Photoproduction of Diffractive Dijets in PYTHIA 8 DIS 2021

Photoproduction of diffractive dijets in PYTHIA 8 DIS 2021

Ilkka Helenius April 15th, 2021

In collaboration with Christine O. Rasmussen Motivation: Diffractive dijets at HERA [H1: JHEP 1505 (2015) 056]

H1 VFPS data H1 VFPS data H1 VFPS data AFG γ-PDF H1 VFPS data AFG γ-PDF

NLO H12006 Fit-B × 0.83 × (1+δ ) NLO H12006 Fit-B × 0.83 × (1+δNLO) H12006 Fit-B × 0.83 × (1+δ ) NLO H12006 Fit-B × 0.83 × (1+δ ) hadr hadr hadr hadr H1 4000 H1 H1 H1 [pb] [pb] 100 [pb] [pb] 60000 IP IP DIS IP 1000 DIS γp IP γp

/dz 3000 /dz /dx /dx σ σ σ σ 40000 d d d d 50 2000 500 20000 1000

1.5 1.5 1.5 1.5 1 1 1 1 0.5 0.5 0.5 0.5 ratio to NLO ratio to NLO ratio to NLO ratio to NLO 0 0.2 0.4 0.6 0.8 0.01 0.015 0.020 0.2 0.4 0.6 0.8 0.01 0.015 0.02 z x zIP xIP IP IP 2 • H1 data andH1 VFPS factorization-based data NLO calculationH1 VFPS data in DISH1 (highVFPS data Q ) inAFG γ agreement-PDF H1 VFPS data AFG γ-PDF

NLO H12006 Fit-B × 0.83 × (1+δ ) NLO H12006 Fit-B × 0.83 × (1+δNLO) H12006 Fit-B2 × 0.83 × (1+δ ) NLO H12006 Fit-B × 0.83 × (1+δ ) • NLO calculation overshoothadr the data in photoproductionhadr (low Q ) hadr hadr ] 2 10 1500 2 H1 H1 ⇒ H1 H1 [pb] Factorization broken in hard diffraction at low Q similarly as in pp γ DIS 1500 DIS γp 1 γp /dy [pb] /dy [pb] /dx

100 σ

σ 1000 σ d d d [pb/GeV 2 1 1000 /dQ

50 σ 500

d 500 10-1

1.5 1.5 1.5 1.5 1 1 1 1 0.5 0.5 0.5 0.5 ratio to NLO ratio to NLO ratio to NLO ratio to NLO 0.2 0.3 0.4 0.5 0.6 0.7 5 6 10 20 300.2 0.3 0.4 0.5 0.6 0.7 0 0.2 0.4 0.6 0.8 1 y x y Q2 [GeV2] γ

Figure 6: Diffractive2 dijet ep cross sections in the photoproduction kinematic range differential Figure 4: Diffractive dijet DIS cross sections differentialinzIP , xIP , y and Q .Theinnererror bars represent the statistical errors. The outer error bars indicatein zIP , x theIP , statisticaly and xγ .Theinnererrorbarsrepresentthestatisticalerrors.The and systematic outer error bars errors added in quadrature. The overall normalisation unceindicatertainty the of 6% statisticalis not shown. and systematic NLO errors added in quadrature. The overall normalisation QCD predictions based on the H12006 Fit-B DPDF set, correcteuncertaintydtothelevelofstablehadrons, of 6% is not shown. NLO QCD predictions based on the H12006 Fit-B DPDF set are shown as a white line. They are scaled by a factor 0.83andto the account GRV forγ-PDF contributions set, corrected from to the level of stable hadrons, are shown as a white line. proton-dissociation which are present in the DPDF ﬁt butThey not arein the scaled data. by aThe factor inner,0.83 lightto account for contributions from proton-dissociation which shaded band indicates the size of the DPDF uncertainties andare presenthadronisation in the DPDF corrections ﬁt but added not in the data. The inner, light shaded band indicates the size in quadrature. The outer, dark shaded band indicates the totofal the NLO DPDF uncertainty, uncertainties also includingand hadronisation corrections added in quadrature. The outer, dark scale variations by a factor of 0.5 to 2.Foreachvariable,thecrosssectionisshownintheuppershaded band indicates the total NLO uncertainty, also including scale variations by a factor of panel, whereas the ratio to the NLO prediction is shown in0 the.5 tolower2.AvariantoftheNLOcalculationusingtheAFG panel. γ-PDF set is shown as a dashed line. For each variable, the cross section is shown in the upper panel, whereas the ratio to the NLO prediction is shown in the lower panel.

