The LHeC Project(Large Hadron-electron Collider)

Cyrille Marquet

Centre de Physique Théorique

Ecole Polytechnique

Contents

• Accelerator design

• Detector considerations

• Small-x and e+A physics

• Other SM and BSM physics

Accelerator design(slides stolen from P. Newman)

• Previously considered as `QCD explorer’ (also THERA)

• Main advantages: low interference with LHC, high and stageable Ee, highlepton polarisation, LC relation?

• Main difficulties: obtaining high positron intensities, no previous experience exists

• First considered (as LEPxLHC)in 1984 ECFA workshop

• Main advantages: high peaklumi, tunnelling (mostly) exists

• Main difficulties: building round existing LHC, e beam energy and lifetime limited by synchrotron radiation

LINAC-RING

RING-RING

How to do DIS using the LHC ?while allowing simultaneous ep(eA) and pp(AA) running

design constraint: power < 100 MW Ee = 60 GeV @ 1033 cm-2 s-1

• Two 10 GeV linacs, • 3 returns, 20 MV/m• Energy recovery insame structures[CERN plans energy recovery prototype]

• ep Lumi ~ 1033 cm-2 s-1

corresponds to ~10 fb-1 per year (~ 100 fb-1 total) • eD and eA collisions have always been integral to programme• e-nucleon Lumi estimates ~ 1031 (1032) cm-2 s-1 for eD (ePb)

Baseline Design (electron Linac)

Design parameter summaryelectron beam RR LR LRe- energy at IP[GeV] 60 60 140luminosity [1032 cm-2s-1] 17 10 0.44polarization [%] 40 90 90bunch population [109] 26 2.0 1.6e- bunch length [mm] 10 0.3 0.3bunch interval [ns] 25 50 50transv. emit. x,y [mm] 0.58, 0.29 0.05 0.1

rms IP beam size x,y [m] 30, 16 7 7

e- IP beta funct. *x,y [m] 0.18, 0.10 0.12 0.14

full crossing angle [mrad] 0.93 0 0geometric reduction Hhg 0.77 0.91 0.94

repetition rate [Hz] N/A N/A 10beam pulse length [ms] N/A N/A 5ER efficiency N/A 94% N/Aaverage current [mA] 131 6.6 5.4tot. wall plug power[MW] 100 100 100

proton beam RR LRbunch pop. [1011] 1.7 1.7tr.emit.x,y [m] 3.75 3.75

spot size x,y [m] 30, 16 7

*x,y [m] 1.8,0.5 0.1

bunch spacing [ns] 25 25

RR= Ring – RingLR =Linac –Ring

Include deuterons (new) and lead (exists)

10 fb-1 per yearlooks possible

… ~ 100 fb-1 total

From 2012 Chamonix LHC

performance workshop summary

[See also NuPeCC long range

plan]Current mandate from CERN is to aim for TDR by ~ 2015.… requires detailed further study and prototyping of accelerator components (including CERN ERL LHeC test facility), but also an experimental collaboration to develop the detector concept

How and when might LHeC fit ?

Currently approved future of high-energy DIS

On LHeC @ IP2 after LS3

• Rather it seems that LHeC is seen as a threat by ALICE

(from the naïve perspective of a theorist)

don’t rush into a decision, thoroughly consider both options

• When I first heard of it I thought: what a fantastic opportunity for the ALICE community!They will embrace this project and secure heavy-ion physics at the LHC for many years beyond LS3

I was naïve … this was not the predominant response

and they will likely fight the project

• Embrace or fight the project ?

