Making Precision Measurements at Hadron Collidershep.uchicago.edu/~frisch/talks/lakelouise.pdf · Two Lectures on Making Precision Measurements at Hadron ... Lake Louise Winter Institute

Two Lectures on Making Precision Measurements at Hadron Colliders

Making Precision Measurements at Hadron CollidersHenry Frisch

University of Chicago

Lake Louise Winter Institute, Feb. 17-23, 2006

Contents

1 Lecture I: The Electroweak Scale: Top, the W and Z, and the Higgs via MW and Mtop 3

2 Purpose 3

3 Some History and Cultural Background 43.1 Instrumental Sensitivity: Orders of Magnitude . . . . . . . . . . . . . . . . . . . . . . . . . 43.2 Hubris: The 50 GeV Top Quark and No Quarkonia . . . . . . . . . . . . . . . . . . . . . . 6

4 The Tevatron and the LHC 7

5 The Anatomy of Detectors at Hadron Collider: Basics 85.1 Basics: Kinematics and Coverage: pT vs P|| . . . . . . . . . . . . . . . . . . . . . . . . . . 85.2 Basics: Particle Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

6 Calibration Techniques 126.1 Momentum and Energy Scales: E/p . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126.2 Higher-order momentum and energy corrections . . . . . . . . . . . . . . . . . . . . . . . . 14

7 W and Z0Production as Archetypes 14

8 ‘QCD’- Jet Production, Quark and Gluons, ISR, FSR 19

9 The MTop −MW Plane and the Higgs Mass 249.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249.2 What limits the precision on the W mass and the top mass measurements? . . . . . . . . . 24

10 Measuring the Top Quark Mass and Cross-section 2710.1 tt Production: Measuring the Top Cross-section Precisely . . . . . . . . . . . . . . . . . . . 2710.2 Total Cross-section for tt Production: Parsing the CDF and DØ Summary Plots . . . . . . 2910.3 Properties of the tt system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3210.4 Top Decays: ‘Lepton+Jets’, ‘Dileptons’, ‘All-Hadronic’ . . . . . . . . . . . . . . . . . . . . 3410.5 Precision Measurement of the Top Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

11 Lecture II: Searching for Physics Beyond the SM, and Some Challenges for the Audi-ence 3511.1 Strategies: Signature-Based vs Model-Directed, Blind vs A Priori vs Myopic, etc. . . . . . . 3711.2 Lepton+Gamma+X: The `γ 6Et and ``γ Signatures . . . . . . . . . . . . . . . . . . . . . . . 39

HJF Lake Louise Winter Institute Feb. 17-23, 2006


11.3 Gamma+Gamma+X: The ``γ Signature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4411.4 Inclusive High Pt W’s and Z’s: A Weak Boson Signature . . . . . . . . . . . . . . . . . . . 4611.5 The Tail of the W: Above the Pole- Wprimes . . . . . . . . . . . . . . . . . . . . . . . . . . 4911.6 An Indirect Search: Asymmetries above the Pole (CDF+D0) . . . . . . . . . . . . . . . . . 4911.7 A Classic SUSY Search: Trileptons at D0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4911.8 A Classic SUSY Search: Met + Jets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

12 Direct Search for the Higgs 49

13 Expert Topics: Black and Double Black: Challenges for Students 5113.1 Fragmentation Near z = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5113.2 Photon and Tau Fake Rates: Gluon and Quark Jets . . . . . . . . . . . . . . . . . . . . . . 5113.3 Monte Carlo Issues: QCD and QED, NLO and beyond . . . . . . . . . . . . . . . . . . . . 5113.4 B-jet Momentum Scale: Gamma-bjet Balancing . . . . . . . . . . . . . . . . . . . . . . . . 52

14 Beyond Expert: Out-of Boundary Area Topics: Challenges for Expert Groups 5314.1 Book-keeping: Rethinking Luminosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5314.2 Rethinking Analysis Code and Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5314.3 Changing the Paradigm: W/Z ratios, Color Singlet/Color Triplet Ratios, and Other New

Precision Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5314.4 Particle ID: Distinguishing W → cs from W → ud, bb from b in Top Decays . . . . . . . . . 5314.5 Discrete Symmetry Tests: C, CP, and T above the W and Z poles . . . . . . . . . . . . . . 53

15 Credits 54



1 Lecture I: The Electroweak Scale: Top, the W and Z, and

the Higgs via MW and Mtop

2 Purpose

These two lectures are purely pedagogical. My intent is to enable non-experts

to get something out of the individual presentations on collider physics that

will follow- the Higgs, the W,Z, top, searches for SUSY, LED’s, etc. We are

presented with so many measurements and so much detail that we often forget

that we are talking about instruments and the measurements they have made.

The suprise is how precise the detectors themselves are; the challenge will be to

exploit that precision in the regime where statistics is no longer a problem, and

everything is dominated by the performance of the detector (‘systematics’).

