ssc. -o.a" ,°rces

SSC DETECTOR R&D PROPOSAL

DEVELOPMENT OF TECHNOLOGYFOR PIXEL VERTEX DETECTOR

University of California at Berkeley -SpaceSciencesLaboratory,

University of California at Davis,HughesAircraft Company,

University of IowaIowa State University,

Lawrence Berkeley Laboratory,University of Oklahoma

University of PennsylvaniaPrinceton University

Stanford Linear Accelerator CenterYale University

October2, 1989

1

Collaboration Members

1. UCB-SSL: John Arens,GarrettJernigan

2. UC Davis: David Pellett

3. Hughes Aircraft Company: Frank Augustine, William Flaugh, GordonKramer,Carl Pfeiffer, Ken Reyzer,David Wiggins, John Winterberg

4. University of Iowa: Ed McCliment

5. Iowa State University: JohnHauptman

6. Lawrence BerkeleyLaboratory: SteveHolland, David Nygren, HelmuthSpieler,Michael Wright

7. University of Oklahoma: Phil Cutierrez,Pat Skubic

8. University of Pennsylvania:Nigel Lockyer

9. Princeton University: Kirk MacDonald

10. SLAC: StephenShapiro

II. Yale University: Paul Karchin

ContactPersons

2

Contents

1 ExecutiveSummary 7

2 Introduction

3 Relevanceof Pixel Vertex Technologyto the SSC 11

3.1 Overview 11

3.2 IntermediateP Physicsand VertexDetection . . 11

3.2.1 BCD Vertex Detector 13

3.3 High P Region and Vertex Detection 16

3.3.1 SDE Vertex Detector 16

4 Technical Discussion 20

4.1 SystemDesignProcess 20

4.2 Design,Fabrication,and Testof a PrototypePixel Detector 22

4.3 Designand Fabricationof the Control and DataAcquisition Electronics 23

4.4 Pixel DetectorTests 24

4.4.1 Laboratory Tests 24

4.4.2 RadiationDamageStudies 25

3

4.4.3 BeamTests . 26

S Technical ProgressSummary 29

5.1 UCB/SLAC/HUGHESGenericProgram 29

5.1.1 DeviceSuimnary 29

5.1.2 DeviceEvaluation 29

5.1.3 DeviceDesign and Simulation 32

5.1.4 Conclusion 32

5.2 LBL TechnicalProgress 37

5.2.1 A RobustHigh-TemperatureDetectorFabricationProcess 37

5.2.2 Monolithic Integrationof Circuitry and Detectors . . . . 37

5.2.3 Low-power Low-noise AnalogCircuitry Optimization 38

5.2.4 RadiationDamageStudiesof Detectorsand Circuitry . 38

5.2.5 PIXSTRIP, a designstudy for a smartpixel array . . . 38

6 Statementof Work and Budget Summary 40

6.1 HughesAircraft Company 40

6.2 LawrenceBerkeleyLaboratory 40

4

6.3 UC Berkeley - SpaceSciencesLaboratory 41

6.4 Stanford Linear AcceleratorCenter 42

6.5 University of Oklahoma,Yale University, and University ofIowa 42

6.6 UC Davis and Iowa StateUniversity 43

6.7 University of Pennsylvania,PrincetonUniversity 43

7 Budgetsand Schedules 45

8 Appendices 50

8.1 HUGHES AIRCRAFT COMPANY CAPABILITIES 50

8.1.1 MOSAIC ARRAY DEVELOPMENT 55

8.1.2 ORGANIZATIONAL RESOURCES 56

8.1.3 FACILITIES 56

8.1.4 KEY PERSONNELRESUMES 63

8.2 A READOUT DESIGN FOR HIGH FLUX ENVIRONMENTS 66

8.3 SILICON PIN ARRAY PAPER 67

8.4 A SET OF POSSIBLEREADOUT DESIGNS 68

5

1 Executive Summary

* ProposalObjectives:

This proposalcoversthe12 monthfirst phaseof a 36 monthprogramto demonstratekey technicalperformancerequirementsfor silicon pixel vertexdetectorsfor the SSC.The effort addressesdiffering needsfor high P1 andintermediateP experimentssuchas thosestudiedby two different groups,the SolenoidalDetector Experiment SDE, and the Bottom Collider Detector BCD.

During the first year, systemdesignconceptswill be developed,a preliminarysystemcost analysiswill be performed,andprototypehybrid detectorswill bedesignedandfabricated. Beamtestson the pixel detectorswill be performed,andspecialtestequipmentwill be designedandbuilt to operatethe prototypearrays. Theprincipalobjectiveof this first phaseis to demonstratethetechnicalandcost feasibility of pixel detectorsystemsfor the BCD and the SDE. Afterthefull 36 month term, final designverificationandkey hardwareelementswillbe demonstrated;beamtestsof detectorarrayswill demonstratereadinessforfull scaleengineeringdevelopment.

* TeamDescription:The SSC placessevereperformancerequirementsupon detectortechnology,requiring new and imaginative solutions in many areassuch as tracking andvertexdetection.A teamapproachcouplingthehigh energyphysicscommunityand industry is essentialto establishrealistic technical requirements,developan appropriatesystemsarchitecture,provide rapid evaluationand test beamresults,andoptimize the overall systemperformancefor an acceptablecost.

The industrial partner, HughesAircraft Company, is a recognizedleader inboth pixel array designand sophisticatedelectronicsystemfabrication. Pixeldetectorarraysin a 10x64 format andin a 256x256format havealreadybeenbuilt and initial test datais presentedin this proposal. The physics researchmembersof the teamcombinetechnicalexpertisein semiconductordevicesandextensiveandvaried experiencein high energy physicsresearch.BCD and SDE,the major SSC initiatives reflecting this experience,are well-representedin thiscollaboration.

6

* Key Elementsof the ProposedProgram:

- Early appraisalof systemfeasibility and cost issuesreducesrisk;

- Early involvementof industry will increaserateof progresstoward a finalsystemand introducesophisticatedmanagementpractices;

- First yeareffort addresseswide rangeof physicsrequirements;

- Coordinatedeffort by all key membersof physics communityeliminatesduplicationof effort and ensuresgood communications.

- Prototypepixel arrayswith sparsereadoutcircuitry will be designedandbuilt for testingand evaluation.

* Total PhaseOneCost for ProposedProgram: $1,190,000

7

2 Introduction

For theSSC,silicon pixel detectorsoffer theattractiveprospectof highlyunambiguouspatternrecognition,high spatial resolution,high radiation tolerance,and fast timeresponse,which are essentialqualitiesfor vertex detectionand tracking closeto thebeampipe. The extremelychallengingphysicsand backgroundenvironmentof theSSC would appearto make all other known detectorconceptsinferior to a pixeldetectorfor this purpose. Much of the basic semiconductortechnologyneededtorealizepixel detectorssuitablefor the SSCalreadyexists, and significant advancesin pixel detectordesignshave beenachievedduring the currentSSC GenericR&Dprogram. Nevertheless,to meetprojectedSSCschedulesfor detectorconstruction,&substantialeffort is needednow to realizetheprogressneededto achievetherequiredtechnical performanceand to establish budgetaryrequirementsfor thesedetectorsystems.

Within the broadestcontextof SSC physics,thereare perhapsthreeor four distinct categoriesfor pixel detectorapplications.High P physics, with a wide varietyof physics signatures,requiresa complextwo-level trigger structure,that hassubstantialimpacton thepixel arrayarchitecture.IntermediateP physics,on theotherhand, placesemphasison very rapid readoutof detectorinformation after a singlerelatively short trigger decisioninterval. A third categoryincludesmorespecializedphysicsgoals such as exotic particlesearcheswherehigh quality ionization densityinformationas well as spatialresolution is required. Finally, it hasbeenrecognizedthat the pixel arra3?swe intend to developarevery well matchedto a possiblesolution in conjunctionwith an imageintensifier for the ratherdauntingreadoutsystemrequirementsfor a scintillatingfiber detectorat the SSC. While thepixel arraycharacteristicsneededfor all four categoriesmay overlapsubstantially,only the first twocategoriesareaddresseddirectly in this proposal.

To facilitate rapid progressin the developmentof silicon pixel vertexdetectors,asizablecollaborationincluding anexperiencedindustrialpartnerhasbeenformed. Inlargepart, the proponentsareassociatedwith the BCD and SDE groups,which haveundertakensubstantialefforts to evaluatethe detectorrequirementsfor the intermediateand high P1 SSC physicscontextsrespectively.Throughtheseassociations,the technicalgoalsof this collaborationwill maintain the closestpossiblerelationship to therequirementsassociatedwith thehigh andintermediateP1 physicsof the

8

SSC.The collaborationis enteringinto a teamingagreementexpressingourcollectiveintentionto work togetherthrough the constructionphaseof SSC detectorsystems.

Thecurrentproposalpresentsin detailour planfor thefirst yearof what is foreseenas a threeyearprogramto reacha rathercompleteengineeringdemonstration.Thedevelopmentprogramfor the following two years,including an approximateestimateof budgetaryrequirements,is given in broadoutline.

9

3 Relevance of Pixel Vertex Technology to thessc

3.1 Overview

TheSSCwill providea window into anentirelynewandunchartedphysicsdomain. Inorder to bestexploit the discoverypotentialof this instrument,interactionsmust bestudiedin sufficient detail to separaterareor unexpectedprocessesfrom theferociousbackgroundscharacteristicof theSSC.New physicsprocessesareexpectedto generateheavyquarksand leptonsasdecayproductsmuchmorefrequentlythanknown QCDprocesses.Heavy quarkswill be copiouslyproduced,offering an opportunity for thestudy of the mechanismof CP violation and raredecays.On a moremundanelevel,a consequenceof the high design luminosity is a high probability of more than oneinteractionperbeamcrossing. It is clearthat theability to measurevertexstructureis an essentialalthoughextremelydifficult technicalchallenge.

There exists to date no silicon vertex detectorin an e4C machineor hadroncollider, though efforts are underwayto install suchdevicesin severalexperiments,like CDF at Fermilab, MarkIl at SLAC, and Delphi and Aleph at CERN. Thesefirst generationdetectorsall use silicon strip technologyand newly developedVLSIcircuits mountedon thedetectorfor amplificationand readoutprocessing.TheSLDvertexdetectorat SLAC will be a CCD devicewith very a slow readout.

3.2 Intermediate P2 Physicsand Vertex Detection

The BCD experimentis a programof physicsthat will study intermediateand lowP1 physics at the SSC.It is hoped that the BCD can begin taking dataat Fermilabin the mid 1990’s with some of the detectorsystems. The completeexperimentwill be a first round detectorat the SSC. The physics goal of this experimentisthe completeand thoroughstudy of the CP violating decaymodesof the B meson.Only B mesonsthat decayto all chargedstatesareconsidered. CP violation hasgreat importancewithin the standardmodel and is a subjectof much interest for

10

theoristsandexperimentalists.Thefollowing is a list of someof the interestingphysicsquestionsthat canbe addressedwith a detectorat the SSCthat hasa powerful vertexdetector.

1. CosmologicalmodelsinvokeCP to helpexplainthe matter-antimatterasymrnetry in the universe.Thus theexistenceof the universeis thoughtto be relatedto CPin someway.

