j. huston qcd backgrounds to new...
TRANSCRIPT
TeVU 8/29/02J. Huston
See http://www.pa.msu.edu/~huston/tevu/tevu.pdf
QCD Backgrounds to New Physics
J. Huston
…thanks to Weiming Yao, John Campbell, Bruce KnutesonNigel Glover for transparencies
TeVU 8/29/02J. Huston
but first some commercials
TeVU 8/29/02J. Huston
Run 2 Monte Carlo Workshop
Transparencies, videolinks to individual talksand links to programscan all be found athttp://www-theory.fnal.gov/runiimc/
orhttp://thpc20.fnal.gov/runiimc/
I’ll be referring to someof these programs inthe course of my talk
TeVU 8/29/02J. Huston
Les Houches
Two workshops on “Physics at TeVColliders” have been held so far, in1999 and 2001 (May 21-June 1)
Working groups on QCD/SM, Higgs,Beyond Standard Model
See web page:
http://wwwlapp.in2p3.fr/conferences/LesHouches/Houches2001/
especially for links to writeups from 1999and 2001
QCD 1999 writeup (hep-ph/0005114)is an excellent pedagogical reviewfor new students
QCD 2001 writeup (hep-ph/0204316)is a good treatment of the state ofthe art for pdfs, NLO calculations,Monte Carlos
Les Houches 2003 will have more ofa concentration on EW/top physics
TeVU 8/29/02J. Huston
Les Houches 2001 Writeups
The QCD/SM Working Group: SummaryReport
◆ hep-ph/0204316
The Higgs Working Group: Summary Report(2001)
◆ hep-ph/0203056
The Beyond the Standard Model WorkingGroup: Summary Report
◆ hep-ph/0204031
TeVU 8/29/02J. Huston
Les Houches 2001
200th bottle of wineconsumed at theworkshop
TeVU 8/29/02J. Huston
Other useful references (for pdfs)
LHC Guide to Parton Distributions and Cross Sections, J. Huston;http://www.pa.msu.edu/~huston/lhc/lhc_pdfnote.ps
The QCD and Standard Model Working Group: Summary Report from LesHouches; hep-ph/0005114
Parton Distributions Working Group, Tevatron Run 2 Workshop; hep-ph/0006300
A QCD Analysis of HERA and fixed target structure functiondata, M. Botje; hep-ph/9912439
Global fit to the charged leptons DIS data…, S. Alekhin; hep-ph/0011002
Walter Giele’s presentation to the QCD group on Jan. 12; http://www-cdf.fnal.gov/internal/physics/qcd/qcd99_internal_meetings.html
Uncertainties of Predictions from Parton Distributions I: the Lagrange MultiplierMethod, D. Stump, J. Huston et al.;hep-ph/0101051
Uncertainties of Predictions from Parton Distributions II: the Hessian Method, J.Pumplin, J. Huston et al.; hep-ph/0101032
Error Estimates on Parton Density Distributions, M. Botje; hep-ph/0110123
New Generation of Parton Distributions with Uncertainties from GlobalQCD Analysis (CTEQ6); hep-ph/0201195
TeVU 8/29/02J. Huston
Other references…
…to the ability to calculate QCD backgrounds …to the need to do so
TeVU 8/29/02J. Huston
Monojets in UA1
UA1 monojets (1983-1984)
◆ Possible signature of new physics(SUSY, etc)
◆ A number of backgrounds wereidentified, but each was noted asbeing too small to account for theobserved signal
▲ pp->Z + jets
|_ νν
▲ pp->W + jets
|_ τ + ν
|_ hadrons + ν
▲ pp->W + jets
|_ l + ν
▲ pp->W + jets
|_ τ + ν
|_ l + ν
jet
…but the sum was not
◆ “The sum of many small things is abig thing.” G. Altarelli
Can calculate from first principles orcalibrate to observed cross sectionsfor Z->e+e- and W->eν
Ellis, Kleiss, Stirling PL 167B, 1986.
TeVU 8/29/02J. Huston
Signatures of New Physics
Ws, jets, γs, b quarks, ET
…pretty much the same as signatures for SM physics
How do we find new physics? By showing that its not old physics.
◆ can be modifications to the rate of production
◆ …or modification to the kinematics, e.g.angular distributions
Crucial to understand the QCD dynamics and normalization of bothbackgrounds to any new physics and to the new physics itself
Some backgrounds can be measured in situ
◆ …but may still want to predict in advance, e.g. QCD backgrounds toH->γγ
For some backgrounds, need to rely on theoretical calculations, e.g. ttbbbackgrounds to ttH
TeVU 8/29/02J. Huston
Theoretical Predictions for New (Old) Physics
There are a variety of programs available forcomparison of data to theory and/or predictions.
◆ Tree level
◆ Leading log Monte Carlo
◆ NnLO
◆ Resummed
Important to know strengths/weaknesses of each.
In general, agree quite well…but beforeyou appeal to new physics, check theME. (for example using Comphep)Can have ME corrections to MC or MC corrections to ME. (in CDF->HERPRT)
Perhaps biggest effort…include NLO MEcorrections in Monte Carlo programs…correct normalizations. Correct shapes. NnLO needed for precision physics.
