Observation of strong leakage reduction in crystal assisted collimation of the SPS beam

Physics Letters B 748 (2015) 451–454

Contents lists available atScienceDirect

Physics

Letters

B

www.elsevier.com/locate/physletb

Observation

of

strong

leakage

reduction

in

crystal

assisted

collimation

of

the

SPS

beam

W. Scandale

a

,

b

,

e

,

G. Arduini

a

,

M. Butcher

a

,

F. Cerutti

a

,

M. Garattini

a

,

S. Gilardoni

a

,

A. Lechner

a

,

R. Losito

a

,

A. Masi

a

,

A. Mereghetti

a

,

E. Metral

a

,

D. Mirarchi

a

,

j

,

S. Montesano

a

,

S. Redaelli

a

,

R. Rossi

a

,

e

,

P. Schoofs

a

,

G. Smirnov

a

,

E. Bagli

c

,

L. Bandiera

c

,

S. Baricordi

c

,

P. Dalpiaz

c

,

G. Germogli

c

,

V. Guidi

c

,

A. Mazzolari

c

,

D. Vincenzi

c

,

G. Claps

d

,

S. Dabagov

d

,

k

,

l

,

D. Hampai

d

,

F. Murtas

d

,

G. Cavoto

e

,

F. Iacoangeli

e

,

L. Ludovici

e

,

R. Santacesaria

e

,

P. Valente

e

,

F. Galluccio

f

,

A.G. Afonin

g

,

Yu.A. Chesnokov

g

,

A.A. Durum

g

,

V.A. Maisheev

g

,

Yu.E. Sandomirskiy

g

,

A.A. Yanovich

g

,

A.D. Kovalenko

h

,

A.M. Taratin

h

,

∗

,

Yu.A. Gavrikov

i

,

Yu.M. Ivanov

i

,

L.P. Lapina

i

,

J. Fulcher

j

,

G. Hall

j

,

M. Pesaresi

j

,

M. Raymond

j

a_{CERN,EuropeanOrganizationforNuclearResearch,CH-1211Geneva23,Switzerland} b_{Laboratoiredel’AccelerateurLineaire(LAL),UniversiteParisSudOrsay,Orsay,France} c_{INFNSezionediFerrara,DipartimentodiFisica,UniversitàdiFerrara,Ferrara,Italy} d_{INFNLNF,ViaE.Fermi,4000044Frascati(Roma),Italy}

e_{INFNSezionediRoma,PiazzaleAldoMoro2,00185Rome,Italy} f_{INFNSezionediNapoli,Italy}

g_{InstituteofHighEnergyPhysics,MoscowRegion,RU-142284Protvino,Russia}

h_{JointInstituteforNuclearResearch,Joliot-Curie6,141980,Dubna,MoscowRegion,Russia} i_{PetersburgNuclearPhysicsInstitute,188300Gatchina,LeningradRegion,Russia} j_{ImperialCollege,London,UnitedKingdom}

k_{RASP.N.LebedevPhysicalInstitute,Moscow,Russia} l_{NRNuclearUniversityMEPhI,Moscow,Russia}

a

r

t

i

c

l

e

i

n

f

o

a

b

s

t

r

a

c

t

Articlehistory:

Received4June2015

Receivedinrevisedform8July2015 Accepted17July2015

Availableonline23July2015 Editor:L.Rolandi Keywords: Accelerator Beamcollimation Crystal Channeling

In ideal two-stage collimation systems, the secondary collimator–absorber should have its length sufficient to exclude practicallythe exit ofhalo particles with large impact parameters. In the UA9 experiments onthe crystal assisted collimation ofthe SPS beam a60 cm longtungstenbar is used as asecondary collimator–absorberwhich is insufficient for the full absorption of the halo protons. Multi-turn simulation studies of the collimation allowed to select the position for the beam loss monitor downstreamthe collimationarea wherethe contributionofparticlesdeflectedbythe crystal inchannelingregimebutemergingfromthesecondarycollimator–absorberisconsiderablyreduced.This allowedobservationofastrongleakagereductionofhaloprotonsfromtheSPSbeamcollimationarea, therebyapproachingthecasewithanidealabsorber.