2. Multiple Higgsbosonscan leadto relativecomplexphasesthat in turn haveCPviolating effects. Thus theunderstandingof massgenerationand CP is related.

3. CP violation is important to the generationpuzzle. With two families thereis no CP violation in the StandardModel. All existing datais consistentwiththreefamilies or one complex phase. If therearefour families, thentherearethreecomplexphasesand new CP phenomenaare expected.

4. Measurementsof CP violation in the B systemcan overconstrainthe C-K-Mmatrix.

5. Left-right symmetricmodelspredictsmall CP violating effectsin the B system.

6. Thereis somehope that if the C-K-M elementsare well determinedit will bepossibleto observesymmetriesin themassmatricesof thequarksandthereforeYukawacouplings.

7. The StandardModel is a parametri±ationof experimentaldata. It providesnoinsight into the origin of CP violation. Furtherstudy of CP violation in the Ksystemand information from a new systemmay providevaluableinformationfar beyondthe standardmodel.

8. Thesearchfor Higgs particlesthat decayinto M is madepossiblewith a vertexdetector. Technicolormesonscan decayto bb aswell.

9. In summary,manyfundamentallyimportantphysicstopicsarewithin reachofstudy at theSSCif a vertexdetectorof sufficientperformancecanhelpextractthe B signal.

11

3.2.1 BCD Vertex Detector

Introduction The emergingopportunitiesto study B physics at the SSCareprimarily a resultof theadvancementof silicon vertexdetectortechnology. It hasbeenunderstoodfor sometime that the SSCcouldproduceof order 1012 bb pairsper year.The experimentalchallengeis to fully reconstructthe B mesoninvariantmassfromits decayproducts.Thereare2-6 chargedtracksfrom the B mesonand typically over100 chargedtracks in the event. Fortunatelythe B lifetime is sufficiently long thatthe decayvertex can be separatedfrom the beamcollision point using a precisionvertexdetector.Theinvariantmassis thencalculatedusingonly tracks comingfroma secondaryvertex. Techniqueslike this havegivenrejectionfactors of up to 106 insomedecaychannelsin low ratefixed targetexperiments.

The most relevantwork for BCD is that of the CDF group. However, in orderto reconstructfully the B mesonmasswith high efficiency over a largesolid angle,the BCD vertex detectordesignis significantly morecomplex. The requirementsaredescribedin the following sections.

Vertex Detector Requirements The BCD must build the bestvertex detectorpossibleif it is to achievesuccessat the SSC. If the vertexdetectordoesnot workwell, theexperimentwill be severelycompromised.The BCD vertexdetectorshoulddeterminethe impactparameterof chargedtrackswith respectto thedecayvertextoless than 10pm rms error. The multiple scatteringerroron the track is proportionalto R/Pt.JL/Lo, whereIt is theradius to thefirst detectorplane,P1 is the transversemomentumand L/L0 is thepercentradiationlength. To meetthe requiredresolutiongoals, the vertex detectorinner layer must be placedat It-’ 1.25 cm. This shortdistanceis necessarybecausethe averagePr of tracks from B decayat the SSC isless than 1 GeV/c. Furthermore,there is much incentiveto minimize the amountofmaterialand the Z of the material. Beryllium is thereforea common choicefor thesupport material.

The radiationdamagetolerancefor gammarays must be greater thanor equalto l0Mrads. If luminositiesabovelO32cnr2sec1arerequiredto study smallerthanexpectedCP asymmetries,a higher radiationresistancewill be important. Theneutron fiwi in BCD may be less than detectorswith 4r steradianhadroncalorimetry

12

coverage,but BCD has substantialiron in the magnet yokes and muon detectionsystem.A betterestimateof the neutronflux in BCD is needed.

An important design factor will be the hit occupancy. This may ultimately determine the pixel readoutsystemand the highest luminosity possiblein the vertexdetector.A sophisticatedsimulationof the silicon vertexdetectoris underwayby theBCD groupand will contributeto designconsiderationsin thenext coupleof months.The patternrecognitionstrengthsof a small pixel devicemust be tradedagainstthedifficulty in manufacturingthe electronicsfor eachpixel asthe sizedecreasesbelowbOx100 pm2.

The following is a list of mechanicalissuesthat the BCD vertexdetectorgroupmust addressin the coming monthswith simulation and tests. This overall systemmechanicaldesignspecificationis underway. We list issuesand raisequestionsthatrequireexaminationduring the performanceperiodof this proposal.

1. Thevertexdetectoris envisionedasbeingconstructedwith self-supportingmodules. Thesilicon detectorswould be gluedinto a polygon matrix. The moduleswould thenbe mountedto an externallysupportedstructurecalled a "gutter".It is the gutterphilosophyand alternativesthat requireexaminationas part ofthe work of this proposal. Ref. FermilabTM-1616 Jostleinet al.

2. The alignmentaccuracycanbe consideredin two parts. Theinitial alignmentof detectorswith respectto eachotherand the support structurethe gutterduring and after assemblymust be specified. The long and short term driftsdueto heatingfrom theamplifiersmustbesimulatedin orderto understandthelimits of an acceptabletolerancespecification.It is estimatedthe x-y accuracymust be 4pm , while the accuracyin Z may be 200pmover the lengthofthegutter.

3. The assemblyprocedureneedsto be studied.

4. Thecooling requirementsmay feedbackinto theelectronicsdesign. An efficientcooling system,capableof extracting2-5 kW of powerwill permit a lower noiseamplifier design. The cooling will begasand liquid.

5. The issue of operating below room temperature,e.g. zero degrees,must bestudied. This will reduceleakagecurrentsin the detector.

13

6. Material from connectorsand cablesare a major sourceof scattering. Fiberopticsshouldbestudiedasan alternativesignaltransfertechniqueevenif powerdissipationis higher.

The electronicsissuesfor BCD are similar but perhapsmore severethan thosefor SDE, primarily becauseof two reasons.First, the requiredresolutionfor BCDmust be 10pm. This suggestsa pixel size of 30 x 30pm2. The subsequentareaavailable for electronicsis roughly an order of magnitudeless for BCD thanSDE.Secondly,the noise level requiredis driven by the dipped track problem,ie. smalldepositionof chargefor high angleof incidencetracks.A high efficiencyfor detectingtracks that deposit 6000 electronsis necessary.Even so, this placesa 45° trackcutoff angleof incidence.Theremainingissuesand questionsarelisted.

1. A further considerationdriving the low noise goals of the electronics, 51electronsrins noiseperpixel, is the very largenumberof noisehits in a systemof 10" pixels.

2. The radiationdamageeffectsaresimilar for BCD and SDE, but if BCD runsat higher luminosity, this changes.The higher luminosity running will be determinedby the magnitudeof the CP asymmetries.Also, the proximity of thesilicon to thebeamduring injection may give largedamage,asexperiencedwithIJA2.

3. The BCD wishesto read out the vertexdetectorevery few microseconds,incontrastto SDE. The readoutdesignsmay be different.

4. Theresettingphilosophyof thepixels for the two experimentsmaybedifferent.If the resetfrequencyis too small, pixel occupancybecomesexcessive.Thismay be moreseverefor SDE if the level 2 wait time is S0psec.

5. The bunchcrossingtaggingproceduremustbe defined.

6. The advantagesof driving analogversusdigital signalsneedsinvestigation.

7. Centroid finding techniquesareof great interest to BCD becausethe dippedtracksgive up to 60 hit pixels.

14

3.3 High P2 Region and Vertex Detection

For a completetracking detector, a high resolution, radiationhard detectorarray is needednear the SSC beam to tag b’s and r’s and to allow lifetimemeasurementsof newlong-livedparticles[1}. Suchdetectorsareneededfor trackreconstructionand accuratedE/dxmeasurementsto searchfor free quarksinthe cores of jets[2]. In addition, the detectorprovidesaccurateseparationoftrackscomingfrom different primaryinteractionverticesin a singlebeambunchcrossing.

The following are some representativephysics goals for the vertex detectorinthecentralrapidity region:

a Taggingof secondaryverticesin the searchfor an intermediate-massHiggsbosondecayingvia the bb or ri’ channels.Thedetectoris alsovaluableforreducingbackgroundsby anti-taggingin otherHiggs massregimes.

b Tagging isolatedleptonsfrom W decaysto reducebackgroundfrom QCDjets andphotonconversionsto e+c pairs. This is needed,for example,tostudy WW scatteringor heavy Higgs decayvia the channelW-+ lv +W-. jets.

c Improvementin the signal to noiseratio in high-masstop quark searches.Possibledecaychainsof interest are, e.g., ii -p bevbqqor ii -* evpvbb.The b’s may be fairly soft, so thepixel trackercanhelp here.

d Accuratemeasurementsof the physics processesinvolving the copiouslyproduced& quarks.

3.3.1 SDE Vertex Detector

Introduction In large part the general characteristicsof the BCD vertexdetectoralso apply to the SDE vertexdetector.Ratherthan repeatthose,weelaborateon the specific characteristicsneededfor the SDE.

A major difference is that the SDE vertexdetectormust coexistwith and beintegratedwith a much larger and very complex tracker, which we presumehereto be a silicon-basedsystem. As such,it may be expectedthat systemsengineeringconsiderationswill play the dominant role in the ultimatephysicalconfigurationand generalelectronicarchitecture.

15

A useful definition of a pixel detectoris that the subdivisionof detectionelementshasbeenincreasedto the maximumpracticalextentconsistentwith theintrinsic spatial resolution. A consequenceof this extremesubdivision is thatthe hit rate of individual pixel cells is typically a few hundredHz. This is trueevenfor rather largepixels suchas100 x 100 pm’ closeto the beamline at fulldesignluminosity. In turn, this low hit rate allows the pixel to be usedas itsown buffer for eitheranalogor digital information. This powerful but paradoxical fact is thecentralconceptfor thesystemsarchitectureof the pixel detectorsconsideredin this proposal. At the same time, this featureplacesstringentlimitations on the length of time datacanbe maintainedwithout compromisefrom pileup effects; it also introducescomplexity in the relationshipsamongdeadtime,pixel reset,and circuit design.

A "smartpixel" concepthasbeendevelopedby D. ft. Nygrenand H. Spielertomatch the SDE requirements;a further elaborationof this conceptand otherissuescan be found in appendix8.2, which is a reprintof a reportgiven at theSnowmass1988 meeting.

Vertex Detector Requirements The SDE vertexdetectoris envisagedtoprovide tracking within a rangeof ± 2.5 units of rapidity, matchingthe general rangeof tracking for the detectorsystemsat larger radii. Physically, thistranslatesinto an extentalongthebeainlineof about± 50cmfor a 10 cm outerradiusfor the pixel vertexdetector.The inner radius may well be constrainedby backgroundand ratelimitations to a radius not less than4 cm.

Theminimumnumberof layersneededis generallyregardedasthree;this number in principleprovidessomeconstraintsfor elementarytrack finding. As morerealisticsimulationsareperformed,addingbackgrounds,delta rays,deadtime,andotherreal-lifeeffects,it canbeexpectedthat theminimumnumberof layersincreasesto four. Detailed and carefully executedsimulationsarean essentialpart of the processleadingto a stablesystemarchitecture.