Resummed description describes soft gluoneffects (better than MC’s)…has correctnormalization (but need HO to get it); resummedpredictions include non-perturbative effectscorrectly…may have to be put in by hand in MC’sthreshold kT
W,Z, Higgsdijet, direct γ
b space(ResBos)
qt space
Where possible, normalize to existing data.…in addition, worry about pdf, fragmentation uncertainties
TeVU 8/29/02J. Huston
W + Jet(s) at the Tevatron
Good testing ground for parton showers,matrix elements, NLO
Background for new physics
◆ or old physics (top production)
Reasonable agreement for the leadingorder comparisons using VECBOS (butlarge scale dependence)
-1.5
-1
-0.5
0
0.5
1
0 10 20 30 40 50 60 70 80 90 100
Systematic uncertainties
Leading Jet ET (GeV)
dσ
(W +
≥1
jets
) / d
ET
(D
ata −
QC
D)
/ QC
D
CDF PRELIMINARY
CDF Data (108 pb-1)0.4 Jet Cones, |ηd| < 2.4
DYRAD NLO QCD Predictionswith jet smearingMRSA/
Qr2 = Qf
2 = (0.5 MW)2
Qr2 = Qf
2 = MW2
Qr2 = Qf
2 = (2.0 MW)2
Good agreement with NLO (and smaller scale dependence) for W + >= 1 jet
TeVU 8/29/02J. Huston
W + jets
For W + >=n jet production, typicallyuse Herwig (Herprt) for additionalgluon radiation and for hadronization
Can also start off with n-1 jets andgenerate additional jets using Herwig
10-1
1
10
10 2
10 3
0 20 40 60 80 100 120 140 160 180 2
(a)
(b)
(c)
+ CDF Preliminary Data (108 pb-1)
− VECBOS + HERWIG Fragmentation+ Detector Simulation
CTEQ3M, Q2 = < pT >2
(Normalized to data)
Jet ET (GeV)
Eve
nts
/ 5 G
eV
(a) ET2 (W + ≥2 Jets Data)
VECBOS W + 1 Jet
(b) ET3 (W + ≥3 Jets Data)
VECBOS W + 2 Jets
(c) ET4 (W + ≥4 Jets Data)
VECBOS W + 3 Jets
10-1
1
10
10 2
10 3
0 20 40 60 80 100 120 140 160 180 200
(a)
(b)
(c)
(d)
+ CDF Preliminary Data (108 pb-1)
− VECBOS + HERWIG Fragmentation+ Detector Simulation
CTEQ3M, Q2 = < pT >2
(Normalized to data)
Jet ET (GeV)
Eve
nts
/ 5 G
eV
(a) ET1 (W + ≥1 Jet)
(b) ET2 (W + ≥2 Jets)
(c) ET3 (W + ≥3 Jets)
(d) ET4 (W + ≥4 Jets)
TeVU 8/29/02J. Huston
More Comparisons (VECBOS and HERWIG)
Start with W + n jets from VECBOS
0
100
200
300
400
500
0 1 2 3 4 5
+ CDF PreliminaryData (108 pb-1)
− VECBOS + HERWIG+ Detector SimulationCTEQ3M, Q2 = < pT >(Normalized to data)
Eve
nts
/ 0.5
W + ≥2 Jets DataVECBOS W + 1 Jet
0
20
40
60
80
100
0 1 2 3 4 5∆Rjj of Leading Two Jets
Eve
nts
/ 0.5
W + ≥3 Jets DataVECBOS W + 2 Jet
0
100
200
300
400
500
0 1 2 3 4 5 6
+ CDF PreliminaryData (108 pb-1)
− VECBOS + HERWIG+ Detector SimulationCTEQ3M, Q2 = < pT >2
(Normalized to data)
Eve
nts
/ 0.5
W + ≥2 Jets
0
20
40
60
80
100
0 1 2 3 4 5 6∆Rjj of Leading Two Jets
Eve
nts
/ 0.5
W + ≥3 Jets
•Start with W + (n-1) jets from VECBOS
TeVU 8/29/02J. Huston
More Comparisons
Start with W + n jets from VECBOS Start with W + (n-1) jets fromVECBOS
1
10
10 2
10 3
0 50 100 150 200 250 300 350 400
+ CDF Preliminary Data (108 pb-1)− VECBOS + HERWIG Fragmentation
+ Detector SimulationCTEQ3M, Q2 = < pT >2
(Normalized to data)
Eve
nts
/ (10
GeV
/c2 )
W + ≥2 Jets DataVECBOS W + 1 Jet
10-1
1
10
10 2
0 50 100 150 200 250 300 350 400Mjj of Leading Two Jets (GeV/c2)
Eve
nts
/ (10
GeV
/c2 )
W + ≥3 Jets DataVECBOS W + 2 Jets
1
10
10 2
10 3
0 50 100 150 200 250 300 350 400
+ CDF Preliminary Data (108 pb-1)− VECBOS + HERWIG Fragmentation
+ Detector SimulationCTEQ3M, Q2 = < pT >2
(Normalized to data)
Eve
nts
/ (10
GeV
/c2 )
W + ≥2 Jets
10-1
1
10
10 2
0 50 100 150 200 250 300 350 400Mjj of Leading Two Jets (GeV/c2)
Eve
nts
/ (10
GeV
/c2 )
W + ≥3 Jets
TeVU 8/29/02J. Huston
When good Monte Carlos go bad
Consider W + jet(s) sample
Compare data (Run 0 CDF) toVECBOS+HERPRT (Herwigradiation+hadronization interfaceto VECBOS) normalized to W+Xjets
Starting with W + 1 jet rate indata, Herwig predicts 1 W+ >4 jetevents; in data observe 10
…factor of ~2 every jet
◆ very dependent on kinematicsituation, though
▲ jet ET cuts
▲ center-of-mass energy
▲ etc
#events >1 jet >2 jets >3 jets >4 jet
pT>10 GeV/c
Data 920 213 42 10
VECBOS + HERPRT (Q=<pT>)
W + 1jet 920 178 21 1
W + 2jet ----- 213 43 6
W + 3jet ----- ----- 42 10
VECBOS+HERPRT(Q=mW)
W + 1jet 920 176 24 2
W + 2jet ---- 213 46 6
W + 3 jet ---- ----- 42 7
TeVU 8/29/02J. Huston
Factors get worse at the LHC
factor of 50
TeVU 8/29/02J. Huston
Why?
Some reasons given by the experts (Mangano, Yuan,Ilyin…)◆ Herwig (any Monte Carlo) only has collinear part of matrix element for gluon
emission; underestimate for the wide angle emission that leads to widelyseparated jets
◆ phase space: Herwig has ordering in virtualities for gluon emission while thisis not present in exact matrix element calculations; more phase space forgluon emission in exact matrix element calculations
◆ in case of exact matrix element, there are interferences among all of thedifferent diagrams; these interferences become large when emissions takeplace at large angles (don’t know a priori whether interference is positive ornegative)
◆ unitarity of Herwig evolution: multijet events in Herwig will always be afraction of the 2 jet rate, since multijet events all start from the 2-jet hardprocess
▲ all “K-factors” from higher order are missed.
TeVU 8/29/02J. Huston
Tree Level Calculations
Leading order matrixelement calculationsdescribe multi-bodyconfigurations betterthan parton showers
Many programs existfor calculation of multi-body final states attree-level
◆ References: see Run 2MC workshop
CompHep
◆ includes SM Lagrangian and several othermodels, including MSSM
◆ deals with matrix elements squared
◆ calculates leading order 2->4-6 in the finalstate taking into account all of QCD andEW diagrams
◆ color flow information; interface exits toPythia
◆ great user interface
Grace
◆ similar to CompHep
Madgraph
◆ SM + MSSM
◆ deals with helicity amplitudes
◆ “unlimited” external particles (12?)
◆ color flow information
◆ not much user interfacing yet
Alpha + O’Mega
◆ does not use Feynman diagrams
◆ gg->10 g (5,348,843,500 diagrams)
TeVU 8/29/02J. Huston
Monte Carlo Interfaces
To obtain full predictability for atheoretical calculation, would liketo interface to a Monte Carloprogram (Herwig, Pythia, Isajet)
◆ parton showering (additional jets)
◆ hadronization
◆ detector simulation
Some interfaces already exist
◆ VECBOS->Herwig (HERPRT)
◆ CompHep->Pythia
A general interface accord wasreached at the 2001 LesHouches workshop
All of the matrix elementprograms mentioned will output4-vector and color flowinformation in such a way as tobe universally readable by allMonte Carlo programs
CompHep, Grace, Madgraph,Alpha, etc, etc
->Herwig, Pythia, Isajet
TeVU 8/29/02J. Huston
Les Houches and Monte Carlos
Much of the time during meetingwas spent developing a genericprocess interface from matrixelement to Monte Carloprograms
This interface allows:◆ arbitrary hard subprocesses to
be plugged intoshower/hadronizationgenerators.