Detector considerations(slides stolen from P. Newman)

Access to Q2=1 GeV2 in ep mode for all x > 5 x 10-7 requires scattered electron acceptance to 179o

Similarly, need 1o acceptancein outgoing proton directionto contain hadrons at high x(essential for good kinematicreconstruction)

Detector acceptance requirements

Forward/backward asymmetry in energy deposited and thus in geometry and technologyPresent dimensions: LxD =14x9m2 [CMS 21 x 15m2 , ATLAS 45 x 25 m2]Taggers at -62m (e),100m (γ,LR), -22.4m (γ,RR), +100m (n), +420m (p)

e p

Overview (full acceptance version)

EM Calorimeter

[encased in3.5T solenoid

field]

Transverse momentumΔpt/p2

t 6 10-4 GeV-1

transverseimpact parameter 10μm

• Full angular coverage, long tracking region 1o acceptance• Several technologies under discussion

Tracking region

Liquid Argon EM Calorimeter [accordion geometry, inside coil] Barrel: Pb, 20 X0 , 11m3 FEC: Si -W, 30 X0 BEC: Si -Pb, 25 XHadronic Tile Calorimeter [modular, outside coil: flux return]

Calorimeters

A GEANT4 simulated high-x event

In the absence of a detailed simulation set-up, simulated `pseudo-data’ produced with reasonable assumptions on systematics (typically 2x better than H1 and ZEUS at HERA).

Assumed systematic presicion

Small-x and e+A physics

Deep inelastic scattering (DIS)

e+A @ EIC e+Pb @ LHeC

*A center-of-mass energyW2 = (q+p)2

photon virtualityQ2 = - (k-k’)2 = - q2 > 0

What we know about small x• fundamental consequence of QCD dynamics:

at asymptotically small x:- QCD evolution becomes non-linear- particle production becomes non-linear- QCD stays weakly coupled

• the energy dependence of the saturation scale, and more generally of observables, can be computed from first principles

although in practice, the predictivity will depend on the level of accuracy of thecalculation (LO vs NLO, amount of non-perturbative inputs needed, …)

both in terms of practical applicabilityand phenomenological success

• the Color Glass Condensate (CGC) has emerged as the best candidateto approximate QCD in the saturationregime

A big open question• is this relevant at today’s colliders ?

- for each of these observables, there are alternatives explanations

- the applicability of the theory can be questioned when values of QS start to drop below 1 GeV (e.g. p+p and peripheral d+Au at RHIC)

• the CGC is not widely accepted because

in other words: can we get away with using sucha gluon distribution (with ad hoc cutoff if necessary) ?

or do we need to properly take into account

the QCD dynamics at kT ~ QS and below ?

the CGC phenomenology is successfulfor every collider process that involves

small-x partons and kT ~ QS , i.e. for abroad range for high-energy observables:

multiplicities in p+p, d+Au, Au+Au and Pb+Pb; forward spectra and correlationsin p+p and d+Au; total, diffractive and exclusive cross sections in e+p and e+A, …

Bigger open question

• the impact parameter dependence of the gluon density and of QS

what is done in the most advanced CGC phenomenologicalstudies, is to treat the nucleus as a collection of Woods-Saxondistributed CGCs, and to evolve (down in x) the resulting gluon

density at different impact parameters independently

but is this good enough ? (in principle not)

this has always been the main non-perturbative input in CGC calculations

modeling

in the case of a proton, using an impact-parameter averaged saturationscale is enough most of the time, but in the case of a nucleus it is not

Why QGP physicists (should) care• bulk observables in heavy-ion collisions reflect the properties of the initial state as much as those of the hydro evolution of the QGP

new sources of uncertainties keepemerging, for instance even two CGCmodels predict different eccentricities

• the main source of error in the extraction of medium parameters (e.g. η/s) is our insufficient understanding of initial state fluctuations

QGP properties cannot be precisely extracted from data without a properunderstanding of the initial state; e+A collisions: access to a precise picture

Schenke, Tribedy, Venugopalan

Inclusive structure functions

NLO DGLAP cannot simultaneously accommodate F2 and FL

LHeC data if saturation sets in according to current models

precisely measuring FL is crucial, and this requires an e+A energy ( ) scan

Albacete

measures quark distributions gluon distribution

@ LHeC

Exclusive Vector Meson production

through a Fourier transformation, one canextract the spatial gluon distribution (andcorrelations), this is not feasible in p+A

energy dependence

momentum transfer dependence

Cold nuclear matter effects• hard probes (esp. jets) in heavy-ion collisions need calibration

what is the effect of cold nuclear matteron parton branching ? on hadronization ?

