11You’ve got a state-of-the-art SEM……Take nice pictures!(Elimination of gasps and giggles)IR HarveyFall, 2008Artifact EliminationPublished Image!Hitachi Reliability Handbook2Artifact EliminationDirty or charge-sensitivesamples: Focus elsewhere;then capture images from lo-mag to hi-mag(Au nanoparticles on Si; oxide field)Artifact Elimination“Mains lock”50Hz (EU mfg) converted to 60 Hz(parks the beam on each scan)3Artifact EliminationSome SEM images of UEA (Chola Varma)Artifact EliminationThe squared marks here aredifferential charging artifactsfrom prior scans:Image lo-to-hi mag…A very large file full of “junk”…“streaks” indicate charging4Artifact Elimination“Keep your grubby mitts off…”8UnderlyingAssumption:There is much information to begained from the interaction of thebeam with your virgin sample5Artifact EliminationDon’t focus/stig on the area ofinterest. Image a virgin areathen step scan fields in photocapture modeArtifact EliminationA decent image saved in toocoarse file format: Pixelated6Artifact EliminationNot all samples are so easy…Gerton groupC-nanotube onAFM tipArtifact EliminationKey: start with a clean sample…“or else”!Hint: may want to invest(timewise) in plasma clean7Artifact EliminationLayingnano-pavement14EnvironmentalEffectsWhere and how you install yourinstrument is very important……Do a good job with the site survey!8Artifact EliminationSearching for the noise source…Is it vibration? Possibly the pumps?Artifact EliminationFloor-coupled vibration &acoustic noiseTo verify, mount sample directly to pole piece. Thisplaces sample in same vib mode as columnInstrument in exhibition hall9Artifact Elimination…not floor vibration relatedArtifact EliminationGround loop effects (steady-state)Two system components don’t share the sameprimary circuit10Artifact EliminationIntermittent source as afunction of WD or Vacc:8 mm12 mm15 mm20 mmArtifact Elimination60 Hz: Mains lock helps…11Artifact EliminationSynch 60 Hz noise to scan speed1µs 3µs 10µs100nsArtifact Elimination60 Hz EMI and harmonics(try again with a better site survey)Green line is with active EMI field cancellation ON12Artifact EliminationEffect of Active EMI fieldcancellationArtifact EliminationFine-tuning EMI sensor position1325Artifact MiscellanyArtifact EliminationScan distortion at low magBe careful about taking measurements!14Artifact EliminationPrecision Cross-sectioning By hand… By machine (Multi-prep) By FIB…http://www.accurel.comArtifact EliminationCross-sectioning Si-baseddevices; careful else inducemany artifacts Best if correlated to top-view feature Mount to thin glass to enable this Many polishing styles incorporating: Silk Silk with polishing media Ground glass “lapping film”15Artifact EliminationGold coating artifactsKnow what your sample looks like in virgin state, then coat…Post Oxide etchPost Nitride etchas-cleavedlite-BOEstainPost Oxide etchPost Nitride etchArtifact EliminationScan-drift tilt:Coat if you need to measure…(photoresist image drift in slow scan)0.30µm0.25µm 0.35µm 0.40µmDUV Pattern Linearity00.10.20.30.40.50.60.70.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55Mask CD (µm)Drawn CD (µm)Dense ContactsIsolated Contacts16IIBB=100pA=100pAVaccVacc.. : : 1.5kV1.5kVMagMag.. : : x200kx200kFast Scan stabilized imagingUncoated photoresistImaged at E2and scannedat TV rate andthenintegrated overmany frames(fuzzy, but notscan-drifttilted)Slide courtesy of David JoyArtifact Elimination“fractal films” does it help toincrease mag?17Artifact EliminationVacuum-induced deformation==>cracking of surface coatingArtifact EliminationThese are “spheres”. Whereare the shadows?18Artifact EliminationWhy the shadows?Artifact Eliminationcharging…and toomany sigfigs…19Artifact EliminationCharginginhomogeneityArtifact EliminationElectrostaticmirror (chargedsample)20Artifact EliminationImaging charging specimens… Lower probe current (emission, spot, aperture) Lower Vacc Tilt the sample Raise Vacc Scan faster (frame integration) Use BSE detector LV-mode (BSE detector-- classical microscopy) LV-mode (SE detector-- FEI implementation) ESEM mode (higher pressure, water vapor ambient) Ion source (FIB surface implantation, or trickling +veions (or water vapor) in the vicinity of scan -- Zeiss)…Do you know WHY?40SEM is notnondestructiveCase study in IC failure analysis21Artifact EliminationExample from IC failure analysisDeprocessed to Metal-2 layer(RIE depassivation)Liquid Crystal Hot SpotArtifact EliminationContinue Chemical Deprocessing…IMO removed by BOEM2 removed chemically22Artifact EliminationDeprocessing…BPSG removed, poly+ contactsM1 removed / BPSG remainsArtifact EliminationSEM is “destructive to”subsequent (de)processingpost-Wright etch of Si(but 2 areas maskedby prior SEM imaging)Focused here becauseof “crossed features”,took 2-3 micrographs2345Electrostaticartifacts(with dynamicovertones!)Imaging non-conductors On a new SEM this will be thelowest available energy On older machines you mustdecide how low to go before theperformance becomes too poor tobe useful for the purpose intended The goal is to avoid implantingcharge deep beneath the surface. Ifthis is allowed to occur then stableimaging may never be achieved. Step #1 - Set the SEM tothe lowest operatingenergy (2 keV is “good”)Slide courtesy of David JoyHarvey image with Erickson nose hair!24Artifact EliminationArtifact Elimination25Artifact EliminationLo-VacHi-VacArtifact