two scenarios for the lhc luminosity upgrade walter scandale, frank zimmermann special paf meeting...

Two Scenarios for the LHC Luminosity Upgrade

Walter Scandale, Frank Zimmermann

Special PAF meeting 10.04.2007

We acknowledge the support of the European Community-Research Infrastructure Activity under the FP6 "Structuring the European Research Area" programme (CARE, contract number RII3-CT-2003-506395)

outline

• beam parameters • features, IR layouts, merits and challenges

of both scenarios• luminosity evolution• bunch structures• luminosity leveling• summary & recommendations• appendix - LUMI’06 outcome, effect of off-center collisions,

shorter bunches vs crab cavities, Super-LHCb, leveling equations

x

zcR

2 ;

1

12

Piwinski angle

luminosity reduction factor from crossing angle

nominal LHC

Name Event DateName Event Date44

W. Scandale/F. Zimmermann, 10.04.2007

parameterparameter symbolsymbol nominalnominal ultimateultimate 12.5 ns, short

transverse emittancetransverse emittance [[m]m] 3.753.75 3.753.75 3.75

protons per bunchprotons per bunch NNbb [10 [101111]] 1.151.15 1.71.7 1.7

bunch spacingbunch spacing t [ns]t [ns] 2525 2525 12.5

beam currentbeam current I [A]I [A] 0.580.58 0.860.86 1.72

longitudinal profilelongitudinal profile GaussGauss GaussGauss Gauss

rms bunch lengthrms bunch length zz [cm] [cm] 7.557.55 7.557.55 3.78

beta* at IP1&5beta* at IP1&5 [m][m] 0.550.55 0.50.5 0.25

full crossing anglefull crossing angle c c [[rad]rad] 285285 315315 445

Piwinski parameterPiwinski parameter cczz/(2*/(2*xx*)*) 0.640.64 0.750.75 0.75

peak luminositypeak luminosity LL [10 [103434 cm cm-2-2ss-1-1]] 11 2.32.3 9.2

peak events per crossingpeak events per crossing 1919 4444 88

initial lumi lifetimeinitial lumi lifetime LL [h] [h] 2222 1414 7.2

effective luminosity effective luminosity (T(Tturnaroundturnaround=10 h)=10 h)

LLeff eff [10[103434 cm cm-2-2ss-1-1]] 0.460.46 0.910.91 2.7

TTrun,optrun,opt [h] [h] 21.221.2 17.017.0 12.0


LLeff eff [10[103434 cm cm-2-2ss-1-1]] 0.560.56 1.151.15 3.6

TTrun,optrun,opt [h] [h] 15.015.0 12.012.0 8.5

e-c heat SEY=1.4(1.3)e-c heat SEY=1.4(1.3) P [W/m]P [W/m] 1.07 (0.44)1.07 (0.44) 1.04 (0.59)1.04 (0.59) 13.34 (7.85)

SR heat load 4.6-20 KSR heat load 4.6-20 K PPSRSR [W/m] [W/m] 0.170.17 0.250.25 0.5

image current heat image current heat PPICIC [W/m] [W/m] 0.150.15 0.330.33 1.87

gas-s. 100 h (10 h) gas-s. 100 h (10 h) bb PPgasgas [W/m] [W/m] 0.04 (0.38)0.04 (0.38) 0.06 (0.56)0.06 (0.56) 0.113 (1.13)

extent luminous regionextent luminous region l [cm] 4.54.5 4.34.3 2.1

commentcommentpartial wire

c.total heat far exceeds max. local cooling capacity of 2.4 W/m

baselineupgradeparameters2001-2005

abandonedatLUMI’06

(SR and image current heat load well known)