28 30 Outline

Outline 1. Event generation in PYTHIA 8 2. Photoproduction, direct and resolved processes 3. Dynamical rapidity gap survival model for hard diffraction 4. Comparison to HERA data 5. Predictions for EIC and UPCs at the LHC [Figure: S. Prestel] 6. Summary & Outlook →

2 PYTHIA 8: A general-purpose Monte Carlo event generator 1. Hard scattering • Convolution of LO partonic cross sections and PDFs 2. Parton showers • Generate Initial and Final State Radiation (ISR & FSR) 3. Multiparton interactions (MPIs) • Use regularized QCD 2 → 2 cross sections 4. Beam remnants • Minimal number of partons to conserve colour and flavour [Figure: S. Prestel] 5. Hadronization → • Lund string model with color reconnection 3 Event generation in photoproduction

Direct processes b b • Photon initiator of the hard process (DIS-like) x k p σˆ • Convolute photon flux fγ with proton PDFs fi and d l → → bp kl+X b ⊗ p 2 ⊗ γi kl xp dσ = fγ (x) f (xp, µ ) dσˆ i p remn. • Generate FSR and ISR for proton side b Resolved processes b x remn. • Convolute also with photon PDFs xγ k bp→kl+X b ⊗ γ 2 ⊗ p 2 ⊗ ij→kl dσ = fγ (x) f (xγ, µ ) f (xp, µ ) dσ l j i xp 2 p remn. • Sample x and Q , setup γp sub-system with Wγp • Evolve γp as any hadronic collision (including MPIs) 4 Comparion to HERA photoproduction data

ZEUS dijet measurement 2000 b 2 [pb] ZEUS 2 < . b

• Q 1 0 GeV obs γ Pythia 8.226 x

d resolved 1500 σ/ direct • 134 < Wγp < 277 GeV d jet1 17 < ET < 25 GeV jet1 jet2 • ET > 14 GeV, ET > 11 GeV 1000 b

, b − jet1 2 b • 1 < η < 2.4 b b 500 b Two contributions b 0 1.4 • Momentum fraction of partons in 1.3 1.2 1.1

1.0 b b photon b b jet1 ηjet1 jet2 ηjet2 0.9 b b E e + E e 0.8 b obs T T 0.7 b x = ≈ xγ ratio to Pythia 0.6 γ 0.5 2yEe 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 obs xγ • Sensitivity to process type [ZEUS: Eur.Phys.J. C23 (2002) 615-631] obs • At high-xγ direct processes dominate 5 Hard diffraction in PYTHIA 8

Direct: Hard diffraction in photoproduction b b • Process with a hard scale, desribed with a γ jet colour-neutral Pomeron (IP) exchange jet remn. • Experimentally identified from rapidity gap P Factorization of the diffractive cross section p p • Direct: Pomeron flux and diffractive PDFs Resolved: 2jets b γj→2jets IP 2 p dσ = f (x) ⊗ dσ ⊗ f (zIP, µ ) ⊗ f (xIP, t) direct γ j IP b b • Resolved: photon PDFs γ remn. jet σ2jets = b( ) ⊗ γ ( , µ2) ⊗ σij→2jets ⊗ IP( , µ2) ⊗ p( , ) d resolved fγ x fi xγ d fj zIP fIP xIP t jet remn. P p p 6 2. Reject events where MPIs in γp system (MPI) 3. Evolve γIP system, allow MPIs

Hard diffraction in PYTHIA 8

Direct: Hard diffraction in photoproduction b b • Process with a hard scale, desribed with a γ jet colour-neutral Pomeron (IP) exchange jet remn. • Experimentally identified from rapidity gap P Dynamical rapidity gap survival model p p 1. Generate diffractive events with dPDFs (PDF) Resolved: b b γ remn. jet jet remn. P p p 6 3. Evolve γIP system, allow MPIs

Hard diffraction in PYTHIA 8

Direct: Hard diffraction in photoproduction b b • Process with a hard scale, desribed with a γ jet colour-neutral Pomeron (IP) exchange jet remn. • Experimentally identified from rapidity gap P Dynamical rapidity gap survival model p p 1. Generate diffractive events with dPDFs (PDF) Resolved: 2. Reject events where MPIs in γp system (MPI) b b γ remn. jet jet remn. P p p 6 Hard diffraction in PYTHIA 8

Direct: Hard diffraction in photoproduction b b • Process with a hard scale, desribed with a γ jet colour-neutral Pomeron (IP) exchange jet remn. • Experimentally identified from rapidity gap P Dynamical rapidity gap survival model p p 1. Generate diffractive events with dPDFs (PDF) Resolved: 2. Reject events where MPIs in γp system (MPI) b b 3. Evolve γIP system, allow MPIs γ remn. jet jet remn. P p p 6 Hard diffraction in PYTHIA 8

