EndpointWorks. Plasma-Therm LLC

Transcription

1 EndpointWorks Plasma-Therm LLC

2 Outline Introduction Overview of EndpointWorks Endpoint Techniques User Interface - Menus EndpointWorks Modules Input Module Data Source Data Processing Endpoint Detection Building an Endpoint recipe

3 Introduction Overview of EndpointWorks Fully integrated endpoint detection system designed for VLN and ME. Versions for VLR, SLR and 790 systems also available Open Architecture Allows input from various data sources Different hardware options available multi-wavelength spectrometer, laser with a CCD camera etc. Open Ended provision for future expansion using other data sources that produce an analog signal Dedicated computer (Beckoff) running Win XP that accepts input from hardware of choice, processes the data and communicates with the VLN software via Ethernet

4 Introduction Endpoint Techniques Universal front end for all techniques Different hardware interface for different endpoint techniques Multi-wavelength spectrometer Optical emission spectrometry (OEI), Optical emission Interferometry (OEI) Laser with CCD camera Laser interferometry Processing algorithms and endpoint detection modules may differ for different techniques

5 User Interface File Menu Options for loading and saving endpoint recipes

6 User Interface Configuration Menu Data logging and Triggering options

7 Configuration Menu - Logging Saves complete spectral or interferometric data for each process Saves jpeg image of endpoint trace Saves endpoint trace as data file Options can be enabled\disabled using toggle switches

8 Configuration Menu - Triggering Toggle switches to enable/disable automatic endpoint detection on Plasma On or Plasma Off Threshold for triggering Plasma On condition Actual signal amplitude from data source

9 EndpointWorks Modules Input Module Data Source Selection Data Source Provides input device selection (spectrometer, laser, stored data file for playback etc.) Control parameters Averaging, integration time Processing Algorithm Module Algorithm selection for signal processing Algorithm specific settings, tools (smoothing, spectral workbench etc.) Endpoint Detection Module Algorithm selection for analyzing processed data (feature detector, peak counter etc.) Setup of endpoint conditions

10 EndpointWorks Modules Data Source Module Currently configured with 4 different input choices (2 active and 2 file types for offline analysis): Spectrometer 4000 for OES and OEI (Ocean Optics USB Spectrometer 4000) Laser For Interferometry File Spectrum For analyzing logged spectra File Scalar For analyzing logged laser reflectance measurements Choice of Input device decides available options in data processing and endpoint detection modules

11 Data Source Module Spectrometer Input Device Selection sat Spectral Information 1. Integration time Product of 1 and 2 2. No. of spectra to be averaged

12 Data Source Module Spectrometer Omni Gain Should be set as high as possible without the signal saturating (red sat box appears on display. See next slide) No. of samples per scan adjusted to achieve desired scan period Lower values (more samples per second) for shorter processes Higher values (less samples per second) for longer processes Higher the number of files larger is the saved datalog file (if logging is enabled)

13 Data Source Module Spectrometer Signal Saturation Saturated Spectrum. Gain too high

14 Data Source Module Laser Interferometer Input Device Selection Real time output of laser reflectance voltage Data Acquisition Update time

15 Processing Algorithm Module Algorithms process data from input source and output a simple, time varying analog signal The options available depend on type of Data Source Selected. Unsuitable options grayed out Not available for selection Options Available Smoothing When Data Source is Interferometer or File Scalar Spectral Bands When Data source is Spectrometer or File Spectrum

16 Processing Algorithm Module Smoothing Unsmoothed/Raw signal from Data Source Number of points in running point average Raw signal or Differential Smoothed Output Signal sent to the Endpoint Detector Module

17 Data Processing Module Spectral Bands Used with Spectrometer data (OES and OEI) 1. Displays real time spectrum with color coded wavelength regions 2. Assignment of wavelength regions using Spectral Workbench 3. Ability to save snapshots of spectra 4. Algebraic expressions to operate on wavelength regions selected for maximizing the signal to noise ratio with graphical display of formula output 5. Smoothing (running point average) for output of algebraic expression. Note: Smoothing effectively delays the signal. Smoothing should be as low as possible 6. Choice of raw or differential signal

18 Data Processing Module Spectral Bands

19 Spectral Workbench Assignment of Wavelength Regions Spectral Analysis Spectral Spy

20 Spectral Workbench Assignment of Wavelength Regions 1. 5 different wavelengths (R1- R5) can be specified 2. Specific wavelengths (numerical values) can be entered directly in boxes on bottom left 3. If not known, display live spectrum and use vertical color coded lines and move as required. The chosen regions are displayed in the bottom left corner 4. To activate a specific colored line click on appropriate button on the row below the spectrum 4 4 2

21 Spectral Workbench Spectral Analysis 1. Useful for evaluating differences between spectra, either at different points in time in the same run or run to run 2. Used for identifying wavelength regions that change during a process 3. Previously saved spectra can be loaded using the log and snapshot buttons in the top left corner 4. The Index time scale in the top right can be used to playback the entire spectral log from a process 3 4

22 Spectral Workbench Spectral Analysis Modes Delta will display the difference between the two spectra (specrum2 spectrum 1)

23 Spectral Workbench Spectral Analysis Modes Delta% displays the difference between the two spectra (specrum2 spectrum 1) expressed as a percentage

24 Spectral Workbench Spectral Analysis Modes Overlay displays the two spectra simultaneously spectrum 1 is displayed in white, spectrum 2 in orange.