parameterparameter symbolsymbol 25 ns, small * 50 ns, long

transverse emittancetransverse emittance [[m]m] 3.75 3.75

protons per bunchprotons per bunch NNbb [10 [101111]] 1.7 4.9

bunch spacingbunch spacing t [ns]t [ns] 25 50

beam currentbeam current I [A]I [A] 0.86 1.22

longitudinal profilelongitudinal profile Gauss Flat

rms bunch lengthrms bunch length zz [cm] [cm] 7.55 11.8

beta* at IP1&5beta* at IP1&5 [m][m] 0.08 0.25

full crossing anglefull crossing angle c c [[rad]rad] 0 381

Piwinski parameterPiwinski parameter cczz/(2*/(2*xx*)*) 0 2.0

hourglass reduction hourglass reduction 0.86 0.99

peak luminositypeak luminosity LL [10 [103434 cm cm-2-2ss-1-1]] 15.5 10.7

peak events per crossingpeak events per crossing 294 403

initial lumi lifetimeinitial lumi lifetime LL [h] [h] 2.2 4.5


LLeff eff [10[103434 cm cm-2-2ss-1-1]] 2.4 2.5

TTrun,optrun,opt [h] [h] 6.6 9.5


LLeff eff [10[103434 cm cm-2-2ss-1-1]] 3.6 3.5

TTrun,optrun,opt [h] [h] 4.6 6.7

e-c heat SEY=1.4(1.3)e-c heat SEY=1.4(1.3) P [W/m]P [W/m] 1.04 (0.59) 0.36 (0.1)

SR heat load 4.6-20 KSR heat load 4.6-20 K PPSRSR [W/m] [W/m] 0.25 0.36

image current heat image current heat PPICIC [W/m] [W/m] 0.33 0.78

gas-s. 100 h (10 h) gas-s. 100 h (10 h) bb PPgasgas [W/m] [W/m] 0.06 (0.56) 0.09 (0.9)

extent luminous regionextent luminous region l [cm] 3.7 5.3

commentcomment D0 + crab (+ Q0) wire comp.

two newupgradescenarios

compromisesbetweenheat loadand # pile upevents

PAF/POFPA Meeting 20 November 2006LHC Upgrade Beam Parameters, Frank ZimmermannW. Scandale/F. Zimmermann, 10.04.2007

for operation at beam-beam limitwith alternating planes of crossing at two IPs, luminosity equation can be written as

ghprofilebb

p

revb FFQ

r

fnL 22

*21

↓↓ 25 ns ↑↑ 50 ns↓ 50 ns

↓ 50 ns

where Qbb = total beam-beam tune shift

(hourglass effect is neglected above)



25-ns low-25-ns low- upgrade scenario upgrade scenario• stay with ultimate LHC beam (1.7x10stay with ultimate LHC beam (1.7x101111

protons/bunch, 25 spacing)protons/bunch, 25 spacing)

• squeeze squeeze * to ~10 cm in ATLAS & CMS * to ~10 cm in ATLAS & CMS

• add early-separation dipoles in detectors add early-separation dipoles in detectors starting at ~ 3 m from IP starting at ~ 3 m from IP

• possibly also add quadrupole-doublet possibly also add quadrupole-doublet inside detector at ~13 m from IP inside detector at ~13 m from IP

• and add crab cavities (and add crab cavities (PiwinskiPiwinski~ 0), and/or ~ 0), and/or shorten bunches with massive addt’l rfshorten bunches with massive addt’l rf

→ → new hardware inside ATLAS & CMS new hardware inside ATLAS & CMS detectors, first hadron-beam crab cavities detectors, first hadron-beam crab cavities

(J.-P. Koutchouk et al)


CMS & ATLAS IR layout for 25-ns option

ultimate bunches & near head-on collision

stronger triplet magnetsD0 dipole

small-angle

crab cavity

Q0 quad’s


challenges: D0 dipole deep inside detector (~3 m from IP),Q0 doublet inside detector (~13 m from IP),crab cavity for hadron beams (emittance growth),

or shorter bunches (requires much more RF)4 parasitic collisions at 4-5 separation,“chromatic beam-beam” Q’eff~z/(4*),poor beam and luminosity lifetime ~*.

merits:negligible long-range collisions,no geometric luminosity loss,no increase in beam current beyond ultimate

25-ns scenario assessment



50-ns higher 50-ns higher * upgrade * upgrade scenarioscenario• double bunch spacingdouble bunch spacing

• longer & more intense bunches with longer & more intense bunches with PiwinskiPiwinski~ 2~ 2

• keep keep *~25 cm (achieved by stronger low-*~25 cm (achieved by stronger low- quads alone)quads alone)

• do not add any elements inside detectorsdo not add any elements inside detectors

