ChargedKaonMassMeasurementusingtheCherenkovEffect N.Graff A.Lebedevd R.J.Abramsk,n U.Akgunj G.Aydinj W.Bakerc P.D.Barnes,Jr. g 0 T.Bergfeld(cid:96) L.Beverlyc A.Bujakh D.Careyc C.Dukesm F.Duruj G.J.Feldmand A.Godley(cid:96) 1 0 E.Gu¨lmezj,o Y.O.Gu¨naydınj H.R.Gustafsonk L.Gutayh E.Hartounig P.Hanlete S.Hansenc 2M.HeffnergC.JohnstonecD.KaplaneO.KamaeveJ.KilmercJ.KlaygM.KostincD.LangegJ.Ling(cid:96) n M.J.Longok L.C.Lum C.Materniakm M.D.Messierf H.Meyerc,p D.E.Millerh S.R.Mishra(cid:96) a K.Nelsonm T.Nigmanovk A.Normanm Y.Onelj J.M.Paleyf H.K.Parkk A.Penzoj R.J.Petersoni J R.Rajac D.Rajaramk D.Ratnikove C.Rosenfeld(cid:96) H.Rubine S.Seund N.Solomeye,p R.Soltzg 4 E.Swallowb R.Schmittc P.Subbaraok Y.Torune T.E.Topec K.Wilson(cid:96) D.Wrightg K.Wu(cid:96) ] x aBrookhavenNationalLaboratory,UptonNY11973 e bElmhurstCollege,Elmhurst,IL60126 - cFermiNationalAcceleratorLaboratory,Batavia,IL60510 p dHarvardUniversity,Cambridge,MA02138 e eIllinoisInstituteofTechnology,Chicago,IL60616 h fIndianaUniversity,Bloomington,IN47403 [ gLawrenceLivermoreNationalLaboratory,Livermore,CA94550 2 hPurdueUniversty,WestLafayette,IN47907 v iUniversityofColorado,Boulder,CO80309 1 jUniversityofIowa,IowaCity,IA52242 7 kUniversityofMichigan,AnnArbor,MI48109 9 (cid:96)UniversityofSouthCarolina,Columbia,SC29201 0 mUniversityofVirginia,Charlottesville,VA22904 . nCurrentlyatMuons,Inc.,Batavia,IL60510 9 oAlsoatBogaziciUniversity,Istanbul,Turkey 0 pCurrentlyatWichitaStateUniversity,Wichita,KS67260 9 0 : v iAbstract X rThetwomostrecentandprecisemeasurementsofthechargedkaonmassuseX-raysfromkaonicatomsandreportuncertaintiesof a 14ppmand22ppmyetdifferfromeachotherby122ppm.Wedescribethepossibilityofanindependentmassmeasurementusing the measurement of Cherenkov light from a narrow-band beam of kaons, pions, and protons. This technique was demonstrated usingdatatakenopportunisticallybytheMainInjectorParticleProductionexperimentatFermiNationalAcceleratorLaboratory which recorded beams of protons, kaons, and pions ranging in momentum from +37 GeV/c to +63 GeV/c. The measured value is491.3±1.7MeV/c2,whichiswithin1.4σ oftheworldaverage.Animprovementoftwoordersofmagnitudeinprecisionwould makethistechniqueusefulforresolvingtheambiguityintheX-raydataandmaybeachievableinadedicatedexperiment. 1. Introduction X-ray energies from kaonic atoms. While these measure- ments report uncertainties of 14 and 22 ppm they differ The charged kaon mass is an important input in deter- by 122 ppm (4.6σ). In this article we explore one possibil- mining the CKM matrix element V from measurements ity to resolve this discrepancy using an independent tech- us of the branching ratio of K+ → π0e+ν. The value of the nique for measuring the charged kaon mass based on the chargedkaonmassreportedbytheParticleDataGroupis Cherenkoveffect.Thewellknownpionandprotonmasses 493.677MeV/c2withanuncertaintyof26partspermillion are used as references. The technique is demonstrated us- (ppm)[1].Thisvalueisaweightedaverageofsixmeasure- ing data taken opportunistically using the Ring Imaging mentsbutisdominatedbythetwomostrecentandprecise Cherenkov (RICH) sub-detector of the Main Injector Par- measurementsfromDenisov[2]andGall[3]whichmeasure ticleProduction(MIPP)experimentatFermilab[4]. PreprintsubmittedtoNuclearInstrumentsandMethods 4January2010 2. MeasurementConcept the differences