PicoQuant MultiHarp 150 - Count Rates and Single Photon Statistics

PicoQuantGmbH MultiHarpSoftwareV.3.1.0.0
registrationofonephotoninvolvesthefollowingsteps:first,thetimedifferencebetweenthephotoneventand
thecorrespondingexcitationpulsemustbemeasured.Forthispurposebothopticalsignalsareconvertedto
electricalsignals.Forthefluorescencephotonthisisdoneviathesinglephotondetectormentionedpreviously.
Fortheexcitationpulseitmaybedoneviaanotherdetectorifthereisnoelectricalsyncsignalsuppliedbythe
laserdirectly.Obviously,allconversiontoelectricalpulsesmustpreservetheprecisetimingofthesignalsas
accuratelyaspossible.Theactualtimedifferencemeasurementisdonebymeansoffastelectronicswhich
provideadigitaltimingresult.Thisdigitaltimingresultisthenusedtoaddressthehistogrammemorysothat
eachpossibletimingvalue corresponds toone memorycell orhistogramchannel. Finallythe addressed
histogramcellisincremented.Allstepsarecarriedoutbyfastelectronicssothattheprocessingtimerequired
foreachphotoneventisasshortaspossible.Whensufficientcountshavebeencollected,thehistogram
memorycanbereadout.Thehistogramdatacanthenbeusedfordisplayande.g.,fluorescencelifetime
calculation.Inthefollowingwewillexpandonthevariousstepsinvolvedinthemethodandrelatedissuesof
importance.
2.1. Count Rates and Single Photon Statistics
Itwasmentionedthatinthepastitwasnecessarytomaintainalowprobabilityofregisteringmorethanone
photonpercycle.Thisconditionensuredthatthehistogramofphotonarrivalsrepresentsthetimedecayone
wouldhaveobtainedfromasingleshottime–resolvedanalogrecording.Thereasonforthisisbrieflythefollow-
ing:Duetodeadtimesofdetectorandelectronicslastingatleastsometensofnanosecondsafteraphoton
event,historicalTCSPCsystemscouldregisteronlyonephotonperexcitation/emissioncycle.Ifthenthenum-
berofphotonsoccurringinoneexcitationcycleweretypically>1,thesystemwouldveryoftenregisterthefirst
photonbutmissthefollowingoneormore.Thiswouldleadtoanover-representationofearlyphotonsinthehis-
togram,aneffectcalled‘pile-up’.Thisleadstodistortionsofthefluorescencedecay,typicallymakingthefluo-
rescencelifetimeappearshorter.Itwasthereforecrucialtokeeptheprobabilityofcycleswithmorethanone
detectedphotonsufficientlylow,unlessonetoleratessomeerrorand/orcorrectsforitindataanalysis.
Toquantifythe“safe”countratelimitfortraditionalTCSPC,onehadtosetacceptableerrorlimitsandapply
somemathematicalstatistics.Forpracticalpurposesonewouldapplyaruleofthumb:Inordertomaintainsin-
glephotonstatistics,onaverageonlyonein20to100excitationpulsesshouldgenerateacountatthedetector.
Inotherwords:theaveragecountrateatthedetectorshouldbeatmost1%to5%oftheexcitationrate.Using
e.g.,adiodelaserfromPicoQuant'sPDLSeries,pulsedat80MHzrepetitionrate,theaveragedetectorcount
ratewouldthennotexceed4Mcps.Typicallyonewantshighcountrates,inordertoacquirefluorescencede-
cayhistogramsquickly.Thisisofparticularimportancewheredynamiclifetimechangesorfastmoleculetransi-
tionsaretobestudiedorwherelargenumbersoflifetimesamplesmustbecollected(e.g.,inimagescanning).
Thisiswhyhighlaserrates(suchas40or80MHzforthePDLSeries)areimportant.PMTscansafelyhandle
TCSPCcountratesofover10Mcps.MostactivelyquenchedSPADsmayalsooperateupto10Mcpsbutsome
typessufferresolutiondegradationwhenoperatedtoofast.OldNIMbasedTCSPCelectronicscouldonlyhan-
dleamaximumof50to500kcps,whilenewerintegratedTCSPCelectronicstypicallyreachcountratesof5to
40Mcps.WiththeMultiHarpandmodernHybridPhotodetectors(HPD)countratesupto78Mcpscanbecol-
lected.
Itisworthnotingthatthephotonarrivaltimesaretypicallyrandomsothattherecanbeburstsofhighcountrate
andperiodsoflowcountrates.Burstsmayexceedtheaveragerate.Theaverage(sustained)ratetheinstru-
mentcandealwithisimportantwhenlossescannotbetolerated,notablyintimetaggingapplications.Dueto
thehighbandwidthof USB3.0 the MultiHarp canhandlesustained timetagging ratesashighas 80Mcps
throughthisinterface.TheexternalFPGAinterfaceofthenewMultiHarp160canevenhandle1.6Gcps.When
comparingcountratesconsideredhereandelsewherepleasepayattentiontothedetails.Thespecificationsfor
TCSPCsystemsofteninterprettheirmaximumcountratesdifferently.Thisiswhyinthiscontextdeadtimeis
alsoofinterest.Itdescribesthetimethesystemcannotregisterphotonswhileitisprocessingapreviousphoton
event.Thetermisapplicablebothtodetectorsandelectronics.TheMultiHarphasanextremelyshortdeadtime
ofabout650ps,imposingthesmallestlossesamongcomparableinstrumentstoday.Theshortdeadtimeofthe
MultiHarptogetherwithitsmulti-stopcapabilitynowallowsmeasurementscenarioswheretheclassicpile-up
limitisnolongercritical.ThisisofspecialinterestforveryfastFLIMmeasurementsandPicoQuant'sconceptof
rapidFLIMexploitstheidea.UsinganHPD(e.g.,PicoQuantPMAHybrid)thatpermitscountratesashighas80
McpsonecanobtainFLIMimagesatunprecedentedspeed.Whileclassicpile-upduetodeadtimeisnolonger
anissue,onemustnowconsidertheeffectofpulse-pile-up.Thiseffectoccurswhenphotonsarriveatthedetec-
torwithveryshorttemporalspacing,suchthatthedetectorcannolongerproduceseparateoutputpulses.This
mergingofdetectorpulsesleadstoanothertypeofartifactinthedecayhistograms.Anumericalcorrectionof
theseartifactsduringdataanalysisensuresthatlifetimesandamplitudesarestillobtainedcorrectlyforveryfast
quantitativemeasurements.Seesection2.5orwww.picoquant.comforrelatedpublications.
Page9