February2,2008 8 9 9 1 The high-intensity hyperon beam at CERN n a J 7 ] h p Yu.A.Alexandrov1, M.Clement2, F.Dropmann3, A.Fournier4 P.Grafstro¨m.2, E.Hubbard3, S.Paul3,a, - c H.W.Siebert5,A.Trombini3 M.Zavertiaev1 c a s. 1MoscowLebedevPhysicsInst.,RU-117924,Moscow,Russia. c 2CERN;CH-1211Gene`ve23,Switzerland. i s 3HeidelbergMax-Planck-Inst.fu¨rKernphysik#;Postfach103980,D-69029Heidelberg,Germany. y h 4GrenobleISN;53AvenuedesMartyrs;F-38026GrenobleCEDEX,France p 5Heidelberg,Univ.,Physikal.Inst.#;Philosophenweg12,D-69120Heidelberg,Germany. [ 1 Abstract v 6 0 A high-intensity hyperon beam was constructed at CERN to deliver Σ− to experiment WA89 at the 0 Omega facility and operated from 1989 to 1994. The setup allowed rapid changeover between hyperon 1 0 and conventional hadron beam configurations. The beam provided a Σ− flux of 1.4 × 105 per burst at 8 mean momenta between 330 and 345 GeV/c produced by about 3×1010 protons of 450 GeV/c . Atthe 9 experiment target the beam had a Σ−/π− ratio close to 0.4and asize of 1.6 ×3.7 cm2. Thebeam particle / s trajectoriesandtheirmomentaweremeasuredwithascintillatingfibrehodoscopeinthebeamchannelanda c i siliconmicrostripdetectorattheexitofthechannel. Afasttransitionradiationdetectorwasusedtoidentify s y thepioncomponent ofthebeam. h p : v i X r a (SubmittedtoNuclearInstruments andMethodsA) a)NowatnowatTechnischeUniversita¨tMu¨nchen,Graching,Germany #)supportedbytheBundesministeriumfu¨rBildung,Wissenschaft,ForschungundTechnologie,Germany,undercon- 1 Introduction Our present experimental knowledge of the properties and interactions of hyperons comes mainly from hyperon beam experiments. In contrast to normal hadron beam lines, hyperon beams are characterized byveryshortlengthsoforder10m,whichrepresent acompromisebetweentheneedforadequate shielding betweentheprimaryhyperonproductiontargetandtheexperimentalapparatusandtheneedtoreducedecay losses of the hyperons passing along the beam line. At a momentum of 100 GeV/c, the hyperon decay − lengths vary from 7.1m for Λ to 1.5m for Ω , which leads to severe reductions of the hyperon fluxes available for experiments. This has induced a continuous quest for higher hyperon beam momenta, which started at about 25 GeV/c at the Brookhaven AGS and the CERN PS in the early seventies. The advent of the super proton synchrotrons at CERN and at FERMILAB in the mid-seventies led to the immediate construction ofnewhyperonbeamlinestoexploitthenewlyavailablehighenergies. At CERN,a charged hyperon beam with momenta between 70 and 135 GeV/c was operated between − − − 1976and1982intheCERNWestHall. Itachievedfluxesof12,000Σ ,400Ξ and10Ω perbeampulse, allofwhichweresuccessfully usedinexperiments. Forreviewsoftheseactivitiessee[1]and[2]. − ThequestforhigherΣ fluxesthenledtotheconstructionofanewhyperonbeamlineatCERN,capable of delivering in excess of 105 Σ− per beam pulse at 340 GeV/c to the Omega spectrometer facility in the West Hall. The beam was brought into operation in 1989-90 and was used for physics runs in 1991, 1993 and1994. Thedesignandoperation ofthebeamarethesubject ofthisarticle. − AtFERMILABthedevelopment ofhyperon beamsculminated ina600GeV/cbeam delivering Σ to theSELEXexperiment in1997[3]. 2 Design considerations 2.1 Boundary conditions Budget and manpower restrictions led to the decision to base the hyperon beam experiment WA89 on the existing Omega spectrometer facility in the CERN West Area [4], which was used by several different experiments each year. This imposed quite severe restrictions on the design of the beam, which are listed below: • TheOmegaspectrometer wassituated atthe end of theH1beam line inthe WestArea. H1isahigh energy hadron/electron beam used by a number of different experiments. The new hyperon beam had to be designed allowing for a short changeover time between the different experiments and the different modesofoperating H1. • InadditiontotheH1beamtherearefivemorebeamsintheWestAreacomplexoperatingatthesame