Weir PW Litho Works PEB Critical Feature and

Weir PW & Litho. Works PEB Critical Feature and Process Tuning Products Weir PW and Litho. Works PEB are a new generation of lithography process analysis software providing: • critical feature analysis, • Reticle-to-wafer correlation • behavioral visualization and • automated modeling for process optimization. • Thermal analysis and correlation TEA Systems Corp. 2004 Weir PW and Litho. Works PEB 65 Schlossburg St. Alburtis, PA 18011 610 682 4146 TZavecz@enter. net

Product Highlights Program Weir PW (CD mapping & Process Window) Litho. Works PEB (Thermal bake and chill-plate modeling) Operation Mode Engineering – analytical tools Manufacturing – Daily Monitor templates Automation – APC model engine Platform Windows 2000, XP etc. Data Input Any metrology including CD-sem, Overlay and Ellipsometric measurements. Sens. Array, On. Wafer etc Thermal Sensors Data Storage Microsoft Excel workbooks. Organization “Open” system with easy access to data and analysis results Applications Metrology verification and setup Reticle verification Exposure tool ? setup and analysis. ? Matching ? Uniformity analysis Process ? Optimization ? Window calculation ? MEEF control ? OPC performance analysis ? PEB modeling and Thermal-flow analysis 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 2

Agenda • Weir PW Overview • Data acquisition & setup • Analysis Options ? Sub-set selection ? Metrology analysis • Automated error budget calculation • Element Covariance ? Spatial analysis • Wafer and Field models • Removing systematic components • “What if” yield analysis • Graphic aids ? Process Window Analysis • Focus, Dose and the process window • Multiple element selection • Analysis scope control • Application example ? Reticle to wafer pattern fidelity • Analysis Automation with Weir DM • Summary 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 3

TEA Systems - Product Summary Litho. Works PW Weir Litho. Works PEB Weir DM Automation & Trend Charts • Process Window • Weir PW – Lite • PLUS full-field analysis • Low-cost 2004 Weir PSFM Weir PW • Focus • Focal Plane & Aberration • Full-wafer models • Full-field • Stepper & scanner • Benchmark PSFM, FOCAL, PGM • Process Window • Feature models • Field models • Full-wafer models • Covariance • Precision • Reticle trace Weir PW and Litho. Works PEB TEA Systems Corp. Confidential • Thermal & Feature models • PEB, Chill-Plate • Thermal-flow • Feature correlations • Matching • Weir PW models 4

Data Import • manual data import ? Data is selected from the analysis import screen • Weir can import Binary, ASCII or Excel data formats ? The data is stored in an Excel spreadsheet. ? Uses user-defined variable names from the raw data set. Therma-Wave Optiprobe CD Supports any metrology data including overlay, CD, thickness, thermal etc. • “Test” numbers correlate to information shown on the “Sites” spreadsheet. • The information sheet contains additional setup parameters 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 5

Exposure Layout optimizer • Provides ? Layout fine tuning for centering and parameter updates. • Point-and-click updates of scan, NA, focus, dose coherence etc. ? Basic metrology with statistics and visualization graphics such as histograms, XY-scatter, wafer, field plots, contour and 3 D plots ? Exposure tool library maintenance and selection ? Access to Weir focus, process window and daily monitor modules. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 6

CD Analysis – data sub-set control Data variable and product family currently under analysis Controls for setting data range restrictions Full-wafer data distribution. The mouse can be used to select and cull individual points or full areas. Exposure-Field data distribution. The mouse can be used to select and cull individual sites or full areas. • The “Data Selection” screen provides tools for sub-set selection and data culling 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential • Use the mouse to box the top and bottom (shown) die and then select “Cull” from the pop-up menu. • Data could also have been selected by use of the controls at the top of the screen 7

Viewing selected areas • Drag mouse across die to be viewed, plotted or removed by culling. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 8

Viewing variations across wafer diameter • Use mouse to drag and outline a section. • Plot graphic by x column position • Notice astigmatism and how it increases at wafer’s edge. Xfocus Yfocus and Average 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 9

Box Plots & Color Profiles • Color each data point according to it’s location in the population density. ? Red = 50% median ? Red => Blue as 50% => 0% and 50% goes to 100% • Color peaks at 50% level ? Flyers are displayed in Black This plot displays the population density for the same data. • Box. Plots have been “turned-off” • Three contour profiles have been selected for display. • Data point sizes were set to 90 • Plot background was set to grey 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential • Use the “Display” tab controls. 10

