parton distributions for the lhc - university of … · 2008-11-20 · parton distributions for the...
TRANSCRIPT
Parton Distributions for the LHC
James StirlingCambridge University
(with Alan Martin, Robert Thorne, Graeme Watt)
deep inelastic scattering
qµ
pµ
X
electron
proton
• variablesQ2 = –q2
x = Q2 /2p∙q (Bjorken x)
( y = Q2 /x s )
• resolution
at HERA, Q2 < 105 GeV2 ⇒ λ > 1018 m = rp/1000
• inelasticity
⇒ 0 < x ≤ 1
QQh GeVm102 16−×==λ
222
2
pX MMQQx
−+=
•in general, we can write
where the Fi(x,Q2) are called structure functions
0 .0 0 .1 0 .2 0 .3 0 .4 0 .5 0 .6 0 .7 0 .80 .0
0 .2
0 .4
0 .6
0 .8
1 .0
1 .2
Q 2 (G e V 2 ) 1 .5 3 .0 5 .0 8 .0 1 1 .0 8 .7 5 2 4 .5 2 3 0 8 0 8 0 0 8 0 0 0
F2(x
,Q2 )Bjorken
Scaling
4
• photon scatters incoherently off massless, pointlike, spin1/2 quarks
• probability that a quark carries fraction ξ of parent proton’s momentum is q(ξ), (0< ξ < 1)
the parton model (Feynman 1969)
• the functions u(x), d(x), s(x), … are called parton distribution functions (pdfs) they encode information about the proton’s deep structure
...)(91)(
91)(
94
)()()()( 2
,
21
0,
2
+++=
=−= ∑∫∑
xsxxdxxux
xqxexqedxF qqq
qqq
ξδξξξ
infinitemomentumframe
5
extracting pdfs from experiment• different beams (e,µ,ν,…) &
targets (H,D,Fe,…) measure different combinations of quark pdfs
• thus the individual q(x) can be extracted from a set of structure function measurements
• gluon not measured directly, but carries about 1/2 the proton’s momentum
[ ][ ]...2
...2
...)(91)(
94)(
91
...)(91)(
91)(
94
2
2
2
2
+++=
+++=
++++++=
++++++=
sduFusdF
ssdduuF
ssdduuF
n
p
en
ep
ν
ν
( ) 55.0)()(1
0=+∫∑ xqxqxdx
q
eNN FFss 22 365 −== ν
7
40 years of Deep Inelastic Scattering
0 .0 0 .1 0 .2 0 .3 0 .4 0 .5 0 .6 0 .7 0 .80 .0
0 .2
0 .4
0 .6
0 .8
1 .0
1 .2
Q 2 (G e V 2 ) 1 .5 3 .0 5 .0 8 .0 1 1 .0 8 .7 5 2 4 .5 2 3 0 8 0 8 0 0 8 0 0 0
F2(x
,Q2)
x
8
scaling violations and QCD
quarks emit gluons!0 .0 0 .2 0 .4 0 .6 0 .8 1 .0
0
2
4
6
x
F2
Q1
Q2 > Q1
The structure function data exhibit systematic violations of Bjorken scaling:
DGLAPequations
9
197277 197780 2004
going to higher orders in pQCD is straightforward in principle, since the above structure for F2 and for DGLAP generalises in a straightforward way:
The 2004 calculation of the complete set of P(2) splitting functions by Moch, Vermaseren and Vogt (hepph/0403192,0404111) completes the calculational tools for a consistent NNLO pQCD treatment of Tevatron & LHC hardscattering cross sections
beyond lowest order in pQCD
see above see book see next slide!
11
summary: how pdfs are obtained• choose a factorisation scheme (e.g. MSbar), an order in
perturbation theory (see below, e.g. LO, NLO, NNLO) and a ‘starting scale’ Q0 where pQCD applies (e.g. 12 GeV)
• parametrise the quark and gluon distributions at Q0,, e.g.
• solve DGLAP equations to obtain the pdfs at any x and scale Q > Q0 ; fit data for parameters {Ai,ai, …� S}
• approximate the exact solutions (e.g. interpolation grids, expansions in polynomials etc) for ease of use; thus the output ‘global fits’ are available ‘off the shelf”, e.g.
input | output
SUBROUTINE PDF(X,Q,U,UBAR,D,DBAR,…,BBAR,GLU)
12
the asymmetric sea•the sea presumably arises when ‘primordial‘ valence quarks emit gluons which in turn split into quarkantiquark pairs, with suppressed splitting into heavier quark pairs
•so we naively expect
• but why such a big du asymmetry? Meson cloud, Pauli exclusion, …?
