Scanning Surface Inspection System with Defect …

1 Scanning Surface Inspection System with Defect -review SEM and Analysis System Solutions 78 Scanning Surface Inspection System with Defect -reviewSEM and Analysis System SolutionsOVERVIEW: Data obtained from wafer- Surface Inspection tools are limitedto Defect maps, Defect number, etc. In any case, in order to swiftly executemeasures to decrease the amount of defects , it is very important to determinewhere the defects are being generated. Accordingly, more concreteinformation on defects , such as shape and constitution data, is becomingmore necessary. Hitachi High-Technologies Corporation is therefore offeringa line-up of wafer- Surface Defect Inspection tools and review scanningelectron microscopes, and Hitachi Kenki FineTech Co.

2 , Ltd. is offering aline-up of atomic-force microscopes. In the present work, a System ofimproving yields by providing Surface -analysis information aboutunpatterned silicon wafers called smart root cause analysis wasdeveloped. From now onwards, although Defect observation based on theoutput coordinates from Surface Inspection of unpatterned silicon wafers isextremely difficult owing to a lack of an alignment basis, the developedsystem can address this difficulty and make a key contribution to crucialdefect countermeasures with high OtaMasayuki HachiyaYoji IchiyasuToru KurenumaINTRODUCTIONAPPLICATIONS of unpatterned silicon-wafersubstrates have two broad classifications.

3 The first(fundamental) application is the Inspection for startingmaterial before a circuit pattern is formed on thesubstrate Surface , and a semiconductor device isprocessed. In the second, defects generated bysemiconductor device-processsing equipment namely, film-deposition equipment, etching tools,CMP (chemical-mechanical polishing), and cleaningequipment are investigated. In the former case, thewafer is distinguished as a prime wafer, and in theFig. 1 Linkage of SurfaceInspection System and Defect -review SEM based ondefect-location coordinates obtainedfrom SSIS and by observation,classification, and EDS elementalanalysis or 3D observation withAFM, data that contributes toimproving yields of customers production systems are image (2D) EDS analysis Smart Root Cause Analysis From Defect detection to cause pinpointingWork flow of measuresfor reducing Surface defectsInspectionDefectreviewMeasureGoal Information on the countermeasure is processCause toolPrompt analysis of Defect mechanismDefect-review SEM Wide-area AFM Wafer-surfacedefect-detection toolDefect mapAFM image (3D) Defect -generationcause pinpointingSSIS.

4 Scanning Surface Inspection systemEDS: energy dispersive X-ray spectrometerSEM: Scanning electron microscopeAFM: atomic force microscopeHitachi Review Vol. 55 (2006), No. 2 79such as moving wafers under actual process conditionssuch as film formation ( dummy deposition) (seeTable 1).In the rest of this paper, equipment for detectingwafer defects is described, and the above-mentionedanalysis methods are explained (see Fig. 1).SSIS ( Scanning Surface INSPECTIONSYSTEM)As regards defects on unpatterned silicon wafersubstrates, the principle of laser scattering is appliedin detecting the defects . The detection principle utilizedin Hitachi High-Technologies s SSIS is shownschematically in Fig. 2.

5 This Defect -detection principleis as follows: a laser is illuminated onto the wafersurface from above, scattered light from a Defect iscollected by a receiving lens, and the scattered light isconverted to an electrical signal by a detector. Thewafer is set on the rotating stage, and by moving thestage in the radial direction while it rotates, the wholewafer Surface can be inspected at high speed. And byfixing an encoder to the stage, positional data of thewafer Defect can be attainable sensitivity of our latest model ofSSIS is 36 nm on a bare wafer OF SSIS AND Defect -REVIEWTOOLSD efect-review SEMA lthough information obtained from SSIS isrestricted to Defect maps, Defect number, etc.

