Recent Experimental Results on QCD Factorization Breaking of Nonperturbative Functions Joe Osborn University of Michigan Based on work in arXiv:1609.04769, submitted to Phys. Rev. D November 21, 2016 J.D. Osborn UMich HEP Seminar 11/21/2016 1 / 37
Recent Experimental Results on QCD FactorizationBreaking of Nonperturbative Functions
Joe OsbornUniversity of Michigan
Based on work in arXiv:1609.04769, submitted to Phys. Rev. D
November 21, 2016
J.D. Osborn UMich HEP Seminar 11/21/2016 1 / 37
Outline
Why study the 3-D structure of nucleon
Physics in the transverse momentum dependent framework
RHIC, PHENIX, and two particle angular correlations
Recent results on factorization breaking
Future factorization breaking measurements
J.D. Osborn UMich HEP Seminar 11/21/2016 2 / 37
Mapping the Structure of the Proton
Historically have usedsemi-inclusivedeep-inelastic-scattering(SIDIS) and Drell-Yan(DY) as probes of hadronstructure
SIDIS measurementsshowed there is structureto the proton
Longitudinal structure ofproton in terms ofx = pquark/pproton
Well mapped out overlarge range of x and Q2
J.D. Osborn UMich HEP Seminar 11/21/2016 3 / 37
1-D Structure
Collected data has led toincredible precision forpartonic structure ofnucleons in thelongitudinal direction!
Collinear partondistribution functions(PDFs) are very wellconstrained over 4 ordersof magnitude in x
Figures taken from http://nnpdf.hepforge.org
J.D. Osborn UMich HEP Seminar 11/21/2016 4 / 37
QCD Cross Sections
To account for boundstate nature ofhadrons, cross sectionsare factorized
Nonperturbative partondistribution andfragmentationfunctions (PDFs andFFs) are used todescribe the individualpartons within a hadron
Functions arenonperturbative, mustbe constrained by data!
Taken to be processindependent anduncorrelated
σ = f1(x ,Q2)⊗f2(x ,Q2)⊗d σ̂
dt⊗Dh
q (z ,Q2)
J.D. Osborn UMich HEP Seminar 11/21/2016 5 / 37
Multidimensional Proton Structure
What does thebound-state proton looklike in terms of the quarksand gluons inside it?
PositionMomentumSpinFlavorChargeColor (!)
What about transversemomentum degrees offreedom?
J.D. Osborn UMich HEP Seminar 11/21/2016 6 / 37
1D vs. 3D Nonperturbative Functions
Historicallynonperturbative functionsare approximated as onlydependent on the collinearmomentum fraction x
In reality there must betransverse structure dueto the confined nature ofthe partons and theadditional possibility ofgluon radiation
The unintegrated kT distributions are explicitly dependent ontransverse momentum
Parton Distribution Functions: f (x)→ f (x , kT )
Fragmentation Functions: D(z)→ D(z , jT )
We can also add spin into the picture...
J.D. Osborn UMich HEP Seminar 11/21/2016 7 / 37
Transverse-Momentum-Dependent PDF Zoo
Transverse-Momentum-Dependent(TMD) PDFs
N - Nucleonq - QuarkU - UnpolarizedL - Longitudinally polarizedT - Transversely polarized
8 TMD PDFs at twist-2 describing transverse partonic structure,spin-spin, and spin-momentum correlations!
Image taken from Alexei Prokudin Spin 2016
J.D. Osborn UMich HEP Seminar 11/21/2016 8 / 37
Transverse-Momentum-Dependent Functions in Nature
Many transverse-momentum-dependentPDFs and FFs correlatespin with momentum
Beginning to reallyexplore partoniccorrelations within thenucleon! New era ofnucleon structure
Are these correlationsreally present in nature?Absolutely!