...csdu >>>≈
The ratio of DrellYan cross sections for pp,pn � ¡+μ- + X provides a measure of the difference between the u and d sea quark distributions
13
strangeearliest pdf fits had SU(3) symmetry:
later relaxed to include (constant) strange suppression (cf. fragmentation):
with κ = 0.4 – 0.5
nowadays, dimuon production in υN DIS (CCFR, NuTeV) allows ‘direct’ determination:
in the range 0.01 < x < 0.4
data seem to prefer
theoretical explanation?!
16
charm, bottomconsidered sufficiently massive to allow pQCD treatment:
distinguish two regimes:(i) include full mH dependence to get correct threshold behaviour(ii) treat as ~massless partons to resum αS
nlogn(Q2/mH2) via DGLAP
FFNS: OK for (i) only ZMVFNS: OK for (ii) only
consistent GM(=general mass)VFNS now available (e.g. ACOT(χ), RobertsThorne) which interpolates smoothly between the two regimes
Note: definition of these is tricky and nonunique (ambiguity in assignment of O(mH
2//Q2) contributions), and the implementation of improved treatment (e.g. in going from MRST2006 to MSTW 2008) can have a big effect on light partons
19
the MRS/MRST/MSTW project
• since 1987 (with Alan Martin, Dick Roberts, Robert Thorne, Graeme Watt), to produce ‘stateoftheart’ pdfs
• combine experimental data with theoretical formalism to perform ‘global fits' to data to extract the pdfs in userfriendly form for the particle physics community
• currently widely used at HERA and the Fermilab
Tevatron, and in physics simulations for the LHC
• currently, the only available NNLO pdf sets with rigorous treatment of heavy quark flavours
1985
1990
1995
2000
2005
MRS(E,B)MRS(E’,B’)
MRS(D,D0,S0)
MRS(A)MRS(A’,G)MRS(Rn)MRS(J,J’)MRST1998MRST1999MRSTS2001E
MRST2004MRST2004QED
MRST2006
MSTW2008
20
MSTW 2008 update
• new data (see next slide)
• new theory/infrastructure
− δ fi from new dynamic tolerance method− new definition of αS (no more ΛQCD)− new GMVFNS for c, b (see Martin et al., arXiv:0706.0459)− new fitting codes: FEWZ, fastNLO− new grids: denser, broader coverage− slightly extended parametrisation at Q0
2 :344=30 free parameters including αS
22
MSTW input parametrisation
Note: 20 parameters allowed to go free for eigenvector PDF sets, cf. 15 for MRST sets
27
Scattering processes at high energy hadron colliders can be classified as either HARD or SOFT
Quantum Chromodynamics (QCD) is the underlying theory for all such processes, but the approach (and the level of understanding) is very different for the two cases
For HARD processes, e.g. W or highET jet production, the rates and event properties can be predicted with some precision using perturbation theory
For SOFT processes, e.g. the total cross section or diffractive processes, the rates and properties are dominated by nonperturbative QCD effects, which are much less well understood
What can we calculate?
WJS
29
where X=W, Z, H, highET jets, SUSY sparticles, black hole, …, and Q is the ‘hard scale’ (e.g. = MX), usually µF = µR = Q, and σ is known …
• to some fixed order in pQCD, e.g. highET jets
• or in some leading logarithm approximation (LL, NLL, …) to all orders via resummation
the QCD factorization theorem for hardscattering (shortdistance) inclusive processes
^
σ̂
30
DGLAP evolution
momentum fractions x1 and x2 determined by mass and rapidity of X
x1P
proton
x2P
proton
M
31
pdfs at LHC – the issues• high precision cross section predictions require accurate
knowledge of pdfs: δσth = δσpdf + …
, improved signal and background predictions ∙ easier to spot new physics
• ‘standard candle’ processes (e.g. σZ) to– check formalism (factorisation, DGLAP, …)– measure machine luminosity?
• learning more about pdfs from LHC measurements. e.g. – highET jets gluon? – W+,W–,Z0 quarks? – forward DY ¿ small x?– …
32
how important is pdf precision?
