A 1.2MeV, 100mA Proton Implanter

13
1 A 1.2MeV, 100mA Proton Implanter Geoff Ryding, Ted Smick, Bill Park, Joe Gillespie, Paul Eide, Takao Sakase, Kan Ota, Ron Horner, Drew Arnold Abstract A high energy, high current proton implanter has been developed for producing thin layers of c‐Si crystalline silicon with thicknesses in the range of 5‐20 µm. The implanter described is designed to operate at energies between 0.4‐1.2 MeV with proton beam currents up to 100 mA. This system can also be used to exfoliate other materials such as GaAs and SiC. Introduction There is an increasing demand for renewable energy using photovoltaic technology. In particular, photovoltaic cells are commonly fabricated on crystalline silicon wafers which are conventionally produced by slicing an ingot of single crystal silicon. This method of producing photovoltaic cells accounts for approximately 80% of the present worldwide production of solar cells. Unfortunately 50% or more of the silicon is lost as a result of the kerf consumed in the slicing process. Furthermore, the minimum wafer thickness that can be produced by sawing or slicing is ≥115 µm, which is much thicker than needed for optimum photovoltaic devices (1). Thinner silicon lamina can be made by implanting hydrogen ions to the desired depth below the surface of the silicon wafer. After a suitable number of ions have been implanted, a thin layer of micro bubbles is formed and on subsequent heating of the wafer, the top layer exfoliates to produce a thin layer of silicon with a thickness precisely equal to the depth of the original implanted layer (2). Using this technique, as many as 10 or more layers of silicon can be peeled from a standard silicon wafer, each having the ideal thickness for the fabrication of an efficient photovoltaic cell. In this way the cost of the single crystal silicon material, which is generally the dominant cost component of the finished cell, can be reduced by 80%. An added benefit is the fact that the cells made from the thin lamina are flexible. In the last 3 years we have investigated the potential advantages of lamina based photovoltaic cells in detail (1). Early work was based on an air insulated DC accelerator operating with a ribbon beam of protons in the energy range of 350‐420 keV and beam currents of 10‐75 mA (3). Critical process steps including light trapping schemes, and metal contact techniques have been developed using the 4.5 µm lamina produced by this machine. As the lamina thickness increases the cell efficiency can increase but the cost of increasing the energy and operating voltage of the accelerator also increases. The implanter described here operates routinely at energies between 0.4‐1.2 MV and beam currents up to 100 mA. These basic parameters represent an optimum compromise between cell performance and total implanter operating cost. An ion implanter suitable for solar cell production of the type described has three primary requirements: an acceleration scheme which will accelerate the ions to the velocity required for penetration to the appropriate depth below the surface, a beam current intensity that results in the required system productivity and high speed wafer handling.

Transcript of A 1.2MeV, 100mA Proton Implanter

Page 1: A 1.2MeV, 100mA Proton Implanter

1

A 1.2MeV, 100mA Proton Implanter GeoffRyding,TedSmick,BillPark,JoeGillespie,PaulEide,

TakaoSakase,KanOta,RonHorner,DrewArnold

Abstract Ahighenergy,highcurrentprotonimplanterhasbeendevelopedforproducingthinlayersofc‐Si

crystallinesiliconwiththicknessesintherangeof5‐20µm.Theimplanterdescribedisdesignedtooperateatenergiesbetween0.4‐1.2MeVwithprotonbeamcurrentsupto100mA.ThissystemcanalsobeusedtoexfoliateothermaterialssuchasGaAsandSiC.