Sivers TMD PDF
arXiv:1609.07374, COMPASS at CERN
J.D. Osborn UMich HEP Seminar 11/21/2016 9 / 37
Transverse-Momentum-Dependent Functions in Nature
Boer Mulders TMD PDF
PRD 87, 012010(2013), HERMES at HERA
J.D. Osborn UMich HEP Seminar 11/21/2016 10 / 37
Transverse-Momentum-Dependent Functions in Nature
Transversity TMD PDF
PLB 717, (2012) 376, COMPASS at CERN
J.D. Osborn UMich HEP Seminar 11/21/2016 11 / 37
Transverse-Momentum-Dependent Functions in Nature
PRD 78, 032011 (2008), Belle at KEK
Collins TMD FF
J.D. Osborn UMich HEP Seminar 11/21/2016 12 / 37
Transverse-Momentum-Dependent Phenomenology
In the collinear framework, nonperturbative functions are taken to beuncorrelated, universal, process independent functions
In the transverse-momentum-dependent framework, it has beennecessary to re-check these assumptions
This has led to very interesting predictions...
J.D. Osborn UMich HEP Seminar 11/21/2016 13 / 37
Universality in Transverse-Momentum-Dependent Functions
Drell-Yan Semi-Inclusive DIS
Sign change in Sivers transverse-momentum-dependent PDFpredicted due to initial-state vs. final-state gluon exchange withproton remnants between Drell-Yan and semi-inclusive DIS: processdependent PDF!
Factorization of transverse-momentum-dependent PDFs andfragmentation functions still predicted to hold in these QED processes
J.D. Osborn UMich HEP Seminar 11/21/2016 14 / 37
First Measurement of Possible Modified Universality
Semi-inclusive DIS Sivers asymmetries have been measured, e.g. byHERMES and COMPASS collaborations
First measurement of Drell-Yan type process from STAR at RHIC!
Data support prediction of process dependenttransverse-momentum-dependent PDF
PRL 116, 132301(2016)
J.D. Osborn UMich HEP Seminar 11/21/2016 15 / 37
Factorization of Transverse-Momentum-Dependent Functions
Factorization is still predicted to hold in semi-inclusive DIS andDrell-Yan
σ = f1(x , kT ,Q2)⊗ f2(x , kT ,Q
2)⊗ d σ̂
dt⊗ Dh
q (z , jT ,Q2)
What about leading-order QCD processes where a colored quark orgluon is exchanged?
Color present in both the initial and final state - therefore soft gluonexchange possible in both the initial and final state
J.D. Osborn UMich HEP Seminar 11/21/2016 16 / 37
Factorization of Transverse-Momentum-Dependent Functions
Factorization breakingpredicted in a transverse-momentum-dependent(TMD) framework forp + p → h1 + h2 (PRD 81,094006 (2010))
TMD nonperturbativefunctions no longer defined -partons are quantummechanically correlatedacross colliding hadrons!
Consequence of soft gluonexchanges in both the initialand final state
Predicted modifieduniversality of certain TMDPDFs and factorizationbreaking from same physicalprocess - consequences ofcolor flow in action!
Consequence of QCD as anon-Abelian gauge theory
J.D. Osborn UMich HEP Seminar 11/21/2016 17 / 37
Looking for Factorization Breaking
An obvious way to look foreffects is by comparingmeasurement to acalculation which assumesfactorization
Problem: calculationsrequire good knowledge ofthe transverse-momentum-dependent nonperturbativefunctions. Collinear pQCDcalculations still have∼10-40% errors
What about observingchange of functions with thehard scattering scale?