©2015TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense (http://creativecommons.org/licenses/by/4.0/).FundedbySCOAP3_.

1. Introduction

Amulti-stage collimationsystem is usedin the LargeHadron Collider(LHC)toabsorbhaloparticlesofthecirculatingbeam

[1]

. A crystal-assistedcollimation scheme for LHC is presently under study

[2]

.Abentcrystalusedasaprimary collimatorinsteadofa

*

Correspondingauthor.

E-mailaddress:alexander.taratin@cern.ch(A.M. Taratin).

heavysolidtargetdeflectshaloparticlesinthechannelingregime, directing them into a secondary collimator–absorber far fromits internaledge.

Fig. 1showsthedistributionsoftheimpactparametersof pro-tonswiththe absorber obtainedby multi-turnsimulation, which should be realized inthe experiment discussed belowon crystal assisted collimationoftheSuperProton Synchrotron(SPS) proton beam.Inthecaseofcrystalorientationoptimalforchanneling (2), the halo fraction, which hits the absorber edge, is considerably

http://dx.doi.org/10.1016/j.physletb.2015.07.040

0370-2693/©2015TheAuthors.PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/).Fundedby SCOAP3_.

452 W. Scandale et al. / Physics Letters B 748 (2015) 451–454

Fig. 1. (Coloronline.)Thecalculateddistributionsofimpactparameterswiththe secondarycollimator–absorbermeasuredfromitsinternaledgeforacrystalassisted collimationoftheSPSbeamof270 GeV/cprotonsinthecaseofamorphous(1)and channeling(2)orientationsofthecrystal.

smallerthan inthe caseofan amorphouscrystalorientation (1). Asaresult, thenumberofparticlesreturningback intothebeam duetoscatteringatthesurfaceoftheabsorbershouldbestrongly reduced,thusincreasingthecollimationefficiency.Thefirst exper-imentsoncrystalassistedcollimationwereperformedattheIHEP synchrotron

[3]

,RHIC

[4]

andTevatron[5].

TheUA9experimentalstudies

[6–9]

oncrystalassisted collima-tionofthe SPSbeamwhichstarteda fewyearsago showedthat crystalalignmentwiththecirculatedbeamhalocouldbeobtained quicklyusingthebeamlossmonitors(BLM1)installeddownstream ofthe crystal, asshownby theschematic layout in Fig. 2. Chan-neledparticleswithsmalloscillationamplitudesinthecrystal pla-nar channelsdo nothave closecollisions withthe crystalnuclei, and, consequently, do not experiencenuclear interactions. There-fore,thebeamlossesinthealignedcrystalarestronglydecreased incomparisonwiththecaseofitsamorphousorientation.

Inourpreviousexperiments,thecollimationleakagewas

mea-sured by the monitor BLM2 installed in the first high

disper-sion(HD)areadownstreamofthecollimator–absorberwhere

off-momentum particles produced in the collimation area have the

firstpossibility tohit thebeampipe.A considerablereductionof the collimation leakage was always observed for the channeling orientation of a crystal. Forthe case of an SPS beam of Pbions with270 GeV/cmomentumperunitcharge,thelossreduction ob-servedintheHDareabyBLM2waspracticallythesameasinthe crystalbecausetheprobabilityofbackscatteringfromthetungsten absorber is very small for Pb ions due to their high probability ofnuclear interactions. In thecase ofprotons, thebeam loss

re-duction detected by BLM2 was always smaller than detected by

BLM1inthecrystalbecauseofthecontributionofparticles emerg-ingfromtheabsorber

[8]

.

IntheUA9experiments,a60 cmlongtungsten barisusedas a secondary collimator–absorber.It is insufficientforthe full ab-sorptionofthehaloprotons.Thenuclearinelasticcross-sectionfor 270 GeV/cprotonsintungsten

σ

in

=

1

.