time. A hyperon beam always implies dumping a large number of protons close to the experiment, whichcreatesanunavoidable muonbackground intheexperimental hall. Itwasmandatorytoreduce themuonfluxtotheneighbouring experiments asmuchaspossible. • The West Area, where the Omega spectrometer was located, is an open unshielded hall . It was necessarytocovermorethan100moftheH1beamandinparticularthesurroundings ofthehyperon production target with shielding. The maximum amount of shielding that could be installed without interfering too much with existing installations limited the proton flux to 5×1010 protons per pulse thereby imposingalimitontheintensity ofthehyperonbeam. 1 • Economically itwasinconceivable tobuild anynewbeamlinemagnetsandwethushadtorelyupon thosemagneticelementsbeingavailable. 2.2 Basicdesign Theconditions listed above left little freedom in the basic design [5]. The firstcondition made awobbling configuration anatural choice inorder topreserve thedirection oftheH1lineandnottobeforcedtomove theOmegaspectrometer. Thesecondconditionsuggestedabendingofthebeamintheverticalplaneinorder − not to sweep the muons horizontally into other beam lines. Condition three together with the wanted Σ fluxfixedtheacceptance ofthehyperon channel. Thefourth condition limitedthebending poweravailable forthebeam,thestrongestbendingmagnetsavailablebeingthreeH-typemagnets,3.6meterslongandeach withamaximumfieldof2.4T. At an early stage it was decided not to use any quadrupoles. This decision was dictated by space limitations andbythelossofintensity duetoanincreased beamlength. The choice of the beam momentum was a compromise between conflicting requirements. On the one − − hand, the Σ /π production ratio increases rapidly with the ratio of the hyperon beam momentum p Y and the primary proton momentum, x ≈p /p with p = 450 GeV/c , and the Σ− decay losses are F Y p p − reducedbyincreasingp . Ontheotherhand,theΣ productioncrosssectiondecreasesrapidlywithx as Y F σ ∝ (1−x )3. The limits on the primary proton flux mentioned above led us to the choice of p = 355 F c GeV/cforthemomentumonthecentraltrajectory ofthecollimator, whichcorresponds tox =0.8. F Fig. 1shows the basic design of the beam line. The 450 GeV/c protons are incident from the left and bentdown7mradbeforeinteracting intheproductiontarget. Downstreamofthetargettherewerethree2.4 T magnets all filled with brass and tungsten collimators. A channel with small transverse dimensions was embedded in the tungsten defining the passage of the beam. The first magnet contained the dump region discussedbelow. Themomentumspreadofthebeamwasdefinedbythechannelinthesecondmagnet. The thirdmagnetbentthebeambackontothegeometrical axisoftheOmegaspectrometer. 2.3 Flux calculations − − − − The Σ flux of the beam was estimated using the known π -flux and measured values of the Σ /π ratio − as a function of x (see Fig. 2). For the π flux we used measurements in the H1 and H8 beams both F − made at x = 0.8. As the momentum bite of the beam is quite large the x dependence of the π flux F F also enters into the flux calculation. Here we assumed an x -dependence of the form (1 − x )4. The F F result ofthe fluxcalculation asafunction of the central momentum ofthe channel is shown inFig. 3. The basic design is compared with a solution without wobbling. Also shown in figure 3 is the gain that would have been achieved with 3.0 T magnets had they been available. With these stronger magnets the required momentum spread of δp/p = 15% (FWHM) would have allowed for a larger angular acceptance of the magnetic channel andtherefore anincreased flux. Thisincrease influxisalso seeninfig. 3. Atthe chosen − − valueofthecentralmomentum,p = 355GeV/c,thecalculated Σ /π ratiowassomewhatbelow1. The c designparameters ofthebeamaregivenintable1. 