JES)
CERN-PH-EP-2015-234
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 43
MC Mass
Pole Mass
√s = 7 TeV
√s = 8 TeVDilepton Channel @ 8 TeV
(Updated!) Measured Result:
Consider ee, µµ, eµ channels (modest background [~18%], largely D-Y) Employ an Analytic Matrix Weighting Technique (AMWT) Require ≥ 1 b-tagged jet (if < 2 also take non-tagged jet with leading pT)
mtop extracted from a likelihood fit based on templates constructed from simulation
Consider large number of possible neutrino momenta and weight solution accordingly via:
where:
7 TeV Result:
Updated since Moriond 2015!
Renorm/Factor. Scales b Fragmentation JEC b Flavour Comp
(∆mtop = 0.75 GeV) [0.4%] (∆mtop = 0.69 GeV) [0.4%] (∆mtop = 0.34 GeV) [0.2%]
Dominant Sources of Uncertainty
Associated 7 TeV Result
Updated/Combined Channels
Eur.Phys. J.C. 72 (2012) 2202
CERN-PH-EP-2015-234
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 44
MC Mass
Pole Mass
√s = 7 TeV
√s = 8 TeVDilepton Channel @ 8 TeVAssociated 7 TeV Result
Updated/Combined Channels
Eur.Phys. J.C. 72 (2012) 2202
CERN-PH-EP-2015-234
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 45
MC Mass
Pole Mass
√s = 7 TeV
√s = 8 TeVDilepton Channel @ 8 TeV Physics Analysis Summary (PAS)
CMS-PAS-TOP-14-014
Separate analysis of dileptonic tt channel data @ 8 TeV 1D template method using mlb observable (defined below) to probe mtop:
-
Measured Result:
Top Quark pT b Fragmentation
(∆mtop = 0.66 GeV) [0.4%] (∆mtop = 0.62 GeV) [0.4%]
Dominant Sources of Uncertainty
Base signal MC employs:MADGRAPH 5.1.5.11 for ME generator MADSPIN for heavy resonance decays Pythia 6.426 for PS and hadronization
Require two opposite-sign leptons (specifically eµ) Forward detector folding of theoretical prediction to allow for comparison with measured data
angle between lepton-bin W rest frame
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 46
Pole Mass √s = 8 TeV
MC Mass √s = 7 TeVEur.Phys.J. C74 (2014) 3109Pole Mass from Cross-Section at √s = 7 & 8 TeV
Measured Result:
Predominantly a high-precision cross-section calculation measurement Pole mass (mtop) can simultaneously be extracted from dependence on XS Work specifically in the eµ channel (oppositely charged leptons)
where
}Solve for&
Allows for solution of system of equations below Simultaneous measurements:
Look @ Number of Data Events with 1 or 2 b-tagged Jets (N1 & N2, respectively)
Inclusive cross-section ( ) Efficiency to reconstruct and tag a b-quark jet ( )
Luminosity Parton Distribution Functions (PDFs)
Dominant Uncertainties (Largely Theoretical):Uncertainties on 𝛼s
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 47
MC Mass
Pole MassPole Mass from Cross-Section at √s = 7 & 8 TeV
Measurements of σtt together with NNLO theoretical prediction allow for the extraction of the top quark pole mass (mt) Relies on a given choice of PDF sets and 𝛼s (quoted result uses NNPDF3.0, 𝛼s = 0.118 ± 0.001)
Consistent results using CT14 and MMHT2014 Parameterize functional dependence of σtt on mt as exponential function
Maximization of final likelihood yields measured value (after 7/8 TeV combination):
Submitted to JHEP
√s = 8 TeV
√s = 7 TeV
Precision cross-section measurements performed at √s = 7 & 8:
Great improvement previous measurement:
Above: Measured values (black markers) with the vertical bars indicating the total 1-σ uncertainty. The theoretical predictions are shown as the coloured bands.
Three points above correspond to different mt hypotheses(for which σtt fit is repeated) ∆mt/mt ~ 1%
Top quark pole mass precision of
arXiv:1603.02303
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 48
MC Mass
√s = 8 TeVPole Mass
√s = 7 TeVPhys.Lett.B. 738 (2014) 526Pole Mass from Cross-Section at √s = 7 TeV
Measured Result:
!Measured value taken from dileptonic (eµ) channel result (most precise) Result gives a Baysian contour interval on mtop under external constraint of 𝛼s
Baysian analysis to determine a value of mtop (pole mass) based on measured and theoretical σtt
Very good precision on 𝛼s:
Sensitivity to Uncertainty on 𝛼s
But even shifting by this amount greatly affects ∆mtop:
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 49
Alternate Measurements
Advantages:Can take advantage of ‘tag and probe’ method of lepton + jets channel (relatively low backgrounds) Complimentary to direct top quark mass measurements as there no distinction is made between top/anti-top
Very sensitive to detector’s b- vs. b-jet response Differences in W- / W+ production also significant systematic uncertainty
Disadvantages:
Top/Anti-top Quark Mass Difference
Top Quark Mass Difference
t W +
bb_
t_
W _
qq_
+mt-
mt
_
Phys. Lett. B. 728C (2014)
JHEP 06 (2012) I 09
7 TeV
7 TeV
semileptonic (e/µ) ∆mtop
semileptonic (e/µ) ∆mtop
= Covered in this Talk= See Backup
e + jets µ + jets
e + jets µ + jets
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 50
√s = 7 TeVPhys. Lett. B. 728C (2014)Pole Mass √s = 8 TeV
MC MassTop/Anti-top Quark Mass Difference √s = 7 TeV
∆mtop = 0.67 GeV ± 0.61 (stat) ± 0.41 (syst) GeV
Measured Result:
Result Consistent with Standard Model
Predictions and CPT invariance
Require two b-tagged jets Event reconstruction: kinematic fit ∆mm calculated on per-event basis Maximum likelihood fit to final distribution
fit
}e/µ charge from
leptonic top quark decay
Semileptonic channel with 1 hadronically / 1 leptonically decaying top quark
∆m [GeV]fit
Choice of b fragmentation model ➙ Different Responses to b/b-Jets
Dominant Uncertainty:
_
Data statistics
T.G.McCarthy - Top Quark Mass @ ATLAS & CMS Rencontres de Moriond QCD Session, March 20, 2016 51
√s = 7 TeV
√s = 8 TeV
MC Mass
Pole MassTop/Anti-top Quark Mass Difference √s = 7 TeV
Performed separately for e+jets and µ+jets channels Final likelihood-based fit performed separately for each distribution (as shown above) Results show ∆mt = |mtop - manti-top| consistent with 0 as required by CPT theorem
Average mt also consistent with other measurements
e+jets:µ+jets:
combined:
∆mt = -0.44 GeV ± 0.46 (stat)
± 0.27 (syst) GeV
Measured Result:
JHEP 06 (2012) I 09
top related