CompHEP
Grace Herwig
MadGraph z Isajet
VecBos Pythia
Wbbgen
◆ ->Les Houches accord (#1)
“Les Houches” User ProcessInterface
for Event Generators
hep-ph/0109068
E. Boos, M. Dobbs, W. Giele, I. Hinchliffe, J. Huston,
V. Ilyin, J. Kanzaki, K. Kato, Y. Kurihara,
L. Lonnblad, M. Mangano, S. Mrenna, F. Paige, E. Richter-Was,
M. Seymour, T. Sjostrand, B. Webber, D. Zeppenfeld
Possible because one or more authors from
each of these programs was present
at Les Houches
◆ Matt Dobbs has been the front man for
coordinating the disputes/discussions
◆ literally hundreds of email exchanges
TeVU 8/29/02J. Huston
Universal Interface
This interface will allow for amore complete predictability forME programs
◆ parton showering (additional jets)
◆ hadronization
◆ detector simulation
Some specialized interfacesalready exist
◆ VECBOSzHerwig (HERPRT)
◆ WbbgenzHerwig
◆ CompHepzPythia
This interface should supercedethem.
Specialize in the ‘generic’ parts of the event.
f(x,Q2) f(x,Q2)PartonDistributions
HardSubProcess
PartonCascade
Hadronization
Decay
+Minimum BiasCollisions
TeVU 8/29/02J. Huston
Interface
Provides information on parton 4-vectors,mother-daughterrelationships, spins/helicities andcolor flow
◆ also points to intermediate particleswhose mass should be preserved inthe parton showering
Not intended as a replacement forHEPEVT
◆ addresses communication betweenevent generators only, not betweenevent generators and the outsideworld
Partonic information is in 2 Fortrancommon blocks
◆ run info
◆ specific event info
Interface Structure
<<Container for RUN related information>>common /HepRUP/
+paramter MAXPUP: integer = 100+IDBMUP(2): integer+EBMUP(2): double+PDFGUP(2): integer+PDFSUP(2): integer+IDWTUP: integer+NPRUP: integer+XSECUP(MAXPUP): double+XERRUP(MAXPUP): double+XMAXUP(MAXPUP): double+LPRUP(MAXPUP): integer
<<Container for EVENT related information>>common /HepEUP/
+parameter MAXNUP: integer = 500, max num particle entries+NUP: integer = number entries this event+IDPRUP: integer = process id+XWGTUP: double = event weight+SCALUP: double = scale [GeV]+AQEDUP: double = QED coupling for this event+AQCDUP: double = QCD coupling for this event+IDUP(MAXNUP): integer = particle id+ISTUP(MAXNUP): integer = particle status+MOTHUP(2,MAXNUP): integer = pointer to parents+ICOLUP(2,MAXNUP): integer = particle (anit)color indices+PUP(5,MAXNUP): double = particle momentum, energy, mass+VTIMUP(MAXNUP): double = particle invariant lifetime+SPINUP(MAXNUP): double = spin vector angle (usually +1,-1)
<<called by SHG to for HepRUP info>>subroutine UPINIT()
<<called by SHG for HepEUP info>>subroutine UPEVNT()
integer MAXPUP parameter ( MAXPUP=100 ) integer IDBMUP, PDFGUP,PDFSUP, IDWUP, NPRUP, LPRUP double precision EBMUP,XSECUP, XERRUP, XMAXUP common /HEPRUP/ IDBMUP(2), EBMUP(2), PDFGUP(2),PDFSUP(2), + IDWTUP, NPRUP, XSECUP(MAXPUP), XERRUP(MAXNUP), + XMAXUP(MAXNUP), LPRUP(MAXPUP)
(Specialized for each matrix element)
TeVU 8/29/02J. Huston
Subroutines
Each stage (run and event)associated with own subroutine,called from the showergenerator, where information isplaced in the respective commonblock, based on output from thematrix element generator
Subroutine names (in Pythia 6.2)are:
◆ UPINIT
◆ UPEVNT
◆ note no PY prefixes
Other authors should use thesame convention
Interface Structure
<<Container for RUN related information>>common /HepRUP/
+paramter MAXPUP: integer = 100+IDBMUP(2): integer+EBMUP(2): double+PDFGUP(2): integer+PDFSUP(2): integer+IDWTUP: integer+NPRUP: integer+XSECUP(MAXPUP): double+XERRUP(MAXPUP): double+XMAXUP(MAXPUP): double+LPRUP(MAXPUP): integer
<<Container for EVENT related information>>common /HepEUP/
+parameter MAXNUP: integer = 500, max num particle entries+NUP: integer = number entries this event+IDPRUP: integer = process id+XWGTUP: double = event weight+SCALUP: double = scale [GeV]+AQEDUP: double = QED coupling for this event+AQCDUP: double = QCD coupling for this event+IDUP(MAXNUP): integer = particle id+ISTUP(MAXNUP): integer = particle status+MOTHUP(2,MAXNUP): integer = pointer to parents+ICOLUP(2,MAXNUP): integer = particle (anit)color indices+PUP(5,MAXNUP): double = particle momentum, energy, mass+VTIMUP(MAXNUP): double = particle invariant lifetime+SPINUP(MAXNUP): double = spin vector angle (usually +1,-1)
<<called by SHG to for HepRUP info>>subroutine UPINIT()
<<called by SHG for HepEUP info>>subroutine UPEVNT()
integer MAXPUP parameter ( MAXPUP=100 ) integer IDBMUP, PDFGUP,PDFSUP, IDWUP, NPRUP, LPRUP double precision EBMUP,XSECUP, XERRUP, XMAXUP common /HEPRUP/ IDBMUP(2), EBMUP(2), PDFGUP(2),PDFSUP(2), + IDWTUP, NPRUP, XSECUP(MAXPUP), XERRUP(MAXNUP), + XMAXUP(MAXNUP), LPRUP(MAXPUP)
(Specialized for each matrix element)
TeVU 8/29/02J. Huston
Interface Structure
integer MAXPUP parameter ( MAXPUP=100 ) integer IDBMUP, PDFGUP,PDFSUP, IDWUP, NPRUP, LPRUP double precision EBMUP,XSECUP, XERRUP, XMAXUP common /HEPRUP/ IDBMUP(2), EBMUP(2), PDFGUP(2),PDFSUP(2), + IDWTUP, NPRUP, XSECUP(MAXPUP), XERRUP(MAXNUP), + XMAXUP(MAXNUP), LPRUP(MAXPUP)
Unweighting
Shower generator can unweightevents from matrix elementgenerator, mix differentsubprocesses from matrix elementgenerator, or just read eventsstraight from a file
◆ if unweighting/mixing is needed thenshower generator needs info aboutsubprocess cross sections and/ormaximum weights
If extra information is needed forspecific user implementation, thenimplementation-specific commonblock has to be created
Note that a lot of the technicalitiesare intended for ME/MC authors, notfor users; in most cases, thesedetails will be invisible to the casualuser