3 Layout 3.1 Proton beam TheprimaryprotonbeamoftheSPSwithanenergyof450GeV/cwasextractedtotheWestAreainaslow extraction mode with a spill duration of 2.6 s and a cycle time of 14.4 s. The intensity was typically a few times1012 protonsperspill. Afirstintensity attenuation oftheprimarybeamwasprovidedbyaproduction 2 target (T1). This 500 mm long Be target produced secondaries for the neighbouring H3 beam. A further reduction in intensity by roughly a factor 10 was obtained by a fixed hole collimator 20m downstream of theT1target. Thiscollimatorconsistedof4ironblockswithtransversedimensionsof 80cm×120cmand with a length of 40 cm each in the direction of the beam. The blocks were equipped with tungsten inserts having small ”pinholes” of diameter 3.3 mm. The fine tuning of the intensity was achieved by a movable four-jaw collimator located 35 m further downstream. It should be pointed out that the T1 target and the subsequent collimatorswerelocatedinthehorizontal planeoftheacceleratorringi.e. some20mbelowthe surface. This means that the muon background coming from the proton beam collimation is negligible at thesurface, wherethehyperon beamandtheexperimentwerelocated. The distance between the T1 target and the hyperon production target was approximately 800 meters. The protons were transported over this distance via the H1 beam line. This was a general purpose, high energy and high resolution beam (for details see [6] ). The last 100 meters of the beam were modified in ordertomakespaceforthehyperonbeamupstream oftheOmegaspectrometer. The H1beam line is equipped with MWPCsand scintillator filament scanners normally used for beam tuning. Obviously, thosedetectors couldnotbeusedforbeamintensities ofseveral1010 protons perpulse. In order to tune the proton beam, attenuators consisting of 120 cm Be, i.e three interaction lengths, could beintroduced justdownstream oftheT1target. Togetherwiththeabovementioned four-jaw collimator this allowed to reduce the intensity down to the range of 107 protons per pulse suitable for tuning . Once the beamwastunedtheabsorbersweretakenawayandthecollimatorsopenedtoobtainthefullprotonintensity onthehyperon production target. Severaltypes ofsecondary emission monitors wereinstalled justinfront of the target in order to measure and monitor the proton flux and the position of the proton beam on the target. Fig. 4.a and 4.b show the horizontal and vertical profiles of the proton beam as measured with one ofthe secondary emission monitors just upstream ofthe target. Thehorizontal andvertical spot sizes were 250µm and 450µm, respectively (standard deviation). Thedivergence of the proton beam at thetarget was about0.1mradhorizontally andvertically. 3.2 Target choice Thetargetconsisted offourBerodseach100mmlongand2mmindiameter. Therodsweresuspended on thinaluminium foils(25µm)ateach end. Thetarget wasmotorized andcouldbemovedintwoorthogonal directions in a plane transverse to the beam. There is an interest to keep the diameter of the target as small as possible as long as the proton beam is well inside the target diameter. This minimizes secondary interactions which create low energy pions subsequently decaying into muons. The proton beam was well contained within the diameter of the target as can beseen from figures 4.a and 4.b. Beryllium was chosen as target material because the total inelastic cross section increases faster with A than the leading particle crosssection. Thismeansthat,perinteracting proton,thefraction ofhyperons islargestforlowAtargets. 