This challenge also extends to the theoretical community- to look for

something new we will need to understand the non-new, i.e. the SM predictions,

at an unprecedented level of precision. Some amount of this can be done with

control samples- it is always best to use data rather than Monte Carlo, but

it’s not always possible. The detectors are already better than the theoretical

predictions- the theory community needs to catch up.

I work on CDF, and have used mostly CDF plots just because I know

them. No slight to DØ or the LHC experiments is meant. I have cut some

corners in places and been a little provocative in others, as teachers will. All

views presented here are my own.

I have intentionally used older public results from CDF and DØ instead

of the hot-off-the-press results generated for the 2006 ‘winter conferences’ so as

not to steal the thunder of the invited speakers who are here to present new

results from CDF and D0. The idea is to provide the understanding so that

you can ask them the hard questions, and to provoke discussion. This is going

to be really different from a raporteur’s talk...



Figure 1: A history of high-energy (no ISR) hadron colliders: integrated luminosity by year.

3 Some History and Cultural Background

3.1 Instrumental Sensitivity: Orders of Magnitude

A brief history of luminosity, starting with the SPPS and the race to discover

the W and the Z0, and then the race to discover the top, is shown in Figure 1.

At the Tevatron we feel we are in a race now to discover whatever is next-

for objects in the several hundred GeV mass range it’s all in accumulating

luminosity.



’88: Inverse Nanobarns ’06: Inverse Femtobarns

Figure 2: The integrated luminosity in the 1987 Tevatron run (Left), in Inverse Nanobarns, and in Run II(Right), in Inverse Femtobarns. Note that 1 fb−1= 103 pb−1= 106 nb−1. Note also the efficiency to tapehas improved substantially.

Figure 2 shows the luminosity ‘delivered’ and ‘to tape’ from the current Run

II, in inverse femtobarns (right), and from the 1987 run, in inverse nanobarns.

As a quick reminder, the W± → e±ν cross-section times BR is about 2.2 nb

for the left-hand plot, so 30 nb−1means that 66 W± → e±ν decays were cre-

ated in the recorded exposure. The cross-section for a 115 GeV Higgs in

W± → e±ν + H production is xxx fb, and so the right hand plot indicates

that xxx W± → e±ν +H events were created.



Figure 3: Left: The 1984 Top ‘discovery’; Right: The 1974 ‘no discovery’ announcement of the J/ψ andUpsilons.

3.2 Hubris: The 50 GeV Top Quark and No Quarkonia

Figure 3 is an historical reminder both that we should not be over-confident

about what we know, and that Nature has a rich menu of surprises. The left-

hand page is the discovery of something that did not exist- a top quark with

mass less than 50 GeV (it was largely W+jets, as shown by Steve Ellis). The

right-hand page is a prediction that there are no narrow states with masses

between 3 and 10 GeV decaying into lepton pairs (note both these guys did

well- Nature gave more chances!).



4 The Tevatron and the LHC

By now everybody should know about the Tevatron and LHC. I will spare

you pictures and boilerplate; The main differences that everybody, including

theorists, should know are:

Tevatron LHC

Parton Source Antiproton-Proton Proton-proton

Energy (TeV) 1.96 (not 2!) 14

Peak Luminosity (cm−2s−1) 2× 1032 1× 1034

Crossing Spacing (ns) 396 24.95

Peak Interactions/Crossing 5 19

Luminous Line σ (cm) 30 4.5 [3]

Luminosity Lifetime (hours) 3.8/23 [4] 15

< x > at MW 0.04 0.006

< x > at 2MT 0.18 0.025

An LHC upgrade to 1× 1035 is planned.

Figure 4: The CTEQ6.1M PDF’s at Q=100 (Joey Huston).



5 The Anatomy of Detectors at Hadron Collider: Basics

I start with a brief elementary introduction. Working at a hadron collider is

really different from at an e+e− machine!

5.1 Basics: Kinematics and Coverage: pT vs P||

The phase space for particle production at a hadron collider is traditionally

described in cylindrical coordinates with the z axis along the beam direction,

the radial direction called ‘transverse’, as in ‘Transverse Momentum’ (pT), and

the polar angle expressed as Pseudo-rapidity η, where η ≡ −ln(tanθ/2)).

Pseudo-rapidity is a substitute for the Lorentz-boost variable, y, where y ≡1/2ln(E + pz)/(E − pz) ≡ tanh−1(pz/E). Since in most cases one does not

know the mass of a particle produced in a hadron collision (most are light- pions,

kaons, baryons,..), we use pseudo-rapidity. (This is a common trap when doing

complex kinematics with W’s, Z’s, and top, where the mass truly matters).

Figure 5 shows an early sketch of the proposed coverage in η for CDF; note that

the big central detector seems very small, while the little luminosity monitors

seem big.