MAXUP: maximum number of different processes to be interfaced at one time
TeVU 8/29/02J. Huston
Run related information
Each stage (run and eventassociated with own subroutine)
Run subroutine◆ IDWTUP: master switch indicating
how the event weights (XWGTUP)are interpreted (some examplesbelow)
▲ +1: events are weighted on inputand SHG is asked to produceevents with weight +1 on output
▲ -1: same as above but eventweights may be either positive ornegative; SHG will produceevents with weights +1 or -1 onoutput
▲ +3: events are unweighted oninput so SHG only asks for nextevent
▲ -3: same as above but eventweights may be either +1 or -1
<<Container for RUN related information>>common /HepRUP/
+paramter MAXPUP: integer = 100+IDBMUP(2): integer+EBMUP(2): double+PDFGUP(2): integer+PDFSUP(2): integer+IDWTUP: integer+NPRUP: integer+XSECUP(MAXPUP): double+XERRUP(MAXPUP): double+XMAXUP(MAXPUP): double+LPRUP(MAXPUP): integer
<<called by SHG to for HepRUP info>>subroutine UPINIT()
(Specialized for
TeVU 8/29/02J. Huston
Event related information
NUP: number of particle entries for thisevent
IDPRUP: ID of the process for this event
XWGTUP: event weight
IDUP: particle ID (non-physical particlesassigned IDUP=0)
ISTUP: status code
◆ -1: incoming particle
◆ +1: outgoing particle
◆ -2: intermediate space-like propagatordefining an x and Q2 which should bepreserved (DIS-specific)
◆ +2: intermediate resonance, mass shouldbe preserved
▲ recoil from parton shower needs to beabsorbed by particles in the event
◆ +3: intermediate resonance, fordocumentation only
◆ -9: incoming beam particles
<<Container for EVENT related information>>common /HepEUP/
+parameter MAXNUP: integer = 500, max num particle entries+NUP: integer = number entries this event+IDPRUP: integer = process id+XWGTUP: double = event weight+SCALUP: double = scale [GeV]+AQEDUP: double = QED coupling for this event+AQCDUP: double = QCD coupling for this event+IDUP(MAXNUP): integer = particle id+ISTUP(MAXNUP): integer = particle status+MOTHUP(2,MAXNUP): integer = pointer to parents+ICOLUP(2,MAXNUP): integer = particle (anit)color indices+PUP(5,MAXNUP): double = particle momentum, energy, mass+VTIMUP(MAXNUP): double = particle invariant lifetime+SPINUP(MAXNUP): double = spin vector angle (usually +1,-1)
<<called by SHG for HepEUP info>>subroutine UPEVNT()
each matrix element)
TeVU 8/29/02J. Huston
Event info
MOTHUP(2,I): index of first and lastmother
◆ For decays, daughter particles willonly have 1 mother
◆ For 2->n, daughter particles will have2 mothers
Color flow: specific choice of colorflow for a particular event is oftenunphysical, due to interferenceeffects, but SHGs require specificcolor state from which to beginshower
◆ ICOLUP(1,I): integer tag for colorflow line passing through color of theparticle
◆ Integer tag fro color flow line passingthrough anti-color of tag
<<Container for EVENT related information>>common /HepEUP/
+parameter MAXNUP: integer = 500, max num particle entries+NUP: integer = number entries this event+IDPRUP: integer = process id+XWGTUP: double = event weight+SCALUP: double = scale [GeV]+AQEDUP: double = QED coupling for this event+AQCDUP: double = QCD coupling for this event+IDUP(MAXNUP): integer = particle id+ISTUP(MAXNUP): integer = particle status+MOTHUP(2,MAXNUP): integer = pointer to parents+ICOLUP(2,MAXNUP): integer = particle (anit)color indices+PUP(5,MAXNUP): double = particle momentum, energy, mass+VTIMUP(MAXNUP): double = particle invariant lifetime+SPINUP(MAXNUP): double = spin vector angle (usually +1,-1)
<<called by SHG for HepEUP info>>subroutine UPEVNT()
each matrix element)
TeVU 8/29/02J. Huston
Example (gg->gg)
1
2
3
4
501
503
502 504
I ISTUP(I) IDUP(I) MOTHUP(1,I) MOTHUP(2,I) ICOLUP(1,I) ICOLUP(2,I)1 –1 21 (g) 501 5022 –1 21 (g) 502 5033 +1 21 (g) 1 2 501 5044 +1 21 (g) 1 2 504 503
initial/final state
particle code
mother/daughter relationships
color flow
TeVU 8/29/02J. Huston
Consider ttbar production
t and tbar given ISTUP=+2, whichinforms SHG to preserve theirinvariant masses when showeringand hadronizing the event
Intermediate s-channel gluon hasbeen drawn, but no entry becausecannot be distinguished from t-channel
Definition of color or anti-color linedepends on orientation of graph
◆ define color and anti-color accordingto physical time order
◆ quark will always have color tagICOLUP(1,I) filled, but never its anti-color tag ICOLUP(2,I); reverse foranti-quark; gluon has info in bothtags
Example: hadronic tt production
1
24
t 7b
8W
3t
5b
6W
501
502502
503
503
I ISTUP(I) IDUP(I) MOTHUP(1,I) MOTHUP(2,I) ICOLUP(1,I) ICOLUP(2,I)1 –1 21 (g) 0 0 501 5022 –1 21 (g) 0 0 503 5013 +2 –6 (t) 1 2 0 5024 +2 6 (t) 1 2 503 05 +1 –5 (b) 3 3 0 5026 +1 –24 (W−) 3 3 0 07 +1 5 (b) 4 4 503 08 +1 24 (W+) 4 4 0 0
The t and t are given ISTUP=+2, which informs the SHG to preserve their invariant masseswhen showering and hadronizing the event. An intermediate s-channel gluon has been drawnin the diagram, but since this graph cannot be usefully distinguished from the one with at-channel top exchange, an entry has not been included for it in the event record.
The definition of a line as ‘color’ or ‘anti-color’ depends on the orientation of the graph.This ambiguity is resolved by defining color and anti-color according to the physical timeorder. A quark will always have its color tag ICOLUP(1,I) filled, but never its anti-color tagICOLUP(2,I). The reverse is true for an anti-quark, and a gluon will always have informationin both ICOLUP(1,I) and ICOLUP(2,I) tags.