Introduction Thereisanincreasingdemandforrenewableenergyusingphotovoltaictechnology.Inparticular,

photovoltaiccellsarecommonlyfabricatedoncrystallinesiliconwaferswhichareconventionallyproducedbyslicinganingotofsinglecrystalsilicon.Thismethodofproducingphotovoltaiccellsaccountsforapproximately80%ofthepresentworldwideproductionofsolarcells.Unfortunately50%

ormoreofthesiliconislostasaresultofthekerfconsumedintheslicingprocess.Furthermore,theminimumwaferthicknessthatcanbeproducedbysawingorslicingis≥115µm,whichismuchthickerthanneededforoptimumphotovoltaicdevices(1).Thinnersiliconlaminacanbemadebyimplanting

hydrogenionstothedesireddepthbelowthesurfaceofthesiliconwafer.Afterasuitablenumberofionshavebeenimplanted,athinlayerofmicrobubblesisformedandonsubsequentheatingofthewafer,thetoplayerexfoliatestoproduceathinlayerofsiliconwithathicknesspreciselyequaltothe

depthoftheoriginalimplantedlayer(2).Usingthistechnique,asmanyas10ormorelayersofsiliconcanbepeeledfromastandardsiliconwafer,eachhavingtheidealthicknessforthefabricationofanefficientphotovoltaiccell.Inthiswaythecostofthesinglecrystalsiliconmaterial,whichisgenerallythe

dominantcostcomponentofthefinishedcell,canbereducedby80%.Anaddedbenefitisthefactthatthecellsmadefromthethinlaminaareflexible.

Inthelast3yearswehaveinvestigatedthepotentialadvantagesoflaminabasedphotovoltaiccellsindetail(1).EarlyworkwasbasedonanairinsulatedDCacceleratoroperatingwitharibbonbeamof

protonsintheenergyrangeof350‐420keVandbeamcurrentsof10‐75mA(3).Criticalprocessstepsincludinglighttrappingschemes,andmetalcontacttechniqueshavebeendevelopedusingthe4.5µmlaminaproducedbythismachine.Asthelaminathicknessincreasesthecellefficiencycanincreasebut

thecostofincreasingtheenergyandoperatingvoltageoftheacceleratoralsoincreases.Theimplanterdescribedhereoperatesroutinelyatenergiesbetween0.4‐1.2MVandbeamcurrentsupto100mA.Thesebasicparametersrepresentanoptimumcompromisebetweencellperformanceandtotal

implanteroperatingcost.

Anionimplantersuitableforsolarcellproductionofthetypedescribedhasthreeprimaryrequirements:anaccelerationschemewhichwillacceleratetheionstothevelocityrequiredforpenetrationtothe

appropriatedepthbelowthesurface,abeamcurrentintensitythatresultsintherequiredsystemproductivityandhighspeedwaferhandling.

Page 2: A 1.2MeV, 100mA Proton Implanter

2

Thenovelimplanterdescribedoperateswithprotoncurrentsupto100mAandenergiesupto1.2MeVandthewaferhandlingnon‐productivetimeis<8%ofthetotalcycletime.

System Architecture and performance targets Thehighbeampowerrequirementsoutlinedpresenttwoprimarychallenges.Firstthe120kWionbeam

mustbegeneratedwithreasonablepowerefficiencyandsecondtheprocesschambermustbecapableofabsorbingthispowerwithoutexcessiveheatingofthetargetwafers.

ThechallengeshavebeenmetbyadoptingthesystemarchitectureshowninFigure1whichidentifiesthemajorcomponentsoftheimplanter.Thereare4basictypesofionacceleratorstructuresthatcanbe

usedtoaccelerateionstotheMeVenergyrange:linearhighfrequency(LINAC),circularhighfrequency(Cyclotron,synchrotron),RFQ(RFquadrupoleaccelerators)andDC(SingleendedorTandem).Thepowerconversionefficiencyofthesesystemscanbedefinedastheratiooffinalbeampowertototal

inputpower.InthecaseofRFtypeacceleratorsthepowerefficiencyisgenerallylow(≤20%)asaresultoflossesintheamplifierandresonantcavitycomponents.ForthisreasonaDCapproachwasadoptedusingstandardcommercialhighvoltagepowersupplies