Catani et al,
• Example 1: : (MH=120 GeV) @ LHCδδ pdf ≈ ±2%, δδ ptNNL0 ≈ ± 10%
δδ ptNNLL ≈ ± 8% � δδ theory ≈ ± 10%
• Example 2: : (Z0) @ LHCδδ pdf ≈ ±2%, δδ ptNNL0 ≈ ± 2%
δδ theory ≈ ± 3%
±2%
MSTW
33
• Example 3: : (tt) @ LHCδδ pdf ≈ ±2%, δδ ptNNL0approx ≈ ± 3%
δδ theory ≈ ± 4%
• Example 4: quantitative limits on New Physics depend on pdfs
Moch, Uwer
34
• LED accelerate the running of r S as the compactification scale Mc is approached
• sensitivity attentuated by pdf uncertainties in SM prediction
sensitivity of dijet cross section at LHC to large extra dimensions
Ferrag (ATLAS), hepph/0407303
35
parton luminosity functions• a quick and easy way to assess the mass and collider energy dependence of production cross sections
s Ma
b
• i.e. all the mass and energy dependence is contained in the Xindependent parton luminosity function in [ ]• useful combinations are • and also useful for assessing the uncertainty on cross sections due to uncertainties in the pdfs
37
parton luminosity uncertainties at LHC
1 0 2 1 0 31 5
1 0
5
0
5
1 0
1 5
g g ® X q q b a r ® X G G ® X |y
X| < 2 .5
G = g + 4 /9 ¡q(q + q b a r)
p a r to n lu m in o s ity u n c e r ta in t ie sa t L H C (M S T W 2 0 0 8 N L O )
lum
ino
sity
un
cert
ain
ty (
pe
rce
nt)
MX (G e V )
39
future hadron colliders: energy vs luminosity?
1 02
1 03
1 04
1 01
1 00
1 01
1 02
1 03
1 04
1 05
1 06
1 07
1 08
1 4 T e V
4 0 T e V
p a r to n lu m in o s ity : g g → X
glu
ong
luo
n lu
min
osi
ty
[p
b]
MX [G e V ]
for MX > O(1 TeV), energy × 3 is better than luminosity × 10 (everything else assumed equal!)
recall partonparton luminosity:
so that
with τ = MX2/s
40
pdfs at LHC – the issues• high precision cross section predictions require accurate
knowledge of pdfs: δσth = δσpdf + …
, improved signal and background predictions ∙ easier to spot new physics
• ‘standard candle’ processes (e.g. σZ) to– check formalism (factorisation, DGLAP, …)– measure machine luminosity?
• learning more about pdfs from LHC measurements. e.g. – highET jets gluon? – W+,W–,Z0 quarks? – forward DY ¿ small x?– …
41
comparison with measured Tevatron cross sections
data errors dominated by ±6% systematic error from luminosity uncertainty!
• Absolute luminosity – from the parameters of the LHC machine
(immediate, but only 1015% accuracy)– rate of pp� Z0, W± ± l+ l , l etc.– from elastic scattering in the Coulomb region
(ALFA = roman pots at 240m), high β* running, eventually 23% accuracy– combinations of above
• Relative luminosity– LUCID Cerenkov monitor, large dynamic range,
excellent linearity
R
50 1 2 43 6 7 8 y109 TAS
Barrel
FCAL LUCIDTracking
EndCap
RP ZDC/TAN
Diffraction/Proton Tagging Regionµ chambers
pp
bgdobs
LANN
⋅⋅−
=ε
σ exp
( )2
2
0
2
42 t
eit
aFFLdtdN EM
NCt
B−
≈
++−≈+= L tot
πππ
ATLAS aims for 23% accuracy in L, eventually
Luminosity measurements in ATLAS
43
L from a fit to the tspectrum
( ) ( )( )
++−=
+=
−−
2
222/
2
22
2
1614
ce
te
tcL
FFLdtdN
tBtot
tBtot
NC
μ
πρσαρσπα
π
4.24%
0.59%
0.74%
1.5 %
error
B
� tot
L
92%0.15020.15
64%17.95 Gev218 Gev2
99%101.1 mb100 mb
8.162 10268.124 1026
correlation fitinput
Simulating 10 M events,running 100 hrsfit range 0.000550.03
large stat.correlation between L and other parameters
from a talk by: Hasko Stenzel on behalf of
the ATLAS Luminosity & forward physics WG
44
• cross sections (total and rapidity distributions) known to NNLO pQCD and NLO EW; perturbation series seems to be converging quickly
• EW parameters well measured at LEP • samples pdfs where they are well measured (in x) in DIS
• … although the mix of quark flavours is different: F2 and σ(W,Z) probe different combinations of u,d,s,c,b � sea quark distributions important
• precise measurement of cross section ratios at LHC (e.g. σ(W+)/σ(W), σ(W±)/σ(Z)) will allow these subtle effects to be explored further