Note the difference in the treatment by the parton shower of the above example, and anidentical final state, where the intermediate particles are not specified:
9
TeVU 8/29/02J. Huston
q q
e e
1
3
2
54
501 501
Another example:little pink elephant exchange
I ISTUP(I) IDUP(I) MOTHUP(1,I) MOTHUP(2,I) ICOLUP(1,I) ICOLU1 –1 –2 (u) 0 0 0 502 –1 2 (u) 0 0 501 03 +2 0 (pink elephant) 1 2 0 04 +1 11 (e−) 3 3 0 05 +1 –11 (e+) 3 3 0 0
TeVU 8/29/02J. Huston
What Les Houches doesn’t do
Specify the exact formof output format
◆ nominally details aresupposed to be invisibleto casual user
Specify correct Q scalefor parton showering
◆ imagine that W + 5 jetevents probe moredeeply into shower thanW + 1 jet events
▲ would need lower scale
Interface Structure
<<Container for RUN related information>>common /HepRUP/
+paramter MAXPUP: integer = 100+IDBMUP(2): integer+EBMUP(2): double+PDFGUP(2): integer+PDFSUP(2): integer+IDWTUP: integer+NPRUP: integer+XSECUP(MAXPUP): double+XERRUP(MAXPUP): double+XMAXUP(MAXPUP): double+LPRUP(MAXPUP): integer
<<Container for EVENT related information>>common /HepEUP/
+parameter MAXNUP: integer = 500, max num particle entries+NUP: integer = number entries this event+IDPRUP: integer = process id+XWGTUP: double = event weight+SCALUP: double = scale [GeV]+AQEDUP: double = QED coupling for this event+AQCDUP: double = QCD coupling for this event+IDUP(MAXNUP): integer = particle id+ISTUP(MAXNUP): integer = particle status+MOTHUP(2,MAXNUP): integer = pointer to parents+ICOLUP(2,MAXNUP): integer = particle (anit)color indices+PUP(5,MAXNUP): double = particle momentum, energy, mass+VTIMUP(MAXNUP): double = particle invariant lifetime+SPINUP(MAXNUP): double = spin vector angle (usually +1,-1)
<<called by SHG to for HepRUP info>>subroutine UPINIT()
<<called by SHG for HepEUP info>>subroutine UPEVNT()
integer MAXPUP parameter ( MAXPUP=100 ) integer IDBMUP, PDFGUP,PDFSUP, IDWUP, NPRUP, LPRUP double precision EBMUP,XSECUP, XERRUP, XMAXUP common /HEPRUP/ IDBMUP(2), EBMUP(2), PDFGUP(2),PDFSUP(2), + IDWTUP, NPRUP, XSECUP(MAXPUP), XERRUP(MAXNUP), + XMAXUP(MAXNUP), LPRUP(MAXPUP)
(Specialized for each matrix element)
TeVU 8/29/02J. Huston
The proper way
The proper way of taking care of this would be togenerate parton showers starting at full hard scalebut vetoing those emissions that populate samephase space as exact ME
◆ See for example, Krauss, Catani, Kuhn andWebber, hep-ph/00109231
◆ Sjostrand argued for this type of specification atLes Houches but was out-shouted
◆ Herwig doesn’t need this information since it usesthe color flow to start the showering
Need to look for the next best solution
TeVU 8/29/02J. Huston
Comparisons of ME and MC calculations
Tree level ME predictions should agreewith each other (modulo αs choices, etc)
Parton showering Monte Carlos takeslightly different roads, though and maynot necessarily agree
◆ e.g. gg->H
Through implementation of colorcoherence, Herwig reproduces the LLterms and part of NLL terms
◆ see C. Balazs, J. Huston, I. Puljak,PRD; Ditto + S. Mrenna, LesHouches 2001 + paper in preparation
Pythia approximates color coherence andhas approximate agreement with logstructure
More global comparisons in progress
0 10 20 30 40 50 60 70 80 90 1000
0.2
0.4
0.6
0.8
1
pT (GeV)0 10 20 30 40 50 60 70 80 90 100
dp
T (
pb
/Gev
)⁄
dσ
0
0.2
0.4
0.6
0.8
1
0 20 40 60 80 100 120 140 160 180 20010
-3
10-2
10-1
1
pT (GeV)0 10 20 30 40 50 60 70 80 90 100
dp
T (
pb
/Gev
)⁄
dσ
0
0.2
0.4
0.6
0.8
1 gg → H + X at LHCmH = 125 GeV, CTEQ5M, √s = 14 TeVcurves: ResBos NLL: B(1,2), A(1,2), C(0,1) LL: B(1), A(1), C(0)histograms: PYTHIA 5.7 PYTHIA 6.136 PYTHIA 6.203 HERWIG 6.3
pT (GeV)0 20 40 60 80 100 120 140 160 180 200
dp
T (
pb
/Gev
)⁄
dσ
10-3
10-2
10-1
1
TeVU 8/29/02J. Huston
Logs that we know and love
A1, B1 and (a bit of) A2 areeffectively in Monte Carlos(especially Herwig)
A1,A2 and B1 for Higgsproduction are in currentoff-the-shelf version ofResBos
◆ …as are C0 and C1 whichcontrol the NLO normalization
The B2 term has recentlybeen calculated for gg->H
C. Balazs, Higgs production with soft-gluon effects, les Houches ’99 14
TeVU 8/29/02J. Huston
The need for higher order…
0
0.2
0.4
0.6
0.8
1
0 0.5 1 1.5 2
LONLO
NNLO
arX
iv:h
ep-p
h/01
0606
96
Jun
2001
TeVU 8/29/02J. Huston
What would we like?
Bruce Knuteson’s wishlist from the Run 2 Monte Carlo workshop
…all at NLO
TeVU 8/29/02J. Huston
What are we likely to get?
TeVU 8/29/02J. Huston
MCFM (Monte Carlo for FemtobarnProcesses) J. Campbell and K.Ellis
Goal is to provide a unifieddescription of processes involvingheavy quarks, leptons and missingenergy at NLO accuracy
There have so far been three main applicationsof this Monte Carlo, each associated with adifferent paper.
◆ Calculation of the Wbb background to a WHsignal at the Tevatron.
R.K.Ellis, Sinisa Veseli, Phys. Rev. D60:011501(1999), hep-ph/9810489.
◆ Vector boson pair production at the Tevatron,including all spin correlations of the boson decayproducts.
J.M.Campbell, R.K.Ellis, Phys. Rev. D60:113006(1999), hep-ph/9905386.
◆ Calculation of the Zbb and other backgrounds to aZH signal at the Tevatron.
J.M.Campbell, R.K.Ellis, FERMILAB-PUB-00-145-T, June 2000, hep-ph/0006304.
The last of these references contains the mostdetails of our method.
TeVU 8/29/02J. Huston
Higgs backgrounds using MCFM
TeVU 8/29/02J. Huston
Wbbar and Zbbar
0
0.5
1
1.5
2
2.5
0 20 40 60 80 100 120 140 160
dσ /
dMbb
[fb/
GeV
]
Mbb [GeV] (with Gaussian smearing of 10 GeV)
pp− → νe e+ bb− + X
√s=2 TeV MRS98
NLO
LO
TeVU 8/29/02J. Huston
Recent example of data vs Monte Carlo
There is a discovery potential atthe Tevatron during Run 2 for arelatively light Higgs (especially ifHiggs mass is 115 GeV)
…but small signal to backgroundratio makes understanding ofbackgrounds very important
CDF and ATLAS recently wentthrough similar exercisesregarding this background
◆ CDF using Run 1 data
◆ ATLAS using Monte Carlopredictions
ν
b
b-
ν
TeVU 8/29/02J. Huston
Data vs Monte Carlo
TeVU 8/29/02J. Huston
Sleuth strategy
Consider recent major discoveries inhep
◆ W,Z bosonsCERN 1983
◆ top quark Fermilab 1995
◆ tau neutrino Fermilab 2000
◆ Higgs Boson? CERN 2000
In all cases, predictions weredefinite, aside from mass
Plethora of models that appear dailyon hep-ph
Is it possible to perform a “generic”search?
