HERA for LHC Roberto Carlin University of Padova

HERA for LHC Roberto Carlin University of Padova and INFN-PD R. Carlin 2006 Aspen Winter Conference

LHC: pp collider, √s = 14 Te. V Discovery machine: Higgs, Susy. . But, it is a parton-parton collider: need the structure of the proton and QCD to get precise results, understand the backgrounds to new physics, measure the luminosity. . Need to control QCD effects at the best R. Carlin 2006 Aspen Winter Conference

HERA: ep collider, √s = 320 Ge. V Mostly a QCD machine. . Parton densities, jets, hard diffraction, heavy flavors production… The photon can be used as an hadronic probe (“resolved” photon in photoproduction): multiple intraction, gap survival in diffraction… HERA is the ideal machine to make precision QCD studies and test new QCD predictions R. Carlin 2006 Aspen Winter Conference

Topics for this talk • What is the impact of the measurements done so far (HERA-I)? • Can we do better in HERA-II for what concernes LHC? • What should be measured really to have an impact? • Structure functions and parton distribution functions. • Diffraction: diffractive PDFs and generalised parton distributions. • Multiple interactions. • Heavy flavor production. Very detailed documentation in the proceedings of the HERA-LHC workshop, recently issued as CERN-2005 -014 yellow report Apologies to the original authors for the plots without the proper reference R. Carlin 2006 Aspen Winter Conference

Parton distributions: HERA results PDFs fits without and with HERA data at different Q 2: uncertainty vastly reduced, particularly at low-x. NB, same fit, sameparametrization and DGLAP evolution, just no HERA data R. Carlin 2006 Aspen Winter Conference

Example: the effect of HERA PDFs in the determination of the W+ production at LHC 0. 0001<x<0. 1 Process is important for the measurement of the luminosity in LHC W+ rapidity • Uncertainty from a specific PDF fit goes from ~16% to ~3. 5% using HERA data • Low- to midle-x gluon knowledge is essential • Is our knowledge of PDFsit good enough? R. Carlin 2006 Aspen Winter Conference W+ rapidity

Example of differences between PDFs Results of H 1 and ZEUS are compatible But there are clear differences, particularly at low-x in the gluon. The differences are mostly from the dat sets, not from the fits and prameterizations NB, the gluon is not directly measured in HERA, enters in the PDFs from F 2/ ln. Q 2 use exclusive measurements senitive to the gluons (jets, charm…) R. Carlin 2006 Aspen Winter Conference

• Spread of existing PDFs gives 10% uncertainty on Higgs cross sections at LHC • Same in the W production (8%) • Need to do better than this R. Carlin 2006 Aspen Winter Conference Djouadi & Ferrag

Not only experimental data, but also QCD calculations need progress (NNLO…) for the PDF fits Large evolution from HERA to LHC, to high-Q 2 and low-x • Is NLO (NNLO) DGLAP good enough? • Collinear factorization, kt effects • BFKL ln(1/x) resummation • Saturation, non linear evolution. Precise data from HERA should help models to disentangle these effects R. Carlin 2006 Aspen Winter Conference

How can we improve with HERA-2 the uncertainty in the PDFs? HERA-2 is expecting an integrated luminosity ~700 pb-1 (H 1 and ZEUS have ~130 pb-1 on tape from HERA-1). Various improvements are possible: • Use Jet data in the fits, with higher statistics. • Increased statistics on high Q 2 data. • Measure FL. Improvements are also possible with present data: • Improved systematics • Combined data fits R. Carlin 2006 Aspen Winter Conference

Inclusion of DIS Jet and High-Et dijet photoproduction data in the PDFs fits gluon fractional error C. Gwenlan, A. Cooper-Sarkar, C. Targett-Adams Improves the mid- to high-x gluon distributions, up to high scales Allowes to fit without the low energy heavy target data (target corrections, systematics. . ) The use of high statitics HERA -2 data brings a significative improvement Even better using full statistics AND optimised jet cross sections measurements designed to optimise the sensitivity to the gluon density R. Carlin 2006 Aspen Winter Conference