in the Cherenkov angles are largest, while stayingaboveprotonthreshold. Cherenkov light is emitted when a relativistic charged In a RICH detector, the angle θ can be determined on particle of mass m, momentum p, and speed β = atrack-by-trackbasisfromthepatternofCherenkovpho- 1/(cid:112)1+(m/p)2 travels through a radiator volume of in- tons recorded. However, the light for a single ring will be dex of refraction n with β > 1/n. (As is customary in distributed about the central angle θ due to the variation high energy physics we work in units in which the speed of the index of refraction of the radiator medium over the of light, c, is 1.) Neglecting dispersive effects for the mo- wavelengths at which Cherenkov photons are produced. ment, the light is emitted in a cone at an angle θ given by Thisgivesascontributiontotheuncertaintyintheaverage cosθ =1/βn[5,6]whichisapproximately angleθ determinedfromasingletrackof (cid:114) 1 1 (cid:18) 1 (cid:19)2 θ = 2(1− ) (1) σ(cid:48)2 = δ2, (7) nβ θi Nh θin2βi n for small angles. Now, consider two particles i and j with where δ is the amount of dispersion over the photomul- n identicalmomentapbutdifferentmassesm ,m andspeeds tiplier tube (PMT) wavelength acceptance and N is the i j h β ,β .TheywillemitCherenkovlightatanglesθ ,θ which numberofPMThitsinthering. i j i j arerelatedbytheexpression Inabeamline,particleswillbeacceptediftheirmomen- tum lies in a narrow window about some central value p. β θ2−β θ2 =2(β −β ). (2) i i j j i j Thefinitesizeofthismomentumacceptancewindowintro- duces an additional uncertainty in the average angle mea- In the relativistic limit p (cid:29) m, β ≈ 1− m2. This, when 2p2 suredfromasingletrack.AveragingoverN rings,themo- combinedwithEqn.2,gives r mentumspreadcontributesanuncertainty θ2−θ2 = m2j −m2i, (3) 1 (cid:34) (cid:18)m2β (cid:19)2 (cid:35) i j p2 σ2 = σ(cid:48)2 + i i (σ /p)2 (8) θi N θi θ np2 p r i where we have neglected the small difference between βθ andθ.Iftheparticlesiandjarepions,protons,andkaons, to the measurement of θ, where σ is the spread in the p wehavetwoindependentangle-squareddifferencesthatcan beamparticlemomentaabouttheircentralvalue.Wetake bemeasured the specific case of an experiment using CO as the radi- 2 ator (n = 1.00045, δ = 3×10−5) and a beam of central m2 −m2 m2−m2 n θ2 −θ2 = K π, and θ2 −θ2 = p π. (4) momentum 40 GeV/c with a width of σ /p=0.01. Under π K p2 π p p2 p these assumptions, σ values are in good agreement with θi Usingthese,thekaonmassisgivenby observed widths of ring radius distributions. This gives a statisticaluncertaintyinthekaonmassof θ2 −θ2 m2 =m2 +∆ π K, (5) K π pπ θ2 −θ2 σ2 31.02 44.12 18.12 π p mK =0.62+0.92+ + + [ppm]2, (9) m2 N N N where ∆ ≡ m2 −m2. Notice that for a monochromatic K π K p pπ p π beam in the absence of dispersion the index of refraction where the first term results from uncertainties in the pion n and momenta p drop out. The kaon mass can be deter- mass,thesecondfromuncertaintiesintheprotonmass,and mined,inprinciple,throughmeasurementsofthepionand thefinalthreetermsresultfromuncertaintiesintheangles proton masses and the Cherenkov angles of the three par- θ , θ , and θ with N , N , and N being the number π K p π K p ticles. The proton and pion masses are known to 0.9 ppm ofmillionsofpion,kaon,andprotonringsrecorded.Using and 2.5 ppm respectively