• long-range beam-beam wire compensation long-range beam-beam wire compensation

→ → novel operating regime for hadron novel operating regime for hadron colliderscolliders


CMS & ATLAS IR layout for 50-ns option

long bunches & nonzero crossing angle & wire compensation

wire

compensator

larger-aperture triplet magnets


merits:no elements in detector, no crab cavities,lower chromaticity,less demand on IR quadrupoles (NbTi possible),could be adapted to crab waist collisions (LNF/FP7)

challenges: operation with large Piwinski parameter unproven for hadron beams, high bunch charge,beam production and acceleration through SPS,“chromatic beam-beam” Q’eff~z/(4*),larger beam current,wire compensation (almost etablished)

50-ns scenario assessment


IR upgrade optics“compact low-gradient” NbTi, *=25 cm

<75 T/m (Riccardo De Maria, Oliver Bruning)

“modular low gradient” NbTi, *=25 cm <90 T/m (Riccardo De Maria, Oliver Bruning)

“low max low-gradient” NbTi, *=25 cm

<125 T/m (Riccardo De Maria, Oliver Bruning)

standard Nb3Sn upgrade, *=25 cm ~200 T/m, 2 versions with different magnet parameters(Tanaji Sen et al, Emmanuel Laface, Walter Scandale)

+ crab-waist sextupole insertions? (LNF/FP7)

early separation with *=8 cm, Nb3Sn includes D0; either triplet closer to IP or Q0; being prepared for PAC’07 (Jean-Pierre Koutchouk et al)

compatiblewith50-nsupgradepath


crab waist scheme

cyI

xpH

2

4

1 2

minimizes at s=-x/c

focal plane

realization:add sextupoles at right phase distance from IP

initiated and ledby LNF in the frame of FP7;first beam testsat DAFNE later in 2007

Hamiltonian


25 ns spacing

50 ns spacing

IP1& 5 luminosity evolution for 25-ns and 50-ns spacing

averageluminosity

initial luminosity peakmay not be useful for physics(set up & tuning?)


25 ns spacing

50 ns spacing

IP1& 5 event pile up for 25-ns and 50-ns spacing


old upgrade bunch structure

25 ns

12.5 ns

nominal

25 ns

ultimate

12.5-ns upgrade

abandonedat LUMI’06


new upgrade bunch structures

25 ns

50 ns

nominal

25 ns

ultimate& 25-ns upgrade

50-ns upgrade,no collisions @S-LHCb!

50 ns

50-ns upgradewith 25-ns collisionsin LHCb

25 ns

new alternative!

new baseline!

PAF/POFPA Meeting 20 November 2006LHC Upgrade Beam Parameters, Frank ZimmermannF. Zimmermann, W. Scandale, 10.04.2007

luminosity leveling in IP1&5experiments prefer more constant luminosity, less pile up at the start of run, higher luminosity at end

how could we achieve this?

25-ns low- scheme: dynamic squeeze

50-ns higher- scheme:dynamic squeeze, and/ordynamic reduction in bunch length

(less invasive)


dynamic squeeze for 25-ns option

N150

N75


dynamic squeeze for 50-ns option

N150

we might also reduce the charge/bunchand go for shorter bunches


dynamic bunch length change for 50-ns option


25 ns, low *,

with leveling

50 ns, long bunches,

with levelingevents/crossing 300 300run time N/A 2.5 h

av. luminosity N/A 2.6x1034s-1cm-2

events/crossing 150 150run time 2.5 h 14.8 h

av. luminosity 2.6x1034s-1cm-2 2.9x1034s-1cm-2

events/crossing 75 75

run time 9.9 h 26.4 h

av. luminosity 2.6x1034s-1cm-2 1.7x1034s-1cm-2

assuming 5 h turn-around time


25 ns spacing

50 ns spacing

IP1& 5 luminosity evolution for 25-ns and 50-ns spacingwith leveling

averageluminosity


25 ns spacing

50 ns spacing

IP1& 5 event pile up for 25-ns and 50-ns spacingwith leveling


150 evts/Xing

example tune shifts with luminosity leveling


average luminosity & run time vs. final for 25-ns option with dynamic * squeeze


average luminosity & run time vs. final for 50-ns option with dynamic * squeeze


average luminosity & run time vs. final z

for 50-ns option with dynamic z change



summary summary • two scenarios of L~10two scenarios of L~103535 cm cm-2-2ss-1-1 for which heat load for which heat load

