Laser Physics &Nonlinear OpticsReal-Time Ellipsometry for Studying Cesium-Telluride PhotocathodesPeter van der Slot Rolf Loch Mark Luttikhof Liviu ProdanMesa+ Institute for Nanotechnology

ContentIntroduction Setup and previous results Improvements Verification of setup

IntroductionReal time ellipsometry seems viable to deliver inline information aboutLayer thicknessCompositionKnowledge about these is important forCathode productionCathode performance monitoringInitial measurements performed by us showsVariation in the ellipsometric variablesComplete characterisation of measurement setup required

IntroductionFrom previous measurements we had insufficient information to retrieve the ellipsometric variables and Window preparation chamber was not stress-freeWe did not characterize window of preparation chamberAngles of various optical components could not be measured with sufficient accuracy

QE versus preparation timeLonger preparation times results in lower quantum efficiencies and lead to more robust cathodes (also reported by other groups)Typical value INFN cathodes

Ellipsometry (1)

Ellipsometry (2)Fresnel reflection coefficient:Total amplitude reflection for s and p polarisaton:

Ellipsometry (3)This ratio is complextanY measures the modulus of the amplitude reflection ratioThe phase difference between p- and s-polarised reflected light is given by D

AdvantagesIt measures the ratio of two values so is highly accurate and reproducible, does not need a reference sample, and is not so susceptible to light source fluctuationSince it measures phase, it is highly sensitive to the presence of ultrathin films (down to submonolayer coverage).It provides two pieces of data at each wavelength. More film properties can be determined.

Delta-Psi trajectoriesStarting point (no film):Substrate = silicon

Data analysisRegression AnalysisThe measured data are compared with the data generated from a theoretical model.Unknown parameters in the optical model, such as film thickness or optical constants, are varied to try to produce a best fit to experimental data.Regression algorithms, such as Levenberg-Marquadt, are used to vary unknown parameters and minimize the difference between experimental and model-generated data.Physical parameters are obtained once a good fit is achieved.

Rotating Compensator EllipsometryCompensator (QWP) rotates continuously Sample properties influence reflected beam characteristicsReflected beam characteristics influence intensity after analyzerCorrelation between compensator angle and detector signal gives information about sample propertiesPolarizerAnalyzerHeNe laserQWPFaraday IsolatorBSD1SampleCopper mirrorHWPWindow

Stokes vector & Mueller matrixIntentsity I=(Rp2 +Rs2)I02Polarization angle Polarization ellipticity =a/bRotation directionStokes vectorlightMueller matrixComponent

Rotating CompensatorTotal Mueller matrix of Rotating Compensator Ellipsometer:Solve for intensity:Fourier coefficients Ai and Bi depend on and , and polarizer angles A and P.

Experimental setupPreparation ChamberEllipsometerPolarizerAnalyzerHeNe laserQWPFaraday IsolatorBSSampleCopper mirrorHWPWindow

Cs:Te Photocathode PreparationSubstrate at 120CDeposit Tellurium by Physical Vapor Deposition (PVD) for about 30 minutesDeposit Cesium by PVD until cathode is completedCs and Te mixing produces multiple CsxTey layers

Different Cs deposition rates

Different Te layers

New setup: Rotating Analyzer EllipsometerDetector: Si photodiodeAnalyzer mounted on rotation stage (PA)Sample mounted on rotation stageDiaphragmPolarizer mounted onrotation stage (Pp) set at 45HeNe laserNDM1P0P0 is used to set the input polarizationequal to the p-polarization for the sample

Rotating Analyzer EllipsometerHeNe laserNDM1P0DPADetPPDSample

Improvements on setupRotating Analyzer Ellipsometer (RAE)Motorized rotation stagesConstant rotation velocityAccurate positioning through build-in encodersAccurate determination of relative anglesAlignment procedure for optical elementsFocused on accurate determinations of relative angles

Improvements (continued)Build additional electronics to synchronize angle measurement with detector signalRequires two-channel oscilloscopeTrigger pulse & sample rate determine angle resolution

AlignmentFirst use glass plate (n=1.5225) under Brewster angle to set P0 to p-polarizationThen align PP and PA with respect to the p-polarization (determine angle of fast axis with respect to p-polarization)Insert PP and set angle PP to 45Mount sample and determine 0.Set sample to desired anglePlace PA and align, place detectorPerform measurement

ResultsMeasurement of relative angles with resolution better than 0.05Absolute angle measurement is0.15 for angle of incidence for polarizer and analyzer~ 0.5 for angle of incidence on sampleSample = SOI wafer nSi = 3.87732537 + i 0.0213285228 (632 nm)nSiO2 = 1.45707806

Results (continued)SOI wafer1.5 m Si top layer3 m SiO2 insulation layer522 m Si substrateFrom these values we expect = 24. 7 and = 163.2 (632 nm, i = 60)

Measurement on SOI waferCurve fit to measured data: fit coefficients are used to calculated and

Results for SOI waferFit to function

a0 = 0.0386, a2 = -0.0256, b2 = -0.0284

This results in = 24.25 and = 168.4Theoretical = 24.7 and = 163.2

ConclusionsRotating Analyzer Ellipsometer build and testedObtained ellipsometric parameters agree well with theoretical valuesObtained ellipsometric parameters compare well with commercial ellipsometerStill required:Modification to preparation chamberStress free input/output windowsMechanical setup for accurate alignment/ measurement of angle of incidence (highly sensitive)

AcknowledgementThis work is supported byEuropean Community FP6

Research Infrastructure Activity Structuring the European Research Area programme

CARE, contract number RII3-CT-2003-506395