and will not be the limiting fac- Eqn.9wefindthattheuncertaintyisminimizedif32%of torsintheexperiment. the data is collected using protons, 23% using pions, and Using Eqn. 5 we can estimate the uncertainty in m2 45% using kaons. This result is only weakly dependent on K measuredusingthismethodas: momentumasshowninFigure1whichplotstheexpected statisticalprecisionofthechargedkaonmassasafunction (cid:18)θ2 −θ2 (cid:19)2 of momentum choice and total number of rings recorded. σ2 =σ2 + π K σ2 + m2K m2π θπ2 −θp2 ∆pπ With 10 million rings at 40 GeV/c, we expect a statistical (cid:34) (cid:35) precisionof30ppmusingthistechnique. ∆2 ∆2 4p4 θ2 pKσ2 +θ2 σ2 +θ2 Kπσ2 . (6) π ∆2 θπ K θK p ∆2 θp pπ pπ 3. RICHdetectoroverview Thefirsttwotermsareduetotheuncertaintiesinthepion and proton masses and are small. The third term grows The RICH detector used by MIPP was built by the SE- with momentum and suggests that it is best to conduct LEXCollaboration[7]foruseinthatexperiment.Wesum- themeasurementataslowamomentumaspossiblewhere marize here only the most important features of the de- 2 m] 30 pion p p Error [200 1200 Mass 150 M E vents -100 1 -20 -30 100 3M Events 50 10M Events 30 kaon 30M Events 20 10 m 0 c0 35 40 45 50 55 60 65 -10 Momentum [GeV/c] -20 -30 Fig.1.Expectedstatisticaluncertaintyforkaonmassmeasurement. 30 proton tector as deployed for the MIPP experiment and refer the 20 readerto[8–11]fordetails. 10 0 Phototube Array -10 -20 -30 Beam -80 -60 -40 -20 c0m 20 40 60 80 BWeinadmow Rings 105 Fig.2.Schem1amticoftheRICHdetecMft =oir r1ro0[rm8]. Reconstructed 110034 ThegeometryoftheRICHcounterisshowninFigure2. 102 Thedetectorwasconstructedfromalowcarboncylindrical 10 steel vessel 10.22 m in length and 93 in. in diameter with a wall thickness of 1 in. The ends were sealed with 1.5 in. 1 2 10 15 20 25 30 35 thickaluminumflangesthatwerecutouttoholdthinbeam Radius [cm] windows at each end and a photomultiplier tube holder plate at the upstream end. The vessel was wrapped by a Fig. 3. Sample event displays of 40 GeV pion (top), kaon (second water line carrying chilled water and 15 cm of building from top), and proton (second from bottom) rings in the RICH counter. Small circles indicate hit PMTs. The large dashed circles insulationtoregulatethetemperature. show the rings reconstructed from the PMT hits. The bottom plot A2.4m×1.2marrayof16mirrorsmountedatthedown- isthedistributionofreconstructedringradiifor40GeV. streamendofthecounterfocusedCherenkovlightonanar- rayof1/2in.photomultipliertubes.Onaveragethemirrors of PMTs were used: Hamamatsu R760 and Russian made had a radius of curvature of 1980 cm with variations less FEU60. The FEU60 tubes were coated with a wavelength than5cmandareflectivityofmorethan85%at160nm. shifter to match the acceptance region of the R760’s. The MIPPusedCO2gasheldatjustaboveatmosphericpres- R760(FEU60)tubeshaveamaximumquantumefficiency sure as the radiator. At STP, CO2 has an index of refrac- of about 25% (11%) at 350 nm. Of the available 89 PMT tion of 1.00045 at λ = 300 nm giving thresholds of p = columns,68wereused,withR760’sinstalledin15columns 4.5GeV/cforpions,17GeV/cforkaons,and31GeV/cfor and FEU60’s installed in 53 columns. The front-end elec- protons. The gas temperature and pressure were continu- tronics of the RICH detector was re-designed and rebuilt