and #events/crossing are acceptableand #events/crossing are acceptable• 25-ns option25-ns option: pushes : pushes *; requires slim magnets *; requires slim magnets

inside detector, crab cavities, & Nbinside detector, crab cavities, & Nb33Sn Sn quadrupolesquadrupoles and/or Q0 doublet; attractive if total and/or Q0 doublet; attractive if total beam current is limited; transformed to a 50-ns beam current is limited; transformed to a 50-ns spacing by keeping only ½ the number of bunches spacing by keeping only ½ the number of bunches

• 50-ns option:50-ns option: has fewer longer bunches of higher has fewer longer bunches of higher charge ; can be realized with NbTi technologycharge ; can be realized with NbTi technology if if needed ; compatible with LHCb ; open issues are needed ; compatible with LHCb ; open issues are SPS & beam-beam effects at large Piwinski SPS & beam-beam effects at large Piwinski angleangle; luminosity leveling may be done via bunch ; luminosity leveling may be done via bunch lengthlength and via and via **

bb



recommendationsrecommendations• luminosity levelingluminosity leveling should be seriously should be seriously

considered: considered: → → higher quality events, moderatehigher quality events, moderatedecrease in average luminositydecrease in average luminosity

• it seems it seems long-bunch 50-ns option entails less risklong-bunch 50-ns option entails less risk and less uncertainties; however not w/o problemsand less uncertainties; however not w/o problems

• leaving the leaving the 25-ns option as back up25-ns option as back up until we have until we have gained some experience with the real LHC may gained some experience with the real LHC may be wisebe wise

• needed for both scenarios are needed for both scenarios are concrete optics concrete optics solutionssolutions, , beam-beam tracking studiesbeam-beam tracking studies, and , and beam-beam machine experimentsbeam-beam machine experiments


appendix



1) quadrupole 1st preferred over dipole 1st

2) pushed NbTi or Nb3Sn still pursued, or hybrid solution - new

3) slim magnets inside detector (“D0 and Q0”) – new

4) wire compensation ~established; electron lens – new

5) crab cavities: large angle rejected; small-angle – new

6) 12.5-ns scenario strongly deprecated

7) e-cloud/pile-up compromise: 25-ns w *~8 cm, or 50-ns spacing, *=0.25 m, long bunches – new

Reminder of LUMI’06 OutcomeReminder of LUMI’06 OutcomeIR upgrade and beam parametersIR upgrade and beam parameters


4 parasitic collisions at 4-5 offset in 25-ns low- case

concerns:

• poor beam lifetime • enhanced detector background

discouraging experience at RHIC, SPS, HERA and Tevatron


RHIC experiments in 2005 and 2006

single off-center collision

one collision with 5-6 offset strongly reduces RHIC beam lifetime; worse at smaller offsets

(W. Fischer et al.)

24 GeV

100 GeV


proton background with 1 head-on and 1 off-center collision vs beam-beam separation (K. Cornelis, LHC99);significantly affected by single LR collision at 3(W.Herrsee also PhD thesis M.Meddahi, CERN SL/91-30, Fig. 22

SPS collider ~1980s


HERA ~1992proton beam lifetime drops from 50 h to 1-5 h for single off-center collision with beam-beam separation between 0.3 and 2 (F. Willeke & R. Brinkmann, PAC 93; T. Limberg, LHC’99)

Tevatron 2006removal of the four closest long-range collisions at about 6.2separation has increased integrated Tevatron luminosity per run by up to 30%(V. Shiltsev, private communication)


shorter bunches for 25 ns?