Figure 5: An early planning document (Hans Jensen) for the coverage in rapitidy for CDF

Two simple equations contain much of the physics for the production of heavy

states at a collider: the mass and longitudinal momentum of the heavy state

(e.g. a W, Z, tt pair, or WH) are determined by the fraction of the beam

momentum carried by the interacting partons. Note that for a heavy object

typically has a velocity β << 1, even though the longitudinal momentum is

typically not small (we’re not in the c.m! of the collision.). Note also that

the transverse momentum of the system is determined by the competition of

falling parton distribution functions (PDF’s- also known as structure functions)

as the total invariant mass of the system rises, and the increase in phase space

as the momentum of the system increases. The production thus peaks with a

total system energy above threshold by an amount characteristic of the slope

in x1 ∗ x2.

m2 = x1 ∗ x2s pz = (x1 − x2)pbeam (1)



5.2 Basics: Particle Detection

Here I deal with high-momentum particle detection. Low-momentum– typically

up to a few GeV– charged particles can be identified by processes that depend

on their velocity, β, as a simultaneous measurement of p = βγm and β allows

extracting the mass. However for momenta above a few GeV pions, kaons, and

protons cannot be separated. However electrons, muons, hadrons, and neutrinos

interact differently, as shown in Figure 5.2. The measurement of their energies

and/or momenta stem from their different modes of interaction.



(a) Identifying a high ET-electron

(b) Identifying a high pT-muon

(c) Identifying a jet (d) Identifying a neutrino



6 Calibration Techniques

6.1 Momentum and Energy Scales: E/p

The Tevatron and the LHC are as different from LEP and other e+e− colliders

as night and day- it is a big disadvantage to have worked at LEP(!). One

key difference is that the overall mass (energy) scale is not set by the beam

energy- there is a continuum of c.m. energies in the parton-parton collisions.

Moreover the hard scattering is not at rest either longitudinally nor transverse

in the lab system- there is ‘intrinsic Kt’ as well as initial-state radiation (ISR).

Finally, the beam spot is a line and not a spot- the vertex point, used to

calculate transverse energies, has to be determined from the event, including

for neutrinos and photons for which no track is observed.

Dealing first with the issue of setting the scale for momentum, energy, and

mass measurements. All current detectors consist of a magnetic spectrometer

followed by calorimeters.

The magnetic spectrometer uses a precisely measured (NMR) magnetic field

and the precise geometry of the tracking chambers to measure the curvature

(1/PT )of the tracks of charged particles. This is an absolute measurement- if

perfect one has the momentum scale. One can then use particles with mea-

sured momentum as an in situ ‘test beam’ to calibrate the energy scale of the

calorimeters.

The momentum scale can be checked by measuring the masses of some calibra-

tion ‘lines’ thoughtfully provided by Mother Nature- the J/Psi and Υ systems,

and the Z0in its Z0 → µ+µ− decays (Z0 → e+e− doesn’t work for momentum

calibration!). Fig. 6 shows measured distributions from CDF. However the mo-

mentum scale can be incorrect due to mis-alignments in the tracking chamber.

The combination of a calorimeter and a magnetic spectrometer allows one to

remove the 1st-order errors in both [5] by measuring ‘E’ (calorimeter energy)

over ‘p’ (spectrometer moementum. With perfect resolution, no energy loss,



Figure 6: Left: The reconstructed JΨ invariant mass in dimuons (CDF). Right: The similar plot for theUpsilon system.

and no radiation these two should be equal: E/p = 1.0. Figure 7 shows the

measured spectrum in E/p for electrons.

The 1st-order error in momentum is due to a ‘false-curvature’- that is that

a straight line (zero-curvature= ∞ momentum) is reconstructed with a finite

momentum. The 1st-order error in calorimeter energy is an offset in the energy

scale, and does not depend on the sign (±) of the particle [6]. Expanding both

the curvature and calorimeter energies to first order:

1/p = 1/ptrue + 1/pfalse (µ+) 1/p = 1/ptrue − 1/pfalse (µ−) (2)

E = Etrue ∗ (1 + ε) (e+) E = Etrue ∗ (1− ε) (e−) (3)

The first-order false curvature pfalse then is derived by measuring E/p for pos-

itive and negative electrons with the same E

1/pfalse = ((E/p(e+)− E/p(e−))/2E (4)

The first-order calibration scale error ε then is removed by setting the calorime-

ter scale for electrons so that E/p agrees with expectations. In CDF, this is

done initially to make the calorimeter response uniform in φ− η.

1/pfalse = ((E/p(e+) + E/p(e−))/2 (5)



Figure 7:

6.2 Higher-order momentum and energy corrections

The momentum and energy calibrations at this point are good enough for ev-

erything at present exposures except the W mass measurement. There are three

higher-order effects that are taken care of at present:

1. ‘Twist’ between the 2 end-plates of the tracking chamber;

2. Systematic scale change in the z-measurements in the chamber;

3. Non-linearity of the calorimeter due to e(E/2) + γ(E/2) 6= e(E)

Figure 8 shows the use of the J/Ψ mass to correct for the first two of these

effects. What is plotted is the correction to the momentum scale versus the

cotan of the difference in polar (from the beam axis) angle of the two muons.

There is a linear correction to the curvature of δc = 6×10−7cot(θ) that corrects

for the twist between the endplates, and a change in the scale of the z-coordinate

by 2 parts in 104, zscale = 0.9998 ± 0.0001. This is precision tuning of a large

but exceptionally precise instrument!