At high-Q 2, high-x the NC/CC measurement is still limited by statistics Expect impact in the valence quarks distributions The measurements at low-Q 2 are already limited by the systematic uncertainties O(23%) • H 1 is planning to reduce the systematics at low-Q 2 using HERA-I yr 2000 data. • Goal is to reach 1% level, will effect mostly low-x gluon. R. Carlin 2006 Aspen Winter Conference

Simulation: 500 pb-1 in the e+ and e- channels Improvement up to a factor 2 in the valence. NB, this is a ZEUS fit with jets cross sections and no low energy data, results still improves w. r. t. global fits There also improvements in x. F 3~Σ(q-qbar), up to a factor of three R. Carlin 2006 Aspen Winter Conference

Impact on LHC from improved PDFs (I) R. Carlin 2006 Aspen Winter Conference

Impact on LHC from improved PDFs (II) Uncertainty on jet cross section at LHC significantly reduced R. Carlin 2006 Aspen Winter Conference

Measurement of FL M. Klein FL~ sxg(x, Q 2) direct measurement of the gluon Can be performed with a set of lowered p energies at HERA-2 (~10 pb-1): • σr~F 2(x, Q 2)-f(y)·FL(x, Q 2); y=Q 2/sx • measure cross sections at high-y, and change y at fixed x, Q 2 ( change s) R. Carlin 2006 Aspen Winter Conference

Result of a simulation (M. Klein): • Can get ΔFL ~ 10 -20% • The measurement of FL can have an impact, mostly in the low-x gluon, low-Q • Important test of QCD consistency • HERA is the only place to mesure it. • Trade off with luminosity lost at high energy • Luminosity goes as Ep 2 • Setup the machine • Problem: are the present detectors well tuned to the measurement (low-x, low. Q 2)? R. Carlin 2006 Aspen Winter Conference

Overall improvement from HERA-II statistics R. Carlin 2006 Aspen Winter Conference

Combining H 1 and ZEUS present data Using H 1 and ZEUS data together in standard QCD fits. The improvement is marginal, errors are dominated by systematics, the data sets are not incompatible but pull against each other… R. Carlin 2006 Aspen Winter Conference

Different method (Glazov). Global (Hessian) fit of H 1 and ZEUS cross section data. The idea is that we are measuring the same thing, and the systematics of one experiment can be solved by the other, taking into account correlations. Indeed, systematics are quite different between ZEUS and H 1. Example on how the combined “HERA” data set after the fit is better than each individual point Beware, still preliminary, all pitfalls have to be worked out It is a duty of the HERA community to provide the best PDFs R. Carlin 2006 Aspen Winter Conference

Diffractive Higgs production at LHC 2 -10 fb O(100) fb fi(z, XIP, t, Q 2) • Clean signal, very good mass resolution for the exclusive process • Need knowledge of the diffractive PDFs and GPD R. Carlin 2006 Aspen Winter Conference

HERA is measuring DVCS and eleastic VM production (ρ, φ, J/Ψ, Υ) which are in principle sensitive to GPDs • VM data sensitive to gluon distribution at x~(MV 2+Q 2)/W 2 (neglecting skeweness of the two gluons) R. Carlin 2006 Aspen Winter Conference

Example, exclusive J/Ψ production in DIS • Smallest theoretical error in DVCS and Υ • Rates are low, need HERA-2 R. Carlin 2006 Aspen Winter Conference

Diffraction at HERA • ~10% of DIS NC events are diffractive (rapidity gap) • Hard process, should be described by QCD (2 g exchange) • QCD factorization: a hard process and a structure function F 2 D(4)(x. IP, t, β, Q 2). X • Measure the diffractive PDFs • Background to the exclusive diffraction at LH • Understand the gap survival probability R. Carlin 2006 Aspen Winter Conference

First compilation of HERA F 2 D(3) Different methods: • tagged proton: LPS, FPS (clean but low statistics) • rapidity gap (H 1) • Mx method (Zeus), based on the distribution of ln. MX 2 Result: • Reasonable agreement between H 1 and ZEUS • There are regions of significant discrepancies R. Carlin 2006 Aspen Winter Conference