Direct: Hard diffraction in photoproduction b b • Process with a hard scale, desribed with a γ jet colour-neutral Pomeron (IP) exchange jet remn. • Experimentally identified from rapidity gap P Dynamical rapidity gap survival model p p 1. Generate diffractive events with dPDFs (PDF) Resolved: 2. Reject events where MPIs in γp system (MPI) b b 3. Evolve γIP system, allow MPIs γ remn. jet jet Implemented from PYTHIA 8.235 onwards remn. [I.H. and C.O. Rasmussen, EPJC 79 (2019) no.5, 413] P Same idea applied for pp collisions at the LHC p p [C.O. Rasmussen and T. Sjöstrand, JHEP 1602 (2016) 142] 6 Comparisons to HERA data

H1: [EPJC 51 (2007) 549] ZEUS: [EPJC 55 (2008) 177] Cuts H1 ZEUS 12 2 2 b Data 5 Q [GeV ] 0.01 1.0 10 PDF max [pb/GeV] [pb/GeV] MPI

4 X jet1 W b

M 8 b /d E [GeV] 5.0 7.5 σ /d , b b d T min 3 b σ d 6 b b jet2

2 b Data 4 E [GeV] 4.0 6.5 b T,min PDF 1 MPI 2 b b xmax 0.03 0.025 0 0 IP 2 2 1.5 1.5 1 b b b b 1 b b b b b b PYTHIA setup MC/Data 0.5 MC/Data 0.5

170 180 190 200 210 220 230 240 15 20 25 30 35 40 45 W [GeV] MX [GeV] • dPDFs from H1 fit B LO

• PDF selection overshoots the data by 20–50 % • γPDFs from CJKL ref • Impact of the MPI rejection increases with W • pT0 = 3.00 GeV/c (Tuned to inclusive • Stronger suppression in H1 analysis due to jets charged particle data looser cuts on E and xIP ⇒More MPIs T from γp at HERA) 7 Comparisons to HERA data

H1: [EPJC 51 (2007) 549] ZEUS: [EPJC 55 (2008) 177] Cuts H1 ZEUS

500 2 2 b b

[pb] Data [pb] Data 600 Q [GeV ] 0.01 1.0 PDF PDF max obs γ obs γ

x x 400 500 MPI MPI /d /d b jet1 σ σ d d 400 300 E [GeV] 5.0 7.5 b T,min b 300 jet2 b 200 E [GeV] 4.0 6.5 200 T,min b b 100 100 b b max b x 0.03 0.025 0 2.02 IP 2 2 1.8 1.5 1.6 1.4 1 b b b b 1.2 1 b b b b b MC/Data 0 5 MC/Data 0.8 . 0.6 0.4 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 2 xobs xobs χ γ γ analysis PDF MPI

obs H1 5.2 1.4 • Stronger suppression at low-xγ (more MPIs) ZEUS 9.6 5.1 • ZEUS cuts select events at high-xobs region γ H1 & ZEUS 7.6 3.4 • Some theoretical uncertainty from γPDFs, (with all data points) dPDFs and scale variation 8 Predictions for EIC

Repeat the H1 analysis at EIC kinematics (Ee = 18 GeV, Ep = 275 GeV)

0.9 [pb] [pb] PDF

0.8 PDF γ 120 W x MPI MPI /d 0.7 /d σ σ 100 d d 0.6 80 0.5 0.4 60 0.3 40 0.2 20 0.1 0 0 2 2 1.5 1.5

1 1 Ratio Ratio 0.5 0.5

60 70 80 90 100 110 120 130 140 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 W xγ • Only up to ∼ 10% effects in the considered W range • Noticeable suppression only at low xγ where cross section small ⇒ Available energy and kinematical cuts typically applied for diffraction push the kinematics to region where no room for MPIs 9 Diffractive dijets in UPCs

• Apply the dynamical rapidity gap survival p p model to UPCs in pp and pPb γ gap • In pPb the photon flux from Pb ∝ 2 X dominates ( Z ), p neglected P gap Kinematics similar to HERA p p jet1(2) | jet1,2| • ET > 8(6) GeV, η < 4.4

• Mjets > 14 GeV, xIP < 0.025 Pb Pb gap PYTHIA setup γ • Same PDFs as for HERA X P • Vary MPI parameter: gap ref p p pT0 = 3.00 GeV (HERA γp) ref pT0 = 2.28 GeV (LHC pp) 10 Predictions for diffractive dijets in UPC √ √ pPb s = 5.0 TeV pp s = 13 TeV