Transparencies from Bruce Knuteson talk at Moriond 2001
TeVU 8/29/02J. Huston
W2j
We consider exclusive final statesWe consider exclusive final states
We assume the existence of standard object definitions
These define e, _, ττττ, γγγγ, j, b, ET, W, and Z fi
All events that contain the same numbers of each of theseobjects belong to the same final state
Step 1: Exclusive final statesSleuth Bruce Knuteson
e_ET
Z4j
eET jj eE
T 3jW3j eeγγγγeγγγγγγγγ
ZγγγγWγγγγγγγγ
__jj e_ET j
γγγγγγγγγγγγ ___eee
TeVU 8/29/02J. Huston
Results
Results agree well with expectationNo evidence of new physics is observed
DØ data
Search for regions of excess (more data Search for regions of excess (more data events than expected from background)events than expected from background) within that variable space within that variable space
probability to be SM
TeVU 8/29/02J. Huston
Fragmentation Uncertainties: Higgs->γγ and
Backgrounds
One of the most useful searchmodes for the discovery of theHiggs in the 100-150 GeV massrange at the LHC is in the twophoton mode
Higgs->γγ has very large
backgrounds from QCD sources
◆ Diphoton production
◆ γπo and πoπo production; jetsfragmenting into very high z πo’s
With excellent diphoton massresolution, can try to resolveHiggs “bump”
Still important to understand levelof background
10000
12500
15000
17500
20000
105 120 135
mγγ (GeV)
Eve
nts
/ 2 G
eV
0
500
1000
1500
105 120 135
mγγ (GeV)
Sign
al-b
ackg
roun
d, e
vent
s / 2
GeV
TeVU 8/29/02J. Huston
Diphoton Backgrounds in ATLAS
Again, for a H->γγ search at the
LHC, face irreducible backgrounds
from QCD γγ and reducible
backgrounds from γπo and πoπo
◆ in range from 70 to 170 GeV
jet-jet cross section is estimated
to be a factor of 2E6 times the γγ
cross section and γ-jet a factor
of 8E2 larger
◆ Need rejection factors of 2E7 and
8E3 respectively
◆ PYTHIA results seem to indicate
that reducible backgrounds are
comfortably less than reducible
ones
◆ …but how to normalize PYTHIApredictions for very high z fragmentation ofjets; fragmentation not known well at highz and certainly not for gluon jets
ATLAS detector and physics performance Volume IITechnical Design Report 25 May 1999
from higher-order corrections and from uncertainties in jet fragmentation. In order to reducethese backgrounds to a level well below that of the irreducible γγ continuum, rejection factors of2x107 and 8x103 are required.
These rejection factors have been realisticallyevaluated by using large samples of fully sim-ulated two-jet events, as described inSection 7.6. The rejection factors were then ap-plied to estimate the cross-sections for the re-ducible jet-jet and γ-jet backgrounds relative tothe irreducible γγ-background. The results areshown in Figure 19-2 as a function of the two-photon invariant mass mγγ. After applying thefull photon identification cuts from the calo-rimeter and the Inner Detector, the residualjet-jet and γ-jet backgrounds are found to be atthe level of approximately 15% and 20%, re-spectively, of the irreducible γγ backgroundover the mass range relevant to theH → γγ search.
Z → ee background
h
Figure 19-2 Expected ratios of the residual reduciblejet-jet and γ-jet backgrounds to the irreducible γγ-con-tinuum background as a function of the invariant massof the pair of photon candidates at high luminosity.
0
0.1
0.2
0.3
0.4
80 100 120 140 160
γ-jet/γγ
jet-jet/γγ
mγγ (GeV)
Rat
io
TeVU 8/29/02J. Huston
Different models predict different high z fragmentation
Backgrounds to γγ production inHiggs mass region arise fromfragmentation of jets to high z πo’s
◆ Pythia and Herwig predict verydifferent rates for high z
◆ all fragmentation is not equal
Example of a background that canbe measured in situ, but nice to beable to predict the environmentbeforehand
DIPHOX (see Run 2 MC workshop)program can calculate γγ, γπo, andπoπo cross sections to NLO
◆ comparisons underway to Tevatrondata
gg->γγ and qqbar->γγ at NNLO maybe available soon
10-3
10-2
10-1
1
10
10 2
(1/N
Jet)
dN/d
x E(c
h)
Y topology
Quark jetsGluon jets
DELPHI
00.20.40.60.8
11.21.4
0 0.2 0.4 0.6 0.8
Gluon / Quark
xE(ch)
Glu
on/Q
uark
10-3
10-2
10-1
1
10
10 2
(1/N
Jet)
dN/d
x E(c
h)
Jetset 7.4
00.20.40.60.8
11.21.4
0 0.2 0.4 0.6 0.8
Gluon / Quark
xE(ch)
Glu
on/Q
uark
10-3
10-2
10-1
1
10
10 2
(1/N
Jet)
dN/d
x E(c
h)
Ariadne 4.08
00.20.40.60.8
11.21.4
0 0.2 0.4 0.6 0.8
Gluon / Quark
xE(ch)
Glu
on/Q
uark
10-3
10-2
10-1
1
10
10 2
(1/N
Jet)
dN/d
x E(c
h)
Herwig 5.8C
00.20.40.60.8
11.21.4
0 0.2 0.4 0.6 0.8
Gluon / Quark
xE(ch)
Glu
on/Q
uark
B. Webber, hep-ph/9912399
TeVU 8/29/02J. Huston
PDF Uncertainties
What’s unknown aboutPDF’s
◆ the gluon distribution
◆ strange and anti-strangequarks
◆ details in the {u,d} quarksector; up/downdifferences and ratios
◆ heavy quarkdistributions
Σ of quark distributions (q + qbar) is well-
determined over wide range of x and Q2
◆ Quark distributions primarily determinedfrom DIS and DY data sets which havelarge statistics and systematic errors infew percent range (±3% for 10-4<x<0.75)
◆ Individual quark flavors, though may haveuncertainties larger than that on the sum;important, for example, for W asymmetry
information on dbar and ubar comes atsmall x from HERA and at medium x fromfixed target DY production on H2 and D2
targets
◆ Note dbar≠ubar
strange quark sea determined fromdimuon production in ν DIS (CCFR)
d/u at large x comes from FT DYproduction on H2 and D2 and leptonasymmetry in W production
TeVU 8/29/02J. Huston
Example:Jets at the Tevatron
Both experiments compare to NLO QCDcalculations
◆ D0: JETRAD, modified Snowmassclustering(Rsep=1.3, µF=µR=ETmax/2
◆ CDF: EKS, Snowmass clustering(Rsep=1.3 (2.0 in some previouscomparisons), µF=µR=ETjet/2
•In Run 1a, CDF observed an excess in thejet cross section at high ET, outside the range of the theoretical uncertainties shown
TeVU 8/29/02J. Huston
Similar excess observed in Run 1B
TeVU 8/29/02J. Huston
Exotic explanations
Transverse Energy
Cross section for various compositeness scales ΛC
1
1.5
2
2.5
3
3.5
4
4.5
5
0 50 100 150 200 250 300 350 400 45
Rat
io
Λ1300/Λ∞
Λ1400/Λ∞
Λ1500/Λ∞
Λ1600/Λ∞
Λ1700/Λ∞
Data 92/Λ∞
CDF Preliminary
Only statistical errors are plotted
TeVU 8/29/02J. Huston
Non-exotic explanations
Modify the gluon distribution at high x
TeVU 8/29/02J. Huston
Tevatron Jets and the high x gluon
Best fit to CDF and D0 central jet cross sections provided byCTEQ5HJ pdf’s