7 W and Z0Production as Archetypes

Let us consider the production of the W and Z0vector bosons as archetypes of

hard processes. Figure 10 shows the dominant diagram and a ‘cartoon’ of the



Figure 8: Left: The correction to the momentum scale versus the cotan of the difference in polar angleof the two muons in J/psi decay before corrections: Right: The same after correcting the curvature byδc = 6× 10−7cot(θ) the scale of the z-coordinate by 2 parts in 104.

Figure 9: Measuring a higher-order correction to track curvature: the calorimeter to momentum ratio E/pversus cotθ for e+ and e−, before and after the curvature and z-scale corrections.

production process. Both the W and Z are observed in their leptonic decays

W± → l±ν and Z0 → ``. W and Z production thus provide a precise measure

of the up and down quark parton distribution functions (PDF’s). Since we

measure W’s and Z’s in their leptonic modes, the kinematics of the decay also

matter. Consider the W’s: they are polarized, as the u and d quarks are light

and couple through V-A so quarks have helicity -1 and antiquarks +1. The W

decays also by V-A, so the charged leptons come out opposite to the helicity

direction. However, the dominant effect, at least at the Tevatron, is that the



Figure 10: The dominant diagrams and a ‘cartoon’ of the production process for W and Z production.

W is moving in the rest frame, and since the (valence) u quark momentum is

generally higher than the (sea) d anti-quark; W+ go in the proton direction,

and W− in the p direction (the LHC, being proton-proton, doesn’t have this

useful asymmetry).

Figure 11 shows the distribution in the difference of e+ and e− versus η (pseudo-

rapidity) of the electron (e±) measured by CDF. The left-hand plot shows the

full range as well as the experimental uncertainty band; the right-hand plot



Figure 11: Left: The forward-backward charge asymmetry in W± → e±ν decays plotted versus pseudo-rapidity. The blue error band gives the experimental uncertainty; also shown is the prediction using theCTEQ5L parton distribution functions. Right: The same data, folded around zero in η (remember this ispp), compared to a prediction using the RESBOS MC generator and the CTEQ6.1M PDF’s.

shows a comparison with the predictions using the CTEQ6 PDF’s. One can

see that the PDF’s do not fit well, and so we are learning about the u and d

quark distributions from the W asymmetry.

The W and Z longitudinal momenta are determined by the structure functions;

the transverse momenta are determined by initial state radiation off of the in-

coming quarks (radiation off of the outgoing remnants is suppressed). Figure 12



Figure 12: Left: Right:



Figure 13: Left: Right:

8 ‘QCD’- Jet Production, Quark and Gluons, ISR, FSR

The dominant feature in the hadron collider landscape is the production of jets-

the hard scattering of partons. Figure 13 reproduces two pages from a seminal

paper in 1971, when the idea of partons was brand new, by Berman, Bjorken,

and Kogut, pointing out that the existence of partons would lead to point-like

scatterings and hence high pT phenomena, including ‘cores’ (jets).



Figure 14:



Figure 15:



Figure 16:



Figure 17:



9 The MTop −MW Plane and the Higgs Mass

9.1 Motivation

The top quark is remarkable for its physics and useful as a tool for calibration.

It may also be a window into the world of heavy weakly-interacting particles

(such as a Higgs of one sort or another) in that it is produced strongly (i.e

with coupling Oαs) in pairs, but due to its strongly-conserved flavor quantum

number (top-ness), has to decay electro-weakly. Due to radiative corrections,

the masses of the W, Z, Higgs, and top quark are related in the SM; precise

measurements of the W and top quark masses determine the predicted Higgs

mass.

Figure 18: Left: The MW vs MT plane as of March 1998. Right: The MW vs MT plane as of the summerof 2005. Note the difference in the scales of the abscissas.

9.2 What limits the precision on the W mass and the top mass measure-ments?

Figure 20 gives the history of the uncertainty on the W mass as a function of

the square-root of luminosity. The statistical uncertainty is expected to scale



Figure 19: Left: The measured allowed region at 68% (1σ) in the MW −Mtop plane (the intersection ofinside the blue and solid-red contours), and the predicted dependence of the MW and Mtop on the SMHiggs mass. Right: The fit for the mass of the SM Higgs, showing the region excluded at 68% C.L.

as∫Ldt−1. The systematic uncertainties will be discussed below when we get

to the measurement of the W mass; however it is interesting to note that since

the systematics are studied with data, they also seem to scale with luminosity.

If the control of systematic uncertainties continues to scale with statistics as∫Ldt−1 the Tevatron can do as well as LHC projections [9], and with very

different systematics.



Figure 20: The total uncertainty on the W mass as measured at the Tevatron, versus integrated luminosity.If the control of systematic uncertainties continues to scale with statistics as

∫Ldt−1 the Tevatron can do

as well as LHC projections, and with different systematics.