0.012 4.5 ref = ref = PDF, p⊥0 3.00 GeV PDF, p⊥0 3.00 GeV 4 ref = ref = MPI, p⊥0 3.00 GeV 0.01 MPI, p⊥0 3.00 GeV ref = ref = [nb/GeV] 3.5 PDF, p⊥0 2.28 GeV [nb/GeV] PDF, p⊥0 2.28 GeV ref = ref = W MPI, p⊥ 2.28 GeV W MPI, p⊥ 2.28 GeV 3 0 0.008 0 /d /d σ σ

d 2.5 d 0.006 2 1.5 0.004 1 0.002 0.5 0 0 1.4 1.4 1.2 1.2 1 1 0.8 0.8 0.6 0.6 0.4 0.4 Ratio to PDF 0.2 Ratio to PDF 0.2 0 0 0 200 400 600 800 1000 1200 1400 0 500 1000 1500 2000 W [GeV] W [GeV] • Extended W range wrt. HERA, especially in pp (harder flux) • Stronger suppression from MPIs than at HERA ⇒ Ideal process to study factorization-breaking effects in hard diffraction 11 Summary & Conclusions

Photoproduction in PYTHIA 8 • Full simulations of direct and resolved contributions • Good description of different HERA data • Can be applied also to ultra-peripheral collisions Diffractive dijets in photoproduction • Implemented dynamical rapidity gap survival model for γp (and γγ), originally introduced for pp ⇒ Uniform framework to describe the observed factorization breaking for hard diffraction in pp and ep relying only on MPI description in PYTHIA • Support from HERA data • Only mild effects expected at EIC energies • Pronounced suppression predicted in UPCs at the LHC 12 Backup slides VOLUME 84, NUMBER 22 PHYSICALREVIEWLETTERS 29MAY 2000

The result is shown in Fig. 4. The solid curve is a fit D 2n to the data of the form F˜ jj ͑b͒ ෇ B͑b͞0.1͒ in the range ͑1023͞j͒ ,b,0.5, which corresponds to the 23 region 10 , x , 0.5jmin of Fig. 3. For our average D j of 0.065 the value of b ෇ 0.1, for which F˜ jj ෇ B, corresponds to x ෇ 0.0065, for which R˜ ෇ R0. This fit yields B ෇ 1.12 6 0.01 and n ෇ 1.08 6 0.01 with x2͞d.o.f. ෇ 1.7. The systematic uncertainty in B is 60.28, carried over from that in R0. The lower and upper boundaries of the filled band surrounding the data points represent the b distributions obtained by using only the two leading jets or up to four jets of ET . 5 GeV, respectively, in the evaluation of x. The dashed (dotted) D curve is the expectation for F˜ jj ͑b͒ calculated from fit 2 (fit 3) of the H1 diffractive structure function [1] evaluated at Q2 ෇ 75 GeV2, which approximately corresponds to jet 2 the average value of ͑ET ͒ of our data. The H1 structure function, which was derived from data in the region of j,0.04, has two terms, presumed to be due to Pomeron (IP ) and Reggeon (IR ) exchanges. Each term consists of the structure function of the exchanged Pomeron/Reggeon FIG. 3. Ratio of diffractive to nondiffractive dijet event rates as a function of x (momentum fraction of parton in p¯ ). The solid multiplied by the corresponding flux factor, f͑IP ,IR ͒͞p¯ ͑j, t͒: 2r ෇ lines are fits to the form R˜ ͑x͒ ෇ R0͑x͞0.0065͒ for b,0.5. tmin j 0.095 Motivation:F˜ D͑b͒ ෇ DiffractiveC ? f ͑j dijets, t͒ in hadronic collisions jj Z Z i i͞p¯ i෇XIP ,IR t෇21 j෇0.035 of the diffractive to ND parton densities of the antiproton, i as “viewed” by dijet production. We will denote the asso- ? Fjj͑b͒ dj dt . ciated structure functions by F ͑x͒ ෇ x͓g͑x͒ 1 4 q͑x͔͒, jj 9 [CDF: PRL 84 (2000) 5043-5048] where g͑x͒ is the gluon and q͑x͒ is the quark density; the • Dijets in ultra-peripheral 4 latter is multiplied by 9 to account for color factors. The shape of the R˜ ͑x͒ distribution exhibits no significant j dependence. A fit to all the data in the region 0.035 ,j, collisions at the LHC 23 6 0.095 yields R0 ෇ ͑6.1 6 0.1͒ 3 10 and r ෇ 0.45 6 10 2 0.02 with x ͞d.o.f. ෇ 0.76. The exponent r is insensitive ATLAS Preliminary anti-k t R=0.4 jets 2015 Pb+Pb data, 0.38 nb-1 plead > 20 GeV to systematic uncertainties in jet energy calibration, which 104 T s = 5.02 TeV mjets > 35 GeV jet b / GeV ] NN

generally depend on h . A 30% change in the SD or ND µ

[ 42 < H < 50 GeV underlying event energy values results in a 14% change A T 2 x 10 in R ; adding in quadrature an estimated 20% normaliza- ∼ σ