50 100 150 200 250 300 350 400pT
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Incl. Jet : pt7 * dσ/dpt
(Error bars: statistical only)
14% < Corr. Sys. Err. < 27%
Ratio: Prel. data / NLO QCD (CTEQ5M | CTEQ5HJ)
norm. facor :CTEQ5HJ: 1.04CTEQ5M : 1.00
CDF
Data / CTEQ5MCTEQ5HJ / CTEQ5MCDF Data ( Prel. )CTEQ5HJCTEQ5M
100 200 300 400 pT
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Incl. Jet : pt7 * dσ/dpt
Error bars: statistical only
8% < Corr. Sys. Err. < 30%
Ratio: Data / NLO QCD (CTEQ5M | CTEQ5HJ)
χ2 = norm. factor :CTEQ5HJ: 25/24 1.08CTEQ5M : 24/24 1.04
D0
Data / CTEQ5MCTEQ5HJ / CTEQ5MD0 dataCTEQ5HJCTEQ5M
TeVU 8/29/02J. Huston
D0 jet cross section as function of rapidity
10-1
1
10
10 2
10 3
10 4
10 5
10 6
10 7
10 8
0 100 200 300 400 500
ET (GeV)
⟨⟨⟨⟨ dσσσσ /
/// dE
T dηηηη⟩⟩⟩⟩
(fb/
GeV
) 0.0 ≤≤≤≤ ||||ηηηη|||| <<<< 0.5 0.5 ≤≤≤≤ ||||ηηηη|||| <<<< 1.0 1.0 ≤≤≤≤ ||||ηηηη|||| <<<< 1.5 1.5 ≤≤≤≤ ||||ηηηη|||| <<<< 2.0 2.0 ≤≤≤≤ ||||ηηηη|||| <<<< 3.0
DØ PreliminaryDØ Preliminary√√√√↵↵↵↵��������
���� ====
−−−−
����1
pb 92 Ldt √√√√↵↵↵↵��������
���� ====
−−−−
����1
pb 92 LdtRun 1BRun 1B
Nominal cross sections & statistical errors only
-1
0
1
50 100 150 200 250 300 350 400
-1
0
1
50 100 150 200 250 300 350 400
-1
0
1
50 100 150 200 250 300 350 400 450 500
-1
0
1
50 100 150 200 250 300 350 400 450 500
-1
0
1
50 100 150 200 250 300 350 400 450 500
JETRADµ=ET
max/2
CTEQ4HJ provides bestdescription of data
How reliable is NLO theory inthis region?K-factors?
TeVU 8/29/02J. Huston
Chisquares for recent pdf’s
PDF χχχχ2222 χχχχ2222//// dof Prob
CTEQ3M 121.56 1.35 0.01
CTEQ4M 92.46 1.03 0.41
CTEQ4HJ 59.38 0.66 0.99
MRST 113.78 1.26 0.05
MRSTgD 155.52 1.73 <0.01
MRSTgU 85.09 0.95 0.631
10
10 2
10 3
10 4
10 5
10 6
10 7
50 100 150 200 250 300 350 400 450 500
0.0 ≤ |η| < 0.50.5 ≤ |η| < 1.01.0 ≤ |η| < 1.51.5 ≤ |η| < 2.02.0 ≤ |η| < 3.0
•For 90 data points, are the chisquaresfor CTEQ4M and MRSTgU “good”?•Compared to CTEQ4HJ?
TeVU 8/29/02J. Huston
D0 jet cross section
CTEQ4 and CTEQ5 had CDFand D0 central jet cross sectionsin fit
Statistical power not greatenough to strongly influence highx gluon
◆ CTEQ4HJ/5HJ required aspecial emphasis to be given tohigh ET data points
Central fit for CTEQ6 is naturallyHJ-like
χ2 for CDF+D0 jet data is 113 for
123 data points
dataΗ theory theory versus pT GeVerror bars Η stat.Η syst.
100 200 300 400 500
Η 0.5
0
0.5
1
1.52ΗΗΗ3
100 200 300 400 500
Η 0.5
0
0.5
1
1.51.5ΗΗΗ2
100 200 300 400 500
Η 0.5
0
0.5
1
1.51ΗΗΗ1.5
100 200 300 400 500
Η 0.5
0
0.5
1
1.50.5ΗΗΗ1
100 200 300 400 500
Η 0.5
0
0.5
1
1.50ΗΗΗ0.5
Figure 19: Comparison between theory and the D0 jet data. The error bars are statistical
and systematic errors combined.
33
TeVU 8/29/02J. Huston
PDF Uncertainties: included in CTEQ6M sets inLHAPDF
Use Hessian technique (T=10)
TeVU 8/29/02J. Huston
Gluon Uncertainty
Gluon is fairly well-constrained up to an x-value of 0.3
New gluon is stiffer thanCTEQ5M; not quite as stiffas CTEQ5HJ
TeVU 8/29/02J. Huston
Luminosity function uncertainties at the Tevatron
102
√s (GeV)
Luminosity function at TeV RunII
Frac
tiona
l Unc
erta
inty
0.1
0.1
0.1
-0.1
-0.1
-0.1
0.1
Q-G --> Z
0
0
0
^
≈
≈ ≈
≈
200 405020
-0.1
0
≈
Q-G --> W+
Q-G --> W-
Q-G --> γ*
≈
102
√s (GeV)
Luminosity function at TeV RunII
Frac
tiona
l Unc
erta
inty
0.1
0.1
0.1
-0.1
-0.1
-0.2
0.2
Q-Q --> W+ (W-)
Q-Q --> γ* (Z)
G-G
0
0
0
^
≈
≈ ≈
≈−
−
0.3
0.4
200 4005020
±
TeVU 8/29/02J. Huston
Luminosity Function Uncertainties at the LHC
102 10
√s (GeV)
Luminosity function at LHC
Frac
tiona
l Unc
erta
inty
0.1
0.1
0.1
-0.1
-0.1
-0.1
0.1
0
0
0
0
-0.1
Q-G --> W+
Q-G --> W-
Q-G --> Z
Q-G --> γ*
^50 500200
≈ ≈
≈≈
≈≈
102 103
√s (GeV)
Luminosity function at LHC
Frac
tiona
l Unc
erta
inty
0.1
0.1
0.1
-0.1
-0.1
-0.2
0.2
Q-Q --> W+
Q-Q --> W-
G-G
0
0
0
≈
≈ ≈
≈−
−
±
50 200 500
^
TeVU 8/29/02J. Huston
Effective use of pdf uncertainties
PDF uncertainties are important both for precision measurements (W/Z crosssections) as well as for studies of potential new physics (a la jet cross sections athigh ET)
Most Monte Carlo/matrix element programs have “central” pdf’s built in, or caneasily interface to PDFLIB
Determining the pdf uncertainty for a particular cross section/distribution mightrequire the use of many pdf’s
◆ CTEQ Hessian pdf errors require using 33 pdf’s
◆ GKK on the order of 100
Too clumsy to attempt to includes grids for calculation of all of these pdf’s withthe MC programs
->Les Houches accord #2◆ Each pdf can be specified by a few lines of information, if MC programs can perform
the evolution
◆ Fast evolution routine will be included in new releases to construct grids for each pdf
NB: pdf uncertainties make most sense in the context of NLOcalculations; current MC programs are basically leading order and LOpdfs should be used when available
◆ NNB: CTEQ6L is a leading order fit to the data but using the 2-loop αs, since somehigher order corrections are in MC programs like Pythia, Herwig, etc
TeVU 8/29/02J. Huston
Les Houches accord #2
Using the interface is aseasy as using PDFLIB (andmuch easier to update)
First version has CTEQ6M,CTEQ6L, all of CTEQ6error pdfs and MRST2001pdfs
See pdf.fnal.gov (and talkby Walter Giele at thisconference)
call InitiPDFset(name)
◆ called once at the beginningof the code; name is the filename of external PDF file thatdefines PDF set
call InitPDF(mem)
◆ mem specifies individualmember of pdf set
call evolvePDF(x,Q,f)
◆ returns pdf momentumdensities for flavor f atmomentum fraction x andscale Q
TeVU 8/29/02J. Huston
The Big Idea
Reminder: the big idea:◆ The Les Houches accords will be
implemented in all ME/MC programsthat experimentalists/theorists use
◆ They will make it easy to generatethe multi-parton final states crucial tomuch of the Run 2/HERA/LHCphysics program and to compare theresults from different programs
◆ experimentalists/theorists can allshare common MC data sets
◆ They will make it possible togenerate the pdf uncertainties for anycross sections
TeVU 8/29/02J. Huston
Les Houches accords
Les Houches accord #1 (ME->MC)
◆ accord implemented in Pythia 6.2
◆ accord implemented in CompHEP
▲ CDF top dilepton group has beengenerating ttbar events withCompHEP/Madgraph + Pythia
◆ accord implemented in ALPGEN
▲ hep-ph/0206293
◆ accord implemented in Madgraph
▲ MADCUP:http://pheno.physics.wisc.edu/Software/MadCUP/}.