Figure 1  System architecture 

Theionsource,theaccelerationtubeandthehighvoltagegeneratorarecontainedwithinapressurevesselwithadiameterof2.1mandalengthof4.1m.ThevesselcontainscompressedSF6gasatapressureintherange50‐100psitofacilitateoperationatvoltagesupto1.2MV.Theionbeamwhich

emergesfromthisbeamgenerationsystemisfocusedbyamagneticquadrupolelensanddirectedintoamagneticscannersystemwhichdeflectsandscansthebeaminthehorizontalplaneasshown.The

Page 3: A 1.2MeV, 100mA Proton Implanter

3

scannedbeamisthendeflectedandanalyzedbyamagneticdipolemagnetwhichservestofilteroutunwantedionbeamsanddirectthecollimatedbeamontotheinsideofalargediameterrotating

processdrum.Thepseudo‐squaresiliconwafersaremountedontheinsideofthedrumduringimplantandbetweenimplants,highspeedrobotsareusedtoloadandunloadthesewafersthroughavacuumload‐locksystem.

ThebasicperformancetargetsforthesystemaresummarizedinTable1.

Table 1  Implanter design targets 

LaminaThickness 5‐20micronsLaminaProductionrate 150/hourSystemEnergyrequirements <2kWhrs/laminaIonBeamEnergy 0.4‐1.2MeVProtonbeamcurrent 100mAPowerefficiency(Beampower/totalpower) >40%Machineutilization >90%

Voltage Generator MosthighpowerDCionacceleratorshaveincorporatedsomeformofcascade‐rectifiervoltagemultiplierbasedontheconceptsofJohnCockroftandErnestWaltontogeneratethehighpotentials

requiredtoacceleratetheions(4).Theoutputofthegeneratorisconnectedtotheionsourceanddeterminesthefinalenergyoftheions.Ionsextractedfromthesourcepassintoanaccelerationtubeconsistingofseveralelectrodesandtheappropriatebiasvoltagesfortheseelectrodesisgenerally

derivedfrompotentialdividerresistorsconnectedtotheoutputofthehighvoltagegenerator.Asthebeamcurrentinsuchasystemisincreased,severaldifficultiescanbeencountered.ThetwoprimarychallengesaresummarizedinTable2.

Table 2  Fundamental challenges for high current, high voltage DC accelerators 

1.Highvoltagegenerationandpowerefficiency

Sincetheoutputpowermustflowthrougheachmultiplierstageinseries,theavailableoutputpower,thevoltageregulationandthesystemefficiencydeclinesasmorestagesareadded.

2.Productionofastableelectrostaticaccelerationfield

Smallbeamcurrentsinterceptingtheaccelerationelectrodescanresultinlargeperturbationsoftheacceleratingfieldwhichinturnleadstobreakdownandavalancheeffects.Theseinterceptedbeamsmayoriginateasaresultofsmallaberrationsorchargeexchangewithresidualgasmoleculesinthevacuumsystem.Attemptstominimizetheseeffectsbydecreasingthevalueofresistivedividersresultsinwastedpowerandfurtherlossofefficiency.

Wehavedevelopedanovelpowersupplysystemwhichovercomesthetwoprimarychallengesoutlinedabove.ThesystemconceptisshownschematicallyinFigure2.

Page 4: A 1.2MeV, 100mA Proton Implanter

4

Figure 2  Schematic of the high voltage generator concept 

Asystemisshownschematicallywithfiveelectricallyisolatedalternatorsmountedonacommon

insulatingdriveshaftwhichinturnisdrivenbyacommondrivemotor.Eachalternatorprovidesanelectricallyisolatedsourceofpowerforitscorrespondinghighvoltagepowersupply.Asshown,thesepowersuppliesareconnectedinserieswiththefinaloutputattachedtotheionsourceandeach

intermediateoutputisconnectedtothecorrespondingelectrodeoftheaccelerationtube.ThemainadvantagesofthisarchitecturearesummarizedinTable3.