725b [10] andthe

interac-tionlengthSin

=

9

.

18 cm correspondstotheattenuation

probabil-ityoftheprotonbeam Pin

=

exp

(

−

L

/

Sin

)

=

1

.

45

×

10−3.Besides,

protons afterlosing a small partof their momentum by diffrac-tive scatteringcan alsoremaininthe beam. Protonsdeflectedby acrystaldeeplyintotheabsorberbutemergingfromitwithsome momentumlossgivealargecontributiontothebeamlosses mea-suredbythemonitorBLM2.Thus,theimperfectabsorptionofhalo protonsinourcollimator–absorberleadstoanunderestimationof the efficiencywhichisachievablewitha crystalassisted collima-tion system. Thesituation maybeconsiderably improvedalready witha1 mlongtungstenabsorber, Pin

=

1

.

86

×

10−5.

Multi-turnsimulationofthecrystalassistedcollimationofthe SPS beamhalowithaSixTrackcodeandrealbeampipe aperture

[14] allowed predicting the azimuth in the first high dispersion area where the lossreduction forthe crystalchanneling orienta-tion is considerably larger than that observed in the position of BLM2.Beamlossmonitorshavethereforebeeninstalledatthis az-imuth,BLM3in

Fig. 2

.

In thispaper the resultsof theexperiment on thecrystal

as-sisted collimation of the CERN SPS beam where the collimation

leakage reductionobservedwithBLM3 isconsiderablylargerthan the lossreduction inthe crystalare described.The situationwas closetotheidealcasewhenfullabsorptionofparticleswithlarge impactparametersoccursinthesecondarycollimator.

2. Theexperimentdescription

Fig. 2showstheschematic layoutoftheUA9experimentwith only devices used in the presentmeasurements. The crystal pri-mary collimator andthesecondary collimator–absorber(TAL) are installed at the SPS azimuths with relative horizontal betatron phase advance closeto90 degreesandwitha large value ofthe horizontalbetafunction.ThesiliconstripcrystalC1produced us-ing techniques described in [11,12] was used asa primary colli-mator.Thecrystalparametersarepresentedin

Table 1

.Thecrystal miscut angle between the crystal surface and the (110)

crystal-lographic planes is about 10 μrad, which is much smaller than

forthecrystalsusedinourearlierexperiments.Thisfeaturehelps toreduce theparticlelossesatthecrystalchannelingorientation. ProtonsfromtheSPSbeamhalohitthecrystalandsecondary par-ticlesproducedinnuclearinelasticinteractionsaredetectedinthe beamlossmonitorBLM1.ThegoniometerproducedbyIHEPallows adjustingthecrystalorientationrelativetothebeamhalodirection withanangularaccuracyofabout

±

10 μrad.

Downstreamoftheabsorber TALthereisthefirsthigh

disper-sion area. Off-momentum particles with momentum p and

suf-ficiently large

δ

=

p

/

po

−

1, where po is the momentum of the

synchronous particle,emergingfromeitherthe crystalorthe ab-sorber have displacements from the orbit here, xδ

=

Dx

δ

, where

Table 1 ParametersofcrystalC1. Length (mm) Bendangleα (μrad) BendradiusR (m) Miscastangleθm (μrad) 1.87 165 11.33 10

Fig. 2. (Coloronline.)AschematiclayoutoftheUA9experiment.ThecrystalprimarycollimatorC1islocatedupstreamofthequadrupoleQF518(QF1).TheTALactingasa secondarycollimator–absorberisupstreamofthequadrupoleQF520(QF2).ThebeamlossmonitorBLM1isusedtofindchannelinginthecrystalandBLM2andBLM3are usedtodetectparticlesleakingoutfromthecollimationarea.

W. Scandale et al. / Physics Letters B 748 (2015) 451–454 453

Fig. 3. (Coloronline.) Thedependenceofthedispersionfunctiononthedistance alongthecollimationareaandthefirstHDareadownstreamoftheabsorber.