3.3 Magneticchannel Fig. 5showstheshapeofthemagneticchannelinboththebendandnon-bendplane. Notethatthehorizontal axisisinmetersandtheverticalincm. Thechannelisembeddedintungsten. Tungstenischosen duetoits veryshortinteraction lengththusstopping asmanyparticles aspossible beforethedecay. Thebasic shape in both planes isawedge opening towards the exit. Sometolerance has been added to takeintoaccount possible misalignments betweensuccessive magnets. The shape of the first part of the channel is designed trying to minimise the muon background that originates fromsecondary pionsandkaonscreatedintheproduction targetanddecaying intomuonsbefore being absorbed in the walls of the channel. It turns out that the lateral dimensions are of small importance and what really matters is the longitudinal distance between the target and the dump. The distance can 3 − however not be too small in order to ensure the necessary lateral separation between the Σ -beam and the protonsatthepointwheretheyaredumped. Takingthisintoaccountthedistance waschosentobe2.5m. In the more downstream part of the vertical section ”pockets” have been added. The purpose of these pocketsistodiminishthenumberofinteractions inthewallsandtotrappartoftheshowersinthetungsten. There is a disadvantage with the wobbling configuration of the beam as can be seen from the shape of thechannelinthebendplane. Neutralparticlescanreachtheexitofthechannelfrompointsfarupstreamin thechannel. Thegeometryactuallyallowsneutralstoescapefromthechannelfrominteractionsinthewalls starting atadistance ofonly3.5mfromtheproduction target. Alocalphoton background wasobserved in theexperiment ataplace corresponding toanangle of7mradabovethebeam. Ascanbeseenfrom Fig. 5 thisisconsistent withthegeometryofthechannel. Toconfirmthatthisneutralbackground originates from the upstream part of the channel, 2 radiation lengths of lead were introduced between the second and the thirdmagnets. Thisreduced thebackground byafactor2-3. Moreover, iftheproduction target isremoved, thebackground disappears completely, showingthatthebackground doesnotoriginate fromthedump. After the first beam period in 1998-90 the shape of the channel was partially modified to diminish this neutral background. The modifications are indicated in fig. 5. The second pocket was moved upstream and the walls were shaped in such a way that neutral particles from the target can not reach the upstream wall of the pocket which is seen by the exit and that the downstream part where neutral particles may hit the channel wall can not be seen from the exit. The point where the protons are dumped was also moved slightly upstream in order to preserve the distance between the dump and the second pocket to avoid any − leakage into the pocket. The separation of the proton beam and the Σ -beam at the dump was in this way reducedfrom8mmto6mm. Fig. 6 shows a transversal cross section of the bending magnets. The pole gaps were filled with brass and tungsten collimators as mentioned previously. Each magnet had a fixed part and a removable part of inserts. The fixed part was made of brass for the two downstream magnets and of tungsten for the most upstream magnet. The removable part consisted of 90 tungsten block totally (30/magnet) with each block having its machined rectangular hole. The dimension of those blocks were 13 cm in the direction of the beamand3.5×18cm2 transverse tothebeam. 3.4 Detection andidentification ofbeam particles Thebeammomentumwasmeasuredwiththreehodoscopes,namelyonescintillatingfibrehodoscopeplaced beforethelastbendingmagnetandtwosetsofµ-stripsilicondetectorslocatedattwopositionsdownstream of the magnetic channel (figure 1) A scintillating fibre hodoscope was chosen to allow the measurement of the particle position in the bending plane (z) with 250/p(12)µm accuracy in the area of strong stray magneticfieldandinthestrongradiation environment ofthemagneticchannel. Thehodoscopewasconstructedofscintillatingfibres(KuraraySCSF-38)withrectangularcrosssection (1x0.5mm2). Thefibres werecovered bya55µmthick cladding towhichawhitepigment wasmixedto prevent crosstalk. Threefibreswereputtogether toform onechannel (3×0.5mm2,figure7). Thesethree were glued to one clear fibre (1.5 mm × 1.5 mm × 1 m) as light guide which transmitted the light to one pixel of a 64-channel photo-multiplier. The channels were arranged in two rows which were displaced by half a fibre width (0.25 mm). The hodoscope had a total of 64 channels, 32 in each row. A valid hit was a coincidencebetweentwofibresfromeachrow. Topreventcrosstalkatthetubecausingwrongcoincidences, neighbouring fibres were distributed over the 64 channels of the PM. Two such hodoscopes were used to obtainfullefficiency. At the exit of the beam channel the beam had to pass a hole of 5 × 5 cm2 defined by a veto counter consisting of2cmleadsheetsand5mmscintillator whichwasusedtosuppress thebeamhalo. − To reject high momentum π in the beam at an early stage of the trigger a fast transition radiation detector (TRD) was placed approximately 1 m downstream of the veto counter. This detector is described 4 indetailin[9]. Twoscintillators eachwithanactiveareaof3.5×3.5cm2 incoincidence placedinfrontof theTRDandtheexperiments targetwereusedtotriggeronbeamparticles. The trajectory of the beam particle in front of the experimental target was measured by a set of 8 high resolutionsiliconmicrostripdetectorswithanactiveareaof5.1×5.1cm2andwithapitchof 50µm. Aset of4planes wasplaced infront oftheTRDdetector, 4planes wereplaced behind the TRDdirectly infront of the experiment target. The detectors measured in two orthogonal directions. The lower part of Figure 8 showsthesetuparoundtheexperiment target. ParticlesemergingfromtheinteractionweredetectedintheWA89spectrometershownintheupperpart of figure 8. The secondary particles close to the interaction point were detected by 20 silicon microstrip planes with 25 and 50µm pitch. Two scintillators (3.5×3.5 cm 2 each) were placed in the middle of the vertex detector to measure the multiplicity of particles coming from the interaction. These detectors were usedinthetrigger. Positioning the target about 14 m upstream of the centre of the Ω-spectrometer provided a 10 m long decay area for short-living strange particles. The products of these decays along with the particles coming directly from the target were detected by 40 planes of drift chambers with a spatial resolution of about 300 µm. Special MWPC chambers (20 planes with 1mm wire spacing) were used in the central region of high particle fluxes. Inorder toimprovethe track matching between the target region andthe decay region threesetsof4MWPCseachwithapitchof1mmwereinstalledabout2mbehindthetarget. The particle momenta were measured by the Ω-spectrometer consisting of a super-conducting magnet with a field integral of 7.5 Tm and a tracking detector consisting of 45 MWPC planes inside the field area and 12 drift chamber planes at the exit of the magnet. The momentum resolution was σ(p)/p2 ≈ 10−4 (GeV/c)−1. Charged particles were identified using a ring imaging Cherenkov (RICH)detector . It had a threshold ofγ = 42andprovidedπ/pseparationuptoabout150GeV/c. DownstreamoftheRICHaleadglasselec- tromagnetic calorimeter was positioned for photon and electron detection . This calorimeter was followed byahadroncalorimeter. Foramoredetailed description see[10]andreferences therein. 4 Beam distributions Figures 9.c and 9.d show the beam profile in horizontal (y) and vertical (z) direction at the position of the experiment target about 3 m from the exit of the beam channel. The beam had a width of 1.6 cm in y and a width of 3.7 cm in z direction. The shape of the beam profile in z reflects the momentum distribution of thebeam. Thebeamprofileinthenonbending planeisaresultoftheprofileoftheproton beamandofthe position oftheberyllium target. Figure9aandbshowthecorrelationsbetweenthebeamparticleimpactpointanditsangleinthebending andnonbending plane. Allparticletrajectories inthenarrowbandsinfigure9pointbacktotheproduction − − target. Particles outside the correlation bands are π from Σ decays, particles elastically scattered in the beam channel or muons. The distribution in figures 9a and 9.b were used to measure the amount of background inthebeamfromthesesources(seebelow). Themomentum distribution ofthebeam wasmeasured