PRD 91, 032001 (2015)
PRD 80, 111108 (2009)
J.D. Osborn UMich HEP Seminar 11/21/2016 18 / 37
Collins-Soper-Sterman (CSS) Evolution
CSS evolution first published in 1985. Similar to DGLAP evolution equation,but includes small transverse momentum scale
Has been used to successfully describe global Drell-Yan and Tevatron Z0
cross sections
Clear qualitative prediction - momentum widths sensitive to nonperturbativetransverse momentum increase with increasing hard scale
Due to increased phase space for hard gluon radiation
PRD 67, 073016 (2003)
J.D. Osborn UMich HEP Seminar 11/21/2016 19 / 37
Drell-Yan/Z and Semi-Inclusive DIS in CSS Evolution
Measurements show thatDrell-Yan andsemi-inclusive DIS followtheoretical prediction -widths rise with hardscattering scale
The theoretical evolutionprediction comes directlyout of the derivation fortransverse-momentum-dependentfactorization
SIDIS - Eur. Phys. J. C(2013) 73,2531
2 [GeV/c]2Q1 2 3 4 5 6 7 8
2 [G
eV/c
]⟩
2 Tp⟨
0.15
0.2
0.25
0.3
0.35
0.4
0.20<z<0.250.30<z<0.350.40<z<0.50
+ h-
h
]2 [GeV/cµµM0 5 10 15 20 25
[GeV
/c]
⟩Tp⟨
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
= 21 GeV, PRL 38,1334sp+C, = 25 GeV, PRD 23,604sp+Cu,
= 44 GeV, PRL 47,12sp+p, = 62 GeV, PRL 47,12sp+p,
FNAL fixed target DY
J.D. Osborn UMich HEP Seminar 11/21/2016 20 / 37
Drell-Yan/Z and Semi-Inclusive DIS in CSS Evolution
Phenomenological studiesconfirm that Drell-Yan andsemi-inclusive DIS followtheoretical prediction
The evolution predictioncomes directly out of thederivation for transverse-momentum-dependent(TMD) factorization
If TMD factorization,then CSS evolution. Ifnot CSS evolution,then not TMDfactorization!
Drell-Yan and semi-inclusiveDIS clearly follow theoreticalprediction
DY/Z - PLB 633, 710 (2006)
SIDIS - PRD 89, 094002 (2014)
J.D. Osborn UMich HEP Seminar 11/21/2016 21 / 37
Direct Photons and Dihadrons
Direct photon-hadron anddihadron correlations bothpredicted to be sensitive tofactorization breaking effectsin PHENIX
Assuming factorization,direct photon-hadrons probethree nonperturbativefunctions, while dihadronsprobe four
Direct photons offer one lessavenue for gluon exchangein the final-state:fewer/different effects?
J.D. Osborn UMich HEP Seminar 11/21/2016 22 / 37
Relativistic Heavy Ion Collider - RHIC at Brookhaven National Laboratory
J.D. Osborn UMich HEP Seminar 11/21/2016 23 / 37
PHENIX Detector
PHENIX central arms
∆φ ∼ π|η| <0.35
ElectromagneticCalorimeter (PbSc/PbGl)provides isolated directphoton and π0 → γγdetection