The rather long luminous region s 7 cm along the beam direction hasanimpacton pixel geometry.Particularlyin the very centralregion,a largerangeof track anglesthrougha given pixel will exist,on theorder of ±45°. Givenastandarddetectorthicknessof 250-300pm, this rangetranslatesinto a similar"natural length" for depositedenergyin the beamdirection. Ionizationfluctuationswill be substantialalongthe projectedtrack length. Theconsequenceisthat a pixel cell substantiallysmallerthan200 pm will receiverelatively small

16

fluctuating signals that cannotbe used to improve the resolution. From thisperspective,a pixel cell hasa similar naturallength that canbe usedfor neededcircuit functionality.

On the otherhand, resolutionsof 4-5 pm are desiredin the spatialdimensionstransverseto the beam. Experiencewith strip detectorsystemshas indicatedthat this level is achievablewith widths on the order of 50 pm. For detectorsizes of this scale,diffusion and ionization processescontributeto an intrinsicspreadingof chargethat can be usedto improveresolution using analoginformation. Therefore,a rectangularpixel on the order of 40 x 200 pm is anattractivepossibility.

Theorientationof thepixel arraycanbe chosento benormalto thecenteroftheluminous region to minimize track obliquity. This leadsto anotherbeneficialresult for the distribution of the pixel arraysalong the beamdirection. Forprogressivelylargervaluesof rapidity, the pixel arraysare increasinglyangled,leadingto a kind of "Fresnelcylinder" configurationthat permitsa considerablegap betweenone ring of arraysand its neighborswithout loss of efficiency forall but the very lowest momenta. This significantly reducesthe numberofarraysneededto providecoverageas comparedwith a configurationin whichthe arraysarenot angled.Unfortunately,anadequatedrawingillustrating thisFresnelconceptis not yet available.

Calculatedpowerdissipationfor eachlayerof this pixel array is on the orderof100W. Removalof this heat with a reliable low-masstechniqueis a very serious mechanicalengineeringchallenge.At the sametime, the low-masssupportstructuremust incorporatecompletepower distribution, trigger, timing, databus, etc. connectivity. Probablythe successfulrealizationof neededfunctionality within the severephysicalconstraintswill be the most challengingaspectof a completeSSCvertexdetector.

Thetriggerfor the SDEwill be a two-levelhierarchyasfar asdetectorsareconcerned.Level onemay occuron averageevery 100 crossingswith a synchronousaccept/rejectoccurringabout oneor two microsecondsafter the crossing.Thegreaterthe delayin forming this trigger, the more silicon must be devotedtostoring pattern information. It is imperativethat this time interval be keptto a realisticand stableminimum, and madeimmuneto gradualincreasesdueto "improvements". Level two is expectedto provideanotherfactor of about100 in rejection,but thelatencyis variable,perhapsevengreaterthan30 psec.Givencurrentuncertaintiesin knowledgeof backgroundrates,thereis a danger

17

that pixel information could be contaminatedfor very long level two triggerlatency. Simulationof this will be an importantpart of the informationneededto choosesensiblesystemarchitecturesand pixel cell designs.

If pixelscanstore information,they must alsohavethemeansto clearinformation at appropriateintervals. A system-wideresetcanbe easilyenvisaged,butthis must occur more frequently thanthe level two accept,as the arrayswilllikely becometoo heavily "occupied". A system-wideresetoccurringafterleveloneacceptskeepingthe neededinformationwill introduceexcessivedeadtime.It appearspossibleto selectivelyresetonly thosepixels falling out of theaccepttime window. Clearly, the resetquestionwill be an importantissue.

The issueof radiationdamagehastwo main branches,one for thedetectorsandthe other for the readoutIC’s, which havequite different aspects.

After SSP operation is underway, the leakagecurrent generatedby the detectors will becomehundredsof times greater than at first. Oneconcernisstability againstbreakdown,which will benefit from a robustprocesssuchasthat developedat LBL. Anotherconcernis that thedarkcurrentwill contributeintegratedchargecomparableto or greaterthanthat of chargedparticletracks.This indicatesthe needfor detectorswith AC coupling to the preamps.LBLis developingthe meansto provide robusthigh resistancepolysilicon resistorsto bleed off the pixel leakagecurrent to groundand the capacitorsneededtoisolatepreampvoltageoffset. At the extremelyhigh fluencescorrespondingtoyearsof SSC operation,there is little informationabout chargecollectionandtrappingeffects. LBL, in collaborationwith others, is pursuing a systematicradiation damageevaluationprogram. -

References

[1] Reportof Large SolenoidDetectorGroup in Proceedingsof the Workshopon Experiments,Detectors,and ExperimentalAreasfor the Supercollider,p. 340 1987.

[2] Exotics Group SummaryReport in Proceedingsof the Workshopon Experiments,Detectors,andExperimentalAreasfor the Supercollider,p. 8531987.

18

4 Technical Discussion

4.1 SystemDesign Process

To determinethe feasibility of a pixel vertexdetectorfor the SSC severalbasicquestionsmust be answered:

a What are the radiation backgroundcharacteristicsin which the vertexdetectormust operateand how do thesebackgroundfields determineradiation hardnessrequirements?

b What arethe physicalconstraintse.g. multiple coulomb scatteringandphotonconversionlimits,

...that a vertexdetectormust match?

c What arethe spatialresolutionand track-pairrequirements?

d What arethe timing resolutionrequirementsfor tracktagging?

e What are the trigger requirementsand how do they affect systemarchitecture?

f What are the datareadoutrequirementsand how do they affect systemarchitecture?

The answersto thesequestionsareat leastpartially known; given this much,thenthe questionsresolveto technicalissues:

a Are high speed,low power, radiationhardeneddetectorand readoutassembliespracticalwith currenttechnology?

b Can an efficient data collection and processingsystembe devisedwhichcanaccommodatetheseassemblies?

c Can a stablelow-massmechanicaland cooling systembebuilt?

Theproposedwork addressesthesequestionsfrom both a technicaland a costviewpoint.

Theproposedprogramwill consistof threemain activities: the design,simulation, andfabricationof a prototypedetectorand readoutarray,the simulationanddevelopmentof a signalanddataprocessingarchitecture,anda conceptualdesign for the entire system. The efforts for the pixel array developmentarediscussedin detail elsewhere,but theprototypedesignand simulationwill influence and be impactedby the work donein the otherparts of the proposed

19

effort. For example,the functionsand featureswhich will be incorporatedintothe initial prototypewill be thosewhich arenecessaryto validatetheprototypeas a proof of concept. Thesefeatureswill be determinedby a review of theexisting conceptsby the team. This task is complicatedby the diversity ofphysical processesinvestigatedby the different collaborators. Theseconceptswill be examinedto distill the essentialtechnicalfeatureswhich arecommonto the various conceptsand which will enablea decisionto be madeasto thefeasibility of the pixel arrayvertexdetectors.

With an initial pixel array architecturedetermined,the questionof systemfeasibility can be addressed.In order to perform this evaluationa candidatesystemdesign must be used. Representativequantitieswill be usedto arriveat one or moresystemconfigurations,from which a designmay proceed.Thefirst step in developinga systemdesign is determininga datacollection plan.This plan will detail how the datacollectedby the pixel arrays will be usedto accomplishthe track reconstruction. This plan will be the roadmspforthe signal and data processingdesign. The data collection plan will addressthe various trigger configurationsso that such quantitiesas datarates andbuffer sizes may be determined. In addition, the datacollection plan mustaddressthe track reconstructionalgorithms. The reconstructionalgorithmswill influence such systemlevel componentsas pixel size, numberof pixels,signal to noise ratio, and digitization accuracy. By developingsucha datacollectionplan, systemarchitecturesmaybe developedwhich takeadvantageofsignalpreprocessing,therebyreducingsubsequentdataacquisitionand storagerequirements.After thedatacollectionplan is defineda simulationmodel willbedevelopedwhich will allow designtradeoffstudiesto be conducted.Thegoalof suchanalysiswill be to determinetheminimal systemconfigurationnecessaryto achievethe physicsgoals.

Finally, a completesystemdesignwill be performedto assesscost and lifetimeconsiderations.The systemdesignwill cover mechanicalpackaging, detectorlayout, cabling, cooling, shielding, and other mattersthat needto be studiedto identify technologyand cost drivers.

The first phase12 monthprogramproposedhereinis shown in Figure 2. Provision is made for regular technical interchangemeetingsat 2 and 3 monthintervalswith all principalmembersof the teaminvolved in pixel detectordevelopment.Theproposedreadoutdesignwill be thoroughlyreviewedat a conceptdesign review, preliminarydesignreview and final designreviewbefore initiat

20

ing wafer fabrication. This review processwill assurethat the critical readoutrequirementsnecessaryto demonstratetechnicalfeasibility for both the BCDand SDE experimentsarebeingaddressed.

A numberof readoutdesignshavebeenproposedand thesewill be thoroughlyanalyzedso that feasibility can be assessed:

4.2 Design,Fabrication, and Test of a PrototypePixelDetector

The objectiveof the fabricationtask is to fabricateprototypedetectorarrayswhoseperformancecanbemeasuredand comparedagainstSSCgoals. Experiencegainedin the prototypedevelopmentwill thenbe utilized in the development of thefull scalesystem.

Following the developmentof the conceptualdesigndescribedin the previoussection,a detaileddesignmust be produced.The detaileddesigndevelopmentcomprisestwo broadareasof activity, technologyselectionand circuit development. Thereareseveraldevicetechnologycandidateswhich couldbeusedin thedevelopmentof a prototypearray. Silicon MOS technologyoffersvery low powerand reasonablyhigh speed.It is inherentlyhard againstneutronradiation,andcan be hardenedagainsttotal doseradiation. Bipolar technologyoffers speedadvantagesover MOS, but at the expenseof somewhathigherpower. Bipolardevicesareextremelyhardto total dosebut suffersomedegradationin neutronenvironments.Unipolar devices,suchas JFETSand MESFETS,offer furtherspeedenhancementswhenfabricatedin high mobility materialsuchas galliumarsenide.Their radiationhardnessis similar to that of bipolartransistors.Finally, it may be desirableto employ a BICMOS process,which is a mergerofMOS and bipolar technologieson the samechip.

This allows the useof low powerMOS circuitry in powercritical areas,whilehavingbipolar transistorsavailablefor portionsof thecircuit which requirehighdrive capability.

The following requirementswere agreedupon for the first prototype hybridpixel readout. Theserequirementswere consideredgeneric to both the SDEand BCD experiments.They areappropriatefor a first yeardemonstrationtodemonstrationto demonstratethe technicalfeasibility of pixel detectorarrays;

21

a Powerfrom 100 mW per squarecentimeterto 1000 mW per squarecentimeter.

b Readouttime of approximately1 ps per hit

c Tag time about 10 ns

d Radiationhardnessin excessof 1/2 megarad

e Noise level of 200 electronsrms

Thesespecificationsarebelievedto be a reasonableand conservativeset of firstyear goals. Substantiallybetter performanceis expectedover a longer timeperiodasprogressis achieved.