Figure 21:

10 Measuring the Top Quark Mass and Cross-section

I will discuss two specific measurements as pedagogic examples of some specific

difficulties (challenges is the polite word) of doing precision measurements - the

measurements of the top cross-section and the top mass. The idea is make

it possible for you to ask really hard questions when you see these standard

busy-busy plots that speakers expect you to just let go by. First some basics.

10.1 tt Production: Measuring the Top Cross-section Precisely

The prime motivation for a precise measurement of the top cross-section is

that new physics could provide an additional source for the production (leading

to a larger cross-section than expected) or additional decay channels (leading



to a smaller measured cross-section into b) [8]. More prosaically, the cross-

section is a well-defined and in-principle easy-to-measure quantity that tests

many aspects of QCD and the underlying universe of hadron collider physics-

the PDF’s, LO, NLO and NNLO calculations, and provides a calibration point

for calorimeters and the energy scale (will be a key calibration for LHC). Lastly,

and less defensible scientifically, is the uneasy feeling that too low a cross-section

(e.g.) means that the top mass is really lighter than we measure, and the crucial

EWK fits and limits on the Higgs mass are thus probably not correct.

Figure 22 shows the dominant diagrams for top production. At the Tevatron

(left) the tt system, with a mass 400 GeV, samples the structure functions at

a typical x given by < x1x2 >= m2/√s =∼ (400/1960)2, giving < x > =

∼ 0.20,

well into the valence quark region. At the LHC, the corresponding value is

< x > =∼ 0.04, i.e. in a region dominated by gluons.

Figure 22: Left: The dominant diagram for tt production at the Tevatron; Right: The dominant diagramat the LHC. (from F. Maltoni [7]).



10.2 Total Cross-section for tt Production: Parsing the CDF and DØ Sum-mary Plots

Figure 23: From CDF: The measured and predicted top cross-sections versus mass with approximately200 pb−1(Left) and now with approximately 350 pb−1(Right).

‘200’ pb-1 ‘350’ pb-1



Figure 24: Left: The differential spectrum dN/dPT of t-quarks in tt production; Right:The differentialspectrum dN/dη, again for t-quarks.

‘200’ pb-1 ‘350’ pb-1



Figure 25: Left: The distribution in ∆φ between the t and the t in tt production; Right: the analogousdistribution in ∆η.



10.3 Properties of the tt system

The tt system is particularly interesting, as there may be new resonances decaying directly into tt or newpairs of particles each with a decay into top plus something. Either way there would be a feature in the ttmass spectrum and a change in shape in the tt pT spectrum. Figure26

Figure 26: Left: CDF’s ttbar mass spectrum from 320 pb−1. Right: The ttbar mass spectrum as measuredin 370 pb−1by DØ .



What is the probability for a lower-mass pair to be reconstructed at a higher

mass? Input a mean value for the pair, and look at the output. Abcissa runs

from 0 to 1200 GeV.

Figure 27: The output templates for an input ttbar pair mass; the abscissa runs between 0 and 1200 GeVin each plot (CDF).

450 GeV 500 GeV

550 GeV 600 GeV

650 GeV 700 GeV

750 GeV 800 GeV

850 GeV 900 GeV



10.4 Top Decays: ‘Lepton+Jets’, ‘Dileptons’, ‘All-Hadronic’

The expectation is that the Top quark decays t → W+ + b (t → W−b); i.e.

Vtd = 1. The experimental limit on Vtd is Vtd = xx± xx [].

10.5 Precision Measurement of the Top Mass



Figure 28:

11 Lecture II: Searching for Physics Beyond the SM, and Some

Challenges for the Audience

Our hope at the Tevatron is, of course, that we find something new before the

LHC. We had hints of new things in Run I:

1. the top dilepton sample looked odd (too many e-mu events, e-mu close in

phi, some odd kinematics;

2. The eeγγ 6Et event and the 2.8σ excess in ` + γ + 6Et

3. Perhaps more top to tau events than we deserved?



4. Top mass in dileptons was consistently lower than in lepton+jets



11.1 Strategies: Signature-Based vs Model-Directed, Blind vs A Priori vsMyopic, etc.

None of these was significant statistically- but made one want more data. We

now have 10-times the data! What to do: There are two major kinds of di-

rect searches, and in each three kinds of strategies have been followed (all this

categorization is arguable):

1. Signature-Based:

2. Model-Directed

Avoiding biases is important (see next slide)- two strategies are followed.

1. A Priori- use the same cuts as published in Run I, or in the 1st 1/3rd of

the data; then run on rest of data without changing anything (my favorite

for signature-based searches- look hard at your data!).

2. Blind- this is heavily used now-very useful and appropriate in some cases

(e.g. precision measurements: W mass, B lifetimes and masses, and classic

well-defined searches: B → µµ,...