0 d

2 -1 T 50 < H < 59 GeV ( 10 ) d T ×

tion uncertainty yields an overall systematic uncertainty of H d 625%. Another uncertainty arises from the sensitivity of 1 -2 59 < H T < 70 GeV ( × 10 ) the parameters R0 and r to the number of jets used in evalu- ෇ -3 ating x. Using only the two leading jets yields R0 −2 70 < H T < 84 GeV ( × 10 ) ͑4.8 6 0.1͒ 3 1023 and r ෇ 0.33 6 0.02 (x2͞d.o.f. ෇ 10 84 < H < 100 GeV ( × 10 -4 ) 1.21), while by using up to four jets with ET . 5 GeV we T 23 −4 obtain R0 ෇ ͑7.0 6 0.1͒ 3 10 and r ෇ 0.48 6 0.02 10 100 < H < 119 GeV ( × 10 -5 ) (x2͞d.o.f. ෇ 0.74). About 48% (23%) of the SD (ND) T events have no jets of ET . 5 GeV, other than the two −6 -6 10 119 < H < 141 GeV ( × 10 ) 23 T leading jets; for these events R0 ෇ ͑9.6 6 0.2͒ 3 10 2 ෇ ෇ -7 and r 0.31 6 0.03 (x ͞d.o.f. 1.18). 141 < H < 168 GeV ( × 10 ) −8 T The diffractive structure function of the antiproton is FIG. 4. Data b distribution (points) compared with expecta- 10 Data tions from the parton densities of the proton extracted from Pythia+STARlight obtained from the equation 168 < H < 200 GeV ( × 10 -8 ) diffractive deep inelastic scattering by the H1 Collaboration. The scaled to data T • A significant 2suppressionn of −10 ˜ D ෇ ˜ ෇ ˜ ND straight line is a fit to the data of the form b . The lower (up- 10 Fjj ͑b͒ R͑x bj͒ 3 Fjj ͑x ! bj͒ . per) boundary of the filled band represents the data distribution Not unfolded for detector response ˜ D 2 obtained by using only the two leading jets (up to four jets of We have evaluated Fjj ͑b͒ for jtj , 1 GeV , diffractive dijets observed in p+p 10−12 ET . 5 GeV) in evaluating b. The dashed (dotted) lines are −3 −2 −1 0.035 ,j,0.095, and ET ͑ jet1, jet2͒ . 7 GeV using expectations from the H1 fit 2 (fit 3). The systematic uncer- 10 10 10 1 the GRV98LO parton density set [13] in F˜ ND͑x ! bj͒. tainty in the normalization of the data is 625%. x jj • Similar results also at the LHC A 5047 [ATLAS-CONF-2017-011] PDFs for resolved photons

Comparison of different photon PDF analysis

Q2 = 10.0 GeV2 CJKL Q2 = 10.0 GeV2 CJKL u-quark GRV gluon GRV SaSgam SaSgam EM EM /α /α ) ) 2 2 x, Q x, Q ( ( xf xf

x x • Some differences between analyses, especially for gluon ⇒ Theoretical uncertainty for resolved processes • CJKL used as a default in PYTHIA 8, others via LHAPDF5 but only for hard-process generation MPIs with resolved photons

Parametrization for γp γγ γp • pT0 values between γγ (using LEP pp data) and pp ) [GeV/c] s √ (

• Relevant energies: T0 p

• HERA: Wγp ≈ 200 GeV ≈ √ • eRHIC: Wγp 100 GeV s [GeV]

Number of MPIs in different RHIC: pp 200 GeV HERA: ep 300 GeV colliders eRHIC: ep 145 GeV • Non-diffractive events with [A.U.] resolved photons • Less MPIs in ep than pp

• Larger pT0 Number of interactions/Event • Point-like PDF in PS Charged particle pT spectra in ep collisions at HERA

H1 measurement 3 10 b b b

[nb] b b b H1 = = . b • E 820 GeV, E 27 5 GeV 2 T b p e b p 2 b Pythia 8.226 d 10 b η b b d resolved ≈ b • < W > 200 GeV b γp σ/ b direct 2 1 b ref d 10 b pT0 =3.00 GeV/c 2 2 b < . b |η| < 1 • Qγ 0 01 GeV b b

1 b b Comparison to PYTHIA 8

b −1 10 b • Resolved contribution dominates

b 10−2 • Good agreement with the data using 1.6 1.4 b pref = 3.00 GeV/c b T0 1.2 b b b b b b b b b b b b b b b b b b 1 b b b b b ⇒ 0.8 b MPI probability between pp and γγ 0.6 0.4 ratio to Pythia 0.2 0 2 4 6 8 10 12 pT [GeV/c]

[H1: Eur.Phys.J. C10 (1999) 363-372] Charged particle pT spectra in ep collisions at HERA