Drift Chamber (DC) andPad Chambers (PC)provide nonidentifiedcharged hadron detection
New results from2012/2013
√s=510 GeV
p+p runs
J.D. Osborn UMich HEP Seminar 11/21/2016 24 / 37
Angular Correlation Observables
Direct photon-hadron production
Dihadron production
pout = passocT sin ∆φ
J.D. Osborn UMich HEP Seminar 11/21/2016 25 / 37
∆φ Correlations for π0-h± and Direct γ-h±
-1 0 1 2 3 4 50
0.5
<15 GeV/ctrig
T12<p
<2 GeV/cassocT
1<p
-1 0 1 2 3 4 50
0.1 <15 GeV/ctrig
T12<p
<4 GeV/cassocT
3<p
-1 0 1 2 3 4 50
0.1
<15 GeV/ctrig
T12<p
<10 GeV/cassocT
5<p
-1 0 1 2 3 4 50
0.5
<12 GeV/ctrig
T9<p
<2 GeV/cassoc
T1<p
-1 0 1 2 3 4 50
0.1
<12 GeV/ctrig
T9<p
<4 GeV/cassoc
T3<p
-1 0 1 2 3 4 50
0.1
<12 GeV/ctrig
T9<p
<10 GeV/cassoc
T5<p
PHENIX
-1 0 1 2 3 4 50
0.5
<9 GeV/ctrig
T8<p
<2 GeV/cassocT
1<p
-1 0 1 2 3 4 50
0.1
<9 GeV/ctrig
T8<p
<4 GeV/cassocT
3<p
=510 GeVsp+p at
|<0.35η|
-1 0 1 2 3 4 50
0.1
<9 GeV/ctrig
T8<p
<10 GeV/cassocT
5<p
±-hγIsolated Direct ±-h0π
Underlying Event
[rad]φ∆
-1 [
rad
]φ∆dd
Ntr
igN
1
Two jet structure visible for π0-h±, isolation cut on near side for direct γ-h±
Direct γ-h± probes smaller jet energy due to emerging from hard scattering at LO
J.D. Osborn UMich HEP Seminar 11/21/2016 26 / 37
√〈p2
out〉 Extracted from Fits to ∆φ Correlations
[GeV/c]trig
Tp
4 5 6 7 8 9 10 11 12 13 14
[GeV
/c]
⟩2 ou
t p⟨
0
1
2
3
4
5±-hγ
|<0.35η|
=510 GeVsp+p at
PHENIX
±-h0π<3 GeV/cassoc
T2<p
<4 GeV/cassocT
3<p
√〈p2
out〉 characterizes away-side jet width in momentum space
Decreases with hard scattering scale ptrigT , opposite of semi-inclusiveDIS and Drell-Yan!
Sensitive to perturbative and nonperturbative kT and jT ; fits are toentire away-side jet
J.D. Osborn UMich HEP Seminar 11/21/2016 27 / 37
pout Distributions
pout shows twodistinct regions:Gaussian and powerlaw
Gaussian fits clearlyfail past ∼1.3GeV/c
Indicates transitionfromnonperturbative toperturbative kT andjT
[GeV/c]out
p-8 -6 -4 -2 0 2 4 6 8
-1 [G
eV/c
]ou
tdpdN
trig
N1
-1110
-1010
-910
-810
-710
-610
-510
-410
-310
-210
-110
1
10
PHENIX
=510 GeVsp+p at
|<0.35η|
3π4<φ∆<3
π2
<10 GeV/cassocT
0.7<p
±-h0π 4-5
)-1 5-6 (x10)-2 6-7 (x10)-3 7-8 (x10)-4 8-9 (x10
)-5 9-12 (x10)-6 12-15 (x10
[GeV/c]trigT
p
±-hγ
Note: Curves are Kaplan andGaussian fits, not calculations!!
J.D. Osborn UMich HEP Seminar 11/21/2016 28 / 37
Gaussian Widths of pout
Extract Gaussian widthsof pout vs. ptrigT , hardscattering scale
Sensitive to onlynonperturbative kT andjT in the nearlyback-to-back region∆φ ∼ πGaussian widths decreasewith ptrigT also, consistent
with√〈p2
out〉 andopposite of semi-inclusiveDIS and Drell-Yan!