The completedprototypearrayswill be subjectedto a varietyof testsin orderto verify their performance.Initially, the readoutICs, are testedat the waferlevel in order to determinewhetherthe designis functioning correctlyas wellas to identify candidateparts for dicing, packagingand performancetesting.Following wafer level testing, candidateparts will be hybridizedto detectorsand packagedfor performancetesting. The deviceswill be operatedat speedusing a pulsed laser as an excitation source. Such a setupwill enablea fullrangeof diagnosticand performancetesting to be performed.

Following the diagnostictestingand performanceverification, partswill bedelivered to the appropriateuniversityandlaboratoryresearchersfor furthertesting andinitial tracking experiments.

4.3 Design and Fabrication of the Control and DataAcquisition Electronics

Electronicsto powerand clock the existing 10x64 arrayswill be built for longdurationradiationtests. Thesetestswill beperformedat UC Davis and oneormoreof thefollowing institutions: SLAC, Los Alamos, Universityof Tennesseeat Knoxville, and HughesAircraft as describedin the Pixel Detector Testssection. It is too costly to assignthe complex dataacquisitionsystemto thistask asit must be availablefor beamtests.

Electronicswill also be designedand built to operatethe new detectorarraysthat will be built aspart of the work describedin this proposal. The sparsereadoutschemeof the new detectorswill be quite different from the randomaccessreadoutschemesusedin the existing 10x64 and 256x256arrays. Work

22

on designingthiscontrol anddataacquisitionelectronicsmustproceedsimultaneouslywith the pixel array readoutelectronics.The electronicswill be basedasmuchaspossibleon theelectronicsdesignthat is usedto operatetheexistingdetectors. The existing system matchesa set of computersand processorstothe severaltasks. Calculationspeedand real time capability arethe most importantfeaturesof the processors.Custombuilt circuits using statemachines,shift registers,andotherdigital componentswill be built if speedrequirementsmakethem necessary.The software for the systemwill be written in parallelwith the hardwaredesignand fabricationwork to morefully ensurethesystemwill work and to completethe systemrapidly.

4.4 Pixel Detector Tests

The testingprogramfor pixel detectorswill involve

* laboratory tests,

* radiationtests,and

* beamtests.

We proposeto combinerelevantongoing individual genericR&D programsinorder to bring a coherentand efficient effort to bear on the problems. Beamtestsusing pixel detectorsand silicon microstripswill allow us to evaluatetheperformanceof the pixel detectorsthemselvesas well as that of a combinedsystemof pixels and microstripsin an acceleratorenvironment.

4.4.1 Laboratory Tests

Testingin the laboratoryis aimedprimarily at characterizingthe performanceof thedevicesandoptimizing themfor high energyphysicsapplications.Typicalissueshere are:

* signal to noiseratio

* stability of performancecharacteristics

* spatialand time resolution

* effectivenessof readoutarchitecture.

23

Continuedtestingof theexisting devicesfrom Hughes,a radiationhard 10 x 64array with 120 pm squarepixels, and a 256 x 256 array with 30 pm squarepixels, is an essentialpreliminarystepin this effort. This will be funded in partby the existinggenericR&D program.

The new pixel device must be testedto insure that the presenceof data issignaledsufficiently rapidly within iOns and that datais capableof beingread out at speedsconsistentwith SSC requirements.A new dataacquisitionand control systemis requiredfor thesetests in order to run the device atdesignspeed. Funding for the developmentof this systemincluding both thehardwareand the extensivesoftwareis containedin the SSL budget.

4.4.2 Radiation DamageStudies

Radiationtestingof the devicesis necessaryto insurethat theywill be capableof enduringthehigh neutronandchargedparticlefluenëesexpectedin theSSCenvironment.The chargedparticledoserateat a radiusof 10 cm from an SSChigh luminosity intersectionpoint is estimatedto be 3x iO Gy/year.Thefluxof neutronsback-scatteredfrom the surroundingcalorimeteris estimatedto becomparablein magnitudeandmoreseriousin termsof radiationdamage.Hencethe performanceof the the detectorsand thehybrid readoutcircuitry must bereliably characterizedwith respectto both displacementdamageandlong termionization effects. Thesetests would proceedmost rapidly and efficiently ifthey utilized the devicetestbeamsat the 68 MeV UCD cyclotronand proposedradiationtestingfacilities at SLAC. Thelatterwould providea 6°Co sourceanda 252Cf neutronsourceplus personnelfor managementof dosimetry[1]. If thisdoes not materialize, then an alternativesuchas Los Alaxnos may be neededfor long-termneutronexposuresandfacilities at the Universityof TennesseeatKnoxville UTK or Hughesfor 6°Co sourcesor X-rays.

Irradiation by neutronsand by gammarays canbe viewedasthe two extremetests. Irradiation by fairly ionizing protonsis usefulin that it producesamix ofionization and displacementeffects,aswill be thecaseat the SSC.Furthermore,recentstudiesof displacementdamageas a function of energywith a varietyof incidentparticlesindicate that in manysituations,the resultingeffects canbe accuratelyscaledfrom one energyand particletype to anotherl2]. Henceaprogramof testingat the UCD cyclotron is proposed.This facility is regularlyemployedfor testsof radiationdamageto semiconductordevicesby a variety

24

of users. Proton energiesare availableup to 68 MeV with a maximum fluxof 2.5 x 10" protons/s. Typical doseratesin actual testsareof the order of20 Gy/s spreaduniformly over a 40 cm2 area. Well-calibratedinstrumentationexistsfor dosimetryand monitoring the arealuniformity of the beam.Readoutand control electronicscan be placed in the radiation cavebehind shieldingblocksnearthe equipmentundertest. Remotecontrol systemscanbe locatedwithin rangeof a 30 m cablerun. A further potentiallyuseful diagnostictool istheproton microprobe-a4.5 MeV beam5 pm in diameterwith a flux of up to1.3 x 1010 protons/s.Work is taking placeat presentto extendthe microprobeenergyto 40 MeV althoughwith an increaseddiameterof 30 pm. Sucha beamcouldbe usedto testspecificareason a chip for sensitivityto radiation. Variousneutronbeamsarealso available.

The existing radiationhard 10 x 64 arraywill be theprincipal subjectof thesetestsduring the first yearof this program.

The funding for thesetestsis containedin thebudgetrequestsof UCD, SLACand OU. It shouldbe mentionedhere that if long-durationradiationtestsareto be performed,dedicatedelectronicsto power and clock the devicesmust befabricated. It is too costly to assign the completedataacquisitionsystemtothis task,asit must be availablefor the beamtestsdescribedbelow.

Radiationtestingof thenewSSCdevicesresultingfrom thepresentproposalwillfollow a similar path. Additional powerandclocking circuits maybe necessary,however.Thesetestswill be scheduledfor thesecondyearof the proposal.

4.4.3 Beam Tests

Testing of the pixel arrays in particle beamsis requiredto demonstratethefunctionalityof the readoutsystem,to studyalgorithmsrelating to the ultimatespatialresolutionachievablewith thesedevices,to measurechargesharingandits effects, and to study the responseto particles which strike the device atnon-normalincidence.

Testsof Existing Arrays Threeseparatetestsof the existing pixel arraysareproposed:one at theUCD cyclotron,oneat theHyperonBeamat Fermilaband one at the FermilabMTest Beam.

25

The UCD cyclotron canbe usedfor initial chargedparticlebeamtestsof the10 x 64 array. Responseof thedetectorwould bemeasuredin aparallelbeamof68 MeV protons. This would also facilitate thestudy of detectorperformanceasa function of radiationdose. Thesetestscould be further refined if thelocation of eachtrackwere accuratelyknown asit entersthe array. OU expectstoassembledouble-sidedmicrostrip detectorswith SVX-D readoutswhich wouldbe well suited to sucha test,sincethey providecorrelatedx,y positionmeasurementswith a minimumof scatteringmaterialandin a compactgeometricalarrangement.A portabletest setupwould accessthe SVX CAMAC modulesandotherdiagnosticequipmentusinga Macintoshcomputer,aNationalInstrumentsOPIB interfaceand a GPIB/CAMAC cratecontroller. This apparatuswith the requiredsoftwarewould be providedjointly by OU andUCD. A Sunworkstationwill be requiredat UCD to control the dataacquisitionsystemforthe 10 x 64 pixel array itself. OU is requestingfunds in the presentproposalfor travelmoneyto conductthetests,for mechanicalsupportsand interfacePCcardsfor the microstrip detectorsand for supportof a graduatestudentduringthenext fiscal year.

Testsof the non- radiationhard 256 x 256 arrays as part of more sophisticatedtrackingsystemsin beamsof minimum ionizing particleswill be doneatFermilab.

Thefirst of two suchstudieswill takeplacein the E781 testbeamduringMayandJuneof 1990. Theapparatuswill occupya 350 GeV/cE beamofintensity

i0 particles/s12% byperons.Two pixel arrayswill be used, followed byanumberof silicon strip detectors.Though this is primarily a testrun, we arehopefulof recordinganumberof charmedcascadesandlambdas,andproducingapublishablepieceof physics. Thefunding for this testwill comeprimarily fromgenericR&D fundswhich havecometo SSL and SLAC as well as funds fromE781 sources.

Immediatelyfollowing the E781 test, we proposeto movethe 256 x 256 pixelarraysto the Meson Lab at Fermilabwherethe BCD collaborationwill beperforming testsof microstrip detectorsin the MTest beamduring the 1990 fixedtargetrun. Here,onecanobtain a5 cm squarebeamspotof 1-3 x 1O particles perspill at energiesbetween75 and250 3eV. The OU microstrip systemwill be readout with CAMAC-interfacedSRS and SDA modulescontrolledbya VAX3100 computerworkstation. Detectorarrayswill be usedto measureposition resolutionsand noisefor a variety of detectortypes. We plan to take

26

advantageof this setupto further evaluatethe performanceof the 256 x 256array and to measurethe improvementin vertex resolutionbrought about bythe useof pixel detectors. Funding for the microstrip testshasalreadybeenobtainedand the computerdataacquisitionsystemhasbeeninstalled. In thisproposal,we requestonly thoseadditionalfundsrequiredfor testingof thepixeldetectorsthemselves.

SystemTestof New Array As indicatedabove,thenewarray,whichwill bedesignedandbuilt aspart of thework proposedhere,will be subjectedinitiallyto a seriesof laboratory and radiation tests similar to those of the originalradiation-hard10 x 64 array. Following this, OU proposesto useit in a vertexdetectorsystemcombiningmicrostripandpixel devices.The goal of thesetestsis to measurethe position resolutionfor reconstructingverticeswith a detectorsystemsimilar to that requiredfor an actualSSC experiment.This will allow adirect evaluationof the capabilitiesof sucha combinedpixel/microstripsystemin an acceleratorenvironment.

Fundsfor themicrostrip testswill be requestedin a separatesubsystemsproposal. Timewill be requestedin theCO interactionregionat Fermilabduringthe1992-93 colliding beamrun. OU requestsadditional funds in this proposalfora Sun-4 computersystemand associatedsoftwareto be stationedat Fermilab,for mechanicalsupportsfor the pixel detectorsand for incrementaloperatingandtravel expensesinvolving the pixel testsat Fermilab.