what is the x,Q2 dependenceof nuclear quarks and gluons?

answering these questions can help understanding jet suppression in HIC

• the complementarity of e+A with respect to p+A can help

especially when coldmatter effects in p+A

collisions are “strangerthan expected” 

recent PHENIX data

Ri = Nuclear PDF i / (A*proton PDF i)

Early LHC data (e.g. inclusive J/) suggest low x assumptions inadequate

Nuclear parton densities don’t scale with A

Nuclear pdfs: current knowledge

• Simulated LHeC ePb F2 measurementhas huge impact on uncertainties

• Most striking effect for sea & gluons

• High x gluon uncertainty still large

Valence

Sea

Glue

[Example pseudo-datafrom single Q2 Value]

[Effects on EPS09nPDF fit]

Impact of eA F2 LHeC data

Sample of other SM and BSM physics at LHeC

- Least constrained fundamental coupling by far (known to ~1%)- Do coupling constants unify (with a little help from SUSY)?- (Why) is DIS result historically low?

Red = current world average

Black = LHeC projected

[MSSM40.2.5]

- Simulated LHeC precision fromfitting inclusive data per-mille (experimental) also requires improved theory

Measuring αs

Gluon

Sea

d valence

Full simulation of inclusive NC and CC DIS data, includingsystematics NLO DGLAP fit using HERA technology…

… impact at low x (kinematic range) and high x (luminosity)

… precise light quark vector, axial couplings, weak mixing angle

… full flavour decomposition

PDF constraints at LHeC

Current uncertainties due to PDFsfor particles on LHC rapidity plateau (NLO):- Most precise for quark initiated processes around EW scale- Gluon initiated processes lesswell known- All uncertainties explode for largest masses

PDFs at LHC

Ancient history (HERA, Tevatron)

- Apparent excess in large ET jets at Tevatron turned out to be explained by too low high x gluon density in PDF sets

- Confirmation of (non-resonant) new physics near LHC kinematic limit relies on breakdown of factorisation between ep and pp

PRL 77 (1996) 438

Searches near LHC kinematic boundary may ultimately belimited by knowledge of PDFs (especially gluon as x 1)

Do we need to care ?

Summary: nothing on scale of 1 TeV … need to pushsensitivity to higher masses (also non-SUSY searches)

Status of LHC SUSY searches

- Both signal & background uncertainties driven by error on gluon density …Essentially unknown formasses much beyond 2 TeV

- Similar conclusions for other non-resonant LHC signals involving high x partons (e.g. contact interactions signal in Drell-Yan)

- Signature is excess @ large invariant mass - Expected SM background (e.g. gg gg)

poorly known for s-hat > 1 TeV.

e.g. high mass gluino production

• The (pp) LHC has much better discovery potential than LHeC (unless Ee increases to ~500 GeV and Lumi to 1034 cm-2 s-1)

e.g. Expected quark compositeness limitsbelow 10-19 m at LHeC

… big improvement on HERA, but already beaten by LHC

• LHeC is competitive with LHC in cases where initial state lepton is an advantage and offers cleaner final states

e

q

e

q~

0

Direct sensitivity to new physics

Mass range of LQ sensitivity to ~ 2 TeV … similar to LHCSingle production gives access to LQ quantum numbers:- fermion number (below) - spin (decay angular distributions) - chiral couplings (beam lepton polarisation asymmetry)

Leptoquark quantum numbers

• LHC is a totally new world of energy and luminosity, alreadymaking discoveries. LHeCproposal aims to exploit it for lepton-hadron scattering… ep complementing LHC andnext generation ee facility forfull Terascale exploration

• ECFA/CERN/NuPECC workshop gathered many accelerator, theory & experimental colleagues

Conceptual Design Report published. Moving to TDR phase Awaiting outcome of European strategy exercise Build collaboration for detector development

[More at http://cern.ch/lhec]

Summary

… with thanks to many colleagues working on LHeC …