NLO QCD fit Gluon dominates over the singlet Almost a factor 2 difference in the gluon between Zeus and H 1 More work needed. More data will come from FPS (H 1) and exclusive processes R. Carlin 2006 Aspen Winter Conference

Diffractive PDFs: are they process independent functions (hard scattering factorization)? QCD fit to inclusive diffractive data can predict the exclusive diffractive dijets production in DIS. Same for diffractive DIS charm production. Some confidence in the process. R. Carlin 2006 Aspen Winter Conference

HERA diffractive PDFs fail to describe diffractive hadron collisions at the Tevatron. To be understood and quantified to export results to LHC! Factorization is broken by the interctions of the spectator quarks (multiple interactions). Soft interactions between the spectator fill the gap (gap survival probability reduced) R. Carlin 2006 Aspen Winter Conference

At HERA we can have hadron-hadron interactions: “resolved” photon in photoproduction, where the photon hadronizes before the hard scattering. So we should see that the rapidity gap fills in these interactions NB, at HERA we can switch ON and OFF the hadron-hadron studying direct vs resolved R. Carlin 2006 Aspen Winter Conference

For jets in photoproduction, the NLO fit overestimates the data by a factor 2 in both “resolved” and “direct” photon regimes, contrary to what expected from “gap survival” models (suppression in resolved only). More work is needed. Use ratio beween diffractive and inclusive Jets? R. Carlin 2006 Aspen Winter Conference

Multiple Interactions MI (additional partonic remnant-remnant interactions) will certainly be an issue at LHC. Has HERA anything to say about that? R. Carlin 2006 Aspen Winter Conference

4 jets event in photoproduction at HERA • • High order photoproduction process At low xγ (“resolved”) it can be interpreted as hadron-hadron collisions • The inclusive sample is well described by standard PS models + MI (with hard scattering contribution to the underlying event) • Not conclusive, but suggestive R. Carlin 2006 Aspen Winter Conference

Multiple Interaction, Diffraction, saturation relationship: AGK cutting rule in QCD Bartels, Kowalski, Sabio-Vera Diffraction Saturation MI R. Carlin 2006 Aspen Winter Conference

Try to find observables sensitive to non-DGLAP evolution: Measure forward jets in DIS at HERA Study non-DGLAP QCD evolution at low-x: • low xbj • xjet>>xbj • E 2 T, jet~Q 2 • k. T ordering (DGLAP, DISENT) • x ordering (BFKL) • angular ordering (CCFM) R. Carlin 2006 Aspen Winter Conference

Standard p. QCD gives poor description at low-x forward jets, but hadronization and scale uncertainties are large • Need more statistics and phenomenological progress to improve the understanding x R. Carlin 2006 Aspen Winter Conference

Open Heavy Quark Production at HERA, impact on LHC • At very high-Q 2, heavy quarks distributions are of size comprable to light sea quarks – Impact on lumi processes (Z production) • The b quarks PDFs enter in a number of interesting process at LHC (as signal or background) – Higgs production – Searches R. Carlin 2006 Aspen Winter Conference

• 5% difference in the pp z production if b PDFs are neglected • Need to know F 2 b to 20% accuracy to get Z production at 1% level R. Carlin 2006 Aspen Winter Conference

open b production dijet in photoproduction open b production, jet in DIS • Presently the b production cross section at HERA is relatively well described by NLO QCD, some discrepancy in selected regions • No more a factor 2 -3 of overall excess ! • Still limited statistics from HERA-I (b into μ channel) R. Carlin 2006 Aspen Winter Conference

• H 1 has shown the first F 2 b measurements using the lifetime; still limited in statistics (57 pb-1) • In HERA-II ZEUS has a new microvertex (MVD) • The increased statistics and the use of vertex detectors will improve significantly (order of magnitude) the heavy flavours measurements at HERA-II R. Carlin 2006 Aspen Winter Conference

Conclusions • The HERA experiments have performed many measurements essential for LHC • Many more have to come with HERA-II • Many things I did not mention: – Lot of theoretical/phenomenological developments in QCD triggered by HERA – MC tools developments – Design of very forward region (leading proton tagging) • HERA will stop in July 2007, still lot of data to come, we are getting ready for a prompt analisys R. Carlin 2006 Aspen Winter Conference