A brief anecdote about a blind analysis around 1900:

There was a controversy over 2 conflicting measurements of a line in the

solar spectrum. The famous spectroscopist at Princeton asked his machin-

ist to rule a grating at a non-standard (blind!) lines/inch, and to put the

value in a sealed envelope. The Prof. then measured the line in terms

of an unknown dispersion, wrote a Phys Rev with an accompanying letter

that said ‘under separate cover you will receive the grating spacing from

my machinist, Mr. Smith; take this number, multiply it by my number,

put it in the blank space in the paper, and publish it’. Now, that’s blind.



Much as in the search for the W and Z, there is a defining energy scale for new

physics beyond the SM. In the case of the W, Fermi’s ‘Standard Model’ (i.e. ‘ef-

fective field theory’) of a 4-fermion interaction predicted that νe+e− → νe+e

−

scattering violated S-wave unitarity at a c.m. energy =∼ 300 GeV (see Commins

and Bucksbaum, Chapter 1.6, e.g.). For the SM, it’s more complicated (see,

e.g. Gunion et al. in the Higgs Hunter’s Guide), but the conclusion is the

same- there must be something new at the TeV scale. We experimentalists

are consequently primed to find something new at the Tevatron and/or LHC.

New means comparing data to precise predictions of the SM. Figure 29 shows

what can happen when eagerness combines with insufficiently understood SM

predictions.

Figure 29: An example of why the careful calculation of SM predictions is so crucial: the announcementof the ‘discovery’ of SUSY at the 1986 Aspen Conference. The right explanation (S. Ellis) turned out tobe a cocktail of SM processes, in particular W+jets and Z+jets.



11.2 Lepton+Gamma+X: The `γ 6Et and ``γ Signatures

One of the anomalies of Run I was the famous eeγγ 6Et event. This spawned the

advent of ‘signature-based’ searches at the Tevatron. In particular there were

two follow-ups: γγ +X (Toback) and `γ +X (Berryhill). The `γ +X search

resulted in a 2.7σ excess over SM expectations.

Figure 30:

The analysis is being repeated with exactly the same kinematic cuts so this

time it is a priori- (i.e. not self-selected to be interesting).



Andrei Loginov Search for Lepton-Photon-X Events

Photon-Electron Flow-Chart

Lepton-Photon Sample1 Lepton and 1 Photon

ET > 25 GeV508 Events

??

Exactly 1 LeptonExactly 1 Photon∆φlγ > 1506ET < 25

397 Events

?

?

Inclusive Multi-Body Events(All Other Photon-Lepton)

111 Events

??

?

Z-Like lepton-photon81 Gev < Meγ < 101 Gev(Background Calibration)

209 Events

Exactly 1 LeptonExactly 1 Photon∆φlγ < 1506ET < 25 GeV

67 Events

Two-Body Events188 Events

Multi-Body lγET

Events

6ET > 25 GeV

25 Events

Multi-Photon and

Multi-Lepton Events

0 and 19 Events, resp.

Figure 2: Photon-Electron Sample: the subsets of inclusive γl events analyzed

Exotics Meeting -8- July 14, 2005

Andrei Loginov Search for Lepton-Photon-X Events

Photon-Muon Flow-Chart

Lepton-Photon Sample1 Lepton and 1 Photon

ET > 25 GeV71 Events

??

Exactly 1 LeptonExactly 1 Photon∆φlγ > 1506ET < 2528 Events

?

?

Inclusive Multi-Body Events(All Other Photon-Lepton)

43 Events

??

?

Z-Like lepton-photon81 Gev < Meγ < 101 Gev(Background Calibration)

10 Events

Exactly 1 LeptonExactly 1 Photon∆φlγ < 1506ET < 25 GeV

13 Events

Two-Body Events18 Events

Multi-Body lγET

Events

6ET > 25 GeV

18 Events

Multi-Photon and

Multi-Lepton Events

0 and 12 Events, resp.

Figure 1: Photon-Muon Sample: the subsets of inclusive γl events analyzed in this paper

Exotics Meeting -7- July 14, 2005

Figure 31: Left: The flow of the `+ γ +X signature based search in electrons. Right:The flow in muons.



Lepton+Photon+ 6ET Predicted Events

SM Source eγ 6ET µγ 6ET (e+ µ)γ 6ET

W±γ 11.9 ± 2.0 9.0 ± 1.4 20.9 ± 2.8Z0/γ + γ 1.2 ± 0.3 4.2 ± 0.7 5.4 ± 1.0W±γγ, Z0/γ + γγ 0.14 ± 0.02 0.18 ± 0.02 0.32 ± 0.04(W±γ or W±)→ τγ 0.7 ± 0.2 0.3 ± 0.1 1.0 ± 0.2Jet faking γ 2.8 ± 2.8 1.6 ± 1.6 4.4 ± 4.4Z0/γ → e+e−, e→ γ 2.5 ± 0.2 - 2.5 ± 0.2Jets faking `+ 6ET 0.6 ± 0.1 < 0.1 0.6 ± 0.1