H1 measurement 4 10 b [nb] H1 ref 2 • E = 820 GeV, E = 27.5 GeV 2 T 3 p e

p b pT,0 =2.28 GeV, χ /n = 9.90 10 b d b ref 2 b

η b b pT,0 =2.70 GeV, χ /n = 1.85 d b ref 2 ≈ b • < W > 200 GeV 2 b pT 0 =3.00 GeV, χ /n = 0.79 γp σ/ b , 10 b 2 b ref 2

d b pT,0 =3.30 GeV, χ /n = 1.69 b 2 2 2 b 1 b MPI oﬀ, χ /n = 2.48 • Qγ < 0.01 GeV b 10 b b

b b Comparison to PYTHIA 8 1 b b b −1 • Resolved contribution dominates 10 b

b −2 10 • Good agreement with the data using 1.4 ref 1.3 p = 3.00 GeV/c 1.2 T0 1.1 1 bbbbbbbbbbbbbbbbbbbb b b b b b b ⇒ γγ 0.9 MPI probability between pp and

MC/Data 0.8 0.7 0.6 0.5 2 4 6 8 10 12 pT [GeV/c]

[H1: Eur.Phys.J. C10 (1999) 363-372] Charged particle pT spectra in ep collisions at HERA

H1 measurement 4 10 b [nb] H1 ref 2 • E = 820 GeV, E = 27.5 GeV 2 T 3 p e

p b pT,0 =2.28 GeV, χ /n = 9.90 10 b d b ref 2 b

η b b pT,0 =2.70 GeV, χ /n = 1.85 d b ref 2 ≈ b • < W > 200 GeV 2 b pT 0 =3.00 GeV, χ /n = 0.79 γp σ/ b , 10 b 2 b ref 2

d b pT,0 =3.30 GeV, χ /n = 1.69 b 2 2 2 b 1 b MPI oﬀ, χ /n = 2.48 • Qγ < 0.01 GeV b 10 b b

b b Comparison to PYTHIA 8 1 b b b −1 • Resolved contribution dominates 10 γγ b

γp b −2 • Good agreement with the data using 10 pp 1.4 ref 1.3 p = 3.00 GeV/c 1.2 T0 1.1 1 bbbbbbbbbbbbbbbbbbbb b b b b b b ⇒ ) [GeV/c] MPI probability between pp and γγ

s 0.9

MC/Data 0.8 √

( 0.7 0.6 T0 0.5 p 2 4 6 8 10 12 pT [GeV/c]

√ s [GeV] Charged-particle η dependence

3000 800

b b

[nb] H1 [nb] H1 700 η 2500 Pythia 8.226 η Pythia 8.226 /d /d

σ resolved σ 600 resolved d d 2000 direct direct

> b 500 > pT 2.0 GeV/c b b b pT 3.0 GeV/c b b b b b b 1500 b 400

b b b b b b b b 300 b 1000 200 500 100

0 0 1.6 1.6 1.4 1.4

1 2 b 1 2 . b b b . b b b b b b b b b b b 1 1 b b b b b 0.8 0.8 0.6 0.6 0.4 0.4 ratio to Pythia 0.2 ratio to Pythia 0.2 0 0 -1 -0.5 0 0.5 1 -1 -0.5 0 0.5 1 η η

[H1: Eur.Phys.J. C10 (1999) 363-372] • Good agreement also for charged-particle η dependence • Resolved contribution dominates the cross section Dijet in ep collisions at HERA

Pseudorapidity dependence of dijets [Eur.Phys.J. C23 (2002) 615-631]

700

b b [pb] ZEUS [pb] 350 ZEUS 600 CJKL b CJKL jet2 jet2

η η 300

d GRV d GRV 500 b σ/ SaSgam σ/ SaSgam d obs d 250 b obs

xγ < 0.75 b b xγ > 0.75 400 1 < ηjet1 < 2.4 1 < ηjet1 < 2.4 b 200

b 300 b 150

b 200 100 b b 100 50 b

b b 0 0 1.4 1.4 1.3 1.3 1.2 1.2 1.1 1.1 b 1.0 b 1.0 b b b b b b b b 0.9 0.9 b b b 0.8 0.8 b 0.7 0.7 ratio to0.6 CJKL ratio to0.6 CJKL 0.5 0.5 -1 -0.5 0 0.5 1 1.5 2 -1 -0.5 0 0.5 1 1.5 2 ηjet2 ηjet2 • Simulations tend to overshoot the dijet data by ∼10 % ∼ obs • 10 % uncertainty from photon PDFs for xγ < 0.75 Predictions for dijets in UPCs

Event selection similar to HERA √ PbPb, sNN = 5.5 TeV Pythia anti-kT,R = 0.4 NNPDF2.3 = . lead EPPS16 • anti-kT with R 0 4 pT > 8 GeV/c mjets > 14 GeV resolved lead jets direct • pT > 8 GeV, pT > 6 GeV [nb]

jets A | | x • η < 4.4, mjets > 14 GeV d σ/ • Event-level variables: d m − = Σ = √jets yjets • HT ipTi, xA s e Results from PYTHIA 8