6 , in orderto swiftly carry out measures for reducing defects , itis important to find out where the defects weregenerated. This location is thus determined by defectinspection. Based on the location coordinates obtainedfrom this Inspection , Defect observation by high-latter case, the wafer is distinguished as a test wafer, a process monitor wafer or a dummy wafer. As regards the amount of test wafers consumed, Defect measurement using test wafers is widelyperformed with about the same number of wafers asplatform wafers used in processing line with the trend of device scaling-down, the objectsize in required Defect measurement is getting smalleryear by year, and detection and analysis of defects aregetting even the Defect -analysis method used for testwafers, analyses of wafer-transport particle andprocess-generated particle are carried out.

7 Analysis onwafer-transport particle PWP (particles per wafer pass)is the operation of investigating particle that becomesattached to a wafer when it is in actual processingequipment and when it is transferred between pro-cesses. Analysis on process-generated particle PIDs(process-induced defects ) is the operation ofinvestigating defects generated after wafer processingTABLE 1. Object of Surface - Defect Analysis, and ObjectUnpatterned Wafer and UserSurface- Defect analysis requires many users involved insemiconductor 2 Overview of Detection laser is illuminated on the barewafer, and defects are detected athigh moving stageLaser illuminationWaferWaferShift from wafer center to perimeterDefectShift from wafer center to perimeterRevolutionWafer position dataScatteringReceiving lensDetectorDefect detected signalA/D converter( Defect XY coordinates)Signal processingData processingInspection-results display ( Defect map)A/D.

8 Analog-to-digitalPWP measurement(wafer transportation dust) PID measurement(process generated dust) Surface - Defect measurement of silicon wafer Bare waferCoated waferBare waferDevice manufacturerEquipment and materials supplierWafer manufacturerAimObject unpatternedwaferUserPWP: particle per wafer passPIP: process-induced defectScanning Surface Inspection System with Defect -review SEM and Analysis System Solutions 80magnification tools, such as SEM ( Scanning electronmicroscope), is an extremely effective these circumstances, Hitachi s latest Defect -review SEMs were developed for reviewing at highefficiency in a short time the vast amounts ofinspection results obtained from the improved-sensitivity Defect - Inspection tools in use over Regarding Defect Review ofUnpatterned WafersThe four main methods used in SSIS and defectanalysis are given as follows.

9 (1) Defect detection by SSIS, and output of Defect -location coordinates(2) Coordinate matching alignment of SSIS tool andreview SEM (see Fig. 3)(3) Defect search and observation by review SEM(4) Defect analysis by EDS (energy dispersive X-rayspectrometer) ( elemental analysis)While these methods are recognized as effectivefor Defect analysis, in execution of Defect observation,there is a barrier to reacquiring defects by the methodson the SEM side, (2) and (3) the case of a patterned wafer, a chip die origin isused as an alignment point, and since the alignmentcan relatively easily come off, defects are easy , with an unpatterned silicon wafer, sincea unique alignment method is used for fixing the profileof a wafer, it is difficult to get high-precision matchingbetween the coordinate System on the Inspection -toolside and that on the review SEM Optical-microscope FunctionTo address the above-described issues, HitachiHigh-Technologies has fitted its Defect -review SEMswith a function for wide-view optical microscope,thereby vastly improving the efficiency of observa-tion of bare wafers with large alignment error (seeFigs.)

10 4 and 5).It is currently possible to automatically pick updefects with a size of 70 nm in a 600- m2 FOV (fieldof view) on a wafer Surface . This gives enough leewayfor a condition where detection of total alignment errorwith standard Defect -detection tools of less than 200 m, and matching with each detection tool are possible(see Fig. 6).Fig. 3 Example ofAlignment of BareWa f e total of five points(three points on theperiphery and two atthe mouth of a V-notch) are 4 Making Observation of Bare Wafer Surface MoreEfficient by Wide-view Optical efficiency is significantly improved (about 100times).Fig. 5 FOV of Wide-FOV Optical can be searched for over a 6 Example of Defect Observed by Using Wide-FOVO ptical-microscope , defects at the macro level are found by opticalmicroscope, then micro observations are done by SEM.