Table 2

Relevantacceleratorparameters.

Parameter C1 TAL QF3

βx(m) 90.945 87.660 107.023

σx(mm) 0.9047 0.888 0.981

μxfrom C1(2π) 0 0.2405 0.4909

Dx(m) −0.857 −6.8×10−4 3.772 Dx isthedispersionfunction,and, hence,they mightbelost.The

targetslimitingthe accelerator aperture installed in theHD area were not used in this experiment.The beamloss monitor BLM2 usedinour previous experiments

[8,9]

detected secondary parti-clesgeneratedby protonsinthepipe walls. Onemorebeamloss

monitor BLM3 has been installed upstream the quadrupole QF4.

Fig. 3showsthedispersionfunctionchangeinthecollimationarea andinthefirstHDareadownstreamoftheabsorber.Thepositions ofthemonitorsBLM2 andBLM3areclosetothefirstandthe sec-ond dispersion maximums, respectively. Additionally, it is shown

that the betatron phase advance between the absorber TAL and

themonitorBLM2 equals approximately90◦.Itis veryimportant thatprotonsstronglyscatteredintheTALshould acquirea maxi-malbetatrondeviationfromtheorbitnearBLM2.

IntheSPS,the beamofprotons wasacceleratedto270 GeV/c withnominalbetatrontunes QH

=

26

.

13 and QV

=

26

.

18.Inthe

positionofthemonitorBLM3thebeamlossesaresmall.Therefore, a beamconsisting of 12 buncheswith a total intensity of 1

.

3

×

1012protonswasusedinthisexperiment.Therelevantaccelerator

parameters at the azimuths of some UA9 elements are listed in

Table 2, where

β

x is the horizontalbeta-function,

σ

x is theRMS

valueofthehorizontalbeamsize(herefortheRMSemittance

ε

=

0

.

009 μm rad),and



μ

x isthehorizontalphase advancebetween

theelements.

3. Experimentalresults

At the beginning of the measurement the two-sided collima-torCOL was centered relative to the closedorbit. The collimator halfgap X1/2 determinedthe referencebeamenvelope.Then the alignmentpositionsforthecrystalCR1andtheabsorberTALwere determined by fixing their positions at the edge of the collima-torshadow.Afterthatthecrystalandtheabsorberwereplacedat adistance XC 1

=

4

.

07 mm (4

.

5

σ

x) and XTAL

=

7

.

05 mm (7

.

94

σ

x)

fromthe orbit, respectively. Hence, the TAL offsetrelative to the crystalwas Xoff

=

3

.

06 mm.Undertheseconditionsascanofthe

horizontalangularpositionsofthecrystalwasperformedwiththe

Fig. 4. (Coloronline.)TheresultsofcrystalassistedcollimationoftheSPSbeamhalo of270 GeV/c protons.Curve(1)shows thedependenceofbeamlossesobserved inthecrystal(a)andintheHDareawiththeBLM3 monitor(b)ontheangular positionofthecrystalC1normalizedtoitsvalueforamorphousorientationofthe crystal(dot-dashedline).Curves(2)and(3)showthedependenceofthenumber ofnuclearinelasticinteractionsofprotonsinthecrystalonitsorientationangle obtainedbysimulationaccordingto[13]and[14],respectively.

goniometer.

Fig. 4

a showsthe dependenceofbeam lossesin the crystal observed with BLM1 (curve 1). The left minimum corre-spondstothecrystalorientationoptimalforchannelingwherethe beamlossesoccurmainlyduetothenon-channeledfraction. The ratioofthebeamlossesattheamorphousandchanneling orienta-tionsofthecrystaldeterminesthebeamlossreductionRbl.The

re-ductionmeasuredfortheareabehindthecrystalis Rbl

(

1

)

=

11

.

8.