Table 3  Advantages of the system architecture 

1 Theelectrostaticfieldintheacceleratortubeispreciselycontrolledbytheregulatedoutputsettingsoftheindividualpowersupplieswhicharecapableofregulatingvoltageoverawiderangeofcurrentloads.Theinfluenceofspuriousbeamstrikeorscatteredbeamandsecondaryparticlesisvirtuallyeliminated.

2 Fieldgradientsalongthetubecanbeeasilycontrolledtogiveoptimumionopticalpropertiessimplybyadjustingtheoutputsettingsofeachindividualpowersupply.

3 Individualsectionsoftheaccelerationtubecanbe‘voltageconditioned’withoutinfluencefromotherregionsofthetube.

4 Themaximumvoltageandcurrentcanbescaledbyselectingtheappropriatenumberofpowersuppliesandalternators.

5 Standardcommercialpowersuppliescanbeused

Page 5: A 1.2MeV, 100mA Proton Implanter

5

Thehighpowerimplanteritselfhasbeenconstructedusingthreehighvoltagegeneratorassembliesinsidethehighpressurecontainmenttank.Eachgeneratorstackconsistsofa100HP,3‐phaseAC

inductionmotorwithanefficiencyof94%,five12kVA,permanentmagnet,3‐phaseACalternatorsoperatingat80%efficiencyandfiveseriesCockroft‐Waltonpowersupplies(5)operatingatvoltagesupto80kV,currentsupto125mAandefficienciesof85%.

OneofthesegeneratorassembliesisshowninFigure3.

Figure 3  High voltage generator assembly 

AphotographofthecompletehighvoltagegeneratorsectionofthemachineisshowninFigure4.

Page 6: A 1.2MeV, 100mA Proton Implanter

6

Figure 4  The high voltage generator 

Inthisconfigurationthereare15isolatedalternatorsand15independenthighvoltagepowersupplies

connectedinserieswhichresultinamaximumcombinedcapabilityof1.2MVand125mA.

Ion Source and acceleration tube optics Theprotonbeamisdevelopedusingahighcurrentelectroncyclotronresonance(ECR)microwaveionsourcecoupleddirectlyintoaconventionalelectrostaticaccelerationcolumn.Figure5showsaschematicofthesourceandcolumn.

Page 7: A 1.2MeV, 100mA Proton Implanter

7

Figure 5 Schematic of ion source and acceleration column 

ThesourceconsistsofthreesolenoidelectromagnetssurroundingacircularplasmachamberfedwithH2

gas.Theplasmachamberisexcitedbyupto1kWof2.45GHzmicrowavesfedthroughathreelayervacuumwindow(6)intothechamber.Athreestubtunerandridgedtaperedwaveguidematchtheimpedanceofthemicrowavesystemtotheplasma.Thesourcearchitectureisbasedonpreviousion

sourcesdevelopedatChalkRiver(7)andLosAlamos(8).ThesourcecomponentsaredisplayedontheleftsideofFigure5.Themagneticfieldisadjustedsuchthatthefieldisslightlyaboveresonanceatthevacuumwindow,taperingtothe875Gresonantfieldtowardstheexitofthechamber.Theboundary

elementcodeLorentz‐3D(9)wasusedtomodelthesolenoidalfieldsandoptimizethedesigntoachievethedesiredfieldshape.

Figure 6  PBGUNS simulation of the beam extraction for a 70kV, 100mA beam 

ThebeamisextractedataDCvoltageofupto80kVatcurrentsupto125mA(powersupplylimited)usingatetrodeextractiongeometrywithelectrostaticsuppressionofbackstreamingelectrons.The2D

codePBGUNS(10)wasusedtomodeltheextractiongeometryandoptimizethebeamemittanceandcurrenttobefedintotheaccelerationcolumn.AscreenshotfromPBGUNSisshowninFigure6.