25 Spectral Workbench Spectral Spy Provides access to a library of atomic and molecular species and the wavelengths at which their most prominent emission occurs. Used to assign particular spectral features to species within the plasma Selected by clicking on the?? button -activates vertical red line to the center of spectral display Line can be moved by clicking and dragging Can be used with zoom functions for better wavelength resolution Known species for selected wavelengths displayed

26 Endpoint Detection Modules Feature Detector Used with simple analog signal output by the Process Algorithm module Commonly used with OES Ideal signal is a step function (low signal going high or vice versa) Three tier detection system Two threshold detectors and one peak/valley detector Settings include startup delays and choice of one or more of the detection systems Individual settings for the threshold and peak detectors include signal level, level type, direction and time

27 Endpoint Detection Modules Feature Detector

28 Feature Detector Threshold and Peak Detectors The threshold and peak detectors operate sequentially Threshold 1, Peak, Threshold 2 Each can be individually enabled/disabled Endpoint detected only after condition set by last detector is satisfied

29 Feature Detector Threshold Detector Settings Signal - The threshold detector can operate on either the raw signal or on the differential signal: this selection is made in the signal field. Level - The desired threshold level is entered: this can be either a positive or negative number. The meaning of the value depends on the level type Level type - The level type has options of absolute, relative and percent, which determine how the detector will interpret the value entered as the level value. Absolute - threshold level is numerically equal to the entered value Relative/Percentage -threshold level is referenced to the signal value saved at the end of the delay time Direction Specification of direction from which signal approaches threshold Time Verification time for threshold detector. Used to preventing false triggering

30 Feature Detector Peak/Valley Detector Settings Signal - The threshold detector can operate on either the raw signal or on the differential signal: this selection is made in the signal field. Type Peak or Valley Time Verification Time for peak detector

31 Endpoint Detection Modules Peak Counter Used to determine changes in film thickness when used with an input that generates a sinusoidal signal variation due to interference effects as the film thickness changes Can be used with etching where thickness decreases with time or with deposition where thickness increases with time The data source can be interference with a fixed wavelength laser or a multi-wavelength spectrometer with wavelengths selected

32 Endpoint Detection Modules Peak Counter As film thickness changes, intensity of reflected light changes sinusoidally, where one period of the signal is given by d = λ/2n d = change in film thickness λ = wavelength of light source n = refractive index of film Finds the maxima and minima in the sinusoidal signal; the difference between a successive maximum and minimum is then equal to a half period Only monitors changes in film thickness; the absolute film thickness is not known except in the specific instance of deposition onto a bare substrate where the starting thickness is zero.

33 Endpoint Detection Modules Peak Counter Smoothed Interference signal Peak Counter Settings Real time thickness monitor

34 Endpoint Detection Modules Peak Counter Settings Wafer Start State Switch Bare or Film. Should be set to bare for deposition, where the monitored substrate has no previous film deposited. For etch processes, as a general rule, the switch should be set to Film Desired Thickness Film thickness to be deposited or etched in Ǻ Index of Refraction Refractive Index of film to be etched or deposited This value has to be the refractive index of the material at the wavelength being monitored Wavelength Wavelength at which emission is monitored. For spectra, this value is the center wavelength of the range being monitored Approx. Rate The expected deposition or etch rate. This value is used to calculate the expected interference cycle time. This in turn is used to disable the max/min detector for a short period to avoid false triggering.

35 Building an Endpoint Recipe Spectrometer (OES) Step 1 : Choose Spectrometer as Data Source (Slide 11) Step 2 : Enable Raw data logging (Slide 7) Step 3 : Run desired process to completion (including overetch for etch applications) Step 4 : Choose Spectral Bands and Feature Detector (Slides 18 & 27) Step 4 : Manually start endpoint detection using the green Start button This will save the entire spectral information as a log file Step 5 : Use the saved log file to identify wavelengths of choice and come up with algebraic expressions for endpoint detection (Slide 21) Step 6 : Setup the feature detector using optimal settings (Slide 27) Step 7 : Save all conditions as an endpoint recipe (Slide 5) Step 8 : Use the File Spectrum data source to select saved log file as data source and replay the save spectrum using the endpoint detection recipe for verification (Slide 10)

36 Building an Endpoint Recipe Spectrometer (OEI) Step 1 : Choose Spectrometer as Data Source (Slide 11) Step 2 : Enable Raw data logging (Slide 7) Step 3 : Run desired process to completion (including overetch for etch applications) Step 4 : Choose Spectral Bands and Feature Detector (Slides 18 & 27) Step 4 : Manually start endpoint detection using the green Start button This will save the entire spectral information as a log file Step 5 : Use the saved log file to identify wavelengths of choice and come up with algebraic expressions for endpoint detection (Slide 21) Step 6 : Setup the Peak Counter using optimal settings (Slide 33) Step 7 : Save all conditions as an endpoint recipe (Slide 5) Step 8 : Use the File Spectrum data source to select saved log file as data source and replay the save spectrum using the endpoint detection recipe for verification (Slide 10)

37 Building an Endpoint Recipe Laser Interferometry Step 1 : Choose Interferometer as Data Source (Slide Step 2 : Enable Raw data logging (Slide 7) Step 3 : Run desired process to completion (including overetch for etch applications) Step 4 : Choose Smoothing and Peak Counter (Slides 16 & 33) Step 4 : Manually start endpoint detection using the green Start button This will save the entire interference information as a log file Step 5 : Use File Scalar data source to replay the process and setup parameters for peak detector (Slide 10)