• Resolved dominant at high-xA, direct at low-xA L = 1 nb−1 5.0 TeV L = 13 nb−1 5.5 TeV • Sensitive to nuclear PDFs Ratio to NNPDF2.3 xA • Statistical uncertainty estimated at [I.H., arXiv:1811.10931] different luminosities [also Guzey, Klasen, arXiv:1902.05126] • Hard diffraction in DIS

e e Diffractive dijets γ∗ jet • Virtual photon interacts with Pomeron zP jet from proton producing jets remn. P xP • Signature: scattered proton or a rapidity p t p gap between proton and Pomeron remnant

Factorized cross section for diffractive dijets 2jets+X IP 2 ⊗ p ⊗ ie→2jets • DIS: dσ = fi (zIP, µ ) fIP(xIP, t) dσ p IP where fIP is Pomeron flux and fj diffractive PDF (dPDF) • Factorization verifed by H1 and ZEUS at HERA obs • Better agreement for the shape of zIP with ZEUS-SJ

Theoretical uncertainties

Largest uncertainties arise from • LO ME (vary factorization and renormalization scales) • diffractive PDFs (H1fitB, ZEUS-SJ and GKG18A) ZEUS 2008: ZEUS 2008:

12 500

b b

Data [pb] Data 10 central central obs P

z 400

[pb/GeV] Combined scale uncertainty Combined scale uncertainty /d X σ M

8 d 300 /d b b σ

d 6 b

200 b b 4 b b

b 100 b 2 b

0 0 2 2 1.5 1.5

1 b b b b b b 1 b b b b

MC/Data 0.5 MC/Data 0.5

15 20 25 30 35 40 45 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 obs • Scale uncertainty aroundMX [GeV] 20 % zP Theoretical uncertainties

Largest uncertainties arise from • LO ME (vary factorization and renormalization scales) • diffractive PDFs (H1fitB, ZEUS-SJ and GKG18A) ZEUS 2008: ZEUS 2008:

12 500

b b

Data [pb] Data 10 GKG LO Fit A GKG LO Fit A obs P

z 400

[pb/GeV] ZEUS NLO Fit SJ ZEUS NLO Fit SJ /d X

H1 LO Fit B σ H1 LO Fit B M

8 d 300 /d b b σ

d 6 b

200 b b 4 b b

b 100 b 2 b

0 0 2 2 1.5 1.5

1 b b b b b b 1 b b b b

MC/Data 0.5 MC/Data 0.5

15 20 25 30 35 40 45 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 obs • Scale uncertainty aroundMX [GeV] 20 % zP obs • Better agreement for the shape of zIP with ZEUS-SJ obs zIP distributions

H1 2007: ZEUS 2008: 500

b [pb] 500 [pb] Data PDF obs P obs P z

z 400 MPI /d 400 /d σ σ d d b 300 b 300 b

200 b b

200 b Data b b PDF 100 b 100 MPI

0 0 2 2 1.5 1.5

1 b b b b 1 b b b b MC/Data 0.5 MC/Data 0.5

0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 obs obs zP zP

obs • MPI suppression not dependent on zIP • Better agreement with H1 data after MPI rejection • Shape a bit off in both cases, observable sensitive to • dPDFs, Jet reconstruction 12

applied to account for theﬁnite resolution of the reconstructed variables used in the analysis. They are evaluated with POMWIG,PYTHIA8 4C and PYTHIA8 CUETP8M1. The average between the results is taken as the nominal value in the analysis. The measured cross sections are obtained by unfolding the data using the D’Agostini method with early stopping [38]. In this method the regularisation parameter is the number of iterations used, which is optimized to obtain a relativeχ 2 variation between iterations lower than 5%. Figure 4 shows the differential cross section as a function oft andξ, integrated over the con- Diffractivejugate variable. dijets in The pp results from events in which the proton is detected in either side of the interaction point are averaged.

37.5 nb-1 (8 TeV) 37.5 nb-1 (8 TeV) 4 CMS+TOTEM Preliminary CMS+TOTEM Preliminary ) 10 2 Data 2

POMWIG ( ~~= 1) (nb)~~

~~2 ξ d POMWIG ( ~~= 7.4%) σ d 3 103 PYTHIA8 4C 10 (nb/GeV PYTHIA8 CUETP8M1 σ d~~~~