ouslymonitoredenablingcalibrationoftheindexofrefrac- bytheMIPPcollaboration. tion.Aβ =1particleproducedaringofradius29.5cmand an average of 30 PMT hits providing 3σ π/K separation up to 80 GeV/c and 3σ p/K separation up to 120 GeV/c. 4. DataAnalysis Figure3showssampleeventdisplaysforeachbeamspecies showingPMThitsandreconstructedrings. In outlining the measurement concept, we made several The PMTs were mounted in a hexagonal array behind simplifyingassumptionswhichmustbeaccountedfordur- a 2 mm thick quartz window which provided a gas-tight ing analysis. Due to dispersion, Cherenkov light is not ob- sealbetweenthephototubesandtheradiatorvolume.They servedatasingleangle,butasadistributionacrossseveral were read out in threshold (on/off) mode. Two models angles.Inouranalysis,wemeasuredtheaveragePMToc- 3 cupanciesforpion,proton,andkaonringsasafunctionof Cherenkovangleindataandcomparedthemtocalculations absorption in radiator 1 which incorporated the kaon mass as a free parameter. In quartz efficiency total 12 million rings were recorded for this measurement mirror reflectivity usingpositivelychargedbeamsofpions,kaons,andprotons cone reflectivity rangingfrom37GeV/cto63GeV/cinmomentum. In the data, wire chambers upstream and downstream of the RICH counter were used to reconstruct the particle nor m ttwarhaaPasjtMeacpsTtrsoeirgdtynoiceatfidnirodtenopwferoaaerscdhtichcoPtemMtchpeTeunttcaeeendrndptaeotsrhsaieotfiafoutvnhne,ercaatRigoCeInChpeHorrofebrnCiankhbogeiv.lrieUtanynskigfnoolvger 0.5 PMT efficiaelinzecd yphotoelectrons angle. This procedure was done separately for each PMT twyapse,mmaidrerowr,itahndbemaommpenatrutimclesse,ttliignhgt.Astsruthckemonelaysutrheemtewnot normalized photons 0 centralmirrors,labeled8and9. 200 400 600 To compare to the data, we calculated the expected oc- Wavelength [nm] cupancyofeachPMTintheRICHarrayforpion,proton, and kaon rings. The calculation starts with the number of Fig.4.Overlayofallefficiencyfunctionsincorporatedintothecalcu- Cherenkovphotonsproducedperunitpathlengthbetween lationofdetectorphotoelectronsasafunctionofangle.Alsoshown wavelengthsλandλ+dλ[5,6]: aretheinitialCherenkovphotonproductionspectrumandfinalpho- toelectronyieldnormalizedtopeakat1forinclusionintheplot. d2N 2πα dλdpxh = λ2 sin2θC, (10) P =1−exp(−Npe−b), (13) where α is the fine structure constant. These photons wherebaccountsforthePMTdarknoiserate. travel through the radiator medium, reflect off mirrors, The measured and calculated PMT occupancies were pass through a quartz window, and are collected by a re- comparedusingaχ2statistic.Inthecomparisonseveralpa- flectiveconebeforetheyareincidentonthephotodetector. rameterswereallowedtovarytominimizeχ2.Thesewere: The transmission probability for a Cherenkov photon of i. Intrinsicdetectorsmearingwidth,σ 0 wavelength λ for each of these steps is plotted in Figure 4 ii. Dispersivesmearingwidth,σ N along with the photodetector efficiency. Combining these iii. Densityratioscalingfactor factors, the average number of photoelectrons detected by iv. Densityratiooffset PMTiisgivenby v. LevelofaircontaminationinCO 2 vi. Centralpionmomentum (cid:90)L(cid:90)θ2(cid:90)λ22πα(cid:18) 1 (cid:19) vii. Widthofmomentumdistribution Npie= λ2 1− n2(λ)β2 e−µ(λ)(FL+x)⊗ viii. Kaonmass The density ratio is defined as the gas density inside the 0 θ1 λ1 S(θ,θ (λ))(cid:15)(λ)G (θ)dλdθdx (11) RICHdividedbydensityatSTPandwasvariedtoaccount C i foruncertaintiesinthecalibrationofthepressureandtem- where(cid:15)(λ)istheproductofallwavelength-dependenteffi- peraturemonitorsinstalledinthecounter. ciencyfactors,µ(λ)istheabsorptioncoefficientofCO2,FL Themomentumacceptanceofthebeamlinewasmodeled is the mirror focal length, and Gi is the geometric accep- as a Gaussian and uncertainties in the mean and width of tanceoftheith PMT.Scatteringoflightfromangularbin thisacceptancewereincorporatedintothecalculation.The θC to bin θ is accounted for by the function S(θ,θC). The average proton and kaon momenta accepted by the beam- lightscatteringwasmodeledasaGaussianofwidthσwith linedifferslightlyfromtheaveragepionmomentumdueto threecomponents.Theseincludeanintrinsictermindepen- variations in particle production at the secondary target. dentofwavelengthandawavelength-dependentdispersive Thiseffectwasincorporatedusingasimulationofthecop- term[12].Athirdtermaccountsformultiplescatteringof persecondarytargetusingtheFLUKA[13,14]production thebeamparticleintheRICHradiator: model. After all known effects were accounted for there were σ2 σ2 =σ2+ N +σ2 (12) O(10%) differences between the measured and calculated 0 tan2θ ms C PMT occupancies, due presumably to our incomplete un- where σ is the intrinsic scattering width and σ is the derstandingofthefactorswhichcontrolCherenkovphoton 0 n dispersivescatteringwidth. Thecalculatedshapesareim- production and transport in our detector. To account for plicitly functions of the particle masses through β. Using these differences, the uncertainties in the PMT occupan- Poisson statistics, the probability for a PMT seeing N cies were increased until χ2/ndf = 1 was achieved at the pe photoelectronstobeonis: minimum.Thescalingfactorrangedfromabout25to100 4 for the different combinations of PMT type, mirror, and beammomentum.Systematicuncertaintiesoverwhelmthe Occupancy0.4 DPraetadiction Occupancy0.15 DPraetadiction s4t0aGtisetVic/aclduantcaeristasihnotwiens fionrFtihgiusrdea5t.aPseertc.eAntsdaimffeprleencfietbfoer- Mirror 9 Hamamatsu 0.2 Mirror 9 FEU 0.00.51 tTnlawahretgeeeeydnradtbghyarateankeaatowhnniedthrspitinnragetds5is.i-tc1Tit5ceha%dilsouodcnvciecffureeprrttaheanneinccyateyniigssinustlhyeaoaprwcirchnaanalilnngyugF1uei5lgailurtluirmbemien6is-.. V] 0 0.028 0.03 0.032 0 0.028 0.03 0.032 Me Fit Result Pion Ring Angle [rad] Pion Ring Angle [rad] s [ World Avg. s a M Occupancy0.2 DPraetadiction Occupancy0.1 DPraetadiction Kaon 500 Mirror 9 Hamamatsu 0.1 Mirror 9 FEU 0.05 450 0 0 40 50 60 0.015 0.02 0.015 0.02 Proton Ring Angle [rad] Proton Ring Angle [rad] Momentum [GeV] Occupancy 0.4 DPraetadiction Occupancy0.15 DPraetadiction Fbeiga.m7.mSoummenmtaaraynodfmkairornorsm.aAstseraecshultmsofmorendtautma sseetttsinugs,inrgesuvaltrsiofuosr Mirror 9 Hamamatsu 0.2 Mirror 9 FEU 0.00.51 mmtfrhoiierrmrrfioontrrha98elakParaerDoenpGlpom.ltoattestdsedrsehssiufhtlietfd,teasdnlidgshltitghlhyetdtlyootttthoeedtrhliigenhetle.isfTt,thheweshaoiclleicderpleitnseuedlstvhsaolfwuoesr KaonmassresultsforeachdatasetareshowninFigure7. Using just the low momentum data sets results in a kaon 0 0 massmeasurementof491.9±1.1MeV;usingjustthehigh 0.024 0.026 0.028 0.03 0.024 0.026 0.028 0.03 Kaon Ring Angle [rad] Kaon Ring Angle [rad] momentumdatasetsgives486.7±3.0MeV.Theseresults