Page 8: A 1.2MeV, 100mA Proton Implanter

8

TheECRsourceprovidesmanyadvantagesoverconventionalionsourcesincludingahighprotonfractionof90%ormore,highcurrentdensityof200mA/cm2andabove,lowgasloadtothevacuum

system,andlongperiodsbetweenserviceswithnofilamentstowearout.Thehighprotonfractionallowsthebeamtobeinjecteddirectlyintotheaccelerationcolumnwithoutpre‐accelerationmagneticanalysis,greatlysimplifyingtheterminaldesign.

Theaccelerationcolumnconsistsof14accelerationstages(the15thbeingthesourceitself),each

capableofacceleratingthebeamupto80keVpersection.Sinceeachstageisdrivenbyanindependenthighvoltagepowersupply,thevoltageoneachelectrodecanbeeasilyadjustedtooptimizethetuneofthesystemandalsoaidintheinitialhighvoltageconditioningofthecolumn.Simulationofthecolumn

opticswereagainstudiedusingLorentz‐3D.Figure7showsaraytracingsimulationoftheaccelerationcolumnincludingspacechargeeffectfora1.2MeV,100mAprotonbeam.

Figure 7  Beam trajectories simulation for a 1.2 MeV, 100mA proton beam using Lorentz‐3D 

Thecolumnhasconsiderablevacuumpumpingareathroughoutitslengthwithameasuredconductanceinexcessof2000L/secforhydrogen.Thishighconductance,combinedwiththelowgasloadfromtheECRsourceallowstheentiresystemtobevacuumpumpedbyturbo‐molecularpumpsatground

potential.EliminatingvacuumpumpsintheSF6environmentatterminalpotentialgreatlysimplifiestheoverallsystem.Aluminaceramicbushingsareusedtofurtherreducethegasloadtothesystemcomparedwithepoxyresinbushings.

Thecentralportionofeachelectrodeismadeoftitaniumwitha100mmaperturetominimizebeam

strikeandprovidegoodpumping.Overlappingstainlesssteelstressringswithembeddedsamariumcobaltpermanentmagnetssurroundthecentralapertureandblockthelineofsightbetweenthebeamandthebushingstopreventcoatingandeventualshortingofthebushing.Theembeddedmagnets

suppressbackstreamingelectronsinthetubebylimitingthetotalenergygainedbyanelectrontooneortwogaps.Thisminimizesthebremsstrahlungradiationaswellastheloadonthehighvoltagesupplies.Themagneticfieldsrotate90degreesateachstagewithmagneticadjustmentsmadeatthe

entranceandexitsuchthatthebeamcenterlineremainsclosetothecolumnaxisthroughoutitslengthandexitsoncenterandparalleltotheaxis.Electrostaticsuppressionispresentattheexitofthecolumn

Page 9: A 1.2MeV, 100mA Proton Implanter

9

tofurthersuppresselectronsenteringthecolumnandprovideafieldfreeregionbeyondthecolumnforbeamneutralization.

Beam Scan System Themagneticscansystemperformstwokeyfunctions.Bymagneticallydeflectingthebeamthrough

approximately80degrees,unwantedspeciesandlowenergycontaminantsareremovedfromthebeampriortoimplant.Second,byincorporatingahighspeedoscillatingscannermagnet,thebeamisrepeatedlysweptacrossthewafersinacontrolledfashion,providingveryhighdoseuniformityforthe

implant(11).Asetofmagneticquadrupolescontroltheshapeofthebeamspotonwafertominimizeoverscan,instantaneousheating,andchanneling.ThemagneticscanningsystemisshownschematicallyinFigure8.

Figure 8  Magnetic scanning system 

Toavoidlargeinstantaneoustemperaturevariationsonthewafercausedbytheveryhighpowerbeam(upto135kWonwafer),thescanspeedcanbeincreasedupto2kHz,almostanorderofmagnitude

fasterthantypicalmagneticallyscannedimplanters..