Metrology: Automated Error Budget Variable names derived from data import file • Tab provides two functions ? Precision or automated errorbudget analysis ? Covariance analysis of variable interaction • Precision screen ? Component Precision values are specified in one-sigma estimates. ? CD’s at 5% threshold have about the same performance as those at the 50% threshold for the lot. However, they show greater variance on a field basis. ? As expected, pitch is stable across the field 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 11

Metrology: Variable Covariance • Understanding the covariance of thresholds and measured values is critical. • Below is a covariance matrix of multiple features measured in the data set. • The plot shows the covariance of two selected variables • Graph selections are easily made using the drop-down and check-boxes provided. Top & bottom die sites 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 12

CD 5% and vs SWA & Resist Thickness • CD v SWA (Top) ? Interesting behavior ? Peaks at a CD=73. 566 nm/ 90. 8 deg. • Natural point? ? Anomalous points to right of plot. ? Retrograde profiles? • Would have expected 87 to 89 deg. • CD vs Resist Thickness (bottom) ? Resist seems to be a good predictor for good/bad CD results. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 13

Covariance • Weir Produts are “Open” systems. • Data, model and statistics tables are stored in Excel workbooks. • Covariance analysis can be used for metrology program evaluation, process sensitivity comparison and many other comparative analyses critical to reducing the process window. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 14

Spatial analysis of variables • This analysis provides a complete visualization and modeling interface for data sensitivity to position and tool travel variations. • Raw, modeled and residual data is easily analyzed. • Screen tabs provide: ? Display selection • Graphic selection and data culling including vector, range, bulls-eye, histogram, contour, 3 D and XY scatter. ? Model components • View all or only the contribution of selected elements. • “What-If” analysis shows tool performance if selected components are corrected. ? Component removal • Eliminate the average or modeled field to see the influence of stage travel or scan-direction on the data. • Extended to modeled, average or userspecified inputs. ? Plot formats • View plots of average field, Field-byfield IFD, Maximum values etc. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 15

Spatial comparison of parameters ARC Coat SWA v Wafer Radius • Examine and compare systematic variations across the wafer PR • Provides information of thickness process sensitivity • Systematic errors can be removed from data to enhance analysis resolution. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential CD Threshold 16

Analysis and interactive screen analysis 2004 • Critical model elements are determined using automated culling, singular-value-decomposition and automated coefficient validation. • CD variation after proper removal of selected wafer systematic errors can be very repeatable resulting in enhance control factors. • Cuts along two field exposure rows as shown illustrates the repeatability of the exposure tool for the CD. Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 17

Data Range Reduction for focus Site Location Scaled data value 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential • Focus Matrix Data was reduced to only include the limited focus range of the scanner • Data on this analysis has also eliminated all but the fieldcenter column 18

Working with the data • Removed the center-site value for each field from that fields data. ? Allows us to work with a field average and ignore filed-to-field variation offsets. • Can now see the change in BCD from field-to-field due to process variations. • The variation of each field across the wafer is shown below. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 19

Classic Process Window and more! • Enjoy greater access and ease of use in classic analysis of the process window. • Independent or simultaneous analysis of random and systematic process window. • Multiple graphics for response testing • Test variation across every site in the field (below) 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 20

Full Window Dose sensitivity Field with 121 data sites Field with two data sites • Process window analysis contains extended controls to identify elements, variables and their ranges for analysis. • Go beyond simple multiple feature analysis to examine each tools full-exposure field response to the process window • Mouse sensitive graphics provide complete data selection and drill down capability. • Control response surface viewport by any combination of variables. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 21

Site 1 v 2: OPC response to dose & focus • Here is the dose response of site 1 v 2 near optimum focus (-0. 35) 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 22

Other process window graphics Depth-of-Focus and exposure latitude interactive response analysis 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 23

Application Example • Program: • Sample Analysis • Problem Weir PW Reticle to wafer pattern fidelity ? Wafer metrology of the reticle from the wafer patterns is masked by systematic errors of the process, sub-layers, exposure variability and random errors. • Objective ? Illustrate the first few steps in measuring the Mask Enhancement Factor for a feature family. Import & Layout of Data 2004 Remove selected wafer systematics Remove selected Field systematics (per-field basis) Weir PW and Litho. Works PEB TEA Systems Corp. Confidential Subtract reticle values MEF – Regress residuals 24

Partial Example - Signatures Raw Reticle after Field random error removal Next slide scans data area shown in box for reticle and wafer • Signatures between CD’s and reticle are very close. • The similarity of reticle SEM measurement & feature behavior tracks. • We can map the isolated to dense transformation as function of (size, period, duty cycle) 3 fields from 3 corners of wafer Note similarity of signature. Need to model systematic behavior properly in order to successfully trace feature size from the wafer back to the reticle. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 25