References

Li] D.E. Groom, private communication.[2] G.P.Summers,ci a4 Correlationof Particle-InducedDisplacementDamage

in Silicon, IEEE Trans. Nuc. Sci. NS-34, 1134 1987.

27

5 Technical ProgressSummary

5.1 UCB/SLAC/HIJGHES Generic Program

Developmentof hybrid vertexdetectorshasbeenthegoal of theteamcomprisedof Eric Arens and GarrettJerniganof the SpaceSciencesLaboratoryat UCBSSL and StephenShapiroof SLAC for the pastthreeyears. In collaborationwith theHughesAircraft Company,we havedesignedandfabricatedtwo hybridpixel arrays.Thesensorportionof the hybrid arrayswerefabricatedby MicronSemiconductorLimited.

5.1.1 DeviceSummary

The propertiesof thesetwo arraysaredescribedin Table I.

TABLE 1Summaryof DeviceParameters

Array Dimension 10x64 256x256Pixel Size 120 pm 30 pmDetectorMaterial Silicon SiliconNumberof Readoutchannels 10 2PowerDuring "Write" Cycle 0 mW 0 mWPowerDuring ReadCycle 10 mW 2 mWClock Speed 1 MHz 1 MHzReadoutMode RandomAccess RandomAccessRadiationHardness 1 MR.ad ?Noiseat Room Temperature 300 e- rms

5.1.2 Device Evaluation

To evaluatethesedevices,two separatedataacquisitionsystemsto clock andreadout thearraysweredeveloped.Theonefor the 10x64 arraywas developedfirst, while the second,more sophisticatedsystemfor the 256x256 array wascompletedonly recently. The10x64systemis describedin our first publication1] while the secondis discussedbriefly in a papergivenby SteveCaalemaat the

28

1989InternationalIndustrialSymposiumon theSuperCollider in New Orleansearlier this year, and which is appendedto this proposal. Figure 1 is a blockdiagramof this system. At New Orleans, the hybrid conceptwas discussed,and its applicationto SSChighlighted. At that time we had just takensomepreliminarydatawith the 10x64 array, andit waspresented.

Testingof thesedeviceswith radioactivesourcesstartedin mid-september,1989.This additionaldata,takenwith the10x64 array,will be briefly presentedhere.A 1°6Rubetasourceandan 24Mm alphasourcewereusedto irradiatethedevice.Thepixel capacitanceis 80 x 10_is F. The pixel is 120 pm square.Spectraforboth the alphasand thebetaswereobtained.

The dataarepresentedin Figure 2. They havebeencastin a fashion whichallows, from an analysisof thesespectraonly, thecalculationof four quantities;signal to noise, noise,size of the chargecloud,and spatialresolution.

In our geometry,particlesenterthedeviceon thecathodeof the PIN diode, theside farthestfrom the bump bonds. Thus, muchof the chargecollectedmustdrift acrossthe entiredepletiondistanceof 300 pm. The chargecloud, spreadby scatteringof the initial radiationand by diffusion, will havea finite lateralsize. If the particlewere closer to the edgeof a pixel than to the center,onewould expectchargeto besharedby adjacentpixels. On average,we seechargespreadover nine pixels; a 3x3 array. If this spreadingis correlatedto the initalpositionof the particlewithin the pixel then a well definedlocus of points willbe evidentin Figure2 ratherthana randomdistribution. The axesin Figure2representthe ratio of the chargein the center-toplabeledearly pixel and thecenter-bottomlabeledlate pixel divided by the chargein the centercolumnof the 3x3 array.

The cluster of points near the origin representeventswhich have no chargedepositionin the off-center or adjacentpixels. Therefore, the varianceof thiscluster is a measureof the fluctuation of the ratio of the signal of a typicalnon-hit pixel and the total signal from the threepixels in a hit column. Thisvarianceis the inverseof the averagesignal to noiseratio. The signal to noiseratio for betasdeterminedfrom this varianceof 0.029 ± 0.007 is 34 ± 8. Theaveragesignal sizefor this dataset is roughly 6300 electrons.Thus, thenoiseis about 185 electronsrms. A similar calculationfor the alphadatapresentedin Figure 2 yields a varianceof 0.008 ± 0.001, a signal to noiseof 125 ± 15,and a noise level of 287 electronsrms.

Thesenoise measurementsareconsistentwith a calculationof KTC noiseand

29

dark currentnoise, which are the dominantsourcesof noise at room temperature. For our devicestheseare 110 electronsrms and 120 electronsrmsrespectivelyyielding a combinednoiseof 165 electronsrms.

From the ratio of the numberof datapoints which lie near the axes and thenumber of datapoints in the cluster near the origin, one can determinetheaveragesizeof the chargecloudproducedby an incident chargedparticle. Forbetasthis sizeis 19 pm 1 o’ and for alphasthis size is 28 pm 1 o.

To determinespatial resolutiononeusesthefact that the clusterof datapointsnear the origin correspondsto eventswhich depositnearlyall of their chargein the centralpixel. Thus, within this region,no interpolationis possible.Forthoseeventswhich arenearthe edgeof a pixel one canusethe variancenotedearlierto estimatetheerror in thepositionwithin thepixel. For the betaswhichstrikewithin 30 pm of anedge,the centroidof the 19 pm 1 or distributioncanbe locatedto within about2 pm. A similar resultsfor alphasis 2-3 pm acrossthe entire 120 p pixel, dueto thelargersizeof the alphainducedchargecloud.

Thesemeasurements,though preliminary, arethe first measurementsdemonstrating good performanceof a pixel arrayat room temperature.

The majoreffort of thelastsix monthswas thecompletionofthereadoutsystemfor the256x256array. Late in September,thelast bugswere removed,andthe256 x256 devicesin our possessioncould becharacterized.We haverun a curveof cathodesvoltage againstchargecollected,and convincedourselvesthat wefully depletethe detector. We also were able to measurethe dark current inour devicefor one settingof readoutspeed48 ms to readout theentirechip.At this speed,the dark currentwas within a factor of two of that measuredbyMicron Semiconductorfor thehigh purity siliconwaler usedin producingthesedevices. We havenot yet had the opportunity to calibratethe capacitanceofeachpixel so in the calculationwe used that of the 10x64 device. Figure 3shows the operationof the 256x256 PIN array. The deviceis operatedby thegenerationof clockpulsesto driverow andcolumnshift registerswhich scanthefull array. The electronicsthat control this operation is diagramedin Figure1. The upper panelof Figure 3 is a displayof a portion of the full readoutconsistingof two 128 row samples. The operationis shown for threedifferentsettingsof the detectorbias voltageand clearly indicatesthe increasein darkcurrent in responseto increasesin PIN bias voltage. The lower panelis anexpandedview of a 19 pixel segmentof a row. Theclock traceshowsthetimeof pixel resetwhich occursevery 2.7 ps in the exampleshown.

30

5.1.3 Device Design and Simulation

Last year, in anticipationof this proposal,we proposed[2 the creationof aprototypearrayby the HughesAircraft Co. which met someof the SSCspecifications. Thoughthis proposalwaspremature,its preparation,andthe designwork and simulationwhich followed haveprovideduseful ideasfor the presenteffort. Figure 4 is a schematicof our proposeddesignfor a devicewhich wasto be radiation hard, responsiveto chargedparticlesin 15 ns and which hada readoutschemewhich allowed the completeanaloginformation to be readout in severalhundredmicroseconds.An alternativearrayarchitectureis beingstudiedwhich requiresmore electronicson chip and less softwarefor readout.Such a designcanbereadoutfasterthansimpler alternatives.A descriptionofthis designis included as an appendixand representswork in progress.A preliminary studyof the interactionof thesedeviceswith a digital signalprocessorwascarriedout during the lastyear. Thestudyrevealedthat somearchitectureswould requireseveralhundredmicrosecondsfor thereadoutand calculationofthe centriod of chargeclustersinduced by particlehits. The parallel designwill operatesignificantly fasterthat the simplier serial designshownin Figure4 sinceit presentsthe datato the digital signal processorin a form readyforanalysis.

5.1.4 Conclusion

The UCB/SLAC/HUGHESgenericprogramhas reachedtheprimaryobjectiveof fabricatinganddemonstratingtheoperationof hybrid PIN diodearrayswitha room temperaturenoise of less than 300 electronsrms. Such devicescanbe usedfor somehigh energyphysics experimentstodayand are a significantstep towardsthe developmentof deviceswhich meet SSC requirements. Wehavealso developeda prototypesystemincludingboth hardwareand softwarenecessaryfor the operationof thesedevices. This systemservesas guide forthedevelopmentof a SSCcontrol systemaswell asprovidingthe capability toevaluatedevicesboth in the laboratoryandin thefield. We havealsodevelopedsomepreliminaryconceptsfor new deviceswhich approachSSC requirementsand havecarriedout a modestprogramof simulations.More work is necessaryto properlycalibrateand characterizethesedevices.

31

Figure 1: Block diagramof the control anddataacquisitionsystems

32

.3 I I I I I I I I I I I I I I I I I I I I

I cxx

.3 - xi

F- -a -I-.

2- x -#4-

4x

- Wrx -xx x x -- a xx xxx 1’ x0 ----p-

-.1 III! 11111111111111111111111-j 0 .1 .2 .3 .4 .5

LATE/TOTAL

liii liii IIIIIIIIJI1I1l III

Al -

A- A

-- AJ

ILA.3- -

- a. A -Alas

IA -F- - I

‘ .2-- A A’ AA -

A

- I AA

££

AA -I - A A A A --

s4 AA £ A -A I A A A A A£A £ A

A A AA AAAA AtI4AA Ao -7rr*-.r-r

A -

- AA

I A --* r -

-.1 11111I

11111111111 IIII I I I I I II I

-j 0 .1 .2 .3 .4 .5LATE/TOTAL

Figure2: Scatterplots of chargesharingfor a and fi particles

33

256 x 256 PIN Array

Low Response

Fair Response

Good Response

=

- __ - - - .--

- - - -- - - - -

‘F _tRow of 128 Pixels

TIME

ExpandedView in Time

, ,

_. ,A ‘.A_A’A A -

DetectorBias Voltage

2v

7v

22v

CorrelatedDouble Sample

Pixel Reset

Every 2.7 microseconds

TIME

Figure3: Oscilloscopetracesof clocksandoutputof 256x256 array

34

Figure4: Oneof the proposedreadoutarchitectures

Analog Out

Fast DigitalOR

Column Out

256 Row Out

Fast Clock In

Column PatternMemory

35

5.2 LBL TechnicalProgress

In this section,a very brief outlineof relevantcurrentactivities and progressisgiven. Effort hasbeendirectedin five areas:

a A RobustHigh-TemperatureDetectorFabricationProcess;

b Monolithic Integrationof Circuitry and Detectors;

c Low-power low-noise Analog Circuitry Optimization;

d RadiationDamageStudiesof Detectorsand Circuitry;

e PIXSTRIP, a designstudy for a smart pixel array.

Thesetopics aredescribedin greaterdetail in Appendix8.2, a reportpresentedat the 1988 SnowmassMeeting. Referencesareincludedthere.