Process chamber and wafer drum Thesiliconexfoliationprocessrequiresanimplanttemperatureintherange30‐140°Cwithatemperaturevariationacrossthewaferoflessthan20°C.Controlofwafertemperatureinthisrangeisa

formidablechallengewhenbeampowersupto120kWareused.Inordertoreducebeamheatingeffects,thewellknowntechniqueofbeampowersharinghasbeenadoptedbyimplantingthewafersin

Page 10: A 1.2MeV, 100mA Proton Implanter

10

largebatches.Ineachbatch60wafersaremountedontheinsideofa3.1mdiameter,watercooleddrum.Thefacetsofthedrumonwhichthewafersaremountedarecoatedwithacompliantpolymerof

suitablethermalconductivityandthewafersarepressedagainstthismaterialbythecentrifugalforcesgeneratedwhenthedrumisrotated.Thetotalareaexposedtothescannedbeamisapproximately1.5E4cm²andinordertomaintainatemperaturerisebelow120°C,athermalcontactbetweenthe

waferandthedrumofapproximately70mW/cm²°Kisrequired.Thedrumcanrotateaboutitsaxisatspeedsupto400rpmwhichprovidesradialaccelerationsofthewaferupto280Gs.Theresultingclampingforceofthewaferagainsttheelastomericcoatingresultsintherequiredimplanttemperature.

ThedrumandthesystemwhichuniformlyscansanddeflectsthebeamontotheinsideofthedrumisshownschematicallyinFigure1.AndaphotographofthedrumitselfisshowninFigure9.

Figure 9  Photograph of the wafer drum 

Onecomplicationofthisarrangementisthefactthattheangleatwhichthebeamstrikeseachwafervariesfromtheedgeofthewafertothecenterofthewaferasaconsequenceofitsrotationasthe

drumrotates.Thiseffect(12)canresultinundesirablechannelingoftheionsdownaxialorplanarchannelsofthesinglecrystalsiliconandcanoccurinbothrotatingdiskandrotatingdrumprocessstationswhenevertheincidentionbeamaxisisnotparalleltotheaxisifrotation.Itisoftenreferredto

asthe‘coneangle’effect.Inthesystemdescribedthevariationinangleisapproximately±3°andby

Page 11: A 1.2MeV, 100mA Proton Implanter

11

selectingtheappropriateaveragetiltandtwistangleofthewaferswithrespecttothebeam,allaxialandplanarchannelingeffectshavebeenavoided.

Wafer handling Aftereachwaferbatchhasbeenimplantedtotherequireddoseforsuccessfulexfoliation,thewafers

mustbeunloadedfromthedrumandreplacedbyanewbatchofwafersinreadinessforthenextimplant.Duringthisexchangesequencetheionbeamisnotbeingusedanditisthereforeverydesirabletominimizethetimetakenforanywaferhandling.

Tominimizewaferexchangetimesallwaferhandlingtasksaredoneinparallel.Thesystemshownin

Figure10isusedtotransferimplantedwafersfromthedrumandintoastackingelevatorlocatedinavacuumloadlock.Itconsistsofapick/placerobotmechanismandaconveyorbeltsystemasshown.Asecondsystemconsistingofidenticalcomponentsisusedforthereverseprocesswherebyun‐implanted

wafersfromasecondstackingelevatorlocatedinthevacuumloadlockaretransferredbytheconveyorsystemtoasecondpick/placerobotwhichplacesthemonthedrum.Withthisarrangement,onesystemisdedicatedtounloadingimplantedwafersfromthedrumandanothersystemisdedicatedto

loadingun‐implantedwafersontothedrum.Forthishighlyparallelarchitecturetofunctionproperlyallactionsmustbeexecutedwithhighprecision.On‐the‐flyconveyortrajectorymodificationtechniquesandactivewaferalignmentmechanismsareutilizedtocorrectanypositionalvariabilityofthewafers.