Total SM

Prediction 19.8 ± 3.2 15.3 ± 2.2 35.1 ± 5.3

Observed in Data 25 18 43

Multi-Lepton+Photon Predicted Events

SM Source eeγ µµγ llγZ0/γ + γ 12.5 ± 2.3 7.3 ± 1.7 19.8 ± 4.0Z0/γ + γγ 0.24 ± 0.03 0.12 ± 0.02 0.36 ± 0.04Z0/γ+ Jet faking γ 0.3 ± 0.3 0.2 ± 0.2 0.5 ± 0.5Jets faking `+ 6ET 0.5 ± 0.1 < 0.1 0.5 ± 0.1

Total SM

Prediction 13.6 ± 2.3 7.6 ± 1.7 21.2 ± 4.0

Observed in Data 19 12 31

Table 1: A comparison of the numbers of events predicted by the Standard Model and the observations for the `γ 6ET and ``γ searches.The SM predictions for the two searches are dominated by Wγ and Zγ production, respectively [?, ?, ?]. Other contributions comefrom the tri-boson processes Wγγ and Zγγ, leptonic τ decays, and misidentified leptons, photons, or 6ET.

1

Figure 32:



(GeV)TPhoton E20 40 60 80 100 120 140

Eve

nts

/10

GeV

0

5

10

15

20

25CDF Run II Preliminary

-1), 307 pbµ Data(e+TEγlγW

γZγe fake

γγ, Wγγ, QCD, ZγτW jet,

(a)

(GeV)TLepton E20 40 60 80 100 120 140 160 180 200

Eve

nts

/10

GeV

0

5

10

15

20

25(b)

(GeV)TE0 20 40 60 80 100 120 140

Eve

nts

/10

GeV

0

5

10

15

20

25(c)

) (GeV)γ, TE (l, TM0 50 100 150 200 250

Eve

nts

/15

GeV

02468

101214161820 (d)

(GeV)TPhoton E20 40 60 80 100 120 140

Eve

nts

/10

GeV

0

5

10

15

20

25 CDF Run II Preliminary-1), 307 pbµ Data(e+γll

γZ

γγZ jet, QCD, Z

(a)

(GeV)TLepton E20 40 60 80 100 120 140 160 180 200

Eve

nts

/10

GeV

0

5

10

15

20

25(b)

M (l, l) (GeV)0 50 100 150 200 250 300

Eve

nts

/10

GeV

0123456789 (c)

) (GeV)γM (l, l, 0 50 100 150 200 250 300 350 400

Eve

nts

/10

GeV

0

2

4

6

8

10 (d)

Figure 33:

(GeV)TE0 5 10 15 20 25 30 35

Eve

nts

/5 G

eV

0

2

4

6

8

10 CDF Run II Preliminary-1), 307 pbµData(

γZ

γγZ jet, Z

(a)

(GeV)TE0 5 10 15 20 25 30 35

Eve

nts

/5 G

eV

0

2

4

6

8

10 CDF Run II Preliminary-1Data(e), 307 pb

γZ

γγZ jet, QCD, Z

(b)

Figure 34:

No more ``γγ 6Et events with > 3 times the data and higher energy. Have

another factor of 3 in data ready.



However, this has proved another educational example of MC predictions being

the limiting factor in speed and sensitivity. We do not have a control sample-

depend on SM predictions, largely Wγ and Zγ. Have 2 MC generators- Mad-

Graph [?] and a program from UliBaur [?]. They agree beautifully.

Figure 35:

However after running them through Pythia they disagreed by 15% in yield,

including a different identification efficiency for muons (!). Problems were in the

interface (diagnosed by Loginov and Tsuno) for both- the Les Houches accord

format is not precisely defined. Lessons:

1. Always use 2 MC’s- you may find both samples are flawed.

2. CDF has lost huge amounts of time to the generator interfacing- needs

re-examination by the theoretical community.



11.3 Gamma+Gamma+X: The ``γ Signature

Figure 36:



Figure 37:



11.4 Inclusive High Pt W’s and Z’s: A Weak Boson Signature

Idea: Many models of new physics- Extra Dimensions [?], Z-primes, Excited

Top, t′ → Wb, SUSY, Right-handed Quarks ([?]) naturally give a signature of

a high-Pt EWK boson- W, Z, or photon. Natural in strong production of pairs-

if decays, decays weakly. E.g. top

PtWEntries 1026

Mean 135

RMS 64.82

Underflow 0

Overflow 0

0 50 100 150 200 250 300 350 4000

20

40

60

80

100

120

PtWEntries 1026

Mean 135

RMS 64.82

Underflow 0

Overflow 0

Transverse Momentum of the W

Figure 38: The pT spectrum for Z’s from the decay of a 300 GeV right-handed singlet down quark QQ→uWdZ in the Bjorken-Pakvasa-Tuan model.

Figure 39: The inclusive search for high-¶T Z+X production (CDF). The cuts are frozen on the first 0.3fb−1: the rest will then be a priori.



However the inclusive Z+X is dominated by SM Z+jets- we cannot yet predict

this at the level needed. Figure 40

Figure 40: Inclusive high pT Z production and 3 monte-carlo predictions, showing that we cannot yet apriori test the data against the SM (work in progress).