[GeV/c]trig
Tp
4 5 6 7 8 9 10 11 12 13 14
Gau
ssia
n W
idth
[GeV
/c]
0.45
0.5
0.55
0.6
0.65±-h0π
±-hγ Linear Fit±-h0π
Linear Fit±-hγ<10 GeV/cassoc
T0.7<p
PHENIX
|<0.35η|=510 GeVsp+p at
J.D. Osborn UMich HEP Seminar 11/21/2016 29 / 37
PYTHIA p+p Event Simulation
To make a comparison,used PYTHIA eventgenerator simulation
PYTHIA reproducesexpectation from CSSevolution in Drell-Yanover large range of Mµµ
]2 [GeV/cµµM5 10 15 20 25 30
Gau
ssia
n W
idth
[GeV
/c]
1
2
3
PYTHIA Perugia0 SimulationDrell-Yan Dilepton
>1 GeV/clepT
p|<0.35η|
=510 GeVsp+p at
]2 [GeV/cµµM20 30 40 50 60 70 80 90 100
Gau
ssia
n W
idth
[GeV
/c]
2
3
4
5PYTHIA Perugia0 SimulationDrell-Yan Dilepton
>4 GeV/clepT
p|<0.35η|
=510 GeVsp+p at
J.D. Osborn UMich HEP Seminar 11/21/2016 30 / 37
PYTHIA Event Simulation
[GeV/c]out
p-5 -4 -3 -2 -1 0 1 2 3 4 5
-1 [G
eV/c
]ou
tdpdN
trig
N1
-910
-810
-710
-610
-510
-410
-310
-210
-110
1
10
±-h0π 4-5
)-1 5-6 (x10)-2 6-7 (x10)-3 7-8 (x10)-4 8-9 (x10
)-5 9-12 (x10)-6 12-15 (x10
PYTHIA Perugia0 Simulation
=510 GeVsp+p at <10 GeV/cassoc
T0.7<p
3π4<φ∆<
3π2
|<0.35η|
±-hγ [GeV/c]trig
Tp
Can construct poutdistributions fordirect photons anddihadrons inPYTHIA as well fordirect comparison
PYTHIA replicatesthe nonperturbativeto perturbativetransition in thepout distributions
J.D. Osborn UMich HEP Seminar 11/21/2016 31 / 37
PYTHIA Event Simulation
PYTHIA alsoreplicates thenegative slope ofthe gaussian widthsin γ−hadron andπ0−hadron!
Magnitudes ofwidths fromPYTHIA show∼15% differencefrom data despiteslope beingreplicated
[GeV/c]trig
Tp
4 5 6 7 8 9 10 11 12 13 14
Gau
ssia
n W
idth
[GeV
/c]
0.5
0.6
0.7
0.8 Linear Fit±-h0πPHENIX
Linear Fit±-hγPHENIX Linear Fit±-h0πPYTHIA
Linear Fit±-hγPYTHIA
|<0.35η|=510 GeVsp+p at <10 GeV/cassoc
T0.7<p
±-h0πPHENIX ±-hγPHENIX
±-h0πPYTHIA Perugia0 ±-hγPYTHIA Perugia0
J.D. Osborn UMich HEP Seminar 11/21/2016 32 / 37
PYTHIA Simulation
WHY does PYTHIA replicate both increasing behavior in p+p → `¯̀ anddecreasing behavior in p+p →h+X??
Unlike analytical pQCD calculation, PYTHIA forces all particles to colorneutralize in the event, including remnants
PYTHIA allows initial and final state soft gluon exchanges!
PYTHIA authors confirm that it is plausible that PYTHIA would besensitive to such effects
Image taken from http://home.thep.lu.se/ torbjorn/talks/karlsruhe10a.pdf
J.D. Osborn UMich HEP Seminar 11/21/2016 33 / 37
Relations to Other QCD Studies?
Color coherence studies atthe Tevatron
Phys. Rev. D 50, 5562 (1994) (CDF)Phys. Lett. B 414, 419-427 (1997) (D0)
Color coherence studies atthe LHC
Eur. Phys. J. C74 (2014) no.6,2901
J.D. Osborn UMich HEP Seminar 11/21/2016 34 / 37
Future Measurements
Recent RHIC run in 2015delivered one of the mostunique data sets toPHENIX and STAR
RHIC collided p↑+p,p↑+Au, and p↑+Al at√s = 200 GeV
PHENIX recorded ∼10xthe amount of
√s = 200
GeV data from previousanalysis
Possibility to comparep+p at
√s = 200 and
510 GeV
Possibility to comparep+p to p+A (strongergluon fields in nucleus??)