agreewithinuncertaintiesandarecombinedtogiveafinal Fig. 5. Comparison of measured PMT occupancies to best-fit pre- resultforthechargedkaonmassof491.3±1.7MeV(3500 dictions for 40 GeV data set [15]. Pions are shown in the top row, ppm).Theconventionofscalingerrorbarswhencombining protons in the middle row, and kaons in the bottom row. In each measurements so that χ2/ndf = 1 is followed resulting in row,theleftpanelshowstheresultsforR760PMTswhiletheright a larger final uncertainty. This result is within 1.4σ of the panelshowstheresultsforFEU60PMTs. PDGvalue[1]. See[15]foradetaileddiscussionoftheanalysispresented inthispaper. % Difference 1105 % Difference 105 5. SuggestionsforImprovement matsu 5 FEU 0 In the analysis above we achieved an understanding of ma 0 the PMT occupancies as a function of Cherenkov angle at Ha -5 -5 the 10% level resulting in a final uncertainty in the kaon -10 -10 mass of 3500 ppm. If the uncertainties in the PMT occu- pancies could be reduced to 1% the contribution of this -15 -15 source of uncertainty in the kaon mass would be reduced 0.026 0.027 0.028 0.029 0.026 0.027 0.028 0.029 Kaon Ring Angle [rad] Kaon Ring Angle [rad] to the level of the statistical uncertainties and make this technique useful for resolving the 122 ppm discrepancy in theX-raymeasurements. Fig.6.Percentdifferencebetweendataandpredictedoccupancyfor 40 GeV/c kaons. Hamamatsu tubes are on the left, FEU tubes are The largest uncertainty in our calculation of the PMT ontheright. occupancyasafunctionofangleresultsfromknowledgeof the acceptance of the PMT array which was complicated 5 by cross-talk in the readout electronics – a problem which sure the charged kaon mass with 3500 ppm precision us- manifesteditselfduringthehighrateconditionsthedetec- ing an existing RICH detector. We believe that there are torwasoperatedunderforthismeasurement.Occasionally, significantopportunitiesforafutureexperimentdedicated a single PMT hit would cause all 16 channels sharing the to the kaon mass measurement to reduce systematic un- samereadoutboardtoregisterahit.Inouranalysiswere- certainties to the level where this technique may be useful jectedPMThitsthatappearedtobeduetocross-talk.As to resolve the 122 ppm discrepancy in the X-ray measure- thecross-talkwasdifficulttomodel,itsaffectonthePMT mentsofthechargedkaonmass. acceptances has large uncertainties accountingfor most of the ∼10% uncertainty in PMT occupancy as a function of Acknowledgments angle. Additional uncertainties in the occupancies result from Fermilab is operated by Fermi Research Alliance, LLC our treatment of the detector response as a function of underContractNo.DE-AC02-07CH11359withtheUnited wavelength. A measurement of the PMT response on a StatesDepartmentofEnergy. tube-bytubebasiswouldhaverequireddis-assemblyofthe detector and was not undertaken, rather we used a single averageresponseforallPMTs.Astheresponseofindivid- References ual PMTs typically differ from the average by about 15%, thisintroducesanuncertaintyof∼6%inthecalculationof [1] C.Amsleretal.,“Particledatagroup,”Phys.Lett.B6671 occupancyversusangle.Likewise,weassumedidenticalre- (2008)704. flectionefficienciesforthetworeflectingmirrors.Takento- [2] A.Denisovetal.,“NewmeasurementofthemassoftheK− gether,theuncertaintiesinthePMTacceptancesandspec- meson,”JEPTLett.54(1991)558. [3] K.Galletal.,“PrecisionmeasurementoftheK− andσ− tralresponseoftheopticalsystemcontribute3400ppmto masses,”Phys.Rev.Lett.60(1988)186. ourmeasurementofthekaonmass. [4] “MIPPCollaboration.”