5.2.1 A Robust High-TemperatureDetector FabricationProcess

In this work, a backsidephosphorus-dopedpolysilicon layer hasbeenpostulatedby S. Holland of LBL to offer excellentgetteringcharacteristicsfor heavymetalcontaminantsthat causeexcessiveleakagecurrentsand/or early breakdown behaviorin detectors.The techniquehasbeenfound to work extremelywell in practice,producingvery uniform andreproducibledeviceswith highperformance. This gettering techniqueactually turns the high temperaturestepsusedin conventionalIC fabrication to advantage,as the diffusive mobility ofthe contaminantsis high enoughat thosetemperaturesto captureessentiallyall contaminantswithin the electrically inert back layer. Strip detectorshavebeenproducedwith high yields, andexceptionalperformance.Thesearebeingutilized in the activitiesdiscussedbelow, as well as in a relatedeffort aimed ata silicon trackerwith optimizedpowerdissipation.

5.2.2 Monolithic Integration of Circuitry and Detectors

Building on the resultsmentionedabove, PMOS transistorshaveSbeen integratedon high-resistivitysilicon with high-qualitydetectors.Thesedeviceshaveboth channelstopand thresholdimplants, and comparefavorably with bettercommercialtransistors. A completeCMOS processis regardedas a straightforward extensionof this work, but effort will be neededto fully optimizethe

36

P-well characteristics.A novel JFET structurehasbeenanalyzed,and wouldappearto offer low-noise rad-hardperformance. All processingstepsare conventional,permitting commercialfoundriesto fabricatedeviceswith low risk.This is thefirst clearexampleof a techniquethat eliminatesthe traditional incompatibility of detectorfabricationand integratedcircuitry. A strip detectorwith an integratedPMOS amplifier/MPX hasbeenfabricated.

5.2.3 Low-power Low-noise Analog Circuitry Optimization

In this task,an intimate knowledgeof devicephysics, circuit design principles,radiationdamageeffects,and measurementsto definescalingbehaviorarecombined to synthesizeguidelinesand principlesfor optimizedlow-power low-noiseanalog circuitry. Both CMOS and bipolar technologieshave beenincludedintheseanalyses.Thesehaveapplicationto both pixel detectorsand strip detectors, and lead to novel configurationswith greatly reducedpower dissipationrelative to extensionsof currentdesign practice. The impact of thesestudieson overall feasibility for a silicon tracker is quite substantial. For pixel cells,designshavebeenevolvedthat areexpectedto demonstrateless than 100 electrons rms noise,with power dissipationsless than 10 pW perpixel. The useofCR-lW shapingwith peakdetectionhasbeendevelopedto minimize shot noiseunderconditionsof heavyradiationdamageto the detector.A pixel cell designwith optimized analogperformanceis scheduledfor implementationin siliconthis winter.

5.2.4 Radiation DamageStudies of Detectorsand Circuitry

In collaborationwith UCSC and Los Alamos, a set of exposureshasbeenmadeunder conditions that will allow the determinationof damageconstantsfordiffering resistivity, crystal orientation,and processsteps. The exposurescorrespondroughly to one week and one yearof SSC operation.The preliminaryresultsdo not supportthe prevailingmythology that thedamageconstantsdepend sensitivelyon the resistivity. Other aspectsof this work are systematicnoisemeasurementsof commercialradhardMOSFETs,an areaof great interest and little systematicinformation. A theoreticalanalysisof the relationshipbetweendetectorsize areaand thickness,signal/noiseratio, electronicintegration time, damageconstantand fluencehasbeenmade,which permits thestraightforwardevaluationof systemperformance.

37

5.2.5 PIXSTRIP, a designstudy for a smart pixel array

This device is a linear 128 channelarray implementedin silicon to test someof the logic and sparsedatareadoutfeaturesof the smartpixel arrayconcept.The sparsedatafeatureefficiently scanslarge addressfields to find true hitsbelonging to the time-slice of interest. A major result of this study is thatthe sparsedatascanproceedsto find a column or row with hit data at anextraordinaryscan speedof 660 MHz. On average,a column with a valid hit isfound in about 100 nanoseconds.In a two-dimensionalversion,logic is includedfor theinterrogationof row elementsto identify only true hits. Ambiguitiesarethus suppressedat the chip level. This built-in logic is capableof expressingthe two-dimensionaladdressof a hit pixel in about two hundrednanoseconds.Much of the circuit designconceptis an evolution of the successfulLBL SVXsilicon strip readoutIC.

References

[1] StephenL. Shapiro, William M. Dunwoodie,John F. Arens, J. GarrettJernigan,StephenGaalema,Silicon PIN Diode Array Hybridsfor ChargedParticleDetection,NIM, A275, 580 1989.

[2] StephenL. Shapiro,JohnF. Arens,J. GarrettJernigan,StephenGaalema,A VertexDetectorSystemfor theSuperconductingSuperCollider, submitted to Departmentof Energy,September,1988.

38

6 Statement of Work and Budget Summary

The following statementsof work and budgetsfor the various teamactivitiesareproposed.

6.1 Hughes Aircraft Company

a HughesAircraft Companybudgetrequest-$487K

b Perform systemlevel architectureand design studiesfor pixel detectorsystems,for vertextrackingfor both SDE and BCD pixel experiments;

c Performa preliminarycost analysisof the proposedpixel detectorsystems;

d Design fab and test pixel array addressingthe critical requirementsfortechnicalfeasibility.

e Personnel

i. Seeresumesin appendix8.1.4

6.2 Lawrence Berkeley Laboratory

a LawrenceBerkeleyLaboratory budgetrequest- $280K

b Providedesignsupport and systemanalysisto HughesAircraft Companyincluding a review of proposeddesigns;

c Continuedevelopmentwork on LBL detectorprocessto provideappropriatepixel detectorarraysfor HughesAircraft Company;

d Perform analysisof radiation damagestudieson detectors,circuitry, andhybrid arrayswith othercollaborationmembers.

e Personnel

i. SteveHollandwill devote40% of his time to thetaskof detectordevelopmentandfabrication,usingthenew LBL high temperaturedetectorprocess. Thesedetectorswill be matchedto the pixel cell geometry,to bedefinedand will incorporateAC coupling usingbuilt-in capacitors and polysiliconresistors.As a variety of detectorgeometriesareneeded,this effort level representsa minimum.

39

ii. David Nygren will devote80% of his time to physics direction andcoordinationof this entire proposedactivity. His salary is supportedby the LBL HE? budget.

iii. Helmuth Spielerwill devote40%of his time to thetasksof pixel arrayarchitecturedevelopment,circuit and systemanalysis for prototypesmart pixels, and detectorprocessdevelopment.Theseactivities willhave a closerelationshipto the effort describedin a SSC major subsystemsproposalfor a silicon tracking system,which will requirea50% effort. Theremaining10% is devotedto a genericR&D programemphasizingradiationdamagestudies.

iv. Michael Wright will devote40% of his time to the task of CMOSpixel cell circuit designand analysis,underthe guidanceof HelmuthSpieler. Pixel cell designswill be laid out, and prototypesfabricatedand tested.Thesedesignswill supportthe effort undertakenat HAC,and are also intendedto providean avenueto understandalternateconceptperformance. Another 50% of his time will be dedicatedtothe work for a silicon trackingsystem.The remaining10% is allocatedto a varietyof CMOS designprojects.

6.3 UC Berkeley - Space SciencesLaboratory

a UC Berkeley - SSL budgetrequest- $211 Kb Continuetestingof the currentpixel arraysthat havebeenprepared;

c Statemachineor digital signalprocessordesign;

d Assist in detectorreadoutdesign

e Testsystemdevelopment.

f Personnel

i. John Arens will devote50% of his time to helping designa readoutarchitectureand testing detectors,mainly in the SSL lab. He willparticipatein designingtheelectronicsfor operatingthedetectorsthatwill be designedand built aspartof this proposal.He will coordinateactivitiesbetweenthephysicsgroupsand Hughes.

ii. GarrettJerniganwill devote50%of his timeto helpingdesignthe dataacquisitionanddatastoragesystemand to designingandbuilding the

40

electronicsfor operatingthe detectorsthat will be designedand builtas part of this proposal. The designwork for the electronicscannotbegin until the architectureof the readoutchip is completed.

6.4 Stanford Linear Accelerator Center

a Stanford Linear Acceleratorbudgetrequest- $61K

b Performlaboratory,radiation,andbeamtestswith thecurrentpixel arraysandthosebeingdeveloped;

c Perform mechanicaland cooling studiesof proposedvertex detectorsystems.

d Assist in detectorreadoutandpixel architecturedesign

e Personnel

i. SteveShapirowill devote 100% of his time to testing detectorsandinvestigatingmechanicaland cooling systems. The testing will beperformedin two or threeareas:

A. Testing the 10x64 and 256x256 pixel arrays that were recentlyproducedby HughesAircraft and Micron Semiconductorat SSL,at accelerators,and in high radiationflux facilities.

B. Using the 256x256 array in the testsetupof Jim Russ’ E781 experimentat Fermilab. This work will be fundedby Russ.

C. Testingthedetectorarrays that will be designedand built aspartof the work of this proposal.

6.5 University of Oklahoma, Yale University, and University of Iowa

a Budgetrequest- $85K

b RadiationDamageStudies;

c Beamtestsat Fermilab;

d Preparationfor BeamTestsof a Vertex DetectorSubsystem.e Personnel -

41

1. Phil Gutierrezwill devote33% of his researchtime to testingstrip andpixel detectors. He will test both the existing 10x64 and 256x256arrayssadalso the detectorarraysthat will be designedand built aspart of the work describedin this proposal. Hewill haveno teachingresponsibilitiesin the spring of 1990.

ii. Pat Skubic will devote33% of his researchtime to testing strip anpixel detectors. He will test both the existing 10x64 and 256x256arraysand also the detectorarraysthat will bedesignedandbuilt aspart of thework describedin this proposal. Hewill haveno teachingresponsibilitiesin the spring of 1990. He will coordinateactivities atOak Ridgeif radiationsourcesthereareused.He will alsohandlethetravel accountsfor the University of Pennsylvania,Princeton,Yale,and the University of Iowa if the SSC choosesto consolidatesmallfinancial grants.

iii. Paul Karchin will devote5% of his time to reviewingdesignsand testresults.

iv. Ed MeClimentwill devote10% of his time to partkipation in beamtestingat FNAL.

6.6 UC Davis and Iowa State Universitya Budget request- $66K

b Physicssimulationsof systemperformancein SSC environment;

c Proton beamtestingof pixel detectorarrays.

d Personnel -

1. David Pellett UC Davis will devote20% of his researcheffort totesting the existing lOx 64 array and the arraysthat will be designedand built aspart of the work proposedhere. He will coordinatetestsat the U.C. Davis68 MeV cyclotron.Altogether, half of Pellett’s researcheffort will be devotedto SSCstudies under the existing UC Davis generic R&D program plus amajorsubsystemsproposalon silicon tracking,if thelatter is approved.in addition to pixel devicetesting, this programincludessimulationand design studiesof a pixel vertex detectorfor the central rapidityregion. Theseefforts, which involve additional personnel,complementthework of the presentproposal.