Sophisticatedstate‐machinealgorithmsareimplementedsothatanactionstartsassoonastherightconditionsaremet.Atthebeginningofeachaction,sensorsareutilizedtoensurethesoftwarestate

matchesthehardware.Alltheaxesareservo‐controlled;thereforeanydeviationfromthedesignedtrajectoryisimmediatelydetectedandreported.Attheendofeachaction,sensorsareutilizedtoconfirmsuccessfulcompletion.Asaresult,allaxesworkharmoniouslyinparallelwiththestate‐

machinegeneratingtheoptimumsequenceinreal‐time.

Finiteelementanalysis,modalanalysis,andkinematic/dynamicsimulationsoftwarehavebeenusedextensivelytooptimizethehardwareforhighspeedautomation.Complexcalculationshavebeendonetodevelopadvancedhighspeedmotioncontrolprofilesforallaxes.

Theresultisahighspeedwaferhandlingsystemwhichreliablyunloadsandloads60pseudo‐square

wafersfromtheimplantdrumandtransfersthemintothevacuumloadlockin90seconds.

Thestackingelevatorsintheloadlockareinturnunloadedandloadedbyanexternal6axisrobot(13)whichcanbeprogrammedtointerfacewithavarietyoffactorydeliverysystems.

Page 12: A 1.2MeV, 100mA Proton Implanter

12

Figure 10  The three main wafer transfer mechanisms of the drum load or unload system 

Performance results Theimplanterdescribedhasbeensuccessfullyoperatedwith45mAprotonbeamsatenergiesupto800keV.ThesystemisnowbeingshippedtotheTwinCreeksLaminarbasedsolarfactoryinSenatobia,MS.

Conclusion AnovelacceleratorarchitecturehasbeendemonstratedwhichiscapableofproducinghighpowerprotonbeamswithMeVenergiesandcurrentsupto100mA.Thesystemhasbeentestedusingalarge

area,highspeedprocessstationwhichiscapableofmaintainingwafertemperaturesbelow140°Catbeampowersupto120kW.Thisimplanternowprovidesapracticalanduniquesolutionforthefabricationoflowcost,flexible,singlecrystalsiliconsolarcellswithefficiencies≥16%.

References 1.Petti, Christopher.Optimal thickness for crystalline silicon solar cells. (2011).

2.Terrealt, Bernard.Hydrogen blistering of silicon: Progress in fundamental understanding. Phys.Stat.

Sol.(a)Vol.204No.7(2007).pp.2129‐2184.

3.US Patents. 7939812, 7982197, 7989784, 8044374, 8058626, 8089050 

4.Cottereau, E.DC accelerators. ParticleAcceleratorsforMedicineandIndustry,Pruhonice,CzechRepublic,May2001.

Page 13: A 1.2MeV, 100mA Proton Implanter

13

5.HiTek Power. [Online]www.hitekpower.com.

6.Torii, Y., et al.A high‐current density and long lifetime ECR source for oxygen implanters. Rev.Sci.Instrum.Vol.61No.1(1990).pp.253‐255.

7.Taylor, Terence and Wills, John S.C.A high‐current low‐emittance dc ECR proton source. Nucl.Instru.

MethodsPhys.Res.Vol.A309(1991).pp.37‐42.

8.Sherman, Joe, et al.A 75 keV, 140 mA proton injector. Rev.Sci.Instrum.Vol.73No.2(2002).pp.917‐921.

9.Integrated Engineering Software. [Online]http://www.integratedsoft.com/.

10.Boers, Dr. Jack E.PBGUNS (Particle Beam GUN Simulations). [Online]http://www.far‐tech.com/pbguns/index.html.

11.US Patents. 5311208, 5438203 

12.Simonton, R. and Rubin, L.Chapter7.[ed.]J.F.Ziegler.Ion Implantation Science and Technology. 

2004.

13.Staubli. [Online]http://www.staubli.com/.