The angularregion in Fig. 4 with reducedlosseson the right

of the channeling minimum is due to volume reflection(VR) of

particles by bent crystal planes,which allows them to reach the TALapertureinasmallernumberofpassagesthroughthecrystal than occursdueto multiplescatteringforamorphouscrystal ori-entations.ThesecondminimumintheVRregionobservedalsoin ourpreviousexperiments

[7,8]

isclearlyseenhere.Thisminimum isobservedatanangulardistancefromthechannelingorientation equalaboutthecrystalbendangle.Asalreadyexplained

[6]

,inthis casethewholeVR regionison thesameside ofthebeam enve-lopedirection.Therefore,angularkicksduetoVR alwaysincrease theoscillationamplitudesofparticlesandtheymorequicklyreach theabsorber.

Curve2 showsthe dependenceofnuclearinelasticinteraction number in the crystal on its orientation obtained by multi-turn simulationofthecollimationprocesswiththedetailedcalculation of particle trajectories in the crystal, see [13]. Linear 6-D trans-fer matrices M

(

6

,

6

)

were used to transport particles in theSPS. The simulation for a given particlewas finished when it hit the

454 W. Scandale et al. / Physics Letters B 748 (2015) 451–454

TALorwhenanuclearinelasticinteractionoccurredinthecrystal. Thecalculateddependencedescribeswellboththewidthandthe shapeoftheexperimentaldependencebutgivessmallervaluesof the beamlosses for the channelingas well as forthe VR orien-tations of the crystal. The distributions of impact parameters of protonswiththeTALobtainedbysimulationforsome amorphous crystalorientation (1) andfor the alignedcase(2) are shownin

Fig. 1.Protons deflected by the crystal inthe channeling regime hittheTALatadistanceofabout8mmfromitsedge.

Curve 3 shows the multi-turn simulation results obtained by usingtheSixTrackcodetotransport particlesintheSPS.The pro-cessofdiffractivescatteringofprotonsinthecrystalandTALwas takenintoaccount.Theinteractionofprotonswiththecrystalwas consideredusingapproximationsfordifferentprocessesdescribed in [14]. In this case, the calculated width of the angular depen-denceis smallerthan in theexperiment becauseof the approxi-mationsused for the description of the protoninteractions with thecrystal. However, the loss value inthe VR regionis inbetter agreement with the experiment. The contribution of diffraction-scattered protons in repeated passages through the crystal may explainthelossesinthisVRregion.

Thereduction ofnuclearinelasticinteraction rateofhalo pro-tonsinthecrystalforits optimalchannelingorientation obtained inboth simulations isconsiderablylarger thanthe beamloss re-ductionobservedwithBLM1(about100and60,respectively).The discrepancymaybepartlyconnectedwiththegoniometer inaccu-racyof

±

10 μrad. Simulations combiningthedetailedcalculation of particle trajectories in the crystalwith the application of the SixTrackcodeandtakingintoconsiderationthediffractive scatter-ing of protons inthe crystaland TAL are plannedfor ourfuture collimationexperiments.

The beam loss reduction observed in the HD area with the

monitorBLM2wasRbl

(

2

)

=

8

.

3,whichisconsiderablysmallerthan

behind thecrystal asinour previous experiments.Fig. 4b shows thebeamloss dependenceon thecrystalorientation observedin thenewposition withmonitorBLM3.The angulardependence is

the same as behind the crystal but the beam loss reduction in

channelingisconsiderably larger, Rbl

(

3

)

=

18

.

1,which isingood

agreement with the simulation prediction mentioned above. The

reduction valuesofthe protonlossesonthe beampipe obtained

in our simulation with the SixTrack code are 9.1 and 20.1

up-streamBLM2 andBLM3,respectively.The observationofthelarge reductionofcollimationleakageispossiblewithBLM3becausethe numberofparticlesdeflectedbythecrystalinchannelingregime deeplyinside theTALbutemerging fromitconsiderably reduced ontheirwaytoBLM3.