• reduced longitudinal emittance 2.5 eVs → 1.78 eVs (loss of Landau damping if z

5frf3Vrf < const Nb Im(Z||/n);

also stronger IBS) • 43 MV rf voltage at 1.2 GHz

(nominal LHC: 16 MV at 400 MHz)• not sufficient to avoid large luminosity loss

(for which crab cavities are needed anyhow)

bunch shortening rf voltage:

c

rfrf

ccrab Rfe

cE

Rfe

cEV

12

0

12

0

42

2/tan

crab cavity rf voltage:

unfavorable scaling as 4th power of crossing angle and inverse 4th power of IP beam size; can be decreased byreducing the longitudinal emittance; inversely proportionalto rf frequency

proportional to crossing angle & independent of IP beam size;scales with 1/R12; also inversely proportional to rf frequency

4*4

4

0

32||,

40

32||,

16 7.02

1

2 x

c

rf

rms

zrf

rmsrf fE

Cc

fE

CcV

LUMI’05


LHCb recipe for 50-ns scenario

• add satellites at 25 ns spacing• these can be produced by highly asymmetric bunch

splitting in the PS (possibly large fluctuation)• in LHCb satellites collide with main bunches • satellite intensity should be lower than 3x1010 p/bunch

to add <5% to beam-beam tune shift and to avoide-cloud problems; 3x1010 ~ 1/16th of main-bunch charge

• function of ~3 m would result in desired luminosity equivalent to 2x1033 cm-2s-1; easily possible with present IR magnets & layout


• here head-on collisions unavoidably contribute to beam-beam tune shift of the bunches colliding in ATLAS & CMS

• two potential ways out:– collisions with transverse offset; concerns: offset

stability, interference with collimation, poor beam lifetime, detector background

– collide at LHCb only in later part of each store, when the beam-beam tune shift from IP1 & 5 has decreased (H. Dijkstra)

LHCb schemes for 25-ns scenario


L = L0 exp (-d2/(42))

LHCb collisions with transverse offset d

luminosity:

Q LHCb = 2 QIP1or5 / (d/2tune shift:

suppose tune shift from LHCb should be less than 10% of that from CMS or ATLAS → d>4.5 then luminosity L ~ 0.006 L0

if we wish LLHCb~0.01 LIP1or5 (~1-2x1033 cm-2s-1)

we need * ~0.08 m → IR triplet upgrade!

offset collisions w/o IR upgrade LLHCb ~ 4x1031 cm-2s-1

(for Gaussian distribution)


LHCb luminosity for 25 ns with offset & 50 ns

25 ns spacing,4.5 offset,*~0.08 m

50 ns spacing,satellites

LHCb 50-ns luminosity decays 2x more slowlythan 25-ns luminosity or that at ATLAS and CMS


25 ns spacing

50 ns spacing

tune shift during store for 25-ns & 50-ns spacing

changeQ ~-0.0033

LHCb 25-ns collisions from middle of each store?! *~3 m (5 h turnaround time is assumed)


LHCb luminosity for 25-ns late collisions & 50 ns

25 ns spacing,* ~ 3 m,no transverseoffset

50 ns spacing,*~3 m,satellites

(5 h turnaround time is assumed)


S-LHCb collision parametersparameter symbol 25 ns, offset 25 ns, late collision 50 ns, satellites

collision spacing Tcoll 25 ns 25 ns 25 ns

protons per bunch Nb [1011] 1.7 1.7 4.9 & 0.3

longitudinal profile Gaussian Gaussian flat

rms bunch length z [cm] 7.55 7.55 11.8

beta* at LHCb [m] 0.08 3 3

rms beam size x,y* [m] 6 40 40

rms divergence x’,y’* [rad] 80 13 13

full crossing angle c [urad] 550 180 180

Piwinski parameter cz/(2*x*) 3.3 0.18 0.28

peak luminosity L [1033 cm-2s-1] 1.13 2.1 2.4

effective luminosity (5 h turnaround time) Leff [1033 cm-2s-1] 0.25 0.35 0.67

initial lumi lifetime L [h] 1.8 2.8 9

length of lum. region l [cm] 1.6 5.3 8.0

rms length of luminous region:

2

,*

2

222

21

yx

c

zl


constLL 0constn

LXingevents

b

inel 0/

IPtot

brun nL

nNt

max

aroundturnb

IPtotave

TnN

nLL

L

max

0

0

1

tn

nLNN

b

IPtot00

beam intensitydecays linearly

length of run average luminosity

leveling equations

two scenarios for the lhc luminosity upgrade walter scandale, frank zimmermann special paf meeting...

Documents