withtheΩspectrometer. Forthismeasurement it was required that exactly one track was detected in the spectrometer to reject events with interactions. Themomentum measurement forthese non-interacting events wassubsequently used tocalibrate thebeam − hodoscope tobeable tomeasure thebeam particle momentaalsoforevents wheretheΣ interacted inthe target. The momentum distribution of beam particles for the 1994 data taking period is shown in figure 10. The solid line corresponds to the value measured by the beam hodoscope. The dashed line shows the momentum measurement in the spectrometer. This spectrum shows an additional component below 150 GeV/c stemming from the decays Σ−→nπ−. The measurement of the beam momentum gave average 5 momenta of345GeV/c forthe1993 beam timeand 330 GeV/c forthe beam timein1994 andawidthof ∆p/p=18%(FWHM)whichisclosetothebeamlinedesignvalueof∆p/p=15%. 5 Beam composition 5.1 Σ− identification − In the TRD high momentum π were identified by their characteristic transition radiation, which was de- tected in ten double planes of multiwire proportional chambers. If three of the ten planes gave a signal abovethreshold thebeamparticle wasidentified asahighmomentum pion andtheeventwassubsequently rejected intheexperiment trigger. Thepulseheightofeachchamberwasrecorded formonitoring purposes and to allow a more detailed offline analysis at a later stage. Offline the pulse height information from the TRD chambers along the track measured in the the microstrip detectors was used to calculate a likelihood for the different particle hypotheses. Thedistribution oflikelihoods forbeam particles before and after the TRD decision is shown in fig. 11. From a two-component fit to the distribution of likelihoods before the TRD decision (fig 11.a) a Σ−/π− ratio of 0.38 ± 0.04 at the position of the TRD can be deduced, which is in very good agreement with the design value of the beam line. With the help of the TRD information thepioncomponent foreventsintheexperiment triggerwasreduced toaΣ−/π− ratioof10±1withaΣ− detection efficiency of 82±1% Byacut on the TRDlikelihood inthe offline analysis a Σ− purity of 98.5 − %wasachievedwhilemorethan95%oftheΣ identifiedonlinewereretained. 5.2 Ξ− inthe beam − TomeasuretheamountofΞ inthebeam,asampleofeventswithbeamparticlespassingthroughthetarget − − without interaction was analysed. In this data sample we identified Σ and Ξ decays by the observation of decay kinks between the incoming beam particle and a negative particle detected in the spectrometer from the decays Σ− → nπ− and Ξ− → Λπ−, respectively. The decay lengths of Σ− and Ξ− at the same momentum differ by only 1% and therefore no acceptance or efficiency corrections are needed for a given − − fiducialvolumetoevaluatetheΞ /Σ ratio. Inparta)offig.12theexpectedcorrelation(generatedwiththeGEANTMonteCarlo)betweenmomen- − − tumandkinkangleforΣ andΞ decaysisshown. Theexperimentaldataforthecorrelationaredisplayed − in fig. 12b. A second band originating from Ξ decays can be discerned there. For a quantitative analysis weprojected thisdistribution along theangle-momentum correlation markedbythedashed andsolid lines. − Thisprojection isshownbythesolidhistogram inpartc)offig.12. Besides thedominating peak fromΣ − decays,awellseparatedmaximumcorresponding toΞ decayscanbeidentified. Theobserveddistribution can be reproduced by Monte Carlo simulations assuming aΞ− contribution of (1.3± 0.1)% relative to the − numberofΣ (seecross-hatched histogram infig.12.c). Thisratioisingood agreement withthenumbers givenin[8]. 