▲ MADGRAPH 2: within a few weeks
◆ work proceeding on Herwig; in release 6.5Sept 2002 (private version made availablefor debugging)
◆ Implemented in Grace
◆ in AcerMC:hep-ph/0201302
Les Houches accord #2 (pdfs inME/MC)
◆ version of pdf interface has beendeveloped
▲ available athttp://pdf.fnal.gov
◆ commitment for beingimplemented in MCFM
◆ commitment for beingimplemented in your name here
TeVU 8/29/02J. Huston
LO vs NLO
In some cases, differencebetween LO and NLO pdfs maybe larger than NLO uncertaintieson pdfs, e.g. the gluondistribution at high x
Ambiguity in LO pdfs◆ use 1-loop αs or 2-loop αs
▲ CTEQ6L uses 2-loop
▲ now have a 1-loop 6L inLHAPDF
▲ difference between the two isprobably a better estimate ofuncertainty than choosing a pdffrom another set
Will think about the idea of LOpdf uncertainties
TeVU 8/29/02J. Huston
Using pdf uncertainties
Currently need to run 40 pdfs tocalculate pdf uncertainties
Could be time-consuming (andunnecessary)
Think of 3 steps of decreasinglaziness
◆ look at plots of pdf luminosityuncertainty for given massand subprocess (and rapidity)
◆ use pdfs that are extremes fora particular subprocess
▲ particular eigenvectors probethat direction
◆ generate full cross sectionusing central pdf and storethe ratio of pdf luminositiesfor the 40 error pdf’s
theoretical model with many parameters, as well as its own sources of inherent uncertaintie
In [9, 10, 11], we formulated two methods, the Hessian and the Lagrange, which enab
the systematic characterization of the neighborhood of the global minimum of the χ2 function
These allow a systematic way of assessing the compatibility of the data sets in the framewor
of the theoretical (PQCD) model [36], as well as of estimating the uncertainties of the PDF
and their physical predictions if the experimental inputs are assumed to be compatibl
within certain practical tolerance. The basic ideas are summarized in the accompanyin
illustration (adapted from [10]):
(a)Original parameter basis
(b)Orthonormal eigenvector basis
zk
Tdiagonalization and
rescaling bythe iterative method
ul
ai
2-dim (i,j) rendition of d-dim (~20) PDF parameter space
Hessian eigenvector basis sets
ajul
p(i)
s0s0
contours of constant 2global
ul: eigenvector in the l-directionp(i): point of largest ai with tolerance T
s0: global minimum p(i)zl
The behavior of the global χ2 function in the neighborhood of the global minimum in th
PDF parameter space is encapsulated in 2Np sets of eigenvector PDFs (with Np ∼ 20 bein
the number of initial PDF parameters), obtained by an iterative procedure to diagonaliz
the Hessian matrix and to adjust the step sizes of the numerical calculation to match th
natural physical scales. This procedure efficiently overcomes a number of critical obstacles
encountered when applying standard tools to perform error propagation in the global χ
minimization approach. Details are given in [9, 10]
The analysis done in this work make full use of this method. The resulting 2Np +
PDF sets, consisting of the best fit and the eigenvector basis sets, allow us to calculate th
best estimate, and the range of uncertainty, of the PDF themselves as well as any physic
3For our analysis, there are ∼ 2000 data points from ∼ 16 different experiments of very different systematics and a wide range of precision.
4These are due to difficulties in calculating physically meaningful error matrices, by finite differences,the face of: (i) vastly different scales (∼ 106−7) in different, a priori unknown, directions in the high dimensioparameter space; and (ii) numerical fluctuations due to integration errors in the theoretical (PQCD) modcalculation as ell as round offs
TeVU 8/29/02J. Huston
Workshop in Cambridge (and at Tevatron)
Right before Amsterdam, there was a workshop held in Cambridge on TeV-scale Physics
◆ http://www.hep.phy.cam.ac.uk/theory/webber/camws.html
Original idea of workshop was to examine the implications of the several hundred pb-1 ofdata available from CDF and D0
Instead emphasis was more on ME/MC tools, especially NLO MC’s
◆ MC@NLO available now for diboson production, and soon will follow with moresubprocesses
◆ also talks by John Collins, Steve Mrenna, Dave Soper on their ideas for NLO MCs
◆ Most of the talks I’ve linked to the website:http://www.pa.msu.edu/~huston/cambridge_tevscale/cambridge_program.html
May be followup workshop in Durham this winter (January)
Also, there will be a 1-day mini-workshop with CDF/D0/theory at Fermi on issues of:
◆ Common MC tunings
◆ ME/MC comparisons and what else is needed
◆ Practical pdf uncertainties
Friday Oct 4 in 1-West
TeVU 8/29/02J. Huston
Conclusions
Great opportunity at Run 2 at theTevatron for discovery of new physics;even better opportunity when the LHCturns on
In order to be believeable, we mustunderstand the QCD backgrounds to anynew physics
◆ Don’t rely totally on Monte Carlosand certainly not on one Monte Carloalone
◆ In the words of Ronald Reagan,“Trust but Verify”, if possible,theoretical predictions/formalismswith data
▲ existing Run 1 data/Run 2 data
▲ background” data to be taken at theLHC
If no data, then verify with more completetheoretical treatments
Many new tools/links between old toolsare now being developed to make this jobeasier for experimenters
Hopefully, we’ll find many more of thetype of the event on the right to try them
t