~~dt PYTHIA8 DG 0.6 GeV ±-2Exp. fit b = 6.6 2 102 10~~

H1 fit B Data 2 j1j2 POMWIG ( = 1) H1 fit B p > 40 GeV 2 10 10 T j1j2 POMWIG ( ~~= 7.4%) pj1j2 > 40 GeV |η| < 4.4 T j1j2 PYTHIA8 4C |η| < 4.4 < 0.1ξ0 < PYTHIA8 CUETP8M1 2 < 0.1ξ0 < 0.03 < |t| < 1.0 GeV PYTHIA8 DG 2 1 0.03 < |t| < 1.0 GeV 1~~

~~2 2~~

~~1 1.5~~

~~MC/Data 0 MC/Data 1~~

~~0.3 2 Data/POMWIG ( = 1) 0.1 Data/POMWIG ( ~~= 1) 0.2 0.08 0.1~~~~

Data/MC 0 Data/MC 0.06 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.02 0.04 0.06 0.08 0.1 -t (GeV2) ξ • Dynamical rapidity gap survival model in PYTHIA 8 (DG) provide a good Figure 4: Differential cross section as a function oft (left) and as a function ofξ (right) for single- diffractivedescription dijet production,of the measurement compared to the (Survival predictions probability from POMWIG< 10,P%YTHIA) 8 4C, PYTHIA8 CUETP8M1 and PYTHIA8 Dynamic Gap (DG). POMWIG is shown with no correction for the rapidity gap survival probability ( S2 = 1) and with a correction of S2 = 7.4%. The vertical bars indicate the statistical uncertainties and the yellow band indicates the total systematic � � � � uncertainty. The average of the results for events in which the proton is detected in either side of the interaction point is shown.

The data are compared to POMWIG,PYTHIA8 4C, PYTHIA8 CUETP8M1 and PYTHIA8 Dy- namic Gap (DG). POMWIG is shown for two values of the suppression of the diffractive cross section, i.e. the rapidity gap survival probability, represented by S2 . When S2 = 1, no correction is applied. The resulting cross sections are higher than the data by roughly an order � � � � of magnitude, in agreement with the Tevatron results [5–7]. POMWIG is also shown with the correction S2 = 7.4%, calculated from the ratio of the measured diffractive cross section and the MC prediction, as discussed below. After this correction, POMWIG gives a good description � � of the data. POMWIG is shown in Fig. 4 as the sum of the Pomeron (pIP), Reggeon (pIR) and Pomeron-Pomeron (IPIP) exchange contributions while PYTHIA8 includes only the Pomeron (pIP) contribution. PYTHIA8 4C and PYTHIA8 CUETP8M1 show cross sections higher than the data by up to a factor of two. The PYTHIA8 Dynamic Gap model shows overall a good agreement with the data. No correction is applied to the normalisation of the PYTHIA8 sam- ples.

~~The ratio of the data yields and the POMWIG predictions is shown in the bottom of the left Ultra-peripheral collisions (UPCs) (main70.cc)~~

~~Photon flux from protons 105 leptons 4 • Take the proton form factor into account 10 protons Pb nucleus [ ] 103~~

~~2 M α ( + ( − ) ) E p em 1 1 x 11 3 3 1 2 fγ (x) = log(A) − + − + / 10 2 3 * π )~~

~~2 x 6 A 2A 3A x 1~~

~~( 10 i f x 0 where A = 1 + Q2 /Q2 and Q2 = 0.71 GeV2 10 0 min 0 10 1~~

~~10 2 • The form factor suppress contribution from 10 4 10 3 10 2 10 1 100 x high-Q2 ⇒ photoproduction regime UPCs with heavy ions~~

 • Define photon flux in impact-parameter space     b > 2RA to reject events where colliding nuclei overlap    [ ]  ( )  2α Z2 ξ2  f A(x) = EM ξ K (ξ)K (ξ) − K2(ξ) − K2 (ξ) γ xπ 1 0 2 1 0

~~where Z charge, ξ = bminxm Soft QCD photoproduction~~

~~Soft QCD process implemented for photoproduction Elastic ρ production at ⟨ ⟩ • Based on Schuler and Sjöstrand model in PYTHIA Wγp = 70 GeV ] 2~~

~~b b 6 2 b Data 10 b b Pythia 8 b/GeV µ [ b | t • Vector meson dominance (VMD) with ρ, ω, ϕ and | b~~

~~/d b ρ el σ 1 b d 10 J/Ψ mesons for b~~

~~• Soft diffraction (high- and low-mass) 1~~

~~1.4 1.3 • Elastic scattering 1.2 1.1 1 b b b b b b b b b b 0.9~~

MC/Data 0.8 • Non-diffractive from MPI machinery 0.7 0.6 0.5 0 0.1 0.2 0.3 0.4 0.5 • Total Cross section parametrized as |t| [GeV2] − [Data from ZEUS: σAB(s) = XAB sϵ + YAB s η tot Z.Phys. C69 (1995) 39-54] where ϵ = 0.0808 and η = 0.4525 are universal, XAB and YAB process-dependent