. During the run the temperature and pressure of the ra- http://ppd.fnal.gov/experiments/e907/. diator gas varied considerably and the index of refraction [5] V.Zrelov,CherenkovRadiationinHigh-EnergyPhysics. Israel hadtobecalibratedseveraltimesaday.Weestimatethat ProgramforScientificTranslations,Ltd.,1970. [6] J.Jelley,CherenkovRadiationanditsApplications. Pergamon uncertainties in the knowledge of index of refraction con- Press,1958. tributed 500 ppm in our kaon mass measurement. The [7] “SELEXCollaboration.”.http://fn781a.fnal.gov. spreadonthebeammomentumaboutitscentralvaluecon- [8] J.Engelfriedetal.,“TheSELEXphototubeRICHdetector,” tributedroughly50ppmtothekaonmassmeasurement. Nucl.Instr.andMeth.A431(1999)53–69. Toachieve40ppmprecisionwiththistechniqueeachof [9] J.Engelfriedetal.,“TheRICHdetectoroftheSELEX experiment,”Nucl.Instr.andMeth.A433(1999)149–152. theuncertaintieslistedabovewouldneedtobeaddressed. [10]J.Engelfriedetal.,“Thee781(SELEX)RICHdetector,” Thelargestimprovementsaretobegainedbyusingasim- Nucl.Instr.andMeth.A409(1998)439–442. pleropticalsystemwhichcouldbecompletelycharacterized [11]L.Stutte,J.Engelfried,andJ.Kilmer,“Amethodtoevaluate and operated at a more consistent pressure and tempera- mirrorsforCherenkovcounters,”Nucl.Instr.andMeth.A turefreefromelectroniccross-talkproblems.Forexample, 369(1996)69. [12]R.FortyandO.Scheider,“RICHpatternrecognition.”Lhcb theopticalsystemwouldbesimplerifonlyasingleprimary technicalnotenumberlhcb/98040,1998. http: mirror could be used and the transmission windows elim- //lhcb.web.cern.ch/lhcb-rich/html/lhcb_rich_notes.htm. inated. Use of a fine-grained optical detector would allow [13]A.Fasso’,A.Ferrari,J.Ranft,andP.Sala,“FLUKA:a forthedevicetobemademuchmorecompactallowingfor multi-particletransportcode.”Cern2005-10,infn/tc 05/11, better control of pressure and temperature. A more com- slac-r-773,2005. [14]A.Fasso’etal.,“ThephysicsmodelsofFLUKA:statusand pact device could be built to run with large excursions in recentdevelopments,”inComputinginHighEnergyand gaspressureandtemperatureallowingforthevariationin NuclearPhysicsConference. 2003. the index of refraction with temperature and pressure to [15]N.J.Graf,MeasurementoftheChargedKaonMasswiththe be studied in situ. Finally, a charge based readout (rather MIPPRICH. PhDthesis,IndianaUniversity,Bloomington, than the on/off readout used in the above analysis) would IN,August,2008. [16]C.D.R.Azevedoetal.,“TowardsTHGEMUV-photon allow for better pixel-by-pixel knowledge of the detector detectorsforRICH:onsinlge-photondetectionefficiencyin response. This configuration could be achieved, for exam- Ne/CH4andNe/CF4,”arXiv:0909.5357[physics.ins-det]. ple, by replacing the secondary mirror in the MIPP beam Cherenkov counter with a GEM-based photo-cathode [16] which would afford 30 micron resolution of the Cherenkov rings in a device with is 45.7 cm in diameter and 22.9 m long. A complete assessment of the performance of such a devicewouldrequireasignificantresearcheffortandawaits an experimental proposal. We hope that this initial work willinspirethesefutureinvestigations. Inconclusion,wehaveusedtheCherenkoveffecttomea- 6