42

ii. John HauptmanIowa StateUniversity will devote25% of his timeto detectorsimulationsand will participatein detectortests. Fundingis requestedfor travel only.

6.7 University of Pennsylvania, Princeton University

a No funding requested;

b Providedesignanalysisand simulationfor BCD experiment.

c Personnel

i. Nigel Lockyer will devote40% of his researchtime to the designandsimulation of the BCD detector. About 1/3 of this time is usedforstudyingmechanicalandelectricalissuesrelatedto thevertexdetector.He will install the pixel geometryand gutter support structureintothe Geant simulationprogram. He will provide Monte Carlo eventanalysesto Hughesin order to study readout schemes. Raw dataratesper tile and clustersof tiles will be provided. He will simulatealignment effects on B reconstructionefficienciesand on backgroundrejection. He will ensuregood communicationbetweenthis projectand the BCD dataacquisitionsystembeing designedby Ed Barsottiof Fermilab.Nigel Lockyerassumesthecomputingresourcesof this project will beprovidedby theFarmproposalof Lockyer,McDonald,et al submitteedto the SSC laboratory. The overlapin the softwaredevelopmentforthepixel projectandtheFarmstudiesarenearly100%. Fundsfor thiswork are requestedin theFarm proposal.

ii. Kirk McDonald will devote10% of his time to simulatingeventsandreviewingdesignsand test results.

Total Cost: *1190K

43

7 Budgets and Schedules

The budgetsfor the proposedwork are given below for each institution. Aschedulefor afive yearprogramto producea completeprototypevertexdetectorfor theSSCis givenin Figure 1. Theschedulefor thework proposedherecoversoneyearand is shownin Figure 2.

Hughes Aircraft Company Budget Summary

Total Material $ 23,600

Direct Labor Elements-Hours

Membertechnicalstaff 1,234Associateengineer 156Researchassistant 48Seniormemberof technicalstaff 1,776Senior technicalengineer 98Total Direct Labor hours: 3,310

Other Direct Costs-Dollars

Waler processing $ 60,708Dataprocessing 3,000Travel 6,954Total Other $ 70,662

Total EstimatedCost: $ 425,870

Total Costof Money: $18,414Total EstimatedCost Plus COM: 444,284Feeor Profit: 42,601

TOTAL PRICE: $ 486,885

44

Lawrence Berkeley Laboratory Budget Summary

Scientific/EngineeringPersonnel$ 150,000/FTE:

Helmuth Spieler 0.4 FTESteveHolland 0.4 FTEMichael Wright 0.4 FTETotal 1.2 FTE $ 180,000

TechnicalPersonnel$ 100,000/FTE:

Distributedeffort 0.5 FTE $ 50,000

Equipment: nonerequested

Suppliesand Expenses:

MOSIS submissions $ 15,000Travel 12,000Shops 6,000Supplies 6,000Noncapitalequipment 11,000Total Suppliesand Expenses: $ 50,000

Total Support Request: . $ 280,000

45

University of California-Space SciencesLaboratoryBudget Summary

Labor

Arens 25% $ 22,000Jernigan 25% 16,000Engineer 50% 30,000Technical 25% 12,000Programmer 50% 22,000Total Labor $101,000

Travel 10,000

CapitalEquipment 18,000

Overhead 57,000

Total Support Request: $ 211,000

SLAC Budget Summary

Material

Mountinghardwarefor 256 x 256 device $1,850Powersuppliesand clock driversfor radiationtests 3,000

Laborhours

Mount for 256 x 256 device 160 $ 4,000Powersuppliesanddock drivers 125 3,000Softwaresupport-simulationof mounting, cooling 500 15,000Softwaredevelopment 250 7,500

Travel 8,000

Overhead 18,150

Total Support Request: $ 60,500

46

UC Davis Budget Summary

StudentLabor $10,000

Equipment-SUNworkstationetc. for radiation tests 15,000

Supplies 2,000

Beamtime for radiationdamagestudies 10,000

Travel including budgetfor J. Hauptman 14,000

Overhead 15,000

Total Support Request: $ 66,000

47

SSC- LONG RANGE PIXEL DETECTORDEVELOPMENTPLAN

SCHEDULE 7/89 1/90 1/91 1/92 1/93

PIXEL ARRAY DEVELOPMENT* PREUMINARY DESIGNS

* INTERIM PROTOTYPE A DEMO ‘I.

* INTERIM PROTOTYPE B A /

* FINAL. PROTOTYPE 4*BEAMTESTDEMO

< >

SIGNAL PROCESSING* CONCEPT DESIGN, SIMULATION 4’* BRASS BOARD DEMO

* INTERIM PROTOTYPE A-VlSI CIRCUIT DESIGN/FAB

C * ANAL PROTOTYPE DESIGN/MB

* BEAM TEST DEMO

.

I .

1.I

A

t

A

J IA

‘1>SYSTEM DESIGN

* CONCEPTUAL EXP LOCKYERDESIGN

i A %

EXP TRIWNG

* MECHANICAL MOUN11NG

* COOLING

* PERIPHERAL ELECTRONICS

* KEY HARDWARE DEMO

* FINAL DESIGN .

. 4’ ‘

FULL SCALEIJJGINEERINGDEVELOPMENT

.

/ .S £ t

HAC BUDGET S 5K + 4I 2.51.1 3.51.1FUNDING SOURCE SLAC SLAC + DOE DOE DOE

TEAMTOTAL 81070K 3.SM 4.5*4

1 EXISTING FUNDING

2 ANTICIPATED FUNDING3 THIS PROPOSAL

I

Figure 1. SSClong rangepixel detectordevelopmentplan.

University of Oklahoma Budget Summary

Material

Mechanicalsupportsand cables $ 6,000

Mechanicalsupports,interfacecards,cables 3,000

Irradiationchargesand instrumentation 5,000pixel testsat UTK or Hughes

Pixel detectormechanicalsupportsandcables 13,000"CO" tests

Labor

GraduateResearchAssistant,1/2 year 12,000

Capitalequipment

SUN-4 coliaputerand support hardware 18,000

Travel 7,400

Overhead 5,044

Total Support Request: $ 69,444

University of Iowa Budget Summary

Travel 7,500

Total Support Request: $ 7,500

Yale University Budget Summary

Travel 7,500

Total Support Request: $ 7,500

48

Themilestoneschedulefor theHugheseffort is shownin Figure2. Regularmeetingswill beheldat2 and3 monthintervalsamongall teammembersto ensureaccurateandup todatecommunicationandtechniCalinterchange.A ConceptDesignReview,Preliminary DesignReview,andFinal DesignReviewwill be heldto ensurethat thereadoutdesignwill satisfytheappropriatesystem requirements.

SSC PROGRAM - PIXEL DEVELOPMENT - HUGHES AIRCRAFT COMPANY

SCHEDULEMONThS

o i a jjK

GROUP MEETINGS

SYSTEM CONCEPTDESIGN/COST ANALYSIS

SYSTEMREQUIREMENTSFLOWOOWN/ANALYSIS

CONCEPT DESIGNREVIEW

PREUMINARY DESIGN’LAYOUT

PREUMINARY DESIGNREVIEW

FINAL DESIGN

FINAL DESIGN REVIEW

WAFER FABRICATION

DETECTORASSEMBLY/TEST

DEU VERY OF HYBRIDARRAYS’

FINAL REPORT

CKOFF

>a

0

0

0 0 0-----A

V

0 I

INI FROM TEAM COLLABORATORS

* MN OF SIX HYBRID ARRAYS DEUVERED TO TEAM COLLABORATORS

A

Figure 2. HugheS Pixel Vertex Detector SystemDevelopment for SCC program-Miestoneschedule.

8 Appendices

8.1 HUGHESAIRCRAFT COMPANY CAPABILITIES

The electronic readoutand hybrid technology,which are critical to the developmentof hybrid pixel detectorarraysfor the SSC,exist at Hughesasa resultof the work in long wave infrared LWIR sensortechnology. The SSC similarly requiressensitive,radiation hardened,high speedpixel detectorarrays.Readoutarraysand their hybridizationhave alreadybeenmadeavailable toSLAC and UCB on an existing SSCdetectordevelopmentproject "Silicon PinDiode Hybrid Arrays for ChargedParticleDetection: Building Blocks for Vertex Detectorsat theSSC." 0. Krameret al., presentedat the 1989 InternationalIndustrialSymposiumon theSuperCollider, New Orleans,LA, February8-10,1989.

Someof the key technologyrequirementsfor the SSC detectorare:

a Thin substratedetectorsto minimizemultiple scatteringeffects;

b High speedoperation to matchbeamcrossingand particlerates;

c Radiationhardness.

Ourcapabilityto "thin" detectorchipsis demonstratedin Figure 1, which showsdetectorchips thinned to 50 m and hybridized to a readoutchip. Likewise,readoutchips can be thinned.

Hughestechnologycan meet the high speedoperationneedsfor the SSC asshown by the performanceof shift registers. The latest Hugheshigh speedsilicon integratedcircuit was recently tested. It uses the lowest risk siliconon sapphireSOS processFigure 2 technologyat Hughes. The chip is adigital fully pipelined 6x6 bit multiplier that operatesat speedsgreaterthan132 MHz. The chip has8340 metaloxide semiconductorfield effect transistorsMOSFETs, and the latest lot has an 85 percentoperationalyield. Usingthis exacttechnology,onecouldexpecta digital shift registercircuit to operateat speedsbetween300 and 500 MHz. The circuit uses 1.5 pm gatesand 5-Vsupplies. Figure 3 shows the scopephotos of the chip operation. The clockis running the chip at 125 MHz, and the chip responseis at 100 MHz. Fromthe photo, it is clear that, if the clock rate could be increased,the chip couldrespondat 250 MHz. TheSOS technologyis radiationhard to greaterthan 1MradSi of total dose.

49

Figure 1: Thinnedchip technology. Top chip on hybrid array is thinnedto 50 pm.

50

In thepast4 years,morethan$50M of contractfundsfor hybrid detectorarraytechnologyhas beenreceived by Hughes. The experiencegainedfrom theseefforts, in addition to companyfunded IR.&D, will ensurethe successof theproposedSSC detectordevelopmentprogram.

Hughesis a leader in the developmentof LWIR sensortechnologyfor the Departmentof Defense,havingcompiledmore than95 percentof all LWIR spacebackgroundflight dataand performedmore that 70 percentof all successfultargetmeasurementflights using Companysensors.In 1964, Hughesdevelopedthefar infraredsearchtracksensor,the first LWIR sensorto be twice flown successfullyin space1967, followed in 1970 by the first ballistic missile defensefly along infrared FAIR I sensors. Thesesensorssuccessfullygatheredthefirst comprehensiveLWIR radiometricmeasurementsof ballistic missile targetsin flight. Hughesprovidedthe first mosaicsurveillancefocal planeto be flighttested,HICAMP I in 1978, and the largestmosaicsurveillanceinfrared focalplanechip to be flight tested,HICAMP 11 in 1983.