Theparticles werelostatthebeampipenearthefirst

disper-sion maximum where the betatron phase advance from the TAL

exit is about 90◦. Actually, the RMS deflection due to multiple Coulomb scatteringin theTALtakingintoaccount nuclearelastic scattering, islarge,

θ

ms

=

0

.

742 mrad,and theaverage ionization

lossesareestimatedas

δ

= −

7

.

62

×

10−3.Thebetatronamplitudes forprotons deflectedby thecrystalin thechannelingregime are about15 mm at the TAL entrance face. At the TAL exit,protons deflectedthrough

θ

msduetomultiplescatteringwillhavean

am-plitude of Xm

=

68

.

6 mm andtheamplitudenearthequadrupole

QF3willbe Xm

=

75 mm.Besides, theaverageshiftnearQF3for

theseparticlesduetohighdispersionis Xδ

= δ

Dx

= −

28

.

75 mm.

The horizontaland vertical dimensions ofthe SPS beam pipe in

the quadrupolesQFs are76 mm and19.15 mm,respectively. The

simulationsshowthatalargerfractionofparticlesdeflectedbythe crystalbutwhichavoidedabsorption intheTALshould belostin thispartofthepipe.

4. Conclusions

The position selected for the beam loss monitor downstream the collimation area helps to reduce considerably the contribu-tionofparticlesdeflectedbythecrystalinchannelingregimebut

emerging from the secondary collimator–absorber. This allowed

observationofastrongleakagereductionofhaloprotonsfromthe SPS beamcollimationarea, therebyapproaching thecasewithan idealabsorber.

Acknowledgements

WewishtoacknowledgethestrongsupportoftheCERNEN-STI

and BE-ABPgroups. We also acknowledge the partial supportby

theRussianFoundationforBasicResearchGrants05-02-17622and 06-02-16912, the RF President Foundation Grant SS-3383.2010.2,

the “LHC Program of Presidiumof Russian Academy ofSciences”

and the grant RFBR-CERN12-02-91532. G. Cavoto and F.

Iacoan-geli acknowledge the supportfrom ERC IdeasConsolidator Grant

No.615089 “CRYSBEAM”. S. Dabagov acknowledges the support

by the Ministry of Education and Science of RF in the frames

of Competitiveness Growth Program of NRNU MEPhI, Agreement

02.A03.21.0005. Work supported by the EuCARD program GA

227579, within the “Collimators and Materials for high power

beams” work package (Colmat-WP). The Imperial College group

gratefullyacknowledgessupportfromtheUKScienceand Technol-ogyFacilitiesCouncil.

References

[1]R.Assmann,etal.,RequirementsfortheLHCcollimationsystem, LHC-PROJECT-REPORT-599in:8thEuropeanParticleAcceleratorConference:AEurophysics Conference,LaVilette,Paris,France,Jun2002,pp. 3–7.

[2]R.Assmann,S.Redaelli,W.Scandale,in:EPACProceedings,Edinburgh,2006, p. 1526.

[3]A.G.Afonin,etal.,Phys.Rev.Lett.87(2001)094802. [4]R.P.Fliller,etal.,Nucl.Instrum.MethodsB234(2005)47. [5] R.A.CarriganJr.,etal.,Fermilab-CONF-06-309-AD. [6]W.Scandale,etal.,Phys.Lett.B692(2010)78. [7]W.Scandale,etal.,Phys.Lett.B703(2011)547. [8]W.Scandale,etal.,Phys.Lett.B714(2012)231. [9]W.Scandale,etal.,Phys.Lett.B726(2013)182. [10]V.Uzhinsky,A.Galoyan,Phys.Lett.B721(2013)68. [11]S.Baricordi,etal.,Appl.Phys.Lett.91(2007)061908. [12]S.Baricordi,etal.,J.Phys.D,Appl.Phys.41(2008)245501. [13]A.M.Taratin,W.Scandale,Nucl.Instrum.MethodsB313(2013)26.

[14]D.Mirarchi,G. Hall,S. Redaelli,W.Scandale, Nucl.Instrum. MethodsB35 (2015)378.