Does transverse spinchange anything??
J.D. Osborn UMich HEP Seminar 11/21/2016 35 / 37
Future Measurements - sPHENIX
Future detector at RHIC,sPHENIX, has beenproposed
Dedicated jet detector atRHIC
Golden measurement - γ-jet
Allows full kinematic eventreconstruction
J.D. Osborn UMich HEP Seminar 11/21/2016 36 / 37
Conclusions
Extending the knowledge of nucleon structure from 1 dimension to 3dimensions (and more!)
Transverse-momentum-dependent nucleon structure offers a richerdescription of the nucleon with many interesting phenomenologicalpredictions
Factorization breaking has been predicted in hadronic collisions wherea final-state hadron is measured in atransverse-momentum-dependent framework
PHENIX has just released the first measurement studying thesepredicted effects - arXiv:1609.04769
Data show the opposite evolution trend in the nonperturbativemomentum widths from semi-inclusive DIS and Drell-Yan, wherefactorization is predicted to hold
More measurements planned in the future... stay tuned!
J.D. Osborn UMich HEP Seminar 11/21/2016 37 / 37
Back Up
J.D. Osborn UMich HEP Seminar 11/21/2016 1 / 11
√s=200 GeV Results from PHENIX
Previous PHENIX resultat√s=200 GeV with
larger errors (Phys. Rev.D 82, 072001 (2010))
Next step: analyze recentRun 15
√s=200 GeV
p+p and p+A data fromRHIC!
6x luminosity in Run 15p+p, as well as first resultfrom p+A
Can also look attransverse spindependence in Run 15! 2 < passocT < 5
J.D. Osborn UMich HEP Seminar 11/21/2016 2 / 11
√s=200 GeV Results from RHIC
Previous PHENIX resultat√s=200 GeV to lower
ptrigT (PRD 81, 012002(2010))
Shows√〈p2
out〉 over lower
range of ptrigT
Also can plot away-sidewidth in angular space -same trend over largerange of ptrigT
[GeV/c]trig
Tp
2 4 6 8 10 12 14
[GeV
/c]
⟩2 ou
tp⟨
0
0.5
1
1.5
<5 GeV/c (PRD 74, 072002)assoc
T, 1.4<p±-h0π
<5 GeV/c (PRD 82, 072001)assoc
T, 2<p±-h0π
<5 GeV/c (PRD 81, 012002)assoc
T, 2<p±-h0π
<5 GeV/c (PRD 82, 072001)assoc
T, 2<p±-hγIsolated Direct
=200 GeVsp+p at |<0.35η|
[GeV/c]trig
Tp
5 10 15 20 25 30 35
Aw
ay-s
ide
Wid
th [r
ad]
0.2
0.3
0.4
0.5
0.6 <3 GeV/c (PRL 104, 252301)assoc
T 2<p±-h0π
<5 GeV/c (PRD 74, 072002)assoc
T 1.4<p±-h0π
<17 GeV/c (PRL 112, 122301)assoc
T 0.2<p±jet-h
=200 GeVsp+p at
J.D. Osborn UMich HEP Seminar 11/21/2016 3 / 11
〈zT 〉 with Gaussian Widths
〈zT 〉 ptrigT correction wasalso applied to Gaussianwidths vs. ptrigT
〈zT 〉 more or less amountsto a scale factor of 2difference in the slope
[GeV/c]jetT
p6 8 10 12 14 16 18 20 22 24
Gau
ssia
n W
idth
[GeV
/c]
0.45
0.5
0.55
0.6
0.65±-h0π
±-hγ Linear Fit±-h0π
Linear Fit±-hγ<10 GeV/cassoc
T0.7<p
PHENIX
|<0.35η|=510 GeVsp+p at
J.D. Osborn UMich HEP Seminar 11/21/2016 4 / 11
√〈p2
out〉 vs. pjetT
[GeV/c]jet
Tp
6 8 10 12 14 16 18 20 22 24
[GeV
/c]
⟩2 ou
t p⟨
0
1
2
3
4
5±-hγ
|<0.35η|
=510 GeVsp+p at
PHENIX
±-h0π<3 GeV/cassoc
T2<p
<4 GeV/cassocT
3<p
Examined√〈p2
out〉 as a function of pjetT as well
pjetT = ptrigT for direct photons
pjetT = ptrigT /〈zT 〉 for π0s, with 〈zT 〉 estimated using PYTHIA
〈zT 〉 =ptrigT
p̂trigT
The√〈p2
out〉 distributions almost form a continuous function?