The HI-STAR sensor,designedand developedby Hughesin 1971, completed11 flights on Aerobeesoundingrockets,gatheringtheLWIR stellar backgrounddatathat todayremainsthedefinitivecomprehensivedatareferencebasein thiscountry. The HughesCelestialMapping Programsaw thefirst orbital flight ofan LWIR sensorcooledby a closed cycle refrigerator;manyorbits and manystellar backgroundmapswere completed. The P11 sensor,developedin 1971,wassuccessfullyoperatedin the field from 1974 through 1978 to obtain LWIRsignaturemeasurementsfrom an airborneplatform.

The Hughesdesignatingoptical trackersensors,with five successfulflights todate,provide the basic instrumentsfor the only continuinglow backgroundLWIR exoatmospherictarget radiometric signaturemeasurementprograminexistence.TheHughesminiaturevehiclesensoris thefirst plannedoperationalLWIR sensor,and the spaceinfraredexperimentrepresentsthe only advancedsatellite-basedLWIR measurementsensorcooled by a closedcycle refrigerator.The Airborne Optical Adjunct AOA sensorprogram, which hasdevelopedthe most advancedlit sensorever built, is in final test. Delivery to systemintegrationwascompletedin July 1988.

Each year Hughesinvests heavily in independentresearchand developmentIR&D programsand is strongly committedto advancementsin the field ofsensortechnology. Special emphasisis given to spacerelatedfocal planedevelopment,including detectors,readouts,and signalprocessing.Much of this

51

Figure 2: Sos6x6 bit multiplier, composedof 8340 MOSFETs,operatesat 132 MHz.

52

CLOCKIN! 125MHZ OPERAIJIC AT -100 MHZ

tWPER UMfl. - no MHzWITH FASTER CLOCK

Figure 3: sosdigital multiplier demonstratesability to operateat 250 MHz.

CLOCKNUT .u-2S CHIP RESPONSE iuIt 2 s

53

IRScD will directly benefit SSCprograms.

8.1.1 MOSAIC ARRAY DEVELOPMENT

The mosaicinfraredsensortechnologyMIST and forward acquisitionsensorFAS optical sensorprogramsrequiredintensiveefforts to develophigh densitymosaicfocal planearrays. Hugheshasled in thedevelopmentof state-of-the-arttechnologiesnecessaryto support major improvementsrequiredin next generation infrared sensorperformance;the monolithic focal planearray MFPAwas demonstratedin 1973, the first monolithic extrinsicsilicon infraredsensitive chip in 1974, and the first incorporationand flight of a many thousandelementMFPA in 1976. Theseadvanceshavepacedthe industry. The DefenseAdvancedResearchProjectsAgency ChargeCoupledDevice program,the Mini-HALO program,the Air Force’s Forward Looking Infrared Technology Developmentprogram,andthe Army’s TaákBreakermissileseekerall useHughesmosaictechnology. The USAF spacebasedsurveillancesystemdetector technologyprogramand the Ballistic Missile DefenseAdvancedTechnologyCenter’sMIST and FAS programshaveled to the designevolutionof the highperformancesourcefollower perdetectorSFD concept,which is thekey to thelow noise integrationand readoutof the tensof thousandsof low noise, background limited detectorsrequiredfor spacesurveillanceand tracking systemdevelopment.

The AOA programdevelopedand manufactureda focal planewith 15 modules, eachconsistingof 2,560 Si:Gadetectorelementsand devicesthat measurecontaminationlevel and sensoroptical image quality. In addition, the SensorExperimentalEvaluationand ReviewSEER programhassuccessfullydemonstratedthe feasibility of producinghardenedsensorchip assemblySCA hybrids with LWIR Si:As detectorarraysand SFD multiplexer readouts. ThePrecursorAbove the Horizon SensorPATHS programhasnot only demonstratedreproducibility by fabricatingSi:As impurity band conductionhybridswith high yield but hasalsodevelopedhigh performancesiliconcapacitivefeedback transimpedanceamplifier andgermaniumSFD multiplexerreadouts.Thisprogramis also establishinga documentedpre-pilot processcapability for IBCdetectorsand silicon readouts.The experienceandcapabilityestablishedfromthe SEER and PATHS programswill be directly applicableto the technologydevelopmentof the SSCDetectorSystem.

54

I

In addition, developmentefforts on GaAs integratedcircuits for both digitaland radiationhardenedanalogreadoutcircuits arebeingpursuedat Hughes.

8.1.2 ORGANIZATIONAL RESOURCES

Hugheshasthe organizationaldepth and commitmentto successfullyconductthePixel VertexDetectorSystemDevelopmentfor SSCprogram.Thecorporatestructureof Hughesis shown in Figure 4. Managementof the Companyisvestedin M.R. Currie, Chairmanand ChiefExecutiveOfficer, andD. H. White,Presidentand Chief OperatingOfficer. TheElectro-OpticalandDataSystemsGroup EDSG, which employsover 10,000 personnel,is responsiblefor theoverallmanagementofresearch,development,test,manufacture,andevaluationof tactical, laser,and spaceelectro-opticalsystems. EDSG, headedby R. D.Brandes,Senior Vice Presidentand Group President,will be responsibleforthe overallmanagementand coordinationof the Pixel Vertex DetectorSystemDevelopmentfor SCCprogram.

The HughesMicroelectronicsCenterHMC hasextensiveexperiencein advancedsilicon detectorsand on-focal-planesignalprocessingelectronics.Thus,it can developand manufactureadvancedhybrid arrays to meet SSC systemrequirements.Hugheshas developedand demonstratedthe advancedanalogradiationhard CMOS processfor readouts.

The HughesResearchLaboratoriesHR.L directly support the fundamentaltechnologyrefinementand processoptimization with their extensiveanalysisandmensurationcapabilities.

Hughesis strongly committedto thedesignanddevelopmentof superconductingsupercolliderdetectorsystems. A highly skilled teamhasalreadybeenformedasthe nucleusof the proposedprogramand is organizedasshown in Figure 5.Hughesis fully committedto providing the personneland facility resourcestoensurethe successof this program.

8.1.3 FACILITIES

Electro-Optical and Data Systems Group Complete facilities for thePixel Vertex DetectorSystemDevelopmentfor SSC programareavailableatEDSG’smodernengineeringandmanufacturingfacility. This 1,750,000-square-foot facility, locatedin El Segundo,CA, is specifically designedand equipped

55

to facilitateall aspectsof electro-opticalproductionand technologicaldevelopment.

56

HUGHESAIRCRAFT COMPANY

ioec -.MRaJat *Nt.*TN0" ,Amsn. It TtsP J SCiaTZ $ CFASMTH *DIlWIAtj

S

-a In__v I -_ It II *11MA1 I-=_I L..J i=i Lt=iI s-wC-

--aON wrnU

- 00I

AM mtIa- -4.n *11 1tM J10

Jrato IWIICS 1*00*11* WOW-fl Av.a

--a-- ...t.’r....

‘itvia

S - Pl.4*A. Itt ii - a 0IMI- It_

I

C,’-1

=- .a --I £PiC’ I *NmOfl I 000Mt

__________

L 1 - I PAO

I

ua.w3t?a..p.s... a* * SO. IM OMC’* SUON

"I. Pnd- P44t0.U41.0I0".- - p...S

N0401* ** P40.1* I S -l tWfla

._ - -

Inn vasle.’flhI

At A0014IaWI -

-ra in n.ss

At IiSMin

- pM_t.,= ow.

*3

A P

- .0*-

In.-- IS a

Figure4 Hughes corporateorganization.

Integratedtesting laboratories,computer and processingfacilities, and engineeringpersonneloffices are available to support hybrid array development.Facilities and equipmentare available to measureand evaluateall the criticalperformanceparametersof integrateddetectorassemblies.Cobalt radiationtestequipmentfor radiationhardnessverification is also available. Someof thekey facilities andequipmentarelocatedin theSpaceand StrategicEngineeringDivision’s IntegratedFocal PlaneArray IFPA Laboratory. This laboratoryincludesa dedicatedcomputer-aidedfacility for testingandevaluatingthe performanceof detectors,readouts,and signal processingelectronics. It has fourpermanent,fully equippedtestconsoles.Flying spot infraredscannersmeasuredetector/readoutinfraredperformanceovera wide rangeof dwell timesandsignal and backgroundphoton fluxes in thesensorsystemsbeingsimulated.Thisfacility hasbeenfurtherenhancedby the additionof a computerandlaboratoryinterfacesystemthat canbe connectedto the test instrumentationto log dataand control testsequence.

Hughes Microelectronics Center HMC, with a staff of approximately600 people,is locatedin a 150,000square-footfacility in Carlsbad,CA. Varioussemiconductordevicesare producedhere, including silicon infrared detectorarrays,advancedreadouts,SOS/CMOS,germanium,andbipolar CMOS mixedtechnologychips. Facilities exist to design, process,test, and packageverycomplexanalogand digital integratedcircuits. The programssupportedrangefrom low volumeprototypepartsfor systemsin theearly stagesof developmentto parts fabricatedaccordingto appropriatemilitary standardproceduresforspace/flightprograms. The facility containsthe VHSIC 1 pilot line.

HMC deliversapproximately4000 parts per year for developmentand ffightprograms. In the digital SOS/CMOS area, the following programsare representativeof this capability: the Miniature Vehicle Sensor,Thornton SecureCommunication,MILSTAR signalprocessing,M-1 Abramstank, forwardlooking infrared multiplexers,USASDC VLSI radiation hardening,and very highspeedintegratedcircuits. For theanalogcommandand control electronicsandsilicon focal planearrays FPAs, successfultechnologydemonstrationshavebeenachievedin theAdvancedAnti- tankWeaponSystem-Medium,FiberOpticGuidedMissile, High EndoatmosphericDefenseInterceptor,Highly CalibratedAirborne Measurementsprogrammodule usingdopedsilicon areaarrays,Thdent programusing a visible imaging array, and a variety of other programssuchas AdvancedSensorDemonstration,FAS Joint ServicesSeeker,PATHS,

58

HAC SSC TEAM

SPACE & STRATEGICENGINEERING DIVISION

B. SKEHAJ’IDIVSJOA Ib

[EDSG

SPACE & TECHNOLOGYPROGRAMS

J.W. DAIBER_

[EDSG

PROGRAM MANAGER

G. KRMER

__

[SG

CONTRACTtYSTEM

ADMINISTRATIONENG*4EENG

________ ________

WFI..AUGH

P. FOSTER C. PFEFFERJ. WV9TERSERGEDSO FEDSG

BUSINESSMANGER

______

N. FENTa’J

I EDSGCIRC*AT DESIGN I

MB/TESTK. REYZER

flat

____

D.wGwaj

Lfl]

EDSG * EL.ECTROPTICAJ. Ale DATA SYSTEMS GRQJP1aCWE5AICRCELECThCWJTER

Figure5: Hughesorganizationfor the PixelVertex DetectorSystemDevelopmentfor

SSC program.

59

and AOA. Theseefforts demonstratea unique ability to designand fabricatestate-of-the-arthardware.

Processing:The HMC processinglaboratory is responsiblefor the fabricationof prototype linear integratedcircuit devices,as well as the developmentofstate-of-the-artprocessingtechnologyin support of Hughes’ numerousMFPAprograms.High quality workmanshipand stringentprocesscontrol