standard candles: σ(W,Z) @ LHC
47
• MRST/MSTW NNLO: 2008 ~ 2006 > 2004 mainly due to changes in treatment of charm
• CTEQ: 6.6 ~ 6.5 > 6.1 due to changes in treatment of s,c,b• NLO: CTEQ6.6 2% higher than MSTW 2008 at LHC, because of
slight differences in quark (u,d,s,c) pdfs, difference within quoted uncertainty
predictions for σ(W,Z) @ Tevatron, LHC
49
Note: at NNLO, factorisation and renormalisation scale variation M/2 � 2M gives an additional ± 2% change in the LHC cross sections
50
predictions for σ(W,Z) @ LHC (Tevatron)
MSTW
2.001 (0.2543)21.32 (2.733)Alekhin 2002 NLO1.977 (0.2611)21.13 (2.805)Alekhin 2002 NNLO
2.043 (0.2393)21.58 (2.599)CTEQ6.6 NLO1.917 (0.2519)20.23 (2.724)MRST 2004 NNLO1.964 (0.2424)20.61 (2.632)MRST 2004 NLO2.044 (0.2535)21.51 (2.759)MRST 2006 NNLO2.018 (0.2426)21.21 (2.645)MRST 2006 NLO
2.051 (0.2507)21.72 (2.747)MSTW 2008 NNLO2.001 (0.2426)21.17 (2.659)MSTW 2008 NLO
Bll .σZ (nb)Blυ .σW (nb)
53
• high precision cross section predictions require accurate knowledge of pdfs: δσth = δσpdf + …
, improved signal and background predictions ∙ easier to spot new physics
• ‘standard candle’ processes (e.g. σZ) to– check formalism (factorisation, DGLAP, …)– measure machine luminosity?
• learning more about pdfs from LHC measurements. e.g. – highET jets gluon? – W+,W–,Z0 quarks? – forward DY ¿ small x?– …
pdfs at LHC – the issues
54
impact of LHC measurements on pdfs• the standard candles:
central σ(W,Z,tt,jets) as a probe and test of pdfs in the x ~ 10 2±1, Q2 ~ 1046 GeV2 range where most New Physics is expected (H, SUSY, ….)
• forward production of (relatively) lowmass states (e.g. γ *,dijets,…) to access partons at x<<1 (and x~1)
W,Z
γ*
56
Unique features
• pseudorapidity range 1.9 4.9– 1.9 2.5 complementary to ATLAS/CMS– > 2.5 unique to LHCb
• beam defocused at LHCb: 1 year of running = 2 fb1
• trigger on low momentum muons: p > 8 GeV, pT > 1 GeV
access to unique range of (x,Q2)
LHCb
59
Impact of 1 fb1 LHCb data for forward Z and γ* (M = 14 GeV) production on the gluon distribution uncertainty
60
summary• precision phenomenology at highenergy colliders such
as the LHC requires an accurate knowledge of the distribution functions of partons in hadrons
• determining pdfs from global fits to data is now a major industry… the MSTW collaboration is about to release its latest (2008) LO, NLO, NNLO sets
• pdf uncertainty for ‘new physics’ cross sections not expected to be too important (few % level), apart from at very high mass
• ongoing highprecision studies of standard candle cross sections and ratios
• potential of LHCb to access very small x via lowmass DrellYan lepton pair production
62
PDF eigenvector sets
• the Hessian matrix
• diagonalise the covariance matrix C ≡ H1
• produce eigenvector pdf sets Sk± with parameters ai shifted from the
global minimum
with t adjusted to give the desired tolerance
• then calculate uncertainties on a quantity F with
CTEQ
63
criteria for choice of tolerance TParameterfitting criterion
• T2 = 1 for 68% (1σ) c.l., T2 = 2.71 for 90% c.l., etc• appropriate if fitting consistent data sets with ideal Gaussian errors to
a welldefined theory• in practice: minor inconsistencies between fitted data sets, and
unknown experimental and theoretical uncertainties, so• therefore not appropriate for global PDF analysis
Hypothesistesting criterion (CTEQ)
• much weaker than the parameterfitting criterion: treat eigenvector pdf sets as alternative hypotheses
• determine T2 from the criterion that each data set should be described within its 90% c.l. limit
70
summary of DIS data+ neutrino FT DIS data Note: must impose cuts on
DIS data to ensure validity of leadingtwist DGLAP formalism in the global analysis, typically:
Q2 > 2 4 GeV2
W2 = (1x)/x Q2 > 10 15 GeV2