Signature Correlation of reticle to CD data Reticle CD Note scale of abscissa. The average CD value of the field’s center site has been removed so only perturbations remain. Wafer-measured data with the wafer-systematic errors removed (CD@50%) is plotted along with reticle final size measurements (Reticle CD). Reticle data can now be successfully subtracted to obtain MEF performance across the lens and to quantify the anomalous exposure behavior. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 26

Litho. Works PEB • Hot-plate setup, matching and characterization. • Thermal-Models ? Heat-flow ? Across wafer uniformity ? Rise/Fall • Feature Models ? Full-wafer, field, slit & scan ? Weir PW • Correlation ? Model PEB/CD response ? matching A 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential B C D 27

Litho. Works PEB v PW 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 28

Thermal & Feature Models 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 29

Bake versus Feature correlations 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 30

Weir DM: Analysis Automation Any analysis sequence can be automated for implementation in the production environment using the Weir Daily Monitor (DM) • Weir DM Setup Screen • entered through the Weir Engineer “Tools/Daily Monitor Setup” menu item. • Statistics displayed will be specified in the second “Graphics Display” tab. Reticle data or Component (Mean or Best Focus fields) can be removed. 2004 • Statistics and trend charts maintained are specified in the “Statistics Display” tab. Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 31

Running Weir DM The Weir DM is a “stand-alone” program that can also be called from the Weir Engineering interface. Current DM Templates The analysis for a calibration uses the layout specified in the template. Files in the data directory. Sorted alphabetically. Starts the calibration. • pre-selected templates in dropdown listings “One-Click” Analysis • Data files, sorted and pre-selected in a drop-down listing. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 32

Weir DM Trend Charts • Here we observe “Mean” feature settings. Vertical (X) and Horizontal (Y) focus data can also be viewed. • The Weir DM software plots one template chart for each data variable selected. Above is shown the data for the “Mean 3 sigma statistic”. Corresponding data for the Best Focus, or data mean is shown on the right 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential Measurement date shown on the abscissa. 33

Field Focus • Field modeled best focus is changing with wafer number (right). • If we look at vertical (slit) focus, we can see the focus drift until about the 6 th wafer. • Horizontal (scan) focus remains constant. • This is consistent with lens heating effects. Average Lens (slit) focus 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential Scan focus 34

Astigmatism • Astigmatism, also sensitive to lens heating, is shown to settle after the 8 th wafer. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 35

Summary of Weir PW • Any metrology data can be imported ? Weir PW contains a layout optimizer and basic metrology analysis capabilities ? Any user-defined variable names may be used ? Data is stored in Microsoft Excel® worksheet and workbooks • Multiple data sets can be combined and data sub-sets selected for analysis ? Point-and-click mouse interaction for data viewing and culling • “Metrology” analysis ? Automated metrology covariance ? Automated error-budget analysis. • Spatial analysis of variables ? Provides information on variation with scan direction and film thickness variation ? Point-and-click drill down capability for sub-field graphics, data viewing and analysis ? Remove critical systematic components to see the true exposure performance • Process Window analysis and more! ? Extend beyond the single feature/family analysis. ? Examine process window performance variation across the full field of exposure. • Automation ? Weir DM can be used to automate process-monitor points for any variable series. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 36

Extra Wafer Maps of 6 th order parameter variation (next section) These slides (32 -35) illustrate the variation of measured feature size across the wafer and field. More importantly, they illustrate that feature sizes can be modeled as systematic errors across each. Note that when these systematic errors are removed, the residuals track the variation of underlying film thickness (such as. anti-reflective coatings) and thermal-variation (Post Exposure Bake) across the wafer during processing. Slide 34 is a good example. Weir therefore provides a direct method of measuring both feature and film-variation influence on feature size. 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 37

Photoresist All terms 2004 4 th & 6 th only Weir PW and Litho. Works PEB TEA Systems Corp. Confidential Residuals to all terms 38

ARC Wafer Model 2004 Residuals to all Wafer Coefs Weir PW and Litho. Works PEB TEA Systems Corp. Confidential Residuals to Wafer Model Mean Field Removed 39

CD 95% Threshold Wafer Model 2004 Residuals to Wafer Model; mean field removed. Weir PW and Litho. Works PEB TEA Systems Corp. Confidential Wafer model; 4 th & 6 th order only 40

SWA 2004 Weir PW and Litho. Works PEB TEA Systems Corp. Confidential 41