To increase sensitivity, add objects to the signature- subsignatures of Z+Njets,

Z + γ, Z + `,... For example: a Z with 200 GeV Pt balanced by a photon with

200 GeV Pt from Run I (100 pb−1):

Figure 41:

From Run II- Z+N(photons) (300 pb−1now- soon 1000).

# of Photons0 1 2 3 4 5 6 7 8 9 10

en

trie

s

-210

-110

1

10

210

310

410

# of Photons0 1 2 3 4 5 6 7 8 9 10

en

trie

s

-210

-110

1

10

210

310

410 Inclusive Z’s

# of Photons0 1 2 3 4 5 6 7 8 9 10

en

trie

s

-210

-110

1

10

210

310

# of Photons0 1 2 3 4 5 6 7 8 9 10

en

trie

s

-210

-110

1

10

210

310

(Z) > 60 GeVTP

# of Photons0 1 2 3 4 5 6 7 8 9 10

en

trie

s

-210

-110

1

10

# of Photons0 1 2 3 4 5 6 7 8 9 10

en

trie

s

-210

-110

1

10(Z) > 120 GeVTP

-µ+µ →Z

DATA

Z + jets

WZ

ZZ

WW

ττ →Z

tt

Figure 42:



11.5 The Tail of the W: Above the Pole- Wprimes

Figure 43: The predicted tail of the W way above the pole (from D0).

11.6 An Indirect Search: Asymmetries above the Pole (CDF+D0)

11.7 A Classic SUSY Search: Trileptons at D0

11.8 A Classic SUSY Search: Met + Jets

12 Direct Search for the Higgs

We saw in Lecture I that the EWK precision data favor a light Higgs (too light,

even). I briefly summarize the current status of Higgs searches:



Figure 44: The cross-section limits from direct searches for the Higgs as of Sept 05 from CDF and D0

Figure 45: The ratio of cross-section limits from direct searches to SM predictions for the Higgs as of Sept05 from CDF and D0



13 Expert Topics: Black and Double Black: Challenges for Stu-

dents

13.1 Fragmentation Near z = 1

13.2 Photon and Tau Fake Rates: Gluon and Quark Jets

13.3 Monte Carlo Issues: QCD and QED, NLO and beyond



13.4 B-jet Momentum Scale: Gamma-bjet Balancing

The response of the calorimeter to the b-quark jets from top decay is critical for

the top mass; sharpening the resolution is also critical for discovering the Higgs.

One source of b’s of known momentum is Z0 → bb; even at the Tevatron this

is very difficult as the rate of 2-jet production prohibits an unprescaled trigger

threshold well below MZ/2. At the LHC this will be hopeless, I predict. How-

ever the ‘Compton’ process gluonb → γb will give a photon opposite a b-jet.

Figure 46 shows the flux of b-quarks versus x at Q = 100 GeV (CTEQ6.1M);

one can see that at x=0.01 (pT = 70 GeV at the LHC) the b-quark flux is

predicted to be only a factor of 3 lower than the gluon flux.

Figure 46: The PDF’s at Q = 100 GeV (CTEQ6.1M) showing that the b-quark flux is only half that ofthe u flux (Plot from Joey Huston).



14 Beyond Expert: Out-of Boundary Area Topics: Challenges

for Expert Groups

14.1 Book-keeping: Rethinking Luminosity

14.2 Rethinking Analysis Code and Structure

14.3 Changing the Paradigm: W/Z ratios, Color Singlet/Color Triplet Ra-tios, and Other New Precision Tests

14.4 Particle ID: Distinguishing W → cs from W → ud, bb from b in TopDecays

14.5 Discrete Symmetry Tests: C, CP, and T above the W and Z poles



15 Credits



References

[1] review

[2] orig bbk

[3] LHC Design Report CERN-2004-003 (June 2004), Section 2. I have taken the 7.75 cm quoted for theRMS bunch length, multiplied by the geometric luminosity reduction factor of 0.836, and divided by√

2. I hope this is correct.

[4] The initial luminosity has a lifetime of 3.8 hours, which crosses the longer lifetime after 2 hours, atwhich point the luminosity is half the peak.

[5] I first learned of this method from Aseet Mukherjee and Barry Wicklund, who used it in the CDFearly precise (at that time) measurement of the Z0mass.

[6] J. D. Jackson and R. McCarthy; ”Z3 Corrections to Energy Loss and Range”, Phys. Rev. B6,4131(1972).

[7] Fabio Maltoni, Top Physics: Theoretical Issues and Aims at theTevatron and LHC, HCP2005, July 8, Les Diablerets, Switz.;http://indico.cern.ch/getFile.py/access?contribId=51&sessionId=13&resId=0&materialId=slides&confId=0512

[8] Kane Mrenna top prod and dec

[9] Both are projections- time will tell.

[10]

[11]

[12]


Documents

Making Precision Measurements at Hadron Collidershep.uchicago.edu/~frisch/talks/lakelouise.pdf · Two Lectures on Making Precision Measurements at Hadron ... Lake Louise Winter Institute