5.3 Ω− andK− inthe beam − − TheproductionofΩ isexpectedtobeevenmoresuppressedthantheproductionofΞ duetotheadditional − unit of strangeness. In figure 12b a sizeable contribution of Ω would be visible in a band above the − − correlation for Σ in the momentum range of 70-160 GeV/c . In this region of the plot no Ω can be − − identifiedabovethebackground. TheexpectedΩ /Ξ ratioatthechannelexitcalculatedfromthenumbers givenin[8]is2×10−3 whichisbelowthesensitivity ofthemeasurement presented infigure12. − − − TheK component ofthebeam canbeestimated onthebasis oftheK toπ production ratio, which wasmeasured tobe(1.15±0.02)% [11]in300GeV/c p-Becollisions . SincetheK− arebelowtheTRD 6 − − threshold theywerenotrejected bythisdetector. FromtheK /π ratioatproduction wededuce thatthere − − wasa2.1%admixtureofK inthebeamaftertheonlinerejection ofthehigh-momentum π component. 5.4 Muon halo There are two different origins of the muon background. One part comes from showers created by the protons being dumped in the tungsten. The second part is due to secondary pions and kaons created in the production target whichdecayinto muonsbefore beingabsorbed inthewallsofthechannel. Thefirstkind of muon background is minimized by choosing a very dense material with a very short interaction length, such as tungsten. The second contribution depends on the geometry of the beginning of the channel as discussed in section 3.3. To obtain a rough idea of the muon background we have used the Monte-Carlo program HALO[12]. Thisprogram generates parent particles decaying intomuonsandthemuonsarethen tracked through the coils and yokes of the different magnets. The result of this calculation is shown in table 2, where muon densities for 2× 1010 protons are given at the entrance to the Omega spectrometer. Contributions from the dump and from the target have been separated. It can be seen that at the beam axis weexpectabout2×105 muons/m2 andthatthedominating partcomesfromsecondaries fromthetarget. Thelevelofthemuonbackgroundwasmeasuredwithscintillatorhodoscopeslocatedacoupleofmeters downstream ofthe exit ofthe Omegaspectrometer. About 2×105 muons/m2 per pulse weremeasured, in good agreement withthe calculations. Takingout the production target reduced thisrate byafactor of2.5, confirmingthetrendsfoundinthecalculations. 5.5 Backgrounds inthe beam Thetotalamountofbackground inthebeam canbeestimated fromtheplotsinfigure9. Inthisplotsabout 21 % of all measured particle trajectories do not fall into the angle-impact correlation which is given for − − particles coming from the production target. Of this 21 % about half could be assigned to π from Σ decaysasshownbyaMonteCarlosimulationofthebeamline. Therestcorrespondstomuonsandparticles elastically scattered inthecollimators. To measure the effect of the veto counter data were taken in a special run where the veto counter was not included in the experiment trigger. Removing the veto counter resulted in an increase of 5 % of the multiplicity dependent partofthetrigger rate. Thisincreasewascausedbyadditional particles crossing the apparatus together with the particles produced in the experiment target causing larger pulse heights in the multiplicity counters. 5.6 Interactions from Pionsand Neutrons − Withthegivenbeam composition itwaspossible toobtain inaddition totheevents withΣ induced inter- − actionsdatasamplesoninteractions inducedbyπ andneutrons: Toobtain aclean sample ofevents withpion interaction data wastaken with aninverted TRDdecision − for a small fraction of the runing time. In this data sample the Σ component of the beam was suppressed onlinebytheTRDtoalevelbelow2%. − ForΣ decayingupstreamofthecentreofthelastbendingmagnetthedecaypionfellintheacceptance ofthebeam trigger counters. Sincethemomentum ofthedecay pion isbelow 150GeV/c thiseventswere notrejectedbytheTRDandwereidentifiedwiththeangle-impact correlationasmentionedabove. Insome fraction of this events however an interaction was induced by the decay neutron. It was possible to isolate a clean sample of such events by requiring that the beam particle trajectory was outside the angle-impact correlationandnotpointingtothereconstructedinteractionvertex. Additionalyitwasrequiredthatthepion trackcouldbefollowedintothespectrometer andhadamomentumbelow150GeV/c. 7 − Neutron and π induced interactions were used as control data sets to be able to compare to other experiments usingthesebeamstypes[10]. 