J.D. Osborn UMich HEP Seminar 11/21/2016 5 / 11
More about Color Coherence
Radiation“drags” coloraway from vertex
Destructive interferenceoccurs away from emittedgluons
Soft radiation inhibited incertain areas
Leads to certain regionsof phase space wheregluons constructively ordestructively interfere
See the followingreferences
Phys. Rev. D 50,5562(1994)Phys. Lett. B 414(1997) 419-427Dokshitzer, Yuri.Basics of PerturbativeQCD (EditionsFrontieres, 1991)Chapters 4,5,9
J.D. Osborn UMich HEP Seminar 11/21/2016 6 / 11
SIDIS and e+e− Annihilation Momentum Widths
PRD 61, 014003
Z. Phys. C 21:37
J.D. Osborn UMich HEP Seminar 11/21/2016 7 / 11
Partonic Contributions to Processes at LO
[GeV/c]0π
Tp
2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Fra
ctio
nal C
ontr
ibut
ion
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1g+g→g+gq+q→q+qq+g→q+g
qq+→g+g
+X0π→p+p|<0.35η|=510 GeVs
CTEQ6L1 PDFs at LO
(a)
π0 contribution changesfrom gluon dominated atlow pT to mix of quarkand gluons at high pT
[GeV/c]γT
p2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Fra
ctio
nal C
ontr
ibut
ion
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
+qγ→q+g
+gγ→qq++Xγ→p+p
|<0.35η|=510 GeVs
CTEQ6L1 PDFs at LO
(b)
Direct photoncontribution dominatedby QCD Comptonscattering at all pT
NLO corrections small atmidrapidity (Phys. Lett.B 140,87)
J.D. Osborn UMich HEP Seminar 11/21/2016 8 / 11
Analysis Methods
Correlated π0−h± orisolated γ−h± are collectedand corrected with:
Charged hadronefficiencyAcceptance correction
Direct photons undergoadditional statisticalsubtraction to remove decayphoton background,estimated with Monte Carloprobability functions
Isolation and tagging cutsremove decay photonbackground and NLOfragmentation photons
Probability for a π0 to decay to aphoton which could not be tagged
with 5 < pT < 7 GeV/c in PHENIX
Y isodir =
1
R isoγ − 1
(R isoγ Y iso
inc − Y isodec
)PRD 82,072001 (2010)PRC 80,024908 (2009)
J.D. Osborn UMich HEP Seminar 11/21/2016 9 / 11
R isoγ Measurement at
√s=510 GeV
R isoγ measured for
statistical subtraction ofisolated decay photoncontribution
Rγ measured in PHENIXand corrected by taggingand isolation efficiencies
R isoγ >1 indicates isolated
direct photon production [GeV/c]γ
Tp
7 8 9 10 11 12 13 14
iso
γR
1
1.2
1.4
1.6
1.8
2
2.2
PHENIX
=510 GeVsp+p at |<0.35η|
R isoγ =
Rγ
(1− εtagdec)(1− εnisodec )
N isoinc
Ninc
J.D. Osborn UMich HEP Seminar 11/21/2016 10 / 11
PDF Review
J.D. Osborn UMich HEP Seminar 11/21/2016 11 / 11