layersonthewafersurfaces.WehavenotattemptedamagnitudeoftheerrorderivedonthebasisofFig.3Bisdetailedinvestigationofathree-layer(oxide-silicon-ox-approximately1015atoms-cm-3.SinceCzochralskisil-ide)specimenduetothevariabilityinsuchparametersiconwafersgenerallycontainfrom10TMto1017carbonasoxidethicknessesandindicesofrefraction.However,atoms-cm-3,thiserrorcanrangefrom1to10%ofthewehaveobservedthattheamplitudeofthetertiarycarbonconcentration.interferogramsdoesdependupontheactualreflectanceofthewafersurface.II.CONCLUSIONTheobservedinterferogramscanbecorrectedbeforeWehavedemonstratedtheexistenceoftertiaryinter-transformationinordertoeliminateFabry-Perotfringes.ferogramsandhaveshownthattheyareproducedbyHirschfeldandMantz5havedevelopedaprocedureformultiplereflectionsbothbetweenthesurfacesofaspec-eliminatingtheeffectsofsecondaryinterferogramsbyimenandbetweenthewaferandthemirrorsoftheremovingtheinterferometersignalfromtheregionininterferometer.Wehavealsodemonstratedamethodforwhichtheamplitudeoftheextraneousinterferogramiscorrectingtheobservedinterferograminordertomea-greaterthantheamplitudeofthemaininterferogram.surethespecimenabsorptivitymoreaccurately.TheThesignalinthisregionisthenreplacedbythesignalmagnitudeoftheerrorintroducedbynotcorrectingforfromthesameregionofaninterferogramobtainedwithtertiaryandsecondaryinterferogramsincreasesastheadifferentspecimenthickness.specimenthicknessisreduced.SincethetrendintheWeusedthisprocedureforthesecondaryandthesemiconductorindustryistowardmeasuringimpuritytertiaryinterferogramsindeterminingthecarboncon-contentsofthinnerwafers,thepotentialerrorsintro-tentofa0.869-mmthicksiliconwaferbymeasuringtheducedbyboththesecondaryandtertiaryinterferogramsintensityofthesubstitutionalcarbonabsorptionlineatwillincrease.607607cm-1atroomtemperature.Ahigh-purity2.0-mmthickvacuumfloat-zonedwaferwasusedasareference.Fig.3showstheeffectofthetertiaryinterferogram.In1.D.G.MeadandS.R.Lowry,Appl.Spectrosc.34,167(1980).2.2.R.J.B