6 Effect of the vertical target position Theverticaltargetpositiondeterminedthepositionoftheproducedparticlesattheentranceofthemagnetic channelandthereforethecurvatureofthecentraltrajectorypassingthroughthecollimatorsofthemagnetic channel. By moving the target up and down it was hence possible to tune the average beam momentum. For each given target position the proton beam could be steered onto the production target by means of a trimmermagnet. Thetargetpositionwithmaximumchanneltransmissionwasgivenbythegeometricalacceptanceofthe − − collimatorsystem. SincetheΣ /π productionratioisstronglydependentontheaveragebeammomentum − thistargetsteering alsohadanimpactontheΣ content ofthebeam. Tofindanoptimalpointofoperationthetargetwasmoved±0.4cmrelativetoitsnominalzeroposition and the corresponding beam momenta and particle fluxes were measured. Figure 13 shows the results of thesemeasurements. Theaverageparticlemomentumvariedfrom310to360GeV/cforthisrangeoftarget positions (fig 13.a). The flux variation is shown in figure 13.b . At maximum transmission the flux was 5×105 particles (Σ− and π− ) for 3.0×1010 incident protons. In 1993 the beam was operated with the − − targetatposition-0.05whichgaveanaveragemomentumof345GeV/c,aΣ /π ratioof0.4andafluxof 1.38×105 Σ− particles atthechannel exit. 7 Conclusions A high-intensity double-bend hyperon beam was successfully operated at momenta between 330 and 345 GeV/c delivering 1.4× 105 Σ− to experiment WA89 in the CERN West Area. At the experiment target 3 m downstream of the channel exit the beam had a Σ−/π− ratio of 0.38 ± 0.04 and a size of 1.6 × 3.7 cm2. TheΞ−/Σ− componentofthebeamwasmeasuredtobe1.3±0.1%ingoodagreementwithprevious measurements. The muon and neutral backgrounds were substantially reduced by a careful design of the proton beam dump areaandthe shape ofthe collimators. Byanadjustment ofthe vertical target position it − − waspossible totunethebeammomentum,theparticlefluxandtheΣ /π ratiotoanoptimalvalue. 8 Acknowledgement The members of the WA89 collaboration are grateful to the EA group for their continuous support in the construction and operation of the hyperon beam line, to the SPS staff for providing good beam conditions andtotheOmegagroupfortheirhelpinrunningtheOmegaspectrometer. WewouldliketothankJ.Zimmer forhisinvaluable support duringallphasesofdetectordevelopment andsetup. References [1] M.BourquinandJ.-P.Repellin,PhysicsReports114,(1984) 99 [2] L.Pondrom,PhysicsReports122,(1985) 57 [3] R.Edelsteinetal.,E781proposal, Fermilab,1987 [4] Proposal for a new hyperon beam experiment at the CERN-SPS using the Omega facility, CERN/SPSC-87/43,SPSC-P/233 8 − [5] P. Grafstro¨m, A study of a high-energy Σ beam for the Omega spectrometer in the West Area. CERN/SPSC88-4andCERN/SL90-104 [6] D.E.Plane,TheWestAreaBeamsCERN/SPS/83-22(EBS) [7] M.Bourquinetal.,Nucl.Phys.B153,(1979) 13 [8] T.R.Cardelloetal.,Phys.RevD32,(1985)1 [9] W.Bru¨ckneretal.,Nucl.Instrum. MethodsA378,451(1996) [10] M.I.Adamovich.etal.,Z.Phys.C76,35-44(1997) [11] H.W.Athertonetal.,CERN80-07,CERN,Geneva,Switzerland(1980) [12] Ch.Iselin,HALO - AcomputerprogrammetocalculateMuonHalo.CERN74-17 Lengthofthebeamfromtarget toexitofchannel(m) 12.7 B×l(Tm) +17.2;-8.6 Deflection(mrad) +13.8;-6.9 Centralmomentum(GeV/c) 355 Acceptance (±2σ) α (mrad) ±1.5 H α (mrad) ±0.6 V α (mrad) ±12 V Spotsizeatexit(x×y mm2) 37×16 FluxofΣ− for3×1010 ppp 1.4×105 Σ−/π− ratioatchannelexit 0.38±0.04 − Table1: ParametersoftheΣ beam Muons/m2 atthebeamaxis Muons/m2 0.85mabove(µ−) overanareaof1×1m2 andbelow(µ+)thebeamaxis overanareaof0.3×0.3m2 µ+ µ− µ+ µ− Fromtarget 2.2×104 1.3×105 1.3×106 9.6×105 Fromdump 0.3×104 0.2×105 0.1×106 0.5×105 Total 2.5×104 1.5×105 1.4×106 1.0×106 Table2: Calculated muons/m2 attheentranceoftheOmegaspectrometer for2×1010 incidentprotons 9